rasterizer_cache: Remove reliance on the System singleton

Rather than have a transparent dependency, we can make it explicit in
the interface. This also gets rid of the need to put the core include in
a header.

src/audio_core/cubeb_sink.cpp

@@ -121,7 +121,8 @@ CubebSink::CubebSink(std::string target_device_name) {
             const auto collection_end{collection.device + collection.count};
             const auto device{
                 std::find_if(collection.device, collection_end, [&](const cubeb_device_info& info) {
-                    return target_device_name == info.friendly_name;
+                    return info.friendly_name != nullptr &&
+                           target_device_name == info.friendly_name;
                 })};
             if (device != collection_end) {
                 output_device = device->devid;

src/audio_core/stream.cpp

@@ -11,7 +11,6 @@
 #include "audio_core/stream.h"
 #include "common/assert.h"
 #include "common/logging/log.h"
-#include "common/microprofile.h"
 #include "core/core_timing.h"
 #include "core/core_timing_util.h"
 #include "core/settings.h"
@@ -104,10 +103,7 @@ void Stream::PlayNextBuffer() {
     CoreTiming::ScheduleEventThreadsafe(GetBufferReleaseCycles(*active_buffer), release_event, {});
 }
 
-MICROPROFILE_DEFINE(AudioOutput, "Audio", "ReleaseActiveBuffer", MP_RGB(100, 100, 255));
-
 void Stream::ReleaseActiveBuffer() {
-    MICROPROFILE_SCOPE(AudioOutput);
     ASSERT(active_buffer);
     released_buffers.push(std::move(active_buffer));
     release_callback();

src/audio_core/time_stretch.cpp

@@ -10,8 +10,7 @@
 
 namespace AudioCore {
 
-TimeStretcher::TimeStretcher(u32 sample_rate, u32 channel_count)
-    : m_sample_rate(sample_rate), m_channel_count(channel_count) {
+TimeStretcher::TimeStretcher(u32 sample_rate, u32 channel_count) : m_sample_rate{sample_rate} {
     m_sound_touch.setChannels(channel_count);
     m_sound_touch.setSampleRate(sample_rate);
     m_sound_touch.setPitch(1.0);
@@ -33,10 +32,10 @@ std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out,
     // We were given actual_samples number of samples, and num_samples were requested from us.
     double current_ratio = static_cast<double>(num_in) / static_cast<double>(num_out);
 
-    const double max_latency = 1.0; // seconds
+    const double max_latency = 0.25; // seconds
     const double max_backlog = m_sample_rate * max_latency;
     const double backlog_fullness = m_sound_touch.numSamples() / max_backlog;
-    if (backlog_fullness > 5.0) {
+    if (backlog_fullness > 4.0) {
         // Too many samples in backlog: Don't push anymore on
         num_in = 0;
     }
@@ -50,7 +49,7 @@ std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out,
 
     // This low-pass filter smoothes out variance in the calculated stretch ratio.
     // The time-scale determines how responsive this filter is.
-    constexpr double lpf_time_scale = 2.0; // seconds
+    constexpr double lpf_time_scale = 0.712; // seconds
     const double lpf_gain = 1.0 - std::exp(-time_delta / lpf_time_scale);
     m_stretch_ratio += lpf_gain * (current_ratio - m_stretch_ratio);
 

src/audio_core/time_stretch.h

@@ -27,7 +27,6 @@ public:
 
 private:
     u32 m_sample_rate;
-    u32 m_channel_count;
     soundtouch::SoundTouch m_sound_touch;
     double m_stretch_ratio = 1.0;
 };

src/common/logging/backend.cpp

@@ -12,7 +12,8 @@
 #include <thread>
 #include <vector>
 #ifdef _WIN32
-#include <share.h> // For _SH_DENYWR
+#include <share.h>   // For _SH_DENYWR
+#include <windows.h> // For OutputDebugStringA
 #else
 #define _SH_DENYWR 0
 #endif
@@ -139,12 +140,18 @@ void FileBackend::Write(const Entry& entry) {
     if (!file.IsOpen() || bytes_written > MAX_BYTES_WRITTEN) {
         return;
     }
-    bytes_written += file.WriteString(FormatLogMessage(entry) + '\n');
+    bytes_written += file.WriteString(FormatLogMessage(entry).append(1, '\n'));
     if (entry.log_level >= Level::Error) {
         file.Flush();
     }
 }
 
+void DebuggerBackend::Write(const Entry& entry) {
+#ifdef _WIN32
+    ::OutputDebugStringA(FormatLogMessage(entry).append(1, '\n').c_str());
+#endif
+}
+
 /// Macro listing all log classes. Code should define CLS and SUB as desired before invoking this.
 #define ALL_LOG_CLASSES()                                                                          \
     CLS(Log)                                                                                       \
@@ -196,6 +203,7 @@ void FileBackend::Write(const Entry& entry) {
     SUB(Service, NFP)                                                                              \
     SUB(Service, NIFM)                                                                             \
     SUB(Service, NIM)                                                                              \
+    SUB(Service, NPNS)                                                                             \
     SUB(Service, NS)                                                                               \
     SUB(Service, NVDRV)                                                                            \
     SUB(Service, PCIE)                                                                             \
@@ -204,10 +212,12 @@ void FileBackend::Write(const Entry& entry) {
     SUB(Service, PM)                                                                               \
     SUB(Service, PREPO)                                                                            \
     SUB(Service, PSC)                                                                              \
+    SUB(Service, PSM)                                                                              \
     SUB(Service, SET)                                                                              \
     SUB(Service, SM)                                                                               \
     SUB(Service, SPL)                                                                              \
     SUB(Service, SSL)                                                                              \
+    SUB(Service, TCAP)                                                                             \
     SUB(Service, Time)                                                                             \
     SUB(Service, USB)                                                                              \
     SUB(Service, VI)                                                                               \

src/common/logging/backend.h

@@ -103,6 +103,20 @@ private:
     std::size_t bytes_written;
 };
 
+/**
+ * Backend that writes to Visual Studio's output window
+ */
+class DebuggerBackend : public Backend {
+public:
+    static const char* Name() {
+        return "debugger";
+    }
+    const char* GetName() const override {
+        return Name();
+    }
+    void Write(const Entry& entry) override;
+};
+
 void AddBackend(std::unique_ptr<Backend> backend);
 
 void RemoveBackend(std::string_view backend_name);

src/common/logging/log.h

@@ -83,6 +83,7 @@ enum class Class : ClassType {
     Service_NFP,       ///< The NFP service
     Service_NIFM,      ///< The NIFM (Network interface) service
     Service_NIM,       ///< The NIM service
+    Service_NPNS,      ///< The NPNS service
     Service_NS,        ///< The NS services
     Service_NVDRV,     ///< The NVDRV (Nvidia driver) service
     Service_PCIE,      ///< The PCIe service
@@ -96,6 +97,7 @@ enum class Class : ClassType {
     Service_SM,        ///< The SM (Service manager) service
     Service_SPL,       ///< The SPL service
     Service_SSL,       ///< The SSL service
+    Service_TCAP,      ///< The TCAP service.
     Service_Time,      ///< The time service
     Service_USB,       ///< The USB (Universal Serial Bus) service
     Service_VI,        ///< The VI (Video interface) service

src/common/telemetry.h

@@ -153,6 +153,7 @@ struct VisitorInterface : NonCopyable {
 
     /// Completion method, called once all fields have been visited
     virtual void Complete() = 0;
+    virtual bool SubmitTestcase() = 0;
 };
 
 /**
@@ -178,6 +179,9 @@ struct NullVisitor : public VisitorInterface {
     void Visit(const Field<std::chrono::microseconds>& /*field*/) override {}
 
     void Complete() override {}
+    bool SubmitTestcase() override {
+        return false;
+    }
 };
 
 /// Appends build-specific information to the given FieldCollection,

src/core/core.cpp

@@ -185,7 +185,7 @@ struct System::Impl {
             LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath);
             return ResultStatus::ErrorGetLoader;
         }
-        std::pair<boost::optional<u32>, Loader::ResultStatus> system_mode =
+        std::pair<std::optional<u32>, Loader::ResultStatus> system_mode =
             app_loader->LoadKernelSystemMode();
 
         if (system_mode.second != Loader::ResultStatus::Success) {
@@ -312,6 +312,10 @@ Cpu& System::CurrentCpuCore() {
     return impl->CurrentCpuCore();
 }
 
+const Cpu& System::CurrentCpuCore() const {
+    return impl->CurrentCpuCore();
+}
+
 System::ResultStatus System::RunLoop(bool tight_loop) {
     return impl->RunLoop(tight_loop);
 }
@@ -342,7 +346,11 @@ PerfStatsResults System::GetAndResetPerfStats() {
     return impl->GetAndResetPerfStats();
 }
 
-Core::TelemetrySession& System::TelemetrySession() const {
+TelemetrySession& System::TelemetrySession() {
+    return *impl->telemetry_session;
+}
+
+const TelemetrySession& System::TelemetrySession() const {
     return *impl->telemetry_session;
 }
 
@@ -350,7 +358,11 @@ ARM_Interface& System::CurrentArmInterface() {
     return CurrentCpuCore().ArmInterface();
 }
 
-std::size_t System::CurrentCoreIndex() {
+const ARM_Interface& System::CurrentArmInterface() const {
+    return CurrentCpuCore().ArmInterface();
+}
+
+std::size_t System::CurrentCoreIndex() const {
     return CurrentCpuCore().CoreIndex();
 }
 
@@ -358,6 +370,10 @@ Kernel::Scheduler& System::CurrentScheduler() {
     return CurrentCpuCore().Scheduler();
 }
 
+const Kernel::Scheduler& System::CurrentScheduler() const {
+    return CurrentCpuCore().Scheduler();
+}
+
 Kernel::Scheduler& System::Scheduler(std::size_t core_index) {
     return CpuCore(core_index).Scheduler();
 }
@@ -378,6 +394,10 @@ ARM_Interface& System::ArmInterface(std::size_t core_index) {
     return CpuCore(core_index).ArmInterface();
 }
 
+const ARM_Interface& System::ArmInterface(std::size_t core_index) const {
+    return CpuCore(core_index).ArmInterface();
+}
+
 Cpu& System::CpuCore(std::size_t core_index) {
     ASSERT(core_index < NUM_CPU_CORES);
     return *impl->cpu_cores[core_index];
@@ -392,6 +412,10 @@ ExclusiveMonitor& System::Monitor() {
     return *impl->cpu_exclusive_monitor;
 }
 
+const ExclusiveMonitor& System::Monitor() const {
+    return *impl->cpu_exclusive_monitor;
+}
+
 Tegra::GPU& System::GPU() {
     return *impl->gpu_core;
 }

src/core/core.h

@@ -129,11 +129,11 @@ public:
      */
     bool IsPoweredOn() const;
 
-    /**
-     * Returns a reference to the telemetry session for this emulation session.
-     * @returns Reference to the telemetry session.
-     */
-    Core::TelemetrySession& TelemetrySession() const;
+    /// Gets a reference to the telemetry session for this emulation session.
+    Core::TelemetrySession& TelemetrySession();
+
+    /// Gets a reference to the telemetry session for this emulation session.
+    const Core::TelemetrySession& TelemetrySession() const;
 
     /// Prepare the core emulation for a reschedule
     void PrepareReschedule();
@@ -144,24 +144,36 @@ public:
     /// Gets an ARM interface to the CPU core that is currently running
     ARM_Interface& CurrentArmInterface();
 
+    /// Gets an ARM interface to the CPU core that is currently running
+    const ARM_Interface& CurrentArmInterface() const;
+
     /// Gets the index of the currently running CPU core
-    std::size_t CurrentCoreIndex();
+    std::size_t CurrentCoreIndex() const;
 
     /// Gets the scheduler for the CPU core that is currently running
     Kernel::Scheduler& CurrentScheduler();
 
-    /// Gets an ARM interface to the CPU core with the specified index
+    /// Gets the scheduler for the CPU core that is currently running
+    const Kernel::Scheduler& CurrentScheduler() const;
+
+    /// Gets a reference to an ARM interface for the CPU core with the specified index
     ARM_Interface& ArmInterface(std::size_t core_index);
 
+    /// Gets a const reference to an ARM interface from the CPU core with the specified index
+    const ARM_Interface& ArmInterface(std::size_t core_index) const;
+
     /// Gets a CPU interface to the CPU core with the specified index
     Cpu& CpuCore(std::size_t core_index);
 
     /// Gets a CPU interface to the CPU core with the specified index
     const Cpu& CpuCore(std::size_t core_index) const;
 
-    /// Gets the exclusive monitor
+    /// Gets a reference to the exclusive monitor
     ExclusiveMonitor& Monitor();
 
+    /// Gets a constant reference to the exclusive monitor
+    const ExclusiveMonitor& Monitor() const;
+
     /// Gets a mutable reference to the GPU interface
     Tegra::GPU& GPU();
 
@@ -230,6 +242,9 @@ private:
     /// Returns the currently running CPU core
     Cpu& CurrentCpuCore();
 
+    /// Returns the currently running CPU core
+    const Cpu& CurrentCpuCore() const;
+
     /**
      * Initialize the emulated system.
      * @param emu_window Reference to the host-system window used for video output and keyboard

src/core/crypto/key_manager.cpp

@@ -141,28 +141,28 @@ Key128 DeriveKeyblobMACKey(const Key128& keyblob_key, const Key128& mac_source)
     return mac_key;
 }
 
-boost::optional<Key128> DeriveSDSeed() {
+std::optional<Key128> DeriveSDSeed() {
     const FileUtil::IOFile save_43(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
                                        "/system/save/8000000000000043",
                                    "rb+");
     if (!save_43.IsOpen())
-        return boost::none;
+        return {};
 
     const FileUtil::IOFile sd_private(
         FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir) + "/Nintendo/Contents/private", "rb+");
     if (!sd_private.IsOpen())
-        return boost::none;
+        return {};
 
     std::array<u8, 0x10> private_seed{};
     if (sd_private.ReadBytes(private_seed.data(), private_seed.size()) != private_seed.size()) {
-        return boost::none;
+        return {};
     }
 
     std::array<u8, 0x10> buffer{};
     std::size_t offset = 0;
     for (; offset + 0x10 < save_43.GetSize(); ++offset) {
         if (!save_43.Seek(offset, SEEK_SET)) {
-            return boost::none;
+            return {};
         }
 
         save_43.ReadBytes(buffer.data(), buffer.size());
@@ -172,12 +172,12 @@ boost::optional<Key128> DeriveSDSeed() {
     }
 
     if (!save_43.Seek(offset + 0x10, SEEK_SET)) {
-        return boost::none;
+        return {};
     }
 
     Key128 seed{};
     if (save_43.ReadBytes(seed.data(), seed.size()) != seed.size()) {
-        return boost::none;
+        return {};
     }
     return seed;
 }
@@ -291,26 +291,26 @@ static std::array<u8, target_size> MGF1(const std::array<u8, in_size>& seed) {
 }
 
 template <size_t size>
-static boost::optional<u64> FindTicketOffset(const std::array<u8, size>& data) {
+static std::optional<u64> FindTicketOffset(const std::array<u8, size>& data) {
     u64 offset = 0;
     for (size_t i = 0x20; i < data.size() - 0x10; ++i) {
         if (data[i] == 0x1) {
             offset = i + 1;
             break;
         } else if (data[i] != 0x0) {
-            return boost::none;
+            return {};
         }
     }
 
     return offset;
 }
 
-boost::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
-                                                       const RSAKeyPair<2048>& key) {
+std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
+                                                     const RSAKeyPair<2048>& key) {
     u32 cert_authority;
     std::memcpy(&cert_authority, ticket.data() + 0x140, sizeof(cert_authority));
     if (cert_authority == 0)
-        return boost::none;
+        return {};
     if (cert_authority != Common::MakeMagic('R', 'o', 'o', 't')) {
         LOG_INFO(Crypto,
                  "Attempting to parse ticket with non-standard certificate authority {:08X}.",
@@ -321,7 +321,7 @@ boost::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
     std::memcpy(rights_id.data(), ticket.data() + 0x2A0, sizeof(Key128));
 
     if (rights_id == Key128{})
-        return boost::none;
+        return {};
 
     Key128 key_temp{};
 
@@ -356,17 +356,17 @@ boost::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
     std::memcpy(m_2.data(), rsa_step.data() + 0x21, m_2.size());
 
     if (m_0 != 0)
-        return boost::none;
+        return {};
 
     m_1 = m_1 ^ MGF1<0x20>(m_2);
     m_2 = m_2 ^ MGF1<0xDF>(m_1);
 
     const auto offset = FindTicketOffset(m_2);
-    if (offset == boost::none)
-        return boost::none;
-    ASSERT(offset.get() > 0);
+    if (!offset)
+        return {};
+    ASSERT(*offset > 0);
 
-    std::memcpy(key_temp.data(), m_2.data() + offset.get(), key_temp.size());
+    std::memcpy(key_temp.data(), m_2.data() + *offset, key_temp.size());
 
     return std::make_pair(rights_id, key_temp);
 }
@@ -395,7 +395,7 @@ static bool ValidCryptoRevisionString(std::string_view base, size_t begin, size_
     if (base.size() < begin + length)
         return false;
     return std::all_of(base.begin() + begin, base.begin() + begin + length,
-                       [](u8 c) { return std::isdigit(c); });
+                       [](u8 c) { return std::isxdigit(c); });
 }
 
 void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {
@@ -661,8 +661,8 @@ void KeyManager::DeriveSDSeedLazy() {
         return;
 
     const auto res = DeriveSDSeed();
-    if (res != boost::none)
-        SetKey(S128KeyType::SDSeed, res.get());
+    if (res)
+        SetKey(S128KeyType::SDSeed, *res);
 }
 
 static Key128 CalculateCMAC(const u8* source, size_t size, const Key128& key) {
@@ -713,7 +713,6 @@ void KeyManager::DeriveBase() {
 
     const auto sbk = GetKey(S128KeyType::SecureBoot);
     const auto tsec = GetKey(S128KeyType::TSEC);
-    const auto master_source = GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::Master));
 
     for (size_t i = 0; i < revisions.size(); ++i) {
         if (!revisions[i])
@@ -890,9 +889,9 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) {
 
     for (const auto& raw : res) {
         const auto pair = ParseTicket(raw, rsa_key);
-        if (pair == boost::none)
+        if (!pair)
             continue;
-        const auto& [rid, key] = pair.value();
+        const auto& [rid, key] = *pair;
         u128 rights_id;
         std::memcpy(rights_id.data(), rid.data(), rid.size());
         SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);

src/core/crypto/key_manager.h

@@ -6,9 +6,10 @@
 
 #include <array>
 #include <map>
+#include <optional>
 #include <string>
+
 #include <boost/container/flat_map.hpp>
-#include <boost/optional.hpp>
 #include <fmt/format.h>
 #include "common/common_types.h"
 #include "core/crypto/partition_data_manager.h"
@@ -191,14 +192,14 @@ Key128 DeriveMasterKey(const std::array<u8, 0x90>& keyblob, const Key128& master
 std::array<u8, 0x90> DecryptKeyblob(const std::array<u8, 0xB0>& encrypted_keyblob,
                                     const Key128& key);
 
-boost::optional<Key128> DeriveSDSeed();
+std::optional<Key128> DeriveSDSeed();
 Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& keys);
 
 std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save);
 
 // Returns a pair of {rights_id, titlekey}. Fails if the ticket has no certificate authority (offset
 // 0x140-0x144 is zero)
-boost::optional<std::pair<Key128, Key128>> ParseTicket(
-    const TicketRaw& ticket, const RSAKeyPair<2048>& eticket_extended_key);
+std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
+                                                     const RSAKeyPair<2048>& eticket_extended_key);
 
 } // namespace Core::Crypto

src/core/file_sys/card_image.cpp

@@ -168,10 +168,6 @@ VirtualDir XCI::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool XCI::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
-
 Loader::ResultStatus XCI::AddNCAFromPartition(XCIPartition part) {
     if (partitions[static_cast<std::size_t>(part)] == nullptr) {
         return Loader::ResultStatus::ErrorXCIMissingPartition;

src/core/file_sys/card_image.h

@@ -94,9 +94,6 @@ public:
 
     VirtualDir GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     Loader::ResultStatus AddNCAFromPartition(XCIPartition part);
 

src/core/file_sys/content_archive.cpp

@@ -4,10 +4,9 @@
 
 #include <algorithm>
 #include <cstring>
+#include <optional>
 #include <utility>
 
-#include <boost/optional.hpp>
-
 #include "common/logging/log.h"
 #include "core/crypto/aes_util.h"
 #include "core/crypto/ctr_encryption_layer.h"
@@ -306,18 +305,18 @@ bool NCA::ReadRomFSSection(const NCASectionHeader& section, const NCASectionTabl
         subsection_buckets.back().entries.push_back({section.bktr.relocation.offset, {0}, ctr_low});
         subsection_buckets.back().entries.push_back({size, {0}, 0});
 
-        boost::optional<Core::Crypto::Key128> key = boost::none;
+        std::optional<Core::Crypto::Key128> key = {};
         if (encrypted) {
             if (has_rights_id) {
                 status = Loader::ResultStatus::Success;
                 key = GetTitlekey();
-                if (key == boost::none) {
+                if (!key) {
                     status = Loader::ResultStatus::ErrorMissingTitlekey;
                     return false;
                 }
             } else {
                 key = GetKeyAreaKey(NCASectionCryptoType::BKTR);
-                if (key == boost::none) {
+                if (!key) {
                     status = Loader::ResultStatus::ErrorMissingKeyAreaKey;
                     return false;
                 }
@@ -332,7 +331,7 @@ bool NCA::ReadRomFSSection(const NCASectionHeader& section, const NCASectionTabl
         auto bktr = std::make_shared<BKTR>(
             bktr_base_romfs, std::make_shared<OffsetVfsFile>(file, romfs_size, base_offset),
             relocation_block, relocation_buckets, subsection_block, subsection_buckets, encrypted,
-            encrypted ? key.get() : Core::Crypto::Key128{}, base_offset, bktr_base_ivfc_offset,
+            encrypted ? *key : Core::Crypto::Key128{}, base_offset, bktr_base_ivfc_offset,
             section.raw.section_ctr);
 
         // BKTR applies to entire IVFC, so make an offset version to level 6
@@ -388,11 +387,11 @@ u8 NCA::GetCryptoRevision() const {
     return master_key_id;
 }
 
-boost::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const {
+std::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const {
     const auto master_key_id = GetCryptoRevision();
 
     if (!keys.HasKey(Core::Crypto::S128KeyType::KeyArea, master_key_id, header.key_index))
-        return boost::none;
+        return {};
 
     std::vector<u8> key_area(header.key_area.begin(), header.key_area.end());
     Core::Crypto::AESCipher<Core::Crypto::Key128> cipher(
@@ -416,25 +415,25 @@ boost::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType ty
     return out;
 }
 
-boost::optional<Core::Crypto::Key128> NCA::GetTitlekey() {
+std::optional<Core::Crypto::Key128> NCA::GetTitlekey() {
     const auto master_key_id = GetCryptoRevision();
 
     u128 rights_id{};
     memcpy(rights_id.data(), header.rights_id.data(), 16);
     if (rights_id == u128{}) {
         status = Loader::ResultStatus::ErrorInvalidRightsID;
-        return boost::none;
+        return {};
     }
 
     auto titlekey = keys.GetKey(Core::Crypto::S128KeyType::Titlekey, rights_id[1], rights_id[0]);
     if (titlekey == Core::Crypto::Key128{}) {
         status = Loader::ResultStatus::ErrorMissingTitlekey;
-        return boost::none;
+        return {};
     }
 
     if (!keys.HasKey(Core::Crypto::S128KeyType::Titlekek, master_key_id)) {
         status = Loader::ResultStatus::ErrorMissingTitlekek;
-        return boost::none;
+        return {};
     }
 
     Core::Crypto::AESCipher<Core::Crypto::Key128> cipher(
@@ -458,25 +457,25 @@ VirtualFile NCA::Decrypt(const NCASectionHeader& s_header, VirtualFile in, u64 s
     case NCASectionCryptoType::BKTR:
         LOG_DEBUG(Crypto, "called with mode=CTR, starting_offset={:016X}", starting_offset);
         {
-            boost::optional<Core::Crypto::Key128> key = boost::none;
+            std::optional<Core::Crypto::Key128> key = {};
             if (has_rights_id) {
                 status = Loader::ResultStatus::Success;
                 key = GetTitlekey();
-                if (key == boost::none) {
+                if (!key) {
                     if (status == Loader::ResultStatus::Success)
                         status = Loader::ResultStatus::ErrorMissingTitlekey;
                     return nullptr;
                 }
             } else {
                 key = GetKeyAreaKey(NCASectionCryptoType::CTR);
-                if (key == boost::none) {
+                if (!key) {
                     status = Loader::ResultStatus::ErrorMissingKeyAreaKey;
                     return nullptr;
                 }
             }
 
-            auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(
-                std::move(in), key.value(), starting_offset);
+            auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(std::move(in), *key,
+                                                                          starting_offset);
             std::vector<u8> iv(16);
             for (u8 i = 0; i < 8; ++i)
                 iv[i] = s_header.raw.section_ctr[0x8 - i - 1];
@@ -546,7 +545,4 @@ u64 NCA::GetBaseIVFCOffset() const {
     return ivfc_offset;
 }
 
-bool NCA::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys

src/core/file_sys/content_archive.h

@@ -6,9 +6,10 @@
 
 #include <array>
 #include <memory>
+#include <optional>
 #include <string>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/swap.h"
@@ -100,9 +101,6 @@ public:
     // Returns the base ivfc offset used in BKTR patching.
     u64 GetBaseIVFCOffset() const;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     bool CheckSupportedNCA(const NCAHeader& header);
     bool HandlePotentialHeaderDecryption();
@@ -114,8 +112,8 @@ private:
     bool ReadPFS0Section(const NCASectionHeader& section, const NCASectionTableEntry& entry);
 
     u8 GetCryptoRevision() const;
-    boost::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const;
-    boost::optional<Core::Crypto::Key128> GetTitlekey();
+    std::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const;
+    std::optional<Core::Crypto::Key128> GetTitlekey();
     VirtualFile Decrypt(const NCASectionHeader& header, VirtualFile in, u64 starting_offset);
 
     std::vector<VirtualDir> dirs;

src/core/file_sys/fsmitm_romfsbuild.h

@@ -27,7 +27,6 @@
 #include <map>
 #include <memory>
 #include <string>
-#include <boost/detail/container_fwd.hpp>
 #include "common/common_types.h"
 #include "core/file_sys/vfs.h"
 

src/core/file_sys/ips_layer.cpp

@@ -99,16 +99,16 @@ VirtualFile PatchIPS(const VirtualFile& in, const VirtualFile& ips) {
             u16 rle_size{};
             if (ips->ReadObject(&rle_size, offset) != sizeof(u16))
                 return nullptr;
-            rle_size = Common::swap16(data_size);
+            rle_size = Common::swap16(rle_size);
             offset += sizeof(u16);
 
             const auto data = ips->ReadByte(offset++);
-            if (data == boost::none)
+            if (!data)
                 return nullptr;
 
             if (real_offset + rle_size > in_data.size())
                 rle_size = static_cast<u16>(in_data.size() - real_offset);
-            std::memset(in_data.data() + real_offset, data.get(), rle_size);
+            std::memset(in_data.data() + real_offset, *data, rle_size);
         } else { // Standard Patch
             auto read = data_size;
             if (real_offset + read > in_data.size())

src/core/file_sys/partition_filesystem.cpp

@@ -83,7 +83,7 @@ std::vector<std::shared_ptr<VfsFile>> PartitionFilesystem::GetFiles() const {
 }
 
 std::vector<std::shared_ptr<VfsDirectory>> PartitionFilesystem::GetSubdirectories() const {
-    return pfs_dirs;
+    return {};
 }
 
 std::string PartitionFilesystem::GetName() const {
@@ -103,18 +103,4 @@ void PartitionFilesystem::PrintDebugInfo() const {
                   pfs_files[i]->GetName(), pfs_files[i]->GetSize());
     }
 }
-
-bool PartitionFilesystem::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    const auto iter = std::find(pfs_files.begin(), pfs_files.end(), file);
-    if (iter == pfs_files.end())
-        return false;
-
-    const std::ptrdiff_t offset = std::distance(pfs_files.begin(), iter);
-    pfs_files[offset] = std::move(pfs_files.back());
-    pfs_files.pop_back();
-
-    pfs_dirs.emplace_back(std::move(dir));
-
-    return true;
-}
 } // namespace FileSys

src/core/file_sys/partition_filesystem.h

@@ -35,9 +35,6 @@ public:
     std::shared_ptr<VfsDirectory> GetParentDirectory() const override;
     void PrintDebugInfo() const;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     struct Header {
         u32_le magic;
@@ -84,7 +81,6 @@ private:
     std::size_t content_offset = 0;
 
     std::vector<VirtualFile> pfs_files;
-    std::vector<VirtualDir> pfs_dirs;
 };
 
 } // namespace FileSys

src/core/file_sys/patch_manager.cpp

@@ -61,13 +61,12 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
     // Game Updates
     const auto update_tid = GetUpdateTitleID(title_id);
     const auto update = installed->GetEntry(update_tid, ContentRecordType::Program);
-    if (update != nullptr) {
-        if (update->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS &&
-            update->GetExeFS() != nullptr) {
-            LOG_INFO(Loader, "    ExeFS: Update ({}) applied successfully",
-                     FormatTitleVersion(installed->GetEntryVersion(update_tid).get_value_or(0)));
-            exefs = update->GetExeFS();
-        }
+
+    if (update != nullptr && update->GetExeFS() != nullptr &&
+        update->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) {
+        LOG_INFO(Loader, "    ExeFS: Update ({}) applied successfully",
+                 FormatTitleVersion(installed->GetEntryVersion(update_tid).value_or(0)));
+        exefs = update->GetExeFS();
     }
 
     return exefs;
@@ -237,7 +236,7 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content
         if (new_nca->GetStatus() == Loader::ResultStatus::Success &&
             new_nca->GetRomFS() != nullptr) {
             LOG_INFO(Loader, "    RomFS: Update ({}) applied successfully",
-                     FormatTitleVersion(installed->GetEntryVersion(update_tid).get_value_or(0)));
+                     FormatTitleVersion(installed->GetEntryVersion(update_tid).value_or(0)));
             romfs = new_nca->GetRomFS();
         }
     } else if (update_raw != nullptr) {
@@ -281,12 +280,11 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
     } else {
         if (installed->HasEntry(update_tid, ContentRecordType::Program)) {
             const auto meta_ver = installed->GetEntryVersion(update_tid);
-            if (meta_ver == boost::none || meta_ver.get() == 0) {
+            if (meta_ver.value_or(0) == 0) {
                 out.insert_or_assign("Update", "");
             } else {
                 out.insert_or_assign(
-                    "Update",
-                    FormatTitleVersion(meta_ver.get(), TitleVersionFormat::ThreeElements));
+                    "Update", FormatTitleVersion(*meta_ver, TitleVersionFormat::ThreeElements));
             }
         } else if (update_raw != nullptr) {
             out.insert_or_assign("Update", "PACKED");

src/core/file_sys/registered_cache.cpp

@@ -159,28 +159,28 @@ VirtualFile RegisteredCache::GetFileAtID(NcaID id) const {
     return file;
 }
 
-static boost::optional<NcaID> CheckMapForContentRecord(
+static std::optional<NcaID> CheckMapForContentRecord(
     const boost::container::flat_map<u64, CNMT>& map, u64 title_id, ContentRecordType type) {
     if (map.find(title_id) == map.end())
-        return boost::none;
+        return {};
 
     const auto& cnmt = map.at(title_id);
 
     const auto iter = std::find_if(cnmt.GetContentRecords().begin(), cnmt.GetContentRecords().end(),
                                    [type](const ContentRecord& rec) { return rec.type == type; });
     if (iter == cnmt.GetContentRecords().end())
-        return boost::none;
+        return {};
 
-    return boost::make_optional(iter->nca_id);
+    return std::make_optional(iter->nca_id);
 }
 
-boost::optional<NcaID> RegisteredCache::GetNcaIDFromMetadata(u64 title_id,
-                                                             ContentRecordType type) const {
+std::optional<NcaID> RegisteredCache::GetNcaIDFromMetadata(u64 title_id,
+                                                           ContentRecordType type) const {
     if (type == ContentRecordType::Meta && meta_id.find(title_id) != meta_id.end())
         return meta_id.at(title_id);
 
     const auto res1 = CheckMapForContentRecord(yuzu_meta, title_id, type);
-    if (res1 != boost::none)
+    if (res1)
         return res1;
     return CheckMapForContentRecord(meta, title_id, type);
 }
@@ -283,17 +283,14 @@ bool RegisteredCache::HasEntry(RegisteredCacheEntry entry) const {
 
 VirtualFile RegisteredCache::GetEntryUnparsed(u64 title_id, ContentRecordType type) const {
     const auto id = GetNcaIDFromMetadata(title_id, type);
-    if (id == boost::none)
-        return nullptr;
-
-    return GetFileAtID(id.get());
+    return id ? GetFileAtID(*id) : nullptr;
 }
 
 VirtualFile RegisteredCache::GetEntryUnparsed(RegisteredCacheEntry entry) const {
     return GetEntryUnparsed(entry.title_id, entry.type);
 }
 
-boost::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const {
+std::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const {
     const auto meta_iter = meta.find(title_id);
     if (meta_iter != meta.end())
         return meta_iter->second.GetTitleVersion();
@@ -302,15 +299,12 @@ boost::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const {
     if (yuzu_meta_iter != yuzu_meta.end())
         return yuzu_meta_iter->second.GetTitleVersion();
 
-    return boost::none;
+    return {};
 }
 
 VirtualFile RegisteredCache::GetEntryRaw(u64 title_id, ContentRecordType type) const {
     const auto id = GetNcaIDFromMetadata(title_id, type);
-    if (id == boost::none)
-        return nullptr;
-
-    return parser(GetFileAtID(id.get()), id.get());
+    return id ? parser(GetFileAtID(*id), *id) : nullptr;
 }
 
 VirtualFile RegisteredCache::GetEntryRaw(RegisteredCacheEntry entry) const {
@@ -364,8 +358,8 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntries() const {
 }
 
 std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter(
-    boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type,
-    boost::optional<u64> title_id) const {
+    std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type,
+    std::optional<u64> title_id) const {
     std::vector<RegisteredCacheEntry> out;
     IterateAllMetadata<RegisteredCacheEntry>(
         out,
@@ -373,11 +367,11 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter(
             return RegisteredCacheEntry{c.GetTitleID(), r.type};
         },
         [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) {
-            if (title_type != boost::none && title_type.get() != c.GetType())
+            if (title_type && *title_type != c.GetType())
                 return false;
-            if (record_type != boost::none && record_type.get() != r.type)
+            if (record_type && *record_type != r.type)
                 return false;
-            if (title_id != boost::none && title_id.get() != c.GetTitleID())
+            if (title_id && *title_id != c.GetTitleID())
                 return false;
             return true;
         });
@@ -459,7 +453,7 @@ InstallResult RegisteredCache::InstallEntry(std::shared_ptr<NCA> nca, TitleType
 
 InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy,
                                              bool overwrite_if_exists,
-                                             boost::optional<NcaID> override_id) {
+                                             std::optional<NcaID> override_id) {
     const auto in = nca->GetBaseFile();
     Core::Crypto::SHA256Hash hash{};
 
@@ -468,12 +462,12 @@ InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const Vfs
     // game is massive), we're going to cheat and only hash the first MB of the NCA.
     // Also, for XCIs the NcaID matters, so if the override id isn't none, use that.
     NcaID id{};
-    if (override_id == boost::none) {
+    if (override_id) {
+        id = *override_id;
+    } else {
         const auto& data = in->ReadBytes(0x100000);
         mbedtls_sha256(data.data(), data.size(), hash.data(), 0);
         memcpy(id.data(), hash.data(), 16);
-    } else {
-        id = override_id.get();
     }
 
     std::string path = GetRelativePathFromNcaID(id, false, true);
@@ -543,14 +537,14 @@ bool RegisteredCacheUnion::HasEntry(RegisteredCacheEntry entry) const {
     return HasEntry(entry.title_id, entry.type);
 }
 
-boost::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const {
+std::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const {
     for (const auto& c : caches) {
         const auto res = c->GetEntryVersion(title_id);
-        if (res != boost::none)
+        if (res)
             return res;
     }
 
-    return boost::none;
+    return {};
 }
 
 VirtualFile RegisteredCacheUnion::GetEntryUnparsed(u64 title_id, ContentRecordType type) const {
@@ -609,8 +603,8 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntries() const {
 }
 
 std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter(
-    boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type,
-    boost::optional<u64> title_id) const {
+    std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type,
+    std::optional<u64> title_id) const {
     std::vector<RegisteredCacheEntry> out;
     for (const auto& c : caches) {
         c->IterateAllMetadata<RegisteredCacheEntry>(
@@ -619,11 +613,11 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter(
                 return RegisteredCacheEntry{c.GetTitleID(), r.type};
             },
             [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) {
-                if (title_type != boost::none && title_type.get() != c.GetType())
+                if (title_type && *title_type != c.GetType())
                     return false;
-                if (record_type != boost::none && record_type.get() != r.type)
+                if (record_type && *record_type != r.type)
                     return false;
-                if (title_id != boost::none && title_id.get() != c.GetTitleID())
+                if (title_id && *title_id != c.GetTitleID())
                     return false;
                 return true;
             });

src/core/file_sys/registered_cache.h

@@ -84,7 +84,7 @@ public:
     bool HasEntry(u64 title_id, ContentRecordType type) const;
     bool HasEntry(RegisteredCacheEntry entry) const;
 
-    boost::optional<u32> GetEntryVersion(u64 title_id) const;
+    std::optional<u32> GetEntryVersion(u64 title_id) const;
 
     VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const;
     VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const;
@@ -96,11 +96,10 @@ public:
     std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const;
 
     std::vector<RegisteredCacheEntry> ListEntries() const;
-    // If a parameter is not boost::none, it will be filtered for from all entries.
+    // If a parameter is not std::nullopt, it will be filtered for from all entries.
     std::vector<RegisteredCacheEntry> ListEntriesFilter(
-        boost::optional<TitleType> title_type = boost::none,
-        boost::optional<ContentRecordType> record_type = boost::none,
-        boost::optional<u64> title_id = boost::none) const;
+        std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {},
+        std::optional<u64> title_id = {}) const;
 
     // Raw copies all the ncas from the xci/nsp to the csache. Does some quick checks to make sure
     // there is a meta NCA and all of them are accessible.
@@ -125,12 +124,11 @@ private:
     std::vector<NcaID> AccumulateFiles() const;
     void ProcessFiles(const std::vector<NcaID>& ids);
     void AccumulateYuzuMeta();
-    boost::optional<NcaID> GetNcaIDFromMetadata(u64 title_id, ContentRecordType type) const;
+    std::optional<NcaID> GetNcaIDFromMetadata(u64 title_id, ContentRecordType type) const;
     VirtualFile GetFileAtID(NcaID id) const;
     VirtualFile OpenFileOrDirectoryConcat(const VirtualDir& dir, std::string_view path) const;
     InstallResult RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy,
-                                bool overwrite_if_exists,
-                                boost::optional<NcaID> override_id = boost::none);
+                                bool overwrite_if_exists, std::optional<NcaID> override_id = {});
     bool RawInstallYuzuMeta(const CNMT& cnmt);
 
     VirtualDir dir;
@@ -153,7 +151,7 @@ public:
     bool HasEntry(u64 title_id, ContentRecordType type) const;
     bool HasEntry(RegisteredCacheEntry entry) const;
 
-    boost::optional<u32> GetEntryVersion(u64 title_id) const;
+    std::optional<u32> GetEntryVersion(u64 title_id) const;
 
     VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const;
     VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const;
@@ -165,11 +163,10 @@ public:
     std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const;
 
     std::vector<RegisteredCacheEntry> ListEntries() const;
-    // If a parameter is not boost::none, it will be filtered for from all entries.
+    // If a parameter is not std::nullopt, it will be filtered for from all entries.
     std::vector<RegisteredCacheEntry> ListEntriesFilter(
-        boost::optional<TitleType> title_type = boost::none,
-        boost::optional<ContentRecordType> record_type = boost::none,
-        boost::optional<u64> title_id = boost::none) const;
+        std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {},
+        std::optional<u64> title_id = {}) const;
 
 private:
     std::vector<RegisteredCache*> caches;

src/core/file_sys/submission_package.cpp

@@ -205,10 +205,6 @@ VirtualDir NSP::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool NSP::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
-
 void NSP::InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files) {
     exefs = pfs;
 

src/core/file_sys/submission_package.h

@@ -55,9 +55,6 @@ public:
 
     VirtualDir GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     void InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files);
     void ReadNCAs(const std::vector<VirtualFile>& files);

src/core/file_sys/vfs.cpp

@@ -167,13 +167,13 @@ std::string VfsFile::GetExtension() const {
 
 VfsDirectory::~VfsDirectory() = default;
 
-boost::optional<u8> VfsFile::ReadByte(std::size_t offset) const {
+std::optional<u8> VfsFile::ReadByte(std::size_t offset) const {
     u8 out{};
     std::size_t size = Read(&out, 1, offset);
     if (size == 1)
         return out;
 
-    return boost::none;
+    return {};
 }
 
 std::vector<u8> VfsFile::ReadBytes(std::size_t size, std::size_t offset) const {
@@ -472,10 +472,14 @@ bool VfsRawCopy(const VirtualFile& src, const VirtualFile& dest, std::size_t blo
     std::vector<u8> temp(std::min(block_size, src->GetSize()));
     for (std::size_t i = 0; i < src->GetSize(); i += block_size) {
         const auto read = std::min(block_size, src->GetSize() - i);
-        const auto block = src->Read(temp.data(), read, i);
 
-        if (dest->Write(temp.data(), read, i) != read)
+        if (src->Read(temp.data(), read, i) != read) {
             return false;
+        }
+
+        if (dest->Write(temp.data(), read, i) != read) {
+            return false;
+        }
     }
 
     return true;

src/core/file_sys/vfs.h

@@ -4,13 +4,15 @@
 
 #pragma once
 
+#include <functional>
 #include <map>
 #include <memory>
+#include <optional>
 #include <string>
 #include <string_view>
 #include <type_traits>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/common_types.h"
 #include "core/file_sys/vfs_types.h"
 
@@ -103,8 +105,8 @@ public:
     // into file. Returns number of bytes successfully written.
     virtual std::size_t Write(const u8* data, std::size_t length, std::size_t offset = 0) = 0;
 
-    // Reads exactly one byte at the offset provided, returning boost::none on error.
-    virtual boost::optional<u8> ReadByte(std::size_t offset = 0) const;
+    // Reads exactly one byte at the offset provided, returning std::nullopt on error.
+    virtual std::optional<u8> ReadByte(std::size_t offset = 0) const;
     // Reads size bytes starting at offset in file into a vector.
     virtual std::vector<u8> ReadBytes(std::size_t size, std::size_t offset = 0) const;
     // Reads all the bytes from the file into a vector. Equivalent to 'file->Read(file->GetSize(),
@@ -262,36 +264,8 @@ public:
     // item name -> type.
     virtual std::map<std::string, VfsEntryType, std::less<>> GetEntries() const;
 
-    // Interprets the file with name file instead as a directory of type directory.
-    // The directory must have a constructor that takes a single argument of type
-    // std::shared_ptr<VfsFile>. Allows to reinterpret container files (i.e NCA, zip, XCI, etc) as a
-    // subdirectory in one call.
-    template <typename Directory>
-    bool InterpretAsDirectory(std::string_view file) {
-        auto file_p = GetFile(file);
-
-        if (file_p == nullptr) {
-            return false;
-        }
-
-        return ReplaceFileWithSubdirectory(file_p, std::make_shared<Directory>(file_p));
-    }
-
-    bool InterpretAsDirectory(const std::function<VirtualDir(VirtualFile)>& function,
-                              const std::string& file) {
-        auto file_p = GetFile(file);
-        if (file_p == nullptr)
-            return false;
-        return ReplaceFileWithSubdirectory(file_p, function(file_p));
-    }
-
     // Returns the full path of this directory as a string, recursively
     virtual std::string GetFullPath() const;
-
-protected:
-    // Backend for InterpretAsDirectory.
-    // Removes all references to file and adds a reference to dir in the directory's implementation.
-    virtual bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) = 0;
 };
 
 // A convenience partial-implementation of VfsDirectory that stubs out methods that should only work

src/core/file_sys/vfs_layered.cpp

@@ -126,7 +126,4 @@ bool LayeredVfsDirectory::Rename(std::string_view name_) {
     return true;
 }
 
-bool LayeredVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys

src/core/file_sys/vfs_layered.h

@@ -39,9 +39,6 @@ public:
     bool DeleteFile(std::string_view name) override;
     bool Rename(std::string_view name) override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     std::vector<VirtualDir> dirs;
     std::string name;

src/core/file_sys/vfs_offset.cpp

@@ -57,11 +57,11 @@ std::size_t OffsetVfsFile::Write(const u8* data, std::size_t length, std::size_t
     return file->Write(data, TrimToFit(length, r_offset), offset + r_offset);
 }
 
-boost::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const {
+std::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const {
     if (r_offset < size)
         return file->ReadByte(offset + r_offset);
 
-    return boost::none;
+    return {};
 }
 
 std::vector<u8> OffsetVfsFile::ReadBytes(std::size_t r_size, std::size_t r_offset) const {

src/core/file_sys/vfs_offset.h

@@ -29,7 +29,7 @@ public:
     bool IsReadable() const override;
     std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override;
     std::size_t Write(const u8* data, std::size_t length, std::size_t offset) override;
-    boost::optional<u8> ReadByte(std::size_t offset) const override;
+    std::optional<u8> ReadByte(std::size_t offset) const override;
     std::vector<u8> ReadBytes(std::size_t size, std::size_t offset) const override;
     std::vector<u8> ReadAllBytes() const override;
     bool WriteByte(u8 data, std::size_t offset) override;

src/core/file_sys/vfs_real.cpp

@@ -430,7 +430,4 @@ std::map<std::string, VfsEntryType, std::less<>> RealVfsDirectory::GetEntries()
     return out;
 }
 
-bool RealVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys

src/core/file_sys/vfs_real.h

@@ -100,9 +100,6 @@ public:
     std::string GetFullPath() const override;
     std::map<std::string, VfsEntryType, std::less<>> GetEntries() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     RealVfsDirectory(RealVfsFilesystem& base, const std::string& path, Mode perms = Mode::Read);
 

src/core/file_sys/vfs_static.h

@@ -53,10 +53,10 @@ public:
         return 0;
     }
 
-    boost::optional<u8> ReadByte(std::size_t offset) const override {
+    std::optional<u8> ReadByte(std::size_t offset) const override {
         if (offset < size)
             return value;
-        return boost::none;
+        return {};
     }
 
     std::vector<u8> ReadBytes(std::size_t length, std::size_t offset) const override {

src/core/file_sys/vfs_vector.cpp

@@ -132,11 +132,4 @@ void VectorVfsDirectory::AddFile(VirtualFile file) {
 void VectorVfsDirectory::AddDirectory(VirtualDir dir) {
     dirs.push_back(std::move(dir));
 }
-
-bool VectorVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    if (!DeleteFile(file->GetName()))
-        return false;
-    dirs.emplace_back(std::move(dir));
-    return true;
-}
 } // namespace FileSys

src/core/file_sys/vfs_vector.h

@@ -57,9 +57,6 @@ public:
     virtual void AddFile(VirtualFile file);
     virtual void AddDirectory(VirtualDir dir);
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     std::vector<VirtualFile> files;
     std::vector<VirtualDir> dirs;

src/core/file_sys/xts_archive.cpp

@@ -163,7 +163,4 @@ std::shared_ptr<VfsDirectory> NAX::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool NAX::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys

src/core/file_sys/xts_archive.h

@@ -51,9 +51,6 @@ public:
 
     std::shared_ptr<VfsDirectory> GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     Loader::ResultStatus Parse(std::string_view path);
 

src/core/hle/ipc.h

@@ -14,11 +14,6 @@ namespace IPC {
 /// Size of the command buffer area, in 32-bit words.
 constexpr std::size_t COMMAND_BUFFER_LENGTH = 0x100 / sizeof(u32);
 
-// These errors are commonly returned by invalid IPC translations, so alias them here for
-// convenience.
-// TODO(yuriks): These will probably go away once translation is implemented inside the kernel.
-constexpr auto ERR_INVALID_HANDLE = Kernel::ERR_INVALID_HANDLE_OS;
-
 enum class ControlCommand : u32 {
     ConvertSessionToDomain = 0,
     ConvertDomainToSession = 1,

src/core/hle/ipc_helpers.h

@@ -117,8 +117,7 @@ public:
 
         AlignWithPadding();
 
-        const bool request_has_domain_header{context.GetDomainMessageHeader() != nullptr};
-        if (context.Session()->IsDomain() && request_has_domain_header) {
+        if (context.Session()->IsDomain() && context.HasDomainMessageHeader()) {
             IPC::DomainMessageHeader domain_header{};
             domain_header.num_objects = num_domain_objects;
             PushRaw(domain_header);

src/core/hle/kernel/errors.h

@@ -10,11 +10,6 @@ namespace Kernel {
 
 namespace ErrCodes {
 enum {
-    // TODO(Subv): Remove these 3DS OS error codes.
-    SessionClosedByRemote = 26,
-    NoPendingSessions = 35,
-    InvalidBufferDescriptor = 48,
-
     // Confirmed Switch OS error codes
     MaxConnectionsReached = 7,
     InvalidSize = 101,
@@ -26,6 +21,7 @@ enum {
     InvalidThreadPriority = 112,
     InvalidProcessorId = 113,
     InvalidHandle = 114,
+    InvalidPointer = 115,
     InvalidCombination = 116,
     Timeout = 117,
     SynchronizationCanceled = 118,
@@ -33,6 +29,7 @@ enum {
     InvalidEnumValue = 120,
     NoSuchEntry = 121,
     AlreadyRegistered = 122,
+    SessionClosed = 123,
     InvalidState = 125,
     ResourceLimitExceeded = 132,
 };
@@ -41,18 +38,14 @@ enum {
 // WARNING: The kernel is quite inconsistent in it's usage of errors code. Make sure to always
 // double check that the code matches before re-using the constant.
 
-// TODO(bunnei): Replace -1 with correct errors for Switch OS
 constexpr ResultCode ERR_HANDLE_TABLE_FULL(ErrorModule::Kernel, ErrCodes::HandleTableFull);
-constexpr ResultCode ERR_SESSION_CLOSED_BY_REMOTE(-1);
+constexpr ResultCode ERR_SESSION_CLOSED_BY_REMOTE(ErrorModule::Kernel, ErrCodes::SessionClosed);
 constexpr ResultCode ERR_PORT_NAME_TOO_LONG(ErrorModule::Kernel, ErrCodes::TooLarge);
 constexpr ResultCode ERR_MAX_CONNECTIONS_REACHED(ErrorModule::Kernel,
                                                  ErrCodes::MaxConnectionsReached);
 constexpr ResultCode ERR_INVALID_ENUM_VALUE(ErrorModule::Kernel, ErrCodes::InvalidEnumValue);
-constexpr ResultCode ERR_INVALID_ENUM_VALUE_FND(-1);
-constexpr ResultCode ERR_INVALID_COMBINATION(-1);
 constexpr ResultCode ERR_INVALID_COMBINATION_KERNEL(ErrorModule::Kernel,
                                                     ErrCodes::InvalidCombination);
-constexpr ResultCode ERR_OUT_OF_MEMORY(-1);
 constexpr ResultCode ERR_INVALID_ADDRESS(ErrorModule::Kernel, ErrCodes::InvalidAddress);
 constexpr ResultCode ERR_INVALID_ADDRESS_STATE(ErrorModule::Kernel, ErrCodes::InvalidMemoryState);
 constexpr ResultCode ERR_INVALID_MEMORY_PERMISSIONS(ErrorModule::Kernel,
@@ -65,14 +58,8 @@ constexpr ResultCode ERR_ALREADY_REGISTERED(ErrorModule::Kernel, ErrCodes::Alrea
 constexpr ResultCode ERR_INVALID_STATE(ErrorModule::Kernel, ErrCodes::InvalidState);
 constexpr ResultCode ERR_INVALID_THREAD_PRIORITY(ErrorModule::Kernel,
                                                  ErrCodes::InvalidThreadPriority);
-constexpr ResultCode ERR_INVALID_POINTER(-1);
-constexpr ResultCode ERR_INVALID_OBJECT_ADDR(-1);
-constexpr ResultCode ERR_NOT_AUTHORIZED(-1);
-/// Alternate code returned instead of ERR_INVALID_HANDLE in some code paths.
-constexpr ResultCode ERR_INVALID_HANDLE_OS(-1);
+constexpr ResultCode ERR_INVALID_POINTER(ErrorModule::Kernel, ErrCodes::InvalidPointer);
 constexpr ResultCode ERR_NOT_FOUND(ErrorModule::Kernel, ErrCodes::NoSuchEntry);
 constexpr ResultCode RESULT_TIMEOUT(ErrorModule::Kernel, ErrCodes::Timeout);
-/// Returned when Accept() is called on a port with no sessions to be accepted.
-constexpr ResultCode ERR_NO_PENDING_SESSIONS(-1);
 
 } // namespace Kernel

src/core/hle/kernel/hle_ipc.h

@@ -161,8 +161,12 @@ public:
         return buffer_c_desciptors;
     }
 
-    const std::shared_ptr<IPC::DomainMessageHeader>& GetDomainMessageHeader() const {
-        return domain_message_header;
+    const IPC::DomainMessageHeader* GetDomainMessageHeader() const {
+        return domain_message_header.get();
+    }
+
+    bool HasDomainMessageHeader() const {
+        return domain_message_header != nullptr;
     }
 
     /// Helper function to read a buffer using the appropriate buffer descriptor

src/core/hle/kernel/kernel.cpp

@@ -32,7 +32,7 @@ namespace Kernel {
  */
 static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_late) {
     const auto proper_handle = static_cast<Handle>(thread_handle);
-    auto& system = Core::System::GetInstance();
+    const auto& system = Core::System::GetInstance();
 
     // Lock the global kernel mutex when we enter the kernel HLE.
     std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
@@ -90,7 +90,7 @@ static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_
 /// The timer callback event, called when a timer is fired
 static void TimerCallback(u64 timer_handle, int cycles_late) {
     const auto proper_handle = static_cast<Handle>(timer_handle);
-    auto& system = Core::System::GetInstance();
+    const auto& system = Core::System::GetInstance();
     SharedPtr<Timer> timer = system.Kernel().RetrieveTimerFromCallbackHandleTable(proper_handle);
 
     if (timer == nullptr) {

src/core/hle/kernel/mutex.cpp

@@ -6,8 +6,6 @@
 #include <utility>
 #include <vector>
 
-#include <boost/range/algorithm_ext/erase.hpp>
-
 #include "common/assert.h"
 #include "core/core.h"
 #include "core/hle/kernel/errors.h"

src/core/hle/kernel/process.cpp

@@ -232,6 +232,12 @@ void Process::LoadModule(CodeSet module_, VAddr base_addr) {
     MapSegment(module_.CodeSegment(), VMAPermission::ReadExecute, MemoryState::CodeStatic);
     MapSegment(module_.RODataSegment(), VMAPermission::Read, MemoryState::CodeMutable);
     MapSegment(module_.DataSegment(), VMAPermission::ReadWrite, MemoryState::CodeMutable);
+
+    // Clear instruction cache in CPU JIT
+    Core::System::GetInstance().ArmInterface(0).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(1).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(2).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(3).ClearInstructionCache();
 }
 
 ResultVal<VAddr> Process::HeapAllocate(VAddr target, u64 size, VMAPermission perms) {

src/core/hle/kernel/process.h

@@ -202,6 +202,16 @@ public:
         return is_64bit_process;
     }
 
+    /// Gets the total running time of the process instance in ticks.
+    u64 GetCPUTimeTicks() const {
+        return total_process_running_time_ticks;
+    }
+
+    /// Updates the total running time, adding the given ticks to it.
+    void UpdateCPUTimeTicks(u64 ticks) {
+        total_process_running_time_ticks += ticks;
+    }
+
     /**
      * Loads process-specifics configuration info with metadata provided
      * by an executable.
@@ -305,6 +315,9 @@ private:
     /// specified by metadata provided to the process during loading.
     bool is_64bit_process = true;
 
+    /// Total running time for the process in ticks.
+    u64 total_process_running_time_ticks = 0;
+
     /// Per-process handle table for storing created object handles in.
     HandleTable handle_table;
 

src/core/hle/kernel/scheduler.cpp

@@ -9,6 +9,7 @@
 #include "common/logging/log.h"
 #include "core/arm/arm_interface.h"
 #include "core/core.h"
+#include "core/core_timing.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/scheduler.h"
@@ -34,6 +35,10 @@ Thread* Scheduler::GetCurrentThread() const {
     return current_thread.get();
 }
 
+u64 Scheduler::GetLastContextSwitchTicks() const {
+    return last_context_switch_time;
+}
+
 Thread* Scheduler::PopNextReadyThread() {
     Thread* next = nullptr;
     Thread* thread = GetCurrentThread();
@@ -54,7 +59,10 @@ Thread* Scheduler::PopNextReadyThread() {
 }
 
 void Scheduler::SwitchContext(Thread* new_thread) {
-    Thread* previous_thread = GetCurrentThread();
+    Thread* const previous_thread = GetCurrentThread();
+    Process* const previous_process = Core::CurrentProcess();
+
+    UpdateLastContextSwitchTime(previous_thread, previous_process);
 
     // Save context for previous thread
     if (previous_thread) {
@@ -78,8 +86,6 @@ void Scheduler::SwitchContext(Thread* new_thread) {
         // Cancel any outstanding wakeup events for this thread
         new_thread->CancelWakeupTimer();
 
-        auto* const previous_process = Core::CurrentProcess();
-
         current_thread = new_thread;
 
         ready_queue.remove(new_thread->GetPriority(), new_thread);
@@ -102,6 +108,22 @@ void Scheduler::SwitchContext(Thread* new_thread) {
     }
 }
 
+void Scheduler::UpdateLastContextSwitchTime(Thread* thread, Process* process) {
+    const u64 prev_switch_ticks = last_context_switch_time;
+    const u64 most_recent_switch_ticks = CoreTiming::GetTicks();
+    const u64 update_ticks = most_recent_switch_ticks - prev_switch_ticks;
+
+    if (thread != nullptr) {
+        thread->UpdateCPUTimeTicks(update_ticks);
+    }
+
+    if (process != nullptr) {
+        process->UpdateCPUTimeTicks(update_ticks);
+    }
+
+    last_context_switch_time = most_recent_switch_ticks;
+}
+
 void Scheduler::Reschedule() {
     std::lock_guard<std::mutex> lock(scheduler_mutex);
 

src/core/hle/kernel/scheduler.h

@@ -17,6 +17,8 @@ class ARM_Interface;
 
 namespace Kernel {
 
+class Process;
+
 class Scheduler final {
 public:
     explicit Scheduler(Core::ARM_Interface& cpu_core);
@@ -31,6 +33,9 @@ public:
     /// Gets the current running thread
     Thread* GetCurrentThread() const;
 
+    /// Gets the timestamp for the last context switch in ticks.
+    u64 GetLastContextSwitchTicks() const;
+
     /// Adds a new thread to the scheduler
     void AddThread(SharedPtr<Thread> thread, u32 priority);
 
@@ -64,6 +69,19 @@ private:
      */
     void SwitchContext(Thread* new_thread);
 
+    /**
+     * Called on every context switch to update the internal timestamp
+     * This also updates the running time ticks for the given thread and
+     * process using the following difference:
+     *
+     * ticks += most_recent_ticks - last_context_switch_ticks
+     *
+     * The internal tick timestamp for the scheduler is simply the
+     * most recent tick count retrieved. No special arithmetic is
+     * applied to it.
+     */
+    void UpdateLastContextSwitchTime(Thread* thread, Process* process);
+
     /// Lists all thread ids that aren't deleted/etc.
     std::vector<SharedPtr<Thread>> thread_list;
 
@@ -73,6 +91,7 @@ private:
     SharedPtr<Thread> current_thread = nullptr;
 
     Core::ARM_Interface& cpu_core;
+    u64 last_context_switch_time = 0;
 
     static std::mutex scheduler_mutex;
 };

src/core/hle/kernel/server_port.cpp

@@ -18,7 +18,7 @@ ServerPort::~ServerPort() = default;
 
 ResultVal<SharedPtr<ServerSession>> ServerPort::Accept() {
     if (pending_sessions.empty()) {
-        return ERR_NO_PENDING_SESSIONS;
+        return ERR_NOT_FOUND;
     }
 
     auto session = std::move(pending_sessions.back());
@@ -28,7 +28,7 @@ ResultVal<SharedPtr<ServerSession>> ServerPort::Accept() {
 
 bool ServerPort::ShouldWait(Thread* thread) const {
     // If there are no pending sessions, we wait until a new one is added.
-    return pending_sessions.size() == 0;
+    return pending_sessions.empty();
 }
 
 void ServerPort::Acquire(Thread* thread) {

src/core/hle/kernel/server_session.cpp

@@ -63,7 +63,7 @@ void ServerSession::Acquire(Thread* thread) {
 }
 
 ResultCode ServerSession::HandleDomainSyncRequest(Kernel::HLERequestContext& context) {
-    auto& domain_message_header = context.GetDomainMessageHeader();
+    auto* const domain_message_header = context.GetDomainMessageHeader();
     if (domain_message_header) {
         // Set domain handlers in HLE context, used for domain objects (IPC interfaces) as inputs
         context.SetDomainRequestHandlers(domain_request_handlers);
@@ -111,7 +111,7 @@ ResultCode ServerSession::HandleSyncRequest(SharedPtr<Thread> thread) {
 
     ResultCode result = RESULT_SUCCESS;
     // If the session has been converted to a domain, handle the domain request
-    if (IsDomain() && context.GetDomainMessageHeader()) {
+    if (IsDomain() && context.HasDomainMessageHeader()) {
         result = HandleDomainSyncRequest(context);
         // If there is no domain header, the regular session handler is used
     } else if (hle_handler != nullptr) {

src/core/hle/kernel/shared_memory.cpp

@@ -80,20 +80,19 @@ SharedPtr<SharedMemory> SharedMemory::CreateForApplet(
 
 ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermission permissions,
                              MemoryPermission other_permissions) {
-
-    MemoryPermission own_other_permissions =
+    const MemoryPermission own_other_permissions =
         target_process == owner_process ? this->permissions : this->other_permissions;
 
     // Automatically allocated memory blocks can only be mapped with other_permissions = DontCare
     if (base_address == 0 && other_permissions != MemoryPermission::DontCare) {
-        return ERR_INVALID_COMBINATION;
+        return ERR_INVALID_MEMORY_PERMISSIONS;
     }
 
     // Error out if the requested permissions don't match what the creator process allows.
     if (static_cast<u32>(permissions) & ~static_cast<u32>(own_other_permissions)) {
         LOG_ERROR(Kernel, "cannot map id={}, address=0x{:X} name={}, permissions don't match",
                   GetObjectId(), address, name);
-        return ERR_INVALID_COMBINATION;
+        return ERR_INVALID_MEMORY_PERMISSIONS;
     }
 
     // Error out if the provided permissions are not compatible with what the creator process needs.

src/core/hle/kernel/svc.cpp

@@ -395,16 +395,42 @@ struct BreakReason {
 /// Break program execution
 static void Break(u32 reason, u64 info1, u64 info2) {
     BreakReason break_reason{reason};
+    bool has_dumped_buffer{};
 
+    const auto handle_debug_buffer = [&](VAddr addr, u64 sz) {
+        if (sz == 0 || addr == 0 || has_dumped_buffer) {
+            return;
+        }
+
+        // This typically is an error code so we're going to assume this is the case
+        if (sz == sizeof(u32)) {
+            LOG_CRITICAL(Debug_Emulated, "debug_buffer_err_code={:X}", Memory::Read32(addr));
+        } else {
+            // We don't know what's in here so we'll hexdump it
+            std::vector<u8> debug_buffer(sz);
+            Memory::ReadBlock(addr, debug_buffer.data(), sz);
+            std::string hexdump;
+            for (std::size_t i = 0; i < debug_buffer.size(); i++) {
+                hexdump += fmt::format("{:02X} ", debug_buffer[i]);
+                if (i != 0 && i % 16 == 0) {
+                    hexdump += '\n';
+                }
+            }
+            LOG_CRITICAL(Debug_Emulated, "debug_buffer=\n{}", hexdump);
+        }
+        has_dumped_buffer = true;
+    };
     switch (break_reason.break_type) {
     case BreakType::Panic:
         LOG_CRITICAL(Debug_Emulated, "Signalling debugger, PANIC! info1=0x{:016X}, info2=0x{:016X}",
                      info1, info2);
+        handle_debug_buffer(info1, info2);
         break;
     case BreakType::AssertionFailed:
         LOG_CRITICAL(Debug_Emulated,
                      "Signalling debugger, Assertion failed! info1=0x{:016X}, info2=0x{:016X}",
                      info1, info2);
+        handle_debug_buffer(info1, info2);
         break;
     case BreakType::PreNROLoad:
         LOG_WARNING(
@@ -433,6 +459,7 @@ static void Break(u32 reason, u64 info1, u64 info2) {
             Debug_Emulated,
             "Signalling debugger, Unknown break reason {}, info1=0x{:016X}, info2=0x{:016X}",
             static_cast<u32>(break_reason.break_type.Value()), info1, info2);
+        handle_debug_buffer(info1, info2);
         break;
     }
 
@@ -441,6 +468,7 @@ static void Break(u32 reason, u64 info1, u64 info2) {
             Debug_Emulated,
             "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}",
             reason, info1, info2);
+        handle_debug_buffer(info1, info2);
         ASSERT(false);
 
         Core::CurrentProcess()->PrepareForTermination();
@@ -467,6 +495,37 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
     LOG_TRACE(Kernel_SVC, "called info_id=0x{:X}, info_sub_id=0x{:X}, handle=0x{:08X}", info_id,
               info_sub_id, handle);
 
+    enum class GetInfoType : u64 {
+        // 1.0.0+
+        AllowedCpuIdBitmask = 0,
+        AllowedThreadPrioBitmask = 1,
+        MapRegionBaseAddr = 2,
+        MapRegionSize = 3,
+        HeapRegionBaseAddr = 4,
+        HeapRegionSize = 5,
+        TotalMemoryUsage = 6,
+        TotalHeapUsage = 7,
+        IsCurrentProcessBeingDebugged = 8,
+        ResourceHandleLimit = 9,
+        IdleTickCount = 10,
+        RandomEntropy = 11,
+        PerformanceCounter = 0xF0000002,
+        // 2.0.0+
+        ASLRRegionBaseAddr = 12,
+        ASLRRegionSize = 13,
+        NewMapRegionBaseAddr = 14,
+        NewMapRegionSize = 15,
+        // 3.0.0+
+        IsVirtualAddressMemoryEnabled = 16,
+        PersonalMmHeapUsage = 17,
+        TitleId = 18,
+        // 4.0.0+
+        PrivilegedProcessId = 19,
+        // 5.0.0+
+        UserExceptionContextAddr = 20,
+        ThreadTickCount = 0xF0000002,
+    };
+
     const auto* current_process = Core::CurrentProcess();
     const auto& vm_manager = current_process->VMManager();
 
@@ -529,6 +588,36 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
                     "(STUBBED) Attempted to query user exception context address, returned 0");
         *result = 0;
         break;
+    case GetInfoType::ThreadTickCount: {
+        constexpr u64 num_cpus = 4;
+        if (info_sub_id != 0xFFFFFFFFFFFFFFFF && info_sub_id >= num_cpus) {
+            return ERR_INVALID_COMBINATION_KERNEL;
+        }
+
+        const auto thread =
+            current_process->GetHandleTable().Get<Thread>(static_cast<Handle>(handle));
+        if (!thread) {
+            return ERR_INVALID_HANDLE;
+        }
+
+        const auto& system = Core::System::GetInstance();
+        const auto& scheduler = system.CurrentScheduler();
+        const auto* const current_thread = scheduler.GetCurrentThread();
+        const bool same_thread = current_thread == thread;
+
+        const u64 prev_ctx_ticks = scheduler.GetLastContextSwitchTicks();
+        u64 out_ticks = 0;
+        if (same_thread && info_sub_id == 0xFFFFFFFFFFFFFFFF) {
+            const u64 thread_ticks = current_thread->GetTotalCPUTimeTicks();
+
+            out_ticks = thread_ticks + (CoreTiming::GetTicks() - prev_ctx_ticks);
+        } else if (same_thread && info_sub_id == system.CurrentCoreIndex()) {
+            out_ticks = CoreTiming::GetTicks() - prev_ctx_ticks;
+        }
+
+        *result = out_ticks;
+        break;
+    }
     default:
         UNIMPLEMENTED();
     }
@@ -594,16 +683,17 @@ static ResultCode SetThreadPriority(Handle handle, u32 priority) {
     }
 
     const auto* const current_process = Core::CurrentProcess();
-    SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
-    if (!thread) {
-        return ERR_INVALID_HANDLE;
-    }
 
     // Note: The kernel uses the current process's resource limit instead of
     // the one from the thread owner's resource limit.
     const ResourceLimit& resource_limit = current_process->GetResourceLimit();
     if (resource_limit.GetMaxResourceValue(ResourceType::Priority) > priority) {
-        return ERR_NOT_AUTHORIZED;
+        return ERR_INVALID_THREAD_PRIORITY;
+    }
+
+    SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
+    if (!thread) {
+        return ERR_INVALID_HANDLE;
     }
 
     thread->SetPriority(priority);
@@ -745,7 +835,7 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
     auto* const current_process = Core::CurrentProcess();
     const ResourceLimit& resource_limit = current_process->GetResourceLimit();
     if (resource_limit.GetMaxResourceValue(ResourceType::Priority) > priority) {
-        return ERR_NOT_AUTHORIZED;
+        return ERR_INVALID_THREAD_PRIORITY;
     }
 
     if (processor_id == THREADPROCESSORID_DEFAULT) {

src/core/hle/kernel/svc.h

@@ -24,37 +24,6 @@ struct PageInfo {
     u64 flags;
 };
 
-/// Values accepted by svcGetInfo
-enum class GetInfoType : u64 {
-    // 1.0.0+
-    AllowedCpuIdBitmask = 0,
-    AllowedThreadPrioBitmask = 1,
-    MapRegionBaseAddr = 2,
-    MapRegionSize = 3,
-    HeapRegionBaseAddr = 4,
-    HeapRegionSize = 5,
-    TotalMemoryUsage = 6,
-    TotalHeapUsage = 7,
-    IsCurrentProcessBeingDebugged = 8,
-    ResourceHandleLimit = 9,
-    IdleTickCount = 10,
-    RandomEntropy = 11,
-    PerformanceCounter = 0xF0000002,
-    // 2.0.0+
-    ASLRRegionBaseAddr = 12,
-    ASLRRegionSize = 13,
-    NewMapRegionBaseAddr = 14,
-    NewMapRegionSize = 15,
-    // 3.0.0+
-    IsVirtualAddressMemoryEnabled = 16,
-    PersonalMmHeapUsage = 17,
-    TitleId = 18,
-    // 4.0.0+
-    PrivilegedProcessId = 19,
-    // 5.0.0+
-    UserExceptionContextAddr = 20,
-};
-
 void CallSVC(u32 immediate);
 
 } // namespace Kernel

src/core/hle/kernel/thread.cpp

@@ -4,9 +4,9 @@
 
 #include <algorithm>
 #include <cinttypes>
+#include <optional>
 #include <vector>
 
-#include <boost/optional.hpp>
 #include <boost/range/algorithm_ext/erase.hpp>
 
 #include "common/assert.h"
@@ -94,7 +94,7 @@ void Thread::CancelWakeupTimer() {
     CoreTiming::UnscheduleEventThreadsafe(kernel.ThreadWakeupCallbackEventType(), callback_handle);
 }
 
-static boost::optional<s32> GetNextProcessorId(u64 mask) {
+static std::optional<s32> GetNextProcessorId(u64 mask) {
     for (s32 index = 0; index < Core::NUM_CPU_CORES; ++index) {
         if (mask & (1ULL << index)) {
             if (!Core::System::GetInstance().Scheduler(index).GetCurrentThread()) {
@@ -142,7 +142,7 @@ void Thread::ResumeFromWait() {
 
     status = ThreadStatus::Ready;
 
-    boost::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask);
+    std::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask);
     if (!new_processor_id) {
         new_processor_id = processor_id;
     }
@@ -369,7 +369,7 @@ void Thread::ChangeCore(u32 core, u64 mask) {
         return;
     }
 
-    boost::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
+    std::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
 
     if (!new_processor_id) {
         new_processor_id = processor_id;

src/core/hle/kernel/thread.h

@@ -258,6 +258,14 @@ public:
         return last_running_ticks;
     }
 
+    u64 GetTotalCPUTimeTicks() const {
+        return total_cpu_time_ticks;
+    }
+
+    void UpdateCPUTimeTicks(u64 ticks) {
+        total_cpu_time_ticks += ticks;
+    }
+
     s32 GetProcessorID() const {
         return processor_id;
     }
@@ -378,7 +386,8 @@ private:
     u32 nominal_priority = 0; ///< Nominal thread priority, as set by the emulated application
     u32 current_priority = 0; ///< Current thread priority, can be temporarily changed
 
-    u64 last_running_ticks = 0; ///< CPU tick when thread was last running
+    u64 total_cpu_time_ticks = 0; ///< Total CPU running ticks.
+    u64 last_running_ticks = 0;   ///< CPU tick when thread was last running
 
     s32 processor_id = 0;
 

src/core/hle/kernel/vm_manager.cpp

@@ -143,6 +143,26 @@ ResultVal<VMManager::VMAHandle> VMManager::MapBackingMemory(VAddr target, u8* me
     return MakeResult<VMAHandle>(MergeAdjacent(vma_handle));
 }
 
+ResultVal<VAddr> VMManager::FindFreeRegion(u64 size) const {
+    // Find the first Free VMA.
+    const VAddr base = GetASLRRegionBaseAddress();
+    const VMAHandle vma_handle = std::find_if(vma_map.begin(), vma_map.end(), [&](const auto& vma) {
+        if (vma.second.type != VMAType::Free)
+            return false;
+
+        const VAddr vma_end = vma.second.base + vma.second.size;
+        return vma_end > base && vma_end >= base + size;
+    });
+
+    if (vma_handle == vma_map.end()) {
+        // TODO(Subv): Find the correct error code here.
+        return ResultCode(-1);
+    }
+
+    const VAddr target = std::max(base, vma_handle->second.base);
+    return MakeResult<VAddr>(target);
+}
+
 ResultVal<VMManager::VMAHandle> VMManager::MapMMIO(VAddr target, PAddr paddr, u64 size,
                                                    MemoryState state,
                                                    Memory::MemoryHookPointer mmio_handler) {

src/core/hle/kernel/vm_manager.h

@@ -157,6 +157,14 @@ public:
      */
     ResultVal<VMAHandle> MapBackingMemory(VAddr target, u8* memory, u64 size, MemoryState state);
 
+    /**
+     * Finds the first free address that can hold a region of the desired size.
+     *
+     * @param size Size of the desired region.
+     * @return The found free address.
+     */
+    ResultVal<VAddr> FindFreeRegion(u64 size) const;
+
     /**
      * Maps a memory-mapped IO region at a given address.
      *

src/core/hle/service/acc/acc.cpp

@@ -2,9 +2,13 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include <array>
+#include "common/common_paths.h"
 #include "common/common_types.h"
+#include "common/file_util.h"
 #include "common/logging/log.h"
+#include "common/string_util.h"
 #include "common/swap.h"
 #include "core/core_timing.h"
 #include "core/hle/ipc_helpers.h"
@@ -16,6 +20,7 @@
 #include "core/hle/service/acc/profile_manager.h"
 
 namespace Service::Account {
+
 // TODO: RE this structure
 struct UserData {
     INSERT_PADDING_WORDS(1);
@@ -27,6 +32,29 @@ struct UserData {
 };
 static_assert(sizeof(UserData) == 0x80, "UserData structure has incorrect size");
 
+// Smallest JPEG https://github.com/mathiasbynens/small/blob/master/jpeg.jpg
+// used as a backup should the one on disk not exist
+constexpr u32 backup_jpeg_size = 107;
+constexpr std::array<u8, backup_jpeg_size> backup_jpeg{{
+    0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x02, 0x02,
+    0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04, 0x08, 0x06, 0x06, 0x05,
+    0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a, 0x0c, 0x0f, 0x0c, 0x0a, 0x0b, 0x0e,
+    0x0b, 0x09, 0x09, 0x0d, 0x11, 0x0d, 0x0e, 0x0f, 0x10, 0x10, 0x11, 0x10, 0x0a, 0x0c, 0x12, 0x13,
+    0x12, 0x10, 0x13, 0x0f, 0x10, 0x10, 0x10, 0xff, 0xc9, 0x00, 0x0b, 0x08, 0x00, 0x01, 0x00, 0x01,
+    0x01, 0x01, 0x11, 0x00, 0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08,
+    0x01, 0x01, 0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
+}};
+
+static std::string GetImagePath(UUID uuid) {
+    return FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+           "/system/save/8000000000000010/su/avators/" + uuid.FormatSwitch() + ".jpg";
+}
+
+static constexpr u32 SanitizeJPEGSize(std::size_t size) {
+    constexpr std::size_t max_jpeg_image_size = 0x20000;
+    return static_cast<u32>(std::min(size, max_jpeg_image_size));
+}
+
 class IProfile final : public ServiceFramework<IProfile> {
 public:
     explicit IProfile(UUID user_id, ProfileManager& profile_manager)
@@ -73,32 +101,42 @@ private:
     }
 
     void LoadImage(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_ACC, "(STUBBED) called");
-        // smallest jpeg https://github.com/mathiasbynens/small/blob/master/jpeg.jpg
-        // TODO(mailwl): load actual profile image from disk, width 256px, max size 0x20000
-        constexpr u32 jpeg_size = 107;
-        static constexpr std::array<u8, jpeg_size> jpeg{
-            0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03,
-            0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04,
-            0x08, 0x06, 0x06, 0x05, 0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a,
-            0x0c, 0x0f, 0x0c, 0x0a, 0x0b, 0x0e, 0x0b, 0x09, 0x09, 0x0d, 0x11, 0x0d, 0x0e, 0x0f,
-            0x10, 0x10, 0x11, 0x10, 0x0a, 0x0c, 0x12, 0x13, 0x12, 0x10, 0x13, 0x0f, 0x10, 0x10,
-            0x10, 0xff, 0xc9, 0x00, 0x0b, 0x08, 0x00, 0x01, 0x00, 0x01, 0x01, 0x01, 0x11, 0x00,
-            0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08, 0x01, 0x01,
-            0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
-        };
-        ctx.WriteBuffer(jpeg);
+        LOG_DEBUG(Service_ACC, "called");
+
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(jpeg_size);
+
+        const FileUtil::IOFile image(GetImagePath(user_id), "rb");
+        if (!image.IsOpen()) {
+            LOG_WARNING(Service_ACC,
+                        "Failed to load user provided image! Falling back to built-in backup...");
+            ctx.WriteBuffer(backup_jpeg);
+            rb.Push<u32>(backup_jpeg_size);
+            return;
+        }
+
+        const u32 size = SanitizeJPEGSize(image.GetSize());
+        std::vector<u8> buffer(size);
+        image.ReadBytes(buffer.data(), buffer.size());
+
+        ctx.WriteBuffer(buffer.data(), buffer.size());
+        rb.Push<u32>(size);
     }
 
     void GetImageSize(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_ACC, "(STUBBED) called");
-        constexpr u32 jpeg_size = 107;
+        LOG_DEBUG(Service_ACC, "called");
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(jpeg_size);
+
+        const FileUtil::IOFile image(GetImagePath(user_id), "rb");
+
+        if (!image.IsOpen()) {
+            LOG_WARNING(Service_ACC,
+                        "Failed to load user provided image! Falling back to built-in backup...");
+            rb.Push<u32>(backup_jpeg_size);
+        } else {
+            rb.Push<u32>(SanitizeJPEGSize(image.GetSize()));
+        }
     }
 
     const ProfileManager& profile_manager;
@@ -204,6 +242,28 @@ void Module::Interface::GetBaasAccountManagerForApplication(Kernel::HLERequestCo
     LOG_DEBUG(Service_ACC, "called");
 }
 
+void Module::Interface::TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_ACC, "called");
+    // A u8 is passed into this function which we can safely ignore. It's to determine if we have
+    // access to use the network or not by the looks of it
+    IPC::ResponseBuilder rb{ctx, 6};
+    if (profile_manager->GetUserCount() != 1) {
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u128>(INVALID_UUID);
+        return;
+    }
+    auto user_list = profile_manager->GetAllUsers();
+    if (user_list.empty()) {
+        rb.Push(ResultCode(-1)); // TODO(ogniK): Find the correct error code
+        rb.PushRaw<u128>(INVALID_UUID);
+        return;
+    }
+
+    // Select the first user we have
+    rb.Push(RESULT_SUCCESS);
+    rb.PushRaw<u128>(profile_manager->GetUser(0)->uuid);
+}
+
 Module::Interface::Interface(std::shared_ptr<Module> module,
                              std::shared_ptr<ProfileManager> profile_manager, const char* name)
     : ServiceFramework(name), module(std::move(module)),

src/core/hle/service/acc/acc.h

@@ -27,6 +27,7 @@ public:
         void InitializeApplicationInfo(Kernel::HLERequestContext& ctx);
         void GetBaasAccountManagerForApplication(Kernel::HLERequestContext& ctx);
         void IsUserRegistrationRequestPermitted(Kernel::HLERequestContext& ctx);
+        void TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx);
 
     protected:
         std::shared_ptr<Module> module;

src/core/hle/service/acc/acc_su.cpp

@@ -17,7 +17,7 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
         {5, &ACC_SU::GetProfile, "GetProfile"},
         {6, nullptr, "GetProfileDigest"},
         {50, &ACC_SU::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"},
-        {51, nullptr, "TrySelectUserWithoutInteraction"},
+        {51, &ACC_SU::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"},
         {60, nullptr, "ListOpenContextStoredUsers"},
         {100, nullptr, "GetUserRegistrationNotifier"},
         {101, nullptr, "GetUserStateChangeNotifier"},

src/core/hle/service/acc/acc_u0.cpp

@@ -17,7 +17,7 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
         {5, &ACC_U0::GetProfile, "GetProfile"},
         {6, nullptr, "GetProfileDigest"},
         {50, &ACC_U0::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"},
-        {51, nullptr, "TrySelectUserWithoutInteraction"},
+        {51, &ACC_U0::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"},
         {60, nullptr, "ListOpenContextStoredUsers"},
         {100, &ACC_U0::InitializeApplicationInfo, "InitializeApplicationInfo"},
         {101, &ACC_U0::GetBaasAccountManagerForApplication, "GetBaasAccountManagerForApplication"},

src/core/hle/service/acc/acc_u1.cpp

@@ -17,7 +17,7 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
         {5, &ACC_U1::GetProfile, "GetProfile"},
         {6, nullptr, "GetProfileDigest"},
         {50, &ACC_U1::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"},
-        {51, nullptr, "TrySelectUserWithoutInteraction"},
+        {51, &ACC_U1::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"},
         {60, nullptr, "ListOpenContextStoredUsers"},
         {100, nullptr, "GetUserRegistrationNotifier"},
         {101, nullptr, "GetUserStateChangeNotifier"},

src/core/hle/service/acc/profile_manager.cpp

@@ -3,41 +3,66 @@
 // Refer to the license.txt file included.
 
 #include <random>
-#include <boost/optional.hpp>
+
+#include "common/file_util.h"
 #include "core/hle/service/acc/profile_manager.h"
 #include "core/settings.h"
 
 namespace Service::Account {
+
+struct UserRaw {
+    UUID uuid;
+    UUID uuid2;
+    u64 timestamp;
+    ProfileUsername username;
+    INSERT_PADDING_BYTES(0x80);
+};
+static_assert(sizeof(UserRaw) == 0xC8, "UserRaw has incorrect size.");
+
+struct ProfileDataRaw {
+    INSERT_PADDING_BYTES(0x10);
+    std::array<UserRaw, MAX_USERS> users;
+};
+static_assert(sizeof(ProfileDataRaw) == 0x650, "ProfileDataRaw has incorrect size.");
+
 // TODO(ogniK): Get actual error codes
 constexpr ResultCode ERROR_TOO_MANY_USERS(ErrorModule::Account, -1);
 constexpr ResultCode ERROR_USER_ALREADY_EXISTS(ErrorModule::Account, -2);
 constexpr ResultCode ERROR_ARGUMENT_IS_NULL(ErrorModule::Account, 20);
 
-const UUID& UUID::Generate() {
+constexpr char ACC_SAVE_AVATORS_BASE_PATH[] = "/system/save/8000000000000010/su/avators/";
+
+UUID UUID::Generate() {
     std::random_device device;
     std::mt19937 gen(device());
     std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max());
-    uuid[0] = distribution(gen);
-    uuid[1] = distribution(gen);
-    return *this;
+    return UUID{distribution(gen), distribution(gen)};
 }
 
 ProfileManager::ProfileManager() {
-    // TODO(ogniK): Create the default user we have for now until loading/saving users is added
-    auto user_uuid = UUID{1, 0};
-    ASSERT(CreateNewUser(user_uuid, Settings::values.username).IsSuccess());
-    OpenUser(user_uuid);
+    ParseUserSaveFile();
+
+    if (user_count == 0)
+        CreateNewUser(UUID::Generate(), "yuzu");
+
+    auto current = std::clamp<int>(Settings::values.current_user, 0, MAX_USERS - 1);
+    if (UserExistsIndex(current))
+        current = 0;
+
+    OpenUser(*GetUser(current));
 }
 
-ProfileManager::~ProfileManager() = default;
+ProfileManager::~ProfileManager() {
+    WriteUserSaveFile();
+}
 
 /// After a users creation it needs to be "registered" to the system. AddToProfiles handles the
 /// internal management of the users profiles
-boost::optional<std::size_t> ProfileManager::AddToProfiles(const ProfileInfo& user) {
+std::optional<std::size_t> ProfileManager::AddToProfiles(const ProfileInfo& profile) {
     if (user_count >= MAX_USERS) {
-        return boost::none;
+        return {};
     }
-    profiles[user_count] = user;
+    profiles[user_count] = profile;
     return user_count++;
 }
 
@@ -56,7 +81,7 @@ bool ProfileManager::RemoveProfileAtIndex(std::size_t index) {
 
 /// Helper function to register a user to the system
 ResultCode ProfileManager::AddUser(const ProfileInfo& user) {
-    if (AddToProfiles(user) == boost::none) {
+    if (!AddToProfiles(user)) {
         return ERROR_TOO_MANY_USERS;
     }
     return RESULT_SUCCESS;
@@ -101,31 +126,40 @@ ResultCode ProfileManager::CreateNewUser(UUID uuid, const std::string& username)
     return CreateNewUser(uuid, username_output);
 }
 
+std::optional<UUID> ProfileManager::GetUser(std::size_t index) const {
+    if (index >= MAX_USERS) {
+        return {};
+    }
+
+    return profiles[index].user_uuid;
+}
+
 /// Returns a users profile index based on their user id.
-boost::optional<std::size_t> ProfileManager::GetUserIndex(const UUID& uuid) const {
+std::optional<std::size_t> ProfileManager::GetUserIndex(const UUID& uuid) const {
     if (!uuid) {
-        return boost::none;
+        return {};
     }
-    auto iter = std::find_if(profiles.begin(), profiles.end(),
-                             [&uuid](const ProfileInfo& p) { return p.user_uuid == uuid; });
+
+    const auto iter = std::find_if(profiles.begin(), profiles.end(),
+                                   [&uuid](const ProfileInfo& p) { return p.user_uuid == uuid; });
     if (iter == profiles.end()) {
-        return boost::none;
+        return {};
     }
+
     return static_cast<std::size_t>(std::distance(profiles.begin(), iter));
 }
 
 /// Returns a users profile index based on their profile
-boost::optional<std::size_t> ProfileManager::GetUserIndex(const ProfileInfo& user) const {
+std::optional<std::size_t> ProfileManager::GetUserIndex(const ProfileInfo& user) const {
     return GetUserIndex(user.user_uuid);
 }
 
 /// Returns the data structure used by the switch when GetProfileBase is called on acc:*
-bool ProfileManager::GetProfileBase(boost::optional<std::size_t> index,
-                                    ProfileBase& profile) const {
-    if (index == boost::none || index >= MAX_USERS) {
+bool ProfileManager::GetProfileBase(std::optional<std::size_t> index, ProfileBase& profile) const {
+    if (!index || index >= MAX_USERS) {
         return false;
     }
-    const auto& prof_info = profiles[index.get()];
+    const auto& prof_info = profiles[*index];
     profile.user_uuid = prof_info.user_uuid;
     profile.username = prof_info.username;
     profile.timestamp = prof_info.creation_time;
@@ -134,7 +168,7 @@ bool ProfileManager::GetProfileBase(boost::optional<std::size_t> index,
 
 /// Returns the data structure used by the switch when GetProfileBase is called on acc:*
 bool ProfileManager::GetProfileBase(UUID uuid, ProfileBase& profile) const {
-    auto idx = GetUserIndex(uuid);
+    const auto idx = GetUserIndex(uuid);
     return GetProfileBase(idx, profile);
 }
 
@@ -161,26 +195,34 @@ std::size_t ProfileManager::GetOpenUserCount() const {
 
 /// Checks if a user id exists in our profile manager
 bool ProfileManager::UserExists(UUID uuid) const {
-    return (GetUserIndex(uuid) != boost::none);
+    return GetUserIndex(uuid).has_value();
+}
+
+bool ProfileManager::UserExistsIndex(std::size_t index) const {
+    if (index >= MAX_USERS)
+        return false;
+    return profiles[index].user_uuid.uuid != INVALID_UUID;
 }
 
 /// Opens a specific user
 void ProfileManager::OpenUser(UUID uuid) {
-    auto idx = GetUserIndex(uuid);
-    if (idx == boost::none) {
+    const auto idx = GetUserIndex(uuid);
+    if (!idx) {
         return;
     }
-    profiles[idx.get()].is_open = true;
+
+    profiles[*idx].is_open = true;
     last_opened_user = uuid;
 }
 
 /// Closes a specific user
 void ProfileManager::CloseUser(UUID uuid) {
-    auto idx = GetUserIndex(uuid);
-    if (idx == boost::none) {
+    const auto idx = GetUserIndex(uuid);
+    if (!idx) {
         return;
     }
-    profiles[idx.get()].is_open = false;
+
+    profiles[*idx].is_open = false;
 }
 
 /// Gets all valid user ids on the system
@@ -210,10 +252,10 @@ UUID ProfileManager::GetLastOpenedUser() const {
 }
 
 /// Return the users profile base and the unknown arbitary data.
-bool ProfileManager::GetProfileBaseAndData(boost::optional<std::size_t> index, ProfileBase& profile,
+bool ProfileManager::GetProfileBaseAndData(std::optional<std::size_t> index, ProfileBase& profile,
                                            ProfileData& data) const {
     if (GetProfileBase(index, profile)) {
-        data = profiles[index.get()].data;
+        data = profiles[*index].data;
         return true;
     }
     return false;
@@ -222,7 +264,7 @@ bool ProfileManager::GetProfileBaseAndData(boost::optional<std::size_t> index, P
 /// Return the users profile base and the unknown arbitary data.
 bool ProfileManager::GetProfileBaseAndData(UUID uuid, ProfileBase& profile,
                                            ProfileData& data) const {
-    auto idx = GetUserIndex(uuid);
+    const auto idx = GetUserIndex(uuid);
     return GetProfileBaseAndData(idx, profile, data);
 }
 
@@ -239,4 +281,96 @@ bool ProfileManager::CanSystemRegisterUser() const {
     // emulate qlaunch. Update this to dynamically change.
 }
 
+bool ProfileManager::RemoveUser(UUID uuid) {
+    const auto index = GetUserIndex(uuid);
+    if (!index) {
+        return false;
+    }
+
+    profiles[*index] = ProfileInfo{};
+    std::stable_partition(profiles.begin(), profiles.end(),
+                          [](const ProfileInfo& profile) { return profile.user_uuid; });
+    return true;
+}
+
+bool ProfileManager::SetProfileBase(UUID uuid, const ProfileBase& profile_new) {
+    const auto index = GetUserIndex(uuid);
+    if (!index || profile_new.user_uuid == UUID(INVALID_UUID)) {
+        return false;
+    }
+
+    auto& profile = profiles[*index];
+    profile.user_uuid = profile_new.user_uuid;
+    profile.username = profile_new.username;
+    profile.creation_time = profile_new.timestamp;
+
+    return true;
+}
+
+void ProfileManager::ParseUserSaveFile() {
+    FileUtil::IOFile save(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                              ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat",
+                          "rb");
+
+    if (!save.IsOpen()) {
+        LOG_WARNING(Service_ACC, "Failed to load profile data from save data... Generating new "
+                                 "user 'yuzu' with random UUID.");
+        return;
+    }
+
+    ProfileDataRaw data;
+    if (save.ReadBytes(&data, sizeof(ProfileDataRaw)) != sizeof(ProfileDataRaw)) {
+        LOG_WARNING(Service_ACC, "profiles.dat is smaller than expected... Generating new user "
+                                 "'yuzu' with random UUID.");
+        return;
+    }
+
+    for (std::size_t i = 0; i < MAX_USERS; ++i) {
+        const auto& user = data.users[i];
+
+        if (user.uuid != UUID(INVALID_UUID))
+            AddUser({user.uuid, user.username, user.timestamp, {}, false});
+    }
+
+    std::stable_partition(profiles.begin(), profiles.end(),
+                          [](const ProfileInfo& profile) { return profile.user_uuid; });
+}
+
+void ProfileManager::WriteUserSaveFile() {
+    ProfileDataRaw raw{};
+
+    for (std::size_t i = 0; i < MAX_USERS; ++i) {
+        raw.users[i].username = profiles[i].username;
+        raw.users[i].uuid2 = profiles[i].user_uuid;
+        raw.users[i].uuid = profiles[i].user_uuid;
+        raw.users[i].timestamp = profiles[i].creation_time;
+    }
+
+    const auto raw_path =
+        FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + "/system/save/8000000000000010";
+    if (FileUtil::Exists(raw_path) && !FileUtil::IsDirectory(raw_path))
+        FileUtil::Delete(raw_path);
+
+    const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                      ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat";
+
+    if (!FileUtil::CreateFullPath(path)) {
+        LOG_WARNING(Service_ACC, "Failed to create full path of profiles.dat. Create the directory "
+                                 "nand/system/save/8000000000000010/su/avators to mitigate this "
+                                 "issue.");
+        return;
+    }
+
+    FileUtil::IOFile save(path, "wb");
+
+    if (!save.IsOpen()) {
+        LOG_WARNING(Service_ACC, "Failed to write save data to file... No changes to user data "
+                                 "made in current session will be saved.");
+        return;
+    }
+
+    save.Resize(sizeof(ProfileDataRaw));
+    save.WriteBytes(&raw, sizeof(ProfileDataRaw));
+}
+
 }; // namespace Service::Account

src/core/hle/service/acc/profile_manager.h

@@ -5,8 +5,8 @@
 #pragma once
 
 #include <array>
+#include <optional>
 
-#include "boost/optional.hpp"
 #include "common/common_types.h"
 #include "common/swap.h"
 #include "core/hle/result.h"
@@ -36,7 +36,7 @@ struct UUID {
     }
 
     // TODO(ogniK): Properly generate uuids based on RFC-4122
-    const UUID& Generate();
+    static UUID Generate();
 
     // Set the UUID to {0,0} to be considered an invalid user
     void Invalidate() {
@@ -45,10 +45,20 @@ struct UUID {
     std::string Format() const {
         return fmt::format("0x{:016X}{:016X}", uuid[1], uuid[0]);
     }
+
+    std::string FormatSwitch() const {
+        std::array<u8, 16> s{};
+        std::memcpy(s.data(), uuid.data(), sizeof(u128));
+        return fmt::format("{:02x}{:02x}{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{"
+                           ":02x}{:02x}{:02x}{:02x}{:02x}",
+                           s[0], s[1], s[2], s[3], s[4], s[5], s[6], s[7], s[8], s[9], s[10], s[11],
+                           s[12], s[13], s[14], s[15]);
+    }
 };
 static_assert(sizeof(UUID) == 16, "UUID is an invalid size!");
 
-using ProfileUsername = std::array<u8, 0x20>;
+constexpr std::size_t profile_username_size = 32;
+using ProfileUsername = std::array<u8, profile_username_size>;
 using ProfileData = std::array<u8, MAX_DATA>;
 using UserIDArray = std::array<UUID, MAX_USERS>;
 
@@ -81,18 +91,19 @@ static_assert(sizeof(ProfileBase) == 0x38, "ProfileBase is an invalid size");
 /// objects
 class ProfileManager {
 public:
-    ProfileManager(); // TODO(ogniK): Load from system save
+    ProfileManager();
     ~ProfileManager();
 
     ResultCode AddUser(const ProfileInfo& user);
     ResultCode CreateNewUser(UUID uuid, const ProfileUsername& username);
     ResultCode CreateNewUser(UUID uuid, const std::string& username);
-    boost::optional<std::size_t> GetUserIndex(const UUID& uuid) const;
-    boost::optional<std::size_t> GetUserIndex(const ProfileInfo& user) const;
-    bool GetProfileBase(boost::optional<std::size_t> index, ProfileBase& profile) const;
+    std::optional<UUID> GetUser(std::size_t index) const;
+    std::optional<std::size_t> GetUserIndex(const UUID& uuid) const;
+    std::optional<std::size_t> GetUserIndex(const ProfileInfo& user) const;
+    bool GetProfileBase(std::optional<std::size_t> index, ProfileBase& profile) const;
     bool GetProfileBase(UUID uuid, ProfileBase& profile) const;
     bool GetProfileBase(const ProfileInfo& user, ProfileBase& profile) const;
-    bool GetProfileBaseAndData(boost::optional<std::size_t> index, ProfileBase& profile,
+    bool GetProfileBaseAndData(std::optional<std::size_t> index, ProfileBase& profile,
                                ProfileData& data) const;
     bool GetProfileBaseAndData(UUID uuid, ProfileBase& profile, ProfileData& data) const;
     bool GetProfileBaseAndData(const ProfileInfo& user, ProfileBase& profile,
@@ -100,6 +111,7 @@ public:
     std::size_t GetUserCount() const;
     std::size_t GetOpenUserCount() const;
     bool UserExists(UUID uuid) const;
+    bool UserExistsIndex(std::size_t index) const;
     void OpenUser(UUID uuid);
     void CloseUser(UUID uuid);
     UserIDArray GetOpenUsers() const;
@@ -108,11 +120,17 @@ public:
 
     bool CanSystemRegisterUser() const;
 
+    bool RemoveUser(UUID uuid);
+    bool SetProfileBase(UUID uuid, const ProfileBase& profile_new);
+
 private:
+    void ParseUserSaveFile();
+    void WriteUserSaveFile();
+    std::optional<std::size_t> AddToProfiles(const ProfileInfo& profile);
+    bool RemoveProfileAtIndex(std::size_t index);
+
     std::array<ProfileInfo, MAX_USERS> profiles{};
     std::size_t user_count = 0;
-    boost::optional<std::size_t> AddToProfiles(const ProfileInfo& profile);
-    bool RemoveProfileAtIndex(std::size_t index);
     UUID last_opened_user{INVALID_UUID};
 };
 

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

@@ -4,11 +4,13 @@
 
 #include <array>
 #include <cinttypes>
+#include <cstring>
 #include <stack>
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/process.h"
+#include "core/hle/service/acc/profile_manager.h"
 #include "core/hle/service/am/am.h"
 #include "core/hle/service/am/applet_ae.h"
 #include "core/hle/service/am/applet_oe.h"
@@ -26,6 +28,16 @@
 
 namespace Service::AM {
 
+constexpr u32 POP_LAUNCH_PARAMETER_MAGIC = 0xC79497CA;
+
+struct LaunchParameters {
+    u32_le magic;
+    u32_le is_account_selected;
+    u128 current_user;
+    INSERT_PADDING_BYTES(0x70);
+};
+static_assert(sizeof(LaunchParameters) == 0x88);
+
 IWindowController::IWindowController() : ServiceFramework("IWindowController") {
     // clang-format off
     static const FunctionInfo functions[] = {
@@ -724,20 +736,23 @@ void IApplicationFunctions::EndBlockingHomeButton(Kernel::HLERequestContext& ctx
 }
 
 void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) {
-    constexpr std::array<u8, 0x88> data{{
-        0xca, 0x97, 0x94, 0xc7, // Magic
-        1,    0,    0,    0,    // IsAccountSelected (bool)
-        1,    0,    0,    0,    // User Id (word 0)
-        0,    0,    0,    0,    // User Id (word 1)
-        0,    0,    0,    0,    // User Id (word 2)
-        0,    0,    0,    0     // User Id (word 3)
-    }};
+    LaunchParameters params{};
 
-    std::vector<u8> buffer(data.begin(), data.end());
+    params.magic = POP_LAUNCH_PARAMETER_MAGIC;
+    params.is_account_selected = 1;
+
+    Account::ProfileManager profile_manager{};
+    const auto uuid = profile_manager.GetUser(Settings::values.current_user);
+    ASSERT(uuid);
+    params.current_user = uuid->uuid;
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
 
     rb.Push(RESULT_SUCCESS);
+
+    std::vector<u8> buffer(sizeof(LaunchParameters));
+    std::memcpy(buffer.data(), &params, buffer.size());
+
     rb.PushIpcInterface<AM::IStorage>(buffer);
 
     LOG_DEBUG(Service_AM, "called");

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

@@ -24,8 +24,8 @@ namespace Service::AOC {
 constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000;
 constexpr u64 DLC_BASE_TO_AOC_ID = 0x1000;
 
-static bool CheckAOCTitleIDMatchesBase(u64 base, u64 aoc) {
-    return (aoc & DLC_BASE_TITLE_ID_MASK) == base;
+static bool CheckAOCTitleIDMatchesBase(u64 title_id, u64 base) {
+    return (title_id & DLC_BASE_TITLE_ID_MASK) == base;
 }
 
 static std::vector<u64> AccumulateAOCTitleIDs() {
@@ -74,7 +74,7 @@ void AOC_U::CountAddOnContent(Kernel::HLERequestContext& ctx) {
     const auto current = Core::System::GetInstance().CurrentProcess()->GetTitleID();
     rb.Push<u32>(static_cast<u32>(
         std::count_if(add_on_content.begin(), add_on_content.end(),
-                      [&current](u64 tid) { return (tid & DLC_BASE_TITLE_ID_MASK) == current; })));
+                      [current](u64 tid) { return CheckAOCTitleIDMatchesBase(tid, current); })));
 }
 
 void AOC_U::ListAddOnContent(Kernel::HLERequestContext& ctx) {

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

@@ -161,7 +161,7 @@ void HwOpus::OpenOpusDecoder(Kernel::HLERequestContext& ctx) {
     ASSERT_MSG(channel_count == 1 || channel_count == 2, "Invalid channel count");
 
     std::size_t worker_sz = WorkerBufferSize(channel_count);
-    ASSERT_MSG(buffer_sz < worker_sz, "Worker buffer too large");
+    ASSERT_MSG(buffer_sz >= worker_sz, "Worker buffer too large");
     std::unique_ptr<OpusDecoder, OpusDeleter> decoder{
         static_cast<OpusDecoder*>(operator new(worker_sz))};
     if (opus_decoder_init(decoder.get(), sample_rate, channel_count)) {

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

@@ -273,8 +273,8 @@ public:
             {0, &IFileSystem::CreateFile, "CreateFile"},
             {1, &IFileSystem::DeleteFile, "DeleteFile"},
             {2, &IFileSystem::CreateDirectory, "CreateDirectory"},
-            {3, nullptr, "DeleteDirectory"},
-            {4, nullptr, "DeleteDirectoryRecursively"},
+            {3, &IFileSystem::DeleteDirectory, "DeleteDirectory"},
+            {4, &IFileSystem::DeleteDirectoryRecursively, "DeleteDirectoryRecursively"},
             {5, &IFileSystem::RenameFile, "RenameFile"},
             {6, nullptr, "RenameDirectory"},
             {7, &IFileSystem::GetEntryType, "GetEntryType"},
@@ -329,6 +329,30 @@ public:
         rb.Push(backend.CreateDirectory(name));
     }
 
+    void DeleteDirectory(Kernel::HLERequestContext& ctx) {
+        const IPC::RequestParser rp{ctx};
+
+        const auto file_buffer = ctx.ReadBuffer();
+        std::string name = Common::StringFromBuffer(file_buffer);
+
+        LOG_DEBUG(Service_FS, "called directory {}", name);
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(backend.DeleteDirectory(name));
+    }
+
+    void DeleteDirectoryRecursively(Kernel::HLERequestContext& ctx) {
+        const IPC::RequestParser rp{ctx};
+
+        const auto file_buffer = ctx.ReadBuffer();
+        std::string name = Common::StringFromBuffer(file_buffer);
+
+        LOG_DEBUG(Service_FS, "called directory {}", name);
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(backend.DeleteDirectoryRecursively(name));
+    }
+
     void RenameFile(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
 

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

@@ -392,8 +392,10 @@ std::size_t Controller_NPad::GetSupportedNPadIdTypesSize() const {
 }
 
 void Controller_NPad::SetHoldType(NpadHoldType joy_hold_type) {
+    styleset_changed_event->Signal();
     hold_type = joy_hold_type;
 }
+
 Controller_NPad::NpadHoldType Controller_NPad::GetHoldType() const {
     return hold_type;
 }
@@ -427,6 +429,9 @@ void Controller_NPad::VibrateController(const std::vector<u32>& controller_ids,
 }
 
 Kernel::SharedPtr<Kernel::Event> Controller_NPad::GetStyleSetChangedEvent() const {
+    // TODO(ogniK): Figure out the best time to signal this event. This event seems that it should
+    // be signalled at least once, and signaled after a new controller is connected?
+    styleset_changed_event->Signal();
     return styleset_changed_event;
 }
 

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

@@ -96,6 +96,8 @@ public:
         // TODO(shinyquagsire23): Other update callbacks? (accel, gyro?)
 
         CoreTiming::ScheduleEvent(pad_update_ticks, pad_update_event);
+
+        ReloadInputDevices();
     }
 
     void ActivateController(HidController controller) {

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

@@ -3,9 +3,13 @@
 // Refer to the license.txt file included.
 
 #include <memory>
+#include <fmt/format.h>
 
+#include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/process.h"
 #include "core/hle/service/ldr/ldr.h"
 #include "core/hle/service/service.h"
+#include "core/loader/nro.h"
 
 namespace Service::LDR {
 
@@ -59,16 +63,58 @@ public:
     explicit RelocatableObject() : ServiceFramework{"ldr:ro"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "LoadNro"},
+            {0, &RelocatableObject::LoadNro, "LoadNro"},
             {1, nullptr, "UnloadNro"},
-            {2, nullptr, "LoadNrr"},
+            {2, &RelocatableObject::LoadNrr, "LoadNrr"},
             {3, nullptr, "UnloadNrr"},
-            {4, nullptr, "Initialize"},
+            {4, &RelocatableObject::Initialize, "Initialize"},
         };
         // clang-format on
 
         RegisterHandlers(functions);
     }
+
+    void LoadNrr(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
+
+    void LoadNro(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        rp.Skip(2, false);
+        const VAddr nro_addr{rp.Pop<VAddr>()};
+        const u64 nro_size{rp.Pop<u64>()};
+        const VAddr bss_addr{rp.Pop<VAddr>()};
+        const u64 bss_size{rp.Pop<u64>()};
+
+        // Read NRO data from memory
+        std::vector<u8> nro_data(nro_size);
+        Memory::ReadBlock(nro_addr, nro_data.data(), nro_size);
+
+        // Load NRO as new executable module
+        const VAddr addr{*Core::CurrentProcess()->VMManager().FindFreeRegion(nro_size + bss_size)};
+        Loader::AppLoader_NRO::LoadNro(nro_data, fmt::format("nro-{:08x}", addr), addr);
+
+        // TODO(bunnei): This is an incomplete implementation. It was tested with Super Mario Party.
+        // It is currently missing:
+        // - Signature checks with LoadNRR
+        // - Checking if a module has already been loaded
+        // - Using/validating BSS, etc. params (these are used from NRO header instead)
+        // - Error checking
+        // - ...Probably other things
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push(addr);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
+
+    void Initialize(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
 };
 
 void InstallInterfaces(SM::ServiceManager& sm) {

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

@@ -328,13 +328,15 @@ void Module::Interface::CreateUserInterface(Kernel::HLERequestContext& ctx) {
     rb.PushIpcInterface<IUser>(*this);
 }
 
-void Module::Interface::LoadAmiibo(const std::vector<u8>& buffer) {
+bool Module::Interface::LoadAmiibo(const std::vector<u8>& buffer) {
     std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
     if (buffer.size() < sizeof(AmiiboFile)) {
-        return; // Failed to load file
+        return false;
     }
+
     std::memcpy(&amiibo, buffer.data(), sizeof(amiibo));
     nfc_tag_load->Signal();
+    return true;
 }
 const Kernel::SharedPtr<Kernel::Event>& Module::Interface::GetNFCEvent() const {
     return nfc_tag_load;

src/core/hle/service/nfp/nfp.h

@@ -32,7 +32,7 @@ public:
         static_assert(sizeof(AmiiboFile) == 0x94, "AmiiboFile is an invalid size");
 
         void CreateUserInterface(Kernel::HLERequestContext& ctx);
-        void LoadAmiibo(const std::vector<u8>& buffer);
+        bool LoadAmiibo(const std::vector<u8>& buffer);
         const Kernel::SharedPtr<Kernel::Event>& GetNFCEvent() const;
         const AmiiboFile& GetAmiiboBuffer() const;
 

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

@@ -31,7 +31,7 @@ void BufferQueue::SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer)
     buffer_wait_event->Signal();
 }
 
-boost::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) {
+std::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) {
     auto itr = std::find_if(queue.begin(), queue.end(), [&](const Buffer& buffer) {
         // Only consider free buffers. Buffers become free once again after they've been Acquired
         // and Released by the compositor, see the NVFlinger::Compose method.
@@ -44,7 +44,7 @@ boost::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) {
     });
 
     if (itr == queue.end()) {
-        return boost::none;
+        return {};
     }
 
     itr->status = Buffer::Status::Dequeued;
@@ -70,12 +70,12 @@ void BufferQueue::QueueBuffer(u32 slot, BufferTransformFlags transform,
     itr->crop_rect = crop_rect;
 }
 
-boost::optional<const BufferQueue::Buffer&> BufferQueue::AcquireBuffer() {
+std::optional<std::reference_wrapper<const BufferQueue::Buffer>> BufferQueue::AcquireBuffer() {
     auto itr = std::find_if(queue.begin(), queue.end(), [](const Buffer& buffer) {
         return buffer.status == Buffer::Status::Queued;
     });
     if (itr == queue.end())
-        return boost::none;
+        return {};
     itr->status = Buffer::Status::Acquired;
     return *itr;
 }

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

@@ -4,8 +4,9 @@
 
 #pragma once
 
+#include <optional>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/common_funcs.h"
 #include "common/math_util.h"
 #include "common/swap.h"
@@ -57,9 +58,9 @@ public:
         /// Rotate source image 90 degrees clockwise
         Rotate90 = 0x04,
         /// Rotate source image 180 degrees
-        Roate180 = 0x03,
+        Rotate180 = 0x03,
         /// Rotate source image 270 degrees clockwise
-        Roate270 = 0x07,
+        Rotate270 = 0x07,
     };
 
     struct Buffer {
@@ -73,11 +74,11 @@ public:
     };
 
     void SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer);
-    boost::optional<u32> DequeueBuffer(u32 width, u32 height);
+    std::optional<u32> DequeueBuffer(u32 width, u32 height);
     const IGBPBuffer& RequestBuffer(u32 slot) const;
     void QueueBuffer(u32 slot, BufferTransformFlags transform,
                      const MathUtil::Rectangle<int>& crop_rect);
-    boost::optional<const Buffer&> AcquireBuffer();
+    std::optional<std::reference_wrapper<const Buffer>> AcquireBuffer();
     void ReleaseBuffer(u32 slot);
     u32 Query(QueryType type);
 

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

@@ -3,7 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
-#include <boost/optional.hpp>
+#include <optional>
 
 #include "common/alignment.h"
 #include "common/assert.h"
@@ -134,7 +134,7 @@ void NVFlinger::Compose() {
 
         MicroProfileFlip();
 
-        if (buffer == boost::none) {
+        if (!buffer) {
             auto& system_instance = Core::System::GetInstance();
 
             // There was no queued buffer to draw, render previous frame
@@ -143,7 +143,7 @@ void NVFlinger::Compose() {
             continue;
         }
 
-        auto& igbp_buffer = buffer->igbp_buffer;
+        auto& igbp_buffer = buffer->get().igbp_buffer;
 
         // Now send the buffer to the GPU for drawing.
         // TODO(Subv): Support more than just disp0. The display device selection is probably based
@@ -152,10 +152,10 @@ void NVFlinger::Compose() {
         ASSERT(nvdisp);
 
         nvdisp->flip(igbp_buffer.gpu_buffer_id, igbp_buffer.offset, igbp_buffer.format,
-                     igbp_buffer.width, igbp_buffer.height, igbp_buffer.stride, buffer->transform,
-                     buffer->crop_rect);
+                     igbp_buffer.width, igbp_buffer.height, igbp_buffer.stride,
+                     buffer->get().transform, buffer->get().crop_rect);
 
-        buffer_queue->ReleaseBuffer(buffer->slot);
+        buffer_queue->ReleaseBuffer(buffer->get().slot);
     }
 }
 

src/core/hle/service/usb/usb.cpp

@@ -67,15 +67,15 @@ public:
     explicit IClientEpSession() : ServiceFramework{"IClientEpSession"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "Unknown1"},
-            {1, nullptr, "Unknown2"},
-            {2, nullptr, "Unknown3"},
-            {3, nullptr, "Unknown4"},
+            {0, nullptr, "Open"},
+            {1, nullptr, "Close"},
+            {2, nullptr, "Unknown1"},
+            {3, nullptr, "Populate"},
             {4, nullptr, "PostBufferAsync"},
-            {5, nullptr, "Unknown5"},
-            {6, nullptr, "Unknown6"},
-            {7, nullptr, "Unknown7"},
-            {8, nullptr, "Unknown8"},
+            {5, nullptr, "GetXferReport"},
+            {6, nullptr, "Unknown2"},
+            {7, nullptr, "Unknown3"},
+            {8, nullptr, "Unknown4"},
         };
         // clang-format on
 
@@ -89,15 +89,15 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "Unknown1"},
-            {1, nullptr, "Unknown2"},
-            {2, nullptr, "Unknown3"},
-            {3, nullptr, "Unknown4"},
-            {4, nullptr, "Unknown5"},
+            {1, nullptr, "SetInterface"},
+            {2, nullptr, "GetInterface"},
+            {3, nullptr, "GetAlternateInterface"},
+            {4, nullptr, "GetCurrentFrame"},
             {5, nullptr, "CtrlXferAsync"},
-            {6, nullptr, "Unknown6"},
+            {6, nullptr, "Unknown2"},
             {7, nullptr, "GetCtrlXferReport"},
-            {8, nullptr, "Unknown7"},
-            {9, nullptr, "GetClientEpSession"},
+            {8, nullptr, "ResetDevice"},
+            {9, nullptr, "OpenUsbEp"},
         };
         // clang-format on
 
@@ -111,13 +111,14 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "BindClientProcess"},
-            {1, nullptr, "Unknown1"},
-            {2, nullptr, "Unknown2"},
-            {3, nullptr, "Unknown3"},
-            {4, nullptr, "Unknown4"},
-            {5, nullptr, "Unknown5"},
+            {1, nullptr, "QueryAllInterfaces"},
+            {2, nullptr, "QueryAvailableInterfaces"},
+            {3, nullptr, "QueryAcquiredInterfaces"},
+            {4, nullptr, "CreateInterfaceAvailableEvent"},
+            {5, nullptr, "DestroyInterfaceAvailableEvent"},
             {6, nullptr, "GetInterfaceStateChangeEvent"},
-            {7, nullptr, "GetClientIfSession"},
+            {7, nullptr, "AcquireUsbIf"},
+            {8, nullptr, "Unknown1"},
         };
         // clang-format on
 
@@ -131,11 +132,11 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "BindNoticeEvent"},
-            {1, nullptr, "Unknown1"},
+            {1, nullptr, "UnbindNoticeEvent"},
             {2, nullptr, "GetStatus"},
             {3, nullptr, "GetNotice"},
-            {4, nullptr, "Unknown2"},
-            {5, nullptr, "Unknown3"},
+            {4, nullptr, "EnablePowerRequestNotice"},
+            {5, nullptr, "DisablePowerRequestNotice"},
             {6, nullptr, "ReplyPowerRequest"},
         };
         // clang-format on

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

@@ -6,9 +6,10 @@
 #include <array>
 #include <cstring>
 #include <memory>
+#include <optional>
 #include <type_traits>
 #include <utility>
-#include <boost/optional.hpp>
+
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "common/common_funcs.h"
@@ -506,9 +507,9 @@ private:
             IGBPDequeueBufferRequestParcel request{ctx.ReadBuffer()};
             const u32 width{request.data.width};
             const u32 height{request.data.height};
-            boost::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
+            std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
 
-            if (slot != boost::none) {
+            if (slot) {
                 // Buffer is available
                 IGBPDequeueBufferResponseParcel response{*slot};
                 ctx.WriteBuffer(response.Serialize());
@@ -520,7 +521,7 @@ private:
                         Kernel::ThreadWakeupReason reason) {
                         // Repeat TransactParcel DequeueBuffer when a buffer is available
                         auto buffer_queue = nv_flinger->GetBufferQueue(id);
-                        boost::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
+                        std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
                         IGBPDequeueBufferResponseParcel response{*slot};
                         ctx.WriteBuffer(response.Serialize());
                         IPC::ResponseBuilder rb{ctx, 2};

src/core/loader/loader.h

@@ -6,10 +6,11 @@
 
 #include <iosfwd>
 #include <memory>
+#include <optional>
 #include <string>
 #include <utility>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/common_types.h"
 #include "core/file_sys/vfs.h"
 
@@ -145,7 +146,7 @@ public:
      * information.
      * @returns A pair with the optional system mode, and and the status.
      */
-    virtual std::pair<boost::optional<u32>, ResultStatus> LoadKernelSystemMode() {
+    virtual std::pair<std::optional<u32>, ResultStatus> LoadKernelSystemMode() {
         // 96MB allocated to the application.
         return std::make_pair(2, ResultStatus::Success);
     }

src/core/loader/nro.cpp

@@ -127,18 +127,23 @@ static constexpr u32 PageAlignSize(u32 size) {
     return (size + Memory::PAGE_MASK) & ~Memory::PAGE_MASK;
 }
 
-bool AppLoader_NRO::LoadNro(const FileSys::VfsFile& file, VAddr load_base) {
-    // Read NSO header
-    NroHeader nro_header{};
-    if (sizeof(NroHeader) != file.ReadObject(&nro_header)) {
+/*static*/ bool AppLoader_NRO::LoadNro(const std::vector<u8>& data, const std::string& name,
+                                       VAddr load_base) {
+
+    if (data.size() < sizeof(NroHeader)) {
         return {};
     }
+
+    // Read NSO header
+    NroHeader nro_header{};
+    std::memcpy(&nro_header, data.data(), sizeof(NroHeader));
     if (nro_header.magic != Common::MakeMagic('N', 'R', 'O', '0')) {
         return {};
     }
 
     // Build program image
-    std::vector<u8> program_image = file.ReadBytes(PageAlignSize(nro_header.file_size));
+    std::vector<u8> program_image(PageAlignSize(nro_header.file_size));
+    std::memcpy(program_image.data(), data.data(), program_image.size());
     if (program_image.size() != PageAlignSize(nro_header.file_size)) {
         return {};
     }
@@ -182,11 +187,15 @@ bool AppLoader_NRO::LoadNro(const FileSys::VfsFile& file, VAddr load_base) {
     Core::CurrentProcess()->LoadModule(std::move(codeset), load_base);
 
     // Register module with GDBStub
-    GDBStub::RegisterModule(file.GetName(), load_base, load_base);
+    GDBStub::RegisterModule(name, load_base, load_base);
 
     return true;
 }
 
+bool AppLoader_NRO::LoadNro(const FileSys::VfsFile& file, VAddr load_base) {
+    return AppLoader_NRO::LoadNro(file.ReadAllBytes(), file.GetName(), load_base);
+}
+
 ResultStatus AppLoader_NRO::Load(Kernel::Process& process) {
     if (is_loaded) {
         return ResultStatus::ErrorAlreadyLoaded;

src/core/loader/nro.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <string>
+#include <vector>
 #include "common/common_types.h"
 #include "core/loader/linker.h"
 #include "core/loader/loader.h"
@@ -40,6 +41,8 @@ public:
     ResultStatus ReadTitle(std::string& title) override;
     bool IsRomFSUpdatable() const override;
 
+    static bool LoadNro(const std::vector<u8>& data, const std::string& name, VAddr load_base);
+
 private:
     bool LoadNro(const FileSys::VfsFile& file, VAddr load_base);
 

src/core/loader/nsp.cpp

@@ -36,6 +36,16 @@ AppLoader_NSP::AppLoader_NSP(FileSys::VirtualFile file)
 
     std::tie(nacp_file, icon_file) =
         FileSys::PatchManager(nsp->GetProgramTitleID()).ParseControlNCA(*control_nca);
+
+    if (nsp->IsExtractedType()) {
+        secondary_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(nsp->GetExeFS());
+    } else {
+        if (title_id == 0)
+            return;
+
+        secondary_loader = std::make_unique<AppLoader_NCA>(
+            nsp->GetNCAFile(title_id, FileSys::ContentRecordType::Program));
+    }
 }
 
 AppLoader_NSP::~AppLoader_NSP() = default;
@@ -67,26 +77,19 @@ ResultStatus AppLoader_NSP::Load(Kernel::Process& process) {
         return ResultStatus::ErrorAlreadyLoaded;
     }
 
-    if (nsp->IsExtractedType()) {
-        secondary_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(nsp->GetExeFS());
-    } else {
-        if (title_id == 0)
-            return ResultStatus::ErrorNSPMissingProgramNCA;
+    if (title_id == 0)
+        return ResultStatus::ErrorNSPMissingProgramNCA;
 
-        secondary_loader = std::make_unique<AppLoader_NCA>(
-            nsp->GetNCAFile(title_id, FileSys::ContentRecordType::Program));
+    if (nsp->GetStatus() != ResultStatus::Success)
+        return nsp->GetStatus();
 
-        if (nsp->GetStatus() != ResultStatus::Success)
-            return nsp->GetStatus();
+    if (nsp->GetProgramStatus(title_id) != ResultStatus::Success)
+        return nsp->GetProgramStatus(title_id);
 
-        if (nsp->GetProgramStatus(title_id) != ResultStatus::Success)
-            return nsp->GetProgramStatus(title_id);
-
-        if (nsp->GetNCA(title_id, FileSys::ContentRecordType::Program) == nullptr) {
-            if (!Core::Crypto::KeyManager::KeyFileExists(false))
-                return ResultStatus::ErrorMissingProductionKeyFile;
-            return ResultStatus::ErrorNSPMissingProgramNCA;
-        }
+    if (nsp->GetNCA(title_id, FileSys::ContentRecordType::Program) == nullptr) {
+        if (!Core::Crypto::KeyManager::KeyFileExists(false))
+            return ResultStatus::ErrorMissingProductionKeyFile;
+        return ResultStatus::ErrorNSPMissingProgramNCA;
     }
 
     const auto result = secondary_loader->Load(process);

src/core/memory.cpp

@@ -4,9 +4,9 @@
 
 #include <algorithm>
 #include <cstring>
+#include <optional>
 #include <utility>
 
-#include <boost/optional.hpp>
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "common/logging/log.h"

src/core/memory_hook.h

@@ -5,7 +5,8 @@
 #pragma once
 
 #include <memory>
-#include <boost/optional.hpp>
+#include <optional>
+
 #include "common/common_types.h"
 
 namespace Memory {
@@ -18,19 +19,19 @@ namespace Memory {
  *
  * A hook may be mapped to multiple regions of memory.
  *
- * If a boost::none or false is returned from a function, the read/write request is passed through
+ * If a std::nullopt or false is returned from a function, the read/write request is passed through
  * to the underlying memory region.
  */
 class MemoryHook {
 public:
     virtual ~MemoryHook();
 
-    virtual boost::optional<bool> IsValidAddress(VAddr addr) = 0;
+    virtual std::optional<bool> IsValidAddress(VAddr addr) = 0;
 
-    virtual boost::optional<u8> Read8(VAddr addr) = 0;
-    virtual boost::optional<u16> Read16(VAddr addr) = 0;
-    virtual boost::optional<u32> Read32(VAddr addr) = 0;
-    virtual boost::optional<u64> Read64(VAddr addr) = 0;
+    virtual std::optional<u8> Read8(VAddr addr) = 0;
+    virtual std::optional<u16> Read16(VAddr addr) = 0;
+    virtual std::optional<u32> Read32(VAddr addr) = 0;
+    virtual std::optional<u64> Read64(VAddr addr) = 0;
 
     virtual bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) = 0;
 

src/core/perf_stats.cpp

@@ -74,10 +74,6 @@ double PerfStats::GetLastFrameTimeScale() {
 }
 
 void FrameLimiter::DoFrameLimiting(microseconds current_system_time_us) {
-    // Max lag caused by slow frames. Can be adjusted to compensate for too many slow frames. Higher
-    // values increase the time needed to recover and limit framerate again after spikes.
-    constexpr microseconds MAX_LAG_TIME_US = 25000us;
-
     if (!Settings::values.use_frame_limit) {
         return;
     }

src/core/settings.h

@@ -114,7 +114,7 @@ struct Values {
     // System
     bool use_docked_mode;
     bool enable_nfc;
-    std::string username;
+    int current_user;
     int language_index;
 
     // Controls

src/core/telemetry_session.cpp

@@ -184,4 +184,13 @@ TelemetrySession::~TelemetrySession() {
     backend = nullptr;
 }
 
+bool TelemetrySession::SubmitTestcase() {
+#ifdef ENABLE_WEB_SERVICE
+    field_collection.Accept(*backend);
+    return backend->SubmitTestcase();
+#else
+    return false;
+#endif
+}
+
 } // namespace Core

src/core/telemetry_session.h

@@ -31,6 +31,12 @@ public:
         field_collection.AddField(type, name, std::move(value));
     }
 
+    /**
+     * Submits a Testcase.
+     * @returns A bool indicating whether the submission succeeded
+     */
+    bool SubmitTestcase();
+
 private:
     Telemetry::FieldCollection field_collection; ///< Tracks all added fields for the session
     std::unique_ptr<Telemetry::VisitorInterface> backend; ///< Backend interface that logs fields

src/tests/core/arm/arm_test_common.cpp

@@ -64,11 +64,11 @@ void TestEnvironment::ClearWriteRecords() {
 
 TestEnvironment::TestMemory::~TestMemory() {}
 
-boost::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) {
+std::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) {
     return true;
 }
 
-boost::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
+std::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
     const auto iter = data.find(addr);
 
     if (iter == data.end()) {
@@ -79,15 +79,15 @@ boost::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
     return iter->second;
 }
 
-boost::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) {
+std::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) {
     return *Read8(addr) | static_cast<u16>(*Read8(addr + 1)) << 8;
 }
 
-boost::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) {
+std::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) {
     return *Read16(addr) | static_cast<u32>(*Read16(addr + 2)) << 16;
 }
 
-boost::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) {
+std::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) {
     return *Read32(addr) | static_cast<u64>(*Read32(addr + 4)) << 32;
 }
 

src/tests/core/arm/arm_test_common.h

@@ -64,12 +64,12 @@ private:
 
         ~TestMemory() override;
 
-        boost::optional<bool> IsValidAddress(VAddr addr) override;
+        std::optional<bool> IsValidAddress(VAddr addr) override;
 
-        boost::optional<u8> Read8(VAddr addr) override;
-        boost::optional<u16> Read16(VAddr addr) override;
-        boost::optional<u32> Read32(VAddr addr) override;
-        boost::optional<u64> Read64(VAddr addr) override;
+        std::optional<u8> Read8(VAddr addr) override;
+        std::optional<u16> Read16(VAddr addr) override;
+        std::optional<u32> Read32(VAddr addr) override;
+        std::optional<u64> Read64(VAddr addr) override;
 
         bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) override;
 

src/video_core/CMakeLists.txt

@@ -33,6 +33,7 @@ add_library(video_core STATIC
     renderer_opengl/gl_rasterizer.h
     renderer_opengl/gl_rasterizer_cache.cpp
     renderer_opengl/gl_rasterizer_cache.h
+    renderer_opengl/gl_resource_manager.cpp
     renderer_opengl/gl_resource_manager.h
     renderer_opengl/gl_shader_cache.cpp
     renderer_opengl/gl_shader_cache.h
@@ -51,6 +52,10 @@ add_library(video_core STATIC
     renderer_opengl/maxwell_to_gl.h
     renderer_opengl/renderer_opengl.cpp
     renderer_opengl/renderer_opengl.h
+    renderer_opengl/utils.cpp
+    renderer_opengl/utils.h
+    surface.cpp
+    surface.h
     textures/astc.cpp
     textures/astc.h
     textures/decoders.cpp

src/video_core/command_processor.cpp

@@ -81,7 +81,7 @@ void GPU::ProcessCommandLists(const std::vector<CommandListHeader>& commands) {
     for (auto entry : commands) {
         Tegra::GPUVAddr address = entry.Address();
         u32 size = entry.sz;
-        const boost::optional<VAddr> head_address = memory_manager->GpuToCpuAddress(address);
+        const std::optional<VAddr> head_address = memory_manager->GpuToCpuAddress(address);
         VAddr current_addr = *head_address;
         while (current_addr < *head_address + size * sizeof(CommandHeader)) {
             const CommandHeader header = {Memory::Read32(current_addr)};

src/video_core/engines/maxwell_3d.cpp

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <cinttypes>
+#include <cstring>
 #include "common/assert.h"
 #include "core/core.h"
 #include "core/core_timing.h"
@@ -19,21 +20,56 @@ namespace Tegra::Engines {
 constexpr u32 MacroRegistersStart = 0xE00;
 
 Maxwell3D::Maxwell3D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
-    : memory_manager(memory_manager), rasterizer{rasterizer}, macro_interpreter(*this) {}
+    : memory_manager(memory_manager), rasterizer{rasterizer}, macro_interpreter(*this) {
+    InitializeRegisterDefaults();
+}
+
+void Maxwell3D::InitializeRegisterDefaults() {
+    // Initializes registers to their default values - what games expect them to be at boot. This is
+    // for certain registers that may not be explicitly set by games.
+
+    // Reset all registers to zero
+    std::memset(&regs, 0, sizeof(regs));
+
+    // Depth range near/far is not always set, but is expected to be the default 0.0f, 1.0f. This is
+    // needed for ARMS.
+    for (std::size_t viewport{}; viewport < Regs::NumViewports; ++viewport) {
+        regs.viewport[viewport].depth_range_near = 0.0f;
+        regs.viewport[viewport].depth_range_far = 1.0f;
+    }
+    // Doom and Bomberman seems to use the uninitialized registers and just enable blend
+    // so initialize blend registers with sane values
+    regs.blend.equation_rgb = Regs::Blend::Equation::Add;
+    regs.blend.factor_source_rgb = Regs::Blend::Factor::One;
+    regs.blend.factor_dest_rgb = Regs::Blend::Factor::Zero;
+    regs.blend.equation_a = Regs::Blend::Equation::Add;
+    regs.blend.factor_source_a = Regs::Blend::Factor::One;
+    regs.blend.factor_dest_a = Regs::Blend::Factor::Zero;
+    for (std::size_t blend_index = 0; blend_index < Regs::NumRenderTargets; blend_index++) {
+        regs.independent_blend[blend_index].equation_rgb = Regs::Blend::Equation::Add;
+        regs.independent_blend[blend_index].factor_source_rgb = Regs::Blend::Factor::One;
+        regs.independent_blend[blend_index].factor_dest_rgb = Regs::Blend::Factor::Zero;
+        regs.independent_blend[blend_index].equation_a = Regs::Blend::Equation::Add;
+        regs.independent_blend[blend_index].factor_source_a = Regs::Blend::Factor::One;
+        regs.independent_blend[blend_index].factor_dest_a = Regs::Blend::Factor::Zero;
+    }
+}
 
 void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
     // Reset the current macro.
     executing_macro = 0;
 
-    // The requested macro must have been uploaded already.
-    auto macro_code = uploaded_macros.find(method);
-    if (macro_code == uploaded_macros.end()) {
-        LOG_ERROR(HW_GPU, "Macro {:04X} was not uploaded", method);
+    // Lookup the macro offset
+    const u32 entry{(method - MacroRegistersStart) >> 1};
+    const auto& search{macro_offsets.find(entry)};
+    if (search == macro_offsets.end()) {
+        LOG_CRITICAL(HW_GPU, "macro not found for method 0x{:X}!", method);
+        UNREACHABLE();
         return;
     }
 
     // Execute the current macro.
-    macro_interpreter.Execute(macro_code->second, std::move(parameters));
+    macro_interpreter.Execute(search->second, std::move(parameters));
 }
 
 void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
@@ -79,6 +115,10 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
         ProcessMacroUpload(value);
         break;
     }
+    case MAXWELL3D_REG_INDEX(macros.bind): {
+        ProcessMacroBind(value);
+        break;
+    }
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]):
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]):
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]):
@@ -140,22 +180,25 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
 }
 
 void Maxwell3D::ProcessMacroUpload(u32 data) {
-    // Store the uploaded macro code to interpret them when they're called.
-    auto& macro = uploaded_macros[regs.macros.entry * 2 + MacroRegistersStart];
-    macro.push_back(data);
+    ASSERT_MSG(regs.macros.upload_address < macro_memory.size(),
+               "upload_address exceeded macro_memory size!");
+    macro_memory[regs.macros.upload_address++] = data;
+}
+
+void Maxwell3D::ProcessMacroBind(u32 data) {
+    macro_offsets[regs.macros.entry] = data;
 }
 
 void Maxwell3D::ProcessQueryGet() {
     GPUVAddr sequence_address = regs.query.QueryAddress();
     // Since the sequence address is given as a GPU VAddr, we have to convert it to an application
     // VAddr before writing.
-    boost::optional<VAddr> address = memory_manager.GpuToCpuAddress(sequence_address);
+    std::optional<VAddr> address = memory_manager.GpuToCpuAddress(sequence_address);
 
     // TODO(Subv): Support the other query units.
     ASSERT_MSG(regs.query.query_get.unit == Regs::QueryUnit::Crop,
                "Units other than CROP are unimplemented");
 
-    u32 value = Memory::Read32(*address);
     u64 result = 0;
 
     // TODO(Subv): Support the other query variables
@@ -268,7 +311,7 @@ void Maxwell3D::ProcessCBData(u32 value) {
     // Don't allow writing past the end of the buffer.
     ASSERT(regs.const_buffer.cb_pos + sizeof(u32) <= regs.const_buffer.cb_size);
 
-    boost::optional<VAddr> address =
+    std::optional<VAddr> address =
         memory_manager.GpuToCpuAddress(buffer_address + regs.const_buffer.cb_pos);
 
     Memory::Write32(*address, value);
@@ -281,7 +324,7 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const {
     GPUVAddr tic_base_address = regs.tic.TICAddress();
 
     GPUVAddr tic_address_gpu = tic_base_address + tic_index * sizeof(Texture::TICEntry);
-    boost::optional<VAddr> tic_address_cpu = memory_manager.GpuToCpuAddress(tic_address_gpu);
+    std::optional<VAddr> tic_address_cpu = memory_manager.GpuToCpuAddress(tic_address_gpu);
 
     Texture::TICEntry tic_entry;
     Memory::ReadBlock(*tic_address_cpu, &tic_entry, sizeof(Texture::TICEntry));
@@ -305,7 +348,7 @@ Texture::TSCEntry Maxwell3D::GetTSCEntry(u32 tsc_index) const {
     GPUVAddr tsc_base_address = regs.tsc.TSCAddress();
 
     GPUVAddr tsc_address_gpu = tsc_base_address + tsc_index * sizeof(Texture::TSCEntry);
-    boost::optional<VAddr> tsc_address_cpu = memory_manager.GpuToCpuAddress(tsc_address_gpu);
+    std::optional<VAddr> tsc_address_cpu = memory_manager.GpuToCpuAddress(tsc_address_gpu);
 
     Texture::TSCEntry tsc_entry;
     Memory::ReadBlock(*tsc_address_cpu, &tsc_entry, sizeof(Texture::TSCEntry));
@@ -369,7 +412,7 @@ Texture::FullTextureInfo Maxwell3D::GetStageTexture(Regs::ShaderStage stage,
 
     ASSERT(tex_info_address < tex_info_buffer.address + tex_info_buffer.size);
 
-    boost::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address);
+    std::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address);
     Texture::TextureHandle tex_handle{Memory::Read32(*tex_address_cpu)};
 
     Texture::FullTextureInfo tex_info{};

src/video_core/engines/maxwell_3d.h

@@ -462,6 +462,16 @@ public:
             }
         };
 
+        struct ColorMask {
+            union {
+                u32 raw;
+                BitField<0, 4, u32> R;
+                BitField<4, 4, u32> G;
+                BitField<8, 4, u32> B;
+                BitField<12, 4, u32> A;
+            };
+        };
+
         bool IsShaderConfigEnabled(std::size_t index) const {
             // The VertexB is always enabled.
             if (index == static_cast<std::size_t>(Regs::ShaderProgram::VertexB)) {
@@ -475,12 +485,13 @@ public:
                 INSERT_PADDING_WORDS(0x45);
 
                 struct {
-                    INSERT_PADDING_WORDS(1);
+                    u32 upload_address;
                     u32 data;
                     u32 entry;
+                    u32 bind;
                 } macros;
 
-                INSERT_PADDING_WORDS(0x189);
+                INSERT_PADDING_WORDS(0x188);
 
                 u32 tfb_enabled;
 
@@ -570,7 +581,11 @@ public:
                 u32 stencil_back_mask;
                 u32 stencil_back_func_mask;
 
-                INSERT_PADDING_WORDS(0x13);
+                INSERT_PADDING_WORDS(0xC);
+
+                u32 color_mask_common;
+
+                INSERT_PADDING_WORDS(0x6);
 
                 u32 rt_separate_frag_data;
 
@@ -645,8 +660,14 @@ public:
                 ComparisonOp depth_test_func;
                 float alpha_test_ref;
                 ComparisonOp alpha_test_func;
-
-                INSERT_PADDING_WORDS(0x9);
+                u32 draw_tfb_stride;
+                struct {
+                    float r;
+                    float g;
+                    float b;
+                    float a;
+                } blend_color;
+                INSERT_PADDING_WORDS(0x4);
 
                 struct {
                     u32 separate_alpha;
@@ -723,7 +744,11 @@ public:
                 StencilOp stencil_back_op_zpass;
                 ComparisonOp stencil_back_func_func;
 
-                INSERT_PADDING_WORDS(0x17);
+                INSERT_PADDING_WORDS(0x4);
+
+                u32 framebuffer_srgb;
+
+                INSERT_PADDING_WORDS(0x12);
 
                 union {
                     BitField<2, 1, u32> coord_origin;
@@ -751,7 +776,14 @@ public:
                     };
                 } draw;
 
-                INSERT_PADDING_WORDS(0x6B);
+                INSERT_PADDING_WORDS(0xA);
+
+                struct {
+                    u32 enabled;
+                    u32 index;
+                } primitive_restart;
+
+                INSERT_PADDING_WORDS(0x5F);
 
                 struct {
                     u32 start_addr_high;
@@ -829,8 +861,9 @@ public:
                     BitField<6, 4, u32> RT;
                     BitField<10, 11, u32> layer;
                 } clear_buffers;
-
-                INSERT_PADDING_WORDS(0x4B);
+                INSERT_PADDING_WORDS(0xB);
+                std::array<ColorMask, NumRenderTargets> color_mask;
+                INSERT_PADDING_WORDS(0x38);
 
                 struct {
                     u32 query_address_high;
@@ -983,10 +1016,25 @@ public:
     /// Returns the texture information for a specific texture in a specific shader stage.
     Texture::FullTextureInfo GetStageTexture(Regs::ShaderStage stage, std::size_t offset) const;
 
+    /// Memory for macro code - it's undetermined how big this is, however 1MB is much larger than
+    /// we've seen used.
+    using MacroMemory = std::array<u32, 0x40000>;
+
+    /// Gets a reference to macro memory.
+    const MacroMemory& GetMacroMemory() const {
+        return macro_memory;
+    }
+
 private:
+    void InitializeRegisterDefaults();
+
     VideoCore::RasterizerInterface& rasterizer;
 
-    std::unordered_map<u32, std::vector<u32>> uploaded_macros;
+    /// Start offsets of each macro in macro_memory
+    std::unordered_map<u32, u32> macro_offsets;
+
+    /// Memory for macro code
+    MacroMemory macro_memory;
 
     /// Macro method that is currently being executed / being fed parameters.
     u32 executing_macro = 0;
@@ -1009,9 +1057,12 @@ private:
      */
     void CallMacroMethod(u32 method, std::vector<u32> parameters);
 
-    /// Handles writes to the macro uploading registers.
+    /// Handles writes to the macro uploading register.
     void ProcessMacroUpload(u32 data);
 
+    /// Handles writes to the macro bind register.
+    void ProcessMacroBind(u32 data);
+
     /// Handles a write to the CLEAR_BUFFERS register.
     void ProcessClearBuffers();
 
@@ -1045,6 +1096,7 @@ ASSERT_REG_POSITION(scissor_test, 0x380);
 ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5);
 ASSERT_REG_POSITION(stencil_back_mask, 0x3D6);
 ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7);
+ASSERT_REG_POSITION(color_mask_common, 0x3E4);
 ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB);
 ASSERT_REG_POSITION(zeta, 0x3F8);
 ASSERT_REG_POSITION(vertex_attrib_format, 0x458);
@@ -1057,6 +1109,10 @@ ASSERT_REG_POSITION(depth_write_enabled, 0x4BA);
 ASSERT_REG_POSITION(alpha_test_enabled, 0x4BB);
 ASSERT_REG_POSITION(d3d_cull_mode, 0x4C2);
 ASSERT_REG_POSITION(depth_test_func, 0x4C3);
+ASSERT_REG_POSITION(alpha_test_ref, 0x4C4);
+ASSERT_REG_POSITION(alpha_test_func, 0x4C5);
+ASSERT_REG_POSITION(draw_tfb_stride, 0x4C6);
+ASSERT_REG_POSITION(blend_color, 0x4C7);
 ASSERT_REG_POSITION(blend, 0x4CF);
 ASSERT_REG_POSITION(stencil_enable, 0x4E0);
 ASSERT_REG_POSITION(stencil_front_op_fail, 0x4E1);
@@ -1077,14 +1133,17 @@ ASSERT_REG_POSITION(stencil_back_op_fail, 0x566);
 ASSERT_REG_POSITION(stencil_back_op_zfail, 0x567);
 ASSERT_REG_POSITION(stencil_back_op_zpass, 0x568);
 ASSERT_REG_POSITION(stencil_back_func_func, 0x569);
+ASSERT_REG_POSITION(framebuffer_srgb, 0x56E);
 ASSERT_REG_POSITION(point_coord_replace, 0x581);
 ASSERT_REG_POSITION(code_address, 0x582);
 ASSERT_REG_POSITION(draw, 0x585);
+ASSERT_REG_POSITION(primitive_restart, 0x591);
 ASSERT_REG_POSITION(index_array, 0x5F2);
 ASSERT_REG_POSITION(instanced_arrays, 0x620);
 ASSERT_REG_POSITION(cull, 0x646);
 ASSERT_REG_POSITION(logic_op, 0x671);
 ASSERT_REG_POSITION(clear_buffers, 0x674);
+ASSERT_REG_POSITION(color_mask, 0x680);
 ASSERT_REG_POSITION(query, 0x6C0);
 ASSERT_REG_POSITION(vertex_array[0], 0x700);
 ASSERT_REG_POSITION(independent_blend, 0x780);

src/video_core/engines/shader_bytecode.h

@@ -5,12 +5,11 @@
 #pragma once
 
 #include <bitset>
+#include <optional>
 #include <string>
 #include <tuple>
 #include <vector>
 
-#include <boost/optional.hpp>
-
 #include "common/assert.h"
 #include "common/bit_field.h"
 #include "common/common_types.h"
@@ -79,6 +78,7 @@ union Attribute {
     constexpr explicit Attribute(u64 value) : value(value) {}
 
     enum class Index : u64 {
+        PointSize = 6,
         Position = 7,
         Attribute_0 = 8,
         Attribute_31 = 39,
@@ -207,6 +207,16 @@ enum class UniformType : u64 {
     Double = 5,
 };
 
+enum class StoreType : u64 {
+    Unsigned8 = 0,
+    Signed8 = 1,
+    Unsigned16 = 2,
+    Signed16 = 3,
+    Bytes32 = 4,
+    Bytes64 = 5,
+    Bytes128 = 6,
+};
+
 enum class IMinMaxExchange : u64 {
     None = 0,
     XLo = 1,
@@ -567,6 +577,10 @@ union Instruction {
         BitField<55, 1, u64> saturate;
     } fmul32;
 
+    union {
+        BitField<52, 1, u64> generates_cc;
+    } op_32;
+
     union {
         BitField<48, 1, u64> is_signed;
     } shift;
@@ -746,6 +760,18 @@ union Instruction {
         BitField<44, 2, u64> unknown;
     } ld_c;
 
+    union {
+        BitField<48, 3, StoreType> type;
+    } ldst_sl;
+
+    union {
+        BitField<44, 2, u64> unknown;
+    } ld_l;
+
+    union {
+        BitField<44, 2, u64> unknown;
+    } st_l;
+
     union {
         BitField<0, 3, u64> pred0;
         BitField<3, 3, u64> pred3;
@@ -1208,6 +1234,8 @@ union Instruction {
     BitField<61, 1, u64> is_b_imm;
     BitField<60, 1, u64> is_b_gpr;
     BitField<59, 1, u64> is_c_gpr;
+    BitField<20, 24, s64> smem_imm;
+    BitField<0, 5, ControlCode> flow_control_code;
 
     Attribute attribute;
     Sampler sampler;
@@ -1231,8 +1259,12 @@ public:
         BRA,
         PBK,
         LD_A,
+        LD_L,
+        LD_S,
         LD_C,
         ST_A,
+        ST_L,
+        ST_S,
         LDG, // Load from global memory
         STG, // Store in global memory
         TEX,
@@ -1428,7 +1460,7 @@ public:
         Type type;
     };
 
-    static boost::optional<const Matcher&> Decode(Instruction instr) {
+    static std::optional<std::reference_wrapper<const Matcher>> Decode(Instruction instr) {
         static const auto table{GetDecodeTable()};
 
         const auto matches_instruction = [instr](const auto& matcher) {
@@ -1436,7 +1468,8 @@ public:
         };
 
         auto iter = std::find_if(table.begin(), table.end(), matches_instruction);
-        return iter != table.end() ? boost::optional<const Matcher&>(*iter) : boost::none;
+        return iter != table.end() ? std::optional<std::reference_wrapper<const Matcher>>(*iter)
+                                   : std::nullopt;
     }
 
 private:
@@ -1489,8 +1522,12 @@ private:
             INST("111000110100---", Id::BRK, Type::Flow, "BRK"),
             INST("1111000011110---", Id::DEPBAR, Type::Synch, "DEPBAR"),
             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"),
             INST("1110111110010---", Id::LD_C, Type::Memory, "LD_C"),
             INST("1110111111110---", Id::ST_A, Type::Memory, "ST_A"),
+            INST("1110111101011---", Id::ST_S, Type::Memory, "ST_S"),
+            INST("1110111101010---", Id::ST_L, Type::Memory, "ST_L"),
             INST("1110111011010---", Id::LDG, Type::Memory, "LDG"),
             INST("1110111011011---", Id::STG, Type::Memory, "STG"),
             INST("110000----111---", Id::TEX, Type::Memory, "TEX"),
@@ -1626,4 +1663,4 @@ private:
     }
 };
 
-} // namespace Tegra::Shader
+} // namespace Tegra::Shader

src/video_core/engines/shader_header.h

@@ -96,6 +96,11 @@ struct Header {
             }
         } ps;
     };
+
+    u64 GetLocalMemorySize() {
+        return (common1.shader_local_memory_low_size |
+                (common2.shader_local_memory_high_size << 24));
+    }
 };
 
 static_assert(sizeof(Header) == 0x50, "Incorrect structure size");

src/video_core/macro_interpreter.cpp

@@ -11,7 +11,7 @@ namespace Tegra {
 
 MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {}
 
-void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> parameters) {
+void MacroInterpreter::Execute(u32 offset, std::vector<u32> parameters) {
     Reset();
     registers[1] = parameters[0];
     this->parameters = std::move(parameters);
@@ -19,7 +19,7 @@ void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> pa
     // Execute the code until we hit an exit condition.
     bool keep_executing = true;
     while (keep_executing) {
-        keep_executing = Step(code, false);
+        keep_executing = Step(offset, false);
     }
 
     // Assert the the macro used all the input parameters
@@ -29,7 +29,7 @@ void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> pa
 void MacroInterpreter::Reset() {
     registers = {};
     pc = 0;
-    delayed_pc = boost::none;
+    delayed_pc = {};
     method_address.raw = 0;
     parameters.clear();
     // The next parameter index starts at 1, because $r1 already has the value of the first
@@ -37,17 +37,17 @@ void MacroInterpreter::Reset() {
     next_parameter_index = 1;
 }
 
-bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
+bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
     u32 base_address = pc;
 
-    Opcode opcode = GetOpcode(code);
+    Opcode opcode = GetOpcode(offset);
     pc += 4;
 
     // Update the program counter if we were delayed
-    if (delayed_pc != boost::none) {
+    if (delayed_pc) {
         ASSERT(is_delay_slot);
         pc = *delayed_pc;
-        delayed_pc = boost::none;
+        delayed_pc = {};
     }
 
     switch (opcode.operation) {
@@ -108,7 +108,7 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
 
             delayed_pc = base_address + opcode.GetBranchTarget();
             // Execute one more instruction due to the delay slot.
-            return Step(code, true);
+            return Step(offset, true);
         }
         break;
     }
@@ -121,17 +121,18 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
         // Exit has a delay slot, execute the next instruction
         // Note: Executing an exit during a branch delay slot will cause the instruction at the
         // branch target to be executed before exiting.
-        Step(code, true);
+        Step(offset, true);
         return false;
     }
 
     return true;
 }
 
-MacroInterpreter::Opcode MacroInterpreter::GetOpcode(const std::vector<u32>& code) const {
+MacroInterpreter::Opcode MacroInterpreter::GetOpcode(u32 offset) const {
+    const auto& macro_memory{maxwell3d.GetMacroMemory()};
     ASSERT((pc % sizeof(u32)) == 0);
-    ASSERT(pc < code.size() * sizeof(u32));
-    return {code[pc / sizeof(u32)]};
+    ASSERT((pc + offset) < macro_memory.size() * sizeof(u32));
+    return {macro_memory[offset + pc / sizeof(u32)]};
 }
 
 u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const {

src/video_core/macro_interpreter.h

@@ -5,8 +5,9 @@
 #pragma once
 
 #include <array>
+#include <optional>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/bit_field.h"
 #include "common/common_types.h"
 
@@ -21,10 +22,10 @@ public:
 
     /**
      * Executes the macro code with the specified input parameters.
-     * @param code The macro byte code to execute
-     * @param parameters The parameters of the macro
+     * @param offset Offset to start execution at.
+     * @param parameters The parameters of the macro.
      */
-    void Execute(const std::vector<u32>& code, std::vector<u32> parameters);
+    void Execute(u32 offset, std::vector<u32> parameters);
 
 private:
     enum class Operation : u32 {
@@ -109,11 +110,11 @@ private:
     /**
      * Executes a single macro instruction located at the current program counter. Returns whether
      * the interpreter should keep running.
-     * @param code The macro code to execute.
+     * @param offset Offset to start execution at.
      * @param is_delay_slot Whether the current step is being executed due to a delay slot in a
      * previous instruction.
      */
-    bool Step(const std::vector<u32>& code, bool is_delay_slot);
+    bool Step(u32 offset, bool is_delay_slot);
 
     /// Calculates the result of an ALU operation. src_a OP src_b;
     u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const;
@@ -126,7 +127,7 @@ private:
     bool EvaluateBranchCondition(BranchCondition cond, u32 value) const;
 
     /// Reads an opcode at the current program counter location.
-    Opcode GetOpcode(const std::vector<u32>& code) const;
+    Opcode GetOpcode(u32 offset) const;
 
     /// Returns the specified register's value. Register 0 is hardcoded to always return 0.
     u32 GetRegister(u32 register_id) const;
@@ -149,7 +150,7 @@ private:
     Engines::Maxwell3D& maxwell3d;
 
     u32 pc; ///< Current program counter
-    boost::optional<u32>
+    std::optional<u32>
         delayed_pc; ///< Program counter to execute at after the delay slot is executed.
 
     static constexpr std::size_t NumMacroRegisters = 8;

src/video_core/memory_manager.cpp

@@ -4,18 +4,21 @@
 
 #include "common/alignment.h"
 #include "common/assert.h"
+#include "common/logging/log.h"
 #include "video_core/memory_manager.h"
 
 namespace Tegra {
 
 GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
-    boost::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, align);
-    ASSERT(gpu_addr);
+    const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, align, PageStatus::Unmapped)};
 
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(*gpu_addr + offset);
+    ASSERT_MSG(gpu_addr, "unable to find available GPU memory");
+
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(*gpu_addr + offset)};
 
         ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
+
         slot = static_cast<u64>(PageStatus::Allocated);
     }
 
@@ -23,10 +26,11 @@ GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
 }
 
 GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) {
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(gpu_addr + offset);
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(gpu_addr + offset)};
 
         ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
+
         slot = static_cast<u64>(PageStatus::Allocated);
     }
 
@@ -34,17 +38,19 @@ GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) {
 }
 
 GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) {
-    boost::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, PAGE_SIZE);
-    ASSERT(gpu_addr);
+    const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, PAGE_SIZE, PageStatus::Unmapped)};
 
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(*gpu_addr + offset);
+    ASSERT_MSG(gpu_addr, "unable to find available GPU memory");
+
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(*gpu_addr + offset)};
 
         ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
+
         slot = cpu_addr + offset;
     }
 
-    MappedRegion region{cpu_addr, *gpu_addr, size};
+    const MappedRegion region{cpu_addr, *gpu_addr, size};
     mapped_regions.push_back(region);
 
     return *gpu_addr;
@@ -53,14 +59,31 @@ GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) {
 GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size) {
     ASSERT((gpu_addr & PAGE_MASK) == 0);
 
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(gpu_addr + offset);
+    if (PageSlot(gpu_addr) != static_cast<u64>(PageStatus::Allocated)) {
+        // Page has been already mapped. In this case, we must find a new area of memory to use that
+        // is different than the specified one. Super Mario Odyssey hits this scenario when changing
+        // areas, but we do not want to overwrite the old pages.
+        // TODO(bunnei): We need to write a hardware test to confirm this behavior.
+
+        LOG_ERROR(HW_GPU, "attempting to map addr 0x{:016X}, which is not available!", gpu_addr);
+
+        const std::optional<GPUVAddr> new_gpu_addr{
+            FindFreeBlock(gpu_addr, size, PAGE_SIZE, PageStatus::Allocated)};
+
+        ASSERT_MSG(new_gpu_addr, "unable to find available GPU memory");
+
+        gpu_addr = *new_gpu_addr;
+    }
+
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(gpu_addr + offset)};
 
         ASSERT(slot == static_cast<u64>(PageStatus::Allocated));
+
         slot = cpu_addr + offset;
     }
 
-    MappedRegion region{cpu_addr, gpu_addr, size};
+    const MappedRegion region{cpu_addr, gpu_addr, size};
     mapped_regions.push_back(region);
 
     return gpu_addr;
@@ -69,11 +92,12 @@ GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size)
 GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) {
     ASSERT((gpu_addr & PAGE_MASK) == 0);
 
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(gpu_addr + offset);
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(gpu_addr + offset)};
 
         ASSERT(slot != static_cast<u64>(PageStatus::Allocated) &&
                slot != static_cast<u64>(PageStatus::Unmapped));
+
         slot = static_cast<u64>(PageStatus::Unmapped);
     }
 
@@ -97,13 +121,14 @@ GPUVAddr MemoryManager::GetRegionEnd(GPUVAddr region_start) const {
     return {};
 }
 
-boost::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
-    GPUVAddr gpu_addr = 0;
-    u64 free_space = 0;
+std::optional<GPUVAddr> MemoryManager::FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,
+                                                     PageStatus status) {
+    GPUVAddr gpu_addr{region_start};
+    u64 free_space{};
     align = (align + PAGE_MASK) & ~PAGE_MASK;
 
     while (gpu_addr + free_space < MAX_ADDRESS) {
-        if (!IsPageMapped(gpu_addr + free_space)) {
+        if (PageSlot(gpu_addr + free_space) == static_cast<u64>(status)) {
             free_space += PAGE_SIZE;
             if (free_space >= size) {
                 return gpu_addr;
@@ -118,8 +143,8 @@ boost::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
     return {};
 }
 
-boost::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) {
-    VAddr base_addr = PageSlot(gpu_addr);
+std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) {
+    const VAddr base_addr{PageSlot(gpu_addr)};
 
     if (base_addr == static_cast<u64>(PageStatus::Allocated) ||
         base_addr == static_cast<u64>(PageStatus::Unmapped)) {
@@ -133,19 +158,15 @@ std::vector<GPUVAddr> MemoryManager::CpuToGpuAddress(VAddr cpu_addr) const {
     std::vector<GPUVAddr> results;
     for (const auto& region : mapped_regions) {
         if (cpu_addr >= region.cpu_addr && cpu_addr < (region.cpu_addr + region.size)) {
-            u64 offset = cpu_addr - region.cpu_addr;
+            const u64 offset{cpu_addr - region.cpu_addr};
             results.push_back(region.gpu_addr + offset);
         }
     }
     return results;
 }
 
-bool MemoryManager::IsPageMapped(GPUVAddr gpu_addr) {
-    return PageSlot(gpu_addr) != static_cast<u64>(PageStatus::Unmapped);
-}
-
 VAddr& MemoryManager::PageSlot(GPUVAddr gpu_addr) {
-    auto& block = page_table[(gpu_addr >> (PAGE_BITS + PAGE_TABLE_BITS)) & PAGE_TABLE_MASK];
+    auto& block{page_table[(gpu_addr >> (PAGE_BITS + PAGE_TABLE_BITS)) & PAGE_TABLE_MASK]};
     if (!block) {
         block = std::make_unique<PageBlock>();
         block->fill(static_cast<VAddr>(PageStatus::Unmapped));