diff options
166 files changed, 5439 insertions, 2538 deletions
diff --git a/.ci/templates/build-msvc.yml b/.ci/templates/build-msvc.yml index 721179550..097367018 100644 --- a/.ci/templates/build-msvc.yml +++ b/.ci/templates/build-msvc.yml @@ -8,7 +8,7 @@ steps: displayName: 'Install vulkan-sdk' - script: python -m pip install --upgrade pip conan displayName: 'Install conan' -- script: refreshenv && mkdir build && cd build && cmake -G "Visual Studio 16 2019" -A x64 --config Release -DYUZU_USE_BUNDLED_QT=1 -DYUZU_USE_QT_WEB_ENGINE=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${COMPAT} -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON -DDISPLAY_VERSION=${{ parameters['version'] }} .. && cd .. +- script: refreshenv && mkdir build && cd build && cmake -G "Visual Studio 16 2019" -A x64 -DYUZU_USE_BUNDLED_QT=1 -DYUZU_USE_QT_WEB_ENGINE=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${COMPAT} -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON -DDISPLAY_VERSION=${{ parameters['version'] }} .. && cmake --install . --config Release && cd .. displayName: 'Configure CMake' - task: MSBuild@1 displayName: 'Build' diff --git a/CMakeLists.txt b/CMakeLists.txt index ac7c3ce90..6ea6c650e 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -172,6 +172,8 @@ macro(yuzu_find_packages) "nlohmann_json 3.8 nlohmann_json/3.8.0" "ZLIB 1.2 zlib/1.2.11" "zstd 1.4 zstd/1.4.8" + # can't use opus until AVX check is fixed: https://github.com/yuzu-emu/yuzu/pull/4068 + #"opus 1.3 opus/1.3.1" ) foreach(PACKAGE ${REQUIRED_LIBS}) diff --git a/externals/CMakeLists.txt b/externals/CMakeLists.txt index c629bbc5c..891a47c3c 100644 --- a/externals/CMakeLists.txt +++ b/externals/CMakeLists.txt @@ -64,8 +64,8 @@ endif() add_subdirectory(sirit) # libzip -find_package(Libzip 1.5) -if (NOT LIBZIP_FOUND) +find_package(libzip 1.5) +if (NOT libzip_FOUND) message(STATUS "libzip 1.5 or newer not found, falling back to externals") add_subdirectory(libzip EXCLUDE_FROM_ALL) endif() @@ -97,4 +97,8 @@ if (ENABLE_WEB_SERVICE) endif() # Opus -add_subdirectory(opus) +find_package(opus 1.3) +if (NOT opus_FOUND) + message(STATUS "opus 1.3 or newer not found, falling back to externals") + add_subdirectory(opus EXCLUDE_FROM_ALL) +endif() diff --git a/externals/dynarmic b/externals/dynarmic -Subproject 646fd0592091c5c1e5899fa715bd7b7fcc977a3 +Subproject c788bcdf17e6bc1d1a1dd315106b952013f5ecb diff --git a/externals/find-modules/FindLibzip.cmake b/externals/find-modules/FindLibzip.cmake deleted file mode 100644 index f36b1687a..000000000 --- a/externals/find-modules/FindLibzip.cmake +++ /dev/null @@ -1,72 +0,0 @@ - -find_package(PkgConfig QUIET) -pkg_check_modules(PC_LIBZIP QUIET libzip) - -find_path(LIBZIP_INCLUDE_DIR - NAMES zip.h - PATHS ${PC_LIBZIP_INCLUDE_DIRS} - "$ENV{LIB_DIR}/include" - "$ENV{INCLUDE}" - /usr/local/include - /usr/include -) -find_path(LIBZIP_INCLUDE_DIR_ZIPCONF - NAMES zipconf.h - HINTS ${PC_LIBZIP_INCLUDE_DIRS} - "$ENV{LIB_DIR}/include" - "$ENV{LIB_DIR}/lib/libzip/include" - "$ENV{LIB}/lib/libzip/include" - /usr/local/lib/libzip/include - /usr/lib/libzip/include - /usr/local/include - /usr/include - "$ENV{INCLUDE}" -) -find_library(LIBZIP_LIBRARY - NAMES zip - PATHS ${PC_LIBZIP_LIBRARY_DIRS} - "$ENV{LIB_DIR}/lib" "$ENV{LIB}" /usr/local/lib /usr/lib -) - -if (LIBZIP_INCLUDE_DIR_ZIPCONF) - FILE(READ "${LIBZIP_INCLUDE_DIR_ZIPCONF}/zipconf.h" _LIBZIP_VERSION_CONTENTS) - if (_LIBZIP_VERSION_CONTENTS) - STRING(REGEX REPLACE ".*#define LIBZIP_VERSION \"([0-9.]+)\".*" "\\1" LIBZIP_VERSION "${_LIBZIP_VERSION_CONTENTS}") - endif() - unset(_LIBZIP_VERSION_CONTENTS) -endif() - -set(LIBZIP_VERSION ${LIBZIP_VERSION} CACHE STRING "Version number of libzip") - -include(FindPackageHandleStandardArgs) -find_package_handle_standard_args(Libzip - FOUND_VAR LIBZIP_FOUND - REQUIRED_VARS - LIBZIP_LIBRARY - LIBZIP_INCLUDE_DIR - LIBZIP_INCLUDE_DIR_ZIPCONF - LIBZIP_VERSION - VERSION_VAR LIBZIP_VERSION -) - -if(LIBZIP_FOUND) - set(LIBZIP_LIBRARIES ${LIBZIP_LIBRARY}) - set(LIBZIP_INCLUDE_DIRS ${LIBZIP_INCLUDE_DIR}) - set(LIBZIP_DEFINITIONS ${PC_LIBZIP_CFLAGS_OTHER}) -endif() - -if(LIBZIP_FOUND AND NOT TARGET libzip::libzip) - add_library(libzip::libzip UNKNOWN IMPORTED) - set_target_properties(libzip::libzip PROPERTIES - IMPORTED_LOCATION "${LIBZIP_LIBRARY}" - INTERFACE_COMPILE_OPTIONS "${PC_LIBZIP_CFLAGS_OTHER}" - INTERFACE_INCLUDE_DIRECTORIES "${LIBZIP_INCLUDE_DIR}" - ) -endif() - -mark_as_advanced( - LIBZIP_INCLUDE_DIR - LIBZIP_INCLUDE_DIR_ZIPCONF - LIBZIP_LIBRARY - LIBZIP_VERSION -) diff --git a/externals/find-modules/Findlibzip.cmake b/externals/find-modules/Findlibzip.cmake new file mode 100644 index 000000000..8934de3b8 --- /dev/null +++ b/externals/find-modules/Findlibzip.cmake @@ -0,0 +1,72 @@ + +find_package(PkgConfig QUIET) +pkg_check_modules(PC_libzip QUIET libzip) + +find_path(libzip_INCLUDE_DIR + NAMES zip.h + PATHS ${PC_libzip_INCLUDE_DIRS} + "$ENV{LIB_DIR}/include" + "$ENV{INCLUDE}" + /usr/local/include + /usr/include +) +find_path(libzip_INCLUDE_DIR_ZIPCONF + NAMES zipconf.h + HINTS ${PC_libzip_INCLUDE_DIRS} + "$ENV{LIB_DIR}/include" + "$ENV{LIB_DIR}/lib/libzip/include" + "$ENV{LIB}/lib/libzip/include" + /usr/local/lib/libzip/include + /usr/lib/libzip/include + /usr/local/include + /usr/include + "$ENV{INCLUDE}" +) +find_library(libzip_LIBRARY + NAMES zip + PATHS ${PC_libzip_LIBRARY_DIRS} + "$ENV{LIB_DIR}/lib" "$ENV{LIB}" /usr/local/lib /usr/lib +) + +if (libzip_INCLUDE_DIR_ZIPCONF) + FILE(READ "${libzip_INCLUDE_DIR_ZIPCONF}/zipconf.h" _libzip_VERSION_CONTENTS) + if (_libzip_VERSION_CONTENTS) + STRING(REGEX REPLACE ".*#define LIBZIP_VERSION \"([0-9.]+)\".*" "\\1" libzip_VERSION "${_libzip_VERSION_CONTENTS}") + endif() + unset(_libzip_VERSION_CONTENTS) +endif() + +set(libzip_VERSION ${libzip_VERSION} CACHE STRING "Version number of libzip") + +include(FindPackageHandleStandardArgs) +find_package_handle_standard_args(libzip + FOUND_VAR libzip_FOUND + REQUIRED_VARS + libzip_LIBRARY + libzip_INCLUDE_DIR + libzip_INCLUDE_DIR_ZIPCONF + libzip_VERSION + VERSION_VAR libzip_VERSION +) + +if(libzip_FOUND) + set(libzip_LIBRARIES ${libzip_LIBRARY}) + set(libzip_INCLUDE_DIRS ${libzip_INCLUDE_DIR}) + set(libzip_DEFINITIONS ${PC_libzip_CFLAGS_OTHER}) +endif() + +if(libzip_FOUND AND NOT TARGET libzip::libzip) + add_library(libzip::libzip UNKNOWN IMPORTED) + set_target_properties(libzip::libzip PROPERTIES + IMPORTED_LOCATION "${libzip_LIBRARY}" + INTERFACE_COMPILE_OPTIONS "${PC_libzip_CFLAGS_OTHER}" + INTERFACE_INCLUDE_DIRECTORIES "${libzip_INCLUDE_DIR}" + ) +endif() + +mark_as_advanced( + libzip_INCLUDE_DIR + libzip_INCLUDE_DIR_ZIPCONF + libzip_LIBRARY + libzip_VERSION +) diff --git a/externals/find-modules/Findopus.cmake b/externals/find-modules/Findopus.cmake index de84bd995..2bce56122 100644 --- a/externals/find-modules/Findopus.cmake +++ b/externals/find-modules/Findopus.cmake @@ -28,7 +28,7 @@ if(opus_FOUND) endif() if(opus_FOUND AND NOT TARGET Opus::Opus) - add_library(Opus::Opus UNKNOWN IMPORTED) + add_library(Opus::Opus UNKNOWN IMPORTED GLOBAL) set_target_properties(Opus::Opus PROPERTIES IMPORTED_LOCATION "${opus_LIBRARY}" INTERFACE_COMPILE_OPTIONS "${PC_opus_CFLAGS_OTHER}" diff --git a/externals/opus/CMakeLists.txt b/externals/opus/CMakeLists.txt index 94a86551f..16f5af9f2 100644 --- a/externals/opus/CMakeLists.txt +++ b/externals/opus/CMakeLists.txt @@ -252,3 +252,5 @@ PRIVATE opus/silk/float opus/src ) + +add_library(Opus::Opus ALIAS opus) diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt index 788516ded..9f8dafa3b 100644 --- a/src/common/CMakeLists.txt +++ b/src/common/CMakeLists.txt @@ -97,6 +97,7 @@ add_custom_command(OUTPUT scm_rev.cpp add_library(common STATIC algorithm.h alignment.h + assert.cpp assert.h atomic_ops.h detached_tasks.cpp @@ -109,6 +110,7 @@ add_library(common STATIC cityhash.h common_funcs.h common_paths.h + common_sizes.h common_types.h concepts.h div_ceil.h diff --git a/src/common/assert.cpp b/src/common/assert.cpp new file mode 100644 index 000000000..d7d91b96b --- /dev/null +++ b/src/common/assert.cpp @@ -0,0 +1,11 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" + +#include "common/common_funcs.h" + +void assert_handle_failure() { + Crash(); +} diff --git a/src/common/assert.h b/src/common/assert.h index 06d7b5612..b3ba35c0f 100644 --- a/src/common/assert.h +++ b/src/common/assert.h @@ -4,10 +4,13 @@ #pragma once -#include <cstdlib> -#include "common/common_funcs.h" #include "common/logging/log.h" +// Sometimes we want to try to continue even after hitting an assert. +// However touching this file yields a global recompilation as this header is included almost +// everywhere. So let's just move the handling of the failed assert to a single cpp file. +void assert_handle_failure(); + // For asserts we'd like to keep all the junk executed when an assert happens away from the // important code in the function. One way of doing this is to put all the relevant code inside a // lambda and force the compiler to not inline it. Unfortunately, MSVC seems to have no syntax to @@ -17,15 +20,14 @@ // enough for our purposes. template <typename Fn> #if defined(_MSC_VER) -[[msvc::noinline, noreturn]] +[[msvc::noinline]] #elif defined(__GNUC__) -[[gnu::cold, gnu::noinline, noreturn]] +[[gnu::cold, gnu::noinline]] #endif static void assert_noinline_call(const Fn& fn) { fn(); - Crash(); - exit(1); // Keeps GCC's mouth shut about this actually returning + assert_handle_failure(); } #define ASSERT(_a_) \ diff --git a/src/common/common_sizes.h b/src/common/common_sizes.h new file mode 100644 index 000000000..7e9fd968b --- /dev/null +++ b/src/common/common_sizes.h @@ -0,0 +1,43 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <limits> + +#include "common/common_types.h" + +namespace Common { + +enum : u64 { + Size_1_KB = 0x400ULL, + Size_64_KB = 64ULL * Size_1_KB, + Size_128_KB = 128ULL * Size_1_KB, + Size_1_MB = 0x100000ULL, + Size_2_MB = 2ULL * Size_1_MB, + Size_4_MB = 4ULL * Size_1_MB, + Size_5_MB = 5ULL * Size_1_MB, + Size_14_MB = 14ULL * Size_1_MB, + Size_32_MB = 32ULL * Size_1_MB, + Size_33_MB = 33ULL * Size_1_MB, + Size_128_MB = 128ULL * Size_1_MB, + Size_448_MB = 448ULL * Size_1_MB, + Size_507_MB = 507ULL * Size_1_MB, + Size_562_MB = 562ULL * Size_1_MB, + Size_1554_MB = 1554ULL * Size_1_MB, + Size_2048_MB = 2048ULL * Size_1_MB, + Size_2193_MB = 2193ULL * Size_1_MB, + Size_3285_MB = 3285ULL * Size_1_MB, + Size_4916_MB = 4916ULL * Size_1_MB, + Size_1_GB = 0x40000000ULL, + Size_2_GB = 2ULL * Size_1_GB, + Size_4_GB = 4ULL * Size_1_GB, + Size_6_GB = 6ULL * Size_1_GB, + Size_8_GB = 8ULL * Size_1_GB, + Size_64_GB = 64ULL * Size_1_GB, + Size_512_GB = 512ULL * Size_1_GB, + Size_Invalid = std::numeric_limits<u64>::max(), +}; + +} // namespace Common diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp index 2d4d2e9e7..4575df24d 100644 --- a/src/common/logging/backend.cpp +++ b/src/common/logging/backend.cpp @@ -212,6 +212,7 @@ void DebuggerBackend::Write(const Entry& entry) { SUB(Service, ARP) \ SUB(Service, BCAT) \ SUB(Service, BPC) \ + SUB(Service, BGTC) \ SUB(Service, BTDRV) \ SUB(Service, BTM) \ SUB(Service, Capture) \ diff --git a/src/common/logging/log.h b/src/common/logging/log.h index 835894918..3d7b7dab7 100644 --- a/src/common/logging/log.h +++ b/src/common/logging/log.h @@ -66,6 +66,7 @@ enum class Class : ClassType { Service_ARP, ///< The ARP service Service_Audio, ///< The Audio (Audio control) service Service_BCAT, ///< The BCAT service + Service_BGTC, ///< The BGTC (Background Task Controller) service Service_BPC, ///< The BPC service Service_BTDRV, ///< The Bluetooth driver service Service_BTM, ///< The BTM service diff --git a/src/common/threadsafe_queue.h b/src/common/threadsafe_queue.h index a4647314a..ad04df8ca 100644 --- a/src/common/threadsafe_queue.h +++ b/src/common/threadsafe_queue.h @@ -83,11 +83,15 @@ public: return true; } - T PopWait() { + void Wait() { if (Empty()) { std::unique_lock lock{cv_mutex}; cv.wait(lock, [this]() { return !Empty(); }); } + } + + T PopWait() { + Wait(); T t; Pop(t); return t; @@ -156,6 +160,10 @@ public: return spsc_queue.Pop(t); } + void Wait() { + spsc_queue.Wait(); + } + T PopWait() { return spsc_queue.PopWait(); } diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt index 17f251c37..167ee13f3 100644 --- a/src/core/CMakeLists.txt +++ b/src/core/CMakeLists.txt @@ -141,6 +141,9 @@ add_library(core STATIC hardware_interrupt_manager.h hle/ipc.h hle/ipc_helpers.h + hle/kernel/board/nintendo/nx/k_system_control.cpp + hle/kernel/board/nintendo/nx/k_system_control.h + hle/kernel/board/nintendo/nx/secure_monitor.h hle/kernel/client_port.cpp hle/kernel/client_port.h hle/kernel/client_session.cpp @@ -169,9 +172,13 @@ add_library(core STATIC hle/kernel/k_memory_block.h hle/kernel/k_memory_block_manager.cpp hle/kernel/k_memory_block_manager.h + hle/kernel/k_memory_layout.cpp + hle/kernel/k_memory_layout.board.nintendo_nx.cpp hle/kernel/k_memory_layout.h hle/kernel/k_memory_manager.cpp hle/kernel/k_memory_manager.h + hle/kernel/k_memory_region.h + hle/kernel/k_memory_region_type.h hle/kernel/k_page_bitmap.h hle/kernel/k_page_heap.cpp hle/kernel/k_page_heap.h @@ -196,11 +203,11 @@ add_library(core STATIC hle/kernel/k_spin_lock.h hle/kernel/k_synchronization_object.cpp hle/kernel/k_synchronization_object.h - hle/kernel/k_system_control.cpp hle/kernel/k_system_control.h hle/kernel/k_thread.cpp hle/kernel/k_thread.h hle/kernel/k_thread_queue.h + hle/kernel/k_trace.h hle/kernel/k_writable_event.cpp hle/kernel/k_writable_event.h hle/kernel/kernel.cpp @@ -666,7 +673,7 @@ endif() create_target_directory_groups(core) target_link_libraries(core PUBLIC common PRIVATE audio_core video_core) -target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt::fmt nlohmann_json::nlohmann_json mbedtls opus zip) +target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt::fmt nlohmann_json::nlohmann_json mbedtls Opus::Opus zip) if (YUZU_ENABLE_BOXCAT) target_compile_definitions(core PRIVATE -DYUZU_ENABLE_BOXCAT) diff --git a/src/core/arm/dynarmic/arm_dynarmic_32.cpp b/src/core/arm/dynarmic/arm_dynarmic_32.cpp index 53d78de32..08d889135 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_32.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic_32.cpp @@ -114,18 +114,17 @@ public: static constexpr u64 minimum_run_cycles = 1000U; }; -std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable& page_table, - std::size_t address_space_bits) const { +std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable* page_table) const { Dynarmic::A32::UserConfig config; config.callbacks = cb.get(); - // TODO(bunnei): Implement page table for 32-bit - // config.page_table = &page_table.pointers; config.coprocessors[15] = cp15; config.define_unpredictable_behaviour = true; static constexpr std::size_t PAGE_BITS = 12; static constexpr std::size_t NUM_PAGE_TABLE_ENTRIES = 1 << (32 - PAGE_BITS); - config.page_table = reinterpret_cast<std::array<std::uint8_t*, NUM_PAGE_TABLE_ENTRIES>*>( - page_table.pointers.data()); + if (page_table) { + config.page_table = reinterpret_cast<std::array<std::uint8_t*, NUM_PAGE_TABLE_ENTRIES>*>( + page_table->pointers.data()); + } config.absolute_offset_page_table = true; config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS; config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128; @@ -138,6 +137,10 @@ std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable& // Timing config.wall_clock_cntpct = uses_wall_clock; + // Code cache size + config.code_cache_size = 512 * 1024 * 1024; + config.far_code_offset = 256 * 1024 * 1024; + // Safe optimizations if (Settings::values.cpu_accuracy == Settings::CPUAccuracy::DebugMode) { if (!Settings::values.cpuopt_page_tables) { @@ -201,7 +204,8 @@ ARM_Dynarmic_32::ARM_Dynarmic_32(System& system, CPUInterrupts& interrupt_handle : ARM_Interface{system, interrupt_handlers, uses_wall_clock}, cb(std::make_unique<DynarmicCallbacks32>(*this)), cp15(std::make_shared<DynarmicCP15>(*this)), core_index{core_index}, - exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {} + exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)}, + jit(MakeJit(nullptr)) {} ARM_Dynarmic_32::~ARM_Dynarmic_32() = default; @@ -256,9 +260,6 @@ void ARM_Dynarmic_32::ChangeProcessorID(std::size_t new_core_id) { } void ARM_Dynarmic_32::SaveContext(ThreadContext32& ctx) { - if (!jit) { - return; - } Dynarmic::A32::Context context; jit->SaveContext(context); ctx.cpu_registers = context.Regs(); @@ -268,9 +269,6 @@ void ARM_Dynarmic_32::SaveContext(ThreadContext32& ctx) { } void ARM_Dynarmic_32::LoadContext(const ThreadContext32& ctx) { - if (!jit) { - return; - } Dynarmic::A32::Context context; context.Regs() = ctx.cpu_registers; context.ExtRegs() = ctx.extension_registers; @@ -284,23 +282,14 @@ void ARM_Dynarmic_32::PrepareReschedule() { } void ARM_Dynarmic_32::ClearInstructionCache() { - if (!jit) { - return; - } jit->ClearCache(); } void ARM_Dynarmic_32::InvalidateCacheRange(VAddr addr, std::size_t size) { - if (!jit) { - return; - } jit->InvalidateCacheRange(static_cast<u32>(addr), size); } void ARM_Dynarmic_32::ClearExclusiveState() { - if (!jit) { - return; - } jit->ClearExclusiveState(); } @@ -316,7 +305,7 @@ void ARM_Dynarmic_32::PageTableChanged(Common::PageTable& page_table, LoadContext(ctx); return; } - jit = MakeJit(page_table, new_address_space_size_in_bits); + jit = MakeJit(&page_table); LoadContext(ctx); jit_cache.emplace(key, jit); } diff --git a/src/core/arm/dynarmic/arm_dynarmic_32.h b/src/core/arm/dynarmic/arm_dynarmic_32.h index f6c4d4db9..d40aef7a9 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_32.h +++ b/src/core/arm/dynarmic/arm_dynarmic_32.h @@ -68,8 +68,7 @@ public: std::size_t new_address_space_size_in_bits) override; private: - std::shared_ptr<Dynarmic::A32::Jit> MakeJit(Common::PageTable& page_table, - std::size_t address_space_bits) const; + std::shared_ptr<Dynarmic::A32::Jit> MakeJit(Common::PageTable* page_table) const; using JitCacheKey = std::pair<Common::PageTable*, std::size_t>; using JitCacheType = @@ -80,10 +79,10 @@ private: std::unique_ptr<DynarmicCallbacks32> cb; JitCacheType jit_cache; - std::shared_ptr<Dynarmic::A32::Jit> jit; std::shared_ptr<DynarmicCP15> cp15; std::size_t core_index; DynarmicExclusiveMonitor& exclusive_monitor; + std::shared_ptr<Dynarmic::A32::Jit> jit; }; } // namespace Core diff --git a/src/core/arm/dynarmic/arm_dynarmic_64.cpp b/src/core/arm/dynarmic/arm_dynarmic_64.cpp index b36b7d918..e12e50658 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_64.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic_64.cpp @@ -142,7 +142,7 @@ public: static constexpr u64 minimum_run_cycles = 1000U; }; -std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& page_table, +std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable* page_table, std::size_t address_space_bits) const { Dynarmic::A64::UserConfig config; @@ -150,13 +150,15 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& config.callbacks = cb.get(); // Memory - config.page_table = reinterpret_cast<void**>(page_table.pointers.data()); - config.page_table_address_space_bits = address_space_bits; - config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS; - config.silently_mirror_page_table = false; - config.absolute_offset_page_table = true; - config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128; - config.only_detect_misalignment_via_page_table_on_page_boundary = true; + if (page_table) { + config.page_table = reinterpret_cast<void**>(page_table->pointers.data()); + config.page_table_address_space_bits = address_space_bits; + config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS; + config.silently_mirror_page_table = false; + config.absolute_offset_page_table = true; + config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128; + config.only_detect_misalignment_via_page_table_on_page_boundary = true; + } // Multi-process state config.processor_id = core_index; @@ -175,6 +177,10 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& // Timing config.wall_clock_cntpct = uses_wall_clock; + // Code cache size + config.code_cache_size = 512 * 1024 * 1024; + config.far_code_offset = 256 * 1024 * 1024; + // Safe optimizations if (Settings::values.cpu_accuracy == Settings::CPUAccuracy::DebugMode) { if (!Settings::values.cpuopt_page_tables) { @@ -237,7 +243,8 @@ ARM_Dynarmic_64::ARM_Dynarmic_64(System& system, CPUInterrupts& interrupt_handle std::size_t core_index) : ARM_Interface{system, interrupt_handlers, uses_wall_clock}, cb(std::make_unique<DynarmicCallbacks64>(*this)), core_index{core_index}, - exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {} + exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)}, + jit(MakeJit(nullptr, 48)) {} ARM_Dynarmic_64::~ARM_Dynarmic_64() = default; @@ -294,9 +301,6 @@ void ARM_Dynarmic_64::ChangeProcessorID(std::size_t new_core_id) { } void ARM_Dynarmic_64::SaveContext(ThreadContext64& ctx) { - if (!jit) { - return; - } ctx.cpu_registers = jit->GetRegisters(); ctx.sp = jit->GetSP(); ctx.pc = jit->GetPC(); @@ -308,9 +312,6 @@ void ARM_Dynarmic_64::SaveContext(ThreadContext64& ctx) { } void ARM_Dynarmic_64::LoadContext(const ThreadContext64& ctx) { - if (!jit) { - return; - } jit->SetRegisters(ctx.cpu_registers); jit->SetSP(ctx.sp); jit->SetPC(ctx.pc); @@ -326,23 +327,14 @@ void ARM_Dynarmic_64::PrepareReschedule() { } void ARM_Dynarmic_64::ClearInstructionCache() { - if (!jit) { - return; - } jit->ClearCache(); } void ARM_Dynarmic_64::InvalidateCacheRange(VAddr addr, std::size_t size) { - if (!jit) { - return; - } jit->InvalidateCacheRange(addr, size); } void ARM_Dynarmic_64::ClearExclusiveState() { - if (!jit) { - return; - } jit->ClearExclusiveState(); } @@ -358,7 +350,7 @@ void ARM_Dynarmic_64::PageTableChanged(Common::PageTable& page_table, LoadContext(ctx); return; } - jit = MakeJit(page_table, new_address_space_size_in_bits); + jit = MakeJit(&page_table, new_address_space_size_in_bits); LoadContext(ctx); jit_cache.emplace(key, jit); } diff --git a/src/core/arm/dynarmic/arm_dynarmic_64.h b/src/core/arm/dynarmic/arm_dynarmic_64.h index 329b59a32..edef04376 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_64.h +++ b/src/core/arm/dynarmic/arm_dynarmic_64.h @@ -61,7 +61,7 @@ public: std::size_t new_address_space_size_in_bits) override; private: - std::shared_ptr<Dynarmic::A64::Jit> MakeJit(Common::PageTable& page_table, + std::shared_ptr<Dynarmic::A64::Jit> MakeJit(Common::PageTable* page_table, std::size_t address_space_bits) const; using JitCacheKey = std::pair<Common::PageTable*, std::size_t>; @@ -71,10 +71,11 @@ private: friend class DynarmicCallbacks64; std::unique_ptr<DynarmicCallbacks64> cb; JitCacheType jit_cache; - std::shared_ptr<Dynarmic::A64::Jit> jit; std::size_t core_index; DynarmicExclusiveMonitor& exclusive_monitor; + + std::shared_ptr<Dynarmic::A64::Jit> jit; }; } // namespace Core diff --git a/src/core/core.cpp b/src/core/core.cpp index 305f56ff1..56b47e671 100644 --- a/src/core/core.cpp +++ b/src/core/core.cpp @@ -296,7 +296,7 @@ struct System::Impl { exit_lock = false; if (gpu_core) { - gpu_core->WaitIdle(); + gpu_core->ShutDown(); } services.reset(); diff --git a/src/core/hle/kernel/arch/arm64/k_memory_region_device_types.inc b/src/core/hle/kernel/arch/arm64/k_memory_region_device_types.inc new file mode 100644 index 000000000..857b512ba --- /dev/null +++ b/src/core/hle/kernel/arch/arm64/k_memory_region_device_types.inc @@ -0,0 +1,20 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// All architectures must define NumArchitectureDeviceRegions. +constexpr inline const auto NumArchitectureDeviceRegions = 3; + +constexpr inline const auto KMemoryRegionType_Uart = + KMemoryRegionType_ArchDeviceBase.DeriveSparse(0, NumArchitectureDeviceRegions, 0); +constexpr inline const auto KMemoryRegionType_InterruptCpuInterface = + KMemoryRegionType_ArchDeviceBase.DeriveSparse(0, NumArchitectureDeviceRegions, 1) + .SetAttribute(KMemoryRegionAttr_NoUserMap); +constexpr inline const auto KMemoryRegionType_InterruptDistributor = + KMemoryRegionType_ArchDeviceBase.DeriveSparse(0, NumArchitectureDeviceRegions, 2) + .SetAttribute(KMemoryRegionAttr_NoUserMap); +static_assert(KMemoryRegionType_Uart.GetValue() == (0x1D)); +static_assert(KMemoryRegionType_InterruptCpuInterface.GetValue() == + (0x2D | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_InterruptDistributor.GetValue() == + (0x4D | KMemoryRegionAttr_NoUserMap)); diff --git a/src/core/hle/kernel/board/nintendo/nx/k_memory_region_device_types.inc b/src/core/hle/kernel/board/nintendo/nx/k_memory_region_device_types.inc new file mode 100644 index 000000000..58d6c0b16 --- /dev/null +++ b/src/core/hle/kernel/board/nintendo/nx/k_memory_region_device_types.inc @@ -0,0 +1,52 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// All architectures must define NumBoardDeviceRegions. +constexpr inline const auto NumBoardDeviceRegions = 6; +// UNUSED: .Derive(NumBoardDeviceRegions, 0); +constexpr inline const auto KMemoryRegionType_MemoryController = + KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 1) + .SetAttribute(KMemoryRegionAttr_NoUserMap); +constexpr inline const auto KMemoryRegionType_MemoryController1 = + KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 2) + .SetAttribute(KMemoryRegionAttr_NoUserMap); +constexpr inline const auto KMemoryRegionType_MemoryController0 = + KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 3) + .SetAttribute(KMemoryRegionAttr_NoUserMap); +constexpr inline const auto KMemoryRegionType_PowerManagementController = + KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 4).DeriveTransition(); +constexpr inline const auto KMemoryRegionType_LegacyLpsDevices = + KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 5); +static_assert(KMemoryRegionType_MemoryController.GetValue() == + (0x55 | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_MemoryController1.GetValue() == + (0x65 | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_MemoryController0.GetValue() == + (0x95 | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_PowerManagementController.GetValue() == (0x1A5)); + +static_assert(KMemoryRegionType_LegacyLpsDevices.GetValue() == 0xC5); + +constexpr inline const auto NumLegacyLpsDevices = 7; +constexpr inline const auto KMemoryRegionType_LegacyLpsExceptionVectors = + KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 0); +constexpr inline const auto KMemoryRegionType_LegacyLpsIram = + KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 1); +constexpr inline const auto KMemoryRegionType_LegacyLpsFlowController = + KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 2); +constexpr inline const auto KMemoryRegionType_LegacyLpsPrimaryICtlr = + KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 3); +constexpr inline const auto KMemoryRegionType_LegacyLpsSemaphore = + KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 4); +constexpr inline const auto KMemoryRegionType_LegacyLpsAtomics = + KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 5); +constexpr inline const auto KMemoryRegionType_LegacyLpsClkRst = + KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 6); +static_assert(KMemoryRegionType_LegacyLpsExceptionVectors.GetValue() == 0x3C5); +static_assert(KMemoryRegionType_LegacyLpsIram.GetValue() == 0x5C5); +static_assert(KMemoryRegionType_LegacyLpsFlowController.GetValue() == 0x6C5); +static_assert(KMemoryRegionType_LegacyLpsPrimaryICtlr.GetValue() == 0x9C5); +static_assert(KMemoryRegionType_LegacyLpsSemaphore.GetValue() == 0xAC5); +static_assert(KMemoryRegionType_LegacyLpsAtomics.GetValue() == 0xCC5); +static_assert(KMemoryRegionType_LegacyLpsClkRst.GetValue() == 0x11C5); diff --git a/src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp b/src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp new file mode 100644 index 000000000..86472b5ce --- /dev/null +++ b/src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp @@ -0,0 +1,164 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <random> + +#include "common/common_sizes.h" +#include "core/hle/kernel/board/nintendo/nx/k_system_control.h" +#include "core/hle/kernel/board/nintendo/nx/secure_monitor.h" +#include "core/hle/kernel/k_trace.h" + +namespace Kernel::Board::Nintendo::Nx { + +namespace impl { + +constexpr const std::size_t RequiredNonSecureSystemMemorySizeVi = 0x2238 * 4 * 1024; +constexpr const std::size_t RequiredNonSecureSystemMemorySizeNvservices = 0x710 * 4 * 1024; +constexpr const std::size_t RequiredNonSecureSystemMemorySizeMisc = 0x80 * 4 * 1024; + +} // namespace impl + +constexpr const std::size_t RequiredNonSecureSystemMemorySize = + impl::RequiredNonSecureSystemMemorySizeVi + impl::RequiredNonSecureSystemMemorySizeNvservices + + impl::RequiredNonSecureSystemMemorySizeMisc; + +namespace { + +u32 GetMemoryModeForInit() { + return 0x01; +} + +u32 GetMemorySizeForInit() { + return 0; +} + +Smc::MemoryArrangement GetMemoryArrangeForInit() { + switch (GetMemoryModeForInit() & 0x3F) { + case 0x01: + default: + return Smc::MemoryArrangement_4GB; + case 0x02: + return Smc::MemoryArrangement_4GBForAppletDev; + case 0x03: + return Smc::MemoryArrangement_4GBForSystemDev; + case 0x11: + return Smc::MemoryArrangement_6GB; + case 0x12: + return Smc::MemoryArrangement_6GBForAppletDev; + case 0x21: + return Smc::MemoryArrangement_8GB; + } +} +} // namespace + +// Initialization. +size_t KSystemControl::Init::GetIntendedMemorySize() { + switch (GetMemorySizeForInit()) { + case Smc::MemorySize_4GB: + default: // All invalid modes should go to 4GB. + return Common::Size_4_GB; + case Smc::MemorySize_6GB: + return Common::Size_6_GB; + case Smc::MemorySize_8GB: + return Common::Size_8_GB; + } +} + +PAddr KSystemControl::Init::GetKernelPhysicalBaseAddress(u64 base_address) { + return base_address; +} + +bool KSystemControl::Init::ShouldIncreaseThreadResourceLimit() { + return true; +} + +std::size_t KSystemControl::Init::GetApplicationPoolSize() { + // Get the base pool size. + const size_t base_pool_size = []() -> size_t { + switch (GetMemoryArrangeForInit()) { + case Smc::MemoryArrangement_4GB: + default: + return Common::Size_3285_MB; + case Smc::MemoryArrangement_4GBForAppletDev: + return Common::Size_2048_MB; + case Smc::MemoryArrangement_4GBForSystemDev: + return Common::Size_3285_MB; + case Smc::MemoryArrangement_6GB: + return Common::Size_4916_MB; + case Smc::MemoryArrangement_6GBForAppletDev: + return Common::Size_3285_MB; + case Smc::MemoryArrangement_8GB: + return Common::Size_4916_MB; + } + }(); + + // Return (possibly) adjusted size. + return base_pool_size; +} + +size_t KSystemControl::Init::GetAppletPoolSize() { + // Get the base pool size. + const size_t base_pool_size = []() -> size_t { + switch (GetMemoryArrangeForInit()) { + case Smc::MemoryArrangement_4GB: + default: + return Common::Size_507_MB; + case Smc::MemoryArrangement_4GBForAppletDev: + return Common::Size_1554_MB; + case Smc::MemoryArrangement_4GBForSystemDev: + return Common::Size_448_MB; + case Smc::MemoryArrangement_6GB: + return Common::Size_562_MB; + case Smc::MemoryArrangement_6GBForAppletDev: + return Common::Size_2193_MB; + case Smc::MemoryArrangement_8GB: + return Common::Size_2193_MB; + } + }(); + + // Return (possibly) adjusted size. + constexpr size_t ExtraSystemMemoryForAtmosphere = Common::Size_33_MB; + return base_pool_size - ExtraSystemMemoryForAtmosphere - KTraceBufferSize; +} + +size_t KSystemControl::Init::GetMinimumNonSecureSystemPoolSize() { + // Verify that our minimum is at least as large as Nintendo's. + constexpr size_t MinimumSize = RequiredNonSecureSystemMemorySize; + static_assert(MinimumSize >= 0x29C8000); + + return MinimumSize; +} + +namespace { +template <typename F> +u64 GenerateUniformRange(u64 min, u64 max, F f) { + // Handle the case where the difference is too large to represent. + if (max == std::numeric_limits<u64>::max() && min == std::numeric_limits<u64>::min()) { + return f(); + } + + // Iterate until we get a value in range. + const u64 range_size = ((max + 1) - min); + const u64 effective_max = (std::numeric_limits<u64>::max() / range_size) * range_size; + while (true) { + if (const u64 rnd = f(); rnd < effective_max) { + return min + (rnd % range_size); + } + } +} + +} // Anonymous namespace + +u64 KSystemControl::GenerateRandomU64() { + static std::random_device device; + static std::mt19937 gen(device()); + static std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max()); + return distribution(gen); +} + +u64 KSystemControl::GenerateRandomRange(u64 min, u64 max) { + return GenerateUniformRange(min, max, GenerateRandomU64); +} + +} // namespace Kernel::Board::Nintendo::Nx diff --git a/src/core/hle/kernel/board/nintendo/nx/k_system_control.h b/src/core/hle/kernel/board/nintendo/nx/k_system_control.h new file mode 100644 index 000000000..52f230ced --- /dev/null +++ b/src/core/hle/kernel/board/nintendo/nx/k_system_control.h @@ -0,0 +1,28 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" + +namespace Kernel::Board::Nintendo::Nx { + +class KSystemControl { +public: + class Init { + public: + // Initialization. + static std::size_t GetIntendedMemorySize(); + static PAddr GetKernelPhysicalBaseAddress(u64 base_address); + static bool ShouldIncreaseThreadResourceLimit(); + static std::size_t GetApplicationPoolSize(); + static std::size_t GetAppletPoolSize(); + static std::size_t GetMinimumNonSecureSystemPoolSize(); + }; + + static u64 GenerateRandomRange(u64 min, u64 max); + static u64 GenerateRandomU64(); +}; + +} // namespace Kernel::Board::Nintendo::Nx diff --git a/src/core/hle/kernel/board/nintendo/nx/secure_monitor.h b/src/core/hle/kernel/board/nintendo/nx/secure_monitor.h new file mode 100644 index 000000000..0c366b252 --- /dev/null +++ b/src/core/hle/kernel/board/nintendo/nx/secure_monitor.h @@ -0,0 +1,26 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" + +namespace Kernel::Board::Nintendo::Nx::Smc { + +enum MemorySize { + MemorySize_4GB = 0, + MemorySize_6GB = 1, + MemorySize_8GB = 2, +}; + +enum MemoryArrangement { + MemoryArrangement_4GB = 0, + MemoryArrangement_4GBForAppletDev = 1, + MemoryArrangement_4GBForSystemDev = 2, + MemoryArrangement_6GB = 3, + MemoryArrangement_6GBForAppletDev = 4, + MemoryArrangement_8GB = 5, +}; + +} // namespace Kernel::Board::Nintendo::Nx::Smc diff --git a/src/core/hle/kernel/k_address_space_info.cpp b/src/core/hle/kernel/k_address_space_info.cpp index 24944d15b..c7549f7a2 100644 --- a/src/core/hle/kernel/k_address_space_info.cpp +++ b/src/core/hle/kernel/k_address_space_info.cpp @@ -5,45 +5,34 @@ #include <array> #include "common/assert.h" +#include "common/common_sizes.h" #include "core/hle/kernel/k_address_space_info.h" namespace Kernel { namespace { -enum : u64 { - Size_1_MB = 0x100000, - Size_2_MB = 2 * Size_1_MB, - Size_128_MB = 128 * Size_1_MB, - Size_1_GB = 0x40000000, - Size_2_GB = 2 * Size_1_GB, - Size_4_GB = 4 * Size_1_GB, - Size_6_GB = 6 * Size_1_GB, - Size_64_GB = 64 * Size_1_GB, - Size_512_GB = 512 * Size_1_GB, - Invalid = std::numeric_limits<u64>::max(), -}; - // clang-format off constexpr std::array<KAddressSpaceInfo, 13> AddressSpaceInfos{{ - { .bit_width = 32, .address = Size_2_MB , .size = Size_1_GB - Size_2_MB , .type = KAddressSpaceInfo::Type::MapSmall, }, - { .bit_width = 32, .address = Size_1_GB , .size = Size_4_GB - Size_1_GB , .type = KAddressSpaceInfo::Type::MapLarge, }, - { .bit_width = 32, .address = Invalid , .size = Size_1_GB , .type = KAddressSpaceInfo::Type::Heap, }, - { .bit_width = 32, .address = Invalid , .size = Size_1_GB , .type = KAddressSpaceInfo::Type::Alias, }, - { .bit_width = 36, .address = Size_128_MB, .size = Size_2_GB - Size_128_MB, .type = KAddressSpaceInfo::Type::MapSmall, }, - { .bit_width = 36, .address = Size_2_GB , .size = Size_64_GB - Size_2_GB , .type = KAddressSpaceInfo::Type::MapLarge, }, - { .bit_width = 36, .address = Invalid , .size = Size_6_GB , .type = KAddressSpaceInfo::Type::Heap, }, - { .bit_width = 36, .address = Invalid , .size = Size_6_GB , .type = KAddressSpaceInfo::Type::Alias, }, - { .bit_width = 39, .address = Size_128_MB, .size = Size_512_GB - Size_128_MB, .type = KAddressSpaceInfo::Type::Map39Bit, }, - { .bit_width = 39, .address = Invalid , .size = Size_64_GB , .type = KAddressSpaceInfo::Type::MapSmall }, - { .bit_width = 39, .address = Invalid , .size = Size_6_GB , .type = KAddressSpaceInfo::Type::Heap, }, - { .bit_width = 39, .address = Invalid , .size = Size_64_GB , .type = KAddressSpaceInfo::Type::Alias, }, - { .bit_width = 39, .address = Invalid , .size = Size_2_GB , .type = KAddressSpaceInfo::Type::Stack, }, + { .bit_width = 32, .address = Common::Size_2_MB , .size = Common::Size_1_GB - Common::Size_2_MB , .type = KAddressSpaceInfo::Type::MapSmall, }, + { .bit_width = 32, .address = Common::Size_1_GB , .size = Common::Size_4_GB - Common::Size_1_GB , .type = KAddressSpaceInfo::Type::MapLarge, }, + { .bit_width = 32, .address = Common::Size_Invalid, .size = Common::Size_1_GB , .type = KAddressSpaceInfo::Type::Alias, }, + { .bit_width = 32, .address = Common::Size_Invalid, .size = Common::Size_1_GB , .type = KAddressSpaceInfo::Type::Heap, }, + { .bit_width = 36, .address = Common::Size_128_MB , .size = Common::Size_2_GB - Common::Size_128_MB, .type = KAddressSpaceInfo::Type::MapSmall, }, + { .bit_width = 36, .address = Common::Size_2_GB , .size = Common::Size_64_GB - Common::Size_2_GB , .type = KAddressSpaceInfo::Type::MapLarge, }, + { .bit_width = 36, .address = Common::Size_Invalid, .size = Common::Size_6_GB , .type = KAddressSpaceInfo::Type::Heap, }, + { .bit_width = 36, .address = Common::Size_Invalid, .size = Common::Size_6_GB , .type = KAddressSpaceInfo::Type::Alias, }, + { .bit_width = 39, .address = Common::Size_128_MB , .size = Common::Size_512_GB - Common::Size_128_MB, .type = KAddressSpaceInfo::Type::Map39Bit, }, + { .bit_width = 39, .address = Common::Size_Invalid, .size = Common::Size_64_GB , .type = KAddressSpaceInfo::Type::MapSmall }, + { .bit_width = 39, .address = Common::Size_Invalid, .size = Common::Size_6_GB , .type = KAddressSpaceInfo::Type::Heap, }, + { .bit_width = 39, .address = Common::Size_Invalid, .size = Common::Size_64_GB , .type = KAddressSpaceInfo::Type::Alias, }, + { .bit_width = 39, .address = Common::Size_Invalid, .size = Common::Size_2_GB , .type = KAddressSpaceInfo::Type::Stack, }, }}; // clang-format on constexpr bool IsAllowedIndexForAddress(std::size_t index) { - return index < AddressSpaceInfos.size() && AddressSpaceInfos[index].address != Invalid; + return index < AddressSpaceInfos.size() && + AddressSpaceInfos[index].address != Common::Size_Invalid; } using IndexArray = diff --git a/src/core/hle/kernel/k_memory_layout.board.nintendo_nx.cpp b/src/core/hle/kernel/k_memory_layout.board.nintendo_nx.cpp new file mode 100644 index 000000000..a78551291 --- /dev/null +++ b/src/core/hle/kernel/k_memory_layout.board.nintendo_nx.cpp @@ -0,0 +1,199 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/alignment.h" +#include "core/hle/kernel/k_memory_layout.h" +#include "core/hle/kernel/k_memory_manager.h" +#include "core/hle/kernel/k_system_control.h" +#include "core/hle/kernel/k_trace.h" + +namespace Kernel { + +namespace { + +constexpr size_t CarveoutAlignment = 0x20000; +constexpr size_t CarveoutSizeMax = (512ULL * 1024 * 1024) - CarveoutAlignment; + +bool SetupPowerManagementControllerMemoryRegion(KMemoryLayout& memory_layout) { + // Above firmware 2.0.0, the PMC is not mappable. + return memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x7000E000, 0x400, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap) && + memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x7000E400, 0xC00, + KMemoryRegionType_PowerManagementController | KMemoryRegionAttr_NoUserMap); +} + +void InsertPoolPartitionRegionIntoBothTrees(KMemoryLayout& memory_layout, size_t start, size_t size, + KMemoryRegionType phys_type, + KMemoryRegionType virt_type, u32& cur_attr) { + const u32 attr = cur_attr++; + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(start, size, + static_cast<u32>(phys_type), attr)); + const KMemoryRegion* phys = memory_layout.GetPhysicalMemoryRegionTree().FindByTypeAndAttribute( + static_cast<u32>(phys_type), attr); + ASSERT(phys != nullptr); + ASSERT(phys->GetEndAddress() != 0); + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert(phys->GetPairAddress(), size, + static_cast<u32>(virt_type), attr)); +} + +} // namespace + +namespace Init { + +void SetupDevicePhysicalMemoryRegions(KMemoryLayout& memory_layout) { + ASSERT(SetupPowerManagementControllerMemoryRegion(memory_layout)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x70019000, 0x1000, KMemoryRegionType_MemoryController | KMemoryRegionAttr_NoUserMap)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x7001C000, 0x1000, KMemoryRegionType_MemoryController0 | KMemoryRegionAttr_NoUserMap)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x7001D000, 0x1000, KMemoryRegionType_MemoryController1 | KMemoryRegionAttr_NoUserMap)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x50040000, 0x1000, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x50041000, 0x1000, + KMemoryRegionType_InterruptDistributor | KMemoryRegionAttr_ShouldKernelMap)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x50042000, 0x1000, + KMemoryRegionType_InterruptCpuInterface | KMemoryRegionAttr_ShouldKernelMap)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x50043000, 0x1D000, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap)); + + // Map IRAM unconditionally, to support debug-logging-to-iram build config. + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x40000000, 0x40000, KMemoryRegionType_LegacyLpsIram | KMemoryRegionAttr_ShouldKernelMap)); + + // Above firmware 2.0.0, prevent mapping the bpmp exception vectors or the ipatch region. + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x6000F000, 0x1000, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + 0x6001DC00, 0x400, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap)); +} + +void SetupDramPhysicalMemoryRegions(KMemoryLayout& memory_layout) { + const size_t intended_memory_size = KSystemControl::Init::GetIntendedMemorySize(); + const PAddr physical_memory_base_address = + KSystemControl::Init::GetKernelPhysicalBaseAddress(DramPhysicalAddress); + + // Insert blocks into the tree. + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + physical_memory_base_address, intended_memory_size, KMemoryRegionType_Dram)); + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + physical_memory_base_address, ReservedEarlyDramSize, KMemoryRegionType_DramReservedEarly)); + + // Insert the KTrace block at the end of Dram, if KTrace is enabled. + static_assert(!IsKTraceEnabled || KTraceBufferSize > 0); + if constexpr (IsKTraceEnabled) { + const PAddr ktrace_buffer_phys_addr = + physical_memory_base_address + intended_memory_size - KTraceBufferSize; + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + ktrace_buffer_phys_addr, KTraceBufferSize, KMemoryRegionType_KernelTraceBuffer)); + } +} + +void SetupPoolPartitionMemoryRegions(KMemoryLayout& memory_layout) { + // Start by identifying the extents of the DRAM memory region. + const auto dram_extents = memory_layout.GetMainMemoryPhysicalExtents(); + ASSERT(dram_extents.GetEndAddress() != 0); + + // Determine the end of the pool region. + const u64 pool_end = dram_extents.GetEndAddress() - KTraceBufferSize; + + // Find the start of the kernel DRAM region. + const KMemoryRegion* kernel_dram_region = + memory_layout.GetPhysicalMemoryRegionTree().FindFirstDerived( + KMemoryRegionType_DramKernelBase); + ASSERT(kernel_dram_region != nullptr); + + const u64 kernel_dram_start = kernel_dram_region->GetAddress(); + ASSERT(Common::IsAligned(kernel_dram_start, CarveoutAlignment)); + + // Find the start of the pool partitions region. + const KMemoryRegion* pool_partitions_region = + memory_layout.GetPhysicalMemoryRegionTree().FindByTypeAndAttribute( + KMemoryRegionType_DramPoolPartition, 0); + ASSERT(pool_partitions_region != nullptr); + const u64 pool_partitions_start = pool_partitions_region->GetAddress(); + + // Setup the pool partition layouts. + // On 5.0.0+, setup modern 4-pool-partition layout. + + // Get Application and Applet pool sizes. + const size_t application_pool_size = KSystemControl::Init::GetApplicationPoolSize(); + const size_t applet_pool_size = KSystemControl::Init::GetAppletPoolSize(); + const size_t unsafe_system_pool_min_size = + KSystemControl::Init::GetMinimumNonSecureSystemPoolSize(); + + // Decide on starting addresses for our pools. + const u64 application_pool_start = pool_end - application_pool_size; + const u64 applet_pool_start = application_pool_start - applet_pool_size; + const u64 unsafe_system_pool_start = std::min( + kernel_dram_start + CarveoutSizeMax, + Common::AlignDown(applet_pool_start - unsafe_system_pool_min_size, CarveoutAlignment)); + const size_t unsafe_system_pool_size = applet_pool_start - unsafe_system_pool_start; + + // We want to arrange application pool depending on where the middle of dram is. + const u64 dram_midpoint = (dram_extents.GetAddress() + dram_extents.GetEndAddress()) / 2; + u32 cur_pool_attr = 0; + size_t total_overhead_size = 0; + if (dram_extents.GetEndAddress() <= dram_midpoint || dram_midpoint <= application_pool_start) { + InsertPoolPartitionRegionIntoBothTrees( + memory_layout, application_pool_start, application_pool_size, + KMemoryRegionType_DramApplicationPool, KMemoryRegionType_VirtualDramApplicationPool, + cur_pool_attr); + total_overhead_size += + KMemoryManager::CalculateManagementOverheadSize(application_pool_size); + } else { + const size_t first_application_pool_size = dram_midpoint - application_pool_start; + const size_t second_application_pool_size = + application_pool_start + application_pool_size - dram_midpoint; + InsertPoolPartitionRegionIntoBothTrees( + memory_layout, application_pool_start, first_application_pool_size, + KMemoryRegionType_DramApplicationPool, KMemoryRegionType_VirtualDramApplicationPool, + cur_pool_attr); + InsertPoolPartitionRegionIntoBothTrees( + memory_layout, dram_midpoint, second_application_pool_size, + KMemoryRegionType_DramApplicationPool, KMemoryRegionType_VirtualDramApplicationPool, + cur_pool_attr); + total_overhead_size += + KMemoryManager::CalculateManagementOverheadSize(first_application_pool_size); + total_overhead_size += + KMemoryManager::CalculateManagementOverheadSize(second_application_pool_size); + } + + // Insert the applet pool. + InsertPoolPartitionRegionIntoBothTrees(memory_layout, applet_pool_start, applet_pool_size, + KMemoryRegionType_DramAppletPool, + KMemoryRegionType_VirtualDramAppletPool, cur_pool_attr); + total_overhead_size += KMemoryManager::CalculateManagementOverheadSize(applet_pool_size); + + // Insert the nonsecure system pool. + InsertPoolPartitionRegionIntoBothTrees( + memory_layout, unsafe_system_pool_start, unsafe_system_pool_size, + KMemoryRegionType_DramSystemNonSecurePool, KMemoryRegionType_VirtualDramSystemNonSecurePool, + cur_pool_attr); + total_overhead_size += KMemoryManager::CalculateManagementOverheadSize(unsafe_system_pool_size); + + // Insert the pool management region. + total_overhead_size += KMemoryManager::CalculateManagementOverheadSize( + (unsafe_system_pool_start - pool_partitions_start) - total_overhead_size); + const u64 pool_management_start = unsafe_system_pool_start - total_overhead_size; + const size_t pool_management_size = total_overhead_size; + u32 pool_management_attr = 0; + InsertPoolPartitionRegionIntoBothTrees( + memory_layout, pool_management_start, pool_management_size, + KMemoryRegionType_DramPoolManagement, KMemoryRegionType_VirtualDramPoolManagement, + pool_management_attr); + + // Insert the system pool. + const u64 system_pool_size = pool_management_start - pool_partitions_start; + InsertPoolPartitionRegionIntoBothTrees(memory_layout, pool_partitions_start, system_pool_size, + KMemoryRegionType_DramSystemPool, + KMemoryRegionType_VirtualDramSystemPool, cur_pool_attr); +} + +} // namespace Init + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_memory_layout.cpp b/src/core/hle/kernel/k_memory_layout.cpp new file mode 100644 index 000000000..fb1e2435f --- /dev/null +++ b/src/core/hle/kernel/k_memory_layout.cpp @@ -0,0 +1,166 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <array> + +#include "common/alignment.h" +#include "core/hle/kernel/k_memory_layout.h" +#include "core/hle/kernel/k_system_control.h" + +namespace Kernel { + +namespace { + +template <typename... Args> +KMemoryRegion* AllocateRegion(KMemoryRegionAllocator& memory_region_allocator, Args&&... args) { + return memory_region_allocator.Allocate(std::forward<Args>(args)...); +} + +} // namespace + +KMemoryRegionTree::KMemoryRegionTree(KMemoryRegionAllocator& memory_region_allocator_) + : memory_region_allocator{memory_region_allocator_} {} + +void KMemoryRegionTree::InsertDirectly(u64 address, u64 last_address, u32 attr, u32 type_id) { + this->insert(*AllocateRegion(memory_region_allocator, address, last_address, attr, type_id)); +} + +bool KMemoryRegionTree::Insert(u64 address, size_t size, u32 type_id, u32 new_attr, u32 old_attr) { + // Locate the memory region that contains the address. + KMemoryRegion* found = this->FindModifiable(address); + + // We require that the old attr is correct. + if (found->GetAttributes() != old_attr) { + return false; + } + + // We further require that the region can be split from the old region. + const u64 inserted_region_end = address + size; + const u64 inserted_region_last = inserted_region_end - 1; + if (found->GetLastAddress() < inserted_region_last) { + return false; + } + + // Further, we require that the type id is a valid transformation. + if (!found->CanDerive(type_id)) { + return false; + } + + // Cache information from the region before we remove it. + const u64 old_address = found->GetAddress(); + const u64 old_last = found->GetLastAddress(); + const u64 old_pair = found->GetPairAddress(); + const u32 old_type = found->GetType(); + + // Erase the existing region from the tree. + this->erase(this->iterator_to(*found)); + + // Insert the new region into the tree. + if (old_address == address) { + // Reuse the old object for the new region, if we can. + found->Reset(address, inserted_region_last, old_pair, new_attr, type_id); + this->insert(*found); + } else { + // If we can't re-use, adjust the old region. + found->Reset(old_address, address - 1, old_pair, old_attr, old_type); + this->insert(*found); + + // Insert a new region for the split. + const u64 new_pair = (old_pair != std::numeric_limits<u64>::max()) + ? old_pair + (address - old_address) + : old_pair; + this->insert(*AllocateRegion(memory_region_allocator, address, inserted_region_last, + new_pair, new_attr, type_id)); + } + + // If we need to insert a region after the region, do so. + if (old_last != inserted_region_last) { + const u64 after_pair = (old_pair != std::numeric_limits<u64>::max()) + ? old_pair + (inserted_region_end - old_address) + : old_pair; + this->insert(*AllocateRegion(memory_region_allocator, inserted_region_end, old_last, + after_pair, old_attr, old_type)); + } + + return true; +} + +VAddr KMemoryRegionTree::GetRandomAlignedRegion(size_t size, size_t alignment, u32 type_id) { + // We want to find the total extents of the type id. + const auto extents = this->GetDerivedRegionExtents(static_cast<KMemoryRegionType>(type_id)); + + // Ensure that our alignment is correct. + ASSERT(Common::IsAligned(extents.GetAddress(), alignment)); + + const u64 first_address = extents.GetAddress(); + const u64 last_address = extents.GetLastAddress(); + + const u64 first_index = first_address / alignment; + const u64 last_index = last_address / alignment; + + while (true) { + const u64 candidate = + KSystemControl::GenerateRandomRange(first_index, last_index) * alignment; + + // Ensure that the candidate doesn't overflow with the size. + if (!(candidate < candidate + size)) { + continue; + } + + const u64 candidate_last = candidate + size - 1; + + // Ensure that the candidate fits within the region. + if (candidate_last > last_address) { + continue; + } + + // Locate the candidate region, and ensure it fits and has the correct type id. + if (const auto& candidate_region = *this->Find(candidate); + !(candidate_last <= candidate_region.GetLastAddress() && + candidate_region.GetType() == type_id)) { + continue; + } + + return candidate; + } +} + +KMemoryLayout::KMemoryLayout() + : virtual_tree{memory_region_allocator}, physical_tree{memory_region_allocator}, + virtual_linear_tree{memory_region_allocator}, physical_linear_tree{memory_region_allocator} {} + +void KMemoryLayout::InitializeLinearMemoryRegionTrees(PAddr aligned_linear_phys_start, + VAddr linear_virtual_start) { + // Set static differences. + linear_phys_to_virt_diff = linear_virtual_start - aligned_linear_phys_start; + linear_virt_to_phys_diff = aligned_linear_phys_start - linear_virtual_start; + + // Initialize linear trees. + for (auto& region : GetPhysicalMemoryRegionTree()) { + if (region.HasTypeAttribute(KMemoryRegionAttr_LinearMapped)) { + GetPhysicalLinearMemoryRegionTree().InsertDirectly( + region.GetAddress(), region.GetLastAddress(), region.GetAttributes(), + region.GetType()); + } + } + + for (auto& region : GetVirtualMemoryRegionTree()) { + if (region.IsDerivedFrom(KMemoryRegionType_Dram)) { + GetVirtualLinearMemoryRegionTree().InsertDirectly( + region.GetAddress(), region.GetLastAddress(), region.GetAttributes(), + region.GetType()); + } + } +} + +size_t KMemoryLayout::GetResourceRegionSizeForInit() { + // Calculate resource region size based on whether we allow extra threads. + const bool use_extra_resources = KSystemControl::Init::ShouldIncreaseThreadResourceLimit(); + size_t resource_region_size = + KernelResourceSize + (use_extra_resources ? KernelSlabHeapAdditionalSize : 0); + + return resource_region_size; +} + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_memory_layout.h b/src/core/hle/kernel/k_memory_layout.h index 0821d2d8c..288642d9a 100644 --- a/src/core/hle/kernel/k_memory_layout.h +++ b/src/core/hle/kernel/k_memory_layout.h @@ -1,23 +1,69 @@ -// Copyright 2020 yuzu Emulator Project +// Copyright 2021 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. #pragma once +#include <utility> + +#include "common/alignment.h" +#include "common/common_sizes.h" #include "common/common_types.h" #include "core/device_memory.h" +#include "core/hle/kernel/k_memory_region.h" +#include "core/hle/kernel/k_memory_region_type.h" +#include "core/hle/kernel/memory_types.h" namespace Kernel { -constexpr std::size_t KernelAslrAlignment = 2 * 1024 * 1024; +constexpr std::size_t L1BlockSize = Common::Size_1_GB; +constexpr std::size_t L2BlockSize = Common::Size_2_MB; + +constexpr std::size_t GetMaximumOverheadSize(std::size_t size) { + return (Common::DivideUp(size, L1BlockSize) + Common::DivideUp(size, L2BlockSize)) * PageSize; +} + +constexpr std::size_t MainMemorySize = Common::Size_4_GB; +constexpr std::size_t MainMemorySizeMax = Common::Size_8_GB; + +constexpr std::size_t ReservedEarlyDramSize = 0x60000; +constexpr std::size_t DramPhysicalAddress = 0x80000000; + +constexpr std::size_t KernelAslrAlignment = Common::Size_2_MB; constexpr std::size_t KernelVirtualAddressSpaceWidth = 1ULL << 39; constexpr std::size_t KernelPhysicalAddressSpaceWidth = 1ULL << 48; + constexpr std::size_t KernelVirtualAddressSpaceBase = 0ULL - KernelVirtualAddressSpaceWidth; constexpr std::size_t KernelVirtualAddressSpaceEnd = KernelVirtualAddressSpaceBase + (KernelVirtualAddressSpaceWidth - KernelAslrAlignment); -constexpr std::size_t KernelVirtualAddressSpaceLast = KernelVirtualAddressSpaceEnd - 1; +constexpr std::size_t KernelVirtualAddressSpaceLast = KernelVirtualAddressSpaceEnd - 1ULL; constexpr std::size_t KernelVirtualAddressSpaceSize = KernelVirtualAddressSpaceEnd - KernelVirtualAddressSpaceBase; +constexpr std::size_t KernelVirtualAddressCodeBase = KernelVirtualAddressSpaceBase; +constexpr std::size_t KernelVirtualAddressCodeSize = 0x62000; +constexpr std::size_t KernelVirtualAddressCodeEnd = + KernelVirtualAddressCodeBase + KernelVirtualAddressCodeSize; + +constexpr std::size_t KernelPhysicalAddressSpaceBase = 0ULL; +constexpr std::size_t KernelPhysicalAddressSpaceEnd = + KernelPhysicalAddressSpaceBase + KernelPhysicalAddressSpaceWidth; +constexpr std::size_t KernelPhysicalAddressSpaceLast = KernelPhysicalAddressSpaceEnd - 1ULL; +constexpr std::size_t KernelPhysicalAddressSpaceSize = + KernelPhysicalAddressSpaceEnd - KernelPhysicalAddressSpaceBase; +constexpr std::size_t KernelPhysicalAddressCodeBase = DramPhysicalAddress + ReservedEarlyDramSize; + +constexpr std::size_t KernelPageTableHeapSize = GetMaximumOverheadSize(MainMemorySizeMax); +constexpr std::size_t KernelInitialPageHeapSize = Common::Size_128_KB; + +constexpr std::size_t KernelSlabHeapDataSize = Common::Size_5_MB; +constexpr std::size_t KernelSlabHeapGapsSize = Common::Size_2_MB - Common::Size_64_KB; +constexpr std::size_t KernelSlabHeapSize = KernelSlabHeapDataSize + KernelSlabHeapGapsSize; + +// NOTE: This is calculated from KThread slab counts, assuming KThread size <= 0x860. +constexpr std::size_t KernelSlabHeapAdditionalSize = 0x68000ULL; + +constexpr std::size_t KernelResourceSize = + KernelPageTableHeapSize + KernelInitialPageHeapSize + KernelSlabHeapSize; constexpr bool IsKernelAddressKey(VAddr key) { return KernelVirtualAddressSpaceBase <= key && key <= KernelVirtualAddressSpaceLast; @@ -27,64 +73,327 @@ constexpr bool IsKernelAddress(VAddr address) { return KernelVirtualAddressSpaceBase <= address && address < KernelVirtualAddressSpaceEnd; } -class KMemoryRegion final { - friend class KMemoryLayout; - +class KMemoryLayout final { public: - constexpr PAddr StartAddress() const { - return start_address; + KMemoryLayout(); + + KMemoryRegionTree& GetVirtualMemoryRegionTree() { + return virtual_tree; + } + const KMemoryRegionTree& GetVirtualMemoryRegionTree() const { + return virtual_tree; + } + KMemoryRegionTree& GetPhysicalMemoryRegionTree() { + return physical_tree; + } + const KMemoryRegionTree& GetPhysicalMemoryRegionTree() const { + return physical_tree; + } + KMemoryRegionTree& GetVirtualLinearMemoryRegionTree() { + return virtual_linear_tree; + } + const KMemoryRegionTree& GetVirtualLinearMemoryRegionTree() const { + return virtual_linear_tree; + } + KMemoryRegionTree& GetPhysicalLinearMemoryRegionTree() { + return physical_linear_tree; + } + const KMemoryRegionTree& GetPhysicalLinearMemoryRegionTree() const { + return physical_linear_tree; + } + + VAddr GetLinearVirtualAddress(PAddr address) const { + return address + linear_phys_to_virt_diff; + } + PAddr GetLinearPhysicalAddress(VAddr address) const { + return address + linear_virt_to_phys_diff; + } + + const KMemoryRegion* FindVirtual(VAddr address) const { + return Find(address, GetVirtualMemoryRegionTree()); + } + const KMemoryRegion* FindPhysical(PAddr address) const { + return Find(address, GetPhysicalMemoryRegionTree()); + } + + const KMemoryRegion* FindVirtualLinear(VAddr address) const { + return Find(address, GetVirtualLinearMemoryRegionTree()); + } + const KMemoryRegion* FindPhysicalLinear(PAddr address) const { + return Find(address, GetPhysicalLinearMemoryRegionTree()); + } + + VAddr GetMainStackTopAddress(s32 core_id) const { + return GetStackTopAddress(core_id, KMemoryRegionType_KernelMiscMainStack); + } + VAddr GetIdleStackTopAddress(s32 core_id) const { + return GetStackTopAddress(core_id, KMemoryRegionType_KernelMiscIdleStack); + } + VAddr GetExceptionStackTopAddress(s32 core_id) const { + return GetStackTopAddress(core_id, KMemoryRegionType_KernelMiscExceptionStack); + } + + VAddr GetSlabRegionAddress() const { + return Dereference(GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_KernelSlab)) + .GetAddress(); + } + + const KMemoryRegion& GetDeviceRegion(KMemoryRegionType type) const { + return Dereference(GetPhysicalMemoryRegionTree().FindFirstDerived(type)); + } + PAddr GetDevicePhysicalAddress(KMemoryRegionType type) const { + return GetDeviceRegion(type).GetAddress(); + } + VAddr GetDeviceVirtualAddress(KMemoryRegionType type) const { + return GetDeviceRegion(type).GetPairAddress(); + } + + const KMemoryRegion& GetPoolManagementRegion() const { + return Dereference( + GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_VirtualDramPoolManagement)); + } + const KMemoryRegion& GetPageTableHeapRegion() const { + return Dereference( + GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_VirtualDramKernelPtHeap)); + } + const KMemoryRegion& GetKernelStackRegion() const { + return Dereference(GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_KernelStack)); + } + const KMemoryRegion& GetTempRegion() const { + return Dereference(GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_KernelTemp)); + } + + const KMemoryRegion& GetKernelTraceBufferRegion() const { + return Dereference(GetVirtualLinearMemoryRegionTree().FindByType( + KMemoryRegionType_VirtualDramKernelTraceBuffer)); + } + + const KMemoryRegion& GetVirtualLinearRegion(VAddr address) const { + return Dereference(FindVirtualLinear(address)); + } + + const KMemoryRegion* GetPhysicalKernelTraceBufferRegion() const { + return GetPhysicalMemoryRegionTree().FindFirstDerived(KMemoryRegionType_KernelTraceBuffer); + } + const KMemoryRegion* GetPhysicalOnMemoryBootImageRegion() const { + return GetPhysicalMemoryRegionTree().FindFirstDerived(KMemoryRegionType_OnMemoryBootImage); + } + const KMemoryRegion* GetPhysicalDTBRegion() const { + return GetPhysicalMemoryRegionTree().FindFirstDerived(KMemoryRegionType_DTB); + } + + bool IsHeapPhysicalAddress(const KMemoryRegion*& region, PAddr address) const { + return IsTypedAddress(region, address, GetPhysicalLinearMemoryRegionTree(), + KMemoryRegionType_DramUserPool); + } + bool IsHeapVirtualAddress(const KMemoryRegion*& region, VAddr address) const { + return IsTypedAddress(region, address, GetVirtualLinearMemoryRegionTree(), + KMemoryRegionType_VirtualDramUserPool); + } + + bool IsHeapPhysicalAddress(const KMemoryRegion*& region, PAddr address, size_t size) const { + return IsTypedAddress(region, address, size, GetPhysicalLinearMemoryRegionTree(), + KMemoryRegionType_DramUserPool); + } + bool IsHeapVirtualAddress(const KMemoryRegion*& region, VAddr address, size_t size) const { + return IsTypedAddress(region, address, size, GetVirtualLinearMemoryRegionTree(), + KMemoryRegionType_VirtualDramUserPool); + } + + bool IsLinearMappedPhysicalAddress(const KMemoryRegion*& region, PAddr address) const { + return IsTypedAddress(region, address, GetPhysicalLinearMemoryRegionTree(), + static_cast<KMemoryRegionType>(KMemoryRegionAttr_LinearMapped)); + } + bool IsLinearMappedPhysicalAddress(const KMemoryRegion*& region, PAddr address, + size_t size) const { + return IsTypedAddress(region, address, size, GetPhysicalLinearMemoryRegionTree(), + static_cast<KMemoryRegionType>(KMemoryRegionAttr_LinearMapped)); + } + + std::pair<size_t, size_t> GetTotalAndKernelMemorySizes() const { + size_t total_size = 0, kernel_size = 0; + for (const auto& region : GetPhysicalMemoryRegionTree()) { + if (region.IsDerivedFrom(KMemoryRegionType_Dram)) { + total_size += region.GetSize(); + if (!region.IsDerivedFrom(KMemoryRegionType_DramUserPool)) { + kernel_size += region.GetSize(); + } + } + } + return std::make_pair(total_size, kernel_size); + } + + void InitializeLinearMemoryRegionTrees(PAddr aligned_linear_phys_start, + VAddr linear_virtual_start); + static size_t GetResourceRegionSizeForInit(); + + auto GetKernelRegionExtents() const { + return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_Kernel); + } + auto GetKernelCodeRegionExtents() const { + return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_KernelCode); + } + auto GetKernelStackRegionExtents() const { + return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_KernelStack); + } + auto GetKernelMiscRegionExtents() const { + return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_KernelMisc); + } + auto GetKernelSlabRegionExtents() const { + return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_KernelSlab); + } + + auto GetLinearRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionAttr_LinearMapped); + } + + auto GetLinearRegionVirtualExtents() const { + const auto physical = GetLinearRegionPhysicalExtents(); + return KMemoryRegion(GetLinearVirtualAddress(physical.GetAddress()), + GetLinearVirtualAddress(physical.GetLastAddress()), 0, + KMemoryRegionType_None); + } + + auto GetMainMemoryPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_Dram); + } + auto GetCarveoutRegionExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionAttr_CarveoutProtected); + } + + auto GetKernelRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramKernelBase); + } + auto GetKernelCodeRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramKernelCode); + } + auto GetKernelSlabRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramKernelSlab); + } + auto GetKernelPageTableHeapRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramKernelPtHeap); + } + auto GetKernelInitPageTableRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramKernelInitPt); + } + + auto GetKernelPoolManagementRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramPoolManagement); + } + auto GetKernelPoolPartitionRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramPoolPartition); + } + auto GetKernelSystemPoolRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramSystemPool); + } + auto GetKernelSystemNonSecurePoolRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramSystemNonSecurePool); + } + auto GetKernelAppletPoolRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramAppletPool); + } + auto GetKernelApplicationPoolRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_DramApplicationPool); } - constexpr PAddr EndAddress() const { - return end_address; + auto GetKernelTraceBufferRegionPhysicalExtents() const { + return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionType_KernelTraceBuffer); } private: - constexpr KMemoryRegion() = default; - constexpr KMemoryRegion(PAddr start_address, PAddr end_address) - : start_address{start_address}, end_address{end_address} {} + template <typename AddressType> + static bool IsTypedAddress(const KMemoryRegion*& region, AddressType address, + const KMemoryRegionTree& tree, KMemoryRegionType type) { + // Check if the cached region already contains the address. + if (region != nullptr && region->Contains(address)) { + return true; + } - const PAddr start_address{}; - const PAddr end_address{}; -}; + // Find the containing region, and update the cache. + if (const KMemoryRegion* found = tree.Find(address); + found != nullptr && found->IsDerivedFrom(type)) { + region = found; + return true; + } else { + return false; + } + } -class KMemoryLayout final { -public: - constexpr const KMemoryRegion& Application() const { - return application; + template <typename AddressType> + static bool IsTypedAddress(const KMemoryRegion*& region, AddressType address, size_t size, + const KMemoryRegionTree& tree, KMemoryRegionType type) { + // Get the end of the checked region. + const u64 last_address = address + size - 1; + + // Walk the tree to verify the region is correct. + const KMemoryRegion* cur = + (region != nullptr && region->Contains(address)) ? region : tree.Find(address); + while (cur != nullptr && cur->IsDerivedFrom(type)) { + if (last_address <= cur->GetLastAddress()) { + region = cur; + return true; + } + + cur = cur->GetNext(); + } + return false; } - constexpr const KMemoryRegion& Applet() const { - return applet; + template <typename AddressType> + static const KMemoryRegion* Find(AddressType address, const KMemoryRegionTree& tree) { + return tree.Find(address); } - constexpr const KMemoryRegion& System() const { - return system; + static KMemoryRegion& Dereference(KMemoryRegion* region) { + ASSERT(region != nullptr); + return *region; } - static constexpr KMemoryLayout GetDefaultLayout() { - constexpr std::size_t application_size{0xcd500000}; - constexpr std::size_t applet_size{0x1fb00000}; - constexpr PAddr application_start_address{Core::DramMemoryMap::End - application_size}; - constexpr PAddr application_end_address{Core::DramMemoryMap::End}; - constexpr PAddr applet_start_address{application_start_address - applet_size}; - constexpr PAddr applet_end_address{applet_start_address + applet_size}; - constexpr PAddr system_start_address{Core::DramMemoryMap::SlabHeapEnd}; - constexpr PAddr system_end_address{applet_start_address}; - return {application_start_address, application_end_address, applet_start_address, - applet_end_address, system_start_address, system_end_address}; + static const KMemoryRegion& Dereference(const KMemoryRegion* region) { + ASSERT(region != nullptr); + return *region; + } + + VAddr GetStackTopAddress(s32 core_id, KMemoryRegionType type) const { + const auto& region = Dereference( + GetVirtualMemoryRegionTree().FindByTypeAndAttribute(type, static_cast<u32>(core_id))); + ASSERT(region.GetEndAddress() != 0); + return region.GetEndAddress(); } private: - constexpr KMemoryLayout(PAddr application_start_address, std::size_t application_size, - PAddr applet_start_address, std::size_t applet_size, - PAddr system_start_address, std::size_t system_size) - : application{application_start_address, application_size}, - applet{applet_start_address, applet_size}, system{system_start_address, system_size} {} - - const KMemoryRegion application; - const KMemoryRegion applet; - const KMemoryRegion system; + u64 linear_phys_to_virt_diff{}; + u64 linear_virt_to_phys_diff{}; + KMemoryRegionAllocator memory_region_allocator; + KMemoryRegionTree virtual_tree; + KMemoryRegionTree physical_tree; + KMemoryRegionTree virtual_linear_tree; + KMemoryRegionTree physical_linear_tree; }; +namespace Init { + +// These should be generic, regardless of board. +void SetupPoolPartitionMemoryRegions(KMemoryLayout& memory_layout); + +// These may be implemented in a board-specific manner. +void SetupDevicePhysicalMemoryRegions(KMemoryLayout& memory_layout); +void SetupDramPhysicalMemoryRegions(KMemoryLayout& memory_layout); + +} // namespace Init + } // namespace Kernel diff --git a/src/core/hle/kernel/k_memory_manager.cpp b/src/core/hle/kernel/k_memory_manager.cpp index 9027602bf..aa71697b2 100644 --- a/src/core/hle/kernel/k_memory_manager.cpp +++ b/src/core/hle/kernel/k_memory_manager.cpp @@ -173,4 +173,16 @@ ResultCode KMemoryManager::Free(KPageLinkedList& page_list, std::size_t num_page return RESULT_SUCCESS; } +std::size_t KMemoryManager::Impl::CalculateManagementOverheadSize(std::size_t region_size) { + const std::size_t ref_count_size = (region_size / PageSize) * sizeof(u16); + const std::size_t optimize_map_size = + (Common::AlignUp((region_size / PageSize), Common::BitSize<u64>()) / + Common::BitSize<u64>()) * + sizeof(u64); + const std::size_t manager_meta_size = + Common::AlignUp(optimize_map_size + ref_count_size, PageSize); + const std::size_t page_heap_size = KPageHeap::CalculateManagementOverheadSize(region_size); + return manager_meta_size + page_heap_size; +} + } // namespace Kernel diff --git a/src/core/hle/kernel/k_memory_manager.h b/src/core/hle/kernel/k_memory_manager.h index ae9f683b8..ac840b3d0 100644 --- a/src/core/hle/kernel/k_memory_manager.h +++ b/src/core/hle/kernel/k_memory_manager.h @@ -29,6 +29,10 @@ public: Shift = 4, Mask = (0xF << Shift), + + // Aliases. + Unsafe = Application, + Secure = System, }; enum class Direction : u32 { @@ -56,6 +60,10 @@ public: static constexpr std::size_t MaxManagerCount = 10; public: + static std::size_t CalculateManagementOverheadSize(std::size_t region_size) { + return Impl::CalculateManagementOverheadSize(region_size); + } + static constexpr u32 EncodeOption(Pool pool, Direction dir) { return (static_cast<u32>(pool) << static_cast<u32>(Pool::Shift)) | (static_cast<u32>(dir) << static_cast<u32>(Direction::Shift)); @@ -86,6 +94,16 @@ private: Pool pool{}; public: + static std::size_t CalculateManagementOverheadSize(std::size_t region_size); + + static constexpr std::size_t CalculateOptimizedProcessOverheadSize( + std::size_t region_size) { + return (Common::AlignUp((region_size / PageSize), Common::BitSize<u64>()) / + Common::BitSize<u64>()) * + sizeof(u64); + } + + public: Impl() = default; std::size_t Initialize(Pool new_pool, u64 start_address, u64 end_address); diff --git a/src/core/hle/kernel/k_memory_region.h b/src/core/hle/kernel/k_memory_region.h new file mode 100644 index 000000000..a861c04ab --- /dev/null +++ b/src/core/hle/kernel/k_memory_region.h @@ -0,0 +1,350 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/assert.h" +#include "common/common_types.h" +#include "common/intrusive_red_black_tree.h" +#include "core/hle/kernel/k_memory_region_type.h" + +namespace Kernel { + +class KMemoryRegionAllocator; + +class KMemoryRegion final : public Common::IntrusiveRedBlackTreeBaseNode<KMemoryRegion>, + NonCopyable { + friend class KMemoryRegionTree; + +public: + constexpr KMemoryRegion() = default; + constexpr KMemoryRegion(u64 address_, u64 last_address_) + : address{address_}, last_address{last_address_} {} + constexpr KMemoryRegion(u64 address_, u64 last_address_, u64 pair_address_, u32 attributes_, + u32 type_id_) + : address(address_), last_address(last_address_), pair_address(pair_address_), + attributes(attributes_), type_id(type_id_) {} + constexpr KMemoryRegion(u64 address_, u64 last_address_, u32 attributes_, u32 type_id_) + : KMemoryRegion(address_, last_address_, std::numeric_limits<u64>::max(), attributes_, + type_id_) {} + + static constexpr int Compare(const KMemoryRegion& lhs, const KMemoryRegion& rhs) { + if (lhs.GetAddress() < rhs.GetAddress()) { + return -1; + } else if (lhs.GetAddress() <= rhs.GetLastAddress()) { + return 0; + } else { + return 1; + } + } + +private: + constexpr void Reset(u64 a, u64 la, u64 p, u32 r, u32 t) { + address = a; + pair_address = p; + last_address = la; + attributes = r; + type_id = t; + } + +public: + constexpr u64 GetAddress() const { + return address; + } + + constexpr u64 GetPairAddress() const { + return pair_address; + } + + constexpr u64 GetLastAddress() const { + return last_address; + } + + constexpr u64 GetEndAddress() const { + return this->GetLastAddress() + 1; + } + + constexpr size_t GetSize() const { + return this->GetEndAddress() - this->GetAddress(); + } + + constexpr u32 GetAttributes() const { + return attributes; + } + + constexpr u32 GetType() const { + return type_id; + } + + constexpr void SetType(u32 type) { + ASSERT(this->CanDerive(type)); + type_id = type; + } + + constexpr bool Contains(u64 address) const { + ASSERT(this->GetEndAddress() != 0); + return this->GetAddress() <= address && address <= this->GetLastAddress(); + } + + constexpr bool IsDerivedFrom(u32 type) const { + return (this->GetType() | type) == this->GetType(); + } + + constexpr bool HasTypeAttribute(u32 attr) const { + return (this->GetType() | attr) == this->GetType(); + } + + constexpr bool CanDerive(u32 type) const { + return (this->GetType() | type) == type; + } + + constexpr void SetPairAddress(u64 a) { + pair_address = a; + } + + constexpr void SetTypeAttribute(u32 attr) { + type_id |= attr; + } + +private: + u64 address{}; + u64 last_address{}; + u64 pair_address{}; + u32 attributes{}; + u32 type_id{}; +}; + +class KMemoryRegionTree final : NonCopyable { +public: + struct DerivedRegionExtents { + const KMemoryRegion* first_region{}; + const KMemoryRegion* last_region{}; + + constexpr DerivedRegionExtents() = default; + + constexpr u64 GetAddress() const { + return this->first_region->GetAddress(); + } + + constexpr u64 GetLastAddress() const { + return this->last_region->GetLastAddress(); + } + + constexpr u64 GetEndAddress() const { + return this->GetLastAddress() + 1; + } + + constexpr size_t GetSize() const { + return this->GetEndAddress() - this->GetAddress(); + } + }; + +private: + using TreeType = + Common::IntrusiveRedBlackTreeBaseTraits<KMemoryRegion>::TreeType<KMemoryRegion>; + +public: + using value_type = TreeType::value_type; + using size_type = TreeType::size_type; + using difference_type = TreeType::difference_type; + using pointer = TreeType::pointer; + using const_pointer = TreeType::const_pointer; + using reference = TreeType::reference; + using const_reference = TreeType::const_reference; + using iterator = TreeType::iterator; + using const_iterator = TreeType::const_iterator; + +private: + TreeType m_tree{}; + KMemoryRegionAllocator& memory_region_allocator; + +public: + explicit KMemoryRegionTree(KMemoryRegionAllocator& memory_region_allocator_); + +public: + KMemoryRegion* FindModifiable(u64 address) { + if (auto it = this->find(KMemoryRegion(address, address, 0, 0)); it != this->end()) { + return std::addressof(*it); + } else { + return nullptr; + } + } + + const KMemoryRegion* Find(u64 address) const { + if (auto it = this->find(KMemoryRegion(address, address, 0, 0)); it != this->cend()) { + return std::addressof(*it); + } else { + return nullptr; + } + } + + const KMemoryRegion* FindByType(KMemoryRegionType type_id) const { + for (auto it = this->cbegin(); it != this->cend(); ++it) { + if (it->GetType() == static_cast<u32>(type_id)) { + return std::addressof(*it); + } + } + return nullptr; + } + + const KMemoryRegion* FindByTypeAndAttribute(u32 type_id, u32 attr) const { + for (auto it = this->cbegin(); it != this->cend(); ++it) { + if (it->GetType() == type_id && it->GetAttributes() == attr) { + return std::addressof(*it); + } + } + return nullptr; + } + + const KMemoryRegion* FindFirstDerived(KMemoryRegionType type_id) const { + for (auto it = this->cbegin(); it != this->cend(); it++) { + if (it->IsDerivedFrom(type_id)) { + return std::addressof(*it); + } + } + return nullptr; + } + + const KMemoryRegion* FindLastDerived(KMemoryRegionType type_id) const { + const KMemoryRegion* region = nullptr; + for (auto it = this->begin(); it != this->end(); it++) { + if (it->IsDerivedFrom(type_id)) { + region = std::addressof(*it); + } + } + return region; + } + + DerivedRegionExtents GetDerivedRegionExtents(KMemoryRegionType type_id) const { + DerivedRegionExtents extents; + + ASSERT(extents.first_region == nullptr); + ASSERT(extents.last_region == nullptr); + + for (auto it = this->cbegin(); it != this->cend(); it++) { + if (it->IsDerivedFrom(type_id)) { + if (extents.first_region == nullptr) { + extents.first_region = std::addressof(*it); + } + extents.last_region = std::addressof(*it); + } + } + + ASSERT(extents.first_region != nullptr); + ASSERT(extents.last_region != nullptr); + + return extents; + } + + DerivedRegionExtents GetDerivedRegionExtents(u32 type_id) const { + return GetDerivedRegionExtents(static_cast<KMemoryRegionType>(type_id)); + } + +public: + void InsertDirectly(u64 address, u64 last_address, u32 attr = 0, u32 type_id = 0); + bool Insert(u64 address, size_t size, u32 type_id, u32 new_attr = 0, u32 old_attr = 0); + + VAddr GetRandomAlignedRegion(size_t size, size_t alignment, u32 type_id); + + VAddr GetRandomAlignedRegionWithGuard(size_t size, size_t alignment, u32 type_id, + size_t guard_size) { + return this->GetRandomAlignedRegion(size + 2 * guard_size, alignment, type_id) + guard_size; + } + +public: + // Iterator accessors. + iterator begin() { + return m_tree.begin(); + } + + const_iterator begin() const { + return m_tree.begin(); + } + + iterator end() { + return m_tree.end(); + } + + const_iterator end() const { + return m_tree.end(); + } + + const_iterator cbegin() const { + return this->begin(); + } + + const_iterator cend() const { + return this->end(); + } + + iterator iterator_to(reference ref) { + return m_tree.iterator_to(ref); + } + + const_iterator iterator_to(const_reference ref) const { + return m_tree.iterator_to(ref); + } + + // Content management. + bool empty() const { + return m_tree.empty(); + } + + reference back() { + return m_tree.back(); + } + + const_reference back() const { + return m_tree.back(); + } + + reference front() { + return m_tree.front(); + } + + const_reference front() const { + return m_tree.front(); + } + + iterator insert(reference ref) { + return m_tree.insert(ref); + } + + iterator erase(iterator it) { + return m_tree.erase(it); + } + + iterator find(const_reference ref) const { + return m_tree.find(ref); + } + + iterator nfind(const_reference ref) const { + return m_tree.nfind(ref); + } +}; + +class KMemoryRegionAllocator final : NonCopyable { +public: + static constexpr size_t MaxMemoryRegions = 200; + + constexpr KMemoryRegionAllocator() = default; + + template <typename... Args> + KMemoryRegion* Allocate(Args&&... args) { + // Ensure we stay within the bounds of our heap. + ASSERT(this->num_regions < MaxMemoryRegions); + + // Create the new region. + KMemoryRegion* region = std::addressof(this->region_heap[this->num_regions++]); + new (region) KMemoryRegion(std::forward<Args>(args)...); + + return region; + } + +private: + std::array<KMemoryRegion, MaxMemoryRegions> region_heap{}; + size_t num_regions{}; +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_memory_region_type.h b/src/core/hle/kernel/k_memory_region_type.h new file mode 100644 index 000000000..a05e66677 --- /dev/null +++ b/src/core/hle/kernel/k_memory_region_type.h @@ -0,0 +1,338 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/bit_util.h" +#include "common/common_funcs.h" +#include "common/common_types.h" + +#define ARCH_ARM64 +#define BOARD_NINTENDO_NX + +namespace Kernel { + +enum KMemoryRegionType : u32 { + KMemoryRegionAttr_CarveoutProtected = 0x04000000, + KMemoryRegionAttr_DidKernelMap = 0x08000000, + KMemoryRegionAttr_ShouldKernelMap = 0x10000000, + KMemoryRegionAttr_UserReadOnly = 0x20000000, + KMemoryRegionAttr_NoUserMap = 0x40000000, + KMemoryRegionAttr_LinearMapped = 0x80000000, +}; +DECLARE_ENUM_FLAG_OPERATORS(KMemoryRegionType); + +namespace impl { + +constexpr size_t BitsForDeriveSparse(size_t n) { + return n + 1; +} + +constexpr size_t BitsForDeriveDense(size_t n) { + size_t low = 0, high = 1; + for (size_t i = 0; i < n - 1; ++i) { + if ((++low) == high) { + ++high; + low = 0; + } + } + return high + 1; +} + +class KMemoryRegionTypeValue { +public: + using ValueType = std::underlying_type_t<KMemoryRegionType>; + + constexpr KMemoryRegionTypeValue() = default; + + constexpr operator KMemoryRegionType() const { + return static_cast<KMemoryRegionType>(m_value); + } + + constexpr ValueType GetValue() const { + return m_value; + } + + constexpr const KMemoryRegionTypeValue& Finalize() { + m_finalized = true; + return *this; + } + + constexpr const KMemoryRegionTypeValue& SetSparseOnly() { + m_sparse_only = true; + return *this; + } + + constexpr const KMemoryRegionTypeValue& SetDenseOnly() { + m_dense_only = true; + return *this; + } + + constexpr KMemoryRegionTypeValue& SetAttribute(u32 attr) { + m_value |= attr; + return *this; + } + + constexpr KMemoryRegionTypeValue DeriveInitial( + size_t i, size_t next = Common::BitSize<ValueType>()) const { + KMemoryRegionTypeValue new_type = *this; + new_type.m_value = (ValueType{1} << i); + new_type.m_next_bit = next; + return new_type; + } + + constexpr KMemoryRegionTypeValue DeriveAttribute(u32 attr) const { + KMemoryRegionTypeValue new_type = *this; + new_type.m_value |= attr; + return new_type; + } + + constexpr KMemoryRegionTypeValue DeriveTransition(size_t ofs = 0, size_t adv = 1) const { + KMemoryRegionTypeValue new_type = *this; + new_type.m_value |= (ValueType{1} << (m_next_bit + ofs)); + new_type.m_next_bit += adv; + return new_type; + } + + constexpr KMemoryRegionTypeValue DeriveSparse(size_t ofs, size_t n, size_t i) const { + KMemoryRegionTypeValue new_type = *this; + new_type.m_value |= (ValueType{1} << (m_next_bit + ofs)); + new_type.m_value |= (ValueType{1} << (m_next_bit + ofs + 1 + i)); + new_type.m_next_bit += ofs + n + 1; + return new_type; + } + + constexpr KMemoryRegionTypeValue Derive(size_t n, size_t i) const { + size_t low = 0, high = 1; + for (size_t j = 0; j < i; ++j) { + if ((++low) == high) { + ++high; + low = 0; + } + } + + KMemoryRegionTypeValue new_type = *this; + new_type.m_value |= (ValueType{1} << (m_next_bit + low)); + new_type.m_value |= (ValueType{1} << (m_next_bit + high)); + new_type.m_next_bit += BitsForDeriveDense(n); + return new_type; + } + + constexpr KMemoryRegionTypeValue Advance(size_t n) const { + KMemoryRegionTypeValue new_type = *this; + new_type.m_next_bit += n; + return new_type; + } + + constexpr bool IsAncestorOf(ValueType v) const { + return (m_value | v) == v; + } + +private: + constexpr KMemoryRegionTypeValue(ValueType v) : m_value(v) {} + +private: + ValueType m_value{}; + size_t m_next_bit{}; + bool m_finalized{}; + bool m_sparse_only{}; + bool m_dense_only{}; +}; + +} // namespace impl + +constexpr auto KMemoryRegionType_None = impl::KMemoryRegionTypeValue(); +constexpr auto KMemoryRegionType_Kernel = KMemoryRegionType_None.DeriveInitial(0, 2); +constexpr auto KMemoryRegionType_Dram = KMemoryRegionType_None.DeriveInitial(1, 2); +static_assert(KMemoryRegionType_Kernel.GetValue() == 0x1); +static_assert(KMemoryRegionType_Dram.GetValue() == 0x2); + +constexpr auto KMemoryRegionType_DramKernelBase = + KMemoryRegionType_Dram.DeriveSparse(0, 3, 0) + .SetAttribute(KMemoryRegionAttr_NoUserMap) + .SetAttribute(KMemoryRegionAttr_CarveoutProtected); +constexpr auto KMemoryRegionType_DramReservedBase = KMemoryRegionType_Dram.DeriveSparse(0, 3, 1); +constexpr auto KMemoryRegionType_DramHeapBase = + KMemoryRegionType_Dram.DeriveSparse(0, 3, 2).SetAttribute(KMemoryRegionAttr_LinearMapped); +static_assert(KMemoryRegionType_DramKernelBase.GetValue() == + (0xE | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_DramReservedBase.GetValue() == (0x16)); +static_assert(KMemoryRegionType_DramHeapBase.GetValue() == (0x26 | KMemoryRegionAttr_LinearMapped)); + +constexpr auto KMemoryRegionType_DramKernelCode = + KMemoryRegionType_DramKernelBase.DeriveSparse(0, 4, 0); +constexpr auto KMemoryRegionType_DramKernelSlab = + KMemoryRegionType_DramKernelBase.DeriveSparse(0, 4, 1); +constexpr auto KMemoryRegionType_DramKernelPtHeap = + KMemoryRegionType_DramKernelBase.DeriveSparse(0, 4, 2).SetAttribute( + KMemoryRegionAttr_LinearMapped); +constexpr auto KMemoryRegionType_DramKernelInitPt = + KMemoryRegionType_DramKernelBase.DeriveSparse(0, 4, 3).SetAttribute( + KMemoryRegionAttr_LinearMapped); +static_assert(KMemoryRegionType_DramKernelCode.GetValue() == + (0xCE | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_DramKernelSlab.GetValue() == + (0x14E | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_DramKernelPtHeap.GetValue() == + (0x24E | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap | + KMemoryRegionAttr_LinearMapped)); +static_assert(KMemoryRegionType_DramKernelInitPt.GetValue() == + (0x44E | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap | + KMemoryRegionAttr_LinearMapped)); + +constexpr auto KMemoryRegionType_DramReservedEarly = + KMemoryRegionType_DramReservedBase.DeriveAttribute(KMemoryRegionAttr_NoUserMap); +static_assert(KMemoryRegionType_DramReservedEarly.GetValue() == + (0x16 | KMemoryRegionAttr_NoUserMap)); + +constexpr auto KMemoryRegionType_KernelTraceBuffer = + KMemoryRegionType_DramReservedBase.DeriveSparse(0, 3, 0) + .SetAttribute(KMemoryRegionAttr_LinearMapped) + .SetAttribute(KMemoryRegionAttr_UserReadOnly); +constexpr auto KMemoryRegionType_OnMemoryBootImage = + KMemoryRegionType_DramReservedBase.DeriveSparse(0, 3, 1); +constexpr auto KMemoryRegionType_DTB = KMemoryRegionType_DramReservedBase.DeriveSparse(0, 3, 2); +static_assert(KMemoryRegionType_KernelTraceBuffer.GetValue() == + (0xD6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_UserReadOnly)); +static_assert(KMemoryRegionType_OnMemoryBootImage.GetValue() == 0x156); +static_assert(KMemoryRegionType_DTB.GetValue() == 0x256); + +constexpr auto KMemoryRegionType_DramPoolPartition = + KMemoryRegionType_DramHeapBase.DeriveAttribute(KMemoryRegionAttr_NoUserMap); +static_assert(KMemoryRegionType_DramPoolPartition.GetValue() == + (0x26 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap)); + +constexpr auto KMemoryRegionType_DramPoolManagement = + KMemoryRegionType_DramPoolPartition.DeriveTransition(0, 2).DeriveTransition().SetAttribute( + KMemoryRegionAttr_CarveoutProtected); +constexpr auto KMemoryRegionType_DramUserPool = + KMemoryRegionType_DramPoolPartition.DeriveTransition(1, 2).DeriveTransition(); +static_assert(KMemoryRegionType_DramPoolManagement.GetValue() == + (0x166 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap | + KMemoryRegionAttr_CarveoutProtected)); +static_assert(KMemoryRegionType_DramUserPool.GetValue() == + (0x1A6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap)); + +constexpr auto KMemoryRegionType_DramApplicationPool = KMemoryRegionType_DramUserPool.Derive(4, 0); +constexpr auto KMemoryRegionType_DramAppletPool = KMemoryRegionType_DramUserPool.Derive(4, 1); +constexpr auto KMemoryRegionType_DramSystemNonSecurePool = + KMemoryRegionType_DramUserPool.Derive(4, 2); +constexpr auto KMemoryRegionType_DramSystemPool = + KMemoryRegionType_DramUserPool.Derive(4, 3).SetAttribute(KMemoryRegionAttr_CarveoutProtected); +static_assert(KMemoryRegionType_DramApplicationPool.GetValue() == + (0x7A6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_DramAppletPool.GetValue() == + (0xBA6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_DramSystemNonSecurePool.GetValue() == + (0xDA6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap)); +static_assert(KMemoryRegionType_DramSystemPool.GetValue() == + (0x13A6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap | + KMemoryRegionAttr_CarveoutProtected)); + +constexpr auto KMemoryRegionType_VirtualDramHeapBase = KMemoryRegionType_Dram.DeriveSparse(1, 3, 0); +constexpr auto KMemoryRegionType_VirtualDramKernelPtHeap = + KMemoryRegionType_Dram.DeriveSparse(1, 3, 1); +constexpr auto KMemoryRegionType_VirtualDramKernelTraceBuffer = + KMemoryRegionType_Dram.DeriveSparse(1, 3, 2); +static_assert(KMemoryRegionType_VirtualDramHeapBase.GetValue() == 0x1A); +static_assert(KMemoryRegionType_VirtualDramKernelPtHeap.GetValue() == 0x2A); +static_assert(KMemoryRegionType_VirtualDramKernelTraceBuffer.GetValue() == 0x4A); + +constexpr auto KMemoryRegionType_VirtualDramKernelInitPt = + KMemoryRegionType_VirtualDramHeapBase.Derive(3, 0); +constexpr auto KMemoryRegionType_VirtualDramPoolManagement = + KMemoryRegionType_VirtualDramHeapBase.Derive(3, 1); +constexpr auto KMemoryRegionType_VirtualDramUserPool = + KMemoryRegionType_VirtualDramHeapBase.Derive(3, 2); +static_assert(KMemoryRegionType_VirtualDramKernelInitPt.GetValue() == 0x19A); +static_assert(KMemoryRegionType_VirtualDramPoolManagement.GetValue() == 0x29A); +static_assert(KMemoryRegionType_VirtualDramUserPool.GetValue() == 0x31A); + +// NOTE: For unknown reason, the pools are derived out-of-order here. It's worth eventually trying +// to understand why Nintendo made this choice. +// UNUSED: .Derive(6, 0); +// UNUSED: .Derive(6, 1); +constexpr auto KMemoryRegionType_VirtualDramAppletPool = + KMemoryRegionType_VirtualDramUserPool.Derive(6, 2); +constexpr auto KMemoryRegionType_VirtualDramApplicationPool = + KMemoryRegionType_VirtualDramUserPool.Derive(6, 3); +constexpr auto KMemoryRegionType_VirtualDramSystemNonSecurePool = + KMemoryRegionType_VirtualDramUserPool.Derive(6, 4); +constexpr auto KMemoryRegionType_VirtualDramSystemPool = + KMemoryRegionType_VirtualDramUserPool.Derive(6, 5); +static_assert(KMemoryRegionType_VirtualDramAppletPool.GetValue() == 0x1B1A); +static_assert(KMemoryRegionType_VirtualDramApplicationPool.GetValue() == 0x271A); +static_assert(KMemoryRegionType_VirtualDramSystemNonSecurePool.GetValue() == 0x2B1A); +static_assert(KMemoryRegionType_VirtualDramSystemPool.GetValue() == 0x331A); + +constexpr auto KMemoryRegionType_ArchDeviceBase = + KMemoryRegionType_Kernel.DeriveTransition(0, 1).SetSparseOnly(); +constexpr auto KMemoryRegionType_BoardDeviceBase = + KMemoryRegionType_Kernel.DeriveTransition(0, 2).SetDenseOnly(); +static_assert(KMemoryRegionType_ArchDeviceBase.GetValue() == 0x5); +static_assert(KMemoryRegionType_BoardDeviceBase.GetValue() == 0x5); + +#if defined(ARCH_ARM64) +#include "core/hle/kernel/arch/arm64/k_memory_region_device_types.inc" +#elif defined(ARCH_ARM) +#error "Unimplemented" +#else +// Default to no architecture devices. +constexpr auto NumArchitectureDeviceRegions = 0; +#endif +static_assert(NumArchitectureDeviceRegions >= 0); + +#if defined(BOARD_NINTENDO_NX) +#include "core/hle/kernel/board/nintendo/nx/k_memory_region_device_types.inc" +#else +// Default to no board devices. +constexpr auto NumBoardDeviceRegions = 0; +#endif +static_assert(NumBoardDeviceRegions >= 0); + +constexpr auto KMemoryRegionType_KernelCode = KMemoryRegionType_Kernel.DeriveSparse(1, 4, 0); +constexpr auto KMemoryRegionType_KernelStack = KMemoryRegionType_Kernel.DeriveSparse(1, 4, 1); +constexpr auto KMemoryRegionType_KernelMisc = KMemoryRegionType_Kernel.DeriveSparse(1, 4, 2); +constexpr auto KMemoryRegionType_KernelSlab = KMemoryRegionType_Kernel.DeriveSparse(1, 4, 3); +static_assert(KMemoryRegionType_KernelCode.GetValue() == 0x19); +static_assert(KMemoryRegionType_KernelStack.GetValue() == 0x29); +static_assert(KMemoryRegionType_KernelMisc.GetValue() == 0x49); +static_assert(KMemoryRegionType_KernelSlab.GetValue() == 0x89); + +constexpr auto KMemoryRegionType_KernelMiscDerivedBase = + KMemoryRegionType_KernelMisc.DeriveTransition(); +static_assert(KMemoryRegionType_KernelMiscDerivedBase.GetValue() == 0x149); + +// UNUSED: .Derive(7, 0); +constexpr auto KMemoryRegionType_KernelMiscMainStack = + KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 1); +constexpr auto KMemoryRegionType_KernelMiscMappedDevice = + KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 2); +constexpr auto KMemoryRegionType_KernelMiscExceptionStack = + KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 3); +constexpr auto KMemoryRegionType_KernelMiscUnknownDebug = + KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 4); +// UNUSED: .Derive(7, 5); +constexpr auto KMemoryRegionType_KernelMiscIdleStack = + KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 6); +static_assert(KMemoryRegionType_KernelMiscMainStack.GetValue() == 0xB49); +static_assert(KMemoryRegionType_KernelMiscMappedDevice.GetValue() == 0xD49); +static_assert(KMemoryRegionType_KernelMiscExceptionStack.GetValue() == 0x1349); +static_assert(KMemoryRegionType_KernelMiscUnknownDebug.GetValue() == 0x1549); +static_assert(KMemoryRegionType_KernelMiscIdleStack.GetValue() == 0x2349); + +constexpr auto KMemoryRegionType_KernelTemp = KMemoryRegionType_Kernel.Advance(2).Derive(2, 0); +static_assert(KMemoryRegionType_KernelTemp.GetValue() == 0x31); + +constexpr KMemoryRegionType GetTypeForVirtualLinearMapping(u32 type_id) { + if (KMemoryRegionType_KernelTraceBuffer.IsAncestorOf(type_id)) { + return KMemoryRegionType_VirtualDramKernelTraceBuffer; + } else if (KMemoryRegionType_DramKernelPtHeap.IsAncestorOf(type_id)) { + return KMemoryRegionType_VirtualDramKernelPtHeap; + } else { + return KMemoryRegionType_Dram; + } +} + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_scheduler.cpp b/src/core/hle/kernel/k_scheduler.cpp index e7de48476..d1df97305 100644 --- a/src/core/hle/kernel/k_scheduler.cpp +++ b/src/core/hle/kernel/k_scheduler.cpp @@ -62,7 +62,7 @@ void KScheduler::RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedul } u64 KScheduler::UpdateHighestPriorityThread(KThread* highest_thread) { - std::scoped_lock lock{guard}; + KScopedSpinLock lk{guard}; if (KThread* prev_highest_thread = state.highest_priority_thread; prev_highest_thread != highest_thread) { if (prev_highest_thread != nullptr) { @@ -637,11 +637,11 @@ void KScheduler::RescheduleCurrentCore() { if (phys_core.IsInterrupted()) { phys_core.ClearInterrupt(); } - guard.lock(); + guard.Lock(); if (state.needs_scheduling.load()) { Schedule(); } else { - guard.unlock(); + guard.Unlock(); } } @@ -669,7 +669,7 @@ void KScheduler::Unload(KThread* thread) { } else { prev_thread = nullptr; } - thread->context_guard.unlock(); + thread->context_guard.Unlock(); } } @@ -713,7 +713,7 @@ void KScheduler::ScheduleImpl() { // If we're not actually switching thread, there's nothing to do. if (next_thread == current_thread.load()) { - guard.unlock(); + guard.Unlock(); return; } @@ -732,7 +732,7 @@ void KScheduler::ScheduleImpl() { } else { old_context = &idle_thread->GetHostContext(); } - guard.unlock(); + guard.Unlock(); Common::Fiber::YieldTo(*old_context, *switch_fiber); /// When a thread wakes up, the scheduler may have changed to other in another core. @@ -748,24 +748,24 @@ void KScheduler::OnSwitch(void* this_scheduler) { void KScheduler::SwitchToCurrent() { while (true) { { - std::scoped_lock lock{guard}; + KScopedSpinLock lk{guard}; current_thread.store(state.highest_priority_thread); state.needs_scheduling.store(false); } const auto is_switch_pending = [this] { - std::scoped_lock lock{guard}; + KScopedSpinLock lk{guard}; return state.needs_scheduling.load(); }; do { auto next_thread = current_thread.load(); if (next_thread != nullptr) { - next_thread->context_guard.lock(); + next_thread->context_guard.Lock(); if (next_thread->GetRawState() != ThreadState::Runnable) { - next_thread->context_guard.unlock(); + next_thread->context_guard.Unlock(); break; } if (next_thread->GetActiveCore() != core_id) { - next_thread->context_guard.unlock(); + next_thread->context_guard.Unlock(); break; } } diff --git a/src/core/hle/kernel/k_scheduler.h b/src/core/hle/kernel/k_scheduler.h index f595b9a5c..8e32865aa 100644 --- a/src/core/hle/kernel/k_scheduler.h +++ b/src/core/hle/kernel/k_scheduler.h @@ -2,19 +2,16 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -// This file references various implementation details from Atmosphere, an open-source firmware for -// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. - #pragma once #include <atomic> #include "common/common_types.h" -#include "common/spin_lock.h" #include "core/hle/kernel/global_scheduler_context.h" #include "core/hle/kernel/k_priority_queue.h" #include "core/hle/kernel/k_scheduler_lock.h" #include "core/hle/kernel/k_scoped_lock.h" +#include "core/hle/kernel/k_spin_lock.h" namespace Common { class Fiber; @@ -195,12 +192,12 @@ private: u64 last_context_switch_time{}; const s32 core_id; - Common::SpinLock guard{}; + KSpinLock guard{}; }; -class KScopedSchedulerLock : KScopedLock<GlobalSchedulerContext::LockType> { +class [[nodiscard]] KScopedSchedulerLock : KScopedLock<GlobalSchedulerContext::LockType> { public: - explicit KScopedSchedulerLock(KernelCore& kernel); + explicit KScopedSchedulerLock(KernelCore & kernel); ~KScopedSchedulerLock(); }; diff --git a/src/core/hle/kernel/k_scheduler_lock.h b/src/core/hle/kernel/k_scheduler_lock.h index 169455d18..47e315555 100644 --- a/src/core/hle/kernel/k_scheduler_lock.h +++ b/src/core/hle/kernel/k_scheduler_lock.h @@ -2,14 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -// This file references various implementation details from Atmosphere, an open-source firmware for -// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. - #pragma once #include "common/assert.h" -#include "common/spin_lock.h" #include "core/hardware_properties.h" +#include "core/hle/kernel/k_spin_lock.h" #include "core/hle/kernel/k_thread.h" #include "core/hle/kernel/kernel.h" @@ -34,7 +31,7 @@ public: } else { // Otherwise, we want to disable scheduling and acquire the spinlock. SchedulerType::DisableScheduling(kernel); - spin_lock.lock(); + spin_lock.Lock(); // For debug, ensure that our state is valid. ASSERT(lock_count == 0); @@ -58,7 +55,7 @@ public: // Note that we no longer hold the lock, and unlock the spinlock. owner_thread = nullptr; - spin_lock.unlock(); + spin_lock.Unlock(); // Enable scheduling, and perform a rescheduling operation. SchedulerType::EnableScheduling(kernel, cores_needing_scheduling); @@ -67,7 +64,7 @@ public: private: KernelCore& kernel; - Common::SpinLock spin_lock{}; + KAlignedSpinLock spin_lock{}; s32 lock_count{}; KThread* owner_thread{}; }; diff --git a/src/core/hle/kernel/k_scoped_lock.h b/src/core/hle/kernel/k_scoped_lock.h index d7cc557b2..72c3b0252 100644 --- a/src/core/hle/kernel/k_scoped_lock.h +++ b/src/core/hle/kernel/k_scoped_lock.h @@ -20,19 +20,22 @@ concept KLockable = !std::is_reference_v<T> && requires(T & t) { }; template <typename T> -requires KLockable<T> class KScopedLock { +requires KLockable<T> class [[nodiscard]] KScopedLock { public: - explicit KScopedLock(T* l) : lock_ptr(l) { + explicit KScopedLock(T * l) : lock_ptr(l) { this->lock_ptr->Lock(); } - explicit KScopedLock(T& l) : KScopedLock(std::addressof(l)) { /* ... */ - } + explicit KScopedLock(T & l) : KScopedLock(std::addressof(l)) {} + ~KScopedLock() { this->lock_ptr->Unlock(); } KScopedLock(const KScopedLock&) = delete; - KScopedLock(KScopedLock&&) = delete; + KScopedLock& operator=(const KScopedLock&) = delete; + + KScopedLock(KScopedLock &&) = delete; + KScopedLock& operator=(KScopedLock&&) = delete; private: T* lock_ptr; diff --git a/src/core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h b/src/core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h index f8189e107..ebecf0c77 100644 --- a/src/core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h +++ b/src/core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h @@ -15,9 +15,9 @@ namespace Kernel { -class KScopedSchedulerLockAndSleep { +class [[nodiscard]] KScopedSchedulerLockAndSleep { public: - explicit KScopedSchedulerLockAndSleep(KernelCore& kernel, KThread* t, s64 timeout) + explicit KScopedSchedulerLockAndSleep(KernelCore & kernel, KThread * t, s64 timeout) : kernel(kernel), thread(t), timeout_tick(timeout) { // Lock the scheduler. kernel.GlobalSchedulerContext().scheduler_lock.Lock(); diff --git a/src/core/hle/kernel/k_spin_lock.h b/src/core/hle/kernel/k_spin_lock.h index 12c4b2e88..4d87d006a 100644 --- a/src/core/hle/kernel/k_spin_lock.h +++ b/src/core/hle/kernel/k_spin_lock.h @@ -28,6 +28,12 @@ private: std::atomic_flag lck = ATOMIC_FLAG_INIT; }; +// TODO(bunnei): Alias for now, in case we want to implement these accurately in the future. +using KAlignedSpinLock = KSpinLock; +using KNotAlignedSpinLock = KSpinLock; + using KScopedSpinLock = KScopedLock<KSpinLock>; +using KScopedAlignedSpinLock = KScopedLock<KAlignedSpinLock>; +using KScopedNotAlignedSpinLock = KScopedLock<KNotAlignedSpinLock>; } // namespace Kernel diff --git a/src/core/hle/kernel/k_system_control.cpp b/src/core/hle/kernel/k_system_control.cpp deleted file mode 100644 index aa1682f69..000000000 --- a/src/core/hle/kernel/k_system_control.cpp +++ /dev/null @@ -1,42 +0,0 @@ -// Copyright 2021 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <random> - -#include "core/hle/kernel/k_system_control.h" - -namespace Kernel { - -namespace { -template <typename F> -u64 GenerateUniformRange(u64 min, u64 max, F f) { - // Handle the case where the difference is too large to represent. - if (max == std::numeric_limits<u64>::max() && min == std::numeric_limits<u64>::min()) { - return f(); - } - - // Iterate until we get a value in range. - const u64 range_size = ((max + 1) - min); - const u64 effective_max = (std::numeric_limits<u64>::max() / range_size) * range_size; - while (true) { - if (const u64 rnd = f(); rnd < effective_max) { - return min + (rnd % range_size); - } - } -} - -} // Anonymous namespace - -u64 KSystemControl::GenerateRandomU64() { - static std::random_device device; - static std::mt19937 gen(device()); - static std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max()); - return distribution(gen); -} - -u64 KSystemControl::GenerateRandomRange(u64 min, u64 max) { - return GenerateUniformRange(min, max, GenerateRandomU64); -} - -} // namespace Kernel diff --git a/src/core/hle/kernel/k_system_control.h b/src/core/hle/kernel/k_system_control.h index 1d5b64ffa..d755082c2 100644 --- a/src/core/hle/kernel/k_system_control.h +++ b/src/core/hle/kernel/k_system_control.h @@ -6,14 +6,18 @@ #include "common/common_types.h" -namespace Kernel { +#define BOARD_NINTENDO_NX + +#ifdef BOARD_NINTENDO_NX -class KSystemControl { -public: - KSystemControl() = default; +#include "core/hle/kernel/board/nintendo/nx/k_system_control.h" - static u64 GenerateRandomRange(u64 min, u64 max); - static u64 GenerateRandomU64(); -}; +namespace Kernel { + +using Kernel::Board::Nintendo::Nx::KSystemControl; } // namespace Kernel + +#else +#error "Unknown board for KSystemControl" +#endif diff --git a/src/core/hle/kernel/k_thread.h b/src/core/hle/kernel/k_thread.h index 1c19b23dc..1c86fdd20 100644 --- a/src/core/hle/kernel/k_thread.h +++ b/src/core/hle/kernel/k_thread.h @@ -14,10 +14,10 @@ #include "common/common_types.h" #include "common/intrusive_red_black_tree.h" -#include "common/spin_lock.h" #include "core/arm/arm_interface.h" #include "core/hle/kernel/k_affinity_mask.h" #include "core/hle/kernel/k_light_lock.h" +#include "core/hle/kernel/k_spin_lock.h" #include "core/hle/kernel/k_synchronization_object.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/svc_common.h" @@ -732,7 +732,7 @@ private: s8 priority_inheritance_count{}; bool resource_limit_release_hint{}; StackParameters stack_parameters{}; - Common::SpinLock context_guard{}; + KSpinLock context_guard{}; // For emulation std::shared_ptr<Common::Fiber> host_context{}; diff --git a/src/core/hle/kernel/k_trace.h b/src/core/hle/kernel/k_trace.h new file mode 100644 index 000000000..91ebf9ab2 --- /dev/null +++ b/src/core/hle/kernel/k_trace.h @@ -0,0 +1,12 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +namespace Kernel { + +constexpr bool IsKTraceEnabled = false; +constexpr std::size_t KTraceBufferSize = IsKTraceEnabled ? 16 * 1024 * 1024 : 0; + +} // namespace Kernel diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp index 780008b08..8fd990577 100644 --- a/src/core/hle/kernel/kernel.cpp +++ b/src/core/hle/kernel/kernel.cpp @@ -1,4 +1,4 @@ -// Copyright 2014 Citra Emulator Project +// Copyright 2021 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. @@ -12,6 +12,7 @@ #include <utility> #include "common/assert.h" +#include "common/common_sizes.h" #include "common/logging/log.h" #include "common/microprofile.h" #include "common/thread.h" @@ -68,9 +69,9 @@ struct KernelCore::Impl { InitializePhysicalCores(); InitializeSystemResourceLimit(kernel, system); InitializeMemoryLayout(); - InitializePreemption(kernel); InitializeSchedulers(); InitializeSuspendThreads(); + InitializePreemption(kernel); } void InitializeCores() { @@ -143,10 +144,10 @@ struct KernelCore::Impl { ASSERT(system_resource_limit->SetLimitValue(LimitableResource::PhysicalMemory, 0x100000000) .IsSuccess()); ASSERT(system_resource_limit->SetLimitValue(LimitableResource::Threads, 800).IsSuccess()); - ASSERT(system_resource_limit->SetLimitValue(LimitableResource::Events, 700).IsSuccess()); + ASSERT(system_resource_limit->SetLimitValue(LimitableResource::Events, 900).IsSuccess()); ASSERT(system_resource_limit->SetLimitValue(LimitableResource::TransferMemory, 200) .IsSuccess()); - ASSERT(system_resource_limit->SetLimitValue(LimitableResource::Sessions, 933).IsSuccess()); + ASSERT(system_resource_limit->SetLimitValue(LimitableResource::Sessions, 1133).IsSuccess()); // Derived from recent software updates. The kernel reserves 27MB constexpr u64 kernel_size{0x1b00000}; @@ -268,45 +269,314 @@ struct KernelCore::Impl { return schedulers[thread_id]->GetCurrentThread(); } + void DeriveInitialMemoryLayout(KMemoryLayout& memory_layout) { + // Insert the root region for the virtual memory tree, from which all other regions will + // derive. + memory_layout.GetVirtualMemoryRegionTree().InsertDirectly( + KernelVirtualAddressSpaceBase, + KernelVirtualAddressSpaceBase + KernelVirtualAddressSpaceSize - 1); + + // Insert the root region for the physical memory tree, from which all other regions will + // derive. + memory_layout.GetPhysicalMemoryRegionTree().InsertDirectly( + KernelPhysicalAddressSpaceBase, + KernelPhysicalAddressSpaceBase + KernelPhysicalAddressSpaceSize - 1); + + // Save start and end for ease of use. + const VAddr code_start_virt_addr = KernelVirtualAddressCodeBase; + const VAddr code_end_virt_addr = KernelVirtualAddressCodeEnd; + + // Setup the containing kernel region. + constexpr size_t KernelRegionSize = Common::Size_1_GB; + constexpr size_t KernelRegionAlign = Common::Size_1_GB; + constexpr VAddr kernel_region_start = + Common::AlignDown(code_start_virt_addr, KernelRegionAlign); + size_t kernel_region_size = KernelRegionSize; + if (!(kernel_region_start + KernelRegionSize - 1 <= KernelVirtualAddressSpaceLast)) { + kernel_region_size = KernelVirtualAddressSpaceEnd - kernel_region_start; + } + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert( + kernel_region_start, kernel_region_size, KMemoryRegionType_Kernel)); + + // Setup the code region. + constexpr size_t CodeRegionAlign = PageSize; + constexpr VAddr code_region_start = + Common::AlignDown(code_start_virt_addr, CodeRegionAlign); + constexpr VAddr code_region_end = Common::AlignUp(code_end_virt_addr, CodeRegionAlign); + constexpr size_t code_region_size = code_region_end - code_region_start; + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert( + code_region_start, code_region_size, KMemoryRegionType_KernelCode)); + + // Setup board-specific device physical regions. + Init::SetupDevicePhysicalMemoryRegions(memory_layout); + + // Determine the amount of space needed for the misc region. + size_t misc_region_needed_size; + { + // Each core has a one page stack for all three stack types (Main, Idle, Exception). + misc_region_needed_size = Core::Hardware::NUM_CPU_CORES * (3 * (PageSize + PageSize)); + + // Account for each auto-map device. + for (const auto& region : memory_layout.GetPhysicalMemoryRegionTree()) { + if (region.HasTypeAttribute(KMemoryRegionAttr_ShouldKernelMap)) { + // Check that the region is valid. + ASSERT(region.GetEndAddress() != 0); + + // Account for the region. + misc_region_needed_size += + PageSize + (Common::AlignUp(region.GetLastAddress(), PageSize) - + Common::AlignDown(region.GetAddress(), PageSize)); + } + } + + // Multiply the needed size by three, to account for the need for guard space. + misc_region_needed_size *= 3; + } + + // Decide on the actual size for the misc region. + constexpr size_t MiscRegionAlign = KernelAslrAlignment; + constexpr size_t MiscRegionMinimumSize = Common::Size_32_MB; + const size_t misc_region_size = Common::AlignUp( + std::max(misc_region_needed_size, MiscRegionMinimumSize), MiscRegionAlign); + ASSERT(misc_region_size > 0); + + // Setup the misc region. + const VAddr misc_region_start = + memory_layout.GetVirtualMemoryRegionTree().GetRandomAlignedRegion( + misc_region_size, MiscRegionAlign, KMemoryRegionType_Kernel); + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert( + misc_region_start, misc_region_size, KMemoryRegionType_KernelMisc)); + + // Setup the stack region. + constexpr size_t StackRegionSize = Common::Size_14_MB; + constexpr size_t StackRegionAlign = KernelAslrAlignment; + const VAddr stack_region_start = + memory_layout.GetVirtualMemoryRegionTree().GetRandomAlignedRegion( + StackRegionSize, StackRegionAlign, KMemoryRegionType_Kernel); + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert( + stack_region_start, StackRegionSize, KMemoryRegionType_KernelStack)); + + // Determine the size of the resource region. + const size_t resource_region_size = memory_layout.GetResourceRegionSizeForInit(); + + // Determine the size of the slab region. + const size_t slab_region_size = Common::AlignUp(KernelSlabHeapSize, PageSize); + ASSERT(slab_region_size <= resource_region_size); + + // Setup the slab region. + const PAddr code_start_phys_addr = KernelPhysicalAddressCodeBase; + const PAddr code_end_phys_addr = code_start_phys_addr + code_region_size; + const PAddr slab_start_phys_addr = code_end_phys_addr; + const PAddr slab_end_phys_addr = slab_start_phys_addr + slab_region_size; + constexpr size_t SlabRegionAlign = KernelAslrAlignment; + const size_t slab_region_needed_size = + Common::AlignUp(code_end_phys_addr + slab_region_size, SlabRegionAlign) - + Common::AlignDown(code_end_phys_addr, SlabRegionAlign); + const VAddr slab_region_start = + memory_layout.GetVirtualMemoryRegionTree().GetRandomAlignedRegion( + slab_region_needed_size, SlabRegionAlign, KMemoryRegionType_Kernel) + + (code_end_phys_addr % SlabRegionAlign); + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert( + slab_region_start, slab_region_size, KMemoryRegionType_KernelSlab)); + + // Setup the temp region. + constexpr size_t TempRegionSize = Common::Size_128_MB; + constexpr size_t TempRegionAlign = KernelAslrAlignment; + const VAddr temp_region_start = + memory_layout.GetVirtualMemoryRegionTree().GetRandomAlignedRegion( + TempRegionSize, TempRegionAlign, KMemoryRegionType_Kernel); + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert(temp_region_start, TempRegionSize, + KMemoryRegionType_KernelTemp)); + + // Automatically map in devices that have auto-map attributes. + for (auto& region : memory_layout.GetPhysicalMemoryRegionTree()) { + // We only care about kernel regions. + if (!region.IsDerivedFrom(KMemoryRegionType_Kernel)) { + continue; + } + + // Check whether we should map the region. + if (!region.HasTypeAttribute(KMemoryRegionAttr_ShouldKernelMap)) { + continue; + } + + // If this region has already been mapped, no need to consider it. + if (region.HasTypeAttribute(KMemoryRegionAttr_DidKernelMap)) { + continue; + } + + // Check that the region is valid. + ASSERT(region.GetEndAddress() != 0); + + // Set the attribute to note we've mapped this region. + region.SetTypeAttribute(KMemoryRegionAttr_DidKernelMap); + + // Create a virtual pair region and insert it into the tree. + const PAddr map_phys_addr = Common::AlignDown(region.GetAddress(), PageSize); + const size_t map_size = + Common::AlignUp(region.GetEndAddress(), PageSize) - map_phys_addr; + const VAddr map_virt_addr = + memory_layout.GetVirtualMemoryRegionTree().GetRandomAlignedRegionWithGuard( + map_size, PageSize, KMemoryRegionType_KernelMisc, PageSize); + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert( + map_virt_addr, map_size, KMemoryRegionType_KernelMiscMappedDevice)); + region.SetPairAddress(map_virt_addr + region.GetAddress() - map_phys_addr); + } + + Init::SetupDramPhysicalMemoryRegions(memory_layout); + + // Insert a physical region for the kernel code region. + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + code_start_phys_addr, code_region_size, KMemoryRegionType_DramKernelCode)); + + // Insert a physical region for the kernel slab region. + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + slab_start_phys_addr, slab_region_size, KMemoryRegionType_DramKernelSlab)); + + // Determine size available for kernel page table heaps, requiring > 8 MB. + const PAddr resource_end_phys_addr = slab_start_phys_addr + resource_region_size; + const size_t page_table_heap_size = resource_end_phys_addr - slab_end_phys_addr; + ASSERT(page_table_heap_size / Common::Size_4_MB > 2); + + // Insert a physical region for the kernel page table heap region + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + slab_end_phys_addr, page_table_heap_size, KMemoryRegionType_DramKernelPtHeap)); + + // All DRAM regions that we haven't tagged by this point will be mapped under the linear + // mapping. Tag them. + for (auto& region : memory_layout.GetPhysicalMemoryRegionTree()) { + if (region.GetType() == KMemoryRegionType_Dram) { + // Check that the region is valid. + ASSERT(region.GetEndAddress() != 0); + + // Set the linear map attribute. + region.SetTypeAttribute(KMemoryRegionAttr_LinearMapped); + } + } + + // Get the linear region extents. + const auto linear_extents = + memory_layout.GetPhysicalMemoryRegionTree().GetDerivedRegionExtents( + KMemoryRegionAttr_LinearMapped); + ASSERT(linear_extents.GetEndAddress() != 0); + + // Setup the linear mapping region. + constexpr size_t LinearRegionAlign = Common::Size_1_GB; + const PAddr aligned_linear_phys_start = + Common::AlignDown(linear_extents.GetAddress(), LinearRegionAlign); + const size_t linear_region_size = + Common::AlignUp(linear_extents.GetEndAddress(), LinearRegionAlign) - + aligned_linear_phys_start; + const VAddr linear_region_start = + memory_layout.GetVirtualMemoryRegionTree().GetRandomAlignedRegionWithGuard( + linear_region_size, LinearRegionAlign, KMemoryRegionType_None, LinearRegionAlign); + + const u64 linear_region_phys_to_virt_diff = linear_region_start - aligned_linear_phys_start; + + // Map and create regions for all the linearly-mapped data. + { + PAddr cur_phys_addr = 0; + u64 cur_size = 0; + for (auto& region : memory_layout.GetPhysicalMemoryRegionTree()) { + if (!region.HasTypeAttribute(KMemoryRegionAttr_LinearMapped)) { + continue; + } + + ASSERT(region.GetEndAddress() != 0); + + if (cur_size == 0) { + cur_phys_addr = region.GetAddress(); + cur_size = region.GetSize(); + } else if (cur_phys_addr + cur_size == region.GetAddress()) { + cur_size += region.GetSize(); + } else { + cur_phys_addr = region.GetAddress(); + cur_size = region.GetSize(); + } + + const VAddr region_virt_addr = + region.GetAddress() + linear_region_phys_to_virt_diff; + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert( + region_virt_addr, region.GetSize(), + GetTypeForVirtualLinearMapping(region.GetType()))); + region.SetPairAddress(region_virt_addr); + + KMemoryRegion* virt_region = + memory_layout.GetVirtualMemoryRegionTree().FindModifiable(region_virt_addr); + ASSERT(virt_region != nullptr); + virt_region->SetPairAddress(region.GetAddress()); + } + } + + // Insert regions for the initial page table region. + ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert( + resource_end_phys_addr, KernelPageTableHeapSize, KMemoryRegionType_DramKernelInitPt)); + ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert( + resource_end_phys_addr + linear_region_phys_to_virt_diff, KernelPageTableHeapSize, + KMemoryRegionType_VirtualDramKernelInitPt)); + + // All linear-mapped DRAM regions that we haven't tagged by this point will be allocated to + // some pool partition. Tag them. + for (auto& region : memory_layout.GetPhysicalMemoryRegionTree()) { + if (region.GetType() == (KMemoryRegionType_Dram | KMemoryRegionAttr_LinearMapped)) { + region.SetType(KMemoryRegionType_DramPoolPartition); + } + } + + // Setup all other memory regions needed to arrange the pool partitions. + Init::SetupPoolPartitionMemoryRegions(memory_layout); + + // Cache all linear regions in their own trees for faster access, later. + memory_layout.InitializeLinearMemoryRegionTrees(aligned_linear_phys_start, + linear_region_start); + } + void InitializeMemoryLayout() { - // Initialize memory layout - constexpr KMemoryLayout layout{KMemoryLayout::GetDefaultLayout()}; + // Derive the initial memory layout from the emulated board + KMemoryLayout memory_layout; + DeriveInitialMemoryLayout(memory_layout); + + const auto system_pool = memory_layout.GetKernelSystemPoolRegionPhysicalExtents(); + const auto applet_pool = memory_layout.GetKernelAppletPoolRegionPhysicalExtents(); + const auto application_pool = memory_layout.GetKernelApplicationPoolRegionPhysicalExtents(); + + // Initialize memory managers + memory_manager = std::make_unique<KMemoryManager>(); + memory_manager->InitializeManager(KMemoryManager::Pool::Application, + application_pool.GetAddress(), + application_pool.GetEndAddress()); + memory_manager->InitializeManager(KMemoryManager::Pool::Applet, applet_pool.GetAddress(), + applet_pool.GetEndAddress()); + memory_manager->InitializeManager(KMemoryManager::Pool::System, system_pool.GetAddress(), + system_pool.GetEndAddress()); + + // Setup memory regions for emulated processes + // TODO(bunnei): These should not be hardcoded regions initialized within the kernel constexpr std::size_t hid_size{0x40000}; constexpr std::size_t font_size{0x1100000}; constexpr std::size_t irs_size{0x8000}; constexpr std::size_t time_size{0x1000}; - constexpr PAddr hid_addr{layout.System().StartAddress()}; - constexpr PAddr font_pa{layout.System().StartAddress() + hid_size}; - constexpr PAddr irs_addr{layout.System().StartAddress() + hid_size + font_size}; - constexpr PAddr time_addr{layout.System().StartAddress() + hid_size + font_size + irs_size}; - // Initialize memory manager - memory_manager = std::make_unique<KMemoryManager>(); - memory_manager->InitializeManager(KMemoryManager::Pool::Application, - layout.Application().StartAddress(), - layout.Application().EndAddress()); - memory_manager->InitializeManager(KMemoryManager::Pool::Applet, - layout.Applet().StartAddress(), - layout.Applet().EndAddress()); - memory_manager->InitializeManager(KMemoryManager::Pool::System, - layout.System().StartAddress(), - layout.System().EndAddress()); + const PAddr hid_phys_addr{system_pool.GetAddress()}; + const PAddr font_phys_addr{system_pool.GetAddress() + hid_size}; + const PAddr irs_phys_addr{system_pool.GetAddress() + hid_size + font_size}; + const PAddr time_phys_addr{system_pool.GetAddress() + hid_size + font_size + irs_size}; hid_shared_mem = Kernel::KSharedMemory::Create( - system.Kernel(), system.DeviceMemory(), nullptr, {hid_addr, hid_size / PageSize}, - KMemoryPermission::None, KMemoryPermission::Read, hid_addr, hid_size, + system.Kernel(), system.DeviceMemory(), nullptr, {hid_phys_addr, hid_size / PageSize}, + KMemoryPermission::None, KMemoryPermission::Read, hid_phys_addr, hid_size, "HID:SharedMemory"); font_shared_mem = Kernel::KSharedMemory::Create( - system.Kernel(), system.DeviceMemory(), nullptr, {font_pa, font_size / PageSize}, - KMemoryPermission::None, KMemoryPermission::Read, font_pa, font_size, + system.Kernel(), system.DeviceMemory(), nullptr, {font_phys_addr, font_size / PageSize}, + KMemoryPermission::None, KMemoryPermission::Read, font_phys_addr, font_size, "Font:SharedMemory"); irs_shared_mem = Kernel::KSharedMemory::Create( - system.Kernel(), system.DeviceMemory(), nullptr, {irs_addr, irs_size / PageSize}, - KMemoryPermission::None, KMemoryPermission::Read, irs_addr, irs_size, + system.Kernel(), system.DeviceMemory(), nullptr, {irs_phys_addr, irs_size / PageSize}, + KMemoryPermission::None, KMemoryPermission::Read, irs_phys_addr, irs_size, "IRS:SharedMemory"); time_shared_mem = Kernel::KSharedMemory::Create( - system.Kernel(), system.DeviceMemory(), nullptr, {time_addr, time_size / PageSize}, - KMemoryPermission::None, KMemoryPermission::Read, time_addr, time_size, + system.Kernel(), system.DeviceMemory(), nullptr, {time_phys_addr, time_size / PageSize}, + KMemoryPermission::None, KMemoryPermission::Read, time_phys_addr, time_size, "Time:SharedMemory"); // Allocate slab heaps diff --git a/src/core/hle/kernel/kernel.h b/src/core/hle/kernel/kernel.h index 56906f2da..a500e63bc 100644 --- a/src/core/hle/kernel/kernel.h +++ b/src/core/hle/kernel/kernel.h @@ -1,4 +1,4 @@ -// Copyright 2014 Citra Emulator Project / PPSSPP Project +// Copyright 2021 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. diff --git a/src/core/hle/kernel/process_capability.cpp b/src/core/hle/kernel/process_capability.cpp index 3fc326eab..1006ee50c 100644 --- a/src/core/hle/kernel/process_capability.cpp +++ b/src/core/hle/kernel/process_capability.cpp @@ -281,11 +281,6 @@ ResultCode ProcessCapabilities::HandleSyscallFlags(u32& set_svc_bits, u32 flags) continue; } - if (svc_number >= svc_capabilities.size()) { - LOG_ERROR(Kernel, "Process svc capability is out of range! svc_number={}", svc_number); - return ResultOutOfRange; - } - svc_capabilities[svc_number] = true; } diff --git a/src/core/hle/kernel/process_capability.h b/src/core/hle/kernel/process_capability.h index 73ad197fa..b7a9b2e45 100644 --- a/src/core/hle/kernel/process_capability.h +++ b/src/core/hle/kernel/process_capability.h @@ -68,7 +68,7 @@ enum class ProgramType { class ProcessCapabilities { public: using InterruptCapabilities = std::bitset<1024>; - using SyscallCapabilities = std::bitset<128>; + using SyscallCapabilities = std::bitset<192>; ProcessCapabilities() = default; ProcessCapabilities(const ProcessCapabilities&) = delete; diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index 326d3b9ec..fcffc746d 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -2455,6 +2455,74 @@ static const FunctionDef SVC_Table_32[] = { {0x79, nullptr, "Unknown"}, {0x7A, nullptr, "Unknown"}, {0x7B, nullptr, "TerminateProcess32"}, + {0x7C, nullptr, "GetProcessInfo32"}, + {0x7D, nullptr, "CreateResourceLimit32"}, + {0x7E, nullptr, "SetResourceLimitLimitValue32"}, + {0x7F, nullptr, "CallSecureMonitor32"}, + {0x80, nullptr, "Unknown"}, + {0x81, nullptr, "Unknown"}, + {0x82, nullptr, "Unknown"}, + {0x83, nullptr, "Unknown"}, + {0x84, nullptr, "Unknown"}, + {0x85, nullptr, "Unknown"}, + {0x86, nullptr, "Unknown"}, + {0x87, nullptr, "Unknown"}, + {0x88, nullptr, "Unknown"}, + {0x89, nullptr, "Unknown"}, + {0x8A, nullptr, "Unknown"}, + {0x8B, nullptr, "Unknown"}, + {0x8C, nullptr, "Unknown"}, + {0x8D, nullptr, "Unknown"}, + {0x8E, nullptr, "Unknown"}, + {0x8F, nullptr, "Unknown"}, + {0x90, nullptr, "Unknown"}, + {0x91, nullptr, "Unknown"}, + {0x92, nullptr, "Unknown"}, + {0x93, nullptr, "Unknown"}, + {0x94, nullptr, "Unknown"}, + {0x95, nullptr, "Unknown"}, + {0x96, nullptr, "Unknown"}, + {0x97, nullptr, "Unknown"}, + {0x98, nullptr, "Unknown"}, + {0x99, nullptr, "Unknown"}, + {0x9A, nullptr, "Unknown"}, + {0x9B, nullptr, "Unknown"}, + {0x9C, nullptr, "Unknown"}, + {0x9D, nullptr, "Unknown"}, + {0x9E, nullptr, "Unknown"}, + {0x9F, nullptr, "Unknown"}, + {0xA0, nullptr, "Unknown"}, + {0xA1, nullptr, "Unknown"}, + {0xA2, nullptr, "Unknown"}, + {0xA3, nullptr, "Unknown"}, + {0xA4, nullptr, "Unknown"}, + {0xA5, nullptr, "Unknown"}, + {0xA6, nullptr, "Unknown"}, + {0xA7, nullptr, "Unknown"}, + {0xA8, nullptr, "Unknown"}, + {0xA9, nullptr, "Unknown"}, + {0xAA, nullptr, "Unknown"}, + {0xAB, nullptr, "Unknown"}, + {0xAC, nullptr, "Unknown"}, + {0xAD, nullptr, "Unknown"}, + {0xAE, nullptr, "Unknown"}, + {0xAF, nullptr, "Unknown"}, + {0xB0, nullptr, "Unknown"}, + {0xB1, nullptr, "Unknown"}, + {0xB2, nullptr, "Unknown"}, + {0xB3, nullptr, "Unknown"}, + {0xB4, nullptr, "Unknown"}, + {0xB5, nullptr, "Unknown"}, + {0xB6, nullptr, "Unknown"}, + {0xB7, nullptr, "Unknown"}, + {0xB8, nullptr, "Unknown"}, + {0xB9, nullptr, "Unknown"}, + {0xBA, nullptr, "Unknown"}, + {0xBB, nullptr, "Unknown"}, + {0xBC, nullptr, "Unknown"}, + {0xBD, nullptr, "Unknown"}, + {0xBE, nullptr, "Unknown"}, + {0xBF, nullptr, "Unknown"}, }; static const FunctionDef SVC_Table_64[] = { @@ -2586,6 +2654,70 @@ static const FunctionDef SVC_Table_64[] = { {0x7D, SvcWrap64<CreateResourceLimit>, "CreateResourceLimit"}, {0x7E, SvcWrap64<SetResourceLimitLimitValue>, "SetResourceLimitLimitValue"}, {0x7F, nullptr, "CallSecureMonitor"}, + {0x80, nullptr, "Unknown"}, + {0x81, nullptr, "Unknown"}, + {0x82, nullptr, "Unknown"}, + {0x83, nullptr, "Unknown"}, + {0x84, nullptr, "Unknown"}, + {0x85, nullptr, "Unknown"}, + {0x86, nullptr, "Unknown"}, + {0x87, nullptr, "Unknown"}, + {0x88, nullptr, "Unknown"}, + {0x89, nullptr, "Unknown"}, + {0x8A, nullptr, "Unknown"}, + {0x8B, nullptr, "Unknown"}, + {0x8C, nullptr, "Unknown"}, + {0x8D, nullptr, "Unknown"}, + {0x8E, nullptr, "Unknown"}, + {0x8F, nullptr, "Unknown"}, + {0x90, nullptr, "Unknown"}, + {0x91, nullptr, "Unknown"}, + {0x92, nullptr, "Unknown"}, + {0x93, nullptr, "Unknown"}, + {0x94, nullptr, "Unknown"}, + {0x95, nullptr, "Unknown"}, + {0x96, nullptr, "Unknown"}, + {0x97, nullptr, "Unknown"}, + {0x98, nullptr, "Unknown"}, + {0x99, nullptr, "Unknown"}, + {0x9A, nullptr, "Unknown"}, + {0x9B, nullptr, "Unknown"}, + {0x9C, nullptr, "Unknown"}, + {0x9D, nullptr, "Unknown"}, + {0x9E, nullptr, "Unknown"}, + {0x9F, nullptr, "Unknown"}, + {0xA0, nullptr, "Unknown"}, + {0xA1, nullptr, "Unknown"}, + {0xA2, nullptr, "Unknown"}, + {0xA3, nullptr, "Unknown"}, + {0xA4, nullptr, "Unknown"}, + {0xA5, nullptr, "Unknown"}, + {0xA6, nullptr, "Unknown"}, + {0xA7, nullptr, "Unknown"}, + {0xA8, nullptr, "Unknown"}, + {0xA9, nullptr, "Unknown"}, + {0xAA, nullptr, "Unknown"}, + {0xAB, nullptr, "Unknown"}, + {0xAC, nullptr, "Unknown"}, + {0xAD, nullptr, "Unknown"}, + {0xAE, nullptr, "Unknown"}, + {0xAF, nullptr, "Unknown"}, + {0xB0, nullptr, "Unknown"}, + {0xB1, nullptr, "Unknown"}, + {0xB2, nullptr, "Unknown"}, + {0xB3, nullptr, "Unknown"}, + {0xB4, nullptr, "Unknown"}, + {0xB5, nullptr, "Unknown"}, + {0xB6, nullptr, "Unknown"}, + {0xB7, nullptr, "Unknown"}, + {0xB8, nullptr, "Unknown"}, + {0xB9, nullptr, "Unknown"}, + {0xBA, nullptr, "Unknown"}, + {0xBB, nullptr, "Unknown"}, + {0xBC, nullptr, "Unknown"}, + {0xBD, nullptr, "Unknown"}, + {0xBE, nullptr, "Unknown"}, + {0xBF, nullptr, "Unknown"}, }; static const FunctionDef* GetSVCInfo32(u32 func_num) { diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp index 615e20a54..52535ecc0 100644 --- a/src/core/hle/service/acc/acc.cpp +++ b/src/core/hle/service/acc/acc.cpp @@ -610,12 +610,17 @@ public: explicit DAUTH_O(Core::System& system_, Common::UUID) : ServiceFramework{system_, "dauth:o"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "EnsureAuthenticationTokenCacheAsync"}, // [5.0.0-5.1.0] GeneratePostData - {1, nullptr, "LoadAuthenticationTokenCache"}, // 6.0.0+ - {2, nullptr, "InvalidateAuthenticationTokenCache"}, // 6.0.0+ - {10, nullptr, "EnsureEdgeTokenCacheAsync"}, // 6.0.0+ - {11, nullptr, "LoadEdgeTokenCache"}, // 6.0.0+ - {12, nullptr, "InvalidateEdgeTokenCache"}, // 6.0.0+ + {0, nullptr, "EnsureAuthenticationTokenCacheAsync"}, + {1, nullptr, "LoadAuthenticationTokenCache"}, + {2, nullptr, "InvalidateAuthenticationTokenCache"}, + {10, nullptr, "EnsureEdgeTokenCacheAsync"}, + {11, nullptr, "LoadEdgeTokenCache"}, + {12, nullptr, "InvalidateEdgeTokenCache"}, + {20, nullptr, "EnsureApplicationAuthenticationCacheAsync"}, + {21, nullptr, "LoadApplicationAuthenticationTokenCache"}, + {22, nullptr, "LoadApplicationNetworkServiceClientConfigCache"}, + {23, nullptr, "IsApplicationAuthenticationCacheAvailable"}, + {24, nullptr, "InvalidateApplicationAuthenticationCache"}, }; // clang-format on diff --git a/src/core/hle/service/acc/acc_su.cpp b/src/core/hle/service/acc/acc_su.cpp index 49b22583e..bb6118abf 100644 --- a/src/core/hle/service/acc/acc_su.cpp +++ b/src/core/hle/service/acc/acc_su.cpp @@ -17,28 +17,30 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {3, &ACC_SU::ListOpenUsers, "ListOpenUsers"}, {4, &ACC_SU::GetLastOpenedUser, "GetLastOpenedUser"}, {5, &ACC_SU::GetProfile, "GetProfile"}, - {6, nullptr, "GetProfileDigest"}, // 3.0.0+ + {6, nullptr, "GetProfileDigest"}, {50, &ACC_SU::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, {51, &ACC_SU::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, - {60, &ACC_SU::ListOpenContextStoredUsers, "ListOpenContextStoredUsers"}, // 5.0.0 - 5.1.0 - {99, nullptr, "DebugActivateOpenContextRetention"}, // 6.0.0+ + {60, &ACC_SU::ListOpenContextStoredUsers, "ListOpenContextStoredUsers"}, + {99, nullptr, "DebugActivateOpenContextRetention"}, {100, nullptr, "GetUserRegistrationNotifier"}, {101, nullptr, "GetUserStateChangeNotifier"}, {102, nullptr, "GetBaasAccountManagerForSystemService"}, {103, nullptr, "GetBaasUserAvailabilityChangeNotifier"}, {104, nullptr, "GetProfileUpdateNotifier"}, - {105, nullptr, "CheckNetworkServiceAvailabilityAsync"}, // 4.0.0+ - {106, nullptr, "GetProfileSyncNotifier"}, // 9.0.0+ + {105, nullptr, "CheckNetworkServiceAvailabilityAsync"}, + {106, nullptr, "GetProfileSyncNotifier"}, {110, &ACC_SU::StoreSaveDataThumbnailSystem, "StoreSaveDataThumbnail"}, {111, nullptr, "ClearSaveDataThumbnail"}, {112, nullptr, "LoadSaveDataThumbnail"}, - {113, nullptr, "GetSaveDataThumbnailExistence"}, // 5.0.0+ - {120, nullptr, "ListOpenUsersInApplication"}, // 10.0.0+ - {130, nullptr, "ActivateOpenContextRetention"}, // 6.0.0+ - {140, &ACC_SU::ListQualifiedUsers, "ListQualifiedUsers"}, // 6.0.0+ - {150, nullptr, "AuthenticateApplicationAsync"}, // 10.0.0+ - {190, nullptr, "GetUserLastOpenedApplication"}, // 1.0.0 - 9.2.0 - {191, nullptr, "ActivateOpenContextHolder"}, // 7.0.0+ + {113, nullptr, "GetSaveDataThumbnailExistence"}, + {120, nullptr, "ListOpenUsersInApplication"}, + {130, nullptr, "ActivateOpenContextRetention"}, + {140, &ACC_SU::ListQualifiedUsers, "ListQualifiedUsers"}, + {150, nullptr, "AuthenticateApplicationAsync"}, + {151, nullptr, "Unknown151"}, + {152, nullptr, "Unknown152"}, + {190, nullptr, "GetUserLastOpenedApplication"}, + {191, nullptr, "ActivateOpenContextHolder"}, {200, nullptr, "BeginUserRegistration"}, {201, nullptr, "CompleteUserRegistration"}, {202, nullptr, "CancelUserRegistration"}, @@ -46,15 +48,15 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {204, nullptr, "SetUserPosition"}, {205, &ACC_SU::GetProfileEditor, "GetProfileEditor"}, {206, nullptr, "CompleteUserRegistrationForcibly"}, - {210, nullptr, "CreateFloatingRegistrationRequest"}, // 3.0.0+ - {211, nullptr, "CreateProcedureToRegisterUserWithNintendoAccount"}, // 8.0.0+ - {212, nullptr, "ResumeProcedureToRegisterUserWithNintendoAccount"}, // 8.0.0+ + {210, nullptr, "CreateFloatingRegistrationRequest"}, + {211, nullptr, "CreateProcedureToRegisterUserWithNintendoAccount"}, + {212, nullptr, "ResumeProcedureToRegisterUserWithNintendoAccount"}, {230, nullptr, "AuthenticateServiceAsync"}, {250, nullptr, "GetBaasAccountAdministrator"}, {290, nullptr, "ProxyProcedureForGuestLoginWithNintendoAccount"}, - {291, nullptr, "ProxyProcedureForFloatingRegistrationWithNintendoAccount"}, // 3.0.0+ + {291, nullptr, "ProxyProcedureForFloatingRegistrationWithNintendoAccount"}, {299, nullptr, "SuspendBackgroundDaemon"}, - {997, nullptr, "DebugInvalidateTokenCacheForUser"}, // 3.0.0+ + {997, nullptr, "DebugInvalidateTokenCacheForUser"}, {998, nullptr, "DebugSetUserStateClose"}, {999, nullptr, "DebugSetUserStateOpen"}, }; diff --git a/src/core/hle/service/acc/acc_u1.cpp b/src/core/hle/service/acc/acc_u1.cpp index 951081cd0..71982ad5a 100644 --- a/src/core/hle/service/acc/acc_u1.cpp +++ b/src/core/hle/service/acc/acc_u1.cpp @@ -17,29 +17,31 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {3, &ACC_U1::ListOpenUsers, "ListOpenUsers"}, {4, &ACC_U1::GetLastOpenedUser, "GetLastOpenedUser"}, {5, &ACC_U1::GetProfile, "GetProfile"}, - {6, nullptr, "GetProfileDigest"}, // 3.0.0+ + {6, nullptr, "GetProfileDigest"}, {50, &ACC_U1::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, {51, &ACC_U1::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, - {60, &ACC_U1::ListOpenContextStoredUsers, "ListOpenContextStoredUsers"}, // 5.0.0 - 5.1.0 - {99, nullptr, "DebugActivateOpenContextRetention"}, // 6.0.0+ + {60, &ACC_U1::ListOpenContextStoredUsers, "ListOpenContextStoredUsers"}, + {99, nullptr, "DebugActivateOpenContextRetention"}, {100, nullptr, "GetUserRegistrationNotifier"}, {101, nullptr, "GetUserStateChangeNotifier"}, {102, nullptr, "GetBaasAccountManagerForSystemService"}, {103, nullptr, "GetBaasUserAvailabilityChangeNotifier"}, {104, nullptr, "GetProfileUpdateNotifier"}, - {105, nullptr, "CheckNetworkServiceAvailabilityAsync"}, // 4.0.0+ - {106, nullptr, "GetProfileSyncNotifier"}, // 9.0.0+ + {105, nullptr, "CheckNetworkServiceAvailabilityAsync"}, + {106, nullptr, "GetProfileSyncNotifier"}, {110, &ACC_U1::StoreSaveDataThumbnailApplication, "StoreSaveDataThumbnail"}, {111, nullptr, "ClearSaveDataThumbnail"}, {112, nullptr, "LoadSaveDataThumbnail"}, - {113, nullptr, "GetSaveDataThumbnailExistence"}, // 5.0.0+ - {120, nullptr, "ListOpenUsersInApplication"}, // 10.0.0+ - {130, nullptr, "ActivateOpenContextRetention"}, // 6.0.0+ - {140, &ACC_U1::ListQualifiedUsers, "ListQualifiedUsers"}, // 6.0.0+ - {150, nullptr, "AuthenticateApplicationAsync"}, // 10.0.0+ - {190, nullptr, "GetUserLastOpenedApplication"}, // 1.0.0 - 9.2.0 - {191, nullptr, "ActivateOpenContextHolder"}, // 7.0.0+ - {997, nullptr, "DebugInvalidateTokenCacheForUser"}, // 3.0.0+ + {113, nullptr, "GetSaveDataThumbnailExistence"}, + {120, nullptr, "ListOpenUsersInApplication"}, + {130, nullptr, "ActivateOpenContextRetention"}, + {140, &ACC_U1::ListQualifiedUsers, "ListQualifiedUsers"}, + {150, nullptr, "AuthenticateApplicationAsync"}, + {151, nullptr, "Unknown151"}, + {152, nullptr, "Unknown152"}, + {190, nullptr, "GetUserLastOpenedApplication"}, + {191, nullptr, "ActivateOpenContextHolder"}, + {997, nullptr, "DebugInvalidateTokenCacheForUser"}, {998, nullptr, "DebugSetUserStateClose"}, {999, nullptr, "DebugSetUserStateOpen"}, }; diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp index 8e1fe9438..4374487a3 100644 --- a/src/core/hle/service/am/am.cpp +++ b/src/core/hle/service/am/am.cpp @@ -231,6 +231,7 @@ IDebugFunctions::IDebugFunctions(Core::System& system_) {10, nullptr, "PerformSystemButtonPressing"}, {20, nullptr, "InvalidateTransitionLayer"}, {30, nullptr, "RequestLaunchApplicationWithUserAndArgumentForDebug"}, + {31, nullptr, "RequestLaunchApplicationByApplicationLaunchInfoForDebug"}, {40, nullptr, "GetAppletResourceUsageInfo"}, {100, nullptr, "SetCpuBoostModeForApplet"}, {101, nullptr, "CancelCpuBoostModeForApplet"}, @@ -242,6 +243,7 @@ IDebugFunctions::IDebugFunctions(Core::System& system_) {130, nullptr, "FriendInvitationSetApplicationParameter"}, {131, nullptr, "FriendInvitationClearApplicationParameter"}, {132, nullptr, "FriendInvitationPushApplicationParameter"}, + {900, nullptr, "GetGrcProcessLaunchedSystemEvent"}, }; // clang-format on @@ -295,8 +297,9 @@ ISelfController::ISelfController(Core::System& system_, NVFlinger::NVFlinger& nv {80, nullptr, "SetWirelessPriorityMode"}, {90, &ISelfController::GetAccumulatedSuspendedTickValue, "GetAccumulatedSuspendedTickValue"}, {91, &ISelfController::GetAccumulatedSuspendedTickChangedEvent, "GetAccumulatedSuspendedTickChangedEvent"}, - {100, nullptr, "SetAlbumImageTakenNotificationEnabled"}, + {100, &ISelfController::SetAlbumImageTakenNotificationEnabled, "SetAlbumImageTakenNotificationEnabled"}, {110, nullptr, "SetApplicationAlbumUserData"}, + {120, nullptr, "SaveCurrentScreenshot"}, {1000, nullptr, "GetDebugStorageChannel"}, }; // clang-format on @@ -560,6 +563,21 @@ void ISelfController::GetAccumulatedSuspendedTickChangedEvent(Kernel::HLERequest rb.PushCopyObjects(accumulated_suspended_tick_changed_event->GetReadableEvent()); } +void ISelfController::SetAlbumImageTakenNotificationEnabled(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + // This service call sets an internal flag whether a notification is shown when an image is + // captured. Currently we do not support capturing images via the capture button, so this can be + // stubbed for now. + const bool album_image_taken_notification_enabled = rp.Pop<bool>(); + + LOG_WARNING(Service_AM, "(STUBBED) called. album_image_taken_notification_enabled={}", + album_image_taken_notification_enabled); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + AppletMessageQueue::AppletMessageQueue(Kernel::KernelCore& kernel) { on_new_message = Kernel::KEvent::Create(kernel, "AMMessageQueue:OnMessageReceived"); on_new_message->Initialize(); @@ -630,6 +648,7 @@ ICommonStateGetter::ICommonStateGetter(Core::System& system_, {11, nullptr, "ReleaseSleepLock"}, {12, nullptr, "ReleaseSleepLockTransiently"}, {13, nullptr, "GetAcquiredSleepLockEvent"}, + {14, nullptr, "GetWakeupCount"}, {20, nullptr, "PushToGeneralChannel"}, {30, nullptr, "GetHomeButtonReaderLockAccessor"}, {31, nullptr, "GetReaderLockAccessorEx"}, @@ -641,6 +660,7 @@ ICommonStateGetter::ICommonStateGetter(Core::System& system_, {53, &ICommonStateGetter::BeginVrModeEx, "BeginVrModeEx"}, {54, &ICommonStateGetter::EndVrModeEx, "EndVrModeEx"}, {55, nullptr, "IsInControllerFirmwareUpdateSection"}, + {59, nullptr, "SetVrPositionForDebug"}, {60, &ICommonStateGetter::GetDefaultDisplayResolution, "GetDefaultDisplayResolution"}, {61, &ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent, "GetDefaultDisplayResolutionChangeEvent"}, {62, nullptr, "GetHdcpAuthenticationState"}, @@ -649,14 +669,21 @@ ICommonStateGetter::ICommonStateGetter(Core::System& system_, {65, nullptr, "GetApplicationIdByContentActionName"}, {66, &ICommonStateGetter::SetCpuBoostMode, "SetCpuBoostMode"}, {67, nullptr, "CancelCpuBoostMode"}, + {68, nullptr, "GetBuiltInDisplayType"}, {80, nullptr, "PerformSystemButtonPressingIfInFocus"}, {90, nullptr, "SetPerformanceConfigurationChangedNotification"}, {91, nullptr, "GetCurrentPerformanceConfiguration"}, {100, nullptr, "SetHandlingHomeButtonShortPressedEnabled"}, + {110, nullptr, "OpenMyGpuErrorHandler"}, {200, nullptr, "GetOperationModeSystemInfo"}, {300, nullptr, "GetSettingsPlatformRegion"}, {400, nullptr, "ActivateMigrationService"}, {401, nullptr, "DeactivateMigrationService"}, + {500, nullptr, "DisableSleepTillShutdown"}, + {501, nullptr, "SuppressDisablingSleepTemporarily"}, + {502, nullptr, "IsSleepEnabled"}, + {503, nullptr, "IsDisablingSleepSuppressed"}, + {900, nullptr, "SetRequestExitToLibraryAppletAtExecuteNextProgramEnabled"}, }; // clang-format on @@ -1188,11 +1215,14 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {25, &IApplicationFunctions::ExtendSaveData, "ExtendSaveData"}, {26, &IApplicationFunctions::GetSaveDataSize, "GetSaveDataSize"}, {27, nullptr, "CreateCacheStorage"}, + {28, nullptr, "GetSaveDataSizeMax"}, + {29, nullptr, "GetCacheStorageMax"}, {30, &IApplicationFunctions::BeginBlockingHomeButtonShortAndLongPressed, "BeginBlockingHomeButtonShortAndLongPressed"}, {31, &IApplicationFunctions::EndBlockingHomeButtonShortAndLongPressed, "EndBlockingHomeButtonShortAndLongPressed"}, {32, &IApplicationFunctions::BeginBlockingHomeButton, "BeginBlockingHomeButton"}, {33, &IApplicationFunctions::EndBlockingHomeButton, "EndBlockingHomeButton"}, {34, nullptr, "SelectApplicationLicense"}, + {35, nullptr, "GetDeviceSaveDataSizeMax"}, {40, &IApplicationFunctions::NotifyRunning, "NotifyRunning"}, {50, &IApplicationFunctions::GetPseudoDeviceId, "GetPseudoDeviceId"}, {60, nullptr, "SetMediaPlaybackStateForApplication"}, @@ -1216,6 +1246,7 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {123, &IApplicationFunctions::GetPreviousProgramIndex, "GetPreviousProgramIndex"}, {124, nullptr, "EnableApplicationAllThreadDumpOnCrash"}, {130, &IApplicationFunctions::GetGpuErrorDetectedSystemEvent, "GetGpuErrorDetectedSystemEvent"}, + {131, nullptr, "SetDelayTimeToAbortOnGpuError"}, {140, &IApplicationFunctions::GetFriendInvitationStorageChannelEvent, "GetFriendInvitationStorageChannelEvent"}, {141, &IApplicationFunctions::TryPopFromFriendInvitationStorageChannel, "TryPopFromFriendInvitationStorageChannel"}, {150, nullptr, "GetNotificationStorageChannelEvent"}, @@ -1224,6 +1255,8 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {170, nullptr, "SetHdcpAuthenticationActivated"}, {180, nullptr, "GetLaunchRequiredVersion"}, {181, nullptr, "UpgradeLaunchRequiredVersion"}, + {190, nullptr, "SendServerMaintenanceOverlayNotification"}, + {200, nullptr, "GetLastApplicationExitReason"}, {500, nullptr, "StartContinuousRecordingFlushForDebug"}, {1000, nullptr, "CreateMovieMaker"}, {1001, nullptr, "PrepareForJit"}, @@ -1690,9 +1723,12 @@ IHomeMenuFunctions::IHomeMenuFunctions(Core::System& system_) {21, &IHomeMenuFunctions::GetPopFromGeneralChannelEvent, "GetPopFromGeneralChannelEvent"}, {30, nullptr, "GetHomeButtonWriterLockAccessor"}, {31, nullptr, "GetWriterLockAccessorEx"}, + {40, nullptr, "IsSleepEnabled"}, + {41, nullptr, "IsRebootEnabled"}, {100, nullptr, "PopRequestLaunchApplicationForDebug"}, {110, nullptr, "IsForceTerminateApplicationDisabledForDebug"}, {200, nullptr, "LaunchDevMenu"}, + {1000, nullptr, "SetLastApplicationExitReason"}, }; // clang-format on @@ -1736,6 +1772,7 @@ IGlobalStateController::IGlobalStateController(Core::System& system_) {13, nullptr, "UpdateDefaultDisplayResolution"}, {14, nullptr, "ShouldSleepOnBoot"}, {15, nullptr, "GetHdcpAuthenticationFailedEvent"}, + {30, nullptr, "OpenCradleFirmwareUpdater"}, }; // clang-format on diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h index 6911f0d6e..f6a453ab7 100644 --- a/src/core/hle/service/am/am.h +++ b/src/core/hle/service/am/am.h @@ -146,6 +146,7 @@ private: void IsAutoSleepDisabled(Kernel::HLERequestContext& ctx); void GetAccumulatedSuspendedTickValue(Kernel::HLERequestContext& ctx); void GetAccumulatedSuspendedTickChangedEvent(Kernel::HLERequestContext& ctx); + void SetAlbumImageTakenNotificationEnabled(Kernel::HLERequestContext& ctx); enum class ScreenshotPermission : u32 { Inherit = 0, diff --git a/src/core/hle/service/aoc/aoc_u.cpp b/src/core/hle/service/aoc/aoc_u.cpp index 8d657c0bf..0f51e5871 100644 --- a/src/core/hle/service/aoc/aoc_u.cpp +++ b/src/core/hle/service/aoc/aoc_u.cpp @@ -118,8 +118,10 @@ AOC_U::AOC_U(Core::System& system_) {7, &AOC_U::PrepareAddOnContent, "PrepareAddOnContent"}, {8, &AOC_U::GetAddOnContentListChangedEvent, "GetAddOnContentListChangedEvent"}, {9, nullptr, "GetAddOnContentLostErrorCode"}, + {10, nullptr, "GetAddOnContentListChangedEventWithProcessId"}, {100, &AOC_U::CreateEcPurchasedEventManager, "CreateEcPurchasedEventManager"}, {101, &AOC_U::CreatePermanentEcPurchasedEventManager, "CreatePermanentEcPurchasedEventManager"}, + {110, nullptr, "CreateContentsServiceManager"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audin_a.cpp b/src/core/hle/service/audio/audin_a.cpp index 79c3aa920..10acaad19 100644 --- a/src/core/hle/service/audio/audin_a.cpp +++ b/src/core/hle/service/audio/audin_a.cpp @@ -9,10 +9,10 @@ namespace Service::Audio { AudInA::AudInA(Core::System& system_) : ServiceFramework{system_, "audin:a"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "RequestSuspendAudioIns"}, - {1, nullptr, "RequestResumeAudioIns"}, - {2, nullptr, "GetAudioInsProcessMasterVolume"}, - {3, nullptr, "SetAudioInsProcessMasterVolume"}, + {0, nullptr, "RequestSuspend"}, + {1, nullptr, "RequestResume"}, + {2, nullptr, "GetProcessMasterVolume"}, + {3, nullptr, "SetProcessMasterVolume"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audin_u.cpp b/src/core/hle/service/audio/audin_u.cpp index 26a6deddf..ecd05e4a6 100644 --- a/src/core/hle/service/audio/audin_u.cpp +++ b/src/core/hle/service/audio/audin_u.cpp @@ -15,19 +15,19 @@ public: // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetAudioInState"}, - {1, nullptr, "StartAudioIn"}, - {2, nullptr, "StopAudioIn"}, + {1, nullptr, "Start"}, + {2, nullptr, "Stop"}, {3, nullptr, "AppendAudioInBuffer"}, {4, nullptr, "RegisterBufferEvent"}, {5, nullptr, "GetReleasedAudioInBuffer"}, {6, nullptr, "ContainsAudioInBuffer"}, - {7, nullptr, "AppendAudioInBufferWithUserEvent"}, + {7, nullptr, "AppendUacInBuffer"}, {8, nullptr, "AppendAudioInBufferAuto"}, - {9, nullptr, "GetReleasedAudioInBufferAuto"}, - {10, nullptr, "AppendAudioInBufferWithUserEventAuto"}, + {9, nullptr, "GetReleasedAudioInBuffersAuto"}, + {10, nullptr, "AppendUacInBufferAuto"}, {11, nullptr, "GetAudioInBufferCount"}, - {12, nullptr, "SetAudioInDeviceGain"}, - {13, nullptr, "GetAudioInDeviceGain"}, + {12, nullptr, "SetDeviceGain"}, + {13, nullptr, "GetDeviceGain"}, {14, nullptr, "FlushAudioInBuffers"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audout_a.cpp b/src/core/hle/service/audio/audout_a.cpp index 19825fd5d..3ee522b50 100644 --- a/src/core/hle/service/audio/audout_a.cpp +++ b/src/core/hle/service/audio/audout_a.cpp @@ -9,12 +9,12 @@ namespace Service::Audio { AudOutA::AudOutA(Core::System& system_) : ServiceFramework{system_, "audout:a"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "RequestSuspendAudioOuts"}, - {1, nullptr, "RequestResumeAudioOuts"}, - {2, nullptr, "GetAudioOutsProcessMasterVolume"}, - {3, nullptr, "SetAudioOutsProcessMasterVolume"}, - {4, nullptr, "GetAudioOutsProcessRecordVolume"}, - {5, nullptr, "SetAudioOutsProcessRecordVolume"}, + {0, nullptr, "RequestSuspend"}, + {1, nullptr, "RequestResume"}, + {2, nullptr, "GetProcessMasterVolume"}, + {3, nullptr, "SetProcessMasterVolume"}, + {4, nullptr, "GetProcessRecordVolume"}, + {5, nullptr, "SetProcessRecordVolume"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp index 5ed9cb20e..5f51fca9a 100644 --- a/src/core/hle/service/audio/audout_u.cpp +++ b/src/core/hle/service/audio/audout_u.cpp @@ -49,11 +49,11 @@ public: // clang-format off static const FunctionInfo functions[] = { {0, &IAudioOut::GetAudioOutState, "GetAudioOutState"}, - {1, &IAudioOut::StartAudioOut, "StartAudioOut"}, - {2, &IAudioOut::StopAudioOut, "StopAudioOut"}, + {1, &IAudioOut::StartAudioOut, "Start"}, + {2, &IAudioOut::StopAudioOut, "Stop"}, {3, &IAudioOut::AppendAudioOutBufferImpl, "AppendAudioOutBuffer"}, {4, &IAudioOut::RegisterBufferEvent, "RegisterBufferEvent"}, - {5, &IAudioOut::GetReleasedAudioOutBufferImpl, "GetReleasedAudioOutBuffer"}, + {5, &IAudioOut::GetReleasedAudioOutBufferImpl, "GetReleasedAudioOutBuffers"}, {6, &IAudioOut::ContainsAudioOutBuffer, "ContainsAudioOutBuffer"}, {7, &IAudioOut::AppendAudioOutBufferImpl, "AppendAudioOutBufferAuto"}, {8, &IAudioOut::GetReleasedAudioOutBufferImpl, "GetReleasedAudioOutBufferAuto"}, diff --git a/src/core/hle/service/audio/audrec_a.cpp b/src/core/hle/service/audio/audrec_a.cpp index c5ab7cad4..70fc17ae2 100644 --- a/src/core/hle/service/audio/audrec_a.cpp +++ b/src/core/hle/service/audio/audrec_a.cpp @@ -9,8 +9,8 @@ namespace Service::Audio { AudRecA::AudRecA(Core::System& system_) : ServiceFramework{system_, "audrec:a"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "RequestSuspendFinalOutputRecorders"}, - {1, nullptr, "RequestResumeFinalOutputRecorders"}, + {0, nullptr, "RequestSuspend"}, + {1, nullptr, "RequestResume"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audrec_u.cpp b/src/core/hle/service/audio/audrec_u.cpp index eb5c63c62..74a65ccff 100644 --- a/src/core/hle/service/audio/audrec_u.cpp +++ b/src/core/hle/service/audio/audrec_u.cpp @@ -13,16 +13,17 @@ public: // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetFinalOutputRecorderState"}, - {1, nullptr, "StartFinalOutputRecorder"}, - {2, nullptr, "StopFinalOutputRecorder"}, + {1, nullptr, "Start"}, + {2, nullptr, "Stop"}, {3, nullptr, "AppendFinalOutputRecorderBuffer"}, {4, nullptr, "RegisterBufferEvent"}, - {5, nullptr, "GetReleasedFinalOutputRecorderBuffer"}, + {5, nullptr, "GetReleasedFinalOutputRecorderBuffers"}, {6, nullptr, "ContainsFinalOutputRecorderBuffer"}, {7, nullptr, "GetFinalOutputRecorderBufferEndTime"}, {8, nullptr, "AppendFinalOutputRecorderBufferAuto"}, {9, nullptr, "GetReleasedFinalOutputRecorderBufferAuto"}, {10, nullptr, "FlushFinalOutputRecorderBuffers"}, + {11, nullptr, "AttachWorkBuffer"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audren_a.cpp b/src/core/hle/service/audio/audren_a.cpp index 5e9f866f0..cf8c34a15 100644 --- a/src/core/hle/service/audio/audren_a.cpp +++ b/src/core/hle/service/audio/audren_a.cpp @@ -9,14 +9,14 @@ namespace Service::Audio { AudRenA::AudRenA(Core::System& system_) : ServiceFramework{system_, "audren:a"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "RequestSuspendAudioRenderers"}, - {1, nullptr, "RequestResumeAudioRenderers"}, - {2, nullptr, "GetAudioRenderersProcessMasterVolume"}, - {3, nullptr, "SetAudioRenderersProcessMasterVolume"}, + {0, nullptr, "RequestSuspend"}, + {1, nullptr, "RequestResume"}, + {2, nullptr, "GetProcessMasterVolume"}, + {3, nullptr, "SetProcessMasterVolume"}, {4, nullptr, "RegisterAppletResourceUserId"}, {5, nullptr, "UnregisterAppletResourceUserId"}, - {6, nullptr, "GetAudioRenderersProcessRecordVolume"}, - {7, nullptr, "SetAudioRenderersProcessRecordVolume"}, + {6, nullptr, "GetProcessRecordVolume"}, + {7, nullptr, "SetProcessRecordVolume"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp index b2b2ffc5a..572be8e00 100644 --- a/src/core/hle/service/audio/audren_u.cpp +++ b/src/core/hle/service/audio/audren_u.cpp @@ -332,9 +332,9 @@ AudRenU::AudRenU(Core::System& system_) : ServiceFramework{system_, "audren:u"} // clang-format off static const FunctionInfo functions[] = { {0, &AudRenU::OpenAudioRenderer, "OpenAudioRenderer"}, - {1, &AudRenU::GetAudioRendererWorkBufferSize, "GetAudioRendererWorkBufferSize"}, + {1, &AudRenU::GetAudioRendererWorkBufferSize, "GetWorkBufferSize"}, {2, &AudRenU::GetAudioDeviceService, "GetAudioDeviceService"}, - {3, &AudRenU::OpenAudioRendererAuto, "OpenAudioRendererAuto"}, + {3, &AudRenU::OpenAudioRendererForManualExecution, "OpenAudioRendererForManualExecution"}, {4, &AudRenU::GetAudioDeviceServiceWithRevisionInfo, "GetAudioDeviceServiceWithRevisionInfo"}, }; // clang-format on @@ -665,7 +665,7 @@ void AudRenU::GetAudioDeviceService(Kernel::HLERequestContext& ctx) { rb.PushIpcInterface<IAudioDevice>(system, Common::MakeMagic('R', 'E', 'V', '1')); } -void AudRenU::OpenAudioRendererAuto(Kernel::HLERequestContext& ctx) { +void AudRenU::OpenAudioRendererForManualExecution(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Audio, "called"); OpenAudioRendererImpl(ctx); diff --git a/src/core/hle/service/audio/audren_u.h b/src/core/hle/service/audio/audren_u.h index d693dc406..37e8b4716 100644 --- a/src/core/hle/service/audio/audren_u.h +++ b/src/core/hle/service/audio/audren_u.h @@ -25,7 +25,7 @@ private: void OpenAudioRenderer(Kernel::HLERequestContext& ctx); void GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx); void GetAudioDeviceService(Kernel::HLERequestContext& ctx); - void OpenAudioRendererAuto(Kernel::HLERequestContext& ctx); + void OpenAudioRendererForManualExecution(Kernel::HLERequestContext& ctx); void GetAudioDeviceServiceWithRevisionInfo(Kernel::HLERequestContext& ctx); void OpenAudioRendererImpl(Kernel::HLERequestContext& ctx); diff --git a/src/core/hle/service/audio/codecctl.cpp b/src/core/hle/service/audio/codecctl.cpp index 94afec1b6..42961d908 100644 --- a/src/core/hle/service/audio/codecctl.cpp +++ b/src/core/hle/service/audio/codecctl.cpp @@ -8,19 +8,19 @@ namespace Service::Audio { CodecCtl::CodecCtl(Core::System& system_) : ServiceFramework{system_, "codecctl"} { static const FunctionInfo functions[] = { - {0, nullptr, "InitializeCodecController"}, - {1, nullptr, "FinalizeCodecController"}, - {2, nullptr, "SleepCodecController"}, - {3, nullptr, "WakeCodecController"}, - {4, nullptr, "SetCodecVolume"}, - {5, nullptr, "GetCodecVolumeMax"}, - {6, nullptr, "GetCodecVolumeMin"}, - {7, nullptr, "SetCodecActiveTarget"}, - {8, nullptr, "GetCodecActiveTarget"}, - {9, nullptr, "BindCodecHeadphoneMicJackInterrupt"}, - {10, nullptr, "IsCodecHeadphoneMicJackInserted"}, - {11, nullptr, "ClearCodecHeadphoneMicJackInterrupt"}, - {12, nullptr, "IsCodecDeviceRequested"}, + {0, nullptr, "Initialize"}, + {1, nullptr, "Finalize"}, + {2, nullptr, "Sleep"}, + {3, nullptr, "Wake"}, + {4, nullptr, "SetVolume"}, + {5, nullptr, "GetVolumeMax"}, + {6, nullptr, "GetVolumeMin"}, + {7, nullptr, "SetActiveTarget"}, + {8, nullptr, "GetActiveTarget"}, + {9, nullptr, "BindHeadphoneMicJackInterrupt"}, + {10, nullptr, "IsHeadphoneMicJackInserted"}, + {11, nullptr, "ClearHeadphoneMicJackInterrupt"}, + {12, nullptr, "IsRequested"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/audio/hwopus.cpp b/src/core/hle/service/audio/hwopus.cpp index ea3414fd2..19c578b3a 100644 --- a/src/core/hle/service/audio/hwopus.cpp +++ b/src/core/hle/service/audio/hwopus.cpp @@ -297,6 +297,10 @@ HwOpus::HwOpus(Core::System& system_) : ServiceFramework{system_, "hwopus"} { {1, &HwOpus::GetWorkBufferSize, "GetWorkBufferSize"}, {2, nullptr, "OpenOpusDecoderForMultiStream"}, {3, nullptr, "GetWorkBufferSizeForMultiStream"}, + {4, nullptr, "OpenHardwareOpusDecoderEx"}, + {5, nullptr, "GetWorkBufferSizeEx"}, + {6, nullptr, "OpenHardwareOpusDecoderForMultiStreamEx"}, + {7, nullptr, "GetWorkBufferSizeForMultiStreamEx"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/bcat/module.cpp b/src/core/hle/service/bcat/module.cpp index 503109fdd..b68e2c345 100644 --- a/src/core/hle/service/bcat/module.cpp +++ b/src/core/hle/service/bcat/module.cpp @@ -155,10 +155,12 @@ public: {30210, nullptr, "SetDeliveryTaskTimer"}, {30300, nullptr, "RegisterSystemApplicationDeliveryTasks"}, {90100, nullptr, "EnumerateBackgroundDeliveryTask"}, + {90101, nullptr, "Unknown90101"}, {90200, nullptr, "GetDeliveryList"}, {90201, &IBcatService::ClearDeliveryCacheStorage, "ClearDeliveryCacheStorage"}, {90202, nullptr, "ClearDeliveryTaskSubscriptionStatus"}, {90300, nullptr, "GetPushNotificationLog"}, + {90301, nullptr, "Unknown90301"}, }; // clang-format on RegisterHandlers(functions); diff --git a/src/core/hle/service/bpc/bpc.cpp b/src/core/hle/service/bpc/bpc.cpp index e4630320e..78e01d8d8 100644 --- a/src/core/hle/service/bpc/bpc.cpp +++ b/src/core/hle/service/bpc/bpc.cpp @@ -29,8 +29,8 @@ public: {11, nullptr, "CreateWakeupTimerEx"}, {12, nullptr, "GetLastEnabledWakeupTimerType"}, {13, nullptr, "CleanAllWakeupTimers"}, - {14, nullptr, "Unknown"}, - {15, nullptr, "Unknown2"}, + {14, nullptr, "GetPowerButton"}, + {15, nullptr, "SetEnableWakeupTimer"}, }; // clang-format on diff --git a/src/core/hle/service/btdrv/btdrv.cpp b/src/core/hle/service/btdrv/btdrv.cpp index 17a2ac899..af3a5842d 100644 --- a/src/core/hle/service/btdrv/btdrv.cpp +++ b/src/core/hle/service/btdrv/btdrv.cpp @@ -156,6 +156,25 @@ public: {97, nullptr, "RegisterBleHidEvent"}, {98, nullptr, "SetBleScanParameter"}, {99, nullptr, "MoveToSecondaryPiconet"}, + {100, nullptr, "IsBluetoothEnabled"}, + {128, nullptr, "AcquireAudioEvent"}, + {129, nullptr, "GetAudioEventInfo"}, + {130, nullptr, "OpenAudioConnection"}, + {131, nullptr, "CloseAudioConnection"}, + {132, nullptr, "OpenAudioOut"}, + {133, nullptr, "CloseAudioOut"}, + {134, nullptr, "AcquireAudioOutStateChangedEvent"}, + {135, nullptr, "StartAudioOut"}, + {136, nullptr, "StopAudioOut"}, + {137, nullptr, "GetAudioOutState"}, + {138, nullptr, "GetAudioOutFeedingCodec"}, + {139, nullptr, "GetAudioOutFeedingParameter"}, + {140, nullptr, "AcquireAudioOutBufferAvailableEvent"}, + {141, nullptr, "SendAudioData"}, + {142, nullptr, "AcquireAudioControlInputStateChangedEvent"}, + {143, nullptr, "GetAudioControlInputState"}, + {144, nullptr, "AcquireAudioConnectionStateChangedEvent"}, + {145, nullptr, "GetConnectedAudioDevice"}, {256, nullptr, "IsManufacturingMode"}, {257, nullptr, "EmulateBluetoothCrash"}, {258, nullptr, "GetBleChannelMap"}, diff --git a/src/core/hle/service/btm/btm.cpp b/src/core/hle/service/btm/btm.cpp index 9cf2ee92a..d1ebc2388 100644 --- a/src/core/hle/service/btm/btm.cpp +++ b/src/core/hle/service/btm/btm.cpp @@ -223,6 +223,7 @@ public: {10, nullptr, "GetGattClientDisconnectionReason"}, {11, nullptr, "GetBleConnectionParameter"}, {12, nullptr, "GetBleConnectionParameterRequest"}, + {13, nullptr, "Unknown13"}, }; // clang-format on diff --git a/src/core/hle/service/caps/caps_a.cpp b/src/core/hle/service/caps/caps_a.cpp index 1fe4f0e14..6220e9f77 100644 --- a/src/core/hle/service/caps/caps_a.cpp +++ b/src/core/hle/service/caps/caps_a.cpp @@ -49,6 +49,7 @@ CAPS_A::CAPS_A(Core::System& system_) : ServiceFramework{system_, "caps:a"} { {16, nullptr, "GetAlbumMountResult"}, {17, nullptr, "GetAlbumUsage16"}, {18, nullptr, "Unknown18"}, + {19, nullptr, "Unknown19"}, {100, nullptr, "GetAlbumFileCountEx0"}, {101, nullptr, "GetAlbumFileListEx0"}, {202, nullptr, "SaveEditedScreenShot"}, diff --git a/src/core/hle/service/caps/caps_u.cpp b/src/core/hle/service/caps/caps_u.cpp index 842316a2e..10b8d54b1 100644 --- a/src/core/hle/service/caps/caps_u.cpp +++ b/src/core/hle/service/caps/caps_u.cpp @@ -43,6 +43,7 @@ CAPS_U::CAPS_U(Core::System& system_) : ServiceFramework{system_, "caps:u"} { {141, nullptr, "GetAlbumFileList2AafeUidAruidDeprecated"}, {142, &CAPS_U::GetAlbumFileList3AaeAruid, "GetAlbumFileList3AaeAruid"}, {143, nullptr, "GetAlbumFileList4AaeUidAruid"}, + {144, nullptr, "GetAllAlbumFileList3AaeAruid"}, {60002, nullptr, "OpenAccessorSessionForApplication"}, }; // clang-format on diff --git a/src/core/hle/service/erpt/erpt.cpp b/src/core/hle/service/erpt/erpt.cpp index 4924c61c3..c767926a4 100644 --- a/src/core/hle/service/erpt/erpt.cpp +++ b/src/core/hle/service/erpt/erpt.cpp @@ -16,7 +16,7 @@ public: // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SubmitContext"}, - {1, nullptr, "CreateReport"}, + {1, nullptr, "CreateReportV0"}, {2, nullptr, "SetInitialLaunchSettingsCompletionTime"}, {3, nullptr, "ClearInitialLaunchSettingsCompletionTime"}, {4, nullptr, "UpdatePowerOnTime"}, @@ -26,6 +26,11 @@ public: {8, nullptr, "ClearApplicationLaunchTime"}, {9, nullptr, "SubmitAttachment"}, {10, nullptr, "CreateReportWithAttachments"}, + {11, nullptr, "CreateReport"}, + {20, nullptr, "RegisterRunningApplet"}, + {21, nullptr, "UnregisterRunningApplet"}, + {22, nullptr, "UpdateAppletSuspendedDuration"}, + {30, nullptr, "InvalidateForcedShutdownDetection"}, }; // clang-format on diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp index 9cc260515..a0215c4d7 100644 --- a/src/core/hle/service/filesystem/fsp_srv.cpp +++ b/src/core/hle/service/filesystem/fsp_srv.cpp @@ -118,9 +118,13 @@ public: explicit IFile(Core::System& system_, FileSys::VirtualFile backend_) : ServiceFramework{system_, "IFile"}, backend(std::move(backend_)) { static const FunctionInfo functions[] = { - {0, &IFile::Read, "Read"}, {1, &IFile::Write, "Write"}, - {2, &IFile::Flush, "Flush"}, {3, &IFile::SetSize, "SetSize"}, - {4, &IFile::GetSize, "GetSize"}, {5, nullptr, "OperateRange"}, + {0, &IFile::Read, "Read"}, + {1, &IFile::Write, "Write"}, + {2, &IFile::Flush, "Flush"}, + {3, &IFile::SetSize, "SetSize"}, + {4, &IFile::GetSize, "GetSize"}, + {5, nullptr, "OperateRange"}, + {6, nullptr, "OperateRangeWithBuffer"}, }; RegisterHandlers(functions); } @@ -708,7 +712,10 @@ FSP_SRV::FSP_SRV(Core::System& system_) {84, nullptr, "ListApplicationAccessibleSaveDataOwnerId"}, {85, nullptr, "OpenSaveDataTransferManagerForSaveDataRepair"}, {86, nullptr, "OpenSaveDataMover"}, + {87, nullptr, "OpenSaveDataTransferManagerForRepair"}, {100, nullptr, "OpenImageDirectoryFileSystem"}, + {101, nullptr, "OpenBaseFileSystem"}, + {102, nullptr, "FormatBaseFileSystem"}, {110, nullptr, "OpenContentStorageFileSystem"}, {120, nullptr, "OpenCloudBackupWorkStorageFileSystem"}, {130, nullptr, "OpenCustomStorageFileSystem"}, @@ -764,10 +771,12 @@ FSP_SRV::FSP_SRV(Core::System& system_) {1008, nullptr, "OpenRegisteredUpdatePartition"}, {1009, nullptr, "GetAndClearMemoryReportInfo"}, {1010, nullptr, "SetDataStorageRedirectTarget"}, - {1011, &FSP_SRV::GetAccessLogVersionInfo, "GetAccessLogVersionInfo"}, + {1011, &FSP_SRV::GetProgramIndexForAccessLog, "GetProgramIndexForAccessLog"}, {1012, nullptr, "GetFsStackUsage"}, {1013, nullptr, "UnsetSaveDataRootPath"}, {1014, nullptr, "OutputMultiProgramTagAccessLog"}, + {1016, nullptr, "FlushAccessLogOnSdCard"}, + {1017, nullptr, "OutputApplicationInfoAccessLog"}, {1100, nullptr, "OverrideSaveDataTransferTokenSignVerificationKey"}, {1110, nullptr, "CorruptSaveDataFileSystemBySaveDataSpaceId2"}, {1200, &FSP_SRV::OpenMultiCommitManager, "OpenMultiCommitManager"}, @@ -1051,7 +1060,7 @@ void FSP_SRV::OutputAccessLogToSdCard(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); } -void FSP_SRV::GetAccessLogVersionInfo(Kernel::HLERequestContext& ctx) { +void FSP_SRV::GetProgramIndexForAccessLog(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_FS, "called"); IPC::ResponseBuilder rb{ctx, 4}; diff --git a/src/core/hle/service/filesystem/fsp_srv.h b/src/core/hle/service/filesystem/fsp_srv.h index 8ed933279..b01b924eb 100644 --- a/src/core/hle/service/filesystem/fsp_srv.h +++ b/src/core/hle/service/filesystem/fsp_srv.h @@ -53,7 +53,7 @@ private: void SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx); void GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx); void OutputAccessLogToSdCard(Kernel::HLERequestContext& ctx); - void GetAccessLogVersionInfo(Kernel::HLERequestContext& ctx); + void GetProgramIndexForAccessLog(Kernel::HLERequestContext& ctx); void OpenMultiCommitManager(Kernel::HLERequestContext& ctx); FileSystemController& fsc; diff --git a/src/core/hle/service/friend/friend.cpp b/src/core/hle/service/friend/friend.cpp index 72a877d68..a35979053 100644 --- a/src/core/hle/service/friend/friend.cpp +++ b/src/core/hle/service/friend/friend.cpp @@ -38,7 +38,7 @@ public: {10600, nullptr, "DeclareOpenOnlinePlaySession"}, {10601, &IFriendService::DeclareCloseOnlinePlaySession, "DeclareCloseOnlinePlaySession"}, {10610, &IFriendService::UpdateUserPresence, "UpdateUserPresence"}, - {10700, nullptr, "GetPlayHistoryRegistrationKey"}, + {10700, &IFriendService::GetPlayHistoryRegistrationKey, "GetPlayHistoryRegistrationKey"}, {10701, nullptr, "GetPlayHistoryRegistrationKeyWithNetworkServiceAccountId"}, {10702, nullptr, "AddPlayHistory"}, {11000, nullptr, "GetProfileImageUrl"}, @@ -133,7 +133,7 @@ private: void GetBlockedUserListIds(Kernel::HLERequestContext& ctx) { // This is safe to stub, as there should be no adverse consequences from reporting no // blocked users. - LOG_WARNING(Service_ACC, "(STUBBED) called"); + LOG_WARNING(Service_Friend, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); rb.Push<u32>(0); // Indicates there are no blocked users @@ -141,14 +141,26 @@ private: void DeclareCloseOnlinePlaySession(Kernel::HLERequestContext& ctx) { // Stub used by Splatoon 2 - LOG_WARNING(Service_ACC, "(STUBBED) called"); + LOG_WARNING(Service_Friend, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } void UpdateUserPresence(Kernel::HLERequestContext& ctx) { // Stub used by Retro City Rampage - LOG_WARNING(Service_ACC, "(STUBBED) called"); + LOG_WARNING(Service_Friend, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + void GetPlayHistoryRegistrationKey(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto local_play = rp.Pop<bool>(); + const auto uuid = rp.PopRaw<Common::UUID>(); + + LOG_WARNING(Service_Friend, "(STUBBED) called local_play={} uuid={}", local_play, + uuid.Format()); + IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -159,7 +171,7 @@ private: const auto uuid = rp.PopRaw<Common::UUID>(); [[maybe_unused]] const auto filter = rp.PopRaw<SizedFriendFilter>(); const auto pid = rp.Pop<u64>(); - LOG_WARNING(Service_ACC, "(STUBBED) called, offset={}, uuid={}, pid={}", friend_offset, + LOG_WARNING(Service_Friend, "(STUBBED) called, offset={}, uuid={}, pid={}", friend_offset, uuid.Format(), pid); IPC::ResponseBuilder rb{ctx, 3}; @@ -191,7 +203,7 @@ public: private: void GetEvent(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_ACC, "called"); + LOG_DEBUG(Service_Friend, "called"); IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); @@ -199,7 +211,7 @@ private: } void Clear(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_ACC, "called"); + LOG_DEBUG(Service_Friend, "called"); while (!notifications.empty()) { notifications.pop(); } @@ -210,10 +222,10 @@ private: } void Pop(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_ACC, "called"); + LOG_DEBUG(Service_Friend, "called"); if (notifications.empty()) { - LOG_ERROR(Service_ACC, "No notifications in queue!"); + LOG_ERROR(Service_Friend, "No notifications in queue!"); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_NO_NOTIFICATIONS); return; @@ -231,7 +243,8 @@ private: break; default: // HOS seems not have an error case for an unknown notification - LOG_WARNING(Service_ACC, "Unknown notification {:08X}", notification.notification_type); + LOG_WARNING(Service_Friend, "Unknown notification {:08X}", + notification.notification_type); break; } @@ -269,14 +282,14 @@ void Module::Interface::CreateFriendService(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); rb.PushIpcInterface<IFriendService>(system); - LOG_DEBUG(Service_ACC, "called"); + LOG_DEBUG(Service_Friend, "called"); } void Module::Interface::CreateNotificationService(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; auto uuid = rp.PopRaw<Common::UUID>(); - LOG_DEBUG(Service_ACC, "called, uuid={}", uuid.Format()); + LOG_DEBUG(Service_Friend, "called, uuid={}", uuid.Format()); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); diff --git a/src/core/hle/service/glue/arp.cpp b/src/core/hle/service/glue/arp.cpp index fc77e7286..322125135 100644 --- a/src/core/hle/service/glue/arp.cpp +++ b/src/core/hle/service/glue/arp.cpp @@ -41,6 +41,12 @@ ARP_R::ARP_R(Core::System& system_, const ARPManager& manager_) {1, &ARP_R::GetApplicationLaunchPropertyWithApplicationId, "GetApplicationLaunchPropertyWithApplicationId"}, {2, &ARP_R::GetApplicationControlProperty, "GetApplicationControlProperty"}, {3, &ARP_R::GetApplicationControlPropertyWithApplicationId, "GetApplicationControlPropertyWithApplicationId"}, + {4, nullptr, "GetApplicationInstanceUnregistrationNotifier"}, + {5, nullptr, "ListApplicationInstanceId"}, + {6, nullptr, "GetMicroApplicationInstanceId"}, + {7, nullptr, "GetApplicationCertificate"}, + {9998, nullptr, "GetPreomiaApplicationLaunchProperty"}, + {9999, nullptr, "GetPreomiaApplicationControlProperty"}, }; // clang-format on @@ -243,7 +249,8 @@ ARP_W::ARP_W(Core::System& system_, ARPManager& manager_) // clang-format off static const FunctionInfo functions[] = { {0, &ARP_W::AcquireRegistrar, "AcquireRegistrar"}, - {1, &ARP_W::DeleteProperties, "DeleteProperties"}, + {1, &ARP_W::UnregisterApplicationInstance , "UnregisterApplicationInstance "}, + {2, nullptr, "AcquireUpdater"}, }; // clang-format on @@ -270,7 +277,7 @@ void ARP_W::AcquireRegistrar(Kernel::HLERequestContext& ctx) { rb.PushIpcInterface(registrar); } -void ARP_W::DeleteProperties(Kernel::HLERequestContext& ctx) { +void ARP_W::UnregisterApplicationInstance(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto process_id = rp.PopRaw<u64>(); diff --git a/src/core/hle/service/glue/arp.h b/src/core/hle/service/glue/arp.h index 34b412e26..0df3c5e1f 100644 --- a/src/core/hle/service/glue/arp.h +++ b/src/core/hle/service/glue/arp.h @@ -32,7 +32,7 @@ public: private: void AcquireRegistrar(Kernel::HLERequestContext& ctx); - void DeleteProperties(Kernel::HLERequestContext& ctx); + void UnregisterApplicationInstance(Kernel::HLERequestContext& ctx); ARPManager& manager; std::shared_ptr<IRegistrar> registrar; diff --git a/src/core/hle/service/glue/bgtc.cpp b/src/core/hle/service/glue/bgtc.cpp index a478b68e1..daecfff15 100644 --- a/src/core/hle/service/glue/bgtc.cpp +++ b/src/core/hle/service/glue/bgtc.cpp @@ -2,6 +2,9 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/logging/log.h" +#include "core/core.h" +#include "core/hle/ipc_helpers.h" #include "core/hle/service/glue/bgtc.h" namespace Service::Glue { @@ -9,6 +12,26 @@ namespace Service::Glue { BGTC_T::BGTC_T(Core::System& system_) : ServiceFramework{system_, "bgtc:t"} { // clang-format off static const FunctionInfo functions[] = { + {100, &BGTC_T::OpenTaskService, "OpenTaskService"}, + }; + // clang-format on + + RegisterHandlers(functions); +} + +BGTC_T::~BGTC_T() = default; + +void BGTC_T::OpenTaskService(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_BGTC, "called"); + + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<ITaskService>(system); +} + +ITaskService::ITaskService(Core::System& system_) : ServiceFramework{system_, "ITaskService"} { + // clang-format off + static const FunctionInfo functions[] = { {1, nullptr, "NotifyTaskStarting"}, {2, nullptr, "NotifyTaskFinished"}, {3, nullptr, "GetTriggerEvent"}, @@ -20,16 +43,18 @@ BGTC_T::BGTC_T(Core::System& system_) : ServiceFramework{system_, "bgtc:t"} { {13, nullptr, "UnscheduleTask"}, {14, nullptr, "GetScheduleEvent"}, {15, nullptr, "SchedulePeriodicTask"}, + {16, nullptr, "Unknown16"}, {101, nullptr, "GetOperationMode"}, {102, nullptr, "WillDisconnectNetworkWhenEnteringSleep"}, {103, nullptr, "WillStayHalfAwakeInsteadSleep"}, + {200, nullptr, "Unknown200"}, }; // clang-format on RegisterHandlers(functions); } -BGTC_T::~BGTC_T() = default; +ITaskService::~ITaskService() = default; BGTC_SC::BGTC_SC(Core::System& system_) : ServiceFramework{system_, "bgtc:sc"} { // clang-format off diff --git a/src/core/hle/service/glue/bgtc.h b/src/core/hle/service/glue/bgtc.h index 906116ba6..4c0142fd5 100644 --- a/src/core/hle/service/glue/bgtc.h +++ b/src/core/hle/service/glue/bgtc.h @@ -16,6 +16,14 @@ class BGTC_T final : public ServiceFramework<BGTC_T> { public: explicit BGTC_T(Core::System& system_); ~BGTC_T() override; + + void OpenTaskService(Kernel::HLERequestContext& ctx); +}; + +class ITaskService final : public ServiceFramework<ITaskService> { +public: + explicit ITaskService(Core::System& system_); + ~ITaskService() override; }; class BGTC_SC final : public ServiceFramework<BGTC_SC> { diff --git a/src/core/hle/service/hid/controllers/gesture.cpp b/src/core/hle/service/hid/controllers/gesture.cpp index e7063f8ef..93c43a203 100644 --- a/src/core/hle/service/hid/controllers/gesture.cpp +++ b/src/core/hle/service/hid/controllers/gesture.cpp @@ -4,6 +4,7 @@ #include <cstring> #include "common/common_types.h" +#include "common/logging/log.h" #include "core/core_timing.h" #include "core/frontend/emu_window.h" #include "core/hle/service/hid/controllers/gesture.h" @@ -19,9 +20,9 @@ Controller_Gesture::~Controller_Gesture() = default; void Controller_Gesture::OnInit() { for (std::size_t id = 0; id < MAX_FINGERS; ++id) { - mouse_finger_id[id] = MAX_FINGERS; - keyboard_finger_id[id] = MAX_FINGERS; - udp_finger_id[id] = MAX_FINGERS; + mouse_finger_id[id] = MAX_POINTS; + keyboard_finger_id[id] = MAX_POINTS; + udp_finger_id[id] = MAX_POINTS; } } @@ -142,6 +143,10 @@ std::optional<std::size_t> Controller_Gesture::GetUnusedFingerID() const { std::size_t Controller_Gesture::UpdateTouchInputEvent( const std::tuple<float, float, bool>& touch_input, std::size_t finger_id) { const auto& [x, y, pressed] = touch_input; + if (finger_id > MAX_POINTS) { + LOG_ERROR(Service_HID, "Invalid finger id {}", finger_id); + return MAX_POINTS; + } if (pressed) { if (finger_id == MAX_POINTS) { const auto first_free_id = GetUnusedFingerID(); diff --git a/src/core/hle/service/hid/controllers/npad.cpp b/src/core/hle/service/hid/controllers/npad.cpp index 70b9f3824..673db68c7 100644 --- a/src/core/hle/service/hid/controllers/npad.cpp +++ b/src/core/hle/service/hid/controllers/npad.cpp @@ -413,12 +413,16 @@ void Controller_NPad::RequestPadStateUpdate(u32 npad_id) { lstick_entry.y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX); } - if (controller_type == NPadControllerType::JoyLeft || - controller_type == NPadControllerType::JoyRight) { + if (controller_type == NPadControllerType::JoyLeft) { pad_state.left_sl.Assign(button_state[SL - BUTTON_HID_BEGIN]->GetStatus()); pad_state.left_sr.Assign(button_state[SR - BUTTON_HID_BEGIN]->GetStatus()); } + if (controller_type == NPadControllerType::JoyRight) { + pad_state.right_sl.Assign(button_state[SL - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.right_sr.Assign(button_state[SR - BUTTON_HID_BEGIN]->GetStatus()); + } + if (controller_type == NPadControllerType::GameCube) { trigger_entry.l_analog = static_cast<s32>( button_state[ZL - BUTTON_HID_BEGIN]->GetStatus() ? HID_TRIGGER_MAX : 0); @@ -1134,6 +1138,10 @@ void Controller_NPad::SetUnintendedHomeButtonInputProtectionEnabled(bool is_prot unintended_home_button_input_protection[NPadIdToIndex(npad_id)] = is_protection_enabled; } +void Controller_NPad::SetAnalogStickUseCenterClamp(bool use_center_clamp) { + analog_stick_use_center_clamp = use_center_clamp; +} + void Controller_NPad::ClearAllConnectedControllers() { for (auto& controller : connected_controllers) { if (controller.is_connected && controller.type != NPadControllerType::None) { diff --git a/src/core/hle/service/hid/controllers/npad.h b/src/core/hle/service/hid/controllers/npad.h index bc2e6779d..873a0a1e2 100644 --- a/src/core/hle/service/hid/controllers/npad.h +++ b/src/core/hle/service/hid/controllers/npad.h @@ -219,6 +219,7 @@ public: LedPattern GetLedPattern(u32 npad_id); bool IsUnintendedHomeButtonInputProtectionEnabled(u32 npad_id) const; void SetUnintendedHomeButtonInputProtectionEnabled(bool is_protection_enabled, u32 npad_id); + void SetAnalogStickUseCenterClamp(bool use_center_clamp); void ClearAllConnectedControllers(); void DisconnectAllConnectedControllers(); void ConnectAllDisconnectedControllers(); @@ -577,6 +578,7 @@ private: std::array<std::array<bool, 2>, 10> vibration_devices_mounted{}; std::array<ControllerHolder, 10> connected_controllers{}; std::array<bool, 10> unintended_home_button_input_protection{}; + bool analog_stick_use_center_clamp{}; GyroscopeZeroDriftMode gyroscope_zero_drift_mode{GyroscopeZeroDriftMode::Standard}; bool sixaxis_sensors_enabled{true}; f32 sixaxis_fusion_parameter1{}; diff --git a/src/core/hle/service/hid/controllers/touchscreen.cpp b/src/core/hle/service/hid/controllers/touchscreen.cpp index 5219f2dad..be60492a4 100644 --- a/src/core/hle/service/hid/controllers/touchscreen.cpp +++ b/src/core/hle/service/hid/controllers/touchscreen.cpp @@ -5,6 +5,7 @@ #include <algorithm> #include <cstring> #include "common/common_types.h" +#include "common/logging/log.h" #include "core/core_timing.h" #include "core/frontend/emu_window.h" #include "core/frontend/input.h" @@ -118,6 +119,10 @@ std::optional<std::size_t> Controller_Touchscreen::GetUnusedFingerID() const { std::size_t Controller_Touchscreen::UpdateTouchInputEvent( const std::tuple<float, float, bool>& touch_input, std::size_t finger_id) { const auto& [x, y, pressed] = touch_input; + if (finger_id > MAX_FINGERS) { + LOG_ERROR(Service_HID, "Invalid finger id {}", finger_id); + return MAX_FINGERS; + } if (pressed) { Attributes attribute{}; if (finger_id == MAX_FINGERS) { diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp index ba27bbb05..a1a779cc0 100644 --- a/src/core/hle/service/hid/hid.cpp +++ b/src/core/hle/service/hid/hid.cpp @@ -263,7 +263,7 @@ Hid::Hid(Core::System& system_) : ServiceFramework{system_, "hid"} { {131, &Hid::IsUnintendedHomeButtonInputProtectionEnabled, "IsUnintendedHomeButtonInputProtectionEnabled"}, {132, &Hid::EnableUnintendedHomeButtonInputProtection, "EnableUnintendedHomeButtonInputProtection"}, {133, nullptr, "SetNpadJoyAssignmentModeSingleWithDestination"}, - {134, nullptr, "SetNpadAnalogStickUseCenterClamp"}, + {134, &Hid::SetNpadAnalogStickUseCenterClamp, "SetNpadAnalogStickUseCenterClamp"}, {135, nullptr, "SetNpadCaptureButtonAssignment"}, {136, nullptr, "ClearNpadCaptureButtonAssignment"}, {200, &Hid::GetVibrationDeviceInfo, "GetVibrationDeviceInfo"}, @@ -278,6 +278,7 @@ Hid::Hid(Core::System& system_) : ServiceFramework{system_, "hid"} { {209, &Hid::BeginPermitVibrationSession, "BeginPermitVibrationSession"}, {210, &Hid::EndPermitVibrationSession, "EndPermitVibrationSession"}, {211, &Hid::IsVibrationDeviceMounted, "IsVibrationDeviceMounted"}, + {212, nullptr, "SendVibrationValueInBool"}, {300, &Hid::ActivateConsoleSixAxisSensor, "ActivateConsoleSixAxisSensor"}, {301, &Hid::StartConsoleSixAxisSensor, "StartConsoleSixAxisSensor"}, {302, &Hid::StopConsoleSixAxisSensor, "StopConsoleSixAxisSensor"}, @@ -1087,6 +1088,27 @@ void Hid::EnableUnintendedHomeButtonInputProtection(Kernel::HLERequestContext& c rb.Push(RESULT_SUCCESS); } +void Hid::SetNpadAnalogStickUseCenterClamp(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + bool analog_stick_use_center_clamp; + u64 applet_resource_user_id; + }; + static_assert(sizeof(Parameters) == 0x10, "Parameters has incorrect size."); + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetAnalogStickUseCenterClamp(parameters.analog_stick_use_center_clamp); + + LOG_WARNING(Service_HID, + "(STUBBED) called, analog_stick_use_center_clamp={}, applet_resource_user_id={}", + parameters.analog_stick_use_center_clamp, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + void Hid::GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto vibration_device_handle{rp.PopRaw<Controller_NPad::DeviceHandle>()}; @@ -1553,6 +1575,7 @@ public: {11, nullptr, "SetTouchScreenAutoPilotState"}, {12, nullptr, "UnsetTouchScreenAutoPilotState"}, {13, nullptr, "GetTouchScreenConfiguration"}, + {14, nullptr, "ProcessTouchScreenAutoTune"}, {20, nullptr, "DeactivateMouse"}, {21, nullptr, "SetMouseAutoPilotState"}, {22, nullptr, "UnsetMouseAutoPilotState"}, @@ -1562,6 +1585,7 @@ public: {50, nullptr, "DeactivateXpad"}, {51, nullptr, "SetXpadAutoPilotState"}, {52, nullptr, "UnsetXpadAutoPilotState"}, + {53, nullptr, "DeactivateJoyXpad"}, {60, nullptr, "ClearNpadSystemCommonPolicy"}, {61, nullptr, "DeactivateNpad"}, {62, nullptr, "ForceDisconnectNpad"}, @@ -1632,6 +1656,11 @@ public: {244, nullptr, "RequestKuinaFirmwareVersion"}, {245, nullptr, "GetKuinaFirmwareVersion"}, {246, nullptr, "GetVidPid"}, + {247, nullptr, "GetAnalogStickCalibrationValue"}, + {248, nullptr, "GetUniquePadIdsFull"}, + {249, nullptr, "ConnectUniquePad"}, + {250, nullptr, "IsVirtual"}, + {251, nullptr, "GetAnalogStickModuleParam"}, {301, nullptr, "GetAbstractedPadHandles"}, {302, nullptr, "GetAbstractedPadState"}, {303, nullptr, "GetAbstractedPadsState"}, @@ -1652,12 +1681,16 @@ public: {401, nullptr, "DisableRailDeviceFiltering"}, {402, nullptr, "EnableWiredPairing"}, {403, nullptr, "EnableShipmentModeAutoClear"}, + {404, nullptr, "SetRailEnabled"}, {500, nullptr, "SetFactoryInt"}, {501, nullptr, "IsFactoryBootEnabled"}, {550, nullptr, "SetAnalogStickModelDataTemporarily"}, {551, nullptr, "GetAnalogStickModelData"}, {552, nullptr, "ResetAnalogStickModelData"}, {600, nullptr, "ConvertPadState"}, + {650, nullptr, "AddButtonPlayData"}, + {651, nullptr, "StartButtonPlayData"}, + {652, nullptr, "StopButtonPlayData"}, {2000, nullptr, "DeactivateDigitizer"}, {2001, nullptr, "SetDigitizerAutoPilotState"}, {2002, nullptr, "UnsetDigitizerAutoPilotState"}, @@ -1689,6 +1722,8 @@ public: {215, nullptr, "IsNfcActivated"}, {230, nullptr, "AcquireIrSensorEventHandle"}, {231, nullptr, "ActivateIrSensor"}, + {232, nullptr, "GetIrSensorState"}, + {233, nullptr, "GetXcdHandleForNpadWithIrSensor"}, {301, nullptr, "ActivateNpadSystem"}, {303, nullptr, "ApplyNpadSystemCommonPolicy"}, {304, nullptr, "EnableAssigningSingleOnSlSrPress"}, @@ -1703,9 +1738,16 @@ public: {313, nullptr, "GetNpadCaptureButtonAssignment"}, {314, nullptr, "GetAppletFooterUiType"}, {315, nullptr, "GetAppletDetailedUiType"}, + {316, nullptr, "GetNpadInterfaceType"}, + {317, nullptr, "GetNpadLeftRightInterfaceType"}, + {318, nullptr, "HasBattery"}, + {319, nullptr, "HasLeftRightBattery"}, {321, nullptr, "GetUniquePadsFromNpad"}, {322, nullptr, "GetIrSensorState"}, {323, nullptr, "GetXcdHandleForNpadWithIrSensor"}, + {324, nullptr, "GetUniquePadButtonSet"}, + {325, nullptr, "GetUniquePadColor"}, + {326, nullptr, "GetUniquePadAppletDetailedUiType"}, {500, nullptr, "SetAppletResourceUserId"}, {501, nullptr, "RegisterAppletResourceUserId"}, {502, nullptr, "UnregisterAppletResourceUserId"}, @@ -1716,10 +1758,13 @@ public: {511, nullptr, "GetVibrationMasterVolume"}, {512, nullptr, "BeginPermitVibrationSession"}, {513, nullptr, "EndPermitVibrationSession"}, + {514, nullptr, "Unknown514"}, {520, nullptr, "EnableHandheldHids"}, {521, nullptr, "DisableHandheldHids"}, {522, nullptr, "SetJoyConRailEnabled"}, {523, nullptr, "IsJoyConRailEnabled"}, + {524, nullptr, "IsHandheldHidsEnabled"}, + {525, nullptr, "IsJoyConAttachedOnAllRail"}, {540, nullptr, "AcquirePlayReportControllerUsageUpdateEvent"}, {541, nullptr, "GetPlayReportControllerUsages"}, {542, nullptr, "AcquirePlayReportRegisteredDeviceUpdateEvent"}, @@ -1795,6 +1840,65 @@ public: {1154, nullptr, "IsFirmwareAvailableForNotification"}, {1155, nullptr, "SetForceHandheldStyleVibration"}, {1156, nullptr, "SendConnectionTriggerWithoutTimeoutEvent"}, + {1157, nullptr, "CancelConnectionTrigger"}, + {1200, nullptr, "IsButtonConfigSupported"}, + {1201, nullptr, "IsButtonConfigEmbeddedSupported"}, + {1202, nullptr, "DeleteButtonConfig"}, + {1203, nullptr, "DeleteButtonConfigEmbedded"}, + {1204, nullptr, "SetButtonConfigEnabled"}, + {1205, nullptr, "SetButtonConfigEmbeddedEnabled"}, + {1206, nullptr, "IsButtonConfigEnabled"}, + {1207, nullptr, "IsButtonConfigEmbeddedEnabled"}, + {1208, nullptr, "SetButtonConfigEmbedded"}, + {1209, nullptr, "SetButtonConfigFull"}, + {1210, nullptr, "SetButtonConfigLeft"}, + {1211, nullptr, "SetButtonConfigRight"}, + {1212, nullptr, "GetButtonConfigEmbedded"}, + {1213, nullptr, "GetButtonConfigFull"}, + {1214, nullptr, "GetButtonConfigLeft"}, + {1215, nullptr, "GetButtonConfigRight"}, + {1250, nullptr, "IsCustomButtonConfigSupported"}, + {1251, nullptr, "IsDefaultButtonConfigEmbedded"}, + {1252, nullptr, "IsDefaultButtonConfigFull"}, + {1253, nullptr, "IsDefaultButtonConfigLeft"}, + {1254, nullptr, "IsDefaultButtonConfigRight"}, + {1255, nullptr, "IsButtonConfigStorageEmbeddedEmpty"}, + {1256, nullptr, "IsButtonConfigStorageFullEmpty"}, + {1257, nullptr, "IsButtonConfigStorageLeftEmpty"}, + {1258, nullptr, "IsButtonConfigStorageRightEmpty"}, + {1259, nullptr, "GetButtonConfigStorageEmbeddedDeprecated"}, + {1260, nullptr, "GetButtonConfigStorageFullDeprecated"}, + {1261, nullptr, "GetButtonConfigStorageLeftDeprecated"}, + {1262, nullptr, "GetButtonConfigStorageRightDeprecated"}, + {1263, nullptr, "SetButtonConfigStorageEmbeddedDeprecated"}, + {1264, nullptr, "SetButtonConfigStorageFullDeprecated"}, + {1265, nullptr, "SetButtonConfigStorageLeftDeprecated"}, + {1266, nullptr, "SetButtonConfigStorageRightDeprecated"}, + {1267, nullptr, "DeleteButtonConfigStorageEmbedded"}, + {1268, nullptr, "DeleteButtonConfigStorageFull"}, + {1269, nullptr, "DeleteButtonConfigStorageLeft"}, + {1270, nullptr, "DeleteButtonConfigStorageRight"}, + {1271, nullptr, "IsUsingCustomButtonConfig"}, + {1272, nullptr, "IsAnyCustomButtonConfigEnabled"}, + {1273, nullptr, "SetAllCustomButtonConfigEnabled"}, + {1274, nullptr, "SetDefaultButtonConfig"}, + {1275, nullptr, "SetAllDefaultButtonConfig"}, + {1276, nullptr, "SetHidButtonConfigEmbedded"}, + {1277, nullptr, "SetHidButtonConfigFull"}, + {1278, nullptr, "SetHidButtonConfigLeft"}, + {1279, nullptr, "SetHidButtonConfigRight"}, + {1280, nullptr, "GetHidButtonConfigEmbedded"}, + {1281, nullptr, "GetHidButtonConfigFull"}, + {1282, nullptr, "GetHidButtonConfigLeft"}, + {1283, nullptr, "GetHidButtonConfigRight"}, + {1284, nullptr, "GetButtonConfigStorageEmbedded"}, + {1285, nullptr, "GetButtonConfigStorageFull"}, + {1286, nullptr, "GetButtonConfigStorageLeft"}, + {1287, nullptr, "GetButtonConfigStorageRight"}, + {1288, nullptr, "SetButtonConfigStorageEmbedded"}, + {1289, nullptr, "SetButtonConfigStorageFull"}, + {1290, nullptr, "DeleteButtonConfigStorageRight"}, + {1291, nullptr, "DeleteButtonConfigStorageRight"}, }; // clang-format on diff --git a/src/core/hle/service/hid/hid.h b/src/core/hle/service/hid/hid.h index 36ed228c8..c2bdd39a3 100644 --- a/src/core/hle/service/hid/hid.h +++ b/src/core/hle/service/hid/hid.h @@ -129,6 +129,7 @@ private: void SwapNpadAssignment(Kernel::HLERequestContext& ctx); void IsUnintendedHomeButtonInputProtectionEnabled(Kernel::HLERequestContext& ctx); void EnableUnintendedHomeButtonInputProtection(Kernel::HLERequestContext& ctx); + void SetNpadAnalogStickUseCenterClamp(Kernel::HLERequestContext& ctx); void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx); void SendVibrationValue(Kernel::HLERequestContext& ctx); void GetActualVibrationValue(Kernel::HLERequestContext& ctx); diff --git a/src/core/hle/service/hid/xcd.cpp b/src/core/hle/service/hid/xcd.cpp index 43a8840d0..b1efa3d05 100644 --- a/src/core/hle/service/hid/xcd.cpp +++ b/src/core/hle/service/hid/xcd.cpp @@ -28,6 +28,8 @@ XCD_SYS::XCD_SYS(Core::System& system_) : ServiceFramework{system_, "xcd:sys"} { {20, nullptr, "StartMifareWrite"}, {101, nullptr, "GetAwakeTriggerReasonForLeftRail"}, {102, nullptr, "GetAwakeTriggerReasonForRightRail"}, + {103, nullptr, "GetAwakeTriggerBatteryLevelTransitionForLeftRail"}, + {104, nullptr, "GetAwakeTriggerBatteryLevelTransitionForRightRail"}, }; // clang-format on diff --git a/src/core/hle/service/ldr/ldr.cpp b/src/core/hle/service/ldr/ldr.cpp index d111c1357..c8bc60ad1 100644 --- a/src/core/hle/service/ldr/ldr.cpp +++ b/src/core/hle/service/ldr/ldr.cpp @@ -118,9 +118,9 @@ public: explicit DebugMonitor(Core::System& system_) : ServiceFramework{system_, "ldr:dmnt"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "AddProcessToDebugLaunchQueue"}, - {1, nullptr, "ClearDebugLaunchQueue"}, - {2, nullptr, "GetNsoInfos"}, + {0, nullptr, "SetProgramArgument"}, + {1, nullptr, "FlushArguments"}, + {2, nullptr, "GetProcessModuleInfo"}, }; // clang-format on @@ -135,8 +135,8 @@ public: static const FunctionInfo functions[] = { {0, nullptr, "CreateProcess"}, {1, nullptr, "GetProgramInfo"}, - {2, nullptr, "RegisterTitle"}, - {3, nullptr, "UnregisterTitle"}, + {2, nullptr, "PinProgram"}, + {3, nullptr, "UnpinProgram"}, {4, nullptr, "SetEnabledProgramVerification"}, }; // clang-format on @@ -150,8 +150,8 @@ public: explicit Shell(Core::System& system_) : ServiceFramework{system_, "ldr:shel"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "AddProcessToLaunchQueue"}, - {1, nullptr, "ClearLaunchQueue"}, + {0, nullptr, "SetProgramArgument"}, + {1, nullptr, "FlushArguments"}, }; // clang-format on @@ -164,19 +164,19 @@ public: explicit RelocatableObject(Core::System& system_) : ServiceFramework{system_, "ldr:ro"} { // clang-format off static const FunctionInfo functions[] = { - {0, &RelocatableObject::LoadNro, "LoadNro"}, - {1, &RelocatableObject::UnloadNro, "UnloadNro"}, - {2, &RelocatableObject::LoadNrr, "LoadNrr"}, - {3, &RelocatableObject::UnloadNrr, "UnloadNrr"}, + {0, &RelocatableObject::LoadModule, "LoadModule"}, + {1, &RelocatableObject::UnloadModule, "UnloadModule"}, + {2, &RelocatableObject::RegisterModuleInfo, "RegisterModuleInfo"}, + {3, &RelocatableObject::UnregisterModuleInfo, "UnregisterModuleInfo"}, {4, &RelocatableObject::Initialize, "Initialize"}, - {10, nullptr, "LoadNrrEx"}, + {10, nullptr, "RegisterModuleInfo2"}, }; // clang-format on RegisterHandlers(functions); } - void LoadNrr(Kernel::HLERequestContext& ctx) { + void RegisterModuleInfo(Kernel::HLERequestContext& ctx) { struct Parameters { u64_le process_id; u64_le nrr_address; @@ -273,7 +273,7 @@ public: rb.Push(RESULT_SUCCESS); } - void UnloadNrr(Kernel::HLERequestContext& ctx) { + void UnregisterModuleInfo(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto pid = rp.Pop<u64>(); const auto nrr_address = rp.Pop<VAddr>(); @@ -408,7 +408,7 @@ public: data_start, bss_end_addr - data_start, Kernel::KMemoryPermission::ReadAndWrite); } - void LoadNro(Kernel::HLERequestContext& ctx) { + void LoadModule(Kernel::HLERequestContext& ctx) { struct Parameters { u64_le process_id; u64_le image_address; @@ -546,7 +546,7 @@ public: return RESULT_SUCCESS; } - void UnloadNro(Kernel::HLERequestContext& ctx) { + void UnloadModule(Kernel::HLERequestContext& ctx) { if (!initialized) { LOG_ERROR(Service_LDR, "LDR:RO not initialized before use!"); IPC::ResponseBuilder rb{ctx, 2}; diff --git a/src/core/hle/service/nim/nim.cpp b/src/core/hle/service/nim/nim.cpp index f3be0b878..fee360ab9 100644 --- a/src/core/hle/service/nim/nim.cpp +++ b/src/core/hle/service/nim/nim.cpp @@ -125,51 +125,51 @@ public: {39, nullptr, "PrepareShutdown"}, {40, nullptr, "ListApplyDeltaTask"}, {41, nullptr, "ClearNotEnoughSpaceStateOfApplyDeltaTask"}, - {42, nullptr, "Unknown42"}, - {43, nullptr, "Unknown43"}, - {44, nullptr, "Unknown44"}, - {45, nullptr, "Unknown45"}, - {46, nullptr, "Unknown46"}, - {47, nullptr, "Unknown47"}, - {48, nullptr, "Unknown48"}, - {49, nullptr, "Unknown49"}, - {50, nullptr, "Unknown50"}, - {51, nullptr, "Unknown51"}, - {52, nullptr, "Unknown52"}, - {53, nullptr, "Unknown53"}, - {54, nullptr, "Unknown54"}, - {55, nullptr, "Unknown55"}, - {56, nullptr, "Unknown56"}, - {57, nullptr, "Unknown57"}, - {58, nullptr, "Unknown58"}, - {59, nullptr, "Unknown59"}, - {60, nullptr, "Unknown60"}, - {61, nullptr, "Unknown61"}, - {62, nullptr, "Unknown62"}, - {63, nullptr, "Unknown63"}, - {64, nullptr, "Unknown64"}, - {65, nullptr, "Unknown65"}, - {66, nullptr, "Unknown66"}, - {67, nullptr, "Unknown67"}, - {68, nullptr, "Unknown68"}, - {69, nullptr, "Unknown69"}, - {70, nullptr, "Unknown70"}, - {71, nullptr, "Unknown71"}, - {72, nullptr, "Unknown72"}, - {73, nullptr, "Unknown73"}, - {74, nullptr, "Unknown74"}, - {75, nullptr, "Unknown75"}, - {76, nullptr, "Unknown76"}, - {77, nullptr, "Unknown77"}, - {78, nullptr, "Unknown78"}, - {79, nullptr, "Unknown79"}, - {80, nullptr, "Unknown80"}, - {81, nullptr, "Unknown81"}, - {82, nullptr, "Unknown82"}, - {83, nullptr, "Unknown83"}, + {42, nullptr, "CreateApplyDeltaTaskFromDownloadTask"}, + {43, nullptr, "GetBackgroundApplyDeltaStressTaskInfo"}, + {44, nullptr, "GetApplyDeltaTaskRequiredStorage"}, + {45, nullptr, "CalculateNetworkInstallTaskContentsSize"}, + {46, nullptr, "PrepareShutdownForSystemUpdate"}, + {47, nullptr, "FindMaxRequiredApplicationVersionOfTask"}, + {48, nullptr, "CommitNetworkInstallTaskPartially"}, + {49, nullptr, "ListNetworkInstallTaskCommittedContentMeta"}, + {50, nullptr, "ListNetworkInstallTaskNotCommittedContentMeta"}, + {51, nullptr, "FindMaxRequiredSystemVersionOfTask"}, + {52, nullptr, "GetNetworkInstallTaskErrorContext"}, + {53, nullptr, "CreateLocalCommunicationReceiveApplicationTask"}, + {54, nullptr, "DestroyLocalCommunicationReceiveApplicationTask"}, + {55, nullptr, "ListLocalCommunicationReceiveApplicationTask"}, + {56, nullptr, "RequestLocalCommunicationReceiveApplicationTaskRun"}, + {57, nullptr, "GetLocalCommunicationReceiveApplicationTaskInfo"}, + {58, nullptr, "CommitLocalCommunicationReceiveApplicationTask"}, + {59, nullptr, "ListLocalCommunicationReceiveApplicationTaskContentMeta"}, + {60, nullptr, "CreateLocalCommunicationSendApplicationTask"}, + {61, nullptr, "RequestLocalCommunicationSendApplicationTaskRun"}, + {62, nullptr, "GetLocalCommunicationReceiveApplicationTaskErrorContext"}, + {63, nullptr, "GetLocalCommunicationSendApplicationTaskInfo"}, + {64, nullptr, "DestroyLocalCommunicationSendApplicationTask"}, + {65, nullptr, "GetLocalCommunicationSendApplicationTaskErrorContext"}, + {66, nullptr, "CalculateLocalCommunicationReceiveApplicationTaskRequiredSize"}, + {67, nullptr, "ListApplicationLocalCommunicationReceiveApplicationTask"}, + {68, nullptr, "ListApplicationLocalCommunicationSendApplicationTask"}, + {69, nullptr, "CreateLocalCommunicationReceiveSystemUpdateTask"}, + {70, nullptr, "DestroyLocalCommunicationReceiveSystemUpdateTask"}, + {71, nullptr, "ListLocalCommunicationReceiveSystemUpdateTask"}, + {72, nullptr, "RequestLocalCommunicationReceiveSystemUpdateTaskRun"}, + {73, nullptr, "GetLocalCommunicationReceiveSystemUpdateTaskInfo"}, + {74, nullptr, "CommitLocalCommunicationReceiveSystemUpdateTask"}, + {75, nullptr, "GetLocalCommunicationReceiveSystemUpdateTaskErrorContext"}, + {76, nullptr, "CreateLocalCommunicationSendSystemUpdateTask"}, + {77, nullptr, "RequestLocalCommunicationSendSystemUpdateTaskRun"}, + {78, nullptr, "GetLocalCommunicationSendSystemUpdateTaskInfo"}, + {79, nullptr, "DestroyLocalCommunicationSendSystemUpdateTask"}, + {80, nullptr, "GetLocalCommunicationSendSystemUpdateTaskErrorContext"}, + {81, nullptr, "ListLocalCommunicationSendSystemUpdateTask"}, + {82, nullptr, "GetReceivedSystemDataPath"}, + {83, nullptr, "CalculateApplyDeltaTaskOccupiedSize"}, {84, nullptr, "Unknown84"}, - {85, nullptr, "Unknown85"}, - {86, nullptr, "Unknown86"}, + {85, nullptr, "ListNetworkInstallTaskContentMetaFromInstallMeta"}, + {86, nullptr, "ListNetworkInstallTaskOccupiedSize"}, {87, nullptr, "Unknown87"}, {88, nullptr, "Unknown88"}, {89, nullptr, "Unknown89"}, @@ -202,6 +202,17 @@ public: {116, nullptr, "Unknown116"}, {117, nullptr, "Unknown117"}, {118, nullptr, "Unknown118"}, + {119, nullptr, "Unknown119"}, + {120, nullptr, "Unknown120"}, + {121, nullptr, "Unknown121"}, + {122, nullptr, "Unknown122"}, + {123, nullptr, "Unknown123"}, + {124, nullptr, "Unknown124"}, + {125, nullptr, "Unknown125"}, + {126, nullptr, "Unknown126"}, + {127, nullptr, "Unknown127"}, + {128, nullptr, "Unknown128"}, + {129, nullptr, "Unknown129"}, }; // clang-format on diff --git a/src/core/hle/service/npns/npns.cpp b/src/core/hle/service/npns/npns.cpp index f7a58f659..e4c703da4 100644 --- a/src/core/hle/service/npns/npns.cpp +++ b/src/core/hle/service/npns/npns.cpp @@ -49,6 +49,8 @@ public: {151, nullptr, "GetStateWithHandover"}, {152, nullptr, "GetStateChangeEventWithHandover"}, {153, nullptr, "GetDropEventWithHandover"}, + {154, nullptr, "CreateTokenAsync"}, + {155, nullptr, "CreateTokenAsyncWithApplicationId"}, {161, nullptr, "GetRequestChangeStateCancelEvent"}, {162, nullptr, "RequestChangeStateForceTimedWithCancelEvent"}, {201, nullptr, "RequestChangeStateForceTimed"}, @@ -84,6 +86,7 @@ public: {151, nullptr, "GetStateWithHandover"}, {152, nullptr, "GetStateChangeEventWithHandover"}, {153, nullptr, "GetDropEventWithHandover"}, + {154, nullptr, "CreateTokenAsync"}, }; // clang-format on diff --git a/src/core/hle/service/ns/ns.cpp b/src/core/hle/service/ns/ns.cpp index 6ccf8995c..5fe7a9189 100644 --- a/src/core/hle/service/ns/ns.cpp +++ b/src/core/hle/service/ns/ns.cpp @@ -55,6 +55,7 @@ IApplicationManagerInterface::IApplicationManagerInterface(Core::System& system_ {26, nullptr, "BeginInstallApplication"}, {27, nullptr, "DeleteApplicationRecord"}, {30, nullptr, "RequestApplicationUpdateInfo"}, + {31, nullptr, "Unknown31"}, {32, nullptr, "CancelApplicationDownload"}, {33, nullptr, "ResumeApplicationDownload"}, {35, nullptr, "UpdateVersionList"}, @@ -182,6 +183,7 @@ IApplicationManagerInterface::IApplicationManagerInterface(Core::System& system_ {913, nullptr, "ListAllApplicationRecord"}, {914, nullptr, "HideApplicationRecord"}, {915, nullptr, "ShowApplicationRecord"}, + {916, nullptr, "IsApplicationAutoDeleteDisabled"}, {1000, nullptr, "RequestVerifyApplicationDeprecated"}, {1001, nullptr, "CorruptApplicationForDebug"}, {1002, nullptr, "RequestVerifyAddOnContentsRights"}, @@ -201,6 +203,8 @@ IApplicationManagerInterface::IApplicationManagerInterface(Core::System& system_ {1310, nullptr, "RequestMoveApplicationEntity"}, {1311, nullptr, "EstimateSizeToMove"}, {1312, nullptr, "HasMovableEntity"}, + {1313, nullptr, "CleanupOrphanContents"}, + {1314, nullptr, "CheckPreconditionSatisfiedToMove"}, {1400, nullptr, "PrepareShutdown"}, {1500, nullptr, "FormatSdCard"}, {1501, nullptr, "NeedsSystemUpdateToFormatSdCard"}, @@ -215,6 +219,7 @@ IApplicationManagerInterface::IApplicationManagerInterface(Core::System& system_ {1702, nullptr, "GetApplicationDownloadTaskStatus"}, {1703, nullptr, "GetApplicationViewDownloadErrorContext"}, {1704, nullptr, "GetApplicationViewWithPromotionInfo"}, + {1705, nullptr, "IsPatchAutoDeletableApplication"}, {1800, nullptr, "IsNotificationSetupCompleted"}, {1801, nullptr, "GetLastNotificationInfoCount"}, {1802, nullptr, "ListLastNotificationInfo"}, @@ -269,6 +274,9 @@ IApplicationManagerInterface::IApplicationManagerInterface(Core::System& system_ {2351, nullptr, "RequestNoDownloadRightsErrorResolution"}, {2352, nullptr, "RequestResolveNoDownloadRightsError"}, {2353, nullptr, "GetApplicationDownloadTaskInfo"}, + {2354, nullptr, "PrioritizeApplicationBackgroundTask"}, + {2355, nullptr, "Unknown2355"}, + {2356, nullptr, "Unknown2356"}, {2400, nullptr, "GetPromotionInfo"}, {2401, nullptr, "CountPromotionInfo"}, {2402, nullptr, "ListPromotionInfo"}, @@ -282,6 +290,21 @@ IApplicationManagerInterface::IApplicationManagerInterface(Core::System& system_ {2515, nullptr, "CleanupAllPlaceHolderAndFragmentsIfNoTask"}, {2516, nullptr, "EnsureApplicationCertificate"}, {2800, nullptr, "GetApplicationIdOfPreomia"}, + {3000, nullptr, "RegisterDeviceLockKey"}, + {3001, nullptr, "UnregisterDeviceLockKey"}, + {3002, nullptr, "VerifyDeviceLockKey"}, + {3003, nullptr, "HideApplicationIcon"}, + {3004, nullptr, "ShowApplicationIcon"}, + {3005, nullptr, "HideApplicationTitle"}, + {3006, nullptr, "ShowApplicationTitle"}, + {3007, nullptr, "EnableGameCard"}, + {3008, nullptr, "DisableGameCard"}, + {3009, nullptr, "EnableLocalContentShare"}, + {3010, nullptr, "DisableLocalContentShare"}, + {3011, nullptr, "IsApplicationIconHidden"}, + {3012, nullptr, "IsApplicationTitleHidden"}, + {3013, nullptr, "IsGameCardEnabled"}, + {3014, nullptr, "IsLocalContentShareEnabled"}, {9999, nullptr, "GetApplicationCertificate"}, }; // clang-format on @@ -441,7 +464,11 @@ IApplicationVersionInterface::IApplicationVersionInterface(Core::System& system_ {800, nullptr, "RequestVersionList"}, {801, nullptr, "ListVersionList"}, {802, nullptr, "RequestVersionListData"}, + {900, nullptr, "ImportAutoUpdatePolicyJsonForDebug"}, + {901, nullptr, "ListDefaultAutoUpdatePolicy"}, + {902, nullptr, "ListAutoUpdatePolicyForSpecificApplication"}, {1000, nullptr, "PerformAutoUpdate"}, + {1001, nullptr, "ListAutoUpdateSchedule"}, }; // clang-format on @@ -547,6 +574,9 @@ IFactoryResetInterface::~IFactoryResetInterface() = default; NS::NS(const char* name, Core::System& system_) : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { + {7988, nullptr, "GetDynamicRightsInterface"}, + {7989, nullptr, "GetReadOnlyApplicationControlDataInterface"}, + {7991, nullptr, "GetReadOnlyApplicationRecordInterface"}, {7992, &NS::PushInterface<IECommerceInterface>, "GetECommerceInterface"}, {7993, &NS::PushInterface<IApplicationVersionInterface>, "GetApplicationVersionInterface"}, {7994, &NS::PushInterface<IFactoryResetInterface>, "GetFactoryResetInterface"}, @@ -575,18 +605,22 @@ public: {0, nullptr, "LaunchProgram"}, {1, nullptr, "TerminateProcess"}, {2, nullptr, "TerminateProgram"}, - {4, nullptr, "GetShellEventHandle"}, + {4, nullptr, "GetShellEvent"}, {5, nullptr, "GetShellEventInfo"}, {6, nullptr, "TerminateApplication"}, {7, nullptr, "PrepareLaunchProgramFromHost"}, - {8, nullptr, "LaunchApplication"}, + {8, nullptr, "LaunchApplicationFromHost"}, {9, nullptr, "LaunchApplicationWithStorageIdForDevelop"}, {10, nullptr, "IsSystemMemoryResourceLimitBoosted"}, {11, nullptr, "GetRunningApplicationProcessIdForDevelop"}, - {12, nullptr, "SetCurrentApplicationRightsEnvironmentCanBeActive"}, + {12, nullptr, "SetCurrentApplicationRightsEnvironmentCanBeActiveForDevelop"}, {13, nullptr, "CreateApplicationResourceForDevelop"}, {14, nullptr, "IsPreomiaForDevelop"}, {15, nullptr, "GetApplicationProgramIdFromHost"}, + {16, nullptr, "RefreshCachedDebugValues"}, + {17, nullptr, "PrepareLaunchApplicationFromHost"}, + {18, nullptr, "GetLaunchEvent"}, + {19, nullptr, "GetLaunchResult"}, }; // clang-format on @@ -699,6 +733,7 @@ void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system std::make_shared<NS>("ns:rid", system)->InstallAsService(service_manager); std::make_shared<NS>("ns:rt", system)->InstallAsService(service_manager); std::make_shared<NS>("ns:web", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:ro", system)->InstallAsService(service_manager); std::make_shared<NS_DEV>(system)->InstallAsService(service_manager); std::make_shared<NS_SU>(system)->InstallAsService(service_manager); diff --git a/src/core/hle/service/ns/pl_u.cpp b/src/core/hle/service/ns/pl_u.cpp index fcd15d81f..da139fdc4 100644 --- a/src/core/hle/service/ns/pl_u.cpp +++ b/src/core/hle/service/ns/pl_u.cpp @@ -154,6 +154,10 @@ PL_U::PL_U(Core::System& system_) {100, nullptr, "RequestApplicationFunctionAuthorization"}, {101, nullptr, "RequestApplicationFunctionAuthorizationByProcessId"}, {102, nullptr, "RequestApplicationFunctionAuthorizationByApplicationId"}, + {103, nullptr, "RefreshApplicationFunctionBlackListDebugRecord"}, + {104, nullptr, "RequestApplicationFunctionAuthorizationByProgramId"}, + {105, nullptr, "GetFunctionBlackListSystemVersionToAuthorize"}, + {106, nullptr, "GetFunctionBlackListVersion"}, {1000, nullptr, "LoadNgWordDataForPlatformRegionChina"}, {1001, nullptr, "GetNgWordDataSizeForPlatformRegionChina"}, }; diff --git a/src/core/hle/service/nvdrv/devices/nvdevice.h b/src/core/hle/service/nvdrv/devices/nvdevice.h index 5681599ba..b37f023df 100644 --- a/src/core/hle/service/nvdrv/devices/nvdevice.h +++ b/src/core/hle/service/nvdrv/devices/nvdevice.h @@ -31,7 +31,7 @@ public: * @param output A buffer where the output data will be written to. * @returns The result code of the ioctl. */ - virtual NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, + virtual NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) = 0; /** @@ -42,7 +42,7 @@ public: * @param output A buffer where the output data will be written to. * @returns The result code of the ioctl. */ - virtual NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + virtual NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) = 0; /** @@ -53,8 +53,20 @@ public: * @param inline_output A buffer where the inlined output data will be written to. * @returns The result code of the ioctl. */ - virtual NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) = 0; + virtual NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) = 0; + + /** + * Called once a device is openned + * @param fd The device fd + */ + virtual void OnOpen(DeviceFD fd) = 0; + + /** + * Called once a device is closed + * @param fd The device fd + */ + virtual void OnClose(DeviceFD fd) = 0; protected: Core::System& system; diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp index ce615c758..5ab7e39b0 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp @@ -18,24 +18,27 @@ nvdisp_disp0::nvdisp_disp0(Core::System& system, std::shared_ptr<nvmap> nvmap_de : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} nvdisp_disp0 ::~nvdisp_disp0() = default; -NvResult nvdisp_disp0::Ioctl1(Ioctl command, const std::vector<u8>& input, +NvResult nvdisp_disp0::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvdisp_disp0::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvdisp_disp0::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvdisp_disp0::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) { +NvResult nvdisp_disp0::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } +void nvdisp_disp0::OnOpen(DeviceFD fd) {} +void nvdisp_disp0::OnClose(DeviceFD fd) {} + void nvdisp_disp0::flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u32 height, u32 stride, NVFlinger::BufferQueue::BufferTransformFlags transform, const Common::Rectangle<int>& crop_rect) { diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h index 55a33b7e4..59c9b6101 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h @@ -20,11 +20,15 @@ public: explicit nvdisp_disp0(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); ~nvdisp_disp0() override; - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; /// Performs a screen flip, drawing the buffer pointed to by the handle. void flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u32 height, u32 stride, diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp index 485ac5f50..f7b3dc317 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp @@ -21,7 +21,7 @@ nvhost_as_gpu::nvhost_as_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_ : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} nvhost_as_gpu::~nvhost_as_gpu() = default; -NvResult nvhost_as_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_as_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { switch (command.group) { case 'A': @@ -54,14 +54,14 @@ NvResult nvhost_as_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, return NvResult::NotImplemented; } -NvResult nvhost_as_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_as_gpu::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvhost_as_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) { +NvResult nvhost_as_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { switch (command.group) { case 'A': switch (command.cmd) { @@ -78,6 +78,9 @@ NvResult nvhost_as_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, std: return NvResult::NotImplemented; } +void nvhost_as_gpu::OnOpen(DeviceFD fd) {} +void nvhost_as_gpu::OnClose(DeviceFD fd) {} + NvResult nvhost_as_gpu::AllocAsEx(const std::vector<u8>& input, std::vector<u8>& output) { IoctlAllocAsEx params{}; std::memcpy(¶ms, input.data(), input.size()); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h index 9ee60e060..d86a9cab6 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h @@ -33,11 +33,15 @@ public: explicit nvhost_as_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); ~nvhost_as_gpu() override; - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; private: class BufferMap final { diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp index f6129ef10..9f00d5cb0 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp @@ -20,7 +20,8 @@ nvhost_ctrl::nvhost_ctrl(Core::System& system, EventInterface& events_interface, : nvdevice(system), events_interface{events_interface}, syncpoint_manager{syncpoint_manager} {} nvhost_ctrl::~nvhost_ctrl() = default; -NvResult nvhost_ctrl::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { switch (command.group) { case 0x0: switch (command.cmd) { @@ -46,18 +47,21 @@ NvResult nvhost_ctrl::Ioctl1(Ioctl command, const std::vector<u8>& input, std::v return NvResult::NotImplemented; } -NvResult nvhost_ctrl::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_ctrl::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvhost_ctrl::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_outpu) { +NvResult nvhost_ctrl::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_outpu) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } +void nvhost_ctrl::OnOpen(DeviceFD fd) {} +void nvhost_ctrl::OnClose(DeviceFD fd) {} + NvResult nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output) { IocGetConfigParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h index c5aa1362a..9178789c3 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h @@ -18,11 +18,15 @@ public: SyncpointManager& syncpoint_manager); ~nvhost_ctrl() override; - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; private: struct IocSyncptReadParams { diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp index 0320d3ae2..2edd803f3 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp @@ -15,7 +15,7 @@ namespace Service::Nvidia::Devices { nvhost_ctrl_gpu::nvhost_ctrl_gpu(Core::System& system) : nvdevice(system) {} nvhost_ctrl_gpu::~nvhost_ctrl_gpu() = default; -NvResult nvhost_ctrl_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_ctrl_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { switch (command.group) { case 'G': @@ -47,13 +47,13 @@ NvResult nvhost_ctrl_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, return NvResult::NotImplemented; } -NvResult nvhost_ctrl_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_ctrl_gpu::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvhost_ctrl_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_ctrl_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, std::vector<u8>& inline_output) { switch (command.group) { case 'G': @@ -73,6 +73,9 @@ NvResult nvhost_ctrl_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, return NvResult::NotImplemented; } +void nvhost_ctrl_gpu::OnOpen(DeviceFD fd) {} +void nvhost_ctrl_gpu::OnClose(DeviceFD fd) {} + NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); @@ -245,7 +248,13 @@ NvResult nvhost_ctrl_gpu::ZBCSetTable(const std::vector<u8>& input, std::vector< IoctlZbcSetTable params{}; std::memcpy(¶ms, input.data(), input.size()); // TODO(ogniK): What does this even actually do? - std::memcpy(output.data(), ¶ms, output.size()); + + // Prevent null pointer being passed as arg 1 + if (output.empty()) { + LOG_WARNING(Service_NVDRV, "Avoiding passing null pointer to memcpy"); + } else { + std::memcpy(output.data(), ¶ms, output.size()); + } return NvResult::Success; } diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h index 137b88238..f98aa841a 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h @@ -16,11 +16,15 @@ public: explicit nvhost_ctrl_gpu(Core::System& system); ~nvhost_ctrl_gpu() override; - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; private: struct IoctlGpuCharacteristics { diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp index af8b3d9f1..e83aaa798 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp @@ -23,7 +23,8 @@ nvhost_gpu::nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, nvhost_gpu::~nvhost_gpu() = default; -NvResult nvhost_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { switch (command.group) { case 0x0: switch (command.cmd) { @@ -74,7 +75,7 @@ NvResult nvhost_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, std::ve return NvResult::NotImplemented; }; -NvResult nvhost_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_gpu::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { switch (command.group) { case 'H': @@ -88,12 +89,15 @@ NvResult nvhost_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, return NvResult::NotImplemented; } -NvResult nvhost_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) { +NvResult nvhost_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } +void nvhost_gpu::OnOpen(DeviceFD fd) {} +void nvhost_gpu::OnClose(DeviceFD fd) {} + NvResult nvhost_gpu::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetNvmapFD params{}; std::memcpy(¶ms, input.data(), input.size()); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h index e0298b4fe..12a1a1133 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h @@ -26,11 +26,15 @@ public: SyncpointManager& syncpoint_manager); ~nvhost_gpu() override; - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; private: enum class CtxObjects : u32_le { diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp index ecba1dba1..c8031970b 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp @@ -16,7 +16,7 @@ nvhost_nvdec::nvhost_nvdec(Core::System& system, std::shared_ptr<nvmap> nvmap_de : nvhost_nvdec_common(system, std::move(nvmap_dev), syncpoint_manager) {} nvhost_nvdec::~nvhost_nvdec() = default; -NvResult nvhost_nvdec::Ioctl1(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_nvdec::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { switch (command.group) { case 0x0: @@ -57,16 +57,19 @@ NvResult nvhost_nvdec::Ioctl1(Ioctl command, const std::vector<u8>& input, return NvResult::NotImplemented; } -NvResult nvhost_nvdec::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_nvdec::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvhost_nvdec::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) { +NvResult nvhost_nvdec::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } +void nvhost_nvdec::OnOpen(DeviceFD fd) {} +void nvhost_nvdec::OnClose(DeviceFD fd) {} + } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h index 77ef53cdd..6c38a8c24 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h @@ -15,11 +15,15 @@ public: SyncpointManager& syncpoint_manager); ~nvhost_nvdec() override; - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp index 4898dc27a..c2f152190 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp @@ -23,17 +23,22 @@ namespace { template <typename T> std::size_t SpliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count, std::size_t offset) { - std::memcpy(dst.data(), input.data() + offset, count * sizeof(T)); - offset += count * sizeof(T); - return offset; + if (!dst.empty()) { + std::memcpy(dst.data(), input.data() + offset, count * sizeof(T)); + } + return 0; } // Write vectors will write data to the output buffer template <typename T> std::size_t WriteVectors(std::vector<u8>& dst, const std::vector<T>& src, std::size_t offset) { - std::memcpy(dst.data() + offset, src.data(), src.size() * sizeof(T)); - offset += src.size() * sizeof(T); - return offset; + if (src.empty()) { + return 0; + } else { + std::memcpy(dst.data() + offset, src.data(), src.size() * sizeof(T)); + offset += src.size() * sizeof(T); + return offset; + } } } // Anonymous namespace diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp index 2d06955c0..0a9c35c01 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp @@ -13,7 +13,7 @@ namespace Service::Nvidia::Devices { nvhost_nvjpg::nvhost_nvjpg(Core::System& system) : nvdevice(system) {} nvhost_nvjpg::~nvhost_nvjpg() = default; -NvResult nvhost_nvjpg::Ioctl1(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_nvjpg::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { switch (command.group) { case 'H': @@ -32,18 +32,21 @@ NvResult nvhost_nvjpg::Ioctl1(Ioctl command, const std::vector<u8>& input, return NvResult::NotImplemented; } -NvResult nvhost_nvjpg::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_nvjpg::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvhost_nvjpg::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) { +NvResult nvhost_nvjpg::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } +void nvhost_nvjpg::OnOpen(DeviceFD fd) {} +void nvhost_nvjpg::OnClose(DeviceFD fd) {} + NvResult nvhost_nvjpg::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetNvmapFD params{}; std::memcpy(¶ms, input.data(), input.size()); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h index 43948d18d..1f97b642f 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h @@ -16,11 +16,15 @@ public: explicit nvhost_nvjpg(Core::System& system); ~nvhost_nvjpg() override; - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; private: struct IoctlSetNvmapFD { diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp index 70849a9bd..0421fb956 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp @@ -16,7 +16,8 @@ nvhost_vic::nvhost_vic(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, nvhost_vic::~nvhost_vic() = default; -NvResult nvhost_vic::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_vic::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { switch (command.group) { case 0x0: switch (command.cmd) { @@ -55,16 +56,19 @@ NvResult nvhost_vic::Ioctl1(Ioctl command, const std::vector<u8>& input, std::ve return NvResult::NotImplemented; } -NvResult nvhost_vic::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvhost_vic::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvhost_vic::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) { +NvResult nvhost_vic::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } +void nvhost_vic::OnOpen(DeviceFD fd) {} +void nvhost_vic::OnClose(DeviceFD fd) {} + } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.h b/src/core/hle/service/nvdrv/devices/nvhost_vic.h index f401c61fa..cebefad71 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_vic.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.h @@ -14,10 +14,14 @@ public: SyncpointManager& syncpoint_manager); ~nvhost_vic(); - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvmap.cpp b/src/core/hle/service/nvdrv/devices/nvmap.cpp index 4015a2740..dd1355522 100644 --- a/src/core/hle/service/nvdrv/devices/nvmap.cpp +++ b/src/core/hle/service/nvdrv/devices/nvmap.cpp @@ -19,7 +19,8 @@ nvmap::nvmap(Core::System& system) : nvdevice(system) { nvmap::~nvmap() = default; -NvResult nvmap::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { switch (command.group) { case 0x1: switch (command.cmd) { @@ -47,18 +48,21 @@ NvResult nvmap::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector< return NvResult::NotImplemented; } -NvResult nvmap::Ioctl2(Ioctl command, const std::vector<u8>& input, +NvResult nvmap::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvmap::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) { +NvResult nvmap::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } +void nvmap::OnOpen(DeviceFD fd) {} +void nvmap::OnClose(DeviceFD fd) {} + VAddr nvmap::GetObjectAddress(u32 handle) const { auto object = GetObject(handle); ASSERT(object); diff --git a/src/core/hle/service/nvdrv/devices/nvmap.h b/src/core/hle/service/nvdrv/devices/nvmap.h index 4484bd79f..208875845 100644 --- a/src/core/hle/service/nvdrv/devices/nvmap.h +++ b/src/core/hle/service/nvdrv/devices/nvmap.h @@ -19,11 +19,15 @@ public: explicit nvmap(Core::System& system); ~nvmap() override; - NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; - NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) override; + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, const std::vector<u8>& inline_input, std::vector<u8>& output) override; - NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& inline_output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) override; + + void OnOpen(DeviceFD fd) override; + void OnClose(DeviceFD fd) override; /// Returns the allocated address of an nvmap object given its handle. VAddr GetObjectAddress(u32 handle) const; diff --git a/src/core/hle/service/nvdrv/nvdrv.cpp b/src/core/hle/service/nvdrv/nvdrv.cpp index abba80112..ede77858a 100644 --- a/src/core/hle/service/nvdrv/nvdrv.cpp +++ b/src/core/hle/service/nvdrv/nvdrv.cpp @@ -89,6 +89,8 @@ DeviceFD Module::Open(const std::string& device_name) { auto device = devices[device_name]; const DeviceFD fd = next_fd++; + device->OnOpen(fd); + open_files[fd] = std::move(device); return fd; @@ -108,7 +110,7 @@ NvResult Module::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input return NvResult::NotImplemented; } - return itr->second->Ioctl1(command, input, output); + return itr->second->Ioctl1(fd, command, input, output); } NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, @@ -125,7 +127,7 @@ NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input return NvResult::NotImplemented; } - return itr->second->Ioctl2(command, input, inline_input, output); + return itr->second->Ioctl2(fd, command, input, inline_input, output); } NvResult Module::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, @@ -142,7 +144,7 @@ NvResult Module::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input return NvResult::NotImplemented; } - return itr->second->Ioctl3(command, input, output, inline_output); + return itr->second->Ioctl3(fd, command, input, output, inline_output); } NvResult Module::Close(DeviceFD fd) { @@ -158,6 +160,8 @@ NvResult Module::Close(DeviceFD fd) { return NvResult::NotImplemented; } + itr->second->OnClose(fd); + open_files.erase(itr); return NvResult::Success; diff --git a/src/core/hle/service/olsc/olsc.cpp b/src/core/hle/service/olsc/olsc.cpp index e2ac71fa1..b066c3417 100644 --- a/src/core/hle/service/olsc/olsc.cpp +++ b/src/core/hle/service/olsc/olsc.cpp @@ -26,6 +26,7 @@ public: {22, nullptr, "DeleteSaveDataBackupAsync"}, {25, nullptr, "ListDownloadableSaveDataBackupInfoAsync"}, {26, nullptr, "DownloadSaveDataBackupAsync"}, + {27, nullptr, "UploadSaveDataBackupAsync"}, {9010, nullptr, "VerifySaveDataBackupLicenseAsyncForDebug"}, {9013, nullptr, "GetSaveDataBackupSettingForDebug"}, {9014, nullptr, "SetSaveDataBackupSettingEnabledForDebug"}, diff --git a/src/core/hle/service/pcie/pcie.cpp b/src/core/hle/service/pcie/pcie.cpp index f6686fc4d..9bc851591 100644 --- a/src/core/hle/service/pcie/pcie.cpp +++ b/src/core/hle/service/pcie/pcie.cpp @@ -37,7 +37,7 @@ public: {19, nullptr, "SetIrqEnable"}, {20, nullptr, "SetAspmEnable"}, {21, nullptr, "SetResetUponResumeEnable"}, - {22, nullptr, "Unknown22"}, + {22, nullptr, "ResetFunction"}, {23, nullptr, "Unknown23"}, }; // clang-format on diff --git a/src/core/hle/service/pctl/module.cpp b/src/core/hle/service/pctl/module.cpp index f17c1cf7f..dc59702f1 100644 --- a/src/core/hle/service/pctl/module.cpp +++ b/src/core/hle/service/pctl/module.cpp @@ -64,6 +64,7 @@ public: {1046, nullptr, "DisableFeaturesForReset"}, {1047, nullptr, "NotifyApplicationDownloadStarted"}, {1048, nullptr, "NotifyNetworkProfileCreated"}, + {1049, nullptr, "ResetFreeCommunicationApplicationList"}, {1061, &IParentalControlService::ConfirmStereoVisionRestrictionConfigurable, "ConfirmStereoVisionRestrictionConfigurable"}, {1062, &IParentalControlService::GetStereoVisionRestriction, "GetStereoVisionRestriction"}, {1063, &IParentalControlService::SetStereoVisionRestriction, "SetStereoVisionRestriction"}, @@ -83,6 +84,8 @@ public: {1421, nullptr, "GetAccountNickname"}, {1424, nullptr, "GetAccountState"}, {1425, nullptr, "RequestPostEvents"}, + {1426, nullptr, "GetPostEventInterval"}, + {1427, nullptr, "SetPostEventInterval"}, {1432, nullptr, "GetSynchronizationEvent"}, {1451, nullptr, "StartPlayTimer"}, {1452, nullptr, "StopPlayTimer"}, diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp index 1da56bc27..aec399076 100644 --- a/src/core/hle/service/service.cpp +++ b/src/core/hle/service/service.cpp @@ -70,6 +70,7 @@ #include "core/hle/service/vi/vi.h" #include "core/hle/service/wlan/wlan.h" #include "core/reporter.h" +#include "core/settings.h" namespace Service { @@ -146,6 +147,11 @@ void ServiceFrameworkBase::ReportUnimplementedFunction(Kernel::HLERequestContext system.GetReporter().SaveUnimplementedFunctionReport(ctx, ctx.GetCommand(), function_name, service_name); UNIMPLEMENTED_MSG("Unknown / unimplemented {}", fmt::to_string(buf)); + if (Settings::values.use_auto_stub) { + LOG_WARNING(Service, "Using auto stub fallback!"); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } } void ServiceFrameworkBase::InvokeRequest(Kernel::HLERequestContext& ctx) { diff --git a/src/core/hle/service/set/set_sys.cpp b/src/core/hle/service/set/set_sys.cpp index b58b2c8c5..5909fdd85 100644 --- a/src/core/hle/service/set/set_sys.cpp +++ b/src/core/hle/service/set/set_sys.cpp @@ -261,6 +261,10 @@ SET_SYS::SET_SYS(Core::System& system_) : ServiceFramework{system_, "set:sys"} { {155, nullptr, "SetAccountOnlineStorageSettings"}, {156, nullptr, "GetPctlReadyFlag"}, {157, nullptr, "SetPctlReadyFlag"}, + {158, nullptr, "GetAnalogStickUserCalibrationL"}, + {159, nullptr, "SetAnalogStickUserCalibrationL"}, + {160, nullptr, "GetAnalogStickUserCalibrationR"}, + {161, nullptr, "SetAnalogStickUserCalibrationR"}, {162, nullptr, "GetPtmBatteryVersion"}, {163, nullptr, "SetPtmBatteryVersion"}, {164, nullptr, "GetUsb30HostEnableFlag"}, @@ -302,6 +306,8 @@ SET_SYS::SET_SYS(Core::System& system_) : ServiceFramework{system_, "set:sys"} { {200, nullptr, "SetButtonConfigRegisteredSettings"}, {201, nullptr, "GetFieldTestingFlag"}, {202, nullptr, "SetFieldTestingFlag"}, + {203, nullptr, "GetPanelCrcMode"}, + {204, nullptr, "SetPanelCrcMode"}, }; // clang-format on diff --git a/src/core/hle/service/sm/sm.cpp b/src/core/hle/service/sm/sm.cpp index 2b91a89d1..94608d529 100644 --- a/src/core/hle/service/sm/sm.cpp +++ b/src/core/hle/service/sm/sm.cpp @@ -190,10 +190,11 @@ SM::SM(std::shared_ptr<ServiceManager> service_manager_, Core::System& system_) : ServiceFramework{system_, "sm:", 4}, service_manager{std::move(service_manager_)}, kernel{system_.Kernel()} { static const FunctionInfo functions[] = { - {0x00000000, &SM::Initialize, "Initialize"}, - {0x00000001, &SM::GetService, "GetService"}, - {0x00000002, &SM::RegisterService, "RegisterService"}, - {0x00000003, &SM::UnregisterService, "UnregisterService"}, + {0, &SM::Initialize, "Initialize"}, + {1, &SM::GetService, "GetService"}, + {2, &SM::RegisterService, "RegisterService"}, + {3, &SM::UnregisterService, "UnregisterService"}, + {4, nullptr, "DetachClient"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/sockets/ethc.cpp b/src/core/hle/service/sockets/ethc.cpp index 05681ca2d..899a64c2f 100644 --- a/src/core/hle/service/sockets/ethc.cpp +++ b/src/core/hle/service/sockets/ethc.cpp @@ -15,6 +15,7 @@ ETHC_C::ETHC_C(Core::System& system_) : ServiceFramework{system_, "ethc:c"} { {3, nullptr, "GetMediaList"}, {4, nullptr, "SetMediaType"}, {5, nullptr, "GetMediaType"}, + {6, nullptr, "Unknown6"}, }; // clang-format on diff --git a/src/core/hle/service/sockets/nsd.cpp b/src/core/hle/service/sockets/nsd.cpp index 51c3739bb..1159debc5 100644 --- a/src/core/hle/service/sockets/nsd.cpp +++ b/src/core/hle/service/sockets/nsd.cpp @@ -9,6 +9,7 @@ namespace Service::Sockets { NSD::NSD(Core::System& system_, const char* name) : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { + {5, nullptr, "GetSettingUrl"}, {10, nullptr, "GetSettingName"}, {11, nullptr, "GetEnvironmentIdentifier"}, {12, nullptr, "GetDeviceId"}, diff --git a/src/core/hle/service/sockets/sfdnsres.cpp b/src/core/hle/service/sockets/sfdnsres.cpp index 3a6329f56..5c71f423c 100644 --- a/src/core/hle/service/sockets/sfdnsres.cpp +++ b/src/core/hle/service/sockets/sfdnsres.cpp @@ -9,8 +9,8 @@ namespace Service::Sockets { SFDNSRES::SFDNSRES(Core::System& system_) : ServiceFramework{system_, "sfdnsres"} { static const FunctionInfo functions[] = { - {0, nullptr, "SetDnsAddressesPrivate"}, - {1, nullptr, "GetDnsAddressPrivate"}, + {0, nullptr, "SetDnsAddressesPrivateRequest"}, + {1, nullptr, "GetDnsAddressPrivateRequest"}, {2, nullptr, "GetHostByNameRequest"}, {3, nullptr, "GetHostByAddrRequest"}, {4, nullptr, "GetHostStringErrorRequest"}, diff --git a/src/core/hle/service/spl/spl.cpp b/src/core/hle/service/spl/spl.cpp index 4e212610f..fff3f3c42 100644 --- a/src/core/hle/service/spl/spl.cpp +++ b/src/core/hle/service/spl/spl.cpp @@ -60,6 +60,8 @@ SPL_FS::SPL_FS(Core::System& system_, std::shared_ptr<Module> module_) {4, nullptr, "GenerateAesKey"}, {5, nullptr, "SetConfig"}, {7, &SPL::GetRandomBytes, "GenerateRandomBytes"}, + {9, nullptr, "ImportLotusKey"}, + {10, nullptr, "DecryptLotusMessage"}, {11, nullptr, "IsDevelopment"}, {12, nullptr, "GenerateSpecificAesKey"}, {14, nullptr, "DecryptAesKey"}, @@ -123,6 +125,7 @@ SPL_ES::SPL_ES(Core::System& system_, std::shared_ptr<Module> module_) {14, nullptr, "DecryptAesKey"}, {15, nullptr, "CryptAesCtr"}, {16, nullptr, "ComputeCmac"}, + {17, nullptr, "ImportEsKey"}, {18, nullptr, "UnwrapTitleKey"}, {20, nullptr, "PrepareEsCommonKey"}, {21, nullptr, "AllocateAesKeyslot"}, diff --git a/src/core/hle/service/time/time_zone_service.cpp b/src/core/hle/service/time/time_zone_service.cpp index 25cecbc83..3117627cf 100644 --- a/src/core/hle/service/time/time_zone_service.cpp +++ b/src/core/hle/service/time/time_zone_service.cpp @@ -20,6 +20,7 @@ ITimeZoneService ::ITimeZoneService(Core::System& system_, {3, nullptr, "LoadLocationNameList"}, {4, &ITimeZoneService::LoadTimeZoneRule, "LoadTimeZoneRule"}, {5, nullptr, "GetTimeZoneRuleVersion"}, + {6, nullptr, "GetDeviceLocationNameAndUpdatedTime"}, {100, &ITimeZoneService::ToCalendarTime, "ToCalendarTime"}, {101, &ITimeZoneService::ToCalendarTimeWithMyRule, "ToCalendarTimeWithMyRule"}, {201, &ITimeZoneService::ToPosixTime, "ToPosixTime"}, diff --git a/src/core/hle/service/usb/usb.cpp b/src/core/hle/service/usb/usb.cpp index 579de83e4..b3b230a8c 100644 --- a/src/core/hle/service/usb/usb.cpp +++ b/src/core/hle/service/usb/usb.cpp @@ -69,15 +69,15 @@ public: : ServiceFramework{system_, "IClientEpSession"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Open"}, + {0, nullptr, "ReOpen"}, {1, nullptr, "Close"}, - {2, nullptr, "Unknown2"}, - {3, nullptr, "Populate"}, + {2, nullptr, "GetCompletionEvent"}, + {3, nullptr, "PopulateRing"}, {4, nullptr, "PostBufferAsync"}, {5, nullptr, "GetXferReport"}, {6, nullptr, "PostBufferMultiAsync"}, - {7, nullptr, "Unknown7"}, - {8, nullptr, "Unknown8"}, + {7, nullptr, "CreateSmmuSpace"}, + {8, nullptr, "ShareReportRing"}, }; // clang-format on @@ -91,7 +91,7 @@ public: : ServiceFramework{system_, "IClientIfSession"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown0"}, + {0, nullptr, "GetStateChangeEvent"}, {1, nullptr, "SetInterface"}, {2, nullptr, "GetInterface"}, {3, nullptr, "GetAlternateInterface"}, @@ -176,15 +176,15 @@ public: : ServiceFramework{system_, "IPdCradleSession"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "VdmUserWrite"}, - {1, nullptr, "VdmUserRead"}, - {2, nullptr, "Vdm20Init"}, - {3, nullptr, "GetFwType"}, - {4, nullptr, "GetFwRevision"}, - {5, nullptr, "GetManufacturerId"}, - {6, nullptr, "GetDeviceId"}, - {7, nullptr, "Unknown7"}, - {8, nullptr, "Unknown8"}, + {0, nullptr, "SetCradleVdo"}, + {1, nullptr, "GetCradleVdo"}, + {2, nullptr, "ResetCradleUsbHub"}, + {3, nullptr, "GetHostPdcFirmwareType"}, + {4, nullptr, "GetHostPdcFirmwareRevision"}, + {5, nullptr, "GetHostPdcManufactureId"}, + {6, nullptr, "GetHostPdcDeviceId"}, + {7, nullptr, "AwakeCradle"}, + {8, nullptr, "SleepCradle"}, }; // clang-format on @@ -219,12 +219,12 @@ public: explicit USB_PM(Core::System& system_) : ServiceFramework{system_, "usb:pm"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown0"}, - {1, nullptr, "Unknown1"}, - {2, nullptr, "Unknown2"}, - {3, nullptr, "Unknown3"}, - {4, nullptr, "Unknown4"}, - {5, nullptr, "Unknown5"}, + {0, nullptr, "GetPowerEvent"}, + {1, nullptr, "GetPowerState"}, + {2, nullptr, "GetDataEvent"}, + {3, nullptr, "GetDataRole"}, + {4, nullptr, "SetDiagData"}, + {5, nullptr, "GetDiagData"}, }; // clang-format on diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp index 7423287ea..a1a7ac987 100644 --- a/src/core/hle/service/vi/vi.cpp +++ b/src/core/hle/service/vi/vi.cpp @@ -695,6 +695,7 @@ public: {2205, &ISystemDisplayService::SetLayerZ, "SetLayerZ"}, {2207, &ISystemDisplayService::SetLayerVisibility, "SetLayerVisibility"}, {2209, nullptr, "SetLayerAlpha"}, + {2210, nullptr, "SetLayerPositionAndSize"}, {2312, nullptr, "CreateStrayLayer"}, {2400, nullptr, "OpenIndirectLayer"}, {2401, nullptr, "CloseIndirectLayer"}, @@ -718,6 +719,7 @@ public: {3215, nullptr, "SetDisplayGamma"}, {3216, nullptr, "GetDisplayCmuLuma"}, {3217, nullptr, "SetDisplayCmuLuma"}, + {3218, nullptr, "SetDisplayCrcMode"}, {6013, nullptr, "GetLayerPresentationSubmissionTimestamps"}, {8225, nullptr, "GetSharedBufferMemoryHandleId"}, {8250, nullptr, "OpenSharedLayer"}, @@ -729,6 +731,7 @@ public: {8256, nullptr, "GetSharedFrameBufferAcquirableEvent"}, {8257, nullptr, "FillSharedFrameBufferColor"}, {8258, nullptr, "CancelSharedFrameBuffer"}, + {9000, nullptr, "GetDp2hdmiController"}, }; RegisterHandlers(functions); } @@ -808,10 +811,15 @@ public: {2402, nullptr, "GetDisplayHotplugState"}, {2501, nullptr, "GetCompositorErrorInfo"}, {2601, nullptr, "GetDisplayErrorEvent"}, + {2701, nullptr, "GetDisplayFatalErrorEvent"}, {4201, nullptr, "SetDisplayAlpha"}, {4203, nullptr, "SetDisplayLayerStack"}, {4205, nullptr, "SetDisplayPowerState"}, {4206, nullptr, "SetDefaultDisplay"}, + {4207, nullptr, "ResetDisplayPanel"}, + {4208, nullptr, "SetDisplayFatalErrorEnabled"}, + {4209, nullptr, "IsDisplayPanelOn"}, + {4300, nullptr, "GetInternalPanelId"}, {6000, &IManagerDisplayService::AddToLayerStack, "AddToLayerStack"}, {6001, nullptr, "RemoveFromLayerStack"}, {6002, &IManagerDisplayService::SetLayerVisibility, "SetLayerVisibility"}, diff --git a/src/core/hle/service/wlan/wlan.cpp b/src/core/hle/service/wlan/wlan.cpp index ddbf04069..44957e01d 100644 --- a/src/core/hle/service/wlan/wlan.cpp +++ b/src/core/hle/service/wlan/wlan.cpp @@ -46,6 +46,13 @@ public: {28, nullptr, "Unknown28"}, {29, nullptr, "Unknown29"}, {30, nullptr, "Unknown30"}, + {31, nullptr, "Unknown31"}, + {32, nullptr, "Unknown32"}, + {33, nullptr, "Unknown33"}, + {34, nullptr, "Unknown34"}, + {35, nullptr, "Unknown35"}, + {36, nullptr, "Unknown36"}, + {37, nullptr, "Unknown37"}, }; // clang-format on diff --git a/src/core/settings.h b/src/core/settings.h index d849dded3..a81016b23 100644 --- a/src/core/settings.h +++ b/src/core/settings.h @@ -222,6 +222,7 @@ struct Values { bool quest_flag; bool disable_macro_jit; bool extended_logging; + bool use_auto_stub; // Miscellaneous std::string log_filter; diff --git a/src/input_common/udp/client.cpp b/src/input_common/udp/client.cpp index df73f9ff7..e72df924b 100644 --- a/src/input_common/udp/client.cpp +++ b/src/input_common/udp/client.cpp @@ -27,11 +27,9 @@ class Socket { public: using clock = std::chrono::system_clock; - explicit Socket(const std::string& host, u16 port, std::size_t pad_index_, - SocketCallback callback_) + explicit Socket(const std::string& host, u16 port, SocketCallback callback_) : callback(std::move(callback_)), timer(io_service), - socket(io_service, udp::endpoint(udp::v4(), 0)), client_id(GenerateRandomClientId()), - pad_index(pad_index_) { + socket(io_service, udp::endpoint(udp::v4(), 0)), client_id(GenerateRandomClientId()) { boost::system::error_code ec{}; auto ipv4 = boost::asio::ip::make_address_v4(host, ec); if (ec.value() != boost::system::errc::success) { @@ -99,15 +97,15 @@ private: void HandleSend(const boost::system::error_code&) { boost::system::error_code _ignored{}; // Send a request for getting port info for the pad - const Request::PortInfo port_info{1, {static_cast<u8>(pad_index), 0, 0, 0}}; + const Request::PortInfo port_info{4, {0, 1, 2, 3}}; const auto port_message = Request::Create(port_info, client_id); std::memcpy(&send_buffer1, &port_message, PORT_INFO_SIZE); socket.send_to(boost::asio::buffer(send_buffer1), send_endpoint, {}, _ignored); // Send a request for getting pad data for the pad const Request::PadData pad_data{ - Request::PadData::Flags::Id, - static_cast<u8>(pad_index), + Request::PadData::Flags::AllPorts, + 0, EMPTY_MAC_ADDRESS, }; const auto pad_message = Request::Create(pad_data, client_id); @@ -122,7 +120,6 @@ private: udp::socket socket; const u32 client_id; - std::size_t pad_index{}; static constexpr std::size_t PORT_INFO_SIZE = sizeof(Message<Request::PortInfo>); static constexpr std::size_t PAD_DATA_SIZE = sizeof(Message<Request::PadData>); @@ -150,34 +147,32 @@ Client::~Client() { Reset(); } -Client::ClientData::ClientData() = default; +Client::ClientConnection::ClientConnection() = default; -Client::ClientData::~ClientData() = default; +Client::ClientConnection::~ClientConnection() = default; std::vector<Common::ParamPackage> Client::GetInputDevices() const { std::vector<Common::ParamPackage> devices; - for (std::size_t client = 0; client < clients.size(); client++) { - if (!DeviceConnected(client)) { + for (std::size_t pad = 0; pad < pads.size(); pad++) { + if (!DeviceConnected(pad)) { continue; } - std::string name = fmt::format("UDP Controller {}", client); + std::string name = fmt::format("UDP Controller {}", pad); devices.emplace_back(Common::ParamPackage{ {"class", "cemuhookudp"}, {"display", std::move(name)}, - {"port", std::to_string(client)}, + {"port", std::to_string(pad)}, }); } return devices; } -bool Client::DeviceConnected(std::size_t client) const { +bool Client::DeviceConnected(std::size_t pad) const { // Use last timestamp to detect if the socket has stopped sending data const auto now = std::chrono::steady_clock::now(); - const auto time_difference = - static_cast<u64>(std::chrono::duration_cast<std::chrono::milliseconds>( - now - clients[client].last_motion_update) - .count()); - return time_difference < 1000 && clients[client].active == 1; + const auto time_difference = static_cast<u64>( + std::chrono::duration_cast<std::chrono::milliseconds>(now - pads[pad].last_update).count()); + return time_difference < 1000 && pads[pad].connected; } void Client::ReloadSockets() { @@ -202,25 +197,21 @@ void Client::ReloadSockets() { continue; } - for (std::size_t pad = 0; pad < 4; ++pad) { - const std::size_t client_number = - GetClientNumber(udp_input_address, udp_input_port, pad); - if (client_number != MAX_UDP_CLIENTS) { - LOG_ERROR(Input, "Duplicated UDP servers found"); - continue; - } - StartCommunication(client++, udp_input_address, udp_input_port, pad); + const std::size_t client_number = GetClientNumber(udp_input_address, udp_input_port); + if (client_number != MAX_UDP_CLIENTS) { + LOG_ERROR(Input, "Duplicated UDP servers found"); + continue; } + StartCommunication(client++, udp_input_address, udp_input_port); } } -std::size_t Client::GetClientNumber(std::string_view host, u16 port, std::size_t pad) const { +std::size_t Client::GetClientNumber(std::string_view host, u16 port) const { for (std::size_t client = 0; client < clients.size(); client++) { if (clients[client].active == -1) { continue; } - if (clients[client].host == host && clients[client].port == port && - clients[client].pad_index == pad) { + if (clients[client].host == host && clients[client].port == port) { return client; } } @@ -236,69 +227,75 @@ void Client::OnPortInfo([[maybe_unused]] Response::PortInfo data) { } void Client::OnPadData(Response::PadData data, std::size_t client) { - // Accept packets only for the correct pad - if (static_cast<u8>(clients[client].pad_index) != data.info.id) { + const std::size_t pad_index = (client * PADS_PER_CLIENT) + data.info.id; + + if (pad_index >= pads.size()) { + LOG_ERROR(Input, "Invalid pad id {}", data.info.id); return; } LOG_TRACE(Input, "PadData packet received"); - if (data.packet_counter == clients[client].packet_sequence) { + if (data.packet_counter == pads[pad_index].packet_sequence) { LOG_WARNING( Input, "PadData packet dropped because its stale info. Current count: {} Packet count: {}", - clients[client].packet_sequence, data.packet_counter); + pads[pad_index].packet_sequence, data.packet_counter); + pads[pad_index].connected = false; return; } - clients[client].active = static_cast<s8>(data.info.is_pad_active); - clients[client].packet_sequence = data.packet_counter; + + clients[client].active = 1; + pads[pad_index].connected = true; + pads[pad_index].packet_sequence = data.packet_counter; + const auto now = std::chrono::steady_clock::now(); - const auto time_difference = - static_cast<u64>(std::chrono::duration_cast<std::chrono::microseconds>( - now - clients[client].last_motion_update) - .count()); - clients[client].last_motion_update = now; + const auto time_difference = static_cast<u64>( + std::chrono::duration_cast<std::chrono::microseconds>(now - pads[pad_index].last_update) + .count()); + pads[pad_index].last_update = now; + const Common::Vec3f raw_gyroscope = {data.gyro.pitch, data.gyro.roll, -data.gyro.yaw}; - clients[client].motion.SetAcceleration({data.accel.x, -data.accel.z, data.accel.y}); + pads[pad_index].motion.SetAcceleration({data.accel.x, -data.accel.z, data.accel.y}); // Gyroscope values are not it the correct scale from better joy. // Dividing by 312 allows us to make one full turn = 1 turn // This must be a configurable valued called sensitivity - clients[client].motion.SetGyroscope(raw_gyroscope / 312.0f); - clients[client].motion.UpdateRotation(time_difference); - clients[client].motion.UpdateOrientation(time_difference); + pads[pad_index].motion.SetGyroscope(raw_gyroscope / 312.0f); + pads[pad_index].motion.UpdateRotation(time_difference); + pads[pad_index].motion.UpdateOrientation(time_difference); { - std::lock_guard guard(clients[client].status.update_mutex); - clients[client].status.motion_status = clients[client].motion.GetMotion(); + std::lock_guard guard(pads[pad_index].status.update_mutex); + pads[pad_index].status.motion_status = pads[pad_index].motion.GetMotion(); for (std::size_t id = 0; id < data.touch.size(); ++id) { UpdateTouchInput(data.touch[id], client, id); } if (configuring) { - const Common::Vec3f gyroscope = clients[client].motion.GetGyroscope(); - const Common::Vec3f accelerometer = clients[client].motion.GetAcceleration(); - UpdateYuzuSettings(client, accelerometer, gyroscope); + const Common::Vec3f gyroscope = pads[pad_index].motion.GetGyroscope(); + const Common::Vec3f accelerometer = pads[pad_index].motion.GetAcceleration(); + UpdateYuzuSettings(client, data.info.id, accelerometer, gyroscope); } } } -void Client::StartCommunication(std::size_t client, const std::string& host, u16 port, - std::size_t pad_index) { +void Client::StartCommunication(std::size_t client, const std::string& host, u16 port) { SocketCallback callback{[this](Response::Version version) { OnVersion(version); }, [this](Response::PortInfo info) { OnPortInfo(info); }, [this, client](Response::PadData data) { OnPadData(data, client); }}; - LOG_INFO(Input, "Starting communication with UDP input server on {}:{}:{}", host, port, - pad_index); + LOG_INFO(Input, "Starting communication with UDP input server on {}:{}", host, port); clients[client].host = host; clients[client].port = port; - clients[client].pad_index = pad_index; clients[client].active = 0; - clients[client].socket = std::make_unique<Socket>(host, port, pad_index, callback); + clients[client].socket = std::make_unique<Socket>(host, port, callback); clients[client].thread = std::thread{SocketLoop, clients[client].socket.get()}; + // Set motion parameters // SetGyroThreshold value should be dependent on GyroscopeZeroDriftMode // Real HW values are unknown, 0.0001 is an approximate to Standard - clients[client].motion.SetGyroThreshold(0.0001f); + for (std::size_t pad = 0; pad < PADS_PER_CLIENT; pad++) { + pads[client * PADS_PER_CLIENT + pad].motion.SetGyroThreshold(0.0001f); + } } void Client::Reset() { @@ -311,8 +308,8 @@ void Client::Reset() { } } -void Client::UpdateYuzuSettings(std::size_t client, const Common::Vec3<float>& acc, - const Common::Vec3<float>& gyro) { +void Client::UpdateYuzuSettings(std::size_t client, std::size_t pad_index, + const Common::Vec3<float>& acc, const Common::Vec3<float>& gyro) { if (gyro.Length() > 0.2f) { LOG_DEBUG(Input, "UDP Controller {}: gyro=({}, {}, {}), accel=({}, {}, {})", client, gyro[0], gyro[1], gyro[2], acc[0], acc[1], acc[2]); @@ -320,7 +317,7 @@ void Client::UpdateYuzuSettings(std::size_t client, const Common::Vec3<float>& a UDPPadStatus pad{ .host = clients[client].host, .port = clients[client].port, - .pad_index = clients[client].pad_index, + .pad_index = pad_index, }; for (std::size_t i = 0; i < 3; ++i) { if (gyro[i] > 5.0f || gyro[i] < -5.0f) { @@ -391,19 +388,19 @@ void Client::EndConfiguration() { } DeviceStatus& Client::GetPadState(const std::string& host, u16 port, std::size_t pad) { - const std::size_t client_number = GetClientNumber(host, port, pad); - if (client_number == MAX_UDP_CLIENTS) { - return clients[0].status; + const std::size_t client_number = GetClientNumber(host, port); + if (client_number == MAX_UDP_CLIENTS || pad >= PADS_PER_CLIENT) { + return pads[0].status; } - return clients[client_number].status; + return pads[(client_number * PADS_PER_CLIENT) + pad].status; } const DeviceStatus& Client::GetPadState(const std::string& host, u16 port, std::size_t pad) const { - const std::size_t client_number = GetClientNumber(host, port, pad); - if (client_number == MAX_UDP_CLIENTS) { - return clients[0].status; + const std::size_t client_number = GetClientNumber(host, port); + if (client_number == MAX_UDP_CLIENTS || pad >= PADS_PER_CLIENT) { + return pads[0].status; } - return clients[client_number].status; + return pads[(client_number * PADS_PER_CLIENT) + pad].status; } Input::TouchStatus& Client::GetTouchState() { @@ -422,7 +419,7 @@ const Common::SPSCQueue<UDPPadStatus>& Client::GetPadQueue() const { return pad_queue; } -void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, +void TestCommunication(const std::string& host, u16 port, const std::function<void()>& success_callback, const std::function<void()>& failure_callback) { std::thread([=] { @@ -432,9 +429,10 @@ void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, .port_info = [](Response::PortInfo) {}, .pad_data = [&](Response::PadData) { success_event.Set(); }, }; - Socket socket{host, port, pad_index, std::move(callback)}; + Socket socket{host, port, std::move(callback)}; std::thread worker_thread{SocketLoop, &socket}; - const bool result = success_event.WaitFor(std::chrono::seconds(5)); + const bool result = + success_event.WaitUntil(std::chrono::steady_clock::now() + std::chrono::seconds(10)); socket.Stop(); worker_thread.join(); if (result) { @@ -446,8 +444,7 @@ void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, } CalibrationConfigurationJob::CalibrationConfigurationJob( - const std::string& host, u16 port, std::size_t pad_index, - std::function<void(Status)> status_callback, + const std::string& host, u16 port, std::function<void(Status)> status_callback, std::function<void(u16, u16, u16, u16)> data_callback) { std::thread([=, this] { @@ -491,7 +488,7 @@ CalibrationConfigurationJob::CalibrationConfigurationJob( complete_event.Set(); } }}; - Socket socket{host, port, pad_index, std::move(callback)}; + Socket socket{host, port, std::move(callback)}; std::thread worker_thread{SocketLoop, &socket}; complete_event.Wait(); socket.Stop(); diff --git a/src/input_common/udp/client.h b/src/input_common/udp/client.h index e9e438e88..a11ea3068 100644 --- a/src/input_common/udp/client.h +++ b/src/input_common/udp/client.h @@ -84,7 +84,7 @@ public: std::vector<Common::ParamPackage> GetInputDevices() const; - bool DeviceConnected(std::size_t client) const; + bool DeviceConnected(std::size_t pad) const; void ReloadSockets(); Common::SPSCQueue<UDPPadStatus>& GetPadQueue(); @@ -97,38 +97,40 @@ public: const Input::TouchStatus& GetTouchState() const; private: - struct ClientData { - ClientData(); - ~ClientData(); - - std::string host{"127.0.0.1"}; - u16 port{26760}; + struct PadData { std::size_t pad_index{}; - std::unique_ptr<Socket> socket; + bool connected{}; DeviceStatus status; - std::thread thread; u64 packet_sequence{}; - s8 active{-1}; // Realtime values // motion is initalized with PID values for drift correction on joycons InputCommon::MotionInput motion{0.3f, 0.005f, 0.0f}; - std::chrono::time_point<std::chrono::steady_clock> last_motion_update; + std::chrono::time_point<std::chrono::steady_clock> last_update; + }; + + struct ClientConnection { + ClientConnection(); + ~ClientConnection(); + std::string host{"127.0.0.1"}; + u16 port{26760}; + s8 active{-1}; + std::unique_ptr<Socket> socket; + std::thread thread; }; // For shutting down, clear all data, join all threads, release usb void Reset(); // Translates configuration to client number - std::size_t GetClientNumber(std::string_view host, u16 port, std::size_t pad) const; + std::size_t GetClientNumber(std::string_view host, u16 port) const; void OnVersion(Response::Version); void OnPortInfo(Response::PortInfo); void OnPadData(Response::PadData, std::size_t client); - void StartCommunication(std::size_t client, const std::string& host, u16 port, - std::size_t pad_index); - void UpdateYuzuSettings(std::size_t client, const Common::Vec3<float>& acc, - const Common::Vec3<float>& gyro); + void StartCommunication(std::size_t client, const std::string& host, u16 port); + void UpdateYuzuSettings(std::size_t client, std::size_t pad_index, + const Common::Vec3<float>& acc, const Common::Vec3<float>& gyro); // Returns an unused finger id, if there is no fingers available std::nullopt will be // returned @@ -140,10 +142,12 @@ private: bool configuring = false; // Allocate clients for 8 udp servers - static constexpr std::size_t MAX_UDP_CLIENTS = 4 * 8; + static constexpr std::size_t MAX_UDP_CLIENTS = 8; + static constexpr std::size_t PADS_PER_CLIENT = 4; // Each client can have up 2 touch inputs static constexpr std::size_t MAX_TOUCH_FINGERS = MAX_UDP_CLIENTS * 2; - std::array<ClientData, MAX_UDP_CLIENTS> clients{}; + std::array<PadData, MAX_UDP_CLIENTS * PADS_PER_CLIENT> pads{}; + std::array<ClientConnection, MAX_UDP_CLIENTS> clients{}; Common::SPSCQueue<UDPPadStatus> pad_queue{}; Input::TouchStatus touch_status{}; std::array<std::size_t, MAX_TOUCH_FINGERS> finger_id{}; @@ -164,7 +168,7 @@ public: * @param status_callback Callback for job status updates * @param data_callback Called when calibration data is ready */ - explicit CalibrationConfigurationJob(const std::string& host, u16 port, std::size_t pad_index, + explicit CalibrationConfigurationJob(const std::string& host, u16 port, std::function<void(Status)> status_callback, std::function<void(u16, u16, u16, u16)> data_callback); ~CalibrationConfigurationJob(); @@ -174,7 +178,7 @@ private: Common::Event complete_event; }; -void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, +void TestCommunication(const std::string& host, u16 port, const std::function<void()>& success_callback, const std::function<void()>& failure_callback); diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index 9b931976a..47190c464 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -236,7 +236,6 @@ add_library(video_core STATIC texture_cache/types.h texture_cache/util.cpp texture_cache/util.h - textures/astc.cpp textures/astc.h textures/decoders.cpp textures/decoders.h diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 59e586695..29bb31418 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -2,8 +2,9 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cstring> // for std::memcpy +#include <algorithm> // for std::copy #include <numeric> +#include "common/assert.h" #include "video_core/command_classes/codecs/vp9.h" #include "video_core/gpu.h" #include "video_core/memory_manager.h" @@ -362,7 +363,8 @@ Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) // surface_luma_offset[0:3] contains the address of the reference frame offsets in the following // order: last, golden, altref, current. It may be worthwhile to track the updates done here // to avoid buffering frame data needed for reference frame updating in the header composition. - std::memcpy(vp9_info.frame_offsets.data(), state.surface_luma_offset.data(), 4 * sizeof(u64)); + std::copy(state.surface_luma_offset.begin(), state.surface_luma_offset.begin() + 4, + vp9_info.frame_offsets.begin()); return vp9_info; } @@ -821,11 +823,11 @@ const std::vector<u8>& VP9::ComposeFrameHeader(const NvdecCommon::NvdecRegisters // Write headers and frame to buffer frame.resize(uncompressed_header.size() + compressed_header.size() + bitstream.size()); - std::memcpy(frame.data(), uncompressed_header.data(), uncompressed_header.size()); - std::memcpy(frame.data() + uncompressed_header.size(), compressed_header.data(), - compressed_header.size()); - std::memcpy(frame.data() + uncompressed_header.size() + compressed_header.size(), - bitstream.data(), bitstream.size()); + std::copy(uncompressed_header.begin(), uncompressed_header.end(), frame.begin()); + std::copy(compressed_header.begin(), compressed_header.end(), + frame.begin() + uncompressed_header.size()); + std::copy(bitstream.begin(), bitstream.end(), + frame.begin() + uncompressed_header.size() + compressed_header.size()); // keep track of frame number current_frame_number++; diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp index c61f44619..009c6f574 100644 --- a/src/video_core/gpu.cpp +++ b/src/video_core/gpu.cpp @@ -517,8 +517,8 @@ void GPU::TriggerCpuInterrupt(const u32 syncpoint_id, const u32 value) const { interrupt_manager.GPUInterruptSyncpt(syncpoint_id, value); } -void GPU::WaitIdle() const { - gpu_thread.WaitIdle(); +void GPU::ShutDown() { + gpu_thread.ShutDown(); } void GPU::OnCommandListEnd() { diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h index b2ee45496..ecab35d3b 100644 --- a/src/video_core/gpu.h +++ b/src/video_core/gpu.h @@ -219,8 +219,8 @@ public: return *shader_notify; } - // Waits for the GPU to finish working - void WaitIdle() const; + // Stops the GPU execution and waits for the GPU to finish working + void ShutDown(); /// Allows the CPU/NvFlinger to wait on the GPU before presenting a frame. void WaitFence(u32 syncpoint_id, u32 value); diff --git a/src/video_core/gpu_thread.cpp b/src/video_core/gpu_thread.cpp index 99353f15f..7addfbc7b 100644 --- a/src/video_core/gpu_thread.cpp +++ b/src/video_core/gpu_thread.cpp @@ -29,8 +29,7 @@ static void RunThread(Core::System& system, VideoCore::RendererBase& renderer, system.RegisterHostThread(); // Wait for first GPU command before acquiring the window context - while (state.queue.Empty()) - ; + state.queue.Wait(); // If emulation was stopped during disk shader loading, abort before trying to acquire context if (!state.is_running) { @@ -57,11 +56,17 @@ static void RunThread(Core::System& system, VideoCore::RendererBase& renderer, } else if (const auto* invalidate = std::get_if<InvalidateRegionCommand>(&next.data)) { rasterizer->OnCPUWrite(invalidate->addr, invalidate->size); } else if (std::holds_alternative<EndProcessingCommand>(next.data)) { - return; + ASSERT(state.is_running == false); } else { UNREACHABLE(); } state.signaled_fence.store(next.fence); + if (next.block) { + // We have to lock the write_lock to ensure that the condition_variable wait not get a + // race between the check and the lock itself. + std::lock_guard lk(state.write_lock); + state.cv.notify_all(); + } } } @@ -69,13 +74,7 @@ ThreadManager::ThreadManager(Core::System& system_, bool is_async_) : system{system_}, is_async{is_async_} {} ThreadManager::~ThreadManager() { - if (!thread.joinable()) { - return; - } - - // Notify GPU thread that a shutdown is pending - PushCommand(EndProcessingCommand()); - thread.join(); + ShutDown(); } void ThreadManager::StartThread(VideoCore::RendererBase& renderer, @@ -112,9 +111,8 @@ void ThreadManager::FlushRegion(VAddr addr, u64 size) { case Settings::GPUAccuracy::Extreme: { auto& gpu = system.GPU(); u64 fence = gpu.RequestFlush(addr, size); - PushCommand(GPUTickCommand()); - while (fence > gpu.CurrentFlushRequestFence()) { - } + PushCommand(GPUTickCommand(), true); + ASSERT(fence <= gpu.CurrentFlushRequestFence()); break; } default: @@ -131,23 +129,45 @@ void ThreadManager::FlushAndInvalidateRegion(VAddr addr, u64 size) { rasterizer->OnCPUWrite(addr, size); } -void ThreadManager::WaitIdle() const { - while (state.last_fence > state.signaled_fence.load(std::memory_order_relaxed) && - system.IsPoweredOn()) { +void ThreadManager::ShutDown() { + if (!state.is_running) { + return; } + + { + std::lock_guard lk(state.write_lock); + state.is_running = false; + state.cv.notify_all(); + } + + if (!thread.joinable()) { + return; + } + + // Notify GPU thread that a shutdown is pending + PushCommand(EndProcessingCommand()); + thread.join(); } void ThreadManager::OnCommandListEnd() { PushCommand(OnCommandListEndCommand()); } -u64 ThreadManager::PushCommand(CommandData&& command_data) { - const u64 fence{++state.last_fence}; - state.queue.Push(CommandDataContainer(std::move(command_data), fence)); - +u64 ThreadManager::PushCommand(CommandData&& command_data, bool block) { if (!is_async) { // In synchronous GPU mode, block the caller until the command has executed - WaitIdle(); + block = true; + } + + std::unique_lock lk(state.write_lock); + const u64 fence{++state.last_fence}; + state.queue.Push(CommandDataContainer(std::move(command_data), fence, block)); + + if (block) { + state.cv.wait(lk, [this, fence] { + return fence <= state.signaled_fence.load(std::memory_order_relaxed) || + !state.is_running; + }); } return fence; diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index 18269e51c..11a648f38 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -90,21 +90,24 @@ using CommandData = struct CommandDataContainer { CommandDataContainer() = default; - explicit CommandDataContainer(CommandData&& data_, u64 next_fence_) - : data{std::move(data_)}, fence{next_fence_} {} + explicit CommandDataContainer(CommandData&& data_, u64 next_fence_, bool block_) + : data{std::move(data_)}, fence{next_fence_}, block(block_) {} CommandData data; u64 fence{}; + bool block{}; }; /// Struct used to synchronize the GPU thread struct SynchState final { std::atomic_bool is_running{true}; - using CommandQueue = Common::MPSCQueue<CommandDataContainer>; + using CommandQueue = Common::SPSCQueue<CommandDataContainer>; + std::mutex write_lock; CommandQueue queue; u64 last_fence{}; std::atomic<u64> signaled_fence{}; + std::condition_variable cv; }; /// Class used to manage the GPU thread @@ -132,14 +135,14 @@ public: /// Notify rasterizer that any caches of the specified region should be flushed and invalidated void FlushAndInvalidateRegion(VAddr addr, u64 size); - // Wait until the gpu thread is idle. - void WaitIdle() const; + // Stops the GPU execution and waits for the GPU to finish working + void ShutDown(); void OnCommandListEnd(); private: /// Pushes a command to be executed by the GPU thread - u64 PushCommand(CommandData&& command_data); + u64 PushCommand(CommandData&& command_data, bool block = false); Core::System& system; const bool is_async; diff --git a/src/video_core/host_shaders/CMakeLists.txt b/src/video_core/host_shaders/CMakeLists.txt index 3494318ca..2208e1922 100644 --- a/src/video_core/host_shaders/CMakeLists.txt +++ b/src/video_core/host_shaders/CMakeLists.txt @@ -1,4 +1,5 @@ set(SHADER_FILES + astc_decoder.comp block_linear_unswizzle_2d.comp block_linear_unswizzle_3d.comp convert_depth_to_float.frag diff --git a/src/video_core/host_shaders/StringShaderHeader.cmake b/src/video_core/host_shaders/StringShaderHeader.cmake index c0fc49768..1b4bc6103 100644 --- a/src/video_core/host_shaders/StringShaderHeader.cmake +++ b/src/video_core/host_shaders/StringShaderHeader.cmake @@ -6,7 +6,27 @@ get_filename_component(CONTENTS_NAME ${SOURCE_FILE} NAME) string(REPLACE "." "_" CONTENTS_NAME ${CONTENTS_NAME}) string(TOUPPER ${CONTENTS_NAME} CONTENTS_NAME) -file(READ ${SOURCE_FILE} CONTENTS) +FILE(READ ${SOURCE_FILE} line_contents) + +# Replace double quotes with single quotes, +# as double quotes will be used to wrap the lines +STRING(REGEX REPLACE "\"" "'" line_contents "${line_contents}") + +# CMake separates list elements with semicolons, but semicolons +# are used extensively in the shader code. +# Replace with a temporary marker, to be reverted later. +STRING(REGEX REPLACE ";" "{{SEMICOLON}}" line_contents "${line_contents}") + +# Make every line an individual element in the CMake list. +STRING(REGEX REPLACE "\n" ";" line_contents "${line_contents}") + +# Build the shader string, wrapping each line in double quotes. +foreach(line IN LISTS line_contents) + string(CONCAT CONTENTS "${CONTENTS}" \"${line}\\n\"\n) +endforeach() + +# Revert the original semicolons in the source. +STRING(REGEX REPLACE "{{SEMICOLON}}" ";" CONTENTS "${CONTENTS}") get_filename_component(OUTPUT_DIR ${HEADER_FILE} DIRECTORY) make_directory(${OUTPUT_DIR}) diff --git a/src/video_core/host_shaders/astc_decoder.comp b/src/video_core/host_shaders/astc_decoder.comp new file mode 100644 index 000000000..703e34587 --- /dev/null +++ b/src/video_core/host_shaders/astc_decoder.comp @@ -0,0 +1,1339 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 + +#ifdef VULKAN + +#define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants { +#define END_PUSH_CONSTANTS }; +#define UNIFORM(n) +#define BINDING_INPUT_BUFFER 0 +#define BINDING_ENC_BUFFER 1 +#define BINDING_6_TO_8_BUFFER 2 +#define BINDING_7_TO_8_BUFFER 3 +#define BINDING_8_TO_8_BUFFER 4 +#define BINDING_BYTE_TO_16_BUFFER 5 +#define BINDING_SWIZZLE_BUFFER 6 +#define BINDING_OUTPUT_IMAGE 7 + +#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv + +#define BEGIN_PUSH_CONSTANTS +#define END_PUSH_CONSTANTS +#define UNIFORM(n) layout(location = n) uniform +#define BINDING_SWIZZLE_BUFFER 0 +#define BINDING_INPUT_BUFFER 1 +#define BINDING_ENC_BUFFER 2 +#define BINDING_6_TO_8_BUFFER 3 +#define BINDING_7_TO_8_BUFFER 4 +#define BINDING_8_TO_8_BUFFER 5 +#define BINDING_BYTE_TO_16_BUFFER 6 +#define BINDING_OUTPUT_IMAGE 0 + +#endif + +layout(local_size_x = 32, local_size_y = 32, local_size_z = 1) in; + +BEGIN_PUSH_CONSTANTS +UNIFORM(1) uvec2 block_dims; + +UNIFORM(2) uint bytes_per_block_log2; +UNIFORM(3) uint layer_stride; +UNIFORM(4) uint block_size; +UNIFORM(5) uint x_shift; +UNIFORM(6) uint block_height; +UNIFORM(7) uint block_height_mask; +END_PUSH_CONSTANTS + +struct EncodingData { + uint encoding; + uint num_bits; + uint bit_value; + uint quint_trit_value; +}; + +struct TexelWeightParams { + uvec2 size; + uint max_weight; + bool dual_plane; + bool error_state; + bool void_extent_ldr; + bool void_extent_hdr; +}; + +// Swizzle data +layout(binding = BINDING_SWIZZLE_BUFFER, std430) readonly buffer SwizzleTable { + uint swizzle_table[]; +}; + +layout(binding = BINDING_INPUT_BUFFER, std430) readonly buffer InputBufferU32 { + uint astc_data[]; +}; + +// ASTC Encodings data +layout(binding = BINDING_ENC_BUFFER, std430) readonly buffer EncodingsValues { + EncodingData encoding_values[]; +}; +// ASTC Precompiled tables +layout(binding = BINDING_6_TO_8_BUFFER, std430) readonly buffer REPLICATE_6_BIT_TO_8 { + uint REPLICATE_6_BIT_TO_8_TABLE[]; +}; +layout(binding = BINDING_7_TO_8_BUFFER, std430) readonly buffer REPLICATE_7_BIT_TO_8 { + uint REPLICATE_7_BIT_TO_8_TABLE[]; +}; +layout(binding = BINDING_8_TO_8_BUFFER, std430) readonly buffer REPLICATE_8_BIT_TO_8 { + uint REPLICATE_8_BIT_TO_8_TABLE[]; +}; +layout(binding = BINDING_BYTE_TO_16_BUFFER, std430) readonly buffer REPLICATE_BYTE_TO_16 { + uint REPLICATE_BYTE_TO_16_TABLE[]; +}; + +layout(binding = BINDING_OUTPUT_IMAGE, rgba8) uniform writeonly image2DArray dest_image; + +const uint GOB_SIZE_X = 64; +const uint GOB_SIZE_Y = 8; +const uint GOB_SIZE_Z = 1; +const uint GOB_SIZE = GOB_SIZE_X * GOB_SIZE_Y * GOB_SIZE_Z; + +const uint GOB_SIZE_X_SHIFT = 6; +const uint GOB_SIZE_Y_SHIFT = 3; +const uint GOB_SIZE_Z_SHIFT = 0; +const uint GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHIFT; + +const uvec2 SWIZZLE_MASK = uvec2(GOB_SIZE_X - 1, GOB_SIZE_Y - 1); + +const int BLOCK_SIZE_IN_BYTES = 16; + +const int BLOCK_INFO_ERROR = 0; +const int BLOCK_INFO_VOID_EXTENT_HDR = 1; +const int BLOCK_INFO_VOID_EXTENT_LDR = 2; +const int BLOCK_INFO_NORMAL = 3; + +const int JUST_BITS = 0; +const int QUINT = 1; +const int TRIT = 2; + +// The following constants are expanded variants of the Replicate() +// function calls corresponding to the following arguments: +// value: index into the generated table +// num_bits: the after "REPLICATE" in the table name. i.e. 4 is num_bits in REPLICATE_4. +// to_bit: the integer after "TO_" +const uint REPLICATE_BIT_TO_7_TABLE[2] = uint[](0, 127); +const uint REPLICATE_1_BIT_TO_9_TABLE[2] = uint[](0, 511); + +const uint REPLICATE_1_BIT_TO_8_TABLE[2] = uint[](0, 255); +const uint REPLICATE_2_BIT_TO_8_TABLE[4] = uint[](0, 85, 170, 255); +const uint REPLICATE_3_BIT_TO_8_TABLE[8] = uint[](0, 36, 73, 109, 146, 182, 219, 255); +const uint REPLICATE_4_BIT_TO_8_TABLE[16] = + uint[](0, 17, 34, 51, 68, 85, 102, 119, 136, 153, 170, 187, 204, 221, 238, 255); +const uint REPLICATE_5_BIT_TO_8_TABLE[32] = + uint[](0, 8, 16, 24, 33, 41, 49, 57, 66, 74, 82, 90, 99, 107, 115, 123, 132, 140, 148, 156, 165, + 173, 181, 189, 198, 206, 214, 222, 231, 239, 247, 255); +const uint REPLICATE_1_BIT_TO_6_TABLE[2] = uint[](0, 63); +const uint REPLICATE_2_BIT_TO_6_TABLE[4] = uint[](0, 21, 42, 63); +const uint REPLICATE_3_BIT_TO_6_TABLE[8] = uint[](0, 9, 18, 27, 36, 45, 54, 63); +const uint REPLICATE_4_BIT_TO_6_TABLE[16] = + uint[](0, 4, 8, 12, 17, 21, 25, 29, 34, 38, 42, 46, 51, 55, 59, 63); +const uint REPLICATE_5_BIT_TO_6_TABLE[32] = + uint[](0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 33, 35, 37, 39, 41, 43, 45, + 47, 49, 51, 53, 55, 57, 59, 61, 63); + +// Input ASTC texture globals +uint current_index = 0; +int bitsread = 0; +uint total_bitsread = 0; +uint local_buff[16]; + +// Color data globals +uint color_endpoint_data[16]; +int color_bitsread = 0; +uint total_color_bitsread = 0; +int color_index = 0; + +// Four values, two endpoints, four maximum paritions +uint color_values[32]; +int colvals_index = 0; + +// Weight data globals +uint texel_weight_data[16]; +int texel_bitsread = 0; +uint total_texel_bitsread = 0; +int texel_index = 0; + +bool texel_flag = false; + +// Global "vectors" to be pushed into when decoding +EncodingData result_vector[100]; +int result_index = 0; + +EncodingData texel_vector[100]; +int texel_vector_index = 0; + +uint unquantized_texel_weights[2][144]; + +uint SwizzleOffset(uvec2 pos) { + pos = pos & SWIZZLE_MASK; + return swizzle_table[pos.y * 64 + pos.x]; +} + +uint ReadTexel(uint offset) { + // extract the 8-bit value from the 32-bit packed data. + return bitfieldExtract(astc_data[offset / 4], int((offset * 8) & 24), 8); +} + +// Replicates low num_bits such that [(to_bit - 1):(to_bit - 1 - from_bit)] +// is the same as [(num_bits - 1):0] and repeats all the way down. +uint Replicate(uint val, uint num_bits, uint to_bit) { + if (num_bits == 0 || to_bit == 0) { + return 0; + } + const uint v = val & uint((1 << num_bits) - 1); + uint res = v; + uint reslen = num_bits; + while (reslen < to_bit) { + uint comp = 0; + if (num_bits > to_bit - reslen) { + uint newshift = to_bit - reslen; + comp = num_bits - newshift; + num_bits = newshift; + } + res = uint(res << num_bits); + res = uint(res | (v >> comp)); + reslen += num_bits; + } + return res; +} + +uvec4 ReplicateByteTo16(uvec4 value) { + return uvec4(REPLICATE_BYTE_TO_16_TABLE[value.x], REPLICATE_BYTE_TO_16_TABLE[value.y], + REPLICATE_BYTE_TO_16_TABLE[value.z], REPLICATE_BYTE_TO_16_TABLE[value.w]); +} + +uint ReplicateBitTo7(uint value) { + return REPLICATE_BIT_TO_7_TABLE[value]; +} + +uint ReplicateBitTo9(uint value) { + return REPLICATE_1_BIT_TO_9_TABLE[value]; +} + +uint FastReplicateTo8(uint value, uint num_bits) { + switch (num_bits) { + case 1: + return REPLICATE_1_BIT_TO_8_TABLE[value]; + case 2: + return REPLICATE_2_BIT_TO_8_TABLE[value]; + case 3: + return REPLICATE_3_BIT_TO_8_TABLE[value]; + case 4: + return REPLICATE_4_BIT_TO_8_TABLE[value]; + case 5: + return REPLICATE_5_BIT_TO_8_TABLE[value]; + case 6: + return REPLICATE_6_BIT_TO_8_TABLE[value]; + case 7: + return REPLICATE_7_BIT_TO_8_TABLE[value]; + case 8: + return REPLICATE_8_BIT_TO_8_TABLE[value]; + } + return Replicate(value, num_bits, 8); +} + +uint FastReplicateTo6(uint value, uint num_bits) { + switch (num_bits) { + case 1: + return REPLICATE_1_BIT_TO_6_TABLE[value]; + case 2: + return REPLICATE_2_BIT_TO_6_TABLE[value]; + case 3: + return REPLICATE_3_BIT_TO_6_TABLE[value]; + case 4: + return REPLICATE_4_BIT_TO_6_TABLE[value]; + case 5: + return REPLICATE_5_BIT_TO_6_TABLE[value]; + } + return Replicate(value, num_bits, 6); +} + +uint Div3Floor(uint v) { + return (v * 0x5556) >> 16; +} + +uint Div3Ceil(uint v) { + return Div3Floor(v + 2); +} + +uint Div5Floor(uint v) { + return (v * 0x3334) >> 16; +} + +uint Div5Ceil(uint v) { + return Div5Floor(v + 4); +} + +uint Hash52(uint p) { + p ^= p >> 15; + p -= p << 17; + p += p << 7; + p += p << 4; + p ^= p >> 5; + p += p << 16; + p ^= p >> 7; + p ^= p >> 3; + p ^= p << 6; + p ^= p >> 17; + return p; +} + +uint SelectPartition(uint seed, uint x, uint y, uint z, uint partition_count, bool small_block) { + if (partition_count == 1) { + return 0; + } + if (small_block) { + x <<= 1; + y <<= 1; + z <<= 1; + } + + seed += (partition_count - 1) * 1024; + + uint rnum = Hash52(uint(seed)); + uint seed1 = uint(rnum & 0xF); + uint seed2 = uint((rnum >> 4) & 0xF); + uint seed3 = uint((rnum >> 8) & 0xF); + uint seed4 = uint((rnum >> 12) & 0xF); + uint seed5 = uint((rnum >> 16) & 0xF); + uint seed6 = uint((rnum >> 20) & 0xF); + uint seed7 = uint((rnum >> 24) & 0xF); + uint seed8 = uint((rnum >> 28) & 0xF); + uint seed9 = uint((rnum >> 18) & 0xF); + uint seed10 = uint((rnum >> 22) & 0xF); + uint seed11 = uint((rnum >> 26) & 0xF); + uint seed12 = uint(((rnum >> 30) | (rnum << 2)) & 0xF); + + seed1 = (seed1 * seed1); + seed2 = (seed2 * seed2); + seed3 = (seed3 * seed3); + seed4 = (seed4 * seed4); + seed5 = (seed5 * seed5); + seed6 = (seed6 * seed6); + seed7 = (seed7 * seed7); + seed8 = (seed8 * seed8); + seed9 = (seed9 * seed9); + seed10 = (seed10 * seed10); + seed11 = (seed11 * seed11); + seed12 = (seed12 * seed12); + + int sh1, sh2, sh3; + if ((seed & 1) > 0) { + sh1 = (seed & 2) > 0 ? 4 : 5; + sh2 = (partition_count == 3) ? 6 : 5; + } else { + sh1 = (partition_count == 3) ? 6 : 5; + sh2 = (seed & 2) > 0 ? 4 : 5; + } + sh3 = (seed & 0x10) > 0 ? sh1 : sh2; + + seed1 = (seed1 >> sh1); + seed2 = (seed2 >> sh2); + seed3 = (seed3 >> sh1); + seed4 = (seed4 >> sh2); + seed5 = (seed5 >> sh1); + seed6 = (seed6 >> sh2); + seed7 = (seed7 >> sh1); + seed8 = (seed8 >> sh2); + seed9 = (seed9 >> sh3); + seed10 = (seed10 >> sh3); + seed11 = (seed11 >> sh3); + seed12 = (seed12 >> sh3); + + uint a = seed1 * x + seed2 * y + seed11 * z + (rnum >> 14); + uint b = seed3 * x + seed4 * y + seed12 * z + (rnum >> 10); + uint c = seed5 * x + seed6 * y + seed9 * z + (rnum >> 6); + uint d = seed7 * x + seed8 * y + seed10 * z + (rnum >> 2); + + a &= 0x3F; + b &= 0x3F; + c &= 0x3F; + d &= 0x3F; + + if (partition_count < 4) { + d = 0; + } + if (partition_count < 3) { + c = 0; + } + + if (a >= b && a >= c && a >= d) { + return 0; + } else if (b >= c && b >= d) { + return 1; + } else if (c >= d) { + return 2; + } else { + return 3; + } +} + +uint Select2DPartition(uint seed, uint x, uint y, uint partition_count, bool small_block) { + return SelectPartition(seed, x, y, 0, partition_count, small_block); +} + +uint ReadBit() { + if (current_index >= local_buff.length()) { + return 0; + } + uint bit = bitfieldExtract(local_buff[current_index], bitsread, 1); + ++bitsread; + ++total_bitsread; + if (bitsread == 8) { + ++current_index; + bitsread = 0; + } + return bit; +} + +uint StreamBits(uint num_bits) { + uint ret = 0; + for (uint i = 0; i < num_bits; i++) { + ret |= ((ReadBit() & 1) << i); + } + return ret; +} + +uint ReadColorBit() { + uint bit = 0; + if (texel_flag) { + bit = bitfieldExtract(texel_weight_data[texel_index], texel_bitsread, 1); + ++texel_bitsread; + ++total_texel_bitsread; + if (texel_bitsread == 8) { + ++texel_index; + texel_bitsread = 0; + } + } else { + bit = bitfieldExtract(color_endpoint_data[color_index], color_bitsread, 1); + ++color_bitsread; + ++total_color_bitsread; + if (color_bitsread == 8) { + ++color_index; + color_bitsread = 0; + } + } + return bit; +} + +uint StreamColorBits(uint num_bits) { + uint ret = 0; + for (uint i = 0; i < num_bits; i++) { + ret |= ((ReadColorBit() & 1) << i); + } + return ret; +} + +void ResultEmplaceBack(EncodingData val) { + if (texel_flag) { + texel_vector[texel_vector_index] = val; + ++texel_vector_index; + } else { + result_vector[result_index] = val; + ++result_index; + } +} + +// Returns the number of bits required to encode n_vals values. +uint GetBitLength(uint n_vals, uint encoding_index) { + uint total_bits = encoding_values[encoding_index].num_bits * n_vals; + if (encoding_values[encoding_index].encoding == TRIT) { + total_bits += Div5Ceil(n_vals * 8); + } else if (encoding_values[encoding_index].encoding == QUINT) { + total_bits += Div3Ceil(n_vals * 7); + } + return total_bits; +} + +uint GetNumWeightValues(uvec2 size, bool dual_plane) { + uint n_vals = size.x * size.y; + if (dual_plane) { + n_vals *= 2; + } + return n_vals; +} + +uint GetPackedBitSize(uvec2 size, bool dual_plane, uint max_weight) { + uint n_vals = GetNumWeightValues(size, dual_plane); + return GetBitLength(n_vals, max_weight); +} + +uint BitsBracket(uint bits, uint pos) { + return ((bits >> pos) & 1); +} + +uint BitsOp(uint bits, uint start, uint end) { + if (start == end) { + return BitsBracket(bits, start); + } else if (start > end) { + uint t = start; + start = end; + end = t; + } + + uint mask = (1 << (end - start + 1)) - 1; + return ((bits >> start) & mask); +} + +void DecodeQuintBlock(uint num_bits) { + uint m[3]; + uint q[3]; + uint Q; + m[0] = StreamColorBits(num_bits); + Q = StreamColorBits(3); + m[1] = StreamColorBits(num_bits); + Q |= StreamColorBits(2) << 3; + m[2] = StreamColorBits(num_bits); + Q |= StreamColorBits(2) << 5; + if (BitsOp(Q, 1, 2) == 3 && BitsOp(Q, 5, 6) == 0) { + q[0] = 4; + q[1] = 4; + q[2] = (BitsBracket(Q, 0) << 2) | ((BitsBracket(Q, 4) & ~BitsBracket(Q, 0)) << 1) | + (BitsBracket(Q, 3) & ~BitsBracket(Q, 0)); + } else { + uint C = 0; + if (BitsOp(Q, 1, 2) == 3) { + q[2] = 4; + C = (BitsOp(Q, 3, 4) << 3) | ((~BitsOp(Q, 5, 6) & 3) << 1) | BitsBracket(Q, 0); + } else { + q[2] = BitsOp(Q, 5, 6); + C = BitsOp(Q, 0, 4); + } + if (BitsOp(C, 0, 2) == 5) { + q[1] = 4; + q[0] = BitsOp(C, 3, 4); + } else { + q[1] = BitsOp(C, 3, 4); + q[0] = BitsOp(C, 0, 2); + } + } + for (uint i = 0; i < 3; i++) { + EncodingData val; + val.encoding = QUINT; + val.num_bits = num_bits; + val.bit_value = m[i]; + val.quint_trit_value = q[i]; + ResultEmplaceBack(val); + } +} + +void DecodeTritBlock(uint num_bits) { + uint m[5]; + uint t[5]; + uint T; + m[0] = StreamColorBits(num_bits); + T = StreamColorBits(2); + m[1] = StreamColorBits(num_bits); + T |= StreamColorBits(2) << 2; + m[2] = StreamColorBits(num_bits); + T |= StreamColorBits(1) << 4; + m[3] = StreamColorBits(num_bits); + T |= StreamColorBits(2) << 5; + m[4] = StreamColorBits(num_bits); + T |= StreamColorBits(1) << 7; + uint C = 0; + if (BitsOp(T, 2, 4) == 7) { + C = (BitsOp(T, 5, 7) << 2) | BitsOp(T, 0, 1); + t[4] = 2; + t[3] = 2; + } else { + C = BitsOp(T, 0, 4); + if (BitsOp(T, 5, 6) == 3) { + t[4] = 2; + t[3] = BitsBracket(T, 7); + } else { + t[4] = BitsBracket(T, 7); + t[3] = BitsOp(T, 5, 6); + } + } + if (BitsOp(C, 0, 1) == 3) { + t[2] = 2; + t[1] = BitsBracket(C, 4); + t[0] = (BitsBracket(C, 3) << 1) | (BitsBracket(C, 2) & ~BitsBracket(C, 3)); + } else if (BitsOp(C, 2, 3) == 3) { + t[2] = 2; + t[1] = 2; + t[0] = BitsOp(C, 0, 1); + } else { + t[2] = BitsBracket(C, 4); + t[1] = BitsOp(C, 2, 3); + t[0] = (BitsBracket(C, 1) << 1) | (BitsBracket(C, 0) & ~BitsBracket(C, 1)); + } + for (uint i = 0; i < 5; i++) { + EncodingData val; + val.encoding = TRIT; + val.num_bits = num_bits; + val.bit_value = m[i]; + val.quint_trit_value = t[i]; + ResultEmplaceBack(val); + } +} + +void DecodeIntegerSequence(uint max_range, uint num_values) { + EncodingData val = encoding_values[max_range]; + uint vals_decoded = 0; + while (vals_decoded < num_values) { + switch (val.encoding) { + case QUINT: + DecodeQuintBlock(val.num_bits); + vals_decoded += 3; + break; + case TRIT: + DecodeTritBlock(val.num_bits); + vals_decoded += 5; + break; + case JUST_BITS: + val.bit_value = StreamColorBits(val.num_bits); + ResultEmplaceBack(val); + vals_decoded++; + break; + } + } +} + +void DecodeColorValues(uvec4 modes, uint num_partitions, uint color_data_bits) { + uint num_values = 0; + for (uint i = 0; i < num_partitions; i++) { + num_values += ((modes[i] >> 2) + 1) << 1; + } + int range = 256; + while (--range > 0) { + EncodingData val = encoding_values[range]; + uint bit_length = GetBitLength(num_values, range); + if (bit_length <= color_data_bits) { + while (--range > 0) { + EncodingData newval = encoding_values[range]; + if (newval.encoding != val.encoding && newval.num_bits != val.num_bits) { + break; + } + } + ++range; + break; + } + } + DecodeIntegerSequence(range, num_values); + uint out_index = 0; + for (int itr = 0; itr < result_index; ++itr) { + if (out_index >= num_values) { + break; + } + EncodingData val = result_vector[itr]; + uint bitlen = val.num_bits; + uint bitval = val.bit_value; + uint A = 0, B = 0, C = 0, D = 0; + A = ReplicateBitTo9((bitval & 1)); + switch (val.encoding) { + case JUST_BITS: + color_values[out_index++] = FastReplicateTo8(bitval, bitlen); + break; + case TRIT: { + D = val.quint_trit_value; + switch (bitlen) { + case 1: + C = 204; + break; + case 2: { + C = 93; + uint b = (bitval >> 1) & 1; + B = (b << 8) | (b << 4) | (b << 2) | (b << 1); + break; + } + case 3: { + C = 44; + uint cb = (bitval >> 1) & 3; + B = (cb << 7) | (cb << 2) | cb; + break; + } + case 4: { + C = 22; + uint dcb = (bitval >> 1) & 7; + B = (dcb << 6) | dcb; + break; + } + case 5: { + C = 11; + uint edcb = (bitval >> 1) & 0xF; + B = (edcb << 5) | (edcb >> 2); + break; + } + case 6: { + C = 5; + uint fedcb = (bitval >> 1) & 0x1F; + B = (fedcb << 4) | (fedcb >> 4); + break; + } + } + break; + } + case QUINT: { + D = val.quint_trit_value; + switch (bitlen) { + case 1: + C = 113; + break; + case 2: { + C = 54; + uint b = (bitval >> 1) & 1; + B = (b << 8) | (b << 3) | (b << 2); + break; + } + case 3: { + C = 26; + uint cb = (bitval >> 1) & 3; + B = (cb << 7) | (cb << 1) | (cb >> 1); + break; + } + case 4: { + C = 13; + uint dcb = (bitval >> 1) & 7; + B = (dcb << 6) | (dcb >> 1); + break; + } + case 5: { + C = 6; + uint edcb = (bitval >> 1) & 0xF; + B = (edcb << 5) | (edcb >> 3); + break; + } + } + break; + } + } + if (val.encoding != JUST_BITS) { + uint T = (D * C) + B; + T ^= A; + T = (A & 0x80) | (T >> 2); + color_values[out_index++] = T; + } + } +} + +ivec2 BitTransferSigned(int a, int b) { + ivec2 transferred; + transferred.y = b >> 1; + transferred.y |= a & 0x80; + transferred.x = a >> 1; + transferred.x &= 0x3F; + if ((transferred.x & 0x20) > 0) { + transferred.x -= 0x40; + } + return transferred; +} + +uvec4 ClampByte(ivec4 color) { + for (uint i = 0; i < 4; ++i) { + color[i] = (color[i] < 0) ? 0 : ((color[i] > 255) ? 255 : color[i]); + } + return uvec4(color); +} + +ivec4 BlueContract(int a, int r, int g, int b) { + return ivec4(a, (r + b) >> 1, (g + b) >> 1, b); +} + +void ComputeEndpoints(out uvec4 ep1, out uvec4 ep2, uint color_endpoint_mode) { +#define READ_UINT_VALUES(N) \ + uint v[N]; \ + for (uint i = 0; i < N; i++) { \ + v[i] = color_values[colvals_index++]; \ + } + +#define READ_INT_VALUES(N) \ + int v[N]; \ + for (uint i = 0; i < N; i++) { \ + v[i] = int(color_values[colvals_index++]); \ + } + + switch (color_endpoint_mode) { + case 0: { + READ_UINT_VALUES(2) + ep1 = uvec4(0xFF, v[0], v[0], v[0]); + ep2 = uvec4(0xFF, v[1], v[1], v[1]); + break; + } + case 1: { + READ_UINT_VALUES(2) + uint L0 = (v[0] >> 2) | (v[1] & 0xC0); + uint L1 = max(L0 + (v[1] & 0x3F), 0xFFU); + ep1 = uvec4(0xFF, L0, L0, L0); + ep2 = uvec4(0xFF, L1, L1, L1); + break; + } + case 4: { + READ_UINT_VALUES(4) + ep1 = uvec4(v[2], v[0], v[0], v[0]); + ep2 = uvec4(v[3], v[1], v[1], v[1]); + break; + } + case 5: { + READ_INT_VALUES(4) + ivec2 transferred = BitTransferSigned(v[1], v[0]); + v[1] = transferred.x; + v[0] = transferred.y; + transferred = BitTransferSigned(v[3], v[2]); + v[3] = transferred.x; + v[2] = transferred.y; + ep1 = ClampByte(ivec4(v[2], v[0], v[0], v[0])); + ep2 = ClampByte(ivec4(v[2] + v[3], v[0] + v[1], v[0] + v[1], v[0] + v[1])); + break; + } + case 6: { + READ_UINT_VALUES(4) + ep1 = uvec4(0xFF, (v[0] * v[3]) >> 8, (v[1] * v[3]) >> 8, (v[2] * v[3]) >> 8); + ep2 = uvec4(0xFF, v[0], v[1], v[2]); + break; + } + case 8: { + READ_UINT_VALUES(6) + if ((v[1] + v[3] + v[5]) >= (v[0] + v[2] + v[4])) { + ep1 = uvec4(0xFF, v[0], v[2], v[4]); + ep2 = uvec4(0xFF, v[1], v[3], v[5]); + } else { + ep1 = uvec4(BlueContract(0xFF, int(v[1]), int(v[3]), int(v[5]))); + ep2 = uvec4(BlueContract(0xFF, int(v[0]), int(v[2]), int(v[4]))); + } + break; + } + case 9: { + READ_INT_VALUES(6) + ivec2 transferred = BitTransferSigned(v[1], v[0]); + v[1] = transferred.x; + v[0] = transferred.y; + transferred = BitTransferSigned(v[3], v[2]); + v[3] = transferred.x; + v[2] = transferred.y; + transferred = BitTransferSigned(v[5], v[4]); + v[5] = transferred.x; + v[4] = transferred.y; + if ((v[1] + v[3] + v[5]) >= 0) { + ep1 = ClampByte(ivec4(0xFF, v[0], v[2], v[4])); + ep2 = ClampByte(ivec4(0xFF, v[0] + v[1], v[2] + v[3], v[4] + v[5])); + } else { + ep1 = ClampByte(BlueContract(0xFF, v[0] + v[1], v[2] + v[3], v[4] + v[5])); + ep2 = ClampByte(BlueContract(0xFF, v[0], v[2], v[4])); + } + break; + } + case 10: { + READ_UINT_VALUES(6) + ep1 = uvec4(v[4], (v[0] * v[3]) >> 8, (v[1] * v[3]) >> 8, (v[2] * v[3]) >> 8); + ep2 = uvec4(v[5], v[0], v[1], v[2]); + break; + } + case 12: { + READ_UINT_VALUES(8) + if ((v[1] + v[3] + v[5]) >= (v[0] + v[2] + v[4])) { + ep1 = uvec4(v[6], v[0], v[2], v[4]); + ep2 = uvec4(v[7], v[1], v[3], v[5]); + } else { + ep1 = uvec4(BlueContract(int(v[7]), int(v[1]), int(v[3]), int(v[5]))); + ep2 = uvec4(BlueContract(int(v[6]), int(v[0]), int(v[2]), int(v[4]))); + } + break; + } + case 13: { + READ_INT_VALUES(8) + ivec2 transferred = BitTransferSigned(v[1], v[0]); + v[1] = transferred.x; + v[0] = transferred.y; + transferred = BitTransferSigned(v[3], v[2]); + v[3] = transferred.x; + v[2] = transferred.y; + + transferred = BitTransferSigned(v[5], v[4]); + v[5] = transferred.x; + v[4] = transferred.y; + + transferred = BitTransferSigned(v[7], v[6]); + v[7] = transferred.x; + v[6] = transferred.y; + + if ((v[1] + v[3] + v[5]) >= 0) { + ep1 = ClampByte(ivec4(v[6], v[0], v[2], v[4])); + ep2 = ClampByte(ivec4(v[7] + v[6], v[0] + v[1], v[2] + v[3], v[4] + v[5])); + } else { + ep1 = ClampByte(BlueContract(v[6] + v[7], v[0] + v[1], v[2] + v[3], v[4] + v[5])); + ep2 = ClampByte(BlueContract(v[6], v[0], v[2], v[4])); + } + break; + } + default: { + // HDR mode, or more likely a bug computing the color_endpoint_mode + ep1 = uvec4(0xFF, 0xFF, 0, 0); + ep2 = uvec4(0xFF, 0xFF, 0, 0); + break; + } + } +#undef READ_UINT_VALUES +#undef READ_INT_VALUES +} + +uint UnquantizeTexelWeight(EncodingData val) { + uint bitval = val.bit_value; + uint bitlen = val.num_bits; + uint A = ReplicateBitTo7((bitval & 1)); + uint B = 0, C = 0, D = 0; + uint result = 0; + switch (val.encoding) { + case JUST_BITS: + result = FastReplicateTo6(bitval, bitlen); + break; + case TRIT: { + D = val.quint_trit_value; + switch (bitlen) { + case 0: { + uint results[3] = {0, 32, 63}; + result = results[D]; + break; + } + case 1: { + C = 50; + break; + } + case 2: { + C = 23; + uint b = (bitval >> 1) & 1; + B = (b << 6) | (b << 2) | b; + break; + } + case 3: { + C = 11; + uint cb = (bitval >> 1) & 3; + B = (cb << 5) | cb; + break; + } + default: + break; + } + break; + } + case QUINT: { + D = val.quint_trit_value; + switch (bitlen) { + case 0: { + uint results[5] = {0, 16, 32, 47, 63}; + result = results[D]; + break; + } + case 1: { + C = 28; + break; + } + case 2: { + C = 13; + uint b = (bitval >> 1) & 1; + B = (b << 6) | (b << 1); + break; + } + } + break; + } + } + if (val.encoding != JUST_BITS && bitlen > 0) { + result = D * C + B; + result ^= A; + result = (A & 0x20) | (result >> 2); + } + if (result > 32) { + result += 1; + } + return result; +} + +void UnquantizeTexelWeights(bool dual_plane, uvec2 size) { + uint weight_idx = 0; + uint unquantized[2][144]; + uint area = size.x * size.y; + for (uint itr = 0; itr < texel_vector_index; itr++) { + unquantized[0][weight_idx] = UnquantizeTexelWeight(texel_vector[itr]); + if (dual_plane) { + ++itr; + unquantized[1][weight_idx] = UnquantizeTexelWeight(texel_vector[itr]); + if (itr == texel_vector_index) { + break; + } + } + if (++weight_idx >= (area)) + break; + } + + const uint Ds = uint((block_dims.x * 0.5f + 1024) / (block_dims.x - 1)); + const uint Dt = uint((block_dims.y * 0.5f + 1024) / (block_dims.y - 1)); + const uint k_plane_scale = dual_plane ? 2 : 1; + for (uint plane = 0; plane < k_plane_scale; plane++) { + for (uint t = 0; t < block_dims.y; t++) { + for (uint s = 0; s < block_dims.x; s++) { + uint cs = Ds * s; + uint ct = Dt * t; + uint gs = (cs * (size.x - 1) + 32) >> 6; + uint gt = (ct * (size.y - 1) + 32) >> 6; + uint js = gs >> 4; + uint fs = gs & 0xF; + uint jt = gt >> 4; + uint ft = gt & 0x0F; + uint w11 = (fs * ft + 8) >> 4; + uint w10 = ft - w11; + uint w01 = fs - w11; + uint w00 = 16 - fs - ft + w11; + uvec4 w = uvec4(w00, w01, w10, w11); + uint v0 = jt * size.x + js; + + uvec4 p = uvec4(0); + if (v0 < area) { + p.x = unquantized[plane][v0]; + } + if ((v0 + 1) < (area)) { + p.y = unquantized[plane][v0 + 1]; + } + if ((v0 + size.x) < (area)) { + p.z = unquantized[plane][(v0 + size.x)]; + } + if ((v0 + size.x + 1) < (area)) { + p.w = unquantized[plane][(v0 + size.x + 1)]; + } + unquantized_texel_weights[plane][t * block_dims.x + s] = (uint(dot(p, w)) + 8) >> 4; + } + } + } +} + +int FindLayout(uint mode) { + if ((mode & 3) != 0) { + if ((mode & 8) != 0) { + if ((mode & 4) != 0) { + if ((mode & 0x100) != 0) { + return 4; + } + return 3; + } + return 2; + } + if ((mode & 4) != 0) { + return 1; + } + return 0; + } + if ((mode & 0x100) != 0) { + if ((mode & 0x80) != 0) { + if ((mode & 0x20) != 0) { + return 8; + } + return 7; + } + return 9; + } + if ((mode & 0x80) != 0) { + return 6; + } + return 5; +} + +TexelWeightParams DecodeBlockInfo(uint block_index) { + TexelWeightParams params = TexelWeightParams(uvec2(0), 0, false, false, false, false); + uint mode = StreamBits(11); + if ((mode & 0x1ff) == 0x1fc) { + if ((mode & 0x200) != 0) { + params.void_extent_hdr = true; + } else { + params.void_extent_ldr = true; + } + if ((mode & 0x400) == 0 || StreamBits(1) == 0) { + params.error_state = true; + } + return params; + } + if ((mode & 0xf) == 0) { + params.error_state = true; + return params; + } + if ((mode & 3) == 0 && (mode & 0x1c0) == 0x1c0) { + params.error_state = true; + return params; + } + uint A, B; + uint mode_layout = FindLayout(mode); + switch (mode_layout) { + case 0: + A = (mode >> 5) & 0x3; + B = (mode >> 7) & 0x3; + params.size = uvec2(B + 4, A + 2); + break; + case 1: + A = (mode >> 5) & 0x3; + B = (mode >> 7) & 0x3; + params.size = uvec2(B + 8, A + 2); + break; + case 2: + A = (mode >> 5) & 0x3; + B = (mode >> 7) & 0x3; + params.size = uvec2(A + 2, B + 8); + break; + case 3: + A = (mode >> 5) & 0x3; + B = (mode >> 7) & 0x1; + params.size = uvec2(A + 2, B + 6); + break; + case 4: + A = (mode >> 5) & 0x3; + B = (mode >> 7) & 0x1; + params.size = uvec2(B + 2, A + 2); + break; + case 5: + A = (mode >> 5) & 0x3; + params.size = uvec2(12, A + 2); + break; + case 6: + A = (mode >> 5) & 0x3; + params.size = uvec2(A + 2, 12); + break; + case 7: + params.size = uvec2(6, 10); + break; + case 8: + params.size = uvec2(10, 6); + break; + case 9: + A = (mode >> 5) & 0x3; + B = (mode >> 9) & 0x3; + params.size = uvec2(A + 6, B + 6); + break; + default: + params.error_state = true; + break; + } + params.dual_plane = (mode_layout != 9) && ((mode & 0x400) != 0); + uint weight_index = (mode & 0x10) != 0 ? 1 : 0; + if (mode_layout < 5) { + weight_index |= (mode & 0x3) << 1; + } else { + weight_index |= (mode & 0xc) >> 1; + } + weight_index -= 2; + if ((mode_layout != 9) && ((mode & 0x200) != 0)) { + const int max_weights[6] = int[6](9, 11, 15, 19, 23, 31); + params.max_weight = max_weights[weight_index]; + } else { + const int max_weights[6] = int[6](1, 2, 3, 4, 5, 7); + params.max_weight = max_weights[weight_index]; + } + return params; +} + +void FillError(ivec3 coord) { + for (uint j = 0; j < block_dims.y; j++) { + for (uint i = 0; i < block_dims.x; i++) { + imageStore(dest_image, coord + ivec3(i, j, 0), vec4(1.0, 1.0, 0.0, 1.0)); + } + } +} + +void FillVoidExtentLDR(ivec3 coord) { + StreamBits(52); + uint r_u = StreamBits(16); + uint g_u = StreamBits(16); + uint b_u = StreamBits(16); + uint a_u = StreamBits(16); + float a = float(a_u) / 65535.0f; + float r = float(r_u) / 65535.0f; + float g = float(g_u) / 65535.0f; + float b = float(b_u) / 65535.0f; + for (uint j = 0; j < block_dims.y; j++) { + for (uint i = 0; i < block_dims.x; i++) { + imageStore(dest_image, coord + ivec3(i, j, 0), vec4(r, g, b, a)); + } + } +} + +void DecompressBlock(ivec3 coord, uint block_index) { + TexelWeightParams params = DecodeBlockInfo(block_index); + if (params.error_state) { + FillError(coord); + return; + } + if (params.void_extent_hdr) { + FillError(coord); + return; + } + if (params.void_extent_ldr) { + FillVoidExtentLDR(coord); + return; + } + if ((params.size.x > block_dims.x) || (params.size.y > block_dims.y)) { + FillError(coord); + return; + } + uint num_partitions = StreamBits(2) + 1; + if (num_partitions > 4 || (num_partitions == 4 && params.dual_plane)) { + FillError(coord); + return; + } + int plane_index = -1; + uint partition_index = 1; + uvec4 color_endpoint_mode = uvec4(0); + uint ced_pointer = 0; + uint base_cem = 0; + if (num_partitions == 1) { + color_endpoint_mode.x = StreamBits(4); + partition_index = 0; + } else { + partition_index = StreamBits(10); + base_cem = StreamBits(6); + } + uint base_mode = base_cem & 3; + uint weight_bits = GetPackedBitSize(params.size, params.dual_plane, params.max_weight); + uint remaining_bits = 128 - weight_bits - total_bitsread; + uint extra_cem_bits = 0; + if (base_mode > 0) { + switch (num_partitions) { + case 2: + extra_cem_bits += 2; + break; + case 3: + extra_cem_bits += 5; + break; + case 4: + extra_cem_bits += 8; + break; + default: + return; + } + } + remaining_bits -= extra_cem_bits; + uint plane_selector_bits = 0; + if (params.dual_plane) { + plane_selector_bits = 2; + } + remaining_bits -= plane_selector_bits; + if (remaining_bits > 128) { + // Bad data, more remaining bits than 4 bytes + // return early + return; + } + // Read color data... + uint color_data_bits = remaining_bits; + while (remaining_bits > 0) { + int nb = int(min(remaining_bits, 8U)); + uint b = StreamBits(nb); + color_endpoint_data[ced_pointer] = uint(bitfieldExtract(b, 0, nb)); + ++ced_pointer; + remaining_bits -= nb; + } + plane_index = int(StreamBits(plane_selector_bits)); + if (base_mode > 0) { + uint extra_cem = StreamBits(extra_cem_bits); + uint cem = (extra_cem << 6) | base_cem; + cem >>= 2; + uvec4 C = uvec4(0); + for (uint i = 0; i < num_partitions; i++) { + C[i] = (cem & 1); + cem >>= 1; + } + uvec4 M = uvec4(0); + for (uint i = 0; i < num_partitions; i++) { + M[i] = cem & 3; + cem >>= 2; + } + for (uint i = 0; i < num_partitions; i++) { + color_endpoint_mode[i] = base_mode; + if (C[i] == 0) { + --color_endpoint_mode[i]; + } + color_endpoint_mode[i] <<= 2; + color_endpoint_mode[i] |= M[i]; + } + } else if (num_partitions > 1) { + uint cem = base_cem >> 2; + for (uint i = 0; i < num_partitions; i++) { + color_endpoint_mode[i] = cem; + } + } + DecodeColorValues(color_endpoint_mode, num_partitions, color_data_bits); + + uvec4 endpoints[4][2]; + for (uint i = 0; i < num_partitions; i++) { + ComputeEndpoints(endpoints[i][0], endpoints[i][1], color_endpoint_mode[i]); + } + + for (uint i = 0; i < 16; i++) { + texel_weight_data[i] = local_buff[i]; + } + for (uint i = 0; i < 8; i++) { +#define REVERSE_BYTE(b) ((b * 0x0802U & 0x22110U) | (b * 0x8020U & 0x88440U)) * 0x10101U >> 16 + uint a = REVERSE_BYTE(texel_weight_data[i]); + uint b = REVERSE_BYTE(texel_weight_data[15 - i]); +#undef REVERSE_BYTE + texel_weight_data[i] = uint(bitfieldExtract(b, 0, 8)); + texel_weight_data[15 - i] = uint(bitfieldExtract(a, 0, 8)); + } + uint clear_byte_start = + (GetPackedBitSize(params.size, params.dual_plane, params.max_weight) >> 3) + 1; + texel_weight_data[clear_byte_start - 1] = + texel_weight_data[clear_byte_start - 1] & + uint( + ((1 << (GetPackedBitSize(params.size, params.dual_plane, params.max_weight) % 8)) - 1)); + for (uint i = 0; i < 16 - clear_byte_start; i++) { + texel_weight_data[clear_byte_start + i] = 0U; + } + texel_flag = true; // use texel "vector" and bit stream in integer decoding + DecodeIntegerSequence(params.max_weight, GetNumWeightValues(params.size, params.dual_plane)); + + UnquantizeTexelWeights(params.dual_plane, params.size); + + for (uint j = 0; j < block_dims.y; j++) { + for (uint i = 0; i < block_dims.x; i++) { + uint local_partition = Select2DPartition(partition_index, i, j, num_partitions, + (block_dims.y * block_dims.x) < 32); + vec4 p; + uvec4 C0 = ReplicateByteTo16(endpoints[local_partition][0]); + uvec4 C1 = ReplicateByteTo16(endpoints[local_partition][1]); + uvec4 plane_vec = uvec4(0); + uvec4 weight_vec = uvec4(0); + for (uint c = 0; c < 4; c++) { + if (params.dual_plane && (((plane_index + 1) & 3) == c)) { + plane_vec[c] = 1; + } + weight_vec[c] = unquantized_texel_weights[plane_vec[c]][j * block_dims.x + i]; + } + vec4 Cf = vec4((C0 * (uvec4(64) - weight_vec) + C1 * weight_vec + uvec4(32)) / 64); + p = (Cf / 65535.0); + imageStore(dest_image, coord + ivec3(i, j, 0), p.gbar); + } + } +} + +void main() { + uvec3 pos = gl_GlobalInvocationID; + pos.x <<= bytes_per_block_log2; + + // Read as soon as possible due to its latency + const uint swizzle = SwizzleOffset(pos.xy); + + const uint block_y = pos.y >> GOB_SIZE_Y_SHIFT; + + uint offset = 0; + offset += pos.z * layer_stride; + offset += (block_y >> block_height) * block_size; + offset += (block_y & block_height_mask) << GOB_SIZE_SHIFT; + offset += (pos.x >> GOB_SIZE_X_SHIFT) << x_shift; + offset += swizzle; + + const ivec3 coord = ivec3(gl_GlobalInvocationID * uvec3(block_dims, 1)); + uint block_index = + pos.z * gl_WorkGroupSize.x * gl_WorkGroupSize.y + pos.y * gl_WorkGroupSize.x + pos.x; + + current_index = 0; + bitsread = 0; + for (int i = 0; i < 16; i++) { + local_buff[i] = ReadTexel(offset + i); + } + DecompressBlock(coord, block_index); +} diff --git a/src/video_core/host_shaders/source_shader.h.in b/src/video_core/host_shaders/source_shader.h.in index ccdb0d2a9..929dec39b 100644 --- a/src/video_core/host_shaders/source_shader.h.in +++ b/src/video_core/host_shaders/source_shader.h.in @@ -4,6 +4,8 @@ namespace HostShaders { -constexpr std::string_view @CONTENTS_NAME@ = R"(@CONTENTS@)"; +constexpr std::string_view @CONTENTS_NAME@ = { +@CONTENTS@ +}; } // namespace HostShaders diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index e028677e9..623b43d8a 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -307,7 +307,8 @@ void ApplySwizzle(GLuint handle, PixelFormat format, std::array<SwizzleSource, 4 [[nodiscard]] bool CanBeAccelerated(const TextureCacheRuntime& runtime, const VideoCommon::ImageInfo& info) { - // Disable accelerated uploads for now as they don't implement swizzled uploads + return !runtime.HasNativeASTC() && IsPixelFormatASTC(info.format); + // Disable other accelerated uploads for now as they don't implement swizzled uploads return false; switch (info.type) { case ImageType::e2D: @@ -569,7 +570,11 @@ void TextureCacheRuntime::AccelerateImageUpload(Image& image, const ImageBufferM std::span<const SwizzleParameters> swizzles) { switch (image.info.type) { case ImageType::e2D: - return util_shaders.BlockLinearUpload2D(image, map, swizzles); + if (IsPixelFormatASTC(image.info.format)) { + return util_shaders.ASTCDecode(image, map, swizzles); + } else { + return util_shaders.BlockLinearUpload2D(image, map, swizzles); + } case ImageType::e3D: return util_shaders.BlockLinearUpload3D(image, map, swizzles); case ImageType::Linear: @@ -599,6 +604,10 @@ FormatProperties TextureCacheRuntime::FormatInfo(ImageType type, GLenum internal } } +bool TextureCacheRuntime::HasNativeASTC() const noexcept { + return device.HasASTC(); +} + TextureCacheRuntime::StagingBuffers::StagingBuffers(GLenum storage_flags_, GLenum map_flags_) : storage_flags{storage_flags_}, map_flags{map_flags_} {} diff --git a/src/video_core/renderer_opengl/gl_texture_cache.h b/src/video_core/renderer_opengl/gl_texture_cache.h index 3fbaa102f..3c871541b 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.h +++ b/src/video_core/renderer_opengl/gl_texture_cache.h @@ -95,6 +95,8 @@ public: return has_broken_texture_view_formats; } + bool HasNativeASTC() const noexcept; + private: struct StagingBuffers { explicit StagingBuffers(GLenum storage_flags_, GLenum map_flags_); diff --git a/src/video_core/renderer_opengl/util_shaders.cpp b/src/video_core/renderer_opengl/util_shaders.cpp index 2fe4799bc..47fddcb6e 100644 --- a/src/video_core/renderer_opengl/util_shaders.cpp +++ b/src/video_core/renderer_opengl/util_shaders.cpp @@ -2,7 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <bit> #include <span> #include <string_view> @@ -11,6 +10,7 @@ #include "common/assert.h" #include "common/common_types.h" #include "common/div_ceil.h" +#include "video_core/host_shaders/astc_decoder_comp.h" #include "video_core/host_shaders/block_linear_unswizzle_2d_comp.h" #include "video_core/host_shaders/block_linear_unswizzle_3d_comp.h" #include "video_core/host_shaders/opengl_copy_bc4_comp.h" @@ -20,16 +20,18 @@ #include "video_core/renderer_opengl/gl_shader_manager.h" #include "video_core/renderer_opengl/gl_texture_cache.h" #include "video_core/renderer_opengl/util_shaders.h" -#include "video_core/surface.h" #include "video_core/texture_cache/accelerated_swizzle.h" #include "video_core/texture_cache/types.h" #include "video_core/texture_cache/util.h" +#include "video_core/textures/astc.h" #include "video_core/textures/decoders.h" namespace OpenGL { using namespace HostShaders; +using namespace Tegra::Texture::ASTC; +using VideoCommon::Extent2D; using VideoCommon::Extent3D; using VideoCommon::ImageCopy; using VideoCommon::ImageType; @@ -57,7 +59,7 @@ size_t NumPixelsInCopy(const VideoCommon::ImageCopy& copy) { } // Anonymous namespace UtilShaders::UtilShaders(ProgramManager& program_manager_) - : program_manager{program_manager_}, + : program_manager{program_manager_}, astc_decoder_program(MakeProgram(ASTC_DECODER_COMP)), block_linear_unswizzle_2d_program(MakeProgram(BLOCK_LINEAR_UNSWIZZLE_2D_COMP)), block_linear_unswizzle_3d_program(MakeProgram(BLOCK_LINEAR_UNSWIZZLE_3D_COMP)), pitch_unswizzle_program(MakeProgram(PITCH_UNSWIZZLE_COMP)), @@ -65,11 +67,79 @@ UtilShaders::UtilShaders(ProgramManager& program_manager_) copy_bc4_program(MakeProgram(OPENGL_COPY_BC4_COMP)) { const auto swizzle_table = Tegra::Texture::MakeSwizzleTable(); swizzle_table_buffer.Create(); + astc_buffer.Create(); glNamedBufferStorage(swizzle_table_buffer.handle, sizeof(swizzle_table), &swizzle_table, 0); + glNamedBufferStorage(astc_buffer.handle, sizeof(ASTC_BUFFER_DATA), &ASTC_BUFFER_DATA, 0); } UtilShaders::~UtilShaders() = default; +void UtilShaders::ASTCDecode(Image& image, const ImageBufferMap& map, + std::span<const VideoCommon::SwizzleParameters> swizzles) { + static constexpr GLuint BINDING_SWIZZLE_BUFFER = 0; + static constexpr GLuint BINDING_INPUT_BUFFER = 1; + static constexpr GLuint BINDING_ENC_BUFFER = 2; + + static constexpr GLuint BINDING_6_TO_8_BUFFER = 3; + static constexpr GLuint BINDING_7_TO_8_BUFFER = 4; + static constexpr GLuint BINDING_8_TO_8_BUFFER = 5; + static constexpr GLuint BINDING_BYTE_TO_16_BUFFER = 6; + + static constexpr GLuint BINDING_OUTPUT_IMAGE = 0; + + const Extent2D tile_size{ + .width = VideoCore::Surface::DefaultBlockWidth(image.info.format), + .height = VideoCore::Surface::DefaultBlockHeight(image.info.format), + }; + program_manager.BindHostCompute(astc_decoder_program.handle); + glBindBufferBase(GL_SHADER_STORAGE_BUFFER, BINDING_SWIZZLE_BUFFER, swizzle_table_buffer.handle); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_ENC_BUFFER, astc_buffer.handle, + offsetof(AstcBufferData, encoding_values), + sizeof(AstcBufferData::encoding_values)); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_6_TO_8_BUFFER, astc_buffer.handle, + offsetof(AstcBufferData, replicate_6_to_8), + sizeof(AstcBufferData::replicate_6_to_8)); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_7_TO_8_BUFFER, astc_buffer.handle, + offsetof(AstcBufferData, replicate_7_to_8), + sizeof(AstcBufferData::replicate_7_to_8)); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_8_TO_8_BUFFER, astc_buffer.handle, + offsetof(AstcBufferData, replicate_8_to_8), + sizeof(AstcBufferData::replicate_8_to_8)); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_BYTE_TO_16_BUFFER, astc_buffer.handle, + offsetof(AstcBufferData, replicate_byte_to_16), + sizeof(AstcBufferData::replicate_byte_to_16)); + + glFlushMappedNamedBufferRange(map.buffer, map.offset, image.guest_size_bytes); + glUniform2ui(1, tile_size.width, tile_size.height); + // Ensure buffer data is valid before dispatching + glFlush(); + for (const SwizzleParameters& swizzle : swizzles) { + const size_t input_offset = swizzle.buffer_offset + map.offset; + const u32 num_dispatches_x = Common::DivCeil(swizzle.num_tiles.width, 32U); + const u32 num_dispatches_y = Common::DivCeil(swizzle.num_tiles.height, 32U); + + const auto params = MakeBlockLinearSwizzle2DParams(swizzle, image.info); + ASSERT(params.origin == (std::array<u32, 3>{0, 0, 0})); + ASSERT(params.destination == (std::array<s32, 3>{0, 0, 0})); + + glUniform1ui(2, params.bytes_per_block_log2); + glUniform1ui(3, params.layer_stride); + glUniform1ui(4, params.block_size); + glUniform1ui(5, params.x_shift); + glUniform1ui(6, params.block_height); + glUniform1ui(7, params.block_height_mask); + + glBindImageTexture(BINDING_OUTPUT_IMAGE, image.StorageHandle(), swizzle.level, GL_TRUE, 0, + GL_WRITE_ONLY, GL_RGBA8); + // ASTC texture data + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_INPUT_BUFFER, map.buffer, input_offset, + image.guest_size_bytes - swizzle.buffer_offset); + + glDispatchCompute(num_dispatches_x, num_dispatches_y, image.info.resources.layers); + } + program_manager.RestoreGuestCompute(); +} + void UtilShaders::BlockLinearUpload2D(Image& image, const ImageBufferMap& map, std::span<const SwizzleParameters> swizzles) { static constexpr Extent3D WORKGROUP_SIZE{32, 32, 1}; diff --git a/src/video_core/renderer_opengl/util_shaders.h b/src/video_core/renderer_opengl/util_shaders.h index 93b009743..53d65f368 100644 --- a/src/video_core/renderer_opengl/util_shaders.h +++ b/src/video_core/renderer_opengl/util_shaders.h @@ -40,6 +40,9 @@ public: explicit UtilShaders(ProgramManager& program_manager); ~UtilShaders(); + void ASTCDecode(Image& image, const ImageBufferMap& map, + std::span<const VideoCommon::SwizzleParameters> swizzles); + void BlockLinearUpload2D(Image& image, const ImageBufferMap& map, std::span<const VideoCommon::SwizzleParameters> swizzles); @@ -59,7 +62,9 @@ private: ProgramManager& program_manager; OGLBuffer swizzle_table_buffer; + OGLBuffer astc_buffer; + OGLProgram astc_decoder_program; OGLProgram block_linear_unswizzle_2d_program; OGLProgram block_linear_unswizzle_3d_program; OGLProgram pitch_unswizzle_program; diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp index 19aaf034f..f088447e9 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp @@ -166,7 +166,7 @@ struct FormatTuple { {VK_FORMAT_R16G16_SINT, Attachable | Storage}, // R16G16_SINT {VK_FORMAT_R16G16_SNORM, Attachable | Storage}, // R16G16_SNORM {VK_FORMAT_UNDEFINED}, // R32G32B32_FLOAT - {VK_FORMAT_R8G8B8A8_SRGB, Attachable}, // A8B8G8R8_SRGB + {VK_FORMAT_A8B8G8R8_SRGB_PACK32, Attachable}, // A8B8G8R8_SRGB {VK_FORMAT_R8G8_UNORM, Attachable | Storage}, // R8G8_UNORM {VK_FORMAT_R8G8_SNORM, Attachable | Storage}, // R8G8_SNORM {VK_FORMAT_R8G8_SINT, Attachable | Storage}, // R8G8_SINT diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.cpp b/src/video_core/renderer_vulkan/vk_compute_pass.cpp index 2f9a7b028..e11406e58 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pass.cpp @@ -11,18 +11,39 @@ #include "common/assert.h" #include "common/common_types.h" #include "common/div_ceil.h" +#include "video_core/host_shaders/astc_decoder_comp_spv.h" #include "video_core/host_shaders/vulkan_quad_indexed_comp_spv.h" #include "video_core/host_shaders/vulkan_uint8_comp_spv.h" #include "video_core/renderer_vulkan/vk_compute_pass.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" +#include "video_core/renderer_vulkan/vk_texture_cache.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" +#include "video_core/texture_cache/accelerated_swizzle.h" +#include "video_core/texture_cache/types.h" +#include "video_core/textures/astc.h" +#include "video_core/textures/decoders.h" #include "video_core/vulkan_common/vulkan_device.h" #include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { + +using Tegra::Texture::SWIZZLE_TABLE; +using Tegra::Texture::ASTC::EncodingsValues; +using namespace Tegra::Texture::ASTC; + namespace { + +constexpr u32 ASTC_BINDING_INPUT_BUFFER = 0; +constexpr u32 ASTC_BINDING_ENC_BUFFER = 1; +constexpr u32 ASTC_BINDING_6_TO_8_BUFFER = 2; +constexpr u32 ASTC_BINDING_7_TO_8_BUFFER = 3; +constexpr u32 ASTC_BINDING_8_TO_8_BUFFER = 4; +constexpr u32 ASTC_BINDING_BYTE_TO_16_BUFFER = 5; +constexpr u32 ASTC_BINDING_SWIZZLE_BUFFER = 6; +constexpr u32 ASTC_BINDING_OUTPUT_IMAGE = 7; + VkPushConstantRange BuildComputePushConstantRange(std::size_t size) { return { .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, @@ -50,6 +71,67 @@ std::array<VkDescriptorSetLayoutBinding, 2> BuildInputOutputDescriptorSetBinding }}; } +std::array<VkDescriptorSetLayoutBinding, 8> BuildASTCDescriptorSetBindings() { + return {{ + { + .binding = ASTC_BINDING_INPUT_BUFFER, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = ASTC_BINDING_ENC_BUFFER, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = ASTC_BINDING_6_TO_8_BUFFER, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = ASTC_BINDING_7_TO_8_BUFFER, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = ASTC_BINDING_8_TO_8_BUFFER, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = ASTC_BINDING_BYTE_TO_16_BUFFER, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = ASTC_BINDING_SWIZZLE_BUFFER, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = ASTC_BINDING_OUTPUT_IMAGE, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + }}; +} + VkDescriptorUpdateTemplateEntryKHR BuildInputOutputDescriptorUpdateTemplate() { return { .dstBinding = 0, @@ -61,6 +143,94 @@ VkDescriptorUpdateTemplateEntryKHR BuildInputOutputDescriptorUpdateTemplate() { }; } +std::array<VkDescriptorUpdateTemplateEntryKHR, 8> BuildASTCPassDescriptorUpdateTemplateEntry() { + return {{ + { + .dstBinding = ASTC_BINDING_INPUT_BUFFER, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = ASTC_BINDING_INPUT_BUFFER * sizeof(DescriptorUpdateEntry), + .stride = sizeof(DescriptorUpdateEntry), + }, + { + .dstBinding = ASTC_BINDING_ENC_BUFFER, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = ASTC_BINDING_ENC_BUFFER * sizeof(DescriptorUpdateEntry), + .stride = sizeof(DescriptorUpdateEntry), + }, + { + .dstBinding = ASTC_BINDING_6_TO_8_BUFFER, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = ASTC_BINDING_6_TO_8_BUFFER * sizeof(DescriptorUpdateEntry), + .stride = sizeof(DescriptorUpdateEntry), + }, + { + .dstBinding = ASTC_BINDING_7_TO_8_BUFFER, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = ASTC_BINDING_7_TO_8_BUFFER * sizeof(DescriptorUpdateEntry), + .stride = sizeof(DescriptorUpdateEntry), + }, + { + .dstBinding = ASTC_BINDING_8_TO_8_BUFFER, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = ASTC_BINDING_8_TO_8_BUFFER * sizeof(DescriptorUpdateEntry), + .stride = sizeof(DescriptorUpdateEntry), + }, + { + .dstBinding = ASTC_BINDING_BYTE_TO_16_BUFFER, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = ASTC_BINDING_BYTE_TO_16_BUFFER * sizeof(DescriptorUpdateEntry), + .stride = sizeof(DescriptorUpdateEntry), + }, + { + .dstBinding = ASTC_BINDING_SWIZZLE_BUFFER, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = ASTC_BINDING_SWIZZLE_BUFFER * sizeof(DescriptorUpdateEntry), + .stride = sizeof(DescriptorUpdateEntry), + }, + { + .dstBinding = ASTC_BINDING_OUTPUT_IMAGE, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, + .offset = ASTC_BINDING_OUTPUT_IMAGE * sizeof(DescriptorUpdateEntry), + .stride = sizeof(DescriptorUpdateEntry), + }, + }}; +} + +struct AstcPushConstants { + std::array<u32, 2> blocks_dims; + u32 bytes_per_block_log2; + u32 layer_stride; + u32 block_size; + u32 x_shift; + u32 block_height; + u32 block_height_mask; +}; + +struct AstcBufferData { + decltype(SWIZZLE_TABLE) swizzle_table_buffer = SWIZZLE_TABLE; + decltype(EncodingsValues) encoding_values = EncodingsValues; + decltype(REPLICATE_6_BIT_TO_8_TABLE) replicate_6_to_8 = REPLICATE_6_BIT_TO_8_TABLE; + decltype(REPLICATE_7_BIT_TO_8_TABLE) replicate_7_to_8 = REPLICATE_7_BIT_TO_8_TABLE; + decltype(REPLICATE_8_BIT_TO_8_TABLE) replicate_8_to_8 = REPLICATE_8_BIT_TO_8_TABLE; + decltype(REPLICATE_BYTE_TO_16_TABLE) replicate_byte_to_16 = REPLICATE_BYTE_TO_16_TABLE; +} constexpr ASTC_BUFFER_DATA; + } // Anonymous namespace VKComputePass::VKComputePass(const Device& device, VKDescriptorPool& descriptor_pool, @@ -238,4 +408,167 @@ std::pair<VkBuffer, VkDeviceSize> QuadIndexedPass::Assemble( return {staging.buffer, staging.offset}; } +ASTCDecoderPass::ASTCDecoderPass(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + StagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + MemoryAllocator& memory_allocator_) + : VKComputePass(device_, descriptor_pool_, BuildASTCDescriptorSetBindings(), + BuildASTCPassDescriptorUpdateTemplateEntry(), + BuildComputePushConstantRange(sizeof(AstcPushConstants)), + ASTC_DECODER_COMP_SPV), + device{device_}, scheduler{scheduler_}, staging_buffer_pool{staging_buffer_pool_}, + update_descriptor_queue{update_descriptor_queue_}, memory_allocator{memory_allocator_} {} + +ASTCDecoderPass::~ASTCDecoderPass() = default; + +void ASTCDecoderPass::MakeDataBuffer() { + constexpr size_t TOTAL_BUFFER_SIZE = sizeof(ASTC_BUFFER_DATA) + sizeof(SWIZZLE_TABLE); + data_buffer = device.GetLogical().CreateBuffer(VkBufferCreateInfo{ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .size = TOTAL_BUFFER_SIZE, + .usage = VK_BUFFER_USAGE_STORAGE_BUFFER_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + }); + data_buffer_commit = memory_allocator.Commit(data_buffer, MemoryUsage::Upload); + + const auto staging_ref = staging_buffer_pool.Request(TOTAL_BUFFER_SIZE, MemoryUsage::Upload); + std::memcpy(staging_ref.mapped_span.data(), &ASTC_BUFFER_DATA, sizeof(ASTC_BUFFER_DATA)); + // Tack on the swizzle table at the end of the buffer + std::memcpy(staging_ref.mapped_span.data() + sizeof(ASTC_BUFFER_DATA), &SWIZZLE_TABLE, + sizeof(SWIZZLE_TABLE)); + + scheduler.Record([src = staging_ref.buffer, offset = staging_ref.offset, dst = *data_buffer, + TOTAL_BUFFER_SIZE](vk::CommandBuffer cmdbuf) { + cmdbuf.CopyBuffer(src, dst, + VkBufferCopy{ + .srcOffset = offset, + .dstOffset = 0, + .size = TOTAL_BUFFER_SIZE, + }); + cmdbuf.PipelineBarrier( + VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, 0, + VkMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = 0, + .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_TRANSFER_WRITE_BIT, + }); + }); +} + +void ASTCDecoderPass::Assemble(Image& image, const StagingBufferRef& map, + std::span<const VideoCommon::SwizzleParameters> swizzles) { + using namespace VideoCommon::Accelerated; + const std::array<u32, 2> block_dims{ + VideoCore::Surface::DefaultBlockWidth(image.info.format), + VideoCore::Surface::DefaultBlockHeight(image.info.format), + }; + scheduler.RequestOutsideRenderPassOperationContext(); + if (!data_buffer) { + MakeDataBuffer(); + } + const VkPipeline vk_pipeline = *pipeline; + const VkImageAspectFlags aspect_mask = image.AspectMask(); + const VkImage vk_image = image.Handle(); + const bool is_initialized = image.ExchangeInitialization(); + scheduler.Record( + [vk_pipeline, vk_image, aspect_mask, is_initialized](vk::CommandBuffer cmdbuf) { + const VkImageMemoryBarrier image_barrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT, + .oldLayout = is_initialized ? VK_IMAGE_LAYOUT_GENERAL : VK_IMAGE_LAYOUT_UNDEFINED, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = vk_image, + .subresourceRange{ + .aspectMask = aspect_mask, + .baseMipLevel = 0, + .levelCount = VK_REMAINING_MIP_LEVELS, + .baseArrayLayer = 0, + .layerCount = VK_REMAINING_ARRAY_LAYERS, + }, + }; + cmdbuf.PipelineBarrier(is_initialized ? VK_PIPELINE_STAGE_ALL_COMMANDS_BIT : 0, + VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, 0, image_barrier); + cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_COMPUTE, vk_pipeline); + }); + for (const VideoCommon::SwizzleParameters& swizzle : swizzles) { + const size_t input_offset = swizzle.buffer_offset + map.offset; + const u32 num_dispatches_x = Common::DivCeil(swizzle.num_tiles.width, 32U); + const u32 num_dispatches_y = Common::DivCeil(swizzle.num_tiles.height, 32U); + const u32 num_dispatches_z = image.info.resources.layers; + + update_descriptor_queue.Acquire(); + update_descriptor_queue.AddBuffer(map.buffer, input_offset, + image.guest_size_bytes - swizzle.buffer_offset); + update_descriptor_queue.AddBuffer(*data_buffer, offsetof(AstcBufferData, encoding_values), + sizeof(AstcBufferData::encoding_values)); + update_descriptor_queue.AddBuffer(*data_buffer, offsetof(AstcBufferData, replicate_6_to_8), + sizeof(AstcBufferData::replicate_6_to_8)); + update_descriptor_queue.AddBuffer(*data_buffer, offsetof(AstcBufferData, replicate_7_to_8), + sizeof(AstcBufferData::replicate_7_to_8)); + update_descriptor_queue.AddBuffer(*data_buffer, offsetof(AstcBufferData, replicate_8_to_8), + sizeof(AstcBufferData::replicate_8_to_8)); + update_descriptor_queue.AddBuffer(*data_buffer, + offsetof(AstcBufferData, replicate_byte_to_16), + sizeof(AstcBufferData::replicate_byte_to_16)); + update_descriptor_queue.AddBuffer(*data_buffer, sizeof(AstcBufferData), + sizeof(SWIZZLE_TABLE)); + update_descriptor_queue.AddImage(image.StorageImageView(swizzle.level)); + + const VkDescriptorSet set = CommitDescriptorSet(update_descriptor_queue); + const VkPipelineLayout vk_layout = *layout; + + // To unswizzle the ASTC data + const auto params = MakeBlockLinearSwizzle2DParams(swizzle, image.info); + ASSERT(params.origin == (std::array<u32, 3>{0, 0, 0})); + ASSERT(params.destination == (std::array<s32, 3>{0, 0, 0})); + scheduler.Record([vk_layout, num_dispatches_x, num_dispatches_y, num_dispatches_z, + block_dims, params, set](vk::CommandBuffer cmdbuf) { + const AstcPushConstants uniforms{ + .blocks_dims = block_dims, + .bytes_per_block_log2 = params.bytes_per_block_log2, + .layer_stride = params.layer_stride, + .block_size = params.block_size, + .x_shift = params.x_shift, + .block_height = params.block_height, + .block_height_mask = params.block_height_mask, + }; + cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_COMPUTE, vk_layout, 0, set, {}); + cmdbuf.PushConstants(vk_layout, VK_SHADER_STAGE_COMPUTE_BIT, uniforms); + cmdbuf.Dispatch(num_dispatches_x, num_dispatches_y, num_dispatches_z); + }); + } + scheduler.Record([vk_image, aspect_mask](vk::CommandBuffer cmdbuf) { + const VkImageMemoryBarrier image_barrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = vk_image, + .subresourceRange{ + .aspectMask = aspect_mask, + .baseMipLevel = 0, + .levelCount = VK_REMAINING_MIP_LEVELS, + .baseArrayLayer = 0, + .layerCount = VK_REMAINING_ARRAY_LAYERS, + }, + }; + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, + VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, 0, image_barrier); + }); +} + } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.h b/src/video_core/renderer_vulkan/vk_compute_pass.h index 17d781d99..5ea187c30 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.h +++ b/src/video_core/renderer_vulkan/vk_compute_pass.h @@ -11,14 +11,21 @@ #include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" +#include "video_core/vulkan_common/vulkan_memory_allocator.h" #include "video_core/vulkan_common/vulkan_wrapper.h" +namespace VideoCommon { +struct SwizzleParameters; +} + namespace Vulkan { class Device; class StagingBufferPool; class VKScheduler; class VKUpdateDescriptorQueue; +class Image; +struct StagingBufferRef; class VKComputePass { public: @@ -77,4 +84,29 @@ private: VKUpdateDescriptorQueue& update_descriptor_queue; }; +class ASTCDecoderPass final : public VKComputePass { +public: + explicit ASTCDecoderPass(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + StagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + MemoryAllocator& memory_allocator_); + ~ASTCDecoderPass(); + + void Assemble(Image& image, const StagingBufferRef& map, + std::span<const VideoCommon::SwizzleParameters> swizzles); + +private: + void MakeDataBuffer(); + + const Device& device; + VKScheduler& scheduler; + StagingBufferPool& staging_buffer_pool; + VKUpdateDescriptorQueue& update_descriptor_queue; + MemoryAllocator& memory_allocator; + + vk::Buffer data_buffer; + MemoryCommit data_buffer_commit; +}; + } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index dfd38f575..df5b7b172 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -241,7 +241,10 @@ RasterizerVulkan::RasterizerVulkan(Core::Frontend::EmuWindow& emu_window_, Tegra staging_pool(device, memory_allocator, scheduler), descriptor_pool(device, scheduler), update_descriptor_queue(device, scheduler), blit_image(device, scheduler, state_tracker, descriptor_pool), - texture_cache_runtime{device, scheduler, memory_allocator, staging_pool, blit_image}, + astc_decoder_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue, + memory_allocator), + texture_cache_runtime{device, scheduler, memory_allocator, + staging_pool, blit_image, astc_decoder_pass}, texture_cache(texture_cache_runtime, *this, maxwell3d, kepler_compute, gpu_memory), buffer_cache_runtime(device, memory_allocator, scheduler, staging_pool, update_descriptor_queue, descriptor_pool), diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.h b/src/video_core/renderer_vulkan/vk_rasterizer.h index acea1ba2d..235afc6f3 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.h +++ b/src/video_core/renderer_vulkan/vk_rasterizer.h @@ -173,6 +173,7 @@ private: VKDescriptorPool descriptor_pool; VKUpdateDescriptorQueue update_descriptor_queue; BlitImageHelper blit_image; + ASTCDecoderPass astc_decoder_pass; GraphicsPipelineCacheKey graphics_key; diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index 22a1014a9..18155e449 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -10,6 +10,7 @@ #include "video_core/engines/fermi_2d.h" #include "video_core/renderer_vulkan/blit_image.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" +#include "video_core/renderer_vulkan/vk_compute_pass.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" @@ -807,7 +808,7 @@ Image::Image(TextureCacheRuntime& runtime, const ImageInfo& info_, GPUVAddr gpu_ commit = runtime.memory_allocator.Commit(buffer, MemoryUsage::DeviceLocal); } if (IsPixelFormatASTC(info.format) && !runtime.device.IsOptimalAstcSupported()) { - flags |= VideoCommon::ImageFlagBits::Converted; + flags |= VideoCommon::ImageFlagBits::AcceleratedUpload; } if (runtime.device.HasDebuggingToolAttached()) { if (image) { @@ -816,6 +817,38 @@ Image::Image(TextureCacheRuntime& runtime, const ImageInfo& info_, GPUVAddr gpu_ buffer.SetObjectNameEXT(VideoCommon::Name(*this).c_str()); } } + static constexpr VkImageViewUsageCreateInfo storage_image_view_usage_create_info{ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_USAGE_CREATE_INFO, + .pNext = nullptr, + .usage = VK_IMAGE_USAGE_STORAGE_BIT, + }; + if (IsPixelFormatASTC(info.format) && !runtime.device.IsOptimalAstcSupported()) { + const auto& device = runtime.device.GetLogical(); + storage_image_views.reserve(info.resources.levels); + for (s32 level = 0; level < info.resources.levels; ++level) { + storage_image_views.push_back(device.CreateImageView(VkImageViewCreateInfo{ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, + .pNext = &storage_image_view_usage_create_info, + .flags = 0, + .image = *image, + .viewType = VK_IMAGE_VIEW_TYPE_2D_ARRAY, + .format = VK_FORMAT_A8B8G8R8_UNORM_PACK32, + .components{ + .r = VK_COMPONENT_SWIZZLE_IDENTITY, + .g = VK_COMPONENT_SWIZZLE_IDENTITY, + .b = VK_COMPONENT_SWIZZLE_IDENTITY, + .a = VK_COMPONENT_SWIZZLE_IDENTITY, + }, + .subresourceRange{ + .aspectMask = aspect_mask, + .baseMipLevel = static_cast<u32>(level), + .levelCount = 1, + .baseArrayLayer = 0, + .layerCount = VK_REMAINING_ARRAY_LAYERS, + }, + })); + } + } } void Image::UploadMemory(const StagingBufferRef& map, std::span<const BufferImageCopy> copies) { @@ -918,7 +951,6 @@ ImageView::ImageView(TextureCacheRuntime& runtime, const VideoCommon::ImageViewI } } const auto format_info = MaxwellToVK::SurfaceFormat(*device, FormatType::Optimal, true, format); - const VkFormat vk_format = format_info.format; const VkImageViewUsageCreateInfo image_view_usage{ .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_USAGE_CREATE_INFO, .pNext = nullptr, @@ -930,7 +962,7 @@ ImageView::ImageView(TextureCacheRuntime& runtime, const VideoCommon::ImageViewI .flags = 0, .image = image.Handle(), .viewType = VkImageViewType{}, - .format = vk_format, + .format = format_info.format, .components{ .r = ComponentSwizzle(swizzle[0]), .g = ComponentSwizzle(swizzle[1]), @@ -982,7 +1014,7 @@ ImageView::ImageView(TextureCacheRuntime& runtime, const VideoCommon::ImageViewI .pNext = nullptr, .flags = 0, .buffer = image.Buffer(), - .format = vk_format, + .format = format_info.format, .offset = 0, // TODO: Redesign buffer cache to support this .range = image.guest_size_bytes, }); @@ -1167,4 +1199,13 @@ Framebuffer::Framebuffer(TextureCacheRuntime& runtime, std::span<ImageView*, NUM } } +void TextureCacheRuntime::AccelerateImageUpload( + Image& image, const StagingBufferRef& map, + std::span<const VideoCommon::SwizzleParameters> swizzles) { + if (IsPixelFormatASTC(image.info.format)) { + return astc_decoder_pass.Assemble(image, map, swizzles); + } + UNREACHABLE(); +} + } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h index 3aee27ce0..628785d5e 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.h +++ b/src/video_core/renderer_vulkan/vk_texture_cache.h @@ -20,6 +20,7 @@ using VideoCommon::Offset2D; using VideoCommon::RenderTargets; using VideoCore::Surface::PixelFormat; +class ASTCDecoderPass; class BlitImageHelper; class Device; class Image; @@ -60,6 +61,7 @@ struct TextureCacheRuntime { MemoryAllocator& memory_allocator; StagingBufferPool& staging_buffer_pool; BlitImageHelper& blit_image_helper; + ASTCDecoderPass& astc_decoder_pass; std::unordered_map<RenderPassKey, vk::RenderPass> renderpass_cache{}; void Finish(); @@ -83,9 +85,7 @@ struct TextureCacheRuntime { } void AccelerateImageUpload(Image&, const StagingBufferRef&, - std::span<const VideoCommon::SwizzleParameters>) { - UNREACHABLE(); - } + std::span<const VideoCommon::SwizzleParameters>); void InsertUploadMemoryBarrier() {} @@ -121,15 +121,26 @@ public: return *buffer; } - [[nodiscard]] VkImageCreateFlags AspectMask() const noexcept { + [[nodiscard]] VkImageAspectFlags AspectMask() const noexcept { return aspect_mask; } + [[nodiscard]] VkImageView StorageImageView(s32 level) const noexcept { + return *storage_image_views[level]; + } + + /// Returns true when the image is already initialized and mark it as initialized + [[nodiscard]] bool ExchangeInitialization() noexcept { + return std::exchange(initialized, true); + } + private: VKScheduler* scheduler; vk::Image image; vk::Buffer buffer; MemoryCommit commit; + vk::ImageView image_view; + std::vector<vk::ImageView> storage_image_views; VkImageAspectFlags aspect_mask = 0; bool initialized = false; }; diff --git a/src/video_core/texture_cache/accelerated_swizzle.h b/src/video_core/texture_cache/accelerated_swizzle.h index 6ec5c78c4..a11c924e1 100644 --- a/src/video_core/texture_cache/accelerated_swizzle.h +++ b/src/video_core/texture_cache/accelerated_swizzle.h @@ -13,8 +13,8 @@ namespace VideoCommon::Accelerated { struct BlockLinearSwizzle2DParams { - std::array<u32, 3> origin; - std::array<s32, 3> destination; + alignas(16) std::array<u32, 3> origin; + alignas(16) std::array<s32, 3> destination; u32 bytes_per_block_log2; u32 layer_stride; u32 block_size; diff --git a/src/video_core/texture_cache/util.cpp b/src/video_core/texture_cache/util.cpp index 2c42d1449..c22dd0148 100644 --- a/src/video_core/texture_cache/util.cpp +++ b/src/video_core/texture_cache/util.cpp @@ -47,7 +47,6 @@ #include "video_core/texture_cache/formatter.h" #include "video_core/texture_cache/samples_helper.h" #include "video_core/texture_cache/util.h" -#include "video_core/textures/astc.h" #include "video_core/textures/decoders.h" namespace VideoCommon { @@ -879,17 +878,8 @@ void ConvertImage(std::span<const u8> input, const ImageInfo& info, std::span<u8 ASSERT(copy.image_extent == mip_size); ASSERT(copy.buffer_row_length == Common::AlignUp(mip_size.width, tile_size.width)); ASSERT(copy.buffer_image_height == Common::AlignUp(mip_size.height, tile_size.height)); - - if (IsPixelFormatASTC(info.format)) { - ASSERT(copy.image_extent.depth == 1); - Tegra::Texture::ASTC::Decompress(input.subspan(copy.buffer_offset), - copy.image_extent.width, copy.image_extent.height, - copy.image_subresource.num_layers, tile_size.width, - tile_size.height, output.subspan(output_offset)); - } else { - DecompressBC4(input.subspan(copy.buffer_offset), copy.image_extent, - output.subspan(output_offset)); - } + DecompressBC4(input.subspan(copy.buffer_offset), copy.image_extent, + output.subspan(output_offset)); copy.buffer_offset = output_offset; copy.buffer_row_length = mip_size.width; copy.buffer_image_height = mip_size.height; diff --git a/src/video_core/textures/astc.cpp b/src/video_core/textures/astc.cpp deleted file mode 100644 index 3625b666c..000000000 --- a/src/video_core/textures/astc.cpp +++ /dev/null @@ -1,1710 +0,0 @@ -// Copyright 2016 The University of North Carolina at Chapel Hill -// -// Licensed under the Apache License, Version 2.0 (the "License"); -// you may not use this file except in compliance with the License. -// You may obtain a copy of the License at -// -// http://www.apache.org/licenses/LICENSE-2.0 -// -// Unless required by applicable law or agreed to in writing, software -// distributed under the License is distributed on an "AS IS" BASIS, -// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -// See the License for the specific language governing permissions and -// limitations under the License. -// -// Please send all BUG REPORTS to <pavel@cs.unc.edu>. -// <http://gamma.cs.unc.edu/FasTC/> - -#include <algorithm> -#include <cassert> -#include <cstring> -#include <span> -#include <vector> - -#include <boost/container/static_vector.hpp> - -#include "common/common_types.h" - -#include "video_core/textures/astc.h" - -namespace { - -/// Count the number of bits set in a number. -constexpr u32 Popcnt(u32 n) { - u32 c = 0; - for (; n; c++) { - n &= n - 1; - } - return c; -} - -} // Anonymous namespace - -class InputBitStream { -public: - constexpr explicit InputBitStream(std::span<const u8> data, size_t start_offset = 0) - : cur_byte{data.data()}, total_bits{data.size()}, next_bit{start_offset % 8} {} - - constexpr size_t GetBitsRead() const { - return bits_read; - } - - constexpr bool ReadBit() { - if (bits_read >= total_bits * 8) { - return 0; - } - const bool bit = ((*cur_byte >> next_bit) & 1) != 0; - ++next_bit; - while (next_bit >= 8) { - next_bit -= 8; - ++cur_byte; - } - ++bits_read; - return bit; - } - - constexpr u32 ReadBits(std::size_t nBits) { - u32 ret = 0; - for (std::size_t i = 0; i < nBits; ++i) { - ret |= (ReadBit() & 1) << i; - } - return ret; - } - - template <std::size_t nBits> - constexpr u32 ReadBits() { - u32 ret = 0; - for (std::size_t i = 0; i < nBits; ++i) { - ret |= (ReadBit() & 1) << i; - } - return ret; - } - -private: - const u8* cur_byte; - size_t total_bits = 0; - size_t next_bit = 0; - size_t bits_read = 0; -}; - -class OutputBitStream { -public: - constexpr explicit OutputBitStream(u8* ptr, std::size_t bits = 0, std::size_t start_offset = 0) - : cur_byte{ptr}, num_bits{bits}, next_bit{start_offset % 8} {} - - constexpr std::size_t GetBitsWritten() const { - return bits_written; - } - - constexpr void WriteBitsR(u32 val, u32 nBits) { - for (u32 i = 0; i < nBits; i++) { - WriteBit((val >> (nBits - i - 1)) & 1); - } - } - - constexpr void WriteBits(u32 val, u32 nBits) { - for (u32 i = 0; i < nBits; i++) { - WriteBit((val >> i) & 1); - } - } - -private: - constexpr void WriteBit(bool b) { - if (bits_written >= num_bits) { - return; - } - - const u32 mask = 1 << next_bit++; - - // clear the bit - *cur_byte &= static_cast<u8>(~mask); - - // Write the bit, if necessary - if (b) - *cur_byte |= static_cast<u8>(mask); - - // Next byte? - if (next_bit >= 8) { - cur_byte += 1; - next_bit = 0; - } - } - - u8* cur_byte; - std::size_t num_bits; - std::size_t bits_written = 0; - std::size_t next_bit = 0; -}; - -template <typename IntType> -class Bits { -public: - explicit Bits(const IntType& v) : m_Bits(v) {} - - Bits(const Bits&) = delete; - Bits& operator=(const Bits&) = delete; - - u8 operator[](u32 bitPos) const { - return static_cast<u8>((m_Bits >> bitPos) & 1); - } - - IntType operator()(u32 start, u32 end) const { - if (start == end) { - return (*this)[start]; - } else if (start > end) { - u32 t = start; - start = end; - end = t; - } - - u64 mask = (1 << (end - start + 1)) - 1; - return (m_Bits >> start) & static_cast<IntType>(mask); - } - -private: - const IntType& m_Bits; -}; - -enum class IntegerEncoding { JustBits, Qus32, Trit }; - -struct IntegerEncodedValue { - constexpr IntegerEncodedValue() = default; - - constexpr IntegerEncodedValue(IntegerEncoding encoding_, u32 num_bits_) - : encoding{encoding_}, num_bits{num_bits_} {} - - constexpr bool MatchesEncoding(const IntegerEncodedValue& other) const { - return encoding == other.encoding && num_bits == other.num_bits; - } - - // Returns the number of bits required to encode nVals values. - u32 GetBitLength(u32 nVals) const { - u32 totalBits = num_bits * nVals; - if (encoding == IntegerEncoding::Trit) { - totalBits += (nVals * 8 + 4) / 5; - } else if (encoding == IntegerEncoding::Qus32) { - totalBits += (nVals * 7 + 2) / 3; - } - return totalBits; - } - - IntegerEncoding encoding{}; - u32 num_bits = 0; - u32 bit_value = 0; - union { - u32 qus32_value = 0; - u32 trit_value; - }; -}; -using IntegerEncodedVector = boost::container::static_vector< - IntegerEncodedValue, 256, - boost::container::static_vector_options< - boost::container::inplace_alignment<alignof(IntegerEncodedValue)>, - boost::container::throw_on_overflow<false>>::type>; - -static void DecodeTritBlock(InputBitStream& bits, IntegerEncodedVector& result, u32 nBitsPerValue) { - // Implement the algorithm in section C.2.12 - std::array<u32, 5> m; - std::array<u32, 5> t; - u32 T; - - // Read the trit encoded block according to - // table C.2.14 - m[0] = bits.ReadBits(nBitsPerValue); - T = bits.ReadBits<2>(); - m[1] = bits.ReadBits(nBitsPerValue); - T |= bits.ReadBits<2>() << 2; - m[2] = bits.ReadBits(nBitsPerValue); - T |= bits.ReadBit() << 4; - m[3] = bits.ReadBits(nBitsPerValue); - T |= bits.ReadBits<2>() << 5; - m[4] = bits.ReadBits(nBitsPerValue); - T |= bits.ReadBit() << 7; - - u32 C = 0; - - Bits<u32> Tb(T); - if (Tb(2, 4) == 7) { - C = (Tb(5, 7) << 2) | Tb(0, 1); - t[4] = t[3] = 2; - } else { - C = Tb(0, 4); - if (Tb(5, 6) == 3) { - t[4] = 2; - t[3] = Tb[7]; - } else { - t[4] = Tb[7]; - t[3] = Tb(5, 6); - } - } - - Bits<u32> Cb(C); - if (Cb(0, 1) == 3) { - t[2] = 2; - t[1] = Cb[4]; - t[0] = (Cb[3] << 1) | (Cb[2] & ~Cb[3]); - } else if (Cb(2, 3) == 3) { - t[2] = 2; - t[1] = 2; - t[0] = Cb(0, 1); - } else { - t[2] = Cb[4]; - t[1] = Cb(2, 3); - t[0] = (Cb[1] << 1) | (Cb[0] & ~Cb[1]); - } - - for (std::size_t i = 0; i < 5; ++i) { - IntegerEncodedValue& val = result.emplace_back(IntegerEncoding::Trit, nBitsPerValue); - val.bit_value = m[i]; - val.trit_value = t[i]; - } -} - -static void DecodeQus32Block(InputBitStream& bits, IntegerEncodedVector& result, - u32 nBitsPerValue) { - // Implement the algorithm in section C.2.12 - u32 m[3]; - u32 q[3]; - u32 Q; - - // Read the trit encoded block according to - // table C.2.15 - m[0] = bits.ReadBits(nBitsPerValue); - Q = bits.ReadBits<3>(); - m[1] = bits.ReadBits(nBitsPerValue); - Q |= bits.ReadBits<2>() << 3; - m[2] = bits.ReadBits(nBitsPerValue); - Q |= bits.ReadBits<2>() << 5; - - Bits<u32> Qb(Q); - if (Qb(1, 2) == 3 && Qb(5, 6) == 0) { - q[0] = q[1] = 4; - q[2] = (Qb[0] << 2) | ((Qb[4] & ~Qb[0]) << 1) | (Qb[3] & ~Qb[0]); - } else { - u32 C = 0; - if (Qb(1, 2) == 3) { - q[2] = 4; - C = (Qb(3, 4) << 3) | ((~Qb(5, 6) & 3) << 1) | Qb[0]; - } else { - q[2] = Qb(5, 6); - C = Qb(0, 4); - } - - Bits<u32> Cb(C); - if (Cb(0, 2) == 5) { - q[1] = 4; - q[0] = Cb(3, 4); - } else { - q[1] = Cb(3, 4); - q[0] = Cb(0, 2); - } - } - - for (std::size_t i = 0; i < 3; ++i) { - IntegerEncodedValue& val = result.emplace_back(IntegerEncoding::Qus32, nBitsPerValue); - val.bit_value = m[i]; - val.qus32_value = q[i]; - } -} - -// Returns a new instance of this struct that corresponds to the -// can take no more than maxval values -static constexpr IntegerEncodedValue CreateEncoding(u32 maxVal) { - while (maxVal > 0) { - u32 check = maxVal + 1; - - // Is maxVal a power of two? - if (!(check & (check - 1))) { - return IntegerEncodedValue(IntegerEncoding::JustBits, Popcnt(maxVal)); - } - - // Is maxVal of the type 3*2^n - 1? - if ((check % 3 == 0) && !((check / 3) & ((check / 3) - 1))) { - return IntegerEncodedValue(IntegerEncoding::Trit, Popcnt(check / 3 - 1)); - } - - // Is maxVal of the type 5*2^n - 1? - if ((check % 5 == 0) && !((check / 5) & ((check / 5) - 1))) { - return IntegerEncodedValue(IntegerEncoding::Qus32, Popcnt(check / 5 - 1)); - } - - // Apparently it can't be represented with a bounded integer sequence... - // just iterate. - maxVal--; - } - return IntegerEncodedValue(IntegerEncoding::JustBits, 0); -} - -static constexpr std::array<IntegerEncodedValue, 256> MakeEncodedValues() { - std::array<IntegerEncodedValue, 256> encodings{}; - for (std::size_t i = 0; i < encodings.size(); ++i) { - encodings[i] = CreateEncoding(static_cast<u32>(i)); - } - return encodings; -} - -static constexpr std::array EncodingsValues = MakeEncodedValues(); - -// Fills result with the values that are encoded in the given -// bitstream. We must know beforehand what the maximum possible -// value is, and how many values we're decoding. -static void DecodeIntegerSequence(IntegerEncodedVector& result, InputBitStream& bits, u32 maxRange, - u32 nValues) { - // Determine encoding parameters - IntegerEncodedValue val = EncodingsValues[maxRange]; - - // Start decoding - u32 nValsDecoded = 0; - while (nValsDecoded < nValues) { - switch (val.encoding) { - case IntegerEncoding::Qus32: - DecodeQus32Block(bits, result, val.num_bits); - nValsDecoded += 3; - break; - - case IntegerEncoding::Trit: - DecodeTritBlock(bits, result, val.num_bits); - nValsDecoded += 5; - break; - - case IntegerEncoding::JustBits: - val.bit_value = bits.ReadBits(val.num_bits); - result.push_back(val); - nValsDecoded++; - break; - } - } -} - -namespace ASTCC { - -struct TexelWeightParams { - u32 m_Width = 0; - u32 m_Height = 0; - bool m_bDualPlane = false; - u32 m_MaxWeight = 0; - bool m_bError = false; - bool m_bVoidExtentLDR = false; - bool m_bVoidExtentHDR = false; - - u32 GetPackedBitSize() const { - // How many indices do we have? - u32 nIdxs = m_Height * m_Width; - if (m_bDualPlane) { - nIdxs *= 2; - } - - return EncodingsValues[m_MaxWeight].GetBitLength(nIdxs); - } - - u32 GetNumWeightValues() const { - u32 ret = m_Width * m_Height; - if (m_bDualPlane) { - ret *= 2; - } - return ret; - } -}; - -static TexelWeightParams DecodeBlockInfo(InputBitStream& strm) { - TexelWeightParams params; - - // Read the entire block mode all at once - u16 modeBits = static_cast<u16>(strm.ReadBits<11>()); - - // Does this match the void extent block mode? - if ((modeBits & 0x01FF) == 0x1FC) { - if (modeBits & 0x200) { - params.m_bVoidExtentHDR = true; - } else { - params.m_bVoidExtentLDR = true; - } - - // Next two bits must be one. - if (!(modeBits & 0x400) || !strm.ReadBit()) { - params.m_bError = true; - } - - return params; - } - - // First check if the last four bits are zero - if ((modeBits & 0xF) == 0) { - params.m_bError = true; - return params; - } - - // If the last two bits are zero, then if bits - // [6-8] are all ones, this is also reserved. - if ((modeBits & 0x3) == 0 && (modeBits & 0x1C0) == 0x1C0) { - params.m_bError = true; - return params; - } - - // Otherwise, there is no error... Figure out the layout - // of the block mode. Layout is determined by a number - // between 0 and 9 corresponding to table C.2.8 of the - // ASTC spec. - u32 layout = 0; - - if ((modeBits & 0x1) || (modeBits & 0x2)) { - // layout is in [0-4] - if (modeBits & 0x8) { - // layout is in [2-4] - if (modeBits & 0x4) { - // layout is in [3-4] - if (modeBits & 0x100) { - layout = 4; - } else { - layout = 3; - } - } else { - layout = 2; - } - } else { - // layout is in [0-1] - if (modeBits & 0x4) { - layout = 1; - } else { - layout = 0; - } - } - } else { - // layout is in [5-9] - if (modeBits & 0x100) { - // layout is in [7-9] - if (modeBits & 0x80) { - // layout is in [7-8] - assert((modeBits & 0x40) == 0U); - if (modeBits & 0x20) { - layout = 8; - } else { - layout = 7; - } - } else { - layout = 9; - } - } else { - // layout is in [5-6] - if (modeBits & 0x80) { - layout = 6; - } else { - layout = 5; - } - } - } - - assert(layout < 10); - - // Determine R - u32 R = !!(modeBits & 0x10); - if (layout < 5) { - R |= (modeBits & 0x3) << 1; - } else { - R |= (modeBits & 0xC) >> 1; - } - assert(2 <= R && R <= 7); - - // Determine width & height - switch (layout) { - case 0: { - u32 A = (modeBits >> 5) & 0x3; - u32 B = (modeBits >> 7) & 0x3; - params.m_Width = B + 4; - params.m_Height = A + 2; - break; - } - - case 1: { - u32 A = (modeBits >> 5) & 0x3; - u32 B = (modeBits >> 7) & 0x3; - params.m_Width = B + 8; - params.m_Height = A + 2; - break; - } - - case 2: { - u32 A = (modeBits >> 5) & 0x3; - u32 B = (modeBits >> 7) & 0x3; - params.m_Width = A + 2; - params.m_Height = B + 8; - break; - } - - case 3: { - u32 A = (modeBits >> 5) & 0x3; - u32 B = (modeBits >> 7) & 0x1; - params.m_Width = A + 2; - params.m_Height = B + 6; - break; - } - - case 4: { - u32 A = (modeBits >> 5) & 0x3; - u32 B = (modeBits >> 7) & 0x1; - params.m_Width = B + 2; - params.m_Height = A + 2; - break; - } - - case 5: { - u32 A = (modeBits >> 5) & 0x3; - params.m_Width = 12; - params.m_Height = A + 2; - break; - } - - case 6: { - u32 A = (modeBits >> 5) & 0x3; - params.m_Width = A + 2; - params.m_Height = 12; - break; - } - - case 7: { - params.m_Width = 6; - params.m_Height = 10; - break; - } - - case 8: { - params.m_Width = 10; - params.m_Height = 6; - break; - } - - case 9: { - u32 A = (modeBits >> 5) & 0x3; - u32 B = (modeBits >> 9) & 0x3; - params.m_Width = A + 6; - params.m_Height = B + 6; - break; - } - - default: - assert(false && "Don't know this layout..."); - params.m_bError = true; - break; - } - - // Determine whether or not we're using dual planes - // and/or high precision layouts. - bool D = (layout != 9) && (modeBits & 0x400); - bool H = (layout != 9) && (modeBits & 0x200); - - if (H) { - const u32 maxWeights[6] = {9, 11, 15, 19, 23, 31}; - params.m_MaxWeight = maxWeights[R - 2]; - } else { - const u32 maxWeights[6] = {1, 2, 3, 4, 5, 7}; - params.m_MaxWeight = maxWeights[R - 2]; - } - - params.m_bDualPlane = D; - - return params; -} - -static void FillVoidExtentLDR(InputBitStream& strm, std::span<u32> outBuf, u32 blockWidth, - u32 blockHeight) { - // Don't actually care about the void extent, just read the bits... - for (s32 i = 0; i < 4; ++i) { - strm.ReadBits<13>(); - } - - // Decode the RGBA components and renormalize them to the range [0, 255] - u16 r = static_cast<u16>(strm.ReadBits<16>()); - u16 g = static_cast<u16>(strm.ReadBits<16>()); - u16 b = static_cast<u16>(strm.ReadBits<16>()); - u16 a = static_cast<u16>(strm.ReadBits<16>()); - - u32 rgba = (r >> 8) | (g & 0xFF00) | (static_cast<u32>(b) & 0xFF00) << 8 | - (static_cast<u32>(a) & 0xFF00) << 16; - - for (u32 j = 0; j < blockHeight; j++) { - for (u32 i = 0; i < blockWidth; i++) { - outBuf[j * blockWidth + i] = rgba; - } - } -} - -static void FillError(std::span<u32> outBuf, u32 blockWidth, u32 blockHeight) { - for (u32 j = 0; j < blockHeight; j++) { - for (u32 i = 0; i < blockWidth; i++) { - outBuf[j * blockWidth + i] = 0xFFFF00FF; - } - } -} - -// Replicates low numBits such that [(toBit - 1):(toBit - 1 - fromBit)] -// is the same as [(numBits - 1):0] and repeats all the way down. -template <typename IntType> -static constexpr IntType Replicate(IntType val, u32 numBits, u32 toBit) { - if (numBits == 0) { - return 0; - } - if (toBit == 0) { - return 0; - } - const IntType v = val & static_cast<IntType>((1 << numBits) - 1); - IntType res = v; - u32 reslen = numBits; - while (reslen < toBit) { - u32 comp = 0; - if (numBits > toBit - reslen) { - u32 newshift = toBit - reslen; - comp = numBits - newshift; - numBits = newshift; - } - res = static_cast<IntType>(res << numBits); - res = static_cast<IntType>(res | (v >> comp)); - reslen += numBits; - } - return res; -} - -static constexpr std::size_t NumReplicateEntries(u32 num_bits) { - return std::size_t(1) << num_bits; -} - -template <typename IntType, u32 num_bits, u32 to_bit> -static constexpr auto MakeReplicateTable() { - std::array<IntType, NumReplicateEntries(num_bits)> table{}; - for (IntType value = 0; value < static_cast<IntType>(std::size(table)); ++value) { - table[value] = Replicate(value, num_bits, to_bit); - } - return table; -} - -static constexpr auto REPLICATE_BYTE_TO_16_TABLE = MakeReplicateTable<u32, 8, 16>(); -static constexpr u32 ReplicateByteTo16(std::size_t value) { - return REPLICATE_BYTE_TO_16_TABLE[value]; -} - -static constexpr auto REPLICATE_BIT_TO_7_TABLE = MakeReplicateTable<u32, 1, 7>(); -static constexpr u32 ReplicateBitTo7(std::size_t value) { - return REPLICATE_BIT_TO_7_TABLE[value]; -} - -static constexpr auto REPLICATE_BIT_TO_9_TABLE = MakeReplicateTable<u32, 1, 9>(); -static constexpr u32 ReplicateBitTo9(std::size_t value) { - return REPLICATE_BIT_TO_9_TABLE[value]; -} - -static constexpr auto REPLICATE_1_BIT_TO_8_TABLE = MakeReplicateTable<u32, 1, 8>(); -static constexpr auto REPLICATE_2_BIT_TO_8_TABLE = MakeReplicateTable<u32, 2, 8>(); -static constexpr auto REPLICATE_3_BIT_TO_8_TABLE = MakeReplicateTable<u32, 3, 8>(); -static constexpr auto REPLICATE_4_BIT_TO_8_TABLE = MakeReplicateTable<u32, 4, 8>(); -static constexpr auto REPLICATE_5_BIT_TO_8_TABLE = MakeReplicateTable<u32, 5, 8>(); -static constexpr auto REPLICATE_6_BIT_TO_8_TABLE = MakeReplicateTable<u32, 6, 8>(); -static constexpr auto REPLICATE_7_BIT_TO_8_TABLE = MakeReplicateTable<u32, 7, 8>(); -static constexpr auto REPLICATE_8_BIT_TO_8_TABLE = MakeReplicateTable<u32, 8, 8>(); -/// Use a precompiled table with the most common usages, if it's not in the expected range, fallback -/// to the runtime implementation -static constexpr u32 FastReplicateTo8(u32 value, u32 num_bits) { - switch (num_bits) { - case 1: - return REPLICATE_1_BIT_TO_8_TABLE[value]; - case 2: - return REPLICATE_2_BIT_TO_8_TABLE[value]; - case 3: - return REPLICATE_3_BIT_TO_8_TABLE[value]; - case 4: - return REPLICATE_4_BIT_TO_8_TABLE[value]; - case 5: - return REPLICATE_5_BIT_TO_8_TABLE[value]; - case 6: - return REPLICATE_6_BIT_TO_8_TABLE[value]; - case 7: - return REPLICATE_7_BIT_TO_8_TABLE[value]; - case 8: - return REPLICATE_8_BIT_TO_8_TABLE[value]; - default: - return Replicate(value, num_bits, 8); - } -} - -static constexpr auto REPLICATE_1_BIT_TO_6_TABLE = MakeReplicateTable<u32, 1, 6>(); -static constexpr auto REPLICATE_2_BIT_TO_6_TABLE = MakeReplicateTable<u32, 2, 6>(); -static constexpr auto REPLICATE_3_BIT_TO_6_TABLE = MakeReplicateTable<u32, 3, 6>(); -static constexpr auto REPLICATE_4_BIT_TO_6_TABLE = MakeReplicateTable<u32, 4, 6>(); -static constexpr auto REPLICATE_5_BIT_TO_6_TABLE = MakeReplicateTable<u32, 5, 6>(); -static constexpr u32 FastReplicateTo6(u32 value, u32 num_bits) { - switch (num_bits) { - case 1: - return REPLICATE_1_BIT_TO_6_TABLE[value]; - case 2: - return REPLICATE_2_BIT_TO_6_TABLE[value]; - case 3: - return REPLICATE_3_BIT_TO_6_TABLE[value]; - case 4: - return REPLICATE_4_BIT_TO_6_TABLE[value]; - case 5: - return REPLICATE_5_BIT_TO_6_TABLE[value]; - default: - return Replicate(value, num_bits, 6); - } -} - -class Pixel { -protected: - using ChannelType = s16; - u8 m_BitDepth[4] = {8, 8, 8, 8}; - s16 color[4] = {}; - -public: - Pixel() = default; - Pixel(u32 a, u32 r, u32 g, u32 b, u32 bitDepth = 8) - : m_BitDepth{u8(bitDepth), u8(bitDepth), u8(bitDepth), u8(bitDepth)}, - color{static_cast<ChannelType>(a), static_cast<ChannelType>(r), - static_cast<ChannelType>(g), static_cast<ChannelType>(b)} {} - - // Changes the depth of each pixel. This scales the values to - // the appropriate bit depth by either truncating the least - // significant bits when going from larger to smaller bit depth - // or by repeating the most significant bits when going from - // smaller to larger bit depths. - void ChangeBitDepth() { - for (u32 i = 0; i < 4; i++) { - Component(i) = ChangeBitDepth(Component(i), m_BitDepth[i]); - m_BitDepth[i] = 8; - } - } - - template <typename IntType> - static float ConvertChannelToFloat(IntType channel, u8 bitDepth) { - float denominator = static_cast<float>((1 << bitDepth) - 1); - return static_cast<float>(channel) / denominator; - } - - // Changes the bit depth of a single component. See the comment - // above for how we do this. - static ChannelType ChangeBitDepth(Pixel::ChannelType val, u8 oldDepth) { - assert(oldDepth <= 8); - - if (oldDepth == 8) { - // Do nothing - return val; - } else if (oldDepth == 0) { - return static_cast<ChannelType>((1 << 8) - 1); - } else if (8 > oldDepth) { - return static_cast<ChannelType>(FastReplicateTo8(static_cast<u32>(val), oldDepth)); - } else { - // oldDepth > newDepth - const u8 bitsWasted = static_cast<u8>(oldDepth - 8); - u16 v = static_cast<u16>(val); - v = static_cast<u16>((v + (1 << (bitsWasted - 1))) >> bitsWasted); - v = ::std::min<u16>(::std::max<u16>(0, v), static_cast<u16>((1 << 8) - 1)); - return static_cast<u8>(v); - } - - assert(false && "We shouldn't get here."); - return 0; - } - - const ChannelType& A() const { - return color[0]; - } - ChannelType& A() { - return color[0]; - } - const ChannelType& R() const { - return color[1]; - } - ChannelType& R() { - return color[1]; - } - const ChannelType& G() const { - return color[2]; - } - ChannelType& G() { - return color[2]; - } - const ChannelType& B() const { - return color[3]; - } - ChannelType& B() { - return color[3]; - } - const ChannelType& Component(u32 idx) const { - return color[idx]; - } - ChannelType& Component(u32 idx) { - return color[idx]; - } - - void GetBitDepth(u8 (&outDepth)[4]) const { - for (s32 i = 0; i < 4; i++) { - outDepth[i] = m_BitDepth[i]; - } - } - - // Take all of the components, transform them to their 8-bit variants, - // and then pack each channel into an R8G8B8A8 32-bit integer. We assume - // that the architecture is little-endian, so the alpha channel will end - // up in the most-significant byte. - u32 Pack() const { - Pixel eightBit(*this); - eightBit.ChangeBitDepth(); - - u32 r = 0; - r |= eightBit.A(); - r <<= 8; - r |= eightBit.B(); - r <<= 8; - r |= eightBit.G(); - r <<= 8; - r |= eightBit.R(); - return r; - } - - // Clamps the pixel to the range [0,255] - void ClampByte() { - for (u32 i = 0; i < 4; i++) { - color[i] = (color[i] < 0) ? 0 : ((color[i] > 255) ? 255 : color[i]); - } - } - - void MakeOpaque() { - A() = 255; - } -}; - -static void DecodeColorValues(u32* out, std::span<u8> data, const u32* modes, const u32 nPartitions, - const u32 nBitsForColorData) { - // First figure out how many color values we have - u32 nValues = 0; - for (u32 i = 0; i < nPartitions; i++) { - nValues += ((modes[i] >> 2) + 1) << 1; - } - - // Then based on the number of values and the remaining number of bits, - // figure out the max value for each of them... - u32 range = 256; - while (--range > 0) { - IntegerEncodedValue val = EncodingsValues[range]; - u32 bitLength = val.GetBitLength(nValues); - if (bitLength <= nBitsForColorData) { - // Find the smallest possible range that matches the given encoding - while (--range > 0) { - IntegerEncodedValue newval = EncodingsValues[range]; - if (!newval.MatchesEncoding(val)) { - break; - } - } - - // Return to last matching range. - range++; - break; - } - } - - // We now have enough to decode our integer sequence. - IntegerEncodedVector decodedColorValues; - - InputBitStream colorStream(data, 0); - DecodeIntegerSequence(decodedColorValues, colorStream, range, nValues); - - // Once we have the decoded values, we need to dequantize them to the 0-255 range - // This procedure is outlined in ASTC spec C.2.13 - u32 outIdx = 0; - for (auto itr = decodedColorValues.begin(); itr != decodedColorValues.end(); ++itr) { - // Have we already decoded all that we need? - if (outIdx >= nValues) { - break; - } - - const IntegerEncodedValue& val = *itr; - u32 bitlen = val.num_bits; - u32 bitval = val.bit_value; - - assert(bitlen >= 1); - - u32 A = 0, B = 0, C = 0, D = 0; - // A is just the lsb replicated 9 times. - A = ReplicateBitTo9(bitval & 1); - - switch (val.encoding) { - // Replicate bits - case IntegerEncoding::JustBits: - out[outIdx++] = FastReplicateTo8(bitval, bitlen); - break; - - // Use algorithm in C.2.13 - case IntegerEncoding::Trit: { - - D = val.trit_value; - - switch (bitlen) { - case 1: { - C = 204; - } break; - - case 2: { - C = 93; - // B = b000b0bb0 - u32 b = (bitval >> 1) & 1; - B = (b << 8) | (b << 4) | (b << 2) | (b << 1); - } break; - - case 3: { - C = 44; - // B = cb000cbcb - u32 cb = (bitval >> 1) & 3; - B = (cb << 7) | (cb << 2) | cb; - } break; - - case 4: { - C = 22; - // B = dcb000dcb - u32 dcb = (bitval >> 1) & 7; - B = (dcb << 6) | dcb; - } break; - - case 5: { - C = 11; - // B = edcb000ed - u32 edcb = (bitval >> 1) & 0xF; - B = (edcb << 5) | (edcb >> 2); - } break; - - case 6: { - C = 5; - // B = fedcb000f - u32 fedcb = (bitval >> 1) & 0x1F; - B = (fedcb << 4) | (fedcb >> 4); - } break; - - default: - assert(false && "Unsupported trit encoding for color values!"); - break; - } // switch(bitlen) - } // case IntegerEncoding::Trit - break; - - case IntegerEncoding::Qus32: { - - D = val.qus32_value; - - switch (bitlen) { - case 1: { - C = 113; - } break; - - case 2: { - C = 54; - // B = b0000bb00 - u32 b = (bitval >> 1) & 1; - B = (b << 8) | (b << 3) | (b << 2); - } break; - - case 3: { - C = 26; - // B = cb0000cbc - u32 cb = (bitval >> 1) & 3; - B = (cb << 7) | (cb << 1) | (cb >> 1); - } break; - - case 4: { - C = 13; - // B = dcb0000dc - u32 dcb = (bitval >> 1) & 7; - B = (dcb << 6) | (dcb >> 1); - } break; - - case 5: { - C = 6; - // B = edcb0000e - u32 edcb = (bitval >> 1) & 0xF; - B = (edcb << 5) | (edcb >> 3); - } break; - - default: - assert(false && "Unsupported quint encoding for color values!"); - break; - } // switch(bitlen) - } // case IntegerEncoding::Qus32 - break; - } // switch(val.encoding) - - if (val.encoding != IntegerEncoding::JustBits) { - u32 T = D * C + B; - T ^= A; - T = (A & 0x80) | (T >> 2); - out[outIdx++] = T; - } - } - - // Make sure that each of our values is in the proper range... - for (u32 i = 0; i < nValues; i++) { - assert(out[i] <= 255); - } -} - -static u32 UnquantizeTexelWeight(const IntegerEncodedValue& val) { - u32 bitval = val.bit_value; - u32 bitlen = val.num_bits; - - u32 A = ReplicateBitTo7(bitval & 1); - u32 B = 0, C = 0, D = 0; - - u32 result = 0; - switch (val.encoding) { - case IntegerEncoding::JustBits: - result = FastReplicateTo6(bitval, bitlen); - break; - - case IntegerEncoding::Trit: { - D = val.trit_value; - assert(D < 3); - - switch (bitlen) { - case 0: { - u32 results[3] = {0, 32, 63}; - result = results[D]; - } break; - - case 1: { - C = 50; - } break; - - case 2: { - C = 23; - u32 b = (bitval >> 1) & 1; - B = (b << 6) | (b << 2) | b; - } break; - - case 3: { - C = 11; - u32 cb = (bitval >> 1) & 3; - B = (cb << 5) | cb; - } break; - - default: - assert(false && "Invalid trit encoding for texel weight"); - break; - } - } break; - - case IntegerEncoding::Qus32: { - D = val.qus32_value; - assert(D < 5); - - switch (bitlen) { - case 0: { - u32 results[5] = {0, 16, 32, 47, 63}; - result = results[D]; - } break; - - case 1: { - C = 28; - } break; - - case 2: { - C = 13; - u32 b = (bitval >> 1) & 1; - B = (b << 6) | (b << 1); - } break; - - default: - assert(false && "Invalid quint encoding for texel weight"); - break; - } - } break; - } - - if (val.encoding != IntegerEncoding::JustBits && bitlen > 0) { - // Decode the value... - result = D * C + B; - result ^= A; - result = (A & 0x20) | (result >> 2); - } - - assert(result < 64); - - // Change from [0,63] to [0,64] - if (result > 32) { - result += 1; - } - - return result; -} - -static void UnquantizeTexelWeights(u32 out[2][144], const IntegerEncodedVector& weights, - const TexelWeightParams& params, const u32 blockWidth, - const u32 blockHeight) { - u32 weightIdx = 0; - u32 unquantized[2][144]; - - for (auto itr = weights.begin(); itr != weights.end(); ++itr) { - unquantized[0][weightIdx] = UnquantizeTexelWeight(*itr); - - if (params.m_bDualPlane) { - ++itr; - unquantized[1][weightIdx] = UnquantizeTexelWeight(*itr); - if (itr == weights.end()) { - break; - } - } - - if (++weightIdx >= (params.m_Width * params.m_Height)) - break; - } - - // Do infill if necessary (Section C.2.18) ... - u32 Ds = (1024 + (blockWidth / 2)) / (blockWidth - 1); - u32 Dt = (1024 + (blockHeight / 2)) / (blockHeight - 1); - - const u32 kPlaneScale = params.m_bDualPlane ? 2U : 1U; - for (u32 plane = 0; plane < kPlaneScale; plane++) - for (u32 t = 0; t < blockHeight; t++) - for (u32 s = 0; s < blockWidth; s++) { - u32 cs = Ds * s; - u32 ct = Dt * t; - - u32 gs = (cs * (params.m_Width - 1) + 32) >> 6; - u32 gt = (ct * (params.m_Height - 1) + 32) >> 6; - - u32 js = gs >> 4; - u32 fs = gs & 0xF; - - u32 jt = gt >> 4; - u32 ft = gt & 0x0F; - - u32 w11 = (fs * ft + 8) >> 4; - u32 w10 = ft - w11; - u32 w01 = fs - w11; - u32 w00 = 16 - fs - ft + w11; - - u32 v0 = js + jt * params.m_Width; - -#define FIND_TEXEL(tidx, bidx) \ - u32 p##bidx = 0; \ - do { \ - if ((tidx) < (params.m_Width * params.m_Height)) { \ - p##bidx = unquantized[plane][(tidx)]; \ - } \ - } while (0) - - FIND_TEXEL(v0, 00); - FIND_TEXEL(v0 + 1, 01); - FIND_TEXEL(v0 + params.m_Width, 10); - FIND_TEXEL(v0 + params.m_Width + 1, 11); - -#undef FIND_TEXEL - - out[plane][t * blockWidth + s] = - (p00 * w00 + p01 * w01 + p10 * w10 + p11 * w11 + 8) >> 4; - } -} - -// Transfers a bit as described in C.2.14 -static inline void BitTransferSigned(s32& a, s32& b) { - b >>= 1; - b |= a & 0x80; - a >>= 1; - a &= 0x3F; - if (a & 0x20) - a -= 0x40; -} - -// Adds more precision to the blue channel as described -// in C.2.14 -static inline Pixel BlueContract(s32 a, s32 r, s32 g, s32 b) { - return Pixel(static_cast<s16>(a), static_cast<s16>((r + b) >> 1), - static_cast<s16>((g + b) >> 1), static_cast<s16>(b)); -} - -// Partition selection functions as specified in -// C.2.21 -static inline u32 hash52(u32 p) { - p ^= p >> 15; - p -= p << 17; - p += p << 7; - p += p << 4; - p ^= p >> 5; - p += p << 16; - p ^= p >> 7; - p ^= p >> 3; - p ^= p << 6; - p ^= p >> 17; - return p; -} - -static u32 SelectPartition(s32 seed, s32 x, s32 y, s32 z, s32 partitionCount, s32 smallBlock) { - if (1 == partitionCount) - return 0; - - if (smallBlock) { - x <<= 1; - y <<= 1; - z <<= 1; - } - - seed += (partitionCount - 1) * 1024; - - u32 rnum = hash52(static_cast<u32>(seed)); - u8 seed1 = static_cast<u8>(rnum & 0xF); - u8 seed2 = static_cast<u8>((rnum >> 4) & 0xF); - u8 seed3 = static_cast<u8>((rnum >> 8) & 0xF); - u8 seed4 = static_cast<u8>((rnum >> 12) & 0xF); - u8 seed5 = static_cast<u8>((rnum >> 16) & 0xF); - u8 seed6 = static_cast<u8>((rnum >> 20) & 0xF); - u8 seed7 = static_cast<u8>((rnum >> 24) & 0xF); - u8 seed8 = static_cast<u8>((rnum >> 28) & 0xF); - u8 seed9 = static_cast<u8>((rnum >> 18) & 0xF); - u8 seed10 = static_cast<u8>((rnum >> 22) & 0xF); - u8 seed11 = static_cast<u8>((rnum >> 26) & 0xF); - u8 seed12 = static_cast<u8>(((rnum >> 30) | (rnum << 2)) & 0xF); - - seed1 = static_cast<u8>(seed1 * seed1); - seed2 = static_cast<u8>(seed2 * seed2); - seed3 = static_cast<u8>(seed3 * seed3); - seed4 = static_cast<u8>(seed4 * seed4); - seed5 = static_cast<u8>(seed5 * seed5); - seed6 = static_cast<u8>(seed6 * seed6); - seed7 = static_cast<u8>(seed7 * seed7); - seed8 = static_cast<u8>(seed8 * seed8); - seed9 = static_cast<u8>(seed9 * seed9); - seed10 = static_cast<u8>(seed10 * seed10); - seed11 = static_cast<u8>(seed11 * seed11); - seed12 = static_cast<u8>(seed12 * seed12); - - s32 sh1, sh2, sh3; - if (seed & 1) { - sh1 = (seed & 2) ? 4 : 5; - sh2 = (partitionCount == 3) ? 6 : 5; - } else { - sh1 = (partitionCount == 3) ? 6 : 5; - sh2 = (seed & 2) ? 4 : 5; - } - sh3 = (seed & 0x10) ? sh1 : sh2; - - seed1 = static_cast<u8>(seed1 >> sh1); - seed2 = static_cast<u8>(seed2 >> sh2); - seed3 = static_cast<u8>(seed3 >> sh1); - seed4 = static_cast<u8>(seed4 >> sh2); - seed5 = static_cast<u8>(seed5 >> sh1); - seed6 = static_cast<u8>(seed6 >> sh2); - seed7 = static_cast<u8>(seed7 >> sh1); - seed8 = static_cast<u8>(seed8 >> sh2); - seed9 = static_cast<u8>(seed9 >> sh3); - seed10 = static_cast<u8>(seed10 >> sh3); - seed11 = static_cast<u8>(seed11 >> sh3); - seed12 = static_cast<u8>(seed12 >> sh3); - - s32 a = seed1 * x + seed2 * y + seed11 * z + (rnum >> 14); - s32 b = seed3 * x + seed4 * y + seed12 * z + (rnum >> 10); - s32 c = seed5 * x + seed6 * y + seed9 * z + (rnum >> 6); - s32 d = seed7 * x + seed8 * y + seed10 * z + (rnum >> 2); - - a &= 0x3F; - b &= 0x3F; - c &= 0x3F; - d &= 0x3F; - - if (partitionCount < 4) - d = 0; - if (partitionCount < 3) - c = 0; - - if (a >= b && a >= c && a >= d) - return 0; - else if (b >= c && b >= d) - return 1; - else if (c >= d) - return 2; - return 3; -} - -static inline u32 Select2DPartition(s32 seed, s32 x, s32 y, s32 partitionCount, s32 smallBlock) { - return SelectPartition(seed, x, y, 0, partitionCount, smallBlock); -} - -// Section C.2.14 -static void ComputeEndpos32s(Pixel& ep1, Pixel& ep2, const u32*& colorValues, - u32 colorEndpos32Mode) { -#define READ_UINT_VALUES(N) \ - u32 v[N]; \ - for (u32 i = 0; i < N; i++) { \ - v[i] = *(colorValues++); \ - } - -#define READ_INT_VALUES(N) \ - s32 v[N]; \ - for (u32 i = 0; i < N; i++) { \ - v[i] = static_cast<s32>(*(colorValues++)); \ - } - - switch (colorEndpos32Mode) { - case 0: { - READ_UINT_VALUES(2) - ep1 = Pixel(0xFF, v[0], v[0], v[0]); - ep2 = Pixel(0xFF, v[1], v[1], v[1]); - } break; - - case 1: { - READ_UINT_VALUES(2) - u32 L0 = (v[0] >> 2) | (v[1] & 0xC0); - u32 L1 = std::max(L0 + (v[1] & 0x3F), 0xFFU); - ep1 = Pixel(0xFF, L0, L0, L0); - ep2 = Pixel(0xFF, L1, L1, L1); - } break; - - case 4: { - READ_UINT_VALUES(4) - ep1 = Pixel(v[2], v[0], v[0], v[0]); - ep2 = Pixel(v[3], v[1], v[1], v[1]); - } break; - - case 5: { - READ_INT_VALUES(4) - BitTransferSigned(v[1], v[0]); - BitTransferSigned(v[3], v[2]); - ep1 = Pixel(v[2], v[0], v[0], v[0]); - ep2 = Pixel(v[2] + v[3], v[0] + v[1], v[0] + v[1], v[0] + v[1]); - ep1.ClampByte(); - ep2.ClampByte(); - } break; - - case 6: { - READ_UINT_VALUES(4) - ep1 = Pixel(0xFF, v[0] * v[3] >> 8, v[1] * v[3] >> 8, v[2] * v[3] >> 8); - ep2 = Pixel(0xFF, v[0], v[1], v[2]); - } break; - - case 8: { - READ_UINT_VALUES(6) - if (v[1] + v[3] + v[5] >= v[0] + v[2] + v[4]) { - ep1 = Pixel(0xFF, v[0], v[2], v[4]); - ep2 = Pixel(0xFF, v[1], v[3], v[5]); - } else { - ep1 = BlueContract(0xFF, v[1], v[3], v[5]); - ep2 = BlueContract(0xFF, v[0], v[2], v[4]); - } - } break; - - case 9: { - READ_INT_VALUES(6) - BitTransferSigned(v[1], v[0]); - BitTransferSigned(v[3], v[2]); - BitTransferSigned(v[5], v[4]); - if (v[1] + v[3] + v[5] >= 0) { - ep1 = Pixel(0xFF, v[0], v[2], v[4]); - ep2 = Pixel(0xFF, v[0] + v[1], v[2] + v[3], v[4] + v[5]); - } else { - ep1 = BlueContract(0xFF, v[0] + v[1], v[2] + v[3], v[4] + v[5]); - ep2 = BlueContract(0xFF, v[0], v[2], v[4]); - } - ep1.ClampByte(); - ep2.ClampByte(); - } break; - - case 10: { - READ_UINT_VALUES(6) - ep1 = Pixel(v[4], v[0] * v[3] >> 8, v[1] * v[3] >> 8, v[2] * v[3] >> 8); - ep2 = Pixel(v[5], v[0], v[1], v[2]); - } break; - - case 12: { - READ_UINT_VALUES(8) - if (v[1] + v[3] + v[5] >= v[0] + v[2] + v[4]) { - ep1 = Pixel(v[6], v[0], v[2], v[4]); - ep2 = Pixel(v[7], v[1], v[3], v[5]); - } else { - ep1 = BlueContract(v[7], v[1], v[3], v[5]); - ep2 = BlueContract(v[6], v[0], v[2], v[4]); - } - } break; - - case 13: { - READ_INT_VALUES(8) - BitTransferSigned(v[1], v[0]); - BitTransferSigned(v[3], v[2]); - BitTransferSigned(v[5], v[4]); - BitTransferSigned(v[7], v[6]); - if (v[1] + v[3] + v[5] >= 0) { - ep1 = Pixel(v[6], v[0], v[2], v[4]); - ep2 = Pixel(v[7] + v[6], v[0] + v[1], v[2] + v[3], v[4] + v[5]); - } else { - ep1 = BlueContract(v[6] + v[7], v[0] + v[1], v[2] + v[3], v[4] + v[5]); - ep2 = BlueContract(v[6], v[0], v[2], v[4]); - } - ep1.ClampByte(); - ep2.ClampByte(); - } break; - - default: - assert(false && "Unsupported color endpoint mode (is it HDR?)"); - break; - } - -#undef READ_UINT_VALUES -#undef READ_INT_VALUES -} - -static void DecompressBlock(std::span<const u8, 16> inBuf, const u32 blockWidth, - const u32 blockHeight, std::span<u32, 12 * 12> outBuf) { - InputBitStream strm(inBuf); - TexelWeightParams weightParams = DecodeBlockInfo(strm); - - // Was there an error? - if (weightParams.m_bError) { - assert(false && "Invalid block mode"); - FillError(outBuf, blockWidth, blockHeight); - return; - } - - if (weightParams.m_bVoidExtentLDR) { - FillVoidExtentLDR(strm, outBuf, blockWidth, blockHeight); - return; - } - - if (weightParams.m_bVoidExtentHDR) { - assert(false && "HDR void extent blocks are unsupported!"); - FillError(outBuf, blockWidth, blockHeight); - return; - } - - if (weightParams.m_Width > blockWidth) { - assert(false && "Texel weight grid width should be smaller than block width"); - FillError(outBuf, blockWidth, blockHeight); - return; - } - - if (weightParams.m_Height > blockHeight) { - assert(false && "Texel weight grid height should be smaller than block height"); - FillError(outBuf, blockWidth, blockHeight); - return; - } - - // Read num partitions - u32 nPartitions = strm.ReadBits<2>() + 1; - assert(nPartitions <= 4); - - if (nPartitions == 4 && weightParams.m_bDualPlane) { - assert(false && "Dual plane mode is incompatible with four partition blocks"); - FillError(outBuf, blockWidth, blockHeight); - return; - } - - // Based on the number of partitions, read the color endpos32 mode for - // each partition. - - // Determine partitions, partition index, and color endpos32 modes - s32 planeIdx = -1; - u32 partitionIndex; - u32 colorEndpos32Mode[4] = {0, 0, 0, 0}; - - // Define color data. - u8 colorEndpos32Data[16]; - memset(colorEndpos32Data, 0, sizeof(colorEndpos32Data)); - OutputBitStream colorEndpos32Stream(colorEndpos32Data, 16 * 8, 0); - - // Read extra config data... - u32 baseCEM = 0; - if (nPartitions == 1) { - colorEndpos32Mode[0] = strm.ReadBits<4>(); - partitionIndex = 0; - } else { - partitionIndex = strm.ReadBits<10>(); - baseCEM = strm.ReadBits<6>(); - } - u32 baseMode = (baseCEM & 3); - - // Remaining bits are color endpos32 data... - u32 nWeightBits = weightParams.GetPackedBitSize(); - s32 remainingBits = 128 - nWeightBits - static_cast<s32>(strm.GetBitsRead()); - - // Consider extra bits prior to texel data... - u32 extraCEMbits = 0; - if (baseMode) { - switch (nPartitions) { - case 2: - extraCEMbits += 2; - break; - case 3: - extraCEMbits += 5; - break; - case 4: - extraCEMbits += 8; - break; - default: - assert(false); - break; - } - } - remainingBits -= extraCEMbits; - - // Do we have a dual plane situation? - u32 planeSelectorBits = 0; - if (weightParams.m_bDualPlane) { - planeSelectorBits = 2; - } - remainingBits -= planeSelectorBits; - - // Read color data... - u32 colorDataBits = remainingBits; - while (remainingBits > 0) { - u32 nb = std::min(remainingBits, 8); - u32 b = strm.ReadBits(nb); - colorEndpos32Stream.WriteBits(b, nb); - remainingBits -= 8; - } - - // Read the plane selection bits - planeIdx = strm.ReadBits(planeSelectorBits); - - // Read the rest of the CEM - if (baseMode) { - u32 extraCEM = strm.ReadBits(extraCEMbits); - u32 CEM = (extraCEM << 6) | baseCEM; - CEM >>= 2; - - bool C[4] = {0}; - for (u32 i = 0; i < nPartitions; i++) { - C[i] = CEM & 1; - CEM >>= 1; - } - - u8 M[4] = {0}; - for (u32 i = 0; i < nPartitions; i++) { - M[i] = CEM & 3; - CEM >>= 2; - assert(M[i] <= 3); - } - - for (u32 i = 0; i < nPartitions; i++) { - colorEndpos32Mode[i] = baseMode; - if (!(C[i])) - colorEndpos32Mode[i] -= 1; - colorEndpos32Mode[i] <<= 2; - colorEndpos32Mode[i] |= M[i]; - } - } else if (nPartitions > 1) { - u32 CEM = baseCEM >> 2; - for (u32 i = 0; i < nPartitions; i++) { - colorEndpos32Mode[i] = CEM; - } - } - - // Make sure everything up till here is sane. - for (u32 i = 0; i < nPartitions; i++) { - assert(colorEndpos32Mode[i] < 16); - } - assert(strm.GetBitsRead() + weightParams.GetPackedBitSize() == 128); - - // Decode both color data and texel weight data - u32 colorValues[32]; // Four values, two endpos32s, four maximum paritions - DecodeColorValues(colorValues, colorEndpos32Data, colorEndpos32Mode, nPartitions, - colorDataBits); - - Pixel endpos32s[4][2]; - const u32* colorValuesPtr = colorValues; - for (u32 i = 0; i < nPartitions; i++) { - ComputeEndpos32s(endpos32s[i][0], endpos32s[i][1], colorValuesPtr, colorEndpos32Mode[i]); - } - - // Read the texel weight data.. - std::array<u8, 16> texelWeightData; - std::ranges::copy(inBuf, texelWeightData.begin()); - - // Reverse everything - for (u32 i = 0; i < 8; i++) { -// Taken from http://graphics.stanford.edu/~seander/bithacks.html#ReverseByteWith64Bits -#define REVERSE_BYTE(b) (((b)*0x80200802ULL) & 0x0884422110ULL) * 0x0101010101ULL >> 32 - u8 a = static_cast<u8>(REVERSE_BYTE(texelWeightData[i])); - u8 b = static_cast<u8>(REVERSE_BYTE(texelWeightData[15 - i])); -#undef REVERSE_BYTE - - texelWeightData[i] = b; - texelWeightData[15 - i] = a; - } - - // Make sure that higher non-texel bits are set to zero - const u32 clearByteStart = (weightParams.GetPackedBitSize() >> 3) + 1; - if (clearByteStart > 0 && clearByteStart <= texelWeightData.size()) { - texelWeightData[clearByteStart - 1] &= - static_cast<u8>((1 << (weightParams.GetPackedBitSize() % 8)) - 1); - std::memset(texelWeightData.data() + clearByteStart, 0, - std::min(16U - clearByteStart, 16U)); - } - - IntegerEncodedVector texelWeightValues; - - InputBitStream weightStream(texelWeightData); - - DecodeIntegerSequence(texelWeightValues, weightStream, weightParams.m_MaxWeight, - weightParams.GetNumWeightValues()); - - // Blocks can be at most 12x12, so we can have as many as 144 weights - u32 weights[2][144]; - UnquantizeTexelWeights(weights, texelWeightValues, weightParams, blockWidth, blockHeight); - - // Now that we have endpos32s and weights, we can s32erpolate and generate - // the proper decoding... - for (u32 j = 0; j < blockHeight; j++) - for (u32 i = 0; i < blockWidth; i++) { - u32 partition = Select2DPartition(partitionIndex, i, j, nPartitions, - (blockHeight * blockWidth) < 32); - assert(partition < nPartitions); - - Pixel p; - for (u32 c = 0; c < 4; c++) { - u32 C0 = endpos32s[partition][0].Component(c); - C0 = ReplicateByteTo16(C0); - u32 C1 = endpos32s[partition][1].Component(c); - C1 = ReplicateByteTo16(C1); - - u32 plane = 0; - if (weightParams.m_bDualPlane && (((planeIdx + 1) & 3) == c)) { - plane = 1; - } - - u32 weight = weights[plane][j * blockWidth + i]; - u32 C = (C0 * (64 - weight) + C1 * weight + 32) / 64; - if (C == 65535) { - p.Component(c) = 255; - } else { - double Cf = static_cast<double>(C); - p.Component(c) = static_cast<u16>(255.0 * (Cf / 65536.0) + 0.5); - } - } - - outBuf[j * blockWidth + i] = p.Pack(); - } -} - -} // namespace ASTCC - -namespace Tegra::Texture::ASTC { - -void Decompress(std::span<const uint8_t> data, uint32_t width, uint32_t height, uint32_t depth, - uint32_t block_width, uint32_t block_height, std::span<uint8_t> output) { - u32 block_index = 0; - std::size_t depth_offset = 0; - for (u32 z = 0; z < depth; z++) { - for (u32 y = 0; y < height; y += block_height) { - for (u32 x = 0; x < width; x += block_width) { - const std::span<const u8, 16> blockPtr{data.subspan(block_index * 16, 16)}; - - // Blocks can be at most 12x12 - std::array<u32, 12 * 12> uncompData; - ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData); - - u32 decompWidth = std::min(block_width, width - x); - u32 decompHeight = std::min(block_height, height - y); - - const std::span<u8> outRow = output.subspan(depth_offset + (y * width + x) * 4); - for (u32 jj = 0; jj < decompHeight; jj++) { - std::memcpy(outRow.data() + jj * width * 4, - uncompData.data() + jj * block_width, decompWidth * 4); - } - ++block_index; - } - } - depth_offset += height * width * 4; - } -} - -} // namespace Tegra::Texture::ASTC diff --git a/src/video_core/textures/astc.h b/src/video_core/textures/astc.h index 9105119bc..c1c73fda5 100644 --- a/src/video_core/textures/astc.h +++ b/src/video_core/textures/astc.h @@ -4,11 +4,129 @@ #pragma once -#include <cstdint> +#include <bit> +#include "common/common_types.h" namespace Tegra::Texture::ASTC { -void Decompress(std::span<const uint8_t> data, uint32_t width, uint32_t height, uint32_t depth, - uint32_t block_width, uint32_t block_height, std::span<uint8_t> output); +enum class IntegerEncoding { JustBits, Quint, Trit }; + +struct IntegerEncodedValue { + constexpr IntegerEncodedValue() = default; + + constexpr IntegerEncodedValue(IntegerEncoding encoding_, u32 num_bits_) + : encoding{encoding_}, num_bits{num_bits_} {} + + constexpr bool MatchesEncoding(const IntegerEncodedValue& other) const { + return encoding == other.encoding && num_bits == other.num_bits; + } + + // Returns the number of bits required to encode num_vals values. + u32 GetBitLength(u32 num_vals) const { + u32 total_bits = num_bits * num_vals; + if (encoding == IntegerEncoding::Trit) { + total_bits += (num_vals * 8 + 4) / 5; + } else if (encoding == IntegerEncoding::Quint) { + total_bits += (num_vals * 7 + 2) / 3; + } + return total_bits; + } + + IntegerEncoding encoding{}; + u32 num_bits = 0; + u32 bit_value = 0; + union { + u32 quint_value = 0; + u32 trit_value; + }; +}; + +// Returns a new instance of this struct that corresponds to the +// can take no more than mav_value values +constexpr IntegerEncodedValue CreateEncoding(u32 mav_value) { + while (mav_value > 0) { + u32 check = mav_value + 1; + + // Is mav_value a power of two? + if (!(check & (check - 1))) { + return IntegerEncodedValue(IntegerEncoding::JustBits, std::popcount(mav_value)); + } + + // Is mav_value of the type 3*2^n - 1? + if ((check % 3 == 0) && !((check / 3) & ((check / 3) - 1))) { + return IntegerEncodedValue(IntegerEncoding::Trit, std::popcount(check / 3 - 1)); + } + + // Is mav_value of the type 5*2^n - 1? + if ((check % 5 == 0) && !((check / 5) & ((check / 5) - 1))) { + return IntegerEncodedValue(IntegerEncoding::Quint, std::popcount(check / 5 - 1)); + } + + // Apparently it can't be represented with a bounded integer sequence... + // just iterate. + mav_value--; + } + return IntegerEncodedValue(IntegerEncoding::JustBits, 0); +} + +constexpr std::array<IntegerEncodedValue, 256> MakeEncodedValues() { + std::array<IntegerEncodedValue, 256> encodings{}; + for (std::size_t i = 0; i < encodings.size(); ++i) { + encodings[i] = CreateEncoding(static_cast<u32>(i)); + } + return encodings; +} + +constexpr std::array<IntegerEncodedValue, 256> EncodingsValues = MakeEncodedValues(); + +// Replicates low num_bits such that [(to_bit - 1):(to_bit - 1 - from_bit)] +// is the same as [(num_bits - 1):0] and repeats all the way down. +template <typename IntType> +constexpr IntType Replicate(IntType val, u32 num_bits, u32 to_bit) { + if (num_bits == 0 || to_bit == 0) { + return 0; + } + const IntType v = val & static_cast<IntType>((1 << num_bits) - 1); + IntType res = v; + u32 reslen = num_bits; + while (reslen < to_bit) { + u32 comp = 0; + if (num_bits > to_bit - reslen) { + u32 newshift = to_bit - reslen; + comp = num_bits - newshift; + num_bits = newshift; + } + res = static_cast<IntType>(res << num_bits); + res = static_cast<IntType>(res | (v >> comp)); + reslen += num_bits; + } + return res; +} + +constexpr std::size_t NumReplicateEntries(u32 num_bits) { + return std::size_t(1) << num_bits; +} + +template <typename IntType, u32 num_bits, u32 to_bit> +constexpr auto MakeReplicateTable() { + std::array<IntType, NumReplicateEntries(num_bits)> table{}; + for (IntType value = 0; value < static_cast<IntType>(std::size(table)); ++value) { + table[value] = Replicate(value, num_bits, to_bit); + } + return table; +} + +constexpr auto REPLICATE_BYTE_TO_16_TABLE = MakeReplicateTable<u32, 8, 16>(); +constexpr auto REPLICATE_6_BIT_TO_8_TABLE = MakeReplicateTable<u32, 6, 8>(); +constexpr auto REPLICATE_7_BIT_TO_8_TABLE = MakeReplicateTable<u32, 7, 8>(); +constexpr auto REPLICATE_8_BIT_TO_8_TABLE = MakeReplicateTable<u32, 8, 8>(); + +struct AstcBufferData { + decltype(EncodingsValues) encoding_values = EncodingsValues; + decltype(REPLICATE_6_BIT_TO_8_TABLE) replicate_6_to_8 = REPLICATE_6_BIT_TO_8_TABLE; + decltype(REPLICATE_7_BIT_TO_8_TABLE) replicate_7_to_8 = REPLICATE_7_BIT_TO_8_TABLE; + decltype(REPLICATE_8_BIT_TO_8_TABLE) replicate_8_to_8 = REPLICATE_8_BIT_TO_8_TABLE; + decltype(REPLICATE_BYTE_TO_16_TABLE) replicate_byte_to_16 = REPLICATE_BYTE_TO_16_TABLE; +} constexpr ASTC_BUFFER_DATA; } // namespace Tegra::Texture::ASTC diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp index 62685a183..3a463d5db 100644 --- a/src/video_core/textures/decoders.cpp +++ b/src/video_core/textures/decoders.cpp @@ -17,26 +17,7 @@ #include "video_core/textures/texture.h" namespace Tegra::Texture { - namespace { -/** - * This table represents the internal swizzle of a gob, in format 16 bytes x 2 sector packing. - * Calculates the offset of an (x, y) position within a swizzled texture. - * Taken from the Tegra X1 Technical Reference Manual. pages 1187-1188 - */ -constexpr SwizzleTable MakeSwizzleTableConst() { - SwizzleTable table{}; - for (u32 y = 0; y < table.size(); ++y) { - for (u32 x = 0; x < table[0].size(); ++x) { - table[y][x] = ((x % 64) / 32) * 256 + ((y % 8) / 2) * 64 + ((x % 32) / 16) * 32 + - (y % 2) * 16 + (x % 16); - } - } - return table; -} - -constexpr SwizzleTable SWIZZLE_TABLE = MakeSwizzleTableConst(); - template <bool TO_LINEAR> void Swizzle(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, u32 width, u32 height, u32 depth, u32 block_height, u32 block_depth, u32 stride_alignment) { @@ -91,10 +72,6 @@ void Swizzle(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixe } } // Anonymous namespace -SwizzleTable MakeSwizzleTable() { - return SWIZZLE_TABLE; -} - void UnswizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, u32 width, u32 height, u32 depth, u32 block_height, u32 block_depth, u32 stride_alignment) { diff --git a/src/video_core/textures/decoders.h b/src/video_core/textures/decoders.h index d7cdc81e8..4c14cefbf 100644 --- a/src/video_core/textures/decoders.h +++ b/src/video_core/textures/decoders.h @@ -23,8 +23,22 @@ constexpr u32 GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_ using SwizzleTable = std::array<std::array<u32, GOB_SIZE_X>, GOB_SIZE_Y>; -/// Returns a z-order swizzle table -SwizzleTable MakeSwizzleTable(); +/** + * This table represents the internal swizzle of a gob, in format 16 bytes x 2 sector packing. + * Calculates the offset of an (x, y) position within a swizzled texture. + * Taken from the Tegra X1 Technical Reference Manual. pages 1187-1188 + */ +constexpr SwizzleTable MakeSwizzleTable() { + SwizzleTable table{}; + for (u32 y = 0; y < table.size(); ++y) { + for (u32 x = 0; x < table[0].size(); ++x) { + table[y][x] = ((x % 64) / 32) * 256 + ((y % 8) / 2) * 64 + ((x % 32) / 16) * 32 + + (y % 2) * 16 + (x % 16); + } + } + return table; +} +constexpr SwizzleTable SWIZZLE_TABLE = MakeSwizzleTable(); /// Unswizzles a block linear texture into linear memory. void UnswizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, diff --git a/src/video_core/vulkan_common/vulkan_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp index 34d396434..697cb16b9 100644 --- a/src/video_core/vulkan_common/vulkan_device.cpp +++ b/src/video_core/vulkan_common/vulkan_device.cpp @@ -51,7 +51,7 @@ constexpr std::array REQUIRED_EXTENSIONS{ #ifdef _WIN32 VK_KHR_EXTERNAL_MEMORY_WIN32_EXTENSION_NAME, #endif -#ifdef __linux__ +#ifdef __unix__ VK_KHR_EXTERNAL_MEMORY_FD_EXTENSION_NAME, #endif }; diff --git a/src/video_core/vulkan_common/vulkan_memory_allocator.cpp b/src/video_core/vulkan_common/vulkan_memory_allocator.cpp index 2a8b7a907..fa37aa79a 100644 --- a/src/video_core/vulkan_common/vulkan_memory_allocator.cpp +++ b/src/video_core/vulkan_common/vulkan_memory_allocator.cpp @@ -62,7 +62,7 @@ public: : memory{std::move(memory_)}, allocation_size{allocation_size_}, property_flags{properties}, shifted_memory_type{1U << type} {} -#if defined(_WIN32) || defined(__linux__) +#if defined(_WIN32) || defined(__unix__) ~MemoryAllocation() { if (owning_opengl_handle != 0) { glDeleteMemoryObjectsEXT(1, &owning_opengl_handle); @@ -114,7 +114,7 @@ public: } return owning_opengl_handle; } -#elif __linux__ +#elif __unix__ [[nodiscard]] u32 ExportOpenGLHandle() { if (!owning_opengl_handle) { glCreateMemoryObjectsEXT(1, &owning_opengl_handle); @@ -165,7 +165,7 @@ private: const u32 shifted_memory_type; ///< Shifted Vulkan memory type. std::vector<Range> commits; ///< All commit ranges done from this allocation. std::span<u8> memory_mapped_span; ///< Memory mapped span. Empty if not queried before. -#if defined(_WIN32) || defined(__linux__) +#if defined(_WIN32) || defined(__unix__) u32 owning_opengl_handle{}; ///< Owning OpenGL memory object handle. #endif }; @@ -249,7 +249,7 @@ void MemoryAllocator::AllocMemory(VkMemoryPropertyFlags flags, u32 type_mask, u6 .pNext = nullptr, #ifdef _WIN32 .handleTypes = VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32_BIT, -#elif __linux__ +#elif __unix__ .handleTypes = VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD_BIT, #else .handleTypes = 0, diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp index 1bac57bb2..1d6155999 100644 --- a/src/yuzu/configuration/config.cpp +++ b/src/yuzu/configuration/config.cpp @@ -641,6 +641,7 @@ void Config::ReadDebuggingValues() { ReadSetting(QStringLiteral("disable_macro_jit"), false).toBool(); Settings::values.extended_logging = ReadSetting(QStringLiteral("extended_logging"), false).toBool(); + Settings::values.use_auto_stub = ReadSetting(QStringLiteral("use_auto_stub"), false).toBool(); qt_config->endGroup(); } diff --git a/src/yuzu/configuration/configure_debug.cpp b/src/yuzu/configuration/configure_debug.cpp index 121873f95..2a5b3f5e7 100644 --- a/src/yuzu/configuration/configure_debug.cpp +++ b/src/yuzu/configuration/configure_debug.cpp @@ -34,6 +34,7 @@ void ConfigureDebug::SetConfiguration() { ui->homebrew_args_edit->setText(QString::fromStdString(Settings::values.program_args)); ui->reporting_services->setChecked(Settings::values.reporting_services); ui->quest_flag->setChecked(Settings::values.quest_flag); + ui->use_auto_stub->setChecked(Settings::values.use_auto_stub); ui->enable_graphics_debugging->setEnabled(!Core::System::GetInstance().IsPoweredOn()); ui->enable_graphics_debugging->setChecked(Settings::values.renderer_debug); ui->disable_macro_jit->setEnabled(!Core::System::GetInstance().IsPoweredOn()); @@ -47,6 +48,7 @@ void ConfigureDebug::ApplyConfiguration() { Settings::values.program_args = ui->homebrew_args_edit->text().toStdString(); Settings::values.reporting_services = ui->reporting_services->isChecked(); Settings::values.quest_flag = ui->quest_flag->isChecked(); + Settings::values.use_auto_stub = ui->use_auto_stub->isChecked(); Settings::values.renderer_debug = ui->enable_graphics_debugging->isChecked(); Settings::values.disable_macro_jit = ui->disable_macro_jit->isChecked(); Settings::values.extended_logging = ui->extended_logging->isChecked(); diff --git a/src/yuzu/configuration/configure_debug.ui b/src/yuzu/configuration/configure_debug.ui index 9186aa732..ae48b728c 100644 --- a/src/yuzu/configuration/configure_debug.ui +++ b/src/yuzu/configuration/configure_debug.ui @@ -185,6 +185,28 @@ </property> </widget> </item> + <item> + <widget class="QCheckBox" name="use_auto_stub"> + <property name="text"> + <string>Enable Auto-Stub</string> + </property> + </widget> + </item> + <item> + <widget class="QLabel" name="label_3"> + <property name="font"> + <font> + <italic>true</italic> + </font> + </property> + <property name="text"> + <string>This will be reset automatically when yuzu closes.</string> + </property> + <property name="indent"> + <number>20</number> + </property> + </widget> + </item> </layout> </widget> </item> diff --git a/src/yuzu/configuration/configure_graphics.cpp b/src/yuzu/configuration/configure_graphics.cpp index 9ff32aec4..49acc48b2 100644 --- a/src/yuzu/configuration/configure_graphics.cpp +++ b/src/yuzu/configuration/configure_graphics.cpp @@ -227,7 +227,7 @@ void ConfigureGraphics::RetrieveVulkanDevices() try { vulkan_devices.clear(); vulkan_devices.reserve(physical_devices.size()); for (const VkPhysicalDevice device : physical_devices) { - const char* const name = vk::PhysicalDevice(device, dld).GetProperties().deviceName; + const std::string name = vk::PhysicalDevice(device, dld).GetProperties().deviceName; vulkan_devices.push_back(QString::fromStdString(name)); } diff --git a/src/yuzu/configuration/configure_motion_touch.cpp b/src/yuzu/configuration/configure_motion_touch.cpp index 52fdf7265..083d1ea43 100644 --- a/src/yuzu/configuration/configure_motion_touch.cpp +++ b/src/yuzu/configuration/configure_motion_touch.cpp @@ -23,8 +23,7 @@ #include "yuzu/configuration/configure_touch_from_button.h" CalibrationConfigurationDialog::CalibrationConfigurationDialog(QWidget* parent, - const std::string& host, u16 port, - u8 pad_index) + const std::string& host, u16 port) : QDialog(parent) { layout = new QVBoxLayout; status_label = new QLabel(tr("Communicating with the server...")); @@ -41,7 +40,7 @@ CalibrationConfigurationDialog::CalibrationConfigurationDialog(QWidget* parent, using namespace InputCommon::CemuhookUDP; job = std::make_unique<CalibrationConfigurationJob>( - host, port, pad_index, + host, port, [this](CalibrationConfigurationJob::Status status) { QString text; switch (status) { @@ -217,7 +216,7 @@ void ConfigureMotionTouch::OnCemuhookUDPTest() { ui->udp_test->setText(tr("Testing")); udp_test_in_progress = true; InputCommon::CemuhookUDP::TestCommunication( - ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toInt()), 0, + ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toInt()), [this] { LOG_INFO(Frontend, "UDP input test success"); QMetaObject::invokeMethod(this, "ShowUDPTestResult", Q_ARG(bool, true)); @@ -232,7 +231,7 @@ void ConfigureMotionTouch::OnConfigureTouchCalibration() { ui->touch_calibration_config->setEnabled(false); ui->touch_calibration_config->setText(tr("Configuring")); CalibrationConfigurationDialog dialog(this, ui->udp_server->text().toStdString(), - static_cast<u16>(ui->udp_port->text().toUInt()), 0); + static_cast<u16>(ui->udp_port->text().toUInt())); dialog.exec(); if (dialog.completed) { min_x = dialog.min_x; diff --git a/src/yuzu/configuration/configure_motion_touch.h b/src/yuzu/configuration/configure_motion_touch.h index d76bc8154..8b707d2ff 100644 --- a/src/yuzu/configuration/configure_motion_touch.h +++ b/src/yuzu/configuration/configure_motion_touch.h @@ -29,8 +29,7 @@ class ConfigureMotionTouch; class CalibrationConfigurationDialog : public QDialog { Q_OBJECT public: - explicit CalibrationConfigurationDialog(QWidget* parent, const std::string& host, u16 port, - u8 pad_index); + explicit CalibrationConfigurationDialog(QWidget* parent, const std::string& host, u16 port); ~CalibrationConfigurationDialog() override; private: diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index 24bfa4d34..06445b993 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -320,6 +320,34 @@ GMainWindow::GMainWindow() continue; } + // Launch game with a specific user + if (args[i] == QStringLiteral("-u")) { + if (i >= args.size() - 1) { + continue; + } + + if (args[i + 1].startsWith(QChar::fromLatin1('-'))) { + continue; + } + + bool argument_ok; + const std::size_t selected_user = args[++i].toUInt(&argument_ok); + + if (!argument_ok) { + LOG_ERROR(Frontend, "Invalid user argument"); + continue; + } + + const Service::Account::ProfileManager manager; + if (!manager.UserExistsIndex(selected_user)) { + LOG_ERROR(Frontend, "Selected user doesn't exist"); + continue; + } + + Settings::values.current_user = selected_user; + continue; + } + // Launch game at path if (args[i] == QStringLiteral("-g")) { if (i >= args.size() - 1) { |