25 files changed, 300 insertions, 285 deletions

diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index 39d038493..39ae573b2 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -36,7 +36,6 @@ if (MSVC)
     # /GT                 - Supports fiber safety for data allocated using static thread-local storage
     add_compile_options(
         /MP
-        /Zi
         /Zm200
         /Zo
         /permissive-
@@ -79,6 +78,13 @@ if (MSVC)
         /we5245 # 'function': unreferenced function with internal linkage has been removed
     )
 
+    if (USE_CCACHE)
+    # when caching, we need to use /Z7 to downgrade debug info to use an older but more cachable format
+        add_compile_options(/Z7)
+    else()
+        add_compile_options(/Zi)
+    endif()
+
     if (ARCHITECTURE_x86_64)
         add_compile_options(/QIntel-jcc-erratum)
     endif()
diff --git a/src/common/fiber.cpp b/src/common/fiber.cpp
index f9aeb692a..bc92b360b 100644
--- a/src/common/fiber.cpp
+++ b/src/common/fiber.cpp
@@ -20,10 +20,8 @@ struct Fiber::FiberImpl {
     VirtualBuffer<u8> rewind_stack;
 
     std::mutex guard;
-    std::function<void(void*)> entry_point;
-    std::function<void(void*)> rewind_point;
-    void* rewind_parameter{};
-    void* start_parameter{};
+    std::function<void()> entry_point;
+    std::function<void()> rewind_point;
     std::shared_ptr<Fiber> previous_fiber;
     bool is_thread_fiber{};
     bool released{};
@@ -34,13 +32,8 @@ struct Fiber::FiberImpl {
     boost::context::detail::fcontext_t rewind_context{};
 };
 
-void Fiber::SetStartParameter(void* new_parameter) {
-    impl->start_parameter = new_parameter;
-}
-
-void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param) {
+void Fiber::SetRewindPoint(std::function<void()>&& rewind_func) {
     impl->rewind_point = std::move(rewind_func);
-    impl->rewind_parameter = rewind_param;
 }
 
 void Fiber::Start(boost::context::detail::transfer_t& transfer) {
@@ -48,7 +41,7 @@ void Fiber::Start(boost::context::detail::transfer_t& transfer) {
     impl->previous_fiber->impl->context = transfer.fctx;
     impl->previous_fiber->impl->guard.unlock();
     impl->previous_fiber.reset();
-    impl->entry_point(impl->start_parameter);
+    impl->entry_point();
     UNREACHABLE();
 }
 
@@ -59,7 +52,7 @@ void Fiber::OnRewind([[maybe_unused]] boost::context::detail::transfer_t& transf
     u8* tmp = impl->stack_limit;
     impl->stack_limit = impl->rewind_stack_limit;
     impl->rewind_stack_limit = tmp;
-    impl->rewind_point(impl->rewind_parameter);
+    impl->rewind_point();
     UNREACHABLE();
 }
 
@@ -73,10 +66,8 @@ void Fiber::RewindStartFunc(boost::context::detail::transfer_t transfer) {
     fiber->OnRewind(transfer);
 }
 
-Fiber::Fiber(std::function<void(void*)>&& entry_point_func, void* start_parameter)
-    : impl{std::make_unique<FiberImpl>()} {
+Fiber::Fiber(std::function<void()>&& entry_point_func) : impl{std::make_unique<FiberImpl>()} {
     impl->entry_point = std::move(entry_point_func);
-    impl->start_parameter = start_parameter;
     impl->stack_limit = impl->stack.data();
     impl->rewind_stack_limit = impl->rewind_stack.data();
     u8* stack_base = impl->stack_limit + default_stack_size;
diff --git a/src/common/fiber.h b/src/common/fiber.h
index 873604bc6..f24d333a3 100644
--- a/src/common/fiber.h
+++ b/src/common/fiber.h
@@ -29,7 +29,7 @@ namespace Common {
  */
 class Fiber {
 public:
-    Fiber(std::function<void(void*)>&& entry_point_func, void* start_parameter);
+    Fiber(std::function<void()>&& entry_point_func);
     ~Fiber();
 
     Fiber(const Fiber&) = delete;
@@ -43,16 +43,13 @@ public:
     static void YieldTo(std::weak_ptr<Fiber> weak_from, Fiber& to);
     [[nodiscard]] static std::shared_ptr<Fiber> ThreadToFiber();
 
-    void SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param);
+    void SetRewindPoint(std::function<void()>&& rewind_func);
 
     void Rewind();
 
     /// Only call from main thread's fiber
     void Exit();
 
-    /// Changes the start parameter of the fiber. Has no effect if the fiber already started
-    void SetStartParameter(void* new_parameter);
-
 private:
     Fiber();
 
diff --git a/src/common/thread.cpp b/src/common/thread.cpp
index f932a7290..919e33af9 100644
--- a/src/common/thread.cpp
+++ b/src/common/thread.cpp
@@ -47,6 +47,9 @@ void SetCurrentThreadPriority(ThreadPriority new_priority) {
     case ThreadPriority::VeryHigh:
         windows_priority = THREAD_PRIORITY_HIGHEST;
         break;
+    case ThreadPriority::Critical:
+        windows_priority = THREAD_PRIORITY_TIME_CRITICAL;
+        break;
     default:
         windows_priority = THREAD_PRIORITY_NORMAL;
         break;
@@ -59,9 +62,10 @@ void SetCurrentThreadPriority(ThreadPriority new_priority) {
 void SetCurrentThreadPriority(ThreadPriority new_priority) {
     pthread_t this_thread = pthread_self();
 
-    s32 max_prio = sched_get_priority_max(SCHED_OTHER);
-    s32 min_prio = sched_get_priority_min(SCHED_OTHER);
-    u32 level = static_cast<u32>(new_priority) + 1;
+    const auto scheduling_type = SCHED_OTHER;
+    s32 max_prio = sched_get_priority_max(scheduling_type);
+    s32 min_prio = sched_get_priority_min(scheduling_type);
+    u32 level = std::max(static_cast<u32>(new_priority) + 1, 4U);
 
     struct sched_param params;
     if (max_prio > min_prio) {
@@ -70,7 +74,7 @@ void SetCurrentThreadPriority(ThreadPriority new_priority) {
         params.sched_priority = min_prio - ((min_prio - max_prio) * level) / 4;
     }
 
-    pthread_setschedparam(this_thread, SCHED_OTHER, &params);
+    pthread_setschedparam(this_thread, scheduling_type, &params);
 }
 
 #endif
diff --git a/src/common/thread.h b/src/common/thread.h
index a63122516..1552f58e0 100644
--- a/src/common/thread.h
+++ b/src/common/thread.h
@@ -92,6 +92,7 @@ enum class ThreadPriority : u32 {
     Normal = 1,
     High = 2,
     VeryHigh = 3,
+    Critical = 4,
 };
 
 void SetCurrentThreadPriority(ThreadPriority new_priority);
diff --git a/src/common/uint128.h b/src/common/uint128.h
index f890ffec2..199d0f55e 100644
--- a/src/common/uint128.h
+++ b/src/common/uint128.h
@@ -31,12 +31,17 @@ namespace Common {
     return _udiv128(r[1], r[0], d, &remainder);
 #endif
 #else
+#ifdef __SIZEOF_INT128__
+    const auto product = static_cast<unsigned __int128>(a) * static_cast<unsigned __int128>(b);
+    return static_cast<u64>(product / d);
+#else
     const u64 diva = a / d;
     const u64 moda = a % d;
     const u64 divb = b / d;
     const u64 modb = b % d;
     return diva * b + moda * divb + moda * modb / d;
 #endif
+#endif
 }
 
 // This function multiplies 2 u64 values and produces a u128 value;
diff --git a/src/common/x64/native_clock.cpp b/src/common/x64/native_clock.cpp
index 1b7194503..6aaa8cdf9 100644
--- a/src/common/x64/native_clock.cpp
+++ b/src/common/x64/native_clock.cpp
@@ -75,8 +75,8 @@ NativeClock::NativeClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequen
 }
 
 u64 NativeClock::GetRTSC() {
-    TimePoint new_time_point{};
     TimePoint current_time_point{};
+    TimePoint new_time_point{};
 
     current_time_point.pack = Common::AtomicLoad128(time_point.pack.data());
     do {
@@ -89,8 +89,7 @@ u64 NativeClock::GetRTSC() {
         new_time_point.inner.accumulated_ticks = current_time_point.inner.accumulated_ticks + diff;
     } while (!Common::AtomicCompareAndSwap(time_point.pack.data(), new_time_point.pack,
                                            current_time_point.pack, current_time_point.pack));
-    /// The clock cannot be more precise than the guest timer, remove the lower bits
-    return new_time_point.inner.accumulated_ticks & inaccuracy_mask;
+    return new_time_point.inner.accumulated_ticks;
 }
 
 void NativeClock::Pause(bool is_paused) {
diff --git a/src/common/x64/native_clock.h b/src/common/x64/native_clock.h
index 30d2ba2e9..38ae7a462 100644
--- a/src/common/x64/native_clock.h
+++ b/src/common/x64/native_clock.h
@@ -37,12 +37,8 @@ private:
         } inner;
     };
 
-    /// value used to reduce the native clocks accuracy as some apss rely on
-    /// undefined behavior where the level of accuracy in the clock shouldn't
-    /// be higher.
-    static constexpr u64 inaccuracy_mask = ~(UINT64_C(0x400) - 1);
-
     TimePoint time_point;
+
     // factors
     u64 clock_rtsc_factor{};
     u64 cpu_rtsc_factor{};
diff --git a/src/core/arm/arm_interface.cpp b/src/core/arm/arm_interface.cpp
index 8e095cdcd..0efc3732f 100644
--- a/src/core/arm/arm_interface.cpp
+++ b/src/core/arm/arm_interface.cpp
@@ -119,16 +119,23 @@ void ARM_Interface::Run() {
         }
         system.ExitDynarmicProfile();
 
-        // Notify the debugger and go to sleep if a breakpoint was hit.
-        if (Has(hr, breakpoint)) {
+        // Notify the debugger and go to sleep if a breakpoint was hit,
+        // or if the thread is unable to continue for any reason.
+        if (Has(hr, breakpoint) || Has(hr, no_execute)) {
             RewindBreakpointInstruction();
-            system.GetDebugger().NotifyThreadStopped(current_thread);
-            current_thread->RequestSuspend(SuspendType::Debug);
+            if (system.DebuggerEnabled()) {
+                system.GetDebugger().NotifyThreadStopped(current_thread);
+            }
+            current_thread->RequestSuspend(Kernel::SuspendType::Debug);
             break;
         }
+
+        // Notify the debugger and go to sleep if a watchpoint was hit.
         if (Has(hr, watchpoint)) {
             RewindBreakpointInstruction();
-            system.GetDebugger().NotifyThreadWatchpoint(current_thread, *HaltedWatchpoint());
+            if (system.DebuggerEnabled()) {
+                system.GetDebugger().NotifyThreadWatchpoint(current_thread, *HaltedWatchpoint());
+            }
             current_thread->RequestSuspend(SuspendType::Debug);
             break;
         }
diff --git a/src/core/arm/arm_interface.h b/src/core/arm/arm_interface.h
index 4e431e27a..8a066ed91 100644
--- a/src/core/arm/arm_interface.h
+++ b/src/core/arm/arm_interface.h
@@ -204,6 +204,7 @@ public:
     static constexpr Dynarmic::HaltReason svc_call = Dynarmic::HaltReason::UserDefined3;
     static constexpr Dynarmic::HaltReason breakpoint = Dynarmic::HaltReason::UserDefined4;
     static constexpr Dynarmic::HaltReason watchpoint = Dynarmic::HaltReason::UserDefined5;
+    static constexpr Dynarmic::HaltReason no_execute = Dynarmic::HaltReason::UserDefined6;
 
 protected:
     /// System context that this ARM interface is running under.
diff --git a/src/core/arm/dynarmic/arm_dynarmic_32.cpp b/src/core/arm/dynarmic/arm_dynarmic_32.cpp
index 8c90c8be0..10cf72a45 100644
--- a/src/core/arm/dynarmic/arm_dynarmic_32.cpp
+++ b/src/core/arm/dynarmic/arm_dynarmic_32.cpp
@@ -48,6 +48,12 @@ public:
         CheckMemoryAccess(vaddr, 8, Kernel::DebugWatchpointType::Read);
         return memory.Read64(vaddr);
     }
+    std::optional<u32> MemoryReadCode(u32 vaddr) override {
+        if (!memory.IsValidVirtualAddressRange(vaddr, sizeof(u32))) {
+            return std::nullopt;
+        }
+        return MemoryRead32(vaddr);
+    }
 
     void MemoryWrite8(u32 vaddr, u8 value) override {
         if (CheckMemoryAccess(vaddr, 1, Kernel::DebugWatchpointType::Write)) {
@@ -89,21 +95,28 @@ public:
 
     void InterpreterFallback(u32 pc, std::size_t num_instructions) override {
         parent.LogBacktrace();
-        UNIMPLEMENTED_MSG("This should never happen, pc = {:08X}, code = {:08X}", pc,
-                          MemoryReadCode(pc));
+        LOG_ERROR(Core_ARM,
+                  "Unimplemented instruction @ 0x{:X} for {} instructions (instr = {:08X})", pc,
+                  num_instructions, MemoryRead32(pc));
     }
 
     void ExceptionRaised(u32 pc, Dynarmic::A32::Exception exception) override {
-        if (debugger_enabled) {
-            parent.SaveContext(parent.breakpoint_context);
-            parent.jit.load()->HaltExecution(ARM_Interface::breakpoint);
+        switch (exception) {
+        case Dynarmic::A32::Exception::NoExecuteFault:
+            LOG_CRITICAL(Core_ARM, "Cannot execute instruction at unmapped address {:#08x}", pc);
+            ReturnException(pc, ARM_Interface::no_execute);
             return;
-        }
+        default:
+            if (debugger_enabled) {
+                ReturnException(pc, ARM_Interface::breakpoint);
+                return;
+            }
 
-        parent.LogBacktrace();
-        LOG_CRITICAL(Core_ARM,
-                     "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X}, thumb = {})",
-                     exception, pc, MemoryReadCode(pc), parent.IsInThumbMode());
+            parent.LogBacktrace();
+            LOG_CRITICAL(Core_ARM,
+                         "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X}, thumb = {})",
+                         exception, pc, MemoryRead32(pc), parent.IsInThumbMode());
+        }
     }
 
     void CallSVC(u32 swi) override {
@@ -141,15 +154,20 @@ public:
 
         const auto match{parent.MatchingWatchpoint(addr, size, type)};
         if (match) {
-            parent.SaveContext(parent.breakpoint_context);
-            parent.jit.load()->HaltExecution(ARM_Interface::watchpoint);
             parent.halted_watchpoint = match;
+            ReturnException(parent.jit.load()->Regs()[15], ARM_Interface::watchpoint);
             return false;
         }
 
         return true;
     }
 
+    void ReturnException(u32 pc, Dynarmic::HaltReason hr) {
+        parent.SaveContext(parent.breakpoint_context);
+        parent.breakpoint_context.cpu_registers[15] = pc;
+        parent.jit.load()->HaltExecution(hr);
+    }
+
     ARM_Dynarmic_32& parent;
     Core::Memory::Memory& memory;
     std::size_t num_interpreted_instructions{};
diff --git a/src/core/arm/dynarmic/arm_dynarmic_64.cpp b/src/core/arm/dynarmic/arm_dynarmic_64.cpp
index 4370ca294..92266aa9e 100644
--- a/src/core/arm/dynarmic/arm_dynarmic_64.cpp
+++ b/src/core/arm/dynarmic/arm_dynarmic_64.cpp
@@ -52,6 +52,12 @@ public:
         CheckMemoryAccess(vaddr, 16, Kernel::DebugWatchpointType::Read);
         return {memory.Read64(vaddr), memory.Read64(vaddr + 8)};
     }
+    std::optional<u32> MemoryReadCode(u64 vaddr) override {
+        if (!memory.IsValidVirtualAddressRange(vaddr, sizeof(u32))) {
+            return std::nullopt;
+        }
+        return MemoryRead32(vaddr);
+    }
 
     void MemoryWrite8(u64 vaddr, u8 value) override {
         if (CheckMemoryAccess(vaddr, 1, Kernel::DebugWatchpointType::Write)) {
@@ -105,7 +111,7 @@ public:
         parent.LogBacktrace();
         LOG_ERROR(Core_ARM,
                   "Unimplemented instruction @ 0x{:X} for {} instructions (instr = {:08X})", pc,
-                  num_instructions, MemoryReadCode(pc));
+                  num_instructions, MemoryRead32(pc));
     }
 
     void InstructionCacheOperationRaised(Dynarmic::A64::InstructionCacheOperation op,
@@ -138,16 +144,19 @@ public:
         case Dynarmic::A64::Exception::SendEventLocal:
         case Dynarmic::A64::Exception::Yield:
             return;
+        case Dynarmic::A64::Exception::NoExecuteFault:
+            LOG_CRITICAL(Core_ARM, "Cannot execute instruction at unmapped address {:#016x}", pc);
+            ReturnException(pc, ARM_Interface::no_execute);
+            return;
         default:
             if (debugger_enabled) {
-                parent.SaveContext(parent.breakpoint_context);
-                parent.jit.load()->HaltExecution(ARM_Interface::breakpoint);
+                ReturnException(pc, ARM_Interface::breakpoint);
                 return;
             }
 
             parent.LogBacktrace();
-            ASSERT_MSG(false, "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X})",
-                       static_cast<std::size_t>(exception), pc, MemoryReadCode(pc));
+            LOG_CRITICAL(Core_ARM, "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X})",
+                         static_cast<std::size_t>(exception), pc, MemoryRead32(pc));
         }
     }
 
@@ -188,15 +197,20 @@ public:
 
         const auto match{parent.MatchingWatchpoint(addr, size, type)};
         if (match) {
-            parent.SaveContext(parent.breakpoint_context);
-            parent.jit.load()->HaltExecution(ARM_Interface::watchpoint);
             parent.halted_watchpoint = match;
+            ReturnException(parent.jit.load()->GetPC(), ARM_Interface::watchpoint);
             return false;
         }
 
         return true;
     }
 
+    void ReturnException(u64 pc, Dynarmic::HaltReason hr) {
+        parent.SaveContext(parent.breakpoint_context);
+        parent.breakpoint_context.pc = pc;
+        parent.jit.load()->HaltExecution(hr);
+    }
+
     ARM_Dynarmic_64& parent;
     Core::Memory::Memory& memory;
     u64 tpidrro_el0 = 0;
diff --git a/src/core/core_timing.cpp b/src/core/core_timing.cpp
index 29e7dba9b..140578069 100644
--- a/src/core/core_timing.cpp
+++ b/src/core/core_timing.cpp
@@ -6,7 +6,9 @@
 #include <string>
 #include <tuple>
 
+#include "common/logging/log.h"
 #include "common/microprofile.h"
+#include "common/thread.h"
 #include "core/core_timing.h"
 #include "core/core_timing_util.h"
 #include "core/hardware_properties.h"
@@ -41,11 +43,11 @@ CoreTiming::CoreTiming()
 
 CoreTiming::~CoreTiming() = default;
 
-void CoreTiming::ThreadEntry(CoreTiming& instance) {
-    constexpr char name[] = "yuzu:HostTiming";
-    MicroProfileOnThreadCreate(name);
-    Common::SetCurrentThreadName(name);
-    Common::SetCurrentThreadPriority(Common::ThreadPriority::VeryHigh);
+void CoreTiming::ThreadEntry(CoreTiming& instance, size_t id) {
+    const std::string name = "yuzu:HostTiming_" + std::to_string(id);
+    MicroProfileOnThreadCreate(name.c_str());
+    Common::SetCurrentThreadName(name.c_str());
+    Common::SetCurrentThreadPriority(Common::ThreadPriority::Critical);
     instance.on_thread_init();
     instance.ThreadLoop();
     MicroProfileOnThreadExit();
@@ -59,68 +61,97 @@ void CoreTiming::Initialize(std::function<void()>&& on_thread_init_) {
     const auto empty_timed_callback = [](std::uintptr_t, std::chrono::nanoseconds) {};
     ev_lost = CreateEvent("_lost_event", empty_timed_callback);
     if (is_multicore) {
-        timer_thread = std::make_unique<std::thread>(ThreadEntry, std::ref(*this));
+        const auto hardware_concurrency = std::thread::hardware_concurrency();
+        size_t id = 0;
+        worker_threads.emplace_back(ThreadEntry, std::ref(*this), id++);
+        if (hardware_concurrency > 8) {
+            worker_threads.emplace_back(ThreadEntry, std::ref(*this), id++);
+        }
     }
 }
 
 void CoreTiming::Shutdown() {
-    paused = true;
+    is_paused = true;
     shutting_down = true;
-    pause_event.Set();
-    event.Set();
-    if (timer_thread) {
-        timer_thread->join();
+    std::atomic_thread_fence(std::memory_order_release);
+
+    event_cv.notify_all();
+    wait_pause_cv.notify_all();
+    for (auto& thread : worker_threads) {
+        thread.join();
     }
+    worker_threads.clear();
     ClearPendingEvents();
-    timer_thread.reset();
     has_started = false;
 }
 
-void CoreTiming::Pause(bool is_paused) {
-    paused = is_paused;
-    pause_event.Set();
+void CoreTiming::Pause(bool is_paused_) {
+    std::unique_lock main_lock(event_mutex);
+    if (is_paused_ == paused_state.load(std::memory_order_relaxed)) {
+        return;
+    }
+    if (is_multicore) {
+        is_paused = is_paused_;
+        event_cv.notify_all();
+        if (!is_paused_) {
+            wait_pause_cv.notify_all();
+        }
+    }
+    paused_state.store(is_paused_, std::memory_order_relaxed);
 }
 
-void CoreTiming::SyncPause(bool is_paused) {
-    if (is_paused == paused && paused_set == paused) {
+void CoreTiming::SyncPause(bool is_paused_) {
+    std::unique_lock main_lock(event_mutex);
+    if (is_paused_ == paused_state.load(std::memory_order_relaxed)) {
         return;
     }
-    Pause(is_paused);
-    if (timer_thread) {
-        if (!is_paused) {
-            pause_event.Set();
+
+    if (is_multicore) {
+        is_paused = is_paused_;
+        event_cv.notify_all();
+        if (!is_paused_) {
+            wait_pause_cv.notify_all();
+        }
+    }
+    paused_state.store(is_paused_, std::memory_order_relaxed);
+    if (is_multicore) {
+        if (is_paused_) {
+            wait_signal_cv.wait(main_lock, [this] { return pause_count == worker_threads.size(); });
+        } else {
+            wait_signal_cv.wait(main_lock, [this] { return pause_count == 0; });
         }
-        event.Set();
-        while (paused_set != is_paused)
-            ;
     }
 }
 
 bool CoreTiming::IsRunning() const {
-    return !paused_set;
+    return !paused_state.load(std::memory_order_acquire);
 }
 
 bool CoreTiming::HasPendingEvents() const {
-    return !(wait_set && event_queue.empty());
+    std::unique_lock main_lock(event_mutex);
+    return !event_queue.empty() || pending_events.load(std::memory_order_relaxed) != 0;
 }
 
 void CoreTiming::ScheduleEvent(std::chrono::nanoseconds ns_into_future,
                                const std::shared_ptr<EventType>& event_type,
                                std::uintptr_t user_data) {
-    {
-        std::scoped_lock scope{basic_lock};
-        const u64 timeout = static_cast<u64>((GetGlobalTimeNs() + ns_into_future).count());
 
-        event_queue.emplace_back(Event{timeout, event_fifo_id++, user_data, event_type});
+    std::unique_lock main_lock(event_mutex);
+    const u64 timeout = static_cast<u64>((GetGlobalTimeNs() + ns_into_future).count());
+
+    event_queue.emplace_back(Event{timeout, event_fifo_id++, user_data, event_type});
+    pending_events.fetch_add(1, std::memory_order_relaxed);
 
-        std::push_heap(event_queue.begin(), event_queue.end(), std::greater<>());
+    std::push_heap(event_queue.begin(), event_queue.end(), std::greater<>());
+
+    if (is_multicore) {
+        event_cv.notify_one();
     }
-    event.Set();
 }
 
 void CoreTiming::UnscheduleEvent(const std::shared_ptr<EventType>& event_type,
                                  std::uintptr_t user_data) {
-    std::scoped_lock scope{basic_lock};
+    std::unique_lock main_lock(event_mutex);
     const auto itr = std::remove_if(event_queue.begin(), event_queue.end(), [&](const Event& e) {
         return e.type.lock().get() == event_type.get() && e.user_data == user_data;
     });
@@ -129,6 +160,7 @@ void CoreTiming::UnscheduleEvent(const std::shared_ptr<EventType>& event_type,
     if (itr != event_queue.end()) {
         event_queue.erase(itr, event_queue.end());
         std::make_heap(event_queue.begin(), event_queue.end(), std::greater<>());
+        pending_events.fetch_sub(1, std::memory_order_relaxed);
     }
 }
 
@@ -168,11 +200,12 @@ u64 CoreTiming::GetClockTicks() const {
 }
 
 void CoreTiming::ClearPendingEvents() {
+    std::unique_lock main_lock(event_mutex);
     event_queue.clear();
 }
 
 void CoreTiming::RemoveEvent(const std::shared_ptr<EventType>& event_type) {
-    std::scoped_lock lock{basic_lock};
+    std::unique_lock main_lock(event_mutex);
 
     const auto itr = std::remove_if(event_queue.begin(), event_queue.end(), [&](const Event& e) {
         return e.type.lock().get() == event_type.get();
@@ -186,21 +219,28 @@ void CoreTiming::RemoveEvent(const std::shared_ptr<EventType>& event_type) {
 }
 
 std::optional<s64> CoreTiming::Advance() {
-    std::scoped_lock lock{advance_lock, basic_lock};
     global_timer = GetGlobalTimeNs().count();
 
+    std::unique_lock main_lock(event_mutex);
     while (!event_queue.empty() && event_queue.front().time <= global_timer) {
         Event evt = std::move(event_queue.front());
         std::pop_heap(event_queue.begin(), event_queue.end(), std::greater<>());
         event_queue.pop_back();
-        basic_lock.unlock();
 
         if (const auto event_type{evt.type.lock()}) {
-            event_type->callback(
-                evt.user_data, std::chrono::nanoseconds{static_cast<s64>(global_timer - evt.time)});
+            sequence_mutex.lock();
+            event_mutex.unlock();
+
+            event_type->guard.lock();
+            sequence_mutex.unlock();
+            const s64 delay = static_cast<s64>(GetGlobalTimeNs().count() - evt.time);
+            event_type->callback(evt.user_data, std::chrono::nanoseconds{delay});
+            event_type->guard.unlock();
+
+            event_mutex.lock();
+            pending_events.fetch_sub(1, std::memory_order_relaxed);
         }
 
-        basic_lock.lock();
         global_timer = GetGlobalTimeNs().count();
     }
 
@@ -213,26 +253,34 @@ std::optional<s64> CoreTiming::Advance() {
 }
 
 void CoreTiming::ThreadLoop() {
+    const auto predicate = [this] { return !event_queue.empty() || is_paused; };
     has_started = true;
     while (!shutting_down) {
-        while (!paused) {
-            paused_set = false;
+        while (!is_paused && !shutting_down) {
             const auto next_time = Advance();
             if (next_time) {
                 if (*next_time > 0) {
                     std::chrono::nanoseconds next_time_ns = std::chrono::nanoseconds(*next_time);
-                    event.WaitFor(next_time_ns);
+                    std::unique_lock main_lock(event_mutex);
+                    event_cv.wait_for(main_lock, next_time_ns, predicate);
                 }
             } else {
-                wait_set = true;
-                event.Wait();
+                std::unique_lock main_lock(event_mutex);
+                event_cv.wait(main_lock, predicate);
             }
-            wait_set = false;
         }
-        paused_set = true;
-        clock->Pause(true);
-        pause_event.Wait();
-        clock->Pause(false);
+        std::unique_lock main_lock(event_mutex);
+        pause_count++;
+        if (pause_count == worker_threads.size()) {
+            clock->Pause(true);
+            wait_signal_cv.notify_all();
+        }
+        wait_pause_cv.wait(main_lock, [this] { return !is_paused || shutting_down; });
+        pause_count--;
+        if (pause_count == 0) {
+            clock->Pause(false);
+            wait_signal_cv.notify_all();
+        }
     }
 }
 
diff --git a/src/core/core_timing.h b/src/core/core_timing.h
index d27773009..a86553e08 100644
--- a/src/core/core_timing.h
+++ b/src/core/core_timing.h
@@ -5,6 +5,7 @@
 
 #include <atomic>
 #include <chrono>
+#include <condition_variable>
 #include <functional>
 #include <memory>
 #include <mutex>
@@ -14,7 +15,6 @@
 #include <vector>
 
 #include "common/common_types.h"
-#include "common/thread.h"
 #include "common/wall_clock.h"
 
 namespace Core::Timing {
@@ -32,6 +32,7 @@ struct EventType {
     TimedCallback callback;
     /// A pointer to the name of the event.
     const std::string name;
+    mutable std::mutex guard;
 };
 
 /**
@@ -131,7 +132,7 @@ private:
     /// Clear all pending events. This should ONLY be done on exit.
     void ClearPendingEvents();
 
-    static void ThreadEntry(CoreTiming& instance);
+    static void ThreadEntry(CoreTiming& instance, size_t id);
     void ThreadLoop();
 
     std::unique_ptr<Common::WallClock> clock;
@@ -144,21 +145,25 @@ private:
     // accomodated by the standard adaptor class.
     std::vector<Event> event_queue;
     u64 event_fifo_id = 0;
+    std::atomic<size_t> pending_events{};
 
     std::shared_ptr<EventType> ev_lost;
-    Common::Event event{};
-    Common::Event pause_event{};
-    std::mutex basic_lock;
-    std::mutex advance_lock;
-    std::unique_ptr<std::thread> timer_thread;
-    std::atomic<bool> paused{};
-    std::atomic<bool> paused_set{};
-    std::atomic<bool> wait_set{};
-    std::atomic<bool> shutting_down{};
     std::atomic<bool> has_started{};
     std::function<void()> on_thread_init{};
 
+    std::vector<std::thread> worker_threads;
+
+    std::condition_variable event_cv;
+    std::condition_variable wait_pause_cv;
+    std::condition_variable wait_signal_cv;
+    mutable std::mutex event_mutex;
+    mutable std::mutex sequence_mutex;
+
+    std::atomic<bool> paused_state{};
+    bool is_paused{};
+    bool shutting_down{};
     bool is_multicore{};
+    size_t pause_count{};
 
     /// Cycle timing
     u64 ticks{};
diff --git a/src/core/cpu_manager.cpp b/src/core/cpu_manager.cpp
index fd6928105..37d3d83b9 100644
--- a/src/core/cpu_manager.cpp
+++ b/src/core/cpu_manager.cpp
@@ -41,51 +41,32 @@ void CpuManager::Shutdown() {
     }
 }
 
-std::function<void(void*)> CpuManager::GetGuestThreadStartFunc() {
-    return GuestThreadFunction;
-}
-
-std::function<void(void*)> CpuManager::GetIdleThreadStartFunc() {
-    return IdleThreadFunction;
-}
-
-std::function<void(void*)> CpuManager::GetShutdownThreadStartFunc() {
-    return ShutdownThreadFunction;
-}
-
-void CpuManager::GuestThreadFunction(void* cpu_manager_) {
-    CpuManager* cpu_manager = static_cast<CpuManager*>(cpu_manager_);
-    if (cpu_manager->is_multicore) {
-        cpu_manager->MultiCoreRunGuestThread();
+void CpuManager::GuestThreadFunction() {
+    if (is_multicore) {
+        MultiCoreRunGuestThread();
     } else {
-        cpu_manager->SingleCoreRunGuestThread();
+        SingleCoreRunGuestThread();
     }
 }
 
-void CpuManager::GuestRewindFunction(void* cpu_manager_) {
-    CpuManager* cpu_manager = static_cast<CpuManager*>(cpu_manager_);
-    if (cpu_manager->is_multicore) {
-        cpu_manager->MultiCoreRunGuestLoop();
+void CpuManager::GuestRewindFunction() {
+    if (is_multicore) {
+        MultiCoreRunGuestLoop();
     } else {
-        cpu_manager->SingleCoreRunGuestLoop();
+        SingleCoreRunGuestLoop();
     }
 }
 
-void CpuManager::IdleThreadFunction(void* cpu_manager_) {
-    CpuManager* cpu_manager = static_cast<CpuManager*>(cpu_manager_);
-    if (cpu_manager->is_multicore) {
-        cpu_manager->MultiCoreRunIdleThread();
+void CpuManager::IdleThreadFunction() {
+    if (is_multicore) {
+        MultiCoreRunIdleThread();
     } else {
-        cpu_manager->SingleCoreRunIdleThread();
+        SingleCoreRunIdleThread();
     }
 }
 
-void CpuManager::ShutdownThreadFunction(void* cpu_manager) {
-    static_cast<CpuManager*>(cpu_manager)->ShutdownThread();
-}
-
-void* CpuManager::GetStartFuncParameter() {
-    return this;
+void CpuManager::ShutdownThreadFunction() {
+    ShutdownThread();
 }
 
 ///////////////////////////////////////////////////////////////////////////////
@@ -97,7 +78,7 @@ void CpuManager::MultiCoreRunGuestThread() {
     kernel.CurrentScheduler()->OnThreadStart();
     auto* thread = kernel.CurrentScheduler()->GetSchedulerCurrentThread();
     auto& host_context = thread->GetHostContext();
-    host_context->SetRewindPoint(GuestRewindFunction, this);
+    host_context->SetRewindPoint([this] { GuestRewindFunction(); });
     MultiCoreRunGuestLoop();
 }
 
@@ -134,7 +115,7 @@ void CpuManager::SingleCoreRunGuestThread() {
     kernel.CurrentScheduler()->OnThreadStart();
     auto* thread = kernel.CurrentScheduler()->GetSchedulerCurrentThread();
     auto& host_context = thread->GetHostContext();
-    host_context->SetRewindPoint(GuestRewindFunction, this);
+    host_context->SetRewindPoint([this] { GuestRewindFunction(); });
     SingleCoreRunGuestLoop();
 }
 
@@ -194,7 +175,9 @@ void CpuManager::PreemptSingleCore(bool from_running_enviroment) {
     {
         auto& scheduler = system.Kernel().Scheduler(current_core);
         scheduler.Reload(scheduler.GetSchedulerCurrentThread());
-        idle_count = 0;
+        if (!scheduler.IsIdle()) {
+            idle_count = 0;
+        }
     }
 }
 
diff --git a/src/core/cpu_manager.h b/src/core/cpu_manager.h
index f0751fc58..76dc58ee1 100644
--- a/src/core/cpu_manager.h
+++ b/src/core/cpu_manager.h
@@ -50,10 +50,15 @@ public:
     void Initialize();
     void Shutdown();
 
-    static std::function<void(void*)> GetGuestThreadStartFunc();
-    static std::function<void(void*)> GetIdleThreadStartFunc();
-    static std::function<void(void*)> GetShutdownThreadStartFunc();
-    void* GetStartFuncParameter();
+    std::function<void()> GetGuestThreadStartFunc() {
+        return [this] { GuestThreadFunction(); };
+    }
+    std::function<void()> GetIdleThreadStartFunc() {
+        return [this] { IdleThreadFunction(); };
+    }
+    std::function<void()> GetShutdownThreadStartFunc() {
+        return [this] { ShutdownThreadFunction(); };
+    }
 
     void PreemptSingleCore(bool from_running_enviroment = true);
 
@@ -62,10 +67,10 @@ public:
     }
 
 private:
-    static void GuestThreadFunction(void* cpu_manager);
-    static void GuestRewindFunction(void* cpu_manager);
-    static void IdleThreadFunction(void* cpu_manager);
-    static void ShutdownThreadFunction(void* cpu_manager);
+    void GuestThreadFunction();
+    void GuestRewindFunction();
+    void IdleThreadFunction();
+    void ShutdownThreadFunction();
 
     void MultiCoreRunGuestThread();
     void MultiCoreRunGuestLoop();
diff --git a/src/core/debugger/gdbstub_arch.cpp b/src/core/debugger/gdbstub_arch.cpp
index 750c353b9..4bef09bd7 100644
--- a/src/core/debugger/gdbstub_arch.cpp
+++ b/src/core/debugger/gdbstub_arch.cpp
@@ -191,8 +191,10 @@ std::string GDBStubA64::RegRead(const Kernel::KThread* thread, size_t id) const
     const auto& gprs{context.cpu_registers};
     const auto& fprs{context.vector_registers};
 
-    if (id <= SP_REGISTER) {
+    if (id < SP_REGISTER) {
         return ValueToHex(gprs[id]);
+    } else if (id == SP_REGISTER) {
+        return ValueToHex(context.sp);
     } else if (id == PC_REGISTER) {
         return ValueToHex(context.pc);
     } else if (id == PSTATE_REGISTER) {
@@ -215,8 +217,10 @@ void GDBStubA64::RegWrite(Kernel::KThread* thread, size_t id, std::string_view v
 
     auto& context{thread->GetContext64()};
 
-    if (id <= SP_REGISTER) {
+    if (id < SP_REGISTER) {
         context.cpu_registers[id] = HexToValue<u64>(value);
+    } else if (id == SP_REGISTER) {
+        context.sp = HexToValue<u64>(value);
     } else if (id == PC_REGISTER) {
         context.pc = HexToValue<u64>(value);
     } else if (id == PSTATE_REGISTER) {
diff --git a/src/core/hle/kernel/k_scheduler.cpp b/src/core/hle/kernel/k_scheduler.cpp
index d586b3f5c..d599d2bcb 100644
--- a/src/core/hle/kernel/k_scheduler.cpp
+++ b/src/core/hle/kernel/k_scheduler.cpp
@@ -622,7 +622,7 @@ void KScheduler::YieldToAnyThread(KernelCore& kernel) {
 }
 
 KScheduler::KScheduler(Core::System& system_, s32 core_id_) : system{system_}, core_id{core_id_} {
-    switch_fiber = std::make_shared<Common::Fiber>(OnSwitch, this);
+    switch_fiber = std::make_shared<Common::Fiber>([this] { SwitchToCurrent(); });
     state.needs_scheduling.store(true);
     state.interrupt_task_thread_runnable = false;
     state.should_count_idle = false;
@@ -778,11 +778,6 @@ void KScheduler::ScheduleImpl() {
     next_scheduler.SwitchContextStep2();
 }
 
-void KScheduler::OnSwitch(void* this_scheduler) {
-    KScheduler* sched = static_cast<KScheduler*>(this_scheduler);
-    sched->SwitchToCurrent();
-}
-
 void KScheduler::SwitchToCurrent() {
     while (true) {
         {
diff --git a/src/core/hle/kernel/k_scheduler.h b/src/core/hle/kernel/k_scheduler.h
index 3f90656ee..6a4760eca 100644
--- a/src/core/hle/kernel/k_scheduler.h
+++ b/src/core/hle/kernel/k_scheduler.h
@@ -55,6 +55,11 @@ public:
         return idle_thread;
     }
 
+    /// Returns true if the scheduler is idle
+    [[nodiscard]] bool IsIdle() const {
+        return GetSchedulerCurrentThread() == idle_thread;
+    }
+
     /// Gets the timestamp for the last context switch in ticks.
     [[nodiscard]] u64 GetLastContextSwitchTicks() const;
 
@@ -165,7 +170,6 @@ private:
      */
     void UpdateLastContextSwitchTime(KThread* thread, KProcess* process);
 
-    static void OnSwitch(void* this_scheduler);
     void SwitchToCurrent();
 
     KThread* prev_thread{};
diff --git a/src/core/hle/kernel/k_thread.cpp b/src/core/hle/kernel/k_thread.cpp
index 8d7faa662..23bf7425a 100644
--- a/src/core/hle/kernel/k_thread.cpp
+++ b/src/core/hle/kernel/k_thread.cpp
@@ -246,14 +246,12 @@ Result KThread::Initialize(KThreadFunction func, uintptr_t arg, VAddr user_stack
 
 Result KThread::InitializeThread(KThread* thread, KThreadFunction func, uintptr_t arg,
                                  VAddr user_stack_top, s32 prio, s32 core, KProcess* owner,
-                                 ThreadType type, std::function<void(void*)>&& init_func,
-                                 void* init_func_parameter) {
+                                 ThreadType type, std::function<void()>&& init_func) {
     // Initialize the thread.
     R_TRY(thread->Initialize(func, arg, user_stack_top, prio, core, owner, type));
 
     // Initialize emulation parameters.
-    thread->host_context =
-        std::make_shared<Common::Fiber>(std::move(init_func), init_func_parameter);
+    thread->host_context = std::make_shared<Common::Fiber>(std::move(init_func));
     thread->is_single_core = !Settings::values.use_multi_core.GetValue();
 
     return ResultSuccess;
@@ -265,15 +263,13 @@ Result KThread::InitializeDummyThread(KThread* thread) {
 
 Result KThread::InitializeIdleThread(Core::System& system, KThread* thread, s32 virt_core) {
     return InitializeThread(thread, {}, {}, {}, IdleThreadPriority, virt_core, {}, ThreadType::Main,
-                            Core::CpuManager::GetIdleThreadStartFunc(),
-                            system.GetCpuManager().GetStartFuncParameter());
+                            system.GetCpuManager().GetIdleThreadStartFunc());
 }
 
 Result KThread::InitializeHighPriorityThread(Core::System& system, KThread* thread,
                                              KThreadFunction func, uintptr_t arg, s32 virt_core) {
     return InitializeThread(thread, func, arg, {}, {}, virt_core, nullptr, ThreadType::HighPriority,
-                            Core::CpuManager::GetShutdownThreadStartFunc(),
-                            system.GetCpuManager().GetStartFuncParameter());
+                            system.GetCpuManager().GetShutdownThreadStartFunc());
 }
 
 Result KThread::InitializeUserThread(Core::System& system, KThread* thread, KThreadFunction func,
@@ -281,8 +277,7 @@ Result KThread::InitializeUserThread(Core::System& system, KThread* thread, KThr
                                      KProcess* owner) {
     system.Kernel().GlobalSchedulerContext().AddThread(thread);
     return InitializeThread(thread, func, arg, user_stack_top, prio, virt_core, owner,
-                            ThreadType::User, Core::CpuManager::GetGuestThreadStartFunc(),
-                            system.GetCpuManager().GetStartFuncParameter());
+                            ThreadType::User, system.GetCpuManager().GetGuestThreadStartFunc());
 }
 
 void KThread::PostDestroy(uintptr_t arg) {
diff --git a/src/core/hle/kernel/k_thread.h b/src/core/hle/kernel/k_thread.h
index 94c4cd1c8..28cd7ecb0 100644
--- a/src/core/hle/kernel/k_thread.h
+++ b/src/core/hle/kernel/k_thread.h
@@ -729,8 +729,7 @@ private:
     [[nodiscard]] static Result InitializeThread(KThread* thread, KThreadFunction func,
                                                  uintptr_t arg, VAddr user_stack_top, s32 prio,
                                                  s32 core, KProcess* owner, ThreadType type,
-                                                 std::function<void(void*)>&& init_func,
-                                                 void* init_func_parameter);
+                                                 std::function<void()>&& init_func);
 
     static void RestorePriority(KernelCore& kernel_ctx, KThread* thread);
 
diff --git a/src/tests/common/fibers.cpp b/src/tests/common/fibers.cpp
index cfc84d423..4e29f9199 100644
--- a/src/tests/common/fibers.cpp
+++ b/src/tests/common/fibers.cpp
@@ -43,7 +43,15 @@ class TestControl1 {
 public:
     TestControl1() = default;
 
-    void DoWork();
+    void DoWork() {
+        const u32 id = thread_ids.Get();
+        u32 value = items[id];
+        for (u32 i = 0; i < id; i++) {
+            value++;
+        }
+        results[id] = value;
+        Fiber::YieldTo(work_fibers[id], *thread_fibers[id]);
+    }
 
     void ExecuteThread(u32 id);
 
@@ -54,35 +62,16 @@ public:
     std::vector<u32> results;
 };
 
-static void WorkControl1(void* control) {
-    auto* test_control = static_cast<TestControl1*>(control);
-    test_control->DoWork();
-}
-
-void TestControl1::DoWork() {
-    const u32 id = thread_ids.Get();
-    u32 value = items[id];
-    for (u32 i = 0; i < id; i++) {
-        value++;
-    }
-    results[id] = value;
-    Fiber::YieldTo(work_fibers[id], *thread_fibers[id]);
-}
-
 void TestControl1::ExecuteThread(u32 id) {
     thread_ids.Register(id);
     auto thread_fiber = Fiber::ThreadToFiber();
     thread_fibers[id] = thread_fiber;
-    work_fibers[id] = std::make_shared<Fiber>(std::function<void(void*)>{WorkControl1}, this);
+    work_fibers[id] = std::make_shared<Fiber>([this] { DoWork(); });
     items[id] = rand() % 256;
     Fiber::YieldTo(thread_fibers[id], *work_fibers[id]);
     thread_fibers[id]->Exit();
 }
 
-static void ThreadStart1(u32 id, TestControl1& test_control) {
-    test_control.ExecuteThread(id);
-}
-
 /** This test checks for fiber setup configuration and validates that fibers are
  *  doing all the work required.
  */
@@ -95,7 +84,7 @@ TEST_CASE("Fibers::Setup", "[common]") {
     test_control.results.resize(num_threads, 0);
     std::vector<std::thread> threads;
     for (u32 i = 0; i < num_threads; i++) {
-        threads.emplace_back(ThreadStart1, i, std::ref(test_control));
+        threads.emplace_back([&test_control, i] { test_control.ExecuteThread(i); });
     }
     for (u32 i = 0; i < num_threads; i++) {
         threads[i].join();
@@ -167,21 +156,6 @@ public:
     std::shared_ptr<Common::Fiber> fiber3;
 };
 
-static void WorkControl2_1(void* control) {
-    auto* test_control = static_cast<TestControl2*>(control);
-    test_control->DoWork1();
-}
-
-static void WorkControl2_2(void* control) {
-    auto* test_control = static_cast<TestControl2*>(control);
-    test_control->DoWork2();
-}
-
-static void WorkControl2_3(void* control) {
-    auto* test_control = static_cast<TestControl2*>(control);
-    test_control->DoWork3();
-}
-
 void TestControl2::ExecuteThread(u32 id) {
     thread_ids.Register(id);
     auto thread_fiber = Fiber::ThreadToFiber();
@@ -193,18 +167,6 @@ void TestControl2::Exit() {
     thread_fibers[id]->Exit();
 }
 
-static void ThreadStart2_1(u32 id, TestControl2& test_control) {
-    test_control.ExecuteThread(id);
-    test_control.CallFiber1();
-    test_control.Exit();
-}
-
-static void ThreadStart2_2(u32 id, TestControl2& test_control) {
-    test_control.ExecuteThread(id);
-    test_control.CallFiber2();
-    test_control.Exit();
-}
-
 /** This test checks for fiber thread exchange configuration and validates that fibers are
  *  that a fiber has been successfully transferred from one thread to another and that the TLS
  *  region of the thread is kept while changing fibers.
@@ -212,14 +174,19 @@ static void ThreadStart2_2(u32 id, TestControl2& test_control) {
 TEST_CASE("Fibers::InterExchange", "[common]") {
     TestControl2 test_control{};
     test_control.thread_fibers.resize(2);
-    test_control.fiber1 =
-        std::make_shared<Fiber>(std::function<void(void*)>{WorkControl2_1}, &test_control);
-    test_control.fiber2 =
-        std::make_shared<Fiber>(std::function<void(void*)>{WorkControl2_2}, &test_control);
-    test_control.fiber3 =
-        std::make_shared<Fiber>(std::function<void(void*)>{WorkControl2_3}, &test_control);
-    std::thread thread1(ThreadStart2_1, 0, std::ref(test_control));
-    std::thread thread2(ThreadStart2_2, 1, std::ref(test_control));
+    test_control.fiber1 = std::make_shared<Fiber>([&test_control] { test_control.DoWork1(); });
+    test_control.fiber2 = std::make_shared<Fiber>([&test_control] { test_control.DoWork2(); });
+    test_control.fiber3 = std::make_shared<Fiber>([&test_control] { test_control.DoWork3(); });
+    std::thread thread1{[&test_control] {
+        test_control.ExecuteThread(0);
+        test_control.CallFiber1();
+        test_control.Exit();
+    }};
+    std::thread thread2{[&test_control] {
+        test_control.ExecuteThread(1);
+        test_control.CallFiber2();
+        test_control.Exit();
+    }};
     thread1.join();
     thread2.join();
     REQUIRE(test_control.assert1);
@@ -270,16 +237,6 @@ public:
     std::shared_ptr<Common::Fiber> fiber2;
 };
 
-static void WorkControl3_1(void* control) {
-    auto* test_control = static_cast<TestControl3*>(control);
-    test_control->DoWork1();
-}
-
-static void WorkControl3_2(void* control) {
-    auto* test_control = static_cast<TestControl3*>(control);
-    test_control->DoWork2();
-}
-
 void TestControl3::ExecuteThread(u32 id) {
     thread_ids.Register(id);
     auto thread_fiber = Fiber::ThreadToFiber();
@@ -291,12 +248,6 @@ void TestControl3::Exit() {
     thread_fibers[id]->Exit();
 }
 
-static void ThreadStart3(u32 id, TestControl3& test_control) {
-    test_control.ExecuteThread(id);
-    test_control.CallFiber1();
-    test_control.Exit();
-}
-
 /** This test checks for one two threads racing for starting the same fiber.
  *  It checks execution occurred in an ordered manner and by no time there were
  *  two contexts at the same time.
@@ -304,12 +255,15 @@ static void ThreadStart3(u32 id, TestControl3& test_control) {
 TEST_CASE("Fibers::StartRace", "[common]") {
     TestControl3 test_control{};
     test_control.thread_fibers.resize(2);
-    test_control.fiber1 =
-        std::make_shared<Fiber>(std::function<void(void*)>{WorkControl3_1}, &test_control);
-    test_control.fiber2 =
-        std::make_shared<Fiber>(std::function<void(void*)>{WorkControl3_2}, &test_control);
-    std::thread thread1(ThreadStart3, 0, std::ref(test_control));
-    std::thread thread2(ThreadStart3, 1, std::ref(test_control));
+    test_control.fiber1 = std::make_shared<Fiber>([&test_control] { test_control.DoWork1(); });
+    test_control.fiber2 = std::make_shared<Fiber>([&test_control] { test_control.DoWork2(); });
+    const auto race_function{[&test_control](u32 id) {
+        test_control.ExecuteThread(id);
+        test_control.CallFiber1();
+        test_control.Exit();
+    }};
+    std::thread thread1([&] { race_function(0); });
+    std::thread thread2([&] { race_function(1); });
     thread1.join();
     thread2.join();
     REQUIRE(test_control.value1 == 1);
@@ -319,12 +273,10 @@ TEST_CASE("Fibers::StartRace", "[common]") {
 
 class TestControl4;
 
-static void WorkControl4(void* control);
-
 class TestControl4 {
 public:
     TestControl4() {
-        fiber1 = std::make_shared<Fiber>(std::function<void(void*)>{WorkControl4}, this);
+        fiber1 = std::make_shared<Fiber>([this] { DoWork(); });
         goal_reached = false;
         rewinded = false;
     }
@@ -336,7 +288,7 @@ public:
     }
 
     void DoWork() {
-        fiber1->SetRewindPoint(std::function<void(void*)>{WorkControl4}, this);
+        fiber1->SetRewindPoint([this] { DoWork(); });
         if (rewinded) {
             goal_reached = true;
             Fiber::YieldTo(fiber1, *thread_fiber);
@@ -351,11 +303,6 @@ public:
     bool rewinded;
 };
 
-static void WorkControl4(void* control) {
-    auto* test_control = static_cast<TestControl4*>(control);
-    test_control->DoWork();
-}
-
 TEST_CASE("Fibers::Rewind", "[common]") {
     TestControl4 test_control{};
     test_control.Execute();
diff --git a/src/tests/core/core_timing.cpp b/src/tests/core/core_timing.cpp
index 8358d36b5..e687416a8 100644
--- a/src/tests/core/core_timing.cpp
+++ b/src/tests/core/core_timing.cpp
@@ -8,6 +8,7 @@
 #include <chrono>
 #include <cstdlib>
 #include <memory>
+#include <mutex>
 #include <string>
 
 #include "core/core.h"
@@ -21,13 +22,14 @@ std::array<s64, 5> delays{};
 
 std::bitset<CB_IDS.size()> callbacks_ran_flags;
 u64 expected_callback = 0;
+std::mutex control_mutex;
 
 template <unsigned int IDX>
 void HostCallbackTemplate(std::uintptr_t user_data, std::chrono::nanoseconds ns_late) {
+    std::unique_lock<std::mutex> lk(control_mutex);
     static_assert(IDX < CB_IDS.size(), "IDX out of range");
     callbacks_ran_flags.set(IDX);
     REQUIRE(CB_IDS[IDX] == user_data);
-    REQUIRE(CB_IDS[IDX] == CB_IDS[calls_order[expected_callback]]);
     delays[IDX] = ns_late.count();
     ++expected_callback;
 }
diff --git a/src/video_core/vulkan_common/vulkan_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp
index 11ce865a7..743ac09f6 100644
--- a/src/video_core/vulkan_common/vulkan_device.cpp
+++ b/src/video_core/vulkan_common/vulkan_device.cpp
@@ -669,17 +669,6 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR
     const bool is_amd =
         driver_id == VK_DRIVER_ID_AMD_PROPRIETARY || driver_id == VK_DRIVER_ID_AMD_OPEN_SOURCE;
     if (is_amd) {
-        // TODO(lat9nq): Add an upper bound when AMD fixes their VK_KHR_push_descriptor
-        const bool has_broken_push_descriptor = VK_VERSION_MAJOR(properties.driverVersion) == 2 &&
-                                                VK_VERSION_MINOR(properties.driverVersion) == 0 &&
-                                                VK_VERSION_PATCH(properties.driverVersion) >= 226;
-        if (khr_push_descriptor && has_broken_push_descriptor) {
-            LOG_WARNING(
-                Render_Vulkan,
-                "Disabling AMD driver 2.0.226 and later from broken VK_KHR_push_descriptor");
-            khr_push_descriptor = false;
-        }
-
         // AMD drivers need a higher amount of Sets per Pool in certain circunstances like in XC2.
         sets_per_pool = 96;
         // Disable VK_IMAGE_CREATE_CUBE_COMPATIBLE_BIT on AMD GCN4 and lower as it is broken.
diff --git a/src/yuzu/applets/qt_web_browser.cpp b/src/yuzu/applets/qt_web_browser.cpp
index 283c04cd5..790edbb2a 100644
--- a/src/yuzu/applets/qt_web_browser.cpp
+++ b/src/yuzu/applets/qt_web_browser.cpp
@@ -52,8 +52,8 @@ QtNXWebEngineView::QtNXWebEngineView(QWidget* parent, Core::System& system,
     : QWebEngineView(parent), input_subsystem{input_subsystem_},
       url_interceptor(std::make_unique<UrlRequestInterceptor>()),
       input_interpreter(std::make_unique<InputInterpreter>(system)),
-      default_profile{QWebEngineProfile::defaultProfile()},
-      global_settings{QWebEngineSettings::globalSettings()} {
+      default_profile{QWebEngineProfile::defaultProfile()}, global_settings{
+                                                                default_profile->settings()} {
     default_profile->setPersistentStoragePath(QString::fromStdString(Common::FS::PathToUTF8String(
         Common::FS::GetYuzuPath(Common::FS::YuzuPath::YuzuDir) / "qtwebengine")));
 
@@ -78,7 +78,7 @@ QtNXWebEngineView::QtNXWebEngineView(QWidget* parent, Core::System& system,
     default_profile->scripts()->insert(gamepad);
     default_profile->scripts()->insert(window_nx);
 
-    default_profile->setRequestInterceptor(url_interceptor.get());
+    default_profile->setUrlRequestInterceptor(url_interceptor.get());
 
     global_settings->setAttribute(QWebEngineSettings::LocalContentCanAccessRemoteUrls, true);
     global_settings->setAttribute(QWebEngineSettings::FullScreenSupportEnabled, true);