last clang-format fix

gl_shader_decompiler: Implement TEX, fixes for TEXS.

src/common/logging/backend.cpp

@@ -35,6 +35,7 @@ namespace Log {
     SUB(Service, AM)                                                                               \
     SUB(Service, AOC)                                                                              \
     SUB(Service, APM)                                                                              \
+    SUB(Service, BCAT)                                                                             \
     SUB(Service, Fatal)                                                                            \
     SUB(Service, Friend)                                                                           \
     SUB(Service, FS)                                                                               \

src/common/logging/log.h

@@ -55,6 +55,7 @@ enum class Class : ClassType {
     Service_AOC,       ///< The AOC (AddOn Content) service
     Service_APM,       ///< The APM (Performance) service
     Service_Audio,     ///< The Audio (Audio control) service
+    Service_BCAT,      ///< The BCAT service
     Service_Fatal,     ///< The Fatal service
     Service_Friend,    ///< The friend service
     Service_FS,        ///< The FS (Filesystem) service

src/core/CMakeLists.txt

@@ -4,6 +4,8 @@ add_library(core STATIC
     arm/unicorn/arm_unicorn.h
     core.cpp
     core.h
+    core_cpu.cpp
+    core_cpu.h
     core_timing.cpp
     core_timing.h
     file_sys/directory.h
@@ -122,6 +124,10 @@ add_library(core STATIC
     hle/service/audio/audren_u.h
     hle/service/audio/codecctl.cpp
     hle/service/audio/codecctl.h
+    hle/service/bcat/module.cpp
+    hle/service/bcat/module.h
+    hle/service/bcat/bcat.cpp
+    hle/service/bcat/bcat.h
     hle/service/fatal/fatal.cpp
     hle/service/fatal/fatal.h
     hle/service/fatal/fatal_p.cpp
@@ -169,6 +175,8 @@ add_library(core STATIC
     hle/service/nvdrv/devices/nvhost_ctrl_gpu.h
     hle/service/nvdrv/devices/nvhost_gpu.cpp
     hle/service/nvdrv/devices/nvhost_gpu.h
+    hle/service/nvdrv/devices/nvhost_nvdec.cpp
+    hle/service/nvdrv/devices/nvhost_nvdec.h
     hle/service/nvdrv/devices/nvmap.cpp
     hle/service/nvdrv/devices/nvmap.h
     hle/service/nvdrv/interface.cpp

src/core/arm/unicorn/arm_unicorn.cpp

@@ -52,7 +52,7 @@ static void InterruptHook(uc_engine* uc, u32 intNo, void* user_data) {
 static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int size, u64 value,
                                void* user_data) {
     ARM_Interface::ThreadContext ctx{};
-    Core::CPU().SaveContext(ctx);
+    Core::CurrentArmInterface().SaveContext(ctx);
     ASSERT_MSG(false, "Attempted to read from unmapped memory: 0x{:X}, pc=0x{:X}, lr=0x{:X}", addr,
                ctx.pc, ctx.cpu_registers[30]);
     return {};

src/core/core.cpp

@@ -5,10 +5,6 @@
 #include <memory>
 #include <utility>
 #include "common/logging/log.h"
-#ifdef ARCHITECTURE_x86_64
-#include "core/arm/dynarmic/arm_dynarmic.h"
-#endif
-#include "core/arm/unicorn/arm_unicorn.h"
 #include "core/core.h"
 #include "core/core_timing.h"
 #include "core/gdbstub/gdbstub.h"
@@ -31,11 +27,31 @@ namespace Core {
 
 System::~System() = default;
 
+/// Runs a CPU core while the system is powered on
+static void RunCpuCore(std::shared_ptr<Cpu> cpu_state) {
+    while (Core::System().GetInstance().IsPoweredOn()) {
+        cpu_state->RunLoop(true);
+    }
+}
+
+Cpu& System::CurrentCpuCore() {
+    // If multicore is enabled, use host thread to figure out the current CPU core
+    if (Settings::values.use_multi_core) {
+        const auto& search = thread_to_cpu.find(std::this_thread::get_id());
+        ASSERT(search != thread_to_cpu.end());
+        ASSERT(search->second);
+        return *search->second;
+    }
+
+    // Otherwise, use single-threaded mode active_core variable
+    return *cpu_cores[active_core];
+}
+
 System::ResultStatus System::RunLoop(bool tight_loop) {
     status = ResultStatus::Success;
-    if (!cpu_core) {
-        return ResultStatus::ErrorNotInitialized;
-    }
+
+    // Update thread_to_cpu in case Core 0 is run from a different host thread
+    thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
 
     if (GDBStub::IsServerEnabled()) {
         GDBStub::HandlePacket();
@@ -52,25 +68,14 @@ System::ResultStatus System::RunLoop(bool tight_loop) {
         }
     }
 
-    // If we don't have a currently active thread then don't execute instructions,
-    // instead advance to the next event and try to yield to the next thread
-    if (Kernel::GetCurrentThread() == nullptr) {
-        NGLOG_TRACE(Core_ARM, "Idling");
-        CoreTiming::Idle();
-        CoreTiming::Advance();
-        PrepareReschedule();
-    } else {
-        CoreTiming::Advance();
-        if (tight_loop) {
-            cpu_core->Run();
-        } else {
-            cpu_core->Step();
+    for (active_core = 0; active_core < NUM_CPU_CORES; ++active_core) {
+        cpu_cores[active_core]->RunLoop(tight_loop);
+        if (Settings::values.use_multi_core) {
+            // Cores 1-3 are run on other threads in this mode
+            break;
         }
     }
 
-    HW::Update();
-    Reschedule();
-
     return status;
 }
 
@@ -133,21 +138,26 @@ System::ResultStatus System::Load(EmuWindow* emu_window, const std::string& file
 }
 
 void System::PrepareReschedule() {
-    cpu_core->PrepareReschedule();
-    reschedule_pending = true;
+    CurrentCpuCore().PrepareReschedule();
 }
 
 PerfStats::Results System::GetAndResetPerfStats() {
     return perf_stats.GetAndResetStats(CoreTiming::GetGlobalTimeUs());
 }
 
-void System::Reschedule() {
-    if (!reschedule_pending) {
-        return;
-    }
+const std::shared_ptr<Kernel::Scheduler>& System::Scheduler(size_t core_index) {
+    ASSERT(core_index < NUM_CPU_CORES);
+    return cpu_cores[core_index]->Scheduler();
+}
 
-    reschedule_pending = false;
-    Core::System::GetInstance().Scheduler().Reschedule();
+ARM_Interface& System::ArmInterface(size_t core_index) {
+    ASSERT(core_index < NUM_CPU_CORES);
+    return cpu_cores[core_index]->ArmInterface();
+}
+
+Cpu& System::CpuCore(size_t core_index) {
+    ASSERT(core_index < NUM_CPU_CORES);
+    return *cpu_cores[core_index];
 }
 
 System::ResultStatus System::Init(EmuWindow* emu_window, u32 system_mode) {
@@ -157,26 +167,17 @@ System::ResultStatus System::Init(EmuWindow* emu_window, u32 system_mode) {
 
     current_process = Kernel::Process::Create("main");
 
-    if (Settings::values.use_cpu_jit) {
-#ifdef ARCHITECTURE_x86_64
-        cpu_core = std::make_shared<ARM_Dynarmic>();
-#else
-        cpu_core = std::make_shared<ARM_Unicorn>();
-        NGLOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
-#endif
-    } else {
-        cpu_core = std::make_shared<ARM_Unicorn>();
+    cpu_barrier = std::make_shared<CpuBarrier>();
+    for (size_t index = 0; index < cpu_cores.size(); ++index) {
+        cpu_cores[index] = std::make_shared<Cpu>(cpu_barrier, index);
     }
 
     gpu_core = std::make_unique<Tegra::GPU>();
-
     telemetry_session = std::make_unique<Core::TelemetrySession>();
-
     service_manager = std::make_shared<Service::SM::ServiceManager>();
 
     HW::Init();
     Kernel::Init(system_mode);
-    scheduler = std::make_unique<Kernel::Scheduler>(cpu_core.get());
     Service::Init(service_manager);
     GDBStub::Init();
 
@@ -184,6 +185,17 @@ System::ResultStatus System::Init(EmuWindow* emu_window, u32 system_mode) {
         return ResultStatus::ErrorVideoCore;
     }
 
+    // Create threads for CPU cores 1-3, and build thread_to_cpu map
+    // CPU core 0 is run on the main thread
+    thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
+    if (Settings::values.use_multi_core) {
+        for (size_t index = 0; index < cpu_core_threads.size(); ++index) {
+            cpu_core_threads[index] =
+                std::make_unique<std::thread>(RunCpuCore, cpu_cores[index + 1]);
+            thread_to_cpu[cpu_core_threads[index]->get_id()] = cpu_cores[index + 1];
+        }
+    }
+
     NGLOG_DEBUG(Core, "Initialized OK");
 
     // Reset counters and set time origin to current frame
@@ -207,15 +219,30 @@ void System::Shutdown() {
     VideoCore::Shutdown();
     GDBStub::Shutdown();
     Service::Shutdown();
-    scheduler.reset();
     Kernel::Shutdown();
     HW::Shutdown();
     service_manager.reset();
     telemetry_session.reset();
     gpu_core.reset();
-    cpu_core.reset();
+
+    // Close all CPU/threading state
+    cpu_barrier->NotifyEnd();
+    if (Settings::values.use_multi_core) {
+        for (auto& thread : cpu_core_threads) {
+            thread->join();
+            thread.reset();
+        }
+    }
+    thread_to_cpu.clear();
+    for (auto& cpu_core : cpu_cores) {
+        cpu_core.reset();
+    }
+    cpu_barrier.reset();
+
+    // Close core timing
     CoreTiming::Shutdown();
 
+    // Close app loader
     app_loader.reset();
 
     NGLOG_DEBUG(Core, "Shutdown OK");

src/core/core.h

@@ -4,9 +4,12 @@
 
 #pragma once
 
+#include <array>
 #include <memory>
 #include <string>
+#include <thread>
 #include "common/common_types.h"
+#include "core/core_cpu.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/scheduler.h"
 #include "core/loader/loader.h"
@@ -89,7 +92,7 @@ public:
      * @returns True if the emulated system is powered on, otherwise false.
      */
     bool IsPoweredOn() const {
-        return cpu_core != nullptr;
+        return cpu_barrier && cpu_barrier->IsAlive();
     }
 
     /**
@@ -103,24 +106,34 @@ public:
     /// Prepare the core emulation for a reschedule
     void PrepareReschedule();
 
+    /// Gets and resets core performance statistics
     PerfStats::Results GetAndResetPerfStats();
 
-    /**
-     * Gets a reference to the emulated CPU.
-     * @returns A reference to the emulated CPU.
-     */
-    ARM_Interface& CPU() {
-        return *cpu_core;
+    /// Gets an ARM interface to the CPU core that is currently running
+    ARM_Interface& CurrentArmInterface() {
+        return CurrentCpuCore().ArmInterface();
     }
 
+    /// Gets an ARM interface to the CPU core with the specified index
+    ARM_Interface& ArmInterface(size_t core_index);
+
+    /// Gets a CPU interface to the CPU core with the specified index
+    Cpu& CpuCore(size_t core_index);
+
+    /// Gets the GPU interface
     Tegra::GPU& GPU() {
         return *gpu_core;
     }
 
-    Kernel::Scheduler& Scheduler() {
-        return *scheduler;
+    /// Gets the scheduler for the CPU core that is currently running
+    Kernel::Scheduler& CurrentScheduler() {
+        return *CurrentCpuCore().Scheduler();
     }
 
+    /// Gets the scheduler for the CPU core with the specified index
+    const std::shared_ptr<Kernel::Scheduler>& Scheduler(size_t core_index);
+
+    /// Gets the current process
     Kernel::SharedPtr<Kernel::Process>& CurrentProcess() {
         return current_process;
     }
@@ -155,6 +168,9 @@ public:
     }
 
 private:
+    /// Returns the currently running CPU core
+    Cpu& CurrentCpuCore();
+
     /**
      * Initialize the emulated system.
      * @param emu_window Pointer to the host-system window used for video output and keyboard input.
@@ -163,22 +179,15 @@ private:
      */
     ResultStatus Init(EmuWindow* emu_window, u32 system_mode);
 
-    /// Reschedule the core emulation
-    void Reschedule();
-
     /// AppLoader used to load the current executing application
     std::unique_ptr<Loader::AppLoader> app_loader;
-
-    std::shared_ptr<ARM_Interface> cpu_core;
-    std::unique_ptr<Kernel::Scheduler> scheduler;
     std::unique_ptr<Tegra::GPU> gpu_core;
-
     std::shared_ptr<Tegra::DebugContext> debug_context;
-
     Kernel::SharedPtr<Kernel::Process> current_process;
-
-    /// When true, signals that a reschedule should happen
-    bool reschedule_pending{};
+    std::shared_ptr<CpuBarrier> cpu_barrier;
+    std::array<std::shared_ptr<Cpu>, NUM_CPU_CORES> cpu_cores;
+    std::array<std::unique_ptr<std::thread>, NUM_CPU_CORES - 1> cpu_core_threads;
+    size_t active_core{}; ///< Active core, only used in single thread mode
 
     /// Service manager
     std::shared_ptr<Service::SM::ServiceManager> service_manager;
@@ -190,10 +199,13 @@ private:
 
     ResultStatus status = ResultStatus::Success;
     std::string status_details = "";
+
+    /// Map of guest threads to CPU cores
+    std::map<std::thread::id, std::shared_ptr<Cpu>> thread_to_cpu;
 };
 
-inline ARM_Interface& CPU() {
-    return System::GetInstance().CPU();
+inline ARM_Interface& CurrentArmInterface() {
+    return System::GetInstance().CurrentArmInterface();
 }
 
 inline TelemetrySession& Telemetry() {

src/core/core_cpu.cpp

@@ -0,0 +1,119 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <condition_variable>
+#include <mutex>
+
+#include "common/logging/log.h"
+#ifdef ARCHITECTURE_x86_64
+#include "core/arm/dynarmic/arm_dynarmic.h"
+#endif
+#include "core/arm/unicorn/arm_unicorn.h"
+#include "core/core_cpu.h"
+#include "core/core_timing.h"
+#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/scheduler.h"
+#include "core/hle/kernel/thread.h"
+#include "core/settings.h"
+
+namespace Core {
+
+void CpuBarrier::NotifyEnd() {
+    std::unique_lock<std::mutex> lock(mutex);
+    end = true;
+    condition.notify_all();
+}
+
+bool CpuBarrier::Rendezvous() {
+    if (!Settings::values.use_multi_core) {
+        // Meaningless when running in single-core mode
+        return true;
+    }
+
+    if (!end) {
+        std::unique_lock<std::mutex> lock(mutex);
+
+        --cores_waiting;
+        if (!cores_waiting) {
+            cores_waiting = NUM_CPU_CORES;
+            condition.notify_all();
+            return true;
+        }
+
+        condition.wait(lock);
+        return true;
+    }
+
+    return false;
+}
+
+Cpu::Cpu(std::shared_ptr<CpuBarrier> cpu_barrier, size_t core_index)
+    : cpu_barrier{std::move(cpu_barrier)}, core_index{core_index} {
+
+    if (Settings::values.use_cpu_jit) {
+#ifdef ARCHITECTURE_x86_64
+        arm_interface = std::make_shared<ARM_Dynarmic>();
+#else
+        cpu_core = std::make_shared<ARM_Unicorn>();
+        NGLOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
+#endif
+    } else {
+        arm_interface = std::make_shared<ARM_Unicorn>();
+    }
+
+    scheduler = std::make_shared<Kernel::Scheduler>(arm_interface.get());
+}
+
+void Cpu::RunLoop(bool tight_loop) {
+    // Wait for all other CPU cores to complete the previous slice, such that they run in lock-step
+    if (!cpu_barrier->Rendezvous()) {
+        // If rendezvous failed, session has been killed
+        return;
+    }
+
+    // If we don't have a currently active thread then don't execute instructions,
+    // instead advance to the next event and try to yield to the next thread
+    if (Kernel::GetCurrentThread() == nullptr) {
+        NGLOG_TRACE(Core, "Core-{} idling", core_index);
+
+        if (IsMainCore()) {
+            CoreTiming::Idle();
+            CoreTiming::Advance();
+        }
+
+        PrepareReschedule();
+    } else {
+        if (IsMainCore()) {
+            CoreTiming::Advance();
+        }
+
+        if (tight_loop) {
+            arm_interface->Run();
+        } else {
+            arm_interface->Step();
+        }
+    }
+
+    Reschedule();
+}
+
+void Cpu::SingleStep() {
+    return RunLoop(false);
+}
+
+void Cpu::PrepareReschedule() {
+    arm_interface->PrepareReschedule();
+    reschedule_pending = true;
+}
+
+void Cpu::Reschedule() {
+    if (!reschedule_pending) {
+        return;
+    }
+
+    reschedule_pending = false;
+    scheduler->Reschedule();
+}
+
+} // namespace Core

src/core/core_cpu.h

@@ -0,0 +1,78 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <atomic>
+#include <condition_variable>
+#include <memory>
+#include <mutex>
+#include <string>
+#include "common/common_types.h"
+
+class ARM_Interface;
+
+namespace Kernel {
+class Scheduler;
+}
+
+namespace Core {
+
+constexpr unsigned NUM_CPU_CORES{4};
+
+class CpuBarrier {
+public:
+    bool IsAlive() const {
+        return !end;
+    }
+
+    void NotifyEnd();
+
+    bool Rendezvous();
+
+private:
+    unsigned cores_waiting{NUM_CPU_CORES};
+    std::mutex mutex;
+    std::condition_variable condition;
+    std::atomic<bool> end{};
+};
+
+class Cpu {
+public:
+    Cpu(std::shared_ptr<CpuBarrier> cpu_barrier, size_t core_index);
+
+    void RunLoop(bool tight_loop = true);
+
+    void SingleStep();
+
+    void PrepareReschedule();
+
+    ARM_Interface& ArmInterface() {
+        return *arm_interface;
+    }
+
+    const ARM_Interface& ArmInterface() const {
+        return *arm_interface;
+    }
+
+    const std::shared_ptr<Kernel::Scheduler>& Scheduler() const {
+        return scheduler;
+    }
+
+    bool IsMainCore() const {
+        return core_index == 0;
+    }
+
+private:
+    void Reschedule();
+
+    std::shared_ptr<ARM_Interface> arm_interface;
+    std::shared_ptr<CpuBarrier> cpu_barrier;
+    std::shared_ptr<Kernel::Scheduler> scheduler;
+
+    bool reschedule_pending{};
+    size_t core_index;
+};
+
+} // namespace Core

src/core/gdbstub/gdbstub.cpp

@@ -598,11 +598,11 @@ static void ReadRegister() {
     }
 
     if (id <= SP_REGISTER) {
-        LongToGdbHex(reply, Core::CPU().GetReg(static_cast<int>(id)));
+        LongToGdbHex(reply, Core::CurrentArmInterface().GetReg(static_cast<int>(id)));
     } else if (id == PC_REGISTER) {
-        LongToGdbHex(reply, Core::CPU().GetPC());
+        LongToGdbHex(reply, Core::CurrentArmInterface().GetPC());
     } else if (id == CPSR_REGISTER) {
-        IntToGdbHex(reply, Core::CPU().GetCPSR());
+        IntToGdbHex(reply, Core::CurrentArmInterface().GetCPSR());
     } else {
         return SendReply("E01");
     }
@@ -618,16 +618,16 @@ static void ReadRegisters() {
     u8* bufptr = buffer;
 
     for (int reg = 0; reg <= SP_REGISTER; reg++) {
-        LongToGdbHex(bufptr + reg * 16, Core::CPU().GetReg(reg));
+        LongToGdbHex(bufptr + reg * 16, Core::CurrentArmInterface().GetReg(reg));
     }
 
     bufptr += (32 * 16);
 
-    LongToGdbHex(bufptr, Core::CPU().GetPC());
+    LongToGdbHex(bufptr, Core::CurrentArmInterface().GetPC());
 
     bufptr += 16;
 
-    IntToGdbHex(bufptr, Core::CPU().GetCPSR());
+    IntToGdbHex(bufptr, Core::CurrentArmInterface().GetCPSR());
 
     bufptr += 8;
 
@@ -646,11 +646,11 @@ static void WriteRegister() {
     }
 
     if (id <= SP_REGISTER) {
-        Core::CPU().SetReg(id, GdbHexToLong(buffer_ptr));
+        Core::CurrentArmInterface().SetReg(id, GdbHexToLong(buffer_ptr));
     } else if (id == PC_REGISTER) {
-        Core::CPU().SetPC(GdbHexToLong(buffer_ptr));
+        Core::CurrentArmInterface().SetPC(GdbHexToLong(buffer_ptr));
     } else if (id == CPSR_REGISTER) {
-        Core::CPU().SetCPSR(GdbHexToInt(buffer_ptr));
+        Core::CurrentArmInterface().SetCPSR(GdbHexToInt(buffer_ptr));
     } else {
         return SendReply("E01");
     }
@@ -667,11 +667,11 @@ static void WriteRegisters() {
 
     for (int i = 0, reg = 0; reg <= CPSR_REGISTER; i++, reg++) {
         if (reg <= SP_REGISTER) {
-            Core::CPU().SetReg(reg, GdbHexToLong(buffer_ptr + i * 16));
+            Core::CurrentArmInterface().SetReg(reg, GdbHexToLong(buffer_ptr + i * 16));
         } else if (reg == PC_REGISTER) {
-            Core::CPU().SetPC(GdbHexToLong(buffer_ptr + i * 16));
+            Core::CurrentArmInterface().SetPC(GdbHexToLong(buffer_ptr + i * 16));
         } else if (reg == CPSR_REGISTER) {
-            Core::CPU().SetCPSR(GdbHexToInt(buffer_ptr + i * 16));
+            Core::CurrentArmInterface().SetCPSR(GdbHexToInt(buffer_ptr + i * 16));
         } else {
             UNIMPLEMENTED();
         }

src/core/hle/ipc.h

@@ -29,9 +29,14 @@ enum class ControlCommand : u32 {
 };
 
 enum class CommandType : u32 {
+    Invalid = 0,
+    LegacyRequest = 1,
     Close = 2,
+    LegacyControl = 3,
     Request = 4,
     Control = 5,
+    RequestWithContext = 6,
+    ControlWithContext = 7,
     Unspecified,
 };
 

src/core/hle/kernel/errors.h

@@ -21,7 +21,9 @@ enum {
 
     // Confirmed Switch OS error codes
     MisalignedAddress = 102,
+    InvalidProcessorId = 113,
     InvalidHandle = 114,
+    InvalidCombination = 116,
     Timeout = 117,
     SynchronizationCanceled = 118,
     TooLarge = 119,

src/core/hle/kernel/hle_ipc.cpp

@@ -110,7 +110,9 @@ void HLERequestContext::ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming) {
     // Padding to align to 16 bytes
     rp.AlignWithPadding();
 
-    if (Session()->IsDomain() && (command_header->type == IPC::CommandType::Request || !incoming)) {
+    if (Session()->IsDomain() && ((command_header->type == IPC::CommandType::Request ||
+                                   command_header->type == IPC::CommandType::RequestWithContext) ||
+                                  !incoming)) {
         // If this is an incoming message, only CommandType "Request" has a domain header
         // All outgoing domain messages have the domain header, if only incoming has it
         if (incoming || domain_message_header) {

src/core/hle/kernel/mutex.cpp

@@ -104,7 +104,6 @@ ResultCode Mutex::Release(VAddr address) {
 
     // There are no more threads waiting for the mutex, release it completely.
     if (thread == nullptr) {
-        ASSERT(GetCurrentThread()->wait_mutex_threads.empty());
         Memory::Write32(address, 0);
         return RESULT_SUCCESS;
     }

src/core/hle/kernel/scheduler.cpp

@@ -9,6 +9,8 @@
 
 namespace Kernel {
 
+std::mutex Scheduler::scheduler_mutex;
+
 Scheduler::Scheduler(ARM_Interface* cpu_core) : cpu_core(cpu_core) {}
 
 Scheduler::~Scheduler() {
@@ -18,6 +20,7 @@ Scheduler::~Scheduler() {
 }
 
 bool Scheduler::HaveReadyThreads() {
+    std::lock_guard<std::mutex> lock(scheduler_mutex);
     return ready_queue.get_first() != nullptr;
 }
 
@@ -90,6 +93,8 @@ void Scheduler::SwitchContext(Thread* new_thread) {
 }
 
 void Scheduler::Reschedule() {
+    std::lock_guard<std::mutex> lock(scheduler_mutex);
+
     Thread* cur = GetCurrentThread();
     Thread* next = PopNextReadyThread();
 
@@ -105,26 +110,36 @@ void Scheduler::Reschedule() {
 }
 
 void Scheduler::AddThread(SharedPtr<Thread> thread, u32 priority) {
+    std::lock_guard<std::mutex> lock(scheduler_mutex);
+
     thread_list.push_back(thread);
     ready_queue.prepare(priority);
 }
 
 void Scheduler::RemoveThread(Thread* thread) {
+    std::lock_guard<std::mutex> lock(scheduler_mutex);
+
     thread_list.erase(std::remove(thread_list.begin(), thread_list.end(), thread),
                       thread_list.end());
 }
 
 void Scheduler::ScheduleThread(Thread* thread, u32 priority) {
+    std::lock_guard<std::mutex> lock(scheduler_mutex);
+
     ASSERT(thread->status == THREADSTATUS_READY);
     ready_queue.push_back(priority, thread);
 }
 
 void Scheduler::UnscheduleThread(Thread* thread, u32 priority) {
+    std::lock_guard<std::mutex> lock(scheduler_mutex);
+
     ASSERT(thread->status == THREADSTATUS_READY);
     ready_queue.remove(priority, thread);
 }
 
 void Scheduler::SetThreadPriority(Thread* thread, u32 priority) {
+    std::lock_guard<std::mutex> lock(scheduler_mutex);
+
     // If thread was ready, adjust queues
     if (thread->status == THREADSTATUS_READY)
         ready_queue.move(thread, thread->current_priority, priority);

src/core/hle/kernel/scheduler.h

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include <mutex>
 #include <vector>
 #include "common/common_types.h"
 #include "common/thread_queue_list.h"
@@ -68,6 +69,8 @@ private:
     SharedPtr<Thread> current_thread = nullptr;
 
     ARM_Interface* cpu_core;
+
+    static std::mutex scheduler_mutex;
 };
 
 } // namespace Kernel

src/core/hle/kernel/svc.cpp

@@ -145,36 +145,6 @@ static bool DefaultThreadWakeupCallback(ThreadWakeupReason reason, SharedPtr<Thr
     return true;
 };
 
-/// Wait for a kernel object to synchronize, timeout after the specified nanoseconds
-static ResultCode WaitSynchronization1(
-    SharedPtr<WaitObject> object, Thread* thread, s64 nano_seconds = -1,
-    std::function<Thread::WakeupCallback> wakeup_callback = DefaultThreadWakeupCallback) {
-
-    if (!object) {
-        return ERR_INVALID_HANDLE;
-    }
-
-    if (object->ShouldWait(thread)) {
-        if (nano_seconds == 0) {
-            return RESULT_TIMEOUT;
-        }
-
-        thread->wait_objects = {object};
-        object->AddWaitingThread(thread);
-        thread->status = THREADSTATUS_WAIT_SYNCH_ANY;
-
-        // Create an event to wake the thread up after the specified nanosecond delay has passed
-        thread->WakeAfterDelay(nano_seconds);
-        thread->wakeup_callback = wakeup_callback;
-
-        Core::System::GetInstance().PrepareReschedule();
-    } else {
-        object->Acquire(thread);
-    }
-
-    return RESULT_SUCCESS;
-}
-
 /// Wait for the given handles to synchronize, timeout after the specified nanoseconds
 static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64 handle_count,
                                       s64 nano_seconds) {
@@ -232,7 +202,7 @@ static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64
     thread->WakeAfterDelay(nano_seconds);
     thread->wakeup_callback = DefaultThreadWakeupCallback;
 
-    Core::System::GetInstance().PrepareReschedule();
+    Core::System::GetInstance().CpuCore(thread->processor_id).PrepareReschedule();
 
     return RESULT_TIMEOUT;
 }
@@ -395,14 +365,14 @@ static ResultCode SetThreadPriority(Handle handle, u32 priority) {
 
     thread->SetPriority(priority);
 
-    Core::System::GetInstance().PrepareReschedule();
+    Core::System::GetInstance().CpuCore(thread->processor_id).PrepareReschedule();
     return RESULT_SUCCESS;
 }
 
 /// Get which CPU core is executing the current thread
 static u32 GetCurrentProcessorNumber() {
-    NGLOG_WARNING(Kernel_SVC, "(STUBBED) called, defaulting to processor 0");
-    return 0;
+    NGLOG_TRACE(Kernel_SVC, "called");
+    return GetCurrentThread()->processor_id;
 }
 
 static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 size,
@@ -485,22 +455,28 @@ static void ExitProcess() {
 
     Core::CurrentProcess()->status = ProcessStatus::Exited;
 
-    // Stop all the process threads that are currently waiting for objects.
-    auto& thread_list = Core::System::GetInstance().Scheduler().GetThreadList();
-    for (auto& thread : thread_list) {
-        if (thread->owner_process != Core::CurrentProcess())
-            continue;
+    auto stop_threads = [](const std::vector<SharedPtr<Thread>>& thread_list) {
+        for (auto& thread : thread_list) {
+            if (thread->owner_process != Core::CurrentProcess())
+                continue;
 
-        if (thread == GetCurrentThread())
-            continue;
+            if (thread == GetCurrentThread())
+                continue;
 
-        // TODO(Subv): When are the other running/ready threads terminated?
-        ASSERT_MSG(thread->status == THREADSTATUS_WAIT_SYNCH_ANY ||
-                       thread->status == THREADSTATUS_WAIT_SYNCH_ALL,
-                   "Exiting processes with non-waiting threads is currently unimplemented");
+            // TODO(Subv): When are the other running/ready threads terminated?
+            ASSERT_MSG(thread->status == THREADSTATUS_WAIT_SYNCH_ANY ||
+                           thread->status == THREADSTATUS_WAIT_SYNCH_ALL,
+                       "Exiting processes with non-waiting threads is currently unimplemented");
 
-        thread->Stop();
-    }
+            thread->Stop();
+        }
+    };
+
+    auto& system = Core::System::GetInstance();
+    stop_threads(system.Scheduler(0)->GetThreadList());
+    stop_threads(system.Scheduler(1)->GetThreadList());
+    stop_threads(system.Scheduler(2)->GetThreadList());
+    stop_threads(system.Scheduler(3)->GetThreadList());
 
     // Kill the current thread
     GetCurrentThread()->Stop();
@@ -530,14 +506,9 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
 
     switch (processor_id) {
     case THREADPROCESSORID_0:
-        break;
     case THREADPROCESSORID_1:
     case THREADPROCESSORID_2:
     case THREADPROCESSORID_3:
-        // TODO(bunnei): Implement support for other processor IDs
-        NGLOG_ERROR(Kernel_SVC,
-                    "Newly created thread must run in another thread ({}), unimplemented.",
-                    processor_id);
         break;
     default:
         ASSERT_MSG(false, "Unsupported thread processor ID: {}", processor_id);
@@ -551,6 +522,7 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
     *out_handle = thread->guest_handle;
 
     Core::System::GetInstance().PrepareReschedule();
+    Core::System::GetInstance().CpuCore(thread->processor_id).PrepareReschedule();
 
     NGLOG_TRACE(Kernel_SVC,
                 "called entrypoint=0x{:08X} ({}), arg=0x{:08X}, stacktop=0x{:08X}, "
@@ -569,14 +541,17 @@ static ResultCode StartThread(Handle thread_handle) {
         return ERR_INVALID_HANDLE;
     }
 
+    ASSERT(thread->status == THREADSTATUS_DORMANT);
+
     thread->ResumeFromWait();
+    Core::System::GetInstance().CpuCore(thread->processor_id).PrepareReschedule();
 
     return RESULT_SUCCESS;
 }
 
 /// Called when a thread exits
 static void ExitThread() {
-    NGLOG_TRACE(Kernel_SVC, "called, pc=0x{:08X}", Core::CPU().GetPC());
+    NGLOG_TRACE(Kernel_SVC, "called, pc=0x{:08X}", Core::CurrentArmInterface().GetPC());
 
     ExitCurrentThread();
     Core::System::GetInstance().PrepareReschedule();
@@ -588,7 +563,7 @@ static void SleepThread(s64 nanoseconds) {
 
     // Don't attempt to yield execution if there are no available threads to run,
     // this way we avoid a useless reschedule to the idle thread.
-    if (nanoseconds == 0 && !Core::System::GetInstance().Scheduler().HaveReadyThreads())
+    if (nanoseconds == 0 && !Core::System::GetInstance().CurrentScheduler().HaveReadyThreads())
         return;
 
     // Sleep current thread and check for next thread to schedule
@@ -624,7 +599,7 @@ static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_var
 
     // Note: Deliberately don't attempt to inherit the lock owner's priority.
 
-    Core::System::GetInstance().PrepareReschedule();
+    Core::System::GetInstance().CpuCore(current_thread->processor_id).PrepareReschedule();
     return RESULT_SUCCESS;
 }
 
@@ -633,17 +608,43 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target
     NGLOG_TRACE(Kernel_SVC, "called, condition_variable_addr=0x{:X}, target=0x{:08X}",
                 condition_variable_addr, target);
 
-    u32 processed = 0;
-    auto& thread_list = Core::System::GetInstance().Scheduler().GetThreadList();
+    auto RetrieveWaitingThreads =
+        [](size_t core_index, std::vector<SharedPtr<Thread>>& waiting_threads, VAddr condvar_addr) {
+            const auto& scheduler = Core::System::GetInstance().Scheduler(core_index);
+            auto& thread_list = scheduler->GetThreadList();
 
-    for (auto& thread : thread_list) {
-        if (thread->condvar_wait_address != condition_variable_addr)
-            continue;
+            for (auto& thread : thread_list) {
+                if (thread->condvar_wait_address == condvar_addr)
+                    waiting_threads.push_back(thread);
+            }
+        };
 
-        // Only process up to 'target' threads, unless 'target' is -1, in which case process
-        // them all.
-        if (target != -1 && processed >= target)
-            break;
+    // Retrieve a list of all threads that are waiting for this condition variable.
+    std::vector<SharedPtr<Thread>> waiting_threads;
+    RetrieveWaitingThreads(0, waiting_threads, condition_variable_addr);
+    RetrieveWaitingThreads(1, waiting_threads, condition_variable_addr);
+    RetrieveWaitingThreads(2, waiting_threads, condition_variable_addr);
+    RetrieveWaitingThreads(3, waiting_threads, condition_variable_addr);
+    // Sort them by priority, such that the highest priority ones come first.
+    std::sort(waiting_threads.begin(), waiting_threads.end(),
+              [](const SharedPtr<Thread>& lhs, const SharedPtr<Thread>& rhs) {
+                  return lhs->current_priority < rhs->current_priority;
+              });
+
+    // Only process up to 'target' threads, unless 'target' is -1, in which case process
+    // them all.
+    size_t last = waiting_threads.size();
+    if (target != -1)
+        last = target;
+
+    // If there are no threads waiting on this condition variable, just exit
+    if (last > waiting_threads.size())
+        return RESULT_SUCCESS;
+
+    for (size_t index = 0; index < last; ++index) {
+        auto& thread = waiting_threads[index];
+
+        ASSERT(thread->condvar_wait_address == condition_variable_addr);
 
         // If the mutex is not yet acquired, acquire it.
         u32 mutex_val = Memory::Read32(thread->mutex_wait_address);
@@ -676,10 +677,8 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target
 
             owner->AddMutexWaiter(thread);
 
-            Core::System::GetInstance().PrepareReschedule();
+            Core::System::GetInstance().CpuCore(thread->processor_id).PrepareReschedule();
         }
-
-        ++processed;
     }
 
     return RESULT_SUCCESS;
@@ -718,16 +717,56 @@ static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32
     return RESULT_SUCCESS;
 }
 
-static ResultCode GetThreadCoreMask(Handle handle, u32* mask, u64* unknown) {
-    NGLOG_WARNING(Kernel_SVC, "(STUBBED) called, handle=0x{:08X}", handle);
-    *mask = 0x0;
-    *unknown = 0xf;
+static ResultCode GetThreadCoreMask(Handle thread_handle, u32* core, u64* mask) {
+    NGLOG_TRACE(Kernel_SVC, "called, handle=0x{:08X}", thread_handle);
+
+    const SharedPtr<Thread> thread = g_handle_table.Get<Thread>(thread_handle);
+    if (!thread) {
+        return ERR_INVALID_HANDLE;
+    }
+
+    *core = thread->ideal_core;
+    *mask = thread->affinity_mask;
+
     return RESULT_SUCCESS;
 }
 
-static ResultCode SetThreadCoreMask(Handle handle, u32 mask, u64 unknown) {
-    NGLOG_WARNING(Kernel_SVC, "(STUBBED) called, handle=0x{:08X}, mask=0x{:08X}, unknown=0x{:X}",
-                  handle, mask, unknown);
+static ResultCode SetThreadCoreMask(Handle thread_handle, u32 core, u64 mask) {
+    NGLOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, mask=0x{:16X}, core=0x{:X}", thread_handle,
+                mask, core);
+
+    const SharedPtr<Thread> thread = g_handle_table.Get<Thread>(thread_handle);
+    if (!thread) {
+        return ERR_INVALID_HANDLE;
+    }
+
+    if (core == THREADPROCESSORID_DEFAULT) {
+        ASSERT(thread->owner_process->ideal_processor != THREADPROCESSORID_DEFAULT);
+        // Set the target CPU to the one specified in the process' exheader.
+        core = thread->owner_process->ideal_processor;
+        mask = 1 << core;
+    }
+
+    if (mask == 0) {
+        return ResultCode(ErrorModule::Kernel, ErrCodes::InvalidCombination);
+    }
+
+    /// This value is used to only change the affinity mask without changing the current ideal core.
+    static constexpr u32 OnlyChangeMask = static_cast<u32>(-3);
+
+    if (core == OnlyChangeMask) {
+        core = thread->ideal_core;
+    } else if (core >= Core::NUM_CPU_CORES && core != -1) {
+        return ResultCode(ErrorModule::Kernel, ErrCodes::InvalidProcessorId);
+    }
+
+    // Error out if the input core isn't enabled in the input mask.
+    if (core < Core::NUM_CPU_CORES && (mask & (1 << core)) == 0) {
+        return ResultCode(ErrorModule::Kernel, ErrCodes::InvalidCombination);
+    }
+
+    thread->ChangeCore(core, mask);
+
     return RESULT_SUCCESS;
 }
 

src/core/hle/kernel/svc.h

@@ -47,9 +47,12 @@ enum class GetInfoType : u64 {
     NewMapRegionSize = 15,
     // 3.0.0+
     IsVirtualAddressMemoryEnabled = 16,
+    PersonalMmHeapUsage = 17,
     TitleId = 18,
     // 4.0.0+
     PrivilegedProcessId = 19,
+    // 5.0.0+
+    UserExceptionContextAddr = 20,
 };
 
 void CallSVC(u32 immediate);

src/core/hle/kernel/svc_wrap.h

@@ -13,14 +13,14 @@
 
 namespace Kernel {
 
-#define PARAM(n) Core::CPU().GetReg(n)
+#define PARAM(n) Core::CurrentArmInterface().GetReg(n)
 
 /**
  * HLE a function return from the current ARM userland process
  * @param res Result to return
  */
 static inline void FuncReturn(u64 res) {
-    Core::CPU().SetReg(0, res);
+    Core::CurrentArmInterface().SetReg(0, res);
 }
 
 ////////////////////////////////////////////////////////////////////////////////////////////////////
@@ -45,7 +45,7 @@ template <ResultCode func(u32*, u32)>
 void SvcWrap() {
     u32 param_1 = 0;
     u32 retval = func(&param_1, (u32)PARAM(1)).raw;
-    Core::CPU().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
 
@@ -53,7 +53,7 @@ template <ResultCode func(u32*, u64)>
 void SvcWrap() {
     u32 param_1 = 0;
     u32 retval = func(&param_1, PARAM(1)).raw;
-    Core::CPU().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
 
@@ -66,7 +66,7 @@ template <ResultCode func(u64*, u64)>
 void SvcWrap() {
     u64 param_1 = 0;
     u32 retval = func(&param_1, PARAM(1)).raw;
-    Core::CPU().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
 
@@ -85,8 +85,8 @@ void SvcWrap() {
     u32 param_1 = 0;
     u64 param_2 = 0;
     ResultCode retval = func((u32)(PARAM(2) & 0xFFFFFFFF), &param_1, &param_2);
-    Core::CPU().SetReg(1, param_1);
-    Core::CPU().SetReg(2, param_2);
+    Core::CurrentArmInterface().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(2, param_2);
     FuncReturn(retval.raw);
 }
 
@@ -120,7 +120,7 @@ template <ResultCode func(u32*, u64, u64, s64)>
 void SvcWrap() {
     u32 param_1 = 0;
     ResultCode retval = func(&param_1, PARAM(1), (u32)(PARAM(2) & 0xFFFFFFFF), (s64)PARAM(3));
-    Core::CPU().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval.raw);
 }
 
@@ -133,7 +133,7 @@ template <ResultCode func(u64*, u64, u64, u64)>
 void SvcWrap() {
     u64 param_1 = 0;
     u32 retval = func(&param_1, PARAM(1), PARAM(2), PARAM(3)).raw;
-    Core::CPU().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
 
@@ -143,7 +143,7 @@ void SvcWrap() {
     u32 retval =
         func(&param_1, PARAM(1), PARAM(2), PARAM(3), (u32)PARAM(4), (s32)(PARAM(5) & 0xFFFFFFFF))
             .raw;
-    Core::CPU().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
 
@@ -166,7 +166,7 @@ template <ResultCode func(u32*, u64, u64, u32)>
 void SvcWrap() {
     u32 param_1 = 0;
     u32 retval = func(&param_1, PARAM(1), PARAM(2), (u32)(PARAM(3) & 0xFFFFFFFF)).raw;
-    Core::CPU().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
 
@@ -175,7 +175,7 @@ void SvcWrap() {
     u32 param_1 = 0;
     u32 retval =
         func(&param_1, PARAM(1), (u32)(PARAM(2) & 0xFFFFFFFF), (u32)(PARAM(3) & 0xFFFFFFFF)).raw;
-    Core::CPU().SetReg(1, param_1);
+    Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
 

src/core/hle/kernel/thread.cpp

@@ -64,7 +64,7 @@ void Thread::Stop() {
     // Clean up thread from ready queue
     // This is only needed when the thread is termintated forcefully (SVC TerminateProcess)
     if (status == THREADSTATUS_READY) {
-        Core::System::GetInstance().Scheduler().UnscheduleThread(this, current_priority);
+        scheduler->UnscheduleThread(this, current_priority);
     }
 
     status = THREADSTATUS_DEAD;
@@ -92,7 +92,7 @@ void WaitCurrentThread_Sleep() {
 void ExitCurrentThread() {
     Thread* thread = GetCurrentThread();
     thread->Stop();
-    Core::System::GetInstance().Scheduler().RemoveThread(thread);
+    Core::System::GetInstance().CurrentScheduler().RemoveThread(thread);
 }
 
 /**
@@ -133,8 +133,11 @@ static void ThreadWakeupCallback(u64 thread_handle, int cycles_late) {
 
         auto lock_owner = thread->lock_owner;
         // Threads waking up by timeout from WaitProcessWideKey do not perform priority inheritance
-        // and don't have a lock owner.
-        ASSERT(lock_owner == nullptr);
+        // and don't have a lock owner unless SignalProcessWideKey was called first and the thread
+        // wasn't awakened due to the mutex already being acquired.
+        if (lock_owner) {
+            lock_owner->RemoveMutexWaiter(thread);
+        }
     }
 
     if (resume)
@@ -154,6 +157,18 @@ void Thread::CancelWakeupTimer() {
     CoreTiming::UnscheduleEvent(ThreadWakeupEventType, callback_handle);
 }
 
+static boost::optional<s32> GetNextProcessorId(u64 mask) {
+    for (s32 index = 0; index < Core::NUM_CPU_CORES; ++index) {
+        if (mask & (1ULL << index)) {
+            if (!Core::System().GetInstance().Scheduler(index)->GetCurrentThread()) {
+                // Core is enabled and not running any threads, use this one
+                return index;
+            }
+        }
+    }
+    return {};
+}
+
 void Thread::ResumeFromWait() {
     ASSERT_MSG(wait_objects.empty(), "Thread is waking up while waiting for objects");
 
@@ -188,8 +203,37 @@ void Thread::ResumeFromWait() {
     wakeup_callback = nullptr;
 
     status = THREADSTATUS_READY;
-    Core::System::GetInstance().Scheduler().ScheduleThread(this, current_priority);
-    Core::System::GetInstance().PrepareReschedule();
+
+    boost::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask);
+    if (!new_processor_id) {
+        new_processor_id = processor_id;
+    }
+    if (ideal_core != -1 &&
+        Core::System().GetInstance().Scheduler(ideal_core)->GetCurrentThread() == nullptr) {
+        new_processor_id = ideal_core;
+    }
+
+    ASSERT(*new_processor_id < 4);
+
+    // Add thread to new core's scheduler
+    auto& next_scheduler = Core::System().GetInstance().Scheduler(*new_processor_id);
+
+    if (*new_processor_id != processor_id) {
+        // Remove thread from previous core's scheduler
+        scheduler->RemoveThread(this);
+        next_scheduler->AddThread(this, current_priority);
+    }
+
+    processor_id = *new_processor_id;
+
+    // If the thread was ready, unschedule from the previous core and schedule on the new core
+    scheduler->UnscheduleThread(this, current_priority);
+    next_scheduler->ScheduleThread(this, current_priority);
+
+    // Change thread's scheduler
+    scheduler = next_scheduler;
+
+    Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule();
 }
 
 /**
@@ -259,8 +303,6 @@ ResultVal<SharedPtr<Thread>> Thread::Create(std::string name, VAddr entry_point,
 
     SharedPtr<Thread> thread(new Thread);
 
-    Core::System::GetInstance().Scheduler().AddThread(thread, priority);
-
     thread->thread_id = NewThreadId();
     thread->status = THREADSTATUS_DORMANT;
     thread->entry_point = entry_point;
@@ -268,6 +310,8 @@ ResultVal<SharedPtr<Thread>> Thread::Create(std::string name, VAddr entry_point,
     thread->nominal_priority = thread->current_priority = priority;
     thread->last_running_ticks = CoreTiming::GetTicks();
     thread->processor_id = processor_id;
+    thread->ideal_core = processor_id;
+    thread->affinity_mask = 1ULL << processor_id;
     thread->wait_objects.clear();
     thread->mutex_wait_address = 0;
     thread->condvar_wait_address = 0;
@@ -275,6 +319,8 @@ ResultVal<SharedPtr<Thread>> Thread::Create(std::string name, VAddr entry_point,
     thread->name = std::move(name);
     thread->callback_handle = wakeup_callback_handle_table.Create(thread).Unwrap();
     thread->owner_process = owner_process;
+    thread->scheduler = Core::System().GetInstance().Scheduler(processor_id);
+    thread->scheduler->AddThread(thread, priority);
 
     // Find the next available TLS index, and mark it as used
     auto& tls_slots = owner_process->tls_slots;
@@ -337,7 +383,7 @@ void Thread::SetPriority(u32 priority) {
 }
 
 void Thread::BoostPriority(u32 priority) {
-    Core::System::GetInstance().Scheduler().SetThreadPriority(this, priority);
+    scheduler->SetThreadPriority(this, priority);
     current_priority = priority;
 }
 
@@ -406,7 +452,7 @@ void Thread::UpdatePriority() {
     if (new_priority == current_priority)
         return;
 
-    Core::System::GetInstance().Scheduler().SetThreadPriority(this, new_priority);
+    scheduler->SetThreadPriority(this, new_priority);
 
     current_priority = new_priority;
 
@@ -415,13 +461,54 @@ void Thread::UpdatePriority() {
         lock_owner->UpdatePriority();
 }
 
+void Thread::ChangeCore(u32 core, u64 mask) {
+    ideal_core = core;
+    affinity_mask = mask;
+
+    if (status != THREADSTATUS_READY) {
+        return;
+    }
+
+    boost::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
+
+    if (!new_processor_id) {
+        new_processor_id = processor_id;
+    }
+    if (ideal_core != -1 &&
+        Core::System().GetInstance().Scheduler(ideal_core)->GetCurrentThread() == nullptr) {
+        new_processor_id = ideal_core;
+    }
+
+    ASSERT(*new_processor_id < 4);
+
+    // Add thread to new core's scheduler
+    auto& next_scheduler = Core::System().GetInstance().Scheduler(*new_processor_id);
+
+    if (*new_processor_id != processor_id) {
+        // Remove thread from previous core's scheduler
+        scheduler->RemoveThread(this);
+        next_scheduler->AddThread(this, current_priority);
+    }
+
+    processor_id = *new_processor_id;
+
+    // If the thread was ready, unschedule from the previous core and schedule on the new core
+    scheduler->UnscheduleThread(this, current_priority);
+    next_scheduler->ScheduleThread(this, current_priority);
+
+    // Change thread's scheduler
+    scheduler = next_scheduler;
+
+    Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule();
+}
+
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 
 /**
  * Gets the current thread
  */
 Thread* GetCurrentThread() {
-    return Core::System::GetInstance().Scheduler().GetCurrentThread();
+    return Core::System::GetInstance().CurrentScheduler().GetCurrentThread();
 }
 
 void ThreadingInit() {

src/core/hle/kernel/thread.h

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include <memory>
 #include <string>
 #include <unordered_map>
 #include <vector>
@@ -56,6 +57,7 @@ enum class ThreadWakeupReason {
 namespace Kernel {
 
 class Process;
+class Scheduler;
 
 class Thread final : public WaitObject {
 public:
@@ -118,6 +120,9 @@ public:
     /// Recalculates the current priority taking into account priority inheritance.
     void UpdatePriority();
 
+    /// Changes the core that the thread is running or scheduled to run on.
+    void ChangeCore(u32 core, u64 mask);
+
     /**
      * Gets the thread's thread ID
      * @return The thread's ID
@@ -240,6 +245,11 @@ public:
     // available. In case of a timeout, the object will be nullptr.
     std::function<WakeupCallback> wakeup_callback;
 
+    std::shared_ptr<Scheduler> scheduler;
+
+    u32 ideal_core{0xFFFFFFFF};
+    u64 affinity_mask{0x1};
+
 private:
     Thread();
     ~Thread() override;

src/core/hle/kernel/vm_manager.cpp

@@ -104,8 +104,15 @@ ResultVal<VMManager::VMAHandle> VMManager::MapMemoryBlock(VAddr target,
     VirtualMemoryArea& final_vma = vma_handle->second;
     ASSERT(final_vma.size == size);
 
-    Core::CPU().MapBackingMemory(target, size, block->data() + offset,
-                                 VMAPermission::ReadWriteExecute);
+    auto& system = Core::System::GetInstance();
+    system.ArmInterface(0).MapBackingMemory(target, size, block->data() + offset,
+                                            VMAPermission::ReadWriteExecute);
+    system.ArmInterface(1).MapBackingMemory(target, size, block->data() + offset,
+                                            VMAPermission::ReadWriteExecute);
+    system.ArmInterface(2).MapBackingMemory(target, size, block->data() + offset,
+                                            VMAPermission::ReadWriteExecute);
+    system.ArmInterface(3).MapBackingMemory(target, size, block->data() + offset,
+                                            VMAPermission::ReadWriteExecute);
 
     final_vma.type = VMAType::AllocatedMemoryBlock;
     final_vma.permissions = VMAPermission::ReadWrite;
@@ -126,7 +133,11 @@ ResultVal<VMManager::VMAHandle> VMManager::MapBackingMemory(VAddr target, u8* me
     VirtualMemoryArea& final_vma = vma_handle->second;
     ASSERT(final_vma.size == size);
 
-    Core::CPU().MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
+    auto& system = Core::System::GetInstance();
+    system.ArmInterface(0).MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
+    system.ArmInterface(1).MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
+    system.ArmInterface(2).MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
+    system.ArmInterface(3).MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
 
     final_vma.type = VMAType::BackingMemory;
     final_vma.permissions = VMAPermission::ReadWrite;
@@ -184,7 +195,11 @@ ResultCode VMManager::UnmapRange(VAddr target, u64 size) {
 
     ASSERT(FindVMA(target)->second.size >= size);
 
-    Core::CPU().UnmapMemory(target, size);
+    auto& system = Core::System::GetInstance();
+    system.ArmInterface(0).UnmapMemory(target, size);
+    system.ArmInterface(1).UnmapMemory(target, size);
+    system.ArmInterface(2).UnmapMemory(target, size);
+    system.ArmInterface(3).UnmapMemory(target, size);
 
     return RESULT_SUCCESS;
 }

src/core/hle/result.h

@@ -32,7 +32,8 @@ enum class ErrorModule : u32 {
     Common = 0,
     Kernel = 1,
     FS = 2,
-    NvidiaTransferMemory = 3,
+    OS = 3, // used for Memory, Thread, Mutex, Nvidia
+    HTCS = 4,
     NCM = 5,
     DD = 6,
     LR = 8,
@@ -42,41 +43,80 @@ enum class ErrorModule : u32 {
     PM = 15,
     NS = 16,
     HTC = 18,
+    NCMContent = 20,
     SM = 21,
     RO = 22,
     SDMMC = 24,
+    OVLN = 25,
     SPL = 26,
     ETHC = 100,
     I2C = 101,
+    GPIO = 102,
+    UART = 103,
     Settings = 105,
+    WLAN = 107,
+    XCD = 108,
     NIFM = 110,
-    Display = 114,
-    NTC = 116,
+    Hwopus = 111,
+    Bluetooth = 113,
+    VI = 114,
+    NFP = 115,
+    Time = 116,
     FGM = 117,
-    PCIE = 120,
+    OE = 118,
+    PCIe = 120,
     Friends = 121,
+    BCAT = 122,
     SSL = 123,
     Account = 124,
+    News = 125,
     Mii = 126,
+    NFC = 127,
     AM = 128,
     PlayReport = 129,
+    AHID = 130,
+    Qlaunch = 132,
     PCV = 133,
     OMM = 134,
+    BPC = 135,
+    PSM = 136,
     NIM = 137,
     PSC = 138,
+    TC = 139,
     USB = 140,
+    NSD = 141,
+    PCTL = 142,
     BTM = 143,
+    ETicket = 145,
+    NGC = 146,
     ERPT = 147,
     APM = 148,
+    Profiler = 150,
+    ErrorUpload = 151,
+    Audio = 153,
     NPNS = 154,
+    NPNSHTTPSTREAM = 155,
     ARP = 157,
-    BOOT = 158,
-    NFC = 161,
+    SWKBD = 158,
+    BOOT = 159,
+    NFCMifare = 161,
     UserlandAssert = 162,
+    Fatal = 163,
+    NIMShop = 164,
+    SPSM = 165,
+    BGTC = 167,
     UserlandCrash = 168,
-    HID = 203,
+    SREPO = 180,
+    Dauth = 181,
+    HID = 202,
+    LDN = 203,
+    Irsensor = 205,
     Capture = 206,
-    TC = 651,
+    Manu = 208,
+    ATK = 209,
+    GRC = 212,
+    Migration = 216,
+    MigrationLdcServ = 217,
     GeneralWebApplet = 800,
     WifiWebAuthApplet = 809,
     WhitelistedApplet = 810,

src/core/hle/service/am/am.cpp

@@ -74,7 +74,40 @@ void IAudioController::GetLibraryAppletExpectedMasterVolume(Kernel::HLERequestCo
     rb.Push(volume);
 }
 
-IDisplayController::IDisplayController() : ServiceFramework("IDisplayController") {}
+IDisplayController::IDisplayController() : ServiceFramework("IDisplayController") {
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetLastForegroundCaptureImage"},
+        {1, nullptr, "UpdateLastForegroundCaptureImage"},
+        {2, nullptr, "GetLastApplicationCaptureImage"},
+        {3, nullptr, "GetCallerAppletCaptureImage"},
+        {4, nullptr, "UpdateCallerAppletCaptureImage"},
+        {5, nullptr, "GetLastForegroundCaptureImageEx"},
+        {6, nullptr, "GetLastApplicationCaptureImageEx"},
+        {7, nullptr, "GetCallerAppletCaptureImageEx"},
+        {8, nullptr, "TakeScreenShotOfOwnLayer"},  // 2.0.0+
+        {9, nullptr, "CopyBetweenCaptureBuffers"}, // 5.0.0+
+        {10, nullptr, "AcquireLastApplicationCaptureBuffer"},
+        {11, nullptr, "ReleaseLastApplicationCaptureBuffer"},
+        {12, nullptr, "AcquireLastForegroundCaptureBuffer"},
+        {13, nullptr, "ReleaseLastForegroundCaptureBuffer"},
+        {14, nullptr, "AcquireCallerAppletCaptureBuffer"},
+        {15, nullptr, "ReleaseCallerAppletCaptureBuffer"},
+        {16, nullptr, "AcquireLastApplicationCaptureBufferEx"},
+        {17, nullptr, "AcquireLastForegroundCaptureBufferEx"},
+        {18, nullptr, "AcquireCallerAppletCaptureBufferEx"},
+        // 2.0.0+
+        {20, nullptr, "ClearCaptureBuffer"},
+        {21, nullptr, "ClearAppletTransitionBuffer"},
+        // 4.0.0+
+        {22, nullptr, "AcquireLastApplicationCaptureSharedBuffer"},
+        {23, nullptr, "ReleaseLastApplicationCaptureSharedBuffer"},
+        {24, nullptr, "AcquireLastForegroundCaptureSharedBuffer"},
+        {25, nullptr, "ReleaseLastForegroundCaptureSharedBuffer"},
+        {26, nullptr, "AcquireCallerAppletCaptureSharedBuffer"},
+        {27, nullptr, "ReleaseCallerAppletCaptureSharedBuffer"},
+    };
+    RegisterHandlers(functions);
+}
 
 IDebugFunctions::IDebugFunctions() : ServiceFramework("IDebugFunctions") {}
 
@@ -457,7 +490,7 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
         {20, &IApplicationFunctions::EnsureSaveData, "EnsureSaveData"},
         {21, &IApplicationFunctions::GetDesiredLanguage, "GetDesiredLanguage"},
         {22, &IApplicationFunctions::SetTerminateResult, "SetTerminateResult"},
-        {23, nullptr, "GetDisplayVersion"},
+        {23, &IApplicationFunctions::GetDisplayVersion, "GetDisplayVersion"},
         {24, nullptr, "GetLaunchStorageInfoForDebug"},
         {25, nullptr, "ExtendSaveData"},
         {26, nullptr, "GetSaveDataSize"},
@@ -552,6 +585,14 @@ void IApplicationFunctions::SetTerminateResult(Kernel::HLERequestContext& ctx) {
     NGLOG_WARNING(Service_AM, "(STUBBED) called, result=0x{:08X}", result);
 }
 
+void IApplicationFunctions::GetDisplayVersion(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 6};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push<u64>(1);
+    rb.Push<u64>(0);
+    NGLOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
 void IApplicationFunctions::GetDesiredLanguage(Kernel::HLERequestContext& ctx) {
     // TODO(bunnei): This should be configurable
     IPC::ResponseBuilder rb{ctx, 4};

src/core/hle/service/am/am.h

@@ -18,10 +18,25 @@ class NVFlinger;
 
 namespace AM {
 
-// TODO: Add more languages
 enum SystemLanguage {
     Japanese = 0,
-    English = 1,
+    English = 1, // en-US
+    French = 2,
+    German = 3,
+    Italian = 4,
+    Spanish = 5,
+    Chinese = 6,
+    Korean = 7,
+    Dutch = 8,
+    Portuguese = 9,
+    Russian = 10,
+    Taiwanese = 11,
+    BritishEnglish = 12, // en-GB
+    CanadianFrench = 13,
+    LatinAmericanSpanish = 14, // es-419
+    // 4.0.0+
+    SimplifiedChinese = 15,
+    TraditionalChinese = 16,
 };
 
 class IWindowController final : public ServiceFramework<IWindowController> {
@@ -117,6 +132,7 @@ private:
     void CreateApplicationAndRequestToStartForQuest(Kernel::HLERequestContext& ctx);
     void EnsureSaveData(Kernel::HLERequestContext& ctx);
     void SetTerminateResult(Kernel::HLERequestContext& ctx);
+    void GetDisplayVersion(Kernel::HLERequestContext& ctx);
     void GetDesiredLanguage(Kernel::HLERequestContext& ctx);
     void InitializeGamePlayRecording(Kernel::HLERequestContext& ctx);
     void SetGamePlayRecordingState(Kernel::HLERequestContext& ctx);

src/core/hle/service/am/applet_oe.cpp

@@ -98,7 +98,7 @@ void AppletOE::OpenApplicationProxy(Kernel::HLERequestContext& ctx) {
 AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     : ServiceFramework("appletOE"), nvflinger(std::move(nvflinger)) {
     static const FunctionInfo functions[] = {
-        {0x00000000, &AppletOE::OpenApplicationProxy, "OpenApplicationProxy"},
+        {0, &AppletOE::OpenApplicationProxy, "OpenApplicationProxy"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/audio/audin_u.cpp

@@ -35,10 +35,8 @@ public:
 
 AudInU::AudInU() : ServiceFramework("audin:u") {
     static const FunctionInfo functions[] = {
-        {0, nullptr, "ListAudioIns"},
-        {1, nullptr, "OpenAudioIn"},
-        {3, nullptr, "OpenAudioInAuto"},
-        {4, nullptr, "ListAudioInsAuto"},
+        {0, nullptr, "ListAudioIns"},    {1, nullptr, "OpenAudioIn"},      {2, nullptr, "Unknown"},
+        {3, nullptr, "OpenAudioInAuto"}, {4, nullptr, "ListAudioInsAuto"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/audio/audout_u.cpp

@@ -196,10 +196,10 @@ void AudOutU::OpenAudioOut(Kernel::HLERequestContext& ctx) {
 }
 
 AudOutU::AudOutU() : ServiceFramework("audout:u") {
-    static const FunctionInfo functions[] = {{0x00000000, &AudOutU::ListAudioOuts, "ListAudioOuts"},
-                                             {0x00000001, &AudOutU::OpenAudioOut, "OpenAudioOut"},
-                                             {0x00000002, nullptr, "ListAudioOutsAuto"},
-                                             {0x00000003, nullptr, "OpenAudioOutAuto"}};
+    static const FunctionInfo functions[] = {{0, &AudOutU::ListAudioOuts, "ListAudioOuts"},
+                                             {1, &AudOutU::OpenAudioOut, "OpenAudioOut"},
+                                             {2, nullptr, "ListAudioOutsAuto"},
+                                             {3, nullptr, "OpenAudioOutAuto"}};
     RegisterHandlers(functions);
 }
 

src/core/hle/service/audio/audrec_u.cpp

@@ -20,6 +20,7 @@ public:
             {4, nullptr, "RegisterBufferEvent"},
             {5, nullptr, "GetReleasedFinalOutputRecorderBuffer"},
             {6, nullptr, "ContainsFinalOutputRecorderBuffer"},
+            {7, nullptr, "Unknown"},
             {8, nullptr, "AppendFinalOutputRecorderBufferAuto"},
             {9, nullptr, "GetReleasedFinalOutputRecorderBufferAuto"},
         };
@@ -30,7 +31,7 @@ public:
 
 AudRecU::AudRecU() : ServiceFramework("audrec:u") {
     static const FunctionInfo functions[] = {
-        {0x00000000, nullptr, "OpenFinalOutputRecorder"},
+        {0, nullptr, "OpenFinalOutputRecorder"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/audio/audren_u.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include "common/alignment.h"
 #include "common/logging/log.h"
 #include "core/core_timing.h"
 #include "core/hle/ipc_helpers.h"
@@ -156,19 +157,20 @@ class IAudioDevice final : public ServiceFramework<IAudioDevice> {
 public:
     IAudioDevice() : ServiceFramework("IAudioDevice") {
         static const FunctionInfo functions[] = {
-            {0x0, &IAudioDevice::ListAudioDeviceName, "ListAudioDeviceName"},
-            {0x1, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolume"},
-            {0x2, nullptr, "GetAudioDeviceOutputVolume"},
-            {0x3, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceName"},
-            {0x4, &IAudioDevice::QueryAudioDeviceSystemEvent, "QueryAudioDeviceSystemEvent"},
-            {0x5, &IAudioDevice::GetActiveChannelCount, "GetActiveChannelCount"},
-            {0x6, &IAudioDevice::ListAudioDeviceName,
+            {0, &IAudioDevice::ListAudioDeviceName, "ListAudioDeviceName"},
+            {1, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolume"},
+            {2, nullptr, "GetAudioDeviceOutputVolume"},
+            {3, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceName"},
+            {4, &IAudioDevice::QueryAudioDeviceSystemEvent, "QueryAudioDeviceSystemEvent"},
+            {5, &IAudioDevice::GetActiveChannelCount, "GetActiveChannelCount"},
+            {6, &IAudioDevice::ListAudioDeviceName,
              "ListAudioDeviceNameAuto"}, // TODO(ogniK): Confirm if autos are identical to non auto
-            {0x7, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolumeAuto"},
-            {0x8, nullptr, "GetAudioDeviceOutputVolumeAuto"},
-            {0xa, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceNameAuto"},
-            {0xb, nullptr, "QueryAudioDeviceInputEvent"},
-            {0xc, nullptr, "QueryAudioDeviceOutputEvent"}};
+            {7, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolumeAuto"},
+            {8, nullptr, "GetAudioDeviceOutputVolumeAuto"},
+            {10, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceNameAuto"},
+            {11, nullptr, "QueryAudioDeviceInputEvent"},
+            {12, nullptr, "QueryAudioDeviceOutputEvent"},
+        };
         RegisterHandlers(functions);
 
         buffer_event =
@@ -255,12 +257,62 @@ void AudRenU::OpenAudioRenderer(Kernel::HLERequestContext& ctx) {
 }
 
 void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    auto params = rp.PopRaw<WorkerBufferParameters>();
+
+    u64 buffer_sz = Common::AlignUp(4 * params.unknown8, 0x40);
+    buffer_sz += params.unknownC * 1024;
+    buffer_sz += 0x940 * (params.unknownC + 1);
+    buffer_sz += 0x3F0 * params.voice_count;
+    buffer_sz += Common::AlignUp(8 * (params.unknownC + 1), 0x10);
+    buffer_sz += Common::AlignUp(8 * params.voice_count, 0x10);
+    buffer_sz +=
+        Common::AlignUp((0x3C0 * (params.sink_count + params.unknownC) + 4 * params.sample_count) *
+                            (params.unknown8 + 6),
+                        0x40);
+
+    if (IsFeatureSupported(AudioFeatures::Splitter, params.magic)) {
+        u32 count = params.unknownC + 1;
+        u64 node_count = Common::AlignUp(count, 0x40);
+        u64 node_state_buffer_sz =
+            4 * (node_count * node_count) + 0xC * node_count + 2 * (node_count / 8);
+        u64 edge_matrix_buffer_sz = 0;
+        node_count = Common::AlignUp(count * count, 0x40);
+        if (node_count >> 31 != 0) {
+            edge_matrix_buffer_sz = (node_count | 7) / 8;
+        } else {
+            edge_matrix_buffer_sz = node_count / 8;
+        }
+        buffer_sz += Common::AlignUp(node_state_buffer_sz + edge_matrix_buffer_sz, 0x10);
+    }
+
+    buffer_sz += 0x20 * (params.effect_count + 4 * params.voice_count) + 0x50;
+    if (IsFeatureSupported(AudioFeatures::Splitter, params.magic)) {
+        buffer_sz += 0xE0 * params.unknown2c;
+        buffer_sz += 0x20 * params.splitter_count;
+        buffer_sz += Common::AlignUp(4 * params.unknown2c, 0x10);
+    }
+    buffer_sz = Common::AlignUp(buffer_sz, 0x40) + 0x170 * params.sink_count;
+    u64 output_sz = buffer_sz + 0x280 * params.sink_count + 0x4B0 * params.effect_count +
+                    ((params.voice_count * 256) | 0x40);
+
+    if (params.unknown1c >= 1) {
+        output_sz = Common::AlignUp(((16 * params.sink_count + 16 * params.effect_count +
+                                      16 * params.voice_count + 16) +
+                                     0x658) *
+                                            (params.unknown1c + 1) +
+                                        0xc0,
+                                    0x40) +
+                    output_sz;
+    }
+    output_sz = Common::AlignUp(output_sz + 0x1807e, 0x1000);
+
     IPC::ResponseBuilder rb{ctx, 4};
 
     rb.Push(RESULT_SUCCESS);
-    rb.Push<u64>(0x4000);
+    rb.Push<u64>(output_sz);
 
-    NGLOG_WARNING(Service_Audio, "(STUBBED) called");
+    NGLOG_DEBUG(Service_Audio, "called, buffer_size=0x{:X}", output_sz);
 }
 
 void AudRenU::GetAudioDevice(Kernel::HLERequestContext& ctx) {
@@ -272,4 +324,14 @@ void AudRenU::GetAudioDevice(Kernel::HLERequestContext& ctx) {
     NGLOG_DEBUG(Service_Audio, "called");
 }
 
+bool AudRenU::IsFeatureSupported(AudioFeatures feature, u32_le revision) const {
+    u32_be version_num = (revision - Common::MakeMagic('R', 'E', 'V', '0')); // Byte swap
+    switch (feature) {
+    case AudioFeatures::Splitter:
+        return version_num >= 2;
+    default:
+        return false;
+    }
+}
+
 } // namespace Service::Audio

src/core/hle/service/audio/audren_u.h

@@ -21,6 +21,31 @@ private:
     void OpenAudioRenderer(Kernel::HLERequestContext& ctx);
     void GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx);
     void GetAudioDevice(Kernel::HLERequestContext& ctx);
+
+    struct WorkerBufferParameters {
+        u32_le sample_rate;
+        u32_le sample_count;
+        u32_le unknown8;
+        u32_le unknownC;
+        u32_le voice_count;
+        u32_le sink_count;
+        u32_le effect_count;
+        u32_le unknown1c;
+        u8 unknown20;
+        u8 padding1[3];
+        u32_le splitter_count;
+        u32_le unknown2c;
+        u8 padding2[4];
+        u32_le magic;
+    };
+    static_assert(sizeof(WorkerBufferParameters) == 52,
+                  "WorkerBufferParameters is an invalid size");
+
+    enum class AudioFeatures : u32 {
+        Splitter,
+    };
+
+    bool IsFeatureSupported(AudioFeatures feature, u32_le revision) const;
 };
 
 } // namespace Service::Audio

src/core/hle/service/audio/codecctl.cpp

@@ -11,19 +11,19 @@ namespace Service::Audio {
 
 CodecCtl::CodecCtl() : ServiceFramework("codecctl") {
     static const FunctionInfo functions[] = {
-        {0x00000000, nullptr, "InitializeCodecController"},
-        {0x00000001, nullptr, "FinalizeCodecController"},
-        {0x00000002, nullptr, "SleepCodecController"},
-        {0x00000003, nullptr, "WakeCodecController"},
-        {0x00000004, nullptr, "SetCodecVolume"},
-        {0x00000005, nullptr, "GetCodecVolumeMax"},
-        {0x00000006, nullptr, "GetCodecVolumeMin"},
-        {0x00000007, nullptr, "SetCodecActiveTarget"},
-        {0x00000008, nullptr, "Unknown"},
-        {0x00000009, nullptr, "BindCodecHeadphoneMicJackInterrupt"},
-        {0x00000010, nullptr, "IsCodecHeadphoneMicJackInserted"},
-        {0x00000011, nullptr, "ClearCodecHeadphoneMicJackInterrupt"},
-        {0x00000012, nullptr, "IsCodecDeviceRequested"},
+        {0, nullptr, "InitializeCodecController"},
+        {1, nullptr, "FinalizeCodecController"},
+        {2, nullptr, "SleepCodecController"},
+        {3, nullptr, "WakeCodecController"},
+        {4, nullptr, "SetCodecVolume"},
+        {5, nullptr, "GetCodecVolumeMax"},
+        {6, nullptr, "GetCodecVolumeMin"},
+        {7, nullptr, "SetCodecActiveTarget"},
+        {8, nullptr, "GetCodecActiveTarget"},
+        {9, nullptr, "BindCodecHeadphoneMicJackInterrupt"},
+        {10, nullptr, "IsCodecHeadphoneMicJackInserted"},
+        {11, nullptr, "ClearCodecHeadphoneMicJackInterrupt"},
+        {12, nullptr, "IsCodecDeviceRequested"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/bcat/bcat.cpp

@@ -0,0 +1,16 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/hle/service/bcat/bcat.h"
+
+namespace Service::BCAT {
+
+BCAT::BCAT(std::shared_ptr<Module> module, const char* name)
+    : Module::Interface(std::move(module), name) {
+    static const FunctionInfo functions[] = {
+        {0, &BCAT::CreateBcatService, "CreateBcatService"},
+    };
+    RegisterHandlers(functions);
+}
+} // namespace Service::BCAT

src/core/hle/service/bcat/bcat.h

@@ -0,0 +1,16 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/service/bcat/module.h"
+
+namespace Service::BCAT {
+
+class BCAT final : public Module::Interface {
+public:
+    explicit BCAT(std::shared_ptr<Module> module, const char* name);
+};
+
+} // namespace Service::BCAT

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

@@ -0,0 +1,53 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/logging/log.h"
+#include "core/hle/ipc_helpers.h"
+#include "core/hle/service/bcat/bcat.h"
+#include "core/hle/service/bcat/module.h"
+
+namespace Service::BCAT {
+
+class IBcatService final : public ServiceFramework<IBcatService> {
+public:
+    IBcatService() : ServiceFramework("IBcatService") {
+        static const FunctionInfo functions[] = {
+            {10100, nullptr, "RequestSyncDeliveryCache"},
+            {10101, nullptr, "RequestSyncDeliveryCacheWithDirectoryName"},
+            {10200, nullptr, "CancelSyncDeliveryCacheRequest"},
+            {20100, nullptr, "RequestSyncDeliveryCacheWithApplicationId"},
+            {20101, nullptr, "RequestSyncDeliveryCacheWithApplicationIdAndDirectoryName"},
+            {30100, nullptr, "SetPassphrase"},
+            {30200, nullptr, "RegisterBackgroundDeliveryTask"},
+            {30201, nullptr, "UnregisterBackgroundDeliveryTask"},
+            {30202, nullptr, "BlockDeliveryTask"},
+            {30203, nullptr, "UnblockDeliveryTask"},
+            {90100, nullptr, "EnumerateBackgroundDeliveryTask"},
+            {90200, nullptr, "GetDeliveryList"},
+            {90201, nullptr, "ClearDeliveryCacheStorage"},
+            {90300, nullptr, "GetPushNotificationLog"},
+        };
+        RegisterHandlers(functions);
+    }
+};
+
+void Module::Interface::CreateBcatService(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushIpcInterface<IBcatService>();
+    NGLOG_DEBUG(Service_BCAT, "called");
+}
+
+Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
+    : ServiceFramework(name), module(std::move(module)) {}
+
+void InstallInterfaces(SM::ServiceManager& service_manager) {
+    auto module = std::make_shared<Module>();
+    std::make_shared<BCAT>(module, "bcat:a")->InstallAsService(service_manager);
+    std::make_shared<BCAT>(module, "bcat:m")->InstallAsService(service_manager);
+    std::make_shared<BCAT>(module, "bcat:u")->InstallAsService(service_manager);
+    std::make_shared<BCAT>(module, "bcat:s")->InstallAsService(service_manager);
+}
+
+} // namespace Service::BCAT

src/core/hle/service/bcat/module.h

@@ -0,0 +1,27 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/service/service.h"
+
+namespace Service::BCAT {
+
+class Module final {
+public:
+    class Interface : public ServiceFramework<Interface> {
+    public:
+        Interface(std::shared_ptr<Module> module, const char* name);
+
+        void CreateBcatService(Kernel::HLERequestContext& ctx);
+
+    protected:
+        std::shared_ptr<Module> module;
+    };
+};
+
+/// Registers all BCAT services with the specified service manager.
+void InstallInterfaces(SM::ServiceManager& service_manager);
+
+} // namespace Service::BCAT

src/core/hle/service/fatal/fatal.cpp

@@ -13,7 +13,7 @@ namespace Service::Fatal {
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
     : ServiceFramework(name), module(std::move(module)) {}
 
-void Module::Interface::FatalSimple(Kernel::HLERequestContext& ctx) {
+void Module::Interface::ThrowFatalWithPolicy(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp(ctx);
     u32 error_code = rp.Pop<u32>();
     NGLOG_WARNING(Service_Fatal, "(STUBBED) called, error_code=0x{:X}", error_code);
@@ -21,7 +21,7 @@ void Module::Interface::FatalSimple(Kernel::HLERequestContext& ctx) {
     rb.Push(RESULT_SUCCESS);
 }
 
-void Module::Interface::TransitionToFatalError(Kernel::HLERequestContext& ctx) {
+void Module::Interface::ThrowFatalWithCpuContext(Kernel::HLERequestContext& ctx) {
     NGLOG_WARNING(Service_Fatal, "(STUBBED) called");
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(RESULT_SUCCESS);

src/core/hle/service/fatal/fatal.h

@@ -14,8 +14,8 @@ public:
     public:
         Interface(std::shared_ptr<Module> module, const char* name);
 
-        void FatalSimple(Kernel::HLERequestContext& ctx);
-        void TransitionToFatalError(Kernel::HLERequestContext& ctx);
+        void ThrowFatalWithPolicy(Kernel::HLERequestContext& ctx);
+        void ThrowFatalWithCpuContext(Kernel::HLERequestContext& ctx);
 
     protected:
         std::shared_ptr<Module> module;

src/core/hle/service/fatal/fatal_u.cpp

@@ -8,8 +8,9 @@ namespace Service::Fatal {
 
 Fatal_U::Fatal_U(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "fatal:u") {
     static const FunctionInfo functions[] = {
-        {1, &Fatal_U::FatalSimple, "FatalSimple"},
-        {2, &Fatal_U::TransitionToFatalError, "TransitionToFatalError"},
+        {0, nullptr, "ThrowFatal"},
+        {1, &Fatal_U::ThrowFatalWithPolicy, "ThrowFatalWithPolicy"},
+        {2, &Fatal_U::ThrowFatalWithCpuContext, "ThrowFatalWithCpuContext"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/hid/hid.cpp

@@ -329,6 +329,7 @@ public:
             {130, nullptr, "SwapNpadAssignment"},
             {131, nullptr, "IsUnintendedHomeButtonInputProtectionEnabled"},
             {132, nullptr, "EnableUnintendedHomeButtonInputProtection"},
+            {133, nullptr, "SetNpadJoyAssignmentModeSingleWithDestination"},
             {200, &Hid::GetVibrationDeviceInfo, "GetVibrationDeviceInfo"},
             {201, &Hid::SendVibrationValue, "SendVibrationValue"},
             {202, &Hid::GetActualVibrationValue, "GetActualVibrationValue"},
@@ -336,12 +337,41 @@ public:
             {204, nullptr, "PermitVibration"},
             {205, nullptr, "IsVibrationPermitted"},
             {206, &Hid::SendVibrationValues, "SendVibrationValues"},
+            {207, nullptr, "SendVibrationGcErmCommand"},
+            {208, nullptr, "GetActualVibrationGcErmCommand"},
+            {209, nullptr, "BeginPermitVibrationSession"},
+            {210, nullptr, "EndPermitVibrationSession"},
             {300, nullptr, "ActivateConsoleSixAxisSensor"},
             {301, nullptr, "StartConsoleSixAxisSensor"},
             {302, nullptr, "StopConsoleSixAxisSensor"},
+            {303, nullptr, "ActivateSevenSixAxisSensor"},
+            {304, nullptr, "StartSevenSixAxisSensor"},
+            {305, nullptr, "StopSevenSixAxisSensor"},
+            {306, nullptr, "InitializeSevenSixAxisSensor"},
+            {307, nullptr, "FinalizeSevenSixAxisSensor"},
+            {308, nullptr, "SetSevenSixAxisSensorFusionStrength"},
+            {309, nullptr, "GetSevenSixAxisSensorFusionStrength"},
             {400, nullptr, "IsUsbFullKeyControllerEnabled"},
             {401, nullptr, "EnableUsbFullKeyController"},
             {402, nullptr, "IsUsbFullKeyControllerConnected"},
+            {403, nullptr, "HasBattery"},
+            {404, nullptr, "HasLeftRightBattery"},
+            {405, nullptr, "GetNpadInterfaceType"},
+            {406, nullptr, "GetNpadLeftRightInterfaceType"},
+            {500, nullptr, "GetPalmaConnectionHandle"},
+            {501, nullptr, "InitializePalma"},
+            {502, nullptr, "AcquirePalmaOperationCompleteEvent"},
+            {503, nullptr, "GetPalmaOperationInfo"},
+            {504, nullptr, "PlayPalmaActivity"},
+            {505, nullptr, "SetPalmaFrModeType"},
+            {506, nullptr, "ReadPalmaStep"},
+            {507, nullptr, "EnablePalmaStep"},
+            {508, nullptr, "SuspendPalmaStep"},
+            {509, nullptr, "ResetPalmaStep"},
+            {510, nullptr, "ReadPalmaApplicationSection"},
+            {511, nullptr, "WritePalmaApplicationSection"},
+            {512, nullptr, "ReadPalmaUniqueCode"},
+            {513, nullptr, "SetPalmaUniqueCodeInvalid"},
             {1000, nullptr, "SetNpadCommunicationMode"},
             {1001, nullptr, "GetNpadCommunicationMode"},
         };

src/core/hle/service/hid/hid.h

@@ -12,7 +12,7 @@ namespace Service::HID {
 // Begin enums and output structs
 
 constexpr u32 HID_NUM_ENTRIES = 17;
-constexpr u32 HID_NUM_LAYOUTS = 7;
+constexpr u32 HID_NUM_LAYOUTS = 2;
 constexpr s32 HID_JOYSTICK_MAX = 0x8000;
 constexpr s32 HID_JOYSTICK_MIN = -0x8000;
 

src/core/hle/service/nfp/nfp.cpp

@@ -17,30 +17,30 @@ public:
     IUser() : ServiceFramework("IUser") {
         static const FunctionInfo functions[] = {
             {0, &IUser::Initialize, "Initialize"},
-            {1, nullptr, "Unknown1"},
-            {2, nullptr, "Unknown2"},
-            {3, nullptr, "Unknown3"},
-            {4, nullptr, "Unknown4"},
-            {5, nullptr, "Unknown5"},
-            {6, nullptr, "Unknown6"},
-            {7, nullptr, "Unknown7"},
-            {8, nullptr, "Unknown8"},
-            {9, nullptr, "Unknown9"},
-            {10, nullptr, "Unknown10"},
-            {11, nullptr, "Unknown11"},
-            {12, nullptr, "Unknown12"},
-            {13, nullptr, "Unknown13"},
-            {14, nullptr, "Unknown14"},
-            {15, nullptr, "Unknown15"},
-            {16, nullptr, "Unknown16"},
-            {17, nullptr, "Unknown17"},
-            {18, nullptr, "Unknown18"},
-            {19, nullptr, "Unknown19"},
-            {20, nullptr, "Unknown20"},
-            {21, nullptr, "Unknown21"},
-            {22, nullptr, "Unknown22"},
-            {23, nullptr, "Unknown23"},
-            {24, nullptr, "Unknown24"},
+            {1, nullptr, "Finalize"},
+            {2, nullptr, "ListDevices"},
+            {3, nullptr, "StartDetection"},
+            {4, nullptr, "StopDetection"},
+            {5, nullptr, "Mount"},
+            {6, nullptr, "Unmount"},
+            {7, nullptr, "OpenApplicationArea"},
+            {8, nullptr, "GetApplicationArea"},
+            {9, nullptr, "SetApplicationArea"},
+            {10, nullptr, "Flush"},
+            {11, nullptr, "Restore"},
+            {12, nullptr, "CreateApplicationArea"},
+            {13, nullptr, "GetTagInfo"},
+            {14, nullptr, "GetRegisterInfo"},
+            {15, nullptr, "GetCommonInfo"},
+            {16, nullptr, "GetModelInfo"},
+            {17, nullptr, "AttachActivateEvent"},
+            {18, nullptr, "AttachDeactivateEvent"},
+            {19, nullptr, "GetState"},
+            {20, nullptr, "GetDeviceState"},
+            {21, nullptr, "GetNpadId"},
+            {22, nullptr, "GetApplicationArea2"},
+            {23, nullptr, "AttachAvailabilityChangeEvent"},
+            {24, nullptr, "RecreateApplicationArea"},
         };
         RegisterHandlers(functions);
     }

src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp

@@ -26,6 +26,8 @@ u32 nvhost_as_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vecto
         return BindChannel(input, output);
     case IoctlCommand::IocGetVaRegionsCommand:
         return GetVARegions(input, output);
+    case IoctlCommand::IocUnmapBufferCommand:
+        return UnmapBuffer(input, output);
     }
 
     if (static_cast<IoctlCommand>(command.cmd.Value()) == IoctlCommand::IocRemapCommand)
@@ -125,6 +127,37 @@ u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& ou
         params.offset = gpu.memory_manager->MapBufferEx(object->addr, object->size);
     }
 
+    // Create a new mapping entry for this operation.
+    ASSERT_MSG(buffer_mappings.find(params.offset) == buffer_mappings.end(),
+               "Offset is already mapped");
+
+    BufferMapping mapping{};
+    mapping.nvmap_handle = params.nvmap_handle;
+    mapping.offset = params.offset;
+    mapping.size = object->size;
+
+    buffer_mappings[params.offset] = mapping;
+
+    std::memcpy(output.data(), &params, output.size());
+    return 0;
+}
+
+u32 nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlUnmapBuffer params{};
+    std::memcpy(&params, input.data(), input.size());
+
+    NGLOG_DEBUG(Service_NVDRV, "called, offset=0x{:X}", params.offset);
+
+    auto& gpu = Core::System::GetInstance().GPU();
+
+    auto itr = buffer_mappings.find(params.offset);
+
+    ASSERT_MSG(itr != buffer_mappings.end(), "Tried to unmap invalid mapping");
+
+    params.offset = gpu.memory_manager->UnmapBuffer(params.offset, itr->second.size);
+
+    buffer_mappings.erase(itr->second.offset);
+
     std::memcpy(output.data(), &params, output.size());
     return 0;
 }

src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <unordered_map>
 #include <utility>
 #include <vector>
 #include "common/common_types.h"
@@ -30,6 +31,7 @@ private:
         IocMapBufferExCommand = 0xC0284106,
         IocBindChannelCommand = 0x40044101,
         IocGetVaRegionsCommand = 0xC0404108,
+        IocUnmapBufferCommand = 0xC0084105,
     };
 
     struct IoctlInitalizeEx {
@@ -76,6 +78,11 @@ private:
     };
     static_assert(sizeof(IoctlMapBufferEx) == 40, "IoctlMapBufferEx is incorrect size");
 
+    struct IoctlUnmapBuffer {
+        u64_le offset;
+    };
+    static_assert(sizeof(IoctlUnmapBuffer) == 8, "IoctlUnmapBuffer is incorrect size");
+
     struct IoctlBindChannel {
         u32_le fd;
     };
@@ -98,12 +105,22 @@ private:
     static_assert(sizeof(IoctlGetVaRegions) == 16 + sizeof(IoctlVaRegion) * 2,
                   "IoctlGetVaRegions is incorrect size");
 
+    struct BufferMapping {
+        u64 offset;
+        u64 size;
+        u32 nvmap_handle;
+    };
+
+    /// Map containing the nvmap object mappings in GPU memory.
+    std::unordered_map<u64, BufferMapping> buffer_mappings;
+
     u32 channel{};
 
     u32 InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output);
     u32 AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output);
     u32 Remap(const std::vector<u8>& input, std::vector<u8>& output);
     u32 MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
     u32 BindChannel(const std::vector<u8>& input, std::vector<u8>& output);
     u32 GetVARegions(const std::vector<u8>& input, std::vector<u8>& output);
 

src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp

@@ -16,7 +16,11 @@ u32 nvhost_ctrl::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<
     case IoctlCommand::IocGetConfigCommand:
         return NvOsGetConfigU32(input, output);
     case IoctlCommand::IocCtrlEventWaitCommand:
-        return IocCtrlEventWait(input, output);
+        return IocCtrlEventWait(input, output, false);
+    case IoctlCommand::IocCtrlEventWaitAsyncCommand:
+        return IocCtrlEventWait(input, output, true);
+    case IoctlCommand::IocCtrlEventRegisterCommand:
+        return IocCtrlEventRegister(input, output);
     }
     UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
@@ -36,7 +40,7 @@ u32 nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>&
         } else if (!strcmp(params.param_str.data(), "NVRM_GPU_PREVENT_USE")) {
             params.config_str[0] = '0';
         } else {
-            params.config_str[0] = '0';
+            params.config_str[0] = '\0';
         }
     } else {
         UNIMPLEMENTED(); // unknown domain? Only nv has been seen so far on hardware
@@ -45,11 +49,13 @@ u32 nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>&
     return 0;
 }
 
-u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output) {
+u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output,
+                                  bool is_async) {
     IocCtrlEventWaitParams params{};
     std::memcpy(&params, input.data(), sizeof(params));
-    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called, syncpt_id={} threshold={} timeout={}",
-                  params.syncpt_id, params.threshold, params.timeout);
+    NGLOG_WARNING(Service_NVDRV,
+                  "(STUBBED) called, syncpt_id={}, threshold={}, timeout={}, is_async={}",
+                  params.syncpt_id, params.threshold, params.timeout, is_async);
 
     // TODO(Subv): Implement actual syncpt waiting.
     params.value = 0;
@@ -57,4 +63,10 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>&
     return 0;
 }
 
+u32 nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output) {
+    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called");
+    // TODO(bunnei): Implement this.
+    return 0;
+}
+
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_ctrl.h

@@ -26,12 +26,64 @@ private:
         IocSyncptIncrCommand = 0x40040015,
         IocSyncptWaitCommand = 0xC00C0016,
         IocModuleMutexCommand = 0x40080017,
-        IocModuleRegRDWRCommand = 0xC008010E,
+        IocModuleRegRDWRCommand = 0xC0180018,
         IocSyncptWaitexCommand = 0xC0100019,
         IocSyncptReadMaxCommand = 0xC008001A,
-        IocCtrlEventWaitCommand = 0xC010001D,
         IocGetConfigCommand = 0xC183001B,
+        IocCtrlEventSignalCommand = 0xC004001C,
+        IocCtrlEventWaitCommand = 0xC010001D,
+        IocCtrlEventWaitAsyncCommand = 0xC010001E,
+        IocCtrlEventRegisterCommand = 0xC004001F,
+        IocCtrlEventUnregisterCommand = 0xC0040020,
+        IocCtrlEventKillCommand = 0x40080021,
     };
+    struct IocSyncptReadParams {
+        u32_le id;
+        u32_le value;
+    };
+    static_assert(sizeof(IocSyncptReadParams) == 8, "IocSyncptReadParams is incorrect size");
+
+    struct IocSyncptIncrParams {
+        u32_le id;
+    };
+    static_assert(sizeof(IocSyncptIncrParams) == 4, "IocSyncptIncrParams is incorrect size");
+
+    struct IocSyncptWaitParams {
+        u32_le id;
+        u32_le thresh;
+        s32_le timeout;
+    };
+    static_assert(sizeof(IocSyncptWaitParams) == 12, "IocSyncptWaitParams is incorrect size");
+
+    struct IocModuleMutexParams {
+        u32_le id;
+        u32_le lock; // (0 = unlock and 1 = lock)
+    };
+    static_assert(sizeof(IocModuleMutexParams) == 8, "IocModuleMutexParams is incorrect size");
+
+    struct IocModuleRegRDWRParams {
+        u32_le id;
+        u32_le num_offsets;
+        u32_le block_size;
+        u32_le offsets;
+        u32_le values;
+        u32_le write;
+    };
+    static_assert(sizeof(IocModuleRegRDWRParams) == 24, "IocModuleRegRDWRParams is incorrect size");
+
+    struct IocSyncptWaitexParams {
+        u32_le id;
+        u32_le thresh;
+        s32_le timeout;
+        u32_le value;
+    };
+    static_assert(sizeof(IocSyncptWaitexParams) == 16, "IocSyncptWaitexParams is incorrect size");
+
+    struct IocSyncptReadMaxParams {
+        u32_le id;
+        u32_le value;
+    };
+    static_assert(sizeof(IocSyncptReadMaxParams) == 8, "IocSyncptReadMaxParams is incorrect size");
 
     struct IocGetConfigParams {
         std::array<char, 0x41> domain_str;
@@ -40,6 +92,12 @@ private:
     };
     static_assert(sizeof(IocGetConfigParams) == 387, "IocGetConfigParams is incorrect size");
 
+    struct IocCtrlEventSignalParams {
+        u32_le user_event_id;
+    };
+    static_assert(sizeof(IocCtrlEventSignalParams) == 4,
+                  "IocCtrlEventSignalParams is incorrect size");
+
     struct IocCtrlEventWaitParams {
         u32_le syncpt_id;
         u32_le threshold;
@@ -48,9 +106,37 @@ private:
     };
     static_assert(sizeof(IocCtrlEventWaitParams) == 16, "IocCtrlEventWaitParams is incorrect size");
 
+    struct IocCtrlEventWaitAsyncParams {
+        u32_le syncpt_id;
+        u32_le threshold;
+        u32_le timeout;
+        u32_le value;
+    };
+    static_assert(sizeof(IocCtrlEventWaitAsyncParams) == 16,
+                  "IocCtrlEventWaitAsyncParams is incorrect size");
+
+    struct IocCtrlEventRegisterParams {
+        u32_le user_event_id;
+    };
+    static_assert(sizeof(IocCtrlEventRegisterParams) == 4,
+                  "IocCtrlEventRegisterParams is incorrect size");
+
+    struct IocCtrlEventUnregisterParams {
+        u32_le user_event_id;
+    };
+    static_assert(sizeof(IocCtrlEventUnregisterParams) == 4,
+                  "IocCtrlEventUnregisterParams is incorrect size");
+
+    struct IocCtrlEventKill {
+        u64_le user_events;
+    };
+    static_assert(sizeof(IocCtrlEventKill) == 8, "IocCtrlEventKill is incorrect size");
+
     u32 NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output);
 
-    u32 IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, bool is_async);
+
+    u32 IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output);
 };
 
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp

@@ -24,6 +24,8 @@ u32 nvhost_ctrl_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vec
         return ZCullGetCtxSize(input, output);
     case IoctlCommand::IocZcullGetInfo:
         return ZCullGetInfo(input, output);
+    case IoctlCommand::IocZbcSetTable:
+        return ZBCSetTable(input, output);
     }
     UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
@@ -77,9 +79,13 @@ u32 nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vecto
 u32 nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output) {
     IoctlGpuGetTpcMasksArgs params{};
     std::memcpy(&params, input.data(), input.size());
-    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called, mask=0x{:X}, mask_buf_addr=0x{:X}",
-                  params.mask_buf_size, params.mask_buf_addr);
-    params.unk = 0xcafe; // TODO(ogniK): Needs to be non 0, what does this actually do?
+    NGLOG_INFO(Service_NVDRV, "called, mask=0x{:X}, mask_buf_addr=0x{:X}", params.mask_buf_size,
+               params.mask_buf_addr);
+    // TODO(ogniK): Confirm value on hardware
+    if (params.mask_buf_size)
+        params.tpc_mask_size = 4 * 1; // 4 * num_gpc
+    else
+        params.tpc_mask_size = 0;
     std::memcpy(output.data(), &params, sizeof(params));
     return 0;
 }
@@ -121,4 +127,13 @@ u32 nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>&
     return 0;
 }
 
+u32 nvhost_ctrl_gpu::ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output) {
+    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called");
+    IoctlZbcSetTable params{};
+    std::memcpy(&params, input.data(), input.size());
+    // TODO(ogniK): What does this even actually do?
+    std::memcpy(output.data(), &params, output.size());
+    return 0;
+}
+
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h

@@ -25,6 +25,7 @@ private:
         IocGetActiveSlotMaskCommand = 0x80084714,
         IocZcullGetCtxSizeCommand = 0x80044701,
         IocZcullGetInfo = 0x80284702,
+        IocZbcSetTable = 0x402C4703,
     };
 
     struct IoctlGpuCharacteristics {
@@ -86,7 +87,7 @@ private:
         /// [in]  pointer to TPC mask buffer. It will receive one 32-bit TPC mask per GPC or 0 if
         /// GPC is not enabled or not present. This parameter is ignored if mask_buf_size is 0.
         u64_le mask_buf_addr;
-        u64_le unk; // Nintendo add this?
+        u64_le tpc_mask_size; // Nintendo add this?
     };
     static_assert(sizeof(IoctlGpuGetTpcMasksArgs) == 24,
                   "IoctlGpuGetTpcMasksArgs is incorrect size");
@@ -117,11 +118,21 @@ private:
     static_assert(sizeof(IoctlNvgpuGpuZcullGetInfoArgs) == 40,
                   "IoctlNvgpuGpuZcullGetInfoArgs is incorrect size");
 
+    struct IoctlZbcSetTable {
+        u32_le color_ds[4];
+        u32_le color_l2[4];
+        u32_le depth;
+        u32_le format;
+        u32_le type;
+    };
+    static_assert(sizeof(IoctlZbcSetTable) == 44, "IoctlZbcSetTable is incorrect size");
+
     u32 GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output);
     u32 GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output);
     u32 GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output);
     u32 ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output);
     u32 ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output);
 };
 
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp

@@ -34,6 +34,8 @@ u32 nvhost_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u
         return AllocateObjectContext(input, output);
     case IoctlCommand::IocChannelGetWaitbaseCommand:
         return GetWaitbase(input, output);
+    case IoctlCommand::IocChannelSetTimeoutCommand:
+        return ChannelSetTimeout(input, output);
     }
 
     if (command.group == NVGPU_IOCTL_MAGIC) {
@@ -149,4 +151,11 @@ u32 nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& outpu
     return 0;
 }
 
+u32 nvhost_gpu::ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlChannelSetTimeout params{};
+    std::memcpy(&params, input.data(), sizeof(IoctlChannelSetTimeout));
+    NGLOG_INFO(Service_NVDRV, "called, timeout=0x{:X}", params.timeout);
+    return 0;
+}
+
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_gpu.h

@@ -26,14 +26,23 @@ public:
 private:
     enum class IoctlCommand : u32_le {
         IocSetNVMAPfdCommand = 0x40044801,
+        IocAllocGPFIFOCommand = 0x40084805,
         IocSetClientDataCommand = 0x40084714,
         IocGetClientDataCommand = 0x80084715,
         IocZCullBind = 0xc010480b,
         IocSetErrorNotifierCommand = 0xC018480C,
         IocChannelSetPriorityCommand = 0x4004480D,
+        IocEnableCommand = 0x0000480E,
+        IocDisableCommand = 0x0000480F,
+        IocPreemptCommand = 0x00004810,
+        IocForceResetCommand = 0x00004811,
+        IocEventIdControlCommand = 0x40084812,
+        IocGetErrorNotificationCommand = 0xC0104817,
+        IocAllocGPFIFOExCommand = 0x40204818,
         IocAllocGPFIFOEx2Command = 0xC020481A,
         IocAllocObjCtxCommand = 0xC0104809,
         IocChannelGetWaitbaseCommand = 0xC0080003,
+        IocChannelSetTimeoutCommand = 0x40044803,
     };
 
     enum class CtxObjects : u32_le {
@@ -50,6 +59,17 @@ private:
     };
     static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size");
 
+    struct IoctlChannelSetTimeout {
+        u32_le timeout;
+    };
+    static_assert(sizeof(IoctlChannelSetTimeout) == 4, "IoctlChannelSetTimeout is incorrect size");
+
+    struct IoctlAllocGPFIFO {
+        u32_le num_entries;
+        u32_le flags;
+    };
+    static_assert(sizeof(IoctlAllocGPFIFO) == 8, "IoctlAllocGPFIFO is incorrect size");
+
     struct IoctlClientData {
         u64_le data;
     };
@@ -70,12 +90,45 @@ private:
     };
     static_assert(sizeof(IoctlSetErrorNotifier) == 24, "IoctlSetErrorNotifier is incorrect size");
 
+    struct IoctlChannelSetPriority {
+        u32_le priority;
+    };
+    static_assert(sizeof(IoctlChannelSetPriority) == 4,
+                  "IoctlChannelSetPriority is incorrect size");
+
+    struct IoctlEventIdControl {
+        u32_le cmd; // 0=disable, 1=enable, 2=clear
+        u32_le id;
+    };
+    static_assert(sizeof(IoctlEventIdControl) == 8, "IoctlEventIdControl is incorrect size");
+
+    struct IoctlGetErrorNotification {
+        u64_le timestamp;
+        u32_le info32;
+        u16_le info16;
+        u16_le status; // always 0xFFFF
+    };
+    static_assert(sizeof(IoctlGetErrorNotification) == 16,
+                  "IoctlGetErrorNotification is incorrect size");
+
     struct IoctlFence {
         u32_le id;
         u32_le value;
     };
     static_assert(sizeof(IoctlFence) == 8, "IoctlFence is incorrect size");
 
+    struct IoctlAllocGpfifoEx {
+        u32_le num_entries;
+        u32_le flags;
+        u32_le unk0;
+        u32_le unk1;
+        u32_le unk2;
+        u32_le unk3;
+        u32_le unk4;
+        u32_le unk5;
+    };
+    static_assert(sizeof(IoctlAllocGpfifoEx) == 32, "IoctlAllocGpfifoEx is incorrect size");
+
     struct IoctlAllocGpfifoEx2 {
         u32_le num_entries;   // in
         u32_le flags;         // in
@@ -141,6 +194,7 @@ private:
     u32 AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output);
     u32 SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& output);
     u32 GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output);
 
     std::shared_ptr<nvmap> nvmap_dev;
 };

src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp

@@ -0,0 +1,32 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/assert.h"
+#include "common/logging/log.h"
+#include "core/hle/service/nvdrv/devices/nvhost_nvdec.h"
+
+namespace Service::Nvidia::Devices {
+
+u32 nvhost_nvdec::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
+    NGLOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}",
+                command.raw, input.size(), output.size());
+
+    switch (static_cast<IoctlCommand>(command.raw)) {
+    case IoctlCommand::IocSetNVMAPfdCommand:
+        return SetNVMAPfd(input, output);
+    }
+
+    UNIMPLEMENTED_MSG("Unimplemented ioctl");
+    return 0;
+}
+
+u32 nvhost_nvdec::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlSetNvmapFD params{};
+    std::memcpy(&params, input.data(), input.size());
+    NGLOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd);
+    nvmap_fd = params.nvmap_fd;
+    return 0;
+}
+
+} // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_nvdec.h

@@ -0,0 +1,38 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <cstdlib>
+#include <cstring>
+#include <vector>
+#include "common/common_types.h"
+#include "core/hle/service/nvdrv/devices/nvdevice.h"
+
+namespace Service::Nvidia::Devices {
+
+class nvhost_nvdec final : public nvdevice {
+public:
+    nvhost_nvdec() = default;
+    ~nvhost_nvdec() override = default;
+
+    u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
+
+private:
+    enum class IoctlCommand : u32_le {
+        IocSetNVMAPfdCommand = 0x40044801,
+    };
+
+    struct IoctlSetNvmapFD {
+        u32_le nvmap_fd;
+    };
+    static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size");
+
+    u32_le nvmap_fd{};
+
+    u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output);
+};
+
+} // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvmap.cpp

@@ -30,6 +30,8 @@ u32 nvmap::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& o
         return IocFromId(input, output);
     case IoctlCommand::Param:
         return IocParam(input, output);
+    case IoctlCommand::Free:
+        return IocFree(input, output);
     }
 
     UNIMPLEMENTED_MSG("Unimplemented ioctl");
@@ -45,6 +47,7 @@ u32 nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) {
     object->id = next_id++;
     object->size = params.size;
     object->status = Object::Status::Created;
+    object->refcount = 1;
 
     u32 handle = next_handle++;
     handles[handle] = std::move(object);
@@ -101,6 +104,8 @@ u32 nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) {
                             [&](const auto& entry) { return entry.second->id == params.id; });
     ASSERT(itr != handles.end());
 
+    itr->second->refcount++;
+
     // Return the existing handle instead of creating a new one.
     params.handle = itr->first;
 
@@ -114,25 +119,25 @@ u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) {
     IocParamParams params;
     std::memcpy(&params, input.data(), sizeof(params));
 
-    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called type={}", params.type);
+    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called type={}", params.param);
 
     auto object = GetObject(params.handle);
     ASSERT(object);
     ASSERT(object->status == Object::Status::Allocated);
 
-    switch (static_cast<ParamTypes>(params.type)) {
+    switch (static_cast<ParamTypes>(params.param)) {
     case ParamTypes::Size:
-        params.value = object->size;
+        params.result = object->size;
         break;
     case ParamTypes::Alignment:
-        params.value = object->align;
+        params.result = object->align;
         break;
     case ParamTypes::Heap:
         // TODO(Subv): Seems to be a hardcoded value?
-        params.value = 0x40000000;
+        params.result = 0x40000000;
         break;
     case ParamTypes::Kind:
-        params.value = object->kind;
+        params.result = object->kind;
         break;
     default:
         UNIMPLEMENTED();
@@ -142,4 +147,34 @@ u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) {
     return 0;
 }
 
+u32 nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) {
+    enum FreeFlags {
+        Freed = 0,
+        NotFreedYet = 1,
+    };
+
+    IocFreeParams params;
+    std::memcpy(&params, input.data(), sizeof(params));
+
+    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called");
+
+    auto itr = handles.find(params.handle);
+    ASSERT(itr != handles.end());
+
+    itr->second->refcount--;
+
+    params.refcount = itr->second->refcount;
+    params.size = itr->second->size;
+
+    if (itr->second->refcount == 0)
+        params.flags = Freed;
+    else
+        params.flags = NotFreedYet;
+
+    handles.erase(params.handle);
+
+    std::memcpy(output.data(), &params, sizeof(params));
+    return 0;
+}
+
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvmap.h

@@ -34,6 +34,7 @@ public:
         u8 kind;
         VAddr addr;
         Status status;
+        u32 refcount;
     };
 
     std::shared_ptr<Object> GetObject(u32 handle) const {
@@ -58,16 +59,25 @@ private:
         Create = 0xC0080101,
         FromId = 0xC0080103,
         Alloc = 0xC0200104,
+        Free = 0xC0180105,
         Param = 0xC00C0109,
-        GetId = 0xC008010E
+        GetId = 0xC008010E,
     };
-
     struct IocCreateParams {
         // Input
         u32_le size;
         // Output
         u32_le handle;
     };
+    static_assert(sizeof(IocCreateParams) == 8, "IocCreateParams has wrong size");
+
+    struct IocFromIdParams {
+        // Input
+        u32_le id;
+        // Output
+        u32_le handle;
+    };
+    static_assert(sizeof(IocFromIdParams) == 8, "IocFromIdParams has wrong size");
 
     struct IocAllocParams {
         // Input
@@ -79,6 +89,25 @@ private:
         INSERT_PADDING_BYTES(7);
         u64_le addr;
     };
+    static_assert(sizeof(IocAllocParams) == 32, "IocAllocParams has wrong size");
+
+    struct IocFreeParams {
+        u32_le handle;
+        INSERT_PADDING_BYTES(4);
+        u64_le refcount;
+        u32_le size;
+        u32_le flags;
+    };
+    static_assert(sizeof(IocFreeParams) == 24, "IocFreeParams has wrong size");
+
+    struct IocParamParams {
+        // Input
+        u32_le handle;
+        u32_le param;
+        // Output
+        u32_le result;
+    };
+    static_assert(sizeof(IocParamParams) == 12, "IocParamParams has wrong size");
 
     struct IocGetIdParams {
         // Output
@@ -86,27 +115,14 @@ private:
         // Input
         u32_le handle;
     };
-
-    struct IocFromIdParams {
-        // Input
-        u32_le id;
-        // Output
-        u32_le handle;
-    };
-
-    struct IocParamParams {
-        // Input
-        u32_le handle;
-        u32_le type;
-        // Output
-        u32_le value;
-    };
+    static_assert(sizeof(IocGetIdParams) == 8, "IocGetIdParams has wrong size");
 
     u32 IocCreate(const std::vector<u8>& input, std::vector<u8>& output);
     u32 IocAlloc(const std::vector<u8>& input, std::vector<u8>& output);
     u32 IocGetId(const std::vector<u8>& input, std::vector<u8>& output);
     u32 IocFromId(const std::vector<u8>& input, std::vector<u8>& output);
     u32 IocParam(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 IocFree(const std::vector<u8>& input, std::vector<u8>& output);
 };
 
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/nvdrv.cpp

@@ -9,6 +9,7 @@
 #include "core/hle/service/nvdrv/devices/nvhost_ctrl.h"
 #include "core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h"
 #include "core/hle/service/nvdrv/devices/nvhost_gpu.h"
+#include "core/hle/service/nvdrv/devices/nvhost_nvdec.h"
 #include "core/hle/service/nvdrv/devices/nvmap.h"
 #include "core/hle/service/nvdrv/interface.h"
 #include "core/hle/service/nvdrv/nvdrv.h"
@@ -36,6 +37,7 @@ Module::Module() {
     devices["/dev/nvmap"] = nvmap_dev;
     devices["/dev/nvdisp_disp0"] = std::make_shared<Devices::nvdisp_disp0>(nvmap_dev);
     devices["/dev/nvhost-ctrl"] = std::make_shared<Devices::nvhost_ctrl>();
+    devices["/dev/nvhost-nvdec"] = std::make_shared<Devices::nvhost_nvdec>();
 }
 
 u32 Module::Open(std::string device_name) {

src/core/hle/service/service.cpp

@@ -20,6 +20,7 @@
 #include "core/hle/service/aoc/aoc_u.h"
 #include "core/hle/service/apm/apm.h"
 #include "core/hle/service/audio/audio.h"
+#include "core/hle/service/bcat/bcat.h"
 #include "core/hle/service/fatal/fatal.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/hle/service/friend/friend.h"
@@ -144,10 +145,12 @@ ResultCode ServiceFrameworkBase::HandleSyncRequest(Kernel::HLERequestContext& co
         rb.Push(RESULT_SUCCESS);
         return ResultCode(ErrorModule::HIPC, ErrorDescription::RemoteProcessDead);
     }
+    case IPC::CommandType::ControlWithContext:
     case IPC::CommandType::Control: {
         Core::System::GetInstance().ServiceManager().InvokeControlRequest(context);
         break;
     }
+    case IPC::CommandType::RequestWithContext:
     case IPC::CommandType::Request: {
         InvokeRequest(context);
         break;
@@ -181,6 +184,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm) {
     AM::InstallInterfaces(*sm, nv_flinger);
     AOC::InstallInterfaces(*sm);
     APM::InstallInterfaces(*sm);
+    BCAT::InstallInterfaces(*sm);
     Audio::InstallInterfaces(*sm);
     Fatal::InstallInterfaces(*sm);
     FileSystem::InstallInterfaces(*sm);

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

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <chrono>
+#include <ctime>
 #include "common/logging/log.h"
 #include "core/core_timing.h"
 #include "core/hle/ipc_helpers.h"
@@ -77,7 +78,7 @@ public:
             {3, nullptr, "LoadLocationNameList"},
             {4, &ITimeZoneService::LoadTimeZoneRule, "LoadTimeZoneRule"},
             {5, nullptr, "GetTimeZoneRuleVersion"},
-            {100, nullptr, "ToCalendarTime"},
+            {100, &ITimeZoneService::ToCalendarTime, "ToCalendarTime"},
             {101, &ITimeZoneService::ToCalendarTimeWithMyRule, "ToCalendarTimeWithMyRule"},
             {200, nullptr, "ToPosixTime"},
             {201, nullptr, "ToPosixTimeWithMyRule"},
@@ -86,9 +87,11 @@ public:
     }
 
 private:
+    LocationName location_name{"UTC"};
+    TimeZoneRule my_time_zone_rule{};
+
     void GetDeviceLocationName(Kernel::HLERequestContext& ctx) {
-        NGLOG_WARNING(Service_Time, "(STUBBED) called");
-        LocationName location_name{};
+        NGLOG_DEBUG(Service_Time, "called");
         IPC::ResponseBuilder rb{ctx, (sizeof(LocationName) / 4) + 2};
         rb.Push(RESULT_SUCCESS);
         rb.PushRaw(location_name);
@@ -103,23 +106,70 @@ private:
 
     void LoadTimeZoneRule(Kernel::HLERequestContext& ctx) {
         NGLOG_WARNING(Service_Time, "(STUBBED) called");
+
+        ctx.WriteBuffer(&my_time_zone_rule, sizeof(TimeZoneRule));
+
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
     }
 
+    void ToCalendarTime(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const u64 posix_time = rp.Pop<u64>();
+
+        NGLOG_WARNING(Service_Time, "(STUBBED) called, posix_time=0x{:016X}", posix_time);
+
+        TimeZoneRule time_zone_rule{};
+        auto buffer = ctx.ReadBuffer();
+        std::memcpy(&time_zone_rule, buffer.data(), buffer.size());
+
+        CalendarTime calendar_time{2018, 1, 1, 0, 0, 0};
+        CalendarAdditionalInfo additional_info{};
+
+        PosixToCalendar(posix_time, calendar_time, additional_info, time_zone_rule);
+
+        IPC::ResponseBuilder rb{ctx, 10};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw(calendar_time);
+        rb.PushRaw(additional_info);
+    }
+
     void ToCalendarTimeWithMyRule(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
-        u64 posix_time = rp.Pop<u64>();
+        const u64 posix_time = rp.Pop<u64>();
 
         NGLOG_WARNING(Service_Time, "(STUBBED) called, posix_time=0x{:016X}", posix_time);
 
         CalendarTime calendar_time{2018, 1, 1, 0, 0, 0};
         CalendarAdditionalInfo additional_info{};
+
+        PosixToCalendar(posix_time, calendar_time, additional_info, my_time_zone_rule);
+
         IPC::ResponseBuilder rb{ctx, 10};
         rb.Push(RESULT_SUCCESS);
         rb.PushRaw(calendar_time);
         rb.PushRaw(additional_info);
     }
+
+    void PosixToCalendar(u64 posix_time, CalendarTime& calendar_time,
+                         CalendarAdditionalInfo& additional_info, const TimeZoneRule& /*rule*/) {
+        std::time_t t(posix_time);
+        std::tm* tm = std::localtime(&t);
+        if (!tm) {
+            return;
+        }
+        calendar_time.year = tm->tm_year + 1900;
+        calendar_time.month = tm->tm_mon + 1;
+        calendar_time.day = tm->tm_mday;
+        calendar_time.hour = tm->tm_hour;
+        calendar_time.minute = tm->tm_min;
+        calendar_time.second = tm->tm_sec;
+
+        additional_info.day_of_week = tm->tm_wday;
+        additional_info.day_of_year = tm->tm_yday;
+        std::memcpy(additional_info.name.data(), "UTC", sizeof("UTC"));
+        additional_info.utc_offset = 0;
+    }
 };
 
 void Module::Interface::GetStandardUserSystemClock(Kernel::HLERequestContext& ctx) {

src/core/hle/service/time/time.h

@@ -4,13 +4,13 @@
 
 #pragma once
 
+#include <array>
 #include "core/hle/service/service.h"
 
 namespace Service::Time {
 
-// TODO(Rozelette) RE this structure
 struct LocationName {
-    INSERT_PADDING_BYTES(0x24);
+    std::array<u8, 0x24> name;
 };
 static_assert(sizeof(LocationName) == 0x24, "LocationName is incorrect size");
 
@@ -25,26 +25,34 @@ struct CalendarTime {
 };
 static_assert(sizeof(CalendarTime) == 0x8, "CalendarTime structure has incorrect size");
 
-// TODO(Rozelette) RE this structure
 struct CalendarAdditionalInfo {
-    INSERT_PADDING_BYTES(0x18);
+    u32_le day_of_week;
+    u32_le day_of_year;
+    std::array<u8, 8> name;
+    INSERT_PADDING_BYTES(1);
+    s32_le utc_offset;
 };
 static_assert(sizeof(CalendarAdditionalInfo) == 0x18,
               "CalendarAdditionalInfo structure has incorrect size");
 
-// TODO(bunnei) RE this structure
-struct SystemClockContext {
-    INSERT_PADDING_BYTES(0x20);
+// TODO(mailwl) RE this structure
+struct TimeZoneRule {
+    INSERT_PADDING_BYTES(0x4000);
 };
-static_assert(sizeof(SystemClockContext) == 0x20,
-              "SystemClockContext structure has incorrect size");
 
 struct SteadyClockTimePoint {
-    u64 value;
+    u64_le value;
     INSERT_PADDING_WORDS(4);
 };
 static_assert(sizeof(SteadyClockTimePoint) == 0x18, "SteadyClockTimePoint is incorrect size");
 
+struct SystemClockContext {
+    u64_le offset;
+    SteadyClockTimePoint time_point;
+};
+static_assert(sizeof(SystemClockContext) == 0x20,
+              "SystemClockContext structure has incorrect size");
+
 class Module final {
 public:
     class Interface : public ServiceFramework<Interface> {

src/core/memory.cpp

@@ -28,8 +28,13 @@ static PageTable* current_page_table = nullptr;
 
 void SetCurrentPageTable(PageTable* page_table) {
     current_page_table = page_table;
-    if (Core::System::GetInstance().IsPoweredOn()) {
-        Core::CPU().PageTableChanged();
+
+    auto& system = Core::System::GetInstance();
+    if (system.IsPoweredOn()) {
+        system.ArmInterface(0).PageTableChanged();
+        system.ArmInterface(1).PageTableChanged();
+        system.ArmInterface(2).PageTableChanged();
+        system.ArmInterface(3).PageTableChanged();
     }
 }
 

src/core/settings.h

@@ -121,6 +121,7 @@ struct Values {
 
     // Core
     bool use_cpu_jit;
+    bool use_multi_core;
 
     // Data Storage
     bool use_virtual_sd;

src/core/telemetry_session.cpp

@@ -155,6 +155,8 @@ TelemetrySession::TelemetrySession() {
 
     // Log user configuration information
     AddField(Telemetry::FieldType::UserConfig, "Core_UseCpuJit", Settings::values.use_cpu_jit);
+    AddField(Telemetry::FieldType::UserConfig, "Core_UseMultiCore",
+             Settings::values.use_multi_core);
     AddField(Telemetry::FieldType::UserConfig, "Renderer_ResolutionFactor",
              Settings::values.resolution_factor);
     AddField(Telemetry::FieldType::UserConfig, "Renderer_ToggleFramelimit",

src/video_core/command_processor.cpp

@@ -36,7 +36,6 @@ void GPU::WriteReg(u32 method, u32 subchannel, u32 value, u32 remaining_params)
     if (method == static_cast<u32>(BufferMethods::BindObject)) {
         // Bind the current subchannel to the desired engine id.
         NGLOG_DEBUG(HW_GPU, "Binding subchannel {} to engine {}", subchannel, value);
-        ASSERT(bound_engines.find(subchannel) == bound_engines.end());
         bound_engines[subchannel] = static_cast<EngineID>(value);
         return;
     }

src/video_core/engines/shader_bytecode.h

@@ -1,4 +1,4 @@
-// Copyright 2018 yuzu Emulator Project
+// Copyright 2018 yuzu Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
@@ -75,6 +75,10 @@ union Attribute {
     enum class Index : u64 {
         Position = 7,
         Attribute_0 = 8,
+        // This attribute contains a tuple of (~, ~, InstanceId, VertexId) when inside a vertex
+        // shader, and a tuple of (TessCoord.x, TessCoord.y, TessCoord.z, ~) when inside a Tess Eval
+        // shader.
+        TessCoordInstanceIDVertexID = 47,
     };
 
     union {
@@ -193,6 +197,11 @@ union Instruction {
         BitField<50, 1, u64> abs_d;
         BitField<56, 1, u64> negate_imm;
 
+        union {
+            BitField<39, 3, u64> pred;
+            BitField<42, 1, u64> negate_pred;
+        } fmnmx;
+
         float GetImm20_19() const {
             float result{};
             u32 imm{static_cast<u32>(imm20_19)};
@@ -252,6 +261,19 @@ union Instruction {
         BitField<50, 1, u64> saturate_a;
     } conversion;
 
+    union {
+        BitField<31, 4, u64> component_mask;
+
+        bool IsComponentEnabled(size_t component) const {
+            return ((1 << component) & component_mask) != 0;
+        }
+    } tex;
+
+    union {
+        // TODO(bunnei): This is just a guess, needs to be verified
+        BitField<52, 1, u64> enable_g_component;
+    } texs;
+
     BitField<61, 1, u64> is_b_imm;
     BitField<60, 1, u64> is_b_gpr;
     BitField<59, 1, u64> is_c_gpr;
@@ -272,6 +294,7 @@ public:
         KIL,
         LD_A,
         ST_A,
+        TEX,
         TEXQ, // Texture Query
         TEXS, // Texture Fetch with scalar/non-vec4 source/destinations
         TLDS, // Texture Load with scalar/non-vec4 source/destinations
@@ -288,8 +311,10 @@ public:
         FMUL_R,
         FMUL_IMM,
         FMUL32_IMM,
-        MUFU, // Multi-Function Operator
-        RRO,  // Range Reduction Operator
+        MUFU,  // Multi-Function Operator
+        RRO_C, // Range Reduction Operator
+        RRO_R,
+        RRO_IMM,
         F2F_C,
         F2F_R,
         F2F_IMM,
@@ -310,7 +335,9 @@ public:
         SHR_C,
         SHR_R,
         SHR_IMM,
-        FMNMX,
+        FMNMX_C,
+        FMNMX_R,
+        FMNMX_IMM,
         FSETP_C, // Set Predicate
         FSETP_R,
         FSETP_IMM,
@@ -320,6 +347,7 @@ public:
         ISETP_C,
         ISETP_IMM,
         ISETP_R,
+        PSETP,
     };
 
     enum class Type {
@@ -331,6 +359,7 @@ public:
         FloatSet,
         FloatSetPredicate,
         IntegerSetPredicate,
+        PredicateSetPredicate,
         Conversion,
         Unknown,
     };
@@ -429,6 +458,7 @@ private:
             INST("111000110011----", Id::KIL, Type::Flow, "KIL"),
             INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
             INST("1110111111110---", Id::ST_A, Type::Memory, "ST_A"),
+            INST("1100000000111---", Id::TEX, Type::Memory, "TEX"),
             INST("1101111101001---", Id::TEXQ, Type::Memory, "TEXQ"),
             INST("1101100---------", Id::TEXS, Type::Memory, "TEXS"),
             INST("1101101---------", Id::TLDS, Type::Memory, "TLDS"),
@@ -446,10 +476,12 @@ private:
             INST("0011100-01101---", Id::FMUL_IMM, Type::Arithmetic, "FMUL_IMM"),
             INST("00011110--------", Id::FMUL32_IMM, Type::Arithmetic, "FMUL32_IMM"),
             INST("0101000010000---", Id::MUFU, Type::Arithmetic, "MUFU"),
-            INST("0101110010010---", Id::RRO, Type::Arithmetic, "RRO"),
-            INST("0100110010101---", Id::F2F_C, Type::Arithmetic, "F2F_C"),
-            INST("0101110010101---", Id::F2F_R, Type::Arithmetic, "F2F_R"),
-            INST("0011100-10101---", Id::F2F_IMM, Type::Arithmetic, "F2F_IMM"),
+            INST("0100110010010---", Id::RRO_C, Type::Arithmetic, "RRO_C"),
+            INST("0101110010010---", Id::RRO_R, Type::Arithmetic, "RRO_R"),
+            INST("0011100-10010---", Id::RRO_IMM, Type::Arithmetic, "RRO_IMM"),
+            INST("0100110010101---", Id::F2F_C, Type::Conversion, "F2F_C"),
+            INST("0101110010101---", Id::F2F_R, Type::Conversion, "F2F_R"),
+            INST("0011100-10101---", Id::F2F_IMM, Type::Conversion, "F2F_IMM"),
             INST("0100110010110---", Id::F2I_C, Type::Arithmetic, "F2I_C"),
             INST("0101110010110---", Id::F2I_R, Type::Arithmetic, "F2I_R"),
             INST("0011100-10110---", Id::F2I_IMM, Type::Arithmetic, "F2I_IMM"),
@@ -461,7 +493,9 @@ private:
             INST("0100110000101---", Id::SHR_C, Type::Arithmetic, "SHR_C"),
             INST("0101110000101---", Id::SHR_R, Type::Arithmetic, "SHR_R"),
             INST("0011100-00101---", Id::SHR_IMM, Type::Arithmetic, "SHR_IMM"),
-            INST("0101110001100---", Id::FMNMX, Type::Arithmetic, "FMNMX"),
+            INST("0100110001100---", Id::FMNMX_C, Type::Arithmetic, "FMNMX_C"),
+            INST("0101110001100---", Id::FMNMX_R, Type::Arithmetic, "FMNMX_R"),
+            INST("0011100-01100---", Id::FMNMX_IMM, Type::Arithmetic, "FMNMX_IMM"),
             INST("0100110011100---", Id::I2I_C, Type::Conversion, "I2I_C"),
             INST("0101110011100---", Id::I2I_R, Type::Conversion, "I2I_R"),
             INST("01110001-1000---", Id::I2I_IMM, Type::Conversion, "I2I_IMM"),
@@ -477,6 +511,7 @@ private:
             INST("010010110110----", Id::ISETP_C, Type::IntegerSetPredicate, "ISETP_C"),
             INST("010110110110----", Id::ISETP_R, Type::IntegerSetPredicate, "ISETP_R"),
             INST("0011011-0110----", Id::ISETP_IMM, Type::IntegerSetPredicate, "ISETP_IMM"),
+            INST("0101000010010---", Id::PSETP, Type::PredicateSetPredicate, "PSETP"),
         };
 #undef INST
         std::stable_sort(table.begin(), table.end(), [](const auto& a, const auto& b) {

src/video_core/memory_manager.cpp

@@ -58,6 +58,25 @@ GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size)
     return gpu_addr;
 }
 
+GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) {
+    ASSERT((gpu_addr & PAGE_MASK) == 0);
+
+    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
+        ASSERT(PageSlot(gpu_addr + offset) != static_cast<u64>(PageStatus::Allocated) &&
+               PageSlot(gpu_addr + offset) != static_cast<u64>(PageStatus::Unmapped));
+        PageSlot(gpu_addr + offset) = static_cast<u64>(PageStatus::Unmapped);
+    }
+
+    // Delete the region mappings that are contained within the unmapped region
+    mapped_regions.erase(std::remove_if(mapped_regions.begin(), mapped_regions.end(),
+                                        [&](const MappedRegion& region) {
+                                            return region.gpu_addr <= gpu_addr &&
+                                                   region.gpu_addr + region.size < gpu_addr + size;
+                                        }),
+                         mapped_regions.end());
+    return gpu_addr;
+}
+
 boost::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
     GPUVAddr gpu_addr = 0;
     u64 free_space = 0;

src/video_core/memory_manager.h

@@ -25,6 +25,7 @@ public:
     GPUVAddr AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align);
     GPUVAddr MapBufferEx(VAddr cpu_addr, u64 size);
     GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size);
+    GPUVAddr UnmapBuffer(GPUVAddr gpu_addr, u64 size);
     boost::optional<VAddr> GpuToCpuAddress(GPUVAddr gpu_addr);
     std::vector<GPUVAddr> CpuToGpuAddress(VAddr cpu_addr) const;
 

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -75,14 +75,11 @@ RasterizerOpenGL::RasterizerOpenGL() {
     // Clipping plane 0 is always enabled for PICA fixed clip plane z <= 0
     state.clip_distance[0] = true;
 
-    // Generate VBO, VAO and UBO
-    vertex_buffer = OGLStreamBuffer::MakeBuffer(GLAD_GL_ARB_buffer_storage, GL_ARRAY_BUFFER);
-    vertex_buffer->Create(VERTEX_BUFFER_SIZE, VERTEX_BUFFER_SIZE / 2);
+    // Generate VAO and UBO
     sw_vao.Create();
     uniform_buffer.Create();
 
     state.draw.vertex_array = sw_vao.handle;
-    state.draw.vertex_buffer = vertex_buffer->GetHandle();
     state.draw.uniform_buffer = uniform_buffer.handle;
     state.Apply();
 
@@ -90,7 +87,6 @@ RasterizerOpenGL::RasterizerOpenGL() {
     framebuffer.Create();
 
     hw_vao.Create();
-    hw_vao_enabled_attributes.fill(false);
 
     stream_buffer = OGLStreamBuffer::MakeBuffer(has_ARB_buffer_storage, GL_ARRAY_BUFFER);
     stream_buffer->Create(STREAM_BUFFER_SIZE, STREAM_BUFFER_SIZE / 2);
@@ -181,8 +177,6 @@ std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr,
         glVertexAttribFormat(index, attrib.ComponentCount(), MaxwellToGL::VertexType(attrib),
                              attrib.IsNormalized() ? GL_TRUE : GL_FALSE, attrib.offset);
         glVertexAttribBinding(index, attrib.buffer);
-
-        hw_vao_enabled_attributes[index] = true;
     }
 
     return {array_ptr, buffer_offset};

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -134,21 +134,17 @@ private:
     std::unique_ptr<GLShader::ProgramManager> shader_program_manager;
     OGLVertexArray sw_vao;
     OGLVertexArray hw_vao;
-    std::array<bool, 16> hw_vao_enabled_attributes;
 
     std::array<SamplerInfo, GLShader::NumTextureSamplers> texture_samplers;
     std::array<std::array<OGLBuffer, Tegra::Engines::Maxwell3D::Regs::MaxConstBuffers>,
                Tegra::Engines::Maxwell3D::Regs::MaxShaderStage>
         ssbos;
 
-    static constexpr size_t VERTEX_BUFFER_SIZE = 128 * 1024 * 1024;
-    std::unique_ptr<OGLStreamBuffer> vertex_buffer;
+    static constexpr size_t STREAM_BUFFER_SIZE = 128 * 1024 * 1024;
+    std::unique_ptr<OGLStreamBuffer> stream_buffer;
     OGLBuffer uniform_buffer;
     OGLFramebuffer framebuffer;
 
-    static constexpr size_t STREAM_BUFFER_SIZE = 4 * 1024 * 1024;
-    std::unique_ptr<OGLStreamBuffer> stream_buffer;
-
     size_t CalculateVertexArraysSize() const;
 
     std::pair<u8*, GLintptr> SetupVertexArrays(u8* array_ptr, GLintptr buffer_offset);

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -45,19 +45,23 @@ struct FormatTuple {
 
 static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{
     {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, false},                    // ABGR8
-    {GL_RGB, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, false},                       // B5G6R5
+    {GL_RGB, GL_RGB, GL_UNSIGNED_SHORT_5_6_5, false},                           // B5G6R5
     {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV, false},              // A2B10G10R10
+    {GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_1_5_5_5_REV, false},                // A1B5G5R5
+    {GL_R8, GL_RED, GL_UNSIGNED_BYTE, false},                                   // R8
+    {GL_RGBA16F, GL_RGBA, GL_HALF_FLOAT, false},                                // RGBA16F
     {GL_COMPRESSED_RGB_S3TC_DXT1_EXT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, true},   // DXT1
     {GL_COMPRESSED_RGBA_S3TC_DXT3_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, true}, // DXT23
     {GL_COMPRESSED_RGBA_S3TC_DXT5_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, true}, // DXT45
+    {GL_COMPRESSED_RED_RGTC1, GL_RED, GL_UNSIGNED_INT_8_8_8_8, true},           // DXN1
 }};
 
 static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType component_type) {
     const SurfaceType type = SurfaceParams::GetFormatType(pixel_format);
     if (type == SurfaceType::ColorTexture) {
         ASSERT(static_cast<size_t>(pixel_format) < tex_format_tuples.size());
-        // For now only UNORM components are supported
-        ASSERT(component_type == ComponentType::UNorm);
+        // For now only UNORM components are supported, or RGBA16F which is type FLOAT
+        ASSERT(component_type == ComponentType::UNorm || pixel_format == PixelFormat::RGBA16F);
         return tex_format_tuples[static_cast<unsigned int>(pixel_format)];
     } else if (type == SurfaceType::Depth || type == SurfaceType::DepthStencil) {
         // TODO(Subv): Implement depth formats
@@ -107,8 +111,10 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, Tegra::GPUVAddr, Tegra:
                             SurfaceParams::MaxPixelFormat>
     morton_to_gl_fns = {
         MortonCopy<true, PixelFormat::ABGR8>,       MortonCopy<true, PixelFormat::B5G6R5>,
-        MortonCopy<true, PixelFormat::A2B10G10R10>, MortonCopy<true, PixelFormat::DXT1>,
-        MortonCopy<true, PixelFormat::DXT23>,       MortonCopy<true, PixelFormat::DXT45>,
+        MortonCopy<true, PixelFormat::A2B10G10R10>, MortonCopy<true, PixelFormat::A1B5G5R5>,
+        MortonCopy<true, PixelFormat::R8>,          MortonCopy<true, PixelFormat::RGBA16F>,
+        MortonCopy<true, PixelFormat::DXT1>,        MortonCopy<true, PixelFormat::DXT23>,
+        MortonCopy<true, PixelFormat::DXT45>,       MortonCopy<true, PixelFormat::DXN1>,
 };
 
 static constexpr std::array<void (*)(u32, u32, u32, u8*, Tegra::GPUVAddr, Tegra::GPUVAddr,
@@ -118,7 +124,11 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, Tegra::GPUVAddr, Tegra:
         MortonCopy<false, PixelFormat::ABGR8>,
         MortonCopy<false, PixelFormat::B5G6R5>,
         MortonCopy<false, PixelFormat::A2B10G10R10>,
-        // TODO(Subv): Swizzling the DXT1/DXT23/DXT45 formats is not yet supported
+        MortonCopy<false, PixelFormat::A1B5G5R5>,
+        MortonCopy<false, PixelFormat::R8>,
+        MortonCopy<false, PixelFormat::RGBA16F>,
+        // TODO(Subv): Swizzling the DXT1/DXT23/DXT45/DXN1 formats is not yet supported
+        nullptr,
         nullptr,
         nullptr,
         nullptr,

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -57,9 +57,13 @@ struct SurfaceParams {
         ABGR8 = 0,
         B5G6R5 = 1,
         A2B10G10R10 = 2,
-        DXT1 = 3,
-        DXT23 = 4,
-        DXT45 = 5,
+        A1B5G5R5 = 3,
+        R8 = 4,
+        RGBA16F = 5,
+        DXT1 = 6,
+        DXT23 = 7,
+        DXT45 = 8,
+        DXN1 = 9, // This is also known as BC4
 
         Max,
         Invalid = 255,
@@ -98,9 +102,13 @@ struct SurfaceParams {
             1, // ABGR8
             1, // B5G6R5
             1, // A2B10G10R10
+            1, // A1B5G5R5
+            1, // R8
+            2, // RGBA16F
             4, // DXT1
             4, // DXT23
             4, // DXT45
+            4, // DXN1
         }};
 
         ASSERT(static_cast<size_t>(format) < compression_factor_table.size());
@@ -118,9 +126,13 @@ struct SurfaceParams {
             32,  // ABGR8
             16,  // B5G6R5
             32,  // A2B10G10R10
+            16,  // A1B5G5R5
+            8,   // R8
+            64,  // RGBA16F
             64,  // DXT1
             128, // DXT23
             128, // DXT45
+            64,  // DXN1
         }};
 
         ASSERT(static_cast<size_t>(format) < bpp_table.size());
@@ -133,9 +145,12 @@ struct SurfaceParams {
     static PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
         switch (format) {
         case Tegra::RenderTargetFormat::RGBA8_UNORM:
+        case Tegra::RenderTargetFormat::RGBA8_SRGB:
             return PixelFormat::ABGR8;
         case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
             return PixelFormat::A2B10G10R10;
+        case Tegra::RenderTargetFormat::RGBA16_FLOAT:
+            return PixelFormat::RGBA16F;
         default:
             NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
@@ -161,12 +176,20 @@ struct SurfaceParams {
             return PixelFormat::B5G6R5;
         case Tegra::Texture::TextureFormat::A2B10G10R10:
             return PixelFormat::A2B10G10R10;
+        case Tegra::Texture::TextureFormat::A1B5G5R5:
+            return PixelFormat::A1B5G5R5;
+        case Tegra::Texture::TextureFormat::R8:
+            return PixelFormat::R8;
+        case Tegra::Texture::TextureFormat::R16_G16_B16_A16:
+            return PixelFormat::RGBA16F;
         case Tegra::Texture::TextureFormat::DXT1:
             return PixelFormat::DXT1;
         case Tegra::Texture::TextureFormat::DXT23:
             return PixelFormat::DXT23;
         case Tegra::Texture::TextureFormat::DXT45:
             return PixelFormat::DXT45;
+        case Tegra::Texture::TextureFormat::DXN1:
+            return PixelFormat::DXN1;
         default:
             NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
@@ -182,12 +205,20 @@ struct SurfaceParams {
             return Tegra::Texture::TextureFormat::B5G6R5;
         case PixelFormat::A2B10G10R10:
             return Tegra::Texture::TextureFormat::A2B10G10R10;
+        case PixelFormat::A1B5G5R5:
+            return Tegra::Texture::TextureFormat::A1B5G5R5;
+        case PixelFormat::R8:
+            return Tegra::Texture::TextureFormat::R8;
+        case PixelFormat::RGBA16F:
+            return Tegra::Texture::TextureFormat::R16_G16_B16_A16;
         case PixelFormat::DXT1:
             return Tegra::Texture::TextureFormat::DXT1;
         case PixelFormat::DXT23:
             return Tegra::Texture::TextureFormat::DXT23;
         case PixelFormat::DXT45:
             return Tegra::Texture::TextureFormat::DXT45;
+        case PixelFormat::DXN1:
+            return Tegra::Texture::TextureFormat::DXN1;
         default:
             UNREACHABLE();
         }
@@ -208,8 +239,11 @@ struct SurfaceParams {
         // TODO(Subv): Implement more render targets
         switch (format) {
         case Tegra::RenderTargetFormat::RGBA8_UNORM:
+        case Tegra::RenderTargetFormat::RGBA8_SRGB:
         case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
             return ComponentType::UNorm;
+        case Tegra::RenderTargetFormat::RGBA16_FLOAT:
+            return ComponentType::Float;
         default:
             NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -299,7 +299,7 @@ public:
      * are stored as floats, so this may require conversion.
      * @param reg The destination register to use.
      * @param elem The element to use for the operation.
-     * @param attribute The input attibute to use as the source value.
+     * @param attribute The input attribute to use as the source value.
      */
     void SetRegisterToInputAttibute(const Register& reg, u64 elem, Attribute::Index attribute) {
         std::string dest = GetRegisterAsFloat(reg);
@@ -451,6 +451,12 @@ private:
         switch (attribute) {
         case Attribute::Index::Position:
             return "position";
+        case Attribute::Index::TessCoordInstanceIDVertexID:
+            // TODO(Subv): Find out what the values are for the first two elements when inside a
+            // vertex shader, and what's the value of the fourth element when inside a Tess Eval
+            // shader.
+            ASSERT(stage == Maxwell3D::Regs::ShaderStage::Vertex);
+            return "vec4(0, 0, gl_InstanceID, gl_VertexID)";
         default:
             const u32 index{static_cast<u32>(attribute) -
                             static_cast<u32>(Attribute::Index::Attribute_0)};
@@ -580,20 +586,62 @@ private:
      * @param instr Instruction to generate the if condition for.
      * @returns string containing the predicate condition.
      */
-    std::string GetPredicateCondition(Instruction instr) const {
+    std::string GetPredicateCondition(u64 index, bool negate) const {
         using Tegra::Shader::Pred;
-        ASSERT(instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex));
+        std::string variable;
 
-        std::string variable =
-            'p' + std::to_string(static_cast<u64>(instr.pred.pred_index.Value()));
+        // Index 7 is used as an 'Always True' condition.
+        if (index == static_cast<u64>(Pred::UnusedIndex))
+            variable = "true";
+        else
+            variable = 'p' + std::to_string(index);
 
-        if (instr.negate_pred) {
+        if (negate) {
             return "!(" + variable + ')';
         }
 
         return variable;
     }
 
+    /**
+     * Returns the comparison string to use to compare two values in the 'set' family of
+     * instructions.
+     * @params condition The condition used in the 'set'-family instruction.
+     * @returns String corresponding to the GLSL operator that matches the desired comparison.
+     */
+    std::string GetPredicateComparison(Tegra::Shader::PredCondition condition) const {
+        using Tegra::Shader::PredCondition;
+        static const std::unordered_map<PredCondition, const char*> PredicateComparisonStrings = {
+            {PredCondition::LessThan, "<"},      {PredCondition::Equal, "=="},
+            {PredCondition::LessEqual, "<="},    {PredCondition::GreaterThan, ">"},
+            {PredCondition::GreaterEqual, ">="},
+        };
+
+        auto comparison = PredicateComparisonStrings.find(condition);
+        ASSERT_MSG(comparison != PredicateComparisonStrings.end(),
+                   "Unknown predicate comparison operation");
+        return comparison->second;
+    }
+
+    /**
+     * Returns the operator string to use to combine two predicates in the 'setp' family of
+     * instructions.
+     * @params operation The operator used in the 'setp'-family instruction.
+     * @returns String corresponding to the GLSL operator that matches the desired operator.
+     */
+    std::string GetPredicateCombiner(Tegra::Shader::PredOperation operation) const {
+        using Tegra::Shader::PredOperation;
+        static const std::unordered_map<PredOperation, const char*> PredicateOperationStrings = {
+            {PredOperation::And, "&&"},
+            {PredOperation::Or, "||"},
+            {PredOperation::Xor, "^^"},
+        };
+
+        auto op = PredicateOperationStrings.find(operation);
+        ASSERT_MSG(op != PredicateOperationStrings.end(), "Unknown predicate operation");
+        return op->second;
+    }
+
     /*
      * Returns whether the instruction at the specified offset is a 'sched' instruction.
      * Sched instructions always appear before a sequence of 3 instructions.
@@ -634,7 +682,9 @@ private:
                    "NeverExecute predicate not implemented");
 
         if (instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex)) {
-            shader.AddLine("if (" + GetPredicateCondition(instr) + ')');
+            shader.AddLine("if (" +
+                           GetPredicateCondition(instr.pred.pred_index, instr.negate_pred != 0) +
+                           ')');
             shader.AddLine('{');
             ++shader.scope;
         }
@@ -730,8 +780,25 @@ private:
                 }
                 break;
             }
-            case OpCode::Id::RRO: {
-                NGLOG_DEBUG(HW_GPU, "Skipping RRO instruction");
+            case OpCode::Id::FMNMX_C:
+            case OpCode::Id::FMNMX_R:
+            case OpCode::Id::FMNMX_IMM: {
+                std::string condition =
+                    GetPredicateCondition(instr.alu.fmnmx.pred, instr.alu.fmnmx.negate_pred != 0);
+                std::string parameters = op_a + ',' + op_b;
+                regs.SetRegisterToFloat(instr.gpr0, 0,
+                                        '(' + condition + ") ? min(" + parameters + ") : max(" +
+                                            parameters + ')',
+                                        1, 1);
+                break;
+            }
+            case OpCode::Id::RRO_C:
+            case OpCode::Id::RRO_R:
+            case OpCode::Id::RRO_IMM: {
+                // Currently RRO is only implemented as a register move.
+                // Usage of `abs_b` and `negate_b` here should also be correct.
+                regs.SetRegisterToFloat(instr.gpr0, 0, op_b, 1, 1);
+                NGLOG_WARNING(HW_GPU, "RRO instruction is incomplete");
                 break;
             }
             default: {
@@ -778,13 +845,14 @@ private:
         }
         case OpCode::Type::Conversion: {
             ASSERT_MSG(instr.conversion.size == Register::Size::Word, "Unimplemented");
-            ASSERT_MSG(!instr.conversion.selector, "Unimplemented");
             ASSERT_MSG(!instr.conversion.negate_a, "Unimplemented");
             ASSERT_MSG(!instr.conversion.saturate_a, "Unimplemented");
 
             switch (opcode->GetId()) {
             case OpCode::Id::I2I_R:
             case OpCode::Id::I2F_R: {
+                ASSERT_MSG(!instr.conversion.selector, "Unimplemented");
+
                 std::string op_a =
                     regs.GetRegisterAsInteger(instr.gpr20, 0, instr.conversion.is_signed);
 
@@ -795,6 +863,16 @@ private:
                 regs.SetRegisterToInteger(instr.gpr0, instr.conversion.is_signed, 0, op_a, 1, 1);
                 break;
             }
+            case OpCode::Id::F2F_R: {
+                std::string op_a = regs.GetRegisterAsFloat(instr.gpr20);
+
+                if (instr.conversion.abs_a) {
+                    op_a = "abs(" + op_a + ')';
+                }
+
+                regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
+                break;
+            }
             default: {
                 NGLOG_CRITICAL(HW_GPU, "Unhandled conversion instruction: {}", opcode->GetName());
                 UNREACHABLE();
@@ -818,10 +896,10 @@ private:
                                                   instr.gpr0);
                 break;
             }
-            case OpCode::Id::TEXS: {
+            case OpCode::Id::TEX: {
                 ASSERT_MSG(instr.attribute.fmt20.size == 4, "untested");
                 const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20);
+                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
                 const std::string sampler = GetSampler(instr.sampler);
                 const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
                 // Add an extra scope and declare the texture coords inside to prevent overwriting
@@ -830,8 +908,49 @@ private:
                 ++shader.scope;
                 shader.AddLine(coord);
                 const std::string texture = "texture(" + sampler + ", coords)";
-                for (unsigned elem = 0; elem < instr.attribute.fmt20.size; ++elem) {
-                    regs.SetRegisterToFloat(instr.gpr0, elem, texture, 1, 4, false, elem);
+
+                size_t dest_elem{};
+                for (size_t elem = 0; elem < instr.attribute.fmt20.size; ++elem) {
+                    if (!instr.tex.IsComponentEnabled(elem)) {
+                        // Skip disabled components
+                        continue;
+                    }
+                    regs.SetRegisterToFloat(instr.gpr0, elem, texture, 1, 4, false, dest_elem);
+                    ++dest_elem;
+                }
+                --shader.scope;
+                shader.AddLine("}");
+                break;
+            }
+            case OpCode::Id::TEXS: {
+                ASSERT_MSG(instr.attribute.fmt20.size == 4, "untested");
+                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
+                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20);
+                const std::string sampler = GetSampler(instr.sampler);
+                const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                // Add an extra scope and declare the texture coords inside to prevent
+                // overwriting them in case they are used as outputs of the texs instruction.
+                shader.AddLine("{");
+                ++shader.scope;
+                shader.AddLine(coord);
+                const std::string texture = "texture(" + sampler + ", coords)";
+
+                // TEXS has two destination registers. RG goes into gpr0+0 and gpr0+1, and BA goes
+                // into gpr28+0 and gpr28+1
+                size_t offset{};
+                for (const auto& dest : {instr.gpr0.Value(), instr.gpr28.Value()}) {
+                    for (unsigned elem = 0; elem < 2; ++elem) {
+                        if (dest + elem >= Register::ZeroIndex) {
+                            // Skip invalid register values
+                            break;
+                        }
+                        regs.SetRegisterToFloat(dest, elem + offset, texture, 1, 4, false, elem);
+                        if (!instr.texs.enable_g_component) {
+                            // Skip the second component
+                            break;
+                        }
+                    }
+                    offset += 2;
                 }
                 --shader.scope;
                 shader.AddLine("}");
@@ -873,28 +992,25 @@ private:
             }
 
             using Tegra::Shader::Pred;
-            ASSERT_MSG(instr.fsetp.pred0 == static_cast<u64>(Pred::UnusedIndex) &&
-                           instr.fsetp.pred39 == static_cast<u64>(Pred::UnusedIndex),
-                       "Compound predicates are not implemented");
-
             // We can't use the constant predicate as destination.
             ASSERT(instr.fsetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
 
-            using Tegra::Shader::PredCondition;
-            switch (instr.fsetp.cond) {
-            case PredCondition::LessThan:
-                SetPredicate(instr.fsetp.pred3, '(' + op_a + ") < (" + op_b + ')');
-                break;
-            case PredCondition::Equal:
-                SetPredicate(instr.fsetp.pred3, '(' + op_a + ") == (" + op_b + ')');
-                break;
-            case PredCondition::LessEqual:
-                SetPredicate(instr.fsetp.pred3, '(' + op_a + ") <= (" + op_b + ')');
-                break;
-            default:
-                NGLOG_CRITICAL(HW_GPU, "Unhandled predicate condition: {} (a: {}, b: {})",
-                               static_cast<unsigned>(instr.fsetp.cond.Value()), op_a, op_b);
-                UNREACHABLE();
+            std::string second_pred =
+                GetPredicateCondition(instr.fsetp.pred39, instr.fsetp.neg_pred != 0);
+
+            std::string comparator = GetPredicateComparison(instr.fsetp.cond);
+            std::string combiner = GetPredicateCombiner(instr.fsetp.op);
+
+            std::string predicate = '(' + op_a + ") " + comparator + " (" + op_b + ')';
+            // Set the primary predicate to the result of Predicate OP SecondPredicate
+            SetPredicate(instr.fsetp.pred3,
+                         '(' + predicate + ") " + combiner + " (" + second_pred + ')');
+
+            if (instr.fsetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) {
+                // Set the secondary predicate to the result of !Predicate OP SecondPredicate,
+                // if enabled
+                SetPredicate(instr.fsetp.pred0,
+                             "!(" + predicate + ") " + combiner + " (" + second_pred + ')');
             }
             break;
         }
@@ -926,35 +1042,18 @@ private:
                 op_b = "abs(" + op_b + ')';
             }
 
-            using Tegra::Shader::Pred;
-            ASSERT_MSG(instr.fset.pred39 == static_cast<u64>(Pred::UnusedIndex),
-                       "Compound predicates are not implemented");
-
             // The fset instruction sets a register to 1.0 if the condition is true, and to 0
             // otherwise.
-            using Tegra::Shader::PredCondition;
-            switch (instr.fset.cond) {
-            case PredCondition::LessThan:
-                regs.SetRegisterToFloat(instr.gpr0, 0,
-                                        "((" + op_a + ") < (" + op_b + ")) ? 1.0 : 0", 1, 1);
-                break;
-            case PredCondition::Equal:
-                regs.SetRegisterToFloat(instr.gpr0, 0,
-                                        "((" + op_a + ") == (" + op_b + ")) ? 1.0 : 0", 1, 1);
-                break;
-            case PredCondition::LessEqual:
-                regs.SetRegisterToFloat(instr.gpr0, 0,
-                                        "((" + op_a + ") <= (" + op_b + ")) ? 1.0 : 0", 1, 1);
-                break;
-            case PredCondition::GreaterThan:
-                regs.SetRegisterToFloat(instr.gpr0, 0,
-                                        "((" + op_a + ") > (" + op_b + ")) ? 1.0 : 0", 1, 1);
-                break;
-            default:
-                NGLOG_CRITICAL(HW_GPU, "Unhandled predicate condition: {} (a: {}, b: {})",
-                               static_cast<unsigned>(instr.fset.cond.Value()), op_a, op_b);
-                UNREACHABLE();
-            }
+            std::string second_pred =
+                GetPredicateCondition(instr.fset.pred39, instr.fset.neg_pred != 0);
+
+            std::string comparator = GetPredicateComparison(instr.fset.cond);
+            std::string combiner = GetPredicateCombiner(instr.fset.op);
+
+            std::string predicate = "(((" + op_a + ") " + comparator + " (" + op_b + ")) " +
+                                    combiner + " (" + second_pred + "))";
+
+            regs.SetRegisterToFloat(instr.gpr0, 0, predicate + " ? 1.0 : 0.0", 1, 1);
             break;
         }
         default: {

src/video_core/renderer_opengl/gl_shader_util.h

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include <string>
 #include <vector>
 #include <glad/glad.h>
 #include "common/assert.h"
@@ -11,6 +12,27 @@
 
 namespace GLShader {
 
+/**
+ * Utility function to log the source code of a list of shaders.
+ * @param shaders The OpenGL shaders whose source we will print.
+ */
+template <typename... T>
+void LogShaderSource(T... shaders) {
+    auto shader_list = {shaders...};
+
+    for (const auto& shader : shader_list) {
+        if (shader == 0)
+            continue;
+
+        GLint source_length;
+        glGetShaderiv(shader, GL_SHADER_SOURCE_LENGTH, &source_length);
+
+        std::string source(source_length, ' ');
+        glGetShaderSource(shader, source_length, nullptr, &source[0]);
+        NGLOG_INFO(Render_OpenGL, "Shader source {}", source);
+    }
+}
+
 /**
  * Utility function to create and compile an OpenGL GLSL shader
  * @param source String of the GLSL shader program
@@ -55,6 +77,11 @@ GLuint LoadProgram(bool separable_program, T... shaders) {
         }
     }
 
+    if (result == GL_FALSE) {
+        // There was a problem linking the shader, print the source for debugging purposes.
+        LogShaderSource(shaders...);
+    }
+
     ASSERT_MSG(result == GL_TRUE, "Shader not linked");
 
     ((shaders == 0 ? (void)0 : glDetachShader(program_id, shaders)), ...);

src/video_core/textures/decoders.cpp

@@ -46,6 +46,7 @@ void CopySwizzledData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_
 u32 BytesPerPixel(TextureFormat format) {
     switch (format) {
     case TextureFormat::DXT1:
+    case TextureFormat::DXN1:
         // In this case a 'pixel' actually refers to a 4x4 tile.
         return 8;
     case TextureFormat::DXT23:
@@ -55,8 +56,13 @@ u32 BytesPerPixel(TextureFormat format) {
     case TextureFormat::A8R8G8B8:
     case TextureFormat::A2B10G10R10:
         return 4;
+    case TextureFormat::A1B5G5R5:
     case TextureFormat::B5G6R5:
         return 2;
+    case TextureFormat::R8:
+        return 1;
+    case TextureFormat::R16_G16_B16_A16:
+        return 8;
     default:
         UNIMPLEMENTED_MSG("Format not implemented");
         break;
@@ -74,13 +80,18 @@ std::vector<u8> UnswizzleTexture(VAddr address, TextureFormat format, u32 width,
     case TextureFormat::DXT1:
     case TextureFormat::DXT23:
     case TextureFormat::DXT45:
-        // In the DXT formats, each 4x4 tile is swizzled instead of just individual pixel values.
+    case TextureFormat::DXN1:
+        // In the DXT and DXN formats, each 4x4 tile is swizzled instead of just individual pixel
+        // values.
         CopySwizzledData(width / 4, height / 4, bytes_per_pixel, bytes_per_pixel, data,
                          unswizzled_data.data(), true, block_height);
         break;
     case TextureFormat::A8R8G8B8:
     case TextureFormat::A2B10G10R10:
+    case TextureFormat::A1B5G5R5:
     case TextureFormat::B5G6R5:
+    case TextureFormat::R8:
+    case TextureFormat::R16_G16_B16_A16:
         CopySwizzledData(width, height, bytes_per_pixel, bytes_per_pixel, data,
                          unswizzled_data.data(), true, block_height);
         break;
@@ -101,9 +112,12 @@ std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat
     case TextureFormat::DXT1:
     case TextureFormat::DXT23:
     case TextureFormat::DXT45:
+    case TextureFormat::DXN1:
     case TextureFormat::A8R8G8B8:
     case TextureFormat::A2B10G10R10:
+    case TextureFormat::A1B5G5R5:
     case TextureFormat::B5G6R5:
+    case TextureFormat::R8:
         // TODO(Subv): For the time being just forward the same data without any decoding.
         rgba_data = texture_data;
         break;

src/video_core/textures/texture.h

@@ -14,12 +14,82 @@ namespace Tegra {
 namespace Texture {
 
 enum class TextureFormat : u32 {
-    A8R8G8B8 = 0x8,
-    A2B10G10R10 = 0x9,
+    R32_G32_B32_A32 = 0x01,
+    R32_G32_B32 = 0x02,
+    R16_G16_B16_A16 = 0x03,
+    R32_G32 = 0x04,
+    R32_B24G8 = 0x05,
+    ETC2_RGB = 0x06,
+    X8B8G8R8 = 0x07,
+    A8R8G8B8 = 0x08,
+    A2B10G10R10 = 0x09,
+    ETC2_RGB_PTA = 0x0a,
+    ETC2_RGBA = 0x0b,
+    R16_G16 = 0x0c,
+    G8R24 = 0x0d,
+    G24R8 = 0x0e,
+    R32 = 0x0f,
+    BC6H_SF16 = 0x10,
+    BC6H_UF16 = 0x11,
+    A4B4G4R4 = 0x12,
+    A5B5G5R1 = 0x13,
+    A1B5G5R5 = 0x14,
     B5G6R5 = 0x15,
+    B6G5R5 = 0x16,
+    BC7U = 0x17,
+    G8R8 = 0x18,
+    EAC = 0x19,
+    EACX2 = 0x1a,
+    R16 = 0x1b,
+    Y8_VIDEO = 0x1c,
+    R8 = 0x1d,
+    G4R4 = 0x1e,
+    R1 = 0x1f,
+    E5B9G9R9_SHAREDEXP = 0x20,
+    BF10GF11RF11 = 0x21,
+    G8B8G8R8 = 0x22,
+    B8G8R8G8 = 0x23,
     DXT1 = 0x24,
     DXT23 = 0x25,
     DXT45 = 0x26,
+    DXN1 = 0x27,
+    DXN2 = 0x28,
+    Z24S8 = 0x29,
+    X8Z24 = 0x2a,
+    S8Z24 = 0x2b,
+    X4V4Z24__COV4R4V = 0x2c,
+    X4V4Z24__COV8R8V = 0x2d,
+    V8Z24__COV4R12V = 0x2e,
+    ZF32 = 0x2f,
+    ZF32_X24S8 = 0x30,
+    X8Z24_X20V4S8__COV4R4V = 0x31,
+    X8Z24_X20V4S8__COV8R8V = 0x32,
+    ZF32_X20V4X8__COV4R4V = 0x33,
+    ZF32_X20V4X8__COV8R8V = 0x34,
+    ZF32_X20V4S8__COV4R4V = 0x35,
+    ZF32_X20V4S8__COV8R8V = 0x36,
+    X8Z24_X16V8S8__COV4R12V = 0x37,
+    ZF32_X16V8X8__COV4R12V = 0x38,
+    ZF32_X16V8S8__COV4R12V = 0x39,
+    Z16 = 0x3a,
+    V8Z24__COV8R24V = 0x3b,
+    X8Z24_X16V8S8__COV8R24V = 0x3c,
+    ZF32_X16V8X8__COV8R24V = 0x3d,
+    ZF32_X16V8S8__COV8R24V = 0x3e,
+    ASTC_2D_4X4 = 0x40,
+    ASTC_2D_5X5 = 0x41,
+    ASTC_2D_6X6 = 0x42,
+    ASTC_2D_8X8 = 0x44,
+    ASTC_2D_10X10 = 0x45,
+    ASTC_2D_12X12 = 0x46,
+    ASTC_2D_5X4 = 0x50,
+    ASTC_2D_6X5 = 0x51,
+    ASTC_2D_8X6 = 0x52,
+    ASTC_2D_10X8 = 0x53,
+    ASTC_2D_12X10 = 0x54,
+    ASTC_2D_8X5 = 0x55,
+    ASTC_2D_10X5 = 0x56,
+    ASTC_2D_10X6 = 0x57,
 };
 
 enum class TextureType : u32 {

src/yuzu/configuration/config.cpp

@@ -78,6 +78,7 @@ void Config::ReadValues() {
 
     qt_config->beginGroup("Core");
     Settings::values.use_cpu_jit = qt_config->value("use_cpu_jit", true).toBool();
+    Settings::values.use_multi_core = qt_config->value("use_multi_core", false).toBool();
     qt_config->endGroup();
 
     qt_config->beginGroup("Renderer");
@@ -177,6 +178,7 @@ void Config::SaveValues() {
 
     qt_config->beginGroup("Core");
     qt_config->setValue("use_cpu_jit", Settings::values.use_cpu_jit);
+    qt_config->setValue("use_multi_core", Settings::values.use_multi_core);
     qt_config->endGroup();
 
     qt_config->beginGroup("Renderer");

src/yuzu/configuration/configure_general.cpp

@@ -20,6 +20,7 @@ ConfigureGeneral::ConfigureGeneral(QWidget* parent)
     this->setConfiguration();
 
     ui->use_cpu_jit->setEnabled(!Core::System::GetInstance().IsPoweredOn());
+    ui->use_multi_core->setEnabled(!Core::System::GetInstance().IsPoweredOn());
     ui->use_docked_mode->setEnabled(!Core::System::GetInstance().IsPoweredOn());
 }
 
@@ -30,6 +31,7 @@ void ConfigureGeneral::setConfiguration() {
     ui->toggle_check_exit->setChecked(UISettings::values.confirm_before_closing);
     ui->theme_combobox->setCurrentIndex(ui->theme_combobox->findData(UISettings::values.theme));
     ui->use_cpu_jit->setChecked(Settings::values.use_cpu_jit);
+    ui->use_multi_core->setChecked(Settings::values.use_multi_core);
     ui->use_docked_mode->setChecked(Settings::values.use_docked_mode);
 }
 
@@ -40,6 +42,7 @@ void ConfigureGeneral::applyConfiguration() {
         ui->theme_combobox->itemData(ui->theme_combobox->currentIndex()).toString();
 
     Settings::values.use_cpu_jit = ui->use_cpu_jit->isChecked();
+    Settings::values.use_multi_core = ui->use_multi_core->isChecked();
     Settings::values.use_docked_mode = ui->use_docked_mode->isChecked();
     Settings::Apply();
 }

src/yuzu/configuration/configure_general.ui

@@ -58,6 +58,13 @@
             </property>
            </widget>
           </item>
+          <item>
+           <widget class="QCheckBox" name="use_multi_core">
+            <property name="text">
+             <string>Enable multi-core</string>
+            </property>
+           </widget>
+          </item>
          </layout>
         </item>
        </layout>

src/yuzu/debugger/registers.cpp

@@ -63,7 +63,7 @@ void RegistersWidget::OnDebugModeEntered() {
 
     for (int i = 0; i < core_registers->childCount(); ++i)
         core_registers->child(i)->setText(
-            1, QString("0x%1").arg(Core::CPU().GetReg(i), 8, 16, QLatin1Char('0')));
+            1, QString("0x%1").arg(Core::CurrentArmInterface().GetReg(i), 8, 16, QLatin1Char('0')));
 
     UpdateCPSRValues();
 }
@@ -122,7 +122,7 @@ void RegistersWidget::CreateCPSRChildren() {
 }
 
 void RegistersWidget::UpdateCPSRValues() {
-    const u32 cpsr_val = Core::CPU().GetCPSR();
+    const u32 cpsr_val = Core::CurrentArmInterface().GetCPSR();
 
     cpsr->setText(1, QString("0x%1").arg(cpsr_val, 8, 16, QLatin1Char('0')));
     cpsr->child(0)->setText(

src/yuzu/debugger/wait_tree.cpp

@@ -51,13 +51,21 @@ std::size_t WaitTreeItem::Row() const {
 }
 
 std::vector<std::unique_ptr<WaitTreeThread>> WaitTreeItem::MakeThreadItemList() {
-    const auto& threads = Core::System::GetInstance().Scheduler().GetThreadList();
     std::vector<std::unique_ptr<WaitTreeThread>> item_list;
-    item_list.reserve(threads.size());
-    for (std::size_t i = 0; i < threads.size(); ++i) {
-        item_list.push_back(std::make_unique<WaitTreeThread>(*threads[i]));
-        item_list.back()->row = i;
-    }
+    std::size_t row = 0;
+    auto add_threads = [&](const std::vector<Kernel::SharedPtr<Kernel::Thread>>& threads) {
+        for (std::size_t i = 0; i < threads.size(); ++i) {
+            item_list.push_back(std::make_unique<WaitTreeThread>(*threads[i]));
+            item_list.back()->row = row;
+            ++row;
+        }
+    };
+
+    add_threads(Core::System::GetInstance().Scheduler(0)->GetThreadList());
+    add_threads(Core::System::GetInstance().Scheduler(1)->GetThreadList());
+    add_threads(Core::System::GetInstance().Scheduler(2)->GetThreadList());
+    add_threads(Core::System::GetInstance().Scheduler(3)->GetThreadList());
+
     return item_list;
 }
 
@@ -90,6 +98,30 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeMutexInfo::GetChildren() cons
     return list;
 }
 
+WaitTreeCallstack::WaitTreeCallstack(const Kernel::Thread& thread) : thread(thread) {}
+
+QString WaitTreeCallstack::GetText() const {
+    return tr("Call stack");
+}
+
+std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeCallstack::GetChildren() const {
+    std::vector<std::unique_ptr<WaitTreeItem>> list;
+
+    constexpr size_t BaseRegister = 29;
+    u64 base_pointer = thread.context.cpu_registers[BaseRegister];
+
+    while (base_pointer != 0) {
+        u64 lr = Memory::Read64(base_pointer + sizeof(u64));
+        if (lr == 0)
+            break;
+        list.push_back(
+            std::make_unique<WaitTreeText>(tr("0x%1").arg(lr - sizeof(u32), 16, 16, QChar('0'))));
+        base_pointer = Memory::Read64(base_pointer);
+    }
+
+    return list;
+}
+
 WaitTreeWaitObject::WaitTreeWaitObject(const Kernel::WaitObject& o) : object(o) {}
 
 bool WaitTreeExpandableItem::IsExpandable() const {
@@ -240,6 +272,9 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeThread::GetChildren() const {
     }
 
     list.push_back(std::make_unique<WaitTreeText>(tr("processor = %1").arg(processor)));
+    list.push_back(std::make_unique<WaitTreeText>(tr("ideal core = %1").arg(thread.ideal_core)));
+    list.push_back(
+        std::make_unique<WaitTreeText>(tr("affinity mask = %1").arg(thread.affinity_mask)));
     list.push_back(std::make_unique<WaitTreeText>(tr("thread id = %1").arg(thread.GetThreadId())));
     list.push_back(std::make_unique<WaitTreeText>(tr("priority = %1(current) / %2(normal)")
                                                       .arg(thread.current_priority)
@@ -258,6 +293,8 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeThread::GetChildren() const {
                                                             thread.IsSleepingOnWaitAll()));
     }
 
+    list.push_back(std::make_unique<WaitTreeCallstack>(thread));
+
     return list;
 }
 

src/yuzu/debugger/wait_tree.h

@@ -73,6 +73,17 @@ private:
     Kernel::SharedPtr<Kernel::Thread> owner;
 };
 
+class WaitTreeCallstack : public WaitTreeExpandableItem {
+    Q_OBJECT
+public:
+    explicit WaitTreeCallstack(const Kernel::Thread& thread);
+    QString GetText() const override;
+    std::vector<std::unique_ptr<WaitTreeItem>> GetChildren() const override;
+
+private:
+    const Kernel::Thread& thread;
+};
+
 class WaitTreeWaitObject : public WaitTreeExpandableItem {
     Q_OBJECT
 public:

src/yuzu_cmd/config.cpp

@@ -91,6 +91,7 @@ void Config::ReadValues() {
 
     // Core
     Settings::values.use_cpu_jit = sdl2_config->GetBoolean("Core", "use_cpu_jit", true);
+    Settings::values.use_multi_core = sdl2_config->GetBoolean("Core", "use_multi_core", false);
 
     // Renderer
     Settings::values.resolution_factor =

src/yuzu_cmd/default_ini.h

@@ -80,6 +80,10 @@ touch_device=
 # 0: Interpreter (slow), 1 (default): JIT (fast)
 use_cpu_jit =
 
+# Whether to use multi-core for CPU emulation
+# 0 (default): Disabled, 1: Enabled
+use_multi_core=
+
 [Renderer]
 # Whether to use software or hardware rendering.
 # 0: Software, 1 (default): Hardware

src/yuzu_cmd/yuzu.cpp

@@ -7,8 +7,19 @@
 #include <string>
 #include <thread>
 
-// This needs to be included before getopt.h because the latter #defines symbols used by it
+#include "common/logging/backend.h"
+#include "common/logging/filter.h"
+#include "common/logging/log.h"
 #include "common/microprofile.h"
+#include "common/scm_rev.h"
+#include "common/scope_exit.h"
+#include "common/string_util.h"
+#include "core/core.h"
+#include "core/gdbstub/gdbstub.h"
+#include "core/loader/loader.h"
+#include "core/settings.h"
+#include "yuzu_cmd/config.h"
+#include "yuzu_cmd/emu_window/emu_window_sdl2.h"
 
 #ifdef _MSC_VER
 #include <getopt.h>
@@ -24,19 +35,6 @@
 #include <shellapi.h>
 #endif
 
-#include "common/logging/backend.h"
-#include "common/logging/filter.h"
-#include "common/logging/log.h"
-#include "common/scm_rev.h"
-#include "common/scope_exit.h"
-#include "common/string_util.h"
-#include "core/core.h"
-#include "core/gdbstub/gdbstub.h"
-#include "core/loader/loader.h"
-#include "core/settings.h"
-#include "yuzu_cmd/config.h"
-#include "yuzu_cmd/emu_window/emu_window_sdl2.h"
-
 #ifdef _WIN32
 extern "C" {
 // tells Nvidia and AMD drivers to use the dedicated GPU by default on laptops with switchable