12 files changed, 1553 insertions, 767 deletions

diff --git a/src/core/hle/kernel/hle_ipc.h b/src/core/hle/kernel/hle_ipc.h
index b8a746882..9f764c79a 100644
--- a/src/core/hle/kernel/hle_ipc.h
+++ b/src/core/hle/kernel/hle_ipc.h
@@ -126,9 +126,6 @@ public:
         return server_session;
     }
 
-    using WakeupCallback = std::function<void(
-        std::shared_ptr<KThread> thread, HLERequestContext& context, ThreadWakeupReason reason)>;
-
     /// Populates this context with data from the requesting process/thread.
     ResultCode PopulateFromIncomingCommandBuffer(const HandleTable& handle_table,
                                                  u32_le* src_cmdbuf);
diff --git a/src/core/hle/kernel/k_scheduler.cpp b/src/core/hle/kernel/k_scheduler.cpp
index 233022023..5bdbd9a9b 100644
--- a/src/core/hle/kernel/k_scheduler.cpp
+++ b/src/core/hle/kernel/k_scheduler.cpp
@@ -31,11 +31,15 @@ static void IncrementScheduledCount(Kernel::KThread* thread) {
     }
 }
 
-void KScheduler::RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedule,
-                                 Core::EmuThreadHandle global_thread) {
-    const u32 current_core = global_thread.host_handle;
-    bool must_context_switch = global_thread.guest_handle != InvalidHandle &&
-                               (current_core < Core::Hardware::NUM_CPU_CORES);
+void KScheduler::RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedule) {
+    auto scheduler = kernel.CurrentScheduler();
+
+    u32 current_core{0xF};
+    bool must_context_switch{};
+    if (scheduler) {
+        current_core = scheduler->core_id;
+        must_context_switch = true;
+    }
 
     while (cores_pending_reschedule != 0) {
         const auto core = static_cast<u32>(std::countr_zero(cores_pending_reschedule));
@@ -58,26 +62,25 @@ void KScheduler::RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedul
 
 u64 KScheduler::UpdateHighestPriorityThread(KThread* highest_thread) {
     std::scoped_lock lock{guard};
-    if (KThread* prev_highest_thread = this->state.highest_priority_thread;
+    if (KThread* prev_highest_thread = state.highest_priority_thread;
         prev_highest_thread != highest_thread) {
         if (prev_highest_thread != nullptr) {
             IncrementScheduledCount(prev_highest_thread);
             prev_highest_thread->SetLastScheduledTick(system.CoreTiming().GetCPUTicks());
         }
-        if (this->state.should_count_idle) {
+        if (state.should_count_idle) {
             if (highest_thread != nullptr) {
-                // if (Process* process = highest_thread->GetOwnerProcess(); process != nullptr) {
-                //    process->SetRunningThread(this->core_id, highest_thread,
-                //                              this->state.idle_count);
-                //}
+                if (Process* process = highest_thread->GetOwnerProcess(); process != nullptr) {
+                    process->SetRunningThread(core_id, highest_thread, state.idle_count);
+                }
             } else {
-                this->state.idle_count++;
+                state.idle_count++;
             }
         }
 
-        this->state.highest_priority_thread = highest_thread;
-        this->state.needs_scheduling = true;
-        return (1ULL << this->core_id);
+        state.highest_priority_thread = highest_thread;
+        state.needs_scheduling = true;
+        return (1ULL << core_id);
     } else {
         return 0;
     }
@@ -99,7 +102,20 @@ u64 KScheduler::UpdateHighestPriorityThreadsImpl(KernelCore& kernel) {
         KThread* top_thread = priority_queue.GetScheduledFront(static_cast<s32>(core_id));
         if (top_thread != nullptr) {
             // If the thread has no waiters, we need to check if the process has a thread pinned.
-            // TODO(bunnei): Implement thread pinning
+            if (top_thread->GetNumKernelWaiters() == 0) {
+                if (Process* parent = top_thread->GetOwnerProcess(); parent != nullptr) {
+                    if (KThread* pinned = parent->GetPinnedThread(static_cast<s32>(core_id));
+                        pinned != nullptr && pinned != top_thread) {
+                        // We prefer our parent's pinned thread if possible. However, we also don't
+                        // want to schedule un-runnable threads.
+                        if (pinned->GetRawState() == ThreadState::Runnable) {
+                            top_thread = pinned;
+                        } else {
+                            top_thread = nullptr;
+                        }
+                    }
+                }
+            }
         } else {
             idle_cores |= (1ULL << core_id);
         }
@@ -182,6 +198,19 @@ u64 KScheduler::UpdateHighestPriorityThreadsImpl(KernelCore& kernel) {
     return cores_needing_scheduling;
 }
 
+void KScheduler::ClearPreviousThread(KernelCore& kernel, KThread* thread) {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+    for (size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; ++i) {
+        // Get an atomic reference to the core scheduler's previous thread.
+        std::atomic_ref<KThread*> prev_thread(kernel.Scheduler(static_cast<s32>(i)).prev_thread);
+        static_assert(std::atomic_ref<KThread*>::is_always_lock_free);
+
+        // Atomically clear the previous thread if it's our target.
+        KThread* compare = thread;
+        prev_thread.compare_exchange_strong(compare, nullptr);
+    }
+}
+
 void KScheduler::OnThreadStateChanged(KernelCore& kernel, KThread* thread, ThreadState old_state) {
     ASSERT(kernel.GlobalSchedulerContext().IsLocked());
 
@@ -352,12 +381,14 @@ void KScheduler::DisableScheduling(KernelCore& kernel) {
     }
 }
 
-void KScheduler::EnableScheduling(KernelCore& kernel, u64 cores_needing_scheduling,
-                                  Core::EmuThreadHandle global_thread) {
+void KScheduler::EnableScheduling(KernelCore& kernel, u64 cores_needing_scheduling) {
     if (auto* scheduler = kernel.CurrentScheduler(); scheduler) {
-        scheduler->GetCurrentThread()->EnableDispatch();
+        ASSERT(scheduler->GetCurrentThread()->GetDisableDispatchCount() >= 1);
+        if (scheduler->GetCurrentThread()->GetDisableDispatchCount() >= 1) {
+            scheduler->GetCurrentThread()->EnableDispatch();
+        }
     }
-    RescheduleCores(kernel, cores_needing_scheduling, global_thread);
+    RescheduleCores(kernel, cores_needing_scheduling);
 }
 
 u64 KScheduler::UpdateHighestPriorityThreads(KernelCore& kernel) {
@@ -372,15 +403,13 @@ KSchedulerPriorityQueue& KScheduler::GetPriorityQueue(KernelCore& kernel) {
     return kernel.GlobalSchedulerContext().priority_queue;
 }
 
-void KScheduler::YieldWithoutCoreMigration() {
-    auto& kernel = system.Kernel();
-
+void KScheduler::YieldWithoutCoreMigration(KernelCore& kernel) {
     // Validate preconditions.
     ASSERT(CanSchedule(kernel));
     ASSERT(kernel.CurrentProcess() != nullptr);
 
     // Get the current thread and process.
-    KThread& cur_thread = *GetCurrentThread();
+    KThread& cur_thread = Kernel::GetCurrentThread(kernel);
     Process& cur_process = *kernel.CurrentProcess();
 
     // If the thread's yield count matches, there's nothing for us to do.
@@ -413,15 +442,13 @@ void KScheduler::YieldWithoutCoreMigration() {
     }
 }
 
-void KScheduler::YieldWithCoreMigration() {
-    auto& kernel = system.Kernel();
-
+void KScheduler::YieldWithCoreMigration(KernelCore& kernel) {
     // Validate preconditions.
     ASSERT(CanSchedule(kernel));
     ASSERT(kernel.CurrentProcess() != nullptr);
 
     // Get the current thread and process.
-    KThread& cur_thread = *GetCurrentThread();
+    KThread& cur_thread = Kernel::GetCurrentThread(kernel);
     Process& cur_process = *kernel.CurrentProcess();
 
     // If the thread's yield count matches, there's nothing for us to do.
@@ -503,15 +530,13 @@ void KScheduler::YieldWithCoreMigration() {
     }
 }
 
-void KScheduler::YieldToAnyThread() {
-    auto& kernel = system.Kernel();
-
+void KScheduler::YieldToAnyThread(KernelCore& kernel) {
     // Validate preconditions.
     ASSERT(CanSchedule(kernel));
     ASSERT(kernel.CurrentProcess() != nullptr);
 
     // Get the current thread and process.
-    KThread& cur_thread = *GetCurrentThread();
+    KThread& cur_thread = Kernel::GetCurrentThread(kernel);
     Process& cur_process = *kernel.CurrentProcess();
 
     // If the thread's yield count matches, there's nothing for us to do.
@@ -581,15 +606,14 @@ void KScheduler::YieldToAnyThread() {
     }
 }
 
-KScheduler::KScheduler(Core::System& system, std::size_t core_id)
-    : system(system), core_id(core_id) {
+KScheduler::KScheduler(Core::System& system, s32 core_id) : system(system), core_id(core_id) {
     switch_fiber = std::make_shared<Common::Fiber>(OnSwitch, this);
-    this->state.needs_scheduling = true;
-    this->state.interrupt_task_thread_runnable = false;
-    this->state.should_count_idle = false;
-    this->state.idle_count = 0;
-    this->state.idle_thread_stack = nullptr;
-    this->state.highest_priority_thread = nullptr;
+    state.needs_scheduling = true;
+    state.interrupt_task_thread_runnable = false;
+    state.should_count_idle = false;
+    state.idle_count = 0;
+    state.idle_thread_stack = nullptr;
+    state.highest_priority_thread = nullptr;
 }
 
 KScheduler::~KScheduler() = default;
@@ -613,7 +637,7 @@ void KScheduler::RescheduleCurrentCore() {
         phys_core.ClearInterrupt();
     }
     guard.lock();
-    if (this->state.needs_scheduling) {
+    if (state.needs_scheduling) {
         Schedule();
     } else {
         guard.unlock();
@@ -625,32 +649,34 @@ void KScheduler::OnThreadStart() {
 }
 
 void KScheduler::Unload(KThread* thread) {
+    LOG_TRACE(Kernel, "core {}, unload thread {}", core_id, thread ? thread->GetName() : "nullptr");
+
     if (thread) {
-        thread->SetIsRunning(false);
-        if (thread->IsContinuousOnSVC()) {
+        if (thread->IsCallingSvc()) {
             system.ArmInterface(core_id).ExceptionalExit();
-            thread->SetContinuousOnSVC(false);
+            thread->ClearIsCallingSvc();
         }
-        if (!thread->HasExited()) {
+        if (!thread->IsTerminationRequested()) {
+            prev_thread = thread;
+
             Core::ARM_Interface& cpu_core = system.ArmInterface(core_id);
             cpu_core.SaveContext(thread->GetContext32());
             cpu_core.SaveContext(thread->GetContext64());
             // Save the TPIDR_EL0 system register in case it was modified.
             thread->SetTPIDR_EL0(cpu_core.GetTPIDR_EL0());
             cpu_core.ClearExclusiveState();
+        } else {
+            prev_thread = nullptr;
         }
-        thread->context_guard.unlock();
     }
 }
 
 void KScheduler::Reload(KThread* thread) {
+    LOG_TRACE(Kernel, "core {}, reload thread {}", core_id, thread ? thread->GetName() : "nullptr");
+
     if (thread) {
         ASSERT_MSG(thread->GetState() == ThreadState::Runnable, "Thread must be runnable.");
 
-        // Cancel any outstanding wakeup events for this thread
-        thread->SetIsRunning(true);
-        thread->SetWasRunning(false);
-
         auto* const thread_owner_process = thread->GetOwnerProcess();
         if (thread_owner_process != nullptr) {
             system.Kernel().MakeCurrentProcess(thread_owner_process);
@@ -676,7 +702,7 @@ void KScheduler::ScheduleImpl() {
     KThread* previous_thread = current_thread;
     current_thread = state.highest_priority_thread;
 
-    this->state.needs_scheduling = false;
+    state.needs_scheduling = false;
 
     if (current_thread == previous_thread) {
         guard.unlock();
@@ -714,7 +740,7 @@ void KScheduler::SwitchToCurrent() {
         {
             std::scoped_lock lock{guard};
             current_thread = state.highest_priority_thread;
-            this->state.needs_scheduling = false;
+            state.needs_scheduling = false;
         }
         const auto is_switch_pending = [this] {
             std::scoped_lock lock{guard};
@@ -722,13 +748,10 @@ void KScheduler::SwitchToCurrent() {
         };
         do {
             if (current_thread != nullptr) {
-                current_thread->context_guard.lock();
                 if (current_thread->GetRawState() != ThreadState::Runnable) {
-                    current_thread->context_guard.unlock();
                     break;
                 }
-                if (static_cast<u32>(current_thread->GetProcessorID()) != core_id) {
-                    current_thread->context_guard.unlock();
+                if (static_cast<u32>(current_thread->GetActiveCore()) != core_id) {
                     break;
                 }
             }
@@ -749,7 +772,7 @@ void KScheduler::UpdateLastContextSwitchTime(KThread* thread, Process* process)
     const u64 update_ticks = most_recent_switch_ticks - prev_switch_ticks;
 
     if (thread != nullptr) {
-        thread->UpdateCPUTimeTicks(update_ticks);
+        thread->AddCpuTime(core_id, update_ticks);
     }
 
     if (process != nullptr) {
@@ -763,15 +786,10 @@ void KScheduler::Initialize() {
     std::string name = "Idle Thread Id:" + std::to_string(core_id);
     std::function<void(void*)> init_func = Core::CpuManager::GetIdleThreadStartFunc();
     void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater();
-    auto thread_res = KThread::Create(system, ThreadType::Kernel, name, 0,
+    auto thread_res = KThread::Create(system, ThreadType::Main, name, 0,
                                       KThread::IdleThreadPriority, 0, static_cast<u32>(core_id), 0,
                                       nullptr, std::move(init_func), init_func_parameter);
     idle_thread = thread_res.Unwrap().get();
-
-    {
-        KScopedSchedulerLock lock{system.Kernel()};
-        idle_thread->SetState(ThreadState::Runnable);
-    }
 }
 
 KScopedSchedulerLock::KScopedSchedulerLock(KernelCore& kernel)
diff --git a/src/core/hle/kernel/k_scheduler.h b/src/core/hle/kernel/k_scheduler.h
index 157373934..2308a55be 100644
--- a/src/core/hle/kernel/k_scheduler.h
+++ b/src/core/hle/kernel/k_scheduler.h
@@ -33,15 +33,14 @@ class KThread;
 
 class KScheduler final {
 public:
-    explicit KScheduler(Core::System& system, std::size_t core_id);
+    explicit KScheduler(Core::System& system, s32 core_id);
     ~KScheduler();
 
     /// Reschedules to the next available thread (call after current thread is suspended)
     void RescheduleCurrentCore();
 
     /// Reschedules cores pending reschedule, to be called on EnableScheduling.
-    static void RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedule,
-                                Core::EmuThreadHandle global_thread);
+    static void RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedule);
 
     /// The next two are for SingleCore Only.
     /// Unload current thread before preempting core.
@@ -53,6 +52,11 @@ public:
     /// Gets the current running thread
     [[nodiscard]] KThread* GetCurrentThread() const;
 
+    /// Returns true if the scheduler is idle
+    [[nodiscard]] bool IsIdle() const {
+        return GetCurrentThread() == idle_thread;
+    }
+
     /// Gets the timestamp for the last context switch in ticks.
     [[nodiscard]] u64 GetLastContextSwitchTicks() const;
 
@@ -79,7 +83,7 @@ public:
      *
      * @note This operation can be redundant and no scheduling is changed if marked as so.
      */
-    void YieldWithoutCoreMigration();
+    static void YieldWithoutCoreMigration(KernelCore& kernel);
 
     /**
      * Takes a thread and moves it to the back of the it's priority list.
@@ -88,7 +92,7 @@ public:
      *
      * @note This operation can be redundant and no scheduling is changed if marked as so.
      */
-    void YieldWithCoreMigration();
+    static void YieldWithCoreMigration(KernelCore& kernel);
 
     /**
      * Takes a thread and moves it out of the scheduling queue.
@@ -97,7 +101,9 @@ public:
      *
      * @note This operation can be redundant and no scheduling is changed if marked as so.
      */
-    void YieldToAnyThread();
+    static void YieldToAnyThread(KernelCore& kernel);
+
+    static void ClearPreviousThread(KernelCore& kernel, KThread* thread);
 
     /// Notify the scheduler a thread's status has changed.
     static void OnThreadStateChanged(KernelCore& kernel, KThread* thread, ThreadState old_state);
@@ -114,8 +120,7 @@ public:
     static void SetSchedulerUpdateNeeded(KernelCore& kernel);
     static void ClearSchedulerUpdateNeeded(KernelCore& kernel);
     static void DisableScheduling(KernelCore& kernel);
-    static void EnableScheduling(KernelCore& kernel, u64 cores_needing_scheduling,
-                                 Core::EmuThreadHandle global_thread);
+    static void EnableScheduling(KernelCore& kernel, u64 cores_needing_scheduling);
     [[nodiscard]] static u64 UpdateHighestPriorityThreads(KernelCore& kernel);
 
 private:
@@ -168,6 +173,7 @@ private:
     static void OnSwitch(void* this_scheduler);
     void SwitchToCurrent();
 
+    KThread* prev_thread{};
     KThread* current_thread{};
     KThread* idle_thread{};
 
@@ -186,7 +192,7 @@ private:
 
     Core::System& system;
     u64 last_context_switch_time{};
-    const std::size_t core_id;
+    const s32 core_id;
 
     Common::SpinLock guard{};
 };
diff --git a/src/core/hle/kernel/k_thread.cpp b/src/core/hle/kernel/k_thread.cpp
index e5be849bb..f021b0550 100644
--- a/src/core/hle/kernel/k_thread.cpp
+++ b/src/core/hle/kernel/k_thread.cpp
@@ -1,4 +1,4 @@
-// Copyright 2014 Citra Emulator Project / PPSSPP Project
+// Copyright 2021 yuzu Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
@@ -8,10 +8,12 @@
 #include <vector>
 
 #include "common/assert.h"
+#include "common/bit_util.h"
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/fiber.h"
 #include "common/logging/log.h"
+#include "common/scope_exit.h"
 #include "common/thread_queue_list.h"
 #include "core/core.h"
 #include "core/cpu_manager.h"
@@ -22,10 +24,12 @@
 #include "core/hle/kernel/k_scheduler.h"
 #include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h"
 #include "core/hle/kernel/k_thread.h"
+#include "core/hle/kernel/k_thread_queue.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/memory/memory_layout.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
+#include "core/hle/kernel/resource_limit.h"
 #include "core/hle/kernel/svc_results.h"
 #include "core/hle/kernel/time_manager.h"
 #include "core/hle/result.h"
@@ -36,185 +40,734 @@
 #include "core/arm/dynarmic/arm_dynarmic_64.h"
 #endif
 
-namespace Kernel {
+namespace {
+static void ResetThreadContext32(Core::ARM_Interface::ThreadContext32& context, u32 stack_top,
+                                 u32 entry_point, u32 arg) {
+    context = {};
+    context.cpu_registers[0] = arg;
+    context.cpu_registers[15] = entry_point;
+    context.cpu_registers[13] = stack_top;
+}
 
-bool KThread::IsSignaled() const {
-    return signaled;
+static void ResetThreadContext64(Core::ARM_Interface::ThreadContext64& context, VAddr stack_top,
+                                 VAddr entry_point, u64 arg) {
+    context = {};
+    context.cpu_registers[0] = arg;
+    context.pc = entry_point;
+    context.sp = stack_top;
+    // TODO(merry): Perform a hardware test to determine the below value.
+    context.fpcr = 0;
 }
+} // namespace
+
+namespace Kernel {
 
-KThread::KThread(KernelCore& kernel) : KSynchronizationObject{kernel} {}
+KThread::KThread(KernelCore& kernel)
+    : KSynchronizationObject{kernel}, activity_pause_lock{kernel} {}
 KThread::~KThread() = default;
 
-void KThread::Stop() {
-    {
-        KScopedSchedulerLock lock(kernel);
-        SetState(ThreadState::Terminated);
-        signaled = true;
-        NotifyAvailable();
-        kernel.GlobalHandleTable().Close(global_handle);
+ResultCode KThread::Initialize(KThreadFunction func, uintptr_t arg, VAddr user_stack_top, s32 prio,
+                               s32 virt_core, Process* owner, ThreadType type) {
+    // Assert parameters are valid.
+    ASSERT((type == ThreadType::Main) ||
+           (Svc::HighestThreadPriority <= prio && prio <= Svc::LowestThreadPriority));
+    ASSERT((owner != nullptr) || (type != ThreadType::User));
+    ASSERT(0 <= virt_core && virt_core < static_cast<s32>(Common::BitSize<u64>()));
+
+    // Convert the virtual core to a physical core.
+    const s32 phys_core = Core::Hardware::VirtualToPhysicalCoreMap[virt_core];
+    ASSERT(0 <= phys_core && phys_core < static_cast<s32>(Core::Hardware::NUM_CPU_CORES));
+
+    // First, clear the TLS address.
+    tls_address = {};
+
+    // Next, assert things based on the type.
+    switch (type) {
+    case ThreadType::Main:
+        ASSERT(arg == 0);
+        [[fallthrough]];
+    case ThreadType::HighPriority:
+        [[fallthrough]];
+    case ThreadType::User:
+        ASSERT(((owner == nullptr) ||
+                (owner->GetCoreMask() | (1ULL << virt_core)) == owner->GetCoreMask()));
+        ASSERT(((owner == nullptr) ||
+                (owner->GetPriorityMask() | (1ULL << prio)) == owner->GetPriorityMask()));
+        break;
+    case ThreadType::Kernel:
+        UNIMPLEMENTED();
+        break;
+    default:
+        UNREACHABLE_MSG("KThread::Initialize: Unknown ThreadType {}", static_cast<u32>(type));
+        break;
+    }
+
+    // Set the ideal core ID and affinity mask.
+    virtual_ideal_core_id = virt_core;
+    physical_ideal_core_id = phys_core;
+    virtual_affinity_mask = (static_cast<u64>(1) << virt_core);
+    physical_affinity_mask.SetAffinity(phys_core, true);
+
+    // Set the thread state.
+    thread_state = (type == ThreadType::Main) ? ThreadState::Runnable : ThreadState::Initialized;
+
+    // Set TLS address.
+    tls_address = 0;
+
+    // Set parent and condvar tree.
+    parent = nullptr;
+    condvar_tree = nullptr;
+
+    // Set sync booleans.
+    signaled = false;
+    termination_requested = false;
+    wait_cancelled = false;
+    cancellable = false;
+
+    // Set core ID and wait result.
+    core_id = phys_core;
+    wait_result = Svc::ResultNoSynchronizationObject;
+
+    // Set priorities.
+    priority = prio;
+    base_priority = prio;
+
+    // Set sync object and waiting lock to null.
+    synced_object = nullptr;
+
+    // Initialize sleeping queue.
+    sleeping_queue = nullptr;
+
+    // Set suspend flags.
+    suspend_request_flags = 0;
+    suspend_allowed_flags = static_cast<u32>(ThreadState::SuspendFlagMask);
+
+    // We're neither debug attached, nor are we nesting our priority inheritance.
+    debug_attached = false;
+    priority_inheritance_count = 0;
+
+    // We haven't been scheduled, and we have done no light IPC.
+    schedule_count = -1;
+    last_scheduled_tick = 0;
+    light_ipc_data = nullptr;
+
+    // We're not waiting for a lock, and we haven't disabled migration.
+    lock_owner = nullptr;
+    num_core_migration_disables = 0;
+
+    // We have no waiters, but we do have an entrypoint.
+    num_kernel_waiters = 0;
+
+    // Set our current core id.
+    current_core_id = phys_core;
+
+    // We haven't released our resource limit hint, and we've spent no time on the cpu.
+    resource_limit_release_hint = false;
+    cpu_time = 0;
+
+    // Clear our stack parameters.
+    std::memset(static_cast<void*>(std::addressof(GetStackParameters())), 0,
+                sizeof(StackParameters));
+
+    // Setup the TLS, if needed.
+    if (type == ThreadType::User) {
+        tls_address = owner->CreateTLSRegion();
+    }
+
+    // Set parent, if relevant.
+    if (owner != nullptr) {
+        parent = owner;
+        parent->IncrementThreadCount();
+    }
 
-        if (owner_process) {
-            owner_process->UnregisterThread(this);
+    // Initialize thread context.
+    ResetThreadContext64(thread_context_64, user_stack_top, func, arg);
+    ResetThreadContext32(thread_context_32, static_cast<u32>(user_stack_top),
+                         static_cast<u32>(func), static_cast<u32>(arg));
 
-            // Mark the TLS slot in the thread's page as free.
-            owner_process->FreeTLSRegion(tls_address);
+    // Setup the stack parameters.
+    StackParameters& sp = GetStackParameters();
+    sp.cur_thread = this;
+    sp.disable_count = 1;
+    SetInExceptionHandler();
+
+    // Set thread ID.
+    thread_id = kernel.CreateNewThreadID();
+
+    // We initialized!
+    initialized = true;
+
+    // Register ourselves with our parent process.
+    if (parent != nullptr) {
+        parent->RegisterThread(this);
+        if (parent->IsSuspended()) {
+            RequestSuspend(SuspendType::Process);
         }
-        has_exited = true;
     }
-    global_handle = 0;
-}
 
-void KThread::Wakeup() {
-    KScopedSchedulerLock lock(kernel);
-    SetState(ThreadState::Runnable);
+    return RESULT_SUCCESS;
 }
 
-ResultCode KThread::Start() {
-    KScopedSchedulerLock lock(kernel);
-    SetState(ThreadState::Runnable);
+ResultCode KThread::InitializeThread(KThread* thread, KThreadFunction func, uintptr_t arg,
+                                     VAddr user_stack_top, s32 prio, s32 core, Process* owner,
+                                     ThreadType type) {
+    // Initialize the thread.
+    R_TRY(thread->Initialize(func, arg, user_stack_top, prio, core, owner, type));
+
     return RESULT_SUCCESS;
 }
 
-void KThread::CancelWait() {
-    KScopedSchedulerLock lock(kernel);
-    if (GetState() != ThreadState::Waiting || !is_cancellable) {
-        is_sync_cancelled = true;
-        return;
+void KThread::Finalize() {
+    // If the thread has an owner process, unregister it.
+    if (parent != nullptr) {
+        parent->UnregisterThread(this);
+    }
+
+    // If the thread has a local region, delete it.
+    if (tls_address != 0) {
+        parent->FreeTLSRegion(tls_address);
+    }
+
+    // Release any waiters.
+    {
+        ASSERT(lock_owner == nullptr);
+        KScopedSchedulerLock sl{kernel};
+
+        auto it = waiter_list.begin();
+        while (it != waiter_list.end()) {
+            // The thread shouldn't be a kernel waiter.
+            it->SetLockOwner(nullptr);
+            it->SetSyncedObject(nullptr, Svc::ResultInvalidState);
+            it->Wakeup();
+            it = waiter_list.erase(it);
+        }
+    }
+
+    // Decrement the parent process's thread count.
+    if (parent != nullptr) {
+        parent->DecrementThreadCount();
     }
-    // TODO(Blinkhawk): Implement cancel of server session
-    is_sync_cancelled = false;
-    SetSynchronizationResults(nullptr, ERR_SYNCHRONIZATION_CANCELED);
-    SetState(ThreadState::Runnable);
 }
 
-static void ResetThreadContext32(Core::ARM_Interface::ThreadContext32& context, u32 stack_top,
-                                 u32 entry_point, u32 arg) {
-    context = {};
-    context.cpu_registers[0] = arg;
-    context.cpu_registers[15] = entry_point;
-    context.cpu_registers[13] = stack_top;
+bool KThread::IsSignaled() const {
+    return signaled;
 }
 
-static void ResetThreadContext64(Core::ARM_Interface::ThreadContext64& context, VAddr stack_top,
-                                 VAddr entry_point, u64 arg) {
-    context = {};
-    context.cpu_registers[0] = arg;
-    context.pc = entry_point;
-    context.sp = stack_top;
-    // TODO(merry): Perform a hardware test to determine the below value.
-    context.fpcr = 0;
+void KThread::Wakeup() {
+    KScopedSchedulerLock sl{kernel};
+
+    if (GetState() == ThreadState::Waiting) {
+        if (sleeping_queue != nullptr) {
+            sleeping_queue->WakeupThread(this);
+        } else {
+            SetState(ThreadState::Runnable);
+        }
+    }
 }
 
-std::shared_ptr<Common::Fiber>& KThread::GetHostContext() {
-    return host_context;
+void KThread::StartTermination() {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+
+    // Release user exception and unpin, if relevant.
+    if (parent != nullptr) {
+        parent->ReleaseUserException(this);
+        if (parent->GetPinnedThread(GetCurrentCoreId(kernel)) == this) {
+            parent->UnpinCurrentThread();
+        }
+    }
+
+    // Set state to terminated.
+    SetState(ThreadState::Terminated);
+
+    // Clear the thread's status as running in parent.
+    if (parent != nullptr) {
+        parent->ClearRunningThread(this);
+    }
+
+    // Signal.
+    signaled = true;
+    NotifyAvailable();
+
+    // Clear previous thread in KScheduler.
+    KScheduler::ClearPreviousThread(kernel, this);
+
+    // Register terminated dpc flag.
+    RegisterDpc(DpcFlag::Terminated);
 }
 
-ResultVal<std::shared_ptr<KThread>> KThread::Create(Core::System& system, ThreadType type_flags,
-                                                    std::string name, VAddr entry_point,
-                                                    u32 priority, u64 arg, s32 processor_id,
-                                                    VAddr stack_top, Process* owner_process) {
-    std::function<void(void*)> init_func = Core::CpuManager::GetGuestThreadStartFunc();
-    void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater();
-    return Create(system, type_flags, name, entry_point, priority, arg, processor_id, stack_top,
-                  owner_process, std::move(init_func), init_func_parameter);
+void KThread::Pin() {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+
+    // Set ourselves as pinned.
+    GetStackParameters().is_pinned = true;
+
+    // Disable core migration.
+    ASSERT(num_core_migration_disables == 0);
+    {
+        ++num_core_migration_disables;
+
+        // Save our ideal state to restore when we're unpinned.
+        original_physical_ideal_core_id = physical_ideal_core_id;
+        original_physical_affinity_mask = physical_affinity_mask;
+
+        // Bind ourselves to this core.
+        const s32 active_core = GetActiveCore();
+        const s32 current_core = GetCurrentCoreId(kernel);
+
+        SetActiveCore(current_core);
+        physical_ideal_core_id = current_core;
+        physical_affinity_mask.SetAffinityMask(1ULL << current_core);
+
+        if (active_core != current_core || physical_affinity_mask.GetAffinityMask() !=
+                                               original_physical_affinity_mask.GetAffinityMask()) {
+            KScheduler::OnThreadAffinityMaskChanged(kernel, this, original_physical_affinity_mask,
+                                                    active_core);
+        }
+    }
+
+    // Disallow performing thread suspension.
+    {
+        // Update our allow flags.
+        suspend_allowed_flags &= ~(1 << (static_cast<u32>(SuspendType::Thread) +
+                                         static_cast<u32>(ThreadState::SuspendShift)));
+
+        // Update our state.
+        const ThreadState old_state = thread_state;
+        thread_state = static_cast<ThreadState>(GetSuspendFlags() |
+                                                static_cast<u32>(old_state & ThreadState::Mask));
+        if (thread_state != old_state) {
+            KScheduler::OnThreadStateChanged(kernel, this, old_state);
+        }
+    }
+
+    // TODO(bunnei): Update our SVC access permissions.
+    ASSERT(parent != nullptr);
 }
 
-ResultVal<std::shared_ptr<KThread>> KThread::Create(Core::System& system, ThreadType type_flags,
-                                                    std::string name, VAddr entry_point,
-                                                    u32 priority, u64 arg, s32 processor_id,
-                                                    VAddr stack_top, Process* owner_process,
-                                                    std::function<void(void*)>&& thread_start_func,
-                                                    void* thread_start_parameter) {
-    auto& kernel = system.Kernel();
+void KThread::Unpin() {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+
+    // Set ourselves as unpinned.
+    GetStackParameters().is_pinned = false;
+
+    // Enable core migration.
+    ASSERT(num_core_migration_disables == 1);
+    {
+        --num_core_migration_disables;
+
+        // Restore our original state.
+        const KAffinityMask old_mask = physical_affinity_mask;
 
-    if (owner_process) {
-        if (!system.Memory().IsValidVirtualAddress(*owner_process, entry_point)) {
-            LOG_ERROR(Kernel_SVC, "(name={}): invalid entry {:016X}", name, entry_point);
-            // TODO (bunnei): Find the correct error code to use here
-            return RESULT_UNKNOWN;
+        physical_ideal_core_id = original_physical_ideal_core_id;
+        physical_affinity_mask = original_physical_affinity_mask;
+
+        if (physical_affinity_mask.GetAffinityMask() != old_mask.GetAffinityMask()) {
+            const s32 active_core = GetActiveCore();
+
+            if (!physical_affinity_mask.GetAffinity(active_core)) {
+                if (physical_ideal_core_id >= 0) {
+                    SetActiveCore(physical_ideal_core_id);
+                } else {
+                    SetActiveCore(static_cast<s32>(
+                        Common::BitSize<u64>() - 1 -
+                        std::countl_zero(physical_affinity_mask.GetAffinityMask())));
+                }
+            }
+            KScheduler::OnThreadAffinityMaskChanged(kernel, this, old_mask, active_core);
         }
     }
 
-    std::shared_ptr<KThread> thread = std::make_shared<KThread>(kernel);
+    // Allow performing thread suspension (if termination hasn't been requested).
+    {
+        // Update our allow flags.
+        if (!IsTerminationRequested()) {
+            suspend_allowed_flags |= (1 << (static_cast<u32>(SuspendType::Thread) +
+                                            static_cast<u32>(ThreadState::SuspendShift)));
+        }
 
-    thread->thread_id = kernel.CreateNewThreadID();
-    thread->thread_state = ThreadState::Initialized;
-    thread->entry_point = entry_point;
-    thread->stack_top = stack_top;
-    thread->disable_count = 1;
-    thread->tpidr_el0 = 0;
-    thread->current_priority = priority;
-    thread->base_priority = priority;
-    thread->lock_owner = nullptr;
-    thread->schedule_count = -1;
-    thread->last_scheduled_tick = 0;
-    thread->processor_id = processor_id;
-    thread->ideal_core = processor_id;
-    thread->affinity_mask.SetAffinity(processor_id, true);
-    thread->name = std::move(name);
-    thread->global_handle = kernel.GlobalHandleTable().Create(thread).Unwrap();
-    thread->owner_process = owner_process;
-    thread->type = type_flags;
-    thread->signaled = false;
+        // Update our state.
+        const ThreadState old_state = thread_state;
+        thread_state = static_cast<ThreadState>(GetSuspendFlags() |
+                                                static_cast<u32>(old_state & ThreadState::Mask));
+        if (thread_state != old_state) {
+            KScheduler::OnThreadStateChanged(kernel, this, old_state);
+        }
+    }
 
-    auto& scheduler = kernel.GlobalSchedulerContext();
-    scheduler.AddThread(thread);
+    // TODO(bunnei): Update our SVC access permissions.
+    ASSERT(parent != nullptr);
+
+    // Resume any threads that began waiting on us while we were pinned.
+    for (auto it = pinned_waiter_list.begin(); it != pinned_waiter_list.end(); ++it) {
+        if (it->GetState() == ThreadState::Waiting) {
+            it->SetState(ThreadState::Runnable);
+        }
+    }
+}
+
+ResultCode KThread::GetCoreMask(s32* out_ideal_core, u64* out_affinity_mask) {
+    KScopedSchedulerLock sl{kernel};
+
+    // Get the virtual mask.
+    *out_ideal_core = virtual_ideal_core_id;
+    *out_affinity_mask = virtual_affinity_mask;
+
+    return RESULT_SUCCESS;
+}
 
-    if (owner_process) {
-        thread->tls_address = thread->owner_process->CreateTLSRegion();
-        thread->owner_process->RegisterThread(thread.get());
+ResultCode KThread::GetPhysicalCoreMask(s32* out_ideal_core, u64* out_affinity_mask) {
+    KScopedSchedulerLock sl{kernel};
+    ASSERT(num_core_migration_disables >= 0);
+
+    // Select between core mask and original core mask.
+    if (num_core_migration_disables == 0) {
+        *out_ideal_core = physical_ideal_core_id;
+        *out_affinity_mask = physical_affinity_mask.GetAffinityMask();
     } else {
-        thread->tls_address = 0;
+        *out_ideal_core = original_physical_ideal_core_id;
+        *out_affinity_mask = original_physical_affinity_mask.GetAffinityMask();
     }
 
-    ResetThreadContext32(thread->context_32, static_cast<u32>(stack_top),
-                         static_cast<u32>(entry_point), static_cast<u32>(arg));
-    ResetThreadContext64(thread->context_64, stack_top, entry_point, arg);
+    return RESULT_SUCCESS;
+}
 
-    thread->host_context =
-        std::make_shared<Common::Fiber>(std::move(thread_start_func), thread_start_parameter);
+ResultCode KThread::SetCoreMask(s32 core_id, u64 v_affinity_mask) {
+    ASSERT(parent != nullptr);
+    ASSERT(v_affinity_mask != 0);
+    KScopedLightLock lk{activity_pause_lock};
 
-    return MakeResult<std::shared_ptr<KThread>>(std::move(thread));
+    // Set the core mask.
+    u64 p_affinity_mask = 0;
+    {
+        KScopedSchedulerLock sl{kernel};
+        ASSERT(num_core_migration_disables >= 0);
+
+        // If the core id is no-update magic, preserve the ideal core id.
+        if (core_id == Svc::IdealCoreNoUpdate) {
+            core_id = virtual_ideal_core_id;
+            R_UNLESS(((1ULL << core_id) & v_affinity_mask) != 0, Svc::ResultInvalidCombination);
+        }
+
+        // Set the virtual core/affinity mask.
+        virtual_ideal_core_id = core_id;
+        virtual_affinity_mask = v_affinity_mask;
+
+        // Translate the virtual core to a physical core.
+        if (core_id >= 0) {
+            core_id = Core::Hardware::VirtualToPhysicalCoreMap[core_id];
+        }
+
+        // Translate the virtual affinity mask to a physical one.
+        while (v_affinity_mask != 0) {
+            const u64 next = std::countr_zero(v_affinity_mask);
+            v_affinity_mask &= ~(1ULL << next);
+            p_affinity_mask |= (1ULL << Core::Hardware::VirtualToPhysicalCoreMap[next]);
+        }
+
+        // If we haven't disabled migration, perform an affinity change.
+        if (num_core_migration_disables == 0) {
+            const KAffinityMask old_mask = physical_affinity_mask;
+
+            // Set our new ideals.
+            physical_ideal_core_id = core_id;
+            physical_affinity_mask.SetAffinityMask(p_affinity_mask);
+
+            if (physical_affinity_mask.GetAffinityMask() != old_mask.GetAffinityMask()) {
+                const s32 active_core = GetActiveCore();
+
+                if (active_core >= 0 && !physical_affinity_mask.GetAffinity(active_core)) {
+                    const s32 new_core = static_cast<s32>(
+                        physical_ideal_core_id >= 0
+                            ? physical_ideal_core_id
+                            : Common::BitSize<u64>() - 1 -
+                                  std::countl_zero(physical_affinity_mask.GetAffinityMask()));
+                    SetActiveCore(new_core);
+                }
+                KScheduler::OnThreadAffinityMaskChanged(kernel, this, old_mask, active_core);
+            }
+        } else {
+            // Otherwise, we edit the original affinity for restoration later.
+            original_physical_ideal_core_id = core_id;
+            original_physical_affinity_mask.SetAffinityMask(p_affinity_mask);
+        }
+    }
+
+    // Update the pinned waiter list.
+    {
+        bool retry_update = false;
+        bool thread_is_pinned = false;
+        do {
+            // Lock the scheduler.
+            KScopedSchedulerLock sl{kernel};
+
+            // Don't do any further management if our termination has been requested.
+            R_SUCCEED_IF(IsTerminationRequested());
+
+            // By default, we won't need to retry.
+            retry_update = false;
+
+            // Check if the thread is currently running.
+            bool thread_is_current = false;
+            s32 thread_core;
+            for (thread_core = 0; thread_core < static_cast<s32>(Core::Hardware::NUM_CPU_CORES);
+                 ++thread_core) {
+                if (kernel.Scheduler(thread_core).GetCurrentThread() == this) {
+                    thread_is_current = true;
+                    break;
+                }
+            }
+
+            // If the thread is currently running, check whether it's no longer allowed under the
+            // new mask.
+            if (thread_is_current && ((1ULL << thread_core) & p_affinity_mask) == 0) {
+                // If the thread is pinned, we want to wait until it's not pinned.
+                if (GetStackParameters().is_pinned) {
+                    // Verify that the current thread isn't terminating.
+                    R_UNLESS(!GetCurrentThread(kernel).IsTerminationRequested(),
+                             Svc::ResultTerminationRequested);
+
+                    // Note that the thread was pinned.
+                    thread_is_pinned = true;
+
+                    // Wait until the thread isn't pinned any more.
+                    pinned_waiter_list.push_back(GetCurrentThread(kernel));
+                    GetCurrentThread(kernel).SetState(ThreadState::Waiting);
+                } else {
+                    // If the thread isn't pinned, release the scheduler lock and retry until it's
+                    // not current.
+                    retry_update = true;
+                }
+            }
+        } while (retry_update);
+
+        // If the thread was pinned, it no longer is, and we should remove the current thread from
+        // our waiter list.
+        if (thread_is_pinned) {
+            // Lock the scheduler.
+            KScopedSchedulerLock sl{kernel};
+
+            // Remove from the list.
+            pinned_waiter_list.erase(pinned_waiter_list.iterator_to(GetCurrentThread(kernel)));
+        }
+    }
+
+    return RESULT_SUCCESS;
 }
 
-void KThread::SetBasePriority(u32 priority) {
-    ASSERT(Svc::HighestThreadPriority <= priority && priority <= Svc::LowestThreadPriority);
+void KThread::SetBasePriority(s32 value) {
+    ASSERT(Svc::HighestThreadPriority <= value && value <= Svc::LowestThreadPriority);
 
-    KScopedSchedulerLock lock(kernel);
+    KScopedSchedulerLock sl{kernel};
 
     // Change our base priority.
-    base_priority = priority;
+    base_priority = value;
 
     // Perform a priority restoration.
     RestorePriority(kernel, this);
 }
 
-void KThread::SetSynchronizationResults(KSynchronizationObject* object, ResultCode result) {
-    signaling_object = object;
-    signaling_result = result;
-}
+void KThread::RequestSuspend(SuspendType type) {
+    KScopedSchedulerLock sl{kernel};
 
-VAddr KThread::GetCommandBufferAddress() const {
-    // Offset from the start of TLS at which the IPC command buffer begins.
-    constexpr u64 command_header_offset = 0x80;
-    return GetTLSAddress() + command_header_offset;
+    // Note the request in our flags.
+    suspend_request_flags |=
+        (1u << (static_cast<u32>(ThreadState::SuspendShift) + static_cast<u32>(type)));
+
+    // Try to perform the suspend.
+    TrySuspend();
 }
 
-void KThread::SetState(ThreadState state) {
-    KScopedSchedulerLock sl(kernel);
+void KThread::Resume(SuspendType type) {
+    KScopedSchedulerLock sl{kernel};
 
-    // Clear debugging state
-    SetMutexWaitAddressForDebugging({});
-    SetWaitReasonForDebugging({});
+    // Clear the request in our flags.
+    suspend_request_flags &=
+        ~(1u << (static_cast<u32>(ThreadState::SuspendShift) + static_cast<u32>(type)));
 
+    // Update our state.
     const ThreadState old_state = thread_state;
-    thread_state =
-        static_cast<ThreadState>((old_state & ~ThreadState::Mask) | (state & ThreadState::Mask));
+    thread_state = static_cast<ThreadState>(GetSuspendFlags() |
+                                            static_cast<u32>(old_state & ThreadState::Mask));
     if (thread_state != old_state) {
         KScheduler::OnThreadStateChanged(kernel, this, old_state);
     }
 }
 
+void KThread::WaitCancel() {
+    KScopedSchedulerLock sl{kernel};
+
+    // Check if we're waiting and cancellable.
+    if (GetState() == ThreadState::Waiting && cancellable) {
+        if (sleeping_queue != nullptr) {
+            sleeping_queue->WakeupThread(this);
+            wait_cancelled = true;
+        } else {
+            SetSyncedObject(nullptr, Svc::ResultCancelled);
+            SetState(ThreadState::Runnable);
+            wait_cancelled = false;
+        }
+    } else {
+        // Otherwise, note that we cancelled a wait.
+        wait_cancelled = true;
+    }
+}
+
+void KThread::TrySuspend() {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+    ASSERT(IsSuspendRequested());
+
+    // Ensure that we have no waiters.
+    if (GetNumKernelWaiters() > 0) {
+        return;
+    }
+    ASSERT(GetNumKernelWaiters() == 0);
+
+    // Perform the suspend.
+    Suspend();
+}
+
+void KThread::Suspend() {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+    ASSERT(IsSuspendRequested());
+
+    // Set our suspend flags in state.
+    const auto old_state = thread_state;
+    thread_state = static_cast<ThreadState>(GetSuspendFlags()) | (old_state & ThreadState::Mask);
+
+    // Note the state change in scheduler.
+    KScheduler::OnThreadStateChanged(kernel, this, old_state);
+}
+
+void KThread::Continue() {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+
+    // Clear our suspend flags in state.
+    const auto old_state = thread_state;
+    thread_state = old_state & ThreadState::Mask;
+
+    // Note the state change in scheduler.
+    KScheduler::OnThreadStateChanged(kernel, this, old_state);
+}
+
+ResultCode KThread::SetActivity(Svc::ThreadActivity activity) {
+    // Lock ourselves.
+    KScopedLightLock lk(activity_pause_lock);
+
+    // Set the activity.
+    {
+        // Lock the scheduler.
+        KScopedSchedulerLock sl{kernel};
+
+        // Verify our state.
+        const auto cur_state = GetState();
+        R_UNLESS((cur_state == ThreadState::Waiting || cur_state == ThreadState::Runnable),
+                 Svc::ResultInvalidState);
+
+        // Either pause or resume.
+        if (activity == Svc::ThreadActivity::Paused) {
+            // Verify that we're not suspended.
+            R_UNLESS(!IsSuspendRequested(SuspendType::Thread), Svc::ResultInvalidState);
+
+            // Suspend.
+            RequestSuspend(SuspendType::Thread);
+        } else {
+            ASSERT(activity == Svc::ThreadActivity::Runnable);
+
+            // Verify that we're suspended.
+            R_UNLESS(IsSuspendRequested(SuspendType::Thread), Svc::ResultInvalidState);
+
+            // Resume.
+            Resume(SuspendType::Thread);
+        }
+    }
+
+    // If the thread is now paused, update the pinned waiter list.
+    if (activity == Svc::ThreadActivity::Paused) {
+        bool thread_is_pinned = false;
+        bool thread_is_current;
+        do {
+            // Lock the scheduler.
+            KScopedSchedulerLock sl{kernel};
+
+            // Don't do any further management if our termination has been requested.
+            R_SUCCEED_IF(IsTerminationRequested());
+
+            // Check whether the thread is pinned.
+            if (GetStackParameters().is_pinned) {
+                // Verify that the current thread isn't terminating.
+                R_UNLESS(!GetCurrentThread(kernel).IsTerminationRequested(),
+                         Svc::ResultTerminationRequested);
+
+                // Note that the thread was pinned and not current.
+                thread_is_pinned = true;
+                thread_is_current = false;
+
+                // Wait until the thread isn't pinned any more.
+                pinned_waiter_list.push_back(GetCurrentThread(kernel));
+                GetCurrentThread(kernel).SetState(ThreadState::Waiting);
+            } else {
+                // Check if the thread is currently running.
+                // If it is, we'll need to retry.
+                thread_is_current = false;
+
+                for (auto i = 0; i < static_cast<s32>(Core::Hardware::NUM_CPU_CORES); ++i) {
+                    if (kernel.Scheduler(i).GetCurrentThread() == this) {
+                        thread_is_current = true;
+                        break;
+                    }
+                }
+            }
+        } while (thread_is_current);
+
+        // If the thread was pinned, it no longer is, and we should remove the current thread from
+        // our waiter list.
+        if (thread_is_pinned) {
+            // Lock the scheduler.
+            KScopedSchedulerLock sl{kernel};
+
+            // Remove from the list.
+            pinned_waiter_list.erase(pinned_waiter_list.iterator_to(GetCurrentThread(kernel)));
+        }
+    }
+
+    return RESULT_SUCCESS;
+}
+
+ResultCode KThread::GetThreadContext3(std::vector<u8>& out) {
+    // Lock ourselves.
+    KScopedLightLock lk{activity_pause_lock};
+
+    // Get the context.
+    {
+        // Lock the scheduler.
+        KScopedSchedulerLock sl{kernel};
+
+        // Verify that we're suspended.
+        R_UNLESS(IsSuspendRequested(SuspendType::Thread), Svc::ResultInvalidState);
+
+        // If we're not terminating, get the thread's user context.
+        if (!IsTerminationRequested()) {
+            if (parent->Is64BitProcess()) {
+                // Mask away mode bits, interrupt bits, IL bit, and other reserved bits.
+                auto context = GetContext64();
+                context.pstate &= 0xFF0FFE20;
+
+                out.resize(sizeof(context));
+                std::memcpy(out.data(), &context, sizeof(context));
+            } else {
+                // Mask away mode bits, interrupt bits, IL bit, and other reserved bits.
+                auto context = GetContext32();
+                context.cpsr &= 0xFF0FFE20;
+
+                out.resize(sizeof(context));
+                std::memcpy(out.data(), &context, sizeof(context));
+            }
+        }
+    }
+
+    return RESULT_SUCCESS;
+}
+
 void KThread::AddWaiterImpl(KThread* thread) {
     ASSERT(kernel.GlobalSchedulerContext().IsLocked());
 
@@ -345,104 +898,150 @@ KThread* KThread::RemoveWaiterByKey(s32* out_num_waiters, VAddr key) {
     return next_lock_owner;
 }
 
-ResultCode KThread::SetActivity(ThreadActivity value) {
-    KScopedSchedulerLock lock(kernel);
+ResultCode KThread::Run() {
+    while (true) {
+        KScopedSchedulerLock lk{kernel};
 
-    auto sched_status = GetState();
+        // If either this thread or the current thread are requesting termination, note it.
+        R_UNLESS(!IsTerminationRequested(), Svc::ResultTerminationRequested);
+        R_UNLESS(!GetCurrentThread(kernel).IsTerminationRequested(),
+                 Svc::ResultTerminationRequested);
 
-    if (sched_status != ThreadState::Runnable && sched_status != ThreadState::Waiting) {
-        return ERR_INVALID_STATE;
-    }
+        // Ensure our thread state is correct.
+        R_UNLESS(GetState() == ThreadState::Initialized, Svc::ResultInvalidState);
 
-    if (IsTerminationRequested()) {
+        // If the current thread has been asked to suspend, suspend it and retry.
+        if (GetCurrentThread(kernel).IsSuspended()) {
+            GetCurrentThread(kernel).Suspend();
+            continue;
+        }
+
+        // If we're not a kernel thread and we've been asked to suspend, suspend ourselves.
+        if (IsUserThread() && IsSuspended()) {
+            Suspend();
+        }
+
+        // Set our state and finish.
+        SetState(ThreadState::Runnable);
         return RESULT_SUCCESS;
     }
+}
 
-    if (value == ThreadActivity::Paused) {
-        if ((pausing_state & static_cast<u32>(ThreadSchedFlags::ThreadPauseFlag)) != 0) {
-            return ERR_INVALID_STATE;
-        }
-        AddSchedulingFlag(ThreadSchedFlags::ThreadPauseFlag);
-    } else {
-        if ((pausing_state & static_cast<u32>(ThreadSchedFlags::ThreadPauseFlag)) == 0) {
-            return ERR_INVALID_STATE;
-        }
-        RemoveSchedulingFlag(ThreadSchedFlags::ThreadPauseFlag);
+void KThread::Exit() {
+    ASSERT(this == GetCurrentThreadPointer(kernel));
+
+    // Release the thread resource hint from parent.
+    if (parent != nullptr) {
+        // TODO(bunnei): Hint that the resource is about to be released.
+        resource_limit_release_hint = true;
+    }
+
+    // Perform termination.
+    {
+        KScopedSchedulerLock sl{kernel};
+
+        // Disallow all suspension.
+        suspend_allowed_flags = 0;
+
+        // Start termination.
+        StartTermination();
     }
-    return RESULT_SUCCESS;
 }
 
-ResultCode KThread::Sleep(s64 nanoseconds) {
-    Handle event_handle{};
+ResultCode KThread::Sleep(s64 timeout) {
+    ASSERT(!kernel.GlobalSchedulerContext().IsLocked());
+    ASSERT(this == GetCurrentThreadPointer(kernel));
+    ASSERT(timeout > 0);
+
     {
-        KScopedSchedulerLockAndSleep lock(kernel, event_handle, this, nanoseconds);
+        // Setup the scheduling lock and sleep.
+        KScopedSchedulerLockAndSleep slp{kernel, this, timeout};
+
+        // Check if the thread should terminate.
+        if (IsTerminationRequested()) {
+            slp.CancelSleep();
+            return Svc::ResultTerminationRequested;
+        }
+
+        // Mark the thread as waiting.
         SetState(ThreadState::Waiting);
         SetWaitReasonForDebugging(ThreadWaitReasonForDebugging::Sleep);
     }
 
-    if (event_handle != InvalidHandle) {
-        auto& time_manager = kernel.TimeManager();
-        time_manager.UnscheduleTimeEvent(event_handle);
-    }
+    // The lock/sleep is done.
+
+    // Cancel the timer.
+    kernel.TimeManager().UnscheduleTimeEvent(this);
+
     return RESULT_SUCCESS;
 }
 
-void KThread::AddSchedulingFlag(ThreadSchedFlags flag) {
-    const auto old_state = GetRawState();
-    pausing_state |= static_cast<u32>(flag);
-    const auto base_scheduling = GetState();
-    thread_state = base_scheduling | static_cast<ThreadState>(pausing_state);
-    KScheduler::OnThreadStateChanged(kernel, this, old_state);
+void KThread::SetState(ThreadState state) {
+    KScopedSchedulerLock sl{kernel};
+
+    // Clear debugging state
+    SetMutexWaitAddressForDebugging({});
+    SetWaitReasonForDebugging({});
+
+    const ThreadState old_state = thread_state;
+    thread_state =
+        static_cast<ThreadState>((old_state & ~ThreadState::Mask) | (state & ThreadState::Mask));
+    if (thread_state != old_state) {
+        KScheduler::OnThreadStateChanged(kernel, this, old_state);
+    }
 }
 
-void KThread::RemoveSchedulingFlag(ThreadSchedFlags flag) {
-    const auto old_state = GetRawState();
-    pausing_state &= ~static_cast<u32>(flag);
-    const auto base_scheduling = GetState();
-    thread_state = base_scheduling | static_cast<ThreadState>(pausing_state);
-    KScheduler::OnThreadStateChanged(kernel, this, old_state);
+std::shared_ptr<Common::Fiber>& KThread::GetHostContext() {
+    return host_context;
 }
 
-ResultCode KThread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) {
-    KScopedSchedulerLock lock(kernel);
-    const auto HighestSetCore = [](u64 mask, u32 max_cores) {
-        for (s32 core = static_cast<s32>(max_cores - 1); core >= 0; core--) {
-            if (((mask >> core) & 1) != 0) {
-                return core;
-            }
-        }
-        return -1;
-    };
-
-    const bool use_override = affinity_override_count != 0;
-    if (new_core == Svc::IdealCoreNoUpdate) {
-        new_core = use_override ? ideal_core_override : ideal_core;
-        if ((new_affinity_mask & (1ULL << new_core)) == 0) {
-            LOG_ERROR(Kernel, "New affinity mask is incorrect! new_core={}, new_affinity_mask={}",
-                      new_core, new_affinity_mask);
-            return ERR_INVALID_COMBINATION;
-        }
-    }
-    if (use_override) {
-        ideal_core_override = new_core;
-    } else {
-        const auto old_affinity_mask = affinity_mask;
-        affinity_mask.SetAffinityMask(new_affinity_mask);
-        ideal_core = new_core;
-        if (old_affinity_mask.GetAffinityMask() != new_affinity_mask) {
-            const s32 old_core = processor_id;
-            if (processor_id >= 0 && !affinity_mask.GetAffinity(processor_id)) {
-                if (static_cast<s32>(ideal_core) < 0) {
-                    processor_id = HighestSetCore(affinity_mask.GetAffinityMask(),
-                                                  Core::Hardware::NUM_CPU_CORES);
-                } else {
-                    processor_id = ideal_core;
-                }
-            }
-            KScheduler::OnThreadAffinityMaskChanged(kernel, this, old_affinity_mask, old_core);
-        }
+ResultVal<std::shared_ptr<KThread>> KThread::Create(Core::System& system, ThreadType type_flags,
+                                                    std::string name, VAddr entry_point,
+                                                    u32 priority, u64 arg, s32 processor_id,
+                                                    VAddr stack_top, Process* owner_process) {
+    std::function<void(void*)> init_func = Core::CpuManager::GetGuestThreadStartFunc();
+    void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater();
+    return Create(system, type_flags, name, entry_point, priority, arg, processor_id, stack_top,
+                  owner_process, std::move(init_func), init_func_parameter);
+}
+
+ResultVal<std::shared_ptr<KThread>> KThread::Create(Core::System& system, ThreadType type_flags,
+                                                    std::string name, VAddr entry_point,
+                                                    u32 priority, u64 arg, s32 processor_id,
+                                                    VAddr stack_top, Process* owner_process,
+                                                    std::function<void(void*)>&& thread_start_func,
+                                                    void* thread_start_parameter) {
+    auto& kernel = system.Kernel();
+
+    std::shared_ptr<KThread> thread = std::make_shared<KThread>(kernel);
+
+    thread->InitializeThread(thread.get(), entry_point, arg, stack_top, priority, processor_id,
+                             owner_process, type_flags);
+    thread->name = name;
+
+    auto& scheduler = kernel.GlobalSchedulerContext();
+    scheduler.AddThread(thread);
+
+    thread->host_context =
+        std::make_shared<Common::Fiber>(std::move(thread_start_func), thread_start_parameter);
+
+    return MakeResult<std::shared_ptr<KThread>>(std::move(thread));
+}
+
+KThread* GetCurrentThreadPointer(KernelCore& kernel) {
+    if (!kernel.CurrentScheduler()) {
+        // We are not called from a core thread
+        return {};
     }
-    return RESULT_SUCCESS;
+    return kernel.CurrentScheduler()->GetCurrentThread();
+}
+
+KThread& GetCurrentThread(KernelCore& kernel) {
+    return *GetCurrentThreadPointer(kernel);
+}
+
+s32 GetCurrentCoreId(KernelCore& kernel) {
+    return GetCurrentThread(kernel).GetCurrentCore();
 }
 
 } // namespace Kernel
diff --git a/src/core/hle/kernel/k_thread.h b/src/core/hle/kernel/k_thread.h
index ef2313f87..7845821ba 100644
--- a/src/core/hle/kernel/k_thread.h
+++ b/src/core/hle/kernel/k_thread.h
@@ -1,11 +1,10 @@
-// Copyright 2014 Citra Emulator Project / PPSSPP Project
+// Copyright 2021 yuzu Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
 #pragma once
 
 #include <array>
-#include <functional>
 #include <span>
 #include <string>
 #include <utility>
@@ -18,9 +17,11 @@
 #include "common/spin_lock.h"
 #include "core/arm/arm_interface.h"
 #include "core/hle/kernel/k_affinity_mask.h"
+#include "core/hle/kernel/k_light_lock.h"
 #include "core/hle/kernel/k_synchronization_object.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/svc_common.h"
+#include "core/hle/kernel/svc_types.h"
 #include "core/hle/result.h"
 
 namespace Common {
@@ -38,6 +39,9 @@ class GlobalSchedulerContext;
 class KernelCore;
 class Process;
 class KScheduler;
+class KThreadQueue;
+
+using KThreadFunction = VAddr;
 
 enum class ThreadType : u32 {
     Main = 0,
@@ -47,6 +51,16 @@ enum class ThreadType : u32 {
 };
 DECLARE_ENUM_FLAG_OPERATORS(ThreadType);
 
+enum class SuspendType : u32 {
+    Process = 0,
+    Thread = 1,
+    Debug = 2,
+    Backtrace = 3,
+    Init = 4,
+
+    Count,
+};
+
 enum class ThreadState : u16 {
     Initialized = 0,
     Waiting = 1,
@@ -66,21 +80,9 @@ enum class ThreadState : u16 {
 };
 DECLARE_ENUM_FLAG_OPERATORS(ThreadState);
 
-enum class ThreadWakeupReason {
-    Signal, // The thread was woken up by WakeupAllWaitingThreads due to an object signal.
-    Timeout // The thread was woken up due to a wait timeout.
-};
-
-enum class ThreadActivity : u32 {
-    Normal = 0,
-    Paused = 1,
-};
-
-enum class ThreadSchedFlags : u32 {
-    ProcessPauseFlag = 1 << 4,
-    ThreadPauseFlag = 1 << 5,
-    ProcessDebugPauseFlag = 1 << 6,
-    KernelInitPauseFlag = 1 << 8,
+enum class DpcFlag : u32 {
+    Terminating = (1 << 0),
+    Terminated = (1 << 1),
 };
 
 enum class ThreadWaitReasonForDebugging : u32 {
@@ -93,21 +95,25 @@ enum class ThreadWaitReasonForDebugging : u32 {
     Suspended,       ///< Thread is waiting due to process suspension
 };
 
+[[nodiscard]] KThread* GetCurrentThreadPointer(KernelCore& kernel);
+[[nodiscard]] KThread& GetCurrentThread(KernelCore& kernel);
+[[nodiscard]] s32 GetCurrentCoreId(KernelCore& kernel);
+
 class KThread final : public KSynchronizationObject, public boost::intrusive::list_base_hook<> {
     friend class KScheduler;
     friend class Process;
 
 public:
     static constexpr s32 DefaultThreadPriority = 44;
-    static constexpr s32 IdleThreadPriority = 64;
+    static constexpr s32 IdleThreadPriority = Svc::LowestThreadPriority + 1;
 
     explicit KThread(KernelCore& kernel);
     ~KThread() override;
 
-    using MutexWaitingThreads = std::vector<std::shared_ptr<KThread>>;
-
+public:
     using ThreadContext32 = Core::ARM_Interface::ThreadContext32;
     using ThreadContext64 = Core::ARM_Interface::ThreadContext64;
+    using WaiterList = boost::intrusive::list<KThread>;
 
     /**
      * Creates and returns a new thread. The new thread is immediately scheduled
@@ -121,10 +127,9 @@ public:
      * @param owner_process The parent process for the thread, if null, it's a kernel thread
      * @return A shared pointer to the newly created thread
      */
-    static ResultVal<std::shared_ptr<KThread>> Create(Core::System& system, ThreadType type_flags,
-                                                      std::string name, VAddr entry_point,
-                                                      u32 priority, u64 arg, s32 processor_id,
-                                                      VAddr stack_top, Process* owner_process);
+    [[nodiscard]] static ResultVal<std::shared_ptr<KThread>> Create(
+        Core::System& system, ThreadType type_flags, std::string name, VAddr entry_point,
+        u32 priority, u64 arg, s32 processor_id, VAddr stack_top, Process* owner_process);
 
     /**
      * Creates and returns a new thread. The new thread is immediately scheduled
@@ -140,12 +145,12 @@ public:
      * @param thread_start_parameter The parameter which will passed to host context on init
      * @return A shared pointer to the newly created thread
      */
-    static ResultVal<std::shared_ptr<KThread>> Create(
+    [[nodiscard]] static ResultVal<std::shared_ptr<KThread>> Create(
         Core::System& system, ThreadType type_flags, std::string name, VAddr entry_point,
         u32 priority, u64 arg, s32 processor_id, VAddr stack_top, Process* owner_process,
         std::function<void(void*)>&& thread_start_func, void* thread_start_parameter);
 
-    std::string GetName() const override {
+    [[nodiscard]] std::string GetName() const override {
         return name;
     }
 
@@ -153,12 +158,12 @@ public:
         name = std::move(new_name);
     }
 
-    std::string GetTypeName() const override {
+    [[nodiscard]] std::string GetTypeName() const override {
         return "Thread";
     }
 
     static constexpr HandleType HANDLE_TYPE = HandleType::Thread;
-    HandleType GetHandleType() const override {
+    [[nodiscard]] HandleType GetHandleType() const override {
         return HANDLE_TYPE;
     }
 
@@ -167,15 +172,15 @@ public:
      * @return The current thread's priority
      */
     [[nodiscard]] s32 GetPriority() const {
-        return current_priority;
+        return priority;
     }
 
     /**
      * Sets the thread's current priority.
      * @param priority The new priority.
      */
-    void SetPriority(s32 priority) {
-        current_priority = priority;
+    void SetPriority(s32 value) {
+        priority = value;
     }
 
     /**
@@ -187,15 +192,6 @@ public:
     }
 
     /**
-     * Sets the thread's nominal priority.
-     * @param priority The new priority.
-     */
-    void SetBasePriority(u32 priority);
-
-    /// Changes the core that the thread is running or scheduled to run on.
-    [[nodiscard]] ResultCode SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask);
-
-    /**
      * Gets the thread's thread ID
      * @return The thread's ID
      */
@@ -203,46 +199,67 @@ public:
         return thread_id;
     }
 
-    /// Resumes a thread from waiting
+    void ContinueIfHasKernelWaiters() {
+        if (GetNumKernelWaiters() > 0) {
+            Continue();
+        }
+    }
+
     void Wakeup();
 
-    ResultCode Start();
+    void SetBasePriority(s32 value);
 
-    virtual bool IsSignaled() const override;
+    [[nodiscard]] ResultCode Run();
 
-    /// Cancels a waiting operation that this thread may or may not be within.
-    ///
-    /// When the thread is within a waiting state, this will set the thread's
-    /// waiting result to signal a canceled wait. The function will then resume
-    /// this thread.
-    ///
-    void CancelWait();
+    void Exit();
 
-    void SetSynchronizationResults(KSynchronizationObject* object, ResultCode result);
+    [[nodiscard]] u32 GetSuspendFlags() const {
+        return suspend_allowed_flags & suspend_request_flags;
+    }
 
-    void SetSyncedObject(KSynchronizationObject* object, ResultCode result) {
-        SetSynchronizationResults(object, result);
+    [[nodiscard]] bool IsSuspended() const {
+        return GetSuspendFlags() != 0;
     }
 
-    ResultCode GetWaitResult(KSynchronizationObject** out) const {
-        *out = signaling_object;
-        return signaling_result;
+    [[nodiscard]] bool IsSuspendRequested(SuspendType type) const {
+        return (suspend_request_flags &
+                (1u << (static_cast<u32>(ThreadState::SuspendShift) + static_cast<u32>(type)))) !=
+               0;
     }
 
-    ResultCode GetSignalingResult() const {
-        return signaling_result;
+    [[nodiscard]] bool IsSuspendRequested() const {
+        return suspend_request_flags != 0;
     }
 
-    /**
-     * Stops a thread, invalidating it from further use
-     */
-    void Stop();
+    void RequestSuspend(SuspendType type);
+
+    void Resume(SuspendType type);
+
+    void TrySuspend();
+
+    void Continue();
+
+    void Suspend();
+
+    void Finalize() override;
+
+    bool IsSignaled() const override;
+
+    void SetSyncedObject(KSynchronizationObject* obj, ResultCode wait_res) {
+        synced_object = obj;
+        wait_result = wait_res;
+    }
+
+    [[nodiscard]] ResultCode GetWaitResult(KSynchronizationObject** out) const {
+        *out = synced_object;
+        return wait_result;
+    }
 
     /*
      * Returns the Thread Local Storage address of the current thread
      * @returns VAddr of the thread's TLS
      */
-    VAddr GetTLSAddress() const {
+    [[nodiscard]] VAddr GetTLSAddress() const {
         return tls_address;
     }
 
@@ -250,62 +267,45 @@ public:
      * Returns the value of the TPIDR_EL0 Read/Write system register for this thread.
      * @returns The value of the TPIDR_EL0 register.
      */
-    u64 GetTPIDR_EL0() const {
-        return tpidr_el0;
+    [[nodiscard]] u64 GetTPIDR_EL0() const {
+        return thread_context_64.tpidr;
     }
 
     /// Sets the value of the TPIDR_EL0 Read/Write system register for this thread.
     void SetTPIDR_EL0(u64 value) {
-        tpidr_el0 = value;
+        thread_context_64.tpidr = value;
+        thread_context_32.tpidr = static_cast<u32>(value);
     }
 
-    /*
-     * Returns the address of the current thread's command buffer, located in the TLS.
-     * @returns VAddr of the thread's command buffer.
-     */
-    VAddr GetCommandBufferAddress() const;
-
-    ThreadContext32& GetContext32() {
-        return context_32;
-    }
-
-    const ThreadContext32& GetContext32() const {
-        return context_32;
+    [[nodiscard]] ThreadContext32& GetContext32() {
+        return thread_context_32;
     }
 
-    ThreadContext64& GetContext64() {
-        return context_64;
+    [[nodiscard]] const ThreadContext32& GetContext32() const {
+        return thread_context_32;
     }
 
-    const ThreadContext64& GetContext64() const {
-        return context_64;
+    [[nodiscard]] ThreadContext64& GetContext64() {
+        return thread_context_64;
     }
 
-    bool IsKernelThread() const {
-        return type == ThreadType::Kernel;
+    [[nodiscard]] const ThreadContext64& GetContext64() const {
+        return thread_context_64;
     }
 
-    bool WasRunning() const {
-        return was_running;
-    }
-
-    void SetWasRunning(bool value) {
-        was_running = value;
-    }
+    [[nodiscard]] std::shared_ptr<Common::Fiber>& GetHostContext();
 
-    std::shared_ptr<Common::Fiber>& GetHostContext();
-
-    ThreadState GetState() const {
+    [[nodiscard]] ThreadState GetState() const {
         return thread_state & ThreadState::Mask;
     }
 
-    ThreadState GetRawState() const {
+    [[nodiscard]] ThreadState GetRawState() const {
         return thread_state;
     }
 
     void SetState(ThreadState state);
 
-    s64 GetLastScheduledTick() const {
+    [[nodiscard]] s64 GetLastScheduledTick() const {
         return last_scheduled_tick;
     }
 
@@ -313,43 +313,44 @@ public:
         last_scheduled_tick = tick;
     }
 
-    u64 GetTotalCPUTimeTicks() const {
-        return total_cpu_time_ticks;
+    void AddCpuTime([[maybe_unused]] s32 core_id_, s64 amount) {
+        cpu_time += amount;
+        // TODO(bunnei): Debug kernels track per-core tick counts. Should we?
     }
 
-    void UpdateCPUTimeTicks(u64 ticks) {
-        total_cpu_time_ticks += ticks;
+    [[nodiscard]] s64 GetCpuTime() const {
+        return cpu_time;
     }
 
-    s32 GetProcessorID() const {
-        return processor_id;
+    [[nodiscard]] s32 GetActiveCore() const {
+        return core_id;
     }
 
-    s32 GetActiveCore() const {
-        return GetProcessorID();
+    void SetActiveCore(s32 core) {
+        core_id = core;
     }
 
-    void SetProcessorID(s32 new_core) {
-        processor_id = new_core;
+    [[nodiscard]] s32 GetCurrentCore() const {
+        return current_core_id;
     }
 
-    void SetActiveCore(s32 new_core) {
-        processor_id = new_core;
+    void SetCurrentCore(s32 core) {
+        current_core_id = core;
     }
 
-    Process* GetOwnerProcess() {
-        return owner_process;
+    [[nodiscard]] Process* GetOwnerProcess() {
+        return parent;
     }
 
-    const Process* GetOwnerProcess() const {
-        return owner_process;
+    [[nodiscard]] const Process* GetOwnerProcess() const {
+        return parent;
     }
 
-    const MutexWaitingThreads& GetMutexWaitingThreads() const {
-        return wait_mutex_threads;
+    [[nodiscard]] bool IsUserThread() const {
+        return parent != nullptr;
     }
 
-    KThread* GetLockOwner() const {
+    [[nodiscard]] KThread* GetLockOwner() const {
         return lock_owner;
     }
 
@@ -357,20 +358,21 @@ public:
         lock_owner = owner;
     }
 
-    u32 GetIdealCore() const {
-        return ideal_core;
+    [[nodiscard]] const KAffinityMask& GetAffinityMask() const {
+        return physical_affinity_mask;
     }
 
-    const KAffinityMask& GetAffinityMask() const {
-        return affinity_mask;
-    }
+    [[nodiscard]] ResultCode GetCoreMask(s32* out_ideal_core, u64* out_affinity_mask);
+
+    [[nodiscard]] ResultCode GetPhysicalCoreMask(s32* out_ideal_core, u64* out_affinity_mask);
 
-    ResultCode SetActivity(ThreadActivity value);
+    [[nodiscard]] ResultCode SetCoreMask(s32 core_id, u64 v_affinity_mask);
 
-    /// Sleeps this thread for the given amount of nanoseconds.
-    ResultCode Sleep(s64 nanoseconds);
+    [[nodiscard]] ResultCode SetActivity(Svc::ThreadActivity activity);
 
-    s64 GetYieldScheduleCount() const {
+    [[nodiscard]] ResultCode Sleep(s64 timeout);
+
+    [[nodiscard]] s64 GetYieldScheduleCount() const {
         return schedule_count;
     }
 
@@ -378,56 +380,49 @@ public:
         schedule_count = count;
     }
 
-    bool IsRunning() const {
-        return is_running;
-    }
-
-    void SetIsRunning(bool value) {
-        is_running = value;
-    }
-
-    bool IsWaitCancelled() const {
-        return is_sync_cancelled;
-    }
+    void WaitCancel();
 
-    void ClearWaitCancelled() {
-        is_sync_cancelled = false;
+    [[nodiscard]] bool IsWaitCancelled() const {
+        return wait_cancelled;
     }
 
-    Handle GetGlobalHandle() const {
-        return global_handle;
+    [[nodiscard]] void ClearWaitCancelled() {
+        wait_cancelled = false;
     }
 
-    bool IsCancellable() const {
-        return is_cancellable;
+    [[nodiscard]] bool IsCancellable() const {
+        return cancellable;
     }
 
     void SetCancellable() {
-        is_cancellable = true;
+        cancellable = true;
     }
 
     void ClearCancellable() {
-        is_cancellable = false;
-    }
-
-    bool IsTerminationRequested() const {
-        return will_be_terminated || GetRawState() == ThreadState::Terminated;
+        cancellable = false;
     }
 
-    bool IsPaused() const {
-        return pausing_state != 0;
+    [[nodiscard]] bool IsTerminationRequested() const {
+        return termination_requested || GetRawState() == ThreadState::Terminated;
     }
 
-    bool IsContinuousOnSVC() const {
-        return is_continuous_on_svc;
-    }
+    struct StackParameters {
+        u8 svc_permission[0x10];
+        std::atomic<u8> dpc_flags;
+        u8 current_svc_id;
+        bool is_calling_svc;
+        bool is_in_exception_handler;
+        bool is_pinned;
+        s32 disable_count;
+        KThread* cur_thread;
+    };
 
-    void SetContinuousOnSVC(bool is_continuous) {
-        is_continuous_on_svc = is_continuous;
+    [[nodiscard]] StackParameters& GetStackParameters() {
+        return stack_parameters;
     }
 
-    bool HasExited() const {
-        return has_exited;
+    [[nodiscard]] const StackParameters& GetStackParameters() const {
+        return stack_parameters;
     }
 
     class QueueEntry {
@@ -457,26 +452,78 @@ public:
         KThread* next{};
     };
 
-    QueueEntry& GetPriorityQueueEntry(s32 core) {
+    [[nodiscard]] QueueEntry& GetPriorityQueueEntry(s32 core) {
         return per_core_priority_queue_entry[core];
     }
 
-    const QueueEntry& GetPriorityQueueEntry(s32 core) const {
+    [[nodiscard]] const QueueEntry& GetPriorityQueueEntry(s32 core) const {
         return per_core_priority_queue_entry[core];
     }
 
-    s32 GetDisableDispatchCount() const {
-        return disable_count;
+    void SetSleepingQueue(KThreadQueue* q) {
+        sleeping_queue = q;
+    }
+
+    [[nodiscard]] s32 GetDisableDispatchCount() const {
+        return this->GetStackParameters().disable_count;
     }
 
     void DisableDispatch() {
-        ASSERT(GetDisableDispatchCount() >= 0);
-        disable_count++;
+        ASSERT(GetCurrentThread(kernel).GetDisableDispatchCount() >= 0);
+        this->GetStackParameters().disable_count++;
     }
 
     void EnableDispatch() {
-        ASSERT(GetDisableDispatchCount() > 0);
-        disable_count--;
+        ASSERT(GetCurrentThread(kernel).GetDisableDispatchCount() > 0);
+        this->GetStackParameters().disable_count--;
+    }
+
+    void Pin();
+
+    void Unpin();
+
+    void SetInExceptionHandler() {
+        this->GetStackParameters().is_in_exception_handler = true;
+    }
+
+    void ClearInExceptionHandler() {
+        this->GetStackParameters().is_in_exception_handler = false;
+    }
+
+    [[nodiscard]] bool IsInExceptionHandler() const {
+        return this->GetStackParameters().is_in_exception_handler;
+    }
+
+    void SetIsCallingSvc() {
+        this->GetStackParameters().is_calling_svc = true;
+    }
+
+    void ClearIsCallingSvc() {
+        this->GetStackParameters().is_calling_svc = false;
+    }
+
+    [[nodiscard]] bool IsCallingSvc() const {
+        return this->GetStackParameters().is_calling_svc;
+    }
+
+    [[nodiscard]] u8 GetSvcId() const {
+        return this->GetStackParameters().current_svc_id;
+    }
+
+    void RegisterDpc(DpcFlag flag) {
+        this->GetStackParameters().dpc_flags |= static_cast<u8>(flag);
+    }
+
+    void ClearDpc(DpcFlag flag) {
+        this->GetStackParameters().dpc_flags &= ~static_cast<u8>(flag);
+    }
+
+    [[nodiscard]] u8 GetDpc() const {
+        return this->GetStackParameters().dpc_flags;
+    }
+
+    [[nodiscard]] bool HasDpc() const {
+        return this->GetDpc() != 0;
     }
 
     void SetWaitReasonForDebugging(ThreadWaitReasonForDebugging reason) {
@@ -507,10 +554,16 @@ public:
         return mutex_wait_address_for_debugging;
     }
 
+    [[nodiscard]] s32 GetIdealCoreForDebugging() const {
+        return virtual_ideal_core_id;
+    }
+
     void AddWaiter(KThread* thread);
 
     void RemoveWaiter(KThread* thread);
 
+    [[nodiscard]] ResultCode GetThreadContext3(std::vector<u8>& out);
+
     [[nodiscard]] KThread* RemoveWaiterByKey(s32* out_num_waiters, VAddr key);
 
     [[nodiscard]] VAddr GetAddressKey() const {
@@ -530,6 +583,22 @@ public:
         address_key_value = val;
     }
 
+    [[nodiscard]] bool HasWaiters() const {
+        return !waiter_list.empty();
+    }
+
+    [[nodiscard]] s32 GetNumKernelWaiters() const {
+        return num_kernel_waiters;
+    }
+
+    [[nodiscard]] u64 GetConditionVariableKey() const {
+        return condvar_key;
+    }
+
+    [[nodiscard]] u64 GetAddressArbiterKey() const {
+        return condvar_key;
+    }
+
 private:
     static constexpr size_t PriorityInheritanceCountMax = 10;
     union SyncObjectBuffer {
@@ -560,8 +629,8 @@ private:
             std::same_as<T, KThread> ||
             std::same_as<T, LightCompareType>) static constexpr int Compare(const T& lhs,
                                                                             const KThread& rhs) {
-            const uintptr_t l_key = lhs.GetConditionVariableKey();
-            const uintptr_t r_key = rhs.GetConditionVariableKey();
+            const u64 l_key = lhs.GetConditionVariableKey();
+            const u64 r_key = rhs.GetConditionVariableKey();
 
             if (l_key < r_key) {
                 // Sort first by key
@@ -575,26 +644,88 @@ private:
         }
     };
 
-    Common::IntrusiveRedBlackTreeNode condvar_arbiter_tree_node{};
+    void AddWaiterImpl(KThread* thread);
+
+    void RemoveWaiterImpl(KThread* thread);
 
+    void StartTermination();
+
+    [[nodiscard]] ResultCode Initialize(KThreadFunction func, uintptr_t arg, VAddr user_stack_top,
+                                        s32 prio, s32 virt_core, Process* owner, ThreadType type);
+
+    [[nodiscard]] static ResultCode InitializeThread(KThread* thread, KThreadFunction func,
+                                                     uintptr_t arg, VAddr user_stack_top, s32 prio,
+                                                     s32 core, Process* owner, ThreadType type);
+
+    static void RestorePriority(KernelCore& kernel, KThread* thread);
+
+    // For core KThread implementation
+    ThreadContext32 thread_context_32{};
+    ThreadContext64 thread_context_64{};
+    Common::IntrusiveRedBlackTreeNode condvar_arbiter_tree_node{};
+    s32 priority{};
     using ConditionVariableThreadTreeTraits =
         Common::IntrusiveRedBlackTreeMemberTraitsDeferredAssert<
             &KThread::condvar_arbiter_tree_node>;
     using ConditionVariableThreadTree =
         ConditionVariableThreadTreeTraits::TreeType<ConditionVariableComparator>;
+    ConditionVariableThreadTree* condvar_tree{};
+    u64 condvar_key{};
+    u64 virtual_affinity_mask{};
+    KAffinityMask physical_affinity_mask{};
+    u64 thread_id{};
+    std::atomic<s64> cpu_time{};
+    KSynchronizationObject* synced_object{};
+    VAddr address_key{};
+    Process* parent{};
+    VAddr kernel_stack_top{};
+    u32* light_ipc_data{};
+    VAddr tls_address{};
+    KLightLock activity_pause_lock;
+    s64 schedule_count{};
+    s64 last_scheduled_tick{};
+    std::array<QueueEntry, Core::Hardware::NUM_CPU_CORES> per_core_priority_queue_entry{};
+    KThreadQueue* sleeping_queue{};
+    WaiterList waiter_list{};
+    WaiterList pinned_waiter_list{};
+    KThread* lock_owner{};
+    u32 address_key_value{};
+    u32 suspend_request_flags{};
+    u32 suspend_allowed_flags{};
+    ResultCode wait_result{RESULT_SUCCESS};
+    s32 base_priority{};
+    s32 physical_ideal_core_id{};
+    s32 virtual_ideal_core_id{};
+    s32 num_kernel_waiters{};
+    s32 current_core_id{};
+    s32 core_id{};
+    KAffinityMask original_physical_affinity_mask{};
+    s32 original_physical_ideal_core_id{};
+    s32 num_core_migration_disables{};
+    ThreadState thread_state{};
+    std::atomic<bool> termination_requested{};
+    bool wait_cancelled{};
+    bool cancellable{};
+    bool signaled{};
+    bool initialized{};
+    bool debug_attached{};
+    s8 priority_inheritance_count{};
+    bool resource_limit_release_hint{};
+    StackParameters stack_parameters{};
 
-public:
-    using ConditionVariableThreadTreeType = ConditionVariableThreadTree;
+    // For emulation
+    std::shared_ptr<Common::Fiber> host_context{};
 
-    [[nodiscard]] uintptr_t GetConditionVariableKey() const {
-        return condvar_key;
-    }
+    // For debugging
+    std::vector<KSynchronizationObject*> wait_objects_for_debugging;
+    VAddr mutex_wait_address_for_debugging{};
+    ThreadWaitReasonForDebugging wait_reason_for_debugging{};
+    std::string name;
 
-    [[nodiscard]] uintptr_t GetAddressArbiterKey() const {
-        return condvar_key;
-    }
+public:
+    using ConditionVariableThreadTreeType = ConditionVariableThreadTree;
 
-    void SetConditionVariable(ConditionVariableThreadTree* tree, VAddr address, uintptr_t cv_key,
+    void SetConditionVariable(ConditionVariableThreadTree* tree, VAddr address, u64 cv_key,
                               u32 value) {
         condvar_tree = tree;
         condvar_key = cv_key;
@@ -610,7 +741,7 @@ public:
         return condvar_tree != nullptr;
     }
 
-    void SetAddressArbiter(ConditionVariableThreadTree* tree, uintptr_t address) {
+    void SetAddressArbiter(ConditionVariableThreadTree* tree, u64 address) {
         condvar_tree = tree;
         condvar_key = address;
     }
@@ -626,111 +757,6 @@ public:
     [[nodiscard]] ConditionVariableThreadTree* GetConditionVariableTree() const {
         return condvar_tree;
     }
-
-    [[nodiscard]] bool HasWaiters() const {
-        return !waiter_list.empty();
-    }
-
-private:
-    void AddSchedulingFlag(ThreadSchedFlags flag);
-    void RemoveSchedulingFlag(ThreadSchedFlags flag);
-    void AddWaiterImpl(KThread* thread);
-    void RemoveWaiterImpl(KThread* thread);
-    static void RestorePriority(KernelCore& kernel, KThread* thread);
-
-    Common::SpinLock context_guard{};
-    ThreadContext32 context_32{};
-    ThreadContext64 context_64{};
-    std::shared_ptr<Common::Fiber> host_context{};
-
-    ThreadState thread_state = ThreadState::Initialized;
-
-    u64 thread_id = 0;
-
-    VAddr entry_point = 0;
-    VAddr stack_top = 0;
-    std::atomic_int disable_count = 0;
-
-    ThreadType type;
-
-    /// Nominal thread priority, as set by the emulated application.
-    /// The nominal priority is the thread priority without priority
-    /// inheritance taken into account.
-    s32 base_priority{};
-
-    /// Current thread priority. This may change over the course of the
-    /// thread's lifetime in order to facilitate priority inheritance.
-    s32 current_priority{};
-
-    u64 total_cpu_time_ticks = 0; ///< Total CPU running ticks.
-    s64 schedule_count{};
-    s64 last_scheduled_tick{};
-
-    s32 processor_id = 0;
-
-    VAddr tls_address = 0; ///< Virtual address of the Thread Local Storage of the thread
-    u64 tpidr_el0 = 0;     ///< TPIDR_EL0 read/write system register.
-
-    /// Process that owns this thread
-    Process* owner_process;
-
-    /// Objects that the thread is waiting on, in the same order as they were
-    /// passed to WaitSynchronization. This is used for debugging only.
-    std::vector<KSynchronizationObject*> wait_objects_for_debugging;
-
-    /// The current mutex wait address. This is used for debugging only.
-    VAddr mutex_wait_address_for_debugging{};
-
-    /// The reason the thread is waiting. This is used for debugging only.
-    ThreadWaitReasonForDebugging wait_reason_for_debugging{};
-
-    KSynchronizationObject* signaling_object;
-    ResultCode signaling_result{RESULT_SUCCESS};
-
-    /// List of threads that are waiting for a mutex that is held by this thread.
-    MutexWaitingThreads wait_mutex_threads;
-
-    /// Thread that owns the lock that this thread is waiting for.
-    KThread* lock_owner{};
-
-    /// Handle used as userdata to reference this object when inserting into the CoreTiming queue.
-    Handle global_handle = 0;
-
-    KScheduler* scheduler = nullptr;
-
-    std::array<QueueEntry, Core::Hardware::NUM_CPU_CORES> per_core_priority_queue_entry{};
-
-    u32 ideal_core{0xFFFFFFFF};
-    KAffinityMask affinity_mask{};
-
-    s32 ideal_core_override = -1;
-    u32 affinity_override_count = 0;
-
-    u32 pausing_state = 0;
-    bool is_running = false;
-    bool is_cancellable = false;
-    bool is_sync_cancelled = false;
-
-    bool is_continuous_on_svc = false;
-
-    bool will_be_terminated = false;
-    bool has_exited = false;
-
-    bool was_running = false;
-
-    bool signaled{};
-
-    ConditionVariableThreadTree* condvar_tree{};
-    uintptr_t condvar_key{};
-    VAddr address_key{};
-    u32 address_key_value{};
-    s32 num_kernel_waiters{};
-
-    using WaiterList = boost::intrusive::list<KThread>;
-    WaiterList waiter_list{};
-    WaiterList pinned_waiter_list{};
-
-    std::string name;
 };
 
 } // namespace Kernel
diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp
index 39d5122f5..d61659453 100644
--- a/src/core/hle/kernel/kernel.cpp
+++ b/src/core/hle/kernel/kernel.cpp
@@ -117,14 +117,14 @@ struct KernelCore::Impl {
     void InitializePhysicalCores() {
         exclusive_monitor =
             Core::MakeExclusiveMonitor(system.Memory(), Core::Hardware::NUM_CPU_CORES);
-        for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) {
+        for (s32 i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) {
             schedulers[i] = std::make_unique<Kernel::KScheduler>(system, i);
             cores.emplace_back(i, system, *schedulers[i], interrupts);
         }
     }
 
     void InitializeSchedulers() {
-        for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) {
+        for (s32 i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) {
             cores[i].Scheduler().Initialize();
         }
     }
@@ -169,9 +169,9 @@ struct KernelCore::Impl {
             std::string name = "Suspend Thread Id:" + std::to_string(i);
             std::function<void(void*)> init_func = Core::CpuManager::GetSuspendThreadStartFunc();
             void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater();
-            auto thread_res = KThread::Create(system, ThreadType::Kernel, std::move(name), 0, 0, 0,
-                                              static_cast<u32>(i), 0, nullptr, std::move(init_func),
-                                              init_func_parameter);
+            auto thread_res = KThread::Create(system, ThreadType::HighPriority, std::move(name), 0,
+                                              0, 0, static_cast<u32>(i), 0, nullptr,
+                                              std::move(init_func), init_func_parameter);
 
             suspend_threads[i] = std::move(thread_res).Unwrap();
         }
diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp
index 819e275ff..9f4583b49 100644
--- a/src/core/hle/kernel/process.cpp
+++ b/src/core/hle/kernel/process.cpp
@@ -136,6 +136,23 @@ std::shared_ptr<ResourceLimit> Process::GetResourceLimit() const {
     return resource_limit;
 }
 
+void Process::IncrementThreadCount() {
+    ASSERT(num_threads >= 0);
+    ++num_created_threads;
+
+    if (const auto count = ++num_threads; count > peak_num_threads) {
+        peak_num_threads = count;
+    }
+}
+
+void Process::DecrementThreadCount() {
+    ASSERT(num_threads > 0);
+
+    if (const auto count = --num_threads; count == 0) {
+        UNIMPLEMENTED_MSG("Process termination is not implemented!");
+    }
+}
+
 u64 Process::GetTotalPhysicalMemoryAvailable() const {
     const u64 capacity{resource_limit->GetCurrentResourceValue(ResourceType::PhysicalMemory) +
                        page_table->GetTotalHeapSize() + GetSystemResourceSize() + image_size +
@@ -161,6 +178,61 @@ u64 Process::GetTotalPhysicalMemoryUsedWithoutSystemResource() const {
     return GetTotalPhysicalMemoryUsed() - GetSystemResourceUsage();
 }
 
+bool Process::ReleaseUserException(KThread* thread) {
+    KScopedSchedulerLock sl{kernel};
+
+    if (exception_thread == thread) {
+        exception_thread = nullptr;
+
+        // Remove waiter thread.
+        s32 num_waiters{};
+        KThread* next = thread->RemoveWaiterByKey(
+            std::addressof(num_waiters),
+            reinterpret_cast<uintptr_t>(std::addressof(exception_thread)));
+        if (next != nullptr) {
+            if (next->GetState() == ThreadState::Waiting) {
+                next->SetState(ThreadState::Runnable);
+            } else {
+                KScheduler::SetSchedulerUpdateNeeded(kernel);
+            }
+        }
+
+        return true;
+    } else {
+        return false;
+    }
+}
+
+void Process::PinCurrentThread() {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+
+    // Get the current thread.
+    const s32 core_id = GetCurrentCoreId(kernel);
+    KThread* cur_thread = GetCurrentThreadPointer(kernel);
+
+    // Pin it.
+    PinThread(core_id, cur_thread);
+    cur_thread->Pin();
+
+    // An update is needed.
+    KScheduler::SetSchedulerUpdateNeeded(kernel);
+}
+
+void Process::UnpinCurrentThread() {
+    ASSERT(kernel.GlobalSchedulerContext().IsLocked());
+
+    // Get the current thread.
+    const s32 core_id = GetCurrentCoreId(kernel);
+    KThread* cur_thread = GetCurrentThreadPointer(kernel);
+
+    // Unpin it.
+    cur_thread->Unpin();
+    UnpinThread(core_id, cur_thread);
+
+    // An update is needed.
+    KScheduler::SetSchedulerUpdateNeeded(kernel);
+}
+
 void Process::RegisterThread(const KThread* thread) {
     thread_list.push_back(thread);
 }
@@ -278,7 +350,7 @@ void Process::PrepareForTermination() {
             ASSERT_MSG(thread->GetState() == ThreadState::Waiting,
                        "Exiting processes with non-waiting threads is currently unimplemented");
 
-            thread->Stop();
+            thread->Exit();
         }
     };
 
diff --git a/src/core/hle/kernel/process.h b/src/core/hle/kernel/process.h
index 917babfb4..11d78f3a8 100644
--- a/src/core/hle/kernel/process.h
+++ b/src/core/hle/kernel/process.h
@@ -217,6 +217,14 @@ public:
         return is_64bit_process;
     }
 
+    [[nodiscard]] bool IsSuspended() const {
+        return is_suspended;
+    }
+
+    void SetSuspended(bool suspended) {
+        is_suspended = suspended;
+    }
+
     /// Gets the total running time of the process instance in ticks.
     u64 GetCPUTimeTicks() const {
         return total_process_running_time_ticks;
@@ -237,6 +245,33 @@ public:
         ++schedule_count;
     }
 
+    void IncrementThreadCount();
+    void DecrementThreadCount();
+
+    void SetRunningThread(s32 core, KThread* thread, u64 idle_count) {
+        running_threads[core] = thread;
+        running_thread_idle_counts[core] = idle_count;
+    }
+
+    void ClearRunningThread(KThread* thread) {
+        for (size_t i = 0; i < running_threads.size(); ++i) {
+            if (running_threads[i] == thread) {
+                running_threads[i] = nullptr;
+            }
+        }
+    }
+
+    [[nodiscard]] KThread* GetRunningThread(s32 core) const {
+        return running_threads[core];
+    }
+
+    bool ReleaseUserException(KThread* thread);
+
+    [[nodiscard]] KThread* GetPinnedThread(s32 core_id) const {
+        ASSERT(0 <= core_id && core_id < static_cast<s32>(Core::Hardware::NUM_CPU_CORES));
+        return pinned_threads[core_id];
+    }
+
     /// Gets 8 bytes of random data for svcGetInfo RandomEntropy
     u64 GetRandomEntropy(std::size_t index) const {
         return random_entropy.at(index);
@@ -310,6 +345,9 @@ public:
 
     void Finalize() override {}
 
+    void PinCurrentThread();
+    void UnpinCurrentThread();
+
     ///////////////////////////////////////////////////////////////////////////////////////////////
     // Thread-local storage management
 
@@ -320,6 +358,20 @@ public:
     void FreeTLSRegion(VAddr tls_address);
 
 private:
+    void PinThread(s32 core_id, KThread* thread) {
+        ASSERT(0 <= core_id && core_id < static_cast<s32>(Core::Hardware::NUM_CPU_CORES));
+        ASSERT(thread != nullptr);
+        ASSERT(pinned_threads[core_id] == nullptr);
+        pinned_threads[core_id] = thread;
+    }
+
+    void UnpinThread(s32 core_id, KThread* thread) {
+        ASSERT(0 <= core_id && core_id < static_cast<s32>(Core::Hardware::NUM_CPU_CORES));
+        ASSERT(thread != nullptr);
+        ASSERT(pinned_threads[core_id] == thread);
+        pinned_threads[core_id] = nullptr;
+    }
+
     /// Changes the process status. If the status is different
     /// from the current process status, then this will trigger
     /// a process signal.
@@ -408,6 +460,17 @@ private:
     s64 schedule_count{};
 
     bool is_signaled{};
+    bool is_suspended{};
+
+    std::atomic<s32> num_created_threads{};
+    std::atomic<u16> num_threads{};
+    u16 peak_num_threads{};
+
+    std::array<KThread*, Core::Hardware::NUM_CPU_CORES> running_threads{};
+    std::array<u64, Core::Hardware::NUM_CPU_CORES> running_thread_idle_counts{};
+    std::array<KThread*, Core::Hardware::NUM_CPU_CORES> pinned_threads{};
+
+    KThread* exception_thread{};
 
     /// System context
     Core::System& system;
diff --git a/src/core/hle/kernel/server_session.cpp b/src/core/hle/kernel/server_session.cpp
index 280c9b5f6..790dbb998 100644
--- a/src/core/hle/kernel/server_session.cpp
+++ b/src/core/hle/kernel/server_session.cpp
@@ -154,7 +154,7 @@ ResultCode ServerSession::CompleteSyncRequest(HLERequestContext& context) {
         KScopedSchedulerLock lock(kernel);
         if (!context.IsThreadWaiting()) {
             context.GetThread().Wakeup();
-            context.GetThread().SetSynchronizationResults(nullptr, result);
+            context.GetThread().SetSyncedObject(nullptr, result);
         }
     }
 
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index 2512bfd98..dbef854f8 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -351,7 +351,8 @@ static ResultCode SendSyncRequest(Core::System& system, Handle handle) {
         session->SendSyncRequest(SharedFrom(thread), system.Memory(), system.CoreTiming());
     }
 
-    return thread->GetSignalingResult();
+    KSynchronizationObject* dummy{};
+    return thread->GetWaitResult(std::addressof(dummy));
 }
 
 static ResultCode SendSyncRequest32(Core::System& system, Handle handle) {
@@ -359,27 +360,26 @@ static ResultCode SendSyncRequest32(Core::System& system, Handle handle) {
 }
 
 /// Get the ID for the specified thread.
-static ResultCode GetThreadId(Core::System& system, u64* thread_id, Handle thread_handle) {
+static ResultCode GetThreadId(Core::System& system, u64* out_thread_id, Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
 
+    // Get the thread from its handle.
     const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
     const std::shared_ptr<KThread> thread = handle_table.Get<KThread>(thread_handle);
-    if (!thread) {
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", thread_handle);
-        return ERR_INVALID_HANDLE;
-    }
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
 
-    *thread_id = thread->GetThreadID();
+    // Get the thread's id.
+    *out_thread_id = thread->GetThreadID();
     return RESULT_SUCCESS;
 }
 
-static ResultCode GetThreadId32(Core::System& system, u32* thread_id_low, u32* thread_id_high,
-                                Handle thread_handle) {
-    u64 thread_id{};
-    const ResultCode result{GetThreadId(system, &thread_id, thread_handle)};
+static ResultCode GetThreadId32(Core::System& system, u32* out_thread_id_low,
+                                u32* out_thread_id_high, Handle thread_handle) {
+    u64 out_thread_id{};
+    const ResultCode result{GetThreadId(system, &out_thread_id, thread_handle)};
 
-    *thread_id_low = static_cast<u32>(thread_id >> 32);
-    *thread_id_high = static_cast<u32>(thread_id & std::numeric_limits<u32>::max());
+    *out_thread_id_low = static_cast<u32>(out_thread_id >> 32);
+    *out_thread_id_high = static_cast<u32>(out_thread_id & std::numeric_limits<u32>::max());
 
     return result;
 }
@@ -473,15 +473,13 @@ static ResultCode WaitSynchronization32(Core::System& system, u32 timeout_low, u
 static ResultCode CancelSynchronization(Core::System& system, Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:X}", thread_handle);
 
+    // Get the thread from its handle.
     const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
     std::shared_ptr<KThread> thread = handle_table.Get<KThread>(thread_handle);
-    if (!thread) {
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
-                  thread_handle);
-        return ERR_INVALID_HANDLE;
-    }
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
 
-    thread->CancelWait();
+    // Cancel the thread's wait.
+    thread->WaitCancel();
     return RESULT_SUCCESS;
 }
 
@@ -630,7 +628,7 @@ static void Break(Core::System& system, u32 reason, u64 info1, u64 info2) {
         handle_debug_buffer(info1, info2);
 
         auto* const current_thread = system.Kernel().CurrentScheduler()->GetCurrentThread();
-        const auto thread_processor_id = current_thread->GetProcessorID();
+        const auto thread_processor_id = current_thread->GetActiveCore();
         system.ArmInterface(static_cast<std::size_t>(thread_processor_id)).LogBacktrace();
     }
 }
@@ -888,7 +886,7 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha
         const u64 prev_ctx_ticks = scheduler.GetLastContextSwitchTicks();
         u64 out_ticks = 0;
         if (same_thread && info_sub_id == 0xFFFFFFFFFFFFFFFF) {
-            const u64 thread_ticks = current_thread->GetTotalCPUTimeTicks();
+            const u64 thread_ticks = current_thread->GetCpuTime();
 
             out_ticks = thread_ticks + (core_timing.GetCPUTicks() - prev_ctx_ticks);
         } else if (same_thread && info_sub_id == system.CurrentCoreIndex()) {
@@ -1025,127 +1023,109 @@ static ResultCode UnmapPhysicalMemory32(Core::System& system, u32 addr, u32 size
     return UnmapPhysicalMemory(system, addr, size);
 }
 
-/// Sets the thread activity
-static ResultCode SetThreadActivity(Core::System& system, Handle handle, u32 activity) {
-    LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, activity=0x{:08X}", handle, activity);
-    if (activity > static_cast<u32>(ThreadActivity::Paused)) {
-        return ERR_INVALID_ENUM_VALUE;
+constexpr bool IsValidThreadActivity(Svc::ThreadActivity thread_activity) {
+    switch (thread_activity) {
+    case Svc::ThreadActivity::Runnable:
+    case Svc::ThreadActivity::Paused:
+        return true;
+    default:
+        return false;
     }
+}
 
-    const auto* current_process = system.Kernel().CurrentProcess();
-    const std::shared_ptr<KThread> thread = current_process->GetHandleTable().Get<KThread>(handle);
-    if (!thread) {
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle);
-        return ERR_INVALID_HANDLE;
-    }
+/// Sets the thread activity
+static ResultCode SetThreadActivity(Core::System& system, Handle thread_handle,
+                                    Svc::ThreadActivity thread_activity) {
+    LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, activity=0x{:08X}", thread_handle,
+              thread_activity);
 
-    if (thread->GetOwnerProcess() != current_process) {
-        LOG_ERROR(Kernel_SVC,
-                  "The current process does not own the current thread, thread_handle={:08X} "
-                  "thread_pid={}, "
-                  "current_process_pid={}",
-                  handle, thread->GetOwnerProcess()->GetProcessID(),
-                  current_process->GetProcessID());
-        return ERR_INVALID_HANDLE;
-    }
+    // Validate the activity.
+    R_UNLESS(IsValidThreadActivity(thread_activity), Svc::ResultInvalidEnumValue);
 
-    if (thread.get() == system.Kernel().CurrentScheduler()->GetCurrentThread()) {
-        LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread");
-        return ERR_BUSY;
-    }
+    // Get the thread from its handle.
+    auto& kernel = system.Kernel();
+    const auto& handle_table = kernel.CurrentProcess()->GetHandleTable();
+    const std::shared_ptr<KThread> thread = handle_table.Get<KThread>(thread_handle);
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
+
+    // Check that the activity is being set on a non-current thread for the current process.
+    R_UNLESS(thread->GetOwnerProcess() == kernel.CurrentProcess(), Svc::ResultInvalidHandle);
+    R_UNLESS(thread.get() != GetCurrentThreadPointer(kernel), Svc::ResultBusy);
+
+    // Set the activity.
+    R_TRY(thread->SetActivity(thread_activity));
 
-    return thread->SetActivity(static_cast<ThreadActivity>(activity));
+    return RESULT_SUCCESS;
 }
 
-static ResultCode SetThreadActivity32(Core::System& system, Handle handle, u32 activity) {
-    return SetThreadActivity(system, handle, activity);
+static ResultCode SetThreadActivity32(Core::System& system, Handle thread_handle,
+                                      Svc::ThreadActivity thread_activity) {
+    return SetThreadActivity(system, thread_handle, thread_activity);
 }
 
 /// Gets the thread context
-static ResultCode GetThreadContext(Core::System& system, VAddr thread_context, Handle handle) {
-    LOG_DEBUG(Kernel_SVC, "called, context=0x{:08X}, thread=0x{:X}", thread_context, handle);
+static ResultCode GetThreadContext(Core::System& system, VAddr out_context, Handle thread_handle) {
+    LOG_DEBUG(Kernel_SVC, "called, out_context=0x{:08X}, thread_handle=0x{:X}", out_context,
+              thread_handle);
 
+    // Get the thread from its handle.
     const auto* current_process = system.Kernel().CurrentProcess();
-    const std::shared_ptr<KThread> thread = current_process->GetHandleTable().Get<KThread>(handle);
-    if (!thread) {
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle);
-        return ERR_INVALID_HANDLE;
-    }
+    const std::shared_ptr<KThread> thread =
+        current_process->GetHandleTable().Get<KThread>(thread_handle);
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
 
-    if (thread->GetOwnerProcess() != current_process) {
-        LOG_ERROR(Kernel_SVC,
-                  "The current process does not own the current thread, thread_handle={:08X} "
-                  "thread_pid={}, "
-                  "current_process_pid={}",
-                  handle, thread->GetOwnerProcess()->GetProcessID(),
-                  current_process->GetProcessID());
-        return ERR_INVALID_HANDLE;
-    }
+    // Require the handle be to a non-current thread in the current process.
+    R_UNLESS(thread->GetOwnerProcess() == current_process, Svc::ResultInvalidHandle);
+    R_UNLESS(thread.get() != system.Kernel().CurrentScheduler()->GetCurrentThread(),
+             Svc::ResultBusy);
 
-    if (thread.get() == system.Kernel().CurrentScheduler()->GetCurrentThread()) {
-        LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread");
-        return ERR_BUSY;
-    }
+    // Get the thread context.
+    std::vector<u8> context;
+    R_TRY(thread->GetThreadContext3(context));
 
-    Core::ARM_Interface::ThreadContext64 ctx = thread->GetContext64();
-    // Mask away mode bits, interrupt bits, IL bit, and other reserved bits.
-    ctx.pstate &= 0xFF0FFE20;
+    // Copy the thread context to user space.
+    system.Memory().WriteBlock(out_context, context.data(), context.size());
 
-    // If 64-bit, we can just write the context registers directly and we're good.
-    // However, if 32-bit, we have to ensure some registers are zeroed out.
-    if (!current_process->Is64BitProcess()) {
-        std::fill(ctx.cpu_registers.begin() + 15, ctx.cpu_registers.end(), 0);
-        std::fill(ctx.vector_registers.begin() + 16, ctx.vector_registers.end(), u128{});
-    }
-
-    system.Memory().WriteBlock(thread_context, &ctx, sizeof(ctx));
     return RESULT_SUCCESS;
 }
 
-static ResultCode GetThreadContext32(Core::System& system, u32 thread_context, Handle handle) {
-    return GetThreadContext(system, thread_context, handle);
+static ResultCode GetThreadContext32(Core::System& system, u32 out_context, Handle thread_handle) {
+    return GetThreadContext(system, out_context, thread_handle);
 }
 
 /// Gets the priority for the specified thread
-static ResultCode GetThreadPriority(Core::System& system, u32* priority, Handle handle) {
+static ResultCode GetThreadPriority(Core::System& system, u32* out_priority, Handle handle) {
     LOG_TRACE(Kernel_SVC, "called");
 
+    // Get the thread from its handle.
     const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
     const std::shared_ptr<KThread> thread = handle_table.Get<KThread>(handle);
-    if (!thread) {
-        *priority = 0;
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle);
-        return ERR_INVALID_HANDLE;
-    }
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
 
-    *priority = thread->GetPriority();
+    // Get the thread's priority.
+    *out_priority = thread->GetPriority();
     return RESULT_SUCCESS;
 }
 
-static ResultCode GetThreadPriority32(Core::System& system, u32* priority, Handle handle) {
-    return GetThreadPriority(system, priority, handle);
+static ResultCode GetThreadPriority32(Core::System& system, u32* out_priority, Handle handle) {
+    return GetThreadPriority(system, out_priority, handle);
 }
 
 /// Sets the priority for the specified thread
 static ResultCode SetThreadPriority(Core::System& system, Handle handle, u32 priority) {
     LOG_TRACE(Kernel_SVC, "called");
 
-    if (priority > Svc::LowestThreadPriority) {
-        LOG_ERROR(Kernel_SVC, "An invalid priority was specified {} for thread_handle={:08X}",
-                  priority, handle);
-        return ERR_INVALID_THREAD_PRIORITY;
-    }
-
-    const auto* const current_process = system.Kernel().CurrentProcess();
+    // Validate the priority.
+    R_UNLESS(Svc::HighestThreadPriority <= priority && priority <= Svc::LowestThreadPriority,
+             Svc::ResultInvalidPriority);
 
-    std::shared_ptr<KThread> thread = current_process->GetHandleTable().Get<KThread>(handle);
-    if (!thread) {
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle);
-        return ERR_INVALID_HANDLE;
-    }
+    // Get the thread from its handle.
+    const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
+    const std::shared_ptr<KThread> thread = handle_table.Get<KThread>(handle);
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
 
+    // Set the thread priority.
     thread->SetBasePriority(priority);
-
     return RESULT_SUCCESS;
 }
 
@@ -1436,7 +1416,7 @@ static void ExitProcess(Core::System& system) {
     current_process->PrepareForTermination();
 
     // Kill the current thread
-    system.Kernel().CurrentScheduler()->GetCurrentThread()->Stop();
+    system.Kernel().CurrentScheduler()->GetCurrentThread()->Exit();
 }
 
 static void ExitProcess32(Core::System& system) {
@@ -1500,17 +1480,15 @@ static ResultCode CreateThread32(Core::System& system, Handle* out_handle, u32 p
 static ResultCode StartThread(Core::System& system, Handle thread_handle) {
     LOG_DEBUG(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
 
+    // Get the thread from its handle.
     const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
     const std::shared_ptr<KThread> thread = handle_table.Get<KThread>(thread_handle);
-    if (!thread) {
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
-                  thread_handle);
-        return ERR_INVALID_HANDLE;
-    }
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
 
-    ASSERT(thread->GetState() == ThreadState::Initialized);
+    // Try to start the thread.
+    R_TRY(thread->Run());
 
-    return thread->Start();
+    return RESULT_SUCCESS;
 }
 
 static ResultCode StartThread32(Core::System& system, Handle thread_handle) {
@@ -1523,7 +1501,7 @@ static void ExitThread(Core::System& system) {
 
     auto* const current_thread = system.Kernel().CurrentScheduler()->GetCurrentThread();
     system.GlobalSchedulerContext().RemoveThread(SharedFrom(current_thread));
-    current_thread->Stop();
+    current_thread->Exit();
 }
 
 static void ExitThread32(Core::System& system) {
@@ -1532,34 +1510,28 @@ static void ExitThread32(Core::System& system) {
 
 /// Sleep the current thread
 static void SleepThread(Core::System& system, s64 nanoseconds) {
-    LOG_TRACE(Kernel_SVC, "called nanoseconds={}", nanoseconds);
+    auto& kernel = system.Kernel();
+    const auto yield_type = static_cast<Svc::YieldType>(nanoseconds);
 
-    enum class SleepType : s64 {
-        YieldWithoutCoreMigration = 0,
-        YieldWithCoreMigration = -1,
-        YieldAndWaitForLoadBalancing = -2,
-    };
+    LOG_TRACE(Kernel_SVC, "called nanoseconds={}", nanoseconds);
 
-    auto& scheduler = *system.Kernel().CurrentScheduler();
-    if (nanoseconds <= 0) {
-        switch (static_cast<SleepType>(nanoseconds)) {
-        case SleepType::YieldWithoutCoreMigration: {
-            scheduler.YieldWithoutCoreMigration();
-            break;
-        }
-        case SleepType::YieldWithCoreMigration: {
-            scheduler.YieldWithCoreMigration();
-            break;
-        }
-        case SleepType::YieldAndWaitForLoadBalancing: {
-            scheduler.YieldToAnyThread();
-            break;
-        }
-        default:
-            UNREACHABLE_MSG("Unimplemented sleep yield type '{:016X}'!", nanoseconds);
-        }
+    // When the input tick is positive, sleep.
+    if (nanoseconds > 0) {
+        // Convert the timeout from nanoseconds to ticks.
+        // NOTE: Nintendo does not use this conversion logic in WaitSynchronization...
+
+        // Sleep.
+        // NOTE: Nintendo does not check the result of this sleep.
+        static_cast<void>(GetCurrentThread(kernel).Sleep(nanoseconds));
+    } else if (yield_type == Svc::YieldType::WithoutCoreMigration) {
+        KScheduler::YieldWithoutCoreMigration(kernel);
+    } else if (yield_type == Svc::YieldType::WithCoreMigration) {
+        KScheduler::YieldWithCoreMigration(kernel);
+    } else if (yield_type == Svc::YieldType::ToAnyThread) {
+        KScheduler::YieldToAnyThread(kernel);
     } else {
-        scheduler.GetCurrentThread()->Sleep(nanoseconds);
+        // Nintendo does nothing at all if an otherwise invalid value is passed.
+        UNREACHABLE_MSG("Unimplemented sleep yield type '{:016X}'!", nanoseconds);
     }
 }
 
@@ -1822,95 +1794,72 @@ static ResultCode CreateTransferMemory32(Core::System& system, Handle* handle, u
     return CreateTransferMemory(system, handle, addr, size, permissions);
 }
 
-static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle, u32* core,
-                                    u64* mask) {
+static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle, s32* out_core_id,
+                                    u64* out_affinity_mask) {
     LOG_TRACE(Kernel_SVC, "called, handle=0x{:08X}", thread_handle);
 
+    // Get the thread from its handle.
     const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
     const std::shared_ptr<KThread> thread = handle_table.Get<KThread>(thread_handle);
-    if (!thread) {
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
-                  thread_handle);
-        *core = 0;
-        *mask = 0;
-        return ERR_INVALID_HANDLE;
-    }
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
 
-    *core = thread->GetIdealCore();
-    *mask = thread->GetAffinityMask().GetAffinityMask();
+    // Get the core mask.
+    R_TRY(thread->GetCoreMask(out_core_id, out_affinity_mask));
 
     return RESULT_SUCCESS;
 }
 
-static ResultCode GetThreadCoreMask32(Core::System& system, Handle thread_handle, u32* core,
-                                      u32* mask_low, u32* mask_high) {
-    u64 mask{};
-    const auto result = GetThreadCoreMask(system, thread_handle, core, &mask);
-    *mask_high = static_cast<u32>(mask >> 32);
-    *mask_low = static_cast<u32>(mask);
+static ResultCode GetThreadCoreMask32(Core::System& system, Handle thread_handle, s32* out_core_id,
+                                      u32* out_affinity_mask_low, u32* out_affinity_mask_high) {
+    u64 out_affinity_mask{};
+    const auto result = GetThreadCoreMask(system, thread_handle, out_core_id, &out_affinity_mask);
+    *out_affinity_mask_high = static_cast<u32>(out_affinity_mask >> 32);
+    *out_affinity_mask_low = static_cast<u32>(out_affinity_mask);
     return result;
 }
 
-static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, u32 core,
+static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, s32 core_id,
                                     u64 affinity_mask) {
-    LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, core=0x{:X}, affinity_mask=0x{:016X}",
-              thread_handle, core, affinity_mask);
-
-    const auto* const current_process = system.Kernel().CurrentProcess();
+    LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, core_id=0x{:X}, affinity_mask=0x{:016X}",
+              thread_handle, core_id, affinity_mask);
 
-    if (core == static_cast<u32>(Svc::IdealCoreUseProcessValue)) {
-        const u8 ideal_cpu_core = current_process->GetIdealCoreId();
+    const auto& current_process = *system.Kernel().CurrentProcess();
 
-        ASSERT(ideal_cpu_core != static_cast<u8>(Svc::IdealCoreUseProcessValue));
-
-        // Set the target CPU to the ideal core specified by the process.
-        core = ideal_cpu_core;
-        affinity_mask = 1ULL << core;
+    // Determine the core id/affinity mask.
+    if (core_id == Svc::IdealCoreUseProcessValue) {
+        core_id = current_process.GetIdealCoreId();
+        affinity_mask = (1ULL << core_id);
     } 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::Hardware::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>(Svc::IdealCoreDontCare) &&
-                   core != static_cast<u32>(Svc::IdealCoreNoUpdate)) {
-            LOG_ERROR(Kernel_SVC, "Invalid processor ID specified (core={}).", core);
-            return ERR_INVALID_PROCESSOR_ID;
+        // Validate the affinity mask.
+        const u64 process_core_mask = current_process.GetCoreMask();
+        R_UNLESS((affinity_mask | process_core_mask) == process_core_mask,
+                 Svc::ResultInvalidCoreId);
+        R_UNLESS(affinity_mask != 0, Svc::ResultInvalidCombination);
+
+        // Validate the core id.
+        if (IsValidCoreId(core_id)) {
+            R_UNLESS(((1ULL << core_id) & affinity_mask) != 0, Svc::ResultInvalidCombination);
+        } else {
+            R_UNLESS(core_id == Svc::IdealCoreNoUpdate || core_id == Svc::IdealCoreDontCare,
+                     Svc::ResultInvalidCoreId);
         }
     }
 
-    const auto& handle_table = current_process->GetHandleTable();
+    // Get the thread from its handle.
+    const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
     const std::shared_ptr<KThread> thread = handle_table.Get<KThread>(thread_handle);
-    if (!thread) {
-        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
-                  thread_handle);
-        return ERR_INVALID_HANDLE;
-    }
+    R_UNLESS(thread, Svc::ResultInvalidHandle);
 
-    return thread->SetCoreAndAffinityMask(core, affinity_mask);
+    // Set the core mask.
+    R_TRY(thread->SetCoreMask(core_id, affinity_mask));
+
+    return RESULT_SUCCESS;
 }
 
-static ResultCode SetThreadCoreMask32(Core::System& system, Handle thread_handle, u32 core,
+static ResultCode SetThreadCoreMask32(Core::System& system, Handle thread_handle, s32 core_id,
                                       u32 affinity_mask_low, u32 affinity_mask_high) {
     const auto affinity_mask = u64{affinity_mask_low} | (u64{affinity_mask_high} << 32);
-    return SetThreadCoreMask(system, thread_handle, core, affinity_mask);
+    return SetThreadCoreMask(system, thread_handle, core_id, affinity_mask);
 }
 
 static ResultCode CreateEvent(Core::System& system, Handle* write_handle, Handle* read_handle) {
@@ -2474,7 +2423,7 @@ void Call(Core::System& system, u32 immediate) {
     kernel.EnterSVCProfile();
 
     auto* thread = kernel.CurrentScheduler()->GetCurrentThread();
-    thread->SetContinuousOnSVC(true);
+    thread->SetIsCallingSvc();
 
     const FunctionDef* info = system.CurrentProcess()->Is64BitProcess() ? GetSVCInfo64(immediate)
                                                                         : GetSVCInfo32(immediate);
@@ -2490,7 +2439,7 @@ void Call(Core::System& system, u32 immediate) {
 
     kernel.ExitSVCProfile();
 
-    if (!thread->IsContinuousOnSVC()) {
+    if (!thread->IsCallingSvc()) {
         auto* host_context = thread->GetHostContext().get();
         host_context->Rewind();
     }
diff --git a/src/core/hle/kernel/svc_types.h b/src/core/hle/kernel/svc_types.h
index ded55af9a..ec463b97c 100644
--- a/src/core/hle/kernel/svc_types.h
+++ b/src/core/hle/kernel/svc_types.h
@@ -77,6 +77,12 @@ enum class ArbitrationType : u32 {
     WaitIfEqual = 2,
 };
 
+enum class YieldType : s64 {
+    WithoutCoreMigration = 0,
+    WithCoreMigration = -1,
+    ToAnyThread = -2,
+};
+
 enum class ThreadActivity : u32 {
     Runnable = 0,
     Paused = 1,
diff --git a/src/core/hle/kernel/svc_wrap.h b/src/core/hle/kernel/svc_wrap.h
index a32750ed7..96afd544b 100644
--- a/src/core/hle/kernel/svc_wrap.h
+++ b/src/core/hle/kernel/svc_wrap.h
@@ -58,6 +58,14 @@ void SvcWrap64(Core::System& system) {
         func(system, static_cast<u32>(Param(system, 0)), static_cast<u32>(Param(system, 1))).raw);
 }
 
+// Used by SetThreadActivity
+template <ResultCode func(Core::System&, Handle, Svc::ThreadActivity)>
+void SvcWrap64(Core::System& system) {
+    FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)),
+                            static_cast<Svc::ThreadActivity>(Param(system, 1)))
+                           .raw);
+}
+
 template <ResultCode func(Core::System&, u32, u64, u64, u64)>
 void SvcWrap64(Core::System& system) {
     FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), Param(system, 1),
@@ -158,9 +166,18 @@ void SvcWrap64(Core::System& system) {
                            .raw);
 }
 
-template <ResultCode func(Core::System&, u32, u32*, u64*)>
+// Used by SetThreadCoreMask
+template <ResultCode func(Core::System&, Handle, s32, u64)>
 void SvcWrap64(Core::System& system) {
-    u32 param_1 = 0;
+    FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)),
+                            static_cast<s32>(Param(system, 1)), Param(system, 2))
+                           .raw);
+}
+
+// Used by GetThreadCoreMask
+template <ResultCode func(Core::System&, Handle, s32*, u64*)>
+void SvcWrap64(Core::System& system) {
+    s32 param_1 = 0;
     u64 param_2 = 0;
     const ResultCode retval = func(system, static_cast<u32>(Param(system, 2)), &param_1, &param_2);
 
@@ -473,12 +490,35 @@ void SvcWrap32(Core::System& system) {
     FuncReturn(system, retval);
 }
 
+// Used by GetThreadCoreMask32
+template <ResultCode func(Core::System&, Handle, s32*, u32*, u32*)>
+void SvcWrap32(Core::System& system) {
+    s32 param_1 = 0;
+    u32 param_2 = 0;
+    u32 param_3 = 0;
+
+    const u32 retval = func(system, Param32(system, 2), &param_1, &param_2, &param_3).raw;
+    system.CurrentArmInterface().SetReg(1, param_1);
+    system.CurrentArmInterface().SetReg(2, param_2);
+    system.CurrentArmInterface().SetReg(3, param_3);
+    FuncReturn(system, retval);
+}
+
 // Used by SignalProcessWideKey32
 template <void func(Core::System&, u32, s32)>
 void SvcWrap32(Core::System& system) {
     func(system, static_cast<u32>(Param(system, 0)), static_cast<s32>(Param(system, 1)));
 }
 
+// Used by SetThreadActivity32
+template <ResultCode func(Core::System&, Handle, Svc::ThreadActivity)>
+void SvcWrap32(Core::System& system) {
+    const u32 retval = func(system, static_cast<Handle>(Param(system, 0)),
+                            static_cast<Svc::ThreadActivity>(Param(system, 1)))
+                           .raw;
+    FuncReturn(system, retval);
+}
+
 // Used by SetThreadPriority32
 template <ResultCode func(Core::System&, Handle, u32)>
 void SvcWrap32(Core::System& system) {
@@ -487,7 +527,7 @@ void SvcWrap32(Core::System& system) {
     FuncReturn(system, retval);
 }
 
-// Used by SetThreadCoreMask32
+// Used by SetMemoryAttribute32
 template <ResultCode func(Core::System&, Handle, u32, u32, u32)>
 void SvcWrap32(Core::System& system) {
     const u32 retval =
@@ -497,6 +537,16 @@ void SvcWrap32(Core::System& system) {
     FuncReturn(system, retval);
 }
 
+// Used by SetThreadCoreMask32
+template <ResultCode func(Core::System&, Handle, s32, u32, u32)>
+void SvcWrap32(Core::System& system) {
+    const u32 retval =
+        func(system, static_cast<Handle>(Param(system, 0)), static_cast<s32>(Param(system, 1)),
+             static_cast<u32>(Param(system, 2)), static_cast<u32>(Param(system, 3)))
+            .raw;
+    FuncReturn(system, retval);
+}
+
 // Used by WaitProcessWideKeyAtomic32
 template <ResultCode func(Core::System&, u32, u32, Handle, u32, u32)>
 void SvcWrap32(Core::System& system) {