service/apm: Add the apm:sys service

Adds the basic skeleton of the apm:sys service based off the information
on Switch Brew.

.gitmodules

@@ -25,6 +25,9 @@
 [submodule "unicorn"]
     path = externals/unicorn
     url = https://github.com/yuzu-emu/unicorn
+[submodule "mbedtls"]
+    path = externals/mbedtls
+    url = https://github.com/DarkLordZach/mbedtls
 [submodule "opus"]
-	path = externals/opus
-	url = https://github.com/ogniK5377/opus.git
+    path = externals/opus
+    url = https://github.com/ogniK5377/opus.git

externals/CMakeLists.txt

@@ -35,6 +35,10 @@ set(LZ4_BUNDLED_MODE ON)
 add_subdirectory(lz4/contrib/cmake_unofficial)
 target_include_directories(lz4_static INTERFACE ./lz4/lib)
 
+# mbedtls
+add_subdirectory(mbedtls)
+target_include_directories(mbedtls PUBLIC ./mbedtls/include)
+
 # MicroProfile
 add_library(microprofile INTERFACE)
 target_include_directories(microprofile INTERFACE ./microprofile)

src/audio_core/CMakeLists.txt

@@ -1,9 +1,13 @@
 add_library(audio_core STATIC
     audio_out.cpp
     audio_out.h
+    audio_renderer.cpp
+    audio_renderer.h
     buffer.h
     cubeb_sink.cpp
     cubeb_sink.h
+    codec.cpp
+    codec.h
     null_sink.h
     stream.cpp
     stream.h

src/audio_core/audio_out.cpp

@@ -27,19 +27,19 @@ static Stream::Format ChannelsToStreamFormat(u32 num_channels) {
     return {};
 }
 
-StreamPtr AudioOut::OpenStream(u32 sample_rate, u32 num_channels,
+StreamPtr AudioOut::OpenStream(u32 sample_rate, u32 num_channels, std::string&& name,
                                Stream::ReleaseCallback&& release_callback) {
     if (!sink) {
         const SinkDetails& sink_details = GetSinkDetails(Settings::values.sink_id);
         sink = sink_details.factory(Settings::values.audio_device_id);
     }
 
-    return std::make_shared<Stream>(sample_rate, ChannelsToStreamFormat(num_channels),
-                                    std::move(release_callback),
-                                    sink->AcquireSinkStream(sample_rate, num_channels));
+    return std::make_shared<Stream>(
+        sample_rate, ChannelsToStreamFormat(num_channels), std::move(release_callback),
+        sink->AcquireSinkStream(sample_rate, num_channels, name), std::move(name));
 }
 
-std::vector<u64> AudioOut::GetTagsAndReleaseBuffers(StreamPtr stream, size_t max_count) {
+std::vector<Buffer::Tag> AudioOut::GetTagsAndReleaseBuffers(StreamPtr stream, size_t max_count) {
     return stream->GetTagsAndReleaseBuffers(max_count);
 }
 
@@ -51,7 +51,7 @@ void AudioOut::StopStream(StreamPtr stream) {
     stream->Stop();
 }
 
-bool AudioOut::QueueBuffer(StreamPtr stream, Buffer::Tag tag, std::vector<u8>&& data) {
+bool AudioOut::QueueBuffer(StreamPtr stream, Buffer::Tag tag, std::vector<s16>&& data) {
     return stream->QueueBuffer(std::make_shared<Buffer>(tag, std::move(data)));
 }
 

src/audio_core/audio_out.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <string>
 #include <vector>
 
 #include "audio_core/buffer.h"
@@ -20,11 +21,11 @@ namespace AudioCore {
 class AudioOut {
 public:
     /// Opens a new audio stream
-    StreamPtr OpenStream(u32 sample_rate, u32 num_channels,
+    StreamPtr OpenStream(u32 sample_rate, u32 num_channels, std::string&& name,
                          Stream::ReleaseCallback&& release_callback);
 
     /// Returns a vector of recently released buffers specified by tag for the specified stream
-    std::vector<u64> GetTagsAndReleaseBuffers(StreamPtr stream, size_t max_count);
+    std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(StreamPtr stream, size_t max_count);
 
     /// Starts an audio stream for playback
     void StartStream(StreamPtr stream);
@@ -33,7 +34,7 @@ public:
     void StopStream(StreamPtr stream);
 
     /// Queues a buffer into the specified audio stream, returns true on success
-    bool QueueBuffer(StreamPtr stream, Buffer::Tag tag, std::vector<u8>&& data);
+    bool QueueBuffer(StreamPtr stream, Buffer::Tag tag, std::vector<s16>&& data);
 
 private:
     SinkPtr sink;

src/audio_core/audio_renderer.cpp

@@ -0,0 +1,234 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "audio_core/audio_renderer.h"
+#include "common/assert.h"
+#include "common/logging/log.h"
+#include "core/memory.h"
+
+namespace AudioCore {
+
+constexpr u32 STREAM_SAMPLE_RATE{48000};
+constexpr u32 STREAM_NUM_CHANNELS{2};
+
+AudioRenderer::AudioRenderer(AudioRendererParameter params,
+                             Kernel::SharedPtr<Kernel::Event> buffer_event)
+    : worker_params{params}, buffer_event{buffer_event}, voices(params.voice_count) {
+
+    audio_core = std::make_unique<AudioCore::AudioOut>();
+    stream = audio_core->OpenStream(STREAM_SAMPLE_RATE, STREAM_NUM_CHANNELS, "AudioRenderer",
+                                    [=]() { buffer_event->Signal(); });
+    audio_core->StartStream(stream);
+
+    QueueMixedBuffer(0);
+    QueueMixedBuffer(1);
+    QueueMixedBuffer(2);
+}
+
+std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_params) {
+    // Copy UpdateDataHeader struct
+    UpdateDataHeader config{};
+    std::memcpy(&config, input_params.data(), sizeof(UpdateDataHeader));
+    u32 memory_pool_count = worker_params.effect_count + (worker_params.voice_count * 4);
+
+    // Copy MemoryPoolInfo structs
+    std::vector<MemoryPoolInfo> mem_pool_info(memory_pool_count);
+    std::memcpy(mem_pool_info.data(),
+                input_params.data() + sizeof(UpdateDataHeader) + config.behavior_size,
+                memory_pool_count * sizeof(MemoryPoolInfo));
+
+    // Copy VoiceInfo structs
+    size_t offset{sizeof(UpdateDataHeader) + config.behavior_size + config.memory_pools_size +
+                  config.voice_resource_size};
+    for (auto& voice : voices) {
+        std::memcpy(&voice.Info(), input_params.data() + offset, sizeof(VoiceInfo));
+        offset += sizeof(VoiceInfo);
+    }
+
+    // Update voices
+    for (auto& voice : voices) {
+        voice.UpdateState();
+        if (!voice.GetInfo().is_in_use) {
+            continue;
+        }
+        if (voice.GetInfo().is_new) {
+            voice.SetWaveIndex(voice.GetInfo().wave_buffer_head);
+        }
+    }
+
+    // Update memory pool state
+    std::vector<MemoryPoolEntry> memory_pool(memory_pool_count);
+    for (size_t index = 0; index < memory_pool.size(); ++index) {
+        if (mem_pool_info[index].pool_state == MemoryPoolStates::RequestAttach) {
+            memory_pool[index].state = MemoryPoolStates::Attached;
+        } else if (mem_pool_info[index].pool_state == MemoryPoolStates::RequestDetach) {
+            memory_pool[index].state = MemoryPoolStates::Detached;
+        }
+    }
+
+    // Release previous buffers and queue next ones for playback
+    ReleaseAndQueueBuffers();
+
+    // Copy output header
+    UpdateDataHeader response_data{worker_params};
+    std::vector<u8> output_params(response_data.total_size);
+    std::memcpy(output_params.data(), &response_data, sizeof(UpdateDataHeader));
+
+    // Copy output memory pool entries
+    std::memcpy(output_params.data() + sizeof(UpdateDataHeader), memory_pool.data(),
+                response_data.memory_pools_size);
+
+    // Copy output voice status
+    size_t voice_out_status_offset{sizeof(UpdateDataHeader) + response_data.memory_pools_size};
+    for (const auto& voice : voices) {
+        std::memcpy(output_params.data() + voice_out_status_offset, &voice.GetOutStatus(),
+                    sizeof(VoiceOutStatus));
+        voice_out_status_offset += sizeof(VoiceOutStatus);
+    }
+
+    return output_params;
+}
+
+void AudioRenderer::VoiceState::SetWaveIndex(size_t index) {
+    wave_index = index & 3;
+    is_refresh_pending = true;
+}
+
+std::vector<s16> AudioRenderer::VoiceState::DequeueSamples(size_t sample_count) {
+    if (!IsPlaying()) {
+        return {};
+    }
+
+    if (is_refresh_pending) {
+        RefreshBuffer();
+    }
+
+    const size_t max_size{samples.size() - offset};
+    const size_t dequeue_offset{offset};
+    size_t size{sample_count * STREAM_NUM_CHANNELS};
+    if (size > max_size) {
+        size = max_size;
+    }
+
+    out_status.played_sample_count += size / STREAM_NUM_CHANNELS;
+    offset += size;
+
+    const auto& wave_buffer{info.wave_buffer[wave_index]};
+    if (offset == samples.size()) {
+        offset = 0;
+
+        if (!wave_buffer.is_looping) {
+            SetWaveIndex(wave_index + 1);
+        }
+
+        out_status.wave_buffer_consumed++;
+
+        if (wave_buffer.end_of_stream) {
+            info.play_state = PlayState::Paused;
+        }
+    }
+
+    return {samples.begin() + dequeue_offset, samples.begin() + dequeue_offset + size};
+}
+
+void AudioRenderer::VoiceState::UpdateState() {
+    if (is_in_use && !info.is_in_use) {
+        // No longer in use, reset state
+        is_refresh_pending = true;
+        wave_index = 0;
+        offset = 0;
+        out_status = {};
+    }
+    is_in_use = info.is_in_use;
+}
+
+void AudioRenderer::VoiceState::RefreshBuffer() {
+    std::vector<s16> new_samples(info.wave_buffer[wave_index].buffer_sz / sizeof(s16));
+    Memory::ReadBlock(info.wave_buffer[wave_index].buffer_addr, new_samples.data(),
+                      info.wave_buffer[wave_index].buffer_sz);
+
+    switch (static_cast<Codec::PcmFormat>(info.sample_format)) {
+    case Codec::PcmFormat::Int16: {
+        // PCM16 is played as-is
+        break;
+    }
+    case Codec::PcmFormat::Adpcm: {
+        // Decode ADPCM to PCM16
+        Codec::ADPCM_Coeff coeffs;
+        Memory::ReadBlock(info.additional_params_addr, coeffs.data(), sizeof(Codec::ADPCM_Coeff));
+        new_samples = Codec::DecodeADPCM(reinterpret_cast<u8*>(new_samples.data()),
+                                         new_samples.size() * sizeof(s16), coeffs, adpcm_state);
+        break;
+    }
+    default:
+        LOG_CRITICAL(Audio, "Unimplemented sample_format={}", info.sample_format);
+        UNREACHABLE();
+        break;
+    }
+
+    switch (info.channel_count) {
+    case 1:
+        // 1 channel is upsampled to 2 channel
+        samples.resize(new_samples.size() * 2);
+        for (size_t index = 0; index < new_samples.size(); ++index) {
+            samples[index * 2] = new_samples[index];
+            samples[index * 2 + 1] = new_samples[index];
+        }
+        break;
+    case 2: {
+        // 2 channel is played as is
+        samples = std::move(new_samples);
+        break;
+    }
+    default:
+        LOG_CRITICAL(Audio, "Unimplemented channel_count={}", info.channel_count);
+        UNREACHABLE();
+        break;
+    }
+
+    is_refresh_pending = false;
+}
+
+static constexpr s16 ClampToS16(s32 value) {
+    return static_cast<s16>(std::clamp(value, -32768, 32767));
+}
+
+void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) {
+    constexpr size_t BUFFER_SIZE{512};
+    std::vector<s16> buffer(BUFFER_SIZE * stream->GetNumChannels());
+
+    for (auto& voice : voices) {
+        if (!voice.IsPlaying()) {
+            continue;
+        }
+
+        size_t offset{};
+        s64 samples_remaining{BUFFER_SIZE};
+        while (samples_remaining > 0) {
+            const std::vector<s16> samples{voice.DequeueSamples(samples_remaining)};
+
+            if (samples.empty()) {
+                break;
+            }
+
+            samples_remaining -= samples.size();
+
+            for (const auto& sample : samples) {
+                const s32 buffer_sample{buffer[offset]};
+                buffer[offset++] =
+                    ClampToS16(buffer_sample + static_cast<s32>(sample * voice.GetInfo().volume));
+            }
+        }
+    }
+    audio_core->QueueBuffer(stream, tag, std::move(buffer));
+}
+
+void AudioRenderer::ReleaseAndQueueBuffers() {
+    const auto released_buffers{audio_core->GetTagsAndReleaseBuffers(stream, 2)};
+    for (const auto& tag : released_buffers) {
+        QueueMixedBuffer(tag);
+    }
+}
+
+} // namespace AudioCore

src/audio_core/audio_renderer.h

@@ -0,0 +1,206 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <memory>
+#include <vector>
+
+#include "audio_core/audio_out.h"
+#include "audio_core/codec.h"
+#include "audio_core/stream.h"
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/hle/kernel/event.h"
+
+namespace AudioCore {
+
+enum class PlayState : u8 {
+    Started = 0,
+    Stopped = 1,
+    Paused = 2,
+};
+
+struct AudioRendererParameter {
+    u32_le sample_rate;
+    u32_le sample_count;
+    u32_le unknown_8;
+    u32_le unknown_c;
+    u32_le voice_count;
+    u32_le sink_count;
+    u32_le effect_count;
+    u32_le unknown_1c;
+    u8 unknown_20;
+    INSERT_PADDING_BYTES(3);
+    u32_le splitter_count;
+    u32_le unknown_2c;
+    INSERT_PADDING_WORDS(1);
+    u32_le revision;
+};
+static_assert(sizeof(AudioRendererParameter) == 52, "AudioRendererParameter is an invalid size");
+
+enum class MemoryPoolStates : u32 { // Should be LE
+    Invalid = 0x0,
+    Unknown = 0x1,
+    RequestDetach = 0x2,
+    Detached = 0x3,
+    RequestAttach = 0x4,
+    Attached = 0x5,
+    Released = 0x6,
+};
+
+struct MemoryPoolEntry {
+    MemoryPoolStates state;
+    u32_le unknown_4;
+    u32_le unknown_8;
+    u32_le unknown_c;
+};
+static_assert(sizeof(MemoryPoolEntry) == 0x10, "MemoryPoolEntry has wrong size");
+
+struct MemoryPoolInfo {
+    u64_le pool_address;
+    u64_le pool_size;
+    MemoryPoolStates pool_state;
+    INSERT_PADDING_WORDS(3); // Unknown
+};
+static_assert(sizeof(MemoryPoolInfo) == 0x20, "MemoryPoolInfo has wrong size");
+struct BiquadFilter {
+    u8 enable;
+    INSERT_PADDING_BYTES(1);
+    std::array<s16_le, 3> numerator;
+    std::array<s16_le, 2> denominator;
+};
+static_assert(sizeof(BiquadFilter) == 0xc, "BiquadFilter has wrong size");
+
+struct WaveBuffer {
+    u64_le buffer_addr;
+    u64_le buffer_sz;
+    s32_le start_sample_offset;
+    s32_le end_sample_offset;
+    u8 is_looping;
+    u8 end_of_stream;
+    u8 sent_to_server;
+    INSERT_PADDING_BYTES(5);
+    u64 context_addr;
+    u64 context_sz;
+    INSERT_PADDING_BYTES(8);
+};
+static_assert(sizeof(WaveBuffer) == 0x38, "WaveBuffer has wrong size");
+
+struct VoiceInfo {
+    u32_le id;
+    u32_le node_id;
+    u8 is_new;
+    u8 is_in_use;
+    PlayState play_state;
+    u8 sample_format;
+    u32_le sample_rate;
+    u32_le priority;
+    u32_le sorting_order;
+    u32_le channel_count;
+    float_le pitch;
+    float_le volume;
+    std::array<BiquadFilter, 2> biquad_filter;
+    u32_le wave_buffer_count;
+    u32_le wave_buffer_head;
+    INSERT_PADDING_WORDS(1);
+    u64_le additional_params_addr;
+    u64_le additional_params_sz;
+    u32_le mix_id;
+    u32_le splitter_info_id;
+    std::array<WaveBuffer, 4> wave_buffer;
+    std::array<u32_le, 6> voice_channel_resource_ids;
+    INSERT_PADDING_BYTES(24);
+};
+static_assert(sizeof(VoiceInfo) == 0x170, "VoiceInfo is wrong size");
+
+struct VoiceOutStatus {
+    u64_le played_sample_count;
+    u32_le wave_buffer_consumed;
+    u32_le voice_drops_count;
+};
+static_assert(sizeof(VoiceOutStatus) == 0x10, "VoiceOutStatus has wrong size");
+
+struct UpdateDataHeader {
+    UpdateDataHeader() {}
+
+    explicit UpdateDataHeader(const AudioRendererParameter& config) {
+        revision = Common::MakeMagic('R', 'E', 'V', '4'); // 5.1.0 Revision
+        behavior_size = 0xb0;
+        memory_pools_size = (config.effect_count + (config.voice_count * 4)) * 0x10;
+        voices_size = config.voice_count * 0x10;
+        voice_resource_size = 0x0;
+        effects_size = config.effect_count * 0x10;
+        mixes_size = 0x0;
+        sinks_size = config.sink_count * 0x20;
+        performance_manager_size = 0x10;
+        total_size = sizeof(UpdateDataHeader) + behavior_size + memory_pools_size + voices_size +
+                     effects_size + sinks_size + performance_manager_size;
+    }
+
+    u32_le revision;
+    u32_le behavior_size;
+    u32_le memory_pools_size;
+    u32_le voices_size;
+    u32_le voice_resource_size;
+    u32_le effects_size;
+    u32_le mixes_size;
+    u32_le sinks_size;
+    u32_le performance_manager_size;
+    INSERT_PADDING_WORDS(6);
+    u32_le total_size;
+};
+static_assert(sizeof(UpdateDataHeader) == 0x40, "UpdateDataHeader has wrong size");
+
+class AudioRenderer {
+public:
+    AudioRenderer(AudioRendererParameter params, Kernel::SharedPtr<Kernel::Event> buffer_event);
+    std::vector<u8> UpdateAudioRenderer(const std::vector<u8>& input_params);
+    void QueueMixedBuffer(Buffer::Tag tag);
+    void ReleaseAndQueueBuffers();
+
+private:
+    class VoiceState {
+    public:
+        bool IsPlaying() const {
+            return is_in_use && info.play_state == PlayState::Started;
+        }
+
+        const VoiceOutStatus& GetOutStatus() const {
+            return out_status;
+        }
+
+        const VoiceInfo& GetInfo() const {
+            return info;
+        }
+
+        VoiceInfo& Info() {
+            return info;
+        }
+
+        void SetWaveIndex(size_t index);
+        std::vector<s16> DequeueSamples(size_t sample_count);
+        void UpdateState();
+        void RefreshBuffer();
+
+    private:
+        bool is_in_use{};
+        bool is_refresh_pending{};
+        size_t wave_index{};
+        size_t offset{};
+        Codec::ADPCMState adpcm_state{};
+        std::vector<s16> samples;
+        VoiceOutStatus out_status{};
+        VoiceInfo info{};
+    };
+
+    AudioRendererParameter worker_params;
+    Kernel::SharedPtr<Kernel::Event> buffer_event;
+    std::vector<VoiceState> voices;
+    std::unique_ptr<AudioCore::AudioOut> audio_core;
+    AudioCore::StreamPtr stream;
+};
+
+} // namespace AudioCore

src/audio_core/buffer.h

@@ -18,11 +18,16 @@ class Buffer {
 public:
     using Tag = u64;
 
-    Buffer(Tag tag, std::vector<u8>&& data) : tag{tag}, data{std::move(data)} {}
+    Buffer(Tag tag, std::vector<s16>&& samples) : tag{tag}, samples{std::move(samples)} {}
 
     /// Returns the raw audio data for the buffer
-    const std::vector<u8>& GetData() const {
-        return data;
+    std::vector<s16>& Samples() {
+        return samples;
+    }
+
+    /// Returns the raw audio data for the buffer
+    const std::vector<s16>& GetSamples() const {
+        return samples;
     }
 
     /// Returns the buffer tag, this is provided by the game to the audout service
@@ -32,7 +37,7 @@ public:
 
 private:
     Tag tag;
-    std::vector<u8> data;
+    std::vector<s16> samples;
 };
 
 using BufferPtr = std::shared_ptr<Buffer>;

src/audio_core/codec.cpp

@@ -0,0 +1,77 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <algorithm>
+
+#include "audio_core/codec.h"
+
+namespace AudioCore::Codec {
+
+std::vector<s16> DecodeADPCM(const u8* const data, size_t size, const ADPCM_Coeff& coeff,
+                             ADPCMState& state) {
+    // GC-ADPCM with scale factor and variable coefficients.
+    // Frames are 8 bytes long containing 14 samples each.
+    // Samples are 4 bits (one nibble) long.
+
+    constexpr size_t FRAME_LEN = 8;
+    constexpr size_t SAMPLES_PER_FRAME = 14;
+    constexpr std::array<int, 16> SIGNED_NIBBLES = {
+        {0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1}};
+
+    const size_t sample_count = (size / FRAME_LEN) * SAMPLES_PER_FRAME;
+    const size_t ret_size =
+        sample_count % 2 == 0 ? sample_count : sample_count + 1; // Ensure multiple of two.
+    std::vector<s16> ret(ret_size);
+
+    int yn1 = state.yn1, yn2 = state.yn2;
+
+    const size_t NUM_FRAMES =
+        (sample_count + (SAMPLES_PER_FRAME - 1)) / SAMPLES_PER_FRAME; // Round up.
+    for (size_t framei = 0; framei < NUM_FRAMES; framei++) {
+        const int frame_header = data[framei * FRAME_LEN];
+        const int scale = 1 << (frame_header & 0xF);
+        const int idx = (frame_header >> 4) & 0x7;
+
+        // Coefficients are fixed point with 11 bits fractional part.
+        const int coef1 = coeff[idx * 2 + 0];
+        const int coef2 = coeff[idx * 2 + 1];
+
+        // Decodes an audio sample. One nibble produces one sample.
+        const auto decode_sample = [&](const int nibble) -> s16 {
+            const int xn = nibble * scale;
+            // We first transform everything into 11 bit fixed point, perform the second order
+            // digital filter, then transform back.
+            // 0x400 == 0.5 in 11 bit fixed point.
+            // Filter: y[n] = x[n] + 0.5 + c1 * y[n-1] + c2 * y[n-2]
+            int val = ((xn << 11) + 0x400 + coef1 * yn1 + coef2 * yn2) >> 11;
+            // Clamp to output range.
+            val = std::clamp<s32>(val, -32768, 32767);
+            // Advance output feedback.
+            yn2 = yn1;
+            yn1 = val;
+            return static_cast<s16>(val);
+        };
+
+        size_t outputi = framei * SAMPLES_PER_FRAME;
+        size_t datai = framei * FRAME_LEN + 1;
+        for (size_t i = 0; i < SAMPLES_PER_FRAME && outputi < sample_count; i += 2) {
+            const s16 sample1 = decode_sample(SIGNED_NIBBLES[data[datai] >> 4]);
+            ret[outputi] = sample1;
+            outputi++;
+
+            const s16 sample2 = decode_sample(SIGNED_NIBBLES[data[datai] & 0xF]);
+            ret[outputi] = sample2;
+            outputi++;
+
+            datai++;
+        }
+    }
+
+    state.yn1 = yn1;
+    state.yn2 = yn2;
+
+    return ret;
+}
+
+} // namespace AudioCore::Codec

src/audio_core/codec.h

@@ -0,0 +1,44 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <vector>
+
+#include "common/common_types.h"
+
+namespace AudioCore::Codec {
+
+enum class PcmFormat : u32 {
+    Invalid = 0,
+    Int8 = 1,
+    Int16 = 2,
+    Int24 = 3,
+    Int32 = 4,
+    PcmFloat = 5,
+    Adpcm = 6,
+};
+
+/// See: Codec::DecodeADPCM
+struct ADPCMState {
+    // Two historical samples from previous processed buffer,
+    // required for ADPCM decoding
+    s16 yn1; ///< y[n-1]
+    s16 yn2; ///< y[n-2]
+};
+
+using ADPCM_Coeff = std::array<s16, 16>;
+
+/**
+ * @param data Pointer to buffer that contains ADPCM data to decode
+ * @param size Size of buffer in bytes
+ * @param coeff ADPCM coefficients
+ * @param state ADPCM state, this is updated with new state
+ * @return Decoded stereo signed PCM16 data, sample_count in length
+ */
+std::vector<s16> DecodeADPCM(const u8* const data, size_t size, const ADPCM_Coeff& coeff,
+                             ADPCMState& state);
+
+}; // namespace AudioCore::Codec

src/audio_core/cubeb_sink.cpp

@@ -13,20 +13,30 @@ namespace AudioCore {
 
 class SinkStreamImpl final : public SinkStream {
 public:
-    SinkStreamImpl(cubeb* ctx, cubeb_devid output_device) : ctx{ctx} {
-        cubeb_stream_params params;
-        params.rate = 48000;
-        params.channels = GetNumChannels();
-        params.format = CUBEB_SAMPLE_S16NE;
-        params.layout = CUBEB_LAYOUT_STEREO;
+    SinkStreamImpl(cubeb* ctx, u32 sample_rate, u32 num_channels_, cubeb_devid output_device,
+                   const std::string& name)
+        : ctx{ctx}, num_channels{num_channels_} {
 
-        u32 minimum_latency = 0;
+        if (num_channels == 6) {
+            // 6-channel audio does not seem to work with cubeb + SDL, so we downsample this to 2
+            // channel for now
+            is_6_channel = true;
+            num_channels = 2;
+        }
+
+        cubeb_stream_params params{};
+        params.rate = sample_rate;
+        params.channels = num_channels;
+        params.format = CUBEB_SAMPLE_S16NE;
+        params.layout = num_channels == 1 ? CUBEB_LAYOUT_MONO : CUBEB_LAYOUT_STEREO;
+
+        u32 minimum_latency{};
         if (cubeb_get_min_latency(ctx, &params, &minimum_latency) != CUBEB_OK) {
             LOG_CRITICAL(Audio_Sink, "Error getting minimum latency");
         }
 
-        if (cubeb_stream_init(ctx, &stream_backend, "yuzu Audio Output", nullptr, nullptr,
-                              output_device, &params, std::max(512u, minimum_latency),
+        if (cubeb_stream_init(ctx, &stream_backend, name.c_str(), nullptr, nullptr, output_device,
+                              &params, std::max(512u, minimum_latency),
                               &SinkStreamImpl::DataCallback, &SinkStreamImpl::StateCallback,
                               this) != CUBEB_OK) {
             LOG_CRITICAL(Audio_Sink, "Error initializing cubeb stream");
@@ -51,33 +61,29 @@ public:
         cubeb_stream_destroy(stream_backend);
     }
 
-    void EnqueueSamples(u32 num_channels, const s16* samples, size_t sample_count) override {
+    void EnqueueSamples(u32 num_channels, const std::vector<s16>& samples) override {
         if (!ctx) {
             return;
         }
 
-        queue.reserve(queue.size() + sample_count * GetNumChannels());
+        queue.reserve(queue.size() + samples.size() * GetNumChannels());
 
-        if (num_channels == 2) {
-            // Copy as-is
-            std::copy(samples, samples + sample_count * GetNumChannels(),
-                      std::back_inserter(queue));
-        } else if (num_channels == 6) {
+        if (is_6_channel) {
             // Downsample 6 channels to 2
-            const size_t sample_count_copy_size = sample_count * num_channels * 2;
+            const size_t sample_count_copy_size = samples.size() * 2;
             queue.reserve(sample_count_copy_size);
-            for (size_t i = 0; i < sample_count * num_channels; i += num_channels) {
+            for (size_t i = 0; i < samples.size(); i += num_channels) {
                 queue.push_back(samples[i]);
                 queue.push_back(samples[i + 1]);
             }
         } else {
-            ASSERT_MSG(false, "Unimplemented");
+            // Copy as-is
+            std::copy(samples.begin(), samples.end(), std::back_inserter(queue));
         }
     }
 
     u32 GetNumChannels() const {
-        // Only support 2-channel stereo output for now
-        return 2;
+        return num_channels;
     }
 
 private:
@@ -85,6 +91,8 @@ private:
 
     cubeb* ctx{};
     cubeb_stream* stream_backend{};
+    u32 num_channels{};
+    bool is_6_channel{};
 
     std::vector<s16> queue;
 
@@ -129,8 +137,10 @@ CubebSink::~CubebSink() {
     cubeb_destroy(ctx);
 }
 
-SinkStream& CubebSink::AcquireSinkStream(u32 sample_rate, u32 num_channels) {
-    sink_streams.push_back(std::make_unique<SinkStreamImpl>(ctx, output_device));
+SinkStream& CubebSink::AcquireSinkStream(u32 sample_rate, u32 num_channels,
+                                         const std::string& name) {
+    sink_streams.push_back(
+        std::make_unique<SinkStreamImpl>(ctx, sample_rate, num_channels, output_device, name));
     return *sink_streams.back();
 }
 

src/audio_core/cubeb_sink.h

@@ -18,7 +18,8 @@ public:
     explicit CubebSink(std::string device_id);
     ~CubebSink() override;
 
-    SinkStream& AcquireSinkStream(u32 sample_rate, u32 num_channels) override;
+    SinkStream& AcquireSinkStream(u32 sample_rate, u32 num_channels,
+                                  const std::string& name) override;
 
 private:
     cubeb* ctx{};

src/audio_core/null_sink.h

@@ -13,14 +13,14 @@ public:
     explicit NullSink(std::string){};
     ~NullSink() override = default;
 
-    SinkStream& AcquireSinkStream(u32 /*sample_rate*/, u32 /*num_channels*/) override {
+    SinkStream& AcquireSinkStream(u32 /*sample_rate*/, u32 /*num_channels*/,
+                                  const std::string& /*name*/) override {
         return null_sink_stream;
     }
 
 private:
     struct NullSinkStreamImpl final : SinkStream {
-        void EnqueueSamples(u32 /*num_channels*/, const s16* /*samples*/,
-                            size_t /*sample_count*/) override {}
+        void EnqueueSamples(u32 /*num_channels*/, const std::vector<s16>& /*samples*/) override {}
     } null_sink_stream;
 };
 

src/audio_core/sink.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <string>
 
 #include "audio_core/sink_stream.h"
 #include "common/common_types.h"
@@ -21,7 +22,8 @@ constexpr char auto_device_name[] = "auto";
 class Sink {
 public:
     virtual ~Sink() = default;
-    virtual SinkStream& AcquireSinkStream(u32 sample_rate, u32 num_channels) = 0;
+    virtual SinkStream& AcquireSinkStream(u32 sample_rate, u32 num_channels,
+                                          const std::string& name) = 0;
 };
 
 using SinkPtr = std::unique_ptr<Sink>;

src/audio_core/sink_details.h

@@ -6,6 +6,7 @@
 
 #include <functional>
 #include <memory>
+#include <utility>
 #include <vector>
 
 namespace AudioCore {
@@ -13,16 +14,18 @@ namespace AudioCore {
 class Sink;
 
 struct SinkDetails {
-    SinkDetails(const char* id_, std::function<std::unique_ptr<Sink>(std::string)> factory_,
-                std::function<std::vector<std::string>()> list_devices_)
-        : id(id_), factory(factory_), list_devices(list_devices_) {}
+    using FactoryFn = std::function<std::unique_ptr<Sink>(std::string)>;
+    using ListDevicesFn = std::function<std::vector<std::string>()>;
+
+    SinkDetails(const char* id_, FactoryFn factory_, ListDevicesFn list_devices_)
+        : id(id_), factory(std::move(factory_)), list_devices(std::move(list_devices_)) {}
 
     /// Name for this sink.
     const char* id;
     /// A method to call to construct an instance of this type of sink.
-    std::function<std::unique_ptr<Sink>(std::string device_id)> factory;
+    FactoryFn factory;
     /// A method to call to list available devices.
-    std::function<std::vector<std::string>()> list_devices;
+    ListDevicesFn list_devices;
 };
 
 extern const std::vector<SinkDetails> g_sink_details;

src/audio_core/sink_stream.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <vector>
 
 #include "common/common_types.h"
 
@@ -22,9 +23,8 @@ public:
      * Feed stereo samples to sink.
      * @param num_channels Number of channels used.
      * @param samples Samples in interleaved stereo PCM16 format.
-     * @param sample_count Number of samples.
      */
-    virtual void EnqueueSamples(u32 num_channels, const s16* samples, size_t sample_count) = 0;
+    virtual void EnqueueSamples(u32 num_channels, const std::vector<s16>& samples) = 0;
 };
 
 using SinkStreamPtr = std::unique_ptr<SinkStream>;

src/audio_core/stream.cpp

@@ -32,17 +32,13 @@ u32 Stream::GetNumChannels() const {
     return {};
 }
 
-u32 Stream::GetSampleSize() const {
-    return GetNumChannels() * 2;
-}
-
 Stream::Stream(u32 sample_rate, Format format, ReleaseCallback&& release_callback,
-               SinkStream& sink_stream)
+               SinkStream& sink_stream, std::string&& name_)
     : sample_rate{sample_rate}, format{format}, release_callback{std::move(release_callback)},
-      sink_stream{sink_stream} {
+      sink_stream{sink_stream}, name{std::move(name_)} {
 
     release_event = CoreTiming::RegisterEvent(
-        "Stream::Release", [this](u64 userdata, int cycles_late) { ReleaseActiveBuffer(); });
+        name, [this](u64 userdata, int cycles_late) { ReleaseActiveBuffer(); });
 }
 
 void Stream::Play() {
@@ -55,17 +51,15 @@ void Stream::Stop() {
 }
 
 s64 Stream::GetBufferReleaseCycles(const Buffer& buffer) const {
-    const size_t num_samples{buffer.GetData().size() / GetSampleSize()};
+    const size_t num_samples{buffer.GetSamples().size() / GetNumChannels()};
     return CoreTiming::usToCycles((static_cast<u64>(num_samples) * 1000000) / sample_rate);
 }
 
-static std::vector<s16> GetVolumeAdjustedSamples(const std::vector<u8>& data) {
-    std::vector<s16> samples(data.size() / sizeof(s16));
-    std::memcpy(samples.data(), data.data(), data.size());
+static void VolumeAdjustSamples(std::vector<s16>& samples) {
     const float volume{std::clamp(Settings::values.volume, 0.0f, 1.0f)};
 
     if (volume == 1.0f) {
-        return samples;
+        return;
     }
 
     // Implementation of a volume slider with a dynamic range of 60 dB
@@ -73,8 +67,6 @@ static std::vector<s16> GetVolumeAdjustedSamples(const std::vector<u8>& data) {
     for (auto& sample : samples) {
         sample = static_cast<s16>(sample * volume_scale_factor);
     }
-
-    return samples;
 }
 
 void Stream::PlayNextBuffer() {
@@ -96,14 +88,14 @@ void Stream::PlayNextBuffer() {
     active_buffer = queued_buffers.front();
     queued_buffers.pop();
 
-    const size_t sample_count{active_buffer->GetData().size() / GetSampleSize()};
-    sink_stream.EnqueueSamples(
-        GetNumChannels(), GetVolumeAdjustedSamples(active_buffer->GetData()).data(), sample_count);
+    VolumeAdjustSamples(active_buffer->Samples());
+    sink_stream.EnqueueSamples(GetNumChannels(), active_buffer->GetSamples());
 
     CoreTiming::ScheduleEventThreadsafe(GetBufferReleaseCycles(*active_buffer), release_event, {});
 }
 
 void Stream::ReleaseActiveBuffer() {
+    ASSERT(active_buffer);
     released_buffers.push(std::move(active_buffer));
     release_callback();
     PlayNextBuffer();

src/audio_core/stream.h

@@ -6,6 +6,7 @@
 
 #include <functional>
 #include <memory>
+#include <string>
 #include <vector>
 #include <queue>
 
@@ -33,7 +34,7 @@ public:
     using ReleaseCallback = std::function<void()>;
 
     Stream(u32 sample_rate, Format format, ReleaseCallback&& release_callback,
-           SinkStream& sink_stream);
+           SinkStream& sink_stream, std::string&& name_);
 
     /// Plays the audio stream
     void Play();
@@ -68,9 +69,6 @@ public:
     /// Gets the number of channels
     u32 GetNumChannels() const;
 
-    /// Gets the sample size in bytes
-    u32 GetSampleSize() const;
-
 private:
     /// Current state of the stream
     enum class State {
@@ -96,6 +94,7 @@ private:
     std::queue<BufferPtr> queued_buffers;   ///< Buffers queued to be played in the stream
     std::queue<BufferPtr> released_buffers; ///< Buffers recently released from the stream
     SinkStream& sink_stream;                ///< Output sink for the stream
+    std::string name;                       ///< Name of the stream, must be unique
 };
 
 using StreamPtr = std::shared_ptr<Stream>;

src/common/common_paths.h

@@ -32,6 +32,7 @@
 #define SDMC_DIR "sdmc"
 #define NAND_DIR "nand"
 #define SYSDATA_DIR "sysdata"
+#define KEYS_DIR "keys"
 #define LOG_DIR "log"
 
 // Filenames

src/common/file_util.cpp

@@ -706,6 +706,7 @@ const std::string& GetUserPath(UserPath path, const std::string& new_path) {
         paths.emplace(UserPath::SDMCDir, user_path + SDMC_DIR DIR_SEP);
         paths.emplace(UserPath::NANDDir, user_path + NAND_DIR DIR_SEP);
         paths.emplace(UserPath::SysDataDir, user_path + SYSDATA_DIR DIR_SEP);
+        paths.emplace(UserPath::KeysDir, user_path + KEYS_DIR DIR_SEP);
         // TODO: Put the logs in a better location for each OS
         paths.emplace(UserPath::LogDir, user_path + LOG_DIR DIR_SEP);
     }
@@ -736,6 +737,19 @@ const std::string& GetUserPath(UserPath path, const std::string& new_path) {
     return paths[path];
 }
 
+std::string GetHactoolConfigurationPath() {
+#ifdef _WIN32
+    PWSTR pw_local_path = nullptr;
+    if (SHGetKnownFolderPath(FOLDERID_Profile, 0, nullptr, &pw_local_path) != S_OK)
+        return "";
+    std::string local_path = Common::UTF16ToUTF8(pw_local_path);
+    CoTaskMemFree(pw_local_path);
+    return local_path + "\\.switch";
+#else
+    return GetHomeDirectory() + "/.switch";
+#endif
+}
+
 size_t WriteStringToFile(bool text_file, const std::string& str, const char* filename) {
     return FileUtil::IOFile(filename, text_file ? "w" : "wb").WriteBytes(str.data(), str.size());
 }

src/common/file_util.h

@@ -23,6 +23,7 @@ namespace FileUtil {
 enum class UserPath {
     CacheDir,
     ConfigDir,
+    KeysDir,
     LogDir,
     NANDDir,
     RootDir,
@@ -125,6 +126,8 @@ bool SetCurrentDir(const std::string& directory);
 // directory. To be used in "multi-user" mode (that is, installed).
 const std::string& GetUserPath(UserPath path, const std::string& new_path = "");
 
+std::string GetHactoolConfigurationPath();
+
 // Returns the path to where the sys file are
 std::string GetSysDirectory();
 

src/common/logging/backend.cpp

@@ -168,8 +168,11 @@ void FileBackend::Write(const Entry& entry) {
     SUB(Service, AM)                                                                               \
     SUB(Service, AOC)                                                                              \
     SUB(Service, APM)                                                                              \
+    SUB(Service, ARP)                                                                              \
     SUB(Service, BCAT)                                                                             \
+    SUB(Service, BPC)                                                                              \
     SUB(Service, BTM)                                                                              \
+    SUB(Service, Capture)                                                                          \
     SUB(Service, Fatal)                                                                            \
     SUB(Service, FGM)                                                                              \
     SUB(Service, Friend)                                                                           \
@@ -178,6 +181,7 @@ void FileBackend::Write(const Entry& entry) {
     SUB(Service, LBL)                                                                              \
     SUB(Service, LDN)                                                                              \
     SUB(Service, LM)                                                                               \
+    SUB(Service, Migration)                                                                        \
     SUB(Service, Mii)                                                                              \
     SUB(Service, MM)                                                                               \
     SUB(Service, NCM)                                                                              \
@@ -188,7 +192,9 @@ void FileBackend::Write(const Entry& entry) {
     SUB(Service, NVDRV)                                                                            \
     SUB(Service, PCIE)                                                                             \
     SUB(Service, PCTL)                                                                             \
+    SUB(Service, PCV)                                                                              \
     SUB(Service, PREPO)                                                                            \
+    SUB(Service, PSC)                                                                              \
     SUB(Service, SET)                                                                              \
     SUB(Service, SM)                                                                               \
     SUB(Service, SPL)                                                                              \
@@ -212,6 +218,7 @@ void FileBackend::Write(const Entry& entry) {
     CLS(Input)                                                                                     \
     CLS(Network)                                                                                   \
     CLS(Loader)                                                                                    \
+    CLS(Crypto)                                                                                    \
     CLS(WebService)
 
 // GetClassName is a macro defined by Windows.h, grrr...

src/common/logging/log.h

@@ -54,9 +54,12 @@ enum class Class : ClassType {
     Service_AM,        ///< The AM (Applet manager) service
     Service_AOC,       ///< The AOC (AddOn Content) service
     Service_APM,       ///< The APM (Performance) service
+    Service_ARP,       ///< The ARP service
     Service_Audio,     ///< The Audio (Audio control) service
     Service_BCAT,      ///< The BCAT service
+    Service_BPC,       ///< The BPC service
     Service_BTM,       ///< The BTM service
+    Service_Capture,   ///< The capture service
     Service_Fatal,     ///< The Fatal service
     Service_FGM,       ///< The FGM service
     Service_Friend,    ///< The friend service
@@ -65,6 +68,7 @@ enum class Class : ClassType {
     Service_LBL,       ///< The LBL (LCD backlight) service
     Service_LDN,       ///< The LDN (Local domain network) service
     Service_LM,        ///< The LM (Logger) service
+    Service_Migration, ///< The migration service
     Service_Mii,       ///< The Mii service
     Service_MM,        ///< The MM (Multimedia) service
     Service_NCM,       ///< The NCM service
@@ -75,7 +79,9 @@ enum class Class : ClassType {
     Service_NVDRV,     ///< The NVDRV (Nvidia driver) service
     Service_PCIE,      ///< The PCIe service
     Service_PCTL,      ///< The PCTL (Parental control) service
+    Service_PCV,       ///< The PCV service
     Service_PREPO,     ///< The PREPO (Play report) service
+    Service_PSC,       ///< The PSC service
     Service_SET,       ///< The SET (Settings) service
     Service_SM,        ///< The SM (Service manager) service
     Service_SPL,       ///< The SPL service
@@ -97,6 +103,7 @@ enum class Class : ClassType {
     Audio_DSP,         ///< The HLE implementation of the DSP
     Audio_Sink,        ///< Emulator audio output backend
     Loader,            ///< ROM loader
+    Crypto,            ///< Cryptographic engine/functions
     Input,             ///< Input emulation
     Network,           ///< Network emulation
     WebService,        ///< Interface to yuzu Web Services

src/common/math_util.h

@@ -19,12 +19,12 @@ inline bool IntervalsIntersect(unsigned start0, unsigned length0, unsigned start
 
 template <class T>
 struct Rectangle {
-    T left;
-    T top;
-    T right;
-    T bottom;
+    T left{};
+    T top{};
+    T right{};
+    T bottom{};
 
-    Rectangle() {}
+    Rectangle() = default;
 
     Rectangle(T left, T top, T right, T bottom)
         : left(left), top(top), right(right), bottom(bottom) {}

src/core/CMakeLists.txt

@@ -12,6 +12,16 @@ add_library(core STATIC
     core_timing.h
     core_timing_util.cpp
     core_timing_util.h
+    crypto/aes_util.cpp
+    crypto/aes_util.h
+    crypto/encryption_layer.cpp
+    crypto/encryption_layer.h
+    crypto/key_manager.cpp
+    crypto/key_manager.h
+    crypto/ctr_encryption_layer.cpp
+    crypto/ctr_encryption_layer.h
+    file_sys/card_image.cpp
+    file_sys/card_image.h
     file_sys/content_archive.cpp
     file_sys/content_archive.h
     file_sys/control_metadata.cpp
@@ -63,10 +73,10 @@ add_library(core STATIC
     hle/kernel/hle_ipc.h
     hle/kernel/kernel.cpp
     hle/kernel/kernel.h
-    hle/kernel/memory.cpp
-    hle/kernel/memory.h
     hle/kernel/mutex.cpp
     hle/kernel/mutex.h
+    hle/kernel/object.cpp
+    hle/kernel/object.h
     hle/kernel/process.cpp
     hle/kernel/process.h
     hle/kernel/resource_limit.cpp
@@ -124,6 +134,8 @@ add_library(core STATIC
     hle/service/apm/apm.h
     hle/service/apm/interface.cpp
     hle/service/apm/interface.h
+    hle/service/arp/arp.cpp
+    hle/service/arp/arp.h
     hle/service/audio/audctl.cpp
     hle/service/audio/audctl.h
     hle/service/audio/auddbg.cpp
@@ -154,10 +166,14 @@ add_library(core STATIC
     hle/service/bcat/bcat.h
     hle/service/bcat/module.cpp
     hle/service/bcat/module.h
+    hle/service/bpc/bpc.cpp
+    hle/service/bpc/bpc.h
     hle/service/btdrv/btdrv.cpp
     hle/service/btdrv/btdrv.h
     hle/service/btm/btm.cpp
     hle/service/btm/btm.h
+    hle/service/caps/caps.cpp
+    hle/service/caps/caps.h
     hle/service/erpt/erpt.cpp
     hle/service/erpt/erpt.h
     hle/service/es/es.cpp
@@ -200,6 +216,8 @@ add_library(core STATIC
     hle/service/ldr/ldr.h
     hle/service/lm/lm.cpp
     hle/service/lm/lm.h
+    hle/service/mig/mig.cpp
+    hle/service/mig/mig.h
     hle/service/mii/mii.cpp
     hle/service/mii/mii.h
     hle/service/mm/mm_u.cpp
@@ -251,10 +269,14 @@ add_library(core STATIC
     hle/service/pctl/module.h
     hle/service/pctl/pctl.cpp
     hle/service/pctl/pctl.h
+    hle/service/pcv/pcv.cpp
+    hle/service/pcv/pcv.h
     hle/service/pm/pm.cpp
     hle/service/pm/pm.h
     hle/service/prepo/prepo.cpp
     hle/service/prepo/prepo.h
+    hle/service/psc/psc.cpp
+    hle/service/psc/psc.h
     hle/service/service.cpp
     hle/service/service.h
     hle/service/set/set.cpp
@@ -303,10 +325,6 @@ add_library(core STATIC
     hle/service/vi/vi_u.h
     hle/service/wlan/wlan.cpp
     hle/service/wlan/wlan.h
-    hw/hw.cpp
-    hw/hw.h
-    hw/lcd.cpp
-    hw/lcd.h
     loader/deconstructed_rom_directory.cpp
     loader/deconstructed_rom_directory.h
     loader/elf.cpp
@@ -321,6 +339,8 @@ add_library(core STATIC
     loader/nro.h
     loader/nso.cpp
     loader/nso.h
+    loader/xci.cpp
+    loader/xci.h
     memory.cpp
     memory.h
     memory_hook.cpp
@@ -340,7 +360,7 @@ add_library(core STATIC
 create_target_directory_groups(core)
 
 target_link_libraries(core PUBLIC common PRIVATE audio_core video_core)
-target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt lz4_static opus unicorn)
+target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt lz4_static mbedtls opus unicorn)
 
 if (ARCHITECTURE_x86_64)
     target_sources(core PRIVATE

src/core/arm/unicorn/arm_unicorn.cpp

@@ -203,7 +203,7 @@ void ARM_Unicorn::ExecuteInstructions(int num_instructions) {
         }
         Kernel::Thread* thread = Kernel::GetCurrentThread();
         SaveContext(thread->context);
-        if (last_bkpt_hit || (num_instructions == 1)) {
+        if (last_bkpt_hit || GDBStub::GetCpuStepFlag()) {
             last_bkpt_hit = false;
             GDBStub::Break();
             GDBStub::SendTrap(thread, 5);

src/core/core.cpp

@@ -15,11 +15,10 @@
 #include "core/hle/service/service.h"
 #include "core/hle/service/sm/controller.h"
 #include "core/hle/service/sm/sm.h"
-#include "core/hw/hw.h"
 #include "core/loader/loader.h"
-#include "core/memory_setup.h"
 #include "core/settings.h"
 #include "file_sys/vfs_real.h"
+#include "video_core/renderer_base.h"
 #include "video_core/video_core.h"
 
 namespace Core {
@@ -63,7 +62,6 @@ System::ResultStatus System::RunLoop(bool tight_loop) {
         // execute. Otherwise, get out of the loop function.
         if (GDBStub::GetCpuHaltFlag()) {
             if (GDBStub::GetCpuStepFlag()) {
-                GDBStub::SetCpuStepFlag(false);
                 tight_loop = false;
             } else {
                 return ResultStatus::Success;
@@ -79,6 +77,10 @@ System::ResultStatus System::RunLoop(bool tight_loop) {
         }
     }
 
+    if (GDBStub::IsServerEnabled()) {
+        GDBStub::SetCpuStepFlag(false);
+    }
+
     return status;
 }
 
@@ -86,7 +88,7 @@ System::ResultStatus System::SingleStep() {
     return RunLoop(false);
 }
 
-System::ResultStatus System::Load(EmuWindow* emu_window, const std::string& filepath) {
+System::ResultStatus System::Load(EmuWindow& emu_window, const std::string& filepath) {
     app_loader = Loader::GetLoader(std::make_shared<FileSys::RealVfsFile>(filepath));
 
     if (!app_loader) {
@@ -101,8 +103,10 @@ System::ResultStatus System::Load(EmuWindow* emu_window, const std::string& file
                      static_cast<int>(system_mode.second));
 
         switch (system_mode.second) {
-        case Loader::ResultStatus::ErrorEncrypted:
-            return ResultStatus::ErrorLoader_ErrorEncrypted;
+        case Loader::ResultStatus::ErrorMissingKeys:
+            return ResultStatus::ErrorLoader_ErrorMissingKeys;
+        case Loader::ResultStatus::ErrorDecrypting:
+            return ResultStatus::ErrorLoader_ErrorDecrypting;
         case Loader::ResultStatus::ErrorInvalidFormat:
             return ResultStatus::ErrorLoader_ErrorInvalidFormat;
         case Loader::ResultStatus::ErrorUnsupportedArch:
@@ -112,7 +116,7 @@ System::ResultStatus System::Load(EmuWindow* emu_window, const std::string& file
         }
     }
 
-    ResultStatus init_result{Init(emu_window, system_mode.first.get())};
+    ResultStatus init_result{Init(emu_window)};
     if (init_result != ResultStatus::Success) {
         LOG_CRITICAL(Core, "Failed to initialize system (Error {})!",
                      static_cast<int>(init_result));
@@ -126,8 +130,10 @@ System::ResultStatus System::Load(EmuWindow* emu_window, const std::string& file
         System::Shutdown();
 
         switch (load_result) {
-        case Loader::ResultStatus::ErrorEncrypted:
-            return ResultStatus::ErrorLoader_ErrorEncrypted;
+        case Loader::ResultStatus::ErrorMissingKeys:
+            return ResultStatus::ErrorLoader_ErrorMissingKeys;
+        case Loader::ResultStatus::ErrorDecrypting:
+            return ResultStatus::ErrorLoader_ErrorDecrypting;
         case Loader::ResultStatus::ErrorInvalidFormat:
             return ResultStatus::ErrorLoader_ErrorInvalidFormat;
         case Loader::ResultStatus::ErrorUnsupportedArch:
@@ -163,7 +169,7 @@ Cpu& System::CpuCore(size_t core_index) {
     return *cpu_cores[core_index];
 }
 
-System::ResultStatus System::Init(EmuWindow* emu_window, u32 system_mode) {
+System::ResultStatus System::Init(EmuWindow& emu_window) {
     LOG_DEBUG(HW_Memory, "initialized OK");
 
     CoreTiming::Init();
@@ -176,19 +182,20 @@ System::ResultStatus System::Init(EmuWindow* emu_window, u32 system_mode) {
         cpu_cores[index] = std::make_shared<Cpu>(cpu_exclusive_monitor, cpu_barrier, index);
     }
 
-    gpu_core = std::make_unique<Tegra::GPU>();
     telemetry_session = std::make_unique<Core::TelemetrySession>();
     service_manager = std::make_shared<Service::SM::ServiceManager>();
 
-    HW::Init();
-    Kernel::Init(system_mode);
+    Kernel::Init();
     Service::Init(service_manager);
     GDBStub::Init();
 
-    if (!VideoCore::Init(emu_window)) {
+    renderer = VideoCore::CreateRenderer(emu_window);
+    if (!renderer->Init()) {
         return ResultStatus::ErrorVideoCore;
     }
 
+    gpu_core = std::make_unique<Tegra::GPU>(renderer->Rasterizer());
+
     // Create threads for CPU cores 1-3, and build thread_to_cpu map
     // CPU core 0 is run on the main thread
     thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
@@ -220,11 +227,10 @@ void System::Shutdown() {
                          perf_results.frametime * 1000.0);
 
     // Shutdown emulation session
-    VideoCore::Shutdown();
+    renderer.reset();
     GDBStub::Shutdown();
     Service::Shutdown();
     Kernel::Shutdown();
-    HW::Shutdown();
     service_manager.reset();
     telemetry_session.reset();
     gpu_core.reset();

src/core/core.h

@@ -11,7 +11,7 @@
 #include "common/common_types.h"
 #include "core/arm/exclusive_monitor.h"
 #include "core/core_cpu.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/scheduler.h"
 #include "core/loader/loader.h"
 #include "core/memory.h"
@@ -27,6 +27,10 @@ namespace Service::SM {
 class ServiceManager;
 }
 
+namespace VideoCore {
+class RendererBase;
+}
+
 namespace Core {
 
 class System {
@@ -43,12 +47,14 @@ public:
 
     /// Enumeration representing the return values of the System Initialize and Load process.
     enum class ResultStatus : u32 {
-        Success,                    ///< Succeeded
-        ErrorNotInitialized,        ///< Error trying to use core prior to initialization
-        ErrorGetLoader,             ///< Error finding the correct application loader
-        ErrorSystemMode,            ///< Error determining the system mode
-        ErrorLoader,                ///< Error loading the specified application
-        ErrorLoader_ErrorEncrypted, ///< Error loading the specified application due to encryption
+        Success,                      ///< Succeeded
+        ErrorNotInitialized,          ///< Error trying to use core prior to initialization
+        ErrorGetLoader,               ///< Error finding the correct application loader
+        ErrorSystemMode,              ///< Error determining the system mode
+        ErrorLoader,                  ///< Error loading the specified application
+        ErrorLoader_ErrorMissingKeys, ///< Error because the key/keys needed to run could not be
+                                      ///< found.
+        ErrorLoader_ErrorDecrypting,  ///< Error loading the specified application due to encryption
         ErrorLoader_ErrorInvalidFormat, ///< Error loading the specified application due to an
                                         /// invalid format
         ErrorSystemFiles,               ///< Error in finding system files
@@ -76,16 +82,28 @@ public:
      */
     ResultStatus SingleStep();
 
+    /**
+     * Invalidate the CPU instruction caches
+     * This function should only be used by GDB Stub to support breakpoints, memory updates and
+     * step/continue commands.
+     */
+    void InvalidateCpuInstructionCaches() {
+        for (auto& cpu : cpu_cores) {
+            cpu->ArmInterface().ClearInstructionCache();
+        }
+    }
+
     /// Shutdown the emulated system.
     void Shutdown();
 
     /**
      * Load an executable application.
-     * @param emu_window Pointer to the host-system window used for video output and keyboard input.
+     * @param emu_window Reference to the host-system window used for video output and keyboard
+     *                   input.
      * @param filepath String path to the executable application to load on the host file system.
      * @returns ResultStatus code, indicating if the operation succeeded.
      */
-    ResultStatus Load(EmuWindow* emu_window, const std::string& filepath);
+    ResultStatus Load(EmuWindow& emu_window, const std::string& filepath);
 
     /**
      * Indicates if the emulated system is powered on (all subsystems initialized and able to run an
@@ -126,11 +144,26 @@ public:
     /// Gets a CPU interface to the CPU core with the specified index
     Cpu& CpuCore(size_t core_index);
 
-    /// Gets the GPU interface
+    /// Gets a mutable reference to the GPU interface
     Tegra::GPU& GPU() {
         return *gpu_core;
     }
 
+    /// Gets an immutable reference to the GPU interface.
+    const Tegra::GPU& GPU() const {
+        return *gpu_core;
+    }
+
+    /// Gets a mutable reference to the renderer.
+    VideoCore::RendererBase& Renderer() {
+        return *renderer;
+    }
+
+    /// Gets an immutable reference to the renderer.
+    const VideoCore::RendererBase& Renderer() const {
+        return *renderer;
+    }
+
     /// Gets the scheduler for the CPU core that is currently running
     Kernel::Scheduler& CurrentScheduler() {
         return *CurrentCpuCore().Scheduler();
@@ -186,14 +219,15 @@ private:
 
     /**
      * Initialize the emulated system.
-     * @param emu_window Pointer to the host-system window used for video output and keyboard input.
-     * @param system_mode The system mode.
+     * @param emu_window Reference to the host-system window used for video output and keyboard
+     *                   input.
      * @return ResultStatus code, indicating if the operation succeeded.
      */
-    ResultStatus Init(EmuWindow* emu_window, u32 system_mode);
+    ResultStatus Init(EmuWindow& emu_window);
 
     /// AppLoader used to load the current executing application
     std::unique_ptr<Loader::AppLoader> app_loader;
+    std::unique_ptr<VideoCore::RendererBase> renderer;
     std::unique_ptr<Tegra::GPU> gpu_core;
     std::shared_ptr<Tegra::DebugContext> debug_context;
     Kernel::SharedPtr<Kernel::Process> current_process;

src/core/core_cpu.cpp

@@ -12,7 +12,6 @@
 #include "core/arm/unicorn/arm_unicorn.h"
 #include "core/core_cpu.h"
 #include "core/core_timing.h"
-#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/scheduler.h"
 #include "core/hle/kernel/thread.h"
 #include "core/settings.h"

src/core/core_timing.cpp

@@ -141,7 +141,7 @@ void ScheduleEvent(s64 cycles_into_future, const EventType* event_type, u64 user
         ForceExceptionCheck(cycles_into_future);
 
     event_queue.emplace_back(Event{timeout, event_fifo_id++, userdata, event_type});
-    std::push_heap(event_queue.begin(), event_queue.end(), std::greater<Event>());
+    std::push_heap(event_queue.begin(), event_queue.end(), std::greater<>());
 }
 
 void ScheduleEventThreadsafe(s64 cycles_into_future, const EventType* event_type, u64 userdata) {
@@ -156,7 +156,7 @@ void UnscheduleEvent(const EventType* event_type, u64 userdata) {
     // Removing random items breaks the invariant so we have to re-establish it.
     if (itr != event_queue.end()) {
         event_queue.erase(itr, event_queue.end());
-        std::make_heap(event_queue.begin(), event_queue.end(), std::greater<Event>());
+        std::make_heap(event_queue.begin(), event_queue.end(), std::greater<>());
     }
 }
 
@@ -167,7 +167,7 @@ void RemoveEvent(const EventType* event_type) {
     // Removing random items breaks the invariant so we have to re-establish it.
     if (itr != event_queue.end()) {
         event_queue.erase(itr, event_queue.end());
-        std::make_heap(event_queue.begin(), event_queue.end(), std::greater<Event>());
+        std::make_heap(event_queue.begin(), event_queue.end(), std::greater<>());
     }
 }
 
@@ -190,7 +190,7 @@ void MoveEvents() {
     for (Event ev; ts_queue.Pop(ev);) {
         ev.fifo_order = event_fifo_id++;
         event_queue.emplace_back(std::move(ev));
-        std::push_heap(event_queue.begin(), event_queue.end(), std::greater<Event>());
+        std::push_heap(event_queue.begin(), event_queue.end(), std::greater<>());
     }
 }
 
@@ -205,7 +205,7 @@ void Advance() {
 
     while (!event_queue.empty() && event_queue.front().time <= global_timer) {
         Event evt = std::move(event_queue.front());
-        std::pop_heap(event_queue.begin(), event_queue.end(), std::greater<Event>());
+        std::pop_heap(event_queue.begin(), event_queue.end(), std::greater<>());
         event_queue.pop_back();
         evt.type->callback(evt.userdata, static_cast<int>(global_timer - evt.time));
     }
@@ -226,8 +226,8 @@ void Idle() {
     downcount = 0;
 }
 
-u64 GetGlobalTimeUs() {
-    return GetTicks() * 1000000 / BASE_CLOCK_RATE;
+std::chrono::microseconds GetGlobalTimeUs() {
+    return std::chrono::microseconds{GetTicks() * 1000000 / BASE_CLOCK_RATE};
 }
 
 int GetDowncount() {

src/core/core_timing.h

@@ -17,12 +17,17 @@
  *   ScheduleEvent(periodInCycles - cyclesLate, callback, "whatever")
  */
 
+#include <chrono>
 #include <functional>
 #include <string>
 #include "common/common_types.h"
 
 namespace CoreTiming {
 
+struct EventType;
+
+using TimedCallback = std::function<void(u64 userdata, int cycles_late)>;
+
 /**
  * CoreTiming begins at the boundary of timing slice -1. An initial call to Advance() is
  * required to end slice -1 and start slice 0 before the first cycle of code is executed.
@@ -30,8 +35,6 @@ namespace CoreTiming {
 void Init();
 void Shutdown();
 
-typedef std::function<void(u64 userdata, int cycles_late)> TimedCallback;
-
 /**
  * This should only be called from the emu thread, if you are calling it any other thread, you are
  * doing something evil
@@ -40,8 +43,6 @@ u64 GetTicks();
 u64 GetIdleTicks();
 void AddTicks(u64 ticks);
 
-struct EventType;
-
 /**
  * Returns the event_type identifier. if name is not unique, it will assert.
  */
@@ -86,7 +87,7 @@ void ClearPendingEvents();
 
 void ForceExceptionCheck(s64 cycles);
 
-u64 GetGlobalTimeUs();
+std::chrono::microseconds GetGlobalTimeUs();
 
 int GetDowncount();
 

src/core/crypto/aes_util.cpp

@@ -0,0 +1,115 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <mbedtls/cipher.h>
+#include "common/assert.h"
+#include "common/logging/log.h"
+#include "core/crypto/aes_util.h"
+#include "core/crypto/key_manager.h"
+
+namespace Core::Crypto {
+namespace {
+std::vector<u8> CalculateNintendoTweak(size_t sector_id) {
+    std::vector<u8> out(0x10);
+    for (size_t i = 0xF; i <= 0xF; --i) {
+        out[i] = sector_id & 0xFF;
+        sector_id >>= 8;
+    }
+    return out;
+}
+} // Anonymous namespace
+
+static_assert(static_cast<size_t>(Mode::CTR) == static_cast<size_t>(MBEDTLS_CIPHER_AES_128_CTR),
+              "CTR has incorrect value.");
+static_assert(static_cast<size_t>(Mode::ECB) == static_cast<size_t>(MBEDTLS_CIPHER_AES_128_ECB),
+              "ECB has incorrect value.");
+static_assert(static_cast<size_t>(Mode::XTS) == static_cast<size_t>(MBEDTLS_CIPHER_AES_128_XTS),
+              "XTS has incorrect value.");
+
+// Structure to hide mbedtls types from header file
+struct CipherContext {
+    mbedtls_cipher_context_t encryption_context;
+    mbedtls_cipher_context_t decryption_context;
+};
+
+template <typename Key, size_t KeySize>
+Crypto::AESCipher<Key, KeySize>::AESCipher(Key key, Mode mode)
+    : ctx(std::make_unique<CipherContext>()) {
+    mbedtls_cipher_init(&ctx->encryption_context);
+    mbedtls_cipher_init(&ctx->decryption_context);
+
+    ASSERT_MSG((mbedtls_cipher_setup(
+                    &ctx->encryption_context,
+                    mbedtls_cipher_info_from_type(static_cast<mbedtls_cipher_type_t>(mode))) ||
+                mbedtls_cipher_setup(
+                    &ctx->decryption_context,
+                    mbedtls_cipher_info_from_type(static_cast<mbedtls_cipher_type_t>(mode)))) == 0,
+               "Failed to initialize mbedtls ciphers.");
+
+    ASSERT(
+        !mbedtls_cipher_setkey(&ctx->encryption_context, key.data(), KeySize * 8, MBEDTLS_ENCRYPT));
+    ASSERT(
+        !mbedtls_cipher_setkey(&ctx->decryption_context, key.data(), KeySize * 8, MBEDTLS_DECRYPT));
+    //"Failed to set key on mbedtls ciphers.");
+}
+
+template <typename Key, size_t KeySize>
+AESCipher<Key, KeySize>::~AESCipher() {
+    mbedtls_cipher_free(&ctx->encryption_context);
+    mbedtls_cipher_free(&ctx->decryption_context);
+}
+
+template <typename Key, size_t KeySize>
+void AESCipher<Key, KeySize>::SetIV(std::vector<u8> iv) {
+    ASSERT_MSG((mbedtls_cipher_set_iv(&ctx->encryption_context, iv.data(), iv.size()) ||
+                mbedtls_cipher_set_iv(&ctx->decryption_context, iv.data(), iv.size())) == 0,
+               "Failed to set IV on mbedtls ciphers.");
+}
+
+template <typename Key, size_t KeySize>
+void AESCipher<Key, KeySize>::Transcode(const u8* src, size_t size, u8* dest, Op op) const {
+    auto* const context = op == Op::Encrypt ? &ctx->encryption_context : &ctx->decryption_context;
+
+    mbedtls_cipher_reset(context);
+
+    size_t written = 0;
+    if (mbedtls_cipher_get_cipher_mode(context) == MBEDTLS_MODE_XTS) {
+        mbedtls_cipher_update(context, src, size, dest, &written);
+        if (written != size) {
+            LOG_WARNING(Crypto, "Not all data was decrypted requested={:016X}, actual={:016X}.",
+                        size, written);
+        }
+    } else {
+        const auto block_size = mbedtls_cipher_get_block_size(context);
+
+        for (size_t offset = 0; offset < size; offset += block_size) {
+            auto length = std::min<size_t>(block_size, size - offset);
+            mbedtls_cipher_update(context, src + offset, length, dest + offset, &written);
+            if (written != length) {
+                LOG_WARNING(Crypto, "Not all data was decrypted requested={:016X}, actual={:016X}.",
+                            length, written);
+            }
+        }
+    }
+
+    mbedtls_cipher_finish(context, nullptr, nullptr);
+}
+
+template <typename Key, size_t KeySize>
+void AESCipher<Key, KeySize>::XTSTranscode(const u8* src, size_t size, u8* dest, size_t sector_id,
+                                           size_t sector_size, Op op) {
+    if (size % sector_size > 0) {
+        LOG_CRITICAL(Crypto, "Data size must be a multiple of sector size.");
+        return;
+    }
+
+    for (size_t i = 0; i < size; i += sector_size) {
+        SetIV(CalculateNintendoTweak(sector_id++));
+        Transcode<u8, u8>(src + i, sector_size, dest + i, op);
+    }
+}
+
+template class AESCipher<Key128>;
+template class AESCipher<Key256>;
+} // namespace Core::Crypto

src/core/crypto/aes_util.h

@@ -0,0 +1,64 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <memory>
+#include <type_traits>
+#include <vector>
+#include "common/common_types.h"
+#include "core/file_sys/vfs.h"
+
+namespace Core::Crypto {
+
+struct CipherContext;
+
+enum class Mode {
+    CTR = 11,
+    ECB = 2,
+    XTS = 70,
+};
+
+enum class Op {
+    Encrypt,
+    Decrypt,
+};
+
+template <typename Key, size_t KeySize = sizeof(Key)>
+class AESCipher {
+    static_assert(std::is_same_v<Key, std::array<u8, KeySize>>, "Key must be std::array of u8.");
+    static_assert(KeySize == 0x10 || KeySize == 0x20, "KeySize must be 128 or 256.");
+
+public:
+    AESCipher(Key key, Mode mode);
+
+    ~AESCipher();
+
+    void SetIV(std::vector<u8> iv);
+
+    template <typename Source, typename Dest>
+    void Transcode(const Source* src, size_t size, Dest* dest, Op op) const {
+        static_assert(std::is_trivially_copyable_v<Source> && std::is_trivially_copyable_v<Dest>,
+                      "Transcode source and destination types must be trivially copyable.");
+        Transcode(reinterpret_cast<const u8*>(src), size, reinterpret_cast<u8*>(dest), op);
+    }
+
+    void Transcode(const u8* src, size_t size, u8* dest, Op op) const;
+
+    template <typename Source, typename Dest>
+    void XTSTranscode(const Source* src, size_t size, Dest* dest, size_t sector_id,
+                      size_t sector_size, Op op) {
+        static_assert(std::is_trivially_copyable_v<Source> && std::is_trivially_copyable_v<Dest>,
+                      "XTSTranscode source and destination types must be trivially copyable.");
+        XTSTranscode(reinterpret_cast<const u8*>(src), size, reinterpret_cast<u8*>(dest), sector_id,
+                     sector_size, op);
+    }
+
+    void XTSTranscode(const u8* src, size_t size, u8* dest, size_t sector_id, size_t sector_size,
+                      Op op);
+
+private:
+    std::unique_ptr<CipherContext> ctx;
+};
+} // namespace Core::Crypto

src/core/crypto/ctr_encryption_layer.cpp

@@ -0,0 +1,56 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+#include "common/assert.h"
+#include "core/crypto/ctr_encryption_layer.h"
+
+namespace Core::Crypto {
+
+CTREncryptionLayer::CTREncryptionLayer(FileSys::VirtualFile base_, Key128 key_, size_t base_offset)
+    : EncryptionLayer(std::move(base_)), base_offset(base_offset), cipher(key_, Mode::CTR),
+      iv(16, 0) {}
+
+size_t CTREncryptionLayer::Read(u8* data, size_t length, size_t offset) const {
+    if (length == 0)
+        return 0;
+
+    const auto sector_offset = offset & 0xF;
+    if (sector_offset == 0) {
+        UpdateIV(base_offset + offset);
+        std::vector<u8> raw = base->ReadBytes(length, offset);
+        if (raw.size() != length)
+            return Read(data, raw.size(), offset);
+        cipher.Transcode(raw.data(), length, data, Op::Decrypt);
+        return length;
+    }
+
+    // offset does not fall on block boundary (0x10)
+    std::vector<u8> block = base->ReadBytes(0x10, offset - sector_offset);
+    UpdateIV(base_offset + offset - sector_offset);
+    cipher.Transcode(block.data(), block.size(), block.data(), Op::Decrypt);
+    size_t read = 0x10 - sector_offset;
+
+    if (length + sector_offset < 0x10) {
+        std::memcpy(data, block.data() + sector_offset, std::min<u64>(length, read));
+        return read;
+    }
+    std::memcpy(data, block.data() + sector_offset, read);
+    return read + Read(data + read, length - read, offset + read);
+}
+
+void CTREncryptionLayer::SetIV(const std::vector<u8>& iv_) {
+    const auto length = std::min(iv_.size(), iv.size());
+    iv.assign(iv_.cbegin(), iv_.cbegin() + length);
+}
+
+void CTREncryptionLayer::UpdateIV(size_t offset) const {
+    offset >>= 4;
+    for (size_t i = 0; i < 8; ++i) {
+        iv[16 - i - 1] = offset & 0xFF;
+        offset >>= 8;
+    }
+    cipher.SetIV(iv);
+}
+} // namespace Core::Crypto

src/core/crypto/ctr_encryption_layer.h

@@ -0,0 +1,33 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <vector>
+#include "core/crypto/aes_util.h"
+#include "core/crypto/encryption_layer.h"
+#include "core/crypto/key_manager.h"
+
+namespace Core::Crypto {
+
+// Sits on top of a VirtualFile and provides CTR-mode AES decription.
+class CTREncryptionLayer : public EncryptionLayer {
+public:
+    CTREncryptionLayer(FileSys::VirtualFile base, Key128 key, size_t base_offset);
+
+    size_t Read(u8* data, size_t length, size_t offset) const override;
+
+    void SetIV(const std::vector<u8>& iv);
+
+private:
+    size_t base_offset;
+
+    // Must be mutable as operations modify cipher contexts.
+    mutable AESCipher<Key128> cipher;
+    mutable std::vector<u8> iv;
+
+    void UpdateIV(size_t offset) const;
+};
+
+} // namespace Core::Crypto

src/core/crypto/encryption_layer.cpp

@@ -0,0 +1,42 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/crypto/encryption_layer.h"
+
+namespace Core::Crypto {
+
+EncryptionLayer::EncryptionLayer(FileSys::VirtualFile base_) : base(std::move(base_)) {}
+
+std::string EncryptionLayer::GetName() const {
+    return base->GetName();
+}
+
+size_t EncryptionLayer::GetSize() const {
+    return base->GetSize();
+}
+
+bool EncryptionLayer::Resize(size_t new_size) {
+    return false;
+}
+
+std::shared_ptr<FileSys::VfsDirectory> EncryptionLayer::GetContainingDirectory() const {
+    return base->GetContainingDirectory();
+}
+
+bool EncryptionLayer::IsWritable() const {
+    return false;
+}
+
+bool EncryptionLayer::IsReadable() const {
+    return true;
+}
+
+size_t EncryptionLayer::Write(const u8* data, size_t length, size_t offset) {
+    return 0;
+}
+
+bool EncryptionLayer::Rename(std::string_view name) {
+    return base->Rename(name);
+}
+} // namespace Core::Crypto

src/core/crypto/encryption_layer.h

@@ -0,0 +1,33 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+#include "core/file_sys/vfs.h"
+
+namespace Core::Crypto {
+
+// Basically non-functional class that implements all of the methods that are irrelevant to an
+// EncryptionLayer. Reduces duplicate code.
+class EncryptionLayer : public FileSys::VfsFile {
+public:
+    explicit EncryptionLayer(FileSys::VirtualFile base);
+
+    size_t Read(u8* data, size_t length, size_t offset) const override = 0;
+
+    std::string GetName() const override;
+    size_t GetSize() const override;
+    bool Resize(size_t new_size) override;
+    std::shared_ptr<FileSys::VfsDirectory> GetContainingDirectory() const override;
+    bool IsWritable() const override;
+    bool IsReadable() const override;
+    size_t Write(const u8* data, size_t length, size_t offset) override;
+    bool Rename(std::string_view name) override;
+
+protected:
+    FileSys::VirtualFile base;
+};
+
+} // namespace Core::Crypto

src/core/crypto/key_manager.cpp

@@ -0,0 +1,208 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <algorithm>
+#include <array>
+#include <fstream>
+#include <locale>
+#include <sstream>
+#include <string_view>
+#include "common/common_paths.h"
+#include "common/file_util.h"
+#include "core/crypto/key_manager.h"
+#include "core/settings.h"
+
+namespace Core::Crypto {
+
+static u8 ToHexNibble(char c1) {
+    if (c1 >= 65 && c1 <= 70)
+        return c1 - 55;
+    if (c1 >= 97 && c1 <= 102)
+        return c1 - 87;
+    if (c1 >= 48 && c1 <= 57)
+        return c1 - 48;
+    throw std::logic_error("Invalid hex digit");
+}
+
+template <size_t Size>
+static std::array<u8, Size> HexStringToArray(std::string_view str) {
+    std::array<u8, Size> out{};
+    for (size_t i = 0; i < 2 * Size; i += 2) {
+        auto d1 = str[i];
+        auto d2 = str[i + 1];
+        out[i / 2] = (ToHexNibble(d1) << 4) | ToHexNibble(d2);
+    }
+    return out;
+}
+
+std::array<u8, 16> operator""_array16(const char* str, size_t len) {
+    if (len != 32)
+        throw std::logic_error("Not of correct size.");
+    return HexStringToArray<16>(str);
+}
+
+std::array<u8, 32> operator""_array32(const char* str, size_t len) {
+    if (len != 64)
+        throw std::logic_error("Not of correct size.");
+    return HexStringToArray<32>(str);
+}
+
+KeyManager::KeyManager() {
+    // Initialize keys
+    const std::string hactool_keys_dir = FileUtil::GetHactoolConfigurationPath();
+    const std::string yuzu_keys_dir = FileUtil::GetUserPath(FileUtil::UserPath::KeysDir);
+    if (Settings::values.use_dev_keys) {
+        dev_mode = true;
+        AttemptLoadKeyFile(yuzu_keys_dir, hactool_keys_dir, "dev.keys", false);
+    } else {
+        dev_mode = false;
+        AttemptLoadKeyFile(yuzu_keys_dir, hactool_keys_dir, "prod.keys", false);
+    }
+
+    AttemptLoadKeyFile(yuzu_keys_dir, hactool_keys_dir, "title.keys", true);
+}
+
+void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {
+    std::ifstream file(filename);
+    if (!file.is_open())
+        return;
+
+    std::string line;
+    while (std::getline(file, line)) {
+        std::vector<std::string> out;
+        std::stringstream stream(line);
+        std::string item;
+        while (std::getline(stream, item, '='))
+            out.push_back(std::move(item));
+
+        if (out.size() != 2)
+            continue;
+
+        out[0].erase(std::remove(out[0].begin(), out[0].end(), ' '), out[0].end());
+        out[1].erase(std::remove(out[1].begin(), out[1].end(), ' '), out[1].end());
+
+        if (is_title_keys) {
+            auto rights_id_raw = HexStringToArray<16>(out[0]);
+            u128 rights_id{};
+            std::memcpy(rights_id.data(), rights_id_raw.data(), rights_id_raw.size());
+            Key128 key = HexStringToArray<16>(out[1]);
+            SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
+        } else {
+            std::transform(out[0].begin(), out[0].end(), out[0].begin(), ::tolower);
+            if (s128_file_id.find(out[0]) != s128_file_id.end()) {
+                const auto index = s128_file_id.at(out[0]);
+                Key128 key = HexStringToArray<16>(out[1]);
+                SetKey(index.type, key, index.field1, index.field2);
+            } else if (s256_file_id.find(out[0]) != s256_file_id.end()) {
+                const auto index = s256_file_id.at(out[0]);
+                Key256 key = HexStringToArray<32>(out[1]);
+                SetKey(index.type, key, index.field1, index.field2);
+            }
+        }
+    }
+}
+
+void KeyManager::AttemptLoadKeyFile(const std::string& dir1, const std::string& dir2,
+                                    const std::string& filename, bool title) {
+    if (FileUtil::Exists(dir1 + DIR_SEP + filename))
+        LoadFromFile(dir1 + DIR_SEP + filename, title);
+    else if (FileUtil::Exists(dir2 + DIR_SEP + filename))
+        LoadFromFile(dir2 + DIR_SEP + filename, title);
+}
+
+bool KeyManager::HasKey(S128KeyType id, u64 field1, u64 field2) const {
+    return s128_keys.find({id, field1, field2}) != s128_keys.end();
+}
+
+bool KeyManager::HasKey(S256KeyType id, u64 field1, u64 field2) const {
+    return s256_keys.find({id, field1, field2}) != s256_keys.end();
+}
+
+Key128 KeyManager::GetKey(S128KeyType id, u64 field1, u64 field2) const {
+    if (!HasKey(id, field1, field2))
+        return {};
+    return s128_keys.at({id, field1, field2});
+}
+
+Key256 KeyManager::GetKey(S256KeyType id, u64 field1, u64 field2) const {
+    if (!HasKey(id, field1, field2))
+        return {};
+    return s256_keys.at({id, field1, field2});
+}
+
+void KeyManager::SetKey(S128KeyType id, Key128 key, u64 field1, u64 field2) {
+    s128_keys[{id, field1, field2}] = key;
+}
+
+void KeyManager::SetKey(S256KeyType id, Key256 key, u64 field1, u64 field2) {
+    s256_keys[{id, field1, field2}] = key;
+}
+
+bool KeyManager::KeyFileExists(bool title) {
+    const std::string hactool_keys_dir = FileUtil::GetHactoolConfigurationPath();
+    const std::string yuzu_keys_dir = FileUtil::GetUserPath(FileUtil::UserPath::KeysDir);
+    if (title) {
+        return FileUtil::Exists(hactool_keys_dir + DIR_SEP + "title.keys") ||
+               FileUtil::Exists(yuzu_keys_dir + DIR_SEP + "title.keys");
+    }
+
+    if (Settings::values.use_dev_keys) {
+        return FileUtil::Exists(hactool_keys_dir + DIR_SEP + "dev.keys") ||
+               FileUtil::Exists(yuzu_keys_dir + DIR_SEP + "dev.keys");
+    }
+
+    return FileUtil::Exists(hactool_keys_dir + DIR_SEP + "prod.keys") ||
+           FileUtil::Exists(yuzu_keys_dir + DIR_SEP + "prod.keys");
+}
+
+const std::unordered_map<std::string, KeyIndex<S128KeyType>> KeyManager::s128_file_id = {
+    {"master_key_00", {S128KeyType::Master, 0, 0}},
+    {"master_key_01", {S128KeyType::Master, 1, 0}},
+    {"master_key_02", {S128KeyType::Master, 2, 0}},
+    {"master_key_03", {S128KeyType::Master, 3, 0}},
+    {"master_key_04", {S128KeyType::Master, 4, 0}},
+    {"package1_key_00", {S128KeyType::Package1, 0, 0}},
+    {"package1_key_01", {S128KeyType::Package1, 1, 0}},
+    {"package1_key_02", {S128KeyType::Package1, 2, 0}},
+    {"package1_key_03", {S128KeyType::Package1, 3, 0}},
+    {"package1_key_04", {S128KeyType::Package1, 4, 0}},
+    {"package2_key_00", {S128KeyType::Package2, 0, 0}},
+    {"package2_key_01", {S128KeyType::Package2, 1, 0}},
+    {"package2_key_02", {S128KeyType::Package2, 2, 0}},
+    {"package2_key_03", {S128KeyType::Package2, 3, 0}},
+    {"package2_key_04", {S128KeyType::Package2, 4, 0}},
+    {"titlekek_00", {S128KeyType::Titlekek, 0, 0}},
+    {"titlekek_01", {S128KeyType::Titlekek, 1, 0}},
+    {"titlekek_02", {S128KeyType::Titlekek, 2, 0}},
+    {"titlekek_03", {S128KeyType::Titlekek, 3, 0}},
+    {"titlekek_04", {S128KeyType::Titlekek, 4, 0}},
+    {"eticket_rsa_kek", {S128KeyType::ETicketRSAKek, 0, 0}},
+    {"key_area_key_application_00",
+     {S128KeyType::KeyArea, 0, static_cast<u64>(KeyAreaKeyType::Application)}},
+    {"key_area_key_application_01",
+     {S128KeyType::KeyArea, 1, static_cast<u64>(KeyAreaKeyType::Application)}},
+    {"key_area_key_application_02",
+     {S128KeyType::KeyArea, 2, static_cast<u64>(KeyAreaKeyType::Application)}},
+    {"key_area_key_application_03",
+     {S128KeyType::KeyArea, 3, static_cast<u64>(KeyAreaKeyType::Application)}},
+    {"key_area_key_application_04",
+     {S128KeyType::KeyArea, 4, static_cast<u64>(KeyAreaKeyType::Application)}},
+    {"key_area_key_ocean_00", {S128KeyType::KeyArea, 0, static_cast<u64>(KeyAreaKeyType::Ocean)}},
+    {"key_area_key_ocean_01", {S128KeyType::KeyArea, 1, static_cast<u64>(KeyAreaKeyType::Ocean)}},
+    {"key_area_key_ocean_02", {S128KeyType::KeyArea, 2, static_cast<u64>(KeyAreaKeyType::Ocean)}},
+    {"key_area_key_ocean_03", {S128KeyType::KeyArea, 3, static_cast<u64>(KeyAreaKeyType::Ocean)}},
+    {"key_area_key_ocean_04", {S128KeyType::KeyArea, 4, static_cast<u64>(KeyAreaKeyType::Ocean)}},
+    {"key_area_key_system_00", {S128KeyType::KeyArea, 0, static_cast<u64>(KeyAreaKeyType::System)}},
+    {"key_area_key_system_01", {S128KeyType::KeyArea, 1, static_cast<u64>(KeyAreaKeyType::System)}},
+    {"key_area_key_system_02", {S128KeyType::KeyArea, 2, static_cast<u64>(KeyAreaKeyType::System)}},
+    {"key_area_key_system_03", {S128KeyType::KeyArea, 3, static_cast<u64>(KeyAreaKeyType::System)}},
+    {"key_area_key_system_04", {S128KeyType::KeyArea, 4, static_cast<u64>(KeyAreaKeyType::System)}},
+};
+
+const std::unordered_map<std::string, KeyIndex<S256KeyType>> KeyManager::s256_file_id = {
+    {"header_key", {S256KeyType::Header, 0, 0}},
+    {"sd_card_save_key", {S256KeyType::SDSave, 0, 0}},
+    {"sd_card_nca_key", {S256KeyType::SDNCA, 0, 0}},
+};
+} // namespace Core::Crypto

src/core/crypto/key_manager.h

@@ -0,0 +1,120 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <string>
+#include <type_traits>
+#include <unordered_map>
+#include <vector>
+#include <fmt/format.h>
+#include "common/common_types.h"
+
+namespace Core::Crypto {
+
+using Key128 = std::array<u8, 0x10>;
+using Key256 = std::array<u8, 0x20>;
+using SHA256Hash = std::array<u8, 0x20>;
+
+static_assert(sizeof(Key128) == 16, "Key128 must be 128 bytes big.");
+static_assert(sizeof(Key256) == 32, "Key128 must be 128 bytes big.");
+
+enum class S256KeyType : u64 {
+    Header, //
+    SDSave, //
+    SDNCA,  //
+};
+
+enum class S128KeyType : u64 {
+    Master,        // f1=crypto revision
+    Package1,      // f1=crypto revision
+    Package2,      // f1=crypto revision
+    Titlekek,      // f1=crypto revision
+    ETicketRSAKek, //
+    KeyArea,       // f1=crypto revision f2=type {app, ocean, system}
+    SDSeed,        //
+    Titlekey,      // f1=rights id LSB f2=rights id MSB
+};
+
+enum class KeyAreaKeyType : u8 {
+    Application,
+    Ocean,
+    System,
+};
+
+template <typename KeyType>
+struct KeyIndex {
+    KeyType type;
+    u64 field1;
+    u64 field2;
+
+    std::string DebugInfo() const {
+        u8 key_size = 16;
+        if constexpr (std::is_same_v<KeyType, S256KeyType>)
+            key_size = 32;
+        return fmt::format("key_size={:02X}, key={:02X}, field1={:016X}, field2={:016X}", key_size,
+                           static_cast<u8>(type), field1, field2);
+    }
+};
+
+// The following two (== and hash) are so KeyIndex can be a key in unordered_map
+
+template <typename KeyType>
+bool operator==(const KeyIndex<KeyType>& lhs, const KeyIndex<KeyType>& rhs) {
+    return std::tie(lhs.type, lhs.field1, lhs.field2) == std::tie(rhs.type, rhs.field1, rhs.field2);
+}
+
+template <typename KeyType>
+bool operator!=(const KeyIndex<KeyType>& lhs, const KeyIndex<KeyType>& rhs) {
+    return !operator==(lhs, rhs);
+}
+
+} // namespace Core::Crypto
+
+namespace std {
+template <typename KeyType>
+struct hash<Core::Crypto::KeyIndex<KeyType>> {
+    size_t operator()(const Core::Crypto::KeyIndex<KeyType>& k) const {
+        using std::hash;
+
+        return ((hash<u64>()(static_cast<u64>(k.type)) ^ (hash<u64>()(k.field1) << 1)) >> 1) ^
+               (hash<u64>()(k.field2) << 1);
+    }
+};
+} // namespace std
+
+namespace Core::Crypto {
+
+std::array<u8, 0x10> operator"" _array16(const char* str, size_t len);
+std::array<u8, 0x20> operator"" _array32(const char* str, size_t len);
+
+class KeyManager {
+public:
+    KeyManager();
+
+    bool HasKey(S128KeyType id, u64 field1 = 0, u64 field2 = 0) const;
+    bool HasKey(S256KeyType id, u64 field1 = 0, u64 field2 = 0) const;
+
+    Key128 GetKey(S128KeyType id, u64 field1 = 0, u64 field2 = 0) const;
+    Key256 GetKey(S256KeyType id, u64 field1 = 0, u64 field2 = 0) const;
+
+    void SetKey(S128KeyType id, Key128 key, u64 field1 = 0, u64 field2 = 0);
+    void SetKey(S256KeyType id, Key256 key, u64 field1 = 0, u64 field2 = 0);
+
+    static bool KeyFileExists(bool title);
+
+private:
+    std::unordered_map<KeyIndex<S128KeyType>, Key128> s128_keys;
+    std::unordered_map<KeyIndex<S256KeyType>, Key256> s256_keys;
+
+    bool dev_mode;
+    void LoadFromFile(const std::string& filename, bool is_title_keys);
+    void AttemptLoadKeyFile(const std::string& dir1, const std::string& dir2,
+                            const std::string& filename, bool title);
+
+    static const std::unordered_map<std::string, KeyIndex<S128KeyType>> s128_file_id;
+    static const std::unordered_map<std::string, KeyIndex<S256KeyType>> s256_file_id;
+};
+} // namespace Core::Crypto

src/core/crypto/sha_util.cpp

@@ -0,0 +1,5 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+namespace Crypto {} // namespace Crypto

src/core/crypto/sha_util.h

@@ -0,0 +1,20 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/assert.h"
+#include "core/file_sys/vfs.h"
+#include "key_manager.h"
+#include "mbedtls/cipher.h"
+
+namespace Crypto {
+typedef std::array<u8, 0x20> SHA256Hash;
+
+inline SHA256Hash operator"" _HASH(const char* data, size_t len) {
+    if (len != 0x40)
+        return {};
+}
+
+} // namespace Crypto

src/core/file_sys/card_image.cpp

@@ -0,0 +1,149 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <array>
+#include <string>
+#include <core/loader/loader.h>
+#include "core/file_sys/card_image.h"
+#include "core/file_sys/partition_filesystem.h"
+#include "core/file_sys/vfs_offset.h"
+
+namespace FileSys {
+
+XCI::XCI(VirtualFile file_) : file(std::move(file_)), partitions(0x4) {
+    if (file->ReadObject(&header) != sizeof(GamecardHeader)) {
+        status = Loader::ResultStatus::ErrorInvalidFormat;
+        return;
+    }
+
+    if (header.magic != Common::MakeMagic('H', 'E', 'A', 'D')) {
+        status = Loader::ResultStatus::ErrorInvalidFormat;
+        return;
+    }
+
+    PartitionFilesystem main_hfs(
+        std::make_shared<OffsetVfsFile>(file, header.hfs_size, header.hfs_offset));
+
+    if (main_hfs.GetStatus() != Loader::ResultStatus::Success) {
+        status = main_hfs.GetStatus();
+        return;
+    }
+
+    static constexpr std::array<const char*, 0x4> partition_names = {"update", "normal", "secure",
+                                                                     "logo"};
+
+    for (XCIPartition partition :
+         {XCIPartition::Update, XCIPartition::Normal, XCIPartition::Secure, XCIPartition::Logo}) {
+        auto raw = main_hfs.GetFile(partition_names[static_cast<size_t>(partition)]);
+        if (raw != nullptr)
+            partitions[static_cast<size_t>(partition)] = std::make_shared<PartitionFilesystem>(raw);
+    }
+
+    auto result = AddNCAFromPartition(XCIPartition::Secure);
+    if (result != Loader::ResultStatus::Success) {
+        status = result;
+        return;
+    }
+
+    result = AddNCAFromPartition(XCIPartition::Update);
+    if (result != Loader::ResultStatus::Success) {
+        status = result;
+        return;
+    }
+
+    result = AddNCAFromPartition(XCIPartition::Normal);
+    if (result != Loader::ResultStatus::Success) {
+        status = result;
+        return;
+    }
+
+    if (GetFormatVersion() >= 0x2) {
+        result = AddNCAFromPartition(XCIPartition::Logo);
+        if (result != Loader::ResultStatus::Success) {
+            status = result;
+            return;
+        }
+    }
+
+    status = Loader::ResultStatus::Success;
+}
+
+Loader::ResultStatus XCI::GetStatus() const {
+    return status;
+}
+
+VirtualDir XCI::GetPartition(XCIPartition partition) const {
+    return partitions[static_cast<size_t>(partition)];
+}
+
+VirtualDir XCI::GetSecurePartition() const {
+    return GetPartition(XCIPartition::Secure);
+}
+
+VirtualDir XCI::GetNormalPartition() const {
+    return GetPartition(XCIPartition::Normal);
+}
+
+VirtualDir XCI::GetUpdatePartition() const {
+    return GetPartition(XCIPartition::Update);
+}
+
+VirtualDir XCI::GetLogoPartition() const {
+    return GetPartition(XCIPartition::Logo);
+}
+
+std::shared_ptr<NCA> XCI::GetNCAByType(NCAContentType type) const {
+    const auto iter =
+        std::find_if(ncas.begin(), ncas.end(),
+                     [type](const std::shared_ptr<NCA>& nca) { return nca->GetType() == type; });
+    return iter == ncas.end() ? nullptr : *iter;
+}
+
+VirtualFile XCI::GetNCAFileByType(NCAContentType type) const {
+    auto nca = GetNCAByType(type);
+    if (nca != nullptr)
+        return nca->GetBaseFile();
+    return nullptr;
+}
+
+std::vector<std::shared_ptr<VfsFile>> XCI::GetFiles() const {
+    return {};
+}
+
+std::vector<std::shared_ptr<VfsDirectory>> XCI::GetSubdirectories() const {
+    return std::vector<std::shared_ptr<VfsDirectory>>();
+}
+
+std::string XCI::GetName() const {
+    return file->GetName();
+}
+
+std::shared_ptr<VfsDirectory> 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<size_t>(part)] == nullptr) {
+        return Loader::ResultStatus::ErrorInvalidFormat;
+    }
+
+    for (const VirtualFile& file : partitions[static_cast<size_t>(part)]->GetFiles()) {
+        if (file->GetExtension() != "nca")
+            continue;
+        auto nca = std::make_shared<NCA>(file);
+        if (nca->GetStatus() == Loader::ResultStatus::Success)
+            ncas.push_back(std::move(nca));
+    }
+
+    return Loader::ResultStatus::Success;
+}
+
+u8 XCI::GetFormatVersion() const {
+    return GetLogoPartition() == nullptr ? 0x1 : 0x2;
+}
+} // namespace FileSys

src/core/file_sys/card_image.h

@@ -0,0 +1,96 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <vector>
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/file_sys/content_archive.h"
+#include "core/file_sys/vfs.h"
+#include "core/loader/loader.h"
+
+namespace FileSys {
+
+enum class GamecardSize : u8 {
+    S_1GB = 0xFA,
+    S_2GB = 0xF8,
+    S_4GB = 0xF0,
+    S_8GB = 0xE0,
+    S_16GB = 0xE1,
+    S_32GB = 0xE2,
+};
+
+struct GamecardInfo {
+    std::array<u8, 0x70> data;
+};
+static_assert(sizeof(GamecardInfo) == 0x70, "GamecardInfo has incorrect size.");
+
+struct GamecardHeader {
+    std::array<u8, 0x100> signature;
+    u32_le magic;
+    u32_le secure_area_start;
+    u32_le backup_area_start;
+    u8 kek_index;
+    GamecardSize size;
+    u8 header_version;
+    u8 flags;
+    u64_le package_id;
+    u64_le valid_data_end;
+    u128 info_iv;
+    u64_le hfs_offset;
+    u64_le hfs_size;
+    std::array<u8, 0x20> hfs_header_hash;
+    std::array<u8, 0x20> initial_data_hash;
+    u32_le secure_mode_flag;
+    u32_le title_key_flag;
+    u32_le key_flag;
+    u32_le normal_area_end;
+    GamecardInfo info;
+};
+static_assert(sizeof(GamecardHeader) == 0x200, "GamecardHeader has incorrect size.");
+
+enum class XCIPartition : u8 { Update, Normal, Secure, Logo };
+
+class XCI : public ReadOnlyVfsDirectory {
+public:
+    explicit XCI(VirtualFile file);
+
+    Loader::ResultStatus GetStatus() const;
+
+    u8 GetFormatVersion() const;
+
+    VirtualDir GetPartition(XCIPartition partition) const;
+    VirtualDir GetSecurePartition() const;
+    VirtualDir GetNormalPartition() const;
+    VirtualDir GetUpdatePartition() const;
+    VirtualDir GetLogoPartition() const;
+
+    std::shared_ptr<NCA> GetNCAByType(NCAContentType type) const;
+    VirtualFile GetNCAFileByType(NCAContentType type) const;
+
+    std::vector<std::shared_ptr<VfsFile>> GetFiles() const override;
+
+    std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override;
+
+    std::string GetName() const override;
+
+    std::shared_ptr<VfsDirectory> GetParentDirectory() const override;
+
+protected:
+    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
+
+private:
+    Loader::ResultStatus AddNCAFromPartition(XCIPartition part);
+
+    VirtualFile file;
+    GamecardHeader header{};
+
+    Loader::ResultStatus status;
+
+    std::vector<VirtualDir> partitions;
+    std::vector<std::shared_ptr<NCA>> ncas;
+};
+} // namespace FileSys

src/core/file_sys/content_archive.cpp

@@ -4,12 +4,14 @@
 
 #include <algorithm>
 #include <utility>
-
+#include <boost/optional.hpp>
 #include "common/logging/log.h"
+#include "core/crypto/aes_util.h"
+#include "core/crypto/ctr_encryption_layer.h"
 #include "core/file_sys/content_archive.h"
+#include "core/file_sys/romfs.h"
 #include "core/file_sys/vfs_offset.h"
 #include "core/loader/loader.h"
-#include "romfs.h"
 
 namespace FileSys {
 
@@ -29,11 +31,19 @@ enum class NCASectionFilesystemType : u8 {
 struct NCASectionHeaderBlock {
     INSERT_PADDING_BYTES(3);
     NCASectionFilesystemType filesystem_type;
-    u8 crypto_type;
+    NCASectionCryptoType crypto_type;
     INSERT_PADDING_BYTES(3);
 };
 static_assert(sizeof(NCASectionHeaderBlock) == 0x8, "NCASectionHeaderBlock has incorrect size.");
 
+struct NCASectionRaw {
+    NCASectionHeaderBlock header;
+    std::array<u8, 0x138> block_data;
+    std::array<u8, 0x8> section_ctr;
+    INSERT_PADDING_BYTES(0xB8);
+};
+static_assert(sizeof(NCASectionRaw) == 0x200, "NCASectionRaw has incorrect size.");
+
 struct PFS0Superblock {
     NCASectionHeaderBlock header_block;
     std::array<u8, 0x20> hash;
@@ -43,67 +53,170 @@ struct PFS0Superblock {
     u64_le hash_table_size;
     u64_le pfs0_header_offset;
     u64_le pfs0_size;
-    INSERT_PADDING_BYTES(432);
+    INSERT_PADDING_BYTES(0x1B0);
 };
 static_assert(sizeof(PFS0Superblock) == 0x200, "PFS0Superblock has incorrect size.");
 
 struct RomFSSuperblock {
     NCASectionHeaderBlock header_block;
     IVFCHeader ivfc;
+    INSERT_PADDING_BYTES(0x118);
 };
-static_assert(sizeof(RomFSSuperblock) == 0xE8, "RomFSSuperblock has incorrect size.");
+static_assert(sizeof(RomFSSuperblock) == 0x200, "RomFSSuperblock has incorrect size.");
+
+union NCASectionHeader {
+    NCASectionRaw raw;
+    PFS0Superblock pfs0;
+    RomFSSuperblock romfs;
+};
+static_assert(sizeof(NCASectionHeader) == 0x200, "NCASectionHeader has incorrect size.");
+
+bool IsValidNCA(const NCAHeader& header) {
+    // TODO(DarkLordZach): Add NCA2/NCA0 support.
+    return header.magic == Common::MakeMagic('N', 'C', 'A', '3');
+}
+
+boost::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const {
+    u8 master_key_id = header.crypto_type;
+    if (header.crypto_type_2 > master_key_id)
+        master_key_id = header.crypto_type_2;
+    if (master_key_id > 0)
+        --master_key_id;
+
+    if (!keys.HasKey(Core::Crypto::S128KeyType::KeyArea, master_key_id, header.key_index))
+        return boost::none;
+
+    std::vector<u8> key_area(header.key_area.begin(), header.key_area.end());
+    Core::Crypto::AESCipher<Core::Crypto::Key128> cipher(
+        keys.GetKey(Core::Crypto::S128KeyType::KeyArea, master_key_id, header.key_index),
+        Core::Crypto::Mode::ECB);
+    cipher.Transcode(key_area.data(), key_area.size(), key_area.data(), Core::Crypto::Op::Decrypt);
+
+    Core::Crypto::Key128 out;
+    if (type == NCASectionCryptoType::XTS)
+        std::copy(key_area.begin(), key_area.begin() + 0x10, out.begin());
+    else if (type == NCASectionCryptoType::CTR)
+        std::copy(key_area.begin() + 0x20, key_area.begin() + 0x30, out.begin());
+    else
+        LOG_CRITICAL(Crypto, "Called GetKeyAreaKey on invalid NCASectionCryptoType type={:02X}",
+                     static_cast<u8>(type));
+    u128 out_128{};
+    memcpy(out_128.data(), out.data(), 16);
+    LOG_DEBUG(Crypto, "called with crypto_rev={:02X}, kak_index={:02X}, key={:016X}{:016X}",
+              master_key_id, header.key_index, out_128[1], out_128[0]);
+
+    return out;
+}
+
+VirtualFile NCA::Decrypt(NCASectionHeader header, VirtualFile in, u64 starting_offset) const {
+    if (!encrypted)
+        return in;
+
+    switch (header.raw.header.crypto_type) {
+    case NCASectionCryptoType::NONE:
+        LOG_DEBUG(Crypto, "called with mode=NONE");
+        return in;
+    case NCASectionCryptoType::CTR:
+        LOG_DEBUG(Crypto, "called with mode=CTR, starting_offset={:016X}", starting_offset);
+        {
+            const auto key = GetKeyAreaKey(NCASectionCryptoType::CTR);
+            if (key == boost::none)
+                return nullptr;
+            auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(
+                std::move(in), key.value(), starting_offset);
+            std::vector<u8> iv(16);
+            for (u8 i = 0; i < 8; ++i)
+                iv[i] = header.raw.section_ctr[0x8 - i - 1];
+            out->SetIV(iv);
+            return std::static_pointer_cast<VfsFile>(out);
+        }
+    case NCASectionCryptoType::XTS:
+        // TODO(DarkLordZach): Implement XTSEncryptionLayer and title key encryption.
+    default:
+        LOG_ERROR(Crypto, "called with unhandled crypto type={:02X}",
+                  static_cast<u8>(header.raw.header.crypto_type));
+        return nullptr;
+    }
+}
 
 NCA::NCA(VirtualFile file_) : file(std::move(file_)) {
     if (sizeof(NCAHeader) != file->ReadObject(&header))
-        LOG_CRITICAL(Loader, "File reader errored out during header read.");
+        LOG_ERROR(Loader, "File reader errored out during header read.");
+
+    encrypted = false;
 
     if (!IsValidNCA(header)) {
-        status = Loader::ResultStatus::ErrorInvalidFormat;
-        return;
+        NCAHeader dec_header{};
+        Core::Crypto::AESCipher<Core::Crypto::Key256> cipher(
+            keys.GetKey(Core::Crypto::S256KeyType::Header), Core::Crypto::Mode::XTS);
+        cipher.XTSTranscode(&header, sizeof(NCAHeader), &dec_header, 0, 0x200,
+                            Core::Crypto::Op::Decrypt);
+        if (IsValidNCA(dec_header)) {
+            header = dec_header;
+            encrypted = true;
+        } else {
+            if (!keys.HasKey(Core::Crypto::S256KeyType::Header))
+                status = Loader::ResultStatus::ErrorMissingKeys;
+            else
+                status = Loader::ResultStatus::ErrorDecrypting;
+            return;
+        }
     }
 
-    std::ptrdiff_t number_sections =
+    const std::ptrdiff_t number_sections =
         std::count_if(std::begin(header.section_tables), std::end(header.section_tables),
                       [](NCASectionTableEntry entry) { return entry.media_offset > 0; });
 
+    std::vector<NCASectionHeader> sections(number_sections);
+    const auto length_sections = SECTION_HEADER_SIZE * number_sections;
+
+    if (encrypted) {
+        auto raw = file->ReadBytes(length_sections, SECTION_HEADER_OFFSET);
+        Core::Crypto::AESCipher<Core::Crypto::Key256> cipher(
+            keys.GetKey(Core::Crypto::S256KeyType::Header), Core::Crypto::Mode::XTS);
+        cipher.XTSTranscode(raw.data(), length_sections, sections.data(), 2, SECTION_HEADER_SIZE,
+                            Core::Crypto::Op::Decrypt);
+    } else {
+        file->ReadBytes(sections.data(), length_sections, SECTION_HEADER_OFFSET);
+    }
+
     for (std::ptrdiff_t i = 0; i < number_sections; ++i) {
-        // Seek to beginning of this section.
-        NCASectionHeaderBlock block{};
-        if (sizeof(NCASectionHeaderBlock) !=
-            file->ReadObject(&block, SECTION_HEADER_OFFSET + i * SECTION_HEADER_SIZE))
-            LOG_CRITICAL(Loader, "File reader errored out during header read.");
-
-        if (block.filesystem_type == NCASectionFilesystemType::ROMFS) {
-            RomFSSuperblock sb{};
-            if (sizeof(RomFSSuperblock) !=
-                file->ReadObject(&sb, SECTION_HEADER_OFFSET + i * SECTION_HEADER_SIZE))
-                LOG_CRITICAL(Loader, "File reader errored out during header read.");
+        auto section = sections[i];
 
+        if (section.raw.header.filesystem_type == NCASectionFilesystemType::ROMFS) {
             const size_t romfs_offset =
                 header.section_tables[i].media_offset * MEDIA_OFFSET_MULTIPLIER +
-                sb.ivfc.levels[IVFC_MAX_LEVEL - 1].offset;
-            const size_t romfs_size = sb.ivfc.levels[IVFC_MAX_LEVEL - 1].size;
-            files.emplace_back(std::make_shared<OffsetVfsFile>(file, romfs_size, romfs_offset));
-            romfs = files.back();
-        } else if (block.filesystem_type == NCASectionFilesystemType::PFS0) {
-            PFS0Superblock sb{};
-            // Seek back to beginning of this section.
-            if (sizeof(PFS0Superblock) !=
-                file->ReadObject(&sb, SECTION_HEADER_OFFSET + i * SECTION_HEADER_SIZE))
-                LOG_CRITICAL(Loader, "File reader errored out during header read.");
-
+                section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].offset;
+            const size_t romfs_size = section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].size;
+            auto dec =
+                Decrypt(section, std::make_shared<OffsetVfsFile>(file, romfs_size, romfs_offset),
+                        romfs_offset);
+            if (dec != nullptr) {
+                files.push_back(std::move(dec));
+                romfs = files.back();
+            } else {
+                status = Loader::ResultStatus::ErrorMissingKeys;
+                return;
+            }
+        } else if (section.raw.header.filesystem_type == NCASectionFilesystemType::PFS0) {
             u64 offset = (static_cast<u64>(header.section_tables[i].media_offset) *
                           MEDIA_OFFSET_MULTIPLIER) +
-                         sb.pfs0_header_offset;
+                         section.pfs0.pfs0_header_offset;
             u64 size = MEDIA_OFFSET_MULTIPLIER * (header.section_tables[i].media_end_offset -
                                                   header.section_tables[i].media_offset);
-            auto npfs = std::make_shared<PartitionFilesystem>(
-                std::make_shared<OffsetVfsFile>(file, size, offset));
+            auto dec =
+                Decrypt(section, std::make_shared<OffsetVfsFile>(file, size, offset), offset);
+            if (dec != nullptr) {
+                auto npfs = std::make_shared<PartitionFilesystem>(std::move(dec));
 
-            if (npfs->GetStatus() == Loader::ResultStatus::Success) {
-                dirs.emplace_back(npfs);
-                if (IsDirectoryExeFS(dirs.back()))
-                    exefs = dirs.back();
+                if (npfs->GetStatus() == Loader::ResultStatus::Success) {
+                    dirs.push_back(std::move(npfs));
+                    if (IsDirectoryExeFS(dirs.back()))
+                        exefs = dirs.back();
+                }
+            } else {
+                status = Loader::ResultStatus::ErrorMissingKeys;
+                return;
             }
         }
     }
@@ -153,6 +266,10 @@ VirtualDir NCA::GetExeFS() const {
     return exefs;
 }
 
+VirtualFile NCA::GetBaseFile() const {
+    return file;
+}
+
 bool NCA::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
     return false;
 }

src/core/file_sys/content_archive.h

@@ -8,14 +8,18 @@
 #include <memory>
 #include <string>
 #include <vector>
-
+#include <boost/optional.hpp>
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/swap.h"
+#include "core/crypto/key_manager.h"
 #include "core/file_sys/partition_filesystem.h"
+#include "core/loader/loader.h"
 
 namespace FileSys {
 
+union NCASectionHeader;
+
 enum class NCAContentType : u8 {
     Program = 0,
     Meta = 1,
@@ -24,6 +28,13 @@ enum class NCAContentType : u8 {
     Data = 4,
 };
 
+enum class NCASectionCryptoType : u8 {
+    NONE = 1,
+    XTS = 2,
+    CTR = 3,
+    BKTR = 4,
+};
+
 struct NCASectionTableEntry {
     u32_le media_offset;
     u32_le media_end_offset;
@@ -48,7 +59,7 @@ struct NCAHeader {
     std::array<u8, 0x10> rights_id;
     std::array<NCASectionTableEntry, 0x4> section_tables;
     std::array<std::array<u8, 0x20>, 0x4> hash_tables;
-    std::array<std::array<u8, 0x10>, 0x4> key_area;
+    std::array<u8, 0x40> key_area;
     INSERT_PADDING_BYTES(0xC0);
 };
 static_assert(sizeof(NCAHeader) == 0x400, "NCAHeader has incorrect size.");
@@ -58,10 +69,7 @@ inline bool IsDirectoryExeFS(const std::shared_ptr<VfsDirectory>& pfs) {
     return pfs->GetFile("main") != nullptr && pfs->GetFile("main.npdm") != nullptr;
 }
 
-inline bool IsValidNCA(const NCAHeader& header) {
-    return header.magic == Common::MakeMagic('N', 'C', 'A', '2') ||
-           header.magic == Common::MakeMagic('N', 'C', 'A', '3');
-}
+bool IsValidNCA(const NCAHeader& header);
 
 // An implementation of VfsDirectory that represents a Nintendo Content Archive (NCA) conatiner.
 // After construction, use GetStatus to determine if the file is valid and ready to be used.
@@ -81,10 +89,15 @@ public:
     VirtualFile GetRomFS() const;
     VirtualDir GetExeFS() const;
 
+    VirtualFile GetBaseFile() const;
+
 protected:
     bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
 
 private:
+    boost::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const;
+    VirtualFile Decrypt(NCASectionHeader header, VirtualFile in, u64 starting_offset) const;
+
     std::vector<VirtualDir> dirs;
     std::vector<VirtualFile> files;
 
@@ -95,6 +108,10 @@ private:
     NCAHeader header{};
 
     Loader::ResultStatus status{};
+
+    bool encrypted;
+
+    Core::Crypto::KeyManager keys;
 };
 
 } // namespace FileSys

src/core/file_sys/vfs.cpp

@@ -285,6 +285,26 @@ bool ReadOnlyVfsDirectory::Rename(std::string_view name) {
     return false;
 }
 
+bool DeepEquals(const VirtualFile& file1, const VirtualFile& file2, size_t block_size) {
+    if (file1->GetSize() != file2->GetSize())
+        return false;
+
+    std::vector<u8> f1_v(block_size);
+    std::vector<u8> f2_v(block_size);
+    for (size_t i = 0; i < file1->GetSize(); i += block_size) {
+        auto f1_vs = file1->Read(f1_v.data(), block_size, i);
+        auto f2_vs = file2->Read(f2_v.data(), block_size, i);
+
+        if (f1_vs != f2_vs)
+            return false;
+        auto iters = std::mismatch(f1_v.begin(), f1_v.end(), f2_v.begin(), f2_v.end());
+        if (iters.first != f1_v.end() && iters.second != f2_v.end())
+            return false;
+    }
+
+    return true;
+}
+
 bool VfsRawCopy(VirtualFile src, VirtualFile dest) {
     if (src == nullptr || dest == nullptr)
         return false;

src/core/file_sys/vfs.h

@@ -245,6 +245,9 @@ struct ReadOnlyVfsDirectory : public VfsDirectory {
     bool Rename(std::string_view name) override;
 };
 
+// Compare the two files, byte-for-byte, in increments specificed by block_size
+bool DeepEquals(const VirtualFile& file1, const VirtualFile& file2, size_t block_size = 0x200);
+
 // A method that copies the raw data between two different implementations of VirtualFile. If you
 // are using the same implementation, it is probably better to use the Copy method in the parent
 // directory of src/dest.

src/core/file_sys/vfs_vector.cpp

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
+#include <utility>
 #include "core/file_sys/vfs_vector.h"
 
 namespace FileSys {
@@ -31,16 +32,18 @@ bool VectorVfsDirectory::IsReadable() const {
 std::string VectorVfsDirectory::GetName() const {
     return name;
 }
+
 std::shared_ptr<VfsDirectory> VectorVfsDirectory::GetParentDirectory() const {
     return parent;
 }
 
 template <typename T>
 static bool FindAndRemoveVectorElement(std::vector<T>& vec, std::string_view name) {
-    auto iter = std::find_if(vec.begin(), vec.end(), [name](T e) { return e->GetName() == name; });
+    const auto iter =
+        std::find_if(vec.begin(), vec.end(), [name](const T& e) { return e->GetName() == name; });
     if (iter == vec.end())
         return false;
-    auto old_size = vec.size();
+
     vec.erase(iter);
     return true;
 }
@@ -77,7 +80,7 @@ void VectorVfsDirectory::AddDirectory(VirtualDir dir) {
 bool VectorVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
     if (!DeleteFile(file->GetName()))
         return false;
-    dirs.emplace_back(dir);
+    dirs.emplace_back(std::move(dir));
     return true;
 }
 } // namespace FileSys

src/core/gdbstub/gdbstub.cpp

@@ -37,45 +37,46 @@
 #include "core/core.h"
 #include "core/core_cpu.h"
 #include "core/gdbstub/gdbstub.h"
-#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/scheduler.h"
 #include "core/loader/loader.h"
 #include "core/memory.h"
 
-const int GDB_BUFFER_SIZE = 10000;
+namespace GDBStub {
+namespace {
+constexpr int GDB_BUFFER_SIZE = 10000;
 
-const char GDB_STUB_START = '$';
-const char GDB_STUB_END = '#';
-const char GDB_STUB_ACK = '+';
-const char GDB_STUB_NACK = '-';
+constexpr char GDB_STUB_START = '$';
+constexpr char GDB_STUB_END = '#';
+constexpr char GDB_STUB_ACK = '+';
+constexpr char GDB_STUB_NACK = '-';
 
 #ifndef SIGTRAP
-const u32 SIGTRAP = 5;
+constexpr u32 SIGTRAP = 5;
 #endif
 
 #ifndef SIGTERM
-const u32 SIGTERM = 15;
+constexpr u32 SIGTERM = 15;
 #endif
 
 #ifndef MSG_WAITALL
-const u32 MSG_WAITALL = 8;
+constexpr u32 MSG_WAITALL = 8;
 #endif
 
-const u32 LR_REGISTER = 30;
-const u32 SP_REGISTER = 31;
-const u32 PC_REGISTER = 32;
-const u32 CPSR_REGISTER = 33;
-const u32 UC_ARM64_REG_Q0 = 34;
-const u32 FPSCR_REGISTER = 66;
+constexpr u32 LR_REGISTER = 30;
+constexpr u32 SP_REGISTER = 31;
+constexpr u32 PC_REGISTER = 32;
+constexpr u32 CPSR_REGISTER = 33;
+constexpr u32 UC_ARM64_REG_Q0 = 34;
+constexpr u32 FPSCR_REGISTER = 66;
 
 // TODO/WiP - Used while working on support for FPU
-const u32 TODO_DUMMY_REG_997 = 997;
-const u32 TODO_DUMMY_REG_998 = 998;
+constexpr u32 TODO_DUMMY_REG_997 = 997;
+constexpr u32 TODO_DUMMY_REG_998 = 998;
 
 // For sample XML files see the GDB source /gdb/features
 // GDB also wants the l character at the start
 // This XML defines what the registers are for this specific ARM device
-static const char* target_xml =
+constexpr char target_xml[] =
     R"(l<?xml version="1.0"?>
 <!DOCTYPE target SYSTEM "gdb-target.dtd">
 <target version="1.0">
@@ -141,30 +142,28 @@ static const char* target_xml =
 </target>
 )";
 
-namespace GDBStub {
+int gdbserver_socket = -1;
 
-static int gdbserver_socket = -1;
+u8 command_buffer[GDB_BUFFER_SIZE];
+u32 command_length;
 
-static u8 command_buffer[GDB_BUFFER_SIZE];
-static u32 command_length;
+u32 latest_signal = 0;
+bool memory_break = false;
 
-static u32 latest_signal = 0;
-static bool memory_break = false;
-
-static Kernel::Thread* current_thread = nullptr;
-static u32 current_core = 0;
+Kernel::Thread* current_thread = nullptr;
+u32 current_core = 0;
 
 // Binding to a port within the reserved ports range (0-1023) requires root permissions,
 // so default to a port outside of that range.
-static u16 gdbstub_port = 24689;
+u16 gdbstub_port = 24689;
 
-static bool halt_loop = true;
-static bool step_loop = false;
-static bool send_trap = false;
+bool halt_loop = true;
+bool step_loop = false;
+bool send_trap = false;
 
 // If set to false, the server will never be started and no
 // gdbstub-related functions will be executed.
-static std::atomic<bool> server_enabled(false);
+std::atomic<bool> server_enabled(false);
 
 #ifdef _WIN32
 WSADATA InitData;
@@ -172,23 +171,26 @@ WSADATA InitData;
 
 struct Breakpoint {
     bool active;
-    PAddr addr;
+    VAddr addr;
     u64 len;
+    std::array<u8, 4> inst;
 };
 
-static std::map<u64, Breakpoint> breakpoints_execute;
-static std::map<u64, Breakpoint> breakpoints_read;
-static std::map<u64, Breakpoint> breakpoints_write;
+using BreakpointMap = std::map<VAddr, Breakpoint>;
+BreakpointMap breakpoints_execute;
+BreakpointMap breakpoints_read;
+BreakpointMap breakpoints_write;
 
 struct Module {
     std::string name;
-    PAddr beg;
-    PAddr end;
+    VAddr beg;
+    VAddr end;
 };
 
-static std::vector<Module> modules;
+std::vector<Module> modules;
+} // Anonymous namespace
 
-void RegisterModule(std::string name, PAddr beg, PAddr end, bool add_elf_ext) {
+void RegisterModule(std::string name, VAddr beg, VAddr end, bool add_elf_ext) {
     Module module;
     if (add_elf_ext) {
         Common::SplitPath(name, nullptr, &module.name, nullptr);
@@ -419,11 +421,11 @@ static u8 CalculateChecksum(const u8* buffer, size_t length) {
 }
 
 /**
- * Get the list of breakpoints for a given breakpoint type.
+ * Get the map of breakpoints for a given breakpoint type.
  *
- * @param type Type of breakpoint list.
+ * @param type Type of breakpoint map.
  */
-static std::map<u64, Breakpoint>& GetBreakpointList(BreakpointType type) {
+static BreakpointMap& GetBreakpointMap(BreakpointType type) {
     switch (type) {
     case BreakpointType::Execute:
         return breakpoints_execute;
@@ -442,20 +444,24 @@ static std::map<u64, Breakpoint>& GetBreakpointList(BreakpointType type) {
  * @param type Type of breakpoint.
  * @param addr Address of breakpoint.
  */
-static void RemoveBreakpoint(BreakpointType type, PAddr addr) {
-    std::map<u64, Breakpoint>& p = GetBreakpointList(type);
+static void RemoveBreakpoint(BreakpointType type, VAddr addr) {
+    BreakpointMap& p = GetBreakpointMap(type);
 
-    auto bp = p.find(static_cast<u64>(addr));
-    if (bp != p.end()) {
-        LOG_DEBUG(Debug_GDBStub, "gdb: removed a breakpoint: {:016X} bytes at {:016X} of type {}",
-                  bp->second.len, bp->second.addr, static_cast<int>(type));
-        p.erase(static_cast<u64>(addr));
+    const auto bp = p.find(addr);
+    if (bp == p.end()) {
+        return;
     }
+
+    LOG_DEBUG(Debug_GDBStub, "gdb: removed a breakpoint: {:016X} bytes at {:016X} of type {}",
+              bp->second.len, bp->second.addr, static_cast<int>(type));
+    Memory::WriteBlock(bp->second.addr, bp->second.inst.data(), bp->second.inst.size());
+    Core::System::GetInstance().InvalidateCpuInstructionCaches();
+    p.erase(addr);
 }
 
-BreakpointAddress GetNextBreakpointFromAddress(PAddr addr, BreakpointType type) {
-    std::map<u64, Breakpoint>& p = GetBreakpointList(type);
-    auto next_breakpoint = p.lower_bound(static_cast<u64>(addr));
+BreakpointAddress GetNextBreakpointFromAddress(VAddr addr, BreakpointType type) {
+    const BreakpointMap& p = GetBreakpointMap(type);
+    const auto next_breakpoint = p.lower_bound(addr);
     BreakpointAddress breakpoint;
 
     if (next_breakpoint != p.end()) {
@@ -469,36 +475,38 @@ BreakpointAddress GetNextBreakpointFromAddress(PAddr addr, BreakpointType type)
     return breakpoint;
 }
 
-bool CheckBreakpoint(PAddr addr, BreakpointType type) {
+bool CheckBreakpoint(VAddr addr, BreakpointType type) {
     if (!IsConnected()) {
         return false;
     }
 
-    std::map<u64, Breakpoint>& p = GetBreakpointList(type);
+    const BreakpointMap& p = GetBreakpointMap(type);
+    const auto bp = p.find(addr);
 
-    auto bp = p.find(static_cast<u64>(addr));
-    if (bp != p.end()) {
-        u64 len = bp->second.len;
+    if (bp == p.end()) {
+        return false;
+    }
 
-        // IDA Pro defaults to 4-byte breakpoints for all non-hardware breakpoints
-        // no matter if it's a 4-byte or 2-byte instruction. When you execute a
-        // Thumb instruction with a 4-byte breakpoint set, it will set a breakpoint on
-        // two instructions instead of the single instruction you placed the breakpoint
-        // on. So, as a way to make sure that execution breakpoints are only breaking
-        // on the instruction that was specified, set the length of an execution
-        // breakpoint to 1. This should be fine since the CPU should never begin executing
-        // an instruction anywhere except the beginning of the instruction.
-        if (type == BreakpointType::Execute) {
-            len = 1;
-        }
+    u64 len = bp->second.len;
 
-        if (bp->second.active && (addr >= bp->second.addr && addr < bp->second.addr + len)) {
-            LOG_DEBUG(Debug_GDBStub,
-                      "Found breakpoint type {} @ {:016X}, range: {:016X}"
-                      " - {:016X} ({:X} bytes)",
-                      static_cast<int>(type), addr, bp->second.addr, bp->second.addr + len, len);
-            return true;
-        }
+    // IDA Pro defaults to 4-byte breakpoints for all non-hardware breakpoints
+    // no matter if it's a 4-byte or 2-byte instruction. When you execute a
+    // Thumb instruction with a 4-byte breakpoint set, it will set a breakpoint on
+    // two instructions instead of the single instruction you placed the breakpoint
+    // on. So, as a way to make sure that execution breakpoints are only breaking
+    // on the instruction that was specified, set the length of an execution
+    // breakpoint to 1. This should be fine since the CPU should never begin executing
+    // an instruction anywhere except the beginning of the instruction.
+    if (type == BreakpointType::Execute) {
+        len = 1;
+    }
+
+    if (bp->second.active && (addr >= bp->second.addr && addr < bp->second.addr + len)) {
+        LOG_DEBUG(Debug_GDBStub,
+                  "Found breakpoint type {} @ {:016X}, range: {:016X}"
+                  " - {:016X} ({:X} bytes)",
+                  static_cast<int>(type), addr, bp->second.addr, bp->second.addr + len, len);
+        return true;
     }
 
     return false;
@@ -932,6 +940,7 @@ static void WriteMemory() {
 
     GdbHexToMem(data.data(), len_pos + 1, len);
     Memory::WriteBlock(addr, data.data(), len);
+    Core::System::GetInstance().InvalidateCpuInstructionCaches();
     SendReply("OK");
 }
 
@@ -951,6 +960,7 @@ static void Step() {
     step_loop = true;
     halt_loop = true;
     send_trap = true;
+    Core::System::GetInstance().InvalidateCpuInstructionCaches();
 }
 
 /// Tell the CPU if we hit a memory breakpoint.
@@ -967,6 +977,7 @@ static void Continue() {
     memory_break = false;
     step_loop = false;
     halt_loop = false;
+    Core::System::GetInstance().InvalidateCpuInstructionCaches();
 }
 
 /**
@@ -976,13 +987,17 @@ static void Continue() {
  * @param addr Address of breakpoint.
  * @param len Length of breakpoint.
  */
-static bool CommitBreakpoint(BreakpointType type, PAddr addr, u64 len) {
-    std::map<u64, Breakpoint>& p = GetBreakpointList(type);
+static bool CommitBreakpoint(BreakpointType type, VAddr addr, u64 len) {
+    BreakpointMap& p = GetBreakpointMap(type);
 
     Breakpoint breakpoint;
     breakpoint.active = true;
     breakpoint.addr = addr;
     breakpoint.len = len;
+    Memory::ReadBlock(addr, breakpoint.inst.data(), breakpoint.inst.size());
+    static constexpr std::array<u8, 4> btrap{{0xd4, 0x20, 0x7d, 0x0}};
+    Memory::WriteBlock(addr, btrap.data(), btrap.size());
+    Core::System::GetInstance().InvalidateCpuInstructionCaches();
     p.insert({addr, breakpoint});
 
     LOG_DEBUG(Debug_GDBStub, "gdb: added {} breakpoint: {:016X} bytes at {:016X}",
@@ -1016,7 +1031,7 @@ static void AddBreakpoint() {
 
     auto start_offset = command_buffer + 3;
     auto addr_pos = std::find(start_offset, command_buffer + command_length, ',');
-    PAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset));
+    VAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset));
 
     start_offset = addr_pos + 1;
     u64 len =
@@ -1065,7 +1080,7 @@ static void RemoveBreakpoint() {
 
     auto start_offset = command_buffer + 3;
     auto addr_pos = std::find(start_offset, command_buffer + command_length, ',');
-    PAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset));
+    VAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset));
 
     if (type == BreakpointType::Access) {
         // Access is made up of Read and Write types, so add both breakpoints

src/core/gdbstub/gdbstub.h

@@ -22,7 +22,7 @@ enum class BreakpointType {
 };
 
 struct BreakpointAddress {
-    PAddr address;
+    VAddr address;
     BreakpointType type;
 };
 
@@ -53,7 +53,7 @@ bool IsServerEnabled();
 bool IsConnected();
 
 /// Register module.
-void RegisterModule(std::string name, PAddr beg, PAddr end, bool add_elf_ext = true);
+void RegisterModule(std::string name, VAddr beg, VAddr end, bool add_elf_ext = true);
 
 /**
  * Signal to the gdbstub server that it should halt CPU execution.
@@ -74,7 +74,7 @@ void HandlePacket();
  * @param addr Address to search from.
  * @param type Type of breakpoint.
  */
-BreakpointAddress GetNextBreakpointFromAddress(PAddr addr, GDBStub::BreakpointType type);
+BreakpointAddress GetNextBreakpointFromAddress(VAddr addr, GDBStub::BreakpointType type);
 
 /**
  * Check if a breakpoint of the specified type exists at the given address.
@@ -82,7 +82,7 @@ BreakpointAddress GetNextBreakpointFromAddress(PAddr addr, GDBStub::BreakpointTy
  * @param addr Address of breakpoint.
  * @param type Type of breakpoint.
  */
-bool CheckBreakpoint(PAddr addr, GDBStub::BreakpointType type);
+bool CheckBreakpoint(VAddr addr, GDBStub::BreakpointType type);
 
 /// If set to true, the CPU will halt at the beginning of the next CPU loop.
 bool GetCpuHaltFlag();

src/core/hle/ipc_helpers.h

@@ -5,15 +5,18 @@
 #pragma once
 
 #include <array>
+#include <cstring>
+#include <memory>
 #include <tuple>
 #include <type_traits>
 #include <utility>
+#include "common/assert.h"
+#include "common/common_types.h"
 #include "core/hle/ipc.h"
 #include "core/hle/kernel/client_port.h"
 #include "core/hle/kernel/client_session.h"
-#include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/hle_ipc.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/server_port.h"
 
 namespace IPC {

src/core/hle/kernel/address_arbiter.cpp

@@ -9,7 +9,7 @@
 #include "common/common_types.h"
 #include "core/core.h"
 #include "core/hle/kernel/errors.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/result.h"
@@ -32,9 +32,8 @@ static ResultCode WaitForAddress(VAddr address, s64 timeout) {
 }
 
 // Gets the threads waiting on an address.
-static void GetThreadsWaitingOnAddress(std::vector<SharedPtr<Thread>>& waiting_threads,
-                                       VAddr address) {
-    auto RetrieveWaitingThreads =
+static std::vector<SharedPtr<Thread>> GetThreadsWaitingOnAddress(VAddr address) {
+    const auto RetrieveWaitingThreads =
         [](size_t core_index, std::vector<SharedPtr<Thread>>& waiting_threads, VAddr arb_addr) {
             const auto& scheduler = Core::System::GetInstance().Scheduler(core_index);
             auto& thread_list = scheduler->GetThreadList();
@@ -45,16 +44,20 @@ static void GetThreadsWaitingOnAddress(std::vector<SharedPtr<Thread>>& waiting_t
             }
         };
 
-    // Retrieve a list of all threads that are waiting for this address.
-    RetrieveWaitingThreads(0, waiting_threads, address);
-    RetrieveWaitingThreads(1, waiting_threads, address);
-    RetrieveWaitingThreads(2, waiting_threads, address);
-    RetrieveWaitingThreads(3, waiting_threads, address);
+    // Retrieve all threads that are waiting for this address.
+    std::vector<SharedPtr<Thread>> threads;
+    RetrieveWaitingThreads(0, threads, address);
+    RetrieveWaitingThreads(1, threads, address);
+    RetrieveWaitingThreads(2, threads, address);
+    RetrieveWaitingThreads(3, threads, address);
+
     // Sort them by priority, such that the highest priority ones come first.
-    std::sort(waiting_threads.begin(), waiting_threads.end(),
+    std::sort(threads.begin(), threads.end(),
               [](const SharedPtr<Thread>& lhs, const SharedPtr<Thread>& rhs) {
                   return lhs->current_priority < rhs->current_priority;
               });
+
+    return threads;
 }
 
 // Wake up num_to_wake (or all) threads in a vector.
@@ -76,9 +79,7 @@ static void WakeThreads(std::vector<SharedPtr<Thread>>& waiting_threads, s32 num
 
 // Signals an address being waited on.
 ResultCode SignalToAddress(VAddr address, s32 num_to_wake) {
-    // Get threads waiting on the address.
-    std::vector<SharedPtr<Thread>> waiting_threads;
-    GetThreadsWaitingOnAddress(waiting_threads, address);
+    std::vector<SharedPtr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address);
 
     WakeThreads(waiting_threads, num_to_wake);
     return RESULT_SUCCESS;
@@ -110,12 +111,11 @@ ResultCode ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 valu
     }
 
     // Get threads waiting on the address.
-    std::vector<SharedPtr<Thread>> waiting_threads;
-    GetThreadsWaitingOnAddress(waiting_threads, address);
+    std::vector<SharedPtr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address);
 
     // Determine the modified value depending on the waiting count.
     s32 updated_value;
-    if (waiting_threads.size() == 0) {
+    if (waiting_threads.empty()) {
         updated_value = value - 1;
     } else if (num_to_wake <= 0 || waiting_threads.size() <= static_cast<u32>(num_to_wake)) {
         updated_value = value + 1;

src/core/hle/kernel/client_port.cpp

@@ -8,7 +8,7 @@
 #include "core/hle/kernel/client_session.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/hle_ipc.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/server_port.h"
 #include "core/hle/kernel/server_session.h"
 

src/core/hle/kernel/client_port.h

@@ -6,7 +6,7 @@
 
 #include <string>
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/result.h"
 
 namespace Kernel {

src/core/hle/kernel/client_session.h

@@ -7,7 +7,7 @@
 #include <memory>
 #include <string>
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/result.h"
 
 namespace Kernel {

src/core/hle/kernel/event.cpp

@@ -5,7 +5,7 @@
 #include <algorithm>
 #include "common/assert.h"
 #include "core/hle/kernel/event.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 
 namespace Kernel {

src/core/hle/kernel/event.h

@@ -5,7 +5,7 @@
 #pragma once
 
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/wait_object.h"
 
 namespace Kernel {
@@ -31,10 +31,9 @@ public:
         return HANDLE_TYPE;
     }
 
-    ResetType reset_type; ///< Current ResetType
-
-    bool signaled;    ///< Whether the event has already been signaled
-    std::string name; ///< Name of event (optional)
+    ResetType GetResetType() const {
+        return reset_type;
+    }
 
     bool ShouldWait(Thread* thread) const override;
     void Acquire(Thread* thread) override;
@@ -47,6 +46,11 @@ public:
 private:
     Event();
     ~Event() override;
+
+    ResetType reset_type; ///< Current ResetType
+
+    bool signaled;    ///< Whether the event has already been signaled
+    std::string name; ///< Name of event (optional)
 };
 
 } // namespace Kernel

src/core/hle/kernel/handle_table.cpp

@@ -8,7 +8,6 @@
 #include "core/core.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/handle_table.h"
-#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/thread.h"
 

src/core/hle/kernel/handle_table.h

@@ -7,7 +7,7 @@
 #include <array>
 #include <cstddef>
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/result.h"
 
 namespace Kernel {

src/core/hle/kernel/hle_ipc.cpp

@@ -3,17 +3,21 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
+#include <array>
+#include <sstream>
 #include <utility>
 
 #include <boost/range/algorithm_ext/erase.hpp>
+
 #include "common/assert.h"
 #include "common/common_funcs.h"
 #include "common/common_types.h"
+#include "common/logging/log.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/hle_ipc.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/server_session.h"
 #include "core/memory.h"

src/core/hle/kernel/hle_ipc.h

@@ -13,7 +13,7 @@
 #include "common/common_types.h"
 #include "common/swap.h"
 #include "core/hle/ipc.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/server_session.h"
 #include "core/hle/kernel/thread.h"
 

src/core/hle/kernel/kernel.cpp

@@ -4,7 +4,6 @@
 
 #include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/kernel.h"
-#include "core/hle/kernel/memory.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/resource_limit.h"
 #include "core/hle/kernel/thread.h"
@@ -15,9 +14,7 @@ namespace Kernel {
 unsigned int Object::next_object_id;
 
 /// Initialize the kernel
-void Init(u32 system_mode) {
-    Kernel::MemoryInit(system_mode);
-
+void Init() {
     Kernel::ResourceLimitsInit();
     Kernel::ThreadingInit();
     Kernel::TimersInit();
@@ -37,7 +34,6 @@ void Shutdown() {
 
     Kernel::TimersShutdown();
     Kernel::ResourceLimitsShutdown();
-    Kernel::MemoryShutdown();
 }
 
 } // namespace Kernel

src/core/hle/kernel/kernel.h

@@ -4,122 +4,12 @@
 
 #pragma once
 
-#include <cstddef>
-#include <string>
-#include <utility>
-#include <boost/smart_ptr/intrusive_ptr.hpp>
-#include "common/assert.h"
 #include "common/common_types.h"
 
 namespace Kernel {
 
-using Handle = u32;
-
-enum class HandleType : u32 {
-    Unknown,
-    Event,
-    SharedMemory,
-    Thread,
-    Process,
-    AddressArbiter,
-    Timer,
-    ResourceLimit,
-    CodeSet,
-    ClientPort,
-    ServerPort,
-    ClientSession,
-    ServerSession,
-};
-
-enum class ResetType {
-    OneShot,
-    Sticky,
-    Pulse,
-};
-
-class Object : NonCopyable {
-public:
-    virtual ~Object() {}
-
-    /// Returns a unique identifier for the object. For debugging purposes only.
-    unsigned int GetObjectId() const {
-        return object_id;
-    }
-
-    virtual std::string GetTypeName() const {
-        return "[BAD KERNEL OBJECT TYPE]";
-    }
-    virtual std::string GetName() const {
-        return "[UNKNOWN KERNEL OBJECT]";
-    }
-    virtual Kernel::HandleType GetHandleType() const = 0;
-
-    /**
-     * Check if a thread can wait on the object
-     * @return True if a thread can wait on the object, otherwise false
-     */
-    bool IsWaitable() const {
-        switch (GetHandleType()) {
-        case HandleType::Event:
-        case HandleType::Thread:
-        case HandleType::Timer:
-        case HandleType::ServerPort:
-        case HandleType::ServerSession:
-            return true;
-
-        case HandleType::Unknown:
-        case HandleType::SharedMemory:
-        case HandleType::Process:
-        case HandleType::AddressArbiter:
-        case HandleType::ResourceLimit:
-        case HandleType::CodeSet:
-        case HandleType::ClientPort:
-        case HandleType::ClientSession:
-            return false;
-        }
-
-        UNREACHABLE();
-    }
-
-public:
-    static unsigned int next_object_id;
-
-private:
-    friend void intrusive_ptr_add_ref(Object*);
-    friend void intrusive_ptr_release(Object*);
-
-    unsigned int ref_count = 0;
-    unsigned int object_id = next_object_id++;
-};
-
-// Special functions used by boost::instrusive_ptr to do automatic ref-counting
-inline void intrusive_ptr_add_ref(Object* object) {
-    ++object->ref_count;
-}
-
-inline void intrusive_ptr_release(Object* object) {
-    if (--object->ref_count == 0) {
-        delete object;
-    }
-}
-
-template <typename T>
-using SharedPtr = boost::intrusive_ptr<T>;
-
-/**
- * Attempts to downcast the given Object pointer to a pointer to T.
- * @return Derived pointer to the object, or `nullptr` if `object` isn't of type T.
- */
-template <typename T>
-inline SharedPtr<T> DynamicObjectCast(SharedPtr<Object> object) {
-    if (object != nullptr && object->GetHandleType() == T::HANDLE_TYPE) {
-        return boost::static_pointer_cast<T>(std::move(object));
-    }
-    return nullptr;
-}
-
 /// Initialize the kernel with the specified system mode.
-void Init(u32 system_mode);
+void Init();
 
 /// Shutdown the kernel
 void Shutdown();

src/core/hle/kernel/memory.cpp

@@ -1,90 +0,0 @@
-// Copyright 2014 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#include <algorithm>
-#include <cinttypes>
-#include <memory>
-#include <utility>
-#include <vector>
-#include "common/assert.h"
-#include "common/common_types.h"
-#include "common/logging/log.h"
-#include "core/hle/kernel/memory.h"
-#include "core/hle/kernel/process.h"
-#include "core/hle/kernel/vm_manager.h"
-#include "core/memory.h"
-
-////////////////////////////////////////////////////////////////////////////////////////////////////
-
-namespace Kernel {
-
-MemoryRegionInfo memory_regions[3];
-
-/// Size of the APPLICATION, SYSTEM and BASE memory regions (respectively) for each system
-/// memory configuration type.
-static const u32 memory_region_sizes[8][3] = {
-    // Old 3DS layouts
-    {0x04000000, 0x02C00000, 0x01400000}, // 0
-    {/* This appears to be unused. */},   // 1
-    {0x06000000, 0x00C00000, 0x01400000}, // 2
-    {0x05000000, 0x01C00000, 0x01400000}, // 3
-    {0x04800000, 0x02400000, 0x01400000}, // 4
-    {0x02000000, 0x04C00000, 0x01400000}, // 5
-
-    // New 3DS layouts
-    {0x07C00000, 0x06400000, 0x02000000}, // 6
-    {0x0B200000, 0x02E00000, 0x02000000}, // 7
-};
-
-void MemoryInit(u32 mem_type) {
-    // TODO(yuriks): On the n3DS, all o3DS configurations (<=5) are forced to 6 instead.
-    ASSERT_MSG(mem_type <= 5, "New 3DS memory configuration aren't supported yet!");
-    ASSERT(mem_type != 1);
-
-    // The kernel allocation regions (APPLICATION, SYSTEM and BASE) are laid out in sequence, with
-    // the sizes specified in the memory_region_sizes table.
-    VAddr base = 0;
-    for (int i = 0; i < 3; ++i) {
-        memory_regions[i].base = base;
-        memory_regions[i].size = memory_region_sizes[mem_type][i];
-        memory_regions[i].used = 0;
-        memory_regions[i].linear_heap_memory = std::make_shared<std::vector<u8>>();
-        // Reserve enough space for this region of FCRAM.
-        // We do not want this block of memory to be relocated when allocating from it.
-        memory_regions[i].linear_heap_memory->reserve(memory_regions[i].size);
-
-        base += memory_regions[i].size;
-    }
-
-    // We must've allocated the entire FCRAM by the end
-    ASSERT(base == Memory::FCRAM_SIZE);
-}
-
-void MemoryShutdown() {
-    for (auto& region : memory_regions) {
-        region.base = 0;
-        region.size = 0;
-        region.used = 0;
-        region.linear_heap_memory = nullptr;
-    }
-}
-
-MemoryRegionInfo* GetMemoryRegion(MemoryRegion region) {
-    switch (region) {
-    case MemoryRegion::APPLICATION:
-        return &memory_regions[0];
-    case MemoryRegion::SYSTEM:
-        return &memory_regions[1];
-    case MemoryRegion::BASE:
-        return &memory_regions[2];
-    default:
-        UNREACHABLE();
-    }
-}
-
-void HandleSpecialMapping(VMManager& address_space, const AddressMapping& mapping) {}
-
-void MapSharedPages(VMManager& address_space) {}
-
-} // namespace Kernel

src/core/hle/kernel/memory.h

@@ -1,34 +0,0 @@
-// Copyright 2014 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <memory>
-#include <vector>
-
-#include "common/common_types.h"
-
-namespace Kernel {
-
-class VMManager;
-enum class MemoryRegion : u16;
-struct AddressMapping;
-
-struct MemoryRegionInfo {
-    u64 base; // Not an address, but offset from start of FCRAM
-    u64 size;
-    u64 used;
-
-    std::shared_ptr<std::vector<u8>> linear_heap_memory;
-};
-
-void MemoryInit(u32 mem_type);
-void MemoryShutdown();
-MemoryRegionInfo* GetMemoryRegion(MemoryRegion region);
-
-void HandleSpecialMapping(VMManager& address_space, const AddressMapping& mapping);
-void MapSharedPages(VMManager& address_space);
-
-extern MemoryRegionInfo memory_regions[3];
-} // namespace Kernel

src/core/hle/kernel/mutex.cpp

@@ -12,8 +12,8 @@
 #include "core/core.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/handle_table.h"
-#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/mutex.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/result.h"
 

src/core/hle/kernel/mutex.h

@@ -5,7 +5,7 @@
 #pragma once
 
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 
 union ResultCode;
 

src/core/hle/kernel/object.cpp

@@ -0,0 +1,35 @@
+// Copyright 2018 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/assert.h"
+#include "core/hle/kernel/object.h"
+
+namespace Kernel {
+
+Object::~Object() = default;
+
+bool Object::IsWaitable() const {
+    switch (GetHandleType()) {
+    case HandleType::Event:
+    case HandleType::Thread:
+    case HandleType::Timer:
+    case HandleType::ServerPort:
+    case HandleType::ServerSession:
+        return true;
+
+    case HandleType::Unknown:
+    case HandleType::SharedMemory:
+    case HandleType::Process:
+    case HandleType::AddressArbiter:
+    case HandleType::ResourceLimit:
+    case HandleType::CodeSet:
+    case HandleType::ClientPort:
+    case HandleType::ClientSession:
+        return false;
+    }
+
+    UNREACHABLE();
+}
+
+} // namespace Kernel

src/core/hle/kernel/object.h

@@ -0,0 +1,100 @@
+// Copyright 2018 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <string>
+#include <utility>
+
+#include <boost/smart_ptr/intrusive_ptr.hpp>
+
+#include "common/common_types.h"
+
+namespace Kernel {
+
+using Handle = u32;
+
+enum class HandleType : u32 {
+    Unknown,
+    Event,
+    SharedMemory,
+    Thread,
+    Process,
+    AddressArbiter,
+    Timer,
+    ResourceLimit,
+    CodeSet,
+    ClientPort,
+    ServerPort,
+    ClientSession,
+    ServerSession,
+};
+
+enum class ResetType {
+    OneShot,
+    Sticky,
+    Pulse,
+};
+
+class Object : NonCopyable {
+public:
+    virtual ~Object();
+
+    /// Returns a unique identifier for the object. For debugging purposes only.
+    unsigned int GetObjectId() const {
+        return object_id;
+    }
+
+    virtual std::string GetTypeName() const {
+        return "[BAD KERNEL OBJECT TYPE]";
+    }
+    virtual std::string GetName() const {
+        return "[UNKNOWN KERNEL OBJECT]";
+    }
+    virtual HandleType GetHandleType() const = 0;
+
+    /**
+     * Check if a thread can wait on the object
+     * @return True if a thread can wait on the object, otherwise false
+     */
+    bool IsWaitable() const;
+
+public:
+    static unsigned int next_object_id;
+
+private:
+    friend void intrusive_ptr_add_ref(Object*);
+    friend void intrusive_ptr_release(Object*);
+
+    unsigned int ref_count = 0;
+    unsigned int object_id = next_object_id++;
+};
+
+// Special functions used by boost::instrusive_ptr to do automatic ref-counting
+inline void intrusive_ptr_add_ref(Object* object) {
+    ++object->ref_count;
+}
+
+inline void intrusive_ptr_release(Object* object) {
+    if (--object->ref_count == 0) {
+        delete object;
+    }
+}
+
+template <typename T>
+using SharedPtr = boost::intrusive_ptr<T>;
+
+/**
+ * Attempts to downcast the given Object pointer to a pointer to T.
+ * @return Derived pointer to the object, or `nullptr` if `object` isn't of type T.
+ */
+template <typename T>
+inline SharedPtr<T> DynamicObjectCast(SharedPtr<Object> object) {
+    if (object != nullptr && object->GetHandleType() == T::HANDLE_TYPE) {
+        return boost::static_pointer_cast<T>(std::move(object));
+    }
+    return nullptr;
+}
+
+} // namespace Kernel

src/core/hle/kernel/process.cpp

@@ -8,7 +8,6 @@
 #include "common/common_funcs.h"
 #include "common/logging/log.h"
 #include "core/hle/kernel/errors.h"
-#include "core/hle/kernel/memory.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/resource_limit.h"
 #include "core/hle/kernel/thread.h"
@@ -125,14 +124,6 @@ void Process::Run(VAddr entry_point, s32 main_thread_priority, u32 stack_size) {
                         std::make_shared<std::vector<u8>>(stack_size, 0), 0, stack_size,
                         MemoryState::Mapped)
         .Unwrap();
-    misc_memory_used += stack_size;
-    memory_region->used += stack_size;
-
-    // Map special address mappings
-    MapSharedPages(vm_manager);
-    for (const auto& mapping : address_mappings) {
-        HandleSpecialMapping(vm_manager, mapping);
-    }
 
     vm_manager.LogLayout();
     status = ProcessStatus::Running;
@@ -141,37 +132,19 @@ void Process::Run(VAddr entry_point, s32 main_thread_priority, u32 stack_size) {
 }
 
 void Process::LoadModule(SharedPtr<CodeSet> module_, VAddr base_addr) {
-    memory_region = GetMemoryRegion(flags.memory_region);
-
-    auto MapSegment = [&](CodeSet::Segment& segment, VMAPermission permissions,
-                          MemoryState memory_state) {
+    const auto MapSegment = [&](CodeSet::Segment& segment, VMAPermission permissions,
+                                MemoryState memory_state) {
         auto vma = vm_manager
                        .MapMemoryBlock(segment.addr + base_addr, module_->memory, segment.offset,
                                        segment.size, memory_state)
                        .Unwrap();
         vm_manager.Reprotect(vma, permissions);
-        misc_memory_used += segment.size;
-        memory_region->used += segment.size;
     };
 
     // Map CodeSet segments
-    MapSegment(module_->code, VMAPermission::ReadExecute, MemoryState::CodeStatic);
-    MapSegment(module_->rodata, VMAPermission::Read, MemoryState::CodeMutable);
-    MapSegment(module_->data, VMAPermission::ReadWrite, MemoryState::CodeMutable);
-}
-
-VAddr Process::GetLinearHeapAreaAddress() const {
-    // Starting from system version 8.0.0 a new linear heap layout is supported to allow usage of
-    // the extra RAM in the n3DS.
-    return kernel_version < 0x22C ? Memory::LINEAR_HEAP_VADDR : Memory::NEW_LINEAR_HEAP_VADDR;
-}
-
-VAddr Process::GetLinearHeapBase() const {
-    return GetLinearHeapAreaAddress() + memory_region->base;
-}
-
-VAddr Process::GetLinearHeapLimit() const {
-    return GetLinearHeapBase() + memory_region->size;
+    MapSegment(module_->CodeSegment(), VMAPermission::ReadExecute, MemoryState::CodeStatic);
+    MapSegment(module_->RODataSegment(), VMAPermission::Read, MemoryState::CodeMutable);
+    MapSegment(module_->DataSegment(), VMAPermission::ReadWrite, MemoryState::CodeMutable);
 }
 
 ResultVal<VAddr> Process::HeapAllocate(VAddr target, u64 size, VMAPermission perms) {
@@ -206,7 +179,6 @@ ResultVal<VAddr> Process::HeapAllocate(VAddr target, u64 size, VMAPermission per
     vm_manager.Reprotect(vma, perms);
 
     heap_used = size;
-    memory_region->used += size;
 
     return MakeResult<VAddr>(heap_end - size);
 }
@@ -226,52 +198,6 @@ ResultCode Process::HeapFree(VAddr target, u32 size) {
         return result;
 
     heap_used -= size;
-    memory_region->used -= size;
-
-    return RESULT_SUCCESS;
-}
-
-ResultVal<VAddr> Process::LinearAllocate(VAddr target, u32 size, VMAPermission perms) {
-    UNIMPLEMENTED();
-    return {};
-}
-
-ResultCode Process::LinearFree(VAddr target, u32 size) {
-    auto& linheap_memory = memory_region->linear_heap_memory;
-
-    if (target < GetLinearHeapBase() || target + size > GetLinearHeapLimit() ||
-        target + size < target) {
-
-        return ERR_INVALID_ADDRESS;
-    }
-
-    if (size == 0) {
-        return RESULT_SUCCESS;
-    }
-
-    VAddr heap_end = GetLinearHeapBase() + (u32)linheap_memory->size();
-    if (target + size > heap_end) {
-        return ERR_INVALID_ADDRESS_STATE;
-    }
-
-    ResultCode result = vm_manager.UnmapRange(target, size);
-    if (result.IsError())
-        return result;
-
-    linear_heap_used -= size;
-    memory_region->used -= size;
-
-    if (target + size == heap_end) {
-        // End of linear heap has been freed, so check what's the last allocated block in it and
-        // reduce the size.
-        auto vma = vm_manager.FindVMA(target);
-        ASSERT(vma != vm_manager.vma_map.end());
-        ASSERT(vma->second.type == VMAType::Free);
-        VAddr new_end = vma->second.base;
-        if (new_end >= GetLinearHeapBase()) {
-            linheap_memory->resize(new_end - GetLinearHeapBase());
-        }
-    }
 
     return RESULT_SUCCESS;
 }

src/core/hle/kernel/process.h

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include <array>
 #include <bitset>
 #include <cstddef>
 #include <memory>
@@ -12,7 +13,7 @@
 #include <boost/container/static_vector.hpp>
 #include "common/bit_field.h"
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/vm_manager.h"
 
@@ -53,9 +54,14 @@ union ProcessFlags {
 enum class ProcessStatus { Created, Running, Exited };
 
 class ResourceLimit;
-struct MemoryRegionInfo;
 
 struct CodeSet final : public Object {
+    struct Segment {
+        size_t offset = 0;
+        VAddr addr = 0;
+        u32 size = 0;
+    };
+
     static SharedPtr<CodeSet> Create(std::string name);
 
     std::string GetTypeName() const override {
@@ -70,24 +76,38 @@ struct CodeSet final : public Object {
         return HANDLE_TYPE;
     }
 
-    /// Name of the process
-    std::string name;
+    Segment& CodeSegment() {
+        return segments[0];
+    }
+
+    const Segment& CodeSegment() const {
+        return segments[0];
+    }
+
+    Segment& RODataSegment() {
+        return segments[1];
+    }
+
+    const Segment& RODataSegment() const {
+        return segments[1];
+    }
+
+    Segment& DataSegment() {
+        return segments[2];
+    }
+
+    const Segment& DataSegment() const {
+        return segments[2];
+    }
 
     std::shared_ptr<std::vector<u8>> memory;
 
-    struct Segment {
-        size_t offset = 0;
-        VAddr addr = 0;
-        u32 size = 0;
-    };
-
-    Segment segments[3];
-    Segment& code = segments[0];
-    Segment& rodata = segments[1];
-    Segment& data = segments[2];
-
+    std::array<Segment, 3> segments;
     VAddr entrypoint;
 
+    /// Name of the process
+    std::string name;
+
 private:
     CodeSet();
     ~CodeSet() override;
@@ -163,12 +183,11 @@ public:
     // This makes deallocation and reallocation of holes fast and keeps process memory contiguous
     // in the emulator address space, allowing Memory::GetPointer to be reasonably safe.
     std::shared_ptr<std::vector<u8>> heap_memory;
+
     // The left/right bounds of the address space covered by heap_memory.
-    VAddr heap_start = 0, heap_end = 0;
-
-    u64 heap_used = 0, linear_heap_used = 0, misc_memory_used = 0;
-
-    MemoryRegionInfo* memory_region = nullptr;
+    VAddr heap_start = 0;
+    VAddr heap_end = 0;
+    u64 heap_used = 0;
 
     /// The Thread Local Storage area is allocated as processes create threads,
     /// each TLS area is 0x200 bytes, so one page (0x1000) is split up in 8 parts, and each part
@@ -179,16 +198,9 @@ public:
 
     std::string name;
 
-    VAddr GetLinearHeapAreaAddress() const;
-    VAddr GetLinearHeapBase() const;
-    VAddr GetLinearHeapLimit() const;
-
     ResultVal<VAddr> HeapAllocate(VAddr target, u64 size, VMAPermission perms);
     ResultCode HeapFree(VAddr target, u32 size);
 
-    ResultVal<VAddr> LinearAllocate(VAddr target, u32 size, VMAPermission perms);
-    ResultCode LinearFree(VAddr target, u32 size);
-
     ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size);
 
     ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, u64 size);

src/core/hle/kernel/resource_limit.h

@@ -5,7 +5,7 @@
 #pragma once
 
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 
 namespace Kernel {
 

src/core/hle/kernel/scheduler.h

@@ -8,6 +8,7 @@
 #include <vector>
 #include "common/common_types.h"
 #include "common/thread_queue_list.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 
 class ARM_Interface;

src/core/hle/kernel/server_port.cpp

@@ -6,7 +6,7 @@
 #include "common/assert.h"
 #include "core/hle/kernel/client_port.h"
 #include "core/hle/kernel/errors.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/server_port.h"
 #include "core/hle/kernel/server_session.h"
 #include "core/hle/kernel/thread.h"

src/core/hle/kernel/server_port.h

@@ -9,7 +9,7 @@
 #include <tuple>
 #include <vector>
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/wait_object.h"
 
 namespace Kernel {

src/core/hle/kernel/server_session.h

@@ -9,7 +9,7 @@
 #include <vector>
 
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/wait_object.h"
 #include "core/hle/result.h"
 

src/core/hle/kernel/session.h

@@ -4,7 +4,7 @@
 
 #pragma once
 
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 
 namespace Kernel {
 

src/core/hle/kernel/shared_memory.cpp

@@ -8,7 +8,6 @@
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "core/hle/kernel/errors.h"
-#include "core/hle/kernel/memory.h"
 #include "core/hle/kernel/shared_memory.h"
 #include "core/memory.h"
 
@@ -30,35 +29,17 @@ SharedPtr<SharedMemory> SharedMemory::Create(SharedPtr<Process> owner_process, u
     shared_memory->other_permissions = other_permissions;
 
     if (address == 0) {
-        // We need to allocate a block from the Linear Heap ourselves.
-        // We'll manually allocate some memory from the linear heap in the specified region.
-        MemoryRegionInfo* memory_region = GetMemoryRegion(region);
-        auto& linheap_memory = memory_region->linear_heap_memory;
-
-        ASSERT_MSG(linheap_memory->size() + size <= memory_region->size,
-                   "Not enough space in region to allocate shared memory!");
-
-        shared_memory->backing_block = linheap_memory;
-        shared_memory->backing_block_offset = linheap_memory->size();
-        // Allocate some memory from the end of the linear heap for this region.
-        linheap_memory->insert(linheap_memory->end(), size, 0);
-        memory_region->used += size;
-
-        shared_memory->linear_heap_phys_address =
-            Memory::FCRAM_PADDR + memory_region->base +
-            static_cast<PAddr>(shared_memory->backing_block_offset);
-
-        // Increase the amount of used linear heap memory for the owner process.
-        if (shared_memory->owner_process != nullptr) {
-            shared_memory->owner_process->linear_heap_used += size;
-        }
+        shared_memory->backing_block = std::make_shared<std::vector<u8>>(size);
+        shared_memory->backing_block_offset = 0;
 
         // Refresh the address mappings for the current process.
         if (Core::CurrentProcess() != nullptr) {
-            Core::CurrentProcess()->vm_manager.RefreshMemoryBlockMappings(linheap_memory.get());
+            Core::CurrentProcess()->vm_manager.RefreshMemoryBlockMappings(
+                shared_memory->backing_block.get());
         }
     } else {
         auto& vm_manager = shared_memory->owner_process->vm_manager;
+
         // The memory is already available and mapped in the owner process.
         auto vma = vm_manager.FindVMA(address);
         ASSERT_MSG(vma != vm_manager.vma_map.end(), "Invalid memory address");
@@ -74,6 +55,7 @@ SharedPtr<SharedMemory> SharedMemory::Create(SharedPtr<Process> owner_process, u
     }
 
     shared_memory->base_address = address;
+
     return shared_memory;
 }
 
@@ -124,11 +106,6 @@ ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermi
 
     VAddr target_address = address;
 
-    if (base_address == 0 && target_address == 0) {
-        // Calculate the address at which to map the memory block.
-        target_address = Memory::PhysicalToVirtualAddress(linear_heap_phys_address).value();
-    }
-
     // Map the memory block into the target process
     auto result = target_process->vm_manager.MapMemoryBlock(
         target_address, backing_block, backing_block_offset, size, MemoryState::Shared);

src/core/hle/kernel/shared_memory.h

@@ -9,7 +9,7 @@
 #include <vector>
 
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/result.h"
 
@@ -111,9 +111,6 @@ public:
     SharedPtr<Process> owner_process;
     /// Address of shared memory block in the owner process if specified.
     VAddr base_address;
-    /// Physical address of the shared memory block in the linear heap if no address was specified
-    /// during creation.
-    PAddr linear_heap_phys_address;
     /// Backing memory for this shared memory block.
     std::shared_ptr<std::vector<u8>> backing_block;
     /// Offset into the backing block for this shared memory.

src/core/hle/kernel/svc.cpp

@@ -267,7 +267,7 @@ 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);
 
-    auto& vm_manager = Core::CurrentProcess()->vm_manager;
+    const auto& vm_manager = Core::CurrentProcess()->vm_manager;
 
     switch (static_cast<GetInfoType>(info_id)) {
     case GetInfoType::AllowedCpuIdBitmask:

src/core/hle/kernel/thread.cpp

@@ -20,8 +20,7 @@
 #include "core/core_timing_util.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/handle_table.h"
-#include "core/hle/kernel/kernel.h"
-#include "core/hle/kernel/memory.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/result.h"
@@ -81,8 +80,8 @@ void Thread::Stop() {
     wait_objects.clear();
 
     // Mark the TLS slot in the thread's page as free.
-    u64 tls_page = (tls_address - Memory::TLS_AREA_VADDR) / Memory::PAGE_SIZE;
-    u64 tls_slot =
+    const u64 tls_page = (tls_address - Memory::TLS_AREA_VADDR) / Memory::PAGE_SIZE;
+    const u64 tls_slot =
         ((tls_address - Memory::TLS_AREA_VADDR) % Memory::PAGE_SIZE) / Memory::TLS_ENTRY_SIZE;
     Core::CurrentProcess()->tls_slots[tls_page].reset(tls_slot);
 }
@@ -252,13 +251,14 @@ void Thread::ResumeFromWait() {
  * slot: The index of the first free slot in the indicated page.
  * alloc_needed: Whether there's a need to allocate a new TLS page (All pages are full).
  */
-std::tuple<u32, u32, bool> GetFreeThreadLocalSlot(std::vector<std::bitset<8>>& tls_slots) {
+static std::tuple<std::size_t, std::size_t, bool> GetFreeThreadLocalSlot(
+    const std::vector<std::bitset<8>>& tls_slots) {
     // Iterate over all the allocated pages, and try to find one where not all slots are used.
-    for (unsigned page = 0; page < tls_slots.size(); ++page) {
+    for (std::size_t page = 0; page < tls_slots.size(); ++page) {
         const auto& page_tls_slots = tls_slots[page];
         if (!page_tls_slots.all()) {
             // We found a page with at least one free slot, find which slot it is
-            for (unsigned slot = 0; slot < page_tls_slots.size(); ++slot) {
+            for (std::size_t slot = 0; slot < page_tls_slots.size(); ++slot) {
                 if (!page_tls_slots.test(slot)) {
                     return std::make_tuple(page, slot, false);
                 }
@@ -333,42 +333,22 @@ ResultVal<SharedPtr<Thread>> Thread::Create(std::string name, VAddr entry_point,
 
     // Find the next available TLS index, and mark it as used
     auto& tls_slots = owner_process->tls_slots;
-    bool needs_allocation = true;
-    u32 available_page; // Which allocated page has free space
-    u32 available_slot; // Which slot within the page is free
-
-    std::tie(available_page, available_slot, needs_allocation) = GetFreeThreadLocalSlot(tls_slots);
 
+    auto [available_page, available_slot, needs_allocation] = GetFreeThreadLocalSlot(tls_slots);
     if (needs_allocation) {
-        // There are no already-allocated pages with free slots, lets allocate a new one.
-        // TLS pages are allocated from the BASE region in the linear heap.
-        MemoryRegionInfo* memory_region = GetMemoryRegion(MemoryRegion::BASE);
-        auto& linheap_memory = memory_region->linear_heap_memory;
-
-        if (linheap_memory->size() + Memory::PAGE_SIZE > memory_region->size) {
-            LOG_ERROR(Kernel_SVC,
-                      "Not enough space in region to allocate a new TLS page for thread");
-            return ERR_OUT_OF_MEMORY;
-        }
-
-        size_t offset = linheap_memory->size();
-
-        // Allocate some memory from the end of the linear heap for this region.
-        linheap_memory->insert(linheap_memory->end(), Memory::PAGE_SIZE, 0);
-        memory_region->used += Memory::PAGE_SIZE;
-        owner_process->linear_heap_used += Memory::PAGE_SIZE;
-
         tls_slots.emplace_back(0); // The page is completely available at the start
-        available_page = static_cast<u32>(tls_slots.size() - 1);
+        available_page = tls_slots.size() - 1;
         available_slot = 0; // Use the first slot in the new page
 
-        auto& vm_manager = owner_process->vm_manager;
-        vm_manager.RefreshMemoryBlockMappings(linheap_memory.get());
+        // Allocate some memory from the end of the linear heap for this region.
+        const size_t offset = thread->tls_memory->size();
+        thread->tls_memory->insert(thread->tls_memory->end(), Memory::PAGE_SIZE, 0);
+
+        auto& vm_manager = owner_process->vm_manager;
+        vm_manager.RefreshMemoryBlockMappings(thread->tls_memory.get());
 
-        // Map the page to the current process' address space.
-        // TODO(Subv): Find the correct MemoryState for this region.
         vm_manager.MapMemoryBlock(Memory::TLS_AREA_VADDR + available_page * Memory::PAGE_SIZE,
-                                  linheap_memory, offset, Memory::PAGE_SIZE,
+                                  thread->tls_memory, 0, Memory::PAGE_SIZE,
                                   MemoryState::ThreadLocal);
     }
 

src/core/hle/kernel/thread.h

@@ -11,7 +11,7 @@
 
 #include "common/common_types.h"
 #include "core/arm/arm_interface.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/wait_object.h"
 #include "core/hle/result.h"
 
@@ -265,6 +265,8 @@ public:
 private:
     Thread();
     ~Thread() override;
+
+    std::shared_ptr<std::vector<u8>> tls_memory = std::make_shared<std::vector<u8>>();
 };
 
 /**

src/core/hle/kernel/timer.cpp

@@ -8,7 +8,7 @@
 #include "core/core_timing.h"
 #include "core/core_timing_util.h"
 #include "core/hle/kernel/handle_table.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/timer.h"
 

src/core/hle/kernel/timer.h

@@ -5,7 +5,7 @@
 #pragma once
 
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/wait_object.h"
 
 namespace Kernel {

src/core/hle/kernel/vm_manager.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include <iterator>
 #include <utility>
 #include "common/assert.h"
@@ -175,9 +176,9 @@ VMManager::VMAIter VMManager::Unmap(VMAIter vma_handle) {
 
 ResultCode VMManager::UnmapRange(VAddr target, u64 size) {
     CASCADE_RESULT(VMAIter vma, CarveVMARange(target, size));
-    VAddr target_end = target + size;
+    const VAddr target_end = target + size;
 
-    VMAIter end = vma_map.end();
+    const VMAIter end = vma_map.end();
     // The comparison against the end of the range must be done using addresses since VMAs can be
     // merged during this process, causing invalidation of the iterators.
     while (vma != end && vma->second.base < target_end) {
@@ -207,9 +208,9 @@ VMManager::VMAHandle VMManager::Reprotect(VMAHandle vma_handle, VMAPermission ne
 
 ResultCode VMManager::ReprotectRange(VAddr target, u64 size, VMAPermission new_perms) {
     CASCADE_RESULT(VMAIter vma, CarveVMARange(target, size));
-    VAddr target_end = target + size;
+    const VAddr target_end = target + size;
 
-    VMAIter end = vma_map.end();
+    const VMAIter end = vma_map.end();
     // The comparison against the end of the range must be done using addresses since VMAs can be
     // merged during this process, causing invalidation of the iterators.
     while (vma != end && vma->second.base < target_end) {
@@ -258,14 +259,14 @@ ResultVal<VMManager::VMAIter> VMManager::CarveVMA(VAddr base, u64 size) {
         return ERR_INVALID_ADDRESS;
     }
 
-    VirtualMemoryArea& vma = vma_handle->second;
+    const VirtualMemoryArea& vma = vma_handle->second;
     if (vma.type != VMAType::Free) {
         // Region is already allocated
         return ERR_INVALID_ADDRESS_STATE;
     }
 
-    u64 start_in_vma = base - vma.base;
-    u64 end_in_vma = start_in_vma + size;
+    const VAddr start_in_vma = base - vma.base;
+    const VAddr end_in_vma = start_in_vma + size;
 
     if (end_in_vma > vma.size) {
         // Requested allocation doesn't fit inside VMA
@@ -288,17 +289,16 @@ ResultVal<VMManager::VMAIter> VMManager::CarveVMARange(VAddr target, u64 size) {
     ASSERT_MSG((size & Memory::PAGE_MASK) == 0, "non-page aligned size: 0x{:016X}", size);
     ASSERT_MSG((target & Memory::PAGE_MASK) == 0, "non-page aligned base: 0x{:016X}", target);
 
-    VAddr target_end = target + size;
+    const VAddr target_end = target + size;
     ASSERT(target_end >= target);
     ASSERT(target_end <= MAX_ADDRESS);
     ASSERT(size > 0);
 
     VMAIter begin_vma = StripIterConstness(FindVMA(target));
-    VMAIter i_end = vma_map.lower_bound(target_end);
-    for (auto i = begin_vma; i != i_end; ++i) {
-        if (i->second.type == VMAType::Free) {
-            return ERR_INVALID_ADDRESS_STATE;
-        }
+    const VMAIter i_end = vma_map.lower_bound(target_end);
+    if (std::any_of(begin_vma, i_end,
+                    [](const auto& entry) { return entry.second.type == VMAType::Free; })) {
+        return ERR_INVALID_ADDRESS_STATE;
     }
 
     if (target != begin_vma->second.base) {
@@ -346,7 +346,7 @@ VMManager::VMAIter VMManager::SplitVMA(VMAIter vma_handle, u64 offset_in_vma) {
 }
 
 VMManager::VMAIter VMManager::MergeAdjacent(VMAIter iter) {
-    VMAIter next_vma = std::next(iter);
+    const VMAIter next_vma = std::next(iter);
     if (next_vma != vma_map.end() && iter->second.CanBeMergedWith(next_vma->second)) {
         iter->second.size += next_vma->second.size;
         vma_map.erase(next_vma);
@@ -382,22 +382,22 @@ void VMManager::UpdatePageTableForVMA(const VirtualMemoryArea& vma) {
     }
 }
 
-u64 VMManager::GetTotalMemoryUsage() {
+u64 VMManager::GetTotalMemoryUsage() const {
     LOG_WARNING(Kernel, "(STUBBED) called");
     return 0xF8000000;
 }
 
-u64 VMManager::GetTotalHeapUsage() {
+u64 VMManager::GetTotalHeapUsage() const {
     LOG_WARNING(Kernel, "(STUBBED) called");
     return 0x0;
 }
 
-VAddr VMManager::GetAddressSpaceBaseAddr() {
+VAddr VMManager::GetAddressSpaceBaseAddr() const {
     LOG_WARNING(Kernel, "(STUBBED) called");
     return 0x8000000;
 }
 
-u64 VMManager::GetAddressSpaceSize() {
+u64 VMManager::GetAddressSpaceSize() const {
     LOG_WARNING(Kernel, "(STUBBED) called");
     return MAX_ADDRESS;
 }

src/core/hle/kernel/vm_manager.h

@@ -190,16 +190,16 @@ public:
     void LogLayout() const;
 
     /// Gets the total memory usage, used by svcGetInfo
-    u64 GetTotalMemoryUsage();
+    u64 GetTotalMemoryUsage() const;
 
     /// Gets the total heap usage, used by svcGetInfo
-    u64 GetTotalHeapUsage();
+    u64 GetTotalHeapUsage() const;
 
     /// Gets the total address space base address, used by svcGetInfo
-    VAddr GetAddressSpaceBaseAddr();
+    VAddr GetAddressSpaceBaseAddr() const;
 
     /// Gets the total address space address size, used by svcGetInfo
-    u64 GetAddressSpaceSize();
+    u64 GetAddressSpaceSize() const;
 
     /// Each VMManager has its own page table, which is set as the main one when the owning process
     /// is scheduled.

src/core/hle/kernel/wait_object.cpp

@@ -5,8 +5,7 @@
 #include <algorithm>
 #include "common/assert.h"
 #include "common/logging/log.h"
-#include "core/hle/kernel/kernel.h"
-#include "core/hle/kernel/memory.h"
+#include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/timer.h"

src/core/hle/kernel/wait_object.h

@@ -7,7 +7,7 @@
 #include <vector>
 #include <boost/smart_ptr/intrusive_ptr.hpp>
 #include "common/common_types.h"
-#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/object.h"
 
 namespace Kernel {
 

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

@@ -10,6 +10,7 @@
 #include "core/hle/service/acc/acc_su.h"
 #include "core/hle/service/acc/acc_u0.h"
 #include "core/hle/service/acc/acc_u1.h"
+#include "core/settings.h"
 
 namespace Service::Account {
 
@@ -31,13 +32,14 @@ struct ProfileBase {
 };
 static_assert(sizeof(ProfileBase) == 0x38, "ProfileBase structure has incorrect size");
 
+// TODO(ogniK): Generate a real user id based on username, md5(username) maybe?
 static constexpr u128 DEFAULT_USER_ID{1ull, 0ull};
 
 class IProfile final : public ServiceFramework<IProfile> {
 public:
     explicit IProfile(u128 user_id) : ServiceFramework("IProfile"), user_id(user_id) {
         static const FunctionInfo functions[] = {
-            {0, nullptr, "Get"},
+            {0, &IProfile::Get, "Get"},
             {1, &IProfile::GetBase, "GetBase"},
             {10, nullptr, "GetImageSize"},
             {11, nullptr, "LoadImage"},
@@ -46,14 +48,36 @@ public:
     }
 
 private:
+    void Get(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_ACC, "(STUBBED) called");
+        ProfileBase profile_base{};
+        profile_base.user_id = user_id;
+        if (Settings::values.username.size() > profile_base.username.size()) {
+            std::copy_n(Settings::values.username.begin(), profile_base.username.size(),
+                        profile_base.username.begin());
+        } else {
+            std::copy(Settings::values.username.begin(), Settings::values.username.end(),
+                      profile_base.username.begin());
+        }
+
+        IPC::ResponseBuilder rb{ctx, 16};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw(profile_base);
+    }
+
     void GetBase(Kernel::HLERequestContext& ctx) {
         LOG_WARNING(Service_ACC, "(STUBBED) called");
 
         // TODO(Subv): Retrieve this information from somewhere.
         ProfileBase profile_base{};
         profile_base.user_id = user_id;
-        profile_base.username = {'y', 'u', 'z', 'u'};
-
+        if (Settings::values.username.size() > profile_base.username.size()) {
+            std::copy_n(Settings::values.username.begin(), profile_base.username.size(),
+                        profile_base.username.begin());
+        } else {
+            std::copy(Settings::values.username.begin(), Settings::values.username.end(),
+                      profile_base.username.begin());
+        }
         IPC::ResponseBuilder rb{ctx, 16};
         rb.Push(RESULT_SUCCESS);
         rb.PushRaw(profile_base);

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

@@ -652,7 +652,8 @@ void IApplicationFunctions::GetDesiredLanguage(Kernel::HLERequestContext& ctx) {
     // TODO(bunnei): This should be configurable
     IPC::ResponseBuilder rb{ctx, 4};
     rb.Push(RESULT_SUCCESS);
-    rb.Push(static_cast<u64>(Service::Set::LanguageCode::EN_US));
+    rb.Push(
+        static_cast<u64>(Service::Set::GetLanguageCodeFromIndex(Settings::values.language_index)));
     LOG_DEBUG(Service_AM, "called");
 }
 

src/core/hle/service/apm/apm.cpp

@@ -13,6 +13,7 @@ void InstallInterfaces(SM::ServiceManager& service_manager) {
     auto module_ = std::make_shared<Module>();
     std::make_shared<APM>(module_, "apm")->InstallAsService(service_manager);
     std::make_shared<APM>(module_, "apm:p")->InstallAsService(service_manager);
+    std::make_shared<APM_Sys>()->InstallAsService(service_manager);
 }
 
 } // namespace Service::APM

src/core/hle/service/apm/interface.cpp

@@ -74,6 +74,31 @@ void APM::OpenSession(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
     rb.PushIpcInterface<ISession>();
+
+    LOG_DEBUG(Service_APM, "called");
+}
+
+APM_Sys::APM_Sys() : ServiceFramework{"apm:sys"} {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "RequestPerformanceMode"},
+        {1, &APM_Sys::GetPerformanceEvent, "GetPerformanceEvent"},
+        {2, nullptr, "GetThrottlingState"},
+        {3, nullptr, "GetLastThrottlingState"},
+        {4, nullptr, "ClearLastThrottlingState"},
+        {5, nullptr, "LoadAndApplySettings"},
+    };
+    // clang-format on
+
+    RegisterHandlers(functions);
+}
+
+void APM_Sys::GetPerformanceEvent(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushIpcInterface<ISession>();
+
+    LOG_DEBUG(Service_APM, "called");
 }
 
 } // namespace Service::APM

src/core/hle/service/apm/interface.h

@@ -19,4 +19,12 @@ private:
     std::shared_ptr<Module> apm;
 };
 
+class APM_Sys final : public ServiceFramework<APM_Sys> {
+public:
+    explicit APM_Sys();
+
+private:
+    void GetPerformanceEvent(Kernel::HLERequestContext& ctx);
+};
+
 } // namespace Service::APM

src/core/hle/service/arp/arp.cpp

@@ -0,0 +1,75 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <memory>
+
+#include "common/logging/log.h"
+#include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/hle_ipc.h"
+#include "core/hle/service/arp/arp.h"
+#include "core/hle/service/service.h"
+#include "core/hle/service/sm/sm.h"
+
+namespace Service::ARP {
+
+class ARP_R final : public ServiceFramework<ARP_R> {
+public:
+    explicit ARP_R() : ServiceFramework{"arp:r"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, nullptr, "GetApplicationLaunchProperty"},
+            {1, nullptr, "GetApplicationLaunchPropertyWithApplicationId"},
+            {2, nullptr, "GetApplicationControlProperty"},
+            {3, nullptr, "GetApplicationControlPropertyWithApplicationId"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+class IRegistrar final : public ServiceFramework<IRegistrar> {
+public:
+    explicit IRegistrar() : ServiceFramework{"IRegistrar"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, nullptr, "Issue"},
+            {1, nullptr, "SetApplicationLaunchProperty"},
+            {2, nullptr, "SetApplicationControlProperty"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+class ARP_W final : public ServiceFramework<ARP_W> {
+public:
+    explicit ARP_W() : ServiceFramework{"arp:w"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &ARP_W::AcquireRegistrar, "AcquireRegistrar"},
+            {1, nullptr, "DeleteProperties"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+
+private:
+    void AcquireRegistrar(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IRegistrar>();
+
+        LOG_DEBUG(Service_ARP, "called");
+    }
+};
+
+void InstallInterfaces(SM::ServiceManager& sm) {
+    std::make_shared<ARP_R>()->InstallAsService(sm);
+    std::make_shared<ARP_W>()->InstallAsService(sm);
+}
+
+} // namespace Service::ARP

src/core/hle/service/arp/arp.h

@@ -0,0 +1,16 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+namespace Service::SM {
+class ServiceManager;
+}
+
+namespace Service::ARP {
+
+/// Registers all ARP services with the specified service manager.
+void InstallInterfaces(SM::ServiceManager& sm);
+
+} // namespace Service::ARP

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

@@ -2,8 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#pragma once
-
 #include "core/hle/service/audio/audin_a.h"
 
 namespace Service::Audio {

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

@@ -2,8 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#pragma once
-
 #include "core/hle/service/audio/audout_a.h"
 
 namespace Service::Audio {

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

@@ -4,6 +4,8 @@
 
 #include <array>
 #include <vector>
+
+#include "audio_core/codec.h"
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
@@ -48,7 +50,7 @@ public:
         buffer_event = Kernel::Event::Create(Kernel::ResetType::Sticky, "IAudioOutBufferReleased");
 
         stream = audio_core.OpenStream(audio_params.sample_rate, audio_params.channel_count,
-                                       [=]() { buffer_event->Signal(); });
+                                       "IAudioOut", [=]() { buffer_event->Signal(); });
     }
 
 private:
@@ -111,10 +113,10 @@ private:
         std::memcpy(&audio_buffer, input_buffer.data(), sizeof(AudioBuffer));
         const u64 tag{rp.Pop<u64>()};
 
-        std::vector<u8> data(audio_buffer.buffer_size);
-        Memory::ReadBlock(audio_buffer.buffer, data.data(), data.size());
+        std::vector<s16> samples(audio_buffer.buffer_size / sizeof(s16));
+        Memory::ReadBlock(audio_buffer.buffer, samples.data(), audio_buffer.buffer_size);
 
-        if (!audio_core.QueueBuffer(stream, tag, std::move(data))) {
+        if (!audio_core.QueueBuffer(stream, tag, std::move(samples))) {
             IPC::ResponseBuilder rb{ctx, 2};
             rb.Push(ResultCode(ErrorModule::Audio, ErrCodes::BufferCountExceeded));
         }
@@ -200,7 +202,7 @@ void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {
     rb.Push(RESULT_SUCCESS);
     rb.Push<u32>(DefaultSampleRate);
     rb.Push<u32>(params.channel_count);
-    rb.Push<u32>(static_cast<u32>(PcmFormat::Int16));
+    rb.Push<u32>(static_cast<u32>(AudioCore::Codec::PcmFormat::Int16));
     rb.Push<u32>(static_cast<u32>(AudioState::Stopped));
     rb.PushIpcInterface<Audio::IAudioOut>(audio_out_interface);
 }