blob: 1eefed6d0749d90688fb56c8f23831400e765417 (
plain) (
tree)
|
|
// Copyright 2018 yuzu emulator team
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <condition_variable>
#include <mutex>
#include "common/logging/log.h"
#ifdef ARCHITECTURE_x86_64
#include "core/arm/dynarmic/arm_dynarmic.h"
#endif
#include "core/arm/exclusive_monitor.h"
#include "core/arm/unicorn/arm_unicorn.h"
#include "core/core.h"
#include "core/core_cpu.h"
#include "core/core_timing.h"
#include "core/hle/kernel/scheduler.h"
#include "core/hle/kernel/thread.h"
#include "core/hle/lock.h"
#include "core/settings.h"
namespace Core {
void CpuBarrier::NotifyEnd() {
std::unique_lock<std::mutex> lock(mutex);
end = true;
condition.notify_all();
}
bool CpuBarrier::Rendezvous() {
if (!Settings::values.use_multi_core) {
// Meaningless when running in single-core mode
return true;
}
if (!end) {
std::unique_lock<std::mutex> lock(mutex);
--cores_waiting;
if (!cores_waiting) {
cores_waiting = NUM_CPU_CORES;
condition.notify_all();
return true;
}
condition.wait(lock);
return true;
}
return false;
}
Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_barrier,
std::size_t core_index)
: cpu_barrier{cpu_barrier}, core_timing{system.CoreTiming()}, core_index{core_index} {
if (Settings::values.use_cpu_jit) {
#ifdef ARCHITECTURE_x86_64
arm_interface = std::make_unique<ARM_Dynarmic>(core_timing, exclusive_monitor, core_index);
#else
arm_interface = std::make_unique<ARM_Unicorn>();
LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
#endif
} else {
arm_interface = std::make_unique<ARM_Unicorn>(core_timing);
}
scheduler = std::make_unique<Kernel::Scheduler>(system, *arm_interface);
}
Cpu::~Cpu() = default;
std::unique_ptr<ExclusiveMonitor> Cpu::MakeExclusiveMonitor(std::size_t num_cores) {
if (Settings::values.use_cpu_jit) {
#ifdef ARCHITECTURE_x86_64
return std::make_unique<DynarmicExclusiveMonitor>(num_cores);
#else
return nullptr; // TODO(merry): Passthrough exclusive monitor
#endif
} else {
return nullptr; // TODO(merry): Passthrough exclusive monitor
}
}
void Cpu::RunLoop(bool tight_loop) {
// Wait for all other CPU cores to complete the previous slice, such that they run in lock-step
if (!cpu_barrier.Rendezvous()) {
// If rendezvous failed, session has been killed
return;
}
// If we don't have a currently active thread then don't execute instructions,
// instead advance to the next event and try to yield to the next thread
if (Kernel::GetCurrentThread() == nullptr) {
LOG_TRACE(Core, "Core-{} idling", core_index);
if (IsMainCore()) {
// TODO(Subv): Only let CoreTiming idle if all 4 cores are idling.
core_timing.Idle();
core_timing.Advance();
}
PrepareReschedule();
} else {
if (IsMainCore()) {
core_timing.Advance();
}
if (tight_loop) {
arm_interface->Run();
} else {
arm_interface->Step();
}
}
Reschedule();
}
void Cpu::SingleStep() {
return RunLoop(false);
}
void Cpu::PrepareReschedule() {
arm_interface->PrepareReschedule();
reschedule_pending = true;
}
void Cpu::Reschedule() {
if (!reschedule_pending) {
return;
}
reschedule_pending = false;
// Lock the global kernel mutex when we manipulate the HLE state
std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
scheduler->Reschedule();
}
} // namespace Core
|