Service/Audio: update audren:u service

Build: Fixed some MSVC warnings in various parts of the code.

.travis.yml

@@ -42,3 +42,7 @@ notifications:
   webhooks:
     urls:
       - https://api.yuzu-emu.org/code/travis/notify
+
+cache:
+  directories:
+    - $HOME/.ccache

.travis/linux/build.sh

@@ -1,3 +1,3 @@
 #!/bin/bash -ex
 
-docker run -v $(pwd):/yuzu ubuntu:18.04 /bin/bash /yuzu/.travis/linux/docker.sh
+docker run -e CCACHE_DIR=/ccache -v $HOME/.ccache:/ccache -v $(pwd):/yuzu ubuntu:18.04 /bin/bash /yuzu/.travis/linux/docker.sh

.travis/linux/docker.sh

@@ -1,16 +1,18 @@
 #!/bin/bash -ex
 
 apt-get update
-apt-get install -y build-essential git libqt5opengl5-dev libsdl2-dev libssl-dev python qtbase5-dev wget
-
-# Get a recent version of CMake
-wget https://cmake.org/files/v3.10/cmake-3.10.1-Linux-x86_64.sh
-sh cmake-3.10.1-Linux-x86_64.sh --exclude-subdir --prefix=/ --skip-license
+apt-get install --no-install-recommends -y build-essential git libqt5opengl5-dev libsdl2-dev libssl-dev python qtbase5-dev wget cmake ninja-build ccache
 
 cd /yuzu
 
+export PATH=/usr/lib/ccache:$PATH
+ln -sf /usr/bin/ccache /usr/lib/ccache/cc
+ln -sf /usr/bin/ccache /usr/lib/ccache/c++
 mkdir build && cd build
-cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release
-make -j4
+ccache --show-stats > ccache_before
+cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -G Ninja
+ninja
+ccache --show-stats > ccache_after
+diff -U100 ccache_before ccache_after || true
 
 ctest -VV -C Release

.travis/macos/build.sh

@@ -7,8 +7,12 @@ export Qt5_DIR=$(brew --prefix)/opt/qt5
 export UNICORNDIR=$(pwd)/externals/unicorn
 
 mkdir build && cd build
+export PATH=/usr/local/opt/ccache/libexec:$PATH
+ccache --show-stats > ccache_before
 cmake --version
 cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release
 make -j4
+ccache --show-stats > ccache_after
+diff -U100 ccache_before ccache_after || true
 
 ctest -VV -C Release

.travis/macos/deps.sh

@@ -1,5 +1,5 @@
 #!/bin/sh -ex
 
 brew update
-brew install dylibbundler p7zip qt5 sdl2
+brew install dylibbundler p7zip qt5 sdl2 ccache
 brew outdated cmake || brew upgrade cmake

src/common/logging/backend.cpp

@@ -35,11 +35,13 @@ namespace Log {
     SUB(Service, AM)                                                                               \
     SUB(Service, AOC)                                                                              \
     SUB(Service, APM)                                                                              \
+    SUB(Service, BCAT)                                                                             \
     SUB(Service, Fatal)                                                                            \
     SUB(Service, Friend)                                                                           \
     SUB(Service, FS)                                                                               \
     SUB(Service, HID)                                                                              \
     SUB(Service, LM)                                                                               \
+    SUB(Service, MM)                                                                               \
     SUB(Service, NFP)                                                                              \
     SUB(Service, NIFM)                                                                             \
     SUB(Service, NS)                                                                               \

src/common/logging/log.h

@@ -55,11 +55,13 @@ 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
     Service_HID,       ///< The HID (Human interface device) service
     Service_LM,        ///< The LM (Logger) service
+    Service_MM,        ///< The MM (Multimedia) service
     Service_NFP,       ///< The NFP service
     Service_NIFM,      ///< The NIFM (Network interface) service
     Service_NS,        ///< The NS services

src/common/string_util.cpp

@@ -64,6 +64,10 @@ std::string ArrayToString(const u8* data, size_t size, int line_len, bool spaces
     return oss.str();
 }
 
+std::string StringFromBuffer(const std::vector<u8>& data) {
+    return std::string(data.begin(), std::find(data.begin(), data.end(), '\0'));
+}
+
 // Turns "  hej " into "hej". Also handles tabs.
 std::string StripSpaces(const std::string& str) {
     const size_t s = str.find_first_not_of(" \t\r\n");

src/common/string_util.h

@@ -21,6 +21,8 @@ std::string ToUpper(std::string str);
 
 std::string ArrayToString(const u8* data, size_t size, int line_len = 20, bool spaces = true);
 
+std::string StringFromBuffer(const std::vector<u8>& data);
+
 std::string StripSpaces(const std::string& s);
 std::string StripQuotes(const std::string& s);
 

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
@@ -142,6 +148,8 @@ add_library(core STATIC
     hle/service/hid/hid.h
     hle/service/lm/lm.cpp
     hle/service/lm/lm.h
+    hle/service/mm/mm_u.cpp
+    hle/service/mm/mm_u.h
     hle/service/nifm/nifm.cpp
     hle/service/nifm/nifm.h
     hle/service/nifm/nifm_a.cpp
@@ -169,6 +177,8 @@ 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

@@ -35,6 +35,17 @@ LoadDll LoadDll::g_load_dll;
         }                                                                                          \
     } while (0)
 
+static void CodeHook(uc_engine* uc, uint64_t address, uint32_t size, void* user_data) {
+    GDBStub::BreakpointAddress bkpt =
+        GDBStub::GetNextBreakpointFromAddress(address, GDBStub::BreakpointType::Execute);
+    if (GDBStub::IsMemoryBreak() ||
+        (bkpt.type != GDBStub::BreakpointType::None && address == bkpt.address)) {
+        auto core = static_cast<ARM_Unicorn*>(user_data);
+        core->RecordBreak(bkpt);
+        uc_emu_stop(uc);
+    }
+}
+
 static void InterruptHook(uc_engine* uc, u32 intNo, void* user_data) {
     u32 esr{};
     CHECKED(uc_reg_read(uc, UC_ARM64_REG_ESR, &esr));
@@ -52,7 +63,7 @@ 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 {};
@@ -67,6 +78,10 @@ ARM_Unicorn::ARM_Unicorn() {
     uc_hook hook{};
     CHECKED(uc_hook_add(uc, &hook, UC_HOOK_INTR, (void*)InterruptHook, this, 0, -1));
     CHECKED(uc_hook_add(uc, &hook, UC_HOOK_MEM_INVALID, (void*)UnmappedMemoryHook, this, 0, -1));
+    if (GDBStub::IsServerEnabled()) {
+        CHECKED(uc_hook_add(uc, &hook, UC_HOOK_CODE, (void*)CodeHook, this, 0, -1));
+        last_bkpt_hit = false;
+    }
 }
 
 ARM_Unicorn::~ARM_Unicorn() {
@@ -155,7 +170,11 @@ void ARM_Unicorn::SetTlsAddress(VAddr base) {
 }
 
 void ARM_Unicorn::Run() {
-    ExecuteInstructions(std::max(CoreTiming::GetDowncount(), 0));
+    if (GDBStub::IsServerEnabled()) {
+        ExecuteInstructions(std::max(4000000, 0));
+    } else {
+        ExecuteInstructions(std::max(CoreTiming::GetDowncount(), 0));
+    }
 }
 
 void ARM_Unicorn::Step() {
@@ -168,6 +187,18 @@ void ARM_Unicorn::ExecuteInstructions(int num_instructions) {
     MICROPROFILE_SCOPE(ARM_Jit);
     CHECKED(uc_emu_start(uc, GetPC(), 1ULL << 63, 0, num_instructions));
     CoreTiming::AddTicks(num_instructions);
+    if (GDBStub::IsServerEnabled()) {
+        if (last_bkpt_hit) {
+            uc_reg_write(uc, UC_ARM64_REG_PC, &last_bkpt.address);
+        }
+        Kernel::Thread* thread = Kernel::GetCurrentThread();
+        SaveContext(thread->context);
+        if (last_bkpt_hit) {
+            last_bkpt_hit = false;
+            GDBStub::Break();
+        }
+        GDBStub::SendTrap(thread, 5);
+    }
 }
 
 void ARM_Unicorn::SaveContext(ARM_Interface::ThreadContext& ctx) {
@@ -233,3 +264,8 @@ void ARM_Unicorn::PrepareReschedule() {
 }
 
 void ARM_Unicorn::ClearInstructionCache() {}
+
+void ARM_Unicorn::RecordBreak(GDBStub::BreakpointAddress bkpt) {
+    last_bkpt = bkpt;
+    last_bkpt_hit = true;
+}

src/core/arm/unicorn/arm_unicorn.h

@@ -7,6 +7,7 @@
 #include <unicorn/unicorn.h>
 #include "common/common_types.h"
 #include "core/arm/arm_interface.h"
+#include "core/gdbstub/gdbstub.h"
 
 class ARM_Unicorn final : public ARM_Interface {
 public:
@@ -35,7 +36,10 @@ public:
     void Step() override;
     void ClearInstructionCache() override;
     void PageTableChanged() override{};
+    void RecordBreak(GDBStub::BreakpointAddress bkpt);
 
 private:
     uc_engine* uc{};
+    GDBStub::BreakpointAddress last_bkpt{};
+    bool last_bkpt_hit;
 };

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/file_sys/partition_filesystem.cpp

@@ -73,21 +73,21 @@ u32 PartitionFilesystem::GetNumEntries() const {
     return pfs_header.num_entries;
 }
 
-u64 PartitionFilesystem::GetEntryOffset(int index) const {
+u64 PartitionFilesystem::GetEntryOffset(u32 index) const {
     if (index > GetNumEntries())
         return 0;
 
     return content_offset + pfs_entries[index].fs_entry.offset;
 }
 
-u64 PartitionFilesystem::GetEntrySize(int index) const {
+u64 PartitionFilesystem::GetEntrySize(u32 index) const {
     if (index > GetNumEntries())
         return 0;
 
     return pfs_entries[index].fs_entry.size;
 }
 
-std::string PartitionFilesystem::GetEntryName(int index) const {
+std::string PartitionFilesystem::GetEntryName(u32 index) const {
     if (index > GetNumEntries())
         return "";
 

src/core/file_sys/partition_filesystem.h

@@ -27,9 +27,9 @@ public:
     Loader::ResultStatus Load(const std::vector<u8>& file_data, size_t offset = 0);
 
     u32 GetNumEntries() const;
-    u64 GetEntryOffset(int index) const;
-    u64 GetEntrySize(int index) const;
-    std::string GetEntryName(int index) const;
+    u64 GetEntryOffset(u32 index) const;
+    u64 GetEntrySize(u32 index) const;
+    std::string GetEntryName(u32 index) const;
     u64 GetFileOffset(const std::string& name) const;
     u64 GetFileSize(const std::string& name) const;
 

src/core/gdbstub/gdbstub.cpp

@@ -32,9 +32,13 @@
 
 #include "common/logging/log.h"
 #include "common/string_util.h"
+#include "common/swap.h"
 #include "core/arm/arm_interface.h"
 #include "core/core.h"
+#include "core/core_cpu.h"
 #include "core/gdbstub/gdbstub.h"
+#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/scheduler.h"
 #include "core/loader/loader.h"
 #include "core/memory.h"
 
@@ -137,15 +141,17 @@ static u8 command_buffer[GDB_BUFFER_SIZE];
 static u32 command_length;
 
 static u32 latest_signal = 0;
-static bool step_break = false;
 static bool memory_break = false;
 
+static Kernel::Thread* current_thread = nullptr;
+
 // Binding to a port within the reserved ports range (0-1023) requires root permissions,
 // so default to a port outside of that range.
 static u16 gdbstub_port = 24689;
 
 static bool halt_loop = true;
 static bool step_loop = false;
+static bool send_trap = false;
 
 // If set to false, the server will never be started and no
 // gdbstub-related functions will be executed.
@@ -165,6 +171,53 @@ static std::map<u64, Breakpoint> breakpoints_execute;
 static std::map<u64, Breakpoint> breakpoints_read;
 static std::map<u64, Breakpoint> breakpoints_write;
 
+static Kernel::Thread* FindThreadById(int id) {
+    for (int core = 0; core < Core::NUM_CPU_CORES; core++) {
+        auto threads = Core::System::GetInstance().Scheduler(core)->GetThreadList();
+        for (auto thread : threads) {
+            if (thread->GetThreadId() == id) {
+                current_thread = thread.get();
+                return current_thread;
+            }
+        }
+    }
+    return nullptr;
+}
+
+static u64 RegRead(int id, Kernel::Thread* thread = nullptr) {
+    if (!thread) {
+        return 0;
+    }
+
+    if (id < SP_REGISTER) {
+        return thread->context.cpu_registers[id];
+    } else if (id == SP_REGISTER) {
+        return thread->context.sp;
+    } else if (id == PC_REGISTER) {
+        return thread->context.pc;
+    } else if (id == CPSR_REGISTER) {
+        return thread->context.cpsr;
+    } else {
+        return 0;
+    }
+}
+
+static void RegWrite(int id, u64 val, Kernel::Thread* thread = nullptr) {
+    if (!thread) {
+        return;
+    }
+
+    if (id < SP_REGISTER) {
+        thread->context.cpu_registers[id] = val;
+    } else if (id == SP_REGISTER) {
+        thread->context.sp = val;
+    } else if (id == PC_REGISTER) {
+        thread->context.pc = val;
+    } else if (id == CPSR_REGISTER) {
+        thread->context.cpsr = val;
+    }
+}
+
 /**
  * Turns hex string character into the equivalent byte.
  *
@@ -193,7 +246,7 @@ static u8 NibbleToHex(u8 n) {
     if (n < 0xA) {
         return '0' + n;
     } else {
-        return 'A' + n - 0xA;
+        return 'a' + n - 0xA;
     }
 }
 
@@ -439,6 +492,8 @@ static void SendReply(const char* reply) {
         return;
     }
 
+    NGLOG_DEBUG(Debug_GDBStub, "Reply: {}", reply);
+
     memset(command_buffer, 0, sizeof(command_buffer));
 
     command_length = static_cast<u32>(strlen(reply));
@@ -483,6 +538,22 @@ static void HandleQuery() {
     } else if (strncmp(query, "Xfer:features:read:target.xml:",
                        strlen("Xfer:features:read:target.xml:")) == 0) {
         SendReply(target_xml);
+    } else if (strncmp(query, "Offsets", strlen("Offsets")) == 0) {
+        std::string buffer = fmt::format("TextSeg={:0x}", Memory::PROCESS_IMAGE_VADDR);
+        SendReply(buffer.c_str());
+    } else if (strncmp(query, "fThreadInfo", strlen("fThreadInfo")) == 0) {
+        std::string val = "m";
+        for (int core = 0; core < Core::NUM_CPU_CORES; core++) {
+            auto threads = Core::System::GetInstance().Scheduler(core)->GetThreadList();
+            for (auto thread : threads) {
+                val += fmt::format("{:x}", thread->GetThreadId());
+                val += ",";
+            }
+        }
+        val.pop_back();
+        SendReply(val.c_str());
+    } else if (strncmp(query, "sThreadInfo", strlen("sThreadInfo")) == 0) {
+        SendReply("l");
     } else {
         SendReply("");
     }
@@ -490,11 +561,40 @@ static void HandleQuery() {
 
 /// Handle set thread command from gdb client.
 static void HandleSetThread() {
-    if (memcmp(command_buffer, "Hg0", 3) == 0 || memcmp(command_buffer, "Hc-1", 4) == 0 ||
-        memcmp(command_buffer, "Hc0", 4) == 0 || memcmp(command_buffer, "Hc1", 4) == 0) {
-        return SendReply("OK");
+    if (memcmp(command_buffer, "Hc", 2) == 0 || memcmp(command_buffer, "Hg", 2) == 0) {
+        int thread_id = -1;
+        if (command_buffer[2] != '-') {
+            thread_id = static_cast<int>(HexToInt(
+                command_buffer + 2,
+                command_length - 2 /*strlen(reinterpret_cast<char*>(command_buffer) + 2)*/));
+        }
+        if (thread_id >= 1) {
+            current_thread = FindThreadById(thread_id);
+        }
+        if (!current_thread) {
+            thread_id = 1;
+            current_thread = FindThreadById(thread_id);
+        }
+        if (current_thread) {
+            SendReply("OK");
+            return;
+        }
     }
+    SendReply("E01");
+}
 
+/// Handle thread alive command from gdb client.
+static void HandleThreadAlive() {
+    int thread_id = static_cast<int>(
+        HexToInt(command_buffer + 1,
+                 command_length - 1 /*strlen(reinterpret_cast<char*>(command_buffer) + 1)*/));
+    if (thread_id == 0) {
+        thread_id = 1;
+    }
+    if (FindThreadById(thread_id)) {
+        SendReply("OK");
+        return;
+    }
     SendReply("E01");
 }
 
@@ -503,15 +603,24 @@ static void HandleSetThread() {
  *
  * @param signal Signal to be sent to client.
  */
-static void SendSignal(u32 signal) {
+static void SendSignal(Kernel::Thread* thread, u32 signal, bool full = true) {
     if (gdbserver_socket == -1) {
         return;
     }
 
     latest_signal = signal;
 
-    std::string buffer = fmt::format("T{:02x}", latest_signal);
-    NGLOG_DEBUG(Debug_GDBStub, "Response: {}", buffer);
+    std::string buffer;
+    if (full) {
+        buffer = fmt::format("T{:02x}{:02x}:{:016x};{:02x}:{:016x};", latest_signal, PC_REGISTER,
+                             Common::swap64(RegRead(PC_REGISTER, thread)), SP_REGISTER,
+                             Common::swap64(RegRead(SP_REGISTER, thread)));
+    } else {
+        buffer = fmt::format("T{:02x};", latest_signal);
+    }
+
+    buffer += fmt::format("thread:{:x};", thread->GetThreadId());
+
     SendReply(buffer.c_str());
 }
 
@@ -527,7 +636,7 @@ static void ReadCommand() {
     } else if (c == 0x03) {
         NGLOG_INFO(Debug_GDBStub, "gdb: found break command");
         halt_loop = true;
-        SendSignal(SIGTRAP);
+        SendSignal(current_thread, SIGTRAP);
         return;
     } else if (c != GDB_STUB_START) {
         NGLOG_DEBUG(Debug_GDBStub, "gdb: read invalid byte {:02X}", c);
@@ -598,11 +707,11 @@ static void ReadRegister() {
     }
 
     if (id <= SP_REGISTER) {
-        LongToGdbHex(reply, Core::CPU().GetReg(static_cast<int>(id)));
+        LongToGdbHex(reply, RegRead(id, current_thread));
     } else if (id == PC_REGISTER) {
-        LongToGdbHex(reply, Core::CPU().GetPC());
+        LongToGdbHex(reply, RegRead(id, current_thread));
     } else if (id == CPSR_REGISTER) {
-        IntToGdbHex(reply, Core::CPU().GetCPSR());
+        IntToGdbHex(reply, (u32)RegRead(id, current_thread));
     } else {
         return SendReply("E01");
     }
@@ -618,16 +727,16 @@ 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, RegRead(reg, current_thread));
     }
 
     bufptr += (32 * 16);
 
-    LongToGdbHex(bufptr, Core::CPU().GetPC());
+    LongToGdbHex(bufptr, RegRead(PC_REGISTER, current_thread));
 
     bufptr += 16;
 
-    IntToGdbHex(bufptr, Core::CPU().GetCPSR());
+    IntToGdbHex(bufptr, (u32)RegRead(CPSR_REGISTER, current_thread));
 
     bufptr += 8;
 
@@ -646,11 +755,11 @@ static void WriteRegister() {
     }
 
     if (id <= SP_REGISTER) {
-        Core::CPU().SetReg(id, GdbHexToLong(buffer_ptr));
+        RegWrite(id, GdbHexToLong(buffer_ptr), current_thread);
     } else if (id == PC_REGISTER) {
-        Core::CPU().SetPC(GdbHexToLong(buffer_ptr));
+        RegWrite(id, GdbHexToLong(buffer_ptr), current_thread);
     } else if (id == CPSR_REGISTER) {
-        Core::CPU().SetCPSR(GdbHexToInt(buffer_ptr));
+        RegWrite(id, GdbHexToInt(buffer_ptr), current_thread);
     } else {
         return SendReply("E01");
     }
@@ -667,11 +776,11 @@ 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));
+            RegWrite(reg, GdbHexToLong(buffer_ptr + i * 16), current_thread);
         } else if (reg == PC_REGISTER) {
-            Core::CPU().SetPC(GdbHexToLong(buffer_ptr + i * 16));
+            RegWrite(PC_REGISTER, GdbHexToLong(buffer_ptr + i * 16), current_thread);
         } else if (reg == CPSR_REGISTER) {
-            Core::CPU().SetCPSR(GdbHexToInt(buffer_ptr + i * 16));
+            RegWrite(CPSR_REGISTER, GdbHexToInt(buffer_ptr + i * 16), current_thread);
         } else {
             UNIMPLEMENTED();
         }
@@ -734,7 +843,7 @@ static void WriteMemory() {
 void Break(bool is_memory_break) {
     if (!halt_loop) {
         halt_loop = true;
-        SendSignal(SIGTRAP);
+        send_trap = true;
     }
 
     memory_break = is_memory_break;
@@ -744,10 +853,10 @@ void Break(bool is_memory_break) {
 static void Step() {
     step_loop = true;
     halt_loop = true;
-    step_break = true;
-    SendSignal(SIGTRAP);
+    send_trap = true;
 }
 
+/// Tell the CPU if we hit a memory breakpoint.
 bool IsMemoryBreak() {
     if (IsConnected()) {
         return false;
@@ -759,7 +868,6 @@ bool IsMemoryBreak() {
 /// Tell the CPU to continue executing.
 static void Continue() {
     memory_break = false;
-    step_break = false;
     step_loop = false;
     halt_loop = false;
 }
@@ -898,7 +1006,7 @@ void HandlePacket() {
         HandleSetThread();
         break;
     case '?':
-        SendSignal(latest_signal);
+        SendSignal(current_thread, latest_signal);
         break;
     case 'k':
         Shutdown();
@@ -935,6 +1043,9 @@ void HandlePacket() {
     case 'Z':
         AddBreakpoint();
         break;
+    case 'T':
+        HandleThreadAlive();
+        break;
     default:
         SendReply("");
         break;
@@ -1079,4 +1190,11 @@ bool GetCpuStepFlag() {
 void SetCpuStepFlag(bool is_step) {
     step_loop = is_step;
 }
+
+void SendTrap(Kernel::Thread* thread, int trap) {
+    if (send_trap) {
+        send_trap = false;
+        SendSignal(thread, trap);
+    }
+}
 }; // namespace GDBStub

src/core/gdbstub/gdbstub.h

@@ -7,6 +7,7 @@
 #pragma once
 
 #include "common/common_types.h"
+#include "core/hle/kernel/thread.h"
 
 namespace GDBStub {
 
@@ -91,4 +92,12 @@ bool GetCpuStepFlag();
  * @param is_step
  */
 void SetCpuStepFlag(bool is_step);
+
+/**
+ * Send trap signal from thread back to the gdbstub server.
+ *
+ * @param thread Sending thread.
+ * @param trap Trap no.
+ */
+void SendTrap(Kernel::Thread* thread, int trap);
 } // namespace GDBStub

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;
 }
@@ -346,6 +316,11 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
                       "(STUBBED) Attempted to query privileged process id bounds, returned 0");
         *result = 0;
         break;
+    case GetInfoType::UserExceptionContextAddr:
+        NGLOG_WARNING(Kernel_SVC,
+                      "(STUBBED) Attempted to query user exception context address, returned 0");
+        *result = 0;
+        break;
     default:
         UNIMPLEMENTED();
     }
@@ -395,14 +370,14 @@ 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 +460,28 @@ 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 +511,9 @@ 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 +527,7 @@ 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 +546,17 @@ 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 +568,7 @@ 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 +604,7 @@ 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 +613,43 @@ 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 +682,8 @@ 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 +722,56 @@ 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 = 1ull << 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 & (1ull << 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

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <cinttypes>
+#include <stack>
 #include "core/file_sys/filesystem.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
@@ -74,7 +75,40 @@ 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") {}
 
@@ -103,7 +137,7 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
         {40, &ISelfController::CreateManagedDisplayLayer, "CreateManagedDisplayLayer"},
         {41, nullptr, "IsSystemBufferSharingEnabled"},
         {42, nullptr, "GetSystemSharedLayerHandle"},
-        {50, nullptr, "SetHandlesRequestToDisplay"},
+        {50, &ISelfController::SetHandlesRequestToDisplay, "SetHandlesRequestToDisplay"},
         {51, nullptr, "ApproveToDisplay"},
         {60, nullptr, "OverrideAutoSleepTimeAndDimmingTime"},
         {61, nullptr, "SetMediaPlaybackState"},
@@ -121,7 +155,7 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
     RegisterHandlers(functions);
 
     launchable_event =
-        Kernel::Event::Create(Kernel::ResetType::OneShot, "ISelfController:LaunchableEvent");
+        Kernel::Event::Create(Kernel::ResetType::Sticky, "ISelfController:LaunchableEvent");
 }
 
 void ISelfController::SetFocusHandlingMode(Kernel::HLERequestContext& ctx) {
@@ -229,6 +263,13 @@ void ISelfController::CreateManagedDisplayLayer(Kernel::HLERequestContext& ctx)
     NGLOG_WARNING(Service_AM, "(STUBBED) called");
 }
 
+void ISelfController::SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+
+    NGLOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
 ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter") {
     static const FunctionInfo functions[] = {
         {0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"},
@@ -308,77 +349,13 @@ void ICommonStateGetter::GetPerformanceMode(Kernel::HLERequestContext& ctx) {
     NGLOG_WARNING(Service_AM, "(STUBBED) called");
 }
 
-class ILibraryAppletAccessor final : public ServiceFramework<ILibraryAppletAccessor> {
-public:
-    explicit ILibraryAppletAccessor() : ServiceFramework("ILibraryAppletAccessor") {
-        static const FunctionInfo functions[] = {
-            {0, &ILibraryAppletAccessor::GetAppletStateChangedEvent, "GetAppletStateChangedEvent"},
-            {1, nullptr, "IsCompleted"},
-            {10, nullptr, "Start"},
-            {20, nullptr, "RequestExit"},
-            {25, nullptr, "Terminate"},
-            {30, nullptr, "GetResult"},
-            {50, nullptr, "SetOutOfFocusApplicationSuspendingEnabled"},
-            {100, nullptr, "PushInData"},
-            {101, nullptr, "PopOutData"},
-            {102, nullptr, "PushExtraStorage"},
-            {103, nullptr, "PushInteractiveInData"},
-            {104, nullptr, "PopInteractiveOutData"},
-            {105, nullptr, "GetPopOutDataEvent"},
-            {106, nullptr, "GetPopInteractiveOutDataEvent"},
-            {110, nullptr, "NeedsToExitProcess"},
-            {120, nullptr, "GetLibraryAppletInfo"},
-            {150, nullptr, "RequestForAppletToGetForeground"},
-            {160, nullptr, "GetIndirectLayerConsumerHandle"},
-        };
-        RegisterHandlers(functions);
-
-        state_changed_event = Kernel::Event::Create(Kernel::ResetType::OneShot,
-                                                    "ILibraryAppletAccessor:StateChangedEvent");
-    }
-
-private:
-    void GetAppletStateChangedEvent(Kernel::HLERequestContext& ctx) {
-        state_changed_event->Signal();
-
-        IPC::ResponseBuilder rb{ctx, 2, 1};
-        rb.Push(RESULT_SUCCESS);
-        rb.PushCopyObjects(state_changed_event);
-
-        NGLOG_WARNING(Service_AM, "(STUBBED) called");
-    }
-
-    Kernel::SharedPtr<Kernel::Event> state_changed_event;
-};
-
-ILibraryAppletCreator::ILibraryAppletCreator() : ServiceFramework("ILibraryAppletCreator") {
-    static const FunctionInfo functions[] = {
-        {0, &ILibraryAppletCreator::CreateLibraryApplet, "CreateLibraryApplet"},
-        {1, nullptr, "TerminateAllLibraryApplets"},
-        {2, nullptr, "AreAnyLibraryAppletsLeft"},
-        {10, nullptr, "CreateStorage"},
-        {11, nullptr, "CreateTransferMemoryStorage"},
-        {12, nullptr, "CreateHandleStorage"},
-    };
-    RegisterHandlers(functions);
-}
-
-void ILibraryAppletCreator::CreateLibraryApplet(Kernel::HLERequestContext& ctx) {
-    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
-
-    rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<AM::ILibraryAppletAccessor>();
-
-    NGLOG_DEBUG(Service_AM, "called");
-}
-
 class IStorageAccessor final : public ServiceFramework<IStorageAccessor> {
 public:
     explicit IStorageAccessor(std::vector<u8> buffer)
         : ServiceFramework("IStorageAccessor"), buffer(std::move(buffer)) {
         static const FunctionInfo functions[] = {
             {0, &IStorageAccessor::GetSize, "GetSize"},
-            {10, nullptr, "Write"},
+            {10, &IStorageAccessor::Write, "Write"},
             {11, &IStorageAccessor::Read, "Read"},
         };
         RegisterHandlers(functions);
@@ -396,22 +373,36 @@ private:
         NGLOG_DEBUG(Service_AM, "called");
     }
 
+    void Write(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+
+        const u64 offset{rp.Pop<u64>()};
+        const std::vector<u8> data{ctx.ReadBuffer()};
+
+        ASSERT(offset + data.size() <= buffer.size());
+
+        std::memcpy(&buffer[offset], data.data(), data.size());
+
+        IPC::ResponseBuilder rb{rp.MakeBuilder(2, 0, 0)};
+        rb.Push(RESULT_SUCCESS);
+
+        NGLOG_DEBUG(Service_AM, "called, offset={}", offset);
+    }
+
     void Read(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
 
-        u64 offset = rp.Pop<u64>();
-
+        const u64 offset{rp.Pop<u64>()};
         const size_t size{ctx.GetWriteBufferSize()};
 
         ASSERT(offset + size <= buffer.size());
 
         ctx.WriteBuffer(buffer.data() + offset, size);
 
-        IPC::ResponseBuilder rb{ctx, 2};
-
+        IPC::ResponseBuilder rb{rp.MakeBuilder(2, 0, 0)};
         rb.Push(RESULT_SUCCESS);
 
-        NGLOG_DEBUG(Service_AM, "called");
+        NGLOG_DEBUG(Service_AM, "called, offset={}", offset);
     }
 };
 
@@ -439,17 +430,129 @@ private:
     }
 };
 
+class ILibraryAppletAccessor final : public ServiceFramework<ILibraryAppletAccessor> {
+public:
+    explicit ILibraryAppletAccessor() : ServiceFramework("ILibraryAppletAccessor") {
+        static const FunctionInfo functions[] = {
+            {0, &ILibraryAppletAccessor::GetAppletStateChangedEvent, "GetAppletStateChangedEvent"},
+            {1, nullptr, "IsCompleted"},
+            {10, &ILibraryAppletAccessor::Start, "Start"},
+            {20, nullptr, "RequestExit"},
+            {25, nullptr, "Terminate"},
+            {30, &ILibraryAppletAccessor::GetResult, "GetResult"},
+            {50, nullptr, "SetOutOfFocusApplicationSuspendingEnabled"},
+            {100, &ILibraryAppletAccessor::PushInData, "PushInData"},
+            {101, &ILibraryAppletAccessor::PopOutData, "PopOutData"},
+            {102, nullptr, "PushExtraStorage"},
+            {103, nullptr, "PushInteractiveInData"},
+            {104, nullptr, "PopInteractiveOutData"},
+            {105, nullptr, "GetPopOutDataEvent"},
+            {106, nullptr, "GetPopInteractiveOutDataEvent"},
+            {110, nullptr, "NeedsToExitProcess"},
+            {120, nullptr, "GetLibraryAppletInfo"},
+            {150, nullptr, "RequestForAppletToGetForeground"},
+            {160, nullptr, "GetIndirectLayerConsumerHandle"},
+        };
+        RegisterHandlers(functions);
+
+        state_changed_event = Kernel::Event::Create(Kernel::ResetType::OneShot,
+                                                    "ILibraryAppletAccessor:StateChangedEvent");
+    }
+
+private:
+    void GetAppletStateChangedEvent(Kernel::HLERequestContext& ctx) {
+        state_changed_event->Signal();
+
+        IPC::ResponseBuilder rb{ctx, 2, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushCopyObjects(state_changed_event);
+
+        NGLOG_WARNING(Service_AM, "(STUBBED) called");
+    }
+
+    void GetResult(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+
+        NGLOG_WARNING(Service_AM, "(STUBBED) called");
+    }
+
+    void Start(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+
+        NGLOG_WARNING(Service_AM, "(STUBBED) called");
+    }
+
+    void PushInData(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        storage_stack.push(rp.PopIpcInterface<AM::IStorage>());
+
+        IPC::ResponseBuilder rb{rp.MakeBuilder(2, 0, 0)};
+        rb.Push(RESULT_SUCCESS);
+
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void PopOutData(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<AM::IStorage>(std::move(storage_stack.top()));
+
+        storage_stack.pop();
+
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    std::stack<std::shared_ptr<AM::IStorage>> storage_stack;
+    Kernel::SharedPtr<Kernel::Event> state_changed_event;
+};
+
+ILibraryAppletCreator::ILibraryAppletCreator() : ServiceFramework("ILibraryAppletCreator") {
+    static const FunctionInfo functions[] = {
+        {0, &ILibraryAppletCreator::CreateLibraryApplet, "CreateLibraryApplet"},
+        {1, nullptr, "TerminateAllLibraryApplets"},
+        {2, nullptr, "AreAnyLibraryAppletsLeft"},
+        {10, &ILibraryAppletCreator::CreateStorage, "CreateStorage"},
+        {11, nullptr, "CreateTransferMemoryStorage"},
+        {12, nullptr, "CreateHandleStorage"},
+    };
+    RegisterHandlers(functions);
+}
+
+void ILibraryAppletCreator::CreateLibraryApplet(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+
+    rb.Push(RESULT_SUCCESS);
+    rb.PushIpcInterface<AM::ILibraryAppletAccessor>();
+
+    NGLOG_DEBUG(Service_AM, "called");
+}
+
+void ILibraryAppletCreator::CreateStorage(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const u64 size{rp.Pop<u64>()};
+    std::vector<u8> buffer(size);
+
+    IPC::ResponseBuilder rb{rp.MakeBuilder(2, 0, 1)};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushIpcInterface<AM::IStorage>(std::move(buffer));
+
+    NGLOG_DEBUG(Service_AM, "called, size={}", size);
+}
+
 IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationFunctions") {
     static const FunctionInfo functions[] = {
         {1, &IApplicationFunctions::PopLaunchParameter, "PopLaunchParameter"},
         {10, nullptr, "CreateApplicationAndPushAndRequestToStart"},
         {11, nullptr, "CreateApplicationAndPushAndRequestToStartForQuest"},
         {12, nullptr, "CreateApplicationAndRequestToStart"},
-        {13, nullptr, "CreateApplicationAndRequestToStartForQuest"},
+        {13, &IApplicationFunctions::CreateApplicationAndRequestToStartForQuest,
+         "CreateApplicationAndRequestToStartForQuest"},
         {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"},
@@ -458,7 +561,7 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
         {32, nullptr, "BeginBlockingHomeButton"},
         {33, nullptr, "EndBlockingHomeButton"},
         {40, &IApplicationFunctions::NotifyRunning, "NotifyRunning"},
-        {50, nullptr, "GetPseudoDeviceId"},
+        {50, &IApplicationFunctions::GetPseudoDeviceId, "GetPseudoDeviceId"},
         {60, nullptr, "SetMediaPlaybackStateForApplication"},
         {65, nullptr, "IsGamePlayRecordingSupported"},
         {66, &IApplicationFunctions::InitializeGamePlayRecording, "InitializeGamePlayRecording"},
@@ -502,6 +605,13 @@ void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) {
     NGLOG_DEBUG(Service_AM, "called");
 }
 
+void IApplicationFunctions::CreateApplicationAndRequestToStartForQuest(
+    Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    NGLOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
 void IApplicationFunctions::EnsureSaveData(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     u128 uid = rp.PopRaw<u128>();
@@ -537,6 +647,14 @@ 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};
@@ -566,10 +684,81 @@ void IApplicationFunctions::NotifyRunning(Kernel::HLERequestContext& ctx) {
     NGLOG_WARNING(Service_AM, "(STUBBED) called");
 }
 
+void IApplicationFunctions::GetPseudoDeviceId(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 6};
+    rb.Push(RESULT_SUCCESS);
+
+    // Returns a 128-bit UUID
+    rb.Push<u64>(0);
+    rb.Push<u64>(0);
+
+    NGLOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
 void InstallInterfaces(SM::ServiceManager& service_manager,
                        std::shared_ptr<NVFlinger::NVFlinger> nvflinger) {
     std::make_shared<AppletAE>(nvflinger)->InstallAsService(service_manager);
     std::make_shared<AppletOE>(nvflinger)->InstallAsService(service_manager);
 }
 
+IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions") {
+    static const FunctionInfo functions[] = {
+        {10, &IHomeMenuFunctions::RequestToGetForeground, "RequestToGetForeground"},
+        {11, nullptr, "LockForeground"},
+        {12, nullptr, "UnlockForeground"},
+        {20, nullptr, "PopFromGeneralChannel"},
+        {21, nullptr, "GetPopFromGeneralChannelEvent"},
+        {30, nullptr, "GetHomeButtonWriterLockAccessor"},
+        {31, nullptr, "GetWriterLockAccessorEx"},
+    };
+    RegisterHandlers(functions);
+}
+
+void IHomeMenuFunctions::RequestToGetForeground(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    NGLOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
+IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStateController") {
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "RequestToEnterSleep"},
+        {1, nullptr, "EnterSleep"},
+        {2, nullptr, "StartSleepSequence"},
+        {3, nullptr, "StartShutdownSequence"},
+        {4, nullptr, "StartRebootSequence"},
+        {10, nullptr, "LoadAndApplyIdlePolicySettings"},
+        {11, nullptr, "NotifyCecSettingsChanged"},
+        {12, nullptr, "SetDefaultHomeButtonLongPressTime"},
+        {13, nullptr, "UpdateDefaultDisplayResolution"},
+        {14, nullptr, "ShouldSleepOnBoot"},
+        {15, nullptr, "GetHdcpAuthenticationFailedEvent"},
+    };
+    RegisterHandlers(functions);
+}
+
+IApplicationCreator::IApplicationCreator() : ServiceFramework("IApplicationCreator") {
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "CreateApplication"},
+        {1, nullptr, "PopLaunchRequestedApplication"},
+        {10, nullptr, "CreateSystemApplication"},
+        {100, nullptr, "PopFloatingApplicationForDevelopment"},
+    };
+    RegisterHandlers(functions);
+}
+
+IProcessWindingController::IProcessWindingController()
+    : ServiceFramework("IProcessWindingController") {
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetLaunchReason"},
+        {11, nullptr, "OpenCallingLibraryApplet"},
+        {21, nullptr, "PushContext"},
+        {22, nullptr, "PopContext"},
+        {23, nullptr, "CancelWindingReservation"},
+        {30, nullptr, "WindAndDoReserved"},
+        {40, nullptr, "ReserveToStartAndWaitAndUnwindThis"},
+        {41, nullptr, "ReserveToStartAndWait"},
+    };
+    RegisterHandlers(functions);
+}
 } // namespace Service::AM

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> {
@@ -70,6 +85,7 @@ private:
     void GetLibraryAppletLaunchableEvent(Kernel::HLERequestContext& ctx);
     void CreateManagedDisplayLayer(Kernel::HLERequestContext& ctx);
     void SetScreenShotPermission(Kernel::HLERequestContext& ctx);
+    void SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
     Kernel::SharedPtr<Kernel::Event> launchable_event;
@@ -105,6 +121,7 @@ public:
 
 private:
     void CreateLibraryApplet(Kernel::HLERequestContext& ctx);
+    void CreateStorage(Kernel::HLERequestContext& ctx);
 };
 
 class IApplicationFunctions final : public ServiceFramework<IApplicationFunctions> {
@@ -113,12 +130,38 @@ public:
 
 private:
     void PopLaunchParameter(Kernel::HLERequestContext& ctx);
+    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);
     void NotifyRunning(Kernel::HLERequestContext& ctx);
+    void GetPseudoDeviceId(Kernel::HLERequestContext& ctx);
+};
+
+class IHomeMenuFunctions final : public ServiceFramework<IHomeMenuFunctions> {
+public:
+    IHomeMenuFunctions();
+
+private:
+    void RequestToGetForeground(Kernel::HLERequestContext& ctx);
+};
+
+class IGlobalStateController final : public ServiceFramework<IGlobalStateController> {
+public:
+    IGlobalStateController();
+};
+
+class IApplicationCreator final : public ServiceFramework<IApplicationCreator> {
+public:
+    IApplicationCreator();
+};
+
+class IProcessWindingController final : public ServiceFramework<IProcessWindingController> {
+public:
+    IProcessWindingController();
 };
 
 /// Registers all AM services with the specified service manager.

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

@@ -20,7 +20,7 @@ public:
             {2, &ILibraryAppletProxy::GetWindowController, "GetWindowController"},
             {3, &ILibraryAppletProxy::GetAudioController, "GetAudioController"},
             {4, &ILibraryAppletProxy::GetDisplayController, "GetDisplayController"},
-            {10, nullptr, "GetProcessWindingController"},
+            {10, &ILibraryAppletProxy::GetProcessWindingController, "GetProcessWindingController"},
             {11, &ILibraryAppletProxy::GetLibraryAppletCreator, "GetLibraryAppletCreator"},
             {20, &ILibraryAppletProxy::GetApplicationFunctions, "GetApplicationFunctions"},
             {1000, &ILibraryAppletProxy::GetDebugFunctions, "GetDebugFunctions"},
@@ -28,6 +28,93 @@ public:
         RegisterHandlers(functions);
     }
 
+private:
+    void GetCommonStateGetter(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<ICommonStateGetter>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetSelfController(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<ISelfController>(nvflinger);
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetWindowController(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IWindowController>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetAudioController(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IAudioController>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetDisplayController(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IDisplayController>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetProcessWindingController(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IProcessWindingController>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetDebugFunctions(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IDebugFunctions>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetLibraryAppletCreator(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<ILibraryAppletCreator>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetApplicationFunctions(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IApplicationFunctions>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
+};
+
+class ISystemAppletProxy final : public ServiceFramework<ISystemAppletProxy> {
+public:
+    explicit ISystemAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
+        : ServiceFramework("ISystemAppletProxy"), nvflinger(std::move(nvflinger)) {
+        static const FunctionInfo functions[] = {
+            {0, &ISystemAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"},
+            {1, &ISystemAppletProxy::GetSelfController, "GetSelfController"},
+            {2, &ISystemAppletProxy::GetWindowController, "GetWindowController"},
+            {3, &ISystemAppletProxy::GetAudioController, "GetAudioController"},
+            {4, &ISystemAppletProxy::GetDisplayController, "GetDisplayController"},
+            {10, nullptr, "GetProcessWindingController"},
+            {11, &ISystemAppletProxy::GetLibraryAppletCreator, "GetLibraryAppletCreator"},
+            {20, &ISystemAppletProxy::GetHomeMenuFunctions, "GetHomeMenuFunctions"},
+            {21, &ISystemAppletProxy::GetGlobalStateController, "GetGlobalStateController"},
+            {22, &ISystemAppletProxy::GetApplicationCreator, "GetApplicationCreator"},
+            {1000, &ISystemAppletProxy::GetDebugFunctions, "GetDebugFunctions"},
+        };
+        RegisterHandlers(functions);
+    }
+
 private:
     void GetCommonStateGetter(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 0, 1};
@@ -78,16 +165,43 @@ private:
         NGLOG_DEBUG(Service_AM, "called");
     }
 
-    void GetApplicationFunctions(Kernel::HLERequestContext& ctx) {
+    void GetHomeMenuFunctions(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 0, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushIpcInterface<IApplicationFunctions>();
+        rb.PushIpcInterface<IHomeMenuFunctions>();
         NGLOG_DEBUG(Service_AM, "called");
     }
 
+    void GetGlobalStateController(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IGlobalStateController>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
+
+    void GetApplicationCreator(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IApplicationCreator>();
+        NGLOG_DEBUG(Service_AM, "called");
+    }
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
 };
 
+void AppletAE::OpenSystemAppletProxy(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushIpcInterface<ISystemAppletProxy>(nvflinger);
+    NGLOG_DEBUG(Service_AM, "called");
+}
+
+void AppletAE::OpenLibraryAppletProxy(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger);
+    NGLOG_DEBUG(Service_AM, "called");
+}
+
 void AppletAE::OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
@@ -98,9 +212,9 @@ void AppletAE::OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx) {
 AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     : ServiceFramework("appletAE"), nvflinger(std::move(nvflinger)) {
     static const FunctionInfo functions[] = {
-        {100, nullptr, "OpenSystemAppletProxy"},
+        {100, &AppletAE::OpenSystemAppletProxy, "OpenSystemAppletProxy"},
         {200, &AppletAE::OpenLibraryAppletProxyOld, "OpenLibraryAppletProxyOld"},
-        {201, nullptr, "OpenLibraryAppletProxy"},
+        {201, &AppletAE::OpenLibraryAppletProxy, "OpenLibraryAppletProxy"},
         {300, nullptr, "OpenOverlayAppletProxy"},
         {350, nullptr, "OpenSystemApplicationProxy"},
         {400, nullptr, "CreateSelfLibraryAppletCreatorForDevelop"},

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

@@ -21,6 +21,8 @@ public:
     ~AppletAE() = default;
 
 private:
+    void OpenSystemAppletProxy(Kernel::HLERequestContext& ctx);
+    void OpenLibraryAppletProxy(Kernel::HLERequestContext& ctx);
     void OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;

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"
@@ -56,27 +57,26 @@ private:
     }
 
     void RequestUpdateAudioRenderer(Kernel::HLERequestContext& ctx) {
-        NGLOG_DEBUG(Service_Audio, "{}", ctx.Description());
-        AudioRendererResponseData response_data{};
+        AudioRendererConfig config;
+        auto buf = ctx.ReadBuffer();
+        std::memcpy(&config, buf.data(), sizeof(AudioRendererConfig));
 
-        response_data.section_0_size =
-            static_cast<u32>(response_data.state_entries.size() * sizeof(AudioRendererStateEntry));
-        response_data.section_1_size = static_cast<u32>(response_data.section_1.size());
-        response_data.section_2_size = static_cast<u32>(response_data.section_2.size());
-        response_data.section_3_size = static_cast<u32>(response_data.section_3.size());
-        response_data.section_4_size = static_cast<u32>(response_data.section_4.size());
-        response_data.section_5_size = static_cast<u32>(response_data.section_5.size());
-        response_data.total_size = sizeof(AudioRendererResponseData);
+        AudioRendererResponse response_data{config};
 
-        for (unsigned i = 0; i < response_data.state_entries.size(); i++) {
-            // 4 = Busy and 5 = Ready?
-            response_data.state_entries[i].state = 5;
+        ASSERT(ctx.GetWriteBufferSize() == response_data.total_size);
+
+        std::vector<u8> output(response_data.total_size);
+        std::memcpy(output.data(), &response_data, sizeof(AudioRendererResponse));
+        std::vector<MemoryPoolEntry> memory_pool(config.memory_pools_size / 0x20);
+        for (auto& entry : memory_pool) {
+            entry.state = 5;
         }
+        std::memcpy(output.data() + sizeof(AudioRendererResponse), memory_pool.data(),
+                    response_data.memory_pools_size);
 
-        ctx.WriteBuffer(&response_data, response_data.total_size);
+        ctx.WriteBuffer(output);
 
         IPC::ResponseBuilder rb{ctx, 2};
-
         rb.Push(RESULT_SUCCESS);
 
         NGLOG_WARNING(Service_Audio, "(STUBBED) called");
@@ -108,43 +108,55 @@ private:
         NGLOG_WARNING(Service_Audio, "(STUBBED) called");
     }
 
-    struct AudioRendererStateEntry {
+    struct MemoryPoolEntry {
         u32_le state;
         u32_le unknown_4;
         u32_le unknown_8;
         u32_le unknown_c;
     };
-    static_assert(sizeof(AudioRendererStateEntry) == 0x10,
-                  "AudioRendererStateEntry has wrong size");
+    static_assert(sizeof(MemoryPoolEntry) == 0x10, "MemoryPoolEntry has wrong size");
 
-    struct AudioRendererResponseData {
-        u32_le unknown_0;
-        u32_le section_5_size;
-        u32_le section_0_size;
-        u32_le section_1_size;
-        u32_le unknown_10;
-        u32_le section_2_size;
-        u32_le unknown_18;
-        u32_le section_3_size;
-        u32_le section_4_size;
-        u32_le unknown_24;
-        u32_le unknown_28;
-        u32_le unknown_2c;
-        u32_le unknown_30;
-        u32_le unknown_34;
-        u32_le unknown_38;
-        u32_le total_size;
-
-        std::array<AudioRendererStateEntry, 0x18e> state_entries;
-
-        std::array<u8, 0x600> section_1;
-        std::array<u8, 0xe0> section_2;
-        std::array<u8, 0x20> section_3;
-        std::array<u8, 0x10> section_4;
-        std::array<u8, 0xb0> section_5;
+    struct AudioRendererConfig {
+        u32 revision;
+        u32 behavior_size;
+        u32 memory_pools_size;
+        u32 voices_size;
+        u32 voice_resource_size;
+        u32 effects_size;
+        u32 mixes_size;
+        u32 sinks_size;
+        u32 performance_buffer_size;
+        INSERT_PADDING_WORDS(6);
+        u32 total_size;
     };
-    static_assert(sizeof(AudioRendererResponseData) == 0x20e0,
-                  "AudioRendererResponseData has wrong size");
+    static_assert(sizeof(AudioRendererConfig) == 0x40, "AudioRendererConfig has wrong size");
+
+    struct AudioRendererResponse {
+        AudioRendererResponse(const AudioRendererConfig& config) {
+            revision = config.revision;
+            error_info_size = 0xb0;
+            memory_pools_size = (config.memory_pools_size / 0x20) * 0x10;
+            voices_size = (config.voices_size / 0x170) * 0x10;
+            effects_size = (config.effects_size / 0xC0) * 0x10;
+            sinks_size = (config.sinks_size / 0x140) * 0x20;
+            performance_manager_size = 0x10;
+            total_size = sizeof(AudioRendererResponse) + error_info_size + memory_pools_size +
+                         voices_size + effects_size + sinks_size + performance_manager_size;
+        }
+
+        u32_le revision;
+        u32_le error_info_size;
+        u32_le memory_pools_size;
+        u32_le voices_size;
+        u32_le unknown_10;
+        u32_le effects_size;
+        u32_le unknown_18;
+        u32_le sinks_size;
+        u32_le performance_manager_size;
+        INSERT_PADDING_WORDS(6);
+        u32_le total_size;
+    };
+    static_assert(sizeof(AudioRendererResponse) == 0x40, "AudioRendererResponse has wrong size");
 
     /// This is used to trigger the audio event callback.
     CoreTiming::EventType* audio_event;
@@ -156,19 +168,20 @@ 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 +268,62 @@ void AudRenU::OpenAudioRenderer(Kernel::HLERequestContext& ctx) {
 }
 
 void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    auto params = rp.PopRaw<AudioRendererParameters>();
+
+    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 +335,14 @@ 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 >= 2u;
+    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 AudioRendererParameters {
+        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(AudioRendererParameters) == 52,
+                  "AudioRendererParameters 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/filesystem/fsp_srv.cpp

@@ -4,6 +4,7 @@
 
 #include <cinttypes>
 #include "common/logging/log.h"
+#include "common/string_util.h"
 #include "core/core.h"
 #include "core/file_sys/directory.h"
 #include "core/file_sys/filesystem.h"
@@ -258,9 +259,7 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
-
-        std::string name(file_buffer.begin(), end);
+        std::string name = Common::StringFromBuffer(file_buffer);
 
         u64 mode = rp.Pop<u64>();
         u32 size = rp.Pop<u32>();
@@ -275,9 +274,7 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
-
-        std::string name(file_buffer.begin(), end);
+        std::string name = Common::StringFromBuffer(file_buffer);
 
         NGLOG_DEBUG(Service_FS, "called file {}", name);
 
@@ -289,9 +286,7 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
-
-        std::string name(file_buffer.begin(), end);
+        std::string name = Common::StringFromBuffer(file_buffer);
 
         NGLOG_DEBUG(Service_FS, "called directory {}", name);
 
@@ -305,13 +300,11 @@ public:
         std::vector<u8> buffer;
         buffer.resize(ctx.BufferDescriptorX()[0].Size());
         Memory::ReadBlock(ctx.BufferDescriptorX()[0].Address(), buffer.data(), buffer.size());
-        auto end = std::find(buffer.begin(), buffer.end(), '\0');
-        std::string src_name(buffer.begin(), end);
+        std::string src_name = Common::StringFromBuffer(buffer);
 
         buffer.resize(ctx.BufferDescriptorX()[1].Size());
         Memory::ReadBlock(ctx.BufferDescriptorX()[1].Address(), buffer.data(), buffer.size());
-        end = std::find(buffer.begin(), buffer.end(), '\0');
-        std::string dst_name(buffer.begin(), end);
+        std::string dst_name = Common::StringFromBuffer(buffer);
 
         NGLOG_DEBUG(Service_FS, "called file '{}' to file '{}'", src_name, dst_name);
 
@@ -323,9 +316,7 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
-
-        std::string name(file_buffer.begin(), end);
+        std::string name = Common::StringFromBuffer(file_buffer);
 
         auto mode = static_cast<FileSys::Mode>(rp.Pop<u32>());
 
@@ -349,9 +340,7 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
-
-        std::string name(file_buffer.begin(), end);
+        std::string name = Common::StringFromBuffer(file_buffer);
 
         // TODO(Subv): Implement this filter.
         u32 filter_flags = rp.Pop<u32>();
@@ -376,9 +365,7 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
-
-        std::string name(file_buffer.begin(), end);
+        std::string name = Common::StringFromBuffer(file_buffer);
 
         NGLOG_DEBUG(Service_FS, "called file {}", name);
 

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

@@ -84,6 +84,10 @@ private:
 
         for (size_t controller = 0; controller < mem.controllers.size(); controller++) {
             for (int index = 0; index < HID_NUM_LAYOUTS; index++) {
+                // TODO(DarkLordZach): Is this layout/controller config actually invalid?
+                if (controller == Controller_Handheld && index == Layout_Single)
+                    continue;
+
                 ControllerLayout& layout = mem.controllers[controller].layouts[index];
                 layout.header.num_entries = HID_NUM_ENTRIES;
                 layout.header.max_entry_index = HID_NUM_ENTRIES - 1;
@@ -94,7 +98,6 @@ private:
                 layout.header.latest_entry = (layout.header.latest_entry + 1) % HID_NUM_ENTRIES;
 
                 ControllerInputEntry& entry = layout.entries[layout.header.latest_entry];
-                entry.connection_state = ConnectionState_Connected | ConnectionState_Wired;
                 entry.timestamp++;
                 // TODO(shinyquagsire23): Is this always identical to timestamp?
                 entry.timestamp_2++;
@@ -103,6 +106,8 @@ private:
                 if (controller != Controller_Handheld)
                     continue;
 
+                entry.connection_state = ConnectionState_Connected | ConnectionState_Wired;
+
                 // TODO(shinyquagsire23): Set up some LUTs for each layout mapping in the future?
                 // For now everything is just the default handheld layout, but split Joy-Con will
                 // rotate the face buttons and directions for certain layouts.
@@ -329,6 +334,7 @@ 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 +342,41 @@ 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/mm/mm_u.cpp

@@ -0,0 +1,50 @@
+// 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/kernel/client_session.h"
+#include "core/hle/service/mm/mm_u.h"
+
+namespace Service::MM {
+
+void InstallInterfaces(SM::ServiceManager& service_manager) {
+    std::make_shared<MM_U>()->InstallAsService(service_manager);
+}
+
+void MM_U::Initialize(Kernel::HLERequestContext& ctx) {
+    NGLOG_WARNING(Service_MM, "(STUBBED) called");
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void MM_U::SetAndWait(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    min = rp.Pop<u32>();
+    max = rp.Pop<u32>();
+    current = min;
+
+    NGLOG_WARNING(Service_MM, "(STUBBED) called, min=0x{:X}, max=0x{:X}", min, max);
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void MM_U::Get(Kernel::HLERequestContext& ctx) {
+    NGLOG_WARNING(Service_MM, "(STUBBED) called");
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push(current);
+}
+
+MM_U::MM_U() : ServiceFramework("mm:u") {
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "InitializeOld"},        {1, nullptr, "FinalizeOld"},
+        {2, nullptr, "SetAndWaitOld"},        {3, nullptr, "GetOld"},
+        {4, &MM_U::Initialize, "Initialize"}, {5, nullptr, "Finalize"},
+        {6, &MM_U::SetAndWait, "SetAndWait"}, {7, &MM_U::Get, "Get"},
+    };
+    RegisterHandlers(functions);
+}
+
+} // namespace Service::MM

src/core/hle/service/mm/mm_u.h

@@ -0,0 +1,29 @@
+// 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::MM {
+
+class MM_U final : public ServiceFramework<MM_U> {
+public:
+    MM_U();
+    ~MM_U() = default;
+
+private:
+    void Initialize(Kernel::HLERequestContext& ctx);
+    void SetAndWait(Kernel::HLERequestContext& ctx);
+    void Get(Kernel::HLERequestContext& ctx);
+
+    u32 min{0};
+    u32 max{0};
+    u32 current{0};
+};
+
+/// Registers all MM services with the specified service manager.
+void InstallInterfaces(SM::ServiceManager& service_manager);
+
+} // namespace Service::MM

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

@@ -4,6 +4,8 @@
 
 #include "common/logging/log.h"
 #include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/event.h"
+#include "core/hle/service/hid/hid.h"
 #include "core/hle/service/nfp/nfp.h"
 #include "core/hle/service/nfp/nfp_user.h"
 
@@ -17,40 +19,132 @@ 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, &IUser::ListDevices, "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, &IUser::AttachActivateEvent, "AttachActivateEvent"},
+            {18, &IUser::AttachDeactivateEvent, "AttachDeactivateEvent"},
+            {19, &IUser::GetState, "GetState"},
+            {20, &IUser::GetDeviceState, "GetDeviceState"},
+            {21, &IUser::GetNpadId, "GetNpadId"},
+            {22, nullptr, "GetApplicationArea2"},
+            {23, &IUser::AttachAvailabilityChangeEvent, "AttachAvailabilityChangeEvent"},
+            {24, nullptr, "RecreateApplicationArea"},
         };
         RegisterHandlers(functions);
+
+        activate_event = Kernel::Event::Create(Kernel::ResetType::OneShot, "IUser:ActivateEvent");
+        deactivate_event =
+            Kernel::Event::Create(Kernel::ResetType::OneShot, "IUser:DeactivateEvent");
+        availability_change_event =
+            Kernel::Event::Create(Kernel::ResetType::OneShot, "IUser:AvailabilityChangeEvent");
     }
 
 private:
+    enum class State : u32 {
+        NonInitialized = 0,
+        Initialized = 1,
+    };
+
+    enum class DeviceState : u32 {
+        Initialized = 0,
+    };
+
     void Initialize(Kernel::HLERequestContext& ctx) {
         NGLOG_WARNING(Service_NFP, "(STUBBED) called");
+
+        state = State::Initialized;
+
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
     }
+
+    void ListDevices(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const u32 array_size = rp.Pop<u32>();
+
+        ctx.WriteBuffer(&device_handle, sizeof(device_handle));
+
+        NGLOG_WARNING(Service_NFP, "(STUBBED) called, array_size={}", array_size);
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(0);
+    }
+
+    void AttachActivateEvent(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const u64 dev_handle = rp.Pop<u64>();
+        NGLOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+
+        IPC::ResponseBuilder rb{ctx, 2, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushCopyObjects(activate_event);
+    }
+
+    void AttachDeactivateEvent(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const u64 dev_handle = rp.Pop<u64>();
+        NGLOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+
+        IPC::ResponseBuilder rb{ctx, 2, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushCopyObjects(deactivate_event);
+    }
+
+    void GetState(Kernel::HLERequestContext& ctx) {
+        NGLOG_WARNING(Service_NFP, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(static_cast<u32>(state));
+    }
+
+    void GetDeviceState(Kernel::HLERequestContext& ctx) {
+        NGLOG_WARNING(Service_NFP, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(static_cast<u32>(device_state));
+    }
+
+    void GetNpadId(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const u64 dev_handle = rp.Pop<u64>();
+        NGLOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(npad_id);
+    }
+
+    void AttachAvailabilityChangeEvent(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const u64 dev_handle = rp.Pop<u64>();
+        NGLOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+
+        IPC::ResponseBuilder rb{ctx, 2, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushCopyObjects(availability_change_event);
+    }
+
+    const u64 device_handle{0xDEAD};
+    const HID::ControllerID npad_id{HID::Controller_Player1};
+    State state{State::NonInitialized};
+    DeviceState device_state{DeviceState::Initialized};
+    Kernel::SharedPtr<Kernel::Event> activate_event;
+    Kernel::SharedPtr<Kernel::Event> deactivate_event;
+    Kernel::SharedPtr<Kernel::Event> availability_change_event;
 };
 
 void Module::Interface::CreateUserInterface(Kernel::HLERequestContext& ctx) {

src/core/hle/service/nifm/nifm.cpp

@@ -38,7 +38,7 @@ public:
             {8, nullptr, "SetPriority"},
             {9, nullptr, "SetNetworkProfileId"},
             {10, nullptr, "SetRejectable"},
-            {11, nullptr, "SetConnectionConfirmationOption"},
+            {11, &IRequest::SetConnectionConfirmationOption, "SetConnectionConfirmationOption"},
             {12, nullptr, "SetPersistent"},
             {13, nullptr, "SetInstant"},
             {14, nullptr, "SetSustainable"},
@@ -67,23 +67,32 @@ private:
         rb.Push(RESULT_SUCCESS);
         rb.Push<u32>(0);
     }
+
     void GetResult(Kernel::HLERequestContext& ctx) {
         NGLOG_WARNING(Service_NIFM, "(STUBBED) called");
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
     }
+
     void GetSystemEventReadableHandles(Kernel::HLERequestContext& ctx) {
         NGLOG_WARNING(Service_NIFM, "(STUBBED) called");
         IPC::ResponseBuilder rb{ctx, 2, 2};
         rb.Push(RESULT_SUCCESS);
         rb.PushCopyObjects(event1, event2);
     }
+
     void Cancel(Kernel::HLERequestContext& ctx) {
         NGLOG_WARNING(Service_NIFM, "(STUBBED) called");
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
     }
 
+    void SetConnectionConfirmationOption(Kernel::HLERequestContext& ctx) {
+        NGLOG_WARNING(Service_NIFM, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
     Kernel::SharedPtr<Kernel::Event> event1, event2;
 };
 

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 +24,12 @@ 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);
+    case IoctlCommand::IocZbcQueryTable:
+        return ZBCQueryTable(input, output);
+    case IoctlCommand::IocFlushL2:
+        return FlushL2(input, output);
     }
     UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
@@ -77,9 +83,13 @@ 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 +131,31 @@ 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;
+}
+
+u32 nvhost_ctrl_gpu::ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output) {
+    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called");
+    IoctlZbcQueryTable params{};
+    std::memcpy(&params, input.data(), input.size());
+    // TODO : To implement properly
+    std::memcpy(output.data(), &params, output.size());
+    return 0;
+}
+
+u32 nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>& output) {
+    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called");
+    IoctlFlushL2 params{};
+    std::memcpy(&params, input.data(), input.size());
+    // TODO : To implement properly
+    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 +25,19 @@ private:
         IocGetActiveSlotMaskCommand = 0x80084714,
         IocZcullGetCtxSizeCommand = 0x80044701,
         IocZcullGetInfo = 0x80284702,
+        IocZbcSetTable = 0x402C4703,
+        IocZbcQueryTable = 0xC0344704,
+        IocFlushL2 = 0x40084707,
+        IocInvalICache = 0x4008470D,
+        IocSetMmudebugMode = 0x4008470E,
+        IocSetSmDebugMode = 0x4010470F,
+        IocWaitForPause = 0xC0084710,
+        IocGetTcpExceptionEnStatus = 0x80084711,
+        IocNumVsms = 0x80084712,
+        IocVsmsMapping = 0xC0044713,
+        IocGetErrorChannelUserData = 0xC008471B,
+        IocGetGpuTime = 0xC010471C,
+        IocGetCpuTimeCorrelationInfo = 0xC108471D,
     };
 
     struct IoctlGpuCharacteristics {
@@ -86,7 +99,7 @@ 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 +130,40 @@ 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");
+
+    struct IoctlZbcQueryTable {
+        u32_le color_ds[4];
+        u32_le color_l2[4];
+        u32_le depth;
+        u32_le ref_cnt;
+        u32_le format;
+        u32_le type;
+        u32_le index_size;
+    };
+    static_assert(sizeof(IoctlZbcQueryTable) == 52, "IoctlZbcQueryTable is incorrect size");
+
+    struct IoctlFlushL2 {
+        u32_le flush; // l2_flush | l2_invalidate << 1 | fb_flush << 2
+        u32_le reserved;
+    };
+    static_assert(sizeof(IoctlFlushL2) == 8, "IoctlFlushL2 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);
+    u32 ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 FlushL2(const std::vector<u8>& input, std::vector<u8>& output);
 };
 
 } // namespace Service::Nvidia::Devices

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

@@ -32,6 +32,10 @@ u32 nvhost_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u
         return AllocGPFIFOEx2(input, output);
     case IoctlCommand::IocAllocObjCtxCommand:
         return AllocateObjectContext(input, output);
+    case IoctlCommand::IocChannelGetWaitbaseCommand:
+        return GetWaitbase(input, output);
+    case IoctlCommand::IocChannelSetTimeoutCommand:
+        return ChannelSetTimeout(input, output);
     }
 
     if (command.group == NVGPU_IOCTL_MAGIC) {
@@ -117,8 +121,9 @@ u32 nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector<
 }
 
 u32 nvhost_gpu::SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& output) {
-    if (input.size() < sizeof(IoctlSubmitGpfifo))
+    if (input.size() < sizeof(IoctlSubmitGpfifo)) {
         UNIMPLEMENTED();
+    }
     IoctlSubmitGpfifo params{};
     std::memcpy(&params, input.data(), sizeof(IoctlSubmitGpfifo));
     NGLOG_WARNING(Service_NVDRV, "(STUBBED) called, gpfifo={:X}, num_entries={:X}, flags={:X}",
@@ -138,4 +143,20 @@ u32 nvhost_gpu::SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& outp
     return 0;
 }
 
+u32 nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlGetWaitbase params{};
+    std::memcpy(&params, input.data(), sizeof(IoctlGetWaitbase));
+    NGLOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown);
+    params.value = 0; // Seems to be hard coded at 0
+    std::memcpy(output.data(), &params, output.size());
+    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

@@ -26,13 +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 {
@@ -49,6 +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;
     };
@@ -69,12 +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
@@ -117,7 +171,13 @@ private:
         IoctlFence fence_out; // returned new fence object for others to wait on
     };
     static_assert(sizeof(IoctlSubmitGpfifo) == 16 + sizeof(IoctlFence),
-                  "submit_gpfifo is incorrect size");
+                  "IoctlSubmitGpfifo is incorrect size");
+
+    struct IoctlGetWaitbase {
+        u32 unknown; // seems to be ignored? Nintendo added this
+        u32 value;
+    };
+    static_assert(sizeof(IoctlGetWaitbase) == 8, "IoctlGetWaitbase is incorrect size");
 
     u32_le nvmap_fd{};
     u64_le user_data{};
@@ -133,6 +193,8 @@ private:
     u32 AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output);
     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

@@ -20,11 +20,13 @@
 #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"
 #include "core/hle/service/hid/hid.h"
 #include "core/hle/service/lm/lm.h"
+#include "core/hle/service/mm/mm_u.h"
 #include "core/hle/service/nfp/nfp.h"
 #include "core/hle/service/nifm/nifm.h"
 #include "core/hle/service/ns/ns.h"
@@ -144,10 +146,12 @@ 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;
@@ -181,12 +185,14 @@ 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);
     Friend::InstallInterfaces(*sm);
     HID::InstallInterfaces(*sm);
     LM::InstallInterfaces(*sm);
+    MM::InstallInterfaces(*sm);
     NFP::InstallInterfaces(*sm);
     NIFM::InstallInterfaces(*sm);
     NS::InstallInterfaces(*sm);

src/core/hle/service/set/set.cpp

@@ -12,9 +12,6 @@
 namespace Service::Set {
 
 void SET::GetAvailableLanguageCodes(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
-    u32 id = rp.Pop<u32>();
-
     static constexpr std::array<LanguageCode, 17> available_language_codes = {{
         LanguageCode::JA,
         LanguageCode::EN_US,
@@ -50,7 +47,7 @@ SET::SET() : ServiceFramework("set") {
         {2, nullptr, "MakeLanguageCode"},
         {3, nullptr, "GetAvailableLanguageCodeCount"},
         {4, nullptr, "GetRegionCode"},
-        {5, nullptr, "GetAvailableLanguageCodes2"},
+        {5, &SET::GetAvailableLanguageCodes, "GetAvailableLanguageCodes2"},
         {6, nullptr, "GetAvailableLanguageCodeCount2"},
         {7, nullptr, "GetKeyCodeMap"},
         {8, nullptr, "GetQuestFlag"},

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/CMakeLists.txt

@@ -9,6 +9,8 @@ add_library(video_core STATIC
     engines/maxwell_3d.h
     engines/maxwell_compute.cpp
     engines/maxwell_compute.h
+    engines/maxwell_dma.cpp
+    engines/maxwell_dma.h
     engines/shader_bytecode.h
     gpu.cpp
     gpu.h
@@ -39,6 +41,8 @@ add_library(video_core STATIC
     renderer_opengl/maxwell_to_gl.h
     renderer_opengl/renderer_opengl.cpp
     renderer_opengl/renderer_opengl.h
+    textures/astc.cpp
+    textures/astc.h
     textures/decoders.cpp
     textures/decoders.h
     textures/texture.h

src/video_core/command_processor.cpp

@@ -16,6 +16,7 @@
 #include "video_core/engines/fermi_2d.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/engines/maxwell_compute.h"
+#include "video_core/engines/maxwell_dma.h"
 #include "video_core/gpu.h"
 #include "video_core/renderer_base.h"
 #include "video_core/video_core.h"
@@ -36,7 +37,6 @@ 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;
     }
@@ -61,8 +61,11 @@ void GPU::WriteReg(u32 method, u32 subchannel, u32 value, u32 remaining_params)
     case EngineID::MAXWELL_COMPUTE_B:
         maxwell_compute->WriteReg(method, value);
         break;
+    case EngineID::MAXWELL_DMA_COPY_A:
+        maxwell_dma->WriteReg(method, value);
+        break;
     default:
-        UNIMPLEMENTED();
+        UNIMPLEMENTED_MSG("Unimplemented engine");
     }
 }
 

src/video_core/engines/fermi_2d.cpp

@@ -47,6 +47,7 @@ void Fermi2D::HandleSurfaceCopy() {
 
     if (regs.src.linear == regs.dst.linear) {
         // If the input layout and the output layout are the same, just perform a raw copy.
+        ASSERT(regs.src.BlockHeight() == regs.dst.BlockHeight());
         Memory::CopyBlock(dest_cpu, source_cpu,
                           src_bytes_per_pixel * regs.dst.width * regs.dst.height);
         return;

src/video_core/engines/maxwell_3d.cpp

@@ -156,16 +156,15 @@ void Maxwell3D::ProcessQueryGet() {
     // TODO(Subv): Support the other query units.
     ASSERT_MSG(regs.query.query_get.unit == Regs::QueryUnit::Crop,
                "Units other than CROP are unimplemented");
-    ASSERT_MSG(regs.query.query_get.short_query,
-               "Writing the entire query result structure is unimplemented");
 
     u32 value = Memory::Read32(*address);
-    u32 result = 0;
+    u64 result = 0;
 
     // TODO(Subv): Support the other query variables
     switch (regs.query.query_get.select) {
     case Regs::QuerySelect::Zero:
-        result = 0;
+        // This seems to actually write the query sequence to the query address.
+        result = regs.query.query_sequence;
         break;
     default:
         UNIMPLEMENTED_MSG("Unimplemented query select type {}",
@@ -174,15 +173,31 @@ void Maxwell3D::ProcessQueryGet() {
 
     // TODO(Subv): Research and implement how query sync conditions work.
 
+    struct LongQueryResult {
+        u64_le value;
+        u64_le timestamp;
+    };
+    static_assert(sizeof(LongQueryResult) == 16, "LongQueryResult has wrong size");
+
     switch (regs.query.query_get.mode) {
     case Regs::QueryMode::Write:
     case Regs::QueryMode::Write2: {
-        // Write the current query sequence to the sequence address.
         u32 sequence = regs.query.query_sequence;
-        Memory::Write32(*address, sequence);
-
-        // TODO(Subv): Write the proper query response structure to the address when not using short
-        // mode.
+        if (regs.query.query_get.short_query) {
+            // Write the current query sequence to the sequence address.
+            // TODO(Subv): Find out what happens if you use a long query type but mark it as a short
+            // query.
+            Memory::Write32(*address, sequence);
+        } else {
+            // Write the 128-bit result structure in long mode. Note: We emulate an infinitely fast
+            // GPU, this command may actually take a while to complete in real hardware due to GPU
+            // wait queues.
+            LongQueryResult query_result{};
+            query_result.value = result;
+            // TODO(Subv): Generate a real GPU timestamp and write it here instead of 0
+            query_result.timestamp = 0;
+            Memory::WriteBlock(*address, &query_result, sizeof(query_result));
+        }
         break;
     }
     default:
@@ -313,8 +328,9 @@ std::vector<Texture::FullTextureInfo> Maxwell3D::GetStageTextures(Regs::ShaderSt
 
         Texture::FullTextureInfo tex_info{};
         // TODO(Subv): Use the shader to determine which textures are actually accessed.
-        tex_info.index = (current_texture - tex_info_buffer.address - TextureInfoOffset) /
-                         sizeof(Texture::TextureHandle);
+        tex_info.index =
+            static_cast<u32>(current_texture - tex_info_buffer.address - TextureInfoOffset) /
+            sizeof(Texture::TextureHandle);
 
         // Load the TIC data.
         if (tex_handle.tic_id != 0) {
@@ -339,6 +355,40 @@ std::vector<Texture::FullTextureInfo> Maxwell3D::GetStageTextures(Regs::ShaderSt
     return textures;
 }
 
+Texture::FullTextureInfo Maxwell3D::GetStageTexture(Regs::ShaderStage stage, size_t offset) const {
+    auto& shader = state.shader_stages[static_cast<size_t>(stage)];
+    auto& tex_info_buffer = shader.const_buffers[regs.tex_cb_index];
+    ASSERT(tex_info_buffer.enabled && tex_info_buffer.address != 0);
+
+    GPUVAddr tex_info_address = tex_info_buffer.address + offset * sizeof(Texture::TextureHandle);
+
+    ASSERT(tex_info_address < tex_info_buffer.address + tex_info_buffer.size);
+
+    boost::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address);
+    Texture::TextureHandle tex_handle{Memory::Read32(*tex_address_cpu)};
+
+    Texture::FullTextureInfo tex_info{};
+    tex_info.index = static_cast<u32>(offset);
+
+    // Load the TIC data.
+    if (tex_handle.tic_id != 0) {
+        tex_info.enabled = true;
+
+        auto tic_entry = GetTICEntry(tex_handle.tic_id);
+        // TODO(Subv): Workaround for BitField's move constructor being deleted.
+        std::memcpy(&tex_info.tic, &tic_entry, sizeof(tic_entry));
+    }
+
+    // Load the TSC data
+    if (tex_handle.tsc_id != 0) {
+        auto tsc_entry = GetTSCEntry(tex_handle.tsc_id);
+        // TODO(Subv): Workaround for BitField's move constructor being deleted.
+        std::memcpy(&tex_info.tsc, &tsc_entry, sizeof(tsc_entry));
+    }
+
+    return tex_info;
+}
+
 u32 Maxwell3D::GetRegisterValue(u32 method) const {
     ASSERT_MSG(method < Regs::NUM_REGS, "Invalid Maxwell3D register");
     return regs.reg_array[method];

src/video_core/engines/maxwell_3d.h

@@ -318,6 +318,7 @@ public:
             Equation equation_a;
             Factor factor_source_a;
             Factor factor_dest_a;
+            INSERT_PADDING_WORDS(1);
         };
 
         union {
@@ -354,10 +355,35 @@ public:
                     f32 scale_x;
                     f32 scale_y;
                     f32 scale_z;
-                    u32 translate_x;
-                    u32 translate_y;
-                    u32 translate_z;
+                    f32 translate_x;
+                    f32 translate_y;
+                    f32 translate_z;
                     INSERT_PADDING_WORDS(2);
+
+                    MathUtil::Rectangle<s32> GetRect() const {
+                        return {
+                            GetX(),               // left
+                            GetY() + GetHeight(), // top
+                            GetX() + GetWidth(),  // right
+                            GetY()                // bottom
+                        };
+                    };
+
+                    s32 GetX() const {
+                        return static_cast<s32>(std::max(0.0f, translate_x - std::fabs(scale_x)));
+                    }
+
+                    s32 GetY() const {
+                        return static_cast<s32>(std::max(0.0f, translate_y - std::fabs(scale_y)));
+                    }
+
+                    s32 GetWidth() const {
+                        return static_cast<s32>(translate_x + std::fabs(scale_x)) - GetX();
+                    }
+
+                    s32 GetHeight() const {
+                        return static_cast<s32>(translate_y + std::fabs(scale_y)) - GetY();
+                    }
                 } viewport_transform[NumViewports];
 
                 struct {
@@ -371,15 +397,6 @@ public:
                     };
                     float depth_range_near;
                     float depth_range_far;
-
-                    MathUtil::Rectangle<s32> GetRect() const {
-                        return {
-                            static_cast<s32>(x),          // left
-                            static_cast<s32>(y + height), // top
-                            static_cast<s32>(x + width),  // right
-                            static_cast<s32>(y)           // bottom
-                        };
-                    };
                 } viewport[NumViewports];
 
                 INSERT_PADDING_WORDS(0x1D);
@@ -416,7 +433,27 @@ public:
                     };
                 } rt_control;
 
-                INSERT_PADDING_WORDS(0xCF);
+                INSERT_PADDING_WORDS(0x31);
+
+                u32 independent_blend_enable;
+
+                INSERT_PADDING_WORDS(0x15);
+
+                struct {
+                    u32 separate_alpha;
+                    Blend::Equation equation_rgb;
+                    Blend::Factor factor_source_rgb;
+                    Blend::Factor factor_dest_rgb;
+                    Blend::Equation equation_a;
+                    Blend::Factor factor_source_a;
+                    INSERT_PADDING_WORDS(1);
+                    Blend::Factor factor_dest_a;
+
+                    u32 enable_common;
+                    u32 enable[NumRenderTargets];
+                } blend;
+
+                INSERT_PADDING_WORDS(0x77);
 
                 struct {
                     u32 tsc_address_high;
@@ -541,9 +578,7 @@ public:
 
                 } vertex_array[NumVertexArrays];
 
-                Blend blend;
-
-                INSERT_PADDING_WORDS(0x39);
+                Blend independent_blend[NumRenderTargets];
 
                 struct {
                     u32 limit_high;
@@ -648,6 +683,9 @@ public:
     /// Returns a list of enabled textures for the specified shader stage.
     std::vector<Texture::FullTextureInfo> GetStageTextures(Regs::ShaderStage stage) const;
 
+    /// Returns the texture information for a specific texture in a specific shader stage.
+    Texture::FullTextureInfo GetStageTexture(Regs::ShaderStage stage, size_t offset) const;
+
     /// Returns whether the specified shader stage is enabled or not.
     bool IsShaderStageEnabled(Regs::ShaderStage stage) const;
 
@@ -703,6 +741,8 @@ ASSERT_REG_POSITION(vertex_buffer, 0x35D);
 ASSERT_REG_POSITION(zeta, 0x3F8);
 ASSERT_REG_POSITION(vertex_attrib_format[0], 0x458);
 ASSERT_REG_POSITION(rt_control, 0x487);
+ASSERT_REG_POSITION(independent_blend_enable, 0x4B9);
+ASSERT_REG_POSITION(blend, 0x4CF);
 ASSERT_REG_POSITION(tsc, 0x557);
 ASSERT_REG_POSITION(tic, 0x55D);
 ASSERT_REG_POSITION(code_address, 0x582);
@@ -710,7 +750,7 @@ ASSERT_REG_POSITION(draw, 0x585);
 ASSERT_REG_POSITION(index_array, 0x5F2);
 ASSERT_REG_POSITION(query, 0x6C0);
 ASSERT_REG_POSITION(vertex_array[0], 0x700);
-ASSERT_REG_POSITION(blend, 0x780);
+ASSERT_REG_POSITION(independent_blend, 0x780);
 ASSERT_REG_POSITION(vertex_array_limit[0], 0x7C0);
 ASSERT_REG_POSITION(shader_config[0], 0x800);
 ASSERT_REG_POSITION(const_buffer, 0x8E0);

src/video_core/engines/maxwell_dma.cpp

@@ -0,0 +1,69 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/memory.h"
+#include "video_core/engines/maxwell_dma.h"
+#include "video_core/textures/decoders.h"
+
+namespace Tegra {
+namespace Engines {
+
+MaxwellDMA::MaxwellDMA(MemoryManager& memory_manager) : memory_manager(memory_manager) {}
+
+void MaxwellDMA::WriteReg(u32 method, u32 value) {
+    ASSERT_MSG(method < Regs::NUM_REGS,
+               "Invalid MaxwellDMA register, increase the size of the Regs structure");
+
+    regs.reg_array[method] = value;
+
+#define MAXWELLDMA_REG_INDEX(field_name)                                                           \
+    (offsetof(Tegra::Engines::MaxwellDMA::Regs, field_name) / sizeof(u32))
+
+    switch (method) {
+    case MAXWELLDMA_REG_INDEX(exec): {
+        HandleCopy();
+        break;
+    }
+    }
+
+#undef MAXWELLDMA_REG_INDEX
+}
+
+void MaxwellDMA::HandleCopy() {
+    NGLOG_WARNING(HW_GPU, "Requested a DMA copy");
+
+    const GPUVAddr source = regs.src_address.Address();
+    const GPUVAddr dest = regs.dst_address.Address();
+
+    const VAddr source_cpu = *memory_manager.GpuToCpuAddress(source);
+    const VAddr dest_cpu = *memory_manager.GpuToCpuAddress(dest);
+
+    // TODO(Subv): Perform more research and implement all features of this engine.
+    ASSERT(regs.exec.enable_swizzle == 0);
+    ASSERT(regs.exec.enable_2d == 1);
+    ASSERT(regs.exec.query_mode == Regs::QueryMode::None);
+    ASSERT(regs.exec.query_intr == Regs::QueryIntr::None);
+    ASSERT(regs.exec.copy_mode == Regs::CopyMode::Unk2);
+    ASSERT(regs.src_params.pos_x == 0);
+    ASSERT(regs.src_params.pos_y == 0);
+    ASSERT(regs.dst_params.pos_x == 0);
+    ASSERT(regs.dst_params.pos_y == 0);
+    ASSERT(regs.exec.is_dst_linear != regs.exec.is_src_linear);
+
+    u8* src_buffer = Memory::GetPointer(source_cpu);
+    u8* dst_buffer = Memory::GetPointer(dest_cpu);
+
+    if (regs.exec.is_dst_linear && !regs.exec.is_src_linear) {
+        // If the input is tiled and the output is linear, deswizzle the input and copy it over.
+        Texture::CopySwizzledData(regs.src_params.size_x, regs.src_params.size_y, 1, 1, src_buffer,
+                                  dst_buffer, true, regs.src_params.BlockHeight());
+    } else {
+        // If the input is linear and the output is tiled, swizzle the input and copy it over.
+        Texture::CopySwizzledData(regs.dst_params.size_x, regs.dst_params.size_y, 1, 1, dst_buffer,
+                                  src_buffer, false, regs.dst_params.BlockHeight());
+    }
+}
+
+} // namespace Engines
+} // namespace Tegra

src/video_core/engines/maxwell_dma.h

@@ -0,0 +1,155 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include "common/assert.h"
+#include "common/bit_field.h"
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "video_core/gpu.h"
+#include "video_core/memory_manager.h"
+
+namespace Tegra {
+namespace Engines {
+
+class MaxwellDMA final {
+public:
+    explicit MaxwellDMA(MemoryManager& memory_manager);
+    ~MaxwellDMA() = default;
+
+    /// Write the value to the register identified by method.
+    void WriteReg(u32 method, u32 value);
+
+    struct Regs {
+        static constexpr size_t NUM_REGS = 0x1D6;
+
+        struct Parameters {
+            union {
+                BitField<0, 4, u32> block_depth;
+                BitField<4, 4, u32> block_height;
+                BitField<8, 4, u32> block_width;
+            };
+            u32 size_x;
+            u32 size_y;
+            u32 size_z;
+            u32 pos_z;
+            union {
+                BitField<0, 16, u32> pos_x;
+                BitField<16, 16, u32> pos_y;
+            };
+
+            u32 BlockHeight() const {
+                return 1 << block_height;
+            }
+        };
+
+        static_assert(sizeof(Parameters) == 24, "Parameters has wrong size");
+
+        enum class CopyMode : u32 {
+            None = 0,
+            Unk1 = 1,
+            Unk2 = 2,
+        };
+
+        enum class QueryMode : u32 {
+            None = 0,
+            Short = 1,
+            Long = 2,
+        };
+
+        enum class QueryIntr : u32 {
+            None = 0,
+            Block = 1,
+            NonBlock = 2,
+        };
+
+        union {
+            struct {
+                INSERT_PADDING_WORDS(0xC0);
+
+                struct {
+                    union {
+                        BitField<0, 2, CopyMode> copy_mode;
+                        BitField<2, 1, u32> flush;
+
+                        BitField<3, 2, QueryMode> query_mode;
+                        BitField<5, 2, QueryIntr> query_intr;
+
+                        BitField<7, 1, u32> is_src_linear;
+                        BitField<8, 1, u32> is_dst_linear;
+
+                        BitField<9, 1, u32> enable_2d;
+                        BitField<10, 1, u32> enable_swizzle;
+                    };
+                } exec;
+
+                INSERT_PADDING_WORDS(0x3F);
+
+                struct {
+                    u32 address_high;
+                    u32 address_low;
+
+                    GPUVAddr Address() const {
+                        return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) |
+                                                     address_low);
+                    }
+                } src_address;
+
+                struct {
+                    u32 address_high;
+                    u32 address_low;
+
+                    GPUVAddr Address() const {
+                        return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) |
+                                                     address_low);
+                    }
+                } dst_address;
+
+                u32 src_pitch;
+                u32 dst_pitch;
+                u32 x_count;
+                u32 y_count;
+
+                INSERT_PADDING_WORDS(0xBB);
+
+                Parameters dst_params;
+
+                INSERT_PADDING_WORDS(1);
+
+                Parameters src_params;
+
+                INSERT_PADDING_WORDS(0x13);
+            };
+            std::array<u32, NUM_REGS> reg_array;
+        };
+    } regs{};
+
+    MemoryManager& memory_manager;
+
+private:
+    /// Performs the copy from the source buffer to the destination buffer as configured in the
+    /// registers.
+    void HandleCopy();
+};
+
+#define ASSERT_REG_POSITION(field_name, position)                                                  \
+    static_assert(offsetof(MaxwellDMA::Regs, field_name) == position * 4,                          \
+                  "Field " #field_name " has invalid position")
+
+ASSERT_REG_POSITION(exec, 0xC0);
+ASSERT_REG_POSITION(src_address, 0x100);
+ASSERT_REG_POSITION(dst_address, 0x102);
+ASSERT_REG_POSITION(src_pitch, 0x104);
+ASSERT_REG_POSITION(dst_pitch, 0x105);
+ASSERT_REG_POSITION(x_count, 0x106);
+ASSERT_REG_POSITION(y_count, 0x107);
+ASSERT_REG_POSITION(dst_params, 0x1C3);
+ASSERT_REG_POSITION(src_params, 0x1CA);
+
+#undef ASSERT_REG_POSITION
+
+} // namespace Engines
+} // namespace Tegra

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 {
@@ -105,11 +109,6 @@ union Sampler {
     u64 value{};
 };
 
-union Uniform {
-    BitField<20, 14, u64> offset;
-    BitField<34, 5, u64> index;
-};
-
 } // namespace Shader
 } // namespace Tegra
 
@@ -152,6 +151,13 @@ enum class PredOperation : u64 {
     Xor = 2,
 };
 
+enum class LogicOperation : u64 {
+    And = 0,
+    Or = 1,
+    Xor = 2,
+    PassB = 3,
+};
+
 enum class SubOp : u64 {
     Cos = 0x0,
     Sin = 0x1,
@@ -162,6 +168,31 @@ enum class SubOp : u64 {
     Min = 0x8,
 };
 
+enum class F2iRoundingOp : u64 {
+    None = 0,
+    Floor = 1,
+    Ceil = 2,
+    Trunc = 3,
+};
+
+enum class F2fRoundingOp : u64 {
+    None = 0,
+    Pass = 3,
+    Round = 8,
+    Floor = 9,
+    Ceil = 10,
+    Trunc = 11,
+};
+
+enum class UniformType : u64 {
+    UnsignedByte = 0,
+    SignedByte = 1,
+    UnsignedShort = 2,
+    SignedShort = 3,
+    Single = 4,
+    Double = 5,
+};
+
 union Instruction {
     Instruction& operator=(const Instruction& instr) {
         value = instr.value;
@@ -185,14 +216,33 @@ union Instruction {
 
     union {
         BitField<20, 19, u64> imm20_19;
-        BitField<20, 32, u64> imm20_32;
+        BitField<20, 32, s64> imm20_32;
         BitField<45, 1, u64> negate_b;
         BitField<46, 1, u64> abs_a;
         BitField<48, 1, u64> negate_a;
         BitField<49, 1, u64> abs_b;
-        BitField<50, 1, u64> abs_d;
+        BitField<50, 1, u64> saturate_d;
         BitField<56, 1, u64> negate_imm;
 
+        union {
+            BitField<39, 3, u64> pred;
+            BitField<42, 1, u64> negate_pred;
+        } fmnmx;
+
+        union {
+            BitField<39, 1, u64> invert_a;
+            BitField<40, 1, u64> invert_b;
+            BitField<41, 2, LogicOperation> operation;
+            BitField<44, 2, u64> unk44;
+            BitField<48, 3, Pred> pred48;
+        } lop;
+
+        union {
+            BitField<53, 2, LogicOperation> operation;
+            BitField<55, 1, u64> invert_a;
+            BitField<56, 1, u64> invert_b;
+        } lop32i;
+
         float GetImm20_19() const {
             float result{};
             u32 imm{static_cast<u32>(imm20_19)};
@@ -204,17 +254,55 @@ union Instruction {
 
         float GetImm20_32() const {
             float result{};
-            u32 imm{static_cast<u32>(imm20_32)};
+            s32 imm{static_cast<s32>(imm20_32)};
             std::memcpy(&result, &imm, sizeof(imm));
             return result;
         }
+
+        s32 GetSignedImm20_20() const {
+            u32 immediate = static_cast<u32>(imm20_19 | (negate_imm << 19));
+            // Sign extend the 20-bit value.
+            u32 mask = 1U << (20 - 1);
+            return static_cast<s32>((immediate ^ mask) - mask);
+        }
     } alu;
 
+    union {
+        BitField<48, 1, u64> is_signed;
+    } shift;
+
+    union {
+        BitField<39, 5, u64> shift_amount;
+        BitField<48, 1, u64> negate_b;
+        BitField<49, 1, u64> negate_a;
+    } alu_integer;
+
+    union {
+        BitField<54, 1, u64> saturate;
+        BitField<56, 1, u64> negate_a;
+    } iadd32i;
+
+    union {
+        BitField<20, 8, u64> shift_position;
+        BitField<28, 8, u64> shift_length;
+        BitField<48, 1, u64> negate_b;
+        BitField<49, 1, u64> negate_a;
+
+        u64 GetLeftShiftValue() const {
+            return 32 - (shift_position + shift_length);
+        }
+    } bfe;
+
     union {
         BitField<48, 1, u64> negate_b;
         BitField<49, 1, u64> negate_c;
     } ffma;
 
+    union {
+        BitField<48, 3, UniformType> type;
+        BitField<44, 2, u64> unknown;
+    } ld_c;
+
     union {
         BitField<0, 3, u64> pred0;
         BitField<3, 3, u64> pred3;
@@ -229,6 +317,16 @@ union Instruction {
         BitField<56, 1, u64> neg_b;
     } fsetp;
 
+    union {
+        BitField<0, 3, u64> pred0;
+        BitField<3, 3, u64> pred3;
+        BitField<39, 3, u64> pred39;
+        BitField<42, 1, u64> neg_pred;
+        BitField<45, 2, PredOperation> op;
+        BitField<48, 1, u64> is_signed;
+        BitField<49, 3, PredCondition> cond;
+    } isetp;
+
     union {
         BitField<39, 3, u64> pred39;
         BitField<42, 1, u64> neg_pred;
@@ -236,28 +334,96 @@ union Instruction {
         BitField<44, 1, u64> abs_b;
         BitField<45, 2, PredOperation> op;
         BitField<48, 4, PredCondition> cond;
+        BitField<52, 1, u64> bf;
         BitField<53, 1, u64> neg_b;
         BitField<54, 1, u64> abs_a;
-        BitField<52, 1, u64> bf;
         BitField<55, 1, u64> ftz;
         BitField<56, 1, u64> neg_imm;
     } fset;
 
     union {
-        BitField<10, 2, Register::Size> size;
-        BitField<13, 1, u64> is_signed;
+        BitField<39, 3, u64> pred39;
+        BitField<42, 1, u64> neg_pred;
+        BitField<44, 1, u64> bf;
+        BitField<45, 2, PredOperation> op;
+        BitField<48, 1, u64> is_signed;
+        BitField<49, 3, PredCondition> cond;
+    } iset;
+
+    union {
+        BitField<8, 2, Register::Size> dest_size;
+        BitField<10, 2, Register::Size> src_size;
+        BitField<12, 1, u64> is_output_signed;
+        BitField<13, 1, u64> is_input_signed;
         BitField<41, 2, u64> selector;
         BitField<45, 1, u64> negate_a;
         BitField<49, 1, u64> abs_a;
-        BitField<50, 1, u64> saturate_a;
+
+        union {
+            BitField<39, 2, F2iRoundingOp> rounding;
+        } f2i;
+
+        union {
+            BitField<39, 4, F2fRoundingOp> rounding;
+        } f2f;
     } conversion;
 
+    union {
+        BitField<31, 4, u64> component_mask;
+
+        bool IsComponentEnabled(size_t component) const {
+            return ((1ull << component) & component_mask) != 0;
+        }
+    } tex;
+
+    union {
+        BitField<50, 3, u64> component_mask_selector;
+        BitField<28, 8, Register> gpr28;
+
+        bool HasTwoDestinations() const {
+            return gpr28.Value() != Register::ZeroIndex;
+        }
+
+        bool IsComponentEnabled(size_t component) const {
+            static constexpr std::array<size_t, 5> one_dest_mask{0x1, 0x2, 0x4, 0x8, 0x3};
+            static constexpr std::array<size_t, 5> two_dest_mask{0x7, 0xb, 0xd, 0xe, 0xf};
+            const auto& mask{HasTwoDestinations() ? two_dest_mask : one_dest_mask};
+
+            ASSERT(component_mask_selector < mask.size());
+
+            return ((1ull << component) & mask[component_mask_selector]) != 0;
+        }
+    } texs;
+
+    union {
+        BitField<20, 24, u64> target;
+        BitField<5, 1, u64> constant_buffer;
+
+        s32 GetBranchTarget() const {
+            // Sign extend the branch target offset
+            u32 mask = 1U << (24 - 1);
+            u32 value = static_cast<u32>(target);
+            // The branch offset is relative to the next instruction and is stored in bytes, so
+            // divide it by the size of an instruction and add 1 to it.
+            return static_cast<s32>((value ^ mask) - mask) / sizeof(Instruction) + 1;
+        }
+    } bra;
+
+    union {
+        BitField<20, 14, u64> offset;
+        BitField<34, 5, u64> index;
+    } cbuf34;
+
+    union {
+        BitField<20, 16, s64> offset;
+        BitField<36, 5, u64> index;
+    } cbuf36;
+
     BitField<61, 1, u64> is_b_imm;
     BitField<60, 1, u64> is_b_gpr;
     BitField<59, 1, u64> is_c_gpr;
 
     Attribute attribute;
-    Uniform uniform;
     Sampler sampler;
 
     u64 value;
@@ -270,8 +436,15 @@ class OpCode {
 public:
     enum class Id {
         KIL,
+        SSY,
+        BFE_C,
+        BFE_R,
+        BFE_IMM,
+        BRA,
         LD_A,
+        LD_C,
         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 +461,17 @@ public:
         FMUL_R,
         FMUL_IMM,
         FMUL32_IMM,
-        MUFU, // Multi-Function Operator
-        RRO,  // Range Reduction Operator
+        IADD_C,
+        IADD_R,
+        IADD_IMM,
+        IADD32I,
+        ISCADD_C, // Scale and Add
+        ISCADD_R,
+        ISCADD_IMM,
+        MUFU,  // Multi-Function Operator
+        RRO_C, // Range Reduction Operator
+        RRO_R,
+        RRO_IMM,
         F2F_C,
         F2F_R,
         F2F_IMM,
@@ -302,15 +484,26 @@ public:
         I2I_C,
         I2I_R,
         I2I_IMM,
+        LOP_C,
+        LOP_R,
+        LOP_IMM,
         LOP32I,
         MOV_C,
         MOV_R,
         MOV_IMM,
         MOV32_IMM,
+        SHL_C,
+        SHL_R,
+        SHL_IMM,
         SHR_C,
         SHR_R,
         SHR_IMM,
-        FMNMX,
+        FMNMX_C,
+        FMNMX_R,
+        FMNMX_IMM,
+        IMNMX_C,
+        IMNMX_R,
+        IMNMX_IMM,
         FSETP_C, // Set Predicate
         FSETP_R,
         FSETP_IMM,
@@ -320,17 +513,32 @@ public:
         ISETP_C,
         ISETP_IMM,
         ISETP_R,
+        ISET_R,
+        ISET_C,
+        ISET_IMM,
+        PSETP,
+        XMAD_IMM,
+        XMAD_CR,
+        XMAD_RC,
+        XMAD_RR,
     };
 
     enum class Type {
         Trivial,
         Arithmetic,
+        ArithmeticImmediate,
+        ArithmeticInteger,
+        ArithmeticIntegerImmediate,
+        Bfe,
+        Shift,
         Ffma,
         Flow,
         Memory,
         FloatSet,
         FloatSetPredicate,
+        IntegerSet,
         IntegerSetPredicate,
+        PredicateSetPredicate,
         Conversion,
         Unknown,
     };
@@ -427,8 +635,12 @@ private:
         std::vector<Matcher> table = {
 #define INST(bitstring, op, type, name) Detail::GetMatcher(bitstring, op, type, name)
             INST("111000110011----", Id::KIL, Type::Flow, "KIL"),
+            INST("111000101001----", Id::SSY, Type::Flow, "SSY"),
+            INST("111000100100----", Id::BRA, Type::Flow, "BRA"),
             INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
+            INST("1110111110010---", Id::LD_C, Type::Memory, "LD_C"),
             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"),
@@ -444,24 +656,47 @@ private:
             INST("0100110001101---", Id::FMUL_C, Type::Arithmetic, "FMUL_C"),
             INST("0101110001101---", Id::FMUL_R, Type::Arithmetic, "FMUL_R"),
             INST("0011100-01101---", Id::FMUL_IMM, Type::Arithmetic, "FMUL_IMM"),
-            INST("00011110--------", Id::FMUL32_IMM, Type::Arithmetic, "FMUL32_IMM"),
+            INST("00011110--------", Id::FMUL32_IMM, Type::ArithmeticImmediate, "FMUL32_IMM"),
+            INST("0100110000010---", Id::IADD_C, Type::ArithmeticInteger, "IADD_C"),
+            INST("0101110000010---", Id::IADD_R, Type::ArithmeticInteger, "IADD_R"),
+            INST("0011100-00010---", Id::IADD_IMM, Type::ArithmeticInteger, "IADD_IMM"),
+            INST("0001110---------", Id::IADD32I, Type::ArithmeticIntegerImmediate, "IADD32I"),
+            INST("0100110000011---", Id::ISCADD_C, Type::ArithmeticInteger, "ISCADD_C"),
+            INST("0101110000011---", Id::ISCADD_R, Type::ArithmeticInteger, "ISCADD_R"),
+            INST("0011100-00011---", Id::ISCADD_IMM, Type::ArithmeticInteger, "ISCADD_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("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"),
-            INST("000001----------", Id::LOP32I, Type::Arithmetic, "LOP32I"),
+            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::Conversion, "F2I_C"),
+            INST("0101110010110---", Id::F2I_R, Type::Conversion, "F2I_R"),
+            INST("0011100-10110---", Id::F2I_IMM, Type::Conversion, "F2I_IMM"),
             INST("0100110010011---", Id::MOV_C, Type::Arithmetic, "MOV_C"),
             INST("0101110010011---", Id::MOV_R, Type::Arithmetic, "MOV_R"),
             INST("0011100-10011---", Id::MOV_IMM, Type::Arithmetic, "MOV_IMM"),
-            INST("000000010000----", Id::MOV32_IMM, Type::Arithmetic, "MOV32_IMM"),
-            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("000000010000----", Id::MOV32_IMM, Type::ArithmeticImmediate, "MOV32_IMM"),
+            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("0100110000100---", Id::IMNMX_C, Type::Arithmetic, "FMNMX_IMM"),
+            INST("0101110000100---", Id::IMNMX_R, Type::Arithmetic, "FMNMX_IMM"),
+            INST("0011100-00100---", Id::IMNMX_IMM, Type::Arithmetic, "FMNMX_IMM"),
+            INST("0100110000000---", Id::BFE_C, Type::Bfe, "BFE_C"),
+            INST("0101110000000---", Id::BFE_R, Type::Bfe, "BFE_R"),
+            INST("0011100-00000---", Id::BFE_IMM, Type::Bfe, "BFE_IMM"),
+            INST("0100110001000---", Id::LOP_C, Type::ArithmeticInteger, "LOP_C"),
+            INST("0101110001000---", Id::LOP_R, Type::ArithmeticInteger, "LOP_R"),
+            INST("0011100001000---", Id::LOP_IMM, Type::ArithmeticInteger, "LOP_IMM"),
+            INST("000001----------", Id::LOP32I, Type::ArithmeticIntegerImmediate, "LOP32I"),
+            INST("0100110001001---", Id::SHL_C, Type::Shift, "SHL_C"),
+            INST("0101110001001---", Id::SHL_R, Type::Shift, "SHL_R"),
+            INST("0011100-01001---", Id::SHL_IMM, Type::Shift, "SHL_IMM"),
+            INST("0100110000101---", Id::SHR_C, Type::Shift, "SHR_C"),
+            INST("0101110000101---", Id::SHR_R, Type::Shift, "SHR_R"),
+            INST("0011100-00101---", Id::SHR_IMM, Type::Shift, "SHR_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 +712,14 @@ 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("010110110101----", Id::ISET_R, Type::IntegerSet, "ISET_R"),
+            INST("010010110101----", Id::ISET_C, Type::IntegerSet, "ISET_C"),
+            INST("0011011-0101----", Id::ISET_IMM, Type::IntegerSet, "ISET_IMM"),
+            INST("0101000010010---", Id::PSETP, Type::PredicateSetPredicate, "PSETP"),
+            INST("0011011-00------", Id::XMAD_IMM, Type::Arithmetic, "XMAD_IMM"),
+            INST("0100111---------", Id::XMAD_CR, Type::Arithmetic, "XMAD_CR"),
+            INST("010100010-------", Id::XMAD_RC, Type::Arithmetic, "XMAD_RC"),
+            INST("0101101100------", Id::XMAD_RR, Type::Arithmetic, "XMAD_RR"),
         };
 #undef INST
         std::stable_sort(table.begin(), table.end(), [](const auto& a, const auto& b) {

src/video_core/gpu.cpp

@@ -5,6 +5,7 @@
 #include "video_core/engines/fermi_2d.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/engines/maxwell_compute.h"
+#include "video_core/engines/maxwell_dma.h"
 #include "video_core/gpu.h"
 
 namespace Tegra {
@@ -14,6 +15,7 @@ GPU::GPU() {
     maxwell_3d = std::make_unique<Engines::Maxwell3D>(*memory_manager);
     fermi_2d = std::make_unique<Engines::Fermi2D>(*memory_manager);
     maxwell_compute = std::make_unique<Engines::MaxwellCompute>();
+    maxwell_dma = std::make_unique<Engines::MaxwellDMA>(*memory_manager);
 }
 
 GPU::~GPU() = default;
@@ -26,6 +28,10 @@ u32 RenderTargetBytesPerPixel(RenderTargetFormat format) {
     ASSERT(format != RenderTargetFormat::NONE);
 
     switch (format) {
+    case RenderTargetFormat::RGBA32_FLOAT:
+        return 16;
+    case RenderTargetFormat::RGBA16_FLOAT:
+        return 8;
     case RenderTargetFormat::RGBA8_UNORM:
     case RenderTargetFormat::RGB10_A2_UNORM:
         return 4;

src/video_core/gpu.h

@@ -15,10 +15,12 @@ namespace Tegra {
 
 enum class RenderTargetFormat : u32 {
     NONE = 0x0,
+    RGBA32_FLOAT = 0xC0,
     RGBA16_FLOAT = 0xCA,
     RGB10_A2_UNORM = 0xD1,
     RGBA8_UNORM = 0xD5,
     RGBA8_SRGB = 0xD6,
+    R11G11B10_FLOAT = 0xE0,
 };
 
 /// Returns the number of bytes per pixel of each rendertarget format.
@@ -61,6 +63,7 @@ namespace Engines {
 class Fermi2D;
 class Maxwell3D;
 class MaxwellCompute;
+class MaxwellDMA;
 } // namespace Engines
 
 enum class EngineID {
@@ -101,6 +104,8 @@ private:
     std::unique_ptr<Engines::Fermi2D> fermi_2d;
     /// Compute engine
     std::unique_ptr<Engines::MaxwellCompute> maxwell_compute;
+    /// DMA engine
+    std::unique_ptr<Engines::MaxwellDMA> maxwell_dma;
 };
 
 } // namespace Tegra

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};
@@ -202,8 +196,10 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) {
     auto& gpu = Core::System().GetInstance().GPU().Maxwell3D();
     ASSERT_MSG(!gpu.regs.shader_config[0].enable, "VertexA is unsupported!");
 
-    // Next available bindpoint to use when uploading the const buffers to the GLSL shaders.
+    // Next available bindpoints to use when uploading the const buffers and textures to the GLSL
+    // shaders.
     u32 current_constbuffer_bindpoint = 0;
+    u32 current_texture_bindpoint = 0;
 
     for (unsigned index = 1; index < Maxwell::MaxShaderProgram; ++index) {
         auto& shader_config = gpu.regs.shader_config[index];
@@ -218,13 +214,17 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) {
             continue;
         }
 
+        GLShader::MaxwellUniformData ubo{};
+        ubo.SetFromRegs(gpu.state.shader_stages[stage]);
+        std::memcpy(buffer_ptr, &ubo, sizeof(ubo));
+
+        // Flush the buffer so that the GPU can see the data we just wrote.
+        glFlushMappedBufferRange(GL_ARRAY_BUFFER, buffer_offset, sizeof(ubo));
+
         // Upload uniform data as one UBO per stage
         const GLintptr ubo_offset = buffer_offset;
         copy_buffer(uniform_buffers[stage].handle, ubo_offset,
                     sizeof(GLShader::MaxwellUniformData));
-        GLShader::MaxwellUniformData* ub_ptr =
-            reinterpret_cast<GLShader::MaxwellUniformData*>(buffer_ptr);
-        ub_ptr->SetFromRegs(gpu.state.shader_stages[stage]);
 
         buffer_ptr += sizeof(GLShader::MaxwellUniformData);
         buffer_offset += sizeof(GLShader::MaxwellUniformData);
@@ -264,6 +264,11 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) {
         current_constbuffer_bindpoint =
             SetupConstBuffers(static_cast<Maxwell::ShaderStage>(stage), gl_stage_program,
                               current_constbuffer_bindpoint, shader_resources.const_buffer_entries);
+
+        // Configure the textures for this shader stage.
+        current_texture_bindpoint =
+            SetupTextures(static_cast<Maxwell::ShaderStage>(stage), gl_stage_program,
+                          current_texture_bindpoint, shader_resources.texture_samplers);
     }
 
     shader_program_manager->UseTrivialGeometryShader();
@@ -304,7 +309,7 @@ void RasterizerOpenGL::DrawArrays() {
     const bool has_stencil = false;
     const bool using_color_fb = true;
     const bool using_depth_fb = false;
-    const MathUtil::Rectangle<s32> viewport_rect{regs.viewport[0].GetRect()};
+    const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()};
 
     const bool write_color_fb =
         state.color_mask.red_enabled == GL_TRUE || state.color_mask.green_enabled == GL_TRUE ||
@@ -344,12 +349,12 @@ void RasterizerOpenGL::DrawArrays() {
     // Sync the viewport
     SyncViewport(surfaces_rect, res_scale);
 
+    // Sync the blend state registers
+    SyncBlendState();
+
     // TODO(bunnei): Sync framebuffer_scale uniform here
     // TODO(bunnei): Sync scissorbox uniform(s) here
 
-    // Sync and bind the texture surfaces
-    BindTextures();
-
     // Viewport can have negative offsets or larger dimensions than our framebuffer sub-rect. Enable
     // scissor test to prevent drawing outside of the framebuffer region
     state.scissor.enabled = true;
@@ -453,65 +458,7 @@ void RasterizerOpenGL::DrawArrays() {
     }
 }
 
-void RasterizerOpenGL::BindTextures() {
-    using Regs = Tegra::Engines::Maxwell3D::Regs;
-    auto& maxwell3d = Core::System::GetInstance().GPU().Get3DEngine();
-
-    // Each Maxwell shader stage can have an arbitrary number of textures, but we're limited to a
-    // certain number in OpenGL. We try to only use the minimum amount of host textures by not
-    // keeping a 1:1 relation between guest texture ids and host texture ids, ie, guest texture id 8
-    // can be host texture id 0 if it's the only texture used in the guest shader program.
-    u32 host_texture_index = 0;
-    for (u32 stage = 0; stage < Regs::MaxShaderStage; ++stage) {
-        ASSERT(host_texture_index < texture_samplers.size());
-        const auto textures = maxwell3d.GetStageTextures(static_cast<Regs::ShaderStage>(stage));
-        for (unsigned texture_index = 0; texture_index < textures.size(); ++texture_index) {
-            const auto& texture = textures[texture_index];
-
-            if (texture.enabled) {
-                texture_samplers[host_texture_index].SyncWithConfig(texture.tsc);
-                Surface surface = res_cache.GetTextureSurface(texture);
-                if (surface != nullptr) {
-                    state.texture_units[host_texture_index].texture_2d = surface->texture.handle;
-                } else {
-                    // Can occur when texture addr is null or its memory is unmapped/invalid
-                    state.texture_units[texture_index].texture_2d = 0;
-                }
-
-                ++host_texture_index;
-            } else {
-                state.texture_units[texture_index].texture_2d = 0;
-            }
-        }
-    }
-}
-
-void RasterizerOpenGL::NotifyMaxwellRegisterChanged(u32 method) {
-    const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
-    switch (method) {
-    case MAXWELL3D_REG_INDEX(blend.separate_alpha):
-        ASSERT_MSG(false, "unimplemented");
-        break;
-    case MAXWELL3D_REG_INDEX(blend.equation_rgb):
-        state.blend.rgb_equation = MaxwellToGL::BlendEquation(regs.blend.equation_rgb);
-        break;
-    case MAXWELL3D_REG_INDEX(blend.factor_source_rgb):
-        state.blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_rgb);
-        break;
-    case MAXWELL3D_REG_INDEX(blend.factor_dest_rgb):
-        state.blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_rgb);
-        break;
-    case MAXWELL3D_REG_INDEX(blend.equation_a):
-        state.blend.a_equation = MaxwellToGL::BlendEquation(regs.blend.equation_a);
-        break;
-    case MAXWELL3D_REG_INDEX(blend.factor_source_a):
-        state.blend.src_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_a);
-        break;
-    case MAXWELL3D_REG_INDEX(blend.factor_dest_a):
-        state.blend.dst_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_a);
-        break;
-    }
-}
+void RasterizerOpenGL::NotifyMaxwellRegisterChanged(u32 method) {}
 
 void RasterizerOpenGL::FlushAll() {
     MICROPROFILE_SCOPE(OpenGL_CacheManagement);
@@ -660,7 +607,16 @@ u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, GLuint progr
         buffer_draw_state.bindpoint = current_bindpoint + bindpoint;
 
         boost::optional<VAddr> addr = gpu.memory_manager->GpuToCpuAddress(buffer.address);
-        std::vector<u8> data(used_buffer.GetSize() * sizeof(float));
+
+        std::vector<u8> data;
+        if (used_buffer.IsIndirect()) {
+            // Buffer is accessed indirectly, so upload the entire thing
+            data.resize(buffer.size * sizeof(float));
+        } else {
+            // Buffer is accessed directly, upload just what we use
+            data.resize(used_buffer.GetSize() * sizeof(float));
+        }
+
         Memory::ReadBlock(*addr, data.data(), data.size());
 
         glBindBuffer(GL_SHADER_STORAGE_BUFFER, buffer_draw_state.ssbo);
@@ -677,7 +633,53 @@ u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, GLuint progr
 
     state.Apply();
 
-    return current_bindpoint + entries.size();
+    return current_bindpoint + static_cast<u32>(entries.size());
+}
+
+u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, GLuint program, u32 current_unit,
+                                    const std::vector<GLShader::SamplerEntry>& entries) {
+    auto& gpu = Core::System::GetInstance().GPU();
+    auto& maxwell3d = gpu.Get3DEngine();
+
+    ASSERT_MSG(maxwell3d.IsShaderStageEnabled(stage),
+               "Attempted to upload textures of disabled shader stage");
+
+    ASSERT_MSG(current_unit + entries.size() <= std::size(state.texture_units),
+               "Exceeded the number of active textures.");
+
+    for (u32 bindpoint = 0; bindpoint < entries.size(); ++bindpoint) {
+        const auto& entry = entries[bindpoint];
+        u32 current_bindpoint = current_unit + bindpoint;
+
+        // Bind the uniform to the sampler.
+        GLint uniform = glGetUniformLocation(program, entry.GetName().c_str());
+        ASSERT(uniform != -1);
+        glProgramUniform1i(program, uniform, current_bindpoint);
+
+        const auto texture = maxwell3d.GetStageTexture(entry.GetStage(), entry.GetOffset());
+        ASSERT(texture.enabled);
+
+        texture_samplers[current_bindpoint].SyncWithConfig(texture.tsc);
+        Surface surface = res_cache.GetTextureSurface(texture);
+        if (surface != nullptr) {
+            state.texture_units[current_bindpoint].texture_2d = surface->texture.handle;
+            state.texture_units[current_bindpoint].swizzle.r =
+                MaxwellToGL::SwizzleSource(texture.tic.x_source);
+            state.texture_units[current_bindpoint].swizzle.g =
+                MaxwellToGL::SwizzleSource(texture.tic.y_source);
+            state.texture_units[current_bindpoint].swizzle.b =
+                MaxwellToGL::SwizzleSource(texture.tic.z_source);
+            state.texture_units[current_bindpoint].swizzle.a =
+                MaxwellToGL::SwizzleSource(texture.tic.w_source);
+        } else {
+            // Can occur when texture addr is null or its memory is unmapped/invalid
+            state.texture_units[current_bindpoint].texture_2d = 0;
+        }
+    }
+
+    state.Apply();
+
+    return current_unit + static_cast<u32>(entries.size());
 }
 
 void RasterizerOpenGL::BindFramebufferSurfaces(const Surface& color_surface,
@@ -708,7 +710,7 @@ void RasterizerOpenGL::BindFramebufferSurfaces(const Surface& color_surface,
 
 void RasterizerOpenGL::SyncViewport(const MathUtil::Rectangle<u32>& surfaces_rect, u16 res_scale) {
     const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
-    const MathUtil::Rectangle<s32> viewport_rect{regs.viewport[0].GetRect()};
+    const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()};
 
     state.viewport.x = static_cast<GLint>(surfaces_rect.left) + viewport_rect.left * res_scale;
     state.viewport.y = static_cast<GLint>(surfaces_rect.bottom) + viewport_rect.bottom * res_scale;
@@ -736,14 +738,21 @@ void RasterizerOpenGL::SyncDepthOffset() {
     UNREACHABLE();
 }
 
-void RasterizerOpenGL::SyncBlendEnabled() {
-    UNREACHABLE();
-}
+void RasterizerOpenGL::SyncBlendState() {
+    const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
 
-void RasterizerOpenGL::SyncBlendFuncs() {
-    UNREACHABLE();
-}
+    // TODO(Subv): Support more than just render target 0.
+    state.blend.enabled = regs.blend.enable[0] != 0;
 
-void RasterizerOpenGL::SyncBlendColor() {
-    UNREACHABLE();
+    if (!state.blend.enabled)
+        return;
+
+    ASSERT_MSG(regs.independent_blend_enable == 1, "Only independent blending is implemented");
+    ASSERT_MSG(!regs.independent_blend[0].separate_alpha, "Unimplemented");
+    state.blend.rgb_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_rgb);
+    state.blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_rgb);
+    state.blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_rgb);
+    state.blend.a_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_a);
+    state.blend.src_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_a);
+    state.blend.dst_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_a);
 }

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -80,9 +80,6 @@ private:
     void BindFramebufferSurfaces(const Surface& color_surface, const Surface& depth_surface,
                                  bool has_stencil);
 
-    /// Binds the required textures to OpenGL before drawing a batch.
-    void BindTextures();
-
     /*
      * Configures the current constbuffers to use for the draw command.
      * @param stage The shader stage to configure buffers for.
@@ -95,6 +92,17 @@ private:
                           u32 current_bindpoint,
                           const std::vector<GLShader::ConstBufferEntry>& entries);
 
+    /*
+     * Configures the current textures to use for the draw command.
+     * @param stage The shader stage to configure textures for.
+     * @param program The OpenGL program object that contains the specified stage.
+     * @param current_unit The offset at which to start counting unused texture units.
+     * @param entries Vector describing the textures that are actually used in the guest shader.
+     * @returns The next available bindpoint for use in the next shader stage.
+     */
+    u32 SetupTextures(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, GLuint program,
+                      u32 current_unit, const std::vector<GLShader::SamplerEntry>& entries);
+
     /// Syncs the viewport to match the guest state
     void SyncViewport(const MathUtil::Rectangle<u32>& surfaces_rect, u16 res_scale);
 
@@ -113,14 +121,8 @@ private:
     /// Syncs the depth offset to match the guest state
     void SyncDepthOffset();
 
-    /// Syncs the blend enabled status to match the guest state
-    void SyncBlendEnabled();
-
-    /// Syncs the blend functions to match the guest state
-    void SyncBlendFuncs();
-
-    /// Syncs the blend color to match the guest state
-    void SyncBlendColor();
+    /// Syncs the blend state to match the guest state
+    void SyncBlendState();
 
     bool has_ARB_buffer_storage;
     bool has_ARB_direct_state_access;
@@ -134,21 +136,17 @@ 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

@@ -28,6 +28,7 @@
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_opengl/gl_rasterizer_cache.h"
 #include "video_core/renderer_opengl/gl_state.h"
+#include "video_core/textures/astc.h"
 #include "video_core/textures/decoders.h"
 #include "video_core/utils.h"
 #include "video_core/video_core.h"
@@ -45,19 +46,27 @@ 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_R11F_G11F_B10F, GL_RGB, GL_UNSIGNED_INT_10F_11F_11F_REV, false},        // R11FG11FB10F
     {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
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, false},                               // ASTC_2D_4X4
 }};
 
 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 either R11FG11FB10F or RGBA16F which are
+        // type FLOAT
+        ASSERT(component_type == ComponentType::UNorm || pixel_format == PixelFormat::RGBA16F ||
+               pixel_format == PixelFormat::R11FG11FB10F);
         return tex_format_tuples[static_cast<unsigned int>(pixel_format)];
     } else if (type == SurfaceType::Depth || type == SurfaceType::DepthStencil) {
         // TODO(Subv): Implement depth formats
@@ -79,6 +88,23 @@ static u16 GetResolutionScaleFactor() {
                                 : Settings::values.resolution_factor);
 }
 
+static void ConvertASTCToRGBA8(std::vector<u8>& data, PixelFormat format, u32 width, u32 height) {
+    u32 block_width{};
+    u32 block_height{};
+
+    switch (format) {
+    case PixelFormat::ASTC_2D_4X4:
+        block_width = 4;
+        block_height = 4;
+        break;
+    default:
+        NGLOG_CRITICAL(HW_GPU, "Unhandled format: {}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+
+    data = Tegra::Texture::ASTC::Decompress(data, width, height, block_width, block_height);
+}
+
 template <bool morton_to_gl, PixelFormat format>
 void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, Tegra::GPUVAddr base,
                 Tegra::GPUVAddr start, Tegra::GPUVAddr end) {
@@ -90,6 +116,12 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, Tegra::
         auto data = Tegra::Texture::UnswizzleTexture(
             *gpu.memory_manager->GpuToCpuAddress(base),
             SurfaceParams::TextureFormatFromPixelFormat(format), stride, height, block_height);
+
+        if (SurfaceParams::IsFormatASTC(format)) {
+            // ASTC formats are converted to RGBA8 in software, as most PC GPUs do not support this
+            ConvertASTCToRGBA8(data, format, stride, height);
+        }
+
         std::memcpy(gl_buffer, data.data(), data.size());
     } else {
         // TODO(bunnei): Assumes the default rendering GOB size of 16 (128 lines). We should check
@@ -106,9 +138,12 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, Tegra::GPUVAddr, Tegra:
                                      Tegra::GPUVAddr),
                             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::ABGR8>,        MortonCopy<true, PixelFormat::B5G6R5>,
+        MortonCopy<true, PixelFormat::A2B10G10R10>,  MortonCopy<true, PixelFormat::A1B5G5R5>,
+        MortonCopy<true, PixelFormat::R8>,           MortonCopy<true, PixelFormat::RGBA16F>,
+        MortonCopy<true, PixelFormat::R11FG11FB10F>, MortonCopy<true, PixelFormat::DXT1>,
+        MortonCopy<true, PixelFormat::DXT23>,        MortonCopy<true, PixelFormat::DXT45>,
+        MortonCopy<true, PixelFormat::DXN1>,         MortonCopy<true, PixelFormat::ASTC_2D_4X4>,
 };
 
 static constexpr std::array<void (*)(u32, u32, u32, u8*, Tegra::GPUVAddr, Tegra::GPUVAddr,
@@ -118,10 +153,16 @@ 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>,
+        MortonCopy<false, PixelFormat::R11FG11FB10F>,
+        // TODO(Subv): Swizzling the DXT1/DXT23/DXT45/DXN1 formats is not yet supported
         nullptr,
         nullptr,
         nullptr,
+        nullptr,
+        MortonCopy<false, PixelFormat::ABGR8>,
 };
 
 // Allocate an uninitialized texture of appropriate size and format for the surface
@@ -154,60 +195,10 @@ static void AllocateSurfaceTexture(GLuint texture, const FormatTuple& format_tup
 static bool BlitTextures(GLuint src_tex, const MathUtil::Rectangle<u32>& src_rect, GLuint dst_tex,
                          const MathUtil::Rectangle<u32>& dst_rect, SurfaceType type,
                          GLuint read_fb_handle, GLuint draw_fb_handle) {
-    OpenGLState state = OpenGLState::GetCurState();
-
-    OpenGLState prev_state = state;
-    SCOPE_EXIT({ prev_state.Apply(); });
-
-    // Make sure textures aren't bound to texture units, since going to bind them to framebuffer
-    // components
-    state.ResetTexture(src_tex);
-    state.ResetTexture(dst_tex);
-
-    state.draw.read_framebuffer = read_fb_handle;
-    state.draw.draw_framebuffer = draw_fb_handle;
-    state.Apply();
-
-    u32 buffers = 0;
-
-    if (type == SurfaceType::ColorTexture) {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, src_tex,
-                               0);
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
-                               0);
-
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, dst_tex,
-                               0);
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
-                               0);
-
-        buffers = GL_COLOR_BUFFER_BIT;
-    } else if (type == SurfaceType::Depth) {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D, src_tex, 0);
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
-
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D, dst_tex, 0);
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
-
-        buffers = GL_DEPTH_BUFFER_BIT;
-    } else if (type == SurfaceType::DepthStencil) {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
-                               src_tex, 0);
-
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
-                               dst_tex, 0);
-
-        buffers = GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT;
-    }
-
-    glBlitFramebuffer(src_rect.left, src_rect.bottom, src_rect.right, src_rect.top, dst_rect.left,
-                      dst_rect.bottom, dst_rect.right, dst_rect.top, buffers,
-                      buffers == GL_COLOR_BUFFER_BIT ? GL_LINEAR : GL_NEAREST);
 
+    glCopyImageSubData(src_tex, GL_TEXTURE_2D, 0, src_rect.left, src_rect.bottom, 0, dst_tex,
+                       GL_TEXTURE_2D, 0, dst_rect.left, dst_rect.bottom, 0, src_rect.GetWidth(),
+                       src_rect.GetHeight(), 0);
     return true;
 }
 
@@ -584,7 +575,7 @@ void CachedSurface::UploadGLTexture(const MathUtil::Rectangle<u32>& rect, GLuint
         glCompressedTexImage2D(GL_TEXTURE_2D, 0, tuple.internal_format,
                                static_cast<GLsizei>(rect.GetWidth() * GetCompresssionFactor()),
                                static_cast<GLsizei>(rect.GetHeight() * GetCompresssionFactor()), 0,
-                               size, &gl_buffer[buffer_offset]);
+                               static_cast<GLsizei>(size), &gl_buffer[buffer_offset]);
     } else {
         glTexSubImage2D(GL_TEXTURE_2D, 0, x0, y0, static_cast<GLsizei>(rect.GetWidth()),
                         static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
@@ -923,8 +914,6 @@ Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, ScaleMatc
     // Use GetSurfaceSubRect instead
     ASSERT(params.width == params.stride);
 
-    ASSERT(!params.is_tiled || (params.width % 8 == 0 && params.height % 8 == 0));
-
     // Check for an exact match in existing surfaces
     Surface surface =
         FindMatch<MatchFlags::Exact | MatchFlags::Invalid>(surface_cache, params, match_res_scale);
@@ -1067,8 +1056,11 @@ Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextu
     params.addr = config.tic.Address();
     params.is_tiled = config.tic.IsTiled();
     params.pixel_format = SurfaceParams::PixelFormatFromTextureFormat(config.tic.format);
-    params.width = config.tic.Width() / params.GetCompresssionFactor();
-    params.height = config.tic.Height() / params.GetCompresssionFactor();
+
+    params.width = Common::AlignUp(config.tic.Width(), params.GetCompresssionFactor()) /
+                   params.GetCompresssionFactor();
+    params.height = Common::AlignUp(config.tic.Height(), params.GetCompresssionFactor()) /
+                    params.GetCompresssionFactor();
 
     // TODO(Subv): Different types per component are not supported.
     ASSERT(config.tic.r_type.Value() == config.tic.g_type.Value() &&
@@ -1079,6 +1071,13 @@ Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextu
 
     if (config.tic.IsTiled()) {
         params.block_height = config.tic.BlockHeight();
+
+        // TODO(bunnei): The below align up is a hack. This is here because some compressed textures
+        // are not a multiple of their own compression factor, and so this accounts for that. This
+        // could potentially result in an extra row of 4px being decoded if a texture is not a
+        // multiple of 4.
+        params.width = Common::AlignUp(params.width, 4);
+        params.height = Common::AlignUp(params.height, 4);
     } else {
         // Use the texture-provided stride value if the texture isn't tiled.
         params.stride = static_cast<u32>(params.PixelsInBytes(config.tic.Pitch()));
@@ -1086,23 +1085,6 @@ Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextu
 
     params.UpdateParams();
 
-    if (config.tic.Width() % 8 != 0 || config.tic.Height() % 8 != 0 ||
-        params.stride != params.width) {
-        Surface src_surface;
-        MathUtil::Rectangle<u32> rect;
-        std::tie(src_surface, rect) = GetSurfaceSubRect(params, ScaleMatch::Ignore, true);
-
-        params.res_scale = src_surface->res_scale;
-        Surface tmp_surface = CreateSurface(params);
-        BlitTextures(src_surface->texture.handle, rect, tmp_surface->texture.handle,
-                     tmp_surface->GetScaledRect(),
-                     SurfaceParams::GetFormatType(params.pixel_format), read_framebuffer.handle,
-                     draw_framebuffer.handle);
-
-        remove_surfaces.emplace(tmp_surface);
-        return tmp_surface;
-    }
-
     return GetSurface(params, ScaleMatch::Ignore, true);
 }
 
@@ -1277,7 +1259,7 @@ void RasterizerCacheOpenGL::ValidateSurface(const Surface& surface, Tegra::GPUVA
 
         const auto interval = *it & validate_interval;
         // Look for a valid surface to copy from
-        SurfaceParams params = surface->FromInterval(interval);
+        SurfaceParams params = *surface;
 
         Surface copy_surface =
             FindMatch<MatchFlags::Copy>(surface_cache, params, ScaleMatch::Ignore, interval);

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -57,9 +57,15 @@ struct SurfaceParams {
         ABGR8 = 0,
         B5G6R5 = 1,
         A2B10G10R10 = 2,
-        DXT1 = 3,
-        DXT23 = 4,
-        DXT45 = 5,
+        A1B5G5R5 = 3,
+        R8 = 4,
+        RGBA16F = 5,
+        R11FG11FB10F = 6,
+        DXT1 = 7,
+        DXT23 = 8,
+        DXT45 = 9,
+        DXN1 = 10, // This is also known as BC4
+        ASTC_2D_4X4 = 11,
 
         Max,
         Invalid = 255,
@@ -98,9 +104,15 @@ struct SurfaceParams {
             1, // ABGR8
             1, // B5G6R5
             1, // A2B10G10R10
+            1, // A1B5G5R5
+            1, // R8
+            1, // RGBA16F
+            1, // R11FG11FB10F
             4, // DXT1
             4, // DXT23
             4, // DXT45
+            4, // DXN1
+            1, // ASTC_2D_4X4
         }};
 
         ASSERT(static_cast<size_t>(format) < compression_factor_table.size());
@@ -118,9 +130,15 @@ struct SurfaceParams {
             32,  // ABGR8
             16,  // B5G6R5
             32,  // A2B10G10R10
+            16,  // A1B5G5R5
+            8,   // R8
+            64,  // RGBA16F
+            32,  // R11FG11FB10F
             64,  // DXT1
             128, // DXT23
             128, // DXT45
+            64,  // DXN1
+            32,  // ASTC_2D_4X4
         }};
 
         ASSERT(static_cast<size_t>(format) < bpp_table.size());
@@ -133,15 +151,29 @@ 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;
+        case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
+            return PixelFormat::R11FG11FB10F;
         default:
             NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
         }
     }
 
+    static bool IsFormatASTC(PixelFormat format) {
+        switch (format) {
+        case PixelFormat::ASTC_2D_4X4:
+            return true;
+        default:
+            return false;
+        }
+    }
+
     static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
         switch (format) {
         case Tegra::FramebufferConfig::PixelFormat::ABGR8:
@@ -161,12 +193,24 @@ 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::BF10GF11RF11:
+            return PixelFormat::R11FG11FB10F;
         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;
+        case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
+            return PixelFormat::ASTC_2D_4X4;
         default:
             NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
@@ -182,12 +226,24 @@ 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::R11FG11FB10F:
+            return Tegra::Texture::TextureFormat::BF10GF11RF11;
         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;
+        case PixelFormat::ASTC_2D_4X4:
+            return Tegra::Texture::TextureFormat::ASTC_2D_4X4;
         default:
             UNREACHABLE();
         }
@@ -208,8 +264,12 @@ 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:
+        case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
+            return ComponentType::Float;
         default:
             NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();

src/video_core/renderer_opengl/gl_shader_gen.cpp

@@ -39,6 +39,10 @@ void main() {
     // Viewport can be flipped, which is unsupported by glViewport
     position.xy *= viewport_flip.xy;
     gl_Position = position;
+
+    // TODO(bunnei): This is likely a hack, position.w should be interpolated as 1.0
+    // For now, this is here to bring order in lieu of proper emulation
+    position.w = 1.0;
 }
 )";
     out += program.first;
@@ -62,8 +66,6 @@ layout (std140) uniform fs_config {
     vec4 viewport_flip;
 };
 
-uniform sampler2D tex[32];
-
 void main() {
     exec_shader();
 }

src/video_core/renderer_opengl/gl_shader_gen.h

@@ -22,17 +22,28 @@ class ConstBufferEntry {
     using Maxwell = Tegra::Engines::Maxwell3D::Regs;
 
 public:
-    void MarkAsUsed(unsigned index, unsigned offset, Maxwell::ShaderStage stage) {
+    void MarkAsUsed(u64 index, u64 offset, Maxwell::ShaderStage stage) {
         is_used = true;
-        this->index = index;
+        this->index = static_cast<unsigned>(index);
+        this->stage = stage;
+        max_offset = std::max(max_offset, static_cast<unsigned>(offset));
+    }
+
+    void MarkAsUsedIndirect(u64 index, Maxwell::ShaderStage stage) {
+        is_used = true;
+        is_indirect = true;
+        this->index = static_cast<unsigned>(index);
         this->stage = stage;
-        max_offset = std::max(max_offset, offset);
     }
 
     bool IsUsed() const {
         return is_used;
     }
 
+    bool IsIndirect() const {
+        return is_indirect;
+    }
+
     unsigned GetIndex() const {
         return index;
     }
@@ -51,13 +62,54 @@ private:
     };
 
     bool is_used{};
+    bool is_indirect{};
     unsigned index{};
     unsigned max_offset{};
     Maxwell::ShaderStage stage;
 };
 
+class SamplerEntry {
+    using Maxwell = Tegra::Engines::Maxwell3D::Regs;
+
+public:
+    SamplerEntry(Maxwell::ShaderStage stage, size_t offset, size_t index)
+        : offset(offset), stage(stage), sampler_index(index) {}
+
+    size_t GetOffset() const {
+        return offset;
+    }
+
+    size_t GetIndex() const {
+        return sampler_index;
+    }
+
+    Maxwell::ShaderStage GetStage() const {
+        return stage;
+    }
+
+    std::string GetName() const {
+        return std::string(TextureSamplerNames[static_cast<size_t>(stage)]) + '[' +
+               std::to_string(sampler_index) + ']';
+    }
+
+    static std::string GetArrayName(Maxwell::ShaderStage stage) {
+        return TextureSamplerNames[static_cast<size_t>(stage)];
+    }
+
+private:
+    static constexpr std::array<const char*, Maxwell::MaxShaderStage> TextureSamplerNames = {
+        "tex_vs", "tex_tessc", "tex_tesse", "tex_gs", "tex_fs",
+    };
+    /// Offset in TSC memory from which to read the sampler object, as specified by the sampling
+    /// instruction.
+    size_t offset;
+    Maxwell::ShaderStage stage; ///< Shader stage where this sampler was used.
+    size_t sampler_index;       ///< Value used to index into the generated GLSL sampler array.
+};
+
 struct ShaderEntries {
     std::vector<ConstBufferEntry> const_buffer_entries;
+    std::vector<SamplerEntry> texture_samplers;
 };
 
 using ProgramResult = std::pair<std::string, ShaderEntries>;

src/video_core/renderer_opengl/gl_shader_manager.cpp

@@ -32,33 +32,14 @@ void SetShaderUniformBlockBindings(GLuint shader) {
                                  sizeof(MaxwellUniformData));
 }
 
-void SetShaderSamplerBindings(GLuint shader) {
-    OpenGLState cur_state = OpenGLState::GetCurState();
-    GLuint old_program = std::exchange(cur_state.draw.shader_program, shader);
-    cur_state.Apply();
-
-    // Set the texture samplers to correspond to different texture units
-    for (u32 texture = 0; texture < NumTextureSamplers; ++texture) {
-        // Set the texture samplers to correspond to different texture units
-        std::string uniform_name = "tex[" + std::to_string(texture) + "]";
-        GLint uniform_tex = glGetUniformLocation(shader, uniform_name.c_str());
-        if (uniform_tex != -1) {
-            glUniform1i(uniform_tex, TextureUnits::MaxwellTexture(texture).id);
-        }
-    }
-
-    cur_state.draw.shader_program = old_program;
-    cur_state.Apply();
-}
-
 } // namespace Impl
 
 void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage) {
     const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
 
     // TODO(bunnei): Support more than one viewport
-    viewport_flip[0] = regs.viewport_transform[0].scale_x < 0.0 ? -1.0 : 1.0;
-    viewport_flip[1] = regs.viewport_transform[0].scale_y < 0.0 ? -1.0 : 1.0;
+    viewport_flip[0] = regs.viewport_transform[0].scale_x < 0.0 ? -1.0f : 1.0f;
+    viewport_flip[1] = regs.viewport_transform[0].scale_y < 0.0 ? -1.0f : 1.0f;
 }
 
 } // namespace GLShader

src/video_core/renderer_opengl/gl_shader_manager.h

@@ -45,7 +45,6 @@ public:
         shader.Create(program_result.first.c_str(), type);
         program.Create(true, shader.handle);
         Impl::SetShaderUniformBlockBindings(program.handle);
-        Impl::SetShaderSamplerBindings(program.handle);
         entries = program_result.second;
     }
     GLuint GetHandle() const {

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)), ...);