arm_dynarmic: Always have a 'valid' jit instance

core: arm_dynarmic: Ensure JIT state is saved/restored on page table changes.

.gitmodules

@@ -27,7 +27,7 @@
     url = https://github.com/ReinUsesLisp/sirit
 [submodule "mbedtls"]
     path = externals/mbedtls
-    url = https://github.com/DarkLordZach/mbedtls
+    url = https://github.com/yuzu-emu/mbedtls
 [submodule "libzip"]
     path = externals/libzip/libzip
     url = https://github.com/nih-at/libzip.git

externals/libusb/CMakeLists.txt

@@ -1,3 +1,8 @@
+# Ensure libusb compiles with UTF-8 encoding on MSVC
+if(MSVC)
+    add_compile_options(/utf-8)
+endif()
+
 add_library(usb STATIC EXCLUDE_FROM_ALL
     libusb/libusb/core.c
     libusb/libusb/core.c

src/CMakeLists.txt

@@ -27,6 +27,7 @@ if (MSVC)
     # /Zo                 - Enhanced debug info for optimized builds
     # /permissive-        - Enables stricter C++ standards conformance checks
     # /EHsc               - C++-only exception handling semantics
+    # /utf-8              - Set source and execution character sets to UTF-8
     # /volatile:iso       - Use strict standards-compliant volatile semantics.
     # /Zc:externConstexpr - Allow extern constexpr variables to have external linkage, like the standard mandates
     # /Zc:inline          - Let codegen omit inline functions in object files
@@ -38,6 +39,7 @@ if (MSVC)
         /permissive-
         /EHsc
         /std:c++latest
+        /utf-8
         /volatile:iso
         /Zc:externConstexpr
         /Zc:inline

src/common/fiber.cpp

@@ -11,7 +11,7 @@
 
 namespace Common {
 
-constexpr std::size_t default_stack_size = 256 * 1024;
+constexpr std::size_t default_stack_size = 512 * 1024;
 
 struct Fiber::FiberImpl {
     FiberImpl() : stack{default_stack_size}, rewind_stack{default_stack_size} {}
@@ -116,16 +116,19 @@ void Fiber::Rewind() {
     boost::context::detail::jump_fcontext(impl->rewind_context, this);
 }
 
-void Fiber::YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to) {
-    ASSERT_MSG(from != nullptr, "Yielding fiber is null!");
-    ASSERT_MSG(to != nullptr, "Next fiber is null!");
-    to->impl->guard.lock();
-    to->impl->previous_fiber = from;
-    auto transfer = boost::context::detail::jump_fcontext(to->impl->context, to.get());
-    ASSERT(from->impl->previous_fiber != nullptr);
-    from->impl->previous_fiber->impl->context = transfer.fctx;
-    from->impl->previous_fiber->impl->guard.unlock();
-    from->impl->previous_fiber.reset();
+void Fiber::YieldTo(std::weak_ptr<Fiber> weak_from, Fiber& to) {
+    to.impl->guard.lock();
+    to.impl->previous_fiber = weak_from.lock();
+
+    auto transfer = boost::context::detail::jump_fcontext(to.impl->context, &to);
+
+    // "from" might no longer be valid if the thread was killed
+    if (auto from = weak_from.lock()) {
+        ASSERT(from->impl->previous_fiber != nullptr);
+        from->impl->previous_fiber->impl->context = transfer.fctx;
+        from->impl->previous_fiber->impl->guard.unlock();
+        from->impl->previous_fiber.reset();
+    }
 }
 
 std::shared_ptr<Fiber> Fiber::ThreadToFiber() {

src/common/fiber.h

@@ -41,7 +41,7 @@ public:
 
     /// Yields control from Fiber 'from' to Fiber 'to'
     /// Fiber 'from' must be the currently running fiber.
-    static void YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to);
+    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);

src/core/arm/dynarmic/arm_dynarmic_32.cpp

@@ -114,18 +114,17 @@ public:
     static constexpr u64 minimum_run_cycles = 1000U;
 };
 
-std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable& page_table,
-                                                             std::size_t address_space_bits) const {
+std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable* page_table) const {
     Dynarmic::A32::UserConfig config;
     config.callbacks = cb.get();
-    // TODO(bunnei): Implement page table for 32-bit
-    // config.page_table = &page_table.pointers;
     config.coprocessors[15] = cp15;
     config.define_unpredictable_behaviour = true;
     static constexpr std::size_t PAGE_BITS = 12;
     static constexpr std::size_t NUM_PAGE_TABLE_ENTRIES = 1 << (32 - PAGE_BITS);
-    config.page_table = reinterpret_cast<std::array<std::uint8_t*, NUM_PAGE_TABLE_ENTRIES>*>(
-        page_table.pointers.data());
+    if (page_table) {
+        config.page_table = reinterpret_cast<std::array<std::uint8_t*, NUM_PAGE_TABLE_ENTRIES>*>(
+            page_table->pointers.data());
+    }
     config.absolute_offset_page_table = true;
     config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS;
     config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128;
@@ -201,7 +200,8 @@ ARM_Dynarmic_32::ARM_Dynarmic_32(System& system, CPUInterrupts& interrupt_handle
     : ARM_Interface{system, interrupt_handlers, uses_wall_clock},
       cb(std::make_unique<DynarmicCallbacks32>(*this)),
       cp15(std::make_shared<DynarmicCP15>(*this)), core_index{core_index},
-      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {}
+      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)},
+      jit(MakeJit(nullptr)) {}
 
 ARM_Dynarmic_32::~ARM_Dynarmic_32() = default;
 
@@ -256,9 +256,6 @@ void ARM_Dynarmic_32::ChangeProcessorID(std::size_t new_core_id) {
 }
 
 void ARM_Dynarmic_32::SaveContext(ThreadContext32& ctx) {
-    if (!jit) {
-        return;
-    }
     Dynarmic::A32::Context context;
     jit->SaveContext(context);
     ctx.cpu_registers = context.Regs();
@@ -268,9 +265,6 @@ void ARM_Dynarmic_32::SaveContext(ThreadContext32& ctx) {
 }
 
 void ARM_Dynarmic_32::LoadContext(const ThreadContext32& ctx) {
-    if (!jit) {
-        return;
-    }
     Dynarmic::A32::Context context;
     context.Regs() = ctx.cpu_registers;
     context.ExtRegs() = ctx.extension_registers;
@@ -284,35 +278,31 @@ void ARM_Dynarmic_32::PrepareReschedule() {
 }
 
 void ARM_Dynarmic_32::ClearInstructionCache() {
-    if (!jit) {
-        return;
-    }
     jit->ClearCache();
 }
 
 void ARM_Dynarmic_32::InvalidateCacheRange(VAddr addr, std::size_t size) {
-    if (!jit) {
-        return;
-    }
     jit->InvalidateCacheRange(static_cast<u32>(addr), size);
 }
 
 void ARM_Dynarmic_32::ClearExclusiveState() {
-    if (!jit) {
-        return;
-    }
     jit->ClearExclusiveState();
 }
 
 void ARM_Dynarmic_32::PageTableChanged(Common::PageTable& page_table,
                                        std::size_t new_address_space_size_in_bits) {
+    ThreadContext32 ctx{};
+    SaveContext(ctx);
+
     auto key = std::make_pair(&page_table, new_address_space_size_in_bits);
     auto iter = jit_cache.find(key);
     if (iter != jit_cache.end()) {
         jit = iter->second;
+        LoadContext(ctx);
         return;
     }
-    jit = MakeJit(page_table, new_address_space_size_in_bits);
+    jit = MakeJit(&page_table);
+    LoadContext(ctx);
     jit_cache.emplace(key, jit);
 }
 

src/core/arm/dynarmic/arm_dynarmic_32.h

@@ -68,8 +68,7 @@ public:
                           std::size_t new_address_space_size_in_bits) override;
 
 private:
-    std::shared_ptr<Dynarmic::A32::Jit> MakeJit(Common::PageTable& page_table,
-                                                std::size_t address_space_bits) const;
+    std::shared_ptr<Dynarmic::A32::Jit> MakeJit(Common::PageTable* page_table) const;
 
     using JitCacheKey = std::pair<Common::PageTable*, std::size_t>;
     using JitCacheType =
@@ -80,10 +79,10 @@ private:
 
     std::unique_ptr<DynarmicCallbacks32> cb;
     JitCacheType jit_cache;
-    std::shared_ptr<Dynarmic::A32::Jit> jit;
     std::shared_ptr<DynarmicCP15> cp15;
     std::size_t core_index;
     DynarmicExclusiveMonitor& exclusive_monitor;
+    std::shared_ptr<Dynarmic::A32::Jit> jit;
 };
 
 } // namespace Core

src/core/arm/dynarmic/arm_dynarmic_64.cpp

@@ -142,7 +142,7 @@ public:
     static constexpr u64 minimum_run_cycles = 1000U;
 };
 
-std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& page_table,
+std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable* page_table,
                                                              std::size_t address_space_bits) const {
     Dynarmic::A64::UserConfig config;
 
@@ -150,13 +150,15 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable&
     config.callbacks = cb.get();
 
     // Memory
-    config.page_table = reinterpret_cast<void**>(page_table.pointers.data());
-    config.page_table_address_space_bits = address_space_bits;
-    config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS;
-    config.silently_mirror_page_table = false;
-    config.absolute_offset_page_table = true;
-    config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128;
-    config.only_detect_misalignment_via_page_table_on_page_boundary = true;
+    if (page_table) {
+        config.page_table = reinterpret_cast<void**>(page_table->pointers.data());
+        config.page_table_address_space_bits = address_space_bits;
+        config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS;
+        config.silently_mirror_page_table = false;
+        config.absolute_offset_page_table = true;
+        config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128;
+        config.only_detect_misalignment_via_page_table_on_page_boundary = true;
+    }
 
     // Multi-process state
     config.processor_id = core_index;
@@ -237,7 +239,8 @@ ARM_Dynarmic_64::ARM_Dynarmic_64(System& system, CPUInterrupts& interrupt_handle
                                  std::size_t core_index)
     : ARM_Interface{system, interrupt_handlers, uses_wall_clock},
       cb(std::make_unique<DynarmicCallbacks64>(*this)), core_index{core_index},
-      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {}
+      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)},
+      jit(MakeJit(nullptr, 48)) {}
 
 ARM_Dynarmic_64::~ARM_Dynarmic_64() = default;
 
@@ -294,9 +297,6 @@ void ARM_Dynarmic_64::ChangeProcessorID(std::size_t new_core_id) {
 }
 
 void ARM_Dynarmic_64::SaveContext(ThreadContext64& ctx) {
-    if (!jit) {
-        return;
-    }
     ctx.cpu_registers = jit->GetRegisters();
     ctx.sp = jit->GetSP();
     ctx.pc = jit->GetPC();
@@ -308,9 +308,6 @@ void ARM_Dynarmic_64::SaveContext(ThreadContext64& ctx) {
 }
 
 void ARM_Dynarmic_64::LoadContext(const ThreadContext64& ctx) {
-    if (!jit) {
-        return;
-    }
     jit->SetRegisters(ctx.cpu_registers);
     jit->SetSP(ctx.sp);
     jit->SetPC(ctx.pc);
@@ -326,35 +323,31 @@ void ARM_Dynarmic_64::PrepareReschedule() {
 }
 
 void ARM_Dynarmic_64::ClearInstructionCache() {
-    if (!jit) {
-        return;
-    }
     jit->ClearCache();
 }
 
 void ARM_Dynarmic_64::InvalidateCacheRange(VAddr addr, std::size_t size) {
-    if (!jit) {
-        return;
-    }
     jit->InvalidateCacheRange(addr, size);
 }
 
 void ARM_Dynarmic_64::ClearExclusiveState() {
-    if (!jit) {
-        return;
-    }
     jit->ClearExclusiveState();
 }
 
 void ARM_Dynarmic_64::PageTableChanged(Common::PageTable& page_table,
                                        std::size_t new_address_space_size_in_bits) {
+    ThreadContext64 ctx{};
+    SaveContext(ctx);
+
     auto key = std::make_pair(&page_table, new_address_space_size_in_bits);
     auto iter = jit_cache.find(key);
     if (iter != jit_cache.end()) {
         jit = iter->second;
+        LoadContext(ctx);
         return;
     }
-    jit = MakeJit(page_table, new_address_space_size_in_bits);
+    jit = MakeJit(&page_table, new_address_space_size_in_bits);
+    LoadContext(ctx);
     jit_cache.emplace(key, jit);
 }
 

src/core/arm/dynarmic/arm_dynarmic_64.h

@@ -61,7 +61,7 @@ public:
                           std::size_t new_address_space_size_in_bits) override;
 
 private:
-    std::shared_ptr<Dynarmic::A64::Jit> MakeJit(Common::PageTable& page_table,
+    std::shared_ptr<Dynarmic::A64::Jit> MakeJit(Common::PageTable* page_table,
                                                 std::size_t address_space_bits) const;
 
     using JitCacheKey = std::pair<Common::PageTable*, std::size_t>;
@@ -71,10 +71,11 @@ private:
     friend class DynarmicCallbacks64;
     std::unique_ptr<DynarmicCallbacks64> cb;
     JitCacheType jit_cache;
-    std::shared_ptr<Dynarmic::A64::Jit> jit;
 
     std::size_t core_index;
     DynarmicExclusiveMonitor& exclusive_monitor;
+
+    std::shared_ptr<Dynarmic::A64::Jit> jit;
 };
 
 } // namespace Core

src/core/core.cpp

@@ -299,28 +299,17 @@ struct System::Impl {
             gpu_core->WaitIdle();
         }
 
-        // Shutdown emulation session
         services.reset();
         service_manager.reset();
         cheat_engine.reset();
         telemetry_session.reset();
-
-        // Close all CPU/threading state
         cpu_manager.Shutdown();
-
-        // Release the Time Manager's resources
         time_manager.Shutdown();
-
-        // Shutdown kernel and core timing
         core_timing.Shutdown();
-        kernel.Shutdown();
-
-        // Close app loader
         app_loader.reset();
         gpu_core.reset();
         perf_stats.reset();
-
-        // Clear all applets
+        kernel.Shutdown();
         applet_manager.ClearAll();
 
         LOG_DEBUG(Core, "Shutdown OK");

src/core/cpu_manager.cpp

@@ -148,7 +148,7 @@ void CpuManager::MultiCoreRunSuspendThread() {
         auto core = kernel.GetCurrentHostThreadID();
         auto& scheduler = *kernel.CurrentScheduler();
         Kernel::KThread* current_thread = scheduler.GetCurrentThread();
-        Common::Fiber::YieldTo(current_thread->GetHostContext(), core_data[core].host_context);
+        Common::Fiber::YieldTo(current_thread->GetHostContext(), *core_data[core].host_context);
         ASSERT(scheduler.ContextSwitchPending());
         ASSERT(core == kernel.GetCurrentHostThreadID());
         scheduler.RescheduleCurrentCore();
@@ -245,7 +245,7 @@ void CpuManager::SingleCoreRunSuspendThread() {
         auto core = kernel.GetCurrentHostThreadID();
         auto& scheduler = *kernel.CurrentScheduler();
         Kernel::KThread* current_thread = scheduler.GetCurrentThread();
-        Common::Fiber::YieldTo(current_thread->GetHostContext(), core_data[0].host_context);
+        Common::Fiber::YieldTo(current_thread->GetHostContext(), *core_data[0].host_context);
         ASSERT(scheduler.ContextSwitchPending());
         ASSERT(core == kernel.GetCurrentHostThreadID());
         scheduler.RescheduleCurrentCore();
@@ -271,7 +271,7 @@ void CpuManager::PreemptSingleCore(bool from_running_enviroment) {
         scheduler.Unload(scheduler.GetCurrentThread());
 
         auto& next_scheduler = kernel.Scheduler(current_core);
-        Common::Fiber::YieldTo(current_thread->GetHostContext(), next_scheduler.ControlContext());
+        Common::Fiber::YieldTo(current_thread->GetHostContext(), *next_scheduler.ControlContext());
     }
 
     // May have changed scheduler
@@ -363,7 +363,7 @@ void CpuManager::RunThread(std::size_t core) {
 
         auto current_thread = system.Kernel().CurrentScheduler()->GetCurrentThread();
         data.is_running = true;
-        Common::Fiber::YieldTo(data.host_context, current_thread->GetHostContext());
+        Common::Fiber::YieldTo(data.host_context, *current_thread->GetHostContext());
         data.is_running = false;
         data.is_paused = true;
         data.exit_barrier->Wait();

src/core/crypto/aes_util.cpp

@@ -105,8 +105,6 @@ void AESCipher<Key, KeySize>::Transcode(const u8* src, std::size_t size, u8* des
             }
         }
     }
-
-    mbedtls_cipher_finish(context, nullptr, nullptr);
 }
 
 template <typename Key, std::size_t KeySize>

src/core/file_sys/system_archive/ng_word.cpp

@@ -14,7 +14,7 @@ namespace NgWord1Data {
 constexpr std::size_t NUMBER_WORD_TXT_FILES = 0x10;
 
 // Should this archive replacement mysteriously not work on a future game, consider updating.
-constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x19}; // 5.1.0 System Version
+constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x20}; // 11.0.1 System Version
 
 constexpr std::array<u8, 30> WORD_TXT{
     0xFE, 0xFF, 0x00, 0x5E, 0x00, 0x76, 0x00, 0x65, 0x00, 0x72, 0x00, 0x79, 0x00, 0x62, 0x00,
@@ -43,7 +43,7 @@ namespace NgWord2Data {
 constexpr std::size_t NUMBER_AC_NX_FILES = 0x10;
 
 // Should this archive replacement mysteriously not work on a future game, consider updating.
-constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x15}; // 5.1.0 System Version
+constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x1A}; // 11.0.1 System Version
 
 constexpr std::array<u8, 0x2C> AC_NX_DATA{
     0x1F, 0x8B, 0x08, 0x08, 0xD5, 0x2C, 0x09, 0x5C, 0x04, 0x00, 0x61, 0x63, 0x72, 0x61, 0x77,

src/core/file_sys/system_archive/system_version.cpp

@@ -14,15 +14,15 @@ namespace SystemVersionData {
 
 constexpr u8 VERSION_MAJOR = 11;
 constexpr u8 VERSION_MINOR = 0;
-constexpr u8 VERSION_MICRO = 0;
+constexpr u8 VERSION_MICRO = 1;
 
-constexpr u8 REVISION_MAJOR = 5;
+constexpr u8 REVISION_MAJOR = 1;
 constexpr u8 REVISION_MINOR = 0;
 
 constexpr char PLATFORM_STRING[] = "NX";
-constexpr char VERSION_HASH[] = "34197eba8810e2edd5e9dfcfbde7b340882e856d";
-constexpr char DISPLAY_VERSION[] = "11.0.0";
-constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 11.0.0-5.0";
+constexpr char VERSION_HASH[] = "69103fcb2004dace877094c2f8c29e6113be5dbf";
+constexpr char DISPLAY_VERSION[] = "11.0.1";
+constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 11.0.1-1.0";
 
 } // namespace SystemVersionData
 

src/core/hle/kernel/k_scheduler.cpp

@@ -734,7 +734,7 @@ void KScheduler::ScheduleImpl() {
     }
     guard.unlock();
 
-    Common::Fiber::YieldTo(*old_context, switch_fiber);
+    Common::Fiber::YieldTo(*old_context, *switch_fiber);
     /// When a thread wakes up, the scheduler may have changed to other in another core.
     auto& next_scheduler = *system.Kernel().CurrentScheduler();
     next_scheduler.SwitchContextStep2();
@@ -769,13 +769,8 @@ void KScheduler::SwitchToCurrent() {
                     break;
                 }
             }
-            std::shared_ptr<Common::Fiber>* next_context;
-            if (next_thread != nullptr) {
-                next_context = &next_thread->GetHostContext();
-            } else {
-                next_context = &idle_thread->GetHostContext();
-            }
-            Common::Fiber::YieldTo(switch_fiber, *next_context);
+            auto thread = next_thread ? next_thread : idle_thread;
+            Common::Fiber::YieldTo(switch_fiber, *thread->GetHostContext());
         } while (!is_switch_pending());
     }
 }
@@ -800,9 +795,9 @@ 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::Main, name, 0,
-                                      KThread::IdleThreadPriority, 0, static_cast<u32>(core_id), 0,
-                                      nullptr, std::move(init_func), init_func_parameter);
+    auto thread_res = KThread::CreateThread(
+        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();
 }
 

src/core/hle/kernel/k_thread.cpp

@@ -995,22 +995,11 @@ std::shared_ptr<Common::Fiber>& KThread::GetHostContext() {
     return host_context;
 }
 
-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) {
+ResultVal<std::shared_ptr<KThread>> KThread::CreateThread(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) {
     auto& kernel = system.Kernel();
 
     std::shared_ptr<KThread> thread = std::make_shared<KThread>(kernel);
@@ -1027,12 +1016,35 @@ ResultVal<std::shared_ptr<KThread>> KThread::Create(Core::System& system, Thread
     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));
 }
 
+ResultVal<std::shared_ptr<KThread>> KThread::CreateThread(
+    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 thread_result = CreateThread(system, type_flags, name, entry_point, priority, arg,
+                                      processor_id, stack_top, owner_process);
+
+    if (thread_result.Succeeded()) {
+        (*thread_result)->host_context =
+            std::make_shared<Common::Fiber>(std::move(thread_start_func), thread_start_parameter);
+    }
+
+    return thread_result;
+}
+
+ResultVal<std::shared_ptr<KThread>> KThread::CreateUserThread(
+    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 CreateThread(system, type_flags, name, entry_point, priority, arg, processor_id,
+                        stack_top, owner_process, std::move(init_func), init_func_parameter);
+}
+
 KThread* GetCurrentThreadPointer(KernelCore& kernel) {
     return kernel.GetCurrentEmuThread();
 }

src/core/hle/kernel/k_thread.h

@@ -116,7 +116,7 @@ public:
     using WaiterList = boost::intrusive::list<KThread>;
 
     /**
-     * Creates and returns a new thread. The new thread is immediately scheduled
+     * Creates and returns a new thread.
      * @param system The instance of the whole system
      * @param name The friendly name desired for the thread
      * @param entry_point The address at which the thread should start execution
@@ -127,12 +127,12 @@ 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
      */
-    [[nodiscard]] static ResultVal<std::shared_ptr<KThread>> Create(
+    [[nodiscard]] static ResultVal<std::shared_ptr<KThread>> CreateThread(
         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
+     * Creates and returns a new thread, with a specified entry point.
      * @param system The instance of the whole system
      * @param name The friendly name desired for the thread
      * @param entry_point The address at which the thread should start execution
@@ -145,11 +145,27 @@ public:
      * @param thread_start_parameter The parameter which will passed to host context on init
      * @return A shared pointer to the newly created thread
      */
-    [[nodiscard]] static ResultVal<std::shared_ptr<KThread>> Create(
+    [[nodiscard]] static ResultVal<std::shared_ptr<KThread>> CreateThread(
         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);
 
+    /**
+     * Creates and returns a new thread for the emulated "user" process.
+     * @param system The instance of the whole system
+     * @param name The friendly name desired for the thread
+     * @param entry_point The address at which the thread should start execution
+     * @param priority The thread's priority
+     * @param arg User data to pass to the thread
+     * @param processor_id The ID(s) of the processors on which the thread is desired to be run
+     * @param stack_top The address of the thread's stack top
+     * @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
+     */
+    [[nodiscard]] static ResultVal<std::shared_ptr<KThread>> CreateUserThread(
+        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]] std::string GetName() const override {
         return name;
     }

src/core/hle/kernel/kernel.cpp

@@ -181,9 +181,9 @@ 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::HighPriority, std::move(name), 0,
-                                              0, 0, static_cast<u32>(i), 0, nullptr,
-                                              std::move(init_func), init_func_parameter);
+            auto thread_res = KThread::CreateThread(
+                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();
         }
@@ -221,10 +221,9 @@ struct KernelCore::Impl {
     // Gets the dummy KThread for the caller, allocating a new one if this is the first time
     KThread* GetHostDummyThread() {
         const thread_local auto thread =
-            KThread::Create(
+            KThread::CreateThread(
                 system, ThreadType::Main, fmt::format("DummyThread:{}", GetHostThreadId()), 0,
-                KThread::DefaultThreadPriority, 0, static_cast<u32>(3), 0, nullptr,
-                []([[maybe_unused]] void* arg) { UNREACHABLE(); }, nullptr)
+                KThread::DefaultThreadPriority, 0, static_cast<u32>(3), 0, nullptr)
                 .Unwrap();
         return thread.get();
     }

src/core/hle/kernel/process.cpp

@@ -40,8 +40,9 @@ namespace {
 void SetupMainThread(Core::System& system, Process& owner_process, u32 priority, VAddr stack_top) {
     const VAddr entry_point = owner_process.PageTable().GetCodeRegionStart();
     ASSERT(owner_process.GetResourceLimit()->Reserve(LimitableResource::Threads, 1));
-    auto thread_res = KThread::Create(system, ThreadType::User, "main", entry_point, priority, 0,
-                                      owner_process.GetIdealCoreId(), stack_top, &owner_process);
+    auto thread_res =
+        KThread::CreateUserThread(system, ThreadType::User, "main", entry_point, priority, 0,
+                                  owner_process.GetIdealCoreId(), stack_top, &owner_process);
 
     std::shared_ptr<KThread> thread = std::move(thread_res).Unwrap();
 

src/core/hle/kernel/svc.cpp

@@ -1532,8 +1532,9 @@ static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr e
     std::shared_ptr<KThread> thread;
     {
         KScopedLightLock lk{process.GetStateLock()};
-        CASCADE_RESULT(thread, KThread::Create(system, ThreadType::User, "", entry_point, priority,
-                                               arg, core_id, stack_bottom, &process));
+        CASCADE_RESULT(thread,
+                       KThread::CreateUserThread(system, ThreadType::User, "", entry_point,
+                                                 priority, arg, core_id, stack_bottom, &process));
     }
 
     const auto new_thread_handle = process.GetHandleTable().Create(thread);

src/core/hle/service/sockets/bsd.cpp

@@ -42,7 +42,9 @@ void BSD::PollWork::Execute(BSD* bsd) {
 }
 
 void BSD::PollWork::Response(Kernel::HLERequestContext& ctx) {
-    ctx.WriteBuffer(write_buffer);
+    if (write_buffer.size() > 0) {
+        ctx.WriteBuffer(write_buffer);
+    }
 
     IPC::ResponseBuilder rb{ctx, 4};
     rb.Push(RESULT_SUCCESS);
@@ -55,7 +57,9 @@ void BSD::AcceptWork::Execute(BSD* bsd) {
 }
 
 void BSD::AcceptWork::Response(Kernel::HLERequestContext& ctx) {
-    ctx.WriteBuffer(write_buffer);
+    if (write_buffer.size() > 0) {
+        ctx.WriteBuffer(write_buffer);
+    }
 
     IPC::ResponseBuilder rb{ctx, 5};
     rb.Push(RESULT_SUCCESS);

src/core/hle/service/spl/module.cpp

@@ -43,6 +43,11 @@ void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system
     auto module = std::make_shared<Module>();
     std::make_shared<CSRNG>(system, module)->InstallAsService(service_manager);
     std::make_shared<SPL>(system, module)->InstallAsService(service_manager);
+    std::make_shared<SPL_MIG>(system, module)->InstallAsService(service_manager);
+    std::make_shared<SPL_FS>(system, module)->InstallAsService(service_manager);
+    std::make_shared<SPL_SSL>(system, module)->InstallAsService(service_manager);
+    std::make_shared<SPL_ES>(system, module)->InstallAsService(service_manager);
+    std::make_shared<SPL_MANU>(system, module)->InstallAsService(service_manager);
 }
 
 } // namespace Service::SPL

src/core/hle/service/spl/spl.cpp

@@ -8,6 +8,79 @@ namespace Service::SPL {
 
 SPL::SPL(Core::System& system_, std::shared_ptr<Module> module_)
     : Interface(system_, std::move(module_), "spl:") {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetConfig"},
+        {1, nullptr, "ModularExponentiate"},
+        {5, nullptr, "SetConfig"},
+        {7, &SPL::GetRandomBytes, "GetRandomBytes"},
+        {11, nullptr, "IsDevelopment"},
+        {24, nullptr, "SetBootReason"},
+        {25, nullptr, "GetBootReason"},
+    };
+    // clang-format on
+
+    RegisterHandlers(functions);
+}
+
+SPL_MIG::SPL_MIG(Core::System& system_, std::shared_ptr<Module> module_)
+    : Interface(system_, std::move(module_), "spl:mig") {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetConfig"},
+        {1, nullptr, "ModularExponentiate"},
+        {2, nullptr, "GenerateAesKek"},
+        {3, nullptr, "LoadAesKey"},
+        {4, nullptr, "GenerateAesKey"},
+        {5, nullptr, "SetConfig"},
+        {7, &SPL::GetRandomBytes, "GenerateRandomBytes"},
+        {11, nullptr, "IsDevelopment"},
+        {14, nullptr, "DecryptAesKey"},
+        {15, nullptr, "CryptAesCtr"},
+        {16, nullptr, "ComputeCmac"},
+        {21, nullptr, "AllocateAesKeyslot"},
+        {22, nullptr, "DeallocateAesKeySlot"},
+        {23, nullptr, "GetAesKeyslotAvailableEvent"},
+        {24, nullptr, "SetBootReason"},
+        {25, nullptr, "GetBootReason"},
+    };
+    // clang-format on
+
+    RegisterHandlers(functions);
+}
+
+SPL_FS::SPL_FS(Core::System& system_, std::shared_ptr<Module> module_)
+    : Interface(system_, std::move(module_), "spl:fs") {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetConfig"},
+        {1, nullptr, "ModularExponentiate"},
+        {2, nullptr, "GenerateAesKek"},
+        {3, nullptr, "LoadAesKey"},
+        {4, nullptr, "GenerateAesKey"},
+        {5, nullptr, "SetConfig"},
+        {7, &SPL::GetRandomBytes, "GenerateRandomBytes"},
+        {11, nullptr, "IsDevelopment"},
+        {12, nullptr, "GenerateSpecificAesKey"},
+        {14, nullptr, "DecryptAesKey"},
+        {15, nullptr, "CryptAesCtr"},
+        {16, nullptr, "ComputeCmac"},
+        {19, nullptr, "LoadTitleKey"},
+        {21, nullptr, "AllocateAesKeyslot"},
+        {22, nullptr, "DeallocateAesKeySlot"},
+        {23, nullptr, "GetAesKeyslotAvailableEvent"},
+        {24, nullptr, "SetBootReason"},
+        {25, nullptr, "GetBootReason"},
+        {31, nullptr, "GetPackage2Hash"},
+    };
+    // clang-format on
+
+    RegisterHandlers(functions);
+}
+
+SPL_SSL::SPL_SSL(Core::System& system_, std::shared_ptr<Module> module_)
+    : Interface(system_, std::move(module_), "spl:ssl") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "GetConfig"},
         {1, nullptr, "ModularExponentiate"},
@@ -16,18 +89,11 @@ SPL::SPL(Core::System& system_, std::shared_ptr<Module> module_)
         {4, nullptr, "GenerateAesKey"},
         {5, nullptr, "SetConfig"},
         {7, &SPL::GetRandomBytes, "GetRandomBytes"},
-        {9, nullptr, "ImportLotusKey"},
-        {10, nullptr, "DecryptLotusMessage"},
         {11, nullptr, "IsDevelopment"},
-        {12, nullptr, "GenerateSpecificAesKey"},
         {13, nullptr, "DecryptDeviceUniqueData"},
         {14, nullptr, "DecryptAesKey"},
         {15, nullptr, "CryptAesCtr"},
         {16, nullptr, "ComputeCmac"},
-        {17, nullptr, "ImportEsKey"},
-        {18, nullptr, "UnwrapTitleKey"},
-        {19, nullptr, "LoadTitleKey"},
-        {20, nullptr, "PrepareEsCommonKey"},
         {21, nullptr, "AllocateAesKeyslot"},
         {22, nullptr, "DeallocateAesKeySlot"},
         {23, nullptr, "GetAesKeyslotAvailableEvent"},
@@ -35,15 +101,83 @@ SPL::SPL(Core::System& system_, std::shared_ptr<Module> module_)
         {25, nullptr, "GetBootReason"},
         {26, nullptr, "DecryptAndStoreSslClientCertKey"},
         {27, nullptr, "ModularExponentiateWithSslClientCertKey"},
+    };
+    // clang-format on
+
+    RegisterHandlers(functions);
+}
+
+SPL_ES::SPL_ES(Core::System& system_, std::shared_ptr<Module> module_)
+    : Interface(system_, std::move(module_), "spl:es") {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetConfig"},
+        {1, nullptr, "ModularExponentiate"},
+        {2, nullptr, "GenerateAesKek"},
+        {3, nullptr, "LoadAesKey"},
+        {4, nullptr, "GenerateAesKey"},
+        {5, nullptr, "SetConfig"},
+        {7, &SPL::GetRandomBytes, "GenerateRandomBytes"},
+        {11, nullptr, "IsDevelopment"},
+        {13, nullptr, "DecryptDeviceUniqueData"},
+        {14, nullptr, "DecryptAesKey"},
+        {15, nullptr, "CryptAesCtr"},
+        {16, nullptr, "ComputeCmac"},
+        {18, nullptr, "UnwrapTitleKey"},
+        {20, nullptr, "PrepareEsCommonKey"},
+        {21, nullptr, "AllocateAesKeyslot"},
+        {22, nullptr, "DeallocateAesKeySlot"},
+        {23, nullptr, "GetAesKeyslotAvailableEvent"},
+        {24, nullptr, "SetBootReason"},
+        {25, nullptr, "GetBootReason"},
         {28, nullptr, "DecryptAndStoreDrmDeviceCertKey"},
         {29, nullptr, "ModularExponentiateWithDrmDeviceCertKey"},
-        {30, nullptr, "ReencryptDeviceUniqueData "},
-        {31, nullptr, "PrepareEsArchiveKey"}, // This is also GetPackage2Hash?
+        {31, nullptr, "PrepareEsArchiveKey"},
         {32, nullptr, "LoadPreparedAesKey"},
     };
+    // clang-format on
+
+    RegisterHandlers(functions);
+}
+
+SPL_MANU::SPL_MANU(Core::System& system_, std::shared_ptr<Module> module_)
+    : Interface(system_, std::move(module_), "spl:manu") {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetConfig"},
+        {1, nullptr, "ModularExponentiate"},
+        {2, nullptr, "GenerateAesKek"},
+        {3, nullptr, "LoadAesKey"},
+        {4, nullptr, "GenerateAesKey"},
+        {5, nullptr, "SetConfig"},
+        {7, &SPL::GetRandomBytes, "GetRandomBytes"},
+        {11, nullptr, "IsDevelopment"},
+        {13, nullptr, "DecryptDeviceUniqueData"},
+        {14, nullptr, "DecryptAesKey"},
+        {15, nullptr, "CryptAesCtr"},
+        {16, nullptr, "ComputeCmac"},
+        {21, nullptr, "AllocateAesKeyslot"},
+        {22, nullptr, "DeallocateAesKeySlot"},
+        {23, nullptr, "GetAesKeyslotAvailableEvent"},
+        {24, nullptr, "SetBootReason"},
+        {25, nullptr, "GetBootReason"},
+        {30, nullptr, "ReencryptDeviceUniqueData"},
+    };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
 SPL::~SPL() = default;
 
+SPL_MIG::~SPL_MIG() = default;
+
+SPL_FS::~SPL_FS() = default;
+
+SPL_SSL::~SPL_SSL() = default;
+
+SPL_ES::~SPL_ES() = default;
+
+SPL_MANU::~SPL_MANU() = default;
+
 } // namespace Service::SPL

src/core/hle/service/spl/spl.h

@@ -18,4 +18,34 @@ public:
     ~SPL() override;
 };
 
+class SPL_MIG final : public Module::Interface {
+public:
+    explicit SPL_MIG(Core::System& system_, std::shared_ptr<Module> module_);
+    ~SPL_MIG() override;
+};
+
+class SPL_FS final : public Module::Interface {
+public:
+    explicit SPL_FS(Core::System& system_, std::shared_ptr<Module> module_);
+    ~SPL_FS() override;
+};
+
+class SPL_SSL final : public Module::Interface {
+public:
+    explicit SPL_SSL(Core::System& system_, std::shared_ptr<Module> module_);
+    ~SPL_SSL() override;
+};
+
+class SPL_ES final : public Module::Interface {
+public:
+    explicit SPL_ES(Core::System& system_, std::shared_ptr<Module> module_);
+    ~SPL_ES() override;
+};
+
+class SPL_MANU final : public Module::Interface {
+public:
+    explicit SPL_MANU(Core::System& system_, std::shared_ptr<Module> module_);
+    ~SPL_MANU() override;
+};
+
 } // namespace Service::SPL

src/core/hle/service/time/time.cpp

@@ -140,6 +140,8 @@ ResultCode Module::Interface::GetClockSnapshotFromSystemClockContextInternal(
 
     const auto current_time_point{
         time_manager.GetStandardSteadyClockCore().GetCurrentTimePoint(system)};
+    clock_snapshot.steady_clock_time_point = current_time_point;
+
     if (const ResultCode result{Clock::ClockSnapshot::GetCurrentTime(
             clock_snapshot.user_time, current_time_point, clock_snapshot.user_context)};
         result != RESULT_SUCCESS) {
@@ -341,12 +343,18 @@ void Module::Interface::CalculateStandardUserSystemClockDifferenceByUser(
 void Module::Interface::CalculateSpanBetween(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_Time, "called");
 
-    IPC::RequestParser rp{ctx};
-    const auto snapshot_a = rp.PopRaw<Clock::ClockSnapshot>();
-    const auto snapshot_b = rp.PopRaw<Clock::ClockSnapshot>();
+    Clock::ClockSnapshot snapshot_a;
+    Clock::ClockSnapshot snapshot_b;
+
+    const auto snapshot_a_data = ctx.ReadBuffer(0);
+    const auto snapshot_b_data = ctx.ReadBuffer(1);
+
+    std::memcpy(&snapshot_a, snapshot_a_data.data(), sizeof(Clock::ClockSnapshot));
+    std::memcpy(&snapshot_b, snapshot_b_data.data(), sizeof(Clock::ClockSnapshot));
 
     Clock::TimeSpanType time_span_type{};
     s64 span{};
+
     if (const ResultCode result{snapshot_a.steady_clock_time_point.GetSpanBetween(
             snapshot_b.steady_clock_time_point, span)};
         result != RESULT_SUCCESS) {

src/core/hle/service/vi/vi.cpp

@@ -1217,6 +1217,32 @@ private:
         }
     }
 
+    void GetIndirectLayerImageMap(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto width = rp.Pop<s64>();
+        const auto height = rp.Pop<s64>();
+        const auto indirect_layer_consumer_handle = rp.Pop<u64>();
+        const auto applet_resource_user_id = rp.Pop<u64>();
+
+        LOG_WARNING(Service_VI,
+                    "(STUBBED) called, width={}, height={}, indirect_layer_consumer_handle={}, "
+                    "applet_resource_user_id={}",
+                    width, height, indirect_layer_consumer_handle, applet_resource_user_id);
+
+        std::vector<u8> out_buffer(0x46);
+        ctx.WriteBuffer(out_buffer);
+
+        // TODO: Figure out what these are
+
+        constexpr s64 unknown_result_1 = 0;
+        constexpr s64 unknown_result_2 = 0;
+
+        IPC::ResponseBuilder rb{ctx, 6};
+        rb.Push(unknown_result_1);
+        rb.Push(unknown_result_2);
+        rb.Push(RESULT_SUCCESS);
+    }
+
     void GetIndirectLayerImageRequiredMemoryInfo(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
         const auto width = rp.Pop<u64>();
@@ -1276,7 +1302,7 @@ IApplicationDisplayService::IApplicationDisplayService(Core::System& system_,
         {2031, &IApplicationDisplayService::DestroyStrayLayer, "DestroyStrayLayer"},
         {2101, &IApplicationDisplayService::SetLayerScalingMode, "SetLayerScalingMode"},
         {2102, &IApplicationDisplayService::ConvertScalingMode, "ConvertScalingMode"},
-        {2450, nullptr, "GetIndirectLayerImageMap"},
+        {2450, &IApplicationDisplayService::GetIndirectLayerImageMap, "GetIndirectLayerImageMap"},
         {2451, nullptr, "GetIndirectLayerImageCropMap"},
         {2460, &IApplicationDisplayService::GetIndirectLayerImageRequiredMemoryInfo,
          "GetIndirectLayerImageRequiredMemoryInfo"},

src/input_common/keyboard.cpp

@@ -12,20 +12,39 @@ namespace InputCommon {
 
 class KeyButton final : public Input::ButtonDevice {
 public:
-    explicit KeyButton(std::shared_ptr<KeyButtonList> key_button_list_)
-        : key_button_list(std::move(key_button_list_)) {}
+    explicit KeyButton(std::shared_ptr<KeyButtonList> key_button_list_, bool toggle_)
+        : key_button_list(std::move(key_button_list_)), toggle(toggle_) {}
 
     ~KeyButton() override;
 
     bool GetStatus() const override {
+        if (toggle) {
+            return toggled_status.load(std::memory_order_relaxed);
+        }
         return status.load();
     }
 
+    void ToggleButton() {
+        if (lock) {
+            return;
+        }
+        lock = true;
+        const bool old_toggle_status = toggled_status.load();
+        toggled_status.store(!old_toggle_status);
+    }
+
+    void UnlockButton() {
+        lock = false;
+    }
+
     friend class KeyButtonList;
 
 private:
     std::shared_ptr<KeyButtonList> key_button_list;
     std::atomic<bool> status{false};
+    std::atomic<bool> toggled_status{false};
+    bool lock{false};
+    const bool toggle;
 };
 
 struct KeyButtonPair {
@@ -51,6 +70,11 @@ public:
         for (const KeyButtonPair& pair : list) {
             if (pair.key_code == key_code) {
                 pair.key_button->status.store(pressed);
+                if (pressed) {
+                    pair.key_button->ToggleButton();
+                } else {
+                    pair.key_button->UnlockButton();
+                }
             }
         }
     }
@@ -75,7 +99,8 @@ KeyButton::~KeyButton() {
 
 std::unique_ptr<Input::ButtonDevice> Keyboard::Create(const Common::ParamPackage& params) {
     const int key_code = params.Get("code", 0);
-    std::unique_ptr<KeyButton> button = std::make_unique<KeyButton>(key_button_list);
+    const bool toggle = params.Get("toggle", false);
+    std::unique_ptr<KeyButton> button = std::make_unique<KeyButton>(key_button_list, toggle);
     key_button_list->AddKeyButton(key_code, button.get());
     return button;
 }

src/input_common/mouse/mouse_input.cpp

@@ -59,7 +59,7 @@ void Mouse::UpdateYuzuSettings() {
     });
 }
 
-void Mouse::PressButton(int x, int y, int button_) {
+void Mouse::PressButton(int x, int y, MouseButton button_) {
     const auto button_index = static_cast<std::size_t>(button_);
     if (button_index >= mouse_info.size()) {
         return;
@@ -67,7 +67,7 @@ void Mouse::PressButton(int x, int y, int button_) {
 
     const auto button = 1U << button_index;
     buttons |= static_cast<u16>(button);
-    last_button = static_cast<MouseButton>(button_index);
+    last_button = button_;
 
     mouse_info[button_index].mouse_origin = Common::MakeVec(x, y);
     mouse_info[button_index].last_mouse_position = Common::MakeVec(x, y);
@@ -129,7 +129,7 @@ void Mouse::MouseMove(int x, int y, int center_x, int center_y) {
     }
 }
 
-void Mouse::ReleaseButton(int button_) {
+void Mouse::ReleaseButton(MouseButton button_) {
     const auto button_index = static_cast<std::size_t>(button_);
     if (button_index >= mouse_info.size()) {
         return;
@@ -152,11 +152,52 @@ void Mouse::BeginConfiguration() {
 
 void Mouse::EndConfiguration() {
     buttons = 0;
+    for (MouseInfo& info : mouse_info) {
+        info.tilt_speed = 0;
+        info.data.pressed = false;
+        info.data.axis = {0, 0};
+    }
     last_button = MouseButton::Undefined;
     mouse_queue.Clear();
     configuring = false;
 }
 
+bool Mouse::ToggleButton(std::size_t button_) {
+    if (button_ >= mouse_info.size()) {
+        return false;
+    }
+    const auto button = 1U << button_;
+    const bool button_state = (toggle_buttons & button) != 0;
+    const bool button_lock = (lock_buttons & button) != 0;
+
+    if (button_lock) {
+        return button_state;
+    }
+
+    lock_buttons |= static_cast<u16>(button);
+
+    if (button_state) {
+        toggle_buttons &= static_cast<u16>(0xFF - button);
+    } else {
+        toggle_buttons |= static_cast<u16>(button);
+    }
+
+    return !button_state;
+}
+
+bool Mouse::UnlockButton(std::size_t button_) {
+    if (button_ >= mouse_info.size()) {
+        return false;
+    }
+
+    const auto button = 1U << button_;
+    const bool button_state = (toggle_buttons & button) != 0;
+
+    lock_buttons &= static_cast<u16>(0xFF - button);
+
+    return button_state;
+}
+
 Common::SPSCQueue<MouseStatus>& Mouse::GetMouseQueue() {
     return mouse_queue;
 }

src/input_common/mouse/mouse_input.h

@@ -18,10 +18,12 @@ namespace MouseInput {
 
 enum class MouseButton {
     Left,
-    Wheel,
     Right,
-    Forward,
+    Wheel,
     Backward,
+    Forward,
+    Task,
+    Extra,
     Undefined,
 };
 
@@ -51,7 +53,7 @@ public:
      * @param y the y-coordinate of the cursor
      * @param button_ the button pressed
      */
-    void PressButton(int x, int y, int button_);
+    void PressButton(int x, int y, MouseButton button_);
 
     /**
      * Signals that mouse has moved.
@@ -65,7 +67,10 @@ public:
     /**
      * Signals that a motion sensor tilt has ended.
      */
-    void ReleaseButton(int button_);
+    void ReleaseButton(MouseButton button_);
+
+    [[nodiscard]] bool ToggleButton(std::size_t button_);
+    [[nodiscard]] bool UnlockButton(std::size_t button_);
 
     [[nodiscard]] Common::SPSCQueue<MouseStatus>& GetMouseQueue();
     [[nodiscard]] const Common::SPSCQueue<MouseStatus>& GetMouseQueue() const;
@@ -92,9 +97,11 @@ private:
     };
 
     u16 buttons{};
+    u16 toggle_buttons{};
+    u16 lock_buttons{};
     std::thread update_thread;
     MouseButton last_button{MouseButton::Undefined};
-    std::array<MouseInfo, 5> mouse_info;
+    std::array<MouseInfo, 7> mouse_info;
     Common::SPSCQueue<MouseStatus> mouse_queue;
     bool configuring{false};
     bool update_thread_running{true};

src/input_common/mouse/mouse_poller.cpp

@@ -14,16 +14,25 @@ namespace InputCommon {
 
 class MouseButton final : public Input::ButtonDevice {
 public:
-    explicit MouseButton(u32 button_, const MouseInput::Mouse* mouse_input_)
-        : button(button_), mouse_input(mouse_input_) {}
+    explicit MouseButton(u32 button_, bool toggle_, MouseInput::Mouse* mouse_input_)
+        : button(button_), toggle(toggle_), mouse_input(mouse_input_) {}
 
     bool GetStatus() const override {
-        return mouse_input->GetMouseState(button).pressed;
+        const bool button_state = mouse_input->GetMouseState(button).pressed;
+        if (!toggle) {
+            return button_state;
+        }
+
+        if (button_state) {
+            return mouse_input->ToggleButton(button);
+        }
+        return mouse_input->UnlockButton(button);
     }
 
 private:
     const u32 button;
-    const MouseInput::Mouse* mouse_input;
+    const bool toggle;
+    MouseInput::Mouse* mouse_input;
 };
 
 MouseButtonFactory::MouseButtonFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_)
@@ -32,8 +41,9 @@ MouseButtonFactory::MouseButtonFactory(std::shared_ptr<MouseInput::Mouse> mouse_
 std::unique_ptr<Input::ButtonDevice> MouseButtonFactory::Create(
     const Common::ParamPackage& params) {
     const auto button_id = params.Get("button", 0);
+    const auto toggle = params.Get("toggle", false);
 
-    return std::make_unique<MouseButton>(button_id, mouse_input.get());
+    return std::make_unique<MouseButton>(button_id, toggle, mouse_input.get());
 }
 
 Common::ParamPackage MouseButtonFactory::GetNextInput() const {

src/input_common/udp/client.cpp

@@ -5,6 +5,7 @@
 #include <chrono>
 #include <cstring>
 #include <functional>
+#include <random>
 #include <thread>
 #include <boost/asio.hpp>
 #include "common/logging/log.h"
@@ -26,10 +27,10 @@ class Socket {
 public:
     using clock = std::chrono::system_clock;
 
-    explicit Socket(const std::string& host, u16 port, std::size_t pad_index_, u32 client_id_,
+    explicit Socket(const std::string& host, u16 port, std::size_t pad_index_,
                     SocketCallback callback_)
         : callback(std::move(callback_)), timer(io_service),
-          socket(io_service, udp::endpoint(udp::v4(), 0)), client_id(client_id_),
+          socket(io_service, udp::endpoint(udp::v4(), 0)), client_id(GenerateRandomClientId()),
           pad_index(pad_index_) {
         boost::system::error_code ec{};
         auto ipv4 = boost::asio::ip::make_address_v4(host, ec);
@@ -63,6 +64,11 @@ public:
     }
 
 private:
+    u32 GenerateRandomClientId() const {
+        std::random_device device;
+        return device();
+    }
+
     void HandleReceive(const boost::system::error_code&, std::size_t bytes_transferred) {
         if (auto type = Response::Validate(receive_buffer.data(), bytes_transferred)) {
             switch (*type) {
@@ -115,7 +121,7 @@ private:
     boost::asio::basic_waitable_timer<clock> timer;
     udp::socket socket;
 
-    u32 client_id{};
+    const u32 client_id;
     std::size_t pad_index{};
 
     static constexpr std::size_t PORT_INFO_SIZE = sizeof(Message<Request::PortInfo>);
@@ -203,7 +209,7 @@ void Client::ReloadSockets() {
                 LOG_ERROR(Input, "Duplicated UDP servers found");
                 continue;
             }
-            StartCommunication(client++, udp_input_address, udp_input_port, pad, 24872);
+            StartCommunication(client++, udp_input_address, udp_input_port, pad);
         }
     }
 }
@@ -277,7 +283,7 @@ void Client::OnPadData(Response::PadData data, std::size_t client) {
 }
 
 void Client::StartCommunication(std::size_t client, const std::string& host, u16 port,
-                                std::size_t pad_index, u32 client_id) {
+                                std::size_t pad_index) {
     SocketCallback callback{[this](Response::Version version) { OnVersion(version); },
                             [this](Response::PortInfo info) { OnPortInfo(info); },
                             [this, client](Response::PadData data) { OnPadData(data, client); }};
@@ -287,7 +293,7 @@ void Client::StartCommunication(std::size_t client, const std::string& host, u16
     clients[client].port = port;
     clients[client].pad_index = pad_index;
     clients[client].active = 0;
-    clients[client].socket = std::make_unique<Socket>(host, port, pad_index, client_id, callback);
+    clients[client].socket = std::make_unique<Socket>(host, port, pad_index, callback);
     clients[client].thread = std::thread{SocketLoop, clients[client].socket.get()};
     // Set motion parameters
     // SetGyroThreshold value should be dependent on GyroscopeZeroDriftMode
@@ -416,7 +422,7 @@ const Common::SPSCQueue<UDPPadStatus>& Client::GetPadQueue() const {
     return pad_queue;
 }
 
-void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, u32 client_id,
+void TestCommunication(const std::string& host, u16 port, std::size_t pad_index,
                        const std::function<void()>& success_callback,
                        const std::function<void()>& failure_callback) {
     std::thread([=] {
@@ -426,7 +432,7 @@ void TestCommunication(const std::string& host, u16 port, std::size_t pad_index,
             .port_info = [](Response::PortInfo) {},
             .pad_data = [&](Response::PadData) { success_event.Set(); },
         };
-        Socket socket{host, port, pad_index, client_id, std::move(callback)};
+        Socket socket{host, port, pad_index, std::move(callback)};
         std::thread worker_thread{SocketLoop, &socket};
         const bool result = success_event.WaitFor(std::chrono::seconds(5));
         socket.Stop();
@@ -440,7 +446,7 @@ void TestCommunication(const std::string& host, u16 port, std::size_t pad_index,
 }
 
 CalibrationConfigurationJob::CalibrationConfigurationJob(
-    const std::string& host, u16 port, std::size_t pad_index, u32 client_id,
+    const std::string& host, u16 port, std::size_t pad_index,
     std::function<void(Status)> status_callback,
     std::function<void(u16, u16, u16, u16)> data_callback) {
 
@@ -485,7 +491,7 @@ CalibrationConfigurationJob::CalibrationConfigurationJob(
                                         complete_event.Set();
                                     }
                                 }};
-        Socket socket{host, port, pad_index, client_id, std::move(callback)};
+        Socket socket{host, port, pad_index, std::move(callback)};
         std::thread worker_thread{SocketLoop, &socket};
         complete_event.Wait();
         socket.Stop();

src/input_common/udp/client.h

@@ -126,7 +126,7 @@ private:
     void OnPortInfo(Response::PortInfo);
     void OnPadData(Response::PadData, std::size_t client);
     void StartCommunication(std::size_t client, const std::string& host, u16 port,
-                            std::size_t pad_index, u32 client_id);
+                            std::size_t pad_index);
     void UpdateYuzuSettings(std::size_t client, const Common::Vec3<float>& acc,
                             const Common::Vec3<float>& gyro);
 
@@ -165,7 +165,7 @@ public:
      * @param data_callback Called when calibration data is ready
      */
     explicit CalibrationConfigurationJob(const std::string& host, u16 port, std::size_t pad_index,
-                                         u32 client_id, std::function<void(Status)> status_callback,
+                                         std::function<void(Status)> status_callback,
                                          std::function<void(u16, u16, u16, u16)> data_callback);
     ~CalibrationConfigurationJob();
     void Stop();
@@ -174,7 +174,7 @@ private:
     Common::Event complete_event;
 };
 
-void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, u32 client_id,
+void TestCommunication(const std::string& host, u16 port, std::size_t pad_index,
                        const std::function<void()>& success_callback,
                        const std::function<void()>& failure_callback);
 

src/tests/common/fibers.cpp

@@ -67,7 +67,7 @@ void TestControl1::DoWork() {
         value++;
     }
     results[id] = value;
-    Fiber::YieldTo(work_fibers[id], thread_fibers[id]);
+    Fiber::YieldTo(work_fibers[id], *thread_fibers[id]);
 }
 
 void TestControl1::ExecuteThread(u32 id) {
@@ -76,7 +76,7 @@ void TestControl1::ExecuteThread(u32 id) {
     thread_fibers[id] = thread_fiber;
     work_fibers[id] = std::make_shared<Fiber>(std::function<void(void*)>{WorkControl1}, this);
     items[id] = rand() % 256;
-    Fiber::YieldTo(thread_fibers[id], work_fibers[id]);
+    Fiber::YieldTo(thread_fibers[id], *work_fibers[id]);
     thread_fibers[id]->Exit();
 }
 
@@ -117,11 +117,11 @@ public:
         for (u32 i = 0; i < 12000; i++) {
             value1 += i;
         }
-        Fiber::YieldTo(fiber1, fiber3);
+        Fiber::YieldTo(fiber1, *fiber3);
         const u32 id = thread_ids.Get();
         assert1 = id == 1;
         value2 += 5000;
-        Fiber::YieldTo(fiber1, thread_fibers[id]);
+        Fiber::YieldTo(fiber1, *thread_fibers[id]);
     }
 
     void DoWork2() {
@@ -129,7 +129,7 @@ public:
             ;
         value2 = 2000;
         trap = false;
-        Fiber::YieldTo(fiber2, fiber1);
+        Fiber::YieldTo(fiber2, *fiber1);
         assert3 = false;
     }
 
@@ -137,19 +137,19 @@ public:
         const u32 id = thread_ids.Get();
         assert2 = id == 0;
         value1 += 1000;
-        Fiber::YieldTo(fiber3, thread_fibers[id]);
+        Fiber::YieldTo(fiber3, *thread_fibers[id]);
     }
 
     void ExecuteThread(u32 id);
 
     void CallFiber1() {
         const u32 id = thread_ids.Get();
-        Fiber::YieldTo(thread_fibers[id], fiber1);
+        Fiber::YieldTo(thread_fibers[id], *fiber1);
     }
 
     void CallFiber2() {
         const u32 id = thread_ids.Get();
-        Fiber::YieldTo(thread_fibers[id], fiber2);
+        Fiber::YieldTo(thread_fibers[id], *fiber2);
     }
 
     void Exit();
@@ -241,23 +241,23 @@ public:
 
     void DoWork1() {
         value1 += 1;
-        Fiber::YieldTo(fiber1, fiber2);
+        Fiber::YieldTo(fiber1, *fiber2);
         const u32 id = thread_ids.Get();
         value3 += 1;
-        Fiber::YieldTo(fiber1, thread_fibers[id]);
+        Fiber::YieldTo(fiber1, *thread_fibers[id]);
     }
 
     void DoWork2() {
         value2 += 1;
         const u32 id = thread_ids.Get();
-        Fiber::YieldTo(fiber2, thread_fibers[id]);
+        Fiber::YieldTo(fiber2, *thread_fibers[id]);
     }
 
     void ExecuteThread(u32 id);
 
     void CallFiber1() {
         const u32 id = thread_ids.Get();
-        Fiber::YieldTo(thread_fibers[id], fiber1);
+        Fiber::YieldTo(thread_fibers[id], *fiber1);
     }
 
     void Exit();
@@ -332,7 +332,7 @@ public:
 
     void Execute() {
         thread_fiber = Fiber::ThreadToFiber();
-        Fiber::YieldTo(thread_fiber, fiber1);
+        Fiber::YieldTo(thread_fiber, *fiber1);
         thread_fiber->Exit();
     }
 
@@ -340,7 +340,7 @@ public:
         fiber1->SetRewindPoint(std::function<void(void*)>{WorkControl4}, this);
         if (rewinded) {
             goal_reached = true;
-            Fiber::YieldTo(fiber1, thread_fiber);
+            Fiber::YieldTo(fiber1, *thread_fiber);
         }
         rewinded = true;
         fiber1->Rewind();

src/video_core/buffer_cache/buffer_cache.h

@@ -9,6 +9,7 @@
 #include <deque>
 #include <memory>
 #include <mutex>
+#include <numeric>
 #include <span>
 #include <unordered_map>
 #include <vector>
@@ -91,7 +92,7 @@ class BufferCache {
     };
 
 public:
-    static constexpr u32 SKIP_CACHE_SIZE = 4096;
+    static constexpr u32 DEFAULT_SKIP_CACHE_SIZE = 4096;
 
     explicit BufferCache(VideoCore::RasterizerInterface& rasterizer_,
                          Tegra::Engines::Maxwell3D& maxwell3d_,
@@ -240,9 +241,9 @@ private:
     template <bool insert>
     void ChangeRegister(BufferId buffer_id);
 
-    void SynchronizeBuffer(Buffer& buffer, VAddr cpu_addr, u32 size);
+    bool SynchronizeBuffer(Buffer& buffer, VAddr cpu_addr, u32 size);
 
-    void SynchronizeBufferImpl(Buffer& buffer, VAddr cpu_addr, u32 size);
+    bool SynchronizeBufferImpl(Buffer& buffer, VAddr cpu_addr, u32 size);
 
     void UploadMemory(Buffer& buffer, u64 total_size_bytes, u64 largest_copy,
                       std::span<BufferCopy> copies);
@@ -297,6 +298,11 @@ private:
 
     std::array<u32, NUM_STAGES> fast_bound_uniform_buffers{};
 
+    std::array<u32, 16> uniform_cache_hits{};
+    std::array<u32, 16> uniform_cache_shots{};
+
+    u32 uniform_buffer_skip_cache_size = DEFAULT_SKIP_CACHE_SIZE;
+
     bool has_deleted_buffers = false;
 
     std::conditional_t<HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS, std::array<u32, NUM_STAGES>, Empty>
@@ -328,6 +334,19 @@ BufferCache<P>::BufferCache(VideoCore::RasterizerInterface& rasterizer_,
 
 template <class P>
 void BufferCache<P>::TickFrame() {
+    // Calculate hits and shots and move hit bits to the right
+    const u32 hits = std::reduce(uniform_cache_hits.begin(), uniform_cache_hits.end());
+    const u32 shots = std::reduce(uniform_cache_shots.begin(), uniform_cache_shots.end());
+    std::copy_n(uniform_cache_hits.begin(), uniform_cache_hits.size() - 1,
+                uniform_cache_hits.begin() + 1);
+    std::copy_n(uniform_cache_shots.begin(), uniform_cache_shots.size() - 1,
+                uniform_cache_shots.begin() + 1);
+    uniform_cache_hits[0] = 0;
+    uniform_cache_shots[0] = 0;
+
+    const bool skip_preferred = hits * 256 < shots * 251;
+    uniform_buffer_skip_cache_size = skip_preferred ? DEFAULT_SKIP_CACHE_SIZE : 0;
+
     delayed_destruction_ring.Tick();
 }
 
@@ -671,7 +690,7 @@ void BufferCache<P>::BindHostGraphicsUniformBuffer(size_t stage, u32 index, u32
     const VAddr cpu_addr = binding.cpu_addr;
     const u32 size = binding.size;
     Buffer& buffer = slot_buffers[binding.buffer_id];
-    if (size <= SKIP_CACHE_SIZE && !buffer.IsRegionGpuModified(cpu_addr, size)) {
+    if (size <= uniform_buffer_skip_cache_size && !buffer.IsRegionGpuModified(cpu_addr, size)) {
         if constexpr (IS_OPENGL) {
             if (runtime.HasFastBufferSubData()) {
                 // Fast path for Nvidia
@@ -692,7 +711,12 @@ void BufferCache<P>::BindHostGraphicsUniformBuffer(size_t stage, u32 index, u32
         return;
     }
     // Classic cached path
-    SynchronizeBuffer(buffer, cpu_addr, size);
+    const bool sync_cached = SynchronizeBuffer(buffer, cpu_addr, size);
+    if (sync_cached) {
+        ++uniform_cache_hits[0];
+    }
+    ++uniform_cache_shots[0];
+
     if (!needs_bind && !HasFastUniformBufferBound(stage, binding_index)) {
         // Skip binding if it's not needed and if the bound buffer is not the fast version
         // This exists to avoid instances where the fast buffer is bound and a GPU write happens
@@ -1106,15 +1130,15 @@ void BufferCache<P>::ChangeRegister(BufferId buffer_id) {
 }
 
 template <class P>
-void BufferCache<P>::SynchronizeBuffer(Buffer& buffer, VAddr cpu_addr, u32 size) {
+bool BufferCache<P>::SynchronizeBuffer(Buffer& buffer, VAddr cpu_addr, u32 size) {
     if (buffer.CpuAddr() == 0) {
-        return;
+        return true;
     }
-    SynchronizeBufferImpl(buffer, cpu_addr, size);
+    return SynchronizeBufferImpl(buffer, cpu_addr, size);
 }
 
 template <class P>
-void BufferCache<P>::SynchronizeBufferImpl(Buffer& buffer, VAddr cpu_addr, u32 size) {
+bool BufferCache<P>::SynchronizeBufferImpl(Buffer& buffer, VAddr cpu_addr, u32 size) {
     boost::container::small_vector<BufferCopy, 4> copies;
     u64 total_size_bytes = 0;
     u64 largest_copy = 0;
@@ -1128,10 +1152,11 @@ void BufferCache<P>::SynchronizeBufferImpl(Buffer& buffer, VAddr cpu_addr, u32 s
         largest_copy = std::max(largest_copy, range_size);
     });
     if (total_size_bytes == 0) {
-        return;
+        return true;
     }
     const std::span<BufferCopy> copies_span(copies.data(), copies.size());
     UploadMemory(buffer, total_size_bytes, largest_copy, copies_span);
+    return false;
 }
 
 template <class P>

src/video_core/compatible_formats.cpp

@@ -48,6 +48,15 @@ constexpr std::array VIEW_CLASS_32_BITS{
     PixelFormat::A2B10G10R10_UINT,
 };
 
+constexpr std::array VIEW_CLASS_32_BITS_NO_BGR{
+    PixelFormat::R16G16_FLOAT,      PixelFormat::B10G11R11_FLOAT,  PixelFormat::R32_FLOAT,
+    PixelFormat::A2B10G10R10_UNORM, PixelFormat::R16G16_UINT,      PixelFormat::R32_UINT,
+    PixelFormat::R16G16_SINT,       PixelFormat::R32_SINT,         PixelFormat::A8B8G8R8_UNORM,
+    PixelFormat::R16G16_UNORM,      PixelFormat::A8B8G8R8_SNORM,   PixelFormat::R16G16_SNORM,
+    PixelFormat::A8B8G8R8_SRGB,     PixelFormat::E5B9G9R9_FLOAT,   PixelFormat::A8B8G8R8_UINT,
+    PixelFormat::A8B8G8R8_SINT,     PixelFormat::A2B10G10R10_UINT,
+};
+
 // TODO: How should we handle 24 bits?
 
 constexpr std::array VIEW_CLASS_16_BITS{
@@ -205,7 +214,6 @@ constexpr Table MakeViewTable() {
     EnableRange(view, VIEW_CLASS_128_BITS);
     EnableRange(view, VIEW_CLASS_96_BITS);
     EnableRange(view, VIEW_CLASS_64_BITS);
-    EnableRange(view, VIEW_CLASS_32_BITS);
     EnableRange(view, VIEW_CLASS_16_BITS);
     EnableRange(view, VIEW_CLASS_8_BITS);
     EnableRange(view, VIEW_CLASS_RGTC1_RED);
@@ -231,20 +239,47 @@ constexpr Table MakeCopyTable() {
     EnableRange(copy, COPY_CLASS_64_BITS);
     return copy;
 }
+
+constexpr Table MakeNativeBgrViewTable() {
+    Table copy = MakeViewTable();
+    EnableRange(copy, VIEW_CLASS_32_BITS);
+    return copy;
+}
+
+constexpr Table MakeNonNativeBgrViewTable() {
+    Table copy = MakeViewTable();
+    EnableRange(copy, VIEW_CLASS_32_BITS_NO_BGR);
+    return copy;
+}
+
+constexpr Table MakeNativeBgrCopyTable() {
+    Table copy = MakeCopyTable();
+    EnableRange(copy, VIEW_CLASS_32_BITS);
+    return copy;
+}
+
+constexpr Table MakeNonNativeBgrCopyTable() {
+    Table copy = MakeCopyTable();
+    EnableRange(copy, VIEW_CLASS_32_BITS);
+    return copy;
+}
 } // Anonymous namespace
 
-bool IsViewCompatible(PixelFormat format_a, PixelFormat format_b, bool broken_views) {
+bool IsViewCompatible(PixelFormat format_a, PixelFormat format_b, bool broken_views,
+                      bool native_bgr) {
     if (broken_views) {
         // If format views are broken, only accept formats that are identical.
         return format_a == format_b;
     }
-    static constexpr Table TABLE = MakeViewTable();
-    return IsSupported(TABLE, format_a, format_b);
+    static constexpr Table BGR_TABLE = MakeNativeBgrViewTable();
+    static constexpr Table NO_BGR_TABLE = MakeNonNativeBgrViewTable();
+    return IsSupported(native_bgr ? BGR_TABLE : NO_BGR_TABLE, format_a, format_b);
 }
 
-bool IsCopyCompatible(PixelFormat format_a, PixelFormat format_b) {
-    static constexpr Table TABLE = MakeCopyTable();
-    return IsSupported(TABLE, format_a, format_b);
+bool IsCopyCompatible(PixelFormat format_a, PixelFormat format_b, bool native_bgr) {
+    static constexpr Table BGR_TABLE = MakeNativeBgrCopyTable();
+    static constexpr Table NO_BGR_TABLE = MakeNonNativeBgrCopyTable();
+    return IsSupported(native_bgr ? BGR_TABLE : NO_BGR_TABLE, format_a, format_b);
 }
 
 } // namespace VideoCore::Surface

src/video_core/compatible_formats.h

@@ -8,8 +8,9 @@
 
 namespace VideoCore::Surface {
 
-bool IsViewCompatible(PixelFormat format_a, PixelFormat format_b, bool broken_views);
+bool IsViewCompatible(PixelFormat format_a, PixelFormat format_b, bool broken_views,
+                      bool native_bgr);
 
-bool IsCopyCompatible(PixelFormat format_a, PixelFormat format_b);
+bool IsCopyCompatible(PixelFormat format_a, PixelFormat format_b, bool native_bgr);
 
 } // namespace VideoCore::Surface

src/video_core/gpu.cpp

@@ -459,7 +459,7 @@ void GPU::ProcessSemaphoreAcquire() {
 }
 
 void GPU::Start() {
-    gpu_thread.StartThread(*renderer, renderer->Context(), *dma_pusher, *cdma_pusher);
+    gpu_thread.StartThread(*renderer, renderer->Context(), *dma_pusher);
     cpu_context = renderer->GetRenderWindow().CreateSharedContext();
     cpu_context->MakeCurrent();
 }

src/video_core/gpu_thread.cpp

@@ -19,7 +19,7 @@ namespace VideoCommon::GPUThread {
 /// Runs the GPU thread
 static void RunThread(Core::System& system, VideoCore::RendererBase& renderer,
                       Core::Frontend::GraphicsContext& context, Tegra::DmaPusher& dma_pusher,
-                      SynchState& state, Tegra::CDmaPusher& cdma_pusher) {
+                      SynchState& state) {
     std::string name = "yuzu:GPU";
     MicroProfileOnThreadCreate(name.c_str());
     SCOPE_EXIT({ MicroProfileOnThreadExit(); });
@@ -46,9 +46,6 @@ static void RunThread(Core::System& system, VideoCore::RendererBase& renderer,
         if (auto* submit_list = std::get_if<SubmitListCommand>(&next.data)) {
             dma_pusher.Push(std::move(submit_list->entries));
             dma_pusher.DispatchCalls();
-        } else if (auto* command_list = std::get_if<SubmitChCommandEntries>(&next.data)) {
-            // NVDEC
-            cdma_pusher.ProcessEntries(std::move(command_list->entries));
         } else if (const auto* data = std::get_if<SwapBuffersCommand>(&next.data)) {
             renderer.SwapBuffers(data->framebuffer ? &*data->framebuffer : nullptr);
         } else if (std::holds_alternative<OnCommandListEndCommand>(next.data)) {
@@ -83,20 +80,16 @@ ThreadManager::~ThreadManager() {
 
 void ThreadManager::StartThread(VideoCore::RendererBase& renderer,
                                 Core::Frontend::GraphicsContext& context,
-                                Tegra::DmaPusher& dma_pusher, Tegra::CDmaPusher& cdma_pusher) {
+                                Tegra::DmaPusher& dma_pusher) {
     rasterizer = renderer.ReadRasterizer();
     thread = std::thread(RunThread, std::ref(system), std::ref(renderer), std::ref(context),
-                         std::ref(dma_pusher), std::ref(state), std::ref(cdma_pusher));
+                         std::ref(dma_pusher), std::ref(state));
 }
 
 void ThreadManager::SubmitList(Tegra::CommandList&& entries) {
     PushCommand(SubmitListCommand(std::move(entries)));
 }
 
-void ThreadManager::SubmitCommandBuffer(Tegra::ChCommandHeaderList&& entries) {
-    PushCommand(SubmitChCommandEntries(std::move(entries)));
-}
-
 void ThreadManager::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
     PushCommand(SwapBuffersCommand(framebuffer ? std::make_optional(*framebuffer) : std::nullopt));
 }

src/video_core/gpu_thread.h

@@ -43,14 +43,6 @@ struct SubmitListCommand final {
     Tegra::CommandList entries;
 };
 
-/// Command to signal to the GPU thread that a cdma command list is ready for processing
-struct SubmitChCommandEntries final {
-    explicit SubmitChCommandEntries(Tegra::ChCommandHeaderList&& entries_)
-        : entries{std::move(entries_)} {}
-
-    Tegra::ChCommandHeaderList entries;
-};
-
 /// Command to signal to the GPU thread that a swap buffers is pending
 struct SwapBuffersCommand final {
     explicit SwapBuffersCommand(std::optional<const Tegra::FramebufferConfig> framebuffer_)
@@ -91,9 +83,9 @@ struct OnCommandListEndCommand final {};
 struct GPUTickCommand final {};
 
 using CommandData =
-    std::variant<EndProcessingCommand, SubmitListCommand, SubmitChCommandEntries,
-                 SwapBuffersCommand, FlushRegionCommand, InvalidateRegionCommand,
-                 FlushAndInvalidateRegionCommand, OnCommandListEndCommand, GPUTickCommand>;
+    std::variant<EndProcessingCommand, SubmitListCommand, SwapBuffersCommand, FlushRegionCommand,
+                 InvalidateRegionCommand, FlushAndInvalidateRegionCommand, OnCommandListEndCommand,
+                 GPUTickCommand>;
 
 struct CommandDataContainer {
     CommandDataContainer() = default;
@@ -123,14 +115,11 @@ public:
 
     /// Creates and starts the GPU thread.
     void StartThread(VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context,
-                     Tegra::DmaPusher& dma_pusher, Tegra::CDmaPusher& cdma_pusher);
+                     Tegra::DmaPusher& dma_pusher);
 
     /// Push GPU command entries to be processed
     void SubmitList(Tegra::CommandList&& entries);
 
-    /// Push GPU CDMA command buffer entries to be processed
-    void SubmitCommandBuffer(Tegra::ChCommandHeaderList&& entries);
-
     /// Swap buffers (render frame)
     void SwapBuffers(const Tegra::FramebufferConfig* framebuffer);
 

src/video_core/host_shaders/CMakeLists.txt

@@ -5,6 +5,7 @@ set(SHADER_FILES
     convert_float_to_depth.frag
     full_screen_triangle.vert
     opengl_copy_bc4.comp
+    opengl_copy_bgra.comp
     opengl_present.frag
     opengl_present.vert
     pitch_unswizzle.comp

src/video_core/host_shaders/opengl_copy_bgra.comp

@@ -0,0 +1,15 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#version 430 core
+
+layout (local_size_x = 4, local_size_y = 4) in;
+
+layout(binding = 0, rgba8) readonly uniform image2DArray bgr_input;
+layout(binding = 1, rgba8) writeonly uniform image2DArray bgr_output;
+
+void main() {
+    vec4 color = imageLoad(bgr_input, ivec3(gl_GlobalInvocationID));
+    imageStore(bgr_output, ivec3(gl_GlobalInvocationID), color.bgra);
+}

src/video_core/rasterizer_accelerated.cpp

@@ -2,63 +2,43 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <mutex>
-
-#include <boost/icl/interval_map.hpp>
-#include <boost/range/iterator_range.hpp>
-
 #include "common/assert.h"
 #include "common/common_types.h"
+#include "common/div_ceil.h"
 #include "core/memory.h"
 #include "video_core/rasterizer_accelerated.h"
 
 namespace VideoCore {
 
-namespace {
-
-template <typename Map, typename Interval>
-constexpr auto RangeFromInterval(Map& map, const Interval& interval) {
-    return boost::make_iterator_range(map.equal_range(interval));
-}
-
-} // Anonymous namespace
-
 RasterizerAccelerated::RasterizerAccelerated(Core::Memory::Memory& cpu_memory_)
     : cpu_memory{cpu_memory_} {}
 
 RasterizerAccelerated::~RasterizerAccelerated() = default;
 
 void RasterizerAccelerated::UpdatePagesCachedCount(VAddr addr, u64 size, int delta) {
-    std::lock_guard lock{pages_mutex};
-    const u64 page_start{addr >> Core::Memory::PAGE_BITS};
-    const u64 page_end{(addr + size + Core::Memory::PAGE_SIZE - 1) >> Core::Memory::PAGE_BITS};
+    const auto page_end = Common::DivCeil(addr + size, Core::Memory::PAGE_SIZE);
+    for (auto page = addr >> Core::Memory::PAGE_BITS; page != page_end; ++page) {
+        auto& count = cached_pages.at(page >> 3).Count(page);
 
-    // Interval maps will erase segments if count reaches 0, so if delta is negative we have to
-    // subtract after iterating
-    const auto pages_interval = CachedPageMap::interval_type::right_open(page_start, page_end);
-    if (delta > 0) {
-        cached_pages.add({pages_interval, delta});
-    }
-
-    for (const auto& pair : RangeFromInterval(cached_pages, pages_interval)) {
-        const auto interval = pair.first & pages_interval;
-        const int count = pair.second;
-
-        const VAddr interval_start_addr = boost::icl::first(interval) << Core::Memory::PAGE_BITS;
-        const VAddr interval_end_addr = boost::icl::last_next(interval) << Core::Memory::PAGE_BITS;
-        const u64 interval_size = interval_end_addr - interval_start_addr;
-
-        if (delta > 0 && count == delta) {
-            cpu_memory.RasterizerMarkRegionCached(interval_start_addr, interval_size, true);
-        } else if (delta < 0 && count == -delta) {
-            cpu_memory.RasterizerMarkRegionCached(interval_start_addr, interval_size, false);
+        if (delta > 0) {
+            ASSERT_MSG(count < UINT8_MAX, "Count may overflow!");
+        } else if (delta < 0) {
+            ASSERT_MSG(count > 0, "Count may underflow!");
         } else {
-            ASSERT(count >= 0);
+            ASSERT_MSG(true, "Delta must be non-zero!");
         }
-    }
 
-    if (delta < 0) {
-        cached_pages.add({pages_interval, delta});
+        // Adds or subtracts 1, as count is a unsigned 8-bit value
+        count += static_cast<u8>(delta);
+
+        // Assume delta is either -1 or 1
+        if (count == 0) {
+            cpu_memory.RasterizerMarkRegionCached(page << Core::Memory::PAGE_BITS,
+                                                  Core::Memory::PAGE_SIZE, false);
+        } else if (count == 1 && delta > 0) {
+            cpu_memory.RasterizerMarkRegionCached(page << Core::Memory::PAGE_BITS,
+                                                  Core::Memory::PAGE_SIZE, true);
+        }
     }
 }
 

src/video_core/rasterizer_accelerated.h

@@ -4,9 +4,8 @@
 
 #pragma once
 
-#include <mutex>
-
-#include <boost/icl/interval_map.hpp>
+#include <array>
+#include <atomic>
 
 #include "common/common_types.h"
 #include "video_core/rasterizer_interface.h"
@@ -26,10 +25,24 @@ public:
     void UpdatePagesCachedCount(VAddr addr, u64 size, int delta) override;
 
 private:
-    using CachedPageMap = boost::icl::interval_map<u64, int>;
-    CachedPageMap cached_pages;
-    std::mutex pages_mutex;
+    class CacheEntry final {
+    public:
+        CacheEntry() = default;
 
+        std::atomic_uint8_t& Count(std::size_t page) {
+            return values[page & 7];
+        }
+
+        const std::atomic_uint8_t& Count(std::size_t page) const {
+            return values[page & 7];
+        }
+
+    private:
+        std::array<std::atomic_uint8_t, 8> values{};
+    };
+    static_assert(sizeof(CacheEntry) == 8, "CacheEntry should be 8 bytes!");
+
+    std::array<CacheEntry, 0x800000> cached_pages;
     Core::Memory::Memory& cpu_memory;
 };
 

src/video_core/renderer_opengl/gl_buffer_cache.cpp

@@ -73,7 +73,8 @@ BufferCacheRuntime::BufferCacheRuntime(const Device& device_)
     for (auto& stage_uniforms : fast_uniforms) {
         for (OGLBuffer& buffer : stage_uniforms) {
             buffer.Create();
-            glNamedBufferData(buffer.handle, BufferCache::SKIP_CACHE_SIZE, nullptr, GL_STREAM_DRAW);
+            glNamedBufferData(buffer.handle, BufferCache::DEFAULT_SKIP_CACHE_SIZE, nullptr,
+                              GL_STREAM_DRAW);
         }
     }
     for (auto& stage_uniforms : copy_uniforms) {

src/video_core/renderer_opengl/gl_device.cpp

@@ -210,6 +210,12 @@ Device::Device() {
     const bool is_amd = vendor == "ATI Technologies Inc.";
     const bool is_intel = vendor == "Intel";
 
+#ifdef __linux__
+    const bool is_linux = true;
+#else
+    const bool is_linux = false;
+#endif
+
     bool disable_fast_buffer_sub_data = false;
     if (is_nvidia && version == "4.6.0 NVIDIA 443.24") {
         LOG_WARNING(
@@ -249,7 +255,9 @@ Device::Device() {
                            GLAD_GL_NV_gpu_program5 && GLAD_GL_NV_compute_program5 &&
                            GLAD_GL_NV_transform_feedback && GLAD_GL_NV_transform_feedback2;
 
-    use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue();
+    // Blocks AMD and Intel OpenGL drivers on Windows from using asynchronous shader compilation.
+    use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue() &&
+                               !(is_amd || (is_intel && !is_linux));
     use_driver_cache = is_nvidia;
 
     LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi);
@@ -261,6 +269,10 @@ Device::Device() {
     if (Settings::values.use_assembly_shaders.GetValue() && !use_assembly_shaders) {
         LOG_ERROR(Render_OpenGL, "Assembly shaders enabled but not supported");
     }
+
+    if (Settings::values.use_asynchronous_shaders.GetValue() && !use_asynchronous_shaders) {
+        LOG_WARNING(Render_OpenGL, "Asynchronous shader compilation enabled but not supported");
+    }
 }
 
 Device::Device(std::nullptr_t) {

src/video_core/renderer_opengl/gl_texture_cache.cpp

@@ -96,7 +96,7 @@ constexpr std::array<FormatTuple, MaxPixelFormat> FORMAT_TABLE = {{
     {GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT},                          // BC6H_UFLOAT
     {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT},                            // BC6H_SFLOAT
     {GL_COMPRESSED_RGBA_ASTC_4x4_KHR},                                // ASTC_2D_4X4_UNORM
-    {GL_RGBA8, GL_BGRA, GL_UNSIGNED_BYTE},                            // B8G8R8A8_UNORM
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE},                            // B8G8R8A8_UNORM
     {GL_RGBA32F, GL_RGBA, GL_FLOAT},                                  // R32G32B32A32_FLOAT
     {GL_RGBA32I, GL_RGBA_INTEGER, GL_INT},                            // R32G32B32A32_SINT
     {GL_RG32F, GL_RG, GL_FLOAT},                                      // R32G32_FLOAT
@@ -125,7 +125,7 @@ constexpr std::array<FormatTuple, MaxPixelFormat> FORMAT_TABLE = {{
     {GL_COMPRESSED_RGBA_ASTC_8x8_KHR},                                // ASTC_2D_8X8_UNORM
     {GL_COMPRESSED_RGBA_ASTC_8x5_KHR},                                // ASTC_2D_8X5_UNORM
     {GL_COMPRESSED_RGBA_ASTC_5x4_KHR},                                // ASTC_2D_5X4_UNORM
-    {GL_SRGB8_ALPHA8, GL_BGRA, GL_UNSIGNED_BYTE},                     // B8G8R8A8_UNORM
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE},                     // B8G8R8A8_SRGB
     {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT},                         // BC1_RGBA_SRGB
     {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT},                         // BC2_SRGB
     {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT},                         // BC3_SRGB
@@ -396,6 +396,17 @@ void AttachTexture(GLuint fbo, GLenum attachment, const ImageView* image_view) {
     }
 }
 
+[[nodiscard]] bool IsPixelFormatBGR(PixelFormat format) {
+    switch (format) {
+    case PixelFormat::B5G6R5_UNORM:
+    case PixelFormat::B8G8R8A8_UNORM:
+    case PixelFormat::B8G8R8A8_SRGB:
+        return true;
+    default:
+        return false;
+    }
+}
+
 } // Anonymous namespace
 
 ImageBufferMap::~ImageBufferMap() {
@@ -512,6 +523,9 @@ bool TextureCacheRuntime::CanImageBeCopied(const Image& dst, const Image& src) {
     if (dst.info.type == ImageType::e3D && dst.info.format == PixelFormat::BC4_UNORM) {
         return false;
     }
+    if (IsPixelFormatBGR(dst.info.format) || IsPixelFormatBGR(src.info.format)) {
+        return false;
+    }
     return true;
 }
 
@@ -520,6 +534,8 @@ void TextureCacheRuntime::EmulateCopyImage(Image& dst, Image& src,
     if (dst.info.type == ImageType::e3D && dst.info.format == PixelFormat::BC4_UNORM) {
         ASSERT(src.info.type == ImageType::e3D);
         util_shaders.CopyBC4(dst, src, copies);
+    } else if (IsPixelFormatBGR(dst.info.format) || IsPixelFormatBGR(src.info.format)) {
+        util_shaders.CopyBGR(dst, src, copies);
     } else {
         UNREACHABLE();
     }

src/video_core/renderer_opengl/gl_texture_cache.h

@@ -86,6 +86,11 @@ public:
 
     FormatProperties FormatInfo(VideoCommon::ImageType type, GLenum internal_format) const;
 
+    bool HasNativeBgr() const noexcept {
+        // OpenGL does not have native support for the BGR internal format
+        return false;
+    }
+
     bool HasBrokenTextureViewFormats() const noexcept {
         return has_broken_texture_view_formats;
     }

src/video_core/renderer_opengl/util_shaders.cpp

@@ -14,6 +14,7 @@
 #include "video_core/host_shaders/block_linear_unswizzle_2d_comp.h"
 #include "video_core/host_shaders/block_linear_unswizzle_3d_comp.h"
 #include "video_core/host_shaders/opengl_copy_bc4_comp.h"
+#include "video_core/host_shaders/opengl_copy_bgra_comp.h"
 #include "video_core/host_shaders/pitch_unswizzle_comp.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
 #include "video_core/renderer_opengl/gl_shader_manager.h"
@@ -48,6 +49,11 @@ OGLProgram MakeProgram(std::string_view source) {
     return program;
 }
 
+size_t NumPixelsInCopy(const VideoCommon::ImageCopy& copy) {
+    return static_cast<size_t>(copy.extent.width * copy.extent.height *
+                               copy.src_subresource.num_layers);
+}
+
 } // Anonymous namespace
 
 UtilShaders::UtilShaders(ProgramManager& program_manager_)
@@ -55,6 +61,7 @@ UtilShaders::UtilShaders(ProgramManager& program_manager_)
       block_linear_unswizzle_2d_program(MakeProgram(BLOCK_LINEAR_UNSWIZZLE_2D_COMP)),
       block_linear_unswizzle_3d_program(MakeProgram(BLOCK_LINEAR_UNSWIZZLE_3D_COMP)),
       pitch_unswizzle_program(MakeProgram(PITCH_UNSWIZZLE_COMP)),
+      copy_bgra_program(MakeProgram(OPENGL_COPY_BGRA_COMP)),
       copy_bc4_program(MakeProgram(OPENGL_COPY_BC4_COMP)) {
     const auto swizzle_table = Tegra::Texture::MakeSwizzleTable();
     swizzle_table_buffer.Create();
@@ -205,6 +212,43 @@ void UtilShaders::CopyBC4(Image& dst_image, Image& src_image, std::span<const Im
     program_manager.RestoreGuestCompute();
 }
 
+void UtilShaders::CopyBGR(Image& dst_image, Image& src_image,
+                          std::span<const VideoCommon::ImageCopy> copies) {
+    static constexpr GLuint BINDING_INPUT_IMAGE = 0;
+    static constexpr GLuint BINDING_OUTPUT_IMAGE = 1;
+    static constexpr VideoCommon::Offset3D zero_offset{0, 0, 0};
+    const u32 bytes_per_block = BytesPerBlock(dst_image.info.format);
+    switch (bytes_per_block) {
+    case 2:
+        // BGR565 copy
+        for (const ImageCopy& copy : copies) {
+            ASSERT(copy.src_offset == zero_offset);
+            ASSERT(copy.dst_offset == zero_offset);
+            bgr_copy_pass.Execute(dst_image, src_image, copy);
+        }
+        break;
+    case 4: {
+        // BGRA8 copy
+        program_manager.BindHostCompute(copy_bgra_program.handle);
+        constexpr GLenum FORMAT = GL_RGBA8;
+        for (const ImageCopy& copy : copies) {
+            ASSERT(copy.src_offset == zero_offset);
+            ASSERT(copy.dst_offset == zero_offset);
+            glBindImageTexture(BINDING_INPUT_IMAGE, src_image.StorageHandle(),
+                               copy.src_subresource.base_level, GL_FALSE, 0, GL_READ_ONLY, FORMAT);
+            glBindImageTexture(BINDING_OUTPUT_IMAGE, dst_image.StorageHandle(),
+                               copy.dst_subresource.base_level, GL_FALSE, 0, GL_WRITE_ONLY, FORMAT);
+            glDispatchCompute(copy.extent.width, copy.extent.height, copy.extent.depth);
+        }
+        program_manager.RestoreGuestCompute();
+        break;
+    }
+    default:
+        UNREACHABLE();
+        break;
+    }
+}
+
 GLenum StoreFormat(u32 bytes_per_block) {
     switch (bytes_per_block) {
     case 1:
@@ -222,4 +266,36 @@ GLenum StoreFormat(u32 bytes_per_block) {
     return GL_R8UI;
 }
 
+void Bgr565CopyPass::Execute(const Image& dst_image, const Image& src_image,
+                             const ImageCopy& copy) {
+    if (CopyBufferCreationNeeded(copy)) {
+        CreateNewCopyBuffer(copy, GL_TEXTURE_2D_ARRAY, GL_RGB565);
+    }
+    // Copy from source to PBO
+    glPixelStorei(GL_PACK_ALIGNMENT, 1);
+    glPixelStorei(GL_PACK_ROW_LENGTH, copy.extent.width);
+    glBindBuffer(GL_PIXEL_PACK_BUFFER, bgr16_pbo.handle);
+    glGetTextureSubImage(src_image.Handle(), 0, 0, 0, 0, copy.extent.width, copy.extent.height,
+                         copy.src_subresource.num_layers, GL_RGB, GL_UNSIGNED_SHORT_5_6_5,
+                         static_cast<GLsizei>(bgr16_pbo_size), nullptr);
+
+    // Copy from PBO to destination in reverse order
+    glPixelStorei(GL_UNPACK_ALIGNMENT, 1);
+    glPixelStorei(GL_UNPACK_ROW_LENGTH, copy.extent.width);
+    glBindBuffer(GL_PIXEL_UNPACK_BUFFER, bgr16_pbo.handle);
+    glTextureSubImage3D(dst_image.Handle(), 0, 0, 0, 0, copy.extent.width, copy.extent.height,
+                        copy.dst_subresource.num_layers, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV,
+                        nullptr);
+}
+
+bool Bgr565CopyPass::CopyBufferCreationNeeded(const ImageCopy& copy) {
+    return bgr16_pbo_size < NumPixelsInCopy(copy) * sizeof(u16);
+}
+
+void Bgr565CopyPass::CreateNewCopyBuffer(const ImageCopy& copy, GLenum target, GLuint format) {
+    bgr16_pbo.Create();
+    bgr16_pbo_size = NumPixelsInCopy(copy) * sizeof(u16);
+    glNamedBufferData(bgr16_pbo.handle, bgr16_pbo_size, nullptr, GL_STREAM_COPY);
+}
+
 } // namespace OpenGL

src/video_core/renderer_opengl/util_shaders.h

@@ -19,6 +19,22 @@ class ProgramManager;
 
 struct ImageBufferMap;
 
+class Bgr565CopyPass {
+public:
+    Bgr565CopyPass() = default;
+    ~Bgr565CopyPass() = default;
+
+    void Execute(const Image& dst_image, const Image& src_image,
+                 const VideoCommon::ImageCopy& copy);
+
+private:
+    [[nodiscard]] bool CopyBufferCreationNeeded(const VideoCommon::ImageCopy& copy);
+    void CreateNewCopyBuffer(const VideoCommon::ImageCopy& copy, GLenum target, GLuint format);
+
+    OGLBuffer bgr16_pbo;
+    size_t bgr16_pbo_size{};
+};
+
 class UtilShaders {
 public:
     explicit UtilShaders(ProgramManager& program_manager);
@@ -36,6 +52,9 @@ public:
     void CopyBC4(Image& dst_image, Image& src_image,
                  std::span<const VideoCommon::ImageCopy> copies);
 
+    void CopyBGR(Image& dst_image, Image& src_image,
+                 std::span<const VideoCommon::ImageCopy> copies);
+
 private:
     ProgramManager& program_manager;
 
@@ -44,7 +63,10 @@ private:
     OGLProgram block_linear_unswizzle_2d_program;
     OGLProgram block_linear_unswizzle_3d_program;
     OGLProgram pitch_unswizzle_program;
+    OGLProgram copy_bgra_program;
     OGLProgram copy_bc4_program;
+
+    Bgr565CopyPass bgr_copy_pass;
 };
 
 GLenum StoreFormat(u32 bytes_per_block);

src/video_core/renderer_vulkan/vk_command_pool.cpp

@@ -10,7 +10,7 @@
 
 namespace Vulkan {
 
-constexpr size_t COMMAND_BUFFER_POOL_SIZE = 0x1000;
+constexpr size_t COMMAND_BUFFER_POOL_SIZE = 4;
 
 struct CommandPool::Pool {
     vk::CommandPool handle;

src/video_core/renderer_vulkan/vk_texture_cache.h

@@ -93,6 +93,11 @@ struct TextureCacheRuntime {
         // No known Vulkan driver has broken image views
         return false;
     }
+
+    bool HasNativeBgr() const noexcept {
+        // All known Vulkan drivers can natively handle BGR textures
+        return true;
+    }
 };
 
 class Image : public VideoCommon::ImageBase {

src/video_core/texture_cache/image_base.cpp

@@ -120,9 +120,10 @@ void AddImageAlias(ImageBase& lhs, ImageBase& rhs, ImageId lhs_id, ImageId rhs_i
     if (lhs.info.type == ImageType::Linear) {
         base = SubresourceBase{.level = 0, .layer = 0};
     } else {
-        // We are passing relaxed formats as an option, having broken views or not won't matter
+        // We are passing relaxed formats as an option, having broken views/bgr or not won't matter
         static constexpr bool broken_views = false;
-        base = FindSubresource(rhs.info, lhs, rhs.gpu_addr, OPTIONS, broken_views);
+        static constexpr bool native_bgr = true;
+        base = FindSubresource(rhs.info, lhs, rhs.gpu_addr, OPTIONS, broken_views, native_bgr);
     }
     if (!base) {
         LOG_ERROR(HW_GPU, "Image alias should have been flipped");

src/video_core/texture_cache/image_view_base.cpp

@@ -24,7 +24,7 @@ ImageViewBase::ImageViewBase(const ImageViewInfo& info, const ImageInfo& image_i
           .height = std::max(image_info.size.height >> range.base.level, 1u),
           .depth = std::max(image_info.size.depth >> range.base.level, 1u),
       } {
-    ASSERT_MSG(VideoCore::Surface::IsViewCompatible(image_info.format, info.format, false),
+    ASSERT_MSG(VideoCore::Surface::IsViewCompatible(image_info.format, info.format, false, true),
                "Image view format {} is incompatible with image format {}", info.format,
                image_info.format);
     const bool is_async = Settings::values.use_asynchronous_gpu_emulation.GetValue();

src/video_core/texture_cache/texture_cache.h

@@ -876,6 +876,7 @@ ImageId TextureCache<P>::FindImage(const ImageInfo& info, GPUVAddr gpu_addr,
         return ImageId{};
     }
     const bool broken_views = runtime.HasBrokenTextureViewFormats();
+    const bool native_bgr = runtime.HasNativeBgr();
     ImageId image_id;
     const auto lambda = [&](ImageId existing_image_id, ImageBase& existing_image) {
         if (info.type == ImageType::Linear || existing_image.info.type == ImageType::Linear) {
@@ -885,11 +886,12 @@ ImageId TextureCache<P>::FindImage(const ImageInfo& info, GPUVAddr gpu_addr,
             if (existing_image.gpu_addr == gpu_addr && existing.type == info.type &&
                 existing.pitch == info.pitch &&
                 IsPitchLinearSameSize(existing, info, strict_size) &&
-                IsViewCompatible(existing.format, info.format, broken_views)) {
+                IsViewCompatible(existing.format, info.format, broken_views, native_bgr)) {
                 image_id = existing_image_id;
                 return true;
             }
-        } else if (IsSubresource(info, existing_image, gpu_addr, options, broken_views)) {
+        } else if (IsSubresource(info, existing_image, gpu_addr, options, broken_views,
+                                 native_bgr)) {
             image_id = existing_image_id;
             return true;
         }
@@ -920,6 +922,7 @@ ImageId TextureCache<P>::JoinImages(const ImageInfo& info, GPUVAddr gpu_addr, VA
     ImageInfo new_info = info;
     const size_t size_bytes = CalculateGuestSizeInBytes(new_info);
     const bool broken_views = runtime.HasBrokenTextureViewFormats();
+    const bool native_bgr = runtime.HasNativeBgr();
     std::vector<ImageId> overlap_ids;
     std::vector<ImageId> left_aliased_ids;
     std::vector<ImageId> right_aliased_ids;
@@ -935,8 +938,8 @@ ImageId TextureCache<P>::JoinImages(const ImageInfo& info, GPUVAddr gpu_addr, VA
             return;
         }
         static constexpr bool strict_size = true;
-        const std::optional<OverlapResult> solution =
-            ResolveOverlap(new_info, gpu_addr, cpu_addr, overlap, strict_size, broken_views);
+        const std::optional<OverlapResult> solution = ResolveOverlap(
+            new_info, gpu_addr, cpu_addr, overlap, strict_size, broken_views, native_bgr);
         if (solution) {
             gpu_addr = solution->gpu_addr;
             cpu_addr = solution->cpu_addr;
@@ -946,10 +949,10 @@ ImageId TextureCache<P>::JoinImages(const ImageInfo& info, GPUVAddr gpu_addr, VA
         }
         static constexpr auto options = RelaxedOptions::Size | RelaxedOptions::Format;
         const ImageBase new_image_base(new_info, gpu_addr, cpu_addr);
-        if (IsSubresource(new_info, overlap, gpu_addr, options, broken_views)) {
+        if (IsSubresource(new_info, overlap, gpu_addr, options, broken_views, native_bgr)) {
             left_aliased_ids.push_back(overlap_id);
         } else if (IsSubresource(overlap.info, new_image_base, overlap.gpu_addr, options,
-                                 broken_views)) {
+                                 broken_views, native_bgr)) {
             right_aliased_ids.push_back(overlap_id);
         }
     });

src/video_core/texture_cache/util.cpp

@@ -1035,13 +1035,13 @@ bool IsPitchLinearSameSize(const ImageInfo& lhs, const ImageInfo& rhs, bool stri
 
 std::optional<OverlapResult> ResolveOverlap(const ImageInfo& new_info, GPUVAddr gpu_addr,
                                             VAddr cpu_addr, const ImageBase& overlap,
-                                            bool strict_size, bool broken_views) {
+                                            bool strict_size, bool broken_views, bool native_bgr) {
     ASSERT(new_info.type != ImageType::Linear);
     ASSERT(overlap.info.type != ImageType::Linear);
     if (!IsLayerStrideCompatible(new_info, overlap.info)) {
         return std::nullopt;
     }
-    if (!IsViewCompatible(overlap.info.format, new_info.format, broken_views)) {
+    if (!IsViewCompatible(overlap.info.format, new_info.format, broken_views, native_bgr)) {
         return std::nullopt;
     }
     if (gpu_addr == overlap.gpu_addr) {
@@ -1085,14 +1085,14 @@ bool IsLayerStrideCompatible(const ImageInfo& lhs, const ImageInfo& rhs) {
 
 std::optional<SubresourceBase> FindSubresource(const ImageInfo& candidate, const ImageBase& image,
                                                GPUVAddr candidate_addr, RelaxedOptions options,
-                                               bool broken_views) {
+                                               bool broken_views, bool native_bgr) {
     const std::optional<SubresourceBase> base = image.TryFindBase(candidate_addr);
     if (!base) {
         return std::nullopt;
     }
     const ImageInfo& existing = image.info;
     if (False(options & RelaxedOptions::Format)) {
-        if (!IsViewCompatible(existing.format, candidate.format, broken_views)) {
+        if (!IsViewCompatible(existing.format, candidate.format, broken_views, native_bgr)) {
             return std::nullopt;
         }
     }
@@ -1129,8 +1129,9 @@ std::optional<SubresourceBase> FindSubresource(const ImageInfo& candidate, const
 }
 
 bool IsSubresource(const ImageInfo& candidate, const ImageBase& image, GPUVAddr candidate_addr,
-                   RelaxedOptions options, bool broken_views) {
-    return FindSubresource(candidate, image, candidate_addr, options, broken_views).has_value();
+                   RelaxedOptions options, bool broken_views, bool native_bgr) {
+    return FindSubresource(candidate, image, candidate_addr, options, broken_views, native_bgr)
+        .has_value();
 }
 
 void DeduceBlitImages(ImageInfo& dst_info, ImageInfo& src_info, const ImageBase* dst,

src/video_core/texture_cache/util.h

@@ -87,7 +87,8 @@ void SwizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const Ima
 [[nodiscard]] std::optional<OverlapResult> ResolveOverlap(const ImageInfo& new_info,
                                                           GPUVAddr gpu_addr, VAddr cpu_addr,
                                                           const ImageBase& overlap,
-                                                          bool strict_size, bool broken_views);
+                                                          bool strict_size, bool broken_views,
+                                                          bool native_bgr);
 
 [[nodiscard]] bool IsLayerStrideCompatible(const ImageInfo& lhs, const ImageInfo& rhs);
 
@@ -95,11 +96,11 @@ void SwizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const Ima
                                                              const ImageBase& image,
                                                              GPUVAddr candidate_addr,
                                                              RelaxedOptions options,
-                                                             bool broken_views);
+                                                             bool broken_views, bool native_bgr);
 
 [[nodiscard]] bool IsSubresource(const ImageInfo& candidate, const ImageBase& image,
-                                 GPUVAddr candidate_addr, RelaxedOptions options,
-                                 bool broken_views);
+                                 GPUVAddr candidate_addr, RelaxedOptions options, bool broken_views,
+                                 bool native_bgr);
 
 void DeduceBlitImages(ImageInfo& dst_info, ImageInfo& src_info, const ImageBase* dst,
                       const ImageBase* src);

src/yuzu/bootmanager.cpp

@@ -376,11 +376,34 @@ void GRenderWindow::closeEvent(QCloseEvent* event) {
 }
 
 void GRenderWindow::keyPressEvent(QKeyEvent* event) {
-    input_subsystem->GetKeyboard()->PressKey(event->key());
+    if (!event->isAutoRepeat()) {
+        input_subsystem->GetKeyboard()->PressKey(event->key());
+    }
 }
 
 void GRenderWindow::keyReleaseEvent(QKeyEvent* event) {
-    input_subsystem->GetKeyboard()->ReleaseKey(event->key());
+    if (!event->isAutoRepeat()) {
+        input_subsystem->GetKeyboard()->ReleaseKey(event->key());
+    }
+}
+
+MouseInput::MouseButton GRenderWindow::QtButtonToMouseButton(Qt::MouseButton button) {
+    switch (button) {
+    case Qt::LeftButton:
+        return MouseInput::MouseButton::Left;
+    case Qt::RightButton:
+        return MouseInput::MouseButton::Right;
+    case Qt::MiddleButton:
+        return MouseInput::MouseButton::Wheel;
+    case Qt::BackButton:
+        return MouseInput::MouseButton::Backward;
+    case Qt::ForwardButton:
+        return MouseInput::MouseButton::Forward;
+    case Qt::TaskButton:
+        return MouseInput::MouseButton::Task;
+    default:
+        return MouseInput::MouseButton::Extra;
+    }
 }
 
 void GRenderWindow::mousePressEvent(QMouseEvent* event) {
@@ -391,7 +414,8 @@ void GRenderWindow::mousePressEvent(QMouseEvent* event) {
 
     auto pos = event->pos();
     const auto [x, y] = ScaleTouch(pos);
-    input_subsystem->GetMouse()->PressButton(x, y, event->button());
+    const auto button = QtButtonToMouseButton(event->button());
+    input_subsystem->GetMouse()->PressButton(x, y, button);
 
     if (event->button() == Qt::LeftButton) {
         this->TouchPressed(x, y, 0);
@@ -425,7 +449,8 @@ void GRenderWindow::mouseReleaseEvent(QMouseEvent* event) {
         return;
     }
 
-    input_subsystem->GetMouse()->ReleaseButton(event->button());
+    const auto button = QtButtonToMouseButton(event->button());
+    input_subsystem->GetMouse()->ReleaseButton(button);
 
     if (event->button() == Qt::LeftButton) {
         this->TouchReleased(0);

src/yuzu/bootmanager.h

@@ -28,6 +28,10 @@ namespace InputCommon {
 class InputSubsystem;
 }
 
+namespace MouseInput {
+enum class MouseButton;
+}
+
 namespace VideoCore {
 enum class LoadCallbackStage;
 }
@@ -149,6 +153,9 @@ public:
     void keyPressEvent(QKeyEvent* event) override;
     void keyReleaseEvent(QKeyEvent* event) override;
 
+    /// Converts a Qt mouse button into MouseInput mouse button
+    static MouseInput::MouseButton QtButtonToMouseButton(Qt::MouseButton button);
+
     void mousePressEvent(QMouseEvent* event) override;
     void mouseMoveEvent(QMouseEvent* event) override;
     void mouseReleaseEvent(QMouseEvent* event) override;

src/yuzu/configuration/config.cpp

@@ -235,7 +235,7 @@ const std::array<UISettings::Shortcut, 17> Config::default_hotkeys{{
     {QStringLiteral("Restart Emulation"),        QStringLiteral("Main Window"), {QStringLiteral("F6"), Qt::WindowShortcut}},
     {QStringLiteral("Stop Emulation"),           QStringLiteral("Main Window"), {QStringLiteral("F5"), Qt::WindowShortcut}},
     {QStringLiteral("Toggle Filter Bar"),        QStringLiteral("Main Window"), {QStringLiteral("Ctrl+F"), Qt::WindowShortcut}},
-    {QStringLiteral("Toggle Mouse Panning"),     QStringLiteral("Main Window"), {QStringLiteral("F9"), Qt::ApplicationShortcut}},
+    {QStringLiteral("Toggle Mouse Panning"),     QStringLiteral("Main Window"), {QStringLiteral("Ctrl+F9"), Qt::ApplicationShortcut}},
     {QStringLiteral("Toggle Speed Limit"),       QStringLiteral("Main Window"), {QStringLiteral("Ctrl+Z"), Qt::ApplicationShortcut}},
     {QStringLiteral("Toggle Status Bar"),        QStringLiteral("Main Window"), {QStringLiteral("Ctrl+S"), Qt::WindowShortcut}},
 }};
@@ -508,7 +508,7 @@ void Config::ReadControlValues() {
 
     Settings::values.emulate_analog_keyboard =
         ReadSetting(QStringLiteral("emulate_analog_keyboard"), false).toBool();
-    Settings::values.mouse_panning = ReadSetting(QStringLiteral("mouse_panning"), false).toBool();
+    Settings::values.mouse_panning = false;
     Settings::values.mouse_panning_sensitivity =
         ReadSetting(QStringLiteral("mouse_panning_sensitivity"), 1).toFloat();
 
@@ -648,7 +648,7 @@ void Config::ReadDebuggingValues() {
 void Config::ReadServiceValues() {
     qt_config->beginGroup(QStringLiteral("Services"));
     Settings::values.bcat_backend =
-        ReadSetting(QStringLiteral("bcat_backend"), QStringLiteral("null"))
+        ReadSetting(QStringLiteral("bcat_backend"), QStringLiteral("none"))
             .toString()
             .toStdString();
     Settings::values.bcat_boxcat_local =
@@ -1182,7 +1182,6 @@ void Config::SaveControlValues() {
     WriteSetting(QStringLiteral("keyboard_enabled"), Settings::values.keyboard_enabled, false);
     WriteSetting(QStringLiteral("emulate_analog_keyboard"),
                  Settings::values.emulate_analog_keyboard, false);
-    WriteSetting(QStringLiteral("mouse_panning"), Settings::values.mouse_panning, false);
     WriteSetting(QStringLiteral("mouse_panning_sensitivity"),
                  Settings::values.mouse_panning_sensitivity, 1.0f);
     qt_config->endGroup();
@@ -1239,7 +1238,7 @@ void Config::SaveDebuggingValues() {
 void Config::SaveServiceValues() {
     qt_config->beginGroup(QStringLiteral("Services"));
     WriteSetting(QStringLiteral("bcat_backend"),
-                 QString::fromStdString(Settings::values.bcat_backend), QStringLiteral("null"));
+                 QString::fromStdString(Settings::values.bcat_backend), QStringLiteral("none"));
     WriteSetting(QStringLiteral("bcat_boxcat_local"), Settings::values.bcat_boxcat_local, false);
     qt_config->endGroup();
 }

src/yuzu/configuration/configure_filesystem.cpp

@@ -103,7 +103,10 @@ void ConfigureFilesystem::SetDirectory(DirectoryTarget target, QLineEdit* edit)
         str = QFileDialog::getOpenFileName(this, caption, QFileInfo(edit->text()).dir().path(),
                                            QStringLiteral("NX Gamecard;*.xci"));
     } else {
-        str = QFileDialog::getExistingDirectory(this, caption, edit->text()) + QDir::separator();
+        str = QFileDialog::getExistingDirectory(this, caption, edit->text());
+        if (!str.isNull() && str.back() != QDir::separator()) {
+            str.append(QDir::separator());
+        }
     }
 
     if (str.isEmpty())

src/yuzu/configuration/configure_input_player.cpp

@@ -21,6 +21,7 @@
 #include "input_common/mouse/mouse_poller.h"
 #include "input_common/udp/udp.h"
 #include "ui_configure_input_player.h"
+#include "yuzu/bootmanager.h"
 #include "yuzu/configuration/config.h"
 #include "yuzu/configuration/configure_input_player.h"
 #include "yuzu/configuration/configure_input_player_widget.h"
@@ -104,7 +105,9 @@ QString ButtonToText(const Common::ParamPackage& param) {
     }
 
     if (param.Get("engine", "") == "keyboard") {
-        return GetKeyName(param.Get("code", 0));
+        const QString button_str = GetKeyName(param.Get("code", 0));
+        const QString toggle = QString::fromStdString(param.Get("toggle", false) ? "~" : "");
+        return QObject::tr("%1%2").arg(toggle, button_str);
     }
 
     if (param.Get("engine", "") == "gcpad") {
@@ -156,7 +159,8 @@ QString ButtonToText(const Common::ParamPackage& param) {
     if (param.Get("engine", "") == "mouse") {
         if (param.Has("button")) {
             const QString button_str = QString::number(int(param.Get("button", 0)));
-            return QObject::tr("Click %1").arg(button_str);
+            const QString toggle = QString::fromStdString(param.Get("toggle", false) ? "~" : "");
+            return QObject::tr("%1Click %2").arg(toggle, button_str);
         }
         return GetKeyName(param.Get("code", 0));
     }
@@ -300,6 +304,11 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
                         buttons_param[button_id].Clear();
                         button_map[button_id]->setText(tr("[not set]"));
                     });
+                    context_menu.addAction(tr("Toggle button"), [&] {
+                        const bool toggle_value = !buttons_param[button_id].Get("toggle", false);
+                        buttons_param[button_id].Set("toggle", toggle_value);
+                        button_map[button_id]->setText(ButtonToText(buttons_param[button_id]));
+                    });
                     context_menu.exec(button_map[button_id]->mapToGlobal(menu_location));
                     ui->controllerFrame->SetPlayerInput(player_index, buttons_param, analogs_param);
                 });
@@ -412,6 +421,15 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
                         analogs_param[analog_id].Set("modifier", "");
                         analog_map_modifier_button[analog_id]->setText(tr("[not set]"));
                     });
+                    context_menu.addAction(tr("Toggle button"), [&] {
+                        Common::ParamPackage modifier_param =
+                            Common::ParamPackage{analogs_param[analog_id].Get("modifier", "")};
+                        const bool toggle_value = !modifier_param.Get("toggle", false);
+                        modifier_param.Set("toggle", toggle_value);
+                        analogs_param[analog_id].Set("modifier", modifier_param.Serialize());
+                        analog_map_modifier_button[analog_id]->setText(
+                            ButtonToText(modifier_param));
+                    });
                     context_menu.exec(
                         analog_map_modifier_button[analog_id]->mapToGlobal(menu_location));
                 });
@@ -1345,7 +1363,8 @@ void ConfigureInputPlayer::mousePressEvent(QMouseEvent* event) {
         return;
     }
 
-    input_subsystem->GetMouse()->PressButton(0, 0, event->button());
+    const auto button = GRenderWindow::QtButtonToMouseButton(event->button());
+    input_subsystem->GetMouse()->PressButton(0, 0, button);
 }
 
 void ConfigureInputPlayer::keyPressEvent(QKeyEvent* event) {

src/yuzu/configuration/configure_motion_touch.cpp

@@ -24,7 +24,7 @@
 
 CalibrationConfigurationDialog::CalibrationConfigurationDialog(QWidget* parent,
                                                                const std::string& host, u16 port,
-                                                               u8 pad_index, u16 client_id)
+                                                               u8 pad_index)
     : QDialog(parent) {
     layout = new QVBoxLayout;
     status_label = new QLabel(tr("Communicating with the server..."));
@@ -41,7 +41,7 @@ CalibrationConfigurationDialog::CalibrationConfigurationDialog(QWidget* parent,
 
     using namespace InputCommon::CemuhookUDP;
     job = std::make_unique<CalibrationConfigurationJob>(
-        host, port, pad_index, client_id,
+        host, port, pad_index,
         [this](CalibrationConfigurationJob::Status status) {
             QString text;
             switch (status) {
@@ -218,7 +218,6 @@ void ConfigureMotionTouch::OnCemuhookUDPTest() {
     udp_test_in_progress = true;
     InputCommon::CemuhookUDP::TestCommunication(
         ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toInt()), 0,
-        24872,
         [this] {
             LOG_INFO(Frontend, "UDP input test success");
             QMetaObject::invokeMethod(this, "ShowUDPTestResult", Q_ARG(bool, true));
@@ -233,8 +232,7 @@ void ConfigureMotionTouch::OnConfigureTouchCalibration() {
     ui->touch_calibration_config->setEnabled(false);
     ui->touch_calibration_config->setText(tr("Configuring"));
     CalibrationConfigurationDialog dialog(this, ui->udp_server->text().toStdString(),
-                                          static_cast<u16>(ui->udp_port->text().toUInt()), 0,
-                                          24872);
+                                          static_cast<u16>(ui->udp_port->text().toUInt()), 0);
     dialog.exec();
     if (dialog.completed) {
         min_x = dialog.min_x;

src/yuzu/configuration/configure_motion_touch.h

@@ -30,7 +30,7 @@ class CalibrationConfigurationDialog : public QDialog {
     Q_OBJECT
 public:
     explicit CalibrationConfigurationDialog(QWidget* parent, const std::string& host, u16 port,
-                                            u8 pad_index, u16 client_id);
+                                            u8 pad_index);
     ~CalibrationConfigurationDialog() override;
 
 private:

src/yuzu/main.cpp

@@ -60,6 +60,7 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual
 #include <QPushButton>
 #include <QShortcut>
 #include <QStatusBar>
+#include <QString>
 #include <QSysInfo>
 #include <QUrl>
 #include <QtConcurrent/QtConcurrent>
@@ -854,8 +855,7 @@ void GMainWindow::InitializeHotkeys() {
     connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Toggle Mouse Panning"), this),
             &QShortcut::activated, this, [&] {
                 Settings::values.mouse_panning = !Settings::values.mouse_panning;
-                if (UISettings::values.hide_mouse || Settings::values.mouse_panning) {
-                    mouse_hide_timer.start();
+                if (Settings::values.mouse_panning) {
                     render_window->installEventFilter(render_window);
                     render_window->setAttribute(Qt::WA_Hover, true);
                 }
@@ -1208,11 +1208,14 @@ void GMainWindow::BootGame(const QString& filename, std::size_t program_index) {
     renderer_status_button->setDisabled(true);
 
     if (UISettings::values.hide_mouse || Settings::values.mouse_panning) {
-        mouse_hide_timer.start();
         render_window->installEventFilter(render_window);
         render_window->setAttribute(Qt::WA_Hover, true);
     }
 
+    if (UISettings::values.hide_mouse) {
+        mouse_hide_timer.start();
+    }
+
     std::string title_name;
     std::string title_version;
     const auto res = system.GetGameName(title_name);
@@ -2372,12 +2375,15 @@ void GMainWindow::OnConfigure() {
     if ((UISettings::values.hide_mouse || Settings::values.mouse_panning) && emulation_running) {
         render_window->installEventFilter(render_window);
         render_window->setAttribute(Qt::WA_Hover, true);
-        mouse_hide_timer.start();
     } else {
         render_window->removeEventFilter(render_window);
         render_window->setAttribute(Qt::WA_Hover, false);
     }
 
+    if (UISettings::values.hide_mouse) {
+        mouse_hide_timer.start();
+    }
+
     UpdateStatusButtons();
 }
 
@@ -2615,8 +2621,7 @@ void GMainWindow::UpdateUISettings() {
 }
 
 void GMainWindow::HideMouseCursor() {
-    if (emu_thread == nullptr ||
-        (!UISettings::values.hide_mouse && !Settings::values.mouse_panning)) {
+    if (emu_thread == nullptr && UISettings::values.hide_mouse) {
         mouse_hide_timer.stop();
         ShowMouseCursor();
         return;
@@ -2626,8 +2631,7 @@ void GMainWindow::HideMouseCursor() {
 
 void GMainWindow::ShowMouseCursor() {
     render_window->unsetCursor();
-    if (emu_thread != nullptr &&
-        (UISettings::values.hide_mouse || Settings::values.mouse_panning)) {
+    if (emu_thread != nullptr && UISettings::values.hide_mouse) {
         mouse_hide_timer.start();
     }
 }
@@ -3058,6 +3062,14 @@ int main(int argc, char* argv[]) {
     chdir(bin_path.c_str());
 #endif
 
+#ifdef __linux__
+    // Set the DISPLAY variable in order to open web browsers
+    // TODO (lat9nq): Find a better solution for AppImages to start external applications
+    if (QString::fromLocal8Bit(qgetenv("DISPLAY")).isEmpty()) {
+        qputenv("DISPLAY", ":0");
+    }
+#endif
+
     // Enables the core to make the qt created contexts current on std::threads
     QCoreApplication::setAttribute(Qt::AA_DontCheckOpenGLContextThreadAffinity);
     QApplication app(argc, argv);

src/yuzu_cmd/config.cpp

@@ -457,7 +457,7 @@ void Config::ReadValues() {
     Settings::values.yuzu_token = sdl2_config->Get("WebService", "yuzu_token", "");
 
     // Services
-    Settings::values.bcat_backend = sdl2_config->Get("Services", "bcat_backend", "null");
+    Settings::values.bcat_backend = sdl2_config->Get("Services", "bcat_backend", "none");
     Settings::values.bcat_boxcat_local =
         sdl2_config->GetBoolean("Services", "bcat_boxcat_local", false);
 }

src/yuzu_cmd/emu_window/emu_window_sdl2.cpp

@@ -35,18 +35,36 @@ void EmuWindow_SDL2::OnMouseMotion(s32 x, s32 y) {
     input_subsystem->GetMouse()->MouseMove(x, y, 0, 0);
 }
 
+MouseInput::MouseButton EmuWindow_SDL2::SDLButtonToMouseButton(u32 button) const {
+    switch (button) {
+    case SDL_BUTTON_LEFT:
+        return MouseInput::MouseButton::Left;
+    case SDL_BUTTON_RIGHT:
+        return MouseInput::MouseButton::Right;
+    case SDL_BUTTON_MIDDLE:
+        return MouseInput::MouseButton::Wheel;
+    case SDL_BUTTON_X1:
+        return MouseInput::MouseButton::Backward;
+    case SDL_BUTTON_X2:
+        return MouseInput::MouseButton::Forward;
+    default:
+        return MouseInput::MouseButton::Undefined;
+    }
+}
+
 void EmuWindow_SDL2::OnMouseButton(u32 button, u8 state, s32 x, s32 y) {
+    const auto mouse_button = SDLButtonToMouseButton(button);
     if (button == SDL_BUTTON_LEFT) {
         if (state == SDL_PRESSED) {
             TouchPressed((unsigned)std::max(x, 0), (unsigned)std::max(y, 0), 0);
         } else {
             TouchReleased(0);
         }
-    } else if (button == SDL_BUTTON_RIGHT) {
+    } else {
         if (state == SDL_PRESSED) {
-            input_subsystem->GetMouse()->PressButton(x, y, button);
+            input_subsystem->GetMouse()->PressButton(x, y, mouse_button);
         } else {
-            input_subsystem->GetMouse()->ReleaseButton(button);
+            input_subsystem->GetMouse()->ReleaseButton(mouse_button);
         }
     }
 }

src/yuzu_cmd/emu_window/emu_window_sdl2.h

@@ -18,6 +18,10 @@ namespace InputCommon {
 class InputSubsystem;
 }
 
+namespace MouseInput {
+enum class MouseButton;
+}
+
 class EmuWindow_SDL2 : public Core::Frontend::EmuWindow {
 public:
     explicit EmuWindow_SDL2(InputCommon::InputSubsystem* input_subsystem);
@@ -42,6 +46,9 @@ protected:
     /// Called by WaitEvent when the mouse moves.
     void OnMouseMotion(s32 x, s32 y);
 
+    /// Converts a SDL mouse button into MouseInput mouse button
+    MouseInput::MouseButton SDLButtonToMouseButton(u32 button) const;
+
     /// Called by WaitEvent when a mouse button is pressed or released
     void OnMouseButton(u32 button, u8 state, s32 x, s32 y);