Merge pull request #3689 from MarcZX100/codex/libkernel-spin-reltimedwait-3554

feat(libkernel): implement pthread spinlocks and relative condition wait

core/libs/prx/libkernel/CMakeLists.txt

@@ -36,6 +36,7 @@ add_library(libkernel SHARED EXCLUDE_FROM_ALL
         ${CMAKE_CURRENT_SOURCE_DIR}/Pthread/Posix/Once.cpp
         ${CMAKE_CURRENT_SOURCE_DIR}/Pthread/Posix/Tsd.cpp
         ${CMAKE_CURRENT_SOURCE_DIR}/Pthread/Posix/Sem.cpp
+        ${CMAKE_CURRENT_SOURCE_DIR}/Pthread/Posix/Spinlock.cpp
         ${CMAKE_CURRENT_SOURCE_DIR}/Socket/src/Socket.cpp
         ${CMAKE_CURRENT_SOURCE_DIR}/Socket/src/Udp.cpp
         ${CMAKE_CURRENT_SOURCE_DIR}/Socket/src/AddressConversion.cpp
@@ -314,6 +315,16 @@ if(BUILD_TESTING)
         set_tests_properties(guest_thread_self PROPERTIES ENVIRONMENT_MODIFICATION
             "PATH=path_list_prepend:$<TARGET_FILE_DIR:libkernel>")
     endif()
+    add_test_executable(guest_pthread_spin_tests ${LIBS_INCLUDE_DIR}/tests/GuestPthreadSpin.cpp)
+    target_include_directories(guest_pthread_spin_tests PRIVATE ${LIBS_INCLUDE_DIR})
+    target_link_libraries(guest_pthread_spin_tests PRIVATE libkernel libc)
+    configure_windows_unwind(guest_pthread_spin_tests)
+    add_test(NAME guest_pthread_spin COMMAND guest_pthread_spin_tests)
+    set_tests_properties(guest_pthread_spin PROPERTIES TIMEOUT 20)
+    if(WIN32)
+        set_tests_properties(guest_pthread_spin PROPERTIES ENVIRONMENT_MODIFICATION
+            "PATH=path_list_prepend:$<TARGET_FILE_DIR:libkernel>")
+    endif()
     add_test_executable(guest_pthread_getname_tests ${LIBS_INCLUDE_DIR}/tests/GuestPthreadGetname.cpp)
     target_include_directories(guest_pthread_getname_tests PRIVATE ${LIBS_INCLUDE_DIR})
     target_link_libraries(guest_pthread_getname_tests PRIVATE libkernel libc)

core/libs/prx/libkernel/Pthread/Posix/Cond.cpp

@@ -41,6 +41,10 @@ int APS5_VABI pthread_cond_timedwait_nid_postfix(PthreadCond* cond, PthreadMutex
     return toPosix(CondOperations::AbsoluteTimedwait(cond, mutex, abstime));
 }
 
+int APS5_VABI pthread_cond_reltimedwait_np_nid_postfix(PthreadCond* cond, PthreadMutex* mutex, KernelUseconds usec) {
+    return toPosix(scePthreadCondTimedwait(cond, mutex, usec));
+}
+
 int APS5_VABI pthread_cond_wait_nid_postfix(PthreadCond* cond, PthreadMutex* mutex) {
     return toPosix(scePthreadCondWait(cond, mutex));
 }

core/libs/prx/libkernel/Pthread/Posix/Spinlock.cpp

@@ -0,0 +1,76 @@
+#include "SceTypes.hpp"
+#include "Common.hpp"
+#include <atomic>
+#include <new>
+#include <stdexcept>
+#include <thread>
+
+struct PthreadSpinlockPrivate {
+    std::atomic<std::thread::id> owner{};
+};
+
+namespace {
+
+constexpr int GUEST_EPERM = 1;
+constexpr int GUEST_ENOMEM = 12;
+constexpr int GUEST_EBUSY = 16;
+constexpr int PROCESS_PRIVATE = 0;
+constexpr int PROCESS_SHARED = 1;
+constexpr int SPIN_COUNT = 100000;
+
+bool tryAcquire(PthreadSpinlockPrivate* spinlock) {
+    std::thread::id unowned{};
+    return spinlock->owner.compare_exchange_strong(unowned, std::this_thread::get_id(), std::memory_order_acquire, std::memory_order_relaxed);
+}
+
+}
+
+extern "C" {
+
+int APS5_VABI pthread_spin_init_nid_postfix(PthreadSpinlockPrivate** lock, int pshared) {
+    if (!lock) return PosixThread::GUEST_EINVAL;
+    if (pshared != PROCESS_PRIVATE && pshared != PROCESS_SHARED) return PosixThread::GUEST_EINVAL;
+    auto* spinlock = new (std::nothrow) PthreadSpinlockPrivate;
+    if (!spinlock) return GUEST_ENOMEM;
+    *lock = spinlock;
+    return 0;
+}
+
+int APS5_VABI pthread_spin_destroy_nid_postfix(PthreadSpinlockPrivate** lock) {
+    if (!lock || !*lock) return PosixThread::GUEST_EINVAL;
+    delete *lock;
+    *lock = nullptr;
+    return 0;
+}
+
+int APS5_VABI pthread_spin_trylock_nid_postfix(PthreadSpinlockPrivate** lock) {
+    if (!lock || !*lock) return PosixThread::GUEST_EINVAL;
+    return tryAcquire(*lock) ? 0 : GUEST_EBUSY;
+}
+
+int APS5_VABI pthread_spin_lock_nid_postfix(PthreadSpinlockPrivate** lock) {
+    if (!lock || !*lock) return PosixThread::GUEST_EINVAL;
+    auto* spinlock = *lock;
+    if (spinlock->owner.load(std::memory_order_relaxed) == std::this_thread::get_id())
+        throw std::runtime_error("pthread_spin_lock: the calling thread already holds the lock and would spin forever");
+    int count = SPIN_COUNT;
+    while (!tryAcquire(spinlock)) {
+        while (spinlock->owner.load(std::memory_order_relaxed) != std::thread::id{}) {
+            if (--count <= 0) {
+                count = SPIN_COUNT;
+                std::this_thread::yield();
+            }
+        }
+    }
+    return 0;
+}
+
+int APS5_VABI pthread_spin_unlock_nid_postfix(PthreadSpinlockPrivate** lock) {
+    if (!lock || !*lock) return PosixThread::GUEST_EINVAL;
+    auto owner = std::this_thread::get_id();
+    if (!(*lock)->owner.compare_exchange_strong(owner, std::thread::id{}, std::memory_order_release, std::memory_order_relaxed))
+        return GUEST_EPERM;
+    return 0;
+}
+
+}

core/libs/tests/GuestCondTimedwait.cpp

@@ -1,5 +1,6 @@
 #include "SceTypes.hpp"
 #include <atomic>
+#include <chrono>
 #include <cstdint>
 #include <cstdlib>
 
@@ -10,9 +11,12 @@ int APS5_VABI scePthreadMutexInit(PthreadMutex* mutex, const PthreadMutexattr* a
 int APS5_VABI scePthreadMutexDestroy(PthreadMutex* mutex);
 int APS5_VABI scePthreadMutexLock(PthreadMutex* mutex);
 int APS5_VABI scePthreadMutexUnlock(PthreadMutex* mutex);
+int APS5_VABI scePthreadMutexTrylock(PthreadMutex* mutex);
 int APS5_VABI scePthreadCondInit(PthreadCond* cond, const PthreadCondattr* attr, const char* name);
 int APS5_VABI scePthreadCondDestroy(PthreadCond* cond);
 int APS5_VABI scePthreadCondTimedwait(PthreadCond* cond, PthreadMutex* mutex, unsigned int usec);
+int APS5_VABI scePthreadCondSignal(PthreadCond* cond);
+int APS5_VABI pthread_cond_reltimedwait_np_nid_postfix(PthreadCond* cond, PthreadMutex* mutex, unsigned int usec);
 int APS5_VABI __cxa_atexit_nid_postfix(void (APS5_VABI *)(void*), void*, void*);
 void APS5_VABI __pthread_cxa_finalize_nid_postfix(void*);
 unsigned int APS5_VABI sceKernelSleep(unsigned int seconds);
@@ -23,6 +27,9 @@ static void APS5_VABI Cleanup(void*) { cleanupAt = 0; }
 
 static constexpr int SCE_OK = 0;
 static constexpr int SCE_KERNEL_ERROR_ETIMEDOUT = static_cast<int>(0x8002003C);
+static constexpr int SCE_KERNEL_ERROR_EBUSY = static_cast<int>(0x80020010);
+static constexpr int POSIX_EPERM = 1;
+static constexpr int POSIX_ETIMEDOUT = 60;
 
 static void Require(bool value) { if (!value) std::abort(); }
 
@@ -39,7 +46,67 @@ static void* APS5_VABI Contender(void* arg) {
     return nullptr;
 }
 
+struct Signalled {
+    PthreadMutex mutex = nullptr;
+    PthreadCond cond = nullptr;
+    bool ready = false;
+    int trylock = SCE_OK;
+};
+
+static void* APS5_VABI Trylocker(void* arg) {
+    auto& signalled = *static_cast<Signalled*>(arg);
+    signalled.trylock = scePthreadMutexTrylock(&signalled.mutex);
+    return nullptr;
+}
+
+static void* APS5_VABI Signaller(void* arg) {
+    auto& signalled = *static_cast<Signalled*>(arg);
+    Require(scePthreadMutexLock(&signalled.mutex) == SCE_OK);
+    signalled.ready = true;
+    Require(scePthreadCondSignal(&signalled.cond) == SCE_OK);
+    Require(scePthreadMutexUnlock(&signalled.mutex) == SCE_OK);
+    return nullptr;
+}
+
+static void RequireHeldByCaller(Signalled& signalled) {
+    Pthread thread = nullptr;
+    Require(scePthreadCreate(&thread, nullptr, Trylocker, &signalled, nullptr) == SCE_OK);
+    Require(scePthreadJoin(thread, nullptr) == SCE_OK);
+    Require(signalled.trylock == SCE_KERNEL_ERROR_EBUSY);
+}
+
+static void RelativeWait() {
+    Signalled signalled;
+    Require(scePthreadMutexInit(&signalled.mutex, nullptr, nullptr) == SCE_OK);
+    Require(scePthreadCondInit(&signalled.cond, nullptr, nullptr) == SCE_OK);
+
+    Require(pthread_cond_reltimedwait_np_nid_postfix(&signalled.cond, &signalled.mutex, 1000) == POSIX_EPERM);
+
+    Require(scePthreadMutexLock(&signalled.mutex) == SCE_OK);
+    const auto start = std::chrono::steady_clock::now();
+    Require(pthread_cond_reltimedwait_np_nid_postfix(&signalled.cond, &signalled.mutex, 50000) == POSIX_ETIMEDOUT);
+    Require(std::chrono::steady_clock::now() - start >= std::chrono::milliseconds(40));
+    RequireHeldByCaller(signalled);
+    Require(pthread_cond_reltimedwait_np_nid_postfix(&signalled.cond, &signalled.mutex, 0) == POSIX_ETIMEDOUT);
+    RequireHeldByCaller(signalled);
+
+    Pthread thread = nullptr;
+    Require(scePthreadCreate(&thread, nullptr, Signaller, &signalled, nullptr) == SCE_OK);
+    const auto waitStart = std::chrono::steady_clock::now();
+    while (!signalled.ready)
+        Require(pthread_cond_reltimedwait_np_nid_postfix(&signalled.cond, &signalled.mutex, 10000000) == 0);
+    Require(std::chrono::steady_clock::now() - waitStart < std::chrono::seconds(5));
+    RequireHeldByCaller(signalled);
+    Require(scePthreadMutexUnlock(&signalled.mutex) == SCE_OK);
+    Require(scePthreadJoin(thread, nullptr) == SCE_OK);
+
+    Require(scePthreadCondDestroy(&signalled.cond) == SCE_OK);
+    Require(scePthreadMutexDestroy(&signalled.mutex) == SCE_OK);
+}
+
 int main() {
+    RelativeWait();
+
     Context context;
     PthreadCond cond = nullptr;
     Require(scePthreadMutexInit(&context.mutex, nullptr, nullptr) == SCE_OK);

core/libs/tests/GuestPthreadSpin.cpp

@@ -0,0 +1,122 @@
+#include "SceTypes.hpp"
+#include <cstdlib>
+#include <stdexcept>
+
+struct PthreadSpinlockPrivate;
+using PthreadSpinlock = PthreadSpinlockPrivate*;
+
+extern "C" {
+int APS5_VABI scePthreadCreate(Pthread* thread, const PthreadAttr* attr, PthreadEntry entry, void* arg, const char* name);
+int APS5_VABI scePthreadJoin(Pthread thread, void** retval);
+int APS5_VABI pthread_spin_init_nid_postfix(PthreadSpinlock* lock, int pshared);
+int APS5_VABI pthread_spin_destroy_nid_postfix(PthreadSpinlock* lock);
+int APS5_VABI pthread_spin_lock_nid_postfix(PthreadSpinlock* lock);
+int APS5_VABI pthread_spin_trylock_nid_postfix(PthreadSpinlock* lock);
+int APS5_VABI pthread_spin_unlock_nid_postfix(PthreadSpinlock* lock);
+}
+
+static constexpr int GUEST_EPERM = 1;
+static constexpr int GUEST_EBUSY = 16;
+static constexpr int GUEST_EINVAL = 22;
+static constexpr int PTHREAD_PROCESS_PRIVATE = 0;
+static constexpr int PTHREAD_PROCESS_SHARED = 1;
+static constexpr int Increments = 200000;
+
+static void Require(bool value) { if (!value) std::abort(); }
+
+struct Shared {
+    PthreadSpinlock lock = nullptr;
+    long counter = 0;
+    int trylock = -1;
+    int unlock = -1;
+};
+
+static void* APS5_VABI Contender(void* arg) {
+    auto& shared = *static_cast<Shared*>(arg);
+    shared.trylock = pthread_spin_trylock_nid_postfix(&shared.lock);
+    shared.unlock = pthread_spin_unlock_nid_postfix(&shared.lock);
+    return nullptr;
+}
+
+static void* APS5_VABI Incrementer(void* arg) {
+    auto& shared = *static_cast<Shared*>(arg);
+    for (int i = 0; i < Increments; ++i) {
+        Require(pthread_spin_lock_nid_postfix(&shared.lock) == 0);
+        ++shared.counter;
+        Require(pthread_spin_unlock_nid_postfix(&shared.lock) == 0);
+    }
+    return nullptr;
+}
+
+static void ArgumentErrors() {
+    Require(pthread_spin_init_nid_postfix(nullptr, PTHREAD_PROCESS_PRIVATE) == GUEST_EINVAL);
+    PthreadSpinlock lock = nullptr;
+    Require(pthread_spin_init_nid_postfix(&lock, 2) == GUEST_EINVAL);
+    Require(pthread_spin_init_nid_postfix(&lock, -1) == GUEST_EINVAL);
+    Require(lock == nullptr);
+    Require(pthread_spin_lock_nid_postfix(nullptr) == GUEST_EINVAL);
+    Require(pthread_spin_trylock_nid_postfix(nullptr) == GUEST_EINVAL);
+    Require(pthread_spin_unlock_nid_postfix(nullptr) == GUEST_EINVAL);
+    Require(pthread_spin_destroy_nid_postfix(nullptr) == GUEST_EINVAL);
+    Require(pthread_spin_lock_nid_postfix(&lock) == GUEST_EINVAL);
+    Require(pthread_spin_trylock_nid_postfix(&lock) == GUEST_EINVAL);
+    Require(pthread_spin_unlock_nid_postfix(&lock) == GUEST_EINVAL);
+    Require(pthread_spin_destroy_nid_postfix(&lock) == GUEST_EINVAL);
+}
+
+static void SingleThread(int pshared) {
+    PthreadSpinlock lock = nullptr;
+    Require(pthread_spin_init_nid_postfix(&lock, pshared) == 0);
+    Require(lock != nullptr);
+    Require(pthread_spin_unlock_nid_postfix(&lock) == GUEST_EPERM);
+    Require(pthread_spin_lock_nid_postfix(&lock) == 0);
+    Require(pthread_spin_trylock_nid_postfix(&lock) == GUEST_EBUSY);
+    bool threw = false;
+    try {
+        pthread_spin_lock_nid_postfix(&lock);
+    } catch (const std::runtime_error&) {
+        threw = true;
+    }
+    Require(threw);
+    Require(pthread_spin_unlock_nid_postfix(&lock) == 0);
+    Require(pthread_spin_unlock_nid_postfix(&lock) == GUEST_EPERM);
+    Require(pthread_spin_trylock_nid_postfix(&lock) == 0);
+    Require(pthread_spin_unlock_nid_postfix(&lock) == 0);
+    Require(pthread_spin_destroy_nid_postfix(&lock) == 0);
+    Require(lock == nullptr);
+    Require(pthread_spin_destroy_nid_postfix(&lock) == GUEST_EINVAL);
+}
+
+static void HeldByAnotherThread() {
+    Shared shared;
+    Require(pthread_spin_init_nid_postfix(&shared.lock, PTHREAD_PROCESS_PRIVATE) == 0);
+    Require(pthread_spin_lock_nid_postfix(&shared.lock) == 0);
+    Pthread thread = nullptr;
+    Require(scePthreadCreate(&thread, nullptr, Contender, &shared, nullptr) == 0);
+    Require(scePthreadJoin(thread, nullptr) == 0);
+    Require(shared.trylock == GUEST_EBUSY);
+    Require(shared.unlock == GUEST_EPERM);
+    Require(pthread_spin_unlock_nid_postfix(&shared.lock) == 0);
+    Require(pthread_spin_destroy_nid_postfix(&shared.lock) == 0);
+}
+
+static void MutualExclusion() {
+    Shared shared;
+    Require(pthread_spin_init_nid_postfix(&shared.lock, PTHREAD_PROCESS_PRIVATE) == 0);
+    Pthread first = nullptr;
+    Pthread second = nullptr;
+    Require(scePthreadCreate(&first, nullptr, Incrementer, &shared, nullptr) == 0);
+    Require(scePthreadCreate(&second, nullptr, Incrementer, &shared, nullptr) == 0);
+    Require(scePthreadJoin(first, nullptr) == 0);
+    Require(scePthreadJoin(second, nullptr) == 0);
+    Require(shared.counter == 2L * Increments);
+    Require(pthread_spin_destroy_nid_postfix(&shared.lock) == 0);
+}
+
+int main() {
+    ArgumentErrors();
+    SingleThread(PTHREAD_PROCESS_PRIVATE);
+    SingleThread(PTHREAD_PROCESS_SHARED);
+    HeldByAnotherThread();
+    MutualExclusion();
+}

docs/dev/TechnicalDebt.md

@@ -648,3 +648,5 @@ Throughout the project, every function at every stage either **does exactly what
 - [_WStoul](../../core/libs/prx/libc/src/Strings.cpp), [_Iswctype](../../core/libs/prx/libc/src/RuneLocale.cpp) and [swscanf_s](../../core/libs/prx/libc/src/ScanningWide.cpp) (libc) - `_WStoul` is `wcstoul` and throws for a base other than 0 or 2 to 36. `_Iswctype` takes the class numbers of Dinkumware's [xwcc.h](https://github.com/superna9999/ti-cgt6x/blob/cgt6x_7_3_4/include/xwcc.h) (1 `alnum` to 11 `xdigit`, 12 `blank`) and throws for any other value, including the FreeBSD masks `wctype` returns; the numbering is not confirmed on the console's `libc.prx`. `swscanf_s` follows C11 7.29.2.2 and K.3.9.1.5, but reads the size argument as `unsigned int` like `sscanf_s` where K.3.9.1.5 specifies `rsize_t`; the console's width is unverified. A runtime-constraint violation (null buffer, format or receiving pointer) only returns `EOF`, and a too small array gets a null first element as in `sscanf_s`. A literal mismatch is a matching failure and a field that is only a prefix, such as `0x` or `1e`, does not convert, as C11 requires; FreeBSD's `vfwscanf` reports an input failure for the first and converts the second. Integers out of range are truncated from `strtoll`/`strtoull`. It throws for `%p`, `%S`, `%C`, unknown length modifiers, scanset ranges, `%n` with suppression or width, and narrow `%s`, `%c` or `%[` input that the C-locale `wcrtomb` cannot convert
 - [sceSharePlayGetCurrentConnectionInfoA](../../core/libs/prx/libSceSharePlay/Export.cpp) (libSceSharePlay) - reports no Share Play session, as the library keeps none: status 0 (DORMANT), mode 0, empty online IDs, account IDs 0 and both user IDs -1 (`SCE_USER_SERVICE_USER_ID_INVALID`). A null pointer returns `SCE_SHARE_PLAY_ERROR_INVALID_ARGS` (0x810E0001). The 72-byte `SceSharePlayConnectionInfoA` layout and the error code follow [KytyPS5](https://github.com/KytyPS5/KytyPS5/blob/main/src/libs/libNet.cpp) and [fpPS4](https://github.com/red-prig/fpPS4/blob/trunk/src/ps4_libsceshareplay.pas), which agree on both; KytyPS5 sets the user IDs to -1 and fpPS4 leaves them 0, and which one the console returns is not verified. KytyPS5 also returns `SCE_SHARE_PLAY_ERROR_NOT_INITIALIZED` before `sceSharePlayInitialize`; this library keeps no initialization state and does not.
 - [sceAgcCreatePrimState](../../core/libs/prx/libSceAgc/Shader/src/PrimState.cpp) (libSceAgc) - the `ge_user_vgpr_en` special is passed to the user configuration registers at whatever offset the shader header stores, as the SPRX does according to [OpenAGC](https://github.com/OpenAGC/OpenAGC/blob/3c7e8072f0d7684e930454794fd5c02056946d73/src/game_compat.c#L1181-L1252), which copies all four specials verbatim (KytyPS5 stops on any other offset). PPSA99039, PPSA99505, PPSA99515 and PPSA99997 build their headers by hand with it at 0x25c (`GE_USER_VGPR1`) instead of 0x262 (`GE_USER_VGPR_EN`), value 0, and run on the console; it is not moved to 0x262, and the driver reads neither register. `ge_cntl`, `vgt_shader_stages_en` and a used `vgt_gs_out_prim_type` must still be at their own offsets, since the driver reads those registers there, so a header that stores one elsewhere throws naming the offset, although by OpenAGC the SPRX would copy it as well
+- [pthread_spin_init, pthread_spin_destroy, pthread_spin_lock, pthread_spin_trylock, pthread_spin_unlock](../../core/libs/prx/libkernel/Pthread/Posix/Spinlock.cpp) (libkernel) follow FreeBSD 11.4 libthr [`thr_pspinlock.c`](https://github.com/freebsd/freebsd-src/blob/8c639945a44588f6cfdddfb057bb5aad88735da0/lib/libthr/thread/thr_pspinlock.c) and the owner checks of its [`thr_umtx.h`](https://github.com/freebsd/freebsd-src/blob/8c639945a44588f6cfdddfb057bb5aad88735da0/lib/libthr/thread/thr_umtx.h#L80-L175), unverified on the console: `EINVAL` for a null or destroyed lock and a `pshared` other than 0 or 1, `EBUSY` from `trylock` while any thread holds it, `EPERM` from `unlock` by a thread that does not hold it. `PTHREAD_PROCESS_SHARED` allocates a private lock, as guest code runs in one process. libthr spins forever when the holder locks again; here that call throws
+- [pthread_cond_reltimedwait_np](../../core/libs/prx/libkernel/Pthread/Posix/Cond.cpp) (libkernel) takes `(cond, mutex, usec)`, a relative timeout in microseconds, and is `scePthreadCondTimedwait` with POSIX error numbers. FreeBSD 11.4 has no such function; the signature comes only from shadPS4, which registers one function for this NID and for `scePthreadCondTimedwait` ([`condvar.cpp`](https://github.com/shadps4-emu/shadPS4/blob/3a9f57265d28975658f1ca941daa5c879c7a74bd/src/core/libraries/kernel/threads/condvar.cpp#L229-L234)). shadPS4 reads a 64-bit count; here it is the 32-bit `SceKernelUseconds` of the SCE twin, as for `sem_reltimedwait_np`. Unverified on the console