Merge ea5b35a630 into 299f943202

renderer_opengl: Silence a -Wdangling-else warning in DrawScreenTriangles()

src/core/CMakeLists.txt

@@ -42,8 +42,6 @@ add_library(core STATIC
     hle/kernel/client_port.h
     hle/kernel/client_session.cpp
     hle/kernel/client_session.h
-    hle/kernel/condition_variable.cpp
-    hle/kernel/condition_variable.h
     hle/kernel/errors.h
     hle/kernel/event.cpp
     hle/kernel/event.h

src/core/arm/unicorn/arm_unicorn.cpp

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
+#include <cinttypes>
 #include <unicorn/arm64.h>
 #include "common/assert.h"
 #include "common/microprofile.h"
@@ -53,8 +54,10 @@ static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int si
                                void* user_data) {
     ARM_Interface::ThreadContext ctx{};
     Core::CPU().SaveContext(ctx);
-    ASSERT_MSG(false, "Attempted to read from unmapped memory: 0x%lx, pc=0x%lx, lr=0x%lx", addr,
-               ctx.pc, ctx.cpu_registers[30]);
+    ASSERT_MSG(false,
+               "Attempted to read from unmapped memory: 0x%" PRIx64 ", pc=0x%" PRIx64
+               ", lr=0x%" PRIx64,
+               addr, ctx.pc, ctx.cpu_registers[30]);
     return {};
 }
 

src/core/file_sys/disk_filesystem.cpp

@@ -67,10 +67,16 @@ ResultCode Disk_FileSystem::DeleteFile(const std::string& path) const {
     return RESULT_SUCCESS;
 }
 
-ResultCode Disk_FileSystem::RenameFile(const Path& src_path, const Path& dest_path) const {
-    LOG_WARNING(Service_FS, "(STUBBED) called");
+ResultCode Disk_FileSystem::RenameFile(const std::string& src_path,
+                                       const std::string& dest_path) const {
+    const std::string full_src_path = base_directory + src_path;
+    const std::string full_dest_path = base_directory + dest_path;
+
+    if (!FileUtil::Exists(full_src_path)) {
+        return ERROR_PATH_NOT_FOUND;
+    }
     // TODO(wwylele): Use correct error code
-    return ResultCode(-1);
+    return FileUtil::Rename(full_src_path, full_dest_path) ? RESULT_SUCCESS : ResultCode(-1);
 }
 
 ResultCode Disk_FileSystem::DeleteDirectory(const Path& path) const {

src/core/file_sys/disk_filesystem.h

@@ -26,7 +26,7 @@ public:
     ResultVal<std::unique_ptr<StorageBackend>> OpenFile(const std::string& path,
                                                         Mode mode) const override;
     ResultCode DeleteFile(const std::string& path) const override;
-    ResultCode RenameFile(const Path& src_path, const Path& dest_path) const override;
+    ResultCode RenameFile(const std::string& src_path, const std::string& dest_path) const override;
     ResultCode DeleteDirectory(const Path& path) const override;
     ResultCode DeleteDirectoryRecursively(const Path& path) const override;
     ResultCode CreateFile(const std::string& path, u64 size) const override;

src/core/file_sys/filesystem.h

@@ -126,7 +126,8 @@ public:
      * @param dest_path Destination path relative to the archive
      * @return Result of the operation
      */
-    virtual ResultCode RenameFile(const Path& src_path, const Path& dest_path) const = 0;
+    virtual ResultCode RenameFile(const std::string& src_path,
+                                  const std::string& dest_path) const = 0;
 
     /**
      * Rename a Directory specified by its path

src/core/file_sys/romfs_filesystem.cpp

@@ -27,7 +27,8 @@ ResultCode RomFS_FileSystem::DeleteFile(const std::string& path) const {
     return ResultCode(-1);
 }
 
-ResultCode RomFS_FileSystem::RenameFile(const Path& src_path, const Path& dest_path) const {
+ResultCode RomFS_FileSystem::RenameFile(const std::string& src_path,
+                                        const std::string& dest_path) const {
     LOG_CRITICAL(Service_FS, "Attempted to rename a file within an ROMFS archive (%s).",
                  GetName().c_str());
     // TODO(wwylele): Use correct error code

src/core/file_sys/romfs_filesystem.h

@@ -32,7 +32,7 @@ public:
     ResultVal<std::unique_ptr<StorageBackend>> OpenFile(const std::string& path,
                                                         Mode mode) const override;
     ResultCode DeleteFile(const std::string& path) const override;
-    ResultCode RenameFile(const Path& src_path, const Path& dest_path) const override;
+    ResultCode RenameFile(const std::string& src_path, const std::string& dest_path) const override;
     ResultCode DeleteDirectory(const Path& path) const override;
     ResultCode DeleteDirectoryRecursively(const Path& path) const override;
     ResultCode CreateFile(const std::string& path, u64 size) const override;

src/core/hle/kernel/condition_variable.cpp

@@ -1,64 +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 "core/hle/kernel/condition_variable.h"
-#include "core/hle/kernel/errors.h"
-#include "core/hle/kernel/kernel.h"
-#include "core/hle/kernel/object_address_table.h"
-#include "core/hle/kernel/thread.h"
-
-namespace Kernel {
-
-ConditionVariable::ConditionVariable() {}
-ConditionVariable::~ConditionVariable() {}
-
-ResultVal<SharedPtr<ConditionVariable>> ConditionVariable::Create(VAddr guest_addr,
-                                                                  std::string name) {
-    SharedPtr<ConditionVariable> condition_variable(new ConditionVariable);
-
-    condition_variable->name = std::move(name);
-    condition_variable->guest_addr = guest_addr;
-    condition_variable->mutex_addr = 0;
-
-    // Condition variables are referenced by guest address, so track this in the kernel
-    g_object_address_table.Insert(guest_addr, condition_variable);
-
-    return MakeResult<SharedPtr<ConditionVariable>>(std::move(condition_variable));
-}
-
-bool ConditionVariable::ShouldWait(Thread* thread) const {
-    return GetAvailableCount() <= 0;
-}
-
-void ConditionVariable::Acquire(Thread* thread) {
-    if (GetAvailableCount() <= 0)
-        return;
-
-    SetAvailableCount(GetAvailableCount() - 1);
-}
-
-ResultCode ConditionVariable::Release(s32 target) {
-    if (target == -1) {
-        // When -1, wake up all waiting threads
-        SetAvailableCount(static_cast<s32>(GetWaitingThreads().size()));
-        WakeupAllWaitingThreads();
-    } else {
-        // Otherwise, wake up just a single thread
-        SetAvailableCount(target);
-        WakeupWaitingThread(GetHighestPriorityReadyThread());
-    }
-
-    return RESULT_SUCCESS;
-}
-
-s32 ConditionVariable::GetAvailableCount() const {
-    return Memory::Read32(guest_addr);
-}
-
-void ConditionVariable::SetAvailableCount(s32 value) const {
-    Memory::Write32(guest_addr, value);
-}
-
-} // namespace Kernel

src/core/hle/kernel/condition_variable.h

@@ -1,63 +0,0 @@
-// Copyright 2018 yuzu emulator team
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <string>
-#include <queue>
-#include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
-#include "core/hle/kernel/wait_object.h"
-#include "core/hle/result.h"
-
-namespace Kernel {
-
-class ConditionVariable final : public WaitObject {
-public:
-    /**
-     * Creates a condition variable.
-     * @param guest_addr Address of the object tracking the condition variable in guest memory. If
-     * specified, this condition variable will update the guest object when its state changes.
-     * @param name Optional name of condition variable.
-     * @return The created condition variable.
-     */
-    static ResultVal<SharedPtr<ConditionVariable>> Create(VAddr guest_addr,
-                                                          std::string name = "Unknown");
-
-    std::string GetTypeName() const override {
-        return "ConditionVariable";
-    }
-    std::string GetName() const override {
-        return name;
-    }
-
-    static const HandleType HANDLE_TYPE = HandleType::ConditionVariable;
-    HandleType GetHandleType() const override {
-        return HANDLE_TYPE;
-    }
-
-    s32 GetAvailableCount() const;
-    void SetAvailableCount(s32 value) const;
-
-    std::string name; ///< Name of condition variable (optional)
-    VAddr guest_addr; ///< Address of the guest condition variable value
-    VAddr mutex_addr; ///< (optional) Address of guest mutex value associated with this condition
-                      ///< variable, used for implementing events
-
-    bool ShouldWait(Thread* thread) const override;
-    void Acquire(Thread* thread) override;
-
-    /**
-     * Releases a slot from a condition variable.
-     * @param target The number of threads to wakeup, -1 is all.
-     * @return ResultCode indicating if the operation succeeded.
-     */
-    ResultCode Release(s32 target);
-
-private:
-    ConditionVariable();
-    ~ConditionVariable() override;
-};
-
-} // namespace Kernel

src/core/hle/kernel/errors.h

@@ -20,6 +20,7 @@ enum {
     MaxConnectionsReached = 52,
 
     // Confirmed Switch OS error codes
+    MisalignedAddress = 102,
     InvalidHandle = 114,
     Timeout = 117,
     SynchronizationCanceled = 118,

src/core/hle/kernel/hle_ipc.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cinttypes>
 #include <boost/range/algorithm_ext/erase.hpp>
 #include "common/assert.h"
 #include "common/common_funcs.h"

src/core/hle/kernel/kernel.h

@@ -18,12 +18,10 @@ using Handle = u32;
 enum class HandleType : u32 {
     Unknown,
     Event,
-    Mutex,
     SharedMemory,
     Thread,
     Process,
     AddressArbiter,
-    ConditionVariable,
     Timer,
     ResourceLimit,
     CodeSet,
@@ -63,9 +61,7 @@ public:
     bool IsWaitable() const {
         switch (GetHandleType()) {
         case HandleType::Event:
-        case HandleType::Mutex:
         case HandleType::Thread:
-        case HandleType::ConditionVariable:
         case HandleType::Timer:
         case HandleType::ServerPort:
         case HandleType::ServerSession:

src/core/hle/kernel/mutex.cpp

@@ -7,6 +7,7 @@
 #include <boost/range/algorithm_ext/erase.hpp>
 #include "common/assert.h"
 #include "core/core.h"
+#include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/mutex.h"
@@ -15,124 +16,120 @@
 
 namespace Kernel {
 
-void ReleaseThreadMutexes(Thread* thread) {
-    for (auto& mtx : thread->held_mutexes) {
-        mtx->SetHasWaiters(false);
-        mtx->SetHoldingThread(nullptr);
-        mtx->WakeupAllWaitingThreads();
-    }
-    thread->held_mutexes.clear();
-}
+/// Returns the number of threads that are waiting for a mutex, and the highest priority one among
+/// those.
+static std::pair<SharedPtr<Thread>, u32> GetHighestPriorityMutexWaitingThread(
+    SharedPtr<Thread> current_thread, VAddr mutex_addr) {
 
-Mutex::Mutex() {}
-Mutex::~Mutex() {}
+    SharedPtr<Thread> highest_priority_thread;
+    u32 num_waiters = 0;
 
-SharedPtr<Mutex> Mutex::Create(SharedPtr<Kernel::Thread> holding_thread, VAddr guest_addr,
-                               std::string name) {
-    SharedPtr<Mutex> mutex(new Mutex);
+    for (auto& thread : current_thread->wait_mutex_threads) {
+        if (thread->mutex_wait_address != mutex_addr)
+            continue;
 
-    mutex->guest_addr = guest_addr;
-    mutex->name = std::move(name);
+        ASSERT(thread->status == THREADSTATUS_WAIT_MUTEX);
 
-    // If mutex was initialized with a holding thread, acquire it by the holding thread
-    if (holding_thread) {
-        mutex->Acquire(holding_thread.get());
+        ++num_waiters;
+        if (highest_priority_thread == nullptr ||
+            thread->GetPriority() < highest_priority_thread->GetPriority()) {
+            highest_priority_thread = thread;
+        }
     }
 
-    // Mutexes are referenced by guest address, so track this in the kernel
-    g_object_address_table.Insert(guest_addr, mutex);
-
-    return mutex;
+    return {highest_priority_thread, num_waiters};
 }
 
-bool Mutex::ShouldWait(Thread* thread) const {
-    auto holding_thread = GetHoldingThread();
-    return holding_thread != nullptr && thread != holding_thread;
+/// Update the mutex owner field of all threads waiting on the mutex to point to the new owner.
+static void TransferMutexOwnership(VAddr mutex_addr, SharedPtr<Thread> current_thread,
+                                   SharedPtr<Thread> new_owner) {
+    auto threads = current_thread->wait_mutex_threads;
+    for (auto& thread : threads) {
+        if (thread->mutex_wait_address != mutex_addr)
+            continue;
+
+        ASSERT(thread->lock_owner == current_thread);
+        current_thread->RemoveMutexWaiter(thread);
+        if (new_owner != thread)
+            new_owner->AddMutexWaiter(thread);
+    }
 }
 
-void Mutex::Acquire(Thread* thread) {
-    ASSERT_MSG(!ShouldWait(thread), "object unavailable!");
+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::MisalignedAddress);
+    }
 
-    priority = thread->current_priority;
-    thread->held_mutexes.insert(this);
-    SetHoldingThread(thread);
-    thread->UpdatePriority();
-    Core::System::GetInstance().PrepareReschedule();
-}
+    SharedPtr<Thread> holding_thread = g_handle_table.Get<Thread>(holding_thread_handle);
+    SharedPtr<Thread> requesting_thread = g_handle_table.Get<Thread>(requesting_thread_handle);
 
-ResultCode Mutex::Release(Thread* thread) {
-    auto holding_thread = GetHoldingThread();
-    ASSERT(holding_thread);
+    // TODO(Subv): It is currently unknown if it is possible to lock a mutex in behalf of another
+    // thread.
+    ASSERT(requesting_thread == GetCurrentThread());
 
-    // We can only release the mutex if it's held by the calling thread.
-    ASSERT(thread == holding_thread);
+    u32 addr_value = Memory::Read32(address);
+
+    // If the mutex isn't being held, just return success.
+    if (addr_value != (holding_thread_handle | Mutex::MutexHasWaitersFlag)) {
+        return RESULT_SUCCESS;
+    }
+
+    if (holding_thread == nullptr)
+        return ERR_INVALID_HANDLE;
+
+    // Wait until the mutex is released
+    GetCurrentThread()->mutex_wait_address = address;
+    GetCurrentThread()->wait_handle = requesting_thread_handle;
+
+    GetCurrentThread()->status = THREADSTATUS_WAIT_MUTEX;
+    GetCurrentThread()->wakeup_callback = nullptr;
+
+    // Update the lock holder thread's priority to prevent priority inversion.
+    holding_thread->AddMutexWaiter(GetCurrentThread());
 
-    holding_thread->held_mutexes.erase(this);
-    holding_thread->UpdatePriority();
-    SetHoldingThread(nullptr);
-    SetHasWaiters(!GetWaitingThreads().empty());
-    WakeupAllWaitingThreads();
     Core::System::GetInstance().PrepareReschedule();
 
     return RESULT_SUCCESS;
 }
 
-void Mutex::AddWaitingThread(SharedPtr<Thread> thread) {
-    WaitObject::AddWaitingThread(thread);
-    thread->pending_mutexes.insert(this);
-    SetHasWaiters(true);
-    UpdatePriority();
-}
-
-void Mutex::RemoveWaitingThread(Thread* thread) {
-    WaitObject::RemoveWaitingThread(thread);
-    thread->pending_mutexes.erase(this);
-    if (!GetHasWaiters())
-        SetHasWaiters(!GetWaitingThreads().empty());
-    UpdatePriority();
-}
-
-void Mutex::UpdatePriority() {
-    if (!GetHoldingThread())
-        return;
-
-    u32 best_priority = THREADPRIO_LOWEST;
-    for (auto& waiter : GetWaitingThreads()) {
-        if (waiter->current_priority < best_priority)
-            best_priority = waiter->current_priority;
+ResultCode Mutex::Release(VAddr address) {
+    // The mutex address must be 4-byte aligned
+    if ((address % sizeof(u32)) != 0) {
+        return ResultCode(ErrorModule::Kernel, ErrCodes::MisalignedAddress);
     }
 
-    if (best_priority != priority) {
-        priority = best_priority;
-        GetHoldingThread()->UpdatePriority();
+    auto [thread, num_waiters] = GetHighestPriorityMutexWaitingThread(GetCurrentThread(), 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;
     }
-}
 
-Handle Mutex::GetOwnerHandle() const {
-    GuestState guest_state{Memory::Read32(guest_addr)};
-    return guest_state.holding_thread_handle;
-}
+    // Transfer the ownership of the mutex from the previous owner to the new one.
+    TransferMutexOwnership(address, GetCurrentThread(), thread);
 
-SharedPtr<Thread> Mutex::GetHoldingThread() const {
-    GuestState guest_state{Memory::Read32(guest_addr)};
-    return g_handle_table.Get<Thread>(guest_state.holding_thread_handle);
-}
+    u32 mutex_value = thread->wait_handle;
 
-void Mutex::SetHoldingThread(SharedPtr<Thread> thread) {
-    GuestState guest_state{Memory::Read32(guest_addr)};
-    guest_state.holding_thread_handle.Assign(thread ? thread->guest_handle : 0);
-    Memory::Write32(guest_addr, guest_state.raw);
-}
+    if (num_waiters >= 2) {
+        // Notify the guest that there are still some threads waiting for the mutex
+        mutex_value |= Mutex::MutexHasWaitersFlag;
+    }
 
-bool Mutex::GetHasWaiters() const {
-    GuestState guest_state{Memory::Read32(guest_addr)};
-    return guest_state.has_waiters != 0;
-}
+    // Grant the mutex to the next waiting thread and resume it.
+    Memory::Write32(address, mutex_value);
 
-void Mutex::SetHasWaiters(bool has_waiters) {
-    GuestState guest_state{Memory::Read32(guest_addr)};
-    guest_state.has_waiters.Assign(has_waiters ? 1 : 0);
-    Memory::Write32(guest_addr, guest_state.raw);
-}
+    ASSERT(thread->status == THREADSTATUS_WAIT_MUTEX);
+    thread->ResumeFromWait();
 
+    thread->lock_owner = nullptr;
+    thread->condvar_wait_address = 0;
+    thread->mutex_wait_address = 0;
+    thread->wait_handle = 0;
+
+    return RESULT_SUCCESS;
+}
 } // namespace Kernel

src/core/hle/kernel/mutex.h

@@ -15,87 +15,23 @@ namespace Kernel {
 
 class Thread;
 
-class Mutex final : public WaitObject {
+class Mutex final {
 public:
-    /**
-     * Creates a mutex.
-     * @param holding_thread Specifies a thread already holding the mutex. If not nullptr, this
-     * thread will acquire the mutex.
-     * @param guest_addr Address of the object tracking the mutex in guest memory. If specified,
-     * this mutex will update the guest object when its state changes.
-     * @param name Optional name of mutex
-     * @return Pointer to new Mutex object
-     */
-    static SharedPtr<Mutex> Create(SharedPtr<Kernel::Thread> holding_thread, VAddr guest_addr = 0,
-                                   std::string name = "Unknown");
+    /// Flag that indicates that a mutex still has threads waiting for it.
+    static constexpr u32 MutexHasWaitersFlag = 0x40000000;
+    /// Mask of the bits in a mutex address value that contain the mutex owner.
+    static constexpr u32 MutexOwnerMask = 0xBFFFFFFF;
 
-    std::string GetTypeName() const override {
-        return "Mutex";
-    }
-    std::string GetName() const override {
-        return name;
-    }
+    /// Attempts to acquire a mutex at the specified address.
+    static ResultCode TryAcquire(VAddr address, Handle holding_thread_handle,
+                                 Handle requesting_thread_handle);
 
-    static const HandleType HANDLE_TYPE = HandleType::Mutex;
-    HandleType GetHandleType() const override {
-        return HANDLE_TYPE;
-    }
-
-    u32 priority;     ///< The priority of the mutex, used for priority inheritance.
-    std::string name; ///< Name of mutex (optional)
-    VAddr guest_addr; ///< Address of the guest mutex value
-
-    /**
-     * Elevate the mutex priority to the best priority
-     * among the priorities of all its waiting threads.
-     */
-    void UpdatePriority();
-
-    bool ShouldWait(Thread* thread) const override;
-    void Acquire(Thread* thread) override;
-
-    void AddWaitingThread(SharedPtr<Thread> thread) override;
-    void RemoveWaitingThread(Thread* thread) override;
-
-    /**
-     * Attempts to release the mutex from the specified thread.
-     * @param thread Thread that wants to release the mutex.
-     * @returns The result code of the operation.
-     */
-    ResultCode Release(Thread* thread);
-
-    /// Gets the handle to the holding process stored in the guest state.
-    Handle GetOwnerHandle() const;
-
-    /// Gets the Thread pointed to by the owner handle
-    SharedPtr<Thread> GetHoldingThread() const;
-    /// Sets the holding process handle in the guest state.
-    void SetHoldingThread(SharedPtr<Thread> thread);
-
-    /// Returns the has_waiters bit in the guest state.
-    bool GetHasWaiters() const;
-    /// Sets the has_waiters bit in the guest state.
-    void SetHasWaiters(bool has_waiters);
+    /// Releases the mutex at the specified address.
+    static ResultCode Release(VAddr address);
 
 private:
-    Mutex();
-    ~Mutex() override;
-
-    /// Object in guest memory used to track the mutex state
-    union GuestState {
-        u32_le raw;
-        /// Handle of the thread that currently holds the mutex, 0 if available
-        BitField<0, 30, u32_le> holding_thread_handle;
-        /// 1 when there are threads waiting for this mutex, otherwise 0
-        BitField<30, 1, u32_le> has_waiters;
-    };
-    static_assert(sizeof(GuestState) == 4, "GuestState size is incorrect");
+    Mutex() = default;
+    ~Mutex() = default;
 };
 
-/**
- * Releases all the mutexes held by the specified thread
- * @param thread Thread that is holding the mutexes
- */
-void ReleaseThreadMutexes(Thread* thread);
-
 } // namespace Kernel

src/core/hle/kernel/object_address_table.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cinttypes>
 #include "common/assert.h"
 #include "core/hle/kernel/object_address_table.h"
 
@@ -10,12 +11,14 @@ namespace Kernel {
 ObjectAddressTable g_object_address_table;
 
 void ObjectAddressTable::Insert(VAddr addr, SharedPtr<Object> obj) {
-    ASSERT_MSG(objects.find(addr) == objects.end(), "Object already exists with addr=0x%lx", addr);
+    ASSERT_MSG(objects.find(addr) == objects.end(), "Object already exists with addr=0x%" PRIx64,
+               addr);
     objects[addr] = obj;
 }
 
 void ObjectAddressTable::Close(VAddr addr) {
-    ASSERT_MSG(objects.find(addr) != objects.end(), "Object does not exist with addr=0x%lx", addr);
+    ASSERT_MSG(objects.find(addr) != objects.end(), "Object does not exist with addr=0x%" PRIx64,
+               addr);
     objects.erase(addr);
 }
 

src/core/hle/kernel/shared_memory.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cinttypes>
 #include <cstring>
 #include "common/logging/log.h"
 #include "core/core.h"
@@ -107,7 +108,8 @@ ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermi
 
     // Error out if the requested permissions don't match what the creator process allows.
     if (static_cast<u32>(permissions) & ~static_cast<u32>(own_other_permissions)) {
-        LOG_ERROR(Kernel, "cannot map id=%u, address=0x%lx name=%s, permissions don't match",
+        LOG_ERROR(Kernel,
+                  "cannot map id=%u, address=0x%" PRIx64 " name=%s, permissions don't match",
                   GetObjectId(), address, name.c_str());
         return ERR_INVALID_COMBINATION;
     }
@@ -115,7 +117,8 @@ ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermi
     // Error out if the provided permissions are not compatible with what the creator process needs.
     if (other_permissions != MemoryPermission::DontCare &&
         static_cast<u32>(this->permissions) & ~static_cast<u32>(other_permissions)) {
-        LOG_ERROR(Kernel, "cannot map id=%u, address=0x%lx name=%s, permissions don't match",
+        LOG_ERROR(Kernel,
+                  "cannot map id=%u, address=0x%" PRIx64 " name=%s, permissions don't match",
                   GetObjectId(), address, name.c_str());
         return ERR_WRONG_PERMISSION;
     }
@@ -132,7 +135,8 @@ ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermi
         target_address, backing_block, backing_block_offset, size, MemoryState::Shared);
     if (result.Failed()) {
         LOG_ERROR(Kernel,
-                  "cannot map id=%u, target_address=0x%lx name=%s, error mapping to virtual memory",
+                  "cannot map id=%u, target_address=0x%" PRIx64
+                  " name=%s, error mapping to virtual memory",
                   GetObjectId(), target_address, name.c_str());
         return result.Code();
     }

src/core/hle/kernel/svc.cpp

@@ -13,7 +13,6 @@
 #include "core/core_timing.h"
 #include "core/hle/kernel/client_port.h"
 #include "core/hle/kernel/client_session.h"
-#include "core/hle/kernel/condition_variable.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/mutex.h"
@@ -40,7 +39,7 @@ static ResultCode SetHeapSize(VAddr* heap_addr, u64 heap_size) {
 }
 
 static ResultCode SetMemoryAttribute(VAddr addr, u64 size, u32 state0, u32 state1) {
-    LOG_WARNING(Kernel_SVC, "(STUBBED) called, addr=0x%lx", addr);
+    LOG_WARNING(Kernel_SVC, "(STUBBED) called, addr=0x%" PRIx64, addr);
     return RESULT_SUCCESS;
 }
 
@@ -262,32 +261,14 @@ static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr,
               "requesting_current_thread_handle=0x%08X",
               holding_thread_handle, mutex_addr, requesting_thread_handle);
 
-    SharedPtr<Thread> holding_thread = g_handle_table.Get<Thread>(holding_thread_handle);
-    SharedPtr<Thread> requesting_thread = g_handle_table.Get<Thread>(requesting_thread_handle);
-
-    ASSERT(requesting_thread);
-    ASSERT(requesting_thread == GetCurrentThread());
-
-    SharedPtr<Mutex> mutex = g_object_address_table.Get<Mutex>(mutex_addr);
-    if (!mutex) {
-        // Create a new mutex for the specified address if one does not already exist
-        mutex = Mutex::Create(holding_thread, mutex_addr);
-        mutex->name = Common::StringFromFormat("mutex-%llx", mutex_addr);
-    }
-
-    ASSERT(holding_thread == mutex->GetHoldingThread());
-
-    return WaitSynchronization1(mutex, requesting_thread.get());
+    return Mutex::TryAcquire(mutex_addr, holding_thread_handle, requesting_thread_handle);
 }
 
 /// Unlock a mutex
 static ResultCode ArbitrateUnlock(VAddr mutex_addr) {
     LOG_TRACE(Kernel_SVC, "called mutex_addr=0x%llx", mutex_addr);
 
-    SharedPtr<Mutex> mutex = g_object_address_table.Get<Mutex>(mutex_addr);
-    ASSERT(mutex);
-
-    return mutex->Release(GetCurrentThread());
+    return Mutex::Release(mutex_addr);
 }
 
 /// Break program execution
@@ -412,11 +393,6 @@ static ResultCode SetThreadPriority(Handle handle, u32 priority) {
     }
 
     thread->SetPriority(priority);
-    thread->UpdatePriority();
-
-    // Update the mutexes that this thread is waiting for
-    for (auto& mutex : thread->pending_mutexes)
-        mutex->UpdatePriority();
 
     Core::System::GetInstance().PrepareReschedule();
     return RESULT_SUCCESS;
@@ -634,77 +610,20 @@ static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_var
     SharedPtr<Thread> thread = g_handle_table.Get<Thread>(thread_handle);
     ASSERT(thread);
 
-    SharedPtr<Mutex> mutex = g_object_address_table.Get<Mutex>(mutex_addr);
-    if (!mutex) {
-        // Create a new mutex for the specified address if one does not already exist
-        mutex = Mutex::Create(thread, mutex_addr);
-        mutex->name = Common::StringFromFormat("mutex-%llx", mutex_addr);
-    }
+    CASCADE_CODE(Mutex::Release(mutex_addr));
 
-    SharedPtr<ConditionVariable> condition_variable =
-        g_object_address_table.Get<ConditionVariable>(condition_variable_addr);
-    if (!condition_variable) {
-        // Create a new condition_variable for the specified address if one does not already exist
-        condition_variable = ConditionVariable::Create(condition_variable_addr).Unwrap();
-        condition_variable->name =
-            Common::StringFromFormat("condition-variable-%llx", condition_variable_addr);
-    }
+    SharedPtr<Thread> current_thread = GetCurrentThread();
+    current_thread->condvar_wait_address = condition_variable_addr;
+    current_thread->mutex_wait_address = mutex_addr;
+    current_thread->wait_handle = thread_handle;
+    current_thread->status = THREADSTATUS_WAIT_MUTEX;
+    current_thread->wakeup_callback = nullptr;
 
-    if (condition_variable->mutex_addr) {
-        // Previously created the ConditionVariable using WaitProcessWideKeyAtomic, verify
-        // everything is correct
-        ASSERT(condition_variable->mutex_addr == mutex_addr);
-    } else {
-        // Previously created the ConditionVariable using SignalProcessWideKey, set the mutex
-        // associated with it
-        condition_variable->mutex_addr = mutex_addr;
-    }
+    current_thread->WakeAfterDelay(nano_seconds);
 
-    if (mutex->GetOwnerHandle()) {
-        // Release the mutex if the current thread is holding it
-        mutex->Release(thread.get());
-    }
-
-    auto wakeup_callback = [mutex, nano_seconds](ThreadWakeupReason reason,
-                                                 SharedPtr<Thread> thread,
-                                                 SharedPtr<WaitObject> object, size_t index) {
-        ASSERT(thread->status == THREADSTATUS_WAIT_SYNCH_ANY);
-
-        if (reason == ThreadWakeupReason::Timeout) {
-            thread->SetWaitSynchronizationResult(RESULT_TIMEOUT);
-            return true;
-        }
-
-        ASSERT(reason == ThreadWakeupReason::Signal);
-
-        // Now try to acquire the mutex and don't resume if it's not available.
-        if (!mutex->ShouldWait(thread.get())) {
-            mutex->Acquire(thread.get());
-            thread->SetWaitSynchronizationResult(RESULT_SUCCESS);
-            return true;
-        }
-
-        if (nano_seconds == 0) {
-            thread->SetWaitSynchronizationResult(RESULT_TIMEOUT);
-            return true;
-        }
-
-        thread->wait_objects = {mutex};
-        mutex->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 = DefaultThreadWakeupCallback;
-
-        Core::System::GetInstance().PrepareReschedule();
-
-        return false;
-    };
-    CASCADE_CODE(
-        WaitSynchronization1(condition_variable, thread.get(), nano_seconds, wakeup_callback));
+    // Note: Deliberately don't attempt to inherit the lock owner's priority.
 
+    Core::System::GetInstance().PrepareReschedule();
     return RESULT_SUCCESS;
 }
 
@@ -713,24 +632,53 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target
     LOG_TRACE(Kernel_SVC, "called, condition_variable_addr=0x%llx, target=0x%08x",
               condition_variable_addr, target);
 
-    // Wakeup all or one thread - Any other value is unimplemented
-    ASSERT(target == -1 || target == 1);
+    u32 processed = 0;
+    auto& thread_list = Core::System::GetInstance().Scheduler().GetThreadList();
 
-    SharedPtr<ConditionVariable> condition_variable =
-        g_object_address_table.Get<ConditionVariable>(condition_variable_addr);
-    if (!condition_variable) {
-        // Create a new condition_variable for the specified address if one does not already exist
-        condition_variable = ConditionVariable::Create(condition_variable_addr).Unwrap();
-        condition_variable->name =
-            Common::StringFromFormat("condition-variable-%llx", condition_variable_addr);
-    }
+    for (auto& thread : thread_list) {
+        if (thread->condvar_wait_address != condition_variable_addr)
+            continue;
 
-    CASCADE_CODE(condition_variable->Release(target));
+        // Only process up to 'target' threads, unless 'target' is -1, in which case process
+        // them all.
+        if (target != -1 && processed >= target)
+            break;
 
-    if (condition_variable->mutex_addr) {
-        // If a mutex was created for this condition_variable, wait the current thread on it
-        SharedPtr<Mutex> mutex = g_object_address_table.Get<Mutex>(condition_variable->mutex_addr);
-        return WaitSynchronization1(mutex, GetCurrentThread());
+        // If the mutex is not yet acquired, acquire it.
+        u32 mutex_val = Memory::Read32(thread->mutex_wait_address);
+
+        if (mutex_val == 0) {
+            // We were able to acquire the mutex, resume this thread.
+            Memory::Write32(thread->mutex_wait_address, thread->wait_handle);
+            ASSERT(thread->status == THREADSTATUS_WAIT_MUTEX);
+            thread->ResumeFromWait();
+
+            auto lock_owner = thread->lock_owner;
+            if (lock_owner)
+                lock_owner->RemoveMutexWaiter(thread);
+
+            thread->lock_owner = nullptr;
+            thread->mutex_wait_address = 0;
+            thread->condvar_wait_address = 0;
+            thread->wait_handle = 0;
+        } else {
+            // Couldn't acquire the mutex, block the thread.
+            Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask);
+            auto owner = g_handle_table.Get<Thread>(owner_handle);
+            ASSERT(owner);
+            ASSERT(thread->status != THREADSTATUS_RUNNING);
+            thread->status = THREADSTATUS_WAIT_MUTEX;
+            thread->wakeup_callback = nullptr;
+
+            // Signal that the mutex now has a waiting thread.
+            Memory::Write32(thread->mutex_wait_address, mutex_val | Mutex::MutexHasWaitersFlag);
+
+            owner->AddMutexWaiter(thread);
+
+            Core::System::GetInstance().PrepareReschedule();
+        }
+
+        ++processed;
     }
 
     return RESULT_SUCCESS;
@@ -763,8 +711,8 @@ static ResultCode ResetSignal(Handle handle) {
 
 /// Creates a TransferMemory object
 static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32 permissions) {
-    LOG_WARNING(Kernel_SVC, "(STUBBED) called addr=0x%lx, size=0x%lx, perms=%08X", addr, size,
-                permissions);
+    LOG_WARNING(Kernel_SVC, "(STUBBED) called addr=0x%" PRIx64 ", size=0x%" PRIx64 ", perms=%08X",
+                addr, size, permissions);
     *handle = 0;
     return RESULT_SUCCESS;
 }

src/core/hle/kernel/thread.cpp

@@ -77,9 +77,6 @@ void Thread::Stop() {
     }
     wait_objects.clear();
 
-    // Release all the mutexes that this thread holds
-    ReleaseThreadMutexes(this);
-
     // Mark the TLS slot in the thread's page as free.
     u64 tls_page = (tls_address - Memory::TLS_AREA_VADDR) / Memory::PAGE_SIZE;
     u64 tls_slot =
@@ -126,6 +123,19 @@ static void ThreadWakeupCallback(u64 thread_handle, int cycles_late) {
             resume = thread->wakeup_callback(ThreadWakeupReason::Timeout, thread, nullptr, 0);
     }
 
+    if (thread->mutex_wait_address != 0 || thread->condvar_wait_address != 0 ||
+        thread->wait_handle) {
+        ASSERT(thread->status == THREADSTATUS_WAIT_MUTEX);
+        thread->mutex_wait_address = 0;
+        thread->condvar_wait_address = 0;
+        thread->wait_handle = 0;
+
+        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);
+    }
+
     if (resume)
         thread->ResumeFromWait();
 }
@@ -151,6 +161,7 @@ void Thread::ResumeFromWait() {
     case THREADSTATUS_WAIT_HLE_EVENT:
     case THREADSTATUS_WAIT_SLEEP:
     case THREADSTATUS_WAIT_IPC:
+    case THREADSTATUS_WAIT_MUTEX:
         break;
 
     case THREADSTATUS_READY:
@@ -256,7 +267,9 @@ ResultVal<SharedPtr<Thread>> Thread::Create(std::string name, VAddr entry_point,
     thread->last_running_ticks = CoreTiming::GetTicks();
     thread->processor_id = processor_id;
     thread->wait_objects.clear();
-    thread->wait_address = 0;
+    thread->mutex_wait_address = 0;
+    thread->condvar_wait_address = 0;
+    thread->wait_handle = 0;
     thread->name = std::move(name);
     thread->callback_handle = wakeup_callback_handle_table.Create(thread).Unwrap();
     thread->owner_process = owner_process;
@@ -317,17 +330,8 @@ ResultVal<SharedPtr<Thread>> Thread::Create(std::string name, VAddr entry_point,
 void Thread::SetPriority(u32 priority) {
     ASSERT_MSG(priority <= THREADPRIO_LOWEST && priority >= THREADPRIO_HIGHEST,
                "Invalid priority value.");
-    Core::System::GetInstance().Scheduler().SetThreadPriority(this, priority);
-    nominal_priority = current_priority = priority;
-}
-
-void Thread::UpdatePriority() {
-    u32 best_priority = nominal_priority;
-    for (auto& mutex : held_mutexes) {
-        if (mutex->priority < best_priority)
-            best_priority = mutex->priority;
-    }
-    BoostPriority(best_priority);
+    nominal_priority = priority;
+    UpdatePriority();
 }
 
 void Thread::BoostPriority(u32 priority) {
@@ -377,6 +381,38 @@ VAddr Thread::GetCommandBufferAddress() const {
     return GetTLSAddress() + CommandHeaderOffset;
 }
 
+void Thread::AddMutexWaiter(SharedPtr<Thread> thread) {
+    thread->lock_owner = this;
+    wait_mutex_threads.emplace_back(std::move(thread));
+    UpdatePriority();
+}
+
+void Thread::RemoveMutexWaiter(SharedPtr<Thread> thread) {
+    boost::remove_erase(wait_mutex_threads, thread);
+    thread->lock_owner = nullptr;
+    UpdatePriority();
+}
+
+void Thread::UpdatePriority() {
+    // Find the highest priority among all the threads that are waiting for this thread's lock
+    u32 new_priority = nominal_priority;
+    for (const auto& thread : wait_mutex_threads) {
+        if (thread->nominal_priority < new_priority)
+            new_priority = thread->nominal_priority;
+    }
+
+    if (new_priority == current_priority)
+        return;
+
+    Core::System::GetInstance().Scheduler().SetThreadPriority(this, new_priority);
+
+    current_priority = new_priority;
+
+    // Recursively update the priority of the thread that depends on the priority of this one.
+    if (lock_owner)
+        lock_owner->UpdatePriority();
+}
+
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 
 /**

src/core/hle/kernel/thread.h

@@ -18,7 +18,7 @@
 enum ThreadPriority : u32 {
     THREADPRIO_HIGHEST = 0,       ///< Highest thread priority
     THREADPRIO_USERLAND_MAX = 24, ///< Highest thread priority for userland apps
-    THREADPRIO_DEFAULT = 48,      ///< Default thread priority for userland apps
+    THREADPRIO_DEFAULT = 44,      ///< Default thread priority for userland apps
     THREADPRIO_LOWEST = 63,       ///< Lowest thread priority
 };
 
@@ -43,6 +43,7 @@ enum ThreadStatus {
     THREADSTATUS_WAIT_IPC,       ///< Waiting for the reply from an IPC request
     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_DORMANT,        ///< Created but not yet made ready
     THREADSTATUS_DEAD            ///< Run to completion, or forcefully terminated
 };
@@ -54,7 +55,6 @@ enum class ThreadWakeupReason {
 
 namespace Kernel {
 
-class Mutex;
 class Process;
 
 class Thread final : public WaitObject {
@@ -103,18 +103,21 @@ public:
      */
     void SetPriority(u32 priority);
 
-    /**
-     * Boost's a thread's priority to the best priority among the thread's held mutexes.
-     * This prevents priority inversion via priority inheritance.
-     */
-    void UpdatePriority();
-
     /**
      * Temporarily boosts the thread's priority until the next time it is scheduled
      * @param priority The new priority
      */
     void BoostPriority(u32 priority);
 
+    /// Adds a thread to the list of threads that are waiting for a lock held by this thread.
+    void AddMutexWaiter(SharedPtr<Thread> thread);
+
+    /// Removes a thread from the list of threads that are waiting for a lock held by this thread.
+    void RemoveMutexWaiter(SharedPtr<Thread> thread);
+
+    /// Recalculates the current priority taking into account priority inheritance.
+    void UpdatePriority();
+
     /**
      * Gets the thread's thread ID
      * @return The thread's ID
@@ -205,19 +208,22 @@ public:
 
     VAddr tls_address; ///< Virtual address of the Thread Local Storage of the thread
 
-    /// Mutexes currently held by this thread, which will be released when it exits.
-    boost::container::flat_set<SharedPtr<Mutex>> held_mutexes;
-
-    /// Mutexes that this thread is currently waiting for.
-    boost::container::flat_set<SharedPtr<Mutex>> pending_mutexes;
-
     SharedPtr<Process> owner_process; ///< Process that owns this thread
 
     /// Objects that the thread is waiting on, in the same order as they were
     // passed to WaitSynchronization1/N.
     std::vector<SharedPtr<WaitObject>> wait_objects;
 
-    VAddr wait_address; ///< If waiting on an AddressArbiter, this is the arbitration address
+    /// List of threads that are waiting for a mutex that is held by this thread.
+    std::vector<SharedPtr<Thread>> wait_mutex_threads;
+
+    /// Thread that owns the lock that this thread is waiting for.
+    SharedPtr<Thread> lock_owner;
+
+    // If waiting on a ConditionVariable, this is the ConditionVariable  address
+    VAddr condvar_wait_address;
+    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.
 
     std::string name;
 

src/core/hle/service/filesystem/fsp_srv.cpp

@@ -197,7 +197,7 @@ private:
         IPC::RequestParser rp{ctx};
         const u64 unk = rp.Pop<u64>();
 
-        LOG_DEBUG(Service_FS, "called, unk=0x%llx", unk);
+        LOG_DEBUG(Service_FS, "called, unk=0x%" PRIx64, unk);
 
         // Calculate how many entries we can fit in the output buffer
         u64 count_entries = ctx.GetWriteBufferSize() / sizeof(FileSys::Entry);
@@ -239,7 +239,7 @@ public:
             {2, &IFileSystem::CreateDirectory, "CreateDirectory"},
             {3, nullptr, "DeleteDirectory"},
             {4, nullptr, "DeleteDirectoryRecursively"},
-            {5, nullptr, "RenameFile"},
+            {5, &IFileSystem::RenameFile, "RenameFile"},
             {6, nullptr, "RenameDirectory"},
             {7, &IFileSystem::GetEntryType, "GetEntryType"},
             {8, &IFileSystem::OpenFile, "OpenFile"},
@@ -300,6 +300,26 @@ public:
         rb.Push(backend->CreateDirectory(name));
     }
 
+    void RenameFile(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+
+        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);
+
+        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);
+
+        LOG_DEBUG(Service_FS, "called file '%s' to file '%s'", src_name.c_str(), dst_name.c_str());
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(backend->RenameFile(src_name, dst_name));
+    }
+
     void OpenFile(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
 

src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cinttypes>
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/core.h"
@@ -13,7 +14,7 @@
 namespace Service::Nvidia::Devices {
 
 u32 nvdisp_disp0::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
-    UNIMPLEMENTED();
+    UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
 }
 
@@ -21,8 +22,9 @@ void nvdisp_disp0::flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u3
                         u32 stride, NVFlinger::BufferQueue::BufferTransformFlags transform) {
     VAddr addr = nvmap_dev->GetObjectAddress(buffer_handle);
     LOG_WARNING(Service,
-                "Drawing from address %lx offset %08X Width %u Height %u Stride %u Format %u", addr,
-                offset, width, height, stride, format);
+                "Drawing from address %" PRIx64
+                " offset %08X Width %u Height %u Stride %u Format %u",
+                addr, offset, width, height, stride, format);
 
     using PixelFormat = Tegra::FramebufferConfig::PixelFormat;
     const Tegra::FramebufferConfig framebuffer{

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

@@ -27,6 +27,11 @@ u32 nvhost_as_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vecto
     case IoctlCommand::IocGetVaRegionsCommand:
         return GetVARegions(input, output);
     }
+
+    if (static_cast<IoctlCommand>(command.cmd.Value()) == IoctlCommand::IocRemapCommand)
+        return Remap(input, output);
+
+    UNIMPLEMENTED_MSG("Unimplemented ioctl command");
     return 0;
 }
 
@@ -56,6 +61,36 @@ u32 nvhost_as_gpu::AllocateSpace(const std::vector<u8>& input, std::vector<u8>&
     return 0;
 }
 
+u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) {
+    size_t num_entries = input.size() / sizeof(IoctlRemapEntry);
+
+    NGLOG_WARNING(Service_NVDRV, "(STUBBED) called, num_entries=0x{:X}", num_entries);
+
+    std::vector<IoctlRemapEntry> entries(num_entries);
+    std::memcpy(entries.data(), input.data(), input.size());
+
+    auto& gpu = Core::System::GetInstance().GPU();
+
+    for (const auto& entry : entries) {
+        NGLOG_WARNING(Service_NVDRV, "remap entry, offset=0x{:X} handle=0x{:X} pages=0x{:X}",
+                      entry.offset, entry.nvmap_handle, entry.pages);
+        Tegra::GPUVAddr offset = static_cast<Tegra::GPUVAddr>(entry.offset) << 0x10;
+
+        auto object = nvmap_dev->GetObject(entry.nvmap_handle);
+        ASSERT(object);
+
+        ASSERT(object->status == nvmap::Object::Status::Allocated);
+
+        u64 size = static_cast<u64>(entry.pages) << 0x10;
+        ASSERT(size <= object->size);
+
+        Tegra::GPUVAddr returned = gpu.memory_manager->MapBufferEx(object->addr, offset, size);
+        ASSERT(returned == offset);
+    }
+    std::memcpy(output.data(), entries.data(), output.size());
+    return 0;
+}
+
 u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) {
     IoctlMapBufferEx params{};
     std::memcpy(&params, input.data(), input.size());
@@ -73,6 +108,16 @@ u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& ou
     auto object = nvmap_dev->GetObject(params.nvmap_handle);
     ASSERT(object);
 
+    // We can only map objects that have already been assigned a CPU address.
+    ASSERT(object->status == nvmap::Object::Status::Allocated);
+
+    ASSERT(params.buffer_offset == 0);
+
+    // The real nvservices doesn't make a distinction between handles and ids, and
+    // object can only have one handle and it will be the same as its id. Assert that this is the
+    // case to prevent unexpected behavior.
+    ASSERT(object->id == params.nvmap_handle);
+
     auto& gpu = Core::System::GetInstance().GPU();
 
     if (params.flags & 1) {

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

@@ -26,6 +26,7 @@ private:
     enum class IoctlCommand : u32_le {
         IocInitalizeExCommand = 0x40284109,
         IocAllocateSpaceCommand = 0xC0184102,
+        IocRemapCommand = 0x00000014,
         IocMapBufferExCommand = 0xC0284106,
         IocBindChannelCommand = 0x40044101,
         IocGetVaRegionsCommand = 0xC0404108,
@@ -54,6 +55,16 @@ private:
     };
     static_assert(sizeof(IoctlAllocSpace) == 24, "IoctlInitalizeEx is incorrect size");
 
+    struct IoctlRemapEntry {
+        u16_le flags;
+        u16_le kind;
+        u32_le nvmap_handle;
+        INSERT_PADDING_WORDS(1);
+        u32_le offset;
+        u32_le pages;
+    };
+    static_assert(sizeof(IoctlRemapEntry) == 20, "IoctlRemapEntry is incorrect size");
+
     struct IoctlMapBufferEx {
         u32_le flags; // bit0: fixed_offset, bit2: cacheable
         u32_le kind;  // -1 is default
@@ -91,6 +102,7 @@ private:
 
     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 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

@@ -18,7 +18,7 @@ u32 nvhost_ctrl::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<
     case IoctlCommand::IocCtrlEventWaitCommand:
         return IocCtrlEventWait(input, output);
     }
-    UNIMPLEMENTED();
+    UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
 }
 

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

@@ -25,7 +25,7 @@ u32 nvhost_ctrl_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vec
     case IoctlCommand::IocZcullGetInfo:
         return ZCullGetInfo(input, output);
     }
-    UNIMPLEMENTED();
+    UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
 }
 

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

@@ -40,7 +40,7 @@ u32 nvhost_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u
         }
     }
 
-    UNIMPLEMENTED();
+    UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
 };
 

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

@@ -32,7 +32,7 @@ u32 nvmap::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& o
         return IocParam(input, output);
     }
 
-    UNIMPLEMENTED();
+    UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
 }
 

src/core/hle/service/nvflinger/buffer_queue.cpp

@@ -9,7 +9,8 @@
 #include "core/core_timing.h"
 #include "core/hle/service/nvflinger/buffer_queue.h"
 
-namespace Service::NVFlinger {
+namespace Service {
+namespace NVFlinger {
 
 BufferQueue::BufferQueue(u32 id, u64 layer_id) : id(id), layer_id(layer_id) {
     native_handle = Kernel::Event::Create(Kernel::ResetType::OneShot, "BufferQueue NativeHandle");
@@ -110,4 +111,5 @@ void BufferQueue::SetBufferWaitEvent(Kernel::SharedPtr<Kernel::Event>&& wait_eve
     buffer_wait_event = std::move(wait_event);
 }
 
-} // namespace Service::NVFlinger
+} // namespace NVFlinger
+} // namespace Service

src/core/hle/service/nvflinger/buffer_queue.h

@@ -13,7 +13,8 @@ namespace CoreTiming {
 struct EventType;
 }
 
-namespace Service::NVFlinger {
+namespace Service {
+namespace NVFlinger {
 
 struct IGBPBuffer {
     u32_le magic;
@@ -97,4 +98,5 @@ private:
     Kernel::SharedPtr<Kernel::Event> buffer_wait_event;
 };
 
-} // namespace Service::NVFlinger
+} // namespace NVFlinger
+} // namespace Service

src/core/loader/nro.cpp

@@ -137,7 +137,7 @@ ResultStatus AppLoader_NRO::Load(Kernel::SharedPtr<Kernel::Process>& process) {
     process->address_mappings = default_address_mappings;
     process->resource_limit =
         Kernel::ResourceLimit::GetForCategory(Kernel::ResourceLimitCategory::APPLICATION);
-    process->Run(base_addr, 48, Memory::DEFAULT_STACK_SIZE);
+    process->Run(base_addr, THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE);
 
     is_loaded = true;
     return ResultStatus::Success;

src/core/loader/nso.cpp

@@ -165,7 +165,7 @@ ResultStatus AppLoader_NSO::Load(Kernel::SharedPtr<Kernel::Process>& process) {
     process->address_mappings = default_address_mappings;
     process->resource_limit =
         Kernel::ResourceLimit::GetForCategory(Kernel::ResourceLimitCategory::APPLICATION);
-    process->Run(Memory::PROCESS_IMAGE_VADDR, 48, Memory::DEFAULT_STACK_SIZE);
+    process->Run(Memory::PROCESS_IMAGE_VADDR, THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE);
 
     is_loaded = true;
     return ResultStatus::Success;

src/video_core/engines/maxwell_3d.h

@@ -500,6 +500,11 @@ public:
                         return static_cast<GPUVAddr>((static_cast<GPUVAddr>(start_high) << 32) |
                                                      start_low);
                     }
+
+                    bool IsEnabled() const {
+                        return enable != 0 && StartAddress() != 0;
+                    }
+
                 } vertex_array[NumVertexArrays];
 
                 Blend blend;

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -127,7 +127,8 @@ RasterizerOpenGL::~RasterizerOpenGL() {
     }
 }
 
-void RasterizerOpenGL::SetupVertexArray(u8* array_ptr, GLintptr buffer_offset) {
+std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr,
+                                                             GLintptr buffer_offset) {
     MICROPROFILE_SCOPE(OpenGL_VAO);
     const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
     const auto& memory_manager = Core::System().GetInstance().GPU().memory_manager;
@@ -136,43 +137,59 @@ void RasterizerOpenGL::SetupVertexArray(u8* array_ptr, GLintptr buffer_offset) {
     state.draw.vertex_buffer = stream_buffer->GetHandle();
     state.Apply();
 
-    // TODO(bunnei): Add support for 1+ vertex arrays
-    const auto& vertex_array{regs.vertex_array[0]};
-    const auto& vertex_array_limit{regs.vertex_array_limit[0]};
-    ASSERT_MSG(vertex_array.enable, "vertex array 0 is disabled?");
-    ASSERT_MSG(!vertex_array.divisor, "vertex array 0 divisor is unimplemented!");
-    for (unsigned index = 1; index < Maxwell::NumVertexArrays; ++index) {
-        ASSERT_MSG(!regs.vertex_array[index].enable, "vertex array %d is unimplemented!", index);
+    // Upload all guest vertex arrays sequentially to our buffer
+    for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) {
+        const auto& vertex_array = regs.vertex_array[index];
+        if (!vertex_array.IsEnabled())
+            continue;
+
+        const Tegra::GPUVAddr start = vertex_array.StartAddress();
+        const Tegra::GPUVAddr end = regs.vertex_array_limit[index].LimitAddress();
+
+        ASSERT(end > start);
+        u64 size = end - start + 1;
+
+        // Copy vertex array data
+        const VAddr data_addr{memory_manager->PhysicalToVirtualAddress(start)};
+        res_cache.FlushRegion(data_addr, size, nullptr);
+        Memory::ReadBlock(data_addr, array_ptr, size);
+
+        // Bind the vertex array to the buffer at the current offset.
+        glBindVertexBuffer(index, stream_buffer->GetHandle(), buffer_offset, vertex_array.stride);
+
+        ASSERT_MSG(vertex_array.divisor == 0, "Vertex buffer divisor unimplemented");
+
+        array_ptr += size;
+        buffer_offset += size;
     }
 
     // Use the vertex array as-is, assumes that the data is formatted correctly for OpenGL.
     // Enables the first 16 vertex attributes always, as we don't know which ones are actually used
-    // until shader time. Note, Tegra technically supports 32, but we're cappinig this to 16 for now
+    // until shader time. Note, Tegra technically supports 32, but we're capping this to 16 for now
     // to avoid OpenGL errors.
+    // TODO(Subv): Analyze the shader to identify which attributes are actually used and don't
+    // assume every shader uses them all.
     for (unsigned index = 0; index < 16; ++index) {
         auto& attrib = regs.vertex_attrib_format[index];
         NGLOG_DEBUG(HW_GPU, "vertex attrib {}, count={}, size={}, type={}, offset={}, normalize={}",
                     index, attrib.ComponentCount(), attrib.SizeString(), attrib.TypeString(),
                     attrib.offset.Value(), attrib.IsNormalized());
 
-        glVertexAttribPointer(index, attrib.ComponentCount(), MaxwellToGL::VertexType(attrib),
-                              attrib.IsNormalized() ? GL_TRUE : GL_FALSE, vertex_array.stride,
-                              reinterpret_cast<GLvoid*>(buffer_offset + attrib.offset));
+        auto& buffer = regs.vertex_array[attrib.buffer];
+        ASSERT(buffer.IsEnabled());
+
         glEnableVertexAttribArray(index);
+        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;
     }
 
-    // Copy vertex array data
-    const u64 data_size{vertex_array_limit.LimitAddress() - vertex_array.StartAddress() + 1};
-    const VAddr data_addr{memory_manager->PhysicalToVirtualAddress(vertex_array.StartAddress())};
-    res_cache.FlushRegion(data_addr, data_size, nullptr);
-    Memory::ReadBlock(data_addr, array_ptr, data_size);
-
-    array_ptr += data_size;
-    buffer_offset += data_size;
+    return {array_ptr, buffer_offset};
 }
 
-void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset, size_t ptr_pos) {
+void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) {
     // Helper function for uploading uniform data
     const auto copy_buffer = [&](GLuint handle, GLintptr offset, GLsizeiptr size) {
         if (has_ARB_direct_state_access) {
@@ -190,8 +207,6 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset, size
     u32 current_constbuffer_bindpoint = 0;
 
     for (unsigned index = 1; index < Maxwell::MaxShaderProgram; ++index) {
-        ptr_pos += sizeof(GLShader::MaxwellUniformData);
-
         auto& shader_config = gpu.regs.shader_config[index];
         const Maxwell::ShaderProgram program{static_cast<Maxwell::ShaderProgram>(index)};
 
@@ -205,13 +220,16 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset, size
         }
 
         // Upload uniform data as one UBO per stage
-        const GLintptr ubo_offset = buffer_offset + static_cast<GLintptr>(ptr_pos);
+        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[ptr_pos]);
+            reinterpret_cast<GLShader::MaxwellUniformData*>(buffer_ptr);
         ub_ptr->SetFromRegs(gpu.state.shader_stages[stage]);
 
+        buffer_ptr += sizeof(GLShader::MaxwellUniformData);
+        buffer_offset += sizeof(GLShader::MaxwellUniformData);
+
         // Fetch program code from memory
         GLShader::ProgramCode program_code;
         const u64 gpu_address{gpu.regs.code_address.CodeAddress() + shader_config.offset};
@@ -252,6 +270,24 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset, size
     shader_program_manager->UseTrivialGeometryShader();
 }
 
+size_t RasterizerOpenGL::CalculateVertexArraysSize() const {
+    const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
+
+    size_t size = 0;
+    for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) {
+        if (!regs.vertex_array[index].IsEnabled())
+            continue;
+
+        const Tegra::GPUVAddr start = regs.vertex_array[index].StartAddress();
+        const Tegra::GPUVAddr end = regs.vertex_array_limit[index].LimitAddress();
+
+        ASSERT(end > start);
+        size += end - start + 1;
+    }
+
+    return size;
+}
+
 bool RasterizerOpenGL::AccelerateDrawBatch(bool is_indexed) {
     accelerate_draw = is_indexed ? AccelDraw::Indexed : AccelDraw::Arrays;
     DrawArrays();
@@ -329,44 +365,49 @@ void RasterizerOpenGL::DrawArrays() {
     const u64 index_buffer_size{regs.index_array.count * regs.index_array.FormatSizeInBytes()};
     const unsigned vertex_num{is_indexed ? regs.index_array.count : regs.vertex_buffer.count};
 
-    // TODO(bunnei): Add support for 1+ vertex arrays
-    vs_input_size = vertex_num * regs.vertex_array[0].stride;
-
     state.draw.vertex_buffer = stream_buffer->GetHandle();
     state.Apply();
 
-    size_t buffer_size = static_cast<size_t>(vs_input_size);
+    size_t buffer_size = CalculateVertexArraysSize();
+
     if (is_indexed) {
-        buffer_size = Common::AlignUp(buffer_size, 4) + index_buffer_size;
+        buffer_size = Common::AlignUp<size_t>(buffer_size, 4) + index_buffer_size;
     }
 
     // Uniform space for the 5 shader stages
-    buffer_size += sizeof(GLShader::MaxwellUniformData) * Maxwell::MaxShaderStage;
+    buffer_size = Common::AlignUp<size_t>(buffer_size, 4) +
+                  sizeof(GLShader::MaxwellUniformData) * Maxwell::MaxShaderStage;
 
-    size_t ptr_pos = 0;
     u8* buffer_ptr;
     GLintptr buffer_offset;
     std::tie(buffer_ptr, buffer_offset) =
         stream_buffer->Map(static_cast<GLsizeiptr>(buffer_size), 4);
 
-    SetupVertexArray(buffer_ptr, buffer_offset);
-    ptr_pos += vs_input_size;
+    u8* offseted_buffer;
+    std::tie(offseted_buffer, buffer_offset) = SetupVertexArrays(buffer_ptr, buffer_offset);
+
+    offseted_buffer =
+        reinterpret_cast<u8*>(Common::AlignUp(reinterpret_cast<size_t>(offseted_buffer), 4));
+    buffer_offset = Common::AlignUp<size_t>(buffer_offset, 4);
 
     // If indexed mode, copy the index buffer
     GLintptr index_buffer_offset = 0;
     if (is_indexed) {
-        ptr_pos = Common::AlignUp(ptr_pos, 4);
-
         const auto& memory_manager = Core::System().GetInstance().GPU().memory_manager;
         const VAddr index_data_addr{
             memory_manager->PhysicalToVirtualAddress(regs.index_array.StartAddress())};
-        Memory::ReadBlock(index_data_addr, &buffer_ptr[ptr_pos], index_buffer_size);
+        Memory::ReadBlock(index_data_addr, offseted_buffer, index_buffer_size);
 
-        index_buffer_offset = buffer_offset + static_cast<GLintptr>(ptr_pos);
-        ptr_pos += index_buffer_size;
+        index_buffer_offset = buffer_offset;
+        offseted_buffer += index_buffer_size;
+        buffer_offset += index_buffer_size;
     }
 
-    SetupShaders(buffer_ptr, buffer_offset, ptr_pos);
+    offseted_buffer =
+        reinterpret_cast<u8*>(Common::AlignUp(reinterpret_cast<size_t>(offseted_buffer), 4));
+    buffer_offset = Common::AlignUp<size_t>(buffer_offset, 4);
+
+    SetupShaders(offseted_buffer, buffer_offset);
 
     stream_buffer->Unmap();
 

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -148,13 +148,13 @@ private:
     static constexpr size_t STREAM_BUFFER_SIZE = 4 * 1024 * 1024;
     std::unique_ptr<OGLStreamBuffer> stream_buffer;
 
-    GLsizeiptr vs_input_size;
+    size_t CalculateVertexArraysSize() const;
 
-    void SetupVertexArray(u8* array_ptr, GLintptr buffer_offset);
+    std::pair<u8*, GLintptr> SetupVertexArrays(u8* array_ptr, GLintptr buffer_offset);
 
     std::array<OGLBuffer, Tegra::Engines::Maxwell3D::Regs::MaxShaderStage> uniform_buffers;
 
-    void SetupShaders(u8* buffer_ptr, GLintptr buffer_offset, size_t ptr_pos);
+    void SetupShaders(u8* buffer_ptr, GLintptr buffer_offset);
 
     enum class AccelDraw { Disabled, Arrays, Indexed };
     AccelDraw accelerate_draw;

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -49,6 +49,7 @@ 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, 1},                     // ABGR8
     {GL_RGB, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, false, 1},                        // B5G6R5
+    {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV, false, 1},               // A2B10G10R10
     {GL_COMPRESSED_RGB_S3TC_DXT1_EXT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, true, 16},   // DXT1
     {GL_COMPRESSED_RGBA_S3TC_DXT3_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, true, 16}, // DXT23
     {GL_COMPRESSED_RGBA_S3TC_DXT5_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, true, 16}, // DXT45
@@ -104,9 +105,9 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, VAddr b
 static constexpr std::array<void (*)(u32, u32, u32, u8*, VAddr, VAddr, VAddr),
                             SurfaceParams::MaxPixelFormat>
     morton_to_gl_fns = {
-        MortonCopy<true, PixelFormat::ABGR8>, MortonCopy<true, PixelFormat::B5G6R5>,
-        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::DXT1>,
+        MortonCopy<true, PixelFormat::DXT23>,       MortonCopy<true, PixelFormat::DXT45>,
 };
 
 static constexpr std::array<void (*)(u32, u32, u32, u8*, VAddr, VAddr, VAddr),
@@ -114,6 +115,7 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, VAddr, VAddr, VAddr),
     gl_to_morton_fns = {
         MortonCopy<false, PixelFormat::ABGR8>,
         MortonCopy<false, PixelFormat::B5G6R5>,
+        MortonCopy<false, PixelFormat::A2B10G10R10>,
         // TODO(Subv): Swizzling the DXT1/DXT23/DXT45 formats is not yet supported
         nullptr,
         nullptr,

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -54,9 +54,10 @@ struct SurfaceParams {
     enum class PixelFormat {
         ABGR8 = 0,
         B5G6R5 = 1,
-        DXT1 = 2,
-        DXT23 = 3,
-        DXT45 = 4,
+        A2B10G10R10 = 2,
+        DXT1 = 3,
+        DXT23 = 4,
+        DXT45 = 5,
 
         Max,
         Invalid = 255,
@@ -88,6 +89,7 @@ struct SurfaceParams {
         constexpr std::array<unsigned int, MaxPixelFormat> bpp_table = {
             32,  // ABGR8
             16,  // B5G6R5
+            32,  // A2B10G10R10
             64,  // DXT1
             128, // DXT23
             128, // DXT45
@@ -104,6 +106,8 @@ struct SurfaceParams {
         switch (format) {
         case Tegra::RenderTargetFormat::RGBA8_UNORM:
             return PixelFormat::ABGR8;
+        case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
+            return PixelFormat::A2B10G10R10;
         default:
             NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
@@ -127,6 +131,8 @@ struct SurfaceParams {
             return PixelFormat::ABGR8;
         case Tegra::Texture::TextureFormat::B5G6R5:
             return PixelFormat::B5G6R5;
+        case Tegra::Texture::TextureFormat::A2B10G10R10:
+            return PixelFormat::A2B10G10R10;
         case Tegra::Texture::TextureFormat::DXT1:
             return PixelFormat::DXT1;
         case Tegra::Texture::TextureFormat::DXT23:
@@ -146,6 +152,8 @@ struct SurfaceParams {
             return Tegra::Texture::TextureFormat::A8R8G8B8;
         case PixelFormat::B5G6R5:
             return Tegra::Texture::TextureFormat::B5G6R5;
+        case PixelFormat::A2B10G10R10:
+            return Tegra::Texture::TextureFormat::A2B10G10R10;
         case PixelFormat::DXT1:
             return Tegra::Texture::TextureFormat::DXT1;
         case PixelFormat::DXT23:

src/video_core/renderer_opengl/renderer_opengl.cpp

@@ -295,7 +295,7 @@ void RendererOpenGL::DrawScreenTriangles(const ScreenInfo& screen_info, float x,
     const auto& texcoords = screen_info.display_texcoords;
     auto left = texcoords.left;
     auto right = texcoords.right;
-    if (framebuffer_transform_flags != Tegra::FramebufferConfig::TransformFlags::Unset)
+    if (framebuffer_transform_flags != Tegra::FramebufferConfig::TransformFlags::Unset) {
         if (framebuffer_transform_flags == Tegra::FramebufferConfig::TransformFlags::FlipV) {
             // Flip the framebuffer vertically
             left = texcoords.right;
@@ -303,9 +303,10 @@ void RendererOpenGL::DrawScreenTriangles(const ScreenInfo& screen_info, float x,
         } else {
             // Other transformations are unsupported
             LOG_CRITICAL(Render_OpenGL, "Unsupported framebuffer_transform_flags=%d",
-                         framebuffer_transform_flags);
+                         static_cast<int>(framebuffer_transform_flags));
             UNIMPLEMENTED();
         }
+    }
 
     std::array<ScreenRectVertex, 4> vertices = {{
         ScreenRectVertex(x, y, texcoords.top, left),

src/video_core/textures/decoders.cpp

@@ -53,6 +53,7 @@ u32 BytesPerPixel(TextureFormat format) {
         // In this case a 'pixel' actually refers to a 4x4 tile.
         return 16;
     case TextureFormat::A8R8G8B8:
+    case TextureFormat::A2B10G10R10:
         return 4;
     case TextureFormat::B5G6R5:
         return 2;
@@ -78,6 +79,7 @@ std::vector<u8> UnswizzleTexture(VAddr address, TextureFormat format, u32 width,
                          unswizzled_data.data(), true, block_height);
         break;
     case TextureFormat::A8R8G8B8:
+    case TextureFormat::A2B10G10R10:
     case TextureFormat::B5G6R5:
         CopySwizzledData(width, height, bytes_per_pixel, bytes_per_pixel, data,
                          unswizzled_data.data(), true, block_height);
@@ -100,6 +102,7 @@ std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat
     case TextureFormat::DXT23:
     case TextureFormat::DXT45:
     case TextureFormat::A8R8G8B8:
+    case TextureFormat::A2B10G10R10:
     case TextureFormat::B5G6R5:
         // TODO(Subv): For the time being just forward the same data without any decoding.
         rgba_data = texture_data;

src/video_core/textures/texture.h

@@ -15,6 +15,7 @@ namespace Texture {
 
 enum class TextureFormat : u32 {
     A8R8G8B8 = 0x8,
+    A2B10G10R10 = 0x9,
     B5G6R5 = 0x15,
     DXT1 = 0x24,
     DXT23 = 0x25,

src/yuzu/debugger/graphics/graphics_surface.cpp

@@ -25,6 +25,8 @@ static Tegra::Texture::TextureFormat ConvertToTextureFormat(
     switch (render_target_format) {
     case Tegra::RenderTargetFormat::RGBA8_UNORM:
         return Tegra::Texture::TextureFormat::A8R8G8B8;
+    case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
+        return Tegra::Texture::TextureFormat::A2B10G10R10;
     default:
         UNIMPLEMENTED_MSG("Unimplemented RT format");
     }

src/yuzu/debugger/wait_tree.cpp

@@ -6,8 +6,8 @@
 #include "yuzu/util/util.h"
 
 #include "core/core.h"
-#include "core/hle/kernel/condition_variable.h"
 #include "core/hle/kernel/event.h"
+#include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/mutex.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/timer.h"
@@ -67,6 +67,29 @@ QString WaitTreeText::GetText() const {
     return text;
 }
 
+WaitTreeMutexInfo::WaitTreeMutexInfo(VAddr mutex_address) : mutex_address(mutex_address) {
+    mutex_value = Memory::Read32(mutex_address);
+    owner_handle = static_cast<Kernel::Handle>(mutex_value & Kernel::Mutex::MutexOwnerMask);
+    owner = Kernel::g_handle_table.Get<Kernel::Thread>(owner_handle);
+}
+
+QString WaitTreeMutexInfo::GetText() const {
+    return tr("waiting for mutex 0x%1").arg(mutex_address, 16, 16, QLatin1Char('0'));
+}
+
+std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeMutexInfo::GetChildren() const {
+    std::vector<std::unique_ptr<WaitTreeItem>> list;
+
+    bool has_waiters = (mutex_value & Kernel::Mutex::MutexHasWaitersFlag) != 0;
+
+    list.push_back(std::make_unique<WaitTreeText>(tr("has waiters: %1").arg(has_waiters)));
+    list.push_back(std::make_unique<WaitTreeText>(
+        tr("owner handle: 0x%1").arg(owner_handle, 8, 16, QLatin1Char('0'))));
+    if (owner != nullptr)
+        list.push_back(std::make_unique<WaitTreeThread>(*owner));
+    return list;
+}
+
 WaitTreeWaitObject::WaitTreeWaitObject(const Kernel::WaitObject& o) : object(o) {}
 
 bool WaitTreeExpandableItem::IsExpandable() const {
@@ -84,11 +107,6 @@ std::unique_ptr<WaitTreeWaitObject> WaitTreeWaitObject::make(const Kernel::WaitO
     switch (object.GetHandleType()) {
     case Kernel::HandleType::Event:
         return std::make_unique<WaitTreeEvent>(static_cast<const Kernel::Event&>(object));
-    case Kernel::HandleType::Mutex:
-        return std::make_unique<WaitTreeMutex>(static_cast<const Kernel::Mutex&>(object));
-    case Kernel::HandleType::ConditionVariable:
-        return std::make_unique<WaitTreeConditionVariable>(
-            static_cast<const Kernel::ConditionVariable&>(object));
     case Kernel::HandleType::Timer:
         return std::make_unique<WaitTreeTimer>(static_cast<const Kernel::Timer&>(object));
     case Kernel::HandleType::Thread:
@@ -160,6 +178,9 @@ QString WaitTreeThread::GetText() const {
     case THREADSTATUS_WAIT_SYNCH_ANY:
         status = tr("waiting for objects");
         break;
+    case THREADSTATUS_WAIT_MUTEX:
+        status = tr("waiting for mutex");
+        break;
     case THREADSTATUS_DORMANT:
         status = tr("dormant");
         break;
@@ -186,6 +207,7 @@ QColor WaitTreeThread::GetColor() const {
         return QColor(Qt::GlobalColor::darkYellow);
     case THREADSTATUS_WAIT_SYNCH_ALL:
     case THREADSTATUS_WAIT_SYNCH_ANY:
+    case THREADSTATUS_WAIT_MUTEX:
         return QColor(Qt::GlobalColor::red);
     case THREADSTATUS_DORMANT:
         return QColor(Qt::GlobalColor::darkCyan);
@@ -225,11 +247,11 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeThread::GetChildren() const {
     list.push_back(std::make_unique<WaitTreeText>(
         tr("last running ticks = %1").arg(thread.last_running_ticks)));
 
-    if (thread.held_mutexes.empty()) {
-        list.push_back(std::make_unique<WaitTreeText>(tr("not holding mutex")));
-    } else {
-        list.push_back(std::make_unique<WaitTreeMutexList>(thread.held_mutexes));
-    }
+    if (thread.mutex_wait_address != 0)
+        list.push_back(std::make_unique<WaitTreeMutexInfo>(thread.mutex_wait_address));
+    else
+        list.push_back(std::make_unique<WaitTreeText>(tr("not waiting for mutex")));
+
     if (thread.status == THREADSTATUS_WAIT_SYNCH_ANY ||
         thread.status == THREADSTATUS_WAIT_SYNCH_ALL) {
         list.push_back(std::make_unique<WaitTreeObjectList>(thread.wait_objects,
@@ -250,33 +272,6 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeEvent::GetChildren() const {
     return list;
 }
 
-WaitTreeMutex::WaitTreeMutex(const Kernel::Mutex& object) : WaitTreeWaitObject(object) {}
-
-std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeMutex::GetChildren() const {
-    std::vector<std::unique_ptr<WaitTreeItem>> list(WaitTreeWaitObject::GetChildren());
-
-    const auto& mutex = static_cast<const Kernel::Mutex&>(object);
-    if (mutex.GetHasWaiters()) {
-        list.push_back(std::make_unique<WaitTreeText>(tr("locked by thread:")));
-        list.push_back(std::make_unique<WaitTreeThread>(*mutex.GetHoldingThread()));
-    } else {
-        list.push_back(std::make_unique<WaitTreeText>(tr("free")));
-    }
-    return list;
-}
-
-WaitTreeConditionVariable::WaitTreeConditionVariable(const Kernel::ConditionVariable& object)
-    : WaitTreeWaitObject(object) {}
-
-std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeConditionVariable::GetChildren() const {
-    std::vector<std::unique_ptr<WaitTreeItem>> list(WaitTreeWaitObject::GetChildren());
-
-    const auto& condition_variable = static_cast<const Kernel::ConditionVariable&>(object);
-    list.push_back(std::make_unique<WaitTreeText>(
-        tr("available count = %1").arg(condition_variable.GetAvailableCount())));
-    return list;
-}
-
 WaitTreeTimer::WaitTreeTimer(const Kernel::Timer& object) : WaitTreeWaitObject(object) {}
 
 std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeTimer::GetChildren() const {
@@ -293,21 +288,6 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeTimer::GetChildren() const {
     return list;
 }
 
-WaitTreeMutexList::WaitTreeMutexList(
-    const boost::container::flat_set<Kernel::SharedPtr<Kernel::Mutex>>& list)
-    : mutex_list(list) {}
-
-QString WaitTreeMutexList::GetText() const {
-    return tr("holding mutexes");
-}
-
-std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeMutexList::GetChildren() const {
-    std::vector<std::unique_ptr<WaitTreeItem>> list(mutex_list.size());
-    std::transform(mutex_list.begin(), mutex_list.end(), list.begin(),
-                   [](const auto& t) { return std::make_unique<WaitTreeMutex>(*t); });
-    return list;
-}
-
 WaitTreeThreadList::WaitTreeThreadList(const std::vector<Kernel::SharedPtr<Kernel::Thread>>& list)
     : thread_list(list) {}
 

src/yuzu/debugger/wait_tree.h

@@ -16,8 +16,6 @@ class EmuThread;
 namespace Kernel {
 class WaitObject;
 class Event;
-class Mutex;
-class ConditionVariable;
 class Thread;
 class Timer;
 } // namespace Kernel
@@ -61,6 +59,20 @@ public:
     bool IsExpandable() const override;
 };
 
+class WaitTreeMutexInfo : public WaitTreeExpandableItem {
+    Q_OBJECT
+public:
+    explicit WaitTreeMutexInfo(VAddr mutex_address);
+    QString GetText() const override;
+    std::vector<std::unique_ptr<WaitTreeItem>> GetChildren() const override;
+
+private:
+    VAddr mutex_address;
+    u32 mutex_value;
+    Kernel::Handle owner_handle;
+    Kernel::SharedPtr<Kernel::Thread> owner;
+};
+
 class WaitTreeWaitObject : public WaitTreeExpandableItem {
     Q_OBJECT
 public:
@@ -104,20 +116,6 @@ public:
     std::vector<std::unique_ptr<WaitTreeItem>> GetChildren() const override;
 };
 
-class WaitTreeMutex : public WaitTreeWaitObject {
-    Q_OBJECT
-public:
-    explicit WaitTreeMutex(const Kernel::Mutex& object);
-    std::vector<std::unique_ptr<WaitTreeItem>> GetChildren() const override;
-};
-
-class WaitTreeConditionVariable : public WaitTreeWaitObject {
-    Q_OBJECT
-public:
-    explicit WaitTreeConditionVariable(const Kernel::ConditionVariable& object);
-    std::vector<std::unique_ptr<WaitTreeItem>> GetChildren() const override;
-};
-
 class WaitTreeTimer : public WaitTreeWaitObject {
     Q_OBJECT
 public:
@@ -125,19 +123,6 @@ public:
     std::vector<std::unique_ptr<WaitTreeItem>> GetChildren() const override;
 };
 
-class WaitTreeMutexList : public WaitTreeExpandableItem {
-    Q_OBJECT
-public:
-    explicit WaitTreeMutexList(
-        const boost::container::flat_set<Kernel::SharedPtr<Kernel::Mutex>>& list);
-
-    QString GetText() const override;
-    std::vector<std::unique_ptr<WaitTreeItem>> GetChildren() const override;
-
-private:
-    const boost::container::flat_set<Kernel::SharedPtr<Kernel::Mutex>>& mutex_list;
-};
-
 class WaitTreeThreadList : public WaitTreeExpandableItem {
     Q_OBJECT
 public: