Merge pull request #2410 from lioncash/affinity
kernel/svc: Reorganize and fix up the initial handling of svcSetThreadCoreMask()
This commit is contained in:
commit
13dda1d8ed
|
@ -1342,7 +1342,7 @@ static void ExitProcess(Core::System& system) {
|
||||||
/// Creates a new thread
|
/// Creates a new thread
|
||||||
static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr entry_point, u64 arg,
|
static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr entry_point, u64 arg,
|
||||||
VAddr stack_top, u32 priority, s32 processor_id) {
|
VAddr stack_top, u32 priority, s32 processor_id) {
|
||||||
LOG_TRACE(Kernel_SVC,
|
LOG_DEBUG(Kernel_SVC,
|
||||||
"called entrypoint=0x{:08X}, arg=0x{:08X}, stacktop=0x{:08X}, "
|
"called entrypoint=0x{:08X}, arg=0x{:08X}, stacktop=0x{:08X}, "
|
||||||
"threadpriority=0x{:08X}, processorid=0x{:08X} : created handle=0x{:08X}",
|
"threadpriority=0x{:08X}, processorid=0x{:08X} : created handle=0x{:08X}",
|
||||||
entry_point, arg, stack_top, priority, processor_id, *out_handle);
|
entry_point, arg, stack_top, priority, processor_id, *out_handle);
|
||||||
|
@ -1402,7 +1402,7 @@ static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr e
|
||||||
|
|
||||||
/// Starts the thread for the provided handle
|
/// Starts the thread for the provided handle
|
||||||
static ResultCode StartThread(Core::System& system, Handle thread_handle) {
|
static ResultCode StartThread(Core::System& system, Handle thread_handle) {
|
||||||
LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
|
LOG_DEBUG(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
|
||||||
|
|
||||||
const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
|
const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
|
||||||
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
||||||
|
@ -1425,7 +1425,7 @@ static ResultCode StartThread(Core::System& system, Handle thread_handle) {
|
||||||
|
|
||||||
/// Called when a thread exits
|
/// Called when a thread exits
|
||||||
static void ExitThread(Core::System& system) {
|
static void ExitThread(Core::System& system) {
|
||||||
LOG_TRACE(Kernel_SVC, "called, pc=0x{:08X}", system.CurrentArmInterface().GetPC());
|
LOG_DEBUG(Kernel_SVC, "called, pc=0x{:08X}", system.CurrentArmInterface().GetPC());
|
||||||
|
|
||||||
auto* const current_thread = system.CurrentScheduler().GetCurrentThread();
|
auto* const current_thread = system.CurrentScheduler().GetCurrentThread();
|
||||||
current_thread->Stop();
|
current_thread->Stop();
|
||||||
|
@ -1435,7 +1435,7 @@ static void ExitThread(Core::System& system) {
|
||||||
|
|
||||||
/// Sleep the current thread
|
/// Sleep the current thread
|
||||||
static void SleepThread(Core::System& system, s64 nanoseconds) {
|
static void SleepThread(Core::System& system, s64 nanoseconds) {
|
||||||
LOG_TRACE(Kernel_SVC, "called nanoseconds={}", nanoseconds);
|
LOG_DEBUG(Kernel_SVC, "called nanoseconds={}", nanoseconds);
|
||||||
|
|
||||||
enum class SleepType : s64 {
|
enum class SleepType : s64 {
|
||||||
YieldWithoutLoadBalancing = 0,
|
YieldWithoutLoadBalancing = 0,
|
||||||
|
@ -1880,11 +1880,51 @@ static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle,
|
||||||
}
|
}
|
||||||
|
|
||||||
static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, u32 core,
|
static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, u32 core,
|
||||||
u64 mask) {
|
u64 affinity_mask) {
|
||||||
LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, mask=0x{:016X}, core=0x{:X}", thread_handle,
|
LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, core=0x{:X}, affinity_mask=0x{:016X}",
|
||||||
mask, core);
|
thread_handle, core, affinity_mask);
|
||||||
|
|
||||||
const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
|
const auto* const current_process = system.Kernel().CurrentProcess();
|
||||||
|
|
||||||
|
if (core == static_cast<u32>(THREADPROCESSORID_IDEAL)) {
|
||||||
|
const u8 ideal_cpu_core = current_process->GetIdealCore();
|
||||||
|
|
||||||
|
ASSERT(ideal_cpu_core != static_cast<u8>(THREADPROCESSORID_IDEAL));
|
||||||
|
|
||||||
|
// Set the target CPU to the ideal core specified by the process.
|
||||||
|
core = ideal_cpu_core;
|
||||||
|
affinity_mask = 1ULL << core;
|
||||||
|
} else {
|
||||||
|
const u64 core_mask = current_process->GetCoreMask();
|
||||||
|
|
||||||
|
if ((core_mask | affinity_mask) != core_mask) {
|
||||||
|
LOG_ERROR(
|
||||||
|
Kernel_SVC,
|
||||||
|
"Invalid processor ID specified (core_mask=0x{:08X}, affinity_mask=0x{:016X})",
|
||||||
|
core_mask, affinity_mask);
|
||||||
|
return ERR_INVALID_PROCESSOR_ID;
|
||||||
|
}
|
||||||
|
|
||||||
|
if (affinity_mask == 0) {
|
||||||
|
LOG_ERROR(Kernel_SVC, "Specfified affinity mask is zero.");
|
||||||
|
return ERR_INVALID_COMBINATION;
|
||||||
|
}
|
||||||
|
|
||||||
|
if (core < Core::NUM_CPU_CORES) {
|
||||||
|
if ((affinity_mask & (1ULL << core)) == 0) {
|
||||||
|
LOG_ERROR(Kernel_SVC,
|
||||||
|
"Core is not enabled for the current mask, core={}, mask={:016X}", core,
|
||||||
|
affinity_mask);
|
||||||
|
return ERR_INVALID_COMBINATION;
|
||||||
|
}
|
||||||
|
} else if (core != static_cast<u32>(THREADPROCESSORID_DONT_CARE) &&
|
||||||
|
core != static_cast<u32>(THREADPROCESSORID_DONT_UPDATE)) {
|
||||||
|
LOG_ERROR(Kernel_SVC, "Invalid processor ID specified (core={}).", core);
|
||||||
|
return ERR_INVALID_PROCESSOR_ID;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
const auto& handle_table = current_process->GetHandleTable();
|
||||||
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
||||||
if (!thread) {
|
if (!thread) {
|
||||||
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
|
||||||
|
@ -1892,40 +1932,7 @@ static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle,
|
||||||
return ERR_INVALID_HANDLE;
|
return ERR_INVALID_HANDLE;
|
||||||
}
|
}
|
||||||
|
|
||||||
if (core == static_cast<u32>(THREADPROCESSORID_IDEAL)) {
|
thread->ChangeCore(core, affinity_mask);
|
||||||
const u8 ideal_cpu_core = thread->GetOwnerProcess()->GetIdealCore();
|
|
||||||
|
|
||||||
ASSERT(ideal_cpu_core != static_cast<u8>(THREADPROCESSORID_IDEAL));
|
|
||||||
|
|
||||||
// Set the target CPU to the ideal core specified by the process.
|
|
||||||
core = ideal_cpu_core;
|
|
||||||
mask = 1ULL << core;
|
|
||||||
}
|
|
||||||
|
|
||||||
if (mask == 0) {
|
|
||||||
LOG_ERROR(Kernel_SVC, "Mask is 0");
|
|
||||||
return ERR_INVALID_COMBINATION;
|
|
||||||
}
|
|
||||||
|
|
||||||
/// This value is used to only change the affinity mask without changing the current ideal core.
|
|
||||||
static constexpr u32 OnlyChangeMask = static_cast<u32>(-3);
|
|
||||||
|
|
||||||
if (core == OnlyChangeMask) {
|
|
||||||
core = thread->GetIdealCore();
|
|
||||||
} else if (core >= Core::NUM_CPU_CORES && core != static_cast<u32>(-1)) {
|
|
||||||
LOG_ERROR(Kernel_SVC, "Invalid core specified, got {}", core);
|
|
||||||
return ERR_INVALID_PROCESSOR_ID;
|
|
||||||
}
|
|
||||||
|
|
||||||
// Error out if the input core isn't enabled in the input mask.
|
|
||||||
if (core < Core::NUM_CPU_CORES && (mask & (1ull << core)) == 0) {
|
|
||||||
LOG_ERROR(Kernel_SVC, "Core is not enabled for the current mask, core={}, mask={:016X}",
|
|
||||||
core, mask);
|
|
||||||
return ERR_INVALID_COMBINATION;
|
|
||||||
}
|
|
||||||
|
|
||||||
thread->ChangeCore(core, mask);
|
|
||||||
|
|
||||||
return RESULT_SUCCESS;
|
return RESULT_SUCCESS;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
|
@ -30,12 +30,21 @@ enum ThreadPriority : u32 {
|
||||||
};
|
};
|
||||||
|
|
||||||
enum ThreadProcessorId : s32 {
|
enum ThreadProcessorId : s32 {
|
||||||
THREADPROCESSORID_IDEAL = -2, ///< Run thread on the ideal core specified by the process.
|
/// Indicates that no particular processor core is preferred.
|
||||||
THREADPROCESSORID_0 = 0, ///< Run thread on core 0
|
THREADPROCESSORID_DONT_CARE = -1,
|
||||||
THREADPROCESSORID_1 = 1, ///< Run thread on core 1
|
|
||||||
THREADPROCESSORID_2 = 2, ///< Run thread on core 2
|
/// Run thread on the ideal core specified by the process.
|
||||||
THREADPROCESSORID_3 = 3, ///< Run thread on core 3
|
THREADPROCESSORID_IDEAL = -2,
|
||||||
THREADPROCESSORID_MAX = 4, ///< Processor ID must be less than this
|
|
||||||
|
/// Indicates that the preferred processor ID shouldn't be updated in
|
||||||
|
/// a core mask setting operation.
|
||||||
|
THREADPROCESSORID_DONT_UPDATE = -3,
|
||||||
|
|
||||||
|
THREADPROCESSORID_0 = 0, ///< Run thread on core 0
|
||||||
|
THREADPROCESSORID_1 = 1, ///< Run thread on core 1
|
||||||
|
THREADPROCESSORID_2 = 2, ///< Run thread on core 2
|
||||||
|
THREADPROCESSORID_3 = 3, ///< Run thread on core 3
|
||||||
|
THREADPROCESSORID_MAX = 4, ///< Processor ID must be less than this
|
||||||
|
|
||||||
/// Allowed CPU mask
|
/// Allowed CPU mask
|
||||||
THREADPROCESSORID_DEFAULT_MASK = (1 << THREADPROCESSORID_0) | (1 << THREADPROCESSORID_1) |
|
THREADPROCESSORID_DEFAULT_MASK = (1 << THREADPROCESSORID_0) | (1 << THREADPROCESSORID_1) |
|
||||||
|
|
Loading…
Reference in a new issue