kernel/vm_manager: Reset region attributes when unmapping a VMA

Like the other members related to memory regions, the attributes need to
be reset back to their defaults as well.

src/audio_core/algorithm/interpolate.cpp

@@ -54,8 +54,9 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input,
             double l = 0.0;
             double r = 0.0;
             for (std::size_t j = 0; j < h.size(); j++) {
-                l += Lanczos(taps, pos + j - taps + 1) * h[j][0];
-                r += Lanczos(taps, pos + j - taps + 1) * h[j][1];
+                const double lanczos_calc = Lanczos(taps, pos + j - taps + 1);
+                l += lanczos_calc * h[j][0];
+                r += lanczos_calc * h[j][1];
             }
             output.emplace_back(static_cast<s16>(std::clamp(l, -32768.0, 32767.0)));
             output.emplace_back(static_cast<s16>(std::clamp(r, -32768.0, 32767.0)));

src/audio_core/audio_out.cpp

@@ -30,8 +30,7 @@ static Stream::Format ChannelsToStreamFormat(u32 num_channels) {
 StreamPtr AudioOut::OpenStream(u32 sample_rate, u32 num_channels, std::string&& name,
                                Stream::ReleaseCallback&& release_callback) {
     if (!sink) {
-        const SinkDetails& sink_details = GetSinkDetails(Settings::values.sink_id);
-        sink = sink_details.factory(Settings::values.audio_device_id);
+        sink = CreateSinkFromID(Settings::values.sink_id, Settings::values.audio_device_id);
     }
 
     return std::make_shared<Stream>(

src/audio_core/audio_renderer.cpp

@@ -285,8 +285,11 @@ void AudioRenderer::VoiceState::RefreshBuffer() {
         break;
     }
 
-    samples =
-        Interpolate(interp_state, std::move(samples), GetInfo().sample_rate, STREAM_SAMPLE_RATE);
+    // Only interpolate when necessary, expensive.
+    if (GetInfo().sample_rate != STREAM_SAMPLE_RATE) {
+        samples = Interpolate(interp_state, std::move(samples), GetInfo().sample_rate,
+                              STREAM_SAMPLE_RATE);
+    }
 
     is_refresh_pending = false;
 }

src/audio_core/cubeb_sink.cpp

@@ -107,7 +107,7 @@ private:
     static void StateCallback(cubeb_stream* stream, void* user_data, cubeb_state state);
 };
 
-CubebSink::CubebSink(std::string target_device_name) {
+CubebSink::CubebSink(std::string_view target_device_name) {
     if (cubeb_init(&ctx, "yuzu", nullptr) != CUBEB_OK) {
         LOG_CRITICAL(Audio_Sink, "cubeb_init failed");
         return;

src/audio_core/cubeb_sink.h

@@ -15,7 +15,7 @@ namespace AudioCore {
 
 class CubebSink final : public Sink {
 public:
-    explicit CubebSink(std::string device_id);
+    explicit CubebSink(std::string_view device_id);
     ~CubebSink() override;
 
     SinkStream& AcquireSinkStream(u32 sample_rate, u32 num_channels,

src/audio_core/null_sink.h

@@ -10,7 +10,7 @@ namespace AudioCore {
 
 class NullSink final : public Sink {
 public:
-    explicit NullSink(std::string){};
+    explicit NullSink(std::string_view) {}
     ~NullSink() override = default;
 
     SinkStream& AcquireSinkStream(u32 /*sample_rate*/, u32 /*num_channels*/,

src/audio_core/sink_details.cpp

@@ -14,31 +14,68 @@
 #include "common/logging/log.h"
 
 namespace AudioCore {
+namespace {
+struct SinkDetails {
+    using FactoryFn = std::unique_ptr<Sink> (*)(std::string_view);
+    using ListDevicesFn = std::vector<std::string> (*)();
 
-// g_sink_details is ordered in terms of desirability, with the best choice at the top.
-const std::vector<SinkDetails> g_sink_details = {
+    /// Name for this sink.
+    const char* id;
+    /// A method to call to construct an instance of this type of sink.
+    FactoryFn factory;
+    /// A method to call to list available devices.
+    ListDevicesFn list_devices;
+};
+
+// sink_details is ordered in terms of desirability, with the best choice at the top.
+constexpr SinkDetails sink_details[] = {
 #ifdef HAVE_CUBEB
-    SinkDetails{"cubeb", &std::make_unique<CubebSink, std::string>, &ListCubebSinkDevices},
+    SinkDetails{"cubeb",
+                [](std::string_view device_id) -> std::unique_ptr<Sink> {
+                    return std::make_unique<CubebSink>(device_id);
+                },
+                &ListCubebSinkDevices},
 #endif
-    SinkDetails{"null", &std::make_unique<NullSink, std::string>,
+    SinkDetails{"null",
+                [](std::string_view device_id) -> std::unique_ptr<Sink> {
+                    return std::make_unique<NullSink>(device_id);
+                },
                 [] { return std::vector<std::string>{"null"}; }},
 };
 
 const SinkDetails& GetSinkDetails(std::string_view sink_id) {
     auto iter =
-        std::find_if(g_sink_details.begin(), g_sink_details.end(),
+        std::find_if(std::begin(sink_details), std::end(sink_details),
                      [sink_id](const auto& sink_detail) { return sink_detail.id == sink_id; });
 
-    if (sink_id == "auto" || iter == g_sink_details.end()) {
+    if (sink_id == "auto" || iter == std::end(sink_details)) {
         if (sink_id != "auto") {
             LOG_ERROR(Audio, "AudioCore::SelectSink given invalid sink_id {}", sink_id);
         }
         // Auto-select.
-        // g_sink_details is ordered in terms of desirability, with the best choice at the front.
-        iter = g_sink_details.begin();
+        // sink_details is ordered in terms of desirability, with the best choice at the front.
+        iter = std::begin(sink_details);
     }
 
     return *iter;
 }
+} // Anonymous namespace
+
+std::vector<const char*> GetSinkIDs() {
+    std::vector<const char*> sink_ids(std::size(sink_details));
+
+    std::transform(std::begin(sink_details), std::end(sink_details), std::begin(sink_ids),
+                   [](const auto& sink) { return sink.id; });
+
+    return sink_ids;
+}
+
+std::vector<std::string> GetDeviceListForSink(std::string_view sink_id) {
+    return GetSinkDetails(sink_id).list_devices();
+}
+
+std::unique_ptr<Sink> CreateSinkFromID(std::string_view sink_id, std::string_view device_id) {
+    return GetSinkDetails(sink_id).factory(device_id);
+}
 
 } // namespace AudioCore

src/audio_core/sink_details.h

@@ -4,34 +4,21 @@
 
 #pragma once
 
-#include <functional>
-#include <memory>
 #include <string>
 #include <string_view>
-#include <utility>
 #include <vector>
 
 namespace AudioCore {
 
 class Sink;
 
-struct SinkDetails {
-    using FactoryFn = std::function<std::unique_ptr<Sink>(std::string)>;
-    using ListDevicesFn = std::function<std::vector<std::string>()>;
+/// Retrieves the IDs for all available audio sinks.
+std::vector<const char*> GetSinkIDs();
 
-    SinkDetails(const char* id_, FactoryFn factory_, ListDevicesFn list_devices_)
-        : id(id_), factory(std::move(factory_)), list_devices(std::move(list_devices_)) {}
+/// Gets the list of devices for a particular sink identified by the given ID.
+std::vector<std::string> GetDeviceListForSink(std::string_view sink_id);
 
-    /// Name for this sink.
-    const char* id;
-    /// A method to call to construct an instance of this type of sink.
-    FactoryFn factory;
-    /// A method to call to list available devices.
-    ListDevicesFn list_devices;
-};
-
-extern const std::vector<SinkDetails> g_sink_details;
-
-const SinkDetails& GetSinkDetails(std::string_view sink_id);
+/// Creates an audio sink identified by the given device ID.
+std::unique_ptr<Sink> CreateSinkFromID(std::string_view sink_id, std::string_view device_id);
 
 } // namespace AudioCore

src/common/logging/backend.cpp

@@ -13,7 +13,7 @@
 #include <vector>
 #ifdef _WIN32
 #include <share.h>   // For _SH_DENYWR
-#include <windows.h> // For OutputDebugStringA
+#include <windows.h> // For OutputDebugStringW
 #else
 #define _SH_DENYWR 0
 #endif
@@ -148,7 +148,7 @@ void FileBackend::Write(const Entry& entry) {
 
 void DebuggerBackend::Write(const Entry& entry) {
 #ifdef _WIN32
-    ::OutputDebugStringA(FormatLogMessage(entry).append(1, '\n').c_str());
+    ::OutputDebugStringW(Common::UTF8ToUTF16W(FormatLogMessage(entry).append(1, '\n')).c_str());
 #endif
 }
 

src/common/quaternion.h

@@ -12,7 +12,7 @@ template <typename T>
 class Quaternion {
 public:
     Math::Vec3<T> xyz;
-    T w;
+    T w{};
 
     Quaternion<decltype(-T{})> Inverse() const {
         return {-xyz, w};

src/common/thread_queue_list.h

@@ -49,6 +49,22 @@ struct ThreadQueueList {
         return T();
     }
 
+    template <typename UnaryPredicate>
+    T get_first_filter(UnaryPredicate filter) const {
+        const Queue* cur = first;
+        while (cur != nullptr) {
+            if (!cur->data.empty()) {
+                for (const auto& item : cur->data) {
+                    if (filter(item))
+                        return item;
+                }
+            }
+            cur = cur->next_nonempty;
+        }
+
+        return T();
+    }
+
     T pop_first() {
         Queue* cur = first;
         while (cur != nullptr) {

src/core/CMakeLists.txt

@@ -83,6 +83,8 @@ add_library(core STATIC
     file_sys/vfs_vector.h
     file_sys/xts_archive.cpp
     file_sys/xts_archive.h
+    frontend/applets/profile_select.cpp
+    frontend/applets/profile_select.h
     frontend/applets/software_keyboard.cpp
     frontend/applets/software_keyboard.h
     frontend/emu_window.cpp
@@ -162,6 +164,8 @@ add_library(core STATIC
     hle/service/am/applet_oe.h
     hle/service/am/applets/applets.cpp
     hle/service/am/applets/applets.h
+    hle/service/am/applets/profile_select.cpp
+    hle/service/am/applets/profile_select.h
     hle/service/am/applets/software_keyboard.cpp
     hle/service/am/applets/software_keyboard.h
     hle/service/am/applets/stub_applet.cpp

src/core/arm/dynarmic/arm_dynarmic.cpp

@@ -151,6 +151,7 @@ std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const {
     config.tpidr_el0 = &cb->tpidr_el0;
     config.dczid_el0 = 4;
     config.ctr_el0 = 0x8444c004;
+    config.cntfrq_el0 = 19200000; // Value from fusee.
 
     // Unpredictable instructions
     config.define_unpredictable_behaviour = true;

src/core/core.cpp

@@ -8,6 +8,7 @@
 #include <thread>
 #include <utility>
 
+#include "common/file_util.h"
 #include "common/logging/log.h"
 #include "common/string_util.h"
 #include "core/arm/exclusive_monitor.h"
@@ -40,7 +41,6 @@ namespace Core {
 
 /*static*/ System System::s_instance;
 
-namespace {
 FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
                                          const std::string& path) {
     // To account for split 00+01+etc files.
@@ -69,11 +69,13 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
         return FileSys::ConcatenatedVfsFile::MakeConcatenatedFile(concat, dir->GetName());
     }
 
+    if (FileUtil::IsDirectory(path))
+        return vfs->OpenFile(path + "/" + "main", FileSys::Mode::Read);
+
     return vfs->OpenFile(path, FileSys::Mode::Read);
 }
-} // Anonymous namespace
-
 struct System::Impl {
+
     Cpu& CurrentCpuCore() {
         return cpu_core_manager.GetCurrentCore();
     }
@@ -97,6 +99,8 @@ struct System::Impl {
             virtual_filesystem = std::make_shared<FileSys::RealVfsFilesystem>();
 
         /// Create default implementations of applets if one is not provided.
+        if (profile_selector == nullptr)
+            profile_selector = std::make_unique<Core::Frontend::DefaultProfileSelectApplet>();
         if (software_keyboard == nullptr)
             software_keyboard = std::make_unique<Core::Frontend::DefaultSoftwareKeyboardApplet>();
 
@@ -227,6 +231,7 @@ struct System::Impl {
     bool is_powered_on = false;
 
     /// Frontend applets
+    std::unique_ptr<Core::Frontend::ProfileSelectApplet> profile_selector;
     std::unique_ptr<Core::Frontend::SoftwareKeyboardApplet> software_keyboard;
 
     /// Service manager
@@ -422,6 +427,14 @@ std::shared_ptr<FileSys::VfsFilesystem> System::GetFilesystem() const {
     return impl->virtual_filesystem;
 }
 
+void System::SetProfileSelector(std::unique_ptr<Core::Frontend::ProfileSelectApplet> applet) {
+    impl->profile_selector = std::move(applet);
+}
+
+const Core::Frontend::ProfileSelectApplet& System::GetProfileSelector() const {
+    return *impl->profile_selector;
+}
+
 void System::SetSoftwareKeyboard(std::unique_ptr<Core::Frontend::SoftwareKeyboardApplet> applet) {
     impl->software_keyboard = std::move(applet);
 }

src/core/core.h

@@ -9,7 +9,9 @@
 #include <string>
 
 #include "common/common_types.h"
+#include "core/file_sys/vfs_types.h"
 #include "core/hle/kernel/object.h"
+#include "frontend/applets/profile_select.h"
 
 namespace Core::Frontend {
 class EmuWindow;
@@ -55,6 +57,9 @@ class TelemetrySession;
 
 struct PerfStatsResults;
 
+FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
+                                         const std::string& path);
+
 class System {
 public:
     System(const System&) = delete;
@@ -237,6 +242,10 @@ public:
 
     std::shared_ptr<FileSys::VfsFilesystem> GetFilesystem() const;
 
+    void SetProfileSelector(std::unique_ptr<Core::Frontend::ProfileSelectApplet> applet);
+
+    const Core::Frontend::ProfileSelectApplet& GetProfileSelector() const;
+
     void SetSoftwareKeyboard(std::unique_ptr<Core::Frontend::SoftwareKeyboardApplet> applet);
 
     const Core::Frontend::SoftwareKeyboardApplet& GetSoftwareKeyboard() const;

src/core/file_sys/directory.h

@@ -29,8 +29,8 @@ struct Entry {
         filename[copy_size] = '\0';
     }
 
-    char filename[0x300];
-    INSERT_PADDING_BYTES(4);
+    char filename[0x301];
+    INSERT_PADDING_BYTES(3);
     EntryType type;
     INSERT_PADDING_BYTES(3);
     u64 file_size;

src/core/file_sys/patch_manager.cpp

@@ -56,6 +56,10 @@ PatchManager::PatchManager(u64 title_id) : title_id(title_id) {}
 
 PatchManager::~PatchManager() = default;
 
+u64 PatchManager::GetTitleID() const {
+    return title_id;
+}
+
 VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
     LOG_INFO(Loader, "Patching ExeFS for title_id={:016X}", title_id);
 
@@ -73,11 +77,15 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
 
     const auto installed = Service::FileSystem::GetUnionContents();
 
+    const auto& disabled = Settings::values.disabled_addons[title_id];
+    const auto update_disabled =
+        std::find(disabled.begin(), disabled.end(), "Update") != disabled.end();
+
     // Game Updates
     const auto update_tid = GetUpdateTitleID(title_id);
     const auto update = installed.GetEntry(update_tid, ContentRecordType::Program);
 
-    if (update != nullptr && update->GetExeFS() != nullptr &&
+    if (!update_disabled && update != nullptr && update->GetExeFS() != nullptr &&
         update->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) {
         LOG_INFO(Loader, "    ExeFS: Update ({}) applied successfully",
                  FormatTitleVersion(installed.GetEntryVersion(update_tid).value_or(0)));
@@ -95,6 +103,9 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
         std::vector<VirtualDir> layers;
         layers.reserve(patch_dirs.size() + 1);
         for (const auto& subdir : patch_dirs) {
+            if (std::find(disabled.begin(), disabled.end(), subdir->GetName()) != disabled.end())
+                continue;
+
             auto exefs_dir = subdir->GetSubdirectory("exefs");
             if (exefs_dir != nullptr)
                 layers.push_back(std::move(exefs_dir));
@@ -111,11 +122,16 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
     return exefs;
 }
 
-static std::vector<VirtualFile> CollectPatches(const std::vector<VirtualDir>& patch_dirs,
-                                               const std::string& build_id) {
+std::vector<VirtualFile> PatchManager::CollectPatches(const std::vector<VirtualDir>& patch_dirs,
+                                                      const std::string& build_id) const {
+    const auto& disabled = Settings::values.disabled_addons[title_id];
+
     std::vector<VirtualFile> out;
     out.reserve(patch_dirs.size());
     for (const auto& subdir : patch_dirs) {
+        if (std::find(disabled.begin(), disabled.end(), subdir->GetName()) != disabled.end())
+            continue;
+
         auto exefs_dir = subdir->GetSubdirectory("exefs");
         if (exefs_dir != nullptr) {
             for (const auto& file : exefs_dir->GetFiles()) {
@@ -228,6 +244,7 @@ static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType t
         return;
     }
 
+    const auto& disabled = Settings::values.disabled_addons[title_id];
     auto patch_dirs = load_dir->GetSubdirectories();
     std::sort(patch_dirs.begin(), patch_dirs.end(),
               [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
@@ -237,6 +254,9 @@ static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType t
     layers.reserve(patch_dirs.size() + 1);
     layers_ext.reserve(patch_dirs.size() + 1);
     for (const auto& subdir : patch_dirs) {
+        if (std::find(disabled.begin(), disabled.end(), subdir->GetName()) != disabled.end())
+            continue;
+
         auto romfs_dir = subdir->GetSubdirectory("romfs");
         if (romfs_dir != nullptr)
             layers.push_back(std::move(romfs_dir));
@@ -266,13 +286,12 @@ static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType t
 VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, ContentRecordType type,
                                      VirtualFile update_raw) const {
     const auto log_string = fmt::format("Patching RomFS for title_id={:016X}, type={:02X}",
-                                        title_id, static_cast<u8>(type))
-                                .c_str();
+                                        title_id, static_cast<u8>(type));
 
     if (type == ContentRecordType::Program || type == ContentRecordType::Data)
-        LOG_INFO(Loader, log_string);
+        LOG_INFO(Loader, "{}", log_string);
     else
-        LOG_DEBUG(Loader, log_string);
+        LOG_DEBUG(Loader, "{}", log_string);
 
     if (romfs == nullptr)
         return romfs;
@@ -282,7 +301,12 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content
     // Game Updates
     const auto update_tid = GetUpdateTitleID(title_id);
     const auto update = installed.GetEntryRaw(update_tid, type);
-    if (update != nullptr) {
+
+    const auto& disabled = Settings::values.disabled_addons[title_id];
+    const auto update_disabled =
+        std::find(disabled.begin(), disabled.end(), "Update") != disabled.end();
+
+    if (!update_disabled && update != nullptr) {
         const auto new_nca = std::make_shared<NCA>(update, romfs, ivfc_offset);
         if (new_nca->GetStatus() == Loader::ResultStatus::Success &&
             new_nca->GetRomFS() != nullptr) {
@@ -290,7 +314,7 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content
                      FormatTitleVersion(installed.GetEntryVersion(update_tid).value_or(0)));
             romfs = new_nca->GetRomFS();
         }
-    } else if (update_raw != nullptr) {
+    } else if (!update_disabled && update_raw != nullptr) {
         const auto new_nca = std::make_shared<NCA>(update_raw, romfs, ivfc_offset);
         if (new_nca->GetStatus() == Loader::ResultStatus::Success &&
             new_nca->GetRomFS() != nullptr) {
@@ -320,25 +344,30 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
     VirtualFile update_raw) const {
     std::map<std::string, std::string, std::less<>> out;
     const auto installed = Service::FileSystem::GetUnionContents();
+    const auto& disabled = Settings::values.disabled_addons[title_id];
 
     // Game Updates
     const auto update_tid = GetUpdateTitleID(title_id);
     PatchManager update{update_tid};
     auto [nacp, discard_icon_file] = update.GetControlMetadata();
 
+    const auto update_disabled =
+        std::find(disabled.begin(), disabled.end(), "Update") != disabled.end();
+    const auto update_label = update_disabled ? "[D] Update" : "Update";
+
     if (nacp != nullptr) {
-        out.insert_or_assign("Update", nacp->GetVersionString());
+        out.insert_or_assign(update_label, nacp->GetVersionString());
     } else {
         if (installed.HasEntry(update_tid, ContentRecordType::Program)) {
             const auto meta_ver = installed.GetEntryVersion(update_tid);
             if (meta_ver.value_or(0) == 0) {
-                out.insert_or_assign("Update", "");
+                out.insert_or_assign(update_label, "");
             } else {
                 out.insert_or_assign(
-                    "Update", FormatTitleVersion(*meta_ver, TitleVersionFormat::ThreeElements));
+                    update_label, FormatTitleVersion(*meta_ver, TitleVersionFormat::ThreeElements));
             }
         } else if (update_raw != nullptr) {
-            out.insert_or_assign("Update", "PACKED");
+            out.insert_or_assign(update_label, "PACKED");
         }
     }
 
@@ -378,7 +407,9 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
             if (types.empty())
                 continue;
 
-            out.insert_or_assign(mod->GetName(), types);
+            const auto mod_disabled =
+                std::find(disabled.begin(), disabled.end(), mod->GetName()) != disabled.end();
+            out.insert_or_assign(mod_disabled ? "[D] " + mod->GetName() : mod->GetName(), types);
         }
     }
 
@@ -401,7 +432,9 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
 
         list += fmt::format("{}", dlc_match.back().title_id & 0x7FF);
 
-        out.insert_or_assign("DLC", std::move(list));
+        const auto dlc_disabled =
+            std::find(disabled.begin(), disabled.end(), "DLC") != disabled.end();
+        out.insert_or_assign(dlc_disabled ? "[D] DLC" : "DLC", std::move(list));
     }
 
     return out;

src/core/file_sys/patch_manager.h

@@ -30,6 +30,8 @@ public:
     explicit PatchManager(u64 title_id);
     ~PatchManager();
 
+    u64 GetTitleID() const;
+
     // Currently tracked ExeFS patches:
     // - Game Updates
     VirtualDir PatchExeFS(VirtualDir exefs) const;
@@ -63,6 +65,9 @@ public:
     std::pair<std::unique_ptr<NACP>, VirtualFile> ParseControlNCA(const NCA& nca) const;
 
 private:
+    std::vector<VirtualFile> CollectPatches(const std::vector<VirtualDir>& patch_dirs,
+                                            const std::string& build_id) const;
+
     u64 title_id;
 };
 

src/core/file_sys/savedata_factory.cpp

@@ -18,7 +18,11 @@ std::string SaveDataDescriptor::DebugInfo() const {
                        static_cast<u8>(type), title_id, user_id[1], user_id[0], save_id);
 }
 
-SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save_directory)) {}
+SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save_directory)) {
+    // Delete all temporary storages
+    // On hardware, it is expected that temporary storage be empty at first use.
+    dir->DeleteSubdirectoryRecursive("temp");
+}
 
 SaveDataFactory::~SaveDataFactory() = default;
 
@@ -120,8 +124,11 @@ std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType typ
     case SaveDataType::TemporaryStorage:
         return fmt::format("{}{:016X}/{:016X}{:016X}/{:016X}", out, 0, user_id[1], user_id[0],
                            title_id);
+    case SaveDataType::CacheStorage:
+        return fmt::format("{}save/cache/{:016X}", out, title_id);
     default:
         ASSERT_MSG(false, "Unrecognized SaveDataType: {:02X}", static_cast<u8>(type));
+        return fmt::format("{}save/unknown_{:X}/{:016X}", out, static_cast<u8>(type), title_id);
     }
 }
 

src/core/file_sys/system_archive/ng_word.cpp

@@ -39,4 +39,43 @@ VirtualDir NgWord1() {
     return std::make_shared<VectorVfsDirectory>(files, std::vector<VirtualDir>{}, "data");
 }
 
+namespace NgWord2Data {
+
+constexpr std::size_t NUMBER_AC_NX_FILES = 0x10;
+
+// Should this archive replacement mysteriously not work on a future game, consider updating.
+constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x15}; // 5.1.0 System Version
+
+constexpr std::array<u8, 0x2C> AC_NX_DATA{
+    0x1F, 0x8B, 0x08, 0x08, 0xD5, 0x2C, 0x09, 0x5C, 0x04, 0x00, 0x61, 0x63, 0x72, 0x61, 0x77,
+    0x00, 0xED, 0xC1, 0x01, 0x0D, 0x00, 0x00, 0x00, 0xC2, 0x20, 0xFB, 0xA7, 0xB6, 0xC7, 0x07,
+    0x0C, 0x00, 0x00, 0x00, 0xC8, 0x3B, 0x11, 0x00, 0x1C, 0xC7, 0x00, 0x10, 0x00, 0x00,
+}; // Deserializes to no bad words
+
+} // namespace NgWord2Data
+
+VirtualDir NgWord2() {
+    std::vector<VirtualFile> files(NgWord2Data::NUMBER_AC_NX_FILES * 3);
+
+    for (std::size_t i = 0; i < NgWord2Data::NUMBER_AC_NX_FILES; ++i) {
+        files[3 * i] = std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>(
+            NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_b1_nx", i));
+        files[3 * i + 1] = std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>(
+            NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_b2_nx", i));
+        files[3 * i + 2] = std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>(
+            NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_not_b_nx", i));
+    }
+
+    files.push_back(std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>(
+        NgWord2Data::AC_NX_DATA, "ac_common_b1_nx"));
+    files.push_back(std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>(
+        NgWord2Data::AC_NX_DATA, "ac_common_b2_nx"));
+    files.push_back(std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>(
+        NgWord2Data::AC_NX_DATA, "ac_common_not_b_nx"));
+    files.push_back(std::make_shared<ArrayVfsFile<NgWord2Data::VERSION_DAT.size()>>(
+        NgWord2Data::VERSION_DAT, "version.dat"));
+
+    return std::make_shared<VectorVfsDirectory>(files, std::vector<VirtualDir>{}, "data");
+}
+
 } // namespace FileSys::SystemArchive

src/core/file_sys/system_archive/ng_word.h

@@ -9,5 +9,6 @@
 namespace FileSys::SystemArchive {
 
 VirtualDir NgWord1();
+VirtualDir NgWord2();
 
 } // namespace FileSys::SystemArchive

src/core/file_sys/system_archive/system_archive.cpp

@@ -56,7 +56,7 @@ constexpr std::array<SystemArchiveDescriptor, SYSTEM_ARCHIVE_COUNT> SYSTEM_ARCHI
     {0x0100000000000820, "PlatformConfigCopper", nullptr},
     {0x0100000000000821, "PlatformConfigHoag", nullptr},
     {0x0100000000000822, "ControllerFirmware", nullptr},
-    {0x0100000000000823, "NgWord2", nullptr},
+    {0x0100000000000823, "NgWord2", &NgWord2},
     {0x0100000000000824, "PlatformConfigIcosaMariko", nullptr},
     {0x0100000000000825, "ApplicationBlackList", nullptr},
     {0x0100000000000826, "RebootlessSystemUpdateVersion", nullptr},

src/core/frontend/applets/profile_select.cpp

@@ -0,0 +1,19 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/frontend/applets/profile_select.h"
+#include "core/settings.h"
+
+namespace Core::Frontend {
+
+ProfileSelectApplet::~ProfileSelectApplet() = default;
+
+void DefaultProfileSelectApplet::SelectProfile(
+    std::function<void(std::optional<Service::Account::UUID>)> callback) const {
+    Service::Account::ProfileManager manager;
+    callback(manager.GetUser(Settings::values.current_user).value_or(Service::Account::UUID{}));
+    LOG_INFO(Service_ACC, "called, selecting current user instead of prompting...");
+}
+
+} // namespace Core::Frontend

src/core/frontend/applets/profile_select.h

@@ -0,0 +1,27 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <functional>
+#include <optional>
+#include "core/hle/service/acc/profile_manager.h"
+
+namespace Core::Frontend {
+
+class ProfileSelectApplet {
+public:
+    virtual ~ProfileSelectApplet();
+
+    virtual void SelectProfile(
+        std::function<void(std::optional<Service::Account::UUID>)> callback) const = 0;
+};
+
+class DefaultProfileSelectApplet final : public ProfileSelectApplet {
+public:
+    void SelectProfile(
+        std::function<void(std::optional<Service::Account::UUID>)> callback) const override;
+};
+
+} // namespace Core::Frontend

src/core/frontend/framebuffer_layout.cpp

@@ -6,6 +6,7 @@
 
 #include "common/assert.h"
 #include "core/frontend/framebuffer_layout.h"
+#include "core/settings.h"
 
 namespace Layout {
 
@@ -42,4 +43,18 @@ FramebufferLayout DefaultFrameLayout(unsigned width, unsigned height) {
     return res;
 }
 
+FramebufferLayout FrameLayoutFromResolutionScale(u16 res_scale) {
+    int width, height;
+
+    if (Settings::values.use_docked_mode) {
+        width = ScreenDocked::WidthDocked * res_scale;
+        height = ScreenDocked::HeightDocked * res_scale;
+    } else {
+        width = ScreenUndocked::Width * res_scale;
+        height = ScreenUndocked::Height * res_scale;
+    }
+
+    return DefaultFrameLayout(width, height);
+}
+
 } // namespace Layout

src/core/frontend/framebuffer_layout.h

@@ -9,6 +9,7 @@
 namespace Layout {
 
 enum ScreenUndocked : unsigned { Width = 1280, Height = 720 };
+enum ScreenDocked : unsigned { WidthDocked = 1920, HeightDocked = 1080 };
 
 /// Describes the layout of the window framebuffer
 struct FramebufferLayout {
@@ -34,4 +35,10 @@ struct FramebufferLayout {
  */
 FramebufferLayout DefaultFrameLayout(unsigned width, unsigned height);
 
+/**
+ * Convenience method to get frame layout by resolution scale
+ * @param res_scale resolution scale factor
+ */
+FramebufferLayout FrameLayoutFromResolutionScale(u16 res_scale);
+
 } // namespace Layout

src/core/gdbstub/gdbstub.cpp

@@ -201,11 +201,11 @@ void RegisterModule(std::string name, VAddr beg, VAddr end, bool add_elf_ext) {
     modules.push_back(std::move(module));
 }
 
-static Kernel::Thread* FindThreadById(int id) {
+static Kernel::Thread* FindThreadById(s64 id) {
     for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
         const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
         for (auto& thread : threads) {
-            if (thread->GetThreadID() == static_cast<u32>(id)) {
+            if (thread->GetThreadID() == static_cast<u64>(id)) {
                 current_core = core;
                 return thread.get();
             }

src/core/hle/kernel/errors.h

@@ -27,7 +27,7 @@ constexpr ResultCode ERR_SYNCHRONIZATION_CANCELED{ErrorModule::Kernel, 118};
 constexpr ResultCode ERR_OUT_OF_RANGE{ErrorModule::Kernel, 119};
 constexpr ResultCode ERR_INVALID_ENUM_VALUE{ErrorModule::Kernel, 120};
 constexpr ResultCode ERR_NOT_FOUND{ErrorModule::Kernel, 121};
-constexpr ResultCode ERR_ALREADY_REGISTERED{ErrorModule::Kernel, 122};
+constexpr ResultCode ERR_BUSY{ErrorModule::Kernel, 122};
 constexpr ResultCode ERR_SESSION_CLOSED_BY_REMOTE{ErrorModule::Kernel, 123};
 constexpr ResultCode ERR_INVALID_STATE{ErrorModule::Kernel, 125};
 constexpr ResultCode ERR_RESOURCE_LIMIT_EXCEEDED{ErrorModule::Kernel, 132};

src/core/hle/kernel/kernel.cpp

@@ -112,7 +112,7 @@ struct KernelCore::Impl {
 
     void Shutdown() {
         next_object_id = 0;
-        next_process_id = 10;
+        next_process_id = Process::ProcessIDMin;
         next_thread_id = 1;
 
         process_list.clear();
@@ -153,10 +153,8 @@ struct KernelCore::Impl {
     }
 
     std::atomic<u32> next_object_id{0};
-    // TODO(Subv): Start the process ids from 10 for now, as lower PIDs are
-    // reserved for low-level services
-    std::atomic<u32> next_process_id{10};
-    std::atomic<u32> next_thread_id{1};
+    std::atomic<u64> next_process_id{Process::ProcessIDMin};
+    std::atomic<u64> next_thread_id{1};
 
     // Lists all processes that exist in the current session.
     std::vector<SharedPtr<Process>> process_list;
@@ -242,11 +240,11 @@ u32 KernelCore::CreateNewObjectID() {
     return impl->next_object_id++;
 }
 
-u32 KernelCore::CreateNewThreadID() {
+u64 KernelCore::CreateNewThreadID() {
     return impl->next_thread_id++;
 }
 
-u32 KernelCore::CreateNewProcessID() {
+u64 KernelCore::CreateNewProcessID() {
     return impl->next_process_id++;
 }
 

src/core/hle/kernel/kernel.h

@@ -88,10 +88,10 @@ private:
     u32 CreateNewObjectID();
 
     /// Creates a new process ID, incrementing the internal process ID counter;
-    u32 CreateNewProcessID();
+    u64 CreateNewProcessID();
 
     /// Creates a new thread ID, incrementing the internal thread ID counter.
-    u32 CreateNewThreadID();
+    u64 CreateNewThreadID();
 
     /// Creates a timer callback handle for the given timer.
     ResultVal<Handle> CreateTimerCallbackHandle(const SharedPtr<Timer>& timer);

src/core/hle/kernel/object.cpp

@@ -15,6 +15,7 @@ bool Object::IsWaitable() const {
     switch (GetHandleType()) {
     case HandleType::ReadableEvent:
     case HandleType::Thread:
+    case HandleType::Process:
     case HandleType::Timer:
     case HandleType::ServerPort:
     case HandleType::ServerSession:
@@ -23,7 +24,6 @@ bool Object::IsWaitable() const {
     case HandleType::Unknown:
     case HandleType::WritableEvent:
     case HandleType::SharedMemory:
-    case HandleType::Process:
     case HandleType::AddressArbiter:
     case HandleType::ResourceLimit:
     case HandleType::ClientPort:
@@ -32,6 +32,7 @@ bool Object::IsWaitable() const {
     }
 
     UNREACHABLE();
+    return false;
 }
 
 } // namespace Kernel

src/core/hle/kernel/process.cpp

@@ -9,6 +9,7 @@
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "core/file_sys/program_metadata.h"
+#include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/resource_limit.h"
@@ -48,8 +49,24 @@ SharedPtr<ResourceLimit> Process::GetResourceLimit() const {
     return resource_limit;
 }
 
+ResultCode Process::ClearSignalState() {
+    if (status == ProcessStatus::Exited) {
+        LOG_ERROR(Kernel, "called on a terminated process instance.");
+        return ERR_INVALID_STATE;
+    }
+
+    if (!is_signaled) {
+        LOG_ERROR(Kernel, "called on a process instance that isn't signaled.");
+        return ERR_INVALID_STATE;
+    }
+
+    is_signaled = false;
+    return RESULT_SUCCESS;
+}
+
 void Process::LoadFromMetadata(const FileSys::ProgramMetadata& metadata) {
     program_id = metadata.GetTitleID();
+    ideal_processor = metadata.GetMainThreadCore();
     is_64bit_process = metadata.Is64BitProgram();
     vm_manager.Reset(metadata.GetAddressSpaceType());
 }
@@ -133,17 +150,17 @@ void Process::Run(VAddr entry_point, s32 main_thread_priority, u32 stack_size) {
     vm_manager
         .MapMemoryBlock(vm_manager.GetTLSIORegionEndAddress() - stack_size,
                         std::make_shared<std::vector<u8>>(stack_size, 0), 0, stack_size,
-                        MemoryState::Mapped)
+                        MemoryState::Stack)
         .Unwrap();
 
     vm_manager.LogLayout();
-    status = ProcessStatus::Running;
+    ChangeStatus(ProcessStatus::Running);
 
     Kernel::SetupMainThread(kernel, entry_point, main_thread_priority, *this);
 }
 
 void Process::PrepareForTermination() {
-    status = ProcessStatus::Exited;
+    ChangeStatus(ProcessStatus::Exiting);
 
     const auto stop_threads = [this](const std::vector<SharedPtr<Thread>>& thread_list) {
         for (auto& thread : thread_list) {
@@ -167,6 +184,8 @@ void Process::PrepareForTermination() {
     stop_threads(system.Scheduler(1).GetThreadList());
     stop_threads(system.Scheduler(2).GetThreadList());
     stop_threads(system.Scheduler(3).GetThreadList());
+
+    ChangeStatus(ProcessStatus::Exited);
 }
 
 /**
@@ -265,7 +284,25 @@ ResultCode Process::UnmapMemory(VAddr dst_addr, VAddr /*src_addr*/, u64 size) {
     return vm_manager.UnmapRange(dst_addr, size);
 }
 
-Kernel::Process::Process(KernelCore& kernel) : Object{kernel} {}
+Kernel::Process::Process(KernelCore& kernel) : WaitObject{kernel} {}
 Kernel::Process::~Process() {}
 
+void Process::Acquire(Thread* thread) {
+    ASSERT_MSG(!ShouldWait(thread), "Object unavailable!");
+}
+
+bool Process::ShouldWait(Thread* thread) const {
+    return !is_signaled;
+}
+
+void Process::ChangeStatus(ProcessStatus new_status) {
+    if (status == new_status) {
+        return;
+    }
+
+    status = new_status;
+    is_signaled = true;
+    WakeupAllWaitingThreads();
+}
+
 } // namespace Kernel

src/core/hle/kernel/process.h

@@ -14,9 +14,10 @@
 #include "common/bit_field.h"
 #include "common/common_types.h"
 #include "core/hle/kernel/handle_table.h"
-#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/vm_manager.h"
+#include "core/hle/kernel/wait_object.h"
+#include "core/hle/result.h"
 
 namespace FileSys {
 class ProgramMetadata;
@@ -117,8 +118,20 @@ struct CodeSet final {
     VAddr entrypoint = 0;
 };
 
-class Process final : public Object {
+class Process final : public WaitObject {
 public:
+    enum : u64 {
+        /// Lowest allowed process ID for a kernel initial process.
+        InitialKIPIDMin = 1,
+        /// Highest allowed process ID for a kernel initial process.
+        InitialKIPIDMax = 80,
+
+        /// Lowest allowed process ID for a userland process.
+        ProcessIDMin = 81,
+        /// Highest allowed process ID for a userland process.
+        ProcessIDMax = 0xFFFFFFFFFFFFFFFF,
+    };
+
     static constexpr std::size_t RANDOM_ENTROPY_SIZE = 4;
 
     static SharedPtr<Process> Create(KernelCore& kernel, std::string&& name);
@@ -161,7 +174,7 @@ public:
     }
 
     /// Gets the unique ID that identifies this particular process.
-    u32 GetProcessID() const {
+    u64 GetProcessID() const {
         return process_id;
     }
 
@@ -212,6 +225,16 @@ public:
         return random_entropy.at(index);
     }
 
+    /// 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
+    ///      terminated process, then ERR_INVALID_STATE will be returned.
+    ///
+    /// @pre The process must be in a signaled state. If this is called on a
+    ///      process instance that is not signaled, ERR_INVALID_STATE will be
+    ///      returned.
+    ResultCode ClearSignalState();
+
     /**
      * Loads process-specifics configuration info with metadata provided
      * by an executable.
@@ -251,8 +274,7 @@ public:
     ResultVal<VAddr> HeapAllocate(VAddr target, u64 size, VMAPermission perms);
     ResultCode HeapFree(VAddr target, u32 size);
 
-    ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size,
-                            MemoryState state = MemoryState::Mapped);
+    ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size, MemoryState state);
 
     ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, u64 size);
 
@@ -260,6 +282,17 @@ private:
     explicit Process(KernelCore& kernel);
     ~Process() override;
 
+    /// Checks if the specified thread should wait until this process is available.
+    bool ShouldWait(Thread* thread) const override;
+
+    /// Acquires/locks this process for the specified thread if it's available.
+    void Acquire(Thread* thread) override;
+
+    /// Changes the process status. If the status is different
+    /// from the current process status, then this will trigger
+    /// a process signal.
+    void ChangeStatus(ProcessStatus new_status);
+
     /// Memory manager for this process.
     Kernel::VMManager vm_manager;
 
@@ -267,10 +300,10 @@ private:
     ProcessStatus status;
 
     /// The ID of this process
-    u32 process_id = 0;
+    u64 process_id = 0;
 
     /// Title ID corresponding to the process
-    u64 program_id;
+    u64 program_id = 0;
 
     /// Resource limit descriptor for this process
     SharedPtr<ResourceLimit> resource_limit;
@@ -305,6 +338,10 @@ private:
     /// specified by metadata provided to the process during loading.
     bool is_64bit_process = true;
 
+    /// Whether or not this process is signaled. This occurs
+    /// upon the process changing to a different state.
+    bool is_signaled = false;
+
     /// Total running time for the process in ticks.
     u64 total_process_running_time_ticks = 0;
 

src/core/hle/kernel/readable_event.cpp

@@ -4,10 +4,10 @@
 
 #include <algorithm>
 #include "common/assert.h"
+#include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/readable_event.h"
 #include "core/hle/kernel/thread.h"
-#include "core/hle/kernel/writable_event.h"
 
 namespace Kernel {
 
@@ -34,6 +34,16 @@ void ReadableEvent::Clear() {
     signaled = false;
 }
 
+ResultCode ReadableEvent::Reset() {
+    if (!signaled) {
+        return ERR_INVALID_STATE;
+    }
+
+    Clear();
+
+    return RESULT_SUCCESS;
+}
+
 void ReadableEvent::WakeupAllWaitingThreads() {
     WaitObject::WakeupAllWaitingThreads();
 

src/core/hle/kernel/readable_event.h

@@ -7,6 +7,8 @@
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/wait_object.h"
 
+union ResultCode;
+
 namespace Kernel {
 
 class KernelCore;
@@ -39,8 +41,17 @@ public:
 
     void WakeupAllWaitingThreads() override;
 
+    /// Unconditionally clears the readable event's state.
     void Clear();
 
+    /// Clears the readable event's state if and only if it
+    /// has already been signaled.
+    ///
+    /// @pre The event must be in a signaled state. If this event
+    ///      is in an unsignaled state and this function is called,
+    ///      then ERR_INVALID_STATE will be returned.
+    ResultCode Reset();
+
 private:
     explicit ReadableEvent(KernelCore& kernel);
 

src/core/hle/kernel/scheduler.cpp

@@ -9,6 +9,7 @@
 #include "common/logging/log.h"
 #include "core/arm/arm_interface.h"
 #include "core/core.h"
+#include "core/core_cpu.h"
 #include "core/core_timing.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
@@ -179,4 +180,69 @@ void Scheduler::SetThreadPriority(Thread* thread, u32 priority) {
         ready_queue.prepare(priority);
 }
 
+Thread* Scheduler::GetNextSuggestedThread(u32 core, u32 maximum_priority) const {
+    std::lock_guard<std::mutex> lock(scheduler_mutex);
+
+    const u32 mask = 1U << core;
+    return ready_queue.get_first_filter([mask, maximum_priority](Thread const* thread) {
+        return (thread->GetAffinityMask() & mask) != 0 && thread->GetPriority() < maximum_priority;
+    });
+}
+
+void Scheduler::YieldWithoutLoadBalancing(Thread* thread) {
+    ASSERT(thread != nullptr);
+    // Avoid yielding if the thread isn't even running.
+    ASSERT(thread->GetStatus() == ThreadStatus::Running);
+
+    // Sanity check that the priority is valid
+    ASSERT(thread->GetPriority() < THREADPRIO_COUNT);
+
+    // Yield this thread -- sleep for zero time and force reschedule to different thread
+    WaitCurrentThread_Sleep();
+    GetCurrentThread()->WakeAfterDelay(0);
+}
+
+void Scheduler::YieldWithLoadBalancing(Thread* thread) {
+    ASSERT(thread != nullptr);
+    const auto priority = thread->GetPriority();
+    const auto core = static_cast<u32>(thread->GetProcessorID());
+
+    // Avoid yielding if the thread isn't even running.
+    ASSERT(thread->GetStatus() == ThreadStatus::Running);
+
+    // Sanity check that the priority is valid
+    ASSERT(priority < THREADPRIO_COUNT);
+
+    // Sleep for zero time to be able to force reschedule to different thread
+    WaitCurrentThread_Sleep();
+    GetCurrentThread()->WakeAfterDelay(0);
+
+    Thread* suggested_thread = nullptr;
+
+    // Search through all of the cpu cores (except this one) for a suggested thread.
+    // Take the first non-nullptr one
+    for (unsigned cur_core = 0; cur_core < Core::NUM_CPU_CORES; ++cur_core) {
+        const auto res =
+            Core::System::GetInstance().CpuCore(cur_core).Scheduler().GetNextSuggestedThread(
+                core, priority);
+
+        // If scheduler provides a suggested thread
+        if (res != nullptr) {
+            // And its better than the current suggested thread (or is the first valid one)
+            if (suggested_thread == nullptr ||
+                suggested_thread->GetPriority() > res->GetPriority()) {
+                suggested_thread = res;
+            }
+        }
+    }
+
+    // If a suggested thread was found, queue that for this core
+    if (suggested_thread != nullptr)
+        suggested_thread->ChangeCore(core, suggested_thread->GetAffinityMask());
+}
+
+void Scheduler::YieldAndWaitForLoadBalancing(Thread* thread) {
+    UNIMPLEMENTED_MSG("Wait for load balancing thread yield type is not implemented!");
+}
+
 } // namespace Kernel

src/core/hle/kernel/scheduler.h

@@ -51,6 +51,75 @@ public:
     /// Sets the priority of a thread in the scheduler
     void SetThreadPriority(Thread* thread, u32 priority);
 
+    /// Gets the next suggested thread for load balancing
+    Thread* GetNextSuggestedThread(u32 core, u32 minimum_priority) const;
+
+    /**
+     * YieldWithoutLoadBalancing -- analogous to normal yield on a system
+     * Moves the thread to the end of the ready queue for its priority, and then reschedules the
+     * system to the new head of the queue.
+     *
+     * Example (Single Core -- but can be extrapolated to multi):
+     * ready_queue[prio=0]: ThreadA, ThreadB, ThreadC (->exec order->)
+     * Currently Running: ThreadR
+     *
+     * ThreadR calls YieldWithoutLoadBalancing
+     *
+     * ThreadR is moved to the end of ready_queue[prio=0]:
+     * ready_queue[prio=0]: ThreadA, ThreadB, ThreadC, ThreadR (->exec order->)
+     * Currently Running: Nothing
+     *
+     * System is rescheduled (ThreadA is popped off of queue):
+     * ready_queue[prio=0]: ThreadB, ThreadC, ThreadR (->exec order->)
+     * Currently Running: ThreadA
+     *
+     * If the queue is empty at time of call, no yielding occurs. This does not cross between cores
+     * or priorities at all.
+     */
+    void YieldWithoutLoadBalancing(Thread* thread);
+
+    /**
+     * YieldWithLoadBalancing -- yield but with better selection of the new running thread
+     * Moves the current thread to the end of the ready queue for its priority, then selects a
+     * 'suggested thread' (a thread on a different core that could run on this core) from the
+     * scheduler, changes its core, and reschedules the current core to that thread.
+     *
+     * Example (Dual Core -- can be extrapolated to Quad Core, this is just normal yield if it were
+     * single core):
+     * ready_queue[core=0][prio=0]: ThreadA, ThreadB (affinities not pictured as irrelevant
+     * ready_queue[core=1][prio=0]: ThreadC[affinity=both], ThreadD[affinity=core1only]
+     * Currently Running: ThreadQ on Core 0 || ThreadP on Core 1
+     *
+     * ThreadQ calls YieldWithLoadBalancing
+     *
+     * ThreadQ is moved to the end of ready_queue[core=0][prio=0]:
+     * ready_queue[core=0][prio=0]: ThreadA, ThreadB
+     * ready_queue[core=1][prio=0]: ThreadC[affinity=both], ThreadD[affinity=core1only]
+     * Currently Running: ThreadQ on Core 0 || ThreadP on Core 1
+     *
+     * A list of suggested threads for each core is compiled
+     * Suggested Threads: {ThreadC on Core 1}
+     * If this were quad core (as the switch is), there could be between 0 and 3 threads in this
+     * list. If there are more than one, the thread is selected by highest prio.
+     *
+     * ThreadC is core changed to Core 0:
+     * ready_queue[core=0][prio=0]: ThreadC, ThreadA, ThreadB, ThreadQ
+     * ready_queue[core=1][prio=0]: ThreadD
+     * Currently Running: None on Core 0 || ThreadP on Core 1
+     *
+     * System is rescheduled (ThreadC is popped off of queue):
+     * ready_queue[core=0][prio=0]: ThreadA, ThreadB, ThreadQ
+     * ready_queue[core=1][prio=0]: ThreadD
+     * Currently Running: ThreadC on Core 0 || ThreadP on Core 1
+     *
+     * If no suggested threads can be found this will behave just as normal yield. If there are
+     * multiple candidates for the suggested thread on a core, the highest prio is taken.
+     */
+    void YieldWithLoadBalancing(Thread* thread);
+
+    /// Currently unknown -- asserts as unimplemented on call
+    void YieldAndWaitForLoadBalancing(Thread* thread);
+
     /// Returns a list of all threads managed by the scheduler
     const std::vector<SharedPtr<Thread>>& GetThreadList() const {
         return thread_list;

src/core/hle/kernel/shared_memory.cpp

@@ -17,13 +17,13 @@ namespace Kernel {
 SharedMemory::SharedMemory(KernelCore& kernel) : Object{kernel} {}
 SharedMemory::~SharedMemory() = default;
 
-SharedPtr<SharedMemory> SharedMemory::Create(KernelCore& kernel, SharedPtr<Process> owner_process,
-                                             u64 size, MemoryPermission permissions,
+SharedPtr<SharedMemory> SharedMemory::Create(KernelCore& kernel, Process* owner_process, u64 size,
+                                             MemoryPermission permissions,
                                              MemoryPermission other_permissions, VAddr address,
                                              MemoryRegion region, std::string name) {
     SharedPtr<SharedMemory> shared_memory(new SharedMemory(kernel));
 
-    shared_memory->owner_process = std::move(owner_process);
+    shared_memory->owner_process = owner_process;
     shared_memory->name = std::move(name);
     shared_memory->size = size;
     shared_memory->permissions = permissions;
@@ -39,15 +39,15 @@ SharedPtr<SharedMemory> SharedMemory::Create(KernelCore& kernel, SharedPtr<Proce
                 shared_memory->backing_block.get());
         }
     } else {
-        auto& vm_manager = shared_memory->owner_process->VMManager();
+        const auto& vm_manager = shared_memory->owner_process->VMManager();
 
         // The memory is already available and mapped in the owner process.
-        auto vma = vm_manager.FindVMA(address);
-        ASSERT_MSG(vma != vm_manager.vma_map.end(), "Invalid memory address");
+        const auto vma = vm_manager.FindVMA(address);
+        ASSERT_MSG(vm_manager.IsValidHandle(vma), "Invalid memory address");
         ASSERT_MSG(vma->second.backing_block, "Backing block doesn't exist for address");
 
         // The returned VMA might be a bigger one encompassing the desired address.
-        auto vma_offset = address - vma->first;
+        const auto vma_offset = address - vma->first;
         ASSERT_MSG(vma_offset + size <= vma->second.size,
                    "Shared memory exceeds bounds of mapped block");
 

src/core/hle/kernel/shared_memory.h

@@ -45,8 +45,8 @@ public:
      * linear heap.
      * @param name Optional object name, used for debugging purposes.
      */
-    static SharedPtr<SharedMemory> Create(KernelCore& kernel, SharedPtr<Process> owner_process,
-                                          u64 size, MemoryPermission permissions,
+    static SharedPtr<SharedMemory> Create(KernelCore& kernel, Process* owner_process, u64 size,
+                                          MemoryPermission permissions,
                                           MemoryPermission other_permissions, VAddr address = 0,
                                           MemoryRegion region = MemoryRegion::BASE,
                                           std::string name = "Unknown");
@@ -139,7 +139,7 @@ private:
     /// Permission restrictions applied to other processes mapping the block.
     MemoryPermission other_permissions{};
     /// Process that created this shared memory block.
-    SharedPtr<Process> owner_process;
+    Process* owner_process;
     /// Address of shared memory block in the owner process if specified.
     VAddr base_address = 0;
     /// Name of shared memory object.

src/core/hle/kernel/svc.cpp

@@ -35,6 +35,7 @@
 #include "core/hle/lock.h"
 #include "core/hle/result.h"
 #include "core/hle/service/service.h"
+#include "core/memory.h"
 
 namespace Kernel {
 namespace {
@@ -239,7 +240,7 @@ static ResultCode SetMemoryPermission(VAddr addr, u64 size, u32 prot) {
     }
 
     const VMManager::VMAHandle iter = vm_manager.FindVMA(addr);
-    if (iter == vm_manager.vma_map.end()) {
+    if (!vm_manager.IsValidHandle(iter)) {
         LOG_ERROR(Kernel_SVC, "Unable to find VMA for address=0x{:016X}", addr);
         return ERR_INVALID_ADDRESS_STATE;
     }
@@ -253,11 +254,52 @@ static ResultCode SetMemoryPermission(VAddr addr, u64 size, u32 prot) {
     return vm_manager.ReprotectRange(addr, size, converted_permissions);
 }
 
-static ResultCode SetMemoryAttribute(VAddr addr, u64 size, u32 state0, u32 state1) {
-    LOG_WARNING(Kernel_SVC,
-                "(STUBBED) called, addr=0x{:X}, size=0x{:X}, state0=0x{:X}, state1=0x{:X}", addr,
-                size, state0, state1);
-    return RESULT_SUCCESS;
+static ResultCode SetMemoryAttribute(VAddr address, u64 size, u32 mask, u32 attribute) {
+    LOG_DEBUG(Kernel_SVC,
+              "called, address=0x{:016X}, size=0x{:X}, mask=0x{:08X}, attribute=0x{:08X}", address,
+              size, mask, attribute);
+
+    if (!Common::Is4KBAligned(address)) {
+        LOG_ERROR(Kernel_SVC, "Address not page aligned (0x{:016X})", address);
+        return ERR_INVALID_ADDRESS;
+    }
+
+    if (size == 0 || !Common::Is4KBAligned(size)) {
+        LOG_ERROR(Kernel_SVC, "Invalid size (0x{:X}). Size must be non-zero and page aligned.",
+                  size);
+        return ERR_INVALID_ADDRESS;
+    }
+
+    if (!IsValidAddressRange(address, size)) {
+        LOG_ERROR(Kernel_SVC, "Address range overflowed (Address: 0x{:016X}, Size: 0x{:016X})",
+                  address, size);
+        return ERR_INVALID_ADDRESS_STATE;
+    }
+
+    const auto mem_attribute = static_cast<MemoryAttribute>(attribute);
+    const auto mem_mask = static_cast<MemoryAttribute>(mask);
+    const auto attribute_with_mask = mem_attribute | mem_mask;
+
+    if (attribute_with_mask != mem_mask) {
+        LOG_ERROR(Kernel_SVC,
+                  "Memory attribute doesn't match the given mask (Attribute: 0x{:X}, Mask: {:X}",
+                  attribute, mask);
+        return ERR_INVALID_COMBINATION;
+    }
+
+    if ((attribute_with_mask | MemoryAttribute::Uncached) != MemoryAttribute::Uncached) {
+        LOG_ERROR(Kernel_SVC, "Specified attribute isn't equal to MemoryAttributeUncached (8).");
+        return ERR_INVALID_COMBINATION;
+    }
+
+    auto& vm_manager = Core::CurrentProcess()->VMManager();
+    if (!IsInsideAddressSpace(vm_manager, address, size)) {
+        LOG_ERROR(Kernel_SVC,
+                  "Given address (0x{:016X}) is outside the bounds of the address space.", address);
+        return ERR_INVALID_ADDRESS_STATE;
+    }
+
+    return vm_manager.SetMemoryAttribute(address, size, mem_mask, mem_attribute);
 }
 
 /// Maps a memory range into a different range.
@@ -273,7 +315,7 @@ static ResultCode MapMemory(VAddr dst_addr, VAddr src_addr, u64 size) {
         return result;
     }
 
-    return current_process->MirrorMemory(dst_addr, src_addr, size);
+    return current_process->MirrorMemory(dst_addr, src_addr, size, MemoryState::Stack);
 }
 
 /// Unmaps a region that was previously mapped with svcMapMemory
@@ -349,7 +391,7 @@ static ResultCode SendSyncRequest(Handle handle) {
 }
 
 /// Get the ID for the specified thread.
-static ResultCode GetThreadId(u32* thread_id, Handle thread_handle) {
+static ResultCode GetThreadId(u64* thread_id, Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
 
     const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
@@ -363,20 +405,33 @@ static ResultCode GetThreadId(u32* thread_id, Handle thread_handle) {
     return RESULT_SUCCESS;
 }
 
-/// Get the ID of the specified process
-static ResultCode GetProcessId(u32* process_id, Handle process_handle) {
-    LOG_TRACE(Kernel_SVC, "called process=0x{:08X}", process_handle);
+/// Gets the ID of the specified process or a specified thread's owning process.
+static ResultCode GetProcessId(u64* process_id, Handle handle) {
+    LOG_DEBUG(Kernel_SVC, "called handle=0x{:08X}", handle);
 
     const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
-    const SharedPtr<Process> process = handle_table.Get<Process>(process_handle);
-    if (!process) {
-        LOG_ERROR(Kernel_SVC, "Process handle does not exist, process_handle=0x{:08X}",
-                  process_handle);
-        return ERR_INVALID_HANDLE;
+    const SharedPtr<Process> process = handle_table.Get<Process>(handle);
+    if (process) {
+        *process_id = process->GetProcessID();
+        return RESULT_SUCCESS;
     }
 
-    *process_id = process->GetProcessID();
-    return RESULT_SUCCESS;
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(handle);
+    if (thread) {
+        const Process* const owner_process = thread->GetOwnerProcess();
+        if (!owner_process) {
+            LOG_ERROR(Kernel_SVC, "Non-existent owning process encountered.");
+            return ERR_INVALID_HANDLE;
+        }
+
+        *process_id = owner_process->GetProcessID();
+        return RESULT_SUCCESS;
+    }
+
+    // NOTE: This should also handle debug objects before returning.
+
+    LOG_ERROR(Kernel_SVC, "Handle does not exist, handle=0x{:08X}", handle);
+    return ERR_INVALID_HANDLE;
 }
 
 /// Default thread wakeup callback for WaitSynchronization
@@ -877,8 +932,35 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
 }
 
 /// Sets the thread activity
-static ResultCode SetThreadActivity(Handle handle, u32 unknown) {
-    LOG_WARNING(Kernel_SVC, "(STUBBED) called, handle=0x{:08X}, unknown=0x{:08X}", handle, unknown);
+static ResultCode SetThreadActivity(Handle handle, u32 activity) {
+    LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, activity=0x{:08X}", handle, activity);
+    if (activity > static_cast<u32>(ThreadActivity::Paused)) {
+        return ERR_INVALID_ENUM_VALUE;
+    }
+
+    const auto* current_process = Core::CurrentProcess();
+    const SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
+    if (!thread) {
+        LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle);
+        return ERR_INVALID_HANDLE;
+    }
+
+    if (thread->GetOwnerProcess() != current_process) {
+        LOG_ERROR(Kernel_SVC,
+                  "The current process does not own the current thread, thread_handle={:08X} "
+                  "thread_pid={}, "
+                  "current_process_pid={}",
+                  handle, thread->GetOwnerProcess()->GetProcessID(),
+                  current_process->GetProcessID());
+        return ERR_INVALID_HANDLE;
+    }
+
+    if (thread == GetCurrentThread()) {
+        LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread");
+        return ERR_BUSY;
+    }
+
+    thread->SetActivity(static_cast<ThreadActivity>(activity));
     return RESULT_SUCCESS;
 }
 
@@ -905,7 +987,7 @@ static ResultCode GetThreadContext(VAddr thread_context, Handle handle) {
 
     if (thread == GetCurrentThread()) {
         LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread");
-        return ERR_ALREADY_REGISTERED;
+        return ERR_BUSY;
     }
 
     Core::ARM_Interface::ThreadContext ctx = thread->GetContext();
@@ -1066,10 +1148,9 @@ static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64
     return shared_memory->Unmap(*current_process, addr);
 }
 
-/// Query process memory
-static ResultCode QueryProcessMemory(MemoryInfo* memory_info, PageInfo* /*page_info*/,
-                                     Handle process_handle, u64 addr) {
-    LOG_TRACE(Kernel_SVC, "called process=0x{:08X} addr={:X}", process_handle, addr);
+static ResultCode QueryProcessMemory(VAddr memory_info_address, VAddr page_info_address,
+                                     Handle process_handle, VAddr address) {
+    LOG_TRACE(Kernel_SVC, "called process=0x{:08X} address={:X}", process_handle, address);
     const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
     SharedPtr<Process> process = handle_table.Get<Process>(process_handle);
     if (!process) {
@@ -1077,26 +1158,34 @@ static ResultCode QueryProcessMemory(MemoryInfo* memory_info, PageInfo* /*page_i
                   process_handle);
         return ERR_INVALID_HANDLE;
     }
-    auto vma = process->VMManager().FindVMA(addr);
-    memory_info->attributes = 0;
-    if (vma == process->VMManager().vma_map.end()) {
-        memory_info->base_address = 0;
-        memory_info->permission = static_cast<u32>(VMAPermission::None);
-        memory_info->size = 0;
-        memory_info->type = static_cast<u32>(MemoryState::Unmapped);
-    } else {
-        memory_info->base_address = vma->second.base;
-        memory_info->permission = static_cast<u32>(vma->second.permissions);
-        memory_info->size = vma->second.size;
-        memory_info->type = static_cast<u32>(vma->second.meminfo_state);
-    }
+
+    const auto& vm_manager = process->VMManager();
+    const MemoryInfo memory_info = vm_manager.QueryMemory(address);
+
+    Memory::Write64(memory_info_address, memory_info.base_address);
+    Memory::Write64(memory_info_address + 8, memory_info.size);
+    Memory::Write32(memory_info_address + 16, memory_info.state);
+    Memory::Write32(memory_info_address + 20, memory_info.attributes);
+    Memory::Write32(memory_info_address + 24, memory_info.permission);
+    Memory::Write32(memory_info_address + 32, memory_info.ipc_ref_count);
+    Memory::Write32(memory_info_address + 28, memory_info.device_ref_count);
+    Memory::Write32(memory_info_address + 36, 0);
+
+    // Page info appears to be currently unused by the kernel and is always set to zero.
+    Memory::Write32(page_info_address, 0);
+
     return RESULT_SUCCESS;
 }
 
-/// Query memory
-static ResultCode QueryMemory(MemoryInfo* memory_info, PageInfo* page_info, VAddr addr) {
-    LOG_TRACE(Kernel_SVC, "called, addr={:X}", addr);
-    return QueryProcessMemory(memory_info, page_info, CurrentProcess, addr);
+static ResultCode QueryMemory(VAddr memory_info_address, VAddr page_info_address,
+                              VAddr query_address) {
+    LOG_TRACE(Kernel_SVC,
+              "called, memory_info_address=0x{:016X}, page_info_address=0x{:016X}, "
+              "query_address=0x{:016X}",
+              memory_info_address, page_info_address, query_address);
+
+    return QueryProcessMemory(memory_info_address, page_info_address, CurrentProcess,
+                              query_address);
 }
 
 /// Exits the current process
@@ -1183,7 +1272,10 @@ static ResultCode StartThread(Handle thread_handle) {
     ASSERT(thread->GetStatus() == ThreadStatus::Dormant);
 
     thread->ResumeFromWait();
-    Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule();
+
+    if (thread->GetStatus() == ThreadStatus::Ready) {
+        Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule();
+    }
 
     return RESULT_SUCCESS;
 }
@@ -1200,18 +1292,38 @@ static void ExitThread() {
 static void SleepThread(s64 nanoseconds) {
     LOG_TRACE(Kernel_SVC, "called nanoseconds={}", nanoseconds);
 
-    // Don't attempt to yield execution if there are no available threads to run,
-    // this way we avoid a useless reschedule to the idle thread.
-    if (nanoseconds == 0 && !Core::System::GetInstance().CurrentScheduler().HaveReadyThreads())
-        return;
+    enum class SleepType : s64 {
+        YieldWithoutLoadBalancing = 0,
+        YieldWithLoadBalancing = -1,
+        YieldAndWaitForLoadBalancing = -2,
+    };
 
-    // Sleep current thread and check for next thread to schedule
-    WaitCurrentThread_Sleep();
+    if (nanoseconds <= 0) {
+        auto& scheduler{Core::System::GetInstance().CurrentScheduler()};
+        switch (static_cast<SleepType>(nanoseconds)) {
+        case SleepType::YieldWithoutLoadBalancing:
+            scheduler.YieldWithoutLoadBalancing(GetCurrentThread());
+            break;
+        case SleepType::YieldWithLoadBalancing:
+            scheduler.YieldWithLoadBalancing(GetCurrentThread());
+            break;
+        case SleepType::YieldAndWaitForLoadBalancing:
+            scheduler.YieldAndWaitForLoadBalancing(GetCurrentThread());
+            break;
+        default:
+            UNREACHABLE_MSG("Unimplemented sleep yield type '{:016X}'!", nanoseconds);
+        }
+    } else {
+        // Sleep current thread and check for next thread to schedule
+        WaitCurrentThread_Sleep();
 
-    // Create an event to wake the thread up after the specified nanosecond delay has passed
-    GetCurrentThread()->WakeAfterDelay(nanoseconds);
+        // Create an event to wake the thread up after the specified nanosecond delay has passed
+        GetCurrentThread()->WakeAfterDelay(nanoseconds);
+    }
 
-    Core::System::GetInstance().PrepareReschedule();
+    // Reschedule all CPU cores
+    for (std::size_t i = 0; i < Core::NUM_CPU_CORES; ++i)
+        Core::System::GetInstance().CpuCore(i).PrepareReschedule();
 }
 
 /// Wait process wide key atomic
@@ -1433,17 +1545,24 @@ static ResultCode CloseHandle(Handle handle) {
     return handle_table.Close(handle);
 }
 
-/// Reset an event
+/// Clears the signaled state of an event or process.
 static ResultCode ResetSignal(Handle handle) {
     LOG_DEBUG(Kernel_SVC, "called handle 0x{:08X}", handle);
 
     const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+
     auto event = handle_table.Get<ReadableEvent>(handle);
+    if (event) {
+        return event->Reset();
+    }
 
-    ASSERT(event != nullptr);
+    auto process = handle_table.Get<Process>(handle);
+    if (process) {
+        return process->ClearSignalState();
+    }
 
-    event->Clear();
-    return RESULT_SUCCESS;
+    LOG_ERROR(Kernel_SVC, "Invalid handle (0x{:08X})", handle);
+    return ERR_INVALID_HANDLE;
 }
 
 /// Creates a TransferMemory object
@@ -1476,9 +1595,9 @@ static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32
     }
 
     auto& kernel = Core::System::GetInstance().Kernel();
-    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
-    const auto shared_mem_handle = SharedMemory::Create(
-        kernel, handle_table.Get<Process>(CurrentProcess), size, perms, perms, addr);
+    auto process = kernel.CurrentProcess();
+    auto& handle_table = process->GetHandleTable();
+    const auto shared_mem_handle = SharedMemory::Create(kernel, process, size, perms, perms, addr);
 
     CASCADE_RESULT(*handle, handle_table.Create(shared_mem_handle));
     return RESULT_SUCCESS;
@@ -1588,10 +1707,9 @@ static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permiss
     }
 
     auto& kernel = Core::System::GetInstance().Kernel();
-    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
-    auto shared_mem_handle =
-        SharedMemory::Create(kernel, handle_table.Get<Process>(KernelHandle::CurrentProcess), size,
-                             local_perms, remote_perms);
+    auto process = kernel.CurrentProcess();
+    auto& handle_table = process->GetHandleTable();
+    auto shared_mem_handle = SharedMemory::Create(kernel, process, size, local_perms, remote_perms);
 
     CASCADE_RESULT(*handle, handle_table.Create(shared_mem_handle));
     return RESULT_SUCCESS;
@@ -1897,7 +2015,7 @@ static const FunctionDef SVC_Table[] = {
     {0x73, nullptr, "SetProcessMemoryPermission"},
     {0x74, nullptr, "MapProcessMemory"},
     {0x75, nullptr, "UnmapProcessMemory"},
-    {0x76, nullptr, "QueryProcessMemory"},
+    {0x76, SvcWrap<QueryProcessMemory>, "QueryProcessMemory"},
     {0x77, nullptr, "MapProcessCodeMemory"},
     {0x78, nullptr, "UnmapProcessCodeMemory"},
     {0x79, nullptr, "CreateProcess"},

src/core/hle/kernel/svc.h

@@ -8,22 +8,6 @@
 
 namespace Kernel {
 
-struct MemoryInfo {
-    u64 base_address;
-    u64 size;
-    u32 type;
-    u32 attributes;
-    u32 permission;
-    u32 device_refcount;
-    u32 ipc_refcount;
-    INSERT_PADDING_WORDS(1);
-};
-static_assert(sizeof(MemoryInfo) == 0x28, "MemoryInfo has incorrect size.");
-
-struct PageInfo {
-    u64 flags;
-};
-
 void CallSVC(u32 immediate);
 
 } // namespace Kernel

src/core/hle/kernel/svc_wrap.h

@@ -7,9 +7,7 @@
 #include "common/common_types.h"
 #include "core/arm/arm_interface.h"
 #include "core/core.h"
-#include "core/hle/kernel/svc.h"
 #include "core/hle/result.h"
-#include "core/memory.h"
 
 namespace Kernel {
 
@@ -75,7 +73,15 @@ void SvcWrap() {
 template <ResultCode func(u32*, u64)>
 void SvcWrap() {
     u32 param_1 = 0;
-    u32 retval = func(&param_1, Param(1)).raw;
+    const u32 retval = func(&param_1, Param(1)).raw;
+    Core::CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(retval);
+}
+
+template <ResultCode func(u64*, u32)>
+void SvcWrap() {
+    u64 param_1 = 0;
+    const u32 retval = func(&param_1, static_cast<u32>(Param(1))).raw;
     Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
@@ -129,7 +135,12 @@ void SvcWrap() {
 template <ResultCode func(u64, u64, u32, u32)>
 void SvcWrap() {
     FuncReturn(
-        func(Param(0), Param(1), static_cast<u32>(Param(3)), static_cast<u32>(Param(3))).raw);
+        func(Param(0), Param(1), static_cast<u32>(Param(2)), static_cast<u32>(Param(3))).raw);
+}
+
+template <ResultCode func(u64, u64, u32, u64)>
+void SvcWrap() {
+    FuncReturn(func(Param(0), Param(1), static_cast<u32>(Param(2)), Param(3)).raw);
 }
 
 template <ResultCode func(u32, u64, u32)>
@@ -191,21 +202,6 @@ void SvcWrap() {
     FuncReturn(retval);
 }
 
-template <ResultCode func(MemoryInfo*, PageInfo*, u64)>
-void SvcWrap() {
-    MemoryInfo memory_info = {};
-    PageInfo page_info = {};
-    u32 retval = func(&memory_info, &page_info, Param(2)).raw;
-
-    Memory::Write64(Param(0), memory_info.base_address);
-    Memory::Write64(Param(0) + 8, memory_info.size);
-    Memory::Write32(Param(0) + 16, memory_info.type);
-    Memory::Write32(Param(0) + 20, memory_info.attributes);
-    Memory::Write32(Param(0) + 24, memory_info.permission);
-
-    FuncReturn(retval);
-}
-
 template <ResultCode func(u32*, u64, u64, u32)>
 void SvcWrap() {
     u32 param_1 = 0;

src/core/hle/kernel/thread.cpp

@@ -50,7 +50,7 @@ void Thread::Stop() {
 
     // Clean up thread from ready queue
     // This is only needed when the thread is terminated forcefully (SVC TerminateProcess)
-    if (status == ThreadStatus::Ready) {
+    if (status == ThreadStatus::Ready || status == ThreadStatus::Paused) {
         scheduler->UnscheduleThread(this, current_priority);
     }
 
@@ -140,6 +140,11 @@ void Thread::ResumeFromWait() {
 
     wakeup_callback = nullptr;
 
+    if (activity == ThreadActivity::Paused) {
+        status = ThreadStatus::Paused;
+        return;
+    }
+
     status = ThreadStatus::Ready;
 
     ChangeScheduler();
@@ -158,6 +163,9 @@ static void ResetThreadContext(Core::ARM_Interface::ThreadContext& context, VAdd
     context.cpu_registers[0] = arg;
     context.pc = entry_point;
     context.sp = stack_top;
+    // TODO(merry): Perform a hardware test to determine the below value.
+    // AHP = 0, DN = 1, FTZ = 1, RMode = Round towards zero
+    context.fpcr = 0x03C00000;
 }
 
 ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name, VAddr entry_point,
@@ -388,6 +396,23 @@ bool Thread::InvokeWakeupCallback(ThreadWakeupReason reason, SharedPtr<Thread> t
     return wakeup_callback(reason, std::move(thread), std::move(object), index);
 }
 
+void Thread::SetActivity(ThreadActivity value) {
+    activity = value;
+
+    if (value == ThreadActivity::Paused) {
+        // Set status if not waiting
+        if (status == ThreadStatus::Ready) {
+            status = ThreadStatus::Paused;
+        } else if (status == ThreadStatus::Running) {
+            status = ThreadStatus::Paused;
+            Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule();
+        }
+    } else if (status == ThreadStatus::Paused) {
+        // Ready to reschedule
+        ResumeFromWait();
+    }
+}
+
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 
 /**

src/core/hle/kernel/thread.h

@@ -26,6 +26,7 @@ enum ThreadPriority : u32 {
     THREADPRIO_USERLAND_MAX = 24, ///< Highest thread priority for userland apps
     THREADPRIO_DEFAULT = 44,      ///< Default thread priority for userland apps
     THREADPRIO_LOWEST = 63,       ///< Lowest thread priority
+    THREADPRIO_COUNT = 64,        ///< Total number of possible thread priorities.
 };
 
 enum ThreadProcessorId : s32 {
@@ -44,6 +45,7 @@ enum ThreadProcessorId : s32 {
 enum class ThreadStatus {
     Running,      ///< Currently running
     Ready,        ///< Ready to run
+    Paused,       ///< Paused by SetThreadActivity or debug
     WaitHLEEvent, ///< Waiting for hle event to finish
     WaitSleep,    ///< Waiting due to a SleepThread SVC
     WaitIPC,      ///< Waiting for the reply from an IPC request
@@ -60,6 +62,11 @@ enum class ThreadWakeupReason {
     Timeout // The thread was woken up due to a wait timeout.
 };
 
+enum class ThreadActivity : u32 {
+    Normal = 0,
+    Paused = 1,
+};
+
 class Thread final : public WaitObject {
 public:
     using TLSMemory = std::vector<u8>;
@@ -150,7 +157,7 @@ public:
      * Gets the thread's thread ID
      * @return The thread's ID
      */
-    u32 GetThreadID() const {
+    u64 GetThreadID() const {
         return thread_id;
     }
 
@@ -370,6 +377,12 @@ public:
         return affinity_mask;
     }
 
+    ThreadActivity GetActivity() const {
+        return activity;
+    }
+
+    void SetActivity(ThreadActivity value);
+
 private:
     explicit Thread(KernelCore& kernel);
     ~Thread() override;
@@ -378,7 +391,7 @@ private:
 
     Core::ARM_Interface::ThreadContext context{};
 
-    u32 thread_id = 0;
+    u64 thread_id = 0;
 
     ThreadStatus status = ThreadStatus::Dormant;
 
@@ -438,6 +451,8 @@ private:
     TLSMemoryPtr tls_memory = std::make_shared<TLSMemory>();
 
     std::string name;
+
+    ThreadActivity activity = ThreadActivity::Normal;
 };
 
 /**

src/core/hle/kernel/vm_manager.cpp

@@ -25,19 +25,19 @@ static const char* GetMemoryStateName(MemoryState state) {
         "CodeMutable",      "Heap",
         "Shared",           "Unknown1",
         "ModuleCodeStatic", "ModuleCodeMutable",
-        "IpcBuffer0",       "Mapped",
+        "IpcBuffer0",       "Stack",
         "ThreadLocal",      "TransferMemoryIsolated",
         "TransferMemory",   "ProcessMemory",
-        "Unknown2",         "IpcBuffer1",
+        "Inaccessible",     "IpcBuffer1",
         "IpcBuffer3",       "KernelStack",
     };
 
-    return names[static_cast<int>(state)];
+    return names[ToSvcMemoryState(state)];
 }
 
 bool VirtualMemoryArea::CanBeMergedWith(const VirtualMemoryArea& next) const {
     ASSERT(base + size == next.base);
-    if (permissions != next.permissions || meminfo_state != next.meminfo_state ||
+    if (permissions != next.permissions || state != next.state || attribute != next.attribute ||
         type != next.type) {
         return false;
     }
@@ -87,6 +87,10 @@ VMManager::VMAHandle VMManager::FindVMA(VAddr target) const {
     }
 }
 
+bool VMManager::IsValidHandle(VMAHandle handle) const {
+    return handle != vma_map.cend();
+}
+
 ResultVal<VMManager::VMAHandle> VMManager::MapMemoryBlock(VAddr target,
                                                           std::shared_ptr<std::vector<u8>> block,
                                                           std::size_t offset, u64 size,
@@ -111,7 +115,7 @@ ResultVal<VMManager::VMAHandle> VMManager::MapMemoryBlock(VAddr target,
 
     final_vma.type = VMAType::AllocatedMemoryBlock;
     final_vma.permissions = VMAPermission::ReadWrite;
-    final_vma.meminfo_state = state;
+    final_vma.state = state;
     final_vma.backing_block = std::move(block);
     final_vma.offset = offset;
     UpdatePageTableForVMA(final_vma);
@@ -136,7 +140,7 @@ ResultVal<VMManager::VMAHandle> VMManager::MapBackingMemory(VAddr target, u8* me
 
     final_vma.type = VMAType::BackingMemory;
     final_vma.permissions = VMAPermission::ReadWrite;
-    final_vma.meminfo_state = state;
+    final_vma.state = state;
     final_vma.backing_memory = memory;
     UpdatePageTableForVMA(final_vma);
 
@@ -173,7 +177,7 @@ ResultVal<VMManager::VMAHandle> VMManager::MapMMIO(VAddr target, PAddr paddr, u6
 
     final_vma.type = VMAType::MMIO;
     final_vma.permissions = VMAPermission::ReadWrite;
-    final_vma.meminfo_state = state;
+    final_vma.state = state;
     final_vma.paddr = paddr;
     final_vma.mmio_handler = std::move(mmio_handler);
     UpdatePageTableForVMA(final_vma);
@@ -185,7 +189,8 @@ VMManager::VMAIter VMManager::Unmap(VMAIter vma_handle) {
     VirtualMemoryArea& vma = vma_handle->second;
     vma.type = VMAType::Free;
     vma.permissions = VMAPermission::None;
-    vma.meminfo_state = MemoryState::Unmapped;
+    vma.state = MemoryState::Unmapped;
+    vma.attribute = MemoryAttribute::None;
 
     vma.backing_block = nullptr;
     vma.offset = 0;
@@ -298,6 +303,54 @@ ResultCode VMManager::HeapFree(VAddr target, u64 size) {
     return RESULT_SUCCESS;
 }
 
+MemoryInfo VMManager::QueryMemory(VAddr address) const {
+    const auto vma = FindVMA(address);
+    MemoryInfo memory_info{};
+
+    if (IsValidHandle(vma)) {
+        memory_info.base_address = vma->second.base;
+        memory_info.attributes = ToSvcMemoryAttribute(vma->second.attribute);
+        memory_info.permission = static_cast<u32>(vma->second.permissions);
+        memory_info.size = vma->second.size;
+        memory_info.state = ToSvcMemoryState(vma->second.state);
+    } else {
+        memory_info.base_address = address_space_end;
+        memory_info.permission = static_cast<u32>(VMAPermission::None);
+        memory_info.size = 0 - address_space_end;
+        memory_info.state = static_cast<u32>(MemoryState::Inaccessible);
+    }
+
+    return memory_info;
+}
+
+ResultCode VMManager::SetMemoryAttribute(VAddr address, u64 size, MemoryAttribute mask,
+                                         MemoryAttribute attribute) {
+    constexpr auto ignore_mask = MemoryAttribute::Uncached | MemoryAttribute::DeviceMapped;
+    constexpr auto attribute_mask = ~ignore_mask;
+
+    const auto result = CheckRangeState(
+        address, size, MemoryState::FlagUncached, MemoryState::FlagUncached, VMAPermission::None,
+        VMAPermission::None, attribute_mask, MemoryAttribute::None, ignore_mask);
+
+    if (result.Failed()) {
+        return result.Code();
+    }
+
+    const auto [prev_state, prev_permissions, prev_attributes] = *result;
+    const auto new_attribute = (prev_attributes & ~mask) | (mask & attribute);
+
+    const auto carve_result = CarveVMARange(address, size);
+    if (carve_result.Failed()) {
+        return carve_result.Code();
+    }
+
+    auto vma_iter = *carve_result;
+    vma_iter->second.attribute = new_attribute;
+
+    MergeAdjacent(vma_iter);
+    return RESULT_SUCCESS;
+}
+
 ResultCode VMManager::MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size, MemoryState state) {
     const auto vma = FindVMA(src_addr);
 
@@ -341,7 +394,7 @@ void VMManager::LogLayout() const {
                   (u8)vma.permissions & (u8)VMAPermission::Read ? 'R' : '-',
                   (u8)vma.permissions & (u8)VMAPermission::Write ? 'W' : '-',
                   (u8)vma.permissions & (u8)VMAPermission::Execute ? 'X' : '-',
-                  GetMemoryStateName(vma.meminfo_state));
+                  GetMemoryStateName(vma.state));
     }
 }
 
@@ -568,6 +621,66 @@ void VMManager::ClearPageTable() {
               Memory::PageType::Unmapped);
 }
 
+VMManager::CheckResults VMManager::CheckRangeState(VAddr address, u64 size, MemoryState state_mask,
+                                                   MemoryState state, VMAPermission permission_mask,
+                                                   VMAPermission permissions,
+                                                   MemoryAttribute attribute_mask,
+                                                   MemoryAttribute attribute,
+                                                   MemoryAttribute ignore_mask) const {
+    auto iter = FindVMA(address);
+
+    // If we don't have a valid VMA handle at this point, then it means this is
+    // being called with an address outside of the address space, which is definitely
+    // indicative of a bug, as this function only operates on mapped memory regions.
+    DEBUG_ASSERT(IsValidHandle(iter));
+
+    const VAddr end_address = address + size - 1;
+    const MemoryAttribute initial_attributes = iter->second.attribute;
+    const VMAPermission initial_permissions = iter->second.permissions;
+    const MemoryState initial_state = iter->second.state;
+
+    while (true) {
+        // The iterator should be valid throughout the traversal. Hitting the end of
+        // the mapped VMA regions is unquestionably indicative of a bug.
+        DEBUG_ASSERT(IsValidHandle(iter));
+
+        const auto& vma = iter->second;
+
+        if (vma.state != initial_state) {
+            return ERR_INVALID_ADDRESS_STATE;
+        }
+
+        if ((vma.state & state_mask) != state) {
+            return ERR_INVALID_ADDRESS_STATE;
+        }
+
+        if (vma.permissions != initial_permissions) {
+            return ERR_INVALID_ADDRESS_STATE;
+        }
+
+        if ((vma.permissions & permission_mask) != permissions) {
+            return ERR_INVALID_ADDRESS_STATE;
+        }
+
+        if ((vma.attribute | ignore_mask) != (initial_attributes | ignore_mask)) {
+            return ERR_INVALID_ADDRESS_STATE;
+        }
+
+        if ((vma.attribute & attribute_mask) != attribute) {
+            return ERR_INVALID_ADDRESS_STATE;
+        }
+
+        if (end_address <= vma.EndAddress()) {
+            break;
+        }
+
+        ++iter;
+    }
+
+    return MakeResult(
+        std::make_tuple(initial_state, initial_permissions, initial_attributes & ~ignore_mask));
+}
+
 u64 VMManager::GetTotalMemoryUsage() const {
     LOG_WARNING(Kernel, "(STUBBED) called");
     return 0xF8000000;

src/core/hle/kernel/vm_manager.h

@@ -6,6 +6,7 @@
 
 #include <map>
 #include <memory>
+#include <tuple>
 #include <vector>
 #include "common/common_types.h"
 #include "core/hle/result.h"
@@ -43,26 +44,211 @@ enum class VMAPermission : u8 {
     ReadWriteExecute = Read | Write | Execute,
 };
 
-/// Set of values returned in MemoryInfo.state by svcQueryMemory.
+constexpr VMAPermission operator|(VMAPermission lhs, VMAPermission rhs) {
+    return static_cast<VMAPermission>(u32(lhs) | u32(rhs));
+}
+
+constexpr VMAPermission operator&(VMAPermission lhs, VMAPermission rhs) {
+    return static_cast<VMAPermission>(u32(lhs) & u32(rhs));
+}
+
+constexpr VMAPermission operator^(VMAPermission lhs, VMAPermission rhs) {
+    return static_cast<VMAPermission>(u32(lhs) ^ u32(rhs));
+}
+
+constexpr VMAPermission operator~(VMAPermission permission) {
+    return static_cast<VMAPermission>(~u32(permission));
+}
+
+constexpr VMAPermission& operator|=(VMAPermission& lhs, VMAPermission rhs) {
+    lhs = lhs | rhs;
+    return lhs;
+}
+
+constexpr VMAPermission& operator&=(VMAPermission& lhs, VMAPermission rhs) {
+    lhs = lhs & rhs;
+    return lhs;
+}
+
+constexpr VMAPermission& operator^=(VMAPermission& lhs, VMAPermission rhs) {
+    lhs = lhs ^ rhs;
+    return lhs;
+}
+
+/// Attribute flags that can be applied to a VMA
+enum class MemoryAttribute : u32 {
+    Mask = 0xFF,
+
+    /// No particular qualities
+    None = 0,
+    /// Memory locked/borrowed for use. e.g. This would be used by transfer memory.
+    Locked = 1,
+    /// Memory locked for use by IPC-related internals.
+    LockedForIPC = 2,
+    /// Mapped as part of the device address space.
+    DeviceMapped = 4,
+    /// Uncached memory
+    Uncached = 8,
+};
+
+constexpr MemoryAttribute operator|(MemoryAttribute lhs, MemoryAttribute rhs) {
+    return static_cast<MemoryAttribute>(u32(lhs) | u32(rhs));
+}
+
+constexpr MemoryAttribute operator&(MemoryAttribute lhs, MemoryAttribute rhs) {
+    return static_cast<MemoryAttribute>(u32(lhs) & u32(rhs));
+}
+
+constexpr MemoryAttribute operator^(MemoryAttribute lhs, MemoryAttribute rhs) {
+    return static_cast<MemoryAttribute>(u32(lhs) ^ u32(rhs));
+}
+
+constexpr MemoryAttribute operator~(MemoryAttribute attribute) {
+    return static_cast<MemoryAttribute>(~u32(attribute));
+}
+
+constexpr MemoryAttribute& operator|=(MemoryAttribute& lhs, MemoryAttribute rhs) {
+    lhs = lhs | rhs;
+    return lhs;
+}
+
+constexpr MemoryAttribute& operator&=(MemoryAttribute& lhs, MemoryAttribute rhs) {
+    lhs = lhs & rhs;
+    return lhs;
+}
+
+constexpr MemoryAttribute& operator^=(MemoryAttribute& lhs, MemoryAttribute rhs) {
+    lhs = lhs ^ rhs;
+    return lhs;
+}
+
+constexpr u32 ToSvcMemoryAttribute(MemoryAttribute attribute) {
+    return static_cast<u32>(attribute & MemoryAttribute::Mask);
+}
+
+// clang-format off
+/// Represents memory states and any relevant flags, as used by the kernel.
+/// svcQueryMemory interprets these by masking away all but the first eight
+/// bits when storing memory state into a MemoryInfo instance.
 enum class MemoryState : u32 {
-    Unmapped = 0x0,
-    Io = 0x1,
-    Normal = 0x2,
-    CodeStatic = 0x3,
-    CodeMutable = 0x4,
-    Heap = 0x5,
-    Shared = 0x6,
-    ModuleCodeStatic = 0x8,
-    ModuleCodeMutable = 0x9,
-    IpcBuffer0 = 0xA,
-    Mapped = 0xB,
-    ThreadLocal = 0xC,
-    TransferMemoryIsolated = 0xD,
-    TransferMemory = 0xE,
-    ProcessMemory = 0xF,
-    IpcBuffer1 = 0x11,
-    IpcBuffer3 = 0x12,
-    KernelStack = 0x13,
+    Mask                            = 0xFF,
+    FlagProtect                     = 1U << 8,
+    FlagDebug                       = 1U << 9,
+    FlagIPC0                        = 1U << 10,
+    FlagIPC3                        = 1U << 11,
+    FlagIPC1                        = 1U << 12,
+    FlagMapped                      = 1U << 13,
+    FlagCode                        = 1U << 14,
+    FlagAlias                       = 1U << 15,
+    FlagModule                      = 1U << 16,
+    FlagTransfer                    = 1U << 17,
+    FlagQueryPhysicalAddressAllowed = 1U << 18,
+    FlagSharedDevice                = 1U << 19,
+    FlagSharedDeviceAligned         = 1U << 20,
+    FlagIPCBuffer                   = 1U << 21,
+    FlagMemoryPoolAllocated         = 1U << 22,
+    FlagMapProcess                  = 1U << 23,
+    FlagUncached                    = 1U << 24,
+    FlagCodeMemory                  = 1U << 25,
+
+    // Convenience flag sets to reduce repetition
+    IPCFlags = FlagIPC0 | FlagIPC3 | FlagIPC1,
+
+    CodeFlags = FlagDebug | IPCFlags | FlagMapped | FlagCode | FlagQueryPhysicalAddressAllowed |
+                FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated,
+
+    DataFlags = FlagProtect | IPCFlags | FlagMapped | FlagAlias | FlagTransfer |
+                FlagQueryPhysicalAddressAllowed | FlagSharedDevice | FlagSharedDeviceAligned |
+                FlagMemoryPoolAllocated | FlagIPCBuffer | FlagUncached,
+
+    Unmapped               = 0x00,
+    Io                     = 0x01 | FlagMapped,
+    Normal                 = 0x02 | FlagMapped | FlagQueryPhysicalAddressAllowed,
+    CodeStatic             = 0x03 | CodeFlags  | FlagMapProcess,
+    CodeMutable            = 0x04 | CodeFlags  | FlagMapProcess | FlagCodeMemory,
+    Heap                   = 0x05 | DataFlags  | FlagCodeMemory,
+    Shared                 = 0x06 | FlagMapped | FlagMemoryPoolAllocated,
+    ModuleCodeStatic       = 0x08 | CodeFlags  | FlagModule | FlagMapProcess,
+    ModuleCodeMutable      = 0x09 | DataFlags  | FlagModule | FlagMapProcess | FlagCodeMemory,
+
+    IpcBuffer0             = 0x0A | FlagMapped | FlagQueryPhysicalAddressAllowed | FlagMemoryPoolAllocated |
+                                    IPCFlags | FlagSharedDevice | FlagSharedDeviceAligned,
+
+    Stack                  = 0x0B | FlagMapped | IPCFlags | FlagQueryPhysicalAddressAllowed |
+                                    FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated,
+
+    ThreadLocal            = 0x0C | FlagMapped | FlagMemoryPoolAllocated,
+
+    TransferMemoryIsolated = 0x0D | IPCFlags | FlagMapped | FlagQueryPhysicalAddressAllowed |
+                                    FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated |
+                                    FlagUncached,
+
+    TransferMemory         = 0x0E | FlagIPC3   | FlagIPC1   | FlagMapped | FlagQueryPhysicalAddressAllowed |
+                                    FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated,
+
+    ProcessMemory          = 0x0F | FlagIPC3   | FlagIPC1   | FlagMapped | FlagMemoryPoolAllocated,
+
+    // Used to signify an inaccessible or invalid memory region with memory queries
+    Inaccessible           = 0x10,
+
+    IpcBuffer1             = 0x11 | FlagIPC3   | FlagIPC1   | FlagMapped | FlagQueryPhysicalAddressAllowed |
+                                    FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated,
+
+    IpcBuffer3             = 0x12 | FlagIPC3   | FlagMapped | FlagQueryPhysicalAddressAllowed |
+                                    FlagSharedDeviceAligned | FlagMemoryPoolAllocated,
+
+    KernelStack            = 0x13 | FlagMapped,
+};
+// clang-format on
+
+constexpr MemoryState operator|(MemoryState lhs, MemoryState rhs) {
+    return static_cast<MemoryState>(u32(lhs) | u32(rhs));
+}
+
+constexpr MemoryState operator&(MemoryState lhs, MemoryState rhs) {
+    return static_cast<MemoryState>(u32(lhs) & u32(rhs));
+}
+
+constexpr MemoryState operator^(MemoryState lhs, MemoryState rhs) {
+    return static_cast<MemoryState>(u32(lhs) ^ u32(rhs));
+}
+
+constexpr MemoryState operator~(MemoryState lhs) {
+    return static_cast<MemoryState>(~u32(lhs));
+}
+
+constexpr MemoryState& operator|=(MemoryState& lhs, MemoryState rhs) {
+    lhs = lhs | rhs;
+    return lhs;
+}
+
+constexpr MemoryState& operator&=(MemoryState& lhs, MemoryState rhs) {
+    lhs = lhs & rhs;
+    return lhs;
+}
+
+constexpr MemoryState& operator^=(MemoryState& lhs, MemoryState rhs) {
+    lhs = lhs ^ rhs;
+    return lhs;
+}
+
+constexpr u32 ToSvcMemoryState(MemoryState state) {
+    return static_cast<u32>(state & MemoryState::Mask);
+}
+
+struct MemoryInfo {
+    u64 base_address;
+    u64 size;
+    u32 state;
+    u32 attributes;
+    u32 permission;
+    u32 ipc_ref_count;
+    u32 device_ref_count;
+};
+static_assert(sizeof(MemoryInfo) == 0x28, "MemoryInfo has incorrect size.");
+
+struct PageInfo {
+    u32 flags;
 };
 
 /**
@@ -71,6 +257,16 @@ enum class MemoryState : u32 {
  * also backed by a single host memory allocation.
  */
 struct VirtualMemoryArea {
+    /// Gets the starting (base) address of this VMA.
+    VAddr StartAddress() const {
+        return base;
+    }
+
+    /// Gets the ending address of this VMA.
+    VAddr EndAddress() const {
+        return base + size - 1;
+    }
+
     /// Virtual base address of the region.
     VAddr base = 0;
     /// Size of the region.
@@ -78,8 +274,8 @@ struct VirtualMemoryArea {
 
     VMAType type = VMAType::Free;
     VMAPermission permissions = VMAPermission::None;
-    /// Tag returned by svcQueryMemory. Not otherwise used.
-    MemoryState meminfo_state = MemoryState::Unmapped;
+    MemoryState state = MemoryState::Unmapped;
+    MemoryAttribute attribute = MemoryAttribute::None;
 
     // Settings for type = AllocatedMemoryBlock
     /// Memory block backing this VMA.
@@ -113,16 +309,10 @@ struct VirtualMemoryArea {
  *  - http://duartes.org/gustavo/blog/post/page-cache-the-affair-between-memory-and-files/
  */
 class VMManager final {
+    using VMAMap = std::map<VAddr, VirtualMemoryArea>;
+
 public:
-    /**
-     * A map covering the entirety of the managed address space, keyed by the `base` field of each
-     * VMA. It must always be modified by splitting or merging VMAs, so that the invariant
-     * `elem.base + elem.size == next.base` is preserved, and mergeable regions must always be
-     * merged when possible so that no two similar and adjacent regions exist that have not been
-     * merged.
-     */
-    std::map<VAddr, VirtualMemoryArea> vma_map;
-    using VMAHandle = decltype(vma_map)::const_iterator;
+    using VMAHandle = VMAMap::const_iterator;
 
     VMManager();
     ~VMManager();
@@ -133,6 +323,9 @@ public:
     /// Finds the VMA in which the given address is included in, or `vma_map.end()`.
     VMAHandle FindVMA(VAddr target) const;
 
+    /// Indicates whether or not the given handle is within the VMA map.
+    bool IsValidHandle(VMAHandle handle) const;
+
     // TODO(yuriks): Should these functions actually return the handle?
 
     /**
@@ -189,8 +382,28 @@ public:
     ResultVal<VAddr> HeapAllocate(VAddr target, u64 size, VMAPermission perms);
     ResultCode HeapFree(VAddr target, u64 size);
 
-    ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size,
-                            MemoryState state = MemoryState::Mapped);
+    ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size, MemoryState state);
+
+    /// Queries the memory manager for information about the given address.
+    ///
+    /// @param address The address to query the memory manager about for information.
+    ///
+    /// @return A MemoryInfo instance containing information about the given address.
+    ///
+    MemoryInfo QueryMemory(VAddr address) const;
+
+    /// Sets an attribute across the given address range.
+    ///
+    /// @param address   The starting address
+    /// @param size      The size of the range to set the attribute on.
+    /// @param mask      The attribute mask
+    /// @param attribute The attribute to set across the given address range
+    ///
+    /// @returns RESULT_SUCCESS if successful
+    /// @returns ERR_INVALID_ADDRESS_STATE if the attribute could not be set.
+    ///
+    ResultCode SetMemoryAttribute(VAddr address, u64 size, MemoryAttribute mask,
+                                  MemoryAttribute attribute);
 
     /**
      * Scans all VMAs and updates the page table range of any that use the given vector as backing
@@ -281,7 +494,7 @@ public:
     Memory::PageTable page_table;
 
 private:
-    using VMAIter = decltype(vma_map)::iterator;
+    using VMAIter = VMAMap::iterator;
 
     /// Converts a VMAHandle to a mutable VMAIter.
     VMAIter StripIterConstness(const VMAHandle& iter);
@@ -328,6 +541,44 @@ private:
     /// Clears out the page table
     void ClearPageTable();
 
+    using CheckResults = ResultVal<std::tuple<MemoryState, VMAPermission, MemoryAttribute>>;
+
+    /// Checks if an address range adheres to the specified states provided.
+    ///
+    /// @param address         The starting address of the address range.
+    /// @param size            The size of the address range.
+    /// @param state_mask      The memory state mask.
+    /// @param state           The state to compare the individual VMA states against,
+    ///                        which is done in the form of: (vma.state & state_mask) != state.
+    /// @param permission_mask The memory permissions mask.
+    /// @param permissions     The permission to compare the individual VMA permissions against,
+    ///                        which is done in the form of:
+    ///                        (vma.permission & permission_mask) != permission.
+    /// @param attribute_mask  The memory attribute mask.
+    /// @param attribute       The memory attributes to compare the individual VMA attributes
+    ///                        against, which is done in the form of:
+    ///                        (vma.attributes & attribute_mask) != attribute.
+    /// @param ignore_mask     The memory attributes to ignore during the check.
+    ///
+    /// @returns If successful, returns a tuple containing the memory attributes
+    ///          (with ignored bits specified by ignore_mask unset), memory permissions, and
+    ///          memory state across the memory range.
+    /// @returns If not successful, returns ERR_INVALID_ADDRESS_STATE.
+    ///
+    CheckResults CheckRangeState(VAddr address, u64 size, MemoryState state_mask, MemoryState state,
+                                 VMAPermission permission_mask, VMAPermission permissions,
+                                 MemoryAttribute attribute_mask, MemoryAttribute attribute,
+                                 MemoryAttribute ignore_mask) const;
+
+    /**
+     * A map covering the entirety of the managed address space, keyed by the `base` field of each
+     * VMA. It must always be modified by splitting or merging VMAs, so that the invariant
+     * `elem.base + elem.size == next.base` is preserved, and mergeable regions must always be
+     * merged when possible so that no two similar and adjacent regions exist that have not been
+     * merged.
+     */
+    VMAMap vma_map;
+
     u32 address_space_width = 0;
     VAddr address_space_base = 0;
     VAddr address_space_end = 0;

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

@@ -19,6 +19,7 @@
 #include "core/hle/service/am/applet_ae.h"
 #include "core/hle/service/am/applet_oe.h"
 #include "core/hle/service/am/applets/applets.h"
+#include "core/hle/service/am/applets/profile_select.h"
 #include "core/hle/service/am/applets/software_keyboard.h"
 #include "core/hle/service/am/applets/stub_applet.h"
 #include "core/hle/service/am/idle.h"
@@ -39,6 +40,7 @@ constexpr ResultCode ERR_NO_DATA_IN_CHANNEL{ErrorModule::AM, 0x2};
 constexpr ResultCode ERR_SIZE_OUT_OF_BOUNDS{ErrorModule::AM, 0x1F7};
 
 enum class AppletId : u32 {
+    ProfileSelect = 0x10,
     SoftwareKeyboard = 0x11,
 };
 
@@ -71,10 +73,13 @@ IWindowController::IWindowController() : ServiceFramework("IWindowController") {
 IWindowController::~IWindowController() = default;
 
 void IWindowController::GetAppletResourceUserId(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    const u64 process_id = Core::System::GetInstance().Kernel().CurrentProcess()->GetProcessID();
+
+    LOG_DEBUG(Service_AM, "called. Process ID=0x{:016X}", process_id);
+
     IPC::ResponseBuilder rb{ctx, 4};
     rb.Push(RESULT_SUCCESS);
-    rb.Push<u64>(0);
+    rb.Push<u64>(process_id);
 }
 
 void IWindowController::AcquireForegroundRights(Kernel::HLERequestContext& ctx) {
@@ -565,7 +570,6 @@ private:
     void GetAppletStateChangedEvent(Kernel::HLERequestContext& ctx) {
         LOG_DEBUG(Service_AM, "called");
 
-        applet->GetBroker().SignalStateChanged();
         const auto event = applet->GetBroker().GetStateChangedEvent();
 
         IPC::ResponseBuilder rb{ctx, 2, 1};
@@ -773,6 +777,8 @@ ILibraryAppletCreator::~ILibraryAppletCreator() = default;
 
 static std::shared_ptr<Applets::Applet> GetAppletFromId(AppletId id) {
     switch (id) {
+    case AppletId::ProfileSelect:
+        return std::make_shared<Applets::ProfileSelect>();
     case AppletId::SoftwareKeyboard:
         return std::make_shared<Applets::SoftwareKeyboard>();
     default:

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

@@ -16,11 +16,11 @@ namespace Service::AM::Applets {
 AppletDataBroker::AppletDataBroker() {
     auto& kernel = Core::System::GetInstance().Kernel();
     state_changed_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::OneShot, "ILibraryAppletAccessor:StateChangedEvent");
+        kernel, Kernel::ResetType::Sticky, "ILibraryAppletAccessor:StateChangedEvent");
     pop_out_data_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::OneShot, "ILibraryAppletAccessor:PopDataOutEvent");
+        kernel, Kernel::ResetType::Sticky, "ILibraryAppletAccessor:PopDataOutEvent");
     pop_interactive_out_data_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::OneShot, "ILibraryAppletAccessor:PopInteractiveDataOutEvent");
+        kernel, Kernel::ResetType::Sticky, "ILibraryAppletAccessor:PopInteractiveDataOutEvent");
 }
 
 AppletDataBroker::~AppletDataBroker() = default;

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

@@ -0,0 +1,77 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+
+#include "common/assert.h"
+#include "common/string_util.h"
+#include "core/core.h"
+#include "core/frontend/applets/software_keyboard.h"
+#include "core/hle/service/am/am.h"
+#include "core/hle/service/am/applets/profile_select.h"
+
+namespace Service::AM::Applets {
+
+constexpr ResultCode ERR_USER_CANCELLED_SELECTION{ErrorModule::Account, 1};
+
+ProfileSelect::ProfileSelect() = default;
+ProfileSelect::~ProfileSelect() = default;
+
+void ProfileSelect::Initialize() {
+    complete = false;
+    status = RESULT_SUCCESS;
+    final_data.clear();
+
+    Applet::Initialize();
+
+    const auto user_config_storage = broker.PopNormalDataToApplet();
+    ASSERT(user_config_storage != nullptr);
+    const auto& user_config = user_config_storage->GetData();
+
+    ASSERT(user_config.size() >= sizeof(UserSelectionConfig));
+    std::memcpy(&config, user_config.data(), sizeof(UserSelectionConfig));
+}
+
+bool ProfileSelect::TransactionComplete() const {
+    return complete;
+}
+
+ResultCode ProfileSelect::GetStatus() const {
+    return status;
+}
+
+void ProfileSelect::ExecuteInteractive() {
+    UNREACHABLE_MSG("Attempted to call interactive execution on non-interactive applet.");
+}
+
+void ProfileSelect::Execute() {
+    if (complete) {
+        broker.PushNormalDataFromApplet(IStorage{final_data});
+        return;
+    }
+
+    const auto& frontend{Core::System::GetInstance().GetProfileSelector()};
+
+    frontend.SelectProfile([this](std::optional<Account::UUID> uuid) { SelectionComplete(uuid); });
+}
+
+void ProfileSelect::SelectionComplete(std::optional<Account::UUID> uuid) {
+    UserSelectionOutput output{};
+
+    if (uuid.has_value() && uuid->uuid != Account::INVALID_UUID) {
+        output.result = 0;
+        output.uuid_selected = uuid->uuid;
+    } else {
+        status = ERR_USER_CANCELLED_SELECTION;
+        output.result = ERR_USER_CANCELLED_SELECTION.raw;
+        output.uuid_selected = Account::INVALID_UUID;
+    }
+
+    final_data = std::vector<u8>(sizeof(UserSelectionOutput));
+    std::memcpy(final_data.data(), &output, final_data.size());
+    broker.PushNormalDataFromApplet(IStorage{final_data});
+    broker.SignalStateChanged();
+}
+
+} // namespace Service::AM::Applets

src/core/hle/service/am/applets/profile_select.h

@@ -0,0 +1,50 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <vector>
+
+#include "common/common_funcs.h"
+#include "core/hle/service/acc/profile_manager.h"
+#include "core/hle/service/am/applets/applets.h"
+
+namespace Service::AM::Applets {
+
+struct UserSelectionConfig {
+    // TODO(DarkLordZach): RE this structure
+    // It seems to be flags and the like that determine the UI of the applet on the switch... from
+    // my research this is safe to ignore for now.
+    INSERT_PADDING_BYTES(0xA0);
+};
+static_assert(sizeof(UserSelectionConfig) == 0xA0, "UserSelectionConfig has incorrect size.");
+
+struct UserSelectionOutput {
+    u64 result;
+    u128 uuid_selected;
+};
+static_assert(sizeof(UserSelectionOutput) == 0x18, "UserSelectionOutput has incorrect size.");
+
+class ProfileSelect final : public Applet {
+public:
+    ProfileSelect();
+    ~ProfileSelect() override;
+
+    void Initialize() override;
+
+    bool TransactionComplete() const override;
+    ResultCode GetStatus() const override;
+    void ExecuteInteractive() override;
+    void Execute() override;
+
+    void SelectionComplete(std::optional<Account::UUID> uuid);
+
+private:
+    UserSelectionConfig config;
+    bool complete = false;
+    ResultCode status = RESULT_SUCCESS;
+    std::vector<u8> final_data;
+};
+
+} // namespace Service::AM::Applets

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

@@ -146,11 +146,10 @@ void SoftwareKeyboard::WriteText(std::optional<std::u16string> text) {
 
         if (complete) {
             broker.PushNormalDataFromApplet(IStorage{output_main});
+            broker.SignalStateChanged();
         } else {
             broker.PushInteractiveDataFromApplet(IStorage{output_sub});
         }
-
-        broker.SignalStateChanged();
     } else {
         output_main[0] = 1;
         complete = true;

src/core/hle/service/aoc/aoc_u.cpp

@@ -20,6 +20,7 @@
 #include "core/hle/service/aoc/aoc_u.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/loader.h"
+#include "core/settings.h"
 
 namespace Service::AOC {
 
@@ -76,6 +77,13 @@ void AOC_U::CountAddOnContent(Kernel::HLERequestContext& ctx) {
     rb.Push(RESULT_SUCCESS);
 
     const auto current = Core::System::GetInstance().CurrentProcess()->GetTitleID();
+
+    const auto& disabled = Settings::values.disabled_addons[current];
+    if (std::find(disabled.begin(), disabled.end(), "DLC") != disabled.end()) {
+        rb.Push<u32>(0);
+        return;
+    }
+
     rb.Push<u32>(static_cast<u32>(
         std::count_if(add_on_content.begin(), add_on_content.end(),
                       [current](u64 tid) { return CheckAOCTitleIDMatchesBase(tid, current); })));
@@ -96,6 +104,10 @@ void AOC_U::ListAddOnContent(Kernel::HLERequestContext& ctx) {
             out.push_back(static_cast<u32>(add_on_content[i] & 0x7FF));
     }
 
+    const auto& disabled = Settings::values.disabled_addons[current];
+    if (std::find(disabled.begin(), disabled.end(), "DLC") != disabled.end())
+        out = {};
+
     if (out.size() < offset) {
         IPC::ResponseBuilder rb{ctx, 2};
         // TODO(DarkLordZach): Find the correct error code.

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

@@ -45,8 +45,12 @@ public:
     explicit IStorage(FileSys::VirtualFile backend_)
         : ServiceFramework("IStorage"), backend(std::move(backend_)) {
         static const FunctionInfo functions[] = {
-            {0, &IStorage::Read, "Read"}, {1, nullptr, "Write"},   {2, nullptr, "Flush"},
-            {3, nullptr, "SetSize"},      {4, nullptr, "GetSize"}, {5, nullptr, "OperateRange"},
+            {0, &IStorage::Read, "Read"},
+            {1, nullptr, "Write"},
+            {2, nullptr, "Flush"},
+            {3, nullptr, "SetSize"},
+            {4, &IStorage::GetSize, "GetSize"},
+            {5, nullptr, "OperateRange"},
         };
         RegisterHandlers(functions);
     }
@@ -83,6 +87,15 @@ private:
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
     }
+
+    void GetSize(Kernel::HLERequestContext& ctx) {
+        const u64 size = backend->GetSize();
+        LOG_DEBUG(Service_FS, "called, size={}", size);
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u64>(size);
+    }
 };
 
 class IFile final : public ServiceFramework<IFile> {
@@ -796,9 +809,18 @@ void FSP_SRV::OpenSaveDataInfoReaderBySaveDataSpaceId(Kernel::HLERequestContext&
 void FSP_SRV::GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_FS, "(STUBBED) called");
 
+    enum class LogMode : u32 {
+        Off,
+        Log,
+        RedirectToSdCard,
+        LogToSdCard = Log | RedirectToSdCard,
+    };
+
+    // Given we always want to receive logging information,
+    // we always specify logging as enabled.
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
-    rb.Push<u32>(5);
+    rb.PushEnum(LogMode::Log);
 }
 
 void FSP_SRV::OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx) {

src/core/hle/service/ldr/ldr.cpp

@@ -408,13 +408,13 @@ private:
     using SHA256Hash = std::array<u8, 0x20>;
 
     struct NROHeader {
-        u32_le entrypoint_insn;
+        INSERT_PADDING_WORDS(1);
         u32_le mod_offset;
         INSERT_PADDING_WORDS(2);
         u32_le magic;
-        INSERT_PADDING_WORDS(1);
+        u32_le version;
         u32_le nro_size;
-        INSERT_PADDING_WORDS(1);
+        u32_le flags;
         u32_le text_offset;
         u32_le text_size;
         u32_le ro_offset;
@@ -430,9 +430,10 @@ private:
 
     struct NRRHeader {
         u32_le magic;
-        INSERT_PADDING_BYTES(0x1C);
+        INSERT_PADDING_BYTES(12);
         u64_le title_id_mask;
         u64_le title_id_pattern;
+        INSERT_PADDING_BYTES(16);
         std::array<u8, 0x100> modulus;
         std::array<u8, 0x100> signature_1;
         std::array<u8, 0x100> signature_2;

src/core/hle/service/nfp/nfp.cpp

@@ -317,8 +317,8 @@ private:
     }
 
     bool has_attached_handle{};
-    const u64 device_handle{Common::MakeMagic('Y', 'U', 'Z', 'U')};
-    const u32 npad_id{0}; // Player 1 controller
+    const u64 device_handle{0}; // Npad device 1
+    const u32 npad_id{0};       // Player 1 controller
     State state{State::NonInitialized};
     DeviceState device_state{DeviceState::Initialized};
     Kernel::EventPair deactivate_event;

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

@@ -137,6 +137,10 @@ u32 nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector<
 }
 
 static void PushGPUEntries(Tegra::CommandList&& entries) {
+    if (entries.empty()) {
+        return;
+    }
+
     auto& dma_pusher{Core::System::GetInstance().GPU().DmaPusher()};
     dma_pusher.Push(std::move(entries));
     dma_pusher.DispatchCalls();

src/core/hle/service/service.cpp

@@ -70,10 +70,6 @@
 #include "core/hle/service/vi/vi.h"
 #include "core/hle/service/wlan/wlan.h"
 
-using Kernel::ClientPort;
-using Kernel::ServerPort;
-using Kernel::SharedPtr;
-
 namespace Service {
 
 /**
@@ -101,33 +97,33 @@ ServiceFrameworkBase::ServiceFrameworkBase(const char* service_name, u32 max_ses
 ServiceFrameworkBase::~ServiceFrameworkBase() = default;
 
 void ServiceFrameworkBase::InstallAsService(SM::ServiceManager& service_manager) {
-    ASSERT(port == nullptr);
-    port = service_manager.RegisterService(service_name, max_sessions).Unwrap();
+    ASSERT(!port_installed);
+
+    auto port = service_manager.RegisterService(service_name, max_sessions).Unwrap();
     port->SetHleHandler(shared_from_this());
+    port_installed = true;
 }
 
 void ServiceFrameworkBase::InstallAsNamedPort() {
-    ASSERT(port == nullptr);
+    ASSERT(!port_installed);
 
     auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<ServerPort> server_port;
-    SharedPtr<ClientPort> client_port;
-    std::tie(server_port, client_port) =
-        ServerPort::CreatePortPair(kernel, max_sessions, service_name);
+    auto [server_port, client_port] =
+        Kernel::ServerPort::CreatePortPair(kernel, max_sessions, service_name);
     server_port->SetHleHandler(shared_from_this());
     kernel.AddNamedPort(service_name, std::move(client_port));
+    port_installed = true;
 }
 
 Kernel::SharedPtr<Kernel::ClientPort> ServiceFrameworkBase::CreatePort() {
-    ASSERT(port == nullptr);
+    ASSERT(!port_installed);
 
     auto& kernel = Core::System::GetInstance().Kernel();
-    Kernel::SharedPtr<Kernel::ServerPort> server_port;
-    Kernel::SharedPtr<Kernel::ClientPort> client_port;
-    std::tie(server_port, client_port) =
+    auto [server_port, client_port] =
         Kernel::ServerPort::CreatePortPair(kernel, max_sessions, service_name);
-    port = MakeResult<Kernel::SharedPtr<Kernel::ServerPort>>(std::move(server_port)).Unwrap();
+    auto port = MakeResult(std::move(server_port)).Unwrap();
     port->SetHleHandler(shared_from_this());
+    port_installed = true;
     return client_port;
 }
 
@@ -152,8 +148,7 @@ void ServiceFrameworkBase::ReportUnimplementedFunction(Kernel::HLERequestContext
     }
     buf.push_back('}');
 
-    LOG_ERROR(Service, "unknown / unimplemented {}", fmt::to_string(buf));
-    UNIMPLEMENTED();
+    UNIMPLEMENTED_MSG("Unknown / unimplemented {}", fmt::to_string(buf));
 }
 
 void ServiceFrameworkBase::InvokeRequest(Kernel::HLERequestContext& ctx) {

src/core/hle/service/service.h

@@ -96,11 +96,9 @@ private:
     /// Maximum number of concurrent sessions that this service can handle.
     u32 max_sessions;
 
-    /**
-     * Port where incoming connections will be received. Only created when InstallAsService() or
-     * InstallAsNamedPort() are called.
-     */
-    Kernel::SharedPtr<Kernel::ServerPort> port;
+    /// Flag to store if a port was already create/installed to detect multiple install attempts,
+    /// which is not supported.
+    bool port_installed = false;
 
     /// Function used to safely up-cast pointers to the derived class before invoking a handler.
     InvokerFn* handler_invoker;

src/core/hle/service/sm/sm.cpp

@@ -54,13 +54,11 @@ ResultVal<Kernel::SharedPtr<Kernel::ServerPort>> ServiceManager::RegisterService
         return ERR_ALREADY_REGISTERED;
 
     auto& kernel = Core::System::GetInstance().Kernel();
-    Kernel::SharedPtr<Kernel::ServerPort> server_port;
-    Kernel::SharedPtr<Kernel::ClientPort> client_port;
-    std::tie(server_port, client_port) =
+    auto [server_port, client_port] =
         Kernel::ServerPort::CreatePortPair(kernel, max_sessions, name);
 
     registered_services.emplace(std::move(name), std::move(client_port));
-    return MakeResult<Kernel::SharedPtr<Kernel::ServerPort>>(std::move(server_port));
+    return MakeResult(std::move(server_port));
 }
 
 ResultCode ServiceManager::UnregisterService(const std::string& name) {
@@ -83,7 +81,7 @@ ResultVal<Kernel::SharedPtr<Kernel::ClientPort>> ServiceManager::GetServicePort(
         return ERR_SERVICE_NOT_REGISTERED;
     }
 
-    return MakeResult<Kernel::SharedPtr<Kernel::ClientPort>>(it->second);
+    return MakeResult(it->second);
 }
 
 ResultVal<Kernel::SharedPtr<Kernel::ClientSession>> ServiceManager::ConnectToService(
@@ -147,12 +145,13 @@ void SM::RegisterService(Kernel::HLERequestContext& ctx) {
 
     const std::string name(name_buf.begin(), end);
 
-    const auto unk_bool = static_cast<bool>(rp.PopRaw<u32>());
-    const auto session_count = rp.PopRaw<u32>();
+    const auto is_light = static_cast<bool>(rp.PopRaw<u32>());
+    const auto max_session_count = rp.PopRaw<u32>();
 
-    LOG_DEBUG(Service_SM, "called with unk_bool={}", unk_bool);
+    LOG_DEBUG(Service_SM, "called with name={}, max_session_count={}, is_light={}", name,
+              max_session_count, is_light);
 
-    auto handle = service_manager->RegisterService(name, session_count);
+    auto handle = service_manager->RegisterService(name, max_session_count);
     if (handle.Failed()) {
         LOG_ERROR(Service_SM, "failed to register service with error_code={:08X}",
                   handle.Code().raw);

src/core/loader/deconstructed_rom_directory.h

@@ -33,7 +33,7 @@ public:
      */
     static FileType IdentifyType(const FileSys::VirtualFile& file);
 
-    FileType GetFileType() override {
+    FileType GetFileType() const override {
         return IdentifyType(file);
     }
 

src/core/loader/elf.h

@@ -22,7 +22,7 @@ public:
      */
     static FileType IdentifyType(const FileSys::VirtualFile& file);
 
-    FileType GetFileType() override {
+    FileType GetFileType() const override {
         return IdentifyType(file);
     }
 

src/core/loader/loader.h

@@ -12,6 +12,7 @@
 #include <vector>
 
 #include "common/common_types.h"
+#include "core/file_sys/control_metadata.h"
 #include "core/file_sys/vfs.h"
 
 namespace Kernel {
@@ -131,7 +132,7 @@ public:
      * Returns the type of this file
      * @return FileType corresponding to the loaded file
      */
-    virtual FileType GetFileType() = 0;
+    virtual FileType GetFileType() const = 0;
 
     /**
      * Load the application and return the created Process instance
@@ -243,6 +244,15 @@ public:
         return ResultStatus::ErrorNotImplemented;
     }
 
+    /**
+     * Get the developer of the application
+     * @param developer Reference to store the application developer into
+     * @return ResultStatus result of function
+     */
+    virtual ResultStatus ReadDeveloper(std::string& developer) {
+        return ResultStatus::ErrorNotImplemented;
+    }
+
 protected:
     FileSys::VirtualFile file;
     bool is_loaded = false;

src/core/loader/nax.cpp

@@ -37,7 +37,7 @@ FileType AppLoader_NAX::IdentifyType(const FileSys::VirtualFile& file) {
     return IdentifyTypeImpl(nax);
 }
 
-FileType AppLoader_NAX::GetFileType() {
+FileType AppLoader_NAX::GetFileType() const {
     return IdentifyTypeImpl(*nax);
 }
 

src/core/loader/nax.h

@@ -31,7 +31,7 @@ public:
      */
     static FileType IdentifyType(const FileSys::VirtualFile& file);
 
-    FileType GetFileType() override;
+    FileType GetFileType() const override;
 
     ResultStatus Load(Kernel::Process& process) override;
 

src/core/loader/nca.h

@@ -29,7 +29,7 @@ public:
      */
     static FileType IdentifyType(const FileSys::VirtualFile& file);
 
-    FileType GetFileType() override {
+    FileType GetFileType() const override {
         return IdentifyType(file);
     }
 

src/core/loader/nro.h

@@ -33,7 +33,7 @@ public:
      */
     static FileType IdentifyType(const FileSys::VirtualFile& file);
 
-    FileType GetFileType() override {
+    FileType GetFileType() const override {
         return IdentifyType(file);
     }
 

src/core/loader/nso.h

@@ -37,7 +37,7 @@ public:
      */
     static FileType IdentifyType(const FileSys::VirtualFile& file);
 
-    FileType GetFileType() override {
+    FileType GetFileType() const override {
         return IdentifyType(file);
     }
 

src/core/loader/nsp.cpp

@@ -151,4 +151,11 @@ ResultStatus AppLoader_NSP::ReadTitle(std::string& title) {
     title = nacp_file->GetApplicationName();
     return ResultStatus::Success;
 }
+
+ResultStatus AppLoader_NSP::ReadDeveloper(std::string& developer) {
+    if (nacp_file == nullptr)
+        return ResultStatus::ErrorNoControl;
+    developer = nacp_file->GetDeveloperName();
+    return ResultStatus::Success;
+}
 } // namespace Loader

src/core/loader/nsp.h

@@ -31,7 +31,7 @@ public:
      */
     static FileType IdentifyType(const FileSys::VirtualFile& file);
 
-    FileType GetFileType() override {
+    FileType GetFileType() const override {
         return IdentifyType(file);
     }
 
@@ -43,6 +43,7 @@ public:
     ResultStatus ReadProgramId(u64& out_program_id) override;
     ResultStatus ReadIcon(std::vector<u8>& buffer) override;
     ResultStatus ReadTitle(std::string& title) override;
+    ResultStatus ReadDeveloper(std::string& developer) override;
 
 private:
     std::unique_ptr<FileSys::NSP> nsp;

src/core/loader/xci.cpp

@@ -120,4 +120,11 @@ ResultStatus AppLoader_XCI::ReadTitle(std::string& title) {
     title = nacp_file->GetApplicationName();
     return ResultStatus::Success;
 }
+
+ResultStatus AppLoader_XCI::ReadDeveloper(std::string& developer) {
+    if (nacp_file == nullptr)
+        return ResultStatus::ErrorNoControl;
+    developer = nacp_file->GetDeveloperName();
+    return ResultStatus::Success;
+}
 } // namespace Loader

src/core/loader/xci.h

@@ -31,7 +31,7 @@ public:
      */
     static FileType IdentifyType(const FileSys::VirtualFile& file);
 
-    FileType GetFileType() override {
+    FileType GetFileType() const override {
         return IdentifyType(file);
     }
 
@@ -43,6 +43,7 @@ public:
     ResultStatus ReadProgramId(u64& out_program_id) override;
     ResultStatus ReadIcon(std::vector<u8>& buffer) override;
     ResultStatus ReadTitle(std::string& title) override;
+    ResultStatus ReadDeveloper(std::string& developer) override;
 
 private:
     std::unique_ptr<FileSys::XCI> xci;

src/core/memory.cpp

@@ -125,14 +125,13 @@ void RemoveDebugHook(PageTable& page_table, VAddr base, u64 size, MemoryHookPoin
  * using a VMA from the current process
  */
 static u8* GetPointerFromVMA(const Kernel::Process& process, VAddr vaddr) {
+    const auto& vm_manager = process.VMManager();
+
+    const auto it = vm_manager.FindVMA(vaddr);
+    DEBUG_ASSERT(vm_manager.IsValidHandle(it));
+
     u8* direct_pointer = nullptr;
-
-    auto& vm_manager = process.VMManager();
-
-    auto it = vm_manager.FindVMA(vaddr);
-    ASSERT(it != vm_manager.vma_map.end());
-
-    auto& vma = it->second;
+    const auto& vma = it->second;
     switch (vma.type) {
     case Kernel::VMAType::AllocatedMemoryBlock:
         direct_pointer = vma.backing_block->data() + vma.offset;
@@ -188,6 +187,7 @@ T Read(const VAddr vaddr) {
     default:
         UNREACHABLE();
     }
+    return {};
 }
 
 template <typename T>

src/core/settings.h

@@ -6,8 +6,10 @@
 
 #include <array>
 #include <atomic>
+#include <map>
 #include <optional>
 #include <string>
+#include <vector>
 #include "common/common_types.h"
 
 namespace Settings {
@@ -411,6 +413,9 @@ struct Values {
     std::string web_api_url;
     std::string yuzu_username;
     std::string yuzu_token;
+
+    // Add-Ons
+    std::map<u64, std::vector<std::string>> disabled_addons;
 } extern values;
 
 void Apply();

src/core/telemetry_session.cpp

@@ -103,13 +103,8 @@ bool VerifyLogin(const std::string& username, const std::string& token) {
 
 TelemetrySession::TelemetrySession() {
 #ifdef ENABLE_WEB_SERVICE
-    if (Settings::values.enable_telemetry) {
-        backend = std::make_unique<WebService::TelemetryJson>(Settings::values.web_api_url,
-                                                              Settings::values.yuzu_username,
-                                                              Settings::values.yuzu_token);
-    } else {
-        backend = std::make_unique<Telemetry::NullVisitor>();
-    }
+    backend = std::make_unique<WebService::TelemetryJson>(
+        Settings::values.web_api_url, Settings::values.yuzu_username, Settings::values.yuzu_token);
 #else
     backend = std::make_unique<Telemetry::NullVisitor>();
 #endif
@@ -180,7 +175,8 @@ TelemetrySession::~TelemetrySession() {
     // This is just a placeholder to wrap up the session once the core completes and this is
     // destroyed. This will be moved elsewhere once we are actually doing real I/O with the service.
     field_collection.Accept(*backend);
-    backend->Complete();
+    if (Settings::values.enable_telemetry)
+        backend->Complete();
     backend = nullptr;
 }
 

src/video_core/engines/maxwell_3d.h

@@ -164,6 +164,7 @@ public:
                     return 3;
                 default:
                     UNREACHABLE();
+                    return 1;
                 }
             }
 
@@ -871,6 +872,7 @@ public:
                             return 4;
                         }
                         UNREACHABLE();
+                        return 1;
                     }
 
                     GPUVAddr StartAddress() const {

src/video_core/engines/shader_bytecode.h

@@ -575,7 +575,7 @@ union Instruction {
 
     union {
         BitField<39, 2, u64> tab5cb8_2;
-        BitField<41, 3, u64> tab5c68_1;
+        BitField<41, 3, u64> postfactor;
         BitField<44, 2, u64> tab5c68_0;
         BitField<48, 1, u64> negate_b;
     } fmul;
@@ -609,7 +609,7 @@ union Instruction {
 
         BitField<31, 1, u64> negate_b;
         BitField<30, 1, u64> abs_b;
-        BitField<47, 2, HalfType> type_b;
+        BitField<28, 2, HalfType> type_b;
 
         BitField<35, 2, HalfType> type_c;
     } alu_half;
@@ -1049,6 +1049,7 @@ union Instruction {
         BitField<49, 1, u64> nodep_flag;
         BitField<50, 3, u64> component_mask_selector;
         BitField<53, 4, u64> texture_info;
+        BitField<59, 1, u64> fp32_flag;
 
         TextureType GetTextureType() const {
             // The TEXS instruction has a weird encoding for the texture type.
@@ -1064,6 +1065,7 @@ union Instruction {
             LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}",
                          static_cast<u32>(texture_info.Value()));
             UNREACHABLE();
+            return TextureType::Texture1D;
         }
 
         TextureProcessMode GetTextureProcessMode() const {
@@ -1144,6 +1146,7 @@ union Instruction {
             LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}",
                          static_cast<u32>(texture_info.Value()));
             UNREACHABLE();
+            return TextureType::Texture1D;
         }
 
         TextureProcessMode GetTextureProcessMode() const {
@@ -1549,7 +1552,7 @@ private:
             INST("1110111011011---", Id::STG, Type::Memory, "STG"),
             INST("110000----111---", Id::TEX, Type::Memory, "TEX"),
             INST("1101111101001---", Id::TXQ, Type::Memory, "TXQ"),
-            INST("1101100---------", Id::TEXS, Type::Memory, "TEXS"),
+            INST("1101-00---------", Id::TEXS, Type::Memory, "TEXS"),
             INST("1101101---------", Id::TLDS, Type::Memory, "TLDS"),
             INST("110010----111---", Id::TLD4, Type::Memory, "TLD4"),
             INST("1101111100------", Id::TLD4S, Type::Memory, "TLD4S"),

src/video_core/gpu.cpp

@@ -102,6 +102,7 @@ u32 RenderTargetBytesPerPixel(RenderTargetFormat format) {
         return 1;
     default:
         UNIMPLEMENTED_MSG("Unimplemented render target format {}", static_cast<u32>(format));
+        return 1;
     }
 }
 
@@ -119,6 +120,7 @@ u32 DepthFormatBytesPerPixel(DepthFormat format) {
         return 2;
     default:
         UNIMPLEMENTED_MSG("Unimplemented Depth format {}", static_cast<u32>(format));
+        return 1;
     }
 }
 

src/video_core/macro_interpreter.cpp

@@ -171,6 +171,7 @@ u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b)
 
     default:
         UNIMPLEMENTED_MSG("Unimplemented ALU operation {}", static_cast<u32>(operation));
+        return 0;
     }
 }
 
@@ -268,6 +269,7 @@ bool MacroInterpreter::EvaluateBranchCondition(BranchCondition cond, u32 value)
         return value != 0;
     }
     UNREACHABLE();
+    return true;
 }
 
 } // namespace Tegra

src/video_core/morton.cpp

@@ -192,6 +192,7 @@ static MortonCopyFn GetSwizzleFunction(MortonSwizzleMode mode, Surface::PixelFor
         return linear_to_morton_fns[static_cast<std::size_t>(format)];
     }
     UNREACHABLE();
+    return morton_to_linear_fns[static_cast<std::size_t>(format)];
 }
 
 /// 8x8 Z-Order coordinate from 2D coordinates

src/video_core/renderer_base.cpp

@@ -27,4 +27,16 @@ void RendererBase::UpdateCurrentFramebufferLayout() {
     render_window.UpdateCurrentFramebufferLayout(layout.width, layout.height);
 }
 
+void RendererBase::RequestScreenshot(void* data, std::function<void()> callback,
+                                     const Layout::FramebufferLayout& layout) {
+    if (renderer_settings.screenshot_requested) {
+        LOG_ERROR(Render, "A screenshot is already requested or in progress, ignoring the request");
+        return;
+    }
+    renderer_settings.screenshot_bits = data;
+    renderer_settings.screenshot_complete_callback = std::move(callback);
+    renderer_settings.screenshot_framebuffer_layout = layout;
+    renderer_settings.screenshot_requested = true;
+}
+
 } // namespace VideoCore

src/video_core/renderer_base.h

@@ -9,6 +9,7 @@
 #include <optional>
 
 #include "common/common_types.h"
+#include "core/frontend/emu_window.h"
 #include "video_core/gpu.h"
 #include "video_core/rasterizer_interface.h"
 
@@ -21,6 +22,12 @@ namespace VideoCore {
 struct RendererSettings {
     std::atomic_bool use_framelimiter{false};
     std::atomic_bool set_background_color{false};
+
+    // Screenshot
+    std::atomic<bool> screenshot_requested{false};
+    void* screenshot_bits;
+    std::function<void()> screenshot_complete_callback;
+    Layout::FramebufferLayout screenshot_framebuffer_layout;
 };
 
 class RendererBase : NonCopyable {
@@ -57,9 +64,29 @@ public:
         return *rasterizer;
     }
 
+    Core::Frontend::EmuWindow& GetRenderWindow() {
+        return render_window;
+    }
+
+    const Core::Frontend::EmuWindow& GetRenderWindow() const {
+        return render_window;
+    }
+
+    RendererSettings& Settings() {
+        return renderer_settings;
+    }
+
+    const RendererSettings& Settings() const {
+        return renderer_settings;
+    }
+
     /// Refreshes the settings common to all renderers
     void RefreshBaseSettings();
 
+    /// Request a screenshot of the next frame
+    void RequestScreenshot(void* data, std::function<void()> callback,
+                           const Layout::FramebufferLayout& layout);
+
 protected:
     Core::Frontend::EmuWindow& render_window; ///< Reference to the render window handle.
     std::unique_ptr<RasterizerInterface> rasterizer;

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -79,6 +79,26 @@ struct DrawParameters {
     }
 };
 
+struct FramebufferCacheKey {
+    bool is_single_buffer = false;
+    bool stencil_enable = false;
+
+    std::array<GLenum, Maxwell::NumRenderTargets> color_attachments{};
+    std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> colors{};
+    u32 colors_count = 0;
+
+    GLuint zeta = 0;
+
+    auto Tie() const {
+        return std::tie(is_single_buffer, stencil_enable, color_attachments, colors, colors_count,
+                        zeta);
+    }
+
+    bool operator<(const FramebufferCacheKey& rhs) const {
+        return Tie() < rhs.Tie();
+    }
+};
+
 RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo& info)
     : res_cache{*this}, shader_cache{*this}, emu_window{window}, screen_info{info},
       buffer_cache(*this, STREAM_BUFFER_SIZE) {
@@ -90,9 +110,6 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
 
     OpenGLState::ApplyDefaultState();
 
-    // Create render framebuffer
-    framebuffer.Create();
-
     shader_program_manager = std::make_unique<GLShader::ProgramManager>();
     state.draw.shader_program = 0;
     state.Apply();
@@ -361,6 +378,44 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) {
     SyncClipEnabled(clip_distances);
 }
 
+void RasterizerOpenGL::SetupCachedFramebuffer(const FramebufferCacheKey& fbkey,
+                                              OpenGLState& current_state) {
+    const auto [entry, is_cache_miss] = framebuffer_cache.try_emplace(fbkey);
+    auto& framebuffer = entry->second;
+
+    if (is_cache_miss)
+        framebuffer.Create();
+
+    current_state.draw.draw_framebuffer = framebuffer.handle;
+    current_state.ApplyFramebufferState();
+
+    if (!is_cache_miss)
+        return;
+
+    if (fbkey.is_single_buffer) {
+        if (fbkey.color_attachments[0] != GL_NONE) {
+            glFramebufferTexture(GL_DRAW_FRAMEBUFFER, fbkey.color_attachments[0], fbkey.colors[0],
+                                 0);
+        }
+        glDrawBuffer(fbkey.color_attachments[0]);
+    } else {
+        for (std::size_t index = 0; index < Maxwell::NumRenderTargets; ++index) {
+            if (fbkey.colors[index]) {
+                glFramebufferTexture(GL_DRAW_FRAMEBUFFER,
+                                     GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(index),
+                                     fbkey.colors[index], 0);
+            }
+        }
+        glDrawBuffers(fbkey.colors_count, fbkey.color_attachments.data());
+    }
+
+    if (fbkey.zeta) {
+        GLenum zeta_attachment =
+            fbkey.stencil_enable ? GL_DEPTH_STENCIL_ATTACHMENT : GL_DEPTH_ATTACHMENT;
+        glFramebufferTexture(GL_DRAW_FRAMEBUFFER, zeta_attachment, fbkey.zeta, 0);
+    }
+}
+
 std::size_t RasterizerOpenGL::CalculateVertexArraysSize() const {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
@@ -444,10 +499,10 @@ void RasterizerOpenGL::ConfigureFramebuffers(OpenGLState& current_state, bool us
     UNIMPLEMENTED_IF(regs.rt_separate_frag_data != 0);
 
     // Bind the framebuffer surfaces
-    current_state.draw.draw_framebuffer = framebuffer.handle;
-    current_state.ApplyFramebufferState();
     current_state.framebuffer_srgb.enabled = regs.framebuffer_srgb != 0;
 
+    FramebufferCacheKey fbkey;
+
     if (using_color_fb) {
         if (single_color_target) {
             // Used when just a single color attachment is enabled, e.g. for clearing a color buffer
@@ -463,14 +518,12 @@ void RasterizerOpenGL::ConfigureFramebuffers(OpenGLState& current_state, bool us
                 state.framebuffer_srgb.enabled |= color_surface->GetSurfaceParams().srgb_conversion;
             }
 
-            glFramebufferTexture2D(
-                GL_DRAW_FRAMEBUFFER,
-                GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(*single_color_target), GL_TEXTURE_2D,
-                color_surface != nullptr ? color_surface->Texture().handle : 0, 0);
-            glDrawBuffer(GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(*single_color_target));
+            fbkey.is_single_buffer = true;
+            fbkey.color_attachments[0] =
+                GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(*single_color_target);
+            fbkey.colors[0] = color_surface != nullptr ? color_surface->Texture().handle : 0;
         } else {
             // Multiple color attachments are enabled
-            std::array<GLenum, Maxwell::NumRenderTargets> buffers;
             for (std::size_t index = 0; index < Maxwell::NumRenderTargets; ++index) {
                 Surface color_surface = res_cache.GetColorBufferSurface(index, preserve_contents);
 
@@ -485,22 +538,17 @@ void RasterizerOpenGL::ConfigureFramebuffers(OpenGLState& current_state, bool us
                         color_surface->GetSurfaceParams().srgb_conversion;
                 }
 
-                buffers[index] = GL_COLOR_ATTACHMENT0 + regs.rt_control.GetMap(index);
-                glFramebufferTexture2D(
-                    GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(index),
-                    GL_TEXTURE_2D, color_surface != nullptr ? color_surface->Texture().handle : 0,
-                    0);
+                fbkey.color_attachments[index] =
+                    GL_COLOR_ATTACHMENT0 + regs.rt_control.GetMap(index);
+                fbkey.colors[index] =
+                    color_surface != nullptr ? color_surface->Texture().handle : 0;
             }
-            glDrawBuffers(regs.rt_control.count, buffers.data());
+            fbkey.is_single_buffer = false;
+            fbkey.colors_count = regs.rt_control.count;
         }
     } else {
-        // No color attachments are enabled - zero out all of them
-        for (std::size_t index = 0; index < Maxwell::NumRenderTargets; ++index) {
-            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER,
-                                   GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(index), GL_TEXTURE_2D,
-                                   0, 0);
-        }
-        glDrawBuffer(GL_NONE);
+        // No color attachments are enabled - leave them as zero
+        fbkey.is_single_buffer = true;
     }
 
     if (depth_surface) {
@@ -508,22 +556,12 @@ void RasterizerOpenGL::ConfigureFramebuffers(OpenGLState& current_state, bool us
         // the shader doesn't actually write to it.
         depth_surface->MarkAsModified(true, res_cache);
 
-        if (regs.stencil_enable) {
-            // Attach both depth and stencil
-            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
-                                   depth_surface->Texture().handle, 0);
-        } else {
-            // Attach depth
-            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D,
-                                   depth_surface->Texture().handle, 0);
-            // Clear stencil attachment
-            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
-        }
-    } else {
-        // Clear both depth and stencil attachment
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
-                               0);
+        fbkey.zeta = depth_surface->Texture().handle;
+        fbkey.stencil_enable = regs.stencil_enable;
     }
+
+    SetupCachedFramebuffer(fbkey, current_state);
+
     SyncViewport(current_state);
 }
 

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -40,6 +40,7 @@ namespace OpenGL {
 
 struct ScreenInfo;
 struct DrawParameters;
+struct FramebufferCacheKey;
 
 class RasterizerOpenGL : public VideoCore::RasterizerInterface {
 public:
@@ -195,11 +196,12 @@ private:
              OGLVertexArray>
         vertex_array_cache;
 
+    std::map<FramebufferCacheKey, OGLFramebuffer> framebuffer_cache;
+
     std::array<SamplerInfo, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> texture_samplers;
 
     static constexpr std::size_t STREAM_BUFFER_SIZE = 128 * 1024 * 1024;
     OGLBufferCache buffer_cache;
-    OGLFramebuffer framebuffer;
     PrimitiveAssembler primitive_assembler{buffer_cache};
     GLint uniform_buffer_alignment;
 
@@ -214,6 +216,8 @@ private:
 
     void SetupShaders(GLenum primitive_mode);
 
+    void SetupCachedFramebuffer(const FramebufferCacheKey& fbkey, OpenGLState& current_state);
+
     enum class AccelDraw { Disabled, Arrays, Indexed };
     AccelDraw accelerate_draw = AccelDraw::Disabled;
 

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -101,8 +101,18 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
     params.srgb_conversion = config.tic.IsSrgbConversionEnabled();
     params.pixel_format = PixelFormatFromTextureFormat(config.tic.format, config.tic.r_type.Value(),
                                                        params.srgb_conversion);
+
+    if (params.pixel_format == PixelFormat::R16U && config.tsc.depth_compare_enabled) {
+        // Some titles create a 'R16U' (normalized 16-bit) texture with depth_compare enabled,
+        // then attempt to sample from it via a shadow sampler. Convert format to Z16 (which also
+        // causes GetFormatType to properly return 'Depth' below).
+        params.pixel_format = PixelFormat::Z16;
+    }
+
     params.component_type = ComponentTypeFromTexture(config.tic.r_type.Value());
     params.type = GetFormatType(params.pixel_format);
+    UNIMPLEMENTED_IF(params.type == SurfaceType::ColorTexture && config.tsc.depth_compare_enabled);
+
     params.width = Common::AlignUp(config.tic.Width(), GetCompressionFactor(params.pixel_format));
     params.height = Common::AlignUp(config.tic.Height(), GetCompressionFactor(params.pixel_format));
     params.unaligned_height = config.tic.Height();
@@ -257,7 +267,7 @@ static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex
     {GL_R8UI, GL_RED_INTEGER, GL_UNSIGNED_BYTE, ComponentType::UInt, false},           // R8UI
     {GL_RGBA16F, GL_RGBA, GL_HALF_FLOAT, ComponentType::Float, false},                 // RGBA16F
     {GL_RGBA16, GL_RGBA, GL_UNSIGNED_SHORT, ComponentType::UNorm, false},              // RGBA16U
-    {GL_RGBA16UI, GL_RGBA, GL_UNSIGNED_SHORT, ComponentType::UInt, false},             // RGBA16UI
+    {GL_RGBA16UI, GL_RGBA_INTEGER, GL_UNSIGNED_SHORT, ComponentType::UInt, false},     // RGBA16UI
     {GL_R11F_G11F_B10F, GL_RGB, GL_UNSIGNED_INT_10F_11F_11F_REV, ComponentType::Float,
      false},                                                                     // R11FG11FB10F
     {GL_RGBA32UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false}, // RGBA32UI

src/video_core/renderer_opengl/gl_shader_cache.cpp

@@ -2,7 +2,9 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <boost/functional/hash.hpp>
 #include "common/assert.h"
+#include "common/hash.h"
 #include "core/core.h"
 #include "core/memory.h"
 #include "video_core/engines/maxwell_3d.h"
@@ -66,14 +68,17 @@ CachedShader::CachedShader(VAddr addr, Maxwell::ShaderProgram program_type)
         // stage here.
         setup.SetProgramB(GetShaderCode(GetShaderAddress(Maxwell::ShaderProgram::VertexB)));
     case Maxwell::ShaderProgram::VertexB:
+        CalculateProperties();
         program_result = GLShader::GenerateVertexShader(setup);
         gl_type = GL_VERTEX_SHADER;
         break;
     case Maxwell::ShaderProgram::Geometry:
+        CalculateProperties();
         program_result = GLShader::GenerateGeometryShader(setup);
         gl_type = GL_GEOMETRY_SHADER;
         break;
     case Maxwell::ShaderProgram::Fragment:
+        CalculateProperties();
         program_result = GLShader::GenerateFragmentShader(setup);
         gl_type = GL_FRAGMENT_SHADER;
         break;
@@ -140,6 +145,46 @@ GLuint CachedShader::LazyGeometryProgram(OGLProgram& target_program,
     return target_program.handle;
 };
 
+static bool IsSchedInstruction(std::size_t offset, std::size_t main_offset) {
+    // sched instructions appear once every 4 instructions.
+    static constexpr std::size_t SchedPeriod = 4;
+    const std::size_t absolute_offset = offset - main_offset;
+    return (absolute_offset % SchedPeriod) == 0;
+}
+
+static std::size_t CalculateProgramSize(const GLShader::ProgramCode& program) {
+    constexpr std::size_t start_offset = 10;
+    std::size_t offset = start_offset;
+    std::size_t size = start_offset * sizeof(u64);
+    while (offset < program.size()) {
+        const u64 inst = program[offset];
+        if (!IsSchedInstruction(offset, start_offset)) {
+            if (inst == 0 || (inst >> 52) == 0x50b) {
+                break;
+            }
+        }
+        size += sizeof(inst);
+        offset++;
+    }
+    return size;
+}
+
+void CachedShader::CalculateProperties() {
+    setup.program.real_size = CalculateProgramSize(setup.program.code);
+    setup.program.real_size_b = 0;
+    setup.program.unique_identifier = Common::CityHash64(
+        reinterpret_cast<const char*>(setup.program.code.data()), setup.program.real_size);
+    if (program_type == Maxwell::ShaderProgram::VertexA) {
+        std::size_t seed = 0;
+        boost::hash_combine(seed, setup.program.unique_identifier);
+        setup.program.real_size_b = CalculateProgramSize(setup.program.code_b);
+        const u64 identifier_b = Common::CityHash64(
+            reinterpret_cast<const char*>(setup.program.code_b.data()), setup.program.real_size_b);
+        boost::hash_combine(seed, identifier_b);
+        setup.program.unique_identifier = static_cast<u64>(seed);
+    }
+}
+
 ShaderCacheOpenGL::ShaderCacheOpenGL(RasterizerOpenGL& rasterizer) : RasterizerCache{rasterizer} {}
 
 Shader ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) {

src/video_core/renderer_opengl/gl_shader_cache.h

@@ -67,6 +67,7 @@ public:
                                        6, "ShaderTrianglesAdjacency");
         default:
             UNREACHABLE_MSG("Unknown primitive mode.");
+            return LazyGeometryProgram(geometry_programs.points, "points", 1, "ShaderPoints");
         }
     }
 
@@ -81,6 +82,8 @@ private:
     GLuint LazyGeometryProgram(OGLProgram& target_program, const std::string& glsl_topology,
                                u32 max_vertices, const std::string& debug_name);
 
+    void CalculateProperties();
+
     VAddr addr;
     std::size_t shader_length;
     Maxwell::ShaderProgram program_type;

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -50,6 +50,14 @@ public:
     using std::runtime_error::runtime_error;
 };
 
+/// Generates code to use for a swizzle operation.
+static std::string GetSwizzle(u64 elem) {
+    ASSERT(elem <= 3);
+    std::string swizzle = ".";
+    swizzle += "xyzw"[elem];
+    return swizzle;
+}
+
 /// Translate topology
 static std::string GetTopologyName(Tegra::Shader::OutputTopology topology) {
     switch (topology) {
@@ -356,6 +364,7 @@ public:
             return value;
         default:
             UNREACHABLE_MSG("Unimplemented conversion size: {}", static_cast<u32>(size));
+            return value;
         }
     }
 
@@ -618,6 +627,7 @@ public:
             return "floatBitsToInt(" + value + ')';
         } else {
             UNREACHABLE();
+            return value;
         }
     }
 
@@ -920,7 +930,7 @@ private:
         case Attribute::Index::FrontFacing:
             // TODO(Subv): Find out what the values are for the other elements.
             ASSERT(stage == Maxwell3D::Regs::ShaderStage::Fragment);
-            return "vec4(0, 0, 0, uintBitsToFloat(gl_FrontFacing ? 1 : 0))";
+            return "vec4(0, 0, 0, intBitsToFloat(gl_FrontFacing ? -1 : 0))";
         default:
             const u32 index{static_cast<u32>(attribute) -
                             static_cast<u32>(Attribute::Index::Attribute_0)};
@@ -1004,14 +1014,6 @@ private:
         }
     }
 
-    /// Generates code to use for a swizzle operation.
-    static std::string GetSwizzle(u64 elem) {
-        ASSERT(elem <= 3);
-        std::string swizzle = ".";
-        swizzle += "xyzw"[elem];
-        return swizzle;
-    }
-
     ShaderWriter& shader;
     ShaderWriter& declarations;
     std::vector<GLSLRegister> regs;
@@ -1343,7 +1345,7 @@ private:
         regs.SetRegisterToInteger(dest, true, 0, result, 1, 1);
     }
 
-    void WriteTexsInstruction(const Instruction& instr, const std::string& texture) {
+    void WriteTexsInstructionFloat(const Instruction& instr, const std::string& texture) {
         // TEXS has two destination registers and a swizzle. The first two elements in the swizzle
         // go into gpr0+0 and gpr0+1, and the rest goes into gpr28+0 and gpr28+1
 
@@ -1368,6 +1370,38 @@ private:
         }
     }
 
+    void WriteTexsInstructionHalfFloat(const Instruction& instr, const std::string& texture) {
+        // TEXS.F16 destionation registers are packed in two registers in pairs (just like any half
+        // float instruction).
+
+        std::array<std::string, 4> components;
+        u32 written_components = 0;
+
+        for (u32 component = 0; component < 4; ++component) {
+            if (!instr.texs.IsComponentEnabled(component))
+                continue;
+            components[written_components++] = texture + GetSwizzle(component);
+        }
+        if (written_components == 0)
+            return;
+
+        const auto BuildComponent = [&](std::string low, std::string high, bool high_enabled) {
+            return "vec2(" + low + ", " + (high_enabled ? high : "0") + ')';
+        };
+
+        regs.SetRegisterToHalfFloat(
+            instr.gpr0, 0, BuildComponent(components[0], components[1], written_components > 1),
+            Tegra::Shader::HalfMerge::H0_H1, 1, 1);
+
+        if (written_components > 2) {
+            ASSERT(instr.texs.HasTwoDestinations());
+            regs.SetRegisterToHalfFloat(
+                instr.gpr28, 0,
+                BuildComponent(components[2], components[3], written_components > 3),
+                Tegra::Shader::HalfMerge::H0_H1, 1, 1);
+        }
+    }
+
     static u32 TextureCoordinates(Tegra::Shader::TextureType texture_type) {
         switch (texture_type) {
         case Tegra::Shader::TextureType::Texture1D:
@@ -1559,23 +1593,21 @@ private:
                                 process_mode == Tegra::Shader::TextureProcessMode::LL ||
                                 process_mode == Tegra::Shader::TextureProcessMode::LLA;
 
+        // LOD selection (either via bias or explicit textureLod) not supported in GL for
+        // sampler2DArrayShadow and samplerCubeArrayShadow.
         const bool gl_lod_supported = !(
             (texture_type == Tegra::Shader::TextureType::Texture2D && is_array && depth_compare) ||
-            (texture_type == Tegra::Shader::TextureType::TextureCube && !is_array &&
-             depth_compare));
+            (texture_type == Tegra::Shader::TextureType::TextureCube && is_array && depth_compare));
 
         const std::string read_method = lod_needed && gl_lod_supported ? "textureLod(" : "texture(";
         std::string texture = read_method + sampler + ", coord";
 
-        if (process_mode != Tegra::Shader::TextureProcessMode::None) {
+        UNIMPLEMENTED_IF(process_mode != Tegra::Shader::TextureProcessMode::None &&
+                         !gl_lod_supported);
+
+        if (process_mode != Tegra::Shader::TextureProcessMode::None && gl_lod_supported) {
             if (process_mode == Tegra::Shader::TextureProcessMode::LZ) {
-                if (gl_lod_supported) {
-                    texture += ", 0";
-                } else {
-                    // Lod 0 is emulated by a big negative bias
-                    // in scenarios that are not supported by glsl
-                    texture += ", -1000";
-                }
+                texture += ", 0.0";
             } else {
                 // If present, lod or bias are always stored in the register indexed by the
                 // gpr20
@@ -1613,15 +1645,15 @@ private:
         if (depth_compare && !is_array && texture_type == Tegra::Shader::TextureType::Texture1D) {
             coord += ",0.0";
         }
+        if (is_array) {
+            coord += ',' + regs.GetRegisterAsInteger(array_register);
+        }
         if (depth_compare) {
             // Depth is always stored in the register signaled by gpr20
             // or in the next register if lod or bias are used
             const u64 depth_register = instr.gpr20.Value() + (lod_bias_enabled ? 1 : 0);
             coord += ',' + regs.GetRegisterAsFloat(depth_register);
         }
-        if (is_array) {
-            coord += ',' + regs.GetRegisterAsInteger(array_register);
-        }
         coord += ");";
         return std::make_pair(
             coord, GetTextureCode(instr, texture_type, process_mode, depth_compare, is_array, 0));
@@ -1649,20 +1681,20 @@ private:
         for (size_t i = 0; i < coord_count; ++i) {
             const bool last = (i == (coord_count - 1)) && (coord_count > 1);
             coord += regs.GetRegisterAsFloat(last ? last_coord_register : coord_register + i);
-            if (!last) {
+            if (i < coord_count - 1) {
                 coord += ',';
             }
         }
 
+        if (is_array) {
+            coord += ',' + regs.GetRegisterAsInteger(array_register);
+        }
         if (depth_compare) {
             // Depth is always stored in the register signaled by gpr20
             // or in the next register if lod or bias are used
             const u64 depth_register = instr.gpr20.Value() + (lod_bias_enabled ? 1 : 0);
             coord += ',' + regs.GetRegisterAsFloat(depth_register);
         }
-        if (is_array) {
-            coord += ',' + regs.GetRegisterAsInteger(array_register);
-        }
         coord += ");";
 
         return std::make_pair(coord,
@@ -1670,6 +1702,99 @@ private:
                                              is_array, (coord_count > 2 ? 1 : 0)));
     }
 
+    std::pair<std::string, std::string> GetTLD4Code(const Instruction& instr,
+                                                    const Tegra::Shader::TextureType texture_type,
+                                                    const bool depth_compare, const bool is_array) {
+
+        const size_t coord_count = TextureCoordinates(texture_type);
+        const size_t total_coord_count = coord_count + (is_array ? 1 : 0);
+        const size_t total_reg_count = total_coord_count + (depth_compare ? 1 : 0);
+
+        constexpr std::array<const char*, 5> coord_container{
+            {"", "", "vec2 coord = vec2(", "vec3 coord = vec3(", "vec4 coord = vec4("}};
+
+        // If enabled arrays index is always stored in the gpr8 field
+        const u64 array_register = instr.gpr8.Value();
+        // First coordinate index is the gpr8 or gpr8 + 1 when arrays are used
+        const u64 coord_register = array_register + (is_array ? 1 : 0);
+
+        std::string coord = coord_container[total_coord_count];
+        for (size_t i = 0; i < coord_count;) {
+            coord += regs.GetRegisterAsFloat(coord_register + i);
+            ++i;
+            if (i != coord_count) {
+                coord += ',';
+            }
+        }
+
+        if (is_array) {
+            coord += ',' + regs.GetRegisterAsInteger(array_register);
+        }
+        coord += ");";
+
+        const std::string sampler =
+            GetSampler(instr.sampler, texture_type, is_array, depth_compare);
+
+        std::string texture = "textureGather(" + sampler + ", coord, ";
+        if (depth_compare) {
+            // Depth is always stored in the register signaled by gpr20
+            texture += regs.GetRegisterAsFloat(instr.gpr20.Value()) + ')';
+        } else {
+            texture += std::to_string(instr.tld4.component) + ')';
+        }
+        return std::make_pair(coord, texture);
+    }
+
+    std::pair<std::string, std::string> GetTLDSCode(const Instruction& instr,
+                                                    const Tegra::Shader::TextureType texture_type,
+                                                    const bool is_array) {
+
+        const size_t coord_count = TextureCoordinates(texture_type);
+        const size_t total_coord_count = coord_count + (is_array ? 1 : 0);
+        const bool lod_enabled =
+            instr.tlds.GetTextureProcessMode() == Tegra::Shader::TextureProcessMode::LL;
+
+        constexpr std::array<const char*, 4> coord_container{
+            {"", "int coords = (", "ivec2 coords = ivec2(", "ivec3 coords = ivec3("}};
+
+        std::string coord = coord_container[total_coord_count];
+
+        // If enabled arrays index is always stored in the gpr8 field
+        const u64 array_register = instr.gpr8.Value();
+
+        // if is array gpr20 is used
+        const u64 coord_register = is_array ? instr.gpr20.Value() : instr.gpr8.Value();
+
+        const u64 last_coord_register =
+            ((coord_count > 2) || (coord_count == 2 && !lod_enabled)) && !is_array
+                ? static_cast<u64>(instr.gpr20.Value())
+                : coord_register + 1;
+
+        for (size_t i = 0; i < coord_count; ++i) {
+            const bool last = (i == (coord_count - 1)) && (coord_count > 1);
+            coord += regs.GetRegisterAsInteger(last ? last_coord_register : coord_register + i);
+            if (i < coord_count - 1) {
+                coord += ',';
+            }
+        }
+        if (is_array) {
+            coord += ',' + regs.GetRegisterAsInteger(array_register);
+        }
+        coord += ");";
+
+        const std::string sampler = GetSampler(instr.sampler, texture_type, is_array, false);
+
+        std::string texture = "texelFetch(" + sampler + ", coords";
+
+        if (lod_enabled) {
+            // When lod is used always is in grp20
+            texture += ", " + regs.GetRegisterAsInteger(instr.gpr20) + ')';
+        } else {
+            texture += ", 0)";
+        }
+        return std::make_pair(coord, texture);
+    }
+
     /**
      * Compiles a single instruction from Tegra to GLSL.
      * @param offset the offset of the Tegra shader instruction.
@@ -1742,9 +1867,6 @@ private:
                 UNIMPLEMENTED_IF_MSG(instr.fmul.tab5cb8_2 != 0,
                                      "FMUL tab5cb8_2({}) is not implemented",
                                      instr.fmul.tab5cb8_2.Value());
-                UNIMPLEMENTED_IF_MSG(instr.fmul.tab5c68_1 != 0,
-                                     "FMUL tab5cb8_1({}) is not implemented",
-                                     instr.fmul.tab5c68_1.Value());
                 UNIMPLEMENTED_IF_MSG(
                     instr.fmul.tab5c68_0 != 1, "FMUL tab5cb8_0({}) is not implemented",
                     instr.fmul.tab5c68_0
@@ -1754,7 +1876,26 @@ private:
 
                 op_b = GetOperandAbsNeg(op_b, false, instr.fmul.negate_b);
 
-                regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b, 1, 1,
+                std::string postfactor_op;
+                if (instr.fmul.postfactor != 0) {
+                    s8 postfactor = static_cast<s8>(instr.fmul.postfactor);
+
+                    // postfactor encoded as 3-bit 1's complement in instruction,
+                    // interpreted with below logic.
+                    if (postfactor >= 4) {
+                        postfactor = 7 - postfactor;
+                    } else {
+                        postfactor = 0 - postfactor;
+                    }
+
+                    if (postfactor > 0) {
+                        postfactor_op = " * " + std::to_string(1 << postfactor);
+                    } else {
+                        postfactor_op = " / " + std::to_string(1 << -postfactor);
+                    }
+                }
+
+                regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b + postfactor_op, 1, 1,
                                         instr.alu.saturate_d, 0, true);
                 break;
             }
@@ -1923,6 +2064,8 @@ private:
                             std::to_string(instr.alu.GetSignedImm20_20())};
                 default:
                     UNREACHABLE();
+                    return {regs.GetRegisterAsInteger(instr.gpr39, 0, false),
+                            std::to_string(instr.alu.GetSignedImm20_20())};
                 }
             }();
             const std::string offset = '(' + packed_shift + " & 0xff)";
@@ -2766,33 +2909,33 @@ private:
                 const bool depth_compare =
                     instr.texs.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC);
                 const auto process_mode = instr.texs.GetTextureProcessMode();
+
                 UNIMPLEMENTED_IF_MSG(instr.texs.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP),
                                      "NODEP is not implemented");
 
                 const auto scope = shader.Scope();
 
-                const auto [coord, texture] =
+                auto [coord, texture] =
                     GetTEXSCode(instr, texture_type, process_mode, depth_compare, is_array);
 
                 shader.AddLine(coord);
 
-                if (!depth_compare) {
-                    shader.AddLine("vec4 texture_tmp = " + texture + ';');
-
-                } else {
-                    shader.AddLine("vec4 texture_tmp = vec4(" + texture + ");");
+                if (depth_compare) {
+                    texture = "vec4(" + texture + ')';
                 }
+                shader.AddLine("vec4 texture_tmp = " + texture + ';');
 
-                WriteTexsInstruction(instr, "texture_tmp");
+                if (instr.texs.fp32_flag) {
+                    WriteTexsInstructionFloat(instr, "texture_tmp");
+                } else {
+                    WriteTexsInstructionHalfFloat(instr, "texture_tmp");
+                }
                 break;
             }
             case OpCode::Id::TLDS: {
                 const Tegra::Shader::TextureType texture_type{instr.tlds.GetTextureType()};
                 const bool is_array{instr.tlds.IsArrayTexture()};
 
-                ASSERT(texture_type == Tegra::Shader::TextureType::Texture2D);
-                ASSERT(is_array == false);
-
                 UNIMPLEMENTED_IF_MSG(instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP),
                                      "NODEP is not implemented");
                 UNIMPLEMENTED_IF_MSG(instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI),
@@ -2800,54 +2943,16 @@ private:
                 UNIMPLEMENTED_IF_MSG(instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::MZ),
                                      "MZ is not implemented");
 
-                u32 extra_op_offset = 0;
+                const auto [coord, texture] = GetTLDSCode(instr, texture_type, is_array);
 
-                ShaderScopedScope scope = shader.Scope();
+                const auto scope = shader.Scope();
 
-                switch (texture_type) {
-                case Tegra::Shader::TextureType::Texture1D: {
-                    const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
-                    shader.AddLine("float coords = " + x + ';');
-                    break;
-                }
-                case Tegra::Shader::TextureType::Texture2D: {
-                    UNIMPLEMENTED_IF_MSG(is_array, "Unhandled 2d array texture");
-
-                    const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
-                    const std::string y = regs.GetRegisterAsInteger(instr.gpr20);
-                    // shader.AddLine("ivec2 coords = ivec2(" + x + ", " + y + ");");
-                    shader.AddLine("ivec2 coords = ivec2(" + x + ", " + y + ");");
-                    extra_op_offset = 1;
-                    break;
-                }
-                default:
-                    UNIMPLEMENTED_MSG("Unhandled texture type {}", static_cast<u32>(texture_type));
-                }
-                const std::string sampler =
-                    GetSampler(instr.sampler, texture_type, is_array, false);
-
-                const std::string texture = [&]() {
-                    switch (instr.tlds.GetTextureProcessMode()) {
-                    case Tegra::Shader::TextureProcessMode::LZ:
-                        return "texelFetch(" + sampler + ", coords, 0)";
-                    case Tegra::Shader::TextureProcessMode::LL:
-                        shader.AddLine(
-                            "float lod = " +
-                            regs.GetRegisterAsInteger(instr.gpr20.Value() + extra_op_offset) + ';');
-                        return "texelFetch(" + sampler + ", coords, lod)";
-                    default:
-                        UNIMPLEMENTED_MSG("Unhandled texture process mode {}",
-                                          static_cast<u32>(instr.tlds.GetTextureProcessMode()));
-                        return "texelFetch(" + sampler + ", coords, 0)";
-                    }
-                }();
-
-                WriteTexsInstruction(instr, texture);
+                shader.AddLine(coord);
+                shader.AddLine("vec4 texture_tmp = " + texture + ';');
+                WriteTexsInstructionFloat(instr, "texture_tmp");
                 break;
             }
             case OpCode::Id::TLD4: {
-                ASSERT(instr.tld4.texture_type == Tegra::Shader::TextureType::Texture2D);
-                ASSERT(instr.tld4.array == 0);
 
                 UNIMPLEMENTED_IF_MSG(instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP),
                                      "NODEP is not implemented");
@@ -2857,56 +2962,29 @@ private:
                                      "NDV is not implemented");
                 UNIMPLEMENTED_IF_MSG(instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::PTP),
                                      "PTP is not implemented");
+
+                auto texture_type = instr.tld4.texture_type.Value();
                 const bool depth_compare =
                     instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC);
-                auto texture_type = instr.tld4.texture_type.Value();
-                u32 num_coordinates = TextureCoordinates(texture_type);
-                if (depth_compare)
-                    num_coordinates += 1;
+                const bool is_array = instr.tld4.array != 0;
+
+                const auto [coord, texture] =
+                    GetTLD4Code(instr, texture_type, depth_compare, is_array);
 
                 const auto scope = shader.Scope();
 
-                switch (num_coordinates) {
-                case 2: {
-                    const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
-                    shader.AddLine("vec2 coords = vec2(" + x + ", " + y + ");");
-                    break;
-                }
-                case 3: {
-                    const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
-                    const std::string z = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
-                    shader.AddLine("vec3 coords = vec3(" + x + ", " + y + ", " + z + ");");
-                    break;
-                }
-                default:
-                    UNIMPLEMENTED_MSG("Unhandled coordinates number {}",
-                                      static_cast<u32>(num_coordinates));
-                    const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
-                    shader.AddLine("vec2 coords = vec2(" + x + ", " + y + ");");
-                    texture_type = Tegra::Shader::TextureType::Texture2D;
-                }
+                shader.AddLine(coord);
+                std::size_t dest_elem{};
 
-                const std::string sampler =
-                    GetSampler(instr.sampler, texture_type, false, depth_compare);
-
-                const std::string texture = "textureGather(" + sampler + ", coords, " +
-                                            std::to_string(instr.tld4.component) + ')';
-
-                if (depth_compare) {
-                    regs.SetRegisterToFloat(instr.gpr0, 0, texture, 1, 1, false);
-                } else {
-                    std::size_t dest_elem{};
-                    for (std::size_t elem = 0; elem < 4; ++elem) {
-                        if (!instr.tex.IsComponentEnabled(elem)) {
-                            // Skip disabled components
-                            continue;
-                        }
-                        regs.SetRegisterToFloat(instr.gpr0, elem, texture, 1, 4, false, dest_elem);
-                        ++dest_elem;
+                shader.AddLine("vec4 texture_tmp = " + texture + ';');
+                for (std::size_t elem = 0; elem < 4; ++elem) {
+                    if (!instr.tex.IsComponentEnabled(elem)) {
+                        // Skip disabled components
+                        continue;
                     }
+                    regs.SetRegisterToFloat(instr.gpr0, elem, "texture_tmp", 1, 4, false,
+                                            dest_elem);
+                    ++dest_elem;
                 }
                 break;
             }
@@ -2920,27 +2998,31 @@ private:
 
                 const auto scope = shader.Scope();
 
+                std::string coords;
+
                 const bool depth_compare =
                     instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC);
-                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20);
-                // TODO(Subv): Figure out how the sampler type is encoded in the TLD4S instruction.
+
                 const std::string sampler = GetSampler(
                     instr.sampler, Tegra::Shader::TextureType::Texture2D, false, depth_compare);
-                if (depth_compare) {
-                    // Note: TLD4S coordinate encoding works just like TEXS's
-                    const std::string op_y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
-                    shader.AddLine("vec3 coords = vec3(" + op_a + ", " + op_y + ", " + op_b + ");");
-                } else {
-                    shader.AddLine("vec2 coords = vec2(" + op_a + ", " + op_b + ");");
-                }
 
-                std::string texture = "textureGather(" + sampler + ", coords, " +
-                                      std::to_string(instr.tld4s.component) + ')';
-                if (depth_compare) {
-                    texture = "vec4(" + texture + ')';
+                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
+                coords = "vec2 coords = vec2(" + op_a + ", ";
+                std::string texture = "textureGather(" + sampler + ", coords, ";
+
+                if (!depth_compare) {
+                    const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20);
+                    coords += op_b + ");";
+                    texture += std::to_string(instr.tld4s.component) + ')';
+                } else {
+                    const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    const std::string op_c = regs.GetRegisterAsFloat(instr.gpr20);
+                    coords += op_b + ");";
+                    texture += op_c + ')';
                 }
-                WriteTexsInstruction(instr, texture);
+                shader.AddLine(coords);
+                shader.AddLine("vec4 texture_tmp = " + texture + ';');
+                WriteTexsInstructionFloat(instr, "texture_tmp");
                 break;
             }
             case OpCode::Id::TXQ: {
@@ -3234,6 +3316,7 @@ private:
                     return std::to_string(instr.r2p.immediate_mask);
                 default:
                     UNREACHABLE();
+                    return std::to_string(instr.r2p.immediate_mask);
                 }
             }();
             const std::string mask = '(' + regs.GetRegisterAsInteger(instr.gpr8, 0, false) +
@@ -3697,7 +3780,10 @@ private:
                 }
                 break;
             }
-            default: { UNIMPLEMENTED_MSG("Unhandled instruction: {}", opcode->get().GetName()); }
+            default: {
+                UNIMPLEMENTED_MSG("Unhandled instruction: {}", opcode->get().GetName());
+                break;
+            }
             }
 
             break;
@@ -3852,4 +3938,4 @@ std::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u
     return {};
 }
 
-} // namespace OpenGL::GLShader::Decompiler
+} // namespace OpenGL::GLShader::Decompiler

src/video_core/renderer_opengl/gl_shader_gen.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <fmt/format.h>
 #include "common/assert.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_opengl/gl_shader_decompiler.h"
@@ -16,6 +17,8 @@ static constexpr u32 PROGRAM_OFFSET{10};
 ProgramResult GenerateVertexShader(const ShaderSetup& setup) {
     std::string out = "#version 430 core\n";
     out += "#extension GL_ARB_separate_shader_objects : enable\n\n";
+    const std::string id = fmt::format("{:016x}", setup.program.unique_identifier);
+    out += "// Shader Unique Id: VS" + id + "\n\n";
     out += Decompiler::GetCommonDeclarations();
 
     out += R"(
@@ -84,6 +87,8 @@ void main() {
 ProgramResult GenerateGeometryShader(const ShaderSetup& setup) {
     // Version is intentionally skipped in shader generation, it's added by the lazy compilation.
     std::string out = "#extension GL_ARB_separate_shader_objects : enable\n\n";
+    const std::string id = fmt::format("{:016x}", setup.program.unique_identifier);
+    out += "// Shader Unique Id: GS" + id + "\n\n";
     out += Decompiler::GetCommonDeclarations();
     out += "bool exec_geometry();\n";
 
@@ -117,6 +122,8 @@ void main() {
 ProgramResult GenerateFragmentShader(const ShaderSetup& setup) {
     std::string out = "#version 430 core\n";
     out += "#extension GL_ARB_separate_shader_objects : enable\n\n";
+    const std::string id = fmt::format("{:016x}", setup.program.unique_identifier);
+    out += "// Shader Unique Id: FS" + id + "\n\n";
     out += Decompiler::GetCommonDeclarations();
     out += "bool exec_fragment();\n";
 

src/video_core/renderer_opengl/gl_shader_gen.h

@@ -177,6 +177,9 @@ struct ShaderSetup {
     struct {
         ProgramCode code;
         ProgramCode code_b; // Used for dual vertex shaders
+        u64 unique_identifier;
+        std::size_t real_size;
+        std::size_t real_size_b;
     } program;
 
     /// Used in scenarios where we have a dual vertex shaders

src/video_core/renderer_opengl/renderer_opengl.cpp

@@ -138,7 +138,12 @@ void RendererOpenGL::SwapBuffers(
 
         // Load the framebuffer from memory, draw it to the screen, and swap buffers
         LoadFBToScreenInfo(*framebuffer);
-        DrawScreen();
+
+        if (renderer_settings.screenshot_requested)
+            CaptureScreenshot();
+
+        DrawScreen(render_window.GetFramebufferLayout());
+
         render_window.SwapBuffers();
     }
 
@@ -383,14 +388,13 @@ void RendererOpenGL::DrawScreenTriangles(const ScreenInfo& screen_info, float x,
 /**
  * Draws the emulated screens to the emulator window.
  */
-void RendererOpenGL::DrawScreen() {
+void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) {
     if (renderer_settings.set_background_color) {
         // Update background color before drawing
         glClearColor(Settings::values.bg_red, Settings::values.bg_green, Settings::values.bg_blue,
                      0.0f);
     }
 
-    const auto& layout = render_window.GetFramebufferLayout();
     const auto& screen = layout.screen;
 
     glViewport(0, 0, layout.width, layout.height);
@@ -414,6 +418,37 @@ void RendererOpenGL::DrawScreen() {
 /// Updates the framerate
 void RendererOpenGL::UpdateFramerate() {}
 
+void RendererOpenGL::CaptureScreenshot() {
+    // Draw the current frame to the screenshot framebuffer
+    screenshot_framebuffer.Create();
+    GLuint old_read_fb = state.draw.read_framebuffer;
+    GLuint old_draw_fb = state.draw.draw_framebuffer;
+    state.draw.read_framebuffer = state.draw.draw_framebuffer = screenshot_framebuffer.handle;
+    state.Apply();
+
+    Layout::FramebufferLayout layout{renderer_settings.screenshot_framebuffer_layout};
+
+    GLuint renderbuffer;
+    glGenRenderbuffers(1, &renderbuffer);
+    glBindRenderbuffer(GL_RENDERBUFFER, renderbuffer);
+    glRenderbufferStorage(GL_RENDERBUFFER, GL_RGB8, layout.width, layout.height);
+    glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_RENDERBUFFER, renderbuffer);
+
+    DrawScreen(layout);
+
+    glReadPixels(0, 0, layout.width, layout.height, GL_BGRA, GL_UNSIGNED_INT_8_8_8_8_REV,
+                 renderer_settings.screenshot_bits);
+
+    screenshot_framebuffer.Release();
+    state.draw.read_framebuffer = old_read_fb;
+    state.draw.draw_framebuffer = old_draw_fb;
+    state.Apply();
+    glDeleteRenderbuffers(1, &renderbuffer);
+
+    renderer_settings.screenshot_complete_callback();
+    renderer_settings.screenshot_requested = false;
+}
+
 static const char* GetSource(GLenum source) {
 #define RET(s)                                                                                     \
     case GL_DEBUG_SOURCE_##s:                                                                      \
@@ -427,6 +462,7 @@ static const char* GetSource(GLenum source) {
         RET(OTHER);
     default:
         UNREACHABLE();
+        return "Unknown source";
     }
 #undef RET
 }
@@ -445,6 +481,7 @@ static const char* GetType(GLenum type) {
         RET(MARKER);
     default:
         UNREACHABLE();
+        return "Unknown type";
     }
 #undef RET
 }

src/video_core/renderer_opengl/renderer_opengl.h

@@ -16,6 +16,10 @@ namespace Core::Frontend {
 class EmuWindow;
 }
 
+namespace Layout {
+class FramebufferLayout;
+}
+
 namespace OpenGL {
 
 /// Structure used for storing information about the textures for the Switch screen
@@ -66,10 +70,12 @@ private:
 
     void ConfigureFramebufferTexture(TextureInfo& texture,
                                      const Tegra::FramebufferConfig& framebuffer);
-    void DrawScreen();
+    void DrawScreen(const Layout::FramebufferLayout& layout);
     void DrawScreenTriangles(const ScreenInfo& screen_info, float x, float y, float w, float h);
     void UpdateFramerate();
 
+    void CaptureScreenshot();
+
     // Loads framebuffer from emulated memory into the display information structure
     void LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuffer);
     // Fills active OpenGL texture with the given RGBA color.
@@ -82,6 +88,7 @@ private:
     OGLVertexArray vertex_array;
     OGLBuffer vertex_buffer;
     OGLProgram shader;
+    OGLFramebuffer screenshot_framebuffer;
 
     /// Display information for Switch screen
     ScreenInfo screen_info;

src/video_core/surface.cpp

@@ -65,6 +65,7 @@ PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
     default:
         LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
         UNREACHABLE();
+        return PixelFormat::S8Z24;
     }
 }
 
@@ -141,6 +142,7 @@ PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format)
     default:
         LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
         UNREACHABLE();
+        return PixelFormat::RGBA8_SRGB;
     }
 }
 
@@ -327,6 +329,7 @@ PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
         LOG_CRITICAL(HW_GPU, "Unimplemented format={}, component_type={}", static_cast<u32>(format),
                      static_cast<u32>(component_type));
         UNREACHABLE();
+        return PixelFormat::ABGR8U;
     }
 }
 
@@ -346,6 +349,7 @@ ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
     default:
         LOG_CRITICAL(HW_GPU, "Unimplemented component type={}", static_cast<u32>(type));
         UNREACHABLE();
+        return ComponentType::UNorm;
     }
 }
 
@@ -393,6 +397,7 @@ ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format) {
     default:
         LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
         UNREACHABLE();
+        return ComponentType::UNorm;
     }
 }
 
@@ -403,6 +408,7 @@ PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat
     default:
         LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
         UNREACHABLE();
+        return PixelFormat::ABGR8U;
     }
 }
 
@@ -418,6 +424,7 @@ ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format) {
     default:
         LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
         UNREACHABLE();
+        return ComponentType::UNorm;
     }
 }
 

src/video_core/textures/decoders.cpp

@@ -226,7 +226,7 @@ u32 BytesPerPixel(TextureFormat format) {
         return 8;
     default:
         UNIMPLEMENTED_MSG("Format not implemented");
-        break;
+        return 1;
     }
 }
 

src/video_core/video_core.cpp

@@ -3,6 +3,8 @@
 // Refer to the license.txt file included.
 
 #include <memory>
+#include "core/core.h"
+#include "core/settings.h"
 #include "video_core/renderer_base.h"
 #include "video_core/renderer_opengl/renderer_opengl.h"
 #include "video_core/video_core.h"
@@ -13,4 +15,10 @@ std::unique_ptr<RendererBase> CreateRenderer(Core::Frontend::EmuWindow& emu_wind
     return std::make_unique<OpenGL::RendererOpenGL>(emu_window);
 }
 
+u16 GetResolutionScaleFactor(const RendererBase& renderer) {
+    return !Settings::values.resolution_factor
+               ? renderer.GetRenderWindow().GetFramebufferLayout().GetScalingRatio()
+               : Settings::values.resolution_factor;
+}
+
 } // namespace VideoCore

src/video_core/video_core.h

@@ -22,4 +22,6 @@ class RendererBase;
  */
 std::unique_ptr<RendererBase> CreateRenderer(Core::Frontend::EmuWindow& emu_window);
 
+u16 GetResolutionScaleFactor(const RendererBase& renderer);
+
 } // namespace VideoCore

src/yuzu/CMakeLists.txt

@@ -7,6 +7,8 @@ add_executable(yuzu
     Info.plist
     about_dialog.cpp
     about_dialog.h
+    applets/profile_select.cpp
+    applets/profile_select.h
     applets/software_keyboard.cpp
     applets/software_keyboard.h
     bootmanager.cpp
@@ -31,10 +33,14 @@ add_executable(yuzu
     configuration/configure_input.h
     configuration/configure_input_player.cpp
     configuration/configure_input_player.h
+    configuration/configure_input_simple.cpp
+    configuration/configure_input_simple.h
     configuration/configure_mouse_advanced.cpp
     configuration/configure_mouse_advanced.h
     configuration/configure_system.cpp
     configuration/configure_system.h
+    configuration/configure_per_general.cpp
+    configuration/configure_per_general.h
     configuration/configure_touchscreen_advanced.cpp
     configuration/configure_touchscreen_advanced.h
     configuration/configure_web.cpp
@@ -85,7 +91,9 @@ set(UIS
     configuration/configure_graphics.ui
     configuration/configure_input.ui
     configuration/configure_input_player.ui
+    configuration/configure_input_simple.ui
     configuration/configure_mouse_advanced.ui
+    configuration/configure_per_general.ui
     configuration/configure_system.ui
     configuration/configure_touchscreen_advanced.ui
     configuration/configure_web.ui

src/yuzu/applets/profile_select.cpp

@@ -0,0 +1,168 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <mutex>
+#include <QDialogButtonBox>
+#include <QLabel>
+#include <QLineEdit>
+#include <QScrollArea>
+#include <QStandardItemModel>
+#include <QVBoxLayout>
+#include "common/file_util.h"
+#include "common/string_util.h"
+#include "core/hle/lock.h"
+#include "yuzu/applets/profile_select.h"
+#include "yuzu/main.h"
+
+// Same backup JPEG used by acc IProfile::GetImage if no jpeg found
+constexpr std::array<u8, 107> backup_jpeg{
+    0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x02, 0x02,
+    0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04, 0x08, 0x06, 0x06, 0x05,
+    0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a, 0x0c, 0x0f, 0x0c, 0x0a, 0x0b, 0x0e,
+    0x0b, 0x09, 0x09, 0x0d, 0x11, 0x0d, 0x0e, 0x0f, 0x10, 0x10, 0x11, 0x10, 0x0a, 0x0c, 0x12, 0x13,
+    0x12, 0x10, 0x13, 0x0f, 0x10, 0x10, 0x10, 0xff, 0xc9, 0x00, 0x0b, 0x08, 0x00, 0x01, 0x00, 0x01,
+    0x01, 0x01, 0x11, 0x00, 0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08,
+    0x01, 0x01, 0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
+};
+
+QString FormatUserEntryText(const QString& username, Service::Account::UUID uuid) {
+    return QtProfileSelectionDialog::tr(
+               "%1\n%2", "%1 is the profile username, %2 is the formatted UUID (e.g. "
+                         "00112233-4455-6677-8899-AABBCCDDEEFF))")
+        .arg(username, QString::fromStdString(uuid.FormatSwitch()));
+}
+
+QString GetImagePath(Service::Account::UUID uuid) {
+    const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                      "/system/save/8000000000000010/su/avators/" + uuid.FormatSwitch() + ".jpg";
+    return QString::fromStdString(path);
+}
+
+QPixmap GetIcon(Service::Account::UUID uuid) {
+    QPixmap icon{GetImagePath(uuid)};
+
+    if (!icon) {
+        icon.fill(Qt::black);
+        icon.loadFromData(backup_jpeg.data(), static_cast<u32>(backup_jpeg.size()));
+    }
+
+    return icon.scaled(64, 64, Qt::IgnoreAspectRatio, Qt::SmoothTransformation);
+}
+
+QtProfileSelectionDialog::QtProfileSelectionDialog(QWidget* parent)
+    : QDialog(parent), profile_manager(std::make_unique<Service::Account::ProfileManager>()) {
+    outer_layout = new QVBoxLayout;
+
+    instruction_label = new QLabel(tr("Select a user:"));
+
+    scroll_area = new QScrollArea;
+
+    buttons = new QDialogButtonBox;
+    buttons->addButton(tr("Cancel"), QDialogButtonBox::RejectRole);
+    buttons->addButton(tr("OK"), QDialogButtonBox::AcceptRole);
+
+    connect(buttons, &QDialogButtonBox::accepted, this, &QtProfileSelectionDialog::accept);
+    connect(buttons, &QDialogButtonBox::rejected, this, &QtProfileSelectionDialog::reject);
+
+    outer_layout->addWidget(instruction_label);
+    outer_layout->addWidget(scroll_area);
+    outer_layout->addWidget(buttons);
+
+    layout = new QVBoxLayout;
+    tree_view = new QTreeView;
+    item_model = new QStandardItemModel(tree_view);
+    tree_view->setModel(item_model);
+
+    tree_view->setAlternatingRowColors(true);
+    tree_view->setSelectionMode(QHeaderView::SingleSelection);
+    tree_view->setSelectionBehavior(QHeaderView::SelectRows);
+    tree_view->setVerticalScrollMode(QHeaderView::ScrollPerPixel);
+    tree_view->setHorizontalScrollMode(QHeaderView::ScrollPerPixel);
+    tree_view->setSortingEnabled(true);
+    tree_view->setEditTriggers(QHeaderView::NoEditTriggers);
+    tree_view->setUniformRowHeights(true);
+    tree_view->setIconSize({64, 64});
+    tree_view->setContextMenuPolicy(Qt::NoContextMenu);
+
+    item_model->insertColumns(0, 1);
+    item_model->setHeaderData(0, Qt::Horizontal, "Users");
+
+    // We must register all custom types with the Qt Automoc system so that we are able to use it
+    // with signals/slots. In this case, QList falls under the umbrells of custom types.
+    qRegisterMetaType<QList<QStandardItem*>>("QList<QStandardItem*>");
+
+    layout->setContentsMargins(0, 0, 0, 0);
+    layout->setSpacing(0);
+    layout->addWidget(tree_view);
+
+    scroll_area->setLayout(layout);
+
+    connect(tree_view, &QTreeView::clicked, this, &QtProfileSelectionDialog::SelectUser);
+
+    const auto& profiles = profile_manager->GetAllUsers();
+    for (const auto& user : profiles) {
+        Service::Account::ProfileBase profile;
+        if (!profile_manager->GetProfileBase(user, profile))
+            continue;
+
+        const auto username = Common::StringFromFixedZeroTerminatedBuffer(
+            reinterpret_cast<const char*>(profile.username.data()), profile.username.size());
+
+        list_items.push_back(QList<QStandardItem*>{new QStandardItem{
+            GetIcon(user), FormatUserEntryText(QString::fromStdString(username), user)}});
+    }
+
+    for (const auto& item : list_items)
+        item_model->appendRow(item);
+
+    setLayout(outer_layout);
+    setWindowTitle(tr("Profile Selector"));
+    resize(550, 400);
+}
+
+QtProfileSelectionDialog::~QtProfileSelectionDialog() = default;
+
+void QtProfileSelectionDialog::accept() {
+    ok = true;
+    QDialog::accept();
+}
+
+void QtProfileSelectionDialog::reject() {
+    ok = false;
+    user_index = 0;
+    QDialog::reject();
+}
+
+bool QtProfileSelectionDialog::GetStatus() const {
+    return ok;
+}
+
+u32 QtProfileSelectionDialog::GetIndex() const {
+    return user_index;
+}
+
+void QtProfileSelectionDialog::SelectUser(const QModelIndex& index) {
+    user_index = index.row();
+}
+
+QtProfileSelector::QtProfileSelector(GMainWindow& parent) {
+    connect(this, &QtProfileSelector::MainWindowSelectProfile, &parent,
+            &GMainWindow::ProfileSelectorSelectProfile, Qt::QueuedConnection);
+    connect(&parent, &GMainWindow::ProfileSelectorFinishedSelection, this,
+            &QtProfileSelector::MainWindowFinishedSelection, Qt::DirectConnection);
+}
+
+QtProfileSelector::~QtProfileSelector() = default;
+
+void QtProfileSelector::SelectProfile(
+    std::function<void(std::optional<Service::Account::UUID>)> callback) const {
+    this->callback = std::move(callback);
+    emit MainWindowSelectProfile();
+}
+
+void QtProfileSelector::MainWindowFinishedSelection(std::optional<Service::Account::UUID> uuid) {
+    // Acquire the HLE mutex
+    std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
+    callback(uuid);
+}

src/yuzu/applets/profile_select.h

@@ -0,0 +1,73 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <vector>
+#include <QDialog>
+#include <QList>
+#include "core/frontend/applets/profile_select.h"
+
+class GMainWindow;
+class QDialogButtonBox;
+class QGraphicsScene;
+class QLabel;
+class QScrollArea;
+class QStandardItem;
+class QStandardItemModel;
+class QTreeView;
+class QVBoxLayout;
+
+class QtProfileSelectionDialog final : public QDialog {
+    Q_OBJECT
+
+public:
+    explicit QtProfileSelectionDialog(QWidget* parent);
+    ~QtProfileSelectionDialog() override;
+
+    void accept() override;
+    void reject() override;
+
+    bool GetStatus() const;
+    u32 GetIndex() const;
+
+private:
+    bool ok = false;
+    u32 user_index = 0;
+
+    void SelectUser(const QModelIndex& index);
+
+    QVBoxLayout* layout;
+    QTreeView* tree_view;
+    QStandardItemModel* item_model;
+    QGraphicsScene* scene;
+
+    std::vector<QList<QStandardItem*>> list_items;
+
+    QVBoxLayout* outer_layout;
+    QLabel* instruction_label;
+    QScrollArea* scroll_area;
+    QDialogButtonBox* buttons;
+
+    std::unique_ptr<Service::Account::ProfileManager> profile_manager;
+};
+
+class QtProfileSelector final : public QObject, public Core::Frontend::ProfileSelectApplet {
+    Q_OBJECT
+
+public:
+    explicit QtProfileSelector(GMainWindow& parent);
+    ~QtProfileSelector() override;
+
+    void SelectProfile(
+        std::function<void(std::optional<Service::Account::UUID>)> callback) const override;
+
+signals:
+    void MainWindowSelectProfile() const;
+
+private:
+    void MainWindowFinishedSelection(std::optional<Service::Account::UUID> uuid);
+
+    mutable std::function<void(std::optional<Service::Account::UUID>)> callback;
+};

src/yuzu/bootmanager.cpp

@@ -14,6 +14,8 @@
 #include "input_common/keyboard.h"
 #include "input_common/main.h"
 #include "input_common/motion_emu.h"
+#include "video_core/renderer_base.h"
+#include "video_core/video_core.h"
 #include "yuzu/bootmanager.h"
 
 EmuThread::EmuThread(GRenderWindow* render_window) : render_window(render_window) {}
@@ -333,6 +335,22 @@ void GRenderWindow::InitRenderTarget() {
     BackupGeometry();
 }
 
+void GRenderWindow::CaptureScreenshot(u16 res_scale, const QString& screenshot_path) {
+    auto& renderer = Core::System::GetInstance().Renderer();
+
+    if (!res_scale)
+        res_scale = VideoCore::GetResolutionScaleFactor(renderer);
+
+    const Layout::FramebufferLayout layout{Layout::FrameLayoutFromResolutionScale(res_scale)};
+    screenshot_image = QImage(QSize(layout.width, layout.height), QImage::Format_RGB32);
+    renderer.RequestScreenshot(screenshot_image.bits(),
+                               [=] {
+                                   screenshot_image.mirrored(false, true).save(screenshot_path);
+                                   LOG_INFO(Frontend, "The screenshot is saved.");
+                               },
+                               layout);
+}
+
 void GRenderWindow::OnMinimalClientAreaChangeRequest(
     const std::pair<unsigned, unsigned>& minimal_size) {
     setMinimumSize(minimal_size.first, minimal_size.second);

src/yuzu/bootmanager.h

@@ -8,6 +8,7 @@
 #include <condition_variable>
 #include <mutex>
 #include <QGLWidget>
+#include <QImage>
 #include <QThread>
 #include "common/thread.h"
 #include "core/core.h"
@@ -139,6 +140,8 @@ public:
 
     void InitRenderTarget();
 
+    void CaptureScreenshot(u16 res_scale, const QString& screenshot_path);
+
 public slots:
     void moveContext(); // overridden
 
@@ -165,6 +168,9 @@ private:
 
     EmuThread* emu_thread;
 
+    /// Temporary storage of the screenshot taken
+    QImage screenshot_image;
+
 protected:
     void showEvent(QShowEvent* event) override;
 };