mirror of
https://github.com/yuzu-emu/yuzu.git
synced 2024-07-04 23:31:19 +01:00
core: Make the live Cpu instances unique_ptrs instead of shared_ptrs
There's no need for shared ownership here, as the only owning class instance of those Cpu instances is the System class itself. We can also make the thread_to_cpu map use regular pointers instead of shared_ptrs, given that the Cpu instances will always outlive the cases where they're used with that map.
This commit is contained in:
parent
aeadbfa790
commit
59f872a8e0
1 changed files with 9 additions and 9 deletions
|
@ -71,9 +71,9 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
|
||||||
}
|
}
|
||||||
|
|
||||||
/// Runs a CPU core while the system is powered on
|
/// Runs a CPU core while the system is powered on
|
||||||
void RunCpuCore(std::shared_ptr<Cpu> cpu_state) {
|
void RunCpuCore(Cpu& cpu_state) {
|
||||||
while (Core::System::GetInstance().IsPoweredOn()) {
|
while (Core::System::GetInstance().IsPoweredOn()) {
|
||||||
cpu_state->RunLoop(true);
|
cpu_state.RunLoop(true);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
} // Anonymous namespace
|
} // Anonymous namespace
|
||||||
|
@ -95,7 +95,7 @@ struct System::Impl {
|
||||||
status = ResultStatus::Success;
|
status = ResultStatus::Success;
|
||||||
|
|
||||||
// Update thread_to_cpu in case Core 0 is run from a different host thread
|
// Update thread_to_cpu in case Core 0 is run from a different host thread
|
||||||
thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
|
thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0].get();
|
||||||
|
|
||||||
if (GDBStub::IsServerEnabled()) {
|
if (GDBStub::IsServerEnabled()) {
|
||||||
GDBStub::HandlePacket();
|
GDBStub::HandlePacket();
|
||||||
|
@ -142,7 +142,7 @@ struct System::Impl {
|
||||||
cpu_barrier = std::make_unique<CpuBarrier>();
|
cpu_barrier = std::make_unique<CpuBarrier>();
|
||||||
cpu_exclusive_monitor = Cpu::MakeExclusiveMonitor(cpu_cores.size());
|
cpu_exclusive_monitor = Cpu::MakeExclusiveMonitor(cpu_cores.size());
|
||||||
for (std::size_t index = 0; index < cpu_cores.size(); ++index) {
|
for (std::size_t index = 0; index < cpu_cores.size(); ++index) {
|
||||||
cpu_cores[index] = std::make_shared<Cpu>(*cpu_exclusive_monitor, *cpu_barrier, index);
|
cpu_cores[index] = std::make_unique<Cpu>(*cpu_exclusive_monitor, *cpu_barrier, index);
|
||||||
}
|
}
|
||||||
|
|
||||||
telemetry_session = std::make_unique<Core::TelemetrySession>();
|
telemetry_session = std::make_unique<Core::TelemetrySession>();
|
||||||
|
@ -160,12 +160,12 @@ struct System::Impl {
|
||||||
|
|
||||||
// Create threads for CPU cores 1-3, and build thread_to_cpu map
|
// Create threads for CPU cores 1-3, and build thread_to_cpu map
|
||||||
// CPU core 0 is run on the main thread
|
// CPU core 0 is run on the main thread
|
||||||
thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
|
thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0].get();
|
||||||
if (Settings::values.use_multi_core) {
|
if (Settings::values.use_multi_core) {
|
||||||
for (std::size_t index = 0; index < cpu_core_threads.size(); ++index) {
|
for (std::size_t index = 0; index < cpu_core_threads.size(); ++index) {
|
||||||
cpu_core_threads[index] =
|
cpu_core_threads[index] =
|
||||||
std::make_unique<std::thread>(RunCpuCore, cpu_cores[index + 1]);
|
std::make_unique<std::thread>(RunCpuCore, std::ref(*cpu_cores[index + 1]));
|
||||||
thread_to_cpu[cpu_core_threads[index]->get_id()] = cpu_cores[index + 1];
|
thread_to_cpu[cpu_core_threads[index]->get_id()] = cpu_cores[index + 1].get();
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -285,7 +285,7 @@ struct System::Impl {
|
||||||
std::shared_ptr<Tegra::DebugContext> debug_context;
|
std::shared_ptr<Tegra::DebugContext> debug_context;
|
||||||
std::unique_ptr<ExclusiveMonitor> cpu_exclusive_monitor;
|
std::unique_ptr<ExclusiveMonitor> cpu_exclusive_monitor;
|
||||||
std::unique_ptr<CpuBarrier> cpu_barrier;
|
std::unique_ptr<CpuBarrier> cpu_barrier;
|
||||||
std::array<std::shared_ptr<Cpu>, NUM_CPU_CORES> cpu_cores;
|
std::array<std::unique_ptr<Cpu>, NUM_CPU_CORES> cpu_cores;
|
||||||
std::array<std::unique_ptr<std::thread>, NUM_CPU_CORES - 1> cpu_core_threads;
|
std::array<std::unique_ptr<std::thread>, NUM_CPU_CORES - 1> cpu_core_threads;
|
||||||
std::size_t active_core{}; ///< Active core, only used in single thread mode
|
std::size_t active_core{}; ///< Active core, only used in single thread mode
|
||||||
|
|
||||||
|
@ -299,7 +299,7 @@ struct System::Impl {
|
||||||
std::string status_details = "";
|
std::string status_details = "";
|
||||||
|
|
||||||
/// Map of guest threads to CPU cores
|
/// Map of guest threads to CPU cores
|
||||||
std::map<std::thread::id, std::shared_ptr<Cpu>> thread_to_cpu;
|
std::map<std::thread::id, Cpu*> thread_to_cpu;
|
||||||
|
|
||||||
Core::PerfStats perf_stats;
|
Core::PerfStats perf_stats;
|
||||||
Core::FrameLimiter frame_limiter;
|
Core::FrameLimiter frame_limiter;
|
||||||
|
|
Loading…
Reference in a new issue