spelling error

This fixes RequestUpdateAudioRenderer deadlocks in games like Puyo Puyo Tetris and games which require a proper section size in games such as Retro City Rampage. This fixes causes various games to start rendering or trying to render

src/common/assert.h

@@ -28,12 +28,19 @@ __declspec(noinline, noreturn)
 }
 
 #define ASSERT(_a_)                                                                                \
-    if (!(_a_)) {                                                                                  \
-    }
+    do                                                                                             \
+        if (!(_a_)) {                                                                              \
+            assert_noinline_call([] { NGLOG_CRITICAL(Debug, "Assertion Failed!"); });              \
+        }                                                                                          \
+    while (0)
 
 #define ASSERT_MSG(_a_, ...)                                                                       \
-    if (!(_a_)) {                                                                                  \
-    }
+    do                                                                                             \
+        if (!(_a_)) {                                                                              \
+            assert_noinline_call(                                                                  \
+                [&] { NGLOG_CRITICAL(Debug, "Assertion Failed!\n" __VA_ARGS__); });                \
+        }                                                                                          \
+    while (0)
 
 #define UNREACHABLE() ASSERT_MSG(false, "Unreachable code!")
 #define UNREACHABLE_MSG(...) ASSERT_MSG(false, __VA_ARGS__)

src/core/CMakeLists.txt

@@ -40,8 +40,6 @@ add_library(core STATIC
     hle/config_mem.h
     hle/ipc.h
     hle/ipc_helpers.h
-    hle/kernel/address_arbiter.cpp
-    hle/kernel/address_arbiter.h
     hle/kernel/client_port.cpp
     hle/kernel/client_port.h
     hle/kernel/client_session.cpp

src/core/hle/kernel/address_arbiter.cpp

@@ -1,173 +0,0 @@
-// 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/common_funcs.h"
-#include "common/common_types.h"
-#include "core/core.h"
-#include "core/hle/kernel/errors.h"
-#include "core/hle/kernel/kernel.h"
-#include "core/hle/kernel/process.h"
-#include "core/hle/kernel/thread.h"
-#include "core/hle/lock.h"
-#include "core/memory.h"
-
-namespace Kernel {
-namespace AddressArbiter {
-
-// Performs actual address waiting logic.
-static ResultCode WaitForAddress(VAddr address, s64 timeout) {
-    SharedPtr<Thread> current_thread = GetCurrentThread();
-    current_thread->arb_wait_address = address;
-    current_thread->status = THREADSTATUS_WAIT_ARB;
-    current_thread->wakeup_callback = nullptr;
-
-    current_thread->WakeAfterDelay(timeout);
-
-    Core::System::GetInstance().CpuCore(current_thread->processor_id).PrepareReschedule();
-    return RESULT_TIMEOUT;
-}
-
-// Gets the threads waiting on an address.
-static void GetThreadsWaitingOnAddress(std::vector<SharedPtr<Thread>>& waiting_threads,
-                                       VAddr address) {
-    auto RetrieveWaitingThreads =
-        [](size_t core_index, std::vector<SharedPtr<Thread>>& waiting_threads, VAddr arb_addr) {
-            const auto& scheduler = Core::System::GetInstance().Scheduler(core_index);
-            auto& thread_list = scheduler->GetThreadList();
-
-            for (auto& thread : thread_list) {
-                if (thread->arb_wait_address == arb_addr)
-                    waiting_threads.push_back(thread);
-            }
-        };
-
-    // Retrieve a list of all threads that are waiting for this address.
-    RetrieveWaitingThreads(0, waiting_threads, address);
-    RetrieveWaitingThreads(1, waiting_threads, address);
-    RetrieveWaitingThreads(2, waiting_threads, address);
-    RetrieveWaitingThreads(3, waiting_threads, address);
-    // 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;
-              });
-}
-
-// Wake up num_to_wake (or all) threads in a vector.
-static void WakeThreads(std::vector<SharedPtr<Thread>>& waiting_threads, s32 num_to_wake) {
-    // Only process up to 'target' threads, unless 'target' is <= 0, in which case process
-    // them all.
-    size_t last = waiting_threads.size();
-    if (num_to_wake > 0)
-        last = num_to_wake;
-
-    // Signal the waiting threads.
-    for (size_t i = 0; i < last; i++) {
-        ASSERT(waiting_threads[i]->status = THREADSTATUS_WAIT_ARB);
-        waiting_threads[i]->SetWaitSynchronizationResult(RESULT_SUCCESS);
-        waiting_threads[i]->arb_wait_address = 0;
-        waiting_threads[i]->ResumeFromWait();
-    }
-}
-
-// Signals an address being waited on.
-ResultCode SignalToAddress(VAddr address, s32 num_to_wake) {
-    // Get threads waiting on the address.
-    std::vector<SharedPtr<Thread>> waiting_threads;
-    GetThreadsWaitingOnAddress(waiting_threads, address);
-
-    WakeThreads(waiting_threads, num_to_wake);
-    return RESULT_SUCCESS;
-}
-
-// Signals an address being waited on and increments its value if equal to the value argument.
-ResultCode IncrementAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake) {
-    // Ensure that we can write to the address.
-    if (!Memory::IsValidVirtualAddress(address)) {
-        return ERR_INVALID_ADDRESS_STATE;
-    }
-
-    if (static_cast<s32>(Memory::Read32(address)) == value) {
-        Memory::Write32(address, static_cast<u32>(value + 1));
-    } else {
-        return ERR_INVALID_STATE;
-    }
-
-    return SignalToAddress(address, num_to_wake);
-}
-
-// Signals an address being waited on and modifies its value based on waiting thread count if equal
-// to the value argument.
-ResultCode ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 value,
-                                                         s32 num_to_wake) {
-    // Ensure that we can write to the address.
-    if (!Memory::IsValidVirtualAddress(address)) {
-        return ERR_INVALID_ADDRESS_STATE;
-    }
-
-    // Get threads waiting on the address.
-    std::vector<SharedPtr<Thread>> waiting_threads;
-    GetThreadsWaitingOnAddress(waiting_threads, address);
-
-    // Determine the modified value depending on the waiting count.
-    s32 updated_value;
-    if (waiting_threads.size() == 0) {
-        updated_value = value - 1;
-    } else if (num_to_wake <= 0 || waiting_threads.size() <= num_to_wake) {
-        updated_value = value + 1;
-    } else {
-        updated_value = value;
-    }
-
-    if (static_cast<s32>(Memory::Read32(address)) == value) {
-        Memory::Write32(address, static_cast<u32>(updated_value));
-    } else {
-        return ERR_INVALID_STATE;
-    }
-
-    WakeThreads(waiting_threads, num_to_wake);
-    return RESULT_SUCCESS;
-}
-
-// Waits on an address if the value passed is less than the argument value, optionally decrementing.
-ResultCode WaitForAddressIfLessThan(VAddr address, s32 value, s64 timeout, bool should_decrement) {
-    // Ensure that we can read the address.
-    if (!Memory::IsValidVirtualAddress(address)) {
-        return ERR_INVALID_ADDRESS_STATE;
-    }
-
-    s32 cur_value = static_cast<s32>(Memory::Read32(address));
-    if (cur_value < value) {
-        Memory::Write32(address, static_cast<u32>(cur_value - 1));
-    } else {
-        return ERR_INVALID_STATE;
-    }
-    // Short-circuit without rescheduling, if timeout is zero.
-    if (timeout == 0) {
-        return RESULT_TIMEOUT;
-    }
-
-    return WaitForAddress(address, timeout);
-}
-
-// Waits on an address if the value passed is equal to the argument value.
-ResultCode WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout) {
-    // Ensure that we can read the address.
-    if (!Memory::IsValidVirtualAddress(address)) {
-        return ERR_INVALID_ADDRESS_STATE;
-    }
-    // Only wait for the address if equal.
-    if (static_cast<s32>(Memory::Read32(address)) != value) {
-        return ERR_INVALID_STATE;
-    }
-    // Short-circuit without rescheduling, if timeout is zero.
-    if (timeout == 0) {
-        return RESULT_TIMEOUT;
-    }
-
-    return WaitForAddress(address, timeout);
-}
-} // namespace AddressArbiter
-} // namespace Kernel

src/core/hle/kernel/address_arbiter.h

@@ -1,32 +0,0 @@
-// Copyright 2018 yuzu emulator team
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include "core/hle/result.h"
-
-namespace Kernel {
-
-namespace AddressArbiter {
-enum class ArbitrationType {
-    WaitIfLessThan = 0,
-    DecrementAndWaitIfLessThan = 1,
-    WaitIfEqual = 2,
-};
-
-enum class SignalType {
-    Signal = 0,
-    IncrementAndSignalIfEqual = 1,
-    ModifyByWaitingCountAndSignalIfEqual = 2,
-};
-
-ResultCode SignalToAddress(VAddr address, s32 num_to_wake);
-ResultCode IncrementAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake);
-ResultCode ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake);
-
-ResultCode WaitForAddressIfLessThan(VAddr address, s32 value, s64 timeout, bool should_decrement);
-ResultCode WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout);
-} // namespace AddressArbiter
-
-} // namespace Kernel

src/core/hle/kernel/errors.h

@@ -20,16 +20,13 @@ enum {
     MaxConnectionsReached = 52,
 
     // Confirmed Switch OS error codes
-    InvalidAddress = 102,
-    InvalidMemoryState = 106,
+    MisalignedAddress = 102,
     InvalidProcessorId = 113,
     InvalidHandle = 114,
     InvalidCombination = 116,
     Timeout = 117,
     SynchronizationCanceled = 118,
     TooLarge = 119,
-    InvalidEnumValue = 120,
-    InvalidState = 125,
 };
 }
 
@@ -42,15 +39,14 @@ constexpr ResultCode ERR_SESSION_CLOSED_BY_REMOTE(-1);
 constexpr ResultCode ERR_PORT_NAME_TOO_LONG(-1);
 constexpr ResultCode ERR_WRONG_PERMISSION(-1);
 constexpr ResultCode ERR_MAX_CONNECTIONS_REACHED(-1);
-constexpr ResultCode ERR_INVALID_ENUM_VALUE(ErrorModule::Kernel, ErrCodes::InvalidEnumValue);
+constexpr ResultCode ERR_INVALID_ENUM_VALUE(-1);
 constexpr ResultCode ERR_INVALID_ENUM_VALUE_FND(-1);
 constexpr ResultCode ERR_INVALID_COMBINATION(-1);
 constexpr ResultCode ERR_INVALID_COMBINATION_KERNEL(-1);
 constexpr ResultCode ERR_OUT_OF_MEMORY(-1);
-constexpr ResultCode ERR_INVALID_ADDRESS(ErrorModule::Kernel, ErrCodes::InvalidAddress);
-constexpr ResultCode ERR_INVALID_ADDRESS_STATE(ErrorModule::Kernel, ErrCodes::InvalidMemoryState);
+constexpr ResultCode ERR_INVALID_ADDRESS(-1);
+constexpr ResultCode ERR_INVALID_ADDRESS_STATE(-1);
 constexpr ResultCode ERR_INVALID_HANDLE(ErrorModule::Kernel, ErrCodes::InvalidHandle);
-constexpr ResultCode ERR_INVALID_STATE(ErrorModule::Kernel, ErrCodes::InvalidState);
 constexpr ResultCode ERR_INVALID_POINTER(-1);
 constexpr ResultCode ERR_INVALID_OBJECT_ADDR(-1);
 constexpr ResultCode ERR_NOT_AUTHORIZED(-1);

src/core/hle/kernel/hle_ipc.cpp

@@ -271,11 +271,6 @@ std::vector<u8> HLERequestContext::ReadBuffer(int buffer_index) const {
 }
 
 size_t HLERequestContext::WriteBuffer(const void* buffer, size_t size, int buffer_index) const {
-    if (size == 0) {
-        NGLOG_WARNING(Core, "skip empty buffer write");
-        return 0;
-    }
-
     const bool is_buffer_b{BufferDescriptorB().size() && BufferDescriptorB()[buffer_index].Size()};
     const size_t buffer_size{GetWriteBufferSize(buffer_index)};
     if (size > buffer_size) {

src/core/hle/kernel/mutex.cpp

@@ -59,7 +59,7 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
                              Handle requesting_thread_handle) {
     // The mutex address must be 4-byte aligned
     if ((address % sizeof(u32)) != 0) {
-        return ResultCode(ErrorModule::Kernel, ErrCodes::InvalidAddress);
+        return ResultCode(ErrorModule::Kernel, ErrCodes::MisalignedAddress);
     }
 
     SharedPtr<Thread> holding_thread = g_handle_table.Get<Thread>(holding_thread_handle);
@@ -97,7 +97,7 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
 ResultCode Mutex::Release(VAddr address) {
     // The mutex address must be 4-byte aligned
     if ((address % sizeof(u32)) != 0) {
-        return ResultCode(ErrorModule::Kernel, ErrCodes::InvalidAddress);
+        return ResultCode(ErrorModule::Kernel, ErrCodes::MisalignedAddress);
     }
 
     auto [thread, num_waiters] = GetHighestPriorityMutexWaitingThread(GetCurrentThread(), address);

src/core/hle/kernel/svc.cpp

@@ -11,7 +11,6 @@
 #include "common/string_util.h"
 #include "core/core.h"
 #include "core/core_timing.h"
-#include "core/hle/kernel/address_arbiter.h"
 #include "core/hle/kernel/client_port.h"
 #include "core/hle/kernel/client_session.h"
 #include "core/hle/kernel/event.h"
@@ -581,7 +580,7 @@ static void SleepThread(s64 nanoseconds) {
     Core::System::GetInstance().PrepareReschedule();
 }
 
-/// Wait process wide key atomic
+/// Signal process wide key atomic
 static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_variable_addr,
                                            Handle thread_handle, s64 nano_seconds) {
     NGLOG_TRACE(
@@ -690,58 +689,6 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target
     return RESULT_SUCCESS;
 }
 
-// Wait for an address (via Address Arbiter)
-static ResultCode WaitForAddress(VAddr address, u32 type, s32 value, s64 timeout) {
-    NGLOG_WARNING(Kernel_SVC, "called, address=0x{:X}, type=0x{:X}, value=0x{:X}, timeout={}",
-                  address, type, value, timeout);
-    // If the passed address is a kernel virtual address, return invalid memory state.
-    if (Memory::IsKernelVirtualAddress(address)) {
-        return ERR_INVALID_ADDRESS_STATE;
-    }
-    // If the address is not properly aligned to 4 bytes, return invalid address.
-    if (address % sizeof(u32) != 0) {
-        return ERR_INVALID_ADDRESS;
-    }
-
-    switch (static_cast<AddressArbiter::ArbitrationType>(type)) {
-    case AddressArbiter::ArbitrationType::WaitIfLessThan:
-        return AddressArbiter::WaitForAddressIfLessThan(address, value, timeout, false);
-    case AddressArbiter::ArbitrationType::DecrementAndWaitIfLessThan:
-        return AddressArbiter::WaitForAddressIfLessThan(address, value, timeout, true);
-    case AddressArbiter::ArbitrationType::WaitIfEqual:
-        return AddressArbiter::WaitForAddressIfEqual(address, value, timeout);
-    default:
-        return ERR_INVALID_ENUM_VALUE;
-    }
-}
-
-// Signals to an address (via Address Arbiter)
-static ResultCode SignalToAddress(VAddr address, u32 type, s32 value, s32 num_to_wake) {
-    NGLOG_WARNING(Kernel_SVC,
-                  "called, address=0x{:X}, type=0x{:X}, value=0x{:X}, num_to_wake=0x{:X}", address,
-                  type, value, num_to_wake);
-    // If the passed address is a kernel virtual address, return invalid memory state.
-    if (Memory::IsKernelVirtualAddress(address)) {
-        return ERR_INVALID_ADDRESS_STATE;
-    }
-    // If the address is not properly aligned to 4 bytes, return invalid address.
-    if (address % sizeof(u32) != 0) {
-        return ERR_INVALID_ADDRESS;
-    }
-
-    switch (static_cast<AddressArbiter::SignalType>(type)) {
-    case AddressArbiter::SignalType::Signal:
-        return AddressArbiter::SignalToAddress(address, num_to_wake);
-    case AddressArbiter::SignalType::IncrementAndSignalIfEqual:
-        return AddressArbiter::IncrementAndSignalToAddressIfEqual(address, value, num_to_wake);
-    case AddressArbiter::SignalType::ModifyByWaitingCountAndSignalIfEqual:
-        return AddressArbiter::ModifyByWaitingCountAndSignalToAddressIfEqual(address, value,
-                                                                             num_to_wake);
-    default:
-        return ERR_INVALID_ENUM_VALUE;
-    }
-}
-
 /// This returns the total CPU ticks elapsed since the CPU was powered-on
 static u64 GetSystemTick() {
     const u64 result{CoreTiming::GetTicks()};
@@ -914,8 +861,8 @@ static const FunctionDef SVC_Table[] = {
     {0x31, nullptr, "GetResourceLimitCurrentValue"},
     {0x32, SvcWrap<SetThreadActivity>, "SetThreadActivity"},
     {0x33, SvcWrap<GetThreadContext>, "GetThreadContext"},
-    {0x34, SvcWrap<WaitForAddress>, "WaitForAddress"},
-    {0x35, SvcWrap<SignalToAddress>, "SignalToAddress"},
+    {0x34, nullptr, "WaitForAddress"},
+    {0x35, nullptr, "SignalToAddress"},
     {0x36, nullptr, "Unknown"},
     {0x37, nullptr, "Unknown"},
     {0x38, nullptr, "Unknown"},

src/core/hle/kernel/svc_wrap.h

@@ -179,20 +179,6 @@ void SvcWrap() {
     FuncReturn(retval);
 }
 
-template <ResultCode func(u64, u32, s32, s64)>
-void SvcWrap() {
-    FuncReturn(
-        func(PARAM(0), (u32)(PARAM(1) & 0xFFFFFFFF), (s32)(PARAM(2) & 0xFFFFFFFF), (s64)PARAM(3))
-            .raw);
-}
-
-template <ResultCode func(u64, u32, s32, s32)>
-void SvcWrap() {
-    FuncReturn(func(PARAM(0), (u32)(PARAM(1) & 0xFFFFFFFF), (s32)(PARAM(2) & 0xFFFFFFFF),
-                    (s32)(PARAM(3) & 0xFFFFFFFF))
-                   .raw);
-}
-
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 // Function wrappers that return type u32
 

src/core/hle/kernel/thread.cpp

@@ -140,11 +140,6 @@ static void ThreadWakeupCallback(u64 thread_handle, int cycles_late) {
         }
     }
 
-    if (thread->arb_wait_address != 0) {
-        ASSERT(thread->status == THREADSTATUS_WAIT_ARB);
-        thread->arb_wait_address = 0;
-    }
-
     if (resume)
         thread->ResumeFromWait();
 }
@@ -184,7 +179,6 @@ void Thread::ResumeFromWait() {
     case THREADSTATUS_WAIT_SLEEP:
     case THREADSTATUS_WAIT_IPC:
     case THREADSTATUS_WAIT_MUTEX:
-    case THREADSTATUS_WAIT_ARB:
         break;
 
     case THREADSTATUS_READY:

src/core/hle/kernel/thread.h

@@ -45,7 +45,6 @@ enum ThreadStatus {
     THREADSTATUS_WAIT_SYNCH_ANY, ///< Waiting due to WaitSynch1 or WaitSynchN with wait_all = false
     THREADSTATUS_WAIT_SYNCH_ALL, ///< Waiting due to WaitSynchronizationN with wait_all = true
     THREADSTATUS_WAIT_MUTEX,     ///< Waiting due to an ArbitrateLock/WaitProcessWideKey svc
-    THREADSTATUS_WAIT_ARB,       ///< Waiting due to a SignalToAddress/WaitForAddress svc
     THREADSTATUS_DORMANT,        ///< Created but not yet made ready
     THREADSTATUS_DEAD            ///< Run to completion, or forcefully terminated
 };
@@ -231,9 +230,6 @@ public:
     VAddr mutex_wait_address; ///< If waiting on a Mutex, this is the mutex address
     Handle wait_handle;       ///< The handle used to wait for the mutex.
 
-    // If waiting for an AddressArbiter, this is the address being waited on.
-    VAddr arb_wait_address{0};
-
     std::string name;
 
     /// Handle used by guest emulated application to access this thread

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

@@ -17,7 +17,7 @@ constexpr u64 audio_ticks{static_cast<u64>(CoreTiming::BASE_CLOCK_RATE / 200)};
 
 class IAudioRenderer final : public ServiceFramework<IAudioRenderer> {
 public:
-    IAudioRenderer(AudioRendererParameter audren_params)
+    IAudioRenderer(AudioRendererParameters audren_params)
         : ServiceFramework("IAudioRenderer"), worker_params(audren_params) {
         static const FunctionInfo functions[] = {
             {0, nullptr, "GetAudioRendererSampleRate"},
@@ -58,22 +58,22 @@ private:
     }
 
     void RequestUpdateAudioRenderer(Kernel::HLERequestContext& ctx) {
-        UpdateDataHeader config{};
+        AudioRendererConfig config;
         auto buf = ctx.ReadBuffer();
-        std::memcpy(&config, buf.data(), sizeof(UpdateDataHeader));
+        std::memcpy(&config, buf.data(), sizeof(AudioRendererConfig));
         u32 memory_pool_count = worker_params.effect_count + (worker_params.voice_count * 4);
 
         std::vector<MemoryPoolInfo> mem_pool_info(memory_pool_count);
         std::memcpy(mem_pool_info.data(),
-                    buf.data() + sizeof(UpdateDataHeader) + config.behavior_size,
+                    buf.data() + sizeof(AudioRendererConfig) + config.behavior_size,
                     memory_pool_count * sizeof(MemoryPoolInfo));
 
-        UpdateDataHeader response_data{worker_params};
+        AudioRendererResponse response_data{worker_params};
 
         ASSERT(ctx.GetWriteBufferSize() == response_data.total_size);
 
         std::vector<u8> output(response_data.total_size);
-        std::memcpy(output.data(), &response_data, sizeof(UpdateDataHeader));
+        std::memcpy(output.data(), &response_data, sizeof(AudioRendererResponse));
         std::vector<MemoryPoolEntry> memory_pool(memory_pool_count);
         for (unsigned i = 0; i < memory_pool.size(); i++) {
             if (mem_pool_info[i].pool_state == MemoryPoolStates::RequestAttach)
@@ -83,7 +83,7 @@ private:
             else
                 memory_pool[i].state = mem_pool_info[i].pool_state;
         }
-        std::memcpy(output.data() + sizeof(UpdateDataHeader), memory_pool.data(),
+        std::memcpy(output.data() + sizeof(AudioRendererResponse), memory_pool.data(),
                     response_data.memory_pools_size);
 
         ctx.WriteBuffer(output);
@@ -146,40 +146,53 @@ private:
     };
     static_assert(sizeof(MemoryPoolInfo) == 0x20, "MemoryPoolInfo has wrong size");
 
-    struct UpdateDataHeader {
-        UpdateDataHeader() {}
+    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(AudioRendererConfig) == 0x40, "AudioRendererConfig has wrong size");
 
-        UpdateDataHeader(const AudioRendererParameter& config) {
-            revision = Common::MakeMagic('R', 'E', 'V', '4'); // 5.1.0 Revision
-            behavior_size = 0xb0;
+    struct AudioRendererResponse {
+        AudioRendererResponse(const AudioRendererParameters& config) {
+            revision = config.revision;
+            error_info_size = 0xb0;
             memory_pools_size = (config.effect_count + (config.voice_count * 4)) * 0x10;
             voices_size = config.voice_count * 0x10;
             effects_size = config.effect_count * 0x10;
             sinks_size = config.sink_count * 0x20;
             performance_manager_size = 0x10;
-            total_size = sizeof(UpdateDataHeader) + behavior_size + memory_pools_size +
+            total_size = sizeof(AudioRendererResponse) + error_info_size + memory_pools_size +
                          voices_size + effects_size + sinks_size + performance_manager_size;
         }
 
         u32_le revision;
-        u32_le behavior_size;
+        u32_le error_info_size;
         u32_le memory_pools_size;
         u32_le voices_size;
-        u32_le voice_resource_size;
+        u32_le unknown_10;
         u32_le effects_size;
-        u32_le mixes_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(UpdateDataHeader) == 0x40, "UpdateDataHeader has wrong size");
+    static_assert(sizeof(AudioRendererResponse) == 0x40, "AudioRendererResponse has wrong size");
 
     /// This is used to trigger the audio event callback.
     CoreTiming::EventType* audio_event;
 
     Kernel::SharedPtr<Kernel::Event> system_event;
-    AudioRendererParameter worker_params;
+    AudioRendererParameters worker_params;
 };
 
 class IAudioDevice final : public ServiceFramework<IAudioDevice> {
@@ -278,7 +291,7 @@ AudRenU::AudRenU() : ServiceFramework("audren:u") {
 
 void AudRenU::OpenAudioRenderer(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
-    auto params = rp.PopRaw<AudioRendererParameter>();
+    auto params = rp.PopRaw<AudioRendererParameters>();
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
 
     rb.Push(RESULT_SUCCESS);
@@ -289,7 +302,7 @@ void AudRenU::OpenAudioRenderer(Kernel::HLERequestContext& ctx) {
 
 void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
-    auto params = rp.PopRaw<AudioRendererParameter>();
+    auto params = rp.PopRaw<AudioRendererParameters>();
 
     u64 buffer_sz = Common::AlignUp(4 * params.unknown_8, 0x40);
     buffer_sz += params.unknown_c * 1024;

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

@@ -12,7 +12,7 @@ class HLERequestContext;
 
 namespace Service::Audio {
 
-struct AudioRendererParameter {
+struct AudioRendererParameters {
     u32_le sample_rate;
     u32_le sample_count;
     u32_le unknown_8;
@@ -28,7 +28,7 @@ struct AudioRendererParameter {
     INSERT_PADDING_WORDS(1);
     u32_le revision;
 };
-static_assert(sizeof(AudioRendererParameter) == 52, "AudioRendererParameter is an invalid size");
+static_assert(sizeof(AudioRendererParameters) == 52, "AudioRendererParameters is an invalid size");
 
 class AudRenU final : public ServiceFramework<AudRenU> {
 public:

src/core/memory.cpp

@@ -241,10 +241,6 @@ bool IsValidVirtualAddress(const VAddr vaddr) {
     return IsValidVirtualAddress(*Core::CurrentProcess(), vaddr);
 }
 
-bool IsKernelVirtualAddress(const VAddr vaddr) {
-    return KERNEL_REGION_VADDR <= vaddr && vaddr < KERNEL_REGION_END;
-}
-
 bool IsValidPhysicalAddress(const PAddr paddr) {
     return GetPhysicalPointer(paddr) != nullptr;
 }

src/core/memory.h

@@ -188,11 +188,6 @@ enum : VAddr {
     MAP_REGION_VADDR = NEW_MAP_REGION_VADDR_END,
     MAP_REGION_SIZE = 0x1000000000,
     MAP_REGION_VADDR_END = MAP_REGION_VADDR + MAP_REGION_SIZE,
-
-    /// Kernel Virtual Address Range
-    KERNEL_REGION_VADDR = 0xFFFFFF8000000000,
-    KERNEL_REGION_SIZE = 0x7FFFE00000,
-    KERNEL_REGION_END = KERNEL_REGION_VADDR + KERNEL_REGION_SIZE,
 };
 
 /// Currently active page table
@@ -202,8 +197,6 @@ PageTable* GetCurrentPageTable();
 /// Determines if the given VAddr is valid for the specified process.
 bool IsValidVirtualAddress(const Kernel::Process& process, const VAddr vaddr);
 bool IsValidVirtualAddress(const VAddr addr);
-/// Determines if the given VAddr is a kernel address
-bool IsKernelVirtualAddress(const VAddr addr);
 
 bool IsValidPhysicalAddress(const PAddr addr);
 

src/video_core/debug_utils/debug_utils.h

@@ -55,10 +55,8 @@ public:
         virtual ~BreakPointObserver() {
             auto context = context_weak.lock();
             if (context) {
-                {
-                    std::unique_lock<std::mutex> lock(context->breakpoint_mutex);
-                    context->breakpoint_observers.remove(this);
-                }
+                std::unique_lock<std::mutex> lock(context->breakpoint_mutex);
+                context->breakpoint_observers.remove(this);
 
                 // If we are the last observer to be destroyed, tell the debugger context that
                 // it is free to continue. In particular, this is required for a proper yuzu

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -197,8 +197,8 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) {
     ASSERT_MSG(!gpu.regs.shader_config[0].enable, "VertexA is unsupported!");
 
     // Next available bindpoints to use when uploading the const buffers and textures to the GLSL
-    // shaders. The constbuffer bindpoint starts after the shader stage configuration bind points.
-    u32 current_constbuffer_bindpoint = uniform_buffers.size();
+    // shaders.
+    u32 current_constbuffer_bindpoint = 0;
     u32 current_texture_bindpoint = 0;
 
     for (unsigned index = 1; index < Maxwell::MaxShaderProgram; ++index) {
@@ -437,7 +437,7 @@ void RasterizerOpenGL::DrawArrays() {
 
     // Unbind textures for potential future use as framebuffer attachments
     for (auto& texture_unit : state.texture_units) {
-        texture_unit.Unbind();
+        texture_unit.texture_2d = 0;
     }
     state.Apply();
 
@@ -608,39 +608,27 @@ u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, GLuint progr
 
         boost::optional<VAddr> addr = gpu.memory_manager->GpuToCpuAddress(buffer.address);
 
-        size_t size = 0;
-
+        std::vector<u8> data;
         if (used_buffer.IsIndirect()) {
             // Buffer is accessed indirectly, so upload the entire thing
-            size = buffer.size * sizeof(float);
-
-            if (size > MaxConstbufferSize) {
-                NGLOG_ERROR(HW_GPU, "indirect constbuffer size {} exceeds maximum {}", size,
-                            MaxConstbufferSize);
-                size = MaxConstbufferSize;
-            }
+            data.resize(buffer.size * sizeof(float));
         } else {
             // Buffer is accessed directly, upload just what we use
-            size = used_buffer.GetSize() * sizeof(float);
+            data.resize(used_buffer.GetSize() * sizeof(float));
         }
 
-        // Align the actual size so it ends up being a multiple of vec4 to meet the OpenGL std140
-        // UBO alignment requirements.
-        size = Common::AlignUp(size, sizeof(GLvec4));
-        ASSERT_MSG(size <= MaxConstbufferSize, "Constbuffer too big");
-
-        std::vector<u8> data(size);
         Memory::ReadBlock(*addr, data.data(), data.size());
 
-        glBindBuffer(GL_UNIFORM_BUFFER, buffer_draw_state.ssbo);
-        glBufferData(GL_UNIFORM_BUFFER, data.size(), data.data(), GL_DYNAMIC_DRAW);
-        glBindBuffer(GL_UNIFORM_BUFFER, 0);
+        glBindBuffer(GL_SHADER_STORAGE_BUFFER, buffer_draw_state.ssbo);
+        glBufferData(GL_SHADER_STORAGE_BUFFER, data.size(), data.data(), GL_DYNAMIC_DRAW);
+        glBindBuffer(GL_SHADER_STORAGE_BUFFER, 0);
 
         // Now configure the bindpoint of the buffer inside the shader
         std::string buffer_name = used_buffer.GetName();
-        GLuint index = glGetProgramResourceIndex(program, GL_UNIFORM_BLOCK, buffer_name.c_str());
+        GLuint index =
+            glGetProgramResourceIndex(program, GL_SHADER_STORAGE_BLOCK, buffer_name.c_str());
         if (index != -1)
-            glUniformBlockBinding(program, index, buffer_draw_state.bindpoint);
+            glShaderStorageBlockBinding(program, index, buffer_draw_state.bindpoint);
     }
 
     state.Apply();

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -54,11 +54,6 @@ public:
         OGLShader shader;
     };
 
-    /// Maximum supported size that a constbuffer can have in bytes.
-    static constexpr size_t MaxConstbufferSize = 0x10000;
-    static_assert(MaxConstbufferSize % sizeof(GLvec4) == 0,
-                  "The maximum size of a constbuffer must be a multiple of the size of GLvec4");
-
 private:
     class SamplerInfo {
     public:

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -46,7 +46,7 @@ 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
@@ -645,7 +645,7 @@ void CachedSurface::DownloadGLTexture(const MathUtil::Rectangle<u32>& rect, GLui
         glActiveTexture(GL_TEXTURE0);
         glGetTexImage(GL_TEXTURE_2D, 0, tuple.format, tuple.type, &gl_buffer[buffer_offset]);
     } else {
-        state.UnbindTexture(texture.handle);
+        state.ResetTexture(texture.handle);
         state.draw.read_framebuffer = read_fb_handle;
         state.Apply();
 

src/video_core/renderer_opengl/gl_resource_manager.h

@@ -38,7 +38,7 @@ public:
         if (handle == 0)
             return;
         glDeleteTextures(1, &handle);
-        OpenGLState::GetCurState().UnbindTexture(handle).Apply();
+        OpenGLState::GetCurState().ResetTexture(handle).Apply();
         handle = 0;
     }
 

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -9,7 +9,6 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "video_core/engines/shader_bytecode.h"
-#include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/gl_shader_decompiler.h"
 
 namespace GLShader {
@@ -398,8 +397,7 @@ public:
     /// Generates code representing a uniform (C buffer) register, interpreted as the input type.
     std::string GetUniform(u64 index, u64 offset, GLSLRegister::Type type) {
         declr_const_buffers[index].MarkAsUsed(index, offset, stage);
-        std::string value = 'c' + std::to_string(index) + '[' + std::to_string(offset / 4) + "][" +
-                            std::to_string(offset % 4) + ']';
+        std::string value = 'c' + std::to_string(index) + '[' + std::to_string(offset) + ']';
 
         if (type == GLSLRegister::Type::Float) {
             return value;
@@ -413,12 +411,8 @@ public:
     std::string GetUniformIndirect(u64 index, s64 offset, const Register& index_reg,
                                    GLSLRegister::Type type) {
         declr_const_buffers[index].MarkAsUsedIndirect(index, stage);
-
-        std::string final_offset = "((floatBitsToInt(" + GetRegister(index_reg, 0) + ") + " +
-                                   std::to_string(offset) + ") / 4)";
-
-        std::string value =
-            'c' + std::to_string(index) + '[' + final_offset + " / 4][" + final_offset + " % 4]";
+        std::string value = 'c' + std::to_string(index) + "[(floatBitsToInt(" +
+                            GetRegister(index_reg, 0) + ") + " + std::to_string(offset) + ") / 4]";
 
         if (type == GLSLRegister::Type::Float) {
             return value;
@@ -460,10 +454,9 @@ public:
 
         unsigned const_buffer_layout = 0;
         for (const auto& entry : GetConstBuffersDeclarations()) {
-            declarations.AddLine("layout(std140) uniform " + entry.GetName());
+            declarations.AddLine("layout(std430) buffer " + entry.GetName());
             declarations.AddLine('{');
-            declarations.AddLine("    vec4 c" + std::to_string(entry.GetIndex()) +
-                                 "[MAX_CONSTBUFFER_ELEMENTS];");
+            declarations.AddLine("    float c" + std::to_string(entry.GetIndex()) + "[];");
             declarations.AddLine("};");
             declarations.AddNewLine();
             ++const_buffer_layout;
@@ -811,7 +804,6 @@ private:
         if (!opcode) {
             NGLOG_CRITICAL(HW_GPU, "Unhandled instruction: {0:x}", instr.value);
             UNREACHABLE();
-            return offset + 1;
         }
 
         shader.AddLine("// " + std::to_string(offset) + ": " + opcode->GetName());
@@ -1721,10 +1713,7 @@ private:
 }; // namespace Decompiler
 
 std::string GetCommonDeclarations() {
-    std::string declarations = "bool exec_shader();\n";
-    declarations += "#define MAX_CONSTBUFFER_ELEMENTS " +
-                    std::to_string(RasterizerOpenGL::MaxConstbufferSize / (sizeof(GLvec4)));
-    return declarations;
+    return "bool exec_shader();";
 }
 
 boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,

src/video_core/renderer_opengl/gl_state.cpp

@@ -48,9 +48,24 @@ OpenGLState::OpenGLState() {
     logic_op = GL_COPY;
 
     for (auto& texture_unit : texture_units) {
-        texture_unit.Reset();
+        texture_unit.texture_2d = 0;
+        texture_unit.sampler = 0;
+        texture_unit.swizzle.r = GL_RED;
+        texture_unit.swizzle.g = GL_GREEN;
+        texture_unit.swizzle.b = GL_BLUE;
+        texture_unit.swizzle.a = GL_ALPHA;
     }
 
+    lighting_lut.texture_buffer = 0;
+
+    fog_lut.texture_buffer = 0;
+
+    proctex_lut.texture_buffer = 0;
+    proctex_diff_lut.texture_buffer = 0;
+    proctex_color_map.texture_buffer = 0;
+    proctex_alpha_map.texture_buffer = 0;
+    proctex_noise_lut.texture_buffer = 0;
+
     draw.read_framebuffer = 0;
     draw.draw_framebuffer = 0;
     draw.vertex_array = 0;
@@ -208,12 +223,54 @@ void OpenGLState::Apply() const {
             if (current.enabled != new_state.enabled || current.bindpoint != new_state.bindpoint ||
                 current.ssbo != new_state.ssbo) {
                 if (new_state.enabled) {
-                    glBindBufferBase(GL_UNIFORM_BUFFER, new_state.bindpoint, new_state.ssbo);
+                    glBindBufferBase(GL_SHADER_STORAGE_BUFFER, new_state.bindpoint, new_state.ssbo);
                 }
             }
         }
     }
 
+    // Lighting LUTs
+    if (lighting_lut.texture_buffer != cur_state.lighting_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::LightingLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, lighting_lut.texture_buffer);
+    }
+
+    // Fog LUT
+    if (fog_lut.texture_buffer != cur_state.fog_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::FogLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, fog_lut.texture_buffer);
+    }
+
+    // ProcTex Noise LUT
+    if (proctex_noise_lut.texture_buffer != cur_state.proctex_noise_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexNoiseLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_noise_lut.texture_buffer);
+    }
+
+    // ProcTex Color Map
+    if (proctex_color_map.texture_buffer != cur_state.proctex_color_map.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexColorMap.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_color_map.texture_buffer);
+    }
+
+    // ProcTex Alpha Map
+    if (proctex_alpha_map.texture_buffer != cur_state.proctex_alpha_map.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexAlphaMap.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_alpha_map.texture_buffer);
+    }
+
+    // ProcTex LUT
+    if (proctex_lut.texture_buffer != cur_state.proctex_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_lut.texture_buffer);
+    }
+
+    // ProcTex Diff LUT
+    if (proctex_diff_lut.texture_buffer != cur_state.proctex_diff_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexDiffLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_diff_lut.texture_buffer);
+    }
+
     // Framebuffer
     if (draw.read_framebuffer != cur_state.draw.read_framebuffer) {
         glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer);
@@ -281,12 +338,26 @@ void OpenGLState::Apply() const {
     cur_state = *this;
 }
 
-OpenGLState& OpenGLState::UnbindTexture(GLuint handle) {
+OpenGLState& OpenGLState::ResetTexture(GLuint handle) {
     for (auto& unit : texture_units) {
         if (unit.texture_2d == handle) {
-            unit.Unbind();
+            unit.texture_2d = 0;
         }
     }
+    if (lighting_lut.texture_buffer == handle)
+        lighting_lut.texture_buffer = 0;
+    if (fog_lut.texture_buffer == handle)
+        fog_lut.texture_buffer = 0;
+    if (proctex_noise_lut.texture_buffer == handle)
+        proctex_noise_lut.texture_buffer = 0;
+    if (proctex_color_map.texture_buffer == handle)
+        proctex_color_map.texture_buffer = 0;
+    if (proctex_alpha_map.texture_buffer == handle)
+        proctex_alpha_map.texture_buffer = 0;
+    if (proctex_lut.texture_buffer == handle)
+        proctex_lut.texture_buffer = 0;
+    if (proctex_diff_lut.texture_buffer == handle)
+        proctex_diff_lut.texture_buffer = 0;
     return *this;
 }
 

src/video_core/renderer_opengl/gl_state.h

@@ -91,21 +91,36 @@ public:
             GLint b; // GL_TEXTURE_SWIZZLE_B
             GLint a; // GL_TEXTURE_SWIZZLE_A
         } swizzle;
-
-        void Unbind() {
-            texture_2d = 0;
-            swizzle.r = GL_RED;
-            swizzle.g = GL_GREEN;
-            swizzle.b = GL_BLUE;
-            swizzle.a = GL_ALPHA;
-        }
-
-        void Reset() {
-            Unbind();
-            sampler = 0;
-        }
     } texture_units[32];
 
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } lighting_lut;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } fog_lut;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_noise_lut;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_color_map;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_alpha_map;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_lut;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_diff_lut;
+
     struct {
         GLuint read_framebuffer; // GL_READ_FRAMEBUFFER_BINDING
         GLuint draw_framebuffer; // GL_DRAW_FRAMEBUFFER_BINDING
@@ -150,7 +165,7 @@ public:
     void Apply() const;
 
     /// Resets any references to the given resource
-    OpenGLState& UnbindTexture(GLuint handle);
+    OpenGLState& ResetTexture(GLuint handle);
     OpenGLState& ResetSampler(GLuint handle);
     OpenGLState& ResetProgram(GLuint handle);
     OpenGLState& ResetPipeline(GLuint handle);

src/yuzu/debugger/wait_tree.cpp

@@ -213,9 +213,6 @@ QString WaitTreeThread::GetText() const {
     case THREADSTATUS_WAIT_MUTEX:
         status = tr("waiting for mutex");
         break;
-    case THREADSTATUS_WAIT_ARB:
-        status = tr("waiting for address arbiter");
-        break;
     case THREADSTATUS_DORMANT:
         status = tr("dormant");
         break;
@@ -243,7 +240,6 @@ QColor WaitTreeThread::GetColor() const {
     case THREADSTATUS_WAIT_SYNCH_ALL:
     case THREADSTATUS_WAIT_SYNCH_ANY:
     case THREADSTATUS_WAIT_MUTEX:
-    case THREADSTATUS_WAIT_ARB:
         return QColor(Qt::GlobalColor::red);
     case THREADSTATUS_DORMANT:
         return QColor(Qt::GlobalColor::darkCyan);

src/yuzu/main.cpp

@@ -334,6 +334,8 @@ bool GMainWindow::SupportsRequiredGLExtensions() {
         unsupported_ext.append("ARB_program_interface_query");
     if (!GLAD_GL_ARB_separate_shader_objects)
         unsupported_ext.append("ARB_separate_shader_objects");
+    if (!GLAD_GL_ARB_shader_storage_buffer_object)
+        unsupported_ext.append("ARB_shader_storage_buffer_object");
     if (!GLAD_GL_ARB_vertex_attrib_binding)
         unsupported_ext.append("ARB_vertex_attrib_binding");