Android #174

kernel: use simple mutex for object list container

README.md

@@ -1,3 +1,14 @@
+| Pull Request | Commit | Title | Author | Merged? |
+|----|----|----|----|----|
+| [12466](https://github.com/yuzu-emu/yuzu//pull/12466) | [`adb2af0a2`](https://github.com/yuzu-emu/yuzu//pull/12466/files) |  core: track separate heap allocation for linux  | [liamwhite](https://github.com/liamwhite/) | Yes |
+| [12479](https://github.com/yuzu-emu/yuzu//pull/12479) | [`20e040723`](https://github.com/yuzu-emu/yuzu//pull/12479/files) | video_core: Fix buffer_row_length for linear compressed textures | [GPUCode](https://github.com/GPUCode/) | Yes |
+| [12487](https://github.com/yuzu-emu/yuzu//pull/12487) | [`d0c60605a`](https://github.com/yuzu-emu/yuzu//pull/12487/files) | shader_recompiler: use default value for clip distances array | [liamwhite](https://github.com/liamwhite/) | Yes |
+
+
+End of merge log. You can find the original README.md below the break.
+
+-----
+
 <!--
 SPDX-FileCopyrightText: 2018 yuzu Emulator Project
 SPDX-License-Identifier: GPL-2.0-or-later

src/common/CMakeLists.txt

@@ -64,6 +64,8 @@ add_library(common STATIC
     fs/path_util.cpp
     fs/path_util.h
     hash.h
+    heap_tracker.cpp
+    heap_tracker.h
     hex_util.cpp
     hex_util.h
     host_memory.cpp

src/common/assert.cpp

@@ -3,16 +3,19 @@
 
 #include "common/assert.h"
 #include "common/common_funcs.h"
+#include "common/logging/backend.h"
 
 #include "common/settings.h"
 
 void assert_fail_impl() {
     if (Settings::values.use_debug_asserts) {
+        Common::Log::Stop();
         Crash();
     }
 }
 
 [[noreturn]] void unreachable_impl() {
+    Common::Log::Stop();
     Crash();
     throw std::runtime_error("Unreachable code");
 }

src/common/heap_tracker.cpp

@@ -0,0 +1,281 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include <fstream>
+#include <vector>
+
+#include "common/heap_tracker.h"
+#include "common/logging/log.h"
+
+namespace Common {
+
+namespace {
+
+s64 GetMaxPermissibleResidentMapCount() {
+    // Default value.
+    s64 value = 65530;
+
+    // Try to read how many mappings we can make.
+    std::ifstream s("/proc/sys/vm/max_map_count");
+    s >> value;
+
+    // Print, for debug.
+    LOG_INFO(HW_Memory, "Current maximum map count: {}", value);
+
+    // Allow 20000 maps for other code and to account for split inaccuracy.
+    return std::max<s64>(value - 20000, 0);
+}
+
+} // namespace
+
+HeapTracker::HeapTracker(Common::HostMemory& buffer)
+    : m_buffer(buffer), m_max_resident_map_count(GetMaxPermissibleResidentMapCount()) {}
+HeapTracker::~HeapTracker() = default;
+
+void HeapTracker::Map(size_t virtual_offset, size_t host_offset, size_t length,
+                      MemoryPermission perm, bool is_separate_heap) {
+    // When mapping other memory, map pages immediately.
+    if (!is_separate_heap) {
+        m_buffer.Map(virtual_offset, host_offset, length, perm, false);
+        return;
+    }
+
+    {
+        // We are mapping part of a separate heap.
+        std::scoped_lock lk{m_lock};
+
+        auto* const map = new SeparateHeapMap{
+            .vaddr = virtual_offset,
+            .paddr = host_offset,
+            .size = length,
+            .tick = m_tick++,
+            .perm = perm,
+            .is_resident = false,
+        };
+
+        // Insert into mappings.
+        m_map_count++;
+        m_mappings.insert(*map);
+    }
+
+    // Finally, map.
+    this->DeferredMapSeparateHeap(virtual_offset);
+}
+
+void HeapTracker::Unmap(size_t virtual_offset, size_t size, bool is_separate_heap) {
+    // If this is a separate heap...
+    if (is_separate_heap) {
+        std::scoped_lock lk{m_lock};
+
+        const SeparateHeapMap key{
+            .vaddr = virtual_offset,
+        };
+
+        // Split at the boundaries of the region we are removing.
+        this->SplitHeapMapLocked(virtual_offset);
+        this->SplitHeapMapLocked(virtual_offset + size);
+
+        // Erase all mappings in range.
+        auto it = m_mappings.find(key);
+        while (it != m_mappings.end() && it->vaddr < virtual_offset + size) {
+            // Get underlying item.
+            auto* const item = std::addressof(*it);
+
+            // If resident, erase from resident map.
+            if (item->is_resident) {
+                ASSERT(--m_resident_map_count >= 0);
+                m_resident_mappings.erase(m_resident_mappings.iterator_to(*item));
+            }
+
+            // Erase from map.
+            ASSERT(--m_map_count >= 0);
+            it = m_mappings.erase(it);
+
+            // Free the item.
+            delete item;
+        }
+    }
+
+    // Unmap pages.
+    m_buffer.Unmap(virtual_offset, size, false);
+}
+
+void HeapTracker::Protect(size_t virtual_offset, size_t size, MemoryPermission perm) {
+    // Ensure no rebuild occurs while reprotecting.
+    std::shared_lock lk{m_rebuild_lock};
+
+    // Split at the boundaries of the region we are reprotecting.
+    this->SplitHeapMap(virtual_offset, size);
+
+    // Declare tracking variables.
+    const VAddr end = virtual_offset + size;
+    VAddr cur = virtual_offset;
+
+    while (cur < end) {
+        VAddr next = cur;
+        bool should_protect = false;
+
+        {
+            std::scoped_lock lk2{m_lock};
+
+            const SeparateHeapMap key{
+                .vaddr = next,
+            };
+
+            // Try to get the next mapping corresponding to this address.
+            const auto it = m_mappings.nfind(key);
+
+            if (it == m_mappings.end()) {
+                // There are no separate heap mappings remaining.
+                next = end;
+                should_protect = true;
+            } else if (it->vaddr == cur) {
+                // We are in range.
+                // Update permission bits.
+                it->perm = perm;
+
+                // Determine next address and whether we should protect.
+                next = cur + it->size;
+                should_protect = it->is_resident;
+            } else /* if (it->vaddr > cur) */ {
+                // We weren't in range, but there is a block coming up that will be.
+                next = it->vaddr;
+                should_protect = true;
+            }
+        }
+
+        // Clamp to end.
+        next = std::min(next, end);
+
+        // Reprotect, if we need to.
+        if (should_protect) {
+            m_buffer.Protect(cur, next - cur, perm);
+        }
+
+        // Advance.
+        cur = next;
+    }
+}
+
+bool HeapTracker::DeferredMapSeparateHeap(u8* fault_address) {
+    if (m_buffer.IsInVirtualRange(fault_address)) {
+        return this->DeferredMapSeparateHeap(fault_address - m_buffer.VirtualBasePointer());
+    }
+
+    return false;
+}
+
+bool HeapTracker::DeferredMapSeparateHeap(size_t virtual_offset) {
+    bool rebuild_required = false;
+
+    {
+        std::scoped_lock lk{m_lock};
+
+        // Check to ensure this was a non-resident separate heap mapping.
+        const auto it = this->GetNearestHeapMapLocked(virtual_offset);
+        if (it == m_mappings.end() || it->is_resident) {
+            return false;
+        }
+
+        // Update tick before possible rebuild.
+        it->tick = m_tick++;
+
+        // Check if we need to rebuild.
+        if (m_resident_map_count > m_max_resident_map_count) {
+            rebuild_required = true;
+        }
+
+        // Map the area.
+        m_buffer.Map(it->vaddr, it->paddr, it->size, it->perm, false);
+
+        // This map is now resident.
+        it->is_resident = true;
+        m_resident_map_count++;
+        m_resident_mappings.insert(*it);
+    }
+
+    if (rebuild_required) {
+        // A rebuild was required, so perform it now.
+        this->RebuildSeparateHeapAddressSpace();
+    }
+
+    return true;
+}
+
+void HeapTracker::RebuildSeparateHeapAddressSpace() {
+    std::scoped_lock lk{m_rebuild_lock, m_lock};
+
+    ASSERT(!m_resident_mappings.empty());
+
+    // Dump half of the mappings.
+    //
+    // Despite being worse in theory, this has proven to be better in practice than more
+    // regularly dumping a smaller amount, because it significantly reduces average case
+    // lock contention.
+    const size_t desired_count = std::min(m_resident_map_count, m_max_resident_map_count) / 2;
+    const size_t evict_count = m_resident_map_count - desired_count;
+    auto it = m_resident_mappings.begin();
+
+    for (size_t i = 0; i < evict_count && it != m_resident_mappings.end(); i++) {
+        // Unmark and unmap.
+        it->is_resident = false;
+        m_buffer.Unmap(it->vaddr, it->size, false);
+
+        // Advance.
+        ASSERT(--m_resident_map_count >= 0);
+        it = m_resident_mappings.erase(it);
+    }
+}
+
+void HeapTracker::SplitHeapMap(VAddr offset, size_t size) {
+    std::scoped_lock lk{m_lock};
+
+    this->SplitHeapMapLocked(offset);
+    this->SplitHeapMapLocked(offset + size);
+}
+
+void HeapTracker::SplitHeapMapLocked(VAddr offset) {
+    const auto it = this->GetNearestHeapMapLocked(offset);
+    if (it == m_mappings.end() || it->vaddr == offset) {
+        // Not contained or no split required.
+        return;
+    }
+
+    // Cache the original values.
+    auto* const left = std::addressof(*it);
+    const size_t orig_size = left->size;
+
+    // Adjust the left map.
+    const size_t left_size = offset - left->vaddr;
+    left->size = left_size;
+
+    // Create the new right map.
+    auto* const right = new SeparateHeapMap{
+        .vaddr = left->vaddr + left_size,
+        .paddr = left->paddr + left_size,
+        .size = orig_size - left_size,
+        .tick = left->tick,
+        .perm = left->perm,
+        .is_resident = left->is_resident,
+    };
+
+    // Insert the new right map.
+    m_map_count++;
+    m_mappings.insert(*right);
+
+    // If resident, also insert into resident map.
+    if (right->is_resident) {
+        m_resident_map_count++;
+        m_resident_mappings.insert(*right);
+    }
+}
+
+HeapTracker::AddrTree::iterator HeapTracker::GetNearestHeapMapLocked(VAddr offset) {
+    const SeparateHeapMap key{
+        .vaddr = offset,
+    };
+
+    return m_mappings.find(key);
+}
+
+} // namespace Common

src/common/heap_tracker.h

@@ -0,0 +1,98 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include <atomic>
+#include <mutex>
+#include <set>
+#include <shared_mutex>
+
+#include "common/host_memory.h"
+#include "common/intrusive_red_black_tree.h"
+
+namespace Common {
+
+struct SeparateHeapMap {
+    Common::IntrusiveRedBlackTreeNode addr_node{};
+    Common::IntrusiveRedBlackTreeNode tick_node{};
+    VAddr vaddr{};
+    PAddr paddr{};
+    size_t size{};
+    size_t tick{};
+    MemoryPermission perm{};
+    bool is_resident{};
+};
+
+struct SeparateHeapMapAddrComparator {
+    static constexpr int Compare(const SeparateHeapMap& lhs, const SeparateHeapMap& rhs) {
+        if (lhs.vaddr < rhs.vaddr) {
+            return -1;
+        } else if (lhs.vaddr <= (rhs.vaddr + rhs.size - 1)) {
+            return 0;
+        } else {
+            return 1;
+        }
+    }
+};
+
+struct SeparateHeapMapTickComparator {
+    static constexpr int Compare(const SeparateHeapMap& lhs, const SeparateHeapMap& rhs) {
+        if (lhs.tick < rhs.tick) {
+            return -1;
+        } else if (lhs.tick > rhs.tick) {
+            return 1;
+        } else {
+            return SeparateHeapMapAddrComparator::Compare(lhs, rhs);
+        }
+    }
+};
+
+class HeapTracker {
+public:
+    explicit HeapTracker(Common::HostMemory& buffer);
+    ~HeapTracker();
+
+    void Map(size_t virtual_offset, size_t host_offset, size_t length, MemoryPermission perm,
+             bool is_separate_heap);
+    void Unmap(size_t virtual_offset, size_t size, bool is_separate_heap);
+    void Protect(size_t virtual_offset, size_t length, MemoryPermission perm);
+    u8* VirtualBasePointer() {
+        return m_buffer.VirtualBasePointer();
+    }
+
+    bool DeferredMapSeparateHeap(u8* fault_address);
+    bool DeferredMapSeparateHeap(size_t virtual_offset);
+
+private:
+    using AddrTreeTraits =
+        Common::IntrusiveRedBlackTreeMemberTraitsDeferredAssert<&SeparateHeapMap::addr_node>;
+    using AddrTree = AddrTreeTraits::TreeType<SeparateHeapMapAddrComparator>;
+
+    using TickTreeTraits =
+        Common::IntrusiveRedBlackTreeMemberTraitsDeferredAssert<&SeparateHeapMap::tick_node>;
+    using TickTree = TickTreeTraits::TreeType<SeparateHeapMapTickComparator>;
+
+    AddrTree m_mappings{};
+    TickTree m_resident_mappings{};
+
+private:
+    void SplitHeapMap(VAddr offset, size_t size);
+    void SplitHeapMapLocked(VAddr offset);
+
+    AddrTree::iterator GetNearestHeapMapLocked(VAddr offset);
+
+    void RebuildSeparateHeapAddressSpace();
+
+private:
+    Common::HostMemory& m_buffer;
+    const s64 m_max_resident_map_count;
+
+    std::shared_mutex m_rebuild_lock{};
+    std::mutex m_lock{};
+    s64 m_map_count{};
+    s64 m_resident_map_count{};
+    size_t m_tick{};
+};
+
+} // namespace Common

src/common/host_memory.cpp

@@ -679,7 +679,7 @@ HostMemory::HostMemory(HostMemory&&) noexcept = default;
 HostMemory& HostMemory::operator=(HostMemory&&) noexcept = default;
 
 void HostMemory::Map(size_t virtual_offset, size_t host_offset, size_t length,
-                     MemoryPermission perms) {
+                     MemoryPermission perms, bool separate_heap) {
     ASSERT(virtual_offset % PageAlignment == 0);
     ASSERT(host_offset % PageAlignment == 0);
     ASSERT(length % PageAlignment == 0);
@@ -691,7 +691,7 @@ void HostMemory::Map(size_t virtual_offset, size_t host_offset, size_t length,
     impl->Map(virtual_offset + virtual_base_offset, host_offset, length, perms);
 }
 
-void HostMemory::Unmap(size_t virtual_offset, size_t length) {
+void HostMemory::Unmap(size_t virtual_offset, size_t length, bool separate_heap) {
     ASSERT(virtual_offset % PageAlignment == 0);
     ASSERT(length % PageAlignment == 0);
     ASSERT(virtual_offset + length <= virtual_size);
@@ -701,14 +701,16 @@ void HostMemory::Unmap(size_t virtual_offset, size_t length) {
     impl->Unmap(virtual_offset + virtual_base_offset, length);
 }
 
-void HostMemory::Protect(size_t virtual_offset, size_t length, bool read, bool write,
-                         bool execute) {
+void HostMemory::Protect(size_t virtual_offset, size_t length, MemoryPermission perm) {
     ASSERT(virtual_offset % PageAlignment == 0);
     ASSERT(length % PageAlignment == 0);
     ASSERT(virtual_offset + length <= virtual_size);
     if (length == 0 || !virtual_base || !impl) {
         return;
     }
+    const bool read = True(perm & MemoryPermission::Read);
+    const bool write = True(perm & MemoryPermission::Write);
+    const bool execute = True(perm & MemoryPermission::Execute);
     impl->Protect(virtual_offset + virtual_base_offset, length, read, write, execute);
 }
 

src/common/host_memory.h

@@ -40,11 +40,12 @@ public:
     HostMemory(HostMemory&& other) noexcept;
     HostMemory& operator=(HostMemory&& other) noexcept;
 
-    void Map(size_t virtual_offset, size_t host_offset, size_t length, MemoryPermission perms);
+    void Map(size_t virtual_offset, size_t host_offset, size_t length, MemoryPermission perms,
+             bool separate_heap);
 
-    void Unmap(size_t virtual_offset, size_t length);
+    void Unmap(size_t virtual_offset, size_t length, bool separate_heap);
 
-    void Protect(size_t virtual_offset, size_t length, bool read, bool write, bool execute = false);
+    void Protect(size_t virtual_offset, size_t length, MemoryPermission perms);
 
     void EnableDirectMappedAddress();
 
@@ -64,6 +65,10 @@ public:
         return virtual_base;
     }
 
+    bool IsInVirtualRange(void* address) const noexcept {
+        return address >= virtual_base && address < virtual_base + virtual_size;
+    }
+
 private:
     size_t backing_size{};
     size_t virtual_size{};

src/common/logging/backend.cpp

@@ -208,6 +208,10 @@ public:
         instance->StartBackendThread();
     }
 
+    static void Stop() {
+        instance->StopBackendThread();
+    }
+
     Impl(const Impl&) = delete;
     Impl& operator=(const Impl&) = delete;
 
@@ -259,6 +263,15 @@ private:
         });
     }
 
+    void StopBackendThread() {
+        backend_thread.request_stop();
+        if (backend_thread.joinable()) {
+            backend_thread.join();
+        }
+
+        ForEachBackend([](Backend& backend) { backend.Flush(); });
+    }
+
     Entry CreateEntry(Class log_class, Level log_level, const char* filename, unsigned int line_nr,
                       const char* function, std::string&& message) const {
         using std::chrono::duration_cast;
@@ -313,6 +326,10 @@ void Start() {
     Impl::Start();
 }
 
+void Stop() {
+    Impl::Stop();
+}
+
 void DisableLoggingInTests() {
     initialization_in_progress_suppress_logging = true;
 }

src/common/logging/backend.h

@@ -14,6 +14,9 @@ void Initialize();
 
 void Start();
 
+/// Explicitly stops the logger thread and flushes the buffers
+void Stop();
+
 void DisableLoggingInTests();
 
 /**

src/common/ring_buffer.h

@@ -103,7 +103,7 @@ private:
     // Having them on the same cache-line would result in false-sharing between them.
     // TODO: Remove this ifdef whenever clang and GCC support
     //       std::hardware_destructive_interference_size.
-#if defined(_MSC_VER) && _MSC_VER >= 1911
+#ifdef __cpp_lib_hardware_interference_size
     alignas(std::hardware_destructive_interference_size) std::atomic_size_t m_read_index{0};
     alignas(std::hardware_destructive_interference_size) std::atomic_size_t m_write_index{0};
 #else

src/core/CMakeLists.txt

@@ -978,6 +978,7 @@ endif()
 
 if (ARCHITECTURE_x86_64 OR ARCHITECTURE_arm64)
     target_sources(core PRIVATE
+        arm/dynarmic/arm_dynarmic.cpp
         arm/dynarmic/arm_dynarmic.h
         arm/dynarmic/arm_dynarmic_64.cpp
         arm/dynarmic/arm_dynarmic_64.h

src/core/arm/dynarmic/arm_dynarmic.cpp

@@ -0,0 +1,49 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#ifdef __linux__
+
+#include "common/signal_chain.h"
+
+#include "core/arm/dynarmic/arm_dynarmic.h"
+#include "core/hle/kernel/k_process.h"
+#include "core/memory.h"
+
+namespace Core {
+
+namespace {
+
+thread_local Core::Memory::Memory* g_current_memory{};
+std::once_flag g_registered{};
+struct sigaction g_old_segv {};
+
+void HandleSigSegv(int sig, siginfo_t* info, void* ctx) {
+    if (g_current_memory && g_current_memory->InvalidateSeparateHeap(info->si_addr)) {
+        return;
+    }
+
+    return g_old_segv.sa_sigaction(sig, info, ctx);
+}
+
+} // namespace
+
+ScopedJitExecution::ScopedJitExecution(Kernel::KProcess* process) {
+    g_current_memory = std::addressof(process->GetMemory());
+}
+
+ScopedJitExecution::~ScopedJitExecution() {
+    g_current_memory = nullptr;
+}
+
+void ScopedJitExecution::RegisterHandler() {
+    std::call_once(g_registered, [] {
+        struct sigaction sa {};
+        sa.sa_sigaction = &HandleSigSegv;
+        sa.sa_flags = SA_SIGINFO | SA_ONSTACK;
+        Common::SigAction(SIGSEGV, std::addressof(sa), std::addressof(g_old_segv));
+    });
+}
+
+} // namespace Core
+
+#endif

src/core/arm/dynarmic/arm_dynarmic.h

@@ -26,4 +26,24 @@ constexpr HaltReason TranslateHaltReason(Dynarmic::HaltReason hr) {
     return static_cast<HaltReason>(hr);
 }
 
+#ifdef __linux__
+
+class ScopedJitExecution {
+public:
+    explicit ScopedJitExecution(Kernel::KProcess* process);
+    ~ScopedJitExecution();
+    static void RegisterHandler();
+};
+
+#else
+
+class ScopedJitExecution {
+public:
+    explicit ScopedJitExecution(Kernel::KProcess* process) {}
+    ~ScopedJitExecution() {}
+    static void RegisterHandler() {}
+};
+
+#endif
+
 } // namespace Core

src/core/arm/dynarmic/arm_dynarmic_32.cpp

@@ -331,11 +331,15 @@ bool ArmDynarmic32::IsInThumbMode() const {
 }
 
 HaltReason ArmDynarmic32::RunThread(Kernel::KThread* thread) {
+    ScopedJitExecution sj(thread->GetOwnerProcess());
+
     m_jit->ClearExclusiveState();
     return TranslateHaltReason(m_jit->Run());
 }
 
 HaltReason ArmDynarmic32::StepThread(Kernel::KThread* thread) {
+    ScopedJitExecution sj(thread->GetOwnerProcess());
+
     m_jit->ClearExclusiveState();
     return TranslateHaltReason(m_jit->Step());
 }
@@ -377,6 +381,7 @@ ArmDynarmic32::ArmDynarmic32(System& system, bool uses_wall_clock, Kernel::KProc
       m_cp15(std::make_shared<DynarmicCP15>(*this)), m_core_index{core_index} {
     auto& page_table_impl = process->GetPageTable().GetBasePageTable().GetImpl();
     m_jit = MakeJit(&page_table_impl);
+    ScopedJitExecution::RegisterHandler();
 }
 
 ArmDynarmic32::~ArmDynarmic32() = default;

src/core/arm/dynarmic/arm_dynarmic_64.cpp

@@ -362,11 +362,15 @@ std::shared_ptr<Dynarmic::A64::Jit> ArmDynarmic64::MakeJit(Common::PageTable* pa
 }
 
 HaltReason ArmDynarmic64::RunThread(Kernel::KThread* thread) {
+    ScopedJitExecution sj(thread->GetOwnerProcess());
+
     m_jit->ClearExclusiveState();
     return TranslateHaltReason(m_jit->Run());
 }
 
 HaltReason ArmDynarmic64::StepThread(Kernel::KThread* thread) {
+    ScopedJitExecution sj(thread->GetOwnerProcess());
+
     m_jit->ClearExclusiveState();
     return TranslateHaltReason(m_jit->Step());
 }
@@ -406,6 +410,7 @@ ArmDynarmic64::ArmDynarmic64(System& system, bool uses_wall_clock, Kernel::KProc
     auto& page_table = process->GetPageTable().GetBasePageTable();
     auto& page_table_impl = page_table.GetImpl();
     m_jit = MakeJit(&page_table_impl, page_table.GetAddressSpaceWidth());
+    ScopedJitExecution::RegisterHandler();
 }
 
 ArmDynarmic64::~ArmDynarmic64() = default;

src/core/hle/kernel/k_auto_object_container.cpp

@@ -8,19 +8,22 @@
 namespace Kernel {
 
 void KAutoObjectWithListContainer::Register(KAutoObjectWithList* obj) {
-    KScopedLightLock lk(m_lock);
+    // KScopedInterruptDisable di;
+    KScopedSpinLock lk(m_lock);
 
     m_object_list.insert_unique(*obj);
 }
 
 void KAutoObjectWithListContainer::Unregister(KAutoObjectWithList* obj) {
-    KScopedLightLock lk(m_lock);
+    // KScopedInterruptDisable di;
+    KScopedSpinLock lk(m_lock);
 
     m_object_list.erase(*obj);
 }
 
 size_t KAutoObjectWithListContainer::GetOwnedCount(KProcess* owner) {
-    KScopedLightLock lk(m_lock);
+    // KScopedInterruptDisable di;
+    KScopedSpinLock lk(m_lock);
 
     return std::count_if(m_object_list.begin(), m_object_list.end(),
                          [&](const auto& obj) { return obj.GetOwner() == owner; });

src/core/hle/kernel/k_auto_object_container.h

@@ -7,7 +7,7 @@
 
 #include "common/common_funcs.h"
 #include "core/hle/kernel/k_auto_object.h"
-#include "core/hle/kernel/k_light_lock.h"
+#include "core/hle/kernel/k_spin_lock.h"
 
 namespace Kernel {
 
@@ -21,32 +21,7 @@ public:
 
     using ListType = boost::intrusive::rbtree<KAutoObjectWithList>;
 
-    class ListAccessor : public KScopedLightLock {
-    public:
-        explicit ListAccessor(KAutoObjectWithListContainer* container)
-            : KScopedLightLock(container->m_lock), m_list(container->m_object_list) {}
-        explicit ListAccessor(KAutoObjectWithListContainer& container)
-            : KScopedLightLock(container.m_lock), m_list(container.m_object_list) {}
-
-        typename ListType::iterator begin() const {
-            return m_list.begin();
-        }
-
-        typename ListType::iterator end() const {
-            return m_list.end();
-        }
-
-        typename ListType::iterator find(typename ListType::const_reference ref) const {
-            return m_list.find(ref);
-        }
-
-    private:
-        ListType& m_list;
-    };
-
-    friend class ListAccessor;
-
-    KAutoObjectWithListContainer(KernelCore& kernel) : m_lock(kernel), m_object_list() {}
+    KAutoObjectWithListContainer(KernelCore& kernel) : m_lock(), m_object_list() {}
 
     void Initialize() {}
     void Finalize() {}
@@ -56,7 +31,7 @@ public:
     size_t GetOwnedCount(KProcess* owner);
 
 private:
-    KLightLock m_lock;
+    KSpinLock m_lock;
     ListType m_object_list;
 };
 

src/core/hle/kernel/k_handle_table.h

@@ -30,7 +30,7 @@ public:
 public:
     explicit KHandleTable(KernelCore& kernel) : m_kernel(kernel) {}
 
-    Result Initialize(KProcess* owner, s32 size) {
+    Result Initialize(s32 size) {
         // Check that the table size is valid.
         R_UNLESS(size <= static_cast<s32>(MaxTableSize), ResultOutOfMemory);
 
@@ -44,7 +44,6 @@ public:
         m_next_linear_id = MinLinearId;
         m_count = 0;
         m_free_head_index = -1;
-        m_owner = owner;
 
         // Free all entries.
         for (s32 i = 0; i < static_cast<s32>(m_table_size); ++i) {
@@ -91,8 +90,7 @@ public:
         // Handle pseudo-handles.
         if constexpr (std::derived_from<KProcess, T>) {
             if (handle == Svc::PseudoHandle::CurrentProcess) {
-                // TODO: this should be the current process
-                auto* const cur_process = m_owner;
+                auto* const cur_process = GetCurrentProcessPointer(m_kernel);
                 ASSERT(cur_process != nullptr);
                 return cur_process;
             }
@@ -302,7 +300,6 @@ private:
 
 private:
     KernelCore& m_kernel;
-    KProcess* m_owner{};
     std::array<EntryInfo, MaxTableSize> m_entry_infos{};
     std::array<KAutoObject*, MaxTableSize> m_objects{};
     mutable KSpinLock m_lock;

src/core/hle/kernel/k_page_table_base.cpp

@@ -434,7 +434,7 @@ Result KPageTableBase::InitializeForProcess(Svc::CreateProcessFlag as_type, bool
 void KPageTableBase::Finalize() {
     auto HostUnmapCallback = [&](KProcessAddress addr, u64 size) {
         if (Settings::IsFastmemEnabled()) {
-            m_system.DeviceMemory().buffer.Unmap(GetInteger(addr), size);
+            m_system.DeviceMemory().buffer.Unmap(GetInteger(addr), size, false);
         }
     };
 
@@ -5243,7 +5243,7 @@ Result KPageTableBase::MapPhysicalMemory(KProcessAddress address, size_t size) {
                                 // Unmap.
                                 R_ASSERT(this->Operate(updater.GetPageList(), cur_address,
                                                        cur_pages, 0, false, unmap_properties,
-                                                       OperationType::Unmap, true));
+                                                       OperationType::UnmapPhysical, true));
                             }
 
                             // Check if we're done.
@@ -5326,7 +5326,7 @@ Result KPageTableBase::MapPhysicalMemory(KProcessAddress address, size_t size) {
                             // Map the papges.
                             R_TRY(this->Operate(updater.GetPageList(), cur_address, map_pages,
                                                 cur_pg, map_properties,
-                                                OperationType::MapFirstGroup, false));
+                                                OperationType::MapFirstGroupPhysical, false));
                         }
                     }
 
@@ -5480,7 +5480,7 @@ Result KPageTableBase::UnmapPhysicalMemory(KProcessAddress address, size_t size)
 
             // Unmap.
             R_ASSERT(this->Operate(updater.GetPageList(), cur_address, cur_pages, 0, false,
-                                   unmap_properties, OperationType::Unmap, false));
+                                   unmap_properties, OperationType::UnmapPhysical, false));
         }
 
         // Check if we're done.
@@ -5655,7 +5655,10 @@ Result KPageTableBase::Operate(PageLinkedList* page_list, KProcessAddress virt_a
     // or free them to the page list, and so it goes unused (along with page properties).
 
     switch (operation) {
-    case OperationType::Unmap: {
+    case OperationType::Unmap:
+    case OperationType::UnmapPhysical: {
+        const bool separate_heap = operation == OperationType::UnmapPhysical;
+
         // Ensure that any pages we track are closed on exit.
         KPageGroup pages_to_close(m_kernel, this->GetBlockInfoManager());
         SCOPE_EXIT({ pages_to_close.CloseAndReset(); });
@@ -5664,7 +5667,7 @@ Result KPageTableBase::Operate(PageLinkedList* page_list, KProcessAddress virt_a
         this->MakePageGroup(pages_to_close, virt_addr, num_pages);
 
         // Unmap.
-        m_memory->UnmapRegion(*m_impl, virt_addr, num_pages * PageSize);
+        m_memory->UnmapRegion(*m_impl, virt_addr, num_pages * PageSize, separate_heap);
 
         R_SUCCEED();
     }
@@ -5672,7 +5675,7 @@ Result KPageTableBase::Operate(PageLinkedList* page_list, KProcessAddress virt_a
         ASSERT(virt_addr != 0);
         ASSERT(Common::IsAligned(GetInteger(virt_addr), PageSize));
         m_memory->MapMemoryRegion(*m_impl, virt_addr, num_pages * PageSize, phys_addr,
-                                  ConvertToMemoryPermission(properties.perm));
+                                  ConvertToMemoryPermission(properties.perm), false);
 
         // Open references to pages, if we should.
         if (this->IsHeapPhysicalAddress(phys_addr)) {
@@ -5711,16 +5714,19 @@ Result KPageTableBase::Operate(PageLinkedList* page_list, KProcessAddress virt_a
 
     switch (operation) {
     case OperationType::MapGroup:
-    case OperationType::MapFirstGroup: {
+    case OperationType::MapFirstGroup:
+    case OperationType::MapFirstGroupPhysical: {
+        const bool separate_heap = operation == OperationType::MapFirstGroupPhysical;
+
         // We want to maintain a new reference to every page in the group.
-        KScopedPageGroup spg(page_group, operation != OperationType::MapFirstGroup);
+        KScopedPageGroup spg(page_group, operation == OperationType::MapGroup);
 
         for (const auto& node : page_group) {
             const size_t size{node.GetNumPages() * PageSize};
 
             // Map the pages.
             m_memory->MapMemoryRegion(*m_impl, virt_addr, size, node.GetAddress(),
-                                      ConvertToMemoryPermission(properties.perm));
+                                      ConvertToMemoryPermission(properties.perm), separate_heap);
 
             virt_addr += size;
         }

src/core/hle/kernel/k_page_table_base.h

@@ -104,6 +104,9 @@ protected:
         ChangePermissionsAndRefresh = 5,
         ChangePermissionsAndRefreshAndFlush = 6,
         Separate = 7,
+
+        MapFirstGroupPhysical = 65000,
+        UnmapPhysical = 65001,
     };
 
     static constexpr size_t MaxPhysicalMapAlignment = 1_GiB;

src/core/hle/kernel/k_process.cpp

@@ -1237,8 +1237,10 @@ void KProcess::LoadModule(CodeSet code_set, KProcessAddress base_addr) {
         auto& buffer = m_kernel.System().DeviceMemory().buffer;
         const auto& code = code_set.CodeSegment();
         const auto& patch = code_set.PatchSegment();
-        buffer.Protect(GetInteger(base_addr + code.addr), code.size, true, true, true);
-        buffer.Protect(GetInteger(base_addr + patch.addr), patch.size, true, true, true);
+        buffer.Protect(GetInteger(base_addr + code.addr), code.size,
+                       Common::MemoryPermission::Read | Common::MemoryPermission::Execute);
+        buffer.Protect(GetInteger(base_addr + patch.addr), patch.size,
+                       Common::MemoryPermission::Read | Common::MemoryPermission::Execute);
         ReprotectSegment(code_set.PatchSegment(), Svc::MemoryPermission::None);
     }
 #endif

src/core/hle/kernel/k_process.h

@@ -552,7 +552,7 @@ private:
 
     Result InitializeHandleTable(s32 size) {
         // Try to initialize the handle table.
-        R_TRY(m_handle_table.Initialize(this, size));
+        R_TRY(m_handle_table.Initialize(size));
 
         // We succeeded, so note that we did.
         m_is_handle_table_initialized = true;

src/core/hle/kernel/k_server_session.cpp

@@ -1147,8 +1147,7 @@ Result KServerSession::ReceiveRequest(uintptr_t server_message, uintptr_t server
         *out_context =
             std::make_shared<Service::HLERequestContext>(m_kernel, memory, this, client_thread);
         (*out_context)->SetSessionRequestManager(manager);
-        (*out_context)
-            ->PopulateFromIncomingCommandBuffer(*client_thread->GetOwnerProcess(), cmd_buf);
+        (*out_context)->PopulateFromIncomingCommandBuffer(cmd_buf);
         // We succeeded.
         R_SUCCEED();
     } else {

src/core/hle/kernel/k_transfer_memory.h

@@ -5,6 +5,7 @@
 
 #include <optional>
 
+#include "core/hle/kernel/k_light_lock.h"
 #include "core/hle/kernel/k_page_group.h"
 #include "core/hle/kernel/slab_helpers.h"
 #include "core/hle/kernel/svc_types.h"

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

@@ -1513,8 +1513,7 @@ void ILibraryAppletCreator::CreateTransferMemoryStorage(HLERequestContext& ctx)
         return;
     }
 
-    auto transfer_mem =
-        system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(handle);
+    auto transfer_mem = ctx.GetObjectFromHandle<Kernel::KTransferMemory>(handle);
 
     if (transfer_mem.IsNull()) {
         LOG_ERROR(Service_AM, "transfer_mem is a nullptr for handle={:08X}", handle);
@@ -1524,8 +1523,7 @@ void ILibraryAppletCreator::CreateTransferMemoryStorage(HLERequestContext& ctx)
     }
 
     std::vector<u8> memory(transfer_mem->GetSize());
-    system.ApplicationMemory().ReadBlock(transfer_mem->GetSourceAddress(), memory.data(),
-                                         memory.size());
+    ctx.GetMemory().ReadBlock(transfer_mem->GetSourceAddress(), memory.data(), memory.size());
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(ResultSuccess);
@@ -1547,8 +1545,7 @@ void ILibraryAppletCreator::CreateHandleStorage(HLERequestContext& ctx) {
         return;
     }
 
-    auto transfer_mem =
-        system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(handle);
+    auto transfer_mem = ctx.GetObjectFromHandle<Kernel::KTransferMemory>(handle);
 
     if (transfer_mem.IsNull()) {
         LOG_ERROR(Service_AM, "transfer_mem is a nullptr for handle={:08X}", handle);
@@ -1558,8 +1555,7 @@ void ILibraryAppletCreator::CreateHandleStorage(HLERequestContext& ctx) {
     }
 
     std::vector<u8> memory(transfer_mem->GetSize());
-    system.ApplicationMemory().ReadBlock(transfer_mem->GetSourceAddress(), memory.data(),
-                                         memory.size());
+    ctx.GetMemory().ReadBlock(transfer_mem->GetSourceAddress(), memory.data(), memory.size());
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(ResultSuccess);

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

@@ -454,10 +454,8 @@ void AudRenU::OpenAudioRenderer(HLERequestContext& ctx) {
         return;
     }
 
-    const auto& handle_table{system.ApplicationProcess()->GetHandleTable()};
-    auto process{handle_table.GetObject<Kernel::KProcess>(process_handle)};
-    auto transfer_memory{
-        process->GetHandleTable().GetObject<Kernel::KTransferMemory>(transfer_memory_handle)};
+    auto process{ctx.GetObjectFromHandle<Kernel::KProcess>(process_handle)};
+    auto transfer_memory{ctx.GetObjectFromHandle<Kernel::KTransferMemory>(transfer_memory_handle)};
 
     const auto session_id{impl->GetSessionId()};
     if (session_id == -1) {

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

@@ -278,9 +278,7 @@ void HwOpus::OpenHardwareOpusDecoder(HLERequestContext& ctx) {
     auto params = rp.PopRaw<OpusParameters>();
     auto transfer_memory_size{rp.Pop<u32>()};
     auto transfer_memory_handle{ctx.GetCopyHandle(0)};
-    auto transfer_memory{
-        system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-            transfer_memory_handle)};
+    auto transfer_memory{ctx.GetObjectFromHandle<Kernel::KTransferMemory>(transfer_memory_handle)};
 
     LOG_DEBUG(Service_Audio, "sample_rate {} channel_count {} transfer_memory_size 0x{:X}",
               params.sample_rate, params.channel_count, transfer_memory_size);
@@ -323,9 +321,7 @@ void HwOpus::OpenHardwareOpusDecoderForMultiStream(HLERequestContext& ctx) {
 
     auto transfer_memory_size{rp.Pop<u32>()};
     auto transfer_memory_handle{ctx.GetCopyHandle(0)};
-    auto transfer_memory{
-        system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-            transfer_memory_handle)};
+    auto transfer_memory{ctx.GetObjectFromHandle<Kernel::KTransferMemory>(transfer_memory_handle)};
 
     LOG_DEBUG(Service_Audio,
               "sample_rate {} channel_count {} total_stream_count {} stereo_stream_count {} "
@@ -374,9 +370,7 @@ void HwOpus::OpenHardwareOpusDecoderEx(HLERequestContext& ctx) {
     auto params = rp.PopRaw<OpusParametersEx>();
     auto transfer_memory_size{rp.Pop<u32>()};
     auto transfer_memory_handle{ctx.GetCopyHandle(0)};
-    auto transfer_memory{
-        system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-            transfer_memory_handle)};
+    auto transfer_memory{ctx.GetObjectFromHandle<Kernel::KTransferMemory>(transfer_memory_handle)};
 
     LOG_DEBUG(Service_Audio, "sample_rate {} channel_count {} transfer_memory_size 0x{:X}",
               params.sample_rate, params.channel_count, transfer_memory_size);
@@ -414,9 +408,7 @@ void HwOpus::OpenHardwareOpusDecoderForMultiStreamEx(HLERequestContext& ctx) {
 
     auto transfer_memory_size{rp.Pop<u32>()};
     auto transfer_memory_handle{ctx.GetCopyHandle(0)};
-    auto transfer_memory{
-        system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-            transfer_memory_handle)};
+    auto transfer_memory{ctx.GetObjectFromHandle<Kernel::KTransferMemory>(transfer_memory_handle)};
 
     LOG_DEBUG(Service_Audio,
               "sample_rate {} channel_count {} total_stream_count {} stereo_stream_count {} "

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

@@ -1850,8 +1850,7 @@ void IHidServer::InitializeSevenSixAxisSensor(HLERequestContext& ctx) {
     ASSERT_MSG(t_mem_1_size == 0x1000, "t_mem_1_size is not 0x1000 bytes");
     ASSERT_MSG(t_mem_2_size == 0x7F000, "t_mem_2_size is not 0x7F000 bytes");
 
-    auto t_mem_1 = system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-        t_mem_1_handle);
+    auto t_mem_1 = ctx.GetObjectFromHandle<Kernel::KTransferMemory>(t_mem_1_handle);
 
     if (t_mem_1.IsNull()) {
         LOG_ERROR(Service_HID, "t_mem_1 is a nullptr for handle=0x{:08X}", t_mem_1_handle);
@@ -1860,8 +1859,7 @@ void IHidServer::InitializeSevenSixAxisSensor(HLERequestContext& ctx) {
         return;
     }
 
-    auto t_mem_2 = system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-        t_mem_2_handle);
+    auto t_mem_2 = ctx.GetObjectFromHandle<Kernel::KTransferMemory>(t_mem_2_handle);
 
     if (t_mem_2.IsNull()) {
         LOG_ERROR(Service_HID, "t_mem_2 is a nullptr for handle=0x{:08X}", t_mem_2_handle);
@@ -2142,8 +2140,7 @@ void IHidServer::WritePalmaWaveEntry(HLERequestContext& ctx) {
 
     ASSERT_MSG(t_mem_size == 0x3000, "t_mem_size is not 0x3000 bytes");
 
-    auto t_mem = system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-        t_mem_handle);
+    auto t_mem = ctx.GetObjectFromHandle<Kernel::KTransferMemory>(t_mem_handle);
 
     if (t_mem.IsNull()) {
         LOG_ERROR(Service_HID, "t_mem is a nullptr for handle=0x{:08X}", t_mem_handle);

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

@@ -448,8 +448,7 @@ void HidBus::EnableJoyPollingReceiveMode(HLERequestContext& ctx) {
 
     ASSERT_MSG(t_mem_size == 0x1000, "t_mem_size is not 0x1000 bytes");
 
-    auto t_mem = system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-        t_mem_handle);
+    auto t_mem = ctx.GetObjectFromHandle<Kernel::KTransferMemory>(t_mem_handle);
 
     if (t_mem.IsNull()) {
         LOG_ERROR(Service_HID, "t_mem is a nullptr for handle=0x{:08X}", t_mem_handle);

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

@@ -197,8 +197,7 @@ void IRS::RunImageTransferProcessor(HLERequestContext& ctx) {
     const auto parameters{rp.PopRaw<Parameters>()};
     const auto t_mem_handle{ctx.GetCopyHandle(0)};
 
-    auto t_mem = system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-        t_mem_handle);
+    auto t_mem = ctx.GetObjectFromHandle<Kernel::KTransferMemory>(t_mem_handle);
 
     if (t_mem.IsNull()) {
         LOG_ERROR(Service_IRS, "t_mem is a nullptr for handle=0x{:08X}", t_mem_handle);
@@ -444,8 +443,7 @@ void IRS::RunImageTransferExProcessor(HLERequestContext& ctx) {
     const auto parameters{rp.PopRaw<Parameters>()};
     const auto t_mem_handle{ctx.GetCopyHandle(0)};
 
-    auto t_mem = system.ApplicationProcess()->GetHandleTable().GetObject<Kernel::KTransferMemory>(
-        t_mem_handle);
+    auto t_mem = ctx.GetObjectFromHandle<Kernel::KTransferMemory>(t_mem_handle);
 
     LOG_INFO(Service_IRS,
              "called, npad_type={}, npad_id={}, transfer_memory_size={}, "

src/core/hle/service/hle_ipc.cpp

@@ -146,10 +146,7 @@ HLERequestContext::HLERequestContext(Kernel::KernelCore& kernel_, Core::Memory::
 
 HLERequestContext::~HLERequestContext() = default;
 
-void HLERequestContext::ParseCommandBuffer(Kernel::KProcess& process, u32_le* src_cmdbuf,
-                                           bool incoming) {
-    client_handle_table = &process.GetHandleTable();
-
+void HLERequestContext::ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming) {
     IPC::RequestParser rp(src_cmdbuf);
     command_header = rp.PopRaw<IPC::CommandHeader>();
 
@@ -162,7 +159,7 @@ void HLERequestContext::ParseCommandBuffer(Kernel::KProcess& process, u32_le* sr
     if (command_header->enable_handle_descriptor) {
         handle_descriptor_header = rp.PopRaw<IPC::HandleDescriptorHeader>();
         if (handle_descriptor_header->send_current_pid) {
-            pid = process.GetProcessId();
+            pid = thread->GetOwnerProcess()->GetProcessId();
             rp.Skip(2, false);
         }
         if (incoming) {
@@ -270,9 +267,10 @@ void HLERequestContext::ParseCommandBuffer(Kernel::KProcess& process, u32_le* sr
     rp.Skip(1, false); // The command is actually an u64, but we don't use the high part.
 }
 
-Result HLERequestContext::PopulateFromIncomingCommandBuffer(Kernel::KProcess& process,
-                                                            u32_le* src_cmdbuf) {
-    ParseCommandBuffer(process, src_cmdbuf, true);
+Result HLERequestContext::PopulateFromIncomingCommandBuffer(u32_le* src_cmdbuf) {
+    client_handle_table = &thread->GetOwnerProcess()->GetHandleTable();
+
+    ParseCommandBuffer(src_cmdbuf, true);
 
     if (command_header->IsCloseCommand()) {
         // Close does not populate the rest of the IPC header
@@ -284,9 +282,9 @@ Result HLERequestContext::PopulateFromIncomingCommandBuffer(Kernel::KProcess& pr
     return ResultSuccess;
 }
 
-Result HLERequestContext::WriteToOutgoingCommandBuffer(Kernel::KThread& requesting_thread) {
+Result HLERequestContext::WriteToOutgoingCommandBuffer() {
     auto current_offset = handles_offset;
-    auto& owner_process = *requesting_thread.GetOwnerProcess();
+    auto& owner_process = *thread->GetOwnerProcess();
     auto& handle_table = owner_process.GetHandleTable();
 
     for (auto& object : outgoing_copy_objects) {
@@ -319,7 +317,7 @@ Result HLERequestContext::WriteToOutgoingCommandBuffer(Kernel::KThread& requesti
     }
 
     // Copy the translated command buffer back into the thread's command buffer area.
-    memory.WriteBlock(requesting_thread.GetTlsAddress(), cmd_buf.data(), write_size * sizeof(u32));
+    memory.WriteBlock(thread->GetTlsAddress(), cmd_buf.data(), write_size * sizeof(u32));
 
     return ResultSuccess;
 }

src/core/hle/service/hle_ipc.h

@@ -17,6 +17,7 @@
 #include "common/concepts.h"
 #include "common/swap.h"
 #include "core/hle/ipc.h"
+#include "core/hle/kernel/k_handle_table.h"
 #include "core/hle/kernel/svc_common.h"
 
 union Result;
@@ -196,10 +197,10 @@ public:
     }
 
     /// Populates this context with data from the requesting process/thread.
-    Result PopulateFromIncomingCommandBuffer(Kernel::KProcess& process, u32_le* src_cmdbuf);
+    Result PopulateFromIncomingCommandBuffer(u32_le* src_cmdbuf);
 
     /// Writes data from this context back to the requesting process/thread.
-    Result WriteToOutgoingCommandBuffer(Kernel::KThread& requesting_thread);
+    Result WriteToOutgoingCommandBuffer();
 
     [[nodiscard]] u32_le GetHipcCommand() const {
         return command;
@@ -359,8 +360,17 @@ public:
         return *thread;
     }
 
-    Kernel::KHandleTable& GetClientHandleTable() {
-        return *client_handle_table;
+    [[nodiscard]] Core::Memory::Memory& GetMemory() const {
+        return memory;
+    }
+
+    template <typename T>
+    Kernel::KScopedAutoObject<T> GetObjectFromHandle(u32 handle) {
+        auto obj = client_handle_table->GetObjectForIpc(handle, thread);
+        if (obj.IsNotNull()) {
+            return obj->DynamicCast<T*>();
+        }
+        return nullptr;
     }
 
     [[nodiscard]] std::shared_ptr<SessionRequestManager> GetManager() const {
@@ -378,7 +388,7 @@ public:
 private:
     friend class IPC::ResponseBuilder;
 
-    void ParseCommandBuffer(Kernel::KProcess& process, u32_le* src_cmdbuf, bool incoming);
+    void ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming);
 
     std::array<u32, IPC::COMMAND_BUFFER_LENGTH> cmd_buf;
     Kernel::KServerSession* server_session{};

src/core/hle/service/jit/jit.cpp

@@ -26,7 +26,7 @@ public:
     explicit IJitEnvironment(Core::System& system_, Kernel::KProcess& process_, CodeRange user_rx,
                              CodeRange user_ro)
         : ServiceFramework{system_, "IJitEnvironment"}, process{&process_},
-          context{system_.ApplicationMemory()} {
+          context{process->GetMemory()} {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IJitEnvironment::GenerateCode, "GenerateCode"},
@@ -188,7 +188,7 @@ public:
             return;
         }
 
-        auto tmem{process->GetHandleTable().GetObject<Kernel::KTransferMemory>(tmem_handle)};
+        auto tmem{ctx.GetObjectFromHandle<Kernel::KTransferMemory>(tmem_handle)};
         if (tmem.IsNull()) {
             LOG_ERROR(Service_JIT, "attempted to load plugin with invalid transfer memory handle");
             IPC::ResponseBuilder rb{ctx, 2};
@@ -356,11 +356,7 @@ public:
             return;
         }
 
-        // Fetch using the handle table for the application process here,
-        // since we are not multiprocess yet.
-        const auto& handle_table{system.ApplicationProcess()->GetHandleTable()};
-
-        auto process{handle_table.GetObject<Kernel::KProcess>(process_handle)};
+        auto process{ctx.GetObjectFromHandle<Kernel::KProcess>(process_handle)};
         if (process.IsNull()) {
             LOG_ERROR(Service_JIT, "process is null for handle=0x{:08X}", process_handle);
             IPC::ResponseBuilder rb{ctx, 2};
@@ -368,7 +364,7 @@ public:
             return;
         }
 
-        auto rx_mem{handle_table.GetObject<Kernel::KCodeMemory>(rx_mem_handle)};
+        auto rx_mem{ctx.GetObjectFromHandle<Kernel::KCodeMemory>(rx_mem_handle)};
         if (rx_mem.IsNull()) {
             LOG_ERROR(Service_JIT, "rx_mem is null for handle=0x{:08X}", rx_mem_handle);
             IPC::ResponseBuilder rb{ctx, 2};
@@ -376,7 +372,7 @@ public:
             return;
         }
 
-        auto ro_mem{handle_table.GetObject<Kernel::KCodeMemory>(ro_mem_handle)};
+        auto ro_mem{ctx.GetObjectFromHandle<Kernel::KCodeMemory>(ro_mem_handle)};
         if (ro_mem.IsNull()) {
             LOG_ERROR(Service_JIT, "ro_mem is null for handle=0x{:08X}", ro_mem_handle);
             IPC::ResponseBuilder rb{ctx, 2};

src/core/hle/service/ro/ro.cpp

@@ -651,10 +651,9 @@ private:
     void RegisterProcessHandle(HLERequestContext& ctx) {
         LOG_DEBUG(Service_LDR, "(called)");
 
-        auto process_h = ctx.GetClientHandleTable().GetObject(ctx.GetCopyHandle(0));
+        auto process = ctx.GetObjectFromHandle<Kernel::KProcess>(ctx.GetCopyHandle(0));
         auto client_pid = ctx.GetPID();
-        auto result = interface.RegisterProcessHandle(client_pid,
-                                                      process_h->DynamicCast<Kernel::KProcess*>());
+        auto result = interface.RegisterProcessHandle(client_pid, process.GetPointerUnsafe());
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(result);
@@ -671,12 +670,11 @@ private:
 
         IPC::RequestParser rp{ctx};
         auto params = rp.PopRaw<InputParameters>();
-        auto process_h = ctx.GetClientHandleTable().GetObject(ctx.GetCopyHandle(0));
+        auto process = ctx.GetObjectFromHandle<Kernel::KProcess>(ctx.GetCopyHandle(0));
 
         auto client_pid = ctx.GetPID();
-        auto result =
-            interface.RegisterProcessModuleInfo(client_pid, params.nrr_address, params.nrr_size,
-                                                process_h->DynamicCast<Kernel::KProcess*>());
+        auto result = interface.RegisterProcessModuleInfo(
+            client_pid, params.nrr_address, params.nrr_size, process.GetPointerUnsafe());
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(result);

src/core/hle/service/service.cpp

@@ -203,7 +203,7 @@ Result ServiceFrameworkBase::HandleSyncRequest(Kernel::KServerSession& session,
     // If emulation was shutdown, we are closing service threads, do not write the response back to
     // memory that may be shutting down as well.
     if (system.IsPoweredOn()) {
-        ctx.WriteToOutgoingCommandBuffer(ctx.GetThread());
+        ctx.WriteToOutgoingCommandBuffer();
     }
 
     return result;

src/core/memory.cpp

@@ -10,6 +10,7 @@
 #include "common/assert.h"
 #include "common/atomic_ops.h"
 #include "common/common_types.h"
+#include "common/heap_tracker.h"
 #include "common/logging/log.h"
 #include "common/page_table.h"
 #include "common/scope_exit.h"
@@ -52,10 +53,18 @@ struct Memory::Impl {
         } else {
             current_page_table->fastmem_arena = nullptr;
         }
+
+#ifdef __linux__
+        heap_tracker.emplace(system.DeviceMemory().buffer);
+        buffer = std::addressof(*heap_tracker);
+#else
+        buffer = std::addressof(system.DeviceMemory().buffer);
+#endif
     }
 
     void MapMemoryRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size,
-                         Common::PhysicalAddress target, Common::MemoryPermission perms) {
+                         Common::PhysicalAddress target, Common::MemoryPermission perms,
+                         bool separate_heap) {
         ASSERT_MSG((size & YUZU_PAGEMASK) == 0, "non-page aligned size: {:016X}", size);
         ASSERT_MSG((base & YUZU_PAGEMASK) == 0, "non-page aligned base: {:016X}", GetInteger(base));
         ASSERT_MSG(target >= DramMemoryMap::Base, "Out of bounds target: {:016X}",
@@ -64,19 +73,20 @@ struct Memory::Impl {
                  Common::PageType::Memory);
 
         if (current_page_table->fastmem_arena) {
-            system.DeviceMemory().buffer.Map(GetInteger(base),
-                                             GetInteger(target) - DramMemoryMap::Base, size, perms);
+            buffer->Map(GetInteger(base), GetInteger(target) - DramMemoryMap::Base, size, perms,
+                        separate_heap);
         }
     }
 
-    void UnmapRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size) {
+    void UnmapRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size,
+                     bool separate_heap) {
         ASSERT_MSG((size & YUZU_PAGEMASK) == 0, "non-page aligned size: {:016X}", size);
         ASSERT_MSG((base & YUZU_PAGEMASK) == 0, "non-page aligned base: {:016X}", GetInteger(base));
         MapPages(page_table, base / YUZU_PAGESIZE, size / YUZU_PAGESIZE, 0,
                  Common::PageType::Unmapped);
 
         if (current_page_table->fastmem_arena) {
-            system.DeviceMemory().buffer.Unmap(GetInteger(base), size);
+            buffer->Unmap(GetInteger(base), size, separate_heap);
         }
     }
 
@@ -89,11 +99,6 @@ struct Memory::Impl {
             return;
         }
 
-        const bool is_r = True(perms & Common::MemoryPermission::Read);
-        const bool is_w = True(perms & Common::MemoryPermission::Write);
-        const bool is_x =
-            True(perms & Common::MemoryPermission::Execute) && Settings::IsNceEnabled();
-
         u64 protect_bytes{};
         u64 protect_begin{};
         for (u64 addr = vaddr; addr < vaddr + size; addr += YUZU_PAGESIZE) {
@@ -102,8 +107,7 @@ struct Memory::Impl {
             switch (page_type) {
             case Common::PageType::RasterizerCachedMemory:
                 if (protect_bytes > 0) {
-                    system.DeviceMemory().buffer.Protect(protect_begin, protect_bytes, is_r, is_w,
-                                                         is_x);
+                    buffer->Protect(protect_begin, protect_bytes, perms);
                     protect_bytes = 0;
                 }
                 break;
@@ -116,7 +120,7 @@ struct Memory::Impl {
         }
 
         if (protect_bytes > 0) {
-            system.DeviceMemory().buffer.Protect(protect_begin, protect_bytes, is_r, is_w, is_x);
+            buffer->Protect(protect_begin, protect_bytes, perms);
         }
     }
 
@@ -486,7 +490,9 @@ struct Memory::Impl {
         }
 
         if (current_page_table->fastmem_arena) {
-            system.DeviceMemory().buffer.Protect(vaddr, size, !debug, !debug);
+            const auto perm{debug ? Common::MemoryPermission{}
+                                  : Common::MemoryPermission::ReadWrite};
+            buffer->Protect(vaddr, size, perm);
         }
 
         // Iterate over a contiguous CPU address space, marking/unmarking the region.
@@ -543,9 +549,14 @@ struct Memory::Impl {
         }
 
         if (current_page_table->fastmem_arena) {
-            const bool is_read_enable =
-                !Settings::values.use_reactive_flushing.GetValue() || !cached;
-            system.DeviceMemory().buffer.Protect(vaddr, size, is_read_enable, !cached);
+            Common::MemoryPermission perm{};
+            if (!Settings::values.use_reactive_flushing.GetValue() || !cached) {
+                perm |= Common::MemoryPermission::Read;
+            }
+            if (!cached) {
+                perm |= Common::MemoryPermission::Write;
+            }
+            buffer->Protect(vaddr, size, perm);
         }
 
         // Iterate over a contiguous CPU address space, which corresponds to the specified GPU
@@ -856,6 +867,13 @@ struct Memory::Impl {
     std::array<GPUDirtyState, Core::Hardware::NUM_CPU_CORES> rasterizer_write_areas{};
     std::span<Core::GPUDirtyMemoryManager> gpu_dirty_managers;
     std::mutex sys_core_guard;
+
+    std::optional<Common::HeapTracker> heap_tracker;
+#ifdef __linux__
+    Common::HeapTracker* buffer{};
+#else
+    Common::HostMemory* buffer{};
+#endif
 };
 
 Memory::Memory(Core::System& system_) : system{system_} {
@@ -873,12 +891,14 @@ void Memory::SetCurrentPageTable(Kernel::KProcess& process) {
 }
 
 void Memory::MapMemoryRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size,
-                             Common::PhysicalAddress target, Common::MemoryPermission perms) {
-    impl->MapMemoryRegion(page_table, base, size, target, perms);
+                             Common::PhysicalAddress target, Common::MemoryPermission perms,
+                             bool separate_heap) {
+    impl->MapMemoryRegion(page_table, base, size, target, perms, separate_heap);
 }
 
-void Memory::UnmapRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size) {
-    impl->UnmapRegion(page_table, base, size);
+void Memory::UnmapRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size,
+                         bool separate_heap) {
+    impl->UnmapRegion(page_table, base, size, separate_heap);
 }
 
 void Memory::ProtectRegion(Common::PageTable& page_table, Common::ProcessAddress vaddr, u64 size,
@@ -1048,7 +1068,9 @@ void Memory::FlushRegion(Common::ProcessAddress dest_addr, size_t size) {
 }
 
 bool Memory::InvalidateNCE(Common::ProcessAddress vaddr, size_t size) {
-    bool mapped = true;
+    [[maybe_unused]] bool mapped = true;
+    [[maybe_unused]] bool rasterizer = false;
+
     u8* const ptr = impl->GetPointerImpl(
         GetInteger(vaddr),
         [&] {
@@ -1056,8 +1078,26 @@ bool Memory::InvalidateNCE(Common::ProcessAddress vaddr, size_t size) {
                       GetInteger(vaddr));
             mapped = false;
         },
-        [&] { impl->system.GPU().InvalidateRegion(GetInteger(vaddr), size); });
+        [&] {
+            impl->system.GPU().InvalidateRegion(GetInteger(vaddr), size);
+            rasterizer = true;
+        });
+
+#ifdef __linux__
+    if (!rasterizer && mapped) {
+        impl->buffer->DeferredMapSeparateHeap(GetInteger(vaddr));
+    }
+#endif
+
     return mapped && ptr != nullptr;
 }
 
+bool Memory::InvalidateSeparateHeap(void* fault_address) {
+#ifdef __linux__
+    return impl->buffer->DeferredMapSeparateHeap(static_cast<u8*>(fault_address));
+#else
+    return false;
+#endif
+}
+
 } // namespace Core::Memory

src/core/memory.h

@@ -86,7 +86,8 @@ public:
      * @param perms      The permissions to map the memory with.
      */
     void MapMemoryRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size,
-                         Common::PhysicalAddress target, Common::MemoryPermission perms);
+                         Common::PhysicalAddress target, Common::MemoryPermission perms,
+                         bool separate_heap);
 
     /**
      * Unmaps a region of the emulated process address space.
@@ -95,7 +96,8 @@ public:
      * @param base       The address to begin unmapping at.
      * @param size       The amount of bytes to unmap.
      */
-    void UnmapRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size);
+    void UnmapRegion(Common::PageTable& page_table, Common::ProcessAddress base, u64 size,
+                     bool separate_heap);
 
     /**
      * Protects a region of the emulated process address space with the new permissions.
@@ -486,6 +488,7 @@ public:
     void SetGPUDirtyManagers(std::span<Core::GPUDirtyMemoryManager> managers);
     void InvalidateRegion(Common::ProcessAddress dest_addr, size_t size);
     bool InvalidateNCE(Common::ProcessAddress vaddr, size_t size);
+    bool InvalidateSeparateHeap(void* fault_address);
     void FlushRegion(Common::ProcessAddress dest_addr, size_t size);
 
 private:

src/shader_recompiler/backend/spirv/spirv_emit_context.cpp

@@ -96,9 +96,9 @@ Id ImageType(EmitContext& ctx, const ImageDescriptor& desc, Id sampled_type) {
 }
 
 Id DefineVariable(EmitContext& ctx, Id type, std::optional<spv::BuiltIn> builtin,
-                  spv::StorageClass storage_class) {
+                  spv::StorageClass storage_class, std::optional<Id> initializer = std::nullopt) {
     const Id pointer_type{ctx.TypePointer(storage_class, type)};
-    const Id id{ctx.AddGlobalVariable(pointer_type, storage_class)};
+    const Id id{ctx.AddGlobalVariable(pointer_type, storage_class, initializer)};
     if (builtin) {
         ctx.Decorate(id, spv::Decoration::BuiltIn, *builtin);
     }
@@ -144,11 +144,12 @@ Id DefineInput(EmitContext& ctx, Id type, bool per_invocation,
 }
 
 Id DefineOutput(EmitContext& ctx, Id type, std::optional<u32> invocations,
-                std::optional<spv::BuiltIn> builtin = std::nullopt) {
+                std::optional<spv::BuiltIn> builtin = std::nullopt,
+                std::optional<Id> initializer = std::nullopt) {
     if (invocations && ctx.stage == Stage::TessellationControl) {
         type = ctx.TypeArray(type, ctx.Const(*invocations));
     }
-    return DefineVariable(ctx, type, builtin, spv::StorageClass::Output);
+    return DefineVariable(ctx, type, builtin, spv::StorageClass::Output, initializer);
 }
 
 void DefineGenericOutput(EmitContext& ctx, size_t index, std::optional<u32> invocations) {
@@ -811,10 +812,14 @@ void EmitContext::DefineAttributeMemAccess(const Info& info) {
             labels.push_back(OpLabel());
         }
         if (info.stores.ClipDistances()) {
-            literals.push_back(static_cast<u32>(IR::Attribute::ClipDistance0) >> 2);
-            labels.push_back(OpLabel());
-            literals.push_back(static_cast<u32>(IR::Attribute::ClipDistance4) >> 2);
-            labels.push_back(OpLabel());
+            if (profile.max_user_clip_distances >= 4) {
+                literals.push_back(static_cast<u32>(IR::Attribute::ClipDistance0) >> 2);
+                labels.push_back(OpLabel());
+            }
+            if (profile.max_user_clip_distances >= 8) {
+                literals.push_back(static_cast<u32>(IR::Attribute::ClipDistance4) >> 2);
+                labels.push_back(OpLabel());
+            }
         }
         OpSelectionMerge(end_block, spv::SelectionControlMask::MaskNone);
         OpSwitch(compare_index, default_label, literals, labels);
@@ -843,17 +848,21 @@ void EmitContext::DefineAttributeMemAccess(const Info& info) {
             ++label_index;
         }
         if (info.stores.ClipDistances()) {
-            AddLabel(labels[label_index]);
-            const Id pointer{OpAccessChain(output_f32, clip_distances, masked_index)};
-            OpStore(pointer, store_value);
-            OpReturn();
-            ++label_index;
-            AddLabel(labels[label_index]);
-            const Id fixed_index{OpIAdd(U32[1], masked_index, Const(4U))};
-            const Id pointer2{OpAccessChain(output_f32, clip_distances, fixed_index)};
-            OpStore(pointer2, store_value);
-            OpReturn();
-            ++label_index;
+            if (profile.max_user_clip_distances >= 4) {
+                AddLabel(labels[label_index]);
+                const Id pointer{OpAccessChain(output_f32, clip_distances, masked_index)};
+                OpStore(pointer, store_value);
+                OpReturn();
+                ++label_index;
+            }
+            if (profile.max_user_clip_distances >= 8) {
+                AddLabel(labels[label_index]);
+                const Id fixed_index{OpIAdd(U32[1], masked_index, Const(4U))};
+                const Id pointer{OpAccessChain(output_f32, clip_distances, fixed_index)};
+                OpStore(pointer, store_value);
+                OpReturn();
+                ++label_index;
+            }
         }
         AddLabel(end_block);
         OpUnreachable();
@@ -1532,9 +1541,16 @@ void EmitContext::DefineOutputs(const IR::Program& program) {
         if (stage == Stage::Fragment) {
             throw NotImplementedException("Storing ClipDistance in fragment stage");
         }
-        const Id type{TypeArray(
-            F32[1], Const(std::min(info.used_clip_distances, profile.max_user_clip_distances)))};
-        clip_distances = DefineOutput(*this, type, invocations, spv::BuiltIn::ClipDistance);
+        if (profile.max_user_clip_distances > 0) {
+            const u32 used{std::min(profile.max_user_clip_distances, 8u)};
+            const std::array<Id, 8> zero{f32_zero_value, f32_zero_value, f32_zero_value,
+                                         f32_zero_value, f32_zero_value, f32_zero_value,
+                                         f32_zero_value, f32_zero_value};
+            const Id type{TypeArray(F32[1], Const(used))};
+            const Id initializer{ConstantComposite(type, std::span(zero).subspan(0, used))};
+            clip_distances =
+                DefineOutput(*this, type, invocations, spv::BuiltIn::ClipDistance, initializer);
+        }
     }
     if (info.stores[IR::Attribute::Layer] &&
         (profile.support_viewport_index_layer_non_geometry || stage == Stage::Geometry)) {

src/tests/common/host_memory.cpp

@@ -12,6 +12,7 @@ using namespace Common::Literals;
 static constexpr size_t VIRTUAL_SIZE = 1ULL << 39;
 static constexpr size_t BACKING_SIZE = 4_GiB;
 static constexpr auto PERMS = Common::MemoryPermission::ReadWrite;
+static constexpr auto HEAP = false;
 
 TEST_CASE("HostMemory: Initialize and deinitialize", "[common]") {
     { HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE); }
@@ -20,7 +21,7 @@ TEST_CASE("HostMemory: Initialize and deinitialize", "[common]") {
 
 TEST_CASE("HostMemory: Simple map", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x5000, 0x8000, 0x1000, PERMS);
+    mem.Map(0x5000, 0x8000, 0x1000, PERMS, HEAP);
 
     volatile u8* const data = mem.VirtualBasePointer() + 0x5000;
     data[0] = 50;
@@ -29,8 +30,8 @@ TEST_CASE("HostMemory: Simple map", "[common]") {
 
 TEST_CASE("HostMemory: Simple mirror map", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x5000, 0x3000, 0x2000, PERMS);
-    mem.Map(0x8000, 0x4000, 0x1000, PERMS);
+    mem.Map(0x5000, 0x3000, 0x2000, PERMS, HEAP);
+    mem.Map(0x8000, 0x4000, 0x1000, PERMS, HEAP);
 
     volatile u8* const mirror_a = mem.VirtualBasePointer() + 0x5000;
     volatile u8* const mirror_b = mem.VirtualBasePointer() + 0x8000;
@@ -40,116 +41,116 @@ TEST_CASE("HostMemory: Simple mirror map", "[common]") {
 
 TEST_CASE("HostMemory: Simple unmap", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x5000, 0x3000, 0x2000, PERMS);
+    mem.Map(0x5000, 0x3000, 0x2000, PERMS, HEAP);
 
     volatile u8* const data = mem.VirtualBasePointer() + 0x5000;
     data[75] = 50;
     REQUIRE(data[75] == 50);
 
-    mem.Unmap(0x5000, 0x2000);
+    mem.Unmap(0x5000, 0x2000, HEAP);
 }
 
 TEST_CASE("HostMemory: Simple unmap and remap", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x5000, 0x3000, 0x2000, PERMS);
+    mem.Map(0x5000, 0x3000, 0x2000, PERMS, HEAP);
 
     volatile u8* const data = mem.VirtualBasePointer() + 0x5000;
     data[0] = 50;
     REQUIRE(data[0] == 50);
 
-    mem.Unmap(0x5000, 0x2000);
+    mem.Unmap(0x5000, 0x2000, HEAP);
 
-    mem.Map(0x5000, 0x3000, 0x2000, PERMS);
+    mem.Map(0x5000, 0x3000, 0x2000, PERMS, HEAP);
     REQUIRE(data[0] == 50);
 
-    mem.Map(0x7000, 0x2000, 0x5000, PERMS);
+    mem.Map(0x7000, 0x2000, 0x5000, PERMS, HEAP);
     REQUIRE(data[0x3000] == 50);
 }
 
 TEST_CASE("HostMemory: Nieche allocation", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x0000, 0, 0x20000, PERMS);
-    mem.Unmap(0x0000, 0x4000);
-    mem.Map(0x1000, 0, 0x2000, PERMS);
-    mem.Map(0x3000, 0, 0x1000, PERMS);
-    mem.Map(0, 0, 0x1000, PERMS);
+    mem.Map(0x0000, 0, 0x20000, PERMS, HEAP);
+    mem.Unmap(0x0000, 0x4000, HEAP);
+    mem.Map(0x1000, 0, 0x2000, PERMS, HEAP);
+    mem.Map(0x3000, 0, 0x1000, PERMS, HEAP);
+    mem.Map(0, 0, 0x1000, PERMS, HEAP);
 }
 
 TEST_CASE("HostMemory: Full unmap", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x8000, 0, 0x4000, PERMS);
-    mem.Unmap(0x8000, 0x4000);
-    mem.Map(0x6000, 0, 0x16000, PERMS);
+    mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
+    mem.Unmap(0x8000, 0x4000, HEAP);
+    mem.Map(0x6000, 0, 0x16000, PERMS, HEAP);
 }
 
 TEST_CASE("HostMemory: Right out of bounds unmap", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x0000, 0, 0x4000, PERMS);
-    mem.Unmap(0x2000, 0x4000);
-    mem.Map(0x2000, 0x80000, 0x4000, PERMS);
+    mem.Map(0x0000, 0, 0x4000, PERMS, HEAP);
+    mem.Unmap(0x2000, 0x4000, HEAP);
+    mem.Map(0x2000, 0x80000, 0x4000, PERMS, HEAP);
 }
 
 TEST_CASE("HostMemory: Left out of bounds unmap", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x8000, 0, 0x4000, PERMS);
-    mem.Unmap(0x6000, 0x4000);
-    mem.Map(0x8000, 0, 0x2000, PERMS);
+    mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
+    mem.Unmap(0x6000, 0x4000, HEAP);
+    mem.Map(0x8000, 0, 0x2000, PERMS, HEAP);
 }
 
 TEST_CASE("HostMemory: Multiple placeholder unmap", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x0000, 0, 0x4000, PERMS);
-    mem.Map(0x4000, 0, 0x1b000, PERMS);
-    mem.Unmap(0x3000, 0x1c000);
-    mem.Map(0x3000, 0, 0x20000, PERMS);
+    mem.Map(0x0000, 0, 0x4000, PERMS, HEAP);
+    mem.Map(0x4000, 0, 0x1b000, PERMS, HEAP);
+    mem.Unmap(0x3000, 0x1c000, HEAP);
+    mem.Map(0x3000, 0, 0x20000, PERMS, HEAP);
 }
 
 TEST_CASE("HostMemory: Unmap between placeholders", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x0000, 0, 0x4000, PERMS);
-    mem.Map(0x4000, 0, 0x4000, PERMS);
-    mem.Unmap(0x2000, 0x4000);
-    mem.Map(0x2000, 0, 0x4000, PERMS);
+    mem.Map(0x0000, 0, 0x4000, PERMS, HEAP);
+    mem.Map(0x4000, 0, 0x4000, PERMS, HEAP);
+    mem.Unmap(0x2000, 0x4000, HEAP);
+    mem.Map(0x2000, 0, 0x4000, PERMS, HEAP);
 }
 
 TEST_CASE("HostMemory: Unmap to origin", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x4000, 0, 0x4000, PERMS);
-    mem.Map(0x8000, 0, 0x4000, PERMS);
-    mem.Unmap(0x4000, 0x4000);
-    mem.Map(0, 0, 0x4000, PERMS);
-    mem.Map(0x4000, 0, 0x4000, PERMS);
+    mem.Map(0x4000, 0, 0x4000, PERMS, HEAP);
+    mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
+    mem.Unmap(0x4000, 0x4000, HEAP);
+    mem.Map(0, 0, 0x4000, PERMS, HEAP);
+    mem.Map(0x4000, 0, 0x4000, PERMS, HEAP);
 }
 
 TEST_CASE("HostMemory: Unmap to right", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x4000, 0, 0x4000, PERMS);
-    mem.Map(0x8000, 0, 0x4000, PERMS);
-    mem.Unmap(0x8000, 0x4000);
-    mem.Map(0x8000, 0, 0x4000, PERMS);
+    mem.Map(0x4000, 0, 0x4000, PERMS, HEAP);
+    mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
+    mem.Unmap(0x8000, 0x4000, HEAP);
+    mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
 }
 
 TEST_CASE("HostMemory: Partial right unmap check bindings", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x4000, 0x10000, 0x4000, PERMS);
+    mem.Map(0x4000, 0x10000, 0x4000, PERMS, HEAP);
 
     volatile u8* const ptr = mem.VirtualBasePointer() + 0x4000;
     ptr[0x1000] = 17;
 
-    mem.Unmap(0x6000, 0x2000);
+    mem.Unmap(0x6000, 0x2000, HEAP);
 
     REQUIRE(ptr[0x1000] == 17);
 }
 
 TEST_CASE("HostMemory: Partial left unmap check bindings", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x4000, 0x10000, 0x4000, PERMS);
+    mem.Map(0x4000, 0x10000, 0x4000, PERMS, HEAP);
 
     volatile u8* const ptr = mem.VirtualBasePointer() + 0x4000;
     ptr[0x3000] = 19;
     ptr[0x3fff] = 12;
 
-    mem.Unmap(0x4000, 0x2000);
+    mem.Unmap(0x4000, 0x2000, HEAP);
 
     REQUIRE(ptr[0x3000] == 19);
     REQUIRE(ptr[0x3fff] == 12);
@@ -157,13 +158,13 @@ TEST_CASE("HostMemory: Partial left unmap check bindings", "[common]") {
 
 TEST_CASE("HostMemory: Partial middle unmap check bindings", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x4000, 0x10000, 0x4000, PERMS);
+    mem.Map(0x4000, 0x10000, 0x4000, PERMS, HEAP);
 
     volatile u8* const ptr = mem.VirtualBasePointer() + 0x4000;
     ptr[0x0000] = 19;
     ptr[0x3fff] = 12;
 
-    mem.Unmap(0x1000, 0x2000);
+    mem.Unmap(0x1000, 0x2000, HEAP);
 
     REQUIRE(ptr[0x0000] == 19);
     REQUIRE(ptr[0x3fff] == 12);
@@ -171,14 +172,14 @@ TEST_CASE("HostMemory: Partial middle unmap check bindings", "[common]") {
 
 TEST_CASE("HostMemory: Partial sparse middle unmap and check bindings", "[common]") {
     HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
-    mem.Map(0x4000, 0x10000, 0x2000, PERMS);
-    mem.Map(0x6000, 0x20000, 0x2000, PERMS);
+    mem.Map(0x4000, 0x10000, 0x2000, PERMS, HEAP);
+    mem.Map(0x6000, 0x20000, 0x2000, PERMS, HEAP);
 
     volatile u8* const ptr = mem.VirtualBasePointer() + 0x4000;
     ptr[0x0000] = 19;
     ptr[0x3fff] = 12;
 
-    mem.Unmap(0x5000, 0x2000);
+    mem.Unmap(0x5000, 0x2000, HEAP);
 
     REQUIRE(ptr[0x0000] == 19);
     REQUIRE(ptr[0x3fff] == 12);

src/video_core/macro/macro_hle.cpp

@@ -327,12 +327,13 @@ public:
     explicit HLE_DrawIndirectByteCount(Maxwell3D& maxwell3d_) : HLEMacroImpl(maxwell3d_) {}
 
     void Execute(const std::vector<u32>& parameters, [[maybe_unused]] u32 method) override {
+        const bool force = maxwell3d.Rasterizer().HasDrawTransformFeedback();
+
         auto topology = static_cast<Maxwell3D::Regs::PrimitiveTopology>(parameters[0] & 0xFFFFU);
-        if (!maxwell3d.AnyParametersDirty() || !IsTopologySafe(topology)) {
+        if (!force && (!maxwell3d.AnyParametersDirty() || !IsTopologySafe(topology))) {
             Fallback(parameters);
             return;
         }
-
         auto& params = maxwell3d.draw_manager->GetIndirectParams();
         params.is_byte_count = true;
         params.is_indexed = false;
@@ -503,6 +504,8 @@ public:
         maxwell3d.CallMethod(static_cast<size_t>(MAXWELL3D_REG_INDEX(launch_dma)), 0x1011, true);
         maxwell3d.CallMethod(static_cast<size_t>(MAXWELL3D_REG_INDEX(inline_data)),
                              regs.transform_feedback.controls[0].stride, true);
+
+        maxwell3d.Rasterizer().RegisterTransformFeedback(regs.upload.dest.Address());
     }
 };
 

src/video_core/rasterizer_interface.h

@@ -173,5 +173,13 @@ public:
     virtual void BindChannel(Tegra::Control::ChannelState& channel) {}
 
     virtual void ReleaseChannel(s32 channel_id) {}
+
+    /// Register the address as a Transform Feedback Object
+    virtual void RegisterTransformFeedback(GPUVAddr tfb_object_addr) {}
+
+    /// Returns true when the rasterizer has Draw Transform Feedback capabilities
+    virtual bool HasDrawTransformFeedback() {
+        return false;
+    }
 };
 } // namespace VideoCore

src/video_core/renderer_opengl/gl_buffer_cache.cpp

@@ -376,4 +376,15 @@ void BufferCacheRuntime::BindImageBuffer(Buffer& buffer, u32 offset, u32 size, P
     *image_handles++ = buffer.View(offset, size, format);
 }
 
+void BufferCacheRuntime::BindTransformFeedbackObject(GPUVAddr tfb_object_addr) {
+    OGLTransformFeedback& tfb_object = tfb_objects[tfb_object_addr];
+    tfb_object.Create();
+    glBindTransformFeedback(GL_TRANSFORM_FEEDBACK, tfb_object.handle);
+}
+
+GLuint BufferCacheRuntime::GetTransformFeedbackObject(GPUVAddr tfb_object_addr) {
+    ASSERT(tfb_objects.contains(tfb_object_addr));
+    return tfb_objects[tfb_object_addr].handle;
+}
+
 } // namespace OpenGL

src/video_core/renderer_opengl/gl_buffer_cache.h

@@ -5,6 +5,7 @@
 
 #include <array>
 #include <span>
+#include <unordered_map>
 
 #include "common/common_types.h"
 #include "video_core/buffer_cache/buffer_cache_base.h"
@@ -121,6 +122,9 @@ public:
     void BindImageBuffer(Buffer& buffer, u32 offset, u32 size,
                          VideoCore::Surface::PixelFormat format);
 
+    void BindTransformFeedbackObject(GPUVAddr tfb_object_addr);
+    GLuint GetTransformFeedbackObject(GPUVAddr tfb_object_addr);
+
     u64 GetDeviceMemoryUsage() const;
 
     void BindFastUniformBuffer(size_t stage, u32 binding_index, u32 size) {
@@ -233,6 +237,7 @@ private:
     u32 index_buffer_offset = 0;
 
     u64 device_access_memory;
+    std::unordered_map<GPUVAddr, OGLTransformFeedback> tfb_objects;
 };
 
 struct BufferCacheParams {

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -309,6 +309,13 @@ void RasterizerOpenGL::DrawIndirect() {
     const auto& params = maxwell3d->draw_manager->GetIndirectParams();
     buffer_cache.SetDrawIndirect(&params);
     PrepareDraw(params.is_indexed, [this, &params](GLenum primitive_mode) {
+        if (params.is_byte_count) {
+            const GPUVAddr tfb_object_base_addr = params.indirect_start_address - 4U;
+            const GLuint tfb_object =
+                buffer_cache_runtime.GetTransformFeedbackObject(tfb_object_base_addr);
+            glDrawTransformFeedback(primitive_mode, tfb_object);
+            return;
+        }
         const auto [buffer, offset] = buffer_cache.GetDrawIndirectBuffer();
         const GLvoid* const gl_offset =
             reinterpret_cast<const GLvoid*>(static_cast<uintptr_t>(offset));
@@ -1371,6 +1378,10 @@ void RasterizerOpenGL::ReleaseChannel(s32 channel_id) {
     query_cache.EraseChannel(channel_id);
 }
 
+void RasterizerOpenGL::RegisterTransformFeedback(GPUVAddr tfb_object_addr) {
+    buffer_cache_runtime.BindTransformFeedbackObject(tfb_object_addr);
+}
+
 AccelerateDMA::AccelerateDMA(BufferCache& buffer_cache_, TextureCache& texture_cache_)
     : buffer_cache{buffer_cache_}, texture_cache{texture_cache_} {}
 

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -139,6 +139,12 @@ public:
 
     void ReleaseChannel(s32 channel_id) override;
 
+    void RegisterTransformFeedback(GPUVAddr tfb_object_addr) override;
+
+    bool HasDrawTransformFeedback() override {
+        return true;
+    }
+
 private:
     static constexpr size_t MAX_TEXTURES = 192;
     static constexpr size_t MAX_IMAGES = 48;

src/video_core/renderer_opengl/gl_resource_manager.cpp

@@ -207,4 +207,21 @@ void OGLQuery::Release() {
     handle = 0;
 }
 
+void OGLTransformFeedback::Create() {
+    if (handle != 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    glCreateTransformFeedbacks(1, &handle);
+}
+
+void OGLTransformFeedback::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteTransformFeedbacks(1, &handle);
+    handle = 0;
+}
+
 } // namespace OpenGL

src/video_core/renderer_opengl/gl_resource_manager.h

@@ -323,4 +323,31 @@ public:
     GLuint handle = 0;
 };
 
+class OGLTransformFeedback final {
+public:
+    YUZU_NON_COPYABLE(OGLTransformFeedback);
+
+    OGLTransformFeedback() = default;
+
+    OGLTransformFeedback(OGLTransformFeedback&& o) noexcept : handle(std::exchange(o.handle, 0)) {}
+
+    ~OGLTransformFeedback() {
+        Release();
+    }
+
+    OGLTransformFeedback& operator=(OGLTransformFeedback&& o) noexcept {
+        Release();
+        handle = std::exchange(o.handle, 0);
+        return *this;
+    }
+
+    /// Creates a new internal OpenGL resource and stores the handle
+    void Create();
+
+    /// Deletes the internal OpenGL resource
+    void Release();
+
+    GLuint handle = 0;
+};
+
 } // namespace OpenGL

src/video_core/texture_cache/decode_bc.cpp

@@ -60,66 +60,72 @@ u32 ConvertedBytesPerBlock(VideoCore::Surface::PixelFormat pixel_format) {
 }
 
 template <auto decompress, PixelFormat pixel_format>
-void DecompressBlocks(std::span<const u8> input, std::span<u8> output, Extent3D extent,
+void DecompressBlocks(std::span<const u8> input, std::span<u8> output, BufferImageCopy& copy,
                       bool is_signed = false) {
     const u32 out_bpp = ConvertedBytesPerBlock(pixel_format);
-    const u32 block_width = std::min(extent.width, BLOCK_SIZE);
-    const u32 block_height = std::min(extent.height, BLOCK_SIZE);
-    const u32 pitch = extent.width * out_bpp;
+    const u32 block_size = BlockSize(pixel_format);
+    const u32 width = copy.image_extent.width;
+    const u32 height = copy.image_extent.height * copy.image_subresource.num_layers;
+    const u32 depth = copy.image_extent.depth;
+    const u32 block_width = std::min(width, BLOCK_SIZE);
+    const u32 block_height = std::min(height, BLOCK_SIZE);
+    const u32 pitch = width * out_bpp;
     size_t input_offset = 0;
     size_t output_offset = 0;
-    for (u32 slice = 0; slice < extent.depth; ++slice) {
-        for (u32 y = 0; y < extent.height; y += block_height) {
-            size_t row_offset = 0;
-            for (u32 x = 0; x < extent.width;
-                 x += block_width, row_offset += block_width * out_bpp) {
-                const u8* src = input.data() + input_offset;
-                u8* const dst = output.data() + output_offset + row_offset;
+    for (u32 slice = 0; slice < depth; ++slice) {
+        for (u32 y = 0; y < height; y += block_height) {
+            size_t src_offset = input_offset;
+            size_t dst_offset = output_offset;
+            for (u32 x = 0; x < width; x += block_width) {
+                const u8* src = input.data() + src_offset;
+                u8* const dst = output.data() + dst_offset;
                 if constexpr (IsSigned(pixel_format)) {
-                    decompress(src, dst, x, y, extent.width, extent.height, is_signed);
+                    decompress(src, dst, x, y, width, height, is_signed);
                 } else {
-                    decompress(src, dst, x, y, extent.width, extent.height);
+                    decompress(src, dst, x, y, width, height);
                 }
-                input_offset += BlockSize(pixel_format);
+                src_offset += block_size;
+                dst_offset += block_width * out_bpp;
             }
+            input_offset += copy.buffer_row_length * block_size / block_width;
             output_offset += block_height * pitch;
         }
     }
 }
 
-void DecompressBCn(std::span<const u8> input, std::span<u8> output, Extent3D extent,
+void DecompressBCn(std::span<const u8> input, std::span<u8> output, BufferImageCopy& copy,
                    VideoCore::Surface::PixelFormat pixel_format) {
     switch (pixel_format) {
     case PixelFormat::BC1_RGBA_UNORM:
     case PixelFormat::BC1_RGBA_SRGB:
-        DecompressBlocks<bcn::DecodeBc1, PixelFormat::BC1_RGBA_UNORM>(input, output, extent);
+        DecompressBlocks<bcn::DecodeBc1, PixelFormat::BC1_RGBA_UNORM>(input, output, copy);
         break;
     case PixelFormat::BC2_UNORM:
     case PixelFormat::BC2_SRGB:
-        DecompressBlocks<bcn::DecodeBc2, PixelFormat::BC2_UNORM>(input, output, extent);
+        DecompressBlocks<bcn::DecodeBc2, PixelFormat::BC2_UNORM>(input, output, copy);
         break;
     case PixelFormat::BC3_UNORM:
     case PixelFormat::BC3_SRGB:
-        DecompressBlocks<bcn::DecodeBc3, PixelFormat::BC3_UNORM>(input, output, extent);
+        DecompressBlocks<bcn::DecodeBc3, PixelFormat::BC3_UNORM>(input, output, copy);
         break;
     case PixelFormat::BC4_SNORM:
     case PixelFormat::BC4_UNORM:
         DecompressBlocks<bcn::DecodeBc4, PixelFormat::BC4_UNORM>(
-            input, output, extent, pixel_format == PixelFormat::BC4_SNORM);
+            input, output, copy, pixel_format == PixelFormat::BC4_SNORM);
         break;
     case PixelFormat::BC5_SNORM:
     case PixelFormat::BC5_UNORM:
         DecompressBlocks<bcn::DecodeBc5, PixelFormat::BC5_UNORM>(
-            input, output, extent, pixel_format == PixelFormat::BC5_SNORM);
+            input, output, copy, pixel_format == PixelFormat::BC5_SNORM);
         break;
     case PixelFormat::BC6H_SFLOAT:
     case PixelFormat::BC6H_UFLOAT:
         DecompressBlocks<bcn::DecodeBc6, PixelFormat::BC6H_UFLOAT>(
-            input, output, extent, pixel_format == PixelFormat::BC6H_SFLOAT);
+            input, output, copy, pixel_format == PixelFormat::BC6H_SFLOAT);
         break;
     case PixelFormat::BC7_SRGB:
     case PixelFormat::BC7_UNORM:
-        DecompressBlocks<bcn::DecodeBc7, PixelFormat::BC7_UNORM>(input, output, extent);
+        DecompressBlocks<bcn::DecodeBc7, PixelFormat::BC7_UNORM>(input, output, copy);
         break;
     default:
         LOG_WARNING(HW_GPU, "Unimplemented BCn decompression {}", pixel_format);

src/video_core/texture_cache/decode_bc.h

@@ -13,7 +13,7 @@ namespace VideoCommon {
 
 [[nodiscard]] u32 ConvertedBytesPerBlock(VideoCore::Surface::PixelFormat pixel_format);
 
-void DecompressBCn(std::span<const u8> input, std::span<u8> output, Extent3D extent,
+void DecompressBCn(std::span<const u8> input, std::span<u8> output, BufferImageCopy& copy,
                    VideoCore::Surface::PixelFormat pixel_format);
 
 } // namespace VideoCommon

src/video_core/texture_cache/util.cpp

@@ -837,6 +837,7 @@ boost::container::small_vector<BufferImageCopy, 16> UnswizzleImage(Tegra::Memory
                                                                    std::span<u8> output) {
     const size_t guest_size_bytes = input.size_bytes();
     const u32 bpp_log2 = BytesPerBlockLog2(info.format);
+    const Extent2D tile_size = DefaultBlockSize(info.format);
     const Extent3D size = info.size;
 
     if (info.type == ImageType::Linear) {
@@ -847,7 +848,7 @@ boost::container::small_vector<BufferImageCopy, 16> UnswizzleImage(Tegra::Memory
         return {{
             .buffer_offset = 0,
             .buffer_size = guest_size_bytes,
-            .buffer_row_length = info.pitch >> bpp_log2,
+            .buffer_row_length = info.pitch * tile_size.width >> bpp_log2,
             .buffer_image_height = size.height,
             .image_subresource =
                 {
@@ -862,7 +863,6 @@ boost::container::small_vector<BufferImageCopy, 16> UnswizzleImage(Tegra::Memory
     const LevelInfo level_info = MakeLevelInfo(info);
     const s32 num_layers = info.resources.layers;
     const s32 num_levels = info.resources.levels;
-    const Extent2D tile_size = DefaultBlockSize(info.format);
     const std::array level_sizes = CalculateLevelSizes(level_info, num_levels);
     const Extent2D gob = GobSize(bpp_log2, info.block.height, info.tile_width_spacing);
     const u32 layer_size = CalculateLevelBytes(level_sizes, num_levels);
@@ -926,8 +926,6 @@ void ConvertImage(std::span<const u8> input, const ImageInfo& info, std::span<u8
 
         const auto input_offset = input.subspan(copy.buffer_offset);
         copy.buffer_offset = output_offset;
-        copy.buffer_row_length = mip_size.width;
-        copy.buffer_image_height = mip_size.height;
 
         const auto recompression_setting = Settings::values.astc_recompression.GetValue();
         const bool astc = IsPixelFormatASTC(info.format);
@@ -972,16 +970,14 @@ void ConvertImage(std::span<const u8> input, const ImageInfo& info, std::span<u8
                 bpp_div;
             output_offset += static_cast<u32>(copy.buffer_size);
         } else {
-            const Extent3D image_extent{
-                .width = copy.image_extent.width,
-                .height = copy.image_extent.height * copy.image_subresource.num_layers,
-                .depth = copy.image_extent.depth,
-            };
-            DecompressBCn(input_offset, output.subspan(output_offset), image_extent, info.format);
+            DecompressBCn(input_offset, output.subspan(output_offset), copy, info.format);
             output_offset += copy.image_extent.width * copy.image_extent.height *
                              copy.image_subresource.num_layers *
                              ConvertedBytesPerBlock(info.format);
         }
+
+        copy.buffer_row_length = mip_size.width;
+        copy.buffer_image_height = mip_size.height;
     }
 }
 

src/video_core/vulkan_common/vulkan_device.cpp

@@ -755,10 +755,10 @@ VkFormat Device::GetSupportedFormat(VkFormat wanted_format, VkFormatFeatureFlags
     // The wanted format is not supported by hardware, search for alternatives
     const VkFormat* alternatives = GetFormatAlternatives(wanted_format);
     if (alternatives == nullptr) {
-        ASSERT_MSG(false,
-                   "Format={} with usage={} and type={} has no defined alternatives and host "
-                   "hardware does not support it",
-                   wanted_format, wanted_usage, format_type);
+        LOG_ERROR(Render_Vulkan,
+                  "Format={} with usage={} and type={} has no defined alternatives and host "
+                  "hardware does not support it",
+                  wanted_format, wanted_usage, format_type);
         return wanted_format;
     }
 
@@ -774,10 +774,10 @@ VkFormat Device::GetSupportedFormat(VkFormat wanted_format, VkFormatFeatureFlags
     }
 
     // No alternatives found, panic
-    ASSERT_MSG(false,
-               "Format={} with usage={} and type={} is not supported by the host hardware and "
-               "doesn't support any of the alternatives",
-               wanted_format, wanted_usage, format_type);
+    LOG_ERROR(Render_Vulkan,
+              "Format={} with usage={} and type={} is not supported by the host hardware and "
+              "doesn't support any of the alternatives",
+              wanted_format, wanted_usage, format_type);
     return wanted_format;
 }
 

src/video_core/vulkan_common/vulkan_wrapper.cpp

@@ -246,7 +246,9 @@ void SetObjectName(const DeviceDispatch* dld, VkDevice device, T handle, VkObjec
         .objectHandle = reinterpret_cast<u64>(handle),
         .pObjectName = name,
     };
-    Check(dld->vkSetDebugUtilsObjectNameEXT(device, &name_info));
+    if (dld->vkSetDebugUtilsObjectNameEXT) {
+        Check(dld->vkSetDebugUtilsObjectNameEXT(device, &name_info));
+    }
 }
 
 } // Anonymous namespace

src/yuzu/main.h

@@ -168,14 +168,6 @@ class GMainWindow : public QMainWindow {
     /// Max number of recently loaded items to keep track of
     static const int max_recent_files_item = 10;
 
-    // TODO: Make use of this!
-    enum {
-        UI_IDLE,
-        UI_EMU_BOOTING,
-        UI_EMU_RUNNING,
-        UI_EMU_STOPPING,
-    };
-
     enum {
         CREATE_SHORTCUT_MSGBOX_FULLSCREEN_YES,
         CREATE_SHORTCUT_MSGBOX_SUCCESS,