yuzu: Remember last selected directory

* Testing LTO (#4)

* Testing LTO

* clang

* linux

* Added the rest of Blinkhawk's optimizations

* Unlikely asserts

* Removing LTO from Linux builds - GCC

* Removing LTO from Linux builds - Clang

.ci/scripts/clang/docker.sh

@@ -11,6 +11,7 @@ ccache -s
 mkdir build || true && cd build
 cmake .. \
       -DCMAKE_BUILD_TYPE=Release \
+      -DCMAKE_CXX_FLAGS="-march=x86-64-v2" \
       -DCMAKE_CXX_COMPILER=/usr/lib/ccache/clang++ \
       -DCMAKE_C_COMPILER=/usr/lib/ccache/clang \
       -DCMAKE_INSTALL_PREFIX="/usr" \

.ci/scripts/linux/docker.sh

@@ -12,6 +12,7 @@ mkdir build || true && cd build
 cmake .. \
       -DBoost_USE_STATIC_LIBS=ON \
       -DCMAKE_BUILD_TYPE=Release \
+      -DCMAKE_CXX_FLAGS="-march=x86-64-v2" \
       -DCMAKE_CXX_COMPILER=/usr/lib/ccache/g++ \
       -DCMAKE_C_COMPILER=/usr/lib/ccache/gcc \
       -DCMAKE_INSTALL_PREFIX="/usr" \

.ci/templates/build-msvc.yml

@@ -9,7 +9,7 @@ parameters:
 steps:
 - script: choco install vulkan-sdk
   displayName: 'Install vulkan-sdk'
-- script: refreshenv && mkdir build && cd build && cmake -G "Visual Studio 17 2022" -A x64 -DYUZU_USE_BUNDLED_QT=1 -DYUZU_USE_BUNDLED_SDL2=1 -DYUZU_USE_QT_WEB_ENGINE=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${COMPAT} -DYUZU_TESTS=OFF -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON -DDISPLAY_VERSION=${{ parameters['version'] }} -DCMAKE_BUILD_TYPE=Release -DYUZU_CRASH_DUMPS=ON .. && cd ..
+- script: refreshenv && mkdir build && cd build && cmake -E env CXXFLAGS="/Gw /GA /Gr /Ob2" cmake -G "Visual Studio 17 2022" -A x64 -DCMAKE_INTERPROCEDURAL_OPTIMIZATION=ON -DCMAKE_POLICY_DEFAULT_CMP0069=NEW -DYUZU_USE_BUNDLED_QT=1 -DYUZU_USE_BUNDLED_SDL2=1 -DYUZU_USE_QT_WEB_ENGINE=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${COMPAT} -DYUZU_TESTS=OFF -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON -DDISPLAY_VERSION=${{ parameters['version'] }} -DCMAKE_BUILD_TYPE=Release -DYUZU_CRASH_DUMPS=ON .. && cd ..
   displayName: 'Configure CMake'
 - task: MSBuild@1
   displayName: 'Build'

CMakeLists.txt

@@ -22,6 +22,8 @@ CMAKE_DEPENDENT_OPTION(YUZU_USE_BUNDLED_SDL2 "Download bundled SDL2 binaries" ON
 # On Linux system SDL2 is likely to be lacking HIDAPI support which have drawbacks but is needed for SDL motion
 CMAKE_DEPENDENT_OPTION(YUZU_USE_EXTERNAL_SDL2 "Compile external SDL2" ON "ENABLE_SDL2;NOT MSVC" OFF)
 
+option(ENABLE_OPENGL "Enable OpenGL" ON)
+mark_as_advanced(FORCE ENABLE_OPENGL)
 option(ENABLE_QT "Enable the Qt frontend" ON)
 option(ENABLE_QT6 "Allow usage of Qt6 to be attempted" OFF)
 set(QT6_LOCATION "" CACHE PATH "Additional Location to search for Qt6 libraries like C:/Qt/6.3.1/msvc2019_64/")
@@ -51,6 +53,8 @@ option(YUZU_USE_BUNDLED_VCPKG "Use vcpkg for yuzu dependencies" "${MSVC}")
 
 option(YUZU_CHECK_SUBMODULES "Check if submodules are present" ON)
 
+CMAKE_DEPENDENT_OPTION(YUZU_USE_FASTER_LD "Check if a faster linker is available" ON "NOT WIN32" OFF)
+
 if (YUZU_USE_BUNDLED_VCPKG)
     if (YUZU_TESTS)
         list(APPEND VCPKG_MANIFEST_FEATURES "yuzu-tests")
@@ -579,6 +583,21 @@ if (MSVC AND CMAKE_GENERATOR STREQUAL "Ninja")
     )
 endif()
 
+if (YUZU_USE_FASTER_LD AND CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
+    # We will assume that if the compiler is GCC, it will attempt to use ld.bfd by default.
+    # Try to pick a faster linker.
+    find_program(LLD lld)
+    find_program(MOLD mold)
+
+    if (MOLD AND CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL "12.1")
+        message(NOTICE "Selecting mold as linker")
+        add_link_options("-fuse-ld=mold")
+    elseif (LLD)
+        message(NOTICE "Selecting lld as linker")
+        add_link_options("-fuse-ld=lld")
+    endif()
+endif()
+
 enable_testing()
 add_subdirectory(externals)
 add_subdirectory(src)

externals/find-modules/FindOpus.cmake

@@ -2,9 +2,7 @@
 # SPDX-License-Identifier: GPL-2.0-or-later
 
 find_package(PkgConfig QUIET)
-if (PKG_CONFIG_FOUND)
-    pkg_search_module(OPUS QUIET IMPORTED_TARGET opus)
-endif()
+pkg_search_module(OPUS QUIET IMPORTED_TARGET opus)
 
 include(FindPackageHandleStandardArgs)
 find_package_handle_standard_args(Opus

externals/find-modules/Findenet.cmake

@@ -3,9 +3,7 @@
 # SPDX-License-Identifier: GPL-3.0-or-later
 
 find_package(PkgConfig QUIET)
-if (PKG_CONFIG_FOUND)
-    pkg_search_module(ENET QUIET IMPORTED_TARGET libenet)
-endif()
+pkg_search_module(ENET QUIET IMPORTED_TARGET libenet)
 
 include(FindPackageHandleStandardArgs)
 find_package_handle_standard_args(enet

externals/find-modules/Findhttplib.cmake

@@ -9,9 +9,7 @@ if (httplib_FOUND)
     find_package_handle_standard_args(httplib CONFIG_MODE)
 else()
     find_package(PkgConfig QUIET)
-    if (PKG_CONFIG_FOUND)
-        pkg_search_module(HTTPLIB QUIET IMPORTED_TARGET cpp-httplib)
-    endif()
+    pkg_search_module(HTTPLIB QUIET IMPORTED_TARGET cpp-httplib)
     find_package_handle_standard_args(httplib
         REQUIRED_VARS HTTPLIB_INCLUDEDIR
         VERSION_VAR HTTPLIB_VERSION

externals/find-modules/Findinih.cmake

@@ -3,9 +3,7 @@
 # SPDX-License-Identifier: GPL-3.0-or-later
 
 find_package(PkgConfig QUIET)
-if (PKG_CONFIG_FOUND)
-    pkg_search_module(INIREADER QUIET IMPORTED_TARGET INIReader)
-endif()
+pkg_search_module(INIREADER QUIET IMPORTED_TARGET INIReader)
 
 include(FindPackageHandleStandardArgs)
 find_package_handle_standard_args(inih

externals/find-modules/Findlibusb.cmake

@@ -3,9 +3,7 @@
 # SPDX-License-Identifier: GPL-3.0-or-later
 
 find_package(PkgConfig QUIET)
-if (PKG_CONFIG_FOUND)
-    pkg_search_module(LIBUSB QUIET IMPORTED_TARGET libusb-1.0)
-endif()
+pkg_search_module(LIBUSB QUIET IMPORTED_TARGET libusb-1.0)
 
 include(FindPackageHandleStandardArgs)
 find_package_handle_standard_args(libusb

externals/find-modules/Findlz4.cmake

@@ -8,9 +8,7 @@ if (lz4_FOUND)
     find_package_handle_standard_args(lz4 CONFIG_MODE)
 else()
     find_package(PkgConfig QUIET)
-    if (PKG_CONFIG_FOUND)
-        pkg_search_module(LZ4 QUIET IMPORTED_TARGET liblz4)
-    endif()
+    pkg_search_module(LZ4 QUIET IMPORTED_TARGET liblz4)
     find_package_handle_standard_args(lz4
         REQUIRED_VARS LZ4_LINK_LIBRARIES
         VERSION_VAR LZ4_VERSION

externals/find-modules/Findzstd.cmake

@@ -8,9 +8,7 @@ if (zstd_FOUND)
     find_package_handle_standard_args(zstd CONFIG_MODE)
 else()
     find_package(PkgConfig QUIET)
-    if (PKG_CONFIG_FOUND)
-        pkg_search_module(ZSTD QUIET IMPORTED_TARGET libzstd)
-    endif()
+    pkg_search_module(ZSTD QUIET IMPORTED_TARGET libzstd)
     find_package_handle_standard_args(zstd
         REQUIRED_VARS ZSTD_LINK_LIBRARIES
         VERSION_VAR ZSTD_VERSION

src/audio_core/device/audio_buffers.h

@@ -91,9 +91,10 @@ public:
      * @param core_timing - The CoreTiming instance
      * @param session     - The device session
      *
-     * @return Is the buffer was released.
+     * @return If any buffer was released.
      */
-    bool ReleaseBuffers(const Core::Timing::CoreTiming& core_timing, const DeviceSession& session) {
+    bool ReleaseBuffers(const Core::Timing::CoreTiming& core_timing, const DeviceSession& session,
+                        bool force) {
         std::scoped_lock l{lock};
         bool buffer_released{false};
         while (registered_count > 0) {
@@ -103,7 +104,8 @@ public:
             }
 
             // Check with the backend if this buffer can be released yet.
-            if (!session.IsBufferConsumed(buffers[index])) {
+            // If we're shutting down, we don't care if it's been played or not.
+            if (!force && !session.IsBufferConsumed(buffers[index])) {
                 break;
             }
 

src/audio_core/device/device_session.cpp

@@ -73,6 +73,12 @@ void DeviceSession::Stop() {
     }
 }
 
+void DeviceSession::ClearBuffers() {
+    if (stream) {
+        stream->ClearQueue();
+    }
+}
+
 void DeviceSession::AppendBuffers(std::span<const AudioBuffer> buffers) const {
     for (const auto& buffer : buffers) {
         Sink::SinkBuffer new_buffer{

src/audio_core/device/device_session.h

@@ -90,6 +90,11 @@ public:
      */
     void Stop();
 
+    /**
+     * Clear out the underlying audio buffers in the backend stream.
+     */
+    void ClearBuffers();
+
     /**
      * Set this device session's volume.
      *

src/audio_core/in/audio_in_system.cpp

@@ -23,7 +23,6 @@ System::~System() {
 void System::Finalize() {
     Stop();
     session->Finalize();
-    buffer_event->Signal();
 }
 
 void System::StartSession() {
@@ -102,6 +101,10 @@ Result System::Stop() {
     if (state == State::Started) {
         session->Stop();
         session->SetVolume(0.0f);
+        session->ClearBuffers();
+        if (buffers.ReleaseBuffers(system.CoreTiming(), *session, true)) {
+            buffer_event->Signal();
+        }
         state = State::Stopped;
     }
 
@@ -138,7 +141,7 @@ void System::RegisterBuffers() {
 }
 
 void System::ReleaseBuffers() {
-    bool signal{buffers.ReleaseBuffers(system.CoreTiming(), *session)};
+    bool signal{buffers.ReleaseBuffers(system.CoreTiming(), *session, false)};
 
     if (signal) {
         // Signal if any buffer was released, or if none are registered, we need more.

src/audio_core/out/audio_out_system.cpp

@@ -24,7 +24,6 @@ System::~System() {
 void System::Finalize() {
     Stop();
     session->Finalize();
-    buffer_event->Signal();
 }
 
 std::string_view System::GetDefaultOutputDeviceName() const {
@@ -102,6 +101,10 @@ Result System::Stop() {
     if (state == State::Started) {
         session->Stop();
         session->SetVolume(0.0f);
+        session->ClearBuffers();
+        if (buffers.ReleaseBuffers(system.CoreTiming(), *session, true)) {
+            buffer_event->Signal();
+        }
         state = State::Stopped;
     }
 
@@ -138,7 +141,7 @@ void System::RegisterBuffers() {
 }
 
 void System::ReleaseBuffers() {
-    bool signal{buffers.ReleaseBuffers(system.CoreTiming(), *session)};
+    bool signal{buffers.ReleaseBuffers(system.CoreTiming(), *session, false)};
     if (signal) {
         // Signal if any buffer was released, or if none are registered, we need more.
         buffer_event->Signal();

src/audio_core/renderer/performance/detail_aspect.h

@@ -16,7 +16,6 @@ class CommandGenerator;
  */
 class DetailAspect {
 public:
-    DetailAspect() = default;
     DetailAspect(CommandGenerator& command_generator, PerformanceEntryType entry_type, s32 node_id,
                  PerformanceDetailType detail_type);
 

src/audio_core/renderer/performance/entry_aspect.h

@@ -16,7 +16,6 @@ class CommandGenerator;
  */
 class EntryAspect {
 public:
-    EntryAspect() = default;
     EntryAspect(CommandGenerator& command_generator, PerformanceEntryType type, s32 node_id);
 
     /// Command generator the command will be generated into

src/common/CMakeLists.txt

@@ -34,8 +34,6 @@ add_library(common STATIC
     bit_util.h
     cityhash.cpp
     cityhash.h
-    cache_management.cpp
-    cache_management.h
     common_funcs.h
     common_precompiled_headers.h
     common_types.h

src/common/assert.h

@@ -69,7 +69,7 @@ void assert_fail_impl();
 #define ASSERT_OR_EXECUTE(_a_, _b_)                                                                \
     do {                                                                                           \
         ASSERT(_a_);                                                                               \
-        if (!(_a_)) {                                                                              \
+        if (!(_a_)) [[unlikely]] {                                                                 \
             _b_                                                                                    \
         }                                                                                          \
     } while (0)
@@ -78,7 +78,7 @@ void assert_fail_impl();
 #define ASSERT_OR_EXECUTE_MSG(_a_, _b_, ...)                                                       \
     do {                                                                                           \
         ASSERT_MSG(_a_, __VA_ARGS__);                                                              \
-        if (!(_a_)) {                                                                              \
+        if (!(_a_)) [[unlikely]] {                                                                 \
             _b_                                                                                    \
         }                                                                                          \
     } while (0)

src/common/cache_management.cpp

@@ -1,59 +0,0 @@
-// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
-// SPDX-License-Identifier: GPL-2.0-or-later
-
-#include <cstdint>
-#include <cstring>
-
-#include "common/cache_management.h"
-
-namespace Common {
-
-#if defined(ARCHITECTURE_x86_64)
-
-// Most cache operations are no-ops on x86
-
-void DataCacheLineCleanByVAToPoU(void* start, size_t size) {}
-void DataCacheLineCleanAndInvalidateByVAToPoC(void* start, size_t size) {}
-void DataCacheLineCleanByVAToPoC(void* start, size_t size) {}
-void DataCacheZeroByVA(void* start, size_t size) {
-    std::memset(start, 0, size);
-}
-
-#elif defined(ARCHITECTURE_arm64)
-
-// BS/DminLine is log2(cache size in words), we want size in bytes
-#define EXTRACT_DMINLINE(ctr_el0) (1 << ((((ctr_el0) >> 16) & 0xf) + 2))
-#define EXTRACT_BS(dczid_el0) (1 << (((dczid_el0)&0xf) + 2))
-
-#define DEFINE_DC_OP(op_name, function_name)                                                       \
-    void function_name(void* start, size_t size) {                                                 \
-        size_t ctr_el0;                                                                            \
-        asm volatile("mrs %[ctr_el0], ctr_el0\n\t" : [ctr_el0] "=r"(ctr_el0));                     \
-        size_t cacheline_size = EXTRACT_DMINLINE(ctr_el0);                                         \
-        uintptr_t va_start = reinterpret_cast<uintptr_t>(start);                                   \
-        uintptr_t va_end = va_start + size;                                                        \
-        for (uintptr_t va = va_start; va < va_end; va += cacheline_size) {                         \
-            asm volatile("dc " #op_name ", %[va]\n\t" : : [va] "r"(va) : "memory");                \
-        }                                                                                          \
-    }
-
-#define DEFINE_DC_OP_DCZID(op_name, function_name)                                                 \
-    void function_name(void* start, size_t size) {                                                 \
-        size_t dczid_el0;                                                                          \
-        asm volatile("mrs %[dczid_el0], dczid_el0\n\t" : [dczid_el0] "=r"(dczid_el0));             \
-        size_t cacheline_size = EXTRACT_BS(dczid_el0);                                             \
-        uintptr_t va_start = reinterpret_cast<uintptr_t>(start);                                   \
-        uintptr_t va_end = va_start + size;                                                        \
-        for (uintptr_t va = va_start; va < va_end; va += cacheline_size) {                         \
-            asm volatile("dc " #op_name ", %[va]\n\t" : : [va] "r"(va) : "memory");                \
-        }                                                                                          \
-    }
-
-DEFINE_DC_OP(cvau, DataCacheLineCleanByVAToPoU);
-DEFINE_DC_OP(civac, DataCacheLineCleanAndInvalidateByVAToPoC);
-DEFINE_DC_OP(cvac, DataCacheLineCleanByVAToPoC);
-DEFINE_DC_OP_DCZID(zva, DataCacheZeroByVA);
-
-#endif
-
-} // namespace Common

src/common/cache_management.h

@@ -1,27 +0,0 @@
-// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
-// SPDX-License-Identifier: GPL-2.0-or-later
-
-#pragma once
-
-#include <cstddef>
-
-namespace Common {
-
-// Data cache instructions enabled at EL0 by SCTLR_EL1.UCI.
-// VA = virtual address
-// PoC = point of coherency
-// PoU = point of unification
-
-// dc cvau
-void DataCacheLineCleanByVAToPoU(void* start, size_t size);
-
-// dc civac
-void DataCacheLineCleanAndInvalidateByVAToPoC(void* start, size_t size);
-
-// dc cvac
-void DataCacheLineCleanByVAToPoC(void* start, size_t size);
-
-// dc zva
-void DataCacheZeroByVA(void* start, size_t size);
-
-} // namespace Common

src/common/settings.cpp

@@ -40,6 +40,7 @@ void LogSettings() {
     LOG_INFO(Config, "yuzu Configuration:");
     log_setting("Controls_UseDockedMode", values.use_docked_mode.GetValue());
     log_setting("System_RngSeed", values.rng_seed.GetValue().value_or(0));
+    log_setting("System_DeviceName", values.device_name.GetValue());
     log_setting("System_CurrentUser", values.current_user.GetValue());
     log_setting("System_LanguageIndex", values.language_index.GetValue());
     log_setting("System_RegionIndex", values.region_index.GetValue());

src/common/settings.h

@@ -76,7 +76,8 @@ enum class ScalingFilter : u32 {
 enum class AntiAliasing : u32 {
     None = 0,
     Fxaa = 1,
-    LastAA = Fxaa,
+    Smaa = 2,
+    LastAA = Smaa,
 };
 
 struct ResolutionScalingInfo {
@@ -457,6 +458,7 @@ struct Values {
 
     // System
     SwitchableSetting<std::optional<u32>> rng_seed{std::optional<u32>(), "rng_seed"};
+    Setting<std::string> device_name{"Yuzu", "device_name"};
     // Measured in seconds since epoch
     std::optional<s64> custom_rtc;
     // Set on game boot, reset on stop. Seconds difference between current time and `custom_rtc`

src/common/thread.h

@@ -11,6 +11,7 @@
 #include <mutex>
 #include <thread>
 #include "common/common_types.h"
+#include "common/polyfill_thread.h"
 
 namespace Common {
 
@@ -69,7 +70,7 @@ public:
     explicit Barrier(std::size_t count_) : count(count_) {}
 
     /// Blocks until all "count" threads have called Sync()
-    void Sync() {
+    bool Sync(std::stop_token token = {}) {
         std::unique_lock lk{mutex};
         const std::size_t current_generation = generation;
 
@@ -77,14 +78,16 @@ public:
             generation++;
             waiting = 0;
             condvar.notify_all();
+            return true;
         } else {
-            condvar.wait(lk,
-                         [this, current_generation] { return current_generation != generation; });
+            CondvarWait(condvar, lk, token,
+                        [this, current_generation] { return current_generation != generation; });
+            return !token.stop_requested();
         }
     }
 
 private:
-    std::condition_variable condvar;
+    std::condition_variable_any condvar;
     std::mutex mutex;
     std::size_t count;
     std::size_t waiting = 0;

src/core/core.cpp

@@ -389,7 +389,9 @@ struct System::Impl {
         kernel.ShutdownCores();
         cpu_manager.Shutdown();
         debugger.reset();
-        services->KillNVNFlinger();
+        if (services) {
+            services->KillNVNFlinger();
+        }
         kernel.CloseServices();
         services.reset();
         service_manager.reset();

src/core/cpu_manager.cpp

@@ -20,23 +20,20 @@ namespace Core {
 CpuManager::CpuManager(System& system_) : system{system_} {}
 CpuManager::~CpuManager() = default;
 
-void CpuManager::ThreadStart(std::stop_token stop_token, CpuManager& cpu_manager,
-                             std::size_t core) {
-    cpu_manager.RunThread(core);
-}
-
 void CpuManager::Initialize() {
     num_cores = is_multicore ? Core::Hardware::NUM_CPU_CORES : 1;
     gpu_barrier = std::make_unique<Common::Barrier>(num_cores + 1);
 
     for (std::size_t core = 0; core < num_cores; core++) {
-        core_data[core].host_thread = std::jthread(ThreadStart, std::ref(*this), core);
+        core_data[core].host_thread =
+            std::jthread([this, core](std::stop_token token) { RunThread(token, core); });
     }
 }
 
 void CpuManager::Shutdown() {
     for (std::size_t core = 0; core < num_cores; core++) {
         if (core_data[core].host_thread.joinable()) {
+            core_data[core].host_thread.request_stop();
             core_data[core].host_thread.join();
         }
     }
@@ -184,7 +181,7 @@ void CpuManager::ShutdownThread() {
     UNREACHABLE();
 }
 
-void CpuManager::RunThread(std::size_t core) {
+void CpuManager::RunThread(std::stop_token token, std::size_t core) {
     /// Initialization
     system.RegisterCoreThread(core);
     std::string name;
@@ -206,7 +203,9 @@ void CpuManager::RunThread(std::size_t core) {
     });
 
     // Running
-    gpu_barrier->Sync();
+    if (!gpu_barrier->Sync(token)) {
+        return;
+    }
 
     if (!is_async_gpu && !is_multicore) {
         system.GPU().ObtainContext();

src/core/cpu_manager.h

@@ -81,12 +81,10 @@ private:
     void SingleCoreRunGuestThread();
     void SingleCoreRunIdleThread();
 
-    static void ThreadStart(std::stop_token stop_token, CpuManager& cpu_manager, std::size_t core);
-
     void GuestActivate();
     void HandleInterrupt();
     void ShutdownThread();
-    void RunThread(std::size_t core);
+    void RunThread(std::stop_token stop_token, std::size_t core);
 
     struct CoreData {
         std::shared_ptr<Common::Fiber> host_context;

src/core/frontend/emu_window.h

@@ -131,6 +131,10 @@ public:
         return active_config;
     }
 
+    bool StrictContextRequired() const {
+        return strict_context_required;
+    }
+
     /**
      * Requests the internal configuration to be replaced by the specified argument at some point in
      * the future.
@@ -207,6 +211,8 @@ protected:
 
     WindowSystemInfo window_info;
 
+    bool strict_context_required = false;
+
 private:
     /**
      * Handler called when the minimal client area was requested to be changed via SetConfig.

src/core/hid/emulated_controller.cpp

@@ -145,6 +145,7 @@ void EmulatedController::LoadDevices() {
     output_params[3].Set("output", true);
 
     LoadTASParams();
+    LoadVirtualGamepadParams();
 
     std::ranges::transform(button_params, button_devices.begin(), Common::Input::CreateInputDevice);
     std::ranges::transform(stick_params, stick_devices.begin(), Common::Input::CreateInputDevice);
@@ -163,6 +164,12 @@ void EmulatedController::LoadDevices() {
                            Common::Input::CreateInputDevice);
     std::ranges::transform(tas_stick_params, tas_stick_devices.begin(),
                            Common::Input::CreateInputDevice);
+
+    // Initialize virtual gamepad devices
+    std::ranges::transform(virtual_button_params, virtual_button_devices.begin(),
+                           Common::Input::CreateInputDevice);
+    std::ranges::transform(virtual_stick_params, virtual_stick_devices.begin(),
+                           Common::Input::CreateInputDevice);
 }
 
 void EmulatedController::LoadTASParams() {
@@ -205,6 +212,46 @@ void EmulatedController::LoadTASParams() {
     tas_stick_params[Settings::NativeAnalog::RStick].Set("axis_y", 3);
 }
 
+void EmulatedController::LoadVirtualGamepadParams() {
+    const auto player_index = NpadIdTypeToIndex(npad_id_type);
+    Common::ParamPackage common_params{};
+    common_params.Set("engine", "virtual_gamepad");
+    common_params.Set("port", static_cast<int>(player_index));
+    for (auto& param : virtual_button_params) {
+        param = common_params;
+    }
+    for (auto& param : virtual_stick_params) {
+        param = common_params;
+    }
+
+    // TODO(german77): Replace this with an input profile or something better
+    virtual_button_params[Settings::NativeButton::A].Set("button", 0);
+    virtual_button_params[Settings::NativeButton::B].Set("button", 1);
+    virtual_button_params[Settings::NativeButton::X].Set("button", 2);
+    virtual_button_params[Settings::NativeButton::Y].Set("button", 3);
+    virtual_button_params[Settings::NativeButton::LStick].Set("button", 4);
+    virtual_button_params[Settings::NativeButton::RStick].Set("button", 5);
+    virtual_button_params[Settings::NativeButton::L].Set("button", 6);
+    virtual_button_params[Settings::NativeButton::R].Set("button", 7);
+    virtual_button_params[Settings::NativeButton::ZL].Set("button", 8);
+    virtual_button_params[Settings::NativeButton::ZR].Set("button", 9);
+    virtual_button_params[Settings::NativeButton::Plus].Set("button", 10);
+    virtual_button_params[Settings::NativeButton::Minus].Set("button", 11);
+    virtual_button_params[Settings::NativeButton::DLeft].Set("button", 12);
+    virtual_button_params[Settings::NativeButton::DUp].Set("button", 13);
+    virtual_button_params[Settings::NativeButton::DRight].Set("button", 14);
+    virtual_button_params[Settings::NativeButton::DDown].Set("button", 15);
+    virtual_button_params[Settings::NativeButton::SL].Set("button", 16);
+    virtual_button_params[Settings::NativeButton::SR].Set("button", 17);
+    virtual_button_params[Settings::NativeButton::Home].Set("button", 18);
+    virtual_button_params[Settings::NativeButton::Screenshot].Set("button", 19);
+
+    virtual_stick_params[Settings::NativeAnalog::LStick].Set("axis_x", 0);
+    virtual_stick_params[Settings::NativeAnalog::LStick].Set("axis_y", 1);
+    virtual_stick_params[Settings::NativeAnalog::RStick].Set("axis_x", 2);
+    virtual_stick_params[Settings::NativeAnalog::RStick].Set("axis_y", 3);
+}
+
 void EmulatedController::ReloadInput() {
     // If you load any device here add the equivalent to the UnloadInput() function
     LoadDevices();
@@ -322,6 +369,35 @@ void EmulatedController::ReloadInput() {
                 },
         });
     }
+
+    // Use a common UUID for Virtual Gamepad
+    static constexpr Common::UUID VIRTUAL_UUID = Common::UUID{
+        {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x7, 0xFF, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}};
+
+    // Register virtual devices. No need to force update
+    for (std::size_t index = 0; index < virtual_button_devices.size(); ++index) {
+        if (!virtual_button_devices[index]) {
+            continue;
+        }
+        virtual_button_devices[index]->SetCallback({
+            .on_change =
+                [this, index](const Common::Input::CallbackStatus& callback) {
+                    SetButton(callback, index, VIRTUAL_UUID);
+                },
+        });
+    }
+
+    for (std::size_t index = 0; index < virtual_stick_devices.size(); ++index) {
+        if (!virtual_stick_devices[index]) {
+            continue;
+        }
+        virtual_stick_devices[index]->SetCallback({
+            .on_change =
+                [this, index](const Common::Input::CallbackStatus& callback) {
+                    SetStick(callback, index, VIRTUAL_UUID);
+                },
+        });
+    }
 }
 
 void EmulatedController::UnloadInput() {
@@ -349,6 +425,12 @@ void EmulatedController::UnloadInput() {
     for (auto& stick : tas_stick_devices) {
         stick.reset();
     }
+    for (auto& button : virtual_button_devices) {
+        button.reset();
+    }
+    for (auto& stick : virtual_stick_devices) {
+        stick.reset();
+    }
     camera_devices.reset();
     nfc_devices.reset();
 }

src/core/hid/emulated_controller.h

@@ -385,6 +385,9 @@ private:
     /// Set the params for TAS devices
     void LoadTASParams();
 
+    /// Set the params for virtual pad devices
+    void LoadVirtualGamepadParams();
+
     /**
      * @param use_temporary_value If true tmp_npad_type will be used
      * @return true if the controller style is fullkey
@@ -500,6 +503,12 @@ private:
     ButtonDevices tas_button_devices;
     StickDevices tas_stick_devices;
 
+    // Virtual gamepad related variables
+    ButtonParams virtual_button_params;
+    StickParams virtual_stick_params;
+    ButtonDevices virtual_button_devices;
+    StickDevices virtual_stick_devices;
+
     mutable std::mutex mutex;
     mutable std::mutex callback_mutex;
     std::unordered_map<int, ControllerUpdateCallback> callback_list;

src/core/hle/kernel/hle_ipc.cpp

@@ -167,6 +167,9 @@ void HLERequestContext::ParseCommandBuffer(const KHandleTable& handle_table, u32
         }
         if (incoming) {
             // Populate the object lists with the data in the IPC request.
+            incoming_copy_handles.reserve(handle_descriptor_header->num_handles_to_copy);
+            incoming_move_handles.reserve(handle_descriptor_header->num_handles_to_move);
+
             for (u32 handle = 0; handle < handle_descriptor_header->num_handles_to_copy; ++handle) {
                 incoming_copy_handles.push_back(rp.Pop<Handle>());
             }
@@ -181,6 +184,11 @@ void HLERequestContext::ParseCommandBuffer(const KHandleTable& handle_table, u32
         }
     }
 
+    buffer_x_desciptors.reserve(command_header->num_buf_x_descriptors);
+    buffer_a_desciptors.reserve(command_header->num_buf_a_descriptors);
+    buffer_b_desciptors.reserve(command_header->num_buf_b_descriptors);
+    buffer_w_desciptors.reserve(command_header->num_buf_w_descriptors);
+
     for (u32 i = 0; i < command_header->num_buf_x_descriptors; ++i) {
         buffer_x_desciptors.push_back(rp.PopRaw<IPC::BufferDescriptorX>());
     }
@@ -318,25 +326,23 @@ Result HLERequestContext::WriteToOutgoingCommandBuffer(KThread& requesting_threa
 }
 
 std::vector<u8> HLERequestContext::ReadBuffer(std::size_t buffer_index) const {
-    std::vector<u8> buffer{};
     const bool is_buffer_a{BufferDescriptorA().size() > buffer_index &&
                            BufferDescriptorA()[buffer_index].Size()};
-
     if (is_buffer_a) {
         ASSERT_OR_EXECUTE_MSG(
-            BufferDescriptorA().size() > buffer_index, { return buffer; },
+            BufferDescriptorA().size() > buffer_index, { return {}; },
             "BufferDescriptorA invalid buffer_index {}", buffer_index);
-        buffer.resize(BufferDescriptorA()[buffer_index].Size());
+        std::vector<u8> buffer(BufferDescriptorA()[buffer_index].Size());
         memory.ReadBlock(BufferDescriptorA()[buffer_index].Address(), buffer.data(), buffer.size());
+        return buffer;
     } else {
         ASSERT_OR_EXECUTE_MSG(
-            BufferDescriptorX().size() > buffer_index, { return buffer; },
+            BufferDescriptorX().size() > buffer_index, { return {}; },
             "BufferDescriptorX invalid buffer_index {}", buffer_index);
-        buffer.resize(BufferDescriptorX()[buffer_index].Size());
+        std::vector<u8> buffer(BufferDescriptorX()[buffer_index].Size());
         memory.ReadBlock(BufferDescriptorX()[buffer_index].Address(), buffer.data(), buffer.size());
+        return buffer;
     }
-
-    return buffer;
 }
 
 std::size_t HLERequestContext::WriteBuffer(const void* buffer, std::size_t size,

src/core/hle/kernel/k_process.cpp

@@ -285,6 +285,17 @@ void KProcess::UnregisterThread(KThread* thread) {
     thread_list.remove(thread);
 }
 
+u64 KProcess::GetFreeThreadCount() const {
+    if (resource_limit != nullptr) {
+        const auto current_value =
+            resource_limit->GetCurrentValue(LimitableResource::ThreadCountMax);
+        const auto limit_value = resource_limit->GetLimitValue(LimitableResource::ThreadCountMax);
+        return limit_value - current_value;
+    } else {
+        return 0;
+    }
+}
+
 Result KProcess::Reset() {
     // Lock the process and the scheduler.
     KScopedLightLock lk(state_lock);

src/core/hle/kernel/k_process.h

@@ -304,6 +304,9 @@ public:
     /// from this process' thread list.
     void UnregisterThread(KThread* thread);
 
+    /// Retrieves the number of available threads for this process.
+    u64 GetFreeThreadCount() const;
+
     /// Clears the signaled state of the process if and only if it's signaled.
     ///
     /// @pre The process must not be already terminated. If this is called on a

src/core/hle/kernel/kernel.cpp

@@ -104,12 +104,16 @@ struct KernelCore::Impl {
     }
 
     void CloseCurrentProcess() {
-        (*current_process).Finalize();
-        // current_process->Close();
-        // TODO: The current process should be destroyed based on accurate ref counting after
+        KProcess* old_process = current_process.exchange(nullptr);
+        if (old_process == nullptr) {
+            return;
+        }
+
+        // old_process->Close();
+        // TODO: The process should be destroyed based on accurate ref counting after
         // calling Close(). Adding a manual Destroy() call instead to avoid a memory leak.
-        (*current_process).Destroy();
-        current_process = nullptr;
+        old_process->Finalize();
+        old_process->Destroy();
     }
 
     void Shutdown() {

src/core/hle/kernel/service_thread.cpp

@@ -163,9 +163,6 @@ ServiceThread::Impl::~Impl() {
     m_wakeup_event->Signal();
     m_host_thread.join();
 
-    // Lock mutex.
-    m_session_mutex.lock();
-
     // Close all remaining sessions.
     for (const auto& [server_session, manager] : m_sessions) {
         server_session->Close();

src/core/hle/kernel/svc.cpp

@@ -784,63 +784,29 @@ static Result GetInfo(Core::System& system, u64* result, u64 info_id, Handle han
     LOG_TRACE(Kernel_SVC, "called info_id=0x{:X}, info_sub_id=0x{:X}, handle=0x{:08X}", info_id,
               info_sub_id, handle);
 
-    enum class GetInfoType : u64 {
-        // 1.0.0+
-        AllowedCPUCoreMask = 0,
-        AllowedThreadPriorityMask = 1,
-        MapRegionBaseAddr = 2,
-        MapRegionSize = 3,
-        HeapRegionBaseAddr = 4,
-        HeapRegionSize = 5,
-        TotalPhysicalMemoryAvailable = 6,
-        TotalPhysicalMemoryUsed = 7,
-        IsCurrentProcessBeingDebugged = 8,
-        RegisterResourceLimit = 9,
-        IdleTickCount = 10,
-        RandomEntropy = 11,
-        ThreadTickCount = 0xF0000002,
-        // 2.0.0+
-        ASLRRegionBaseAddr = 12,
-        ASLRRegionSize = 13,
-        StackRegionBaseAddr = 14,
-        StackRegionSize = 15,
-        // 3.0.0+
-        SystemResourceSize = 16,
-        SystemResourceUsage = 17,
-        TitleId = 18,
-        // 4.0.0+
-        PrivilegedProcessId = 19,
-        // 5.0.0+
-        UserExceptionContextAddr = 20,
-        // 6.0.0+
-        TotalPhysicalMemoryAvailableWithoutSystemResource = 21,
-        TotalPhysicalMemoryUsedWithoutSystemResource = 22,
-
-        // Homebrew only
-        MesosphereCurrentProcess = 65001,
-    };
-
-    const auto info_id_type = static_cast<GetInfoType>(info_id);
+    const auto info_id_type = static_cast<InfoType>(info_id);
 
     switch (info_id_type) {
-    case GetInfoType::AllowedCPUCoreMask:
-    case GetInfoType::AllowedThreadPriorityMask:
-    case GetInfoType::MapRegionBaseAddr:
-    case GetInfoType::MapRegionSize:
-    case GetInfoType::HeapRegionBaseAddr:
-    case GetInfoType::HeapRegionSize:
-    case GetInfoType::ASLRRegionBaseAddr:
-    case GetInfoType::ASLRRegionSize:
-    case GetInfoType::StackRegionBaseAddr:
-    case GetInfoType::StackRegionSize:
-    case GetInfoType::TotalPhysicalMemoryAvailable:
-    case GetInfoType::TotalPhysicalMemoryUsed:
-    case GetInfoType::SystemResourceSize:
-    case GetInfoType::SystemResourceUsage:
-    case GetInfoType::TitleId:
-    case GetInfoType::UserExceptionContextAddr:
-    case GetInfoType::TotalPhysicalMemoryAvailableWithoutSystemResource:
-    case GetInfoType::TotalPhysicalMemoryUsedWithoutSystemResource: {
+    case InfoType::CoreMask:
+    case InfoType::PriorityMask:
+    case InfoType::AliasRegionAddress:
+    case InfoType::AliasRegionSize:
+    case InfoType::HeapRegionAddress:
+    case InfoType::HeapRegionSize:
+    case InfoType::AslrRegionAddress:
+    case InfoType::AslrRegionSize:
+    case InfoType::StackRegionAddress:
+    case InfoType::StackRegionSize:
+    case InfoType::TotalMemorySize:
+    case InfoType::UsedMemorySize:
+    case InfoType::SystemResourceSizeTotal:
+    case InfoType::SystemResourceSizeUsed:
+    case InfoType::ProgramId:
+    case InfoType::UserExceptionContextAddress:
+    case InfoType::TotalNonSystemMemorySize:
+    case InfoType::UsedNonSystemMemorySize:
+    case InfoType::IsApplication:
+    case InfoType::FreeThreadCount: {
         if (info_sub_id != 0) {
             LOG_ERROR(Kernel_SVC, "Info sub id is non zero! info_id={}, info_sub_id={}", info_id,
                       info_sub_id);
@@ -856,79 +822,83 @@ static Result GetInfo(Core::System& system, u64* result, u64 info_id, Handle han
         }
 
         switch (info_id_type) {
-        case GetInfoType::AllowedCPUCoreMask:
+        case InfoType::CoreMask:
             *result = process->GetCoreMask();
             return ResultSuccess;
 
-        case GetInfoType::AllowedThreadPriorityMask:
+        case InfoType::PriorityMask:
             *result = process->GetPriorityMask();
             return ResultSuccess;
 
-        case GetInfoType::MapRegionBaseAddr:
+        case InfoType::AliasRegionAddress:
             *result = process->PageTable().GetAliasRegionStart();
             return ResultSuccess;
 
-        case GetInfoType::MapRegionSize:
+        case InfoType::AliasRegionSize:
             *result = process->PageTable().GetAliasRegionSize();
             return ResultSuccess;
 
-        case GetInfoType::HeapRegionBaseAddr:
+        case InfoType::HeapRegionAddress:
             *result = process->PageTable().GetHeapRegionStart();
             return ResultSuccess;
 
-        case GetInfoType::HeapRegionSize:
+        case InfoType::HeapRegionSize:
             *result = process->PageTable().GetHeapRegionSize();
             return ResultSuccess;
 
-        case GetInfoType::ASLRRegionBaseAddr:
+        case InfoType::AslrRegionAddress:
             *result = process->PageTable().GetAliasCodeRegionStart();
             return ResultSuccess;
 
-        case GetInfoType::ASLRRegionSize:
+        case InfoType::AslrRegionSize:
             *result = process->PageTable().GetAliasCodeRegionSize();
             return ResultSuccess;
 
-        case GetInfoType::StackRegionBaseAddr:
+        case InfoType::StackRegionAddress:
             *result = process->PageTable().GetStackRegionStart();
             return ResultSuccess;
 
-        case GetInfoType::StackRegionSize:
+        case InfoType::StackRegionSize:
             *result = process->PageTable().GetStackRegionSize();
             return ResultSuccess;
 
-        case GetInfoType::TotalPhysicalMemoryAvailable:
+        case InfoType::TotalMemorySize:
             *result = process->GetTotalPhysicalMemoryAvailable();
             return ResultSuccess;
 
-        case GetInfoType::TotalPhysicalMemoryUsed:
+        case InfoType::UsedMemorySize:
             *result = process->GetTotalPhysicalMemoryUsed();
             return ResultSuccess;
 
-        case GetInfoType::SystemResourceSize:
+        case InfoType::SystemResourceSizeTotal:
             *result = process->GetSystemResourceSize();
             return ResultSuccess;
 
-        case GetInfoType::SystemResourceUsage:
+        case InfoType::SystemResourceSizeUsed:
             LOG_WARNING(Kernel_SVC, "(STUBBED) Attempted to query system resource usage");
             *result = process->GetSystemResourceUsage();
             return ResultSuccess;
 
-        case GetInfoType::TitleId:
+        case InfoType::ProgramId:
             *result = process->GetProgramID();
             return ResultSuccess;
 
-        case GetInfoType::UserExceptionContextAddr:
+        case InfoType::UserExceptionContextAddress:
             *result = process->GetProcessLocalRegionAddress();
             return ResultSuccess;
 
-        case GetInfoType::TotalPhysicalMemoryAvailableWithoutSystemResource:
+        case InfoType::TotalNonSystemMemorySize:
             *result = process->GetTotalPhysicalMemoryAvailableWithoutSystemResource();
             return ResultSuccess;
 
-        case GetInfoType::TotalPhysicalMemoryUsedWithoutSystemResource:
+        case InfoType::UsedNonSystemMemorySize:
             *result = process->GetTotalPhysicalMemoryUsedWithoutSystemResource();
             return ResultSuccess;
 
+        case InfoType::FreeThreadCount:
+            *result = process->GetFreeThreadCount();
+            return ResultSuccess;
+
         default:
             break;
         }
@@ -937,11 +907,11 @@ static Result GetInfo(Core::System& system, u64* result, u64 info_id, Handle han
         return ResultInvalidEnumValue;
     }
 
-    case GetInfoType::IsCurrentProcessBeingDebugged:
+    case InfoType::DebuggerAttached:
         *result = 0;
         return ResultSuccess;
 
-    case GetInfoType::RegisterResourceLimit: {
+    case InfoType::ResourceLimit: {
         if (handle != 0) {
             LOG_ERROR(Kernel, "Handle is non zero! handle={:08X}", handle);
             return ResultInvalidHandle;
@@ -969,7 +939,7 @@ static Result GetInfo(Core::System& system, u64* result, u64 info_id, Handle han
         return ResultSuccess;
     }
 
-    case GetInfoType::RandomEntropy:
+    case InfoType::RandomEntropy:
         if (handle != 0) {
             LOG_ERROR(Kernel_SVC, "Process Handle is non zero, expected 0 result but got {:016X}",
                       handle);
@@ -985,13 +955,13 @@ static Result GetInfo(Core::System& system, u64* result, u64 info_id, Handle han
         *result = system.Kernel().CurrentProcess()->GetRandomEntropy(info_sub_id);
         return ResultSuccess;
 
-    case GetInfoType::PrivilegedProcessId:
+    case InfoType::InitialProcessIdRange:
         LOG_WARNING(Kernel_SVC,
                     "(STUBBED) Attempted to query privileged process id bounds, returned 0");
         *result = 0;
         return ResultSuccess;
 
-    case GetInfoType::ThreadTickCount: {
+    case InfoType::ThreadTickCount: {
         constexpr u64 num_cpus = 4;
         if (info_sub_id != 0xFFFFFFFFFFFFFFFF && info_sub_id >= num_cpus) {
             LOG_ERROR(Kernel_SVC, "Core count is out of range, expected {} but got {}", num_cpus,
@@ -1026,7 +996,7 @@ static Result GetInfo(Core::System& system, u64* result, u64 info_id, Handle han
         *result = out_ticks;
         return ResultSuccess;
     }
-    case GetInfoType::IdleTickCount: {
+    case InfoType::IdleTickCount: {
         // Verify the input handle is invalid.
         R_UNLESS(handle == InvalidHandle, ResultInvalidHandle);
 
@@ -1040,7 +1010,7 @@ static Result GetInfo(Core::System& system, u64* result, u64 info_id, Handle han
         *result = system.Kernel().CurrentScheduler()->GetIdleThread()->GetCpuTime();
         return ResultSuccess;
     }
-    case GetInfoType::MesosphereCurrentProcess: {
+    case InfoType::MesosphereCurrentProcess: {
         // Verify the input handle is invalid.
         R_UNLESS(handle == InvalidHandle, ResultInvalidHandle);
 

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

@@ -203,8 +203,9 @@ private:
 };
 
 AudInU::AudInU(Core::System& system_)
-    : ServiceFramework{system_, "audin:u"}, service_context{system_, "AudInU"},
-      impl{std::make_unique<AudioCore::AudioIn::Manager>(system_)} {
+    : ServiceFramework{system_, "audin:u", ServiceThreadType::CreateNew},
+      service_context{system_, "AudInU"}, impl{std::make_unique<AudioCore::AudioIn::Manager>(
+                                              system_)} {
     // clang-format off
     static const FunctionInfo functions[] = {
         {0, &AudInU::ListAudioIns, "ListAudioIns"},

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

@@ -26,8 +26,9 @@ public:
     explicit IAudioOut(Core::System& system_, AudioCore::AudioOut::Manager& manager,
                        size_t session_id, const std::string& device_name,
                        const AudioOutParameter& in_params, u32 handle, u64 applet_resource_user_id)
-        : ServiceFramework{system_, "IAudioOut"}, service_context{system_, "IAudioOut"},
-          event{service_context.CreateEvent("AudioOutEvent")},
+        : ServiceFramework{system_, "IAudioOut", ServiceThreadType::CreateNew},
+          service_context{system_, "IAudioOut"}, event{service_context.CreateEvent(
+                                                     "AudioOutEvent")},
           impl{std::make_shared<AudioCore::AudioOut::Out>(system_, manager, event, session_id)} {
 
         // clang-format off
@@ -220,8 +221,9 @@ private:
 };
 
 AudOutU::AudOutU(Core::System& system_)
-    : ServiceFramework{system_, "audout:u"}, service_context{system_, "AudOutU"},
-      impl{std::make_unique<AudioCore::AudioOut::Manager>(system_)} {
+    : ServiceFramework{system_, "audout:u", ServiceThreadType::CreateNew},
+      service_context{system_, "AudOutU"}, impl{std::make_unique<AudioCore::AudioOut::Manager>(
+                                               system_)} {
     // clang-format off
     static const FunctionInfo functions[] = {
         {0, &AudOutU::ListAudioOuts, "ListAudioOuts"},

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

@@ -35,9 +35,10 @@ public:
                             AudioCore::AudioRendererParameterInternal& params,
                             Kernel::KTransferMemory* transfer_memory, u64 transfer_memory_size,
                             u32 process_handle, u64 applet_resource_user_id, s32 session_id)
-        : ServiceFramework{system_, "IAudioRenderer"}, service_context{system_, "IAudioRenderer"},
-          rendered_event{service_context.CreateEvent("IAudioRendererEvent")}, manager{manager_},
-          impl{std::make_unique<Renderer>(system_, manager, rendered_event)} {
+        : ServiceFramework{system_, "IAudioRenderer", ServiceThreadType::CreateNew},
+          service_context{system_, "IAudioRenderer"}, rendered_event{service_context.CreateEvent(
+                                                          "IAudioRendererEvent")},
+          manager{manager_}, impl{std::make_unique<Renderer>(system_, manager, rendered_event)} {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IAudioRenderer::GetSampleRate, "GetSampleRate"},
@@ -242,8 +243,10 @@ class IAudioDevice final : public ServiceFramework<IAudioDevice> {
 public:
     explicit IAudioDevice(Core::System& system_, u64 applet_resource_user_id, u32 revision,
                           u32 device_num)
-        : ServiceFramework{system_, "IAudioDevice"}, service_context{system_, "IAudioDevice"},
-          impl{std::make_unique<AudioDevice>(system_, applet_resource_user_id, revision)},
+        : ServiceFramework{system_, "IAudioDevice", ServiceThreadType::CreateNew},
+          service_context{system_, "IAudioDevice"}, impl{std::make_unique<AudioDevice>(
+                                                        system_, applet_resource_user_id,
+                                                        revision)},
           event{service_context.CreateEvent(fmt::format("IAudioDeviceEvent-{}", device_num))} {
         static const FunctionInfo functions[] = {
             {0, &IAudioDevice::ListAudioDeviceName, "ListAudioDeviceName"},
@@ -418,7 +421,7 @@ private:
 };
 
 AudRenU::AudRenU(Core::System& system_)
-    : ServiceFramework{system_, "audren:u"},
+    : ServiceFramework{system_, "audren:u", ServiceThreadType::CreateNew},
       service_context{system_, "audren:u"}, impl{std::make_unique<Manager>(system_)} {
     // clang-format off
     static const FunctionInfo functions[] = {

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

@@ -191,6 +191,13 @@ void SET::GetKeyCodeMap2(Kernel::HLERequestContext& ctx) {
     GetKeyCodeMapImpl(ctx);
 }
 
+void SET::GetDeviceNickName(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_SET, "called");
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(ResultSuccess);
+    ctx.WriteBuffer(Settings::values.device_name.GetValue());
+}
+
 SET::SET(Core::System& system_) : ServiceFramework{system_, "set"} {
     // clang-format off
     static const FunctionInfo functions[] = {
@@ -205,7 +212,7 @@ SET::SET(Core::System& system_) : ServiceFramework{system_, "set"} {
         {8, &SET::GetQuestFlag, "GetQuestFlag"},
         {9, &SET::GetKeyCodeMap2, "GetKeyCodeMap2"},
         {10, nullptr, "GetFirmwareVersionForDebug"},
-        {11, nullptr, "GetDeviceNickName"},
+        {11, &SET::GetDeviceNickName, "GetDeviceNickName"},
     };
     // clang-format on
 

src/core/hle/service/set/set.h

@@ -50,6 +50,7 @@ private:
     void GetRegionCode(Kernel::HLERequestContext& ctx);
     void GetKeyCodeMap(Kernel::HLERequestContext& ctx);
     void GetKeyCodeMap2(Kernel::HLERequestContext& ctx);
+    void GetDeviceNickName(Kernel::HLERequestContext& ctx);
 };
 
 } // namespace Service::Set

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

@@ -3,6 +3,7 @@
 
 #include "common/assert.h"
 #include "common/logging/log.h"
+#include "common/settings.h"
 #include "core/file_sys/errors.h"
 #include "core/file_sys/system_archive/system_version.h"
 #include "core/hle/ipc_helpers.h"
@@ -176,6 +177,13 @@ void SET_SYS::GetSettingsItemValue(Kernel::HLERequestContext& ctx) {
     rb.Push(response);
 }
 
+void SET_SYS::GetDeviceNickName(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_SET, "called");
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(ResultSuccess);
+    ctx.WriteBuffer(::Settings::values.device_name.GetValue());
+}
+
 SET_SYS::SET_SYS(Core::System& system_) : ServiceFramework{system_, "set:sys"} {
     // clang-format off
     static const FunctionInfo functions[] = {
@@ -253,7 +261,7 @@ SET_SYS::SET_SYS(Core::System& system_) : ServiceFramework{system_, "set:sys"} {
         {74, nullptr, "SetWirelessLanEnableFlag"},
         {75, nullptr, "GetInitialLaunchSettings"},
         {76, nullptr, "SetInitialLaunchSettings"},
-        {77, nullptr, "GetDeviceNickName"},
+        {77, &SET_SYS::GetDeviceNickName, "GetDeviceNickName"},
         {78, nullptr, "SetDeviceNickName"},
         {79, nullptr, "GetProductModel"},
         {80, nullptr, "GetLdnChannel"},

src/core/hle/service/set/set_sys.h

@@ -29,6 +29,7 @@ private:
     void GetFirmwareVersion2(Kernel::HLERequestContext& ctx);
     void GetColorSetId(Kernel::HLERequestContext& ctx);
     void SetColorSetId(Kernel::HLERequestContext& ctx);
+    void GetDeviceNickName(Kernel::HLERequestContext& ctx);
 
     ColorSet color_set = ColorSet::BasicWhite;
 };

src/core/memory.cpp

@@ -6,7 +6,6 @@
 
 #include "common/assert.h"
 #include "common/atomic_ops.h"
-#include "common/cache_management.h"
 #include "common/common_types.h"
 #include "common/logging/log.h"
 #include "common/page_table.h"
@@ -195,13 +194,11 @@ struct Memory::Impl {
                 break;
             }
             case Common::PageType::Memory: {
-                DEBUG_ASSERT(pointer);
                 u8* mem_ptr = pointer + page_offset + (page_index << YUZU_PAGEBITS);
                 on_memory(copy_amount, mem_ptr);
                 break;
             }
             case Common::PageType::DebugMemory: {
-                DEBUG_ASSERT(pointer);
                 u8* const mem_ptr{GetPointerFromDebugMemory(current_vaddr)};
                 on_memory(copy_amount, mem_ptr);
                 break;
@@ -342,10 +339,9 @@ struct Memory::Impl {
                     LOG_ERROR(HW_Memory, "Unmapped cache maintenance @ {:#018X}", current_vaddr);
                     throw InvalidMemoryException();
                 },
-                [&](const std::size_t block_size, u8* const host_ptr) { cb(block_size, host_ptr); },
+                [&](const std::size_t block_size, u8* const host_ptr) {},
                 [&](const VAddr current_vaddr, const std::size_t block_size, u8* const host_ptr) {
-                    system.GPU().FlushRegion(current_vaddr, block_size);
-                    cb(block_size, host_ptr);
+                    cb(current_vaddr, block_size);
                 },
                 [](const std::size_t block_size) {});
         } catch (InvalidMemoryException&) {
@@ -356,27 +352,30 @@ struct Memory::Impl {
     }
 
     Result InvalidateDataCache(const Kernel::KProcess& process, VAddr dest_addr, std::size_t size) {
-        auto perform = [&](const std::size_t block_size, u8* const host_ptr) {
-            // Do nothing; this operation (dc ivac) cannot be supported
-            // from EL0
+        auto on_rasterizer = [&](const VAddr current_vaddr, const std::size_t block_size) {
+            // dc ivac: Invalidate to point of coherency
+            // GPU flush -> CPU invalidate
+            system.GPU().FlushRegion(current_vaddr, block_size);
         };
-        return PerformCacheOperation(process, dest_addr, size, perform);
+        return PerformCacheOperation(process, dest_addr, size, on_rasterizer);
     }
 
     Result StoreDataCache(const Kernel::KProcess& process, VAddr dest_addr, std::size_t size) {
-        auto perform = [&](const std::size_t block_size, u8* const host_ptr) {
+        auto on_rasterizer = [&](const VAddr current_vaddr, const std::size_t block_size) {
             // dc cvac: Store to point of coherency
-            Common::DataCacheLineCleanByVAToPoC(host_ptr, block_size);
+            // CPU flush -> GPU invalidate
+            system.GPU().InvalidateRegion(current_vaddr, block_size);
         };
-        return PerformCacheOperation(process, dest_addr, size, perform);
+        return PerformCacheOperation(process, dest_addr, size, on_rasterizer);
     }
 
     Result FlushDataCache(const Kernel::KProcess& process, VAddr dest_addr, std::size_t size) {
-        auto perform = [&](const std::size_t block_size, u8* const host_ptr) {
+        auto on_rasterizer = [&](const VAddr current_vaddr, const std::size_t block_size) {
             // dc civac: Store to point of coherency, and invalidate from cache
-            Common::DataCacheLineCleanAndInvalidateByVAToPoC(host_ptr, block_size);
+            // CPU flush -> GPU invalidate
+            system.GPU().InvalidateRegion(current_vaddr, block_size);
         };
-        return PerformCacheOperation(process, dest_addr, size, perform);
+        return PerformCacheOperation(process, dest_addr, size, on_rasterizer);
     }
 
     void MarkRegionDebug(VAddr vaddr, u64 size, bool debug) {

src/input_common/CMakeLists.txt

@@ -20,6 +20,8 @@ add_library(input_common STATIC
     drivers/udp_client.h
     drivers/virtual_amiibo.cpp
     drivers/virtual_amiibo.h
+    drivers/virtual_gamepad.cpp
+    drivers/virtual_gamepad.h
     helpers/stick_from_buttons.cpp
     helpers/stick_from_buttons.h
     helpers/touch_from_buttons.cpp

src/input_common/drivers/camera.cpp

@@ -17,7 +17,7 @@ Camera::Camera(std::string input_engine_) : InputEngine(std::move(input_engine_)
     PreSetController(identifier);
 }
 
-void Camera::SetCameraData(std::size_t width, std::size_t height, std::vector<u32> data) {
+void Camera::SetCameraData(std::size_t width, std::size_t height, std::span<const u32> data) {
     const std::size_t desired_width = getImageWidth();
     const std::size_t desired_height = getImageHeight();
     status.data.resize(desired_width * desired_height);

src/input_common/drivers/camera.h

@@ -3,6 +3,8 @@
 
 #pragma once
 
+#include <span>
+
 #include "input_common/input_engine.h"
 
 namespace InputCommon {
@@ -15,7 +17,7 @@ class Camera final : public InputEngine {
 public:
     explicit Camera(std::string input_engine_);
 
-    void SetCameraData(std::size_t width, std::size_t height, std::vector<u32> data);
+    void SetCameraData(std::size_t width, std::size_t height, std::span<const u32> data);
 
     std::size_t getImageWidth() const;
     std::size_t getImageHeight() const;

src/input_common/drivers/sdl_driver.cpp

@@ -16,6 +16,8 @@ Common::UUID GetGUID(SDL_Joystick* joystick) {
     const SDL_JoystickGUID guid = SDL_JoystickGetGUID(joystick);
     std::array<u8, 16> data{};
     std::memcpy(data.data(), guid.data, sizeof(data));
+    // Clear controller name crc
+    std::memset(data.data() + 2, 0, sizeof(u16));
     return Common::UUID{data};
 }
 } // Anonymous namespace

src/input_common/drivers/virtual_gamepad.cpp

@@ -0,0 +1,78 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include "input_common/drivers/virtual_gamepad.h"
+
+namespace InputCommon {
+constexpr std::size_t PlayerIndexCount = 10;
+
+VirtualGamepad::VirtualGamepad(std::string input_engine_) : InputEngine(std::move(input_engine_)) {
+    for (std::size_t i = 0; i < PlayerIndexCount; i++) {
+        PreSetController(GetIdentifier(i));
+    }
+}
+
+void VirtualGamepad::SetButtonState(std::size_t player_index, int button_id, bool value) {
+    if (player_index > PlayerIndexCount) {
+        return;
+    }
+    const auto identifier = GetIdentifier(player_index);
+    SetButton(identifier, button_id, value);
+}
+
+void VirtualGamepad::SetButtonState(std::size_t player_index, VirtualButton button_id, bool value) {
+    SetButtonState(player_index, static_cast<int>(button_id), value);
+}
+
+void VirtualGamepad::SetStickPosition(std::size_t player_index, int axis_id, float x_value,
+                                      float y_value) {
+    if (player_index > PlayerIndexCount) {
+        return;
+    }
+    const auto identifier = GetIdentifier(player_index);
+    SetAxis(identifier, axis_id * 2, x_value);
+    SetAxis(identifier, (axis_id * 2) + 1, y_value);
+}
+
+void VirtualGamepad::SetStickPosition(std::size_t player_index, VirtualStick axis_id, float x_value,
+                                      float y_value) {
+    SetStickPosition(player_index, static_cast<int>(axis_id), x_value, y_value);
+}
+
+void VirtualGamepad::ResetControllers() {
+    for (std::size_t i = 0; i < PlayerIndexCount; i++) {
+        SetStickPosition(i, VirtualStick::Left, 0.0f, 0.0f);
+        SetStickPosition(i, VirtualStick::Right, 0.0f, 0.0f);
+
+        SetButtonState(i, VirtualButton::ButtonA, false);
+        SetButtonState(i, VirtualButton::ButtonB, false);
+        SetButtonState(i, VirtualButton::ButtonX, false);
+        SetButtonState(i, VirtualButton::ButtonY, false);
+        SetButtonState(i, VirtualButton::StickL, false);
+        SetButtonState(i, VirtualButton::StickR, false);
+        SetButtonState(i, VirtualButton::TriggerL, false);
+        SetButtonState(i, VirtualButton::TriggerR, false);
+        SetButtonState(i, VirtualButton::TriggerZL, false);
+        SetButtonState(i, VirtualButton::TriggerZR, false);
+        SetButtonState(i, VirtualButton::ButtonPlus, false);
+        SetButtonState(i, VirtualButton::ButtonMinus, false);
+        SetButtonState(i, VirtualButton::ButtonLeft, false);
+        SetButtonState(i, VirtualButton::ButtonUp, false);
+        SetButtonState(i, VirtualButton::ButtonRight, false);
+        SetButtonState(i, VirtualButton::ButtonDown, false);
+        SetButtonState(i, VirtualButton::ButtonSL, false);
+        SetButtonState(i, VirtualButton::ButtonSR, false);
+        SetButtonState(i, VirtualButton::ButtonHome, false);
+        SetButtonState(i, VirtualButton::ButtonCapture, false);
+    }
+}
+
+PadIdentifier VirtualGamepad::GetIdentifier(std::size_t player_index) const {
+    return {
+        .guid = Common::UUID{},
+        .port = player_index,
+        .pad = 0,
+    };
+}
+
+} // namespace InputCommon

src/input_common/drivers/virtual_gamepad.h

@@ -0,0 +1,73 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include "input_common/input_engine.h"
+
+namespace InputCommon {
+
+/**
+ * A virtual controller that is always assigned to the game input
+ */
+class VirtualGamepad final : public InputEngine {
+public:
+    enum class VirtualButton {
+        ButtonA,
+        ButtonB,
+        ButtonX,
+        ButtonY,
+        StickL,
+        StickR,
+        TriggerL,
+        TriggerR,
+        TriggerZL,
+        TriggerZR,
+        ButtonPlus,
+        ButtonMinus,
+        ButtonLeft,
+        ButtonUp,
+        ButtonRight,
+        ButtonDown,
+        ButtonSL,
+        ButtonSR,
+        ButtonHome,
+        ButtonCapture,
+    };
+
+    enum class VirtualStick {
+        Left = 0,
+        Right = 1,
+    };
+
+    explicit VirtualGamepad(std::string input_engine_);
+
+    /**
+     * Sets the status of all buttons bound with the key to pressed
+     * @param player_index the player number that will take this action
+     * @param button_id the id of the button
+     * @param value indicates if the button is pressed or not
+     */
+    void SetButtonState(std::size_t player_index, int button_id, bool value);
+    void SetButtonState(std::size_t player_index, VirtualButton button_id, bool value);
+
+    /**
+     * Sets the status of all buttons bound with the key to released
+     * @param player_index the player number that will take this action
+     * @param axis_id the id of the axis to move
+     * @param x_value the position of the stick in the x axis
+     * @param y_value the position of the stick in the y axis
+     */
+    void SetStickPosition(std::size_t player_index, int axis_id, float x_value, float y_value);
+    void SetStickPosition(std::size_t player_index, VirtualStick axis_id, float x_value,
+                          float y_value);
+
+    /// Restores all inputs into the neutral position
+    void ResetControllers();
+
+private:
+    /// Returns the correct identifier corresponding to the player index
+    PadIdentifier GetIdentifier(std::size_t player_index) const;
+};
+
+} // namespace InputCommon

src/input_common/main.cpp

@@ -12,6 +12,7 @@
 #include "input_common/drivers/touch_screen.h"
 #include "input_common/drivers/udp_client.h"
 #include "input_common/drivers/virtual_amiibo.h"
+#include "input_common/drivers/virtual_gamepad.h"
 #include "input_common/helpers/stick_from_buttons.h"
 #include "input_common/helpers/touch_from_buttons.h"
 #include "input_common/input_engine.h"
@@ -85,6 +86,12 @@ struct InputSubsystem::Impl {
         Common::Input::RegisterOutputFactory(virtual_amiibo->GetEngineName(),
                                              virtual_amiibo_output_factory);
 
+        virtual_gamepad = std::make_shared<VirtualGamepad>("virtual_gamepad");
+        virtual_gamepad->SetMappingCallback(mapping_callback);
+        virtual_gamepad_input_factory = std::make_shared<InputFactory>(virtual_gamepad);
+        Common::Input::RegisterInputFactory(virtual_gamepad->GetEngineName(),
+                                            virtual_gamepad_input_factory);
+
 #ifdef HAVE_SDL2
         sdl = std::make_shared<SDLDriver>("sdl");
         sdl->SetMappingCallback(mapping_callback);
@@ -132,6 +139,9 @@ struct InputSubsystem::Impl {
         Common::Input::UnregisterOutputFactory(virtual_amiibo->GetEngineName());
         virtual_amiibo.reset();
 
+        Common::Input::UnregisterInputFactory(virtual_gamepad->GetEngineName());
+        virtual_gamepad.reset();
+
 #ifdef HAVE_SDL2
         Common::Input::UnregisterInputFactory(sdl->GetEngineName());
         Common::Input::UnregisterOutputFactory(sdl->GetEngineName());
@@ -290,6 +300,9 @@ struct InputSubsystem::Impl {
         if (engine == tas_input->GetEngineName()) {
             return true;
         }
+        if (engine == virtual_gamepad->GetEngineName()) {
+            return true;
+        }
 #ifdef HAVE_SDL2
         if (engine == sdl->GetEngineName()) {
             return true;
@@ -338,6 +351,7 @@ struct InputSubsystem::Impl {
     std::shared_ptr<CemuhookUDP::UDPClient> udp_client;
     std::shared_ptr<Camera> camera;
     std::shared_ptr<VirtualAmiibo> virtual_amiibo;
+    std::shared_ptr<VirtualGamepad> virtual_gamepad;
 
     std::shared_ptr<InputFactory> keyboard_factory;
     std::shared_ptr<InputFactory> mouse_factory;
@@ -347,6 +361,7 @@ struct InputSubsystem::Impl {
     std::shared_ptr<InputFactory> tas_input_factory;
     std::shared_ptr<InputFactory> camera_input_factory;
     std::shared_ptr<InputFactory> virtual_amiibo_input_factory;
+    std::shared_ptr<InputFactory> virtual_gamepad_input_factory;
 
     std::shared_ptr<OutputFactory> keyboard_output_factory;
     std::shared_ptr<OutputFactory> mouse_output_factory;
@@ -423,6 +438,14 @@ const VirtualAmiibo* InputSubsystem::GetVirtualAmiibo() const {
     return impl->virtual_amiibo.get();
 }
 
+VirtualGamepad* InputSubsystem::GetVirtualGamepad() {
+    return impl->virtual_gamepad.get();
+}
+
+const VirtualGamepad* InputSubsystem::GetVirtualGamepad() const {
+    return impl->virtual_gamepad.get();
+}
+
 std::vector<Common::ParamPackage> InputSubsystem::GetInputDevices() const {
     return impl->GetInputDevices();
 }

src/input_common/main.h

@@ -34,6 +34,7 @@ class Keyboard;
 class Mouse;
 class TouchScreen;
 class VirtualAmiibo;
+class VirtualGamepad;
 struct MappingData;
 } // namespace InputCommon
 
@@ -108,6 +109,12 @@ public:
     /// Retrieves the underlying virtual amiibo input device.
     [[nodiscard]] const VirtualAmiibo* GetVirtualAmiibo() const;
 
+    /// Retrieves the underlying virtual gamepad input device.
+    [[nodiscard]] VirtualGamepad* GetVirtualGamepad();
+
+    /// Retrieves the underlying virtual gamepad input device.
+    [[nodiscard]] const VirtualGamepad* GetVirtualGamepad() const;
+
     /**
      * Returns all available input devices that this Factory can create a new device with.
      * Each returned ParamPackage should have a `display` field used for display, a `engine` field

src/shader_recompiler/backend/spirv/emit_spirv_context_get_set.cpp

@@ -461,7 +461,7 @@ void EmitSetSampleMask(EmitContext& ctx, Id value) {
 }
 
 void EmitSetFragDepth(EmitContext& ctx, Id value) {
-    if (!ctx.runtime_info.convert_depth_mode) {
+    if (!ctx.runtime_info.convert_depth_mode || ctx.profile.support_native_ndc) {
         ctx.OpStore(ctx.frag_depth, value);
         return;
     }

src/shader_recompiler/backend/spirv/emit_spirv_special.cpp

@@ -116,7 +116,8 @@ void EmitPrologue(EmitContext& ctx) {
 }
 
 void EmitEpilogue(EmitContext& ctx) {
-    if (ctx.stage == Stage::VertexB && ctx.runtime_info.convert_depth_mode) {
+    if (ctx.stage == Stage::VertexB && ctx.runtime_info.convert_depth_mode &&
+        !ctx.profile.support_native_ndc) {
         ConvertDepthMode(ctx);
     }
     if (ctx.stage == Stage::Fragment) {
@@ -125,7 +126,7 @@ void EmitEpilogue(EmitContext& ctx) {
 }
 
 void EmitEmitVertex(EmitContext& ctx, const IR::Value& stream) {
-    if (ctx.runtime_info.convert_depth_mode) {
+    if (ctx.runtime_info.convert_depth_mode && !ctx.profile.support_native_ndc) {
         ConvertDepthMode(ctx);
     }
     if (stream.IsImmediate()) {

src/shader_recompiler/backend/spirv/spirv_emit_context.cpp

@@ -1345,8 +1345,10 @@ void EmitContext::DefineInputs(const IR::Program& program) {
     if (info.uses_fswzadd || info.uses_subgroup_invocation_id || info.uses_subgroup_shuffles ||
         (profile.warp_size_potentially_larger_than_guest &&
          (info.uses_subgroup_vote || info.uses_subgroup_mask))) {
+        AddCapability(spv::Capability::GroupNonUniform);
         subgroup_local_invocation_id =
             DefineInput(*this, U32[1], false, spv::BuiltIn::SubgroupLocalInvocationId);
+        Decorate(subgroup_local_invocation_id, spv::Decoration::Flat);
     }
     if (info.uses_fswzadd) {
         const Id f32_one{Const(1.0f)};

src/shader_recompiler/profile.h

@@ -35,6 +35,7 @@ struct Profile {
     bool support_int64_atomics{};
     bool support_derivative_control{};
     bool support_geometry_shader_passthrough{};
+    bool support_native_ndc{};
     bool support_gl_nv_gpu_shader_5{};
     bool support_gl_amd_gpu_shader_half_float{};
     bool support_gl_texture_shadow_lod{};

src/video_core/CMakeLists.txt

@@ -178,6 +178,8 @@ add_library(video_core STATIC
     renderer_vulkan/vk_scheduler.h
     renderer_vulkan/vk_shader_util.cpp
     renderer_vulkan/vk_shader_util.h
+    renderer_vulkan/vk_smaa.cpp
+    renderer_vulkan/vk_smaa.h
     renderer_vulkan/vk_staging_buffer_pool.cpp
     renderer_vulkan/vk_staging_buffer_pool.h
     renderer_vulkan/vk_state_tracker.cpp
@@ -195,6 +197,8 @@ add_library(video_core STATIC
     shader_environment.h
     shader_notify.cpp
     shader_notify.h
+    smaa_area_tex.h
+    smaa_search_tex.h
     surface.cpp
     surface.h
     texture_cache/accelerated_swizzle.cpp

src/video_core/engines/draw_manager.cpp

@@ -189,7 +189,7 @@ void DrawManager::UpdateTopology() {
 }
 
 void DrawManager::ProcessDraw(bool draw_indexed, u32 instance_count) {
-    LOG_TRACE(HW_GPU, "called, topology={}, count={}", draw_state.topology.Value(),
+    LOG_TRACE(HW_GPU, "called, topology={}, count={}", draw_state.topology,
               draw_indexed ? draw_state.index_buffer.count : draw_state.vertex_buffer.count);
 
     UpdateTopology();

src/video_core/gpu.cpp

@@ -223,8 +223,6 @@ struct GPU::Impl {
     /// core timing events.
     void Start() {
         gpu_thread.StartThread(*renderer, renderer->Context(), *scheduler);
-        cpu_context = renderer->GetRenderWindow().CreateSharedContext();
-        cpu_context->MakeCurrent();
     }
 
     void NotifyShutdown() {
@@ -235,6 +233,9 @@ struct GPU::Impl {
 
     /// Obtain the CPU Context
     void ObtainContext() {
+        if (!cpu_context) {
+            cpu_context = renderer->GetRenderWindow().CreateSharedContext();
+        }
         cpu_context->MakeCurrent();
     }
 

src/video_core/host_shaders/CMakeLists.txt

@@ -26,9 +26,16 @@ set(SHADER_FILES
     opengl_present.frag
     opengl_present.vert
     opengl_present_scaleforce.frag
+    opengl_smaa.glsl
     pitch_unswizzle.comp
     present_bicubic.frag
     present_gaussian.frag
+    smaa_edge_detection.vert
+    smaa_edge_detection.frag
+    smaa_blending_weight_calculation.vert
+    smaa_blending_weight_calculation.frag
+    smaa_neighborhood_blending.vert
+    smaa_neighborhood_blending.frag
     vulkan_blit_color_float.frag
     vulkan_blit_depth_stencil.frag
     vulkan_fidelityfx_fsr_easu_fp16.comp

src/video_core/host_shaders/StringShaderHeader.cmake

@@ -11,10 +11,6 @@ string(TOUPPER ${CONTENTS_NAME} CONTENTS_NAME)
 
 FILE(READ ${SOURCE_FILE} line_contents)
 
-# Replace double quotes with single quotes,
-# as double quotes will be used to wrap the lines
-STRING(REGEX REPLACE "\"" "'" line_contents "${line_contents}")
-
 # CMake separates list elements with semicolons, but semicolons
 # are used extensively in the shader code.
 # Replace with a temporary marker, to be reverted later.
@@ -25,7 +21,7 @@ STRING(REGEX REPLACE "\n" ";" line_contents "${line_contents}")
 
 # Build the shader string, wrapping each line in double quotes.
 foreach(line IN LISTS line_contents)
-    string(CONCAT CONTENTS "${CONTENTS}" \"${line}\\n\"\n)
+    string(CONCAT CONTENTS "${CONTENTS}" "R\"(${line}\n)\" ")
 endforeach()
 
 # Revert the original semicolons in the source.

src/video_core/host_shaders/smaa_blending_weight_calculation.frag

@@ -0,0 +1,36 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#version 460
+
+#extension GL_GOOGLE_include_directive : enable
+
+layout (binding = 0) uniform sampler2D edges_tex;
+layout (binding = 1) uniform sampler2D area_tex;
+layout (binding = 2) uniform sampler2D search_tex;
+
+layout (location = 0) in vec2 tex_coord;
+layout (location = 1) in vec2 pix_coord;
+layout (location = 2) in vec4 offset[3];
+
+layout (location = 0) out vec4 frag_color;
+
+vec4 metrics = vec4(1.0 / textureSize(edges_tex, 0), textureSize(edges_tex, 0));
+#define SMAA_RT_METRICS metrics
+#define SMAA_GLSL_4
+#define SMAA_PRESET_ULTRA
+#define SMAA_INCLUDE_VS 0
+#define SMAA_INCLUDE_PS 1
+
+#include "opengl_smaa.glsl"
+
+void main() {
+    frag_color = SMAABlendingWeightCalculationPS(tex_coord,
+                                       pix_coord,
+                                       offset,
+                                       edges_tex,
+                                       area_tex,
+                                       search_tex,
+                                       vec4(0)
+                                       );
+}

src/video_core/host_shaders/smaa_blending_weight_calculation.vert

@@ -0,0 +1,43 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#version 460
+
+#extension GL_GOOGLE_include_directive : enable
+
+#ifdef VULKAN
+#define VERTEX_ID gl_VertexIndex
+#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv
+#define VERTEX_ID gl_VertexID
+#endif
+
+out gl_PerVertex {
+    vec4 gl_Position;
+};
+
+const vec2 vertices[3] =
+    vec2[3](vec2(-1,-1), vec2(3,-1), vec2(-1, 3));
+
+layout (binding = 0) uniform sampler2D edges_tex;
+layout (binding = 1) uniform sampler2D area_tex;
+layout (binding = 2) uniform sampler2D search_tex;
+
+layout (location = 0) out vec2 tex_coord;
+layout (location = 1) out vec2 pix_coord;
+layout (location = 2) out vec4 offset[3];
+
+vec4 metrics = vec4(1.0 / textureSize(edges_tex, 0), textureSize(edges_tex, 0));
+#define SMAA_RT_METRICS metrics
+#define SMAA_GLSL_4
+#define SMAA_PRESET_ULTRA
+#define SMAA_INCLUDE_VS 1
+#define SMAA_INCLUDE_PS 0
+
+#include "opengl_smaa.glsl"
+
+void main() {
+    vec2 vertex = vertices[VERTEX_ID];
+    gl_Position = vec4(vertex, 0.0, 1.0);
+    tex_coord = (vertex + 1.0) / 2.0;
+    SMAABlendingWeightCalculationVS(tex_coord, pix_coord, offset);
+}

src/video_core/host_shaders/smaa_edge_detection.frag

@@ -0,0 +1,26 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#version 460
+
+#extension GL_GOOGLE_include_directive : enable
+
+layout (binding = 0) uniform sampler2D input_tex;
+
+layout (location = 0) in vec2 tex_coord;
+layout (location = 1) in vec4 offset[3];
+
+layout (location = 0) out vec2 frag_color;
+
+vec4 metrics = vec4(1.0 / textureSize(input_tex, 0), textureSize(input_tex, 0));
+#define SMAA_RT_METRICS metrics
+#define SMAA_GLSL_4
+#define SMAA_PRESET_ULTRA
+#define SMAA_INCLUDE_VS 0
+#define SMAA_INCLUDE_PS 1
+
+#include "opengl_smaa.glsl"
+
+void main() {
+    frag_color = SMAAColorEdgeDetectionPS(tex_coord, offset, input_tex);
+}

src/video_core/host_shaders/smaa_edge_detection.vert

@@ -0,0 +1,40 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#version 460
+
+#extension GL_GOOGLE_include_directive : enable
+
+#ifdef VULKAN
+#define VERTEX_ID gl_VertexIndex
+#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv
+#define VERTEX_ID gl_VertexID
+#endif
+
+out gl_PerVertex {
+    vec4 gl_Position;
+};
+
+const vec2 vertices[3] =
+    vec2[3](vec2(-1,-1), vec2(3,-1), vec2(-1, 3));
+
+layout (binding = 0) uniform sampler2D input_tex;
+
+layout (location = 0) out vec2 tex_coord;
+layout (location = 1) out vec4 offset[3];
+
+vec4 metrics = vec4(1.0 / textureSize(input_tex, 0), textureSize(input_tex, 0));
+#define SMAA_RT_METRICS metrics
+#define SMAA_GLSL_4
+#define SMAA_PRESET_ULTRA
+#define SMAA_INCLUDE_VS 1
+#define SMAA_INCLUDE_PS 0
+
+#include "opengl_smaa.glsl"
+
+void main() {
+    vec2 vertex = vertices[VERTEX_ID];
+    gl_Position = vec4(vertex, 0.0, 1.0);
+    tex_coord = (vertex + 1.0) / 2.0;
+    SMAAEdgeDetectionVS(tex_coord, offset);
+}

src/video_core/host_shaders/smaa_neighborhood_blending.frag

@@ -0,0 +1,31 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#version 460
+
+#extension GL_GOOGLE_include_directive : enable
+
+layout (binding = 0) uniform sampler2D input_tex;
+layout (binding = 1) uniform sampler2D blend_tex;
+
+layout (location = 0) in vec2 tex_coord;
+layout (location = 1) in vec4 offset;
+
+layout (location = 0) out vec4 frag_color;
+
+vec4 metrics = vec4(1.0 / textureSize(input_tex, 0), textureSize(input_tex, 0));
+#define SMAA_RT_METRICS metrics
+#define SMAA_GLSL_4
+#define SMAA_PRESET_ULTRA
+#define SMAA_INCLUDE_VS 0
+#define SMAA_INCLUDE_PS 1
+
+#include "opengl_smaa.glsl"
+
+void main() {
+    frag_color = SMAANeighborhoodBlendingPS(tex_coord,
+                                  offset,
+                                  input_tex,
+                                  blend_tex
+                                  );
+}

src/video_core/host_shaders/smaa_neighborhood_blending.vert

@@ -0,0 +1,41 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#version 460
+
+#extension GL_GOOGLE_include_directive : enable
+
+#ifdef VULKAN
+#define VERTEX_ID gl_VertexIndex
+#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv
+#define VERTEX_ID gl_VertexID
+#endif
+
+out gl_PerVertex {
+    vec4 gl_Position;
+};
+
+const vec2 vertices[3] =
+    vec2[3](vec2(-1,-1), vec2(3,-1), vec2(-1, 3));
+
+layout (binding = 0) uniform sampler2D input_tex;
+layout (binding = 1) uniform sampler2D blend_tex;
+
+layout (location = 0) out vec2 tex_coord;
+layout (location = 1) out vec4 offset;
+
+vec4 metrics = vec4(1.0 / textureSize(input_tex, 0), textureSize(input_tex, 0));
+#define SMAA_RT_METRICS metrics
+#define SMAA_GLSL_4
+#define SMAA_PRESET_ULTRA
+#define SMAA_INCLUDE_VS 1
+#define SMAA_INCLUDE_PS 0
+
+#include "opengl_smaa.glsl"
+
+void main() {
+    vec2 vertex = vertices[VERTEX_ID];
+    gl_Position = vec4(vertex, 0.0, 1.0);
+    tex_coord = (vertex + 1.0) / 2.0;
+    SMAANeighborhoodBlendingVS(tex_coord, offset);
+}

src/video_core/renderer_opengl/gl_device.cpp

@@ -112,7 +112,7 @@ bool IsASTCSupported() {
 }
 } // Anonymous namespace
 
-Device::Device() {
+Device::Device(Core::Frontend::EmuWindow& emu_window) {
     if (!GLAD_GL_VERSION_4_6) {
         LOG_ERROR(Render_OpenGL, "OpenGL 4.6 is not available");
         throw std::runtime_error{"Insufficient version"};
@@ -126,9 +126,9 @@ Device::Device() {
     const bool is_intel = vendor_name == "Intel";
 
 #ifdef __unix__
-    const bool is_linux = true;
+    constexpr bool is_linux = true;
 #else
-    const bool is_linux = false;
+    constexpr bool is_linux = false;
 #endif
 
     bool disable_fast_buffer_sub_data = false;
@@ -193,9 +193,11 @@ Device::Device() {
         }
     }
 
+    strict_context_required = emu_window.StrictContextRequired();
     // Blocks AMD and Intel OpenGL drivers on Windows from using asynchronous shader compilation.
+    // Blocks EGL on Wayland from using asynchronous shader compilation.
     use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue() &&
-                               !(is_amd || (is_intel && !is_linux));
+                               !(is_amd || (is_intel && !is_linux)) && !strict_context_required;
     use_driver_cache = is_nvidia;
 
     LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi);

src/video_core/renderer_opengl/gl_device.h

@@ -5,6 +5,7 @@
 
 #include <cstddef>
 #include "common/common_types.h"
+#include "core/frontend/emu_window.h"
 #include "shader_recompiler/stage.h"
 
 namespace Settings {
@@ -15,7 +16,7 @@ namespace OpenGL {
 
 class Device {
 public:
-    explicit Device();
+    explicit Device(Core::Frontend::EmuWindow& emu_window);
 
     [[nodiscard]] std::string GetVendorName() const;
 
@@ -173,6 +174,10 @@ public:
         return can_report_memory;
     }
 
+    bool StrictContextRequired() const {
+        return strict_context_required;
+    }
+
 private:
     static bool TestVariableAoffi();
     static bool TestPreciseBug();
@@ -216,6 +221,7 @@ private:
     bool has_cbuf_ftou_bug{};
     bool has_bool_ref_bug{};
     bool can_report_memory{};
+    bool strict_context_required{};
 
     std::string vendor_name;
 };

src/video_core/renderer_opengl/gl_shader_cache.cpp

@@ -174,6 +174,7 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo
       texture_cache{texture_cache_}, buffer_cache{buffer_cache_}, program_manager{program_manager_},
       state_tracker{state_tracker_}, shader_notify{shader_notify_},
       use_asynchronous_shaders{device.UseAsynchronousShaders()},
+      strict_context_required{device.StrictContextRequired()},
       profile{
           .supported_spirv = 0x00010000,
 
@@ -203,6 +204,7 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo
           .support_int64_atomics = false,
           .support_derivative_control = device.HasDerivativeControl(),
           .support_geometry_shader_passthrough = device.HasGeometryShaderPassthrough(),
+          .support_native_ndc = true,
           .support_gl_nv_gpu_shader_5 = device.HasNvGpuShader5(),
           .support_gl_amd_gpu_shader_half_float = device.HasAmdShaderHalfFloat(),
           .support_gl_texture_shadow_lod = device.HasTextureShadowLod(),
@@ -255,9 +257,14 @@ void ShaderCache::LoadDiskResources(u64 title_id, std::stop_token stop_loading,
     }
     shader_cache_filename = base_dir / "opengl.bin";
 
-    if (!workers) {
+    if (!workers && !strict_context_required) {
         workers = CreateWorkers();
     }
+    std::optional<Context> strict_context;
+    if (strict_context_required) {
+        strict_context.emplace(emu_window);
+    }
+
     struct {
         std::mutex mutex;
         size_t total{};
@@ -265,44 +272,49 @@ void ShaderCache::LoadDiskResources(u64 title_id, std::stop_token stop_loading,
         bool has_loaded{};
     } state;
 
+    const auto queue_work{[&](Common::UniqueFunction<void, Context*>&& work) {
+        if (strict_context_required) {
+            work(&strict_context.value());
+        } else {
+            workers->QueueWork(std::move(work));
+        }
+    }};
     const auto load_compute{[&](std::ifstream& file, FileEnvironment env) {
         ComputePipelineKey key;
         file.read(reinterpret_cast<char*>(&key), sizeof(key));
-        workers->QueueWork(
-            [this, key, env = std::move(env), &state, &callback](Context* ctx) mutable {
-                ctx->pools.ReleaseContents();
-                auto pipeline{CreateComputePipeline(ctx->pools, key, env)};
-                std::scoped_lock lock{state.mutex};
-                if (pipeline) {
-                    compute_cache.emplace(key, std::move(pipeline));
-                }
-                ++state.built;
-                if (state.has_loaded) {
-                    callback(VideoCore::LoadCallbackStage::Build, state.built, state.total);
-                }
-            });
+        queue_work([this, key, env = std::move(env), &state, &callback](Context* ctx) mutable {
+            ctx->pools.ReleaseContents();
+            auto pipeline{CreateComputePipeline(ctx->pools, key, env)};
+            std::scoped_lock lock{state.mutex};
+            if (pipeline) {
+                compute_cache.emplace(key, std::move(pipeline));
+            }
+            ++state.built;
+            if (state.has_loaded) {
+                callback(VideoCore::LoadCallbackStage::Build, state.built, state.total);
+            }
+        });
         ++state.total;
     }};
     const auto load_graphics{[&](std::ifstream& file, std::vector<FileEnvironment> envs) {
         GraphicsPipelineKey key;
         file.read(reinterpret_cast<char*>(&key), sizeof(key));
-        workers->QueueWork(
-            [this, key, envs = std::move(envs), &state, &callback](Context* ctx) mutable {
-                boost::container::static_vector<Shader::Environment*, 5> env_ptrs;
-                for (auto& env : envs) {
-                    env_ptrs.push_back(&env);
-                }
-                ctx->pools.ReleaseContents();
-                auto pipeline{CreateGraphicsPipeline(ctx->pools, key, MakeSpan(env_ptrs), false)};
-                std::scoped_lock lock{state.mutex};
-                if (pipeline) {
-                    graphics_cache.emplace(key, std::move(pipeline));
-                }
-                ++state.built;
-                if (state.has_loaded) {
-                    callback(VideoCore::LoadCallbackStage::Build, state.built, state.total);
-                }
-            });
+        queue_work([this, key, envs = std::move(envs), &state, &callback](Context* ctx) mutable {
+            boost::container::static_vector<Shader::Environment*, 5> env_ptrs;
+            for (auto& env : envs) {
+                env_ptrs.push_back(&env);
+            }
+            ctx->pools.ReleaseContents();
+            auto pipeline{CreateGraphicsPipeline(ctx->pools, key, MakeSpan(env_ptrs), false)};
+            std::scoped_lock lock{state.mutex};
+            if (pipeline) {
+                graphics_cache.emplace(key, std::move(pipeline));
+            }
+            ++state.built;
+            if (state.has_loaded) {
+                callback(VideoCore::LoadCallbackStage::Build, state.built, state.total);
+            }
+        });
         ++state.total;
     }};
     LoadPipelines(stop_loading, shader_cache_filename, CACHE_VERSION, load_compute, load_graphics);
@@ -314,6 +326,9 @@ void ShaderCache::LoadDiskResources(u64 title_id, std::stop_token stop_loading,
     state.has_loaded = true;
     lock.unlock();
 
+    if (strict_context_required) {
+        return;
+    }
     workers->WaitForRequests(stop_loading);
     if (!use_asynchronous_shaders) {
         workers.reset();

src/video_core/renderer_opengl/gl_shader_cache.h

@@ -69,6 +69,7 @@ private:
     StateTracker& state_tracker;
     VideoCore::ShaderNotify& shader_notify;
     const bool use_asynchronous_shaders;
+    const bool strict_context_required;
 
     GraphicsPipelineKey graphics_key{};
     GraphicsPipeline* current_pipeline{};

src/video_core/renderer_opengl/renderer_opengl.cpp

@@ -22,12 +22,21 @@
 #include "video_core/host_shaders/opengl_present_frag.h"
 #include "video_core/host_shaders/opengl_present_scaleforce_frag.h"
 #include "video_core/host_shaders/opengl_present_vert.h"
+#include "video_core/host_shaders/opengl_smaa_glsl.h"
 #include "video_core/host_shaders/present_bicubic_frag.h"
 #include "video_core/host_shaders/present_gaussian_frag.h"
+#include "video_core/host_shaders/smaa_blending_weight_calculation_frag.h"
+#include "video_core/host_shaders/smaa_blending_weight_calculation_vert.h"
+#include "video_core/host_shaders/smaa_edge_detection_frag.h"
+#include "video_core/host_shaders/smaa_edge_detection_vert.h"
+#include "video_core/host_shaders/smaa_neighborhood_blending_frag.h"
+#include "video_core/host_shaders/smaa_neighborhood_blending_vert.h"
 #include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/gl_shader_manager.h"
 #include "video_core/renderer_opengl/gl_shader_util.h"
 #include "video_core/renderer_opengl/renderer_opengl.h"
+#include "video_core/smaa_area_tex.h"
+#include "video_core/smaa_search_tex.h"
 #include "video_core/textures/decoders.h"
 
 namespace OpenGL {
@@ -131,8 +140,8 @@ RendererOpenGL::RendererOpenGL(Core::TelemetrySession& telemetry_session_,
                                Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_,
                                std::unique_ptr<Core::Frontend::GraphicsContext> context_)
     : RendererBase{emu_window_, std::move(context_)}, telemetry_session{telemetry_session_},
-      emu_window{emu_window_}, cpu_memory{cpu_memory_}, gpu{gpu_}, state_tracker{},
-      program_manager{device},
+      emu_window{emu_window_}, cpu_memory{cpu_memory_}, gpu{gpu_}, device{emu_window_},
+      state_tracker{}, program_manager{device},
       rasterizer(emu_window, gpu, cpu_memory, device, screen_info, program_manager, state_tracker) {
     if (Settings::values.renderer_debug && GLAD_GL_KHR_debug) {
         glEnable(GL_DEBUG_OUTPUT);
@@ -258,6 +267,28 @@ void RendererOpenGL::InitOpenGLObjects() {
     // Create shader programs
     fxaa_vertex = CreateProgram(HostShaders::FXAA_VERT, GL_VERTEX_SHADER);
     fxaa_fragment = CreateProgram(HostShaders::FXAA_FRAG, GL_FRAGMENT_SHADER);
+
+    const auto SmaaShader = [](std::string_view specialized_source, GLenum stage) {
+        std::string shader_source{specialized_source};
+        constexpr std::string_view include_string = "#include \"opengl_smaa.glsl\"";
+        const std::size_t pos = shader_source.find(include_string);
+        ASSERT(pos != std::string::npos);
+        shader_source.replace(pos, include_string.size(), HostShaders::OPENGL_SMAA_GLSL);
+        return CreateProgram(shader_source, stage);
+    };
+
+    smaa_edge_detection_vert = SmaaShader(HostShaders::SMAA_EDGE_DETECTION_VERT, GL_VERTEX_SHADER);
+    smaa_edge_detection_frag =
+        SmaaShader(HostShaders::SMAA_EDGE_DETECTION_FRAG, GL_FRAGMENT_SHADER);
+    smaa_blending_weight_calculation_vert =
+        SmaaShader(HostShaders::SMAA_BLENDING_WEIGHT_CALCULATION_VERT, GL_VERTEX_SHADER);
+    smaa_blending_weight_calculation_frag =
+        SmaaShader(HostShaders::SMAA_BLENDING_WEIGHT_CALCULATION_FRAG, GL_FRAGMENT_SHADER);
+    smaa_neighborhood_blending_vert =
+        SmaaShader(HostShaders::SMAA_NEIGHBORHOOD_BLENDING_VERT, GL_VERTEX_SHADER);
+    smaa_neighborhood_blending_frag =
+        SmaaShader(HostShaders::SMAA_NEIGHBORHOOD_BLENDING_FRAG, GL_FRAGMENT_SHADER);
+
     present_vertex = CreateProgram(HostShaders::OPENGL_PRESENT_VERT, GL_VERTEX_SHADER);
     present_bilinear_fragment = CreateProgram(HostShaders::OPENGL_PRESENT_FRAG, GL_FRAGMENT_SHADER);
     present_bicubic_fragment = CreateProgram(HostShaders::PRESENT_BICUBIC_FRAG, GL_FRAGMENT_SHADER);
@@ -293,7 +324,16 @@ void RendererOpenGL::InitOpenGLObjects() {
     // Clear screen to black
     LoadColorToActiveGLTexture(0, 0, 0, 0, screen_info.texture);
 
-    fxaa_framebuffer.Create();
+    aa_framebuffer.Create();
+
+    smaa_area_tex.Create(GL_TEXTURE_2D);
+    glTextureStorage2D(smaa_area_tex.handle, 1, GL_RG8, AREATEX_WIDTH, AREATEX_HEIGHT);
+    glTextureSubImage2D(smaa_area_tex.handle, 0, 0, 0, AREATEX_WIDTH, AREATEX_HEIGHT, GL_RG,
+                        GL_UNSIGNED_BYTE, areaTexBytes);
+    smaa_search_tex.Create(GL_TEXTURE_2D);
+    glTextureStorage2D(smaa_search_tex.handle, 1, GL_R8, SEARCHTEX_WIDTH, SEARCHTEX_HEIGHT);
+    glTextureSubImage2D(smaa_search_tex.handle, 0, 0, 0, SEARCHTEX_WIDTH, SEARCHTEX_HEIGHT, GL_RED,
+                        GL_UNSIGNED_BYTE, searchTexBytes);
 }
 
 void RendererOpenGL::AddTelemetryFields() {
@@ -346,13 +386,22 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture,
     texture.resource.Release();
     texture.resource.Create(GL_TEXTURE_2D);
     glTextureStorage2D(texture.resource.handle, 1, internal_format, texture.width, texture.height);
-    fxaa_texture.Release();
-    fxaa_texture.Create(GL_TEXTURE_2D);
-    glTextureStorage2D(fxaa_texture.handle, 1, GL_RGBA16F,
+    aa_texture.Release();
+    aa_texture.Create(GL_TEXTURE_2D);
+    glTextureStorage2D(aa_texture.handle, 1, GL_RGBA16F,
+                       Settings::values.resolution_info.ScaleUp(screen_info.texture.width),
+                       Settings::values.resolution_info.ScaleUp(screen_info.texture.height));
+    glNamedFramebufferTexture(aa_framebuffer.handle, GL_COLOR_ATTACHMENT0, aa_texture.handle, 0);
+    smaa_edges_tex.Release();
+    smaa_edges_tex.Create(GL_TEXTURE_2D);
+    glTextureStorage2D(smaa_edges_tex.handle, 1, GL_RG16F,
+                       Settings::values.resolution_info.ScaleUp(screen_info.texture.width),
+                       Settings::values.resolution_info.ScaleUp(screen_info.texture.height));
+    smaa_blend_tex.Release();
+    smaa_blend_tex.Create(GL_TEXTURE_2D);
+    glTextureStorage2D(smaa_blend_tex.handle, 1, GL_RGBA16F,
                        Settings::values.resolution_info.ScaleUp(screen_info.texture.width),
                        Settings::values.resolution_info.ScaleUp(screen_info.texture.height));
-    glNamedFramebufferTexture(fxaa_framebuffer.handle, GL_COLOR_ATTACHMENT0, fxaa_texture.handle,
-                              0);
 }
 
 void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) {
@@ -377,11 +426,6 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) {
 
     state_tracker.ClipControl(GL_LOWER_LEFT, GL_ZERO_TO_ONE);
 
-    // Update background color before drawing
-    glClearColor(Settings::values.bg_red.GetValue() / 255.0f,
-                 Settings::values.bg_green.GetValue() / 255.0f,
-                 Settings::values.bg_blue.GetValue() / 255.0f, 1.0f);
-
     glEnable(GL_CULL_FACE);
     glDisable(GL_COLOR_LOGIC_OP);
     glDisable(GL_DEPTH_TEST);
@@ -394,12 +438,12 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) {
     glCullFace(GL_BACK);
     glFrontFace(GL_CW);
     glColorMaski(0, GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE);
+    glDepthRangeIndexed(0, 0.0, 0.0);
 
     glBindTextureUnit(0, screen_info.display_texture);
 
-    if (Settings::values.anti_aliasing.GetValue() == Settings::AntiAliasing::Fxaa) {
-        program_manager.BindPresentPrograms(fxaa_vertex.handle, fxaa_fragment.handle);
-
+    const auto anti_aliasing = Settings::values.anti_aliasing.GetValue();
+    if (anti_aliasing != Settings::AntiAliasing::None) {
         glEnablei(GL_SCISSOR_TEST, 0);
         auto viewport_width = screen_info.texture.width;
         auto scissor_width = framebuffer_crop_rect.GetWidth();
@@ -420,21 +464,60 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) {
         glScissorIndexed(0, 0, 0, scissor_width, scissor_height);
         glViewportIndexedf(0, 0.0f, 0.0f, static_cast<GLfloat>(viewport_width),
                            static_cast<GLfloat>(viewport_height));
-        glDepthRangeIndexed(0, 0.0, 0.0);
 
         glBindSampler(0, present_sampler.handle);
         GLint old_read_fb;
         GLint old_draw_fb;
         glGetIntegerv(GL_READ_FRAMEBUFFER_BINDING, &old_read_fb);
         glGetIntegerv(GL_DRAW_FRAMEBUFFER_BINDING, &old_draw_fb);
-        glBindFramebuffer(GL_DRAW_FRAMEBUFFER, fxaa_framebuffer.handle);
 
-        glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
+        switch (anti_aliasing) {
+        case Settings::AntiAliasing::Fxaa: {
+            program_manager.BindPresentPrograms(fxaa_vertex.handle, fxaa_fragment.handle);
+            glBindFramebuffer(GL_DRAW_FRAMEBUFFER, aa_framebuffer.handle);
+            glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
+        } break;
+        case Settings::AntiAliasing::Smaa: {
+            glClearColor(0, 0, 0, 0);
+            glFrontFace(GL_CCW);
+            glBindFramebuffer(GL_DRAW_FRAMEBUFFER, aa_framebuffer.handle);
+            glBindSampler(1, present_sampler.handle);
+            glBindSampler(2, present_sampler.handle);
+
+            glNamedFramebufferTexture(aa_framebuffer.handle, GL_COLOR_ATTACHMENT0,
+                                      smaa_edges_tex.handle, 0);
+            glClear(GL_COLOR_BUFFER_BIT);
+            program_manager.BindPresentPrograms(smaa_edge_detection_vert.handle,
+                                                smaa_edge_detection_frag.handle);
+            glDrawArrays(GL_TRIANGLES, 0, 3);
+
+            glBindTextureUnit(0, smaa_edges_tex.handle);
+            glBindTextureUnit(1, smaa_area_tex.handle);
+            glBindTextureUnit(2, smaa_search_tex.handle);
+            glNamedFramebufferTexture(aa_framebuffer.handle, GL_COLOR_ATTACHMENT0,
+                                      smaa_blend_tex.handle, 0);
+            glClear(GL_COLOR_BUFFER_BIT);
+            program_manager.BindPresentPrograms(smaa_blending_weight_calculation_vert.handle,
+                                                smaa_blending_weight_calculation_frag.handle);
+            glDrawArrays(GL_TRIANGLES, 0, 3);
+
+            glBindTextureUnit(0, screen_info.display_texture);
+            glBindTextureUnit(1, smaa_blend_tex.handle);
+            glNamedFramebufferTexture(aa_framebuffer.handle, GL_COLOR_ATTACHMENT0,
+                                      aa_texture.handle, 0);
+            program_manager.BindPresentPrograms(smaa_neighborhood_blending_vert.handle,
+                                                smaa_neighborhood_blending_frag.handle);
+            glDrawArrays(GL_TRIANGLES, 0, 3);
+            glFrontFace(GL_CW);
+        } break;
+        default:
+            UNREACHABLE();
+        }
 
         glBindFramebuffer(GL_READ_FRAMEBUFFER, old_read_fb);
         glBindFramebuffer(GL_DRAW_FRAMEBUFFER, old_draw_fb);
 
-        glBindTextureUnit(0, fxaa_texture.handle);
+        glBindTextureUnit(0, aa_texture.handle);
     }
     const std::array ortho_matrix =
         MakeOrthographicMatrix(static_cast<float>(layout.width), static_cast<float>(layout.height));
@@ -551,6 +634,11 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) {
         glBindSampler(0, present_sampler_nn.handle);
     }
 
+    // Update background color before drawing
+    glClearColor(Settings::values.bg_red.GetValue() / 255.0f,
+                 Settings::values.bg_green.GetValue() / 255.0f,
+                 Settings::values.bg_blue.GetValue() / 255.0f, 1.0f);
+
     glClear(GL_COLOR_BUFFER_BIT);
     glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
 

src/video_core/renderer_opengl/renderer_opengl.h

@@ -127,8 +127,19 @@ private:
 
     /// Display information for Switch screen
     ScreenInfo screen_info;
-    OGLTexture fxaa_texture;
-    OGLFramebuffer fxaa_framebuffer;
+    OGLTexture aa_texture;
+    OGLFramebuffer aa_framebuffer;
+
+    OGLProgram smaa_edge_detection_vert;
+    OGLProgram smaa_blending_weight_calculation_vert;
+    OGLProgram smaa_neighborhood_blending_vert;
+    OGLProgram smaa_edge_detection_frag;
+    OGLProgram smaa_blending_weight_calculation_frag;
+    OGLProgram smaa_neighborhood_blending_frag;
+    OGLTexture smaa_area_tex;
+    OGLTexture smaa_search_tex;
+    OGLTexture smaa_edges_tex;
+    OGLTexture smaa_blend_tex;
 
     /// OpenGL framebuffer data
     std::vector<u8> gl_framebuffer_data;

src/video_core/renderer_vulkan/renderer_vulkan.cpp

@@ -139,23 +139,25 @@ void RendererVulkan::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
     RenderScreenshot(*framebuffer, use_accelerated);
 
     bool has_been_recreated = false;
-    const auto recreate_swapchain = [&] {
+    const auto recreate_swapchain = [&](u32 width, u32 height) {
         if (!has_been_recreated) {
             has_been_recreated = true;
             scheduler.Finish();
         }
-        const Layout::FramebufferLayout layout = render_window.GetFramebufferLayout();
-        swapchain.Create(layout.width, layout.height, is_srgb);
+        swapchain.Create(width, height, is_srgb);
     };
-    if (swapchain.NeedsRecreation(is_srgb)) {
-        recreate_swapchain();
+
+    const Layout::FramebufferLayout layout = render_window.GetFramebufferLayout();
+    if (swapchain.NeedsRecreation(is_srgb) || swapchain.GetWidth() != layout.width ||
+        swapchain.GetHeight() != layout.height) {
+        recreate_swapchain(layout.width, layout.height);
     }
     bool is_outdated;
     do {
         swapchain.AcquireNextImage();
         is_outdated = swapchain.IsOutDated();
         if (is_outdated) {
-            recreate_swapchain();
+            recreate_swapchain(layout.width, layout.height);
         }
     } while (is_outdated);
     if (has_been_recreated) {

src/video_core/renderer_vulkan/vk_blit_screen.cpp

@@ -29,6 +29,7 @@
 #include "video_core/renderer_vulkan/vk_fsr.h"
 #include "video_core/renderer_vulkan/vk_scheduler.h"
 #include "video_core/renderer_vulkan/vk_shader_util.h"
+#include "video_core/renderer_vulkan/vk_smaa.h"
 #include "video_core/renderer_vulkan/vk_swapchain.h"
 #include "video_core/surface.h"
 #include "video_core/textures/decoders.h"
@@ -156,6 +157,7 @@ VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
     scheduler.Wait(resource_ticks[image_index]);
     resource_ticks[image_index] = scheduler.CurrentTick();
 
+    VkImage source_image = use_accelerated ? screen_info.image : *raw_images[image_index];
     VkImageView source_image_view =
         use_accelerated ? screen_info.image_view : *raw_image_views[image_index];
 
@@ -242,7 +244,7 @@ VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
     }
 
     const auto anti_alias_pass = Settings::values.anti_aliasing.GetValue();
-    if (use_accelerated && anti_alias_pass != Settings::AntiAliasing::None) {
+    if (use_accelerated && anti_alias_pass == Settings::AntiAliasing::Fxaa) {
         UpdateAADescriptorSet(image_index, source_image_view, false);
         const u32 up_scale = Settings::values.resolution_info.up_scale;
         const u32 down_shift = Settings::values.resolution_info.down_shift;
@@ -340,7 +342,18 @@ VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
         });
         source_image_view = *aa_image_view;
     }
-
+    if (use_accelerated && anti_alias_pass == Settings::AntiAliasing::Smaa) {
+        if (!smaa) {
+            const u32 up_scale = Settings::values.resolution_info.up_scale;
+            const u32 down_shift = Settings::values.resolution_info.down_shift;
+            const VkExtent2D smaa_size{
+                .width = (up_scale * framebuffer.width) >> down_shift,
+                .height = (up_scale * framebuffer.height) >> down_shift,
+            };
+            CreateSMAA(smaa_size);
+        }
+        source_image_view = smaa->Draw(scheduler, image_index, source_image, source_image_view);
+    }
     if (fsr) {
         auto crop_rect = framebuffer.crop_rect;
         if (crop_rect.GetWidth() == 0) {
@@ -467,6 +480,7 @@ void BlitScreen::CreateDynamicResources() {
     CreateFramebuffers();
     CreateGraphicsPipeline();
     fsr.reset();
+    smaa.reset();
     if (Settings::values.scaling_filter.GetValue() == Settings::ScalingFilter::Fsr) {
         CreateFSR();
     }
@@ -490,6 +504,7 @@ void BlitScreen::RefreshResources(const Tegra::FramebufferConfig& framebuffer) {
     raw_height = framebuffer.height;
     pixel_format = framebuffer.pixel_format;
 
+    smaa.reset();
     ReleaseRawImages();
 
     CreateStagingBuffer(framebuffer);
@@ -1448,6 +1463,10 @@ void BlitScreen::SetVertexData(BufferData& data, const Tegra::FramebufferConfig&
         ScreenRectVertex(x + w, y + h, texcoords.bottom * scale_u, left_start + right * scale_v);
 }
 
+void BlitScreen::CreateSMAA(VkExtent2D smaa_size) {
+    smaa = std::make_unique<SMAA>(device, memory_allocator, image_count, smaa_size);
+}
+
 void BlitScreen::CreateFSR() {
     const auto& layout = render_window.GetFramebufferLayout();
     const VkExtent2D fsr_size{

src/video_core/renderer_vulkan/vk_blit_screen.h

@@ -40,9 +40,11 @@ class Device;
 class FSR;
 class RasterizerVulkan;
 class Scheduler;
+class SMAA;
 class Swapchain;
 
 struct ScreenInfo {
+    VkImage image{};
     VkImageView image_view{};
     u32 width{};
     u32 height{};
@@ -101,6 +103,7 @@ private:
     void SetVertexData(BufferData& data, const Tegra::FramebufferConfig& framebuffer,
                        const Layout::FramebufferLayout layout) const;
 
+    void CreateSMAA(VkExtent2D smaa_size);
     void CreateFSR();
 
     u64 CalculateBufferSize(const Tegra::FramebufferConfig& framebuffer) const;
@@ -163,6 +166,7 @@ private:
     Service::android::PixelFormat pixel_format{};
 
     std::unique_ptr<FSR> fsr;
+    std::unique_ptr<SMAA> smaa;
 };
 
 } // namespace Vulkan

src/video_core/renderer_vulkan/vk_buffer_cache.cpp

@@ -285,6 +285,9 @@ void BufferCacheRuntime::BindQuadArrayIndexBuffer(u32 first, u32 count) {
 
 void BufferCacheRuntime::BindVertexBuffer(u32 index, VkBuffer buffer, u32 offset, u32 size,
                                           u32 stride) {
+    if (index >= device.GetMaxVertexInputBindings()) {
+        return;
+    }
     if (device.IsExtExtendedDynamicStateSupported()) {
         scheduler.Record([index, buffer, offset, size, stride](vk::CommandBuffer cmdbuf) {
             const VkDeviceSize vk_offset = buffer != VK_NULL_HANDLE ? offset : 0;

src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp

@@ -529,7 +529,9 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) {
     static_vector<VkVertexInputBindingDivisorDescriptionEXT, 32> vertex_binding_divisors;
     static_vector<VkVertexInputAttributeDescription, 32> vertex_attributes;
     if (key.state.dynamic_vertex_input) {
-        for (size_t index = 0; index < key.state.attributes.size(); ++index) {
+        const size_t num_vertex_arrays = std::min(
+            key.state.attributes.size(), static_cast<size_t>(device.GetMaxVertexInputBindings()));
+        for (size_t index = 0; index < num_vertex_arrays; ++index) {
             const u32 type = key.state.DynamicAttributeType(index);
             if (!stage_infos[0].loads.Generic(index) || type == 0) {
                 continue;
@@ -551,7 +553,9 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) {
             });
         }
     } else {
-        for (size_t index = 0; index < Maxwell::NumVertexArrays; ++index) {
+        const size_t num_vertex_arrays = std::min(
+            Maxwell::NumVertexArrays, static_cast<size_t>(device.GetMaxVertexInputBindings()));
+        for (size_t index = 0; index < num_vertex_arrays; ++index) {
             const bool instanced = key.state.binding_divisors[index] != 0;
             const auto rate =
                 instanced ? VK_VERTEX_INPUT_RATE_INSTANCE : VK_VERTEX_INPUT_RATE_VERTEX;
@@ -580,6 +584,8 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) {
             });
         }
     }
+    ASSERT(vertex_attributes.size() <= device.GetMaxVertexInputAttributes());
+
     VkPipelineVertexInputStateCreateInfo vertex_input_ci{
         .sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO,
         .pNext = nullptr,
@@ -634,23 +640,33 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) {
     };
     std::array<VkViewportSwizzleNV, Maxwell::NumViewports> swizzles;
     std::ranges::transform(key.state.viewport_swizzles, swizzles.begin(), UnpackViewportSwizzle);
-    const VkPipelineViewportSwizzleStateCreateInfoNV swizzle_ci{
+    VkPipelineViewportSwizzleStateCreateInfoNV swizzle_ci{
         .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_SWIZZLE_STATE_CREATE_INFO_NV,
         .pNext = nullptr,
         .flags = 0,
         .viewportCount = Maxwell::NumViewports,
         .pViewportSwizzles = swizzles.data(),
     };
-    const VkPipelineViewportStateCreateInfo viewport_ci{
+    VkPipelineViewportDepthClipControlCreateInfoEXT ndc_info{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_DEPTH_CLIP_CONTROL_CREATE_INFO_EXT,
+        .pNext = nullptr,
+        .negativeOneToOne = key.state.ndc_minus_one_to_one.Value() != 0 ? VK_TRUE : VK_FALSE,
+    };
+    VkPipelineViewportStateCreateInfo viewport_ci{
         .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_STATE_CREATE_INFO,
-        .pNext = device.IsNvViewportSwizzleSupported() ? &swizzle_ci : nullptr,
+        .pNext = nullptr,
         .flags = 0,
         .viewportCount = Maxwell::NumViewports,
         .pViewports = nullptr,
         .scissorCount = Maxwell::NumViewports,
         .pScissors = nullptr,
     };
-
+    if (device.IsNvViewportSwizzleSupported()) {
+        swizzle_ci.pNext = std::exchange(viewport_ci.pNext, &swizzle_ci);
+    }
+    if (device.IsExtDepthClipControlSupported()) {
+        ndc_info.pNext = std::exchange(viewport_ci.pNext, &ndc_info);
+    }
     VkPipelineRasterizationStateCreateInfo rasterization_ci{
         .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO,
         .pNext = nullptr,

src/video_core/renderer_vulkan/vk_pipeline_cache.cpp

@@ -321,6 +321,7 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device
         .support_int64_atomics = device.IsExtShaderAtomicInt64Supported(),
         .support_derivative_control = true,
         .support_geometry_shader_passthrough = device.IsNvGeometryShaderPassthroughSupported(),
+        .support_native_ndc = device.IsExtDepthClipControlSupported(),
 
         .warp_size_potentially_larger_than_guest = device.IsWarpSizePotentiallyBiggerThanGuest(),
 
@@ -341,6 +342,15 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device
         .support_snorm_render_buffer = true,
         .support_viewport_index_layer = device.IsExtShaderViewportIndexLayerSupported(),
     };
+
+    if (device.GetMaxVertexInputAttributes() < Maxwell::NumVertexAttributes) {
+        LOG_WARNING(Render_Vulkan, "maxVertexInputAttributes is too low: {} < {}",
+                    device.GetMaxVertexInputAttributes(), Maxwell::NumVertexAttributes);
+    }
+    if (device.GetMaxVertexInputBindings() < Maxwell::NumVertexArrays) {
+        LOG_WARNING(Render_Vulkan, "maxVertexInputBindings is too low: {} < {}",
+                    device.GetMaxVertexInputBindings(), Maxwell::NumVertexArrays);
+    }
 }
 
 PipelineCache::~PipelineCache() = default;

src/video_core/renderer_vulkan/vk_rasterizer.cpp

@@ -580,6 +580,7 @@ bool RasterizerVulkan::AccelerateDisplay(const Tegra::FramebufferConfig& config,
     if (!image_view) {
         return false;
     }
+    screen_info.image = image_view->ImageHandle();
     screen_info.image_view = image_view->Handle(Shader::TextureType::Color2D);
     screen_info.width = image_view->size.width;
     screen_info.height = image_view->size.height;
@@ -657,8 +658,7 @@ void RasterizerVulkan::BeginTransformFeedback() {
         return;
     }
     UNIMPLEMENTED_IF(regs.IsShaderConfigEnabled(Maxwell::ShaderType::TessellationInit) ||
-                     regs.IsShaderConfigEnabled(Maxwell::ShaderType::Tessellation) ||
-                     regs.IsShaderConfigEnabled(Maxwell::ShaderType::Geometry));
+                     regs.IsShaderConfigEnabled(Maxwell::ShaderType::Tessellation));
     scheduler.Record(
         [](vk::CommandBuffer cmdbuf) { cmdbuf.BeginTransformFeedbackEXT(0, 0, nullptr, nullptr); });
 }

src/video_core/renderer_vulkan/vk_smaa.cpp

@@ -0,0 +1,761 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include <list>
+
+#include "common/assert.h"
+#include "common/polyfill_ranges.h"
+
+#include "video_core/renderer_vulkan/vk_scheduler.h"
+#include "video_core/renderer_vulkan/vk_shader_util.h"
+#include "video_core/renderer_vulkan/vk_smaa.h"
+#include "video_core/smaa_area_tex.h"
+#include "video_core/smaa_search_tex.h"
+#include "video_core/vulkan_common/vulkan_device.h"
+
+#include "video_core/host_shaders/smaa_blending_weight_calculation_frag_spv.h"
+#include "video_core/host_shaders/smaa_blending_weight_calculation_vert_spv.h"
+#include "video_core/host_shaders/smaa_edge_detection_frag_spv.h"
+#include "video_core/host_shaders/smaa_edge_detection_vert_spv.h"
+#include "video_core/host_shaders/smaa_neighborhood_blending_frag_spv.h"
+#include "video_core/host_shaders/smaa_neighborhood_blending_vert_spv.h"
+
+namespace Vulkan {
+namespace {
+
+#define ARRAY_TO_SPAN(a) std::span(a, (sizeof(a) / sizeof(a[0])))
+
+std::pair<vk::Image, MemoryCommit> CreateWrappedImage(const Device& device,
+                                                      MemoryAllocator& allocator,
+                                                      VkExtent2D dimensions, VkFormat format) {
+    const VkImageCreateInfo image_ci{
+        .sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .imageType = VK_IMAGE_TYPE_2D,
+        .format = format,
+        .extent = {.width = dimensions.width, .height = dimensions.height, .depth = 1},
+        .mipLevels = 1,
+        .arrayLayers = 1,
+        .samples = VK_SAMPLE_COUNT_1_BIT,
+        .tiling = VK_IMAGE_TILING_OPTIMAL,
+        .usage = VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_STORAGE_BIT |
+                 VK_IMAGE_USAGE_SAMPLED_BIT | VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT,
+        .sharingMode = VK_SHARING_MODE_EXCLUSIVE,
+        .queueFamilyIndexCount = 0,
+        .pQueueFamilyIndices = nullptr,
+        .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED,
+    };
+
+    auto image = device.GetLogical().CreateImage(image_ci);
+    auto commit = allocator.Commit(image, Vulkan::MemoryUsage::DeviceLocal);
+
+    return std::make_pair(std::move(image), std::move(commit));
+}
+
+void TransitionImageLayout(vk::CommandBuffer& cmdbuf, VkImage image, VkImageLayout target_layout,
+                           VkImageLayout source_layout = VK_IMAGE_LAYOUT_GENERAL) {
+    constexpr VkFlags flags{VK_ACCESS_COLOR_ATTACHMENT_READ_BIT |
+                            VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | VK_ACCESS_SHADER_READ_BIT};
+    const VkImageMemoryBarrier barrier{
+        .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
+        .pNext = nullptr,
+        .srcAccessMask = flags,
+        .dstAccessMask = flags,
+        .oldLayout = source_layout,
+        .newLayout = target_layout,
+        .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+        .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+        .image = image,
+        .subresourceRange{
+            .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+            .baseMipLevel = 0,
+            .levelCount = 1,
+            .baseArrayLayer = 0,
+            .layerCount = 1,
+        },
+    };
+    cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_ALL_COMMANDS_BIT,
+                           0, barrier);
+}
+
+void UploadImage(const Device& device, MemoryAllocator& allocator, Scheduler& scheduler,
+                 vk::Image& image, VkExtent2D dimensions, VkFormat format,
+                 std::span<const u8> initial_contents = {}) {
+    auto upload_buffer = device.GetLogical().CreateBuffer(VkBufferCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .size = initial_contents.size_bytes(),
+        .usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT,
+        .sharingMode = VK_SHARING_MODE_EXCLUSIVE,
+        .queueFamilyIndexCount = 0,
+        .pQueueFamilyIndices = nullptr,
+    });
+    auto upload_commit = allocator.Commit(upload_buffer, MemoryUsage::Upload);
+    std::ranges::copy(initial_contents, upload_commit.Map().begin());
+
+    const std::array<VkBufferImageCopy, 1> regions{{{
+        .bufferOffset = 0,
+        .bufferRowLength = dimensions.width,
+        .bufferImageHeight = dimensions.height,
+        .imageSubresource{.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+                          .mipLevel = 0,
+                          .baseArrayLayer = 0,
+                          .layerCount = 1},
+        .imageOffset{},
+        .imageExtent{.width = dimensions.width, .height = dimensions.height, .depth = 1},
+    }}};
+
+    scheduler.RequestOutsideRenderPassOperationContext();
+    scheduler.Record([&](vk::CommandBuffer cmdbuf) {
+        TransitionImageLayout(cmdbuf, *image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL,
+                              VK_IMAGE_LAYOUT_UNDEFINED);
+        cmdbuf.CopyBufferToImage(*upload_buffer, *image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL,
+                                 regions);
+        TransitionImageLayout(cmdbuf, *image, VK_IMAGE_LAYOUT_GENERAL,
+                              VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL);
+    });
+    scheduler.Finish();
+
+    // This should go out of scope before the commit
+    auto upload_buffer2 = std::move(upload_buffer);
+}
+
+vk::ImageView CreateWrappedImageView(const Device& device, vk::Image& image, VkFormat format) {
+    return device.GetLogical().CreateImageView(VkImageViewCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .image = *image,
+        .viewType = VK_IMAGE_VIEW_TYPE_2D,
+        .format = format,
+        .components{},
+        .subresourceRange{.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+                          .baseMipLevel = 0,
+                          .levelCount = 1,
+                          .baseArrayLayer = 0,
+                          .layerCount = 1},
+    });
+}
+
+vk::RenderPass CreateWrappedRenderPass(const Device& device, VkFormat format) {
+    const VkAttachmentDescription attachment{
+        .flags = VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT,
+        .format = format,
+        .samples = VK_SAMPLE_COUNT_1_BIT,
+        .loadOp = VK_ATTACHMENT_LOAD_OP_LOAD,
+        .storeOp = VK_ATTACHMENT_STORE_OP_STORE,
+        .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_LOAD,
+        .stencilStoreOp = VK_ATTACHMENT_STORE_OP_STORE,
+        .initialLayout = VK_IMAGE_LAYOUT_GENERAL,
+        .finalLayout = VK_IMAGE_LAYOUT_GENERAL,
+    };
+
+    constexpr VkAttachmentReference color_attachment_ref{
+        .attachment = 0,
+        .layout = VK_IMAGE_LAYOUT_GENERAL,
+    };
+
+    const VkSubpassDescription subpass_description{
+        .flags = 0,
+        .pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS,
+        .inputAttachmentCount = 0,
+        .pInputAttachments = nullptr,
+        .colorAttachmentCount = 1,
+        .pColorAttachments = &color_attachment_ref,
+        .pResolveAttachments = nullptr,
+        .pDepthStencilAttachment = nullptr,
+        .preserveAttachmentCount = 0,
+        .pPreserveAttachments = nullptr,
+    };
+
+    constexpr VkSubpassDependency dependency{
+        .srcSubpass = VK_SUBPASS_EXTERNAL,
+        .dstSubpass = 0,
+        .srcStageMask = VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT,
+        .dstStageMask = VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT,
+        .srcAccessMask = 0,
+        .dstAccessMask = VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT,
+        .dependencyFlags = 0,
+    };
+
+    return device.GetLogical().CreateRenderPass(VkRenderPassCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .attachmentCount = 1,
+        .pAttachments = &attachment,
+        .subpassCount = 1,
+        .pSubpasses = &subpass_description,
+        .dependencyCount = 1,
+        .pDependencies = &dependency,
+    });
+}
+
+vk::Framebuffer CreateWrappedFramebuffer(const Device& device, vk::RenderPass& render_pass,
+                                         vk::ImageView& dest_image, VkExtent2D extent) {
+    return device.GetLogical().CreateFramebuffer(VkFramebufferCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .renderPass = *render_pass,
+        .attachmentCount = 1,
+        .pAttachments = dest_image.address(),
+        .width = extent.width,
+        .height = extent.height,
+        .layers = 1,
+    });
+}
+
+vk::Sampler CreateWrappedSampler(const Device& device) {
+    return device.GetLogical().CreateSampler(VkSamplerCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .magFilter = VK_FILTER_LINEAR,
+        .minFilter = VK_FILTER_LINEAR,
+        .mipmapMode = VK_SAMPLER_MIPMAP_MODE_LINEAR,
+        .addressModeU = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_EDGE,
+        .addressModeV = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_EDGE,
+        .addressModeW = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_EDGE,
+        .mipLodBias = 0.0f,
+        .anisotropyEnable = VK_FALSE,
+        .maxAnisotropy = 0.0f,
+        .compareEnable = VK_FALSE,
+        .compareOp = VK_COMPARE_OP_NEVER,
+        .minLod = 0.0f,
+        .maxLod = 0.0f,
+        .borderColor = VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK,
+        .unnormalizedCoordinates = VK_FALSE,
+    });
+}
+
+vk::ShaderModule CreateWrappedShaderModule(const Device& device, std::span<const u32> code) {
+    return device.GetLogical().CreateShaderModule(VkShaderModuleCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .codeSize = code.size_bytes(),
+        .pCode = code.data(),
+    });
+}
+
+vk::DescriptorPool CreateWrappedDescriptorPool(const Device& device, u32 max_descriptors,
+                                               u32 max_sets) {
+    const VkDescriptorPoolSize pool_size{
+        .type = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER,
+        .descriptorCount = static_cast<u32>(max_descriptors),
+    };
+
+    return device.GetLogical().CreateDescriptorPool(VkDescriptorPoolCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .maxSets = max_sets,
+        .poolSizeCount = 1,
+        .pPoolSizes = &pool_size,
+    });
+}
+
+vk::DescriptorSetLayout CreateWrappedDescriptorSetLayout(const Device& device,
+                                                         u32 max_sampler_bindings) {
+    std::vector<VkDescriptorSetLayoutBinding> bindings(max_sampler_bindings);
+    for (u32 i = 0; i < max_sampler_bindings; i++) {
+        bindings[i] = {
+            .binding = i,
+            .descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER,
+            .descriptorCount = 1,
+            .stageFlags = VK_SHADER_STAGE_VERTEX_BIT | VK_SHADER_STAGE_FRAGMENT_BIT,
+            .pImmutableSamplers = nullptr,
+        };
+    }
+
+    return device.GetLogical().CreateDescriptorSetLayout(VkDescriptorSetLayoutCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .bindingCount = static_cast<u32>(bindings.size()),
+        .pBindings = bindings.data(),
+    });
+}
+
+vk::DescriptorSets CreateWrappedDescriptorSets(vk::DescriptorPool& pool,
+                                               vk::Span<VkDescriptorSetLayout> layouts) {
+    return pool.Allocate(VkDescriptorSetAllocateInfo{
+        .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO,
+        .pNext = nullptr,
+        .descriptorPool = *pool,
+        .descriptorSetCount = layouts.size(),
+        .pSetLayouts = layouts.data(),
+    });
+}
+
+vk::PipelineLayout CreateWrappedPipelineLayout(const Device& device,
+                                               vk::DescriptorSetLayout& layout) {
+    return device.GetLogical().CreatePipelineLayout(VkPipelineLayoutCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .setLayoutCount = 1,
+        .pSetLayouts = layout.address(),
+        .pushConstantRangeCount = 0,
+        .pPushConstantRanges = nullptr,
+    });
+}
+
+vk::Pipeline CreateWrappedPipeline(const Device& device, vk::RenderPass& renderpass,
+                                   vk::PipelineLayout& layout,
+                                   std::tuple<vk::ShaderModule&, vk::ShaderModule&> shaders) {
+    const std::array<VkPipelineShaderStageCreateInfo, 2> shader_stages{{
+        {
+            .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO,
+            .pNext = nullptr,
+            .flags = 0,
+            .stage = VK_SHADER_STAGE_VERTEX_BIT,
+            .module = *std::get<0>(shaders),
+            .pName = "main",
+            .pSpecializationInfo = nullptr,
+        },
+        {
+            .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO,
+            .pNext = nullptr,
+            .flags = 0,
+            .stage = VK_SHADER_STAGE_FRAGMENT_BIT,
+            .module = *std::get<1>(shaders),
+            .pName = "main",
+            .pSpecializationInfo = nullptr,
+        },
+    }};
+
+    constexpr VkPipelineVertexInputStateCreateInfo vertex_input_ci{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .vertexBindingDescriptionCount = 0,
+        .pVertexBindingDescriptions = nullptr,
+        .vertexAttributeDescriptionCount = 0,
+        .pVertexAttributeDescriptions = nullptr,
+    };
+
+    constexpr VkPipelineInputAssemblyStateCreateInfo input_assembly_ci{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .topology = VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP,
+        .primitiveRestartEnable = VK_FALSE,
+    };
+
+    constexpr VkPipelineViewportStateCreateInfo viewport_state_ci{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_STATE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .viewportCount = 1,
+        .pViewports = nullptr,
+        .scissorCount = 1,
+        .pScissors = nullptr,
+    };
+
+    constexpr VkPipelineRasterizationStateCreateInfo rasterization_ci{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .depthClampEnable = VK_FALSE,
+        .rasterizerDiscardEnable = VK_FALSE,
+        .polygonMode = VK_POLYGON_MODE_FILL,
+        .cullMode = VK_CULL_MODE_NONE,
+        .frontFace = VK_FRONT_FACE_CLOCKWISE,
+        .depthBiasEnable = VK_FALSE,
+        .depthBiasConstantFactor = 0.0f,
+        .depthBiasClamp = 0.0f,
+        .depthBiasSlopeFactor = 0.0f,
+        .lineWidth = 1.0f,
+    };
+
+    constexpr VkPipelineMultisampleStateCreateInfo multisampling_ci{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .rasterizationSamples = VK_SAMPLE_COUNT_1_BIT,
+        .sampleShadingEnable = VK_FALSE,
+        .minSampleShading = 0.0f,
+        .pSampleMask = nullptr,
+        .alphaToCoverageEnable = VK_FALSE,
+        .alphaToOneEnable = VK_FALSE,
+    };
+
+    constexpr VkPipelineColorBlendAttachmentState color_blend_attachment{
+        .blendEnable = VK_FALSE,
+        .srcColorBlendFactor = VK_BLEND_FACTOR_ZERO,
+        .dstColorBlendFactor = VK_BLEND_FACTOR_ZERO,
+        .colorBlendOp = VK_BLEND_OP_ADD,
+        .srcAlphaBlendFactor = VK_BLEND_FACTOR_ZERO,
+        .dstAlphaBlendFactor = VK_BLEND_FACTOR_ZERO,
+        .alphaBlendOp = VK_BLEND_OP_ADD,
+        .colorWriteMask = VK_COLOR_COMPONENT_R_BIT | VK_COLOR_COMPONENT_G_BIT |
+                          VK_COLOR_COMPONENT_B_BIT | VK_COLOR_COMPONENT_A_BIT,
+    };
+
+    const VkPipelineColorBlendStateCreateInfo color_blend_ci{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .logicOpEnable = VK_FALSE,
+        .logicOp = VK_LOGIC_OP_COPY,
+        .attachmentCount = 1,
+        .pAttachments = &color_blend_attachment,
+        .blendConstants = {0.0f, 0.0f, 0.0f, 0.0f},
+    };
+
+    constexpr std::array dynamic_states{
+        VK_DYNAMIC_STATE_VIEWPORT,
+        VK_DYNAMIC_STATE_SCISSOR,
+    };
+
+    const VkPipelineDynamicStateCreateInfo dynamic_state_ci{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_DYNAMIC_STATE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .dynamicStateCount = static_cast<u32>(dynamic_states.size()),
+        .pDynamicStates = dynamic_states.data(),
+    };
+
+    return device.GetLogical().CreateGraphicsPipeline(VkGraphicsPipelineCreateInfo{
+        .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .stageCount = static_cast<u32>(shader_stages.size()),
+        .pStages = shader_stages.data(),
+        .pVertexInputState = &vertex_input_ci,
+        .pInputAssemblyState = &input_assembly_ci,
+        .pTessellationState = nullptr,
+        .pViewportState = &viewport_state_ci,
+        .pRasterizationState = &rasterization_ci,
+        .pMultisampleState = &multisampling_ci,
+        .pDepthStencilState = nullptr,
+        .pColorBlendState = &color_blend_ci,
+        .pDynamicState = &dynamic_state_ci,
+        .layout = *layout,
+        .renderPass = *renderpass,
+        .subpass = 0,
+        .basePipelineHandle = 0,
+        .basePipelineIndex = 0,
+    });
+}
+
+VkWriteDescriptorSet CreateWriteDescriptorSet(std::vector<VkDescriptorImageInfo>& images,
+                                              VkSampler sampler, VkImageView view,
+                                              VkDescriptorSet set, u32 binding) {
+    ASSERT(images.capacity() > images.size());
+    auto& image_info = images.emplace_back(VkDescriptorImageInfo{
+        .sampler = sampler,
+        .imageView = view,
+        .imageLayout = VK_IMAGE_LAYOUT_GENERAL,
+    });
+
+    return VkWriteDescriptorSet{
+        .sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET,
+        .pNext = nullptr,
+        .dstSet = set,
+        .dstBinding = binding,
+        .dstArrayElement = 0,
+        .descriptorCount = 1,
+        .descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER,
+        .pImageInfo = &image_info,
+        .pBufferInfo = nullptr,
+        .pTexelBufferView = nullptr,
+    };
+}
+
+void ClearColorImage(vk::CommandBuffer& cmdbuf, VkImage image) {
+    constexpr std::array<VkImageSubresourceRange, 1> subresources{{{
+        .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+        .baseMipLevel = 0,
+        .levelCount = 1,
+        .baseArrayLayer = 0,
+        .layerCount = 1,
+    }}};
+    TransitionImageLayout(cmdbuf, image, VK_IMAGE_LAYOUT_GENERAL, VK_IMAGE_LAYOUT_UNDEFINED);
+    cmdbuf.ClearColorImage(image, VK_IMAGE_LAYOUT_GENERAL, {}, subresources);
+}
+
+void BeginRenderPass(vk::CommandBuffer& cmdbuf, vk::RenderPass& render_pass,
+                     VkFramebuffer framebuffer, VkExtent2D extent) {
+    const VkRenderPassBeginInfo renderpass_bi{
+        .sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO,
+        .pNext = nullptr,
+        .renderPass = *render_pass,
+        .framebuffer = framebuffer,
+        .renderArea{
+            .offset{},
+            .extent = extent,
+        },
+        .clearValueCount = 0,
+        .pClearValues = nullptr,
+    };
+    cmdbuf.BeginRenderPass(renderpass_bi, VK_SUBPASS_CONTENTS_INLINE);
+
+    const VkViewport viewport{
+        .x = 0.0f,
+        .y = 0.0f,
+        .width = static_cast<float>(extent.width),
+        .height = static_cast<float>(extent.height),
+        .minDepth = 0.0f,
+        .maxDepth = 1.0f,
+    };
+    const VkRect2D scissor{
+        .offset = {0, 0},
+        .extent = extent,
+    };
+    cmdbuf.SetViewport(0, viewport);
+    cmdbuf.SetScissor(0, scissor);
+}
+
+} // Anonymous namespace
+
+SMAA::SMAA(const Device& device, MemoryAllocator& allocator, size_t image_count, VkExtent2D extent)
+    : m_device(device), m_allocator(allocator), m_extent(extent),
+      m_image_count(static_cast<u32>(image_count)) {
+    CreateImages();
+    CreateRenderPasses();
+    CreateSampler();
+    CreateShaders();
+    CreateDescriptorPool();
+    CreateDescriptorSetLayouts();
+    CreateDescriptorSets();
+    CreatePipelineLayouts();
+    CreatePipelines();
+}
+
+void SMAA::CreateImages() {
+    constexpr VkExtent2D area_extent{AREATEX_WIDTH, AREATEX_HEIGHT};
+    constexpr VkExtent2D search_extent{SEARCHTEX_WIDTH, SEARCHTEX_HEIGHT};
+
+    std::tie(m_static_images[Area], m_static_buffer_commits[Area]) =
+        CreateWrappedImage(m_device, m_allocator, area_extent, VK_FORMAT_R8G8_UNORM);
+    std::tie(m_static_images[Search], m_static_buffer_commits[Search]) =
+        CreateWrappedImage(m_device, m_allocator, search_extent, VK_FORMAT_R8_UNORM);
+
+    m_static_image_views[Area] =
+        CreateWrappedImageView(m_device, m_static_images[Area], VK_FORMAT_R8G8_UNORM);
+    m_static_image_views[Search] =
+        CreateWrappedImageView(m_device, m_static_images[Search], VK_FORMAT_R8_UNORM);
+
+    for (u32 i = 0; i < m_image_count; i++) {
+        Images& images = m_dynamic_images.emplace_back();
+
+        std::tie(images.images[Blend], images.buffer_commits[Blend]) =
+            CreateWrappedImage(m_device, m_allocator, m_extent, VK_FORMAT_R16G16B16A16_SFLOAT);
+        std::tie(images.images[Edges], images.buffer_commits[Edges]) =
+            CreateWrappedImage(m_device, m_allocator, m_extent, VK_FORMAT_R16G16_SFLOAT);
+        std::tie(images.images[Output], images.buffer_commits[Output]) =
+            CreateWrappedImage(m_device, m_allocator, m_extent, VK_FORMAT_R16G16B16A16_SFLOAT);
+
+        images.image_views[Blend] =
+            CreateWrappedImageView(m_device, images.images[Blend], VK_FORMAT_R16G16B16A16_SFLOAT);
+        images.image_views[Edges] =
+            CreateWrappedImageView(m_device, images.images[Edges], VK_FORMAT_R16G16_SFLOAT);
+        images.image_views[Output] =
+            CreateWrappedImageView(m_device, images.images[Output], VK_FORMAT_R16G16B16A16_SFLOAT);
+    }
+}
+
+void SMAA::CreateRenderPasses() {
+    m_renderpasses[EdgeDetection] = CreateWrappedRenderPass(m_device, VK_FORMAT_R16G16_SFLOAT);
+    m_renderpasses[BlendingWeightCalculation] =
+        CreateWrappedRenderPass(m_device, VK_FORMAT_R16G16B16A16_SFLOAT);
+    m_renderpasses[NeighborhoodBlending] =
+        CreateWrappedRenderPass(m_device, VK_FORMAT_R16G16B16A16_SFLOAT);
+
+    for (auto& images : m_dynamic_images) {
+        images.framebuffers[EdgeDetection] = CreateWrappedFramebuffer(
+            m_device, m_renderpasses[EdgeDetection], images.image_views[Edges], m_extent);
+
+        images.framebuffers[BlendingWeightCalculation] =
+            CreateWrappedFramebuffer(m_device, m_renderpasses[BlendingWeightCalculation],
+                                     images.image_views[Blend], m_extent);
+
+        images.framebuffers[NeighborhoodBlending] = CreateWrappedFramebuffer(
+            m_device, m_renderpasses[NeighborhoodBlending], images.image_views[Output], m_extent);
+    }
+}
+
+void SMAA::CreateSampler() {
+    m_sampler = CreateWrappedSampler(m_device);
+}
+
+void SMAA::CreateShaders() {
+    // These match the order of the SMAAStage enum
+    constexpr std::array vert_shader_sources{
+        ARRAY_TO_SPAN(SMAA_EDGE_DETECTION_VERT_SPV),
+        ARRAY_TO_SPAN(SMAA_BLENDING_WEIGHT_CALCULATION_VERT_SPV),
+        ARRAY_TO_SPAN(SMAA_NEIGHBORHOOD_BLENDING_VERT_SPV),
+    };
+    constexpr std::array frag_shader_sources{
+        ARRAY_TO_SPAN(SMAA_EDGE_DETECTION_FRAG_SPV),
+        ARRAY_TO_SPAN(SMAA_BLENDING_WEIGHT_CALCULATION_FRAG_SPV),
+        ARRAY_TO_SPAN(SMAA_NEIGHBORHOOD_BLENDING_FRAG_SPV),
+    };
+
+    for (size_t i = 0; i < MaxSMAAStage; i++) {
+        m_vertex_shaders[i] = CreateWrappedShaderModule(m_device, vert_shader_sources[i]);
+        m_fragment_shaders[i] = CreateWrappedShaderModule(m_device, frag_shader_sources[i]);
+    }
+}
+
+void SMAA::CreateDescriptorPool() {
+    // Edge detection: 1 descriptor
+    // Blending weight calculation: 3 descriptors
+    // Neighborhood blending: 2 descriptors
+
+    // 6 descriptors, 3 descriptor sets per image
+    m_descriptor_pool = CreateWrappedDescriptorPool(m_device, 6 * m_image_count, 3 * m_image_count);
+}
+
+void SMAA::CreateDescriptorSetLayouts() {
+    m_descriptor_set_layouts[EdgeDetection] = CreateWrappedDescriptorSetLayout(m_device, 1);
+    m_descriptor_set_layouts[BlendingWeightCalculation] =
+        CreateWrappedDescriptorSetLayout(m_device, 3);
+    m_descriptor_set_layouts[NeighborhoodBlending] = CreateWrappedDescriptorSetLayout(m_device, 2);
+}
+
+void SMAA::CreateDescriptorSets() {
+    std::vector<VkDescriptorSetLayout> layouts(m_descriptor_set_layouts.size());
+    std::ranges::transform(m_descriptor_set_layouts, layouts.begin(),
+                           [](auto& layout) { return *layout; });
+
+    for (auto& images : m_dynamic_images) {
+        images.descriptor_sets = CreateWrappedDescriptorSets(m_descriptor_pool, layouts);
+    }
+}
+
+void SMAA::CreatePipelineLayouts() {
+    for (size_t i = 0; i < MaxSMAAStage; i++) {
+        m_pipeline_layouts[i] = CreateWrappedPipelineLayout(m_device, m_descriptor_set_layouts[i]);
+    }
+}
+
+void SMAA::CreatePipelines() {
+    for (size_t i = 0; i < MaxSMAAStage; i++) {
+        m_pipelines[i] =
+            CreateWrappedPipeline(m_device, m_renderpasses[i], m_pipeline_layouts[i],
+                                  std::tie(m_vertex_shaders[i], m_fragment_shaders[i]));
+    }
+}
+
+void SMAA::UpdateDescriptorSets(VkImageView image_view, size_t image_index) {
+    Images& images = m_dynamic_images[image_index];
+    std::vector<VkDescriptorImageInfo> image_infos;
+    std::vector<VkWriteDescriptorSet> updates;
+    image_infos.reserve(6);
+
+    updates.push_back(CreateWriteDescriptorSet(image_infos, *m_sampler, image_view,
+                                               images.descriptor_sets[EdgeDetection], 0));
+
+    updates.push_back(CreateWriteDescriptorSet(image_infos, *m_sampler, *images.image_views[Edges],
+                                               images.descriptor_sets[BlendingWeightCalculation],
+                                               0));
+    updates.push_back(CreateWriteDescriptorSet(image_infos, *m_sampler, *m_static_image_views[Area],
+                                               images.descriptor_sets[BlendingWeightCalculation],
+                                               1));
+    updates.push_back(
+        CreateWriteDescriptorSet(image_infos, *m_sampler, *m_static_image_views[Search],
+                                 images.descriptor_sets[BlendingWeightCalculation], 2));
+
+    updates.push_back(CreateWriteDescriptorSet(image_infos, *m_sampler, image_view,
+                                               images.descriptor_sets[NeighborhoodBlending], 0));
+    updates.push_back(CreateWriteDescriptorSet(image_infos, *m_sampler, *images.image_views[Blend],
+                                               images.descriptor_sets[NeighborhoodBlending], 1));
+
+    m_device.GetLogical().UpdateDescriptorSets(updates, {});
+}
+
+void SMAA::UploadImages(Scheduler& scheduler) {
+    if (m_images_ready) {
+        return;
+    }
+
+    constexpr VkExtent2D area_extent{AREATEX_WIDTH, AREATEX_HEIGHT};
+    constexpr VkExtent2D search_extent{SEARCHTEX_WIDTH, SEARCHTEX_HEIGHT};
+
+    UploadImage(m_device, m_allocator, scheduler, m_static_images[Area], area_extent,
+                VK_FORMAT_R8G8_UNORM, ARRAY_TO_SPAN(areaTexBytes));
+    UploadImage(m_device, m_allocator, scheduler, m_static_images[Search], search_extent,
+                VK_FORMAT_R8_UNORM, ARRAY_TO_SPAN(searchTexBytes));
+
+    scheduler.Record([&](vk::CommandBuffer& cmdbuf) {
+        for (auto& images : m_dynamic_images) {
+            for (size_t i = 0; i < MaxDynamicImage; i++) {
+                ClearColorImage(cmdbuf, *images.images[i]);
+            }
+        }
+    });
+    scheduler.Finish();
+
+    m_images_ready = true;
+}
+
+VkImageView SMAA::Draw(Scheduler& scheduler, size_t image_index, VkImage source_image,
+                       VkImageView source_image_view) {
+    Images& images = m_dynamic_images[image_index];
+
+    VkImage output_image = *images.images[Output];
+    VkImage edges_image = *images.images[Edges];
+    VkImage blend_image = *images.images[Blend];
+
+    VkDescriptorSet edge_detection_descriptor_set = images.descriptor_sets[EdgeDetection];
+    VkDescriptorSet blending_weight_calculation_descriptor_set =
+        images.descriptor_sets[BlendingWeightCalculation];
+    VkDescriptorSet neighborhood_blending_descriptor_set =
+        images.descriptor_sets[NeighborhoodBlending];
+
+    VkFramebuffer edge_detection_framebuffer = *images.framebuffers[EdgeDetection];
+    VkFramebuffer blending_weight_calculation_framebuffer =
+        *images.framebuffers[BlendingWeightCalculation];
+    VkFramebuffer neighborhood_blending_framebuffer = *images.framebuffers[NeighborhoodBlending];
+
+    UploadImages(scheduler);
+    UpdateDescriptorSets(source_image_view, image_index);
+
+    scheduler.RequestOutsideRenderPassOperationContext();
+    scheduler.Record([=, this](vk::CommandBuffer& cmdbuf) {
+        TransitionImageLayout(cmdbuf, source_image, VK_IMAGE_LAYOUT_GENERAL);
+        TransitionImageLayout(cmdbuf, edges_image, VK_IMAGE_LAYOUT_GENERAL);
+        BeginRenderPass(cmdbuf, m_renderpasses[EdgeDetection], edge_detection_framebuffer,
+                        m_extent);
+        cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, *m_pipelines[EdgeDetection]);
+        cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS,
+                                  *m_pipeline_layouts[EdgeDetection], 0,
+                                  edge_detection_descriptor_set, {});
+        cmdbuf.Draw(3, 1, 0, 0);
+        cmdbuf.EndRenderPass();
+
+        TransitionImageLayout(cmdbuf, edges_image, VK_IMAGE_LAYOUT_GENERAL);
+        TransitionImageLayout(cmdbuf, blend_image, VK_IMAGE_LAYOUT_GENERAL);
+        BeginRenderPass(cmdbuf, m_renderpasses[BlendingWeightCalculation],
+                        blending_weight_calculation_framebuffer, m_extent);
+        cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS,
+                            *m_pipelines[BlendingWeightCalculation]);
+        cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS,
+                                  *m_pipeline_layouts[BlendingWeightCalculation], 0,
+                                  blending_weight_calculation_descriptor_set, {});
+        cmdbuf.Draw(3, 1, 0, 0);
+        cmdbuf.EndRenderPass();
+
+        TransitionImageLayout(cmdbuf, blend_image, VK_IMAGE_LAYOUT_GENERAL);
+        TransitionImageLayout(cmdbuf, output_image, VK_IMAGE_LAYOUT_GENERAL);
+        BeginRenderPass(cmdbuf, m_renderpasses[NeighborhoodBlending],
+                        neighborhood_blending_framebuffer, m_extent);
+        cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, *m_pipelines[NeighborhoodBlending]);
+        cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS,
+                                  *m_pipeline_layouts[NeighborhoodBlending], 0,
+                                  neighborhood_blending_descriptor_set, {});
+        cmdbuf.Draw(3, 1, 0, 0);
+        cmdbuf.EndRenderPass();
+        TransitionImageLayout(cmdbuf, output_image, VK_IMAGE_LAYOUT_GENERAL);
+    });
+
+    return *images.image_views[Output];
+}
+
+} // namespace Vulkan

src/video_core/renderer_vulkan/vk_smaa.h

@@ -0,0 +1,86 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include <array>
+#include "video_core/vulkan_common/vulkan_memory_allocator.h"
+#include "video_core/vulkan_common/vulkan_wrapper.h"
+
+namespace Vulkan {
+
+class Device;
+class Scheduler;
+class StagingBufferPool;
+
+class SMAA {
+public:
+    explicit SMAA(const Device& device, MemoryAllocator& allocator, size_t image_count,
+                  VkExtent2D extent);
+    VkImageView Draw(Scheduler& scheduler, size_t image_index, VkImage source_image,
+                     VkImageView source_image_view);
+
+private:
+    enum SMAAStage {
+        EdgeDetection = 0,
+        BlendingWeightCalculation = 1,
+        NeighborhoodBlending = 2,
+        MaxSMAAStage = 3,
+    };
+
+    enum StaticImageType {
+        Area = 0,
+        Search = 1,
+        MaxStaticImage = 2,
+    };
+
+    enum DynamicImageType {
+        Blend = 0,
+        Edges = 1,
+        Output = 2,
+        MaxDynamicImage = 3,
+    };
+
+    void CreateImages();
+    void CreateRenderPasses();
+    void CreateSampler();
+    void CreateShaders();
+    void CreateDescriptorPool();
+    void CreateDescriptorSetLayouts();
+    void CreateDescriptorSets();
+    void CreatePipelineLayouts();
+    void CreatePipelines();
+    void UpdateDescriptorSets(VkImageView image_view, size_t image_index);
+    void UploadImages(Scheduler& scheduler);
+
+    const Device& m_device;
+    MemoryAllocator& m_allocator;
+    const VkExtent2D m_extent;
+    const u32 m_image_count;
+
+    vk::DescriptorPool m_descriptor_pool{};
+    std::array<vk::DescriptorSetLayout, MaxSMAAStage> m_descriptor_set_layouts{};
+    std::array<vk::PipelineLayout, MaxSMAAStage> m_pipeline_layouts{};
+    std::array<vk::ShaderModule, MaxSMAAStage> m_vertex_shaders{};
+    std::array<vk::ShaderModule, MaxSMAAStage> m_fragment_shaders{};
+    std::array<vk::Pipeline, MaxSMAAStage> m_pipelines{};
+    std::array<vk::RenderPass, MaxSMAAStage> m_renderpasses{};
+
+    std::array<MemoryCommit, MaxStaticImage> m_static_buffer_commits;
+    std::array<vk::Image, MaxStaticImage> m_static_images{};
+    std::array<vk::ImageView, MaxStaticImage> m_static_image_views{};
+
+    struct Images {
+        vk::DescriptorSets descriptor_sets{};
+        std::array<MemoryCommit, MaxDynamicImage> buffer_commits;
+        std::array<vk::Image, MaxDynamicImage> images{};
+        std::array<vk::ImageView, MaxDynamicImage> image_views{};
+        std::array<vk::Framebuffer, MaxSMAAStage> framebuffers{};
+    };
+    std::vector<Images> m_dynamic_images{};
+    bool m_images_ready{};
+
+    vk::Sampler m_sampler{};
+};
+
+} // namespace Vulkan

src/video_core/renderer_vulkan/vk_swapchain.cpp

@@ -67,17 +67,19 @@ VkExtent2D ChooseSwapExtent(const VkSurfaceCapabilitiesKHR& capabilities, u32 wi
 
 } // Anonymous namespace
 
-Swapchain::Swapchain(VkSurfaceKHR surface_, const Device& device_, Scheduler& scheduler_, u32 width,
-                     u32 height, bool srgb)
+Swapchain::Swapchain(VkSurfaceKHR surface_, const Device& device_, Scheduler& scheduler_,
+                     u32 width_, u32 height_, bool srgb)
     : surface{surface_}, device{device_}, scheduler{scheduler_} {
-    Create(width, height, srgb);
+    Create(width_, height_, srgb);
 }
 
 Swapchain::~Swapchain() = default;
 
-void Swapchain::Create(u32 width, u32 height, bool srgb) {
+void Swapchain::Create(u32 width_, u32 height_, bool srgb) {
     is_outdated = false;
     is_suboptimal = false;
+    width = width_;
+    height = height_;
 
     const auto physical_device = device.GetPhysical();
     const auto capabilities{physical_device.GetSurfaceCapabilitiesKHR(surface)};
@@ -88,7 +90,7 @@ void Swapchain::Create(u32 width, u32 height, bool srgb) {
     device.GetLogical().WaitIdle();
     Destroy();
 
-    CreateSwapchain(capabilities, width, height, srgb);
+    CreateSwapchain(capabilities, srgb);
     CreateSemaphores();
     CreateImageViews();
 
@@ -148,8 +150,7 @@ void Swapchain::Present(VkSemaphore render_semaphore) {
     }
 }
 
-void Swapchain::CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities, u32 width, u32 height,
-                                bool srgb) {
+void Swapchain::CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities, bool srgb) {
     const auto physical_device{device.GetPhysical()};
     const auto formats{physical_device.GetSurfaceFormatsKHR(surface)};
     const auto present_modes{physical_device.GetSurfacePresentModesKHR(surface)};

src/video_core/renderer_vulkan/vk_swapchain.h

@@ -80,9 +80,16 @@ public:
         return *present_semaphores[frame_index];
     }
 
+    u32 GetWidth() const {
+        return width;
+    }
+
+    u32 GetHeight() const {
+        return height;
+    }
+
 private:
-    void CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities, u32 width, u32 height,
-                         bool srgb);
+    void CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities, bool srgb);
     void CreateSemaphores();
     void CreateImageViews();
 
@@ -105,6 +112,9 @@ private:
     std::vector<u64> resource_ticks;
     std::vector<vk::Semaphore> present_semaphores;
 
+    u32 width;
+    u32 height;
+
     u32 image_index{};
     u32 frame_index{};
 

src/video_core/smaa_search_tex.h

@@ -0,0 +1,88 @@
+// SPDX-FileCopyrightText: 2013 Jorge Jimenez (jorge@iryoku.com)
+// SPDX-FileCopyrightText: 2013 Jose I. Echevarria (joseignacioechevarria@gmail.com)
+// SPDX-FileCopyrightText: 2013 Belen Masia (bmasia@unizar.es)
+// SPDX-FileCopyrightText: 2013 Fernando Navarro (fernandn@microsoft.com)
+// SPDX-FileCopyrightText: 2013 Diego Gutierrez (diegog@unizar.es)
+// SPDX-License-Identifier: MIT
+
+#ifndef SEARCHTEX_H
+#define SEARCHTEX_H
+
+#define SEARCHTEX_WIDTH 64
+#define SEARCHTEX_HEIGHT 16
+#define SEARCHTEX_PITCH SEARCHTEX_WIDTH
+#define SEARCHTEX_SIZE (SEARCHTEX_HEIGHT * SEARCHTEX_PITCH)
+
+/**
+ * Stored in R8 format. Load it in the following format:
+ *  - DX9:  D3DFMT_L8
+ *  - DX10: DXGI_FORMAT_R8_UNORM
+ */
+static const unsigned char searchTexBytes[] = {
+    0xfe, 0xfe, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0xfe, 0xfe, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f,
+    0x7f, 0xfe, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0xfe, 0xfe, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0xfe, 0xfe, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f,
+    0x7f, 0xfe, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0xfe, 0xfe, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0xfe, 0xfe, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f,
+    0x7f, 0xfe, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0xfe, 0xfe, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0xfe, 0xfe, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f,
+    0x7f, 0xfe, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f,
+    0x7f, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f,
+    0x7f, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f,
+    0x7f, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x7f,
+    0x7f, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7f, 0x7f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+};
+
+#endif

src/video_core/textures/texture.cpp

@@ -64,10 +64,11 @@ float TSCEntry::MaxAnisotropy() const noexcept {
         return 1.0f;
     }
     const auto anisotropic_settings = Settings::values.max_anisotropy.GetValue();
-    u32 added_anisotropic{};
+    s32 added_anisotropic{};
     if (anisotropic_settings == 0) {
         added_anisotropic = Settings::values.resolution_info.up_scale >>
                             Settings::values.resolution_info.down_shift;
+        added_anisotropic = std::max(added_anisotropic - 1, 0);
     } else {
         added_anisotropic = Settings::values.max_anisotropy.GetValue() - 1U;
     }

src/video_core/vulkan_common/vulkan_device.cpp

@@ -421,7 +421,7 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR
     VkPhysicalDevice8BitStorageFeatures bit8_storage{
         .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_8BIT_STORAGE_FEATURES,
         .pNext = nullptr,
-        .storageBuffer8BitAccess = false,
+        .storageBuffer8BitAccess = true,
         .uniformAndStorageBuffer8BitAccess = true,
         .storagePushConstant8 = false,
     };
@@ -660,6 +660,16 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR
         LOG_INFO(Render_Vulkan, "Device doesn't support depth range unrestricted");
     }
 
+    VkPhysicalDeviceDepthClipControlFeaturesEXT depth_clip_control_features;
+    if (ext_depth_clip_control) {
+        depth_clip_control_features = {
+            .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DEPTH_CLIP_CONTROL_FEATURES_EXT,
+            .pNext = nullptr,
+            .depthClipControl = VK_TRUE,
+        };
+        SetNext(next, depth_clip_control_features);
+    }
+
     VkDeviceDiagnosticsConfigCreateInfoNV diagnostics_nv;
     if (Settings::values.enable_nsight_aftermath && nv_device_diagnostics_config) {
         nsight_aftermath_tracker = std::make_unique<NsightAftermathTracker>();
@@ -1044,6 +1054,7 @@ void Device::CheckSuitability(bool requires_swapchain) const {
         std::make_pair(bit16_storage.storageBuffer16BitAccess, "storageBuffer16BitAccess"),
         std::make_pair(bit16_storage.uniformAndStorageBuffer16BitAccess,
                        "uniformAndStorageBuffer16BitAccess"),
+        std::make_pair(bit8_storage.storageBuffer8BitAccess, "storageBuffer8BitAccess"),
         std::make_pair(bit8_storage.uniformAndStorageBuffer8BitAccess,
                        "uniformAndStorageBuffer8BitAccess"),
         std::make_pair(host_query_reset.hostQueryReset, "hostQueryReset"),
@@ -1083,6 +1094,7 @@ std::vector<const char*> Device::LoadExtensions(bool requires_surface) {
     bool has_ext_vertex_input_dynamic_state{};
     bool has_ext_line_rasterization{};
     bool has_ext_primitive_topology_list_restart{};
+    bool has_ext_depth_clip_control{};
     for (const std::string& extension : supported_extensions) {
         const auto test = [&](std::optional<std::reference_wrapper<bool>> status, const char* name,
                               bool push) {
@@ -1116,6 +1128,7 @@ std::vector<const char*> Device::LoadExtensions(bool requires_surface) {
         test(ext_shader_stencil_export, VK_EXT_SHADER_STENCIL_EXPORT_EXTENSION_NAME, true);
         test(ext_conservative_rasterization, VK_EXT_CONSERVATIVE_RASTERIZATION_EXTENSION_NAME,
              true);
+        test(has_ext_depth_clip_control, VK_EXT_DEPTH_CLIP_CONTROL_EXTENSION_NAME, false);
         test(has_ext_transform_feedback, VK_EXT_TRANSFORM_FEEDBACK_EXTENSION_NAME, false);
         test(has_ext_custom_border_color, VK_EXT_CUSTOM_BORDER_COLOR_EXTENSION_NAME, false);
         test(has_ext_extended_dynamic_state, VK_EXT_EXTENDED_DYNAMIC_STATE_EXTENSION_NAME, false);
@@ -1279,6 +1292,19 @@ std::vector<const char*> Device::LoadExtensions(bool requires_surface) {
             ext_line_rasterization = true;
         }
     }
+    if (has_ext_depth_clip_control) {
+        VkPhysicalDeviceDepthClipControlFeaturesEXT depth_clip_control_features;
+        depth_clip_control_features.sType =
+            VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DEPTH_CLIP_CONTROL_FEATURES_EXT;
+        depth_clip_control_features.pNext = nullptr;
+        features.pNext = &depth_clip_control_features;
+        physical.GetFeatures2(features);
+
+        if (depth_clip_control_features.depthClipControl) {
+            extensions.push_back(VK_EXT_DEPTH_CLIP_CONTROL_EXTENSION_NAME);
+            ext_depth_clip_control = true;
+        }
+    }
     if (has_khr_workgroup_memory_explicit_layout) {
         VkPhysicalDeviceWorkgroupMemoryExplicitLayoutFeaturesKHR layout;
         layout.sType =
@@ -1380,6 +1406,10 @@ void Device::SetupFeatures() {
     is_shader_storage_image_multisample = features.shaderStorageImageMultisample;
     is_blit_depth_stencil_supported = TestDepthStencilBlits();
     is_optimal_astc_supported = IsOptimalAstcSupported(features);
+
+    const VkPhysicalDeviceLimits& limits{properties.limits};
+    max_vertex_input_attributes = limits.maxVertexInputAttributes;
+    max_vertex_input_bindings = limits.maxVertexInputBindings;
 }
 
 void Device::SetupProperties() {

src/video_core/vulkan_common/vulkan_device.h

@@ -256,6 +256,11 @@ public:
         return ext_depth_range_unrestricted;
     }
 
+    /// Returns true if the device supports VK_EXT_depth_clip_control.
+    bool IsExtDepthClipControlSupported() const {
+        return ext_depth_clip_control;
+    }
+
     /// Returns true if the device supports VK_EXT_shader_viewport_index_layer.
     bool IsExtShaderViewportIndexLayerSupported() const {
         return ext_shader_viewport_index_layer;
@@ -368,6 +373,14 @@ public:
         return must_emulate_bgr565;
     }
 
+    u32 GetMaxVertexInputAttributes() const {
+        return max_vertex_input_attributes;
+    }
+
+    u32 GetMaxVertexInputBindings() const {
+        return max_vertex_input_bindings;
+    }
+
 private:
     /// Checks if the physical device is suitable.
     void CheckSuitability(bool requires_swapchain) const;
@@ -446,6 +459,7 @@ private:
     bool khr_swapchain_mutable_format{};         ///< Support for VK_KHR_swapchain_mutable_format.
     bool ext_index_type_uint8{};                 ///< Support for VK_EXT_index_type_uint8.
     bool ext_sampler_filter_minmax{};            ///< Support for VK_EXT_sampler_filter_minmax.
+    bool ext_depth_clip_control{};               ///< Support for VK_EXT_depth_clip_control
     bool ext_depth_range_unrestricted{};         ///< Support for VK_EXT_depth_range_unrestricted.
     bool ext_shader_viewport_index_layer{}; ///< Support for VK_EXT_shader_viewport_index_layer.
     bool ext_tooling_info{};                ///< Support for VK_EXT_tooling_info.
@@ -467,6 +481,8 @@ private:
     bool supports_d24_depth{};              ///< Supports D24 depth buffers.
     bool cant_blit_msaa{};                  ///< Does not support MSAA<->MSAA blitting.
     bool must_emulate_bgr565{};             ///< Emulates BGR565 by swizzling RGB565 format.
+    u32 max_vertex_input_attributes{};      ///< Max vertex input attributes in pipeline
+    u32 max_vertex_input_bindings{};        ///< Max vertex input buffers in pipeline
 
     // Telemetry parameters
     std::string vendor_name;                       ///< Device's driver name.

src/video_core/vulkan_common/vulkan_wrapper.cpp

@@ -86,6 +86,7 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept {
     X(vkCmdBindVertexBuffers);
     X(vkCmdBlitImage);
     X(vkCmdClearAttachments);
+    X(vkCmdClearColorImage);
     X(vkCmdCopyBuffer);
     X(vkCmdCopyBufferToImage);
     X(vkCmdCopyImage);

src/video_core/vulkan_common/vulkan_wrapper.h

@@ -205,6 +205,7 @@ struct DeviceDispatch : InstanceDispatch {
     PFN_vkCmdBindVertexBuffers2EXT vkCmdBindVertexBuffers2EXT{};
     PFN_vkCmdBlitImage vkCmdBlitImage{};
     PFN_vkCmdClearAttachments vkCmdClearAttachments{};
+    PFN_vkCmdClearColorImage vkCmdClearColorImage{};
     PFN_vkCmdCopyBuffer vkCmdCopyBuffer{};
     PFN_vkCmdCopyBufferToImage vkCmdCopyBufferToImage{};
     PFN_vkCmdCopyImage vkCmdCopyImage{};
@@ -1024,6 +1025,11 @@ public:
                                    rects.data());
     }
 
+    void ClearColorImage(VkImage image, VkImageLayout layout, VkClearColorValue color,
+                         Span<VkImageSubresourceRange> ranges) {
+        dld->vkCmdClearColorImage(handle, image, layout, &color, ranges.size(), ranges.data());
+    }
+
     void BlitImage(VkImage src_image, VkImageLayout src_layout, VkImage dst_image,
                    VkImageLayout dst_layout, Span<VkImageBlit> regions,
                    VkFilter filter) const noexcept {

src/yuzu/CMakeLists.txt

@@ -402,7 +402,7 @@ if (MSVC)
     copy_yuzu_FFmpeg_deps(yuzu)
 endif()
 
-if (NOT APPLE)
+if (NOT APPLE AND ENABLE_OPENGL)
     target_compile_definitions(yuzu PRIVATE HAS_OPENGL)
 endif()
 

src/yuzu/bootmanager.cpp

@@ -44,6 +44,8 @@
 #include "yuzu/bootmanager.h"
 #include "yuzu/main.h"
 
+static Core::Frontend::WindowSystemType GetWindowSystemType();
+
 EmuThread::EmuThread(Core::System& system_) : system{system_} {}
 
 EmuThread::~EmuThread() = default;
@@ -61,8 +63,6 @@ void EmuThread::run() {
 
     // Main process has been loaded. Make the context current to this thread and begin GPU and CPU
     // execution.
-    gpu.Start();
-
     gpu.ObtainContext();
 
     emit LoadProgress(VideoCore::LoadCallbackStage::Prepare, 0, 0);
@@ -77,9 +77,15 @@ void EmuThread::run() {
     emit LoadProgress(VideoCore::LoadCallbackStage::Complete, 0, 0);
 
     gpu.ReleaseContext();
+    gpu.Start();
 
     system.GetCpuManager().OnGpuReady();
 
+    system.RegisterExitCallback([this]() {
+        stop_source.request_stop();
+        SetRunning(false);
+    });
+
     // Holds whether the cpu was running during the last iteration,
     // so that the DebugModeLeft signal can be emitted before the
     // next execution step
@@ -225,6 +231,9 @@ public:
     explicit RenderWidget(GRenderWindow* parent) : QWidget(parent), render_window(parent) {
         setAttribute(Qt::WA_NativeWindow);
         setAttribute(Qt::WA_PaintOnScreen);
+        if (GetWindowSystemType() == Core::Frontend::WindowSystemType::Wayland) {
+            setAttribute(Qt::WA_DontCreateNativeAncestors);
+        }
     }
 
     virtual ~RenderWidget() = default;
@@ -269,12 +278,14 @@ static Core::Frontend::WindowSystemType GetWindowSystemType() {
         return Core::Frontend::WindowSystemType::X11;
     else if (platform_name == QStringLiteral("wayland"))
         return Core::Frontend::WindowSystemType::Wayland;
+    else if (platform_name == QStringLiteral("wayland-egl"))
+        return Core::Frontend::WindowSystemType::Wayland;
     else if (platform_name == QStringLiteral("cocoa"))
         return Core::Frontend::WindowSystemType::Cocoa;
     else if (platform_name == QStringLiteral("android"))
         return Core::Frontend::WindowSystemType::Android;
 
-    LOG_CRITICAL(Frontend, "Unknown Qt platform!");
+    LOG_CRITICAL(Frontend, "Unknown Qt platform {}!", platform_name.toStdString());
     return Core::Frontend::WindowSystemType::Windows;
 }
 
@@ -314,6 +325,9 @@ GRenderWindow::GRenderWindow(GMainWindow* parent, EmuThread* emu_thread_,
     input_subsystem->Initialize();
     this->setMouseTracking(true);
 
+    strict_context_required = QGuiApplication::platformName() == QStringLiteral("wayland") ||
+                              QGuiApplication::platformName() == QStringLiteral("wayland-egl");
+
     connect(this, &GRenderWindow::FirstFrameDisplayed, parent, &GMainWindow::OnLoadComplete);
     connect(this, &GRenderWindow::ExecuteProgramSignal, parent, &GMainWindow::OnExecuteProgram,
             Qt::QueuedConnection);
@@ -750,6 +764,7 @@ void GRenderWindow::InitializeCamera() {
         return;
     }
 
+    camera_data.resize(CAMERA_WIDTH * CAMERA_HEIGHT);
     camera_capture->setCaptureDestination(QCameraImageCapture::CaptureDestination::CaptureToBuffer);
     connect(camera_capture.get(), &QCameraImageCapture::imageCaptured, this,
             &GRenderWindow::OnCameraCapture);
@@ -805,16 +820,13 @@ void GRenderWindow::RequestCameraCapture() {
 }
 
 void GRenderWindow::OnCameraCapture(int requestId, const QImage& img) {
-    constexpr std::size_t camera_width = 320;
-    constexpr std::size_t camera_height = 240;
+    // TODO: Capture directly in the format and resolution needed
     const auto converted =
-        img.scaled(camera_width, camera_height, Qt::AspectRatioMode::IgnoreAspectRatio,
+        img.scaled(CAMERA_WIDTH, CAMERA_HEIGHT, Qt::AspectRatioMode::IgnoreAspectRatio,
                    Qt::TransformationMode::SmoothTransformation)
             .mirrored(false, true);
-    std::vector<u32> camera_data{};
-    camera_data.resize(camera_width * camera_height);
-    std::memcpy(camera_data.data(), converted.bits(), camera_width * camera_height * sizeof(u32));
-    input_subsystem->GetCamera()->SetCameraData(camera_width, camera_height, camera_data);
+    std::memcpy(camera_data.data(), converted.bits(), CAMERA_WIDTH * CAMERA_HEIGHT * sizeof(u32));
+    input_subsystem->GetCamera()->SetCameraData(CAMERA_WIDTH, CAMERA_HEIGHT, camera_data);
     pending_camera_snapshots = 0;
 }
 
@@ -952,6 +964,12 @@ void GRenderWindow::OnMinimalClientAreaChangeRequest(std::pair<u32, u32> minimal
 
 bool GRenderWindow::InitializeOpenGL() {
 #ifdef HAS_OPENGL
+    if (!QOpenGLContext::supportsThreadedOpenGL()) {
+        QMessageBox::warning(this, tr("OpenGL not available!"),
+                             tr("OpenGL shared contexts are not supported."));
+        return false;
+    }
+
     // TODO: One of these flags might be interesting: WA_OpaquePaintEvent, WA_NoBackground,
     // WA_DontShowOnScreen, WA_DeleteOnClose
     auto child = new OpenGLRenderWidget(this);

src/yuzu/bootmanager.h

@@ -84,9 +84,10 @@ public:
     }
 
     /**
-     * Requests for the emulation thread to stop running
+     * Requests for the emulation thread to immediately stop running
      */
-    void RequestStop() {
+    void ForceStop() {
+        LOG_WARNING(Frontend, "Force stopping EmuThread");
         stop_source.request_stop();
         SetRunning(false);
     }
@@ -246,6 +247,9 @@ private:
 #if (QT_VERSION < QT_VERSION_CHECK(6, 0, 0)) && YUZU_USE_QT_MULTIMEDIA
     std::unique_ptr<QCamera> camera;
     std::unique_ptr<QCameraImageCapture> camera_capture;
+    static constexpr std::size_t CAMERA_WIDTH = 320;
+    static constexpr std::size_t CAMERA_HEIGHT = 240;
+    std::vector<u32> camera_data;
 #endif
     std::unique_ptr<QTimer> camera_timer;
 

src/yuzu/configuration/config.cpp

@@ -697,7 +697,6 @@ void Config::ReadRendererValues() {
     ReadGlobalSetting(Settings::values.fsr_sharpening_slider);
     ReadGlobalSetting(Settings::values.anti_aliasing);
     ReadGlobalSetting(Settings::values.max_anisotropy);
-    ReadGlobalSetting(Settings::values.use_speed_limit);
     ReadGlobalSetting(Settings::values.speed_limit);
     ReadGlobalSetting(Settings::values.use_disk_shader_cache);
     ReadGlobalSetting(Settings::values.gpu_accuracy);
@@ -784,6 +783,8 @@ void Config::ReadSystemValues() {
         }
     }
 
+    ReadBasicSetting(Settings::values.device_name);
+
     if (global) {
         ReadBasicSetting(Settings::values.current_user);
         Settings::values.current_user = std::clamp<int>(Settings::values.current_user.GetValue(), 0,
@@ -1328,7 +1329,6 @@ void Config::SaveRendererValues() {
                  static_cast<u32>(Settings::values.anti_aliasing.GetDefault()),
                  Settings::values.anti_aliasing.UsingGlobal());
     WriteGlobalSetting(Settings::values.max_anisotropy);
-    WriteGlobalSetting(Settings::values.use_speed_limit);
     WriteGlobalSetting(Settings::values.speed_limit);
     WriteGlobalSetting(Settings::values.use_disk_shader_cache);
     WriteSetting(QString::fromStdString(Settings::values.gpu_accuracy.GetLabel()),
@@ -1407,6 +1407,7 @@ void Config::SaveSystemValues() {
                  Settings::values.rng_seed.UsingGlobal());
     WriteSetting(QStringLiteral("rng_seed"), Settings::values.rng_seed.GetValue(global).value_or(0),
                  0, Settings::values.rng_seed.UsingGlobal());
+    WriteBasicSetting(Settings::values.device_name);
 
     if (global) {
         WriteBasicSetting(Settings::values.current_user);

src/yuzu/configuration/configure_graphics.ui

@@ -487,6 +487,11 @@
                <string>FXAA</string>
               </property>
              </item>
+             <item>
+              <property name="text">
+               <string>SMAA</string>
+              </property>
+             </item>
             </widget>
            </item>
           </layout>
@@ -524,6 +529,12 @@
            </property>
            <item>
             <layout class="QHBoxLayout" name="fsr_sharpening_label_group">
+             <property name="rightMargin">
+              <number>0</number>
+             </property>
+             <property name="bottomMargin">
+              <number>0</number>
+             </property>
              <item>
               <widget class="QComboBox" name="fsr_sharpening_combobox">
                <property name="sizePolicy">

src/yuzu/configuration/configure_system.cpp

@@ -72,6 +72,8 @@ void ConfigureSystem::SetConfiguration() {
     ui->custom_rtc_checkbox->setChecked(Settings::values.custom_rtc.has_value());
     ui->custom_rtc_edit->setEnabled(Settings::values.custom_rtc.has_value());
     ui->custom_rtc_edit->setDateTime(QDateTime::fromSecsSinceEpoch(rtc_time));
+    ui->device_name_edit->setText(
+        QString::fromUtf8(Settings::values.device_name.GetValue().c_str()));
 
     if (Settings::IsConfiguringGlobal()) {
         ui->combo_language->setCurrentIndex(Settings::values.language_index.GetValue());
@@ -115,6 +117,8 @@ void ConfigureSystem::ApplyConfiguration() {
         }
     }
 
+    Settings::values.device_name = ui->device_name_edit->text().toStdString();
+
     if (!enabled) {
         return;
     }

src/yuzu/configuration/configure_system.ui

@@ -432,6 +432,13 @@
             </property>
            </widget>
           </item>
+          <item row="7" column="0">
+           <widget class="QLabel" name="device_name_label">
+            <property name="text">
+             <string>Device Name</string>
+            </property>
+           </widget>
+          </item>
           <item row="3" column="1">
            <widget class="QComboBox" name="combo_sound">
             <item>
@@ -476,6 +483,13 @@
             </property>
            </widget>
           </item>
+          <item row="7" column="1">
+           <widget class="QLineEdit" name="device_name_edit">
+            <property name="maxLength">
+             <number>128</number>
+            </property>
+           </widget>
+          </item>
           <item row="6" column="1">
            <widget class="QLineEdit" name="rng_seed_edit">
             <property name="sizePolicy">