Android #175

README.md

@@ -1,3 +1,15 @@
+| Pull Request | Commit | Title | Author | Merged? |
+|----|----|----|----|----|
+| [12454](https://github.com/yuzu-emu/yuzu//pull/12454) | [`3a4e7d45f`](https://github.com/yuzu-emu/yuzu//pull/12454/files) |  core_timing: minor refactors | [liamwhite](https://github.com/liamwhite/) | Yes |
+| [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/audio_core/device/device_session.cpp

@@ -18,9 +18,7 @@ constexpr auto INCREMENT_TIME{5ms};
 DeviceSession::DeviceSession(Core::System& system_)
     : system{system_}, thread_event{Core::Timing::CreateEvent(
                            "AudioOutSampleTick",
-                           [this](std::uintptr_t, s64 time, std::chrono::nanoseconds) {
-                               return ThreadFunc();
-                           })} {}
+                           [this](s64 time, std::chrono::nanoseconds) { return ThreadFunc(); })} {}
 
 DeviceSession::~DeviceSession() {
     Finalize();

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/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/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/core_timing.cpp

@@ -29,7 +29,6 @@ std::shared_ptr<EventType> CreateEvent(std::string name, TimedCallback&& callbac
 struct CoreTiming::Event {
     s64 time;
     u64 fifo_order;
-    std::uintptr_t user_data;
     std::weak_ptr<EventType> type;
     s64 reschedule_time;
     heap_t::handle_type handle{};
@@ -67,17 +66,15 @@ void CoreTiming::Initialize(std::function<void()>&& on_thread_init_) {
     event_fifo_id = 0;
     shutting_down = false;
     cpu_ticks = 0;
-    const auto empty_timed_callback = [](std::uintptr_t, u64, std::chrono::nanoseconds)
-        -> std::optional<std::chrono::nanoseconds> { return std::nullopt; };
-    ev_lost = CreateEvent("_lost_event", empty_timed_callback);
     if (is_multicore) {
         timer_thread = std::make_unique<std::jthread>(ThreadEntry, std::ref(*this));
     }
 }
 
 void CoreTiming::ClearPendingEvents() {
-    std::scoped_lock lock{basic_lock};
+    std::scoped_lock lock{advance_lock, basic_lock};
     event_queue.clear();
+    event.Set();
 }
 
 void CoreTiming::Pause(bool is_paused) {
@@ -119,14 +116,12 @@ bool CoreTiming::HasPendingEvents() const {
 }
 
 void CoreTiming::ScheduleEvent(std::chrono::nanoseconds ns_into_future,
-                               const std::shared_ptr<EventType>& event_type,
-                               std::uintptr_t user_data, bool absolute_time) {
+                               const std::shared_ptr<EventType>& event_type, bool absolute_time) {
     {
         std::scoped_lock scope{basic_lock};
         const auto next_time{absolute_time ? ns_into_future : GetGlobalTimeNs() + ns_into_future};
 
-        auto h{event_queue.emplace(
-            Event{next_time.count(), event_fifo_id++, user_data, event_type, 0})};
+        auto h{event_queue.emplace(Event{next_time.count(), event_fifo_id++, event_type, 0})};
         (*h).handle = h;
     }
 
@@ -136,13 +131,13 @@ void CoreTiming::ScheduleEvent(std::chrono::nanoseconds ns_into_future,
 void CoreTiming::ScheduleLoopingEvent(std::chrono::nanoseconds start_time,
                                       std::chrono::nanoseconds resched_time,
                                       const std::shared_ptr<EventType>& event_type,
-                                      std::uintptr_t user_data, bool absolute_time) {
+                                      bool absolute_time) {
     {
         std::scoped_lock scope{basic_lock};
         const auto next_time{absolute_time ? start_time : GetGlobalTimeNs() + start_time};
 
-        auto h{event_queue.emplace(Event{next_time.count(), event_fifo_id++, user_data, event_type,
-                                         resched_time.count()})};
+        auto h{event_queue.emplace(
+            Event{next_time.count(), event_fifo_id++, event_type, resched_time.count()})};
         (*h).handle = h;
     }
 
@@ -150,14 +145,14 @@ void CoreTiming::ScheduleLoopingEvent(std::chrono::nanoseconds start_time,
 }
 
 void CoreTiming::UnscheduleEvent(const std::shared_ptr<EventType>& event_type,
-                                 std::uintptr_t user_data, bool wait) {
+                                 UnscheduleEventType type) {
     {
         std::scoped_lock lk{basic_lock};
 
         std::vector<heap_t::handle_type> to_remove;
         for (auto itr = event_queue.begin(); itr != event_queue.end(); itr++) {
             const Event& e = *itr;
-            if (e.type.lock().get() == event_type.get() && e.user_data == user_data) {
+            if (e.type.lock().get() == event_type.get()) {
                 to_remove.push_back(itr->handle);
             }
         }
@@ -165,10 +160,12 @@ void CoreTiming::UnscheduleEvent(const std::shared_ptr<EventType>& event_type,
         for (auto h : to_remove) {
             event_queue.erase(h);
         }
+
+        event_type->sequence_number++;
     }
 
     // Force any in-progress events to finish
-    if (wait) {
+    if (type == UnscheduleEventType::Wait) {
         std::scoped_lock lk{advance_lock};
     }
 }
@@ -208,28 +205,31 @@ std::optional<s64> CoreTiming::Advance() {
         const Event& evt = event_queue.top();
 
         if (const auto event_type{evt.type.lock()}) {
-            if (evt.reschedule_time == 0) {
-                const auto evt_user_data = evt.user_data;
-                const auto evt_time = evt.time;
+            const auto evt_time = evt.time;
+            const auto evt_sequence_num = event_type->sequence_number;
 
+            if (evt.reschedule_time == 0) {
                 event_queue.pop();
 
                 basic_lock.unlock();
 
                 event_type->callback(
-                    evt_user_data, evt_time,
-                    std::chrono::nanoseconds{GetGlobalTimeNs().count() - evt_time});
+                    evt_time, std::chrono::nanoseconds{GetGlobalTimeNs().count() - evt_time});
 
                 basic_lock.lock();
             } else {
                 basic_lock.unlock();
 
                 const auto new_schedule_time{event_type->callback(
-                    evt.user_data, evt.time,
-                    std::chrono::nanoseconds{GetGlobalTimeNs().count() - evt.time})};
+                    evt_time, std::chrono::nanoseconds{GetGlobalTimeNs().count() - evt_time})};
 
                 basic_lock.lock();
 
+                if (evt_sequence_num != event_type->sequence_number) {
+                    // Heap handle is invalidated after external modification.
+                    continue;
+                }
+
                 const auto next_schedule_time{new_schedule_time.has_value()
                                                   ? new_schedule_time.value().count()
                                                   : evt.reschedule_time};
@@ -241,8 +241,8 @@ std::optional<s64> CoreTiming::Advance() {
                     next_time = pause_end_time + next_schedule_time;
                 }
 
-                event_queue.update(evt.handle, Event{next_time, event_fifo_id++, evt.user_data,
-                                                     evt.type, next_schedule_time, evt.handle});
+                event_queue.update(evt.handle, Event{next_time, event_fifo_id++, evt.type,
+                                                     next_schedule_time, evt.handle});
             }
         }
 

src/core/core_timing.h

@@ -22,17 +22,25 @@ namespace Core::Timing {
 
 /// A callback that may be scheduled for a particular core timing event.
 using TimedCallback = std::function<std::optional<std::chrono::nanoseconds>(
-    std::uintptr_t user_data, s64 time, std::chrono::nanoseconds ns_late)>;
+    s64 time, std::chrono::nanoseconds ns_late)>;
 
 /// Contains the characteristics of a particular event.
 struct EventType {
     explicit EventType(TimedCallback&& callback_, std::string&& name_)
-        : callback{std::move(callback_)}, name{std::move(name_)} {}
+        : callback{std::move(callback_)}, name{std::move(name_)}, sequence_number{0} {}
 
     /// The event's callback function.
     TimedCallback callback;
     /// A pointer to the name of the event.
     const std::string name;
+    /// A monotonic sequence number, incremented when this event is
+    /// changed externally.
+    size_t sequence_number;
+};
+
+enum class UnscheduleEventType {
+    Wait,
+    NoWait,
 };
 
 /**
@@ -89,23 +97,17 @@ public:
 
     /// Schedules an event in core timing
     void ScheduleEvent(std::chrono::nanoseconds ns_into_future,
-                       const std::shared_ptr<EventType>& event_type, std::uintptr_t user_data = 0,
-                       bool absolute_time = false);
+                       const std::shared_ptr<EventType>& event_type, bool absolute_time = false);
 
     /// Schedules an event which will automatically re-schedule itself with the given time, until
     /// unscheduled
     void ScheduleLoopingEvent(std::chrono::nanoseconds start_time,
                               std::chrono::nanoseconds resched_time,
                               const std::shared_ptr<EventType>& event_type,
-                              std::uintptr_t user_data = 0, bool absolute_time = false);
+                              bool absolute_time = false);
 
-    void UnscheduleEvent(const std::shared_ptr<EventType>& event_type, std::uintptr_t user_data,
-                         bool wait = true);
-
-    void UnscheduleEventWithoutWait(const std::shared_ptr<EventType>& event_type,
-                                    std::uintptr_t user_data) {
-        UnscheduleEvent(event_type, user_data, false);
-    }
+    void UnscheduleEvent(const std::shared_ptr<EventType>& event_type,
+                         UnscheduleEventType type = UnscheduleEventType::Wait);
 
     void AddTicks(u64 ticks_to_add);
 
@@ -158,7 +160,6 @@ private:
     heap_t event_queue;
     u64 event_fifo_id = 0;
 
-    std::shared_ptr<EventType> ev_lost;
     Common::Event event{};
     Common::Event pause_event{};
     mutable std::mutex basic_lock;

src/core/hle/kernel/k_hardware_timer.cpp

@@ -10,15 +10,15 @@ namespace Kernel {
 
 void KHardwareTimer::Initialize() {
     // Create the timing callback to register with CoreTiming.
-    m_event_type = Core::Timing::CreateEvent(
-        "KHardwareTimer::Callback", [](std::uintptr_t timer_handle, s64, std::chrono::nanoseconds) {
-            reinterpret_cast<KHardwareTimer*>(timer_handle)->DoTask();
-            return std::nullopt;
-        });
+    m_event_type = Core::Timing::CreateEvent("KHardwareTimer::Callback",
+                                             [this](s64, std::chrono::nanoseconds) {
+                                                 this->DoTask();
+                                                 return std::nullopt;
+                                             });
 }
 
 void KHardwareTimer::Finalize() {
-    m_kernel.System().CoreTiming().UnscheduleEvent(m_event_type, reinterpret_cast<uintptr_t>(this));
+    m_kernel.System().CoreTiming().UnscheduleEvent(m_event_type);
     m_wakeup_time = std::numeric_limits<s64>::max();
     m_event_type.reset();
 }
@@ -57,13 +57,12 @@ void KHardwareTimer::EnableInterrupt(s64 wakeup_time) {
 
     m_wakeup_time = wakeup_time;
     m_kernel.System().CoreTiming().ScheduleEvent(std::chrono::nanoseconds{m_wakeup_time},
-                                                 m_event_type, reinterpret_cast<uintptr_t>(this),
-                                                 true);
+                                                 m_event_type, true);
 }
 
 void KHardwareTimer::DisableInterrupt() {
-    m_kernel.System().CoreTiming().UnscheduleEventWithoutWait(m_event_type,
-                                                              reinterpret_cast<uintptr_t>(this));
+    m_kernel.System().CoreTiming().UnscheduleEvent(m_event_type,
+                                                   Core::Timing::UnscheduleEventType::NoWait);
     m_wakeup_time = std::numeric_limits<s64>::max();
 }
 

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/kernel.cpp

@@ -238,7 +238,7 @@ struct KernelCore::Impl {
     void InitializePreemption(KernelCore& kernel) {
         preemption_event = Core::Timing::CreateEvent(
             "PreemptionCallback",
-            [this, &kernel](std::uintptr_t, s64 time,
+            [this, &kernel](s64 time,
                             std::chrono::nanoseconds) -> std::optional<std::chrono::nanoseconds> {
                 {
                     KScopedSchedulerLock lock(kernel);

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

@@ -49,10 +49,10 @@ HidBus::HidBus(Core::System& system_)
     // Register update callbacks
     hidbus_update_event = Core::Timing::CreateEvent(
         "Hidbus::UpdateCallback",
-        [this](std::uintptr_t user_data, s64 time,
+        [this](s64 time,
                std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
             const auto guard = LockService();
-            UpdateHidbus(user_data, ns_late);
+            UpdateHidbus(ns_late);
             return std::nullopt;
         });
 
@@ -61,10 +61,10 @@ HidBus::HidBus(Core::System& system_)
 }
 
 HidBus::~HidBus() {
-    system.CoreTiming().UnscheduleEvent(hidbus_update_event, 0);
+    system.CoreTiming().UnscheduleEvent(hidbus_update_event);
 }
 
-void HidBus::UpdateHidbus(std::uintptr_t user_data, std::chrono::nanoseconds ns_late) {
+void HidBus::UpdateHidbus(std::chrono::nanoseconds ns_late) {
     if (is_hidbus_enabled) {
         for (std::size_t i = 0; i < devices.size(); ++i) {
             if (!devices[i].is_device_initializated) {

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

@@ -108,7 +108,7 @@ private:
     void DisableJoyPollingReceiveMode(HLERequestContext& ctx);
     void SetStatusManagerType(HLERequestContext& ctx);
 
-    void UpdateHidbus(std::uintptr_t user_data, std::chrono::nanoseconds ns_late);
+    void UpdateHidbus(std::chrono::nanoseconds ns_late);
     std::optional<std::size_t> GetDeviceIndexFromHandle(BusHandle handle) const;
 
     template <typename T>

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

@@ -227,8 +227,7 @@ void ResourceManager::EnableTouchScreen(u64 aruid, bool is_enabled) {
     applet_resource->EnableTouchScreen(aruid, is_enabled);
 }
 
-void ResourceManager::UpdateControllers(std::uintptr_t user_data,
-                                        std::chrono::nanoseconds ns_late) {
+void ResourceManager::UpdateControllers(std::chrono::nanoseconds ns_late) {
     auto& core_timing = system.CoreTiming();
     debug_pad->OnUpdate(core_timing);
     digitizer->OnUpdate(core_timing);
@@ -241,20 +240,19 @@ void ResourceManager::UpdateControllers(std::uintptr_t user_data,
     capture_button->OnUpdate(core_timing);
 }
 
-void ResourceManager::UpdateNpad(std::uintptr_t user_data, std::chrono::nanoseconds ns_late) {
+void ResourceManager::UpdateNpad(std::chrono::nanoseconds ns_late) {
     auto& core_timing = system.CoreTiming();
     npad->OnUpdate(core_timing);
 }
 
-void ResourceManager::UpdateMouseKeyboard(std::uintptr_t user_data,
-                                          std::chrono::nanoseconds ns_late) {
+void ResourceManager::UpdateMouseKeyboard(std::chrono::nanoseconds ns_late) {
     auto& core_timing = system.CoreTiming();
     mouse->OnUpdate(core_timing);
     debug_mouse->OnUpdate(core_timing);
     keyboard->OnUpdate(core_timing);
 }
 
-void ResourceManager::UpdateMotion(std::uintptr_t user_data, std::chrono::nanoseconds ns_late) {
+void ResourceManager::UpdateMotion(std::chrono::nanoseconds ns_late) {
     auto& core_timing = system.CoreTiming();
     six_axis->OnUpdate(core_timing);
     seven_six_axis->OnUpdate(core_timing);
@@ -273,34 +271,34 @@ IAppletResource::IAppletResource(Core::System& system_, std::shared_ptr<Resource
     // Register update callbacks
     npad_update_event = Core::Timing::CreateEvent(
         "HID::UpdatePadCallback",
-        [this, resource](std::uintptr_t user_data, s64 time, std::chrono::nanoseconds ns_late)
-            -> std::optional<std::chrono::nanoseconds> {
+        [this, resource](
+            s64 time, std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
             const auto guard = LockService();
-            resource->UpdateNpad(user_data, ns_late);
+            resource->UpdateNpad(ns_late);
             return std::nullopt;
         });
     default_update_event = Core::Timing::CreateEvent(
         "HID::UpdateDefaultCallback",
-        [this, resource](std::uintptr_t user_data, s64 time, std::chrono::nanoseconds ns_late)
-            -> std::optional<std::chrono::nanoseconds> {
+        [this, resource](
+            s64 time, std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
             const auto guard = LockService();
-            resource->UpdateControllers(user_data, ns_late);
+            resource->UpdateControllers(ns_late);
             return std::nullopt;
         });
     mouse_keyboard_update_event = Core::Timing::CreateEvent(
         "HID::UpdateMouseKeyboardCallback",
-        [this, resource](std::uintptr_t user_data, s64 time, std::chrono::nanoseconds ns_late)
-            -> std::optional<std::chrono::nanoseconds> {
+        [this, resource](
+            s64 time, std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
             const auto guard = LockService();
-            resource->UpdateMouseKeyboard(user_data, ns_late);
+            resource->UpdateMouseKeyboard(ns_late);
             return std::nullopt;
         });
     motion_update_event = Core::Timing::CreateEvent(
         "HID::UpdateMotionCallback",
-        [this, resource](std::uintptr_t user_data, s64 time, std::chrono::nanoseconds ns_late)
-            -> std::optional<std::chrono::nanoseconds> {
+        [this, resource](
+            s64 time, std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
             const auto guard = LockService();
-            resource->UpdateMotion(user_data, ns_late);
+            resource->UpdateMotion(ns_late);
             return std::nullopt;
         });
 
@@ -314,10 +312,10 @@ IAppletResource::IAppletResource(Core::System& system_, std::shared_ptr<Resource
 }
 
 IAppletResource::~IAppletResource() {
-    system.CoreTiming().UnscheduleEvent(npad_update_event, 0);
-    system.CoreTiming().UnscheduleEvent(default_update_event, 0);
-    system.CoreTiming().UnscheduleEvent(mouse_keyboard_update_event, 0);
-    system.CoreTiming().UnscheduleEvent(motion_update_event, 0);
+    system.CoreTiming().UnscheduleEvent(npad_update_event);
+    system.CoreTiming().UnscheduleEvent(default_update_event);
+    system.CoreTiming().UnscheduleEvent(mouse_keyboard_update_event);
+    system.CoreTiming().UnscheduleEvent(motion_update_event);
     resource_manager->FreeAppletResourceId(aruid);
 }
 

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

@@ -81,10 +81,10 @@ public:
     void EnablePadInput(u64 aruid, bool is_enabled);
     void EnableTouchScreen(u64 aruid, bool is_enabled);
 
-    void UpdateControllers(std::uintptr_t user_data, std::chrono::nanoseconds ns_late);
-    void UpdateNpad(std::uintptr_t user_data, std::chrono::nanoseconds ns_late);
-    void UpdateMouseKeyboard(std::uintptr_t user_data, std::chrono::nanoseconds ns_late);
-    void UpdateMotion(std::uintptr_t user_data, std::chrono::nanoseconds ns_late);
+    void UpdateControllers(std::chrono::nanoseconds ns_late);
+    void UpdateNpad(std::chrono::nanoseconds ns_late);
+    void UpdateMouseKeyboard(std::chrono::nanoseconds ns_late);
+    void UpdateMotion(std::chrono::nanoseconds ns_late);
 
 private:
     Result CreateAppletResourceImpl(u64 aruid);

src/core/hle/service/nvnflinger/nvnflinger.cpp

@@ -67,7 +67,7 @@ Nvnflinger::Nvnflinger(Core::System& system_, HosBinderDriverServer& hos_binder_
     // Schedule the screen composition events
     multi_composition_event = Core::Timing::CreateEvent(
         "ScreenComposition",
-        [this](std::uintptr_t, s64 time,
+        [this](s64 time,
                std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
             vsync_signal.Set();
             return std::chrono::nanoseconds(GetNextTicks());
@@ -75,7 +75,7 @@ Nvnflinger::Nvnflinger(Core::System& system_, HosBinderDriverServer& hos_binder_
 
     single_composition_event = Core::Timing::CreateEvent(
         "ScreenComposition",
-        [this](std::uintptr_t, s64 time,
+        [this](s64 time,
                std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
             const auto lock_guard = Lock();
             Compose();
@@ -93,11 +93,11 @@ Nvnflinger::Nvnflinger(Core::System& system_, HosBinderDriverServer& hos_binder_
 
 Nvnflinger::~Nvnflinger() {
     if (system.IsMulticore()) {
-        system.CoreTiming().UnscheduleEvent(multi_composition_event, {});
+        system.CoreTiming().UnscheduleEvent(multi_composition_event);
         vsync_thread.request_stop();
         vsync_signal.Set();
     } else {
-        system.CoreTiming().UnscheduleEvent(single_composition_event, {});
+        system.CoreTiming().UnscheduleEvent(single_composition_event);
     }
 
     ShutdownLayers();

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/core/memory/cheat_engine.cpp

@@ -190,15 +190,15 @@ CheatEngine::CheatEngine(System& system_, std::vector<CheatEntry> cheats_,
 }
 
 CheatEngine::~CheatEngine() {
-    core_timing.UnscheduleEvent(event, 0);
+    core_timing.UnscheduleEvent(event);
 }
 
 void CheatEngine::Initialize() {
     event = Core::Timing::CreateEvent(
         "CheatEngine::FrameCallback::" + Common::HexToString(metadata.main_nso_build_id),
-        [this](std::uintptr_t user_data, s64 time,
+        [this](s64 time,
                std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
-            FrameCallback(user_data, ns_late);
+            FrameCallback(ns_late);
             return std::nullopt;
         });
     core_timing.ScheduleLoopingEvent(CHEAT_ENGINE_NS, CHEAT_ENGINE_NS, event);
@@ -239,7 +239,7 @@ void CheatEngine::Reload(std::vector<CheatEntry> reload_cheats) {
 
 MICROPROFILE_DEFINE(Cheat_Engine, "Add-Ons", "Cheat Engine", MP_RGB(70, 200, 70));
 
-void CheatEngine::FrameCallback(std::uintptr_t, std::chrono::nanoseconds ns_late) {
+void CheatEngine::FrameCallback(std::chrono::nanoseconds ns_late) {
     if (is_pending_reload.exchange(false)) {
         vm.LoadProgram(cheats);
     }

src/core/memory/cheat_engine.h

@@ -70,7 +70,7 @@ public:
     void Reload(std::vector<CheatEntry> reload_cheats);
 
 private:
-    void FrameCallback(std::uintptr_t user_data, std::chrono::nanoseconds ns_late);
+    void FrameCallback(std::chrono::nanoseconds ns_late);
 
     DmntCheatVm vm;
     CheatProcessMetadata metadata;

src/core/tools/freezer.cpp

@@ -51,18 +51,17 @@ void MemoryWriteWidth(Core::Memory::Memory& memory, u32 width, VAddr addr, u64 v
 
 Freezer::Freezer(Core::Timing::CoreTiming& core_timing_, Core::Memory::Memory& memory_)
     : core_timing{core_timing_}, memory{memory_} {
-    event = Core::Timing::CreateEvent(
-        "MemoryFreezer::FrameCallback",
-        [this](std::uintptr_t user_data, s64 time,
-               std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
-            FrameCallback(user_data, ns_late);
-            return std::nullopt;
-        });
+    event = Core::Timing::CreateEvent("MemoryFreezer::FrameCallback",
+                                      [this](s64 time, std::chrono::nanoseconds ns_late)
+                                          -> std::optional<std::chrono::nanoseconds> {
+                                          FrameCallback(ns_late);
+                                          return std::nullopt;
+                                      });
     core_timing.ScheduleEvent(memory_freezer_ns, event);
 }
 
 Freezer::~Freezer() {
-    core_timing.UnscheduleEvent(event, 0);
+    core_timing.UnscheduleEvent(event);
 }
 
 void Freezer::SetActive(bool is_active) {
@@ -159,7 +158,7 @@ Freezer::Entries::const_iterator Freezer::FindEntry(VAddr address) const {
                         [address](const Entry& entry) { return entry.address == address; });
 }
 
-void Freezer::FrameCallback(std::uintptr_t, std::chrono::nanoseconds ns_late) {
+void Freezer::FrameCallback(std::chrono::nanoseconds ns_late) {
     if (!IsActive()) {
         LOG_DEBUG(Common_Memory, "Memory freezer has been deactivated, ending callback events.");
         return;

src/core/tools/freezer.h

@@ -77,7 +77,7 @@ private:
     Entries::iterator FindEntry(VAddr address);
     Entries::const_iterator FindEntry(VAddr address) const;
 
-    void FrameCallback(std::uintptr_t user_data, std::chrono::nanoseconds ns_late);
+    void FrameCallback(std::chrono::nanoseconds ns_late);
     void FillEntryReads();
 
     std::atomic_bool active{false};

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/tests/core/core_timing.cpp

@@ -16,20 +16,16 @@
 
 namespace {
 // Numbers are chosen randomly to make sure the correct one is given.
-constexpr std::array<u64, 5> CB_IDS{{42, 144, 93, 1026, UINT64_C(0xFFFF7FFFF7FFFF)}};
 constexpr std::array<u64, 5> calls_order{{2, 0, 1, 4, 3}};
 std::array<s64, 5> delays{};
-
-std::bitset<CB_IDS.size()> callbacks_ran_flags;
+std::bitset<5> callbacks_ran_flags;
 u64 expected_callback = 0;
 
 template <unsigned int IDX>
-std::optional<std::chrono::nanoseconds> HostCallbackTemplate(std::uintptr_t user_data, s64 time,
+std::optional<std::chrono::nanoseconds> HostCallbackTemplate(s64 time,
                                                              std::chrono::nanoseconds ns_late) {
-    static_assert(IDX < CB_IDS.size(), "IDX out of range");
+    static_assert(IDX < callbacks_ran_flags.size(), "IDX out of range");
     callbacks_ran_flags.set(IDX);
-    REQUIRE(CB_IDS[IDX] == user_data);
-    REQUIRE(CB_IDS[IDX] == CB_IDS[calls_order[expected_callback]]);
     delays[IDX] = ns_late.count();
     ++expected_callback;
     return std::nullopt;
@@ -76,7 +72,7 @@ TEST_CASE("CoreTiming[BasicOrder]", "[core]") {
         const u64 order = calls_order[i];
         const auto future_ns = std::chrono::nanoseconds{static_cast<s64>(i * one_micro + 100)};
 
-        core_timing.ScheduleEvent(future_ns, events[order], CB_IDS[order]);
+        core_timing.ScheduleEvent(future_ns, events[order]);
     }
     /// test pause
     REQUIRE(callbacks_ran_flags.none());
@@ -118,7 +114,7 @@ TEST_CASE("CoreTiming[BasicOrderNoPausing]", "[core]") {
     for (std::size_t i = 0; i < events.size(); i++) {
         const u64 order = calls_order[i];
         const auto future_ns = std::chrono::nanoseconds{static_cast<s64>(i * one_micro + 100)};
-        core_timing.ScheduleEvent(future_ns, events[order], CB_IDS[order]);
+        core_timing.ScheduleEvent(future_ns, events[order]);
     }
 
     const u64 end = core_timing.GetGlobalTimeNs().count();

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;
     }
 }