citra/src/core/hle/kernel/thread.cpp

// Copyright 2014 Citra Emulator Project / PPSSPP Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.

#include <algorithm>
#include <climits>
#include <list>
#include <vector>
#include <boost/serialization/string.hpp>
#include "common/archives.h"
#include "common/assert.h"
#include "common/common_types.h"
#include "common/logging/log.h"
#include "common/math_util.h"
#include "common/serialization/boost_flat_set.h"
#include "core/arm/arm_interface.h"
#include "core/arm/skyeye_common/armstate.h"
#include "core/core.h"
#include "core/hle/kernel/errors.h"
#include "core/hle/kernel/handle_table.h"
#include "core/hle/kernel/kernel.h"
#include "core/hle/kernel/memory.h"
#include "core/hle/kernel/mutex.h"
#include "core/hle/kernel/process.h"
#include "core/hle/kernel/thread.h"
#include "core/hle/result.h"
#include "core/memory.h"

SERIALIZE_EXPORT_IMPL(Kernel::Thread)

namespace Kernel {

template <class Archive>
void Thread::serialize(Archive& ar, const unsigned int file_version) {
    ar& boost::serialization::base_object<WaitObject>(*this);
    ar&* context.get();
    ar& thread_id;
    ar& status;
    ar& entry_point;
    ar& stack_top;
    ar& nominal_priority;
    ar& current_priority;
    ar& last_running_ticks;
    ar& processor_id;
    ar& tls_address;
    ar& held_mutexes;
    ar& pending_mutexes;

    // Note: this is equivalent of what is done in boost/serialization/weak_ptr.hpp, but it's
    // compatible with previous versions of savestates.
    // TODO(SaveStates): When the savestate version is bumped, simplify this again.
    std::shared_ptr<Process> shared_owner_process = owner_process.lock();
    ar& shared_owner_process;
    if (Archive::is_loading::value) {
        owner_process = shared_owner_process;
    }

    ar& wait_objects;
    ar& wait_address;
    ar& name;
    ar& wakeup_callback;
}

SERIALIZE_IMPL(Thread)

bool Thread::ShouldWait(const Thread* thread) const {
    return status != ThreadStatus::Dead;
}

void Thread::Acquire(Thread* thread) {
    ASSERT_MSG(!ShouldWait(thread), "object unavailable!");
}

Thread::Thread(KernelSystem& kernel, u32 core_id)
    : WaitObject(kernel), context(kernel.GetThreadManager(core_id).NewContext()),
      can_schedule(true), core_id(core_id), thread_manager(kernel.GetThreadManager(core_id)) {}
Thread::~Thread() {}

Thread* ThreadManager::GetCurrentThread() const {
    return current_thread.get();
}

void Thread::Stop() {
    // Cancel any outstanding wakeup events for this thread
    thread_manager.kernel.timing.UnscheduleEvent(thread_manager.ThreadWakeupEventType, thread_id);
    thread_manager.wakeup_callback_table.erase(thread_id);

    // Clean up thread from ready queue
    // This is only needed when the thread is termintated forcefully (SVC TerminateProcess)
    if (status == ThreadStatus::Ready) {
        thread_manager.ready_queue.remove(current_priority, this);
    }

    status = ThreadStatus::Dead;

    WakeupAllWaitingThreads();

    // Clean up any dangling references in objects that this thread was waiting for
    for (auto& wait_object : wait_objects) {
        wait_object->RemoveWaitingThread(this);
    }
    wait_objects.clear();

    // Release all the mutexes that this thread holds
    ReleaseThreadMutexes(this);

    // Mark the TLS slot in the thread's page as free.
    u32 tls_page = (tls_address - Memory::TLS_AREA_VADDR) / Memory::CITRA_PAGE_SIZE;
    u32 tls_slot =
        ((tls_address - Memory::TLS_AREA_VADDR) % Memory::CITRA_PAGE_SIZE) / Memory::TLS_ENTRY_SIZE;
    if (auto process = owner_process.lock()) {
        process->tls_slots[tls_page].reset(tls_slot);
    }
}

void ThreadManager::SwitchContext(Thread* new_thread) {
    Thread* previous_thread = GetCurrentThread();
    std::shared_ptr<Process> previous_process = nullptr;

    Core::Timing& timing = kernel.timing;

    // Save context for previous thread
    if (previous_thread) {
        previous_process = previous_thread->owner_process.lock();
        previous_thread->last_running_ticks = cpu->GetTimer().GetTicks();
        cpu->SaveContext(previous_thread->context);

        if (previous_thread->status == ThreadStatus::Running) {
            // This is only the case when a reschedule is triggered without the current thread
            // yielding execution (i.e. an event triggered, system core time-sliced, etc)
            ready_queue.push_front(previous_thread->current_priority, previous_thread);
            previous_thread->status = ThreadStatus::Ready;
        }
    }

    // Load context of new thread
    if (new_thread) {
        ASSERT_MSG(new_thread->status == ThreadStatus::Ready,
                   "Thread must be ready to become running.");

        // Cancel any outstanding wakeup events for this thread
        timing.UnscheduleEvent(ThreadWakeupEventType, new_thread->thread_id);

        current_thread = SharedFrom(new_thread);

        ready_queue.remove(new_thread->current_priority, new_thread);
        new_thread->status = ThreadStatus::Running;

        ASSERT(current_thread->owner_process.lock());
        if (previous_process != current_thread->owner_process.lock()) {
            kernel.SetCurrentProcessForCPU(current_thread->owner_process.lock(), cpu->GetID());
        }

        cpu->LoadContext(new_thread->context);
        cpu->SetCP15Register(CP15_THREAD_URO, new_thread->GetTLSAddress());
    } else {
        current_thread = nullptr;
        // Note: We do not reset the current process and current page table when idling because
        // technically we haven't changed processes, our threads are just paused.
    }
}

Thread* ThreadManager::PopNextReadyThread() {
    Thread* next = nullptr;
    Thread* thread = GetCurrentThread();

    if (thread && thread->status == ThreadStatus::Running) {
        do {
            // We have to do better than the current thread.
            // This call returns null when that's not possible.
            next = ready_queue.pop_first_better(thread->current_priority);
            if (!next) {
                // Otherwise just keep going with the current thread
                next = thread;
                break;
            } else if (!next->can_schedule)
                unscheduled_ready_queue.push_back(next);
        } while (!next->can_schedule);
    } else {
        do {
            next = ready_queue.pop_first();
            if (next && !next->can_schedule)
                unscheduled_ready_queue.push_back(next);
        } while (next && !next->can_schedule);
    }

    while (!unscheduled_ready_queue.empty()) {
        auto t = std::move(unscheduled_ready_queue.back());
        ready_queue.push_back(t->current_priority, t);
        unscheduled_ready_queue.pop_back();
    }

    return next;
}

void ThreadManager::WaitCurrentThread_Sleep() {
    Thread* thread = GetCurrentThread();
    thread->status = ThreadStatus::WaitSleep;
}

void ThreadManager::ExitCurrentThread() {
    Thread* thread = GetCurrentThread();
    thread->Stop();
    thread_list.erase(std::remove_if(thread_list.begin(), thread_list.end(),
                                     [thread](const auto& p) { return p.get() == thread; }),
                      thread_list.end());
}

void ThreadManager::ThreadWakeupCallback(u64 thread_id, s64 cycles_late) {
    std::shared_ptr<Thread> thread = SharedFrom(wakeup_callback_table.at(thread_id));
    if (thread == nullptr) {
        LOG_CRITICAL(Kernel, "Callback fired for invalid thread {:08X}", thread_id);
        return;
    }

    if (thread->status == ThreadStatus::WaitSynchAny ||
        thread->status == ThreadStatus::WaitSynchAll || thread->status == ThreadStatus::WaitArb ||
        thread->status == ThreadStatus::WaitHleEvent) {

        // Invoke the wakeup callback before clearing the wait objects
        if (thread->wakeup_callback)
            thread->wakeup_callback->WakeUp(ThreadWakeupReason::Timeout, thread, nullptr);

        // Remove the thread from each of its waiting objects' waitlists
        for (auto& object : thread->wait_objects)
            object->RemoveWaitingThread(thread.get());
        thread->wait_objects.clear();
    }

    thread->ResumeFromWait();
}

void Thread::WakeAfterDelay(s64 nanoseconds) {
    // Don't schedule a wakeup if the thread wants to wait forever
    if (nanoseconds == -1)
        return;

    thread_manager.kernel.timing.ScheduleEvent(nsToCycles(nanoseconds),
                                               thread_manager.ThreadWakeupEventType, thread_id);
}

void Thread::ResumeFromWait() {
    ASSERT_MSG(wait_objects.empty(), "Thread is waking up while waiting for objects");

    switch (status) {
    case ThreadStatus::WaitSynchAll:
    case ThreadStatus::WaitSynchAny:
    case ThreadStatus::WaitHleEvent:
    case ThreadStatus::WaitArb:
    case ThreadStatus::WaitSleep:
    case ThreadStatus::WaitIPC:
    case ThreadStatus::Dormant:
        break;

    case ThreadStatus::Ready:
        // The thread's wakeup callback must have already been cleared when the thread was first
        // awoken.
        ASSERT(wakeup_callback == nullptr);
        // If the thread is waiting on multiple wait objects, it might be awoken more than once
        // before actually resuming. We can ignore subsequent wakeups if the thread status has
        // already been set to ThreadStatus::Ready.
        return;

    case ThreadStatus::Running:
        DEBUG_ASSERT_MSG(false, "Thread with object id {} has already resumed.", GetObjectId());
        return;
    case ThreadStatus::Dead:
        // This should never happen, as threads must complete before being stopped.
        DEBUG_ASSERT_MSG(false, "Thread with object id {} cannot be resumed because it's DEAD.",
                         GetObjectId());
        return;
    }

    wakeup_callback = nullptr;

    thread_manager.ready_queue.push_back(current_priority, this);
    status = ThreadStatus::Ready;
    thread_manager.kernel.PrepareReschedule();
}

void ThreadManager::DebugThreadQueue() {
    Thread* thread = GetCurrentThread();
    if (!thread) {
        LOG_DEBUG(Kernel, "Current: NO CURRENT THREAD");
    } else {
        LOG_DEBUG(Kernel, "0x{:02X} {} (current)", thread->current_priority,
                  GetCurrentThread()->GetObjectId());
    }

    for (auto& t : thread_list) {
        u32 priority = ready_queue.contains(t.get());
        if (priority != UINT_MAX) {
            LOG_DEBUG(Kernel, "0x{:02X} {}", priority, t->GetObjectId());
        }
    }
}

/**
 * Finds a free location for the TLS section of a thread.
 * @param tls_slots The TLS page array of the thread's owner process.
 * Returns a tuple of (page, slot, alloc_needed) where:
 * page: The index of the first allocated TLS page that has free slots.
 * slot: The index of the first free slot in the indicated page.
 * alloc_needed: Whether there's a need to allocate a new TLS page (All pages are full).
 */
static std::tuple<std::size_t, std::size_t, bool> GetFreeThreadLocalSlot(
    const std::vector<std::bitset<8>>& tls_slots) {
    // Iterate over all the allocated pages, and try to find one where not all slots are used.
    for (std::size_t page = 0; page < tls_slots.size(); ++page) {
        const auto& page_tls_slots = tls_slots[page];
        if (!page_tls_slots.all()) {
            // We found a page with at least one free slot, find which slot it is
            for (std::size_t slot = 0; slot < page_tls_slots.size(); ++slot) {
                if (!page_tls_slots.test(slot)) {
                    return std::make_tuple(page, slot, false);
                }
            }
        }
    }

    return std::make_tuple(0, 0, true);
}

/**
 * Resets a thread context, making it ready to be scheduled and run by the CPU
 * @param context Thread context to reset
 * @param stack_top Address of the top of the stack
 * @param entry_point Address of entry point for execution
 * @param arg User argument for thread
 */
static void ResetThreadContext(const std::unique_ptr<ARM_Interface::ThreadContext>& context,
                               u32 stack_top, u32 entry_point, u32 arg) {
    context->Reset();
    context->SetCpuRegister(0, arg);
    context->SetProgramCounter(entry_point);
    context->SetStackPointer(stack_top);
    context->SetCpsr(USER32MODE | ((entry_point & 1) << 5)); // Usermode and THUMB mode
}

ResultVal<std::shared_ptr<Thread>> KernelSystem::CreateThread(
    std::string name, VAddr entry_point, u32 priority, u32 arg, s32 processor_id, VAddr stack_top,
    std::shared_ptr<Process> owner_process) {
    // Check if priority is in ranged. Lowest priority -> highest priority id.
    if (priority > ThreadPrioLowest) {
        LOG_ERROR(Kernel_SVC, "Invalid thread priority: {}", priority);
        return ERR_OUT_OF_RANGE;
    }

    if (processor_id > ThreadProcessorIdMax) {
        LOG_ERROR(Kernel_SVC, "Invalid processor id: {}", processor_id);
        return ERR_OUT_OF_RANGE_KERNEL;
    }

    // TODO(yuriks): Other checks, returning 0xD9001BEA
    if (!memory.IsValidVirtualAddress(*owner_process, entry_point)) {
        LOG_ERROR(Kernel_SVC, "(name={}): invalid entry {:08x}", name, entry_point);
        // TODO: Verify error
        return ResultCode(ErrorDescription::InvalidAddress, ErrorModule::Kernel,
                          ErrorSummary::InvalidArgument, ErrorLevel::Permanent);
    }

    auto thread{std::make_shared<Thread>(*this, processor_id)};

    thread_managers[processor_id]->thread_list.push_back(thread);
    thread_managers[processor_id]->ready_queue.prepare(priority);

    thread->thread_id = NewThreadId();
    thread->status = ThreadStatus::Dormant;
    thread->entry_point = entry_point;
    thread->stack_top = stack_top;
    thread->nominal_priority = thread->current_priority = priority;
    thread->last_running_ticks = timing.GetTimer(processor_id)->GetTicks();
    thread->processor_id = processor_id;
    thread->wait_objects.clear();
    thread->wait_address = 0;
    thread->name = std::move(name);
    thread_managers[processor_id]->wakeup_callback_table[thread->thread_id] = thread.get();
    thread->owner_process = owner_process;

    // Find the next available TLS index, and mark it as used
    auto& tls_slots = owner_process->tls_slots;

    auto [available_page, available_slot, needs_allocation] = GetFreeThreadLocalSlot(tls_slots);

    if (needs_allocation) {
        // There are no already-allocated pages with free slots, lets allocate a new one.
        // TLS pages are allocated from the BASE region in the linear heap.
        auto memory_region = GetMemoryRegion(MemoryRegion::BASE);

        // Allocate some memory from the end of the linear heap for this region.
        auto offset = memory_region->LinearAllocate(Memory::CITRA_PAGE_SIZE);
        if (!offset) {
            LOG_ERROR(Kernel_SVC,
                      "Not enough space in region to allocate a new TLS page for thread");
            return ERR_OUT_OF_MEMORY;
        }
        owner_process->memory_used += Memory::CITRA_PAGE_SIZE;

        tls_slots.emplace_back(0); // The page is completely available at the start
        available_page = tls_slots.size() - 1;
        available_slot = 0; // Use the first slot in the new page

        auto& vm_manager = owner_process->vm_manager;

        // Map the page to the current process' address space.
        vm_manager.MapBackingMemory(
            Memory::TLS_AREA_VADDR + static_cast<VAddr>(available_page) * Memory::CITRA_PAGE_SIZE,
            memory.GetFCRAMRef(*offset), Memory::CITRA_PAGE_SIZE, MemoryState::Locked);
    }

    // Mark the slot as used
    tls_slots[available_page].set(available_slot);
    thread->tls_address = Memory::TLS_AREA_VADDR +
                          static_cast<VAddr>(available_page) * Memory::CITRA_PAGE_SIZE +
                          static_cast<VAddr>(available_slot) * Memory::TLS_ENTRY_SIZE;

    memory.ZeroBlock(*owner_process, thread->tls_address, Memory::TLS_ENTRY_SIZE);

    // TODO(peachum): move to ScheduleThread() when scheduler is added so selected core is used
    // to initialize the context
    ResetThreadContext(thread->context, stack_top, entry_point, arg);

    thread_managers[processor_id]->ready_queue.push_back(thread->current_priority, thread.get());
    thread->status = ThreadStatus::Ready;

    return MakeResult<std::shared_ptr<Thread>>(std::move(thread));
}

void Thread::SetPriority(u32 priority) {
    ASSERT_MSG(priority <= ThreadPrioLowest && priority >= ThreadPrioHighest,
               "Invalid priority value.");
    // If thread was ready, adjust queues
    if (status == ThreadStatus::Ready)
        thread_manager.ready_queue.move(this, current_priority, priority);
    else
        thread_manager.ready_queue.prepare(priority);

    nominal_priority = current_priority = priority;
}

void Thread::UpdatePriority() {
    u32 best_priority = nominal_priority;
    for (auto& mutex : held_mutexes) {
        if (mutex->priority < best_priority)
            best_priority = mutex->priority;
    }
    BoostPriority(best_priority);
}

void Thread::BoostPriority(u32 priority) {
    // If thread was ready, adjust queues
    if (status == ThreadStatus::Ready)
        thread_manager.ready_queue.move(this, current_priority, priority);
    else
        thread_manager.ready_queue.prepare(priority);
    current_priority = priority;
}

std::shared_ptr<Thread> SetupMainThread(KernelSystem& kernel, u32 entry_point, u32 priority,
                                        std::shared_ptr<Process> owner_process) {
    // Initialize new "main" thread
    auto thread_res =
        kernel.CreateThread("main", entry_point, priority, 0, owner_process->ideal_processor,
                            Memory::HEAP_VADDR_END, owner_process);

    std::shared_ptr<Thread> thread = std::move(thread_res).Unwrap();

    thread->context->SetFpscr(FPSCR_DEFAULT_NAN | FPSCR_FLUSH_TO_ZERO | FPSCR_ROUND_TOZERO |
                              FPSCR_IXC); // 0x03C00010

    // Note: The newly created thread will be run when the scheduler fires.
    return thread;
}

bool ThreadManager::HaveReadyThreads() {
    return ready_queue.get_first() != nullptr;
}

void ThreadManager::Reschedule() {
    Thread* cur = GetCurrentThread();
    Thread* next = PopNextReadyThread();

    if (cur && next) {
        LOG_TRACE(Kernel, "context switch {} -> {}", cur->GetObjectId(), next->GetObjectId());
    } else if (cur) {
        LOG_TRACE(Kernel, "context switch {} -> idle", cur->GetObjectId());
    } else if (next) {
        LOG_TRACE(Kernel, "context switch idle -> {}", next->GetObjectId());
    } else {
        LOG_TRACE(Kernel, "context switch idle -> idle, do nothing");
        return;
    }

    SwitchContext(next);
}

void Thread::SetWaitSynchronizationResult(ResultCode result) {
    context->SetCpuRegister(0, result.raw);
}

void Thread::SetWaitSynchronizationOutput(s32 output) {
    context->SetCpuRegister(1, output);
}

s32 Thread::GetWaitObjectIndex(const WaitObject* object) const {
    ASSERT_MSG(!wait_objects.empty(), "Thread is not waiting for anything");
    const auto match = std::find_if(wait_objects.rbegin(), wait_objects.rend(),
                                    [object](const auto& p) { return p.get() == object; });
    return static_cast<s32>(std::distance(match, wait_objects.rend()) - 1);
}

VAddr Thread::GetCommandBufferAddress() const {
    // Offset from the start of TLS at which the IPC command buffer begins.
    constexpr u32 command_header_offset = 0x80;
    return GetTLSAddress() + command_header_offset;
}

ThreadManager::ThreadManager(Kernel::KernelSystem& kernel, u32 core_id) : kernel(kernel) {
    ThreadWakeupEventType = kernel.timing.RegisterEvent(
        "ThreadWakeupCallback_" + std::to_string(core_id),
        [this](u64 thread_id, s64 cycle_late) { ThreadWakeupCallback(thread_id, cycle_late); });
}

ThreadManager::~ThreadManager() {
    for (auto& t : thread_list) {
        t->Stop();
    }
}

const std::vector<std::shared_ptr<Thread>>& ThreadManager::GetThreadList() {
    return thread_list;
}

} // namespace Kernel