service/am: Remove unnecessary Skip calls

We can simplify these by wrapping the necessary members in structs and
then simply reading out the whole struct.

.ci/scripts/common/post-upload.sh

@@ -4,9 +4,13 @@
 cp license.txt "$DIR_NAME"
 cp README.md "$DIR_NAME"
 
+tar -cJvf "${REV_NAME}-source.tar.xz" src externals CMakeLists.txt README.md license.txt
+cp "${REV_NAME}-source.tar.xz" "$DIR_NAME"
+
 tar $COMPRESSION_FLAGS "$ARCHIVE_NAME" "$DIR_NAME"
 
 mv "$DIR_NAME" $RELEASE_NAME
+mv "${REV_NAME}-source.tar.xz" $RELEASE_NAME
 
 7z a "$REV_NAME.7z" $RELEASE_NAME
 

.ci/scripts/linux/upload.sh

@@ -7,9 +7,9 @@ ARCHIVE_NAME="${REV_NAME}.tar.xz"
 COMPRESSION_FLAGS="-cJvf"
 
 if [ "${RELEASE_NAME}" = "mainline" ]; then
-    DIR_NAME="${REV_NAME}_${RELEASE_NAME}"
-else
     DIR_NAME="${REV_NAME}"
+else
+    DIR_NAME="${REV_NAME}_${RELEASE_NAME}"
 fi
 
 mkdir "$DIR_NAME"

.ci/scripts/windows/upload.ps1

@@ -1,19 +1,23 @@
 param($BUILD_NAME)
 
-$GITDATE = $(git show -s --date=short --format='%ad') -replace "-",""
+$GITDATE = $(git show -s --date=short --format='%ad') -replace "-", ""
 $GITREV = $(git show -s --format='%h')
 
 if ("$BUILD_NAME" -eq "mainline") {
-    $RELEASE_DIST = "yuzu-windows-msvc-$BUILD_NAME"
-} else {
     $RELEASE_DIST = "yuzu-windows-msvc"
 }
+else {
+    $RELEASE_DIST = "yuzu-windows-msvc-$BUILD_NAME"
+}
 
 $MSVC_BUILD_ZIP = "yuzu-windows-msvc-$GITDATE-$GITREV.zip" -replace " ", ""
 $MSVC_BUILD_PDB = "yuzu-windows-msvc-$GITDATE-$GITREV-debugsymbols.zip" -replace " ", ""
 $MSVC_SEVENZIP = "yuzu-windows-msvc-$GITDATE-$GITREV.7z" -replace " ", ""
 $MSVC_TAR = "yuzu-windows-msvc-$GITDATE-$GITREV.tar" -replace " ", ""
 $MSVC_TARXZ = "yuzu-windows-msvc-$GITDATE-$GITREV.tar.xz" -replace " ", ""
+$MSVC_SOURCE = "yuzu-windows-msvc-source-$GITDATE-$GITREV" -replace " ", ""
+$MSVC_SOURCE_TAR = "$MSVC_SOURCE.tar"
+$MSVC_SOURCE_TARXZ = "$MSVC_SOURCE_TAR.xz"
 
 $env:BUILD_ZIP = $MSVC_BUILD_ZIP
 $env:BUILD_SYMBOLS = $MSVC_BUILD_PDB
@@ -21,19 +25,33 @@ $env:BUILD_UPDATE = $MSVC_SEVENZIP
 
 $BUILD_DIR = ".\build\bin\Release"
 
+# Upload debugging symbols
 mkdir pdb
 Get-ChildItem "$BUILD_DIR\" -Recurse -Filter "*.pdb" | Copy-Item -destination .\pdb
 7z a -tzip $MSVC_BUILD_PDB .\pdb\*.pdb
 rm "$BUILD_DIR\*.pdb"
+
+# Create artifact directories
 mkdir $RELEASE_DIST
+mkdir $MSVC_SOURCE
 mkdir "artifacts"
 
+# Build a tar.xz for the source of the release
+Copy-Item .\license.txt -Destination $MSVC_SOURCE
+Copy-Item .\README.md -Destination $MSVC_SOURCE
+Copy-Item .\src -Recurse -Destination $MSVC_SOURCE
+Copy-Item .\externals -Recurse -Destination $MSVC_SOURCE
+Copy-Item .\dist -Recurse -Destination $MSVC_SOURCE
+Copy-Item .\CMakeModules -Recurse -Destination $MSVC_SOURCE
+7z a -r -ttar $MSVC_SOURCE_TAR $MSVC_SOURCE
+7z a -r -txz $MSVC_SOURCE_TARXZ $MSVC_SOURCE_TAR
+
+# Build the final release artifacts
+Copy-Item $MSVC_SOURCE_TARXZ -Destination $RELEASE_DIST
 Copy-Item "$BUILD_DIR\*" -Destination $RELEASE_DIST -Recurse
 rm "$RELEASE_DIST\*.exe"
 Get-ChildItem "$BUILD_DIR" -Recurse -Filter "yuzu*.exe" | Copy-Item -destination $RELEASE_DIST
 Get-ChildItem "$BUILD_DIR" -Recurse -Filter "QtWebEngineProcess*.exe" | Copy-Item -destination $RELEASE_DIST
-Copy-Item .\license.txt -Destination $RELEASE_DIST
-Copy-Item .\README.md -Destination $RELEASE_DIST
 7z a -tzip $MSVC_BUILD_ZIP $RELEASE_DIST\*
 7z a $MSVC_SEVENZIP $RELEASE_DIST
 

.ci/scripts/windows/upload.sh

@@ -7,9 +7,9 @@ ARCHIVE_NAME="${REV_NAME}.tar.gz"
 COMPRESSION_FLAGS="-czvf"
 
 if [ "${RELEASE_NAME}" = "mainline" ]; then
-    DIR_NAME="${REV_NAME}_${RELEASE_NAME}"
-else
     DIR_NAME="${REV_NAME}"
+else
+    DIR_NAME="${REV_NAME}_${RELEASE_NAME}"
 fi
 
 mkdir "$DIR_NAME"

CONTRIBUTING.md

@@ -1 +1 @@
-**The Contributor's Guide has moved to [the Yuzu wiki](https://github.com/yuzu-emu/yuzu/wiki/Contributing).**
+**The Contributor's Guide has moved to [the yuzu wiki](https://github.com/yuzu-emu/yuzu/wiki/Contributing).**

externals/microprofile/microprofile.h

@@ -814,7 +814,7 @@ struct MicroProfile
 
 inline int MicroProfileLogType(MicroProfileLogEntry Index)
 {
-    return ((MP_LOG_BEGIN_MASK & Index)>>62) & 0x3;
+    return (int)(((MP_LOG_BEGIN_MASK & Index)>>62) & 0x3ULL);
 }
 
 inline uint64_t MicroProfileLogTimerIndex(MicroProfileLogEntry Index)
@@ -861,12 +861,12 @@ T MicroProfileMax(T a, T b)
 
 inline int64_t MicroProfileMsToTick(float fMs, int64_t nTicksPerSecond)
 {
-    return (int64_t)(fMs*0.001f*nTicksPerSecond);
+    return (int64_t)(fMs*0.001f*(float)nTicksPerSecond);
 }
 
 inline float MicroProfileTickToMsMultiplier(int64_t nTicksPerSecond)
 {
-    return 1000.f / nTicksPerSecond;
+    return 1000.f / (float)nTicksPerSecond;
 }
 
 inline uint16_t MicroProfileGetGroupIndex(MicroProfileToken t)

src/common/CMakeLists.txt

@@ -3,17 +3,8 @@
 # could affect the result, but much more unlikely than the following files. Keeping a list of files
 # like this allows for much better caching since it doesn't force the user to recompile binary shaders every update
 set(VIDEO_CORE "${CMAKE_SOURCE_DIR}/src/video_core")
-if (DEFINED ENV{CI})
-  if (DEFINED ENV{TRAVIS})
-    set(BUILD_REPOSITORY $ENV{TRAVIS_REPO_SLUG})
-    set(BUILD_TAG $ENV{TRAVIS_TAG})
-  elseif(DEFINED ENV{APPVEYOR})
-    set(BUILD_REPOSITORY $ENV{APPVEYOR_REPO_NAME})
-    set(BUILD_TAG $ENV{APPVEYOR_REPO_TAG_NAME})
-  elseif(DEFINED ENV{AZURE})
-    set(BUILD_REPOSITORY $ENV{AZURE_REPO_NAME})
-    set(BUILD_TAG $ENV{AZURE_REPO_TAG})
-  endif()
+if (DEFINED ENV{AZURECIREPO})
+  set(BUILD_REPOSITORY $ENV{AZURECIREPO})
 endif()
 if (DEFINED ENV{TITLEBARFORMATIDLE})
   set(TITLE_BAR_FORMAT_IDLE $ENV{TITLEBARFORMATIDLE})

src/common/common_funcs.h

@@ -56,7 +56,7 @@ std::string GetLastErrorMsg();
 namespace Common {
 
 constexpr u32 MakeMagic(char a, char b, char c, char d) {
-    return a | b << 8 | c << 16 | d << 24;
+    return u32(a) | u32(b) << 8 | u32(c) << 16 | u32(d) << 24;
 }
 
 } // namespace Common

src/common/hash.h

@@ -35,41 +35,6 @@ static inline u64 ComputeStructHash64(const T& data) {
     return ComputeHash64(&data, sizeof(data));
 }
 
-/// A helper template that ensures the padding in a struct is initialized by memsetting to 0.
-template <typename T>
-struct HashableStruct {
-    // In addition to being trivially copyable, T must also have a trivial default constructor,
-    // because any member initialization would be overridden by memset
-    static_assert(std::is_trivial_v<T>, "Type passed to HashableStruct must be trivial");
-    /*
-     * We use a union because "implicitly-defined copy/move constructor for a union X copies the
-     * object representation of X." and "implicitly-defined copy assignment operator for a union X
-     * copies the object representation (3.9) of X." = Bytewise copy instead of memberwise copy.
-     * This is important because the padding bytes are included in the hash and comparison between
-     * objects.
-     */
-    union {
-        T state;
-    };
-
-    HashableStruct() {
-        // Memset structure to zero padding bits, so that they will be deterministic when hashing
-        std::memset(&state, 0, sizeof(T));
-    }
-
-    bool operator==(const HashableStruct<T>& o) const {
-        return std::memcmp(&state, &o.state, sizeof(T)) == 0;
-    };
-
-    bool operator!=(const HashableStruct<T>& o) const {
-        return !(*this == o);
-    };
-
-    std::size_t Hash() const {
-        return Common::ComputeStructHash64(state);
-    }
-};
-
 struct PairHash {
     template <class T1, class T2>
     std::size_t operator()(const std::pair<T1, T2>& pair) const noexcept {

src/core/crypto/key_manager.cpp

@@ -396,7 +396,7 @@ static std::array<u8, target_size> MGF1(const std::array<u8, in_size>& seed) {
     while (out.size() < target_size) {
         out.resize(out.size() + 0x20);
         seed_exp[in_size + 3] = static_cast<u8>(i);
-        mbedtls_sha256(seed_exp.data(), seed_exp.size(), out.data() + out.size() - 0x20, 0);
+        mbedtls_sha256_ret(seed_exp.data(), seed_exp.size(), out.data() + out.size() - 0x20, 0);
         ++i;
     }
 
@@ -668,23 +668,27 @@ void KeyManager::WriteKeyToFile(KeyCategory category, std::string_view keyname,
                                 const std::array<u8, Size>& key) {
     const std::string yuzu_keys_dir = FileUtil::GetUserPath(FileUtil::UserPath::KeysDir);
     std::string filename = "title.keys_autogenerated";
-    if (category == KeyCategory::Standard)
+    if (category == KeyCategory::Standard) {
         filename = dev_mode ? "dev.keys_autogenerated" : "prod.keys_autogenerated";
-    else if (category == KeyCategory::Console)
+    } else if (category == KeyCategory::Console) {
         filename = "console.keys_autogenerated";
-    const auto add_info_text = !FileUtil::Exists(yuzu_keys_dir + DIR_SEP + filename);
-    FileUtil::CreateFullPath(yuzu_keys_dir + DIR_SEP + filename);
-    std::ofstream file(yuzu_keys_dir + DIR_SEP + filename, std::ios::app);
-    if (!file.is_open())
-        return;
-    if (add_info_text) {
-        file
-            << "# This file is autogenerated by Yuzu\n"
-            << "# It serves to store keys that were automatically generated from the normal keys\n"
-            << "# If you are experiencing issues involving keys, it may help to delete this file\n";
     }
 
-    file << fmt::format("\n{} = {}", keyname, Common::HexToString(key));
+    const auto path = yuzu_keys_dir + DIR_SEP + filename;
+    const auto add_info_text = !FileUtil::Exists(path);
+    FileUtil::CreateFullPath(path);
+    FileUtil::IOFile file{path, "a"};
+    if (!file.IsOpen()) {
+        return;
+    }
+    if (add_info_text) {
+        file.WriteString(
+            "# This file is autogenerated by Yuzu\n"
+            "# It serves to store keys that were automatically generated from the normal keys\n"
+            "# If you are experiencing issues involving keys, it may help to delete this file\n");
+    }
+
+    file.WriteString(fmt::format("\n{} = {}", keyname, Common::HexToString(key)));
     AttemptLoadKeyFile(yuzu_keys_dir, yuzu_keys_dir, filename, category == KeyCategory::Title);
 }
 

src/core/crypto/partition_data_manager.cpp

@@ -161,7 +161,7 @@ std::array<u8, key_size> FindKeyFromHex(const std::vector<u8>& binary,
 
     std::array<u8, 0x20> temp{};
     for (size_t i = 0; i < binary.size() - key_size; ++i) {
-        mbedtls_sha256(binary.data() + i, key_size, temp.data(), 0);
+        mbedtls_sha256_ret(binary.data() + i, key_size, temp.data(), 0);
 
         if (temp != hash)
             continue;
@@ -189,7 +189,7 @@ static std::array<Key128, 0x20> FindEncryptedMasterKeyFromHex(const std::vector<
     AESCipher<Key128> cipher(key, Mode::ECB);
     for (size_t i = 0; i < binary.size() - 0x10; ++i) {
         cipher.Transcode(binary.data() + i, dec_temp.size(), dec_temp.data(), Op::Decrypt);
-        mbedtls_sha256(dec_temp.data(), dec_temp.size(), temp.data(), 0);
+        mbedtls_sha256_ret(dec_temp.data(), dec_temp.size(), temp.data(), 0);
 
         for (size_t k = 0; k < out.size(); ++k) {
             if (temp == master_key_hashes[k]) {

src/core/file_sys/registered_cache.cpp

@@ -62,7 +62,7 @@ static std::string GetRelativePathFromNcaID(const std::array<u8, 16>& nca_id, bo
                            Common::HexToString(nca_id, second_hex_upper));
 
     Core::Crypto::SHA256Hash hash{};
-    mbedtls_sha256(nca_id.data(), nca_id.size(), hash.data(), 0);
+    mbedtls_sha256_ret(nca_id.data(), nca_id.size(), hash.data(), 0);
     return fmt::format(cnmt_suffix ? "/000000{:02X}/{}.cnmt.nca" : "/000000{:02X}/{}.nca", hash[0],
                        Common::HexToString(nca_id, second_hex_upper));
 }
@@ -141,7 +141,7 @@ bool PlaceholderCache::Create(const NcaID& id, u64 size) const {
     }
 
     Core::Crypto::SHA256Hash hash{};
-    mbedtls_sha256(id.data(), id.size(), hash.data(), 0);
+    mbedtls_sha256_ret(id.data(), id.size(), hash.data(), 0);
     const auto dirname = fmt::format("000000{:02X}", hash[0]);
 
     const auto dir2 = GetOrCreateDirectoryRelative(dir, dirname);
@@ -165,7 +165,7 @@ bool PlaceholderCache::Delete(const NcaID& id) const {
     }
 
     Core::Crypto::SHA256Hash hash{};
-    mbedtls_sha256(id.data(), id.size(), hash.data(), 0);
+    mbedtls_sha256_ret(id.data(), id.size(), hash.data(), 0);
     const auto dirname = fmt::format("000000{:02X}", hash[0]);
 
     const auto dir2 = GetOrCreateDirectoryRelative(dir, dirname);
@@ -603,7 +603,7 @@ InstallResult RegisteredCache::InstallEntry(const NCA& nca, TitleType type,
     OptionalHeader opt_header{0, 0};
     ContentRecord c_rec{{}, {}, {}, GetCRTypeFromNCAType(nca.GetType()), {}};
     const auto& data = nca.GetBaseFile()->ReadBytes(0x100000);
-    mbedtls_sha256(data.data(), data.size(), c_rec.hash.data(), 0);
+    mbedtls_sha256_ret(data.data(), data.size(), c_rec.hash.data(), 0);
     memcpy(&c_rec.nca_id, &c_rec.hash, 16);
     const CNMT new_cnmt(header, opt_header, {c_rec}, {});
     if (!RawInstallYuzuMeta(new_cnmt))
@@ -626,7 +626,7 @@ InstallResult RegisteredCache::RawInstallNCA(const NCA& nca, const VfsCopyFuncti
         id = *override_id;
     } else {
         const auto& data = in->ReadBytes(0x100000);
-        mbedtls_sha256(data.data(), data.size(), hash.data(), 0);
+        mbedtls_sha256_ret(data.data(), data.size(), hash.data(), 0);
         memcpy(id.data(), hash.data(), 16);
     }
 

src/core/file_sys/xts_archive.cpp

@@ -64,7 +64,7 @@ NAX::NAX(VirtualFile file_) : header(std::make_unique<NAXHeader>()), file(std::m
 NAX::NAX(VirtualFile file_, std::array<u8, 0x10> nca_id)
     : header(std::make_unique<NAXHeader>()), file(std::move(file_)) {
     Core::Crypto::SHA256Hash hash{};
-    mbedtls_sha256(nca_id.data(), nca_id.size(), hash.data(), 0);
+    mbedtls_sha256_ret(nca_id.data(), nca_id.size(), hash.data(), 0);
     status = Parse(fmt::format("/registered/000000{:02X}/{}.nca", hash[0],
                                Common::HexToString(nca_id, false)));
 }
@@ -93,8 +93,7 @@ Loader::ResultStatus NAX::Parse(std::string_view path) {
     std::size_t i = 0;
     for (; i < sd_keys.size(); ++i) {
         std::array<Core::Crypto::Key128, 2> nax_keys{};
-        if (!CalculateHMAC256(nax_keys.data(), sd_keys[i].data(), 0x10, std::string(path).c_str(),
-                              path.size())) {
+        if (!CalculateHMAC256(nax_keys.data(), sd_keys[i].data(), 0x10, path.data(), path.size())) {
             return Loader::ResultStatus::ErrorNAXKeyHMACFailed;
         }
 

src/core/hle/kernel/hle_ipc.cpp

@@ -58,8 +58,7 @@ SharedPtr<WritableEvent> HLERequestContext::SleepClientThread(
     auto& kernel = Core::System::GetInstance().Kernel();
     if (!writable_event) {
         // Create event if not provided
-        const auto pair = WritableEvent::CreateEventPair(kernel, ResetType::Automatic,
-                                                         "HLE Pause Event: " + reason);
+        const auto pair = WritableEvent::CreateEventPair(kernel, "HLE Pause Event: " + reason);
         writable_event = pair.writable;
     }
 

src/core/hle/kernel/object.h

@@ -32,11 +32,6 @@ enum class HandleType : u32 {
     ServerSession,
 };
 
-enum class ResetType {
-    Automatic, ///< Reset automatically on object acquisition
-    Manual,    ///< Never reset automatically
-};
-
 class Object : NonCopyable {
 public:
     explicit Object(KernelCore& kernel);

src/core/hle/kernel/readable_event.cpp

@@ -20,15 +20,13 @@ bool ReadableEvent::ShouldWait(const Thread* thread) const {
 
 void ReadableEvent::Acquire(Thread* thread) {
     ASSERT_MSG(!ShouldWait(thread), "object unavailable!");
-
-    if (reset_type == ResetType::Automatic) {
-        signaled = false;
-    }
 }
 
 void ReadableEvent::Signal() {
-    signaled = true;
-    WakeupAllWaitingThreads();
+    if (!signaled) {
+        signaled = true;
+        WakeupAllWaitingThreads();
+    };
 }
 
 void ReadableEvent::Clear() {

src/core/hle/kernel/readable_event.h

@@ -27,10 +27,6 @@ public:
         return name;
     }
 
-    ResetType GetResetType() const {
-        return reset_type;
-    }
-
     static constexpr HandleType HANDLE_TYPE = HandleType::ReadableEvent;
     HandleType GetHandleType() const override {
         return HANDLE_TYPE;
@@ -55,8 +51,7 @@ private:
 
     void Signal();
 
-    ResetType reset_type;
-    bool signaled;
+    bool signaled{};
 
     std::string name; ///< Name of event (optional)
 };

src/core/hle/kernel/svc.cpp

@@ -2099,7 +2099,7 @@ static ResultCode CreateEvent(Core::System& system, Handle* write_handle, Handle
 
     auto& kernel = system.Kernel();
     const auto [readable_event, writable_event] =
-        WritableEvent::CreateEventPair(kernel, ResetType::Manual, "CreateEvent");
+        WritableEvent::CreateEventPair(kernel, "CreateEvent");
 
     HandleTable& handle_table = kernel.CurrentProcess()->GetHandleTable();
 

src/core/hle/kernel/writable_event.cpp

@@ -15,8 +15,7 @@ namespace Kernel {
 WritableEvent::WritableEvent(KernelCore& kernel) : Object{kernel} {}
 WritableEvent::~WritableEvent() = default;
 
-EventPair WritableEvent::CreateEventPair(KernelCore& kernel, ResetType reset_type,
-                                         std::string name) {
+EventPair WritableEvent::CreateEventPair(KernelCore& kernel, std::string name) {
     SharedPtr<WritableEvent> writable_event(new WritableEvent(kernel));
     SharedPtr<ReadableEvent> readable_event(new ReadableEvent(kernel));
 
@@ -24,7 +23,6 @@ EventPair WritableEvent::CreateEventPair(KernelCore& kernel, ResetType reset_typ
     writable_event->readable = readable_event;
     readable_event->name = name + ":Readable";
     readable_event->signaled = false;
-    readable_event->reset_type = reset_type;
 
     return {std::move(readable_event), std::move(writable_event)};
 }
@@ -33,10 +31,6 @@ SharedPtr<ReadableEvent> WritableEvent::GetReadableEvent() const {
     return readable;
 }
 
-ResetType WritableEvent::GetResetType() const {
-    return readable->reset_type;
-}
-
 void WritableEvent::Signal() {
     readable->Signal();
 }

src/core/hle/kernel/writable_event.h

@@ -24,11 +24,9 @@ public:
     /**
      * Creates an event
      * @param kernel The kernel instance to create this event under.
-     * @param reset_type ResetType describing how to create event
      * @param name Optional name of event
      */
-    static EventPair CreateEventPair(KernelCore& kernel, ResetType reset_type,
-                                     std::string name = "Unknown");
+    static EventPair CreateEventPair(KernelCore& kernel, std::string name = "Unknown");
 
     std::string GetTypeName() const override {
         return "WritableEvent";
@@ -44,8 +42,6 @@ public:
 
     SharedPtr<ReadableEvent> GetReadableEvent() const;
 
-    ResetType GetResetType() const;
-
     void Signal();
     void Clear();
     bool IsSignaled() const;

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

@@ -289,8 +289,8 @@ ISelfController::ISelfController(Core::System& system,
     RegisterHandlers(functions);
 
     auto& kernel = system.Kernel();
-    launchable_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                              "ISelfController:LaunchableEvent");
+    launchable_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, "ISelfController:LaunchableEvent");
 
     // This event is created by AM on the first time GetAccumulatedSuspendedTickChangedEvent() is
     // called. Yuzu can just create it unconditionally, since it doesn't need to support multiple
@@ -298,7 +298,7 @@ ISelfController::ISelfController(Core::System& system,
     // suspended if the event has previously been created by a call to
     // GetAccumulatedSuspendedTickChangedEvent.
     accumulated_suspended_tick_changed_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "ISelfController:AccumulatedSuspendedTickChangedEvent");
+        kernel, "ISelfController:AccumulatedSuspendedTickChangedEvent");
     accumulated_suspended_tick_changed_event.writable->Signal();
 }
 
@@ -523,10 +523,10 @@ void ISelfController::GetAccumulatedSuspendedTickChangedEvent(Kernel::HLERequest
 }
 
 AppletMessageQueue::AppletMessageQueue(Kernel::KernelCore& kernel) {
-    on_new_message = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                            "AMMessageQueue:OnMessageRecieved");
-    on_operation_mode_changed = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Automatic, "AMMessageQueue:OperationModeChanged");
+    on_new_message =
+        Kernel::WritableEvent::CreateEventPair(kernel, "AMMessageQueue:OnMessageRecieved");
+    on_operation_mode_changed =
+        Kernel::WritableEvent::CreateEventPair(kernel, "AMMessageQueue:OperationModeChanged");
 }
 
 AppletMessageQueue::~AppletMessageQueue() = default;
@@ -1077,6 +1077,7 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_)
         {101, &IApplicationFunctions::SetApplicationCopyrightImage, "SetApplicationCopyrightImage"},
         {102, &IApplicationFunctions::SetApplicationCopyrightVisibility, "SetApplicationCopyrightVisibility"},
         {110, nullptr, "QueryApplicationPlayStatistics"},
+        {111, &IApplicationFunctions::QueryApplicationPlayStatisticsByUid, "QueryApplicationPlayStatisticsByUid"},
         {120, nullptr, "ExecuteProgram"},
         {121, nullptr, "ClearUserChannel"},
         {122, nullptr, "UnpopToUserChannel"},
@@ -1091,7 +1092,7 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_)
 
     auto& kernel = system.Kernel();
     gpu_error_detected_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "IApplicationFunctions:GpuErrorDetectedSystemEvent");
+        kernel, "IApplicationFunctions:GpuErrorDetectedSystemEvent");
 }
 
 IApplicationFunctions::~IApplicationFunctions() = default;
@@ -1335,12 +1336,16 @@ void IApplicationFunctions::GetPseudoDeviceId(Kernel::HLERequestContext& ctx) {
 }
 
 void IApplicationFunctions::ExtendSaveData(Kernel::HLERequestContext& ctx) {
+    struct Parameters {
+        FileSys::SaveDataType type;
+        u128 user_id;
+        u64 new_normal_size;
+        u64 new_journal_size;
+    };
+    static_assert(sizeof(Parameters) == 40);
+
     IPC::RequestParser rp{ctx};
-    const auto type{rp.PopRaw<FileSys::SaveDataType>()};
-    rp.Skip(1, false);
-    const auto user_id{rp.PopRaw<u128>()};
-    const auto new_normal_size{rp.PopRaw<u64>()};
-    const auto new_journal_size{rp.PopRaw<u64>()};
+    const auto [type, user_id, new_normal_size, new_journal_size] = rp.PopRaw<Parameters>();
 
     LOG_DEBUG(Service_AM,
               "called with type={:02X}, user_id={:016X}{:016X}, new_normal={:016X}, "
@@ -1359,10 +1364,14 @@ void IApplicationFunctions::ExtendSaveData(Kernel::HLERequestContext& ctx) {
 }
 
 void IApplicationFunctions::GetSaveDataSize(Kernel::HLERequestContext& ctx) {
+    struct Parameters {
+        FileSys::SaveDataType type;
+        u128 user_id;
+    };
+    static_assert(sizeof(Parameters) == 24);
+
     IPC::RequestParser rp{ctx};
-    const auto type{rp.PopRaw<FileSys::SaveDataType>()};
-    rp.Skip(1, false);
-    const auto user_id{rp.PopRaw<u128>()};
+    const auto [type, user_id] = rp.PopRaw<Parameters>();
 
     LOG_DEBUG(Service_AM, "called with type={:02X}, user_id={:016X}{:016X}", static_cast<u8>(type),
               user_id[1], user_id[0]);
@@ -1384,6 +1393,14 @@ void IApplicationFunctions::GetGpuErrorDetectedSystemEvent(Kernel::HLERequestCon
     rb.PushCopyObjects(gpu_error_detected_event.readable);
 }
 
+void IApplicationFunctions::QueryApplicationPlayStatisticsByUid(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push<u32>(0);
+}
+
 void InstallInterfaces(SM::ServiceManager& service_manager,
                        std::shared_ptr<NVFlinger::NVFlinger> nvflinger, Core::System& system) {
     auto message_queue = std::make_shared<AppletMessageQueue>(system.Kernel());

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

@@ -256,6 +256,7 @@ private:
     void SetApplicationCopyrightImage(Kernel::HLERequestContext& ctx);
     void SetApplicationCopyrightVisibility(Kernel::HLERequestContext& ctx);
     void GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx);
+    void QueryApplicationPlayStatisticsByUid(Kernel::HLERequestContext& ctx);
 
     bool launch_popped_application_specific = false;
     bool launch_popped_account_preselect = false;

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

@@ -24,12 +24,12 @@
 namespace Service::AM::Applets {
 
 AppletDataBroker::AppletDataBroker(Kernel::KernelCore& kernel) {
-    state_changed_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "ILibraryAppletAccessor:StateChangedEvent");
-    pop_out_data_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "ILibraryAppletAccessor:PopDataOutEvent");
+    state_changed_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, "ILibraryAppletAccessor:StateChangedEvent");
+    pop_out_data_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, "ILibraryAppletAccessor:PopDataOutEvent");
     pop_interactive_out_data_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "ILibraryAppletAccessor:PopInteractiveDataOutEvent");
+        kernel, "ILibraryAppletAccessor:PopInteractiveDataOutEvent");
 }
 
 AppletDataBroker::~AppletDataBroker() = default;

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

@@ -67,8 +67,8 @@ AOC_U::AOC_U(Core::System& system)
     RegisterHandlers(functions);
 
     auto& kernel = system.Kernel();
-    aoc_change_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                              "GetAddOnContentListChanged:Event");
+    aoc_change_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, "GetAddOnContentListChanged:Event");
 }
 
 AOC_U::~AOC_U() = default;

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

@@ -65,8 +65,8 @@ public:
         RegisterHandlers(functions);
 
         // This is the event handle used to check if the audio buffer was released
-        buffer_event = Kernel::WritableEvent::CreateEventPair(
-            system.Kernel(), Kernel::ResetType::Manual, "IAudioOutBufferReleased");
+        buffer_event =
+            Kernel::WritableEvent::CreateEventPair(system.Kernel(), "IAudioOutBufferReleased");
 
         stream = audio_core.OpenStream(system.CoreTiming(), audio_params.sample_rate,
                                        audio_params.channel_count, std::move(unique_name),

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

@@ -47,8 +47,8 @@ public:
         // clang-format on
         RegisterHandlers(functions);
 
-        system_event = Kernel::WritableEvent::CreateEventPair(
-            system.Kernel(), Kernel::ResetType::Manual, "IAudioRenderer:SystemEvent");
+        system_event =
+            Kernel::WritableEvent::CreateEventPair(system.Kernel(), "IAudioRenderer:SystemEvent");
         renderer = std::make_unique<AudioCore::AudioRenderer>(
             system.CoreTiming(), audren_params, system_event.writable, instance_number);
     }
@@ -180,17 +180,17 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        buffer_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                              "IAudioOutBufferReleasedEvent");
+        buffer_event =
+            Kernel::WritableEvent::CreateEventPair(kernel, "IAudioOutBufferReleasedEvent");
 
         // Should be similar to audio_output_device_switch_event
         audio_input_device_switch_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IAudioDevice:AudioInputDeviceSwitchedEvent");
+            kernel, "IAudioDevice:AudioInputDeviceSwitchedEvent");
 
         // Should only be signalled when an audio output device has been changed, example: speaker
         // to headset
         audio_output_device_switch_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IAudioDevice:AudioOutputDeviceSwitchedEvent");
+            kernel, "IAudioDevice:AudioOutputDeviceSwitchedEvent");
     }
 
 private:

src/core/hle/service/bcat/backend/backend.cpp

@@ -13,8 +13,7 @@ namespace Service::BCAT {
 ProgressServiceBackend::ProgressServiceBackend(Kernel::KernelCore& kernel,
                                                std::string_view event_name) {
     event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Automatic,
-        std::string("ProgressServiceBackend:UpdateEvent:").append(event_name));
+        kernel, std::string("ProgressServiceBackend:UpdateEvent:").append(event_name));
 }
 
 Kernel::SharedPtr<Kernel::ReadableEvent> ProgressServiceBackend::GetEvent() const {

src/core/hle/service/bcat/backend/boxcat.cpp

@@ -255,7 +255,7 @@ private:
     using Digest = std::array<u8, 0x20>;
     static Digest DigestFile(std::vector<u8> bytes) {
         Digest out{};
-        mbedtls_sha256(bytes.data(), bytes.size(), out.data(), 0);
+        mbedtls_sha256_ret(bytes.data(), bytes.size(), out.data(), 0);
         return out;
     }
 

src/core/hle/service/bcat/module.cpp

@@ -46,7 +46,7 @@ u64 GetCurrentBuildID(const Core::System::CurrentBuildProcessID& id) {
 BCATDigest DigestFile(const FileSys::VirtualFile& file) {
     BCATDigest out{};
     const auto bytes = file->ReadAllBytes();
-    mbedtls_md5(bytes.data(), bytes.size(), out.data());
+    mbedtls_md5_ret(bytes.data(), bytes.size(), out.data());
     return out;
 }
 

src/core/hle/service/btdrv/btdrv.cpp

@@ -34,8 +34,7 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        register_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "BT:RegisterEvent");
+        register_event = Kernel::WritableEvent::CreateEventPair(kernel, "BT:RegisterEvent");
     }
 
 private:

src/core/hle/service/btm/btm.cpp

@@ -57,14 +57,12 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        scan_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                            "IBtmUserCore:ScanEvent");
-        connection_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IBtmUserCore:ConnectionEvent");
-        service_discovery = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IBtmUserCore:Discovery");
-        config_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                              "IBtmUserCore:ConfigEvent");
+        scan_event = Kernel::WritableEvent::CreateEventPair(kernel, "IBtmUserCore:ScanEvent");
+        connection_event =
+            Kernel::WritableEvent::CreateEventPair(kernel, "IBtmUserCore:ConnectionEvent");
+        service_discovery =
+            Kernel::WritableEvent::CreateEventPair(kernel, "IBtmUserCore:Discovery");
+        config_event = Kernel::WritableEvent::CreateEventPair(kernel, "IBtmUserCore:ConfigEvent");
     }
 
 private:

src/core/hle/service/friend/friend.cpp

@@ -162,7 +162,7 @@ public:
         RegisterHandlers(functions);
 
         notification_event = Kernel::WritableEvent::CreateEventPair(
-            system.Kernel(), Kernel::ResetType::Manual, "INotificationService:NotifyEvent");
+            system.Kernel(), "INotificationService:NotifyEvent");
     }
 
 private:

src/core/hle/service/hid/controllers/npad.cpp

@@ -174,7 +174,7 @@ void Controller_NPad::OnInit() {
     auto& kernel = system.Kernel();
     for (std::size_t i = 0; i < styleset_changed_events.size(); i++) {
         styleset_changed_events[i] = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Manual, fmt::format("npad:NpadStyleSetChanged_{}", i));
+            kernel, fmt::format("npad:NpadStyleSetChanged_{}", i));
     }
 
     if (!IsControllerActivated()) {

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

@@ -294,7 +294,7 @@ public:
         Memory::ReadBlock(nro_address, nro_data.data(), nro_size);
 
         SHA256Hash hash{};
-        mbedtls_sha256(nro_data.data(), nro_data.size(), hash.data(), 0);
+        mbedtls_sha256_ret(nro_data.data(), nro_data.size(), hash.data(), 0);
 
         // NRO Hash is already loaded
         if (std::any_of(nro.begin(), nro.end(), [&hash](const std::pair<VAddr, NROInfo>& info) {

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

@@ -26,8 +26,7 @@ constexpr ResultCode ERR_NO_APPLICATION_AREA(ErrorModule::NFP, 152);
 Module::Interface::Interface(std::shared_ptr<Module> module, Core::System& system, const char* name)
     : ServiceFramework(name), module(std::move(module)), system(system) {
     auto& kernel = system.Kernel();
-    nfc_tag_load = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                          "IUser:NFCTagDetected");
+    nfc_tag_load = Kernel::WritableEvent::CreateEventPair(kernel, "IUser:NFCTagDetected");
 }
 
 Module::Interface::~Interface() = default;
@@ -66,10 +65,9 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        deactivate_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IUser:DeactivateEvent");
-        availability_change_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IUser:AvailabilityChangeEvent");
+        deactivate_event = Kernel::WritableEvent::CreateEventPair(kernel, "IUser:DeactivateEvent");
+        availability_change_event =
+            Kernel::WritableEvent::CreateEventPair(kernel, "IUser:AvailabilityChangeEvent");
     }
 
 private:

src/core/hle/service/nifm/nifm.cpp

@@ -69,10 +69,8 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        event1 = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                        "IRequest:Event1");
-        event2 = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                        "IRequest:Event2");
+        event1 = Kernel::WritableEvent::CreateEventPair(kernel, "IRequest:Event1");
+        event2 = Kernel::WritableEvent::CreateEventPair(kernel, "IRequest:Event2");
     }
 
 private:

src/core/hle/service/nim/nim.cpp

@@ -141,8 +141,7 @@ public:
 
         auto& kernel = system.Kernel();
         finished_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic,
-            "IEnsureNetworkClockAvailabilityService:FinishEvent");
+            kernel, "IEnsureNetworkClockAvailabilityService:FinishEvent");
     }
 
 private:

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

@@ -40,8 +40,7 @@ Module::Module(Core::System& system) {
     auto& kernel = system.Kernel();
     for (u32 i = 0; i < MaxNvEvents; i++) {
         std::string event_label = fmt::format("NVDRV::NvEvent_{}", i);
-        events_interface.events[i] =
-            Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual, event_label);
+        events_interface.events[i] = Kernel::WritableEvent::CreateEventPair(kernel, event_label);
         events_interface.status[i] = EventState::Free;
         events_interface.registered[i] = false;
     }

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

@@ -16,8 +16,7 @@ namespace Service::NVFlinger {
 
 BufferQueue::BufferQueue(Kernel::KernelCore& kernel, u32 id, u64 layer_id)
     : id(id), layer_id(layer_id) {
-    buffer_wait_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                               "BufferQueue NativeHandle");
+    buffer_wait_event = Kernel::WritableEvent::CreateEventPair(kernel, "BufferQueue NativeHandle");
 }
 
 BufferQueue::~BufferQueue() = default;

src/core/hle/service/vi/display/vi_display.cpp

@@ -17,8 +17,8 @@ namespace Service::VI {
 
 Display::Display(u64 id, std::string name, Core::System& system) : id{id}, name{std::move(name)} {
     auto& kernel = system.Kernel();
-    vsync_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                         fmt::format("Display VSync Event {}", id));
+    vsync_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, fmt::format("Display VSync Event {}", id));
 }
 
 Display::~Display() = default;

src/video_core/CMakeLists.txt

@@ -180,3 +180,9 @@ target_link_libraries(video_core PRIVATE glad)
 if (ENABLE_VULKAN)
     target_link_libraries(video_core PRIVATE sirit)
 endif()
+
+if (MSVC)
+    target_compile_options(video_core PRIVATE /we4267)
+else()
+    target_compile_options(video_core PRIVATE -Werror=conversion -Wno-error=sign-conversion)
+endif()

src/video_core/buffer_cache/buffer_cache.h

@@ -12,6 +12,10 @@
 #include <utility>
 #include <vector>
 
+#include <boost/icl/interval_map.hpp>
+#include <boost/icl/interval_set.hpp>
+#include <boost/range/iterator_range.hpp>
+
 #include "common/alignment.h"
 #include "common/common_types.h"
 #include "core/core.h"
@@ -30,7 +34,7 @@ public:
     using BufferInfo = std::pair<const TBufferType*, u64>;
 
     BufferInfo UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4,
-                            bool is_written = false) {
+                            bool is_written = false, bool use_fast_cbuf = false) {
         std::lock_guard lock{mutex};
 
         auto& memory_manager = system.GPU().MemoryManager();
@@ -43,9 +47,13 @@ public:
         // Cache management is a big overhead, so only cache entries with a given size.
         // TODO: Figure out which size is the best for given games.
         constexpr std::size_t max_stream_size = 0x800;
-        if (size < max_stream_size) {
+        if (use_fast_cbuf || size < max_stream_size) {
             if (!is_written && !IsRegionWritten(cache_addr, cache_addr + size - 1)) {
-                return StreamBufferUpload(host_ptr, size, alignment);
+                if (use_fast_cbuf) {
+                    return ConstBufferUpload(host_ptr, size);
+                } else {
+                    return StreamBufferUpload(host_ptr, size, alignment);
+                }
             }
         }
 
@@ -152,6 +160,10 @@ protected:
     virtual void CopyBlock(const TBuffer& src, const TBuffer& dst, std::size_t src_offset,
                            std::size_t dst_offset, std::size_t size) = 0;
 
+    virtual BufferInfo ConstBufferUpload(const void* raw_pointer, std::size_t size) {
+        return {};
+    }
+
     /// Register an object into the cache
     void Register(const MapInterval& new_map, bool inherit_written = false) {
         const CacheAddr cache_ptr = new_map->GetStart();

src/video_core/engines/maxwell_3d.cpp

@@ -261,7 +261,8 @@ void Maxwell3D::CallMacroMethod(u32 method, std::size_t num_parameters, const u3
     executing_macro = 0;
 
     // Lookup the macro offset
-    const u32 entry = ((method - MacroRegistersStart) >> 1) % macro_positions.size();
+    const u32 entry =
+        ((method - MacroRegistersStart) >> 1) % static_cast<u32>(macro_positions.size());
 
     // Execute the current macro.
     macro_interpreter.Execute(macro_positions[entry], num_parameters, parameters);

src/video_core/engines/shader_bytecode.h

@@ -615,6 +615,14 @@ union Instruction {
         BitField<34, 13, u64> mask_imm;
     } shfl;
 
+    union {
+        BitField<44, 1, u64> ftz;
+        BitField<39, 2, u64> tab5cb8_2;
+        BitField<38, 1, u64> ndv;
+        BitField<47, 1, u64> cc;
+        BitField<28, 8, u64> swizzle;
+    } fswzadd;
+
     union {
         BitField<8, 8, Register> gpr;
         BitField<20, 24, s64> offset;
@@ -1478,7 +1486,8 @@ union Instruction {
             u32 value = static_cast<u32>(target);
             // The branch offset is relative to the next instruction and is stored in bytes, so
             // divide it by the size of an instruction and add 1 to it.
-            return static_cast<s32>((value ^ mask) - mask) / sizeof(Instruction) + 1;
+            return static_cast<s32>((value ^ mask) - mask) / static_cast<s32>(sizeof(Instruction)) +
+                   1;
         }
     } bra;
 
@@ -1492,7 +1501,8 @@ union Instruction {
             u32 value = static_cast<u32>(target);
             // The branch offset is relative to the next instruction and is stored in bytes, so
             // divide it by the size of an instruction and add 1 to it.
-            return static_cast<s32>((value ^ mask) - mask) / sizeof(Instruction) + 1;
+            return static_cast<s32>((value ^ mask) - mask) / static_cast<s32>(sizeof(Instruction)) +
+                   1;
         }
     } brx;
 
@@ -1590,6 +1600,7 @@ public:
         DEPBAR,
         VOTE,
         SHFL,
+        FSWZADD,
         BFE_C,
         BFE_R,
         BFE_IMM,
@@ -1851,11 +1862,11 @@ private:
                 const std::size_t bit_position = opcode_bitsize - i - 1;
                 switch (bitstring[i]) {
                 case '0':
-                    mask |= 1 << bit_position;
+                    mask |= static_cast<u16>(1U << bit_position);
                     break;
                 case '1':
-                    expect |= 1 << bit_position;
-                    mask |= 1 << bit_position;
+                    expect |= static_cast<u16>(1U << bit_position);
+                    mask |= static_cast<u16>(1U << bit_position);
                     break;
                 default:
                     // Ignore
@@ -1888,6 +1899,7 @@ private:
             INST("1111000011110---", Id::DEPBAR, Type::Synch, "DEPBAR"),
             INST("0101000011011---", Id::VOTE, Type::Warp, "VOTE"),
             INST("1110111100010---", Id::SHFL, Type::Warp, "SHFL"),
+            INST("0101000011111---", Id::FSWZADD, Type::Warp, "FSWZADD"),
             INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
             INST("1110111101001---", Id::LD_S, Type::Memory, "LD_S"),
             INST("1110111101000---", Id::LD_L, Type::Memory, "LD_L"),

src/video_core/renderer_opengl/gl_buffer_cache.cpp

@@ -8,13 +8,17 @@
 
 #include "common/assert.h"
 #include "common/microprofile.h"
+#include "video_core/engines/maxwell_3d.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_opengl/gl_buffer_cache.h"
+#include "video_core/renderer_opengl/gl_device.h"
 #include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
 
 namespace OpenGL {
 
+using Maxwell = Tegra::Engines::Maxwell3D::Regs;
+
 MICROPROFILE_DEFINE(OpenGL_Buffer_Download, "OpenGL", "Buffer Download", MP_RGB(192, 192, 128));
 
 CachedBufferBlock::CachedBufferBlock(CacheAddr cache_addr, const std::size_t size)
@@ -26,11 +30,22 @@ CachedBufferBlock::CachedBufferBlock(CacheAddr cache_addr, const std::size_t siz
 CachedBufferBlock::~CachedBufferBlock() = default;
 
 OGLBufferCache::OGLBufferCache(RasterizerOpenGL& rasterizer, Core::System& system,
-                               std::size_t stream_size)
-    : VideoCommon::BufferCache<Buffer, GLuint, OGLStreamBuffer>{
-          rasterizer, system, std::make_unique<OGLStreamBuffer>(stream_size, true)} {}
+                               const Device& device, std::size_t stream_size)
+    : GenericBufferCache{rasterizer, system, std::make_unique<OGLStreamBuffer>(stream_size, true)} {
+    if (!device.HasFastBufferSubData()) {
+        return;
+    }
 
-OGLBufferCache::~OGLBufferCache() = default;
+    static constexpr auto size = static_cast<GLsizeiptr>(Maxwell::MaxConstBufferSize);
+    glCreateBuffers(static_cast<GLsizei>(std::size(cbufs)), std::data(cbufs));
+    for (const GLuint cbuf : cbufs) {
+        glNamedBufferData(cbuf, size, nullptr, GL_STREAM_DRAW);
+    }
+}
+
+OGLBufferCache::~OGLBufferCache() {
+    glDeleteBuffers(static_cast<GLsizei>(std::size(cbufs)), std::data(cbufs));
+}
 
 Buffer OGLBufferCache::CreateBlock(CacheAddr cache_addr, std::size_t size) {
     return std::make_shared<CachedBufferBlock>(cache_addr, size);
@@ -69,4 +84,12 @@ void OGLBufferCache::CopyBlock(const Buffer& src, const Buffer& dst, std::size_t
                              static_cast<GLsizeiptr>(size));
 }
 
+OGLBufferCache::BufferInfo OGLBufferCache::ConstBufferUpload(const void* raw_pointer,
+                                                             std::size_t size) {
+    DEBUG_ASSERT(cbuf_cursor < std::size(cbufs));
+    const GLuint& cbuf = cbufs[cbuf_cursor++];
+    glNamedBufferSubData(cbuf, 0, static_cast<GLsizeiptr>(size), raw_pointer);
+    return {&cbuf, 0};
+}
+
 } // namespace OpenGL

src/video_core/renderer_opengl/gl_buffer_cache.h

@@ -4,10 +4,12 @@
 
 #pragma once
 
+#include <array>
 #include <memory>
 
 #include "common/common_types.h"
 #include "video_core/buffer_cache/buffer_cache.h"
+#include "video_core/engines/maxwell_3d.h"
 #include "video_core/rasterizer_cache.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
 #include "video_core/renderer_opengl/gl_stream_buffer.h"
@@ -18,12 +20,14 @@ class System;
 
 namespace OpenGL {
 
+class Device;
 class OGLStreamBuffer;
 class RasterizerOpenGL;
 
 class CachedBufferBlock;
 
 using Buffer = std::shared_ptr<CachedBufferBlock>;
+using GenericBufferCache = VideoCommon::BufferCache<Buffer, GLuint, OGLStreamBuffer>;
 
 class CachedBufferBlock : public VideoCommon::BufferBlock {
 public:
@@ -38,14 +42,18 @@ private:
     OGLBuffer gl_buffer{};
 };
 
-class OGLBufferCache final : public VideoCommon::BufferCache<Buffer, GLuint, OGLStreamBuffer> {
+class OGLBufferCache final : public GenericBufferCache {
 public:
     explicit OGLBufferCache(RasterizerOpenGL& rasterizer, Core::System& system,
-                            std::size_t stream_size);
+                            const Device& device, std::size_t stream_size);
     ~OGLBufferCache();
 
     const GLuint* GetEmptyBuffer(std::size_t) override;
 
+    void Acquire() noexcept {
+        cbuf_cursor = 0;
+    }
+
 protected:
     Buffer CreateBlock(CacheAddr cache_addr, std::size_t size) override;
 
@@ -61,6 +69,14 @@ protected:
 
     void CopyBlock(const Buffer& src, const Buffer& dst, std::size_t src_offset,
                    std::size_t dst_offset, std::size_t size) override;
+
+    BufferInfo ConstBufferUpload(const void* raw_pointer, std::size_t size) override;
+
+private:
+    std::size_t cbuf_cursor = 0;
+    std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::MaxConstBuffers *
+                           Tegra::Engines::Maxwell3D::Regs::MaxShaderProgram>
+        cbufs;
 };
 
 } // namespace OpenGL

src/video_core/renderer_opengl/gl_device.cpp

@@ -51,19 +51,24 @@ bool HasExtension(const std::vector<std::string_view>& images, std::string_view
 } // Anonymous namespace
 
 Device::Device() {
+    const std::string_view vendor = reinterpret_cast<const char*>(glGetString(GL_VENDOR));
     const std::vector extensions = GetExtensions();
 
+    const bool is_nvidia = vendor == "NVIDIA Corporation";
+
     uniform_buffer_alignment = GetInteger<std::size_t>(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT);
     shader_storage_alignment = GetInteger<std::size_t>(GL_SHADER_STORAGE_BUFFER_OFFSET_ALIGNMENT);
     max_vertex_attributes = GetInteger<u32>(GL_MAX_VERTEX_ATTRIBS);
     max_varyings = GetInteger<u32>(GL_MAX_VARYING_VECTORS);
     has_warp_intrinsics = GLAD_GL_NV_gpu_shader5 && GLAD_GL_NV_shader_thread_group &&
                           GLAD_GL_NV_shader_thread_shuffle;
+    has_shader_ballot = GLAD_GL_ARB_shader_ballot;
     has_vertex_viewport_layer = GLAD_GL_ARB_shader_viewport_layer_array;
     has_image_load_formatted = HasExtension(extensions, "GL_EXT_shader_image_load_formatted");
     has_variable_aoffi = TestVariableAoffi();
     has_component_indexing_bug = TestComponentIndexingBug();
     has_precise_bug = TestPreciseBug();
+    has_fast_buffer_sub_data = is_nvidia;
 
     LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi);
     LOG_INFO(Render_OpenGL, "Renderer_ComponentIndexingBug: {}", has_component_indexing_bug);
@@ -75,6 +80,7 @@ Device::Device(std::nullptr_t) {
     max_vertex_attributes = 16;
     max_varyings = 15;
     has_warp_intrinsics = true;
+    has_shader_ballot = true;
     has_vertex_viewport_layer = true;
     has_image_load_formatted = true;
     has_variable_aoffi = true;

src/video_core/renderer_opengl/gl_device.h

@@ -34,6 +34,10 @@ public:
         return has_warp_intrinsics;
     }
 
+    bool HasShaderBallot() const {
+        return has_shader_ballot;
+    }
+
     bool HasVertexViewportLayer() const {
         return has_vertex_viewport_layer;
     }
@@ -54,6 +58,10 @@ public:
         return has_precise_bug;
     }
 
+    bool HasFastBufferSubData() const {
+        return has_fast_buffer_sub_data;
+    }
+
 private:
     static bool TestVariableAoffi();
     static bool TestComponentIndexingBug();
@@ -64,11 +72,13 @@ private:
     u32 max_vertex_attributes{};
     u32 max_varyings{};
     bool has_warp_intrinsics{};
+    bool has_shader_ballot{};
     bool has_vertex_viewport_layer{};
     bool has_image_load_formatted{};
     bool has_variable_aoffi{};
     bool has_component_indexing_bug{};
     bool has_precise_bug{};
+    bool has_fast_buffer_sub_data{};
 };
 
 } // namespace OpenGL

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -67,7 +67,7 @@ static std::size_t GetConstBufferSize(const Tegra::Engines::ConstBufferInfo& buf
 RasterizerOpenGL::RasterizerOpenGL(Core::System& system, Core::Frontend::EmuWindow& emu_window,
                                    ScreenInfo& info)
     : texture_cache{system, *this, device}, shader_cache{*this, system, emu_window, device},
-      system{system}, screen_info{info}, buffer_cache{*this, system, STREAM_BUFFER_SIZE} {
+      system{system}, screen_info{info}, buffer_cache{*this, system, device, STREAM_BUFFER_SIZE} {
     shader_program_manager = std::make_unique<GLShader::ProgramManager>();
     state.draw.shader_program = 0;
     state.Apply();
@@ -375,7 +375,7 @@ void RasterizerOpenGL::ConfigureFramebuffers() {
         fbkey.color_attachments[index] = GL_COLOR_ATTACHMENT0 + regs.rt_control.GetMap(index);
         fbkey.colors[index] = std::move(color_surface);
     }
-    fbkey.colors_count = regs.rt_control.count;
+    fbkey.colors_count = static_cast<u16>(regs.rt_control.count);
 
     if (depth_surface) {
         // Assume that a surface will be written to if it is used as a framebuffer, even if
@@ -558,6 +558,8 @@ void RasterizerOpenGL::DrawPrelude() {
     SyncPolygonOffset();
     SyncAlphaTest();
 
+    buffer_cache.Acquire();
+
     // Draw the vertex batch
     const bool is_indexed = accelerate_draw == AccelDraw::Indexed;
 
@@ -879,7 +881,8 @@ void RasterizerOpenGL::SetupConstBuffer(const Tegra::Engines::ConstBufferInfo& b
     const std::size_t size = Common::AlignUp(GetConstBufferSize(buffer, entry), sizeof(GLvec4));
 
     const auto alignment = device.GetUniformBufferAlignment();
-    const auto [cbuf, offset] = buffer_cache.UploadMemory(buffer.address, size, alignment);
+    const auto [cbuf, offset] = buffer_cache.UploadMemory(buffer.address, size, alignment, false,
+                                                          device.HasFastBufferSubData());
     bind_ubo_pushbuffer.Push(cbuf, offset, size);
 }
 

src/video_core/renderer_opengl/gl_shader_cache.cpp

@@ -275,16 +275,25 @@ CachedProgram BuildShader(const Device& device, u64 unique_identifier, ProgramTy
     std::string source = fmt::format(R"(// {}
 #version 430 core
 #extension GL_ARB_separate_shader_objects : enable
-#extension GL_ARB_shader_viewport_layer_array : enable
-#extension GL_EXT_shader_image_load_formatted : enable
-#extension GL_NV_gpu_shader5 : enable
-#extension GL_NV_shader_thread_group : enable
-#extension GL_NV_shader_thread_shuffle : enable
 )",
                                      GetShaderId(unique_identifier, program_type));
     if (is_compute) {
         source += "#extension GL_ARB_compute_variable_group_size : require\n";
     }
+    if (device.HasShaderBallot()) {
+        source += "#extension GL_ARB_shader_ballot : require\n";
+    }
+    if (device.HasVertexViewportLayer()) {
+        source += "#extension GL_ARB_shader_viewport_layer_array : require\n";
+    }
+    if (device.HasImageLoadFormatted()) {
+        source += "#extension GL_EXT_shader_image_load_formatted : require\n";
+    }
+    if (device.HasWarpIntrinsics()) {
+        source += "#extension GL_NV_gpu_shader5 : require\n"
+                  "#extension GL_NV_shader_thread_group : require\n"
+                  "#extension GL_NV_shader_thread_shuffle : require\n";
+    }
     source += '\n';
 
     if (!is_compute) {
@@ -394,7 +403,8 @@ Shader CachedShader::CreateStageFromMemory(const ShaderParameters& params,
     params.disk_cache.SaveRaw(ShaderDiskCacheRaw(
         params.unique_identifier, GetProgramType(program_type), program_code, program_code_b));
 
-    ConstBufferLocker locker(GetEnginesShaderType(GetProgramType(program_type)));
+    ConstBufferLocker locker(GetEnginesShaderType(GetProgramType(program_type)),
+                             params.system.GPU().Maxwell3D());
     const ShaderIR ir(program_code, STAGE_MAIN_OFFSET, COMPILER_SETTINGS, locker);
     // TODO(Rodrigo): Handle VertexA shaders
     // std::optional<ShaderIR> ir_b;
@@ -410,7 +420,8 @@ Shader CachedShader::CreateKernelFromMemory(const ShaderParameters& params, Prog
     params.disk_cache.SaveRaw(
         ShaderDiskCacheRaw(params.unique_identifier, ProgramType::Compute, code));
 
-    ConstBufferLocker locker(Tegra::Engines::ShaderType::Compute);
+    ConstBufferLocker locker(Tegra::Engines::ShaderType::Compute,
+                             params.system.GPU().KeplerCompute());
     const ShaderIR ir(code, KERNEL_MAIN_OFFSET, COMPILER_SETTINGS, locker);
     return std::shared_ptr<CachedShader>(new CachedShader(
         params, ProgramType::Compute, GLShader::GetEntries(ir), std::move(code), {}));

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -1379,6 +1379,26 @@ private:
         return GenerateUnary(operation, "float", Type::Float, type);
     }
 
+    Expression FSwizzleAdd(Operation operation) {
+        const std::string op_a = VisitOperand(operation, 0).AsFloat();
+        const std::string op_b = VisitOperand(operation, 1).AsFloat();
+
+        if (!device.HasShaderBallot()) {
+            LOG_ERROR(Render_OpenGL, "Shader ballot is unavailable but required by the shader");
+            return {fmt::format("{} + {}", op_a, op_b), Type::Float};
+        }
+
+        const std::string instr_mask = VisitOperand(operation, 2).AsUint();
+        const std::string mask = code.GenerateTemporary();
+        code.AddLine("uint {} = ({} >> ((gl_SubGroupInvocationARB & 3) << 1)) & 3;", mask,
+                     instr_mask);
+
+        const std::string modifier_a = fmt::format("fswzadd_modifiers_a[{}]", mask);
+        const std::string modifier_b = fmt::format("fswzadd_modifiers_b[{}]", mask);
+        return {fmt::format("(({} * {}) + ({} * {}))", op_a, modifier_a, op_b, modifier_b),
+                Type::Float};
+    }
+
     Expression ICastFloat(Operation operation) {
         return GenerateUnary(operation, "int", Type::Int, Type::Float);
     }
@@ -1670,7 +1690,7 @@ private:
 
         const auto type = meta->sampler.IsShadow() ? Type::Float : Type::Int;
         return {GenerateTexture(operation, "Gather",
-                                {TextureArgument{type, meta->component}, TextureAoffi{}}) +
+                                {TextureAoffi{}, TextureArgument{type, meta->component}}) +
                     GetSwizzle(meta->element),
                 Type::Float};
     }
@@ -1942,34 +1962,24 @@ private:
         return Vote(operation, "allThreadsEqualNV");
     }
 
-    template <const std::string_view& func>
-    Expression Shuffle(Operation operation) {
-        const std::string value = VisitOperand(operation, 0).AsFloat();
-        if (!device.HasWarpIntrinsics()) {
-            LOG_ERROR(Render_OpenGL, "Nvidia shuffle intrinsics are required by this shader");
-            // On a "single-thread" device we are either on the same thread or out of bounds. Both
-            // cases return the passed value.
-            return {value, Type::Float};
+    Expression ThreadId(Operation operation) {
+        if (!device.HasShaderBallot()) {
+            LOG_ERROR(Render_OpenGL, "Shader ballot is unavailable but required by the shader");
+            return {"0U", Type::Uint};
+        }
+        return {"gl_SubGroupInvocationARB", Type::Uint};
+    }
+
+    Expression ShuffleIndexed(Operation operation) {
+        std::string value = VisitOperand(operation, 0).AsFloat();
+
+        if (!device.HasShaderBallot()) {
+            LOG_ERROR(Render_OpenGL, "Shader ballot is unavailable but required by the shader");
+            return {std::move(value), Type::Float};
         }
 
         const std::string index = VisitOperand(operation, 1).AsUint();
-        const std::string width = VisitOperand(operation, 2).AsUint();
-        return {fmt::format("{}({}, {}, {})", func, value, index, width), Type::Float};
-    }
-
-    template <const std::string_view& func>
-    Expression InRangeShuffle(Operation operation) {
-        const std::string index = VisitOperand(operation, 0).AsUint();
-        const std::string width = VisitOperand(operation, 1).AsUint();
-        if (!device.HasWarpIntrinsics()) {
-            // On a "single-thread" device we are only in bounds when the requested index is 0.
-            return {fmt::format("({} == 0U)", index), Type::Bool};
-        }
-
-        const std::string in_range = code.GenerateTemporary();
-        code.AddLine("bool {};", in_range);
-        code.AddLine("{}(0U, {}, {}, {});", func, index, width, in_range);
-        return {in_range, Type::Bool};
+        return {fmt::format("readInvocationARB({}, {})", value, index), Type::Float};
     }
 
     struct Func final {
@@ -1981,11 +1991,6 @@ private:
         static constexpr std::string_view Or = "Or";
         static constexpr std::string_view Xor = "Xor";
         static constexpr std::string_view Exchange = "Exchange";
-
-        static constexpr std::string_view ShuffleIndexed = "shuffleNV";
-        static constexpr std::string_view ShuffleUp = "shuffleUpNV";
-        static constexpr std::string_view ShuffleDown = "shuffleDownNV";
-        static constexpr std::string_view ShuffleButterfly = "shuffleXorNV";
     };
 
     static constexpr std::array operation_decompilers = {
@@ -2016,6 +2021,7 @@ private:
         &GLSLDecompiler::FTrunc,
         &GLSLDecompiler::FCastInteger<Type::Int>,
         &GLSLDecompiler::FCastInteger<Type::Uint>,
+        &GLSLDecompiler::FSwizzleAdd,
 
         &GLSLDecompiler::Add<Type::Int>,
         &GLSLDecompiler::Mul<Type::Int>,
@@ -2151,15 +2157,8 @@ private:
         &GLSLDecompiler::VoteAny,
         &GLSLDecompiler::VoteEqual,
 
-        &GLSLDecompiler::Shuffle<Func::ShuffleIndexed>,
-        &GLSLDecompiler::Shuffle<Func::ShuffleUp>,
-        &GLSLDecompiler::Shuffle<Func::ShuffleDown>,
-        &GLSLDecompiler::Shuffle<Func::ShuffleButterfly>,
-
-        &GLSLDecompiler::InRangeShuffle<Func::ShuffleIndexed>,
-        &GLSLDecompiler::InRangeShuffle<Func::ShuffleUp>,
-        &GLSLDecompiler::InRangeShuffle<Func::ShuffleDown>,
-        &GLSLDecompiler::InRangeShuffle<Func::ShuffleButterfly>,
+        &GLSLDecompiler::ThreadId,
+        &GLSLDecompiler::ShuffleIndexed,
     };
     static_assert(operation_decompilers.size() == static_cast<std::size_t>(OperationCode::Amount));
 
@@ -2492,6 +2491,9 @@ bvec2 HalfFloatNanComparison(bvec2 comparison, vec2 pair1, vec2 pair2) {
     bvec2 is_nan2 = isnan(pair2);
     return bvec2(comparison.x || is_nan1.x || is_nan2.x, comparison.y || is_nan1.y || is_nan2.y);
 }
+
+const float fswzadd_modifiers_a[] = float[4](-1.0f,  1.0f, -1.0f,  0.0f );
+const float fswzadd_modifiers_b[] = float[4](-1.0f, -1.0f,  1.0f, -1.0f );
 )";
 }
 

src/video_core/renderer_opengl/renderer_opengl.cpp

@@ -323,10 +323,12 @@ void RendererOpenGL::DrawScreenTriangles(const ScreenInfo& screen_info, float x,
     // (e.g. handheld mode) on a 1920x1080 framebuffer.
     f32 scale_u = 1.f, scale_v = 1.f;
     if (framebuffer_crop_rect.GetWidth() > 0) {
-        scale_u = static_cast<f32>(framebuffer_crop_rect.GetWidth()) / screen_info.texture.width;
+        scale_u = static_cast<f32>(framebuffer_crop_rect.GetWidth()) /
+                  static_cast<f32>(screen_info.texture.width);
     }
     if (framebuffer_crop_rect.GetHeight() > 0) {
-        scale_v = static_cast<f32>(framebuffer_crop_rect.GetHeight()) / screen_info.texture.height;
+        scale_v = static_cast<f32>(framebuffer_crop_rect.GetHeight()) /
+                  static_cast<f32>(screen_info.texture.height);
     }
 
     std::array<ScreenRectVertex, 4> vertices = {{

src/video_core/renderer_vulkan/vk_shader_decompiler.cpp

@@ -783,6 +783,11 @@ private:
         return {};
     }
 
+    Id FSwizzleAdd(Operation operation) {
+        UNIMPLEMENTED();
+        return {};
+    }
+
     Id HNegate(Operation operation) {
         UNIMPLEMENTED();
         return {};
@@ -1195,46 +1200,16 @@ private:
         return {};
     }
 
+    Id ThreadId(Operation) {
+        UNIMPLEMENTED();
+        return {};
+    }
+
     Id ShuffleIndexed(Operation) {
         UNIMPLEMENTED();
         return {};
     }
 
-    Id ShuffleUp(Operation) {
-        UNIMPLEMENTED();
-        return {};
-    }
-
-    Id ShuffleDown(Operation) {
-        UNIMPLEMENTED();
-        return {};
-    }
-
-    Id ShuffleButterfly(Operation) {
-        UNIMPLEMENTED();
-        return {};
-    }
-
-    Id InRangeShuffleIndexed(Operation) {
-        UNIMPLEMENTED();
-        return {};
-    }
-
-    Id InRangeShuffleUp(Operation) {
-        UNIMPLEMENTED();
-        return {};
-    }
-
-    Id InRangeShuffleDown(Operation) {
-        UNIMPLEMENTED();
-        return {};
-    }
-
-    Id InRangeShuffleButterfly(Operation) {
-        UNIMPLEMENTED();
-        return {};
-    }
-
     Id DeclareBuiltIn(spv::BuiltIn builtin, spv::StorageClass storage, Id type,
                       const std::string& name) {
         const Id id = OpVariable(type, storage);
@@ -1393,6 +1368,7 @@ private:
         &SPIRVDecompiler::Unary<&Module::OpTrunc, Type::Float>,
         &SPIRVDecompiler::Unary<&Module::OpConvertSToF, Type::Float, Type::Int>,
         &SPIRVDecompiler::Unary<&Module::OpConvertUToF, Type::Float, Type::Uint>,
+        &SPIRVDecompiler::FSwizzleAdd,
 
         &SPIRVDecompiler::Binary<&Module::OpIAdd, Type::Int>,
         &SPIRVDecompiler::Binary<&Module::OpIMul, Type::Int>,
@@ -1528,15 +1504,8 @@ private:
         &SPIRVDecompiler::VoteAny,
         &SPIRVDecompiler::VoteEqual,
 
+        &SPIRVDecompiler::ThreadId,
         &SPIRVDecompiler::ShuffleIndexed,
-        &SPIRVDecompiler::ShuffleUp,
-        &SPIRVDecompiler::ShuffleDown,
-        &SPIRVDecompiler::ShuffleButterfly,
-
-        &SPIRVDecompiler::InRangeShuffleIndexed,
-        &SPIRVDecompiler::InRangeShuffleUp,
-        &SPIRVDecompiler::InRangeShuffleDown,
-        &SPIRVDecompiler::InRangeShuffleButterfly,
     };
     static_assert(operation_decompilers.size() == static_cast<std::size_t>(OperationCode::Amount));
 

src/video_core/shader/control_flow.cpp

@@ -16,7 +16,9 @@
 #include "video_core/shader/shader_ir.h"
 
 namespace VideoCommon::Shader {
+
 namespace {
+
 using Tegra::Shader::Instruction;
 using Tegra::Shader::OpCode;
 
@@ -68,15 +70,15 @@ struct CFGRebuildState {
     const ProgramCode& program_code;
     ConstBufferLocker& locker;
     u32 start{};
-    std::vector<BlockInfo> block_info{};
-    std::list<u32> inspect_queries{};
-    std::list<Query> queries{};
-    std::unordered_map<u32, u32> registered{};
-    std::set<u32> labels{};
-    std::map<u32, u32> ssy_labels{};
-    std::map<u32, u32> pbk_labels{};
-    std::unordered_map<u32, BlockStack> stacks{};
-    ASTManager* manager;
+    std::vector<BlockInfo> block_info;
+    std::list<u32> inspect_queries;
+    std::list<Query> queries;
+    std::unordered_map<u32, u32> registered;
+    std::set<u32> labels;
+    std::map<u32, u32> ssy_labels;
+    std::map<u32, u32> pbk_labels;
+    std::unordered_map<u32, BlockStack> stacks;
+    ASTManager* manager{};
 };
 
 enum class BlockCollision : u32 { None, Found, Inside };
@@ -109,7 +111,7 @@ BlockInfo& CreateBlockInfo(CFGRebuildState& state, u32 start, u32 end) {
 }
 
 Pred GetPredicate(u32 index, bool negated) {
-    return static_cast<Pred>(index + (negated ? 8 : 0));
+    return static_cast<Pred>(static_cast<u64>(index) + (negated ? 8ULL : 0ULL));
 }
 
 /**
@@ -136,15 +138,13 @@ struct BranchIndirectInfo {
     s32 relative_position{};
 };
 
-std::optional<BranchIndirectInfo> TrackBranchIndirectInfo(const CFGRebuildState& state,
-                                                          u32 start_address, u32 current_position) {
-    const u32 shader_start = state.start;
-    u32 pos = current_position;
-    BranchIndirectInfo result{};
-    u64 track_register = 0;
+struct BufferInfo {
+    u32 index;
+    u32 offset;
+};
 
-    // Step 0 Get BRX Info
-    const Instruction instr = {state.program_code[pos]};
+std::optional<std::pair<s32, u64>> GetBRXInfo(const CFGRebuildState& state, u32& pos) {
+    const Instruction instr = state.program_code[pos];
     const auto opcode = OpCode::Decode(instr);
     if (opcode->get().GetId() != OpCode::Id::BRX) {
         return std::nullopt;
@@ -152,86 +152,94 @@ std::optional<BranchIndirectInfo> TrackBranchIndirectInfo(const CFGRebuildState&
     if (instr.brx.constant_buffer != 0) {
         return std::nullopt;
     }
-    track_register = instr.gpr8.Value();
-    result.relative_position = instr.brx.GetBranchExtend();
-    pos--;
-    bool found_track = false;
+    --pos;
+    return std::make_pair(instr.brx.GetBranchExtend(), instr.gpr8.Value());
+}
 
-    // Step 1 Track LDC
-    while (pos >= shader_start) {
-        if (IsSchedInstruction(pos, shader_start)) {
-            pos--;
-            continue;
-        }
-        const Instruction instr = {state.program_code[pos]};
-        const auto opcode = OpCode::Decode(instr);
-        if (opcode->get().GetId() == OpCode::Id::LD_C) {
-            if (instr.gpr0.Value() == track_register &&
-                instr.ld_c.type.Value() == Tegra::Shader::UniformType::Single) {
-                result.buffer = instr.cbuf36.index.Value();
-                result.offset = static_cast<u32>(instr.cbuf36.GetOffset());
-                track_register = instr.gpr8.Value();
-                pos--;
-                found_track = true;
-                break;
-            }
-        }
-        pos--;
-    }
-
-    if (!found_track) {
-        return std::nullopt;
-    }
-    found_track = false;
-
-    // Step 2 Track SHL
-    while (pos >= shader_start) {
-        if (IsSchedInstruction(pos, shader_start)) {
-            pos--;
+template <typename Result, typename TestCallable, typename PackCallable>
+// requires std::predicate<TestCallable, Instruction, const OpCode::Matcher&>
+// requires std::invocable<PackCallable, Instruction, const OpCode::Matcher&>
+std::optional<Result> TrackInstruction(const CFGRebuildState& state, u32& pos, TestCallable test,
+                                       PackCallable pack) {
+    for (; pos >= state.start; --pos) {
+        if (IsSchedInstruction(pos, state.start)) {
             continue;
         }
         const Instruction instr = state.program_code[pos];
         const auto opcode = OpCode::Decode(instr);
-        if (opcode->get().GetId() == OpCode::Id::SHL_IMM) {
-            if (instr.gpr0.Value() == track_register) {
-                track_register = instr.gpr8.Value();
-                pos--;
-                found_track = true;
-                break;
-            }
-        }
-        pos--;
-    }
-
-    if (!found_track) {
-        return std::nullopt;
-    }
-    found_track = false;
-
-    // Step 3 Track IMNMX
-    while (pos >= shader_start) {
-        if (IsSchedInstruction(pos, shader_start)) {
-            pos--;
+        if (!opcode) {
             continue;
         }
-        const Instruction instr = state.program_code[pos];
-        const auto opcode = OpCode::Decode(instr);
-        if (opcode->get().GetId() == OpCode::Id::IMNMX_IMM) {
-            if (instr.gpr0.Value() == track_register) {
-                track_register = instr.gpr8.Value();
-                result.entries = instr.alu.GetSignedImm20_20() + 1;
-                pos--;
-                found_track = true;
-                break;
-            }
+        if (test(instr, opcode->get())) {
+            --pos;
+            return std::make_optional(pack(instr, opcode->get()));
         }
-        pos--;
     }
+    return std::nullopt;
+}
 
-    if (!found_track) {
+std::optional<std::pair<BufferInfo, u64>> TrackLDC(const CFGRebuildState& state, u32& pos,
+                                                   u64 brx_tracked_register) {
+    return TrackInstruction<std::pair<BufferInfo, u64>>(
+        state, pos,
+        [brx_tracked_register](auto instr, const auto& opcode) {
+            return opcode.GetId() == OpCode::Id::LD_C &&
+                   instr.gpr0.Value() == brx_tracked_register &&
+                   instr.ld_c.type.Value() == Tegra::Shader::UniformType::Single;
+        },
+        [](auto instr, const auto& opcode) {
+            const BufferInfo info = {static_cast<u32>(instr.cbuf36.index.Value()),
+                                     static_cast<u32>(instr.cbuf36.GetOffset())};
+            return std::make_pair(info, instr.gpr8.Value());
+        });
+}
+
+std::optional<u64> TrackSHLRegister(const CFGRebuildState& state, u32& pos,
+                                    u64 ldc_tracked_register) {
+    return TrackInstruction<u64>(state, pos,
+                                 [ldc_tracked_register](auto instr, const auto& opcode) {
+                                     return opcode.GetId() == OpCode::Id::SHL_IMM &&
+                                            instr.gpr0.Value() == ldc_tracked_register;
+                                 },
+                                 [](auto instr, const auto&) { return instr.gpr8.Value(); });
+}
+
+std::optional<u32> TrackIMNMXValue(const CFGRebuildState& state, u32& pos,
+                                   u64 shl_tracked_register) {
+    return TrackInstruction<u32>(state, pos,
+                                 [shl_tracked_register](auto instr, const auto& opcode) {
+                                     return opcode.GetId() == OpCode::Id::IMNMX_IMM &&
+                                            instr.gpr0.Value() == shl_tracked_register;
+                                 },
+                                 [](auto instr, const auto&) {
+                                     return static_cast<u32>(instr.alu.GetSignedImm20_20() + 1);
+                                 });
+}
+
+std::optional<BranchIndirectInfo> TrackBranchIndirectInfo(const CFGRebuildState& state, u32 pos) {
+    const auto brx_info = GetBRXInfo(state, pos);
+    if (!brx_info) {
         return std::nullopt;
     }
-    return result;
+    const auto [relative_position, brx_tracked_register] = *brx_info;
+
+    const auto ldc_info = TrackLDC(state, pos, brx_tracked_register);
+    if (!ldc_info) {
+        return std::nullopt;
+    }
+    const auto [buffer_info, ldc_tracked_register] = *ldc_info;
+
+    const auto shl_tracked_register = TrackSHLRegister(state, pos, ldc_tracked_register);
+    if (!shl_tracked_register) {
+        return std::nullopt;
+    }
+
+    const auto entries = TrackIMNMXValue(state, pos, *shl_tracked_register);
+    if (!entries) {
+        return std::nullopt;
+    }
+
+    return BranchIndirectInfo{buffer_info.index, buffer_info.offset, *entries, relative_position};
 }
 
 std::pair<ParseResult, ParseInfo> ParseCode(CFGRebuildState& state, u32 address) {
@@ -420,30 +428,30 @@ std::pair<ParseResult, ParseInfo> ParseCode(CFGRebuildState& state, u32 address)
             break;
         }
         case OpCode::Id::BRX: {
-            auto tmp = TrackBranchIndirectInfo(state, address, offset);
-            if (tmp) {
-                auto result = *tmp;
-                std::vector<CaseBranch> branches{};
-                s32 pc_target = offset + result.relative_position;
-                for (u32 i = 0; i < result.entries; i++) {
-                    auto k = state.locker.ObtainKey(result.buffer, result.offset + i * 4);
-                    if (!k) {
-                        return {ParseResult::AbnormalFlow, parse_info};
-                    }
-                    u32 value = *k;
-                    u32 target = static_cast<u32>((value >> 3) + pc_target);
-                    insert_label(state, target);
-                    branches.emplace_back(value, target);
-                }
-                parse_info.end_address = offset;
-                parse_info.branch_info = MakeBranchInfo<MultiBranch>(
-                    static_cast<u32>(instr.gpr8.Value()), std::move(branches));
-
-                return {ParseResult::ControlCaught, parse_info};
-            } else {
+            const auto tmp = TrackBranchIndirectInfo(state, offset);
+            if (!tmp) {
                 LOG_WARNING(HW_GPU, "BRX Track Unsuccesful");
+                return {ParseResult::AbnormalFlow, parse_info};
             }
-            return {ParseResult::AbnormalFlow, parse_info};
+
+            const auto result = *tmp;
+            const s32 pc_target = offset + result.relative_position;
+            std::vector<CaseBranch> branches;
+            for (u32 i = 0; i < result.entries; i++) {
+                auto key = state.locker.ObtainKey(result.buffer, result.offset + i * 4);
+                if (!key) {
+                    return {ParseResult::AbnormalFlow, parse_info};
+                }
+                u32 value = *key;
+                u32 target = static_cast<u32>((value >> 3) + pc_target);
+                insert_label(state, target);
+                branches.emplace_back(value, target);
+            }
+            parse_info.end_address = offset;
+            parse_info.branch_info = MakeBranchInfo<MultiBranch>(
+                static_cast<u32>(instr.gpr8.Value()), std::move(branches));
+
+            return {ParseResult::ControlCaught, parse_info};
         }
         default:
             break;

src/video_core/shader/decode.cpp

@@ -154,10 +154,10 @@ void ShaderIR::Decode() {
         LOG_CRITICAL(HW_GPU, "Unknown decompilation mode!");
         [[fallthrough]];
     case CompileDepth::BruteForce: {
+        const auto shader_end = static_cast<u32>(program_code.size());
         coverage_begin = main_offset;
-        const std::size_t shader_end = program_code.size();
         coverage_end = shader_end;
-        for (u32 label = main_offset; label < shader_end; label++) {
+        for (u32 label = main_offset; label < shader_end; ++label) {
             basic_blocks.insert({label, DecodeRange(label, label + 1)});
         }
         break;

src/video_core/shader/decode/warp.cpp

@@ -17,6 +17,7 @@ using Tegra::Shader::ShuffleOperation;
 using Tegra::Shader::VoteOperation;
 
 namespace {
+
 OperationCode GetOperationCode(VoteOperation vote_op) {
     switch (vote_op) {
     case VoteOperation::All:
@@ -30,6 +31,7 @@ OperationCode GetOperationCode(VoteOperation vote_op) {
         return OperationCode::VoteAll;
     }
 }
+
 } // Anonymous namespace
 
 u32 ShaderIR::DecodeWarp(NodeBlock& bb, u32 pc) {
@@ -46,50 +48,59 @@ u32 ShaderIR::DecodeWarp(NodeBlock& bb, u32 pc) {
         break;
     }
     case OpCode::Id::SHFL: {
-        Node width = [this, instr] {
-            Node mask = instr.shfl.is_mask_imm ? Immediate(static_cast<u32>(instr.shfl.mask_imm))
-                                               : GetRegister(instr.gpr39);
-
-            // Convert the obscure SHFL mask back into GL_NV_shader_thread_shuffle's width. This has
-            // been done reversing Nvidia's math. It won't work on all cases due to SHFL having
-            // different parameters that don't properly map to GLSL's interface, but it should work
-            // for cases emitted by Nvidia's compiler.
-            if (instr.shfl.operation == ShuffleOperation::Up) {
-                return Operation(
-                    OperationCode::ILogicalShiftRight,
-                    Operation(OperationCode::IAdd, std::move(mask), Immediate(-0x2000)),
-                    Immediate(8));
-            } else {
-                return Operation(OperationCode::ILogicalShiftRight,
-                                 Operation(OperationCode::IAdd, Immediate(0x201F),
-                                           Operation(OperationCode::INegate, std::move(mask))),
-                                 Immediate(8));
-            }
-        }();
-
-        const auto [operation, in_range] = [instr]() -> std::pair<OperationCode, OperationCode> {
-            switch (instr.shfl.operation) {
-            case ShuffleOperation::Idx:
-                return {OperationCode::ShuffleIndexed, OperationCode::InRangeShuffleIndexed};
-            case ShuffleOperation::Up:
-                return {OperationCode::ShuffleUp, OperationCode::InRangeShuffleUp};
-            case ShuffleOperation::Down:
-                return {OperationCode::ShuffleDown, OperationCode::InRangeShuffleDown};
-            case ShuffleOperation::Bfly:
-                return {OperationCode::ShuffleButterfly, OperationCode::InRangeShuffleButterfly};
-            }
-            UNREACHABLE_MSG("Invalid SHFL operation: {}",
-                            static_cast<u64>(instr.shfl.operation.Value()));
-            return {};
-        }();
-
-        // Setting the predicate before the register is intentional to avoid overwriting.
+        Node mask = instr.shfl.is_mask_imm ? Immediate(static_cast<u32>(instr.shfl.mask_imm))
+                                           : GetRegister(instr.gpr39);
         Node index = instr.shfl.is_index_imm ? Immediate(static_cast<u32>(instr.shfl.index_imm))
                                              : GetRegister(instr.gpr20);
-        SetPredicate(bb, instr.shfl.pred48, Operation(in_range, index, width));
+
+        Node thread_id = Operation(OperationCode::ThreadId);
+        Node clamp = Operation(OperationCode::IBitwiseAnd, mask, Immediate(0x1FU));
+        Node seg_mask = BitfieldExtract(mask, 8, 16);
+
+        Node neg_seg_mask = Operation(OperationCode::IBitwiseNot, seg_mask);
+        Node min_thread_id = Operation(OperationCode::IBitwiseAnd, thread_id, seg_mask);
+        Node max_thread_id = Operation(OperationCode::IBitwiseOr, min_thread_id,
+                                       Operation(OperationCode::IBitwiseAnd, clamp, neg_seg_mask));
+
+        Node src_thread_id = [instr, index, neg_seg_mask, min_thread_id, thread_id] {
+            switch (instr.shfl.operation) {
+            case ShuffleOperation::Idx:
+                return Operation(OperationCode::IBitwiseOr,
+                                 Operation(OperationCode::IBitwiseAnd, index, neg_seg_mask),
+                                 min_thread_id);
+            case ShuffleOperation::Down:
+                return Operation(OperationCode::IAdd, thread_id, index);
+            case ShuffleOperation::Up:
+                return Operation(OperationCode::IAdd, thread_id,
+                                 Operation(OperationCode::INegate, index));
+            case ShuffleOperation::Bfly:
+                return Operation(OperationCode::IBitwiseXor, thread_id, index);
+            }
+            UNREACHABLE();
+            return Immediate(0U);
+        }();
+
+        Node in_bounds = [instr, src_thread_id, min_thread_id, max_thread_id] {
+            if (instr.shfl.operation == ShuffleOperation::Up) {
+                return Operation(OperationCode::LogicalIGreaterEqual, src_thread_id, min_thread_id);
+            } else {
+                return Operation(OperationCode::LogicalILessEqual, src_thread_id, max_thread_id);
+            }
+        }();
+
+        SetPredicate(bb, instr.shfl.pred48, in_bounds);
         SetRegister(
             bb, instr.gpr0,
-            Operation(operation, GetRegister(instr.gpr8), std::move(index), std::move(width)));
+            Operation(OperationCode::ShuffleIndexed, GetRegister(instr.gpr8), src_thread_id));
+        break;
+    }
+    case OpCode::Id::FSWZADD: {
+        UNIMPLEMENTED_IF(instr.fswzadd.ndv);
+
+        Node op_a = GetRegister(instr.gpr8);
+        Node op_b = GetRegister(instr.gpr20);
+        Node mask = Immediate(static_cast<u32>(instr.fswzadd.swizzle));
+        SetRegister(bb, instr.gpr0, Operation(OperationCode::FSwizzleAdd, op_a, op_b, mask));
         break;
     }
     default:

src/video_core/shader/node.h

@@ -47,6 +47,7 @@ enum class OperationCode {
     FTrunc,        /// (MetaArithmetic, float a) -> float
     FCastInteger,  /// (MetaArithmetic, int a) -> float
     FCastUInteger, /// (MetaArithmetic, uint a) -> float
+    FSwizzleAdd,   /// (float a, float b, uint mask) -> float
 
     IAdd,                  /// (MetaArithmetic, int a, int b) -> int
     IMul,                  /// (MetaArithmetic, int a, int b) -> int
@@ -181,15 +182,8 @@ enum class OperationCode {
     VoteAny,      /// (bool) -> bool
     VoteEqual,    /// (bool) -> bool
 
-    ShuffleIndexed,   /// (uint value, uint index, uint width) -> uint
-    ShuffleUp,        /// (uint value, uint index, uint width) -> uint
-    ShuffleDown,      /// (uint value, uint index, uint width) -> uint
-    ShuffleButterfly, /// (uint value, uint index, uint width) -> uint
-
-    InRangeShuffleIndexed,   /// (uint index, uint width) -> bool
-    InRangeShuffleUp,        /// (uint index, uint width) -> bool
-    InRangeShuffleDown,      /// (uint index, uint width) -> bool
-    InRangeShuffleButterfly, /// (uint index, uint width) -> bool
+    ThreadId,       /// () -> uint
+    ShuffleIndexed, /// (uint value, uint index) -> uint
 
     Amount,
 };

src/video_core/shader/shader_ir.h

@@ -49,7 +49,7 @@ public:
     }
 
     u32 GetSize() const {
-        return max_offset + sizeof(float);
+        return max_offset + static_cast<u32>(sizeof(float));
     }
 
     u32 GetMaxOffset() const {
@@ -165,8 +165,8 @@ public:
         return program_manager.GetVariables();
     }
 
-    u32 ConvertAddressToNvidiaSpace(const u32 address) const {
-        return (address - main_offset) * sizeof(Tegra::Shader::Instruction);
+    u32 ConvertAddressToNvidiaSpace(u32 address) const {
+        return (address - main_offset) * static_cast<u32>(sizeof(Tegra::Shader::Instruction));
     }
 
     /// Returns a condition code evaluated from internal flags

src/video_core/textures/astc.cpp

@@ -92,11 +92,11 @@ private:
         const unsigned int mask = 1 << m_NextBit++;
 
         // clear the bit
-        *m_CurByte &= ~mask;
+        *m_CurByte &= static_cast<unsigned char>(~mask);
 
         // Write the bit, if necessary
         if (b)
-            *m_CurByte |= mask;
+            *m_CurByte |= static_cast<unsigned char>(mask);
 
         // Next byte?
         if (m_NextBit >= 8) {
@@ -137,7 +137,7 @@ public:
         }
 
         uint64_t mask = (1 << (end - start + 1)) - 1;
-        return (m_Bits >> start) & mask;
+        return (m_Bits >> start) & static_cast<IntType>(mask);
     }
 
 private:
@@ -656,7 +656,7 @@ static IntType Replicate(const IntType& val, uint32_t numBits, uint32_t toBit) {
         return 0;
     if (toBit == 0)
         return 0;
-    IntType v = val & ((1 << numBits) - 1);
+    IntType v = val & static_cast<IntType>((1 << numBits) - 1);
     IntType res = v;
     uint32_t reslen = numBits;
     while (reslen < toBit) {
@@ -666,8 +666,8 @@ static IntType Replicate(const IntType& val, uint32_t numBits, uint32_t toBit) {
             comp = numBits - newshift;
             numBits = newshift;
         }
-        res <<= numBits;
-        res |= v >> comp;
+        res = static_cast<IntType>(res << numBits);
+        res = static_cast<IntType>(res | (v >> comp));
         reslen += numBits;
     }
     return res;
@@ -714,7 +714,7 @@ public:
             // Do nothing
             return val;
         } else if (oldDepth == 0 && newDepth != 0) {
-            return (1 << newDepth) - 1;
+            return static_cast<ChannelType>((1 << newDepth) - 1);
         } else if (newDepth > oldDepth) {
             return Replicate(val, oldDepth, newDepth);
         } else {
@@ -722,10 +722,11 @@ public:
             if (newDepth == 0) {
                 return 0xFF;
             } else {
-                uint8_t bitsWasted = oldDepth - newDepth;
+                uint8_t bitsWasted = static_cast<uint8_t>(oldDepth - newDepth);
                 uint16_t v = static_cast<uint16_t>(val);
-                v = (v + (1 << (bitsWasted - 1))) >> bitsWasted;
-                v = ::std::min<uint16_t>(::std::max<uint16_t>(0, v), (1 << newDepth) - 1);
+                v = static_cast<uint16_t>((v + (1 << (bitsWasted - 1))) >> bitsWasted);
+                v = ::std::min<uint16_t>(::std::max<uint16_t>(0, v),
+                                         static_cast<uint16_t>((1 << newDepth) - 1));
                 return static_cast<uint8_t>(v);
             }
         }
@@ -1191,18 +1192,18 @@ static uint32_t SelectPartition(int32_t seed, int32_t x, int32_t y, int32_t z,
     uint8_t seed11 = static_cast<uint8_t>((rnum >> 26) & 0xF);
     uint8_t seed12 = static_cast<uint8_t>(((rnum >> 30) | (rnum << 2)) & 0xF);
 
-    seed1 *= seed1;
-    seed2 *= seed2;
-    seed3 *= seed3;
-    seed4 *= seed4;
-    seed5 *= seed5;
-    seed6 *= seed6;
-    seed7 *= seed7;
-    seed8 *= seed8;
-    seed9 *= seed9;
-    seed10 *= seed10;
-    seed11 *= seed11;
-    seed12 *= seed12;
+    seed1 = static_cast<uint8_t>(seed1 * seed1);
+    seed2 = static_cast<uint8_t>(seed2 * seed2);
+    seed3 = static_cast<uint8_t>(seed3 * seed3);
+    seed4 = static_cast<uint8_t>(seed4 * seed4);
+    seed5 = static_cast<uint8_t>(seed5 * seed5);
+    seed6 = static_cast<uint8_t>(seed6 * seed6);
+    seed7 = static_cast<uint8_t>(seed7 * seed7);
+    seed8 = static_cast<uint8_t>(seed8 * seed8);
+    seed9 = static_cast<uint8_t>(seed9 * seed9);
+    seed10 = static_cast<uint8_t>(seed10 * seed10);
+    seed11 = static_cast<uint8_t>(seed11 * seed11);
+    seed12 = static_cast<uint8_t>(seed12 * seed12);
 
     int32_t sh1, sh2, sh3;
     if (seed & 1) {
@@ -1214,18 +1215,18 @@ static uint32_t SelectPartition(int32_t seed, int32_t x, int32_t y, int32_t z,
     }
     sh3 = (seed & 0x10) ? sh1 : sh2;
 
-    seed1 >>= sh1;
-    seed2 >>= sh2;
-    seed3 >>= sh1;
-    seed4 >>= sh2;
-    seed5 >>= sh1;
-    seed6 >>= sh2;
-    seed7 >>= sh1;
-    seed8 >>= sh2;
-    seed9 >>= sh3;
-    seed10 >>= sh3;
-    seed11 >>= sh3;
-    seed12 >>= sh3;
+    seed1 = static_cast<uint8_t>(seed1 >> sh1);
+    seed2 = static_cast<uint8_t>(seed2 >> sh2);
+    seed3 = static_cast<uint8_t>(seed3 >> sh1);
+    seed4 = static_cast<uint8_t>(seed4 >> sh2);
+    seed5 = static_cast<uint8_t>(seed5 >> sh1);
+    seed6 = static_cast<uint8_t>(seed6 >> sh2);
+    seed7 = static_cast<uint8_t>(seed7 >> sh1);
+    seed8 = static_cast<uint8_t>(seed8 >> sh2);
+    seed9 = static_cast<uint8_t>(seed9 >> sh3);
+    seed10 = static_cast<uint8_t>(seed10 >> sh3);
+    seed11 = static_cast<uint8_t>(seed11 >> sh3);
+    seed12 = static_cast<uint8_t>(seed12 >> sh3);
 
     int32_t a = seed1 * x + seed2 * y + seed11 * z + (rnum >> 14);
     int32_t b = seed3 * x + seed4 * y + seed12 * z + (rnum >> 10);
@@ -1558,7 +1559,9 @@ static void DecompressBlock(const uint8_t inBuf[16], const uint32_t blockWidth,
 
     // Make sure that higher non-texel bits are set to zero
     const uint32_t clearByteStart = (weightParams.GetPackedBitSize() >> 3) + 1;
-    texelWeightData[clearByteStart - 1] &= (1 << (weightParams.GetPackedBitSize() % 8)) - 1;
+    texelWeightData[clearByteStart - 1] =
+        texelWeightData[clearByteStart - 1] &
+        static_cast<uint8_t>((1 << (weightParams.GetPackedBitSize() % 8)) - 1);
     memset(texelWeightData + clearByteStart, 0, 16 - clearByteStart);
 
     std::vector<IntegerEncodedValue> texelWeightValues;

src/video_core/textures/texture.h

@@ -342,13 +342,14 @@ struct TSCEntry {
     float GetLodBias() const {
         // Sign extend the 13-bit value.
         constexpr u32 mask = 1U << (13 - 1);
-        return static_cast<s32>((mip_lod_bias ^ mask) - mask) / 256.0f;
+        return static_cast<float>(static_cast<s32>((mip_lod_bias ^ mask) - mask)) / 256.0f;
     }
 
     std::array<float, 4> GetBorderColor() const {
         if (srgb_conversion) {
-            return {srgb_border_color_r / 255.0f, srgb_border_color_g / 255.0f,
-                    srgb_border_color_b / 255.0f, border_color[3]};
+            return {static_cast<float>(srgb_border_color_r) / 255.0f,
+                    static_cast<float>(srgb_border_color_g) / 255.0f,
+                    static_cast<float>(srgb_border_color_b) / 255.0f, border_color[3]};
         }
         return border_color;
     }

src/video_core/video_core.cpp

@@ -28,7 +28,7 @@ std::unique_ptr<Tegra::GPU> CreateGPU(Core::System& system) {
 
 u16 GetResolutionScaleFactor(const RendererBase& renderer) {
     return static_cast<u16>(
-        Settings::values.resolution_factor
+        Settings::values.resolution_factor != 0
             ? Settings::values.resolution_factor
             : renderer.GetRenderWindow().GetFramebufferLayout().GetScalingRatio());
 }

src/web_service/web_backend.cpp

@@ -2,10 +2,12 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <array>
 #include <cstdlib>
 #include <mutex>
 #include <string>
 #include <LUrlParser.h>
+#include <fmt/format.h>
 #include <httplib.h>
 #include "common/common_types.h"
 #include "common/logging/log.h"
@@ -16,10 +18,10 @@ namespace WebService {
 
 constexpr std::array<const char, 1> API_VERSION{'1'};
 
-constexpr u32 HTTP_PORT = 80;
-constexpr u32 HTTPS_PORT = 443;
+constexpr int HTTP_PORT = 80;
+constexpr int HTTPS_PORT = 443;
 
-constexpr u32 TIMEOUT_SECONDS = 30;
+constexpr std::size_t TIMEOUT_SECONDS = 30;
 
 struct Client::Impl {
     Impl(std::string host, std::string username, std::string token)
@@ -31,8 +33,9 @@ struct Client::Impl {
     }
 
     /// A generic function handles POST, GET and DELETE request together
-    Common::WebResult GenericJson(const std::string& method, const std::string& path,
-                                  const std::string& data, bool allow_anonymous) {
+    Common::WebResult GenericRequest(const std::string& method, const std::string& path,
+                                     const std::string& data, bool allow_anonymous,
+                                     const std::string& accept) {
         if (jwt.empty()) {
             UpdateJWT();
         }
@@ -43,11 +46,11 @@ struct Client::Impl {
                                      "Credentials needed"};
         }
 
-        auto result = GenericJson(method, path, data, jwt);
+        auto result = GenericRequest(method, path, data, accept, jwt);
         if (result.result_string == "401") {
             // Try again with new JWT
             UpdateJWT();
-            result = GenericJson(method, path, data, jwt);
+            result = GenericRequest(method, path, data, accept, jwt);
         }
 
         return result;
@@ -56,12 +59,13 @@ struct Client::Impl {
     /**
      * A generic function with explicit authentication method specified
      * JWT is used if the jwt parameter is not empty
-     * username + token is used if jwt is empty but username and token are not empty
-     * anonymous if all of jwt, username and token are empty
+     * username + token is used if jwt is empty but username and token are
+     * not empty anonymous if all of jwt, username and token are empty
      */
-    Common::WebResult GenericJson(const std::string& method, const std::string& path,
-                                  const std::string& data, const std::string& jwt = "",
-                                  const std::string& username = "", const std::string& token = "") {
+    Common::WebResult GenericRequest(const std::string& method, const std::string& path,
+                                     const std::string& data, const std::string& accept,
+                                     const std::string& jwt = "", const std::string& username = "",
+                                     const std::string& token = "") {
         if (cli == nullptr) {
             auto parsedUrl = LUrlParser::clParseURL::ParseURL(host);
             int port;
@@ -132,8 +136,7 @@ struct Client::Impl {
             return Common::WebResult{Common::WebResult::Code::WrongContent, ""};
         }
 
-        if (content_type->second.find("application/json") == std::string::npos &&
-            content_type->second.find("text/html; charset=utf-8") == std::string::npos) {
+        if (content_type->second.find(accept) == std::string::npos) {
             LOG_ERROR(WebService, "{} to {} returned wrong content: {}", method, host + path,
                       content_type->second);
             return Common::WebResult{Common::WebResult::Code::WrongContent, "Wrong content"};
@@ -147,7 +150,7 @@ struct Client::Impl {
             return;
         }
 
-        auto result = GenericJson("POST", "/jwt/internal", "", "", username, token);
+        auto result = GenericRequest("POST", "/jwt/internal", "", "text/html", "", username, token);
         if (result.result_code != Common::WebResult::Code::Success) {
             LOG_ERROR(WebService, "UpdateJWT failed");
         } else {
@@ -180,16 +183,29 @@ Client::~Client() = default;
 
 Common::WebResult Client::PostJson(const std::string& path, const std::string& data,
                                    bool allow_anonymous) {
-    return impl->GenericJson("POST", path, data, allow_anonymous);
+    return impl->GenericRequest("POST", path, data, allow_anonymous, "application/json");
 }
 
 Common::WebResult Client::GetJson(const std::string& path, bool allow_anonymous) {
-    return impl->GenericJson("GET", path, "", allow_anonymous);
+    return impl->GenericRequest("GET", path, "", allow_anonymous, "application/json");
 }
 
 Common::WebResult Client::DeleteJson(const std::string& path, const std::string& data,
                                      bool allow_anonymous) {
-    return impl->GenericJson("DELETE", path, data, allow_anonymous);
+    return impl->GenericRequest("DELETE", path, data, allow_anonymous, "application/json");
+}
+
+Common::WebResult Client::GetPlain(const std::string& path, bool allow_anonymous) {
+    return impl->GenericRequest("GET", path, "", allow_anonymous, "text/plain");
+}
+
+Common::WebResult Client::GetImage(const std::string& path, bool allow_anonymous) {
+    return impl->GenericRequest("GET", path, "", allow_anonymous, "image/png");
+}
+
+Common::WebResult Client::GetExternalJWT(const std::string& audience) {
+    return impl->GenericRequest("POST", fmt::format("/jwt/external/{}", audience), "", false,
+                                "text/html");
 }
 
 } // namespace WebService

src/web_service/web_backend.h

@@ -46,6 +46,29 @@ public:
     Common::WebResult DeleteJson(const std::string& path, const std::string& data,
                                  bool allow_anonymous);
 
+    /**
+     * Gets a plain string from the specified path.
+     * @param path the URL segment after the host address.
+     * @param allow_anonymous If true, allow anonymous unauthenticated requests.
+     * @return the result of the request.
+     */
+    Common::WebResult GetPlain(const std::string& path, bool allow_anonymous);
+
+    /**
+     * Gets an PNG image from the specified path.
+     * @param path the URL segment after the host address.
+     * @param allow_anonymous If true, allow anonymous unauthenticated requests.
+     * @return the result of the request.
+     */
+    Common::WebResult GetImage(const std::string& path, bool allow_anonymous);
+
+    /**
+     * Requests an external JWT for the specific audience provided.
+     * @param audience the audience of the JWT requested.
+     * @return the result of the request.
+     */
+    Common::WebResult GetExternalJWT(const std::string& audience);
+
 private:
     struct Impl;
     std::unique_ptr<Impl> impl;

src/yuzu/configuration/configure_web.cpp

@@ -11,6 +11,31 @@
 #include "yuzu/configuration/configure_web.h"
 #include "yuzu/uisettings.h"
 
+static constexpr char token_delimiter{':'};
+
+static std::string GenerateDisplayToken(const std::string& username, const std::string& token) {
+    if (username.empty() || token.empty()) {
+        return {};
+    }
+
+    const std::string unencoded_display_token{username + token_delimiter + token};
+    QByteArray b{unencoded_display_token.c_str()};
+    QByteArray b64 = b.toBase64();
+    return b64.toStdString();
+}
+
+static std::string UsernameFromDisplayToken(const std::string& display_token) {
+    const std::string unencoded_display_token{
+        QByteArray::fromBase64(display_token.c_str()).toStdString()};
+    return unencoded_display_token.substr(0, unencoded_display_token.find(token_delimiter));
+}
+
+static std::string TokenFromDisplayToken(const std::string& display_token) {
+    const std::string unencoded_display_token{
+        QByteArray::fromBase64(display_token.c_str()).toStdString()};
+    return unencoded_display_token.substr(unencoded_display_token.find(token_delimiter) + 1);
+}
+
 ConfigureWeb::ConfigureWeb(QWidget* parent)
     : QWidget(parent), ui(std::make_unique<Ui::ConfigureWeb>()) {
     ui->setupUi(this);
@@ -63,13 +88,18 @@ void ConfigureWeb::SetConfiguration() {
     ui->web_signup_link->setOpenExternalLinks(true);
     ui->web_token_info_link->setOpenExternalLinks(true);
 
+    if (Settings::values.yuzu_username.empty()) {
+        ui->username->setText(tr("Unspecified"));
+    } else {
+        ui->username->setText(QString::fromStdString(Settings::values.yuzu_username));
+    }
+
     ui->toggle_telemetry->setChecked(Settings::values.enable_telemetry);
-    ui->edit_username->setText(QString::fromStdString(Settings::values.yuzu_username));
-    ui->edit_token->setText(QString::fromStdString(Settings::values.yuzu_token));
+    ui->edit_token->setText(QString::fromStdString(
+        GenerateDisplayToken(Settings::values.yuzu_username, Settings::values.yuzu_token)));
 
     // Connect after setting the values, to avoid calling OnLoginChanged now
     connect(ui->edit_token, &QLineEdit::textChanged, this, &ConfigureWeb::OnLoginChanged);
-    connect(ui->edit_username, &QLineEdit::textChanged, this, &ConfigureWeb::OnLoginChanged);
 
     user_verified = true;
 
@@ -80,12 +110,13 @@ void ConfigureWeb::ApplyConfiguration() {
     Settings::values.enable_telemetry = ui->toggle_telemetry->isChecked();
     UISettings::values.enable_discord_presence = ui->toggle_discordrpc->isChecked();
     if (user_verified) {
-        Settings::values.yuzu_username = ui->edit_username->text().toStdString();
-        Settings::values.yuzu_token = ui->edit_token->text().toStdString();
+        Settings::values.yuzu_username =
+            UsernameFromDisplayToken(ui->edit_token->text().toStdString());
+        Settings::values.yuzu_token = TokenFromDisplayToken(ui->edit_token->text().toStdString());
     } else {
-        QMessageBox::warning(this, tr("Username and token not verified"),
-                             tr("Username and token were not verified. The changes to your "
-                                "username and/or token have not been saved."));
+        QMessageBox::warning(
+            this, tr("Token not verified"),
+            tr("Token was not verified. The change to your token has not been saved."));
     }
 }
 
@@ -96,17 +127,15 @@ void ConfigureWeb::RefreshTelemetryID() {
 }
 
 void ConfigureWeb::OnLoginChanged() {
-    if (ui->edit_username->text().isEmpty() && ui->edit_token->text().isEmpty()) {
+    if (ui->edit_token->text().isEmpty()) {
         user_verified = true;
 
         const QPixmap pixmap = QIcon::fromTheme(QStringLiteral("checked")).pixmap(16);
-        ui->label_username_verified->setPixmap(pixmap);
         ui->label_token_verified->setPixmap(pixmap);
     } else {
         user_verified = false;
 
         const QPixmap pixmap = QIcon::fromTheme(QStringLiteral("failed")).pixmap(16);
-        ui->label_username_verified->setPixmap(pixmap);
         ui->label_token_verified->setPixmap(pixmap);
     }
 }
@@ -114,10 +143,11 @@ void ConfigureWeb::OnLoginChanged() {
 void ConfigureWeb::VerifyLogin() {
     ui->button_verify_login->setDisabled(true);
     ui->button_verify_login->setText(tr("Verifying..."));
-    verify_watcher.setFuture(QtConcurrent::run([username = ui->edit_username->text().toStdString(),
-                                                token = ui->edit_token->text().toStdString()] {
-        return Core::VerifyLogin(username, token);
-    }));
+    verify_watcher.setFuture(QtConcurrent::run(
+        [username = UsernameFromDisplayToken(ui->edit_token->text().toStdString()),
+         token = TokenFromDisplayToken(ui->edit_token->text().toStdString())] {
+            return Core::VerifyLogin(username, token);
+        }));
 }
 
 void ConfigureWeb::OnLoginVerified() {
@@ -127,16 +157,15 @@ void ConfigureWeb::OnLoginVerified() {
         user_verified = true;
 
         const QPixmap pixmap = QIcon::fromTheme(QStringLiteral("checked")).pixmap(16);
-        ui->label_username_verified->setPixmap(pixmap);
         ui->label_token_verified->setPixmap(pixmap);
+        ui->username->setText(
+            QString::fromStdString(UsernameFromDisplayToken(ui->edit_token->text().toStdString())));
     } else {
         const QPixmap pixmap = QIcon::fromTheme(QStringLiteral("failed")).pixmap(16);
-        ui->label_username_verified->setPixmap(pixmap);
         ui->label_token_verified->setPixmap(pixmap);
-
-        QMessageBox::critical(
-            this, tr("Verification failed"),
-            tr("Verification failed. Check that you have entered your username and token "
-               "correctly, and that your internet connection is working."));
+        ui->username->setText(tr("Unspecified"));
+        QMessageBox::critical(this, tr("Verification failed"),
+                              tr("Verification failed. Check that you have entered your token "
+                                 "correctly, and that your internet connection is working."));
     }
 }

src/yuzu/configuration/configure_web.ui

@@ -55,11 +55,7 @@
            </widget>
           </item>
           <item row="0" column="1" colspan="3">
-           <widget class="QLineEdit" name="edit_username">
-            <property name="maxLength">
-             <number>36</number>
-            </property>
-           </widget>
+           <widget class="QLabel" name="username" />
           </item>
           <item row="1" column="0">
            <widget class="QLabel" name="label_token">
@@ -79,14 +75,10 @@
             </property>
            </widget>
           </item>
-          <item row="0" column="4">
-           <widget class="QLabel" name="label_username_verified">
-           </widget>
-          </item>
           <item row="1" column="1" colspan="3">
            <widget class="QLineEdit" name="edit_token">
             <property name="maxLength">
-             <number>36</number>
+             <number>80</number>
             </property>
             <property name="echoMode">
              <enum>QLineEdit::Password</enum>

src/yuzu/debugger/wait_tree.cpp

@@ -172,17 +172,6 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeWaitObject::GetChildren() con
     return list;
 }
 
-QString WaitTreeWaitObject::GetResetTypeQString(Kernel::ResetType reset_type) {
-    switch (reset_type) {
-    case Kernel::ResetType::Automatic:
-        return tr("automatic reset");
-    case Kernel::ResetType::Manual:
-        return tr("manual reset");
-    }
-    UNREACHABLE();
-    return {};
-}
-
 WaitTreeObjectList::WaitTreeObjectList(
     const std::vector<Kernel::SharedPtr<Kernel::WaitObject>>& list, bool w_all)
     : object_list(list), wait_all(w_all) {}
@@ -336,16 +325,6 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeThread::GetChildren() const {
 WaitTreeEvent::WaitTreeEvent(const Kernel::ReadableEvent& object) : WaitTreeWaitObject(object) {}
 WaitTreeEvent::~WaitTreeEvent() = default;
 
-std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeEvent::GetChildren() const {
-    std::vector<std::unique_ptr<WaitTreeItem>> list(WaitTreeWaitObject::GetChildren());
-
-    list.push_back(std::make_unique<WaitTreeText>(
-        tr("reset type = %1")
-            .arg(GetResetTypeQString(
-                static_cast<const Kernel::ReadableEvent&>(object).GetResetType()))));
-    return list;
-}
-
 WaitTreeThreadList::WaitTreeThreadList(const std::vector<Kernel::SharedPtr<Kernel::Thread>>& list)
     : thread_list(list) {}
 WaitTreeThreadList::~WaitTreeThreadList() = default;

src/yuzu/debugger/wait_tree.h

@@ -111,8 +111,6 @@ public:
 
 protected:
     const Kernel::WaitObject& object;
-
-    static QString GetResetTypeQString(Kernel::ResetType reset_type);
 };
 
 class WaitTreeObjectList : public WaitTreeExpandableItem {
@@ -146,8 +144,6 @@ class WaitTreeEvent : public WaitTreeWaitObject {
 public:
     explicit WaitTreeEvent(const Kernel::ReadableEvent& object);
     ~WaitTreeEvent() override;
-
-    std::vector<std::unique_ptr<WaitTreeItem>> GetChildren() const override;
 };
 
 class WaitTreeThreadList : public WaitTreeExpandableItem {