diff --git a/.github/workflows/cmake-multi-platform.yml b/.github/workflows/cmake-multi-platform.yml index e162571..9f1cf54 100644 --- a/.github/workflows/cmake-multi-platform.yml +++ b/.github/workflows/cmake-multi-platform.yml @@ -92,6 +92,7 @@ jobs: with: version: ${{ env.QT_VERSION }} cache: true + cache-key-prefix: install-qt-action-${{ github.job }}-${{ matrix.build_type }} arch: linux_gcc_64 - name: Configure @@ -104,7 +105,7 @@ jobs: -DCMAKE_PREFIX_PATH="$QT_ROOT_DIR" - name: Build - run: cmake --build build --parallel 2 + run: cmake --build build --parallel - name: Stage AppDir run: | @@ -198,6 +199,7 @@ jobs: with: version: ${{ env.QT_VERSION }} cache: true + cache-key-prefix: install-qt-action-${{ github.job }}-${{ matrix.build_type }} arch: linux_gcc_arm64 - name: Configure @@ -210,7 +212,7 @@ jobs: -DCMAKE_PREFIX_PATH="$QT_ROOT_DIR" - name: Build - run: cmake --build build --parallel 2 + run: cmake --build build --parallel - name: Stage AppDir run: | @@ -271,19 +273,39 @@ jobs: path: Brokkr-linux-arm64${{ matrix.suffix }}.AppImage # ═══════════════════════════════════════════════════ - # Windows x64 (NSIS installer + portable zip) + # Windows x64 + arm64 (NSIS installer + portable zip) # ═══════════════════════════════════════════════════ windows_installer: - name: Windows installer x64 (${{ matrix.build_type }}) + name: Windows installer ${{ matrix.artifact_arch }} (${{ matrix.build_type }}) needs: [prepare] if: always() && (needs.prepare.result == 'success' || needs.prepare.result == 'skipped') - runs-on: windows-2022 + runs-on: ${{ matrix.runner }} strategy: matrix: include: - - build_type: Release + - runner: windows-2022 + artifact_arch: x64 + msvc_arch: x64 + qt_arch: win64_msvc2019_64 + build_type: Release suffix: "" - - build_type: Debug + - runner: windows-2022 + artifact_arch: x64 + msvc_arch: x64 + qt_arch: win64_msvc2019_64 + build_type: Debug + suffix: "-dbg" + - runner: windows-11-arm + artifact_arch: arm64 + msvc_arch: arm64 + qt_arch: win64_msvc2019_arm64 + build_type: Release + suffix: "" + - runner: windows-11-arm + artifact_arch: arm64 + msvc_arch: arm64 + qt_arch: win64_msvc2019_arm64 + build_type: Debug suffix: "-dbg" steps: @@ -293,7 +315,7 @@ jobs: - uses: ilammy/msvc-dev-cmd@v1 with: - arch: x64 + arch: ${{ matrix.msvc_arch }} - name: Install NSIS + Ninja shell: powershell @@ -304,7 +326,9 @@ jobs: with: version: ${{ env.QT_VERSION }} cache: true - arch: win64_msvc2019_64 + cache-key-prefix: install-qt-action-${{ github.job }}-${{ matrix.artifact_arch }}-${{ matrix.build_type }} + host: windows + arch: ${{ matrix.qt_arch }} - name: Configure shell: bash @@ -317,7 +341,7 @@ jobs: - name: Build shell: bash - run: cmake --build build --parallel 2 + run: cmake --build build --parallel - name: Package (NSIS installer) shell: bash @@ -330,27 +354,27 @@ jobs: exit 1 } installer="$(ls -1 *.exe | head -n1)" - mv "$installer" ../Brokkr-windows-x64${{ matrix.suffix }}.exe + mv "$installer" ../Brokkr-windows-${{ matrix.artifact_arch }}${{ matrix.suffix }}.exe - name: Package (portable zip) shell: bash run: | set -euo pipefail - mkdir -p portable/brokkr - cp build/brokkr.exe portable/brokkr/ - "$QT_ROOT_DIR/bin/windeployqt.exe" portable/brokkr/brokkr.exe + install_root="$PWD/portable/brokkr" + rm -rf "$install_root" + cmake --install build --prefix "$install_root" --config ${{ matrix.build_type }} cd portable - 7z a -tzip ../Brokkr-windows-x64${{ matrix.suffix }}.zip brokkr/ + 7z a -tzip ../Brokkr-windows-${{ matrix.artifact_arch }}${{ matrix.suffix }}.zip brokkr/ - uses: actions/upload-artifact@v4 with: - name: Brokkr-windows-x64${{ matrix.suffix }}.exe - path: Brokkr-windows-x64${{ matrix.suffix }}.exe + name: Brokkr-windows-${{ matrix.artifact_arch }}${{ matrix.suffix }}.exe + path: Brokkr-windows-${{ matrix.artifact_arch }}${{ matrix.suffix }}.exe - uses: actions/upload-artifact@v4 with: - name: Brokkr-windows-x64${{ matrix.suffix }}.zip - path: Brokkr-windows-x64${{ matrix.suffix }}.zip + name: Brokkr-windows-${{ matrix.artifact_arch }}${{ matrix.suffix }}.zip + path: Brokkr-windows-${{ matrix.artifact_arch }}${{ matrix.suffix }}.zip # ═══════════════════════════════════════════════════ # macOS arm64 @@ -385,6 +409,7 @@ jobs: with: version: ${{ env.QT_VERSION }} cache: true + cache-key-prefix: install-qt-action-${{ github.job }}-${{ matrix.build_type }} arch: clang_64 - name: Configure @@ -402,7 +427,7 @@ jobs: sed -i '' 's/ -framework AGL//g; s/-framework AGL //g; s/-framework AGL//g' build/build.ninja - name: Build - run: cmake --build build --parallel 3 + run: cmake --build build --parallel - name: Deploy + Zip if: matrix.build_type == 'Release' @@ -459,6 +484,7 @@ jobs: with: version: ${{ env.QT_VERSION }} cache: true + cache-key-prefix: install-qt-action-${{ github.job }}-${{ matrix.build_type }} arch: clang_64 - name: Configure @@ -476,7 +502,7 @@ jobs: sed -i '' 's/ -framework AGL//g; s/-framework AGL //g; s/-framework AGL//g' build/build.ninja - name: Build - run: cmake --build build --parallel 3 + run: cmake --build build --parallel - name: Deploy + Zip if: matrix.build_type == 'Release' @@ -522,6 +548,11 @@ jobs: path: artifacts merge-multiple: true + - name: Remove debug artifacts from release bundle + run: | + set -euo pipefail + find artifacts -maxdepth 1 -type f -name '*-dbg*' -delete + - name: Generate changelog body run: | NEW="${{ inputs.new_tag }}" diff --git a/CMakeLists.txt b/CMakeLists.txt index 9a01007..f71342e 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -1,5 +1,5 @@ cmake_minimum_required(VERSION 3.22) -project(Brokkr LANGUAGES C CXX VERSION 1.3.4) +project(Brokkr LANGUAGES C CXX VERSION 1.4.5) set(CMAKE_CXX_STANDARD 23) set(CMAKE_CXX_STANDARD_REQUIRED ON) @@ -112,6 +112,7 @@ if (LINUX OR CMAKE_SYSTEM_NAME STREQUAL "Linux") add_library(brokkr-platform INTERFACE) target_compile_definitions(brokkr-platform INTERFACE BROKKR_PLATFORM_LINUX) target_sources(brokkr-platform INTERFACE + src/platform/posix-common/app_dirs.cpp src/platform/posix-common/signal_shield.cpp src/platform/posix-common/single_instance.cpp src/platform/posix-common/tcp_transport.cpp @@ -127,6 +128,7 @@ elseif (WIN32) NOMINMAX ) target_sources(brokkr-platform INTERFACE + src/platform/windows/app_dirs.cpp src/platform/windows/signal_shield.cpp src/platform/windows/single_instance.cpp src/platform/windows/sysfs_usb.cpp @@ -138,6 +140,7 @@ elseif (APPLE) add_library(brokkr-platform INTERFACE) target_compile_definitions(brokkr-platform INTERFACE BROKKR_PLATFORM_MACOS) target_sources(brokkr-platform INTERFACE + src/platform/posix-common/app_dirs.cpp src/platform/posix-common/signal_shield.cpp src/platform/posix-common/single_instance.cpp src/platform/posix-common/tcp_transport.cpp @@ -169,6 +172,7 @@ target_sources(brokkr-lib INTERFACE src/protocol/odin/flash.cpp src/protocol/odin/group_flasher.cpp src/protocol/odin/pit_transfer.cpp + src/app/md5_xxh3_cache.cpp src/app/md5_verify.cpp ) @@ -244,8 +248,10 @@ endif() # ── end macOS framework fixup ───────────────────────────────────── if (WIN32) - add_executable(brokkr WIN32 + add_executable(brokkr src-gui/main_gui.cpp + src/app/cli_mode.hpp + src/app/cli_mode.cpp src-gui/brokkr_wrapper.hpp src-gui/brokkr_wrapper.cpp assets/icon.rc @@ -256,6 +262,8 @@ elseif (APPLE) set_source_files_properties(assets/brokkr.icns PROPERTIES MACOSX_PACKAGE_LOCATION "Resources") add_executable(brokkr MACOSX_BUNDLE src-gui/main_gui.cpp + src/app/cli_mode.hpp + src/app/cli_mode.cpp src-gui/brokkr_wrapper.hpp src-gui/brokkr_wrapper.cpp assets/brokkr.qrc @@ -264,6 +272,8 @@ elseif (APPLE) else() add_executable(brokkr src-gui/main_gui.cpp + src/app/cli_mode.hpp + src/app/cli_mode.cpp src-gui/brokkr_wrapper.hpp src-gui/brokkr_wrapper.cpp assets/brokkr.qrc @@ -300,21 +310,31 @@ target_link_libraries(brokkr PRIVATE set_target_properties(brokkr PROPERTIES AUTOMOC ON AUTORCC ON) -if (MSVC) +if (MSVC AND CMAKE_CXX_COMPILER_ID STREQUAL "MSVC") target_compile_options(brokkr PRIVATE /Zc:__cplusplus) endif() +set(BROKKR_WINDOWS_ARM64 OFF) +if(WIN32 AND (CMAKE_SYSTEM_PROCESSOR MATCHES "^(ARM64|AARCH64|arm64|aarch64)$" + OR CMAKE_VS_PLATFORM_NAME STREQUAL "ARM64")) + set(BROKKR_WINDOWS_ARM64 ON) +endif() + if(WIN32) - get_target_property(_qmake_exec Qt6::qmake IMPORTED_LOCATION) - get_filename_component(_qt_bin_dir "${_qmake_exec}" DIRECTORY) - find_program(WINDEPLOYQT_EXEC windeployqt HINTS "${_qt_bin_dir}") - if(WINDEPLOYQT_EXEC) + if(BROKKR_WINDOWS_ARM64) + message(STATUS "Skipping post-build windeployqt on Windows ARM64; install-time Qt deployment will be used instead.") + else() + get_target_property(_qmake_exec Qt6::qmake IMPORTED_LOCATION) + get_filename_component(_qt_bin_dir "${_qmake_exec}" DIRECTORY) + find_program(WINDEPLOYQT_EXEC windeployqt HINTS "${_qt_bin_dir}") + endif() + if(NOT BROKKR_WINDOWS_ARM64 AND WINDEPLOYQT_EXEC) add_custom_command(TARGET brokkr POST_BUILD COMMAND "${WINDEPLOYQT_EXEC}" - ARGS --no-translations --compiler-runtime "$" + ARGS --no-translations --no-compiler-runtime "$" COMMENT "Running windeployqt to deploy Qt libraries..." ) - else() + elseif(NOT BROKKR_WINDOWS_ARM64) message(WARNING "windeployqt not found! You will need to copy DLLs manually.") endif() endif() @@ -359,7 +379,7 @@ target_compile_definitions(brokkr PRIVATE qt_generate_deploy_app_script( TARGET brokkr - OUTPUT_SCRIPT deploy_script + FILENAME_VARIABLE deploy_script NO_UNSUPPORTED_PLATFORM_ERROR ) install(SCRIPT ${deploy_script}) @@ -378,6 +398,14 @@ add_executable(test_md5 target_include_directories(test_md5 PRIVATE src) add_test(NAME md5 COMMAND test_md5) +add_executable(test_md5_xxh3_cache + tests/test_md5_xxh3_cache.cpp + src/app/md5_xxh3_cache.cpp +) +target_include_directories(test_md5_xxh3_cache PRIVATE src) +target_link_libraries(test_md5_xxh3_cache PRIVATE fmt::fmt-header-only) +add_test(NAME md5_xxh3_cache COMMAND test_md5_xxh3_cache) + add_executable(test_endian tests/test_endian.cpp) target_include_directories(test_endian PRIVATE src) add_test(NAME endian COMMAND test_endian) @@ -388,7 +416,7 @@ set(CPACK_PACKAGE_VERSION "${PROJECT_VERSION}") set(CPACK_PACKAGE_DESCRIPTION_SUMMARY "Open-Source Implementation of Samsung Odin3") set(CPACK_PACKAGE_CONTACT "roynatech@gmail.com") set(CPACK_GENERATOR "ZIP;TGZ") -set(CPACK_PACKAGE_INSTALL_DIRECTORY "${PROJECT_NAME}-v${CPACK_PACKAGE_VERSION}") +set(CPACK_PACKAGE_INSTALL_DIRECTORY "${PROJECT_NAME}") # Resource files – guarded so missing files don't break CPack if(EXISTS "${CMAKE_SOURCE_DIR}/LICENSE") @@ -423,7 +451,10 @@ if(WIN32) set(CPACK_NSIS_INSTALLED_ICON_NAME "bin/brokkr.exe") # Bundle VC++ Redistributable - if(CMAKE_SIZEOF_VOID_P EQUAL 8) + if(CMAKE_SYSTEM_PROCESSOR MATCHES "^(ARM64|AARCH64|arm64|aarch64)$" + OR CMAKE_VS_PLATFORM_NAME STREQUAL "ARM64") + set(_vcredist_name "vc_redist.arm64.exe") + elseif(CMAKE_SIZEOF_VOID_P EQUAL 8) set(_vcredist_name "vc_redist.x64.exe") else() set(_vcredist_name "vc_redist.x86.exe") @@ -437,6 +468,8 @@ if(WIN32) install(FILES "${_vcredist_path}" DESTINATION redist) set(CPACK_NSIS_EXTRA_INSTALL_COMMANDS " ExecWait '\\\"$INSTDIR\\\\redist\\\\${_vcredist_name}\\\" /install /quiet /norestart' + SetOutPath '$INSTDIR' + RMDir /r /REBOOTOK '$INSTDIR\\redist' ") else() message(WARNING "VCToolsRedistDir set but ${_vcredist_name} not found at: ${_vcredist_path}") @@ -455,5 +488,6 @@ endif() set(CPACK_PACKAGE_FILE_NAME "${PROJECT_NAME}-${PROJECT_VERSION}-${CMAKE_SYSTEM_NAME}-${CMAKE_SYSTEM_PROCESSOR}" ) +set(CPACK_VERBATIM_VARIABLES TRUE) include(CPack) diff --git a/README.md b/README.md index 817b007..d9464eb 100644 --- a/README.md +++ b/README.md @@ -1,20 +1,13 @@ # Brokkr Flash -![brokkr](https://raw.githubusercontent.com/Gabriel2392/brokkr-flash/main/assets/brokkr.jpg) +A modern, cross-platform Samsung device flashing utility written in C++23. -A modern, cross-platform Samsung device flashing utility written in C++23. Brokkr provides a command-line interface for flashing firmware partitions to Samsung Android devices using the ODIN protocol. +## Features (why is it better than etc lol) -## Features - -- **Multi-device support**: Flash multiple Samsung devices simultaneously -- **Multiple partition types**: Support for AP (Application Processor), BL (Bootloader), CP (Cellular Processor), CSC (Consumer Software Customization), and USERDATA partitions -- **In-place Decompression support**: Built-in LZ4 decompression for efficient firmware transfers -- **Wireless flashing**: Support for TCP-based flashing for Galaxy Watch and other wireless devices -- **PIT management**: Get, set, and print PIT (Partition Information Table) files -- **Cross-platform**: Native support for Windows and Linux -- **USB and TCP transport**: Direct USB and TCP connectivity options -- **MD5 verification**: Built-in MD5 verification for data integrity -- **Compressed download support**; Odin3/4 decompresses the lz4 stream before uploading no matter how recent is the device. We just send it compressed (if the device supports), allowing for up to 2x speed (depends on compression ratio). +- **Multi-device support**: Flash multiple devices in parallel +- **Wireless flashing support**: Support for TCP-based flashing for Galaxy Watch +- **Cross-platform**: Native support for Windows, Linux, MacOS and Android +- **Compressed download support**; Samsung's Odin decompresses the lz4 stream before uploading no matter how recent is the device. We just send it compressed (if the device supports), allowing for up to 2x speed (depends on compression ratio). ## Requirements @@ -28,9 +21,10 @@ A modern, cross-platform Samsung device flashing utility written in C++23. Brokk ### Runtime Requirements -- **Windows**: Windows 7 or later +- **Windows**: Windows 10 or later - **Linux**: Any modern Linux distribution with USB support -- **macOS**: macOS 10.15 or later +- **macOS**: macOS 11 or later +- **Android**: Android 7.0 or later with USB host support ## Building @@ -56,57 +50,94 @@ ninja The compiled executable will be located at `build/brokkr` -## Project Structure +## Linux Notes + +### USB device opened read-only + +Brokkr needs **read/write** access to the Samsung USB device node (`/dev/bus/usb/...`). If your user +only has read access, Brokkr will fail with `UsbFsDevice: opened read-only` and the GUI will show a +popup asking you to reopen as root. + +Two ways to fix it: + +1. **Quickest** — run as root: + + ```bash + sudo ./ + ``` + +2. **Recommended** — add a udev rule so your normal user can flash without `sudo`. Create + `/etc/udev/rules.d/51-brokkr-samsung.rules` with: + + ``` + # Samsung VID + SUBSYSTEM=="usb", ATTR{idVendor}=="04e8", MODE="0660", GROUP="plugdev", TAG+="uaccess" + ``` + + Then reload: + + ```bash + sudo udevadm control --reload-rules + sudo udevadm trigger + ``` + + Make sure your user is in the `plugdev` group (`sudo usermod -aG plugdev $USER`, then log out + and back in). Replug the device after the rule is in place. + +### Wireless device not detected (firewall) + +The wireless listener binds TCP port **13579** on `0.0.0.0`. If your firewall blocks incoming +connections, the connection won't happen. + +For `ufw`: +```bash +sudo ufw allow 13579/tcp +sudo ufw reload ``` -brokkr-flash/ -├── src/ -│ ├── app/ # Application layer (CLI, main logic) -│ ├── core/ # Core utilities (bytes, threading) -│ ├── crypto/ # Cryptographic functions (MD5) -│ ├── io/ # I/O operations (TAR, LZ4) -│ ├── platform/ # Platform-specific code (Windows/Linux) -│ │ ├── linux/ # Linux implementations -│ │ |─- windows/ # Windows implementations -| | └── macos/ # macOS implementations -│ └── protocol/ # Device communication protocols -│ └── odin/ # ODIN protocol implementation -├── CMakeLists.txt # Build configuration -└── LICENSE # GNU General Public License v3 + +For `firewalld`: + +```bash +sudo firewall-cmd --add-port=13579/tcp --permanent +sudo firewall-cmd --reload ``` -## Architecture +For `iptables`: -### ODIN Protocol +```bash +sudo iptables -A INPUT -p tcp --dport 13579 -j ACCEPT +``` -Brokkr implements the Samsung ODIN (Open Download Interface for Nodes) protocol for device communication. The protocol layer handles: +Both the host and the client must be on the same network (or at least be able to route TCP to +each other on this port). -- Device initialization and handshaking -- Partition table (PIT) management -- Binary transfer and flashing -- Bootloader communication +## macOS Notes -### Transport Layer +### "Brokkr can't be opened because the developer cannot be verified" -Multiple transport backends support different connection types: +Brokkr is not notarized with Apple, so the system Gatekeeper will refuse to launch it on first +run and may also tag it as "damaged" after download. This is not an actual problem with the +binary — it's just the quarantine attribute Safari/Finder add to anything fetched from the +internet. -- **USB Bulk**: Direct USB connection to Samsung devices -- **TCP**: Wireless connections via network interface +To clear the attribute and allow the app to run: -### Device Detection +```bash +sudo xattr -rd com.apple.quarantine +``` + +(Adjust the path to the `.app`.) After this, Brokkr launches normally +and the warning will not return. -Platform-specific device enumeration: -- **Windows**: Uses Windows USB APIs to detect connected devices -- **Linux**: Uses sysfs interface for device detection -- **macOS** Uses IOKit for device detection +Alternatively, on the first launch attempt you can right-click the app, choose **Open**, and +then click **Open** in the dialog — but the `xattr` command is faster and removes the warning +permanently. -## Partition Types +## Android Notes -- **AP (Application Processor)**: Main system firmware -- **BL (Bootloader)**: Bootloader image -- **CP (Cellular Processor)**: Modem/radio firmware -- **CSC (Consumer Software Customization)**: Region-specific customization -- **USERDATA**: User data partition +Not all devices are supported to be flashed in the Android version. For problematic devices, +please use the Windows version instead. ## License @@ -129,23 +160,18 @@ This tool is provided as-is for firmware flashing operations. Users assume full Flashing custom firmware may void device warranties and violate terms of service. Use at your own risk. -## Technical Details - -### Build Configuration - -- **Optimization**: LTO (Link Time Optimization) enabled when supported -- **Target Architecture**: Native architecture optimization -- **Debug Information**: Full debug symbols with hot reload support on MSVC - ### Dependencies -- **LZ4**: Compression library for firmware data -- **Threads**: Standard C++ threading library -- **Platform Libraries**: Windows API (Windows) or none (Linux) +- **Compiler**: C++23-capable toolchain (recent GCC, Clang, or MSVC). +- **CMake**: 3.21 or newer. +- **Qt 6.7.3 or newer** + +### USB backends -### Performance Features +- **Linux**: `usbfs` +- **macOS**: `IOKit/IOUSBLib` +- **Windows**: VCOM (like original Odin, same driver) +- **Android**: `libusb` (as of 2.4.5) -- Thread pool for parallel operations -- Efficient binary I/O with zero-copy spans -- Streaming tar and LZ4 decompression -- Hardware-accelerated crypto operations when available +## Telegram Community +- https://t.me/BrokkrCommunity diff --git a/src-gui/brokkr_wrapper.cpp b/src-gui/brokkr_wrapper.cpp index cf77dca..ecdb0bc 100644 --- a/src-gui/brokkr_wrapper.cpp +++ b/src-gui/brokkr_wrapper.cpp @@ -1,1999 +1,2637 @@ -#include "brokkr_wrapper.hpp" - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include - -#include "app/md5_verify.hpp" -#include "app/version.hpp" -#include "core/str.hpp" -#include "protocol/odin/flash.hpp" -#include "protocol/odin/group_flasher.hpp" -#include "protocol/odin/odin_cmd.hpp" -#include "protocol/odin/pit.hpp" -#include "protocol/odin/pit_transfer.hpp" - -#if defined(BROKKR_PLATFORM_LINUX) - #include - #include - #include - #include -#endif - -#if defined(Q_OS_WIN) - #include - #include -#endif - -class DeviceSquare final : public QWidget { - public: - enum class Variant { Green, Blue, Red }; - - explicit DeviceSquare(QWidget* parent = nullptr) : QWidget(parent) { - setAttribute(Qt::WA_TranslucentBackground, true); - setAutoFillBackground(false); - - auto sp = sizePolicy(); - sp.setHorizontalPolicy(QSizePolicy::Expanding); - sp.setVerticalPolicy(QSizePolicy::Fixed); - setSizePolicy(sp); - setFixedHeight(58); - - setVariant(Variant::Green); - anim_.setEasingCurve(QEasingCurve::InOutQuad); - anim_.setDuration(120); - QObject::connect(&anim_, &QVariantAnimation::valueChanged, this, [this](const QVariant& v) { - fill_ = std::clamp(v.toDouble(), 0.0, 1.0); - updateGeometry(); - update(); - }); - } - - QSize sizeHint() const override { return QSize(30, 18); } - QSize minimumSizeHint() const override { return QSize(14, 18); } - - void setFill(double v) { - fill_ = std::clamp(v, 0.0, 1.0); - updateGeometry(); - update(); - } - - void setFillAnimated(double v, int ms) { - v = std::clamp(v, 0.0, 1.0); - anim_.stop(); - anim_.setDuration(ms); - anim_.setStartValue(fill_); - anim_.setEndValue(v); - anim_.start(); - } - - void setVariant(Variant v) { - var_ = v; - switch (var_) { - case Variant::Green: - top_ = QColor("#b4e051"); - bot_ = QColor("#5ba30a"); - textCol_ = QColor("#ffffff"); - break; - case Variant::Blue: - top_ = QColor("#68b3e4"); - bot_ = QColor("#186ba6"); - textCol_ = QColor("#ffffff"); - break; - case Variant::Red: - top_ = QColor("#d95757"); - bot_ = QColor("#9a0a0a"); - textCol_ = QColor("#ffffff"); - break; - } - update(); - } - - void setText(const QString& s) { - text_ = s; - update(); - } - - protected: - void paintEvent(QPaintEvent*) override { - QPainter p(this); - p.setRenderHint(QPainter::Antialiasing, false); - - const QRect r = rect().adjusted(1, 1, -1, -1); - p.fillRect(r, palette().color(QPalette::Base)); - - const int fillW = static_cast(std::lround(r.width() * fill_)); - if (fillW > 0) { - QRect fr = r; - fr.setWidth(fillW); - QLinearGradient g(fr.topLeft(), fr.bottomLeft()); - g.setColorAt(0.0, top_); - g.setColorAt(1.0, bot_); - p.fillRect(fr, g); - } - - p.setPen(QPen(palette().color(QPalette::Mid), 1)); - p.drawRect(r); - - if (!text_.isEmpty()) { - QFont f("Arial"); - f.setStyleHint(QFont::SansSerif); - f.setBold(true); - - int pt = 10; - for (; pt >= 6; --pt) { - f.setPointSize(pt); - QFontMetrics fm(f); - if (fm.height() <= (r.height() - 2)) break; - } - if (pt < 6) f.setPointSize(6); - - p.setFont(f); - - QFontMetrics fm(f); - const int maxW = std::max(0, r.width() - 4); - const QString shown = fm.elidedText(text_, Qt::ElideMiddle, maxW); - const QRect textR = r.adjusted(2, 0, -2, 0); - - p.setPen(QColor(0, 0, 0, 160)); - p.drawText(textR.translated(1, 1), Qt::AlignCenter, shown); - - p.setPen(textCol_); - p.drawText(textR, Qt::AlignCenter, shown); - } - } - - private: - double fill_ = 0.0; - Variant var_ = Variant::Green; - QColor top_; - QColor bot_; - QColor textCol_; - QString text_; - QVariantAnimation anim_{}; -}; - -namespace { - -static constexpr std::uint16_t SAMSUNG_VID = 0x04E8; -static constexpr std::uint16_t ODIN_PIDS[] = {0x6601, 0x685D, 0x68C3}; - -static std::vector default_pids() { return {std::begin(ODIN_PIDS), std::end(ODIN_PIDS)}; } - -static bool is_pit_name(std::string_view base) { return brokkr::core::ends_with_ci(base, ".pit"); } - -static QString htmlEsc(const QString& s) { return s.toHtmlEscaped(); } - -static QString human_bytes(std::uint64_t b) { - const char* units[] = {"B", "KB", "MB", "GB", "TB"}; - double v = static_cast(b); - int u = 0; - while (v >= 1024.0 && u < 4) { - v /= 1024.0; - ++u; - } - const int prec = (u == 0) ? 0 : 1; - return QString("%1 %2").arg(v, 0, 'f', prec).arg(units[u]); -} - -template -class QtTextSink final : public spdlog::sinks::base_sink { - public: - explicit QtTextSink(BrokkrWrapper* w) : w_(w) {} - - protected: - void sink_it_(const spdlog::details::log_msg& msg) override { - spdlog::memory_buf_t formatted; - this->formatter_->format(msg, formatted); - - QString line = QString::fromUtf8(formatted.data(), static_cast(formatted.size())); - while (!line.isEmpty() && (line.endsWith('\n') || line.endsWith('\r'))) line.chop(1); - if (line.isEmpty()) return; - - const int z = w_ ? w_->logDeviceCountForLog() : 0; - line = QString("<%1> %2").arg(z).arg(line); - - QMetaObject::invokeMethod( - w_, [w = w_, s = htmlEsc(line)]() { w->appendLogLineFromEngine(s); }, Qt::QueuedConnection); - } - - void flush_() override {} - - private: - BrokkrWrapper* w_ = nullptr; -}; - -static std::shared_ptr make_qt_logger(BrokkrWrapper* w) { - auto sink = std::make_shared>(w); - sink->set_pattern("%v"); - auto log = std::make_shared("qt", spdlog::sinks_init_list{sink}); -#ifndef NDEBUG - log->set_level(spdlog::level::debug); -#else - log->set_level(spdlog::level::info); -#endif - return log; -} - -static std::optional select_target(QString sysname_q) { - const std::string sysname = sysname_q.trimmed().toStdString(); - if (sysname.empty()) return {}; - - auto info = brokkr::platform::find_by_sysname(sysname); - if (!info) return {}; - - if (info->vendor != SAMSUNG_VID) return {}; - const auto pids = default_pids(); - if (std::ranges::find(pids, info->product) == pids.end()) return {}; - return info; -} - -static std::vector enumerate_targets() { - brokkr::platform::EnumerateFilter f{.vendor = SAMSUNG_VID, .products = default_pids()}; - return brokkr::platform::enumerate_usb_devices_sysfs(f); -} - -static brokkr::core::Result> read_all_source(brokkr::io::ByteSource& src) noexcept { - constexpr std::uint64_t kMax = 256ull * 1024ull * 1024ull; - const auto sz64 = src.size(); - if (sz64 > kMax) return brokkr::core::fail("Source too large: " + src.display_name()); - - std::vector out(static_cast(sz64)); - for (std::size_t off = 0; off < out.size();) { - const std::size_t got = src.read({out.data() + off, out.size() - off}); - if (!got) { - auto st = src.status(); - if (!st) return brokkr::core::fail(std::move(st.error())); - return brokkr::core::fail("Short read: " + src.display_name()); - } - off += got; - } - return out; -} - -static std::shared_ptr> pit_from_specs(const std::vector& specs) { - const brokkr::odin::ImageSpec* pit = nullptr; - for (const auto& s : specs) - if (is_pit_name(s.basename)) pit = &s; - if (!pit) return {}; - - auto sr = pit->open(); - if (!sr) { - spdlog::error("PIT open failed: {}", sr.error()); - return {}; - } - - auto rr = read_all_source(**sr); - if (!rr) { - spdlog::error("PIT read failed: {}", rr.error()); - return {}; - } - - return std::make_shared>(std::move(*rr)); -} - -static void print_pit_table_to_log(const brokkr::odin::pit::PitTable& t) { - auto d = [&](const std::string& s) { return s.empty() ? "-" : s; }; - - spdlog::info("PIT TABLE"); - spdlog::info("cpu_bl_id: {}", d(t.cpu_bl_id)); - spdlog::info("com_tar2: {}", d(t.com_tar2)); - spdlog::info("lu_count: {}", t.lu_count); - spdlog::info("entries: {}", t.partitions.size()); - spdlog::info(" "); - - for (std::size_t i = 0; i < t.partitions.size(); ++i) { - const auto& p = t.partitions[i]; - spdlog::info("Partition #{}:", i); - spdlog::info("id: {}", p.id); - spdlog::info("dev_type: {}", p.dev_type); - spdlog::info("block_count: {}", p.block_size); - spdlog::info("block_size: {}", p.block_bytes); - spdlog::info("file_size: {}", p.file_size); - spdlog::info("name: {}", d(p.name)); - spdlog::info("file_name: {}", d(p.file_name)); - spdlog::info(" "); - } -} - -static brokkr::odin::OdinCommands::ShutdownMode shutdown_mode_from_ui(int idx) { - if (idx == 1) return brokkr::odin::OdinCommands::ShutdownMode::ReDownload; - if (idx == 2) return brokkr::odin::OdinCommands::ShutdownMode::NoReboot; - return brokkr::odin::OdinCommands::ShutdownMode::Reboot; -} - -static brokkr::core::Status print_pit_over_link(brokkr::core::IByteTransport& link, const brokkr::odin::Cfg& cfg, - brokkr::odin::OdinCommands::ShutdownMode sm) noexcept { - brokkr::odin::OdinCommands odin(link); - link.set_timeout_ms(cfg.preflash_timeout_ms); - - BRK_TRY(odin.handshake(cfg.preflash_retries)); - BRK_TRYV(_, odin.get_version(cfg.preflash_retries)); - - link.set_timeout_ms(cfg.flash_timeout_ms); - - BRK_TRYV(bytes, brokkr::odin::download_pit_bytes(odin, cfg.preflash_retries)); - BRK_TRYV(tbl, brokkr::odin::pit::parse({bytes.data(), bytes.size()})); - - print_pit_table_to_log(tbl); - - return odin.shutdown(sm); -} - -} // namespace - -#if defined(Q_OS_MACOS) -void BrokkrWrapper::macOsUsbDeviceChanged(void* refCon, io_iterator_t iterator) { - bool changed = false; - spdlog::debug("macOS USB device change detected"); - while (io_service_t object = IOIteratorNext(iterator)) { - changed = true; - spdlog::debug("Device changed: {}", object); - IOObjectRelease(object); - } - spdlog::debug("End of device change events"); - if (changed && refCon) { - auto* wrapper = static_cast(refCon); - spdlog::debug("Requesting USB refresh from macOS device change event"); - static_cast(refCon)->requestUsbRefresh_(); - } - spdlog::debug("Finished processing macOS USB device change events"); -} -#endif - -BrokkrWrapper::BrokkrWrapper(QWidget* parent) : QWidget(parent) { - setWindowTitle("Brokkr Flash"); - setWindowIcon(QIcon(":/brokkr.ico")); - resize(850, 600); - baseWindowHeight_ = height(); - - spdlog::set_default_logger(make_qt_logger(this)); - - auto* mainLayout = new QVBoxLayout(this); - mainLayout->setContentsMargins(10, 10, 10, 10); - - headerWidget_ = new QWidget(this); - headerWidget_->setObjectName("headerBanner"); - headerWidget_->setFixedHeight(56); - applyHeaderStyle_(); - - auto* headerLayout = new QHBoxLayout(headerWidget_); - headerLayout->setContentsMargins(24, 0, 24, 0); - - auto* titleLabel = - new QLabel(QString(R"(BROKKR)" - R"(  FLASH TOOL)" - R"(  v%1)") - .arg(QString::fromStdString(brokkr::app::version_string())), - headerWidget_); - headerLayout->addWidget(titleLabel); - headerLayout->addStretch(); - - ledContainer_ = new QWidget(headerWidget_); - auto* ledInner = new QHBoxLayout(ledContainer_); - ledInner->setContentsMargins(0, 0, 0, 0); - ledInner->setSpacing(8); - headerLayout->addWidget(ledContainer_); - - mainLayout->addWidget(headerWidget_); - - idComGroup_ = new QGroupBox("ID:COM", this); - { - auto pal = idComGroup_->palette(); - QColor bg = pal.color(QPalette::Window); - bg = bg.darker(110); - pal.setColor(QPalette::Window, bg); - idComGroup_->setPalette(pal); - idComGroup_->setAutoFillBackground(true); - } - idComLayout_ = new QGridLayout(idComGroup_); - idComLayout_->setSpacing(4); - idComLayout_->setContentsMargins(5, 5, 5, 5); - - rebuildDeviceBoxes_(kBoxesNormal, true); - mainLayout->addWidget(idComGroup_); - - auto* middleLayout = new QHBoxLayout(); - - tabWidget_ = new QTabWidget(this); - tabWidget_->setFixedWidth(360); - - consoleOutput = new QTextEdit(this); - consoleOutput->setReadOnly(true); - consoleOutput->setVerticalScrollBarPolicy(Qt::ScrollBarAlwaysOn); - QFont logFont = QFontDatabase::systemFont(QFontDatabase::GeneralFont); - logFont.setPointSize(9); - consoleOutput->setFont(logFont); - tabWidget_->addTab(consoleOutput, "Log"); - - auto* optTab = new QWidget(); - auto* optLayout = new QVBoxLayout(optTab); - optLayout->setAlignment(Qt::AlignTop); - - auto* targetLayout = new QHBoxLayout(); - targetLayout->addWidget(new QLabel("Target Sysname:", this)); - editTarget = new QLineEdit(this); - editTarget->setPlaceholderText("e.g. COM12 or 1-1.4"); - targetLayout->addWidget(editTarget); - optLayout->addLayout(targetLayout); - - chkWireless = new QCheckBox("Wireless", this); - optLayout->addWidget(chkWireless); - - optLayout->addSpacing(10); - optLayout->addWidget(new QLabel("Post-Action:", this)); - cmbRebootAction = new QComboBox(this); - cmbRebootAction->addItem("Default (Reboot Normally)"); - cmbRebootAction->addItem("Redownload"); - cmbRebootAction->addItem("No Reboot"); - optLayout->addWidget(cmbRebootAction); - - chkAdvanced_ = new QCheckBox("Advanced options", this); - optLayout->addWidget(chkAdvanced_); - - manyRowWidget_ = new QWidget(this); - auto* manyRow = new QHBoxLayout(manyRowWidget_); - manyRow->setContentsMargins(0, 0, 0, 0); - lblDeviceBoxes = new QLabel("Count: 8", this); - sldDeviceBoxes = new QSlider(Qt::Horizontal, this); - sldDeviceBoxes->setRange(kBoxesNormal, kMassDlMaxBoxes); - sldDeviceBoxes->setValue(kMassDlMaxBoxes); - sldDeviceBoxes->setEnabled(false); - manyRow->addWidget(lblDeviceBoxes); - manyRow->addWidget(sldDeviceBoxes, 1); - manyRowWidget_->setVisible(false); - optLayout->addWidget(manyRowWidget_); - - tabWidget_->addTab(optTab, "Options"); - - auto* pitTab = new QWidget(); - auto* pitLayout = new QGridLayout(pitTab); - pitLayout->setAlignment(Qt::AlignTop); - - chkUsePit = new QCheckBox("Use PIT", this); - pitLayout->addWidget(chkUsePit, 0, 0, 1, 3); - - pitLayout->addWidget(new QLabel("PIT File:"), 1, 0); - - editPit = new QLineEdit(this); - editPit->setEnabled(false); - pitLayout->addWidget(editPit, 1, 1); - - btnPitBrowse = new QPushButton("Browse", this); - btnPitBrowse->setEnabled(false); - pitLayout->addWidget(btnPitBrowse, 1, 2); - - tabWidget_->addTab(pitTab, "Pit"); - - connect(chkAdvanced_, &QCheckBox::toggled, this, [this](bool on) { - if (busy_) { - chkAdvanced_->setChecked(!on); - return; - } - if (btnPrintPit) btnPrintPit->setVisible(on); - if (manyRowWidget_) manyRowWidget_->setVisible(on); - }); - - connect(chkUsePit, &QCheckBox::toggled, this, [this](bool checked) { - if (busy_) { - chkUsePit->setChecked(!checked); - return; - } - editPit->setEnabled(checked); - btnPitBrowse->setEnabled(checked); - if (!checked) editPit->clear(); - updateActionButtons_(); - }); - - connect(btnPitBrowse, &QPushButton::clicked, this, [this]() { - if (busy_) return; - const QString file = QFileDialog::getOpenFileName(this, "Select PIT File", lastDir, - "PIT Files (*.pit);;All Files (*)", nullptr, - QFileDialog::DontUseNativeDialog); - if (!file.isEmpty()) { - lastDir = QFileInfo(file).absolutePath(); - editPit->setText(file); - updateActionButtons_(); - } - }); - - btnManyDevices_ = new QPushButton("Mass D/L", this); - btnManyDevices_->setCheckable(true); - btnManyDevices_->setMinimumWidth(135); - btnManyDevices_->setFixedHeight(24); - - connect(btnManyDevices_, &QPushButton::toggled, this, [this](bool checked) { - if (busy_) { - btnManyDevices_->setChecked(!checked); - return; - } - if (chkWireless && chkWireless->isChecked()) { - btnManyDevices_->setChecked(false); - return; - } - - sldDeviceBoxes->setEnabled(checked); - - if (!checked) { - lblDeviceBoxes->setText(QString("Count: %1").arg(kBoxesNormal)); - rebuildDeviceBoxes_(kBoxesNormal, true); - layout()->invalidate(); - layout()->activate(); - setMinimumHeight(0); - resize(width(), baseWindowHeight_); - } else { - const int v = std::min(sldDeviceBoxes->value(), kMassDlMaxBoxes); - lblDeviceBoxes->setText(QString("Count: %1").arg(v)); - rebuildDeviceBoxes_(v, false); - applyWindowHeightToContents_(); - } - - refreshDeviceBoxes_(); - updateActionButtons_(); - }); - - connect(sldDeviceBoxes, &QSlider::valueChanged, this, [this](int v) { - if (busy_) return; - if (!btnManyDevices_ || !btnManyDevices_->isChecked()) return; - if (chkWireless && chkWireless->isChecked()) return; - - v = std::min(v, kMassDlMaxBoxes); - lblDeviceBoxes->setText(QString("Count: %1").arg(v)); - rebuildDeviceBoxes_(v, false); - applyWindowHeightToContents_(); - refreshDeviceBoxes_(); - updateActionButtons_(); - }); - - connect(chkWireless, &QCheckBox::toggled, this, [this](bool checked) { - if (busy_) { - chkWireless->setChecked(!checked); - return; - } - - if (checked) { - if (btnManyDevices_) { - btnManyDevices_->setChecked(false); - btnManyDevices_->setEnabled(false); - } - if (sldDeviceBoxes) sldDeviceBoxes->setEnabled(false); - if (lblDeviceBoxes) lblDeviceBoxes->setText("Count: 1"); - - rebuildDeviceBoxes_(1, true); - layout()->invalidate(); - layout()->activate(); - setMinimumHeight(0); - resize(width(), baseWindowHeight_); - - startWirelessListener_(); - } else { - stopWirelessListener_(); - - if (btnManyDevices_) btnManyDevices_->setEnabled(true); - if (sldDeviceBoxes) sldDeviceBoxes->setEnabled(false); - if (lblDeviceBoxes) lblDeviceBoxes->setText(QString("Count: %1").arg(kBoxesNormal)); - - rebuildDeviceBoxes_(kBoxesNormal, true); - layout()->invalidate(); - layout()->activate(); - setMinimumHeight(0); - resize(width(), baseWindowHeight_); - } - - refreshDeviceBoxes_(); - updateActionButtons_(); - }); - - connect(editTarget, &QLineEdit::textChanged, this, [this](const QString&) { - if (busy_) return; - updateActionButtons_(); - requestUsbRefresh_(); - }); - - middleLayout->addWidget(tabWidget_); - - auto* rightWidget = new QWidget(this); - auto* rightLayout = new QVBoxLayout(rightWidget); - rightLayout->setAlignment(Qt::AlignTop); - - QLabel* tipsLabel = new QLabel( - "Tips - How to download HOME binary\n" - " OLD model : Download one binary ...\n" - " NEW model : Download BL + AP + CP + CSC", - this); - tipsLabel->setFrameStyle(QFrame::StyledPanel | QFrame::Sunken); - tipsLabel->setMargin(6); - QFont tipsFont = tipsLabel->font(); - tipsFont.setPointSize(9); - tipsLabel->setFont(tipsFont); - QPalette tipsPal = tipsLabel->palette(); - tipsPal.setColor(QPalette::WindowText, Qt::gray); - tipsLabel->setPalette(tipsPal); - rightLayout->addWidget(tipsLabel); - - auto* fileLayout = new QGridLayout(); - fileLayout->setVerticalSpacing(8); - - setupOdinFileInput(fileLayout, 0, "BL", editBL); - setupOdinFileInput(fileLayout, 1, "AP", editAP); - setupOdinFileInput(fileLayout, 2, "CP", editCP); - setupOdinFileInput(fileLayout, 3, "CSC", editCSC); - setupOdinFileInput(fileLayout, 4, "USERDATA", editUserData); - - fileLayout->setColumnStretch(0, 0); - fileLayout->setColumnStretch(1, 0); - fileLayout->setColumnStretch(2, 1); - - rightLayout->addLayout(fileLayout); - rightLayout->addStretch(); - middleLayout->addWidget(rightWidget, 1); - - mainLayout->addLayout(middleLayout); - - auto* bottomLayout = new QHBoxLayout(); - - btnRun = new QPushButton("Start", this); - btnReset_ = new QPushButton("Reset", this); - btnPrintPit = new QPushButton("Print PIT", this); - - const int bottomW = 135; - const int bottomH = 32; - btnRun->setMinimumSize(bottomW, bottomH); - btnReset_->setMinimumSize(bottomW, bottomH); - btnPrintPit->setMinimumSize(bottomW, bottomH); - - btnPrintPit->setVisible(false); - - auto* footerLabel = new QLabel( - R"(GitHub Repo)", this); - footerLabel->setOpenExternalLinks(true); - QFont footerFont = footerLabel->font(); - footerFont.setPointSize(10); - footerLabel->setFont(footerFont); - footerLabel->setAlignment(Qt::AlignLeft | Qt::AlignVCenter); - - bottomLayout->addWidget(footerLabel, 0, Qt::AlignBottom); - bottomLayout->addStretch(); - bottomLayout->addWidget(btnPrintPit, 0, Qt::AlignBottom); - bottomLayout->addSpacing(10); - - auto* resetColWidget = new QWidget(this); - auto* resetColLayout = new QVBoxLayout(resetColWidget); - resetColLayout->setContentsMargins(0, 0, 0, 0); - resetColLayout->setSpacing(6); - resetColLayout->addWidget(btnManyDevices_); - resetColLayout->addWidget(btnReset_); - bottomLayout->addWidget(resetColWidget, 0, Qt::AlignBottom); - - bottomLayout->addSpacing(10); - bottomLayout->addWidget(btnRun, 0, Qt::AlignBottom); - - mainLayout->addLayout(bottomLayout); - - connect(btnPrintPit, &QPushButton::clicked, this, [this]() { - if (busy_) return; - QString why; - if (!canRunPrintPit_(&why)) { - showBlocked_("Cannot print PIT", why); - return; - } - startWorkPrintPit_(); - }); - - connect(btnRun, &QPushButton::clicked, this, &BrokkrWrapper::onRunClicked); - - connect(btnReset_, &QPushButton::clicked, this, [this, fileLayout]() { - if (busy_) return; - - editAP->clear(); - editBL->clear(); - editCP->clear(); - editCSC->clear(); - editUserData->clear(); - consoleOutput->clear(); - editTarget->clear(); - editPit->clear(); - chkUsePit->setChecked(false); - cmbRebootAction->setCurrentIndex(0); - - slotFailed_.assign(static_cast(devSquares_.size()), 0); - lastFlashDevices_.clear(); - - setSquaresText_(""); - setSquaresProgress_(0.0, false); - setSquaresActiveColor_(false); - - for (int i = 0; i < fileLayout->count(); ++i) { - if (auto* chk = qobject_cast(fileLayout->itemAt(i)->widget())) chk->setChecked(false); - } - - updateActionButtons_(); - requestUsbRefresh_(); - }); - -#if defined(BROKKR_PLATFORM_LINUX) - { - uevent_fd_ = ::socket(AF_NETLINK, SOCK_DGRAM, NETLINK_KOBJECT_UEVENT); - if (uevent_fd_ >= 0) { - sockaddr_nl addr{}; - addr.nl_family = AF_NETLINK; - addr.nl_groups = 1; - if (::bind(uevent_fd_, reinterpret_cast(&addr), sizeof(addr)) == 0) { - uevent_notifier_ = new QSocketNotifier(uevent_fd_, QSocketNotifier::Read, this); - connect(uevent_notifier_, &QSocketNotifier::activated, this, [this]() { - char buf[4096]; - const ssize_t n = ::recv(uevent_fd_, buf, sizeof(buf) - 1, 0); - if (n <= 0) return; - buf[n] = '\0'; - const QString s = QString::fromLocal8Bit(buf, n); - if (s.contains("SUBSYSTEM=usb") || s.contains("SUBSYSTEM=tty")) { - // SAFETY: n is always positive here due to the check above, and we ensure null-termination at buf[n] = '\0'. - spdlog::debug("Uevent received: {}", std::string_view{buf, static_cast(n)}); - requestUsbRefresh_(); - } - }); - } else { - ::close(uevent_fd_); - uevent_fd_ = -1; - } - } - } -#endif -#if defined(Q_OS_MACOS) - { - mac_notify_port_ = IONotificationPortCreate(kIOMasterPortDefault); - CFRunLoopAddSource(CFRunLoopGetMain(), IONotificationPortGetRunLoopSource(mac_notify_port_), kCFRunLoopDefaultMode); - - CFMutableDictionaryRef matchDict = IOServiceMatching(kIOUSBDeviceClassName); - matchDict = (CFMutableDictionaryRef)CFRetain(matchDict); - - IOServiceAddMatchingNotification(mac_notify_port_, kIOFirstMatchNotification, matchDict, - &BrokkrWrapper::macOsUsbDeviceChanged, this, &mac_added_iter_); - - macOsUsbDeviceChanged(nullptr, mac_added_iter_); - - IOServiceAddMatchingNotification(mac_notify_port_, kIOTerminatedNotification, matchDict, - &BrokkrWrapper::macOsUsbDeviceChanged, this, &mac_removed_iter_); - - macOsUsbDeviceChanged(nullptr, mac_removed_iter_); - } -#endif - - deviceTimer = new QTimer(this); - connect(deviceTimer, &QTimer::timeout, this, [this]() { - if (!usbDirty_.exchange(false)) return; - refreshConnectedDevices_(); - }); - deviceTimer->start(2000); - - requestUsbRefresh_(); - setControlsEnabled_(true); - updateActionButtons_(); -} - -BrokkrWrapper::~BrokkrWrapper() { - stopWirelessListener_(); -#if defined(BROKKR_PLATFORM_LINUX) - if (uevent_notifier_) uevent_notifier_->setEnabled(false); - if (uevent_fd_ >= 0) { - ::close(uevent_fd_); - uevent_fd_ = -1; - } -#endif -#if defined(Q_OS_MACOS) - if (mac_notify_port_) { - CFRunLoopRemoveSource(CFRunLoopGetMain(), IONotificationPortGetRunLoopSource(mac_notify_port_), - kCFRunLoopDefaultMode); - IONotificationPortDestroy(mac_notify_port_); - mac_notify_port_ = nullptr; - } - if (mac_added_iter_) { - IOObjectRelease(mac_added_iter_); - mac_added_iter_ = 0; - } - if (mac_removed_iter_) { - IOObjectRelease(mac_removed_iter_); - mac_removed_iter_ = 0; - } -#endif -} - -void BrokkrWrapper::appendLogLineFromEngine(const QString& html) { appendLogLine_(html); } - -void BrokkrWrapper::closeEvent(QCloseEvent* e) { - if (busy_) { - QMessageBox::warning(this, "Brokkr Flasher", "An operation is in progress. Please wait for it to complete."); - e->ignore(); - return; - } - e->accept(); -} - -void BrokkrWrapper::changeEvent(QEvent* e) { - QWidget::changeEvent(e); - if (e->type() == QEvent::PaletteChange || e->type() == QEvent::ApplicationPaletteChange) { - applyHeaderStyle_(); - - { - auto pal = palette(); - QColor bg = pal.color(QPalette::Window).darker(110); - pal.setColor(QPalette::Window, bg); - idComGroup_->setPalette(pal); - } - - consoleOutput->setPalette(palette()); - refreshDeviceBoxes_(); - } -} - -#if defined(Q_OS_WIN) -bool BrokkrWrapper::nativeEvent(const QByteArray& eventType, void* message, qintptr* result) { - (void)eventType; - (void)result; - MSG* msg = reinterpret_cast(message); - if (!msg) return false; - - if (msg->message == WM_DEVICECHANGE) { - if (msg->wParam == DBT_DEVICEARRIVAL || msg->wParam == DBT_DEVICEREMOVECOMPLETE) requestUsbRefresh_(); - } - return false; -} -#endif - -void BrokkrWrapper::requestUsbRefresh_() noexcept { - spdlog::debug("USB refresh requested"); - usbDirty_.store(true, std::memory_order_relaxed); -} - -void BrokkrWrapper::refreshConnectedDevices_() { - QStringList shown; - QStringList physicalUsb; - - for (const auto& d : enumerate_targets()) physicalUsb << QString::fromStdString(d.sysname); - - bool physicalWireless = false; - QString physicalWirelessId; - { - std::lock_guard lk(wireless_mtx_); - const bool wantWireless = (chkWireless && chkWireless->isChecked()); - if (wantWireless) { - if (wireless_conn_ && wireless_conn_->connected() && !wireless_sysname_.isEmpty()) { - physicalWireless = true; - physicalWirelessId = wireless_sysname_; - } else { - wireless_conn_.reset(); - wireless_sysname_.clear(); - } - } - } - - const int physicalCount = physicalUsb.size() + (physicalWireless ? 1 : 0); - logDevCount_.store(physicalCount, std::memory_order_relaxed); - - const QString tgt = editTarget ? editTarget->text().trimmed() : QString{}; - if (!tgt.isEmpty()) { - auto info = brokkr::platform::find_by_sysname(tgt.toStdString()); - if (info) shown << QString::fromStdString(info->sysname); - } else { - shown = physicalUsb; - } - - if (physicalWireless) shown.prepend(physicalWirelessId); - - if (busy_) return; - - { - const QSet prev = QSet(physicalUsbPrev_.begin(), physicalUsbPrev_.end()); - const QSet now = QSet(physicalUsb.begin(), physicalUsb.end()); - - bool anyAttached = false; - - for (const auto& s : now) { - if (!prev.contains(s)) { - spdlog::info("Connected: {}", s.toStdString()); - anyAttached = true; - } - } - for (const auto& s : prev) - if (!now.contains(s)) spdlog::info("Disconnected: {}", s.toStdString()); - - if (physicalWirelessPrev_ && (!physicalWireless || physicalWirelessIdPrev_ != physicalWirelessId)) { - if (!physicalWirelessIdPrev_.isEmpty()) spdlog::info("Disconnected: {}", physicalWirelessIdPrev_.toStdString()); - } - if (physicalWireless && (!physicalWirelessPrev_ || physicalWirelessIdPrev_ != physicalWirelessId)) { - if (!physicalWirelessId.isEmpty()) { - spdlog::info("Connected: {}", physicalWirelessId.toStdString()); - anyAttached = true; - } - } - - if (anyAttached) lastFlashDevices_.clear(); - - physicalUsbPrev_ = physicalUsb; - physicalWirelessPrev_ = physicalWireless; - physicalWirelessIdPrev_ = physicalWirelessId; - } - - connectedDevices_ = shown; - refreshDeviceBoxes_(); - updateHeaderLeds_(); - updateActionButtons_(); -} - -void BrokkrWrapper::startWirelessListener_() { - if (busy_) return; - - stopWirelessListener_(); - - wireless_sysname_.clear(); - { - std::lock_guard lk(wireless_mtx_); - wireless_conn_.reset(); - wireless_listener_.emplace(); - auto st = wireless_listener_->bind_and_listen("0.0.0.0", 13579); - if (!st) { - spdlog::error("Wireless listen failed: {}", st.error()); - wireless_listener_.reset(); - requestUsbRefresh_(); - return; - } - } - - wireless_thread_ = std::jthread([this](std::stop_token st) { - for (;;) { - if (st.stop_requested()) return; - - { - bool connected = false; - { - std::lock_guard lk(wireless_mtx_); - if (wireless_conn_) { - if (wireless_conn_->connected()) - connected = true; - else { - wireless_conn_.reset(); - wireless_sysname_.clear(); - QMetaObject::invokeMethod(this, [this]() { requestUsbRefresh_(); }, Qt::QueuedConnection); - } - } - } - if (connected) { - std::this_thread::sleep_for(std::chrono::milliseconds(200)); - continue; - } - } - - brokkr::platform::TcpListener* lst = nullptr; - { - std::lock_guard lk(wireless_mtx_); - if (!wireless_listener_) return; - lst = &*wireless_listener_; - } - - auto ar = lst->accept_one(); - if (!ar) { - if (st.stop_requested()) return; - std::this_thread::sleep_for(std::chrono::milliseconds(50)); - continue; - } - - const QString peer = QString::fromStdString(ar->peer_label()); - const int sep = peer.lastIndexOf(':'); - const QString sys = (sep > 0) ? peer.left(sep) : peer; - - { - std::lock_guard lk(wireless_mtx_); - wireless_conn_.emplace(std::move(*ar)); - wireless_sysname_ = sys; - } - - QMetaObject::invokeMethod(this, [this]() { refreshConnectedDevices_(); }, Qt::QueuedConnection); - } - }); - - refreshConnectedDevices_(); -} - -void BrokkrWrapper::stopWirelessListener_() { - brokkr::platform::TcpListener* lst = nullptr; - { - std::lock_guard lk(wireless_mtx_); - if (wireless_listener_) lst = &*wireless_listener_; - } - if (lst) lst->close(); - - if (wireless_thread_.joinable()) { - wireless_thread_.request_stop(); - wireless_thread_.join(); - } - - { - std::lock_guard lk(wireless_mtx_); - wireless_conn_.reset(); - wireless_listener_.reset(); - wireless_sysname_.clear(); - } - requestUsbRefresh_(); -} - -void BrokkrWrapper::appendLogLine_(const QString& html) { - consoleOutput->append(html); - QTextCursor cursor = consoleOutput->textCursor(); - cursor.movePosition(QTextCursor::End); - consoleOutput->setTextCursor(cursor); -} - -void BrokkrWrapper::applyHeaderStyle_() { - if (!headerWidget_) return; - const int luma = palette().color(QPalette::Window).lightness(); - const bool dark = (luma < 128); - - headerWidget_->setStyleSheet(""); // Clear stylesheet to fallback natively - - QPalette pal = headerWidget_->palette(); - QColor bgColor = dark ? QColor("#2d2e30") : QColor("#dce0e4"); - pal.setColor(QPalette::Window, bgColor); - - headerWidget_->setAutoFillBackground(true); - headerWidget_->setPalette(pal); -} - -void BrokkrWrapper::updateHeaderLeds_() { - if (!ledContainer_) return; - - auto* lo = ledContainer_->layout(); - if (!lo) return; - - while (lo->count() > 0) { - QLayoutItem* it = lo->takeAt(0); - if (!it) break; - if (auto* w = it->widget()) delete w; - delete it; - } - - const int count = std::min(connectedDevices_.size(), 12); - for (int i = 0; i < count; ++i) { - auto* led = new QWidget(ledContainer_); - led->setFixedSize(14, 14); - - // NOTE: This single style sheet is maintained because there is no native Mac - // widget for a "glowing orb". Styling a pure QWidget does not break Mac theming. - led->setStyleSheet( - "background: qradialgradient(cx:0.3, cy:0.3, radius:0.7, fx:0.3, fy:0.3," - " stop:0 #aaddff, stop:1 #0066cc);" - "border: 1px solid #000;" - "border-radius: 7px;"); - auto* glow = new QGraphicsDropShadowEffect(led); - glow->setColor(QColor(0, 153, 255, 200)); - glow->setBlurRadius(10); - glow->setOffset(0, 0); - led->setGraphicsEffect(glow); - lo->addWidget(led); - } -} - -void BrokkrWrapper::setControlsEnabled_(bool enabled) { - if (tabWidget_) { - tabWidget_->setEnabled(true); - if (tabWidget_->count() > 0) tabWidget_->setTabEnabled(0, true); - for (int i = 1; i < tabWidget_->count(); ++i) tabWidget_->setTabEnabled(i, enabled); - if (!enabled) tabWidget_->setCurrentIndex(0); - } - if (consoleOutput) consoleOutput->setEnabled(true); - - if (editTarget) editTarget->setEnabled(enabled); - if (cmbRebootAction) cmbRebootAction->setEnabled(enabled); - - if (chkUsePit) chkUsePit->setEnabled(enabled); - if (editPit) editPit->setEnabled(enabled && chkUsePit && chkUsePit->isChecked()); - if (btnPitBrowse) btnPitBrowse->setEnabled(enabled && chkUsePit && chkUsePit->isChecked()); - - if (chkWireless) chkWireless->setEnabled(enabled); - - const bool wireless = (chkWireless && chkWireless->isChecked()); - - if (btnManyDevices_) btnManyDevices_->setEnabled(enabled && !wireless); - if (sldDeviceBoxes) - sldDeviceBoxes->setEnabled(enabled && !wireless && btnManyDevices_ && btnManyDevices_->isChecked()); - - if (chkAdvanced_) chkAdvanced_->setEnabled(enabled); - - for (auto* chk : fileChecks_) - if (chk) chk->setEnabled(enabled); - for (auto* btn : fileButtons_) - if (btn) btn->setEnabled(enabled); - - if (btnReset_) btnReset_->setEnabled(enabled); -} - -void BrokkrWrapper::setBusy_(bool busy) { - busy_ = busy; - if (busy_ && tabWidget_) tabWidget_->setCurrentIndex(0); - setControlsEnabled_(!busy_); - updateActionButtons_(); -} - -void BrokkrWrapper::setSquaresProgress_(double frac, bool animate) { - const int n = devSquares_.size(); - for (int i = 0; i < n; ++i) { - auto* sq = devSquares_[i]; - if (!sq) continue; - if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; - if (static_cast(i) < slotFailed_.size() && slotFailed_[static_cast(i)]) continue; - if (animate) - sq->setFillAnimated(frac, 120); - else - sq->setFill(frac); - } -} - -void BrokkrWrapper::setSquaresText_(const QString& s) { - const int n = devSquares_.size(); - for (int i = 0; i < n; ++i) { - auto* sq = devSquares_[i]; - if (!sq) continue; - if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; - if (static_cast(i) < slotFailed_.size() && slotFailed_[static_cast(i)]) continue; - sq->setText(s); - } -} - -void BrokkrWrapper::setSquaresActiveColor_(bool enhanced) { - enhanced_speed_ = enhanced; - const int n = devSquares_.size(); - for (int i = 0; i < n; ++i) { - auto* sq = devSquares_[i]; - if (!sq) continue; - if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; - if (static_cast(i) < slotFailed_.size() && slotFailed_[static_cast(i)]) continue; - sq->setVariant(enhanced ? DeviceSquare::Variant::Blue : DeviceSquare::Variant::Green); - } -} - -void BrokkrWrapper::setSquaresFinal_(bool ok) { - if (!ok) { - const int n = devSquares_.size(); - for (int i = 0; i < n; ++i) { - auto* sq = devSquares_[i]; - if (!sq) continue; - if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; - sq->setVariant(DeviceSquare::Variant::Red); - sq->setText("FAIL!"); - sq->setFillAnimated(1.0, 250); - } - return; - } - - const auto passV = (enhanced_speed_ ? DeviceSquare::Variant::Green : DeviceSquare::Variant::Blue); - const int n = devSquares_.size(); - for (int i = 0; i < n; ++i) { - auto* sq = devSquares_[i]; - if (!sq) continue; - if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; - const bool failed = (static_cast(i) < slotFailed_.size() && slotFailed_[static_cast(i)]); - if (failed) { - sq->setVariant(DeviceSquare::Variant::Red); - sq->setText("FAIL!"); - sq->setFillAnimated(1.0, 200); - } else { - sq->setVariant(passV); - sq->setText("PASS"); - sq->setFillAnimated(1.0, 350); - } - } -} - -void BrokkrWrapper::applyWindowHeightToContents_() { - if (!layout()) return; - layout()->invalidate(); - layout()->activate(); - const int want = layout()->sizeHint().height() + 30; - if (height() < want) resize(width(), want); -} - -void BrokkrWrapper::rebuildDeviceBoxes_(int boxCount, bool singleRow) { - while (idComLayout_->count() > 0) { - QLayoutItem* it = idComLayout_->takeAt(0); - if (!it) break; - if (auto* w = it->widget()) delete w; - delete it; - } - - for (int c = 0; c < kMassDlMaxBoxes + 2; ++c) idComLayout_->setColumnStretch(c, 0); - for (int r = 0; r < (kMassDlMaxBoxes / std::max(1, kBoxesColsMany)) + 2; ++r) idComLayout_->setRowStretch(r, 0); - - comBoxes.clear(); - devSquares_.clear(); - - const bool singleCellNoStretch = (singleRow && boxCount == 1); - - int fixedCellW = -1; - if (singleCellNoStretch) { - const int cols = kBoxesNormal; - int sp = idComLayout_->horizontalSpacing(); - if (sp < 0) sp = idComLayout_->spacing(); - const auto m = idComLayout_->contentsMargins(); - const int avail = std::max(0, idComGroup_->contentsRect().width() - m.left() - m.right()); - const int totalSp = std::max(0, (cols - 1) * std::max(0, sp)); - fixedCellW = (cols > 0) ? ((avail - totalSp) / cols) : -1; - if (fixedCellW < 30) fixedCellW = 30; - } - - auto make_cell = [&](int row, int col) { - auto* cell = new QWidget(idComGroup_); - auto* v = new QVBoxLayout(cell); - v->setContentsMargins(0, 0, 0, 0); - v->setSpacing(2); - - auto* sq = new DeviceSquare(cell); - - auto* box = new QLineEdit(cell); - box->setReadOnly(true); - box->setAlignment(Qt::AlignCenter); - box->setMinimumHeight(22); - - // Use QFont instead of CSS for text size - QFont boxFont = box->font(); - boxFont.setPointSize(9); - box->setFont(boxFont); - - if (singleCellNoStretch && fixedCellW > 0) { - cell->setFixedWidth(fixedCellW); - box->setFixedWidth(fixedCellW); - sq->setMinimumWidth(fixedCellW); - sq->setMaximumWidth(fixedCellW); - } - - v->addWidget(sq); - v->addWidget(box); - - idComLayout_->addWidget(cell, row, col); - - devSquares_.append(sq); - comBoxes.append(box); - }; - - if (singleRow) { - if (boxCount > 0) make_cell(0, 0); - - if (singleCellNoStretch) { - idComLayout_->setColumnStretch(0, 0); - idComLayout_->addItem(new QSpacerItem(0, 0, QSizePolicy::Expanding, QSizePolicy::Minimum), 0, 1); - idComLayout_->setColumnStretch(1, 1); - } else { - for (int col = 1; col < boxCount; ++col) make_cell(0, col); - for (int col = 0; col < boxCount; ++col) idComLayout_->setColumnStretch(col, 1); - } - - slotFailed_.assign(static_cast(devSquares_.size()), 0); - slotActive_.assign(static_cast(devSquares_.size()), 0); - return; - } - - for (int c = 0; c < kBoxesColsMany; ++c) idComLayout_->setColumnStretch(c, 1); - - for (int i = 0; i < boxCount; ++i) { - const int row = i / kBoxesColsMany; - const int col = i % kBoxesColsMany; - make_cell(row, col); - } - - slotFailed_.assign(static_cast(devSquares_.size()), 0); - slotActive_.assign(static_cast(devSquares_.size()), 0); -} - -void BrokkrWrapper::refreshDeviceBoxes_() { - overflowDevices_ = false; - - for (auto* box : comBoxes) { - box->clear(); - box->setToolTip(QString()); - box->setStyleSheet(""); // Clear any leftover inline styling - - QPalette pal = box->palette(); - pal.setColor(QPalette::Base, palette().color(QPalette::Window)); - pal.setColor(QPalette::Text, palette().color(QPalette::Text)); - box->setPalette(pal); - - QFont f = box->font(); - f.setBold(false); - box->setFont(f); - } - - auto elideFor = [&](QLineEdit* box, const QString& s) { - const int w = std::max(0, box->width() - 10); - return box->fontMetrics().elidedText(s, Qt::ElideMiddle, w); - }; - - const int shown = std::min(connectedDevices_.size(), comBoxes.size()); - for (int i = 0; i < shown; ++i) { - const QString sysname = connectedDevices_[i].trimmed(); - const QString raw = QString("%1:[%2]").arg(i).arg(sysname); - comBoxes[i]->setText(elideFor(comBoxes[i], raw)); - comBoxes[i]->setToolTip(sysname); - - QPalette pal = comBoxes[i]->palette(); - pal.setColor(QPalette::Base, QColor("#4080c0")); // Active blue tint - pal.setColor(QPalette::Text, Qt::white); - comBoxes[i]->setPalette(pal); - - QFont f = comBoxes[i]->font(); - f.setBold(true); - comBoxes[i]->setFont(f); - } - - for (int i = shown; i < comBoxes.size(); ++i) { - if (i < lastFlashDevices_.size()) { - const QString prev = lastFlashDevices_[i].trimmed(); - if (!prev.isEmpty()) { - const QString raw = QString("%1:[%2]").arg(i).arg(prev); - comBoxes[i]->setText(elideFor(comBoxes[i], raw)); - comBoxes[i]->setToolTip(prev); - - QPalette pal = comBoxes[i]->palette(); - pal.setColor(QPalette::Text, palette().color(QPalette::Mid)); - comBoxes[i]->setPalette(pal); - } - } - } - - if (connectedDevices_.size() > comBoxes.size() && !comBoxes.isEmpty()) { - overflowDevices_ = true; - const int extra = connectedDevices_.size() - comBoxes.size(); - auto* last = comBoxes.back(); - const QString raw = QString("... +%1 more").arg(extra); - last->setText(elideFor(last, raw)); - - QPalette pal = last->palette(); - QColor base = palette().color(QPalette::Base); - QColor warn(255, 183, 77); - warn.setAlpha(50); - const int r2 = (base.red() * (255 - warn.alpha()) + warn.red() * warn.alpha()) / 255; - const int g2 = (base.green() * (255 - warn.alpha()) + warn.green() * warn.alpha()) / 255; - const int b2 = (base.blue() * (255 - warn.alpha()) + warn.blue() * warn.alpha()) / 255; - - pal.setColor(QPalette::Base, QColor(r2, g2, b2)); - pal.setColor(QPalette::Text, palette().color(QPalette::Text)); - last->setPalette(pal); - - QFont f = last->font(); - f.setBold(true); - last->setFont(f); - } -} - -bool BrokkrWrapper::canRunStart_(QString* whyNot) const { - const bool wireless = chkWireless->isChecked(); - const bool hasTarget = !editTarget->text().trimmed().isEmpty(); - - const bool hasAnyFile = (!editBL->text().isEmpty() && editBL->isEnabled()) || - (!editAP->text().isEmpty() && editAP->isEnabled()) || - (!editCP->text().isEmpty() && editCP->isEnabled()) || - (!editCSC->text().isEmpty() && editCSC->isEnabled()) || - (!editUserData->text().isEmpty() && editUserData->isEnabled()) || - (chkUsePit->isChecked() && !editPit->text().isEmpty()); - - if (!hasAnyFile) { - if (whyNot) *whyNot = "No files selected."; - return false; - } - - if (wireless && hasTarget) { - if (whyNot) *whyNot = "Wireless cannot be used together with Target Sysname."; - return false; - } - - if (wireless) { - std::lock_guard lk(wireless_mtx_); - if (!wireless_conn_ || !wireless_conn_->connected()) { - if (whyNot) *whyNot = "Wireless is enabled but no device is connected yet."; - return false; - } - return true; - } - - if (hasTarget) return true; - - if (connectedDevices_.isEmpty()) { - if (whyNot) *whyNot = "No connected devices detected."; - return false; - } - if (overflowDevices_) { - if (whyNot) *whyNot = "Too many devices are connected for the current GUI box limit."; - return false; - } - return true; -} - -bool BrokkrWrapper::canRunPrintPit_(QString* whyNot) const { - const bool wireless = chkWireless->isChecked(); - const bool hasTarget = !editTarget->text().trimmed().isEmpty(); - - if (wireless && hasTarget) { - if (whyNot) *whyNot = "Wireless cannot be used together with Target Sysname."; - return false; - } - - if (wireless) { - std::lock_guard lk(wireless_mtx_); - if (!wireless_conn_ || !wireless_conn_->connected()) { - if (whyNot) *whyNot = "Wireless is enabled but no device is connected yet."; - return false; - } - return true; - } - - if (hasTarget) return true; - - if (connectedDevices_.isEmpty()) { - if (whyNot) *whyNot = "No connected devices detected."; - return false; - } - - if (connectedDevices_.size() != 1) { - if (whyNot) *whyNot = "Printing PIT from device requires exactly one connected device."; - return false; - } - - if (overflowDevices_) { - if (whyNot) *whyNot = "Too many devices for PIT print."; - return false; - } - - return true; -} - -void BrokkrWrapper::updateActionButtons_() { - if (busy_) { - btnRun->setEnabled(false); - if (btnPrintPit) btnPrintPit->setEnabled(false); - return; - } - - QString why; - btnRun->setEnabled(canRunStart_(&why)); - if (btnPrintPit) btnPrintPit->setEnabled(canRunPrintPit_(&why)); -} - -void BrokkrWrapper::showBlocked_(const QString& title, const QString& msg) const { - QMessageBox::warning(const_cast(this), title, msg); -} - -void BrokkrWrapper::setupOdinFileInput(QGridLayout* layout, int row, const QString& label, QLineEdit*& lineEdit) { - auto* chk = new QCheckBox(this); - layout->addWidget(chk, row, 0); - - auto* btn = new QPushButton(label, this); - btn->setMinimumWidth(95); - btn->setFixedHeight(28); - layout->addWidget(btn, row, 1); - - lineEdit = new QLineEdit(this); - lineEdit->setReadOnly(true); - layout->addWidget(lineEdit, row, 2); - - fileChecks_.append(chk); - fileButtons_.append(btn); - - connect(btn, &QPushButton::clicked, this, [this, lineEdit, chk]() { - if (busy_) return; - QString file = QFileDialog::getOpenFileName(this, "Select Firmware File", lastDir, - "Firmware Archives (*.tar *.tar.md5);;All Files (*)", nullptr, - QFileDialog::DontUseNativeDialog); - if (!file.isEmpty()) { - lastDir = QFileInfo(file).absolutePath(); - lineEdit->setText(file); - chk->setChecked(true); - } - updateActionButtons_(); - }); - - connect(chk, &QCheckBox::toggled, this, [this, lineEdit](bool checked) { - if (busy_) return; - lineEdit->setEnabled(checked); - updateActionButtons_(); - }); -} - -void BrokkrWrapper::onRunClicked() { - if (busy_) return; - QString why; - if (!canRunStart_(&why)) { - showBlocked_("Cannot start", why); - return; - } - startWorkStart_(); -} - -void BrokkrWrapper::startWorkStart_() { - if (busy_) return; - - lastFlashDevices_ = connectedDevices_; - const QStringList uiDevicesSnapshot = connectedDevices_; - setBusy_(true); - - plan_names_.clear(); - plan_from_names_.clear(); - enhanced_speed_ = false; - - slotFailed_.assign(static_cast(devSquares_.size()), 0); - slotActive_.assign(static_cast(devSquares_.size()), 0); - { - const int activeCount = std::min(uiDevicesSnapshot.size(), devSquares_.size()); - for (int i = 0; i < activeCount; ++i) slotActive_[static_cast(i)] = 1; - } - - setSquaresProgress_(0.0, false); - setSquaresText_(""); - setSquaresActiveColor_(false); - - worker_ = std::jthread([this, uiDevicesSnapshot](std::stop_token) { - auto done_ui = [&] { - QMetaObject::invokeMethod( - this, - [this]() { - setSquaresText_("PASS"); - setSquaresFinal_(true); - setBusy_(false); - }, - Qt::QueuedConnection); - }; - - auto fail_ui = [&](const QString& msg) { - QMetaObject::invokeMethod( - this, - [this, msg]() { - const int z = logDevCount_.load(std::memory_order_relaxed); - appendLogLine_(QString("<%1> FAIL! %2").arg(z).arg(htmlEsc(msg))); - setSquaresFinal_(false); - setBusy_(false); - }, - Qt::QueuedConnection); - }; - - const int actionIndex = cmbRebootAction->currentIndex(); - - const QString tgt = editTarget->text().trimmed(); - const bool wireless = chkWireless->isChecked(); - if (wireless && !tgt.isEmpty()) { - fail_ui("Wireless cannot be used together with Target Sysname."); - return; - } - - QMetaObject::invokeMethod(this, [this]() { setSquaresText_("HANDSHAKE"); }, Qt::QueuedConnection); - - brokkr::odin::Cfg cfg; - cfg.redownload_after = (actionIndex == 1); - cfg.reboot_after = (actionIndex != 2); - - brokkr::odin::Ui ui; - - ui.on_plan = [&](const std::vector& p, std::uint64_t) { - std::vector parts, froms; - parts.reserve(p.size()); - froms.reserve(p.size()); - for (const auto& it : p) { - std::string part = it.part_name; - if (part.empty() && it.kind == brokkr::odin::PlanItem::Kind::Pit) part = "PIT"; - std::string from = it.pit_file_name; - if (from.empty()) from = part; - parts.push_back(QString::fromStdString(part)); - froms.push_back(QString::fromStdString(from)); - } - QMetaObject::invokeMethod( - this, - [this, parts = std::move(parts), froms = std::move(froms)]() mutable { - plan_names_ = std::move(parts); - plan_from_names_ = std::move(froms); - }, - Qt::QueuedConnection); - }; - - ui.on_item_active = [&](std::size_t i) { - QMetaObject::invokeMethod( - this, - [this, i]() { - if (i >= plan_names_.size()) return; - setSquaresText_(plan_names_[i]); - }, - Qt::QueuedConnection); - }; - - ui.on_stage = [&](const std::string& s) { - const QString qs = QString::fromStdString(s); - QMetaObject::invokeMethod( - this, - [this, qs]() { - if (qs.contains("Enhanced", Qt::CaseInsensitive)) - setSquaresActiveColor_(true); - else if (qs.contains("Normal", Qt::CaseInsensitive)) - setSquaresActiveColor_(false); - if (qs.contains("handshake", Qt::CaseInsensitive)) setSquaresText_("HANDSHAKE"); - if (qs.contains("shutdown", Qt::CaseInsensitive) || qs.contains("reboot", Qt::CaseInsensitive) || - qs.contains("reset", Qt::CaseInsensitive) || qs.contains("finalizing", Qt::CaseInsensitive)) - setSquaresText_("RESET"); - }, - Qt::QueuedConnection); - }; - - ui.on_progress = [&](std::uint64_t d, std::uint64_t t, std::uint64_t, std::uint64_t) { - const double frac = (t > 0) ? (static_cast(d) / static_cast(t)) : 0.0; - QMetaObject::invokeMethod(this, [this, frac]() { setSquaresProgress_(frac, true); }, Qt::QueuedConnection); - }; - - ui.on_error = [&](const std::string& s) { - QMetaObject::invokeMethod( - this, - [this, qs = QString::fromStdString(s)]() { - const int z = logDevCount_.load(std::memory_order_relaxed); - - QString shown = qs; - int idx = -1; - - const QString pref = "DEVFAIL idx="; - if (qs.startsWith(pref)) { - int p = pref.size(); - int sp = qs.indexOf(' ', p); - if (sp < 0) sp = qs.size(); - bool okNum = false; - idx = qs.mid(p, sp - p).toInt(&okNum); - QString reason; - if (sp < qs.size()) reason = qs.mid(sp + 1).trimmed(); - - if (okNum && idx >= 0) { - if (static_cast(idx) >= slotFailed_.size()) - slotFailed_.resize(static_cast(idx) + 1, 0); - slotFailed_[static_cast(idx)] = 1; - - if (idx < devSquares_.size() && devSquares_[idx]) { - devSquares_[idx]->setVariant(DeviceSquare::Variant::Red); - devSquares_[idx]->setText("FAIL!"); - devSquares_[idx]->setFill(1.0); - } - - // Keep native but set red palette for failure - if (idx < comBoxes.size() && comBoxes[idx]) { - QPalette pal = comBoxes[idx]->palette(); - pal.setColor(QPalette::Base, QColor("#a02020")); - pal.setColor(QPalette::Text, Qt::white); - comBoxes[idx]->setPalette(pal); - } - - shown = reason.isEmpty() ? "FAIL!" : QString("FAIL! (Device %1) %2").arg(idx).arg(reason); - } - } - - appendLogLine_(QString("<%1> %2").arg(z).arg(htmlEsc(shown))); - }, - Qt::QueuedConnection); - }; - - std::shared_ptr> pit_to_upload; - if (chkUsePit->isChecked() && !editPit->text().isEmpty()) { - const std::filesystem::path p = editPit->text().toStdString(); - std::error_code ec; - const auto sz = std::filesystem::file_size(p, ec); - if (ec) { - fail_ui("Cannot stat PIT file."); - return; - } - - std::vector buf(static_cast(sz)); - std::ifstream in(p, std::ios::binary); - if (!in.is_open()) { - fail_ui("Cannot open PIT file."); - return; - } - if (!buf.empty()) { - in.read(reinterpret_cast(buf.data()), static_cast(buf.size())); - if (!in.good()) { - fail_ui("Failed to read PIT file."); - return; - } - } - pit_to_upload = std::make_shared>(std::move(buf)); - } - - std::vector inputs; - if (editBL->isEnabled() && !editBL->text().isEmpty()) inputs.emplace_back(editBL->text().toStdString()); - if (editAP->isEnabled() && !editAP->text().isEmpty()) inputs.emplace_back(editAP->text().toStdString()); - if (editCP->isEnabled() && !editCP->text().isEmpty()) inputs.emplace_back(editCP->text().toStdString()); - if (editCSC->isEnabled() && !editCSC->text().isEmpty()) inputs.emplace_back(editCSC->text().toStdString()); - if (editUserData->isEnabled() && !editUserData->text().isEmpty()) - inputs.emplace_back(editUserData->text().toStdString()); - - struct Provider { - std::vector> usb; // owns USB transports - std::vector owned; // owns Targets - std::vector ptrs; // passed to engine - }; - - auto make_provider = [&]() -> std::optional { - Provider p; - - if (wireless) { - brokkr::platform::TcpConnection* connp = nullptr; - { - std::lock_guard lk(wireless_mtx_); - if (wireless_conn_) connp = &*wireless_conn_; - } - if (!connp || !connp->connected()) { - fail_ui("No wireless device connected."); - return std::nullopt; - } - - p.owned.push_back(brokkr::odin::Target{.id = wireless_sysname_.toStdString(), .link = connp}); - p.ptrs.push_back(&p.owned.back()); - return p; - } - - std::vector targets; - if (!tgt.isEmpty()) { - auto one = select_target(tgt); - if (!one) { - fail_ui("Target sysname not found or not supported."); - return std::nullopt; - } - targets.push_back(*one); - } else { - for (const auto& sys : uiDevicesSnapshot) - if (auto one = select_target(sys)) targets.push_back(*one); - if (targets.empty()) targets = enumerate_targets(); - } - - if (targets.empty()) { - fail_ui("No supported devices found."); - return std::nullopt; - } - - p.usb.reserve(targets.size()); - p.owned.reserve(targets.size()); - p.ptrs.reserve(targets.size()); - - for (const auto& td : targets) { - auto ut = std::make_unique(td.devnode()); - - auto st = ut->dev.open_and_init(); - if (!st) { - fail_ui(QString::fromStdString(st.error())); - return std::nullopt; - } - - auto cst = ut->conn.open(); - if (!cst) { - fail_ui(QString::fromStdString(cst.error())); - return std::nullopt; - } - - ut->conn.set_timeout_ms(cfg.preflash_timeout_ms); - - p.owned.push_back(brokkr::odin::Target{.id = ut->devnode, .link = &ut->conn}); - p.ptrs.push_back(&p.owned.back()); - p.usb.push_back(std::move(ut)); - } - - return p; - }; - - auto run_engine = [&](std::vector srcs) { - auto prov = make_provider(); - if (!prov) return; - - auto st = brokkr::odin::flash(prov->ptrs, srcs, pit_to_upload, cfg, ui); - (void)st; // UI handles per-device fails via ui.on_error; overall uses setSquaresFinal_(true) - done_ui(); - requestUsbRefresh_(); - }; - - if (inputs.empty()) { - run_engine({}); - return; - } - - auto jobsr = brokkr::app::md5_jobs(inputs); - if (!jobsr) { - fail_ui(QString::fromStdString(jobsr.error())); - return; - } - - std::uint64_t totalBytes = 0; - for (const auto& p : inputs) { - std::error_code ec; - const auto sz = std::filesystem::file_size(p, ec); - if (!ec) totalBytes += static_cast(sz); - } - - spdlog::info("MD5 check ({}), Please wait.", human_bytes(totalBytes).toStdString()); - QMetaObject::invokeMethod( - this, - [this]() { - setSquaresActiveColor_(false); - setSquaresText_("MD5"); - setSquaresProgress_(0.0, false); - }, - Qt::QueuedConnection); - - auto vst = brokkr::app::md5_verify(*jobsr, ui); - if (!vst) { - fail_ui(QString::fromStdString(vst.error())); - return; - } - - auto specs = brokkr::odin::expand_inputs_tar_or_raw(inputs); - if (!specs) { - fail_ui(QString::fromStdString(specs.error())); - return; - } - - const bool dl_mode = std::ranges::any_of(*specs, - [](const brokkr::odin::ImageSpec& s) { return s.download_list_mode; }); - - if (!pit_to_upload && !dl_mode) { - auto pit = pit_from_specs(*specs); - if (pit) pit_to_upload = pit; - } - - std::vector srcs; - for (auto& s : *specs) - if (!is_pit_name(s.basename)) srcs.push_back(std::move(s)); - if (srcs.empty() && !pit_to_upload) { - fail_ui("No valid flashable files."); - return; - } - - QMetaObject::invokeMethod(this, [this]() { setSquaresProgress_(0.0, false); }, Qt::QueuedConnection); - - run_engine(std::move(srcs)); - }); -} - -void BrokkrWrapper::startWorkPrintPit_() { - if (busy_) return; - setBusy_(true); - - slotFailed_.assign(static_cast(devSquares_.size()), 0); - slotActive_.assign(static_cast(devSquares_.size()), 0); - { - const int activeCount = std::min(connectedDevices_.size(), devSquares_.size()); - for (int i = 0; i < activeCount; ++i) slotActive_[static_cast(i)] = 1; - } - - setSquaresProgress_(0.0, false); - setSquaresText_("PIT"); - setSquaresActiveColor_(false); - - const int actionIndex = cmbRebootAction->currentIndex(); - - worker_ = std::jthread([this, actionIndex](std::stop_token) { - auto done_ui = [&] { - QMetaObject::invokeMethod( - this, - [this]() { - setSquaresText_("PASS"); - setSquaresFinal_(true); - setBusy_(false); - }, - Qt::QueuedConnection); - }; - - auto fail_ui = [&](const QString& msg) { - QMetaObject::invokeMethod( - this, - [this, msg]() { - appendLogLine_(QString("%1").arg(htmlEsc(msg))); - setSquaresFinal_(false); - setBusy_(false); - }, - Qt::QueuedConnection); - }; - - const auto sm = shutdown_mode_from_ui(actionIndex); - - const QString tgt = editTarget->text().trimmed(); - const bool wireless = chkWireless->isChecked(); - if (wireless && !tgt.isEmpty()) { - fail_ui("Wireless cannot be used together with Target Sysname."); - return; - } - - brokkr::odin::Cfg cfg; - - auto run_link = [&](brokkr::core::IByteTransport& link) { - QMetaObject::invokeMethod(this, [this]() { setSquaresText_("HANDSHAKE"); }, Qt::QueuedConnection); - - auto st = print_pit_over_link(link, cfg, sm); - if (!st) { - fail_ui(QString::fromStdString(st.error())); - return; - } - - QMetaObject::invokeMethod(this, [this]() { setSquaresText_("RESET"); }, Qt::QueuedConnection); - done_ui(); - }; - - if (wireless) { - brokkr::platform::TcpConnection* connp = nullptr; - { - std::lock_guard lk(wireless_mtx_); - if (wireless_conn_) connp = &*wireless_conn_; - } - if (!connp || !connp->connected()) { - fail_ui("No wireless device connected."); - return; - } - run_link(*connp); - return; - } - - std::vector targets; - if (!tgt.isEmpty()) { - auto one = select_target(tgt); - if (!one) { - fail_ui("Target sysname not found or not supported."); - return; - } - targets.push_back(*one); - } else { - targets = enumerate_targets(); - } - - if (targets.size() != 1) { - fail_ui("Printing PIT requires exactly one device."); - return; - } - - brokkr::platform::UsbFsDevice dev(targets.front().devnode()); - auto dst = dev.open_and_init(); - if (!dst) { - fail_ui(QString::fromStdString(dst.error())); - return; - } - - brokkr::platform::UsbFsConnection conn(dev); - auto cst = conn.open(); - if (!cst) { - fail_ui(QString::fromStdString(cst.error())); - return; - } - - run_link(conn); - }); -} +#include "brokkr_wrapper.hpp" + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#include "app/md5_verify.hpp" +#include "app/version.hpp" +#include "core/str.hpp" +#include "protocol/odin/flash.hpp" +#include "protocol/odin/group_flasher.hpp" + +#if defined(BROKKR_PLATFORM_LINUX) + #include + #include + #include + #include +#endif + +#if defined(Q_OS_WIN) + #include + #include + #include +#endif + +class DeviceSquare final : public QWidget { + public: + enum class Variant { Green, Blue, Red }; + + explicit DeviceSquare(QWidget* parent = nullptr) : QWidget(parent) { + setAttribute(Qt::WA_TranslucentBackground, true); + setAutoFillBackground(false); + + auto sp = sizePolicy(); + sp.setHorizontalPolicy(QSizePolicy::Expanding); + sp.setVerticalPolicy(QSizePolicy::Fixed); + setSizePolicy(sp); + setFixedHeight(58); + + setVariant(Variant::Green); + anim_.setEasingCurve(QEasingCurve::InOutQuad); + anim_.setDuration(120); + QObject::connect(&anim_, &QVariantAnimation::valueChanged, this, [this](const QVariant& v) { + fill_ = std::clamp(v.toDouble(), 0.0, 1.0); + updateGeometry(); + update(); + }); + } + + QSize sizeHint() const override { return QSize(30, 18); } + QSize minimumSizeHint() const override { return QSize(14, 18); } + + void setFill(double v) { + fill_ = std::clamp(v, 0.0, 1.0); + updateGeometry(); + update(); + } + + void setFillAnimated(double v, int ms) { + v = std::clamp(v, 0.0, 1.0); + anim_.stop(); + anim_.setDuration(ms); + anim_.setStartValue(fill_); + anim_.setEndValue(v); + anim_.start(); + } + + void setVariant(Variant v) { + var_ = v; + switch (var_) { + case Variant::Green: + top_ = QColor("#b4e051"); + bot_ = QColor("#5ba30a"); + break; + case Variant::Blue: + top_ = QColor("#68b3e4"); + bot_ = QColor("#186ba6"); + break; + case Variant::Red: + top_ = QColor("#d95757"); + bot_ = QColor("#9a0a0a"); + break; + } + update(); + } + + void setText(const QString& s) { + text_ = s; + update(); + } + + protected: + void paintEvent(QPaintEvent*) override { + QPainter p(this); + p.setRenderHint(QPainter::Antialiasing, false); + + const QRect r = rect().adjusted(1, 1, -1, -1); + p.fillRect(r, palette().color(QPalette::Base)); + + const int fillW = static_cast(std::lround(r.width() * fill_)); + if (fillW > 0) { + QRect fr = r; + fr.setWidth(fillW); + QLinearGradient g(fr.topLeft(), fr.bottomLeft()); + g.setColorAt(0.0, top_); + g.setColorAt(1.0, bot_); + p.fillRect(fr, g); + } + + p.setPen(QPen(palette().color(QPalette::Mid), 1)); + p.drawRect(r); + + if (!text_.isEmpty()) { + QFont f("Arial"); + f.setStyleHint(QFont::SansSerif); + f.setBold(true); + + int pt = 10; + for (; pt >= 6; --pt) { + f.setPointSize(pt); + QFontMetrics fm(f); + if (fm.height() <= (r.height() - 2)) break; + } + if (pt < 6) f.setPointSize(6); + + p.setFont(f); + + QFontMetrics fm(f); + const int maxW = std::max(0, r.width() - 4); + const QString shown = fm.elidedText(text_, Qt::ElideMiddle, maxW); + const QRect textR = r.adjusted(2, 0, -2, 0); + + const bool dark_theme = palette().color(QPalette::Window).lightness() < 128; + const QColor shadow_col = dark_theme ? QColor(0, 0, 0, 160) : QColor(255, 255, 255, 180); + const QColor text_col = dark_theme ? QColor("#ffffff") : QColor("#111111"); + + p.setPen(shadow_col); + p.drawText(textR.translated(1, 1), Qt::AlignCenter, shown); + + p.setPen(text_col); + p.drawText(textR, Qt::AlignCenter, shown); + } + } + + private: + double fill_ = 0.0; + Variant var_ = Variant::Green; + QColor top_; + QColor bot_; + QString text_; + QVariantAnimation anim_{}; +}; + +namespace { + +static constexpr std::uint16_t SAMSUNG_VID = 0x04E8; +static constexpr std::uint16_t ODIN_PIDS[] = {0x6601, 0x685D, 0x68C3}; + +static std::vector default_pids() { return {std::begin(ODIN_PIDS), std::end(ODIN_PIDS)}; } + +static bool is_odin_product(std::uint16_t pid) { + const auto pids = default_pids(); + return std::ranges::find(pids, pid) != pids.end(); +} + +static bool is_pit_name(std::string_view base) { return brokkr::core::ends_with_ci(base, ".pit"); } + +static bool is_pit_drop_name(const QString& file_name) { + return file_name.trimmed().toLower().endsWith(".pit"); +} + +static bool is_firmware_drop_name(const QString& file_name) { + const QString name = file_name.trimmed().toLower(); + return name.endsWith(".tar") || name.endsWith(".md5"); +} + +static bool is_token_char(QChar c) { return c.isLetterOrNumber(); } + +static bool contains_token(const QString& upper_name, const QString& token) { + int from = 0; + for (;;) { + const int pos = upper_name.indexOf(token, from, Qt::CaseSensitive); + if (pos < 0) return false; + + const int left = pos - 1; + const int right = pos + token.size(); + + const bool left_ok = (left < 0) || !is_token_char(upper_name[left]); + const bool right_ok = (right >= upper_name.size()) || !is_token_char(upper_name[right]); + + if (left_ok && right_ok) return true; + from = pos + 1; + } +} + +static constexpr const char* kDropSlotProperty = "brokkr.dropSlot"; + +static QString drop_file_identity_key(const QString& path_like) { + QFileInfo fi(QDir::cleanPath(path_like.trimmed())); + QString key; + + if (fi.exists()) { + key = fi.canonicalFilePath(); + if (key.isEmpty()) key = fi.absoluteFilePath(); + } else { + key = fi.absoluteFilePath(); + } + + key = QDir::cleanPath(key); +#if defined(Q_OS_WIN) + key = key.toLower(); +#endif + return key; +} + +#if defined(Q_OS_WIN) +static void allow_windows_drop_messages(HWND hwnd) { + if (!hwnd) return; + (void)ChangeWindowMessageFilterEx(hwnd, WM_DROPFILES, MSGFLT_ALLOW, nullptr); + (void)ChangeWindowMessageFilterEx(hwnd, WM_COPYDATA, MSGFLT_ALLOW, nullptr); + (void)ChangeWindowMessageFilterEx(hwnd, 0x0049, MSGFLT_ALLOW, nullptr); // WM_COPYGLOBALDATA + DragAcceptFiles(hwnd, TRUE); +} + +static bool is_windows_process_elevated() { + HANDLE tok = nullptr; + if (!OpenProcessToken(GetCurrentProcess(), TOKEN_QUERY, &tok)) return false; + + TOKEN_ELEVATION elev{}; + DWORD out = 0; + const BOOL ok = GetTokenInformation(tok, TokenElevation, &elev, sizeof(elev), &out); + CloseHandle(tok); + return ok && elev.TokenIsElevated; +} + +static QStringList parse_hdrop_files(HDROP hdrop) { + QStringList out; + if (!hdrop) return out; + + const UINT count = DragQueryFileW(hdrop, 0xFFFFFFFF, nullptr, 0); + for (UINT i = 0; i < count; ++i) { + const UINT need = DragQueryFileW(hdrop, i, nullptr, 0); + if (need == 0) continue; + + std::wstring tmp; + tmp.resize(static_cast(need)); + const UINT got = DragQueryFileW(hdrop, i, tmp.data(), need + 1); + if (got == 0) continue; + + out.push_back(QDir::cleanPath(QString::fromWCharArray(tmp.c_str()))); + } + return out; +} + +static QStringList parse_windows_filenamew_payload(const QByteArray& raw) { + QStringList out; + if (raw.isEmpty()) return out; + + const auto* w = reinterpret_cast(raw.constData()); + const int wlen = static_cast(raw.size() / static_cast(sizeof(wchar_t))); + if (wlen <= 0) return out; + + QString packed = QString::fromWCharArray(w, wlen); + packed.replace('\r', '\0'); + packed.replace('\n', '\0'); + + const auto parts = packed.split('\0', Qt::SkipEmptyParts); + for (const auto& p : parts) { + const QString t = QDir::cleanPath(p.trimmed()); + if (!t.isEmpty()) out.push_back(t); + } + return out; +} +#endif + +static QString htmlEsc(const QString& s) { return s.toHtmlEscaped(); } + +static QString human_bytes(std::uint64_t b) { + const char* units[] = {"B", "KB", "MB", "GB", "TB"}; + double v = static_cast(b); + int u = 0; + while (v >= 1024.0 && u < 4) { + v /= 1024.0; + ++u; + } + const int prec = (u == 0) ? 0 : 1; + return QString("%1 %2").arg(v, 0, 'f', prec).arg(units[u]); +} + +template +class QtTextSink final : public spdlog::sinks::base_sink { + public: + explicit QtTextSink(BrokkrWrapper* w) : w_(w) {} + + protected: + void sink_it_(const spdlog::details::log_msg& msg) override { + spdlog::memory_buf_t formatted; + this->formatter_->format(msg, formatted); + + QString line = QString::fromUtf8(formatted.data(), static_cast(formatted.size())); + while (!line.isEmpty() && (line.endsWith('\n') || line.endsWith('\r'))) line.chop(1); + if (line.isEmpty()) return; + + const int z = w_ ? w_->logDeviceCountForLog() : 0; + line = QString("<%1> %2").arg(z).arg(line); + + QMetaObject::invokeMethod( + w_, [w = w_, s = htmlEsc(line)]() { w->appendLogLineFromEngine(s); }, Qt::QueuedConnection); + } + + void flush_() override {} + + private: + BrokkrWrapper* w_ = nullptr; +}; + +static std::shared_ptr make_qt_logger(BrokkrWrapper* w) { + auto sink = std::make_shared>(w); + sink->set_pattern("%v"); + auto log = std::make_shared("qt", spdlog::sinks_init_list{sink}); +#ifndef NDEBUG + log->set_level(spdlog::level::debug); +#else + log->set_level(spdlog::level::info); +#endif + return log; +} + +static std::optional select_samsung_target(QString sysname_q) { + const std::string sysname = sysname_q.trimmed().toStdString(); + if (sysname.empty()) return {}; + + auto info = brokkr::platform::find_by_sysname(sysname); + if (!info) return {}; + + if (info->vendor != SAMSUNG_VID) return {}; + return info; +} + +static std::optional select_odin_target(QString sysname_q) { + auto info = select_samsung_target(std::move(sysname_q)); + if (!info) return {}; + if (!is_odin_product(info->product)) return {}; + return info; +} + +static std::vector enumerate_samsung_targets() { + brokkr::platform::EnumerateFilter f{.vendor = SAMSUNG_VID}; + return brokkr::platform::enumerate_usb_devices_sysfs(f); +} + +static std::vector enumerate_odin_targets() { + brokkr::platform::EnumerateFilter f{.vendor = SAMSUNG_VID, .products = default_pids()}; + return brokkr::platform::enumerate_usb_devices_sysfs(f); +} + +static brokkr::core::Result> read_all_source(brokkr::io::ByteSource& src) noexcept { + constexpr std::uint64_t kMax = 256ull * 1024ull * 1024ull; + const auto sz64 = src.size(); + if (sz64 > kMax) return brokkr::core::fail("Source too large: " + src.display_name()); + + std::vector out(static_cast(sz64)); + for (std::size_t off = 0; off < out.size();) { + const std::size_t got = src.read({out.data() + off, out.size() - off}); + if (!got) { + auto st = src.status(); + if (!st) return brokkr::core::fail(std::move(st.error())); + return brokkr::core::fail("Short read: " + src.display_name()); + } + off += got; + } + return out; +} + +static std::shared_ptr> pit_from_specs(const std::vector& specs) { + const brokkr::odin::ImageSpec* pit = nullptr; + for (const auto& s : specs) + if (is_pit_name(s.basename)) pit = &s; + if (!pit) return {}; + + auto sr = pit->open(); + if (!sr) { + spdlog::error("PIT open failed: {}", sr.error()); + return {}; + } + + auto rr = read_all_source(**sr); + if (!rr) { + spdlog::error("PIT read failed: {}", rr.error()); + return {}; + } + + return std::make_shared>(std::move(*rr)); +} + +} // namespace + +#if defined(Q_OS_MACOS) +void BrokkrWrapper::macOsUsbDeviceChanged(void* refCon, io_iterator_t iterator) { + bool changed = false; + spdlog::debug("macOS USB device change detected"); + while (io_service_t object = IOIteratorNext(iterator)) { + changed = true; + spdlog::debug("Device changed: {}", object); + IOObjectRelease(object); + } + spdlog::debug("End of device change events"); + if (changed && refCon) { + auto* wrapper = static_cast(refCon); + spdlog::debug("Requesting USB refresh from macOS device change event"); + static_cast(refCon)->requestUsbRefresh_(); + } + spdlog::debug("Finished processing macOS USB device change events"); +} +#endif + +BrokkrWrapper::BrokkrWrapper(QWidget* parent) : QWidget(parent) { + setWindowTitle("Brokkr Flash"); + setWindowIcon(QIcon(":/brokkr.ico")); + resize(850, 600); + setAcceptDrops(true); +#if defined(Q_OS_WIN) + allow_windows_drop_messages(reinterpret_cast(winId())); +#endif + baseWindowHeight_ = height(); + + spdlog::set_default_logger(make_qt_logger(this)); + + auto* mainLayout = new QVBoxLayout(this); + mainLayout->setContentsMargins(10, 10, 10, 10); + + headerWidget_ = new QWidget(this); + headerWidget_->setObjectName("headerBanner"); + headerWidget_->setFixedHeight(56); + applyHeaderStyle_(); + + auto* headerLayout = new QHBoxLayout(headerWidget_); + headerLayout->setContentsMargins(24, 0, 24, 0); + + auto* titleLabel = + new QLabel(QString(R"(BROKKR)" + R"(  FLASH TOOL)" + R"(  v%1)") + .arg(QString::fromStdString(brokkr::app::version_string())), + headerWidget_); + headerLayout->addWidget(titleLabel); + headerLayout->addStretch(); + + ledContainer_ = new QWidget(headerWidget_); + auto* ledInner = new QHBoxLayout(ledContainer_); + ledInner->setContentsMargins(0, 0, 0, 0); + ledInner->setSpacing(8); + headerLayout->addWidget(ledContainer_); + + mainLayout->addWidget(headerWidget_); + + idComGroup_ = new QGroupBox("ID:COM", this); + { + auto pal = idComGroup_->palette(); + QColor bg = pal.color(QPalette::Window); + bg = bg.darker(110); + pal.setColor(QPalette::Window, bg); + idComGroup_->setPalette(pal); + idComGroup_->setAutoFillBackground(true); + } + idComLayout_ = new QGridLayout(idComGroup_); + idComLayout_->setSpacing(4); + idComLayout_->setContentsMargins(5, 5, 5, 5); + + rebuildDeviceBoxes_(kBoxesNormal, true); + mainLayout->addWidget(idComGroup_); + + auto* middleLayout = new QHBoxLayout(); + + tabWidget_ = new QTabWidget(this); + tabWidget_->setFixedWidth(360); + + consoleOutput = new QTextEdit(this); + consoleOutput->setReadOnly(true); + consoleOutput->setVerticalScrollBarPolicy(Qt::ScrollBarAlwaysOn); + QFont logFont = QFontDatabase::systemFont(QFontDatabase::GeneralFont); + logFont.setPointSize(9); + consoleOutput->setFont(logFont); + tabWidget_->addTab(consoleOutput, "Log"); + + auto* optTab = new QWidget(); + auto* optLayout = new QVBoxLayout(optTab); + optLayout->setAlignment(Qt::AlignTop); + + auto* targetLayout = new QHBoxLayout(); + targetLayout->addWidget(new QLabel("Target Sysname:", this)); + editTarget = new QLineEdit(this); + editTarget->setPlaceholderText("e.g. COM12 or 1-1.4"); + targetLayout->addWidget(editTarget); + optLayout->addLayout(targetLayout); + + chkWireless = new QCheckBox("Wireless", this); + optLayout->addWidget(chkWireless); + +#if defined(BROKKR_PLATFORM_WINDOWS) + btnRebootDownloadMode_ = new QPushButton("Try to Reboot them into Download Mode", this); + btnRebootDownloadMode_->setEnabled(false); + optLayout->addWidget(btnRebootDownloadMode_); + + connect(btnRebootDownloadMode_, &QPushButton::clicked, this, [this]() { + if (busy_) return; + tryRebootIntoDownloadMode_(); + }); +#endif + + optLayout->addSpacing(10); + optLayout->addWidget(new QLabel("Post-Action:", this)); + cmbRebootAction = new QComboBox(this); + cmbRebootAction->addItem("Default (Reboot Normally)"); + cmbRebootAction->addItem("No Reboot"); + optLayout->addWidget(cmbRebootAction); + + tabWidget_->addTab(optTab, "Options"); + + auto* pitTab = new QWidget(); + auto* pitLayout = new QGridLayout(pitTab); + pitLayout->setAlignment(Qt::AlignTop); + + chkUsePit = new QCheckBox("Use PIT", this); + pitLayout->addWidget(chkUsePit, 0, 0, 1, 3); + + pitLayout->addWidget(new QLabel("PIT File:"), 1, 0); + + editPit = new QLineEdit(this); + editPit->setEnabled(false); + pitLayout->addWidget(editPit, 1, 1); + + btnPitBrowse = new QPushButton("Browse", this); + btnPitBrowse->setEnabled(false); + pitLayout->addWidget(btnPitBrowse, 1, 2); + + bindDropTarget_(pitTab, kDropSlotPIT); + bindDropTarget_(chkUsePit, kDropSlotPIT); + bindDropTarget_(editPit, kDropSlotPIT); + bindDropTarget_(btnPitBrowse, kDropSlotPIT); + + tabWidget_->addTab(pitTab, "Pit"); + +#if defined(Q_OS_MACOS) + connect(tabWidget_, &QTabWidget::currentChanged, this, [this](int) { + if (auto* fw = QApplication::focusWidget()) fw->clearFocus(); + if (tabWidget_) tabWidget_->setFocus(Qt::OtherFocusReason); + }); +#endif + + connect(chkUsePit, &QCheckBox::toggled, this, [this](bool checked) { + if (busy_) { + chkUsePit->setChecked(!checked); + return; + } + editPit->setEnabled(checked); + btnPitBrowse->setEnabled(checked); + if (!checked) editPit->clear(); + updateActionButtons_(); + }); + + connect(btnPitBrowse, &QPushButton::clicked, this, [this]() { + if (busy_) return; + QFileDialog::Options opts; +#if defined(Q_OS_MACOS) || defined(Q_OS_LINUX) + opts |= QFileDialog::DontUseNativeDialog; +#endif + const QString file = QFileDialog::getOpenFileName(this, "Select PIT File", lastDir, + "PIT Files (*.pit);;All Files (*)", + nullptr, opts); + if (!file.isEmpty()) { + lastDir = QFileInfo(file).absolutePath(); + editPit->setText(file); + updateActionButtons_(); + } + }); + + btnManyDevices_ = new QPushButton("Mass D/L", this); + btnManyDevices_->setCheckable(true); + btnManyDevices_->setMinimumWidth(135); +#if defined(Q_OS_MACOS) + btnManyDevices_->setFixedHeight(32); +#else + btnManyDevices_->setFixedHeight(24); +#endif + + connect(btnManyDevices_, &QPushButton::toggled, this, [this](bool checked) { + if (busy_) { + btnManyDevices_->setChecked(!checked); + return; + } + if (chkWireless && chkWireless->isChecked()) { + btnManyDevices_->setChecked(false); + return; + } + + if (!checked) { + rebuildDeviceBoxes_(kBoxesNormal, true); + layout()->invalidate(); + layout()->activate(); + setMinimumHeight(0); + resize(width(), baseWindowHeight_); + } else { + rebuildDeviceBoxes_(kMassDlMaxBoxes, false); + applyWindowHeightToContents_(); + } + + refreshDeviceBoxes_(); + updateActionButtons_(); + }); + + connect(chkWireless, &QCheckBox::toggled, this, [this](bool checked) { + if (busy_) { + chkWireless->setChecked(!checked); + return; + } + + if (checked) { + if (btnManyDevices_) { + btnManyDevices_->setChecked(false); + btnManyDevices_->setEnabled(false); + } + + rebuildDeviceBoxes_(1, true); + layout()->invalidate(); + layout()->activate(); + setMinimumHeight(0); + resize(width(), baseWindowHeight_); + + startWirelessListener_(); + } else { + stopWirelessListener_(); + + if (btnManyDevices_) btnManyDevices_->setEnabled(true); + + rebuildDeviceBoxes_(kBoxesNormal, true); + layout()->invalidate(); + layout()->activate(); + setMinimumHeight(0); + resize(width(), baseWindowHeight_); + } + + refreshDeviceBoxes_(); + updateActionButtons_(); + }); + + connect(editTarget, &QLineEdit::textChanged, this, [this](const QString&) { + if (busy_) return; + updateActionButtons_(); + requestUsbRefresh_(); + }); + + middleLayout->addWidget(tabWidget_); + + auto* rightWidget = new QWidget(this); + auto* rightLayout = new QVBoxLayout(rightWidget); + rightLayout->setAlignment(Qt::AlignTop); + + QLabel* tipsLabel = new QLabel( + "Tips - How to download HOME binary\n" + " OLD model : Download one binary ...\n" + " NEW model : Download BL + AP + CP + CSC", + this); + tipsLabel->setFrameStyle(QFrame::StyledPanel | QFrame::Sunken); + tipsLabel->setMargin(6); + QFont tipsFont = tipsLabel->font(); + tipsFont.setPointSize(9); + tipsLabel->setFont(tipsFont); + QPalette tipsPal = tipsLabel->palette(); + tipsPal.setColor(QPalette::WindowText, Qt::gray); + tipsLabel->setPalette(tipsPal); + rightLayout->addWidget(tipsLabel); + + auto* fileLayout = new QGridLayout(); + fileLayout->setVerticalSpacing(8); + + setupOdinFileInput(fileLayout, 0, "BL", editBL); + setupOdinFileInput(fileLayout, 1, "AP", editAP); + setupOdinFileInput(fileLayout, 2, "CP", editCP); + setupOdinFileInput(fileLayout, 3, "CSC", editCSC); + setupOdinFileInput(fileLayout, 4, "USERDATA", editUserData); + + fileLayout->setColumnStretch(0, 0); + fileLayout->setColumnStretch(1, 0); + fileLayout->setColumnStretch(2, 1); + + rightLayout->addLayout(fileLayout); + rightLayout->addStretch(); + middleLayout->addWidget(rightWidget, 1); + + mainLayout->addLayout(middleLayout); + + auto* bottomLayout = new QHBoxLayout(); + + btnRun = new QPushButton("Start", this); + btnReset_ = new QPushButton("Reset", this); + + const int bottomW = 135; + const int bottomH = 32; + btnRun->setMinimumSize(bottomW, bottomH); + btnReset_->setMinimumSize(bottomW, bottomH); + + auto* footerLabel = new QLabel( + R"(GitHub Repo)", this); + footerLabel->setOpenExternalLinks(true); + QFont footerFont = footerLabel->font(); + footerFont.setPointSize(10); + footerLabel->setFont(footerFont); + footerLabel->setAlignment(Qt::AlignLeft | Qt::AlignVCenter); + + auto* telegramLabel = new QLabel( + R"(Telegram Community)", this); + telegramLabel->setOpenExternalLinks(true); + telegramLabel->setFont(footerFont); + telegramLabel->setAlignment(Qt::AlignLeft | Qt::AlignVCenter); + + bottomLayout->addWidget(footerLabel, 0, Qt::AlignBottom); + bottomLayout->addSpacing(14); + bottomLayout->addWidget(telegramLabel, 0, Qt::AlignBottom); + bottomLayout->addStretch(); + + auto* resetColWidget = new QWidget(this); + auto* resetColLayout = new QVBoxLayout(resetColWidget); + resetColLayout->setContentsMargins(0, 0, 0, 0); + resetColLayout->setSpacing(6); + resetColLayout->addWidget(btnManyDevices_); + resetColLayout->addWidget(btnReset_); + bottomLayout->addWidget(resetColWidget, 0, Qt::AlignBottom); + + bottomLayout->addSpacing(10); + +#if defined(Q_OS_MACOS) + auto* startColWidget = new QWidget(this); + auto* startColLayout = new QVBoxLayout(startColWidget); + startColLayout->setContentsMargins(0, 0, 0, 0); + startColLayout->setSpacing(6); + startColLayout->addSpacing(btnManyDevices_->sizeHint().height()); + startColLayout->addWidget(btnRun); + bottomLayout->addWidget(startColWidget, 0, Qt::AlignBottom); +#else + bottomLayout->addWidget(btnRun, 0, Qt::AlignBottom); +#endif + + mainLayout->addLayout(bottomLayout); + + connect(btnRun, &QPushButton::clicked, this, &BrokkrWrapper::onRunClicked); + + connect(btnReset_, &QPushButton::clicked, this, [this, fileLayout]() { + if (busy_) return; + + editAP->clear(); + editBL->clear(); + editCP->clear(); + editCSC->clear(); + editUserData->clear(); + consoleOutput->clear(); + editTarget->clear(); + editPit->clear(); + chkUsePit->setChecked(false); + cmbRebootAction->setCurrentIndex(0); + + slotFailed_.assign(static_cast(devSquares_.size()), 0); + slotActive_.assign(static_cast(devSquares_.size()), 0); + + for (auto* sq : devSquares_) { + if (!sq) continue; + sq->setVariant(DeviceSquare::Variant::Green); + sq->setText(""); + sq->setFill(0.0); + } + + refreshDeviceBoxes_(); + updateHeaderLeds_(); + + for (int i = 0; i < fileLayout->count(); ++i) { + if (auto* chk = qobject_cast(fileLayout->itemAt(i)->widget())) chk->setChecked(false); + } + + updateActionButtons_(); + requestUsbRefresh_(); + }); + +#if defined(BROKKR_PLATFORM_LINUX) + { + uevent_fd_ = ::socket(AF_NETLINK, SOCK_DGRAM, NETLINK_KOBJECT_UEVENT); + if (uevent_fd_ >= 0) { + sockaddr_nl addr{}; + addr.nl_family = AF_NETLINK; + addr.nl_groups = 1; + if (::bind(uevent_fd_, reinterpret_cast(&addr), sizeof(addr)) == 0) { + uevent_notifier_ = new QSocketNotifier(uevent_fd_, QSocketNotifier::Read, this); + connect(uevent_notifier_, &QSocketNotifier::activated, this, [this]() { + char buf[4096]; + const ssize_t n = ::recv(uevent_fd_, buf, sizeof(buf) - 1, 0); + if (n <= 0) return; + buf[n] = '\0'; + const QString s = QString::fromLocal8Bit(buf, n); + if (s.contains("SUBSYSTEM=usb") || s.contains("SUBSYSTEM=tty")) { + // SAFETY: n is always positive here due to the check above, and we ensure null-termination at buf[n] = '\0'. + spdlog::debug("Uevent received: {}", std::string_view{buf, static_cast(n)}); + requestUsbRefresh_(); + } + }); + } else { + ::close(uevent_fd_); + uevent_fd_ = -1; + } + } + } +#endif +#if defined(Q_OS_MACOS) + { + mac_notify_port_ = IONotificationPortCreate(kIOMasterPortDefault); + CFRunLoopAddSource(CFRunLoopGetMain(), IONotificationPortGetRunLoopSource(mac_notify_port_), kCFRunLoopDefaultMode); + + CFMutableDictionaryRef matchDict = IOServiceMatching(kIOUSBDeviceClassName); + matchDict = (CFMutableDictionaryRef)CFRetain(matchDict); + + IOServiceAddMatchingNotification(mac_notify_port_, kIOFirstMatchNotification, matchDict, + &BrokkrWrapper::macOsUsbDeviceChanged, this, &mac_added_iter_); + + macOsUsbDeviceChanged(nullptr, mac_added_iter_); + + IOServiceAddMatchingNotification(mac_notify_port_, kIOTerminatedNotification, matchDict, + &BrokkrWrapper::macOsUsbDeviceChanged, this, &mac_removed_iter_); + + macOsUsbDeviceChanged(nullptr, mac_removed_iter_); + } +#endif + + deviceTimer = new QTimer(this); + connect(deviceTimer, &QTimer::timeout, this, [this]() { + if (!usbDirty_.exchange(false)) return; + refreshConnectedDevices_(); + }); + deviceTimer->start(2000); + + requestUsbRefresh_(); + setControlsEnabled_(true); + updateActionButtons_(); + + // Ensure drops are handled regardless of the exact child widget under cursor. + bindDropTarget_(this, kDropSlotAny); + for (auto* w : findChildren()) { + if (!w) continue; + if (w->property(kDropSlotProperty).isValid()) continue; + bindDropTarget_(w, kDropSlotAny); + } + +#if defined(Q_OS_WIN) + if (!windowsElevatedHintShown_ && is_windows_process_elevated()) { + windowsElevatedHintShown_ = true; +#ifndef NDEBUG + appendLogLine_( + "Notice: Running as Administrator can block Explorer drag & drop. " + "Run Brokkr normally for reliable DnD."); +#endif + } +#endif +} + +BrokkrWrapper::~BrokkrWrapper() { + stopWirelessListener_(); +#if defined(BROKKR_PLATFORM_LINUX) + if (uevent_notifier_) uevent_notifier_->setEnabled(false); + if (uevent_fd_ >= 0) { + ::close(uevent_fd_); + uevent_fd_ = -1; + } +#endif +#if defined(Q_OS_MACOS) + if (mac_notify_port_) { + CFRunLoopRemoveSource(CFRunLoopGetMain(), IONotificationPortGetRunLoopSource(mac_notify_port_), + kCFRunLoopDefaultMode); + IONotificationPortDestroy(mac_notify_port_); + mac_notify_port_ = nullptr; + } + if (mac_added_iter_) { + IOObjectRelease(mac_added_iter_); + mac_added_iter_ = 0; + } + if (mac_removed_iter_) { + IOObjectRelease(mac_removed_iter_); + mac_removed_iter_ = 0; + } +#endif +} + +void BrokkrWrapper::appendLogLineFromEngine(const QString& html) { appendLogLine_(html); } + +void BrokkrWrapper::closeEvent(QCloseEvent* e) { + if (busy_) { + QMessageBox::warning(this, "Brokkr Flash", "An operation is in progress. Please wait for it to complete."); + e->ignore(); + return; + } + e->accept(); +} + +QString BrokkrWrapper::dropSlotName_(int slotId) { + switch (slotId) { + case kDropSlotBL: return "BL"; + case kDropSlotAP: return "AP"; + case kDropSlotCP: return "CP"; + case kDropSlotCSC: return "CSC"; + case kDropSlotUSERDATA: return "USERDATA"; + case kDropSlotPIT: return "PIT"; + default: return "UNKNOWN"; + } +} + +int BrokkrWrapper::inferDropSlotFromFileName_(const QString& fileName) { + const QString leaf = QFileInfo(fileName).fileName(); + if (leaf.isEmpty()) return kDropSlotUnknown; + + if (is_pit_drop_name(leaf)) return kDropSlotPIT; + + const QString upper = leaf.toUpper(); + + if (contains_token(upper, "HOME_CSC")) return kDropSlotCSC; + if (contains_token(upper, "USERDATA")) return kDropSlotUSERDATA; + if (contains_token(upper, "CSC")) return kDropSlotCSC; + if (contains_token(upper, "BL")) return kDropSlotBL; + if (contains_token(upper, "AP")) return kDropSlotAP; + if (contains_token(upper, "CP")) return kDropSlotCP; + + return kDropSlotUnknown; +} + +bool BrokkrWrapper::isDropFileAllowedForSlot_(int slotId, const QString& fileName) { + if (slotId == kDropSlotPIT) return is_pit_drop_name(fileName); + if (slotId >= kDropSlotBL && slotId <= kDropSlotUSERDATA) return is_firmware_drop_name(fileName); + return false; +} + +QStringList BrokkrWrapper::extractLocalDropFiles_(const QMimeData* mime) const { + QStringList out; + if (!mime) return out; + + QSet seen; + + auto add_if_local = [&](const QString& pathLike) { + const QString p = QDir::cleanPath(pathLike.trimmed()); + if (p.isEmpty()) return; + const QString key = drop_file_identity_key(p); + if (key.isEmpty() || seen.contains(key)) return; + seen.insert(key); + out.push_back(p); + }; + + if (mime->hasUrls()) { + for (const QUrl& u : mime->urls()) { + if (!u.isLocalFile()) continue; + add_if_local(u.toLocalFile()); + } + } + + if (mime->hasText()) { + const auto lines = mime->text().split(QRegularExpression("[\\r\\n]+"), Qt::SkipEmptyParts); + for (const auto& line : lines) { + const QString t = line.trimmed(); + if (t.isEmpty()) continue; + const QUrl maybeUrl(t); + if (maybeUrl.isValid() && maybeUrl.isLocalFile()) { + add_if_local(maybeUrl.toLocalFile()); + } else if (QFileInfo(t).isAbsolute()) { + add_if_local(t); + } + } + } + +#if defined(Q_OS_WIN) + static constexpr const char* kFmtFileNameW = "application/x-qt-windows-mime;value=\"FileNameW\""; + static constexpr const char* kFmtFileNameA = "application/x-qt-windows-mime;value=\"FileName\""; + + if (mime->hasFormat(kFmtFileNameW)) { + for (const auto& p : parse_windows_filenamew_payload(mime->data(kFmtFileNameW))) add_if_local(p); + } + if (mime->hasFormat(kFmtFileNameA)) { + const auto raw = mime->data(kFmtFileNameA); + for (const auto& part : QString::fromLocal8Bit(raw).split(QChar('\0'), Qt::SkipEmptyParts)) add_if_local(part); + } +#endif + + return out; +} + +void BrokkrWrapper::bindDropTarget_(QWidget* widget, int slotId) { + if (!widget) return; + widget->setAcceptDrops(true); + widget->setProperty(kDropSlotProperty, slotId); + widget->removeEventFilter(this); + widget->installEventFilter(this); +} + +bool BrokkrWrapper::assignDroppedFileToSlot_(int slotId, const QString& localPath, QString* reason, bool* replaced) { + if (replaced) *replaced = false; + + QFileInfo fi(localPath); + if (!fi.exists()) { + if (reason) *reason = "File does not exist."; + return false; + } + if (!fi.isFile()) { + if (reason) *reason = "Dropped item is not a regular file."; + return false; + } + if (!fi.isReadable()) { + if (reason) *reason = "File is not readable."; + return false; + } + if (!isDropFileAllowedForSlot_(slotId, fi.fileName())) { + if (reason) { + if (slotId == kDropSlotPIT) + *reason = "PIT slot accepts only .pit files."; + else + *reason = "Firmware slots accept only .tar or .md5 files."; + } + return false; + } + + const QString finalPath = QDir::toNativeSeparators(fi.absoluteFilePath()); + lastDir = fi.absolutePath(); + + if (slotId == kDropSlotPIT) { + if (!chkUsePit || !editPit) { + if (reason) *reason = "PIT controls are unavailable."; + return false; + } + + const QString old = editPit->text().trimmed(); + if (replaced) *replaced = !old.isEmpty() && old.compare(finalPath, Qt::CaseInsensitive) != 0; + + if (!chkUsePit->isChecked()) chkUsePit->setChecked(true); + editPit->setText(finalPath); + return true; + } + + if (slotId < kDropSlotBL || slotId > kDropSlotUSERDATA) { + if (reason) *reason = "Unknown destination slot."; + return false; + } + if (slotId >= fileLineEdits_.size() || slotId >= fileChecks_.size()) { + if (reason) *reason = "Destination slot is not available in this UI."; + return false; + } + + auto* lineEdit = fileLineEdits_[slotId]; + auto* check = fileChecks_[slotId]; + if (!lineEdit || !check) { + if (reason) *reason = "Destination slot controls are not available."; + return false; + } + + const QString old = lineEdit->text().trimmed(); + if (replaced) *replaced = !old.isEmpty() && old.compare(finalPath, Qt::CaseInsensitive) != 0; + + lineEdit->setText(finalPath); + check->setEnabled(true); + check->setChecked(true); + return true; +} + +void BrokkrWrapper::handleDroppedFiles_(const QStringList& localFiles, int forcedSlotId) { + if (busy_) { + showBlocked_("Cannot accept files", "Cannot change files while flashing is in progress."); + return; + } + if (localFiles.isEmpty()) return; + + QStringList accepted; + QStringList rejected; + QSet usedSlots; + QSet seenFiles; + + for (const QString& p : localFiles) { + const QString fileKey = drop_file_identity_key(p); + if (!fileKey.isEmpty() && seenFiles.contains(fileKey)) { + // Ignore duplicate payload entries (e.g., Windows long-path + 8.3 alias of the same file). + continue; + } + if (!fileKey.isEmpty()) seenFiles.insert(fileKey); + + const QFileInfo fi(p); + const QString leaf = fi.fileName().isEmpty() ? p : fi.fileName(); + + int slotId = forcedSlotId; + if (slotId == kDropSlotAny) slotId = inferDropSlotFromFileName_(leaf); + if (slotId == kDropSlotUnknown) { + rejected.push_back( + QString("%1 -> rejected (cannot infer destination: expected AP/BL/CP/CSC/USERDATA token or .pit)") + .arg(leaf)); + continue; + } + + if (usedSlots.contains(slotId)) { + rejected.push_back(QString("%1 -> rejected (%2 already assigned in this drop)").arg(leaf, dropSlotName_(slotId))); + continue; + } + + QString reason; + bool replaced = false; + if (!assignDroppedFileToSlot_(slotId, p, &reason, &replaced)) { + rejected.push_back(QString("%1 -> rejected (%2)").arg(leaf, reason)); + continue; + } + + usedSlots.insert(slotId); + accepted.push_back(QString("%1 -> %2%3") + .arg(leaf, dropSlotName_(slotId), replaced ? " (replaced existing file)" : "")); + } + + updateActionButtons_(); + +#ifndef NDEBUG + for (const auto& msg : accepted) + appendLogLine_(QString("DnD route: %1").arg(htmlEsc(msg))); + for (const auto& msg : rejected) + appendLogLine_(QString("DnD reject: %1").arg(htmlEsc(msg))); +#endif + + if (accepted.isEmpty() && !rejected.isEmpty()) { + showBlocked_("Drop rejected", rejected.join("\n")); + } +} + +void BrokkrWrapper::dragEnterEvent(QDragEnterEvent* e) { + if (busy_) { + e->ignore(); + return; + } + const auto files = extractLocalDropFiles_(e->mimeData()); + if (files.isEmpty()) { + e->ignore(); + return; + } + e->acceptProposedAction(); +} + +void BrokkrWrapper::dragMoveEvent(QDragMoveEvent* e) { + if (busy_) { + e->ignore(); + return; + } + const auto files = extractLocalDropFiles_(e->mimeData()); + if (files.isEmpty()) { + e->ignore(); + return; + } + e->acceptProposedAction(); +} + +void BrokkrWrapper::dropEvent(QDropEvent* e) { + if (busy_) { + e->ignore(); + return; + } + const auto files = extractLocalDropFiles_(e->mimeData()); + if (files.isEmpty()) { +#ifndef NDEBUG + appendLogLine_( + "DnD reject: unsupported payload (drop local files from Explorer)"); +#endif + e->ignore(); + return; + } + handleDroppedFiles_(files, kDropSlotAny); + e->acceptProposedAction(); +} + +bool BrokkrWrapper::eventFilter(QObject* watched, QEvent* event) { + if (!watched || !event) return QWidget::eventFilter(watched, event); + + auto* ww = qobject_cast(watched); + if (!ww) return QWidget::eventFilter(watched, event); + if (ww != this && !isAncestorOf(ww)) return QWidget::eventFilter(watched, event); + + int forcedSlotId = kDropSlotAny; + const QVariant slotVar = ww->property(kDropSlotProperty); + if (slotVar.isValid()) forcedSlotId = slotVar.toInt(); + + switch (event->type()) { + case QEvent::DragEnter: { + auto* de = static_cast(event); + if (!busy_ && !extractLocalDropFiles_(de->mimeData()).isEmpty()) + de->acceptProposedAction(); + else + de->ignore(); + return true; + } + case QEvent::DragMove: { + auto* de = static_cast(event); + if (!busy_ && !extractLocalDropFiles_(de->mimeData()).isEmpty()) + de->acceptProposedAction(); + else + de->ignore(); + return true; + } + case QEvent::Drop: { + auto* de = static_cast(event); + if (busy_) { + de->ignore(); + return true; + } + const auto files = extractLocalDropFiles_(de->mimeData()); + if (files.isEmpty()) { +#ifndef NDEBUG + appendLogLine_( + "DnD reject: unsupported payload (drop local files from Explorer)"); +#endif + de->ignore(); + return true; + } + handleDroppedFiles_(files, forcedSlotId); + de->acceptProposedAction(); + return true; + } + default: break; + } + + return QWidget::eventFilter(watched, event); +} + +void BrokkrWrapper::childEvent(QChildEvent* e) { + QWidget::childEvent(e); + if (!e || !e->added()) return; + + auto* w = qobject_cast(e->child()); + if (!w || w == this) return; + + if (!w->property(kDropSlotProperty).isValid()) bindDropTarget_(w, kDropSlotAny); + + for (auto* sub : w->findChildren()) { + if (!sub || sub == this) continue; + if (sub->property(kDropSlotProperty).isValid()) continue; + bindDropTarget_(sub, kDropSlotAny); + } +} + +void BrokkrWrapper::changeEvent(QEvent* e) { + QWidget::changeEvent(e); + if (e->type() == QEvent::PaletteChange || e->type() == QEvent::ApplicationPaletteChange) { + applyHeaderStyle_(); + + { + auto pal = palette(); + QColor bg = pal.color(QPalette::Window).darker(110); + pal.setColor(QPalette::Window, bg); + idComGroup_->setPalette(pal); + } + + consoleOutput->setPalette(palette()); + refreshDeviceBoxes_(); + } +} + +#if defined(Q_OS_WIN) +bool BrokkrWrapper::nativeEvent(const QByteArray& eventType, void* message, qintptr* result) { + (void)eventType; + MSG* msg = reinterpret_cast(message); + if (!msg) return false; + + if (msg->message == WM_DROPFILES) { + HDROP hdrop = reinterpret_cast(msg->wParam); + int forcedSlotId = kDropSlotAny; + + POINT pt{}; + if (DragQueryPoint(hdrop, &pt)) { + QWidget* hit = childAt(QPoint(pt.x, pt.y)); + while (hit) { + const QVariant v = hit->property(kDropSlotProperty); + if (v.isValid()) { + forcedSlotId = v.toInt(); + break; + } + if (hit == this) break; + hit = hit->parentWidget(); + } + } + + const auto files = parse_hdrop_files(hdrop); + DragFinish(hdrop); + + if (!files.isEmpty()) { + handleDroppedFiles_(files, forcedSlotId); + } else { +#ifndef NDEBUG + appendLogLine_( + "DnD reject: WM_DROPFILES payload did not contain local files"); +#endif + } + + if (result) *result = 0; + return true; + } + + if (msg->message == WM_DEVICECHANGE) { + if (msg->wParam == DBT_DEVICEARRIVAL || msg->wParam == DBT_DEVICEREMOVECOMPLETE) requestUsbRefresh_(); + } + + if (result) *result = 0; + return false; +} +#endif + +void BrokkrWrapper::requestUsbRefresh_() noexcept { + spdlog::debug("USB refresh requested"); + usbDirty_.store(true, std::memory_order_relaxed); +} + +void BrokkrWrapper::refreshConnectedDevices_() { + QStringList shown; + QStringList physicalUsb; + + for (const auto& d : enumerate_samsung_targets()) physicalUsb << QString::fromStdString(d.sysname); + + bool physicalWireless = false; + QString physicalWirelessId; + { + std::lock_guard lk(wireless_mtx_); + const bool wantWireless = (chkWireless && chkWireless->isChecked()); + if (wantWireless) { + if (wireless_conn_ && wireless_conn_->connected() && !wireless_sysname_.isEmpty()) { + physicalWireless = true; + physicalWirelessId = wireless_sysname_; + } else { + wireless_conn_.reset(); + wireless_sysname_.clear(); + } + } + } + + const int physicalCount = physicalUsb.size() + (physicalWireless ? 1 : 0); + logDevCount_.store(physicalCount, std::memory_order_relaxed); + + const QString tgt = editTarget ? editTarget->text().trimmed() : QString{}; + if (!tgt.isEmpty()) { + auto info = brokkr::platform::find_by_sysname(tgt.toStdString()); + if (info) shown << QString::fromStdString(info->sysname); + } else { + shown = physicalUsb; + } + + if (physicalWireless) shown.prepend(physicalWirelessId); + + if (busy_) return; + + { + const QSet prev = QSet(physicalUsbPrev_.begin(), physicalUsbPrev_.end()); + const QSet now = QSet(physicalUsb.begin(), physicalUsb.end()); + + bool anyAttached = false; + + for (const auto& s : now) { + if (!prev.contains(s)) { + spdlog::info("Connected: {}", s.toStdString()); + anyAttached = true; + } + } + for (const auto& s : prev) + if (!now.contains(s)) spdlog::info("Disconnected: {}", s.toStdString()); + + if (physicalWirelessPrev_ && (!physicalWireless || physicalWirelessIdPrev_ != physicalWirelessId)) { + if (!physicalWirelessIdPrev_.isEmpty()) spdlog::info("Disconnected: {}", physicalWirelessIdPrev_.toStdString()); + } + if (physicalWireless && (!physicalWirelessPrev_ || physicalWirelessIdPrev_ != physicalWirelessId)) { + if (!physicalWirelessId.isEmpty()) { + spdlog::info("Connected: {}", physicalWirelessId.toStdString()); + anyAttached = true; + } + } + + physicalUsbPrev_ = physicalUsb; + physicalWirelessPrev_ = physicalWireless; + physicalWirelessIdPrev_ = physicalWirelessId; + } + + const QStringList prevShown = connectedDevices_; + connectedDevices_ = shown; + + auto clearRowState = [this](int i) { + if (i < 0 || i >= devSquares_.size()) return; + if (static_cast(i) < slotFailed_.size()) slotFailed_[static_cast(i)] = 0; + if (static_cast(i) < slotActive_.size()) slotActive_[static_cast(i)] = 0; + + if (auto* sq = devSquares_[i]) { + sq->setVariant(DeviceSquare::Variant::Green); + sq->setText(""); + sq->setFill(0.0); + } + }; + + const int rows = devSquares_.size(); + for (int i = 0; i < rows; ++i) { + const QString prevSys = (i < prevShown.size()) ? prevShown[i].trimmed() : QString{}; + const QString nowSys = (i < connectedDevices_.size()) ? connectedDevices_[i].trimmed() : QString{}; + if (!nowSys.isEmpty() && prevSys != nowSys) clearRowState(i); + } + + refreshDeviceBoxes_(); + updateHeaderLeds_(); + updateActionButtons_(); +} + +void BrokkrWrapper::startWirelessListener_() { + if (busy_) return; + + stopWirelessListener_(); + + wireless_sysname_.clear(); + { + std::lock_guard lk(wireless_mtx_); + wireless_conn_.reset(); + wireless_listener_.emplace(); + auto st = wireless_listener_->bind_and_listen("0.0.0.0", 13579); + if (!st) { + spdlog::error("Wireless listen failed: {}", st.error()); + wireless_listener_.reset(); + requestUsbRefresh_(); + return; + } + } + + wireless_thread_ = std::jthread([this](std::stop_token st) { + for (;;) { + if (st.stop_requested()) return; + + // While the engine owns the socket, do not touch wireless_conn_. + // Otherwise we may reset() it out from under a raw pointer the engine + // is actively using, or contend with its recv() via MSG_PEEK probes. + if (wireless_watcher_paused_.load(std::memory_order_acquire)) { + std::this_thread::sleep_for(std::chrono::milliseconds(200)); + continue; + } + + { + bool connected = false; + { + std::lock_guard lk(wireless_mtx_); + if (wireless_conn_) { + if (wireless_conn_->connected()) + connected = true; + else { + wireless_conn_.reset(); + wireless_sysname_.clear(); + QMetaObject::invokeMethod(this, [this]() { requestUsbRefresh_(); }, Qt::QueuedConnection); + } + } + } + if (connected) { + std::this_thread::sleep_for(std::chrono::milliseconds(200)); + continue; + } + } + + brokkr::platform::TcpListener* lst = nullptr; + { + std::lock_guard lk(wireless_mtx_); + if (!wireless_listener_) return; + lst = &*wireless_listener_; + } + + auto ar = lst->accept_one(); + if (!ar) { + if (st.stop_requested()) return; + std::this_thread::sleep_for(std::chrono::milliseconds(50)); + continue; + } + + const QString peer = QString::fromStdString(ar->peer_label()); + const int sep = peer.lastIndexOf(':'); + const QString sys = (sep > 0) ? peer.left(sep) : peer; + + { + std::lock_guard lk(wireless_mtx_); + wireless_conn_.emplace(std::move(*ar)); + wireless_sysname_ = sys; + } + + QMetaObject::invokeMethod(this, [this]() { refreshConnectedDevices_(); }, Qt::QueuedConnection); + } + }); + + refreshConnectedDevices_(); +} + +void BrokkrWrapper::stopWirelessListener_() { + brokkr::platform::TcpListener* lst = nullptr; + { + std::lock_guard lk(wireless_mtx_); + if (wireless_listener_) lst = &*wireless_listener_; + } + if (lst) lst->close(); + + if (wireless_thread_.joinable()) { + wireless_thread_.request_stop(); + wireless_thread_.join(); + } + + { + std::lock_guard lk(wireless_mtx_); + wireless_conn_.reset(); + wireless_listener_.reset(); + wireless_sysname_.clear(); + } + requestUsbRefresh_(); +} + +void BrokkrWrapper::appendLogLine_(const QString& html) { + consoleOutput->append(html); + QTextCursor cursor = consoleOutput->textCursor(); + cursor.movePosition(QTextCursor::End); + consoleOutput->setTextCursor(cursor); +} + +void BrokkrWrapper::applyHeaderStyle_() { + if (!headerWidget_) return; + const int luma = palette().color(QPalette::Window).lightness(); + const bool dark = (luma < 128); + + headerWidget_->setStyleSheet(""); // Clear stylesheet to fallback natively + + QPalette pal = headerWidget_->palette(); + QColor bgColor = dark ? QColor("#2d2e30") : QColor("#dce0e4"); + pal.setColor(QPalette::Window, bgColor); + + headerWidget_->setAutoFillBackground(true); + headerWidget_->setPalette(pal); +} + +void BrokkrWrapper::updateHeaderLeds_() { + if (!ledContainer_) return; + + auto* lo = ledContainer_->layout(); + if (!lo) return; + + while (lo->count() > 0) { + QLayoutItem* it = lo->takeAt(0); + if (!it) break; + if (auto* w = it->widget()) delete w; + delete it; + } + + const int count = std::min(connectedDevices_.size(), 12); + for (int i = 0; i < count; ++i) { + auto* led = new QWidget(ledContainer_); + led->setFixedSize(14, 14); + + // NOTE: This single style sheet is maintained because there is no native Mac + // widget for a "glowing orb". Styling a pure QWidget does not break Mac theming. + led->setStyleSheet( + "background: qradialgradient(cx:0.3, cy:0.3, radius:0.7, fx:0.3, fy:0.3," + " stop:0 #aaddff, stop:1 #0066cc);" + "border: 1px solid #000;" + "border-radius: 7px;"); + auto* glow = new QGraphicsDropShadowEffect(led); + glow->setColor(QColor(0, 153, 255, 200)); + glow->setBlurRadius(10); + glow->setOffset(0, 0); + led->setGraphicsEffect(glow); + lo->addWidget(led); + } +} + +void BrokkrWrapper::setControlsEnabled_(bool enabled) { + if (tabWidget_) { + tabWidget_->setEnabled(true); + if (tabWidget_->count() > 0) tabWidget_->setTabEnabled(0, true); + for (int i = 1; i < tabWidget_->count(); ++i) tabWidget_->setTabEnabled(i, enabled); + if (!enabled) tabWidget_->setCurrentIndex(0); + } + if (consoleOutput) consoleOutput->setEnabled(true); + + if (editTarget) editTarget->setEnabled(enabled); + if (cmbRebootAction) cmbRebootAction->setEnabled(enabled); + + if (chkUsePit) chkUsePit->setEnabled(enabled); + if (editPit) editPit->setEnabled(enabled && chkUsePit && chkUsePit->isChecked()); + if (btnPitBrowse) btnPitBrowse->setEnabled(enabled && chkUsePit && chkUsePit->isChecked()); + + if (chkWireless) chkWireless->setEnabled(enabled); + + const bool wireless = (chkWireless && chkWireless->isChecked()); + + if (btnManyDevices_) btnManyDevices_->setEnabled(enabled && !wireless); + + for (auto* chk : fileChecks_) + (void)chk; + for (int i = 0; i < fileChecks_.size(); ++i) { + auto* chk = fileChecks_[i]; + auto* edit = (i < fileLineEdits_.size()) ? fileLineEdits_[i] : nullptr; + if (!chk) continue; + const bool has_file = (edit && !edit->text().trimmed().isEmpty()); + chk->setEnabled(enabled && has_file); + } + for (auto* btn : fileButtons_) + if (btn) btn->setEnabled(enabled); + + if (btnReset_) btnReset_->setEnabled(enabled); + + updateRebootDownloadButton_(); +} + +void BrokkrWrapper::setBusy_(bool busy) { + busy_ = busy; + if (busy_ && tabWidget_) tabWidget_->setCurrentIndex(0); + setControlsEnabled_(!busy_); + updateActionButtons_(); +} + +void BrokkrWrapper::setSquaresProgress_(double frac, bool animate) { + const int n = devSquares_.size(); + for (int i = 0; i < n; ++i) { + auto* sq = devSquares_[i]; + if (!sq) continue; + if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; + if (static_cast(i) < slotFailed_.size() && slotFailed_[static_cast(i)]) continue; + if (animate) + sq->setFillAnimated(frac, 120); + else + sq->setFill(frac); + } +} + +void BrokkrWrapper::setSquaresText_(const QString& s) { + const int n = devSquares_.size(); + for (int i = 0; i < n; ++i) { + auto* sq = devSquares_[i]; + if (!sq) continue; + if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; + if (static_cast(i) < slotFailed_.size() && slotFailed_[static_cast(i)]) continue; + sq->setText(s); + } +} + +void BrokkrWrapper::setSquaresActiveColor_(bool enhanced) { + enhanced_speed_ = enhanced; + const int n = devSquares_.size(); + for (int i = 0; i < n; ++i) { + auto* sq = devSquares_[i]; + if (!sq) continue; + if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; + if (static_cast(i) < slotFailed_.size() && slotFailed_[static_cast(i)]) continue; + sq->setVariant(enhanced ? DeviceSquare::Variant::Blue : DeviceSquare::Variant::Green); + } +} + +void BrokkrWrapper::setSquaresFinal_(bool ok) { + if (!ok) { + const int n = devSquares_.size(); + for (int i = 0; i < n; ++i) { + auto* sq = devSquares_[i]; + if (!sq) continue; + if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; + sq->setVariant(DeviceSquare::Variant::Red); + sq->setText("FAIL!"); + sq->setFillAnimated(1.0, 250); + } + return; + } + + const auto passV = (enhanced_speed_ ? DeviceSquare::Variant::Green : DeviceSquare::Variant::Blue); + const int n = devSquares_.size(); + for (int i = 0; i < n; ++i) { + auto* sq = devSquares_[i]; + if (!sq) continue; + if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; + const bool failed = (static_cast(i) < slotFailed_.size() && slotFailed_[static_cast(i)]); + if (failed) { + sq->setVariant(DeviceSquare::Variant::Red); + sq->setText("FAIL!"); + sq->setFillAnimated(1.0, 200); + } else { + sq->setVariant(passV); + sq->setText("PASS"); + sq->setFillAnimated(1.0, 350); + } + } +} + +void BrokkrWrapper::applyWindowHeightToContents_() { + if (!layout()) return; + layout()->invalidate(); + layout()->activate(); + const int want = layout()->sizeHint().height() + 30; + if (height() < want) resize(width(), want); +} + +void BrokkrWrapper::rebuildDeviceBoxes_(int boxCount, bool singleRow) { + while (idComLayout_->count() > 0) { + QLayoutItem* it = idComLayout_->takeAt(0); + if (!it) break; + if (auto* w = it->widget()) delete w; + delete it; + } + + for (int c = 0; c < kMassDlMaxBoxes + 2; ++c) idComLayout_->setColumnStretch(c, 0); + for (int r = 0; r < (kMassDlMaxBoxes / std::max(1, kBoxesColsMany)) + 2; ++r) idComLayout_->setRowStretch(r, 0); + + comBoxes.clear(); + devSquares_.clear(); + + const bool singleCellNoStretch = (singleRow && boxCount == 1); + + int fixedCellW = -1; + if (singleCellNoStretch) { + const int cols = kBoxesNormal; + int sp = idComLayout_->horizontalSpacing(); + if (sp < 0) sp = idComLayout_->spacing(); + const auto m = idComLayout_->contentsMargins(); + const int avail = std::max(0, idComGroup_->contentsRect().width() - m.left() - m.right()); + const int totalSp = std::max(0, (cols - 1) * std::max(0, sp)); + fixedCellW = (cols > 0) ? ((avail - totalSp) / cols) : -1; + if (fixedCellW < 30) fixedCellW = 30; + } + + auto make_cell = [&](int row, int col) { + auto* cell = new QWidget(idComGroup_); + auto* v = new QVBoxLayout(cell); + v->setContentsMargins(0, 0, 0, 0); + v->setSpacing(2); + + auto* sq = new DeviceSquare(cell); + + auto* box = new QLabel(cell); + box->setAlignment(Qt::AlignCenter); + box->setMinimumHeight(22); + box->setAutoFillBackground(true); + box->setBackgroundRole(QPalette::Base); + box->setForegroundRole(QPalette::Text); + box->setFrameShape(QFrame::StyledPanel); + box->setFrameShadow(QFrame::Sunken); + box->setMargin(1); + + // Use QFont instead of CSS for text size + QFont boxFont = box->font(); + boxFont.setPointSize(9); + box->setFont(boxFont); + + if (singleCellNoStretch && fixedCellW > 0) { + cell->setFixedWidth(fixedCellW); + box->setFixedWidth(fixedCellW); + sq->setMinimumWidth(fixedCellW); + sq->setMaximumWidth(fixedCellW); + } + + v->addWidget(sq); + v->addWidget(box); + + idComLayout_->addWidget(cell, row, col); + + devSquares_.append(sq); + comBoxes.append(box); + }; + + if (singleRow) { + if (boxCount > 0) make_cell(0, 0); + + if (singleCellNoStretch) { + idComLayout_->setColumnStretch(0, 0); + idComLayout_->addItem(new QSpacerItem(0, 0, QSizePolicy::Expanding, QSizePolicy::Minimum), 0, 1); + idComLayout_->setColumnStretch(1, 1); + } else { + for (int col = 1; col < boxCount; ++col) make_cell(0, col); + for (int col = 0; col < boxCount; ++col) idComLayout_->setColumnStretch(col, 1); + } + + slotFailed_.assign(static_cast(devSquares_.size()), 0); + slotActive_.assign(static_cast(devSquares_.size()), 0); + return; + } + + for (int c = 0; c < kBoxesColsMany; ++c) idComLayout_->setColumnStretch(c, 1); + + for (int i = 0; i < boxCount; ++i) { + const int row = i / kBoxesColsMany; + const int col = i % kBoxesColsMany; + make_cell(row, col); + } + + slotFailed_.assign(static_cast(devSquares_.size()), 0); + slotActive_.assign(static_cast(devSquares_.size()), 0); +} + +void BrokkrWrapper::refreshDeviceBoxes_() { + overflowDevices_ = false; + + for (auto* box : comBoxes) { + box->clear(); + box->setToolTip(QString()); + box->setStyleSheet(""); // Clear any leftover inline styling + + QPalette pal = box->palette(); + pal.setColor(QPalette::Base, palette().color(QPalette::Base)); + pal.setColor(QPalette::Text, palette().color(QPalette::Text)); + box->setPalette(pal); + + QFont f = box->font(); + f.setBold(false); + box->setFont(f); + } + + auto elideFor = [&](QLabel* box, const QString& s) { + const int w = std::max(0, box->width() - 10); + return box->fontMetrics().elidedText(s, Qt::ElideMiddle, w); + }; + + const int shown = std::min(connectedDevices_.size(), comBoxes.size()); + for (int i = 0; i < shown; ++i) { + const QString sysname = connectedDevices_[i].trimmed(); + const QString raw = QString("%1:[%2]").arg(i).arg(sysname); + comBoxes[i]->setText(elideFor(comBoxes[i], raw)); + comBoxes[i]->setToolTip(sysname); + + bool odin_mode = false; + if (chkWireless && chkWireless->isChecked() && !wireless_sysname_.isEmpty() && sysname == wireless_sysname_) { + odin_mode = true; + } else if (auto info = select_samsung_target(sysname)) { + odin_mode = is_odin_product(info->product); + } + + QPalette pal = comBoxes[i]->palette(); + if (odin_mode) { + pal.setColor(QPalette::Base, QColor("#4080c0")); // Odin mode: blue tint + pal.setColor(QPalette::Text, Qt::white); + } else { + pal.setColor(QPalette::Base, QColor("#d9822b")); // Non-Odin Samsung device: orange tint + pal.setColor(QPalette::Text, Qt::white); + } + comBoxes[i]->setPalette(pal); + + QFont f = comBoxes[i]->font(); + f.setBold(true); + comBoxes[i]->setFont(f); + } + + if (connectedDevices_.size() > comBoxes.size() && !comBoxes.isEmpty()) { + overflowDevices_ = true; + const int extra = connectedDevices_.size() - comBoxes.size(); + auto* last = comBoxes.back(); + const QString raw = QString("... +%1 more").arg(extra); + last->setText(elideFor(last, raw)); + + QPalette pal = last->palette(); + QColor base = palette().color(QPalette::Base); + QColor warn(255, 183, 77); + warn.setAlpha(50); + const int r2 = (base.red() * (255 - warn.alpha()) + warn.red() * warn.alpha()) / 255; + const int g2 = (base.green() * (255 - warn.alpha()) + warn.green() * warn.alpha()) / 255; + const int b2 = (base.blue() * (255 - warn.alpha()) + warn.blue() * warn.alpha()) / 255; + + pal.setColor(QPalette::Base, QColor(r2, g2, b2)); + pal.setColor(QPalette::Text, palette().color(QPalette::Text)); + last->setPalette(pal); + + QFont f = last->font(); + f.setBold(true); + last->setFont(f); + } +} + +bool BrokkrWrapper::canRunStart_(QString* whyNot) const { + const bool wireless = chkWireless->isChecked(); + const bool hasTarget = !editTarget->text().trimmed().isEmpty(); + + auto selected = [&](int idx, const QLineEdit* e) { + const bool checked = (idx < fileChecks_.size() && fileChecks_[idx] && fileChecks_[idx]->isChecked()); + return checked && e && !e->text().trimmed().isEmpty(); + }; + + const bool hasAnyFile = selected(0, editBL) || selected(1, editAP) || selected(2, editCP) || selected(3, editCSC) || + selected(4, editUserData) || + (chkUsePit->isChecked() && !editPit->text().isEmpty()); + + if (!hasAnyFile) { + if (whyNot) *whyNot = "No files selected."; + return false; + } + + if (wireless && hasTarget) { + if (whyNot) *whyNot = "Wireless cannot be used together with Target Sysname."; + return false; + } + + if (wireless) { + std::lock_guard lk(wireless_mtx_); + if (!wireless_conn_ || !wireless_conn_->connected()) { + if (whyNot) *whyNot = "Wireless is enabled but no device is connected yet."; + return false; + } + return true; + } + + if (hasTarget) return true; + + if (connectedDevices_.isEmpty()) { + if (whyNot) *whyNot = "No connected devices detected."; + return false; + } + if (overflowDevices_) { + if (whyNot) *whyNot = "Too many devices are connected for the current GUI box limit."; + return false; + } + return true; +} + +bool BrokkrWrapper::confirmOdinModeDevicesForStart_() { + if (chkWireless && chkWireless->isChecked()) return true; + + auto show_odin_popup = [this](const QString& text, bool allowIgnore) { + QMessageBox box(this); + box.setIcon(QMessageBox::Warning); + box.setWindowTitle("Brokkr Flash"); + box.setText(text); + + auto* ok_btn = box.addButton(QMessageBox::Ok); + #if defined(BROKKR_PLATFORM_WINDOWS) + auto* reboot_btn = box.addButton("Try to Reboot them into Download Mode", QMessageBox::ActionRole); + #endif + QAbstractButton* ignore_btn = nullptr; + if (allowIgnore) ignore_btn = box.addButton("Ignore them", QMessageBox::AcceptRole); + + box.setDefaultButton(qobject_cast(ok_btn)); + box.exec(); + +#if defined(BROKKR_PLATFORM_WINDOWS) + if (box.clickedButton() == reboot_btn) { + tryRebootIntoDownloadMode_(); + return false; + } +#endif + return allowIgnore && ignore_btn && box.clickedButton() == ignore_btn; + }; + + const QString tgt = editTarget ? editTarget->text().trimmed() : QString{}; + if (!tgt.isEmpty()) { + if (auto one = select_samsung_target(tgt); one && !is_odin_product(one->product)) { + return show_odin_popup("The device is not in Odin Mode, reboot to download mode first.", false); + } + } + + const auto all_samsung = enumerate_samsung_targets(); + if (all_samsung.empty()) return true; + + QStringList not_odin; + int odin_count = 0; + for (const auto& d : all_samsung) { + if (is_odin_product(d.product)) { + ++odin_count; + } else { + not_odin << QString::fromStdString(d.sysname); + } + } + + if (not_odin.isEmpty()) return true; + + const int total = static_cast(all_samsung.size()); + + if (total == 1) { + return show_odin_popup("The device is not in Odin Mode, reboot to download mode first.", false); + } + + if (odin_count == 0) { + return show_odin_popup("None of the devices are in Odin Mode", false); + } + + return show_odin_popup(QString("The following device(s) are not in Odin Mode:\n%1").arg(not_odin.join("\n")), + true); +} + +void BrokkrWrapper::updateActionButtons_() { + if (busy_) { + btnRun->setEnabled(false); + updateRebootDownloadButton_(); + return; + } + + QString why; + btnRun->setEnabled(canRunStart_(&why)); + updateRebootDownloadButton_(); +} + +void BrokkrWrapper::updateRebootDownloadButton_() { + if (!btnRebootDownloadMode_) return; +#if !defined(BROKKR_PLATFORM_WINDOWS) + btnRebootDownloadMode_->setEnabled(false); + btnRebootDownloadMode_->hide(); + return; +#endif + if (busy_) { + btnRebootDownloadMode_->setEnabled(false); + return; + } + + bool hasNonOdin = false; + for (const auto& sys : connectedDevices_) { + if (auto one = select_samsung_target(sys); one && !is_odin_product(one->product)) { + hasNonOdin = true; + break; + } + } + + btnRebootDownloadMode_->setEnabled(hasNonOdin); +} + +void BrokkrWrapper::showBlocked_(const QString& title, const QString& msg) const { + QMessageBox::warning(const_cast(this), title, msg); +} + +void BrokkrWrapper::tryRebootIntoDownloadMode_() { + if (busy_) return; + +#if !defined(BROKKR_PLATFORM_WINDOWS) + QMessageBox::information(this, "Brokkr Flash", "This action is available only on Windows builds."); + return; +#else + + int nonOdinCount = 0; + for (const auto& sys : connectedDevices_) { + if (auto one = select_samsung_target(sys); one && !is_odin_product(one->product)) { + ++nonOdinCount; + } + } + + if (nonOdinCount <= 0) { + QMessageBox::information(this, "Brokkr Flash", "No connected device needs reboot into Download Mode."); + return; + } + + const auto r = brokkr::platform::send_suddlmod_to_samsung_serial(SAMSUNG_VID, nonOdinCount); + if (r.ports_seen == 0) { + QMessageBox::warning(this, "Brokkr Flash", "No Samsung serial port found."); + return; + } + + if (r.sent_ok == 0) { + QString detail; + for (const auto& f : r.failures) { + if (!detail.isEmpty()) detail += "\n"; + detail += QString::fromStdString(f); + } + QMessageBox::warning(this, "Brokkr Flash", + detail.isEmpty() ? "Failed to send reboot command to Samsung serial ports." + : QString("Failed to send reboot command to Samsung serial ports.\n%1") + .arg(detail)); + return; + } + + if (r.failures.empty()) { + QMessageBox::information(this, "Brokkr Flash", + QString("Reboot command sent to %1 Samsung serial port(s).") + .arg(r.sent_ok)); + } else { + QString detail; + for (const auto& f : r.failures) { + if (!detail.isEmpty()) detail += "\n"; + detail += QString::fromStdString(f); + } + QMessageBox::warning(this, "Brokkr Flash", + QString("Reboot command sent to %1 port(s), but %2 failed.\n%3") + .arg(r.sent_ok) + .arg(static_cast(r.failures.size())) + .arg(detail)); + } + + requestUsbRefresh_(); +#endif +} + +void BrokkrWrapper::setupOdinFileInput(QGridLayout* layout, int row, const QString& label, QLineEdit*& lineEdit) { + auto* chk = new QCheckBox(this); + chk->setEnabled(false); + layout->addWidget(chk, row, 0); + + auto* btn = new QPushButton(label, this); + btn->setMinimumWidth(95); +#if defined(Q_OS_MACOS) + btn->setFixedHeight(32); +#else + btn->setFixedHeight(28); +#endif + layout->addWidget(btn, row, 1); + + lineEdit = new QLineEdit(this); + lineEdit->setReadOnly(true); + layout->addWidget(lineEdit, row, 2); + + fileChecks_.append(chk); + fileButtons_.append(btn); + fileLineEdits_.append(lineEdit); + + bindDropTarget_(chk, row); + bindDropTarget_(btn, row); + bindDropTarget_(lineEdit, row); + + connect(btn, &QPushButton::clicked, this, [this, lineEdit, chk]() { + if (busy_) return; + QFileDialog::Options opts; +#if defined(Q_OS_MACOS) || defined(Q_OS_LINUX) + opts |= QFileDialog::DontUseNativeDialog; +#endif + QString file = QFileDialog::getOpenFileName(this, "Select Firmware File", lastDir, + "Firmware Archives (*.tar *.md5);;All Files (*)", + nullptr, opts); + if (!file.isEmpty()) { + lastDir = QFileInfo(file).absolutePath(); + lineEdit->setText(file); + chk->setEnabled(true); + chk->setChecked(true); + } + updateActionButtons_(); + }); + + connect(chk, &QCheckBox::toggled, this, [this](bool) { + if (busy_) return; + updateActionButtons_(); + }); + + connect(lineEdit, &QLineEdit::textChanged, this, [this, chk](const QString& txt) { + if (busy_) return; + const bool has_file = !txt.trimmed().isEmpty(); + chk->setEnabled(has_file); + if (!has_file) chk->setChecked(false); + updateActionButtons_(); + }); +} + +void BrokkrWrapper::onRunClicked() { + if (busy_) return; + QString why; + if (!canRunStart_(&why)) { + showBlocked_("Cannot start", why); + return; + } + if (!confirmOdinModeDevicesForStart_()) return; + startWorkStart_(); +} + +void BrokkrWrapper::startWorkStart_() { + if (busy_) return; + + const QStringList uiDevicesSnapshot = connectedDevices_; + setBusy_(true); + + plan_names_.clear(); + plan_from_names_.clear(); + enhanced_speed_ = false; + + slotFailed_.assign(static_cast(devSquares_.size()), 0); + slotActive_.assign(static_cast(devSquares_.size()), 0); + { + const int activeCount = std::min(uiDevicesSnapshot.size(), devSquares_.size()); + for (int i = 0; i < activeCount; ++i) { + bool active = false; + const QString sysname = uiDevicesSnapshot[i].trimmed(); + if (chkWireless && chkWireless->isChecked() && !wireless_sysname_.isEmpty() && sysname == wireless_sysname_) { + active = true; + } else if (auto info = select_odin_target(sysname)) { + (void)info; + active = true; + } + slotActive_[static_cast(i)] = active ? 1 : 0; + } + } + + int progressSteps = 1; + for (int i = 0; i < devSquares_.size(); ++i) { + auto* sq = devSquares_[i]; + if (!sq) continue; + if (static_cast(i) < slotActive_.size() && !slotActive_[static_cast(i)]) continue; + progressSteps = std::max(progressSteps, std::max(sq->width(), sq->sizeHint().width())); + } + progressVisualSteps_.store(progressSteps, std::memory_order_relaxed); + progressVisualBucket_.store(-1, std::memory_order_relaxed); + + setSquaresProgress_(0.0, false); + setSquaresText_(""); + setSquaresActiveColor_(false); + + // Block the wireless watcher from touching the TcpConnection while the + // engine owns it. Cleared from done_ui / fail_ui below. + wireless_watcher_paused_.store(true, std::memory_order_release); + + worker_ = std::jthread([this, uiDevicesSnapshot](std::stop_token) { + auto done_ui = [&] { + QMetaObject::invokeMethod( + this, + [this]() { + wireless_watcher_paused_.store(false, std::memory_order_release); + setSquaresText_("PASS"); + setSquaresFinal_(true); + setBusy_(false); + }, + Qt::QueuedConnection); + }; + + auto fail_ui = [&](const QString& msg, bool showPopup = false) { + QMetaObject::invokeMethod( + this, + [this, msg, showPopup]() { + wireless_watcher_paused_.store(false, std::memory_order_release); + const int z = logDevCount_.load(std::memory_order_relaxed); + appendLogLine_(QString("<%1> FAIL! %2").arg(z).arg(htmlEsc(msg))); + setSquaresFinal_(false); + setBusy_(false); + if (showPopup) QMessageBox::warning(this, "Brokkr Flash", "The connected devices do not match!"); + }, + Qt::QueuedConnection); + }; + + const int actionIndex = cmbRebootAction->currentIndex(); + + const QString tgt = editTarget->text().trimmed(); + const bool wireless = chkWireless->isChecked(); + if (wireless && !tgt.isEmpty()) { + fail_ui("Wireless cannot be used together with Target Sysname."); + return; + } + + QMetaObject::invokeMethod(this, [this]() { setSquaresText_("HANDSHAKE"); }, Qt::QueuedConnection); + + brokkr::odin::Cfg cfg; + cfg.reboot_after = (actionIndex == 0); + + brokkr::odin::Ui ui; + std::atomic_bool sawPerDeviceFail{false}; + std::vector targetToUiSlots; + + auto findUiIndexBySysname = [&](const QString& sysname) { + const QString needle = sysname.trimmed(); + for (int i = 0; i < uiDevicesSnapshot.size(); ++i) { + if (uiDevicesSnapshot[i].trimmed() == needle) return i; + } + return -1; + }; + + ui.on_plan = [&](const std::vector& p, std::uint64_t) { + std::vector parts, froms; + parts.reserve(p.size()); + froms.reserve(p.size()); + for (const auto& it : p) { + std::string part = it.part_name; + if (part.empty() && it.kind == brokkr::odin::PlanItem::Kind::Pit) part = "PIT"; + std::string from = it.pit_file_name; + if (from.empty()) from = part; + parts.push_back(QString::fromStdString(part)); + froms.push_back(QString::fromStdString(from)); + } + QMetaObject::invokeMethod( + this, + [this, parts = std::move(parts), froms = std::move(froms)]() mutable { + plan_names_ = std::move(parts); + plan_from_names_ = std::move(froms); + }, + Qt::QueuedConnection); + }; + + ui.on_item_active = [&](std::size_t i) { + QMetaObject::invokeMethod( + this, + [this, i]() { + if (i >= plan_names_.size()) return; + setSquaresText_(plan_names_[i]); + }, + Qt::QueuedConnection); + }; + + ui.on_stage = [&](const std::string& s) { + const QString qs = QString::fromStdString(s); + QMetaObject::invokeMethod( + this, + [this, qs]() { + if (qs.contains("Enhanced", Qt::CaseInsensitive)) + setSquaresActiveColor_(true); + else if (qs.contains("Normal", Qt::CaseInsensitive)) + setSquaresActiveColor_(false); + if (qs.contains("xxh3", Qt::CaseInsensitive)) setSquaresText_("XXH3"); + if (qs.contains("md5", Qt::CaseInsensitive)) setSquaresText_("MD5"); + if (qs.contains("handshake", Qt::CaseInsensitive)) setSquaresText_("HANDSHAKE"); + if (qs.contains("shutdown", Qt::CaseInsensitive) || qs.contains("reboot", Qt::CaseInsensitive) || + qs.contains("reset", Qt::CaseInsensitive) || qs.contains("finalizing", Qt::CaseInsensitive)) + setSquaresText_("RESET"); + }, + Qt::QueuedConnection); + }; + + ui.on_progress = [&](std::uint64_t d, std::uint64_t t, std::uint64_t, std::uint64_t) { + const double frac = (t > 0) ? (static_cast(d) / static_cast(t)) : 0.0; + const int steps = std::max(1, progressVisualSteps_.load(std::memory_order_relaxed)); + const int bucket = (t > 0) + ? static_cast(std::clamp((d * static_cast(steps)) / t, 0, + static_cast(steps))) + : 0; + + int prev = progressVisualBucket_.load(std::memory_order_relaxed); + while (prev != bucket && + !progressVisualBucket_.compare_exchange_weak(prev, bucket, std::memory_order_relaxed, + std::memory_order_relaxed)) { + } + if (prev == bucket) return; + + QMetaObject::invokeMethod(this, [this, frac]() { setSquaresProgress_(frac, true); }, Qt::QueuedConnection); + }; + + ui.on_error = [&](const std::string& s) { + if (s.rfind("DEVFAIL idx=", 0) == 0) sawPerDeviceFail.store(true, std::memory_order_relaxed); + const std::vector mapSnapshot = targetToUiSlots; + QMetaObject::invokeMethod( + this, + [this, qs = QString::fromStdString(s), mapSnapshot]() { + const int z = logDevCount_.load(std::memory_order_relaxed); + + QString shown = qs; + int idx = -1; + + const QString pref = "DEVFAIL idx="; + if (qs.startsWith(pref)) { + int p = pref.size(); + int sp = qs.indexOf(' ', p); + if (sp < 0) sp = qs.size(); + bool okNum = false; + idx = qs.mid(p, sp - p).toInt(&okNum); + QString reason; + if (sp < qs.size()) reason = qs.mid(sp + 1).trimmed(); + + if (okNum && idx >= 0) { + int uiIdx = -1; + if (static_cast(idx) < mapSnapshot.size()) uiIdx = mapSnapshot[static_cast(idx)]; + + if (uiIdx >= 0) { + if (static_cast(uiIdx) >= slotFailed_.size()) + slotFailed_.resize(static_cast(uiIdx) + 1, 0); + slotFailed_[static_cast(uiIdx)] = 1; + } + + if (uiIdx >= 0 && uiIdx < devSquares_.size() && devSquares_[uiIdx]) { + devSquares_[uiIdx]->setVariant(DeviceSquare::Variant::Red); + devSquares_[uiIdx]->setText("FAIL!"); + devSquares_[uiIdx]->setFill(1.0); + } + + // Keep native but set red palette for failure + if (uiIdx >= 0 && uiIdx < comBoxes.size() && comBoxes[uiIdx]) { + QPalette pal = comBoxes[uiIdx]->palette(); + pal.setColor(QPalette::Base, QColor("#a02020")); + pal.setColor(QPalette::Text, Qt::white); + comBoxes[uiIdx]->setPalette(pal); + } + + if (reason.isEmpty()) { + shown = (uiIdx >= 0) ? QString("FAIL! (Device %1)").arg(uiIdx) : QString("FAIL!"); + } else { + shown = (uiIdx >= 0) ? QString("FAIL! (Device %1) %2").arg(uiIdx).arg(reason) + : QString("FAIL! %1").arg(reason); + } + } + } + + appendLogLine_(QString("<%1> %2").arg(z).arg(htmlEsc(shown))); + }, + Qt::QueuedConnection); + }; + + std::shared_ptr> pit_to_upload; + if (chkUsePit->isChecked() && !editPit->text().isEmpty()) { + const std::filesystem::path p = editPit->text().toStdString(); + std::error_code ec; + const auto sz = std::filesystem::file_size(p, ec); + if (ec) { + fail_ui("Cannot stat PIT file."); + return; + } + + std::vector buf(static_cast(sz)); + std::ifstream in(p, std::ios::binary); + if (!in.is_open()) { + fail_ui("Cannot open PIT file."); + return; + } + if (!buf.empty()) { + in.read(reinterpret_cast(buf.data()), static_cast(buf.size())); + if (!in.good()) { + fail_ui("Failed to read PIT file."); + return; + } + } + pit_to_upload = std::make_shared>(std::move(buf)); + } + + std::vector inputs; + auto push_if_selected = [&](int idx, QLineEdit* e) { + if (!e) return; + const bool checked = (idx < fileChecks_.size() && fileChecks_[idx] && fileChecks_[idx]->isChecked()); + if (checked && !e->text().trimmed().isEmpty()) inputs.emplace_back(e->text().toStdString()); + }; + + push_if_selected(0, editBL); + push_if_selected(1, editAP); + push_if_selected(2, editCP); + push_if_selected(3, editCSC); + push_if_selected(4, editUserData); + + struct Provider { + std::vector> usb; // owns USB transports + std::vector owned; // owns Targets + std::vector ptrs; // passed to engine + }; + + auto make_provider = [&]() -> std::optional { + Provider p; + targetToUiSlots.clear(); + + if (wireless) { + brokkr::platform::TcpConnection* connp = nullptr; + { + std::lock_guard lk(wireless_mtx_); + if (wireless_conn_) connp = &*wireless_conn_; + } + if (!connp || !connp->connected()) { + fail_ui("No wireless device connected."); + return std::nullopt; + } + + p.owned.push_back(brokkr::odin::Target{.id = wireless_sysname_.toStdString(), .link = connp}); + p.ptrs.push_back(&p.owned.back()); + int uiIdx = findUiIndexBySysname(wireless_sysname_); + if (uiIdx < 0 && !devSquares_.isEmpty()) uiIdx = 0; + targetToUiSlots.push_back(uiIdx); + return p; + } + + std::vector targets; + if (!tgt.isEmpty()) { + auto one = select_odin_target(tgt); + if (!one) { + fail_ui("Target sysname not found or not supported."); + return std::nullopt; + } + targets.push_back(*one); + targetToUiSlots.push_back(findUiIndexBySysname(QString::fromStdString(one->sysname))); + } else { + for (int i = 0; i < uiDevicesSnapshot.size(); ++i) { + const QString sys = uiDevicesSnapshot[i].trimmed(); + if (auto one = select_odin_target(sys)) { + targets.push_back(*one); + targetToUiSlots.push_back(i); + } + } + if (targets.empty()) { + targets = enumerate_odin_targets(); + targetToUiSlots.clear(); + for (const auto& td : targets) { + targetToUiSlots.push_back(findUiIndexBySysname(QString::fromStdString(td.sysname))); + } + } + } + + if (targets.empty()) { + fail_ui("No supported devices found."); + return std::nullopt; + } + + p.usb.reserve(targets.size()); + p.owned.reserve(targets.size()); + p.ptrs.reserve(targets.size()); + + for (const auto& td : targets) { + auto ut = std::make_unique(td.devnode()); + + auto st = ut->dev.open_and_init(); + if (!st) { + const QString err = QString::fromStdString(st.error()); +#ifdef Q_OS_LINUX + if (err.contains("opened read-only", Qt::CaseInsensitive)) { + QMetaObject::invokeMethod( + this, + [this]() { + QMessageBox::warning(this, "Brokkr Flash", + "The USB device was opened read-only. Brokkr cannot flash without " + "write access to the device node.\n\n" + "Please close Brokkr and reopen it as root (e.g. via sudo), " + "or configure udev rules to grant your user write access to the " + "Samsung USB device."); + }, + Qt::QueuedConnection); + } +#endif + fail_ui(err); + return std::nullopt; + } + + auto cst = ut->conn.open(); + if (!cst) { + fail_ui(QString::fromStdString(cst.error())); + return std::nullopt; + } + + ut->conn.set_timeout_ms(cfg.preflash_timeout_ms); + + p.owned.push_back(brokkr::odin::Target{.id = ut->devnode, .link = &ut->conn}); + p.ptrs.push_back(&p.owned.back()); + p.usb.push_back(std::move(ut)); + } + + return p; + }; + + auto run_engine = [&](std::vector srcs) { + auto prov = make_provider(); + if (!prov) return; + + auto st = brokkr::odin::flash(prov->ptrs, srcs, pit_to_upload, cfg, ui); + if (!st) { + if (sawPerDeviceFail.load(std::memory_order_relaxed)) { + done_ui(); + } else { + const QString err = QString::fromStdString(st.error()); + const bool mismatch = + err.contains("mismatch across devices", Qt::CaseInsensitive) || + err.contains("differs across devices", Qt::CaseInsensitive); + fail_ui(mismatch ? QString("The connected devices do not match!") : err, mismatch); + } + requestUsbRefresh_(); + return; + } + done_ui(); + requestUsbRefresh_(); + }; + + if (inputs.empty()) { + run_engine({}); + return; + } + + auto jobsr = brokkr::app::md5_jobs(inputs); + if (!jobsr) { + fail_ui(QString::fromStdString(jobsr.error())); + return; + } + + std::uint64_t totalBytes = 0; + for (const auto& p : inputs) { + std::error_code ec; + const auto sz = std::filesystem::file_size(p, ec); + if (!ec) totalBytes += static_cast(sz); + } + + const QString verifyName = QString::fromStdString(std::string(brokkr::app::md5_verify_name(*jobsr))); + + spdlog::info("{} check ({}), Please wait.", verifyName.toStdString(), human_bytes(totalBytes).toStdString()); + QMetaObject::invokeMethod( + this, + [this, verifyName]() { + setSquaresActiveColor_(false); + setSquaresText_(verifyName); + setSquaresProgress_(0.0, false); + }, + Qt::QueuedConnection); + + auto vst = brokkr::app::md5_verify(*jobsr, ui); + if (!vst) { + fail_ui(QString::fromStdString(vst.error())); + return; + } + + auto specs = brokkr::odin::expand_inputs_tar_or_raw(inputs); + if (!specs) { + fail_ui(QString::fromStdString(specs.error())); + return; + } + + if (!pit_to_upload) { + auto pit = pit_from_specs(*specs); + if (pit) pit_to_upload = pit; + } + + std::vector srcs; + for (auto& s : *specs) + if (!is_pit_name(s.basename)) srcs.push_back(std::move(s)); + if (srcs.empty() && !pit_to_upload) { + fail_ui("No valid flashable files."); + return; + } + + progressVisualBucket_.store(-1, std::memory_order_relaxed); + QMetaObject::invokeMethod(this, [this]() { setSquaresProgress_(0.0, false); }, Qt::QueuedConnection); + + run_engine(std::move(srcs)); + }); +} diff --git a/src-gui/brokkr_wrapper.hpp b/src-gui/brokkr_wrapper.hpp index f30bd4c..a050e72 100644 --- a/src-gui/brokkr_wrapper.hpp +++ b/src-gui/brokkr_wrapper.hpp @@ -1,189 +1,219 @@ -#pragma once - -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include - -#include "platform/platform_all.hpp" - -class QGridLayout; -class QGroupBox; -class QTextEdit; -class QCloseEvent; -class QTabWidget; - -#if defined(BROKKR_PLATFORM_LINUX) -class QSocketNotifier; -#endif -#if defined(Q_OS_MACOS) -#include -#include -#include -#endif - -class DeviceSquare; - -class BrokkrWrapper : public QWidget { - public: - explicit BrokkrWrapper(QWidget* parent = nullptr); - ~BrokkrWrapper() override; - - int logDeviceCountForLog() const noexcept { return logDevCount_.load(std::memory_order_relaxed); } - - public slots: - void appendLogLineFromEngine(const QString& html); - - protected: - void closeEvent(QCloseEvent* e) override; - -#if defined(Q_OS_WIN) - bool nativeEvent(const QByteArray& eventType, void* message, qintptr* result) override; -#endif - - void changeEvent(QEvent* e) override; - - private slots: - void onRunClicked(); - - private: - void setupOdinFileInput(QGridLayout* layout, int row, const QString& label, QLineEdit*& lineEdit); - - void rebuildDeviceBoxes_(int boxCount, bool singleRow); - void refreshDeviceBoxes_(); - void updateActionButtons_(); - void applyWindowHeightToContents_(); - - bool canRunStart_(QString* whyNot = nullptr) const; - bool canRunPrintPit_(QString* whyNot = nullptr) const; - - void showBlocked_(const QString& title, const QString& msg) const; - - void requestUsbRefresh_() noexcept; - void refreshConnectedDevices_(); - - void startWirelessListener_(); - void stopWirelessListener_(); - - void startWorkStart_(); - void startWorkPrintPit_(); - - void setBusy_(bool busy); - void setControlsEnabled_(bool enabled); - - void appendLogLine_(const QString& html); - void applyHeaderStyle_(); - void updateHeaderLeds_(); - - void setSquaresProgress_(double frac, bool animate); - void setSquaresText_(const QString& s); - void setSquaresActiveColor_(bool enhanced); - void setSquaresFinal_(bool ok); - - private: - static constexpr int kBoxesNormal = 8; - static constexpr int kMassDlMaxBoxes = 24; - static constexpr int kBoxesColsMany = 8; - - QStringList connectedDevices_; - QStringList lastFlashDevices_; - bool overflowDevices_ = false; - - int baseWindowHeight_ = 600; - - QGroupBox* idComGroup_ = nullptr; - QGridLayout* idComLayout_ = nullptr; - QWidget* headerWidget_ = nullptr; - QWidget* ledContainer_ = nullptr; - - QList devSquares_; - QList comBoxes; - - std::vector slotFailed_; - std::vector slotActive_; - - QTabWidget* tabWidget_ = nullptr; - QTextEdit* consoleOutput = nullptr; - - QLineEdit* editTarget = nullptr; - QCheckBox* chkWireless = nullptr; - - QPushButton* btnManyDevices_ = nullptr; - QSlider* sldDeviceBoxes = nullptr; - QLabel* lblDeviceBoxes = nullptr; - QWidget* manyRowWidget_ = nullptr; - - QCheckBox* chkAdvanced_ = nullptr; - - QComboBox* cmbRebootAction = nullptr; - - QCheckBox* chkUsePit = nullptr; - QLineEdit* editPit = nullptr; - QPushButton* btnPitBrowse = nullptr; - - QLineEdit* editBL = nullptr; - QLineEdit* editAP = nullptr; - QLineEdit* editCP = nullptr; - QLineEdit* editCSC = nullptr; - QLineEdit* editUserData = nullptr; - - QPushButton* btnRun = nullptr; - QPushButton* btnPrintPit = nullptr; - QPushButton* btnReset_ = nullptr; - - QList fileChecks_; - QList fileButtons_; - - QString lastDir; - - QTimer* deviceTimer = nullptr; - std::atomic_bool usbDirty_{true}; - bool busy_ = false; - -#if defined(BROKKR_PLATFORM_LINUX) - int uevent_fd_ = -1; - QSocketNotifier* uevent_notifier_ = nullptr; -#endif -#if defined(Q_OS_MACOS) - IONotificationPortRef mac_notify_port_ = nullptr; - io_iterator_t mac_added_iter_ = 0; - io_iterator_t mac_removed_iter_ = 0; - static void macOsUsbDeviceChanged(void* refCon, io_iterator_t iterator); -#endif - - std::jthread worker_; - - std::jthread wireless_thread_; - mutable std::mutex wireless_mtx_; - std::optional wireless_listener_; - std::optional wireless_conn_; - QString wireless_sysname_; - - std::atomic_int logDevCount_{0}; - - QStringList physicalUsbPrev_; - bool physicalWirelessPrev_ = false; - QString physicalWirelessIdPrev_; - - std::vector plan_names_; - std::vector plan_from_names_; - bool enhanced_speed_ = false; -}; +#pragma once + +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include + +#include "platform/platform_all.hpp" + +class QGridLayout; +class QGroupBox; +class QTextEdit; +class QCloseEvent; +class QTabWidget; +class QMimeData; +class QDragEnterEvent; +class QDragMoveEvent; +class QDropEvent; +class QChildEvent; + +#if defined(BROKKR_PLATFORM_LINUX) +class QSocketNotifier; +#endif +#if defined(Q_OS_MACOS) +#include +#include +#include +#endif + +class DeviceSquare; + +class BrokkrWrapper : public QWidget { + public: + explicit BrokkrWrapper(QWidget* parent = nullptr); + ~BrokkrWrapper() override; + + int logDeviceCountForLog() const noexcept { return logDevCount_.load(std::memory_order_relaxed); } + + public slots: + void appendLogLineFromEngine(const QString& html); + + protected: + void closeEvent(QCloseEvent* e) override; + void dragEnterEvent(QDragEnterEvent* e) override; + void dragMoveEvent(QDragMoveEvent* e) override; + void dropEvent(QDropEvent* e) override; + bool eventFilter(QObject* watched, QEvent* event) override; + void childEvent(QChildEvent* e) override; + +#if defined(Q_OS_WIN) + bool nativeEvent(const QByteArray& eventType, void* message, qintptr* result) override; +#endif + + void changeEvent(QEvent* e) override; + + private slots: + void onRunClicked(); + + private: + static constexpr int kDropSlotAny = -1; + static constexpr int kDropSlotUnknown = -2; + static constexpr int kDropSlotBL = 0; + static constexpr int kDropSlotAP = 1; + static constexpr int kDropSlotCP = 2; + static constexpr int kDropSlotCSC = 3; + static constexpr int kDropSlotUSERDATA = 4; + static constexpr int kDropSlotPIT = 5; + + static QString dropSlotName_(int slotId); + static int inferDropSlotFromFileName_(const QString& fileName); + static bool isDropFileAllowedForSlot_(int slotId, const QString& fileName); + + QStringList extractLocalDropFiles_(const QMimeData* mime) const; + void bindDropTarget_(QWidget* widget, int slotId); + void handleDroppedFiles_(const QStringList& localFiles, int forcedSlotId); + bool assignDroppedFileToSlot_(int slotId, const QString& localPath, QString* reason, bool* replaced); + + void setupOdinFileInput(QGridLayout* layout, int row, const QString& label, QLineEdit*& lineEdit); + + void rebuildDeviceBoxes_(int boxCount, bool singleRow); + void refreshDeviceBoxes_(); + void updateActionButtons_(); + void updateRebootDownloadButton_(); + void applyWindowHeightToContents_(); + + bool canRunStart_(QString* whyNot = nullptr) const; + bool confirmOdinModeDevicesForStart_(); + + void showBlocked_(const QString& title, const QString& msg) const; + + void requestUsbRefresh_() noexcept; + void refreshConnectedDevices_(); + + void startWirelessListener_(); + void stopWirelessListener_(); + + void startWorkStart_(); + void tryRebootIntoDownloadMode_(); + + void setBusy_(bool busy); + void setControlsEnabled_(bool enabled); + + void appendLogLine_(const QString& html); + void applyHeaderStyle_(); + void updateHeaderLeds_(); + + void setSquaresProgress_(double frac, bool animate); + void setSquaresText_(const QString& s); + void setSquaresActiveColor_(bool enhanced); + void setSquaresFinal_(bool ok); + + private: + static constexpr int kBoxesNormal = 8; + static constexpr int kMassDlMaxBoxes = 24; + static constexpr int kBoxesColsMany = 8; + + QStringList connectedDevices_; + bool overflowDevices_ = false; + + int baseWindowHeight_ = 600; + + QGroupBox* idComGroup_ = nullptr; + QGridLayout* idComLayout_ = nullptr; + QWidget* headerWidget_ = nullptr; + QWidget* ledContainer_ = nullptr; + + QList devSquares_; + QList comBoxes; + + std::vector slotFailed_; + std::vector slotActive_; + + QTabWidget* tabWidget_ = nullptr; + QTextEdit* consoleOutput = nullptr; + + QLineEdit* editTarget = nullptr; + QCheckBox* chkWireless = nullptr; + + QPushButton* btnManyDevices_ = nullptr; + QPushButton* btnRebootDownloadMode_ = nullptr; + + QComboBox* cmbRebootAction = nullptr; + + QCheckBox* chkUsePit = nullptr; + QLineEdit* editPit = nullptr; + QPushButton* btnPitBrowse = nullptr; + + QLineEdit* editBL = nullptr; + QLineEdit* editAP = nullptr; + QLineEdit* editCP = nullptr; + QLineEdit* editCSC = nullptr; + QLineEdit* editUserData = nullptr; + + QPushButton* btnRun = nullptr; + QPushButton* btnReset_ = nullptr; + + QList fileChecks_; + QList fileButtons_; + QList fileLineEdits_; + + QString lastDir; + + QTimer* deviceTimer = nullptr; + std::atomic_bool usbDirty_{true}; + bool busy_ = false; + bool windowsElevatedHintShown_ = false; + +#if defined(BROKKR_PLATFORM_LINUX) + int uevent_fd_ = -1; + QSocketNotifier* uevent_notifier_ = nullptr; +#endif +#if defined(Q_OS_MACOS) + IONotificationPortRef mac_notify_port_ = nullptr; + io_iterator_t mac_added_iter_ = 0; + io_iterator_t mac_removed_iter_ = 0; + static void macOsUsbDeviceChanged(void* refCon, io_iterator_t iterator); +#endif + + std::jthread worker_; + + std::jthread wireless_thread_; + mutable std::mutex wireless_mtx_; + std::optional wireless_listener_; + std::optional wireless_conn_; + QString wireless_sysname_; + // While true the watcher thread must NOT touch wireless_conn_ (no + // connected() probes, no reset()), because the flash engine is using the + // underlying socket. Set during a flash run. + std::atomic_bool wireless_watcher_paused_{false}; + + std::atomic_int logDevCount_{0}; + std::atomic_int progressVisualSteps_{1}; + std::atomic_int progressVisualBucket_{-1}; + + QStringList physicalUsbPrev_; + bool physicalWirelessPrev_ = false; + QString physicalWirelessIdPrev_; + + std::vector plan_names_; + std::vector plan_from_names_; + bool enhanced_speed_ = false; +}; diff --git a/src-gui/main_gui.cpp b/src-gui/main_gui.cpp index d8baafd..3625c3d 100644 --- a/src-gui/main_gui.cpp +++ b/src-gui/main_gui.cpp @@ -18,15 +18,33 @@ #include #include +#include "app/cli_mode.hpp" #include "platform/platform_all.hpp" #include "brokkr_wrapper.hpp" +#if defined(_WIN32) + #include +#endif + int main(int argc, char* argv[]) { + const bool cli_mode = brokkr::app::should_run_cli(argc, argv); + + if (cli_mode) { + return brokkr::app::run_cli(argc, argv); + } + +#if defined(_WIN32) + if (HWND cw = ::GetConsoleWindow(); cw) { + ::ShowWindow(cw, SW_HIDE); + (void)::FreeConsole(); + } +#endif + QApplication app(argc, argv); auto lock = brokkr::platform::SingleInstanceLock::try_acquire("brokkr-engine"); if (!lock) { - QMessageBox::critical(nullptr, "Brokkr Flasher", "Another instance is already running."); + QMessageBox::critical(nullptr, "Brokkr Flash", "Another instance is already running."); return 2; } diff --git a/src/app/cli_mode.cpp b/src/app/cli_mode.cpp new file mode 100644 index 0000000..856186a --- /dev/null +++ b/src/app/cli_mode.cpp @@ -0,0 +1,544 @@ +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#include "app/cli_mode.hpp" + +#include "app/md5_verify.hpp" +#include "core/status.hpp" +#include "core/str.hpp" +#include "io/source.hpp" +#include "platform/platform_all.hpp" +#include "protocol/odin/flash.hpp" +#include "protocol/odin/group_flasher.hpp" + +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +namespace brokkr::app { + +namespace { + +constexpr std::uint16_t kSamsungVid = 0x04E8; +constexpr std::uint16_t kOdinPids[] = {0x6601, 0x685D, 0x68C3}; +constexpr std::uint16_t kWirelessPort = 13579; + +struct CliArgs { + bool help = false; + bool list = false; + bool wireless = false; + bool no_reboot = false; + + std::optional target; + std::optional pit; + + std::optional bl; + std::optional ap; + std::optional cp; + std::optional csc; + std::optional userdata; +}; + +struct Provider { + std::vector> usb; + std::vector owned; + std::vector ptrs; + std::unique_ptr wireless_conn; +}; + +bool is_odin_product(std::uint16_t pid) { + return std::find(std::begin(kOdinPids), std::end(kOdinPids), pid) != std::end(kOdinPids); +} + +bool is_cli_trigger(std::string_view arg) { + static const std::unordered_set kTriggers = { + "-h", "--help", "--list", "--wireless", "--no-reboot", "--use-pit", "--target", + "-b", "-a", "-c", "-s", "-u", + }; + return kTriggers.contains(arg); +} + +void configure_cli_logger() { + auto sink = std::make_shared(); + sink->set_pattern("%v"); + auto logger = std::make_shared("cli", spdlog::sinks_init_list{sink}); +#ifndef NDEBUG + logger->set_level(spdlog::level::debug); +#else + logger->set_level(spdlog::level::info); +#endif + spdlog::set_default_logger(std::move(logger)); +} + +void print_usage() { + std::cout + << "Usage:\n" + << " brokkr [CLI options]\n\n" + << "CLI options (any of these switches CLI mode):\n" + << " -h, --help Show this help\n" + << " --list List Samsung devices usable by --target\n" + << " -b BL file\n" + << " -a AP file\n" + << " -c CP file\n" + << " -s CSC file\n" + << " -u USERDATA file\n" + << " --use-pit Optional PIT file\n" + << " --no-reboot Do not reboot at end\n" + << " --wireless Flash via wireless listener\n" + << " --target Same target semantics as GUI\n\n" + << "Notes:\n" + << " - At least one file is required from: -b -a -c -s -u --use-pit\n" + << " - --wireless cannot be used with --target\n" + << " - If no valid CLI option is present, GUI mode is launched\n"; +} + +brokkr::core::Result parse_cli_args(int argc, char* argv[]) { + CliArgs out; + + auto require_value = [&](int& i, const char* flag) -> brokkr::core::Result { + if (i + 1 >= argc) return brokkr::core::fail(std::string("Missing value for ") + flag); + ++i; + return std::string(argv[i]); + }; + + for (int i = 1; i < argc; ++i) { + const std::string arg = argv[i]; + + if (arg == "-h" || arg == "--help") { + out.help = true; + continue; + } + if (arg == "--list") { + out.list = true; + continue; + } + if (arg == "--wireless") { + out.wireless = true; + continue; + } + if (arg == "--no-reboot") { + out.no_reboot = true; + continue; + } + if (arg == "--target") { + BRK_TRYV(v, require_value(i, "--target")); + out.target = std::move(v); + continue; + } + if (arg == "--use-pit") { + BRK_TRYV(v, require_value(i, "--use-pit")); + out.pit = std::move(v); + continue; + } + if (arg == "-b") { + BRK_TRYV(v, require_value(i, "-b")); + out.bl = std::move(v); + continue; + } + if (arg == "-a") { + BRK_TRYV(v, require_value(i, "-a")); + out.ap = std::move(v); + continue; + } + if (arg == "-c") { + BRK_TRYV(v, require_value(i, "-c")); + out.cp = std::move(v); + continue; + } + if (arg == "-s") { + BRK_TRYV(v, require_value(i, "-s")); + out.csc = std::move(v); + continue; + } + if (arg == "-u") { + BRK_TRYV(v, require_value(i, "-u")); + out.userdata = std::move(v); + continue; + } + + return brokkr::core::fail("Unknown argument: " + arg); + } + + return out; +} + +std::vector enumerate_samsung_targets() { + brokkr::platform::EnumerateFilter f{.vendor = kSamsungVid}; + return brokkr::platform::enumerate_usb_devices_sysfs(f); +} + +std::optional select_samsung_target(std::string_view sysname) { + if (sysname.empty()) return std::nullopt; + auto info = brokkr::platform::find_by_sysname(sysname); + if (!info || info->vendor != kSamsungVid) return std::nullopt; + return info; +} + +std::optional select_odin_target(std::string_view sysname) { + auto info = select_samsung_target(sysname); + if (!info || !is_odin_product(info->product)) return std::nullopt; + return info; +} + +std::vector collect_inputs_in_gui_order(const CliArgs& args) { + std::vector out; + if (args.bl) out.emplace_back(*args.bl); + if (args.ap) out.emplace_back(*args.ap); + if (args.cp) out.emplace_back(*args.cp); + if (args.csc) out.emplace_back(*args.csc); + if (args.userdata) out.emplace_back(*args.userdata); + return out; +} + +bool has_any_file_selected(const CliArgs& args) { + return args.bl.has_value() || args.ap.has_value() || args.cp.has_value() || args.csc.has_value() || + args.userdata.has_value() || args.pit.has_value(); +} + +std::shared_ptr> load_pit_if_needed(const CliArgs& args) { + if (!args.pit) return {}; + + const std::filesystem::path p = *args.pit; + std::error_code ec; + const auto sz = std::filesystem::file_size(p, ec); + if (ec) { + spdlog::error("Cannot stat PIT file."); + return {}; + } + + std::vector buf(static_cast(sz)); + std::ifstream in(p, std::ios::binary); + if (!in.is_open()) { + spdlog::error("Cannot open PIT file."); + return {}; + } + + if (!buf.empty()) { + in.read(reinterpret_cast(buf.data()), static_cast(buf.size())); + if (!in.good()) { + spdlog::error("Failed to read PIT file."); + return {}; + } + } + + return std::make_shared>(std::move(buf)); +} + +bool is_pit_name(std::string_view base) noexcept { + return brokkr::core::ends_with_ci(base, ".pit"); +} + +std::shared_ptr> pit_from_specs( + const std::vector& specs) { + const brokkr::odin::ImageSpec* pit = nullptr; + for (const auto& s : specs) + if (is_pit_name(s.basename)) pit = &s; + if (!pit) return {}; + + auto sr = pit->open(); + if (!sr) { + spdlog::error("PIT open failed: {}", sr.error()); + return {}; + } + auto& src = **sr; + + constexpr std::uint64_t kMaxPit = 256ull * 1024ull * 1024ull; + const auto sz64 = src.size(); + if (sz64 > kMaxPit) { + spdlog::error("Embedded PIT too large: {}", src.display_name()); + return {}; + } + + std::vector out(static_cast(sz64)); + for (std::size_t off = 0; off < out.size();) { + const std::size_t got = src.read({out.data() + off, out.size() - off}); + if (!got) { + auto st = src.status(); + if (!st) + spdlog::error("PIT read failed: {}", st.error()); + else + spdlog::error("Short read on embedded PIT: {}", src.display_name()); + return {}; + } + off += got; + } + + return std::make_shared>(std::move(out)); +} + +std::string map_global_error_to_cli_message(const std::string& err) { + const auto has_ic = [&](std::string_view needle) { + auto it = std::search(err.begin(), err.end(), needle.begin(), needle.end(), [](char a, char b) { + return static_cast(std::tolower(static_cast(a))) == + static_cast(std::tolower(static_cast(b))); + }); + return it != err.end(); + }; + + if (has_ic("mismatch across devices") || has_ic("differs across devices")) { + return "The connected devices do not match!"; + } + + return err; +} + +int list_devices_cli() { + const auto devs = enumerate_samsung_targets(); + if (devs.empty()) { + std::cout << "No Samsung devices found.\n"; + return 0; + } + + for (const auto& d : devs) { + std::cout << d.sysname << "\t" << (is_odin_product(d.product) ? "Odin Mode" : "Not in Odin Mode") << "\n"; + } + + return 0; +} + +brokkr::core::Result make_provider(const CliArgs& args, const brokkr::odin::Cfg& cfg) { + Provider p; + + if (args.wireless) { + brokkr::platform::TcpListener listener; + BRK_TRY(listener.bind_and_listen("0.0.0.0", kWirelessPort)); + + spdlog::info("Waiting for connection on port {}", kWirelessPort); + + for (;;) { + auto ar = listener.accept_one(); + if (ar) { + p.wireless_conn = std::make_unique(std::move(*ar)); + break; + } + if (ar.error() != "accept: timeout") return brokkr::core::fail(std::move(ar.error())); + std::this_thread::sleep_for(std::chrono::milliseconds(200)); + } + + p.owned.push_back(brokkr::odin::Target{.id = "wireless", .link = p.wireless_conn.get()}); + p.ptrs.push_back(&p.owned.back()); + return p; + } + + std::vector targets; + if (args.target && !args.target->empty()) { + auto info = select_samsung_target(*args.target); + if (!info) return brokkr::core::fail("Target sysname not found."); + if (!is_odin_product(info->product)) { + return brokkr::core::fail("The device is not in Odin Mode, reboot to download mode first."); + } + targets.push_back(*info); + } else { + const auto all_samsung = enumerate_samsung_targets(); + if (all_samsung.empty()) return brokkr::core::fail("No connected devices detected."); + + for (const auto& d : all_samsung) { + if (is_odin_product(d.product)) { + targets.push_back(d); + } else { + spdlog::info("{} is not in Odin Mode hence ignored", d.sysname); + } + } + + if (targets.empty()) return brokkr::core::fail("None of the devices are in Odin Mode"); + } + + p.usb.reserve(targets.size()); + p.owned.reserve(targets.size()); + p.ptrs.reserve(targets.size()); + + for (const auto& td : targets) { + auto ut = std::make_unique(td.devnode()); + + auto st = ut->dev.open_and_init(); + if (!st) return brokkr::core::fail(std::move(st.error())); + + auto cst = ut->conn.open(); + if (!cst) return brokkr::core::fail(std::move(cst.error())); + + ut->conn.set_timeout_ms(cfg.preflash_timeout_ms); + + p.owned.push_back(brokkr::odin::Target{.id = ut->devnode, .link = &ut->conn}); + p.ptrs.push_back(&p.owned.back()); + p.usb.push_back(std::move(ut)); + } + + return p; +} + +int run_flash_cli(const CliArgs& args) { + if (!has_any_file_selected(args)) { + spdlog::error("No files selected."); + return 2; + } + + if (args.wireless && args.target && !args.target->empty()) { + spdlog::error("Wireless cannot be used together with Target Sysname."); + return 2; + } + + auto pit_to_upload = load_pit_if_needed(args); + if (args.pit && !pit_to_upload) return 1; + + auto inputs = collect_inputs_in_gui_order(args); + + brokkr::odin::Cfg cfg; + cfg.reboot_after = !args.no_reboot; + + std::atomic_bool saw_per_device_fail{false}; + std::optional sig_guard; + bool flash_signal_shield_attempted = false; + brokkr::odin::Ui ui; + ui.on_stage = [&](const std::string& s) { + if (flash_signal_shield_attempted) return; + if (s.rfind("Flashing", 0) != 0) return; + + flash_signal_shield_attempted = true; + sig_guard = brokkr::core::SignalShield::enable([](const char* sig_desc, int count) { + spdlog::warn("{} received during active flash operation - ignoring ({})", sig_desc, count); + }); + if (!sig_guard) { + spdlog::warn("Failed to enable signal shielding; interrupts may terminate this flash run."); + } + }; + ui.on_progress = [](std::uint64_t, std::uint64_t, std::uint64_t, std::uint64_t) { + // CLI intentionally suppresses percentage/progress bars. + }; + ui.on_error = [&](const std::string& s) { + if (s.rfind("DEVFAIL idx=", 0) == 0) saw_per_device_fail.store(true, std::memory_order_relaxed); + spdlog::error("{}", s); + }; + + auto provider_r = make_provider(args, cfg); + if (!provider_r) { + spdlog::error("{}", map_global_error_to_cli_message(provider_r.error())); + return 1; + } + auto provider = std::move(*provider_r); + + std::vector specs; + if (!inputs.empty()) { + auto jobsr = brokkr::app::md5_jobs(inputs); + if (!jobsr) { + spdlog::error("{}", jobsr.error()); + return 1; + } + + for (const auto& job : *jobsr) { + std::string name = job.path.filename().string(); + if (name.empty()) name = job.path.string(); + + if (name.size() >= 11) + name = name.substr(0, 10) + "..."; + + spdlog::info("Checking MD5/XXH3 on {}", name); + } + + auto vst = brokkr::app::md5_verify(*jobsr, ui); + if (!vst) { + spdlog::error("{}", vst.error()); + return 1; + } + + auto specsr = brokkr::odin::expand_inputs_tar_or_raw(inputs); + if (!specsr) { + spdlog::error("{}", specsr.error()); + return 1; + } + specs = std::move(*specsr); + + if (!pit_to_upload) { + if (auto pit = pit_from_specs(specs)) pit_to_upload = std::move(pit); + } + + std::vector filtered; + filtered.reserve(specs.size()); + for (auto& s : specs) + if (!is_pit_name(s.basename)) filtered.push_back(std::move(s)); + specs = std::move(filtered); + + if (specs.empty() && !pit_to_upload) { + spdlog::error("No valid flashable files."); + return 1; + } + } + + auto fst = brokkr::odin::flash(provider.ptrs, specs, pit_to_upload, cfg, ui); + if (!fst) { + if (!saw_per_device_fail.load(std::memory_order_relaxed)) { + spdlog::error("{}", map_global_error_to_cli_message(fst.error())); + } + return 1; + } + + return 0; +} + +} // namespace + +bool should_run_cli(int argc, char* argv[]) noexcept { + for (int i = 1; i < argc; ++i) { + if (is_cli_trigger(argv[i])) return true; + } + return false; +} + +int run_cli(int argc, char* argv[]) { + configure_cli_logger(); + + auto lock = brokkr::platform::SingleInstanceLock::try_acquire("brokkr-engine"); + if (!lock) { + spdlog::error("Another instance is already running."); + return 2; + } + + auto argsr = parse_cli_args(argc, argv); + if (!argsr) { + spdlog::error("{}", argsr.error()); + print_usage(); + return 2; + } + const CliArgs args = std::move(*argsr); + + if (args.help) { + print_usage(); + return 0; + } + + if (args.list) return list_devices_cli(); + + return run_flash_cli(args); +} + +} // namespace brokkr::app diff --git a/src/app/cli_mode.hpp b/src/app/cli_mode.hpp new file mode 100644 index 0000000..0f8e37d --- /dev/null +++ b/src/app/cli_mode.hpp @@ -0,0 +1,25 @@ +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#pragma once + +namespace brokkr::app { + +bool should_run_cli(int argc, char* argv[]) noexcept; +int run_cli(int argc, char* argv[]); + +} // namespace brokkr::app diff --git a/src/app/md5_verify.cpp b/src/app/md5_verify.cpp index 2b875ba..b332483 100644 --- a/src/app/md5_verify.cpp +++ b/src/app/md5_verify.cpp @@ -17,22 +17,42 @@ #include "app/md5_verify.hpp" +#include "app/md5_xxh3_cache.hpp" + #include "core/prefetcher.hpp" +#include "core/str.hpp" #include "core/thread_pool.hpp" #include "io/tar.hpp" +#include "platform/platform_all.hpp" #include "third_party/md5/md5.h" +#include "third_party/xxhash/xxhash_vendor.h" +#include #include #include #include #include +#include +#include #include #include #include #include +#include #include +#if defined(_WIN32) + #include +#else + #include + #include + #include + #if defined(POSIX_FADV_SEQUENTIAL) + #include + #endif +#endif + #include #include @@ -42,6 +62,127 @@ namespace { constexpr std::size_t kTrailerMaxBytes = 16 * 1024; constexpr std::size_t kMd5HexChars = 32; +constexpr std::size_t kMd5Xxh3CacheMaxEntries = 100; +constexpr std::size_t kHashBufBytes = 32 * 1024 * 1024; + +struct CombinedDigest { + std::array md5{}; + std::uint64_t xxh3_64 = 0; +}; + +struct SessionVerifyKey { + std::string identity_path; + std::uint64_t identity_size = 0; + std::int64_t identity_write_time = 0; + std::uint64_t bytes_to_hash = 0; + std::array expected{}; + + bool operator==(const SessionVerifyKey& other) const noexcept { + return identity_path == other.identity_path && identity_size == other.identity_size && + identity_write_time == other.identity_write_time && bytes_to_hash == other.bytes_to_hash && + expected == other.expected; + } +}; + +struct SessionVerifyKeyHash { + std::size_t operator()(const SessionVerifyKey& key) const noexcept { + auto hash_combine = [](std::size_t seed, std::size_t value) noexcept { + return seed ^ (value + 0x9e3779b97f4a7c15ull + (seed << 6) + (seed >> 2)); + }; + + std::size_t seed = std::hash{}(key.identity_path); + seed = hash_combine(seed, std::hash{}(key.identity_size)); + seed = hash_combine(seed, std::hash{}(key.identity_write_time)); + seed = hash_combine(seed, std::hash{}(key.bytes_to_hash)); + for (unsigned char b : key.expected) seed = hash_combine(seed, std::hash{}(b)); + return seed; + } +}; + +std::unordered_set& session_verify_cache() { + static std::unordered_set cache; + return cache; +} + +std::mutex& session_verify_cache_mutex() { + static std::mutex mtx; + return mtx; +} + +class HashFileReader { + public: + explicit HashFileReader(const std::filesystem::path& path) noexcept : path_(path) {} + + brokkr::core::Status open() noexcept { +#if defined(_WIN32) + handle_ = CreateFileW(path_.c_str(), GENERIC_READ, + FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE, nullptr, OPEN_EXISTING, + FILE_ATTRIBUTE_NORMAL | FILE_FLAG_SEQUENTIAL_SCAN, nullptr); + if (handle_ == INVALID_HANDLE_VALUE) { + return brokkr::core::failf("Cannot open for hashing: {}", path_.string()); + } + return {}; +#else + fd_ = ::open(path_.c_str(), O_RDONLY +#if defined(O_CLOEXEC) + | O_CLOEXEC +#endif + ); + if (fd_ < 0) return brokkr::core::failf("Cannot open for hashing: {}", path_.string()); +#if defined(POSIX_FADV_SEQUENTIAL) + (void)::posix_fadvise(fd_, 0, 0, POSIX_FADV_SEQUENTIAL); +#endif +#if defined(POSIX_FADV_WILLNEED) + (void)::posix_fadvise(fd_, 0, 0, POSIX_FADV_WILLNEED); +#endif + return {}; +#endif + } + + brokkr::core::Result read_some(unsigned char* data, std::size_t want) noexcept { +#if defined(_WIN32) + std::size_t total = 0; + while (total < want) { + const DWORD chunk = static_cast(std::min(want - total, static_cast(1u << 30))); + DWORD got = 0; + if (!ReadFile(handle_, data + total, chunk, &got, nullptr)) { + return brokkr::core::failf("Read failed while hashing: {}", path_.string()); + } + if (got == 0) break; + total += static_cast(got); + } + return total; +#else + std::size_t total = 0; + while (total < want) { + const ssize_t got = ::read(fd_, data + total, want - total); + if (got < 0) { + if (errno == EINTR) continue; + return brokkr::core::failf("Read failed while hashing: {}", path_.string()); + } + if (got == 0) break; + total += static_cast(got); + } + return total; +#endif + } + + ~HashFileReader() { +#if defined(_WIN32) + if (handle_ != INVALID_HANDLE_VALUE) CloseHandle(handle_); +#else + if (fd_ >= 0) ::close(fd_); +#endif + } + + private: + std::filesystem::path path_; +#if defined(_WIN32) + HANDLE handle_ = INVALID_HANDLE_VALUE; +#else + int fd_ = -1; +#endif +}; static bool is_hex(unsigned char c) noexcept { return (c >= '0' && c <= '9') || (c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F'); @@ -55,8 +196,8 @@ static int hex_nibble(unsigned char c) noexcept { } static bool parse_md5_hex(std::string_view hex32, std::array& out) noexcept { - if (hex32.size() != 32) return false; - for (std::size_t i = 0; i < 16; ++i) { + if (hex32.size() != kMd5HexChars) return false; + for (std::size_t i = 0; i < out.size(); ++i) { const int hi = hex_nibble(static_cast(hex32[2 * i + 0])); const int lo = hex_nibble(static_cast(hex32[2 * i + 1])); if (hi < 0 || lo < 0) return false; @@ -65,20 +206,185 @@ static bool parse_md5_hex(std::string_view hex32, std::array& return true; } -static std::string md5_hex32(const std::array& d) { - static constexpr char hex[] = "0123456789abcdef"; - std::string out(32, '0'); - for (std::size_t i = 0; i < 16; ++i) { - out[2 * i + 0] = hex[(d[i] >> 4) & 0x0F]; - out[2 * i + 1] = hex[d[i] & 0x0F]; +static bool is_md5_wrapped_tar_name(const std::filesystem::path& path) noexcept { + return brokkr::core::ends_with_ci(path.filename().string(), ".md5"); +} + +static std::filesystem::path session_identity_path(const std::filesystem::path& path) noexcept { + std::error_code ec; + auto canonical = std::filesystem::weakly_canonical(path, ec); + if (!ec) return canonical; + + auto absolute = std::filesystem::absolute(path, ec); + if (!ec) return absolute.lexically_normal(); + + return path.lexically_normal(); +} + +static std::int64_t session_identity_write_time(std::filesystem::file_time_type t) noexcept { + return static_cast(t.time_since_epoch().count()); +} + +static SessionVerifyKey make_session_verify_key(const Md5Job& job) { + SessionVerifyKey key; + key.identity_path = job.identity_path.generic_string(); + key.identity_size = job.identity_size; + key.identity_write_time = job.identity_write_time; + key.bytes_to_hash = job.bytes_to_hash; + key.expected = job.expected; + return key; +} + +static bool session_verify_cache_contains(const Md5Job& job) { + const auto key = make_session_verify_key(job); + std::lock_guard lk(session_verify_cache_mutex()); + return session_verify_cache().contains(key); +} + +static void remember_session_verify_cache(const Md5Job& job) { + std::lock_guard lk(session_verify_cache_mutex()); + session_verify_cache().insert(make_session_verify_key(job)); +} + +struct Xxh3Consumer { + using Digest = std::uint64_t; + static constexpr bool kUsesMd5 = false; + static constexpr bool kUsesXxh3 = true; + + brokkr::core::Status init() noexcept { + state_ = XXH3_createState(); + if (!state_) return brokkr::core::fail("Cannot allocate XXH3 state"); + if (XXH3_64bits_reset(state_) != XXH_OK) return brokkr::core::fail("Cannot reset XXH3 state"); + return {}; + } + + brokkr::core::Status update(const unsigned char* data, std::size_t size) noexcept { + if (XXH3_64bits_update(state_, data, size) != XXH_OK) return brokkr::core::fail("XXH3 update failed"); + return {}; + } + + Digest finish() noexcept { return XXH3_64bits_digest(state_); } + + ~Xxh3Consumer() { + if (state_) XXH3_freeState(state_); + } + + XXH3_state_t* state_ = nullptr; +}; + +struct Md5Xxh3Consumer { + using Digest = CombinedDigest; + static constexpr bool kUsesMd5 = true; + static constexpr bool kUsesXxh3 = true; + + brokkr::core::Status init() noexcept { + md5_init(&md5_); + state_ = XXH3_createState(); + if (!state_) return brokkr::core::fail("Cannot allocate XXH3 state"); + if (XXH3_64bits_reset(state_) != XXH_OK) return brokkr::core::fail("Cannot reset XXH3 state"); + return {}; + } + + brokkr::core::Status update(const unsigned char* data, std::size_t size) noexcept { + md5_update(&md5_, data, size); + if (XXH3_64bits_update(state_, data, size) != XXH_OK) return brokkr::core::fail("XXH3 update failed"); + return {}; + } + + Digest finish() noexcept { + Digest out; + md5_final(&md5_, out.md5.data()); + out.xxh3_64 = XXH3_64bits_digest(state_); + return out; + } + + ~Md5Xxh3Consumer() { + if (state_) XXH3_freeState(state_); + } + + MD5_CTX md5_{}; + XXH3_state_t* state_ = nullptr; +}; + +template +static brokkr::core::Result hash_prefetch(const std::filesystem::path& path, + std::uint64_t bytes_to_hash, + std::atomic_uint64_t& done, + std::uint64_t total, + const brokkr::odin::Ui& ui) noexcept { + struct Slot { + std::vector buf; + std::size_t n = 0; + }; + + HashFileReader reader(path); + BRK_TRY(reader.open()); + + std::uint64_t remaining = bytes_to_hash; + + brokkr::core::TwoSlotPrefetcher pf( + [&](Slot& s, std::stop_token st) -> brokkr::core::Result { + if (st.stop_requested() || !remaining) return false; + + const std::size_t want = static_cast(std::min(remaining, kHashBufBytes)); + BRK_TRYV(got, reader.read_some(s.buf.data(), want)); + if (got != want) return brokkr::core::failf("Short read while hashing: {}", path.string()); + + s.n = got; + remaining -= static_cast(got); + return true; + }, + [&](Slot& s) { s.buf.resize(kHashBufBytes); }); + + Consumer consumer; + if constexpr (Consumer::kUsesMd5) { + spdlog::debug("MD5 start: {} bytes from {}", bytes_to_hash, path.string()); + } + if constexpr (Consumer::kUsesXxh3) { + spdlog::debug("XXH3 start: {} bytes from {}", bytes_to_hash, path.string()); + } + BRK_TRY(consumer.init()); + + std::uint64_t processed = 0; + while (processed < bytes_to_hash) { + auto lease = pf.next(); + if (!lease) break; + + auto& s = lease->get(); + if (!s.n) break; + + BRK_TRY(consumer.update(s.buf.data(), s.n)); + processed += static_cast(s.n); + + const auto new_done = done.fetch_add(static_cast(s.n), std::memory_order_relaxed) + + static_cast(s.n); + if (ui.on_progress) ui.on_progress(new_done, total, new_done, total); + } + + auto pst = pf.status(); + if (!pst) return brokkr::core::fail(std::move(pst.error())); + + if (processed != bytes_to_hash) { + return brokkr::core::failf("Hashing terminated early: {} (processed {}, expected {})", path.string(), processed, + bytes_to_hash); + } + + auto digest = consumer.finish(); + if constexpr (Consumer::kUsesMd5) { + spdlog::debug("MD5 finish: {} bytes from {}", bytes_to_hash, path.string()); + } + if constexpr (Consumer::kUsesXxh3) { + spdlog::debug("XXH3 finish: {} bytes from {}", bytes_to_hash, path.string()); } - return out; + return digest; } static brokkr::core::Result> detect_md5_job(const std::filesystem::path& p) noexcept { std::error_code ec; const std::uint64_t file_size = std::filesystem::file_size(p, ec); if (ec) return brokkr::core::failf("Cannot stat file: {}", p.string()); + const auto write_time = std::filesystem::last_write_time(p, ec); + if (ec) return brokkr::core::failf("Cannot stat file write time: {}", p.string()); if (file_size < (kMd5HexChars + 2)) return std::nullopt; const std::uint64_t tail_off = (file_size > kTrailerMaxBytes) ? (file_size - kTrailerMaxBytes) : 0; @@ -127,81 +433,21 @@ static brokkr::core::Result> detect_md5_job(const std::fil Md5Job j; j.path = p; + j.identity_path = session_identity_path(p); + j.identity_size = file_size; + j.identity_write_time = session_identity_write_time(write_time); j.bytes_to_hash = bytes_to_hash; j.expected = expected; return std::optional(std::move(j)); } -static brokkr::core::Result> md5_hash_prefetch(const std::filesystem::path& path, - std::uint64_t bytes_to_hash, - std::atomic_uint64_t& done, - std::uint64_t total, - const brokkr::odin::Ui& ui) noexcept { - constexpr std::size_t kBuf = 8 * 1024 * 1024; - struct Slot { - std::vector buf; - std::size_t n = 0; - }; - - std::FILE* f = std::fopen(path.string().c_str(), "rb"); - if (!f) return brokkr::core::failf("Cannot open for MD5: {}", path.string()); - std::unique_ptr g(f, &std::fclose); - - std::uint64_t remaining = bytes_to_hash; - - brokkr::core::TwoSlotPrefetcher pf( - [&](Slot& s, std::stop_token st) -> brokkr::core::Result { - if (st.stop_requested() || !remaining) return false; - - const std::size_t want = static_cast(std::min(remaining, kBuf)); - const std::size_t got = std::fread(s.buf.data(), 1, want, f); - if (got != want) return brokkr::core::failf("Short read while hashing: {}", path.string()); - - s.n = got; - remaining -= static_cast(got); - return true; - }, - [&](Slot& s) { s.buf.resize(kBuf); }); - - MD5_CTX ctx{}; - md5_init(&ctx); - - std::uint64_t processed = 0; - while (processed < bytes_to_hash) { - auto lease = pf.next(); - if (!lease) break; - - auto& s = lease->get(); - if (!s.n) break; - - md5_update(&ctx, s.buf.data(), s.n); - processed += static_cast(s.n); - - const auto new_done = done.fetch_add(static_cast(s.n), std::memory_order_relaxed) + - static_cast(s.n); - - if (ui.on_progress) ui.on_progress(new_done, total, new_done, total); - } - - auto pst = pf.status(); - if (!pst) return brokkr::core::fail(std::move(pst.error())); - - if (processed != bytes_to_hash) { - return brokkr::core::failf("MD5 hashing terminated early: {} (processed {}, expected {})", path.string(), processed, - bytes_to_hash); - } - - std::array out{}; - md5_final(&ctx, out.data()); - return out; -} - } // namespace brokkr::core::Result> md5_jobs(const std::vector& inputs) noexcept { std::vector jobs; for (const auto& p : inputs) { + if (!is_md5_wrapped_tar_name(p)) continue; if (!brokkr::io::TarArchive::is_tar_file(p.string())) continue; auto r = detect_md5_job(p); @@ -212,21 +458,62 @@ brokkr::core::Result> md5_jobs(const std::vector& jobs) noexcept { + if (jobs.empty()) return "MD5"; + + auto cache_dir = brokkr::platform::app_cache_dir(); + if (!cache_dir) return "MD5"; + + auto loaded = load_md5_xxh3_cache(md5_xxh3_cache_file(*cache_dir)); + if (!loaded) return "MD5"; + + const auto& cache_entries = *loaded; + const bool all_jobs_cached = std::all_of(jobs.begin(), jobs.end(), [&](const Md5Job& job) { + return std::any_of(cache_entries.begin(), cache_entries.end(), [&](const Md5Xxh3CacheEntry& entry) { + return entry.md5 == job.expected && entry.bytes_to_hash == job.bytes_to_hash; + }); + }); + return all_jobs_cached ? "XXH3" : "MD5"; +} + brokkr::core::Status md5_verify(const std::vector& jobs, const brokkr::odin::Ui& ui) noexcept { if (jobs.empty()) return {}; std::uint64_t total = 0; for (const auto& j : jobs) total += j.bytes_to_hash; + std::atomic_uint64_t done{0}; + + std::filesystem::path cache_file; + std::vector cache_entries; + std::mutex cache_mtx; + bool cache_enabled = false; + bool cache_dirty = false; + + auto cache_dir = brokkr::platform::app_cache_dir(); + if (!cache_dir) { + spdlog::warn("MD5/XXH3 cache disabled: {}", cache_dir.error()); + } else { + cache_file = md5_xxh3_cache_file(*cache_dir); + auto loaded = load_md5_xxh3_cache(cache_file); + if (!loaded) { + spdlog::warn("MD5/XXH3 cache load failed ({}): {}", cache_file.string(), loaded.error()); + } else { + cache_entries = std::move(*loaded); + cache_enabled = true; + } + } - if (ui.on_stage) ui.on_stage("Checking package checksums"); - spdlog::info("Checking MD5 on {} package(s), {} bytes total", jobs.size(), total); + const std::string_view verify_name = md5_verify_name(jobs); + + if (ui.on_stage) ui.on_stage(fmt::format("Checking package {}", verify_name)); + spdlog::debug("Checking {} on {} package(s), {} bytes total", verify_name, jobs.size(), total); if (ui.on_plan) { brokkr::odin::PlanItem pi; pi.kind = brokkr::odin::PlanItem::Kind::Part; pi.part_id = 0; pi.dev_type = 0; - pi.part_name = "Checksums"; + pi.part_name = verify_name; pi.source_base = fmt::format("{} package(s)", jobs.size()); pi.size = total; ui.on_plan({std::move(pi)}, total); @@ -234,25 +521,119 @@ brokkr::core::Status md5_verify(const std::vector& jobs, const brokkr::o if (ui.on_item_active) ui.on_item_active(0); if (ui.on_progress) ui.on_progress(0, total, 0, total); - const std::size_t threads = std::min(jobs.size(), + std::vector pending_jobs; + pending_jobs.reserve(jobs.size()); + for (const auto& j : jobs) { + if (session_verify_cache_contains(j)) { + const auto new_done = done.fetch_add(j.bytes_to_hash, std::memory_order_relaxed) + j.bytes_to_hash; + if (ui.on_progress) ui.on_progress(new_done, total, new_done, total); + spdlog::debug("Session verify cache hit: {}", j.path.string()); + continue; + } + pending_jobs.push_back(j); + } + + if (pending_jobs.empty()) { + if (ui.on_item_done) ui.on_item_done(0); + spdlog::info("{} OK", verify_name); + return {}; + } + + const std::size_t threads = std::min(pending_jobs.size(), std::max(1, std::thread::hardware_concurrency())); brokkr::core::ThreadPool pool(threads); - std::atomic_uint64_t done{0}; - for (const auto& j : jobs) { + auto persist_cache_if_needed = [&]() noexcept { + std::vector cache_snapshot; + bool should_save_cache = false; + if (cache_enabled) { + std::lock_guard lk(cache_mtx); + should_save_cache = cache_dirty; + if (should_save_cache) cache_snapshot = cache_entries; + } + + if (should_save_cache) { + auto cache_st = save_md5_xxh3_cache(cache_file, std::move(cache_snapshot), kMd5Xxh3CacheMaxEntries); + if (!cache_st) spdlog::warn("MD5/XXH3 cache save failed ({}): {}", cache_file.string(), cache_st.error()); + } + }; + + for (const auto& j : pending_jobs) { auto st = pool.submit([&, j]() -> brokkr::core::Status { if (pool.cancelled()) return {}; - auto r = md5_hash_prefetch(j.path, j.bytes_to_hash, done, total, ui); - if (!r) return brokkr::core::fail(std::move(r.error())); + std::optional cached_xxh3; + if (cache_enabled) { + std::lock_guard lk(cache_mtx); + cached_xxh3 = lookup_md5_xxh3_cache(cache_entries, j.expected, j.bytes_to_hash); + if (cached_xxh3) cache_dirty = true; + } + + if (cached_xxh3) { + auto xxh3 = hash_prefetch(j.path, j.bytes_to_hash, done, total, ui); + if (!xxh3) { + if (cache_enabled) { + std::lock_guard lk(cache_mtx); + if (forget_md5_xxh3_cache(cache_entries, j.expected, j.bytes_to_hash)) { + cache_dirty = true; + spdlog::warn("Removed MD5/XXH3 cache entry after XXH3 failure: {}", j.path.string()); + } + } + return brokkr::core::fail(std::move(xxh3.error())); + } + + if (*xxh3 == *cached_xxh3) { + spdlog::debug("MD5/XXH3 cache hit: {}", j.path.string()); + remember_session_verify_cache(j); + return {}; + } + + spdlog::warn("MD5/XXH3 cache mismatch, falling back to full MD5"); + + if (ui.on_stage) ui.on_stage("Checking package MD5"); + if (ui.on_progress) ui.on_progress(0, j.bytes_to_hash, 0, j.bytes_to_hash); + + std::atomic_uint64_t retry_done{0}; + auto retry_ui = ui; + + auto retry = hash_prefetch(j.path, j.bytes_to_hash, retry_done, j.bytes_to_hash, retry_ui); + if (!retry) return brokkr::core::fail(std::move(retry.error())); + + if (std::memcmp(retry->md5.data(), j.expected.data(), j.expected.size()) != 0) { + return brokkr::core::fail("MD5 mismatch: " + j.path.string() + "\n expected: " + md5_hex32(j.expected) + + "\n calculated: " + md5_hex32(retry->md5) + + "\n byte count: " + std::to_string(j.bytes_to_hash)); + } + + if (cache_enabled) { + std::lock_guard lk(cache_mtx); + remember_md5_xxh3_cache(cache_entries, j.expected, j.bytes_to_hash, retry->xxh3_64, + kMd5Xxh3CacheMaxEntries); + cache_dirty = true; + } + remember_session_verify_cache(j); + return {}; + } - const auto& digest = *r; - if (std::memcmp(digest.data(), j.expected.data(), j.expected.size()) != 0) { + auto digest = hash_prefetch(j.path, j.bytes_to_hash, done, total, ui); + if (!digest) return brokkr::core::fail(std::move(digest.error())); + + if (std::memcmp(digest->md5.data(), j.expected.data(), j.expected.size()) != 0) { return brokkr::core::fail("MD5 mismatch: " + j.path.string() + "\n expected: " + md5_hex32(j.expected) + - "\n calculated: " + md5_hex32(digest) + + "\n calculated: " + md5_hex32(digest->md5) + "\n byte count: " + std::to_string(j.bytes_to_hash)); } + + if (cache_enabled) { + std::lock_guard lk(cache_mtx); + remember_md5_xxh3_cache(cache_entries, j.expected, j.bytes_to_hash, digest->xxh3_64, + kMd5Xxh3CacheMaxEntries); + cache_dirty = true; + } + + remember_session_verify_cache(j); + return {}; }); @@ -260,10 +641,11 @@ brokkr::core::Status md5_verify(const std::vector& jobs, const brokkr::o } auto wst = pool.wait(); + persist_cache_if_needed(); if (!wst) return wst; if (ui.on_item_done) ui.on_item_done(0); - spdlog::info("MD5 OK"); + spdlog::info("{} OK", verify_name); return {}; } diff --git a/src/app/md5_verify.hpp b/src/app/md5_verify.hpp index 76b6d65..bbcbd2c 100644 --- a/src/app/md5_verify.hpp +++ b/src/app/md5_verify.hpp @@ -23,17 +23,22 @@ #include #include #include +#include #include namespace brokkr::app { struct Md5Job { std::filesystem::path path; + std::filesystem::path identity_path; + std::uint64_t identity_size = 0; + std::int64_t identity_write_time = 0; std::uint64_t bytes_to_hash = 0; std::array expected{}; }; brokkr::core::Result> md5_jobs(const std::vector& inputs) noexcept; +std::string_view md5_verify_name(const std::vector& jobs) noexcept; brokkr::core::Status md5_verify(const std::vector& jobs, const brokkr::odin::Ui& ui) noexcept; } // namespace brokkr::app diff --git a/src/app/md5_xxh3_cache.cpp b/src/app/md5_xxh3_cache.cpp new file mode 100644 index 0000000..6f0d8f2 --- /dev/null +++ b/src/app/md5_xxh3_cache.cpp @@ -0,0 +1,409 @@ +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#include "app/md5_xxh3_cache.hpp" + +#include +#include +#include +#include +#include +#include +#include +#include + +#if defined(_WIN32) + #ifndef NOMINMAX + #define NOMINMAX + #endif + #include +#endif + +#include + +namespace brokkr::app { +namespace { + +constexpr std::size_t kMd5HexChars = 32; +constexpr std::size_t kXxh3HexChars = 16; +constexpr std::string_view kCacheHeader = "brokkr-md5-xxh3-cache v1"; +constexpr std::size_t kMaxEntriesDefault = 65535; +constexpr std::uintmax_t kMaxCacheFileBytes = 1024 * 1024; + +struct ParsedCacheFile { + std::vector entries; + bool has_header = false; + bool saw_corruption = false; +}; + +int hex_nibble(unsigned char c) noexcept { + if (c >= '0' && c <= '9') return (c - '0'); + if (c >= 'a' && c <= 'f') return 10 + (c - 'a'); + if (c >= 'A' && c <= 'F') return 10 + (c - 'A'); + return -1; +} + +bool parse_md5_hex(std::string_view hex32, std::array& out) noexcept { + if (hex32.size() != kMd5HexChars) return false; + for (std::size_t i = 0; i < out.size(); ++i) { + const int hi = hex_nibble(static_cast(hex32[2 * i + 0])); + const int lo = hex_nibble(static_cast(hex32[2 * i + 1])); + if (hi < 0 || lo < 0) return false; + out[i] = static_cast((hi << 4) | lo); + } + return true; +} + +bool parse_u64(std::string_view sv, std::uint64_t& out) noexcept { + const char* begin = sv.data(); + const char* end = sv.data() + sv.size(); + auto [ptr, ec] = std::from_chars(begin, end, out, 10); + return ec == std::errc{} && ptr == end; +} + +bool parse_u64_hex(std::string_view sv, std::uint64_t& out) noexcept { + const char* begin = sv.data(); + const char* end = sv.data() + sv.size(); + auto [ptr, ec] = std::from_chars(begin, end, out, 16); + return ec == std::errc{} && ptr == end; +} + +std::uint64_t next_touch(const std::vector& entries) noexcept { + std::uint64_t best = 0; + for (const auto& entry : entries) best = std::max(best, entry.touched); + return (best == UINT64_MAX) ? 1 : (best + 1); +} + +void normalize_entries(std::vector& entries, std::size_t max_entries) { + std::sort(entries.begin(), entries.end(), [](const Md5Xxh3CacheEntry& lhs, const Md5Xxh3CacheEntry& rhs) { + return lhs.touched > rhs.touched; + }); + + std::vector deduped; + deduped.reserve(std::min(entries.size(), max_entries)); + for (const auto& entry : entries) { + bool duplicate = false; + for (const auto& kept : deduped) { + if (kept.md5 == entry.md5 && kept.bytes_to_hash == entry.bytes_to_hash) { + duplicate = true; + break; + } + } + if (!duplicate) deduped.push_back(entry); + if (deduped.size() == max_entries) break; + } + + entries = std::move(deduped); + if (entries.size() > max_entries) entries.resize(max_entries); + + for (std::size_t i = 0; i < entries.size(); ++i) { + entries[i].touched = static_cast(entries.size() - i); + } +} + +std::optional find_entry_index(const std::vector& entries, + const std::array& md5, + std::uint64_t bytes_to_hash) noexcept { + for (std::size_t i = 0; i < entries.size(); ++i) { + if (entries[i].md5 == md5 && entries[i].bytes_to_hash == bytes_to_hash) return i; + } + return std::nullopt; +} + +std::filesystem::path backup_cache_file(const std::filesystem::path& cache_file) noexcept { + auto out = cache_file; + out += ".bak"; + return out; +} + +std::filesystem::path temporary_cache_file(const std::filesystem::path& cache_file) noexcept { + auto out = cache_file; + out += ".tmp"; + return out; +} + +#if defined(_WIN32) +brokkr::core::Status ensure_replaceable_now(const std::filesystem::path& path) noexcept { + std::error_code ec; + if (!std::filesystem::exists(path, ec)) { + if (ec) return brokkr::core::failf("Cannot inspect cache file {}: {}", path.string(), ec.message()); + return {}; + } + if (ec) return brokkr::core::failf("Cannot inspect cache file {}: {}", path.string(), ec.message()); + + HANDLE h = CreateFileW(path.c_str(), DELETE, FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE, nullptr, + OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, nullptr); + if (h != INVALID_HANDLE_VALUE) { + CloseHandle(h); + return {}; + } + + const DWORD err = GetLastError(); + if (err == ERROR_SHARING_VIOLATION || err == ERROR_LOCK_VIOLATION || err == ERROR_ACCESS_DENIED) { + return brokkr::core::failf("Cache file busy, skipping save: {}", path.string()); + } + if (err == ERROR_FILE_NOT_FOUND || err == ERROR_PATH_NOT_FOUND) return {}; + return brokkr::core::failf("Cannot access cache file {} for replace: {}", path.string(), err); +} +#endif + +brokkr::core::Result parse_cache_file(const std::filesystem::path& path) noexcept { + std::error_code ec; + if (!std::filesystem::exists(path, ec)) return ParsedCacheFile{}; + if (ec) return brokkr::core::failf("Cannot access cache file {}: {}", path.string(), ec.message()); + + const auto size = std::filesystem::file_size(path, ec); + if (ec) return brokkr::core::failf("Cannot stat cache file {}: {}", path.string(), ec.message()); + if (size > kMaxCacheFileBytes) { + ParsedCacheFile parsed; + parsed.saw_corruption = true; + return parsed; + } + + std::ifstream in(path); + if (!in.is_open()) return brokkr::core::failf("Cannot open cache file {}", path.string()); + + ParsedCacheFile parsed; + std::string line; + std::size_t line_count = 0; + while (std::getline(in, line)) { + ++line_count; + if (line_count > 4096) { + parsed.saw_corruption = true; + break; + } + + if (!parsed.has_header) { + if (line.empty() || line[0] == '#') continue; + if (line != kCacheHeader) { + parsed.saw_corruption = true; + break; + } + parsed.has_header = true; + continue; + } + + if (line.empty() || line[0] == '#') continue; + + std::istringstream iss(line); + std::string touched_str; + std::string bytes_str; + std::string md5_str; + std::string xxh3_str; + std::string extra; + if (!(iss >> touched_str >> bytes_str >> md5_str >> xxh3_str) || (iss >> extra)) { + parsed.saw_corruption = true; + continue; + } + + Md5Xxh3CacheEntry entry; + if (!parse_u64(touched_str, entry.touched)) { + parsed.saw_corruption = true; + continue; + } + if (!parse_u64(bytes_str, entry.bytes_to_hash)) { + parsed.saw_corruption = true; + continue; + } + if (!parse_md5_hex(md5_str, entry.md5)) { + parsed.saw_corruption = true; + continue; + } + if (xxh3_str.size() != kXxh3HexChars || !parse_u64_hex(xxh3_str, entry.xxh3_64)) { + parsed.saw_corruption = true; + continue; + } + + parsed.entries.push_back(std::move(entry)); + } + + if (!in.eof() && in.fail()) return brokkr::core::failf("Cannot read cache file {}", path.string()); + + if (!parsed.has_header && !parsed.entries.empty()) parsed.saw_corruption = true; + + normalize_entries(parsed.entries, kMaxEntriesDefault); + return parsed; +} + +brokkr::core::Status replace_cache_file(const std::filesystem::path& src, + const std::filesystem::path& dst, + const std::filesystem::path& backup) noexcept { +#if defined(_WIN32) + auto dst_ready = ensure_replaceable_now(dst); + if (!dst_ready) return dst_ready; + + auto bak_ready = ensure_replaceable_now(backup); + if (!bak_ready) return bak_ready; + + if (!ReplaceFileW(dst.c_str(), src.c_str(), backup.c_str(), REPLACEFILE_IGNORE_MERGE_ERRORS, nullptr, nullptr)) { + const DWORD replace_error = GetLastError(); + if (replace_error == ERROR_FILE_NOT_FOUND || replace_error == ERROR_PATH_NOT_FOUND) { + if (MoveFileExW(src.c_str(), dst.c_str(), MOVEFILE_REPLACE_EXISTING | MOVEFILE_WRITE_THROUGH)) return {}; + return brokkr::core::failf("Cannot move cache file {} into place: {}", dst.string(), GetLastError()); + } + return brokkr::core::failf("Cannot replace cache file {}: {}", dst.string(), replace_error); + } + return {}; +#else + std::error_code ec; + if (std::filesystem::exists(dst, ec) && !ec) { + std::filesystem::copy_file(dst, backup, std::filesystem::copy_options::overwrite_existing, ec); + if (ec) return brokkr::core::failf("Cannot back up cache file {}: {}", backup.string(), ec.message()); + } else if (ec) { + return brokkr::core::failf("Cannot inspect cache file {}: {}", dst.string(), ec.message()); + } + + std::filesystem::rename(src, dst, ec); + if (ec) return brokkr::core::failf("Cannot replace cache file {}: {}", dst.string(), ec.message()); + return {}; +#endif +} + +} // namespace + +std::string md5_hex32(const std::array& digest) { + static constexpr char hex[] = "0123456789abcdef"; + std::string out(kMd5HexChars, '0'); + for (std::size_t i = 0; i < digest.size(); ++i) { + out[2 * i + 0] = hex[(digest[i] >> 4) & 0x0F]; + out[2 * i + 1] = hex[digest[i] & 0x0F]; + } + return out; +} + +std::string xxh3_hex16(std::uint64_t digest) { return fmt::format("{:016x}", digest); } + +std::filesystem::path md5_xxh3_cache_file(const std::filesystem::path& app_cache_dir) noexcept { + return app_cache_dir / "md5_xxh3_cache.txt"; +} + +brokkr::core::Result> load_md5_xxh3_cache( + const std::filesystem::path& cache_file) noexcept { + const std::array candidates = {cache_file, backup_cache_file(cache_file), temporary_cache_file(cache_file)}; + + bool saw_corruption = false; + for (const auto& candidate : candidates) { + auto parsed = parse_cache_file(candidate); + if (!parsed) return brokkr::core::fail(std::move(parsed.error())); + + if (!parsed->has_header) { + saw_corruption = saw_corruption || parsed->saw_corruption; + continue; + } + + if (parsed->saw_corruption) saw_corruption = true; + return parsed->entries; + } + + if (saw_corruption) return std::vector{}; + return std::vector{}; +} + +brokkr::core::Status save_md5_xxh3_cache(const std::filesystem::path& cache_file, + std::vector entries, + std::size_t max_entries) noexcept { + normalize_entries(entries, max_entries); + + std::error_code ec; + const auto parent = cache_file.parent_path(); + if (!parent.empty()) std::filesystem::create_directories(parent, ec); + if (ec) return brokkr::core::failf("Cannot create cache directory {}: {}", parent.string(), ec.message()); + + const auto tmp_file = temporary_cache_file(cache_file); + const auto bak_file = backup_cache_file(cache_file); + + { + std::ofstream out(tmp_file, std::ios::trunc); + if (!out.is_open()) return brokkr::core::failf("Cannot write cache file {}", tmp_file.string()); + + out << kCacheHeader << '\n'; + for (const auto& entry : entries) { + out << entry.touched << ' ' << entry.bytes_to_hash << ' ' << md5_hex32(entry.md5) << ' ' + << xxh3_hex16(entry.xxh3_64) << '\n'; + } + + out.flush(); + if (!out.good()) return brokkr::core::failf("Cannot flush cache file {}", tmp_file.string()); + } + + auto verify_tmp = parse_cache_file(tmp_file); + if (!verify_tmp) { + std::error_code rm_ec; + std::filesystem::remove(tmp_file, rm_ec); + return brokkr::core::fail(std::move(verify_tmp.error())); + } + if (!verify_tmp->has_header || verify_tmp->saw_corruption) { + std::error_code rm_ec; + std::filesystem::remove(tmp_file, rm_ec); + return brokkr::core::failf("Refusing to install malformed cache file {}", tmp_file.string()); + } + + auto replace_st = replace_cache_file(tmp_file, cache_file, bak_file); + if (!replace_st) { + std::error_code rm_ec; + std::filesystem::remove(tmp_file, rm_ec); + return replace_st; + } + + return {}; +} + +std::optional lookup_md5_xxh3_cache(std::vector& entries, + const std::array& md5, + std::uint64_t bytes_to_hash) noexcept { + const auto idx = find_entry_index(entries, md5, bytes_to_hash); + if (!idx) return std::nullopt; + + entries[*idx].touched = next_touch(entries); + return entries[*idx].xxh3_64; +} + +bool forget_md5_xxh3_cache(std::vector& entries, + const std::array& md5, + std::uint64_t bytes_to_hash) noexcept { + const auto idx = find_entry_index(entries, md5, bytes_to_hash); + if (!idx) return false; + + entries.erase(entries.begin() + static_cast(*idx)); + normalize_entries(entries, entries.size()); + return true; +} + +void remember_md5_xxh3_cache(std::vector& entries, + const std::array& md5, + std::uint64_t bytes_to_hash, + std::uint64_t xxh3_64, + std::size_t max_entries) noexcept { + const auto idx = find_entry_index(entries, md5, bytes_to_hash); + const auto touch = next_touch(entries); + if (idx) { + auto& entry = entries[*idx]; + entry.xxh3_64 = xxh3_64; + entry.touched = touch; + } else { + Md5Xxh3CacheEntry entry; + entry.md5 = md5; + entry.bytes_to_hash = bytes_to_hash; + entry.xxh3_64 = xxh3_64; + entry.touched = touch; + entries.push_back(std::move(entry)); + } + + normalize_entries(entries, max_entries); +} + +} // namespace brokkr::app \ No newline at end of file diff --git a/src/app/md5_xxh3_cache.hpp b/src/app/md5_xxh3_cache.hpp new file mode 100644 index 0000000..94e3715 --- /dev/null +++ b/src/app/md5_xxh3_cache.hpp @@ -0,0 +1,62 @@ +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#pragma once + +#include "core/status.hpp" + +#include +#include +#include +#include +#include +#include +#include + +namespace brokkr::app { + +struct Md5Xxh3CacheEntry { + std::array md5{}; + std::uint64_t bytes_to_hash = 0; + std::uint64_t xxh3_64 = 0; + std::uint64_t touched = 0; +}; + +std::string md5_hex32(const std::array& digest); +std::string xxh3_hex16(std::uint64_t digest); + +std::filesystem::path md5_xxh3_cache_file(const std::filesystem::path& app_cache_dir) noexcept; + +brokkr::core::Result> load_md5_xxh3_cache( + const std::filesystem::path& cache_file) noexcept; +brokkr::core::Status save_md5_xxh3_cache(const std::filesystem::path& cache_file, + std::vector entries, + std::size_t max_entries = 65535) noexcept; + +std::optional lookup_md5_xxh3_cache(std::vector& entries, + const std::array& md5, + std::uint64_t bytes_to_hash) noexcept; +bool forget_md5_xxh3_cache(std::vector& entries, + const std::array& md5, + std::uint64_t bytes_to_hash) noexcept; +void remember_md5_xxh3_cache(std::vector& entries, + const std::array& md5, + std::uint64_t bytes_to_hash, + std::uint64_t xxh3_64, + std::size_t max_entries = 65535) noexcept; + +} // namespace brokkr::app \ No newline at end of file diff --git a/src/core/byte_transport.hpp b/src/core/byte_transport.hpp index 02a2222..62c9cc2 100644 --- a/src/core/byte_transport.hpp +++ b/src/core/byte_transport.hpp @@ -17,6 +17,7 @@ #pragma once +#include #include #include @@ -35,6 +36,9 @@ class IByteTransport { virtual void set_timeout_ms(int ms) noexcept = 0; virtual int timeout_ms() const noexcept = 0; + // Optional hint used by protocols that negotiate preferred transfer packet sizes. + virtual void set_packet_size_hint(std::size_t bytes) noexcept { (void)bytes; } + virtual int send(std::span data, unsigned retries = 8) = 0; virtual int recv(std::span data, unsigned retries = 8) = 0; diff --git a/src/platform/linux/sysfs_usb.cpp b/src/platform/linux/sysfs_usb.cpp index b989420..df05c40 100644 --- a/src/platform/linux/sysfs_usb.cpp +++ b/src/platform/linux/sysfs_usb.cpp @@ -21,8 +21,11 @@ #include #include #include +#include #include #include +#include +#include #include @@ -133,7 +136,7 @@ std::vector enumerate_usb_devices_sysfs(const EnumerateFilte if (info->vendor != filter.vendor) continue; if (!product_allowed(info->product, filter.products)) continue; - spdlog::info("Matched USB device: {}", info->describe()); + spdlog::debug("Matched USB device: {}", info->describe()); out.emplace_back(std::move(*info)); } diff --git a/src/platform/linux/usbfs_conn.cpp b/src/platform/linux/usbfs_conn.cpp index 9b08fea..43582c9 100644 --- a/src/platform/linux/usbfs_conn.cpp +++ b/src/platform/linux/usbfs_conn.cpp @@ -37,6 +37,17 @@ constexpr std::size_t BULK_BUFFER_LENGTH_NO_LIMIT = 128 * 1024; UsbFsConnection::UsbFsConnection(UsbFsDevice& dev) : dev_(dev) {} +void UsbFsConnection::set_packet_size_hint(std::size_t bytes) noexcept { + if (bytes == 0) return; + + if (dev_.has_packet_size_limit()) + max_pack_size_ = std::min(bytes, BULK_BUFFER_LENGTH_LIMIT); + else + max_pack_size_ = bytes; + + if (max_pack_size_ == 0) max_pack_size_ = 1; +} + brokkr::core::Status UsbFsConnection::open() noexcept { if (connected_) return {}; if (!dev_.is_open()) return brokkr::core::fail("UsbFsConnection::open: device not open"); diff --git a/src/platform/linux/usbfs_conn.hpp b/src/platform/linux/usbfs_conn.hpp index 97c6168..484df6f 100644 --- a/src/platform/linux/usbfs_conn.hpp +++ b/src/platform/linux/usbfs_conn.hpp @@ -45,6 +45,7 @@ class UsbFsConnection : public brokkr::core::IByteTransport { void set_timeout_ms(int ms) noexcept override { timeout_ms_ = ms; } int timeout_ms() const noexcept override { return timeout_ms_; } + void set_packet_size_hint(std::size_t bytes) noexcept override; std::size_t max_packet_size() const noexcept { return max_pack_size_; } diff --git a/src/platform/linux/usbfs_device.cpp b/src/platform/linux/usbfs_device.cpp index e20ce27..b644d93 100644 --- a/src/platform/linux/usbfs_device.cpp +++ b/src/platform/linux/usbfs_device.cpp @@ -289,7 +289,7 @@ brokkr::core::Status UsbFsDevice::parse_descriptors_() noexcept { commit_ifc(); - spdlog::info("UsbFsDevice: {} vendor=0x{:04X} product=0x{:04X} ifc={} bulk_in=0x{:02X} bulk_out=0x{:02X}", devnode_, + spdlog::debug("UsbFsDevice: {} vendor=0x{:04X} product=0x{:04X} ifc={} bulk_in=0x{:02X} bulk_out=0x{:02X}", devnode_, ids_.vendor, ids_.product, ifc_num_, eps_.bulk_in, eps_.bulk_out); if (eps_.bulk_in == 0 || eps_.bulk_out == 0) return brokkr::core::fail("UsbFsDevice: missing bulk endpoints"); diff --git a/src/platform/macos/sysfs_usb.cpp b/src/platform/macos/sysfs_usb.cpp index b5e0161..3bfc7fb 100644 --- a/src/platform/macos/sysfs_usb.cpp +++ b/src/platform/macos/sysfs_usb.cpp @@ -105,7 +105,7 @@ void enumerate_class(const char* className, const EnumerateFilter& filter, std:: info.vendor = vendor; info.product = product; - spdlog::info("Matched USB device: {}", info.describe()); + spdlog::debug("Matched USB device: {}", info.describe()); Match m; m.info = std::move(info); diff --git a/src/platform/macos/usbfs_conn.cpp b/src/platform/macos/usbfs_conn.cpp index 00558d9..c7ae155 100644 --- a/src/platform/macos/usbfs_conn.cpp +++ b/src/platform/macos/usbfs_conn.cpp @@ -36,6 +36,17 @@ inline bool ok_or_underrun(IOReturn kr) noexcept { return kr == kIOReturnSuccess UsbFsConnection::UsbFsConnection(UsbFsDevice& dev) : dev_(dev) {} +void UsbFsConnection::set_packet_size_hint(std::size_t bytes) noexcept { + if (bytes == 0) return; + + if (dev_.has_packet_size_limit()) + max_pack_size_ = std::min(bytes, BULK_BUFFER_LENGTH_LIMIT); + else + max_pack_size_ = bytes; + + if (max_pack_size_ == 0) max_pack_size_ = 1; +} + brokkr::core::Status UsbFsConnection::open() noexcept { if (connected_) return {}; if (!dev_.is_open()) return brokkr::core::fail("UsbFsConnection::open: device not open"); diff --git a/src/platform/macos/usbfs_conn.hpp b/src/platform/macos/usbfs_conn.hpp index e8f2e3e..1b673c4 100644 --- a/src/platform/macos/usbfs_conn.hpp +++ b/src/platform/macos/usbfs_conn.hpp @@ -44,6 +44,7 @@ class UsbFsConnection : public brokkr::core::IByteTransport { void set_timeout_ms(int ms) noexcept override { timeout_ms_ = ms; } int timeout_ms() const noexcept override { return timeout_ms_; } + void set_packet_size_hint(std::size_t bytes) noexcept override; std::size_t max_packet_size() const noexcept { return max_pack_size_; } diff --git a/src/platform/macos/usbfs_device.cpp b/src/platform/macos/usbfs_device.cpp index 48f9f3d..48471cd 100644 --- a/src/platform/macos/usbfs_device.cpp +++ b/src/platform/macos/usbfs_device.cpp @@ -237,7 +237,7 @@ brokkr::core::Status UsbFsDevice::open_and_init() noexcept { return brokkr::core::fail("No interface with bulk endpoints found"); } - spdlog::info("Opened USB device at {} (VID: 0x{:04x}, PID: 0x{:04x}, bulk_in: 0x{:02x}, bulk_out: 0x{:02x})", + spdlog::debug("Opened USB device at {} (VID: 0x{:04x}, PID: 0x{:04x}, bulk_in: 0x{:02x}, bulk_out: 0x{:02x})", devnode_, ids_.vendor, ids_.product, eps_.bulk_in, eps_.bulk_out); return {}; diff --git a/src/platform/platform_all.hpp b/src/platform/platform_all.hpp index 6fe10c3..1124bae 100644 --- a/src/platform/platform_all.hpp +++ b/src/platform/platform_all.hpp @@ -16,6 +16,7 @@ */ #if defined(BROKKR_PLATFORM_LINUX) + #include "platform/posix-common/app_dirs.hpp" #include "platform/posix-common/signal_shield.hpp" #include "platform/posix-common/single_instance.hpp" #include "platform/posix-common/tcp_transport.hpp" @@ -30,6 +31,7 @@ using namespace posix_common; } // namespace brokkr::platform #elif defined(BROKKR_PLATFORM_WINDOWS) + #include "platform/windows/app_dirs.hpp" #include "platform/windows/signal_shield.hpp" #include "platform/windows/single_instance.hpp" #include "platform/windows/sysfs_usb.hpp" @@ -42,6 +44,7 @@ using namespace windows; } // namespace brokkr::platform #elif defined(BROKKR_PLATFORM_MACOS) + #include "platform/posix-common/app_dirs.hpp" #include "platform/posix-common/signal_shield.hpp" #include "platform/posix-common/single_instance.hpp" #include "platform/posix-common/tcp_transport.hpp" diff --git a/src/platform/posix-common/app_dirs.cpp b/src/platform/posix-common/app_dirs.cpp new file mode 100644 index 0000000..e818b32 --- /dev/null +++ b/src/platform/posix-common/app_dirs.cpp @@ -0,0 +1,39 @@ +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#include "platform/posix-common/app_dirs.hpp" + +#include + +namespace brokkr::posix_common { + +brokkr::core::Result app_cache_dir() noexcept { +#if defined(__APPLE__) + const char* home = std::getenv("HOME"); + if (!home || !*home) return brokkr::core::fail("HOME is not set"); + return std::filesystem::path(home) / "Library" / "Caches" / "brokkr"; +#else + const char* xdg_cache_home = std::getenv("XDG_CACHE_HOME"); + if (xdg_cache_home && *xdg_cache_home) return std::filesystem::path(xdg_cache_home) / "brokkr"; + + const char* home = std::getenv("HOME"); + if (!home || !*home) return brokkr::core::fail("HOME is not set"); + return std::filesystem::path(home) / ".cache" / "brokkr"; +#endif +} + +} // namespace brokkr::posix_common \ No newline at end of file diff --git a/src/platform/posix-common/app_dirs.hpp b/src/platform/posix-common/app_dirs.hpp new file mode 100644 index 0000000..59f7525 --- /dev/null +++ b/src/platform/posix-common/app_dirs.hpp @@ -0,0 +1,28 @@ +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#pragma once + +#include "core/status.hpp" + +#include + +namespace brokkr::posix_common { + +brokkr::core::Result app_cache_dir() noexcept; + +} // namespace brokkr::posix_common \ No newline at end of file diff --git a/src/platform/posix-common/filehandle.hpp b/src/platform/posix-common/filehandle.hpp index d171939..f43438d 100644 --- a/src/platform/posix-common/filehandle.hpp +++ b/src/platform/posix-common/filehandle.hpp @@ -146,7 +146,7 @@ struct FileHandle { const int rc = ::ioctl(fd, request, arg); if (rc < 0) { // FIX: ioctl signals error with -1; some ioctls return positive values const int e = errno; - spdlog::error("ioctl(fd={}, req={}): {}", fd, req_name, std::strerror(e)); + spdlog::debug("ioctl(fd={}, req={}): {}", fd, req_name, std::strerror(e)); errno = e; } return rc; diff --git a/src/platform/windows/app_dirs.cpp b/src/platform/windows/app_dirs.cpp new file mode 100644 index 0000000..4cff4a7 --- /dev/null +++ b/src/platform/windows/app_dirs.cpp @@ -0,0 +1,30 @@ +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#include "platform/windows/app_dirs.hpp" + +#include + +namespace brokkr::windows { + +brokkr::core::Result app_cache_dir() noexcept { + const wchar_t* local_app_data = _wgetenv(L"LOCALAPPDATA"); + if (!local_app_data || !*local_app_data) return brokkr::core::fail("LOCALAPPDATA is not set"); + return std::filesystem::path(local_app_data) / "Brokkr"; +} + +} // namespace brokkr::windows \ No newline at end of file diff --git a/src/platform/windows/app_dirs.hpp b/src/platform/windows/app_dirs.hpp new file mode 100644 index 0000000..ca33d6b --- /dev/null +++ b/src/platform/windows/app_dirs.hpp @@ -0,0 +1,28 @@ +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#pragma once + +#include "core/status.hpp" + +#include + +namespace brokkr::windows { + +brokkr::core::Result app_cache_dir() noexcept; + +} // namespace brokkr::windows \ No newline at end of file diff --git a/src/platform/windows/sysfs_usb.cpp b/src/platform/windows/sysfs_usb.cpp index 84a73eb..3f1c08e 100644 --- a/src/platform/windows/sysfs_usb.cpp +++ b/src/platform/windows/sysfs_usb.cpp @@ -22,8 +22,10 @@ #include #include #include +#include #include #include +#include #include #include @@ -49,6 +51,13 @@ struct UsbDeviceInfo { namespace { +constexpr std::string_view kSuddlmod = "AT+SUDDLMOD=0,0\r"; +constexpr std::uint16_t kOdinPids[] = {0x6601, 0x685D, 0x68C3}; + +bool is_odin_pid(std::uint16_t pid) { + return std::find(std::begin(kOdinPids), std::end(kOdinPids), pid) != std::end(kOdinPids); +} + uint16_t extract_hex4(const std::string& str, const std::string& key) { std::string lower_str = str; for (char& c : lower_str) c = static_cast(std::tolower(static_cast(c))); @@ -128,6 +137,69 @@ std::vector enumerate_usb_devices_windows(uint16_t target_vid, return result; } +std::string normalize_com_path(std::string port) { + if (port.empty()) return {}; + if (port.rfind("\\\\.\\", 0) == 0) return port; + + std::string upper = port; + for (char& c : upper) c = static_cast(std::toupper(static_cast(c))); + if (upper.rfind("COM", 0) == 0) return std::string("\\\\.\\") + port; + return {}; +} + +std::optional send_suddlmod_win32(std::string_view comPort) { + const std::string path = normalize_com_path(std::string(comPort)); + if (path.empty()) return std::string("invalid port name"); + + HANDLE h = CreateFileA(path.c_str(), GENERIC_READ | GENERIC_WRITE, 0, nullptr, OPEN_EXISTING, 0, nullptr); + if (h == INVALID_HANDLE_VALUE) return std::string("open failed: ") + std::to_string(GetLastError()); + + DCB dcb{}; + dcb.DCBlength = sizeof(DCB); + if (!GetCommState(h, &dcb)) { + const auto e = GetLastError(); + CloseHandle(h); + return std::string("GetCommState failed: ") + std::to_string(e); + } + + dcb.BaudRate = CBR_115200; + dcb.ByteSize = 8; + dcb.Parity = NOPARITY; + dcb.StopBits = ONESTOPBIT; + dcb.fOutxCtsFlow = FALSE; + dcb.fOutxDsrFlow = FALSE; + dcb.fOutX = FALSE; + dcb.fInX = FALSE; + dcb.fRtsControl = RTS_CONTROL_DISABLE; + dcb.fDtrControl = DTR_CONTROL_ENABLE; + + if (!SetCommState(h, &dcb)) { + const auto e = GetLastError(); + CloseHandle(h); + return std::string("SetCommState failed: ") + std::to_string(e); + } + + COMMTIMEOUTS to{}; + to.ReadIntervalTimeout = 50; + to.ReadTotalTimeoutConstant = 100; + to.ReadTotalTimeoutMultiplier = 10; + to.WriteTotalTimeoutConstant = 1000; + to.WriteTotalTimeoutMultiplier = 10; + (void)SetCommTimeouts(h, &to); + + DWORD wrote = 0; + const BOOL ok = WriteFile(h, kSuddlmod.data(), static_cast(kSuddlmod.size()), &wrote, nullptr); + if (!ok || wrote != static_cast(kSuddlmod.size())) { + const auto e = GetLastError(); + CloseHandle(h); + return std::string("write failed: ") + std::to_string(e); + } + + (void)FlushFileBuffers(h); + CloseHandle(h); + return std::nullopt; +} + } // namespace namespace brokkr::windows { @@ -160,4 +232,41 @@ std::optional find_by_sysname(std::string_view sysname) { return std::nullopt; } +SuddlmodResult send_suddlmod_to_samsung_serial(std::uint16_t vendor, int maxTargets) { + SuddlmodResult out; + const auto devices = enumerate_usb_devices_windows(vendor, {}); + + std::set uniquePorts; + for (const auto& d : devices) { + if (is_odin_pid(d.product)) continue; + if (!d.device_path.empty()) uniquePorts.insert(d.device_path); + } + + std::vector ports(uniquePorts.begin(), uniquePorts.end()); + if (maxTargets > 0 && static_cast(ports.size()) > maxTargets) { + ports.resize(static_cast(maxTargets)); + } + out.ports_seen = static_cast(ports.size()); + + std::vector> errors(ports.size()); + std::vector workers; + workers.reserve(ports.size()); + + for (std::size_t i = 0; i < ports.size(); ++i) { + workers.emplace_back([&, i]() { errors[i] = send_suddlmod_win32(ports[i]); }); + } + for (auto& t : workers) + if (t.joinable()) t.join(); + + for (std::size_t i = 0; i < ports.size(); ++i) { + if (errors[i]) { + out.failures.push_back(ports[i] + ": " + *errors[i]); + continue; + } + ++out.sent_ok; + } + + return out; +} + } // namespace brokkr::windows diff --git a/src/platform/windows/sysfs_usb.hpp b/src/platform/windows/sysfs_usb.hpp index e6dff5f..4f94138 100644 --- a/src/platform/windows/sysfs_usb.hpp +++ b/src/platform/windows/sysfs_usb.hpp @@ -39,7 +39,14 @@ struct EnumerateFilter { std::vector products; }; +struct SuddlmodResult { + int ports_seen = 0; + int sent_ok = 0; + std::vector failures; +}; + std::vector enumerate_usb_devices_sysfs(const EnumerateFilter& filter); std::optional find_by_sysname(std::string_view sysname); +SuddlmodResult send_suddlmod_to_samsung_serial(std::uint16_t vendor = 0x04E8, int maxTargets = -1); } // namespace brokkr::windows diff --git a/src/platform/windows/tcp_transport.cpp b/src/platform/windows/tcp_transport.cpp index 7d0fd37..b90e745 100644 --- a/src/platform/windows/tcp_transport.cpp +++ b/src/platform/windows/tcp_transport.cpp @@ -19,6 +19,7 @@ #include #include +#include #include // clang-format off @@ -33,6 +34,42 @@ namespace brokkr::windows { +namespace { + +// Winsock requires WSAStartup before any socket call, and ref-counts the +// startup/cleanup pairs. Both TcpListener and TcpConnection need WSA up for +// their lifetime, but they have independent lifetimes (the listener may be +// destroyed while a TcpConnection it produced is still in use). A single +// process-wide ref-counted scope keeps WSA initialized as long as any of +// them is alive. +struct WsaScope { + static void acquire() noexcept { + static std::mutex m; + std::lock_guard lk(m); + if (refs_++ == 0) { + WSADATA d{}; + const int err = WSAStartup(MAKEWORD(2, 2), &d); + if (err) { + spdlog::error("WSAStartup failed: {}", err); + // Keep refs_ bumped so we don't repeatedly retry; subsequent socket + // calls will surface the error. + } + } + } + static void release() noexcept { + static std::mutex m; + std::lock_guard lk(m); + if (refs_ > 0 && --refs_ == 0) { + WSACleanup(); + } + } + +private: + static inline int refs_ = 0; +}; + +} // namespace + static void set_nonblocking_(SOCKET s, bool on) noexcept { u_long mode = on ? 1UL : 0UL; (void)::ioctlsocket(s, FIONBIO, &mode); @@ -40,6 +77,8 @@ static void set_nonblocking_(SOCKET s, bool on) noexcept { TcpConnection::TcpConnection(SOCKET fd, std::string peer_ip, std::uint16_t peer_port) : fd_(fd), peer_ip_(std::move(peer_ip)), peer_port_(peer_port) { + WsaScope::acquire(); + wsa_init_ = true; if (fd_ != INVALID_SOCKET) set_nonblocking_(fd_, true); set_sock_timeouts_(); @@ -48,19 +87,25 @@ TcpConnection::TcpConnection(SOCKET fd, std::string peer_ip, std::uint16_t peer_ (void)::setsockopt(fd_, IPPROTO_TCP, TCP_NODELAY, reinterpret_cast(&one), sizeof(one)); } -TcpConnection::~TcpConnection() { close_(); } +TcpConnection::~TcpConnection() { + close_(); + if (wsa_init_) WsaScope::release(); +} TcpConnection::TcpConnection(TcpConnection&& o) noexcept { *this = std::move(o); } TcpConnection& TcpConnection::operator=(TcpConnection&& o) noexcept { if (this == &o) return *this; close_(); + if (wsa_init_) WsaScope::release(); fd_ = o.fd_; o.fd_ = INVALID_SOCKET; timeout_ms_ = o.timeout_ms_; peer_ip_ = std::move(o.peer_ip_); peer_port_ = o.peer_port_; o.peer_port_ = 0; + wsa_init_ = o.wsa_init_; + o.wsa_init_ = false; return *this; } @@ -213,13 +258,12 @@ int TcpConnection::recv(std::span data, unsigned /*retries*/) { TcpListener::~TcpListener() { if (fd_ != INVALID_SOCKET) ::closesocket(fd_); fd_ = INVALID_SOCKET; - if (wsa_init_) WSACleanup(); + if (wsa_init_) WsaScope::release(); } brokkr::core::Status TcpListener::bind_and_listen(std::string bind_ip, std::uint16_t port, int backlog) noexcept { if (!wsa_init_) { - const int err = WSAStartup(MAKEWORD(2, 2), &ws_); - if (err) return brokkr::core::failf("WSAStartup failed: {}", err); + WsaScope::acquire(); wsa_init_ = true; } diff --git a/src/platform/windows/tcp_transport.hpp b/src/platform/windows/tcp_transport.hpp index f003fd6..308d138 100644 --- a/src/platform/windows/tcp_transport.hpp +++ b/src/platform/windows/tcp_transport.hpp @@ -66,6 +66,7 @@ class TcpConnection final : public brokkr::core::IByteTransport { std::string peer_ip_; std::uint16_t peer_port_ = 0; + bool wsa_init_ = false; }; class TcpListener { diff --git a/src/platform/windows/usbfs_conn.cpp b/src/platform/windows/usbfs_conn.cpp index df5e8e5..5d1b745 100644 --- a/src/platform/windows/usbfs_conn.cpp +++ b/src/platform/windows/usbfs_conn.cpp @@ -40,6 +40,11 @@ void backoff_10ms() noexcept { ::Sleep(10); } UsbFsConnection::UsbFsConnection(UsbFsDevice& dev) : dev_(dev) {} +void UsbFsConnection::set_packet_size_hint(std::size_t bytes) noexcept { + if (bytes == 0) return; + max_pack_size_ = bytes; +} + brokkr::core::Status UsbFsConnection::open() noexcept { if (connected_ && dev_.is_open()) return {}; diff --git a/src/platform/windows/usbfs_conn.hpp b/src/platform/windows/usbfs_conn.hpp index d8ebf8c..e0ed3c4 100644 --- a/src/platform/windows/usbfs_conn.hpp +++ b/src/platform/windows/usbfs_conn.hpp @@ -30,7 +30,7 @@ class UsbFsConnection : public brokkr::core::IByteTransport { public: explicit UsbFsConnection(UsbFsDevice& dev); - Kind kind() const noexcept override { return Kind::TcpStream; } + Kind kind() const noexcept override { return Kind::UsbBulk; } brokkr::core::Status open() noexcept; void close() noexcept; @@ -43,13 +43,14 @@ class UsbFsConnection : public brokkr::core::IByteTransport { void set_timeout_ms(int ms) noexcept override { timeout_ms_ = ms; } int timeout_ms() const noexcept override { return timeout_ms_; } + void set_packet_size_hint(std::size_t bytes) noexcept override; private: UsbFsDevice& dev_; bool connected_ = false; int timeout_ms_ = 1000; - std::size_t max_pack_size_ = 128 * 1024; + std::size_t max_pack_size_ = 1 * 1024 * 1024; }; } // namespace brokkr::windows diff --git a/src/protocol/odin/flash.cpp b/src/protocol/odin/flash.cpp index 76eb9e1..539c6a6 100644 --- a/src/protocol/odin/flash.cpp +++ b/src/protocol/odin/flash.cpp @@ -110,7 +110,7 @@ static brokkr::core::Result lz4_content_size(const ImageSpec& spe static brokkr::core::Result make_spec(ImageSpec::Kind kind, std::filesystem::path path, io::TarEntry entry, std::string display, std::string source_basename, - std::uint64_t disk_size, bool dl_mode) noexcept { + std::uint64_t disk_size) noexcept { ImageSpec spec; spec.kind = kind; spec.path = std::move(path); @@ -122,7 +122,6 @@ static brokkr::core::Result make_spec(ImageSpec::Kind kind, std::file spec.lz4 = is_lz4_name(spec.source_basename); spec.basename = spec.lz4 ? strip_lz4_suffix(spec.source_basename) : spec.source_basename; - spec.download_list_mode = dl_mode; if (spec.lz4) { BRK_TRYV(sz, lz4_content_size(spec)); @@ -134,22 +133,6 @@ static brokkr::core::Result make_spec(ImageSpec::Kind kind, std::file return spec; } -struct SourceCandidate { - ImageSpec::Kind kind{}; - std::filesystem::path path; - io::TarEntry entry{}; - std::string basename, source_basename, display; - std::uint64_t disk_size = 0; -}; - -static void merge(std::unordered_map& out, SourceCandidate c) { - if (!c.basename.empty()) out.insert_or_assign(c.basename, std::move(c)); -} - -static brokkr::core::Result finalize(const SourceCandidate& c, bool dl_mode) noexcept { - return make_spec(c.kind, c.path, c.entry, c.display, c.source_basename, c.disk_size, dl_mode); -} - } // namespace brokkr::core::Result> ImageSpec::open() const noexcept { @@ -163,9 +146,12 @@ brokkr::core::Result> ImageSpec::open() const no brokkr::core::Result> expand_inputs_tar_or_raw( const std::vector& inputs) noexcept { std::vector out; + + std::vector> tars(inputs.size()); std::optional> dl; - for (const auto& p : inputs) { + for (std::size_t i = 0; i < inputs.size(); ++i) { + const auto& p = inputs[i]; if (!io::TarArchive::is_tar_file(p.string())) continue; BRK_TRYV(tar, io::TarArchive::open(p.string(), true)); @@ -180,81 +166,116 @@ brokkr::core::Result> expand_inputs_tar_or_raw( else if (!lists_equal(*dl, names)) return brokkr::core::fail("Multiple download-list.txt files found with different contents"); } + + tars[i].emplace(std::move(tar)); } + std::unordered_set allow; if (dl) { - std::unordered_map cands; + allow.reserve(dl->size()); + for (const auto& n : *dl) allow.insert(n); + } - for (const auto& p : inputs) { - if (io::TarArchive::is_tar_file(p.string())) { - BRK_TRYV(tar, io::TarArchive::open(p.string(), true)); + auto compute_basename = [](std::string_view sb) -> std::string { + if (sb.empty() || sb.back() == '/') return {}; + return is_lz4_name(sb) ? strip_lz4_suffix(std::string(sb)) : std::string(sb); + }; - for (const auto& e : tar.entries()) { - if (is_download_list_name(e.name)) continue; + struct Coord { std::size_t inp; std::size_t ent; }; + static constexpr std::size_t kRaw = static_cast(-1); + std::unordered_map last_of; - const std::string sb = io::basename(e.name); - if (sb.empty() || sb.back() == '/') continue; + auto record = [&](const std::string& base, std::size_t inp, std::size_t ent) { + auto [it, ins] = last_of.try_emplace(base, Coord{inp, ent}); + if (!ins) { + spdlog::debug("Duplicate image '{}' across inputs \u2014 later occurrence wins", base); + it->second = Coord{inp, ent}; + } + }; + + for (std::size_t i = 0; i < inputs.size(); ++i) { + if (tars[i]) { + const auto& entries = tars[i]->entries(); + for (std::size_t j = 0; j < entries.size(); ++j) { + if (is_download_list_name(entries[j].name)) continue; + const std::string sb = io::basename(entries[j].name); + const std::string base = compute_basename(sb); + if (base.empty()) continue; + record(base, i, j); + } + } else { + const std::string sb = io::basename(inputs[i].string()); + const std::string base = compute_basename(sb); + if (base.empty()) continue; + record(base, i, kRaw); + } + } - const bool lz4 = is_lz4_name(sb); - const std::string base = lz4 ? strip_lz4_suffix(sb) : sb; + std::vector pit_specs; + std::unordered_set emitted; - merge(cands, SourceCandidate{.kind = ImageSpec::Kind::TarEntry, - .path = p, - .entry = e, - .basename = base, - .source_basename = sb, - .display = p.string() + ":" + e.name, - .disk_size = e.size}); - } - continue; - } + for (std::size_t i = 0; i < inputs.size(); ++i) { + const auto& p = inputs[i]; - BRK_TRYV(src, io::open_raw_file(p)); - const std::string sb = io::basename(p.string()); - const bool lz4 = is_lz4_name(sb); - - merge(cands, SourceCandidate{.kind = ImageSpec::Kind::RawFile, - .path = p, - .entry = {}, - .basename = lz4 ? strip_lz4_suffix(sb) : sb, - .source_basename = sb, - .display = p.string(), - .disk_size = src->size()}); - } + if (tars[i]) { + const auto& entries = tars[i]->entries(); + for (std::size_t j = 0; j < entries.size(); ++j) { + const auto& e = entries[j]; + if (is_download_list_name(e.name)) continue; + const std::string sb = io::basename(e.name); + const std::string base = compute_basename(sb); + if (base.empty()) continue; - out.reserve(dl->size()); - for (const auto& name : *dl) { - auto it = cands.find(name); - if (it == cands.end()) return brokkr::core::fail("download-list.txt references missing file: " + name); + const auto lit = last_of.find(base); + if (lit == last_of.end() || lit->second.inp != i || lit->second.ent != j) continue; - BRK_TRYV(spec, finalize(it->second, true)); - out.push_back(std::move(spec)); - } + const bool is_pit = brokkr::core::ends_with_ci(base, ".pit"); + if (dl && !is_pit && !allow.contains(base)) { + spdlog::debug("Skipping {}: not in download-list.txt", sb); + continue; + } - return out; - } + BRK_TRYV(spec, make_spec(ImageSpec::Kind::TarEntry, p, e, p.string() + ":" + e.name, sb, e.size)); + if (is_pit) { + pit_specs.push_back(std::move(spec)); + } else { + emitted.insert(base); + out.push_back(std::move(spec)); + } + } + } else { + const std::string sb = io::basename(p.string()); + const std::string base = compute_basename(sb); + if (base.empty()) continue; - for (const auto& p : inputs) { - if (io::TarArchive::is_tar_file(p.string())) { - BRK_TRYV(tar, io::TarArchive::open(p.string(), true)); + const auto lit = last_of.find(base); + if (lit == last_of.end() || lit->second.inp != i || lit->second.ent != kRaw) continue; - for (const auto& e : tar.entries()) { - if (is_download_list_name(e.name)) continue; - const auto sb = io::basename(e.name); - if (sb.empty()) continue; + const bool is_pit = brokkr::core::ends_with_ci(base, ".pit"); + if (dl && !is_pit && !allow.contains(base)) { + spdlog::debug("Skipping {}: not in download-list.txt", sb); + continue; + } - BRK_TRYV(spec, make_spec(ImageSpec::Kind::TarEntry, p, e, p.string() + ":" + e.name, io::basename(e.name), - e.size, false)); + BRK_TRYV(src, io::open_raw_file(p)); + BRK_TRYV(spec, make_spec(ImageSpec::Kind::RawFile, p, {}, p.string(), sb, src->size())); + if (is_pit) { + pit_specs.push_back(std::move(spec)); + } else { + emitted.insert(base); out.push_back(std::move(spec)); } - continue; } + } - BRK_TRYV(src, io::open_raw_file(p)); - BRK_TRYV(spec, make_spec(ImageSpec::Kind::RawFile, p, {}, p.string(), io::basename(p.string()), src->size(), false)); - out.push_back(std::move(spec)); + if (dl) { + for (const auto& name : *dl) { + if (!emitted.contains(name)) + spdlog::debug("download-list.txt references missing file: {} (skipping)", name); + } } + for (auto& s : pit_specs) out.push_back(std::move(s)); return out; } @@ -277,8 +298,10 @@ brokkr::core::Result> map_to_pit(const pit::PitTable& pit items[it->second] = FlashItem{.part = *part, .spec = s}; } - if (items.empty()) return brokkr::core::fail("No flashable items after PIT mapping"); - spdlog::debug("PIT mapping: {} items", items.size()); + if (items.empty()) + spdlog::warn("No flashable items after PIT mapping"); + else + spdlog::debug("PIT mapping: {} items", items.size()); return items; } diff --git a/src/protocol/odin/flash.hpp b/src/protocol/odin/flash.hpp index 670689d..b8f2fbc 100644 --- a/src/protocol/odin/flash.hpp +++ b/src/protocol/odin/flash.hpp @@ -53,7 +53,6 @@ struct ImageSpec { std::uint64_t disk_size = 0; bool lz4 = false; - bool download_list_mode = false; std::string display; diff --git a/src/protocol/odin/group_flasher.cpp b/src/protocol/odin/group_flasher.cpp index 3bc92d4..44124db 100644 --- a/src/protocol/odin/group_flasher.cpp +++ b/src/protocol/odin/group_flasher.cpp @@ -66,27 +66,26 @@ struct FirstError { } }; -constexpr std::string_view kHandshakeStr = "ODIN handshake"; -constexpr std::string_view kPktFlashStr = "Negotiating transfer options"; -constexpr std::string_view kPitDlStr = "Downloading PIT(s)"; -constexpr std::string_view kPitUpStr = "Uploading PIT"; +namespace stage_label { +constexpr std::string_view kHandshake = "ODIN handshake"; +constexpr std::string_view kPktFlash = "Negotiating transfer options"; +constexpr std::string_view kPitDl = "Downloading PIT(s)"; +constexpr std::string_view kPitUp = "Uploading PIT"; constexpr std::string_view kCpuCheck = "Checking if devices are equal"; constexpr std::string_view kMapCheck = "Verifying PIT mapping"; constexpr std::string_view kTotalSend = "Sending total size"; constexpr std::string_view kFlashFast = "Flashing (Speed: Enhanced)"; constexpr std::string_view kFlashNorm = "Flashing (Speed: Normal)"; -constexpr std::string_view kRebooting = "Rebooting devices"; +constexpr std::string_view kFinalize = "Finalizing"; +constexpr std::string_view kFinalizeReboot = "Finalizing + reboot"; +} // namespace stage_label static std::string_view final_stage(OdinCommands::ShutdownMode m) { - if (m == OdinCommands::ShutdownMode::ReDownload) return "Finalizing + redownload"; - if (m == OdinCommands::ShutdownMode::Reboot) return "Finalizing + reboot"; - return "Finalizing"; + return m == OdinCommands::ShutdownMode::Reboot ? stage_label::kFinalizeReboot : stage_label::kFinalize; } static OdinCommands::ShutdownMode shutdown_mode_final(const Cfg& cfg) { - return cfg.redownload_after ? OdinCommands::ShutdownMode::ReDownload - : cfg.reboot_after ? OdinCommands::ShutdownMode::Reboot - : OdinCommands::ShutdownMode::NoReboot; + return cfg.reboot_after ? OdinCommands::ShutdownMode::Reboot : OdinCommands::ShutdownMode::NoReboot; } static void log_summary(std::size_t total, std::size_t failed) { @@ -96,10 +95,6 @@ static void log_summary(std::size_t total, std::size_t failed) { } static void log_shutdown_action(OdinCommands::ShutdownMode m) { - if (m == OdinCommands::ShutdownMode::ReDownload) { - spdlog::info("Re-download"); - return; - } if (m == OdinCommands::ShutdownMode::Reboot) { spdlog::info("Reset"); return; @@ -107,8 +102,6 @@ static void log_shutdown_action(OdinCommands::ShutdownMode m) { spdlog::info("No Reboot"); } -static void log_speed(bool enhanced) { spdlog::info("Speed: {}", enhanced ? "Enhanced" : "Normal"); } - static void emit_devfail(const Ui& ui, std::size_t orig_idx, const std::string& msg) { if (ui.on_error) ui.on_error("DEVFAIL idx=" + std::to_string(orig_idx) + " " + msg); @@ -138,14 +131,14 @@ static brokkr::core::Result> sources_common_mapping_or_em for (const auto& s : sources) { const auto* ref = devs.front()->pit_table.find_by_file_name(s.basename); if (!ref) { - spdlog::warn("Source '{}' has no matching PIT entry — skipped", s.basename); + spdlog::debug("Source '{}' has no matching PIT entry — skipped", s.basename); continue; } for (auto* d : devs) { const auto* p = d->pit_table.find_by_file_name(s.basename); if (!p) { - spdlog::warn("Source '{}' missing on device '{}' — skipped", s.basename, d->id); + spdlog::debug("Source '{}' missing on one or more devices — skipped", s.basename); goto next; } if (p->id != ref->id || p->dev_type != ref->dev_type) @@ -241,8 +234,7 @@ brokkr::core::Status flash(std::vector& devs, const std::vectorempty(); - enum class Intent : std::uint8_t { RebootOnly, PitOnly, Flash }; - const Intent intent = has_sources ? Intent::Flash : (has_pit ? Intent::PitOnly : Intent::RebootOnly); + if (!has_sources && !has_pit) return brokkr::core::fail("flash: nothing to do (no sources, no PIT)"); const std::size_t total_devices = devs.size(); std::size_t failed_total = 0; @@ -325,465 +317,421 @@ brokkr::core::Status flash(std::vector& devs, const std::vector brokkr::core::Status { log_shutdown_action(m); stage(stg); - return fanout_keep([&](Target& d) { - spdlog::info("(dev={}) Shutdown", d.id); - return OdinCommands(link(d)).shutdown(m); - }); + return fanout_keep([&](Target& d) { return OdinCommands(link(d)).shutdown(m); }); }; auto steps = std::vector>{}; steps.emplace_back([&] { - spdlog::info("HANDSHAKE"); - stage(kHandshakeStr); + spdlog::info("> Odin"); + stage(stage_label::kHandshake); - return fanout_keep([&](Target& d) -> brokkr::core::Status { + auto st = fanout_keep([&](Target& d) -> brokkr::core::Status { auto& c = link(d); c.set_timeout_ms(cfg.preflash_timeout_ms); OdinCommands odin(c); - spdlog::info("(dev={}) Handshake", d.id); BRK_TRY(odin.handshake(cfg.preflash_retries)); BRK_TRYV(vr, odin.get_version(cfg.preflash_retries)); d.init = vr; d.proto = d.init.protocol(); - spdlog::info("(dev={}) Protocol v{}", d.id, static_cast(d.proto)); return {}; }); + if (st) spdlog::info("< Loke"); + return st; }); - if (intent != Intent::RebootOnly) { - steps.emplace_back([&] -> brokkr::core::Status { - if (active.empty()) return brokkr::core::fail("No active devices"); - - pkt = choose_pkt(active, cfg); - stage(kPktFlashStr); + steps.emplace_back([&] -> brokkr::core::Status { + if (active.empty()) return brokkr::core::fail("No active devices"); - auto st = fanout_keep([&](Target& d) -> brokkr::core::Status { - if (d.proto < ProtocolVersion::PROTOCOL_VER2) return {}; - auto& c = link(d); - c.set_timeout_ms(cfg.preflash_timeout_ms); - spdlog::info("(dev={}) Setting transfer options", d.id); - return OdinCommands(c).setup_transfer_options(static_cast(pkt), cfg.preflash_retries); - }); - if (!st) return st; + pkt = choose_pkt(active, cfg); + stage(stage_label::kPktFlash); - set_flash_timeout_active(); - return {}; + auto st = fanout_keep([&](Target& d) -> brokkr::core::Status { + if (d.proto < ProtocolVersion::PROTOCOL_VER2) return {}; + auto& c = link(d); + c.set_timeout_ms(cfg.preflash_timeout_ms); + return OdinCommands(c).setup_transfer_options(static_cast(pkt), cfg.preflash_retries); }); - } + if (!st) return st; + + set_flash_timeout_active(); + return {}; + }); if (has_pit) { steps.emplace_back([&] { spdlog::info("Uploading PIT"); - stage(kPitUpStr); + stage(stage_label::kPitUp); return fanout_keep([&](Target& d) { - spdlog::info("(dev={}) Uploading PIT", d.id); return OdinCommands(link(d)).set_pit({pit_to_upload->data(), pit_to_upload->size()}, cfg.preflash_retries); }); }); } - if (intent != Intent::PitOnly) { - steps.emplace_back([&] { - spdlog::info("Get PIT for mapping"); - stage(kPitDlStr); - set_flash_timeout_active(); - - return fanout_keep([&](Target& d) -> brokkr::core::Status { - spdlog::info("(dev={}) Downloading PIT", d.id); - OdinCommands odin(link(d)); - BRK_TRYV(bytes, download_pit_bytes(odin)); - d.pit_bytes = std::move(bytes); - BRK_TRYV(t, pit::parse({d.pit_bytes.data(), d.pit_bytes.size()})); - d.pit_table = std::move(t); - spdlog::info("(dev={}) PIT OK ({} partitions)", d.id, d.pit_table.partitions.size()); - return {}; - }); - }); - } - - if (intent == Intent::RebootOnly) { - steps.emplace_back([&] { - return shutdown_active( - cfg.reboot_after ? OdinCommands::ShutdownMode::Reboot : OdinCommands::ShutdownMode::NoReboot, kRebooting); - }); - } + steps.emplace_back([&] { + spdlog::info("Get PIT for mapping"); + stage(stage_label::kPitDl); + set_flash_timeout_active(); - if (intent == Intent::PitOnly) { - steps.emplace_back([&] -> brokkr::core::Status { - auto pr = pit::parse({pit_to_upload->data(), pit_to_upload->size()}); - if (!pr) return brokkr::core::fail(std::move(pr.error())); - if (ui.on_model) ui.on_model(pr->cpu_bl_id); - - if (ui.on_plan) { - PlanItem pi{.kind = PlanItem::Kind::Pit, - .part_name = "PIT (repartition)", - .pit_file_name = "PIT", - .source_base = "PIT", - .size = pit_to_upload->size()}; - ui.on_plan({std::move(pi)}, static_cast(pit_to_upload->size())); - } - if (ui.on_item_active) ui.on_item_active(0); - if (ui.on_progress) { - const auto n = static_cast(pit_to_upload->size()); - ui.on_progress(0, n, 0, n); - } - if (ui.on_progress) { - const auto n = static_cast(pit_to_upload->size()); - ui.on_progress(n, n, n, n); - } - if (ui.on_item_done) ui.on_item_done(0); + return fanout_keep([&](Target& d) -> brokkr::core::Status { + OdinCommands odin(link(d)); + BRK_TRYV(bytes, download_pit_bytes(odin)); + d.pit_bytes = std::move(bytes); + BRK_TRYV(t, pit::parse({d.pit_bytes.data(), d.pit_bytes.size()})); + d.pit_table = std::move(t); return {}; }); + }); - steps.emplace_back([&] { return shutdown_active(sm_final, final_stage(sm_final)); }); - } - - if (intent == Intent::Flash) { - steps.emplace_back([&] -> brokkr::core::Status { - if (active.empty()) return brokkr::core::fail("No active devices"); + steps.emplace_back([&] -> brokkr::core::Status { + if (active.empty()) return brokkr::core::fail("No active devices"); - stage(kCpuCheck); - const std::string ref = active.front()->pit_table.cpu_bl_id; - if (ref.empty()) return brokkr::core::fail("PIT cpu_bl_id missing"); - for (auto* d : active) - if (d->pit_table.cpu_bl_id != ref) return brokkr::core::fail("cpu_bl_id mismatch across devices"); - if (ui.on_model) ui.on_model(ref); - return {}; - }); + stage(stage_label::kCpuCheck); + const std::string ref = active.front()->pit_table.cpu_bl_id; + if (ref.empty()) return brokkr::core::fail("PIT cpu_bl_id missing"); + for (auto* d : active) + if (d->pit_table.cpu_bl_id != ref) return brokkr::core::fail("cpu_bl_id mismatch across devices"); + if (ui.on_model) ui.on_model(ref); + return {}; + }); - steps.emplace_back([&] -> brokkr::core::Status { - spdlog::info("Verifying PIT mapping"); - stage(kMapCheck); - - BRK_TRYV(eff, sources_common_mapping_or_empty(active, sources)); - effective_sources = std::move(eff); - - if (effective_sources.empty() && !has_pit) - return brokkr::core::fail("No sources matched any PIT partition — nothing to flash"); - - if (effective_sources.size() < sources.size()) - spdlog::warn("{} of {} source(s) matched PIT entries", effective_sources.size(), sources.size()); - - BRK_TRYV(items2, map_to_pit(active.front()->pit_table, effective_sources)); - items = std::move(items2); - - total = 0; - for (const auto& it : items) BRK_TRY(detail::checked_add_u64(total, it.spec.size, "TOTALSIZE")); - - plan.clear(); - plan.reserve(items.size() + (has_pit ? 1u : 0u)); - - if (has_pit) - plan.push_back(PlanItem{.kind = PlanItem::Kind::Pit, - .part_name = "PIT (repartition)", - .pit_file_name = "PIT", - .source_base = "PIT", - .size = pit_to_upload->size()}); - for (const auto& it : items) { - plan.push_back( - PlanItem{.kind = PlanItem::Kind::Part, - .part_id = it.part.id, - .dev_type = it.part.dev_type, - .part_name = !it.part.name.empty() ? it.part.name : it.part.file_name, - .pit_file_name = it.part.file_name, - .source_base = it.spec.source_basename.empty() ? it.spec.basename : it.spec.source_basename, - .size = it.spec.size}); - } + steps.emplace_back([&] -> brokkr::core::Status { + spdlog::info("Verifying PIT mapping"); + stage(stage_label::kMapCheck); + + BRK_TRYV(eff, sources_common_mapping_or_empty(active, sources)); + effective_sources = std::move(eff); + + if (effective_sources.empty() && !sources.empty()) + spdlog::debug("No sources matched any PIT partition — nothing to flash from sources"); + else if (effective_sources.size() < sources.size()) + spdlog::debug("{} of {} source(s) matched PIT entries", effective_sources.size(), sources.size()); + + BRK_TRYV(items2, map_to_pit(active.front()->pit_table, effective_sources)); + items = std::move(items2); + + total = 0; + for (const auto& it : items) BRK_TRY(detail::checked_add_u64(total, it.spec.size, "TOTALSIZE")); + + plan.clear(); + plan.reserve(items.size() + (has_pit ? 1u : 0u)); + + if (has_pit) + plan.push_back(PlanItem{.kind = PlanItem::Kind::Pit, + .part_name = "PIT", + .pit_file_name = "PIT", + .source_base = "PIT", + .size = pit_to_upload->size()}); + for (const auto& it : items) { + plan.push_back( + PlanItem{.kind = PlanItem::Kind::Part, + .part_id = it.part.id, + .dev_type = it.part.dev_type, + .part_name = !it.part.name.empty() ? it.part.name : it.part.file_name, + .pit_file_name = it.part.file_name, + .source_base = it.spec.source_basename.empty() ? it.spec.basename : it.spec.source_basename, + .size = it.spec.size}); + } - if (ui.on_plan) ui.on_plan(plan, total); - return {}; - }); + if (ui.on_plan) ui.on_plan(plan, total); + return {}; + }); - steps.emplace_back([&] { - stage(kTotalSend); - return fanout_keep( - [&](Target& d) { return OdinCommands(link(d)).send_total_size(total, d.proto, cfg.preflash_retries); }); - }); + steps.emplace_back([&] -> brokkr::core::Status { + if (items.empty()) return {}; + stage(stage_label::kTotalSend); + return fanout_keep( + [&](Target& d) { return OdinCommands(link(d)).send_total_size(total, d.proto, cfg.preflash_retries); }); + }); - steps.emplace_back([&] -> brokkr::core::Status { - spdlog::info("Flashing has begun!"); - - const bool use_lz4 = any_lz4(effective_sources) && std::all_of(active.begin(), active.end(), [](Target* d) { - return d->init.supports_compressed_download(); - }); - - log_speed(use_lz4); - stage(use_lz4 ? kFlashFast : kFlashNorm); - - const std::size_t ndevs = active.size(); - if (!ndevs) return brokkr::core::fail("No active devices"); - - Step cur{}; - std::barrier sync(static_cast(ndevs + 1)); - FirstError berr; - std::atomic_uint32_t failed_count{0}; - std::vector dead(ndevs, 0); - - auto exec = [&](OdinCommands& odin, const Step& s) -> brokkr::core::Status { - if (s.op == Step::Op::Begin) - return s.comp ? odin.begin_download_compressed(static_cast(s.a)) - : odin.begin_download(static_cast(s.a)); - if (s.op == Step::Op::Data) { - BRK_TRY(odin.send_raw({s.base + static_cast(s.off), s.n})); - BRK_TRYV(_, odin.recv_checked_response(static_cast(RqtCommandType::RQT_EMPTY), nullptr)); - return {}; - } - if (s.op == Step::Op::End) - return s.comp ? odin.end_download_compressed(static_cast(s.a), s.part_id, s.dev_type, s.last) - : odin.end_download(static_cast(s.a), s.part_id, s.dev_type, s.last); + steps.emplace_back([&] -> brokkr::core::Status { + const bool use_lz4 = any_lz4(effective_sources) && std::all_of(active.begin(), active.end(), [](Target* d) { + return d->init.supports_compressed_download(); + }); + + stage(use_lz4 ? stage_label::kFlashFast : stage_label::kFlashNorm); + spdlog::info("Flashing has begun!"); + + const std::size_t ndevs = active.size(); + if (!ndevs) return brokkr::core::fail("No active devices"); + + Step cur{}; + std::barrier sync(static_cast(ndevs + 1)); + FirstError berr; + std::atomic_uint32_t failed_count{0}; + std::vector dead(ndevs, 0); + + auto exec = [&](OdinCommands& odin, const Step& s) -> brokkr::core::Status { + if (s.op == Step::Op::Begin) + return s.comp ? odin.begin_download_compressed(static_cast(s.a)) + : odin.begin_download(static_cast(s.a)); + if (s.op == Step::Op::Data) { + BRK_TRY(odin.send_raw({s.base + static_cast(s.off), s.n})); + BRK_TRYV(_, odin.recv_checked_response(static_cast(RqtCommandType::RQT_EMPTY), nullptr)); return {}; - }; - - std::vector workers; - workers.reserve(ndevs); - - for (std::size_t i = 0; i < ndevs; ++i) { - auto* d = active[i]; - const std::size_t orig = active_idx[i]; - - workers.emplace_back([&, d, i, orig](std::stop_token stt) { - spdlog::info("(dev={}) Flash worker started", d->id); - OdinCommands odin(link(*d)); - bool dead_local = false; - - for (;;) { - sync.arrive_and_wait(); - const Step s = cur; - - const bool quit = (s.op == Step::Op::Quit) || stt.stop_requested(); - if (!quit && !dead_local) { - auto rst = exec(odin, s); - if (!rst) { - dead[i] = 1; - failed_count.fetch_add(1, std::memory_order_relaxed); - const auto msg = rst.error(); - berr.set(std::move(rst)); - emit_devfail(ui, orig, msg); - dead_local = true; - } + } + if (s.op == Step::Op::End) + return s.comp ? odin.end_download_compressed(static_cast(s.a), s.part_id, s.dev_type, s.last) + : odin.end_download(static_cast(s.a), s.part_id, s.dev_type, s.last); + return {}; + }; + + std::vector workers; + workers.reserve(ndevs); + + for (std::size_t i = 0; i < ndevs; ++i) { + auto* d = active[i]; + const std::size_t orig = active_idx[i]; + + workers.emplace_back([&, d, i, orig](std::stop_token stt) { + OdinCommands odin(link(*d)); + bool dead_local = false; + + for (;;) { + sync.arrive_and_wait(); + const Step s = cur; + + const bool quit = (s.op == Step::Op::Quit) || stt.stop_requested(); + if (!quit && !dead_local) { + auto rst = exec(odin, s); + if (!rst) { + dead[i] = 1; + failed_count.fetch_add(1, std::memory_order_relaxed); + const auto msg = rst.error(); + berr.set(std::move(rst)); + emit_devfail(ui, orig, msg); + dead_local = true; } - - sync.arrive_and_wait(); - if (quit) break; } - }); + + sync.arrive_and_wait(); + if (quit) break; + } + }); + } + + auto emit = [&](Step s) { + cur = s; + sync.arrive_and_wait(); + sync.arrive_and_wait(); + }; + + auto coordinator = [&]() -> brokkr::core::Status { + u64 overall_done = 0; + + std::size_t plan_off = 0; + if (has_pit) { + if (ui.on_item_active) ui.on_item_active(0); + if (items.empty() && ui.on_progress) { + const auto n = static_cast(pit_to_upload->size()); + ui.on_progress(0, n, 0, n); + ui.on_progress(n, n, n, n); + } + if (ui.on_item_done) ui.on_item_done(0); + plan_off = 1; } - auto emit = [&](Step s) { - cur = s; - sync.arrive_and_wait(); - sync.arrive_and_wait(); - }; + for (std::size_t idx = 0; idx < items.size(); ++idx) { + if (failed_count.load(std::memory_order_relaxed) >= ndevs) break; - auto coordinator = [&]() -> brokkr::core::Status { - u64 overall_done = 0; + const auto& item = items[idx]; + const std::size_t plan_idx = plan_off + idx; - std::size_t plan_off = 0; - if (has_pit) { - if (ui.on_item_active) ui.on_item_active(0); - if (ui.on_item_done) ui.on_item_done(0); - plan_off = 1; - } + const std::string& file_name = + item.spec.source_basename.empty() ? item.spec.basename : item.spec.source_basename; + if (!file_name.empty()) spdlog::info("{}", file_name); - for (std::size_t idx = 0; idx < items.size(); ++idx) { - if (failed_count.load(std::memory_order_relaxed) >= ndevs) break; + if (ui.on_item_active) ui.on_item_active(plan_idx); - const auto& item = items[idx]; - const std::size_t plan_idx = plan_off + idx; + const u64 item_total = item.spec.size; + u64 item_done = 0; - if (ui.on_item_active) ui.on_item_active(plan_idx); + if (item.spec.lz4 && use_lz4) { + struct Slot { + std::vector stream; + u64 begin = 0, end = 0, rounded = 0; + bool last = false; + const std::byte* data() const { return stream.data(); } + }; - const u64 item_total = item.spec.size; - u64 item_done = 0; + BRK_TRYV(src0, item.spec.open()); + BRK_TRYV(reader, io::Lz4BlockStreamReader::open(std::move(src0))); - if (item.spec.lz4 && use_lz4) { - struct Slot { - std::vector stream; - u64 begin = 0, end = 0, rounded = 0; - bool last = false; - const std::byte* data() const { return stream.data(); } - }; + const u64 total_decomp = reader.content_size(); + if (!total_decomp) return brokkr::core::fail("LZ4 content size is zero: " + item.spec.display); - BRK_TRYV(src0, item.spec.open()); - BRK_TRYV(reader, io::Lz4BlockStreamReader::open(std::move(src0))); + const std::size_t max_blocks = detail::lz4_nonfinal_block_limit(cfg.buffer_bytes); + if (!max_blocks) + return brokkr::core::fail("buffer_bytes too small for compressed download (needs >= 1MiB)"); - const u64 total_decomp = reader.content_size(); - if (!total_decomp) return brokkr::core::fail("LZ4 content size is zero: " + item.spec.display); + u64 sent = 0; - const std::size_t max_blocks = detail::lz4_nonfinal_block_limit(cfg.buffer_bytes); - if (!max_blocks) - return brokkr::core::fail("buffer_bytes too small for compressed download (needs >= 1MiB)"); + brokkr::core::TwoSlotPrefetcher pf( + [&](Slot& s, std::stop_token stt) -> brokkr::core::Result { + if (stt.stop_requested() || sent >= total_decomp) return false; - u64 sent = 0; + const u64 rem = total_decomp - sent; + const bool last = rem <= static_cast(max_blocks) * detail::kOneMiB; + const u64 decomp_sz = last ? rem : static_cast(max_blocks) * detail::kOneMiB; - brokkr::core::TwoSlotPrefetcher pf( - [&](Slot& s, std::stop_token stt) -> brokkr::core::Result { - if (stt.stop_requested() || sent >= total_decomp) return false; + const std::size_t blocks = !last ? static_cast(decomp_sz / detail::kOneMiB) + : reader.blocks_remaining_1m(); - const u64 rem = total_decomp - sent; - const bool last = rem <= static_cast(max_blocks) * detail::kOneMiB; - const u64 decomp_sz = last ? rem : static_cast(max_blocks) * detail::kOneMiB; + s.stream.clear(); + s.stream.reserve(blocks * (static_cast(detail::kOneMiB) + 4)); - const std::size_t blocks = !last ? static_cast(decomp_sz / detail::kOneMiB) - : reader.blocks_remaining_1m(); + BRK_TRYV(comp, reader.read_n_blocks(blocks, s.stream)); + const u64 comp_sz = static_cast(comp); + const u64 rounded_sz = detail::round_up64(comp_sz, pkt); - s.stream.clear(); - s.stream.reserve(blocks * (static_cast(detail::kOneMiB) + 4)); + s.stream.resize(static_cast(rounded_sz), std::byte{0}); + s.begin = comp_sz; + s.end = decomp_sz; + s.rounded = rounded_sz; + s.last = last; - BRK_TRYV(comp, reader.read_n_blocks(blocks, s.stream)); - const u64 comp_sz = static_cast(comp); - const u64 rounded_sz = detail::round_up64(comp_sz, pkt); + sent += decomp_sz; + return true; + }, + [&](Slot& s) { s.stream.reserve(max_blocks * (static_cast(detail::kOneMiB) + 4)); }); - s.stream.resize(static_cast(rounded_sz), std::byte{0}); - s.begin = comp_sz; - s.end = decomp_sz; - s.rounded = rounded_sz; - s.last = last; + if (ui.on_progress) ui.on_progress(overall_done, total, item_done, item_total); - sent += decomp_sz; - return true; - }, - [&](Slot& s) { s.stream.reserve(max_blocks * (static_cast(detail::kOneMiB) + 4)); }); + BRK_TRY(send_prefetched(pf, sync, cur, pkt, true, item.part.id, item.part.dev_type, total, item_total, + overall_done, item_done, ui, failed_count, ndevs, [&](const Slot& w, u64 packets) { + return [end = w.end, packets](u64 p) { + const auto c1 = ((p + 1) * end) / packets; + const auto c0 = (p * end) / packets; + return c1 - c0; + }; + })); - if (ui.on_progress) ui.on_progress(overall_done, total, item_done, item_total); + } else { + struct Slot { + std::vector buf; + u64 begin = 0, end = 0; + std::size_t rounded = 0; + bool last = false; + const std::byte* data() const { return buf.data(); } + }; - BRK_TRY(send_prefetched(pf, sync, cur, pkt, true, item.part.id, item.part.dev_type, total, item_total, - overall_done, item_done, ui, failed_count, ndevs, [&](const Slot& w, u64 packets) { - return [end = w.end, packets](u64 p) { - const auto c1 = ((p + 1) * end) / packets; - const auto c0 = (p * end) / packets; - return c1 - c0; - }; - })); + std::unique_ptr src; + if (item.spec.lz4) { + BRK_TRYV(s0, item.spec.open()); + BRK_TRYV(d0, io::open_lz4_decompressed(std::move(s0))); + src = std::move(d0); } else { - struct Slot { - std::vector buf; - u64 begin = 0, end = 0; - std::size_t rounded = 0; - bool last = false; - const std::byte* data() const { return buf.data(); } - }; - - std::unique_ptr src; - - if (item.spec.lz4) { - BRK_TRYV(s0, item.spec.open()); - BRK_TRYV(d0, io::open_lz4_decompressed(std::move(s0))); - src = std::move(d0); - } else { - BRK_TRYV(s0, item.spec.open()); - src = std::move(s0); - } - - const u64 file_sz = src->size(); - if (!file_sz) return brokkr::core::fail("Empty source: " + item.spec.display); - - const std::size_t max_rounded = - static_cast(detail::round_up64(static_cast(cfg.buffer_bytes), static_cast(pkt))); - u64 sent = 0; - - brokkr::core::TwoSlotPrefetcher pf( - [&](Slot& s, std::stop_token stt) -> brokkr::core::Result { - if (stt.stop_requested() || sent >= file_sz) return false; - - const u64 rem = file_sz - sent; - const u64 actual = std::min(rem, cfg.buffer_bytes); - const u64 rounded_u64 = detail::round_up64(actual, pkt); - const auto rounded = static_cast(rounded_u64); - - s.buf.resize(rounded); - - auto rst = io::read_exact(*src, std::span(s.buf.data(), static_cast(actual))); - if (!rst) return brokkr::core::fail(std::move(rst.error())); - - if (rounded_u64 > actual) - std::memset(s.buf.data() + static_cast(actual), 0, - rounded - static_cast(actual)); - - s.rounded = rounded; - s.begin = rounded_u64; - s.end = actual; - s.last = (sent + actual >= file_sz); - - sent += actual; - return true; - }, - [&](Slot& s) { s.buf.reserve(max_rounded); }); - - if (ui.on_progress) ui.on_progress(overall_done, total, item_done, item_total); - - BRK_TRY(send_prefetched(pf, sync, cur, pkt, false, item.part.id, item.part.dev_type, total, item_total, - overall_done, item_done, ui, failed_count, ndevs, - [&](const Slot& w, u64 /*packets*/) { - u64 rem2 = w.end; - const u64 pkt64 = static_cast(pkt); - return [rem2, pkt64](u64 /*p*/) mutable { - const u64 add = std::min(pkt64, rem2); - rem2 -= add; - return add; - }; - })); + BRK_TRYV(s0, item.spec.open()); + src = std::move(s0); } - if (ui.on_item_done) ui.on_item_done(plan_idx); + const u64 file_sz = src->size(); + if (!file_sz) return brokkr::core::fail("Empty source: " + item.spec.display); + + const std::size_t max_rounded = + static_cast(detail::round_up64(static_cast(cfg.buffer_bytes), static_cast(pkt))); + u64 sent = 0; + + brokkr::core::TwoSlotPrefetcher pf( + [&](Slot& s, std::stop_token stt) -> brokkr::core::Result { + if (stt.stop_requested() || sent >= file_sz) return false; + + const u64 rem = file_sz - sent; + const u64 actual = std::min(rem, cfg.buffer_bytes); + const u64 rounded_u64 = detail::round_up64(actual, pkt); + const auto rounded = static_cast(rounded_u64); + + s.buf.resize(rounded); + + auto rst = io::read_exact(*src, std::span(s.buf.data(), static_cast(actual))); + if (!rst) return brokkr::core::fail(std::move(rst.error())); + + if (rounded_u64 > actual) + std::memset(s.buf.data() + static_cast(actual), 0, + rounded - static_cast(actual)); + + s.rounded = rounded; + s.begin = rounded_u64; + s.end = actual; + s.last = (sent + actual >= file_sz); + + sent += actual; + return true; + }, + [&](Slot& s) { s.buf.reserve(max_rounded); }); + + if (ui.on_progress) ui.on_progress(overall_done, total, item_done, item_total); + + BRK_TRY(send_prefetched(pf, sync, cur, pkt, false, item.part.id, item.part.dev_type, total, item_total, + overall_done, item_done, ui, failed_count, ndevs, + [&](const Slot& w, u64 /*packets*/) { + u64 rem2 = w.end; + const u64 pkt64 = static_cast(pkt); + return [rem2, pkt64](u64 /*p*/) mutable { + const u64 add = std::min(pkt64, rem2); + rem2 -= add; + return add; + }; + })); } - return {}; - }; - - auto cst = coordinator(); - if (!cst) berr.set(std::move(cst)); + if (ui.on_item_done) ui.on_item_done(plan_idx); + } - emit({.op = Step::Op::Quit}); - for (auto& t : workers) - if (t.joinable()) t.join(); + return {}; + }; - const std::size_t bad_in_flash = static_cast(failed_count.load(std::memory_order_relaxed)); - failed_total += bad_in_flash; - if (bad_in_flash) first_err.set(berr.status_or_ok()); + auto cst = coordinator(); + if (!cst) berr.set(std::move(cst)); - { - std::vector survivors; - std::vector survivors_idx; - survivors.reserve(active.size()); - survivors_idx.reserve(active.size()); + emit({.op = Step::Op::Quit}); + for (auto& t : workers) + if (t.joinable()) t.join(); - for (std::size_t i = 0; i < active.size(); ++i) - if (!dead[i]) { - survivors.push_back(active[i]); - survivors_idx.push_back(active_idx[i]); - } - active.swap(survivors); - active_idx.swap(survivors_idx); - } + const std::size_t bad_in_flash = static_cast(failed_count.load(std::memory_order_relaxed)); + failed_total += bad_in_flash; + if (bad_in_flash) first_err.set(berr.status_or_ok()); - return {}; - }); + { + std::vector survivors; + std::vector survivors_idx; + survivors.reserve(active.size()); + survivors_idx.reserve(active.size()); - steps.emplace_back([&] -> brokkr::core::Status { - if (!active.empty()) { - auto st = shutdown_active(sm_final, final_stage(sm_final)); - if (!st) { - first_err.set(st); - for (auto idx : active_idx) emit_devfail(ui, idx, st.error()); - failed_total += active.size(); - return st; + for (std::size_t i = 0; i < active.size(); ++i) + if (!dead[i]) { + survivors.push_back(active[i]); + survivors_idx.push_back(active_idx[i]); } + active.swap(survivors); + active_idx.swap(survivors_idx); + } + + return {}; + }); + + steps.emplace_back([&] -> brokkr::core::Status { + if (!active.empty()) { + auto st = shutdown_active(sm_final, final_stage(sm_final)); + if (!st) { + first_err.set(st); + for (auto idx : active_idx) emit_devfail(ui, idx, st.error()); + failed_total += active.size(); + return st; } - return {}; - }); - } + } + return {}; + }); for (auto& fn : steps) { auto st = fn(); - if (!st) return finish(st, intent != Intent::Flash); + if (!st) return finish(st, false); } - return finish({}, intent != Intent::Flash); + return finish({}, false); } } // namespace brokkr::odin diff --git a/src/protocol/odin/group_flasher.hpp b/src/protocol/odin/group_flasher.hpp index fc7caa5..06455a1 100644 --- a/src/protocol/odin/group_flasher.hpp +++ b/src/protocol/odin/group_flasher.hpp @@ -1,104 +1,103 @@ -/* - * Copyright (c) 2026 Gabriel2392 - * - * This program 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. - * - * This program 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 this program. If not, see . - */ - -#pragma once - -#include "core/byte_transport.hpp" -#include "core/status.hpp" -#include "platform/platform_all.hpp" -#include "protocol/odin/flash.hpp" -#include "protocol/odin/odin_cmd.hpp" -#include "protocol/odin/pit.hpp" - -#include -#include -#include -#include -#include -#include - -namespace brokkr::odin { - -struct UsbTarget { - std::string devnode; - brokkr::platform::UsbFsDevice dev; - brokkr::platform::UsbFsConnection conn; - - InitTargetInfo init{}; - ProtocolVersion proto = ProtocolVersion::PROTOCOL_NONE; - - std::vector pit_bytes{}; - pit::PitTable pit_table{}; - - explicit UsbTarget(std::string devnode_path) : devnode(std::move(devnode_path)), dev(devnode), conn(dev) {} -}; - -struct Target { - std::string id; - brokkr::core::IByteTransport* link = nullptr; - - InitTargetInfo init{}; - ProtocolVersion proto = ProtocolVersion::PROTOCOL_NONE; - - std::vector pit_bytes{}; - pit::PitTable pit_table{}; -}; - -struct PlanItem { - enum class Kind { Pit, Part }; - Kind kind = Kind::Part; - - std::int32_t part_id = -1; - std::int32_t dev_type = 0; - - std::string part_name, pit_file_name, source_base; - std::uint64_t size = 0; -}; - -struct Cfg { - std::size_t buffer_bytes = 30ull * 1024 * 1024; - std::size_t pkt_all_v2plus = 1ull * 1024 * 1024; - std::size_t pkt_any_old = 128ull * 1024; - - int preflash_timeout_ms = 1000; - unsigned preflash_retries = 2; - - int flash_timeout_ms = 45'000; - - bool reboot_after = true; - bool redownload_after = false; -}; - -struct Ui { - std::function&)> on_devices; - std::function on_model; - std::function on_stage; - - std::function&, std::uint64_t)> on_plan; - std::function on_item_active; - std::function on_item_done; - - std::function on_progress; - - std::function on_error; - std::function on_done; -}; - -brokkr::core::Status flash(std::vector& devs, const std::vector& sources, - std::shared_ptr> pit_to_upload, const Cfg& cfg, Ui ui) noexcept; - -} // namespace brokkr::odin +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#pragma once + +#include "core/byte_transport.hpp" +#include "core/status.hpp" +#include "platform/platform_all.hpp" +#include "protocol/odin/flash.hpp" +#include "protocol/odin/odin_cmd.hpp" +#include "protocol/odin/pit.hpp" + +#include +#include +#include +#include +#include +#include + +namespace brokkr::odin { + +struct UsbTarget { + std::string devnode; + brokkr::platform::UsbFsDevice dev; + brokkr::platform::UsbFsConnection conn; + + InitTargetInfo init{}; + ProtocolVersion proto = ProtocolVersion::PROTOCOL_NONE; + + std::vector pit_bytes{}; + pit::PitTable pit_table{}; + + explicit UsbTarget(std::string devnode_path) : devnode(std::move(devnode_path)), dev(devnode), conn(dev) {} +}; + +struct Target { + std::string id; + brokkr::core::IByteTransport* link = nullptr; + + InitTargetInfo init{}; + ProtocolVersion proto = ProtocolVersion::PROTOCOL_NONE; + + std::vector pit_bytes{}; + pit::PitTable pit_table{}; +}; + +struct PlanItem { + enum class Kind { Pit, Part }; + Kind kind = Kind::Part; + + std::int32_t part_id = -1; + std::int32_t dev_type = 0; + + std::string part_name, pit_file_name, source_base; + std::uint64_t size = 0; +}; + +struct Cfg { + std::size_t buffer_bytes = 30ull * 1024 * 1024; + std::size_t pkt_all_v2plus = 1ull * 1024 * 1024; + std::size_t pkt_any_old = 128ull * 1024; + + int preflash_timeout_ms = 1000; + unsigned preflash_retries = 2; + + int flash_timeout_ms = 45'000; + + bool reboot_after = true; +}; + +struct Ui { + std::function&)> on_devices; + std::function on_model; + std::function on_stage; + + std::function&, std::uint64_t)> on_plan; + std::function on_item_active; + std::function on_item_done; + + std::function on_progress; + + std::function on_error; + std::function on_done; +}; + +brokkr::core::Status flash(std::vector& devs, const std::vector& sources, + std::shared_ptr> pit_to_upload, const Cfg& cfg, Ui ui) noexcept; + +} // namespace brokkr::odin diff --git a/src/protocol/odin/odin_cmd.cpp b/src/protocol/odin/odin_cmd.cpp index 092359a..e04f679 100644 --- a/src/protocol/odin/odin_cmd.cpp +++ b/src/protocol/odin/odin_cmd.cpp @@ -1,361 +1,345 @@ -/* - * Copyright (c) 2026 Gabriel2392 - * - * This program 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. - * - * This program 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 this program. If not, see . - */ - -#include "protocol/odin/odin_cmd.hpp" -#include "protocol/odin/odin_wire.hpp" - -#include "core/bytes.hpp" - -#include -#include -#include -#include - -#include -#include - -namespace brokkr::odin { - -namespace { - -constexpr std::int32_t BOOTLOADER_FAIL = static_cast(0xffffffff); - -inline brokkr::core::Status require_connected(brokkr::core::IByteTransport& c) noexcept { - return c.connected() ? brokkr::core::Status{} : brokkr::core::fail("transport not connected"); -} - -inline brokkr::core::Status check_resp(std::int32_t expected_id, const ResponseBox& r, std::int32_t* out_ack) noexcept { - if (r.id == BOOTLOADER_FAIL) return brokkr::core::fail("Bootloader returned FAIL"); - if (r.id == std::numeric_limits::min()) return brokkr::core::fail("Invalid response id (INT_MIN)"); - if (r.id != expected_id) return brokkr::core::fail("Unexpected response id"); - if (out_ack) - *out_ack = r.ack; - else if (r.ack < 0) - return brokkr::core::failf("Operation failed ({})", r.ack); - return {}; -} - -static std::int32_t lo32(std::uint64_t v) { - return static_cast(static_cast(v & 0xFFFFFFFFull)); -} -static std::int32_t hi32(std::uint64_t v) { - return static_cast(static_cast((v >> 32) & 0xFFFFFFFFull)); -} - -static brokkr::core::Result require_i32_total(std::uint64_t v) noexcept { - constexpr std::uint64_t max = static_cast(std::numeric_limits::max()); - if (v > max) return brokkr::core::fail("TOTALSIZE exceeds ODIN int32 limit on protocol v0/v1"); - return static_cast(v); -} - -} // namespace - -brokkr::core::Status OdinCommands::send_raw(std::span data, unsigned retries) noexcept { - auto st = require_connected(conn_); - if (!st) return st; - - std::size_t off = 0; - while (off < data.size()) { - const int sent = conn_.send(brokkr::core::u8(data.subspan(off)), retries); - if (sent <= 0) return brokkr::core::fail("send failed"); - off += static_cast(sent); - } - return {}; -} - -brokkr::core::Status OdinCommands::recv_raw(std::span data, unsigned retries) noexcept { - auto st = require_connected(conn_); - if (!st) return st; - - std::size_t off = 0; - while (off < data.size()) { - const int got = conn_.recv(brokkr::core::u8(data.subspan(off)), retries); - if (got <= 0) return brokkr::core::fail("receive failed"); - off += static_cast(got); - } - return {}; -} - -brokkr::core::Status OdinCommands::send_request(const RequestBox& rq, unsigned retries) noexcept { - return send_raw(std::as_bytes(std::span{&rq, 1}), retries); -} - -brokkr::core::Result OdinCommands::recv_checked_response(std::int32_t expected_id, std::int32_t* out_ack, - unsigned retries) noexcept { - ResponseBox r{}; - auto st = recv_raw(std::as_writable_bytes(std::span{&r, 1}), retries); - if (!st) return brokkr::core::fail(std::move(st.error())); - - response_from_le(r); - - st = check_resp(expected_id, r, out_ack); - if (!st) return brokkr::core::fail(std::move(st.error())); - - return r; -} - -brokkr::core::Result OdinCommands::rpc_(RqtCommandType type, RqtCommandParam param, - std::span ints, - std::span chars, std::int32_t* out_ack, - unsigned retries) noexcept { - auto st = send_request(make_request(type, param, ints, chars), retries); - if (!st) return brokkr::core::fail(std::move(st.error())); - return recv_checked_response(static_cast(type), out_ack, retries); -} - -brokkr::core::Status OdinCommands::handshake(unsigned retries) noexcept { - auto st = require_connected(conn_); - if (!st) return st; - - if (conn_.kind() == brokkr::core::IByteTransport::Kind::UsbBulk) { - static constexpr std::array ping{std::byte{'O'}, std::byte{'D'}, std::byte{'I'}, std::byte{'N'}, - std::byte{0}}; - st = send_raw(ping, retries); - } else { - static constexpr std::array ping{std::byte{'O'}, std::byte{'D'}, std::byte{'I'}, std::byte{'N'}}; - st = send_raw(ping, retries); - } - if (!st) return st; - - constexpr std::string_view expected = "LOKE"; - std::array resp{}; - std::size_t have = 0; - - while (have < expected.size()) { - const int got = conn_.recv(brokkr::core::u8(std::span(resp.data() + have, resp.size() - have)), retries); - if (got <= 0) return brokkr::core::fail("Handshake receive failed"); - have += static_cast(got); - } - - if (std::memcmp(resp.data(), expected.data(), expected.size()) != 0) { - spdlog::error("Dump of handshake response ({} bytes):", have); - spdlog::error("{}", fmt::join(resp.begin(), resp.begin() + have, " ")); -#ifndef NDEBUG - std::array as_str{}; - for (std::size_t i = 0; i < have && i < as_str.size() - 1; ++i) { - const std::byte b = resp[i]; - as_str[i] = (b >= std::byte{32} && b <= std::byte{126}) ? static_cast(b) : '.'; - } - spdlog::error("Trying it as a string: {}", as_str.data()); -#endif - return brokkr::core::fail("Handshake failed (expected LOKE)"); - } - - spdlog::debug("ODIN handshake OK"); - return {}; -} - -brokkr::core::Result OdinCommands::get_version(unsigned retries) noexcept { - const std::int32_t ints[] = {static_cast(ProtocolVersion::PROTOCOL_VER5)}; - - std::int32_t ack_i32 = 0; - auto r = rpc_(RqtCommandType::RQT_INIT, RqtCommandParam::RQT_INIT_TARGET, ints, {}, &ack_i32, retries); - if (!r) return brokkr::core::fail(std::move(r.error())); - - InitTargetInfo out; - out.ack_word = static_cast(ack_i32); - spdlog::debug("ODIN target ack word: 0x{:08X} (protocol v{}, compressed download {})", out.ack_word, - static_cast(out.protocol()), out.supports_compressed_download()); - return out; -} - -brokkr::core::Status OdinCommands::setup_transfer_options(std::int32_t packet_size, unsigned retries) noexcept { - const std::int32_t ints[] = {packet_size}; - auto r = rpc_(RqtCommandType::RQT_INIT, RqtCommandParam::RQT_INIT_PACKETSIZE, ints, {}, nullptr, retries); - return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); -} - -brokkr::core::Status OdinCommands::send_total_size(std::uint64_t total_size, ProtocolVersion proto, - unsigned retries) noexcept { - if (proto <= ProtocolVersion::PROTOCOL_VER1) { - auto v = require_i32_total(total_size); - if (!v) return brokkr::core::fail(std::move(v.error())); - const std::int32_t ints[] = {*v}; - auto r = rpc_(RqtCommandType::RQT_INIT, RqtCommandParam::RQT_INIT_TOTALSIZE, ints, {}, nullptr, retries); - return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); - } - - const std::int32_t ints[] = {lo32(total_size), hi32(total_size)}; - auto r = rpc_(RqtCommandType::RQT_INIT, RqtCommandParam::RQT_INIT_TOTALSIZE, ints, {}, nullptr, retries); - return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); -} - -brokkr::core::Result OdinCommands::get_pit_size(unsigned retries) noexcept { - std::int32_t pitSize = 0; - auto r = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_GET, {}, {}, &pitSize, retries); - if (!r) return brokkr::core::fail(std::move(r.error())); - return pitSize; -} - -brokkr::core::Status OdinCommands::get_pit(std::span out, unsigned retries) noexcept { - constexpr std::size_t PIT_TRANSMIT_UNIT = 500; - if (out.empty()) return brokkr::core::fail("PIT output buffer empty"); - - const std::size_t pitSize = out.size(); - const std::size_t parts = ((pitSize - 1) / PIT_TRANSMIT_UNIT) + 1; - - for (std::size_t idx = 0; idx < parts; ++idx) { - const std::int32_t pitIndex = static_cast(idx); - - auto st = send_request( - make_request(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_START, std::span{&pitIndex, 1}), retries); - if (!st) return st; - - const std::size_t sizeToDownload = std::min(PIT_TRANSMIT_UNIT, pitSize - (PIT_TRANSMIT_UNIT * idx)); - const std::size_t off = idx * PIT_TRANSMIT_UNIT; - - st = recv_raw(out.subspan(off, sizeToDownload), retries); - if (!st) return st; - } - - (void)conn_.recv_zlp(); - auto r = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_COMPLETE, {}, {}, nullptr, retries); - return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); -} - -brokkr::core::Status OdinCommands::set_pit(std::span pit, unsigned retries) noexcept { - if (pit.empty()) return brokkr::core::fail("PIT buffer empty"); - if (pit.size() > static_cast(std::numeric_limits::max())) - return brokkr::core::fail("PIT too large for ODIN int32"); - - auto r1 = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_SET, {}, {}, nullptr, retries); - if (!r1) return brokkr::core::fail(std::move(r1.error())); - - const auto pitSize32 = static_cast(pit.size()); - auto r2 = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_START, std::span{&pitSize32, 1}, {}, nullptr, - retries); - if (!r2) return brokkr::core::fail(std::move(r2.error())); - - auto st = send_raw(pit, retries); - if (!st) return st; - - ResponseBox ack{}; - st = recv_raw(std::as_writable_bytes(std::span{&ack, 1}), retries); - if (!st) return st; - - response_from_le(ack); - - auto r3 = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_COMPLETE, std::span{&pitSize32, 1}, {}, nullptr, - retries); - return r3 ? brokkr::core::Status{} : brokkr::core::fail(std::move(r3.error())); -} - -brokkr::core::Status OdinCommands::begin_download(std::int32_t rounded_total_size, unsigned retries) noexcept { - auto r1 = rpc_(RqtCommandType::RQT_XMIT, RqtCommandParam::RQT_XMIT_DOWNLOAD, {}, {}, nullptr, retries); - if (!r1) return brokkr::core::fail(std::move(r1.error())); - auto r2 = rpc_(RqtCommandType::RQT_XMIT, RqtCommandParam::RQT_XMIT_START, std::span{&rounded_total_size, 1}, {}, - nullptr, retries); - return r2 ? brokkr::core::Status{} : brokkr::core::fail(std::move(r2.error())); -} - -brokkr::core::Status OdinCommands::begin_download_compressed(std::int32_t comp_size, unsigned retries) noexcept { - auto r1 = rpc_(RqtCommandType::RQT_XMIT, RqtCommandParam::RQT_XMIT_COMPRESSED_DOWNLOAD, {}, {}, nullptr, retries); - if (!r1) return brokkr::core::fail(std::move(r1.error())); - auto r2 = rpc_(RqtCommandType::RQT_XMIT, RqtCommandParam::RQT_XMIT_COMPRESSED_START, std::span{&comp_size, 1}, {}, - nullptr, retries); - return r2 ? brokkr::core::Status{} : brokkr::core::fail(std::move(r2.error())); -} - -brokkr::core::Status OdinCommands::end_download_impl_(RqtCommandParam complete_param, std::int32_t size_to_flash, - std::int32_t part_id, std::int32_t dev_type, bool is_last, - std::int32_t bin_type, bool efs_clear, bool boot_update, - unsigned retries) noexcept { - std::int32_t data[8]{}; - data[0] = 0; - data[1] = size_to_flash; - data[2] = bin_type; - data[3] = dev_type; - data[4] = part_id; - data[5] = is_last ? 1 : 0; - data[6] = efs_clear ? 1 : 0; - data[7] = boot_update ? 1 : 0; - - auto r = rpc_(RqtCommandType::RQT_XMIT, complete_param, data, {}, nullptr, retries); - return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); -} - -brokkr::core::Status OdinCommands::end_download(std::int32_t size_to_flash, std::int32_t part_id, std::int32_t dev_type, - bool is_last, std::int32_t bin_type, bool efs_clear, bool boot_update, - unsigned retries) noexcept { - return end_download_impl_(RqtCommandParam::RQT_XMIT_COMPLETE, size_to_flash, part_id, dev_type, is_last, bin_type, - efs_clear, boot_update, retries); -} - -brokkr::core::Status OdinCommands::end_download_compressed(std::int32_t decomp_size_to_flash, std::int32_t part_id, - std::int32_t dev_type, bool is_last, std::int32_t bin_type, - bool efs_clear, bool boot_update, - unsigned retries) noexcept { - return end_download_impl_(RqtCommandParam::RQT_XMIT_COMPRESSED_COMPLETE, decomp_size_to_flash, part_id, dev_type, - is_last, bin_type, efs_clear, boot_update, retries); -} - -brokkr::core::Status OdinCommands::shutdown(ShutdownMode mode, unsigned retries) noexcept { - auto st = require_connected(conn_); - if (!st) return st; - - auto _close_cmd = [&](RqtCommandParam p, const char* name) -> brokkr::core::Status { - auto r = rpc_(RqtCommandType::RQT_CLOSE, p, {}, {}, nullptr, retries); - if (!r) { - spdlog::error("Failed to send shutdown command {}: {}", name, r.error()); - } else { - spdlog::debug("Sent shutdown command {}", name); - } - return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); - }; -#define close_cmd(param) _close_cmd(RqtCommandParam::param, #param) - - if (mode == ShutdownMode::NoReboot) { - return close_cmd(RQT_CLOSE_END); - } - if (mode == ShutdownMode::Reboot) { - st = close_cmd(RQT_CLOSE_END); - if (!st) return st; - auto reboot_st = close_cmd(RQT_CLOSE_REBOOT); - if (!reboot_st) - spdlog::warn("Reboot command failed (device likely already rebooting): {}", reboot_st.error()); - return {}; - } - - st = close_cmd(RQT_CLOSE_REDOWNLOAD); - if (!st) return st; - - static constexpr std::string_view kAutoTest = "@#AuToTEstRst@#"; - std::array msg{}; - for (std::size_t i = 0; i < kAutoTest.size(); ++i) msg[i] = static_cast(kAutoTest[i]); - - st = send_raw({msg.data(), msg.size()}, retries); - if (!st) { - spdlog::error("Failed to send re-download trigger message: {}", st.error()); - return st; - } - - const int old_to = conn_.timeout_ms(); - conn_.set_timeout_ms(500); - std::array tmp{}; - if (conn_.recv({tmp.data(), tmp.size()}, 0) < 0) { - spdlog::debug("No data received after re-download trigger, as expected"); - } else { - spdlog::error("Received unexpected data after re-download trigger, something may have gone wrong"); -#ifndef NDEBUG - spdlog::error("Data ({} bytes): {}", tmp.size(), fmt::join(tmp.begin(), tmp.end(), " ")); -#endif - } - conn_.set_timeout_ms(old_to); - return {}; -} - -} // namespace brokkr::odin +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#include "protocol/odin/odin_cmd.hpp" +#include "protocol/odin/odin_wire.hpp" + +#include "core/bytes.hpp" + +#include +#include +#include +#include + +#include +#include + +namespace brokkr::odin { + +namespace { + +constexpr std::int32_t BOOTLOADER_FAIL = static_cast(0xffffffff); + +inline brokkr::core::Status require_connected(brokkr::core::IByteTransport& c) noexcept { + return c.connected() ? brokkr::core::Status{} : brokkr::core::fail("transport not connected"); +} + +inline brokkr::core::Status check_resp(std::int32_t expected_id, const ResponseBox& r, std::int32_t* out_ack) noexcept { + if (r.id == BOOTLOADER_FAIL) return brokkr::core::fail("Bootloader returned FAIL"); + if (r.id == std::numeric_limits::min()) return brokkr::core::fail("Invalid response id (INT_MIN)"); + if (r.id != expected_id) return brokkr::core::fail("Unexpected response id"); + if (out_ack) + *out_ack = r.ack; + else if (r.ack < 0) + return brokkr::core::failf("Operation failed ({})", r.ack); + return {}; +} + +static std::int32_t lo32(std::uint64_t v) { + return static_cast(static_cast(v & 0xFFFFFFFFull)); +} +static std::int32_t hi32(std::uint64_t v) { + return static_cast(static_cast((v >> 32) & 0xFFFFFFFFull)); +} + +static brokkr::core::Result require_i32_total(std::uint64_t v) noexcept { + constexpr std::uint64_t max = static_cast(std::numeric_limits::max()); + if (v > max) return brokkr::core::fail("TOTALSIZE exceeds ODIN int32 limit on protocol v0/v1"); + return static_cast(v); +} + +} // namespace + +brokkr::core::Status OdinCommands::send_raw(std::span data, unsigned retries) noexcept { + auto st = require_connected(conn_); + if (!st) return st; + + std::size_t off = 0; + while (off < data.size()) { + const int sent = conn_.send(brokkr::core::u8(data.subspan(off)), retries); + if (sent <= 0) return brokkr::core::fail("send failed"); + off += static_cast(sent); + } + return {}; +} + +brokkr::core::Status OdinCommands::recv_raw(std::span data, unsigned retries) noexcept { + auto st = require_connected(conn_); + if (!st) return st; + + std::size_t off = 0; + while (off < data.size()) { + const int got = conn_.recv(brokkr::core::u8(data.subspan(off)), retries); + if (got <= 0) return brokkr::core::fail("receive failed"); + off += static_cast(got); + } + return {}; +} + +brokkr::core::Status OdinCommands::send_request(const RequestBox& rq, unsigned retries) noexcept { + return send_raw(std::as_bytes(std::span{&rq, 1}), retries); +} + +brokkr::core::Result OdinCommands::recv_checked_response(std::int32_t expected_id, std::int32_t* out_ack, + unsigned retries) noexcept { + ResponseBox r{}; + auto st = recv_raw(std::as_writable_bytes(std::span{&r, 1}), retries); + if (!st) return brokkr::core::fail(std::move(st.error())); + + response_from_le(r); + + st = check_resp(expected_id, r, out_ack); + if (!st) return brokkr::core::fail(std::move(st.error())); + + return r; +} + +brokkr::core::Result OdinCommands::rpc_(RqtCommandType type, RqtCommandParam param, + std::span ints, + std::span chars, std::int32_t* out_ack, + unsigned retries) noexcept { + auto st = send_request(make_request(type, param, ints, chars), retries); + if (!st) return brokkr::core::fail(std::move(st.error())); + return recv_checked_response(static_cast(type), out_ack, retries); +} + +brokkr::core::Status OdinCommands::handshake(unsigned retries) noexcept { + auto st = require_connected(conn_); + if (!st) return st; + + if (conn_.kind() == brokkr::core::IByteTransport::Kind::UsbBulk) { + static constexpr std::array ping{std::byte{'O'}, std::byte{'D'}, std::byte{'I'}, std::byte{'N'}, + std::byte{0}}; + st = send_raw(ping, retries); + } else { + static constexpr std::array ping{std::byte{'O'}, std::byte{'D'}, std::byte{'I'}, std::byte{'N'}}; + st = send_raw(ping, retries); + } + if (!st) return st; + + constexpr std::string_view expected = "LOKE"; + std::array resp{}; + std::size_t have = 0; + + while (have < expected.size()) { + const int got = conn_.recv(brokkr::core::u8(std::span(resp.data() + have, resp.size() - have)), retries); + if (got <= 0) return brokkr::core::fail("Handshake receive failed"); + have += static_cast(got); + } + + if (std::memcmp(resp.data(), expected.data(), expected.size()) != 0) { + spdlog::error("Dump of handshake response ({} bytes):", have); + spdlog::error("{}", fmt::join(resp.begin(), resp.begin() + have, " ")); +#ifndef NDEBUG + std::array as_str{}; + for (std::size_t i = 0; i < have && i < as_str.size() - 1; ++i) { + const std::byte b = resp[i]; + as_str[i] = (b >= std::byte{32} && b <= std::byte{126}) ? static_cast(b) : '.'; + } + spdlog::error("Trying it as a string: {}", as_str.data()); +#endif + return brokkr::core::fail("Handshake failed (expected LOKE)"); + } + + spdlog::debug("ODIN handshake OK"); + return {}; +} + +brokkr::core::Result OdinCommands::get_version(unsigned retries) noexcept { + const std::int32_t ints[] = {static_cast(ProtocolVersion::PROTOCOL_VER5)}; + + std::int32_t ack_i32 = 0; + auto r = rpc_(RqtCommandType::RQT_INIT, RqtCommandParam::RQT_INIT_TARGET, ints, {}, &ack_i32, retries); + if (!r) return brokkr::core::fail(std::move(r.error())); + + InitTargetInfo out; + out.ack_word = static_cast(ack_i32); + spdlog::debug("ODIN target ack word: 0x{:08X} (protocol v{}, compressed download {})", out.ack_word, + static_cast(out.protocol()), out.supports_compressed_download()); + return out; +} + +brokkr::core::Status OdinCommands::setup_transfer_options(std::int32_t packet_size, unsigned retries) noexcept { + const std::int32_t ints[] = {packet_size}; + auto r = rpc_(RqtCommandType::RQT_INIT, RqtCommandParam::RQT_INIT_PACKETSIZE, ints, {}, nullptr, retries); + if (!r) return brokkr::core::fail(std::move(r.error())); + + if (packet_size > 0) + conn_.set_packet_size_hint(static_cast(static_cast(packet_size))); + + return {}; +} + +brokkr::core::Status OdinCommands::send_total_size(std::uint64_t total_size, ProtocolVersion proto, + unsigned retries) noexcept { + if (proto <= ProtocolVersion::PROTOCOL_VER1) { + auto v = require_i32_total(total_size); + if (!v) return brokkr::core::fail(std::move(v.error())); + const std::int32_t ints[] = {*v}; + auto r = rpc_(RqtCommandType::RQT_INIT, RqtCommandParam::RQT_INIT_TOTALSIZE, ints, {}, nullptr, retries); + return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); + } + + const std::int32_t ints[] = {lo32(total_size), hi32(total_size)}; + auto r = rpc_(RqtCommandType::RQT_INIT, RqtCommandParam::RQT_INIT_TOTALSIZE, ints, {}, nullptr, retries); + return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); +} + +brokkr::core::Result OdinCommands::get_pit_size(unsigned retries) noexcept { + std::int32_t pitSize = 0; + auto r = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_GET, {}, {}, &pitSize, retries); + if (!r) return brokkr::core::fail(std::move(r.error())); + return pitSize; +} + +brokkr::core::Status OdinCommands::get_pit(std::span out, unsigned retries) noexcept { + constexpr std::size_t PIT_TRANSMIT_UNIT = 500; + if (out.empty()) return brokkr::core::fail("PIT output buffer empty"); + + const std::size_t pitSize = out.size(); + const std::size_t parts = ((pitSize - 1) / PIT_TRANSMIT_UNIT) + 1; + + for (std::size_t idx = 0; idx < parts; ++idx) { + const std::int32_t pitIndex = static_cast(idx); + + auto st = send_request( + make_request(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_START, std::span{&pitIndex, 1}), retries); + if (!st) return st; + + const std::size_t sizeToDownload = std::min(PIT_TRANSMIT_UNIT, pitSize - (PIT_TRANSMIT_UNIT * idx)); + const std::size_t off = idx * PIT_TRANSMIT_UNIT; + + st = recv_raw(out.subspan(off, sizeToDownload), retries); + if (!st) return st; + } + + (void)conn_.recv_zlp(); + auto r = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_COMPLETE, {}, {}, nullptr, retries); + return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); +} + +brokkr::core::Status OdinCommands::set_pit(std::span pit, unsigned retries) noexcept { + if (pit.empty()) return brokkr::core::fail("PIT buffer empty"); + if (pit.size() > static_cast(std::numeric_limits::max())) + return brokkr::core::fail("PIT too large for ODIN int32"); + + auto r1 = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_SET, {}, {}, nullptr, retries); + if (!r1) return brokkr::core::fail(std::move(r1.error())); + + const auto pitSize32 = static_cast(pit.size()); + auto r2 = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_START, std::span{&pitSize32, 1}, {}, nullptr, + retries); + if (!r2) return brokkr::core::fail(std::move(r2.error())); + + auto st = send_raw(pit, retries); + if (!st) return st; + + ResponseBox ack{}; + st = recv_raw(std::as_writable_bytes(std::span{&ack, 1}), retries); + if (!st) return st; + + response_from_le(ack); + + auto r3 = rpc_(RqtCommandType::RQT_PIT, RqtCommandParam::RQT_PIT_COMPLETE, std::span{&pitSize32, 1}, {}, nullptr, + retries); + return r3 ? brokkr::core::Status{} : brokkr::core::fail(std::move(r3.error())); +} + +brokkr::core::Status OdinCommands::begin_download(std::int32_t rounded_total_size, unsigned retries) noexcept { + auto r1 = rpc_(RqtCommandType::RQT_XMIT, RqtCommandParam::RQT_XMIT_DOWNLOAD, {}, {}, nullptr, retries); + if (!r1) return brokkr::core::fail(std::move(r1.error())); + auto r2 = rpc_(RqtCommandType::RQT_XMIT, RqtCommandParam::RQT_XMIT_START, std::span{&rounded_total_size, 1}, {}, + nullptr, retries); + return r2 ? brokkr::core::Status{} : brokkr::core::fail(std::move(r2.error())); +} + +brokkr::core::Status OdinCommands::begin_download_compressed(std::int32_t comp_size, unsigned retries) noexcept { + auto r1 = rpc_(RqtCommandType::RQT_XMIT, RqtCommandParam::RQT_XMIT_COMPRESSED_DOWNLOAD, {}, {}, nullptr, retries); + if (!r1) return brokkr::core::fail(std::move(r1.error())); + auto r2 = rpc_(RqtCommandType::RQT_XMIT, RqtCommandParam::RQT_XMIT_COMPRESSED_START, std::span{&comp_size, 1}, {}, + nullptr, retries); + return r2 ? brokkr::core::Status{} : brokkr::core::fail(std::move(r2.error())); +} + +brokkr::core::Status OdinCommands::end_download_impl_(RqtCommandParam complete_param, std::int32_t size_to_flash, + std::int32_t part_id, std::int32_t dev_type, bool is_last, + std::int32_t bin_type, bool efs_clear, bool boot_update, + unsigned retries) noexcept { + std::int32_t data[8]{}; + data[0] = 0; + data[1] = size_to_flash; + data[2] = bin_type; + data[3] = dev_type; + data[4] = part_id; + data[5] = is_last ? 1 : 0; + data[6] = efs_clear ? 1 : 0; + data[7] = boot_update ? 1 : 0; + + auto r = rpc_(RqtCommandType::RQT_XMIT, complete_param, data, {}, nullptr, retries); + return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); +} + +brokkr::core::Status OdinCommands::end_download(std::int32_t size_to_flash, std::int32_t part_id, std::int32_t dev_type, + bool is_last, std::int32_t bin_type, bool efs_clear, bool boot_update, + unsigned retries) noexcept { + return end_download_impl_(RqtCommandParam::RQT_XMIT_COMPLETE, size_to_flash, part_id, dev_type, is_last, bin_type, + efs_clear, boot_update, retries); +} + +brokkr::core::Status OdinCommands::end_download_compressed(std::int32_t decomp_size_to_flash, std::int32_t part_id, + std::int32_t dev_type, bool is_last, std::int32_t bin_type, + bool efs_clear, bool boot_update, + unsigned retries) noexcept { + return end_download_impl_(RqtCommandParam::RQT_XMIT_COMPRESSED_COMPLETE, decomp_size_to_flash, part_id, dev_type, + is_last, bin_type, efs_clear, boot_update, retries); +} + +brokkr::core::Status OdinCommands::shutdown(ShutdownMode mode, unsigned retries) noexcept { + auto st = require_connected(conn_); + if (!st) return st; + + auto _close_cmd = [&](RqtCommandParam p, const char* name) -> brokkr::core::Status { + auto r = rpc_(RqtCommandType::RQT_CLOSE, p, {}, {}, nullptr, retries); + if (!r) { + if (p == RqtCommandParam::RQT_CLOSE_REBOOT) { + spdlog::debug("Failed to send shutdown command {}: {}", name, r.error()); + } else { + spdlog::error("Failed to send shutdown command {}: {}", name, r.error()); + } + } else { + spdlog::debug("Sent shutdown command {}", name); + } + return r ? brokkr::core::Status{} : brokkr::core::fail(std::move(r.error())); + }; +#define close_cmd(param) _close_cmd(RqtCommandParam::param, #param) + + if (mode == ShutdownMode::NoReboot) { + return close_cmd(RQT_CLOSE_END); + } + if (mode == ShutdownMode::Reboot) { + st = close_cmd(RQT_CLOSE_END); + if (!st) return st; + auto reboot_st = close_cmd(RQT_CLOSE_REBOOT); + if (!reboot_st) + spdlog::debug("Reboot command failed (device likely already rebooting): {}", reboot_st.error()); + return {}; + } + + return brokkr::core::fail("Invalid shutdown mode"); +} + +} // namespace brokkr::odin diff --git a/src/protocol/odin/odin_cmd.hpp b/src/protocol/odin/odin_cmd.hpp index 559dd6c..a478e31 100644 --- a/src/protocol/odin/odin_cmd.hpp +++ b/src/protocol/odin/odin_cmd.hpp @@ -1,102 +1,102 @@ -/* - * Copyright (c) 2026 Gabriel2392 - * - * This program 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. - * - * This program 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 this program. If not, see . - */ - -#pragma once - -#include "core/byte_transport.hpp" -#include "core/status.hpp" -#include "protocol/odin/odin_wire.hpp" - -#include -#include -#include -#include - -namespace brokkr::odin { - -struct InitTargetInfo { - std::uint32_t ack_word = 0; - - std::uint16_t proto_raw() const noexcept { return static_cast((ack_word >> 16) & 0xFFFFu); } - - ProtocolVersion protocol() const noexcept { - const auto p = proto_raw(); - if (p == 0) return ProtocolVersion::PROTOCOL_VER1; - return static_cast(static_cast(p)); - } - - bool supports_compressed_download() const noexcept { return (ack_word & 0x8000u) != 0; } -}; - -class OdinCommands { - public: - enum class ShutdownMode { NoReboot, Reboot, ReDownload }; - - explicit OdinCommands(brokkr::core::IByteTransport& c) : conn_(c) {} - - brokkr::core::Status handshake(unsigned retries = 8) noexcept; - brokkr::core::Result get_version(unsigned retries = 8) noexcept; - - brokkr::core::Status setup_transfer_options(std::int32_t packet_size, unsigned retries = 8) noexcept; - - brokkr::core::Status send_total_size(std::uint64_t total_size, ProtocolVersion proto, unsigned retries = 8) noexcept; - - brokkr::core::Result get_pit_size(unsigned retries = 8) noexcept; - brokkr::core::Status get_pit(std::span out, unsigned retries = 8) noexcept; - brokkr::core::Status set_pit(std::span pit, unsigned retries = 8) noexcept; - - brokkr::core::Status begin_download(std::int32_t rounded_total_size, unsigned retries = 8) noexcept; - brokkr::core::Status begin_download_compressed(std::int32_t comp_size, unsigned retries = 8) noexcept; - - brokkr::core::Status end_download(std::int32_t size_to_flash, std::int32_t part_id, std::int32_t dev_type, - bool is_last, std::int32_t bin_type = 0, bool efs_clear = false, - bool boot_update = false, unsigned retries = 8) noexcept; - - brokkr::core::Status end_download_compressed(std::int32_t decomp_size_to_flash, std::int32_t part_id, - std::int32_t dev_type, bool is_last, std::int32_t bin_type = 0, - bool efs_clear = false, bool boot_update = false, - unsigned retries = 8) noexcept; - - brokkr::core::Status shutdown(ShutdownMode mode, unsigned retries = 8) noexcept; - brokkr::core::Status shutdown(bool reboot, unsigned retries = 8) noexcept { - return shutdown(reboot ? ShutdownMode::Reboot : ShutdownMode::NoReboot, retries); - } - - brokkr::core::Status send_raw(std::span data, unsigned retries = 8) noexcept; - brokkr::core::Status recv_raw(std::span data, unsigned retries = 8) noexcept; - - brokkr::core::Result recv_checked_response(std::int32_t expected_id, std::int32_t* out_ack = nullptr, - unsigned retries = 8) noexcept; - - brokkr::core::Status send_request(const RequestBox& rq, unsigned retries = 8) noexcept; - - private: - brokkr::core::Result rpc_(RqtCommandType type, RqtCommandParam param, - std::span ints = {}, - std::span chars = {}, std::int32_t* out_ack = nullptr, - unsigned retries = 8) noexcept; - - brokkr::core::Status end_download_impl_(RqtCommandParam complete_param, std::int32_t size_to_flash, - std::int32_t part_id, std::int32_t dev_type, bool is_last, - std::int32_t bin_type, bool efs_clear, bool boot_update, - unsigned retries) noexcept; - - private: - brokkr::core::IByteTransport& conn_; -}; - -} // namespace brokkr::odin +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#pragma once + +#include "core/byte_transport.hpp" +#include "core/status.hpp" +#include "protocol/odin/odin_wire.hpp" + +#include +#include +#include +#include + +namespace brokkr::odin { + +struct InitTargetInfo { + std::uint32_t ack_word = 0; + + std::uint16_t proto_raw() const noexcept { return static_cast((ack_word >> 16) & 0xFFFFu); } + + ProtocolVersion protocol() const noexcept { + const auto p = proto_raw(); + if (p == 0) return ProtocolVersion::PROTOCOL_VER1; + return static_cast(static_cast(p)); + } + + bool supports_compressed_download() const noexcept { return (ack_word & 0x8000u) != 0; } +}; + +class OdinCommands { + public: + enum class ShutdownMode { NoReboot, Reboot }; + + explicit OdinCommands(brokkr::core::IByteTransport& c) : conn_(c) {} + + brokkr::core::Status handshake(unsigned retries = 8) noexcept; + brokkr::core::Result get_version(unsigned retries = 8) noexcept; + + brokkr::core::Status setup_transfer_options(std::int32_t packet_size, unsigned retries = 8) noexcept; + + brokkr::core::Status send_total_size(std::uint64_t total_size, ProtocolVersion proto, unsigned retries = 8) noexcept; + + brokkr::core::Result get_pit_size(unsigned retries = 8) noexcept; + brokkr::core::Status get_pit(std::span out, unsigned retries = 8) noexcept; + brokkr::core::Status set_pit(std::span pit, unsigned retries = 8) noexcept; + + brokkr::core::Status begin_download(std::int32_t rounded_total_size, unsigned retries = 8) noexcept; + brokkr::core::Status begin_download_compressed(std::int32_t comp_size, unsigned retries = 8) noexcept; + + brokkr::core::Status end_download(std::int32_t size_to_flash, std::int32_t part_id, std::int32_t dev_type, + bool is_last, std::int32_t bin_type = 0, bool efs_clear = false, + bool boot_update = false, unsigned retries = 8) noexcept; + + brokkr::core::Status end_download_compressed(std::int32_t decomp_size_to_flash, std::int32_t part_id, + std::int32_t dev_type, bool is_last, std::int32_t bin_type = 0, + bool efs_clear = false, bool boot_update = false, + unsigned retries = 8) noexcept; + + brokkr::core::Status shutdown(ShutdownMode mode, unsigned retries = 8) noexcept; + brokkr::core::Status shutdown(bool reboot, unsigned retries = 8) noexcept { + return shutdown(reboot ? ShutdownMode::Reboot : ShutdownMode::NoReboot, retries); + } + + brokkr::core::Status send_raw(std::span data, unsigned retries = 8) noexcept; + brokkr::core::Status recv_raw(std::span data, unsigned retries = 8) noexcept; + + brokkr::core::Result recv_checked_response(std::int32_t expected_id, std::int32_t* out_ack = nullptr, + unsigned retries = 8) noexcept; + + brokkr::core::Status send_request(const RequestBox& rq, unsigned retries = 8) noexcept; + + private: + brokkr::core::Result rpc_(RqtCommandType type, RqtCommandParam param, + std::span ints = {}, + std::span chars = {}, std::int32_t* out_ack = nullptr, + unsigned retries = 8) noexcept; + + brokkr::core::Status end_download_impl_(RqtCommandParam complete_param, std::int32_t size_to_flash, + std::int32_t part_id, std::int32_t dev_type, bool is_last, + std::int32_t bin_type, bool efs_clear, bool boot_update, + unsigned retries) noexcept; + + private: + brokkr::core::IByteTransport& conn_; +}; + +} // namespace brokkr::odin diff --git a/src/protocol/odin/odin_wire.hpp b/src/protocol/odin/odin_wire.hpp index 53c5cb2..99c4961 100644 --- a/src/protocol/odin/odin_wire.hpp +++ b/src/protocol/odin/odin_wire.hpp @@ -1,134 +1,133 @@ -/* - * Copyright (c) 2026 Gabriel2392 - * - * This program 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. - * - * This program 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 this program. If not, see . - */ - -#pragma once - -#include "core/endian.hpp" - -#include -#include -#include -#include -#include -#include -#include - -namespace brokkr::odin { - -enum class RqtCommandType : std::int32_t { - RQT_INIT = 100, - RQT_PIT = 101, - RQT_XMIT = 102, - RQT_CLOSE = 103, - RQT_EMPTY = 0 -}; - -enum class RqtCommandParam : std::int32_t { - // INIT - RQT_INIT_TARGET = 0, - RQT_INIT_RESETTIME = 1, - RQT_INIT_TOTALSIZE = 2, - RQT_INIT_OEMSTATE = 3, - RQT_INIT_NOOEMSTATE = 4, - RQT_INIT_PACKETSIZE = 5, - RQT_INIT_XMIT_SIZE = 6, - - // PIT - RQT_PIT_SET = 0, - RQT_PIT_GET = 1, - RQT_PIT_START = 2, - RQT_PIT_COMPLETE = 3, - - // XMIT (uncompressed) - RQT_XMIT_DOWNLOAD = 0, - RQT_XMIT_DUMP = 1, - RQT_XMIT_START = 2, - RQT_XMIT_COMPLETE = 3, - RQT_XMIT_SMD = 4, - - // XMIT (compressed) - RQT_XMIT_COMPRESSED_DOWNLOAD = 5, - RQT_XMIT_COMPRESSED_START = 6, - RQT_XMIT_COMPRESSED_COMPLETE = 7, - - // CLOSE - RQT_CLOSE_END = 0, - RQT_CLOSE_REBOOT = 1, - RQT_CLOSE_DISCONNECT = 2, - RQT_CLOSE_REBOOT_RECOVERY = 3, - RQT_CLOSE_REDOWNLOAD = 4, -}; - -enum class ProtocolVersion : std::int16_t { - PROTOCOL_NONE = 0, - PROTOCOL_VER1 = 1, - PROTOCOL_VER2 = 2, - PROTOCOL_VER3 = 3, - PROTOCOL_VER4 = 4, - PROTOCOL_VER5 = 5, -}; - -#pragma pack(push, 1) -struct ResponseBox { - std::int32_t id; - std::int32_t ack; -}; -#pragma pack(pop) -static_assert(sizeof(ResponseBox) == 8); - -#pragma pack(push, 1) -struct RequestBox { - static constexpr std::size_t DATA_INT_SIZE = 9; - static constexpr std::size_t DATA_CHAR_SIZE = 128; - static constexpr std::size_t MD5_SIZE = 32; - - std::int32_t id; - std::int32_t data; - std::int32_t intData[DATA_INT_SIZE]; - std::int8_t charData[DATA_CHAR_SIZE]; - std::int8_t md5[MD5_SIZE]; - - std::int8_t dummy[1024 - (2 * 4 + DATA_INT_SIZE * 4 + DATA_CHAR_SIZE + MD5_SIZE)]; -}; -#pragma pack(pop) -static_assert(sizeof(RequestBox) == 1024); - -inline void response_from_le(ResponseBox& r) noexcept { - r.id = brokkr::core::le_to_host(r.id); - r.ack = brokkr::core::le_to_host(r.ack); -} - -inline RequestBox make_request(RqtCommandType type, RqtCommandParam param, std::span ints = {}, - std::span chars = {}) { - RequestBox r{}; - r.id = brokkr::core::host_to_le(static_cast(type)); - r.data = brokkr::core::host_to_le(static_cast(param)); - - if (!ints.empty()) { - const auto n = (ints.size() > RequestBox::DATA_INT_SIZE) ? RequestBox::DATA_INT_SIZE : ints.size(); - for (std::size_t i = 0; i < n; ++i) r.intData[i] = brokkr::core::host_to_le(ints[i]); - } - - if (!chars.empty()) { - const auto n = (chars.size() > RequestBox::DATA_CHAR_SIZE) ? RequestBox::DATA_CHAR_SIZE : chars.size(); - std::memcpy(r.charData, chars.data(), n * sizeof(std::int8_t)); - } - - return r; -} - -} // namespace brokkr::odin +/* + * Copyright (c) 2026 Gabriel2392 + * + * This program 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. + * + * This program 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 this program. If not, see . + */ + +#pragma once + +#include "core/endian.hpp" + +#include +#include +#include +#include +#include +#include +#include + +namespace brokkr::odin { + +enum class RqtCommandType : std::int32_t { + RQT_INIT = 100, + RQT_PIT = 101, + RQT_XMIT = 102, + RQT_CLOSE = 103, + RQT_EMPTY = 0 +}; + +enum class RqtCommandParam : std::int32_t { + // INIT + RQT_INIT_TARGET = 0, + RQT_INIT_RESETTIME = 1, + RQT_INIT_TOTALSIZE = 2, + RQT_INIT_OEMSTATE = 3, + RQT_INIT_NOOEMSTATE = 4, + RQT_INIT_PACKETSIZE = 5, + RQT_INIT_XMIT_SIZE = 6, + + // PIT + RQT_PIT_SET = 0, + RQT_PIT_GET = 1, + RQT_PIT_START = 2, + RQT_PIT_COMPLETE = 3, + + // XMIT (uncompressed) + RQT_XMIT_DOWNLOAD = 0, + RQT_XMIT_DUMP = 1, + RQT_XMIT_START = 2, + RQT_XMIT_COMPLETE = 3, + RQT_XMIT_SMD = 4, + + // XMIT (compressed) + RQT_XMIT_COMPRESSED_DOWNLOAD = 5, + RQT_XMIT_COMPRESSED_START = 6, + RQT_XMIT_COMPRESSED_COMPLETE = 7, + + // CLOSE + RQT_CLOSE_END = 0, + RQT_CLOSE_REBOOT = 1, + RQT_CLOSE_DISCONNECT = 2, + RQT_CLOSE_REBOOT_RECOVERY = 3, +}; + +enum class ProtocolVersion : std::int16_t { + PROTOCOL_NONE = 0, + PROTOCOL_VER1 = 1, + PROTOCOL_VER2 = 2, + PROTOCOL_VER3 = 3, + PROTOCOL_VER4 = 4, + PROTOCOL_VER5 = 5, +}; + +#pragma pack(push, 1) +struct ResponseBox { + std::int32_t id; + std::int32_t ack; +}; +#pragma pack(pop) +static_assert(sizeof(ResponseBox) == 8); + +#pragma pack(push, 1) +struct RequestBox { + static constexpr std::size_t DATA_INT_SIZE = 9; + static constexpr std::size_t DATA_CHAR_SIZE = 128; + static constexpr std::size_t MD5_SIZE = 32; + + std::int32_t id; + std::int32_t data; + std::int32_t intData[DATA_INT_SIZE]; + std::int8_t charData[DATA_CHAR_SIZE]; + std::int8_t md5[MD5_SIZE]; + + std::int8_t dummy[1024 - (2 * 4 + DATA_INT_SIZE * 4 + DATA_CHAR_SIZE + MD5_SIZE)]; +}; +#pragma pack(pop) +static_assert(sizeof(RequestBox) == 1024); + +inline void response_from_le(ResponseBox& r) noexcept { + r.id = brokkr::core::le_to_host(r.id); + r.ack = brokkr::core::le_to_host(r.ack); +} + +inline RequestBox make_request(RqtCommandType type, RqtCommandParam param, std::span ints = {}, + std::span chars = {}) { + RequestBox r{}; + r.id = brokkr::core::host_to_le(static_cast(type)); + r.data = brokkr::core::host_to_le(static_cast(param)); + + if (!ints.empty()) { + const auto n = (ints.size() > RequestBox::DATA_INT_SIZE) ? RequestBox::DATA_INT_SIZE : ints.size(); + for (std::size_t i = 0; i < n; ++i) r.intData[i] = brokkr::core::host_to_le(ints[i]); + } + + if (!chars.empty()) { + const auto n = (chars.size() > RequestBox::DATA_CHAR_SIZE) ? RequestBox::DATA_CHAR_SIZE : chars.size(); + std::memcpy(r.charData, chars.data(), n * sizeof(std::int8_t)); + } + + return r; +} + +} // namespace brokkr::odin diff --git a/src/third_party/xxhash/LICENSE b/src/third_party/xxhash/LICENSE new file mode 100644 index 0000000..e4c5da7 --- /dev/null +++ b/src/third_party/xxhash/LICENSE @@ -0,0 +1,26 @@ +xxHash Library +Copyright (c) 2012-2021 Yann Collet +All rights reserved. + +BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) + +Redistribution and use in source and binary forms, with or without modification, +are permitted provided that the following conditions are met: + +* Redistributions of source code must retain the above copyright notice, this + list of conditions and the following disclaimer. + +* 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. diff --git a/src/third_party/xxhash/xxhash.h b/src/third_party/xxhash/xxhash.h new file mode 100644 index 0000000..78fc2e8 --- /dev/null +++ b/src/third_party/xxhash/xxhash.h @@ -0,0 +1,7238 @@ +/* + * xxHash - Extremely Fast Hash algorithm + * Header File + * Copyright (C) 2012-2023 Yann Collet + * + * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are + * met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * 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 + * OWNER 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. + * + * You can contact the author at: + * - xxHash homepage: https://www.xxhash.com + * - xxHash source repository: https://github.com/Cyan4973/xxHash + */ + +/*! + * @mainpage xxHash + * + * xxHash is an extremely fast non-cryptographic hash algorithm, working at RAM speed + * limits. + * + * It is proposed in four flavors, in three families: + * 1. @ref XXH32_family + * - Classic 32-bit hash function. Simple, compact, and runs on almost all + * 32-bit and 64-bit systems. + * 2. @ref XXH64_family + * - Classic 64-bit adaptation of XXH32. Just as simple, and runs well on most + * 64-bit systems (but _not_ 32-bit systems). + * 3. @ref XXH3_family + * - Modern 64-bit and 128-bit hash function family which features improved + * strength and performance across the board, especially on smaller data. + * It benefits greatly from SIMD and 64-bit without requiring it. + * + * Benchmarks + * --- + * The reference system uses an Intel i7-9700K CPU, and runs Ubuntu x64 20.04. + * The open source benchmark program is compiled with clang v10.0 using -O3 flag. + * + * | Hash Name | ISA ext | Width | Large Data Speed | Small Data Velocity | + * | -------------------- | ------- | ----: | ---------------: | ------------------: | + * | XXH3_64bits() | @b AVX2 | 64 | 59.4 GB/s | 133.1 | + * | MeowHash | AES-NI | 128 | 58.2 GB/s | 52.5 | + * | XXH3_128bits() | @b AVX2 | 128 | 57.9 GB/s | 118.1 | + * | CLHash | PCLMUL | 64 | 37.1 GB/s | 58.1 | + * | XXH3_64bits() | @b SSE2 | 64 | 31.5 GB/s | 133.1 | + * | XXH3_128bits() | @b SSE2 | 128 | 29.6 GB/s | 118.1 | + * | RAM sequential read | | N/A | 28.0 GB/s | N/A | + * | ahash | AES-NI | 64 | 22.5 GB/s | 107.2 | + * | City64 | | 64 | 22.0 GB/s | 76.6 | + * | T1ha2 | | 64 | 22.0 GB/s | 99.0 | + * | City128 | | 128 | 21.7 GB/s | 57.7 | + * | FarmHash | AES-NI | 64 | 21.3 GB/s | 71.9 | + * | XXH64() | | 64 | 19.4 GB/s | 71.0 | + * | SpookyHash | | 64 | 19.3 GB/s | 53.2 | + * | Mum | | 64 | 18.0 GB/s | 67.0 | + * | CRC32C | SSE4.2 | 32 | 13.0 GB/s | 57.9 | + * | XXH32() | | 32 | 9.7 GB/s | 71.9 | + * | City32 | | 32 | 9.1 GB/s | 66.0 | + * | Blake3* | @b AVX2 | 256 | 4.4 GB/s | 8.1 | + * | Murmur3 | | 32 | 3.9 GB/s | 56.1 | + * | SipHash* | | 64 | 3.0 GB/s | 43.2 | + * | Blake3* | @b SSE2 | 256 | 2.4 GB/s | 8.1 | + * | HighwayHash | | 64 | 1.4 GB/s | 6.0 | + * | FNV64 | | 64 | 1.2 GB/s | 62.7 | + * | Blake2* | | 256 | 1.1 GB/s | 5.1 | + * | SHA1* | | 160 | 0.8 GB/s | 5.6 | + * | MD5* | | 128 | 0.6 GB/s | 7.8 | + * @note + * - Hashes which require a specific ISA extension are noted. SSE2 is also noted, + * even though it is mandatory on x64. + * - Hashes with an asterisk are cryptographic. Note that MD5 is non-cryptographic + * by modern standards. + * - Small data velocity is a rough average of algorithm's efficiency for small + * data. For more accurate information, see the wiki. + * - More benchmarks and strength tests are found on the wiki: + * https://github.com/Cyan4973/xxHash/wiki + * + * Usage + * ------ + * All xxHash variants use a similar API. Changing the algorithm is a trivial + * substitution. + * + * @pre + * For functions which take an input and length parameter, the following + * requirements are assumed: + * - The range from [`input`, `input + length`) is valid, readable memory. + * - The only exception is if the `length` is `0`, `input` may be `NULL`. + * - For C++, the objects must have the *TriviallyCopyable* property, as the + * functions access bytes directly as if it was an array of `unsigned char`. + * + * @anchor single_shot_example + * **Single Shot** + * + * These functions are stateless functions which hash a contiguous block of memory, + * immediately returning the result. They are the easiest and usually the fastest + * option. + * + * XXH32(), XXH64(), XXH3_64bits(), XXH3_128bits() + * + * @code{.c} + * #include + * #include "xxhash.h" + * + * // Example for a function which hashes a null terminated string with XXH32(). + * XXH32_hash_t hash_string(const char* string, XXH32_hash_t seed) + * { + * // NULL pointers are only valid if the length is zero + * size_t length = (string == NULL) ? 0 : strlen(string); + * return XXH32(string, length, seed); + * } + * @endcode + * + * + * @anchor streaming_example + * **Streaming** + * + * These groups of functions allow incremental hashing of unknown size, even + * more than what would fit in a size_t. + * + * XXH32_reset(), XXH64_reset(), XXH3_64bits_reset(), XXH3_128bits_reset() + * + * @code{.c} + * #include + * #include + * #include "xxhash.h" + * // Example for a function which hashes a FILE incrementally with XXH3_64bits(). + * XXH64_hash_t hashFile(FILE* f) + * { + * // Allocate a state struct. Do not just use malloc() or new. + * XXH3_state_t* state = XXH3_createState(); + * assert(state != NULL && "Out of memory!"); + * // Reset the state to start a new hashing session. + * XXH3_64bits_reset(state); + * char buffer[4096]; + * size_t count; + * // Read the file in chunks + * while ((count = fread(buffer, 1, sizeof(buffer), f)) != 0) { + * // Run update() as many times as necessary to process the data + * XXH3_64bits_update(state, buffer, count); + * } + * // Retrieve the finalized hash. This will not change the state. + * XXH64_hash_t result = XXH3_64bits_digest(state); + * // Free the state. Do not use free(). + * XXH3_freeState(state); + * return result; + * } + * @endcode + * + * Streaming functions generate the xxHash value from an incremental input. + * This method is slower than single-call functions, due to state management. + * For small inputs, prefer `XXH32()` and `XXH64()`, which are better optimized. + * + * An XXH state must first be allocated using `XXH*_createState()`. + * + * Start a new hash by initializing the state with a seed using `XXH*_reset()`. + * + * Then, feed the hash state by calling `XXH*_update()` as many times as necessary. + * + * The function returns an error code, with 0 meaning OK, and any other value + * meaning there is an error. + * + * Finally, a hash value can be produced anytime, by using `XXH*_digest()`. + * This function returns the nn-bits hash as an int or long long. + * + * It's still possible to continue inserting input into the hash state after a + * digest, and generate new hash values later on by invoking `XXH*_digest()`. + * + * When done, release the state using `XXH*_freeState()`. + * + * + * @anchor canonical_representation_example + * **Canonical Representation** + * + * The default return values from XXH functions are unsigned 32, 64 and 128 bit + * integers. + * This the simplest and fastest format for further post-processing. + * + * However, this leaves open the question of what is the order on the byte level, + * since little and big endian conventions will store the same number differently. + * + * The canonical representation settles this issue by mandating big-endian + * convention, the same convention as human-readable numbers (large digits first). + * + * When writing hash values to storage, sending them over a network, or printing + * them, it's highly recommended to use the canonical representation to ensure + * portability across a wider range of systems, present and future. + * + * The following functions allow transformation of hash values to and from + * canonical format. + * + * XXH32_canonicalFromHash(), XXH32_hashFromCanonical(), + * XXH64_canonicalFromHash(), XXH64_hashFromCanonical(), + * XXH128_canonicalFromHash(), XXH128_hashFromCanonical(), + * + * @code{.c} + * #include + * #include "xxhash.h" + * + * // Example for a function which prints XXH32_hash_t in human readable format + * void printXxh32(XXH32_hash_t hash) + * { + * XXH32_canonical_t cano; + * XXH32_canonicalFromHash(&cano, hash); + * size_t i; + * for(i = 0; i < sizeof(cano.digest); ++i) { + * printf("%02x", cano.digest[i]); + * } + * printf("\n"); + * } + * + * // Example for a function which converts XXH32_canonical_t to XXH32_hash_t + * XXH32_hash_t convertCanonicalToXxh32(XXH32_canonical_t cano) + * { + * XXH32_hash_t hash = XXH32_hashFromCanonical(&cano); + * return hash; + * } + * @endcode + * + * + * @file xxhash.h + * xxHash prototypes and implementation + */ + +#if defined (__cplusplus) +extern "C" { +#endif + +/* **************************** + * INLINE mode + ******************************/ +/*! + * @defgroup public Public API + * Contains details on the public xxHash functions. + * @{ + */ +#ifdef XXH_DOXYGEN +/*! + * @brief Gives access to internal state declaration, required for static allocation. + * + * Incompatible with dynamic linking, due to risks of ABI changes. + * + * Usage: + * @code{.c} + * #define XXH_STATIC_LINKING_ONLY + * #include "xxhash.h" + * @endcode + */ +# define XXH_STATIC_LINKING_ONLY +/* Do not undef XXH_STATIC_LINKING_ONLY for Doxygen */ + +/*! + * @brief Gives access to internal definitions. + * + * Usage: + * @code{.c} + * #define XXH_STATIC_LINKING_ONLY + * #define XXH_IMPLEMENTATION + * #include "xxhash.h" + * @endcode + */ +# define XXH_IMPLEMENTATION +/* Do not undef XXH_IMPLEMENTATION for Doxygen */ + +/*! + * @brief Exposes the implementation and marks all functions as `inline`. + * + * Use these build macros to inline xxhash into the target unit. + * Inlining improves performance on small inputs, especially when the length is + * expressed as a compile-time constant: + * + * https://fastcompression.blogspot.com/2018/03/xxhash-for-small-keys-impressive-power.html + * + * It also keeps xxHash symbols private to the unit, so they are not exported. + * + * Usage: + * @code{.c} + * #define XXH_INLINE_ALL + * #include "xxhash.h" + * @endcode + * Do not compile and link xxhash.o as a separate object, as it is not useful. + */ +# define XXH_INLINE_ALL +# undef XXH_INLINE_ALL +/*! + * @brief Exposes the implementation without marking functions as inline. + */ +# define XXH_PRIVATE_API +# undef XXH_PRIVATE_API +/*! + * @brief Emulate a namespace by transparently prefixing all symbols. + * + * If you want to include _and expose_ xxHash functions from within your own + * library, but also want to avoid symbol collisions with other libraries which + * may also include xxHash, you can use @ref XXH_NAMESPACE to automatically prefix + * any public symbol from xxhash library with the value of @ref XXH_NAMESPACE + * (therefore, avoid empty or numeric values). + * + * Note that no change is required within the calling program as long as it + * includes `xxhash.h`: Regular symbol names will be automatically translated + * by this header. + */ +# define XXH_NAMESPACE /* YOUR NAME HERE */ +# undef XXH_NAMESPACE +#endif + +#if (defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API)) \ + && !defined(XXH_INLINE_ALL_31684351384) + /* this section should be traversed only once */ +# define XXH_INLINE_ALL_31684351384 + /* give access to the advanced API, required to compile implementations */ +# undef XXH_STATIC_LINKING_ONLY /* avoid macro redef */ +# define XXH_STATIC_LINKING_ONLY + /* make all functions private */ +# undef XXH_PUBLIC_API +# if defined(__GNUC__) +# define XXH_PUBLIC_API static __inline __attribute__((__unused__)) +# elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) +# define XXH_PUBLIC_API static inline +# elif defined(_MSC_VER) +# define XXH_PUBLIC_API static __inline +# else + /* note: this version may generate warnings for unused static functions */ +# define XXH_PUBLIC_API static +# endif + + /* + * This part deals with the special case where a unit wants to inline xxHash, + * but "xxhash.h" has previously been included without XXH_INLINE_ALL, + * such as part of some previously included *.h header file. + * Without further action, the new include would just be ignored, + * and functions would effectively _not_ be inlined (silent failure). + * The following macros solve this situation by prefixing all inlined names, + * avoiding naming collision with previous inclusions. + */ + /* Before that, we unconditionally #undef all symbols, + * in case they were already defined with XXH_NAMESPACE. + * They will then be redefined for XXH_INLINE_ALL + */ +# undef XXH_versionNumber + /* XXH32 */ +# undef XXH32 +# undef XXH32_createState +# undef XXH32_freeState +# undef XXH32_reset +# undef XXH32_update +# undef XXH32_digest +# undef XXH32_copyState +# undef XXH32_canonicalFromHash +# undef XXH32_hashFromCanonical + /* XXH64 */ +# undef XXH64 +# undef XXH64_createState +# undef XXH64_freeState +# undef XXH64_reset +# undef XXH64_update +# undef XXH64_digest +# undef XXH64_copyState +# undef XXH64_canonicalFromHash +# undef XXH64_hashFromCanonical + /* XXH3_64bits */ +# undef XXH3_64bits +# undef XXH3_64bits_withSecret +# undef XXH3_64bits_withSeed +# undef XXH3_64bits_withSecretandSeed +# undef XXH3_createState +# undef XXH3_freeState +# undef XXH3_copyState +# undef XXH3_64bits_reset +# undef XXH3_64bits_reset_withSeed +# undef XXH3_64bits_reset_withSecret +# undef XXH3_64bits_update +# undef XXH3_64bits_digest +# undef XXH3_generateSecret + /* XXH3_128bits */ +# undef XXH128 +# undef XXH3_128bits +# undef XXH3_128bits_withSeed +# undef XXH3_128bits_withSecret +# undef XXH3_128bits_reset +# undef XXH3_128bits_reset_withSeed +# undef XXH3_128bits_reset_withSecret +# undef XXH3_128bits_reset_withSecretandSeed +# undef XXH3_128bits_update +# undef XXH3_128bits_digest +# undef XXH128_isEqual +# undef XXH128_cmp +# undef XXH128_canonicalFromHash +# undef XXH128_hashFromCanonical + /* Finally, free the namespace itself */ +# undef XXH_NAMESPACE + + /* employ the namespace for XXH_INLINE_ALL */ +# define XXH_NAMESPACE XXH_INLINE_ + /* + * Some identifiers (enums, type names) are not symbols, + * but they must nonetheless be renamed to avoid redeclaration. + * Alternative solution: do not redeclare them. + * However, this requires some #ifdefs, and has a more dispersed impact. + * Meanwhile, renaming can be achieved in a single place. + */ +# define XXH_IPREF(Id) XXH_NAMESPACE ## Id +# define XXH_OK XXH_IPREF(XXH_OK) +# define XXH_ERROR XXH_IPREF(XXH_ERROR) +# define XXH_errorcode XXH_IPREF(XXH_errorcode) +# define XXH32_canonical_t XXH_IPREF(XXH32_canonical_t) +# define XXH64_canonical_t XXH_IPREF(XXH64_canonical_t) +# define XXH128_canonical_t XXH_IPREF(XXH128_canonical_t) +# define XXH32_state_s XXH_IPREF(XXH32_state_s) +# define XXH32_state_t XXH_IPREF(XXH32_state_t) +# define XXH64_state_s XXH_IPREF(XXH64_state_s) +# define XXH64_state_t XXH_IPREF(XXH64_state_t) +# define XXH3_state_s XXH_IPREF(XXH3_state_s) +# define XXH3_state_t XXH_IPREF(XXH3_state_t) +# define XXH128_hash_t XXH_IPREF(XXH128_hash_t) + /* Ensure the header is parsed again, even if it was previously included */ +# undef XXHASH_H_5627135585666179 +# undef XXHASH_H_STATIC_13879238742 +#endif /* XXH_INLINE_ALL || XXH_PRIVATE_API */ + +/* **************************************************************** + * Stable API + *****************************************************************/ +#ifndef XXHASH_H_5627135585666179 +#define XXHASH_H_5627135585666179 1 + +/*! @brief Marks a global symbol. */ +#if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API) +# if defined(_WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT)) +# ifdef XXH_EXPORT +# define XXH_PUBLIC_API __declspec(dllexport) +# elif XXH_IMPORT +# define XXH_PUBLIC_API __declspec(dllimport) +# endif +# else +# define XXH_PUBLIC_API /* do nothing */ +# endif +#endif + +#ifdef XXH_NAMESPACE +# define XXH_CAT(A,B) A##B +# define XXH_NAME2(A,B) XXH_CAT(A,B) +# define XXH_versionNumber XXH_NAME2(XXH_NAMESPACE, XXH_versionNumber) +/* XXH32 */ +# define XXH32 XXH_NAME2(XXH_NAMESPACE, XXH32) +# define XXH32_createState XXH_NAME2(XXH_NAMESPACE, XXH32_createState) +# define XXH32_freeState XXH_NAME2(XXH_NAMESPACE, XXH32_freeState) +# define XXH32_reset XXH_NAME2(XXH_NAMESPACE, XXH32_reset) +# define XXH32_update XXH_NAME2(XXH_NAMESPACE, XXH32_update) +# define XXH32_digest XXH_NAME2(XXH_NAMESPACE, XXH32_digest) +# define XXH32_copyState XXH_NAME2(XXH_NAMESPACE, XXH32_copyState) +# define XXH32_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH32_canonicalFromHash) +# define XXH32_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH32_hashFromCanonical) +/* XXH64 */ +# define XXH64 XXH_NAME2(XXH_NAMESPACE, XXH64) +# define XXH64_createState XXH_NAME2(XXH_NAMESPACE, XXH64_createState) +# define XXH64_freeState XXH_NAME2(XXH_NAMESPACE, XXH64_freeState) +# define XXH64_reset XXH_NAME2(XXH_NAMESPACE, XXH64_reset) +# define XXH64_update XXH_NAME2(XXH_NAMESPACE, XXH64_update) +# define XXH64_digest XXH_NAME2(XXH_NAMESPACE, XXH64_digest) +# define XXH64_copyState XXH_NAME2(XXH_NAMESPACE, XXH64_copyState) +# define XXH64_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH64_canonicalFromHash) +# define XXH64_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH64_hashFromCanonical) +/* XXH3_64bits */ +# define XXH3_64bits XXH_NAME2(XXH_NAMESPACE, XXH3_64bits) +# define XXH3_64bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecret) +# define XXH3_64bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSeed) +# define XXH3_64bits_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecretandSeed) +# define XXH3_createState XXH_NAME2(XXH_NAMESPACE, XXH3_createState) +# define XXH3_freeState XXH_NAME2(XXH_NAMESPACE, XXH3_freeState) +# define XXH3_copyState XXH_NAME2(XXH_NAMESPACE, XXH3_copyState) +# define XXH3_64bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset) +# define XXH3_64bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSeed) +# define XXH3_64bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecret) +# define XXH3_64bits_reset_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecretandSeed) +# define XXH3_64bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_update) +# define XXH3_64bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_digest) +# define XXH3_generateSecret XXH_NAME2(XXH_NAMESPACE, XXH3_generateSecret) +# define XXH3_generateSecret_fromSeed XXH_NAME2(XXH_NAMESPACE, XXH3_generateSecret_fromSeed) +/* XXH3_128bits */ +# define XXH128 XXH_NAME2(XXH_NAMESPACE, XXH128) +# define XXH3_128bits XXH_NAME2(XXH_NAMESPACE, XXH3_128bits) +# define XXH3_128bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSeed) +# define XXH3_128bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecret) +# define XXH3_128bits_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecretandSeed) +# define XXH3_128bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset) +# define XXH3_128bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSeed) +# define XXH3_128bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecret) +# define XXH3_128bits_reset_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecretandSeed) +# define XXH3_128bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_update) +# define XXH3_128bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_digest) +# define XXH128_isEqual XXH_NAME2(XXH_NAMESPACE, XXH128_isEqual) +# define XXH128_cmp XXH_NAME2(XXH_NAMESPACE, XXH128_cmp) +# define XXH128_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH128_canonicalFromHash) +# define XXH128_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH128_hashFromCanonical) +#endif + + +/* ************************************* +* Compiler specifics +***************************************/ + +/* specific declaration modes for Windows */ +#if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API) +# if defined(_WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT)) +# ifdef XXH_EXPORT +# define XXH_PUBLIC_API __declspec(dllexport) +# elif XXH_IMPORT +# define XXH_PUBLIC_API __declspec(dllimport) +# endif +# else +# define XXH_PUBLIC_API /* do nothing */ +# endif +#endif + +#if defined (__GNUC__) +# define XXH_CONSTF __attribute__((__const__)) +# define XXH_PUREF __attribute__((__pure__)) +# define XXH_MALLOCF __attribute__((__malloc__)) +#else +# define XXH_CONSTF /* disable */ +# define XXH_PUREF +# define XXH_MALLOCF +#endif + +/* ************************************* +* Version +***************************************/ +#define XXH_VERSION_MAJOR 0 +#define XXH_VERSION_MINOR 8 +#define XXH_VERSION_RELEASE 3 +/*! @brief Version number, encoded as two digits each */ +#define XXH_VERSION_NUMBER (XXH_VERSION_MAJOR *100*100 + XXH_VERSION_MINOR *100 + XXH_VERSION_RELEASE) + +/*! + * @brief Obtains the xxHash version. + * + * This is mostly useful when xxHash is compiled as a shared library, + * since the returned value comes from the library, as opposed to header file. + * + * @return @ref XXH_VERSION_NUMBER of the invoked library. + */ +XXH_PUBLIC_API XXH_CONSTF unsigned XXH_versionNumber (void); + + +/* **************************** +* Common basic types +******************************/ +#include /* size_t */ +/*! + * @brief Exit code for the streaming API. + */ +typedef enum { + XXH_OK = 0, /*!< OK */ + XXH_ERROR /*!< Error */ +} XXH_errorcode; + + +/*-********************************************************************** +* 32-bit hash +************************************************************************/ +#if defined(XXH_DOXYGEN) /* Don't show include */ +/*! + * @brief An unsigned 32-bit integer. + * + * Not necessarily defined to `uint32_t` but functionally equivalent. + */ +typedef uint32_t XXH32_hash_t; + +#elif !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint32_t XXH32_hash_t; + +#else +# include +# if UINT_MAX == 0xFFFFFFFFUL + typedef unsigned int XXH32_hash_t; +# elif ULONG_MAX == 0xFFFFFFFFUL + typedef unsigned long XXH32_hash_t; +# else +# error "unsupported platform: need a 32-bit type" +# endif +#endif + +/*! + * @} + * + * @defgroup XXH32_family XXH32 family + * @ingroup public + * Contains functions used in the classic 32-bit xxHash algorithm. + * + * @note + * XXH32 is useful for older platforms, with no or poor 64-bit performance. + * Note that the @ref XXH3_family provides competitive speed for both 32-bit + * and 64-bit systems, and offers true 64/128 bit hash results. + * + * @see @ref XXH64_family, @ref XXH3_family : Other xxHash families + * @see @ref XXH32_impl for implementation details + * @{ + */ + +/*! + * @brief Calculates the 32-bit hash of @p input using xxHash32. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 32-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 32-bit xxHash32 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32 (const void* input, size_t length, XXH32_hash_t seed); + +#ifndef XXH_NO_STREAM +/*! + * @typedef struct XXH32_state_s XXH32_state_t + * @brief The opaque state struct for the XXH32 streaming API. + * + * @see XXH32_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH32_state_s XXH32_state_t; + +/*! + * @brief Allocates an @ref XXH32_state_t. + * + * @return An allocated pointer of @ref XXH32_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH32_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_MALLOCF XXH32_state_t* XXH32_createState(void); +/*! + * @brief Frees an @ref XXH32_state_t. + * + * @param statePtr A pointer to an @ref XXH32_state_t allocated with @ref XXH32_createState(). + * + * @return @ref XXH_OK. + * + * @note @p statePtr must be allocated with XXH32_createState(). + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr); +/*! + * @brief Copies one @ref XXH32_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dst_state, const XXH32_state_t* src_state); + +/*! + * @brief Resets an @ref XXH32_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 32-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note This function resets and seeds a state. Call it before @ref XXH32_update(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH32_reset (XXH32_state_t* statePtr, XXH32_hash_t seed); + +/*! + * @brief Consumes a block of @p input to an @ref XXH32_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void* input, size_t length); + +/*! + * @brief Returns the calculated hash value from an @ref XXH32_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated 32-bit xxHash32 value from that state. + * + * @note + * Calling XXH32_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_digest (const XXH32_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! + * @brief Canonical (big endian) representation of @ref XXH32_hash_t. + */ +typedef struct { + unsigned char digest[4]; /*!< Hash bytes, big endian */ +} XXH32_canonical_t; + +/*! + * @brief Converts an @ref XXH32_hash_t to a big endian @ref XXH32_canonical_t. + * + * @param dst The @ref XXH32_canonical_t pointer to be stored to. + * @param hash The @ref XXH32_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash); + +/*! + * @brief Converts an @ref XXH32_canonical_t to a native @ref XXH32_hash_t. + * + * @param src The @ref XXH32_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src); + + +/*! @cond Doxygen ignores this part */ +#ifdef __has_attribute +# define XXH_HAS_ATTRIBUTE(x) __has_attribute(x) +#else +# define XXH_HAS_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * C23 __STDC_VERSION__ number hasn't been specified yet. For now + * leave as `201711L` (C17 + 1). + * TODO: Update to correct value when its been specified. + */ +#define XXH_C23_VN 201711L +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* C-language Attributes are added in C23. */ +#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= XXH_C23_VN) && defined(__has_c_attribute) +# define XXH_HAS_C_ATTRIBUTE(x) __has_c_attribute(x) +#else +# define XXH_HAS_C_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +#if defined(__cplusplus) && defined(__has_cpp_attribute) +# define XXH_HAS_CPP_ATTRIBUTE(x) __has_cpp_attribute(x) +#else +# define XXH_HAS_CPP_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * Define XXH_FALLTHROUGH macro for annotating switch case with the 'fallthrough' attribute + * introduced in CPP17 and C23. + * CPP17 : https://en.cppreference.com/w/cpp/language/attributes/fallthrough + * C23 : https://en.cppreference.com/w/c/language/attributes/fallthrough + */ +#if XXH_HAS_C_ATTRIBUTE(fallthrough) || XXH_HAS_CPP_ATTRIBUTE(fallthrough) +# define XXH_FALLTHROUGH [[fallthrough]] +#elif XXH_HAS_ATTRIBUTE(__fallthrough__) +# define XXH_FALLTHROUGH __attribute__ ((__fallthrough__)) +#else +# define XXH_FALLTHROUGH /* fallthrough */ +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * Define XXH_NOESCAPE for annotated pointers in public API. + * https://clang.llvm.org/docs/AttributeReference.html#noescape + * As of writing this, only supported by clang. + */ +#if XXH_HAS_ATTRIBUTE(noescape) +# define XXH_NOESCAPE __attribute__((__noescape__)) +#else +# define XXH_NOESCAPE +#endif +/*! @endcond */ + + +/*! + * @} + * @ingroup public + * @{ + */ + +#ifndef XXH_NO_LONG_LONG +/*-********************************************************************** +* 64-bit hash +************************************************************************/ +#if defined(XXH_DOXYGEN) /* don't include */ +/*! + * @brief An unsigned 64-bit integer. + * + * Not necessarily defined to `uint64_t` but functionally equivalent. + */ +typedef uint64_t XXH64_hash_t; +#elif !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint64_t XXH64_hash_t; +#else +# include +# if defined(__LP64__) && ULONG_MAX == 0xFFFFFFFFFFFFFFFFULL + /* LP64 ABI says uint64_t is unsigned long */ + typedef unsigned long XXH64_hash_t; +# else + /* the following type must have a width of 64-bit */ + typedef unsigned long long XXH64_hash_t; +# endif +#endif + +/*! + * @} + * + * @defgroup XXH64_family XXH64 family + * @ingroup public + * @{ + * Contains functions used in the classic 64-bit xxHash algorithm. + * + * @note + * XXH3 provides competitive speed for both 32-bit and 64-bit systems, + * and offers true 64/128 bit hash results. + * It provides better speed for systems with vector processing capabilities. + */ + +/*! + * @brief Calculates the 64-bit hash of @p input using xxHash64. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 64-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit xxHash64 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed); + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/*! + * @brief The opaque state struct for the XXH64 streaming API. + * + * @see XXH64_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH64_state_s XXH64_state_t; /* incomplete type */ + +/*! + * @brief Allocates an @ref XXH64_state_t. + * + * @return An allocated pointer of @ref XXH64_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH64_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_MALLOCF XXH64_state_t* XXH64_createState(void); + +/*! + * @brief Frees an @ref XXH64_state_t. + * + * @param statePtr A pointer to an @ref XXH64_state_t allocated with @ref XXH64_createState(). + * + * @return @ref XXH_OK. + * + * @note @p statePtr must be allocated with XXH64_createState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr); + +/*! + * @brief Copies one @ref XXH64_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dst_state, const XXH64_state_t* src_state); + +/*! + * @brief Resets an @ref XXH64_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note This function resets and seeds a state. Call it before @ref XXH64_update(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_reset (XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed); + +/*! + * @brief Consumes a block of @p input to an @ref XXH64_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH_NOESCAPE XXH64_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated hash value from an @ref XXH64_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated 64-bit xxHash64 value from that state. + * + * @note + * Calling XXH64_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_digest (XXH_NOESCAPE const XXH64_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ +/******* Canonical representation *******/ + +/*! + * @brief Canonical (big endian) representation of @ref XXH64_hash_t. + */ +typedef struct { unsigned char digest[sizeof(XXH64_hash_t)]; } XXH64_canonical_t; + +/*! + * @brief Converts an @ref XXH64_hash_t to a big endian @ref XXH64_canonical_t. + * + * @param dst The @ref XXH64_canonical_t pointer to be stored to. + * @param hash The @ref XXH64_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH_NOESCAPE XXH64_canonical_t* dst, XXH64_hash_t hash); + +/*! + * @brief Converts an @ref XXH64_canonical_t to a native @ref XXH64_hash_t. + * + * @param src The @ref XXH64_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_hashFromCanonical(XXH_NOESCAPE const XXH64_canonical_t* src); + +#ifndef XXH_NO_XXH3 + +/*! + * @} + * ************************************************************************ + * @defgroup XXH3_family XXH3 family + * @ingroup public + * @{ + * + * XXH3 is a more recent hash algorithm featuring: + * - Improved speed for both small and large inputs + * - True 64-bit and 128-bit outputs + * - SIMD acceleration + * - Improved 32-bit viability + * + * Speed analysis methodology is explained here: + * + * https://fastcompression.blogspot.com/2019/03/presenting-xxh3.html + * + * Compared to XXH64, expect XXH3 to run approximately + * ~2x faster on large inputs and >3x faster on small ones, + * exact differences vary depending on platform. + * + * XXH3's speed benefits greatly from SIMD and 64-bit arithmetic, + * but does not require it. + * Most 32-bit and 64-bit targets that can run XXH32 smoothly can run XXH3 + * at competitive speeds, even without vector support. Further details are + * explained in the implementation. + * + * XXH3 has a fast scalar implementation, but it also includes accelerated SIMD + * implementations for many common platforms: + * - AVX512 + * - AVX2 + * - SSE2 + * - ARM NEON + * - WebAssembly SIMD128 + * - POWER8 VSX + * - s390x ZVector + * This can be controlled via the @ref XXH_VECTOR macro, but it automatically + * selects the best version according to predefined macros. For the x86 family, an + * automatic runtime dispatcher is included separately in @ref xxh_x86dispatch.c. + * + * XXH3 implementation is portable: + * it has a generic C90 formulation that can be compiled on any platform, + * all implementations generate exactly the same hash value on all platforms. + * Starting from v0.8.0, it's also labelled "stable", meaning that + * any future version will also generate the same hash value. + * + * XXH3 offers 2 variants, _64bits and _128bits. + * + * When only 64 bits are needed, prefer invoking the _64bits variant, as it + * reduces the amount of mixing, resulting in faster speed on small inputs. + * It's also generally simpler to manipulate a scalar return type than a struct. + * + * The API supports one-shot hashing, streaming mode, and custom secrets. + */ + +/*! + * @ingroup tuning + * @brief Possible values for @ref XXH_VECTOR. + * + * Unless set explicitly, determined automatically. + */ +# define XXH_SCALAR 0 /*!< Portable scalar version */ +# define XXH_SSE2 1 /*!< SSE2 for Pentium 4, Opteron, all x86_64. */ +# define XXH_AVX2 2 /*!< AVX2 for Haswell and Bulldozer */ +# define XXH_AVX512 3 /*!< AVX512 for Skylake and Icelake */ +# define XXH_NEON 4 /*!< NEON for most ARMv7-A, all AArch64, and WASM SIMD128 */ +# define XXH_VSX 5 /*!< VSX and ZVector for POWER8/z13 (64-bit) */ +# define XXH_SVE 6 /*!< SVE for some ARMv8-A and ARMv9-A */ +# define XXH_LSX 7 /*!< LSX (128-bit SIMD) for LoongArch64 */ + + +/*-********************************************************************** +* XXH3 64-bit variant +************************************************************************/ + +/*! + * @brief Calculates 64-bit unseeded variant of XXH3 hash of @p input. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @note + * This is equivalent to @ref XXH3_64bits_withSeed() with a seed of `0`, however + * it may have slightly better performance due to constant propagation of the + * defaults. + * + * @see + * XXH3_64bits_withSeed(), XXH3_64bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits(XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Calculates 64-bit seeded variant of XXH3 hash of @p input. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @note + * seed == 0 produces the same results as @ref XXH3_64bits(). + * + * This variant generates a custom secret on the fly based on default secret + * altered using the @p seed value. + * + * While this operation is decently fast, note that it's not completely free. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSeed(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed); + +/*! + * The bare minimum size for a custom secret. + * + * @see + * XXH3_64bits_withSecret(), XXH3_64bits_reset_withSecret(), + * XXH3_128bits_withSecret(), XXH3_128bits_reset_withSecret(). + */ +#define XXH3_SECRET_SIZE_MIN 136 + +/*! + * @brief Calculates 64-bit variant of XXH3 with a custom "secret". + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @pre + * The memory between @p data and @p data + @p len must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p data may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * It's possible to provide any blob of bytes as a "secret" to generate the hash. + * This makes it more difficult for an external actor to prepare an intentional collision. + * The main condition is that @p secretSize *must* be large enough (>= @ref XXH3_SECRET_SIZE_MIN). + * However, the quality of the secret impacts the dispersion of the hash algorithm. + * Therefore, the secret _must_ look like a bunch of random bytes. + * Avoid "trivial" or structured data such as repeated sequences or a text document. + * Whenever in doubt about the "randomness" of the blob of bytes, + * consider employing @ref XXH3_generateSecret() instead (see below). + * It will generate a proper high entropy secret derived from the blob of bytes. + * Another advantage of using XXH3_generateSecret() is that + * it guarantees that all bits within the initial blob of bytes + * will impact every bit of the output. + * This is not necessarily the case when using the blob of bytes directly + * because, when hashing _small_ inputs, only a portion of the secret is employed. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSecret(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize); + + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/* + * Streaming requires state maintenance. + * This operation costs memory and CPU. + * As a consequence, streaming is slower than one-shot hashing. + * For better performance, prefer one-shot functions whenever applicable. + */ + +/*! + * @brief The opaque state struct for the XXH3 streaming API. + * + * @see XXH3_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH3_state_s XXH3_state_t; +XXH_PUBLIC_API XXH_MALLOCF XXH3_state_t* XXH3_createState(void); +XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr); + +/*! + * @brief Copies one @ref XXH3_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOESCAPE const XXH3_state_t* src_state); + +/*! + * @brief Resets an @ref XXH3_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret with default parameters. + * - Call this function before @ref XXH3_64bits_update(). + * - Digest will be equivalent to `XXH3_64bits()`. + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); + +/*! + * @brief Resets an @ref XXH3_state_t with 64-bit seed to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret from `seed`. + * - Call this function before @ref XXH3_64bits_update(). + * - Digest will be equivalent to `XXH3_64bits_withSeed()`. + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); + +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * `secret` is referenced, it _must outlive_ the hash streaming session. + * + * Similar to one-shot API, `secretSize` must be >= @ref XXH3_SECRET_SIZE_MIN, + * and the quality of produced hash values depends on secret's entropy + * (secret's content should look like a bunch of random bytes). + * When in doubt about the randomness of a candidate `secret`, + * consider employing `XXH3_generateSecret()` instead (see below). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); + +/*! + * @brief Consumes a block of @p input to an @ref XXH3_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated XXH3 64-bit hash value from an @ref XXH3_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated XXH3 64-bit hash value from that state. + * + * @note + * Calling XXH3_64bits_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/* note : canonical representation of XXH3 is the same as XXH64 + * since they both produce XXH64_hash_t values */ + + +/*-********************************************************************** +* XXH3 128-bit variant +************************************************************************/ + +/*! + * @brief The return value from 128-bit hashes. + * + * Stored in little endian order, although the fields themselves are in native + * endianness. + */ +typedef struct { + XXH64_hash_t low64; /*!< `value & 0xFFFFFFFFFFFFFFFF` */ + XXH64_hash_t high64; /*!< `value >> 64` */ +} XXH128_hash_t; + +/*! + * @brief Calculates 128-bit unseeded variant of XXH3 of @p data. + * + * @param data The block of data to be hashed, at least @p length bytes in size. + * @param len The length of @p data, in bytes. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * The 128-bit variant of XXH3 has more strength, but it has a bit of overhead + * for shorter inputs. + * + * This is equivalent to @ref XXH3_128bits_withSeed() with a seed of `0`, however + * it may have slightly better performance due to constant propagation of the + * defaults. + * + * @see XXH3_128bits_withSeed(), XXH3_128bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits(XXH_NOESCAPE const void* data, size_t len); +/*! @brief Calculates 128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The block of data to be hashed, at least @p length bytes in size. + * @param len The length of @p data, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * @note + * seed == 0 produces the same results as @ref XXH3_64bits(). + * + * This variant generates a custom secret on the fly based on default secret + * altered using the @p seed value. + * + * While this operation is decently fast, note that it's not completely free. + * + * @see XXH3_128bits(), XXH3_128bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSeed(XXH_NOESCAPE const void* data, size_t len, XXH64_hash_t seed); +/*! + * @brief Calculates 128-bit variant of XXH3 with a custom "secret". + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * It's possible to provide any blob of bytes as a "secret" to generate the hash. + * This makes it more difficult for an external actor to prepare an intentional collision. + * The main condition is that @p secretSize *must* be large enough (>= @ref XXH3_SECRET_SIZE_MIN). + * However, the quality of the secret impacts the dispersion of the hash algorithm. + * Therefore, the secret _must_ look like a bunch of random bytes. + * Avoid "trivial" or structured data such as repeated sequences or a text document. + * Whenever in doubt about the "randomness" of the blob of bytes, + * consider employing @ref XXH3_generateSecret() instead (see below). + * It will generate a proper high entropy secret derived from the blob of bytes. + * Another advantage of using XXH3_generateSecret() is that + * it guarantees that all bits within the initial blob of bytes + * will impact every bit of the output. + * This is not necessarily the case when using the blob of bytes directly + * because, when hashing _small_ inputs, only a portion of the secret is employed. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSecret(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize); + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/* + * Streaming requires state maintenance. + * This operation costs memory and CPU. + * As a consequence, streaming is slower than one-shot hashing. + * For better performance, prefer one-shot functions whenever applicable. + * + * XXH3_128bits uses the same XXH3_state_t as XXH3_64bits(). + * Use already declared XXH3_createState() and XXH3_freeState(). + * + * All reset and streaming functions have same meaning as their 64-bit counterpart. + */ + +/*! + * @brief Resets an @ref XXH3_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret with default parameters. + * - Call it before @ref XXH3_128bits_update(). + * - Digest will be equivalent to `XXH3_128bits()`. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); + +/*! + * @brief Resets an @ref XXH3_state_t with 64-bit seed to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret from `seed`. + * - Call it before @ref XXH3_128bits_update(). + * - Digest will be equivalent to `XXH3_128bits_withSeed()`. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * `secret` is referenced, it _must outlive_ the hash streaming session. + * Similar to one-shot API, `secretSize` must be >= @ref XXH3_SECRET_SIZE_MIN, + * and the quality of produced hash values depends on secret's entropy + * (secret's content should look like a bunch of random bytes). + * When in doubt about the randomness of a candidate `secret`, + * consider employing `XXH3_generateSecret()` instead (see below). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); + +/*! + * @brief Consumes a block of @p input to an @ref XXH3_state_t. + * + * Call this to incrementally consume blocks of data. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated XXH3 128-bit hash value from an @ref XXH3_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated XXH3 128-bit hash value from that state. + * + * @note + * Calling XXH3_128bits_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/* Following helper functions make it possible to compare XXH128_hast_t values. + * Since XXH128_hash_t is a structure, this capability is not offered by the language. + * Note: For better performance, these functions can be inlined using XXH_INLINE_ALL */ + +/*! + * @brief Check equality of two XXH128_hash_t values + * + * @param h1 The 128-bit hash value. + * @param h2 Another 128-bit hash value. + * + * @return `1` if `h1` and `h2` are equal. + * @return `0` if they are not. + */ +XXH_PUBLIC_API XXH_PUREF int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2); + +/*! + * @brief Compares two @ref XXH128_hash_t + * + * This comparator is compatible with stdlib's `qsort()`/`bsearch()`. + * + * @param h128_1 Left-hand side value + * @param h128_2 Right-hand side value + * + * @return >0 if @p h128_1 > @p h128_2 + * @return =0 if @p h128_1 == @p h128_2 + * @return <0 if @p h128_1 < @p h128_2 + */ +XXH_PUBLIC_API XXH_PUREF int XXH128_cmp(XXH_NOESCAPE const void* h128_1, XXH_NOESCAPE const void* h128_2); + + +/******* Canonical representation *******/ +typedef struct { unsigned char digest[sizeof(XXH128_hash_t)]; } XXH128_canonical_t; + + +/*! + * @brief Converts an @ref XXH128_hash_t to a big endian @ref XXH128_canonical_t. + * + * @param dst The @ref XXH128_canonical_t pointer to be stored to. + * @param hash The @ref XXH128_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH128_canonicalFromHash(XXH_NOESCAPE XXH128_canonical_t* dst, XXH128_hash_t hash); + +/*! + * @brief Converts an @ref XXH128_canonical_t to a native @ref XXH128_hash_t. + * + * @param src The @ref XXH128_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH128_hashFromCanonical(XXH_NOESCAPE const XXH128_canonical_t* src); + + +#endif /* !XXH_NO_XXH3 */ +#endif /* XXH_NO_LONG_LONG */ + +/*! + * @} + */ +#endif /* XXHASH_H_5627135585666179 */ + + + +#if defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) +#define XXHASH_H_STATIC_13879238742 +/* **************************************************************************** + * This section contains declarations which are not guaranteed to remain stable. + * They may change in future versions, becoming incompatible with a different + * version of the library. + * These declarations should only be used with static linking. + * Never use them in association with dynamic linking! + ***************************************************************************** */ + +/* + * These definitions are only present to allow static allocation + * of XXH states, on stack or in a struct, for example. + * Never **ever** access their members directly. + */ + +/*! + * @internal + * @brief Structure for XXH32 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. Otherwise it is + * an opaque type. This allows fields to safely be changed. + * + * Typedef'd to @ref XXH32_state_t. + * Do not access the members of this struct directly. + * @see XXH64_state_s, XXH3_state_s + */ +struct XXH32_state_s { + XXH32_hash_t total_len_32; /*!< Total length hashed, modulo 2^32 */ + XXH32_hash_t large_len; /*!< Whether the hash is >= 16 (handles @ref total_len_32 overflow) */ + XXH32_hash_t acc[4]; /*!< Accumulator lanes */ + unsigned char buffer[16]; /*!< Internal buffer for partial reads. */ + XXH32_hash_t bufferedSize; /*!< Amount of data in @ref buffer */ + XXH32_hash_t reserved; /*!< Reserved field. Do not read nor write to it. */ +}; /* typedef'd to XXH32_state_t */ + + +#ifndef XXH_NO_LONG_LONG /* defined when there is no 64-bit support */ + +/*! + * @internal + * @brief Structure for XXH64 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. Otherwise it is + * an opaque type. This allows fields to safely be changed. + * + * Typedef'd to @ref XXH64_state_t. + * Do not access the members of this struct directly. + * @see XXH32_state_s, XXH3_state_s + */ +struct XXH64_state_s { + XXH64_hash_t total_len; /*!< Total length hashed. This is always 64-bit. */ + XXH64_hash_t acc[4]; /*!< Accumulator lanes */ + unsigned char buffer[32]; /*!< Internal buffer for partial reads.. */ + XXH32_hash_t bufferedSize; /*!< Amount of data in @ref buffer */ + XXH32_hash_t reserved32; /*!< Reserved field, needed for padding anyways*/ + XXH64_hash_t reserved64; /*!< Reserved field. Do not read or write to it. */ +}; /* typedef'd to XXH64_state_t */ + +#ifndef XXH_NO_XXH3 + +#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* >= C11 */ +# define XXH_ALIGN(n) _Alignas(n) +#elif defined(__cplusplus) && (__cplusplus >= 201103L) /* >= C++11 */ +/* In C++ alignas() is a keyword */ +# define XXH_ALIGN(n) alignas(n) +#elif defined(__GNUC__) +# define XXH_ALIGN(n) __attribute__ ((aligned(n))) +#elif defined(_MSC_VER) +# define XXH_ALIGN(n) __declspec(align(n)) +#else +# define XXH_ALIGN(n) /* disabled */ +#endif + +/* Old GCC versions only accept the attribute after the type in structures. */ +#if !(defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L)) /* C11+ */ \ + && ! (defined(__cplusplus) && (__cplusplus >= 201103L)) /* >= C++11 */ \ + && defined(__GNUC__) +# define XXH_ALIGN_MEMBER(align, type) type XXH_ALIGN(align) +#else +# define XXH_ALIGN_MEMBER(align, type) XXH_ALIGN(align) type +#endif + +/*! + * @brief The size of the internal XXH3 buffer. + * + * This is the optimal update size for incremental hashing. + * + * @see XXH3_64b_update(), XXH3_128b_update(). + */ +#define XXH3_INTERNALBUFFER_SIZE 256 + +/*! + * @internal + * @brief Default size of the secret buffer (and @ref XXH3_kSecret). + * + * This is the size used in @ref XXH3_kSecret and the seeded functions. + * + * Not to be confused with @ref XXH3_SECRET_SIZE_MIN. + */ +#define XXH3_SECRET_DEFAULT_SIZE 192 + +/*! + * @internal + * @brief Structure for XXH3 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. + * Otherwise it is an opaque type. + * Never use this definition in combination with dynamic library. + * This allows fields to safely be changed in the future. + * + * @note ** This structure has a strict alignment requirement of 64 bytes!! ** + * Do not allocate this with `malloc()` or `new`, + * it will not be sufficiently aligned. + * Use @ref XXH3_createState() and @ref XXH3_freeState(), or stack allocation. + * + * Typedef'd to @ref XXH3_state_t. + * Do never access the members of this struct directly. + * + * @see XXH3_INITSTATE() for stack initialization. + * @see XXH3_createState(), XXH3_freeState(). + * @see XXH32_state_s, XXH64_state_s + */ +struct XXH3_state_s { + XXH_ALIGN_MEMBER(64, XXH64_hash_t acc[8]); + /*!< The 8 accumulators. See @ref XXH32_state_s::v and @ref XXH64_state_s::v */ + XXH_ALIGN_MEMBER(64, unsigned char customSecret[XXH3_SECRET_DEFAULT_SIZE]); + /*!< Used to store a custom secret generated from a seed. */ + XXH_ALIGN_MEMBER(64, unsigned char buffer[XXH3_INTERNALBUFFER_SIZE]); + /*!< The internal buffer. @see XXH32_state_s::mem32 */ + XXH32_hash_t bufferedSize; + /*!< The amount of memory in @ref buffer, @see XXH32_state_s::memsize */ + XXH32_hash_t useSeed; + /*!< Reserved field. Needed for padding on 64-bit. */ + size_t nbStripesSoFar; + /*!< Number or stripes processed. */ + XXH64_hash_t totalLen; + /*!< Total length hashed. 64-bit even on 32-bit targets. */ + size_t nbStripesPerBlock; + /*!< Number of stripes per block. */ + size_t secretLimit; + /*!< Size of @ref customSecret or @ref extSecret */ + XXH64_hash_t seed; + /*!< Seed for _withSeed variants. Must be zero otherwise, @see XXH3_INITSTATE() */ + XXH64_hash_t reserved64; + /*!< Reserved field. */ + const unsigned char* extSecret; + /*!< Reference to an external secret for the _withSecret variants, NULL + * for other variants. */ + /* note: there may be some padding at the end due to alignment on 64 bytes */ +}; /* typedef'd to XXH3_state_t */ + +#undef XXH_ALIGN_MEMBER + +/*! + * @brief Initializes a stack-allocated `XXH3_state_s`. + * + * When the @ref XXH3_state_t structure is merely emplaced on stack, + * it should be initialized with XXH3_INITSTATE() or a memset() + * in case its first reset uses XXH3_NNbits_reset_withSeed(). + * This init can be omitted if the first reset uses default or _withSecret mode. + * This operation isn't necessary when the state is created with XXH3_createState(). + * Note that this doesn't prepare the state for a streaming operation, + * it's still necessary to use XXH3_NNbits_reset*() afterwards. + */ +#define XXH3_INITSTATE(XXH3_state_ptr) \ + do { \ + XXH3_state_t* tmp_xxh3_state_ptr = (XXH3_state_ptr); \ + tmp_xxh3_state_ptr->seed = 0; \ + tmp_xxh3_state_ptr->extSecret = NULL; \ + } while(0) + + +/*! + * @brief Calculates the 128-bit hash of @p data using XXH3. + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param seed The 64-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p data and @p data + @p len must be valid, + * readable, contiguous memory. However, if @p len is `0`, @p data may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 128-bit XXH3 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH128(XXH_NOESCAPE const void* data, size_t len, XXH64_hash_t seed); + + +/* === Experimental API === */ +/* Symbols defined below must be considered tied to a specific library version. */ + +/*! + * @brief Derive a high-entropy secret from any user-defined content, named customSeed. + * + * @param secretBuffer A writable buffer for derived high-entropy secret data. + * @param secretSize Size of secretBuffer, in bytes. Must be >= XXH3_SECRET_SIZE_MIN. + * @param customSeed A user-defined content. + * @param customSeedSize Size of customSeed, in bytes. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * The generated secret can be used in combination with `*_withSecret()` functions. + * The `_withSecret()` variants are useful to provide a higher level of protection + * than 64-bit seed, as it becomes much more difficult for an external actor to + * guess how to impact the calculation logic. + * + * The function accepts as input a custom seed of any length and any content, + * and derives from it a high-entropy secret of length @p secretSize into an + * already allocated buffer @p secretBuffer. + * + * The generated secret can then be used with any `*_withSecret()` variant. + * The functions @ref XXH3_128bits_withSecret(), @ref XXH3_64bits_withSecret(), + * @ref XXH3_128bits_reset_withSecret() and @ref XXH3_64bits_reset_withSecret() + * are part of this list. They all accept a `secret` parameter + * which must be large enough for implementation reasons (>= @ref XXH3_SECRET_SIZE_MIN) + * _and_ feature very high entropy (consist of random-looking bytes). + * These conditions can be a high bar to meet, so @ref XXH3_generateSecret() can + * be employed to ensure proper quality. + * + * @p customSeed can be anything. It can have any size, even small ones, + * and its content can be anything, even "poor entropy" sources such as a bunch + * of zeroes. The resulting `secret` will nonetheless provide all required qualities. + * + * @pre + * - @p secretSize must be >= @ref XXH3_SECRET_SIZE_MIN + * - When @p customSeedSize > 0, supplying NULL as customSeed is undefined behavior. + * + * Example code: + * @code{.c} + * #include + * #include + * #include + * #define XXH_STATIC_LINKING_ONLY // expose unstable API + * #include "xxhash.h" + * // Hashes argv[2] using the entropy from argv[1]. + * int main(int argc, char* argv[]) + * { + * char secret[XXH3_SECRET_SIZE_MIN]; + * if (argv != 3) { return 1; } + * XXH3_generateSecret(secret, sizeof(secret), argv[1], strlen(argv[1])); + * XXH64_hash_t h = XXH3_64bits_withSecret( + * argv[2], strlen(argv[2]), + * secret, sizeof(secret) + * ); + * printf("%016llx\n", (unsigned long long) h); + * } + * @endcode + */ +XXH_PUBLIC_API XXH_errorcode XXH3_generateSecret(XXH_NOESCAPE void* secretBuffer, size_t secretSize, XXH_NOESCAPE const void* customSeed, size_t customSeedSize); + +/*! + * @brief Generate the same secret as the _withSeed() variants. + * + * @param secretBuffer A writable buffer of @ref XXH3_SECRET_DEFAULT_SIZE bytes + * @param seed The 64-bit seed to alter the hash result predictably. + * + * The generated secret can be used in combination with + *`*_withSecret()` and `_withSecretandSeed()` variants. + * + * Example C++ `std::string` hash class: + * @code{.cpp} + * #include + * #define XXH_STATIC_LINKING_ONLY // expose unstable API + * #include "xxhash.h" + * // Slow, seeds each time + * class HashSlow { + * XXH64_hash_t seed; + * public: + * HashSlow(XXH64_hash_t s) : seed{s} {} + * size_t operator()(const std::string& x) const { + * return size_t{XXH3_64bits_withSeed(x.c_str(), x.length(), seed)}; + * } + * }; + * // Fast, caches the seeded secret for future uses. + * class HashFast { + * unsigned char secret[XXH3_SECRET_DEFAULT_SIZE]; + * public: + * HashFast(XXH64_hash_t s) { + * XXH3_generateSecret_fromSeed(secret, seed); + * } + * size_t operator()(const std::string& x) const { + * return size_t{ + * XXH3_64bits_withSecret(x.c_str(), x.length(), secret, sizeof(secret)) + * }; + * } + * }; + * @endcode + */ +XXH_PUBLIC_API void XXH3_generateSecret_fromSeed(XXH_NOESCAPE void* secretBuffer, XXH64_hash_t seed); + +/*! + * @brief Maximum size of "short" key in bytes. + */ +#define XXH3_MIDSIZE_MAX 240 + +/*! + * @brief Calculates 64/128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * These variants generate hash values using either: + * - @p seed for "short" keys (< @ref XXH3_MIDSIZE_MAX = 240 bytes) + * - @p secret for "large" keys (>= @ref XXH3_MIDSIZE_MAX). + * + * This generally benefits speed, compared to `_withSeed()` or `_withSecret()`. + * `_withSeed()` has to generate the secret on the fly for "large" keys. + * It's fast, but can be perceptible for "not so large" keys (< 1 KB). + * `_withSecret()` has to generate the masks on the fly for "small" keys, + * which requires more instructions than _withSeed() variants. + * Therefore, _withSecretandSeed variant combines the best of both worlds. + * + * When @p secret has been generated by XXH3_generateSecret_fromSeed(), + * this variant produces *exactly* the same results as `_withSeed()` variant, + * hence offering only a pure speed benefit on "large" input, + * by skipping the need to regenerate the secret for every large input. + * + * Another usage scenario is to hash the secret to a 64-bit hash value, + * for example with XXH3_64bits(), which then becomes the seed, + * and then employ both the seed and the secret in _withSecretandSeed(). + * On top of speed, an added benefit is that each bit in the secret + * has a 50% chance to swap each bit in the output, via its impact to the seed. + * + * This is not guaranteed when using the secret directly in "small data" scenarios, + * because only portions of the secret are employed for small data. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t +XXH3_64bits_withSecretandSeed(XXH_NOESCAPE const void* data, size_t len, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed); + +/*! + * @brief Calculates 128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The memory segment to be hashed, at least @p len bytes in size. + * @param length The length of @p data, in bytes. + * @param secret The secret used to alter hash result predictably. + * @param secretSize The length of @p secret, in bytes (must be >= XXH3_SECRET_SIZE_MIN) + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(): contract is the same. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t +XXH3_128bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t length, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +#ifndef XXH_NO_STREAM +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(). Contract is identical. + */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(). Contract is identical. + * + * Note: there was a bug in an earlier version of this function (<= v0.8.2) + * that would make it generate an incorrect hash value + * when @p seed == 0 and @p length < XXH3_MIDSIZE_MAX + * and @p secret is different from XXH3_generateSecret_fromSeed(). + * As stated in the contract, the correct hash result must be + * the same as XXH3_128bits_withSeed() when @p length <= XXH3_MIDSIZE_MAX. + * Results generated by this older version are wrong, hence not comparable. + */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +#endif /* !XXH_NO_STREAM */ + +#endif /* !XXH_NO_XXH3 */ +#endif /* XXH_NO_LONG_LONG */ +#if defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) +# define XXH_IMPLEMENTATION +#endif + +#endif /* defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) */ + + +/* ======================================================================== */ +/* ======================================================================== */ +/* ======================================================================== */ + + +/*-********************************************************************** + * xxHash implementation + *-********************************************************************** + * xxHash's implementation used to be hosted inside xxhash.c. + * + * However, inlining requires implementation to be visible to the compiler, + * hence be included alongside the header. + * Previously, implementation was hosted inside xxhash.c, + * which was then #included when inlining was activated. + * This construction created issues with a few build and install systems, + * as it required xxhash.c to be stored in /include directory. + * + * xxHash implementation is now directly integrated within xxhash.h. + * As a consequence, xxhash.c is no longer needed in /include. + * + * xxhash.c is still available and is still useful. + * In a "normal" setup, when xxhash is not inlined, + * xxhash.h only exposes the prototypes and public symbols, + * while xxhash.c can be built into an object file xxhash.o + * which can then be linked into the final binary. + ************************************************************************/ + +#if ( defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) \ + || defined(XXH_IMPLEMENTATION) ) && !defined(XXH_IMPLEM_13a8737387) +# define XXH_IMPLEM_13a8737387 + +/* ************************************* +* Tuning parameters +***************************************/ + +/*! + * @defgroup tuning Tuning parameters + * @{ + * + * Various macros to control xxHash's behavior. + */ +#ifdef XXH_DOXYGEN +/*! + * @brief Define this to disable 64-bit code. + * + * Useful if only using the @ref XXH32_family and you have a strict C90 compiler. + */ +# define XXH_NO_LONG_LONG +# undef XXH_NO_LONG_LONG /* don't actually */ +/*! + * @brief Controls how unaligned memory is accessed. + * + * By default, access to unaligned memory is controlled by `memcpy()`, which is + * safe and portable. + * + * Unfortunately, on some target/compiler combinations, the generated assembly + * is sub-optimal. + * + * The below switch allow selection of a different access method + * in the search for improved performance. + * + * @par Possible options: + * + * - `XXH_FORCE_MEMORY_ACCESS=0` (default): `memcpy` + * @par + * Use `memcpy()`. Safe and portable. Note that most modern compilers will + * eliminate the function call and treat it as an unaligned access. + * + * - `XXH_FORCE_MEMORY_ACCESS=1`: `__attribute__((aligned(1)))` + * @par + * Depends on compiler extensions and is therefore not portable. + * This method is safe _if_ your compiler supports it, + * and *generally* as fast or faster than `memcpy`. + * + * - `XXH_FORCE_MEMORY_ACCESS=2`: Direct cast + * @par + * Casts directly and dereferences. This method doesn't depend on the + * compiler, but it violates the C standard as it directly dereferences an + * unaligned pointer. It can generate buggy code on targets which do not + * support unaligned memory accesses, but in some circumstances, it's the + * only known way to get the most performance. + * + * - `XXH_FORCE_MEMORY_ACCESS=3`: Byteshift + * @par + * Also portable. This can generate the best code on old compilers which don't + * inline small `memcpy()` calls, and it might also be faster on big-endian + * systems which lack a native byteswap instruction. However, some compilers + * will emit literal byteshifts even if the target supports unaligned access. + * + * + * @warning + * Methods 1 and 2 rely on implementation-defined behavior. Use these with + * care, as what works on one compiler/platform/optimization level may cause + * another to read garbage data or even crash. + * + * See https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html for details. + * + * Prefer these methods in priority order (0 > 3 > 1 > 2) + */ +# define XXH_FORCE_MEMORY_ACCESS 0 + +/*! + * @def XXH_SIZE_OPT + * @brief Controls how much xxHash optimizes for size. + * + * xxHash, when compiled, tends to result in a rather large binary size. This + * is mostly due to heavy usage to forced inlining and constant folding of the + * @ref XXH3_family to increase performance. + * + * However, some developers prefer size over speed. This option can + * significantly reduce the size of the generated code. When using the `-Os` + * or `-Oz` options on GCC or Clang, this is defined to 1 by default, + * otherwise it is defined to 0. + * + * Most of these size optimizations can be controlled manually. + * + * This is a number from 0-2. + * - `XXH_SIZE_OPT` == 0: Default. xxHash makes no size optimizations. Speed + * comes first. + * - `XXH_SIZE_OPT` == 1: Default for `-Os` and `-Oz`. xxHash is more + * conservative and disables hacks that increase code size. It implies the + * options @ref XXH_NO_INLINE_HINTS == 1, @ref XXH_FORCE_ALIGN_CHECK == 0, + * and @ref XXH3_NEON_LANES == 8 if they are not already defined. + * - `XXH_SIZE_OPT` == 2: xxHash tries to make itself as small as possible. + * Performance may cry. For example, the single shot functions just use the + * streaming API. + */ +# define XXH_SIZE_OPT 0 + +/*! + * @def XXH_FORCE_ALIGN_CHECK + * @brief If defined to non-zero, adds a special path for aligned inputs (XXH32() + * and XXH64() only). + * + * This is an important performance trick for architectures without decent + * unaligned memory access performance. + * + * It checks for input alignment, and when conditions are met, uses a "fast + * path" employing direct 32-bit/64-bit reads, resulting in _dramatically + * faster_ read speed. + * + * The check costs one initial branch per hash, which is generally negligible, + * but not zero. + * + * Moreover, it's not useful to generate an additional code path if memory + * access uses the same instruction for both aligned and unaligned + * addresses (e.g. x86 and aarch64). + * + * In these cases, the alignment check can be removed by setting this macro to 0. + * Then the code will always use unaligned memory access. + * Align check is automatically disabled on x86, x64, ARM64, and some ARM chips + * which are platforms known to offer good unaligned memory accesses performance. + * + * It is also disabled by default when @ref XXH_SIZE_OPT >= 1. + * + * This option does not affect XXH3 (only XXH32 and XXH64). + */ +# define XXH_FORCE_ALIGN_CHECK 0 + +/*! + * @def XXH_NO_INLINE_HINTS + * @brief When non-zero, sets all functions to `static`. + * + * By default, xxHash tries to force the compiler to inline almost all internal + * functions. + * + * This can usually improve performance due to reduced jumping and improved + * constant folding, but significantly increases the size of the binary which + * might not be favorable. + * + * Additionally, sometimes the forced inlining can be detrimental to performance, + * depending on the architecture. + * + * XXH_NO_INLINE_HINTS marks all internal functions as static, giving the + * compiler full control on whether to inline or not. + * + * When not optimizing (-O0), using `-fno-inline` with GCC or Clang, or if + * @ref XXH_SIZE_OPT >= 1, this will automatically be defined. + */ +# define XXH_NO_INLINE_HINTS 0 + +/*! + * @def XXH3_INLINE_SECRET + * @brief Determines whether to inline the XXH3 withSecret code. + * + * When the secret size is known, the compiler can improve the performance + * of XXH3_64bits_withSecret() and XXH3_128bits_withSecret(). + * + * However, if the secret size is not known, it doesn't have any benefit. This + * happens when xxHash is compiled into a global symbol. Therefore, if + * @ref XXH_INLINE_ALL is *not* defined, this will be defined to 0. + * + * Additionally, this defaults to 0 on GCC 12+, which has an issue with function pointers + * that are *sometimes* force inline on -Og, and it is impossible to automatically + * detect this optimization level. + */ +# define XXH3_INLINE_SECRET 0 + +/*! + * @def XXH32_ENDJMP + * @brief Whether to use a jump for `XXH32_finalize`. + * + * For performance, `XXH32_finalize` uses multiple branches in the finalizer. + * This is generally preferable for performance, + * but depending on exact architecture, a jmp may be preferable. + * + * This setting is only possibly making a difference for very small inputs. + */ +# define XXH32_ENDJMP 0 + +/*! + * @internal + * @brief Redefines old internal names. + * + * For compatibility with code that uses xxHash's internals before the names + * were changed to improve namespacing. There is no other reason to use this. + */ +# define XXH_OLD_NAMES +# undef XXH_OLD_NAMES /* don't actually use, it is ugly. */ + +/*! + * @def XXH_NO_STREAM + * @brief Disables the streaming API. + * + * When xxHash is not inlined and the streaming functions are not used, disabling + * the streaming functions can improve code size significantly, especially with + * the @ref XXH3_family which tends to make constant folded copies of itself. + */ +# define XXH_NO_STREAM +# undef XXH_NO_STREAM /* don't actually */ +#endif /* XXH_DOXYGEN */ +/*! + * @} + */ + +#ifndef XXH_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ + /* prefer __packed__ structures (method 1) for GCC + * < ARMv7 with unaligned access (e.g. Raspbian armhf) still uses byte shifting, so we use memcpy + * which for some reason does unaligned loads. */ +# if defined(__GNUC__) && !(defined(__ARM_ARCH) && __ARM_ARCH < 7 && defined(__ARM_FEATURE_UNALIGNED)) +# define XXH_FORCE_MEMORY_ACCESS 1 +# endif +#endif + +#ifndef XXH_SIZE_OPT + /* default to 1 for -Os or -Oz */ +# if (defined(__GNUC__) || defined(__clang__)) && defined(__OPTIMIZE_SIZE__) +# define XXH_SIZE_OPT 1 +# else +# define XXH_SIZE_OPT 0 +# endif +#endif + +#ifndef XXH_FORCE_ALIGN_CHECK /* can be defined externally */ + /* don't check on sizeopt, x86, aarch64, or arm when unaligned access is available */ +# if XXH_SIZE_OPT >= 1 || \ + defined(__i386) || defined(__x86_64__) || defined(__aarch64__) || defined(__ARM_FEATURE_UNALIGNED) \ + || defined(_M_IX86) || defined(_M_X64) || defined(_M_ARM64) || defined(_M_ARM) /* visual */ +# define XXH_FORCE_ALIGN_CHECK 0 +# else +# define XXH_FORCE_ALIGN_CHECK 1 +# endif +#endif + +#ifndef XXH_NO_INLINE_HINTS +# if XXH_SIZE_OPT >= 1 || defined(__NO_INLINE__) /* -O0, -fno-inline */ +# define XXH_NO_INLINE_HINTS 1 +# else +# define XXH_NO_INLINE_HINTS 0 +# endif +#endif + +#ifndef XXH3_INLINE_SECRET +# if (defined(__GNUC__) && !defined(__clang__) && __GNUC__ >= 12) \ + || !defined(XXH_INLINE_ALL) +# define XXH3_INLINE_SECRET 0 +# else +# define XXH3_INLINE_SECRET 1 +# endif +#endif + +#ifndef XXH32_ENDJMP +/* generally preferable for performance */ +# define XXH32_ENDJMP 0 +#endif + +/*! + * @defgroup impl Implementation + * @{ + */ + + +/* ************************************* +* Includes & Memory related functions +***************************************/ +#if defined(XXH_NO_STREAM) +/* nothing */ +#elif defined(XXH_NO_STDLIB) + +/* When requesting to disable any mention of stdlib, + * the library loses the ability to invoked malloc / free. + * In practice, it means that functions like `XXH*_createState()` + * will always fail, and return NULL. + * This flag is useful in situations where + * xxhash.h is integrated into some kernel, embedded or limited environment + * without access to dynamic allocation. + */ + +static XXH_CONSTF void* XXH_malloc(size_t s) { (void)s; return NULL; } +static void XXH_free(void* p) { (void)p; } + +#else + +/* + * Modify the local functions below should you wish to use + * different memory routines for malloc() and free() + */ +#include + +/*! + * @internal + * @brief Modify this function to use a different routine than malloc(). + */ +static XXH_MALLOCF void* XXH_malloc(size_t s) { return malloc(s); } + +/*! + * @internal + * @brief Modify this function to use a different routine than free(). + */ +static void XXH_free(void* p) { free(p); } + +#endif /* XXH_NO_STDLIB */ + +#include + +/*! + * @internal + * @brief Modify this function to use a different routine than memcpy(). + */ +static void* XXH_memcpy(void* dest, const void* src, size_t size) +{ + return memcpy(dest,src,size); +} + +#include /* ULLONG_MAX */ + + +/* ************************************* +* Compiler Specific Options +***************************************/ +#ifdef _MSC_VER /* Visual Studio warning fix */ +# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ +#endif + +#if XXH_NO_INLINE_HINTS /* disable inlining hints */ +# if defined(__GNUC__) || defined(__clang__) +# define XXH_FORCE_INLINE static __attribute__((__unused__)) +# else +# define XXH_FORCE_INLINE static +# endif +# define XXH_NO_INLINE static +/* enable inlining hints */ +#elif defined(__GNUC__) || defined(__clang__) +# define XXH_FORCE_INLINE static __inline__ __attribute__((__always_inline__, __unused__)) +# define XXH_NO_INLINE static __attribute__((__noinline__)) +#elif defined(_MSC_VER) /* Visual Studio */ +# define XXH_FORCE_INLINE static __forceinline +# define XXH_NO_INLINE static __declspec(noinline) +#elif defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L)) /* C99 */ +# define XXH_FORCE_INLINE static inline +# define XXH_NO_INLINE static +#else +# define XXH_FORCE_INLINE static +# define XXH_NO_INLINE static +#endif + +#if defined(XXH_INLINE_ALL) +# define XXH_STATIC XXH_FORCE_INLINE +#else +# define XXH_STATIC static +#endif + +#if XXH3_INLINE_SECRET +# define XXH3_WITH_SECRET_INLINE XXH_FORCE_INLINE +#else +# define XXH3_WITH_SECRET_INLINE XXH_NO_INLINE +#endif + +#if ((defined(sun) || defined(__sun)) && __cplusplus) /* Solaris includes __STDC_VERSION__ with C++. Tested with GCC 5.5 */ +# define XXH_RESTRICT /* disable */ +#elif defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* >= C99 */ +# define XXH_RESTRICT restrict +#elif (defined (__GNUC__) && ((__GNUC__ > 3) || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1))) \ + || (defined (__clang__)) \ + || (defined (_MSC_VER) && (_MSC_VER >= 1400)) \ + || (defined (__INTEL_COMPILER) && (__INTEL_COMPILER >= 1300)) +/* + * There are a LOT more compilers that recognize __restrict but this + * covers the major ones. + */ +# define XXH_RESTRICT __restrict +#else +# define XXH_RESTRICT /* disable */ +#endif + +/* ************************************* +* Debug +***************************************/ +/*! + * @ingroup tuning + * @def XXH_DEBUGLEVEL + * @brief Sets the debugging level. + * + * XXH_DEBUGLEVEL is expected to be defined externally, typically via the + * compiler's command line options. The value must be a number. + */ +#ifndef XXH_DEBUGLEVEL +# ifdef DEBUGLEVEL /* backwards compat */ +# define XXH_DEBUGLEVEL DEBUGLEVEL +# else +# define XXH_DEBUGLEVEL 0 +# endif +#endif + +#if (XXH_DEBUGLEVEL>=1) +# include /* note: can still be disabled with NDEBUG */ +# define XXH_ASSERT(c) assert(c) +#else +# if defined(__INTEL_COMPILER) +# define XXH_ASSERT(c) XXH_ASSUME((unsigned char) (c)) +# else +# define XXH_ASSERT(c) XXH_ASSUME(c) +# endif +#endif + +/* note: use after variable declarations */ +#ifndef XXH_STATIC_ASSERT +# if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */ +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { _Static_assert((c),m); } while(0) +# elif defined(__cplusplus) && (__cplusplus >= 201103L) /* C++11 */ +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { static_assert((c),m); } while(0) +# else +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { struct xxh_sa { char x[(c) ? 1 : -1]; }; } while(0) +# endif +# define XXH_STATIC_ASSERT(c) XXH_STATIC_ASSERT_WITH_MESSAGE((c),#c) +#endif + +/*! + * @internal + * @def XXH_COMPILER_GUARD(var) + * @brief Used to prevent unwanted optimizations for @p var. + * + * It uses an empty GCC inline assembly statement with a register constraint + * which forces @p var into a general purpose register (eg eax, ebx, ecx + * on x86) and marks it as modified. + * + * This is used in a few places to avoid unwanted autovectorization (e.g. + * XXH32_round()). All vectorization we want is explicit via intrinsics, + * and _usually_ isn't wanted elsewhere. + * + * We also use it to prevent unwanted constant folding for AArch64 in + * XXH3_initCustomSecret_scalar(). + */ +#if defined(__GNUC__) || defined(__clang__) +# define XXH_COMPILER_GUARD(var) __asm__("" : "+r" (var)) +#else +# define XXH_COMPILER_GUARD(var) ((void)0) +#endif + +/* Specifically for NEON vectors which use the "w" constraint, on + * Clang. */ +#if defined(__clang__) && defined(__ARM_ARCH) && !defined(__wasm__) +# define XXH_COMPILER_GUARD_CLANG_NEON(var) __asm__("" : "+w" (var)) +#else +# define XXH_COMPILER_GUARD_CLANG_NEON(var) ((void)0) +#endif + +/* ************************************* +* Basic Types +***************************************/ +#if !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint8_t xxh_u8; +#else + typedef unsigned char xxh_u8; +#endif +typedef XXH32_hash_t xxh_u32; + +#ifdef XXH_OLD_NAMES +# warning "XXH_OLD_NAMES is planned to be removed starting v0.9. If the program depends on it, consider moving away from it by employing newer type names directly" +# define BYTE xxh_u8 +# define U8 xxh_u8 +# define U32 xxh_u32 +#endif + +/* *** Memory access *** */ + +/*! + * @internal + * @fn xxh_u32 XXH_read32(const void* ptr) + * @brief Reads an unaligned 32-bit integer from @p ptr in native endianness. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit native endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readLE32(const void* ptr) + * @brief Reads an unaligned 32-bit little endian integer from @p ptr. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit little endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readBE32(const void* ptr) + * @brief Reads an unaligned 32-bit big endian integer from @p ptr. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit big endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readLE32_align(const void* ptr, XXH_alignment align) + * @brief Like @ref XXH_readLE32(), but has an option for aligned reads. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * Note that when @ref XXH_FORCE_ALIGN_CHECK == 0, the @p align parameter is + * always @ref XXH_alignment::XXH_unaligned. + * + * @param ptr The pointer to read from. + * @param align Whether @p ptr is aligned. + * @pre + * If @p align == @ref XXH_alignment::XXH_aligned, @p ptr must be 4 byte + * aligned. + * @return The 32-bit little endian integer from the bytes at @p ptr. + */ + +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) +/* + * Manual byteshift. Best for old compilers which don't inline memcpy. + * We actually directly use XXH_readLE32 and XXH_readBE32. + */ +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) + +/* + * Force direct memory access. Only works on CPU which support unaligned memory + * access in hardware. + */ +static xxh_u32 XXH_read32(const void* memPtr) { return *(const xxh_u32*) memPtr; } + +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) + +/* + * __attribute__((aligned(1))) is supported by gcc and clang. Originally the + * documentation claimed that it only increased the alignment, but actually it + * can decrease it on gcc, clang, and icc: + * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69502, + * https://gcc.godbolt.org/z/xYez1j67Y. + */ +#ifdef XXH_OLD_NAMES +typedef union { xxh_u32 u32; } __attribute__((__packed__)) unalign; +#endif +static xxh_u32 XXH_read32(const void* ptr) +{ + typedef __attribute__((__aligned__(1))) xxh_u32 xxh_unalign32; + return *((const xxh_unalign32*)ptr); +} + +#else + +/* + * Portable and safe solution. Generally efficient. + * see: https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html + */ +static xxh_u32 XXH_read32(const void* memPtr) +{ + xxh_u32 val; + XXH_memcpy(&val, memPtr, sizeof(val)); + return val; +} + +#endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ + + +/* *** Endianness *** */ + +/*! + * @ingroup tuning + * @def XXH_CPU_LITTLE_ENDIAN + * @brief Whether the target is little endian. + * + * Defined to 1 if the target is little endian, or 0 if it is big endian. + * It can be defined externally, for example on the compiler command line. + * + * If it is not defined, + * a runtime check (which is usually constant folded) is used instead. + * + * @note + * This is not necessarily defined to an integer constant. + * + * @see XXH_isLittleEndian() for the runtime check. + */ +#ifndef XXH_CPU_LITTLE_ENDIAN +/* + * Try to detect endianness automatically, to avoid the nonstandard behavior + * in `XXH_isLittleEndian()` + */ +# if defined(_WIN32) /* Windows is always little endian */ \ + || defined(__LITTLE_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) +# define XXH_CPU_LITTLE_ENDIAN 1 +# elif defined(__BIG_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) +# define XXH_CPU_LITTLE_ENDIAN 0 +# else +/*! + * @internal + * @brief Runtime check for @ref XXH_CPU_LITTLE_ENDIAN. + * + * Most compilers will constant fold this. + */ +static int XXH_isLittleEndian(void) +{ + /* + * Portable and well-defined behavior. + * Don't use static: it is detrimental to performance. + */ + const union { xxh_u32 u; xxh_u8 c[4]; } one = { 1 }; + return one.c[0]; +} +# define XXH_CPU_LITTLE_ENDIAN XXH_isLittleEndian() +# endif +#endif + + + + +/* **************************************** +* Compiler-specific Functions and Macros +******************************************/ +#define XXH_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) + +#ifdef __has_builtin +# define XXH_HAS_BUILTIN(x) __has_builtin(x) +#else +# define XXH_HAS_BUILTIN(x) 0 +#endif + + + +/* + * C23 and future versions have standard "unreachable()". + * Once it has been implemented reliably we can add it as an + * additional case: + * + * ``` + * #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= XXH_C23_VN) + * # include + * # ifdef unreachable + * # define XXH_UNREACHABLE() unreachable() + * # endif + * #endif + * ``` + * + * Note C++23 also has std::unreachable() which can be detected + * as follows: + * ``` + * #if defined(__cpp_lib_unreachable) && (__cpp_lib_unreachable >= 202202L) + * # include + * # define XXH_UNREACHABLE() std::unreachable() + * #endif + * ``` + * NB: `__cpp_lib_unreachable` is defined in the `` header. + * We don't use that as including `` in `extern "C"` blocks + * doesn't work on GCC12 + */ + +#if XXH_HAS_BUILTIN(__builtin_unreachable) +# define XXH_UNREACHABLE() __builtin_unreachable() + +#elif defined(_MSC_VER) +# define XXH_UNREACHABLE() __assume(0) + +#else +# define XXH_UNREACHABLE() +#endif + +#if XXH_HAS_BUILTIN(__builtin_assume) +# define XXH_ASSUME(c) __builtin_assume(c) +#else +# define XXH_ASSUME(c) if (!(c)) { XXH_UNREACHABLE(); } +#endif + +/*! + * @internal + * @def XXH_rotl32(x,r) + * @brief 32-bit rotate left. + * + * @param x The 32-bit integer to be rotated. + * @param r The number of bits to rotate. + * @pre + * @p r > 0 && @p r < 32 + * @note + * @p x and @p r may be evaluated multiple times. + * @return The rotated result. + */ +#if !defined(NO_CLANG_BUILTIN) && XXH_HAS_BUILTIN(__builtin_rotateleft32) \ + && XXH_HAS_BUILTIN(__builtin_rotateleft64) +# define XXH_rotl32 __builtin_rotateleft32 +# define XXH_rotl64 __builtin_rotateleft64 +#elif XXH_HAS_BUILTIN(__builtin_stdc_rotate_left) +# define XXH_rotl32 __builtin_stdc_rotate_left +# define XXH_rotl64 __builtin_stdc_rotate_left +/* Note: although _rotl exists for minGW (GCC under windows), performance seems poor */ +#elif defined(_MSC_VER) +# define XXH_rotl32(x,r) _rotl(x,r) +# define XXH_rotl64(x,r) _rotl64(x,r) +#else +# define XXH_rotl32(x,r) (((x) << (r)) | ((x) >> (32 - (r)))) +# define XXH_rotl64(x,r) (((x) << (r)) | ((x) >> (64 - (r)))) +#endif + +/*! + * @internal + * @fn xxh_u32 XXH_swap32(xxh_u32 x) + * @brief A 32-bit byteswap. + * + * @param x The 32-bit integer to byteswap. + * @return @p x, byteswapped. + */ +#if defined(_MSC_VER) /* Visual Studio */ +# define XXH_swap32 _byteswap_ulong +#elif XXH_GCC_VERSION >= 403 +# define XXH_swap32 __builtin_bswap32 +#else +static xxh_u32 XXH_swap32 (xxh_u32 x) +{ + return ((x << 24) & 0xff000000 ) | + ((x << 8) & 0x00ff0000 ) | + ((x >> 8) & 0x0000ff00 ) | + ((x >> 24) & 0x000000ff ); +} +#endif + + +/* *************************** +* Memory reads +*****************************/ + +/*! + * @internal + * @brief Enum to indicate whether a pointer is aligned. + */ +typedef enum { + XXH_aligned, /*!< Aligned */ + XXH_unaligned /*!< Possibly unaligned */ +} XXH_alignment; + +/* + * XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. + * + * This is ideal for older compilers which don't inline memcpy. + */ +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) + +XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[0] + | ((xxh_u32)bytePtr[1] << 8) + | ((xxh_u32)bytePtr[2] << 16) + | ((xxh_u32)bytePtr[3] << 24); +} + +XXH_FORCE_INLINE xxh_u32 XXH_readBE32(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[3] + | ((xxh_u32)bytePtr[2] << 8) + | ((xxh_u32)bytePtr[1] << 16) + | ((xxh_u32)bytePtr[0] << 24); +} + +#else +XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr)); +} + +static xxh_u32 XXH_readBE32(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr); +} +#endif + +XXH_FORCE_INLINE xxh_u32 +XXH_readLE32_align(const void* ptr, XXH_alignment align) +{ + if (align==XXH_unaligned) { + return XXH_readLE32(ptr); + } else { + return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u32*)ptr : XXH_swap32(*(const xxh_u32*)ptr); + } +} + + +/* ************************************* +* Misc +***************************************/ +/*! @ingroup public */ +XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; } + + +/* ******************************************************************* +* 32-bit hash functions +*********************************************************************/ +/*! + * @} + * @defgroup XXH32_impl XXH32 implementation + * @ingroup impl + * + * Details on the XXH32 implementation. + * @{ + */ + /* #define instead of static const, to be used as initializers */ +#define XXH_PRIME32_1 0x9E3779B1U /*!< 0b10011110001101110111100110110001 */ +#define XXH_PRIME32_2 0x85EBCA77U /*!< 0b10000101111010111100101001110111 */ +#define XXH_PRIME32_3 0xC2B2AE3DU /*!< 0b11000010101100101010111000111101 */ +#define XXH_PRIME32_4 0x27D4EB2FU /*!< 0b00100111110101001110101100101111 */ +#define XXH_PRIME32_5 0x165667B1U /*!< 0b00010110010101100110011110110001 */ + +#ifdef XXH_OLD_NAMES +# define PRIME32_1 XXH_PRIME32_1 +# define PRIME32_2 XXH_PRIME32_2 +# define PRIME32_3 XXH_PRIME32_3 +# define PRIME32_4 XXH_PRIME32_4 +# define PRIME32_5 XXH_PRIME32_5 +#endif + +/*! + * @internal + * @brief Normal stripe processing routine. + * + * This shuffles the bits so that any bit from @p input impacts several bits in + * @p acc. + * + * @param acc The accumulator lane. + * @param input The stripe of input to mix. + * @return The mixed accumulator lane. + */ +static xxh_u32 XXH32_round(xxh_u32 acc, xxh_u32 input) +{ + acc += input * XXH_PRIME32_2; + acc = XXH_rotl32(acc, 13); + acc *= XXH_PRIME32_1; +#if (defined(__SSE4_1__) || defined(__aarch64__) || defined(__wasm_simd128__)) && !defined(XXH_ENABLE_AUTOVECTORIZE) + /* + * UGLY HACK: + * A compiler fence is used to prevent GCC and Clang from + * autovectorizing the XXH32 loop (pragmas and attributes don't work for some + * reason) without globally disabling SSE4.1. + * + * The reason we want to avoid vectorization is because despite working on + * 4 integers at a time, there are multiple factors slowing XXH32 down on + * SSE4: + * - There's a ridiculous amount of lag from pmulld (10 cycles of latency on + * newer chips!) making it slightly slower to multiply four integers at + * once compared to four integers independently. Even when pmulld was + * fastest, Sandy/Ivy Bridge, it is still not worth it to go into SSE + * just to multiply unless doing a long operation. + * + * - Four instructions are required to rotate, + * movqda tmp, v // not required with VEX encoding + * pslld tmp, 13 // tmp <<= 13 + * psrld v, 19 // x >>= 19 + * por v, tmp // x |= tmp + * compared to one for scalar: + * roll v, 13 // reliably fast across the board + * shldl v, v, 13 // Sandy Bridge and later prefer this for some reason + * + * - Instruction level parallelism is actually more beneficial here because + * the SIMD actually serializes this operation: While v1 is rotating, v2 + * can load data, while v3 can multiply. SSE forces them to operate + * together. + * + * This is also enabled on AArch64, as Clang is *very aggressive* in vectorizing + * the loop. NEON is only faster on the A53, and with the newer cores, it is less + * than half the speed. + * + * Additionally, this is used on WASM SIMD128 because it JITs to the same + * SIMD instructions and has the same issue. + */ + XXH_COMPILER_GUARD(acc); +#endif + return acc; +} + +/*! + * @internal + * @brief Mixes all bits to finalize the hash. + * + * The final mix ensures that all input bits have a chance to impact any bit in + * the output digest, resulting in an unbiased distribution. + * + * @param hash The hash to avalanche. + * @return The avalanched hash. + */ +static xxh_u32 XXH32_avalanche(xxh_u32 hash) +{ + hash ^= hash >> 15; + hash *= XXH_PRIME32_2; + hash ^= hash >> 13; + hash *= XXH_PRIME32_3; + hash ^= hash >> 16; + return hash; +} + +#define XXH_get32bits(p) XXH_readLE32_align(p, align) + +/*! + * @internal + * @brief Sets up the initial accumulator state for XXH32(). + */ +XXH_FORCE_INLINE void +XXH32_initAccs(xxh_u32 *acc, xxh_u32 seed) +{ + XXH_ASSERT(acc != NULL); + acc[0] = seed + XXH_PRIME32_1 + XXH_PRIME32_2; + acc[1] = seed + XXH_PRIME32_2; + acc[2] = seed + 0; + acc[3] = seed - XXH_PRIME32_1; +} + +/*! + * @internal + * @brief Consumes a block of data for XXH32(). + * + * @return the end input pointer. + */ +XXH_FORCE_INLINE const xxh_u8 * +XXH32_consumeLong( + xxh_u32 *XXH_RESTRICT acc, + xxh_u8 const *XXH_RESTRICT input, + size_t len, + XXH_alignment align +) +{ + const xxh_u8* const bEnd = input + len; + const xxh_u8* const limit = bEnd - 15; + XXH_ASSERT(acc != NULL); + XXH_ASSERT(input != NULL); + XXH_ASSERT(len >= 16); + do { + acc[0] = XXH32_round(acc[0], XXH_get32bits(input)); input += 4; + acc[1] = XXH32_round(acc[1], XXH_get32bits(input)); input += 4; + acc[2] = XXH32_round(acc[2], XXH_get32bits(input)); input += 4; + acc[3] = XXH32_round(acc[3], XXH_get32bits(input)); input += 4; + } while (input < limit); + + return input; +} + +/*! + * @internal + * @brief Merges the accumulator lanes together for XXH32() + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u32 +XXH32_mergeAccs(const xxh_u32 *acc) +{ + XXH_ASSERT(acc != NULL); + return XXH_rotl32(acc[0], 1) + XXH_rotl32(acc[1], 7) + + XXH_rotl32(acc[2], 12) + XXH_rotl32(acc[3], 18); +} + +/*! + * @internal + * @brief Processes the last 0-15 bytes of @p ptr. + * + * There may be up to 15 bytes remaining to consume from the input. + * This final stage will digest them to ensure that all input bytes are present + * in the final mix. + * + * @param hash The hash to finalize. + * @param ptr The pointer to the remaining input. + * @param len The remaining length, modulo 16. + * @param align Whether @p ptr is aligned. + * @return The finalized hash. + * @see XXH64_finalize(). + */ +static XXH_PUREF xxh_u32 +XXH32_finalize(xxh_u32 hash, const xxh_u8* ptr, size_t len, XXH_alignment align) +{ +#define XXH_PROCESS1 do { \ + hash += (*ptr++) * XXH_PRIME32_5; \ + hash = XXH_rotl32(hash, 11) * XXH_PRIME32_1; \ +} while (0) + +#define XXH_PROCESS4 do { \ + hash += XXH_get32bits(ptr) * XXH_PRIME32_3; \ + ptr += 4; \ + hash = XXH_rotl32(hash, 17) * XXH_PRIME32_4; \ +} while (0) + + if (ptr==NULL) XXH_ASSERT(len == 0); + + /* Compact rerolled version; generally faster */ + if (!XXH32_ENDJMP) { + len &= 15; + while (len >= 4) { + XXH_PROCESS4; + len -= 4; + } + while (len > 0) { + XXH_PROCESS1; + --len; + } + return XXH32_avalanche(hash); + } else { + switch(len&15) /* or switch(bEnd - p) */ { + case 12: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 8: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 4: XXH_PROCESS4; + return XXH32_avalanche(hash); + + case 13: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 9: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 5: XXH_PROCESS4; + XXH_PROCESS1; + return XXH32_avalanche(hash); + + case 14: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 10: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 6: XXH_PROCESS4; + XXH_PROCESS1; + XXH_PROCESS1; + return XXH32_avalanche(hash); + + case 15: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 11: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 7: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 3: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 2: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 1: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 0: return XXH32_avalanche(hash); + } + XXH_ASSERT(0); + return hash; /* reaching this point is deemed impossible */ + } +} + +#ifdef XXH_OLD_NAMES +# define PROCESS1 XXH_PROCESS1 +# define PROCESS4 XXH_PROCESS4 +#else +# undef XXH_PROCESS1 +# undef XXH_PROCESS4 +#endif + +/*! + * @internal + * @brief The implementation for @ref XXH32(). + * + * @param input , len , seed Directly passed from @ref XXH32(). + * @param align Whether @p input is aligned. + * @return The calculated hash. + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u32 +XXH32_endian_align(const xxh_u8* input, size_t len, xxh_u32 seed, XXH_alignment align) +{ + xxh_u32 h32; + + if (input==NULL) XXH_ASSERT(len == 0); + + if (len>=16) { + xxh_u32 acc[4]; + XXH32_initAccs(acc, seed); + + input = XXH32_consumeLong(acc, input, len, align); + + h32 = XXH32_mergeAccs(acc); + } else { + h32 = seed + XXH_PRIME32_5; + } + + h32 += (xxh_u32)len; + + return XXH32_finalize(h32, input, len&15, align); +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32 (const void* input, size_t len, XXH32_hash_t seed) +{ +#if !defined(XXH_NO_STREAM) && XXH_SIZE_OPT >= 2 + /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ + XXH32_state_t state; + XXH32_reset(&state, seed); + XXH32_update(&state, (const xxh_u8*)input, len); + return XXH32_digest(&state); +#else + if (XXH_FORCE_ALIGN_CHECK) { + if ((((size_t)input) & 3) == 0) { /* Input is 4-bytes aligned, leverage the speed benefit */ + return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_aligned); + } } + + return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned); +#endif +} + + + +/******* Hash streaming *******/ +#ifndef XXH_NO_STREAM +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void) +{ + return (XXH32_state_t*)XXH_malloc(sizeof(XXH32_state_t)); +} +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr) +{ + XXH_free(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dstState, const XXH32_state_t* srcState) +{ + XXH_memcpy(dstState, srcState, sizeof(*dstState)); +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, XXH32_hash_t seed) +{ + XXH_ASSERT(statePtr != NULL); + memset(statePtr, 0, sizeof(*statePtr)); + XXH32_initAccs(statePtr->acc, seed); + return XXH_OK; +} + + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode +XXH32_update(XXH32_state_t* state, const void* input, size_t len) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + state->total_len_32 += (XXH32_hash_t)len; + state->large_len |= (XXH32_hash_t)((len>=16) | (state->total_len_32>=16)); + + XXH_ASSERT(state->bufferedSize < sizeof(state->buffer)); + if (len < sizeof(state->buffer) - state->bufferedSize) { /* fill in tmp buffer */ + XXH_memcpy(state->buffer + state->bufferedSize, input, len); + state->bufferedSize += (XXH32_hash_t)len; + return XXH_OK; + } + + { const xxh_u8* xinput = (const xxh_u8*)input; + const xxh_u8* const bEnd = xinput + len; + + if (state->bufferedSize) { /* non-empty buffer: complete first */ + XXH_memcpy(state->buffer + state->bufferedSize, xinput, sizeof(state->buffer) - state->bufferedSize); + xinput += sizeof(state->buffer) - state->bufferedSize; + /* then process one round */ + (void)XXH32_consumeLong(state->acc, state->buffer, sizeof(state->buffer), XXH_aligned); + state->bufferedSize = 0; + } + + XXH_ASSERT(xinput <= bEnd); + if ((size_t)(bEnd - xinput) >= sizeof(state->buffer)) { + /* Process the remaining data */ + xinput = XXH32_consumeLong(state->acc, xinput, (size_t)(bEnd - xinput), XXH_unaligned); + } + + if (xinput < bEnd) { + /* Copy the leftover to the tmp buffer */ + XXH_memcpy(state->buffer, xinput, (size_t)(bEnd-xinput)); + state->bufferedSize = (unsigned)(bEnd-xinput); + } + } + + return XXH_OK; +} + + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32_digest(const XXH32_state_t* state) +{ + xxh_u32 h32; + + if (state->large_len) { + h32 = XXH32_mergeAccs(state->acc); + } else { + h32 = state->acc[2] /* == seed */ + XXH_PRIME32_5; + } + + h32 += state->total_len_32; + + return XXH32_finalize(h32, state->buffer, state->bufferedSize, XXH_aligned); +} +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash); + XXH_memcpy(dst, &hash, sizeof(*dst)); +} +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src) +{ + return XXH_readBE32(src); +} + + +#ifndef XXH_NO_LONG_LONG + +/* ******************************************************************* +* 64-bit hash functions +*********************************************************************/ +/*! + * @} + * @ingroup impl + * @{ + */ +/******* Memory access *******/ + +typedef XXH64_hash_t xxh_u64; + +#ifdef XXH_OLD_NAMES +# define U64 xxh_u64 +#endif + +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) +/* + * Manual byteshift. Best for old compilers which don't inline memcpy. + * We actually directly use XXH_readLE64 and XXH_readBE64. + */ +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) + +/* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */ +static xxh_u64 XXH_read64(const void* memPtr) +{ + return *(const xxh_u64*) memPtr; +} + +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) + +/* + * __attribute__((aligned(1))) is supported by gcc and clang. Originally the + * documentation claimed that it only increased the alignment, but actually it + * can decrease it on gcc, clang, and icc: + * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69502, + * https://gcc.godbolt.org/z/xYez1j67Y. + */ +#ifdef XXH_OLD_NAMES +typedef union { xxh_u32 u32; xxh_u64 u64; } __attribute__((__packed__)) unalign64; +#endif +static xxh_u64 XXH_read64(const void* ptr) +{ + typedef __attribute__((__aligned__(1))) xxh_u64 xxh_unalign64; + return *((const xxh_unalign64*)ptr); +} + +#else + +/* + * Portable and safe solution. Generally efficient. + * see: https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html + */ +static xxh_u64 XXH_read64(const void* memPtr) +{ + xxh_u64 val; + XXH_memcpy(&val, memPtr, sizeof(val)); + return val; +} + +#endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ + +#if defined(_MSC_VER) /* Visual Studio */ +# define XXH_swap64 _byteswap_uint64 +#elif XXH_GCC_VERSION >= 403 +# define XXH_swap64 __builtin_bswap64 +#else +static xxh_u64 XXH_swap64(xxh_u64 x) +{ + return ((x << 56) & 0xff00000000000000ULL) | + ((x << 40) & 0x00ff000000000000ULL) | + ((x << 24) & 0x0000ff0000000000ULL) | + ((x << 8) & 0x000000ff00000000ULL) | + ((x >> 8) & 0x00000000ff000000ULL) | + ((x >> 24) & 0x0000000000ff0000ULL) | + ((x >> 40) & 0x000000000000ff00ULL) | + ((x >> 56) & 0x00000000000000ffULL); +} +#endif + + +/* XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. */ +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) + +XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[0] + | ((xxh_u64)bytePtr[1] << 8) + | ((xxh_u64)bytePtr[2] << 16) + | ((xxh_u64)bytePtr[3] << 24) + | ((xxh_u64)bytePtr[4] << 32) + | ((xxh_u64)bytePtr[5] << 40) + | ((xxh_u64)bytePtr[6] << 48) + | ((xxh_u64)bytePtr[7] << 56); +} + +XXH_FORCE_INLINE xxh_u64 XXH_readBE64(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[7] + | ((xxh_u64)bytePtr[6] << 8) + | ((xxh_u64)bytePtr[5] << 16) + | ((xxh_u64)bytePtr[4] << 24) + | ((xxh_u64)bytePtr[3] << 32) + | ((xxh_u64)bytePtr[2] << 40) + | ((xxh_u64)bytePtr[1] << 48) + | ((xxh_u64)bytePtr[0] << 56); +} + +#else +XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr)); +} + +static xxh_u64 XXH_readBE64(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr); +} +#endif + +XXH_FORCE_INLINE xxh_u64 +XXH_readLE64_align(const void* ptr, XXH_alignment align) +{ + if (align==XXH_unaligned) + return XXH_readLE64(ptr); + else + return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u64*)ptr : XXH_swap64(*(const xxh_u64*)ptr); +} + + +/******* xxh64 *******/ +/*! + * @} + * @defgroup XXH64_impl XXH64 implementation + * @ingroup impl + * + * Details on the XXH64 implementation. + * @{ + */ +/* #define rather that static const, to be used as initializers */ +#define XXH_PRIME64_1 0x9E3779B185EBCA87ULL /*!< 0b1001111000110111011110011011000110000101111010111100101010000111 */ +#define XXH_PRIME64_2 0xC2B2AE3D27D4EB4FULL /*!< 0b1100001010110010101011100011110100100111110101001110101101001111 */ +#define XXH_PRIME64_3 0x165667B19E3779F9ULL /*!< 0b0001011001010110011001111011000110011110001101110111100111111001 */ +#define XXH_PRIME64_4 0x85EBCA77C2B2AE63ULL /*!< 0b1000010111101011110010100111011111000010101100101010111001100011 */ +#define XXH_PRIME64_5 0x27D4EB2F165667C5ULL /*!< 0b0010011111010100111010110010111100010110010101100110011111000101 */ + +#ifdef XXH_OLD_NAMES +# define PRIME64_1 XXH_PRIME64_1 +# define PRIME64_2 XXH_PRIME64_2 +# define PRIME64_3 XXH_PRIME64_3 +# define PRIME64_4 XXH_PRIME64_4 +# define PRIME64_5 XXH_PRIME64_5 +#endif + +/*! @copydoc XXH32_round */ +static xxh_u64 XXH64_round(xxh_u64 acc, xxh_u64 input) +{ + acc += input * XXH_PRIME64_2; + acc = XXH_rotl64(acc, 31); + acc *= XXH_PRIME64_1; +#if (defined(__AVX512F__)) && !defined(XXH_ENABLE_AUTOVECTORIZE) + /* + * DISABLE AUTOVECTORIZATION: + * A compiler fence is used to prevent GCC and Clang from + * autovectorizing the XXH64 loop (pragmas and attributes don't work for some + * reason) without globally disabling AVX512. + * + * Autovectorization of XXH64 tends to be detrimental, + * though the exact outcome may change depending on exact cpu and compiler version. + * For information, it has been reported as detrimental for Skylake-X, + * but possibly beneficial for Zen4. + * + * The default is to disable auto-vectorization, + * but you can select to enable it instead using `XXH_ENABLE_AUTOVECTORIZE` build variable. + */ + XXH_COMPILER_GUARD(acc); +#endif + return acc; +} + +static xxh_u64 XXH64_mergeRound(xxh_u64 acc, xxh_u64 val) +{ + val = XXH64_round(0, val); + acc ^= val; + acc = acc * XXH_PRIME64_1 + XXH_PRIME64_4; + return acc; +} + +/*! @copydoc XXH32_avalanche */ +static xxh_u64 XXH64_avalanche(xxh_u64 hash) +{ + hash ^= hash >> 33; + hash *= XXH_PRIME64_2; + hash ^= hash >> 29; + hash *= XXH_PRIME64_3; + hash ^= hash >> 32; + return hash; +} + + +#define XXH_get64bits(p) XXH_readLE64_align(p, align) + +/*! + * @internal + * @brief Sets up the initial accumulator state for XXH64(). + */ +XXH_FORCE_INLINE void +XXH64_initAccs(xxh_u64 *acc, xxh_u64 seed) +{ + XXH_ASSERT(acc != NULL); + acc[0] = seed + XXH_PRIME64_1 + XXH_PRIME64_2; + acc[1] = seed + XXH_PRIME64_2; + acc[2] = seed + 0; + acc[3] = seed - XXH_PRIME64_1; +} + +/*! + * @internal + * @brief Consumes a block of data for XXH64(). + * + * @return the end input pointer. + */ +XXH_FORCE_INLINE const xxh_u8 * +XXH64_consumeLong( + xxh_u64 *XXH_RESTRICT acc, + xxh_u8 const *XXH_RESTRICT input, + size_t len, + XXH_alignment align +) +{ + const xxh_u8* const bEnd = input + len; + const xxh_u8* const limit = bEnd - 31; + XXH_ASSERT(acc != NULL); + XXH_ASSERT(input != NULL); + XXH_ASSERT(len >= 32); + do { + /* reroll on 32-bit */ + if (sizeof(void *) < sizeof(xxh_u64)) { + size_t i; + for (i = 0; i < 4; i++) { + acc[i] = XXH64_round(acc[i], XXH_get64bits(input)); + input += 8; + } + } else { + acc[0] = XXH64_round(acc[0], XXH_get64bits(input)); input += 8; + acc[1] = XXH64_round(acc[1], XXH_get64bits(input)); input += 8; + acc[2] = XXH64_round(acc[2], XXH_get64bits(input)); input += 8; + acc[3] = XXH64_round(acc[3], XXH_get64bits(input)); input += 8; + } + } while (input < limit); + + return input; +} + +/*! + * @internal + * @brief Merges the accumulator lanes together for XXH64() + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u64 +XXH64_mergeAccs(const xxh_u64 *acc) +{ + XXH_ASSERT(acc != NULL); + { + xxh_u64 h64 = XXH_rotl64(acc[0], 1) + XXH_rotl64(acc[1], 7) + + XXH_rotl64(acc[2], 12) + XXH_rotl64(acc[3], 18); + /* reroll on 32-bit */ + if (sizeof(void *) < sizeof(xxh_u64)) { + size_t i; + for (i = 0; i < 4; i++) { + h64 = XXH64_mergeRound(h64, acc[i]); + } + } else { + h64 = XXH64_mergeRound(h64, acc[0]); + h64 = XXH64_mergeRound(h64, acc[1]); + h64 = XXH64_mergeRound(h64, acc[2]); + h64 = XXH64_mergeRound(h64, acc[3]); + } + return h64; + } +} + +/*! + * @internal + * @brief Processes the last 0-31 bytes of @p ptr. + * + * There may be up to 31 bytes remaining to consume from the input. + * This final stage will digest them to ensure that all input bytes are present + * in the final mix. + * + * @param hash The hash to finalize. + * @param ptr The pointer to the remaining input. + * @param len The remaining length, modulo 32. + * @param align Whether @p ptr is aligned. + * @return The finalized hash + * @see XXH32_finalize(). + */ +XXH_STATIC XXH_PUREF xxh_u64 +XXH64_finalize(xxh_u64 hash, const xxh_u8* ptr, size_t len, XXH_alignment align) +{ + if (ptr==NULL) XXH_ASSERT(len == 0); + len &= 31; + while (len >= 8) { + xxh_u64 const k1 = XXH64_round(0, XXH_get64bits(ptr)); + ptr += 8; + hash ^= k1; + hash = XXH_rotl64(hash,27) * XXH_PRIME64_1 + XXH_PRIME64_4; + len -= 8; + } + if (len >= 4) { + hash ^= (xxh_u64)(XXH_get32bits(ptr)) * XXH_PRIME64_1; + ptr += 4; + hash = XXH_rotl64(hash, 23) * XXH_PRIME64_2 + XXH_PRIME64_3; + len -= 4; + } + while (len > 0) { + hash ^= (*ptr++) * XXH_PRIME64_5; + hash = XXH_rotl64(hash, 11) * XXH_PRIME64_1; + --len; + } + return XXH64_avalanche(hash); +} + +#ifdef XXH_OLD_NAMES +# define PROCESS1_64 XXH_PROCESS1_64 +# define PROCESS4_64 XXH_PROCESS4_64 +# define PROCESS8_64 XXH_PROCESS8_64 +#else +# undef XXH_PROCESS1_64 +# undef XXH_PROCESS4_64 +# undef XXH_PROCESS8_64 +#endif + +/*! + * @internal + * @brief The implementation for @ref XXH64(). + * + * @param input , len , seed Directly passed from @ref XXH64(). + * @param align Whether @p input is aligned. + * @return The calculated hash. + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u64 +XXH64_endian_align(const xxh_u8* input, size_t len, xxh_u64 seed, XXH_alignment align) +{ + xxh_u64 h64; + if (input==NULL) XXH_ASSERT(len == 0); + + if (len>=32) { /* Process a large block of data */ + xxh_u64 acc[4]; + XXH64_initAccs(acc, seed); + + input = XXH64_consumeLong(acc, input, len, align); + + h64 = XXH64_mergeAccs(acc); + } else { + h64 = seed + XXH_PRIME64_5; + } + + h64 += (xxh_u64) len; + + return XXH64_finalize(h64, input, len, align); +} + + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH64_hash_t XXH64 (XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) +{ +#if !defined(XXH_NO_STREAM) && XXH_SIZE_OPT >= 2 + /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ + XXH64_state_t state; + XXH64_reset(&state, seed); + XXH64_update(&state, (const xxh_u8*)input, len); + return XXH64_digest(&state); +#else + if (XXH_FORCE_ALIGN_CHECK) { + if ((((size_t)input) & 7)==0) { /* Input is aligned, let's leverage the speed advantage */ + return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_aligned); + } } + + return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned); + +#endif +} + +/******* Hash Streaming *******/ +#ifndef XXH_NO_STREAM +/*! @ingroup XXH64_family*/ +XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void) +{ + return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t)); +} +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr) +{ + XXH_free(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dstState, const XXH64_state_t* srcState) +{ + XXH_memcpy(dstState, srcState, sizeof(*dstState)); +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed) +{ + XXH_ASSERT(statePtr != NULL); + memset(statePtr, 0, sizeof(*statePtr)); + XXH64_initAccs(statePtr->acc, seed); + return XXH_OK; +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode +XXH64_update (XXH_NOESCAPE XXH64_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + state->total_len += len; + + XXH_ASSERT(state->bufferedSize <= sizeof(state->buffer)); + if (len < sizeof(state->buffer) - state->bufferedSize) { /* fill in tmp buffer */ + XXH_memcpy(state->buffer + state->bufferedSize, input, len); + state->bufferedSize += (XXH32_hash_t)len; + return XXH_OK; + } + + { const xxh_u8* xinput = (const xxh_u8*)input; + const xxh_u8* const bEnd = xinput + len; + + if (state->bufferedSize) { /* non-empty buffer => complete first */ + XXH_memcpy(state->buffer + state->bufferedSize, xinput, sizeof(state->buffer) - state->bufferedSize); + xinput += sizeof(state->buffer) - state->bufferedSize; + /* and process one round */ + (void)XXH64_consumeLong(state->acc, state->buffer, sizeof(state->buffer), XXH_aligned); + state->bufferedSize = 0; + } + + XXH_ASSERT(xinput <= bEnd); + if ((size_t)(bEnd - xinput) >= sizeof(state->buffer)) { + /* Process the remaining data */ + xinput = XXH64_consumeLong(state->acc, xinput, (size_t)(bEnd - xinput), XXH_unaligned); + } + + if (xinput < bEnd) { + /* Copy the leftover to the tmp buffer */ + XXH_memcpy(state->buffer, xinput, (size_t)(bEnd-xinput)); + state->bufferedSize = (unsigned)(bEnd-xinput); + } + } + + return XXH_OK; +} + + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH64_hash_t XXH64_digest(XXH_NOESCAPE const XXH64_state_t* state) +{ + xxh_u64 h64; + + if (state->total_len >= 32) { + h64 = XXH64_mergeAccs(state->acc); + } else { + h64 = state->acc[2] /*seed*/ + XXH_PRIME64_5; + } + + h64 += (xxh_u64) state->total_len; + + return XXH64_finalize(h64, state->buffer, (size_t)state->total_len, XXH_aligned); +} +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH_NOESCAPE XXH64_canonical_t* dst, XXH64_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH64_canonical_t) == sizeof(XXH64_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap64(hash); + XXH_memcpy(dst, &hash, sizeof(*dst)); +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(XXH_NOESCAPE const XXH64_canonical_t* src) +{ + return XXH_readBE64(src); +} + +#ifndef XXH_NO_XXH3 + +/* ********************************************************************* +* XXH3 +* New generation hash designed for speed on small keys and vectorization +************************************************************************ */ +/*! + * @} + * @defgroup XXH3_impl XXH3 implementation + * @ingroup impl + * @{ + */ + +/* === Compiler specifics === */ + + +#if (defined(__GNUC__) && (__GNUC__ >= 3)) \ + || (defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 800)) \ + || defined(__clang__) +# define XXH_likely(x) __builtin_expect(x, 1) +# define XXH_unlikely(x) __builtin_expect(x, 0) +#else +# define XXH_likely(x) (x) +# define XXH_unlikely(x) (x) +#endif + +#ifndef XXH_HAS_INCLUDE +# ifdef __has_include +/* + * Not defined as XXH_HAS_INCLUDE(x) (function-like) because + * this causes segfaults in Apple Clang 4.2 (on Mac OS X 10.7 Lion) + */ +# define XXH_HAS_INCLUDE __has_include +# else +# define XXH_HAS_INCLUDE(x) 0 +# endif +#endif + +#if defined(__GNUC__) || defined(__clang__) +# if defined(__ARM_FEATURE_SVE) +# include +# endif +# if defined(__ARM_NEON__) || defined(__ARM_NEON) \ + || (defined(_M_ARM) && _M_ARM >= 7) \ + || defined(_M_ARM64) || defined(_M_ARM64EC) \ + || (defined(__wasm_simd128__) && XXH_HAS_INCLUDE()) /* WASM SIMD128 via SIMDe */ +# define inline __inline__ /* circumvent a clang bug */ +# include +# undef inline +# elif defined(__AVX2__) +# include +# elif defined(__SSE2__) +# include +# elif defined(__loongarch_sx) +# include +# endif +#endif + +#if defined(_MSC_VER) +# include +#endif + +/* + * One goal of XXH3 is to make it fast on both 32-bit and 64-bit, while + * remaining a true 64-bit/128-bit hash function. + * + * This is done by prioritizing a subset of 64-bit operations that can be + * emulated without too many steps on the average 32-bit machine. + * + * For example, these two lines seem similar, and run equally fast on 64-bit: + * + * xxh_u64 x; + * x ^= (x >> 47); // good + * x ^= (x >> 13); // bad + * + * However, to a 32-bit machine, there is a major difference. + * + * x ^= (x >> 47) looks like this: + * + * x.lo ^= (x.hi >> (47 - 32)); + * + * while x ^= (x >> 13) looks like this: + * + * // note: funnel shifts are not usually cheap. + * x.lo ^= (x.lo >> 13) | (x.hi << (32 - 13)); + * x.hi ^= (x.hi >> 13); + * + * The first one is significantly faster than the second, simply because the + * shift is larger than 32. This means: + * - All the bits we need are in the upper 32 bits, so we can ignore the lower + * 32 bits in the shift. + * - The shift result will always fit in the lower 32 bits, and therefore, + * we can ignore the upper 32 bits in the xor. + * + * Thanks to this optimization, XXH3 only requires these features to be efficient: + * + * - Usable unaligned access + * - A 32-bit or 64-bit ALU + * - If 32-bit, a decent ADC instruction + * - A 32 or 64-bit multiply with a 64-bit result + * - For the 128-bit variant, a decent byteswap helps short inputs. + * + * The first two are already required by XXH32, and almost all 32-bit and 64-bit + * platforms which can run XXH32 can run XXH3 efficiently. + * + * Thumb-1, the classic 16-bit only subset of ARM's instruction set, is one + * notable exception. + * + * First of all, Thumb-1 lacks support for the UMULL instruction which + * performs the important long multiply. This means numerous __aeabi_lmul + * calls. + * + * Second of all, the 8 functional registers are just not enough. + * Setup for __aeabi_lmul, byteshift loads, pointers, and all arithmetic need + * Lo registers, and this shuffling results in thousands more MOVs than A32. + * + * A32 and T32 don't have this limitation. They can access all 14 registers, + * do a 32->64 multiply with UMULL, and the flexible operand allowing free + * shifts is helpful, too. + * + * Therefore, we do a quick sanity check. + * + * If compiling Thumb-1 for a target which supports ARM instructions, we will + * emit a warning, as it is not a "sane" platform to compile for. + * + * Usually, if this happens, it is because of an accident and you probably need + * to specify -march, as you likely meant to compile for a newer architecture. + * + * Credit: large sections of the vectorial and asm source code paths + * have been contributed by @easyaspi314 + */ +#if defined(__thumb__) && !defined(__thumb2__) && defined(__ARM_ARCH_ISA_ARM) +# warning "XXH3 is highly inefficient without ARM or Thumb-2." +#endif + +/* ========================================== + * Vectorization detection + * ========================================== */ + +#ifdef XXH_DOXYGEN +/*! + * @ingroup tuning + * @brief Overrides the vectorization implementation chosen for XXH3. + * + * Can be defined to 0 to disable SIMD or any of the values mentioned in + * @ref XXH_VECTOR_TYPE. + * + * If this is not defined, it uses predefined macros to determine the best + * implementation. + */ +# define XXH_VECTOR XXH_SCALAR +/*! + * @ingroup tuning + * @brief Selects the minimum alignment for XXH3's accumulators. + * + * When using SIMD, this should match the alignment required for said vector + * type, so, for example, 32 for AVX2. + * + * Default: Auto detected. + */ +# define XXH_ACC_ALIGN 8 +#endif + +/* Actual definition */ +#ifndef XXH_DOXYGEN +#endif + +#ifndef XXH_VECTOR /* can be defined on command line */ +# if defined(__ARM_FEATURE_SVE) +# define XXH_VECTOR XXH_SVE +# elif ( \ + defined(__ARM_NEON__) || defined(__ARM_NEON) /* gcc */ \ + || defined(_M_ARM) || defined(_M_ARM64) || defined(_M_ARM64EC) /* msvc */ \ + || (defined(__wasm_simd128__) && XXH_HAS_INCLUDE()) /* wasm simd128 via SIMDe */ \ + ) && ( \ + defined(_WIN32) || defined(__LITTLE_ENDIAN__) /* little endian only */ \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) \ + ) +# define XXH_VECTOR XXH_NEON +# elif defined(__AVX512F__) +# define XXH_VECTOR XXH_AVX512 +# elif defined(__AVX2__) +# define XXH_VECTOR XXH_AVX2 +# elif defined(__SSE2__) || defined(_M_AMD64) || defined(_M_X64) || (defined(_M_IX86_FP) && (_M_IX86_FP == 2)) +# define XXH_VECTOR XXH_SSE2 +# elif (defined(__PPC64__) && defined(__POWER8_VECTOR__)) \ + || (defined(__s390x__) && defined(__VEC__)) \ + && defined(__GNUC__) /* TODO: IBM XL */ +# define XXH_VECTOR XXH_VSX +# elif defined(__loongarch_sx) +# define XXH_VECTOR XXH_LSX +# else +# define XXH_VECTOR XXH_SCALAR +# endif +#endif + +/* __ARM_FEATURE_SVE is only supported by GCC & Clang. */ +#if (XXH_VECTOR == XXH_SVE) && !defined(__ARM_FEATURE_SVE) +# ifdef _MSC_VER +# pragma warning(once : 4606) +# else +# warning "__ARM_FEATURE_SVE isn't supported. Use SCALAR instead." +# endif +# undef XXH_VECTOR +# define XXH_VECTOR XXH_SCALAR +#endif + +/* + * Controls the alignment of the accumulator, + * for compatibility with aligned vector loads, which are usually faster. + */ +#ifndef XXH_ACC_ALIGN +# if defined(XXH_X86DISPATCH) +# define XXH_ACC_ALIGN 64 /* for compatibility with avx512 */ +# elif XXH_VECTOR == XXH_SCALAR /* scalar */ +# define XXH_ACC_ALIGN 8 +# elif XXH_VECTOR == XXH_SSE2 /* sse2 */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_AVX2 /* avx2 */ +# define XXH_ACC_ALIGN 32 +# elif XXH_VECTOR == XXH_NEON /* neon */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_VSX /* vsx */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_AVX512 /* avx512 */ +# define XXH_ACC_ALIGN 64 +# elif XXH_VECTOR == XXH_SVE /* sve */ +# define XXH_ACC_ALIGN 64 +# elif XXH_VECTOR == XXH_LSX /* lsx */ +# define XXH_ACC_ALIGN 64 +# endif +#endif + +#if defined(XXH_X86DISPATCH) || XXH_VECTOR == XXH_SSE2 \ + || XXH_VECTOR == XXH_AVX2 || XXH_VECTOR == XXH_AVX512 +# define XXH_SEC_ALIGN XXH_ACC_ALIGN +#elif XXH_VECTOR == XXH_SVE +# define XXH_SEC_ALIGN XXH_ACC_ALIGN +#else +# define XXH_SEC_ALIGN 8 +#endif + +#if defined(__GNUC__) || defined(__clang__) +# define XXH_ALIASING __attribute__((__may_alias__)) +#else +# define XXH_ALIASING /* nothing */ +#endif + +/* + * UGLY HACK: + * GCC usually generates the best code with -O3 for xxHash. + * + * However, when targeting AVX2, it is overzealous in its unrolling resulting + * in code roughly 3/4 the speed of Clang. + * + * There are other issues, such as GCC splitting _mm256_loadu_si256 into + * _mm_loadu_si128 + _mm256_inserti128_si256. This is an optimization which + * only applies to Sandy and Ivy Bridge... which don't even support AVX2. + * + * That is why when compiling the AVX2 version, it is recommended to use either + * -O2 -mavx2 -march=haswell + * or + * -O2 -mavx2 -mno-avx256-split-unaligned-load + * for decent performance, or to use Clang instead. + * + * Fortunately, we can control the first one with a pragma that forces GCC into + * -O2, but the other one we can't control without "failed to inline always + * inline function due to target mismatch" warnings. + */ +#if XXH_VECTOR == XXH_AVX2 /* AVX2 */ \ + && defined(__GNUC__) && !defined(__clang__) /* GCC, not Clang */ \ + && defined(__OPTIMIZE__) && XXH_SIZE_OPT <= 0 /* respect -O0 and -Os */ +# pragma GCC push_options +# pragma GCC optimize("-O2") +#endif + +#if XXH_VECTOR == XXH_NEON + +/* + * UGLY HACK: While AArch64 GCC on Linux does not seem to care, on macOS, GCC -O3 + * optimizes out the entire hashLong loop because of the aliasing violation. + * + * However, GCC is also inefficient at load-store optimization with vld1q/vst1q, + * so the only option is to mark it as aliasing. + */ +typedef uint64x2_t xxh_aliasing_uint64x2_t XXH_ALIASING; + +/*! + * @internal + * @brief `vld1q_u64` but faster and alignment-safe. + * + * On AArch64, unaligned access is always safe, but on ARMv7-a, it is only + * *conditionally* safe (`vld1` has an alignment bit like `movdq[ua]` in x86). + * + * GCC for AArch64 sees `vld1q_u8` as an intrinsic instead of a load, so it + * prohibits load-store optimizations. Therefore, a direct dereference is used. + * + * Otherwise, `vld1q_u8` is used with `vreinterpretq_u8_u64` to do a safe + * unaligned load. + */ +#if defined(__aarch64__) && defined(__GNUC__) && !defined(__clang__) +XXH_FORCE_INLINE uint64x2_t XXH_vld1q_u64(void const* ptr) /* silence -Wcast-align */ +{ + return *(xxh_aliasing_uint64x2_t const *)ptr; +} +#else +XXH_FORCE_INLINE uint64x2_t XXH_vld1q_u64(void const* ptr) +{ + return vreinterpretq_u64_u8(vld1q_u8((uint8_t const*)ptr)); +} +#endif + +/*! + * @internal + * @brief `vmlal_u32` on low and high halves of a vector. + * + * This is a workaround for AArch64 GCC < 11 which implemented arm_neon.h with + * inline assembly and were therefore incapable of merging the `vget_{low, high}_u32` + * with `vmlal_u32`. + */ +#if defined(__aarch64__) && defined(__GNUC__) && !defined(__clang__) && __GNUC__ < 11 +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_low_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + /* Inline assembly is the only way */ + __asm__("umlal %0.2d, %1.2s, %2.2s" : "+w" (acc) : "w" (lhs), "w" (rhs)); + return acc; +} +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_high_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + /* This intrinsic works as expected */ + return vmlal_high_u32(acc, lhs, rhs); +} +#else +/* Portable intrinsic versions */ +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_low_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + return vmlal_u32(acc, vget_low_u32(lhs), vget_low_u32(rhs)); +} +/*! @copydoc XXH_vmlal_low_u32 + * Assume the compiler converts this to vmlal_high_u32 on aarch64 */ +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_high_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + return vmlal_u32(acc, vget_high_u32(lhs), vget_high_u32(rhs)); +} +#endif + +/*! + * @ingroup tuning + * @brief Controls the NEON to scalar ratio for XXH3 + * + * This can be set to 2, 4, 6, or 8. + * + * ARM Cortex CPUs are _very_ sensitive to how their pipelines are used. + * + * For example, the Cortex-A73 can dispatch 3 micro-ops per cycle, but only 2 of those + * can be NEON. If you are only using NEON instructions, you are only using 2/3 of the CPU + * bandwidth. + * + * This is even more noticeable on the more advanced cores like the Cortex-A76 which + * can dispatch 8 micro-ops per cycle, but still only 2 NEON micro-ops at once. + * + * Therefore, to make the most out of the pipeline, it is beneficial to run 6 NEON lanes + * and 2 scalar lanes, which is chosen by default. + * + * This does not apply to Apple processors or 32-bit processors, which run better with + * full NEON. These will default to 8. Additionally, size-optimized builds run 8 lanes. + * + * This change benefits CPUs with large micro-op buffers without negatively affecting + * most other CPUs: + * + * | Chipset | Dispatch type | NEON only | 6:2 hybrid | Diff. | + * |:----------------------|:--------------------|----------:|-----------:|------:| + * | Snapdragon 730 (A76) | 2 NEON/8 micro-ops | 8.8 GB/s | 10.1 GB/s | ~16% | + * | Snapdragon 835 (A73) | 2 NEON/3 micro-ops | 5.1 GB/s | 5.3 GB/s | ~5% | + * | Marvell PXA1928 (A53) | In-order dual-issue | 1.9 GB/s | 1.9 GB/s | 0% | + * | Apple M1 | 4 NEON/8 micro-ops | 37.3 GB/s | 36.1 GB/s | ~-3% | + * + * It also seems to fix some bad codegen on GCC, making it almost as fast as clang. + * + * When using WASM SIMD128, if this is 2 or 6, SIMDe will scalarize 2 of the lanes meaning + * it effectively becomes worse 4. + * + * @see XXH3_accumulate_512_neon() + */ +# ifndef XXH3_NEON_LANES +# if (defined(__aarch64__) || defined(__arm64__) || defined(_M_ARM64) || defined(_M_ARM64EC)) \ + && !defined(__APPLE__) && XXH_SIZE_OPT <= 0 +# define XXH3_NEON_LANES 6 +# else +# define XXH3_NEON_LANES XXH_ACC_NB +# endif +# endif +#endif /* XXH_VECTOR == XXH_NEON */ + +/* + * VSX and Z Vector helpers. + * + * This is very messy, and any pull requests to clean this up are welcome. + * + * There are a lot of problems with supporting VSX and s390x, due to + * inconsistent intrinsics, spotty coverage, and multiple endiannesses. + */ +#if XXH_VECTOR == XXH_VSX +/* Annoyingly, these headers _may_ define three macros: `bool`, `vector`, + * and `pixel`. This is a problem for obvious reasons. + * + * These keywords are unnecessary; the spec literally says they are + * equivalent to `__bool`, `__vector`, and `__pixel` and may be undef'd + * after including the header. + * + * We use pragma push_macro/pop_macro to keep the namespace clean. */ +# pragma push_macro("bool") +# pragma push_macro("vector") +# pragma push_macro("pixel") +/* silence potential macro redefined warnings */ +# undef bool +# undef vector +# undef pixel + +# if defined(__s390x__) +# include +# else +# include +# endif + +/* Restore the original macro values, if applicable. */ +# pragma pop_macro("pixel") +# pragma pop_macro("vector") +# pragma pop_macro("bool") + +typedef __vector unsigned long long xxh_u64x2; +typedef __vector unsigned char xxh_u8x16; +typedef __vector unsigned xxh_u32x4; + +/* + * UGLY HACK: Similar to aarch64 macOS GCC, s390x GCC has the same aliasing issue. + */ +typedef xxh_u64x2 xxh_aliasing_u64x2 XXH_ALIASING; + +# ifndef XXH_VSX_BE +# if defined(__BIG_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) +# define XXH_VSX_BE 1 +# elif defined(__VEC_ELEMENT_REG_ORDER__) && __VEC_ELEMENT_REG_ORDER__ == __ORDER_BIG_ENDIAN__ +# warning "-maltivec=be is not recommended. Please use native endianness." +# define XXH_VSX_BE 1 +# else +# define XXH_VSX_BE 0 +# endif +# endif /* !defined(XXH_VSX_BE) */ + +# if XXH_VSX_BE +# if defined(__POWER9_VECTOR__) || (defined(__clang__) && defined(__s390x__)) +# define XXH_vec_revb vec_revb +# else +/*! + * A polyfill for POWER9's vec_revb(). + */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_revb(xxh_u64x2 val) +{ + xxh_u8x16 const vByteSwap = { 0x07, 0x06, 0x05, 0x04, 0x03, 0x02, 0x01, 0x00, + 0x0F, 0x0E, 0x0D, 0x0C, 0x0B, 0x0A, 0x09, 0x08 }; + return vec_perm(val, val, vByteSwap); +} +# endif +# endif /* XXH_VSX_BE */ + +/*! + * Performs an unaligned vector load and byte swaps it on big endian. + */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_loadu(const void *ptr) +{ + xxh_u64x2 ret; + XXH_memcpy(&ret, ptr, sizeof(xxh_u64x2)); +# if XXH_VSX_BE + ret = XXH_vec_revb(ret); +# endif + return ret; +} + +/* + * vec_mulo and vec_mule are very problematic intrinsics on PowerPC + * + * These intrinsics weren't added until GCC 8, despite existing for a while, + * and they are endian dependent. Also, their meaning swap depending on version. + * */ +# if defined(__s390x__) + /* s390x is always big endian, no issue on this platform */ +# define XXH_vec_mulo vec_mulo +# define XXH_vec_mule vec_mule +# elif defined(__clang__) && XXH_HAS_BUILTIN(__builtin_altivec_vmuleuw) && !defined(__ibmxl__) +/* Clang has a better way to control this, we can just use the builtin which doesn't swap. */ + /* The IBM XL Compiler (which defined __clang__) only implements the vec_* operations */ +# define XXH_vec_mulo __builtin_altivec_vmulouw +# define XXH_vec_mule __builtin_altivec_vmuleuw +# else +/* gcc needs inline assembly */ +/* Adapted from https://github.com/google/highwayhash/blob/master/highwayhash/hh_vsx.h. */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_mulo(xxh_u32x4 a, xxh_u32x4 b) +{ + xxh_u64x2 result; + __asm__("vmulouw %0, %1, %2" : "=v" (result) : "v" (a), "v" (b)); + return result; +} +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_mule(xxh_u32x4 a, xxh_u32x4 b) +{ + xxh_u64x2 result; + __asm__("vmuleuw %0, %1, %2" : "=v" (result) : "v" (a), "v" (b)); + return result; +} +# endif /* XXH_vec_mulo, XXH_vec_mule */ +#endif /* XXH_VECTOR == XXH_VSX */ + +#if XXH_VECTOR == XXH_SVE +#define ACCRND(acc, offset) \ +do { \ + svuint64_t input_vec = svld1_u64(mask, xinput + offset); \ + svuint64_t secret_vec = svld1_u64(mask, xsecret + offset); \ + svuint64_t mixed = sveor_u64_x(mask, secret_vec, input_vec); \ + svuint64_t swapped = svtbl_u64(input_vec, kSwap); \ + svuint64_t mixed_lo = svextw_u64_x(mask, mixed); \ + svuint64_t mixed_hi = svlsr_n_u64_x(mask, mixed, 32); \ + svuint64_t mul = svmad_u64_x(mask, mixed_lo, mixed_hi, swapped); \ + acc = svadd_u64_x(mask, acc, mul); \ +} while (0) +#endif /* XXH_VECTOR == XXH_SVE */ + +/* prefetch + * can be disabled, by declaring XXH_NO_PREFETCH build macro */ +#if defined(XXH_NO_PREFETCH) +# define XXH_PREFETCH(ptr) (void)(ptr) /* disabled */ +#else +# if XXH_SIZE_OPT >= 1 +# define XXH_PREFETCH(ptr) (void)(ptr) +# elif defined(_MSC_VER) && (defined(_M_X64) || defined(_M_IX86)) /* _mm_prefetch() not defined outside of x86/x64 */ +# include /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */ +# define XXH_PREFETCH(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T0) +# elif defined(__GNUC__) && ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) ) +# define XXH_PREFETCH(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */) +# else +# define XXH_PREFETCH(ptr) (void)(ptr) /* disabled */ +# endif +#endif /* XXH_NO_PREFETCH */ + + +/* ========================================== + * XXH3 default settings + * ========================================== */ + +#define XXH_SECRET_DEFAULT_SIZE 192 /* minimum XXH3_SECRET_SIZE_MIN */ + +#if (XXH_SECRET_DEFAULT_SIZE < XXH3_SECRET_SIZE_MIN) +# error "default keyset is not large enough" +#endif + +/*! Pseudorandom secret taken directly from FARSH. */ +XXH_ALIGN(64) static const xxh_u8 XXH3_kSecret[XXH_SECRET_DEFAULT_SIZE] = { + 0xb8, 0xfe, 0x6c, 0x39, 0x23, 0xa4, 0x4b, 0xbe, 0x7c, 0x01, 0x81, 0x2c, 0xf7, 0x21, 0xad, 0x1c, + 0xde, 0xd4, 0x6d, 0xe9, 0x83, 0x90, 0x97, 0xdb, 0x72, 0x40, 0xa4, 0xa4, 0xb7, 0xb3, 0x67, 0x1f, + 0xcb, 0x79, 0xe6, 0x4e, 0xcc, 0xc0, 0xe5, 0x78, 0x82, 0x5a, 0xd0, 0x7d, 0xcc, 0xff, 0x72, 0x21, + 0xb8, 0x08, 0x46, 0x74, 0xf7, 0x43, 0x24, 0x8e, 0xe0, 0x35, 0x90, 0xe6, 0x81, 0x3a, 0x26, 0x4c, + 0x3c, 0x28, 0x52, 0xbb, 0x91, 0xc3, 0x00, 0xcb, 0x88, 0xd0, 0x65, 0x8b, 0x1b, 0x53, 0x2e, 0xa3, + 0x71, 0x64, 0x48, 0x97, 0xa2, 0x0d, 0xf9, 0x4e, 0x38, 0x19, 0xef, 0x46, 0xa9, 0xde, 0xac, 0xd8, + 0xa8, 0xfa, 0x76, 0x3f, 0xe3, 0x9c, 0x34, 0x3f, 0xf9, 0xdc, 0xbb, 0xc7, 0xc7, 0x0b, 0x4f, 0x1d, + 0x8a, 0x51, 0xe0, 0x4b, 0xcd, 0xb4, 0x59, 0x31, 0xc8, 0x9f, 0x7e, 0xc9, 0xd9, 0x78, 0x73, 0x64, + 0xea, 0xc5, 0xac, 0x83, 0x34, 0xd3, 0xeb, 0xc3, 0xc5, 0x81, 0xa0, 0xff, 0xfa, 0x13, 0x63, 0xeb, + 0x17, 0x0d, 0xdd, 0x51, 0xb7, 0xf0, 0xda, 0x49, 0xd3, 0x16, 0x55, 0x26, 0x29, 0xd4, 0x68, 0x9e, + 0x2b, 0x16, 0xbe, 0x58, 0x7d, 0x47, 0xa1, 0xfc, 0x8f, 0xf8, 0xb8, 0xd1, 0x7a, 0xd0, 0x31, 0xce, + 0x45, 0xcb, 0x3a, 0x8f, 0x95, 0x16, 0x04, 0x28, 0xaf, 0xd7, 0xfb, 0xca, 0xbb, 0x4b, 0x40, 0x7e, +}; + +static const xxh_u64 PRIME_MX1 = 0x165667919E3779F9ULL; /*!< 0b0001011001010110011001111001000110011110001101110111100111111001 */ +static const xxh_u64 PRIME_MX2 = 0x9FB21C651E98DF25ULL; /*!< 0b1001111110110010000111000110010100011110100110001101111100100101 */ + +#ifdef XXH_OLD_NAMES +# define kSecret XXH3_kSecret +#endif + +#ifdef XXH_DOXYGEN +/*! + * @brief Calculates a 32-bit to 64-bit long multiply. + * + * Implemented as a macro. + * + * Wraps `__emulu` on MSVC x86 because it tends to call `__allmul` when it doesn't + * need to (but it shouldn't need to anyways, it is about 7 instructions to do + * a 64x64 multiply...). Since we know that this will _always_ emit `MULL`, we + * use that instead of the normal method. + * + * If you are compiling for platforms like Thumb-1 and don't have a better option, + * you may also want to write your own long multiply routine here. + * + * @param x, y Numbers to be multiplied + * @return 64-bit product of the low 32 bits of @p x and @p y. + */ +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64(xxh_u64 x, xxh_u64 y) +{ + return (x & 0xFFFFFFFF) * (y & 0xFFFFFFFF); +} +#elif defined(_MSC_VER) && defined(_M_IX86) +# define XXH_mult32to64(x, y) __emulu((unsigned)(x), (unsigned)(y)) +#else +/* + * Downcast + upcast is usually better than masking on older compilers like + * GCC 4.2 (especially 32-bit ones), all without affecting newer compilers. + * + * The other method, (x & 0xFFFFFFFF) * (y & 0xFFFFFFFF), will AND both operands + * and perform a full 64x64 multiply -- entirely redundant on 32-bit. + */ +# define XXH_mult32to64(x, y) ((xxh_u64)(xxh_u32)(x) * (xxh_u64)(xxh_u32)(y)) +#endif + +/*! + * @brief Calculates a 64->128-bit long multiply. + * + * Uses `__uint128_t` and `_umul128` if available, otherwise uses a scalar + * version. + * + * @param lhs , rhs The 64-bit integers to be multiplied + * @return The 128-bit result represented in an @ref XXH128_hash_t. + */ +static XXH128_hash_t +XXH_mult64to128(xxh_u64 lhs, xxh_u64 rhs) +{ + /* + * GCC/Clang __uint128_t method. + * + * On most 64-bit targets, GCC and Clang define a __uint128_t type. + * This is usually the best way as it usually uses a native long 64-bit + * multiply, such as MULQ on x86_64 or MUL + UMULH on aarch64. + * + * Usually. + * + * Despite being a 32-bit platform, Clang (and emscripten) define this type + * despite not having the arithmetic for it. This results in a laggy + * compiler builtin call which calculates a full 128-bit multiply. + * In that case it is best to use the portable one. + * https://github.com/Cyan4973/xxHash/issues/211#issuecomment-515575677 + */ +#if (defined(__GNUC__) || defined(__clang__)) && !defined(__wasm__) \ + && defined(__SIZEOF_INT128__) \ + || (defined(_INTEGRAL_MAX_BITS) && _INTEGRAL_MAX_BITS >= 128) + + __uint128_t const product = (__uint128_t)lhs * (__uint128_t)rhs; + XXH128_hash_t r128; + r128.low64 = (xxh_u64)(product); + r128.high64 = (xxh_u64)(product >> 64); + return r128; + + /* + * MSVC for x64's _umul128 method. + * + * xxh_u64 _umul128(xxh_u64 Multiplier, xxh_u64 Multiplicand, xxh_u64 *HighProduct); + * + * This compiles to single operand MUL on x64. + */ +#elif (defined(_M_X64) || defined(_M_IA64)) && !defined(_M_ARM64EC) + +#ifndef _MSC_VER +# pragma intrinsic(_umul128) +#endif + xxh_u64 product_high; + xxh_u64 const product_low = _umul128(lhs, rhs, &product_high); + XXH128_hash_t r128; + r128.low64 = product_low; + r128.high64 = product_high; + return r128; + + /* + * MSVC for ARM64's __umulh method. + * + * This compiles to the same MUL + UMULH as GCC/Clang's __uint128_t method. + */ +#elif defined(_M_ARM64) || defined(_M_ARM64EC) + +#ifndef _MSC_VER +# pragma intrinsic(__umulh) +#endif + XXH128_hash_t r128; + r128.low64 = lhs * rhs; + r128.high64 = __umulh(lhs, rhs); + return r128; + +#else + /* + * Portable scalar method. Optimized for 32-bit and 64-bit ALUs. + * + * This is a fast and simple grade school multiply, which is shown below + * with base 10 arithmetic instead of base 0x100000000. + * + * 9 3 // D2 lhs = 93 + * x 7 5 // D2 rhs = 75 + * ---------- + * 1 5 // D2 lo_lo = (93 % 10) * (75 % 10) = 15 + * 4 5 | // D2 hi_lo = (93 / 10) * (75 % 10) = 45 + * 2 1 | // D2 lo_hi = (93 % 10) * (75 / 10) = 21 + * + 6 3 | | // D2 hi_hi = (93 / 10) * (75 / 10) = 63 + * --------- + * 2 7 | // D2 cross = (15 / 10) + (45 % 10) + 21 = 27 + * + 6 7 | | // D2 upper = (27 / 10) + (45 / 10) + 63 = 67 + * --------- + * 6 9 7 5 // D4 res = (27 * 10) + (15 % 10) + (67 * 100) = 6975 + * + * The reasons for adding the products like this are: + * 1. It avoids manual carry tracking. Just like how + * (9 * 9) + 9 + 9 = 99, the same applies with this for UINT64_MAX. + * This avoids a lot of complexity. + * + * 2. It hints for, and on Clang, compiles to, the powerful UMAAL + * instruction available in ARM's Digital Signal Processing extension + * in 32-bit ARMv6 and later, which is shown below: + * + * void UMAAL(xxh_u32 *RdLo, xxh_u32 *RdHi, xxh_u32 Rn, xxh_u32 Rm) + * { + * xxh_u64 product = (xxh_u64)*RdLo * (xxh_u64)*RdHi + Rn + Rm; + * *RdLo = (xxh_u32)(product & 0xFFFFFFFF); + * *RdHi = (xxh_u32)(product >> 32); + * } + * + * This instruction was designed for efficient long multiplication, and + * allows this to be calculated in only 4 instructions at speeds + * comparable to some 64-bit ALUs. + * + * 3. It isn't terrible on other platforms. Usually this will be a couple + * of 32-bit ADD/ADCs. + */ + + /* First calculate all of the cross products. */ + xxh_u64 const lo_lo = XXH_mult32to64(lhs & 0xFFFFFFFF, rhs & 0xFFFFFFFF); + xxh_u64 const hi_lo = XXH_mult32to64(lhs >> 32, rhs & 0xFFFFFFFF); + xxh_u64 const lo_hi = XXH_mult32to64(lhs & 0xFFFFFFFF, rhs >> 32); + xxh_u64 const hi_hi = XXH_mult32to64(lhs >> 32, rhs >> 32); + + /* Now add the products together. These will never overflow. */ + xxh_u64 const cross = (lo_lo >> 32) + (hi_lo & 0xFFFFFFFF) + lo_hi; + xxh_u64 const upper = (hi_lo >> 32) + (cross >> 32) + hi_hi; + xxh_u64 const lower = (cross << 32) | (lo_lo & 0xFFFFFFFF); + + XXH128_hash_t r128; + r128.low64 = lower; + r128.high64 = upper; + return r128; +#endif +} + +/*! + * @brief Calculates a 64-bit to 128-bit multiply, then XOR folds it. + * + * The reason for the separate function is to prevent passing too many structs + * around by value. This will hopefully inline the multiply, but we don't force it. + * + * @param lhs , rhs The 64-bit integers to multiply + * @return The low 64 bits of the product XOR'd by the high 64 bits. + * @see XXH_mult64to128() + */ +static xxh_u64 +XXH3_mul128_fold64(xxh_u64 lhs, xxh_u64 rhs) +{ + XXH128_hash_t product = XXH_mult64to128(lhs, rhs); + return product.low64 ^ product.high64; +} + +/*! Seems to produce slightly better code on GCC for some reason. */ +XXH_FORCE_INLINE XXH_CONSTF xxh_u64 XXH_xorshift64(xxh_u64 v64, int shift) +{ + XXH_ASSERT(0 <= shift && shift < 64); + return v64 ^ (v64 >> shift); +} + +/* + * This is a fast avalanche stage, + * suitable when input bits are already partially mixed + */ +static XXH64_hash_t XXH3_avalanche(xxh_u64 h64) +{ + h64 = XXH_xorshift64(h64, 37); + h64 *= PRIME_MX1; + h64 = XXH_xorshift64(h64, 32); + return h64; +} + +/* + * This is a stronger avalanche, + * inspired by Pelle Evensen's rrmxmx + * preferable when input has not been previously mixed + */ +static XXH64_hash_t XXH3_rrmxmx(xxh_u64 h64, xxh_u64 len) +{ + /* this mix is inspired by Pelle Evensen's rrmxmx */ + h64 ^= XXH_rotl64(h64, 49) ^ XXH_rotl64(h64, 24); + h64 *= PRIME_MX2; + h64 ^= (h64 >> 35) + len ; + h64 *= PRIME_MX2; + return XXH_xorshift64(h64, 28); +} + + +/* ========================================== + * Short keys + * ========================================== + * One of the shortcomings of XXH32 and XXH64 was that their performance was + * sub-optimal on short lengths. It used an iterative algorithm which strongly + * favored lengths that were a multiple of 4 or 8. + * + * Instead of iterating over individual inputs, we use a set of single shot + * functions which piece together a range of lengths and operate in constant time. + * + * Additionally, the number of multiplies has been significantly reduced. This + * reduces latency, especially when emulating 64-bit multiplies on 32-bit. + * + * Depending on the platform, this may or may not be faster than XXH32, but it + * is almost guaranteed to be faster than XXH64. + */ + +/* + * At very short lengths, there isn't enough input to fully hide secrets, or use + * the entire secret. + * + * There is also only a limited amount of mixing we can do before significantly + * impacting performance. + * + * Therefore, we use different sections of the secret and always mix two secret + * samples with an XOR. This should have no effect on performance on the + * seedless or withSeed variants because everything _should_ be constant folded + * by modern compilers. + * + * The XOR mixing hides individual parts of the secret and increases entropy. + * + * This adds an extra layer of strength for custom secrets. + */ +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_1to3_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(1 <= len && len <= 3); + XXH_ASSERT(secret != NULL); + /* + * len = 1: combined = { input[0], 0x01, input[0], input[0] } + * len = 2: combined = { input[1], 0x02, input[0], input[1] } + * len = 3: combined = { input[2], 0x03, input[0], input[1] } + */ + { xxh_u8 const c1 = input[0]; + xxh_u8 const c2 = input[len >> 1]; + xxh_u8 const c3 = input[len - 1]; + xxh_u32 const combined = ((xxh_u32)c1 << 16) | ((xxh_u32)c2 << 24) + | ((xxh_u32)c3 << 0) | ((xxh_u32)len << 8); + xxh_u64 const bitflip = (XXH_readLE32(secret) ^ XXH_readLE32(secret+4)) + seed; + xxh_u64 const keyed = (xxh_u64)combined ^ bitflip; + return XXH64_avalanche(keyed); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_4to8_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(4 <= len && len <= 8); + seed ^= (xxh_u64)XXH_swap32((xxh_u32)seed) << 32; + { xxh_u32 const input1 = XXH_readLE32(input); + xxh_u32 const input2 = XXH_readLE32(input + len - 4); + xxh_u64 const bitflip = (XXH_readLE64(secret+8) ^ XXH_readLE64(secret+16)) - seed; + xxh_u64 const input64 = input2 + (((xxh_u64)input1) << 32); + xxh_u64 const keyed = input64 ^ bitflip; + return XXH3_rrmxmx(keyed, len); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_9to16_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(9 <= len && len <= 16); + { xxh_u64 const bitflip1 = (XXH_readLE64(secret+24) ^ XXH_readLE64(secret+32)) + seed; + xxh_u64 const bitflip2 = (XXH_readLE64(secret+40) ^ XXH_readLE64(secret+48)) - seed; + xxh_u64 const input_lo = XXH_readLE64(input) ^ bitflip1; + xxh_u64 const input_hi = XXH_readLE64(input + len - 8) ^ bitflip2; + xxh_u64 const acc = len + + XXH_swap64(input_lo) + input_hi + + XXH3_mul128_fold64(input_lo, input_hi); + return XXH3_avalanche(acc); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_0to16_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(len <= 16); + { if (XXH_likely(len > 8)) return XXH3_len_9to16_64b(input, len, secret, seed); + if (XXH_likely(len >= 4)) return XXH3_len_4to8_64b(input, len, secret, seed); + if (len) return XXH3_len_1to3_64b(input, len, secret, seed); + return XXH64_avalanche(seed ^ (XXH_readLE64(secret+56) ^ XXH_readLE64(secret+64))); + } +} + +/* + * DISCLAIMER: There are known *seed-dependent* multicollisions here due to + * multiplication by zero, affecting hashes of lengths 17 to 240. + * + * However, they are very unlikely. + * + * Keep this in mind when using the unseeded XXH3_64bits() variant: As with all + * unseeded non-cryptographic hashes, it does not attempt to defend itself + * against specially crafted inputs, only random inputs. + * + * Compared to classic UMAC where a 1 in 2^31 chance of 4 consecutive bytes + * cancelling out the secret is taken an arbitrary number of times (addressed + * in XXH3_accumulate_512), this collision is very unlikely with random inputs + * and/or proper seeding: + * + * This only has a 1 in 2^63 chance of 8 consecutive bytes cancelling out, in a + * function that is only called up to 16 times per hash with up to 240 bytes of + * input. + * + * This is not too bad for a non-cryptographic hash function, especially with + * only 64 bit outputs. + * + * The 128-bit variant (which trades some speed for strength) is NOT affected + * by this, although it is always a good idea to use a proper seed if you care + * about strength. + */ +XXH_FORCE_INLINE xxh_u64 XXH3_mix16B(const xxh_u8* XXH_RESTRICT input, + const xxh_u8* XXH_RESTRICT secret, xxh_u64 seed64) +{ +#if defined(__GNUC__) && !defined(__clang__) /* GCC, not Clang */ \ + && defined(__i386__) && defined(__SSE2__) /* x86 + SSE2 */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable like XXH32 hack */ + /* + * UGLY HACK: + * GCC for x86 tends to autovectorize the 128-bit multiply, resulting in + * slower code. + * + * By forcing seed64 into a register, we disrupt the cost model and + * cause it to scalarize. See `XXH32_round()` + * + * FIXME: Clang's output is still _much_ faster -- On an AMD Ryzen 3600, + * XXH3_64bits @ len=240 runs at 4.6 GB/s with Clang 9, but 3.3 GB/s on + * GCC 9.2, despite both emitting scalar code. + * + * GCC generates much better scalar code than Clang for the rest of XXH3, + * which is why finding a more optimal codepath is an interest. + */ + XXH_COMPILER_GUARD(seed64); +#endif + { xxh_u64 const input_lo = XXH_readLE64(input); + xxh_u64 const input_hi = XXH_readLE64(input+8); + return XXH3_mul128_fold64( + input_lo ^ (XXH_readLE64(secret) + seed64), + input_hi ^ (XXH_readLE64(secret+8) - seed64) + ); + } +} + +/* For mid range keys, XXH3 uses a Mum-hash variant. */ +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_17to128_64b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(16 < len && len <= 128); + + { xxh_u64 acc = len * XXH_PRIME64_1; +#if XXH_SIZE_OPT >= 1 + /* Smaller and cleaner, but slightly slower. */ + unsigned int i = (unsigned int)(len - 1) / 32; + do { + acc += XXH3_mix16B(input+16 * i, secret+32*i, seed); + acc += XXH3_mix16B(input+len-16*(i+1), secret+32*i+16, seed); + } while (i-- != 0); +#else + if (len > 32) { + if (len > 64) { + if (len > 96) { + acc += XXH3_mix16B(input+48, secret+96, seed); + acc += XXH3_mix16B(input+len-64, secret+112, seed); + } + acc += XXH3_mix16B(input+32, secret+64, seed); + acc += XXH3_mix16B(input+len-48, secret+80, seed); + } + acc += XXH3_mix16B(input+16, secret+32, seed); + acc += XXH3_mix16B(input+len-32, secret+48, seed); + } + acc += XXH3_mix16B(input+0, secret+0, seed); + acc += XXH3_mix16B(input+len-16, secret+16, seed); +#endif + return XXH3_avalanche(acc); + } +} + +XXH_NO_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_129to240_64b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + + #define XXH3_MIDSIZE_STARTOFFSET 3 + #define XXH3_MIDSIZE_LASTOFFSET 17 + + { xxh_u64 acc = len * XXH_PRIME64_1; + xxh_u64 acc_end; + unsigned int const nbRounds = (unsigned int)len / 16; + unsigned int i; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + for (i=0; i<8; i++) { + acc += XXH3_mix16B(input+(16*i), secret+(16*i), seed); + } + /* last bytes */ + acc_end = XXH3_mix16B(input + len - 16, secret + XXH3_SECRET_SIZE_MIN - XXH3_MIDSIZE_LASTOFFSET, seed); + XXH_ASSERT(nbRounds >= 8); + acc = XXH3_avalanche(acc); +#if defined(__clang__) /* Clang */ \ + && (defined(__ARM_NEON) || defined(__ARM_NEON__)) /* NEON */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable */ + /* + * UGLY HACK: + * Clang for ARMv7-A tries to vectorize this loop, similar to GCC x86. + * In everywhere else, it uses scalar code. + * + * For 64->128-bit multiplies, even if the NEON was 100% optimal, it + * would still be slower than UMAAL (see XXH_mult64to128). + * + * Unfortunately, Clang doesn't handle the long multiplies properly and + * converts them to the nonexistent "vmulq_u64" intrinsic, which is then + * scalarized into an ugly mess of VMOV.32 instructions. + * + * This mess is difficult to avoid without turning autovectorization + * off completely, but they are usually relatively minor and/or not + * worth it to fix. + * + * This loop is the easiest to fix, as unlike XXH32, this pragma + * _actually works_ because it is a loop vectorization instead of an + * SLP vectorization. + */ + #pragma clang loop vectorize(disable) +#endif + for (i=8 ; i < nbRounds; i++) { + /* + * Prevents clang for unrolling the acc loop and interleaving with this one. + */ + XXH_COMPILER_GUARD(acc); + acc_end += XXH3_mix16B(input+(16*i), secret+(16*(i-8)) + XXH3_MIDSIZE_STARTOFFSET, seed); + } + return XXH3_avalanche(acc + acc_end); + } +} + + +/* ======= Long Keys ======= */ + +#define XXH_STRIPE_LEN 64 +#define XXH_SECRET_CONSUME_RATE 8 /* nb of secret bytes consumed at each accumulation */ +#define XXH_ACC_NB (XXH_STRIPE_LEN / sizeof(xxh_u64)) + +#ifdef XXH_OLD_NAMES +# define STRIPE_LEN XXH_STRIPE_LEN +# define ACC_NB XXH_ACC_NB +#endif + +#ifndef XXH_PREFETCH_DIST +# ifdef __clang__ +# define XXH_PREFETCH_DIST 320 +# else +# if (XXH_VECTOR == XXH_AVX512) +# define XXH_PREFETCH_DIST 512 +# else +# define XXH_PREFETCH_DIST 384 +# endif +# endif /* __clang__ */ +#endif /* XXH_PREFETCH_DIST */ + +/* + * These macros are to generate an XXH3_accumulate() function. + * The two arguments select the name suffix and target attribute. + * + * The name of this symbol is XXH3_accumulate_() and it calls + * XXH3_accumulate_512_(). + * + * It may be useful to hand implement this function if the compiler fails to + * optimize the inline function. + */ +#define XXH3_ACCUMULATE_TEMPLATE(name) \ +void \ +XXH3_accumulate_##name(xxh_u64* XXH_RESTRICT acc, \ + const xxh_u8* XXH_RESTRICT input, \ + const xxh_u8* XXH_RESTRICT secret, \ + size_t nbStripes) \ +{ \ + size_t n; \ + for (n = 0; n < nbStripes; n++ ) { \ + const xxh_u8* const in = input + n*XXH_STRIPE_LEN; \ + XXH_PREFETCH(in + XXH_PREFETCH_DIST); \ + XXH3_accumulate_512_##name( \ + acc, \ + in, \ + secret + n*XXH_SECRET_CONSUME_RATE); \ + } \ +} + + +XXH_FORCE_INLINE void XXH_writeLE64(void* dst, xxh_u64 v64) +{ + if (!XXH_CPU_LITTLE_ENDIAN) v64 = XXH_swap64(v64); + XXH_memcpy(dst, &v64, sizeof(v64)); +} + +/* Several intrinsic functions below are supposed to accept __int64 as argument, + * as documented in https://software.intel.com/sites/landingpage/IntrinsicsGuide/ . + * However, several environments do not define __int64 type, + * requiring a workaround. + */ +#if !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) + typedef int64_t xxh_i64; +#else + /* the following type must have a width of 64-bit */ + typedef long long xxh_i64; +#endif + + +/* + * XXH3_accumulate_512 is the tightest loop for long inputs, and it is the most optimized. + * + * It is a hardened version of UMAC, based off of FARSH's implementation. + * + * This was chosen because it adapts quite well to 32-bit, 64-bit, and SIMD + * implementations, and it is ridiculously fast. + * + * We harden it by mixing the original input to the accumulators as well as the product. + * + * This means that in the (relatively likely) case of a multiply by zero, the + * original input is preserved. + * + * On 128-bit inputs, we swap 64-bit pairs when we add the input to improve + * cross-pollination, as otherwise the upper and lower halves would be + * essentially independent. + * + * This doesn't matter on 64-bit hashes since they all get merged together in + * the end, so we skip the extra step. + * + * Both XXH3_64bits and XXH3_128bits use this subroutine. + */ + +#if (XXH_VECTOR == XXH_AVX512) \ + || (defined(XXH_DISPATCH_AVX512) && XXH_DISPATCH_AVX512 != 0) + +#ifndef XXH_TARGET_AVX512 +# define XXH_TARGET_AVX512 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_accumulate_512_avx512(void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + __m512i* const xacc = (__m512i *) acc; + XXH_ASSERT((((size_t)acc) & 63) == 0); + XXH_STATIC_ASSERT(XXH_STRIPE_LEN == sizeof(__m512i)); + + { + /* data_vec = input[0]; */ + __m512i const data_vec = _mm512_loadu_si512 (input); + /* key_vec = secret[0]; */ + __m512i const key_vec = _mm512_loadu_si512 (secret); + /* data_key = data_vec ^ key_vec; */ + __m512i const data_key = _mm512_xor_si512 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m512i const data_key_lo = _mm512_srli_epi64 (data_key, 32); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m512i const product = _mm512_mul_epu32 (data_key, data_key_lo); + /* xacc[0] += swap(data_vec); */ + __m512i const data_swap = _mm512_shuffle_epi32(data_vec, (_MM_PERM_ENUM)_MM_SHUFFLE(1, 0, 3, 2)); + __m512i const sum = _mm512_add_epi64(*xacc, data_swap); + /* xacc[0] += product; */ + *xacc = _mm512_add_epi64(product, sum); + } +} +XXH_FORCE_INLINE XXH_TARGET_AVX512 XXH3_ACCUMULATE_TEMPLATE(avx512) + +/* + * XXH3_scrambleAcc: Scrambles the accumulators to improve mixing. + * + * Multiplication isn't perfect, as explained by Google in HighwayHash: + * + * // Multiplication mixes/scrambles bytes 0-7 of the 64-bit result to + * // varying degrees. In descending order of goodness, bytes + * // 3 4 2 5 1 6 0 7 have quality 228 224 164 160 100 96 36 32. + * // As expected, the upper and lower bytes are much worse. + * + * Source: https://github.com/google/highwayhash/blob/0aaf66b/highwayhash/hh_avx2.h#L291 + * + * Since our algorithm uses a pseudorandom secret to add some variance into the + * mix, we don't need to (or want to) mix as often or as much as HighwayHash does. + * + * This isn't as tight as XXH3_accumulate, but still written in SIMD to avoid + * extraction. + * + * Both XXH3_64bits and XXH3_128bits use this subroutine. + */ + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_scrambleAcc_avx512(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 63) == 0); + XXH_STATIC_ASSERT(XXH_STRIPE_LEN == sizeof(__m512i)); + { __m512i* const xacc = (__m512i*) acc; + const __m512i prime32 = _mm512_set1_epi32((int)XXH_PRIME32_1); + + /* xacc[0] ^= (xacc[0] >> 47) */ + __m512i const acc_vec = *xacc; + __m512i const shifted = _mm512_srli_epi64 (acc_vec, 47); + /* xacc[0] ^= secret; */ + __m512i const key_vec = _mm512_loadu_si512 (secret); + __m512i const data_key = _mm512_ternarylogic_epi32(key_vec, acc_vec, shifted, 0x96 /* key_vec ^ acc_vec ^ shifted */); + + /* xacc[0] *= XXH_PRIME32_1; */ + __m512i const data_key_hi = _mm512_srli_epi64 (data_key, 32); + __m512i const prod_lo = _mm512_mul_epu32 (data_key, prime32); + __m512i const prod_hi = _mm512_mul_epu32 (data_key_hi, prime32); + *xacc = _mm512_add_epi64(prod_lo, _mm512_slli_epi64(prod_hi, 32)); + } +} + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_initCustomSecret_avx512(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 63) == 0); + XXH_STATIC_ASSERT(XXH_SEC_ALIGN == 64); + XXH_ASSERT(((size_t)customSecret & 63) == 0); + (void)(&XXH_writeLE64); + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / sizeof(__m512i); + __m512i const seed_pos = _mm512_set1_epi64((xxh_i64)seed64); + __m512i const seed = _mm512_mask_sub_epi64(seed_pos, 0xAA, _mm512_set1_epi8(0), seed_pos); + + const __m512i* const src = (const __m512i*) ((const void*) XXH3_kSecret); + __m512i* const dest = ( __m512i*) customSecret; + int i; + XXH_ASSERT(((size_t)src & 63) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dest & 63) == 0); + for (i=0; i < nbRounds; ++i) { + dest[i] = _mm512_add_epi64(_mm512_load_si512(src + i), seed); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_AVX2) \ + || (defined(XXH_DISPATCH_AVX2) && XXH_DISPATCH_AVX2 != 0) + +#ifndef XXH_TARGET_AVX2 +# define XXH_TARGET_AVX2 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void +XXH3_accumulate_512_avx2( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 31) == 0); + { __m256i* const xacc = (__m256i *) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xinput = (const __m256i *) input; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xsecret = (const __m256i *) secret; + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m256i); i++) { + /* data_vec = xinput[i]; */ + __m256i const data_vec = _mm256_loadu_si256 (xinput+i); + /* key_vec = xsecret[i]; */ + __m256i const key_vec = _mm256_loadu_si256 (xsecret+i); + /* data_key = data_vec ^ key_vec; */ + __m256i const data_key = _mm256_xor_si256 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m256i const data_key_lo = _mm256_srli_epi64 (data_key, 32); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m256i const product = _mm256_mul_epu32 (data_key, data_key_lo); + /* xacc[i] += swap(data_vec); */ + __m256i const data_swap = _mm256_shuffle_epi32(data_vec, _MM_SHUFFLE(1, 0, 3, 2)); + __m256i const sum = _mm256_add_epi64(xacc[i], data_swap); + /* xacc[i] += product; */ + xacc[i] = _mm256_add_epi64(product, sum); + } } +} +XXH_FORCE_INLINE XXH_TARGET_AVX2 XXH3_ACCUMULATE_TEMPLATE(avx2) + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void +XXH3_scrambleAcc_avx2(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 31) == 0); + { __m256i* const xacc = (__m256i*) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xsecret = (const __m256i *) secret; + const __m256i prime32 = _mm256_set1_epi32((int)XXH_PRIME32_1); + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m256i); i++) { + /* xacc[i] ^= (xacc[i] >> 47) */ + __m256i const acc_vec = xacc[i]; + __m256i const shifted = _mm256_srli_epi64 (acc_vec, 47); + __m256i const data_vec = _mm256_xor_si256 (acc_vec, shifted); + /* xacc[i] ^= xsecret; */ + __m256i const key_vec = _mm256_loadu_si256 (xsecret+i); + __m256i const data_key = _mm256_xor_si256 (data_vec, key_vec); + + /* xacc[i] *= XXH_PRIME32_1; */ + __m256i const data_key_hi = _mm256_srli_epi64 (data_key, 32); + __m256i const prod_lo = _mm256_mul_epu32 (data_key, prime32); + __m256i const prod_hi = _mm256_mul_epu32 (data_key_hi, prime32); + xacc[i] = _mm256_add_epi64(prod_lo, _mm256_slli_epi64(prod_hi, 32)); + } + } +} + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void XXH3_initCustomSecret_avx2(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 31) == 0); + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE / sizeof(__m256i)) == 6); + XXH_STATIC_ASSERT(XXH_SEC_ALIGN <= 64); + (void)(&XXH_writeLE64); + XXH_PREFETCH(customSecret); + { __m256i const seed = _mm256_set_epi64x((xxh_i64)(0U - seed64), (xxh_i64)seed64, (xxh_i64)(0U - seed64), (xxh_i64)seed64); + + const __m256i* const src = (const __m256i*) ((const void*) XXH3_kSecret); + __m256i* dest = ( __m256i*) customSecret; + +# if defined(__GNUC__) || defined(__clang__) + /* + * On GCC & Clang, marking 'dest' as modified will cause the compiler: + * - do not extract the secret from sse registers in the internal loop + * - use less common registers, and avoid pushing these reg into stack + */ + XXH_COMPILER_GUARD(dest); +# endif + XXH_ASSERT(((size_t)src & 31) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dest & 31) == 0); + + /* GCC -O2 need unroll loop manually */ + dest[0] = _mm256_add_epi64(_mm256_load_si256(src+0), seed); + dest[1] = _mm256_add_epi64(_mm256_load_si256(src+1), seed); + dest[2] = _mm256_add_epi64(_mm256_load_si256(src+2), seed); + dest[3] = _mm256_add_epi64(_mm256_load_si256(src+3), seed); + dest[4] = _mm256_add_epi64(_mm256_load_si256(src+4), seed); + dest[5] = _mm256_add_epi64(_mm256_load_si256(src+5), seed); + } +} + +#endif + +/* x86dispatch always generates SSE2 */ +#if (XXH_VECTOR == XXH_SSE2) || defined(XXH_X86DISPATCH) + +#ifndef XXH_TARGET_SSE2 +# define XXH_TARGET_SSE2 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void +XXH3_accumulate_512_sse2( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + /* SSE2 is just a half-scale version of the AVX2 version. */ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { __m128i* const xacc = (__m128i *) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xinput = (const __m128i *) input; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xsecret = (const __m128i *) secret; + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m128i); i++) { + /* data_vec = xinput[i]; */ + __m128i const data_vec = _mm_loadu_si128 (xinput+i); + /* key_vec = xsecret[i]; */ + __m128i const key_vec = _mm_loadu_si128 (xsecret+i); + /* data_key = data_vec ^ key_vec; */ + __m128i const data_key = _mm_xor_si128 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m128i const data_key_lo = _mm_shuffle_epi32 (data_key, _MM_SHUFFLE(0, 3, 0, 1)); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m128i const product = _mm_mul_epu32 (data_key, data_key_lo); + /* xacc[i] += swap(data_vec); */ + __m128i const data_swap = _mm_shuffle_epi32(data_vec, _MM_SHUFFLE(1,0,3,2)); + __m128i const sum = _mm_add_epi64(xacc[i], data_swap); + /* xacc[i] += product; */ + xacc[i] = _mm_add_epi64(product, sum); + } } +} +XXH_FORCE_INLINE XXH_TARGET_SSE2 XXH3_ACCUMULATE_TEMPLATE(sse2) + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void +XXH3_scrambleAcc_sse2(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { __m128i* const xacc = (__m128i*) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xsecret = (const __m128i *) secret; + const __m128i prime32 = _mm_set1_epi32((int)XXH_PRIME32_1); + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m128i); i++) { + /* xacc[i] ^= (xacc[i] >> 47) */ + __m128i const acc_vec = xacc[i]; + __m128i const shifted = _mm_srli_epi64 (acc_vec, 47); + __m128i const data_vec = _mm_xor_si128 (acc_vec, shifted); + /* xacc[i] ^= xsecret[i]; */ + __m128i const key_vec = _mm_loadu_si128 (xsecret+i); + __m128i const data_key = _mm_xor_si128 (data_vec, key_vec); + + /* xacc[i] *= XXH_PRIME32_1; */ + __m128i const data_key_hi = _mm_shuffle_epi32 (data_key, _MM_SHUFFLE(0, 3, 0, 1)); + __m128i const prod_lo = _mm_mul_epu32 (data_key, prime32); + __m128i const prod_hi = _mm_mul_epu32 (data_key_hi, prime32); + xacc[i] = _mm_add_epi64(prod_lo, _mm_slli_epi64(prod_hi, 32)); + } + } +} + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void XXH3_initCustomSecret_sse2(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 15) == 0); + (void)(&XXH_writeLE64); + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / sizeof(__m128i); + +# if defined(_MSC_VER) && defined(_M_IX86) && _MSC_VER < 1900 + /* MSVC 32bit mode does not support _mm_set_epi64x before 2015 */ + XXH_ALIGN(16) const xxh_i64 seed64x2[2] = { (xxh_i64)seed64, (xxh_i64)(0U - seed64) }; + __m128i const seed = _mm_load_si128((__m128i const*)seed64x2); +# else + __m128i const seed = _mm_set_epi64x((xxh_i64)(0U - seed64), (xxh_i64)seed64); +# endif + int i; + + const void* const src16 = XXH3_kSecret; + __m128i* dst16 = (__m128i*) customSecret; +# if defined(__GNUC__) || defined(__clang__) + /* + * On GCC & Clang, marking 'dest' as modified will cause the compiler: + * - do not extract the secret from sse registers in the internal loop + * - use less common registers, and avoid pushing these reg into stack + */ + XXH_COMPILER_GUARD(dst16); +# endif + XXH_ASSERT(((size_t)src16 & 15) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dst16 & 15) == 0); + + for (i=0; i < nbRounds; ++i) { + dst16[i] = _mm_add_epi64(_mm_load_si128((const __m128i *)src16+i), seed); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_NEON) + +/* forward declarations for the scalar routines */ +XXH_FORCE_INLINE void +XXH3_scalarRound(void* XXH_RESTRICT acc, void const* XXH_RESTRICT input, + void const* XXH_RESTRICT secret, size_t lane); + +XXH_FORCE_INLINE void +XXH3_scalarScrambleRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT secret, size_t lane); + +/*! + * @internal + * @brief The bulk processing loop for NEON and WASM SIMD128. + * + * The NEON code path is actually partially scalar when running on AArch64. This + * is to optimize the pipelining and can have up to 15% speedup depending on the + * CPU, and it also mitigates some GCC codegen issues. + * + * @see XXH3_NEON_LANES for configuring this and details about this optimization. + * + * NEON's 32-bit to 64-bit long multiply takes a half vector of 32-bit + * integers instead of the other platforms which mask full 64-bit vectors, + * so the setup is more complicated than just shifting right. + * + * Additionally, there is an optimization for 4 lanes at once noted below. + * + * Since, as stated, the most optimal amount of lanes for Cortexes is 6, + * there needs to be *three* versions of the accumulate operation used + * for the remaining 2 lanes. + * + * WASM's SIMD128 uses SIMDe's arm_neon.h polyfill because the intrinsics overlap + * nearly perfectly. + */ + +XXH_FORCE_INLINE void +XXH3_accumulate_512_neon( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + XXH_STATIC_ASSERT(XXH3_NEON_LANES > 0 && XXH3_NEON_LANES <= XXH_ACC_NB && XXH3_NEON_LANES % 2 == 0); + { /* GCC for darwin arm64 does not like aliasing here */ + xxh_aliasing_uint64x2_t* const xacc = (xxh_aliasing_uint64x2_t*) acc; + /* We don't use a uint32x4_t pointer because it causes bus errors on ARMv7. */ + uint8_t const* xinput = (const uint8_t *) input; + uint8_t const* xsecret = (const uint8_t *) secret; + + size_t i; +#ifdef __wasm_simd128__ + /* + * On WASM SIMD128, Clang emits direct address loads when XXH3_kSecret + * is constant propagated, which results in it converting it to this + * inside the loop: + * + * a = v128.load(XXH3_kSecret + 0 + $secret_offset, offset = 0) + * b = v128.load(XXH3_kSecret + 16 + $secret_offset, offset = 0) + * ... + * + * This requires a full 32-bit address immediate (and therefore a 6 byte + * instruction) as well as an add for each offset. + * + * Putting an asm guard prevents it from folding (at the cost of losing + * the alignment hint), and uses the free offset in `v128.load` instead + * of adding secret_offset each time which overall reduces code size by + * about a kilobyte and improves performance. + */ + XXH_COMPILER_GUARD(xsecret); +#endif + /* Scalar lanes use the normal scalarRound routine */ + for (i = XXH3_NEON_LANES; i < XXH_ACC_NB; i++) { + XXH3_scalarRound(acc, input, secret, i); + } + i = 0; + /* 4 NEON lanes at a time. */ + for (; i+1 < XXH3_NEON_LANES / 2; i+=2) { + /* data_vec = xinput[i]; */ + uint64x2_t data_vec_1 = XXH_vld1q_u64(xinput + (i * 16)); + uint64x2_t data_vec_2 = XXH_vld1q_u64(xinput + ((i+1) * 16)); + /* key_vec = xsecret[i]; */ + uint64x2_t key_vec_1 = XXH_vld1q_u64(xsecret + (i * 16)); + uint64x2_t key_vec_2 = XXH_vld1q_u64(xsecret + ((i+1) * 16)); + /* data_swap = swap(data_vec) */ + uint64x2_t data_swap_1 = vextq_u64(data_vec_1, data_vec_1, 1); + uint64x2_t data_swap_2 = vextq_u64(data_vec_2, data_vec_2, 1); + /* data_key = data_vec ^ key_vec; */ + uint64x2_t data_key_1 = veorq_u64(data_vec_1, key_vec_1); + uint64x2_t data_key_2 = veorq_u64(data_vec_2, key_vec_2); + + /* + * If we reinterpret the 64x2 vectors as 32x4 vectors, we can use a + * de-interleave operation for 4 lanes in 1 step with `vuzpq_u32` to + * get one vector with the low 32 bits of each lane, and one vector + * with the high 32 bits of each lane. + * + * The intrinsic returns a double vector because the original ARMv7-a + * instruction modified both arguments in place. AArch64 and SIMD128 emit + * two instructions from this intrinsic. + * + * [ dk11L | dk11H | dk12L | dk12H ] -> [ dk11L | dk12L | dk21L | dk22L ] + * [ dk21L | dk21H | dk22L | dk22H ] -> [ dk11H | dk12H | dk21H | dk22H ] + */ + uint32x4x2_t unzipped = vuzpq_u32( + vreinterpretq_u32_u64(data_key_1), + vreinterpretq_u32_u64(data_key_2) + ); + /* data_key_lo = data_key & 0xFFFFFFFF */ + uint32x4_t data_key_lo = unzipped.val[0]; + /* data_key_hi = data_key >> 32 */ + uint32x4_t data_key_hi = unzipped.val[1]; + /* + * Then, we can split the vectors horizontally and multiply which, as for most + * widening intrinsics, have a variant that works on both high half vectors + * for free on AArch64. A similar instruction is available on SIMD128. + * + * sum = data_swap + (u64x2) data_key_lo * (u64x2) data_key_hi + */ + uint64x2_t sum_1 = XXH_vmlal_low_u32(data_swap_1, data_key_lo, data_key_hi); + uint64x2_t sum_2 = XXH_vmlal_high_u32(data_swap_2, data_key_lo, data_key_hi); + /* + * Clang reorders + * a += b * c; // umlal swap.2d, dkl.2s, dkh.2s + * c += a; // add acc.2d, acc.2d, swap.2d + * to + * c += a; // add acc.2d, acc.2d, swap.2d + * c += b * c; // umlal acc.2d, dkl.2s, dkh.2s + * + * While it would make sense in theory since the addition is faster, + * for reasons likely related to umlal being limited to certain NEON + * pipelines, this is worse. A compiler guard fixes this. + */ + XXH_COMPILER_GUARD_CLANG_NEON(sum_1); + XXH_COMPILER_GUARD_CLANG_NEON(sum_2); + /* xacc[i] = acc_vec + sum; */ + xacc[i] = vaddq_u64(xacc[i], sum_1); + xacc[i+1] = vaddq_u64(xacc[i+1], sum_2); + } + /* Operate on the remaining NEON lanes 2 at a time. */ + for (; i < XXH3_NEON_LANES / 2; i++) { + /* data_vec = xinput[i]; */ + uint64x2_t data_vec = XXH_vld1q_u64(xinput + (i * 16)); + /* key_vec = xsecret[i]; */ + uint64x2_t key_vec = XXH_vld1q_u64(xsecret + (i * 16)); + /* acc_vec_2 = swap(data_vec) */ + uint64x2_t data_swap = vextq_u64(data_vec, data_vec, 1); + /* data_key = data_vec ^ key_vec; */ + uint64x2_t data_key = veorq_u64(data_vec, key_vec); + /* For two lanes, just use VMOVN and VSHRN. */ + /* data_key_lo = data_key & 0xFFFFFFFF; */ + uint32x2_t data_key_lo = vmovn_u64(data_key); + /* data_key_hi = data_key >> 32; */ + uint32x2_t data_key_hi = vshrn_n_u64(data_key, 32); + /* sum = data_swap + (u64x2) data_key_lo * (u64x2) data_key_hi; */ + uint64x2_t sum = vmlal_u32(data_swap, data_key_lo, data_key_hi); + /* Same Clang workaround as before */ + XXH_COMPILER_GUARD_CLANG_NEON(sum); + /* xacc[i] = acc_vec + sum; */ + xacc[i] = vaddq_u64 (xacc[i], sum); + } + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(neon) + +XXH_FORCE_INLINE void +XXH3_scrambleAcc_neon(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + + { xxh_aliasing_uint64x2_t* xacc = (xxh_aliasing_uint64x2_t*) acc; + uint8_t const* xsecret = (uint8_t const*) secret; + + size_t i; + /* WASM uses operator overloads and doesn't need these. */ +#ifndef __wasm_simd128__ + /* { prime32_1, prime32_1 } */ + uint32x2_t const kPrimeLo = vdup_n_u32(XXH_PRIME32_1); + /* { 0, prime32_1, 0, prime32_1 } */ + uint32x4_t const kPrimeHi = vreinterpretq_u32_u64(vdupq_n_u64((xxh_u64)XXH_PRIME32_1 << 32)); +#endif + + /* AArch64 uses both scalar and neon at the same time */ + for (i = XXH3_NEON_LANES; i < XXH_ACC_NB; i++) { + XXH3_scalarScrambleRound(acc, secret, i); + } + for (i=0; i < XXH3_NEON_LANES / 2; i++) { + /* xacc[i] ^= (xacc[i] >> 47); */ + uint64x2_t acc_vec = xacc[i]; + uint64x2_t shifted = vshrq_n_u64(acc_vec, 47); + uint64x2_t data_vec = veorq_u64(acc_vec, shifted); + + /* xacc[i] ^= xsecret[i]; */ + uint64x2_t key_vec = XXH_vld1q_u64(xsecret + (i * 16)); + uint64x2_t data_key = veorq_u64(data_vec, key_vec); + /* xacc[i] *= XXH_PRIME32_1 */ +#ifdef __wasm_simd128__ + /* SIMD128 has multiply by u64x2, use it instead of expanding and scalarizing */ + xacc[i] = data_key * XXH_PRIME32_1; +#else + /* + * Expanded version with portable NEON intrinsics + * + * lo(x) * lo(y) + (hi(x) * lo(y) << 32) + * + * prod_hi = hi(data_key) * lo(prime) << 32 + * + * Since we only need 32 bits of this multiply a trick can be used, reinterpreting the vector + * as a uint32x4_t and multiplying by { 0, prime, 0, prime } to cancel out the unwanted bits + * and avoid the shift. + */ + uint32x4_t prod_hi = vmulq_u32 (vreinterpretq_u32_u64(data_key), kPrimeHi); + /* Extract low bits for vmlal_u32 */ + uint32x2_t data_key_lo = vmovn_u64(data_key); + /* xacc[i] = prod_hi + lo(data_key) * XXH_PRIME32_1; */ + xacc[i] = vmlal_u32(vreinterpretq_u64_u32(prod_hi), data_key_lo, kPrimeLo); +#endif + } + } +} +#endif + +#if (XXH_VECTOR == XXH_VSX) + +XXH_FORCE_INLINE void +XXH3_accumulate_512_vsx( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + /* presumed aligned */ + xxh_aliasing_u64x2* const xacc = (xxh_aliasing_u64x2*) acc; + xxh_u8 const* const xinput = (xxh_u8 const*) input; /* no alignment restriction */ + xxh_u8 const* const xsecret = (xxh_u8 const*) secret; /* no alignment restriction */ + xxh_u64x2 const v32 = { 32, 32 }; + size_t i; + for (i = 0; i < XXH_STRIPE_LEN / sizeof(xxh_u64x2); i++) { + /* data_vec = xinput[i]; */ + xxh_u64x2 const data_vec = XXH_vec_loadu(xinput + 16*i); + /* key_vec = xsecret[i]; */ + xxh_u64x2 const key_vec = XXH_vec_loadu(xsecret + 16*i); + xxh_u64x2 const data_key = data_vec ^ key_vec; + /* shuffled = (data_key << 32) | (data_key >> 32); */ + xxh_u32x4 const shuffled = (xxh_u32x4)vec_rl(data_key, v32); + /* product = ((xxh_u64x2)data_key & 0xFFFFFFFF) * ((xxh_u64x2)shuffled & 0xFFFFFFFF); */ + xxh_u64x2 const product = XXH_vec_mulo((xxh_u32x4)data_key, shuffled); + /* acc_vec = xacc[i]; */ + xxh_u64x2 acc_vec = xacc[i]; + acc_vec += product; + + /* swap high and low halves */ +#ifdef __s390x__ + acc_vec += vec_permi(data_vec, data_vec, 2); +#else + acc_vec += vec_xxpermdi(data_vec, data_vec, 2); +#endif + xacc[i] = acc_vec; + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(vsx) + +XXH_FORCE_INLINE void +XXH3_scrambleAcc_vsx(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + + { xxh_aliasing_u64x2* const xacc = (xxh_aliasing_u64x2*) acc; + const xxh_u8* const xsecret = (const xxh_u8*) secret; + /* constants */ + xxh_u64x2 const v32 = { 32, 32 }; + xxh_u64x2 const v47 = { 47, 47 }; + xxh_u32x4 const prime = { XXH_PRIME32_1, XXH_PRIME32_1, XXH_PRIME32_1, XXH_PRIME32_1 }; + size_t i; + for (i = 0; i < XXH_STRIPE_LEN / sizeof(xxh_u64x2); i++) { + /* xacc[i] ^= (xacc[i] >> 47); */ + xxh_u64x2 const acc_vec = xacc[i]; + xxh_u64x2 const data_vec = acc_vec ^ (acc_vec >> v47); + + /* xacc[i] ^= xsecret[i]; */ + xxh_u64x2 const key_vec = XXH_vec_loadu(xsecret + 16*i); + xxh_u64x2 const data_key = data_vec ^ key_vec; + + /* xacc[i] *= XXH_PRIME32_1 */ + /* prod_lo = ((xxh_u64x2)data_key & 0xFFFFFFFF) * ((xxh_u64x2)prime & 0xFFFFFFFF); */ + xxh_u64x2 const prod_even = XXH_vec_mule((xxh_u32x4)data_key, prime); + /* prod_hi = ((xxh_u64x2)data_key >> 32) * ((xxh_u64x2)prime >> 32); */ + xxh_u64x2 const prod_odd = XXH_vec_mulo((xxh_u32x4)data_key, prime); + xacc[i] = prod_odd + (prod_even << v32); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_SVE) + +XXH_FORCE_INLINE void +XXH3_accumulate_512_sve( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + uint64_t *xacc = (uint64_t *)acc; + const uint64_t *xinput = (const uint64_t *)(const void *)input; + const uint64_t *xsecret = (const uint64_t *)(const void *)secret; + svuint64_t kSwap = sveor_n_u64_z(svptrue_b64(), svindex_u64(0, 1), 1); + uint64_t element_count = svcntd(); + if (element_count >= 8) { + svbool_t mask = svptrue_pat_b64(SV_VL8); + svuint64_t vacc = svld1_u64(mask, xacc); + ACCRND(vacc, 0); + svst1_u64(mask, xacc, vacc); + } else if (element_count == 2) { /* sve128 */ + svbool_t mask = svptrue_pat_b64(SV_VL2); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 2); + svuint64_t acc2 = svld1_u64(mask, xacc + 4); + svuint64_t acc3 = svld1_u64(mask, xacc + 6); + ACCRND(acc0, 0); + ACCRND(acc1, 2); + ACCRND(acc2, 4); + ACCRND(acc3, 6); + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 2, acc1); + svst1_u64(mask, xacc + 4, acc2); + svst1_u64(mask, xacc + 6, acc3); + } else { + svbool_t mask = svptrue_pat_b64(SV_VL4); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 4); + ACCRND(acc0, 0); + ACCRND(acc1, 4); + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 4, acc1); + } +} + +XXH_FORCE_INLINE void +XXH3_accumulate_sve(xxh_u64* XXH_RESTRICT acc, + const xxh_u8* XXH_RESTRICT input, + const xxh_u8* XXH_RESTRICT secret, + size_t nbStripes) +{ + if (nbStripes != 0) { + uint64_t *xacc = (uint64_t *)acc; + const uint64_t *xinput = (const uint64_t *)(const void *)input; + const uint64_t *xsecret = (const uint64_t *)(const void *)secret; + svuint64_t kSwap = sveor_n_u64_z(svptrue_b64(), svindex_u64(0, 1), 1); + uint64_t element_count = svcntd(); + if (element_count >= 8) { + svbool_t mask = svptrue_pat_b64(SV_VL8); + svuint64_t vacc = svld1_u64(mask, xacc + 0); + do { + /* svprfd(svbool_t, void *, enum svfprop); */ + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(vacc, 0); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, vacc); + } else if (element_count == 2) { /* sve128 */ + svbool_t mask = svptrue_pat_b64(SV_VL2); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 2); + svuint64_t acc2 = svld1_u64(mask, xacc + 4); + svuint64_t acc3 = svld1_u64(mask, xacc + 6); + do { + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(acc0, 0); + ACCRND(acc1, 2); + ACCRND(acc2, 4); + ACCRND(acc3, 6); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 2, acc1); + svst1_u64(mask, xacc + 4, acc2); + svst1_u64(mask, xacc + 6, acc3); + } else { + svbool_t mask = svptrue_pat_b64(SV_VL4); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 4); + do { + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(acc0, 0); + ACCRND(acc1, 4); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 4, acc1); + } + } +} + +#endif + +#if (XXH_VECTOR == XXH_LSX) +#define _LSX_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w)) + +XXH_FORCE_INLINE void +XXH3_accumulate_512_lsx( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { + __m128i* const xacc = (__m128i *) acc; + const __m128i* const xinput = (const __m128i *) input; + const __m128i* const xsecret = (const __m128i *) secret; + + for (size_t i = 0; i < XXH_STRIPE_LEN / sizeof(__m128i); i++) { + /* data_vec = xinput[i]; */ + __m128i const data_vec = __lsx_vld(xinput + i, 0); + /* key_vec = xsecret[i]; */ + __m128i const key_vec = __lsx_vld(xsecret + i, 0); + /* data_key = data_vec ^ key_vec; */ + __m128i const data_key = __lsx_vxor_v(data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m128i const data_key_lo = __lsx_vsrli_d(data_key, 32); + // __m128i const data_key_lo = __lsx_vsrli_d(data_key, 32); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m128i const product = __lsx_vmulwev_d_wu(data_key, data_key_lo); + /* xacc[i] += swap(data_vec); */ + __m128i const data_swap = __lsx_vshuf4i_w(data_vec, _LSX_SHUFFLE(1, 0, 3, 2)); + __m128i const sum = __lsx_vadd_d(xacc[i], data_swap); + /* xacc[i] += product; */ + xacc[i] = __lsx_vadd_d(product, sum); + } + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(lsx) + +XXH_FORCE_INLINE void +XXH3_scrambleAcc_lsx(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { + __m128i* const xacc = (__m128i*) acc; + const __m128i* const xsecret = (const __m128i *) secret; + const __m128i prime32 = __lsx_vreplgr2vr_w((int)XXH_PRIME32_1); + + for (size_t i = 0; i < XXH_STRIPE_LEN / sizeof(__m128i); i++) { + /* xacc[i] ^= (xacc[i] >> 47) */ + __m128i const acc_vec = xacc[i]; + __m128i const shifted = __lsx_vsrli_d(acc_vec, 47); + __m128i const data_vec = __lsx_vxor_v(acc_vec, shifted); + /* xacc[i] ^= xsecret[i]; */ + __m128i const key_vec = __lsx_vld(xsecret + i, 0); + __m128i const data_key = __lsx_vxor_v(data_vec, key_vec); + + /* xacc[i] *= XXH_PRIME32_1; */ + __m128i const data_key_hi = __lsx_vsrli_d(data_key, 32); + __m128i const prod_lo = __lsx_vmulwev_d_wu(data_key, prime32); + __m128i const prod_hi = __lsx_vmulwev_d_wu(data_key_hi, prime32); + xacc[i] = __lsx_vadd_d(prod_lo, __lsx_vslli_d(prod_hi, 32)); + } + } +} + +#endif + +/* scalar variants - universal */ + +#if defined(__aarch64__) && (defined(__GNUC__) || defined(__clang__)) +/* + * In XXH3_scalarRound(), GCC and Clang have a similar codegen issue, where they + * emit an excess mask and a full 64-bit multiply-add (MADD X-form). + * + * While this might not seem like much, as AArch64 is a 64-bit architecture, only + * big Cortex designs have a full 64-bit multiplier. + * + * On the little cores, the smaller 32-bit multiplier is used, and full 64-bit + * multiplies expand to 2-3 multiplies in microcode. This has a major penalty + * of up to 4 latency cycles and 2 stall cycles in the multiply pipeline. + * + * Thankfully, AArch64 still provides the 32-bit long multiply-add (UMADDL) which does + * not have this penalty and does the mask automatically. + */ +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64_add64(xxh_u64 lhs, xxh_u64 rhs, xxh_u64 acc) +{ + xxh_u64 ret; + /* note: %x = 64-bit register, %w = 32-bit register */ + __asm__("umaddl %x0, %w1, %w2, %x3" : "=r" (ret) : "r" (lhs), "r" (rhs), "r" (acc)); + return ret; +} +#else +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64_add64(xxh_u64 lhs, xxh_u64 rhs, xxh_u64 acc) +{ + return XXH_mult32to64((xxh_u32)lhs, (xxh_u32)rhs) + acc; +} +#endif + +/*! + * @internal + * @brief Scalar round for @ref XXH3_accumulate_512_scalar(). + * + * This is extracted to its own function because the NEON path uses a combination + * of NEON and scalar. + */ +XXH_FORCE_INLINE void +XXH3_scalarRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT input, + void const* XXH_RESTRICT secret, + size_t lane) +{ + xxh_u64* xacc = (xxh_u64*) acc; + xxh_u8 const* xinput = (xxh_u8 const*) input; + xxh_u8 const* xsecret = (xxh_u8 const*) secret; + XXH_ASSERT(lane < XXH_ACC_NB); + XXH_ASSERT(((size_t)acc & (XXH_ACC_ALIGN-1)) == 0); + { + xxh_u64 const data_val = XXH_readLE64(xinput + lane * 8); + xxh_u64 const data_key = data_val ^ XXH_readLE64(xsecret + lane * 8); + xacc[lane ^ 1] += data_val; /* swap adjacent lanes */ + xacc[lane] = XXH_mult32to64_add64(data_key /* & 0xFFFFFFFF */, data_key >> 32, xacc[lane]); + } +} + +/*! + * @internal + * @brief Processes a 64 byte block of data using the scalar path. + */ +XXH_FORCE_INLINE void +XXH3_accumulate_512_scalar(void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + size_t i; + /* ARM GCC refuses to unroll this loop, resulting in a 24% slowdown on ARMv6. */ +#if defined(__GNUC__) && !defined(__clang__) \ + && (defined(__arm__) || defined(__thumb2__)) \ + && defined(__ARM_FEATURE_UNALIGNED) /* no unaligned access just wastes bytes */ \ + && XXH_SIZE_OPT <= 0 +# pragma GCC unroll 8 +#endif + for (i=0; i < XXH_ACC_NB; i++) { + XXH3_scalarRound(acc, input, secret, i); + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(scalar) + +/*! + * @internal + * @brief Scalar scramble step for @ref XXH3_scrambleAcc_scalar(). + * + * This is extracted to its own function because the NEON path uses a combination + * of NEON and scalar. + */ +XXH_FORCE_INLINE void +XXH3_scalarScrambleRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT secret, + size_t lane) +{ + xxh_u64* const xacc = (xxh_u64*) acc; /* presumed aligned */ + const xxh_u8* const xsecret = (const xxh_u8*) secret; /* no alignment restriction */ + XXH_ASSERT((((size_t)acc) & (XXH_ACC_ALIGN-1)) == 0); + XXH_ASSERT(lane < XXH_ACC_NB); + { + xxh_u64 const key64 = XXH_readLE64(xsecret + lane * 8); + xxh_u64 acc64 = xacc[lane]; + acc64 = XXH_xorshift64(acc64, 47); + acc64 ^= key64; + acc64 *= XXH_PRIME32_1; + xacc[lane] = acc64; + } +} + +/*! + * @internal + * @brief Scrambles the accumulators after a large chunk has been read + */ +XXH_FORCE_INLINE void +XXH3_scrambleAcc_scalar(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + size_t i; + for (i=0; i < XXH_ACC_NB; i++) { + XXH3_scalarScrambleRound(acc, secret, i); + } +} + +XXH_FORCE_INLINE void +XXH3_initCustomSecret_scalar(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + /* + * We need a separate pointer for the hack below, + * which requires a non-const pointer. + * Any decent compiler will optimize this out otherwise. + */ + const xxh_u8* kSecretPtr = XXH3_kSecret; + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 15) == 0); + +#if defined(__GNUC__) && defined(__aarch64__) + /* + * UGLY HACK: + * GCC and Clang generate a bunch of MOV/MOVK pairs for aarch64, and they are + * placed sequentially, in order, at the top of the unrolled loop. + * + * While MOVK is great for generating constants (2 cycles for a 64-bit + * constant compared to 4 cycles for LDR), it fights for bandwidth with + * the arithmetic instructions. + * + * I L S + * MOVK + * MOVK + * MOVK + * MOVK + * ADD + * SUB STR + * STR + * By forcing loads from memory (as the asm line causes the compiler to assume + * that XXH3_kSecretPtr has been changed), the pipelines are used more + * efficiently: + * I L S + * LDR + * ADD LDR + * SUB STR + * STR + * + * See XXH3_NEON_LANES for details on the pipsline. + * + * XXH3_64bits_withSeed, len == 256, Snapdragon 835 + * without hack: 2654.4 MB/s + * with hack: 3202.9 MB/s + */ + XXH_COMPILER_GUARD(kSecretPtr); +#endif + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / 16; + int i; + for (i=0; i < nbRounds; i++) { + /* + * The asm hack causes the compiler to assume that kSecretPtr aliases with + * customSecret, and on aarch64, this prevented LDP from merging two + * loads together for free. Putting the loads together before the stores + * properly generates LDP. + */ + xxh_u64 lo = XXH_readLE64(kSecretPtr + 16*i) + seed64; + xxh_u64 hi = XXH_readLE64(kSecretPtr + 16*i + 8) - seed64; + XXH_writeLE64((xxh_u8*)customSecret + 16*i, lo); + XXH_writeLE64((xxh_u8*)customSecret + 16*i + 8, hi); + } } +} + + +typedef void (*XXH3_f_accumulate)(xxh_u64* XXH_RESTRICT, const xxh_u8* XXH_RESTRICT, const xxh_u8* XXH_RESTRICT, size_t); +typedef void (*XXH3_f_scrambleAcc)(void* XXH_RESTRICT, const void*); +typedef void (*XXH3_f_initCustomSecret)(void* XXH_RESTRICT, xxh_u64); + + +#if (XXH_VECTOR == XXH_AVX512) + +#define XXH3_accumulate_512 XXH3_accumulate_512_avx512 +#define XXH3_accumulate XXH3_accumulate_avx512 +#define XXH3_scrambleAcc XXH3_scrambleAcc_avx512 +#define XXH3_initCustomSecret XXH3_initCustomSecret_avx512 + +#elif (XXH_VECTOR == XXH_AVX2) + +#define XXH3_accumulate_512 XXH3_accumulate_512_avx2 +#define XXH3_accumulate XXH3_accumulate_avx2 +#define XXH3_scrambleAcc XXH3_scrambleAcc_avx2 +#define XXH3_initCustomSecret XXH3_initCustomSecret_avx2 + +#elif (XXH_VECTOR == XXH_SSE2) + +#define XXH3_accumulate_512 XXH3_accumulate_512_sse2 +#define XXH3_accumulate XXH3_accumulate_sse2 +#define XXH3_scrambleAcc XXH3_scrambleAcc_sse2 +#define XXH3_initCustomSecret XXH3_initCustomSecret_sse2 + +#elif (XXH_VECTOR == XXH_NEON) + +#define XXH3_accumulate_512 XXH3_accumulate_512_neon +#define XXH3_accumulate XXH3_accumulate_neon +#define XXH3_scrambleAcc XXH3_scrambleAcc_neon +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#elif (XXH_VECTOR == XXH_VSX) + +#define XXH3_accumulate_512 XXH3_accumulate_512_vsx +#define XXH3_accumulate XXH3_accumulate_vsx +#define XXH3_scrambleAcc XXH3_scrambleAcc_vsx +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#elif (XXH_VECTOR == XXH_SVE) +#define XXH3_accumulate_512 XXH3_accumulate_512_sve +#define XXH3_accumulate XXH3_accumulate_sve +#define XXH3_scrambleAcc XXH3_scrambleAcc_scalar +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#elif (XXH_VECTOR == XXH_LSX) +#define XXH3_accumulate_512 XXH3_accumulate_512_lsx +#define XXH3_accumulate XXH3_accumulate_lsx +#define XXH3_scrambleAcc XXH3_scrambleAcc_lsx +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#else /* scalar */ + +#define XXH3_accumulate_512 XXH3_accumulate_512_scalar +#define XXH3_accumulate XXH3_accumulate_scalar +#define XXH3_scrambleAcc XXH3_scrambleAcc_scalar +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#endif + +#if XXH_SIZE_OPT >= 1 /* don't do SIMD for initialization */ +# undef XXH3_initCustomSecret +# define XXH3_initCustomSecret XXH3_initCustomSecret_scalar +#endif + +XXH_FORCE_INLINE void +XXH3_hashLong_internal_loop(xxh_u64* XXH_RESTRICT acc, + const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + size_t const nbStripesPerBlock = (secretSize - XXH_STRIPE_LEN) / XXH_SECRET_CONSUME_RATE; + size_t const block_len = XXH_STRIPE_LEN * nbStripesPerBlock; + size_t const nb_blocks = (len - 1) / block_len; + + size_t n; + + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); + + for (n = 0; n < nb_blocks; n++) { + f_acc(acc, input + n*block_len, secret, nbStripesPerBlock); + f_scramble(acc, secret + secretSize - XXH_STRIPE_LEN); + } + + /* last partial block */ + XXH_ASSERT(len > XXH_STRIPE_LEN); + { size_t const nbStripes = ((len - 1) - (block_len * nb_blocks)) / XXH_STRIPE_LEN; + XXH_ASSERT(nbStripes <= (secretSize / XXH_SECRET_CONSUME_RATE)); + f_acc(acc, input + nb_blocks*block_len, secret, nbStripes); + + /* last stripe */ + { const xxh_u8* const p = input + len - XXH_STRIPE_LEN; +#define XXH_SECRET_LASTACC_START 7 /* not aligned on 8, last secret is different from acc & scrambler */ + XXH3_accumulate_512(acc, p, secret + secretSize - XXH_STRIPE_LEN - XXH_SECRET_LASTACC_START); + } } +} + +XXH_FORCE_INLINE xxh_u64 +XXH3_mix2Accs(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret) +{ + return XXH3_mul128_fold64( + acc[0] ^ XXH_readLE64(secret), + acc[1] ^ XXH_readLE64(secret+8) ); +} + +static XXH_PUREF XXH64_hash_t +XXH3_mergeAccs(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, xxh_u64 start) +{ + xxh_u64 result64 = start; + size_t i = 0; + + for (i = 0; i < 4; i++) { + result64 += XXH3_mix2Accs(acc+2*i, secret + 16*i); +#if defined(__clang__) /* Clang */ \ + && (defined(__arm__) || defined(__thumb__)) /* ARMv7 */ \ + && (defined(__ARM_NEON) || defined(__ARM_NEON__)) /* NEON */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable */ + /* + * UGLY HACK: + * Prevent autovectorization on Clang ARMv7-a. Exact same problem as + * the one in XXH3_len_129to240_64b. Speeds up shorter keys > 240b. + * XXH3_64bits, len == 256, Snapdragon 835: + * without hack: 2063.7 MB/s + * with hack: 2560.7 MB/s + */ + XXH_COMPILER_GUARD(result64); +#endif + } + + return XXH3_avalanche(result64); +} + +/* do not align on 8, so that the secret is different from the accumulator */ +#define XXH_SECRET_MERGEACCS_START 11 + +static XXH_PUREF XXH64_hash_t +XXH3_finalizeLong_64b(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, xxh_u64 len) +{ + return XXH3_mergeAccs(acc, secret + XXH_SECRET_MERGEACCS_START, len * XXH_PRIME64_1); +} + +#define XXH3_INIT_ACC { XXH_PRIME32_3, XXH_PRIME64_1, XXH_PRIME64_2, XXH_PRIME64_3, \ + XXH_PRIME64_4, XXH_PRIME32_2, XXH_PRIME64_5, XXH_PRIME32_1 } + +XXH_FORCE_INLINE XXH64_hash_t +XXH3_hashLong_64b_internal(const void* XXH_RESTRICT input, size_t len, + const void* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[XXH_ACC_NB] = XXH3_INIT_ACC; + + XXH3_hashLong_internal_loop(acc, (const xxh_u8*)input, len, (const xxh_u8*)secret, secretSize, f_acc, f_scramble); + + /* converge into final hash */ + XXH_STATIC_ASSERT(sizeof(acc) == 64); + XXH_ASSERT(secretSize >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + return XXH3_finalizeLong_64b(acc, (const xxh_u8*)secret, (xxh_u64)len); +} + +/* + * It's important for performance to transmit secret's size (when it's static) + * so that the compiler can properly optimize the vectorized loop. + * This makes a big performance difference for "medium" keys (<1 KB) when using AVX instruction set. + * When the secret size is unknown, or on GCC 12 where the mix of NO_INLINE and FORCE_INLINE + * breaks -Og, this is XXH_NO_INLINE. + */ +XXH3_WITH_SECRET_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSecret(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; + return XXH3_hashLong_64b_internal(input, len, secret, secretLen, XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * It's preferable for performance that XXH3_hashLong is not inlined, + * as it results in a smaller function for small data, easier to the instruction cache. + * Note that inside this no_inline function, we do inline the internal loop, + * and provide a statically defined secret size to allow optimization of vector loop. + */ +XXH_NO_INLINE XXH_PUREF XXH64_hash_t +XXH3_hashLong_64b_default(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; (void)secret; (void)secretLen; + return XXH3_hashLong_64b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * XXH3_hashLong_64b_withSeed(): + * Generate a custom key based on alteration of default XXH3_kSecret with the seed, + * and then use this key for long mode hashing. + * + * This operation is decently fast but nonetheless costs a little bit of time. + * Try to avoid it whenever possible (typically when seed==0). + * + * It's important for performance that XXH3_hashLong is not inlined. Not sure + * why (uop cache maybe?), but the difference is large and easily measurable. + */ +XXH_FORCE_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSeed_internal(const void* input, size_t len, + XXH64_hash_t seed, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble, + XXH3_f_initCustomSecret f_initSec) +{ +#if XXH_SIZE_OPT <= 0 + if (seed == 0) + return XXH3_hashLong_64b_internal(input, len, + XXH3_kSecret, sizeof(XXH3_kSecret), + f_acc, f_scramble); +#endif + { XXH_ALIGN(XXH_SEC_ALIGN) xxh_u8 secret[XXH_SECRET_DEFAULT_SIZE]; + f_initSec(secret, seed); + return XXH3_hashLong_64b_internal(input, len, secret, sizeof(secret), + f_acc, f_scramble); + } +} + +/* + * It's important for performance that XXH3_hashLong is not inlined. + */ +XXH_NO_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSeed(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)secret; (void)secretLen; + return XXH3_hashLong_64b_withSeed_internal(input, len, seed, + XXH3_accumulate, XXH3_scrambleAcc, XXH3_initCustomSecret); +} + + +typedef XXH64_hash_t (*XXH3_hashLong64_f)(const void* XXH_RESTRICT, size_t, + XXH64_hash_t, const xxh_u8* XXH_RESTRICT, size_t); + +XXH_FORCE_INLINE XXH64_hash_t +XXH3_64bits_internal(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen, + XXH3_hashLong64_f f_hashLong) +{ + XXH_ASSERT(secretLen >= XXH3_SECRET_SIZE_MIN); + /* + * If an action is to be taken if `secretLen` condition is not respected, + * it should be done here. + * For now, it's a contract pre-condition. + * Adding a check and a branch here would cost performance at every hash. + * Also, note that function signature doesn't offer room to return an error. + */ + if (len <= 16) + return XXH3_len_0to16_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, seed64); + if (len <= 128) + return XXH3_len_17to128_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_len_129to240_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + return f_hashLong(input, len, seed64, (const xxh_u8*)secret, secretLen); +} + + +/* === Public entry point === */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t XXH3_64bits(XXH_NOESCAPE const void* input, size_t length) +{ + return XXH3_64bits_internal(input, length, 0, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_default); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSecret(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_64bits_internal(input, length, 0, secret, secretSize, XXH3_hashLong_64b_withSecret); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSeed(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed) +{ + return XXH3_64bits_internal(input, length, seed, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_withSeed); +} + +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + if (length <= XXH3_MIDSIZE_MAX) + return XXH3_64bits_internal(input, length, seed, XXH3_kSecret, sizeof(XXH3_kSecret), NULL); + return XXH3_hashLong_64b_withSecret(input, length, seed, (const xxh_u8*)secret, secretSize); +} + + +/* === XXH3 streaming === */ +#ifndef XXH_NO_STREAM +/* + * Malloc's a pointer that is always aligned to @align. + * + * This must be freed with `XXH_alignedFree()`. + * + * malloc typically guarantees 16 byte alignment on 64-bit systems and 8 byte + * alignment on 32-bit. This isn't enough for the 32 byte aligned loads in AVX2 + * or on 32-bit, the 16 byte aligned loads in SSE2 and NEON. + * + * This underalignment previously caused a rather obvious crash which went + * completely unnoticed due to XXH3_createState() not actually being tested. + * Credit to RedSpah for noticing this bug. + * + * The alignment is done manually: Functions like posix_memalign or _mm_malloc + * are avoided: To maintain portability, we would have to write a fallback + * like this anyways, and besides, testing for the existence of library + * functions without relying on external build tools is impossible. + * + * The method is simple: Overallocate, manually align, and store the offset + * to the original behind the returned pointer. + * + * Align must be a power of 2 and 8 <= align <= 128. + */ +static XXH_MALLOCF void* XXH_alignedMalloc(size_t s, size_t align) +{ + XXH_ASSERT(align <= 128 && align >= 8); /* range check */ + XXH_ASSERT((align & (align-1)) == 0); /* power of 2 */ + XXH_ASSERT(s != 0 && s < (s + align)); /* empty/overflow */ + { /* Overallocate to make room for manual realignment and an offset byte */ + xxh_u8* base = (xxh_u8*)XXH_malloc(s + align); + if (base != NULL) { + /* + * Get the offset needed to align this pointer. + * + * Even if the returned pointer is aligned, there will always be + * at least one byte to store the offset to the original pointer. + */ + size_t offset = align - ((size_t)base & (align - 1)); /* base % align */ + /* Add the offset for the now-aligned pointer */ + xxh_u8* ptr = base + offset; + + XXH_ASSERT((size_t)ptr % align == 0); + + /* Store the offset immediately before the returned pointer. */ + ptr[-1] = (xxh_u8)offset; + return ptr; + } + return NULL; + } +} +/* + * Frees an aligned pointer allocated by XXH_alignedMalloc(). Don't pass + * normal malloc'd pointers, XXH_alignedMalloc has a specific data layout. + */ +static void XXH_alignedFree(void* p) +{ + if (p != NULL) { + xxh_u8* ptr = (xxh_u8*)p; + /* Get the offset byte we added in XXH_malloc. */ + xxh_u8 offset = ptr[-1]; + /* Free the original malloc'd pointer */ + xxh_u8* base = ptr - offset; + XXH_free(base); + } +} +/*! @ingroup XXH3_family */ +/*! + * @brief Allocate an @ref XXH3_state_t. + * + * @return An allocated pointer of @ref XXH3_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH3_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH3_state_t* XXH3_createState(void) +{ + XXH3_state_t* const state = (XXH3_state_t*)XXH_alignedMalloc(sizeof(XXH3_state_t), 64); + if (state==NULL) return NULL; + XXH3_INITSTATE(state); + return state; +} + +/*! @ingroup XXH3_family */ +/*! + * @brief Frees an @ref XXH3_state_t. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * + * @return @ref XXH_OK. + * + * @note Must be allocated with XXH3_createState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr) +{ + XXH_alignedFree(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API void +XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOESCAPE const XXH3_state_t* src_state) +{ + XXH_memcpy(dst_state, src_state, sizeof(*dst_state)); +} + +static void +XXH3_reset_internal(XXH3_state_t* statePtr, + XXH64_hash_t seed, + const void* secret, size_t secretSize) +{ + size_t const initStart = offsetof(XXH3_state_t, bufferedSize); + size_t const initLength = offsetof(XXH3_state_t, nbStripesPerBlock) - initStart; + XXH_ASSERT(offsetof(XXH3_state_t, nbStripesPerBlock) > initStart); + XXH_ASSERT(statePtr != NULL); + /* set members from bufferedSize to nbStripesPerBlock (excluded) to 0 */ + memset((char*)statePtr + initStart, 0, initLength); + statePtr->acc[0] = XXH_PRIME32_3; + statePtr->acc[1] = XXH_PRIME64_1; + statePtr->acc[2] = XXH_PRIME64_2; + statePtr->acc[3] = XXH_PRIME64_3; + statePtr->acc[4] = XXH_PRIME64_4; + statePtr->acc[5] = XXH_PRIME32_2; + statePtr->acc[6] = XXH_PRIME64_5; + statePtr->acc[7] = XXH_PRIME32_1; + statePtr->seed = seed; + statePtr->useSeed = (seed != 0); + statePtr->extSecret = (const unsigned char*)secret; + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); + statePtr->secretLimit = secretSize - XXH_STRIPE_LEN; + statePtr->nbStripesPerBlock = statePtr->secretLimit / XXH_SECRET_CONSUME_RATE; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr) +{ + if (statePtr == NULL) return XXH_ERROR; + XXH3_reset_internal(statePtr, 0, XXH3_kSecret, XXH_SECRET_DEFAULT_SIZE); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + if (statePtr == NULL) return XXH_ERROR; + XXH3_reset_internal(statePtr, 0, secret, secretSize); + if (secret == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed) +{ + if (statePtr == NULL) return XXH_ERROR; + if (seed==0) return XXH3_64bits_reset(statePtr); + if ((seed != statePtr->seed) || (statePtr->extSecret != NULL)) + XXH3_initCustomSecret(statePtr->customSecret, seed); + XXH3_reset_internal(statePtr, seed, NULL, XXH_SECRET_DEFAULT_SIZE); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed64) +{ + if (statePtr == NULL) return XXH_ERROR; + if (secret == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; + XXH3_reset_internal(statePtr, seed64, secret, secretSize); + statePtr->useSeed = 1; /* always, even if seed64==0 */ + return XXH_OK; +} + +/*! + * @internal + * @brief Processes a large input for XXH3_update() and XXH3_digest_long(). + * + * Unlike XXH3_hashLong_internal_loop(), this can process data that overlaps a block. + * + * @param acc Pointer to the 8 accumulator lanes + * @param nbStripesSoFarPtr In/out pointer to the number of leftover stripes in the block* + * @param nbStripesPerBlock Number of stripes in a block + * @param input Input pointer + * @param nbStripes Number of stripes to process + * @param secret Secret pointer + * @param secretLimit Offset of the last block in @p secret + * @param f_acc Pointer to an XXH3_accumulate implementation + * @param f_scramble Pointer to an XXH3_scrambleAcc implementation + * @return Pointer past the end of @p input after processing + */ +XXH_FORCE_INLINE const xxh_u8 * +XXH3_consumeStripes(xxh_u64* XXH_RESTRICT acc, + size_t* XXH_RESTRICT nbStripesSoFarPtr, size_t nbStripesPerBlock, + const xxh_u8* XXH_RESTRICT input, size_t nbStripes, + const xxh_u8* XXH_RESTRICT secret, size_t secretLimit, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + const xxh_u8* initialSecret = secret + *nbStripesSoFarPtr * XXH_SECRET_CONSUME_RATE; + /* Process full blocks */ + if (nbStripes >= (nbStripesPerBlock - *nbStripesSoFarPtr)) { + /* Process the initial partial block... */ + size_t nbStripesThisIter = nbStripesPerBlock - *nbStripesSoFarPtr; + + do { + /* Accumulate and scramble */ + f_acc(acc, input, initialSecret, nbStripesThisIter); + f_scramble(acc, secret + secretLimit); + input += nbStripesThisIter * XXH_STRIPE_LEN; + nbStripes -= nbStripesThisIter; + /* Then continue the loop with the full block size */ + nbStripesThisIter = nbStripesPerBlock; + initialSecret = secret; + } while (nbStripes >= nbStripesPerBlock); + *nbStripesSoFarPtr = 0; + } + /* Process a partial block */ + if (nbStripes > 0) { + f_acc(acc, input, initialSecret, nbStripes); + input += nbStripes * XXH_STRIPE_LEN; + *nbStripesSoFarPtr += nbStripes; + } + /* Return end pointer */ + return input; +} + +#ifndef XXH3_STREAM_USE_STACK +# if XXH_SIZE_OPT <= 0 && !defined(__clang__) /* clang doesn't need additional stack space */ +# define XXH3_STREAM_USE_STACK 1 +# endif +#endif +/* + * Both XXH3_64bits_update and XXH3_128bits_update use this routine. + */ +XXH_FORCE_INLINE XXH_errorcode +XXH3_update(XXH3_state_t* XXH_RESTRICT const state, + const xxh_u8* XXH_RESTRICT input, size_t len, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + XXH_ASSERT(state != NULL); + { const xxh_u8* const bEnd = input + len; + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; +#if defined(XXH3_STREAM_USE_STACK) && XXH3_STREAM_USE_STACK >= 1 + /* For some reason, gcc and MSVC seem to suffer greatly + * when operating accumulators directly into state. + * Operating into stack space seems to enable proper optimization. + * clang, on the other hand, doesn't seem to need this trick */ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[8]; + XXH_memcpy(acc, state->acc, sizeof(acc)); +#else + xxh_u64* XXH_RESTRICT const acc = state->acc; +#endif + state->totalLen += len; + XXH_ASSERT(state->bufferedSize <= XXH3_INTERNALBUFFER_SIZE); + + /* small input : just fill in tmp buffer */ + if (len <= XXH3_INTERNALBUFFER_SIZE - state->bufferedSize) { + XXH_memcpy(state->buffer + state->bufferedSize, input, len); + state->bufferedSize += (XXH32_hash_t)len; + return XXH_OK; + } + + /* total input is now > XXH3_INTERNALBUFFER_SIZE */ + #define XXH3_INTERNALBUFFER_STRIPES (XXH3_INTERNALBUFFER_SIZE / XXH_STRIPE_LEN) + XXH_STATIC_ASSERT(XXH3_INTERNALBUFFER_SIZE % XXH_STRIPE_LEN == 0); /* clean multiple */ + + /* + * Internal buffer is partially filled (always, except at beginning) + * Complete it, then consume it. + */ + if (state->bufferedSize) { + size_t const loadSize = XXH3_INTERNALBUFFER_SIZE - state->bufferedSize; + XXH_memcpy(state->buffer + state->bufferedSize, input, loadSize); + input += loadSize; + XXH3_consumeStripes(acc, + &state->nbStripesSoFar, state->nbStripesPerBlock, + state->buffer, XXH3_INTERNALBUFFER_STRIPES, + secret, state->secretLimit, + f_acc, f_scramble); + state->bufferedSize = 0; + } + XXH_ASSERT(input < bEnd); + if (bEnd - input > XXH3_INTERNALBUFFER_SIZE) { + size_t nbStripes = (size_t)(bEnd - 1 - input) / XXH_STRIPE_LEN; + input = XXH3_consumeStripes(acc, + &state->nbStripesSoFar, state->nbStripesPerBlock, + input, nbStripes, + secret, state->secretLimit, + f_acc, f_scramble); + XXH_memcpy(state->buffer + sizeof(state->buffer) - XXH_STRIPE_LEN, input - XXH_STRIPE_LEN, XXH_STRIPE_LEN); + + } + /* Some remaining input (always) : buffer it */ + XXH_ASSERT(input < bEnd); + XXH_ASSERT(bEnd - input <= XXH3_INTERNALBUFFER_SIZE); + XXH_ASSERT(state->bufferedSize == 0); + XXH_memcpy(state->buffer, input, (size_t)(bEnd-input)); + state->bufferedSize = (XXH32_hash_t)(bEnd-input); +#if defined(XXH3_STREAM_USE_STACK) && XXH3_STREAM_USE_STACK >= 1 + /* save stack accumulators into state */ + XXH_memcpy(state->acc, acc, sizeof(acc)); +#endif + } + + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_update(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_update(state, (const xxh_u8*)input, len, + XXH3_accumulate, XXH3_scrambleAcc); +} + + +XXH_FORCE_INLINE void +XXH3_digest_long (XXH64_hash_t* acc, + const XXH3_state_t* state, + const unsigned char* secret) +{ + xxh_u8 lastStripe[XXH_STRIPE_LEN]; + const xxh_u8* lastStripePtr; + + /* + * Digest on a local copy. This way, the state remains unaltered, and it can + * continue ingesting more input afterwards. + */ + XXH_memcpy(acc, state->acc, sizeof(state->acc)); + if (state->bufferedSize >= XXH_STRIPE_LEN) { + /* Consume remaining stripes then point to remaining data in buffer */ + size_t const nbStripes = (state->bufferedSize - 1) / XXH_STRIPE_LEN; + size_t nbStripesSoFar = state->nbStripesSoFar; + XXH3_consumeStripes(acc, + &nbStripesSoFar, state->nbStripesPerBlock, + state->buffer, nbStripes, + secret, state->secretLimit, + XXH3_accumulate, XXH3_scrambleAcc); + lastStripePtr = state->buffer + state->bufferedSize - XXH_STRIPE_LEN; + } else { /* bufferedSize < XXH_STRIPE_LEN */ + /* Copy to temp buffer */ + size_t const catchupSize = XXH_STRIPE_LEN - state->bufferedSize; + XXH_ASSERT(state->bufferedSize > 0); /* there is always some input buffered */ + XXH_memcpy(lastStripe, state->buffer + sizeof(state->buffer) - catchupSize, catchupSize); + XXH_memcpy(lastStripe + catchupSize, state->buffer, state->bufferedSize); + lastStripePtr = lastStripe; + } + /* Last stripe */ + XXH3_accumulate_512(acc, + lastStripePtr, + secret + state->secretLimit - XXH_SECRET_LASTACC_START); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* state) +{ + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; + if (state->totalLen > XXH3_MIDSIZE_MAX) { + XXH_ALIGN(XXH_ACC_ALIGN) XXH64_hash_t acc[XXH_ACC_NB]; + XXH3_digest_long(acc, state, secret); + return XXH3_finalizeLong_64b(acc, secret, (xxh_u64)state->totalLen); + } + /* totalLen <= XXH3_MIDSIZE_MAX: digesting a short input */ + if (state->useSeed) + return XXH3_64bits_withSeed(state->buffer, (size_t)state->totalLen, state->seed); + return XXH3_64bits_withSecret(state->buffer, (size_t)(state->totalLen), + secret, state->secretLimit + XXH_STRIPE_LEN); +} +#endif /* !XXH_NO_STREAM */ + + +/* ========================================== + * XXH3 128 bits (a.k.a XXH128) + * ========================================== + * XXH3's 128-bit variant has better mixing and strength than the 64-bit variant, + * even without counting the significantly larger output size. + * + * For example, extra steps are taken to avoid the seed-dependent collisions + * in 17-240 byte inputs (See XXH3_mix16B and XXH128_mix32B). + * + * This strength naturally comes at the cost of some speed, especially on short + * lengths. Note that longer hashes are about as fast as the 64-bit version + * due to it using only a slight modification of the 64-bit loop. + * + * XXH128 is also more oriented towards 64-bit machines. It is still extremely + * fast for a _128-bit_ hash on 32-bit (it usually clears XXH64). + */ + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_1to3_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + /* A doubled version of 1to3_64b with different constants. */ + XXH_ASSERT(input != NULL); + XXH_ASSERT(1 <= len && len <= 3); + XXH_ASSERT(secret != NULL); + /* + * len = 1: combinedl = { input[0], 0x01, input[0], input[0] } + * len = 2: combinedl = { input[1], 0x02, input[0], input[1] } + * len = 3: combinedl = { input[2], 0x03, input[0], input[1] } + */ + { xxh_u8 const c1 = input[0]; + xxh_u8 const c2 = input[len >> 1]; + xxh_u8 const c3 = input[len - 1]; + xxh_u32 const combinedl = ((xxh_u32)c1 <<16) | ((xxh_u32)c2 << 24) + | ((xxh_u32)c3 << 0) | ((xxh_u32)len << 8); + xxh_u32 const combinedh = XXH_rotl32(XXH_swap32(combinedl), 13); + xxh_u64 const bitflipl = (XXH_readLE32(secret) ^ XXH_readLE32(secret+4)) + seed; + xxh_u64 const bitfliph = (XXH_readLE32(secret+8) ^ XXH_readLE32(secret+12)) - seed; + xxh_u64 const keyed_lo = (xxh_u64)combinedl ^ bitflipl; + xxh_u64 const keyed_hi = (xxh_u64)combinedh ^ bitfliph; + XXH128_hash_t h128; + h128.low64 = XXH64_avalanche(keyed_lo); + h128.high64 = XXH64_avalanche(keyed_hi); + return h128; + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_4to8_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(4 <= len && len <= 8); + seed ^= (xxh_u64)XXH_swap32((xxh_u32)seed) << 32; + { xxh_u32 const input_lo = XXH_readLE32(input); + xxh_u32 const input_hi = XXH_readLE32(input + len - 4); + xxh_u64 const input_64 = input_lo + ((xxh_u64)input_hi << 32); + xxh_u64 const bitflip = (XXH_readLE64(secret+16) ^ XXH_readLE64(secret+24)) + seed; + xxh_u64 const keyed = input_64 ^ bitflip; + + /* Shift len to the left to ensure it is even, this avoids even multiplies. */ + XXH128_hash_t m128 = XXH_mult64to128(keyed, XXH_PRIME64_1 + (len << 2)); + + m128.high64 += (m128.low64 << 1); + m128.low64 ^= (m128.high64 >> 3); + + m128.low64 = XXH_xorshift64(m128.low64, 35); + m128.low64 *= PRIME_MX2; + m128.low64 = XXH_xorshift64(m128.low64, 28); + m128.high64 = XXH3_avalanche(m128.high64); + return m128; + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_9to16_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(9 <= len && len <= 16); + { xxh_u64 const bitflipl = (XXH_readLE64(secret+32) ^ XXH_readLE64(secret+40)) - seed; + xxh_u64 const bitfliph = (XXH_readLE64(secret+48) ^ XXH_readLE64(secret+56)) + seed; + xxh_u64 const input_lo = XXH_readLE64(input); + xxh_u64 input_hi = XXH_readLE64(input + len - 8); + XXH128_hash_t m128 = XXH_mult64to128(input_lo ^ input_hi ^ bitflipl, XXH_PRIME64_1); + /* + * Put len in the middle of m128 to ensure that the length gets mixed to + * both the low and high bits in the 128x64 multiply below. + */ + m128.low64 += (xxh_u64)(len - 1) << 54; + input_hi ^= bitfliph; + /* + * Add the high 32 bits of input_hi to the high 32 bits of m128, then + * add the long product of the low 32 bits of input_hi and XXH_PRIME32_2 to + * the high 64 bits of m128. + * + * The best approach to this operation is different on 32-bit and 64-bit. + */ + if (sizeof(void *) < sizeof(xxh_u64)) { /* 32-bit */ + /* + * 32-bit optimized version, which is more readable. + * + * On 32-bit, it removes an ADC and delays a dependency between the two + * halves of m128.high64, but it generates an extra mask on 64-bit. + */ + m128.high64 += (input_hi & 0xFFFFFFFF00000000ULL) + XXH_mult32to64((xxh_u32)input_hi, XXH_PRIME32_2); + } else { + /* + * 64-bit optimized (albeit more confusing) version. + * + * Uses some properties of addition and multiplication to remove the mask: + * + * Let: + * a = input_hi.lo = (input_hi & 0x00000000FFFFFFFF) + * b = input_hi.hi = (input_hi & 0xFFFFFFFF00000000) + * c = XXH_PRIME32_2 + * + * a + (b * c) + * Inverse Property: x + y - x == y + * a + (b * (1 + c - 1)) + * Distributive Property: x * (y + z) == (x * y) + (x * z) + * a + (b * 1) + (b * (c - 1)) + * Identity Property: x * 1 == x + * a + b + (b * (c - 1)) + * + * Substitute a, b, and c: + * input_hi.hi + input_hi.lo + ((xxh_u64)input_hi.lo * (XXH_PRIME32_2 - 1)) + * + * Since input_hi.hi + input_hi.lo == input_hi, we get this: + * input_hi + ((xxh_u64)input_hi.lo * (XXH_PRIME32_2 - 1)) + */ + m128.high64 += input_hi + XXH_mult32to64((xxh_u32)input_hi, XXH_PRIME32_2 - 1); + } + /* m128 ^= XXH_swap64(m128 >> 64); */ + m128.low64 ^= XXH_swap64(m128.high64); + + { /* 128x64 multiply: h128 = m128 * XXH_PRIME64_2; */ + XXH128_hash_t h128 = XXH_mult64to128(m128.low64, XXH_PRIME64_2); + h128.high64 += m128.high64 * XXH_PRIME64_2; + + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = XXH3_avalanche(h128.high64); + return h128; + } } +} + +/* + * Assumption: `secret` size is >= XXH3_SECRET_SIZE_MIN + */ +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_0to16_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(len <= 16); + { if (len > 8) return XXH3_len_9to16_128b(input, len, secret, seed); + if (len >= 4) return XXH3_len_4to8_128b(input, len, secret, seed); + if (len) return XXH3_len_1to3_128b(input, len, secret, seed); + { XXH128_hash_t h128; + xxh_u64 const bitflipl = XXH_readLE64(secret+64) ^ XXH_readLE64(secret+72); + xxh_u64 const bitfliph = XXH_readLE64(secret+80) ^ XXH_readLE64(secret+88); + h128.low64 = XXH64_avalanche(seed ^ bitflipl); + h128.high64 = XXH64_avalanche( seed ^ bitfliph); + return h128; + } } +} + +/* + * A bit slower than XXH3_mix16B, but handles multiply by zero better. + */ +XXH_FORCE_INLINE XXH128_hash_t +XXH128_mix32B(XXH128_hash_t acc, const xxh_u8* input_1, const xxh_u8* input_2, + const xxh_u8* secret, XXH64_hash_t seed) +{ + acc.low64 += XXH3_mix16B (input_1, secret+0, seed); + acc.low64 ^= XXH_readLE64(input_2) + XXH_readLE64(input_2 + 8); + acc.high64 += XXH3_mix16B (input_2, secret+16, seed); + acc.high64 ^= XXH_readLE64(input_1) + XXH_readLE64(input_1 + 8); + return acc; +} + + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_17to128_128b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(16 < len && len <= 128); + + { XXH128_hash_t acc; + acc.low64 = len * XXH_PRIME64_1; + acc.high64 = 0; + +#if XXH_SIZE_OPT >= 1 + { + /* Smaller, but slightly slower. */ + unsigned int i = (unsigned int)(len - 1) / 32; + do { + acc = XXH128_mix32B(acc, input+16*i, input+len-16*(i+1), secret+32*i, seed); + } while (i-- != 0); + } +#else + if (len > 32) { + if (len > 64) { + if (len > 96) { + acc = XXH128_mix32B(acc, input+48, input+len-64, secret+96, seed); + } + acc = XXH128_mix32B(acc, input+32, input+len-48, secret+64, seed); + } + acc = XXH128_mix32B(acc, input+16, input+len-32, secret+32, seed); + } + acc = XXH128_mix32B(acc, input, input+len-16, secret, seed); +#endif + { XXH128_hash_t h128; + h128.low64 = acc.low64 + acc.high64; + h128.high64 = (acc.low64 * XXH_PRIME64_1) + + (acc.high64 * XXH_PRIME64_4) + + ((len - seed) * XXH_PRIME64_2); + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = (XXH64_hash_t)0 - XXH3_avalanche(h128.high64); + return h128; + } + } +} + +XXH_NO_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_129to240_128b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + + { XXH128_hash_t acc; + unsigned i; + acc.low64 = len * XXH_PRIME64_1; + acc.high64 = 0; + /* + * We set as `i` as offset + 32. We do this so that unchanged + * `len` can be used as upper bound. This reaches a sweet spot + * where both x86 and aarch64 get simple agen and good codegen + * for the loop. + */ + for (i = 32; i < 160; i += 32) { + acc = XXH128_mix32B(acc, + input + i - 32, + input + i - 16, + secret + i - 32, + seed); + } + acc.low64 = XXH3_avalanche(acc.low64); + acc.high64 = XXH3_avalanche(acc.high64); + /* + * NB: `i <= len` will duplicate the last 32-bytes if + * len % 32 was zero. This is an unfortunate necessity to keep + * the hash result stable. + */ + for (i=160; i <= len; i += 32) { + acc = XXH128_mix32B(acc, + input + i - 32, + input + i - 16, + secret + XXH3_MIDSIZE_STARTOFFSET + i - 160, + seed); + } + /* last bytes */ + acc = XXH128_mix32B(acc, + input + len - 16, + input + len - 32, + secret + XXH3_SECRET_SIZE_MIN - XXH3_MIDSIZE_LASTOFFSET - 16, + (XXH64_hash_t)0 - seed); + + { XXH128_hash_t h128; + h128.low64 = acc.low64 + acc.high64; + h128.high64 = (acc.low64 * XXH_PRIME64_1) + + (acc.high64 * XXH_PRIME64_4) + + ((len - seed) * XXH_PRIME64_2); + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = (XXH64_hash_t)0 - XXH3_avalanche(h128.high64); + return h128; + } + } +} + +static XXH_PUREF XXH128_hash_t +XXH3_finalizeLong_128b(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, xxh_u64 len) +{ + XXH128_hash_t h128; + h128.low64 = XXH3_finalizeLong_64b(acc, secret, len); + h128.high64 = XXH3_mergeAccs(acc, secret + secretSize + - XXH_STRIPE_LEN - XXH_SECRET_MERGEACCS_START, + ~(len * XXH_PRIME64_2)); + return h128; +} + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_hashLong_128b_internal(const void* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[XXH_ACC_NB] = XXH3_INIT_ACC; + + XXH3_hashLong_internal_loop(acc, (const xxh_u8*)input, len, secret, secretSize, f_acc, f_scramble); + + /* converge into final hash */ + XXH_STATIC_ASSERT(sizeof(acc) == 64); + XXH_ASSERT(secretSize >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + return XXH3_finalizeLong_128b(acc, secret, secretSize, (xxh_u64)len); +} + +/* + * It's important for performance that XXH3_hashLong() is not inlined. + */ +XXH_NO_INLINE XXH_PUREF XXH128_hash_t +XXH3_hashLong_128b_default(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; (void)secret; (void)secretLen; + return XXH3_hashLong_128b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * It's important for performance to pass @p secretLen (when it's static) + * to the compiler, so that it can properly optimize the vectorized loop. + * + * When the secret size is unknown, or on GCC 12 where the mix of NO_INLINE and FORCE_INLINE + * breaks -Og, this is XXH_NO_INLINE. + */ +XXH3_WITH_SECRET_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSecret(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; + return XXH3_hashLong_128b_internal(input, len, (const xxh_u8*)secret, secretLen, + XXH3_accumulate, XXH3_scrambleAcc); +} + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSeed_internal(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble, + XXH3_f_initCustomSecret f_initSec) +{ + if (seed64 == 0) + return XXH3_hashLong_128b_internal(input, len, + XXH3_kSecret, sizeof(XXH3_kSecret), + f_acc, f_scramble); + { XXH_ALIGN(XXH_SEC_ALIGN) xxh_u8 secret[XXH_SECRET_DEFAULT_SIZE]; + f_initSec(secret, seed64); + return XXH3_hashLong_128b_internal(input, len, (const xxh_u8*)secret, sizeof(secret), + f_acc, f_scramble); + } +} + +/* + * It's important for performance that XXH3_hashLong is not inlined. + */ +XXH_NO_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSeed(const void* input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)secret; (void)secretLen; + return XXH3_hashLong_128b_withSeed_internal(input, len, seed64, + XXH3_accumulate, XXH3_scrambleAcc, XXH3_initCustomSecret); +} + +typedef XXH128_hash_t (*XXH3_hashLong128_f)(const void* XXH_RESTRICT, size_t, + XXH64_hash_t, const void* XXH_RESTRICT, size_t); + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_128bits_internal(const void* input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen, + XXH3_hashLong128_f f_hl128) +{ + XXH_ASSERT(secretLen >= XXH3_SECRET_SIZE_MIN); + /* + * If an action is to be taken if `secret` conditions are not respected, + * it should be done here. + * For now, it's a contract pre-condition. + * Adding a check and a branch here would cost performance at every hash. + */ + if (len <= 16) + return XXH3_len_0to16_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, seed64); + if (len <= 128) + return XXH3_len_17to128_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_len_129to240_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + return f_hl128(input, len, seed64, secret, secretLen); +} + + +/* === Public XXH128 API === */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t XXH3_128bits(XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_128bits_internal(input, len, 0, + XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_hashLong_128b_default); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSecret(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_128bits_internal(input, len, 0, + (const xxh_u8*)secret, secretSize, + XXH3_hashLong_128b_withSecret); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSeed(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) +{ + return XXH3_128bits_internal(input, len, seed, + XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_hashLong_128b_withSeed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_128bits_internal(input, len, seed, XXH3_kSecret, sizeof(XXH3_kSecret), NULL); + return XXH3_hashLong_128b_withSecret(input, len, seed, secret, secretSize); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH128(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) +{ + return XXH3_128bits_withSeed(input, len, seed); +} + + +/* === XXH3 128-bit streaming === */ +#ifndef XXH_NO_STREAM +/* + * All initialization and update functions are identical to 64-bit streaming variant. + * The only difference is the finalization routine. + */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr) +{ + return XXH3_64bits_reset(statePtr); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_64bits_reset_withSecret(statePtr, secret, secretSize); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed) +{ + return XXH3_64bits_reset_withSeed(statePtr, seed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + return XXH3_64bits_reset_withSecretandSeed(statePtr, secret, secretSize, seed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_update(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_64bits_update(state, input, len); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_digest (XXH_NOESCAPE const XXH3_state_t* state) +{ + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; + if (state->totalLen > XXH3_MIDSIZE_MAX) { + XXH_ALIGN(XXH_ACC_ALIGN) XXH64_hash_t acc[XXH_ACC_NB]; + XXH3_digest_long(acc, state, secret); + XXH_ASSERT(state->secretLimit + XXH_STRIPE_LEN >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + return XXH3_finalizeLong_128b(acc, secret, state->secretLimit + XXH_STRIPE_LEN, (xxh_u64)state->totalLen); + } + /* len <= XXH3_MIDSIZE_MAX : short code */ + if (state->useSeed) + return XXH3_128bits_withSeed(state->buffer, (size_t)state->totalLen, state->seed); + return XXH3_128bits_withSecret(state->buffer, (size_t)(state->totalLen), + secret, state->secretLimit + XXH_STRIPE_LEN); +} +#endif /* !XXH_NO_STREAM */ +/* 128-bit utility functions */ + +#include /* memcmp, memcpy */ + +/* return : 1 is equal, 0 if different */ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2) +{ + /* note : XXH128_hash_t is compact, it has no padding byte */ + return !(memcmp(&h1, &h2, sizeof(h1))); +} + +/* This prototype is compatible with stdlib's qsort(). + * @return : >0 if *h128_1 > *h128_2 + * <0 if *h128_1 < *h128_2 + * =0 if *h128_1 == *h128_2 */ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API int XXH128_cmp(XXH_NOESCAPE const void* h128_1, XXH_NOESCAPE const void* h128_2) +{ + XXH128_hash_t const h1 = *(const XXH128_hash_t*)h128_1; + XXH128_hash_t const h2 = *(const XXH128_hash_t*)h128_2; + int const hcmp = (h1.high64 > h2.high64) - (h2.high64 > h1.high64); + /* note : bets that, in most cases, hash values are different */ + if (hcmp) return hcmp; + return (h1.low64 > h2.low64) - (h2.low64 > h1.low64); +} + + +/*====== Canonical representation ======*/ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API void +XXH128_canonicalFromHash(XXH_NOESCAPE XXH128_canonical_t* dst, XXH128_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH128_canonical_t) == sizeof(XXH128_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) { + hash.high64 = XXH_swap64(hash.high64); + hash.low64 = XXH_swap64(hash.low64); + } + XXH_memcpy(dst, &hash.high64, sizeof(hash.high64)); + XXH_memcpy((char*)dst + sizeof(hash.high64), &hash.low64, sizeof(hash.low64)); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH128_hashFromCanonical(XXH_NOESCAPE const XXH128_canonical_t* src) +{ + XXH128_hash_t h; + h.high64 = XXH_readBE64(src); + h.low64 = XXH_readBE64(src->digest + 8); + return h; +} + + + +/* ========================================== + * Secret generators + * ========================================== + */ +#define XXH_MIN(x, y) (((x) > (y)) ? (y) : (x)) + +XXH_FORCE_INLINE void XXH3_combine16(void* dst, XXH128_hash_t h128) +{ + XXH_writeLE64( dst, XXH_readLE64(dst) ^ h128.low64 ); + XXH_writeLE64( (char*)dst+8, XXH_readLE64((char*)dst+8) ^ h128.high64 ); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_generateSecret(XXH_NOESCAPE void* secretBuffer, size_t secretSize, XXH_NOESCAPE const void* customSeed, size_t customSeedSize) +{ +#if (XXH_DEBUGLEVEL >= 1) + XXH_ASSERT(secretBuffer != NULL); + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); +#else + /* production mode, assert() are disabled */ + if (secretBuffer == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; +#endif + + if (customSeedSize == 0) { + customSeed = XXH3_kSecret; + customSeedSize = XXH_SECRET_DEFAULT_SIZE; + } +#if (XXH_DEBUGLEVEL >= 1) + XXH_ASSERT(customSeed != NULL); +#else + if (customSeed == NULL) return XXH_ERROR; +#endif + + /* Fill secretBuffer with a copy of customSeed - repeat as needed */ + { size_t pos = 0; + while (pos < secretSize) { + size_t const toCopy = XXH_MIN((secretSize - pos), customSeedSize); + memcpy((char*)secretBuffer + pos, customSeed, toCopy); + pos += toCopy; + } } + + { size_t const nbSeg16 = secretSize / 16; + size_t n; + XXH128_canonical_t scrambler; + XXH128_canonicalFromHash(&scrambler, XXH128(customSeed, customSeedSize, 0)); + for (n=0; n. + */ + +#include "app/md5_xxh3_cache.hpp" + +#include +#include +#include +#include +#include +#include +#include + +#if defined(_WIN32) + #ifndef NOMINMAX + #define NOMINMAX + #endif + #include +#endif + +using brokkr::app::Md5Xxh3CacheEntry; + +static int g_pass = 0; +static int g_fail = 0; + +static std::array make_md5(unsigned char seed) { + std::array out{}; + for (std::size_t i = 0; i < out.size(); ++i) out[i] = static_cast(seed + i); + return out; +} + +static std::filesystem::path unique_test_dir() { + const auto stamp = std::chrono::steady_clock::now().time_since_epoch().count(); + return std::filesystem::temp_directory_path() / ("brokkr-md5-xxh3-" + std::to_string(stamp)); +} + +static void fail_msg(const char* label, const std::string& msg) { + std::fprintf(stderr, "FAIL %s: %s\n", label, msg.c_str()); + ++g_fail; +} + +static void pass() { ++g_pass; } + +static void write_text_file(const std::filesystem::path& path, const std::string& content) { + std::filesystem::create_directories(path.parent_path()); + std::ofstream out(path, std::ios::trunc); + out << content; +} + +static void test_roundtrip_save_load() { + const auto dir = unique_test_dir(); + const auto cache_file = brokkr::app::md5_xxh3_cache_file(dir); + + std::vector entries; + brokkr::app::remember_md5_xxh3_cache(entries, make_md5(0x10), 1024, 0x1111222233334444ULL); + brokkr::app::remember_md5_xxh3_cache(entries, make_md5(0x20), 2048, 0x5555666677778888ULL); + + auto save_st = brokkr::app::save_md5_xxh3_cache(cache_file, entries); + if (!save_st) { + fail_msg("roundtrip_save_load", save_st.error()); + std::filesystem::remove_all(dir); + return; + } + + auto loaded = brokkr::app::load_md5_xxh3_cache(cache_file); + if (!loaded) { + fail_msg("roundtrip_save_load", loaded.error()); + std::filesystem::remove_all(dir); + return; + } + + if (loaded->size() != 2) { + fail_msg("roundtrip_save_load", "expected 2 entries after reload"); + std::filesystem::remove_all(dir); + return; + } + + auto hit = brokkr::app::lookup_md5_xxh3_cache(*loaded, make_md5(0x20), 2048); + if (!hit || *hit != 0x5555666677778888ULL) { + fail_msg("roundtrip_save_load", "lookup did not return persisted XXH3 value"); + std::filesystem::remove_all(dir); + return; + } + + pass(); + std::filesystem::remove_all(dir); +} + +static void test_eviction_keeps_most_recent_65535() { + std::vector entries; + for (std::uint64_t i = 0; i < 65540; ++i) { + brokkr::app::remember_md5_xxh3_cache(entries, make_md5(static_cast(i)), i, 0xABC00000ULL + i); + } + + if (entries.size() != 65535) { + fail_msg("eviction_keeps_most_recent_65535", "cache did not clamp to 65535 entries"); + return; + } + + if (brokkr::app::lookup_md5_xxh3_cache(entries, make_md5(0), 0).has_value()) { + fail_msg("eviction_keeps_most_recent_65535", "oldest entry was not evicted"); + return; + } + + auto newest = + brokkr::app::lookup_md5_xxh3_cache(entries, make_md5(static_cast(65539)), 65539); + if (!newest || *newest != 0xABC00000ULL + 65539) { + fail_msg("eviction_keeps_most_recent_65535", "newest entry missing after eviction"); + return; + } + + pass(); +} + +static void test_remember_updates_existing_pair() { + std::vector entries; + const auto md5 = make_md5(0x33); + + brokkr::app::remember_md5_xxh3_cache(entries, md5, 4096, 0x1111ULL); + brokkr::app::remember_md5_xxh3_cache(entries, md5, 4096, 0x2222ULL); + + if (entries.size() != 1) { + fail_msg("remember_updates_existing_pair", "duplicate cache entries were created"); + return; + } + + auto hit = brokkr::app::lookup_md5_xxh3_cache(entries, md5, 4096); + if (!hit || *hit != 0x2222ULL) { + fail_msg("remember_updates_existing_pair", "existing cache entry was not updated"); + return; + } + + pass(); +} + +static void test_forget_removes_existing_pair() { + std::vector entries; + const auto md5 = make_md5(0x66); + + brokkr::app::remember_md5_xxh3_cache(entries, md5, 4096, 0x1111ULL); + brokkr::app::remember_md5_xxh3_cache(entries, make_md5(0x67), 8192, 0x2222ULL); + + if (!brokkr::app::forget_md5_xxh3_cache(entries, md5, 4096)) { + fail_msg("forget_removes_existing_pair", "existing cache entry was not removed"); + return; + } + + if (brokkr::app::lookup_md5_xxh3_cache(entries, md5, 4096).has_value()) { + fail_msg("forget_removes_existing_pair", "removed cache entry was still found"); + return; + } + + if (entries.size() != 1) { + fail_msg("forget_removes_existing_pair", "wrong cache entry count after removal"); + return; + } + + pass(); +} + +static void test_corrupt_primary_falls_back_to_backup() { + const auto dir = unique_test_dir(); + const auto cache_file = brokkr::app::md5_xxh3_cache_file(dir); + auto backup_file = cache_file; + backup_file += ".bak"; + + std::vector entries; + brokkr::app::remember_md5_xxh3_cache(entries, make_md5(0x44), 8192, 0xABCDEF1234567890ULL); + + auto save_st = brokkr::app::save_md5_xxh3_cache(cache_file, entries); + if (!save_st) { + fail_msg("corrupt_primary_falls_back_to_backup", save_st.error()); + std::filesystem::remove_all(dir); + return; + } + + brokkr::app::remember_md5_xxh3_cache(entries, make_md5(0x55), 16384, 0x0123456789ABCDEFULL); + save_st = brokkr::app::save_md5_xxh3_cache(cache_file, entries); + if (!save_st) { + fail_msg("corrupt_primary_falls_back_to_backup", save_st.error()); + std::filesystem::remove_all(dir); + return; + } + + if (!std::filesystem::exists(backup_file)) { + fail_msg("corrupt_primary_falls_back_to_backup", "backup cache file was not created"); + std::filesystem::remove_all(dir); + return; + } + + write_text_file(cache_file, "not a valid cache\n123\n"); + + auto loaded = brokkr::app::load_md5_xxh3_cache(cache_file); + if (!loaded) { + fail_msg("corrupt_primary_falls_back_to_backup", loaded.error()); + std::filesystem::remove_all(dir); + return; + } + + auto hit = brokkr::app::lookup_md5_xxh3_cache(*loaded, make_md5(0x44), 8192); + if (!hit || *hit != 0xABCDEF1234567890ULL) { + fail_msg("corrupt_primary_falls_back_to_backup", "did not recover valid entry from backup cache"); + std::filesystem::remove_all(dir); + return; + } + + pass(); + std::filesystem::remove_all(dir); +} + +static void test_headerless_cache_is_rejected() { + const auto dir = unique_test_dir(); + const auto cache_file = brokkr::app::md5_xxh3_cache_file(dir); + + write_text_file(cache_file, "1 4096 00112233445566778899aabbccddeeff 0123456789abcdef\n"); + + auto loaded = brokkr::app::load_md5_xxh3_cache(cache_file); + if (!loaded) { + fail_msg("headerless_cache_is_rejected", loaded.error()); + std::filesystem::remove_all(dir); + return; + } + + if (!loaded->empty()) { + fail_msg("headerless_cache_is_rejected", "headerless cache content should be ignored"); + std::filesystem::remove_all(dir); + return; + } + + pass(); + std::filesystem::remove_all(dir); +} + +#if defined(_WIN32) +static void test_locked_cache_file_fails_fast() { + const auto dir = unique_test_dir(); + const auto cache_file = brokkr::app::md5_xxh3_cache_file(dir); + + std::vector entries; + brokkr::app::remember_md5_xxh3_cache(entries, make_md5(0x77), 4096, 0xABCDEF1234567890ULL); + + auto save_st = brokkr::app::save_md5_xxh3_cache(cache_file, entries); + if (!save_st) { + fail_msg("locked_cache_file_fails_fast", save_st.error()); + std::filesystem::remove_all(dir); + return; + } + + HANDLE h = CreateFileW(cache_file.c_str(), GENERIC_READ, FILE_SHARE_READ | FILE_SHARE_WRITE, nullptr, OPEN_EXISTING, + FILE_ATTRIBUTE_NORMAL, nullptr); + if (h == INVALID_HANDLE_VALUE) { + fail_msg("locked_cache_file_fails_fast", "could not open cache file for lock simulation"); + std::filesystem::remove_all(dir); + return; + } + + brokkr::app::remember_md5_xxh3_cache(entries, make_md5(0x78), 8192, 0x0123456789ABCDEFULL); + save_st = brokkr::app::save_md5_xxh3_cache(cache_file, entries); + CloseHandle(h); + + if (save_st) { + fail_msg("locked_cache_file_fails_fast", "save unexpectedly succeeded while cache file was locked"); + std::filesystem::remove_all(dir); + return; + } + + if (save_st.error().find("Cache file busy") == std::string::npos) { + fail_msg("locked_cache_file_fails_fast", "locked cache did not report a busy-file failure"); + std::filesystem::remove_all(dir); + return; + } + + pass(); + std::filesystem::remove_all(dir); +} +#endif + +int main() { + test_roundtrip_save_load(); + test_eviction_keeps_most_recent_65535(); + test_remember_updates_existing_pair(); + test_forget_removes_existing_pair(); + test_corrupt_primary_falls_back_to_backup(); + test_headerless_cache_is_rejected(); +#if defined(_WIN32) + test_locked_cache_file_fails_fast(); +#endif + + std::fprintf(stdout, "md5_xxh3_cache: %d passed, %d failed\n", g_pass, g_fail); + return g_fail ? 1 : 0; +} \ No newline at end of file