276 files changed, 13526 insertions, 4583 deletions

diff --git a/src/audio_core/audio_renderer.cpp b/src/audio_core/audio_renderer.cpp
index 6f0ff953a..23e5d3f10 100644
--- a/src/audio_core/audio_renderer.cpp
+++ b/src/audio_core/audio_renderer.cpp
@@ -30,7 +30,7 @@ public:
         return info;
     }
 
-    VoiceInfo& Info() {
+    VoiceInfo& GetInfo() {
         return info;
     }
 
@@ -51,9 +51,30 @@ private:
     VoiceInfo info{};
 };
 
+class AudioRenderer::EffectState {
+public:
+    const EffectOutStatus& GetOutStatus() const {
+        return out_status;
+    }
+
+    const EffectInStatus& GetInfo() const {
+        return info;
+    }
+
+    EffectInStatus& GetInfo() {
+        return info;
+    }
+
+    void UpdateState();
+
+private:
+    EffectOutStatus out_status{};
+    EffectInStatus info{};
+};
 AudioRenderer::AudioRenderer(AudioRendererParameter params,
                              Kernel::SharedPtr<Kernel::Event> buffer_event)
-    : worker_params{params}, buffer_event{buffer_event}, voices(params.voice_count) {
+    : worker_params{params}, buffer_event{buffer_event}, voices(params.voice_count),
+      effects(params.effect_count) {
 
     audio_out = std::make_unique<AudioCore::AudioOut>();
     stream = audio_out->OpenStream(STREAM_SAMPLE_RATE, STREAM_NUM_CHANNELS, "AudioRenderer",
@@ -96,11 +117,29 @@ std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_
                 memory_pool_count * sizeof(MemoryPoolInfo));
 
     // Copy VoiceInfo structs
-    std::size_t offset{sizeof(UpdateDataHeader) + config.behavior_size + config.memory_pools_size +
-                       config.voice_resource_size};
+    std::size_t voice_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);
+        std::memcpy(&voice.GetInfo(), input_params.data() + voice_offset, sizeof(VoiceInfo));
+        voice_offset += sizeof(VoiceInfo);
+    }
+
+    std::size_t effect_offset{sizeof(UpdateDataHeader) + config.behavior_size +
+                              config.memory_pools_size + config.voice_resource_size +
+                              config.voices_size};
+    for (auto& effect : effects) {
+        std::memcpy(&effect.GetInfo(), input_params.data() + effect_offset, sizeof(EffectInStatus));
+        effect_offset += sizeof(EffectInStatus);
+    }
+
+    // Update memory pool state
+    std::vector<MemoryPoolEntry> memory_pool(memory_pool_count);
+    for (std::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;
+        }
     }
 
     // Update voices
@@ -114,14 +153,8 @@ std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_
         }
     }
 
-    // Update memory pool state
-    std::vector<MemoryPoolEntry> memory_pool(memory_pool_count);
-    for (std::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;
-        }
+    for (auto& effect : effects) {
+        effect.UpdateState();
     }
 
     // Release previous buffers and queue next ones for playback
@@ -144,6 +177,14 @@ std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_
         voice_out_status_offset += sizeof(VoiceOutStatus);
     }
 
+    std::size_t effect_out_status_offset{
+        sizeof(UpdateDataHeader) + response_data.memory_pools_size + response_data.voices_size +
+        response_data.voice_resource_size};
+    for (const auto& effect : effects) {
+        std::memcpy(output_params.data() + effect_out_status_offset, &effect.GetOutStatus(),
+                    sizeof(EffectOutStatus));
+        effect_out_status_offset += sizeof(EffectOutStatus);
+    }
     return output_params;
 }
 
@@ -244,11 +285,29 @@ void AudioRenderer::VoiceState::RefreshBuffer() {
         break;
     }
 
-    samples = Interpolate(interp_state, std::move(samples), Info().sample_rate, STREAM_SAMPLE_RATE);
+    samples =
+        Interpolate(interp_state, std::move(samples), GetInfo().sample_rate, STREAM_SAMPLE_RATE);
 
     is_refresh_pending = false;
 }
 
+void AudioRenderer::EffectState::UpdateState() {
+    if (info.is_new) {
+        out_status.state = EffectStatus::New;
+    } else {
+        if (info.type == Effect::Aux) {
+            ASSERT_MSG(Memory::Read32(info.aux_info.return_buffer_info) == 0,
+                       "Aux buffers tried to update");
+            ASSERT_MSG(Memory::Read32(info.aux_info.send_buffer_info) == 0,
+                       "Aux buffers tried to update");
+            ASSERT_MSG(Memory::Read32(info.aux_info.return_buffer_base) == 0,
+                       "Aux buffers tried to update");
+            ASSERT_MSG(Memory::Read32(info.aux_info.send_buffer_base) == 0,
+                       "Aux buffers tried to update");
+        }
+    }
+}
+
 static constexpr s16 ClampToS16(s32 value) {
     return static_cast<s16>(std::clamp(value, -32768, 32767));
 }
diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h
index dfef89e1d..046417da3 100644
--- a/src/audio_core/audio_renderer.h
+++ b/src/audio_core/audio_renderer.h
@@ -28,6 +28,16 @@ enum class PlayState : u8 {
     Paused = 2,
 };
 
+enum class Effect : u8 {
+    None = 0,
+    Aux = 2,
+};
+
+enum class EffectStatus : u8 {
+    None = 0,
+    New = 1,
+};
+
 struct AudioRendererParameter {
     u32_le sample_rate;
     u32_le sample_count;
@@ -128,6 +138,43 @@ struct VoiceOutStatus {
 };
 static_assert(sizeof(VoiceOutStatus) == 0x10, "VoiceOutStatus has wrong size");
 
+struct AuxInfo {
+    std::array<u8, 24> input_mix_buffers;
+    std::array<u8, 24> output_mix_buffers;
+    u32_le mix_buffer_count;
+    u32_le sample_rate; // Stored in the aux buffer currently
+    u32_le sampe_count;
+    u64_le send_buffer_info;
+    u64_le send_buffer_base;
+
+    u64_le return_buffer_info;
+    u64_le return_buffer_base;
+};
+static_assert(sizeof(AuxInfo) == 0x60, "AuxInfo is an invalid size");
+
+struct EffectInStatus {
+    Effect type;
+    u8 is_new;
+    u8 is_enabled;
+    INSERT_PADDING_BYTES(1);
+    u32_le mix_id;
+    u64_le buffer_base;
+    u64_le buffer_sz;
+    s32_le priority;
+    INSERT_PADDING_BYTES(4);
+    union {
+        std::array<u8, 0xa0> raw;
+        AuxInfo aux_info;
+    };
+};
+static_assert(sizeof(EffectInStatus) == 0xc0, "EffectInStatus is an invalid size");
+
+struct EffectOutStatus {
+    EffectStatus state;
+    INSERT_PADDING_BYTES(0xf);
+};
+static_assert(sizeof(EffectOutStatus) == 0x10, "EffectOutStatus is an invalid size");
+
 struct UpdateDataHeader {
     UpdateDataHeader() {}
 
@@ -173,11 +220,13 @@ public:
     Stream::State GetStreamState() const;
 
 private:
+    class EffectState;
     class VoiceState;
 
     AudioRendererParameter worker_params;
     Kernel::SharedPtr<Kernel::Event> buffer_event;
     std::vector<VoiceState> voices;
+    std::vector<EffectState> effects;
     std::unique_ptr<AudioOut> audio_out;
     AudioCore::StreamPtr stream;
 };
diff --git a/src/audio_core/cubeb_sink.cpp b/src/audio_core/cubeb_sink.cpp
index 392039688..d31a1c844 100644
--- a/src/audio_core/cubeb_sink.cpp
+++ b/src/audio_core/cubeb_sink.cpp
@@ -121,7 +121,8 @@ CubebSink::CubebSink(std::string target_device_name) {
             const auto collection_end{collection.device + collection.count};
             const auto device{
                 std::find_if(collection.device, collection_end, [&](const cubeb_device_info& info) {
-                    return target_device_name == info.friendly_name;
+                    return info.friendly_name != nullptr &&
+                           target_device_name == info.friendly_name;
                 })};
             if (device != collection_end) {
                 output_device = device->devid;
diff --git a/src/audio_core/time_stretch.cpp b/src/audio_core/time_stretch.cpp
index d72d67994..2fe0b3aef 100644
--- a/src/audio_core/time_stretch.cpp
+++ b/src/audio_core/time_stretch.cpp
@@ -10,8 +10,7 @@
 
 namespace AudioCore {
 
-TimeStretcher::TimeStretcher(u32 sample_rate, u32 channel_count)
-    : m_sample_rate(sample_rate), m_channel_count(channel_count) {
+TimeStretcher::TimeStretcher(u32 sample_rate, u32 channel_count) : m_sample_rate{sample_rate} {
     m_sound_touch.setChannels(channel_count);
     m_sound_touch.setSampleRate(sample_rate);
     m_sound_touch.setPitch(1.0);
@@ -33,10 +32,10 @@ std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out,
     // We were given actual_samples number of samples, and num_samples were requested from us.
     double current_ratio = static_cast<double>(num_in) / static_cast<double>(num_out);
 
-    const double max_latency = 1.0; // seconds
+    const double max_latency = 0.25; // seconds
     const double max_backlog = m_sample_rate * max_latency;
     const double backlog_fullness = m_sound_touch.numSamples() / max_backlog;
-    if (backlog_fullness > 5.0) {
+    if (backlog_fullness > 4.0) {
         // Too many samples in backlog: Don't push anymore on
         num_in = 0;
     }
@@ -50,7 +49,7 @@ std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out,
 
     // This low-pass filter smoothes out variance in the calculated stretch ratio.
     // The time-scale determines how responsive this filter is.
-    constexpr double lpf_time_scale = 2.0; // seconds
+    constexpr double lpf_time_scale = 0.712; // seconds
     const double lpf_gain = 1.0 - std::exp(-time_delta / lpf_time_scale);
     m_stretch_ratio += lpf_gain * (current_ratio - m_stretch_ratio);
 
diff --git a/src/audio_core/time_stretch.h b/src/audio_core/time_stretch.h
index decd760f1..bb2270b96 100644
--- a/src/audio_core/time_stretch.h
+++ b/src/audio_core/time_stretch.h
@@ -27,7 +27,6 @@ public:
 
 private:
     u32 m_sample_rate;
-    u32 m_channel_count;
     soundtouch::SoundTouch m_sound_touch;
     double m_stretch_ratio = 1.0;
 };
diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt
index 8985e4367..eccd8f64a 100644
--- a/src/common/CMakeLists.txt
+++ b/src/common/CMakeLists.txt
@@ -29,7 +29,7 @@ if ($ENV{CI})
       if (BUILD_VERSION)
         # This leaves a trailing space on the last word, but we actually want that
         # because of how it's styled in the title bar.
-        set(BUILD_FULLNAME "${REPO_NAME} #${BUILD_VERSION} ")
+        set(BUILD_FULLNAME "${REPO_NAME} ${BUILD_VERSION} ")
       else()
         set(BUILD_FULLNAME "")
       endif()
@@ -64,8 +64,6 @@ add_library(common STATIC
     logging/text_formatter.cpp
     logging/text_formatter.h
     math_util.h
-    memory_util.cpp
-    memory_util.h
     microprofile.cpp
     microprofile.h
     microprofileui.h
diff --git a/src/common/alignment.h b/src/common/alignment.h
index 225770fab..d94a2291f 100644
--- a/src/common/alignment.h
+++ b/src/common/alignment.h
@@ -19,4 +19,16 @@ constexpr T AlignDown(T value, std::size_t size) {
     return static_cast<T>(value - value % size);
 }
 
+template <typename T>
+constexpr bool Is4KBAligned(T value) {
+    static_assert(std::is_unsigned_v<T>, "T must be an unsigned value.");
+    return (value & 0xFFF) == 0;
+}
+
+template <typename T>
+constexpr bool IsWordAligned(T value) {
+    static_assert(std::is_unsigned_v<T>, "T must be an unsigned value.");
+    return (value & 0b11) == 0;
+}
+
 } // namespace Common
diff --git a/src/common/file_util.cpp b/src/common/file_util.cpp
index 548463787..b52492da6 100644
--- a/src/common/file_util.cpp
+++ b/src/common/file_util.cpp
@@ -15,21 +15,24 @@
 #ifdef _WIN32
 #include <windows.h>
 // windows.h needs to be included before other windows headers
-#include <commdlg.h> // for GetSaveFileName
-#include <direct.h>  // getcwd
+#include <direct.h> // getcwd
 #include <io.h>
 #include <shellapi.h>
 #include <shlobj.h> // for SHGetFolderPath
 #include <tchar.h>
 #include "common/string_util.h"
 
-// 64 bit offsets for windows
+#ifdef _MSC_VER
+// 64 bit offsets for MSVC
 #define fseeko _fseeki64
 #define ftello _ftelli64
-#define atoll _atoi64
+#define fileno _fileno
+#endif
+
+// 64 bit offsets for MSVC and MinGW. MinGW also needs this for using _wstat64
 #define stat _stat64
 #define fstat _fstat64
-#define fileno _fileno
+
 #else
 #ifdef __APPLE__
 #include <sys/param.h>
diff --git a/src/common/hex_util.cpp b/src/common/hex_util.cpp
index 589ae5cbf..5b63f9e81 100644
--- a/src/common/hex_util.cpp
+++ b/src/common/hex_util.cpp
@@ -18,6 +18,25 @@ u8 ToHexNibble(char c1) {
     return 0;
 }
 
+std::vector<u8> HexStringToVector(std::string_view str, bool little_endian) {
+    std::vector<u8> out(str.size() / 2);
+    if (little_endian) {
+        for (std::size_t i = str.size() - 2; i <= str.size(); i -= 2)
+            out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]);
+    } else {
+        for (std::size_t i = 0; i < str.size(); i += 2)
+            out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]);
+    }
+    return out;
+}
+
+std::string HexVectorToString(const std::vector<u8>& vector, bool upper) {
+    std::string out;
+    for (u8 c : vector)
+        out += fmt::format(upper ? "{:02X}" : "{:02x}", c);
+    return out;
+}
+
 std::array<u8, 16> operator""_array16(const char* str, std::size_t len) {
     if (len != 32) {
         LOG_ERROR(Common,
diff --git a/src/common/hex_util.h b/src/common/hex_util.h
index 863a5ccd9..68f003cb6 100644
--- a/src/common/hex_util.h
+++ b/src/common/hex_util.h
@@ -7,6 +7,7 @@
 #include <array>
 #include <cstddef>
 #include <string>
+#include <vector>
 #include <fmt/format.h>
 #include "common/common_types.h"
 
@@ -14,6 +15,8 @@ namespace Common {
 
 u8 ToHexNibble(char c1);
 
+std::vector<u8> HexStringToVector(std::string_view str, bool little_endian);
+
 template <std::size_t Size, bool le = false>
 std::array<u8, Size> HexStringToArray(std::string_view str) {
     std::array<u8, Size> out{};
@@ -27,6 +30,8 @@ std::array<u8, Size> HexStringToArray(std::string_view str) {
     return out;
 }
 
+std::string HexVectorToString(const std::vector<u8>& vector, bool upper = true);
+
 template <std::size_t Size>
 std::string HexArrayToString(std::array<u8, Size> array, bool upper = true) {
     std::string out;
diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp
index 31ad72f38..5753b871a 100644
--- a/src/common/logging/backend.cpp
+++ b/src/common/logging/backend.cpp
@@ -203,6 +203,7 @@ void DebuggerBackend::Write(const Entry& entry) {
     SUB(Service, NFP)                                                                              \
     SUB(Service, NIFM)                                                                             \
     SUB(Service, NIM)                                                                              \
+    SUB(Service, NPNS)                                                                             \
     SUB(Service, NS)                                                                               \
     SUB(Service, NVDRV)                                                                            \
     SUB(Service, PCIE)                                                                             \
@@ -211,10 +212,12 @@ void DebuggerBackend::Write(const Entry& entry) {
     SUB(Service, PM)                                                                               \
     SUB(Service, PREPO)                                                                            \
     SUB(Service, PSC)                                                                              \
+    SUB(Service, PSM)                                                                              \
     SUB(Service, SET)                                                                              \
     SUB(Service, SM)                                                                               \
     SUB(Service, SPL)                                                                              \
     SUB(Service, SSL)                                                                              \
+    SUB(Service, TCAP)                                                                             \
     SUB(Service, Time)                                                                             \
     SUB(Service, USB)                                                                              \
     SUB(Service, VI)                                                                               \
diff --git a/src/common/logging/log.h b/src/common/logging/log.h
index abbd056ee..d4ec31ec3 100644
--- a/src/common/logging/log.h
+++ b/src/common/logging/log.h
@@ -83,6 +83,7 @@ enum class Class : ClassType {
     Service_NFP,       ///< The NFP service
     Service_NIFM,      ///< The NIFM (Network interface) service
     Service_NIM,       ///< The NIM service
+    Service_NPNS,      ///< The NPNS service
     Service_NS,        ///< The NS services
     Service_NVDRV,     ///< The NVDRV (Nvidia driver) service
     Service_PCIE,      ///< The PCIe service
@@ -91,10 +92,12 @@ enum class Class : ClassType {
     Service_PM,        ///< The PM service
     Service_PREPO,     ///< The PREPO (Play report) service
     Service_PSC,       ///< The PSC service
+    Service_PSM,       ///< The PSM service
     Service_SET,       ///< The SET (Settings) service
     Service_SM,        ///< The SM (Service manager) service
     Service_SPL,       ///< The SPL service
     Service_SSL,       ///< The SSL service
+    Service_TCAP,      ///< The TCAP service.
     Service_Time,      ///< The time service
     Service_USB,       ///< The USB (Universal Serial Bus) service
     Service_VI,        ///< The VI (Video interface) service
diff --git a/src/common/memory_util.cpp b/src/common/memory_util.cpp
deleted file mode 100644
index 9736fb12a..000000000
--- a/src/common/memory_util.cpp
+++ /dev/null
@@ -1,177 +0,0 @@
-// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#include "common/logging/log.h"
-#include "common/memory_util.h"
-
-#ifdef _WIN32
-#include <windows.h>
-// Windows.h needs to be included before psapi.h
-#include <psapi.h>
-#include "common/common_funcs.h"
-#include "common/string_util.h"
-#else
-#include <cstdlib>
-#include <sys/mman.h>
-#endif
-
-#if !defined(_WIN32) && defined(ARCHITECTURE_x86_64) && !defined(MAP_32BIT)
-#include <unistd.h>
-#define PAGE_MASK (getpagesize() - 1)
-#define round_page(x) ((((unsigned long)(x)) + PAGE_MASK) & ~(PAGE_MASK))
-#endif
-
-// This is purposely not a full wrapper for virtualalloc/mmap, but it
-// provides exactly the primitive operations that Dolphin needs.
-
-void* AllocateExecutableMemory(std::size_t size, bool low) {
-#if defined(_WIN32)
-    void* ptr = VirtualAlloc(nullptr, size, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
-#else
-    static char* map_hint = nullptr;
-#if defined(ARCHITECTURE_x86_64) && !defined(MAP_32BIT)
-    // This OS has no flag to enforce allocation below the 4 GB boundary,
-    // but if we hint that we want a low address it is very likely we will
-    // get one.
-    // An older version of this code used MAP_FIXED, but that has the side
-    // effect of discarding already mapped pages that happen to be in the
-    // requested virtual memory range (such as the emulated RAM, sometimes).
-    if (low && (!map_hint))
-        map_hint = (char*)round_page(512 * 1024 * 1024); /* 0.5 GB rounded up to the next page */
-#endif
-    void* ptr = mmap(map_hint, size, PROT_READ | PROT_WRITE | PROT_EXEC,
-                     MAP_ANON | MAP_PRIVATE
-#if defined(ARCHITECTURE_x86_64) && defined(MAP_32BIT)
-                         | (low ? MAP_32BIT : 0)
-#endif
-                         ,
-                     -1, 0);
-#endif /* defined(_WIN32) */
-
-#ifdef _WIN32
-    if (ptr == nullptr) {
-#else
-    if (ptr == MAP_FAILED) {
-        ptr = nullptr;
-#endif
-        LOG_ERROR(Common_Memory, "Failed to allocate executable memory");
-    }
-#if !defined(_WIN32) && defined(ARCHITECTURE_x86_64) && !defined(MAP_32BIT)
-    else {
-        if (low) {
-            map_hint += size;
-            map_hint = (char*)round_page(map_hint); /* round up to the next page */
-        }
-    }
-#endif
-
-#if EMU_ARCH_BITS == 64
-    if ((u64)ptr >= 0x80000000 && low == true)
-        LOG_ERROR(Common_Memory, "Executable memory ended up above 2GB!");
-#endif
-
-    return ptr;
-}
-
-void* AllocateMemoryPages(std::size_t size) {
-#ifdef _WIN32
-    void* ptr = VirtualAlloc(nullptr, size, MEM_COMMIT, PAGE_READWRITE);
-#else
-    void* ptr = mmap(nullptr, size, PROT_READ | PROT_WRITE, MAP_ANON | MAP_PRIVATE, -1, 0);
-
-    if (ptr == MAP_FAILED)
-        ptr = nullptr;
-#endif
-
-    if (ptr == nullptr)
-        LOG_ERROR(Common_Memory, "Failed to allocate raw memory");
-
-    return ptr;
-}
-
-void* AllocateAlignedMemory(std::size_t size, std::size_t alignment) {
-#ifdef _WIN32
-    void* ptr = _aligned_malloc(size, alignment);
-#else
-    void* ptr = nullptr;
-#ifdef ANDROID
-    ptr = memalign(alignment, size);
-#else
-    if (posix_memalign(&ptr, alignment, size) != 0)
-        LOG_ERROR(Common_Memory, "Failed to allocate aligned memory");
-#endif
-#endif
-
-    if (ptr == nullptr)
-        LOG_ERROR(Common_Memory, "Failed to allocate aligned memory");
-
-    return ptr;
-}
-
-void FreeMemoryPages(void* ptr, std::size_t size) {
-    if (ptr) {
-#ifdef _WIN32
-        if (!VirtualFree(ptr, 0, MEM_RELEASE))
-            LOG_ERROR(Common_Memory, "FreeMemoryPages failed!\n{}", GetLastErrorMsg());
-#else
-        munmap(ptr, size);
-#endif
-    }
-}
-
-void FreeAlignedMemory(void* ptr) {
-    if (ptr) {
-#ifdef _WIN32
-        _aligned_free(ptr);
-#else
-        free(ptr);
-#endif
-    }
-}
-
-void WriteProtectMemory(void* ptr, std::size_t size, bool allowExecute) {
-#ifdef _WIN32
-    DWORD oldValue;
-    if (!VirtualProtect(ptr, size, allowExecute ? PAGE_EXECUTE_READ : PAGE_READONLY, &oldValue))
-        LOG_ERROR(Common_Memory, "WriteProtectMemory failed!\n{}", GetLastErrorMsg());
-#else
-    mprotect(ptr, size, allowExecute ? (PROT_READ | PROT_EXEC) : PROT_READ);
-#endif
-}
-
-void UnWriteProtectMemory(void* ptr, std::size_t size, bool allowExecute) {
-#ifdef _WIN32
-    DWORD oldValue;
-    if (!VirtualProtect(ptr, size, allowExecute ? PAGE_EXECUTE_READWRITE : PAGE_READWRITE,
-                        &oldValue))
-        LOG_ERROR(Common_Memory, "UnWriteProtectMemory failed!\n{}", GetLastErrorMsg());
-#else
-    mprotect(ptr, size,
-             allowExecute ? (PROT_READ | PROT_WRITE | PROT_EXEC) : PROT_WRITE | PROT_READ);
-#endif
-}
-
-std::string MemUsage() {
-#ifdef _WIN32
-#pragma comment(lib, "psapi")
-    DWORD processID = GetCurrentProcessId();
-    HANDLE hProcess;
-    PROCESS_MEMORY_COUNTERS pmc;
-    std::string Ret;
-
-    // Print information about the memory usage of the process.
-
-    hProcess = OpenProcess(PROCESS_QUERY_INFORMATION | PROCESS_VM_READ, FALSE, processID);
-    if (nullptr == hProcess)
-        return "MemUsage Error";
-
-    if (GetProcessMemoryInfo(hProcess, &pmc, sizeof(pmc)))
-        Ret = fmt::format("{} K", Common::ThousandSeparate(pmc.WorkingSetSize / 1024, 7));
-
-    CloseHandle(hProcess);
-    return Ret;
-#else
-    return "";
-#endif
-}
diff --git a/src/common/memory_util.h b/src/common/memory_util.h
deleted file mode 100644
index aad071979..000000000
--- a/src/common/memory_util.h
+++ /dev/null
@@ -1,21 +0,0 @@
-// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <cstddef>
-#include <string>
-
-void* AllocateExecutableMemory(std::size_t size, bool low = true);
-void* AllocateMemoryPages(std::size_t size);
-void FreeMemoryPages(void* ptr, std::size_t size);
-void* AllocateAlignedMemory(std::size_t size, std::size_t alignment);
-void FreeAlignedMemory(void* ptr);
-void WriteProtectMemory(void* ptr, std::size_t size, bool executable = false);
-void UnWriteProtectMemory(void* ptr, std::size_t size, bool allowExecute = false);
-std::string MemUsage();
-
-inline int GetPageSize() {
-    return 4096;
-}
diff --git a/src/common/param_package.cpp b/src/common/param_package.cpp
index 9526ca0c6..b916b4866 100644
--- a/src/common/param_package.cpp
+++ b/src/common/param_package.cpp
@@ -20,7 +20,15 @@ constexpr char KEY_VALUE_SEPARATOR_ESCAPE[] = "$0";
 constexpr char PARAM_SEPARATOR_ESCAPE[] = "$1";
 constexpr char ESCAPE_CHARACTER_ESCAPE[] = "$2";
 
+/// A placeholder for empty param packages to avoid empty strings
+/// (they may be recognized as "not set" by some frontend libraries like qt)
+constexpr char EMPTY_PLACEHOLDER[] = "[empty]";
+
 ParamPackage::ParamPackage(const std::string& serialized) {
+    if (serialized == EMPTY_PLACEHOLDER) {
+        return;
+    }
+
     std::vector<std::string> pairs;
     Common::SplitString(serialized, PARAM_SEPARATOR, pairs);
 
@@ -46,7 +54,7 @@ ParamPackage::ParamPackage(std::initializer_list<DataType::value_type> list) : d
 
 std::string ParamPackage::Serialize() const {
     if (data.empty())
-        return "";
+        return EMPTY_PLACEHOLDER;
 
     std::string result;
 
@@ -120,4 +128,12 @@ bool ParamPackage::Has(const std::string& key) const {
     return data.find(key) != data.end();
 }
 
+void ParamPackage::Erase(const std::string& key) {
+    data.erase(key);
+}
+
+void ParamPackage::Clear() {
+    data.clear();
+}
+
 } // namespace Common
diff --git a/src/common/param_package.h b/src/common/param_package.h
index 7842cd4ef..6a0a9b656 100644
--- a/src/common/param_package.h
+++ b/src/common/param_package.h
@@ -32,6 +32,8 @@ public:
     void Set(const std::string& key, int value);
     void Set(const std::string& key, float value);
     bool Has(const std::string& key) const;
+    void Erase(const std::string& key);
+    void Clear();
 
 private:
     DataType data;
diff --git a/src/common/telemetry.h b/src/common/telemetry.h
index 8d6ab986b..854a73fae 100644
--- a/src/common/telemetry.h
+++ b/src/common/telemetry.h
@@ -153,6 +153,7 @@ struct VisitorInterface : NonCopyable {
 
     /// Completion method, called once all fields have been visited
     virtual void Complete() = 0;
+    virtual bool SubmitTestcase() = 0;
 };
 
 /**
@@ -178,6 +179,9 @@ struct NullVisitor : public VisitorInterface {
     void Visit(const Field<std::chrono::microseconds>& /*field*/) override {}
 
     void Complete() override {}
+    bool SubmitTestcase() override {
+        return false;
+    }
 };
 
 /// Appends build-specific information to the given FieldCollection,
diff --git a/src/common/web_result.h b/src/common/web_result.h
index 969926674..8bfa2141d 100644
--- a/src/common/web_result.h
+++ b/src/common/web_result.h
@@ -5,6 +5,7 @@
 #pragma once
 
 #include <string>
+#include "common/common_types.h"
 
 namespace Common {
 struct WebResult {
diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt
index e4a676e91..64fdf38cd 100644
--- a/src/core/CMakeLists.txt
+++ b/src/core/CMakeLists.txt
@@ -18,6 +18,8 @@ add_library(core STATIC
     crypto/encryption_layer.h
     crypto/key_manager.cpp
     crypto/key_manager.h
+    crypto/partition_data_manager.cpp
+    crypto/partition_data_manager.h
     crypto/ctr_encryption_layer.cpp
     crypto/ctr_encryption_layer.h
     crypto/xts_encryption_layer.cpp
@@ -70,6 +72,7 @@ add_library(core STATIC
     file_sys/vfs_real.cpp
     file_sys/vfs_real.h
     file_sys/vfs_static.h
+    file_sys/vfs_types.h
     file_sys/vfs_vector.cpp
     file_sys/vfs_vector.h
     file_sys/xts_archive.cpp
@@ -153,6 +156,8 @@ add_library(core STATIC
     hle/service/am/omm.h
     hle/service/am/spsm.cpp
     hle/service/am/spsm.h
+    hle/service/am/tcap.cpp
+    hle/service/am/tcap.h
     hle/service/aoc/aoc_u.cpp
     hle/service/aoc/aoc_u.h
     hle/service/apm/apm.cpp
@@ -233,6 +238,24 @@ add_library(core STATIC
     hle/service/hid/irs.h
     hle/service/hid/xcd.cpp
     hle/service/hid/xcd.h
+    hle/service/hid/controllers/controller_base.cpp
+    hle/service/hid/controllers/controller_base.h
+    hle/service/hid/controllers/debug_pad.cpp
+    hle/service/hid/controllers/debug_pad.h
+    hle/service/hid/controllers/gesture.cpp
+    hle/service/hid/controllers/gesture.h
+    hle/service/hid/controllers/keyboard.cpp
+    hle/service/hid/controllers/keyboard.h
+    hle/service/hid/controllers/mouse.cpp
+    hle/service/hid/controllers/mouse.h
+    hle/service/hid/controllers/npad.cpp
+    hle/service/hid/controllers/npad.h
+    hle/service/hid/controllers/stubbed.cpp
+    hle/service/hid/controllers/stubbed.h
+    hle/service/hid/controllers/touchscreen.cpp
+    hle/service/hid/controllers/touchscreen.h
+    hle/service/hid/controllers/xpad.cpp
+    hle/service/hid/controllers/xpad.h
     hle/service/lbl/lbl.cpp
     hle/service/lbl/lbl.h
     hle/service/ldn/ldn.cpp
@@ -259,6 +282,8 @@ add_library(core STATIC
     hle/service/nifm/nifm.h
     hle/service/nim/nim.cpp
     hle/service/nim/nim.h
+    hle/service/npns/npns.cpp
+    hle/service/npns/npns.h
     hle/service/ns/ns.cpp
     hle/service/ns/ns.h
     hle/service/ns/pl_u.cpp
@@ -306,6 +331,8 @@ add_library(core STATIC
     hle/service/prepo/prepo.h
     hle/service/psc/psc.cpp
     hle/service/psc/psc.h
+    hle/service/ptm/psm.cpp
+    hle/service/ptm/psm.h
     hle/service/service.cpp
     hle/service/service.h
     hle/service/set/set.cpp
@@ -397,8 +424,8 @@ 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 mbedtls opus unicorn open_source_archives)
 if (ENABLE_WEB_SERVICE)
-    add_definitions(-DENABLE_WEB_SERVICE)
-    target_link_libraries(core PUBLIC json-headers web_service)
+    target_compile_definitions(core PRIVATE -DENABLE_WEB_SERVICE)
+    target_link_libraries(core PRIVATE web_service)
 endif()
 
 if (ARCHITECTURE_x86_64)
diff --git a/src/core/arm/dynarmic/arm_dynarmic.cpp b/src/core/arm/dynarmic/arm_dynarmic.cpp
index 7e978cf7a..4d2491870 100644
--- a/src/core/arm/dynarmic/arm_dynarmic.cpp
+++ b/src/core/arm/dynarmic/arm_dynarmic.cpp
@@ -129,7 +129,7 @@ public:
 };
 
 std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const {
-    auto& current_process = Core::CurrentProcess();
+    auto* current_process = Core::CurrentProcess();
     auto** const page_table = current_process->VMManager().page_table.pointers.data();
 
     Dynarmic::A64::UserConfig config;
@@ -144,7 +144,7 @@ std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const {
 
     // Multi-process state
     config.processor_id = core_index;
-    config.global_monitor = &exclusive_monitor->monitor;
+    config.global_monitor = &exclusive_monitor.monitor;
 
     // System registers
     config.tpidrro_el0 = &cb->tpidrro_el0;
@@ -171,10 +171,9 @@ void ARM_Dynarmic::Step() {
     cb->InterpreterFallback(jit->GetPC(), 1);
 }
 
-ARM_Dynarmic::ARM_Dynarmic(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
-                           std::size_t core_index)
+ARM_Dynarmic::ARM_Dynarmic(ExclusiveMonitor& exclusive_monitor, std::size_t core_index)
     : cb(std::make_unique<ARM_Dynarmic_Callbacks>(*this)), core_index{core_index},
-      exclusive_monitor{std::dynamic_pointer_cast<DynarmicExclusiveMonitor>(exclusive_monitor)} {
+      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {
     ThreadContext ctx{};
     inner_unicorn.SaveContext(ctx);
     PageTableChanged();
diff --git a/src/core/arm/dynarmic/arm_dynarmic.h b/src/core/arm/dynarmic/arm_dynarmic.h
index 4ee92ee27..512bf8ce9 100644
--- a/src/core/arm/dynarmic/arm_dynarmic.h
+++ b/src/core/arm/dynarmic/arm_dynarmic.h
@@ -23,7 +23,7 @@ class DynarmicExclusiveMonitor;
 
 class ARM_Dynarmic final : public ARM_Interface {
 public:
-    ARM_Dynarmic(std::shared_ptr<ExclusiveMonitor> exclusive_monitor, std::size_t core_index);
+    ARM_Dynarmic(ExclusiveMonitor& exclusive_monitor, std::size_t core_index);
     ~ARM_Dynarmic();
 
     void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
@@ -62,7 +62,7 @@ private:
     ARM_Unicorn inner_unicorn;
 
     std::size_t core_index;
-    std::shared_ptr<DynarmicExclusiveMonitor> exclusive_monitor;
+    DynarmicExclusiveMonitor& exclusive_monitor;
 
     Memory::PageTable* current_page_table = nullptr;
 };
diff --git a/src/core/core.cpp b/src/core/core.cpp
index b6acfb3e4..6d5b5a2d0 100644
--- a/src/core/core.cpp
+++ b/src/core/core.cpp
@@ -71,9 +71,9 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
 }
 
 /// Runs a CPU core while the system is powered on
-void RunCpuCore(std::shared_ptr<Cpu> cpu_state) {
+void RunCpuCore(Cpu& cpu_state) {
     while (Core::System::GetInstance().IsPoweredOn()) {
-        cpu_state->RunLoop(true);
+        cpu_state.RunLoop(true);
     }
 }
 } // Anonymous namespace
@@ -95,7 +95,7 @@ struct System::Impl {
         status = ResultStatus::Success;
 
         // Update thread_to_cpu in case Core 0 is run from a different host thread
-        thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
+        thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0].get();
 
         if (GDBStub::IsServerEnabled()) {
             GDBStub::HandlePacket();
@@ -136,18 +136,19 @@ struct System::Impl {
         if (virtual_filesystem == nullptr)
             virtual_filesystem = std::make_shared<FileSys::RealVfsFilesystem>();
 
-        kernel.MakeCurrentProcess(Kernel::Process::Create(kernel, "main"));
+        auto main_process = Kernel::Process::Create(kernel, "main");
+        kernel.MakeCurrentProcess(main_process.get());
 
-        cpu_barrier = std::make_shared<CpuBarrier>();
+        cpu_barrier = std::make_unique<CpuBarrier>();
         cpu_exclusive_monitor = Cpu::MakeExclusiveMonitor(cpu_cores.size());
         for (std::size_t index = 0; index < cpu_cores.size(); ++index) {
-            cpu_cores[index] = std::make_shared<Cpu>(cpu_exclusive_monitor, cpu_barrier, index);
+            cpu_cores[index] = std::make_unique<Cpu>(*cpu_exclusive_monitor, *cpu_barrier, index);
         }
 
         telemetry_session = std::make_unique<Core::TelemetrySession>();
         service_manager = std::make_shared<Service::SM::ServiceManager>();
 
-        Service::Init(service_manager, virtual_filesystem);
+        Service::Init(service_manager, *virtual_filesystem);
         GDBStub::Init();
 
         renderer = VideoCore::CreateRenderer(emu_window);
@@ -159,12 +160,12 @@ struct System::Impl {
 
         // 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];
+        thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0].get();
         if (Settings::values.use_multi_core) {
             for (std::size_t index = 0; index < cpu_core_threads.size(); ++index) {
                 cpu_core_threads[index] =
-                    std::make_unique<std::thread>(RunCpuCore, cpu_cores[index + 1]);
-                thread_to_cpu[cpu_core_threads[index]->get_id()] = cpu_cores[index + 1];
+                    std::make_unique<std::thread>(RunCpuCore, std::ref(*cpu_cores[index + 1]));
+                thread_to_cpu[cpu_core_threads[index]->get_id()] = cpu_cores[index + 1].get();
             }
         }
 
@@ -184,7 +185,7 @@ struct System::Impl {
             LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath);
             return ResultStatus::ErrorGetLoader;
         }
-        std::pair<boost::optional<u32>, Loader::ResultStatus> system_mode =
+        std::pair<std::optional<u32>, Loader::ResultStatus> system_mode =
             app_loader->LoadKernelSystemMode();
 
         if (system_mode.second != Loader::ResultStatus::Success) {
@@ -244,6 +245,7 @@ struct System::Impl {
         for (auto& cpu_core : cpu_cores) {
             cpu_core.reset();
         }
+        cpu_exclusive_monitor.reset();
         cpu_barrier.reset();
 
         // Shutdown kernel and core timing
@@ -281,9 +283,9 @@ struct System::Impl {
     std::unique_ptr<VideoCore::RendererBase> renderer;
     std::unique_ptr<Tegra::GPU> gpu_core;
     std::shared_ptr<Tegra::DebugContext> debug_context;
-    std::shared_ptr<ExclusiveMonitor> cpu_exclusive_monitor;
-    std::shared_ptr<CpuBarrier> cpu_barrier;
-    std::array<std::shared_ptr<Cpu>, NUM_CPU_CORES> cpu_cores;
+    std::unique_ptr<ExclusiveMonitor> cpu_exclusive_monitor;
+    std::unique_ptr<CpuBarrier> cpu_barrier;
+    std::array<std::unique_ptr<Cpu>, NUM_CPU_CORES> cpu_cores;
     std::array<std::unique_ptr<std::thread>, NUM_CPU_CORES - 1> cpu_core_threads;
     std::size_t active_core{}; ///< Active core, only used in single thread mode
 
@@ -297,7 +299,7 @@ struct System::Impl {
     std::string status_details = "";
 
     /// Map of guest threads to CPU cores
-    std::map<std::thread::id, std::shared_ptr<Cpu>> thread_to_cpu;
+    std::map<std::thread::id, Cpu*> thread_to_cpu;
 
     Core::PerfStats perf_stats;
     Core::FrameLimiter frame_limiter;
@@ -310,6 +312,10 @@ Cpu& System::CurrentCpuCore() {
     return impl->CurrentCpuCore();
 }
 
+const Cpu& System::CurrentCpuCore() const {
+    return impl->CurrentCpuCore();
+}
+
 System::ResultStatus System::RunLoop(bool tight_loop) {
     return impl->RunLoop(tight_loop);
 }
@@ -340,7 +346,11 @@ PerfStatsResults System::GetAndResetPerfStats() {
     return impl->GetAndResetPerfStats();
 }
 
-Core::TelemetrySession& System::TelemetrySession() const {
+TelemetrySession& System::TelemetrySession() {
+    return *impl->telemetry_session;
+}
+
+const TelemetrySession& System::TelemetrySession() const {
     return *impl->telemetry_session;
 }
 
@@ -348,30 +358,44 @@ ARM_Interface& System::CurrentArmInterface() {
     return CurrentCpuCore().ArmInterface();
 }
 
-std::size_t System::CurrentCoreIndex() {
+const ARM_Interface& System::CurrentArmInterface() const {
+    return CurrentCpuCore().ArmInterface();
+}
+
+std::size_t System::CurrentCoreIndex() const {
     return CurrentCpuCore().CoreIndex();
 }
 
 Kernel::Scheduler& System::CurrentScheduler() {
-    return *CurrentCpuCore().Scheduler();
+    return CurrentCpuCore().Scheduler();
 }
 
-const std::shared_ptr<Kernel::Scheduler>& System::Scheduler(std::size_t core_index) {
-    ASSERT(core_index < NUM_CPU_CORES);
-    return impl->cpu_cores[core_index]->Scheduler();
+const Kernel::Scheduler& System::CurrentScheduler() const {
+    return CurrentCpuCore().Scheduler();
+}
+
+Kernel::Scheduler& System::Scheduler(std::size_t core_index) {
+    return CpuCore(core_index).Scheduler();
 }
 
-Kernel::SharedPtr<Kernel::Process>& System::CurrentProcess() {
+const Kernel::Scheduler& System::Scheduler(std::size_t core_index) const {
+    return CpuCore(core_index).Scheduler();
+}
+
+Kernel::Process* System::CurrentProcess() {
     return impl->kernel.CurrentProcess();
 }
 
-const Kernel::SharedPtr<Kernel::Process>& System::CurrentProcess() const {
+const Kernel::Process* System::CurrentProcess() const {
     return impl->kernel.CurrentProcess();
 }
 
 ARM_Interface& System::ArmInterface(std::size_t core_index) {
-    ASSERT(core_index < NUM_CPU_CORES);
-    return impl->cpu_cores[core_index]->ArmInterface();
+    return CpuCore(core_index).ArmInterface();
+}
+
+const ARM_Interface& System::ArmInterface(std::size_t core_index) const {
+    return CpuCore(core_index).ArmInterface();
 }
 
 Cpu& System::CpuCore(std::size_t core_index) {
@@ -379,10 +403,19 @@ Cpu& System::CpuCore(std::size_t core_index) {
     return *impl->cpu_cores[core_index];
 }
 
+const Cpu& System::CpuCore(std::size_t core_index) const {
+    ASSERT(core_index < NUM_CPU_CORES);
+    return *impl->cpu_cores[core_index];
+}
+
 ExclusiveMonitor& System::Monitor() {
     return *impl->cpu_exclusive_monitor;
 }
 
+const ExclusiveMonitor& System::Monitor() const {
+    return *impl->cpu_exclusive_monitor;
+}
+
 Tegra::GPU& System::GPU() {
     return *impl->gpu_core;
 }
diff --git a/src/core/core.h b/src/core/core.h
index f9a3e97e3..cfacceb81 100644
--- a/src/core/core.h
+++ b/src/core/core.h
@@ -129,11 +129,11 @@ public:
      */
     bool IsPoweredOn() const;
 
-    /**
-     * Returns a reference to the telemetry session for this emulation session.
-     * @returns Reference to the telemetry session.
-     */
-    Core::TelemetrySession& TelemetrySession() const;
+    /// Gets a reference to the telemetry session for this emulation session.
+    Core::TelemetrySession& TelemetrySession();
+
+    /// Gets a reference to the telemetry session for this emulation session.
+    const Core::TelemetrySession& TelemetrySession() const;
 
     /// Prepare the core emulation for a reschedule
     void PrepareReschedule();
@@ -144,21 +144,36 @@ public:
     /// Gets an ARM interface to the CPU core that is currently running
     ARM_Interface& CurrentArmInterface();
 
+    /// Gets an ARM interface to the CPU core that is currently running
+    const ARM_Interface& CurrentArmInterface() const;
+
     /// Gets the index of the currently running CPU core
-    std::size_t CurrentCoreIndex();
+    std::size_t CurrentCoreIndex() const;
 
     /// Gets the scheduler for the CPU core that is currently running
     Kernel::Scheduler& CurrentScheduler();
 
-    /// Gets an ARM interface to the CPU core with the specified index
+    /// Gets the scheduler for the CPU core that is currently running
+    const Kernel::Scheduler& CurrentScheduler() const;
+
+    /// Gets a reference to an ARM interface for the CPU core with the specified index
     ARM_Interface& ArmInterface(std::size_t core_index);
 
+    /// Gets a const reference to an ARM interface from the CPU core with the specified index
+    const ARM_Interface& ArmInterface(std::size_t core_index) const;
+
     /// Gets a CPU interface to the CPU core with the specified index
     Cpu& CpuCore(std::size_t core_index);
 
-    /// Gets the exclusive monitor
+    /// Gets a CPU interface to the CPU core with the specified index
+    const Cpu& CpuCore(std::size_t core_index) const;
+
+    /// Gets a reference to the exclusive monitor
     ExclusiveMonitor& Monitor();
 
+    /// Gets a constant reference to the exclusive monitor
+    const ExclusiveMonitor& Monitor() const;
+
     /// Gets a mutable reference to the GPU interface
     Tegra::GPU& GPU();
 
@@ -172,13 +187,16 @@ public:
     const VideoCore::RendererBase& Renderer() const;
 
     /// Gets the scheduler for the CPU core with the specified index
-    const std::shared_ptr<Kernel::Scheduler>& Scheduler(std::size_t core_index);
+    Kernel::Scheduler& Scheduler(std::size_t core_index);
 
-    /// Provides a reference to the current process
-    Kernel::SharedPtr<Kernel::Process>& CurrentProcess();
+    /// Gets the scheduler for the CPU core with the specified index
+    const Kernel::Scheduler& Scheduler(std::size_t core_index) const;
 
-    /// Provides a constant reference to the current process.
-    const Kernel::SharedPtr<Kernel::Process>& CurrentProcess() const;
+    /// Provides a pointer to the current process
+    Kernel::Process* CurrentProcess();
+
+    /// Provides a constant pointer to the current process.
+    const Kernel::Process* CurrentProcess() const;
 
     /// Provides a reference to the kernel instance.
     Kernel::KernelCore& Kernel();
@@ -224,6 +242,9 @@ private:
     /// Returns the currently running CPU core
     Cpu& CurrentCpuCore();
 
+    /// Returns the currently running CPU core
+    const Cpu& CurrentCpuCore() const;
+
     /**
      * Initialize the emulated system.
      * @param emu_window Reference to the host-system window used for video output and keyboard
@@ -246,7 +267,7 @@ inline TelemetrySession& Telemetry() {
     return System::GetInstance().TelemetrySession();
 }
 
-inline Kernel::SharedPtr<Kernel::Process>& CurrentProcess() {
+inline Kernel::Process* CurrentProcess() {
     return System::GetInstance().CurrentProcess();
 }
 
diff --git a/src/core/core_cpu.cpp b/src/core/core_cpu.cpp
index 265f8ed9c..fffda8a99 100644
--- a/src/core/core_cpu.cpp
+++ b/src/core/core_cpu.cpp
@@ -49,10 +49,8 @@ bool CpuBarrier::Rendezvous() {
     return false;
 }
 
-Cpu::Cpu(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
-         std::shared_ptr<CpuBarrier> cpu_barrier, std::size_t core_index)
-    : cpu_barrier{std::move(cpu_barrier)}, core_index{core_index} {
-
+Cpu::Cpu(ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_barrier, std::size_t core_index)
+    : cpu_barrier{cpu_barrier}, core_index{core_index} {
     if (Settings::values.use_cpu_jit) {
 #ifdef ARCHITECTURE_x86_64
         arm_interface = std::make_unique<ARM_Dynarmic>(exclusive_monitor, core_index);
@@ -64,15 +62,15 @@ Cpu::Cpu(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
         arm_interface = std::make_unique<ARM_Unicorn>();
     }
 
-    scheduler = std::make_shared<Kernel::Scheduler>(*arm_interface);
+    scheduler = std::make_unique<Kernel::Scheduler>(*arm_interface);
 }
 
 Cpu::~Cpu() = default;
 
-std::shared_ptr<ExclusiveMonitor> Cpu::MakeExclusiveMonitor(std::size_t num_cores) {
+std::unique_ptr<ExclusiveMonitor> Cpu::MakeExclusiveMonitor(std::size_t num_cores) {
     if (Settings::values.use_cpu_jit) {
 #ifdef ARCHITECTURE_x86_64
-        return std::make_shared<DynarmicExclusiveMonitor>(num_cores);
+        return std::make_unique<DynarmicExclusiveMonitor>(num_cores);
 #else
         return nullptr; // TODO(merry): Passthrough exclusive monitor
 #endif
@@ -83,7 +81,7 @@ std::shared_ptr<ExclusiveMonitor> Cpu::MakeExclusiveMonitor(std::size_t num_core
 
 void Cpu::RunLoop(bool tight_loop) {
     // Wait for all other CPU cores to complete the previous slice, such that they run in lock-step
-    if (!cpu_barrier->Rendezvous()) {
+    if (!cpu_barrier.Rendezvous()) {
         // If rendezvous failed, session has been killed
         return;
     }
diff --git a/src/core/core_cpu.h b/src/core/core_cpu.h
index ee7e04abc..1d2bdc6cd 100644
--- a/src/core/core_cpu.h
+++ b/src/core/core_cpu.h
@@ -41,8 +41,7 @@ private:
 
 class Cpu {
 public:
-    Cpu(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
-        std::shared_ptr<CpuBarrier> cpu_barrier, std::size_t core_index);
+    Cpu(ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_barrier, std::size_t core_index);
     ~Cpu();
 
     void RunLoop(bool tight_loop = true);
@@ -59,8 +58,12 @@ public:
         return *arm_interface;
     }
 
-    const std::shared_ptr<Kernel::Scheduler>& Scheduler() const {
-        return scheduler;
+    Kernel::Scheduler& Scheduler() {
+        return *scheduler;
+    }
+
+    const Kernel::Scheduler& Scheduler() const {
+        return *scheduler;
     }
 
     bool IsMainCore() const {
@@ -71,14 +74,14 @@ public:
         return core_index;
     }
 
-    static std::shared_ptr<ExclusiveMonitor> MakeExclusiveMonitor(std::size_t num_cores);
+    static std::unique_ptr<ExclusiveMonitor> MakeExclusiveMonitor(std::size_t num_cores);
 
 private:
     void Reschedule();
 
     std::unique_ptr<ARM_Interface> arm_interface;
-    std::shared_ptr<CpuBarrier> cpu_barrier;
-    std::shared_ptr<Kernel::Scheduler> scheduler;
+    CpuBarrier& cpu_barrier;
+    std::unique_ptr<Kernel::Scheduler> scheduler;
 
     std::atomic<bool> reschedule_pending = false;
     std::size_t core_index;
diff --git a/src/core/crypto/key_manager.cpp b/src/core/crypto/key_manager.cpp
index bf3a70944..904afa039 100644
--- a/src/core/crypto/key_manager.cpp
+++ b/src/core/crypto/key_manager.cpp
@@ -4,23 +4,56 @@
 
 #include <algorithm>
 #include <array>
+#include <bitset>
+#include <cctype>
 #include <fstream>
 #include <locale>
+#include <map>
 #include <sstream>
 #include <string_view>
 #include <tuple>
 #include <vector>
+#include <mbedtls/bignum.h>
+#include <mbedtls/cipher.h>
+#include <mbedtls/cmac.h>
+#include <mbedtls/sha256.h>
+#include "common/common_funcs.h"
 #include "common/common_paths.h"
 #include "common/file_util.h"
 #include "common/hex_util.h"
 #include "common/logging/log.h"
 #include "core/crypto/aes_util.h"
 #include "core/crypto/key_manager.h"
+#include "core/crypto/partition_data_manager.h"
+#include "core/file_sys/content_archive.h"
+#include "core/file_sys/nca_metadata.h"
+#include "core/file_sys/partition_filesystem.h"
+#include "core/file_sys/registered_cache.h"
+#include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/loader.h"
 #include "core/settings.h"
 
 namespace Core::Crypto {
 
+constexpr u64 CURRENT_CRYPTO_REVISION = 0x5;
+
+using namespace Common;
+
+const std::array<SHA256Hash, 2> eticket_source_hashes{
+    "B71DB271DC338DF380AA2C4335EF8873B1AFD408E80B3582D8719FC81C5E511C"_array32, // eticket_rsa_kek_source
+    "E8965A187D30E57869F562D04383C996DE487BBA5761363D2D4D32391866A85C"_array32, // eticket_rsa_kekek_source
+};
+
+const std::map<std::pair<S128KeyType, u64>, std::string> KEYS_VARIABLE_LENGTH{
+    {{S128KeyType::Master, 0}, "master_key_"},
+    {{S128KeyType::Package1, 0}, "package1_key_"},
+    {{S128KeyType::Package2, 0}, "package2_key_"},
+    {{S128KeyType::Titlekek, 0}, "titlekek_"},
+    {{S128KeyType::Source, static_cast<u64>(SourceKeyType::Keyblob)}, "keyblob_key_source_"},
+    {{S128KeyType::Keyblob, 0}, "keyblob_key_"},
+    {{S128KeyType::KeyblobMAC, 0}, "keyblob_mac_key_"},
+};
+
 Key128 GenerateKeyEncryptionKey(Key128 source, Key128 master, Key128 kek_seed, Key128 key_seed) {
     Key128 out{};
 
@@ -37,57 +70,136 @@ Key128 GenerateKeyEncryptionKey(Key128 source, Key128 master, Key128 kek_seed, K
     return out;
 }
 
-boost::optional<Key128> DeriveSDSeed() {
+Key128 DeriveKeyblobKey(const Key128& sbk, const Key128& tsec, Key128 source) {
+    AESCipher<Key128> sbk_cipher(sbk, Mode::ECB);
+    AESCipher<Key128> tsec_cipher(tsec, Mode::ECB);
+    tsec_cipher.Transcode(source.data(), source.size(), source.data(), Op::Decrypt);
+    sbk_cipher.Transcode(source.data(), source.size(), source.data(), Op::Decrypt);
+    return source;
+}
+
+Key128 DeriveMasterKey(const std::array<u8, 0x90>& keyblob, const Key128& master_source) {
+    Key128 master_root;
+    std::memcpy(master_root.data(), keyblob.data(), sizeof(Key128));
+
+    AESCipher<Key128> master_cipher(master_root, Mode::ECB);
+
+    Key128 master{};
+    master_cipher.Transcode(master_source.data(), master_source.size(), master.data(), Op::Decrypt);
+    return master;
+}
+
+std::array<u8, 144> DecryptKeyblob(const std::array<u8, 176>& encrypted_keyblob,
+                                   const Key128& key) {
+    std::array<u8, 0x90> keyblob;
+    AESCipher<Key128> cipher(key, Mode::CTR);
+    cipher.SetIV(std::vector<u8>(encrypted_keyblob.data() + 0x10, encrypted_keyblob.data() + 0x20));
+    cipher.Transcode(encrypted_keyblob.data() + 0x20, keyblob.size(), keyblob.data(), Op::Decrypt);
+    return keyblob;
+}
+
+void KeyManager::DeriveGeneralPurposeKeys(std::size_t crypto_revision) {
+    const auto kek_generation_source =
+        GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration));
+    const auto key_generation_source =
+        GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration));
+
+    if (HasKey(S128KeyType::Master, crypto_revision)) {
+        for (auto kak_type :
+             {KeyAreaKeyType::Application, KeyAreaKeyType::Ocean, KeyAreaKeyType::System}) {
+            if (HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey),
+                       static_cast<u64>(kak_type))) {
+                const auto source =
+                    GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey),
+                           static_cast<u64>(kak_type));
+                const auto kek =
+                    GenerateKeyEncryptionKey(source, GetKey(S128KeyType::Master, crypto_revision),
+                                             kek_generation_source, key_generation_source);
+                SetKey(S128KeyType::KeyArea, kek, crypto_revision, static_cast<u64>(kak_type));
+            }
+        }
+
+        AESCipher<Key128> master_cipher(GetKey(S128KeyType::Master, crypto_revision), Mode::ECB);
+        for (auto key_type : {SourceKeyType::Titlekek, SourceKeyType::Package2}) {
+            if (HasKey(S128KeyType::Source, static_cast<u64>(key_type))) {
+                Key128 key{};
+                master_cipher.Transcode(
+                    GetKey(S128KeyType::Source, static_cast<u64>(key_type)).data(), key.size(),
+                    key.data(), Op::Decrypt);
+                SetKey(key_type == SourceKeyType::Titlekek ? S128KeyType::Titlekek
+                                                           : S128KeyType::Package2,
+                       key, crypto_revision);
+            }
+        }
+    }
+}
+
+Key128 DeriveKeyblobMACKey(const Key128& keyblob_key, const Key128& mac_source) {
+    AESCipher<Key128> mac_cipher(keyblob_key, Mode::ECB);
+    Key128 mac_key{};
+    mac_cipher.Transcode(mac_source.data(), mac_key.size(), mac_key.data(), Op::Decrypt);
+    return mac_key;
+}
+
+std::optional<Key128> DeriveSDSeed() {
     const FileUtil::IOFile save_43(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
                                        "/system/save/8000000000000043",
                                    "rb+");
     if (!save_43.IsOpen())
-        return boost::none;
+        return {};
+
     const FileUtil::IOFile sd_private(
         FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir) + "/Nintendo/Contents/private", "rb+");
     if (!sd_private.IsOpen())
-        return boost::none;
+        return {};
 
-    sd_private.Seek(0, SEEK_SET);
     std::array<u8, 0x10> private_seed{};
-    if (sd_private.ReadBytes(private_seed.data(), private_seed.size()) != 0x10)
-        return boost::none;
+    if (sd_private.ReadBytes(private_seed.data(), private_seed.size()) != private_seed.size()) {
+        return {};
+    }
 
     std::array<u8, 0x10> buffer{};
     std::size_t offset = 0;
     for (; offset + 0x10 < save_43.GetSize(); ++offset) {
-        save_43.Seek(offset, SEEK_SET);
+        if (!save_43.Seek(offset, SEEK_SET)) {
+            return {};
+        }
+
         save_43.ReadBytes(buffer.data(), buffer.size());
-        if (buffer == private_seed)
+        if (buffer == private_seed) {
             break;
+        }
     }
 
-    if (offset + 0x10 >= save_43.GetSize())
-        return boost::none;
+    if (!save_43.Seek(offset + 0x10, SEEK_SET)) {
+        return {};
+    }
 
     Key128 seed{};
-    save_43.Seek(offset + 0x10, SEEK_SET);
-    save_43.ReadBytes(seed.data(), seed.size());
+    if (save_43.ReadBytes(seed.data(), seed.size()) != seed.size()) {
+        return {};
+    }
     return seed;
 }
 
-Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, const KeyManager& keys) {
-    if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKEK)))
+Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& keys) {
+    if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKek)))
         return Loader::ResultStatus::ErrorMissingSDKEKSource;
-    if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKEKGeneration)))
+    if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration)))
         return Loader::ResultStatus::ErrorMissingAESKEKGenerationSource;
     if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration)))
         return Loader::ResultStatus::ErrorMissingAESKeyGenerationSource;
 
     const auto sd_kek_source =
-        keys.GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKEK));
+        keys.GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKek));
     const auto aes_kek_gen =
-        keys.GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKEKGeneration));
+        keys.GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration));
     const auto aes_key_gen =
         keys.GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration));
     const auto master_00 = keys.GetKey(S128KeyType::Master);
     const auto sd_kek =
         GenerateKeyEncryptionKey(sd_kek_source, master_00, aes_kek_gen, aes_key_gen);
+    keys.SetKey(S128KeyType::SDKek, sd_kek);
 
     if (!keys.HasKey(S128KeyType::SDSeed))
         return Loader::ResultStatus::ErrorMissingSDSeed;
@@ -118,9 +230,147 @@ Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, const KeyManag
                        return source; ///< Return unaltered source to satisfy output requirement.
                    });
 
+    keys.SetKey(S256KeyType::SDKey, sd_keys[0], static_cast<u64>(SDKeyType::Save));
+    keys.SetKey(S256KeyType::SDKey, sd_keys[1], static_cast<u64>(SDKeyType::NCA));
+
     return Loader::ResultStatus::Success;
 }
 
+std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save) {
+    if (!ticket_save.IsOpen())
+        return {};
+
+    std::vector<u8> buffer(ticket_save.GetSize());
+    if (ticket_save.ReadBytes(buffer.data(), buffer.size()) != buffer.size()) {
+        return {};
+    }
+
+    std::vector<TicketRaw> out;
+    u32 magic{};
+    for (std::size_t offset = 0; offset + 0x4 < buffer.size(); ++offset) {
+        if (buffer[offset] == 0x4 && buffer[offset + 1] == 0x0 && buffer[offset + 2] == 0x1 &&
+            buffer[offset + 3] == 0x0) {
+            out.emplace_back();
+            auto& next = out.back();
+            std::memcpy(&next, buffer.data() + offset, sizeof(TicketRaw));
+            offset += next.size();
+        }
+    }
+
+    return out;
+}
+
+template <size_t size>
+static std::array<u8, size> operator^(const std::array<u8, size>& lhs,
+                                      const std::array<u8, size>& rhs) {
+    std::array<u8, size> out{};
+    std::transform(lhs.begin(), lhs.end(), rhs.begin(), out.begin(), std::bit_xor<>());
+    return out;
+}
+
+template <size_t target_size, size_t in_size>
+static std::array<u8, target_size> MGF1(const std::array<u8, in_size>& seed) {
+    // Avoids truncation overflow within the loop below.
+    static_assert(target_size <= 0xFF);
+
+    std::array<u8, in_size + 4> seed_exp{};
+    std::memcpy(seed_exp.data(), seed.data(), in_size);
+
+    std::vector<u8> out;
+    size_t i = 0;
+    while (out.size() < target_size) {
+        out.resize(out.size() + 0x20);
+        seed_exp[in_size + 3] = static_cast<u8>(i);
+        mbedtls_sha256(seed_exp.data(), seed_exp.size(), out.data() + out.size() - 0x20, 0);
+        ++i;
+    }
+
+    std::array<u8, target_size> target;
+    std::memcpy(target.data(), out.data(), target_size);
+    return target;
+}
+
+template <size_t size>
+static std::optional<u64> FindTicketOffset(const std::array<u8, size>& data) {
+    u64 offset = 0;
+    for (size_t i = 0x20; i < data.size() - 0x10; ++i) {
+        if (data[i] == 0x1) {
+            offset = i + 1;
+            break;
+        } else if (data[i] != 0x0) {
+            return {};
+        }
+    }
+
+    return offset;
+}
+
+std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
+                                                     const RSAKeyPair<2048>& key) {
+    u32 cert_authority;
+    std::memcpy(&cert_authority, ticket.data() + 0x140, sizeof(cert_authority));
+    if (cert_authority == 0)
+        return {};
+    if (cert_authority != Common::MakeMagic('R', 'o', 'o', 't')) {
+        LOG_INFO(Crypto,
+                 "Attempting to parse ticket with non-standard certificate authority {:08X}.",
+                 cert_authority);
+    }
+
+    Key128 rights_id;
+    std::memcpy(rights_id.data(), ticket.data() + 0x2A0, sizeof(Key128));
+
+    if (rights_id == Key128{})
+        return {};
+
+    Key128 key_temp{};
+
+    if (!std::any_of(ticket.begin() + 0x190, ticket.begin() + 0x280, [](u8 b) { return b != 0; })) {
+        std::memcpy(key_temp.data(), ticket.data() + 0x180, key_temp.size());
+        return std::make_pair(rights_id, key_temp);
+    }
+
+    mbedtls_mpi D; // RSA Private Exponent
+    mbedtls_mpi N; // RSA Modulus
+    mbedtls_mpi S; // Input
+    mbedtls_mpi M; // Output
+
+    mbedtls_mpi_init(&D);
+    mbedtls_mpi_init(&N);
+    mbedtls_mpi_init(&S);
+    mbedtls_mpi_init(&M);
+
+    mbedtls_mpi_read_binary(&D, key.decryption_key.data(), key.decryption_key.size());
+    mbedtls_mpi_read_binary(&N, key.modulus.data(), key.modulus.size());
+    mbedtls_mpi_read_binary(&S, ticket.data() + 0x180, 0x100);
+
+    mbedtls_mpi_exp_mod(&M, &S, &D, &N, nullptr);
+
+    std::array<u8, 0x100> rsa_step;
+    mbedtls_mpi_write_binary(&M, rsa_step.data(), rsa_step.size());
+
+    u8 m_0 = rsa_step[0];
+    std::array<u8, 0x20> m_1;
+    std::memcpy(m_1.data(), rsa_step.data() + 0x01, m_1.size());
+    std::array<u8, 0xDF> m_2;
+    std::memcpy(m_2.data(), rsa_step.data() + 0x21, m_2.size());
+
+    if (m_0 != 0)
+        return {};
+
+    m_1 = m_1 ^ MGF1<0x20>(m_2);
+    m_2 = m_2 ^ MGF1<0xDF>(m_1);
+
+    const auto offset = FindTicketOffset(m_2);
+    if (!offset)
+        return {};
+    ASSERT(*offset > 0);
+
+    std::memcpy(key_temp.data(), m_2.data() + *offset, key_temp.size());
+
+    return std::make_pair(rights_id, key_temp);
+}
+
 KeyManager::KeyManager() {
     // Initialize keys
     const std::string hactool_keys_dir = FileUtil::GetHactoolConfigurationPath();
@@ -137,6 +387,15 @@ KeyManager::KeyManager() {
 
     AttemptLoadKeyFile(yuzu_keys_dir, hactool_keys_dir, "title.keys", true);
     AttemptLoadKeyFile(yuzu_keys_dir, yuzu_keys_dir, "title.keys_autogenerated", true);
+    AttemptLoadKeyFile(yuzu_keys_dir, hactool_keys_dir, "console.keys", false);
+    AttemptLoadKeyFile(yuzu_keys_dir, yuzu_keys_dir, "console.keys_autogenerated", false);
+}
+
+static bool ValidCryptoRevisionString(std::string_view base, size_t begin, size_t length) {
+    if (base.size() < begin + length)
+        return false;
+    return std::all_of(base.begin() + begin, base.begin() + begin + length,
+                       [](u8 c) { return std::isxdigit(c); });
 }
 
 void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {
@@ -158,6 +417,9 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {
         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 (out[0].compare(0, 1, "#") == 0)
+            continue;
+
         if (is_title_keys) {
             auto rights_id_raw = Common::HexStringToArray<16>(out[0]);
             u128 rights_id{};
@@ -174,6 +436,50 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {
                 const auto index = s256_file_id.at(out[0]);
                 Key256 key = Common::HexStringToArray<32>(out[1]);
                 s256_keys[{index.type, index.field1, index.field2}] = key;
+            } else if (out[0].compare(0, 8, "keyblob_") == 0 &&
+                       out[0].compare(0, 9, "keyblob_k") != 0) {
+                if (!ValidCryptoRevisionString(out[0], 8, 2))
+                    continue;
+
+                const auto index = std::stoul(out[0].substr(8, 2), nullptr, 16);
+                keyblobs[index] = Common::HexStringToArray<0x90>(out[1]);
+            } else if (out[0].compare(0, 18, "encrypted_keyblob_") == 0) {
+                if (!ValidCryptoRevisionString(out[0], 18, 2))
+                    continue;
+
+                const auto index = std::stoul(out[0].substr(18, 2), nullptr, 16);
+                encrypted_keyblobs[index] = Common::HexStringToArray<0xB0>(out[1]);
+            } else {
+                for (const auto& kv : KEYS_VARIABLE_LENGTH) {
+                    if (!ValidCryptoRevisionString(out[0], kv.second.size(), 2))
+                        continue;
+                    if (out[0].compare(0, kv.second.size(), kv.second) == 0) {
+                        const auto index =
+                            std::stoul(out[0].substr(kv.second.size(), 2), nullptr, 16);
+                        const auto sub = kv.first.second;
+                        if (sub == 0) {
+                            s128_keys[{kv.first.first, index, 0}] =
+                                Common::HexStringToArray<16>(out[1]);
+                        } else {
+                            s128_keys[{kv.first.first, kv.first.second, index}] =
+                                Common::HexStringToArray<16>(out[1]);
+                        }
+
+                        break;
+                    }
+                }
+
+                static constexpr std::array<const char*, 3> kak_names = {
+                    "key_area_key_application_", "key_area_key_ocean_", "key_area_key_system_"};
+                for (size_t j = 0; j < kak_names.size(); ++j) {
+                    const auto& match = kak_names[j];
+                    if (out[0].compare(0, std::strlen(match), match) == 0) {
+                        const auto index =
+                            std::stoul(out[0].substr(std::strlen(match), 2), nullptr, 16);
+                        s128_keys[{S128KeyType::KeyArea, index, j}] =
+                            Common::HexStringToArray<16>(out[1]);
+                    }
+                }
             }
         }
     }
@@ -187,6 +493,28 @@ void KeyManager::AttemptLoadKeyFile(const std::string& dir1, const std::string&
         LoadFromFile(dir2 + DIR_SEP + filename, title);
 }
 
+bool KeyManager::BaseDeriveNecessary() const {
+    const auto check_key_existence = [this](auto key_type, u64 index1 = 0, u64 index2 = 0) {
+        return !HasKey(key_type, index1, index2);
+    };
+
+    if (check_key_existence(S256KeyType::Header))
+        return true;
+
+    for (size_t i = 0; i < CURRENT_CRYPTO_REVISION; ++i) {
+        if (check_key_existence(S128KeyType::Master, i) ||
+            check_key_existence(S128KeyType::KeyArea, i,
+                                static_cast<u64>(KeyAreaKeyType::Application)) ||
+            check_key_existence(S128KeyType::KeyArea, i, static_cast<u64>(KeyAreaKeyType::Ocean)) ||
+            check_key_existence(S128KeyType::KeyArea, i,
+                                static_cast<u64>(KeyAreaKeyType::System)) ||
+            check_key_existence(S128KeyType::Titlekek, i))
+            return true;
+    }
+
+    return false;
+}
+
 bool KeyManager::HasKey(S128KeyType id, u64 field1, u64 field2) const {
     return s128_keys.find({id, field1, field2}) != s128_keys.end();
 }
@@ -207,13 +535,30 @@ Key256 KeyManager::GetKey(S256KeyType id, u64 field1, u64 field2) const {
     return s256_keys.at({id, field1, field2});
 }
 
-template <std::size_t Size>
-void KeyManager::WriteKeyToFile(bool title_key, std::string_view keyname,
+Key256 KeyManager::GetBISKey(u8 partition_id) const {
+    Key256 out{};
+
+    for (const auto& bis_type : {BISKeyType::Crypto, BISKeyType::Tweak}) {
+        if (HasKey(S128KeyType::BIS, partition_id, static_cast<u64>(bis_type))) {
+            std::memcpy(
+                out.data() + sizeof(Key128) * static_cast<u64>(bis_type),
+                s128_keys.at({S128KeyType::BIS, partition_id, static_cast<u64>(bis_type)}).data(),
+                sizeof(Key128));
+        }
+    }
+
+    return out;
+}
+
+template <size_t Size>
+void KeyManager::WriteKeyToFile(KeyCategory category, std::string_view keyname,
                                 const std::array<u8, Size>& key) {
     const std::string yuzu_keys_dir = FileUtil::GetUserPath(FileUtil::UserPath::KeysDir);
     std::string filename = "title.keys_autogenerated";
-    if (!title_key)
+    if (category == KeyCategory::Standard)
         filename = dev_mode ? "dev.keys_autogenerated" : "prod.keys_autogenerated";
+    else if (category == KeyCategory::Console)
+        filename = "console.keys_autogenerated";
     const auto add_info_text = !FileUtil::Exists(yuzu_keys_dir + DIR_SEP + filename);
     FileUtil::CreateFullPath(yuzu_keys_dir + DIR_SEP + filename);
     std::ofstream file(yuzu_keys_dir + DIR_SEP + filename, std::ios::app);
@@ -227,7 +572,7 @@ void KeyManager::WriteKeyToFile(bool title_key, std::string_view keyname,
     }
 
     file << fmt::format("\n{} = {}", keyname, Common::HexArrayToString(key));
-    AttemptLoadKeyFile(yuzu_keys_dir, yuzu_keys_dir, filename, title_key);
+    AttemptLoadKeyFile(yuzu_keys_dir, yuzu_keys_dir, filename, category == KeyCategory::Title);
 }
 
 void KeyManager::SetKey(S128KeyType id, Key128 key, u64 field1, u64 field2) {
@@ -237,8 +582,15 @@ void KeyManager::SetKey(S128KeyType id, Key128 key, u64 field1, u64 field2) {
         Key128 rights_id;
         std::memcpy(rights_id.data(), &field2, sizeof(u64));
         std::memcpy(rights_id.data() + sizeof(u64), &field1, sizeof(u64));
-        WriteKeyToFile(true, Common::HexArrayToString(rights_id), key);
+        WriteKeyToFile(KeyCategory::Title, Common::HexArrayToString(rights_id), key);
+    }
+
+    auto category = KeyCategory::Standard;
+    if (id == S128KeyType::Keyblob || id == S128KeyType::KeyblobMAC || id == S128KeyType::TSEC ||
+        id == S128KeyType::SecureBoot || id == S128KeyType::SDSeed || id == S128KeyType::BIS) {
+        category = KeyCategory::Console;
     }
+
     const auto iter2 = std::find_if(
         s128_file_id.begin(), s128_file_id.end(),
         [&id, &field1, &field2](const std::pair<std::string, KeyIndex<S128KeyType>> elem) {
@@ -246,7 +598,30 @@ void KeyManager::SetKey(S128KeyType id, Key128 key, u64 field1, u64 field2) {
                    std::tie(id, field1, field2);
         });
     if (iter2 != s128_file_id.end())
-        WriteKeyToFile(false, iter2->first, key);
+        WriteKeyToFile(category, iter2->first, key);
+
+    // Variable cases
+    if (id == S128KeyType::KeyArea) {
+        static constexpr std::array<const char*, 3> kak_names = {"key_area_key_application_{:02X}",
+                                                                 "key_area_key_ocean_{:02X}",
+                                                                 "key_area_key_system_{:02X}"};
+        WriteKeyToFile(category, fmt::format(kak_names.at(field2), field1), key);
+    } else if (id == S128KeyType::Master) {
+        WriteKeyToFile(category, fmt::format("master_key_{:02X}", field1), key);
+    } else if (id == S128KeyType::Package1) {
+        WriteKeyToFile(category, fmt::format("package1_key_{:02X}", field1), key);
+    } else if (id == S128KeyType::Package2) {
+        WriteKeyToFile(category, fmt::format("package2_key_{:02X}", field1), key);
+    } else if (id == S128KeyType::Titlekek) {
+        WriteKeyToFile(category, fmt::format("titlekek_{:02X}", field1), key);
+    } else if (id == S128KeyType::Keyblob) {
+        WriteKeyToFile(category, fmt::format("keyblob_key_{:02X}", field1), key);
+    } else if (id == S128KeyType::KeyblobMAC) {
+        WriteKeyToFile(category, fmt::format("keyblob_mac_key_{:02X}", field1), key);
+    } else if (id == S128KeyType::Source && field1 == static_cast<u64>(SourceKeyType::Keyblob)) {
+        WriteKeyToFile(category, fmt::format("keyblob_key_source_{:02X}", field2), key);
+    }
+
     s128_keys[{id, field1, field2}] = key;
 }
 
@@ -260,7 +635,7 @@ void KeyManager::SetKey(S256KeyType id, Key256 key, u64 field1, u64 field2) {
                    std::tie(id, field1, field2);
         });
     if (iter != s256_file_id.end())
-        WriteKeyToFile(false, iter->first, key);
+        WriteKeyToFile(KeyCategory::Standard, iter->first, key);
     s256_keys[{id, field1, field2}] = key;
 }
 
@@ -286,63 +661,391 @@ void KeyManager::DeriveSDSeedLazy() {
         return;
 
     const auto res = DeriveSDSeed();
-    if (res != boost::none)
-        SetKey(S128KeyType::SDSeed, res.get());
+    if (res)
+        SetKey(S128KeyType::SDSeed, *res);
+}
+
+static Key128 CalculateCMAC(const u8* source, size_t size, const Key128& key) {
+    Key128 out{};
+
+    mbedtls_cipher_cmac(mbedtls_cipher_info_from_type(MBEDTLS_CIPHER_AES_128_ECB), key.data(),
+                        key.size() * 8, source, size, out.data());
+    return out;
+}
+
+void KeyManager::DeriveBase() {
+    if (!BaseDeriveNecessary())
+        return;
+
+    if (!HasKey(S128KeyType::SecureBoot) || !HasKey(S128KeyType::TSEC))
+        return;
+
+    const auto has_bis = [this](u64 id) {
+        return HasKey(S128KeyType::BIS, id, static_cast<u64>(BISKeyType::Crypto)) &&
+               HasKey(S128KeyType::BIS, id, static_cast<u64>(BISKeyType::Tweak));
+    };
+
+    const auto copy_bis = [this](u64 id_from, u64 id_to) {
+        SetKey(S128KeyType::BIS,
+               GetKey(S128KeyType::BIS, id_from, static_cast<u64>(BISKeyType::Crypto)), id_to,
+               static_cast<u64>(BISKeyType::Crypto));
+
+        SetKey(S128KeyType::BIS,
+               GetKey(S128KeyType::BIS, id_from, static_cast<u64>(BISKeyType::Tweak)), id_to,
+               static_cast<u64>(BISKeyType::Tweak));
+    };
+
+    if (has_bis(2) && !has_bis(3))
+        copy_bis(2, 3);
+    else if (has_bis(3) && !has_bis(2))
+        copy_bis(3, 2);
+
+    std::bitset<32> revisions(0xFFFFFFFF);
+    for (size_t i = 0; i < revisions.size(); ++i) {
+        if (!HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::Keyblob), i) ||
+            encrypted_keyblobs[i] == std::array<u8, 0xB0>{}) {
+            revisions.reset(i);
+        }
+    }
+
+    if (!revisions.any())
+        return;
+
+    const auto sbk = GetKey(S128KeyType::SecureBoot);
+    const auto tsec = GetKey(S128KeyType::TSEC);
+
+    for (size_t i = 0; i < revisions.size(); ++i) {
+        if (!revisions[i])
+            continue;
+
+        // Derive keyblob key
+        const auto key = DeriveKeyblobKey(
+            sbk, tsec, GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::Keyblob), i));
+
+        SetKey(S128KeyType::Keyblob, key, i);
+
+        // Derive keyblob MAC key
+        if (!HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyblobMAC)))
+            continue;
+
+        const auto mac_key = DeriveKeyblobMACKey(
+            key, GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyblobMAC)));
+        SetKey(S128KeyType::KeyblobMAC, mac_key, i);
+
+        Key128 cmac = CalculateCMAC(encrypted_keyblobs[i].data() + 0x10, 0xA0, mac_key);
+        if (std::memcmp(cmac.data(), encrypted_keyblobs[i].data(), cmac.size()) != 0)
+            continue;
+
+        // Decrypt keyblob
+        if (keyblobs[i] == std::array<u8, 0x90>{}) {
+            keyblobs[i] = DecryptKeyblob(encrypted_keyblobs[i], key);
+            WriteKeyToFile<0x90>(KeyCategory::Console, fmt::format("keyblob_{:02X}", i),
+                                 keyblobs[i]);
+        }
+
+        Key128 package1;
+        std::memcpy(package1.data(), keyblobs[i].data() + 0x80, sizeof(Key128));
+        SetKey(S128KeyType::Package1, package1, i);
+
+        // Derive master key
+        if (HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::Master))) {
+            SetKey(S128KeyType::Master,
+                   DeriveMasterKey(keyblobs[i], GetKey(S128KeyType::Source,
+                                                       static_cast<u64>(SourceKeyType::Master))),
+                   i);
+        }
+    }
+
+    revisions.set();
+    for (size_t i = 0; i < revisions.size(); ++i) {
+        if (!HasKey(S128KeyType::Master, i))
+            revisions.reset(i);
+    }
+
+    if (!revisions.any())
+        return;
+
+    for (size_t i = 0; i < revisions.size(); ++i) {
+        if (!revisions[i])
+            continue;
+
+        // Derive general purpose keys
+        DeriveGeneralPurposeKeys(i);
+    }
+
+    if (HasKey(S128KeyType::Master, 0) &&
+        HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration)) &&
+        HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration)) &&
+        HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::HeaderKek)) &&
+        HasKey(S256KeyType::HeaderSource)) {
+        const auto header_kek = GenerateKeyEncryptionKey(
+            GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::HeaderKek)),
+            GetKey(S128KeyType::Master, 0),
+            GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration)),
+            GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration)));
+        SetKey(S128KeyType::HeaderKek, header_kek);
+
+        AESCipher<Key128> header_cipher(header_kek, Mode::ECB);
+        Key256 out = GetKey(S256KeyType::HeaderSource);
+        header_cipher.Transcode(out.data(), out.size(), out.data(), Op::Decrypt);
+        SetKey(S256KeyType::Header, out);
+    }
+}
+
+void KeyManager::DeriveETicket(PartitionDataManager& data) {
+    // ETicket keys
+    const auto es = Service::FileSystem::GetUnionContents()->GetEntry(
+        0x0100000000000033, FileSys::ContentRecordType::Program);
+
+    if (es == nullptr)
+        return;
+
+    const auto exefs = es->GetExeFS();
+    if (exefs == nullptr)
+        return;
+
+    const auto main = exefs->GetFile("main");
+    if (main == nullptr)
+        return;
+
+    const auto bytes = main->ReadAllBytes();
+
+    const auto eticket_kek = FindKeyFromHex16(bytes, eticket_source_hashes[0]);
+    const auto eticket_kekek = FindKeyFromHex16(bytes, eticket_source_hashes[1]);
+
+    const auto seed3 = data.GetRSAKekSeed3();
+    const auto mask0 = data.GetRSAKekMask0();
+
+    if (eticket_kek != Key128{})
+        SetKey(S128KeyType::Source, eticket_kek, static_cast<size_t>(SourceKeyType::ETicketKek));
+    if (eticket_kekek != Key128{}) {
+        SetKey(S128KeyType::Source, eticket_kekek,
+               static_cast<size_t>(SourceKeyType::ETicketKekek));
+    }
+    if (seed3 != Key128{})
+        SetKey(S128KeyType::RSAKek, seed3, static_cast<size_t>(RSAKekType::Seed3));
+    if (mask0 != Key128{})
+        SetKey(S128KeyType::RSAKek, mask0, static_cast<size_t>(RSAKekType::Mask0));
+    if (eticket_kek == Key128{} || eticket_kekek == Key128{} || seed3 == Key128{} ||
+        mask0 == Key128{}) {
+        return;
+    }
+
+    Key128 rsa_oaep_kek{};
+    std::transform(seed3.begin(), seed3.end(), mask0.begin(), rsa_oaep_kek.begin(),
+                   std::bit_xor<>());
+
+    if (rsa_oaep_kek == Key128{})
+        return;
+
+    SetKey(S128KeyType::Source, rsa_oaep_kek,
+           static_cast<u64>(SourceKeyType::RSAOaepKekGeneration));
+
+    Key128 temp_kek{};
+    Key128 temp_kekek{};
+    Key128 eticket_final{};
+
+    // Derive ETicket RSA Kek
+    AESCipher<Key128> es_master(GetKey(S128KeyType::Master), Mode::ECB);
+    es_master.Transcode(rsa_oaep_kek.data(), rsa_oaep_kek.size(), temp_kek.data(), Op::Decrypt);
+    AESCipher<Key128> es_kekek(temp_kek, Mode::ECB);
+    es_kekek.Transcode(eticket_kekek.data(), eticket_kekek.size(), temp_kekek.data(), Op::Decrypt);
+    AESCipher<Key128> es_kek(temp_kekek, Mode::ECB);
+    es_kek.Transcode(eticket_kek.data(), eticket_kek.size(), eticket_final.data(), Op::Decrypt);
+
+    if (eticket_final == Key128{})
+        return;
+
+    SetKey(S128KeyType::ETicketRSAKek, eticket_final);
+
+    // Titlekeys
+    data.DecryptProdInfo(GetBISKey(0));
+
+    const auto eticket_extended_kek = data.GetETicketExtendedKek();
+
+    std::vector<u8> extended_iv(0x10);
+    std::memcpy(extended_iv.data(), eticket_extended_kek.data(), extended_iv.size());
+    std::array<u8, 0x230> extended_dec{};
+    AESCipher<Key128> rsa_1(eticket_final, Mode::CTR);
+    rsa_1.SetIV(extended_iv);
+    rsa_1.Transcode(eticket_extended_kek.data() + 0x10, eticket_extended_kek.size() - 0x10,
+                    extended_dec.data(), Op::Decrypt);
+
+    RSAKeyPair<2048> rsa_key{};
+    std::memcpy(rsa_key.decryption_key.data(), extended_dec.data(), rsa_key.decryption_key.size());
+    std::memcpy(rsa_key.modulus.data(), extended_dec.data() + 0x100, rsa_key.modulus.size());
+    std::memcpy(rsa_key.exponent.data(), extended_dec.data() + 0x200, rsa_key.exponent.size());
+
+    const FileUtil::IOFile save1(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                                     "/system/save/80000000000000e1",
+                                 "rb+");
+    const FileUtil::IOFile save2(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                                     "/system/save/80000000000000e2",
+                                 "rb+");
+
+    const auto blob2 = GetTicketblob(save2);
+    auto res = GetTicketblob(save1);
+    res.insert(res.end(), blob2.begin(), blob2.end());
+
+    for (const auto& raw : res) {
+        const auto pair = ParseTicket(raw, rsa_key);
+        if (!pair)
+            continue;
+        const auto& [rid, key] = *pair;
+        u128 rights_id;
+        std::memcpy(rights_id.data(), rid.data(), rid.size());
+        SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
+    }
+}
+
+void KeyManager::SetKeyWrapped(S128KeyType id, Key128 key, u64 field1, u64 field2) {
+    if (key == Key128{})
+        return;
+    SetKey(id, key, field1, field2);
+}
+
+void KeyManager::SetKeyWrapped(S256KeyType id, Key256 key, u64 field1, u64 field2) {
+    if (key == Key256{})
+        return;
+    SetKey(id, key, field1, field2);
+}
+
+void KeyManager::PopulateFromPartitionData(PartitionDataManager& data) {
+    if (!BaseDeriveNecessary())
+        return;
+
+    if (!data.HasBoot0())
+        return;
+
+    for (size_t i = 0; i < encrypted_keyblobs.size(); ++i) {
+        if (encrypted_keyblobs[i] != std::array<u8, 0xB0>{})
+            continue;
+        encrypted_keyblobs[i] = data.GetEncryptedKeyblob(i);
+        WriteKeyToFile<0xB0>(KeyCategory::Console, fmt::format("encrypted_keyblob_{:02X}", i),
+                             encrypted_keyblobs[i]);
+    }
+
+    SetKeyWrapped(S128KeyType::Source, data.GetPackage2KeySource(),
+                  static_cast<u64>(SourceKeyType::Package2));
+    SetKeyWrapped(S128KeyType::Source, data.GetAESKekGenerationSource(),
+                  static_cast<u64>(SourceKeyType::AESKekGeneration));
+    SetKeyWrapped(S128KeyType::Source, data.GetTitlekekSource(),
+                  static_cast<u64>(SourceKeyType::Titlekek));
+    SetKeyWrapped(S128KeyType::Source, data.GetMasterKeySource(),
+                  static_cast<u64>(SourceKeyType::Master));
+    SetKeyWrapped(S128KeyType::Source, data.GetKeyblobMACKeySource(),
+                  static_cast<u64>(SourceKeyType::KeyblobMAC));
+
+    for (size_t i = 0; i < PartitionDataManager::MAX_KEYBLOB_SOURCE_HASH; ++i) {
+        SetKeyWrapped(S128KeyType::Source, data.GetKeyblobKeySource(i),
+                      static_cast<u64>(SourceKeyType::Keyblob), i);
+    }
+
+    if (data.HasFuses())
+        SetKeyWrapped(S128KeyType::SecureBoot, data.GetSecureBootKey());
+
+    DeriveBase();
+
+    Key128 latest_master{};
+    for (s8 i = 0x1F; i >= 0; --i) {
+        if (GetKey(S128KeyType::Master, static_cast<u8>(i)) != Key128{}) {
+            latest_master = GetKey(S128KeyType::Master, static_cast<u8>(i));
+            break;
+        }
+    }
+
+    const auto masters = data.GetTZMasterKeys(latest_master);
+    for (size_t i = 0; i < masters.size(); ++i) {
+        if (masters[i] != Key128{} && !HasKey(S128KeyType::Master, i))
+            SetKey(S128KeyType::Master, masters[i], i);
+    }
+
+    DeriveBase();
+
+    if (!data.HasPackage2())
+        return;
+
+    std::array<Key128, 0x20> package2_keys{};
+    for (size_t i = 0; i < package2_keys.size(); ++i) {
+        if (HasKey(S128KeyType::Package2, i))
+            package2_keys[i] = GetKey(S128KeyType::Package2, i);
+    }
+    data.DecryptPackage2(package2_keys, Package2Type::NormalMain);
+
+    SetKeyWrapped(S128KeyType::Source, data.GetKeyAreaKeyApplicationSource(),
+                  static_cast<u64>(SourceKeyType::KeyAreaKey),
+                  static_cast<u64>(KeyAreaKeyType::Application));
+    SetKeyWrapped(S128KeyType::Source, data.GetKeyAreaKeyOceanSource(),
+                  static_cast<u64>(SourceKeyType::KeyAreaKey),
+                  static_cast<u64>(KeyAreaKeyType::Ocean));
+    SetKeyWrapped(S128KeyType::Source, data.GetKeyAreaKeySystemSource(),
+                  static_cast<u64>(SourceKeyType::KeyAreaKey),
+                  static_cast<u64>(KeyAreaKeyType::System));
+    SetKeyWrapped(S128KeyType::Source, data.GetSDKekSource(),
+                  static_cast<u64>(SourceKeyType::SDKek));
+    SetKeyWrapped(S256KeyType::SDKeySource, data.GetSDSaveKeySource(),
+                  static_cast<u64>(SDKeyType::Save));
+    SetKeyWrapped(S256KeyType::SDKeySource, data.GetSDNCAKeySource(),
+                  static_cast<u64>(SDKeyType::NCA));
+    SetKeyWrapped(S128KeyType::Source, data.GetHeaderKekSource(),
+                  static_cast<u64>(SourceKeyType::HeaderKek));
+    SetKeyWrapped(S256KeyType::HeaderSource, data.GetHeaderKeySource());
+    SetKeyWrapped(S128KeyType::Source, data.GetAESKeyGenerationSource(),
+                  static_cast<u64>(SourceKeyType::AESKeyGeneration));
+
+    DeriveBase();
 }
 
 const boost::container::flat_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)}},
-    {"sd_card_kek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKEK), 0}},
+    {"eticket_rsa_kek_source",
+     {S128KeyType::Source, static_cast<u64>(SourceKeyType::ETicketKek), 0}},
+    {"eticket_rsa_kekek_source",
+     {S128KeyType::Source, static_cast<u64>(SourceKeyType::ETicketKekek), 0}},
+    {"rsa_kek_mask_0", {S128KeyType::RSAKek, static_cast<u64>(RSAKekType::Mask0), 0}},
+    {"rsa_kek_seed_3", {S128KeyType::RSAKek, static_cast<u64>(RSAKekType::Seed3), 0}},
+    {"rsa_oaep_kek_generation_source",
+     {S128KeyType::Source, static_cast<u64>(SourceKeyType::RSAOaepKekGeneration), 0}},
+    {"sd_card_kek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKek), 0}},
     {"aes_kek_generation_source",
-     {S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKEKGeneration), 0}},
+     {S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration), 0}},
     {"aes_key_generation_source",
      {S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration), 0}},
+    {"package2_key_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Package2), 0}},
+    {"master_key_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Master), 0}},
+    {"header_kek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::HeaderKek), 0}},
+    {"key_area_key_application_source",
+     {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey),
+      static_cast<u64>(KeyAreaKeyType::Application)}},
+    {"key_area_key_ocean_source",
+     {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey),
+      static_cast<u64>(KeyAreaKeyType::Ocean)}},
+    {"key_area_key_system_source",
+     {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey),
+      static_cast<u64>(KeyAreaKeyType::System)}},
+    {"titlekek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Titlekek), 0}},
+    {"keyblob_mac_key_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyblobMAC)}},
+    {"tsec_key", {S128KeyType::TSEC, 0, 0}},
+    {"secure_boot_key", {S128KeyType::SecureBoot, 0, 0}},
     {"sd_seed", {S128KeyType::SDSeed, 0, 0}},
+    {"bis_key_0_crypt", {S128KeyType::BIS, 0, static_cast<u64>(BISKeyType::Crypto)}},
+    {"bis_key_0_tweak", {S128KeyType::BIS, 0, static_cast<u64>(BISKeyType::Tweak)}},
+    {"bis_key_1_crypt", {S128KeyType::BIS, 1, static_cast<u64>(BISKeyType::Crypto)}},
+    {"bis_key_1_tweak", {S128KeyType::BIS, 1, static_cast<u64>(BISKeyType::Tweak)}},
+    {"bis_key_2_crypt", {S128KeyType::BIS, 2, static_cast<u64>(BISKeyType::Crypto)}},
+    {"bis_key_2_tweak", {S128KeyType::BIS, 2, static_cast<u64>(BISKeyType::Tweak)}},
+    {"bis_key_3_crypt", {S128KeyType::BIS, 3, static_cast<u64>(BISKeyType::Crypto)}},
+    {"bis_key_3_tweak", {S128KeyType::BIS, 3, static_cast<u64>(BISKeyType::Tweak)}},
+    {"header_kek", {S128KeyType::HeaderKek, 0, 0}},
+    {"sd_card_kek", {S128KeyType::SDKek, 0, 0}},
 };
 
 const boost::container::flat_map<std::string, KeyIndex<S256KeyType>> KeyManager::s256_file_id = {
     {"header_key", {S256KeyType::Header, 0, 0}},
     {"sd_card_save_key_source", {S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::Save), 0}},
     {"sd_card_nca_key_source", {S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::NCA), 0}},
+    {"header_key_source", {S256KeyType::HeaderSource, 0, 0}},
+    {"sd_card_save_key", {S256KeyType::SDKey, static_cast<u64>(SDKeyType::Save), 0}},
+    {"sd_card_nca_key", {S256KeyType::SDKey, static_cast<u64>(SDKeyType::NCA), 0}},
 };
 } // namespace Core::Crypto
diff --git a/src/core/crypto/key_manager.h b/src/core/crypto/key_manager.h
index 978eec8dc..22f268c65 100644
--- a/src/core/crypto/key_manager.h
+++ b/src/core/crypto/key_manager.h
@@ -5,11 +5,19 @@
 #pragma once
 
 #include <array>
+#include <map>
+#include <optional>
 #include <string>
+
 #include <boost/container/flat_map.hpp>
-#include <boost/optional.hpp>
 #include <fmt/format.h>
 #include "common/common_types.h"
+#include "core/crypto/partition_data_manager.h"
+#include "core/file_sys/vfs_types.h"
+
+namespace FileUtil {
+class IOFile;
+}
 
 namespace Loader {
 enum class ResultStatus : u16;
@@ -22,13 +30,30 @@ constexpr u64 TICKET_FILE_TITLEKEY_OFFSET = 0x180;
 using Key128 = std::array<u8, 0x10>;
 using Key256 = std::array<u8, 0x20>;
 using SHA256Hash = std::array<u8, 0x20>;
+using TicketRaw = std::array<u8, 0x400>;
 
 static_assert(sizeof(Key128) == 16, "Key128 must be 128 bytes big.");
-static_assert(sizeof(Key256) == 32, "Key128 must be 128 bytes big.");
+static_assert(sizeof(Key256) == 32, "Key256 must be 256 bytes big.");
+
+template <size_t bit_size, size_t byte_size = (bit_size >> 3)>
+struct RSAKeyPair {
+    std::array<u8, byte_size> encryption_key;
+    std::array<u8, byte_size> decryption_key;
+    std::array<u8, byte_size> modulus;
+    std::array<u8, 4> exponent;
+};
+
+enum class KeyCategory : u8 {
+    Standard,
+    Title,
+    Console,
+};
 
 enum class S256KeyType : u64 {
-    Header,      //
-    SDKeySource, // f1=SDKeyType
+    SDKey,        // f1=SDKeyType
+    Header,       //
+    SDKeySource,  // f1=SDKeyType
+    HeaderSource, //
 };
 
 enum class S128KeyType : u64 {
@@ -41,6 +66,14 @@ enum class S128KeyType : u64 {
     SDSeed,        //
     Titlekey,      // f1=rights id LSB f2=rights id MSB
     Source,        // f1=source type, f2= sub id
+    Keyblob,       // f1=crypto revision
+    KeyblobMAC,    // f1=crypto revision
+    TSEC,          //
+    SecureBoot,    //
+    BIS,           // f1=partition (0-3), f2=type {crypt, tweak}
+    HeaderKek,     //
+    SDKek,         //
+    RSAKek,        //
 };
 
 enum class KeyAreaKeyType : u8 {
@@ -50,9 +83,19 @@ enum class KeyAreaKeyType : u8 {
 };
 
 enum class SourceKeyType : u8 {
-    SDKEK,
-    AESKEKGeneration,
-    AESKeyGeneration,
+    SDKek,                //
+    AESKekGeneration,     //
+    AESKeyGeneration,     //
+    RSAOaepKekGeneration, //
+    Master,               //
+    Keyblob,              // f2=crypto revision
+    KeyAreaKey,           // f2=KeyAreaKeyType
+    Titlekek,             //
+    Package2,             //
+    HeaderKek,            //
+    KeyblobMAC,           //
+    ETicketKek,           //
+    ETicketKekek,         //
 };
 
 enum class SDKeyType : u8 {
@@ -60,6 +103,16 @@ enum class SDKeyType : u8 {
     NCA,
 };
 
+enum class BISKeyType : u8 {
+    Crypto,
+    Tweak,
+};
+
+enum class RSAKekType : u8 {
+    Mask0,
+    Seed3,
+};
+
 template <typename KeyType>
 struct KeyIndex {
     KeyType type;
@@ -91,6 +144,8 @@ public:
     Key128 GetKey(S128KeyType id, u64 field1 = 0, u64 field2 = 0) const;
     Key256 GetKey(S256KeyType id, u64 field1 = 0, u64 field2 = 0) const;
 
+    Key256 GetBISKey(u8 partition_id) 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);
 
@@ -100,23 +155,51 @@ public:
     // 8*43 and the private file to exist.
     void DeriveSDSeedLazy();
 
+    bool BaseDeriveNecessary() const;
+    void DeriveBase();
+    void DeriveETicket(PartitionDataManager& data);
+
+    void PopulateFromPartitionData(PartitionDataManager& data);
+
 private:
-    boost::container::flat_map<KeyIndex<S128KeyType>, Key128> s128_keys;
-    boost::container::flat_map<KeyIndex<S256KeyType>, Key256> s256_keys;
+    std::map<KeyIndex<S128KeyType>, Key128> s128_keys;
+    std::map<KeyIndex<S256KeyType>, Key256> s256_keys;
+
+    std::array<std::array<u8, 0xB0>, 0x20> encrypted_keyblobs{};
+    std::array<std::array<u8, 0x90>, 0x20> keyblobs{};
 
     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);
-    template <std::size_t Size>
-    void WriteKeyToFile(bool title_key, std::string_view keyname, const std::array<u8, Size>& key);
+    template <size_t Size>
+    void WriteKeyToFile(KeyCategory category, std::string_view keyname,
+                        const std::array<u8, Size>& key);
+
+    void DeriveGeneralPurposeKeys(std::size_t crypto_revision);
+
+    void SetKeyWrapped(S128KeyType id, Key128 key, u64 field1 = 0, u64 field2 = 0);
+    void SetKeyWrapped(S256KeyType id, Key256 key, u64 field1 = 0, u64 field2 = 0);
 
     static const boost::container::flat_map<std::string, KeyIndex<S128KeyType>> s128_file_id;
     static const boost::container::flat_map<std::string, KeyIndex<S256KeyType>> s256_file_id;
 };
 
 Key128 GenerateKeyEncryptionKey(Key128 source, Key128 master, Key128 kek_seed, Key128 key_seed);
-boost::optional<Key128> DeriveSDSeed();
-Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, const KeyManager& keys);
+Key128 DeriveKeyblobKey(const Key128& sbk, const Key128& tsec, Key128 source);
+Key128 DeriveKeyblobMACKey(const Key128& keyblob_key, const Key128& mac_source);
+Key128 DeriveMasterKey(const std::array<u8, 0x90>& keyblob, const Key128& master_source);
+std::array<u8, 0x90> DecryptKeyblob(const std::array<u8, 0xB0>& encrypted_keyblob,
+                                    const Key128& key);
+
+std::optional<Key128> DeriveSDSeed();
+Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& keys);
+
+std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save);
+
+// Returns a pair of {rights_id, titlekey}. Fails if the ticket has no certificate authority (offset
+// 0x140-0x144 is zero)
+std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
+                                                     const RSAKeyPair<2048>& eticket_extended_key);
 
 } // namespace Core::Crypto
diff --git a/src/core/crypto/partition_data_manager.cpp b/src/core/crypto/partition_data_manager.cpp
new file mode 100644
index 000000000..ed0775444
--- /dev/null
+++ b/src/core/crypto/partition_data_manager.cpp
@@ -0,0 +1,594 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+// NOTE TO FUTURE MAINTAINERS:
+// When a new version of switch cryptography is released,
+// hash the new keyblob source and master key and add the hashes to
+// the arrays below.
+
+#include <algorithm>
+#include <array>
+#include <cctype>
+#include <cstring>
+#include <mbedtls/sha256.h>
+#include "common/assert.h"
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "common/hex_util.h"
+#include "common/logging/log.h"
+#include "common/string_util.h"
+#include "common/swap.h"
+#include "core/crypto/key_manager.h"
+#include "core/crypto/partition_data_manager.h"
+#include "core/crypto/xts_encryption_layer.h"
+#include "core/file_sys/vfs.h"
+#include "core/file_sys/vfs_offset.h"
+
+using namespace Common;
+
+namespace Core::Crypto {
+
+struct Package2Header {
+    std::array<u8, 0x100> signature;
+    Key128 header_ctr;
+    std::array<Key128, 4> section_ctr;
+    u32_le magic;
+    u32_le base_offset;
+    INSERT_PADDING_BYTES(4);
+    u8 version_max;
+    u8 version_min;
+    INSERT_PADDING_BYTES(2);
+    std::array<u32_le, 4> section_size;
+    std::array<u32_le, 4> section_offset;
+    std::array<SHA256Hash, 4> section_hash;
+};
+static_assert(sizeof(Package2Header) == 0x200, "Package2Header has incorrect size.");
+
+struct INIHeader {
+    u32_le magic;
+    u32_le size;
+    u32_le process_count;
+    INSERT_PADDING_BYTES(4);
+};
+static_assert(sizeof(INIHeader) == 0x10, "INIHeader has incorrect size.");
+
+struct SectionHeader {
+    u32_le offset;
+    u32_le size_decompressed;
+    u32_le size_compressed;
+    u32_le attribute;
+};
+static_assert(sizeof(SectionHeader) == 0x10, "SectionHeader has incorrect size.");
+
+struct KIPHeader {
+    u32_le magic;
+    std::array<char, 12> name;
+    u64_le title_id;
+    u32_le category;
+    u8 priority;
+    u8 core;
+    INSERT_PADDING_BYTES(1);
+    u8 flags;
+    std::array<SectionHeader, 6> sections;
+    std::array<u32, 0x20> capabilities;
+};
+static_assert(sizeof(KIPHeader) == 0x100, "KIPHeader has incorrect size.");
+
+const std::array<SHA256Hash, 0x10> source_hashes{
+    "B24BD293259DBC7AC5D63F88E60C59792498E6FC5443402C7FFE87EE8B61A3F0"_array32, // keyblob_mac_key_source
+    "7944862A3A5C31C6720595EFD302245ABD1B54CCDCF33000557681E65C5664A4"_array32, // master_key_source
+    "21E2DF100FC9E094DB51B47B9B1D6E94ED379DB8B547955BEF8FE08D8DD35603"_array32, // package2_key_source
+    "FC02B9D37B42D7A1452E71444F1F700311D1132E301A83B16062E72A78175085"_array32, // aes_kek_generation_source
+    "FBD10056999EDC7ACDB96098E47E2C3606230270D23281E671F0F389FC5BC585"_array32, // aes_key_generation_source
+    "C48B619827986C7F4E3081D59DB2B460C84312650E9A8E6B458E53E8CBCA4E87"_array32, // titlekek_source
+    "04AD66143C726B2A139FB6B21128B46F56C553B2B3887110304298D8D0092D9E"_array32, // key_area_key_application_source
+    "FD434000C8FF2B26F8E9A9D2D2C12F6BE5773CBB9DC86300E1BD99F8EA33A417"_array32, // key_area_key_ocean_source
+    "1F17B1FD51AD1C2379B58F152CA4912EC2106441E51722F38700D5937A1162F7"_array32, // key_area_key_system_source
+    "6B2ED877C2C52334AC51E59ABFA7EC457F4A7D01E46291E9F2EAA45F011D24B7"_array32, // sd_card_kek_source
+    "D482743563D3EA5DCDC3B74E97C9AC8A342164FA041A1DC80F17F6D31E4BC01C"_array32, // sd_card_save_key_source
+    "2E751CECF7D93A2B957BD5FFCB082FD038CC2853219DD3092C6DAB9838F5A7CC"_array32, // sd_card_nca_key_source
+    "1888CAED5551B3EDE01499E87CE0D86827F80820EFB275921055AA4E2ABDFFC2"_array32, // header_kek_source
+    "8F783E46852DF6BE0BA4E19273C4ADBAEE16380043E1B8C418C4089A8BD64AA6"_array32, // header_key_source
+    "D1757E52F1AE55FA882EC690BC6F954AC46A83DC22F277F8806BD55577C6EED7"_array32, // rsa_kek_seed3
+    "FC02B9D37B42D7A1452E71444F1F700311D1132E301A83B16062E72A78175085"_array32, // rsa_kek_mask0
+};
+
+const std::array<SHA256Hash, 0x20> keyblob_source_hashes{
+    "8A06FE274AC491436791FDB388BCDD3AB9943BD4DEF8094418CDAC150FD73786"_array32, // keyblob_key_source_00
+    "2D5CAEB2521FEF70B47E17D6D0F11F8CE2C1E442A979AD8035832C4E9FBCCC4B"_array32, // keyblob_key_source_01
+    "61C5005E713BAE780641683AF43E5F5C0E03671117F702F401282847D2FC6064"_array32, // keyblob_key_source_02
+    "8E9795928E1C4428E1B78F0BE724D7294D6934689C11B190943923B9D5B85903"_array32, // keyblob_key_source_03
+    "95FA33AF95AFF9D9B61D164655B32710ED8D615D46C7D6CC3CC70481B686B402"_array32, // keyblob_key_source_04
+    "3F5BE7B3C8B1ABD8C10B4B703D44766BA08730562C172A4FE0D6B866B3E2DB3E"_array32, // keyblob_key_source_05
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_06
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_07
+
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_08
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_09
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0A
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0B
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0C
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0D
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0E
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0F
+
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_10
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_11
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_12
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_13
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_14
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_15
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_16
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_17
+
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_18
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_19
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1A
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1B
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1C
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1D
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1E
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1F
+};
+
+const std::array<SHA256Hash, 0x20> master_key_hashes{
+    "0EE359BE3C864BB0782E1D70A718A0342C551EED28C369754F9C4F691BECF7CA"_array32, // master_key_00
+    "4FE707B7E4ABDAF727C894AAF13B1351BFE2AC90D875F73B2E20FA94B9CC661E"_array32, // master_key_01
+    "79277C0237A2252EC3DFAC1F7C359C2B3D121E9DB15BB9AB4C2B4408D2F3AE09"_array32, // master_key_02
+    "4F36C565D13325F65EE134073C6A578FFCB0008E02D69400836844EAB7432754"_array32, // master_key_03
+    "75FF1D95D26113550EE6FCC20ACB58E97EDEB3A2FF52543ED5AEC63BDCC3DA50"_array32, // master_key_04
+    "EBE2BCD6704673EC0F88A187BB2AD9F1CC82B718C389425941BDC194DC46B0DD"_array32, // master_key_05
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_06
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_07
+
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_08
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_09
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0A
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0B
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0C
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0D
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0E
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0F
+
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_10
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_11
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_12
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_13
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_14
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_15
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_16
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_17
+
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_18
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_19
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1A
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1B
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1C
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1D
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1E
+    "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1F
+};
+
+static std::vector<u8> DecompressBLZ(const std::vector<u8>& in) {
+    const auto data_size = in.size() - 0xC;
+
+    u32 compressed_size{};
+    u32 init_index{};
+    u32 additional_size{};
+    std::memcpy(&compressed_size, in.data() + data_size, sizeof(u32));
+    std::memcpy(&init_index, in.data() + data_size + 0x4, sizeof(u32));
+    std::memcpy(&additional_size, in.data() + data_size + 0x8, sizeof(u32));
+
+    std::vector<u8> out(in.size() + additional_size);
+
+    if (compressed_size == in.size())
+        std::memcpy(out.data(), in.data() + in.size() - compressed_size, compressed_size);
+    else
+        std::memcpy(out.data(), in.data(), compressed_size);
+
+    auto index = in.size() - init_index;
+    auto out_index = out.size();
+
+    while (out_index > 0) {
+        --index;
+        auto control = in[index];
+        for (size_t i = 0; i < 8; ++i) {
+            if ((control & 0x80) > 0) {
+                ASSERT(index >= 2);
+                index -= 2;
+                u64 segment_offset = in[index] | in[index + 1] << 8;
+                u64 segment_size = ((segment_offset >> 12) & 0xF) + 3;
+                segment_offset &= 0xFFF;
+                segment_offset += 3;
+
+                if (out_index < segment_size)
+                    segment_size = out_index;
+
+                ASSERT(out_index >= segment_size);
+
+                out_index -= segment_size;
+
+                for (size_t j = 0; j < segment_size; ++j) {
+                    ASSERT(out_index + j + segment_offset < out.size());
+                    out[out_index + j] = out[out_index + j + segment_offset];
+                }
+            } else {
+                ASSERT(out_index >= 1);
+                --out_index;
+                --index;
+                out[out_index] = in[index];
+            }
+
+            control <<= 1;
+            if (out_index == 0)
+                return out;
+        }
+    }
+
+    return out;
+}
+
+static u8 CalculateMaxKeyblobSourceHash() {
+    for (s8 i = 0x1F; i >= 0; --i) {
+        if (keyblob_source_hashes[i] != SHA256Hash{})
+            return static_cast<u8>(i + 1);
+    }
+
+    return 0;
+}
+
+const u8 PartitionDataManager::MAX_KEYBLOB_SOURCE_HASH = CalculateMaxKeyblobSourceHash();
+
+template <size_t key_size = 0x10>
+std::array<u8, key_size> FindKeyFromHex(const std::vector<u8>& binary,
+                                        const std::array<u8, 0x20>& hash) {
+    if (binary.size() < key_size)
+        return {};
+
+    std::array<u8, 0x20> temp{};
+    for (size_t i = 0; i < binary.size() - key_size; ++i) {
+        mbedtls_sha256(binary.data() + i, key_size, temp.data(), 0);
+
+        if (temp != hash)
+            continue;
+
+        std::array<u8, key_size> out{};
+        std::memcpy(out.data(), binary.data() + i, key_size);
+        return out;
+    }
+
+    return {};
+}
+
+std::array<u8, 16> FindKeyFromHex16(const std::vector<u8>& binary, std::array<u8, 32> hash) {
+    return FindKeyFromHex<0x10>(binary, hash);
+}
+
+static std::array<Key128, 0x20> FindEncryptedMasterKeyFromHex(const std::vector<u8>& binary,
+                                                              const Key128& key) {
+    if (binary.size() < 0x10)
+        return {};
+
+    SHA256Hash temp{};
+    Key128 dec_temp{};
+    std::array<Key128, 0x20> out{};
+    AESCipher<Key128> cipher(key, Mode::ECB);
+    for (size_t i = 0; i < binary.size() - 0x10; ++i) {
+        cipher.Transcode(binary.data() + i, dec_temp.size(), dec_temp.data(), Op::Decrypt);
+        mbedtls_sha256(dec_temp.data(), dec_temp.size(), temp.data(), 0);
+
+        for (size_t k = 0; k < out.size(); ++k) {
+            if (temp == master_key_hashes[k]) {
+                out[k] = dec_temp;
+                break;
+            }
+        }
+    }
+
+    return out;
+}
+
+FileSys::VirtualFile FindFileInDirWithNames(const FileSys::VirtualDir& dir,
+                                            const std::string& name) {
+    auto upper = name;
+    std::transform(upper.begin(), upper.end(), upper.begin(), [](u8 c) { return std::toupper(c); });
+    for (const auto& fname : {name, name + ".bin", upper, upper + ".BIN"}) {
+        if (dir->GetFile(fname) != nullptr)
+            return dir->GetFile(fname);
+    }
+
+    return nullptr;
+}
+
+PartitionDataManager::PartitionDataManager(const FileSys::VirtualDir& sysdata_dir)
+    : boot0(FindFileInDirWithNames(sysdata_dir, "BOOT0")),
+      fuses(FindFileInDirWithNames(sysdata_dir, "fuses")),
+      kfuses(FindFileInDirWithNames(sysdata_dir, "kfuses")),
+      package2({
+          FindFileInDirWithNames(sysdata_dir, "BCPKG2-1-Normal-Main"),
+          FindFileInDirWithNames(sysdata_dir, "BCPKG2-2-Normal-Sub"),
+          FindFileInDirWithNames(sysdata_dir, "BCPKG2-3-SafeMode-Main"),
+          FindFileInDirWithNames(sysdata_dir, "BCPKG2-4-SafeMode-Sub"),
+          FindFileInDirWithNames(sysdata_dir, "BCPKG2-5-Repair-Main"),
+          FindFileInDirWithNames(sysdata_dir, "BCPKG2-6-Repair-Sub"),
+      }),
+      prodinfo(FindFileInDirWithNames(sysdata_dir, "PRODINFO")),
+      secure_monitor(FindFileInDirWithNames(sysdata_dir, "secmon")),
+      package1_decrypted(FindFileInDirWithNames(sysdata_dir, "pkg1_decr")),
+      secure_monitor_bytes(secure_monitor == nullptr ? std::vector<u8>{}
+                                                     : secure_monitor->ReadAllBytes()),
+      package1_decrypted_bytes(package1_decrypted == nullptr ? std::vector<u8>{}
+                                                             : package1_decrypted->ReadAllBytes()) {
+}
+
+PartitionDataManager::~PartitionDataManager() = default;
+
+bool PartitionDataManager::HasBoot0() const {
+    return boot0 != nullptr;
+}
+
+FileSys::VirtualFile PartitionDataManager::GetBoot0Raw() const {
+    return boot0;
+}
+
+PartitionDataManager::EncryptedKeyBlob PartitionDataManager::GetEncryptedKeyblob(
+    std::size_t index) const {
+    if (HasBoot0() && index < NUM_ENCRYPTED_KEYBLOBS)
+        return GetEncryptedKeyblobs()[index];
+    return {};
+}
+
+PartitionDataManager::EncryptedKeyBlobs PartitionDataManager::GetEncryptedKeyblobs() const {
+    if (!HasBoot0())
+        return {};
+
+    EncryptedKeyBlobs out{};
+    for (size_t i = 0; i < out.size(); ++i)
+        boot0->Read(out[i].data(), out[i].size(), 0x180000 + i * 0x200);
+    return out;
+}
+
+std::vector<u8> PartitionDataManager::GetSecureMonitor() const {
+    return secure_monitor_bytes;
+}
+
+std::array<u8, 16> PartitionDataManager::GetPackage2KeySource() const {
+    return FindKeyFromHex(secure_monitor_bytes, source_hashes[2]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetAESKekGenerationSource() const {
+    return FindKeyFromHex(secure_monitor_bytes, source_hashes[3]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetTitlekekSource() const {
+    return FindKeyFromHex(secure_monitor_bytes, source_hashes[5]);
+}
+
+std::array<std::array<u8, 16>, 32> PartitionDataManager::GetTZMasterKeys(
+    std::array<u8, 0x10> master_key) const {
+    return FindEncryptedMasterKeyFromHex(secure_monitor_bytes, master_key);
+}
+
+std::array<u8, 16> PartitionDataManager::GetRSAKekSeed3() const {
+    return FindKeyFromHex(secure_monitor_bytes, source_hashes[14]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetRSAKekMask0() const {
+    return FindKeyFromHex(secure_monitor_bytes, source_hashes[15]);
+}
+
+std::vector<u8> PartitionDataManager::GetPackage1Decrypted() const {
+    return package1_decrypted_bytes;
+}
+
+std::array<u8, 16> PartitionDataManager::GetMasterKeySource() const {
+    return FindKeyFromHex(package1_decrypted_bytes, source_hashes[1]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetKeyblobMACKeySource() const {
+    return FindKeyFromHex(package1_decrypted_bytes, source_hashes[0]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetKeyblobKeySource(std::size_t revision) const {
+    if (keyblob_source_hashes[revision] == SHA256Hash{}) {
+        LOG_WARNING(Crypto,
+                    "No keyblob source hash for crypto revision {:02X}! Cannot derive keys...",
+                    revision);
+    }
+    return FindKeyFromHex(package1_decrypted_bytes, keyblob_source_hashes[revision]);
+}
+
+bool PartitionDataManager::HasFuses() const {
+    return fuses != nullptr;
+}
+
+FileSys::VirtualFile PartitionDataManager::GetFusesRaw() const {
+    return fuses;
+}
+
+std::array<u8, 16> PartitionDataManager::GetSecureBootKey() const {
+    if (!HasFuses())
+        return {};
+    Key128 out{};
+    fuses->Read(out.data(), out.size(), 0xA4);
+    return out;
+}
+
+bool PartitionDataManager::HasKFuses() const {
+    return kfuses != nullptr;
+}
+
+FileSys::VirtualFile PartitionDataManager::GetKFusesRaw() const {
+    return kfuses;
+}
+
+bool PartitionDataManager::HasPackage2(Package2Type type) const {
+    return package2.at(static_cast<size_t>(type)) != nullptr;
+}
+
+FileSys::VirtualFile PartitionDataManager::GetPackage2Raw(Package2Type type) const {
+    return package2.at(static_cast<size_t>(type));
+}
+
+bool AttemptDecrypt(const std::array<u8, 16>& key, Package2Header& header) {
+
+    const std::vector<u8> iv(header.header_ctr.begin(), header.header_ctr.end());
+    Package2Header temp = header;
+    AESCipher<Key128> cipher(key, Mode::CTR);
+    cipher.SetIV(iv);
+    cipher.Transcode(&temp.header_ctr, sizeof(Package2Header) - 0x100, &temp.header_ctr,
+                     Op::Decrypt);
+    if (temp.magic == Common::MakeMagic('P', 'K', '2', '1')) {
+        header = temp;
+        return true;
+    }
+
+    return false;
+}
+
+void PartitionDataManager::DecryptPackage2(const std::array<Key128, 0x20>& package2_keys,
+                                           Package2Type type) {
+    FileSys::VirtualFile file = std::make_shared<FileSys::OffsetVfsFile>(
+        package2[static_cast<size_t>(type)],
+        package2[static_cast<size_t>(type)]->GetSize() - 0x4000, 0x4000);
+
+    Package2Header header{};
+    if (file->ReadObject(&header) != sizeof(Package2Header))
+        return;
+
+    std::size_t revision = 0xFF;
+    if (header.magic != Common::MakeMagic('P', 'K', '2', '1')) {
+        for (std::size_t i = 0; i < package2_keys.size(); ++i) {
+            if (AttemptDecrypt(package2_keys[i], header)) {
+                revision = i;
+            }
+        }
+    }
+
+    if (header.magic != Common::MakeMagic('P', 'K', '2', '1'))
+        return;
+
+    const auto a = std::make_shared<FileSys::OffsetVfsFile>(
+        file, header.section_size[1], header.section_size[0] + sizeof(Package2Header));
+
+    auto c = a->ReadAllBytes();
+
+    AESCipher<Key128> cipher(package2_keys[revision], Mode::CTR);
+    cipher.SetIV({header.section_ctr[1].begin(), header.section_ctr[1].end()});
+    cipher.Transcode(c.data(), c.size(), c.data(), Op::Decrypt);
+
+    INIHeader ini;
+    std::memcpy(&ini, c.data(), sizeof(INIHeader));
+    if (ini.magic != Common::MakeMagic('I', 'N', 'I', '1'))
+        return;
+
+    u64 offset = sizeof(INIHeader);
+    for (size_t i = 0; i < ini.process_count; ++i) {
+        KIPHeader kip;
+        std::memcpy(&kip, c.data() + offset, sizeof(KIPHeader));
+        if (kip.magic != Common::MakeMagic('K', 'I', 'P', '1'))
+            return;
+
+        const auto name =
+            Common::StringFromFixedZeroTerminatedBuffer(kip.name.data(), kip.name.size());
+
+        if (name != "FS" && name != "spl") {
+            offset += sizeof(KIPHeader) + kip.sections[0].size_compressed +
+                      kip.sections[1].size_compressed + kip.sections[2].size_compressed;
+            continue;
+        }
+
+        const u64 initial_offset = sizeof(KIPHeader) + offset;
+        const auto text_begin = c.cbegin() + initial_offset;
+        const auto text_end = text_begin + kip.sections[0].size_compressed;
+        const std::vector<u8> text = DecompressBLZ({text_begin, text_end});
+
+        const auto rodata_end = text_end + kip.sections[1].size_compressed;
+        const std::vector<u8> rodata = DecompressBLZ({text_end, rodata_end});
+
+        const auto data_end = rodata_end + kip.sections[2].size_compressed;
+        const std::vector<u8> data = DecompressBLZ({rodata_end, data_end});
+
+        std::vector<u8> out;
+        out.reserve(text.size() + rodata.size() + data.size());
+        out.insert(out.end(), text.begin(), text.end());
+        out.insert(out.end(), rodata.begin(), rodata.end());
+        out.insert(out.end(), data.begin(), data.end());
+
+        offset += sizeof(KIPHeader) + kip.sections[0].size_compressed +
+                  kip.sections[1].size_compressed + kip.sections[2].size_compressed;
+
+        if (name == "FS")
+            package2_fs[static_cast<size_t>(type)] = std::move(out);
+        else if (name == "spl")
+            package2_spl[static_cast<size_t>(type)] = std::move(out);
+    }
+}
+
+const std::vector<u8>& PartitionDataManager::GetPackage2FSDecompressed(Package2Type type) const {
+    return package2_fs.at(static_cast<size_t>(type));
+}
+
+std::array<u8, 16> PartitionDataManager::GetKeyAreaKeyApplicationSource(Package2Type type) const {
+    return FindKeyFromHex(package2_fs.at(static_cast<size_t>(type)), source_hashes[6]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetKeyAreaKeyOceanSource(Package2Type type) const {
+    return FindKeyFromHex(package2_fs.at(static_cast<size_t>(type)), source_hashes[7]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetKeyAreaKeySystemSource(Package2Type type) const {
+    return FindKeyFromHex(package2_fs.at(static_cast<size_t>(type)), source_hashes[8]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetSDKekSource(Package2Type type) const {
+    return FindKeyFromHex(package2_fs.at(static_cast<size_t>(type)), source_hashes[9]);
+}
+
+std::array<u8, 32> PartitionDataManager::GetSDSaveKeySource(Package2Type type) const {
+    return FindKeyFromHex<0x20>(package2_fs.at(static_cast<size_t>(type)), source_hashes[10]);
+}
+
+std::array<u8, 32> PartitionDataManager::GetSDNCAKeySource(Package2Type type) const {
+    return FindKeyFromHex<0x20>(package2_fs.at(static_cast<size_t>(type)), source_hashes[11]);
+}
+
+std::array<u8, 16> PartitionDataManager::GetHeaderKekSource(Package2Type type) const {
+    return FindKeyFromHex(package2_fs.at(static_cast<size_t>(type)), source_hashes[12]);
+}
+
+std::array<u8, 32> PartitionDataManager::GetHeaderKeySource(Package2Type type) const {
+    return FindKeyFromHex<0x20>(package2_fs.at(static_cast<size_t>(type)), source_hashes[13]);
+}
+
+const std::vector<u8>& PartitionDataManager::GetPackage2SPLDecompressed(Package2Type type) const {
+    return package2_spl.at(static_cast<size_t>(type));
+}
+
+std::array<u8, 16> PartitionDataManager::GetAESKeyGenerationSource(Package2Type type) const {
+    return FindKeyFromHex(package2_spl.at(static_cast<size_t>(type)), source_hashes[4]);
+}
+
+bool PartitionDataManager::HasProdInfo() const {
+    return prodinfo != nullptr;
+}
+
+FileSys::VirtualFile PartitionDataManager::GetProdInfoRaw() const {
+    return prodinfo;
+}
+
+void PartitionDataManager::DecryptProdInfo(std::array<u8, 0x20> bis_key) {
+    if (prodinfo == nullptr)
+        return;
+
+    prodinfo_decrypted = std::make_shared<XTSEncryptionLayer>(prodinfo, bis_key);
+}
+
+std::array<u8, 576> PartitionDataManager::GetETicketExtendedKek() const {
+    std::array<u8, 0x240> out{};
+    if (prodinfo_decrypted != nullptr)
+        prodinfo_decrypted->Read(out.data(), out.size(), 0x3890);
+    return out;
+}
+} // namespace Core::Crypto
diff --git a/src/core/crypto/partition_data_manager.h b/src/core/crypto/partition_data_manager.h
new file mode 100644
index 000000000..0ad007c72
--- /dev/null
+++ b/src/core/crypto/partition_data_manager.h
@@ -0,0 +1,109 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <vector>
+#include "common/common_types.h"
+#include "core/file_sys/vfs_types.h"
+
+namespace Core::Crypto {
+
+enum class Package2Type {
+    NormalMain,
+    NormalSub,
+    SafeModeMain,
+    SafeModeSub,
+    RepairMain,
+    RepairSub,
+};
+
+class PartitionDataManager {
+public:
+    static const u8 MAX_KEYBLOB_SOURCE_HASH;
+    static constexpr std::size_t NUM_ENCRYPTED_KEYBLOBS = 32;
+    static constexpr std::size_t ENCRYPTED_KEYBLOB_SIZE = 0xB0;
+
+    using EncryptedKeyBlob = std::array<u8, ENCRYPTED_KEYBLOB_SIZE>;
+    using EncryptedKeyBlobs = std::array<EncryptedKeyBlob, NUM_ENCRYPTED_KEYBLOBS>;
+
+    explicit PartitionDataManager(const FileSys::VirtualDir& sysdata_dir);
+    ~PartitionDataManager();
+
+    // BOOT0
+    bool HasBoot0() const;
+    FileSys::VirtualFile GetBoot0Raw() const;
+    EncryptedKeyBlob GetEncryptedKeyblob(std::size_t index) const;
+    EncryptedKeyBlobs GetEncryptedKeyblobs() const;
+    std::vector<u8> GetSecureMonitor() const;
+    std::array<u8, 0x10> GetPackage2KeySource() const;
+    std::array<u8, 0x10> GetAESKekGenerationSource() const;
+    std::array<u8, 0x10> GetTitlekekSource() const;
+    std::array<std::array<u8, 0x10>, 0x20> GetTZMasterKeys(std::array<u8, 0x10> master_key) const;
+    std::array<u8, 0x10> GetRSAKekSeed3() const;
+    std::array<u8, 0x10> GetRSAKekMask0() const;
+    std::vector<u8> GetPackage1Decrypted() const;
+    std::array<u8, 0x10> GetMasterKeySource() const;
+    std::array<u8, 0x10> GetKeyblobMACKeySource() const;
+    std::array<u8, 0x10> GetKeyblobKeySource(std::size_t revision) const;
+
+    // Fuses
+    bool HasFuses() const;
+    FileSys::VirtualFile GetFusesRaw() const;
+    std::array<u8, 0x10> GetSecureBootKey() const;
+
+    // K-Fuses
+    bool HasKFuses() const;
+    FileSys::VirtualFile GetKFusesRaw() const;
+
+    // Package2
+    bool HasPackage2(Package2Type type = Package2Type::NormalMain) const;
+    FileSys::VirtualFile GetPackage2Raw(Package2Type type = Package2Type::NormalMain) const;
+    void DecryptPackage2(const std::array<std::array<u8, 16>, 0x20>& package2_keys,
+                         Package2Type type);
+    const std::vector<u8>& GetPackage2FSDecompressed(
+        Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x10> GetKeyAreaKeyApplicationSource(
+        Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x10> GetKeyAreaKeyOceanSource(
+        Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x10> GetKeyAreaKeySystemSource(
+        Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x10> GetSDKekSource(Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x20> GetSDSaveKeySource(Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x20> GetSDNCAKeySource(Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x10> GetHeaderKekSource(Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x20> GetHeaderKeySource(Package2Type type = Package2Type::NormalMain) const;
+    const std::vector<u8>& GetPackage2SPLDecompressed(
+        Package2Type type = Package2Type::NormalMain) const;
+    std::array<u8, 0x10> GetAESKeyGenerationSource(
+        Package2Type type = Package2Type::NormalMain) const;
+
+    // PRODINFO
+    bool HasProdInfo() const;
+    FileSys::VirtualFile GetProdInfoRaw() const;
+    void DecryptProdInfo(std::array<u8, 0x20> bis_key);
+    std::array<u8, 0x240> GetETicketExtendedKek() const;
+
+private:
+    FileSys::VirtualFile boot0;
+    FileSys::VirtualFile fuses;
+    FileSys::VirtualFile kfuses;
+    std::array<FileSys::VirtualFile, 6> package2;
+    FileSys::VirtualFile prodinfo;
+    FileSys::VirtualFile secure_monitor;
+    FileSys::VirtualFile package1_decrypted;
+
+    // Processed
+    std::array<FileSys::VirtualFile, 6> package2_decrypted;
+    FileSys::VirtualFile prodinfo_decrypted;
+    std::vector<u8> secure_monitor_bytes;
+    std::vector<u8> package1_decrypted_bytes;
+    std::array<std::vector<u8>, 6> package2_fs;
+    std::array<std::vector<u8>, 6> package2_spl;
+};
+
+std::array<u8, 0x10> FindKeyFromHex16(const std::vector<u8>& binary, std::array<u8, 0x20> hash);
+
+} // namespace Core::Crypto
diff --git a/src/core/file_sys/bis_factory.cpp b/src/core/file_sys/bis_factory.cpp
index 6102ef476..76a2b7e86 100644
--- a/src/core/file_sys/bis_factory.cpp
+++ b/src/core/file_sys/bis_factory.cpp
@@ -10,19 +10,19 @@ namespace FileSys {
 
 BISFactory::BISFactory(VirtualDir nand_root_, VirtualDir load_root_)
     : nand_root(std::move(nand_root_)), load_root(std::move(load_root_)),
-      sysnand_cache(std::make_shared<RegisteredCache>(
+      sysnand_cache(std::make_unique<RegisteredCache>(
           GetOrCreateDirectoryRelative(nand_root, "/system/Contents/registered"))),
-      usrnand_cache(std::make_shared<RegisteredCache>(
+      usrnand_cache(std::make_unique<RegisteredCache>(
           GetOrCreateDirectoryRelative(nand_root, "/user/Contents/registered"))) {}
 
 BISFactory::~BISFactory() = default;
 
-std::shared_ptr<RegisteredCache> BISFactory::GetSystemNANDContents() const {
-    return sysnand_cache;
+RegisteredCache* BISFactory::GetSystemNANDContents() const {
+    return sysnand_cache.get();
 }
 
-std::shared_ptr<RegisteredCache> BISFactory::GetUserNANDContents() const {
-    return usrnand_cache;
+RegisteredCache* BISFactory::GetUserNANDContents() const {
+    return usrnand_cache.get();
 }
 
 VirtualDir BISFactory::GetModificationLoadRoot(u64 title_id) const {
diff --git a/src/core/file_sys/bis_factory.h b/src/core/file_sys/bis_factory.h
index c352e0925..364d309bd 100644
--- a/src/core/file_sys/bis_factory.h
+++ b/src/core/file_sys/bis_factory.h
@@ -20,8 +20,8 @@ public:
     explicit BISFactory(VirtualDir nand_root, VirtualDir load_root);
     ~BISFactory();
 
-    std::shared_ptr<RegisteredCache> GetSystemNANDContents() const;
-    std::shared_ptr<RegisteredCache> GetUserNANDContents() const;
+    RegisteredCache* GetSystemNANDContents() const;
+    RegisteredCache* GetUserNANDContents() const;
 
     VirtualDir GetModificationLoadRoot(u64 title_id) const;
 
@@ -29,8 +29,8 @@ private:
     VirtualDir nand_root;
     VirtualDir load_root;
 
-    std::shared_ptr<RegisteredCache> sysnand_cache;
-    std::shared_ptr<RegisteredCache> usrnand_cache;
+    std::unique_ptr<RegisteredCache> sysnand_cache;
+    std::unique_ptr<RegisteredCache> usrnand_cache;
 };
 
 } // namespace FileSys
diff --git a/src/core/file_sys/card_image.cpp b/src/core/file_sys/card_image.cpp
index 8f5142a07..1ece55731 100644
--- a/src/core/file_sys/card_image.cpp
+++ b/src/core/file_sys/card_image.cpp
@@ -122,14 +122,16 @@ u64 XCI::GetProgramTitleID() const {
     return secure_partition->GetProgramTitleID();
 }
 
-std::shared_ptr<NCA> XCI::GetProgramNCA() const {
-    return program;
+bool XCI::HasProgramNCA() const {
+    return program != nullptr;
 }
 
 VirtualFile XCI::GetProgramNCAFile() const {
-    if (GetProgramNCA() == nullptr)
+    if (!HasProgramNCA()) {
         return nullptr;
-    return GetProgramNCA()->GetBaseFile();
+    }
+
+    return program->GetBaseFile();
 }
 
 const std::vector<std::shared_ptr<NCA>>& XCI::GetNCAs() const {
@@ -166,10 +168,6 @@ VirtualDir XCI::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool XCI::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
-
 Loader::ResultStatus XCI::AddNCAFromPartition(XCIPartition part) {
     if (partitions[static_cast<std::size_t>(part)] == nullptr) {
         return Loader::ResultStatus::ErrorXCIMissingPartition;
diff --git a/src/core/file_sys/card_image.h b/src/core/file_sys/card_image.h
index ce514dfa0..8f62571cf 100644
--- a/src/core/file_sys/card_image.h
+++ b/src/core/file_sys/card_image.h
@@ -80,7 +80,7 @@ public:
 
     u64 GetProgramTitleID() const;
 
-    std::shared_ptr<NCA> GetProgramNCA() const;
+    bool HasProgramNCA() const;
     VirtualFile GetProgramNCAFile() const;
     const std::vector<std::shared_ptr<NCA>>& GetNCAs() const;
     std::shared_ptr<NCA> GetNCAByType(NCAContentType type) const;
@@ -94,9 +94,6 @@ public:
 
     VirtualDir GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     Loader::ResultStatus AddNCAFromPartition(XCIPartition part);
 
diff --git a/src/core/file_sys/content_archive.cpp b/src/core/file_sys/content_archive.cpp
index aa1b3c17d..b46fe893c 100644
--- a/src/core/file_sys/content_archive.cpp
+++ b/src/core/file_sys/content_archive.cpp
@@ -4,10 +4,9 @@
 
 #include <algorithm>
 #include <cstring>
+#include <optional>
 #include <utility>
 
-#include <boost/optional.hpp>
-
 #include "common/logging/log.h"
 #include "core/crypto/aes_util.h"
 #include "core/crypto/ctr_encryption_layer.h"
@@ -97,11 +96,288 @@ union NCASectionHeader {
 };
 static_assert(sizeof(NCASectionHeader) == 0x200, "NCASectionHeader has incorrect size.");
 
-bool IsValidNCA(const NCAHeader& header) {
+static bool IsValidNCA(const NCAHeader& header) {
     // TODO(DarkLordZach): Add NCA2/NCA0 support.
     return header.magic == Common::MakeMagic('N', 'C', 'A', '3');
 }
 
+NCA::NCA(VirtualFile file_, VirtualFile bktr_base_romfs_, u64 bktr_base_ivfc_offset)
+    : file(std::move(file_)), bktr_base_romfs(std::move(bktr_base_romfs_)) {
+    if (file == nullptr) {
+        status = Loader::ResultStatus::ErrorNullFile;
+        return;
+    }
+
+    if (sizeof(NCAHeader) != file->ReadObject(&header)) {
+        LOG_ERROR(Loader, "File reader errored out during header read.");
+        status = Loader::ResultStatus::ErrorBadNCAHeader;
+        return;
+    }
+
+    if (!HandlePotentialHeaderDecryption()) {
+        return;
+    }
+
+    has_rights_id = std::any_of(header.rights_id.begin(), header.rights_id.end(),
+                                [](char c) { return c != '\0'; });
+
+    const std::vector<NCASectionHeader> sections = ReadSectionHeaders();
+    is_update = std::any_of(sections.begin(), sections.end(), [](const NCASectionHeader& header) {
+        return header.raw.header.crypto_type == NCASectionCryptoType::BKTR;
+    });
+
+    if (!ReadSections(sections, bktr_base_ivfc_offset)) {
+        return;
+    }
+
+    status = Loader::ResultStatus::Success;
+}
+
+NCA::~NCA() = default;
+
+bool NCA::CheckSupportedNCA(const NCAHeader& nca_header) {
+    if (nca_header.magic == Common::MakeMagic('N', 'C', 'A', '2')) {
+        status = Loader::ResultStatus::ErrorNCA2;
+        return false;
+    }
+
+    if (nca_header.magic == Common::MakeMagic('N', 'C', 'A', '0')) {
+        status = Loader::ResultStatus::ErrorNCA0;
+        return false;
+    }
+
+    return true;
+}
+
+bool NCA::HandlePotentialHeaderDecryption() {
+    if (IsValidNCA(header)) {
+        return true;
+    }
+
+    if (!CheckSupportedNCA(header)) {
+        return false;
+    }
+
+    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 (!CheckSupportedNCA(dec_header)) {
+            return false;
+        }
+
+        if (keys.HasKey(Core::Crypto::S256KeyType::Header)) {
+            status = Loader::ResultStatus::ErrorIncorrectHeaderKey;
+        } else {
+            status = Loader::ResultStatus::ErrorMissingHeaderKey;
+        }
+        return false;
+    }
+
+    return true;
+}
+
+std::vector<NCASectionHeader> NCA::ReadSectionHeaders() const {
+    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);
+    }
+
+    return sections;
+}
+
+bool NCA::ReadSections(const std::vector<NCASectionHeader>& sections, u64 bktr_base_ivfc_offset) {
+    for (std::size_t i = 0; i < sections.size(); ++i) {
+        const auto& section = sections[i];
+
+        if (section.raw.header.filesystem_type == NCASectionFilesystemType::ROMFS) {
+            if (!ReadRomFSSection(section, header.section_tables[i], bktr_base_ivfc_offset)) {
+                return false;
+            }
+        } else if (section.raw.header.filesystem_type == NCASectionFilesystemType::PFS0) {
+            if (!ReadPFS0Section(section, header.section_tables[i])) {
+                return false;
+            }
+        }
+    }
+
+    return true;
+}
+
+bool NCA::ReadRomFSSection(const NCASectionHeader& section, const NCASectionTableEntry& entry,
+                           u64 bktr_base_ivfc_offset) {
+    const std::size_t base_offset = entry.media_offset * MEDIA_OFFSET_MULTIPLIER;
+    ivfc_offset = section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].offset;
+    const std::size_t romfs_offset = base_offset + ivfc_offset;
+    const std::size_t romfs_size = section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].size;
+    auto raw = std::make_shared<OffsetVfsFile>(file, romfs_size, romfs_offset);
+    auto dec = Decrypt(section, raw, romfs_offset);
+
+    if (dec == nullptr) {
+        if (status != Loader::ResultStatus::Success)
+            return false;
+        if (has_rights_id)
+            status = Loader::ResultStatus::ErrorIncorrectTitlekeyOrTitlekek;
+        else
+            status = Loader::ResultStatus::ErrorIncorrectKeyAreaKey;
+        return false;
+    }
+
+    if (section.raw.header.crypto_type == NCASectionCryptoType::BKTR) {
+        if (section.bktr.relocation.magic != Common::MakeMagic('B', 'K', 'T', 'R') ||
+            section.bktr.subsection.magic != Common::MakeMagic('B', 'K', 'T', 'R')) {
+            status = Loader::ResultStatus::ErrorBadBKTRHeader;
+            return false;
+        }
+
+        if (section.bktr.relocation.offset + section.bktr.relocation.size !=
+            section.bktr.subsection.offset) {
+            status = Loader::ResultStatus::ErrorBKTRSubsectionNotAfterRelocation;
+            return false;
+        }
+
+        const u64 size = MEDIA_OFFSET_MULTIPLIER * (entry.media_end_offset - entry.media_offset);
+        if (section.bktr.subsection.offset + section.bktr.subsection.size != size) {
+            status = Loader::ResultStatus::ErrorBKTRSubsectionNotAtEnd;
+            return false;
+        }
+
+        const u64 offset = section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].offset;
+        RelocationBlock relocation_block{};
+        if (dec->ReadObject(&relocation_block, section.bktr.relocation.offset - offset) !=
+            sizeof(RelocationBlock)) {
+            status = Loader::ResultStatus::ErrorBadRelocationBlock;
+            return false;
+        }
+        SubsectionBlock subsection_block{};
+        if (dec->ReadObject(&subsection_block, section.bktr.subsection.offset - offset) !=
+            sizeof(RelocationBlock)) {
+            status = Loader::ResultStatus::ErrorBadSubsectionBlock;
+            return false;
+        }
+
+        std::vector<RelocationBucketRaw> relocation_buckets_raw(
+            (section.bktr.relocation.size - sizeof(RelocationBlock)) / sizeof(RelocationBucketRaw));
+        if (dec->ReadBytes(relocation_buckets_raw.data(),
+                           section.bktr.relocation.size - sizeof(RelocationBlock),
+                           section.bktr.relocation.offset + sizeof(RelocationBlock) - offset) !=
+            section.bktr.relocation.size - sizeof(RelocationBlock)) {
+            status = Loader::ResultStatus::ErrorBadRelocationBuckets;
+            return false;
+        }
+
+        std::vector<SubsectionBucketRaw> subsection_buckets_raw(
+            (section.bktr.subsection.size - sizeof(SubsectionBlock)) / sizeof(SubsectionBucketRaw));
+        if (dec->ReadBytes(subsection_buckets_raw.data(),
+                           section.bktr.subsection.size - sizeof(SubsectionBlock),
+                           section.bktr.subsection.offset + sizeof(SubsectionBlock) - offset) !=
+            section.bktr.subsection.size - sizeof(SubsectionBlock)) {
+            status = Loader::ResultStatus::ErrorBadSubsectionBuckets;
+            return false;
+        }
+
+        std::vector<RelocationBucket> relocation_buckets(relocation_buckets_raw.size());
+        std::transform(relocation_buckets_raw.begin(), relocation_buckets_raw.end(),
+                       relocation_buckets.begin(), &ConvertRelocationBucketRaw);
+        std::vector<SubsectionBucket> subsection_buckets(subsection_buckets_raw.size());
+        std::transform(subsection_buckets_raw.begin(), subsection_buckets_raw.end(),
+                       subsection_buckets.begin(), &ConvertSubsectionBucketRaw);
+
+        u32 ctr_low;
+        std::memcpy(&ctr_low, section.raw.section_ctr.data(), sizeof(ctr_low));
+        subsection_buckets.back().entries.push_back({section.bktr.relocation.offset, {0}, ctr_low});
+        subsection_buckets.back().entries.push_back({size, {0}, 0});
+
+        std::optional<Core::Crypto::Key128> key = {};
+        if (encrypted) {
+            if (has_rights_id) {
+                status = Loader::ResultStatus::Success;
+                key = GetTitlekey();
+                if (!key) {
+                    status = Loader::ResultStatus::ErrorMissingTitlekey;
+                    return false;
+                }
+            } else {
+                key = GetKeyAreaKey(NCASectionCryptoType::BKTR);
+                if (!key) {
+                    status = Loader::ResultStatus::ErrorMissingKeyAreaKey;
+                    return false;
+                }
+            }
+        }
+
+        if (bktr_base_romfs == nullptr) {
+            status = Loader::ResultStatus::ErrorMissingBKTRBaseRomFS;
+            return false;
+        }
+
+        auto bktr = std::make_shared<BKTR>(
+            bktr_base_romfs, std::make_shared<OffsetVfsFile>(file, romfs_size, base_offset),
+            relocation_block, relocation_buckets, subsection_block, subsection_buckets, encrypted,
+            encrypted ? *key : Core::Crypto::Key128{}, base_offset, bktr_base_ivfc_offset,
+            section.raw.section_ctr);
+
+        // BKTR applies to entire IVFC, so make an offset version to level 6
+        files.push_back(std::make_shared<OffsetVfsFile>(
+            bktr, romfs_size, section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].offset));
+    } else {
+        files.push_back(std::move(dec));
+    }
+
+    romfs = files.back();
+    return true;
+}
+
+bool NCA::ReadPFS0Section(const NCASectionHeader& section, const NCASectionTableEntry& entry) {
+    const u64 offset = (static_cast<u64>(entry.media_offset) * MEDIA_OFFSET_MULTIPLIER) +
+                       section.pfs0.pfs0_header_offset;
+    const u64 size = MEDIA_OFFSET_MULTIPLIER * (entry.media_end_offset - entry.media_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.push_back(std::move(npfs));
+            if (IsDirectoryExeFS(dirs.back()))
+                exefs = dirs.back();
+        } else {
+            if (has_rights_id)
+                status = Loader::ResultStatus::ErrorIncorrectTitlekeyOrTitlekek;
+            else
+                status = Loader::ResultStatus::ErrorIncorrectKeyAreaKey;
+            return false;
+        }
+    } else {
+        if (status != Loader::ResultStatus::Success)
+            return false;
+        if (has_rights_id)
+            status = Loader::ResultStatus::ErrorIncorrectTitlekeyOrTitlekek;
+        else
+            status = Loader::ResultStatus::ErrorIncorrectKeyAreaKey;
+        return false;
+    }
+
+    return true;
+}
+
 u8 NCA::GetCryptoRevision() const {
     u8 master_key_id = header.crypto_type;
     if (header.crypto_type_2 > master_key_id)
@@ -111,11 +387,11 @@ u8 NCA::GetCryptoRevision() const {
     return master_key_id;
 }
 
-boost::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const {
+std::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const {
     const auto master_key_id = GetCryptoRevision();
 
     if (!keys.HasKey(Core::Crypto::S128KeyType::KeyArea, master_key_id, header.key_index))
-        return boost::none;
+        return {};
 
     std::vector<u8> key_area(header.key_area.begin(), header.key_area.end());
     Core::Crypto::AESCipher<Core::Crypto::Key128> cipher(
@@ -133,31 +409,31 @@ boost::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType ty
                      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}",
+    LOG_TRACE(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;
 }
 
-boost::optional<Core::Crypto::Key128> NCA::GetTitlekey() {
+std::optional<Core::Crypto::Key128> NCA::GetTitlekey() {
     const auto master_key_id = GetCryptoRevision();
 
     u128 rights_id{};
     memcpy(rights_id.data(), header.rights_id.data(), 16);
     if (rights_id == u128{}) {
         status = Loader::ResultStatus::ErrorInvalidRightsID;
-        return boost::none;
+        return {};
     }
 
     auto titlekey = keys.GetKey(Core::Crypto::S128KeyType::Titlekey, rights_id[1], rights_id[0]);
     if (titlekey == Core::Crypto::Key128{}) {
         status = Loader::ResultStatus::ErrorMissingTitlekey;
-        return boost::none;
+        return {};
     }
 
     if (!keys.HasKey(Core::Crypto::S128KeyType::Titlekek, master_key_id)) {
         status = Loader::ResultStatus::ErrorMissingTitlekek;
-        return boost::none;
+        return {};
     }
 
     Core::Crypto::AESCipher<Core::Crypto::Key128> cipher(
@@ -167,7 +443,7 @@ boost::optional<Core::Crypto::Key128> NCA::GetTitlekey() {
     return titlekey;
 }
 
-VirtualFile NCA::Decrypt(NCASectionHeader s_header, VirtualFile in, u64 starting_offset) {
+VirtualFile NCA::Decrypt(const NCASectionHeader& s_header, VirtualFile in, u64 starting_offset) {
     if (!encrypted)
         return in;
 
@@ -181,25 +457,25 @@ VirtualFile NCA::Decrypt(NCASectionHeader s_header, VirtualFile in, u64 starting
     case NCASectionCryptoType::BKTR:
         LOG_DEBUG(Crypto, "called with mode=CTR, starting_offset={:016X}", starting_offset);
         {
-            boost::optional<Core::Crypto::Key128> key = boost::none;
+            std::optional<Core::Crypto::Key128> key = {};
             if (has_rights_id) {
                 status = Loader::ResultStatus::Success;
                 key = GetTitlekey();
-                if (key == boost::none) {
+                if (!key) {
                     if (status == Loader::ResultStatus::Success)
                         status = Loader::ResultStatus::ErrorMissingTitlekey;
                     return nullptr;
                 }
             } else {
                 key = GetKeyAreaKey(NCASectionCryptoType::CTR);
-                if (key == boost::none) {
+                if (!key) {
                     status = Loader::ResultStatus::ErrorMissingKeyAreaKey;
                     return nullptr;
                 }
             }
 
-            auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(
-                std::move(in), key.value(), starting_offset);
+            auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(std::move(in), *key,
+                                                                          starting_offset);
             std::vector<u8> iv(16);
             for (u8 i = 0; i < 8; ++i)
                 iv[i] = s_header.raw.section_ctr[0x8 - i - 1];
@@ -215,256 +491,6 @@ VirtualFile NCA::Decrypt(NCASectionHeader s_header, VirtualFile in, u64 starting
     }
 }
 
-NCA::NCA(VirtualFile file_, VirtualFile bktr_base_romfs_, u64 bktr_base_ivfc_offset)
-    : file(std::move(file_)),
-      bktr_base_romfs(bktr_base_romfs_ ? std::move(bktr_base_romfs_) : nullptr) {
-    status = Loader::ResultStatus::Success;
-
-    if (file == nullptr) {
-        status = Loader::ResultStatus::ErrorNullFile;
-        return;
-    }
-
-    if (sizeof(NCAHeader) != file->ReadObject(&header)) {
-        LOG_ERROR(Loader, "File reader errored out during header read.");
-        status = Loader::ResultStatus::ErrorBadNCAHeader;
-        return;
-    }
-
-    encrypted = false;
-
-    if (!IsValidNCA(header)) {
-        if (header.magic == Common::MakeMagic('N', 'C', 'A', '2')) {
-            status = Loader::ResultStatus::ErrorNCA2;
-            return;
-        }
-        if (header.magic == Common::MakeMagic('N', 'C', 'A', '0')) {
-            status = Loader::ResultStatus::ErrorNCA0;
-            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 (dec_header.magic == Common::MakeMagic('N', 'C', 'A', '2')) {
-                status = Loader::ResultStatus::ErrorNCA2;
-                return;
-            }
-            if (dec_header.magic == Common::MakeMagic('N', 'C', 'A', '0')) {
-                status = Loader::ResultStatus::ErrorNCA0;
-                return;
-            }
-
-            if (!keys.HasKey(Core::Crypto::S256KeyType::Header))
-                status = Loader::ResultStatus::ErrorMissingHeaderKey;
-            else
-                status = Loader::ResultStatus::ErrorIncorrectHeaderKey;
-            return;
-        }
-    }
-
-    has_rights_id = std::find_if_not(header.rights_id.begin(), header.rights_id.end(),
-                                     [](char c) { return c == '\0'; }) != header.rights_id.end();
-
-    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);
-    }
-
-    is_update = std::find_if(sections.begin(), sections.end(), [](const NCASectionHeader& header) {
-                    return header.raw.header.crypto_type == NCASectionCryptoType::BKTR;
-                }) != sections.end();
-    ivfc_offset = 0;
-
-    for (std::ptrdiff_t i = 0; i < number_sections; ++i) {
-        auto section = sections[i];
-
-        if (section.raw.header.filesystem_type == NCASectionFilesystemType::ROMFS) {
-            const std::size_t base_offset =
-                header.section_tables[i].media_offset * MEDIA_OFFSET_MULTIPLIER;
-            ivfc_offset = section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].offset;
-            const std::size_t romfs_offset = base_offset + ivfc_offset;
-            const std::size_t romfs_size = section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].size;
-            auto raw = std::make_shared<OffsetVfsFile>(file, romfs_size, romfs_offset);
-            auto dec = Decrypt(section, raw, romfs_offset);
-
-            if (dec == nullptr) {
-                if (status != Loader::ResultStatus::Success)
-                    return;
-                if (has_rights_id)
-                    status = Loader::ResultStatus::ErrorIncorrectTitlekeyOrTitlekek;
-                else
-                    status = Loader::ResultStatus::ErrorIncorrectKeyAreaKey;
-                return;
-            }
-
-            if (section.raw.header.crypto_type == NCASectionCryptoType::BKTR) {
-                if (section.bktr.relocation.magic != Common::MakeMagic('B', 'K', 'T', 'R') ||
-                    section.bktr.subsection.magic != Common::MakeMagic('B', 'K', 'T', 'R')) {
-                    status = Loader::ResultStatus::ErrorBadBKTRHeader;
-                    return;
-                }
-
-                if (section.bktr.relocation.offset + section.bktr.relocation.size !=
-                    section.bktr.subsection.offset) {
-                    status = Loader::ResultStatus::ErrorBKTRSubsectionNotAfterRelocation;
-                    return;
-                }
-
-                const u64 size =
-                    MEDIA_OFFSET_MULTIPLIER * (header.section_tables[i].media_end_offset -
-                                               header.section_tables[i].media_offset);
-                if (section.bktr.subsection.offset + section.bktr.subsection.size != size) {
-                    status = Loader::ResultStatus::ErrorBKTRSubsectionNotAtEnd;
-                    return;
-                }
-
-                const u64 offset = section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].offset;
-                RelocationBlock relocation_block{};
-                if (dec->ReadObject(&relocation_block, section.bktr.relocation.offset - offset) !=
-                    sizeof(RelocationBlock)) {
-                    status = Loader::ResultStatus::ErrorBadRelocationBlock;
-                    return;
-                }
-                SubsectionBlock subsection_block{};
-                if (dec->ReadObject(&subsection_block, section.bktr.subsection.offset - offset) !=
-                    sizeof(RelocationBlock)) {
-                    status = Loader::ResultStatus::ErrorBadSubsectionBlock;
-                    return;
-                }
-
-                std::vector<RelocationBucketRaw> relocation_buckets_raw(
-                    (section.bktr.relocation.size - sizeof(RelocationBlock)) /
-                    sizeof(RelocationBucketRaw));
-                if (dec->ReadBytes(relocation_buckets_raw.data(),
-                                   section.bktr.relocation.size - sizeof(RelocationBlock),
-                                   section.bktr.relocation.offset + sizeof(RelocationBlock) -
-                                       offset) !=
-                    section.bktr.relocation.size - sizeof(RelocationBlock)) {
-                    status = Loader::ResultStatus::ErrorBadRelocationBuckets;
-                    return;
-                }
-
-                std::vector<SubsectionBucketRaw> subsection_buckets_raw(
-                    (section.bktr.subsection.size - sizeof(SubsectionBlock)) /
-                    sizeof(SubsectionBucketRaw));
-                if (dec->ReadBytes(subsection_buckets_raw.data(),
-                                   section.bktr.subsection.size - sizeof(SubsectionBlock),
-                                   section.bktr.subsection.offset + sizeof(SubsectionBlock) -
-                                       offset) !=
-                    section.bktr.subsection.size - sizeof(SubsectionBlock)) {
-                    status = Loader::ResultStatus::ErrorBadSubsectionBuckets;
-                    return;
-                }
-
-                std::vector<RelocationBucket> relocation_buckets(relocation_buckets_raw.size());
-                std::transform(relocation_buckets_raw.begin(), relocation_buckets_raw.end(),
-                               relocation_buckets.begin(), &ConvertRelocationBucketRaw);
-                std::vector<SubsectionBucket> subsection_buckets(subsection_buckets_raw.size());
-                std::transform(subsection_buckets_raw.begin(), subsection_buckets_raw.end(),
-                               subsection_buckets.begin(), &ConvertSubsectionBucketRaw);
-
-                u32 ctr_low;
-                std::memcpy(&ctr_low, section.raw.section_ctr.data(), sizeof(ctr_low));
-                subsection_buckets.back().entries.push_back(
-                    {section.bktr.relocation.offset, {0}, ctr_low});
-                subsection_buckets.back().entries.push_back({size, {0}, 0});
-
-                boost::optional<Core::Crypto::Key128> key = boost::none;
-                if (encrypted) {
-                    if (has_rights_id) {
-                        status = Loader::ResultStatus::Success;
-                        key = GetTitlekey();
-                        if (key == boost::none) {
-                            status = Loader::ResultStatus::ErrorMissingTitlekey;
-                            return;
-                        }
-                    } else {
-                        key = GetKeyAreaKey(NCASectionCryptoType::BKTR);
-                        if (key == boost::none) {
-                            status = Loader::ResultStatus::ErrorMissingKeyAreaKey;
-                            return;
-                        }
-                    }
-                }
-
-                if (bktr_base_romfs == nullptr) {
-                    status = Loader::ResultStatus::ErrorMissingBKTRBaseRomFS;
-                    return;
-                }
-
-                auto bktr = std::make_shared<BKTR>(
-                    bktr_base_romfs, std::make_shared<OffsetVfsFile>(file, romfs_size, base_offset),
-                    relocation_block, relocation_buckets, subsection_block, subsection_buckets,
-                    encrypted, encrypted ? key.get() : Core::Crypto::Key128{}, base_offset,
-                    bktr_base_ivfc_offset, section.raw.section_ctr);
-
-                // BKTR applies to entire IVFC, so make an offset version to level 6
-
-                files.push_back(std::make_shared<OffsetVfsFile>(
-                    bktr, romfs_size, section.romfs.ivfc.levels[IVFC_MAX_LEVEL - 1].offset));
-                romfs = files.back();
-            } else {
-                files.push_back(std::move(dec));
-                romfs = files.back();
-            }
-        } else if (section.raw.header.filesystem_type == NCASectionFilesystemType::PFS0) {
-            u64 offset = (static_cast<u64>(header.section_tables[i].media_offset) *
-                          MEDIA_OFFSET_MULTIPLIER) +
-                         section.pfs0.pfs0_header_offset;
-            u64 size = MEDIA_OFFSET_MULTIPLIER * (header.section_tables[i].media_end_offset -
-                                                  header.section_tables[i].media_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.push_back(std::move(npfs));
-                    if (IsDirectoryExeFS(dirs.back()))
-                        exefs = dirs.back();
-                } else {
-                    if (has_rights_id)
-                        status = Loader::ResultStatus::ErrorIncorrectTitlekeyOrTitlekek;
-                    else
-                        status = Loader::ResultStatus::ErrorIncorrectKeyAreaKey;
-                    return;
-                }
-            } else {
-                if (status != Loader::ResultStatus::Success)
-                    return;
-                if (has_rights_id)
-                    status = Loader::ResultStatus::ErrorIncorrectTitlekeyOrTitlekek;
-                else
-                    status = Loader::ResultStatus::ErrorIncorrectKeyAreaKey;
-                return;
-            }
-        }
-    }
-
-    status = Loader::ResultStatus::Success;
-}
-
-NCA::~NCA() = default;
-
 Loader::ResultStatus NCA::GetStatus() const {
     return status;
 }
@@ -519,7 +545,4 @@ u64 NCA::GetBaseIVFCOffset() const {
     return ivfc_offset;
 }
 
-bool NCA::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys
diff --git a/src/core/file_sys/content_archive.h b/src/core/file_sys/content_archive.h
index f9f66cae9..4bba55607 100644
--- a/src/core/file_sys/content_archive.h
+++ b/src/core/file_sys/content_archive.h
@@ -6,9 +6,10 @@
 
 #include <array>
 #include <memory>
+#include <optional>
 #include <string>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/swap.h"
@@ -73,8 +74,6 @@ inline bool IsDirectoryExeFS(const std::shared_ptr<VfsDirectory>& pfs) {
     return pfs->GetFile("main") != nullptr && pfs->GetFile("main.npdm") != nullptr;
 }
 
-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.
 class NCA : public ReadOnlyVfsDirectory {
@@ -102,14 +101,20 @@ public:
     // Returns the base ivfc offset used in BKTR patching.
     u64 GetBaseIVFCOffset() const;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
+    bool CheckSupportedNCA(const NCAHeader& header);
+    bool HandlePotentialHeaderDecryption();
+
+    std::vector<NCASectionHeader> ReadSectionHeaders() const;
+    bool ReadSections(const std::vector<NCASectionHeader>& sections, u64 bktr_base_ivfc_offset);
+    bool ReadRomFSSection(const NCASectionHeader& section, const NCASectionTableEntry& entry,
+                          u64 bktr_base_ivfc_offset);
+    bool ReadPFS0Section(const NCASectionHeader& section, const NCASectionTableEntry& entry);
+
     u8 GetCryptoRevision() const;
-    boost::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const;
-    boost::optional<Core::Crypto::Key128> GetTitlekey();
-    VirtualFile Decrypt(NCASectionHeader header, VirtualFile in, u64 starting_offset);
+    std::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const;
+    std::optional<Core::Crypto::Key128> GetTitlekey();
+    VirtualFile Decrypt(const NCASectionHeader& header, VirtualFile in, u64 starting_offset);
 
     std::vector<VirtualDir> dirs;
     std::vector<VirtualFile> files;
@@ -118,15 +123,15 @@ private:
     VirtualDir exefs = nullptr;
     VirtualFile file;
     VirtualFile bktr_base_romfs;
-    u64 ivfc_offset;
+    u64 ivfc_offset = 0;
 
     NCAHeader header{};
     bool has_rights_id{};
 
     Loader::ResultStatus status{};
 
-    bool encrypted;
-    bool is_update;
+    bool encrypted = false;
+    bool is_update = false;
 
     Core::Crypto::KeyManager keys;
 };
diff --git a/src/core/file_sys/control_metadata.cpp b/src/core/file_sys/control_metadata.cpp
index 5b1177a03..a012c2be9 100644
--- a/src/core/file_sys/control_metadata.cpp
+++ b/src/core/file_sys/control_metadata.cpp
@@ -17,11 +17,13 @@ const std::array<const char*, 15> LANGUAGE_NAMES = {
 };
 
 std::string LanguageEntry::GetApplicationName() const {
-    return Common::StringFromFixedZeroTerminatedBuffer(application_name.data(), 0x200);
+    return Common::StringFromFixedZeroTerminatedBuffer(application_name.data(),
+                                                       application_name.size());
 }
 
 std::string LanguageEntry::GetDeveloperName() const {
-    return Common::StringFromFixedZeroTerminatedBuffer(developer_name.data(), 0x100);
+    return Common::StringFromFixedZeroTerminatedBuffer(developer_name.data(),
+                                                       developer_name.size());
 }
 
 NACP::NACP(VirtualFile file) : raw(std::make_unique<RawNACP>()) {
@@ -56,7 +58,12 @@ u64 NACP::GetTitleId() const {
     return raw->title_id;
 }
 
+u64 NACP::GetDLCBaseTitleId() const {
+    return raw->dlc_base_title_id;
+}
+
 std::string NACP::GetVersionString() const {
-    return Common::StringFromFixedZeroTerminatedBuffer(raw->version_string.data(), 0x10);
+    return Common::StringFromFixedZeroTerminatedBuffer(raw->version_string.data(),
+                                                       raw->version_string.size());
 }
 } // namespace FileSys
diff --git a/src/core/file_sys/control_metadata.h b/src/core/file_sys/control_metadata.h
index 43d6f0719..141f7e056 100644
--- a/src/core/file_sys/control_metadata.h
+++ b/src/core/file_sys/control_metadata.h
@@ -79,6 +79,7 @@ public:
     std::string GetApplicationName(Language language = Language::Default) const;
     std::string GetDeveloperName(Language language = Language::Default) const;
     u64 GetTitleId() const;
+    u64 GetDLCBaseTitleId() const;
     std::string GetVersionString() const;
 
 private:
diff --git a/src/core/file_sys/fsmitm_romfsbuild.cpp b/src/core/file_sys/fsmitm_romfsbuild.cpp
index 2a913ce82..47b7526c7 100644
--- a/src/core/file_sys/fsmitm_romfsbuild.cpp
+++ b/src/core/file_sys/fsmitm_romfsbuild.cpp
@@ -26,6 +26,7 @@
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "core/file_sys/fsmitm_romfsbuild.h"
+#include "core/file_sys/ips_layer.h"
 #include "core/file_sys/vfs.h"
 #include "core/file_sys/vfs_vector.h"
 
@@ -123,7 +124,7 @@ static u64 romfs_get_hash_table_count(u64 num_entries) {
     return count;
 }
 
-void RomFSBuildContext::VisitDirectory(VirtualDir root_romfs,
+void RomFSBuildContext::VisitDirectory(VirtualDir root_romfs, VirtualDir ext,
                                        std::shared_ptr<RomFSBuildDirectoryContext> parent) {
     std::vector<std::shared_ptr<RomFSBuildDirectoryContext>> child_dirs;
 
@@ -144,6 +145,9 @@ void RomFSBuildContext::VisitDirectory(VirtualDir root_romfs,
             child->path_len = child->cur_path_ofs + static_cast<u32>(kv.first.size());
             child->path = parent->path + "/" + kv.first;
 
+            if (ext != nullptr && ext->GetFileRelative(child->path + ".stub") != nullptr)
+                continue;
+
             // Sanity check on path_len
             ASSERT(child->path_len < FS_MAX_PATH);
 
@@ -157,11 +161,24 @@ void RomFSBuildContext::VisitDirectory(VirtualDir root_romfs,
             child->path_len = child->cur_path_ofs + static_cast<u32>(kv.first.size());
             child->path = parent->path + "/" + kv.first;
 
+            if (ext != nullptr && ext->GetFileRelative(child->path + ".stub") != nullptr)
+                continue;
+
             // Sanity check on path_len
             ASSERT(child->path_len < FS_MAX_PATH);
 
             child->source = root_romfs->GetFileRelative(child->path);
 
+            if (ext != nullptr) {
+                const auto ips = ext->GetFileRelative(child->path + ".ips");
+
+                if (ips != nullptr) {
+                    auto patched = PatchIPS(child->source, ips);
+                    if (patched != nullptr)
+                        child->source = std::move(patched);
+                }
+            }
+
             child->size = child->source->GetSize();
 
             AddFile(parent, child);
@@ -169,7 +186,7 @@ void RomFSBuildContext::VisitDirectory(VirtualDir root_romfs,
     }
 
     for (auto& child : child_dirs) {
-        this->VisitDirectory(root_romfs, child);
+        this->VisitDirectory(root_romfs, ext, child);
     }
 }
 
@@ -208,14 +225,15 @@ bool RomFSBuildContext::AddFile(std::shared_ptr<RomFSBuildDirectoryContext> pare
     return true;
 }
 
-RomFSBuildContext::RomFSBuildContext(VirtualDir base_) : base(std::move(base_)) {
+RomFSBuildContext::RomFSBuildContext(VirtualDir base_, VirtualDir ext_)
+    : base(std::move(base_)), ext(std::move(ext_)) {
     root = std::make_shared<RomFSBuildDirectoryContext>();
     root->path = "\0";
     directories.emplace(root->path, root);
     num_dirs = 1;
     dir_table_size = 0x18;
 
-    VisitDirectory(base, root);
+    VisitDirectory(base, ext, root);
 }
 
 RomFSBuildContext::~RomFSBuildContext() = default;
diff --git a/src/core/file_sys/fsmitm_romfsbuild.h b/src/core/file_sys/fsmitm_romfsbuild.h
index b0c3c123b..a62502193 100644
--- a/src/core/file_sys/fsmitm_romfsbuild.h
+++ b/src/core/file_sys/fsmitm_romfsbuild.h
@@ -27,7 +27,6 @@
 #include <map>
 #include <memory>
 #include <string>
-#include <boost/detail/container_fwd.hpp>
 #include "common/common_types.h"
 #include "core/file_sys/vfs.h"
 
@@ -40,7 +39,7 @@ struct RomFSFileEntry;
 
 class RomFSBuildContext {
 public:
-    explicit RomFSBuildContext(VirtualDir base);
+    explicit RomFSBuildContext(VirtualDir base, VirtualDir ext = nullptr);
     ~RomFSBuildContext();
 
     // This finalizes the context.
@@ -48,6 +47,7 @@ public:
 
 private:
     VirtualDir base;
+    VirtualDir ext;
     std::shared_ptr<RomFSBuildDirectoryContext> root;
     std::map<std::string, std::shared_ptr<RomFSBuildDirectoryContext>, std::less<>> directories;
     std::map<std::string, std::shared_ptr<RomFSBuildFileContext>, std::less<>> files;
@@ -59,7 +59,8 @@ private:
     u64 file_hash_table_size = 0;
     u64 file_partition_size = 0;
 
-    void VisitDirectory(VirtualDir filesys, std::shared_ptr<RomFSBuildDirectoryContext> parent);
+    void VisitDirectory(VirtualDir filesys, VirtualDir ext,
+                        std::shared_ptr<RomFSBuildDirectoryContext> parent);
 
     bool AddDirectory(std::shared_ptr<RomFSBuildDirectoryContext> parent_dir_ctx,
                       std::shared_ptr<RomFSBuildDirectoryContext> dir_ctx);
diff --git a/src/core/file_sys/ips_layer.cpp b/src/core/file_sys/ips_layer.cpp
index df933ee36..485c4913a 100644
--- a/src/core/file_sys/ips_layer.cpp
+++ b/src/core/file_sys/ips_layer.cpp
@@ -2,7 +2,15 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include "common/assert.h"
+#include <algorithm>
+#include <cstring>
+#include <map>
+#include <sstream>
+#include <string>
+#include <utility>
+
+#include "common/hex_util.h"
+#include "common/logging/log.h"
 #include "common/swap.h"
 #include "core/file_sys/ips_layer.h"
 #include "core/file_sys/vfs_vector.h"
@@ -15,16 +23,48 @@ enum class IPSFileType {
     Error,
 };
 
+constexpr std::array<std::pair<const char*, const char*>, 11> ESCAPE_CHARACTER_MAP{{
+    {"\\a", "\a"},
+    {"\\b", "\b"},
+    {"\\f", "\f"},
+    {"\\n", "\n"},
+    {"\\r", "\r"},
+    {"\\t", "\t"},
+    {"\\v", "\v"},
+    {"\\\\", "\\"},
+    {"\\\'", "\'"},
+    {"\\\"", "\""},
+    {"\\\?", "\?"},
+}};
+
 static IPSFileType IdentifyMagic(const std::vector<u8>& magic) {
-    if (magic.size() != 5)
+    if (magic.size() != 5) {
         return IPSFileType::Error;
-    if (magic == std::vector<u8>{'P', 'A', 'T', 'C', 'H'})
+    }
+
+    constexpr std::array<u8, 5> patch_magic{{'P', 'A', 'T', 'C', 'H'}};
+    if (std::equal(magic.begin(), magic.end(), patch_magic.begin())) {
         return IPSFileType::IPS;
-    if (magic == std::vector<u8>{'I', 'P', 'S', '3', '2'})
+    }
+
+    constexpr std::array<u8, 5> ips32_magic{{'I', 'P', 'S', '3', '2'}};
+    if (std::equal(magic.begin(), magic.end(), ips32_magic.begin())) {
         return IPSFileType::IPS32;
+    }
+
     return IPSFileType::Error;
 }
 
+static bool IsEOF(IPSFileType type, const std::vector<u8>& data) {
+    constexpr std::array<u8, 3> eof{{'E', 'O', 'F'}};
+    if (type == IPSFileType::IPS && std::equal(data.begin(), data.end(), eof.begin())) {
+        return true;
+    }
+
+    constexpr std::array<u8, 4> eeof{{'E', 'E', 'O', 'F'}};
+    return type == IPSFileType::IPS32 && std::equal(data.begin(), data.end(), eeof.begin());
+}
+
 VirtualFile PatchIPS(const VirtualFile& in, const VirtualFile& ips) {
     if (in == nullptr || ips == nullptr)
         return nullptr;
@@ -39,8 +79,7 @@ VirtualFile PatchIPS(const VirtualFile& in, const VirtualFile& ips) {
     u64 offset = 5; // After header
     while (ips->Read(temp.data(), temp.size(), offset) == temp.size()) {
         offset += temp.size();
-        if (type == IPSFileType::IPS32 && temp == std::vector<u8>{'E', 'E', 'O', 'F'} ||
-            type == IPSFileType::IPS && temp == std::vector<u8>{'E', 'O', 'F'}) {
+        if (IsEOF(type, temp)) {
             break;
         }
 
@@ -60,29 +99,240 @@ VirtualFile PatchIPS(const VirtualFile& in, const VirtualFile& ips) {
             u16 rle_size{};
             if (ips->ReadObject(&rle_size, offset) != sizeof(u16))
                 return nullptr;
-            rle_size = Common::swap16(data_size);
+            rle_size = Common::swap16(rle_size);
             offset += sizeof(u16);
 
             const auto data = ips->ReadByte(offset++);
-            if (data == boost::none)
+            if (!data)
                 return nullptr;
 
             if (real_offset + rle_size > in_data.size())
-                rle_size = in_data.size() - real_offset;
-            std::memset(in_data.data() + real_offset, data.get(), rle_size);
+                rle_size = static_cast<u16>(in_data.size() - real_offset);
+            std::memset(in_data.data() + real_offset, *data, rle_size);
         } else { // Standard Patch
             auto read = data_size;
             if (real_offset + read > in_data.size())
-                read = in_data.size() - real_offset;
+                read = static_cast<u16>(in_data.size() - real_offset);
             if (ips->Read(in_data.data() + real_offset, read, offset) != data_size)
                 return nullptr;
             offset += data_size;
         }
     }
 
-    if (temp != std::vector<u8>{'E', 'E', 'O', 'F'} && temp != std::vector<u8>{'E', 'O', 'F'})
+    if (!IsEOF(type, temp)) {
         return nullptr;
-    return std::make_shared<VectorVfsFile>(in_data, in->GetName(), in->GetContainingDirectory());
+    }
+
+    return std::make_shared<VectorVfsFile>(std::move(in_data), in->GetName(),
+                                           in->GetContainingDirectory());
+}
+
+struct IPSwitchCompiler::IPSwitchPatch {
+    std::string name;
+    bool enabled;
+    std::map<u32, std::vector<u8>> records;
+};
+
+IPSwitchCompiler::IPSwitchCompiler(VirtualFile patch_text_) : patch_text(std::move(patch_text_)) {
+    Parse();
+}
+
+IPSwitchCompiler::~IPSwitchCompiler() = default;
+
+std::array<u8, 32> IPSwitchCompiler::GetBuildID() const {
+    return nso_build_id;
+}
+
+bool IPSwitchCompiler::IsValid() const {
+    return valid;
+}
+
+static bool StartsWith(std::string_view base, std::string_view check) {
+    return base.size() >= check.size() && base.substr(0, check.size()) == check;
+}
+
+static std::string EscapeStringSequences(std::string in) {
+    for (const auto& seq : ESCAPE_CHARACTER_MAP) {
+        for (auto index = in.find(seq.first); index != std::string::npos;
+             index = in.find(seq.first, index)) {
+            in.replace(index, std::strlen(seq.first), seq.second);
+            index += std::strlen(seq.second);
+        }
+    }
+
+    return in;
+}
+
+void IPSwitchCompiler::ParseFlag(const std::string& line) {
+    if (StartsWith(line, "@flag offset_shift ")) {
+        // Offset Shift Flag
+        offset_shift = std::stoll(line.substr(19), nullptr, 0);
+    } else if (StartsWith(line, "@little-endian")) {
+        // Set values to read as little endian
+        is_little_endian = true;
+    } else if (StartsWith(line, "@big-endian")) {
+        // Set values to read as big endian
+        is_little_endian = false;
+    } else if (StartsWith(line, "@flag print_values")) {
+        // Force printing of applied values
+        print_values = true;
+    }
+}
+
+void IPSwitchCompiler::Parse() {
+    const auto bytes = patch_text->ReadAllBytes();
+    std::stringstream s;
+    s.write(reinterpret_cast<const char*>(bytes.data()), bytes.size());
+
+    std::vector<std::string> lines;
+    std::string stream_line;
+    while (std::getline(s, stream_line)) {
+        // Remove a trailing \r
+        if (!stream_line.empty() && stream_line.back() == '\r')
+            stream_line.pop_back();
+        lines.push_back(std::move(stream_line));
+    }
+
+    for (std::size_t i = 0; i < lines.size(); ++i) {
+        auto line = lines[i];
+
+        // Remove midline comments
+        std::size_t comment_index = std::string::npos;
+        bool within_string = false;
+        for (std::size_t k = 0; k < line.size(); ++k) {
+            if (line[k] == '\"' && (k > 0 && line[k - 1] != '\\')) {
+                within_string = !within_string;
+            } else if (line[k] == '\\' && (k < line.size() - 1 && line[k + 1] == '\\')) {
+                comment_index = k;
+                break;
+            }
+        }
+
+        if (!StartsWith(line, "//") && comment_index != std::string::npos) {
+            last_comment = line.substr(comment_index + 2);
+            line = line.substr(0, comment_index);
+        }
+
+        if (StartsWith(line, "@stop")) {
+            // Force stop
+            break;
+        } else if (StartsWith(line, "@nsobid-")) {
+            // NSO Build ID Specifier
+            auto raw_build_id = line.substr(8);
+            if (raw_build_id.size() != 0x40)
+                raw_build_id.resize(0x40, '0');
+            nso_build_id = Common::HexStringToArray<0x20>(raw_build_id);
+        } else if (StartsWith(line, "#")) {
+            // Mandatory Comment
+            LOG_INFO(Loader, "[IPSwitchCompiler ('{}')] Forced output comment: {}",
+                     patch_text->GetName(), line.substr(1));
+        } else if (StartsWith(line, "//")) {
+            // Normal Comment
+            last_comment = line.substr(2);
+            if (last_comment.find_first_not_of(' ') == std::string::npos)
+                continue;
+            if (last_comment.find_first_not_of(' ') != 0)
+                last_comment = last_comment.substr(last_comment.find_first_not_of(' '));
+        } else if (StartsWith(line, "@enabled") || StartsWith(line, "@disabled")) {
+            // Start of patch
+            const auto enabled = StartsWith(line, "@enabled");
+            if (i == 0)
+                return;
+            LOG_INFO(Loader, "[IPSwitchCompiler ('{}')] Parsing patch '{}' ({})",
+                     patch_text->GetName(), last_comment, line.substr(1));
+
+            IPSwitchPatch patch{last_comment, enabled, {}};
+
+            // Read rest of patch
+            while (true) {
+                if (i + 1 >= lines.size())
+                    break;
+                const auto patch_line = lines[++i];
+
+                // Start of new patch
+                if (StartsWith(patch_line, "@enabled") || StartsWith(patch_line, "@disabled")) {
+                    --i;
+                    break;
+                }
+
+                // Check for a flag
+                if (StartsWith(patch_line, "@")) {
+                    ParseFlag(patch_line);
+                    continue;
+                }
+
+                // 11 - 8 hex digit offset + space + minimum two digit overwrite val
+                if (patch_line.length() < 11)
+                    break;
+                auto offset = std::stoul(patch_line.substr(0, 8), nullptr, 16);
+                offset += static_cast<unsigned long>(offset_shift);
+
+                std::vector<u8> replace;
+                // 9 - first char of replacement val
+                if (patch_line[9] == '\"') {
+                    // string replacement
+                    auto end_index = patch_line.find('\"', 10);
+                    if (end_index == std::string::npos || end_index < 10)
+                        return;
+                    while (patch_line[end_index - 1] == '\\') {
+                        end_index = patch_line.find('\"', end_index + 1);
+                        if (end_index == std::string::npos || end_index < 10)
+                            return;
+                    }
+
+                    auto value = patch_line.substr(10, end_index - 10);
+                    value = EscapeStringSequences(value);
+                    replace.reserve(value.size());
+                    std::copy(value.begin(), value.end(), std::back_inserter(replace));
+                } else {
+                    // hex replacement
+                    const auto value = patch_line.substr(9);
+                    replace.reserve(value.size() / 2);
+                    replace = Common::HexStringToVector(value, is_little_endian);
+                }
+
+                if (print_values) {
+                    LOG_INFO(Loader,
+                             "[IPSwitchCompiler ('{}')]     - Patching value at offset 0x{:08X} "
+                             "with byte string '{}'",
+                             patch_text->GetName(), offset, Common::HexVectorToString(replace));
+                }
+
+                patch.records.insert_or_assign(offset, std::move(replace));
+            }
+
+            patches.push_back(std::move(patch));
+        } else if (StartsWith(line, "@")) {
+            ParseFlag(line);
+        }
+    }
+
+    valid = true;
+}
+
+VirtualFile IPSwitchCompiler::Apply(const VirtualFile& in) const {
+    if (in == nullptr || !valid)
+        return nullptr;
+
+    auto in_data = in->ReadAllBytes();
+
+    for (const auto& patch : patches) {
+        if (!patch.enabled)
+            continue;
+
+        for (const auto& record : patch.records) {
+            if (record.first >= in_data.size())
+                continue;
+            auto replace_size = record.second.size();
+            if (record.first + replace_size > in_data.size())
+                replace_size = in_data.size() - record.first;
+            for (std::size_t i = 0; i < replace_size; ++i)
+                in_data[i + record.first] = record.second[i];
+        }
+    }
+
+    return std::make_shared<VectorVfsFile>(std::move(in_data), in->GetName(),
+                                           in->GetContainingDirectory());
 }
 
 } // namespace FileSys
diff --git a/src/core/file_sys/ips_layer.h b/src/core/file_sys/ips_layer.h
index 81c163494..450b2f71e 100644
--- a/src/core/file_sys/ips_layer.h
+++ b/src/core/file_sys/ips_layer.h
@@ -4,12 +4,41 @@
 
 #pragma once
 
+#include <array>
 #include <memory>
+#include <vector>
 
+#include "common/common_types.h"
 #include "core/file_sys/vfs.h"
 
 namespace FileSys {
 
 VirtualFile PatchIPS(const VirtualFile& in, const VirtualFile& ips);
 
+class IPSwitchCompiler {
+public:
+    explicit IPSwitchCompiler(VirtualFile patch_text);
+    ~IPSwitchCompiler();
+
+    std::array<u8, 0x20> GetBuildID() const;
+    bool IsValid() const;
+    VirtualFile Apply(const VirtualFile& in) const;
+
+private:
+    struct IPSwitchPatch;
+
+    void ParseFlag(const std::string& flag);
+    void Parse();
+
+    bool valid = false;
+
+    VirtualFile patch_text;
+    std::vector<IPSwitchPatch> patches;
+    std::array<u8, 0x20> nso_build_id{};
+    bool is_little_endian = false;
+    s64 offset_shift = 0;
+    bool print_values = false;
+    std::string last_comment = "";
+};
+
 } // namespace FileSys
diff --git a/src/core/file_sys/partition_filesystem.cpp b/src/core/file_sys/partition_filesystem.cpp
index 5791c76ff..a5259a593 100644
--- a/src/core/file_sys/partition_filesystem.cpp
+++ b/src/core/file_sys/partition_filesystem.cpp
@@ -83,7 +83,7 @@ std::vector<std::shared_ptr<VfsFile>> PartitionFilesystem::GetFiles() const {
 }
 
 std::vector<std::shared_ptr<VfsDirectory>> PartitionFilesystem::GetSubdirectories() const {
-    return pfs_dirs;
+    return {};
 }
 
 std::string PartitionFilesystem::GetName() const {
@@ -103,18 +103,4 @@ void PartitionFilesystem::PrintDebugInfo() const {
                   pfs_files[i]->GetName(), pfs_files[i]->GetSize());
     }
 }
-
-bool PartitionFilesystem::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    const auto iter = std::find(pfs_files.begin(), pfs_files.end(), file);
-    if (iter == pfs_files.end())
-        return false;
-
-    const std::ptrdiff_t offset = std::distance(pfs_files.begin(), iter);
-    pfs_files[offset] = std::move(pfs_files.back());
-    pfs_files.pop_back();
-
-    pfs_dirs.emplace_back(std::move(dir));
-
-    return true;
-}
 } // namespace FileSys
diff --git a/src/core/file_sys/partition_filesystem.h b/src/core/file_sys/partition_filesystem.h
index 739c63a7f..248fdfdeb 100644
--- a/src/core/file_sys/partition_filesystem.h
+++ b/src/core/file_sys/partition_filesystem.h
@@ -35,9 +35,6 @@ public:
     std::shared_ptr<VfsDirectory> GetParentDirectory() const override;
     void PrintDebugInfo() const;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     struct Header {
         u32_le magic;
@@ -84,7 +81,6 @@ private:
     std::size_t content_offset = 0;
 
     std::vector<VirtualFile> pfs_files;
-    std::vector<VirtualDir> pfs_dirs;
 };
 
 } // namespace FileSys
diff --git a/src/core/file_sys/patch_manager.cpp b/src/core/file_sys/patch_manager.cpp
index 539698f6e..0c1156989 100644
--- a/src/core/file_sys/patch_manager.cpp
+++ b/src/core/file_sys/patch_manager.cpp
@@ -61,39 +61,49 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
     // Game Updates
     const auto update_tid = GetUpdateTitleID(title_id);
     const auto update = installed->GetEntry(update_tid, ContentRecordType::Program);
-    if (update != nullptr) {
-        if (update->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS &&
-            update->GetExeFS() != nullptr) {
-            LOG_INFO(Loader, "    ExeFS: Update ({}) applied successfully",
-                     FormatTitleVersion(installed->GetEntryVersion(update_tid).get_value_or(0)));
-            exefs = update->GetExeFS();
-        }
+
+    if (update != nullptr && update->GetExeFS() != nullptr &&
+        update->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) {
+        LOG_INFO(Loader, "    ExeFS: Update ({}) applied successfully",
+                 FormatTitleVersion(installed->GetEntryVersion(update_tid).value_or(0)));
+        exefs = update->GetExeFS();
     }
 
     return exefs;
 }
 
-static std::vector<VirtualFile> CollectIPSPatches(const std::vector<VirtualDir>& patch_dirs,
-                                                  const std::string& build_id) {
-    std::vector<VirtualFile> ips;
-    ips.reserve(patch_dirs.size());
+static std::vector<VirtualFile> CollectPatches(const std::vector<VirtualDir>& patch_dirs,
+                                               const std::string& build_id) {
+    std::vector<VirtualFile> out;
+    out.reserve(patch_dirs.size());
     for (const auto& subdir : patch_dirs) {
         auto exefs_dir = subdir->GetSubdirectory("exefs");
         if (exefs_dir != nullptr) {
             for (const auto& file : exefs_dir->GetFiles()) {
-                if (file->GetExtension() != "ips")
-                    continue;
-                auto name = file->GetName();
-                const auto p1 = name.substr(0, name.find('.'));
-                const auto this_build_id = p1.substr(0, p1.find_last_not_of('0') + 1);
-
-                if (build_id == this_build_id)
-                    ips.push_back(file);
+                if (file->GetExtension() == "ips") {
+                    auto name = file->GetName();
+                    const auto p1 = name.substr(0, name.find('.'));
+                    const auto this_build_id = p1.substr(0, p1.find_last_not_of('0') + 1);
+
+                    if (build_id == this_build_id)
+                        out.push_back(file);
+                } else if (file->GetExtension() == "pchtxt") {
+                    IPSwitchCompiler compiler{file};
+                    if (!compiler.IsValid())
+                        continue;
+
+                    auto this_build_id = Common::HexArrayToString(compiler.GetBuildID());
+                    this_build_id =
+                        this_build_id.substr(0, this_build_id.find_last_not_of('0') + 1);
+
+                    if (build_id == this_build_id)
+                        out.push_back(file);
+                }
             }
         }
     }
 
-    return ips;
+    return out;
 }
 
 std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso) const {
@@ -115,15 +125,24 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso) const {
     auto patch_dirs = load_dir->GetSubdirectories();
     std::sort(patch_dirs.begin(), patch_dirs.end(),
               [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
-    const auto ips = CollectIPSPatches(patch_dirs, build_id);
+    const auto patches = CollectPatches(patch_dirs, build_id);
 
     auto out = nso;
-    for (const auto& ips_file : ips) {
-        LOG_INFO(Loader, "    - Appling IPS patch from mod \"{}\"",
-                 ips_file->GetContainingDirectory()->GetParentDirectory()->GetName());
-        const auto patched = PatchIPS(std::make_shared<VectorVfsFile>(out), ips_file);
-        if (patched != nullptr)
-            out = patched->ReadAllBytes();
+    for (const auto& patch_file : patches) {
+        if (patch_file->GetExtension() == "ips") {
+            LOG_INFO(Loader, "    - Applying IPS patch from mod \"{}\"",
+                     patch_file->GetContainingDirectory()->GetParentDirectory()->GetName());
+            const auto patched = PatchIPS(std::make_shared<VectorVfsFile>(out), patch_file);
+            if (patched != nullptr)
+                out = patched->ReadAllBytes();
+        } else if (patch_file->GetExtension() == "pchtxt") {
+            LOG_INFO(Loader, "    - Applying IPSwitch patch from mod \"{}\"",
+                     patch_file->GetContainingDirectory()->GetParentDirectory()->GetName());
+            const IPSwitchCompiler compiler{patch_file};
+            const auto patched = compiler.Apply(std::make_shared<VectorVfsFile>(out));
+            if (patched != nullptr)
+                out = patched->ReadAllBytes();
+        }
     }
 
     if (out.size() < 0x100)
@@ -143,12 +162,13 @@ bool PatchManager::HasNSOPatch(const std::array<u8, 32>& build_id_) const {
     std::sort(patch_dirs.begin(), patch_dirs.end(),
               [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
 
-    return !CollectIPSPatches(patch_dirs, build_id).empty();
+    return !CollectPatches(patch_dirs, build_id).empty();
 }
 
 static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType type) {
     const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
-    if (type != ContentRecordType::Program || load_dir == nullptr || load_dir->GetSize() <= 0) {
+    if ((type != ContentRecordType::Program && type != ContentRecordType::Data) ||
+        load_dir == nullptr || load_dir->GetSize() <= 0) {
         return;
     }
 
@@ -162,11 +182,17 @@ static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType t
               [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
 
     std::vector<VirtualDir> layers;
+    std::vector<VirtualDir> layers_ext;
     layers.reserve(patch_dirs.size() + 1);
+    layers_ext.reserve(patch_dirs.size() + 1);
     for (const auto& subdir : patch_dirs) {
         auto romfs_dir = subdir->GetSubdirectory("romfs");
         if (romfs_dir != nullptr)
             layers.push_back(std::move(romfs_dir));
+
+        auto ext_dir = subdir->GetSubdirectory("romfs_ext");
+        if (ext_dir != nullptr)
+            layers_ext.push_back(std::move(ext_dir));
     }
     layers.push_back(std::move(extracted));
 
@@ -175,7 +201,9 @@ static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType t
         return;
     }
 
-    auto packed = CreateRomFS(std::move(layered));
+    auto layered_ext = LayeredVfsDirectory::MakeLayeredDirectory(std::move(layers_ext));
+
+    auto packed = CreateRomFS(std::move(layered), std::move(layered_ext));
     if (packed == nullptr) {
         return;
     }
@@ -184,10 +212,16 @@ static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType t
     romfs = std::move(packed);
 }
 
-VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset,
-                                     ContentRecordType type) const {
-    LOG_INFO(Loader, "Patching RomFS for title_id={:016X}, type={:02X}", title_id,
-             static_cast<u8>(type));
+VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, ContentRecordType type,
+                                     VirtualFile update_raw) const {
+    const auto log_string = fmt::format("Patching RomFS for title_id={:016X}, type={:02X}",
+                                        title_id, static_cast<u8>(type))
+                                .c_str();
+
+    if (type == ContentRecordType::Program || type == ContentRecordType::Data)
+        LOG_INFO(Loader, log_string);
+    else
+        LOG_DEBUG(Loader, log_string);
 
     if (romfs == nullptr)
         return romfs;
@@ -202,7 +236,14 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset,
         if (new_nca->GetStatus() == Loader::ResultStatus::Success &&
             new_nca->GetRomFS() != nullptr) {
             LOG_INFO(Loader, "    RomFS: Update ({}) applied successfully",
-                     FormatTitleVersion(installed->GetEntryVersion(update_tid).get_value_or(0)));
+                     FormatTitleVersion(installed->GetEntryVersion(update_tid).value_or(0)));
+            romfs = new_nca->GetRomFS();
+        }
+    } else if (update_raw != nullptr) {
+        const auto new_nca = std::make_shared<NCA>(update_raw, romfs, ivfc_offset);
+        if (new_nca->GetStatus() == Loader::ResultStatus::Success &&
+            new_nca->GetRomFS() != nullptr) {
+            LOG_INFO(Loader, "    RomFS: Update (PACKED) applied successfully");
             romfs = new_nca->GetRomFS();
         }
     }
@@ -224,7 +265,8 @@ static bool IsDirValidAndNonEmpty(const VirtualDir& dir) {
     return dir != nullptr && (!dir->GetFiles().empty() || !dir->GetSubdirectories().empty());
 }
 
-std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNames() const {
+std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNames(
+    VirtualFile update_raw) const {
     std::map<std::string, std::string, std::less<>> out;
     const auto installed = Service::FileSystem::GetUnionContents();
 
@@ -238,13 +280,14 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
     } else {
         if (installed->HasEntry(update_tid, ContentRecordType::Program)) {
             const auto meta_ver = installed->GetEntryVersion(update_tid);
-            if (meta_ver == boost::none || meta_ver.get() == 0) {
+            if (meta_ver.value_or(0) == 0) {
                 out.insert_or_assign("Update", "");
             } else {
                 out.insert_or_assign(
-                    "Update",
-                    FormatTitleVersion(meta_ver.get(), TitleVersionFormat::ThreeElements));
+                    "Update", FormatTitleVersion(*meta_ver, TitleVersionFormat::ThreeElements));
             }
+        } else if (update_raw != nullptr) {
+            out.insert_or_assign("Update", "PACKED");
         }
     }
 
@@ -253,8 +296,24 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
     if (mod_dir != nullptr && mod_dir->GetSize() > 0) {
         for (const auto& mod : mod_dir->GetSubdirectories()) {
             std::string types;
-            if (IsDirValidAndNonEmpty(mod->GetSubdirectory("exefs")))
-                AppendCommaIfNotEmpty(types, "IPS");
+
+            const auto exefs_dir = mod->GetSubdirectory("exefs");
+            if (IsDirValidAndNonEmpty(exefs_dir)) {
+                bool ips = false;
+                bool ipswitch = false;
+
+                for (const auto& file : exefs_dir->GetFiles()) {
+                    if (file->GetExtension() == "ips")
+                        ips = true;
+                    else if (file->GetExtension() == "pchtxt")
+                        ipswitch = true;
+                }
+
+                if (ips)
+                    AppendCommaIfNotEmpty(types, "IPS");
+                if (ipswitch)
+                    AppendCommaIfNotEmpty(types, "IPSwitch");
+            }
             if (IsDirValidAndNonEmpty(mod->GetSubdirectory("romfs")))
                 AppendCommaIfNotEmpty(types, "LayeredFS");
 
@@ -291,23 +350,22 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
     return out;
 }
 
-std::pair<std::shared_ptr<NACP>, VirtualFile> PatchManager::GetControlMetadata() const {
-    const auto& installed{Service::FileSystem::GetUnionContents()};
+std::pair<std::unique_ptr<NACP>, VirtualFile> PatchManager::GetControlMetadata() const {
+    const auto installed{Service::FileSystem::GetUnionContents()};
 
     const auto base_control_nca = installed->GetEntry(title_id, ContentRecordType::Control);
     if (base_control_nca == nullptr)
         return {};
 
-    return ParseControlNCA(base_control_nca);
+    return ParseControlNCA(*base_control_nca);
 }
 
-std::pair<std::shared_ptr<NACP>, VirtualFile> PatchManager::ParseControlNCA(
-    const std::shared_ptr<NCA>& nca) const {
-    const auto base_romfs = nca->GetRomFS();
+std::pair<std::unique_ptr<NACP>, VirtualFile> PatchManager::ParseControlNCA(const NCA& nca) const {
+    const auto base_romfs = nca.GetRomFS();
     if (base_romfs == nullptr)
         return {};
 
-    const auto romfs = PatchRomFS(base_romfs, nca->GetBaseIVFCOffset(), ContentRecordType::Control);
+    const auto romfs = PatchRomFS(base_romfs, nca.GetBaseIVFCOffset(), ContentRecordType::Control);
     if (romfs == nullptr)
         return {};
 
@@ -319,7 +377,7 @@ std::pair<std::shared_ptr<NACP>, VirtualFile> PatchManager::ParseControlNCA(
     if (nacp_file == nullptr)
         nacp_file = extracted->GetFile("Control.nacp");
 
-    const auto nacp = nacp_file == nullptr ? nullptr : std::make_shared<NACP>(nacp_file);
+    auto nacp = nacp_file == nullptr ? nullptr : std::make_unique<NACP>(nacp_file);
 
     VirtualFile icon_file;
     for (const auto& language : FileSys::LANGUAGE_NAMES) {
@@ -328,6 +386,6 @@ std::pair<std::shared_ptr<NACP>, VirtualFile> PatchManager::ParseControlNCA(
             break;
     }
 
-    return {nacp, icon_file};
+    return {std::move(nacp), icon_file};
 }
 } // namespace FileSys
diff --git a/src/core/file_sys/patch_manager.h b/src/core/file_sys/patch_manager.h
index 6a864ec43..7d168837f 100644
--- a/src/core/file_sys/patch_manager.h
+++ b/src/core/file_sys/patch_manager.h
@@ -36,6 +36,7 @@ public:
 
     // Currently tracked NSO patches:
     // - IPS
+    // - IPSwitch
     std::vector<u8> PatchNSO(const std::vector<u8>& nso) const;
 
     // Checks to see if PatchNSO() will have any effect given the NSO's build ID.
@@ -46,19 +47,20 @@ public:
     // - Game Updates
     // - LayeredFS
     VirtualFile PatchRomFS(VirtualFile base, u64 ivfc_offset,
-                           ContentRecordType type = ContentRecordType::Program) const;
+                           ContentRecordType type = ContentRecordType::Program,
+                           VirtualFile update_raw = nullptr) const;
 
     // Returns a vector of pairs between patch names and patch versions.
     // i.e. Update 3.2.2 will return {"Update", "3.2.2"}
-    std::map<std::string, std::string, std::less<>> GetPatchVersionNames() const;
+    std::map<std::string, std::string, std::less<>> GetPatchVersionNames(
+        VirtualFile update_raw = nullptr) const;
 
     // Given title_id of the program, attempts to get the control data of the update and parse it,
     // falling back to the base control data.
-    std::pair<std::shared_ptr<NACP>, VirtualFile> GetControlMetadata() const;
+    std::pair<std::unique_ptr<NACP>, VirtualFile> GetControlMetadata() const;
 
     // Version of GetControlMetadata that takes an arbitrary NCA
-    std::pair<std::shared_ptr<NACP>, VirtualFile> ParseControlNCA(
-        const std::shared_ptr<NCA>& nca) const;
+    std::pair<std::unique_ptr<NACP>, VirtualFile> ParseControlNCA(const NCA& nca) const;
 
 private:
     u64 title_id;
diff --git a/src/core/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp
index e9b040689..96302a241 100644
--- a/src/core/file_sys/registered_cache.cpp
+++ b/src/core/file_sys/registered_cache.cpp
@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include <regex>
 #include <mbedtls/sha256.h>
 #include "common/assert.h"
@@ -30,6 +31,14 @@ bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs)
     return (lhs.title_id < rhs.title_id) || (lhs.title_id == rhs.title_id && lhs.type < rhs.type);
 }
 
+bool operator==(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs) {
+    return std::tie(lhs.title_id, lhs.type) == std::tie(rhs.title_id, rhs.type);
+}
+
+bool operator!=(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs) {
+    return !operator==(lhs, rhs);
+}
+
 static bool FollowsTwoDigitDirFormat(std::string_view name) {
     static const std::regex two_digit_regex("000000[0-9A-F]{2}", std::regex_constants::ECMAScript |
                                                                      std::regex_constants::icase);
@@ -150,28 +159,28 @@ VirtualFile RegisteredCache::GetFileAtID(NcaID id) const {
     return file;
 }
 
-static boost::optional<NcaID> CheckMapForContentRecord(
+static std::optional<NcaID> CheckMapForContentRecord(
     const boost::container::flat_map<u64, CNMT>& map, u64 title_id, ContentRecordType type) {
     if (map.find(title_id) == map.end())
-        return boost::none;
+        return {};
 
     const auto& cnmt = map.at(title_id);
 
     const auto iter = std::find_if(cnmt.GetContentRecords().begin(), cnmt.GetContentRecords().end(),
                                    [type](const ContentRecord& rec) { return rec.type == type; });
     if (iter == cnmt.GetContentRecords().end())
-        return boost::none;
+        return {};
 
-    return boost::make_optional(iter->nca_id);
+    return std::make_optional(iter->nca_id);
 }
 
-boost::optional<NcaID> RegisteredCache::GetNcaIDFromMetadata(u64 title_id,
-                                                             ContentRecordType type) const {
+std::optional<NcaID> RegisteredCache::GetNcaIDFromMetadata(u64 title_id,
+                                                           ContentRecordType type) const {
     if (type == ContentRecordType::Meta && meta_id.find(title_id) != meta_id.end())
         return meta_id.at(title_id);
 
     const auto res1 = CheckMapForContentRecord(yuzu_meta, title_id, type);
-    if (res1 != boost::none)
+    if (res1)
         return res1;
     return CheckMapForContentRecord(meta, title_id, type);
 }
@@ -274,17 +283,14 @@ bool RegisteredCache::HasEntry(RegisteredCacheEntry entry) const {
 
 VirtualFile RegisteredCache::GetEntryUnparsed(u64 title_id, ContentRecordType type) const {
     const auto id = GetNcaIDFromMetadata(title_id, type);
-    if (id == boost::none)
-        return nullptr;
-
-    return GetFileAtID(id.get());
+    return id ? GetFileAtID(*id) : nullptr;
 }
 
 VirtualFile RegisteredCache::GetEntryUnparsed(RegisteredCacheEntry entry) const {
     return GetEntryUnparsed(entry.title_id, entry.type);
 }
 
-boost::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const {
+std::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const {
     const auto meta_iter = meta.find(title_id);
     if (meta_iter != meta.end())
         return meta_iter->second.GetTitleVersion();
@@ -293,29 +299,26 @@ boost::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const {
     if (yuzu_meta_iter != yuzu_meta.end())
         return yuzu_meta_iter->second.GetTitleVersion();
 
-    return boost::none;
+    return {};
 }
 
 VirtualFile RegisteredCache::GetEntryRaw(u64 title_id, ContentRecordType type) const {
     const auto id = GetNcaIDFromMetadata(title_id, type);
-    if (id == boost::none)
-        return nullptr;
-
-    return parser(GetFileAtID(id.get()), id.get());
+    return id ? parser(GetFileAtID(*id), *id) : nullptr;
 }
 
 VirtualFile RegisteredCache::GetEntryRaw(RegisteredCacheEntry entry) const {
     return GetEntryRaw(entry.title_id, entry.type);
 }
 
-std::shared_ptr<NCA> RegisteredCache::GetEntry(u64 title_id, ContentRecordType type) const {
+std::unique_ptr<NCA> RegisteredCache::GetEntry(u64 title_id, ContentRecordType type) const {
     const auto raw = GetEntryRaw(title_id, type);
     if (raw == nullptr)
         return nullptr;
-    return std::make_shared<NCA>(raw);
+    return std::make_unique<NCA>(raw);
 }
 
-std::shared_ptr<NCA> RegisteredCache::GetEntry(RegisteredCacheEntry entry) const {
+std::unique_ptr<NCA> RegisteredCache::GetEntry(RegisteredCacheEntry entry) const {
     return GetEntry(entry.title_id, entry.type);
 }
 
@@ -355,8 +358,8 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntries() const {
 }
 
 std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter(
-    boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type,
-    boost::optional<u64> title_id) const {
+    std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type,
+    std::optional<u64> title_id) const {
     std::vector<RegisteredCacheEntry> out;
     IterateAllMetadata<RegisteredCacheEntry>(
         out,
@@ -364,11 +367,11 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter(
             return RegisteredCacheEntry{c.GetTitleID(), r.type};
         },
         [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) {
-            if (title_type != boost::none && title_type.get() != c.GetType())
+            if (title_type && *title_type != c.GetType())
                 return false;
-            if (record_type != boost::none && record_type.get() != r.type)
+            if (record_type && *record_type != r.type)
                 return false;
-            if (title_id != boost::none && title_id.get() != c.GetTitleID())
+            if (title_id && *title_id != c.GetTitleID())
                 return false;
             return true;
         });
@@ -450,7 +453,7 @@ InstallResult RegisteredCache::InstallEntry(std::shared_ptr<NCA> nca, TitleType
 
 InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy,
                                              bool overwrite_if_exists,
-                                             boost::optional<NcaID> override_id) {
+                                             std::optional<NcaID> override_id) {
     const auto in = nca->GetBaseFile();
     Core::Crypto::SHA256Hash hash{};
 
@@ -459,12 +462,12 @@ InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const Vfs
     // game is massive), we're going to cheat and only hash the first MB of the NCA.
     // Also, for XCIs the NcaID matters, so if the override id isn't none, use that.
     NcaID id{};
-    if (override_id == boost::none) {
+    if (override_id) {
+        id = *override_id;
+    } else {
         const auto& data = in->ReadBytes(0x100000);
         mbedtls_sha256(data.data(), data.size(), hash.data(), 0);
         memcpy(id.data(), hash.data(), 16);
-    } else {
-        id = override_id.get();
     }
 
     std::string path = GetRelativePathFromNcaID(id, false, true);
@@ -516,7 +519,7 @@ bool RegisteredCache::RawInstallYuzuMeta(const CNMT& cnmt) {
                         }) != yuzu_meta.end();
 }
 
-RegisteredCacheUnion::RegisteredCacheUnion(std::vector<std::shared_ptr<RegisteredCache>> caches)
+RegisteredCacheUnion::RegisteredCacheUnion(std::vector<RegisteredCache*> caches)
     : caches(std::move(caches)) {}
 
 void RegisteredCacheUnion::Refresh() {
@@ -534,14 +537,14 @@ bool RegisteredCacheUnion::HasEntry(RegisteredCacheEntry entry) const {
     return HasEntry(entry.title_id, entry.type);
 }
 
-boost::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const {
+std::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const {
     for (const auto& c : caches) {
         const auto res = c->GetEntryVersion(title_id);
-        if (res != boost::none)
+        if (res)
             return res;
     }
 
-    return boost::none;
+    return {};
 }
 
 VirtualFile RegisteredCacheUnion::GetEntryUnparsed(u64 title_id, ContentRecordType type) const {
@@ -572,14 +575,14 @@ VirtualFile RegisteredCacheUnion::GetEntryRaw(RegisteredCacheEntry entry) const
     return GetEntryRaw(entry.title_id, entry.type);
 }
 
-std::shared_ptr<NCA> RegisteredCacheUnion::GetEntry(u64 title_id, ContentRecordType type) const {
+std::unique_ptr<NCA> RegisteredCacheUnion::GetEntry(u64 title_id, ContentRecordType type) const {
     const auto raw = GetEntryRaw(title_id, type);
     if (raw == nullptr)
         return nullptr;
-    return std::make_shared<NCA>(raw);
+    return std::make_unique<NCA>(raw);
 }
 
-std::shared_ptr<NCA> RegisteredCacheUnion::GetEntry(RegisteredCacheEntry entry) const {
+std::unique_ptr<NCA> RegisteredCacheUnion::GetEntry(RegisteredCacheEntry entry) const {
     return GetEntry(entry.title_id, entry.type);
 }
 
@@ -593,12 +596,15 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntries() const {
             },
             [](const CNMT& c, const ContentRecord& r) { return true; });
     }
+
+    std::sort(out.begin(), out.end());
+    out.erase(std::unique(out.begin(), out.end()), out.end());
     return out;
 }
 
 std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter(
-    boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type,
-    boost::optional<u64> title_id) const {
+    std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type,
+    std::optional<u64> title_id) const {
     std::vector<RegisteredCacheEntry> out;
     for (const auto& c : caches) {
         c->IterateAllMetadata<RegisteredCacheEntry>(
@@ -607,15 +613,18 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter(
                 return RegisteredCacheEntry{c.GetTitleID(), r.type};
             },
             [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) {
-                if (title_type != boost::none && title_type.get() != c.GetType())
+                if (title_type && *title_type != c.GetType())
                     return false;
-                if (record_type != boost::none && record_type.get() != r.type)
+                if (record_type && *record_type != r.type)
                     return false;
-                if (title_id != boost::none && title_id.get() != c.GetTitleID())
+                if (title_id && *title_id != c.GetTitleID())
                     return false;
                 return true;
             });
     }
+
+    std::sort(out.begin(), out.end());
+    out.erase(std::unique(out.begin(), out.end()), out.end());
     return out;
 }
 } // namespace FileSys
diff --git a/src/core/file_sys/registered_cache.h b/src/core/file_sys/registered_cache.h
index c0cd59fc5..6cfb16017 100644
--- a/src/core/file_sys/registered_cache.h
+++ b/src/core/file_sys/registered_cache.h
@@ -50,6 +50,10 @@ constexpr u64 GetUpdateTitleID(u64 base_title_id) {
 // boost flat_map requires operator< for O(log(n)) lookups.
 bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs);
 
+// std unique requires operator== to identify duplicates.
+bool operator==(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs);
+bool operator!=(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs);
+
 /*
  * A class that catalogues NCAs in the registered directory structure.
  * Nintendo's registered format follows this structure:
@@ -60,8 +64,8 @@ bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs)
  *         | 00
  *         | 01 <- Actual content split along 4GB boundaries. (optional)
  *
- * (This impl also supports substituting the nca dir for an nca file, as that's more convenient when
- * 4GB splitting can be ignored.)
+ * (This impl also supports substituting the nca dir for an nca file, as that's more convenient
+ * when 4GB splitting can be ignored.)
  */
 class RegisteredCache {
     friend class RegisteredCacheUnion;
@@ -80,7 +84,7 @@ public:
     bool HasEntry(u64 title_id, ContentRecordType type) const;
     bool HasEntry(RegisteredCacheEntry entry) const;
 
-    boost::optional<u32> GetEntryVersion(u64 title_id) const;
+    std::optional<u32> GetEntryVersion(u64 title_id) const;
 
     VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const;
     VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const;
@@ -88,15 +92,14 @@ public:
     VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const;
     VirtualFile GetEntryRaw(RegisteredCacheEntry entry) const;
 
-    std::shared_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const;
-    std::shared_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const;
+    std::unique_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const;
+    std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const;
 
     std::vector<RegisteredCacheEntry> ListEntries() const;
-    // If a parameter is not boost::none, it will be filtered for from all entries.
+    // If a parameter is not std::nullopt, it will be filtered for from all entries.
     std::vector<RegisteredCacheEntry> ListEntriesFilter(
-        boost::optional<TitleType> title_type = boost::none,
-        boost::optional<ContentRecordType> record_type = boost::none,
-        boost::optional<u64> title_id = boost::none) const;
+        std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {},
+        std::optional<u64> title_id = {}) const;
 
     // Raw copies all the ncas from the xci/nsp to the csache. Does some quick checks to make sure
     // there is a meta NCA and all of them are accessible.
@@ -121,12 +124,11 @@ private:
     std::vector<NcaID> AccumulateFiles() const;
     void ProcessFiles(const std::vector<NcaID>& ids);
     void AccumulateYuzuMeta();
-    boost::optional<NcaID> GetNcaIDFromMetadata(u64 title_id, ContentRecordType type) const;
+    std::optional<NcaID> GetNcaIDFromMetadata(u64 title_id, ContentRecordType type) const;
     VirtualFile GetFileAtID(NcaID id) const;
     VirtualFile OpenFileOrDirectoryConcat(const VirtualDir& dir, std::string_view path) const;
     InstallResult RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy,
-                                bool overwrite_if_exists,
-                                boost::optional<NcaID> override_id = boost::none);
+                                bool overwrite_if_exists, std::optional<NcaID> override_id = {});
     bool RawInstallYuzuMeta(const CNMT& cnmt);
 
     VirtualDir dir;
@@ -142,14 +144,14 @@ private:
 // Combines multiple RegisteredCaches (i.e. SysNAND, UserNAND, SDMC) into one interface.
 class RegisteredCacheUnion {
 public:
-    explicit RegisteredCacheUnion(std::vector<std::shared_ptr<RegisteredCache>> caches);
+    explicit RegisteredCacheUnion(std::vector<RegisteredCache*> caches);
 
     void Refresh();
 
     bool HasEntry(u64 title_id, ContentRecordType type) const;
     bool HasEntry(RegisteredCacheEntry entry) const;
 
-    boost::optional<u32> GetEntryVersion(u64 title_id) const;
+    std::optional<u32> GetEntryVersion(u64 title_id) const;
 
     VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const;
     VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const;
@@ -157,18 +159,17 @@ public:
     VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const;
     VirtualFile GetEntryRaw(RegisteredCacheEntry entry) const;
 
-    std::shared_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const;
-    std::shared_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const;
+    std::unique_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const;
+    std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const;
 
     std::vector<RegisteredCacheEntry> ListEntries() const;
-    // If a parameter is not boost::none, it will be filtered for from all entries.
+    // If a parameter is not std::nullopt, it will be filtered for from all entries.
     std::vector<RegisteredCacheEntry> ListEntriesFilter(
-        boost::optional<TitleType> title_type = boost::none,
-        boost::optional<ContentRecordType> record_type = boost::none,
-        boost::optional<u64> title_id = boost::none) const;
+        std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {},
+        std::optional<u64> title_id = {}) const;
 
 private:
-    std::vector<std::shared_ptr<RegisteredCache>> caches;
+    std::vector<RegisteredCache*> caches;
 };
 
 } // namespace FileSys
diff --git a/src/core/file_sys/romfs.cpp b/src/core/file_sys/romfs.cpp
index 5910f7046..81e1f66ac 100644
--- a/src/core/file_sys/romfs.cpp
+++ b/src/core/file_sys/romfs.cpp
@@ -129,11 +129,11 @@ VirtualDir ExtractRomFS(VirtualFile file, RomFSExtractionType type) {
     return out;
 }
 
-VirtualFile CreateRomFS(VirtualDir dir) {
+VirtualFile CreateRomFS(VirtualDir dir, VirtualDir ext) {
     if (dir == nullptr)
         return nullptr;
 
-    RomFSBuildContext ctx{dir};
+    RomFSBuildContext ctx{dir, ext};
     return ConcatenatedVfsFile::MakeConcatenatedFile(0, ctx.Build(), dir->GetName());
 }
 
diff --git a/src/core/file_sys/romfs.h b/src/core/file_sys/romfs.h
index ecd1eb725..0ec404731 100644
--- a/src/core/file_sys/romfs.h
+++ b/src/core/file_sys/romfs.h
@@ -44,6 +44,6 @@ VirtualDir ExtractRomFS(VirtualFile file,
 
 // Converts a VFS filesystem into a RomFS binary
 // Returns nullptr on failure
-VirtualFile CreateRomFS(VirtualDir dir);
+VirtualFile CreateRomFS(VirtualDir dir, VirtualDir ext = nullptr);
 
 } // namespace FileSys
diff --git a/src/core/file_sys/romfs_factory.cpp b/src/core/file_sys/romfs_factory.cpp
index 4994c2532..0b645b106 100644
--- a/src/core/file_sys/romfs_factory.cpp
+++ b/src/core/file_sys/romfs_factory.cpp
@@ -30,12 +30,17 @@ RomFSFactory::RomFSFactory(Loader::AppLoader& app_loader) {
 
 RomFSFactory::~RomFSFactory() = default;
 
+void RomFSFactory::SetPackedUpdate(VirtualFile update_raw) {
+    this->update_raw = std::move(update_raw);
+}
+
 ResultVal<VirtualFile> RomFSFactory::OpenCurrentProcess() {
     if (!updatable)
         return MakeResult<VirtualFile>(file);
 
     const PatchManager patch_manager(Core::CurrentProcess()->GetTitleID());
-    return MakeResult<VirtualFile>(patch_manager.PatchRomFS(file, ivfc_offset));
+    return MakeResult<VirtualFile>(
+        patch_manager.PatchRomFS(file, ivfc_offset, ContentRecordType::Program, update_raw));
 }
 
 ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage, ContentRecordType type) {
diff --git a/src/core/file_sys/romfs_factory.h b/src/core/file_sys/romfs_factory.h
index 2cace8180..7724c0b23 100644
--- a/src/core/file_sys/romfs_factory.h
+++ b/src/core/file_sys/romfs_factory.h
@@ -32,11 +32,13 @@ public:
     explicit RomFSFactory(Loader::AppLoader& app_loader);
     ~RomFSFactory();
 
+    void SetPackedUpdate(VirtualFile update_raw);
     ResultVal<VirtualFile> OpenCurrentProcess();
     ResultVal<VirtualFile> Open(u64 title_id, StorageId storage, ContentRecordType type);
 
 private:
     VirtualFile file;
+    VirtualFile update_raw;
     bool updatable;
     u64 ivfc_offset;
 };
diff --git a/src/core/file_sys/savedata_factory.cpp b/src/core/file_sys/savedata_factory.cpp
index 47f2ab9e0..ef1aaebbb 100644
--- a/src/core/file_sys/savedata_factory.cpp
+++ b/src/core/file_sys/savedata_factory.cpp
@@ -51,6 +51,13 @@ ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, SaveDataDescr
                     meta.title_id);
     }
 
+    if (meta.type == SaveDataType::DeviceSaveData && meta.user_id != u128{0, 0}) {
+        LOG_WARNING(Service_FS,
+                    "Possibly incorrect SaveDataDescriptor, type is DeviceSaveData but user_id is "
+                    "non-zero ({:016X}{:016X})",
+                    meta.user_id[1], meta.user_id[0]);
+    }
+
     std::string save_directory =
         GetFullPath(space, meta.type, meta.title_id, meta.user_id, meta.save_id);
 
@@ -92,6 +99,9 @@ std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType typ
     case SaveDataSpaceId::NandUser:
         out = "/user/";
         break;
+    case SaveDataSpaceId::TemporaryStorage:
+        out = "/temp/";
+        break;
     default:
         ASSERT_MSG(false, "Unrecognized SaveDataSpaceId: {:02X}", static_cast<u8>(space));
     }
@@ -100,10 +110,11 @@ std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType typ
     case SaveDataType::SystemSaveData:
         return fmt::format("{}save/{:016X}/{:016X}{:016X}", out, save_id, user_id[1], user_id[0]);
     case SaveDataType::SaveData:
+    case SaveDataType::DeviceSaveData:
         return fmt::format("{}save/{:016X}/{:016X}{:016X}/{:016X}", out, 0, user_id[1], user_id[0],
                            title_id);
     case SaveDataType::TemporaryStorage:
-        return fmt::format("{}temp/{:016X}/{:016X}{:016X}/{:016X}", out, 0, user_id[1], user_id[0],
+        return fmt::format("{}{:016X}/{:016X}{:016X}/{:016X}", out, 0, user_id[1], user_id[0],
                            title_id);
     default:
         ASSERT_MSG(false, "Unrecognized SaveDataType: {:02X}", static_cast<u8>(type));
diff --git a/src/core/file_sys/sdmc_factory.cpp b/src/core/file_sys/sdmc_factory.cpp
index d66a9c9a4..bd3a57058 100644
--- a/src/core/file_sys/sdmc_factory.cpp
+++ b/src/core/file_sys/sdmc_factory.cpp
@@ -10,10 +10,10 @@
 namespace FileSys {
 
 SDMCFactory::SDMCFactory(VirtualDir dir_)
-    : dir(std::move(dir_)), contents(std::make_shared<RegisteredCache>(
+    : dir(std::move(dir_)), contents(std::make_unique<RegisteredCache>(
                                 GetOrCreateDirectoryRelative(dir, "/Nintendo/Contents/registered"),
                                 [](const VirtualFile& file, const NcaID& id) {
-                                    return std::make_shared<NAX>(file, id)->GetDecrypted();
+                                    return NAX{file, id}.GetDecrypted();
                                 })) {}
 
 SDMCFactory::~SDMCFactory() = default;
@@ -22,8 +22,8 @@ ResultVal<VirtualDir> SDMCFactory::Open() {
     return MakeResult<VirtualDir>(dir);
 }
 
-std::shared_ptr<RegisteredCache> SDMCFactory::GetSDMCContents() const {
-    return contents;
+RegisteredCache* SDMCFactory::GetSDMCContents() const {
+    return contents.get();
 }
 
 } // namespace FileSys
diff --git a/src/core/file_sys/sdmc_factory.h b/src/core/file_sys/sdmc_factory.h
index ea12149de..42794ba5b 100644
--- a/src/core/file_sys/sdmc_factory.h
+++ b/src/core/file_sys/sdmc_factory.h
@@ -19,12 +19,12 @@ public:
     ~SDMCFactory();
 
     ResultVal<VirtualDir> Open();
-    std::shared_ptr<RegisteredCache> GetSDMCContents() const;
+    RegisteredCache* GetSDMCContents() const;
 
 private:
     VirtualDir dir;
 
-    std::shared_ptr<RegisteredCache> contents;
+    std::unique_ptr<RegisteredCache> contents;
 };
 
 } // namespace FileSys
diff --git a/src/core/file_sys/submission_package.cpp b/src/core/file_sys/submission_package.cpp
index 09bf077cd..2aaba4179 100644
--- a/src/core/file_sys/submission_package.cpp
+++ b/src/core/file_sys/submission_package.cpp
@@ -205,10 +205,6 @@ VirtualDir NSP::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool NSP::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
-
 void NSP::InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files) {
     exefs = pfs;
 
@@ -259,8 +255,11 @@ void NSP::ReadNCAs(const std::vector<VirtualFile>& files) {
                 auto next_nca = std::make_shared<NCA>(next_file);
                 if (next_nca->GetType() == NCAContentType::Program)
                     program_status[cnmt.GetTitleID()] = next_nca->GetStatus();
-                if (next_nca->GetStatus() == Loader::ResultStatus::Success)
+                if (next_nca->GetStatus() == Loader::ResultStatus::Success ||
+                    (next_nca->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS &&
+                     (cnmt.GetTitleID() & 0x800) != 0)) {
                     ncas_title[rec.type] = std::move(next_nca);
+                }
             }
 
             break;
diff --git a/src/core/file_sys/submission_package.h b/src/core/file_sys/submission_package.h
index da3dc5e9f..338080b7e 100644
--- a/src/core/file_sys/submission_package.h
+++ b/src/core/file_sys/submission_package.h
@@ -55,9 +55,6 @@ public:
 
     VirtualDir GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     void InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files);
     void ReadNCAs(const std::vector<VirtualFile>& files);
diff --git a/src/core/file_sys/vfs.cpp b/src/core/file_sys/vfs.cpp
index bfe50da73..7b584de7f 100644
--- a/src/core/file_sys/vfs.cpp
+++ b/src/core/file_sys/vfs.cpp
@@ -167,13 +167,13 @@ std::string VfsFile::GetExtension() const {
 
 VfsDirectory::~VfsDirectory() = default;
 
-boost::optional<u8> VfsFile::ReadByte(std::size_t offset) const {
+std::optional<u8> VfsFile::ReadByte(std::size_t offset) const {
     u8 out{};
     std::size_t size = Read(&out, 1, offset);
     if (size == 1)
         return out;
 
-    return boost::none;
+    return {};
 }
 
 std::vector<u8> VfsFile::ReadBytes(std::size_t size, std::size_t offset) const {
@@ -472,10 +472,14 @@ bool VfsRawCopy(const VirtualFile& src, const VirtualFile& dest, std::size_t blo
     std::vector<u8> temp(std::min(block_size, src->GetSize()));
     for (std::size_t i = 0; i < src->GetSize(); i += block_size) {
         const auto read = std::min(block_size, src->GetSize() - i);
-        const auto block = src->Read(temp.data(), read, i);
 
-        if (dest->Write(temp.data(), read, i) != read)
+        if (src->Read(temp.data(), read, i) != read) {
             return false;
+        }
+
+        if (dest->Write(temp.data(), read, i) != read) {
+            return false;
+        }
     }
 
     return true;
diff --git a/src/core/file_sys/vfs.h b/src/core/file_sys/vfs.h
index 270291631..002f99d4e 100644
--- a/src/core/file_sys/vfs.h
+++ b/src/core/file_sys/vfs.h
@@ -4,28 +4,22 @@
 
 #pragma once
 
+#include <functional>
 #include <map>
 #include <memory>
+#include <optional>
 #include <string>
 #include <string_view>
 #include <type_traits>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/common_types.h"
+#include "core/file_sys/vfs_types.h"
 
 namespace FileSys {
 
-class VfsDirectory;
-class VfsFile;
-class VfsFilesystem;
-
 enum class Mode : u32;
 
-// Convenience typedefs to use Vfs* interfaces
-using VirtualFilesystem = std::shared_ptr<VfsFilesystem>;
-using VirtualDir = std::shared_ptr<VfsDirectory>;
-using VirtualFile = std::shared_ptr<VfsFile>;
-
 // An enumeration representing what can be at the end of a path in a VfsFilesystem
 enum class VfsEntryType {
     None,
@@ -111,8 +105,8 @@ public:
     // into file. Returns number of bytes successfully written.
     virtual std::size_t Write(const u8* data, std::size_t length, std::size_t offset = 0) = 0;
 
-    // Reads exactly one byte at the offset provided, returning boost::none on error.
-    virtual boost::optional<u8> ReadByte(std::size_t offset = 0) const;
+    // Reads exactly one byte at the offset provided, returning std::nullopt on error.
+    virtual std::optional<u8> ReadByte(std::size_t offset = 0) const;
     // Reads size bytes starting at offset in file into a vector.
     virtual std::vector<u8> ReadBytes(std::size_t size, std::size_t offset = 0) const;
     // Reads all the bytes from the file into a vector. Equivalent to 'file->Read(file->GetSize(),
@@ -270,36 +264,8 @@ public:
     // item name -> type.
     virtual std::map<std::string, VfsEntryType, std::less<>> GetEntries() const;
 
-    // Interprets the file with name file instead as a directory of type directory.
-    // The directory must have a constructor that takes a single argument of type
-    // std::shared_ptr<VfsFile>. Allows to reinterpret container files (i.e NCA, zip, XCI, etc) as a
-    // subdirectory in one call.
-    template <typename Directory>
-    bool InterpretAsDirectory(std::string_view file) {
-        auto file_p = GetFile(file);
-
-        if (file_p == nullptr) {
-            return false;
-        }
-
-        return ReplaceFileWithSubdirectory(file_p, std::make_shared<Directory>(file_p));
-    }
-
-    bool InterpretAsDirectory(const std::function<VirtualDir(VirtualFile)>& function,
-                              const std::string& file) {
-        auto file_p = GetFile(file);
-        if (file_p == nullptr)
-            return false;
-        return ReplaceFileWithSubdirectory(file_p, function(file_p));
-    }
-
     // Returns the full path of this directory as a string, recursively
     virtual std::string GetFullPath() const;
-
-protected:
-    // Backend for InterpretAsDirectory.
-    // Removes all references to file and adds a reference to dir in the directory's implementation.
-    virtual bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) = 0;
 };
 
 // A convenience partial-implementation of VfsDirectory that stubs out methods that should only work
diff --git a/src/core/file_sys/vfs_layered.cpp b/src/core/file_sys/vfs_layered.cpp
index bfee01725..338e398da 100644
--- a/src/core/file_sys/vfs_layered.cpp
+++ b/src/core/file_sys/vfs_layered.cpp
@@ -126,7 +126,4 @@ bool LayeredVfsDirectory::Rename(std::string_view name_) {
     return true;
 }
 
-bool LayeredVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys
diff --git a/src/core/file_sys/vfs_layered.h b/src/core/file_sys/vfs_layered.h
index d85310f57..8a25c3428 100644
--- a/src/core/file_sys/vfs_layered.h
+++ b/src/core/file_sys/vfs_layered.h
@@ -39,9 +39,6 @@ public:
     bool DeleteFile(std::string_view name) override;
     bool Rename(std::string_view name) override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     std::vector<VirtualDir> dirs;
     std::string name;
diff --git a/src/core/file_sys/vfs_offset.cpp b/src/core/file_sys/vfs_offset.cpp
index a4c6719a0..c96f88488 100644
--- a/src/core/file_sys/vfs_offset.cpp
+++ b/src/core/file_sys/vfs_offset.cpp
@@ -57,11 +57,11 @@ std::size_t OffsetVfsFile::Write(const u8* data, std::size_t length, std::size_t
     return file->Write(data, TrimToFit(length, r_offset), offset + r_offset);
 }
 
-boost::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const {
+std::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const {
     if (r_offset < size)
         return file->ReadByte(offset + r_offset);
 
-    return boost::none;
+    return {};
 }
 
 std::vector<u8> OffsetVfsFile::ReadBytes(std::size_t r_size, std::size_t r_offset) const {
diff --git a/src/core/file_sys/vfs_offset.h b/src/core/file_sys/vfs_offset.h
index 8062702a7..f7b7a3256 100644
--- a/src/core/file_sys/vfs_offset.h
+++ b/src/core/file_sys/vfs_offset.h
@@ -29,7 +29,7 @@ public:
     bool IsReadable() const override;
     std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override;
     std::size_t Write(const u8* data, std::size_t length, std::size_t offset) override;
-    boost::optional<u8> ReadByte(std::size_t offset) const override;
+    std::optional<u8> ReadByte(std::size_t offset) const override;
     std::vector<u8> ReadBytes(std::size_t size, std::size_t offset) const override;
     std::vector<u8> ReadAllBytes() const override;
     bool WriteByte(u8 data, std::size_t offset) override;
diff --git a/src/core/file_sys/vfs_real.cpp b/src/core/file_sys/vfs_real.cpp
index 9defad04c..e21300a7c 100644
--- a/src/core/file_sys/vfs_real.cpp
+++ b/src/core/file_sys/vfs_real.cpp
@@ -430,7 +430,4 @@ std::map<std::string, VfsEntryType, std::less<>> RealVfsDirectory::GetEntries()
     return out;
 }
 
-bool RealVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys
diff --git a/src/core/file_sys/vfs_real.h b/src/core/file_sys/vfs_real.h
index 5b61db90d..a0a857a31 100644
--- a/src/core/file_sys/vfs_real.h
+++ b/src/core/file_sys/vfs_real.h
@@ -100,9 +100,6 @@ public:
     std::string GetFullPath() const override;
     std::map<std::string, VfsEntryType, std::less<>> GetEntries() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     RealVfsDirectory(RealVfsFilesystem& base, const std::string& path, Mode perms = Mode::Read);
 
diff --git a/src/core/file_sys/vfs_static.h b/src/core/file_sys/vfs_static.h
index 44fab51d1..9f5a90b1b 100644
--- a/src/core/file_sys/vfs_static.h
+++ b/src/core/file_sys/vfs_static.h
@@ -53,10 +53,10 @@ public:
         return 0;
     }
 
-    boost::optional<u8> ReadByte(std::size_t offset) const override {
+    std::optional<u8> ReadByte(std::size_t offset) const override {
         if (offset < size)
             return value;
-        return boost::none;
+        return {};
     }
 
     std::vector<u8> ReadBytes(std::size_t length, std::size_t offset) const override {
diff --git a/src/core/file_sys/vfs_types.h b/src/core/file_sys/vfs_types.h
new file mode 100644
index 000000000..6215ed7af
--- /dev/null
+++ b/src/core/file_sys/vfs_types.h
@@ -0,0 +1,21 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <memory>
+
+namespace FileSys {
+
+class VfsDirectory;
+class VfsFile;
+class VfsFilesystem;
+
+// Declarations for Vfs* pointer types
+
+using VirtualDir = std::shared_ptr<VfsDirectory>;
+using VirtualFile = std::shared_ptr<VfsFile>;
+using VirtualFilesystem = std::shared_ptr<VfsFilesystem>;
+
+} // namespace FileSys
diff --git a/src/core/file_sys/vfs_vector.cpp b/src/core/file_sys/vfs_vector.cpp
index 389c7e003..808f31e81 100644
--- a/src/core/file_sys/vfs_vector.cpp
+++ b/src/core/file_sys/vfs_vector.cpp
@@ -132,11 +132,4 @@ void VectorVfsDirectory::AddFile(VirtualFile file) {
 void VectorVfsDirectory::AddDirectory(VirtualDir dir) {
     dirs.push_back(std::move(dir));
 }
-
-bool VectorVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    if (!DeleteFile(file->GetName()))
-        return false;
-    dirs.emplace_back(std::move(dir));
-    return true;
-}
 } // namespace FileSys
diff --git a/src/core/file_sys/vfs_vector.h b/src/core/file_sys/vfs_vector.h
index 48a414c98..3e3f790c3 100644
--- a/src/core/file_sys/vfs_vector.h
+++ b/src/core/file_sys/vfs_vector.h
@@ -57,9 +57,6 @@ public:
     virtual void AddFile(VirtualFile file);
     virtual void AddDirectory(VirtualDir dir);
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     std::vector<VirtualFile> files;
     std::vector<VirtualDir> dirs;
diff --git a/src/core/file_sys/xts_archive.cpp b/src/core/file_sys/xts_archive.cpp
index b2b164368..eec51c64e 100644
--- a/src/core/file_sys/xts_archive.cpp
+++ b/src/core/file_sys/xts_archive.cpp
@@ -163,7 +163,4 @@ std::shared_ptr<VfsDirectory> NAX::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool NAX::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys
diff --git a/src/core/file_sys/xts_archive.h b/src/core/file_sys/xts_archive.h
index 8fedd8585..7704dee90 100644
--- a/src/core/file_sys/xts_archive.h
+++ b/src/core/file_sys/xts_archive.h
@@ -51,9 +51,6 @@ public:
 
     std::shared_ptr<VfsDirectory> GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     Loader::ResultStatus Parse(std::string_view path);
 
diff --git a/src/core/gdbstub/gdbstub.cpp b/src/core/gdbstub/gdbstub.cpp
index e961ef121..bdcc889e0 100644
--- a/src/core/gdbstub/gdbstub.cpp
+++ b/src/core/gdbstub/gdbstub.cpp
@@ -207,7 +207,7 @@ void RegisterModule(std::string name, VAddr beg, VAddr end, bool add_elf_ext) {
 
 static Kernel::Thread* FindThreadById(int id) {
     for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-        const auto& threads = Core::System::GetInstance().Scheduler(core)->GetThreadList();
+        const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
         for (auto& thread : threads) {
             if (thread->GetThreadID() == static_cast<u32>(id)) {
                 current_core = core;
@@ -597,7 +597,7 @@ static void HandleQuery() {
     } else if (strncmp(query, "fThreadInfo", strlen("fThreadInfo")) == 0) {
         std::string val = "m";
         for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-            const auto& threads = Core::System::GetInstance().Scheduler(core)->GetThreadList();
+            const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
             for (const auto& thread : threads) {
                 val += fmt::format("{:x}", thread->GetThreadID());
                 val += ",";
@@ -612,7 +612,7 @@ static void HandleQuery() {
         buffer += "l<?xml version=\"1.0\"?>";
         buffer += "<threads>";
         for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-            const auto& threads = Core::System::GetInstance().Scheduler(core)->GetThreadList();
+            const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
             for (const auto& thread : threads) {
                 buffer +=
                     fmt::format(R"*(<thread id="{:x}" core="{:d}" name="Thread {:x}"></thread>)*",
diff --git a/src/core/hle/ipc.h b/src/core/hle/ipc.h
index 419f45896..ed84197b3 100644
--- a/src/core/hle/ipc.h
+++ b/src/core/hle/ipc.h
@@ -14,11 +14,6 @@ namespace IPC {
 /// Size of the command buffer area, in 32-bit words.
 constexpr std::size_t COMMAND_BUFFER_LENGTH = 0x100 / sizeof(u32);
 
-// These errors are commonly returned by invalid IPC translations, so alias them here for
-// convenience.
-// TODO(yuriks): These will probably go away once translation is implemented inside the kernel.
-constexpr auto ERR_INVALID_HANDLE = Kernel::ERR_INVALID_HANDLE_OS;
-
 enum class ControlCommand : u32 {
     ConvertSessionToDomain = 0,
     ConvertDomainToSession = 1,
diff --git a/src/core/hle/ipc_helpers.h b/src/core/hle/ipc_helpers.h
index a4bfe2eb0..0a7142ada 100644
--- a/src/core/hle/ipc_helpers.h
+++ b/src/core/hle/ipc_helpers.h
@@ -117,8 +117,7 @@ public:
 
         AlignWithPadding();
 
-        const bool request_has_domain_header{context.GetDomainMessageHeader() != nullptr};
-        if (context.Session()->IsDomain() && request_has_domain_header) {
+        if (context.Session()->IsDomain() && context.HasDomainMessageHeader()) {
             IPC::DomainMessageHeader domain_header{};
             domain_header.num_objects = num_domain_objects;
             PushRaw(domain_header);
diff --git a/src/core/hle/kernel/address_arbiter.cpp b/src/core/hle/kernel/address_arbiter.cpp
index ebf193930..57157beb4 100644
--- a/src/core/hle/kernel/address_arbiter.cpp
+++ b/src/core/hle/kernel/address_arbiter.cpp
@@ -39,7 +39,7 @@ static std::vector<SharedPtr<Thread>> GetThreadsWaitingOnAddress(VAddr address)
                                            std::vector<SharedPtr<Thread>>& waiting_threads,
                                            VAddr arb_addr) {
         const auto& scheduler = Core::System::GetInstance().Scheduler(core_index);
-        const auto& thread_list = scheduler->GetThreadList();
+        const auto& thread_list = scheduler.GetThreadList();
 
         for (const auto& thread : thread_list) {
             if (thread->GetArbiterWaitAddress() == arb_addr)
diff --git a/src/core/hle/kernel/errors.h b/src/core/hle/kernel/errors.h
index e5fa67ae8..ee698c8a7 100644
--- a/src/core/hle/kernel/errors.h
+++ b/src/core/hle/kernel/errors.h
@@ -10,11 +10,6 @@ namespace Kernel {
 
 namespace ErrCodes {
 enum {
-    // TODO(Subv): Remove these 3DS OS error codes.
-    SessionClosedByRemote = 26,
-    NoPendingSessions = 35,
-    InvalidBufferDescriptor = 48,
-
     // Confirmed Switch OS error codes
     MaxConnectionsReached = 7,
     InvalidSize = 101,
@@ -22,9 +17,11 @@ enum {
     HandleTableFull = 105,
     InvalidMemoryState = 106,
     InvalidMemoryPermissions = 108,
+    InvalidMemoryRange = 110,
     InvalidThreadPriority = 112,
     InvalidProcessorId = 113,
     InvalidHandle = 114,
+    InvalidPointer = 115,
     InvalidCombination = 116,
     Timeout = 117,
     SynchronizationCanceled = 118,
@@ -32,6 +29,7 @@ enum {
     InvalidEnumValue = 120,
     NoSuchEntry = 121,
     AlreadyRegistered = 122,
+    SessionClosed = 123,
     InvalidState = 125,
     ResourceLimitExceeded = 132,
 };
@@ -40,22 +38,19 @@ enum {
 // WARNING: The kernel is quite inconsistent in it's usage of errors code. Make sure to always
 // double check that the code matches before re-using the constant.
 
-// TODO(bunnei): Replace -1 with correct errors for Switch OS
 constexpr ResultCode ERR_HANDLE_TABLE_FULL(ErrorModule::Kernel, ErrCodes::HandleTableFull);
-constexpr ResultCode ERR_SESSION_CLOSED_BY_REMOTE(-1);
+constexpr ResultCode ERR_SESSION_CLOSED_BY_REMOTE(ErrorModule::Kernel, ErrCodes::SessionClosed);
 constexpr ResultCode ERR_PORT_NAME_TOO_LONG(ErrorModule::Kernel, ErrCodes::TooLarge);
 constexpr ResultCode ERR_MAX_CONNECTIONS_REACHED(ErrorModule::Kernel,
                                                  ErrCodes::MaxConnectionsReached);
 constexpr ResultCode ERR_INVALID_ENUM_VALUE(ErrorModule::Kernel, ErrCodes::InvalidEnumValue);
-constexpr ResultCode ERR_INVALID_ENUM_VALUE_FND(-1);
-constexpr ResultCode ERR_INVALID_COMBINATION(-1);
 constexpr ResultCode ERR_INVALID_COMBINATION_KERNEL(ErrorModule::Kernel,
                                                     ErrCodes::InvalidCombination);
-constexpr ResultCode ERR_OUT_OF_MEMORY(-1);
 constexpr ResultCode ERR_INVALID_ADDRESS(ErrorModule::Kernel, ErrCodes::InvalidAddress);
 constexpr ResultCode ERR_INVALID_ADDRESS_STATE(ErrorModule::Kernel, ErrCodes::InvalidMemoryState);
 constexpr ResultCode ERR_INVALID_MEMORY_PERMISSIONS(ErrorModule::Kernel,
                                                     ErrCodes::InvalidMemoryPermissions);
+constexpr ResultCode ERR_INVALID_MEMORY_RANGE(ErrorModule::Kernel, ErrCodes::InvalidMemoryRange);
 constexpr ResultCode ERR_INVALID_HANDLE(ErrorModule::Kernel, ErrCodes::InvalidHandle);
 constexpr ResultCode ERR_INVALID_PROCESSOR_ID(ErrorModule::Kernel, ErrCodes::InvalidProcessorId);
 constexpr ResultCode ERR_INVALID_SIZE(ErrorModule::Kernel, ErrCodes::InvalidSize);
@@ -63,14 +58,8 @@ constexpr ResultCode ERR_ALREADY_REGISTERED(ErrorModule::Kernel, ErrCodes::Alrea
 constexpr ResultCode ERR_INVALID_STATE(ErrorModule::Kernel, ErrCodes::InvalidState);
 constexpr ResultCode ERR_INVALID_THREAD_PRIORITY(ErrorModule::Kernel,
                                                  ErrCodes::InvalidThreadPriority);
-constexpr ResultCode ERR_INVALID_POINTER(-1);
-constexpr ResultCode ERR_INVALID_OBJECT_ADDR(-1);
-constexpr ResultCode ERR_NOT_AUTHORIZED(-1);
-/// Alternate code returned instead of ERR_INVALID_HANDLE in some code paths.
-constexpr ResultCode ERR_INVALID_HANDLE_OS(-1);
+constexpr ResultCode ERR_INVALID_POINTER(ErrorModule::Kernel, ErrCodes::InvalidPointer);
 constexpr ResultCode ERR_NOT_FOUND(ErrorModule::Kernel, ErrCodes::NoSuchEntry);
 constexpr ResultCode RESULT_TIMEOUT(ErrorModule::Kernel, ErrCodes::Timeout);
-/// Returned when Accept() is called on a port with no sessions to be accepted.
-constexpr ResultCode ERR_NO_PENDING_SESSIONS(-1);
 
 } // namespace Kernel
diff --git a/src/core/hle/kernel/hle_ipc.cpp b/src/core/hle/kernel/hle_ipc.cpp
index edad5f1b1..68d5376cb 100644
--- a/src/core/hle/kernel/hle_ipc.cpp
+++ b/src/core/hle/kernel/hle_ipc.cpp
@@ -77,7 +77,8 @@ HLERequestContext::HLERequestContext(SharedPtr<Kernel::ServerSession> server_ses
 
 HLERequestContext::~HLERequestContext() = default;
 
-void HLERequestContext::ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming) {
+void HLERequestContext::ParseCommandBuffer(const HandleTable& handle_table, u32_le* src_cmdbuf,
+                                           bool incoming) {
     IPC::RequestParser rp(src_cmdbuf);
     command_header = std::make_shared<IPC::CommandHeader>(rp.PopRaw<IPC::CommandHeader>());
 
@@ -94,8 +95,6 @@ void HLERequestContext::ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming) {
             rp.Skip(2, false);
         }
         if (incoming) {
-            auto& handle_table = Core::System::GetInstance().Kernel().HandleTable();
-
             // Populate the object lists with the data in the IPC request.
             for (u32 handle = 0; handle < handle_descriptor_header->num_handles_to_copy; ++handle) {
                 copy_objects.push_back(handle_table.GetGeneric(rp.Pop<Handle>()));
@@ -189,10 +188,9 @@ void HLERequestContext::ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming) {
     rp.Skip(1, false); // The command is actually an u64, but we don't use the high part.
 }
 
-ResultCode HLERequestContext::PopulateFromIncomingCommandBuffer(u32_le* src_cmdbuf,
-                                                                Process& src_process,
-                                                                HandleTable& src_table) {
-    ParseCommandBuffer(src_cmdbuf, true);
+ResultCode HLERequestContext::PopulateFromIncomingCommandBuffer(const HandleTable& handle_table,
+                                                                u32_le* src_cmdbuf) {
+    ParseCommandBuffer(handle_table, src_cmdbuf, true);
     if (command_header->type == IPC::CommandType::Close) {
         // Close does not populate the rest of the IPC header
         return RESULT_SUCCESS;
@@ -207,14 +205,17 @@ ResultCode HLERequestContext::PopulateFromIncomingCommandBuffer(u32_le* src_cmdb
     return RESULT_SUCCESS;
 }
 
-ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(const Thread& thread) {
+ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(Thread& thread) {
+    auto& owner_process = *thread.GetOwnerProcess();
+    auto& handle_table = owner_process.GetHandleTable();
+
     std::array<u32, IPC::COMMAND_BUFFER_LENGTH> dst_cmdbuf;
-    Memory::ReadBlock(*thread.GetOwnerProcess(), thread.GetTLSAddress(), dst_cmdbuf.data(),
+    Memory::ReadBlock(owner_process, thread.GetTLSAddress(), dst_cmdbuf.data(),
                       dst_cmdbuf.size() * sizeof(u32));
 
     // The header was already built in the internal command buffer. Attempt to parse it to verify
     // the integrity and then copy it over to the target command buffer.
-    ParseCommandBuffer(cmd_buf.data(), false);
+    ParseCommandBuffer(handle_table, cmd_buf.data(), false);
 
     // The data_size already includes the payload header, the padding and the domain header.
     std::size_t size = data_payload_offset + command_header->data_size -
@@ -236,8 +237,6 @@ ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(const Thread& thread)
         ASSERT(copy_objects.size() == handle_descriptor_header->num_handles_to_copy);
         ASSERT(move_objects.size() == handle_descriptor_header->num_handles_to_move);
 
-        auto& handle_table = Core::System::GetInstance().Kernel().HandleTable();
-
         // We don't make a distinction between copy and move handles when translating since HLE
         // services don't deal with handles directly. However, the guest applications might check
         // for specific values in each of these descriptors.
@@ -268,7 +267,7 @@ ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(const Thread& thread)
     }
 
     // Copy the translated command buffer back into the thread's command buffer area.
-    Memory::WriteBlock(*thread.GetOwnerProcess(), thread.GetTLSAddress(), dst_cmdbuf.data(),
+    Memory::WriteBlock(owner_process, thread.GetTLSAddress(), dst_cmdbuf.data(),
                        dst_cmdbuf.size() * sizeof(u32));
 
     return RESULT_SUCCESS;
diff --git a/src/core/hle/kernel/hle_ipc.h b/src/core/hle/kernel/hle_ipc.h
index 894479ee0..a38e34b74 100644
--- a/src/core/hle/kernel/hle_ipc.h
+++ b/src/core/hle/kernel/hle_ipc.h
@@ -24,10 +24,10 @@ class ServiceFrameworkBase;
 namespace Kernel {
 
 class Domain;
+class Event;
 class HandleTable;
 class HLERequestContext;
 class Process;
-class Event;
 
 /**
  * Interface implemented by HLE Session handlers.
@@ -126,13 +126,12 @@ public:
                                        u64 timeout, WakeupCallback&& callback,
                                        Kernel::SharedPtr<Kernel::Event> event = nullptr);
 
-    void ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming);
-
     /// Populates this context with data from the requesting process/thread.
-    ResultCode PopulateFromIncomingCommandBuffer(u32_le* src_cmdbuf, Process& src_process,
-                                                 HandleTable& src_table);
+    ResultCode PopulateFromIncomingCommandBuffer(const HandleTable& handle_table,
+                                                 u32_le* src_cmdbuf);
+
     /// Writes data from this context back to the requesting process/thread.
-    ResultCode WriteToOutgoingCommandBuffer(const Thread& thread);
+    ResultCode WriteToOutgoingCommandBuffer(Thread& thread);
 
     u32_le GetCommand() const {
         return command;
@@ -162,8 +161,12 @@ public:
         return buffer_c_desciptors;
     }
 
-    const std::shared_ptr<IPC::DomainMessageHeader>& GetDomainMessageHeader() const {
-        return domain_message_header;
+    const IPC::DomainMessageHeader* GetDomainMessageHeader() const {
+        return domain_message_header.get();
+    }
+
+    bool HasDomainMessageHeader() const {
+        return domain_message_header != nullptr;
     }
 
     /// Helper function to read a buffer using the appropriate buffer descriptor
@@ -255,6 +258,8 @@ public:
     std::string Description() const;
 
 private:
+    void ParseCommandBuffer(const HandleTable& handle_table, u32_le* src_cmdbuf, bool incoming);
+
     std::array<u32, IPC::COMMAND_BUFFER_LENGTH> cmd_buf;
     SharedPtr<Kernel::ServerSession> server_session;
     // TODO(yuriks): Check common usage of this and optimize size accordingly
diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp
index 98eb74298..1fd4ba5d2 100644
--- a/src/core/hle/kernel/kernel.cpp
+++ b/src/core/hle/kernel/kernel.cpp
@@ -32,7 +32,7 @@ namespace Kernel {
  */
 static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_late) {
     const auto proper_handle = static_cast<Handle>(thread_handle);
-    auto& system = Core::System::GetInstance();
+    const auto& system = Core::System::GetInstance();
 
     // Lock the global kernel mutex when we enter the kernel HLE.
     std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
@@ -90,7 +90,7 @@ static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_
 /// The timer callback event, called when a timer is fired
 static void TimerCallback(u64 timer_handle, int cycles_late) {
     const auto proper_handle = static_cast<Handle>(timer_handle);
-    auto& system = Core::System::GetInstance();
+    const auto& system = Core::System::GetInstance();
     SharedPtr<Timer> timer = system.Kernel().RetrieveTimerFromCallbackHandleTable(proper_handle);
 
     if (timer == nullptr) {
@@ -116,9 +116,8 @@ struct KernelCore::Impl {
         next_thread_id = 1;
 
         process_list.clear();
-        current_process.reset();
+        current_process = nullptr;
 
-        handle_table.Clear();
         resource_limits.fill(nullptr);
 
         thread_wakeup_callback_handle_table.Clear();
@@ -207,9 +206,8 @@ struct KernelCore::Impl {
 
     // Lists all processes that exist in the current session.
     std::vector<SharedPtr<Process>> process_list;
-    SharedPtr<Process> current_process;
+    Process* current_process = nullptr;
 
-    Kernel::HandleTable handle_table;
     std::array<SharedPtr<ResourceLimit>, 4> resource_limits;
 
     /// The event type of the generic timer callback event
@@ -241,14 +239,6 @@ void KernelCore::Shutdown() {
     impl->Shutdown();
 }
 
-Kernel::HandleTable& KernelCore::HandleTable() {
-    return impl->handle_table;
-}
-
-const Kernel::HandleTable& KernelCore::HandleTable() const {
-    return impl->handle_table;
-}
-
 SharedPtr<ResourceLimit> KernelCore::ResourceLimitForCategory(
     ResourceLimitCategory category) const {
     return impl->resource_limits.at(static_cast<std::size_t>(category));
@@ -266,15 +256,15 @@ void KernelCore::AppendNewProcess(SharedPtr<Process> process) {
     impl->process_list.push_back(std::move(process));
 }
 
-void KernelCore::MakeCurrentProcess(SharedPtr<Process> process) {
-    impl->current_process = std::move(process);
+void KernelCore::MakeCurrentProcess(Process* process) {
+    impl->current_process = process;
 }
 
-SharedPtr<Process>& KernelCore::CurrentProcess() {
+Process* KernelCore::CurrentProcess() {
     return impl->current_process;
 }
 
-const SharedPtr<Process>& KernelCore::CurrentProcess() const {
+const Process* KernelCore::CurrentProcess() const {
     return impl->current_process;
 }
 
diff --git a/src/core/hle/kernel/kernel.h b/src/core/hle/kernel/kernel.h
index c0771ecf0..7f822d524 100644
--- a/src/core/hle/kernel/kernel.h
+++ b/src/core/hle/kernel/kernel.h
@@ -47,12 +47,6 @@ public:
     /// Clears all resources in use by the kernel instance.
     void Shutdown();
 
-    /// Provides a reference to the handle table.
-    Kernel::HandleTable& HandleTable();
-
-    /// Provides a const reference to the handle table.
-    const Kernel::HandleTable& HandleTable() const;
-
     /// Retrieves a shared pointer to a ResourceLimit identified by the given category.
     SharedPtr<ResourceLimit> ResourceLimitForCategory(ResourceLimitCategory category) const;
 
@@ -66,13 +60,13 @@ public:
     void AppendNewProcess(SharedPtr<Process> process);
 
     /// Makes the given process the new current process.
-    void MakeCurrentProcess(SharedPtr<Process> process);
+    void MakeCurrentProcess(Process* process);
 
-    /// Retrieves a reference to the current process.
-    SharedPtr<Process>& CurrentProcess();
+    /// Retrieves a pointer to the current process.
+    Process* CurrentProcess();
 
-    /// Retrieves a const reference to the current process.
-    const SharedPtr<Process>& CurrentProcess() const;
+    /// Retrieves a const pointer to the current process.
+    const Process* CurrentProcess() const;
 
     /// Adds a port to the named port table
     void AddNamedPort(std::string name, SharedPtr<ClientPort> port);
diff --git a/src/core/hle/kernel/mutex.cpp b/src/core/hle/kernel/mutex.cpp
index dd541ffcc..0743670ad 100644
--- a/src/core/hle/kernel/mutex.cpp
+++ b/src/core/hle/kernel/mutex.cpp
@@ -6,8 +6,6 @@
 #include <utility>
 #include <vector>
 
-#include <boost/range/algorithm_ext/erase.hpp>
-
 #include "common/assert.h"
 #include "core/core.h"
 #include "core/hle/kernel/errors.h"
diff --git a/src/core/hle/kernel/object.cpp b/src/core/hle/kernel/object.cpp
index d51562d92..d87a62bb9 100644
--- a/src/core/hle/kernel/object.cpp
+++ b/src/core/hle/kernel/object.cpp
@@ -25,7 +25,6 @@ bool Object::IsWaitable() const {
     case HandleType::Process:
     case HandleType::AddressArbiter:
     case HandleType::ResourceLimit:
-    case HandleType::CodeSet:
     case HandleType::ClientPort:
     case HandleType::ClientSession:
         return false;
diff --git a/src/core/hle/kernel/object.h b/src/core/hle/kernel/object.h
index 9eb72315c..c9f4d0bb3 100644
--- a/src/core/hle/kernel/object.h
+++ b/src/core/hle/kernel/object.h
@@ -26,7 +26,6 @@ enum class HandleType : u32 {
     AddressArbiter,
     Timer,
     ResourceLimit,
-    CodeSet,
     ClientPort,
     ServerPort,
     ClientSession,
diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp
index fb0027a71..420218d59 100644
--- a/src/core/hle/kernel/process.cpp
+++ b/src/core/hle/kernel/process.cpp
@@ -20,13 +20,7 @@
 
 namespace Kernel {
 
-SharedPtr<CodeSet> CodeSet::Create(KernelCore& kernel, std::string name) {
-    SharedPtr<CodeSet> codeset(new CodeSet(kernel));
-    codeset->name = std::move(name);
-    return codeset;
-}
-
-CodeSet::CodeSet(KernelCore& kernel) : Object{kernel} {}
+CodeSet::CodeSet() = default;
 CodeSet::~CodeSet() = default;
 
 SharedPtr<Process> Process::Create(KernelCore& kernel, std::string&& name) {
@@ -159,11 +153,11 @@ void Process::PrepareForTermination() {
         }
     };
 
-    auto& system = Core::System::GetInstance();
-    stop_threads(system.Scheduler(0)->GetThreadList());
-    stop_threads(system.Scheduler(1)->GetThreadList());
-    stop_threads(system.Scheduler(2)->GetThreadList());
-    stop_threads(system.Scheduler(3)->GetThreadList());
+    const auto& system = Core::System::GetInstance();
+    stop_threads(system.Scheduler(0).GetThreadList());
+    stop_threads(system.Scheduler(1).GetThreadList());
+    stop_threads(system.Scheduler(2).GetThreadList());
+    stop_threads(system.Scheduler(3).GetThreadList());
 }
 
 /**
@@ -224,20 +218,26 @@ void Process::FreeTLSSlot(VAddr tls_address) {
     tls_slots[tls_page].reset(tls_slot);
 }
 
-void Process::LoadModule(SharedPtr<CodeSet> module_, VAddr base_addr) {
+void Process::LoadModule(CodeSet module_, VAddr base_addr) {
     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();
+        const auto vma = vm_manager
+                             .MapMemoryBlock(segment.addr + base_addr, module_.memory,
+                                             segment.offset, segment.size, memory_state)
+                             .Unwrap();
         vm_manager.Reprotect(vma, permissions);
     };
 
     // Map CodeSet segments
-    MapSegment(module_->CodeSegment(), VMAPermission::ReadExecute, MemoryState::CodeStatic);
-    MapSegment(module_->RODataSegment(), VMAPermission::Read, MemoryState::CodeMutable);
-    MapSegment(module_->DataSegment(), VMAPermission::ReadWrite, MemoryState::CodeMutable);
+    MapSegment(module_.CodeSegment(), VMAPermission::ReadExecute, MemoryState::CodeStatic);
+    MapSegment(module_.RODataSegment(), VMAPermission::Read, MemoryState::CodeMutable);
+    MapSegment(module_.DataSegment(), VMAPermission::ReadWrite, MemoryState::CodeMutable);
+
+    // Clear instruction cache in CPU JIT
+    Core::System::GetInstance().ArmInterface(0).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(1).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(2).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(3).ClearInstructionCache();
 }
 
 ResultVal<VAddr> Process::HeapAllocate(VAddr target, u64 size, VMAPermission perms) {
diff --git a/src/core/hle/kernel/process.h b/src/core/hle/kernel/process.h
index 590e0c73d..8d2616c79 100644
--- a/src/core/hle/kernel/process.h
+++ b/src/core/hle/kernel/process.h
@@ -13,6 +13,7 @@
 #include <boost/container/static_vector.hpp>
 #include "common/bit_field.h"
 #include "common/common_types.h"
+#include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/vm_manager.h"
@@ -24,6 +25,7 @@ class ProgramMetadata;
 namespace Kernel {
 
 class KernelCore;
+class ResourceLimit;
 
 struct AddressMapping {
     // Address and size must be page-aligned
@@ -57,30 +59,33 @@ union ProcessFlags {
     BitField<12, 1, u16> loaded_high; ///< Application loaded high (not at 0x00100000).
 };
 
-enum class ProcessStatus { Created, Running, Exited };
-
-class ResourceLimit;
+/**
+ * Indicates the status of a Process instance.
+ *
+ * @note These match the values as used by kernel,
+ *       so new entries should only be added if RE
+ *       shows that a new value has been introduced.
+ */
+enum class ProcessStatus {
+    Created,
+    CreatedWithDebuggerAttached,
+    Running,
+    WaitingForDebuggerToAttach,
+    DebuggerAttached,
+    Exiting,
+    Exited,
+    DebugBreak,
+};
 
-struct CodeSet final : public Object {
+struct CodeSet final {
     struct Segment {
         std::size_t offset = 0;
         VAddr addr = 0;
         u32 size = 0;
     };
 
-    static SharedPtr<CodeSet> Create(KernelCore& kernel, std::string name);
-
-    std::string GetTypeName() const override {
-        return "CodeSet";
-    }
-    std::string GetName() const override {
-        return name;
-    }
-
-    static const HandleType HANDLE_TYPE = HandleType::CodeSet;
-    HandleType GetHandleType() const override {
-        return HANDLE_TYPE;
-    }
+    explicit CodeSet();
+    ~CodeSet();
 
     Segment& CodeSegment() {
         return segments[0];
@@ -109,14 +114,7 @@ struct CodeSet final : public Object {
     std::shared_ptr<std::vector<u8>> memory;
 
     std::array<Segment, 3> segments;
-    VAddr entrypoint;
-
-    /// Name of the process
-    std::string name;
-
-private:
-    explicit CodeSet(KernelCore& kernel);
-    ~CodeSet() override;
+    VAddr entrypoint = 0;
 };
 
 class Process final : public Object {
@@ -145,6 +143,16 @@ public:
         return vm_manager;
     }
 
+    /// Gets a reference to the process' handle table.
+    HandleTable& GetHandleTable() {
+        return handle_table;
+    }
+
+    /// Gets a const reference to the process' handle table.
+    const HandleTable& GetHandleTable() const {
+        return handle_table;
+    }
+
     /// Gets the current status of the process
     ProcessStatus GetStatus() const {
         return status;
@@ -194,6 +202,16 @@ public:
         return is_64bit_process;
     }
 
+    /// Gets the total running time of the process instance in ticks.
+    u64 GetCPUTimeTicks() const {
+        return total_process_running_time_ticks;
+    }
+
+    /// Updates the total running time, adding the given ticks to it.
+    void UpdateCPUTimeTicks(u64 ticks) {
+        total_process_running_time_ticks += ticks;
+    }
+
     /**
      * Loads process-specifics configuration info with metadata provided
      * by an executable.
@@ -219,7 +237,7 @@ public:
      */
     void PrepareForTermination();
 
-    void LoadModule(SharedPtr<CodeSet> module_, VAddr base_addr);
+    void LoadModule(CodeSet module_, VAddr base_addr);
 
     ///////////////////////////////////////////////////////////////////////////////////////////////
     // Memory Management
@@ -297,6 +315,12 @@ private:
     /// specified by metadata provided to the process during loading.
     bool is_64bit_process = true;
 
+    /// Total running time for the process in ticks.
+    u64 total_process_running_time_ticks = 0;
+
+    /// Per-process handle table for storing created object handles in.
+    HandleTable handle_table;
+
     std::string name;
 };
 
diff --git a/src/core/hle/kernel/scheduler.cpp b/src/core/hle/kernel/scheduler.cpp
index cfd6e1bad..5a5f4cef1 100644
--- a/src/core/hle/kernel/scheduler.cpp
+++ b/src/core/hle/kernel/scheduler.cpp
@@ -10,6 +10,7 @@
 #include "core/arm/arm_interface.h"
 #include "core/core.h"
 #include "core/core_timing.h"
+#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/scheduler.h"
 
@@ -34,6 +35,10 @@ Thread* Scheduler::GetCurrentThread() const {
     return current_thread.get();
 }
 
+u64 Scheduler::GetLastContextSwitchTicks() const {
+    return last_context_switch_time;
+}
+
 Thread* Scheduler::PopNextReadyThread() {
     Thread* next = nullptr;
     Thread* thread = GetCurrentThread();
@@ -54,7 +59,10 @@ Thread* Scheduler::PopNextReadyThread() {
 }
 
 void Scheduler::SwitchContext(Thread* new_thread) {
-    Thread* previous_thread = GetCurrentThread();
+    Thread* const previous_thread = GetCurrentThread();
+    Process* const previous_process = Core::CurrentProcess();
+
+    UpdateLastContextSwitchTime(previous_thread, previous_process);
 
     // Save context for previous thread
     if (previous_thread) {
@@ -78,16 +86,14 @@ void Scheduler::SwitchContext(Thread* new_thread) {
         // Cancel any outstanding wakeup events for this thread
         new_thread->CancelWakeupTimer();
 
-        auto previous_process = Core::CurrentProcess();
-
         current_thread = new_thread;
 
         ready_queue.remove(new_thread->GetPriority(), new_thread);
         new_thread->SetStatus(ThreadStatus::Running);
 
-        const auto thread_owner_process = current_thread->GetOwnerProcess();
+        auto* const thread_owner_process = current_thread->GetOwnerProcess();
         if (previous_process != thread_owner_process) {
-            Core::CurrentProcess() = thread_owner_process;
+            Core::System::GetInstance().Kernel().MakeCurrentProcess(thread_owner_process);
             SetCurrentPageTable(&Core::CurrentProcess()->VMManager().page_table);
         }
 
@@ -102,6 +108,22 @@ void Scheduler::SwitchContext(Thread* new_thread) {
     }
 }
 
+void Scheduler::UpdateLastContextSwitchTime(Thread* thread, Process* process) {
+    const u64 prev_switch_ticks = last_context_switch_time;
+    const u64 most_recent_switch_ticks = CoreTiming::GetTicks();
+    const u64 update_ticks = most_recent_switch_ticks - prev_switch_ticks;
+
+    if (thread != nullptr) {
+        thread->UpdateCPUTimeTicks(update_ticks);
+    }
+
+    if (process != nullptr) {
+        process->UpdateCPUTimeTicks(update_ticks);
+    }
+
+    last_context_switch_time = most_recent_switch_ticks;
+}
+
 void Scheduler::Reschedule() {
     std::lock_guard<std::mutex> lock(scheduler_mutex);
 
diff --git a/src/core/hle/kernel/scheduler.h b/src/core/hle/kernel/scheduler.h
index 2c94641ec..c63032b7d 100644
--- a/src/core/hle/kernel/scheduler.h
+++ b/src/core/hle/kernel/scheduler.h
@@ -17,6 +17,8 @@ class ARM_Interface;
 
 namespace Kernel {
 
+class Process;
+
 class Scheduler final {
 public:
     explicit Scheduler(Core::ARM_Interface& cpu_core);
@@ -31,6 +33,9 @@ public:
     /// Gets the current running thread
     Thread* GetCurrentThread() const;
 
+    /// Gets the timestamp for the last context switch in ticks.
+    u64 GetLastContextSwitchTicks() const;
+
     /// Adds a new thread to the scheduler
     void AddThread(SharedPtr<Thread> thread, u32 priority);
 
@@ -64,6 +69,19 @@ private:
      */
     void SwitchContext(Thread* new_thread);
 
+    /**
+     * Called on every context switch to update the internal timestamp
+     * This also updates the running time ticks for the given thread and
+     * process using the following difference:
+     *
+     * ticks += most_recent_ticks - last_context_switch_ticks
+     *
+     * The internal tick timestamp for the scheduler is simply the
+     * most recent tick count retrieved. No special arithmetic is
+     * applied to it.
+     */
+    void UpdateLastContextSwitchTime(Thread* thread, Process* process);
+
     /// Lists all thread ids that aren't deleted/etc.
     std::vector<SharedPtr<Thread>> thread_list;
 
@@ -73,6 +91,7 @@ private:
     SharedPtr<Thread> current_thread = nullptr;
 
     Core::ARM_Interface& cpu_core;
+    u64 last_context_switch_time = 0;
 
     static std::mutex scheduler_mutex;
 };
diff --git a/src/core/hle/kernel/server_port.cpp b/src/core/hle/kernel/server_port.cpp
index 3792e3e18..d6ceeb2da 100644
--- a/src/core/hle/kernel/server_port.cpp
+++ b/src/core/hle/kernel/server_port.cpp
@@ -18,7 +18,7 @@ ServerPort::~ServerPort() = default;
 
 ResultVal<SharedPtr<ServerSession>> ServerPort::Accept() {
     if (pending_sessions.empty()) {
-        return ERR_NO_PENDING_SESSIONS;
+        return ERR_NOT_FOUND;
     }
 
     auto session = std::move(pending_sessions.back());
@@ -28,7 +28,7 @@ ResultVal<SharedPtr<ServerSession>> ServerPort::Accept() {
 
 bool ServerPort::ShouldWait(Thread* thread) const {
     // If there are no pending sessions, we wait until a new one is added.
-    return pending_sessions.size() == 0;
+    return pending_sessions.empty();
 }
 
 void ServerPort::Acquire(Thread* thread) {
diff --git a/src/core/hle/kernel/server_session.cpp b/src/core/hle/kernel/server_session.cpp
index 1ece691c7..80897f3a4 100644
--- a/src/core/hle/kernel/server_session.cpp
+++ b/src/core/hle/kernel/server_session.cpp
@@ -63,7 +63,7 @@ void ServerSession::Acquire(Thread* thread) {
 }
 
 ResultCode ServerSession::HandleDomainSyncRequest(Kernel::HLERequestContext& context) {
-    auto& domain_message_header = context.GetDomainMessageHeader();
+    auto* const domain_message_header = context.GetDomainMessageHeader();
     if (domain_message_header) {
         // Set domain handlers in HLE context, used for domain objects (IPC interfaces) as inputs
         context.SetDomainRequestHandlers(domain_request_handlers);
@@ -107,12 +107,11 @@ ResultCode ServerSession::HandleSyncRequest(SharedPtr<Thread> thread) {
     // similar.
     Kernel::HLERequestContext context(this);
     u32* cmd_buf = (u32*)Memory::GetPointer(thread->GetTLSAddress());
-    context.PopulateFromIncomingCommandBuffer(cmd_buf, *Core::CurrentProcess(),
-                                              kernel.HandleTable());
+    context.PopulateFromIncomingCommandBuffer(kernel.CurrentProcess()->GetHandleTable(), cmd_buf);
 
     ResultCode result = RESULT_SUCCESS;
     // If the session has been converted to a domain, handle the domain request
-    if (IsDomain() && context.GetDomainMessageHeader()) {
+    if (IsDomain() && context.HasDomainMessageHeader()) {
         result = HandleDomainSyncRequest(context);
         // If there is no domain header, the regular session handler is used
     } else if (hle_handler != nullptr) {
diff --git a/src/core/hle/kernel/shared_memory.cpp b/src/core/hle/kernel/shared_memory.cpp
index d061e6155..a016a86b6 100644
--- a/src/core/hle/kernel/shared_memory.cpp
+++ b/src/core/hle/kernel/shared_memory.cpp
@@ -80,20 +80,19 @@ SharedPtr<SharedMemory> SharedMemory::CreateForApplet(
 
 ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermission permissions,
                              MemoryPermission other_permissions) {
-
-    MemoryPermission own_other_permissions =
+    const MemoryPermission own_other_permissions =
         target_process == owner_process ? this->permissions : this->other_permissions;
 
     // Automatically allocated memory blocks can only be mapped with other_permissions = DontCare
     if (base_address == 0 && other_permissions != MemoryPermission::DontCare) {
-        return ERR_INVALID_COMBINATION;
+        return ERR_INVALID_MEMORY_PERMISSIONS;
     }
 
     // Error out if the requested permissions don't match what the creator process allows.
     if (static_cast<u32>(permissions) & ~static_cast<u32>(own_other_permissions)) {
         LOG_ERROR(Kernel, "cannot map id={}, address=0x{:X} name={}, permissions don't match",
                   GetObjectId(), address, name);
-        return ERR_INVALID_COMBINATION;
+        return ERR_INVALID_MEMORY_PERMISSIONS;
     }
 
     // Error out if the provided permissions are not compatible with what the creator process needs.
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index 6c4af7e47..c7c579aaf 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -8,6 +8,7 @@
 #include <mutex>
 #include <vector>
 
+#include "common/alignment.h"
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "common/microprofile.h"
@@ -36,8 +37,72 @@
 
 namespace Kernel {
 namespace {
-constexpr bool Is4KBAligned(VAddr address) {
-    return (address & 0xFFF) == 0;
+
+// Checks if address + size is greater than the given address
+// This can return false if the size causes an overflow of a 64-bit type
+// or if the given size is zero.
+constexpr bool IsValidAddressRange(VAddr address, u64 size) {
+    return address + size > address;
+}
+
+// Checks if a given address range lies within a larger address range.
+constexpr bool IsInsideAddressRange(VAddr address, u64 size, VAddr address_range_begin,
+                                    VAddr address_range_end) {
+    const VAddr end_address = address + size - 1;
+    return address_range_begin <= address && end_address <= address_range_end - 1;
+}
+
+bool IsInsideAddressSpace(const VMManager& vm, VAddr address, u64 size) {
+    return IsInsideAddressRange(address, size, vm.GetAddressSpaceBaseAddress(),
+                                vm.GetAddressSpaceEndAddress());
+}
+
+bool IsInsideNewMapRegion(const VMManager& vm, VAddr address, u64 size) {
+    return IsInsideAddressRange(address, size, vm.GetNewMapRegionBaseAddress(),
+                                vm.GetNewMapRegionEndAddress());
+}
+
+// Helper function that performs the common sanity checks for svcMapMemory
+// and svcUnmapMemory. This is doable, as both functions perform their sanitizing
+// in the same order.
+ResultCode MapUnmapMemorySanityChecks(const VMManager& vm_manager, VAddr dst_addr, VAddr src_addr,
+                                      u64 size) {
+    if (!Common::Is4KBAligned(dst_addr) || !Common::Is4KBAligned(src_addr)) {
+        return ERR_INVALID_ADDRESS;
+    }
+
+    if (size == 0 || !Common::Is4KBAligned(size)) {
+        return ERR_INVALID_SIZE;
+    }
+
+    if (!IsValidAddressRange(dst_addr, size)) {
+        return ERR_INVALID_ADDRESS_STATE;
+    }
+
+    if (!IsValidAddressRange(src_addr, size)) {
+        return ERR_INVALID_ADDRESS_STATE;
+    }
+
+    if (!IsInsideAddressSpace(vm_manager, src_addr, size)) {
+        return ERR_INVALID_ADDRESS_STATE;
+    }
+
+    if (!IsInsideNewMapRegion(vm_manager, dst_addr, size)) {
+        return ERR_INVALID_MEMORY_RANGE;
+    }
+
+    const VAddr dst_end_address = dst_addr + size;
+    if (dst_end_address > vm_manager.GetHeapRegionBaseAddress() &&
+        vm_manager.GetHeapRegionEndAddress() > dst_addr) {
+        return ERR_INVALID_MEMORY_RANGE;
+    }
+
+    if (dst_end_address > vm_manager.GetMapRegionBaseAddress() &&
+        vm_manager.GetMapRegionEndAddress() > dst_addr) {
+        return ERR_INVALID_MEMORY_RANGE;
+    }
+
+    return RESULT_SUCCESS;
 }
 } // Anonymous namespace
 
@@ -69,15 +134,15 @@ static ResultCode MapMemory(VAddr dst_addr, VAddr src_addr, u64 size) {
     LOG_TRACE(Kernel_SVC, "called, dst_addr=0x{:X}, src_addr=0x{:X}, size=0x{:X}", dst_addr,
               src_addr, size);
 
-    if (!Is4KBAligned(dst_addr) || !Is4KBAligned(src_addr)) {
-        return ERR_INVALID_ADDRESS;
-    }
+    auto* const current_process = Core::CurrentProcess();
+    const auto& vm_manager = current_process->VMManager();
 
-    if (size == 0 || !Is4KBAligned(size)) {
-        return ERR_INVALID_SIZE;
+    const auto result = MapUnmapMemorySanityChecks(vm_manager, dst_addr, src_addr, size);
+    if (result != RESULT_SUCCESS) {
+        return result;
     }
 
-    return Core::CurrentProcess()->MirrorMemory(dst_addr, src_addr, size);
+    return current_process->MirrorMemory(dst_addr, src_addr, size);
 }
 
 /// Unmaps a region that was previously mapped with svcMapMemory
@@ -85,15 +150,15 @@ static ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, u64 size) {
     LOG_TRACE(Kernel_SVC, "called, dst_addr=0x{:X}, src_addr=0x{:X}, size=0x{:X}", dst_addr,
               src_addr, size);
 
-    if (!Is4KBAligned(dst_addr) || !Is4KBAligned(src_addr)) {
-        return ERR_INVALID_ADDRESS;
-    }
+    auto* const current_process = Core::CurrentProcess();
+    const auto& vm_manager = current_process->VMManager();
 
-    if (size == 0 || !Is4KBAligned(size)) {
-        return ERR_INVALID_SIZE;
+    const auto result = MapUnmapMemorySanityChecks(vm_manager, dst_addr, src_addr, size);
+    if (result != RESULT_SUCCESS) {
+        return result;
     }
 
-    return Core::CurrentProcess()->UnmapMemory(dst_addr, src_addr, size);
+    return current_process->UnmapMemory(dst_addr, src_addr, size);
 }
 
 /// Connect to an OS service given the port name, returns the handle to the port to out
@@ -124,14 +189,15 @@ static ResultCode ConnectToNamedPort(Handle* out_handle, VAddr port_name_address
     CASCADE_RESULT(client_session, client_port->Connect());
 
     // Return the client session
-    CASCADE_RESULT(*out_handle, kernel.HandleTable().Create(client_session));
+    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    CASCADE_RESULT(*out_handle, handle_table.Create(client_session));
     return RESULT_SUCCESS;
 }
 
 /// Makes a blocking IPC call to an OS service.
 static ResultCode SendSyncRequest(Handle handle) {
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<ClientSession> session = kernel.HandleTable().Get<ClientSession>(handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    SharedPtr<ClientSession> session = handle_table.Get<ClientSession>(handle);
     if (!session) {
         LOG_ERROR(Kernel_SVC, "called with invalid handle=0x{:08X}", handle);
         return ERR_INVALID_HANDLE;
@@ -150,8 +216,8 @@ static ResultCode SendSyncRequest(Handle handle) {
 static ResultCode GetThreadId(u32* thread_id, Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -164,8 +230,8 @@ static ResultCode GetThreadId(u32* thread_id, Handle thread_handle) {
 static ResultCode GetProcessId(u32* process_id, Handle process_handle) {
     LOG_TRACE(Kernel_SVC, "called process=0x{:08X}", process_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Process> process = kernel.HandleTable().Get<Process>(process_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Process> process = handle_table.Get<Process>(process_handle);
     if (!process) {
         return ERR_INVALID_HANDLE;
     }
@@ -208,11 +274,11 @@ static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64
 
     using ObjectPtr = Thread::ThreadWaitObjects::value_type;
     Thread::ThreadWaitObjects objects(handle_count);
-    auto& kernel = Core::System::GetInstance().Kernel();
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
 
     for (u64 i = 0; i < handle_count; ++i) {
         const Handle handle = Memory::Read32(handles_address + i * sizeof(Handle));
-        const auto object = kernel.HandleTable().Get<WaitObject>(handle);
+        const auto object = handle_table.Get<WaitObject>(handle);
 
         if (object == nullptr) {
             return ERR_INVALID_HANDLE;
@@ -260,8 +326,8 @@ static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64
 static ResultCode CancelSynchronization(Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:X}", thread_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -285,7 +351,11 @@ static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr,
         return ERR_INVALID_ADDRESS_STATE;
     }
 
-    auto& handle_table = Core::System::GetInstance().Kernel().HandleTable();
+    if (!Common::IsWordAligned(mutex_addr)) {
+        return ERR_INVALID_ADDRESS;
+    }
+
+    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
     return Mutex::TryAcquire(handle_table, mutex_addr, holding_thread_handle,
                              requesting_thread_handle);
 }
@@ -298,16 +368,87 @@ static ResultCode ArbitrateUnlock(VAddr mutex_addr) {
         return ERR_INVALID_ADDRESS_STATE;
     }
 
+    if (!Common::IsWordAligned(mutex_addr)) {
+        return ERR_INVALID_ADDRESS;
+    }
+
     return Mutex::Release(mutex_addr);
 }
 
+enum class BreakType : u32 {
+    Panic = 0,
+    AssertionFailed = 1,
+    PreNROLoad = 3,
+    PostNROLoad = 4,
+    PreNROUnload = 5,
+    PostNROUnload = 6,
+};
+
+struct BreakReason {
+    union {
+        u32 raw;
+        BitField<0, 30, BreakType> break_type;
+        BitField<31, 1, u32> signal_debugger;
+    };
+};
+
 /// Break program execution
-static void Break(u64 reason, u64 info1, u64 info2) {
-    LOG_CRITICAL(
-        Debug_Emulated,
-        "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}",
-        reason, info1, info2);
-    ASSERT(false);
+static void Break(u32 reason, u64 info1, u64 info2) {
+    BreakReason break_reason{reason};
+
+    switch (break_reason.break_type) {
+    case BreakType::Panic:
+        LOG_CRITICAL(Debug_Emulated, "Signalling debugger, PANIC! info1=0x{:016X}, info2=0x{:016X}",
+                     info1, info2);
+        break;
+    case BreakType::AssertionFailed:
+        LOG_CRITICAL(Debug_Emulated,
+                     "Signalling debugger, Assertion failed! info1=0x{:016X}, info2=0x{:016X}",
+                     info1, info2);
+        break;
+    case BreakType::PreNROLoad:
+        LOG_WARNING(
+            Debug_Emulated,
+            "Signalling debugger, Attempting to load an NRO at 0x{:016X} with size 0x{:016X}",
+            info1, info2);
+        break;
+    case BreakType::PostNROLoad:
+        LOG_WARNING(Debug_Emulated,
+                    "Signalling debugger, Loaded an NRO at 0x{:016X} with size 0x{:016X}", info1,
+                    info2);
+        break;
+    case BreakType::PreNROUnload:
+        LOG_WARNING(
+            Debug_Emulated,
+            "Signalling debugger, Attempting to unload an NRO at 0x{:016X} with size 0x{:016X}",
+            info1, info2);
+        break;
+    case BreakType::PostNROUnload:
+        LOG_WARNING(Debug_Emulated,
+                    "Signalling debugger, Unloaded an NRO at 0x{:016X} with size 0x{:016X}", info1,
+                    info2);
+        break;
+    default:
+        LOG_WARNING(
+            Debug_Emulated,
+            "Signalling debugger, Unknown break reason {}, info1=0x{:016X}, info2=0x{:016X}",
+            static_cast<u32>(break_reason.break_type.Value()), info1, info2);
+        break;
+    }
+
+    if (!break_reason.signal_debugger) {
+        LOG_CRITICAL(
+            Debug_Emulated,
+            "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}",
+            reason, info1, info2);
+        ASSERT(false);
+
+        Core::CurrentProcess()->PrepareForTermination();
+
+        // Kill the current thread
+        GetCurrentThread()->Stop();
+        Core::System::GetInstance().PrepareReschedule();
+    }
 }
 
 /// Used to output a message on a debug hardware unit - does nothing on a retail unit
@@ -326,7 +467,38 @@ 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);
 
-    const auto& current_process = Core::CurrentProcess();
+    enum class GetInfoType : u64 {
+        // 1.0.0+
+        AllowedCpuIdBitmask = 0,
+        AllowedThreadPrioBitmask = 1,
+        MapRegionBaseAddr = 2,
+        MapRegionSize = 3,
+        HeapRegionBaseAddr = 4,
+        HeapRegionSize = 5,
+        TotalMemoryUsage = 6,
+        TotalHeapUsage = 7,
+        IsCurrentProcessBeingDebugged = 8,
+        ResourceHandleLimit = 9,
+        IdleTickCount = 10,
+        RandomEntropy = 11,
+        PerformanceCounter = 0xF0000002,
+        // 2.0.0+
+        ASLRRegionBaseAddr = 12,
+        ASLRRegionSize = 13,
+        NewMapRegionBaseAddr = 14,
+        NewMapRegionSize = 15,
+        // 3.0.0+
+        IsVirtualAddressMemoryEnabled = 16,
+        PersonalMmHeapUsage = 17,
+        TitleId = 18,
+        // 4.0.0+
+        PrivilegedProcessId = 19,
+        // 5.0.0+
+        UserExceptionContextAddr = 20,
+        ThreadTickCount = 0xF0000002,
+    };
+
+    const auto* current_process = Core::CurrentProcess();
     const auto& vm_manager = current_process->VMManager();
 
     switch (static_cast<GetInfoType>(info_id)) {
@@ -360,25 +532,12 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
     case GetInfoType::RandomEntropy:
         *result = 0;
         break;
-    case GetInfoType::AddressSpaceBaseAddr:
-        *result = vm_manager.GetCodeRegionBaseAddress();
+    case GetInfoType::ASLRRegionBaseAddr:
+        *result = vm_manager.GetASLRRegionBaseAddress();
         break;
-    case GetInfoType::AddressSpaceSize: {
-        const u64 width = vm_manager.GetAddressSpaceWidth();
-
-        switch (width) {
-        case 32:
-            *result = 0xFFE00000;
-            break;
-        case 36:
-            *result = 0xFF8000000;
-            break;
-        case 39:
-            *result = 0x7FF8000000;
-            break;
-        }
+    case GetInfoType::ASLRRegionSize:
+        *result = vm_manager.GetASLRRegionSize();
         break;
-    }
     case GetInfoType::NewMapRegionBaseAddr:
         *result = vm_manager.GetNewMapRegionBaseAddress();
         break;
@@ -401,6 +560,36 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
                     "(STUBBED) Attempted to query user exception context address, returned 0");
         *result = 0;
         break;
+    case GetInfoType::ThreadTickCount: {
+        constexpr u64 num_cpus = 4;
+        if (info_sub_id != 0xFFFFFFFFFFFFFFFF && info_sub_id >= num_cpus) {
+            return ERR_INVALID_COMBINATION_KERNEL;
+        }
+
+        const auto thread =
+            current_process->GetHandleTable().Get<Thread>(static_cast<Handle>(handle));
+        if (!thread) {
+            return ERR_INVALID_HANDLE;
+        }
+
+        const auto& system = Core::System::GetInstance();
+        const auto& scheduler = system.CurrentScheduler();
+        const auto* const current_thread = scheduler.GetCurrentThread();
+        const bool same_thread = current_thread == thread;
+
+        const u64 prev_ctx_ticks = scheduler.GetLastContextSwitchTicks();
+        u64 out_ticks = 0;
+        if (same_thread && info_sub_id == 0xFFFFFFFFFFFFFFFF) {
+            const u64 thread_ticks = current_thread->GetTotalCPUTimeTicks();
+
+            out_ticks = thread_ticks + (CoreTiming::GetTicks() - prev_ctx_ticks);
+        } else if (same_thread && info_sub_id == system.CurrentCoreIndex()) {
+            out_ticks = CoreTiming::GetTicks() - prev_ctx_ticks;
+        }
+
+        *result = out_ticks;
+        break;
+    }
     default:
         UNIMPLEMENTED();
     }
@@ -418,13 +607,12 @@ static ResultCode SetThreadActivity(Handle handle, u32 unknown) {
 static ResultCode GetThreadContext(VAddr thread_context, Handle handle) {
     LOG_DEBUG(Kernel_SVC, "called, context=0x{:08X}, thread=0x{:X}", thread_context, handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(handle);
+    const auto* current_process = Core::CurrentProcess();
+    const SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
 
-    const auto current_process = Core::CurrentProcess();
     if (thread->GetOwnerProcess() != current_process) {
         return ERR_INVALID_HANDLE;
     }
@@ -450,10 +638,11 @@ static ResultCode GetThreadContext(VAddr thread_context, Handle handle) {
 
 /// Gets the priority for the specified thread
 static ResultCode GetThreadPriority(u32* priority, Handle handle) {
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(handle);
-    if (!thread)
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(handle);
+    if (!thread) {
         return ERR_INVALID_HANDLE;
+    }
 
     *priority = thread->GetPriority();
     return RESULT_SUCCESS;
@@ -465,16 +654,18 @@ static ResultCode SetThreadPriority(Handle handle, u32 priority) {
         return ERR_INVALID_THREAD_PRIORITY;
     }
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(handle);
-    if (!thread)
-        return ERR_INVALID_HANDLE;
+    const auto* const current_process = Core::CurrentProcess();
 
     // Note: The kernel uses the current process's resource limit instead of
     // the one from the thread owner's resource limit.
-    const ResourceLimit& resource_limit = Core::CurrentProcess()->GetResourceLimit();
+    const ResourceLimit& resource_limit = current_process->GetResourceLimit();
     if (resource_limit.GetMaxResourceValue(ResourceType::Priority) > priority) {
-        return ERR_NOT_AUTHORIZED;
+        return ERR_INVALID_THREAD_PRIORITY;
+    }
+
+    SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
+    if (!thread) {
+        return ERR_INVALID_HANDLE;
     }
 
     thread->SetPriority(priority);
@@ -495,14 +686,18 @@ static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 s
               "called, shared_memory_handle=0x{:X}, addr=0x{:X}, size=0x{:X}, permissions=0x{:08X}",
               shared_memory_handle, addr, size, permissions);
 
-    if (!Is4KBAligned(addr)) {
+    if (!Common::Is4KBAligned(addr)) {
         return ERR_INVALID_ADDRESS;
     }
 
-    if (size == 0 || !Is4KBAligned(size)) {
+    if (size == 0 || !Common::Is4KBAligned(size)) {
         return ERR_INVALID_SIZE;
     }
 
+    if (!IsValidAddressRange(addr, size)) {
+        return ERR_INVALID_ADDRESS_STATE;
+    }
+
     const auto permissions_type = static_cast<MemoryPermission>(permissions);
     if (permissions_type != MemoryPermission::Read &&
         permissions_type != MemoryPermission::ReadWrite) {
@@ -510,46 +705,61 @@ static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 s
         return ERR_INVALID_MEMORY_PERMISSIONS;
     }
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    auto shared_memory = kernel.HandleTable().Get<SharedMemory>(shared_memory_handle);
+    auto* const current_process = Core::CurrentProcess();
+    auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle);
     if (!shared_memory) {
         return ERR_INVALID_HANDLE;
     }
 
-    return shared_memory->Map(Core::CurrentProcess().get(), addr, permissions_type,
-                              MemoryPermission::DontCare);
+    const auto& vm_manager = current_process->VMManager();
+    if (!vm_manager.IsWithinASLRRegion(addr, size)) {
+        return ERR_INVALID_MEMORY_RANGE;
+    }
+
+    return shared_memory->Map(current_process, addr, permissions_type, MemoryPermission::DontCare);
 }
 
 static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 size) {
     LOG_WARNING(Kernel_SVC, "called, shared_memory_handle=0x{:08X}, addr=0x{:X}, size=0x{:X}",
                 shared_memory_handle, addr, size);
 
-    if (!Is4KBAligned(addr)) {
+    if (!Common::Is4KBAligned(addr)) {
         return ERR_INVALID_ADDRESS;
     }
 
-    if (size == 0 || !Is4KBAligned(size)) {
+    if (size == 0 || !Common::Is4KBAligned(size)) {
         return ERR_INVALID_SIZE;
     }
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    auto shared_memory = kernel.HandleTable().Get<SharedMemory>(shared_memory_handle);
+    if (!IsValidAddressRange(addr, size)) {
+        return ERR_INVALID_ADDRESS_STATE;
+    }
 
-    return shared_memory->Unmap(Core::CurrentProcess().get(), addr);
+    auto* const current_process = Core::CurrentProcess();
+    auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle);
+    if (!shared_memory) {
+        return ERR_INVALID_HANDLE;
+    }
+
+    const auto& vm_manager = current_process->VMManager();
+    if (!vm_manager.IsWithinASLRRegion(addr, size)) {
+        return ERR_INVALID_MEMORY_RANGE;
+    }
+
+    return shared_memory->Unmap(current_process, addr);
 }
 
 /// Query process memory
 static ResultCode QueryProcessMemory(MemoryInfo* memory_info, PageInfo* /*page_info*/,
                                      Handle process_handle, u64 addr) {
-
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<Process> process = kernel.HandleTable().Get<Process>(process_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    SharedPtr<Process> process = handle_table.Get<Process>(process_handle);
     if (!process) {
         return ERR_INVALID_HANDLE;
     }
     auto vma = process->VMManager().FindVMA(addr);
     memory_info->attributes = 0;
-    if (vma == Core::CurrentProcess()->VMManager().vma_map.end()) {
+    if (vma == process->VMManager().vma_map.end()) {
         memory_info->base_address = 0;
         memory_info->permission = static_cast<u32>(VMAPermission::None);
         memory_info->size = 0;
@@ -573,7 +783,7 @@ static ResultCode QueryMemory(MemoryInfo* memory_info, PageInfo* page_info, VAdd
 
 /// Exits the current process
 static void ExitProcess() {
-    auto& current_process = Core::CurrentProcess();
+    auto* current_process = Core::CurrentProcess();
 
     LOG_INFO(Kernel_SVC, "Process {} exiting", current_process->GetProcessID());
     ASSERT_MSG(current_process->GetStatus() == ProcessStatus::Running,
@@ -590,20 +800,19 @@ static void ExitProcess() {
 /// Creates a new thread
 static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, VAddr stack_top,
                                u32 priority, s32 processor_id) {
-    std::string name = fmt::format("thread-{:X}", entry_point);
-
     if (priority > THREADPRIO_LOWEST) {
         return ERR_INVALID_THREAD_PRIORITY;
     }
 
-    const ResourceLimit& resource_limit = Core::CurrentProcess()->GetResourceLimit();
+    auto* const current_process = Core::CurrentProcess();
+    const ResourceLimit& resource_limit = current_process->GetResourceLimit();
     if (resource_limit.GetMaxResourceValue(ResourceType::Priority) > priority) {
-        return ERR_NOT_AUTHORIZED;
+        return ERR_INVALID_THREAD_PRIORITY;
     }
 
     if (processor_id == THREADPROCESSORID_DEFAULT) {
         // Set the target CPU to the one specified in the process' exheader.
-        processor_id = Core::CurrentProcess()->GetDefaultProcessorID();
+        processor_id = current_process->GetDefaultProcessorID();
         ASSERT(processor_id != THREADPROCESSORID_DEFAULT);
     }
 
@@ -618,11 +827,13 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
         return ERR_INVALID_PROCESSOR_ID;
     }
 
+    const std::string name = fmt::format("thread-{:X}", entry_point);
     auto& kernel = Core::System::GetInstance().Kernel();
     CASCADE_RESULT(SharedPtr<Thread> thread,
                    Thread::Create(kernel, name, entry_point, priority, arg, processor_id, stack_top,
-                                  Core::CurrentProcess()));
-    const auto new_guest_handle = kernel.HandleTable().Create(thread);
+                                  *current_process));
+
+    const auto new_guest_handle = current_process->GetHandleTable().Create(thread);
     if (new_guest_handle.Failed()) {
         return new_guest_handle.Code();
     }
@@ -643,8 +854,8 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
 static ResultCode StartThread(Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -691,8 +902,8 @@ static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_var
         "called mutex_addr={:X}, condition_variable_addr={:X}, thread_handle=0x{:08X}, timeout={}",
         mutex_addr, condition_variable_addr, thread_handle, nano_seconds);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     ASSERT(thread);
 
     CASCADE_CODE(Mutex::Release(mutex_addr));
@@ -721,7 +932,7 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target
                                            std::vector<SharedPtr<Thread>>& waiting_threads,
                                            VAddr condvar_addr) {
         const auto& scheduler = Core::System::GetInstance().Scheduler(core_index);
-        const auto& thread_list = scheduler->GetThreadList();
+        const auto& thread_list = scheduler.GetThreadList();
 
         for (const auto& thread : thread_list) {
             if (thread->GetCondVarWaitAddress() == condvar_addr)
@@ -803,9 +1014,9 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target
                                                mutex_val | Mutex::MutexHasWaitersFlag));
 
             // The mutex is already owned by some other thread, make this thread wait on it.
-            auto& kernel = Core::System::GetInstance().Kernel();
-            Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask);
-            auto owner = kernel.HandleTable().Get<Thread>(owner_handle);
+            const Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask);
+            const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+            auto owner = handle_table.Get<Thread>(owner_handle);
             ASSERT(owner);
             ASSERT(thread->GetStatus() == ThreadStatus::WaitMutex);
             thread->InvalidateWakeupCallback();
@@ -884,16 +1095,16 @@ static u64 GetSystemTick() {
 static ResultCode CloseHandle(Handle handle) {
     LOG_TRACE(Kernel_SVC, "Closing handle 0x{:08X}", handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    return kernel.HandleTable().Close(handle);
+    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    return handle_table.Close(handle);
 }
 
 /// Reset an event
 static ResultCode ResetSignal(Handle handle) {
     LOG_WARNING(Kernel_SVC, "(STUBBED) called handle 0x{:08X}", handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    auto event = kernel.HandleTable().Get<Event>(handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    auto event = handle_table.Get<Event>(handle);
 
     ASSERT(event != nullptr);
 
@@ -912,8 +1123,8 @@ static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32
 static ResultCode GetThreadCoreMask(Handle thread_handle, u32* core, u64* mask) {
     LOG_TRACE(Kernel_SVC, "called, handle=0x{:08X}", thread_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -928,8 +1139,8 @@ static ResultCode SetThreadCoreMask(Handle thread_handle, u32 core, u64 mask) {
     LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, mask=0x{:16X}, core=0x{:X}", thread_handle,
               mask, core);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -990,7 +1201,7 @@ static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permiss
     }
 
     auto& kernel = Core::System::GetInstance().Kernel();
-    auto& handle_table = kernel.HandleTable();
+    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
     auto shared_mem_handle =
         SharedMemory::Create(kernel, handle_table.Get<Process>(KernelHandle::CurrentProcess), size,
                              local_perms, remote_perms);
@@ -1002,14 +1213,39 @@ static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permiss
 static ResultCode ClearEvent(Handle handle) {
     LOG_TRACE(Kernel_SVC, "called, event=0x{:08X}", handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<Event> evt = kernel.HandleTable().Get<Event>(handle);
-    if (evt == nullptr)
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    SharedPtr<Event> evt = handle_table.Get<Event>(handle);
+    if (evt == nullptr) {
         return ERR_INVALID_HANDLE;
+    }
+
     evt->Clear();
     return RESULT_SUCCESS;
 }
 
+static ResultCode GetProcessInfo(u64* out, Handle process_handle, u32 type) {
+    LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, type=0x{:X}", process_handle, type);
+
+    // This function currently only allows retrieving a process' status.
+    enum class InfoType {
+        Status,
+    };
+
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const auto process = handle_table.Get<Process>(process_handle);
+    if (!process) {
+        return ERR_INVALID_HANDLE;
+    }
+
+    const auto info_type = static_cast<InfoType>(type);
+    if (info_type != InfoType::Status) {
+        return ERR_INVALID_ENUM_VALUE;
+    }
+
+    *out = static_cast<u64>(process->GetStatus());
+    return RESULT_SUCCESS;
+}
+
 namespace {
 struct FunctionDef {
     using Func = void();
@@ -1145,7 +1381,7 @@ static const FunctionDef SVC_Table[] = {
     {0x79, nullptr, "CreateProcess"},
     {0x7A, nullptr, "StartProcess"},
     {0x7B, nullptr, "TerminateProcess"},
-    {0x7C, nullptr, "GetProcessInfo"},
+    {0x7C, SvcWrap<GetProcessInfo>, "GetProcessInfo"},
     {0x7D, nullptr, "CreateResourceLimit"},
     {0x7E, nullptr, "SetResourceLimitLimitValue"},
     {0x7F, nullptr, "CallSecureMonitor"},
diff --git a/src/core/hle/kernel/svc.h b/src/core/hle/kernel/svc.h
index 70148c4fe..b06aac4ec 100644
--- a/src/core/hle/kernel/svc.h
+++ b/src/core/hle/kernel/svc.h
@@ -24,37 +24,6 @@ struct PageInfo {
     u64 flags;
 };
 
-/// Values accepted by svcGetInfo
-enum class GetInfoType : u64 {
-    // 1.0.0+
-    AllowedCpuIdBitmask = 0,
-    AllowedThreadPrioBitmask = 1,
-    MapRegionBaseAddr = 2,
-    MapRegionSize = 3,
-    HeapRegionBaseAddr = 4,
-    HeapRegionSize = 5,
-    TotalMemoryUsage = 6,
-    TotalHeapUsage = 7,
-    IsCurrentProcessBeingDebugged = 8,
-    ResourceHandleLimit = 9,
-    IdleTickCount = 10,
-    RandomEntropy = 11,
-    PerformanceCounter = 0xF0000002,
-    // 2.0.0+
-    AddressSpaceBaseAddr = 12,
-    AddressSpaceSize = 13,
-    NewMapRegionBaseAddr = 14,
-    NewMapRegionSize = 15,
-    // 3.0.0+
-    IsVirtualAddressMemoryEnabled = 16,
-    PersonalMmHeapUsage = 17,
-    TitleId = 18,
-    // 4.0.0+
-    PrivilegedProcessId = 19,
-    // 5.0.0+
-    UserExceptionContextAddr = 20,
-};
-
 void CallSVC(u32 immediate);
 
 } // namespace Kernel
diff --git a/src/core/hle/kernel/svc_wrap.h b/src/core/hle/kernel/svc_wrap.h
index 22712e64f..b09753c80 100644
--- a/src/core/hle/kernel/svc_wrap.h
+++ b/src/core/hle/kernel/svc_wrap.h
@@ -35,18 +35,18 @@ void SvcWrap() {
 
 template <ResultCode func(u32)>
 void SvcWrap() {
-    FuncReturn(func((u32)Param(0)).raw);
+    FuncReturn(func(static_cast<u32>(Param(0))).raw);
 }
 
 template <ResultCode func(u32, u32)>
 void SvcWrap() {
-    FuncReturn(func((u32)Param(0), (u32)Param(1)).raw);
+    FuncReturn(func(static_cast<u32>(Param(0)), static_cast<u32>(Param(1))).raw);
 }
 
 template <ResultCode func(u32*, u32)>
 void SvcWrap() {
     u32 param_1 = 0;
-    u32 retval = func(&param_1, (u32)Param(1)).raw;
+    u32 retval = func(&param_1, static_cast<u32>(Param(1))).raw;
     Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
@@ -61,7 +61,7 @@ void SvcWrap() {
 
 template <ResultCode func(u64, s32)>
 void SvcWrap() {
-    FuncReturn(func(Param(0), (s32)Param(1)).raw);
+    FuncReturn(func(Param(0), static_cast<s32>(Param(1))).raw);
 }
 
 template <ResultCode func(u64, u32)>
@@ -77,21 +77,29 @@ void SvcWrap() {
     FuncReturn(retval);
 }
 
+template <ResultCode func(u64*, u32, u32)>
+void SvcWrap() {
+    u64 param_1 = 0;
+    u32 retval = func(&param_1, static_cast<u32>(Param(1)), static_cast<u32>(Param(2))).raw;
+    Core::CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(retval);
+}
+
 template <ResultCode func(u32, u64)>
 void SvcWrap() {
-    FuncReturn(func((u32)(Param(0) & 0xFFFFFFFF), Param(1)).raw);
+    FuncReturn(func(static_cast<u32>(Param(0)), Param(1)).raw);
 }
 
 template <ResultCode func(u32, u32, u64)>
 void SvcWrap() {
-    FuncReturn(func((u32)(Param(0) & 0xFFFFFFFF), (u32)(Param(1) & 0xFFFFFFFF), Param(2)).raw);
+    FuncReturn(func(static_cast<u32>(Param(0)), static_cast<u32>(Param(1)), Param(2)).raw);
 }
 
 template <ResultCode func(u32, u32*, u64*)>
 void SvcWrap() {
     u32 param_1 = 0;
     u64 param_2 = 0;
-    ResultCode retval = func((u32)(Param(2) & 0xFFFFFFFF), &param_1, &param_2);
+    ResultCode retval = func(static_cast<u32>(Param(2)), &param_1, &param_2);
     Core::CurrentArmInterface().SetReg(1, param_1);
     Core::CurrentArmInterface().SetReg(2, param_2);
     FuncReturn(retval.raw);
@@ -100,12 +108,12 @@ void SvcWrap() {
 template <ResultCode func(u64, u64, u32, u32)>
 void SvcWrap() {
     FuncReturn(
-        func(Param(0), Param(1), (u32)(Param(3) & 0xFFFFFFFF), (u32)(Param(3) & 0xFFFFFFFF)).raw);
+        func(Param(0), Param(1), static_cast<u32>(Param(3)), static_cast<u32>(Param(3))).raw);
 }
 
 template <ResultCode func(u32, u64, u32)>
 void SvcWrap() {
-    FuncReturn(func((u32)Param(0), Param(1), (u32)Param(2)).raw);
+    FuncReturn(func(static_cast<u32>(Param(0)), Param(1), static_cast<u32>(Param(2))).raw);
 }
 
 template <ResultCode func(u64, u64, u64)>
@@ -115,25 +123,28 @@ void SvcWrap() {
 
 template <ResultCode func(u32, u64, u64, u32)>
 void SvcWrap() {
-    FuncReturn(func((u32)Param(0), Param(1), Param(2), (u32)Param(3)).raw);
+    FuncReturn(
+        func(static_cast<u32>(Param(0)), Param(1), Param(2), static_cast<u32>(Param(3))).raw);
 }
 
 template <ResultCode func(u32, u64, u64)>
 void SvcWrap() {
-    FuncReturn(func((u32)Param(0), Param(1), Param(2)).raw);
+    FuncReturn(func(static_cast<u32>(Param(0)), Param(1), Param(2)).raw);
 }
 
 template <ResultCode func(u32*, u64, u64, s64)>
 void SvcWrap() {
     u32 param_1 = 0;
-    ResultCode retval = func(&param_1, Param(1), (u32)(Param(2) & 0xFFFFFFFF), (s64)Param(3));
+    ResultCode retval =
+        func(&param_1, Param(1), static_cast<u32>(Param(2)), static_cast<s64>(Param(3)));
     Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval.raw);
 }
 
 template <ResultCode func(u64, u64, u32, s64)>
 void SvcWrap() {
-    FuncReturn(func(Param(0), Param(1), (u32)Param(2), (s64)Param(3)).raw);
+    FuncReturn(
+        func(Param(0), Param(1), static_cast<u32>(Param(2)), static_cast<s64>(Param(3))).raw);
 }
 
 template <ResultCode func(u64*, u64, u64, u64)>
@@ -147,9 +158,9 @@ void SvcWrap() {
 template <ResultCode func(u32*, u64, u64, u64, u32, s32)>
 void SvcWrap() {
     u32 param_1 = 0;
-    u32 retval =
-        func(&param_1, Param(1), Param(2), Param(3), (u32)Param(4), (s32)(Param(5) & 0xFFFFFFFF))
-            .raw;
+    u32 retval = func(&param_1, Param(1), Param(2), Param(3), static_cast<u32>(Param(4)),
+                      static_cast<s32>(Param(5)))
+                     .raw;
     Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
@@ -172,7 +183,7 @@ void SvcWrap() {
 template <ResultCode func(u32*, u64, u64, u32)>
 void SvcWrap() {
     u32 param_1 = 0;
-    u32 retval = func(&param_1, Param(1), Param(2), (u32)(Param(3) & 0xFFFFFFFF)).raw;
+    u32 retval = func(&param_1, Param(1), Param(2), static_cast<u32>(Param(3))).raw;
     Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
@@ -181,22 +192,22 @@ template <ResultCode func(Handle*, u64, u32, u32)>
 void SvcWrap() {
     u32 param_1 = 0;
     u32 retval =
-        func(&param_1, Param(1), (u32)(Param(2) & 0xFFFFFFFF), (u32)(Param(3) & 0xFFFFFFFF)).raw;
+        func(&param_1, Param(1), static_cast<u32>(Param(2)), static_cast<u32>(Param(3))).raw;
     Core::CurrentArmInterface().SetReg(1, param_1);
     FuncReturn(retval);
 }
 
 template <ResultCode func(u64, u32, s32, s64)>
 void SvcWrap() {
-    FuncReturn(
-        func(Param(0), (u32)(Param(1) & 0xFFFFFFFF), (s32)(Param(2) & 0xFFFFFFFF), (s64)Param(3))
-            .raw);
+    FuncReturn(func(Param(0), static_cast<u32>(Param(1)), static_cast<s32>(Param(2)),
+                    static_cast<s64>(Param(3)))
+                   .raw);
 }
 
 template <ResultCode func(u64, u32, s32, s32)>
 void SvcWrap() {
-    FuncReturn(func(Param(0), (u32)(Param(1) & 0xFFFFFFFF), (s32)(Param(2) & 0xFFFFFFFF),
-                    (s32)(Param(3) & 0xFFFFFFFF))
+    FuncReturn(func(Param(0), static_cast<u32>(Param(1)), static_cast<s32>(Param(2)),
+                    static_cast<s32>(Param(3)))
                    .raw);
 }
 
@@ -226,7 +237,7 @@ void SvcWrap() {
 
 template <void func(s64)>
 void SvcWrap() {
-    func((s64)Param(0));
+    func(static_cast<s64>(Param(0)));
 }
 
 template <void func(u64, u64 len)>
@@ -239,4 +250,9 @@ void SvcWrap() {
     func(Param(0), Param(1), Param(2));
 }
 
+template <void func(u32, u64, u64)>
+void SvcWrap() {
+    func(static_cast<u32>(Param(0)), Param(1), Param(2));
+}
+
 } // namespace Kernel
diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp
index 8e514cf9a..dd5cd9ced 100644
--- a/src/core/hle/kernel/thread.cpp
+++ b/src/core/hle/kernel/thread.cpp
@@ -4,9 +4,9 @@
 
 #include <algorithm>
 #include <cinttypes>
+#include <optional>
 #include <vector>
 
-#include <boost/optional.hpp>
 #include <boost/range/algorithm_ext/erase.hpp>
 
 #include "common/assert.h"
@@ -94,10 +94,10 @@ void Thread::CancelWakeupTimer() {
     CoreTiming::UnscheduleEventThreadsafe(kernel.ThreadWakeupCallbackEventType(), callback_handle);
 }
 
-static boost::optional<s32> GetNextProcessorId(u64 mask) {
+static std::optional<s32> GetNextProcessorId(u64 mask) {
     for (s32 index = 0; index < Core::NUM_CPU_CORES; ++index) {
         if (mask & (1ULL << index)) {
-            if (!Core::System::GetInstance().Scheduler(index)->GetCurrentThread()) {
+            if (!Core::System::GetInstance().Scheduler(index).GetCurrentThread()) {
                 // Core is enabled and not running any threads, use this one
                 return index;
             }
@@ -142,19 +142,19 @@ void Thread::ResumeFromWait() {
 
     status = ThreadStatus::Ready;
 
-    boost::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask);
+    std::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask);
     if (!new_processor_id) {
         new_processor_id = processor_id;
     }
     if (ideal_core != -1 &&
-        Core::System::GetInstance().Scheduler(ideal_core)->GetCurrentThread() == nullptr) {
+        Core::System::GetInstance().Scheduler(ideal_core).GetCurrentThread() == nullptr) {
         new_processor_id = ideal_core;
     }
 
     ASSERT(*new_processor_id < 4);
 
     // Add thread to new core's scheduler
-    auto& next_scheduler = Core::System::GetInstance().Scheduler(*new_processor_id);
+    auto* next_scheduler = &Core::System::GetInstance().Scheduler(*new_processor_id);
 
     if (*new_processor_id != processor_id) {
         // Remove thread from previous core's scheduler
@@ -169,7 +169,7 @@ void Thread::ResumeFromWait() {
     next_scheduler->ScheduleThread(this, current_priority);
 
     // Change thread's scheduler
-    scheduler = next_scheduler.get();
+    scheduler = next_scheduler;
 
     Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule();
 }
@@ -183,18 +183,15 @@ void Thread::ResumeFromWait() {
  */
 static void ResetThreadContext(Core::ARM_Interface::ThreadContext& context, VAddr stack_top,
                                VAddr entry_point, u64 arg) {
-    memset(&context, 0, sizeof(Core::ARM_Interface::ThreadContext));
-
+    context = {};
     context.cpu_registers[0] = arg;
     context.pc = entry_point;
     context.sp = stack_top;
-    context.pstate = 0;
-    context.fpcr = 0;
 }
 
 ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name, VAddr entry_point,
                                             u32 priority, u64 arg, s32 processor_id,
-                                            VAddr stack_top, SharedPtr<Process> owner_process) {
+                                            VAddr stack_top, Process& owner_process) {
     // Check if priority is in ranged. Lowest priority -> highest priority id.
     if (priority > THREADPRIO_LOWEST) {
         LOG_ERROR(Kernel_SVC, "Invalid thread priority: {}", priority);
@@ -208,7 +205,7 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
 
     // TODO(yuriks): Other checks, returning 0xD9001BEA
 
-    if (!Memory::IsValidVirtualAddress(*owner_process, entry_point)) {
+    if (!Memory::IsValidVirtualAddress(owner_process, entry_point)) {
         LOG_ERROR(Kernel_SVC, "(name={}): invalid entry {:016X}", name, entry_point);
         // TODO (bunnei): Find the correct error code to use here
         return ResultCode(-1);
@@ -232,8 +229,8 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
     thread->wait_handle = 0;
     thread->name = std::move(name);
     thread->callback_handle = kernel.ThreadWakeupCallbackHandleTable().Create(thread).Unwrap();
-    thread->owner_process = owner_process;
-    thread->scheduler = Core::System::GetInstance().Scheduler(processor_id).get();
+    thread->owner_process = &owner_process;
+    thread->scheduler = &Core::System::GetInstance().Scheduler(processor_id);
     thread->scheduler->AddThread(thread, priority);
     thread->tls_address = thread->owner_process->MarkNextAvailableTLSSlotAsUsed(*thread);
 
@@ -264,12 +261,12 @@ SharedPtr<Thread> SetupMainThread(KernelCore& kernel, VAddr entry_point, u32 pri
     // Initialize new "main" thread
     const VAddr stack_top = owner_process.VMManager().GetTLSIORegionEndAddress();
     auto thread_res = Thread::Create(kernel, "main", entry_point, priority, 0, THREADPROCESSORID_0,
-                                     stack_top, &owner_process);
+                                     stack_top, owner_process);
 
     SharedPtr<Thread> thread = std::move(thread_res).Unwrap();
 
     // Register 1 must be a handle to the main thread
-    const Handle guest_handle = kernel.HandleTable().Create(thread).Unwrap();
+    const Handle guest_handle = owner_process.GetHandleTable().Create(thread).Unwrap();
     thread->SetGuestHandle(guest_handle);
     thread->GetContext().cpu_registers[1] = guest_handle;
 
@@ -372,20 +369,20 @@ void Thread::ChangeCore(u32 core, u64 mask) {
         return;
     }
 
-    boost::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
+    std::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
 
     if (!new_processor_id) {
         new_processor_id = processor_id;
     }
     if (ideal_core != -1 &&
-        Core::System::GetInstance().Scheduler(ideal_core)->GetCurrentThread() == nullptr) {
+        Core::System::GetInstance().Scheduler(ideal_core).GetCurrentThread() == nullptr) {
         new_processor_id = ideal_core;
     }
 
     ASSERT(*new_processor_id < 4);
 
     // Add thread to new core's scheduler
-    auto& next_scheduler = Core::System::GetInstance().Scheduler(*new_processor_id);
+    auto* next_scheduler = &Core::System::GetInstance().Scheduler(*new_processor_id);
 
     if (*new_processor_id != processor_id) {
         // Remove thread from previous core's scheduler
@@ -400,7 +397,7 @@ void Thread::ChangeCore(u32 core, u64 mask) {
     next_scheduler->ScheduleThread(this, current_priority);
 
     // Change thread's scheduler
-    scheduler = next_scheduler.get();
+    scheduler = next_scheduler;
 
     Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule();
 }
diff --git a/src/core/hle/kernel/thread.h b/src/core/hle/kernel/thread.h
index c6ffbd28c..4a6e11239 100644
--- a/src/core/hle/kernel/thread.h
+++ b/src/core/hle/kernel/thread.h
@@ -89,7 +89,7 @@ public:
     static ResultVal<SharedPtr<Thread>> Create(KernelCore& kernel, std::string name,
                                                VAddr entry_point, u32 priority, u64 arg,
                                                s32 processor_id, VAddr stack_top,
-                                               SharedPtr<Process> owner_process);
+                                               Process& owner_process);
 
     std::string GetName() const override {
         return name;
@@ -258,15 +258,23 @@ public:
         return last_running_ticks;
     }
 
+    u64 GetTotalCPUTimeTicks() const {
+        return total_cpu_time_ticks;
+    }
+
+    void UpdateCPUTimeTicks(u64 ticks) {
+        total_cpu_time_ticks += ticks;
+    }
+
     s32 GetProcessorID() const {
         return processor_id;
     }
 
-    SharedPtr<Process>& GetOwnerProcess() {
+    Process* GetOwnerProcess() {
         return owner_process;
     }
 
-    const SharedPtr<Process>& GetOwnerProcess() const {
+    const Process* GetOwnerProcess() const {
         return owner_process;
     }
 
@@ -378,7 +386,8 @@ private:
     u32 nominal_priority = 0; ///< Nominal thread priority, as set by the emulated application
     u32 current_priority = 0; ///< Current thread priority, can be temporarily changed
 
-    u64 last_running_ticks = 0; ///< CPU tick when thread was last running
+    u64 total_cpu_time_ticks = 0; ///< Total CPU running ticks.
+    u64 last_running_ticks = 0;   ///< CPU tick when thread was last running
 
     s32 processor_id = 0;
 
@@ -386,7 +395,7 @@ private:
     u64 tpidr_el0 = 0;     ///< TPIDR_EL0 read/write system register.
 
     /// Process that owns this thread
-    SharedPtr<Process> owner_process;
+    Process* owner_process;
 
     /// Objects that the thread is waiting on, in the same order as they were
     /// passed to WaitSynchronization1/N.
diff --git a/src/core/hle/kernel/vm_manager.cpp b/src/core/hle/kernel/vm_manager.cpp
index e412309fd..1a92c8f70 100644
--- a/src/core/hle/kernel/vm_manager.cpp
+++ b/src/core/hle/kernel/vm_manager.cpp
@@ -143,6 +143,26 @@ ResultVal<VMManager::VMAHandle> VMManager::MapBackingMemory(VAddr target, u8* me
     return MakeResult<VMAHandle>(MergeAdjacent(vma_handle));
 }
 
+ResultVal<VAddr> VMManager::FindFreeRegion(u64 size) const {
+    // Find the first Free VMA.
+    const VAddr base = GetASLRRegionBaseAddress();
+    const VMAHandle vma_handle = std::find_if(vma_map.begin(), vma_map.end(), [&](const auto& vma) {
+        if (vma.second.type != VMAType::Free)
+            return false;
+
+        const VAddr vma_end = vma.second.base + vma.second.size;
+        return vma_end > base && vma_end >= base + size;
+    });
+
+    if (vma_handle == vma_map.end()) {
+        // TODO(Subv): Find the correct error code here.
+        return ResultCode(-1);
+    }
+
+    const VAddr target = std::max(base, vma_handle->second.base);
+    return MakeResult<VAddr>(target);
+}
+
 ResultVal<VMManager::VMAHandle> VMManager::MapMMIO(VAddr target, PAddr paddr, u64 size,
                                                    MemoryState state,
                                                    Memory::MemoryHookPointer mmio_handler) {
@@ -393,30 +413,35 @@ void VMManager::InitializeMemoryRegionRanges(FileSys::ProgramAddressSpaceType ty
 
     switch (type) {
     case FileSys::ProgramAddressSpaceType::Is32Bit:
+    case FileSys::ProgramAddressSpaceType::Is32BitNoMap:
         address_space_width = 32;
         code_region_base = 0x200000;
         code_region_end = code_region_base + 0x3FE00000;
-        map_region_size = 0x40000000;
-        heap_region_size = 0x40000000;
+        aslr_region_base = 0x200000;
+        aslr_region_end = aslr_region_base + 0xFFE00000;
+        if (type == FileSys::ProgramAddressSpaceType::Is32Bit) {
+            map_region_size = 0x40000000;
+            heap_region_size = 0x40000000;
+        } else {
+            map_region_size = 0;
+            heap_region_size = 0x80000000;
+        }
         break;
     case FileSys::ProgramAddressSpaceType::Is36Bit:
         address_space_width = 36;
         code_region_base = 0x8000000;
         code_region_end = code_region_base + 0x78000000;
+        aslr_region_base = 0x8000000;
+        aslr_region_end = aslr_region_base + 0xFF8000000;
         map_region_size = 0x180000000;
         heap_region_size = 0x180000000;
         break;
-    case FileSys::ProgramAddressSpaceType::Is32BitNoMap:
-        address_space_width = 32;
-        code_region_base = 0x200000;
-        code_region_end = code_region_base + 0x3FE00000;
-        map_region_size = 0;
-        heap_region_size = 0x80000000;
-        break;
     case FileSys::ProgramAddressSpaceType::Is39Bit:
         address_space_width = 39;
         code_region_base = 0x8000000;
         code_region_end = code_region_base + 0x80000000;
+        aslr_region_base = 0x8000000;
+        aslr_region_end = aslr_region_base + 0x7FF8000000;
         map_region_size = 0x1000000000;
         heap_region_size = 0x180000000;
         new_map_region_size = 0x80000000;
@@ -490,6 +515,38 @@ u64 VMManager::GetAddressSpaceWidth() const {
     return address_space_width;
 }
 
+VAddr VMManager::GetASLRRegionBaseAddress() const {
+    return aslr_region_base;
+}
+
+VAddr VMManager::GetASLRRegionEndAddress() const {
+    return aslr_region_end;
+}
+
+u64 VMManager::GetASLRRegionSize() const {
+    return aslr_region_end - aslr_region_base;
+}
+
+bool VMManager::IsWithinASLRRegion(VAddr begin, u64 size) const {
+    const VAddr range_end = begin + size;
+    const VAddr aslr_start = GetASLRRegionBaseAddress();
+    const VAddr aslr_end = GetASLRRegionEndAddress();
+
+    if (aslr_start > begin || begin > range_end || range_end - 1 > aslr_end - 1) {
+        return false;
+    }
+
+    if (range_end > heap_region_base && heap_region_end > begin) {
+        return false;
+    }
+
+    if (range_end > map_region_base && map_region_end > begin) {
+        return false;
+    }
+
+    return true;
+}
+
 VAddr VMManager::GetCodeRegionBaseAddress() const {
     return code_region_base;
 }
diff --git a/src/core/hle/kernel/vm_manager.h b/src/core/hle/kernel/vm_manager.h
index 015559a64..2447cbb8f 100644
--- a/src/core/hle/kernel/vm_manager.h
+++ b/src/core/hle/kernel/vm_manager.h
@@ -158,6 +158,14 @@ public:
     ResultVal<VMAHandle> MapBackingMemory(VAddr target, u8* memory, u64 size, MemoryState state);
 
     /**
+     * Finds the first free address that can hold a region of the desired size.
+     *
+     * @param size Size of the desired region.
+     * @return The found free address.
+     */
+    ResultVal<VAddr> FindFreeRegion(u64 size) const;
+
+    /**
      * Maps a memory-mapped IO region at a given address.
      *
      * @param target The guest address to start the mapping at.
@@ -205,6 +213,18 @@ public:
     /// Gets the address space width in bits.
     u64 GetAddressSpaceWidth() const;
 
+    /// Gets the base address of the ASLR region.
+    VAddr GetASLRRegionBaseAddress() const;
+
+    /// Gets the end address of the ASLR region.
+    VAddr GetASLRRegionEndAddress() const;
+
+    /// Determines whether or not the specified address range is within the ASLR region.
+    bool IsWithinASLRRegion(VAddr address, u64 size) const;
+
+    /// Gets the size of the ASLR region
+    u64 GetASLRRegionSize() const;
+
     /// Gets the base address of the code region.
     VAddr GetCodeRegionBaseAddress() const;
 
@@ -306,6 +326,9 @@ private:
     VAddr address_space_base = 0;
     VAddr address_space_end = 0;
 
+    VAddr aslr_region_base = 0;
+    VAddr aslr_region_end = 0;
+
     VAddr code_region_base = 0;
     VAddr code_region_end = 0;
 
diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp
index e61748ca3..c6437a671 100644
--- a/src/core/hle/service/acc/acc.cpp
+++ b/src/core/hle/service/acc/acc.cpp
@@ -2,9 +2,13 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include <array>
+#include "common/common_paths.h"
 #include "common/common_types.h"
+#include "common/file_util.h"
 #include "common/logging/log.h"
+#include "common/string_util.h"
 #include "common/swap.h"
 #include "core/core_timing.h"
 #include "core/hle/ipc_helpers.h"
@@ -16,6 +20,7 @@
 #include "core/hle/service/acc/profile_manager.h"
 
 namespace Service::Account {
+
 // TODO: RE this structure
 struct UserData {
     INSERT_PADDING_WORDS(1);
@@ -27,6 +32,29 @@ struct UserData {
 };
 static_assert(sizeof(UserData) == 0x80, "UserData structure has incorrect size");
 
+// Smallest JPEG https://github.com/mathiasbynens/small/blob/master/jpeg.jpg
+// used as a backup should the one on disk not exist
+constexpr u32 backup_jpeg_size = 107;
+constexpr std::array<u8, backup_jpeg_size> backup_jpeg{{
+    0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x02, 0x02,
+    0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04, 0x08, 0x06, 0x06, 0x05,
+    0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a, 0x0c, 0x0f, 0x0c, 0x0a, 0x0b, 0x0e,
+    0x0b, 0x09, 0x09, 0x0d, 0x11, 0x0d, 0x0e, 0x0f, 0x10, 0x10, 0x11, 0x10, 0x0a, 0x0c, 0x12, 0x13,
+    0x12, 0x10, 0x13, 0x0f, 0x10, 0x10, 0x10, 0xff, 0xc9, 0x00, 0x0b, 0x08, 0x00, 0x01, 0x00, 0x01,
+    0x01, 0x01, 0x11, 0x00, 0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08,
+    0x01, 0x01, 0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
+}};
+
+static std::string GetImagePath(UUID uuid) {
+    return FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+           "/system/save/8000000000000010/su/avators/" + uuid.FormatSwitch() + ".jpg";
+}
+
+static constexpr u32 SanitizeJPEGSize(std::size_t size) {
+    constexpr std::size_t max_jpeg_image_size = 0x20000;
+    return static_cast<u32>(std::min(size, max_jpeg_image_size));
+}
+
 class IProfile final : public ServiceFramework<IProfile> {
 public:
     explicit IProfile(UUID user_id, ProfileManager& profile_manager)
@@ -73,32 +101,42 @@ private:
     }
 
     void LoadImage(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_ACC, "(STUBBED) called");
-        // smallest jpeg https://github.com/mathiasbynens/small/blob/master/jpeg.jpg
-        // TODO(mailwl): load actual profile image from disk, width 256px, max size 0x20000
-        constexpr u32 jpeg_size = 107;
-        static constexpr std::array<u8, jpeg_size> jpeg{
-            0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03,
-            0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04,
-            0x08, 0x06, 0x06, 0x05, 0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a,
-            0x0c, 0x0f, 0x0c, 0x0a, 0x0b, 0x0e, 0x0b, 0x09, 0x09, 0x0d, 0x11, 0x0d, 0x0e, 0x0f,
-            0x10, 0x10, 0x11, 0x10, 0x0a, 0x0c, 0x12, 0x13, 0x12, 0x10, 0x13, 0x0f, 0x10, 0x10,
-            0x10, 0xff, 0xc9, 0x00, 0x0b, 0x08, 0x00, 0x01, 0x00, 0x01, 0x01, 0x01, 0x11, 0x00,
-            0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08, 0x01, 0x01,
-            0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
-        };
-        ctx.WriteBuffer(jpeg);
+        LOG_DEBUG(Service_ACC, "called");
+
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(jpeg_size);
+
+        const FileUtil::IOFile image(GetImagePath(user_id), "rb");
+        if (!image.IsOpen()) {
+            LOG_WARNING(Service_ACC,
+                        "Failed to load user provided image! Falling back to built-in backup...");
+            ctx.WriteBuffer(backup_jpeg);
+            rb.Push<u32>(backup_jpeg_size);
+            return;
+        }
+
+        const u32 size = SanitizeJPEGSize(image.GetSize());
+        std::vector<u8> buffer(size);
+        image.ReadBytes(buffer.data(), buffer.size());
+
+        ctx.WriteBuffer(buffer.data(), buffer.size());
+        rb.Push<u32>(size);
     }
 
     void GetImageSize(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_ACC, "(STUBBED) called");
-        constexpr u32 jpeg_size = 107;
+        LOG_DEBUG(Service_ACC, "called");
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(jpeg_size);
+
+        const FileUtil::IOFile image(GetImagePath(user_id), "rb");
+
+        if (!image.IsOpen()) {
+            LOG_WARNING(Service_ACC,
+                        "Failed to load user provided image! Falling back to built-in backup...");
+            rb.Push<u32>(backup_jpeg_size);
+        } else {
+            rb.Push<u32>(SanitizeJPEGSize(image.GetSize()));
+        }
     }
 
     const ProfileManager& profile_manager;
diff --git a/src/core/hle/service/acc/profile_manager.cpp b/src/core/hle/service/acc/profile_manager.cpp
index bcb3475db..c08394e4c 100644
--- a/src/core/hle/service/acc/profile_manager.cpp
+++ b/src/core/hle/service/acc/profile_manager.cpp
@@ -3,41 +3,66 @@
 // Refer to the license.txt file included.
 
 #include <random>
-#include <boost/optional.hpp>
+
+#include "common/file_util.h"
 #include "core/hle/service/acc/profile_manager.h"
 #include "core/settings.h"
 
 namespace Service::Account {
+
+struct UserRaw {
+    UUID uuid;
+    UUID uuid2;
+    u64 timestamp;
+    ProfileUsername username;
+    INSERT_PADDING_BYTES(0x80);
+};
+static_assert(sizeof(UserRaw) == 0xC8, "UserRaw has incorrect size.");
+
+struct ProfileDataRaw {
+    INSERT_PADDING_BYTES(0x10);
+    std::array<UserRaw, MAX_USERS> users;
+};
+static_assert(sizeof(ProfileDataRaw) == 0x650, "ProfileDataRaw has incorrect size.");
+
 // TODO(ogniK): Get actual error codes
 constexpr ResultCode ERROR_TOO_MANY_USERS(ErrorModule::Account, -1);
 constexpr ResultCode ERROR_USER_ALREADY_EXISTS(ErrorModule::Account, -2);
 constexpr ResultCode ERROR_ARGUMENT_IS_NULL(ErrorModule::Account, 20);
 
-const UUID& UUID::Generate() {
+constexpr char ACC_SAVE_AVATORS_BASE_PATH[] = "/system/save/8000000000000010/su/avators/";
+
+UUID UUID::Generate() {
     std::random_device device;
     std::mt19937 gen(device());
     std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max());
-    uuid[0] = distribution(gen);
-    uuid[1] = distribution(gen);
-    return *this;
+    return UUID{distribution(gen), distribution(gen)};
 }
 
 ProfileManager::ProfileManager() {
-    // TODO(ogniK): Create the default user we have for now until loading/saving users is added
-    auto user_uuid = UUID{1, 0};
-    ASSERT(CreateNewUser(user_uuid, Settings::values.username).IsSuccess());
-    OpenUser(user_uuid);
+    ParseUserSaveFile();
+
+    if (user_count == 0)
+        CreateNewUser(UUID::Generate(), "yuzu");
+
+    auto current = std::clamp<int>(Settings::values.current_user, 0, MAX_USERS - 1);
+    if (UserExistsIndex(current))
+        current = 0;
+
+    OpenUser(*GetUser(current));
 }
 
-ProfileManager::~ProfileManager() = default;
+ProfileManager::~ProfileManager() {
+    WriteUserSaveFile();
+}
 
 /// After a users creation it needs to be "registered" to the system. AddToProfiles handles the
 /// internal management of the users profiles
-boost::optional<std::size_t> ProfileManager::AddToProfiles(const ProfileInfo& user) {
+std::optional<std::size_t> ProfileManager::AddToProfiles(const ProfileInfo& profile) {
     if (user_count >= MAX_USERS) {
-        return boost::none;
+        return {};
     }
-    profiles[user_count] = user;
+    profiles[user_count] = profile;
     return user_count++;
 }
 
@@ -56,7 +81,7 @@ bool ProfileManager::RemoveProfileAtIndex(std::size_t index) {
 
 /// Helper function to register a user to the system
 ResultCode ProfileManager::AddUser(const ProfileInfo& user) {
-    if (AddToProfiles(user) == boost::none) {
+    if (!AddToProfiles(user)) {
         return ERROR_TOO_MANY_USERS;
     }
     return RESULT_SUCCESS;
@@ -101,31 +126,40 @@ ResultCode ProfileManager::CreateNewUser(UUID uuid, const std::string& username)
     return CreateNewUser(uuid, username_output);
 }
 
+std::optional<UUID> ProfileManager::GetUser(std::size_t index) const {
+    if (index >= MAX_USERS) {
+        return {};
+    }
+
+    return profiles[index].user_uuid;
+}
+
 /// Returns a users profile index based on their user id.
-boost::optional<std::size_t> ProfileManager::GetUserIndex(const UUID& uuid) const {
+std::optional<std::size_t> ProfileManager::GetUserIndex(const UUID& uuid) const {
     if (!uuid) {
-        return boost::none;
+        return {};
     }
-    auto iter = std::find_if(profiles.begin(), profiles.end(),
-                             [&uuid](const ProfileInfo& p) { return p.user_uuid == uuid; });
+
+    const auto iter = std::find_if(profiles.begin(), profiles.end(),
+                                   [&uuid](const ProfileInfo& p) { return p.user_uuid == uuid; });
     if (iter == profiles.end()) {
-        return boost::none;
+        return {};
     }
+
     return static_cast<std::size_t>(std::distance(profiles.begin(), iter));
 }
 
 /// Returns a users profile index based on their profile
-boost::optional<std::size_t> ProfileManager::GetUserIndex(const ProfileInfo& user) const {
+std::optional<std::size_t> ProfileManager::GetUserIndex(const ProfileInfo& user) const {
     return GetUserIndex(user.user_uuid);
 }
 
 /// Returns the data structure used by the switch when GetProfileBase is called on acc:*
-bool ProfileManager::GetProfileBase(boost::optional<std::size_t> index,
-                                    ProfileBase& profile) const {
-    if (index == boost::none || index >= MAX_USERS) {
+bool ProfileManager::GetProfileBase(std::optional<std::size_t> index, ProfileBase& profile) const {
+    if (!index || index >= MAX_USERS) {
         return false;
     }
-    const auto& prof_info = profiles[index.get()];
+    const auto& prof_info = profiles[*index];
     profile.user_uuid = prof_info.user_uuid;
     profile.username = prof_info.username;
     profile.timestamp = prof_info.creation_time;
@@ -134,7 +168,7 @@ bool ProfileManager::GetProfileBase(boost::optional<std::size_t> index,
 
 /// Returns the data structure used by the switch when GetProfileBase is called on acc:*
 bool ProfileManager::GetProfileBase(UUID uuid, ProfileBase& profile) const {
-    auto idx = GetUserIndex(uuid);
+    const auto idx = GetUserIndex(uuid);
     return GetProfileBase(idx, profile);
 }
 
@@ -161,26 +195,34 @@ std::size_t ProfileManager::GetOpenUserCount() const {
 
 /// Checks if a user id exists in our profile manager
 bool ProfileManager::UserExists(UUID uuid) const {
-    return (GetUserIndex(uuid) != boost::none);
+    return GetUserIndex(uuid).has_value();
+}
+
+bool ProfileManager::UserExistsIndex(std::size_t index) const {
+    if (index >= MAX_USERS)
+        return false;
+    return profiles[index].user_uuid.uuid != INVALID_UUID;
 }
 
 /// Opens a specific user
 void ProfileManager::OpenUser(UUID uuid) {
-    auto idx = GetUserIndex(uuid);
-    if (idx == boost::none) {
+    const auto idx = GetUserIndex(uuid);
+    if (!idx) {
         return;
     }
-    profiles[idx.get()].is_open = true;
+
+    profiles[*idx].is_open = true;
     last_opened_user = uuid;
 }
 
 /// Closes a specific user
 void ProfileManager::CloseUser(UUID uuid) {
-    auto idx = GetUserIndex(uuid);
-    if (idx == boost::none) {
+    const auto idx = GetUserIndex(uuid);
+    if (!idx) {
         return;
     }
-    profiles[idx.get()].is_open = false;
+
+    profiles[*idx].is_open = false;
 }
 
 /// Gets all valid user ids on the system
@@ -210,10 +252,10 @@ UUID ProfileManager::GetLastOpenedUser() const {
 }
 
 /// Return the users profile base and the unknown arbitary data.
-bool ProfileManager::GetProfileBaseAndData(boost::optional<std::size_t> index, ProfileBase& profile,
+bool ProfileManager::GetProfileBaseAndData(std::optional<std::size_t> index, ProfileBase& profile,
                                            ProfileData& data) const {
     if (GetProfileBase(index, profile)) {
-        data = profiles[index.get()].data;
+        data = profiles[*index].data;
         return true;
     }
     return false;
@@ -222,7 +264,7 @@ bool ProfileManager::GetProfileBaseAndData(boost::optional<std::size_t> index, P
 /// Return the users profile base and the unknown arbitary data.
 bool ProfileManager::GetProfileBaseAndData(UUID uuid, ProfileBase& profile,
                                            ProfileData& data) const {
-    auto idx = GetUserIndex(uuid);
+    const auto idx = GetUserIndex(uuid);
     return GetProfileBaseAndData(idx, profile, data);
 }
 
@@ -239,4 +281,96 @@ bool ProfileManager::CanSystemRegisterUser() const {
     // emulate qlaunch. Update this to dynamically change.
 }
 
+bool ProfileManager::RemoveUser(UUID uuid) {
+    const auto index = GetUserIndex(uuid);
+    if (!index) {
+        return false;
+    }
+
+    profiles[*index] = ProfileInfo{};
+    std::stable_partition(profiles.begin(), profiles.end(),
+                          [](const ProfileInfo& profile) { return profile.user_uuid; });
+    return true;
+}
+
+bool ProfileManager::SetProfileBase(UUID uuid, const ProfileBase& profile_new) {
+    const auto index = GetUserIndex(uuid);
+    if (!index || profile_new.user_uuid == UUID(INVALID_UUID)) {
+        return false;
+    }
+
+    auto& profile = profiles[*index];
+    profile.user_uuid = profile_new.user_uuid;
+    profile.username = profile_new.username;
+    profile.creation_time = profile_new.timestamp;
+
+    return true;
+}
+
+void ProfileManager::ParseUserSaveFile() {
+    FileUtil::IOFile save(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                              ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat",
+                          "rb");
+
+    if (!save.IsOpen()) {
+        LOG_WARNING(Service_ACC, "Failed to load profile data from save data... Generating new "
+                                 "user 'yuzu' with random UUID.");
+        return;
+    }
+
+    ProfileDataRaw data;
+    if (save.ReadBytes(&data, sizeof(ProfileDataRaw)) != sizeof(ProfileDataRaw)) {
+        LOG_WARNING(Service_ACC, "profiles.dat is smaller than expected... Generating new user "
+                                 "'yuzu' with random UUID.");
+        return;
+    }
+
+    for (std::size_t i = 0; i < MAX_USERS; ++i) {
+        const auto& user = data.users[i];
+
+        if (user.uuid != UUID(INVALID_UUID))
+            AddUser({user.uuid, user.username, user.timestamp, {}, false});
+    }
+
+    std::stable_partition(profiles.begin(), profiles.end(),
+                          [](const ProfileInfo& profile) { return profile.user_uuid; });
+}
+
+void ProfileManager::WriteUserSaveFile() {
+    ProfileDataRaw raw{};
+
+    for (std::size_t i = 0; i < MAX_USERS; ++i) {
+        raw.users[i].username = profiles[i].username;
+        raw.users[i].uuid2 = profiles[i].user_uuid;
+        raw.users[i].uuid = profiles[i].user_uuid;
+        raw.users[i].timestamp = profiles[i].creation_time;
+    }
+
+    const auto raw_path =
+        FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + "/system/save/8000000000000010";
+    if (FileUtil::Exists(raw_path) && !FileUtil::IsDirectory(raw_path))
+        FileUtil::Delete(raw_path);
+
+    const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                      ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat";
+
+    if (!FileUtil::CreateFullPath(path)) {
+        LOG_WARNING(Service_ACC, "Failed to create full path of profiles.dat. Create the directory "
+                                 "nand/system/save/8000000000000010/su/avators to mitigate this "
+                                 "issue.");
+        return;
+    }
+
+    FileUtil::IOFile save(path, "wb");
+
+    if (!save.IsOpen()) {
+        LOG_WARNING(Service_ACC, "Failed to write save data to file... No changes to user data "
+                                 "made in current session will be saved.");
+        return;
+    }
+
+    save.Resize(sizeof(ProfileDataRaw));
+    save.WriteBytes(&raw, sizeof(ProfileDataRaw));
+}
+
 }; // namespace Service::Account
diff --git a/src/core/hle/service/acc/profile_manager.h b/src/core/hle/service/acc/profile_manager.h
index bffd4cf4d..747c46c20 100644
--- a/src/core/hle/service/acc/profile_manager.h
+++ b/src/core/hle/service/acc/profile_manager.h
@@ -5,8 +5,8 @@
 #pragma once
 
 #include <array>
+#include <optional>
 
-#include "boost/optional.hpp"
 #include "common/common_types.h"
 #include "common/swap.h"
 #include "core/hle/result.h"
@@ -36,7 +36,7 @@ struct UUID {
     }
 
     // TODO(ogniK): Properly generate uuids based on RFC-4122
-    const UUID& Generate();
+    static UUID Generate();
 
     // Set the UUID to {0,0} to be considered an invalid user
     void Invalidate() {
@@ -45,10 +45,20 @@ struct UUID {
     std::string Format() const {
         return fmt::format("0x{:016X}{:016X}", uuid[1], uuid[0]);
     }
+
+    std::string FormatSwitch() const {
+        std::array<u8, 16> s{};
+        std::memcpy(s.data(), uuid.data(), sizeof(u128));
+        return fmt::format("{:02x}{:02x}{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{"
+                           ":02x}{:02x}{:02x}{:02x}{:02x}",
+                           s[0], s[1], s[2], s[3], s[4], s[5], s[6], s[7], s[8], s[9], s[10], s[11],
+                           s[12], s[13], s[14], s[15]);
+    }
 };
 static_assert(sizeof(UUID) == 16, "UUID is an invalid size!");
 
-using ProfileUsername = std::array<u8, 0x20>;
+constexpr std::size_t profile_username_size = 32;
+using ProfileUsername = std::array<u8, profile_username_size>;
 using ProfileData = std::array<u8, MAX_DATA>;
 using UserIDArray = std::array<UUID, MAX_USERS>;
 
@@ -81,18 +91,19 @@ static_assert(sizeof(ProfileBase) == 0x38, "ProfileBase is an invalid size");
 /// objects
 class ProfileManager {
 public:
-    ProfileManager(); // TODO(ogniK): Load from system save
+    ProfileManager();
     ~ProfileManager();
 
     ResultCode AddUser(const ProfileInfo& user);
     ResultCode CreateNewUser(UUID uuid, const ProfileUsername& username);
     ResultCode CreateNewUser(UUID uuid, const std::string& username);
-    boost::optional<std::size_t> GetUserIndex(const UUID& uuid) const;
-    boost::optional<std::size_t> GetUserIndex(const ProfileInfo& user) const;
-    bool GetProfileBase(boost::optional<std::size_t> index, ProfileBase& profile) const;
+    std::optional<UUID> GetUser(std::size_t index) const;
+    std::optional<std::size_t> GetUserIndex(const UUID& uuid) const;
+    std::optional<std::size_t> GetUserIndex(const ProfileInfo& user) const;
+    bool GetProfileBase(std::optional<std::size_t> index, ProfileBase& profile) const;
     bool GetProfileBase(UUID uuid, ProfileBase& profile) const;
     bool GetProfileBase(const ProfileInfo& user, ProfileBase& profile) const;
-    bool GetProfileBaseAndData(boost::optional<std::size_t> index, ProfileBase& profile,
+    bool GetProfileBaseAndData(std::optional<std::size_t> index, ProfileBase& profile,
                                ProfileData& data) const;
     bool GetProfileBaseAndData(UUID uuid, ProfileBase& profile, ProfileData& data) const;
     bool GetProfileBaseAndData(const ProfileInfo& user, ProfileBase& profile,
@@ -100,6 +111,7 @@ public:
     std::size_t GetUserCount() const;
     std::size_t GetOpenUserCount() const;
     bool UserExists(UUID uuid) const;
+    bool UserExistsIndex(std::size_t index) const;
     void OpenUser(UUID uuid);
     void CloseUser(UUID uuid);
     UserIDArray GetOpenUsers() const;
@@ -108,11 +120,17 @@ public:
 
     bool CanSystemRegisterUser() const;
 
+    bool RemoveUser(UUID uuid);
+    bool SetProfileBase(UUID uuid, const ProfileBase& profile_new);
+
 private:
+    void ParseUserSaveFile();
+    void WriteUserSaveFile();
+    std::optional<std::size_t> AddToProfiles(const ProfileInfo& profile);
+    bool RemoveProfileAtIndex(std::size_t index);
+
     std::array<ProfileInfo, MAX_USERS> profiles{};
     std::size_t user_count = 0;
-    boost::optional<std::size_t> AddToProfiles(const ProfileInfo& profile);
-    bool RemoveProfileAtIndex(std::size_t index);
     UUID last_opened_user{INVALID_UUID};
 };
 
diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp
index 69bfce1c1..ac3ff9f20 100644
--- a/src/core/hle/service/am/am.cpp
+++ b/src/core/hle/service/am/am.cpp
@@ -4,17 +4,20 @@
 
 #include <array>
 #include <cinttypes>
+#include <cstring>
 #include <stack>
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/process.h"
+#include "core/hle/service/acc/profile_manager.h"
 #include "core/hle/service/am/am.h"
 #include "core/hle/service/am/applet_ae.h"
 #include "core/hle/service/am/applet_oe.h"
 #include "core/hle/service/am/idle.h"
 #include "core/hle/service/am/omm.h"
 #include "core/hle/service/am/spsm.h"
+#include "core/hle/service/am/tcap.h"
 #include "core/hle/service/apm/apm.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/hle/service/nvflinger/nvflinger.h"
@@ -25,14 +28,29 @@
 
 namespace Service::AM {
 
+constexpr u32 POP_LAUNCH_PARAMETER_MAGIC = 0xC79497CA;
+
+struct LaunchParameters {
+    u32_le magic;
+    u32_le is_account_selected;
+    u128 current_user;
+    INSERT_PADDING_BYTES(0x70);
+};
+static_assert(sizeof(LaunchParameters) == 0x88);
+
 IWindowController::IWindowController() : ServiceFramework("IWindowController") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "CreateWindow"},
         {1, &IWindowController::GetAppletResourceUserId, "GetAppletResourceUserId"},
         {10, &IWindowController::AcquireForegroundRights, "AcquireForegroundRights"},
         {11, nullptr, "ReleaseForegroundRights"},
         {12, nullptr, "RejectToChangeIntoBackground"},
+        {20, nullptr, "SetAppletWindowVisibility"},
+        {21, nullptr, "SetAppletGpuTimeSlice"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -87,6 +105,7 @@ void IAudioController::GetLibraryAppletExpectedMasterVolume(Kernel::HLERequestCo
 }
 
 IDisplayController::IDisplayController() : ServiceFramework("IDisplayController") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "GetLastForegroundCaptureImage"},
         {1, nullptr, "UpdateLastForegroundCaptureImage"},
@@ -117,7 +136,11 @@ IDisplayController::IDisplayController() : ServiceFramework("IDisplayController"
         {25, nullptr, "ReleaseLastForegroundCaptureSharedBuffer"},
         {26, nullptr, "AcquireCallerAppletCaptureSharedBuffer"},
         {27, nullptr, "ReleaseCallerAppletCaptureSharedBuffer"},
+        // 6.0.0+
+        {28, nullptr, "TakeScreenShotOfOwnLayerEx"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -128,6 +151,7 @@ IDebugFunctions::~IDebugFunctions() = default;
 
 ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     : ServiceFramework("ISelfController"), nvflinger(std::move(nvflinger)) {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "Exit"},
         {1, &ISelfController::LockExit, "LockExit"},
@@ -136,10 +160,8 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
         {4, nullptr, "LeaveFatalSection"},
         {9, &ISelfController::GetLibraryAppletLaunchableEvent, "GetLibraryAppletLaunchableEvent"},
         {10, &ISelfController::SetScreenShotPermission, "SetScreenShotPermission"},
-        {11, &ISelfController::SetOperationModeChangedNotification,
-         "SetOperationModeChangedNotification"},
-        {12, &ISelfController::SetPerformanceModeChangedNotification,
-         "SetPerformanceModeChangedNotification"},
+        {11, &ISelfController::SetOperationModeChangedNotification, "SetOperationModeChangedNotification"},
+        {12, &ISelfController::SetPerformanceModeChangedNotification, "SetPerformanceModeChangedNotification"},
         {13, &ISelfController::SetFocusHandlingMode, "SetFocusHandlingMode"},
         {14, &ISelfController::SetRestartMessageEnabled, "SetRestartMessageEnabled"},
         {15, nullptr, "SetScreenShotAppletIdentityInfo"},
@@ -165,7 +187,12 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
         {69, nullptr, "IsAutoSleepDisabled"},
         {70, nullptr, "ReportMultimediaError"},
         {80, nullptr, "SetWirelessPriorityMode"},
+        {90, nullptr, "GetAccumulatedSuspendedTickValue"},
+        {91, nullptr, "GetAccumulatedSuspendedTickChangedEvent"},
+        {1000, nullptr, "GetDebugStorageChannel"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 
     auto& kernel = Core::System::GetInstance().Kernel();
@@ -312,6 +339,7 @@ void ISelfController::GetIdleTimeDetectionExtension(Kernel::HLERequestContext& c
 }
 
 ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"},
         {1, &ICommonStateGetter::ReceiveMessage, "ReceiveMessage"},
@@ -336,11 +364,12 @@ ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter"
         {52, nullptr, "SwitchLcdBacklight"},
         {55, nullptr, "IsInControllerFirmwareUpdateSection"},
         {60, &ICommonStateGetter::GetDefaultDisplayResolution, "GetDefaultDisplayResolution"},
-        {61, &ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent,
-         "GetDefaultDisplayResolutionChangeEvent"},
+        {61, &ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent, "GetDefaultDisplayResolutionChangeEvent"},
         {62, nullptr, "GetHdcpAuthenticationState"},
         {63, nullptr, "GetHdcpAuthenticationStateChangeEvent"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 
     auto& kernel = Core::System::GetInstance().Kernel();
@@ -432,11 +461,14 @@ class IStorageAccessor final : public ServiceFramework<IStorageAccessor> {
 public:
     explicit IStorageAccessor(std::vector<u8> buffer)
         : ServiceFramework("IStorageAccessor"), buffer(std::move(buffer)) {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IStorageAccessor::GetSize, "GetSize"},
             {10, &IStorageAccessor::Write, "Write"},
             {11, &IStorageAccessor::Read, "Read"},
         };
+        // clang-format on
+
         RegisterHandlers(functions);
     }
 
@@ -489,10 +521,13 @@ class IStorage final : public ServiceFramework<IStorage> {
 public:
     explicit IStorage(std::vector<u8> buffer)
         : ServiceFramework("IStorage"), buffer(std::move(buffer)) {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IStorage::Open, "Open"},
             {1, nullptr, "OpenTransferStorage"},
         };
+        // clang-format on
+
         RegisterHandlers(functions);
     }
 
@@ -512,6 +547,7 @@ private:
 class ILibraryAppletAccessor final : public ServiceFramework<ILibraryAppletAccessor> {
 public:
     explicit ILibraryAppletAccessor() : ServiceFramework("ILibraryAppletAccessor") {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {0, &ILibraryAppletAccessor::GetAppletStateChangedEvent, "GetAppletStateChangedEvent"},
             {1, nullptr, "IsCompleted"},
@@ -532,6 +568,8 @@ public:
             {150, nullptr, "RequestForAppletToGetForeground"},
             {160, nullptr, "GetIndirectLayerConsumerHandle"},
         };
+        // clang-format on
+
         RegisterHandlers(functions);
 
         auto& kernel = Core::System::GetInstance().Kernel();
@@ -624,13 +662,13 @@ void ILibraryAppletCreator::CreateStorage(Kernel::HLERequestContext& ctx) {
 }
 
 IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationFunctions") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {1, &IApplicationFunctions::PopLaunchParameter, "PopLaunchParameter"},
         {10, nullptr, "CreateApplicationAndPushAndRequestToStart"},
         {11, nullptr, "CreateApplicationAndPushAndRequestToStartForQuest"},
         {12, nullptr, "CreateApplicationAndRequestToStart"},
-        {13, &IApplicationFunctions::CreateApplicationAndRequestToStartForQuest,
-         "CreateApplicationAndRequestToStartForQuest"},
+        {13, &IApplicationFunctions::CreateApplicationAndRequestToStartForQuest, "CreateApplicationAndRequestToStartForQuest"},
         {20, &IApplicationFunctions::EnsureSaveData, "EnsureSaveData"},
         {21, &IApplicationFunctions::GetDesiredLanguage, "GetDesiredLanguage"},
         {22, &IApplicationFunctions::SetTerminateResult, "SetTerminateResult"},
@@ -638,10 +676,10 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
         {24, nullptr, "GetLaunchStorageInfoForDebug"},
         {25, nullptr, "ExtendSaveData"},
         {26, nullptr, "GetSaveDataSize"},
-        {30, nullptr, "BeginBlockingHomeButtonShortAndLongPressed"},
-        {31, nullptr, "EndBlockingHomeButtonShortAndLongPressed"},
-        {32, nullptr, "BeginBlockingHomeButton"},
-        {33, nullptr, "EndBlockingHomeButton"},
+        {30, &IApplicationFunctions::BeginBlockingHomeButtonShortAndLongPressed, "BeginBlockingHomeButtonShortAndLongPressed"},
+        {31, &IApplicationFunctions::EndBlockingHomeButtonShortAndLongPressed, "EndBlockingHomeButtonShortAndLongPressed"},
+        {32, &IApplicationFunctions::BeginBlockingHomeButton, "BeginBlockingHomeButton"},
+        {33, &IApplicationFunctions::EndBlockingHomeButton, "EndBlockingHomeButton"},
         {40, &IApplicationFunctions::NotifyRunning, "NotifyRunning"},
         {50, &IApplicationFunctions::GetPseudoDeviceId, "GetPseudoDeviceId"},
         {60, nullptr, "SetMediaPlaybackStateForApplication"},
@@ -664,26 +702,57 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
         {1000, nullptr, "CreateMovieMaker"},
         {1001, nullptr, "PrepareForJit"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
 IApplicationFunctions::~IApplicationFunctions() = default;
 
+void IApplicationFunctions::BeginBlockingHomeButtonShortAndLongPressed(
+    Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
+void IApplicationFunctions::EndBlockingHomeButtonShortAndLongPressed(
+    Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
+void IApplicationFunctions::BeginBlockingHomeButton(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
+void IApplicationFunctions::EndBlockingHomeButton(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
 void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) {
-    constexpr std::array<u8, 0x88> data{{
-        0xca, 0x97, 0x94, 0xc7, // Magic
-        1,    0,    0,    0,    // IsAccountSelected (bool)
-        1,    0,    0,    0,    // User Id (word 0)
-        0,    0,    0,    0,    // User Id (word 1)
-        0,    0,    0,    0,    // User Id (word 2)
-        0,    0,    0,    0     // User Id (word 3)
-    }};
+    LaunchParameters params{};
 
-    std::vector<u8> buffer(data.begin(), data.end());
+    params.magic = POP_LAUNCH_PARAMETER_MAGIC;
+    params.is_account_selected = 1;
+
+    Account::ProfileManager profile_manager{};
+    const auto uuid = profile_manager.GetUser(Settings::values.current_user);
+    ASSERT(uuid);
+    params.current_user = uuid->uuid;
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
 
     rb.Push(RESULT_SUCCESS);
+
+    std::vector<u8> buffer(sizeof(LaunchParameters));
+    std::memcpy(buffer.data(), &params, buffer.size());
+
     rb.PushIpcInterface<AM::IStorage>(buffer);
 
     LOG_DEBUG(Service_AM, "called");
@@ -776,9 +845,11 @@ void InstallInterfaces(SM::ServiceManager& service_manager,
     std::make_shared<IdleSys>()->InstallAsService(service_manager);
     std::make_shared<OMM>()->InstallAsService(service_manager);
     std::make_shared<SPSM>()->InstallAsService(service_manager);
+    std::make_shared<TCAP>()->InstallAsService(service_manager);
 }
 
 IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {10, &IHomeMenuFunctions::RequestToGetForeground, "RequestToGetForeground"},
         {11, nullptr, "LockForeground"},
@@ -787,7 +858,10 @@ IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions"
         {21, nullptr, "GetPopFromGeneralChannelEvent"},
         {30, nullptr, "GetHomeButtonWriterLockAccessor"},
         {31, nullptr, "GetWriterLockAccessorEx"},
+        {100, nullptr, "PopRequestLaunchApplicationForDebug"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -800,6 +874,7 @@ void IHomeMenuFunctions::RequestToGetForeground(Kernel::HLERequestContext& ctx)
 }
 
 IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStateController") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "RequestToEnterSleep"},
         {1, nullptr, "EnterSleep"},
@@ -813,18 +888,23 @@ IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStat
         {14, nullptr, "ShouldSleepOnBoot"},
         {15, nullptr, "GetHdcpAuthenticationFailedEvent"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
 IGlobalStateController::~IGlobalStateController() = default;
 
 IApplicationCreator::IApplicationCreator() : ServiceFramework("IApplicationCreator") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "CreateApplication"},
         {1, nullptr, "PopLaunchRequestedApplication"},
         {10, nullptr, "CreateSystemApplication"},
         {100, nullptr, "PopFloatingApplicationForDevelopment"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -832,6 +912,7 @@ IApplicationCreator::~IApplicationCreator() = default;
 
 IProcessWindingController::IProcessWindingController()
     : ServiceFramework("IProcessWindingController") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "GetLaunchReason"},
         {11, nullptr, "OpenCallingLibraryApplet"},
@@ -842,6 +923,8 @@ IProcessWindingController::IProcessWindingController()
         {40, nullptr, "ReserveToStartAndWaitAndUnwindThis"},
         {41, nullptr, "ReserveToStartAndWait"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h
index b39b0d838..095f94851 100644
--- a/src/core/hle/service/am/am.h
+++ b/src/core/hle/service/am/am.h
@@ -154,6 +154,10 @@ private:
     void SetGamePlayRecordingState(Kernel::HLERequestContext& ctx);
     void NotifyRunning(Kernel::HLERequestContext& ctx);
     void GetPseudoDeviceId(Kernel::HLERequestContext& ctx);
+    void BeginBlockingHomeButtonShortAndLongPressed(Kernel::HLERequestContext& ctx);
+    void EndBlockingHomeButtonShortAndLongPressed(Kernel::HLERequestContext& ctx);
+    void BeginBlockingHomeButton(Kernel::HLERequestContext& ctx);
+    void EndBlockingHomeButton(Kernel::HLERequestContext& ctx);
 };
 
 class IHomeMenuFunctions final : public ServiceFramework<IHomeMenuFunctions> {
diff --git a/src/core/hle/service/am/applet_ae.cpp b/src/core/hle/service/am/applet_ae.cpp
index 4296c255e..68ea778e8 100644
--- a/src/core/hle/service/am/applet_ae.cpp
+++ b/src/core/hle/service/am/applet_ae.cpp
@@ -211,6 +211,7 @@ void AppletAE::OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx) {
 
 AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     : ServiceFramework("appletAE"), nvflinger(std::move(nvflinger)) {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {100, &AppletAE::OpenSystemAppletProxy, "OpenSystemAppletProxy"},
         {200, &AppletAE::OpenLibraryAppletProxyOld, "OpenLibraryAppletProxyOld"},
@@ -218,7 +219,10 @@ AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
         {300, nullptr, "OpenOverlayAppletProxy"},
         {350, nullptr, "OpenSystemApplicationProxy"},
         {400, nullptr, "CreateSelfLibraryAppletCreatorForDevelop"},
+        {401, nullptr, "GetSystemAppletControllerForDebug"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
diff --git a/src/core/hle/service/am/applet_oe.cpp b/src/core/hle/service/am/applet_oe.cpp
index e45cf6e20..60717afd9 100644
--- a/src/core/hle/service/am/applet_oe.cpp
+++ b/src/core/hle/service/am/applet_oe.cpp
@@ -14,6 +14,7 @@ class IApplicationProxy final : public ServiceFramework<IApplicationProxy> {
 public:
     explicit IApplicationProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
         : ServiceFramework("IApplicationProxy"), nvflinger(std::move(nvflinger)) {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IApplicationProxy::GetCommonStateGetter, "GetCommonStateGetter"},
             {1, &IApplicationProxy::GetSelfController, "GetSelfController"},
@@ -25,6 +26,8 @@ public:
             {20, &IApplicationProxy::GetApplicationFunctions, "GetApplicationFunctions"},
             {1000, &IApplicationProxy::GetDebugFunctions, "GetDebugFunctions"},
         };
+        // clang-format on
+
         RegisterHandlers(functions);
     }
 
diff --git a/src/core/hle/service/am/idle.cpp b/src/core/hle/service/am/idle.cpp
index 0e3088bc8..f814fe2c0 100644
--- a/src/core/hle/service/am/idle.cpp
+++ b/src/core/hle/service/am/idle.cpp
@@ -12,9 +12,9 @@ IdleSys::IdleSys() : ServiceFramework{"idle:sys"} {
         {0, nullptr, "GetAutoPowerDownEvent"},
         {1, nullptr, "Unknown1"},
         {2, nullptr, "Unknown2"},
-        {3, nullptr, "Unknown3"},
-        {4, nullptr, "Unknown4"},
-        {5, nullptr, "Unknown5"},
+        {3, nullptr, "SetHandlingContext"},
+        {4, nullptr, "LoadAndApplySettings"},
+        {5, nullptr, "ReportUserIsActive"},
     };
     // clang-format on
 
diff --git a/src/core/hle/service/am/omm.cpp b/src/core/hle/service/am/omm.cpp
index 1c37f849f..6ab3fb906 100644
--- a/src/core/hle/service/am/omm.cpp
+++ b/src/core/hle/service/am/omm.cpp
@@ -17,22 +17,24 @@ OMM::OMM() : ServiceFramework{"omm"} {
         {5, nullptr, "GetCradleStatus"},
         {6, nullptr, "FadeInDisplay"},
         {7, nullptr, "FadeOutDisplay"},
-        {8, nullptr, "Unknown1"},
-        {9, nullptr, "Unknown2"},
-        {10, nullptr, "Unknown3"},
-        {11, nullptr, "Unknown4"},
-        {12, nullptr, "Unknown5"},
-        {13, nullptr, "Unknown6"},
-        {14, nullptr, "Unknown7"},
-        {15, nullptr, "Unknown8"},
-        {16, nullptr, "Unknown9"},
-        {17, nullptr, "Unknown10"},
-        {18, nullptr, "Unknown11"},
-        {19, nullptr, "Unknown12"},
-        {20, nullptr, "Unknown13"},
-        {21, nullptr, "Unknown14"},
-        {22, nullptr, "Unknown15"},
-        {23, nullptr, "Unknown16"},
+        {8, nullptr, "GetCradleFwVersion"},
+        {9, nullptr, "NotifyCecSettingsChanged"},
+        {10, nullptr, "SetOperationModePolicy"},
+        {11, nullptr, "GetDefaultDisplayResolution"},
+        {12, nullptr, "GetDefaultDisplayResolutionChangeEvent"},
+        {13, nullptr, "UpdateDefaultDisplayResolution"},
+        {14, nullptr, "ShouldSleepOnBoot"},
+        {15, nullptr, "NotifyHdcpApplicationExecutionStarted"},
+        {16, nullptr, "NotifyHdcpApplicationExecutionFinished"},
+        {17, nullptr, "NotifyHdcpApplicationDrawingStarted"},
+        {18, nullptr, "NotifyHdcpApplicationDrawingFinished"},
+        {19, nullptr, "GetHdcpAuthenticationFailedEvent"},
+        {20, nullptr, "GetHdcpAuthenticationFailedEmulationEnabled"},
+        {21, nullptr, "SetHdcpAuthenticationFailedEmulation"},
+        {22, nullptr, "GetHdcpStateChangeEvent"},
+        {23, nullptr, "GetHdcpState"},
+        {24, nullptr, "ShowCardUpdateProcessing"},
+        {25, nullptr, "SetApplicationCecSettingsAndNotifyChanged"},
     };
     // clang-format on
 
diff --git a/src/core/hle/service/am/tcap.cpp b/src/core/hle/service/am/tcap.cpp
new file mode 100644
index 000000000..a75cbdda8
--- /dev/null
+++ b/src/core/hle/service/am/tcap.cpp
@@ -0,0 +1,23 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/hle/service/am/tcap.h"
+
+namespace Service::AM {
+
+TCAP::TCAP() : ServiceFramework{"tcap"} {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetContinuousHighSkinTemperatureEvent"},
+        {1, nullptr, "SetOperationMode"},
+        {2, nullptr, "LoadAndApplySettings"},
+    };
+    // clang-format on
+
+    RegisterHandlers(functions);
+}
+
+TCAP::~TCAP() = default;
+
+} // namespace Service::AM
diff --git a/src/core/hle/service/am/tcap.h b/src/core/hle/service/am/tcap.h
new file mode 100644
index 000000000..2021b55d1
--- /dev/null
+++ b/src/core/hle/service/am/tcap.h
@@ -0,0 +1,17 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/service/service.h"
+
+namespace Service::AM {
+
+class TCAP final : public ServiceFramework<TCAP> {
+public:
+    explicit TCAP();
+    ~TCAP() override;
+};
+
+} // namespace Service::AM
diff --git a/src/core/hle/service/aoc/aoc_u.cpp b/src/core/hle/service/aoc/aoc_u.cpp
index 79580bcd9..54305cf05 100644
--- a/src/core/hle/service/aoc/aoc_u.cpp
+++ b/src/core/hle/service/aoc/aoc_u.cpp
@@ -7,10 +7,13 @@
 #include <vector>
 #include "common/logging/log.h"
 #include "core/file_sys/content_archive.h"
+#include "core/file_sys/control_metadata.h"
 #include "core/file_sys/nca_metadata.h"
 #include "core/file_sys/partition_filesystem.h"
+#include "core/file_sys/patch_manager.h"
 #include "core/file_sys/registered_cache.h"
 #include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/event.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/service/aoc/aoc_u.h"
 #include "core/hle/service/filesystem/filesystem.h"
@@ -19,10 +22,10 @@
 namespace Service::AOC {
 
 constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000;
-constexpr u64 DLC_BASE_TO_AOC_ID_MASK = 0x1000;
+constexpr u64 DLC_BASE_TO_AOC_ID = 0x1000;
 
-static bool CheckAOCTitleIDMatchesBase(u64 base, u64 aoc) {
-    return (aoc & DLC_BASE_TITLE_ID_MASK) == base;
+static bool CheckAOCTitleIDMatchesBase(u64 title_id, u64 base) {
+    return (title_id & DLC_BASE_TITLE_ID_MASK) == base;
 }
 
 static std::vector<u64> AccumulateAOCTitleIDs() {
@@ -53,21 +56,25 @@ AOC_U::AOC_U() : ServiceFramework("aoc:u"), add_on_content(AccumulateAOCTitleIDs
         {5, &AOC_U::GetAddOnContentBaseId, "GetAddOnContentBaseId"},
         {6, nullptr, "PrepareAddOnContentByApplicationId"},
         {7, &AOC_U::PrepareAddOnContent, "PrepareAddOnContent"},
-        {8, nullptr, "GetAddOnContentListChangedEvent"},
+        {8, &AOC_U::GetAddOnContentListChangedEvent, "GetAddOnContentListChangedEvent"},
     };
     RegisterHandlers(functions);
+
+    auto& kernel = Core::System::GetInstance().Kernel();
+    aoc_change_event = Kernel::Event::Create(kernel, Kernel::ResetType::Sticky,
+                                             "GetAddOnContentListChanged:Event");
 }
 
 AOC_U::~AOC_U() = default;
 
 void AOC_U::CountAddOnContent(Kernel::HLERequestContext& ctx) {
-    IPC::ResponseBuilder rb{ctx, 4};
+    IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
 
     const auto current = Core::System::GetInstance().CurrentProcess()->GetTitleID();
-    rb.Push<u32>(std::count_if(add_on_content.begin(), add_on_content.end(), [&current](u64 tid) {
-        return (tid & DLC_BASE_TITLE_ID_MASK) == current;
-    }));
+    rb.Push<u32>(static_cast<u32>(
+        std::count_if(add_on_content.begin(), add_on_content.end(),
+                      [current](u64 tid) { return CheckAOCTitleIDMatchesBase(tid, current); })));
 }
 
 void AOC_U::ListAddOnContent(Kernel::HLERequestContext& ctx) {
@@ -91,20 +98,30 @@ void AOC_U::ListAddOnContent(Kernel::HLERequestContext& ctx) {
         return;
     }
 
-    count = std::min<size_t>(out.size() - offset, count);
+    count = static_cast<u32>(std::min<size_t>(out.size() - offset, count));
     std::rotate(out.begin(), out.begin() + offset, out.end());
     out.resize(count);
 
     ctx.WriteBuffer(out);
 
-    IPC::ResponseBuilder rb{ctx, 2};
+    IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
+    rb.Push(count);
 }
 
 void AOC_U::GetAddOnContentBaseId(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 4};
     rb.Push(RESULT_SUCCESS);
-    rb.Push(Core::System::GetInstance().CurrentProcess()->GetTitleID() | DLC_BASE_TO_AOC_ID_MASK);
+    const auto title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID();
+    FileSys::PatchManager pm{title_id};
+
+    const auto res = pm.GetControlMetadata();
+    if (res.first == nullptr) {
+        rb.Push(title_id + DLC_BASE_TO_AOC_ID);
+        return;
+    }
+
+    rb.Push(res.first->GetDLCBaseTitleId());
 }
 
 void AOC_U::PrepareAddOnContent(Kernel::HLERequestContext& ctx) {
@@ -118,6 +135,14 @@ void AOC_U::PrepareAddOnContent(Kernel::HLERequestContext& ctx) {
     rb.Push(RESULT_SUCCESS);
 }
 
+void AOC_U::GetAddOnContentListChangedEvent(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AOC, "(STUBBED) called");
+
+    IPC::ResponseBuilder rb{ctx, 2, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushCopyObjects(aoc_change_event);
+}
+
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     std::make_shared<AOC_U>()->InstallAsService(service_manager);
 }
diff --git a/src/core/hle/service/aoc/aoc_u.h b/src/core/hle/service/aoc/aoc_u.h
index b3c7cab7a..68d94fdaa 100644
--- a/src/core/hle/service/aoc/aoc_u.h
+++ b/src/core/hle/service/aoc/aoc_u.h
@@ -18,8 +18,10 @@ private:
     void ListAddOnContent(Kernel::HLERequestContext& ctx);
     void GetAddOnContentBaseId(Kernel::HLERequestContext& ctx);
     void PrepareAddOnContent(Kernel::HLERequestContext& ctx);
+    void GetAddOnContentListChangedEvent(Kernel::HLERequestContext& ctx);
 
     std::vector<u64> add_on_content;
+    Kernel::SharedPtr<Kernel::Event> aoc_change_event;
 };
 
 /// Registers all AOC services with the specified service manager.
diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp
index 6073f4ecd..fac6785a5 100644
--- a/src/core/hle/service/audio/audren_u.cpp
+++ b/src/core/hle/service/audio/audren_u.cpp
@@ -22,20 +22,22 @@ class IAudioRenderer final : public ServiceFramework<IAudioRenderer> {
 public:
     explicit IAudioRenderer(AudioCore::AudioRendererParameter audren_params)
         : ServiceFramework("IAudioRenderer") {
+        // clang-format off
         static const FunctionInfo functions[] = {
-            {0, &IAudioRenderer::GetAudioRendererSampleRate, "GetAudioRendererSampleRate"},
-            {1, &IAudioRenderer::GetAudioRendererSampleCount, "GetAudioRendererSampleCount"},
-            {2, &IAudioRenderer::GetAudioRendererMixBufferCount, "GetAudioRendererMixBufferCount"},
-            {3, &IAudioRenderer::GetAudioRendererState, "GetAudioRendererState"},
-            {4, &IAudioRenderer::RequestUpdateAudioRenderer, "RequestUpdateAudioRenderer"},
-            {5, &IAudioRenderer::StartAudioRenderer, "StartAudioRenderer"},
-            {6, &IAudioRenderer::StopAudioRenderer, "StopAudioRenderer"},
+            {0, &IAudioRenderer::GetSampleRate, "GetSampleRate"},
+            {1, &IAudioRenderer::GetSampleCount, "GetSampleCount"},
+            {2, &IAudioRenderer::GetMixBufferCount, "GetMixBufferCount"},
+            {3, &IAudioRenderer::GetState, "GetState"},
+            {4, &IAudioRenderer::RequestUpdate, "RequestUpdate"},
+            {5, &IAudioRenderer::Start, "Start"},
+            {6, &IAudioRenderer::Stop, "Stop"},
             {7, &IAudioRenderer::QuerySystemEvent, "QuerySystemEvent"},
-            {8, nullptr, "SetAudioRendererRenderingTimeLimit"},
-            {9, nullptr, "GetAudioRendererRenderingTimeLimit"},
-            {10, nullptr, "RequestUpdateAudioRendererAuto"},
+            {8, nullptr, "SetRenderingTimeLimit"},
+            {9, nullptr, "GetRenderingTimeLimit"},
+            {10, nullptr, "RequestUpdateAuto"},
             {11, nullptr, "ExecuteAudioRendererRendering"},
         };
+        // clang-format on
         RegisterHandlers(functions);
 
         auto& kernel = Core::System::GetInstance().Kernel();
@@ -49,42 +51,42 @@ private:
         system_event->Signal();
     }
 
-    void GetAudioRendererSampleRate(Kernel::HLERequestContext& ctx) {
+    void GetSampleRate(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push<u32>(renderer->GetSampleRate());
         LOG_DEBUG(Service_Audio, "called");
     }
 
-    void GetAudioRendererSampleCount(Kernel::HLERequestContext& ctx) {
+    void GetSampleCount(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push<u32>(renderer->GetSampleCount());
         LOG_DEBUG(Service_Audio, "called");
     }
 
-    void GetAudioRendererState(Kernel::HLERequestContext& ctx) {
+    void GetState(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push<u32>(static_cast<u32>(renderer->GetStreamState()));
         LOG_DEBUG(Service_Audio, "called");
     }
 
-    void GetAudioRendererMixBufferCount(Kernel::HLERequestContext& ctx) {
+    void GetMixBufferCount(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push<u32>(renderer->GetMixBufferCount());
         LOG_DEBUG(Service_Audio, "called");
     }
 
-    void RequestUpdateAudioRenderer(Kernel::HLERequestContext& ctx) {
+    void RequestUpdate(Kernel::HLERequestContext& ctx) {
         ctx.WriteBuffer(renderer->UpdateAudioRenderer(ctx.ReadBuffer()));
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
         LOG_WARNING(Service_Audio, "(STUBBED) called");
     }
 
-    void StartAudioRenderer(Kernel::HLERequestContext& ctx) {
+    void Start(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2};
 
         rb.Push(RESULT_SUCCESS);
@@ -92,7 +94,7 @@ private:
         LOG_WARNING(Service_Audio, "(STUBBED) called");
     }
 
-    void StopAudioRenderer(Kernel::HLERequestContext& ctx) {
+    void Stop(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2};
 
         rb.Push(RESULT_SUCCESS);
@@ -129,6 +131,7 @@ public:
             {10, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceNameAuto"},
             {11, nullptr, "QueryAudioDeviceInputEvent"},
             {12, nullptr, "QueryAudioDeviceOutputEvent"},
+            {13, nullptr, "GetAudioSystemMasterVolumeSetting"},
         };
         RegisterHandlers(functions);
 
diff --git a/src/core/hle/service/audio/hwopus.cpp b/src/core/hle/service/audio/hwopus.cpp
index fc6067e59..783c39503 100644
--- a/src/core/hle/service/audio/hwopus.cpp
+++ b/src/core/hle/service/audio/hwopus.cpp
@@ -2,8 +2,10 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <chrono>
 #include <cstring>
 #include <memory>
+#include <optional>
 #include <vector>
 
 #include <opus.h>
@@ -33,7 +35,8 @@ public:
             {1, nullptr, "SetContext"},
             {2, nullptr, "DecodeInterleavedForMultiStream"},
             {3, nullptr, "SetContextForMultiStream"},
-            {4, nullptr, "Unknown4"},
+            {4, &IHardwareOpusDecoderManager::DecodeInterleavedWithPerformance,
+             "DecodeInterleavedWithPerformance"},
             {5, nullptr, "Unknown5"},
             {6, nullptr, "Unknown6"},
             {7, nullptr, "Unknown7"},
@@ -59,8 +62,31 @@ private:
         ctx.WriteBuffer(samples.data(), samples.size() * sizeof(s16));
     }
 
-    bool Decoder_DecodeInterleaved(u32& consumed, u32& sample_count, const std::vector<u8>& input,
-                                   std::vector<opus_int16>& output) {
+    void DecodeInterleavedWithPerformance(Kernel::HLERequestContext& ctx) {
+        u32 consumed = 0;
+        u32 sample_count = 0;
+        u64 performance = 0;
+        std::vector<opus_int16> samples(ctx.GetWriteBufferSize() / sizeof(opus_int16));
+        if (!Decoder_DecodeInterleaved(consumed, sample_count, ctx.ReadBuffer(), samples,
+                                       performance)) {
+            IPC::ResponseBuilder rb{ctx, 2};
+            // TODO(ogniK): Use correct error code
+            rb.Push(ResultCode(-1));
+            return;
+        }
+        IPC::ResponseBuilder rb{ctx, 6};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(consumed);
+        rb.Push<u64>(performance);
+        rb.Push<u32>(sample_count);
+        ctx.WriteBuffer(samples.data(), samples.size() * sizeof(s16));
+    }
+
+    bool Decoder_DecodeInterleaved(
+        u32& consumed, u32& sample_count, const std::vector<u8>& input,
+        std::vector<opus_int16>& output,
+        std::optional<std::reference_wrapper<u64>> performance_time = std::nullopt) {
+        const auto start_time = std::chrono::high_resolution_clock::now();
         std::size_t raw_output_sz = output.size() * sizeof(opus_int16);
         if (sizeof(OpusHeader) > input.size())
             return false;
@@ -80,8 +106,13 @@ private:
                         (static_cast<int>(raw_output_sz / sizeof(s16) / channel_count)), 0);
         if (out_sample_count < 0)
             return false;
+        const auto end_time = std::chrono::high_resolution_clock::now() - start_time;
         sample_count = out_sample_count;
         consumed = static_cast<u32>(sizeof(OpusHeader) + hdr.sz);
+        if (performance_time.has_value()) {
+            performance_time->get() =
+                std::chrono::duration_cast<std::chrono::milliseconds>(end_time).count();
+        }
         return true;
     }
 
@@ -130,7 +161,7 @@ void HwOpus::OpenOpusDecoder(Kernel::HLERequestContext& ctx) {
     ASSERT_MSG(channel_count == 1 || channel_count == 2, "Invalid channel count");
 
     std::size_t worker_sz = WorkerBufferSize(channel_count);
-    ASSERT_MSG(buffer_sz < worker_sz, "Worker buffer too large");
+    ASSERT_MSG(buffer_sz >= worker_sz, "Worker buffer too large");
     std::unique_ptr<OpusDecoder, OpusDeleter> decoder{
         static_cast<OpusDecoder*>(operator new(worker_sz))};
     if (opus_decoder_init(decoder.get(), sample_rate, channel_count)) {
diff --git a/src/core/hle/service/es/es.cpp b/src/core/hle/service/es/es.cpp
index d40f18565..6701cb913 100644
--- a/src/core/hle/service/es/es.cpp
+++ b/src/core/hle/service/es/es.cpp
@@ -9,6 +9,7 @@ namespace Service::ES {
 class ETicket final : public ServiceFramework<ETicket> {
 public:
     explicit ETicket() : ServiceFramework{"es"} {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {1, nullptr, "ImportTicket"},
             {2, nullptr, "ImportTicketCertificateSet"},
@@ -37,15 +38,18 @@ public:
             {25, nullptr, "DeletePrepurchaseRecord"},
             {26, nullptr, "DeleteAllPrepurchaseRecord"},
             {27, nullptr, "CountPrepurchaseRecord"},
-            {28, nullptr, "ListPrepurchaseRecord"},
+            {28, nullptr, "ListPrepurchaseRecordRightsIds"},
             {29, nullptr, "ListPrepurchaseRecordInfo"},
-            {30, nullptr, "Unknown1"},
-            {31, nullptr, "Unknown2"},
-            {32, nullptr, "Unknown3"},
-            {33, nullptr, "Unknown4"},
-            {34, nullptr, "Unknown5"},
-            {35, nullptr, "Unknown6"},
+            {30, nullptr, "CountTicket"},
+            {31, nullptr, "ListTicketRightsIds"},
+            {32, nullptr, "CountPrepurchaseRecordEx"},
+            {33, nullptr, "ListPrepurchaseRecordRightsIdsEx"},
+            {34, nullptr, "GetEncryptedTicketSize"},
+            {35, nullptr, "GetEncryptedTicketData"},
+            {36, nullptr, "DeleteAllInactiveELicenseRequiredPersonalizedTicket"},
+            {503, nullptr, "GetTitleKey"},
         };
+        // clang-format on
         RegisterHandlers(functions);
     }
 };
diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp
index aed2abb71..e32a7c48e 100644
--- a/src/core/hle/service/filesystem/filesystem.cpp
+++ b/src/core/hle/service/filesystem/filesystem.cpp
@@ -264,6 +264,15 @@ ResultCode RegisterBIS(std::unique_ptr<FileSys::BISFactory>&& factory) {
     return RESULT_SUCCESS;
 }
 
+void SetPackedUpdate(FileSys::VirtualFile update_raw) {
+    LOG_TRACE(Service_FS, "Setting packed update for romfs");
+
+    if (romfs_factory == nullptr)
+        return;
+
+    romfs_factory->SetPackedUpdate(std::move(update_raw));
+}
+
 ResultVal<FileSys::VirtualFile> OpenRomFSCurrentProcess() {
     LOG_TRACE(Service_FS, "Opening RomFS for current process");
 
@@ -310,13 +319,12 @@ ResultVal<FileSys::VirtualDir> OpenSDMC() {
     return sdmc_factory->Open();
 }
 
-std::shared_ptr<FileSys::RegisteredCacheUnion> GetUnionContents() {
-    return std::make_shared<FileSys::RegisteredCacheUnion>(
-        std::vector<std::shared_ptr<FileSys::RegisteredCache>>{
-            GetSystemNANDContents(), GetUserNANDContents(), GetSDMCContents()});
+std::unique_ptr<FileSys::RegisteredCacheUnion> GetUnionContents() {
+    return std::make_unique<FileSys::RegisteredCacheUnion>(std::vector<FileSys::RegisteredCache*>{
+        GetSystemNANDContents(), GetUserNANDContents(), GetSDMCContents()});
 }
 
-std::shared_ptr<FileSys::RegisteredCache> GetSystemNANDContents() {
+FileSys::RegisteredCache* GetSystemNANDContents() {
     LOG_TRACE(Service_FS, "Opening System NAND Contents");
 
     if (bis_factory == nullptr)
@@ -325,7 +333,7 @@ std::shared_ptr<FileSys::RegisteredCache> GetSystemNANDContents() {
     return bis_factory->GetSystemNANDContents();
 }
 
-std::shared_ptr<FileSys::RegisteredCache> GetUserNANDContents() {
+FileSys::RegisteredCache* GetUserNANDContents() {
     LOG_TRACE(Service_FS, "Opening User NAND Contents");
 
     if (bis_factory == nullptr)
@@ -334,7 +342,7 @@ std::shared_ptr<FileSys::RegisteredCache> GetUserNANDContents() {
     return bis_factory->GetUserNANDContents();
 }
 
-std::shared_ptr<FileSys::RegisteredCache> GetSDMCContents() {
+FileSys::RegisteredCache* GetSDMCContents() {
     LOG_TRACE(Service_FS, "Opening SDMC Contents");
 
     if (sdmc_factory == nullptr)
@@ -352,19 +360,19 @@ FileSys::VirtualDir GetModificationLoadRoot(u64 title_id) {
     return bis_factory->GetModificationLoadRoot(title_id);
 }
 
-void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite) {
+void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite) {
     if (overwrite) {
         bis_factory = nullptr;
         save_data_factory = nullptr;
         sdmc_factory = nullptr;
     }
 
-    auto nand_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir),
-                                             FileSys::Mode::ReadWrite);
-    auto sd_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir),
-                                           FileSys::Mode::ReadWrite);
-    auto load_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir),
-                                             FileSys::Mode::ReadWrite);
+    auto nand_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir),
+                                            FileSys::Mode::ReadWrite);
+    auto sd_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir),
+                                          FileSys::Mode::ReadWrite);
+    auto load_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir),
+                                            FileSys::Mode::ReadWrite);
 
     if (bis_factory == nullptr)
         bis_factory = std::make_unique<FileSys::BISFactory>(nand_directory, load_directory);
@@ -374,7 +382,7 @@ void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite) {
         sdmc_factory = std::make_unique<FileSys::SDMCFactory>(std::move(sd_directory));
 }
 
-void InstallInterfaces(SM::ServiceManager& service_manager, const FileSys::VirtualFilesystem& vfs) {
+void InstallInterfaces(SM::ServiceManager& service_manager, FileSys::VfsFilesystem& vfs) {
     romfs_factory = nullptr;
     CreateFactories(vfs, false);
     std::make_shared<FSP_LDR>()->InstallAsService(service_manager);
diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h
index 7039a2247..6ca5c5636 100644
--- a/src/core/hle/service/filesystem/filesystem.h
+++ b/src/core/hle/service/filesystem/filesystem.h
@@ -39,6 +39,7 @@ ResultCode RegisterSaveData(std::unique_ptr<FileSys::SaveDataFactory>&& factory)
 ResultCode RegisterSDMC(std::unique_ptr<FileSys::SDMCFactory>&& factory);
 ResultCode RegisterBIS(std::unique_ptr<FileSys::BISFactory>&& factory);
 
+void SetPackedUpdate(FileSys::VirtualFile update_raw);
 ResultVal<FileSys::VirtualFile> OpenRomFSCurrentProcess();
 ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId storage_id,
                                           FileSys::ContentRecordType type);
@@ -46,19 +47,19 @@ ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space,
                                             FileSys::SaveDataDescriptor save_struct);
 ResultVal<FileSys::VirtualDir> OpenSDMC();
 
-std::shared_ptr<FileSys::RegisteredCacheUnion> GetUnionContents();
+std::unique_ptr<FileSys::RegisteredCacheUnion> GetUnionContents();
 
-std::shared_ptr<FileSys::RegisteredCache> GetSystemNANDContents();
-std::shared_ptr<FileSys::RegisteredCache> GetUserNANDContents();
-std::shared_ptr<FileSys::RegisteredCache> GetSDMCContents();
+FileSys::RegisteredCache* GetSystemNANDContents();
+FileSys::RegisteredCache* GetUserNANDContents();
+FileSys::RegisteredCache* GetSDMCContents();
 
 FileSys::VirtualDir GetModificationLoadRoot(u64 title_id);
 
 // Creates the SaveData, SDMC, and BIS Factories. Should be called once and before any function
 // above is called.
-void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite = true);
+void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite = true);
 
-void InstallInterfaces(SM::ServiceManager& service_manager, const FileSys::VirtualFilesystem& vfs);
+void InstallInterfaces(SM::ServiceManager& service_manager, FileSys::VfsFilesystem& vfs);
 
 // A class that wraps a VfsDirectory with methods that return ResultVal and ResultCode instead of
 // pointers and booleans. This makes using a VfsDirectory with switch services much easier and
diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp
index d5dced429..c1c83a11d 100644
--- a/src/core/hle/service/filesystem/fsp_srv.cpp
+++ b/src/core/hle/service/filesystem/fsp_srv.cpp
@@ -17,6 +17,7 @@
 #include "core/file_sys/errors.h"
 #include "core/file_sys/mode.h"
 #include "core/file_sys/nca_metadata.h"
+#include "core/file_sys/patch_manager.h"
 #include "core/file_sys/savedata_factory.h"
 #include "core/file_sys/vfs.h"
 #include "core/hle/ipc_helpers.h"
@@ -272,8 +273,8 @@ public:
             {0, &IFileSystem::CreateFile, "CreateFile"},
             {1, &IFileSystem::DeleteFile, "DeleteFile"},
             {2, &IFileSystem::CreateDirectory, "CreateDirectory"},
-            {3, nullptr, "DeleteDirectory"},
-            {4, nullptr, "DeleteDirectoryRecursively"},
+            {3, &IFileSystem::DeleteDirectory, "DeleteDirectory"},
+            {4, &IFileSystem::DeleteDirectoryRecursively, "DeleteDirectoryRecursively"},
             {5, &IFileSystem::RenameFile, "RenameFile"},
             {6, nullptr, "RenameDirectory"},
             {7, &IFileSystem::GetEntryType, "GetEntryType"},
@@ -328,6 +329,30 @@ public:
         rb.Push(backend.CreateDirectory(name));
     }
 
+    void DeleteDirectory(Kernel::HLERequestContext& ctx) {
+        const IPC::RequestParser rp{ctx};
+
+        const auto file_buffer = ctx.ReadBuffer();
+        std::string name = Common::StringFromBuffer(file_buffer);
+
+        LOG_DEBUG(Service_FS, "called directory {}", name);
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(backend.DeleteDirectory(name));
+    }
+
+    void DeleteDirectoryRecursively(Kernel::HLERequestContext& ctx) {
+        const IPC::RequestParser rp{ctx};
+
+        const auto file_buffer = ctx.ReadBuffer();
+        std::string name = Common::StringFromBuffer(file_buffer);
+
+        LOG_DEBUG(Service_FS, "called directory {}", name);
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(backend.DeleteDirectoryRecursively(name));
+    }
+
     void RenameFile(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
 
@@ -630,6 +655,7 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) {
               static_cast<u8>(storage_id), unknown, title_id);
 
     auto data = OpenRomFS(title_id, storage_id, FileSys::ContentRecordType::Data);
+
     if (data.Failed()) {
         // TODO(DarkLordZach): Find the right error code to use here
         LOG_ERROR(Service_FS,
@@ -640,7 +666,9 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) {
         return;
     }
 
-    IStorage storage(std::move(data.Unwrap()));
+    FileSys::PatchManager pm{title_id};
+
+    IStorage storage(pm.PatchRomFS(std::move(data.Unwrap()), 0, FileSys::ContentRecordType::Data));
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
diff --git a/src/core/hle/service/hid/controllers/controller_base.cpp b/src/core/hle/service/hid/controllers/controller_base.cpp
new file mode 100644
index 000000000..0993a7815
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/controller_base.cpp
@@ -0,0 +1,30 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/hle/service/hid/controllers/controller_base.h"
+
+namespace Service::HID {
+
+ControllerBase::ControllerBase() = default;
+ControllerBase::~ControllerBase() = default;
+
+void ControllerBase::ActivateController() {
+    if (is_activated) {
+        OnRelease();
+    }
+    is_activated = true;
+    OnInit();
+}
+
+void ControllerBase::DeactivateController() {
+    if (is_activated) {
+        OnRelease();
+    }
+    is_activated = false;
+}
+
+bool ControllerBase::IsControllerActivated() const {
+    return is_activated;
+}
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/controller_base.h b/src/core/hle/service/hid/controllers/controller_base.h
new file mode 100644
index 000000000..f0e092b1b
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/controller_base.h
@@ -0,0 +1,45 @@
+// 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 "common/swap.h"
+
+namespace Service::HID {
+class ControllerBase {
+public:
+    ControllerBase();
+    virtual ~ControllerBase();
+
+    // Called when the controller is initialized
+    virtual void OnInit() = 0;
+
+    // When the controller is released
+    virtual void OnRelease() = 0;
+
+    // When the controller is requesting an update for the shared memory
+    virtual void OnUpdate(u8* data, std::size_t size) = 0;
+
+    // Called when input devices should be loaded
+    virtual void OnLoadInputDevices() = 0;
+
+    void ActivateController();
+
+    void DeactivateController();
+
+    bool IsControllerActivated() const;
+
+protected:
+    bool is_activated{false};
+
+    struct CommonHeader {
+        s64_le timestamp;
+        s64_le total_entry_count;
+        s64_le last_entry_index;
+        s64_le entry_count;
+    };
+    static_assert(sizeof(CommonHeader) == 0x20, "CommonHeader is an invalid size");
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/debug_pad.cpp b/src/core/hle/service/hid/controllers/debug_pad.cpp
new file mode 100644
index 000000000..3d100763f
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/debug_pad.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 <cstring>
+#include "common/common_types.h"
+#include "core/core_timing.h"
+#include "core/hle/service/hid/controllers/debug_pad.h"
+
+namespace Service::HID {
+
+Controller_DebugPad::Controller_DebugPad() = default;
+Controller_DebugPad::~Controller_DebugPad() = default;
+
+void Controller_DebugPad::OnInit() {}
+
+void Controller_DebugPad::OnRelease() {}
+
+void Controller_DebugPad::OnUpdate(u8* data, std::size_t size) {
+    shared_memory.header.timestamp = CoreTiming::GetTicks();
+    shared_memory.header.total_entry_count = 17;
+
+    if (!IsControllerActivated()) {
+        shared_memory.header.entry_count = 0;
+        shared_memory.header.last_entry_index = 0;
+        return;
+    }
+    shared_memory.header.entry_count = 16;
+
+    const auto& last_entry = shared_memory.pad_states[shared_memory.header.last_entry_index];
+    shared_memory.header.last_entry_index = (shared_memory.header.last_entry_index + 1) % 17;
+    auto& cur_entry = shared_memory.pad_states[shared_memory.header.last_entry_index];
+
+    cur_entry.sampling_number = last_entry.sampling_number + 1;
+    cur_entry.sampling_number2 = cur_entry.sampling_number;
+    // TODO(ogniK): Update debug pad states
+
+    std::memcpy(data, &shared_memory, sizeof(SharedMemory));
+}
+
+void Controller_DebugPad::OnLoadInputDevices() {}
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/debug_pad.h b/src/core/hle/service/hid/controllers/debug_pad.h
new file mode 100644
index 000000000..62b4f2682
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/debug_pad.h
@@ -0,0 +1,56 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/hle/service/hid/controllers/controller_base.h"
+
+namespace Service::HID {
+class Controller_DebugPad final : public ControllerBase {
+public:
+    Controller_DebugPad();
+    ~Controller_DebugPad() override;
+
+    // Called when the controller is initialized
+    void OnInit() override;
+
+    // When the controller is released
+    void OnRelease() override;
+
+    // When the controller is requesting an update for the shared memory
+    void OnUpdate(u8* data, std::size_t size) override;
+
+    // Called when input devices should be loaded
+    void OnLoadInputDevices() override;
+
+private:
+    struct AnalogStick {
+        s32_le x;
+        s32_le y;
+    };
+    static_assert(sizeof(AnalogStick) == 0x8);
+
+    struct PadStates {
+        s64_le sampling_number;
+        s64_le sampling_number2;
+        u32_le attribute;
+        u32_le button_state;
+        AnalogStick r_stick;
+        AnalogStick l_stick;
+    };
+    static_assert(sizeof(PadStates) == 0x28, "PadStates is an invalid state");
+
+    struct SharedMemory {
+        CommonHeader header;
+        std::array<PadStates, 17> pad_states;
+        INSERT_PADDING_BYTES(0x138);
+    };
+    static_assert(sizeof(SharedMemory) == 0x400, "SharedMemory is an invalid size");
+    SharedMemory shared_memory{};
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/gesture.cpp b/src/core/hle/service/hid/controllers/gesture.cpp
new file mode 100644
index 000000000..898572277
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/gesture.cpp
@@ -0,0 +1,43 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+#include "common/common_types.h"
+#include "core/core_timing.h"
+#include "core/hle/service/hid/controllers/gesture.h"
+
+namespace Service::HID {
+constexpr std::size_t SHARED_MEMORY_OFFSET = 0x3BA00;
+
+Controller_Gesture::Controller_Gesture() = default;
+Controller_Gesture::~Controller_Gesture() = default;
+
+void Controller_Gesture::OnInit() {}
+
+void Controller_Gesture::OnRelease() {}
+
+void Controller_Gesture::OnUpdate(u8* data, std::size_t size) {
+    shared_memory.header.timestamp = CoreTiming::GetTicks();
+    shared_memory.header.total_entry_count = 17;
+
+    if (!IsControllerActivated()) {
+        shared_memory.header.entry_count = 0;
+        shared_memory.header.last_entry_index = 0;
+        return;
+    }
+    shared_memory.header.entry_count = 16;
+
+    const auto& last_entry = shared_memory.gesture_states[shared_memory.header.last_entry_index];
+    shared_memory.header.last_entry_index = (shared_memory.header.last_entry_index + 1) % 17;
+    auto& cur_entry = shared_memory.gesture_states[shared_memory.header.last_entry_index];
+
+    cur_entry.sampling_number = last_entry.sampling_number + 1;
+    cur_entry.sampling_number2 = cur_entry.sampling_number;
+    // TODO(ogniK): Update gesture states
+
+    std::memcpy(data + SHARED_MEMORY_OFFSET, &shared_memory, sizeof(SharedMemory));
+}
+
+void Controller_Gesture::OnLoadInputDevices() {}
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/gesture.h b/src/core/hle/service/hid/controllers/gesture.h
new file mode 100644
index 000000000..1056ffbcd
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/gesture.h
@@ -0,0 +1,63 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/hle/service/hid/controllers/controller_base.h"
+
+namespace Service::HID {
+class Controller_Gesture final : public ControllerBase {
+public:
+    Controller_Gesture();
+    ~Controller_Gesture() override;
+
+    // Called when the controller is initialized
+    void OnInit() override;
+
+    // When the controller is released
+    void OnRelease() override;
+
+    // When the controller is requesting an update for the shared memory
+    void OnUpdate(u8* data, size_t size) override;
+
+    // Called when input devices should be loaded
+    void OnLoadInputDevices() override;
+
+private:
+    struct Locations {
+        s32_le x;
+        s32_le y;
+    };
+
+    struct GestureState {
+        s64_le sampling_number;
+        s64_le sampling_number2;
+
+        s64_le detection_count;
+        s32_le type;
+        s32_le dir;
+        s32_le x;
+        s32_le y;
+        s32_le delta_x;
+        s32_le delta_y;
+        f32 vel_x;
+        f32 vel_y;
+        s32_le attributes;
+        f32 scale;
+        f32 rotation;
+        s32_le location_count;
+        std::array<Locations, 4> locations;
+    };
+    static_assert(sizeof(GestureState) == 0x68, "GestureState is an invalid size");
+
+    struct SharedMemory {
+        CommonHeader header;
+        std::array<GestureState, 17> gesture_states;
+    };
+    SharedMemory shared_memory{};
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/keyboard.cpp b/src/core/hle/service/hid/controllers/keyboard.cpp
new file mode 100644
index 000000000..ccfbce9ac
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/keyboard.cpp
@@ -0,0 +1,43 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+#include "common/common_types.h"
+#include "core/core_timing.h"
+#include "core/hle/service/hid/controllers/keyboard.h"
+
+namespace Service::HID {
+constexpr std::size_t SHARED_MEMORY_OFFSET = 0x3800;
+
+Controller_Keyboard::Controller_Keyboard() = default;
+Controller_Keyboard::~Controller_Keyboard() = default;
+
+void Controller_Keyboard::OnInit() {}
+
+void Controller_Keyboard::OnRelease() {}
+
+void Controller_Keyboard::OnUpdate(u8* data, std::size_t size) {
+    shared_memory.header.timestamp = CoreTiming::GetTicks();
+    shared_memory.header.total_entry_count = 17;
+
+    if (!IsControllerActivated()) {
+        shared_memory.header.entry_count = 0;
+        shared_memory.header.last_entry_index = 0;
+        return;
+    }
+    shared_memory.header.entry_count = 16;
+
+    const auto& last_entry = shared_memory.pad_states[shared_memory.header.last_entry_index];
+    shared_memory.header.last_entry_index = (shared_memory.header.last_entry_index + 1) % 17;
+    auto& cur_entry = shared_memory.pad_states[shared_memory.header.last_entry_index];
+
+    cur_entry.sampling_number = last_entry.sampling_number + 1;
+    cur_entry.sampling_number2 = cur_entry.sampling_number;
+    // TODO(ogniK): Update keyboard states
+
+    std::memcpy(data + SHARED_MEMORY_OFFSET, &shared_memory, sizeof(SharedMemory));
+}
+
+void Controller_Keyboard::OnLoadInputDevices() {}
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/keyboard.h b/src/core/hle/service/hid/controllers/keyboard.h
new file mode 100644
index 000000000..493e68fce
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/keyboard.h
@@ -0,0 +1,50 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/hle/service/hid/controllers/controller_base.h"
+
+namespace Service::HID {
+class Controller_Keyboard final : public ControllerBase {
+public:
+    Controller_Keyboard();
+    ~Controller_Keyboard() override;
+
+    // Called when the controller is initialized
+    void OnInit() override;
+
+    // When the controller is released
+    void OnRelease() override;
+
+    // When the controller is requesting an update for the shared memory
+    void OnUpdate(u8* data, std::size_t size) override;
+
+    // Called when input devices should be loaded
+    void OnLoadInputDevices() override;
+
+private:
+    struct KeyboardState {
+        s64_le sampling_number;
+        s64_le sampling_number2;
+
+        s32_le modifier;
+        s32_le attribute;
+        std::array<u8, 32> key;
+    };
+    static_assert(sizeof(KeyboardState) == 0x38, "KeyboardState is an invalid size");
+
+    struct SharedMemory {
+        CommonHeader header;
+        std::array<KeyboardState, 17> pad_states;
+        INSERT_PADDING_BYTES(0x28);
+    };
+    static_assert(sizeof(SharedMemory) == 0x400, "SharedMemory is an invalid size");
+    SharedMemory shared_memory{};
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/mouse.cpp b/src/core/hle/service/hid/controllers/mouse.cpp
new file mode 100644
index 000000000..4e246a57d
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/mouse.cpp
@@ -0,0 +1,43 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+#include "common/common_types.h"
+#include "core/core_timing.h"
+#include "core/hle/service/hid/controllers/mouse.h"
+
+namespace Service::HID {
+constexpr std::size_t SHARED_MEMORY_OFFSET = 0x3400;
+
+Controller_Mouse::Controller_Mouse() = default;
+Controller_Mouse::~Controller_Mouse() = default;
+
+void Controller_Mouse::OnInit() {}
+
+void Controller_Mouse::OnRelease() {}
+
+void Controller_Mouse::OnUpdate(u8* data, std::size_t size) {
+    shared_memory.header.timestamp = CoreTiming::GetTicks();
+    shared_memory.header.total_entry_count = 17;
+
+    if (!IsControllerActivated()) {
+        shared_memory.header.entry_count = 0;
+        shared_memory.header.last_entry_index = 0;
+        return;
+    }
+    shared_memory.header.entry_count = 16;
+
+    auto& last_entry = shared_memory.mouse_states[shared_memory.header.last_entry_index];
+    shared_memory.header.last_entry_index = (shared_memory.header.last_entry_index + 1) % 17;
+    auto& cur_entry = shared_memory.mouse_states[shared_memory.header.last_entry_index];
+
+    cur_entry.sampling_number = last_entry.sampling_number + 1;
+    cur_entry.sampling_number2 = cur_entry.sampling_number;
+    // TODO(ogniK): Update mouse states
+
+    std::memcpy(data + SHARED_MEMORY_OFFSET, &shared_memory, sizeof(SharedMemory));
+}
+
+void Controller_Mouse::OnLoadInputDevices() {}
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/mouse.h b/src/core/hle/service/hid/controllers/mouse.h
new file mode 100644
index 000000000..543b0b71f
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/mouse.h
@@ -0,0 +1,50 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/hle/service/hid/controllers/controller_base.h"
+
+namespace Service::HID {
+class Controller_Mouse final : public ControllerBase {
+public:
+    Controller_Mouse();
+    ~Controller_Mouse() override;
+
+    // Called when the controller is initialized
+    void OnInit() override;
+
+    // When the controller is released
+    void OnRelease() override;
+
+    // When the controller is requesting an update for the shared memory
+    void OnUpdate(u8* data, std::size_t size) override;
+
+    // Called when input devices should be loaded
+    void OnLoadInputDevices() override;
+
+private:
+    struct MouseState {
+        s64_le sampling_number;
+        s64_le sampling_number2;
+        s32_le x;
+        s32_le y;
+        s32_le delta_x;
+        s32_le delta_y;
+        s32_le mouse_wheel;
+        s32_le button;
+        s32_le attribute;
+    };
+    static_assert(sizeof(MouseState) == 0x30, "MouseState is an invalid size");
+
+    struct SharedMemory {
+        CommonHeader header;
+        std::array<MouseState, 17> mouse_states;
+    };
+    SharedMemory shared_memory{};
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/npad.cpp b/src/core/hle/service/hid/controllers/npad.cpp
new file mode 100644
index 000000000..4b4d1324f
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/npad.cpp
@@ -0,0 +1,620 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <algorithm>
+#include <array>
+#include <cstring>
+#include "common/assert.h"
+#include "common/bit_field.h"
+#include "common/common_types.h"
+#include "common/logging/log.h"
+#include "core/core.h"
+#include "core/core_timing.h"
+#include "core/frontend/input.h"
+#include "core/hle/kernel/event.h"
+#include "core/hle/service/hid/controllers/npad.h"
+#include "core/settings.h"
+
+namespace Service::HID {
+
+constexpr u32 JOYCON_BODY_NEON_RED = 0xFF3C28;
+constexpr u32 JOYCON_BUTTONS_NEON_RED = 0x1E0A0A;
+constexpr u32 JOYCON_BODY_NEON_BLUE = 0x0AB9E6;
+constexpr u32 JOYCON_BUTTONS_NEON_BLUE = 0x001E1E;
+constexpr s32 HID_JOYSTICK_MAX = 0x7fff;
+constexpr s32 HID_JOYSTICK_MIN = -0x7fff;
+constexpr std::size_t NPAD_OFFSET = 0x9A00;
+constexpr u32 BATTERY_FULL = 2;
+constexpr u32 NPAD_HANDHELD = 32;
+constexpr u32 NPAD_UNKNOWN = 16; // TODO(ogniK): What is this?
+constexpr u32 MAX_NPAD_ID = 7;
+constexpr Controller_NPad::NPadControllerType PREFERRED_CONTROLLER =
+    Controller_NPad::NPadControllerType::JoyDual;
+constexpr std::array<u32, 10> npad_id_list{
+    0, 1, 2, 3, 4, 5, 6, 7, 32, 16,
+};
+
+enum class JoystickId : std::size_t {
+    Joystick_Left,
+    Joystick_Right,
+};
+
+Controller_NPad::Controller_NPad() = default;
+Controller_NPad::~Controller_NPad() = default;
+
+void Controller_NPad::InitNewlyAddedControler(std::size_t controller_idx) {
+    const auto controller_type = connected_controllers[controller_idx].type;
+    auto& controller = shared_memory_entries[controller_idx];
+    if (controller_type == NPadControllerType::None) {
+        return;
+    }
+    controller.joy_styles.raw = 0; // Zero out
+    controller.device_type.raw = 0;
+    switch (controller_type) {
+    case NPadControllerType::Handheld:
+        controller.joy_styles.handheld.Assign(1);
+        controller.device_type.handheld.Assign(1);
+        controller.pad_assignment = NPadAssignments::Dual;
+        break;
+    case NPadControllerType::JoyDual:
+        controller.joy_styles.joycon_dual.Assign(1);
+        controller.device_type.joycon_left.Assign(1);
+        controller.device_type.joycon_right.Assign(1);
+        controller.pad_assignment = NPadAssignments::Dual;
+        break;
+    case NPadControllerType::JoyLeft:
+        controller.joy_styles.joycon_left.Assign(1);
+        controller.device_type.joycon_left.Assign(1);
+        controller.pad_assignment = NPadAssignments::Dual;
+        break;
+    case NPadControllerType::JoyRight:
+        controller.joy_styles.joycon_right.Assign(1);
+        controller.device_type.joycon_right.Assign(1);
+        controller.pad_assignment = NPadAssignments::Dual;
+        break;
+    case NPadControllerType::Pokeball:
+        controller.joy_styles.pokeball.Assign(1);
+        controller.device_type.pokeball.Assign(1);
+        controller.pad_assignment = NPadAssignments::Single;
+        break;
+    case NPadControllerType::ProController:
+        controller.joy_styles.pro_controller.Assign(1);
+        controller.device_type.pro_controller.Assign(1);
+        controller.pad_assignment = NPadAssignments::Single;
+        break;
+    }
+
+    controller.single_color_error = ColorReadError::ReadOk;
+    controller.single_color.body_color = 0;
+    controller.single_color.button_color = 0;
+
+    controller.dual_color_error = ColorReadError::ReadOk;
+    controller.left_color.body_color = JOYCON_BODY_NEON_BLUE;
+    controller.left_color.button_color = JOYCON_BUTTONS_NEON_BLUE;
+    controller.right_color.body_color = JOYCON_BODY_NEON_RED;
+    controller.right_color.button_color = JOYCON_BUTTONS_NEON_RED;
+
+    controller.properties.is_vertical.Assign(1); // TODO(ogniK): Swap joycons orientations
+    controller.properties.use_plus.Assign(1);
+    controller.properties.use_minus.Assign(1);
+    controller.battery_level[0] = BATTERY_FULL;
+    controller.battery_level[1] = BATTERY_FULL;
+    controller.battery_level[2] = BATTERY_FULL;
+}
+
+void Controller_NPad::OnInit() {
+    auto& kernel = Core::System::GetInstance().Kernel();
+    styleset_changed_event =
+        Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "npad:NpadStyleSetChanged");
+
+    if (!IsControllerActivated()) {
+        return;
+    }
+
+    if (style.raw == 0) {
+        // We want to support all controllers
+        style.handheld.Assign(1);
+        style.joycon_left.Assign(1);
+        style.joycon_right.Assign(1);
+        style.joycon_dual.Assign(1);
+        style.pro_controller.Assign(1);
+        style.pokeball.Assign(1);
+    }
+    if (std::none_of(connected_controllers.begin(), connected_controllers.end(),
+                     [](const ControllerHolder& controller) { return controller.is_connected; })) {
+        supported_npad_id_types.resize(npad_id_list.size());
+        std::memcpy(supported_npad_id_types.data(), npad_id_list.data(),
+                    npad_id_list.size() * sizeof(u32));
+        AddNewController(PREFERRED_CONTROLLER);
+    }
+}
+
+void Controller_NPad::OnLoadInputDevices() {
+    std::transform(Settings::values.buttons.begin() + Settings::NativeButton::BUTTON_HID_BEGIN,
+                   Settings::values.buttons.begin() + Settings::NativeButton::BUTTON_HID_END,
+                   buttons.begin(), Input::CreateDevice<Input::ButtonDevice>);
+    std::transform(Settings::values.analogs.begin() + Settings::NativeAnalog::STICK_HID_BEGIN,
+                   Settings::values.analogs.begin() + Settings::NativeAnalog::STICK_HID_END,
+                   sticks.begin(), Input::CreateDevice<Input::AnalogDevice>);
+}
+
+void Controller_NPad::OnRelease() {}
+
+void Controller_NPad::OnUpdate(u8* data, std::size_t data_len) {
+    if (!IsControllerActivated())
+        return;
+    for (std::size_t i = 0; i < shared_memory_entries.size(); i++) {
+        auto& npad = shared_memory_entries[i];
+        const std::array<NPadGeneric*, 7> controller_npads{&npad.main_controller_states,
+                                                           &npad.handheld_states,
+                                                           &npad.dual_states,
+                                                           &npad.left_joy_states,
+                                                           &npad.right_joy_states,
+                                                           &npad.pokeball_states,
+                                                           &npad.libnx};
+
+        for (auto* main_controller : controller_npads) {
+            main_controller->common.entry_count = 16;
+            main_controller->common.total_entry_count = 17;
+
+            const auto& last_entry =
+                main_controller->npad[main_controller->common.last_entry_index];
+
+            main_controller->common.timestamp = CoreTiming::GetTicks();
+            main_controller->common.last_entry_index =
+                (main_controller->common.last_entry_index + 1) % 17;
+
+            auto& cur_entry = main_controller->npad[main_controller->common.last_entry_index];
+
+            cur_entry.timestamp = last_entry.timestamp + 1;
+            cur_entry.timestamp2 = cur_entry.timestamp;
+        }
+
+        const auto& controller_type = connected_controllers[i].type;
+
+        if (controller_type == NPadControllerType::None || !connected_controllers[i].is_connected) {
+            continue;
+        }
+
+        // Pad states
+        ControllerPadState pad_state{};
+        using namespace Settings::NativeButton;
+        pad_state.a.Assign(buttons[A - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.b.Assign(buttons[B - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.x.Assign(buttons[X - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.y.Assign(buttons[Y - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.l_stick.Assign(buttons[LStick - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.r_stick.Assign(buttons[RStick - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.l.Assign(buttons[L - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.r.Assign(buttons[R - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.zl.Assign(buttons[ZL - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.zr.Assign(buttons[ZR - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.plus.Assign(buttons[Plus - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.minus.Assign(buttons[Minus - BUTTON_HID_BEGIN]->GetStatus());
+
+        pad_state.d_left.Assign(buttons[DLeft - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.d_up.Assign(buttons[DUp - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.d_right.Assign(buttons[DRight - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.d_down.Assign(buttons[DDown - BUTTON_HID_BEGIN]->GetStatus());
+
+        pad_state.l_stick_left.Assign(buttons[LStick_Left - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.l_stick_up.Assign(buttons[LStick_Up - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.l_stick_right.Assign(buttons[LStick_Right - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.l_stick_down.Assign(buttons[LStick_Down - BUTTON_HID_BEGIN]->GetStatus());
+
+        pad_state.r_stick_left.Assign(buttons[RStick_Left - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.r_stick_up.Assign(buttons[RStick_Up - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.r_stick_right.Assign(buttons[RStick_Right - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.r_stick_down.Assign(buttons[RStick_Down - BUTTON_HID_BEGIN]->GetStatus());
+
+        pad_state.sl.Assign(buttons[SL - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.sr.Assign(buttons[SR - BUTTON_HID_BEGIN]->GetStatus());
+
+        AnalogPosition lstick_entry{};
+        AnalogPosition rstick_entry{};
+
+        const auto [stick_l_x_f, stick_l_y_f] =
+            sticks[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetStatus();
+        const auto [stick_r_x_f, stick_r_y_f] =
+            sticks[static_cast<std::size_t>(JoystickId::Joystick_Right)]->GetStatus();
+        lstick_entry.x = static_cast<s32>(stick_l_x_f * HID_JOYSTICK_MAX);
+        lstick_entry.y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX);
+        rstick_entry.x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX);
+        rstick_entry.y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX);
+
+        if (controller_type == NPadControllerType::JoyLeft ||
+            controller_type == NPadControllerType::JoyRight) {
+            if (npad.properties.is_horizontal) {
+                ControllerPadState state{};
+                AnalogPosition temp_lstick_entry{};
+                AnalogPosition temp_rstick_entry{};
+                if (controller_type == NPadControllerType::JoyLeft) {
+                    state.d_down.Assign(pad_state.d_left.Value());
+                    state.d_left.Assign(pad_state.d_up.Value());
+                    state.d_right.Assign(pad_state.d_down.Value());
+                    state.d_up.Assign(pad_state.d_right.Value());
+                    state.l.Assign(pad_state.l.Value() | pad_state.sl.Value());
+                    state.r.Assign(pad_state.r.Value() | pad_state.sr.Value());
+
+                    state.zl.Assign(pad_state.zl.Value());
+                    state.plus.Assign(pad_state.minus.Value());
+
+                    temp_lstick_entry = lstick_entry;
+                    temp_rstick_entry = rstick_entry;
+                    std::swap(temp_lstick_entry.x, temp_lstick_entry.y);
+                    std::swap(temp_rstick_entry.x, temp_rstick_entry.y);
+                    temp_lstick_entry.y *= -1;
+                } else if (controller_type == NPadControllerType::JoyRight) {
+                    state.x.Assign(pad_state.a.Value());
+                    state.a.Assign(pad_state.b.Value());
+                    state.b.Assign(pad_state.y.Value());
+                    state.y.Assign(pad_state.b.Value());
+
+                    state.l.Assign(pad_state.l.Value() | pad_state.sl.Value());
+                    state.r.Assign(pad_state.r.Value() | pad_state.sr.Value());
+                    state.zr.Assign(pad_state.zr.Value());
+                    state.plus.Assign(pad_state.plus.Value());
+
+                    temp_lstick_entry = lstick_entry;
+                    temp_rstick_entry = rstick_entry;
+                    std::swap(temp_lstick_entry.x, temp_lstick_entry.y);
+                    std::swap(temp_rstick_entry.x, temp_rstick_entry.y);
+                    temp_rstick_entry.x *= -1;
+                }
+                pad_state.raw = state.raw;
+                lstick_entry = temp_lstick_entry;
+                rstick_entry = temp_rstick_entry;
+            }
+        }
+
+        auto& main_controller =
+            npad.main_controller_states.npad[npad.main_controller_states.common.last_entry_index];
+        auto& handheld_entry =
+            npad.handheld_states.npad[npad.handheld_states.common.last_entry_index];
+        auto& dual_entry = npad.dual_states.npad[npad.dual_states.common.last_entry_index];
+        auto& left_entry = npad.left_joy_states.npad[npad.left_joy_states.common.last_entry_index];
+        auto& right_entry =
+            npad.right_joy_states.npad[npad.right_joy_states.common.last_entry_index];
+        auto& pokeball_entry =
+            npad.pokeball_states.npad[npad.pokeball_states.common.last_entry_index];
+        auto& libnx_entry = npad.libnx.npad[npad.libnx.common.last_entry_index];
+
+        if (hold_type == NpadHoldType::Horizontal) {
+            // TODO(ogniK): Remap buttons for different orientations
+        }
+        libnx_entry.connection_status.raw = 0;
+
+        switch (controller_type) {
+        case NPadControllerType::Handheld:
+            handheld_entry.connection_status.raw = 0;
+            handheld_entry.connection_status.IsConnected.Assign(1);
+            if (!Settings::values.use_docked_mode) {
+                handheld_entry.connection_status.IsWired.Assign(1);
+            }
+            handheld_entry.pad_states.raw = pad_state.raw;
+            handheld_entry.l_stick = lstick_entry;
+            handheld_entry.r_stick = rstick_entry;
+            break;
+        case NPadControllerType::JoyDual:
+            dual_entry.connection_status.raw = 0;
+
+            dual_entry.connection_status.IsLeftJoyConnected.Assign(1);
+            dual_entry.connection_status.IsRightJoyConnected.Assign(1);
+            dual_entry.connection_status.IsConnected.Assign(1);
+
+            libnx_entry.connection_status.IsLeftJoyConnected.Assign(1);
+            libnx_entry.connection_status.IsRightJoyConnected.Assign(1);
+            libnx_entry.connection_status.IsConnected.Assign(1);
+
+            dual_entry.pad_states.raw = pad_state.raw;
+            dual_entry.l_stick = lstick_entry;
+            dual_entry.r_stick = rstick_entry;
+        case NPadControllerType::JoyLeft:
+            left_entry.connection_status.raw = 0;
+
+            left_entry.connection_status.IsConnected.Assign(1);
+            left_entry.pad_states.raw = pad_state.raw;
+            left_entry.l_stick = lstick_entry;
+            left_entry.r_stick = rstick_entry;
+            break;
+        case NPadControllerType::JoyRight:
+            right_entry.connection_status.raw = 0;
+
+            right_entry.connection_status.IsConnected.Assign(1);
+            right_entry.pad_states.raw = pad_state.raw;
+            right_entry.l_stick = lstick_entry;
+            right_entry.r_stick = rstick_entry;
+            break;
+        case NPadControllerType::Pokeball:
+            pokeball_entry.connection_status.raw = 0;
+
+            pokeball_entry.connection_status.IsConnected.Assign(1);
+            pokeball_entry.connection_status.IsWired.Assign(1);
+
+            pokeball_entry.pad_states.raw = pad_state.raw;
+            pokeball_entry.l_stick = lstick_entry;
+            pokeball_entry.r_stick = rstick_entry;
+            break;
+        case NPadControllerType::ProController:
+            main_controller.connection_status.raw = 0;
+
+            main_controller.connection_status.IsConnected.Assign(1);
+            main_controller.connection_status.IsWired.Assign(1);
+            main_controller.pad_states.raw = pad_state.raw;
+            main_controller.l_stick = lstick_entry;
+            main_controller.r_stick = rstick_entry;
+            break;
+        }
+
+        // LibNX exclusively uses this section, so we always update it since LibNX doesn't activate
+        // any controllers.
+        libnx_entry.pad_states.raw = pad_state.raw;
+        libnx_entry.l_stick = lstick_entry;
+        libnx_entry.r_stick = rstick_entry;
+    }
+    std::memcpy(data + NPAD_OFFSET, shared_memory_entries.data(),
+                shared_memory_entries.size() * sizeof(NPadEntry));
+} // namespace Service::HID
+
+void Controller_NPad::SetSupportedStyleSet(NPadType style_set) {
+    style.raw = style_set.raw;
+}
+
+Controller_NPad::NPadType Controller_NPad::GetSupportedStyleSet() const {
+    return style;
+}
+
+void Controller_NPad::SetSupportedNPadIdTypes(u8* data, std::size_t length) {
+    ASSERT(length > 0 && (length % sizeof(u32)) == 0);
+    supported_npad_id_types.clear();
+    supported_npad_id_types.resize(length / sizeof(u32));
+    std::memcpy(supported_npad_id_types.data(), data, length);
+    for (std::size_t i = 0; i < connected_controllers.size(); i++) {
+        auto& controller = connected_controllers[i];
+        if (!controller.is_connected) {
+            continue;
+        }
+        if (!IsControllerSupported(PREFERRED_CONTROLLER)) {
+            controller.type = DecideBestController(PREFERRED_CONTROLLER);
+            InitNewlyAddedControler(i);
+        }
+    }
+}
+
+void Controller_NPad::GetSupportedNpadIdTypes(u32* data, std::size_t max_length) {
+    ASSERT(max_length < supported_npad_id_types.size());
+    std::memcpy(data, supported_npad_id_types.data(), supported_npad_id_types.size());
+}
+
+std::size_t Controller_NPad::GetSupportedNPadIdTypesSize() const {
+    return supported_npad_id_types.size();
+}
+
+void Controller_NPad::SetHoldType(NpadHoldType joy_hold_type) {
+    hold_type = joy_hold_type;
+}
+Controller_NPad::NpadHoldType Controller_NPad::GetHoldType() const {
+    return hold_type;
+}
+
+void Controller_NPad::SetNpadMode(u32 npad_id, NPadAssignments assignment_mode) {
+    ASSERT(npad_id < shared_memory_entries.size());
+    shared_memory_entries[npad_id].pad_assignment = assignment_mode;
+}
+
+void Controller_NPad::VibrateController(const std::vector<u32>& controller_ids,
+                                        const std::vector<Vibration>& vibrations) {
+    if (!can_controllers_vibrate) {
+        return;
+    }
+    for (std::size_t i = 0; i < controller_ids.size(); i++) {
+        std::size_t controller_pos = i;
+        // Handheld controller conversion
+        if (controller_pos == NPAD_HANDHELD) {
+            controller_pos = 8;
+        }
+        // Unknown controller conversion
+        if (controller_pos == NPAD_UNKNOWN) {
+            controller_pos = 9;
+        }
+        if (connected_controllers[controller_pos].is_connected) {
+            // TODO(ogniK): Vibrate the physical controller
+        }
+    }
+    LOG_WARNING(Service_HID, "(STUBBED) called");
+    last_processed_vibration = vibrations.back();
+}
+
+Kernel::SharedPtr<Kernel::Event> Controller_NPad::GetStyleSetChangedEvent() const {
+    return styleset_changed_event;
+}
+
+Controller_NPad::Vibration Controller_NPad::GetLastVibration() const {
+    return last_processed_vibration;
+}
+void Controller_NPad::AddNewController(NPadControllerType controller) {
+    if (controller == NPadControllerType::Handheld) {
+        connected_controllers[8] = {controller, true};
+        InitNewlyAddedControler(8);
+        return;
+    }
+    const auto pos =
+        std::find_if(connected_controllers.begin(), connected_controllers.end() - 2,
+                     [](const ControllerHolder& holder) { return !holder.is_connected; });
+    if (pos == connected_controllers.end() - 2) {
+        LOG_ERROR(Service_HID, "Cannot connect any more controllers!");
+        return;
+    }
+    const auto controller_id = std::distance(connected_controllers.begin(), pos);
+    connected_controllers[controller_id] = {controller, true};
+    InitNewlyAddedControler(controller_id);
+}
+
+void Controller_NPad::ConnectNPad(u32 npad_id) {
+    if (npad_id >= connected_controllers.size())
+        return;
+    connected_controllers[npad_id].is_connected = true;
+}
+
+void Controller_NPad::DisconnectNPad(u32 npad_id) {
+    if (npad_id >= connected_controllers.size())
+        return;
+    connected_controllers[npad_id].is_connected = false;
+}
+
+Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) {
+    if (npad_id == npad_id_list.back() || npad_id == npad_id_list[npad_id_list.size() - 2]) {
+        // These are controllers without led patterns
+        return LedPattern{0, 0, 0, 0};
+    }
+    switch (npad_id) {
+    case 0:
+        return LedPattern{1, 0, 0, 0};
+    case 1:
+        return LedPattern{0, 1, 0, 0};
+    case 2:
+        return LedPattern{0, 0, 1, 0};
+    case 3:
+        return LedPattern{0, 0, 0, 1};
+    case 4:
+        return LedPattern{1, 0, 0, 1};
+    case 5:
+        return LedPattern{1, 0, 1, 0};
+    case 6:
+        return LedPattern{1, 0, 1, 1};
+    case 7:
+        return LedPattern{0, 1, 1, 0};
+    default:
+        UNIMPLEMENTED_MSG("Unhandled npad_id {}", npad_id);
+        return LedPattern{0, 0, 0, 0};
+    };
+}
+void Controller_NPad::SetVibrationEnabled(bool can_vibrate) {
+    can_controllers_vibrate = can_vibrate;
+}
+
+bool Controller_NPad::IsControllerSupported(NPadControllerType controller) const {
+    const bool support_handheld =
+        std::find(supported_npad_id_types.begin(), supported_npad_id_types.end(), NPAD_HANDHELD) !=
+        supported_npad_id_types.end();
+    if (controller == NPadControllerType::Handheld) {
+        // Handheld is not even a supported type, lets stop here
+        if (!support_handheld) {
+            return false;
+        }
+        // Handheld should not be supported in docked mode
+        if (Settings::values.use_docked_mode) {
+            return false;
+        }
+
+        return true;
+    }
+    if (std::any_of(supported_npad_id_types.begin(), supported_npad_id_types.end(),
+                    [](u32 npad_id) { return npad_id <= MAX_NPAD_ID; })) {
+        switch (controller) {
+        case NPadControllerType::ProController:
+            return style.pro_controller;
+        case NPadControllerType::JoyDual:
+            return style.joycon_dual;
+        case NPadControllerType::JoyLeft:
+            return style.joycon_left;
+        case NPadControllerType::JoyRight:
+            return style.joycon_right;
+        case NPadControllerType::Pokeball:
+            return style.pokeball;
+        default:
+            return false;
+        }
+    }
+    return false;
+}
+
+Controller_NPad::NPadControllerType Controller_NPad::DecideBestController(
+    NPadControllerType priority) const {
+    if (IsControllerSupported(priority)) {
+        return priority;
+    }
+    const auto is_docked = Settings::values.use_docked_mode;
+    if (is_docked && priority == NPadControllerType::Handheld) {
+        priority = NPadControllerType::JoyDual;
+        if (IsControllerSupported(priority)) {
+            return priority;
+        }
+    }
+    std::vector<NPadControllerType> priority_list;
+    switch (priority) {
+    case NPadControllerType::ProController:
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::Handheld:
+        priority_list.push_back(NPadControllerType::JoyDual);
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::JoyDual:
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::JoyLeft:
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::JoyRight:
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::Pokeball:
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        break;
+    default:
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::JoyDual);
+    }
+
+    const auto iter = std::find_if(priority_list.begin(), priority_list.end(),
+                                   [this](auto type) { return IsControllerSupported(type); });
+    if (iter == priority_list.end()) {
+        UNIMPLEMENTED_MSG("Could not find supported controller!");
+        return priority;
+    }
+
+    return *iter;
+}
+
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/npad.h b/src/core/hle/service/hid/controllers/npad.h
new file mode 100644
index 000000000..ac86985ff
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/npad.h
@@ -0,0 +1,289 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include "common/common_types.h"
+#include "core/frontend/input.h"
+#include "core/hle/service/hid/controllers/controller_base.h"
+#include "core/settings.h"
+
+namespace Service::HID {
+
+class Controller_NPad final : public ControllerBase {
+public:
+    Controller_NPad();
+    ~Controller_NPad() override;
+
+    // Called when the controller is initialized
+    void OnInit() override;
+
+    // When the controller is released
+    void OnRelease() override;
+
+    // When the controller is requesting an update for the shared memory
+    void OnUpdate(u8* data, std::size_t size) override;
+
+    // Called when input devices should be loaded
+    void OnLoadInputDevices() override;
+
+    struct NPadType {
+        union {
+            u32_le raw{};
+
+            BitField<0, 1, u32_le> pro_controller;
+            BitField<1, 1, u32_le> handheld;
+            BitField<2, 1, u32_le> joycon_dual;
+            BitField<3, 1, u32_le> joycon_left;
+            BitField<4, 1, u32_le> joycon_right;
+
+            BitField<6, 1, u32_le> pokeball; // TODO(ogniK): Confirm when possible
+        };
+    };
+    static_assert(sizeof(NPadType) == 4, "NPadType is an invalid size");
+
+    struct Vibration {
+        f32 amp_low;
+        f32 freq_low;
+        f32 amp_high;
+        f32 freq_high;
+    };
+    static_assert(sizeof(Vibration) == 0x10, "Vibration is an invalid size");
+
+    enum class NpadHoldType : u64 {
+        Vertical = 0,
+        Horizontal = 1,
+    };
+
+    enum class NPadAssignments : u32_le {
+        Dual = 0,
+        Single = 1,
+    };
+
+    enum class NPadControllerType {
+        None,
+        ProController,
+        Handheld,
+        JoyDual,
+        JoyLeft,
+        JoyRight,
+        Pokeball,
+    };
+
+    struct LedPattern {
+        explicit LedPattern(u64 light1, u64 light2, u64 light3, u64 light4) {
+            position1.Assign(light1);
+            position1.Assign(light2);
+            position1.Assign(light3);
+            position1.Assign(light4);
+        }
+        union {
+            u64 raw{};
+            BitField<0, 1, u64> position1;
+            BitField<1, 1, u64> position2;
+            BitField<2, 1, u64> position3;
+            BitField<3, 1, u64> position4;
+        };
+    };
+
+    void SetSupportedStyleSet(NPadType style_set);
+    NPadType GetSupportedStyleSet() const;
+
+    void SetSupportedNPadIdTypes(u8* data, std::size_t length);
+    void GetSupportedNpadIdTypes(u32* data, std::size_t max_length);
+    std::size_t GetSupportedNPadIdTypesSize() const;
+
+    void SetHoldType(NpadHoldType joy_hold_type);
+    NpadHoldType GetHoldType() const;
+
+    void SetNpadMode(u32 npad_id, NPadAssignments assignment_mode);
+
+    void VibrateController(const std::vector<u32>& controller_ids,
+                           const std::vector<Vibration>& vibrations);
+
+    Kernel::SharedPtr<Kernel::Event> GetStyleSetChangedEvent() const;
+    Vibration GetLastVibration() const;
+
+    void AddNewController(NPadControllerType controller);
+
+    void ConnectNPad(u32 npad_id);
+    void DisconnectNPad(u32 npad_id);
+    LedPattern GetLedPattern(u32 npad_id);
+    void SetVibrationEnabled(bool can_vibrate);
+
+private:
+    struct CommonHeader {
+        s64_le timestamp;
+        s64_le total_entry_count;
+        s64_le last_entry_index;
+        s64_le entry_count;
+    };
+    static_assert(sizeof(CommonHeader) == 0x20, "CommonHeader is an invalid size");
+
+    struct ControllerColor {
+        u32_le body_color;
+        u32_le button_color;
+    };
+    static_assert(sizeof(ControllerColor) == 8, "ControllerColor is an invalid size");
+
+    struct ControllerPadState {
+        union {
+            u64_le raw{};
+            // Button states
+            BitField<0, 1, u64_le> a;
+            BitField<1, 1, u64_le> b;
+            BitField<2, 1, u64_le> x;
+            BitField<3, 1, u64_le> y;
+            BitField<4, 1, u64_le> l_stick;
+            BitField<5, 1, u64_le> r_stick;
+            BitField<6, 1, u64_le> l;
+            BitField<7, 1, u64_le> r;
+            BitField<8, 1, u64_le> zl;
+            BitField<9, 1, u64_le> zr;
+            BitField<10, 1, u64_le> plus;
+            BitField<11, 1, u64_le> minus;
+
+            // D-Pad
+            BitField<12, 1, u64_le> d_left;
+            BitField<13, 1, u64_le> d_up;
+            BitField<14, 1, u64_le> d_right;
+            BitField<15, 1, u64_le> d_down;
+
+            // Left JoyStick
+            BitField<16, 1, u64_le> l_stick_left;
+            BitField<17, 1, u64_le> l_stick_up;
+            BitField<18, 1, u64_le> l_stick_right;
+            BitField<19, 1, u64_le> l_stick_down;
+
+            // Right JoyStick
+            BitField<20, 1, u64_le> r_stick_left;
+            BitField<21, 1, u64_le> r_stick_up;
+            BitField<22, 1, u64_le> r_stick_right;
+            BitField<23, 1, u64_le> r_stick_down;
+
+            // Not always active?
+            BitField<24, 1, u64_le> sl;
+            BitField<25, 1, u64_le> sr;
+        };
+    };
+    static_assert(sizeof(ControllerPadState) == 8, "ControllerPadState is an invalid size");
+
+    struct AnalogPosition {
+        s32_le x;
+        s32_le y;
+    };
+    static_assert(sizeof(AnalogPosition) == 8, "AnalogPosition is an invalid size");
+
+    struct ConnectionState {
+        union {
+            u32_le raw{};
+            BitField<0, 1, u32_le> IsConnected;
+            BitField<1, 1, u32_le> IsWired;
+            BitField<2, 1, u32_le> IsLeftJoyConnected;
+            BitField<3, 1, u32_le> IsLeftJoyWired;
+            BitField<4, 1, u32_le> IsRightJoyConnected;
+            BitField<5, 1, u32_le> IsRightJoyWired;
+        };
+    };
+    static_assert(sizeof(ConnectionState) == 4, "ConnectionState is an invalid size");
+
+    struct GenericStates {
+        s64_le timestamp;
+        s64_le timestamp2;
+        ControllerPadState pad_states;
+        AnalogPosition l_stick;
+        AnalogPosition r_stick;
+        ConnectionState connection_status;
+    };
+    static_assert(sizeof(GenericStates) == 0x30, "NPadGenericStates is an invalid size");
+
+    struct NPadGeneric {
+        CommonHeader common;
+        std::array<GenericStates, 17> npad;
+    };
+    static_assert(sizeof(NPadGeneric) == 0x350, "NPadGeneric is an invalid size");
+
+    enum class ColorReadError : u32_le {
+        ReadOk = 0,
+        ColorDoesntExist = 1,
+        NoController = 2,
+    };
+
+    struct NPadProperties {
+        union {
+            s64_le raw{};
+            BitField<11, 1, s64_le> is_vertical;
+            BitField<12, 1, s64_le> is_horizontal;
+            BitField<13, 1, s64_le> use_plus;
+            BitField<14, 1, s64_le> use_minus;
+        };
+    };
+
+    struct NPadDevice {
+        union {
+            u32_le raw{};
+            BitField<0, 1, s32_le> pro_controller;
+            BitField<1, 1, s32_le> handheld;
+            BitField<2, 1, s32_le> handheld_left;
+            BitField<3, 1, s32_le> handheld_right;
+            BitField<4, 1, s32_le> joycon_left;
+            BitField<5, 1, s32_le> joycon_right;
+            BitField<6, 1, s32_le> pokeball;
+        };
+    };
+
+    struct NPadEntry {
+        NPadType joy_styles;
+        NPadAssignments pad_assignment;
+
+        ColorReadError single_color_error;
+        ControllerColor single_color;
+
+        ColorReadError dual_color_error;
+        ControllerColor left_color;
+        ControllerColor right_color;
+
+        NPadGeneric main_controller_states;
+        NPadGeneric handheld_states;
+        NPadGeneric dual_states;
+        NPadGeneric left_joy_states;
+        NPadGeneric right_joy_states;
+        NPadGeneric pokeball_states;
+        NPadGeneric libnx; // TODO(ogniK): Find out what this actually is, libnx seems to only be
+                           // relying on this for the time being
+        INSERT_PADDING_BYTES(
+            0x708 *
+            6); // TODO(ogniK): SixAxis states, require more information before implementation
+        NPadDevice device_type;
+        NPadProperties properties;
+        INSERT_PADDING_WORDS(1);
+        std::array<u32, 3> battery_level;
+        INSERT_PADDING_BYTES(0x5c);
+        INSERT_PADDING_BYTES(0xdf8);
+    };
+    static_assert(sizeof(NPadEntry) == 0x5000, "NPadEntry is an invalid size");
+
+    struct ControllerHolder {
+        Controller_NPad::NPadControllerType type;
+        bool is_connected;
+    };
+
+    NPadType style{};
+    std::array<NPadEntry, 10> shared_memory_entries{};
+    std::array<std::unique_ptr<Input::ButtonDevice>, Settings::NativeButton::NUM_BUTTONS_HID>
+        buttons;
+    std::array<std::unique_ptr<Input::AnalogDevice>, Settings::NativeAnalog::NUM_STICKS_HID> sticks;
+    std::vector<u32> supported_npad_id_types{};
+    NpadHoldType hold_type{NpadHoldType::Vertical};
+    Kernel::SharedPtr<Kernel::Event> styleset_changed_event;
+    Vibration last_processed_vibration{};
+    std::array<ControllerHolder, 10> connected_controllers{};
+    bool can_controllers_vibrate{true};
+
+    void InitNewlyAddedControler(std::size_t controller_idx);
+    bool IsControllerSupported(NPadControllerType controller) const;
+    NPadControllerType DecideBestController(NPadControllerType priority) const;
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/stubbed.cpp b/src/core/hle/service/hid/controllers/stubbed.cpp
new file mode 100644
index 000000000..02fcfadd9
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/stubbed.cpp
@@ -0,0 +1,39 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+#include "common/common_types.h"
+#include "core/core_timing.h"
+#include "core/hle/service/hid/controllers/stubbed.h"
+
+namespace Service::HID {
+
+Controller_Stubbed::Controller_Stubbed() = default;
+Controller_Stubbed::~Controller_Stubbed() = default;
+
+void Controller_Stubbed::OnInit() {}
+
+void Controller_Stubbed::OnRelease() {}
+
+void Controller_Stubbed::OnUpdate(u8* data, std::size_t size) {
+    if (!smart_update) {
+        return;
+    }
+
+    CommonHeader header{};
+    header.timestamp = CoreTiming::GetTicks();
+    header.total_entry_count = 17;
+    header.entry_count = 0;
+    header.last_entry_index = 0;
+
+    std::memcpy(data + common_offset, &header, sizeof(CommonHeader));
+}
+
+void Controller_Stubbed::OnLoadInputDevices() {}
+
+void Controller_Stubbed::SetCommonHeaderOffset(std::size_t off) {
+    common_offset = off;
+    smart_update = true;
+}
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/stubbed.h b/src/core/hle/service/hid/controllers/stubbed.h
new file mode 100644
index 000000000..4a21c643e
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/stubbed.h
@@ -0,0 +1,34 @@
+// 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/hle/service/hid/controllers/controller_base.h"
+
+namespace Service::HID {
+class Controller_Stubbed final : public ControllerBase {
+public:
+    Controller_Stubbed();
+    ~Controller_Stubbed() override;
+
+    // Called when the controller is initialized
+    void OnInit() override;
+
+    // When the controller is released
+    void OnRelease() override;
+
+    // When the controller is requesting an update for the shared memory
+    void OnUpdate(u8* data, std::size_t size) override;
+
+    // Called when input devices should be loaded
+    void OnLoadInputDevices() override;
+
+    void SetCommonHeaderOffset(std::size_t off);
+
+private:
+    bool smart_update{};
+    std::size_t common_offset{};
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/touchscreen.cpp b/src/core/hle/service/hid/controllers/touchscreen.cpp
new file mode 100644
index 000000000..43efef803
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/touchscreen.cpp
@@ -0,0 +1,65 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+#include "common/common_types.h"
+#include "core/core_timing.h"
+#include "core/frontend/emu_window.h"
+#include "core/frontend/input.h"
+#include "core/hle/service/hid/controllers/touchscreen.h"
+#include "core/settings.h"
+
+namespace Service::HID {
+constexpr std::size_t SHARED_MEMORY_OFFSET = 0x400;
+
+Controller_Touchscreen::Controller_Touchscreen() = default;
+Controller_Touchscreen::~Controller_Touchscreen() = default;
+
+void Controller_Touchscreen::OnInit() {}
+
+void Controller_Touchscreen::OnRelease() {}
+
+void Controller_Touchscreen::OnUpdate(u8* data, std::size_t size) {
+    shared_memory.header.timestamp = CoreTiming::GetTicks();
+    shared_memory.header.total_entry_count = 17;
+
+    if (!IsControllerActivated()) {
+        shared_memory.header.entry_count = 0;
+        shared_memory.header.last_entry_index = 0;
+        return;
+    }
+    shared_memory.header.entry_count = 16;
+
+    const auto& last_entry =
+        shared_memory.shared_memory_entries[shared_memory.header.last_entry_index];
+    shared_memory.header.last_entry_index = (shared_memory.header.last_entry_index + 1) % 17;
+    auto& cur_entry = shared_memory.shared_memory_entries[shared_memory.header.last_entry_index];
+
+    cur_entry.sampling_number = last_entry.sampling_number + 1;
+    cur_entry.sampling_number2 = cur_entry.sampling_number;
+
+    const auto [x, y, pressed] = touch_device->GetStatus();
+    auto& touch_entry = cur_entry.states[0];
+    if (pressed) {
+        touch_entry.x = static_cast<u16>(x * Layout::ScreenUndocked::Width);
+        touch_entry.y = static_cast<u16>(y * Layout::ScreenUndocked::Height);
+        touch_entry.diameter_x = 15;
+        touch_entry.diameter_y = 15;
+        touch_entry.rotation_angle = 0;
+        const u64 tick = CoreTiming::GetTicks();
+        touch_entry.delta_time = tick - last_touch;
+        last_touch = tick;
+        touch_entry.finger = 0;
+        cur_entry.entry_count = 1;
+    } else {
+        cur_entry.entry_count = 0;
+    }
+
+    std::memcpy(data + SHARED_MEMORY_OFFSET, &shared_memory, sizeof(TouchScreenSharedMemory));
+}
+
+void Controller_Touchscreen::OnLoadInputDevices() {
+    touch_device = Input::CreateDevice<Input::TouchDevice>(Settings::values.touch_device);
+}
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/touchscreen.h b/src/core/hle/service/hid/controllers/touchscreen.h
new file mode 100644
index 000000000..e5db6e6ba
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/touchscreen.h
@@ -0,0 +1,63 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/frontend/input.h"
+#include "core/hle/service/hid/controllers/controller_base.h"
+
+namespace Service::HID {
+class Controller_Touchscreen final : public ControllerBase {
+public:
+    Controller_Touchscreen();
+    ~Controller_Touchscreen() override;
+
+    // Called when the controller is initialized
+    void OnInit() override;
+
+    // When the controller is released
+    void OnRelease() override;
+
+    // When the controller is requesting an update for the shared memory
+    void OnUpdate(u8* data, std::size_t size) override;
+
+    // Called when input devices should be loaded
+    void OnLoadInputDevices() override;
+
+private:
+    struct TouchState {
+        u64_le delta_time;
+        u32_le attribute;
+        u32_le finger;
+        u32_le x;
+        u32_le y;
+        u32_le diameter_x;
+        u32_le diameter_y;
+        u32_le rotation_angle;
+    };
+    static_assert(sizeof(TouchState) == 0x28, "Touchstate is an invalid size");
+
+    struct TouchScreenEntry {
+        s64_le sampling_number;
+        s64_le sampling_number2;
+        s32_le entry_count;
+        std::array<TouchState, 16> states;
+    };
+    static_assert(sizeof(TouchScreenEntry) == 0x298, "TouchScreenEntry is an invalid size");
+
+    struct TouchScreenSharedMemory {
+        CommonHeader header;
+        std::array<TouchScreenEntry, 17> shared_memory_entries{};
+        INSERT_PADDING_BYTES(0x3c8);
+    };
+    static_assert(sizeof(TouchScreenSharedMemory) == 0x3000,
+                  "TouchScreenSharedMemory is an invalid size");
+    TouchScreenSharedMemory shared_memory{};
+    std::unique_ptr<Input::TouchDevice> touch_device;
+    s64_le last_touch{};
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/xpad.cpp b/src/core/hle/service/hid/controllers/xpad.cpp
new file mode 100644
index 000000000..cd397c70b
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/xpad.cpp
@@ -0,0 +1,45 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+#include "common/common_types.h"
+#include "core/core_timing.h"
+#include "core/hle/service/hid/controllers/xpad.h"
+
+namespace Service::HID {
+constexpr std::size_t SHARED_MEMORY_OFFSET = 0x3C00;
+
+Controller_XPad::Controller_XPad() = default;
+Controller_XPad::~Controller_XPad() = default;
+
+void Controller_XPad::OnInit() {}
+
+void Controller_XPad::OnRelease() {}
+
+void Controller_XPad::OnUpdate(u8* data, std::size_t size) {
+    for (auto& xpad_entry : shared_memory.shared_memory_entries) {
+        xpad_entry.header.timestamp = CoreTiming::GetTicks();
+        xpad_entry.header.total_entry_count = 17;
+
+        if (!IsControllerActivated()) {
+            xpad_entry.header.entry_count = 0;
+            xpad_entry.header.last_entry_index = 0;
+            return;
+        }
+        xpad_entry.header.entry_count = 16;
+
+        const auto& last_entry = xpad_entry.pad_states[xpad_entry.header.last_entry_index];
+        xpad_entry.header.last_entry_index = (xpad_entry.header.last_entry_index + 1) % 17;
+        auto& cur_entry = xpad_entry.pad_states[xpad_entry.header.last_entry_index];
+
+        cur_entry.sampling_number = last_entry.sampling_number + 1;
+        cur_entry.sampling_number2 = cur_entry.sampling_number;
+    }
+    // TODO(ogniK): Update xpad states
+
+    std::memcpy(data + SHARED_MEMORY_OFFSET, &shared_memory, sizeof(SharedMemory));
+}
+
+void Controller_XPad::OnLoadInputDevices() {}
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/controllers/xpad.h b/src/core/hle/service/hid/controllers/xpad.h
new file mode 100644
index 000000000..ff836989f
--- /dev/null
+++ b/src/core/hle/service/hid/controllers/xpad.h
@@ -0,0 +1,60 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/hle/service/hid/controllers/controller_base.h"
+
+namespace Service::HID {
+class Controller_XPad final : public ControllerBase {
+public:
+    Controller_XPad();
+    ~Controller_XPad() override;
+
+    // Called when the controller is initialized
+    void OnInit() override;
+
+    // When the controller is released
+    void OnRelease() override;
+
+    // When the controller is requesting an update for the shared memory
+    void OnUpdate(u8* data, std::size_t size) override;
+
+    // Called when input devices should be loaded
+    void OnLoadInputDevices() override;
+
+private:
+    struct AnalogStick {
+        s32_le x;
+        s32_le y;
+    };
+    static_assert(sizeof(AnalogStick) == 0x8, "AnalogStick is an invalid size");
+
+    struct XPadState {
+        s64_le sampling_number;
+        s64_le sampling_number2;
+        s32_le attributes;
+        u32_le pad_states;
+        AnalogStick x_stick;
+        AnalogStick y_stick;
+    };
+    static_assert(sizeof(XPadState) == 0x28, "XPadState is an invalid size");
+
+    struct XPadEntry {
+        CommonHeader header;
+        std::array<XPadState, 17> pad_states{};
+        INSERT_PADDING_BYTES(0x138);
+    };
+    static_assert(sizeof(XPadEntry) == 0x400, "XPadEntry is an invalid size");
+
+    struct SharedMemory {
+        std::array<XPadEntry, 4> shared_memory_entries{};
+    };
+    static_assert(sizeof(SharedMemory) == 0x1000, "SharedMemory is an invalid size");
+    SharedMemory shared_memory{};
+};
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp
index 7c6b0a4e6..a9aa9ec78 100644
--- a/src/core/hle/service/hid/hid.cpp
+++ b/src/core/hle/service/hid/hid.cpp
@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <array>
+#include "common/common_types.h"
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "core/core_timing.h"
@@ -19,6 +21,16 @@
 #include "core/hle/service/service.h"
 #include "core/settings.h"
 
+#include "core/hle/service/hid/controllers/controller_base.h"
+#include "core/hle/service/hid/controllers/debug_pad.h"
+#include "core/hle/service/hid/controllers/gesture.h"
+#include "core/hle/service/hid/controllers/keyboard.h"
+#include "core/hle/service/hid/controllers/mouse.h"
+#include "core/hle/service/hid/controllers/npad.h"
+#include "core/hle/service/hid/controllers/stubbed.h"
+#include "core/hle/service/hid/controllers/touchscreen.h"
+#include "core/hle/service/hid/controllers/xpad.h"
+
 namespace Service::HID {
 
 // Updating period for each HID device.
@@ -26,6 +38,22 @@ namespace Service::HID {
 constexpr u64 pad_update_ticks = CoreTiming::BASE_CLOCK_RATE / 100;
 constexpr u64 accelerometer_update_ticks = CoreTiming::BASE_CLOCK_RATE / 100;
 constexpr u64 gyroscope_update_ticks = CoreTiming::BASE_CLOCK_RATE / 100;
+constexpr std::size_t SHARED_MEMORY_SIZE = 0x40000;
+enum class HidController : std::size_t {
+    DebugPad,
+    Touchscreen,
+    Mouse,
+    Keyboard,
+    XPad,
+    Unknown1,
+    Unknown2,
+    Unknown3,
+    SixAxisSensor,
+    NPad,
+    Gesture,
+
+    MaxControllers,
+};
 
 class IAppletResource final : public ServiceFramework<IAppletResource> {
 public:
@@ -37,19 +65,57 @@ public:
 
         auto& kernel = Core::System::GetInstance().Kernel();
         shared_mem = Kernel::SharedMemory::Create(
-            kernel, nullptr, 0x40000, Kernel::MemoryPermission::ReadWrite,
+            kernel, nullptr, SHARED_MEMORY_SIZE, Kernel::MemoryPermission::ReadWrite,
             Kernel::MemoryPermission::Read, 0, Kernel::MemoryRegion::BASE, "HID:SharedMemory");
 
+        MakeController<Controller_DebugPad>(HidController::DebugPad);
+        MakeController<Controller_Touchscreen>(HidController::Touchscreen);
+        MakeController<Controller_Mouse>(HidController::Mouse);
+        MakeController<Controller_Keyboard>(HidController::Keyboard);
+        MakeController<Controller_XPad>(HidController::XPad);
+        MakeController<Controller_Stubbed>(HidController::Unknown1);
+        MakeController<Controller_Stubbed>(HidController::Unknown2);
+        MakeController<Controller_Stubbed>(HidController::Unknown3);
+        MakeController<Controller_Stubbed>(HidController::SixAxisSensor);
+        MakeController<Controller_NPad>(HidController::NPad);
+        MakeController<Controller_Gesture>(HidController::Gesture);
+
+        // Homebrew doesn't try to activate some controllers, so we activate them by default
+        GetController<Controller_NPad>(HidController::NPad).ActivateController();
+        GetController<Controller_Touchscreen>(HidController::Touchscreen).ActivateController();
+
+        GetController<Controller_Stubbed>(HidController::Unknown1).SetCommonHeaderOffset(0x4c00);
+        GetController<Controller_Stubbed>(HidController::Unknown2).SetCommonHeaderOffset(0x4e00);
+        GetController<Controller_Stubbed>(HidController::Unknown3).SetCommonHeaderOffset(0x5000);
+
         // Register update callbacks
         pad_update_event = CoreTiming::RegisterEvent(
             "HID::UpdatePadCallback",
-            [this](u64 userdata, int cycles_late) { UpdatePadCallback(userdata, cycles_late); });
+            [this](u64 userdata, int cycles_late) { UpdateControllers(userdata, cycles_late); });
 
         // TODO(shinyquagsire23): Other update callbacks? (accel, gyro?)
 
         CoreTiming::ScheduleEvent(pad_update_ticks, pad_update_event);
     }
 
+    void ActivateController(HidController controller) {
+        controllers[static_cast<size_t>(controller)]->ActivateController();
+    }
+
+    void DeactivateController(HidController controller) {
+        controllers[static_cast<size_t>(controller)]->DeactivateController();
+    }
+
+    template <typename T>
+    void MakeController(HidController controller) {
+        controllers[static_cast<std::size_t>(controller)] = std::make_unique<T>();
+    }
+
+    template <typename T>
+    T& GetController(HidController controller) {
+        return static_cast<T&>(*controllers[static_cast<size_t>(controller)]);
+    }
+
     ~IAppletResource() {
         CoreTiming::UnscheduleEvent(pad_update_event, 0);
     }
@@ -62,200 +128,15 @@ private:
         LOG_DEBUG(Service_HID, "called");
     }
 
-    void LoadInputDevices() {
-        std::transform(Settings::values.buttons.begin() + Settings::NativeButton::BUTTON_HID_BEGIN,
-                       Settings::values.buttons.begin() + Settings::NativeButton::BUTTON_HID_END,
-                       buttons.begin(), Input::CreateDevice<Input::ButtonDevice>);
-        std::transform(Settings::values.analogs.begin() + Settings::NativeAnalog::STICK_HID_BEGIN,
-                       Settings::values.analogs.begin() + Settings::NativeAnalog::STICK_HID_END,
-                       sticks.begin(), Input::CreateDevice<Input::AnalogDevice>);
-        touch_device = Input::CreateDevice<Input::TouchDevice>(Settings::values.touch_device);
-        // TODO(shinyquagsire23): gyro, mouse, keyboard
-    }
-
-    void UpdatePadCallback(u64 userdata, int cycles_late) {
-        SharedMemory mem{};
-        std::memcpy(&mem, shared_mem->GetPointer(), sizeof(SharedMemory));
-
-        if (Settings::values.is_device_reload_pending.exchange(false))
-            LoadInputDevices();
-
-        // Set up controllers as neon red+blue Joy-Con attached to console
-        ControllerHeader& controller_header = mem.controllers[Controller_Handheld].header;
-        controller_header.type = ControllerType_Handheld;
-        controller_header.single_colors_descriptor = ColorDesc_ColorsNonexistent;
-        controller_header.right_color_body = JOYCON_BODY_NEON_RED;
-        controller_header.right_color_buttons = JOYCON_BUTTONS_NEON_RED;
-        controller_header.left_color_body = JOYCON_BODY_NEON_BLUE;
-        controller_header.left_color_buttons = JOYCON_BUTTONS_NEON_BLUE;
-
-        for (std::size_t controller = 0; controller < mem.controllers.size(); controller++) {
-            for (auto& layout : mem.controllers[controller].layouts) {
-                layout.header.num_entries = HID_NUM_ENTRIES;
-                layout.header.max_entry_index = HID_NUM_ENTRIES - 1;
-
-                // HID shared memory stores the state of the past 17 samples in a circlular buffer,
-                // each with a timestamp in number of samples since boot.
-                const ControllerInputEntry& last_entry = layout.entries[layout.header.latest_entry];
-
-                layout.header.timestamp_ticks = CoreTiming::GetTicks();
-                layout.header.latest_entry = (layout.header.latest_entry + 1) % HID_NUM_ENTRIES;
-
-                ControllerInputEntry& entry = layout.entries[layout.header.latest_entry];
-                entry.timestamp = last_entry.timestamp + 1;
-                // TODO(shinyquagsire23): Is this always identical to timestamp?
-                entry.timestamp_2 = entry.timestamp;
-
-                // TODO(shinyquagsire23): More than just handheld input
-                if (controller != Controller_Handheld)
-                    continue;
-
-                entry.connection_state = ConnectionState_Connected | ConnectionState_Wired;
-
-                // TODO(shinyquagsire23): Set up some LUTs for each layout mapping in the future?
-                // For now everything is just the default handheld layout, but split Joy-Con will
-                // rotate the face buttons and directions for certain layouts.
-                ControllerPadState& state = entry.buttons;
-                using namespace Settings::NativeButton;
-                state.a.Assign(buttons[A - BUTTON_HID_BEGIN]->GetStatus());
-                state.b.Assign(buttons[B - BUTTON_HID_BEGIN]->GetStatus());
-                state.x.Assign(buttons[X - BUTTON_HID_BEGIN]->GetStatus());
-                state.y.Assign(buttons[Y - BUTTON_HID_BEGIN]->GetStatus());
-                state.lstick.Assign(buttons[LStick - BUTTON_HID_BEGIN]->GetStatus());
-                state.rstick.Assign(buttons[RStick - BUTTON_HID_BEGIN]->GetStatus());
-                state.l.Assign(buttons[L - BUTTON_HID_BEGIN]->GetStatus());
-                state.r.Assign(buttons[R - BUTTON_HID_BEGIN]->GetStatus());
-                state.zl.Assign(buttons[ZL - BUTTON_HID_BEGIN]->GetStatus());
-                state.zr.Assign(buttons[ZR - BUTTON_HID_BEGIN]->GetStatus());
-                state.plus.Assign(buttons[Plus - BUTTON_HID_BEGIN]->GetStatus());
-                state.minus.Assign(buttons[Minus - BUTTON_HID_BEGIN]->GetStatus());
-
-                state.dleft.Assign(buttons[DLeft - BUTTON_HID_BEGIN]->GetStatus());
-                state.dup.Assign(buttons[DUp - BUTTON_HID_BEGIN]->GetStatus());
-                state.dright.Assign(buttons[DRight - BUTTON_HID_BEGIN]->GetStatus());
-                state.ddown.Assign(buttons[DDown - BUTTON_HID_BEGIN]->GetStatus());
-
-                state.lstick_left.Assign(buttons[LStick_Left - BUTTON_HID_BEGIN]->GetStatus());
-                state.lstick_up.Assign(buttons[LStick_Up - BUTTON_HID_BEGIN]->GetStatus());
-                state.lstick_right.Assign(buttons[LStick_Right - BUTTON_HID_BEGIN]->GetStatus());
-                state.lstick_down.Assign(buttons[LStick_Down - BUTTON_HID_BEGIN]->GetStatus());
-
-                state.rstick_left.Assign(buttons[RStick_Left - BUTTON_HID_BEGIN]->GetStatus());
-                state.rstick_up.Assign(buttons[RStick_Up - BUTTON_HID_BEGIN]->GetStatus());
-                state.rstick_right.Assign(buttons[RStick_Right - BUTTON_HID_BEGIN]->GetStatus());
-                state.rstick_down.Assign(buttons[RStick_Down - BUTTON_HID_BEGIN]->GetStatus());
-
-                state.sl.Assign(buttons[SL - BUTTON_HID_BEGIN]->GetStatus());
-                state.sr.Assign(buttons[SR - BUTTON_HID_BEGIN]->GetStatus());
-
-                const auto [stick_l_x_f, stick_l_y_f] = sticks[Joystick_Left]->GetStatus();
-                const auto [stick_r_x_f, stick_r_y_f] = sticks[Joystick_Right]->GetStatus();
-                entry.joystick_left_x = static_cast<s32>(stick_l_x_f * HID_JOYSTICK_MAX);
-                entry.joystick_left_y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX);
-                entry.joystick_right_x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX);
-                entry.joystick_right_y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX);
+    void UpdateControllers(u64 userdata, int cycles_late) {
+        const bool should_reload = Settings::values.is_device_reload_pending.exchange(false);
+        for (const auto& controller : controllers) {
+            if (should_reload) {
+                controller->OnLoadInputDevices();
             }
+            controller->OnUpdate(shared_mem->GetPointer(), SHARED_MEMORY_SIZE);
         }
 
-        TouchScreen& touchscreen = mem.touchscreen;
-        const u64 last_entry = touchscreen.header.latest_entry;
-        const u64 curr_entry = (last_entry + 1) % touchscreen.entries.size();
-        const u64 timestamp = CoreTiming::GetTicks();
-        const u64 sample_counter = touchscreen.entries[last_entry].header.timestamp + 1;
-        touchscreen.header.timestamp_ticks = timestamp;
-        touchscreen.header.num_entries = touchscreen.entries.size();
-        touchscreen.header.latest_entry = curr_entry;
-        touchscreen.header.max_entry_index = touchscreen.entries.size();
-        touchscreen.header.timestamp = timestamp;
-        touchscreen.entries[curr_entry].header.timestamp = sample_counter;
-
-        TouchScreenEntryTouch touch_entry{};
-        auto [x, y, pressed] = touch_device->GetStatus();
-        touch_entry.timestamp = timestamp;
-        touch_entry.x = static_cast<u16>(x * Layout::ScreenUndocked::Width);
-        touch_entry.y = static_cast<u16>(y * Layout::ScreenUndocked::Height);
-        touch_entry.touch_index = 0;
-
-        // TODO(DarkLordZach): Maybe try to derive these from EmuWindow?
-        touch_entry.diameter_x = 15;
-        touch_entry.diameter_y = 15;
-        touch_entry.angle = 0;
-
-        // TODO(DarkLordZach): Implement multi-touch support
-        if (pressed) {
-            touchscreen.entries[curr_entry].header.num_touches = 1;
-            touchscreen.entries[curr_entry].touches[0] = touch_entry;
-        } else {
-            touchscreen.entries[curr_entry].header.num_touches = 0;
-        }
-
-        // TODO(shinyquagsire23): Properly implement mouse
-        Mouse& mouse = mem.mouse;
-        const u64 last_mouse_entry = mouse.header.latest_entry;
-        const u64 curr_mouse_entry = (mouse.header.latest_entry + 1) % mouse.entries.size();
-        const u64 mouse_sample_counter = mouse.entries[last_mouse_entry].timestamp + 1;
-        mouse.header.timestamp_ticks = timestamp;
-        mouse.header.num_entries = mouse.entries.size();
-        mouse.header.max_entry_index = mouse.entries.size();
-        mouse.header.latest_entry = curr_mouse_entry;
-
-        mouse.entries[curr_mouse_entry].timestamp = mouse_sample_counter;
-        mouse.entries[curr_mouse_entry].timestamp_2 = mouse_sample_counter;
-
-        // TODO(shinyquagsire23): Properly implement keyboard
-        Keyboard& keyboard = mem.keyboard;
-        const u64 last_keyboard_entry = keyboard.header.latest_entry;
-        const u64 curr_keyboard_entry =
-            (keyboard.header.latest_entry + 1) % keyboard.entries.size();
-        const u64 keyboard_sample_counter = keyboard.entries[last_keyboard_entry].timestamp + 1;
-        keyboard.header.timestamp_ticks = timestamp;
-        keyboard.header.num_entries = keyboard.entries.size();
-        keyboard.header.latest_entry = last_keyboard_entry;
-        keyboard.header.max_entry_index = keyboard.entries.size();
-
-        keyboard.entries[curr_keyboard_entry].timestamp = keyboard_sample_counter;
-        keyboard.entries[curr_keyboard_entry].timestamp_2 = keyboard_sample_counter;
-
-        // TODO(shinyquagsire23): Figure out what any of these are
-        for (auto& input : mem.unk_input_1) {
-            const u64 last_input_entry = input.header.latest_entry;
-            const u64 curr_input_entry = (input.header.latest_entry + 1) % input.entries.size();
-            const u64 input_sample_counter = input.entries[last_input_entry].timestamp + 1;
-
-            input.header.timestamp_ticks = timestamp;
-            input.header.num_entries = input.entries.size();
-            input.header.latest_entry = last_input_entry;
-            input.header.max_entry_index = input.entries.size();
-
-            input.entries[curr_input_entry].timestamp = input_sample_counter;
-            input.entries[curr_input_entry].timestamp_2 = input_sample_counter;
-        }
-
-        for (auto& input : mem.unk_input_2) {
-            input.header.timestamp_ticks = timestamp;
-            input.header.num_entries = 17;
-            input.header.latest_entry = 0;
-            input.header.max_entry_index = 0;
-        }
-
-        UnkInput3& input = mem.unk_input_3;
-        const u64 last_input_entry = input.header.latest_entry;
-        const u64 curr_input_entry = (input.header.latest_entry + 1) % input.entries.size();
-        const u64 input_sample_counter = input.entries[last_input_entry].timestamp + 1;
-
-        input.header.timestamp_ticks = timestamp;
-        input.header.num_entries = input.entries.size();
-        input.header.latest_entry = last_input_entry;
-        input.header.max_entry_index = input.entries.size();
-
-        input.entries[curr_input_entry].timestamp = input_sample_counter;
-        input.entries[curr_input_entry].timestamp_2 = input_sample_counter;
-
-        // TODO(shinyquagsire23): Signal events
-
-        std::memcpy(shared_mem->GetPointer(), &mem, sizeof(SharedMemory));
-
-        // Reschedule recurrent event
         CoreTiming::ScheduleEvent(pad_update_ticks - cycles_late, pad_update_event);
     }
 
@@ -265,11 +146,8 @@ private:
     // CoreTiming update events
     CoreTiming::EventType* pad_update_event;
 
-    // Stored input state info
-    std::array<std::unique_ptr<Input::ButtonDevice>, Settings::NativeButton::NUM_BUTTONS_HID>
-        buttons;
-    std::array<std::unique_ptr<Input::AnalogDevice>, Settings::NativeAnalog::NUM_STICKS_HID> sticks;
-    std::unique_ptr<Input::TouchDevice> touch_device;
+    std::array<std::unique_ptr<ControllerBase>, static_cast<size_t>(HidController::MaxControllers)>
+        controllers{};
 };
 
 class IActiveVibrationDeviceList final : public ServiceFramework<IActiveVibrationDeviceList> {
@@ -299,9 +177,10 @@ public:
             {11, &Hid::ActivateTouchScreen, "ActivateTouchScreen"},
             {21, &Hid::ActivateMouse, "ActivateMouse"},
             {31, &Hid::ActivateKeyboard, "ActivateKeyboard"},
+            {32, nullptr, "SendKeyboardLockKeyEvent"},
             {40, nullptr, "AcquireXpadIdEventHandle"},
             {41, nullptr, "ReleaseXpadIdEventHandle"},
-            {51, nullptr, "ActivateXpad"},
+            {51, &Hid::ActivateXpad, "ActivateXpad"},
             {55, nullptr, "GetXpadIds"},
             {56, nullptr, "ActivateJoyXpad"},
             {58, nullptr, "GetJoyXpadLifoHandle"},
@@ -329,6 +208,7 @@ public:
             {80, nullptr, "GetGyroscopeZeroDriftMode"},
             {81, nullptr, "ResetGyroscopeZeroDriftMode"},
             {82, &Hid::IsSixAxisSensorAtRest, "IsSixAxisSensorAtRest"},
+            {83, nullptr, "IsFirmwareUpdateAvailableForSixAxisSensor"},
             {91, &Hid::ActivateGesture, "ActivateGesture"},
             {100, &Hid::SetSupportedNpadStyleSet, "SetSupportedNpadStyleSet"},
             {101, &Hid::GetSupportedNpadStyleSet, "GetSupportedNpadStyleSet"},
@@ -362,8 +242,8 @@ public:
             {206, &Hid::SendVibrationValues, "SendVibrationValues"},
             {207, nullptr, "SendVibrationGcErmCommand"},
             {208, nullptr, "GetActualVibrationGcErmCommand"},
-            {209, nullptr, "BeginPermitVibrationSession"},
-            {210, nullptr, "EndPermitVibrationSession"},
+            {209, &Hid::BeginPermitVibrationSession, "BeginPermitVibrationSession"},
+            {210, &Hid::EndPermitVibrationSession, "EndPermitVibrationSession"},
             {300, &Hid::ActivateConsoleSixAxisSensor, "ActivateConsoleSixAxisSensor"},
             {301, &Hid::StartConsoleSixAxisSensor, "StartConsoleSixAxisSensor"},
             {302, nullptr, "StopConsoleSixAxisSensor"},
@@ -374,6 +254,7 @@ public:
             {307, nullptr, "FinalizeSevenSixAxisSensor"},
             {308, nullptr, "SetSevenSixAxisSensorFusionStrength"},
             {309, nullptr, "GetSevenSixAxisSensorFusionStrength"},
+            {310, nullptr, "ResetSevenSixAxisSensorTimestamp"},
             {400, nullptr, "IsUsbFullKeyControllerEnabled"},
             {401, nullptr, "EnableUsbFullKeyController"},
             {402, nullptr, "IsUsbFullKeyControllerConnected"},
@@ -389,28 +270,35 @@ public:
             {505, nullptr, "SetPalmaFrModeType"},
             {506, nullptr, "ReadPalmaStep"},
             {507, nullptr, "EnablePalmaStep"},
-            {508, nullptr, "SuspendPalmaStep"},
-            {509, nullptr, "ResetPalmaStep"},
-            {510, nullptr, "ReadPalmaApplicationSection"},
-            {511, nullptr, "WritePalmaApplicationSection"},
-            {512, nullptr, "ReadPalmaUniqueCode"},
-            {513, nullptr, "SetPalmaUniqueCodeInvalid"},
+            {508, nullptr, "ResetPalmaStep"},
+            {509, nullptr, "ReadPalmaApplicationSection"},
+            {510, nullptr, "WritePalmaApplicationSection"},
+            {511, nullptr, "ReadPalmaUniqueCode"},
+            {512, nullptr, "SetPalmaUniqueCodeInvalid"},
+            {513, nullptr, "WritePalmaActivityEntry"},
+            {514, nullptr, "WritePalmaRgbLedPatternEntry"},
+            {515, nullptr, "WritePalmaWaveEntry"},
+            {516, nullptr, "SetPalmaDataBaseIdentificationVersion"},
+            {517, nullptr, "GetPalmaDataBaseIdentificationVersion"},
+            {518, nullptr, "SuspendPalmaFeature"},
+            {519, nullptr, "GetPalmaOperationResult"},
+            {520, nullptr, "ReadPalmaPlayLog"},
+            {521, nullptr, "ResetPalmaPlayLog"},
+            {522, nullptr, "SetIsPalmaAllConnectable"},
+            {523, nullptr, "SetIsPalmaPairedConnectable"},
+            {524, nullptr, "PairPalma"},
+            {525, nullptr, "SetPalmaBoostMode"},
             {1000, nullptr, "SetNpadCommunicationMode"},
             {1001, nullptr, "GetNpadCommunicationMode"},
         };
         // clang-format on
 
         RegisterHandlers(functions);
-
-        auto& kernel = Core::System::GetInstance().Kernel();
-        event = Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "hid:EventHandle");
     }
     ~Hid() = default;
 
 private:
     std::shared_ptr<IAppletResource> applet_resource;
-    u32 joy_hold_type{0};
-    Kernel::SharedPtr<Kernel::Event> event;
 
     void CreateAppletResource(Kernel::HLERequestContext& ctx) {
         if (applet_resource == nullptr) {
@@ -423,31 +311,59 @@ private:
         LOG_DEBUG(Service_HID, "called");
     }
 
+    void ActivateXpad(Kernel::HLERequestContext& ctx) {
+        applet_resource->ActivateController(HidController::XPad);
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_DEBUG(Service_HID, "called");
+    }
+
     void ActivateDebugPad(Kernel::HLERequestContext& ctx) {
+        applet_resource->ActivateController(HidController::DebugPad);
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void ActivateTouchScreen(Kernel::HLERequestContext& ctx) {
+        applet_resource->ActivateController(HidController::Touchscreen);
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void ActivateMouse(Kernel::HLERequestContext& ctx) {
+        applet_resource->ActivateController(HidController::Mouse);
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void ActivateKeyboard(Kernel::HLERequestContext& ctx) {
+        applet_resource->ActivateController(HidController::Keyboard);
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
+    }
+
+    void ActivateGesture(Kernel::HLERequestContext& ctx) {
+        applet_resource->ActivateController(HidController::Gesture);
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_DEBUG(Service_HID, "called");
+    }
+
+    void ActivateNpadWithRevision(Kernel::HLERequestContext& ctx) {
+        // Should have no effect with how our npad sets up the data
+        applet_resource->ActivateController(HidController::NPad);
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void StartSixAxisSensor(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        auto handle = rp.PopRaw<u32>();
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
         LOG_WARNING(Service_HID, "(STUBBED) called");
@@ -468,84 +384,168 @@ private:
     }
 
     void SetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        auto supported_styleset = rp.PopRaw<u32>();
+        applet_resource->GetController<Controller_NPad>(HidController::NPad)
+            .SetSupportedStyleSet({supported_styleset});
+
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void GetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx) {
+        auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(0);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        rb.Push<u32>(controller.GetSupportedStyleSet().raw);
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void SetSupportedNpadIdType(Kernel::HLERequestContext& ctx) {
+        applet_resource->GetController<Controller_NPad>(HidController::NPad)
+            .SetSupportedNPadIdTypes(ctx.ReadBuffer().data(), ctx.GetReadBufferSize());
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void ActivateNpad(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        applet_resource->ActivateController(HidController::NPad);
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void AcquireNpadStyleSetUpdateEventHandle(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        auto npad_id = rp.PopRaw<u32>();
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushCopyObjects(event);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        rb.PushCopyObjects(applet_resource->GetController<Controller_NPad>(HidController::NPad)
+                               .GetStyleSetChangedEvent());
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void DisconnectNpad(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        auto npad_id = rp.PopRaw<u32>();
+        applet_resource->GetController<Controller_NPad>(HidController::NPad)
+            .DisconnectNPad(npad_id);
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void GetPlayerLedPattern(Kernel::HLERequestContext& ctx) {
-        IPC::ResponseBuilder rb{ctx, 2};
+        IPC::RequestParser rp{ctx};
+        auto npad_id = rp.PopRaw<u32>();
+        IPC::ResponseBuilder rb{ctx, 4};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        rb.PushRaw<u64>(applet_resource->GetController<Controller_NPad>(HidController::NPad)
+                            .GetLedPattern(npad_id)
+                            .raw);
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void SetNpadJoyHoldType(Kernel::HLERequestContext& ctx) {
+        auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+        IPC::RequestParser rp{ctx};
+        const auto hold_type = rp.PopRaw<u64>();
+        controller.SetHoldType(Controller_NPad::NpadHoldType{hold_type});
+
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void GetNpadJoyHoldType(Kernel::HLERequestContext& ctx) {
-        IPC::ResponseBuilder rb{ctx, 3};
+        const auto& controller =
+            applet_resource->GetController<Controller_NPad>(HidController::NPad);
+        IPC::ResponseBuilder rb{ctx, 4};
         rb.Push(RESULT_SUCCESS);
-        rb.Push(joy_hold_type);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        rb.Push<u64>(static_cast<u64>(controller.GetHoldType()));
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        auto npad_id = rp.PopRaw<u32>();
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
         LOG_WARNING(Service_HID, "(STUBBED) called");
     }
 
+    void BeginPermitVibrationSession(Kernel::HLERequestContext& ctx) {
+        applet_resource->GetController<Controller_NPad>(HidController::NPad)
+            .SetVibrationEnabled(true);
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_DEBUG(Service_HID, "called");
+    }
+
+    void EndPermitVibrationSession(Kernel::HLERequestContext& ctx) {
+        applet_resource->GetController<Controller_NPad>(HidController::NPad)
+            .SetVibrationEnabled(false);
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_DEBUG(Service_HID, "called");
+    }
+
     void SendVibrationValue(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto controller_id = rp.PopRaw<u32>();
+        const auto vibration_values = rp.PopRaw<Controller_NPad::Vibration>();
+
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+
+        applet_resource->GetController<Controller_NPad>(HidController::NPad)
+            .VibrateController({controller_id}, {vibration_values});
+        LOG_DEBUG(Service_HID, "called");
     }
 
-    void GetActualVibrationValue(Kernel::HLERequestContext& ctx) {
+    void SendVibrationValues(Kernel::HLERequestContext& ctx) {
+        const auto controllers = ctx.ReadBuffer(0);
+        const auto vibrations = ctx.ReadBuffer(1);
+
+        std::vector<u32> controller_list(controllers.size() / sizeof(u32));
+        std::vector<Controller_NPad::Vibration> vibration_list(vibrations.size() /
+                                                               sizeof(Controller_NPad::Vibration));
+
+        std::memcpy(controller_list.data(), controllers.data(), controllers.size());
+        std::memcpy(vibration_list.data(), vibrations.data(), vibrations.size());
+        std::transform(controller_list.begin(), controller_list.end(), controller_list.begin(),
+                       [](u32 controller_id) { return controller_id - 3; });
+
+        applet_resource->GetController<Controller_NPad>(HidController::NPad)
+            .VibrateController(controller_list, vibration_list);
+
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
+    }
+
+    void GetActualVibrationValue(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 6};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<Controller_NPad::Vibration>(
+            applet_resource->GetController<Controller_NPad>(HidController::NPad)
+                .GetLastVibration());
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void SetNpadJoyAssignmentModeDual(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto npad_id = rp.PopRaw<u32>();
+        auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+        controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Dual);
+
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void MergeSingleJoyAsDualJoy(Kernel::HLERequestContext& ctx) {
@@ -555,6 +555,8 @@ private:
     }
 
     void SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        auto mode = rp.PopRaw<u32>();
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
         LOG_WARNING(Service_HID, "(STUBBED) called");
@@ -563,8 +565,9 @@ private:
     void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 4};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u64>(0);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
+        rb.Push<u32>(1);
+        rb.Push<u32>(0);
+        LOG_DEBUG(Service_HID, "called");
     }
 
     void CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx) {
@@ -574,12 +577,6 @@ private:
         LOG_DEBUG(Service_HID, "called");
     }
 
-    void SendVibrationValues(Kernel::HLERequestContext& ctx) {
-        IPC::ResponseBuilder rb{ctx, 2};
-        rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
-    }
-
     void ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
@@ -597,18 +594,6 @@ private:
         rb.Push(RESULT_SUCCESS);
         LOG_WARNING(Service_HID, "(STUBBED) called");
     }
-
-    void ActivateGesture(Kernel::HLERequestContext& ctx) {
-        IPC::ResponseBuilder rb{ctx, 2};
-        rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
-    }
-
-    void ActivateNpadWithRevision(Kernel::HLERequestContext& ctx) {
-        IPC::ResponseBuilder rb{ctx, 2};
-        rb.Push(RESULT_SUCCESS);
-        LOG_WARNING(Service_HID, "(STUBBED) called");
-    }
 };
 
 class HidDbg final : public ServiceFramework<HidDbg> {
@@ -650,6 +635,7 @@ public:
             {140, nullptr, "DeactivateConsoleSixAxisSensor"},
             {141, nullptr, "GetConsoleSixAxisSensorSamplingFrequency"},
             {142, nullptr, "DeactivateSevenSixAxisSensor"},
+            {143, nullptr, "GetConsoleSixAxisSensorCountStates"},
             {201, nullptr, "ActivateFirmwareUpdate"},
             {202, nullptr, "DeactivateFirmwareUpdate"},
             {203, nullptr, "StartFirmwareUpdate"},
@@ -660,12 +646,23 @@ public:
             {208, nullptr, "StartFirmwareUpdateForRevert"},
             {209, nullptr, "GetAvailableFirmwareVersionForRevert"},
             {210, nullptr, "IsFirmwareUpdatingDevice"},
+            {211, nullptr, "StartFirmwareUpdateIndividual"},
+            {215, nullptr, "SetUsbFirmwareForceUpdateEnabled"},
+            {216, nullptr, "SetAllKuinaDevicesToFirmwareUpdateMode"},
             {221, nullptr, "UpdateControllerColor"},
             {222, nullptr, "ConnectUsbPadsAsync"},
             {223, nullptr, "DisconnectUsbPadsAsync"},
             {224, nullptr, "UpdateDesignInfo"},
             {225, nullptr, "GetUniquePadDriverState"},
             {226, nullptr, "GetSixAxisSensorDriverStates"},
+            {227, nullptr, "GetRxPacketHistory"},
+            {228, nullptr, "AcquireOperationEventHandle"},
+            {229, nullptr, "ReadSerialFlash"},
+            {230, nullptr, "WriteSerialFlash"},
+            {231, nullptr, "GetOperationResult"},
+            {232, nullptr, "EnableShipmentMode"},
+            {233, nullptr, "ClearPairingInfo"},
+            {234, nullptr, "GetUniquePadDeviceTypeSetInternal"},
             {301, nullptr, "GetAbstractedPadHandles"},
             {302, nullptr, "GetAbstractedPadState"},
             {303, nullptr, "GetAbstractedPadsState"},
@@ -673,6 +670,8 @@ public:
             {322, nullptr, "UnsetAutoPilotVirtualPadState"},
             {323, nullptr, "UnsetAllAutoPilotVirtualPadState"},
             {350, nullptr, "AddRegisteredDevice"},
+            {400, nullptr, "DisableExternalMcuOnNxDevice"},
+            {401, nullptr, "DisableRailDeviceFiltering"},
         };
         // clang-format on
 
@@ -708,7 +707,9 @@ public:
             {307, nullptr, "GetNpadSystemExtStyle"},
             {308, nullptr, "ApplyNpadSystemCommonPolicyFull"},
             {309, nullptr, "GetNpadFullKeyGripColor"},
+            {310, nullptr, "GetMaskedSupportedNpadStyleSet"},
             {311, nullptr, "SetNpadPlayerLedBlinkingDevice"},
+            {312, nullptr, "SetSupportedNpadStyleSetAll"},
             {321, nullptr, "GetUniquePadsFromNpad"},
             {322, nullptr, "GetIrSensorState"},
             {323, nullptr, "GetXcdHandleForNpadWithIrSensor"},
@@ -733,6 +734,7 @@ public:
             {546, nullptr, "AcquireDeviceRegisteredEventForControllerSupport"},
             {547, nullptr, "GetAllowedBluetoothLinksCount"},
             {548, nullptr, "GetRegisteredDevices"},
+            {549, nullptr, "GetConnectableRegisteredDevices"},
             {700, nullptr, "ActivateUniquePad"},
             {702, nullptr, "AcquireUniquePadConnectionEventHandle"},
             {703, nullptr, "GetUniquePadIds"},
@@ -761,6 +763,7 @@ public:
             {850, nullptr, "IsUsbFullKeyControllerEnabled"},
             {851, nullptr, "EnableUsbFullKeyController"},
             {852, nullptr, "IsUsbConnected"},
+            {870, nullptr, "IsHandheldButtonPressedOnConsoleMode"},
             {900, nullptr, "ActivateInputDetector"},
             {901, nullptr, "NotifyInputDetector"},
             {1000, nullptr, "InitializeFirmwareUpdate"},
@@ -780,6 +783,12 @@ public:
             {1052, nullptr, "CancelSixAxisSensorAccurateUserCalibration"},
             {1053, nullptr, "GetSixAxisSensorAccurateUserCalibrationState"},
             {1100, nullptr, "GetHidbusSystemServiceObject"},
+            {1120, nullptr, "SetFirmwareHotfixUpdateSkipEnabled"},
+            {1130, nullptr, "InitializeUsbFirmwareUpdate"},
+            {1131, nullptr, "FinalizeUsbFirmwareUpdate"},
+            {1132, nullptr, "CheckUsbFirmwareUpdateRequired"},
+            {1133, nullptr, "StartUsbFirmwareUpdate"},
+            {1134, nullptr, "GetUsbFirmwareUpdateState"},
         };
         // clang-format on
 
@@ -818,6 +827,7 @@ public:
             {11, nullptr, "EnableJoyPollingReceiveMode"},
             {12, nullptr, "DisableJoyPollingReceiveMode"},
             {13, nullptr, "GetPollingData"},
+            {14, nullptr, "SetStatusManagerType"},
         };
         // clang-format on
 
diff --git a/src/core/hle/service/hid/hid.h b/src/core/hle/service/hid/hid.h
index 88d926808..773035460 100644
--- a/src/core/hle/service/hid/hid.h
+++ b/src/core/hle/service/hid/hid.h
@@ -4,408 +4,12 @@
 
 #pragma once
 
-#include <array>
-#include "common/bit_field.h"
-#include "common/common_types.h"
-#include "core/hle/service/service.h"
+namespace SM {
+class ServiceManager;
+}
 
 namespace Service::HID {
 
-// Begin enums and output structs
-
-constexpr u32 HID_NUM_ENTRIES = 17;
-constexpr u32 HID_NUM_LAYOUTS = 7;
-constexpr s32 HID_JOYSTICK_MAX = 0x8000;
-constexpr s32 HID_JOYSTICK_MIN = -0x8000;
-
-constexpr u32 JOYCON_BODY_NEON_RED = 0xFF3C28;
-constexpr u32 JOYCON_BUTTONS_NEON_RED = 0x1E0A0A;
-constexpr u32 JOYCON_BODY_NEON_BLUE = 0x0AB9E6;
-constexpr u32 JOYCON_BUTTONS_NEON_BLUE = 0x001E1E;
-
-enum ControllerType : u32 {
-    ControllerType_ProController = 1 << 0,
-    ControllerType_Handheld = 1 << 1,
-    ControllerType_JoyconPair = 1 << 2,
-    ControllerType_JoyconLeft = 1 << 3,
-    ControllerType_JoyconRight = 1 << 4,
-};
-
-enum ControllerLayoutType : u32 {
-    Layout_ProController = 0, // Pro Controller or HID gamepad
-    Layout_Handheld = 1,      // Two Joy-Con docked to rails
-    Layout_Single = 2, // Horizontal single Joy-Con or pair of Joy-Con, adjusted for orientation
-    Layout_Left = 3,   // Only raw left Joy-Con state, no orientation adjustment
-    Layout_Right = 4,  // Only raw right Joy-Con state, no orientation adjustment
-    Layout_DefaultDigital = 5, // Same as next, but sticks have 8-direction values only
-    Layout_Default = 6, // Safe default, single Joy-Con have buttons/sticks rotated for orientation
-};
-
-enum ControllerColorDescription {
-    ColorDesc_ColorsNonexistent = 1 << 1,
-};
-
-enum ControllerConnectionState {
-    ConnectionState_Connected = 1 << 0,
-    ConnectionState_Wired = 1 << 1,
-};
-
-enum ControllerJoystick {
-    Joystick_Left = 0,
-    Joystick_Right = 1,
-};
-
-enum ControllerID {
-    Controller_Player1 = 0,
-    Controller_Player2 = 1,
-    Controller_Player3 = 2,
-    Controller_Player4 = 3,
-    Controller_Player5 = 4,
-    Controller_Player6 = 5,
-    Controller_Player7 = 6,
-    Controller_Player8 = 7,
-    Controller_Handheld = 8,
-    Controller_Unknown = 9,
-};
-
-// End enums and output structs
-
-// Begin UnkInput3
-
-struct UnkInput3Header {
-    u64 timestamp_ticks;
-    u64 num_entries;
-    u64 latest_entry;
-    u64 max_entry_index;
-};
-static_assert(sizeof(UnkInput3Header) == 0x20, "HID UnkInput3 header structure has incorrect size");
-
-struct UnkInput3Entry {
-    u64 timestamp;
-    u64 timestamp_2;
-    u64 unk_8;
-    u64 unk_10;
-    u64 unk_18;
-};
-static_assert(sizeof(UnkInput3Entry) == 0x28, "HID UnkInput3 entry structure has incorrect size");
-
-struct UnkInput3 {
-    UnkInput3Header header;
-    std::array<UnkInput3Entry, 17> entries;
-    std::array<u8, 0x138> padding;
-};
-static_assert(sizeof(UnkInput3) == 0x400, "HID UnkInput3 structure has incorrect size");
-
-// End UnkInput3
-
-// Begin TouchScreen
-
-struct TouchScreenHeader {
-    u64 timestamp_ticks;
-    u64 num_entries;
-    u64 latest_entry;
-    u64 max_entry_index;
-    u64 timestamp;
-};
-static_assert(sizeof(TouchScreenHeader) == 0x28,
-              "HID touch screen header structure has incorrect size");
-
-struct TouchScreenEntryHeader {
-    u64 timestamp;
-    u64 num_touches;
-};
-static_assert(sizeof(TouchScreenEntryHeader) == 0x10,
-              "HID touch screen entry header structure has incorrect size");
-
-struct TouchScreenEntryTouch {
-    u64 timestamp;
-    u32 padding;
-    u32 touch_index;
-    u32 x;
-    u32 y;
-    u32 diameter_x;
-    u32 diameter_y;
-    u32 angle;
-    u32 padding_2;
-};
-static_assert(sizeof(TouchScreenEntryTouch) == 0x28,
-              "HID touch screen touch structure has incorrect size");
-
-struct TouchScreenEntry {
-    TouchScreenEntryHeader header;
-    std::array<TouchScreenEntryTouch, 16> touches;
-    u64 unk;
-};
-static_assert(sizeof(TouchScreenEntry) == 0x298,
-              "HID touch screen entry structure has incorrect size");
-
-struct TouchScreen {
-    TouchScreenHeader header;
-    std::array<TouchScreenEntry, 17> entries;
-    std::array<u8, 0x3c0> padding;
-};
-static_assert(sizeof(TouchScreen) == 0x3000, "HID touch screen structure has incorrect size");
-
-// End TouchScreen
-
-// Begin Mouse
-
-struct MouseHeader {
-    u64 timestamp_ticks;
-    u64 num_entries;
-    u64 latest_entry;
-    u64 max_entry_index;
-};
-static_assert(sizeof(MouseHeader) == 0x20, "HID mouse header structure has incorrect size");
-
-struct MouseButtonState {
-    union {
-        u64 hex{};
-
-        // Buttons
-        BitField<0, 1, u64> left;
-        BitField<1, 1, u64> right;
-        BitField<2, 1, u64> middle;
-        BitField<3, 1, u64> forward;
-        BitField<4, 1, u64> back;
-    };
-};
-
-struct MouseEntry {
-    u64 timestamp;
-    u64 timestamp_2;
-    u32 x;
-    u32 y;
-    u32 velocity_x;
-    u32 velocity_y;
-    u32 scroll_velocity_x;
-    u32 scroll_velocity_y;
-    MouseButtonState buttons;
-};
-static_assert(sizeof(MouseEntry) == 0x30, "HID mouse entry structure has incorrect size");
-
-struct Mouse {
-    MouseHeader header;
-    std::array<MouseEntry, 17> entries;
-    std::array<u8, 0xB0> padding;
-};
-static_assert(sizeof(Mouse) == 0x400, "HID mouse structure has incorrect size");
-
-// End Mouse
-
-// Begin Keyboard
-
-struct KeyboardHeader {
-    u64 timestamp_ticks;
-    u64 num_entries;
-    u64 latest_entry;
-    u64 max_entry_index;
-};
-static_assert(sizeof(KeyboardHeader) == 0x20, "HID keyboard header structure has incorrect size");
-
-struct KeyboardModifierKeyState {
-    union {
-        u64 hex{};
-
-        // Buttons
-        BitField<0, 1, u64> lctrl;
-        BitField<1, 1, u64> lshift;
-        BitField<2, 1, u64> lalt;
-        BitField<3, 1, u64> lmeta;
-        BitField<4, 1, u64> rctrl;
-        BitField<5, 1, u64> rshift;
-        BitField<6, 1, u64> ralt;
-        BitField<7, 1, u64> rmeta;
-        BitField<8, 1, u64> capslock;
-        BitField<9, 1, u64> scrolllock;
-        BitField<10, 1, u64> numlock;
-    };
-};
-
-struct KeyboardEntry {
-    u64 timestamp;
-    u64 timestamp_2;
-    KeyboardModifierKeyState modifier;
-    u32 keys[8];
-};
-static_assert(sizeof(KeyboardEntry) == 0x38, "HID keyboard entry structure has incorrect size");
-
-struct Keyboard {
-    KeyboardHeader header;
-    std::array<KeyboardEntry, 17> entries;
-    std::array<u8, 0x28> padding;
-};
-static_assert(sizeof(Keyboard) == 0x400, "HID keyboard structure has incorrect size");
-
-// End Keyboard
-
-// Begin UnkInput1
-
-struct UnkInput1Header {
-    u64 timestamp_ticks;
-    u64 num_entries;
-    u64 latest_entry;
-    u64 max_entry_index;
-};
-static_assert(sizeof(UnkInput1Header) == 0x20, "HID UnkInput1 header structure has incorrect size");
-
-struct UnkInput1Entry {
-    u64 timestamp;
-    u64 timestamp_2;
-    u64 unk_8;
-    u64 unk_10;
-    u64 unk_18;
-};
-static_assert(sizeof(UnkInput1Entry) == 0x28, "HID UnkInput1 entry structure has incorrect size");
-
-struct UnkInput1 {
-    UnkInput1Header header;
-    std::array<UnkInput1Entry, 17> entries;
-    std::array<u8, 0x138> padding;
-};
-static_assert(sizeof(UnkInput1) == 0x400, "HID UnkInput1 structure has incorrect size");
-
-// End UnkInput1
-
-// Begin UnkInput2
-
-struct UnkInput2Header {
-    u64 timestamp_ticks;
-    u64 num_entries;
-    u64 latest_entry;
-    u64 max_entry_index;
-};
-static_assert(sizeof(UnkInput2Header) == 0x20, "HID UnkInput2 header structure has incorrect size");
-
-struct UnkInput2 {
-    UnkInput2Header header;
-    std::array<u8, 0x1E0> padding;
-};
-static_assert(sizeof(UnkInput2) == 0x200, "HID UnkInput2 structure has incorrect size");
-
-// End UnkInput2
-
-// Begin Controller
-
-struct ControllerMAC {
-    u64 timestamp;
-    std::array<u8, 0x8> mac;
-    u64 unk;
-    u64 timestamp_2;
-};
-static_assert(sizeof(ControllerMAC) == 0x20, "HID controller MAC structure has incorrect size");
-
-struct ControllerHeader {
-    u32 type;
-    u32 is_half;
-    u32 single_colors_descriptor;
-    u32 single_color_body;
-    u32 single_color_buttons;
-    u32 split_colors_descriptor;
-    u32 left_color_body;
-    u32 left_color_buttons;
-    u32 right_color_body;
-    u32 right_color_buttons;
-};
-static_assert(sizeof(ControllerHeader) == 0x28,
-              "HID controller header structure has incorrect size");
-
-struct ControllerLayoutHeader {
-    u64 timestamp_ticks;
-    u64 num_entries;
-    u64 latest_entry;
-    u64 max_entry_index;
-};
-static_assert(sizeof(ControllerLayoutHeader) == 0x20,
-              "HID controller layout header structure has incorrect size");
-
-struct ControllerPadState {
-    union {
-        u64 hex{};
-
-        // Buttons
-        BitField<0, 1, u64> a;
-        BitField<1, 1, u64> b;
-        BitField<2, 1, u64> x;
-        BitField<3, 1, u64> y;
-        BitField<4, 1, u64> lstick;
-        BitField<5, 1, u64> rstick;
-        BitField<6, 1, u64> l;
-        BitField<7, 1, u64> r;
-        BitField<8, 1, u64> zl;
-        BitField<9, 1, u64> zr;
-        BitField<10, 1, u64> plus;
-        BitField<11, 1, u64> minus;
-
-        // D-pad buttons
-        BitField<12, 1, u64> dleft;
-        BitField<13, 1, u64> dup;
-        BitField<14, 1, u64> dright;
-        BitField<15, 1, u64> ddown;
-
-        // Left stick directions
-        BitField<16, 1, u64> lstick_left;
-        BitField<17, 1, u64> lstick_up;
-        BitField<18, 1, u64> lstick_right;
-        BitField<19, 1, u64> lstick_down;
-
-        // Right stick directions
-        BitField<20, 1, u64> rstick_left;
-        BitField<21, 1, u64> rstick_up;
-        BitField<22, 1, u64> rstick_right;
-        BitField<23, 1, u64> rstick_down;
-
-        BitField<24, 1, u64> sl;
-        BitField<25, 1, u64> sr;
-    };
-};
-
-struct ControllerInputEntry {
-    u64 timestamp;
-    u64 timestamp_2;
-    ControllerPadState buttons;
-    s32 joystick_left_x;
-    s32 joystick_left_y;
-    s32 joystick_right_x;
-    s32 joystick_right_y;
-    u64 connection_state;
-};
-static_assert(sizeof(ControllerInputEntry) == 0x30,
-              "HID controller input entry structure has incorrect size");
-
-struct ControllerLayout {
-    ControllerLayoutHeader header;
-    std::array<ControllerInputEntry, 17> entries;
-};
-static_assert(sizeof(ControllerLayout) == 0x350,
-              "HID controller layout structure has incorrect size");
-
-struct Controller {
-    ControllerHeader header;
-    std::array<ControllerLayout, HID_NUM_LAYOUTS> layouts;
-    std::array<u8, 0x2a70> unk_1;
-    ControllerMAC mac_left;
-    ControllerMAC mac_right;
-    std::array<u8, 0xdf8> unk_2;
-};
-static_assert(sizeof(Controller) == 0x5000, "HID controller structure has incorrect size");
-
-// End Controller
-
-struct SharedMemory {
-    UnkInput3 unk_input_3;
-    TouchScreen touchscreen;
-    Mouse mouse;
-    Keyboard keyboard;
-    std::array<UnkInput1, 4> unk_input_1;
-    std::array<UnkInput2, 3> unk_input_2;
-    std::array<u8, 0x800> unk_section_8;
-    std::array<u8, 0x4000> controller_serials;
-    std::array<Controller, 10> controllers;
-    std::array<u8, 0x4600> unk_section_9;
-};
-static_assert(sizeof(SharedMemory) == 0x40000, "HID Shared Memory structure has incorrect size");
-
 /// Reload input devices. Used when input configuration changed
 void ReloadInputDevices();
 
diff --git a/src/core/hle/service/lbl/lbl.cpp b/src/core/hle/service/lbl/lbl.cpp
index 7321584e1..164c57e18 100644
--- a/src/core/hle/service/lbl/lbl.cpp
+++ b/src/core/hle/service/lbl/lbl.cpp
@@ -18,35 +18,35 @@ public:
     explicit LBL() : ServiceFramework{"lbl"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "Unknown1"},
-            {1, nullptr, "Unknown2"},
-            {2, nullptr, "Unknown3"},
-            {3, nullptr, "GetCurrentBacklightLevel"},
-            {4, nullptr, "Unknown4"},
-            {5, nullptr, "GetAlsComputedBacklightLevel"},
-            {6, nullptr, "TurnOffBacklight"},
-            {7, nullptr, "TurnOnBacklight"},
-            {8, nullptr, "GetBacklightStatus"},
-            {9, nullptr, "Unknown5"},
-            {10, nullptr, "Unknown6"},
-            {11, nullptr, "Unknown7"},
-            {12, nullptr, "Unknown8"},
-            {13, nullptr, "Unknown9"},
-            {14, nullptr, "Unknown10"},
-            {15, nullptr, "GetAutoBrightnessSetting"},
-            {16, nullptr, "ReadRawLightSensor"},
-            {17, nullptr, "Unknown11"},
-            {18, nullptr, "Unknown12"},
-            {19, nullptr, "Unknown13"},
-            {20, nullptr, "Unknown14"},
-            {21, nullptr, "Unknown15"},
-            {22, nullptr, "Unknown16"},
-            {23, nullptr, "Unknown17"},
-            {24, nullptr, "Unknown18"},
-            {25, nullptr, "Unknown19"},
+            {0, nullptr, "SaveCurrentSetting"},
+            {1, nullptr, "LoadCurrentSetting"},
+            {2, nullptr, "SetCurrentBrightnessSetting"},
+            {3, nullptr, "GetCurrentBrightnessSetting"},
+            {4, nullptr, "ApplyCurrentBrightnessSettingToBacklight"},
+            {5, nullptr, "GetBrightnessSettingAppliedToBacklight"},
+            {6, nullptr, "SwitchBacklightOn"},
+            {7, nullptr, "SwitchBacklightOff"},
+            {8, nullptr, "GetBacklightSwitchStatus"},
+            {9, nullptr, "EnableDimming"},
+            {10, nullptr, "DisableDimming"},
+            {11, nullptr, "IsDimmingEnabled"},
+            {12, nullptr, "EnableAutoBrightnessControl"},
+            {13, nullptr, "DisableAutoBrightnessControl"},
+            {14, nullptr, "IsAutoBrightnessControlEnabled"},
+            {15, nullptr, "SetAmbientLightSensorValue"},
+            {16, nullptr, "GetAmbientLightSensorValue"},
+            {17, nullptr, "SetBrightnessReflectionDelayLevel"},
+            {18, nullptr, "GetBrightnessReflectionDelayLevel"},
+            {19, nullptr, "SetCurrentBrightnessMapping"},
+            {20, nullptr, "GetCurrentBrightnessMapping"},
+            {21, nullptr, "SetCurrentAmbientLightSensorMapping"},
+            {22, nullptr, "GetCurrentAmbientLightSensorMapping"},
+            {23, nullptr, "IsAmbientLightSensorAvailable"},
+            {24, nullptr, "SetCurrentBrightnessSettingForVrMode"},
+            {25, nullptr, "GetCurrentBrightnessSettingForVrMode"},
             {26, &LBL::EnableVrMode, "EnableVrMode"},
             {27, &LBL::DisableVrMode, "DisableVrMode"},
-            {28, &LBL::GetVrMode, "GetVrMode"},
+            {28, &LBL::IsVrModeEnabled, "IsVrModeEnabled"},
         };
         // clang-format on
 
@@ -72,7 +72,7 @@ private:
         LOG_DEBUG(Service_LBL, "called");
     }
 
-    void GetVrMode(Kernel::HLERequestContext& ctx) {
+    void IsVrModeEnabled(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push(vr_mode_enabled);
diff --git a/src/core/hle/service/ldr/ldr.cpp b/src/core/hle/service/ldr/ldr.cpp
index ec32faf15..d607d985e 100644
--- a/src/core/hle/service/ldr/ldr.cpp
+++ b/src/core/hle/service/ldr/ldr.cpp
@@ -3,9 +3,13 @@
 // Refer to the license.txt file included.
 
 #include <memory>
+#include <fmt/format.h>
 
+#include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/process.h"
 #include "core/hle/service/ldr/ldr.h"
 #include "core/hle/service/service.h"
+#include "core/loader/nro.h"
 
 namespace Service::LDR {
 
@@ -59,16 +63,58 @@ public:
     explicit RelocatableObject() : ServiceFramework{"ldr:ro"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "LoadNro"},
+            {0, &RelocatableObject::LoadNro, "LoadNro"},
             {1, nullptr, "UnloadNro"},
-            {2, nullptr, "LoadNrr"},
+            {2, &RelocatableObject::LoadNrr, "LoadNrr"},
             {3, nullptr, "UnloadNrr"},
-            {4, nullptr, "Initialize"},
+            {4, &RelocatableObject::Initialize, "Initialize"},
         };
         // clang-format on
 
         RegisterHandlers(functions);
     }
+
+    void LoadNrr(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
+
+    void LoadNro(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        rp.Skip(2, false);
+        const VAddr nro_addr{rp.Pop<VAddr>()};
+        const u64 nro_size{rp.Pop<u64>()};
+        const VAddr bss_addr{rp.Pop<VAddr>()};
+        const u64 bss_size{rp.Pop<u64>()};
+
+        // Read NRO data from memory
+        std::vector<u8> nro_data(nro_size);
+        Memory::ReadBlock(nro_addr, nro_data.data(), nro_size);
+
+        // Load NRO as new executable module
+        const VAddr addr{*Core::CurrentProcess()->VMManager().FindFreeRegion(nro_size + bss_size)};
+        Loader::AppLoader_NRO::LoadNro(nro_data, fmt::format("nro-{:08x}", addr), addr);
+
+        // TODO(bunnei): This is an incomplete implementation. It was tested with Super Mario Party.
+        // It is currently missing:
+        // - Signature checks with LoadNRR
+        // - Checking if a module has already been loaded
+        // - Using/validating BSS, etc. params (these are used from NRO header instead)
+        // - Error checking
+        // - ...Probably other things
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push(addr);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
+
+    void Initialize(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
 };
 
 void InstallInterfaces(SM::ServiceManager& sm) {
diff --git a/src/core/hle/service/mm/mm_u.cpp b/src/core/hle/service/mm/mm_u.cpp
index 7b91bb258..e1f17a926 100644
--- a/src/core/hle/service/mm/mm_u.cpp
+++ b/src/core/hle/service/mm/mm_u.cpp
@@ -14,14 +14,14 @@ public:
     explicit MM_U() : ServiceFramework{"mm:u"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, &MM_U::Initialize, "InitializeOld"},
-            {1, &MM_U::Finalize, "FinalizeOld"},
-            {2, &MM_U::SetAndWait, "SetAndWaitOld"},
-            {3, &MM_U::Get, "GetOld"},
-            {4, &MM_U::Initialize, "Initialize"},
-            {5, &MM_U::Finalize, "Finalize"},
-            {6, &MM_U::SetAndWait, "SetAndWait"},
-            {7, &MM_U::Get, "Get"},
+            {0, &MM_U::Initialize, "Initialize"},
+            {1, &MM_U::Finalize, "Finalize"},
+            {2, &MM_U::SetAndWait, "SetAndWait"},
+            {3, &MM_U::Get, "Get"},
+            {4, &MM_U::InitializeWithId, "InitializeWithId"},
+            {5, &MM_U::FinalizeWithId, "FinalizeWithId"},
+            {6, &MM_U::SetAndWaitWithId, "SetAndWaitWithId"},
+            {7, &MM_U::GetWithId, "GetWithId"},
         };
         // clang-format on
 
@@ -59,9 +59,43 @@ private:
         rb.Push(current);
     }
 
+    void InitializeWithId(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_MM, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(id); // Any non zero value
+    }
+
+    void FinalizeWithId(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_MM, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void SetAndWaitWithId(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        u32 input_id = rp.Pop<u32>();
+        min = rp.Pop<u32>();
+        max = rp.Pop<u32>();
+        current = min;
+
+        LOG_WARNING(Service_MM, "(STUBBED) called, input_id=0x{:X}, min=0x{:X}, max=0x{:X}",
+                    input_id, min, max);
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetWithId(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_MM, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push(current);
+    }
+
     u32 min{0};
     u32 max{0};
     u32 current{0};
+    u32 id{1};
 };
 
 void InstallInterfaces(SM::ServiceManager& service_manager) {
diff --git a/src/core/hle/service/nfc/nfc.cpp b/src/core/hle/service/nfc/nfc.cpp
index 8fec97db8..30e542542 100644
--- a/src/core/hle/service/nfc/nfc.cpp
+++ b/src/core/hle/service/nfc/nfc.cpp
@@ -10,12 +10,13 @@
 #include "core/hle/service/nfc/nfc.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/sm/sm.h"
+#include "core/settings.h"
 
 namespace Service::NFC {
 
 class IAm final : public ServiceFramework<IAm> {
 public:
-    explicit IAm() : ServiceFramework{"IAm"} {
+    explicit IAm() : ServiceFramework{"NFC::IAm"} {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "Initialize"},
@@ -52,7 +53,7 @@ private:
 
 class MFIUser final : public ServiceFramework<MFIUser> {
 public:
-    explicit MFIUser() : ServiceFramework{"IUser"} {
+    explicit MFIUser() : ServiceFramework{"NFC::MFIUser"} {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "Initialize"},
@@ -100,13 +101,13 @@ private:
 
 class IUser final : public ServiceFramework<IUser> {
 public:
-    explicit IUser() : ServiceFramework{"IUser"} {
+    explicit IUser() : ServiceFramework{"NFC::IUser"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "Initialize"},
-            {1, nullptr, "Finalize"},
-            {2, nullptr, "GetState"},
-            {3, nullptr, "IsNfcEnabled"},
+            {0, &IUser::InitializeOld, "InitializeOld"},
+            {1, &IUser::FinalizeOld, "FinalizeOld"},
+            {2, &IUser::GetStateOld, "GetStateOld"},
+            {3, &IUser::IsNfcEnabledOld, "IsNfcEnabledOld"},
             {400, nullptr, "Initialize"},
             {401, nullptr, "Finalize"},
             {402, nullptr, "GetState"},
@@ -130,11 +131,47 @@ public:
 
         RegisterHandlers(functions);
     }
+
+private:
+    enum class NfcStates : u32 {
+        Finalized = 6,
+    };
+
+    void InitializeOld(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0};
+        rb.Push(RESULT_SUCCESS);
+
+        // We don't deal with hardware initialization so we can just stub this.
+        LOG_DEBUG(Service_NFC, "called");
+    }
+
+    void IsNfcEnabledOld(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u8>(Settings::values.enable_nfc);
+
+        LOG_DEBUG(Service_NFC, "IsNfcEnabledOld");
+    }
+
+    void GetStateOld(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFC, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushEnum(NfcStates::Finalized); // TODO(ogniK): Figure out if this matches nfp
+    }
+
+    void FinalizeOld(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFC, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
 };
 
 class NFC_U final : public ServiceFramework<NFC_U> {
 public:
-    explicit NFC_U() : ServiceFramework{"nfc:u"} {
+    explicit NFC_U() : ServiceFramework{"nfc:user"} {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, &NFC_U::CreateUserInterface, "CreateUserInterface"},
diff --git a/src/core/hle/service/nfp/nfp.cpp b/src/core/hle/service/nfp/nfp.cpp
index 8c07a05c2..c1af878fe 100644
--- a/src/core/hle/service/nfp/nfp.cpp
+++ b/src/core/hle/service/nfp/nfp.cpp
@@ -2,56 +2,67 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <atomic>
+
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
+#include "core/hle/lock.h"
 #include "core/hle/service/hid/hid.h"
 #include "core/hle/service/nfp/nfp.h"
 #include "core/hle/service/nfp/nfp_user.h"
 
 namespace Service::NFP {
 
+namespace ErrCodes {
+constexpr ResultCode ERR_TAG_FAILED(ErrorModule::NFP,
+                                    -1); // TODO(ogniK): Find the actual error code
+}
+
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
-    : ServiceFramework(name), module(std::move(module)) {}
+    : ServiceFramework(name), module(std::move(module)) {
+    auto& kernel = Core::System::GetInstance().Kernel();
+    nfc_tag_load =
+        Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "IUser:NFCTagDetected");
+}
 
 Module::Interface::~Interface() = default;
 
 class IUser final : public ServiceFramework<IUser> {
 public:
-    IUser() : ServiceFramework("IUser") {
+    IUser(Module::Interface& nfp_interface)
+        : ServiceFramework("NFP::IUser"), nfp_interface(nfp_interface) {
         static const FunctionInfo functions[] = {
             {0, &IUser::Initialize, "Initialize"},
-            {1, nullptr, "Finalize"},
+            {1, &IUser::Finalize, "Finalize"},
             {2, &IUser::ListDevices, "ListDevices"},
-            {3, nullptr, "StartDetection"},
-            {4, nullptr, "StopDetection"},
-            {5, nullptr, "Mount"},
-            {6, nullptr, "Unmount"},
-            {7, nullptr, "OpenApplicationArea"},
-            {8, nullptr, "GetApplicationArea"},
+            {3, &IUser::StartDetection, "StartDetection"},
+            {4, &IUser::StopDetection, "StopDetection"},
+            {5, &IUser::Mount, "Mount"},
+            {6, &IUser::Unmount, "Unmount"},
+            {7, &IUser::OpenApplicationArea, "OpenApplicationArea"},
+            {8, &IUser::GetApplicationArea, "GetApplicationArea"},
             {9, nullptr, "SetApplicationArea"},
             {10, nullptr, "Flush"},
             {11, nullptr, "Restore"},
             {12, nullptr, "CreateApplicationArea"},
-            {13, nullptr, "GetTagInfo"},
-            {14, nullptr, "GetRegisterInfo"},
-            {15, nullptr, "GetCommonInfo"},
-            {16, nullptr, "GetModelInfo"},
+            {13, &IUser::GetTagInfo, "GetTagInfo"},
+            {14, &IUser::GetRegisterInfo, "GetRegisterInfo"},
+            {15, &IUser::GetCommonInfo, "GetCommonInfo"},
+            {16, &IUser::GetModelInfo, "GetModelInfo"},
             {17, &IUser::AttachActivateEvent, "AttachActivateEvent"},
             {18, &IUser::AttachDeactivateEvent, "AttachDeactivateEvent"},
             {19, &IUser::GetState, "GetState"},
             {20, &IUser::GetDeviceState, "GetDeviceState"},
             {21, &IUser::GetNpadId, "GetNpadId"},
-            {22, nullptr, "GetApplicationArea2"},
+            {22, &IUser::GetApplicationAreaSize, "GetApplicationAreaSize"},
             {23, &IUser::AttachAvailabilityChangeEvent, "AttachAvailabilityChangeEvent"},
             {24, nullptr, "RecreateApplicationArea"},
         };
         RegisterHandlers(functions);
 
         auto& kernel = Core::System::GetInstance().Kernel();
-        activate_event =
-            Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "IUser:ActivateEvent");
         deactivate_event =
             Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "IUser:DeactivateEvent");
         availability_change_event = Kernel::Event::Create(kernel, Kernel::ResetType::OneShot,
@@ -59,6 +70,17 @@ public:
     }
 
 private:
+    struct TagInfo {
+        std::array<u8, 10> uuid;
+        u8 uuid_length; // TODO(ogniK): Figure out if this is actual the uuid length or does it
+                        // mean something else
+        INSERT_PADDING_BYTES(0x15);
+        u32_le protocol;
+        u32_le tag_type;
+        INSERT_PADDING_BYTES(0x2c);
+    };
+    static_assert(sizeof(TagInfo) == 0x54, "TagInfo is an invalid size");
+
     enum class State : u32 {
         NonInitialized = 0,
         Initialized = 1,
@@ -66,15 +88,40 @@ private:
 
     enum class DeviceState : u32 {
         Initialized = 0,
+        SearchingForTag = 1,
+        TagFound = 2,
+        TagRemoved = 3,
+        TagNearby = 4,
+        Unknown5 = 5,
+        Finalized = 6
+    };
+
+    struct CommonInfo {
+        u16_be last_write_year;
+        u8 last_write_month;
+        u8 last_write_day;
+        u16_be write_counter;
+        u16_be version;
+        u32_be application_area_size;
+        INSERT_PADDING_BYTES(0x34);
     };
+    static_assert(sizeof(CommonInfo) == 0x40, "CommonInfo is an invalid size");
 
     void Initialize(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_NFP, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2, 0};
+        rb.Push(RESULT_SUCCESS);
 
         state = State::Initialized;
 
-        IPC::ResponseBuilder rb{ctx, 2};
+        LOG_DEBUG(Service_NFC, "called");
+    }
+
+    void GetState(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 3, 0};
         rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(static_cast<u32>(state));
+
+        LOG_DEBUG(Service_NFC, "called");
     }
 
     void ListDevices(Kernel::HLERequestContext& ctx) {
@@ -83,80 +130,219 @@ private:
 
         ctx.WriteBuffer(&device_handle, sizeof(device_handle));
 
-        LOG_WARNING(Service_NFP, "(STUBBED) called, array_size={}", array_size);
+        LOG_DEBUG(Service_NFP, "called, array_size={}", array_size);
 
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(0);
+        rb.Push<u32>(1);
     }
 
-    void AttachActivateEvent(Kernel::HLERequestContext& ctx) {
+    void GetNpadId(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
         const u64 dev_handle = rp.Pop<u64>();
-        LOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+        LOG_DEBUG(Service_NFP, "called, dev_handle=0x{:X}", dev_handle);
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(npad_id);
+    }
 
+    void AttachActivateEvent(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const u64 dev_handle = rp.Pop<u64>();
+        LOG_DEBUG(Service_NFP, "called, dev_handle=0x{:X}", dev_handle);
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushCopyObjects(activate_event);
+        rb.PushCopyObjects(nfp_interface.GetNFCEvent());
+        has_attached_handle = true;
     }
 
     void AttachDeactivateEvent(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
         const u64 dev_handle = rp.Pop<u64>();
-        LOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+        LOG_DEBUG(Service_NFP, "called, dev_handle=0x{:X}", dev_handle);
 
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
         rb.PushCopyObjects(deactivate_event);
     }
 
-    void GetState(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_NFP, "(STUBBED) called");
-        IPC::ResponseBuilder rb{ctx, 3};
+    void StopDetection(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+        switch (device_state) {
+        case DeviceState::TagFound:
+        case DeviceState::TagNearby:
+            deactivate_event->Signal();
+            device_state = DeviceState::Initialized;
+            break;
+        case DeviceState::SearchingForTag:
+        case DeviceState::TagRemoved:
+            device_state = DeviceState::Initialized;
+            break;
+        }
+        IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(static_cast<u32>(state));
     }
 
     void GetDeviceState(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_NFP, "(STUBBED) called");
+        LOG_DEBUG(Service_NFP, "called");
+        auto nfc_event = nfp_interface.GetNFCEvent();
+        if (!nfc_event->ShouldWait(Kernel::GetCurrentThread()) && !has_attached_handle) {
+            device_state = DeviceState::TagFound;
+            nfc_event->Clear();
+        }
+
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push<u32>(static_cast<u32>(device_state));
     }
 
-    void GetNpadId(Kernel::HLERequestContext& ctx) {
-        IPC::RequestParser rp{ctx};
-        const u64 dev_handle = rp.Pop<u64>();
-        LOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
-        IPC::ResponseBuilder rb{ctx, 3};
+    void StartDetection(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        if (device_state == DeviceState::Initialized || device_state == DeviceState::TagRemoved) {
+            device_state = DeviceState::SearchingForTag;
+        }
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetTagInfo(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        auto amiibo = nfp_interface.GetAmiiboBuffer();
+        TagInfo tag_info{};
+        std::memcpy(tag_info.uuid.data(), amiibo.uuid.data(), sizeof(tag_info.uuid.size()));
+        tag_info.uuid_length = static_cast<u8>(tag_info.uuid.size());
+
+        tag_info.protocol = 1; // TODO(ogniK): Figure out actual values
+        tag_info.tag_type = 2;
+        ctx.WriteBuffer(&tag_info, sizeof(TagInfo));
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void Mount(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        device_state = DeviceState::TagNearby;
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetModelInfo(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        auto amiibo = nfp_interface.GetAmiiboBuffer();
+        ctx.WriteBuffer(&amiibo.model_info, sizeof(amiibo.model_info));
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void Unmount(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        device_state = DeviceState::TagFound;
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void Finalize(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        device_state = DeviceState::Finalized;
+
+        IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(npad_id);
     }
 
     void AttachAvailabilityChangeEvent(Kernel::HLERequestContext& ctx) {
-        IPC::RequestParser rp{ctx};
-        const u64 dev_handle = rp.Pop<u64>();
-        LOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
 
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
         rb.PushCopyObjects(availability_change_event);
     }
 
-    const u64 device_handle{0xDEAD};
-    const HID::ControllerID npad_id{HID::Controller_Player1};
+    void GetRegisterInfo(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
+
+        // TODO(ogniK): Pull Mii and owner data from amiibo
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetCommonInfo(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
+
+        // TODO(ogniK): Pull common information from amiibo
+
+        CommonInfo common_info{};
+        common_info.application_area_size = 0;
+        ctx.WriteBuffer(&common_info, sizeof(CommonInfo));
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void OpenApplicationArea(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+        // We don't need to worry about this since we can just open the file
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetApplicationAreaSize(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
+        // We don't need to worry about this since we can just open the file
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(0); // This is from the GetCommonInfo stub
+    }
+
+    void GetApplicationArea(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
+
+        // TODO(ogniK): Pull application area from amiibo
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(0); // This is from the GetCommonInfo stub
+    }
+
+    bool has_attached_handle{};
+    const u64 device_handle{Common::MakeMagic('Y', 'U', 'Z', 'U')};
+    const u32 npad_id{0}; // Player 1 controller
     State state{State::NonInitialized};
     DeviceState device_state{DeviceState::Initialized};
-    Kernel::SharedPtr<Kernel::Event> activate_event;
     Kernel::SharedPtr<Kernel::Event> deactivate_event;
     Kernel::SharedPtr<Kernel::Event> availability_change_event;
+    const Module::Interface& nfp_interface;
 };
 
 void Module::Interface::CreateUserInterface(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_NFP, "called");
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<IUser>();
+    rb.PushIpcInterface<IUser>(*this);
+}
+
+bool Module::Interface::LoadAmiibo(const std::vector<u8>& buffer) {
+    std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
+    if (buffer.size() < sizeof(AmiiboFile)) {
+        return false;
+    }
+
+    std::memcpy(&amiibo, buffer.data(), sizeof(amiibo));
+    nfc_tag_load->Signal();
+    return true;
+}
+const Kernel::SharedPtr<Kernel::Event>& Module::Interface::GetNFCEvent() const {
+    return nfc_tag_load;
+}
+const Module::Interface::AmiiboFile& Module::Interface::GetAmiiboBuffer() const {
+    return amiibo;
 }
 
 void InstallInterfaces(SM::ServiceManager& service_manager) {
diff --git a/src/core/hle/service/nfp/nfp.h b/src/core/hle/service/nfp/nfp.h
index 77df343c4..5c0ae8a54 100644
--- a/src/core/hle/service/nfp/nfp.h
+++ b/src/core/hle/service/nfp/nfp.h
@@ -4,6 +4,9 @@
 
 #pragma once
 
+#include <array>
+#include <vector>
+#include "core/hle/kernel/event.h"
 #include "core/hle/service/service.h"
 
 namespace Service::NFP {
@@ -15,7 +18,27 @@ public:
         explicit Interface(std::shared_ptr<Module> module, const char* name);
         ~Interface() override;
 
+        struct ModelInfo {
+            std::array<u8, 0x8> amiibo_identification_block;
+            INSERT_PADDING_BYTES(0x38);
+        };
+        static_assert(sizeof(ModelInfo) == 0x40, "ModelInfo is an invalid size");
+
+        struct AmiiboFile {
+            std::array<u8, 10> uuid;
+            INSERT_PADDING_BYTES(0x4a);
+            ModelInfo model_info;
+        };
+        static_assert(sizeof(AmiiboFile) == 0x94, "AmiiboFile is an invalid size");
+
         void CreateUserInterface(Kernel::HLERequestContext& ctx);
+        bool LoadAmiibo(const std::vector<u8>& buffer);
+        const Kernel::SharedPtr<Kernel::Event>& GetNFCEvent() const;
+        const AmiiboFile& GetAmiiboBuffer() const;
+
+    private:
+        Kernel::SharedPtr<Kernel::Event> nfc_tag_load{};
+        AmiiboFile amiibo{};
 
     protected:
         std::shared_ptr<Module> module;
diff --git a/src/core/hle/service/nifm/nifm.cpp b/src/core/hle/service/nifm/nifm.cpp
index 10611ed6a..75dcd94a3 100644
--- a/src/core/hle/service/nifm/nifm.cpp
+++ b/src/core/hle/service/nifm/nifm.cpp
@@ -219,6 +219,7 @@ IGeneralService::IGeneralService() : ServiceFramework("IGeneralService") {
         {35, nullptr, "GetScanData"},
         {36, nullptr, "GetCurrentAccessPoint"},
         {37, nullptr, "Shutdown"},
+        {38, nullptr, "GetAllowedChannels"},
     };
     RegisterHandlers(functions);
 }
diff --git a/src/core/hle/service/nim/nim.cpp b/src/core/hle/service/nim/nim.cpp
index 261ad539c..18091c9bb 100644
--- a/src/core/hle/service/nim/nim.cpp
+++ b/src/core/hle/service/nim/nim.cpp
@@ -71,6 +71,22 @@ public:
     }
 };
 
+class NIM_ECA final : public ServiceFramework<NIM_ECA> {
+public:
+    explicit NIM_ECA() : ServiceFramework{"nim:eca"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, nullptr, "CreateServerInterface"},
+            {1, nullptr, "RefreshDebugAvailability"},
+            {2, nullptr, "ClearDebugResponse"},
+            {3, nullptr, "RegisterDebugResponse"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
 class NIM_SHP final : public ServiceFramework<NIM_SHP> {
 public:
     explicit NIM_SHP() : ServiceFramework{"nim:shp"} {
@@ -214,6 +230,7 @@ private:
 
 void InstallInterfaces(SM::ServiceManager& sm) {
     std::make_shared<NIM>()->InstallAsService(sm);
+    std::make_shared<NIM_ECA>()->InstallAsService(sm);
     std::make_shared<NIM_SHP>()->InstallAsService(sm);
     std::make_shared<NTC>()->InstallAsService(sm);
 }
diff --git a/src/core/hle/service/npns/npns.cpp b/src/core/hle/service/npns/npns.cpp
new file mode 100644
index 000000000..ccb6f9da9
--- /dev/null
+++ b/src/core/hle/service/npns/npns.cpp
@@ -0,0 +1,88 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <memory>
+
+#include "core/hle/service/npns/npns.h"
+#include "core/hle/service/service.h"
+#include "core/hle/service/sm/sm.h"
+
+namespace Service::NPNS {
+
+class NPNS_S final : public ServiceFramework<NPNS_S> {
+public:
+    explicit NPNS_S() : ServiceFramework{"npns:s"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {1, nullptr, "ListenAll"},
+            {2, nullptr, "ListenTo"},
+            {3, nullptr, "Receive"},
+            {4, nullptr, "ReceiveRaw"},
+            {5, nullptr, "GetReceiveEvent"},
+            {6, nullptr, "ListenUndelivered"},
+            {7, nullptr, "GetStateChangeEVent"},
+            {11, nullptr, "SubscribeTopic"},
+            {12, nullptr, "UnsubscribeTopic"},
+            {13, nullptr, "QueryIsTopicExist"},
+            {21, nullptr, "CreateToken"},
+            {22, nullptr, "CreateTokenWithApplicationId"},
+            {23, nullptr, "DestroyToken"},
+            {24, nullptr, "DestroyTokenWithApplicationId"},
+            {25, nullptr, "QueryIsTokenValid"},
+            {31, nullptr, "UploadTokenToBaaS"},
+            {32, nullptr, "DestroyTokenForBaaS"},
+            {33, nullptr, "CreateTokenForBaaS"},
+            {34, nullptr, "SetBaaSDeviceAccountIdList"},
+            {101, nullptr, "Suspend"},
+            {102, nullptr, "Resume"},
+            {103, nullptr, "GetState"},
+            {104, nullptr, "GetStatistics"},
+            {105, nullptr, "GetPlayReportRequestEvent"},
+            {111, nullptr, "GetJid"},
+            {112, nullptr, "CreateJid"},
+            {113, nullptr, "DestroyJid"},
+            {114, nullptr, "AttachJid"},
+            {115, nullptr, "DetachJid"},
+            {201, nullptr, "RequestChangeStateForceTimed"},
+            {102, nullptr, "RequestChangeStateForceAsync"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+class NPNS_U final : public ServiceFramework<NPNS_U> {
+public:
+    explicit NPNS_U() : ServiceFramework{"npns:u"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {1, nullptr, "ListenAll"},
+            {2, nullptr, "ListenTo"},
+            {3, nullptr, "Receive"},
+            {4, nullptr, "ReceiveRaw"},
+            {5, nullptr, "GetReceiveEvent"},
+            {7, nullptr, "GetStateChangeEVent"},
+            {21, nullptr, "CreateToken"},
+            {23, nullptr, "DestroyToken"},
+            {25, nullptr, "QueryIsTokenValid"},
+            {26, nullptr, "ListenToMyApplicationId"},
+            {101, nullptr, "Suspend"},
+            {102, nullptr, "Resume"},
+            {103, nullptr, "GetState"},
+            {104, nullptr, "GetStatistics"},
+            {111, nullptr, "GetJid"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+void InstallInterfaces(SM::ServiceManager& sm) {
+    std::make_shared<NPNS_S>()->InstallAsService(sm);
+    std::make_shared<NPNS_U>()->InstallAsService(sm);
+}
+
+} // namespace Service::NPNS
diff --git a/src/core/hle/service/npns/npns.h b/src/core/hle/service/npns/npns.h
new file mode 100644
index 000000000..861cd3e48
--- /dev/null
+++ b/src/core/hle/service/npns/npns.h
@@ -0,0 +1,15 @@
+// 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::NPNS {
+
+void InstallInterfaces(SM::ServiceManager& sm);
+
+} // namespace Service::NPNS
diff --git a/src/core/hle/service/ns/ns.cpp b/src/core/hle/service/ns/ns.cpp
index 98017267c..07c1381fe 100644
--- a/src/core/hle/service/ns/ns.cpp
+++ b/src/core/hle/service/ns/ns.cpp
@@ -93,13 +93,23 @@ public:
             {86, nullptr, "EnableApplicationCrashReport"},
             {87, nullptr, "IsApplicationCrashReportEnabled"},
             {90, nullptr, "BoostSystemMemoryResourceLimit"},
+            {91, nullptr, "Unknown1"},
+            {92, nullptr, "Unknown2"},
+            {93, nullptr, "GetMainApplicationProgramIndex"},
+            {94, nullptr, "LaunchApplication2"},
+            {95, nullptr, "GetApplicationLaunchInfo"},
+            {96, nullptr, "AcquireApplicationLaunchInfo"},
+            {97, nullptr, "GetMainApplicationProgramIndex2"},
+            {98, nullptr, "EnableApplicationAllThreadDumpOnCrash"},
             {100, nullptr, "ResetToFactorySettings"},
             {101, nullptr, "ResetToFactorySettingsWithoutUserSaveData"},
             {102, nullptr, "ResetToFactorySettingsForRefurbishment"},
             {200, nullptr, "CalculateUserSaveDataStatistics"},
             {201, nullptr, "DeleteUserSaveDataAll"},
             {210, nullptr, "DeleteUserSystemSaveData"},
+            {211, nullptr, "DeleteSaveData"},
             {220, nullptr, "UnregisterNetworkServiceAccount"},
+            {221, nullptr, "UnregisterNetworkServiceAccountWithUserSaveDataDeletion"},
             {300, nullptr, "GetApplicationShellEvent"},
             {301, nullptr, "PopApplicationShellEventInfo"},
             {302, nullptr, "LaunchLibraryApplet"},
@@ -114,6 +124,7 @@ public:
             {403, nullptr, "GetMaxApplicationControlCacheCount"},
             {404, nullptr, "InvalidateApplicationControlCache"},
             {405, nullptr, "ListApplicationControlCacheEntryInfo"},
+            {406, nullptr, "GetApplicationControlProperty"},
             {502, nullptr, "RequestCheckGameCardRegistration"},
             {503, nullptr, "RequestGameCardRegistrationGoldPoint"},
             {504, nullptr, "RequestRegisterGameCard"},
@@ -129,6 +140,7 @@ public:
             {604, nullptr, "RegisterContentsExternalKey"},
             {605, nullptr, "ListApplicationContentMetaStatusWithRightsCheck"},
             {606, nullptr, "GetContentMetaStorage"},
+            {607, nullptr, "ListAvailableAddOnContent"},
             {700, nullptr, "PushDownloadTaskList"},
             {701, nullptr, "ClearTaskStatusList"},
             {702, nullptr, "RequestDownloadTaskList"},
@@ -148,6 +160,9 @@ public:
             {907, nullptr, "WithdrawApplicationUpdateRequest"},
             {908, nullptr, "ListApplicationRecordInstalledContentMeta"},
             {909, nullptr, "WithdrawCleanupAddOnContentsWithNoRightsRecommendation"},
+            {910, nullptr, "Unknown3"},
+            {911, nullptr, "SetPreInstalledApplication"},
+            {912, nullptr, "ClearPreInstalledApplicationFlag"},
             {1000, nullptr, "RequestVerifyApplicationDeprecated"},
             {1001, nullptr, "CorruptApplicationForDebug"},
             {1002, nullptr, "RequestVerifyAddOnContentsRights"},
@@ -162,6 +177,8 @@ public:
             {1305, nullptr, "TryDeleteRunningApplicationEntity"},
             {1306, nullptr, "TryDeleteRunningApplicationCompletely"},
             {1307, nullptr, "TryDeleteRunningApplicationContentEntities"},
+            {1308, nullptr, "DeleteApplicationCompletelyForDebug"},
+            {1309, nullptr, "CleanupUnavailableAddOnContents"},
             {1400, nullptr, "PrepareShutdown"},
             {1500, nullptr, "FormatSdCard"},
             {1501, nullptr, "NeedsSystemUpdateToFormatSdCard"},
@@ -199,6 +216,28 @@ public:
             {2015, nullptr, "CompareSystemDeliveryInfo"},
             {2016, nullptr, "ListNotCommittedContentMeta"},
             {2017, nullptr, "CreateDownloadTask"},
+            {2018, nullptr, "Unknown4"},
+            {2050, nullptr, "Unknown5"},
+            {2100, nullptr, "Unknown6"},
+            {2101, nullptr, "Unknown7"},
+            {2150, nullptr, "CreateRightsEnvironment"},
+            {2151, nullptr, "DestroyRightsEnvironment"},
+            {2152, nullptr, "ActivateRightsEnvironment"},
+            {2153, nullptr, "DeactivateRightsEnvironment"},
+            {2154, nullptr, "ForceActivateRightsContextForExit"},
+            {2160, nullptr, "AddTargetApplicationToRightsEnvironment"},
+            {2161, nullptr, "SetUsersToRightsEnvironment"},
+            {2170, nullptr, "GetRightsEnvironmentStatus"},
+            {2171, nullptr, "GetRightsEnvironmentStatusChangedEvent"},
+            {2180, nullptr, "RequestExtendRightsInRightsEnvironment"},
+            {2181, nullptr, "GetLastResultOfExtendRightsInRightsEnvironment"},
+            {2182, nullptr, "SetActiveRightsContextUsingStateToRightsEnvironment"},
+            {2190, nullptr, "GetRightsEnvironmentHandleForApplication"},
+            {2199, nullptr, "GetRightsEnvironmentCountForDebug"},
+            {2200, nullptr, "Unknown8"},
+            {2201, nullptr, "Unknown9"},
+            {2250, nullptr, "Unknown10"},
+            {2300, nullptr, "Unknown11"},
         };
         // clang-format on
 
@@ -348,12 +387,15 @@ public:
             {0, nullptr, "LaunchProgram"},
             {1, nullptr, "TerminateProcess"},
             {2, nullptr, "TerminateProgram"},
-            {3, nullptr, "GetShellEventHandle"},
-            {4, nullptr, "GetShellEventInfo"},
-            {5, nullptr, "TerminateApplication"},
-            {6, nullptr, "PrepareLaunchProgramFromHost"},
-            {7, nullptr, "LaunchApplication"},
-            {8, nullptr, "LaunchApplicationWithStorageId"},
+            {4, nullptr, "GetShellEventHandle"},
+            {5, nullptr, "GetShellEventInfo"},
+            {6, nullptr, "TerminateApplication"},
+            {7, nullptr, "PrepareLaunchProgramFromHost"},
+            {8, nullptr, "LaunchApplication"},
+            {9, nullptr, "LaunchApplicationWithStorageId"},
+            {10, nullptr, "TerminateApplication2"},
+            {11, nullptr, "GetRunningApplicationProcessId"},
+            {12, nullptr, "SetCurrentApplicationRightsEnvironmentCanBeActive"},
         };
         // clang-format on
 
@@ -388,6 +430,7 @@ public:
             {19, nullptr, "GetReceivedEulaDataSize"},
             {20, nullptr, "GetReceivedEulaData"},
             {21, nullptr, "SetupToReceiveSystemUpdate"},
+            {22, nullptr, "RequestCheckLatestUpdateIncludesRebootlessUpdate"},
         };
         // clang-format on
 
diff --git a/src/core/hle/service/ns/pl_u.cpp b/src/core/hle/service/ns/pl_u.cpp
index 4b2f758a8..44accecb7 100644
--- a/src/core/hle/service/ns/pl_u.cpp
+++ b/src/core/hle/service/ns/pl_u.cpp
@@ -161,7 +161,7 @@ PL_U::PL_U() : ServiceFramework("pl:u"), impl{std::make_unique<Impl>()} {
     };
     RegisterHandlers(functions);
     // Attempt to load shared font data from disk
-    const auto nand = FileSystem::GetSystemNANDContents();
+    const auto* nand = FileSystem::GetSystemNANDContents();
     std::size_t offset = 0;
     // Rebuild shared fonts from data ncas
     if (nand->HasEntry(static_cast<u64>(FontArchives::Standard),
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
index d8b8037a8..c41ef7058 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
@@ -15,6 +15,11 @@
 #include "video_core/renderer_base.h"
 
 namespace Service::Nvidia::Devices {
+namespace NvErrCodes {
+enum {
+    InvalidNmapHandle = -22,
+};
+}
 
 nvhost_as_gpu::nvhost_as_gpu(std::shared_ptr<nvmap> nvmap_dev) : nvmap_dev(std::move(nvmap_dev)) {}
 nvhost_as_gpu::~nvhost_as_gpu() = default;
@@ -79,14 +84,16 @@ u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output)
     std::memcpy(entries.data(), input.data(), input.size());
 
     auto& gpu = Core::System::GetInstance().GPU();
-
     for (const auto& entry : entries) {
         LOG_WARNING(Service_NVDRV, "remap entry, offset=0x{:X} handle=0x{:X} pages=0x{:X}",
                     entry.offset, entry.nvmap_handle, entry.pages);
         Tegra::GPUVAddr offset = static_cast<Tegra::GPUVAddr>(entry.offset) << 0x10;
-
         auto object = nvmap_dev->GetObject(entry.nvmap_handle);
-        ASSERT(object);
+        if (!object) {
+            LOG_CRITICAL(Service_NVDRV, "nvmap {} is an invalid handle!", entry.nvmap_handle);
+            std::memcpy(output.data(), entries.data(), output.size());
+            return static_cast<u32>(NvErrCodes::InvalidNmapHandle);
+        }
 
         ASSERT(object->status == nvmap::Object::Status::Allocated);
 
@@ -157,15 +164,21 @@ u32 nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& ou
     LOG_DEBUG(Service_NVDRV, "called, offset=0x{:X}", params.offset);
 
     const auto itr = buffer_mappings.find(params.offset);
-    ASSERT_MSG(itr != buffer_mappings.end(), "Tried to unmap invalid mapping");
+    if (itr == buffer_mappings.end()) {
+        LOG_WARNING(Service_NVDRV, "Tried to unmap an invalid offset 0x{:X}", params.offset);
+        // Hardware tests shows that unmapping an already unmapped buffer always returns successful
+        // and doesn't fail.
+        return 0;
+    }
 
     auto& system_instance = Core::System::GetInstance();
 
     // Remove this memory region from the rasterizer cache.
-    system_instance.Renderer().Rasterizer().FlushAndInvalidateRegion(params.offset,
-                                                                     itr->second.size);
-
     auto& gpu = system_instance.GPU();
+    auto cpu_addr = gpu.MemoryManager().GpuToCpuAddress(params.offset);
+    ASSERT(cpu_addr);
+    system_instance.Renderer().Rasterizer().FlushAndInvalidateRegion(*cpu_addr, itr->second.size);
+
     params.offset = gpu.MemoryManager().UnmapBuffer(params.offset, itr->second.size);
 
     buffer_mappings.erase(itr->second.offset);
diff --git a/src/core/hle/service/nvdrv/devices/nvmap.cpp b/src/core/hle/service/nvdrv/devices/nvmap.cpp
index a2287cc1b..43651d8a6 100644
--- a/src/core/hle/service/nvdrv/devices/nvmap.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvmap.cpp
@@ -11,6 +11,13 @@
 
 namespace Service::Nvidia::Devices {
 
+namespace NvErrCodes {
+enum {
+    OperationNotPermitted = -1,
+    InvalidValue = -22,
+};
+}
+
 nvmap::nvmap() = default;
 nvmap::~nvmap() = default;
 
@@ -44,7 +51,11 @@ u32 nvmap::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& o
 u32 nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) {
     IocCreateParams params;
     std::memcpy(&params, input.data(), sizeof(params));
+    LOG_DEBUG(Service_NVDRV, "size=0x{:08X}", params.size);
 
+    if (!params.size) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
     // Create a new nvmap object and obtain a handle to it.
     auto object = std::make_shared<Object>();
     object->id = next_id++;
@@ -55,8 +66,6 @@ u32 nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) {
     u32 handle = next_handle++;
     handles[handle] = std::move(object);
 
-    LOG_DEBUG(Service_NVDRV, "size=0x{:08X}", params.size);
-
     params.handle = handle;
 
     std::memcpy(output.data(), &params, sizeof(params));
@@ -66,9 +75,29 @@ u32 nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) {
 u32 nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) {
     IocAllocParams params;
     std::memcpy(&params, input.data(), sizeof(params));
+    LOG_DEBUG(Service_NVDRV, "called, addr={:X}", params.addr);
+
+    if (!params.handle) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
+
+    if ((params.align - 1) & params.align) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
+
+    const u32 min_alignment = 0x1000;
+    if (params.align < min_alignment) {
+        params.align = min_alignment;
+    }
 
     auto object = GetObject(params.handle);
-    ASSERT(object);
+    if (!object) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
+
+    if (object->status == Object::Status::Allocated) {
+        return static_cast<u32>(NvErrCodes::OperationNotPermitted);
+    }
 
     object->flags = params.flags;
     object->align = params.align;
@@ -76,8 +105,6 @@ u32 nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) {
     object->addr = params.addr;
     object->status = Object::Status::Allocated;
 
-    LOG_DEBUG(Service_NVDRV, "called, addr={:X}", params.addr);
-
     std::memcpy(output.data(), &params, sizeof(params));
     return 0;
 }
@@ -88,8 +115,14 @@ u32 nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output) {
 
     LOG_WARNING(Service_NVDRV, "called");
 
+    if (!params.handle) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
+
     auto object = GetObject(params.handle);
-    ASSERT(object);
+    if (!object) {
+        return static_cast<u32>(NvErrCodes::OperationNotPermitted);
+    }
 
     params.id = object->id;
 
@@ -105,7 +138,14 @@ u32 nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) {
 
     auto itr = std::find_if(handles.begin(), handles.end(),
                             [&](const auto& entry) { return entry.second->id == params.id; });
-    ASSERT(itr != handles.end());
+    if (itr == handles.end()) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
+
+    auto& object = itr->second;
+    if (object->status != Object::Status::Allocated) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
 
     itr->second->refcount++;
 
@@ -125,8 +165,13 @@ u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_WARNING(Service_NVDRV, "(STUBBED) called type={}", params.param);
 
     auto object = GetObject(params.handle);
-    ASSERT(object);
-    ASSERT(object->status == Object::Status::Allocated);
+    if (!object) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
+
+    if (object->status != Object::Status::Allocated) {
+        return static_cast<u32>(NvErrCodes::OperationNotPermitted);
+    }
 
     switch (static_cast<ParamTypes>(params.param)) {
     case ParamTypes::Size:
@@ -163,9 +208,12 @@ u32 nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_WARNING(Service_NVDRV, "(STUBBED) called");
 
     auto itr = handles.find(params.handle);
-    ASSERT(itr != handles.end());
-
-    ASSERT(itr->second->refcount > 0);
+    if (itr == handles.end()) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
+    if (!itr->second->refcount) {
+        return static_cast<u32>(NvErrCodes::InvalidValue);
+    }
 
     itr->second->refcount--;
 
diff --git a/src/core/hle/service/nvflinger/buffer_queue.cpp b/src/core/hle/service/nvflinger/buffer_queue.cpp
index fd98d541d..630ebbfc7 100644
--- a/src/core/hle/service/nvflinger/buffer_queue.cpp
+++ b/src/core/hle/service/nvflinger/buffer_queue.cpp
@@ -31,7 +31,7 @@ void BufferQueue::SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer)
     buffer_wait_event->Signal();
 }
 
-boost::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) {
+std::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) {
     auto itr = std::find_if(queue.begin(), queue.end(), [&](const Buffer& buffer) {
         // Only consider free buffers. Buffers become free once again after they've been Acquired
         // and Released by the compositor, see the NVFlinger::Compose method.
@@ -44,7 +44,7 @@ boost::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) {
     });
 
     if (itr == queue.end()) {
-        return boost::none;
+        return {};
     }
 
     itr->status = Buffer::Status::Dequeued;
@@ -70,12 +70,12 @@ void BufferQueue::QueueBuffer(u32 slot, BufferTransformFlags transform,
     itr->crop_rect = crop_rect;
 }
 
-boost::optional<const BufferQueue::Buffer&> BufferQueue::AcquireBuffer() {
+std::optional<std::reference_wrapper<const BufferQueue::Buffer>> BufferQueue::AcquireBuffer() {
     auto itr = std::find_if(queue.begin(), queue.end(), [](const Buffer& buffer) {
         return buffer.status == Buffer::Status::Queued;
     });
     if (itr == queue.end())
-        return boost::none;
+        return {};
     itr->status = Buffer::Status::Acquired;
     return *itr;
 }
diff --git a/src/core/hle/service/nvflinger/buffer_queue.h b/src/core/hle/service/nvflinger/buffer_queue.h
index 50b767732..8cff5eb71 100644
--- a/src/core/hle/service/nvflinger/buffer_queue.h
+++ b/src/core/hle/service/nvflinger/buffer_queue.h
@@ -4,8 +4,9 @@
 
 #pragma once
 
+#include <optional>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/common_funcs.h"
 #include "common/math_util.h"
 #include "common/swap.h"
@@ -57,9 +58,9 @@ public:
         /// Rotate source image 90 degrees clockwise
         Rotate90 = 0x04,
         /// Rotate source image 180 degrees
-        Roate180 = 0x03,
+        Rotate180 = 0x03,
         /// Rotate source image 270 degrees clockwise
-        Roate270 = 0x07,
+        Rotate270 = 0x07,
     };
 
     struct Buffer {
@@ -73,11 +74,11 @@ public:
     };
 
     void SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer);
-    boost::optional<u32> DequeueBuffer(u32 width, u32 height);
+    std::optional<u32> DequeueBuffer(u32 width, u32 height);
     const IGBPBuffer& RequestBuffer(u32 slot) const;
     void QueueBuffer(u32 slot, BufferTransformFlags transform,
                      const MathUtil::Rectangle<int>& crop_rect);
-    boost::optional<const Buffer&> AcquireBuffer();
+    std::optional<std::reference_wrapper<const Buffer>> AcquireBuffer();
     void ReleaseBuffer(u32 slot);
     u32 Query(QueryType type);
 
diff --git a/src/core/hle/service/nvflinger/nvflinger.cpp b/src/core/hle/service/nvflinger/nvflinger.cpp
index d47b6f659..214e6d1b3 100644
--- a/src/core/hle/service/nvflinger/nvflinger.cpp
+++ b/src/core/hle/service/nvflinger/nvflinger.cpp
@@ -3,7 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
-#include <boost/optional.hpp>
+#include <optional>
 
 #include "common/alignment.h"
 #include "common/assert.h"
@@ -134,7 +134,7 @@ void NVFlinger::Compose() {
 
         MicroProfileFlip();
 
-        if (buffer == boost::none) {
+        if (!buffer) {
             auto& system_instance = Core::System::GetInstance();
 
             // There was no queued buffer to draw, render previous frame
@@ -143,7 +143,7 @@ void NVFlinger::Compose() {
             continue;
         }
 
-        auto& igbp_buffer = buffer->igbp_buffer;
+        auto& igbp_buffer = buffer->get().igbp_buffer;
 
         // Now send the buffer to the GPU for drawing.
         // TODO(Subv): Support more than just disp0. The display device selection is probably based
@@ -152,10 +152,10 @@ void NVFlinger::Compose() {
         ASSERT(nvdisp);
 
         nvdisp->flip(igbp_buffer.gpu_buffer_id, igbp_buffer.offset, igbp_buffer.format,
-                     igbp_buffer.width, igbp_buffer.height, igbp_buffer.stride, buffer->transform,
-                     buffer->crop_rect);
+                     igbp_buffer.width, igbp_buffer.height, igbp_buffer.stride,
+                     buffer->get().transform, buffer->get().crop_rect);
 
-        buffer_queue->ReleaseBuffer(buffer->slot);
+        buffer_queue->ReleaseBuffer(buffer->get().slot);
     }
 }
 
diff --git a/src/core/hle/service/prepo/prepo.cpp b/src/core/hle/service/prepo/prepo.cpp
index 6a9eccfb5..e4fcee9f8 100644
--- a/src/core/hle/service/prepo/prepo.cpp
+++ b/src/core/hle/service/prepo/prepo.cpp
@@ -14,20 +14,24 @@ public:
     explicit PlayReport(const char* name) : ServiceFramework{name} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {10100, nullptr, "SaveReport"},
-            {10101, &PlayReport::SaveReportWithUser, "SaveReportWithUser"},
+            {10100, nullptr, "SaveReportOld"},
+            {10101, &PlayReport::SaveReportWithUserOld, "SaveReportWithUserOld"},
+            {10102, nullptr, "SaveReport"},
+            {10103, nullptr, "SaveReportWithUser"},
             {10200, nullptr, "RequestImmediateTransmission"},
             {10300, nullptr, "GetTransmissionStatus"},
             {20100, nullptr, "SaveSystemReport"},
-            {20200, nullptr, "SetOperationMode"},
             {20101, nullptr, "SaveSystemReportWithUser"},
+            {20200, nullptr, "SetOperationMode"},
             {30100, nullptr, "ClearStorage"},
+            {30200, nullptr, "ClearStatistics"},
+            {30300, nullptr, "GetStorageUsage"},
+            {30400, nullptr, "GetStatistics"},
+            {30401, nullptr, "GetThroughputHistory"},
+            {30500, nullptr, "GetLastUploadError"},
             {40100, nullptr, "IsUserAgreementCheckEnabled"},
             {40101, nullptr, "SetUserAgreementCheckEnabled"},
-            {90100, nullptr, "GetStorageUsage"},
-            {90200, nullptr, "GetStatistics"},
-            {90201, nullptr, "GetThroughputHistory"},
-            {90300, nullptr, "GetLastUploadError"},
+            {90100, nullptr, "ReadAllReportFiles"},
         };
         // clang-format on
 
@@ -35,7 +39,7 @@ public:
     }
 
 private:
-    void SaveReportWithUser(Kernel::HLERequestContext& ctx) {
+    void SaveReportWithUserOld(Kernel::HLERequestContext& ctx) {
         // TODO(ogniK): Do we want to add play report?
         LOG_WARNING(Service_PREPO, "(STUBBED) called");
 
@@ -46,6 +50,7 @@ private:
 
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     std::make_shared<PlayReport>("prepo:a")->InstallAsService(service_manager);
+    std::make_shared<PlayReport>("prepo:a2")->InstallAsService(service_manager);
     std::make_shared<PlayReport>("prepo:m")->InstallAsService(service_manager);
     std::make_shared<PlayReport>("prepo:s")->InstallAsService(service_manager);
     std::make_shared<PlayReport>("prepo:u")->InstallAsService(service_manager);
diff --git a/src/core/hle/service/ptm/psm.cpp b/src/core/hle/service/ptm/psm.cpp
new file mode 100644
index 000000000..c2d5fda94
--- /dev/null
+++ b/src/core/hle/service/ptm/psm.cpp
@@ -0,0 +1,71 @@
+// 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/service/ptm/psm.h"
+#include "core/hle/service/service.h"
+#include "core/hle/service/sm/sm.h"
+
+namespace Service::PSM {
+
+constexpr u32 BATTERY_FULLY_CHARGED = 100;    // 100% Full
+constexpr u32 BATTERY_CURRENTLY_CHARGING = 1; // Plugged into an official dock
+
+class PSM final : public ServiceFramework<PSM> {
+public:
+    explicit PSM() : ServiceFramework{"psm"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &PSM::GetBatteryChargePercentage, "GetBatteryChargePercentage"},
+            {1, &PSM::GetChargerType, "GetChargerType"},
+            {2, nullptr, "EnableBatteryCharging"},
+            {3, nullptr, "DisableBatteryCharging"},
+            {4, nullptr, "IsBatteryChargingEnabled"},
+            {5, nullptr, "AcquireControllerPowerSupply"},
+            {6, nullptr, "ReleaseControllerPowerSupply"},
+            {7, nullptr, "OpenSession"},
+            {8, nullptr, "EnableEnoughPowerChargeEmulation"},
+            {9, nullptr, "DisableEnoughPowerChargeEmulation"},
+            {10, nullptr, "EnableFastBatteryCharging"},
+            {11, nullptr, "DisableFastBatteryCharging"},
+            {12, nullptr, "GetBatteryVoltageState"},
+            {13, nullptr, "GetRawBatteryChargePercentage"},
+            {14, nullptr, "IsEnoughPowerSupplied"},
+            {15, nullptr, "GetBatteryAgePercentage"},
+            {16, nullptr, "GetBatteryChargeInfoEvent"},
+            {17, nullptr, "GetBatteryChargeInfoFields"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+
+    ~PSM() override = default;
+
+private:
+    void GetBatteryChargePercentage(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_PSM, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(BATTERY_FULLY_CHARGED);
+    }
+
+    void GetChargerType(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_PSM, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(BATTERY_CURRENTLY_CHARGING);
+    }
+};
+
+void InstallInterfaces(SM::ServiceManager& sm) {
+    std::make_shared<PSM>()->InstallAsService(sm);
+}
+
+} // namespace Service::PSM
diff --git a/src/core/hle/service/ptm/psm.h b/src/core/hle/service/ptm/psm.h
new file mode 100644
index 000000000..a286793ae
--- /dev/null
+++ b/src/core/hle/service/ptm/psm.h
@@ -0,0 +1,15 @@
+// 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::PSM {
+
+void InstallInterfaces(SM::ServiceManager& sm);
+
+} // namespace Service::PSM
diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp
index 62f049660..a4cf45267 100644
--- a/src/core/hle/service/service.cpp
+++ b/src/core/hle/service/service.cpp
@@ -22,7 +22,7 @@
 #include "core/hle/service/apm/apm.h"
 #include "core/hle/service/arp/arp.h"
 #include "core/hle/service/audio/audio.h"
-#include "core/hle/service/bcat/bcat.h"
+#include "core/hle/service/bcat/module.h"
 #include "core/hle/service/bpc/bpc.h"
 #include "core/hle/service/btdrv/btdrv.h"
 #include "core/hle/service/btm/btm.h"
@@ -48,15 +48,17 @@
 #include "core/hle/service/nfp/nfp.h"
 #include "core/hle/service/nifm/nifm.h"
 #include "core/hle/service/nim/nim.h"
+#include "core/hle/service/npns/npns.h"
 #include "core/hle/service/ns/ns.h"
 #include "core/hle/service/nvdrv/nvdrv.h"
 #include "core/hle/service/nvflinger/nvflinger.h"
 #include "core/hle/service/pcie/pcie.h"
-#include "core/hle/service/pctl/pctl.h"
+#include "core/hle/service/pctl/module.h"
 #include "core/hle/service/pcv/pcv.h"
 #include "core/hle/service/pm/pm.h"
 #include "core/hle/service/prepo/prepo.h"
 #include "core/hle/service/psc/psc.h"
+#include "core/hle/service/ptm/psm.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/set/settings.h"
 #include "core/hle/service/sm/sm.h"
@@ -197,7 +199,7 @@ ResultCode ServiceFrameworkBase::HandleSyncRequest(Kernel::HLERequestContext& co
 // Module interface
 
 /// Initialize ServiceManager
-void Init(std::shared_ptr<SM::ServiceManager>& sm, const FileSys::VirtualFilesystem& rfs) {
+void Init(std::shared_ptr<SM::ServiceManager>& sm, FileSys::VfsFilesystem& vfs) {
     // NVFlinger needs to be accessed by several services like Vi and AppletOE so we instantiate it
     // here and pass it into the respective InstallInterfaces functions.
     auto nv_flinger = std::make_shared<NVFlinger::NVFlinger>();
@@ -220,7 +222,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, const FileSys::VirtualFilesys
     EUPLD::InstallInterfaces(*sm);
     Fatal::InstallInterfaces(*sm);
     FGM::InstallInterfaces(*sm);
-    FileSystem::InstallInterfaces(*sm, rfs);
+    FileSystem::InstallInterfaces(*sm, vfs);
     Friend::InstallInterfaces(*sm);
     GRC::InstallInterfaces(*sm);
     HID::InstallInterfaces(*sm);
@@ -236,6 +238,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, const FileSys::VirtualFilesys
     NFP::InstallInterfaces(*sm);
     NIFM::InstallInterfaces(*sm);
     NIM::InstallInterfaces(*sm);
+    NPNS::InstallInterfaces(*sm);
     NS::InstallInterfaces(*sm);
     Nvidia::InstallInterfaces(*sm, *nv_flinger);
     PCIe::InstallInterfaces(*sm);
@@ -244,6 +247,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, const FileSys::VirtualFilesys
     PlayReport::InstallInterfaces(*sm);
     PM::InstallInterfaces(*sm);
     PSC::InstallInterfaces(*sm);
+    PSM::InstallInterfaces(*sm);
     Set::InstallInterfaces(*sm);
     Sockets::InstallInterfaces(*sm);
     SPL::InstallInterfaces(*sm);
diff --git a/src/core/hle/service/service.h b/src/core/hle/service/service.h
index 2fc57a82e..98483ecf1 100644
--- a/src/core/hle/service/service.h
+++ b/src/core/hle/service/service.h
@@ -180,8 +180,7 @@ private:
 };
 
 /// Initialize ServiceManager
-void Init(std::shared_ptr<SM::ServiceManager>& sm,
-          const std::shared_ptr<FileSys::VfsFilesystem>& vfs);
+void Init(std::shared_ptr<SM::ServiceManager>& sm, FileSys::VfsFilesystem& vfs);
 
 /// Shutdown ServiceManager
 void Shutdown();
diff --git a/src/core/hle/service/set/set_cal.cpp b/src/core/hle/service/set/set_cal.cpp
index 5af356d10..34654bb07 100644
--- a/src/core/hle/service/set/set_cal.cpp
+++ b/src/core/hle/service/set/set_cal.cpp
@@ -39,7 +39,8 @@ SET_CAL::SET_CAL() : ServiceFramework("set:cal") {
         {29, nullptr, "GetAmiiboEcqvBlsKey"},
         {30, nullptr, "GetAmiiboEcqvBlsCertificate"},
         {31, nullptr, "GetAmiiboEcqvBlsRootCertificate"},
-        {32, nullptr, "GetUnknownId"},
+        {32, nullptr, "GetUsbTypeCPowerSourceCircuitVersion"},
+        {33, nullptr, "GetBatteryVersion"},
     };
     RegisterHandlers(functions);
 }
diff --git a/src/core/hle/service/usb/usb.cpp b/src/core/hle/service/usb/usb.cpp
index e7fb5a419..f0a831d45 100644
--- a/src/core/hle/service/usb/usb.cpp
+++ b/src/core/hle/service/usb/usb.cpp
@@ -67,15 +67,15 @@ public:
     explicit IClientEpSession() : ServiceFramework{"IClientEpSession"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "Unknown1"},
-            {1, nullptr, "Unknown2"},
-            {2, nullptr, "Unknown3"},
-            {3, nullptr, "Unknown4"},
+            {0, nullptr, "Open"},
+            {1, nullptr, "Close"},
+            {2, nullptr, "Unknown1"},
+            {3, nullptr, "Populate"},
             {4, nullptr, "PostBufferAsync"},
-            {5, nullptr, "Unknown5"},
-            {6, nullptr, "Unknown6"},
-            {7, nullptr, "Unknown7"},
-            {8, nullptr, "Unknown8"},
+            {5, nullptr, "GetXferReport"},
+            {6, nullptr, "Unknown2"},
+            {7, nullptr, "Unknown3"},
+            {8, nullptr, "Unknown4"},
         };
         // clang-format on
 
@@ -89,15 +89,15 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "Unknown1"},
-            {1, nullptr, "Unknown2"},
-            {2, nullptr, "Unknown3"},
-            {3, nullptr, "Unknown4"},
-            {4, nullptr, "Unknown5"},
+            {1, nullptr, "SetInterface"},
+            {2, nullptr, "GetInterface"},
+            {3, nullptr, "GetAlternateInterface"},
+            {4, nullptr, "GetCurrentFrame"},
             {5, nullptr, "CtrlXferAsync"},
-            {6, nullptr, "Unknown6"},
+            {6, nullptr, "Unknown2"},
             {7, nullptr, "GetCtrlXferReport"},
-            {8, nullptr, "Unknown7"},
-            {9, nullptr, "GetClientEpSession"},
+            {8, nullptr, "ResetDevice"},
+            {9, nullptr, "OpenUsbEp"},
         };
         // clang-format on
 
@@ -111,13 +111,14 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "BindClientProcess"},
-            {1, nullptr, "Unknown1"},
-            {2, nullptr, "Unknown2"},
-            {3, nullptr, "Unknown3"},
-            {4, nullptr, "Unknown4"},
-            {5, nullptr, "Unknown5"},
+            {1, nullptr, "QueryAllInterfaces"},
+            {2, nullptr, "QueryAvailableInterfaces"},
+            {3, nullptr, "QueryAcquiredInterfaces"},
+            {4, nullptr, "CreateInterfaceAvailableEvent"},
+            {5, nullptr, "DestroyInterfaceAvailableEvent"},
             {6, nullptr, "GetInterfaceStateChangeEvent"},
-            {7, nullptr, "GetClientIfSession"},
+            {7, nullptr, "AcquireUsbIf"},
+            {8, nullptr, "Unknown1"},
         };
         // clang-format on
 
@@ -131,11 +132,11 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "BindNoticeEvent"},
-            {1, nullptr, "Unknown1"},
+            {1, nullptr, "UnbindNoticeEvent"},
             {2, nullptr, "GetStatus"},
             {3, nullptr, "GetNotice"},
-            {4, nullptr, "Unknown2"},
-            {5, nullptr, "Unknown3"},
+            {4, nullptr, "EnablePowerRequestNotice"},
+            {5, nullptr, "DisablePowerRequestNotice"},
             {6, nullptr, "ReplyPowerRequest"},
         };
         // clang-format on
diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp
index bbc02abcc..d764b2406 100644
--- a/src/core/hle/service/vi/vi.cpp
+++ b/src/core/hle/service/vi/vi.cpp
@@ -6,9 +6,10 @@
 #include <array>
 #include <cstring>
 #include <memory>
+#include <optional>
 #include <type_traits>
 #include <utility>
-#include <boost/optional.hpp>
+
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "common/common_funcs.h"
@@ -506,9 +507,9 @@ private:
             IGBPDequeueBufferRequestParcel request{ctx.ReadBuffer()};
             const u32 width{request.data.width};
             const u32 height{request.data.height};
-            boost::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
+            std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
 
-            if (slot != boost::none) {
+            if (slot) {
                 // Buffer is available
                 IGBPDequeueBufferResponseParcel response{*slot};
                 ctx.WriteBuffer(response.Serialize());
@@ -520,7 +521,7 @@ private:
                         Kernel::ThreadWakeupReason reason) {
                         // Repeat TransactParcel DequeueBuffer when a buffer is available
                         auto buffer_queue = nv_flinger->GetBufferQueue(id);
-                        boost::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
+                        std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
                         IGBPDequeueBufferResponseParcel response{*slot};
                         ctx.WriteBuffer(response.Serialize());
                         IPC::ResponseBuilder rb{ctx, 2};
@@ -968,6 +969,54 @@ private:
         rb.PushCopyObjects(vsync_event);
     }
 
+    enum class ConvertedScaleMode : u64 {
+        None = 0, // VI seems to name this as "Unknown" but lots of games pass it, assume it's no
+                  // scaling/default
+        Freeze = 1,
+        ScaleToWindow = 2,
+        Crop = 3,
+        NoCrop = 4,
+    };
+
+    // This struct is different, currently it's 1:1 but this might change in the future.
+    enum class NintendoScaleMode : u32 {
+        None = 0,
+        Freeze = 1,
+        ScaleToWindow = 2,
+        Crop = 3,
+        NoCrop = 4,
+    };
+
+    void ConvertScalingMode(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        auto mode = rp.PopEnum<NintendoScaleMode>();
+        LOG_DEBUG(Service_VI, "called mode={}", static_cast<u32>(mode));
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        switch (mode) {
+        case NintendoScaleMode::None:
+            rb.PushEnum(ConvertedScaleMode::None);
+            break;
+        case NintendoScaleMode::Freeze:
+            rb.PushEnum(ConvertedScaleMode::Freeze);
+            break;
+        case NintendoScaleMode::ScaleToWindow:
+            rb.PushEnum(ConvertedScaleMode::ScaleToWindow);
+            break;
+        case NintendoScaleMode::Crop:
+            rb.PushEnum(ConvertedScaleMode::Crop);
+            break;
+        case NintendoScaleMode::NoCrop:
+            rb.PushEnum(ConvertedScaleMode::NoCrop);
+            break;
+        default:
+            UNIMPLEMENTED_MSG("Unknown scaling mode {}", static_cast<u32>(mode));
+            rb.PushEnum(ConvertedScaleMode::None);
+            break;
+        }
+    }
+
     std::shared_ptr<NVFlinger::NVFlinger> nv_flinger;
 };
 
@@ -991,7 +1040,7 @@ IApplicationDisplayService::IApplicationDisplayService(
         {2030, &IApplicationDisplayService::CreateStrayLayer, "CreateStrayLayer"},
         {2031, &IApplicationDisplayService::DestroyStrayLayer, "DestroyStrayLayer"},
         {2101, &IApplicationDisplayService::SetLayerScalingMode, "SetLayerScalingMode"},
-        {2102, nullptr, "ConvertScalingMode"},
+        {2102, &IApplicationDisplayService::ConvertScalingMode, "ConvertScalingMode"},
         {2450, nullptr, "GetIndirectLayerImageMap"},
         {2451, nullptr, "GetIndirectLayerImageCropMap"},
         {2460, nullptr, "GetIndirectLayerImageRequiredMemoryInfo"},
diff --git a/src/core/loader/deconstructed_rom_directory.cpp b/src/core/loader/deconstructed_rom_directory.cpp
index 9a86e5824..8518dddcb 100644
--- a/src/core/loader/deconstructed_rom_directory.cpp
+++ b/src/core/loader/deconstructed_rom_directory.cpp
@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <cinttypes>
+#include <cstring>
 #include "common/common_funcs.h"
 #include "common/file_util.h"
 #include "common/logging/log.h"
@@ -138,13 +139,22 @@ ResultStatus AppLoader_DeconstructedRomDirectory::Load(Kernel::Process& process)
     for (const auto& module : {"rtld", "main", "subsdk0", "subsdk1", "subsdk2", "subsdk3",
                                "subsdk4", "subsdk5", "subsdk6", "subsdk7", "sdk"}) {
         const FileSys::VirtualFile module_file = dir->GetFile(module);
-        if (module_file != nullptr) {
-            const VAddr load_addr = next_load_addr;
-            next_load_addr = AppLoader_NSO::LoadModule(module_file, load_addr, pm);
-            LOG_DEBUG(Loader, "loaded module {} @ 0x{:X}", module, load_addr);
-            // Register module with GDBStub
-            GDBStub::RegisterModule(module, load_addr, next_load_addr - 1, false);
+        if (module_file == nullptr) {
+            continue;
         }
+
+        const VAddr load_addr = next_load_addr;
+        const bool should_pass_arguments = std::strcmp(module, "rtld") == 0;
+        const auto tentative_next_load_addr =
+            AppLoader_NSO::LoadModule(*module_file, load_addr, should_pass_arguments, pm);
+        if (!tentative_next_load_addr) {
+            return ResultStatus::ErrorLoadingNSO;
+        }
+
+        next_load_addr = *tentative_next_load_addr;
+        LOG_DEBUG(Loader, "loaded module {} @ 0x{:X}", module, load_addr);
+        // Register module with GDBStub
+        GDBStub::RegisterModule(module, load_addr, next_load_addr - 1, false);
     }
 
     process.Run(base_address, metadata.GetMainThreadPriority(), metadata.GetMainThreadStackSize());
diff --git a/src/core/loader/elf.cpp b/src/core/loader/elf.cpp
index e67b49fc9..6057c7f26 100644
--- a/src/core/loader/elf.cpp
+++ b/src/core/loader/elf.cpp
@@ -9,16 +9,11 @@
 #include "common/common_types.h"
 #include "common/file_util.h"
 #include "common/logging/log.h"
-#include "core/core.h"
-#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/vm_manager.h"
 #include "core/loader/elf.h"
 #include "core/memory.h"
 
-using Kernel::CodeSet;
-using Kernel::SharedPtr;
-
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 // ELF Header Constants
 
@@ -211,7 +206,7 @@ public:
     u32 GetFlags() const {
         return (u32)(header->e_flags);
     }
-    SharedPtr<CodeSet> LoadInto(VAddr vaddr);
+    Kernel::CodeSet LoadInto(VAddr vaddr);
 
     int GetNumSegments() const {
         return (int)(header->e_phnum);
@@ -274,7 +269,7 @@ const char* ElfReader::GetSectionName(int section) const {
     return nullptr;
 }
 
-SharedPtr<CodeSet> ElfReader::LoadInto(VAddr vaddr) {
+Kernel::CodeSet ElfReader::LoadInto(VAddr vaddr) {
     LOG_DEBUG(Loader, "String section: {}", header->e_shstrndx);
 
     // Should we relocate?
@@ -302,8 +297,7 @@ SharedPtr<CodeSet> ElfReader::LoadInto(VAddr vaddr) {
     std::vector<u8> program_image(total_image_size);
     std::size_t current_image_position = 0;
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<CodeSet> codeset = CodeSet::Create(kernel, "");
+    Kernel::CodeSet codeset;
 
     for (unsigned int i = 0; i < header->e_phnum; ++i) {
         const Elf32_Phdr* p = &segments[i];
@@ -311,14 +305,14 @@ SharedPtr<CodeSet> ElfReader::LoadInto(VAddr vaddr) {
                   p->p_vaddr, p->p_filesz, p->p_memsz);
 
         if (p->p_type == PT_LOAD) {
-            CodeSet::Segment* codeset_segment;
+            Kernel::CodeSet::Segment* codeset_segment;
             u32 permission_flags = p->p_flags & (PF_R | PF_W | PF_X);
             if (permission_flags == (PF_R | PF_X)) {
-                codeset_segment = &codeset->CodeSegment();
+                codeset_segment = &codeset.CodeSegment();
             } else if (permission_flags == (PF_R)) {
-                codeset_segment = &codeset->RODataSegment();
+                codeset_segment = &codeset.RODataSegment();
             } else if (permission_flags == (PF_R | PF_W)) {
-                codeset_segment = &codeset->DataSegment();
+                codeset_segment = &codeset.DataSegment();
             } else {
                 LOG_ERROR(Loader, "Unexpected ELF PT_LOAD segment id {} with flags {:X}", i,
                           p->p_flags);
@@ -345,8 +339,8 @@ SharedPtr<CodeSet> ElfReader::LoadInto(VAddr vaddr) {
         }
     }
 
-    codeset->entrypoint = base_addr + header->e_entry;
-    codeset->memory = std::make_shared<std::vector<u8>>(std::move(program_image));
+    codeset.entrypoint = base_addr + header->e_entry;
+    codeset.memory = std::make_shared<std::vector<u8>>(std::move(program_image));
 
     LOG_DEBUG(Loader, "Done loading.");
 
@@ -397,11 +391,11 @@ ResultStatus AppLoader_ELF::Load(Kernel::Process& process) {
 
     const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress();
     ElfReader elf_reader(&buffer[0]);
-    SharedPtr<CodeSet> codeset = elf_reader.LoadInto(base_address);
-    codeset->name = file->GetName();
+    Kernel::CodeSet codeset = elf_reader.LoadInto(base_address);
+    const VAddr entry_point = codeset.entrypoint;
 
-    process.LoadModule(codeset, codeset->entrypoint);
-    process.Run(codeset->entrypoint, 48, Memory::DEFAULT_STACK_SIZE);
+    process.LoadModule(std::move(codeset), entry_point);
+    process.Run(entry_point, 48, Memory::DEFAULT_STACK_SIZE);
 
     is_loaded = true;
     return ResultStatus::Success;
diff --git a/src/core/loader/loader.cpp b/src/core/loader/loader.cpp
index f2a183ba1..9cd0b0ccd 100644
--- a/src/core/loader/loader.cpp
+++ b/src/core/loader/loader.cpp
@@ -93,7 +93,7 @@ std::string GetFileTypeString(FileType type) {
     return "unknown";
 }
 
-constexpr std::array<const char*, 58> RESULT_MESSAGES{
+constexpr std::array<const char*, 60> RESULT_MESSAGES{
     "The operation completed successfully.",
     "The loader requested to load is already loaded.",
     "The operation is not implemented.",
@@ -128,6 +128,7 @@ constexpr std::array<const char*, 58> RESULT_MESSAGES{
     "The RomFS could not be found.",
     "The ELF file has incorrect size as determined by the header.",
     "There was a general error loading the NRO into emulated memory.",
+    "There was a general error loading the NSO into emulated memory.",
     "There is no icon available.",
     "There is no control data available.",
     "The NAX file has a bad header.",
@@ -152,6 +153,7 @@ constexpr std::array<const char*, 58> RESULT_MESSAGES{
     "The BKTR-type NCA has a bad Relocation bucket.",
     "The BKTR-type NCA has a bad Subsection bucket.",
     "The BKTR-type NCA is missing the base RomFS.",
+    "The NSP or XCI does not contain an update in addition to the base game.",
 };
 
 std::ostream& operator<<(std::ostream& os, ResultStatus status) {
diff --git a/src/core/loader/loader.h b/src/core/loader/loader.h
index 20e66109b..7686634bf 100644
--- a/src/core/loader/loader.h
+++ b/src/core/loader/loader.h
@@ -6,10 +6,11 @@
 
 #include <iosfwd>
 #include <memory>
+#include <optional>
 #include <string>
 #include <utility>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/common_types.h"
 #include "core/file_sys/vfs.h"
 
@@ -90,6 +91,7 @@ enum class ResultStatus : u16 {
     ErrorNoRomFS,
     ErrorIncorrectELFFileSize,
     ErrorLoadingNRO,
+    ErrorLoadingNSO,
     ErrorNoIcon,
     ErrorNoControl,
     ErrorBadNAXHeader,
@@ -114,6 +116,7 @@ enum class ResultStatus : u16 {
     ErrorBadRelocationBuckets,
     ErrorBadSubsectionBuckets,
     ErrorMissingBKTRBaseRomFS,
+    ErrorNoPackedUpdate,
 };
 
 std::ostream& operator<<(std::ostream& os, ResultStatus status);
@@ -143,7 +146,7 @@ public:
      * information.
      * @returns A pair with the optional system mode, and and the status.
      */
-    virtual std::pair<boost::optional<u32>, ResultStatus> LoadKernelSystemMode() {
+    virtual std::pair<std::optional<u32>, ResultStatus> LoadKernelSystemMode() {
         // 96MB allocated to the application.
         return std::make_pair(2, ResultStatus::Success);
     }
@@ -196,10 +199,19 @@ public:
     /**
      * Get the RomFS of the application
      * Since the RomFS can be huge, we return a file reference instead of copying to a buffer
-     * @param dir The directory containing the RomFS
+     * @param file The directory containing the RomFS
+     * @return ResultStatus result of function
+     */
+    virtual ResultStatus ReadRomFS(FileSys::VirtualFile& file) {
+        return ResultStatus::ErrorNotImplemented;
+    }
+
+    /**
+     * Get the raw update of the application, should it come packed with one
+     * @param file The raw update NCA file (Program-type
      * @return ResultStatus result of function
      */
-    virtual ResultStatus ReadRomFS(FileSys::VirtualFile& dir) {
+    virtual ResultStatus ReadUpdateRaw(FileSys::VirtualFile& file) {
         return ResultStatus::ErrorNotImplemented;
     }
 
diff --git a/src/core/loader/nax.cpp b/src/core/loader/nax.cpp
index 073fb9d2f..42f4a777b 100644
--- a/src/core/loader/nax.cpp
+++ b/src/core/loader/nax.cpp
@@ -72,6 +72,10 @@ ResultStatus AppLoader_NAX::ReadRomFS(FileSys::VirtualFile& dir) {
     return nca_loader->ReadRomFS(dir);
 }
 
+u64 AppLoader_NAX::ReadRomFSIVFCOffset() const {
+    return nca_loader->ReadRomFSIVFCOffset();
+}
+
 ResultStatus AppLoader_NAX::ReadProgramId(u64& out_program_id) {
     return nca_loader->ReadProgramId(out_program_id);
 }
diff --git a/src/core/loader/nax.h b/src/core/loader/nax.h
index fc3c01876..b4d93bd01 100644
--- a/src/core/loader/nax.h
+++ b/src/core/loader/nax.h
@@ -36,6 +36,7 @@ public:
     ResultStatus Load(Kernel::Process& process) override;
 
     ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
+    u64 ReadRomFSIVFCOffset() const override;
     ResultStatus ReadProgramId(u64& out_program_id) override;
 
 private:
diff --git a/src/core/loader/nro.cpp b/src/core/loader/nro.cpp
index c10f826a4..bc8e402a8 100644
--- a/src/core/loader/nro.cpp
+++ b/src/core/loader/nro.cpp
@@ -14,11 +14,12 @@
 #include "core/file_sys/control_metadata.h"
 #include "core/file_sys/vfs_offset.h"
 #include "core/gdbstub/gdbstub.h"
-#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/vm_manager.h"
 #include "core/loader/nro.h"
+#include "core/loader/nso.h"
 #include "core/memory.h"
+#include "core/settings.h"
 
 namespace Loader {
 
@@ -126,28 +127,45 @@ static constexpr u32 PageAlignSize(u32 size) {
     return (size + Memory::PAGE_MASK) & ~Memory::PAGE_MASK;
 }
 
-bool AppLoader_NRO::LoadNro(FileSys::VirtualFile file, VAddr load_base) {
-    // Read NSO header
-    NroHeader nro_header{};
-    if (sizeof(NroHeader) != file->ReadObject(&nro_header)) {
+/*static*/ bool AppLoader_NRO::LoadNro(const std::vector<u8>& data, const std::string& name,
+                                       VAddr load_base) {
+
+    if (data.size() < sizeof(NroHeader)) {
         return {};
     }
+
+    // Read NSO header
+    NroHeader nro_header{};
+    std::memcpy(&nro_header, data.data(), sizeof(NroHeader));
     if (nro_header.magic != Common::MakeMagic('N', 'R', 'O', '0')) {
         return {};
     }
 
     // Build program image
-    auto& kernel = Core::System::GetInstance().Kernel();
-    Kernel::SharedPtr<Kernel::CodeSet> codeset = Kernel::CodeSet::Create(kernel, "");
-    std::vector<u8> program_image = file->ReadBytes(PageAlignSize(nro_header.file_size));
+    std::vector<u8> program_image(PageAlignSize(nro_header.file_size));
+    std::memcpy(program_image.data(), data.data(), program_image.size());
     if (program_image.size() != PageAlignSize(nro_header.file_size)) {
         return {};
     }
 
+    Kernel::CodeSet codeset;
     for (std::size_t i = 0; i < nro_header.segments.size(); ++i) {
-        codeset->segments[i].addr = nro_header.segments[i].offset;
-        codeset->segments[i].offset = nro_header.segments[i].offset;
-        codeset->segments[i].size = PageAlignSize(nro_header.segments[i].size);
+        codeset.segments[i].addr = nro_header.segments[i].offset;
+        codeset.segments[i].offset = nro_header.segments[i].offset;
+        codeset.segments[i].size = PageAlignSize(nro_header.segments[i].size);
+    }
+
+    if (!Settings::values.program_args.empty()) {
+        const auto arg_data = Settings::values.program_args;
+        codeset.DataSegment().size += NSO_ARGUMENT_DATA_ALLOCATION_SIZE;
+        NSOArgumentHeader args_header{
+            NSO_ARGUMENT_DATA_ALLOCATION_SIZE, static_cast<u32_le>(arg_data.size()), {}};
+        const auto end_offset = program_image.size();
+        program_image.resize(static_cast<u32>(program_image.size()) +
+                             NSO_ARGUMENT_DATA_ALLOCATION_SIZE);
+        std::memcpy(program_image.data() + end_offset, &args_header, sizeof(NSOArgumentHeader));
+        std::memcpy(program_image.data() + end_offset + sizeof(NSOArgumentHeader), arg_data.data(),
+                    arg_data.size());
     }
 
     // Read MOD header
@@ -161,20 +179,23 @@ bool AppLoader_NRO::LoadNro(FileSys::VirtualFile file, VAddr load_base) {
         // Resize program image to include .bss section and page align each section
         bss_size = PageAlignSize(mod_header.bss_end_offset - mod_header.bss_start_offset);
     }
-    codeset->DataSegment().size += bss_size;
+    codeset.DataSegment().size += bss_size;
     program_image.resize(static_cast<u32>(program_image.size()) + bss_size);
 
     // Load codeset for current process
-    codeset->name = file->GetName();
-    codeset->memory = std::make_shared<std::vector<u8>>(std::move(program_image));
-    Core::CurrentProcess()->LoadModule(codeset, load_base);
+    codeset.memory = std::make_shared<std::vector<u8>>(std::move(program_image));
+    Core::CurrentProcess()->LoadModule(std::move(codeset), load_base);
 
     // Register module with GDBStub
-    GDBStub::RegisterModule(codeset->name, load_base, load_base);
+    GDBStub::RegisterModule(name, load_base, load_base);
 
     return true;
 }
 
+bool AppLoader_NRO::LoadNro(const FileSys::VfsFile& file, VAddr load_base) {
+    return AppLoader_NRO::LoadNro(file.ReadAllBytes(), file.GetName(), load_base);
+}
+
 ResultStatus AppLoader_NRO::Load(Kernel::Process& process) {
     if (is_loaded) {
         return ResultStatus::ErrorAlreadyLoaded;
@@ -183,7 +204,7 @@ ResultStatus AppLoader_NRO::Load(Kernel::Process& process) {
     // Load NRO
     const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress();
 
-    if (!LoadNro(file, base_address)) {
+    if (!LoadNro(*file, base_address)) {
         return ResultStatus::ErrorLoadingNRO;
     }
 
diff --git a/src/core/loader/nro.h b/src/core/loader/nro.h
index 04b46119a..3e6959302 100644
--- a/src/core/loader/nro.h
+++ b/src/core/loader/nro.h
@@ -5,6 +5,7 @@
 #pragma once
 
 #include <string>
+#include <vector>
 #include "common/common_types.h"
 #include "core/loader/linker.h"
 #include "core/loader/loader.h"
@@ -40,8 +41,10 @@ public:
     ResultStatus ReadTitle(std::string& title) override;
     bool IsRomFSUpdatable() const override;
 
+    static bool LoadNro(const std::vector<u8>& data, const std::string& name, VAddr load_base);
+
 private:
-    bool LoadNro(FileSys::VirtualFile file, VAddr load_base);
+    bool LoadNro(const FileSys::VfsFile& file, VAddr load_base);
 
     std::vector<u8> icon_data;
     std::unique_ptr<FileSys::NACP> nacp;
diff --git a/src/core/loader/nso.cpp b/src/core/loader/nso.cpp
index 2186b02af..68efca5c0 100644
--- a/src/core/loader/nso.cpp
+++ b/src/core/loader/nso.cpp
@@ -12,11 +12,11 @@
 #include "core/core.h"
 #include "core/file_sys/patch_manager.h"
 #include "core/gdbstub/gdbstub.h"
-#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/vm_manager.h"
 #include "core/loader/nso.h"
 #include "core/memory.h"
+#include "core/settings.h"
 
 namespace Loader {
 
@@ -93,36 +93,46 @@ static constexpr u32 PageAlignSize(u32 size) {
     return (size + Memory::PAGE_MASK) & ~Memory::PAGE_MASK;
 }
 
-VAddr AppLoader_NSO::LoadModule(FileSys::VirtualFile file, VAddr load_base,
-                                boost::optional<FileSys::PatchManager> pm) {
-    if (file == nullptr)
-        return {};
-
-    if (file->GetSize() < sizeof(NsoHeader))
+std::optional<VAddr> AppLoader_NSO::LoadModule(const FileSys::VfsFile& file, VAddr load_base,
+                                               bool should_pass_arguments,
+                                               std::optional<FileSys::PatchManager> pm) {
+    if (file.GetSize() < sizeof(NsoHeader))
         return {};
 
     NsoHeader nso_header{};
-    if (sizeof(NsoHeader) != file->ReadObject(&nso_header))
+    if (sizeof(NsoHeader) != file.ReadObject(&nso_header))
         return {};
 
     if (nso_header.magic != Common::MakeMagic('N', 'S', 'O', '0'))
         return {};
 
     // Build program image
-    auto& kernel = Core::System::GetInstance().Kernel();
-    Kernel::SharedPtr<Kernel::CodeSet> codeset = Kernel::CodeSet::Create(kernel, "");
+    Kernel::CodeSet codeset;
     std::vector<u8> program_image;
     for (std::size_t i = 0; i < nso_header.segments.size(); ++i) {
         std::vector<u8> data =
-            file->ReadBytes(nso_header.segments_compressed_size[i], nso_header.segments[i].offset);
+            file.ReadBytes(nso_header.segments_compressed_size[i], nso_header.segments[i].offset);
         if (nso_header.IsSegmentCompressed(i)) {
             data = DecompressSegment(data, nso_header.segments[i]);
         }
         program_image.resize(nso_header.segments[i].location);
         program_image.insert(program_image.end(), data.begin(), data.end());
-        codeset->segments[i].addr = nso_header.segments[i].location;
-        codeset->segments[i].offset = nso_header.segments[i].location;
-        codeset->segments[i].size = PageAlignSize(static_cast<u32>(data.size()));
+        codeset.segments[i].addr = nso_header.segments[i].location;
+        codeset.segments[i].offset = nso_header.segments[i].location;
+        codeset.segments[i].size = PageAlignSize(static_cast<u32>(data.size()));
+    }
+
+    if (should_pass_arguments && !Settings::values.program_args.empty()) {
+        const auto arg_data = Settings::values.program_args;
+        codeset.DataSegment().size += NSO_ARGUMENT_DATA_ALLOCATION_SIZE;
+        NSOArgumentHeader args_header{
+            NSO_ARGUMENT_DATA_ALLOCATION_SIZE, static_cast<u32_le>(arg_data.size()), {}};
+        const auto end_offset = program_image.size();
+        program_image.resize(static_cast<u32>(program_image.size()) +
+                             NSO_ARGUMENT_DATA_ALLOCATION_SIZE);
+        std::memcpy(program_image.data() + end_offset, &args_header, sizeof(NSOArgumentHeader));
+        std::memcpy(program_image.data() + end_offset + sizeof(NSOArgumentHeader), arg_data.data(),
+                    arg_data.size());
     }
 
     // MOD header pointer is at .text offset + 4
@@ -139,12 +149,12 @@ VAddr AppLoader_NSO::LoadModule(FileSys::VirtualFile file, VAddr load_base,
         // Resize program image to include .bss section and page align each section
         bss_size = PageAlignSize(mod_header.bss_end_offset - mod_header.bss_start_offset);
     }
-    codeset->DataSegment().size += bss_size;
+    codeset.DataSegment().size += bss_size;
     const u32 image_size{PageAlignSize(static_cast<u32>(program_image.size()) + bss_size)};
     program_image.resize(image_size);
 
     // Apply patches if necessary
-    if (pm != boost::none && pm->HasNSOPatch(nso_header.build_id)) {
+    if (pm && pm->HasNSOPatch(nso_header.build_id)) {
         std::vector<u8> pi_header(program_image.size() + 0x100);
         std::memcpy(pi_header.data(), &nso_header, sizeof(NsoHeader));
         std::memcpy(pi_header.data() + 0x100, program_image.data(), program_image.size());
@@ -155,12 +165,11 @@ VAddr AppLoader_NSO::LoadModule(FileSys::VirtualFile file, VAddr load_base,
     }
 
     // Load codeset for current process
-    codeset->name = file->GetName();
-    codeset->memory = std::make_shared<std::vector<u8>>(std::move(program_image));
-    Core::CurrentProcess()->LoadModule(codeset, load_base);
+    codeset.memory = std::make_shared<std::vector<u8>>(std::move(program_image));
+    Core::CurrentProcess()->LoadModule(std::move(codeset), load_base);
 
     // Register module with GDBStub
-    GDBStub::RegisterModule(codeset->name, load_base, load_base);
+    GDBStub::RegisterModule(file.GetName(), load_base, load_base);
 
     return load_base + image_size;
 }
@@ -172,7 +181,9 @@ ResultStatus AppLoader_NSO::Load(Kernel::Process& process) {
 
     // Load module
     const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress();
-    LoadModule(file, base_address);
+    if (!LoadModule(*file, base_address, true)) {
+        return ResultStatus::ErrorLoadingNSO;
+    }
     LOG_DEBUG(Loader, "loaded module {} @ 0x{:X}", file->GetName(), base_address);
 
     process.Run(base_address, Kernel::THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE);
diff --git a/src/core/loader/nso.h b/src/core/loader/nso.h
index 05353d4d9..433306139 100644
--- a/src/core/loader/nso.h
+++ b/src/core/loader/nso.h
@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include <optional>
 #include "common/common_types.h"
 #include "core/file_sys/patch_manager.h"
 #include "core/loader/linker.h"
@@ -11,6 +12,15 @@
 
 namespace Loader {
 
+constexpr u64 NSO_ARGUMENT_DATA_ALLOCATION_SIZE = 0x9000;
+
+struct NSOArgumentHeader {
+    u32_le allocated_size;
+    u32_le actual_size;
+    INSERT_PADDING_BYTES(0x18);
+};
+static_assert(sizeof(NSOArgumentHeader) == 0x20, "NSOArgumentHeader has incorrect size.");
+
 /// Loads an NSO file
 class AppLoader_NSO final : public AppLoader, Linker {
 public:
@@ -27,8 +37,9 @@ public:
         return IdentifyType(file);
     }
 
-    static VAddr LoadModule(FileSys::VirtualFile file, VAddr load_base,
-                            boost::optional<FileSys::PatchManager> pm = boost::none);
+    static std::optional<VAddr> LoadModule(const FileSys::VfsFile& file, VAddr load_base,
+                                           bool should_pass_arguments,
+                                           std::optional<FileSys::PatchManager> pm = {});
 
     ResultStatus Load(Kernel::Process& process) override;
 };
diff --git a/src/core/loader/nsp.cpp b/src/core/loader/nsp.cpp
index b7ba77ef4..080d89904 100644
--- a/src/core/loader/nsp.cpp
+++ b/src/core/loader/nsp.cpp
@@ -10,8 +10,10 @@
 #include "core/file_sys/control_metadata.h"
 #include "core/file_sys/nca_metadata.h"
 #include "core/file_sys/patch_manager.h"
+#include "core/file_sys/registered_cache.h"
 #include "core/file_sys/submission_package.h"
 #include "core/hle/kernel/process.h"
+#include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/deconstructed_rom_directory.h"
 #include "core/loader/nca.h"
 #include "core/loader/nsp.h"
@@ -33,7 +35,17 @@ AppLoader_NSP::AppLoader_NSP(FileSys::VirtualFile file)
         return;
 
     std::tie(nacp_file, icon_file) =
-        FileSys::PatchManager(nsp->GetProgramTitleID()).ParseControlNCA(control_nca);
+        FileSys::PatchManager(nsp->GetProgramTitleID()).ParseControlNCA(*control_nca);
+
+    if (nsp->IsExtractedType()) {
+        secondary_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(nsp->GetExeFS());
+    } else {
+        if (title_id == 0)
+            return;
+
+        secondary_loader = std::make_unique<AppLoader_NCA>(
+            nsp->GetNCAFile(title_id, FileSys::ContentRecordType::Program));
+    }
 }
 
 AppLoader_NSP::~AppLoader_NSP() = default;
@@ -65,39 +77,58 @@ ResultStatus AppLoader_NSP::Load(Kernel::Process& process) {
         return ResultStatus::ErrorAlreadyLoaded;
     }
 
-    if (nsp->IsExtractedType()) {
-        secondary_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(nsp->GetExeFS());
-    } else {
-        if (title_id == 0)
-            return ResultStatus::ErrorNSPMissingProgramNCA;
-
-        secondary_loader = std::make_unique<AppLoader_NCA>(
-            nsp->GetNCAFile(title_id, FileSys::ContentRecordType::Program));
+    if (title_id == 0)
+        return ResultStatus::ErrorNSPMissingProgramNCA;
 
-        if (nsp->GetStatus() != ResultStatus::Success)
-            return nsp->GetStatus();
+    if (nsp->GetStatus() != ResultStatus::Success)
+        return nsp->GetStatus();
 
-        if (nsp->GetProgramStatus(title_id) != ResultStatus::Success)
-            return nsp->GetProgramStatus(title_id);
+    if (nsp->GetProgramStatus(title_id) != ResultStatus::Success)
+        return nsp->GetProgramStatus(title_id);
 
-        if (nsp->GetNCA(title_id, FileSys::ContentRecordType::Program) == nullptr) {
-            if (!Core::Crypto::KeyManager::KeyFileExists(false))
-                return ResultStatus::ErrorMissingProductionKeyFile;
-            return ResultStatus::ErrorNSPMissingProgramNCA;
-        }
+    if (nsp->GetNCA(title_id, FileSys::ContentRecordType::Program) == nullptr) {
+        if (!Core::Crypto::KeyManager::KeyFileExists(false))
+            return ResultStatus::ErrorMissingProductionKeyFile;
+        return ResultStatus::ErrorNSPMissingProgramNCA;
     }
 
     const auto result = secondary_loader->Load(process);
     if (result != ResultStatus::Success)
         return result;
 
+    FileSys::VirtualFile update_raw;
+    if (ReadUpdateRaw(update_raw) == ResultStatus::Success && update_raw != nullptr)
+        Service::FileSystem::SetPackedUpdate(std::move(update_raw));
+
     is_loaded = true;
 
     return ResultStatus::Success;
 }
 
-ResultStatus AppLoader_NSP::ReadRomFS(FileSys::VirtualFile& dir) {
-    return secondary_loader->ReadRomFS(dir);
+ResultStatus AppLoader_NSP::ReadRomFS(FileSys::VirtualFile& file) {
+    return secondary_loader->ReadRomFS(file);
+}
+
+u64 AppLoader_NSP::ReadRomFSIVFCOffset() const {
+    return secondary_loader->ReadRomFSIVFCOffset();
+}
+
+ResultStatus AppLoader_NSP::ReadUpdateRaw(FileSys::VirtualFile& file) {
+    if (nsp->IsExtractedType())
+        return ResultStatus::ErrorNoPackedUpdate;
+
+    const auto read =
+        nsp->GetNCAFile(FileSys::GetUpdateTitleID(title_id), FileSys::ContentRecordType::Program);
+
+    if (read == nullptr)
+        return ResultStatus::ErrorNoPackedUpdate;
+    const auto nca_test = std::make_shared<FileSys::NCA>(read);
+
+    if (nca_test->GetStatus() != ResultStatus::ErrorMissingBKTRBaseRomFS)
+        return nca_test->GetStatus();
+
+    file = read;
+    return ResultStatus::Success;
 }
 
 ResultStatus AppLoader_NSP::ReadProgramId(u64& out_program_id) {
diff --git a/src/core/loader/nsp.h b/src/core/loader/nsp.h
index eac9b819a..db91cd01e 100644
--- a/src/core/loader/nsp.h
+++ b/src/core/loader/nsp.h
@@ -37,7 +37,9 @@ public:
 
     ResultStatus Load(Kernel::Process& process) override;
 
-    ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
+    ResultStatus ReadRomFS(FileSys::VirtualFile& file) override;
+    u64 ReadRomFSIVFCOffset() const override;
+    ResultStatus ReadUpdateRaw(FileSys::VirtualFile& file) override;
     ResultStatus ReadProgramId(u64& out_program_id) override;
     ResultStatus ReadIcon(std::vector<u8>& buffer) override;
     ResultStatus ReadTitle(std::string& title) override;
@@ -47,7 +49,7 @@ private:
     std::unique_ptr<AppLoader> secondary_loader;
 
     FileSys::VirtualFile icon_file;
-    std::shared_ptr<FileSys::NACP> nacp_file;
+    std::unique_ptr<FileSys::NACP> nacp_file;
     u64 title_id;
 };
 
diff --git a/src/core/loader/xci.cpp b/src/core/loader/xci.cpp
index eda67a8c8..461607c95 100644
--- a/src/core/loader/xci.cpp
+++ b/src/core/loader/xci.cpp
@@ -9,7 +9,11 @@
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/control_metadata.h"
 #include "core/file_sys/patch_manager.h"
+#include "core/file_sys/registered_cache.h"
+#include "core/file_sys/romfs.h"
+#include "core/file_sys/submission_package.h"
 #include "core/hle/kernel/process.h"
+#include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/nca.h"
 #include "core/loader/xci.h"
 
@@ -26,7 +30,7 @@ AppLoader_XCI::AppLoader_XCI(FileSys::VirtualFile file)
         return;
 
     std::tie(nacp_file, icon_file) =
-        FileSys::PatchManager(xci->GetProgramTitleID()).ParseControlNCA(control_nca);
+        FileSys::PatchManager(xci->GetProgramTitleID()).ParseControlNCA(*control_nca);
 }
 
 AppLoader_XCI::~AppLoader_XCI() = default;
@@ -55,21 +59,48 @@ ResultStatus AppLoader_XCI::Load(Kernel::Process& process) {
     if (xci->GetProgramNCAStatus() != ResultStatus::Success)
         return xci->GetProgramNCAStatus();
 
-    const auto nca = xci->GetProgramNCA();
-    if (nca == nullptr && !Core::Crypto::KeyManager::KeyFileExists(false))
+    if (!xci->HasProgramNCA() && !Core::Crypto::KeyManager::KeyFileExists(false))
         return ResultStatus::ErrorMissingProductionKeyFile;
 
     const auto result = nca_loader->Load(process);
     if (result != ResultStatus::Success)
         return result;
 
+    FileSys::VirtualFile update_raw;
+    if (ReadUpdateRaw(update_raw) == ResultStatus::Success && update_raw != nullptr)
+        Service::FileSystem::SetPackedUpdate(std::move(update_raw));
+
     is_loaded = true;
 
     return ResultStatus::Success;
 }
 
-ResultStatus AppLoader_XCI::ReadRomFS(FileSys::VirtualFile& dir) {
-    return nca_loader->ReadRomFS(dir);
+ResultStatus AppLoader_XCI::ReadRomFS(FileSys::VirtualFile& file) {
+    return nca_loader->ReadRomFS(file);
+}
+
+u64 AppLoader_XCI::ReadRomFSIVFCOffset() const {
+    return nca_loader->ReadRomFSIVFCOffset();
+}
+
+ResultStatus AppLoader_XCI::ReadUpdateRaw(FileSys::VirtualFile& file) {
+    u64 program_id{};
+    nca_loader->ReadProgramId(program_id);
+    if (program_id == 0)
+        return ResultStatus::ErrorXCIMissingProgramNCA;
+
+    const auto read = xci->GetSecurePartitionNSP()->GetNCAFile(
+        FileSys::GetUpdateTitleID(program_id), FileSys::ContentRecordType::Program);
+
+    if (read == nullptr)
+        return ResultStatus::ErrorNoPackedUpdate;
+    const auto nca_test = std::make_shared<FileSys::NCA>(read);
+
+    if (nca_test->GetStatus() != ResultStatus::ErrorMissingBKTRBaseRomFS)
+        return nca_test->GetStatus();
+
+    file = read;
+    return ResultStatus::Success;
 }
 
 ResultStatus AppLoader_XCI::ReadProgramId(u64& out_program_id) {
diff --git a/src/core/loader/xci.h b/src/core/loader/xci.h
index 17e47b658..46f8dfc9e 100644
--- a/src/core/loader/xci.h
+++ b/src/core/loader/xci.h
@@ -37,7 +37,9 @@ public:
 
     ResultStatus Load(Kernel::Process& process) override;
 
-    ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
+    ResultStatus ReadRomFS(FileSys::VirtualFile& file) override;
+    u64 ReadRomFSIVFCOffset() const override;
+    ResultStatus ReadUpdateRaw(FileSys::VirtualFile& file) override;
     ResultStatus ReadProgramId(u64& out_program_id) override;
     ResultStatus ReadIcon(std::vector<u8>& buffer) override;
     ResultStatus ReadTitle(std::string& title) override;
@@ -47,7 +49,7 @@ private:
     std::unique_ptr<AppLoader_NCA> nca_loader;
 
     FileSys::VirtualFile icon_file;
-    std::shared_ptr<FileSys::NACP> nacp_file;
+    std::unique_ptr<FileSys::NACP> nacp_file;
 };
 
 } // namespace Loader
diff --git a/src/core/memory.cpp b/src/core/memory.cpp
index 014298ed6..70abd856a 100644
--- a/src/core/memory.cpp
+++ b/src/core/memory.cpp
@@ -4,9 +4,9 @@
 
 #include <algorithm>
 #include <cstring>
+#include <optional>
 #include <utility>
 
-#include <boost/optional.hpp>
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "common/logging/log.h"
diff --git a/src/core/memory_hook.h b/src/core/memory_hook.h
index 0269c7ff1..940777107 100644
--- a/src/core/memory_hook.h
+++ b/src/core/memory_hook.h
@@ -5,7 +5,8 @@
 #pragma once
 
 #include <memory>
-#include <boost/optional.hpp>
+#include <optional>
+
 #include "common/common_types.h"
 
 namespace Memory {
@@ -18,19 +19,19 @@ namespace Memory {
  *
  * A hook may be mapped to multiple regions of memory.
  *
- * If a boost::none or false is returned from a function, the read/write request is passed through
+ * If a std::nullopt or false is returned from a function, the read/write request is passed through
  * to the underlying memory region.
  */
 class MemoryHook {
 public:
     virtual ~MemoryHook();
 
-    virtual boost::optional<bool> IsValidAddress(VAddr addr) = 0;
+    virtual std::optional<bool> IsValidAddress(VAddr addr) = 0;
 
-    virtual boost::optional<u8> Read8(VAddr addr) = 0;
-    virtual boost::optional<u16> Read16(VAddr addr) = 0;
-    virtual boost::optional<u32> Read32(VAddr addr) = 0;
-    virtual boost::optional<u64> Read64(VAddr addr) = 0;
+    virtual std::optional<u8> Read8(VAddr addr) = 0;
+    virtual std::optional<u16> Read16(VAddr addr) = 0;
+    virtual std::optional<u32> Read32(VAddr addr) = 0;
+    virtual std::optional<u64> Read64(VAddr addr) = 0;
 
     virtual bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) = 0;
 
diff --git a/src/core/perf_stats.cpp b/src/core/perf_stats.cpp
index 7d95816fe..c716a462b 100644
--- a/src/core/perf_stats.cpp
+++ b/src/core/perf_stats.cpp
@@ -74,10 +74,6 @@ double PerfStats::GetLastFrameTimeScale() {
 }
 
 void FrameLimiter::DoFrameLimiting(microseconds current_system_time_us) {
-    // Max lag caused by slow frames. Can be adjusted to compensate for too many slow frames. Higher
-    // values increase the time needed to recover and limit framerate again after spikes.
-    constexpr microseconds MAX_LAG_TIME_US = 25000us;
-
     if (!Settings::values.use_frame_limit) {
         return;
     }
diff --git a/src/core/settings.h b/src/core/settings.h
index 1808f5937..b5aeff29b 100644
--- a/src/core/settings.h
+++ b/src/core/settings.h
@@ -113,7 +113,8 @@ static const std::array<const char*, NumAnalogs> mapping = {{
 struct Values {
     // System
     bool use_docked_mode;
-    std::string username;
+    bool enable_nfc;
+    int current_user;
     int language_index;
 
     // Controls
@@ -136,7 +137,7 @@ struct Values {
     float resolution_factor;
     bool use_frame_limit;
     u16 frame_limit;
-    bool use_accurate_framebuffers;
+    bool use_accurate_gpu_emulation;
 
     float bg_red;
     float bg_green;
@@ -155,6 +156,7 @@ struct Values {
     // Debugging
     bool use_gdbstub;
     u16 gdbstub_port;
+    std::string program_args;
 
     // WebService
     bool enable_telemetry;
diff --git a/src/core/telemetry_session.cpp b/src/core/telemetry_session.cpp
index f29fff1e7..a3b08c740 100644
--- a/src/core/telemetry_session.cpp
+++ b/src/core/telemetry_session.cpp
@@ -2,12 +2,16 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <array>
+
+#include <mbedtls/ctr_drbg.h>
+#include <mbedtls/entropy.h>
+
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "common/file_util.h"
+#include "common/logging/log.h"
 
-#include <mbedtls/ctr_drbg.h>
-#include <mbedtls/entropy.h>
 #include "core/core.h"
 #include "core/file_sys/control_metadata.h"
 #include "core/file_sys/patch_manager.h"
@@ -28,11 +32,11 @@ static u64 GenerateTelemetryId() {
     mbedtls_entropy_context entropy;
     mbedtls_entropy_init(&entropy);
     mbedtls_ctr_drbg_context ctr_drbg;
-    std::string personalization = "yuzu Telemetry ID";
+    constexpr std::array<char, 18> personalization{{"yuzu Telemetry ID"}};
 
     mbedtls_ctr_drbg_init(&ctr_drbg);
     ASSERT(mbedtls_ctr_drbg_seed(&ctr_drbg, mbedtls_entropy_func, &entropy,
-                                 reinterpret_cast<const unsigned char*>(personalization.c_str()),
+                                 reinterpret_cast<const unsigned char*>(personalization.data()),
                                  personalization.size()) == 0);
     ASSERT(mbedtls_ctr_drbg_random(&ctr_drbg, reinterpret_cast<unsigned char*>(&telemetry_id),
                                    sizeof(u64)) == 0);
@@ -159,8 +163,8 @@ TelemetrySession::TelemetrySession() {
     AddField(Telemetry::FieldType::UserConfig, "Renderer_UseFrameLimit",
              Settings::values.use_frame_limit);
     AddField(Telemetry::FieldType::UserConfig, "Renderer_FrameLimit", Settings::values.frame_limit);
-    AddField(Telemetry::FieldType::UserConfig, "Renderer_UseAccurateFramebuffers",
-             Settings::values.use_accurate_framebuffers);
+    AddField(Telemetry::FieldType::UserConfig, "Renderer_UseAccurateGpuEmulation",
+             Settings::values.use_accurate_gpu_emulation);
     AddField(Telemetry::FieldType::UserConfig, "System_UseDockedMode",
              Settings::values.use_docked_mode);
 }
@@ -180,4 +184,13 @@ TelemetrySession::~TelemetrySession() {
     backend = nullptr;
 }
 
+bool TelemetrySession::SubmitTestcase() {
+#ifdef ENABLE_WEB_SERVICE
+    field_collection.Accept(*backend);
+    return backend->SubmitTestcase();
+#else
+    return false;
+#endif
+}
+
 } // namespace Core
diff --git a/src/core/telemetry_session.h b/src/core/telemetry_session.h
index cec271df0..023612b79 100644
--- a/src/core/telemetry_session.h
+++ b/src/core/telemetry_session.h
@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <string>
 #include "common/telemetry.h"
 
 namespace Core {
@@ -30,7 +31,11 @@ public:
         field_collection.AddField(type, name, std::move(value));
     }
 
-    static void FinalizeAsyncJob();
+    /**
+     * Submits a Testcase.
+     * @returns A bool indicating whether the submission succeeded
+     */
+    bool SubmitTestcase();
 
 private:
     Telemetry::FieldCollection field_collection; ///< Tracks all added fields for the session
@@ -53,7 +58,6 @@ u64 RegenerateTelemetryId();
  * Verifies the username and token.
  * @param username yuzu username to use for authentication.
  * @param token yuzu token to use for authentication.
- * @param func A function that gets exectued when the verification is finished
  * @returns Future with bool indicating whether the verification succeeded
  */
 bool VerifyLogin(const std::string& username, const std::string& token);
diff --git a/src/tests/core/arm/arm_test_common.cpp b/src/tests/core/arm/arm_test_common.cpp
index c0a57e71f..9b8a44fa1 100644
--- a/src/tests/core/arm/arm_test_common.cpp
+++ b/src/tests/core/arm/arm_test_common.cpp
@@ -15,7 +15,8 @@ namespace ArmTests {
 TestEnvironment::TestEnvironment(bool mutable_memory_)
     : mutable_memory(mutable_memory_), test_memory(std::make_shared<TestMemory>(this)) {
 
-    Core::CurrentProcess() = Kernel::Process::Create(kernel, "");
+    auto process = Kernel::Process::Create(kernel, "");
+    kernel.MakeCurrentProcess(process.get());
     page_table = &Core::CurrentProcess()->VMManager().page_table;
 
     std::fill(page_table->pointers.begin(), page_table->pointers.end(), nullptr);
@@ -63,11 +64,11 @@ void TestEnvironment::ClearWriteRecords() {
 
 TestEnvironment::TestMemory::~TestMemory() {}
 
-boost::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) {
+std::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) {
     return true;
 }
 
-boost::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
+std::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
     const auto iter = data.find(addr);
 
     if (iter == data.end()) {
@@ -78,15 +79,15 @@ boost::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
     return iter->second;
 }
 
-boost::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) {
+std::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) {
     return *Read8(addr) | static_cast<u16>(*Read8(addr + 1)) << 8;
 }
 
-boost::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) {
+std::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) {
     return *Read16(addr) | static_cast<u32>(*Read16(addr + 2)) << 16;
 }
 
-boost::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) {
+std::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) {
     return *Read32(addr) | static_cast<u64>(*Read32(addr + 4)) << 32;
 }
 
diff --git a/src/tests/core/arm/arm_test_common.h b/src/tests/core/arm/arm_test_common.h
index 5de8dab4e..0b7539601 100644
--- a/src/tests/core/arm/arm_test_common.h
+++ b/src/tests/core/arm/arm_test_common.h
@@ -64,12 +64,12 @@ private:
 
         ~TestMemory() override;
 
-        boost::optional<bool> IsValidAddress(VAddr addr) override;
+        std::optional<bool> IsValidAddress(VAddr addr) override;
 
-        boost::optional<u8> Read8(VAddr addr) override;
-        boost::optional<u16> Read16(VAddr addr) override;
-        boost::optional<u32> Read32(VAddr addr) override;
-        boost::optional<u64> Read64(VAddr addr) override;
+        std::optional<u8> Read8(VAddr addr) override;
+        std::optional<u16> Read16(VAddr addr) override;
+        std::optional<u32> Read32(VAddr addr) override;
+        std::optional<u64> Read64(VAddr addr) override;
 
         bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) override;
 
diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt
index 09ecc5bad..ddb1a1d69 100644
--- a/src/video_core/CMakeLists.txt
+++ b/src/video_core/CMakeLists.txt
@@ -51,6 +51,10 @@ add_library(video_core STATIC
     renderer_opengl/maxwell_to_gl.h
     renderer_opengl/renderer_opengl.cpp
     renderer_opengl/renderer_opengl.h
+    renderer_opengl/utils.cpp
+    renderer_opengl/utils.h
+    surface.cpp
+    surface.h
     textures/astc.cpp
     textures/astc.h
     textures/decoders.cpp
diff --git a/src/video_core/command_processor.cpp b/src/video_core/command_processor.cpp
index f1aa6091b..28e8c13aa 100644
--- a/src/video_core/command_processor.cpp
+++ b/src/video_core/command_processor.cpp
@@ -81,7 +81,7 @@ void GPU::ProcessCommandLists(const std::vector<CommandListHeader>& commands) {
     for (auto entry : commands) {
         Tegra::GPUVAddr address = entry.Address();
         u32 size = entry.sz;
-        const boost::optional<VAddr> head_address = memory_manager->GpuToCpuAddress(address);
+        const std::optional<VAddr> head_address = memory_manager->GpuToCpuAddress(address);
         VAddr current_addr = *head_address;
         while (current_addr < *head_address + size * sizeof(CommandHeader)) {
             const CommandHeader header = {Memory::Read32(current_addr)};
diff --git a/src/video_core/engines/fermi_2d.cpp b/src/video_core/engines/fermi_2d.cpp
index ea1555c5d..74e44c7fe 100644
--- a/src/video_core/engines/fermi_2d.cpp
+++ b/src/video_core/engines/fermi_2d.cpp
@@ -4,11 +4,13 @@
 
 #include "core/memory.h"
 #include "video_core/engines/fermi_2d.h"
+#include "video_core/rasterizer_interface.h"
 #include "video_core/textures/decoders.h"
 
 namespace Tegra::Engines {
 
-Fermi2D::Fermi2D(MemoryManager& memory_manager) : memory_manager(memory_manager) {}
+Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
+    : memory_manager(memory_manager), rasterizer{rasterizer} {}
 
 void Fermi2D::WriteReg(u32 method, u32 value) {
     ASSERT_MSG(method < Regs::NUM_REGS,
@@ -44,27 +46,36 @@ void Fermi2D::HandleSurfaceCopy() {
     u32 src_bytes_per_pixel = RenderTargetBytesPerPixel(regs.src.format);
     u32 dst_bytes_per_pixel = RenderTargetBytesPerPixel(regs.dst.format);
 
-    if (regs.src.linear == regs.dst.linear) {
-        // If the input layout and the output layout are the same, just perform a raw copy.
-        ASSERT(regs.src.BlockHeight() == regs.dst.BlockHeight());
-        Memory::CopyBlock(dest_cpu, source_cpu,
-                          src_bytes_per_pixel * regs.dst.width * regs.dst.height);
-        return;
-    }
-
-    u8* src_buffer = Memory::GetPointer(source_cpu);
-    u8* dst_buffer = Memory::GetPointer(dest_cpu);
-
-    if (!regs.src.linear && regs.dst.linear) {
-        // If the input is tiled and the output is linear, deswizzle the input and copy it over.
-        Texture::CopySwizzledData(regs.src.width, regs.src.height, src_bytes_per_pixel,
-                                  dst_bytes_per_pixel, src_buffer, dst_buffer, true,
-                                  regs.src.BlockHeight());
-    } else {
-        // If the input is linear and the output is tiled, swizzle the input and copy it over.
-        Texture::CopySwizzledData(regs.src.width, regs.src.height, src_bytes_per_pixel,
-                                  dst_bytes_per_pixel, dst_buffer, src_buffer, false,
-                                  regs.dst.BlockHeight());
+    if (!rasterizer.AccelerateSurfaceCopy(regs.src, regs.dst)) {
+        rasterizer.FlushRegion(source_cpu, src_bytes_per_pixel * regs.src.width * regs.src.height);
+        // We have to invalidate the destination region to evict any outdated surfaces from the
+        // cache. We do this before actually writing the new data because the destination address
+        // might contain a dirty surface that will have to be written back to memory.
+        rasterizer.InvalidateRegion(dest_cpu,
+                                    dst_bytes_per_pixel * regs.dst.width * regs.dst.height);
+
+        if (regs.src.linear == regs.dst.linear) {
+            // If the input layout and the output layout are the same, just perform a raw copy.
+            ASSERT(regs.src.BlockHeight() == regs.dst.BlockHeight());
+            Memory::CopyBlock(dest_cpu, source_cpu,
+                              src_bytes_per_pixel * regs.dst.width * regs.dst.height);
+            return;
+        }
+        u8* src_buffer = Memory::GetPointer(source_cpu);
+        u8* dst_buffer = Memory::GetPointer(dest_cpu);
+        if (!regs.src.linear && regs.dst.linear) {
+            // If the input is tiled and the output is linear, deswizzle the input and copy it over.
+            Texture::CopySwizzledData(regs.src.width, regs.src.height, regs.src.depth,
+                                      src_bytes_per_pixel, dst_bytes_per_pixel, src_buffer,
+                                      dst_buffer, true, regs.src.BlockHeight(),
+                                      regs.src.BlockDepth());
+        } else {
+            // If the input is linear and the output is tiled, swizzle the input and copy it over.
+            Texture::CopySwizzledData(regs.src.width, regs.src.height, regs.src.depth,
+                                      src_bytes_per_pixel, dst_bytes_per_pixel, dst_buffer,
+                                      src_buffer, false, regs.dst.BlockHeight(),
+                                      regs.dst.BlockDepth());
+        }
     }
 }
 
diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h
index 021b83eaa..2a6e8bbbb 100644
--- a/src/video_core/engines/fermi_2d.h
+++ b/src/video_core/engines/fermi_2d.h
@@ -12,6 +12,10 @@
 #include "video_core/gpu.h"
 #include "video_core/memory_manager.h"
 
+namespace VideoCore {
+class RasterizerInterface;
+}
+
 namespace Tegra::Engines {
 
 #define FERMI2D_REG_INDEX(field_name)                                                              \
@@ -19,7 +23,7 @@ namespace Tegra::Engines {
 
 class Fermi2D final {
 public:
-    explicit Fermi2D(MemoryManager& memory_manager);
+    explicit Fermi2D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager);
     ~Fermi2D() = default;
 
     /// Write the value to the register identified by method.
@@ -32,9 +36,9 @@ public:
             RenderTargetFormat format;
             BitField<0, 1, u32> linear;
             union {
-                BitField<0, 4, u32> block_depth;
+                BitField<0, 4, u32> block_width;
                 BitField<4, 4, u32> block_height;
-                BitField<8, 4, u32> block_width;
+                BitField<8, 4, u32> block_depth;
             };
             u32 depth;
             u32 layer;
@@ -49,10 +53,20 @@ public:
                                              address_low);
             }
 
+            u32 BlockWidth() const {
+                // The block width is stored in log2 format.
+                return 1 << block_width;
+            }
+
             u32 BlockHeight() const {
                 // The block height is stored in log2 format.
                 return 1 << block_height;
             }
+
+            u32 BlockDepth() const {
+                // The block depth is stored in log2 format.
+                return 1 << block_depth;
+            }
         };
         static_assert(sizeof(Surface) == 0x28, "Surface has incorrect size");
 
@@ -94,6 +108,8 @@ public:
     MemoryManager& memory_manager;
 
 private:
+    VideoCore::RasterizerInterface& rasterizer;
+
     /// Performs the copy from the source surface to the destination surface as configured in the
     /// registers.
     void HandleSurfaceCopy();
diff --git a/src/video_core/engines/kepler_memory.cpp b/src/video_core/engines/kepler_memory.cpp
index 66ae6332d..585290d9f 100644
--- a/src/video_core/engines/kepler_memory.cpp
+++ b/src/video_core/engines/kepler_memory.cpp
@@ -5,10 +5,14 @@
 #include "common/logging/log.h"
 #include "core/memory.h"
 #include "video_core/engines/kepler_memory.h"
+#include "video_core/rasterizer_interface.h"
 
 namespace Tegra::Engines {
 
-KeplerMemory::KeplerMemory(MemoryManager& memory_manager) : memory_manager(memory_manager) {}
+KeplerMemory::KeplerMemory(VideoCore::RasterizerInterface& rasterizer,
+                           MemoryManager& memory_manager)
+    : memory_manager(memory_manager), rasterizer{rasterizer} {}
+
 KeplerMemory::~KeplerMemory() = default;
 
 void KeplerMemory::WriteReg(u32 method, u32 value) {
@@ -37,6 +41,11 @@ void KeplerMemory::ProcessData(u32 data) {
     VAddr dest_address =
         *memory_manager.GpuToCpuAddress(address + state.write_offset * sizeof(u32));
 
+    // We have to invalidate the destination region to evict any outdated surfaces from the cache.
+    // We do this before actually writing the new data because the destination address might contain
+    // a dirty surface that will have to be written back to memory.
+    rasterizer.InvalidateRegion(dest_address, sizeof(u32));
+
     Memory::Write32(dest_address, data);
 
     state.write_offset++;
diff --git a/src/video_core/engines/kepler_memory.h b/src/video_core/engines/kepler_memory.h
index b0d0078cf..bf4a13cff 100644
--- a/src/video_core/engines/kepler_memory.h
+++ b/src/video_core/engines/kepler_memory.h
@@ -11,6 +11,10 @@
 #include "common/common_types.h"
 #include "video_core/memory_manager.h"
 
+namespace VideoCore {
+class RasterizerInterface;
+}
+
 namespace Tegra::Engines {
 
 #define KEPLERMEMORY_REG_INDEX(field_name)                                                         \
@@ -18,7 +22,7 @@ namespace Tegra::Engines {
 
 class KeplerMemory final {
 public:
-    KeplerMemory(MemoryManager& memory_manager);
+    KeplerMemory(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager);
     ~KeplerMemory();
 
     /// Write the value to the register identified by method.
@@ -72,6 +76,7 @@ public:
 
 private:
     MemoryManager& memory_manager;
+    VideoCore::RasterizerInterface& rasterizer;
 
     void ProcessData(u32 data);
 };
diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp
index 8afd26fe9..d79c50919 100644
--- a/src/video_core/engines/maxwell_3d.cpp
+++ b/src/video_core/engines/maxwell_3d.cpp
@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <cinttypes>
+#include <cstring>
 #include "common/assert.h"
 #include "core/core.h"
 #include "core/core_timing.h"
@@ -13,28 +14,46 @@
 #include "video_core/renderer_base.h"
 #include "video_core/textures/texture.h"
 
-namespace Tegra {
-namespace Engines {
+namespace Tegra::Engines {
 
 /// First register id that is actually a Macro call.
 constexpr u32 MacroRegistersStart = 0xE00;
 
 Maxwell3D::Maxwell3D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
-    : memory_manager(memory_manager), rasterizer{rasterizer}, macro_interpreter(*this) {}
+    : memory_manager(memory_manager), rasterizer{rasterizer}, macro_interpreter(*this) {
+    InitializeRegisterDefaults();
+}
+
+void Maxwell3D::InitializeRegisterDefaults() {
+    // Initializes registers to their default values - what games expect them to be at boot. This is
+    // for certain registers that may not be explicitly set by games.
+
+    // Reset all registers to zero
+    std::memset(&regs, 0, sizeof(regs));
+
+    // Depth range near/far is not always set, but is expected to be the default 0.0f, 1.0f. This is
+    // needed for ARMS.
+    for (std::size_t viewport{}; viewport < Regs::NumViewports; ++viewport) {
+        regs.viewport[viewport].depth_range_near = 0.0f;
+        regs.viewport[viewport].depth_range_far = 1.0f;
+    }
+}
 
 void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
     // Reset the current macro.
     executing_macro = 0;
 
-    // The requested macro must have been uploaded already.
-    auto macro_code = uploaded_macros.find(method);
-    if (macro_code == uploaded_macros.end()) {
-        LOG_ERROR(HW_GPU, "Macro {:04X} was not uploaded", method);
+    // Lookup the macro offset
+    const u32 entry{(method - MacroRegistersStart) >> 1};
+    const auto& search{macro_offsets.find(entry)};
+    if (search == macro_offsets.end()) {
+        LOG_CRITICAL(HW_GPU, "macro not found for method 0x{:X}!", method);
+        UNREACHABLE();
         return;
     }
 
     // Execute the current macro.
-    macro_interpreter.Execute(macro_code->second, std::move(parameters));
+    macro_interpreter.Execute(search->second, std::move(parameters));
 }
 
 void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
@@ -80,6 +99,10 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
         ProcessMacroUpload(value);
         break;
     }
+    case MAXWELL3D_REG_INDEX(macros.bind): {
+        ProcessMacroBind(value);
+        break;
+    }
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]):
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]):
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]):
@@ -141,22 +164,25 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
 }
 
 void Maxwell3D::ProcessMacroUpload(u32 data) {
-    // Store the uploaded macro code to interpret them when they're called.
-    auto& macro = uploaded_macros[regs.macros.entry * 2 + MacroRegistersStart];
-    macro.push_back(data);
+    ASSERT_MSG(regs.macros.upload_address < macro_memory.size(),
+               "upload_address exceeded macro_memory size!");
+    macro_memory[regs.macros.upload_address++] = data;
+}
+
+void Maxwell3D::ProcessMacroBind(u32 data) {
+    macro_offsets[regs.macros.entry] = data;
 }
 
 void Maxwell3D::ProcessQueryGet() {
     GPUVAddr sequence_address = regs.query.QueryAddress();
     // Since the sequence address is given as a GPU VAddr, we have to convert it to an application
     // VAddr before writing.
-    boost::optional<VAddr> address = memory_manager.GpuToCpuAddress(sequence_address);
+    std::optional<VAddr> address = memory_manager.GpuToCpuAddress(sequence_address);
 
     // TODO(Subv): Support the other query units.
     ASSERT_MSG(regs.query.query_get.unit == Regs::QueryUnit::Crop,
                "Units other than CROP are unimplemented");
 
-    u32 value = Memory::Read32(*address);
     u64 result = 0;
 
     // TODO(Subv): Support the other query variables
@@ -269,7 +295,7 @@ void Maxwell3D::ProcessCBData(u32 value) {
     // Don't allow writing past the end of the buffer.
     ASSERT(regs.const_buffer.cb_pos + sizeof(u32) <= regs.const_buffer.cb_size);
 
-    boost::optional<VAddr> address =
+    std::optional<VAddr> address =
         memory_manager.GpuToCpuAddress(buffer_address + regs.const_buffer.cb_pos);
 
     Memory::Write32(*address, value);
@@ -282,7 +308,7 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const {
     GPUVAddr tic_base_address = regs.tic.TICAddress();
 
     GPUVAddr tic_address_gpu = tic_base_address + tic_index * sizeof(Texture::TICEntry);
-    boost::optional<VAddr> tic_address_cpu = memory_manager.GpuToCpuAddress(tic_address_gpu);
+    std::optional<VAddr> tic_address_cpu = memory_manager.GpuToCpuAddress(tic_address_gpu);
 
     Texture::TICEntry tic_entry;
     Memory::ReadBlock(*tic_address_cpu, &tic_entry, sizeof(Texture::TICEntry));
@@ -306,7 +332,7 @@ Texture::TSCEntry Maxwell3D::GetTSCEntry(u32 tsc_index) const {
     GPUVAddr tsc_base_address = regs.tsc.TSCAddress();
 
     GPUVAddr tsc_address_gpu = tsc_base_address + tsc_index * sizeof(Texture::TSCEntry);
-    boost::optional<VAddr> tsc_address_cpu = memory_manager.GpuToCpuAddress(tsc_address_gpu);
+    std::optional<VAddr> tsc_address_cpu = memory_manager.GpuToCpuAddress(tsc_address_gpu);
 
     Texture::TSCEntry tsc_entry;
     Memory::ReadBlock(*tsc_address_cpu, &tsc_entry, sizeof(Texture::TSCEntry));
@@ -370,7 +396,7 @@ Texture::FullTextureInfo Maxwell3D::GetStageTexture(Regs::ShaderStage stage,
 
     ASSERT(tex_info_address < tex_info_buffer.address + tex_info_buffer.size);
 
-    boost::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address);
+    std::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address);
     Texture::TextureHandle tex_handle{Memory::Read32(*tex_address_cpu)};
 
     Texture::FullTextureInfo tex_info{};
@@ -408,5 +434,4 @@ void Maxwell3D::ProcessClearBuffers() {
     rasterizer.Clear();
 }
 
-} // namespace Engines
-} // namespace Tegra
+} // namespace Tegra::Engines
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index 4290da33f..50873813e 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -347,6 +347,16 @@ public:
             DecrWrap = 8,
         };
 
+        enum class MemoryLayout : u32 {
+            Linear = 0,
+            BlockLinear = 1,
+        };
+
+        enum class InvMemoryLayout : u32 {
+            BlockLinear = 0,
+            Linear = 1,
+        };
+
         struct Cull {
             enum class FrontFace : u32 {
                 ClockWise = 0x0900,
@@ -432,8 +442,16 @@ public:
             u32 width;
             u32 height;
             Tegra::RenderTargetFormat format;
-            u32 block_dimensions;
-            u32 array_mode;
+            union {
+                BitField<0, 3, u32> block_width;
+                BitField<4, 3, u32> block_height;
+                BitField<8, 3, u32> block_depth;
+                BitField<12, 1, InvMemoryLayout> type;
+            } memory_layout;
+            union {
+                BitField<0, 16, u32> array_mode;
+                BitField<16, 1, u32> volume;
+            };
             u32 layer_stride;
             u32 base_layer;
             INSERT_PADDING_WORDS(7);
@@ -457,12 +475,13 @@ public:
                 INSERT_PADDING_WORDS(0x45);
 
                 struct {
-                    INSERT_PADDING_WORDS(1);
+                    u32 upload_address;
                     u32 data;
                     u32 entry;
+                    u32 bind;
                 } macros;
 
-                INSERT_PADDING_WORDS(0x189);
+                INSERT_PADDING_WORDS(0x188);
 
                 u32 tfb_enabled;
 
@@ -532,7 +551,21 @@ public:
                 INSERT_PADDING_WORDS(0x3);
                 s32 clear_stencil;
 
-                INSERT_PADDING_WORDS(0x6C);
+                INSERT_PADDING_WORDS(0x17);
+
+                struct {
+                    u32 enable;
+                    union {
+                        BitField<0, 16, u32> min_x;
+                        BitField<16, 16, u32> max_x;
+                    };
+                    union {
+                        BitField<0, 16, u32> min_y;
+                        BitField<16, 16, u32> max_y;
+                    };
+                } scissor_test;
+
+                INSERT_PADDING_WORDS(0x52);
 
                 s32 stencil_back_func_ref;
                 u32 stencil_back_mask;
@@ -548,7 +581,12 @@ public:
                     u32 address_high;
                     u32 address_low;
                     Tegra::DepthFormat format;
-                    u32 block_dimensions;
+                    union {
+                        BitField<0, 4, u32> block_width;
+                        BitField<4, 4, u32> block_height;
+                        BitField<8, 4, u32> block_depth;
+                        BitField<20, 1, InvMemoryLayout> type;
+                    } memory_layout;
                     u32 layer_stride;
 
                     GPUVAddr Address() const {
@@ -606,8 +644,10 @@ public:
                 u32 d3d_cull_mode;
 
                 ComparisonOp depth_test_func;
+                float alpha_test_ref;
+                ComparisonOp alpha_test_func;
 
-                INSERT_PADDING_WORDS(0xB);
+                INSERT_PADDING_WORDS(0x9);
 
                 struct {
                     u32 separate_alpha;
@@ -684,7 +724,11 @@ public:
                 StencilOp stencil_back_op_zpass;
                 ComparisonOp stencil_back_func_func;
 
-                INSERT_PADDING_WORDS(0x17);
+                INSERT_PADDING_WORDS(0x4);
+
+                u32 framebuffer_srgb;
+
+                INSERT_PADDING_WORDS(0x12);
 
                 union {
                     BitField<2, 1, u32> coord_origin;
@@ -712,7 +756,14 @@ public:
                     };
                 } draw;
 
-                INSERT_PADDING_WORDS(0x6B);
+                INSERT_PADDING_WORDS(0xA);
+
+                struct {
+                    u32 enabled;
+                    u32 index;
+                } primitive_restart;
+
+                INSERT_PADDING_WORDS(0x5F);
 
                 struct {
                     u32 start_addr_high;
@@ -944,10 +995,25 @@ public:
     /// Returns the texture information for a specific texture in a specific shader stage.
     Texture::FullTextureInfo GetStageTexture(Regs::ShaderStage stage, std::size_t offset) const;
 
+    /// Memory for macro code - it's undetermined how big this is, however 1MB is much larger than
+    /// we've seen used.
+    using MacroMemory = std::array<u32, 0x40000>;
+
+    /// Gets a reference to macro memory.
+    const MacroMemory& GetMacroMemory() const {
+        return macro_memory;
+    }
+
 private:
+    void InitializeRegisterDefaults();
+
     VideoCore::RasterizerInterface& rasterizer;
 
-    std::unordered_map<u32, std::vector<u32>> uploaded_macros;
+    /// Start offsets of each macro in macro_memory
+    std::unordered_map<u32, u32> macro_offsets;
+
+    /// Memory for macro code
+    MacroMemory macro_memory;
 
     /// Macro method that is currently being executed / being fed parameters.
     u32 executing_macro = 0;
@@ -970,9 +1036,12 @@ private:
      */
     void CallMacroMethod(u32 method, std::vector<u32> parameters);
 
-    /// Handles writes to the macro uploading registers.
+    /// Handles writes to the macro uploading register.
     void ProcessMacroUpload(u32 data);
 
+    /// Handles writes to the macro bind register.
+    void ProcessMacroBind(u32 data);
+
     /// Handles a write to the CLEAR_BUFFERS register.
     void ProcessClearBuffers();
 
@@ -1002,6 +1071,7 @@ ASSERT_REG_POSITION(vertex_buffer, 0x35D);
 ASSERT_REG_POSITION(clear_color[0], 0x360);
 ASSERT_REG_POSITION(clear_depth, 0x364);
 ASSERT_REG_POSITION(clear_stencil, 0x368);
+ASSERT_REG_POSITION(scissor_test, 0x380);
 ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5);
 ASSERT_REG_POSITION(stencil_back_mask, 0x3D6);
 ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7);
@@ -1037,9 +1107,11 @@ ASSERT_REG_POSITION(stencil_back_op_fail, 0x566);
 ASSERT_REG_POSITION(stencil_back_op_zfail, 0x567);
 ASSERT_REG_POSITION(stencil_back_op_zpass, 0x568);
 ASSERT_REG_POSITION(stencil_back_func_func, 0x569);
+ASSERT_REG_POSITION(framebuffer_srgb, 0x56E);
 ASSERT_REG_POSITION(point_coord_replace, 0x581);
 ASSERT_REG_POSITION(code_address, 0x582);
 ASSERT_REG_POSITION(draw, 0x585);
+ASSERT_REG_POSITION(primitive_restart, 0x591);
 ASSERT_REG_POSITION(index_array, 0x5F2);
 ASSERT_REG_POSITION(instanced_arrays, 0x620);
 ASSERT_REG_POSITION(cull, 0x646);
diff --git a/src/video_core/engines/maxwell_compute.cpp b/src/video_core/engines/maxwell_compute.cpp
index 59e28b22d..8b5f08351 100644
--- a/src/video_core/engines/maxwell_compute.cpp
+++ b/src/video_core/engines/maxwell_compute.cpp
@@ -6,8 +6,7 @@
 #include "core/core.h"
 #include "video_core/engines/maxwell_compute.h"
 
-namespace Tegra {
-namespace Engines {
+namespace Tegra::Engines {
 
 void MaxwellCompute::WriteReg(u32 method, u32 value) {
     ASSERT_MSG(method < Regs::NUM_REGS,
@@ -26,5 +25,4 @@ void MaxwellCompute::WriteReg(u32 method, u32 value) {
     }
 }
 
-} // namespace Engines
-} // namespace Tegra
+} // namespace Tegra::Engines
diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp
index aa7481b8c..b8a78cf82 100644
--- a/src/video_core/engines/maxwell_dma.cpp
+++ b/src/video_core/engines/maxwell_dma.cpp
@@ -4,12 +4,13 @@
 
 #include "core/memory.h"
 #include "video_core/engines/maxwell_dma.h"
+#include "video_core/rasterizer_interface.h"
 #include "video_core/textures/decoders.h"
 
-namespace Tegra {
-namespace Engines {
+namespace Tegra::Engines {
 
-MaxwellDMA::MaxwellDMA(MemoryManager& memory_manager) : memory_manager(memory_manager) {}
+MaxwellDMA::MaxwellDMA(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
+    : memory_manager(memory_manager), rasterizer{rasterizer} {}
 
 void MaxwellDMA::WriteReg(u32 method, u32 value) {
     ASSERT_MSG(method < Regs::NUM_REGS,
@@ -44,38 +45,77 @@ void MaxwellDMA::HandleCopy() {
     ASSERT(regs.exec.query_mode == Regs::QueryMode::None);
     ASSERT(regs.exec.query_intr == Regs::QueryIntr::None);
     ASSERT(regs.exec.copy_mode == Regs::CopyMode::Unk2);
-    ASSERT(regs.src_params.pos_x == 0);
-    ASSERT(regs.src_params.pos_y == 0);
     ASSERT(regs.dst_params.pos_x == 0);
     ASSERT(regs.dst_params.pos_y == 0);
 
-    if (regs.exec.is_dst_linear == regs.exec.is_src_linear) {
-        std::size_t copy_size = regs.x_count;
+    if (!regs.exec.is_dst_linear && !regs.exec.is_src_linear) {
+        // If both the source and the destination are in block layout, assert.
+        UNREACHABLE_MSG("Tiled->Tiled DMA transfers are not yet implemented");
+        return;
+    }
 
+    if (regs.exec.is_dst_linear && regs.exec.is_src_linear) {
         // When the enable_2d bit is disabled, the copy is performed as if we were copying a 1D
-        // buffer of length `x_count`, otherwise we copy a 2D buffer of size (x_count, y_count).
-        if (regs.exec.enable_2d) {
-            copy_size = copy_size * regs.y_count;
+        // buffer of length `x_count`, otherwise we copy a 2D image of dimensions (x_count,
+        // y_count).
+        if (!regs.exec.enable_2d) {
+            Memory::CopyBlock(dest_cpu, source_cpu, regs.x_count);
+            return;
         }
 
-        Memory::CopyBlock(dest_cpu, source_cpu, copy_size);
+        // If both the source and the destination are in linear layout, perform a line-by-line
+        // copy. We're going to take a subrect of size (x_count, y_count) from the source
+        // rectangle. There is no need to manually flush/invalidate the regions because
+        // CopyBlock does that for us.
+        for (u32 line = 0; line < regs.y_count; ++line) {
+            const VAddr source_line = source_cpu + line * regs.src_pitch;
+            const VAddr dest_line = dest_cpu + line * regs.dst_pitch;
+            Memory::CopyBlock(dest_line, source_line, regs.x_count);
+        }
         return;
     }
 
     ASSERT(regs.exec.enable_2d == 1);
-    u8* src_buffer = Memory::GetPointer(source_cpu);
-    u8* dst_buffer = Memory::GetPointer(dest_cpu);
+
+    const std::size_t copy_size = regs.x_count * regs.y_count;
+
+    const auto FlushAndInvalidate = [&](u32 src_size, u64 dst_size) {
+        // TODO(Subv): For now, manually flush the regions until we implement GPU-accelerated
+        // copying.
+        rasterizer.FlushRegion(source_cpu, src_size);
+
+        // We have to invalidate the destination region to evict any outdated surfaces from the
+        // cache. We do this before actually writing the new data because the destination address
+        // might contain a dirty surface that will have to be written back to memory.
+        rasterizer.InvalidateRegion(dest_cpu, dst_size);
+    };
 
     if (regs.exec.is_dst_linear && !regs.exec.is_src_linear) {
+        ASSERT(regs.src_params.size_z == 1);
         // If the input is tiled and the output is linear, deswizzle the input and copy it over.
-        Texture::CopySwizzledData(regs.src_params.size_x, regs.src_params.size_y, 1, 1, src_buffer,
-                                  dst_buffer, true, regs.src_params.BlockHeight());
+
+        const u32 src_bytes_per_pixel = regs.src_pitch / regs.src_params.size_x;
+
+        FlushAndInvalidate(regs.src_pitch * regs.src_params.size_y,
+                           copy_size * src_bytes_per_pixel);
+
+        Texture::UnswizzleSubrect(regs.x_count, regs.y_count, regs.dst_pitch,
+                                  regs.src_params.size_x, src_bytes_per_pixel, source_cpu, dest_cpu,
+                                  regs.src_params.BlockHeight(), regs.src_params.pos_x,
+                                  regs.src_params.pos_y);
     } else {
+        ASSERT(regs.dst_params.size_z == 1);
+        ASSERT(regs.src_pitch == regs.x_count);
+
+        const u32 src_bpp = regs.src_pitch / regs.x_count;
+
+        FlushAndInvalidate(regs.src_pitch * regs.y_count,
+                           regs.dst_params.size_x * regs.dst_params.size_y * src_bpp);
+
         // If the input is linear and the output is tiled, swizzle the input and copy it over.
-        Texture::CopySwizzledData(regs.dst_params.size_x, regs.dst_params.size_y, 1, 1, dst_buffer,
-                                  src_buffer, false, regs.dst_params.BlockHeight());
+        Texture::SwizzleSubrect(regs.x_count, regs.y_count, regs.src_pitch, regs.dst_params.size_x,
+                                src_bpp, dest_cpu, source_cpu, regs.dst_params.BlockHeight());
     }
 }
 
-} // namespace Engines
-} // namespace Tegra
+} // namespace Tegra::Engines
diff --git a/src/video_core/engines/maxwell_dma.h b/src/video_core/engines/maxwell_dma.h
index 311ccb616..5f3704f05 100644
--- a/src/video_core/engines/maxwell_dma.h
+++ b/src/video_core/engines/maxwell_dma.h
@@ -12,11 +12,15 @@
 #include "video_core/gpu.h"
 #include "video_core/memory_manager.h"
 
+namespace VideoCore {
+class RasterizerInterface;
+}
+
 namespace Tegra::Engines {
 
 class MaxwellDMA final {
 public:
-    explicit MaxwellDMA(MemoryManager& memory_manager);
+    explicit MaxwellDMA(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager);
     ~MaxwellDMA() = default;
 
     /// Write the value to the register identified by method.
@@ -43,6 +47,10 @@ public:
             u32 BlockHeight() const {
                 return 1 << block_height;
             }
+
+            u32 BlockDepth() const {
+                return 1 << block_depth;
+            }
         };
 
         static_assert(sizeof(Parameters) == 24, "Parameters has wrong size");
@@ -129,6 +137,8 @@ public:
     MemoryManager& memory_manager;
 
 private:
+    VideoCore::RasterizerInterface& rasterizer;
+
     /// Performs the copy from the source buffer to the destination buffer as configured in the
     /// registers.
     void HandleCopy();
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index b1f137b9c..83a6fd875 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -5,12 +5,11 @@
 #pragma once
 
 #include <bitset>
+#include <optional>
 #include <string>
 #include <tuple>
 #include <vector>
 
-#include <boost/optional.hpp>
-
 #include "common/assert.h"
 #include "common/bit_field.h"
 #include "common/common_types.h"
@@ -79,6 +78,7 @@ union Attribute {
     constexpr explicit Attribute(u64 value) : value(value) {}
 
     enum class Index : u64 {
+        PointSize = 6,
         Position = 7,
         Attribute_0 = 8,
         Attribute_31 = 39,
@@ -207,6 +207,16 @@ enum class UniformType : u64 {
     Double = 5,
 };
 
+enum class StoreType : u64 {
+    Unsigned8 = 0,
+    Signed8 = 1,
+    Unsigned16 = 2,
+    Signed16 = 3,
+    Bytes32 = 4,
+    Bytes64 = 5,
+    Bytes128 = 6,
+};
+
 enum class IMinMaxExchange : u64 {
     None = 0,
     XLo = 1,
@@ -214,6 +224,18 @@ enum class IMinMaxExchange : u64 {
     XHi = 3,
 };
 
+enum class VideoType : u64 {
+    Size16_Low = 0,
+    Size16_High = 1,
+    Size32 = 2,
+    Invalid = 3,
+};
+
+enum class VmadShr : u64 {
+    Shr7 = 1,
+    Shr15 = 2,
+};
+
 enum class XmadMode : u64 {
     None = 0,
     CLo = 1,
@@ -255,7 +277,7 @@ enum class ControlCode : u64 {
     GTU = 12,
     NEU = 13,
     GEU = 14,
-    //
+    T = 15,
     OFF = 16,
     LO = 17,
     SFF = 18,
@@ -314,6 +336,35 @@ enum class TextureMiscMode : u64 {
     PTP,
 };
 
+enum class IsberdMode : u64 {
+    None = 0,
+    Patch = 1,
+    Prim = 2,
+    Attr = 3,
+};
+
+enum class IsberdShift : u64 { None = 0, U16 = 1, B32 = 2 };
+
+enum class HalfType : u64 {
+    H0_H1 = 0,
+    F32 = 1,
+    H0_H0 = 2,
+    H1_H1 = 3,
+};
+
+enum class HalfMerge : u64 {
+    H0_H1 = 0,
+    F32 = 1,
+    Mrg_H0 = 2,
+    Mrg_H1 = 3,
+};
+
+enum class HalfPrecision : u64 {
+    None = 0,
+    FTZ = 1,
+    FMZ = 2,
+};
+
 enum class IpaInterpMode : u64 {
     Linear = 0,
     Perspective = 1,
@@ -340,6 +391,87 @@ struct IpaMode {
     }
 };
 
+enum class SystemVariable : u64 {
+    LaneId = 0x00,
+    VirtCfg = 0x02,
+    VirtId = 0x03,
+    Pm0 = 0x04,
+    Pm1 = 0x05,
+    Pm2 = 0x06,
+    Pm3 = 0x07,
+    Pm4 = 0x08,
+    Pm5 = 0x09,
+    Pm6 = 0x0a,
+    Pm7 = 0x0b,
+    OrderingTicket = 0x0f,
+    PrimType = 0x10,
+    InvocationId = 0x11,
+    Ydirection = 0x12,
+    ThreadKill = 0x13,
+    ShaderType = 0x14,
+    DirectBeWriteAddressLow = 0x15,
+    DirectBeWriteAddressHigh = 0x16,
+    DirectBeWriteEnabled = 0x17,
+    MachineId0 = 0x18,
+    MachineId1 = 0x19,
+    MachineId2 = 0x1a,
+    MachineId3 = 0x1b,
+    Affinity = 0x1c,
+    InvocationInfo = 0x1d,
+    WscaleFactorXY = 0x1e,
+    WscaleFactorZ = 0x1f,
+    Tid = 0x20,
+    TidX = 0x21,
+    TidY = 0x22,
+    TidZ = 0x23,
+    CtaParam = 0x24,
+    CtaIdX = 0x25,
+    CtaIdY = 0x26,
+    CtaIdZ = 0x27,
+    NtId = 0x28,
+    CirQueueIncrMinusOne = 0x29,
+    Nlatc = 0x2a,
+    SmSpaVersion = 0x2c,
+    MultiPassShaderInfo = 0x2d,
+    LwinHi = 0x2e,
+    SwinHi = 0x2f,
+    SwinLo = 0x30,
+    SwinSz = 0x31,
+    SmemSz = 0x32,
+    SmemBanks = 0x33,
+    LwinLo = 0x34,
+    LwinSz = 0x35,
+    LmemLosz = 0x36,
+    LmemHioff = 0x37,
+    EqMask = 0x38,
+    LtMask = 0x39,
+    LeMask = 0x3a,
+    GtMask = 0x3b,
+    GeMask = 0x3c,
+    RegAlloc = 0x3d,
+    CtxAddr = 0x3e,      // .fmask = F_SM50
+    BarrierAlloc = 0x3e, // .fmask = F_SM60
+    GlobalErrorStatus = 0x40,
+    WarpErrorStatus = 0x42,
+    WarpErrorStatusClear = 0x43,
+    PmHi0 = 0x48,
+    PmHi1 = 0x49,
+    PmHi2 = 0x4a,
+    PmHi3 = 0x4b,
+    PmHi4 = 0x4c,
+    PmHi5 = 0x4d,
+    PmHi6 = 0x4e,
+    PmHi7 = 0x4f,
+    ClockLo = 0x50,
+    ClockHi = 0x51,
+    GlobalTimerLo = 0x52,
+    GlobalTimerHi = 0x53,
+    HwTaskId = 0x60,
+    CircularQueueEntryIndex = 0x61,
+    CircularQueueEntryAddressLow = 0x62,
+    CircularQueueEntryAddressHigh = 0x63,
+};
+
 union Instruction {
     Instruction& operator=(const Instruction& instr) {
         value = instr.value;
@@ -362,6 +494,7 @@ union Instruction {
     BitField<48, 16, u64> opcode;
 
     union {
+        BitField<20, 16, u64> imm20_16;
         BitField<20, 19, u64> imm20_19;
         BitField<20, 32, s64> imm20_32;
         BitField<45, 1, u64> negate_b;
@@ -403,6 +536,10 @@ union Instruction {
             }
         } lop3;
 
+        u16 GetImm20_16() const {
+            return static_cast<u16>(imm20_16);
+        }
+
         u32 GetImm20_19() const {
             u32 imm{static_cast<u32>(imm20_19)};
             imm <<= 12;
@@ -437,6 +574,14 @@ union Instruction {
     } fmul;
 
     union {
+        BitField<55, 1, u64> saturate;
+    } fmul32;
+
+    union {
+        BitField<52, 1, u64> generates_cc;
+    } op_32;
+
+    union {
         BitField<48, 1, u64> is_signed;
     } shift;
 
@@ -447,6 +592,70 @@ union Instruction {
     } alu_integer;
 
     union {
+        BitField<39, 1, u64> ftz;
+        BitField<32, 1, u64> saturate;
+        BitField<49, 2, HalfMerge> merge;
+
+        BitField<43, 1, u64> negate_a;
+        BitField<44, 1, u64> abs_a;
+        BitField<47, 2, HalfType> type_a;
+
+        BitField<31, 1, u64> negate_b;
+        BitField<30, 1, u64> abs_b;
+        BitField<47, 2, HalfType> type_b;
+
+        BitField<35, 2, HalfType> type_c;
+    } alu_half;
+
+    union {
+        BitField<39, 2, HalfPrecision> precision;
+        BitField<39, 1, u64> ftz;
+        BitField<52, 1, u64> saturate;
+        BitField<49, 2, HalfMerge> merge;
+
+        BitField<43, 1, u64> negate_a;
+        BitField<44, 1, u64> abs_a;
+        BitField<47, 2, HalfType> type_a;
+    } alu_half_imm;
+
+    union {
+        BitField<29, 1, u64> first_negate;
+        BitField<20, 9, u64> first;
+
+        BitField<56, 1, u64> second_negate;
+        BitField<30, 9, u64> second;
+
+        u32 PackImmediates() const {
+            // Immediates are half floats shifted.
+            constexpr u32 imm_shift = 6;
+            return static_cast<u32>((first << imm_shift) | (second << (16 + imm_shift)));
+        }
+    } half_imm;
+
+    union {
+        union {
+            BitField<37, 2, HalfPrecision> precision;
+            BitField<32, 1, u64> saturate;
+
+            BitField<30, 1, u64> negate_c;
+            BitField<35, 2, HalfType> type_c;
+        } rr;
+
+        BitField<57, 2, HalfPrecision> precision;
+        BitField<52, 1, u64> saturate;
+
+        BitField<49, 2, HalfMerge> merge;
+
+        BitField<47, 2, HalfType> type_a;
+
+        BitField<56, 1, u64> negate_b;
+        BitField<28, 2, HalfType> type_b;
+
+        BitField<51, 1, u64> negate_c;
+        BitField<53, 2, HalfType> type_reg39;
+    } hfma2;
+
+    union {
         BitField<40, 1, u64> invert;
     } popc;
 
@@ -552,6 +761,18 @@ union Instruction {
     } ld_c;
 
     union {
+        BitField<48, 3, StoreType> type;
+    } ldst_sl;
+
+    union {
+        BitField<44, 2, u64> unknown;
+    } ld_l;
+
+    union {
+        BitField<44, 2, u64> unknown;
+    } st_l;
+
+    union {
         BitField<0, 3, u64> pred0;
         BitField<3, 3, u64> pred3;
         BitField<7, 1, u64> abs_a;
@@ -562,7 +783,6 @@ union Instruction {
         BitField<45, 2, PredOperation> op;
         BitField<47, 1, u64> ftz;
         BitField<48, 4, PredCondition> cond;
-        BitField<56, 1, u64> neg_b;
     } fsetp;
 
     union {
@@ -589,6 +809,14 @@ union Instruction {
     } psetp;
 
     union {
+        BitField<43, 4, PredCondition> cond;
+        BitField<45, 2, PredOperation> op;
+        BitField<3, 3, u64> pred3;
+        BitField<0, 3, u64> pred0;
+        BitField<39, 3, u64> pred39;
+    } vsetp;
+
+    union {
         BitField<12, 3, u64> pred12;
         BitField<15, 1, u64> neg_pred12;
         BitField<24, 2, PredOperation> cond;
@@ -610,6 +838,23 @@ union Instruction {
     } csetp;
 
     union {
+        BitField<35, 4, PredCondition> cond;
+        BitField<49, 1, u64> h_and;
+        BitField<6, 1, u64> ftz;
+        BitField<45, 2, PredOperation> op;
+        BitField<3, 3, u64> pred3;
+        BitField<0, 3, u64> pred0;
+        BitField<43, 1, u64> negate_a;
+        BitField<44, 1, u64> abs_a;
+        BitField<47, 2, HalfType> type_a;
+        BitField<31, 1, u64> negate_b;
+        BitField<30, 1, u64> abs_b;
+        BitField<28, 2, HalfType> type_b;
+        BitField<42, 1, u64> neg_pred;
+        BitField<39, 3, u64> pred39;
+    } hsetp2;
+
+    union {
         BitField<39, 3, u64> pred39;
         BitField<42, 1, u64> neg_pred;
         BitField<43, 1, u64> neg_a;
@@ -620,10 +865,24 @@ union Instruction {
         BitField<53, 1, u64> neg_b;
         BitField<54, 1, u64> abs_a;
         BitField<55, 1, u64> ftz;
-        BitField<56, 1, u64> neg_imm;
     } fset;
 
     union {
+        BitField<49, 1, u64> bf;
+        BitField<35, 3, PredCondition> cond;
+        BitField<50, 1, u64> ftz;
+        BitField<45, 2, PredOperation> op;
+        BitField<43, 1, u64> negate_a;
+        BitField<44, 1, u64> abs_a;
+        BitField<47, 2, HalfType> type_a;
+        BitField<31, 1, u64> negate_b;
+        BitField<30, 1, u64> abs_b;
+        BitField<28, 2, HalfType> type_b;
+        BitField<42, 1, u64> neg_pred;
+        BitField<39, 3, u64> pred39;
+    } hset2;
+
+    union {
         BitField<39, 3, u64> pred39;
         BitField<42, 1, u64> neg_pred;
         BitField<44, 1, u64> bf;
@@ -915,6 +1174,37 @@ union Instruction {
     } bra;
 
     union {
+        BitField<39, 1, u64> emit; // EmitVertex
+        BitField<40, 1, u64> cut;  // EndPrimitive
+    } out;
+
+    union {
+        BitField<31, 1, u64> skew;
+        BitField<32, 1, u64> o;
+        BitField<33, 2, IsberdMode> mode;
+        BitField<47, 2, IsberdShift> shift;
+    } isberd;
+
+    union {
+        BitField<48, 1, u64> signed_a;
+        BitField<38, 1, u64> is_byte_chunk_a;
+        BitField<36, 2, VideoType> type_a;
+        BitField<36, 2, u64> byte_height_a;
+
+        BitField<49, 1, u64> signed_b;
+        BitField<50, 1, u64> use_register_b;
+        BitField<30, 1, u64> is_byte_chunk_b;
+        BitField<28, 2, VideoType> type_b;
+        BitField<28, 2, u64> byte_height_b;
+    } video;
+
+    union {
+        BitField<51, 2, VmadShr> shr;
+        BitField<55, 1, u64> saturate; // Saturates the result (a * b + c)
+        BitField<47, 1, u64> cc;
+    } vmad;
+
+    union {
         BitField<20, 16, u64> imm20_16;
         BitField<36, 1, u64> product_shift_left;
         BitField<37, 1, u64> merge_37;
@@ -936,10 +1226,16 @@ union Instruction {
         BitField<36, 5, u64> index;
     } cbuf36;
 
+    // Unsure about the size of this one.
+    // It's always used with a gpr0, so any size should be fine.
+    BitField<20, 8, SystemVariable> sys20;
+
     BitField<47, 1, u64> generates_cc;
     BitField<61, 1, u64> is_b_imm;
     BitField<60, 1, u64> is_b_gpr;
     BitField<59, 1, u64> is_c_gpr;
+    BitField<20, 24, s64> smem_imm;
+    BitField<0, 5, ControlCode> flow_control_code;
 
     Attribute attribute;
     Sampler sampler;
@@ -955,14 +1251,20 @@ public:
         KIL,
         SSY,
         SYNC,
+        BRK,
         DEPBAR,
         BFE_C,
         BFE_R,
         BFE_IMM,
         BRA,
+        PBK,
         LD_A,
+        LD_L,
+        LD_S,
         LD_C,
         ST_A,
+        ST_L,
+        ST_S,
         LDG, // Load from global memory
         STG, // Store in global memory
         TEX,
@@ -975,6 +1277,10 @@ public:
         TMML,   // Texture Mip Map Level
         EXIT,
         IPA,
+        OUT_R, // Emit vertex/primitive
+        ISBERD,
+        VMAD,
+        VSETP,
         FFMA_IMM, // Fused Multiply and Add
         FFMA_CR,
         FFMA_RC,
@@ -1002,6 +1308,18 @@ public:
         LEA_RZ,
         LEA_IMM,
         LEA_HI,
+        HADD2_C,
+        HADD2_R,
+        HADD2_IMM,
+        HMUL2_C,
+        HMUL2_R,
+        HMUL2_IMM,
+        HFMA2_CR,
+        HFMA2_RC,
+        HFMA2_RR,
+        HFMA2_IMM_R,
+        HSETP2_R,
+        HSET2_R,
         POPC_C,
         POPC_R,
         POPC_IMM,
@@ -1034,6 +1352,7 @@ public:
         MOV_C,
         MOV_R,
         MOV_IMM,
+        MOV_SYS,
         MOV32_IMM,
         SHL_C,
         SHL_R,
@@ -1074,9 +1393,12 @@ public:
         ArithmeticImmediate,
         ArithmeticInteger,
         ArithmeticIntegerImmediate,
+        ArithmeticHalf,
+        ArithmeticHalfImmediate,
         Bfe,
         Shift,
         Ffma,
+        Hfma2,
         Flow,
         Synch,
         Memory,
@@ -1084,6 +1406,8 @@ public:
         FloatSetPredicate,
         IntegerSet,
         IntegerSetPredicate,
+        HalfSet,
+        HalfSetPredicate,
         PredicateSetPredicate,
         PredicateSetRegister,
         Conversion,
@@ -1095,7 +1419,7 @@ public:
     /// conditionally executed).
     static bool IsPredicatedInstruction(Id opcode) {
         // TODO(Subv): Add the rest of unpredicated instructions.
-        return opcode != Id::SSY;
+        return opcode != Id::SSY && opcode != Id::PBK;
     }
 
     class Matcher {
@@ -1136,7 +1460,7 @@ public:
         Type type;
     };
 
-    static boost::optional<const Matcher&> Decode(Instruction instr) {
+    static std::optional<std::reference_wrapper<const Matcher>> Decode(Instruction instr) {
         static const auto table{GetDecodeTable()};
 
         const auto matches_instruction = [instr](const auto& matcher) {
@@ -1144,7 +1468,8 @@ public:
         };
 
         auto iter = std::find_if(table.begin(), table.end(), matches_instruction);
-        return iter != table.end() ? boost::optional<const Matcher&>(*iter) : boost::none;
+        return iter != table.end() ? std::optional<std::reference_wrapper<const Matcher>>(*iter)
+                                   : std::nullopt;
     }
 
 private:
@@ -1191,12 +1516,18 @@ private:
 #define INST(bitstring, op, type, name) Detail::GetMatcher(bitstring, op, type, name)
             INST("111000110011----", Id::KIL, Type::Flow, "KIL"),
             INST("111000101001----", Id::SSY, Type::Flow, "SSY"),
+            INST("111000101010----", Id::PBK, Type::Flow, "PBK"),
             INST("111000100100----", Id::BRA, Type::Flow, "BRA"),
+            INST("1111000011111---", Id::SYNC, Type::Flow, "SYNC"),
+            INST("111000110100---", Id::BRK, Type::Flow, "BRK"),
             INST("1111000011110---", Id::DEPBAR, Type::Synch, "DEPBAR"),
-            INST("1111000011111---", Id::SYNC, Type::Synch, "SYNC"),
             INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
+            INST("1110111101001---", Id::LD_S, Type::Memory, "LD_S"),
+            INST("1110111101000---", Id::LD_L, Type::Memory, "LD_L"),
             INST("1110111110010---", Id::LD_C, Type::Memory, "LD_C"),
             INST("1110111111110---", Id::ST_A, Type::Memory, "ST_A"),
+            INST("1110111101011---", Id::ST_S, Type::Memory, "ST_S"),
+            INST("1110111101010---", Id::ST_L, Type::Memory, "ST_L"),
             INST("1110111011010---", Id::LDG, Type::Memory, "LDG"),
             INST("1110111011011---", Id::STG, Type::Memory, "STG"),
             INST("110000----111---", Id::TEX, Type::Memory, "TEX"),
@@ -1209,6 +1540,10 @@ private:
             INST("1101111101011---", Id::TMML, Type::Memory, "TMML"),
             INST("111000110000----", Id::EXIT, Type::Trivial, "EXIT"),
             INST("11100000--------", Id::IPA, Type::Trivial, "IPA"),
+            INST("1111101111100---", Id::OUT_R, Type::Trivial, "OUT_R"),
+            INST("1110111111010---", Id::ISBERD, Type::Trivial, "ISBERD"),
+            INST("01011111--------", Id::VMAD, Type::Trivial, "VMAD"),
+            INST("0101000011110---", Id::VSETP, Type::Trivial, "VSETP"),
             INST("0011001-1-------", Id::FFMA_IMM, Type::Ffma, "FFMA_IMM"),
             INST("010010011-------", Id::FFMA_CR, Type::Ffma, "FFMA_CR"),
             INST("010100011-------", Id::FFMA_RC, Type::Ffma, "FFMA_RC"),
@@ -1242,6 +1577,18 @@ private:
             INST("001101101101----", Id::LEA_IMM, Type::ArithmeticInteger, "LEA_IMM"),
             INST("010010111101----", Id::LEA_RZ, Type::ArithmeticInteger, "LEA_RZ"),
             INST("00011000--------", Id::LEA_HI, Type::ArithmeticInteger, "LEA_HI"),
+            INST("0111101-1-------", Id::HADD2_C, Type::ArithmeticHalf, "HADD2_C"),
+            INST("0101110100010---", Id::HADD2_R, Type::ArithmeticHalf, "HADD2_R"),
+            INST("0111101-0-------", Id::HADD2_IMM, Type::ArithmeticHalfImmediate, "HADD2_IMM"),
+            INST("0111100-1-------", Id::HMUL2_C, Type::ArithmeticHalf, "HMUL2_C"),
+            INST("0101110100001---", Id::HMUL2_R, Type::ArithmeticHalf, "HMUL2_R"),
+            INST("0111100-0-------", Id::HMUL2_IMM, Type::ArithmeticHalfImmediate, "HMUL2_IMM"),
+            INST("01110---1-------", Id::HFMA2_CR, Type::Hfma2, "HFMA2_CR"),
+            INST("01100---1-------", Id::HFMA2_RC, Type::Hfma2, "HFMA2_RC"),
+            INST("0101110100000---", Id::HFMA2_RR, Type::Hfma2, "HFMA2_RR"),
+            INST("01110---0-------", Id::HFMA2_IMM_R, Type::Hfma2, "HFMA2_R_IMM"),
+            INST("0101110100100---", Id::HSETP2_R, Type::HalfSetPredicate, "HSETP_R"),
+            INST("0101110100011---", Id::HSET2_R, Type::HalfSet, "HSET2_R"),
             INST("0101000010000---", Id::MUFU, Type::Arithmetic, "MUFU"),
             INST("0100110010010---", Id::RRO_C, Type::Arithmetic, "RRO_C"),
             INST("0101110010010---", Id::RRO_R, Type::Arithmetic, "RRO_R"),
@@ -1255,6 +1602,7 @@ private:
             INST("0100110010011---", Id::MOV_C, Type::Arithmetic, "MOV_C"),
             INST("0101110010011---", Id::MOV_R, Type::Arithmetic, "MOV_R"),
             INST("0011100-10011---", Id::MOV_IMM, Type::Arithmetic, "MOV_IMM"),
+            INST("1111000011001---", Id::MOV_SYS, Type::Trivial, "MOV_SYS"),
             INST("000000010000----", Id::MOV32_IMM, Type::ArithmeticImmediate, "MOV32_IMM"),
             INST("0100110001100---", Id::FMNMX_C, Type::Arithmetic, "FMNMX_C"),
             INST("0101110001100---", Id::FMNMX_R, Type::Arithmetic, "FMNMX_R"),
diff --git a/src/video_core/engines/shader_header.h b/src/video_core/engines/shader_header.h
index a885ee3cf..a0e015c4b 100644
--- a/src/video_core/engines/shader_header.h
+++ b/src/video_core/engines/shader_header.h
@@ -96,6 +96,11 @@ struct Header {
             }
         } ps;
     };
+
+    u64 GetLocalMemorySize() {
+        return (common1.shader_local_memory_low_size |
+                (common2.shader_local_memory_high_size << 24));
+    }
 };
 
 static_assert(sizeof(Header) == 0x50, "Incorrect structure size");
diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp
index baa8b63b7..83c7e5b0b 100644
--- a/src/video_core/gpu.cpp
+++ b/src/video_core/gpu.cpp
@@ -25,10 +25,10 @@ u32 FramebufferConfig::BytesPerPixel(PixelFormat format) {
 GPU::GPU(VideoCore::RasterizerInterface& rasterizer) {
     memory_manager = std::make_unique<Tegra::MemoryManager>();
     maxwell_3d = std::make_unique<Engines::Maxwell3D>(rasterizer, *memory_manager);
-    fermi_2d = std::make_unique<Engines::Fermi2D>(*memory_manager);
+    fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer, *memory_manager);
     maxwell_compute = std::make_unique<Engines::MaxwellCompute>();
-    maxwell_dma = std::make_unique<Engines::MaxwellDMA>(*memory_manager);
-    kepler_memory = std::make_unique<Engines::KeplerMemory>(*memory_manager);
+    maxwell_dma = std::make_unique<Engines::MaxwellDMA>(rasterizer, *memory_manager);
+    kepler_memory = std::make_unique<Engines::KeplerMemory>(rasterizer, *memory_manager);
 }
 
 GPU::~GPU() = default;
diff --git a/src/video_core/macro_interpreter.cpp b/src/video_core/macro_interpreter.cpp
index 377bd66ab..335a8d407 100644
--- a/src/video_core/macro_interpreter.cpp
+++ b/src/video_core/macro_interpreter.cpp
@@ -11,7 +11,7 @@ namespace Tegra {
 
 MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {}
 
-void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> parameters) {
+void MacroInterpreter::Execute(u32 offset, std::vector<u32> parameters) {
     Reset();
     registers[1] = parameters[0];
     this->parameters = std::move(parameters);
@@ -19,7 +19,7 @@ void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> pa
     // Execute the code until we hit an exit condition.
     bool keep_executing = true;
     while (keep_executing) {
-        keep_executing = Step(code, false);
+        keep_executing = Step(offset, false);
     }
 
     // Assert the the macro used all the input parameters
@@ -29,7 +29,7 @@ void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> pa
 void MacroInterpreter::Reset() {
     registers = {};
     pc = 0;
-    delayed_pc = boost::none;
+    delayed_pc = {};
     method_address.raw = 0;
     parameters.clear();
     // The next parameter index starts at 1, because $r1 already has the value of the first
@@ -37,17 +37,17 @@ void MacroInterpreter::Reset() {
     next_parameter_index = 1;
 }
 
-bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
+bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
     u32 base_address = pc;
 
-    Opcode opcode = GetOpcode(code);
+    Opcode opcode = GetOpcode(offset);
     pc += 4;
 
     // Update the program counter if we were delayed
-    if (delayed_pc != boost::none) {
+    if (delayed_pc) {
         ASSERT(is_delay_slot);
         pc = *delayed_pc;
-        delayed_pc = boost::none;
+        delayed_pc = {};
     }
 
     switch (opcode.operation) {
@@ -108,7 +108,7 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
 
             delayed_pc = base_address + opcode.GetBranchTarget();
             // Execute one more instruction due to the delay slot.
-            return Step(code, true);
+            return Step(offset, true);
         }
         break;
     }
@@ -121,17 +121,18 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
         // Exit has a delay slot, execute the next instruction
         // Note: Executing an exit during a branch delay slot will cause the instruction at the
         // branch target to be executed before exiting.
-        Step(code, true);
+        Step(offset, true);
         return false;
     }
 
     return true;
 }
 
-MacroInterpreter::Opcode MacroInterpreter::GetOpcode(const std::vector<u32>& code) const {
+MacroInterpreter::Opcode MacroInterpreter::GetOpcode(u32 offset) const {
+    const auto& macro_memory{maxwell3d.GetMacroMemory()};
     ASSERT((pc % sizeof(u32)) == 0);
-    ASSERT(pc < code.size() * sizeof(u32));
-    return {code[pc / sizeof(u32)]};
+    ASSERT((pc + offset) < macro_memory.size() * sizeof(u32));
+    return {macro_memory[offset + pc / sizeof(u32)]};
 }
 
 u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const {
diff --git a/src/video_core/macro_interpreter.h b/src/video_core/macro_interpreter.h
index cee0baaf3..62d1ce289 100644
--- a/src/video_core/macro_interpreter.h
+++ b/src/video_core/macro_interpreter.h
@@ -5,8 +5,9 @@
 #pragma once
 
 #include <array>
+#include <optional>
 #include <vector>
-#include <boost/optional.hpp>
+
 #include "common/bit_field.h"
 #include "common/common_types.h"
 
@@ -21,10 +22,10 @@ public:
 
     /**
      * Executes the macro code with the specified input parameters.
-     * @param code The macro byte code to execute
-     * @param parameters The parameters of the macro
+     * @param offset Offset to start execution at.
+     * @param parameters The parameters of the macro.
      */
-    void Execute(const std::vector<u32>& code, std::vector<u32> parameters);
+    void Execute(u32 offset, std::vector<u32> parameters);
 
 private:
     enum class Operation : u32 {
@@ -109,11 +110,11 @@ private:
     /**
      * Executes a single macro instruction located at the current program counter. Returns whether
      * the interpreter should keep running.
-     * @param code The macro code to execute.
+     * @param offset Offset to start execution at.
      * @param is_delay_slot Whether the current step is being executed due to a delay slot in a
      * previous instruction.
      */
-    bool Step(const std::vector<u32>& code, bool is_delay_slot);
+    bool Step(u32 offset, bool is_delay_slot);
 
     /// Calculates the result of an ALU operation. src_a OP src_b;
     u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const;
@@ -126,7 +127,7 @@ private:
     bool EvaluateBranchCondition(BranchCondition cond, u32 value) const;
 
     /// Reads an opcode at the current program counter location.
-    Opcode GetOpcode(const std::vector<u32>& code) const;
+    Opcode GetOpcode(u32 offset) const;
 
     /// Returns the specified register's value. Register 0 is hardcoded to always return 0.
     u32 GetRegister(u32 register_id) const;
@@ -149,7 +150,7 @@ private:
     Engines::Maxwell3D& maxwell3d;
 
     u32 pc; ///< Current program counter
-    boost::optional<u32>
+    std::optional<u32>
         delayed_pc; ///< Program counter to execute at after the delay slot is executed.
 
     static constexpr std::size_t NumMacroRegisters = 8;
diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp
index ca923d17d..90a8e825d 100644
--- a/src/video_core/memory_manager.cpp
+++ b/src/video_core/memory_manager.cpp
@@ -9,7 +9,7 @@
 namespace Tegra {
 
 GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
-    boost::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, align);
+    std::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, align);
     ASSERT(gpu_addr);
 
     for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
@@ -34,7 +34,7 @@ GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) {
 }
 
 GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) {
-    boost::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, PAGE_SIZE);
+    std::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, PAGE_SIZE);
     ASSERT(gpu_addr);
 
     for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
@@ -87,7 +87,17 @@ GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) {
     return gpu_addr;
 }
 
-boost::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
+GPUVAddr MemoryManager::GetRegionEnd(GPUVAddr region_start) const {
+    for (const auto& region : mapped_regions) {
+        const GPUVAddr region_end{region.gpu_addr + region.size};
+        if (region_start >= region.gpu_addr && region_start < region_end) {
+            return region_end;
+        }
+    }
+    return {};
+}
+
+std::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
     GPUVAddr gpu_addr = 0;
     u64 free_space = 0;
     align = (align + PAGE_MASK) & ~PAGE_MASK;
@@ -108,7 +118,7 @@ boost::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
     return {};
 }
 
-boost::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) {
+std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) {
     VAddr base_addr = PageSlot(gpu_addr);
 
     if (base_addr == static_cast<u64>(PageStatus::Allocated) ||
diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h
index 86765e72a..b1255fd56 100644
--- a/src/video_core/memory_manager.h
+++ b/src/video_core/memory_manager.h
@@ -6,10 +6,9 @@
 
 #include <array>
 #include <memory>
+#include <optional>
 #include <vector>
 
-#include <boost/optional.hpp>
-
 #include "common/common_types.h"
 
 namespace Tegra {
@@ -26,7 +25,8 @@ public:
     GPUVAddr MapBufferEx(VAddr cpu_addr, u64 size);
     GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size);
     GPUVAddr UnmapBuffer(GPUVAddr gpu_addr, u64 size);
-    boost::optional<VAddr> GpuToCpuAddress(GPUVAddr gpu_addr);
+    GPUVAddr GetRegionEnd(GPUVAddr region_start) const;
+    std::optional<VAddr> GpuToCpuAddress(GPUVAddr gpu_addr);
     std::vector<GPUVAddr> CpuToGpuAddress(VAddr cpu_addr) const;
 
     static constexpr u64 PAGE_BITS = 16;
@@ -34,7 +34,7 @@ public:
     static constexpr u64 PAGE_MASK = PAGE_SIZE - 1;
 
 private:
-    boost::optional<GPUVAddr> FindFreeBlock(u64 size, u64 align = 1);
+    std::optional<GPUVAddr> FindFreeBlock(u64 size, u64 align = 1);
     bool IsPageMapped(GPUVAddr gpu_addr);
     VAddr& PageSlot(GPUVAddr gpu_addr);
 
diff --git a/src/video_core/rasterizer_cache.h b/src/video_core/rasterizer_cache.h
index 083b283b0..0a3b3951e 100644
--- a/src/video_core/rasterizer_cache.h
+++ b/src/video_core/rasterizer_cache.h
@@ -11,32 +11,77 @@
 
 #include "common/common_types.h"
 #include "core/core.h"
+#include "core/settings.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_base.h"
 
+class RasterizerCacheObject {
+public:
+    /// Gets the address of the shader in guest memory, required for cache management
+    virtual VAddr GetAddr() const = 0;
+
+    /// Gets the size of the shader in guest memory, required for cache management
+    virtual std::size_t GetSizeInBytes() const = 0;
+
+    /// Wriets any cached resources back to memory
+    virtual void Flush() = 0;
+
+    /// Sets whether the cached object should be considered registered
+    void SetIsRegistered(bool registered) {
+        is_registered = registered;
+    }
+
+    /// Returns true if the cached object is registered
+    bool IsRegistered() const {
+        return is_registered;
+    }
+
+    /// Returns true if the cached object is dirty
+    bool IsDirty() const {
+        return is_dirty;
+    }
+
+    /// Returns ticks from when this cached object was last modified
+    u64 GetLastModifiedTicks() const {
+        return last_modified_ticks;
+    }
+
+    /// Marks an object as recently modified, used to specify whether it is clean or dirty
+    template <class T>
+    void MarkAsModified(bool dirty, T& cache) {
+        is_dirty = dirty;
+        last_modified_ticks = cache.GetModifiedTicks();
+    }
+
+private:
+    bool is_registered{};      ///< Whether the object is currently registered with the cache
+    bool is_dirty{};           ///< Whether the object is dirty (out of sync with guest memory)
+    u64 last_modified_ticks{}; ///< When the object was last modified, used for in-order flushing
+};
+
 template <class T>
 class RasterizerCache : NonCopyable {
+    friend class RasterizerCacheObject;
+
 public:
+    /// Write any cached resources overlapping the specified region back to memory
+    void FlushRegion(Tegra::GPUVAddr addr, size_t size) {
+        const auto& objects{GetSortedObjectsFromRegion(addr, size)};
+        for (auto& object : objects) {
+            FlushObject(object);
+        }
+    }
+
     /// Mark the specified region as being invalidated
     void InvalidateRegion(VAddr addr, u64 size) {
-        if (size == 0)
-            return;
-
-        const ObjectInterval interval{addr, addr + size};
-        for (auto& pair : boost::make_iterator_range(object_cache.equal_range(interval))) {
-            for (auto& cached_object : pair.second) {
-                if (!cached_object)
-                    continue;
-
-                remove_objects.emplace(cached_object);
+        const auto& objects{GetSortedObjectsFromRegion(addr, size)};
+        for (auto& object : objects) {
+            if (!object->IsRegistered()) {
+                // Skip duplicates
+                continue;
             }
+            Unregister(object);
         }
-
-        for (auto& remove_object : remove_objects) {
-            Unregister(remove_object);
-        }
-
-        remove_objects.clear();
     }
 
     /// Invalidates everything in the cache
@@ -62,6 +107,7 @@ protected:
 
     /// Register an object into the cache
     void Register(const T& object) {
+        object->SetIsRegistered(true);
         object_cache.add({GetInterval(object), ObjectSet{object}});
         auto& rasterizer = Core::System::GetInstance().Renderer().Rasterizer();
         rasterizer.UpdatePagesCachedCount(object->GetAddr(), object->GetSizeInBytes(), 1);
@@ -69,12 +115,57 @@ protected:
 
     /// Unregisters an object from the cache
     void Unregister(const T& object) {
+        object->SetIsRegistered(false);
         auto& rasterizer = Core::System::GetInstance().Renderer().Rasterizer();
         rasterizer.UpdatePagesCachedCount(object->GetAddr(), object->GetSizeInBytes(), -1);
+
+        // Only flush if use_accurate_gpu_emulation is enabled, as it incurs a performance hit
+        if (Settings::values.use_accurate_gpu_emulation) {
+            FlushObject(object);
+        }
+
         object_cache.subtract({GetInterval(object), ObjectSet{object}});
     }
 
+    /// Returns a ticks counter used for tracking when cached objects were last modified
+    u64 GetModifiedTicks() {
+        return ++modified_ticks;
+    }
+
 private:
+    /// Returns a list of cached objects from the specified memory region, ordered by access time
+    std::vector<T> GetSortedObjectsFromRegion(VAddr addr, u64 size) {
+        if (size == 0) {
+            return {};
+        }
+
+        std::vector<T> objects;
+        const ObjectInterval interval{addr, addr + size};
+        for (auto& pair : boost::make_iterator_range(object_cache.equal_range(interval))) {
+            for (auto& cached_object : pair.second) {
+                if (!cached_object) {
+                    continue;
+                }
+                objects.push_back(cached_object);
+            }
+        }
+
+        std::sort(objects.begin(), objects.end(), [](const T& a, const T& b) -> bool {
+            return a->GetLastModifiedTicks() < b->GetLastModifiedTicks();
+        });
+
+        return objects;
+    }
+
+    /// Flushes the specified object, updating appropriate cache state as needed
+    void FlushObject(const T& object) {
+        if (!object->IsDirty()) {
+            return;
+        }
+        object->Flush();
+        object->MarkAsModified(false, *this);
+    }
+
     using ObjectSet = std::set<T>;
     using ObjectCache = boost::icl::interval_map<VAddr, ObjectSet>;
     using ObjectInterval = typename ObjectCache::interval_type;
@@ -84,6 +175,6 @@ private:
                                           object->GetAddr() + object->GetSizeInBytes());
     }
 
-    ObjectCache object_cache;
-    ObjectSet remove_objects;
+    ObjectCache object_cache; ///< Cache of objects
+    u64 modified_ticks{};     ///< Counter of cache state ticks, used for in-order flushing
 };
diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h
index cd819d69f..06fc59dbe 100644
--- a/src/video_core/rasterizer_interface.h
+++ b/src/video_core/rasterizer_interface.h
@@ -5,6 +5,7 @@
 #pragma once
 
 #include "common/common_types.h"
+#include "video_core/engines/fermi_2d.h"
 #include "video_core/gpu.h"
 #include "video_core/memory_manager.h"
 
@@ -33,13 +34,9 @@ public:
     /// and invalidated
     virtual void FlushAndInvalidateRegion(VAddr addr, u64 size) = 0;
 
-    /// Attempt to use a faster method to perform a display transfer with is_texture_copy = 0
-    virtual bool AccelerateDisplayTransfer(const void* config) {
-        return false;
-    }
-
-    /// Attempt to use a faster method to perform a display transfer with is_texture_copy = 1
-    virtual bool AccelerateTextureCopy(const void* config) {
+    /// Attempt to use a faster method to perform a surface copy
+    virtual bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
+                                       const Tegra::Engines::Fermi2D::Regs::Surface& dst) {
         return false;
     }
 
diff --git a/src/video_core/renderer_base.h b/src/video_core/renderer_base.h
index 2cd0738ff..669e26e15 100644
--- a/src/video_core/renderer_base.h
+++ b/src/video_core/renderer_base.h
@@ -6,7 +6,8 @@
 
 #include <atomic>
 #include <memory>
-#include <boost/optional.hpp>
+#include <optional>
+
 #include "common/common_types.h"
 #include "video_core/gpu.h"
 #include "video_core/rasterizer_interface.h"
@@ -28,7 +29,8 @@ public:
     virtual ~RendererBase();
 
     /// Swap buffers (render frame)
-    virtual void SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) = 0;
+    virtual void SwapBuffers(
+        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) = 0;
 
     /// Initialize the renderer
     virtual bool Init() = 0;
diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp
index c142095c5..41a54b3e7 100644
--- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp
@@ -17,7 +17,7 @@ OGLBufferCache::OGLBufferCache(std::size_t size) : stream_buffer(GL_ARRAY_BUFFER
 GLintptr OGLBufferCache::UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size,
                                       std::size_t alignment, bool cache) {
     auto& memory_manager = Core::System::GetInstance().GPU().MemoryManager();
-    const boost::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)};
+    const std::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)};
 
     // Cache management is a big overhead, so only cache entries with a given size.
     // TODO: Figure out which size is the best for given games.
diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h
index 965976334..be29dc8be 100644
--- a/src/video_core/renderer_opengl/gl_buffer_cache.h
+++ b/src/video_core/renderer_opengl/gl_buffer_cache.h
@@ -15,15 +15,18 @@
 
 namespace OpenGL {
 
-struct CachedBufferEntry final {
-    VAddr GetAddr() const {
+struct CachedBufferEntry final : public RasterizerCacheObject {
+    VAddr GetAddr() const override {
         return addr;
     }
 
-    std::size_t GetSizeInBytes() const {
+    std::size_t GetSizeInBytes() const override {
         return size;
     }
 
+    // We do not have to flush this cache as things in it are never modified by us.
+    void Flush() override {}
+
     VAddr addr;
     std::size_t size;
     GLintptr offset;
diff --git a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp
index ee1d9601b..741f14bc3 100644
--- a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp
+++ b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp
@@ -45,7 +45,7 @@ GLintptr PrimitiveAssembler::MakeQuadIndexed(Tegra::GPUVAddr gpu_addr, std::size
     auto [dst_pointer, index_offset] = buffer_cache.ReserveMemory(map_size);
 
     auto& memory_manager = Core::System::GetInstance().GPU().MemoryManager();
-    const boost::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)};
+    const std::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)};
     const u8* source{Memory::GetPointer(*cpu_addr)};
 
     for (u32 primitive = 0; primitive < count / 4; ++primitive) {
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index a91bc6dee..a0527fe57 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -30,8 +30,8 @@
 namespace OpenGL {
 
 using Maxwell = Tegra::Engines::Maxwell3D::Regs;
-using PixelFormat = SurfaceParams::PixelFormat;
-using SurfaceType = SurfaceParams::SurfaceType;
+using PixelFormat = VideoCore::Surface::PixelFormat;
+using SurfaceType = VideoCore::Surface::SurfaceType;
 
 MICROPROFILE_DEFINE(OpenGL_VAO, "OpenGL", "Vertex Array Setup", MP_RGB(128, 128, 192));
 MICROPROFILE_DEFINE(OpenGL_Shader, "OpenGL", "Shader Setup", MP_RGB(128, 128, 192));
@@ -104,7 +104,7 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
     }
 
     ASSERT_MSG(has_ARB_separate_shader_objects, "has_ARB_separate_shader_objects is unsupported");
-
+    OpenGLState::ApplyDefaultState();
     // Clipping plane 0 is always enabled for PICA fixed clip plane z <= 0
     state.clip_distance[0] = true;
 
@@ -115,8 +115,6 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
     state.draw.shader_program = 0;
     state.Apply();
 
-    glEnable(GL_BLEND);
-
     glGetIntegerv(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT, &uniform_buffer_alignment);
 
     LOG_CRITICAL(Render_OpenGL, "Sync fixed function OpenGL state here!");
@@ -255,7 +253,7 @@ DrawParameters RasterizerOpenGL::SetupDraw() {
     return params;
 }
 
-void RasterizerOpenGL::SetupShaders() {
+void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) {
     MICROPROFILE_SCOPE(OpenGL_Shader);
     const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
 
@@ -270,6 +268,11 @@ void RasterizerOpenGL::SetupShaders() {
 
         // Skip stages that are not enabled
         if (!gpu.regs.IsShaderConfigEnabled(index)) {
+            switch (program) {
+            case Maxwell::ShaderProgram::Geometry:
+                shader_program_manager->UseTrivialGeometryShader();
+                break;
+            }
             continue;
         }
 
@@ -281,18 +284,26 @@ void RasterizerOpenGL::SetupShaders() {
             &ubo, sizeof(ubo), static_cast<std::size_t>(uniform_buffer_alignment));
 
         // Bind the buffer
-        glBindBufferRange(GL_UNIFORM_BUFFER, stage, buffer_cache.GetHandle(), offset, sizeof(ubo));
+        glBindBufferRange(GL_UNIFORM_BUFFER, static_cast<GLuint>(stage), buffer_cache.GetHandle(),
+                          offset, static_cast<GLsizeiptr>(sizeof(ubo)));
 
         Shader shader{shader_cache.GetStageProgram(program)};
 
         switch (program) {
         case Maxwell::ShaderProgram::VertexA:
         case Maxwell::ShaderProgram::VertexB: {
-            shader_program_manager->UseProgrammableVertexShader(shader->GetProgramHandle());
+            shader_program_manager->UseProgrammableVertexShader(
+                shader->GetProgramHandle(primitive_mode));
+            break;
+        }
+        case Maxwell::ShaderProgram::Geometry: {
+            shader_program_manager->UseProgrammableGeometryShader(
+                shader->GetProgramHandle(primitive_mode));
             break;
         }
         case Maxwell::ShaderProgram::Fragment: {
-            shader_program_manager->UseProgrammableFragmentShader(shader->GetProgramHandle());
+            shader_program_manager->UseProgrammableFragmentShader(
+                shader->GetProgramHandle(primitive_mode));
             break;
         }
         default:
@@ -302,12 +313,13 @@ void RasterizerOpenGL::SetupShaders() {
         }
 
         // Configure the const buffers for this shader stage.
-        current_constbuffer_bindpoint = SetupConstBuffers(static_cast<Maxwell::ShaderStage>(stage),
-                                                          shader, current_constbuffer_bindpoint);
+        current_constbuffer_bindpoint =
+            SetupConstBuffers(static_cast<Maxwell::ShaderStage>(stage), shader, primitive_mode,
+                              current_constbuffer_bindpoint);
 
         // Configure the textures for this shader stage.
         current_texture_bindpoint = SetupTextures(static_cast<Maxwell::ShaderStage>(stage), shader,
-                                                  current_texture_bindpoint);
+                                                  primitive_mode, current_texture_bindpoint);
 
         // When VertexA is enabled, we have dual vertex shaders
         if (program == Maxwell::ShaderProgram::VertexA) {
@@ -317,8 +329,6 @@ void RasterizerOpenGL::SetupShaders() {
     }
 
     state.Apply();
-
-    shader_program_manager->UseTrivialGeometryShader();
 }
 
 std::size_t RasterizerOpenGL::CalculateVertexArraysSize() const {
@@ -389,7 +399,7 @@ void RasterizerOpenGL::UpdatePagesCachedCount(VAddr addr, u64 size, int delta) {
 
 void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_depth_fb,
                                              bool preserve_contents,
-                                             boost::optional<std::size_t> single_color_target) {
+                                             std::optional<std::size_t> single_color_target) {
     MICROPROFILE_SCOPE(OpenGL_Framebuffer);
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
@@ -406,12 +416,23 @@ void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_dep
     // Bind the framebuffer surfaces
     state.draw.draw_framebuffer = framebuffer.handle;
     state.Apply();
+    state.framebuffer_srgb.enabled = regs.framebuffer_srgb != 0;
 
     if (using_color_fb) {
         if (single_color_target) {
             // Used when just a single color attachment is enabled, e.g. for clearing a color buffer
             Surface color_surface =
                 res_cache.GetColorBufferSurface(*single_color_target, preserve_contents);
+
+            if (color_surface) {
+                // Assume that a surface will be written to if it is used as a framebuffer, even if
+                // the shader doesn't actually write to it.
+                color_surface->MarkAsModified(true, res_cache);
+                // Workaround for and issue in nvidia drivers
+                // https://devtalk.nvidia.com/default/topic/776591/opengl/gl_framebuffer_srgb-functions-incorrectly/
+                state.framebuffer_srgb.enabled |= color_surface->GetSurfaceParams().srgb_conversion;
+            }
+
             glFramebufferTexture2D(
                 GL_DRAW_FRAMEBUFFER,
                 GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(*single_color_target), GL_TEXTURE_2D,
@@ -422,6 +443,18 @@ void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_dep
             std::array<GLenum, Maxwell::NumRenderTargets> buffers;
             for (std::size_t index = 0; index < Maxwell::NumRenderTargets; ++index) {
                 Surface color_surface = res_cache.GetColorBufferSurface(index, preserve_contents);
+
+                if (color_surface) {
+                    // Assume that a surface will be written to if it is used as a framebuffer, even
+                    // if the shader doesn't actually write to it.
+                    color_surface->MarkAsModified(true, res_cache);
+                    // Enable sRGB only for supported formats
+                    // Workaround for and issue in nvidia drivers
+                    // https://devtalk.nvidia.com/default/topic/776591/opengl/gl_framebuffer_srgb-functions-incorrectly/
+                    state.framebuffer_srgb.enabled |=
+                        color_surface->GetSurfaceParams().srgb_conversion;
+                }
+
                 buffers[index] = GL_COLOR_ATTACHMENT0 + regs.rt_control.GetMap(index);
                 glFramebufferTexture2D(
                     GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(index),
@@ -441,6 +474,10 @@ void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_dep
     }
 
     if (depth_surface) {
+        // Assume that a surface will be written to if it is used as a framebuffer, even if
+        // the shader doesn't actually write to it.
+        depth_surface->MarkAsModified(true, res_cache);
+
         if (regs.stencil_enable) {
             // Attach both depth and stencil
             glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
@@ -507,7 +544,9 @@ void RasterizerOpenGL::Clear() {
 
     ConfigureFramebuffers(use_color, use_depth || use_stencil, false,
                           regs.clear_buffers.RT.Value());
-
+    // Copy the sRGB setting to the clear state to avoid problem with
+    // specific driver implementations
+    clear_state.framebuffer_srgb.enabled = state.framebuffer_srgb.enabled;
     clear_state.Apply();
 
     if (use_color) {
@@ -540,9 +579,13 @@ void RasterizerOpenGL::DrawArrays() {
     SyncBlendState();
     SyncLogicOpState();
     SyncCullMode();
-    SyncAlphaTest();
+    SyncPrimitiveRestart();
+    SyncDepthRange();
+    SyncScissorTest();
+    // Alpha Testing is synced on shaders.
     SyncTransformFeedback();
     SyncPointState();
+    CheckAlphaTests();
 
     // TODO(bunnei): Sync framebuffer_scale uniform here
     // TODO(bunnei): Sync scissorbox uniform(s) here
@@ -580,7 +623,7 @@ void RasterizerOpenGL::DrawArrays() {
 
     SetupVertexArrays();
     DrawParameters params = SetupDraw();
-    SetupShaders();
+    SetupShaders(params.primitive_mode);
 
     buffer_cache.Unmap();
 
@@ -604,7 +647,14 @@ void RasterizerOpenGL::DrawArrays() {
 
 void RasterizerOpenGL::FlushAll() {}
 
-void RasterizerOpenGL::FlushRegion(VAddr addr, u64 size) {}
+void RasterizerOpenGL::FlushRegion(VAddr addr, u64 size) {
+    MICROPROFILE_SCOPE(OpenGL_CacheManagement);
+
+    if (Settings::values.use_accurate_gpu_emulation) {
+        // Only flush if use_accurate_gpu_emulation is enabled, as it incurs a performance hit
+        res_cache.FlushRegion(addr, size);
+    }
+}
 
 void RasterizerOpenGL::InvalidateRegion(VAddr addr, u64 size) {
     MICROPROFILE_SCOPE(OpenGL_CacheManagement);
@@ -614,17 +664,20 @@ void RasterizerOpenGL::InvalidateRegion(VAddr addr, u64 size) {
 }
 
 void RasterizerOpenGL::FlushAndInvalidateRegion(VAddr addr, u64 size) {
+    FlushRegion(addr, size);
     InvalidateRegion(addr, size);
 }
 
-bool RasterizerOpenGL::AccelerateDisplayTransfer(const void* config) {
+bool RasterizerOpenGL::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
+                                             const Tegra::Engines::Fermi2D::Regs::Surface& dst) {
     MICROPROFILE_SCOPE(OpenGL_Blits);
-    UNREACHABLE();
-    return true;
-}
 
-bool RasterizerOpenGL::AccelerateTextureCopy(const void* config) {
-    UNREACHABLE();
+    if (Settings::values.use_accurate_gpu_emulation) {
+        // Skip the accelerated copy and perform a slow but more accurate copy
+        return false;
+    }
+
+    res_cache.FermiCopySurface(src, dst);
     return true;
 }
 
@@ -648,7 +701,8 @@ bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config,
 
     // Verify that the cached surface is the same size and format as the requested framebuffer
     const auto& params{surface->GetSurfaceParams()};
-    const auto& pixel_format{SurfaceParams::PixelFormatFromGPUPixelFormat(config.pixel_format)};
+    const auto& pixel_format{
+        VideoCore::Surface::PixelFormatFromGPUPixelFormat(config.pixel_format)};
     ASSERT_MSG(params.width == config.width, "Framebuffer width is different");
     ASSERT_MSG(params.height == config.height, "Framebuffer height is different");
     ASSERT_MSG(params.pixel_format == pixel_format, "Framebuffer pixel_format is different");
@@ -662,10 +716,13 @@ void RasterizerOpenGL::SamplerInfo::Create() {
     sampler.Create();
     mag_filter = min_filter = Tegra::Texture::TextureFilter::Linear;
     wrap_u = wrap_v = wrap_p = Tegra::Texture::WrapMode::Wrap;
+    uses_depth_compare = false;
+    depth_compare_func = Tegra::Texture::DepthCompareFunc::Never;
 
     // default is GL_LINEAR_MIPMAP_LINEAR
     glSamplerParameteri(sampler.handle, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
     // Other attributes have correct defaults
+    glSamplerParameteri(sampler.handle, GL_TEXTURE_COMPARE_FUNC, GL_NEVER);
 }
 
 void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntry& config) {
@@ -673,11 +730,15 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr
 
     if (mag_filter != config.mag_filter) {
         mag_filter = config.mag_filter;
-        glSamplerParameteri(s, GL_TEXTURE_MAG_FILTER, MaxwellToGL::TextureFilterMode(mag_filter));
+        glSamplerParameteri(
+            s, GL_TEXTURE_MAG_FILTER,
+            MaxwellToGL::TextureFilterMode(mag_filter, Tegra::Texture::TextureMipmapFilter::None));
     }
-    if (min_filter != config.min_filter) {
+    if (min_filter != config.min_filter || mip_filter != config.mip_filter) {
         min_filter = config.min_filter;
-        glSamplerParameteri(s, GL_TEXTURE_MIN_FILTER, MaxwellToGL::TextureFilterMode(min_filter));
+        mip_filter = config.mip_filter;
+        glSamplerParameteri(s, GL_TEXTURE_MIN_FILTER,
+                            MaxwellToGL::TextureFilterMode(min_filter, mip_filter));
     }
 
     if (wrap_u != config.wrap_u) {
@@ -693,6 +754,21 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr
         glSamplerParameteri(s, GL_TEXTURE_WRAP_R, MaxwellToGL::WrapMode(wrap_p));
     }
 
+    if (uses_depth_compare != (config.depth_compare_enabled == 1)) {
+        uses_depth_compare = (config.depth_compare_enabled == 1);
+        if (uses_depth_compare) {
+            glSamplerParameteri(s, GL_TEXTURE_COMPARE_MODE, GL_COMPARE_REF_TO_TEXTURE);
+        } else {
+            glSamplerParameteri(s, GL_TEXTURE_COMPARE_MODE, GL_NONE);
+        }
+    }
+
+    if (depth_compare_func != config.depth_compare_func) {
+        depth_compare_func = config.depth_compare_func;
+        glSamplerParameteri(s, GL_TEXTURE_COMPARE_FUNC,
+                            MaxwellToGL::DepthCompareFunc(depth_compare_func));
+    }
+
     if (wrap_u == Tegra::Texture::WrapMode::Border || wrap_v == Tegra::Texture::WrapMode::Border ||
         wrap_p == Tegra::Texture::WrapMode::Border) {
         const GLvec4 new_border_color = {{config.border_color_r, config.border_color_g,
@@ -705,7 +781,7 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr
 }
 
 u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, Shader& shader,
-                                        u32 current_bindpoint) {
+                                        GLenum primitive_mode, u32 current_bindpoint) {
     MICROPROFILE_SCOPE(OpenGL_UBO);
     const auto& gpu = Core::System::GetInstance().GPU();
     const auto& maxwell3d = gpu.Maxwell3D();
@@ -757,7 +833,7 @@ u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, Shader& shad
             buffer.address, size, static_cast<std::size_t>(uniform_buffer_alignment));
 
         // Now configure the bindpoint of the buffer inside the shader
-        glUniformBlockBinding(shader->GetProgramHandle(),
+        glUniformBlockBinding(shader->GetProgramHandle(primitive_mode),
                               shader->GetProgramResourceIndex(used_buffer),
                               current_bindpoint + bindpoint);
 
@@ -773,7 +849,8 @@ u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, Shader& shad
     return current_bindpoint + static_cast<u32>(entries.size());
 }
 
-u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader, u32 current_unit) {
+u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
+                                    GLenum primitive_mode, u32 current_unit) {
     MICROPROFILE_SCOPE(OpenGL_Texture);
     const auto& gpu = Core::System::GetInstance().GPU();
     const auto& maxwell3d = gpu.Maxwell3D();
@@ -788,8 +865,8 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
 
         // Bind the uniform to the sampler.
 
-        glProgramUniform1i(shader->GetProgramHandle(), shader->GetUniformLocation(entry),
-                           current_bindpoint);
+        glProgramUniform1i(shader->GetProgramHandle(primitive_mode),
+                           shader->GetUniformLocation(entry), current_bindpoint);
 
         const auto texture = maxwell3d.GetStageTexture(entry.GetStage(), entry.GetOffset());
 
@@ -862,12 +939,18 @@ void RasterizerOpenGL::SyncCullMode() {
     }
 }
 
-void RasterizerOpenGL::SyncDepthScale() {
-    UNREACHABLE();
+void RasterizerOpenGL::SyncPrimitiveRestart() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+
+    state.primitive_restart.enabled = regs.primitive_restart.enabled;
+    state.primitive_restart.index = regs.primitive_restart.index;
 }
 
-void RasterizerOpenGL::SyncDepthOffset() {
-    UNREACHABLE();
+void RasterizerOpenGL::SyncDepthRange() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+
+    state.depth.depth_range_near = regs.viewport->depth_range_near;
+    state.depth.depth_range_far = regs.viewport->depth_range_far;
 }
 
 void RasterizerOpenGL::SyncDepthTestState() {
@@ -947,14 +1030,19 @@ void RasterizerOpenGL::SyncLogicOpState() {
     state.logic_op.operation = MaxwellToGL::LogicOp(regs.logic_op.operation);
 }
 
-void RasterizerOpenGL::SyncAlphaTest() {
+void RasterizerOpenGL::SyncScissorTest() {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
-    // TODO(Rodrigo): Alpha testing is a legacy OpenGL feature, but it can be
-    // implemented with a test+discard in fragment shaders.
-    if (regs.alpha_test_enabled != 0) {
-        LOG_CRITICAL(Render_OpenGL, "Alpha testing is not implemented");
-        UNREACHABLE();
+    state.scissor.enabled = (regs.scissor_test.enable != 0);
+    // TODO(Blinkhawk): Figure if the hardware supports scissor testing per viewport and how it's
+    // implemented.
+    if (regs.scissor_test.enable != 0) {
+        const u32 width = regs.scissor_test.max_x - regs.scissor_test.min_x;
+        const u32 height = regs.scissor_test.max_y - regs.scissor_test.min_y;
+        state.scissor.x = regs.scissor_test.min_x;
+        state.scissor.y = regs.scissor_test.min_y;
+        state.scissor.width = width;
+        state.scissor.height = height;
     }
 }
 
@@ -976,4 +1064,15 @@ void RasterizerOpenGL::SyncPointState() {
     state.point.size = regs.point_size == 0 ? 1 : regs.point_size;
 }
 
+void RasterizerOpenGL::CheckAlphaTests() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+
+    if (regs.alpha_test_enabled != 0 && regs.rt_control.count > 1) {
+        LOG_CRITICAL(
+            Render_OpenGL,
+            "Alpha Testing is enabled with Multiple Render Targets, this behavior is undefined.");
+        UNREACHABLE();
+    }
+}
+
 } // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index bf954bb5d..47097c569 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -8,12 +8,12 @@
 #include <cstddef>
 #include <map>
 #include <memory>
+#include <optional>
 #include <tuple>
 #include <utility>
 #include <vector>
 
 #include <boost/icl/interval_map.hpp>
-#include <boost/optional.hpp>
 #include <boost/range/iterator_range.hpp>
 #include <glad/glad.h>
 
@@ -52,8 +52,8 @@ public:
     void FlushRegion(VAddr addr, u64 size) override;
     void InvalidateRegion(VAddr addr, u64 size) override;
     void FlushAndInvalidateRegion(VAddr addr, u64 size) override;
-    bool AccelerateDisplayTransfer(const void* config) override;
-    bool AccelerateTextureCopy(const void* config) override;
+    bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
+                               const Tegra::Engines::Fermi2D::Regs::Surface& dst) override;
     bool AccelerateFill(const void* config) override;
     bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
                            u32 pixel_stride) override;
@@ -93,9 +93,12 @@ private:
     private:
         Tegra::Texture::TextureFilter mag_filter;
         Tegra::Texture::TextureFilter min_filter;
+        Tegra::Texture::TextureMipmapFilter mip_filter;
         Tegra::Texture::WrapMode wrap_u;
         Tegra::Texture::WrapMode wrap_v;
         Tegra::Texture::WrapMode wrap_p;
+        bool uses_depth_compare;
+        Tegra::Texture::DepthCompareFunc depth_compare_func;
         GLvec4 border_color;
     };
 
@@ -108,7 +111,7 @@ private:
      */
     void ConfigureFramebuffers(bool use_color_fb = true, bool using_depth_fb = true,
                                bool preserve_contents = true,
-                               boost::optional<std::size_t> single_color_target = {});
+                               std::optional<std::size_t> single_color_target = {});
 
     /*
      * Configures the current constbuffers to use for the draw command.
@@ -118,7 +121,7 @@ private:
      * @returns The next available bindpoint for use in the next shader stage.
      */
     u32 SetupConstBuffers(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, Shader& shader,
-                          u32 current_bindpoint);
+                          GLenum primitive_mode, u32 current_bindpoint);
 
     /*
      * Configures the current textures to use for the draw command.
@@ -128,7 +131,7 @@ private:
      * @returns The next available bindpoint for use in the next shader stage.
      */
     u32 SetupTextures(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, Shader& shader,
-                      u32 current_unit);
+                      GLenum primitive_mode, u32 current_unit);
 
     /// Syncs the viewport to match the guest state
     void SyncViewport();
@@ -142,11 +145,11 @@ private:
     /// Syncs the cull mode to match the guest state
     void SyncCullMode();
 
-    /// Syncs the depth scale to match the guest state
-    void SyncDepthScale();
+    /// Syncs the primitve restart to match the guest state
+    void SyncPrimitiveRestart();
 
-    /// Syncs the depth offset to match the guest state
-    void SyncDepthOffset();
+    /// Syncs the depth range to match the guest state
+    void SyncDepthRange();
 
     /// Syncs the depth test state to match the guest state
     void SyncDepthTestState();
@@ -160,8 +163,8 @@ private:
     /// Syncs the LogicOp state to match the guest state
     void SyncLogicOpState();
 
-    /// Syncs the alpha test state to match the guest state
-    void SyncAlphaTest();
+    /// Syncs the scissor test state to match the guest state
+    void SyncScissorTest();
 
     /// Syncs the transform feedback state to match the guest state
     void SyncTransformFeedback();
@@ -169,6 +172,9 @@ private:
     /// Syncs the point state to match the guest state
     void SyncPointState();
 
+    /// Check asserts for alpha testing.
+    void CheckAlphaTests();
+
     bool has_ARB_direct_state_access = false;
     bool has_ARB_multi_bind = false;
     bool has_ARB_separate_shader_objects = false;
@@ -205,7 +211,7 @@ private:
 
     DrawParameters SetupDraw();
 
-    void SetupShaders();
+    void SetupShaders(GLenum primitive_mode);
 
     enum class AccelDraw { Disabled, Arrays, Indexed };
     AccelDraw accelerate_draw = AccelDraw::Disabled;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index ce967c4d6..0e11557c4 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -16,15 +16,21 @@
 #include "core/settings.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_opengl/gl_rasterizer_cache.h"
+#include "video_core/renderer_opengl/utils.h"
+#include "video_core/surface.h"
 #include "video_core/textures/astc.h"
 #include "video_core/textures/decoders.h"
 #include "video_core/utils.h"
 
 namespace OpenGL {
 
-using SurfaceType = SurfaceParams::SurfaceType;
-using PixelFormat = SurfaceParams::PixelFormat;
-using ComponentType = SurfaceParams::ComponentType;
+using VideoCore::Surface::ComponentTypeFromDepthFormat;
+using VideoCore::Surface::ComponentTypeFromRenderTarget;
+using VideoCore::Surface::ComponentTypeFromTexture;
+using VideoCore::Surface::PixelFormatFromDepthFormat;
+using VideoCore::Surface::PixelFormatFromRenderTargetFormat;
+using VideoCore::Surface::PixelFormatFromTextureFormat;
+using VideoCore::Surface::SurfaceTargetFromTextureType;
 
 struct FormatTuple {
     GLint internal_format;
@@ -34,20 +40,62 @@ struct FormatTuple {
     bool compressed;
 };
 
-static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
-    auto& gpu{Core::System::GetInstance().GPU()};
-    const auto cpu_addr{gpu.MemoryManager().GpuToCpuAddress(gpu_addr)};
-    return cpu_addr ? *cpu_addr : 0;
+void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
+    auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
+    const auto cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr_)};
+
+    addr = cpu_addr ? *cpu_addr : 0;
+    gpu_addr = gpu_addr_;
+    size_in_bytes = SizeInBytesRaw();
+
+    if (IsPixelFormatASTC(pixel_format)) {
+        // ASTC is uncompressed in software, in emulated as RGBA8
+        size_in_bytes_gl = width * height * depth * 4;
+    } else {
+        size_in_bytes_gl = SizeInBytesGL();
+    }
+}
+
+std::size_t SurfaceParams::InnerMipmapMemorySize(u32 mip_level, bool force_gl, bool layer_only,
+                                                 bool uncompressed) const {
+    const u32 tile_x{GetDefaultBlockWidth(pixel_format)};
+    const u32 tile_y{GetDefaultBlockHeight(pixel_format)};
+    const u32 bytes_per_pixel{GetBytesPerPixel(pixel_format)};
+    u32 m_depth = (layer_only ? 1U : depth);
+    u32 m_width = MipWidth(mip_level);
+    u32 m_height = MipHeight(mip_level);
+    m_width = uncompressed ? m_width : std::max(1U, (m_width + tile_x - 1) / tile_x);
+    m_height = uncompressed ? m_height : std::max(1U, (m_height + tile_y - 1) / tile_y);
+    m_depth = std::max(1U, m_depth >> mip_level);
+    u32 m_block_height = MipBlockHeight(mip_level);
+    u32 m_block_depth = MipBlockDepth(mip_level);
+    return Tegra::Texture::CalculateSize(force_gl ? false : is_tiled, bytes_per_pixel, m_width,
+                                         m_height, m_depth, m_block_height, m_block_depth);
+}
+
+std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
+                                           bool uncompressed) const {
+    std::size_t block_size_bytes = Tegra::Texture::GetGOBSize() * block_height * block_depth;
+    std::size_t size = 0;
+    for (u32 i = 0; i < max_mip_level; i++) {
+        size += InnerMipmapMemorySize(i, force_gl, layer_only, uncompressed);
+    }
+    if (!force_gl && is_tiled) {
+        size = Common::AlignUp(size, block_size_bytes);
+    }
+    return size;
 }
 
 /*static*/ SurfaceParams SurfaceParams::CreateForTexture(
     const Tegra::Texture::FullTextureInfo& config, const GLShader::SamplerEntry& entry) {
     SurfaceParams params{};
-    params.addr = TryGetCpuAddr(config.tic.Address());
     params.is_tiled = config.tic.IsTiled();
+    params.block_width = params.is_tiled ? config.tic.BlockWidth() : 0,
     params.block_height = params.is_tiled ? config.tic.BlockHeight() : 0,
-    params.pixel_format =
-        PixelFormatFromTextureFormat(config.tic.format, config.tic.r_type.Value());
+    params.block_depth = params.is_tiled ? config.tic.BlockDepth() : 0,
+    params.srgb_conversion = config.tic.IsSrgbConversionEnabled();
+    params.pixel_format = PixelFormatFromTextureFormat(config.tic.format, config.tic.r_type.Value(),
+                                                       params.srgb_conversion);
     params.component_type = ComponentTypeFromTexture(config.tic.r_type.Value());
     params.type = GetFormatType(params.pixel_format);
     params.width = Common::AlignUp(config.tic.Width(), GetCompressionFactor(params.pixel_format));
@@ -85,21 +133,27 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
         break;
     }
 
-    params.size_in_bytes_total = params.SizeInBytesTotal();
-    params.size_in_bytes_2d = params.SizeInBytes2D();
+    params.is_layered = SurfaceTargetIsLayered(params.target);
     params.max_mip_level = config.tic.max_mip_level + 1;
     params.rt = {};
 
+    params.InitCacheParameters(config.tic.Address());
+
     return params;
 }
 
 /*static*/ SurfaceParams SurfaceParams::CreateForFramebuffer(std::size_t index) {
     const auto& config{Core::System::GetInstance().GPU().Maxwell3D().regs.rt[index]};
     SurfaceParams params{};
-    params.addr = TryGetCpuAddr(config.Address());
-    params.is_tiled = true;
-    params.block_height = Tegra::Texture::TICEntry::DefaultBlockHeight;
+
+    params.is_tiled =
+        config.memory_layout.type == Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear;
+    params.block_width = 1 << config.memory_layout.block_width;
+    params.block_height = 1 << config.memory_layout.block_height;
+    params.block_depth = 1 << config.memory_layout.block_depth;
     params.pixel_format = PixelFormatFromRenderTargetFormat(config.format);
+    params.srgb_conversion = config.format == Tegra::RenderTargetFormat::BGRA8_SRGB ||
+                             config.format == Tegra::RenderTargetFormat::RGBA8_SRGB;
     params.component_type = ComponentTypeFromRenderTarget(config.format);
     params.type = GetFormatType(params.pixel_format);
     params.width = config.width;
@@ -107,43 +161,76 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
     params.unaligned_height = config.height;
     params.target = SurfaceTarget::Texture2D;
     params.depth = 1;
-    params.size_in_bytes_total = params.SizeInBytesTotal();
-    params.size_in_bytes_2d = params.SizeInBytes2D();
-    params.max_mip_level = 0;
+    params.max_mip_level = 1;
+    params.is_layered = false;
 
     // Render target specific parameters, not used for caching
     params.rt.index = static_cast<u32>(index);
     params.rt.array_mode = config.array_mode;
     params.rt.layer_stride = config.layer_stride;
+    params.rt.volume = config.volume;
     params.rt.base_layer = config.base_layer;
 
+    params.InitCacheParameters(config.Address());
+
     return params;
 }
 
-/*static*/ SurfaceParams SurfaceParams::CreateForDepthBuffer(u32 zeta_width, u32 zeta_height,
-                                                             Tegra::GPUVAddr zeta_address,
-                                                             Tegra::DepthFormat format) {
+/*static*/ SurfaceParams SurfaceParams::CreateForDepthBuffer(
+    u32 zeta_width, u32 zeta_height, Tegra::GPUVAddr zeta_address, Tegra::DepthFormat format,
+    u32 block_width, u32 block_height, u32 block_depth,
+    Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout type) {
     SurfaceParams params{};
-    params.addr = TryGetCpuAddr(zeta_address);
-    params.is_tiled = true;
-    params.block_height = Tegra::Texture::TICEntry::DefaultBlockHeight;
+
+    params.is_tiled = type == Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear;
+    params.block_width = 1 << std::min(block_width, 5U);
+    params.block_height = 1 << std::min(block_height, 5U);
+    params.block_depth = 1 << std::min(block_depth, 5U);
     params.pixel_format = PixelFormatFromDepthFormat(format);
     params.component_type = ComponentTypeFromDepthFormat(format);
     params.type = GetFormatType(params.pixel_format);
+    params.srgb_conversion = false;
     params.width = zeta_width;
     params.height = zeta_height;
     params.unaligned_height = zeta_height;
     params.target = SurfaceTarget::Texture2D;
     params.depth = 1;
-    params.size_in_bytes_total = params.SizeInBytesTotal();
-    params.size_in_bytes_2d = params.SizeInBytes2D();
-    params.max_mip_level = 0;
+    params.max_mip_level = 1;
+    params.is_layered = false;
+    params.rt = {};
+
+    params.InitCacheParameters(zeta_address);
+
+    return params;
+}
+
+/*static*/ SurfaceParams SurfaceParams::CreateForFermiCopySurface(
+    const Tegra::Engines::Fermi2D::Regs::Surface& config) {
+    SurfaceParams params{};
+
+    params.is_tiled = !config.linear;
+    params.block_width = params.is_tiled ? std::min(config.BlockWidth(), 32U) : 0,
+    params.block_height = params.is_tiled ? std::min(config.BlockHeight(), 32U) : 0,
+    params.block_depth = params.is_tiled ? std::min(config.BlockDepth(), 32U) : 0,
+    params.pixel_format = PixelFormatFromRenderTargetFormat(config.format);
+    params.srgb_conversion = config.format == Tegra::RenderTargetFormat::BGRA8_SRGB ||
+                             config.format == Tegra::RenderTargetFormat::RGBA8_SRGB;
+    params.component_type = ComponentTypeFromRenderTarget(config.format);
+    params.type = GetFormatType(params.pixel_format);
+    params.width = config.width;
+    params.height = config.height;
+    params.unaligned_height = config.height;
+    params.target = SurfaceTarget::Texture2D;
+    params.depth = 1;
+    params.max_mip_level = 1;
     params.rt = {};
 
+    params.InitCacheParameters(config.Address());
+
     return params;
 }
 
-static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{
+static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format_tuples = {{
     {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // ABGR8U
     {GL_RGBA8, GL_RGBA, GL_BYTE, ComponentType::SNorm, false},                     // ABGR8S
     {GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE, ComponentType::UInt, false},   // ABGR8UI
@@ -159,7 +246,7 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
     {GL_R11F_G11F_B10F, GL_RGB, GL_UNSIGNED_INT_10F_11F_11F_REV, ComponentType::Float,
      false},                                                                     // R11FG11FB10F
     {GL_RGBA32UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false}, // RGBA32UI
-    {GL_COMPRESSED_RGB_S3TC_DXT1_EXT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+    {GL_COMPRESSED_RGBA_S3TC_DXT1_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
      true}, // DXT1
     {GL_COMPRESSED_RGBA_S3TC_DXT3_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
      true}, // DXT23
@@ -193,12 +280,31 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
     {GL_RG16I, GL_RG_INTEGER, GL_SHORT, ComponentType::SInt, false},           // RG16I
     {GL_RG16_SNORM, GL_RG, GL_SHORT, ComponentType::SNorm, false},             // RG16S
     {GL_RGB32F, GL_RGB, GL_FLOAT, ComponentType::Float, false},                // RGB32F
-    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // SRGBA8
-    {GL_RG8, GL_RG, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},                       // RG8U
-    {GL_RG8, GL_RG, GL_BYTE, ComponentType::SNorm, false},                                // RG8S
-    {GL_RG32UI, GL_RG_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false},              // RG32UI
-    {GL_R32UI, GL_RED_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false},              // R32UI
-    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X8
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm,
+     false},                                                                   // RGBA8_SRGB
+    {GL_RG8, GL_RG, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},            // RG8U
+    {GL_RG8, GL_RG, GL_BYTE, ComponentType::SNorm, false},                     // RG8S
+    {GL_RG32UI, GL_RG_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false},   // RG32UI
+    {GL_R32UI, GL_RED_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false},   // R32UI
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},        // ASTC_2D_8X8
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},        // ASTC_2D_8X5
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},        // ASTC_2D_5X4
+    {GL_SRGB8_ALPHA8, GL_BGRA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // BGRA8
+    // Compressed sRGB formats
+    {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+     true}, // DXT1_SRGB
+    {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+     true}, // DXT23_SRGB
+    {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+     true}, // DXT45_SRGB
+    {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM_ARB, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8,
+     ComponentType::UNorm, true},                                              // BC7U_SRGB
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_4X4_SRGB
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X8_SRGB
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X5_SRGB
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X4_SRGB
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},        // ASTC_2D_5X5
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X5_SRGB
 
     // Depth formats
     {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F
@@ -214,19 +320,19 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
      ComponentType::Float, false}, // Z32FS8
 }};
 
-static GLenum SurfaceTargetToGL(SurfaceParams::SurfaceTarget target) {
+static GLenum SurfaceTargetToGL(SurfaceTarget target) {
     switch (target) {
-    case SurfaceParams::SurfaceTarget::Texture1D:
+    case SurfaceTarget::Texture1D:
         return GL_TEXTURE_1D;
-    case SurfaceParams::SurfaceTarget::Texture2D:
+    case SurfaceTarget::Texture2D:
         return GL_TEXTURE_2D;
-    case SurfaceParams::SurfaceTarget::Texture3D:
+    case SurfaceTarget::Texture3D:
         return GL_TEXTURE_3D;
-    case SurfaceParams::SurfaceTarget::Texture1DArray:
+    case SurfaceTarget::Texture1DArray:
         return GL_TEXTURE_1D_ARRAY;
-    case SurfaceParams::SurfaceTarget::Texture2DArray:
+    case SurfaceTarget::Texture2DArray:
         return GL_TEXTURE_2D_ARRAY;
-    case SurfaceParams::SurfaceTarget::TextureCubemap:
+    case SurfaceTarget::TextureCubemap:
         return GL_TEXTURE_CUBE_MAP;
     }
     LOG_CRITICAL(Render_OpenGL, "Unimplemented texture target={}", static_cast<u32>(target));
@@ -242,81 +348,44 @@ static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType
     return format;
 }
 
-static bool IsPixelFormatASTC(PixelFormat format) {
-    switch (format) {
-    case PixelFormat::ASTC_2D_4X4:
-    case PixelFormat::ASTC_2D_8X8:
-        return true;
-    default:
-        return false;
-    }
-}
-
-static std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) {
-    switch (format) {
-    case PixelFormat::ASTC_2D_4X4:
-        return {4, 4};
-    case PixelFormat::ASTC_2D_8X8:
-        return {8, 8};
-    default:
-        LOG_CRITICAL(HW_GPU, "Unhandled format: {}", static_cast<u32>(format));
-        UNREACHABLE();
-    }
-}
-
-MathUtil::Rectangle<u32> SurfaceParams::GetRect() const {
-    u32 actual_height{unaligned_height};
+MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
+    u32 actual_height{std::max(1U, unaligned_height >> mip_level)};
     if (IsPixelFormatASTC(pixel_format)) {
         // ASTC formats must stop at the ATSC block size boundary
         actual_height = Common::AlignDown(actual_height, GetASTCBlockSize(pixel_format).second);
     }
-    return {0, actual_height, width, 0};
-}
-
-/// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN
-static bool IsFormatBCn(PixelFormat format) {
-    switch (format) {
-    case PixelFormat::DXT1:
-    case PixelFormat::DXT23:
-    case PixelFormat::DXT45:
-    case PixelFormat::DXN1:
-    case PixelFormat::DXN2SNORM:
-    case PixelFormat::DXN2UNORM:
-    case PixelFormat::BC7U:
-    case PixelFormat::BC6H_UF16:
-    case PixelFormat::BC6H_SF16:
-        return true;
-    }
-    return false;
+    return {0, actual_height, MipWidth(mip_level), 0};
 }
 
 template <bool morton_to_gl, PixelFormat format>
-void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, std::size_t gl_buffer_size,
-                VAddr addr) {
-    constexpr u32 bytes_per_pixel = SurfaceParams::GetFormatBpp(format) / CHAR_BIT;
-    constexpr u32 gl_bytes_per_pixel = CachedSurface::GetGLBytesPerPixel(format);
+void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth, u8* gl_buffer,
+                std::size_t gl_buffer_size, VAddr addr) {
+    constexpr u32 bytes_per_pixel = GetBytesPerPixel(format);
+
+    // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
+    // pixel values.
+    const u32 tile_size_x{GetDefaultBlockWidth(format)};
+    const u32 tile_size_y{GetDefaultBlockHeight(format)};
 
     if (morton_to_gl) {
-        // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
-        // pixel values.
-        const u32 tile_size{IsFormatBCn(format) ? 4U : 1U};
-        const std::vector<u8> data = Tegra::Texture::UnswizzleTexture(
-            addr, tile_size, bytes_per_pixel, stride, height, block_height);
+        const std::vector<u8> data =
+            Tegra::Texture::UnswizzleTexture(addr, tile_size_x, tile_size_y, bytes_per_pixel,
+                                             stride, height, depth, block_height, block_depth);
         const std::size_t size_to_copy{std::min(gl_buffer_size, data.size())};
         memcpy(gl_buffer, data.data(), size_to_copy);
     } else {
-        // TODO(bunnei): Assumes the default rendering GOB size of 16 (128 lines). We should
-        // check the configuration for this and perform more generic un/swizzle
-        LOG_WARNING(Render_OpenGL, "need to use correct swizzle/GOB parameters!");
-        VideoCore::MortonCopyPixels128(stride, height, bytes_per_pixel, gl_bytes_per_pixel,
-                                       Memory::GetPointer(addr), gl_buffer, morton_to_gl);
+        Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x,
+                                         (height + tile_size_y - 1) / tile_size_y, depth,
+                                         bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr),
+                                         gl_buffer, false, block_height, block_depth);
     }
 }
 
-static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
-                            SurfaceParams::MaxPixelFormat>
-    morton_to_gl_fns = {
-        // clang-format off
+using GLConversionArray = std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
+                                     VideoCore::Surface::MaxPixelFormat>;
+
+static constexpr GLConversionArray morton_to_gl_fns = {
+    // clang-format off
         MortonCopy<true, PixelFormat::ABGR8U>,
         MortonCopy<true, PixelFormat::ABGR8S>,
         MortonCopy<true, PixelFormat::ABGR8UI>,
@@ -357,24 +426,35 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
         MortonCopy<true, PixelFormat::RG16I>,
         MortonCopy<true, PixelFormat::RG16S>,
         MortonCopy<true, PixelFormat::RGB32F>,
-        MortonCopy<true, PixelFormat::SRGBA8>,
+        MortonCopy<true, PixelFormat::RGBA8_SRGB>,
         MortonCopy<true, PixelFormat::RG8U>,
         MortonCopy<true, PixelFormat::RG8S>,
         MortonCopy<true, PixelFormat::RG32UI>,
         MortonCopy<true, PixelFormat::R32UI>,
         MortonCopy<true, PixelFormat::ASTC_2D_8X8>,
+        MortonCopy<true, PixelFormat::ASTC_2D_8X5>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X4>,
+        MortonCopy<true, PixelFormat::BGRA8_SRGB>,
+        MortonCopy<true, PixelFormat::DXT1_SRGB>,
+        MortonCopy<true, PixelFormat::DXT23_SRGB>,
+        MortonCopy<true, PixelFormat::DXT45_SRGB>,
+        MortonCopy<true, PixelFormat::BC7U_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_4X4_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X5>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>,
         MortonCopy<true, PixelFormat::Z32F>,
         MortonCopy<true, PixelFormat::Z16>,
         MortonCopy<true, PixelFormat::Z24S8>,
         MortonCopy<true, PixelFormat::S8Z24>,
         MortonCopy<true, PixelFormat::Z32FS8>,
-        // clang-format on
+    // clang-format on
 };
 
-static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
-                            SurfaceParams::MaxPixelFormat>
-    gl_to_morton_fns = {
-        // clang-format off
+static constexpr GLConversionArray gl_to_morton_fns = {
+    // clang-format off
         MortonCopy<false, PixelFormat::ABGR8U>,
         MortonCopy<false, PixelFormat::ABGR8S>,
         MortonCopy<false, PixelFormat::ABGR8UI>,
@@ -388,17 +468,16 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
         MortonCopy<false, PixelFormat::RGBA16UI>,
         MortonCopy<false, PixelFormat::R11FG11FB10F>,
         MortonCopy<false, PixelFormat::RGBA32UI>,
-        // TODO(Subv): Swizzling DXT1/DXT23/DXT45/DXN1/DXN2/BC7U/BC6H_UF16/BC6H_SF16/ASTC_2D_4X4
-        // formats are not supported
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
+        MortonCopy<false, PixelFormat::DXT1>,
+        MortonCopy<false, PixelFormat::DXT23>,
+        MortonCopy<false, PixelFormat::DXT45>,
+        MortonCopy<false, PixelFormat::DXN1>,
+        MortonCopy<false, PixelFormat::DXN2UNORM>,
+        MortonCopy<false, PixelFormat::DXN2SNORM>,
+        MortonCopy<false, PixelFormat::BC7U>,
+        MortonCopy<false, PixelFormat::BC6H_UF16>,
+        MortonCopy<false, PixelFormat::BC6H_SF16>,
+        // TODO(Subv): Swizzling ASTC formats are not supported
         nullptr,
         MortonCopy<false, PixelFormat::G8R8U>,
         MortonCopy<false, PixelFormat::G8R8S>,
@@ -417,20 +496,62 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
         MortonCopy<false, PixelFormat::RG16I>,
         MortonCopy<false, PixelFormat::RG16S>,
         MortonCopy<false, PixelFormat::RGB32F>,
-        MortonCopy<false, PixelFormat::SRGBA8>,
+        MortonCopy<false, PixelFormat::RGBA8_SRGB>,
         MortonCopy<false, PixelFormat::RG8U>,
         MortonCopy<false, PixelFormat::RG8S>,
         MortonCopy<false, PixelFormat::RG32UI>,
         MortonCopy<false, PixelFormat::R32UI>,
         nullptr,
+        nullptr,
+        nullptr,
+        MortonCopy<false, PixelFormat::BGRA8_SRGB>,
+        MortonCopy<false, PixelFormat::DXT1_SRGB>,
+        MortonCopy<false, PixelFormat::DXT23_SRGB>,
+        MortonCopy<false, PixelFormat::DXT45_SRGB>,
+        MortonCopy<false, PixelFormat::BC7U_SRGB>,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
         MortonCopy<false, PixelFormat::Z32F>,
         MortonCopy<false, PixelFormat::Z16>,
         MortonCopy<false, PixelFormat::Z24S8>,
         MortonCopy<false, PixelFormat::S8Z24>,
         MortonCopy<false, PixelFormat::Z32FS8>,
-        // clang-format on
+    // clang-format on
 };
 
+void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params,
+                 std::vector<u8>& gl_buffer, u32 mip_level) {
+    u32 depth = params.MipDepth(mip_level);
+    if (params.target == SurfaceTarget::Texture2D) {
+        // TODO(Blinkhawk): Eliminate this condition once all texture types are implemented.
+        depth = 1U;
+    }
+    if (params.is_layered) {
+        u64 offset = params.GetMipmapLevelOffset(mip_level);
+        u64 offset_gl = 0;
+        u64 layer_size = params.LayerMemorySize();
+        u64 gl_size = params.LayerSizeGL(mip_level);
+        for (u32 i = 0; i < params.depth; i++) {
+            functions[static_cast<std::size_t>(params.pixel_format)](
+                params.MipWidth(mip_level), params.MipBlockHeight(mip_level),
+                params.MipHeight(mip_level), params.MipBlockDepth(mip_level), 1,
+                gl_buffer.data() + offset_gl, gl_size, params.addr + offset);
+            offset += layer_size;
+            offset_gl += gl_size;
+        }
+    } else {
+        u64 offset = params.GetMipmapLevelOffset(mip_level);
+        functions[static_cast<std::size_t>(params.pixel_format)](
+            params.MipWidth(mip_level), params.MipBlockHeight(mip_level),
+            params.MipHeight(mip_level), params.MipBlockDepth(mip_level), depth, gl_buffer.data(),
+            gl_buffer.size(), params.addr + offset);
+    }
+}
+
 static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
                         GLuint read_fb_handle, GLuint draw_fb_handle, GLenum src_attachment = 0,
                         GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
@@ -444,19 +565,21 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
     OpenGLState state;
     state.draw.read_framebuffer = read_fb_handle;
     state.draw.draw_framebuffer = draw_fb_handle;
+    // Set sRGB enabled if the destination surfaces need it
+    state.framebuffer_srgb.enabled = dst_params.srgb_conversion;
     state.Apply();
 
     u32 buffers{};
 
     if (src_params.type == SurfaceType::ColorTexture) {
         switch (src_params.target) {
-        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture2D:
             glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
                                    GL_TEXTURE_2D, src_surface->Texture().handle, 0);
             glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
                                    0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::TextureCubemap:
             glFramebufferTexture2D(
                 GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
                 static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
@@ -465,12 +588,12 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
                 GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
                 static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture2DArray:
             glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
                                       src_surface->Texture().handle, 0, 0);
             glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture3D:
             glFramebufferTexture3D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
                                    SurfaceTargetToGL(src_params.target),
                                    src_surface->Texture().handle, 0, 0);
@@ -486,13 +609,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
         }
 
         switch (dst_params.target) {
-        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture2D:
             glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
                                    GL_TEXTURE_2D, dst_surface->Texture().handle, 0);
             glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
                                    0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::TextureCubemap:
             glFramebufferTexture2D(
                 GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
                 static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
@@ -501,13 +624,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
                 GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
                 static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture2DArray:
             glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
                                       dst_surface->Texture().handle, 0, 0);
             glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
             break;
 
-        case SurfaceParams::SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture3D:
             glFramebufferTexture3D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
                                    SurfaceTargetToGL(dst_params.target),
                                    dst_surface->Texture().handle, 0, 0);
@@ -559,6 +682,18 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
     return true;
 }
 
+static void FastCopySurface(const Surface& src_surface, const Surface& dst_surface) {
+    const auto& src_params{src_surface->GetSurfaceParams()};
+    const auto& dst_params{dst_surface->GetSurfaceParams()};
+
+    const u32 width{std::min(src_params.width, dst_params.width)};
+    const u32 height{std::min(src_params.height, dst_params.height)};
+
+    glCopyImageSubData(src_surface->Texture().handle, SurfaceTargetToGL(src_params.target), 0, 0, 0,
+                       0, dst_surface->Texture().handle, SurfaceTargetToGL(dst_params.target), 0, 0,
+                       0, 0, width, height, 1);
+}
+
 static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
                         GLuint copy_pbo_handle, GLenum src_attachment = 0,
                         GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
@@ -570,22 +705,21 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
     auto source_format = GetFormatTuple(src_params.pixel_format, src_params.component_type);
     auto dest_format = GetFormatTuple(dst_params.pixel_format, dst_params.component_type);
 
-    std::size_t buffer_size =
-        std::max(src_params.size_in_bytes_total, dst_params.size_in_bytes_total);
+    std::size_t buffer_size = std::max(src_params.size_in_bytes, dst_params.size_in_bytes);
 
     glBindBuffer(GL_PIXEL_PACK_BUFFER, copy_pbo_handle);
     glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW_ARB);
     if (source_format.compressed) {
         glGetCompressedTextureImage(src_surface->Texture().handle, src_attachment,
-                                    static_cast<GLsizei>(src_params.size_in_bytes_total), nullptr);
+                                    static_cast<GLsizei>(src_params.size_in_bytes), nullptr);
     } else {
         glGetTextureImage(src_surface->Texture().handle, src_attachment, source_format.format,
-                          source_format.type, static_cast<GLsizei>(src_params.size_in_bytes_total),
+                          source_format.type, static_cast<GLsizei>(src_params.size_in_bytes),
                           nullptr);
     }
     // If the new texture is bigger than the previous one, we need to fill in the rest with data
     // from the CPU.
-    if (src_params.size_in_bytes_total < dst_params.size_in_bytes_total) {
+    if (src_params.size_in_bytes < dst_params.size_in_bytes) {
         // Upload the rest of the memory.
         if (dst_params.is_tiled) {
             // TODO(Subv): We might have to de-tile the subtexture and re-tile it with the rest
@@ -595,12 +729,12 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
             LOG_DEBUG(HW_GPU, "Trying to upload extra texture data from the CPU during "
                               "reinterpretation but the texture is tiled.");
         }
-        std::size_t remaining_size =
-            dst_params.size_in_bytes_total - src_params.size_in_bytes_total;
+        std::size_t remaining_size = dst_params.size_in_bytes - src_params.size_in_bytes;
         std::vector<u8> data(remaining_size);
-        Memory::ReadBlock(dst_params.addr + src_params.size_in_bytes_total, data.data(),
-                          data.size());
-        glBufferSubData(GL_PIXEL_PACK_BUFFER, src_params.size_in_bytes_total, remaining_size,
+        std::memcpy(data.data(), Memory::GetPointer(dst_params.addr + src_params.size_in_bytes),
+                    data.size());
+
+        glBufferSubData(GL_PIXEL_PACK_BUFFER, src_params.size_in_bytes, remaining_size,
                         data.data());
     }
 
@@ -617,21 +751,21 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
         UNREACHABLE();
     } else {
         switch (dst_params.target) {
-        case SurfaceParams::SurfaceTarget::Texture1D:
+        case SurfaceTarget::Texture1D:
             glTextureSubImage1D(dst_surface->Texture().handle, 0, 0, width, dest_format.format,
                                 dest_format.type, nullptr);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture2D:
             glTextureSubImage2D(dst_surface->Texture().handle, 0, 0, 0, width, height,
                                 dest_format.format, dest_format.type, nullptr);
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture2DArray:
             glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0, 0, width, height,
                                 static_cast<GLsizei>(dst_params.depth), dest_format.format,
                                 dest_format.type, nullptr);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::TextureCubemap:
             glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0,
                                 static_cast<GLint>(cubemap_face), width, height, 1,
                                 dest_format.format, dest_format.type, nullptr);
@@ -646,7 +780,8 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
 }
 
 CachedSurface::CachedSurface(const SurfaceParams& params)
-    : params(params), gl_target(SurfaceTargetToGL(params.target)) {
+    : params(params), gl_target(SurfaceTargetToGL(params.target)),
+      cached_size_in_bytes(params.size_in_bytes) {
     texture.Create();
     const auto& rect{params.GetRect()};
 
@@ -667,38 +802,57 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
     if (!format_tuple.compressed) {
         // Only pre-create the texture for non-compressed textures.
         switch (params.target) {
-        case SurfaceParams::SurfaceTarget::Texture1D:
-            glTexStorage1D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format,
-                           rect.GetWidth());
+        case SurfaceTarget::Texture1D:
+            glTexStorage1D(SurfaceTargetToGL(params.target), params.max_mip_level,
+                           format_tuple.internal_format, rect.GetWidth());
             break;
-        case SurfaceParams::SurfaceTarget::Texture2D:
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
-            glTexStorage2D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format,
-                           rect.GetWidth(), rect.GetHeight());
+        case SurfaceTarget::Texture2D:
+        case SurfaceTarget::TextureCubemap:
+            glTexStorage2D(SurfaceTargetToGL(params.target), params.max_mip_level,
+                           format_tuple.internal_format, rect.GetWidth(), rect.GetHeight());
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
-            glTexStorage3D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format,
-                           rect.GetWidth(), rect.GetHeight(), params.depth);
+        case SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture2DArray:
+            glTexStorage3D(SurfaceTargetToGL(params.target), params.max_mip_level,
+                           format_tuple.internal_format, rect.GetWidth(), rect.GetHeight(),
+                           params.depth);
             break;
         default:
             LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
                          static_cast<u32>(params.target));
             UNREACHABLE();
-            glTexStorage2D(GL_TEXTURE_2D, 1, format_tuple.internal_format, rect.GetWidth(),
-                           rect.GetHeight());
+            glTexStorage2D(GL_TEXTURE_2D, params.max_mip_level, format_tuple.internal_format,
+                           rect.GetWidth(), rect.GetHeight());
         }
     }
 
     glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_MIN_FILTER, GL_LINEAR);
+    glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
     glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_MAX_LEVEL,
+                    params.max_mip_level - 1);
+    if (params.max_mip_level == 1) {
+        glTexParameterf(SurfaceTargetToGL(params.target), GL_TEXTURE_LOD_BIAS, 1000.0);
+    }
+
+    LabelGLObject(GL_TEXTURE, texture.handle, params.addr,
+                  SurfaceParams::SurfaceTargetName(params.target));
+
+    // Clamp size to mapped GPU memory region
+    // TODO(bunnei): Super Mario Odyssey maps a 0x40000 byte region and then uses it for a 0x80000
+    // R32F render buffer. We do not yet know if this is a game bug or something else, but this
+    // check is necessary to prevent flushing from overwriting unmapped memory.
 
-    VideoCore::LabelGLObject(GL_TEXTURE, texture.handle, params.addr,
-                             SurfaceParams::SurfaceTargetName(params.target));
+    auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
+    const u64 max_size{memory_manager.GetRegionEnd(params.gpu_addr) - params.gpu_addr};
+    if (cached_size_in_bytes > max_size) {
+        LOG_ERROR(HW_GPU, "Surface size {} exceeds region size {}", params.size_in_bytes, max_size);
+        cached_size_in_bytes = max_size;
+    }
 }
 
-static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) {
+static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height, bool reverse) {
     union S8Z24 {
         BitField<0, 24, u32> z24;
         BitField<24, 8, u32> s8;
@@ -711,22 +865,29 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) {
     };
     static_assert(sizeof(Z24S8) == 4, "Z24S8 is incorrect size");
 
-    S8Z24 input_pixel{};
-    Z24S8 output_pixel{};
-    constexpr auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::S8Z24)};
+    S8Z24 s8z24_pixel{};
+    Z24S8 z24s8_pixel{};
+    constexpr auto bpp{GetBytesPerPixel(PixelFormat::S8Z24)};
     for (std::size_t y = 0; y < height; ++y) {
         for (std::size_t x = 0; x < width; ++x) {
             const std::size_t offset{bpp * (y * width + x)};
-            std::memcpy(&input_pixel, &data[offset], sizeof(S8Z24));
-            output_pixel.s8.Assign(input_pixel.s8);
-            output_pixel.z24.Assign(input_pixel.z24);
-            std::memcpy(&data[offset], &output_pixel, sizeof(Z24S8));
+            if (reverse) {
+                std::memcpy(&z24s8_pixel, &data[offset], sizeof(Z24S8));
+                s8z24_pixel.s8.Assign(z24s8_pixel.s8);
+                s8z24_pixel.z24.Assign(z24s8_pixel.z24);
+                std::memcpy(&data[offset], &s8z24_pixel, sizeof(S8Z24));
+            } else {
+                std::memcpy(&s8z24_pixel, &data[offset], sizeof(S8Z24));
+                z24s8_pixel.s8.Assign(s8z24_pixel.s8);
+                z24s8_pixel.z24.Assign(s8z24_pixel.z24);
+                std::memcpy(&data[offset], &z24s8_pixel, sizeof(Z24S8));
+            }
         }
     }
 }
 
 static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
-    constexpr auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::G8R8U)};
+    constexpr auto bpp{GetBytesPerPixel(PixelFormat::G8R8U)};
     for (std::size_t y = 0; y < height; ++y) {
         for (std::size_t x = 0; x < width; ++x) {
             const std::size_t offset{bpp * (y * width + x)};
@@ -743,20 +904,29 @@ static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
  * typical desktop GPUs.
  */
 static void ConvertFormatAsNeeded_LoadGLBuffer(std::vector<u8>& data, PixelFormat pixel_format,
-                                               u32 width, u32 height) {
+                                               u32 width, u32 height, u32 depth) {
     switch (pixel_format) {
     case PixelFormat::ASTC_2D_4X4:
-    case PixelFormat::ASTC_2D_8X8: {
+    case PixelFormat::ASTC_2D_8X8:
+    case PixelFormat::ASTC_2D_8X5:
+    case PixelFormat::ASTC_2D_5X4:
+    case PixelFormat::ASTC_2D_5X5:
+    case PixelFormat::ASTC_2D_4X4_SRGB:
+    case PixelFormat::ASTC_2D_8X8_SRGB:
+    case PixelFormat::ASTC_2D_8X5_SRGB:
+    case PixelFormat::ASTC_2D_5X4_SRGB:
+    case PixelFormat::ASTC_2D_5X5_SRGB: {
         // Convert ASTC pixel formats to RGBA8, as most desktop GPUs do not support ASTC.
         u32 block_width{};
         u32 block_height{};
         std::tie(block_width, block_height) = GetASTCBlockSize(pixel_format);
-        data = Tegra::Texture::ASTC::Decompress(data, width, height, block_width, block_height);
+        data =
+            Tegra::Texture::ASTC::Decompress(data, width, height, depth, block_width, block_height);
         break;
     }
     case PixelFormat::S8Z24:
         // Convert the S8Z24 depth format to Z24S8, as OpenGL does not support S8Z24.
-        ConvertS8Z24ToZ24S8(data, width, height);
+        ConvertS8Z24ToZ24S8(data, width, height, false);
         break;
 
     case PixelFormat::G8R8U:
@@ -767,79 +937,98 @@ static void ConvertFormatAsNeeded_LoadGLBuffer(std::vector<u8>& data, PixelForma
     }
 }
 
+/**
+ * Helper function to perform software conversion (as needed) when flushing a buffer from OpenGL to
+ * Switch memory. This is for Maxwell pixel formats that cannot be represented as-is in OpenGL or
+ * with typical desktop GPUs.
+ */
+static void ConvertFormatAsNeeded_FlushGLBuffer(std::vector<u8>& data, PixelFormat pixel_format,
+                                                u32 width, u32 height) {
+    switch (pixel_format) {
+    case PixelFormat::G8R8U:
+    case PixelFormat::G8R8S:
+    case PixelFormat::ASTC_2D_4X4:
+    case PixelFormat::ASTC_2D_8X8:
+    case PixelFormat::ASTC_2D_4X4_SRGB:
+    case PixelFormat::ASTC_2D_8X8_SRGB: {
+        LOG_CRITICAL(HW_GPU, "Conversion of format {} after texture flushing is not implemented",
+                     static_cast<u32>(pixel_format));
+        UNREACHABLE();
+        break;
+    }
+    case PixelFormat::S8Z24:
+        // Convert the Z24S8 depth format to S8Z24, as OpenGL does not support S8Z24.
+        ConvertS8Z24ToZ24S8(data, width, height, true);
+        break;
+    }
+}
+
 MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 64, 192));
 void CachedSurface::LoadGLBuffer() {
-    ASSERT(params.type != SurfaceType::Fill);
-
-    const u8* const texture_src_data = Memory::GetPointer(params.addr);
-
-    ASSERT(texture_src_data);
-
-    const u32 bytes_per_pixel = GetGLBytesPerPixel(params.pixel_format);
-    const u32 copy_size = params.width * params.height * bytes_per_pixel;
-    const std::size_t total_size = copy_size * params.depth;
-
     MICROPROFILE_SCOPE(OpenGL_SurfaceLoad);
-
+    gl_buffer.resize(params.max_mip_level);
+    for (u32 i = 0; i < params.max_mip_level; i++)
+        gl_buffer[i].resize(params.GetMipmapSizeGL(i));
     if (params.is_tiled) {
-        gl_buffer.resize(total_size);
-
-        // TODO(bunnei): This only unswizzles and copies a 2D texture - we do not yet know how to do
-        // this for 3D textures, etc.
-        switch (params.target) {
-        case SurfaceParams::SurfaceTarget::Texture2D:
-            // Pass impl. to the fallback code below
-            break;
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
-            for (std::size_t index = 0; index < params.depth; ++index) {
-                const std::size_t offset{index * copy_size};
-                morton_to_gl_fns[static_cast<std::size_t>(params.pixel_format)](
-                    params.width, params.block_height, params.height, gl_buffer.data() + offset,
-                    copy_size, params.addr + offset);
-            }
-            break;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented tiled load for target={}",
-                         static_cast<u32>(params.target));
-            UNREACHABLE();
-        }
-
-        morton_to_gl_fns[static_cast<std::size_t>(params.pixel_format)](
-            params.width, params.block_height, params.height, gl_buffer.data(), copy_size,
-            params.addr);
+        ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
+                   params.block_width, static_cast<u32>(params.target));
+        for (u32 i = 0; i < params.max_mip_level; i++)
+            SwizzleFunc(morton_to_gl_fns, params, gl_buffer[i], i);
     } else {
-        const u8* const texture_src_data_end{texture_src_data + total_size};
-        gl_buffer.assign(texture_src_data, texture_src_data_end);
+        const auto texture_src_data{Memory::GetPointer(params.addr)};
+        const auto texture_src_data_end{texture_src_data + params.size_in_bytes_gl};
+        gl_buffer[0].assign(texture_src_data, texture_src_data_end);
     }
-
-    ConvertFormatAsNeeded_LoadGLBuffer(gl_buffer, params.pixel_format, params.width, params.height);
+    for (u32 i = 0; i < params.max_mip_level; i++)
+        ConvertFormatAsNeeded_LoadGLBuffer(gl_buffer[i], params.pixel_format, params.MipWidth(i),
+                                           params.MipHeight(i), params.MipDepth(i));
 }
 
 MICROPROFILE_DEFINE(OpenGL_SurfaceFlush, "OpenGL", "Surface Flush", MP_RGB(128, 192, 64));
 void CachedSurface::FlushGLBuffer() {
-    ASSERT_MSG(false, "Unimplemented");
-}
+    MICROPROFILE_SCOPE(OpenGL_SurfaceFlush);
 
-MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 64, 192));
-void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle) {
-    if (params.type == SurfaceType::Fill)
-        return;
+    ASSERT_MSG(!IsPixelFormatASTC(params.pixel_format), "Unimplemented");
 
-    MICROPROFILE_SCOPE(OpenGL_TextureUL);
+    // OpenGL temporary buffer needs to be big enough to store raw texture size
+    gl_buffer.resize(1);
+    gl_buffer[0].resize(GetSizeInBytes());
 
-    ASSERT(gl_buffer.size() == static_cast<std::size_t>(params.width) * params.height *
-                                   GetGLBytesPerPixel(params.pixel_format) * params.depth);
+    const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
+    // Ensure no bad interactions with GL_UNPACK_ALIGNMENT
+    ASSERT(params.width * GetBytesPerPixel(params.pixel_format) % 4 == 0);
+    glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(params.width));
+    ASSERT(!tuple.compressed);
+    glBindBuffer(GL_PIXEL_PACK_BUFFER, 0);
+    glGetTextureImage(texture.handle, 0, tuple.format, tuple.type,
+                      static_cast<GLsizei>(gl_buffer[0].size()), gl_buffer[0].data());
+    glPixelStorei(GL_PACK_ROW_LENGTH, 0);
+    ConvertFormatAsNeeded_FlushGLBuffer(gl_buffer[0], params.pixel_format, params.width,
+                                        params.height);
+    ASSERT(params.type != SurfaceType::Fill);
+    const u8* const texture_src_data = Memory::GetPointer(params.addr);
+    ASSERT(texture_src_data);
+    if (params.is_tiled) {
+        ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
+                   params.block_width, static_cast<u32>(params.target));
 
-    const auto& rect{params.GetRect()};
+        SwizzleFunc(gl_to_morton_fns, params, gl_buffer[0], 0);
+    } else {
+        std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer[0].data(), GetSizeInBytes());
+    }
+}
+
+void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
+                                          GLuint draw_fb_handle) {
+    const auto& rect{params.GetRect(mip_map)};
 
     // Load data from memory to the surface
     const GLint x0 = static_cast<GLint>(rect.left);
     const GLint y0 = static_cast<GLint>(rect.bottom);
     std::size_t buffer_offset =
-        static_cast<std::size_t>(static_cast<std::size_t>(y0) * params.width +
+        static_cast<std::size_t>(static_cast<std::size_t>(y0) * params.MipWidth(mip_map) +
                                  static_cast<std::size_t>(x0)) *
-        GetGLBytesPerPixel(params.pixel_format);
+        GetBytesPerPixel(params.pixel_format);
 
     const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
     const GLuint target_tex = texture.handle;
@@ -855,88 +1044,116 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
     cur_state.Apply();
 
     // Ensure no bad interactions with GL_UNPACK_ALIGNMENT
-    ASSERT(params.width * GetGLBytesPerPixel(params.pixel_format) % 4 == 0);
-    glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.width));
+    ASSERT(params.MipWidth(mip_map) * GetBytesPerPixel(params.pixel_format) % 4 == 0);
+    glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.MipWidth(mip_map)));
 
+    GLsizei image_size = static_cast<GLsizei>(params.GetMipmapSizeGL(mip_map, false));
     glActiveTexture(GL_TEXTURE0);
     if (tuple.compressed) {
         switch (params.target) {
-        case SurfaceParams::SurfaceTarget::Texture2D:
-            glCompressedTexImage2D(
-                SurfaceTargetToGL(params.target), 0, tuple.internal_format,
-                static_cast<GLsizei>(params.width), static_cast<GLsizei>(params.height), 0,
-                static_cast<GLsizei>(params.size_in_bytes_2d), &gl_buffer[buffer_offset]);
+        case SurfaceTarget::Texture2D:
+            glCompressedTexImage2D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
+                                   static_cast<GLsizei>(params.MipWidth(mip_map)),
+                                   static_cast<GLsizei>(params.MipHeight(mip_map)), 0, image_size,
+                                   &gl_buffer[mip_map][buffer_offset]);
+            break;
+        case SurfaceTarget::Texture3D:
+            glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
+                                   static_cast<GLsizei>(params.MipWidth(mip_map)),
+                                   static_cast<GLsizei>(params.MipHeight(mip_map)),
+                                   static_cast<GLsizei>(params.MipDepth(mip_map)), 0, image_size,
+                                   &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
-            glCompressedTexImage3D(
-                SurfaceTargetToGL(params.target), 0, tuple.internal_format,
-                static_cast<GLsizei>(params.width), static_cast<GLsizei>(params.height),
-                static_cast<GLsizei>(params.depth), 0,
-                static_cast<GLsizei>(params.size_in_bytes_total), &gl_buffer[buffer_offset]);
+        case SurfaceTarget::Texture2DArray:
+            glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
+                                   static_cast<GLsizei>(params.MipWidth(mip_map)),
+                                   static_cast<GLsizei>(params.MipHeight(mip_map)),
+                                   static_cast<GLsizei>(params.depth), 0, image_size,
+                                   &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::TextureCubemap: {
+            GLsizei layer_size = static_cast<GLsizei>(params.LayerSizeGL(mip_map));
             for (std::size_t face = 0; face < params.depth; ++face) {
                 glCompressedTexImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face),
-                                       0, tuple.internal_format, static_cast<GLsizei>(params.width),
-                                       static_cast<GLsizei>(params.height), 0,
-                                       static_cast<GLsizei>(params.size_in_bytes_2d),
-                                       &gl_buffer[buffer_offset]);
-                buffer_offset += params.size_in_bytes_2d;
+                                       mip_map, tuple.internal_format,
+                                       static_cast<GLsizei>(params.MipWidth(mip_map)),
+                                       static_cast<GLsizei>(params.MipHeight(mip_map)), 0,
+                                       layer_size, &gl_buffer[mip_map][buffer_offset]);
+                buffer_offset += layer_size;
             }
             break;
+        }
         default:
             LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
                          static_cast<u32>(params.target));
             UNREACHABLE();
-            glCompressedTexImage2D(
-                GL_TEXTURE_2D, 0, tuple.internal_format, static_cast<GLsizei>(params.width),
-                static_cast<GLsizei>(params.height), 0,
-                static_cast<GLsizei>(params.size_in_bytes_2d), &gl_buffer[buffer_offset]);
+            glCompressedTexImage2D(GL_TEXTURE_2D, mip_map, tuple.internal_format,
+                                   static_cast<GLsizei>(params.MipWidth(mip_map)),
+                                   static_cast<GLsizei>(params.MipHeight(mip_map)), 0,
+                                   static_cast<GLsizei>(params.size_in_bytes_gl),
+                                   &gl_buffer[mip_map][buffer_offset]);
         }
     } else {
 
         switch (params.target) {
-        case SurfaceParams::SurfaceTarget::Texture1D:
-            glTexSubImage1D(SurfaceTargetToGL(params.target), 0, x0,
+        case SurfaceTarget::Texture1D:
+            glTexSubImage1D(SurfaceTargetToGL(params.target), mip_map, x0,
                             static_cast<GLsizei>(rect.GetWidth()), tuple.format, tuple.type,
-                            &gl_buffer[buffer_offset]);
+                            &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2D:
-            glTexSubImage2D(SurfaceTargetToGL(params.target), 0, x0, y0,
+        case SurfaceTarget::Texture2D:
+            glTexSubImage2D(SurfaceTargetToGL(params.target), mip_map, x0, y0,
                             static_cast<GLsizei>(rect.GetWidth()),
                             static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
-                            &gl_buffer[buffer_offset]);
+                            &gl_buffer[mip_map][buffer_offset]);
+            break;
+        case SurfaceTarget::Texture3D:
+            glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0,
+                            static_cast<GLsizei>(rect.GetWidth()),
+                            static_cast<GLsizei>(rect.GetHeight()), params.MipDepth(mip_map),
+                            tuple.format, tuple.type, &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
-            glTexSubImage3D(SurfaceTargetToGL(params.target), 0, x0, y0, 0,
+        case SurfaceTarget::Texture2DArray:
+            glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0,
                             static_cast<GLsizei>(rect.GetWidth()),
                             static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format,
-                            tuple.type, &gl_buffer[buffer_offset]);
+                            tuple.type, &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::TextureCubemap: {
+            std::size_t start = buffer_offset;
             for (std::size_t face = 0; face < params.depth; ++face) {
-                glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), 0, x0,
-                                y0, static_cast<GLsizei>(rect.GetWidth()),
+                glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), mip_map,
+                                x0, y0, static_cast<GLsizei>(rect.GetWidth()),
                                 static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
-                                &gl_buffer[buffer_offset]);
-                buffer_offset += params.size_in_bytes_2d;
+                                &gl_buffer[mip_map][buffer_offset]);
+                buffer_offset += params.LayerSizeGL(mip_map);
             }
             break;
+        }
         default:
             LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
                          static_cast<u32>(params.target));
             UNREACHABLE();
-            glTexSubImage2D(GL_TEXTURE_2D, 0, x0, y0, static_cast<GLsizei>(rect.GetWidth()),
+            glTexSubImage2D(GL_TEXTURE_2D, mip_map, x0, y0, static_cast<GLsizei>(rect.GetWidth()),
                             static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
-                            &gl_buffer[buffer_offset]);
+                            &gl_buffer[mip_map][buffer_offset]);
         }
     }
 
     glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
 }
 
+MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 64, 192));
+void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle) {
+    if (params.type == SurfaceType::Fill)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_TextureUL);
+
+    for (u32 i = 0; i < params.max_mip_level; i++)
+        UploadGLMipmapTexture(i, read_fb_handle, draw_fb_handle);
+}
+
 RasterizerCacheOpenGL::RasterizerCacheOpenGL() {
     read_framebuffer.Create();
     draw_framebuffer.Create();
@@ -955,7 +1172,9 @@ Surface RasterizerCacheOpenGL::GetDepthBufferSurface(bool preserve_contents) {
     }
 
     SurfaceParams depth_params{SurfaceParams::CreateForDepthBuffer(
-        regs.zeta_width, regs.zeta_height, regs.zeta.Address(), regs.zeta.format)};
+        regs.zeta_width, regs.zeta_height, regs.zeta.Address(), regs.zeta.format,
+        regs.zeta.memory_layout.block_width, regs.zeta.memory_layout.block_height,
+        regs.zeta.memory_layout.block_depth, regs.zeta.memory_layout.type)};
 
     return GetSurface(depth_params, preserve_contents);
 }
@@ -981,10 +1200,7 @@ Surface RasterizerCacheOpenGL::GetColorBufferSurface(std::size_t index, bool pre
 void RasterizerCacheOpenGL::LoadSurface(const Surface& surface) {
     surface->LoadGLBuffer();
     surface->UploadGLTexture(read_framebuffer.handle, draw_framebuffer.handle);
-}
-
-void RasterizerCacheOpenGL::FlushSurface(const Surface& surface) {
-    surface->FlushGLBuffer();
+    surface->MarkAsModified(false, *this);
 }
 
 Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool preserve_contents) {
@@ -1001,8 +1217,8 @@ Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool pres
         } else if (preserve_contents) {
             // If surface parameters changed and we care about keeping the previous data, recreate
             // the surface from the old one
-            Unregister(surface);
             Surface new_surface{RecreateSurface(surface, params)};
+            Unregister(surface);
             Register(new_surface);
             return new_surface;
         } else {
@@ -1033,6 +1249,34 @@ Surface RasterizerCacheOpenGL::GetUncachedSurface(const SurfaceParams& params) {
     return surface;
 }
 
+void RasterizerCacheOpenGL::FermiCopySurface(
+    const Tegra::Engines::Fermi2D::Regs::Surface& src_config,
+    const Tegra::Engines::Fermi2D::Regs::Surface& dst_config) {
+
+    const auto& src_params = SurfaceParams::CreateForFermiCopySurface(src_config);
+    const auto& dst_params = SurfaceParams::CreateForFermiCopySurface(dst_config);
+
+    ASSERT(src_params.width == dst_params.width);
+    ASSERT(src_params.height == dst_params.height);
+    ASSERT(src_params.pixel_format == dst_params.pixel_format);
+    ASSERT(src_params.block_height == dst_params.block_height);
+    ASSERT(src_params.is_tiled == dst_params.is_tiled);
+    ASSERT(src_params.depth == dst_params.depth);
+    ASSERT(src_params.depth == 1); // Currently, FastCopySurface only works with 2D surfaces
+    ASSERT(src_params.target == dst_params.target);
+    ASSERT(src_params.rt.index == dst_params.rt.index);
+
+    FastCopySurface(GetSurface(src_params, true), GetSurface(dst_params, false));
+}
+
+void RasterizerCacheOpenGL::AccurateCopySurface(const Surface& src_surface,
+                                                const Surface& dst_surface) {
+    const auto& src_params{src_surface->GetSurfaceParams()};
+    const auto& dst_params{dst_surface->GetSurfaceParams()};
+    FlushRegion(src_params.addr, dst_params.MemorySize());
+    LoadSurface(dst_surface);
+}
+
 Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
                                                const SurfaceParams& new_params) {
     // Verify surface is compatible for blitting
@@ -1041,57 +1285,39 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
     // Get a new surface with the new parameters, and blit the previous surface to it
     Surface new_surface{GetUncachedSurface(new_params)};
 
+    // With use_accurate_gpu_emulation enabled, do an accurate surface copy
+    if (Settings::values.use_accurate_gpu_emulation) {
+        AccurateCopySurface(old_surface, new_surface);
+        return new_surface;
+    }
+
+    // For compatible surfaces, we can just do fast glCopyImageSubData based copy
+    if (old_params.target == new_params.target && old_params.type == new_params.type &&
+        old_params.depth == new_params.depth && old_params.depth == 1 &&
+        GetFormatBpp(old_params.pixel_format) == GetFormatBpp(new_params.pixel_format)) {
+        FastCopySurface(old_surface, new_surface);
+        return new_surface;
+    }
+
     // If the format is the same, just do a framebuffer blit. This is significantly faster than
     // using PBOs. The is also likely less accurate, as textures will be converted rather than
-    // reinterpreted. When use_accurate_framebuffers setting is enabled, perform a more accurate
+    // reinterpreted. When use_accurate_gpu_emulation setting is enabled, perform a more accurate
     // surface copy, where pixels are reinterpreted as a new format (without conversion). This
     // code path uses OpenGL PBOs and is quite slow.
-    const bool is_blit{old_params.pixel_format == new_params.pixel_format ||
-                       !Settings::values.use_accurate_framebuffers};
+    const bool is_blit{old_params.pixel_format == new_params.pixel_format};
 
     switch (new_params.target) {
-    case SurfaceParams::SurfaceTarget::Texture2D:
+    case SurfaceTarget::Texture2D:
         if (is_blit) {
             BlitSurface(old_surface, new_surface, read_framebuffer.handle, draw_framebuffer.handle);
         } else {
             CopySurface(old_surface, new_surface, copy_pbo.handle);
         }
         break;
-    case SurfaceParams::SurfaceTarget::TextureCubemap: {
-        if (old_params.rt.array_mode != 1) {
-            // TODO(bunnei): This is used by Breath of the Wild, I'm not sure how to implement this
-            // yet (array rendering used as a cubemap texture).
-            LOG_CRITICAL(HW_GPU, "Unhandled rendertarget array_mode {}", old_params.rt.array_mode);
-            UNREACHABLE();
-            return new_surface;
-        }
-
-        // This seems to be used for render-to-cubemap texture
-        ASSERT_MSG(old_params.target == SurfaceParams::SurfaceTarget::Texture2D, "Unexpected");
-        ASSERT_MSG(old_params.pixel_format == new_params.pixel_format, "Unexpected");
-        ASSERT_MSG(old_params.rt.base_layer == 0, "Unimplemented");
-
-        // TODO(bunnei): Verify the below - this stride seems to be in 32-bit words, not pixels.
-        // Tested with Splatoon 2, Super Mario Odyssey, and Breath of the Wild.
-        const std::size_t byte_stride{old_params.rt.layer_stride * sizeof(u32)};
-
-        for (std::size_t index = 0; index < new_params.depth; ++index) {
-            Surface face_surface{TryGetReservedSurface(old_params)};
-            ASSERT_MSG(face_surface, "Unexpected");
-
-            if (is_blit) {
-                BlitSurface(face_surface, new_surface, read_framebuffer.handle,
-                            draw_framebuffer.handle, face_surface->GetSurfaceParams().rt.index,
-                            new_params.rt.index, index);
-            } else {
-                CopySurface(face_surface, new_surface, copy_pbo.handle,
-                            face_surface->GetSurfaceParams().rt.index, new_params.rt.index, index);
-            }
-
-            old_params.addr += byte_stride;
-        }
+    case SurfaceTarget::TextureCubemap:
+    case SurfaceTarget::Texture3D:
+        AccurateCopySurface(old_surface, new_surface);
         break;
-    }
     default:
         LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
                      static_cast<u32>(new_params.target));
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
index 49025a3fe..46ad37897 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
@@ -7,16 +7,20 @@
 #include <array>
 #include <map>
 #include <memory>
+#include <string>
 #include <vector>
 
 #include "common/alignment.h"
 #include "common/common_types.h"
 #include "common/hash.h"
 #include "common/math_util.h"
+#include "video_core/engines/fermi_2d.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/rasterizer_cache.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
 #include "video_core/renderer_opengl/gl_shader_gen.h"
+#include "video_core/surface.h"
+#include "video_core/textures/decoders.h"
 #include "video_core/textures/texture.h"
 
 namespace OpenGL {
@@ -25,122 +29,12 @@ class CachedSurface;
 using Surface = std::shared_ptr<CachedSurface>;
 using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, MathUtil::Rectangle<u32>>;
 
-struct SurfaceParams {
-    enum class PixelFormat {
-        ABGR8U = 0,
-        ABGR8S = 1,
-        ABGR8UI = 2,
-        B5G6R5U = 3,
-        A2B10G10R10U = 4,
-        A1B5G5R5U = 5,
-        R8U = 6,
-        R8UI = 7,
-        RGBA16F = 8,
-        RGBA16U = 9,
-        RGBA16UI = 10,
-        R11FG11FB10F = 11,
-        RGBA32UI = 12,
-        DXT1 = 13,
-        DXT23 = 14,
-        DXT45 = 15,
-        DXN1 = 16, // This is also known as BC4
-        DXN2UNORM = 17,
-        DXN2SNORM = 18,
-        BC7U = 19,
-        BC6H_UF16 = 20,
-        BC6H_SF16 = 21,
-        ASTC_2D_4X4 = 22,
-        G8R8U = 23,
-        G8R8S = 24,
-        BGRA8 = 25,
-        RGBA32F = 26,
-        RG32F = 27,
-        R32F = 28,
-        R16F = 29,
-        R16U = 30,
-        R16S = 31,
-        R16UI = 32,
-        R16I = 33,
-        RG16 = 34,
-        RG16F = 35,
-        RG16UI = 36,
-        RG16I = 37,
-        RG16S = 38,
-        RGB32F = 39,
-        SRGBA8 = 40,
-        RG8U = 41,
-        RG8S = 42,
-        RG32UI = 43,
-        R32UI = 44,
-        ASTC_2D_8X8 = 45,
-
-        MaxColorFormat,
-
-        // Depth formats
-        Z32F = 46,
-        Z16 = 47,
-
-        MaxDepthFormat,
-
-        // DepthStencil formats
-        Z24S8 = 48,
-        S8Z24 = 49,
-        Z32FS8 = 50,
-
-        MaxDepthStencilFormat,
-
-        Max = MaxDepthStencilFormat,
-        Invalid = 255,
-    };
-
-    static constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max);
-
-    enum class ComponentType {
-        Invalid = 0,
-        SNorm = 1,
-        UNorm = 2,
-        SInt = 3,
-        UInt = 4,
-        Float = 5,
-    };
-
-    enum class SurfaceType {
-        ColorTexture = 0,
-        Depth = 1,
-        DepthStencil = 2,
-        Fill = 3,
-        Invalid = 4,
-    };
-
-    enum class SurfaceTarget {
-        Texture1D,
-        Texture2D,
-        Texture3D,
-        Texture1DArray,
-        Texture2DArray,
-        TextureCubemap,
-    };
-
-    static SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type) {
-        switch (texture_type) {
-        case Tegra::Texture::TextureType::Texture1D:
-            return SurfaceTarget::Texture1D;
-        case Tegra::Texture::TextureType::Texture2D:
-        case Tegra::Texture::TextureType::Texture2DNoMipmap:
-            return SurfaceTarget::Texture2D;
-        case Tegra::Texture::TextureType::TextureCubemap:
-            return SurfaceTarget::TextureCubemap;
-        case Tegra::Texture::TextureType::Texture1DArray:
-            return SurfaceTarget::Texture1DArray;
-        case Tegra::Texture::TextureType::Texture2DArray:
-            return SurfaceTarget::Texture2DArray;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", static_cast<u32>(texture_type));
-            UNREACHABLE();
-            return SurfaceTarget::Texture2D;
-        }
-    }
+using SurfaceTarget = VideoCore::Surface::SurfaceTarget;
+using SurfaceType = VideoCore::Surface::SurfaceType;
+using PixelFormat = VideoCore::Surface::PixelFormat;
+using ComponentType = VideoCore::Surface::ComponentType;
 
+struct SurfaceParams {
     static std::string SurfaceTargetName(SurfaceTarget target) {
         switch (target) {
         case SurfaceTarget::Texture1D:
@@ -162,549 +56,123 @@ struct SurfaceParams {
         }
     }
 
-    /**
-     * Gets the compression factor for the specified PixelFormat. This applies to just the
-     * "compressed width" and "compressed height", not the overall compression factor of a
-     * compressed image. This is used for maintaining proper surface sizes for compressed
-     * texture formats.
-     */
-    static constexpr u32 GetCompressionFactor(PixelFormat format) {
-        if (format == PixelFormat::Invalid)
-            return 0;
-
-        constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{
-            1, // ABGR8U
-            1, // ABGR8S
-            1, // ABGR8UI
-            1, // B5G6R5U
-            1, // A2B10G10R10U
-            1, // A1B5G5R5U
-            1, // R8U
-            1, // R8UI
-            1, // RGBA16F
-            1, // RGBA16U
-            1, // RGBA16UI
-            1, // R11FG11FB10F
-            1, // RGBA32UI
-            4, // DXT1
-            4, // DXT23
-            4, // DXT45
-            4, // DXN1
-            4, // DXN2UNORM
-            4, // DXN2SNORM
-            4, // BC7U
-            4, // BC6H_UF16
-            4, // BC6H_SF16
-            4, // ASTC_2D_4X4
-            1, // G8R8U
-            1, // G8R8S
-            1, // BGRA8
-            1, // RGBA32F
-            1, // RG32F
-            1, // R32F
-            1, // R16F
-            1, // R16U
-            1, // R16S
-            1, // R16UI
-            1, // R16I
-            1, // RG16
-            1, // RG16F
-            1, // RG16UI
-            1, // RG16I
-            1, // RG16S
-            1, // RGB32F
-            1, // SRGBA8
-            1, // RG8U
-            1, // RG8S
-            1, // RG32UI
-            1, // R32UI
-            4, // ASTC_2D_8X8
-            1, // Z32F
-            1, // Z16
-            1, // Z24S8
-            1, // S8Z24
-            1, // Z32FS8
-        }};
-
-        ASSERT(static_cast<std::size_t>(format) < compression_factor_table.size());
-        return compression_factor_table[static_cast<std::size_t>(format)];
+    u32 GetFormatBpp() const {
+        return VideoCore::Surface::GetFormatBpp(pixel_format);
     }
 
-    static constexpr u32 GetFormatBpp(PixelFormat format) {
-        if (format == PixelFormat::Invalid)
-            return 0;
-
-        constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
-            32,  // ABGR8U
-            32,  // ABGR8S
-            32,  // ABGR8UI
-            16,  // B5G6R5U
-            32,  // A2B10G10R10U
-            16,  // A1B5G5R5U
-            8,   // R8U
-            8,   // R8UI
-            64,  // RGBA16F
-            64,  // RGBA16U
-            64,  // RGBA16UI
-            32,  // R11FG11FB10F
-            128, // RGBA32UI
-            64,  // DXT1
-            128, // DXT23
-            128, // DXT45
-            64,  // DXN1
-            128, // DXN2UNORM
-            128, // DXN2SNORM
-            128, // BC7U
-            128, // BC6H_UF16
-            128, // BC6H_SF16
-            32,  // ASTC_2D_4X4
-            16,  // G8R8U
-            16,  // G8R8S
-            32,  // BGRA8
-            128, // RGBA32F
-            64,  // RG32F
-            32,  // R32F
-            16,  // R16F
-            16,  // R16U
-            16,  // R16S
-            16,  // R16UI
-            16,  // R16I
-            32,  // RG16
-            32,  // RG16F
-            32,  // RG16UI
-            32,  // RG16I
-            32,  // RG16S
-            96,  // RGB32F
-            32,  // SRGBA8
-            16,  // RG8U
-            16,  // RG8S
-            64,  // RG32UI
-            32,  // R32UI
-            16,  // ASTC_2D_8X8
-            32,  // Z32F
-            16,  // Z16
-            32,  // Z24S8
-            32,  // S8Z24
-            64,  // Z32FS8
-        }};
-
-        ASSERT(static_cast<std::size_t>(format) < bpp_table.size());
-        return bpp_table[static_cast<std::size_t>(format)];
-    }
+    /// Returns the rectangle corresponding to this surface
+    MathUtil::Rectangle<u32> GetRect(u32 mip_level = 0) const;
 
-    u32 GetFormatBpp() const {
-        return GetFormatBpp(pixel_format);
-    }
+    /// Returns the total size of this surface in bytes, adjusted for compression
+    std::size_t SizeInBytesRaw(bool ignore_tiled = false) const {
+        const u32 compression_factor{GetCompressionFactor(pixel_format)};
+        const u32 bytes_per_pixel{GetBytesPerPixel(pixel_format)};
+        const size_t uncompressed_size{
+            Tegra::Texture::CalculateSize((ignore_tiled ? false : is_tiled), bytes_per_pixel, width,
+                                          height, depth, block_height, block_depth)};
 
-    static PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
-        switch (format) {
-        case Tegra::DepthFormat::S8_Z24_UNORM:
-            return PixelFormat::S8Z24;
-        case Tegra::DepthFormat::Z24_S8_UNORM:
-            return PixelFormat::Z24S8;
-        case Tegra::DepthFormat::Z32_FLOAT:
-            return PixelFormat::Z32F;
-        case Tegra::DepthFormat::Z16_UNORM:
-            return PixelFormat::Z16;
-        case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
-            return PixelFormat::Z32FS8;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
+        // Divide by compression_factor^2, as height and width are factored by this
+        return uncompressed_size / (compression_factor * compression_factor);
     }
 
-    static PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
-        switch (format) {
-        // TODO (Hexagon12): Converting SRGBA to RGBA is a hack and doesn't completely correct the
-        // gamma.
-        case Tegra::RenderTargetFormat::RGBA8_SRGB:
-        case Tegra::RenderTargetFormat::RGBA8_UNORM:
-            return PixelFormat::ABGR8U;
-        case Tegra::RenderTargetFormat::RGBA8_SNORM:
-            return PixelFormat::ABGR8S;
-        case Tegra::RenderTargetFormat::RGBA8_UINT:
-            return PixelFormat::ABGR8UI;
-        case Tegra::RenderTargetFormat::BGRA8_SRGB:
-        case Tegra::RenderTargetFormat::BGRA8_UNORM:
-            return PixelFormat::BGRA8;
-        case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
-            return PixelFormat::A2B10G10R10U;
-        case Tegra::RenderTargetFormat::RGBA16_FLOAT:
-            return PixelFormat::RGBA16F;
-        case Tegra::RenderTargetFormat::RGBA16_UNORM:
-            return PixelFormat::RGBA16U;
-        case Tegra::RenderTargetFormat::RGBA16_UINT:
-            return PixelFormat::RGBA16UI;
-        case Tegra::RenderTargetFormat::RGBA32_FLOAT:
-            return PixelFormat::RGBA32F;
-        case Tegra::RenderTargetFormat::RG32_FLOAT:
-            return PixelFormat::RG32F;
-        case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
-            return PixelFormat::R11FG11FB10F;
-        case Tegra::RenderTargetFormat::B5G6R5_UNORM:
-            return PixelFormat::B5G6R5U;
-        case Tegra::RenderTargetFormat::BGR5A1_UNORM:
-            return PixelFormat::A1B5G5R5U;
-        case Tegra::RenderTargetFormat::RGBA32_UINT:
-            return PixelFormat::RGBA32UI;
-        case Tegra::RenderTargetFormat::R8_UNORM:
-            return PixelFormat::R8U;
-        case Tegra::RenderTargetFormat::R8_UINT:
-            return PixelFormat::R8UI;
-        case Tegra::RenderTargetFormat::RG16_FLOAT:
-            return PixelFormat::RG16F;
-        case Tegra::RenderTargetFormat::RG16_UINT:
-            return PixelFormat::RG16UI;
-        case Tegra::RenderTargetFormat::RG16_SINT:
-            return PixelFormat::RG16I;
-        case Tegra::RenderTargetFormat::RG16_UNORM:
-            return PixelFormat::RG16;
-        case Tegra::RenderTargetFormat::RG16_SNORM:
-            return PixelFormat::RG16S;
-        case Tegra::RenderTargetFormat::RG8_UNORM:
-            return PixelFormat::RG8U;
-        case Tegra::RenderTargetFormat::RG8_SNORM:
-            return PixelFormat::RG8S;
-        case Tegra::RenderTargetFormat::R16_FLOAT:
-            return PixelFormat::R16F;
-        case Tegra::RenderTargetFormat::R16_UNORM:
-            return PixelFormat::R16U;
-        case Tegra::RenderTargetFormat::R16_SNORM:
-            return PixelFormat::R16S;
-        case Tegra::RenderTargetFormat::R16_UINT:
-            return PixelFormat::R16UI;
-        case Tegra::RenderTargetFormat::R16_SINT:
-            return PixelFormat::R16I;
-        case Tegra::RenderTargetFormat::R32_FLOAT:
-            return PixelFormat::R32F;
-        case Tegra::RenderTargetFormat::R32_UINT:
-            return PixelFormat::R32UI;
-        case Tegra::RenderTargetFormat::RG32_UINT:
-            return PixelFormat::RG32UI;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
+    /// Returns the size of this surface as an OpenGL texture in bytes
+    std::size_t SizeInBytesGL() const {
+        return SizeInBytesRaw(true);
     }
 
-    static PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
-                                                    Tegra::Texture::ComponentType component_type) {
-        // TODO(Subv): Properly implement this
-        switch (format) {
-        case Tegra::Texture::TextureFormat::A8R8G8B8:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::ABGR8U;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::ABGR8S;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::ABGR8UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::B5G6R5:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::B5G6R5U;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::A2B10G10R10:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::A2B10G10R10U;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::A1B5G5R5:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::A1B5G5R5U;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R8:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::R8U;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::R8UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::G8R8:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::G8R8U;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::G8R8S;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R16_G16_B16_A16:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::RGBA16U;
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RGBA16F;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::BF10GF11RF11:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::R11FG11FB10F;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R32_G32_B32_A32:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RGBA32F;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::RGBA32UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R32_G32:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RG32F;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::RG32UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R32_G32_B32:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RGB32F;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R16:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::R16F;
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::R16U;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::R16S;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::R16UI;
-            case Tegra::Texture::ComponentType::SINT:
-                return PixelFormat::R16I;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R32:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::R32F;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::R32UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::ZF32:
-            return PixelFormat::Z32F;
-        case Tegra::Texture::TextureFormat::Z16:
-            return PixelFormat::Z16;
-        case Tegra::Texture::TextureFormat::Z24S8:
-            return PixelFormat::Z24S8;
-        case Tegra::Texture::TextureFormat::DXT1:
-            return PixelFormat::DXT1;
-        case Tegra::Texture::TextureFormat::DXT23:
-            return PixelFormat::DXT23;
-        case Tegra::Texture::TextureFormat::DXT45:
-            return PixelFormat::DXT45;
-        case Tegra::Texture::TextureFormat::DXN1:
-            return PixelFormat::DXN1;
-        case Tegra::Texture::TextureFormat::DXN2:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::DXN2UNORM;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::DXN2SNORM;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::BC7U:
-            return PixelFormat::BC7U;
-        case Tegra::Texture::TextureFormat::BC6H_UF16:
-            return PixelFormat::BC6H_UF16;
-        case Tegra::Texture::TextureFormat::BC6H_SF16:
-            return PixelFormat::BC6H_SF16;
-        case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
-            return PixelFormat::ASTC_2D_4X4;
-        case Tegra::Texture::TextureFormat::ASTC_2D_8X8:
-            return PixelFormat::ASTC_2D_8X8;
-        case Tegra::Texture::TextureFormat::R16_G16:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RG16F;
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::RG16;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::RG16S;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::RG16UI;
-            case Tegra::Texture::ComponentType::SINT:
-                return PixelFormat::RG16I;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}, component_type={}",
-                         static_cast<u32>(format), static_cast<u32>(component_type));
-            UNREACHABLE();
-        }
+    /// Returns the size of this surface as a cube face in bytes
+    std::size_t SizeInBytesCubeFace() const {
+        return size_in_bytes / 6;
     }
 
-    static ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
-        // TODO(Subv): Implement more component types
-        switch (type) {
-        case Tegra::Texture::ComponentType::UNORM:
-            return ComponentType::UNorm;
-        case Tegra::Texture::ComponentType::FLOAT:
-            return ComponentType::Float;
-        case Tegra::Texture::ComponentType::SNORM:
-            return ComponentType::SNorm;
-        case Tegra::Texture::ComponentType::UINT:
-            return ComponentType::UInt;
-        case Tegra::Texture::ComponentType::SINT:
-            return ComponentType::SInt;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented component type={}", static_cast<u32>(type));
-            UNREACHABLE();
-        }
+    /// Returns the size of this surface as an OpenGL cube face in bytes
+    std::size_t SizeInBytesCubeFaceGL() const {
+        return size_in_bytes_gl / 6;
     }
 
-    static ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format) {
-        // TODO(Subv): Implement more render targets
-        switch (format) {
-        case Tegra::RenderTargetFormat::RGBA8_UNORM:
-        case Tegra::RenderTargetFormat::RGBA8_SRGB:
-        case Tegra::RenderTargetFormat::BGRA8_UNORM:
-        case Tegra::RenderTargetFormat::BGRA8_SRGB:
-        case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
-        case Tegra::RenderTargetFormat::R8_UNORM:
-        case Tegra::RenderTargetFormat::RG16_UNORM:
-        case Tegra::RenderTargetFormat::R16_UNORM:
-        case Tegra::RenderTargetFormat::B5G6R5_UNORM:
-        case Tegra::RenderTargetFormat::BGR5A1_UNORM:
-        case Tegra::RenderTargetFormat::RG8_UNORM:
-        case Tegra::RenderTargetFormat::RGBA16_UNORM:
-            return ComponentType::UNorm;
-        case Tegra::RenderTargetFormat::RGBA8_SNORM:
-        case Tegra::RenderTargetFormat::RG16_SNORM:
-        case Tegra::RenderTargetFormat::R16_SNORM:
-        case Tegra::RenderTargetFormat::RG8_SNORM:
-            return ComponentType::SNorm;
-        case Tegra::RenderTargetFormat::RGBA16_FLOAT:
-        case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
-        case Tegra::RenderTargetFormat::RGBA32_FLOAT:
-        case Tegra::RenderTargetFormat::RG32_FLOAT:
-        case Tegra::RenderTargetFormat::RG16_FLOAT:
-        case Tegra::RenderTargetFormat::R16_FLOAT:
-        case Tegra::RenderTargetFormat::R32_FLOAT:
-            return ComponentType::Float;
-        case Tegra::RenderTargetFormat::RGBA32_UINT:
-        case Tegra::RenderTargetFormat::RGBA16_UINT:
-        case Tegra::RenderTargetFormat::RG16_UINT:
-        case Tegra::RenderTargetFormat::R8_UINT:
-        case Tegra::RenderTargetFormat::R16_UINT:
-        case Tegra::RenderTargetFormat::RG32_UINT:
-        case Tegra::RenderTargetFormat::R32_UINT:
-        case Tegra::RenderTargetFormat::RGBA8_UINT:
-            return ComponentType::UInt;
-        case Tegra::RenderTargetFormat::RG16_SINT:
-        case Tegra::RenderTargetFormat::R16_SINT:
-            return ComponentType::SInt;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
+    /// Returns the exact size of memory occupied by the texture in VRAM, including mipmaps.
+    std::size_t MemorySize() const {
+        std::size_t size = InnerMemorySize(false, is_layered);
+        if (is_layered)
+            return size * depth;
+        return size;
     }
 
-    static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
-        switch (format) {
-        case Tegra::FramebufferConfig::PixelFormat::ABGR8:
-            return PixelFormat::ABGR8U;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
+    /// Returns the exact size of the memory occupied by a layer in a texture in VRAM, including
+    /// mipmaps.
+    std::size_t LayerMemorySize() const {
+        return InnerMemorySize(false, true);
     }
 
-    static ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format) {
-        switch (format) {
-        case Tegra::DepthFormat::Z16_UNORM:
-        case Tegra::DepthFormat::S8_Z24_UNORM:
-        case Tegra::DepthFormat::Z24_S8_UNORM:
-            return ComponentType::UNorm;
-        case Tegra::DepthFormat::Z32_FLOAT:
-        case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
-            return ComponentType::Float;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
+    /// Returns the size of a layer of this surface in OpenGL.
+    std::size_t LayerSizeGL(u32 mip_level) const {
+        return InnerMipmapMemorySize(mip_level, true, is_layered, false);
     }
 
-    static SurfaceType GetFormatType(PixelFormat pixel_format) {
-        if (static_cast<std::size_t>(pixel_format) <
-            static_cast<std::size_t>(PixelFormat::MaxColorFormat)) {
-            return SurfaceType::ColorTexture;
-        }
+    std::size_t GetMipmapSizeGL(u32 mip_level, bool ignore_compressed = true) const {
+        std::size_t size = InnerMipmapMemorySize(mip_level, true, is_layered, ignore_compressed);
+        if (is_layered)
+            return size * depth;
+        return size;
+    }
 
-        if (static_cast<std::size_t>(pixel_format) <
-            static_cast<std::size_t>(PixelFormat::MaxDepthFormat)) {
-            return SurfaceType::Depth;
-        }
+    std::size_t GetMipmapLevelOffset(u32 mip_level) const {
+        std::size_t offset = 0;
+        for (u32 i = 0; i < mip_level; i++)
+            offset += InnerMipmapMemorySize(i, false, is_layered);
+        return offset;
+    }
 
-        if (static_cast<std::size_t>(pixel_format) <
-            static_cast<std::size_t>(PixelFormat::MaxDepthStencilFormat)) {
-            return SurfaceType::DepthStencil;
-        }
+    std::size_t GetMipmapLevelOffsetGL(u32 mip_level) const {
+        std::size_t offset = 0;
+        for (u32 i = 0; i < mip_level; i++)
+            offset += InnerMipmapMemorySize(i, true, is_layered);
+        return offset;
+    }
 
-        // TODO(Subv): Implement the other formats
-        ASSERT(false);
+    u32 MipWidth(u32 mip_level) const {
+        return std::max(1U, width >> mip_level);
+    }
 
-        return SurfaceType::Invalid;
+    u32 MipHeight(u32 mip_level) const {
+        return std::max(1U, height >> mip_level);
     }
 
-    /// Returns the rectangle corresponding to this surface
-    MathUtil::Rectangle<u32> GetRect() const;
+    u32 MipDepth(u32 mip_level) const {
+        return is_layered ? depth : std::max(1U, depth >> mip_level);
+    }
 
-    /// Returns the size of this surface as a 2D texture in bytes, adjusted for compression
-    std::size_t SizeInBytes2D() const {
-        const u32 compression_factor{GetCompressionFactor(pixel_format)};
-        ASSERT(width % compression_factor == 0);
-        ASSERT(height % compression_factor == 0);
-        return (width / compression_factor) * (height / compression_factor) *
-               GetFormatBpp(pixel_format) / CHAR_BIT;
+    // Auto block resizing algorithm from:
+    // https://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/drivers/nouveau/nv50/nv50_miptree.c
+    u32 MipBlockHeight(u32 mip_level) const {
+        if (mip_level == 0)
+            return block_height;
+        u32 alt_height = MipHeight(mip_level);
+        u32 h = GetDefaultBlockHeight(pixel_format);
+        u32 blocks_in_y = (alt_height + h - 1) / h;
+        u32 bh = 16;
+        while (bh > 1 && blocks_in_y <= bh * 4) {
+            bh >>= 1;
+        }
+        return bh;
     }
 
-    /// Returns the total size of this surface in bytes, adjusted for compression
-    std::size_t SizeInBytesTotal() const {
-        return SizeInBytes2D() * depth;
+    u32 MipBlockDepth(u32 mip_level) const {
+        if (mip_level == 0)
+            return block_depth;
+        if (is_layered)
+            return 1;
+        u32 depth = MipDepth(mip_level);
+        u32 bd = 32;
+        while (bd > 1 && depth * 2 <= bd) {
+            bd >>= 1;
+        }
+        if (bd == 32) {
+            u32 bh = MipBlockHeight(mip_level);
+            if (bh >= 4)
+                return 16;
+        }
+        return bd;
     }
 
     /// Creates SurfaceParams from a texture configuration
@@ -715,9 +183,14 @@ struct SurfaceParams {
     static SurfaceParams CreateForFramebuffer(std::size_t index);
 
     /// Creates SurfaceParams for a depth buffer configuration
-    static SurfaceParams CreateForDepthBuffer(u32 zeta_width, u32 zeta_height,
-                                              Tegra::GPUVAddr zeta_address,
-                                              Tegra::DepthFormat format);
+    static SurfaceParams CreateForDepthBuffer(
+        u32 zeta_width, u32 zeta_height, Tegra::GPUVAddr zeta_address, Tegra::DepthFormat format,
+        u32 block_width, u32 block_height, u32 block_depth,
+        Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout type);
+
+    /// Creates SurfaceParams for a Fermi2D surface copy
+    static SurfaceParams CreateForFermiCopySurface(
+        const Tegra::Engines::Fermi2D::Regs::Surface& config);
 
     /// Checks if surfaces are compatible for caching
     bool IsCompatibleSurface(const SurfaceParams& other) const {
@@ -726,9 +199,13 @@ struct SurfaceParams {
                         other.depth);
     }
 
-    VAddr addr;
+    /// Initializes parameters for caching, should be called after everything has been initialized
+    void InitCacheParameters(Tegra::GPUVAddr gpu_addr);
+
     bool is_tiled;
+    u32 block_width;
     u32 block_height;
+    u32 block_depth;
     PixelFormat pixel_format;
     ComponentType component_type;
     SurfaceType type;
@@ -736,18 +213,30 @@ struct SurfaceParams {
     u32 height;
     u32 depth;
     u32 unaligned_height;
-    std::size_t size_in_bytes_total;
-    std::size_t size_in_bytes_2d;
     SurfaceTarget target;
     u32 max_mip_level;
+    bool is_layered;
+    bool srgb_conversion;
+    // Parameters used for caching
+    VAddr addr;
+    Tegra::GPUVAddr gpu_addr;
+    std::size_t size_in_bytes;
+    std::size_t size_in_bytes_gl;
 
     // Render target specific parameters, not used in caching
     struct {
         u32 index;
         u32 array_mode;
+        u32 volume;
         u32 layer_stride;
         u32 base_layer;
     } rt;
+
+private:
+    std::size_t InnerMipmapMemorySize(u32 mip_level, bool force_gl = false, bool layer_only = false,
+                                      bool uncompressed = false) const;
+    std::size_t InnerMemorySize(bool force_gl = false, bool layer_only = false,
+                                bool uncompressed = false) const;
 };
 
 }; // namespace OpenGL
@@ -757,7 +246,8 @@ struct SurfaceReserveKey : Common::HashableStruct<OpenGL::SurfaceParams> {
     static SurfaceReserveKey Create(const OpenGL::SurfaceParams& params) {
         SurfaceReserveKey res;
         res.state = params;
-        res.state.rt = {}; // Ignore rt config in caching
+        res.state.gpu_addr = {}; // Ignore GPU vaddr in caching
+        res.state.rt = {};       // Ignore rt config in caching
         return res;
     }
 };
@@ -772,16 +262,20 @@ struct hash<SurfaceReserveKey> {
 
 namespace OpenGL {
 
-class CachedSurface final {
+class CachedSurface final : public RasterizerCacheObject {
 public:
     CachedSurface(const SurfaceParams& params);
 
-    VAddr GetAddr() const {
+    VAddr GetAddr() const override {
         return params.addr;
     }
 
-    std::size_t GetSizeInBytes() const {
-        return params.size_in_bytes_total;
+    std::size_t GetSizeInBytes() const override {
+        return cached_size_in_bytes;
+    }
+
+    void Flush() override {
+        FlushGLBuffer();
     }
 
     const OGLTexture& Texture() const {
@@ -792,13 +286,6 @@ public:
         return gl_target;
     }
 
-    static constexpr unsigned int GetGLBytesPerPixel(SurfaceParams::PixelFormat format) {
-        if (format == SurfaceParams::PixelFormat::Invalid)
-            return 0;
-
-        return SurfaceParams::GetFormatBpp(format) / CHAR_BIT;
-    }
-
     const SurfaceParams& GetSurfaceParams() const {
         return params;
     }
@@ -811,10 +298,13 @@ public:
     void UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle);
 
 private:
+    void UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, GLuint draw_fb_handle);
+
     OGLTexture texture;
-    std::vector<u8> gl_buffer;
+    std::vector<std::vector<u8>> gl_buffer;
     SurfaceParams params;
     GLenum gl_target;
+    std::size_t cached_size_in_bytes;
 };
 
 class RasterizerCacheOpenGL final : public RasterizerCache<Surface> {
@@ -831,12 +321,13 @@ public:
     /// Get the color surface based on the framebuffer configuration and the specified render target
     Surface GetColorBufferSurface(std::size_t index, bool preserve_contents);
 
-    /// Flushes the surface to Switch memory
-    void FlushSurface(const Surface& surface);
-
     /// Tries to find a framebuffer using on the provided CPU address
     Surface TryFindFramebufferSurface(VAddr addr) const;
 
+    /// Copies the contents of one surface to another
+    void FermiCopySurface(const Tegra::Engines::Fermi2D::Regs::Surface& src_config,
+                          const Tegra::Engines::Fermi2D::Regs::Surface& dst_config);
+
 private:
     void LoadSurface(const Surface& surface);
     Surface GetSurface(const SurfaceParams& params, bool preserve_contents = true);
@@ -853,6 +344,9 @@ private:
     /// Tries to get a reserved surface for the specified parameters
     Surface TryGetReservedSurface(const SurfaceParams& params);
 
+    /// Performs a slow but accurate surface copy, flushing to RAM and reinterpreting the data
+    void AccurateCopySurface(const Surface& src_surface, const Surface& dst_surface);
+
     /// The surface reserve is a "backup" cache, this is where we put unique surfaces that have
     /// previously been used. This is to prevent surfaces from being constantly created and
     /// destroyed when used with different surface parameters.
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 7cd8f91e4..9522fd344 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -8,6 +8,7 @@
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_opengl/gl_shader_cache.h"
 #include "video_core/renderer_opengl/gl_shader_manager.h"
+#include "video_core/renderer_opengl/utils.h"
 #include "video_core/utils.h"
 
 namespace OpenGL {
@@ -68,6 +69,10 @@ CachedShader::CachedShader(VAddr addr, Maxwell::ShaderProgram program_type)
         program_result = GLShader::GenerateVertexShader(setup);
         gl_type = GL_VERTEX_SHADER;
         break;
+    case Maxwell::ShaderProgram::Geometry:
+        program_result = GLShader::GenerateGeometryShader(setup);
+        gl_type = GL_GEOMETRY_SHADER;
+        break;
     case Maxwell::ShaderProgram::Fragment:
         program_result = GLShader::GenerateFragmentShader(setup);
         gl_type = GL_FRAGMENT_SHADER;
@@ -80,11 +85,16 @@ CachedShader::CachedShader(VAddr addr, Maxwell::ShaderProgram program_type)
 
     entries = program_result.second;
 
-    OGLShader shader;
-    shader.Create(program_result.first.c_str(), gl_type);
-    program.Create(true, shader.handle);
-    SetShaderUniformBlockBindings(program.handle);
-    VideoCore::LabelGLObject(GL_PROGRAM, program.handle, addr);
+    if (program_type != Maxwell::ShaderProgram::Geometry) {
+        OGLShader shader;
+        shader.Create(program_result.first.c_str(), gl_type);
+        program.Create(true, shader.handle);
+        SetShaderUniformBlockBindings(program.handle);
+        LabelGLObject(GL_PROGRAM, program.handle, addr);
+    } else {
+        // Store shader's code to lazily build it on draw
+        geometry_programs.code = program_result.first;
+    }
 }
 
 GLuint CachedShader::GetProgramResourceIndex(const GLShader::ConstBufferEntry& buffer) {
@@ -110,6 +120,21 @@ GLint CachedShader::GetUniformLocation(const GLShader::SamplerEntry& sampler) {
     return search->second;
 }
 
+GLuint CachedShader::LazyGeometryProgram(OGLProgram& target_program,
+                                         const std::string& glsl_topology,
+                                         const std::string& debug_name) {
+    if (target_program.handle != 0) {
+        return target_program.handle;
+    }
+    const std::string source{geometry_programs.code + "layout (" + glsl_topology + ") in;\n"};
+    OGLShader shader;
+    shader.Create(source.c_str(), GL_GEOMETRY_SHADER);
+    target_program.Create(true, shader.handle);
+    SetShaderUniformBlockBindings(target_program.handle);
+    LabelGLObject(GL_PROGRAM, target_program.handle, addr, debug_name);
+    return target_program.handle;
+};
+
 Shader ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) {
     const VAddr program_addr{GetShaderAddress(program)};
 
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.h b/src/video_core/renderer_opengl/gl_shader_cache.h
index 9bafe43a9..a210f1731 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.h
+++ b/src/video_core/renderer_opengl/gl_shader_cache.h
@@ -7,6 +7,7 @@
 #include <map>
 #include <memory>
 
+#include "common/assert.h"
 #include "common/common_types.h"
 #include "video_core/rasterizer_cache.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
@@ -18,28 +19,52 @@ class CachedShader;
 using Shader = std::shared_ptr<CachedShader>;
 using Maxwell = Tegra::Engines::Maxwell3D::Regs;
 
-class CachedShader final {
+class CachedShader final : public RasterizerCacheObject {
 public:
     CachedShader(VAddr addr, Maxwell::ShaderProgram program_type);
 
-    /// Gets the address of the shader in guest memory, required for cache management
-    VAddr GetAddr() const {
+    VAddr GetAddr() const override {
         return addr;
     }
 
-    /// Gets the size of the shader in guest memory, required for cache management
-    std::size_t GetSizeInBytes() const {
+    std::size_t GetSizeInBytes() const override {
         return GLShader::MAX_PROGRAM_CODE_LENGTH * sizeof(u64);
     }
 
+    // We do not have to flush this cache as things in it are never modified by us.
+    void Flush() override {}
+
     /// Gets the shader entries for the shader
     const GLShader::ShaderEntries& GetShaderEntries() const {
         return entries;
     }
 
     /// Gets the GL program handle for the shader
-    GLuint GetProgramHandle() const {
-        return program.handle;
+    GLuint GetProgramHandle(GLenum primitive_mode) {
+        if (program_type != Maxwell::ShaderProgram::Geometry) {
+            return program.handle;
+        }
+        switch (primitive_mode) {
+        case GL_POINTS:
+            return LazyGeometryProgram(geometry_programs.points, "points", "ShaderPoints");
+        case GL_LINES:
+        case GL_LINE_STRIP:
+            return LazyGeometryProgram(geometry_programs.lines, "lines", "ShaderLines");
+        case GL_LINES_ADJACENCY:
+        case GL_LINE_STRIP_ADJACENCY:
+            return LazyGeometryProgram(geometry_programs.lines_adjacency, "lines_adjacency",
+                                       "ShaderLinesAdjacency");
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            return LazyGeometryProgram(geometry_programs.triangles, "triangles", "ShaderTriangles");
+        case GL_TRIANGLES_ADJACENCY:
+        case GL_TRIANGLE_STRIP_ADJACENCY:
+            return LazyGeometryProgram(geometry_programs.triangles_adjacency, "triangles_adjacency",
+                                       "ShaderLines");
+        default:
+            UNREACHABLE_MSG("Unknown primitive mode.");
+        }
     }
 
     /// Gets the GL program resource location for the specified resource, caching as needed
@@ -49,12 +74,30 @@ public:
     GLint GetUniformLocation(const GLShader::SamplerEntry& sampler);
 
 private:
+    /// Generates a geometry shader or returns one that already exists.
+    GLuint LazyGeometryProgram(OGLProgram& target_program, const std::string& glsl_topology,
+                               const std::string& debug_name);
+
     VAddr addr;
     Maxwell::ShaderProgram program_type;
     GLShader::ShaderSetup setup;
     GLShader::ShaderEntries entries;
+
+    // Non-geometry program.
     OGLProgram program;
 
+    // Geometry programs. These are needed because GLSL needs an input topology but it's not
+    // declared by the hardware. Workaround this issue by generating a different shader per input
+    // topology class.
+    struct {
+        std::string code;
+        OGLProgram points;
+        OGLProgram lines;
+        OGLProgram lines_adjacency;
+        OGLProgram triangles;
+        OGLProgram triangles_adjacency;
+    } geometry_programs;
+
     std::map<u32, GLuint> resource_cache;
     std::map<u32, GLint> uniform_cache;
 };
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index 579a78702..09b003c59 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -3,9 +3,11 @@
 // Refer to the license.txt file included.
 
 #include <map>
+#include <optional>
 #include <set>
 #include <string>
 #include <string_view>
+#include <unordered_set>
 
 #include <fmt/format.h>
 
@@ -29,11 +31,30 @@ using Tegra::Shader::SubOp;
 constexpr u32 PROGRAM_END = MAX_PROGRAM_CODE_LENGTH;
 constexpr u32 PROGRAM_HEADER_SIZE = sizeof(Tegra::Shader::Header);
 
+constexpr u32 MAX_GEOMETRY_BUFFERS = 6;
+constexpr u32 MAX_ATTRIBUTES = 0x100; // Size in vec4s, this value is untested
+
 class DecompileFail : public std::runtime_error {
 public:
     using std::runtime_error::runtime_error;
 };
 
+/// Translate topology
+static std::string GetTopologyName(Tegra::Shader::OutputTopology topology) {
+    switch (topology) {
+    case Tegra::Shader::OutputTopology::PointList:
+        return "points";
+    case Tegra::Shader::OutputTopology::LineStrip:
+        return "line_strip";
+    case Tegra::Shader::OutputTopology::TriangleStrip:
+        return "triangle_strip";
+    default:
+        LOG_CRITICAL(Render_OpenGL, "Unknown output topology {}", static_cast<u32>(topology));
+        UNREACHABLE();
+        return "points";
+    }
+}
+
 /// Describes the behaviour of code path of a given entry point and a return point.
 enum class ExitMethod {
     Undetermined, ///< Internal value. Only occur when analyzing JMP loop.
@@ -123,7 +144,7 @@ private:
         for (u32 offset = begin; offset != end && offset != PROGRAM_END; ++offset) {
             const Instruction instr = {program_code[offset]};
             if (const auto opcode = OpCode::Decode(instr)) {
-                switch (opcode->GetId()) {
+                switch (opcode->get().GetId()) {
                 case OpCode::Id::EXIT: {
                     // The EXIT instruction can be predicated, which means that the shader can
                     // conditionally end on this instruction. We have to consider the case where the
@@ -143,10 +164,11 @@ private:
                     const ExitMethod jmp = Scan(target, end, labels);
                     return exit_method = ParallelExit(no_jmp, jmp);
                 }
-                case OpCode::Id::SSY: {
-                    // The SSY instruction uses a similar encoding as the BRA instruction.
+                case OpCode::Id::SSY:
+                case OpCode::Id::PBK: {
+                    // The SSY and PBK use a similar encoding as the BRA instruction.
                     ASSERT_MSG(instr.bra.constant_buffer == 0,
-                               "Constant buffer SSY is not supported");
+                               "Constant buffer branching is not supported");
                     const u32 target = offset + instr.bra.GetBranchTarget();
                     labels.insert(target);
                     // Continue scanning for an exit method.
@@ -253,8 +275,10 @@ enum class InternalFlag : u64 {
 class GLSLRegisterManager {
 public:
     GLSLRegisterManager(ShaderWriter& shader, ShaderWriter& declarations,
-                        const Maxwell3D::Regs::ShaderStage& stage, const std::string& suffix)
-        : shader{shader}, declarations{declarations}, stage{stage}, suffix{suffix} {
+                        const Maxwell3D::Regs::ShaderStage& stage, const std::string& suffix,
+                        const Tegra::Shader::Header& header)
+        : shader{shader}, declarations{declarations}, stage{stage}, suffix{suffix}, header{header},
+          fixed_pipeline_output_attributes_used{}, local_memory_size{0} {
         BuildRegisterList();
         BuildInputList();
     }
@@ -317,10 +341,10 @@ public:
      */
     void SetRegisterToFloat(const Register& reg, u64 elem, const std::string& value,
                             u64 dest_num_components, u64 value_num_components,
-                            bool is_saturated = false, u64 dest_elem = 0) {
+                            bool is_saturated = false, u64 dest_elem = 0, bool precise = false) {
 
         SetRegister(reg, elem, is_saturated ? "clamp(" + value + ", 0.0, 1.0)" : value,
-                    dest_num_components, value_num_components, dest_elem);
+                    dest_num_components, value_num_components, dest_elem, precise);
     }
 
     /**
@@ -344,28 +368,94 @@ public:
         const std::string func{is_signed ? "intBitsToFloat" : "uintBitsToFloat"};
 
         SetRegister(reg, elem, func + '(' + ConvertIntegerSize(value, size) + ')',
-                    dest_num_components, value_num_components, dest_elem);
+                    dest_num_components, value_num_components, dest_elem, false);
 
         if (sets_cc) {
             const std::string zero_condition = "( " + ConvertIntegerSize(value, size) + " == 0 )";
             SetInternalFlag(InternalFlag::ZeroFlag, zero_condition);
+            LOG_WARNING(HW_GPU, "Control Codes Imcomplete.");
         }
     }
 
     /**
+     * Writes code that does a register assignment to a half float value operation.
+     * @param reg The destination register to use.
+     * @param elem The element to use for the operation.
+     * @param value The code representing the value to assign. Type has to be half float.
+     * @param merge Half float kind of assignment.
+     * @param dest_num_components Number of components in the destination.
+     * @param value_num_components Number of components in the value.
+     * @param is_saturated Optional, when True, saturates the provided value.
+     * @param dest_elem Optional, the destination element to use for the operation.
+     */
+    void SetRegisterToHalfFloat(const Register& reg, u64 elem, const std::string& value,
+                                Tegra::Shader::HalfMerge merge, u64 dest_num_components,
+                                u64 value_num_components, bool is_saturated = false,
+                                u64 dest_elem = 0) {
+        ASSERT_MSG(!is_saturated, "Unimplemented");
+
+        const std::string result = [&]() {
+            switch (merge) {
+            case Tegra::Shader::HalfMerge::H0_H1:
+                return "uintBitsToFloat(packHalf2x16(" + value + "))";
+            case Tegra::Shader::HalfMerge::F32:
+                // Half float instructions take the first component when doing a float cast.
+                return "float(" + value + ".x)";
+            case Tegra::Shader::HalfMerge::Mrg_H0:
+                // TODO(Rodrigo): I guess Mrg_H0 and Mrg_H1 take their respective component from the
+                // pack. I couldn't test this on hardware but it shouldn't really matter since most
+                // of the time when a Mrg_* flag is used both components will be mirrored. That
+                // being said, it deserves a test.
+                return "((" + GetRegisterAsInteger(reg, 0, false) +
+                       " & 0xffff0000) | (packHalf2x16(" + value + ") & 0x0000ffff))";
+            case Tegra::Shader::HalfMerge::Mrg_H1:
+                return "((" + GetRegisterAsInteger(reg, 0, false) +
+                       " & 0x0000ffff) | (packHalf2x16(" + value + ") & 0xffff0000))";
+            default:
+                UNREACHABLE();
+                return std::string("0");
+            }
+        }();
+
+        SetRegister(reg, elem, result, dest_num_components, value_num_components, dest_elem, false);
+    }
+
+    /**
      * Writes code that does a register assignment to input attribute operation. Input attributes
      * are stored as floats, so this may require conversion.
      * @param reg The destination register to use.
      * @param elem The element to use for the operation.
      * @param attribute The input attribute to use as the source value.
+     * @param input_mode The input mode.
+     * @param vertex The register that decides which vertex to read from (used in GS).
      */
     void SetRegisterToInputAttibute(const Register& reg, u64 elem, Attribute::Index attribute,
-                                    const Tegra::Shader::IpaMode& input_mode) {
+                                    const Tegra::Shader::IpaMode& input_mode,
+                                    std::optional<Register> vertex = {}) {
         const std::string dest = GetRegisterAsFloat(reg);
-        const std::string src = GetInputAttribute(attribute, input_mode) + GetSwizzle(elem);
+        const std::string src = GetInputAttribute(attribute, input_mode, vertex) + GetSwizzle(elem);
         shader.AddLine(dest + " = " + src + ';');
     }
 
+    std::string GetLocalMemoryAsFloat(const std::string& index) {
+        return "lmem[" + index + ']';
+    }
+
+    std::string GetLocalMemoryAsInteger(const std::string& index, bool is_signed = false) {
+        const std::string func{is_signed ? "floatToIntBits" : "floatBitsToUint"};
+        return func + "(lmem[" + index + "])";
+    }
+
+    void SetLocalMemoryAsFloat(const std::string& index, const std::string& value) {
+        shader.AddLine("lmem[" + index + "] = " + value + ';');
+    }
+
+    void SetLocalMemoryAsInteger(const std::string& index, const std::string& value,
+                                 bool is_signed = false) {
+        const std::string func{is_signed ? "intBitsToFloat" : "uintBitsToFloat"};
+        shader.AddLine("lmem[" + index + "] = " + func + '(' + value + ");");
+    }
+
     std::string GetControlCode(const Tegra::Shader::ControlCode cc) const {
         switch (cc) {
         case Tegra::Shader::ControlCode::NEU:
@@ -391,16 +481,34 @@ public:
      * are stored as floats, so this may require conversion.
      * @param attribute The destination output attribute.
      * @param elem The element to use for the operation.
-     * @param reg The register to use as the source value.
+     * @param val_reg The register to use as the source value.
+     * @param buf_reg The register that tells which buffer to write to (used in geometry shaders).
      */
-    void SetOutputAttributeToRegister(Attribute::Index attribute, u64 elem, const Register& reg) {
+    void SetOutputAttributeToRegister(Attribute::Index attribute, u64 elem, const Register& val_reg,
+                                      const Register& buf_reg) {
         const std::string dest = GetOutputAttribute(attribute);
-        const std::string src = GetRegisterAsFloat(reg);
+        const std::string src = GetRegisterAsFloat(val_reg);
 
         if (!dest.empty()) {
             // Can happen with unknown/unimplemented output attributes, in which case we ignore the
             // instruction for now.
-            shader.AddLine(dest + GetSwizzle(elem) + " = " + src + ';');
+            if (stage == Maxwell3D::Regs::ShaderStage::Geometry) {
+                // TODO(Rodrigo): nouveau sets some attributes after setting emitting a geometry
+                // shader. These instructions use a dirty register as buffer index. To avoid some
+                // drivers from complaining for the out of boundary writes, guard them.
+                const std::string buf_index{"min(" + GetRegisterAsInteger(buf_reg) + ", " +
+                                            std::to_string(MAX_GEOMETRY_BUFFERS - 1) + ')'};
+                shader.AddLine("amem[" + buf_index + "][" +
+                               std::to_string(static_cast<u32>(attribute)) + ']' +
+                               GetSwizzle(elem) + " = " + src + ';');
+            } else {
+                if (attribute == Attribute::Index::PointSize) {
+                    fixed_pipeline_output_attributes_used.insert(attribute);
+                    shader.AddLine(dest + " = " + src + ';');
+                } else {
+                    shader.AddLine(dest + GetSwizzle(elem) + " = " + src + ';');
+                }
+            }
         }
     }
 
@@ -441,41 +549,127 @@ public:
         }
     }
 
-    /// Add declarations for registers
+    /// Add declarations.
     void GenerateDeclarations(const std::string& suffix) {
+        GenerateVertex();
+        GenerateRegisters(suffix);
+        GenerateLocalMemory();
+        GenerateInternalFlags();
+        GenerateInputAttrs();
+        GenerateOutputAttrs();
+        GenerateConstBuffers();
+        GenerateSamplers();
+        GenerateGeometry();
+    }
+
+    /// Returns a list of constant buffer declarations.
+    std::vector<ConstBufferEntry> GetConstBuffersDeclarations() const {
+        std::vector<ConstBufferEntry> result;
+        std::copy_if(declr_const_buffers.begin(), declr_const_buffers.end(),
+                     std::back_inserter(result), [](const auto& entry) { return entry.IsUsed(); });
+        return result;
+    }
+
+    /// Returns a list of samplers used in the shader.
+    const std::vector<SamplerEntry>& GetSamplers() const {
+        return used_samplers;
+    }
+
+    /// Returns the GLSL sampler used for the input shader sampler, and creates a new one if
+    /// necessary.
+    std::string AccessSampler(const Sampler& sampler, Tegra::Shader::TextureType type,
+                              bool is_array, bool is_shadow) {
+        const auto offset = static_cast<std::size_t>(sampler.index.Value());
+
+        // If this sampler has already been used, return the existing mapping.
+        const auto itr =
+            std::find_if(used_samplers.begin(), used_samplers.end(),
+                         [&](const SamplerEntry& entry) { return entry.GetOffset() == offset; });
+
+        if (itr != used_samplers.end()) {
+            ASSERT(itr->GetType() == type && itr->IsArray() == is_array &&
+                   itr->IsShadow() == is_shadow);
+            return itr->GetName();
+        }
+
+        // Otherwise create a new mapping for this sampler
+        const std::size_t next_index = used_samplers.size();
+        const SamplerEntry entry{stage, offset, next_index, type, is_array, is_shadow};
+        used_samplers.emplace_back(entry);
+        return entry.GetName();
+    }
+
+    void SetLocalMemory(u64 lmem) {
+        local_memory_size = lmem;
+    }
+
+private:
+    /// Generates declarations for registers.
+    void GenerateRegisters(const std::string& suffix) {
         for (const auto& reg : regs) {
             declarations.AddLine(GLSLRegister::GetTypeString() + ' ' + reg.GetPrefixString() +
                                  std::to_string(reg.GetIndex()) + '_' + suffix + " = 0;");
         }
         declarations.AddNewLine();
+    }
+
+    /// Generates declarations for local memory.
+    void GenerateLocalMemory() {
+        if (local_memory_size > 0) {
+            declarations.AddLine("float lmem[" + std::to_string((local_memory_size - 1 + 4) / 4) +
+                                 "];");
+            declarations.AddNewLine();
+        }
+    }
 
+    /// Generates declarations for internal flags.
+    void GenerateInternalFlags() {
         for (u32 ii = 0; ii < static_cast<u64>(InternalFlag::Amount); ii++) {
             const InternalFlag code = static_cast<InternalFlag>(ii);
             declarations.AddLine("bool " + GetInternalFlag(code) + " = false;");
         }
         declarations.AddNewLine();
+    }
 
+    /// Generates declarations for input attributes.
+    void GenerateInputAttrs() {
         for (const auto element : declr_input_attribute) {
             // TODO(bunnei): Use proper number of elements for these
             u32 idx =
                 static_cast<u32>(element.first) - static_cast<u32>(Attribute::Index::Attribute_0);
-            declarations.AddLine("layout(location = " + std::to_string(idx) + ")" +
-                                 GetInputFlags(element.first) + "in vec4 " +
-                                 GetInputAttribute(element.first, element.second) + ';');
+            if (stage != Maxwell3D::Regs::ShaderStage::Vertex) {
+                // If inputs are varyings, add an offset
+                idx += GENERIC_VARYING_START_LOCATION;
+            }
+
+            std::string attr{GetInputAttribute(element.first, element.second)};
+            if (stage == Maxwell3D::Regs::ShaderStage::Geometry) {
+                attr = "gs_" + attr + "[]";
+            }
+            declarations.AddLine("layout (location = " + std::to_string(idx) + ") " +
+                                 GetInputFlags(element.first) + "in vec4 " + attr + ';');
         }
+
         declarations.AddNewLine();
+    }
 
+    /// Generates declarations for output attributes.
+    void GenerateOutputAttrs() {
         for (const auto& index : declr_output_attribute) {
             // TODO(bunnei): Use proper number of elements for these
-            declarations.AddLine("layout(location = " +
-                                 std::to_string(static_cast<u32>(index) -
-                                                static_cast<u32>(Attribute::Index::Attribute_0)) +
-                                 ") out vec4 " + GetOutputAttribute(index) + ';');
+            const u32 idx = static_cast<u32>(index) -
+                            static_cast<u32>(Attribute::Index::Attribute_0) +
+                            GENERIC_VARYING_START_LOCATION;
+            declarations.AddLine("layout (location = " + std::to_string(idx) + ") out vec4 " +
+                                 GetOutputAttribute(index) + ';');
         }
         declarations.AddNewLine();
+    }
 
+    /// Generates declarations for constant buffers.
+    void GenerateConstBuffers() {
         for (const auto& entry : GetConstBuffersDeclarations()) {
-            declarations.AddLine("layout(std140) uniform " + entry.GetName());
+            declarations.AddLine("layout (std140) uniform " + entry.GetName());
             declarations.AddLine('{');
             declarations.AddLine("    vec4 c" + std::to_string(entry.GetIndex()) +
                                  "[MAX_CONSTBUFFER_ELEMENTS];");
@@ -483,7 +677,10 @@ public:
             declarations.AddNewLine();
         }
         declarations.AddNewLine();
+    }
 
+    /// Generates declarations for samplers.
+    void GenerateSamplers() {
         const auto& samplers = GetSamplers();
         for (const auto& sampler : samplers) {
             declarations.AddLine("uniform " + sampler.GetTypeString() + ' ' + sampler.GetName() +
@@ -492,43 +689,56 @@ public:
         declarations.AddNewLine();
     }
 
-    /// Returns a list of constant buffer declarations
-    std::vector<ConstBufferEntry> GetConstBuffersDeclarations() const {
-        std::vector<ConstBufferEntry> result;
-        std::copy_if(declr_const_buffers.begin(), declr_const_buffers.end(),
-                     std::back_inserter(result), [](const auto& entry) { return entry.IsUsed(); });
-        return result;
-    }
+    /// Generates declarations used for geometry shaders.
+    void GenerateGeometry() {
+        if (stage != Maxwell3D::Regs::ShaderStage::Geometry)
+            return;
 
-    /// Returns a list of samplers used in the shader
-    const std::vector<SamplerEntry>& GetSamplers() const {
-        return used_samplers;
-    }
+        declarations.AddLine(
+            "layout (" + GetTopologyName(header.common3.output_topology) +
+            ", max_vertices = " + std::to_string(header.common4.max_output_vertices) + ") out;");
+        declarations.AddNewLine();
 
-    /// Returns the GLSL sampler used for the input shader sampler, and creates a new one if
-    /// necessary.
-    std::string AccessSampler(const Sampler& sampler, Tegra::Shader::TextureType type,
-                              bool is_array) {
-        const std::size_t offset = static_cast<std::size_t>(sampler.index.Value());
+        declarations.AddLine("vec4 amem[" + std::to_string(MAX_GEOMETRY_BUFFERS) + "][" +
+                             std::to_string(MAX_ATTRIBUTES) + "];");
+        declarations.AddNewLine();
 
-        // If this sampler has already been used, return the existing mapping.
-        const auto itr =
-            std::find_if(used_samplers.begin(), used_samplers.end(),
-                         [&](const SamplerEntry& entry) { return entry.GetOffset() == offset; });
+        constexpr char buffer[] = "amem[output_buffer]";
+        declarations.AddLine("void emit_vertex(uint output_buffer) {");
+        ++declarations.scope;
+        for (const auto element : declr_output_attribute) {
+            declarations.AddLine(GetOutputAttribute(element) + " = " + buffer + '[' +
+                                 std::to_string(static_cast<u32>(element)) + "];");
+        }
+
+        declarations.AddLine("position = " + std::string(buffer) + '[' +
+                             std::to_string(static_cast<u32>(Attribute::Index::Position)) + "];");
+
+        // If a geometry shader is attached, it will always flip (it's the last stage before
+        // fragment). For more info about flipping, refer to gl_shader_gen.cpp.
+        declarations.AddLine("position.xy *= viewport_flip.xy;");
+        declarations.AddLine("gl_Position = position;");
+        declarations.AddLine("position.w = 1.0;");
+        declarations.AddLine("EmitVertex();");
+        --declarations.scope;
+        declarations.AddLine('}');
+        declarations.AddNewLine();
+    }
 
-        if (itr != used_samplers.end()) {
-            ASSERT(itr->GetType() == type && itr->IsArray() == is_array);
-            return itr->GetName();
+    void GenerateVertex() {
+        if (stage != Maxwell3D::Regs::ShaderStage::Vertex)
+            return;
+        declarations.AddLine("out gl_PerVertex {");
+        ++declarations.scope;
+        declarations.AddLine("vec4 gl_Position;");
+        for (auto& o : fixed_pipeline_output_attributes_used) {
+            if (o == Attribute::Index::PointSize)
+                declarations.AddLine("float gl_PointSize;");
         }
-
-        // Otherwise create a new mapping for this sampler
-        const std::size_t next_index = used_samplers.size();
-        const SamplerEntry entry{stage, offset, next_index, type, is_array};
-        used_samplers.emplace_back(entry);
-        return entry.GetName();
+        --declarations.scope;
+        declarations.AddLine("};");
     }
 
-private:
     /// Generates code representing a temporary (GPR) register.
     std::string GetRegister(const Register& reg, unsigned elem) {
         if (reg == Register::ZeroIndex) {
@@ -548,7 +758,8 @@ private:
      * @param dest_elem Optional, the destination element to use for the operation.
      */
     void SetRegister(const Register& reg, u64 elem, const std::string& value,
-                     u64 dest_num_components, u64 value_num_components, u64 dest_elem) {
+                     u64 dest_num_components, u64 value_num_components, u64 dest_elem,
+                     bool precise) {
         if (reg == Register::ZeroIndex) {
             LOG_CRITICAL(HW_GPU, "Cannot set Register::ZeroIndex");
             UNREACHABLE();
@@ -565,7 +776,18 @@ private:
             src += GetSwizzle(elem);
         }
 
-        shader.AddLine(dest + " = " + src + ';');
+        if (precise && stage != Maxwell3D::Regs::ShaderStage::Fragment) {
+            shader.AddLine('{');
+            ++shader.scope;
+            // This avoids optimizations of constant propagation and keeps the code as the original
+            // Sadly using the precise keyword causes "linking" errors on fragment shaders.
+            shader.AddLine("precise float tmp = " + src + ';');
+            shader.AddLine(dest + " = tmp;");
+            --shader.scope;
+            shader.AddLine('}');
+        } else {
+            shader.AddLine(dest + " = " + src + ';');
+        }
     }
 
     /// Build the GLSL register list.
@@ -585,11 +807,19 @@ private:
 
     /// Generates code representing an input attribute register.
     std::string GetInputAttribute(Attribute::Index attribute,
-                                  const Tegra::Shader::IpaMode& input_mode) {
+                                  const Tegra::Shader::IpaMode& input_mode,
+                                  std::optional<Register> vertex = {}) {
+        auto GeometryPass = [&](const std::string& name) {
+            if (stage == Maxwell3D::Regs::ShaderStage::Geometry && vertex) {
+                return "gs_" + name + '[' + GetRegisterAsInteger(*vertex, 0, false) + ']';
+            }
+            return name;
+        };
+
         switch (attribute) {
         case Attribute::Index::Position:
             if (stage != Maxwell3D::Regs::ShaderStage::Fragment) {
-                return "position";
+                return GeometryPass("position");
             } else {
                 return "vec4(gl_FragCoord.x, gl_FragCoord.y, gl_FragCoord.z, 1.0)";
             }
@@ -618,7 +848,7 @@ private:
                         UNREACHABLE();
                     }
                 }
-                return "input_attribute_" + std::to_string(index);
+                return GeometryPass("input_attribute_" + std::to_string(index));
             }
 
             LOG_CRITICAL(HW_GPU, "Unhandled input attribute: {}", static_cast<u32>(attribute));
@@ -671,9 +901,11 @@ private:
         return out;
     }
 
-    /// Generates code representing an output attribute register.
+    /// Generates code representing the declaration name of an output attribute register.
     std::string GetOutputAttribute(Attribute::Index attribute) {
         switch (attribute) {
+        case Attribute::Index::PointSize:
+            return "gl_PointSize";
         case Attribute::Index::Position:
             return "position";
         default:
@@ -707,6 +939,9 @@ private:
     std::vector<SamplerEntry> used_samplers;
     const Maxwell3D::Regs::ShaderStage& stage;
     const std::string& suffix;
+    const Tegra::Shader::Header& header;
+    std::unordered_set<Attribute::Index> fixed_pipeline_output_attributes_used;
+    u64 local_memory_size;
 };
 
 class GLSLGenerator {
@@ -716,6 +951,8 @@ public:
         : subroutines(subroutines), program_code(program_code), main_offset(main_offset),
           stage(stage), suffix(suffix) {
         std::memcpy(&header, program_code.data(), sizeof(Tegra::Shader::Header));
+        local_memory_size = header.GetLocalMemorySize();
+        regs.SetLocalMemory(local_memory_size);
         Generate(suffix);
     }
 
@@ -746,9 +983,23 @@ private:
         return fmt::format("uintBitsToFloat({})", instr.alu.GetImm20_32());
     }
 
+    /// Generates code representing a vec2 pair unpacked from a half float immediate
+    static std::string UnpackHalfImmediate(const Instruction& instr, bool negate) {
+        const std::string immediate = GetHalfFloat(std::to_string(instr.half_imm.PackImmediates()));
+        if (!negate) {
+            return immediate;
+        }
+        const std::string negate_first = instr.half_imm.first_negate != 0 ? "-" : "";
+        const std::string negate_second = instr.half_imm.second_negate != 0 ? "-" : "";
+        const std::string negate_vec = "vec2(" + negate_first + "1, " + negate_second + "1)";
+
+        return '(' + immediate + " * " + negate_vec + ')';
+    }
+
     /// Generates code representing a texture sampler.
-    std::string GetSampler(const Sampler& sampler, Tegra::Shader::TextureType type, bool is_array) {
-        return regs.AccessSampler(sampler, type, is_array);
+    std::string GetSampler(const Sampler& sampler, Tegra::Shader::TextureType type, bool is_array,
+                           bool is_shadow) {
+        return regs.AccessSampler(sampler, type, is_array, is_shadow);
     }
 
     /**
@@ -776,7 +1027,7 @@ private:
         // Can't assign to the constant predicate.
         ASSERT(pred != static_cast<u64>(Pred::UnusedIndex));
 
-        const std::string variable = 'p' + std::to_string(pred) + '_' + suffix;
+        std::string variable = 'p' + std::to_string(pred) + '_' + suffix;
         shader.AddLine(variable + " = " + value + ';');
         declr_predicates.insert(std::move(variable));
     }
@@ -881,6 +1132,41 @@ private:
     }
 
     /*
+     * Transforms the input string GLSL operand into an unpacked half float pair.
+     * @note This function returns a float type pair instead of a half float pair. This is because
+     * real half floats are not standardized in GLSL but unpackHalf2x16 (which returns a vec2) is.
+     * @param operand Input operand. It has to be an unsigned integer.
+     * @param type How to unpack the unsigned integer to a half float pair.
+     * @param abs Get the absolute value of unpacked half floats.
+     * @param neg Get the negative value of unpacked half floats.
+     * @returns String corresponding to a half float pair.
+     */
+    static std::string GetHalfFloat(const std::string& operand,
+                                    Tegra::Shader::HalfType type = Tegra::Shader::HalfType::H0_H1,
+                                    bool abs = false, bool neg = false) {
+        // "vec2" calls emitted in this function are intended to alias components.
+        const std::string value = [&]() {
+            switch (type) {
+            case Tegra::Shader::HalfType::H0_H1:
+                return "unpackHalf2x16(" + operand + ')';
+            case Tegra::Shader::HalfType::F32:
+                return "vec2(uintBitsToFloat(" + operand + "))";
+            case Tegra::Shader::HalfType::H0_H0:
+            case Tegra::Shader::HalfType::H1_H1: {
+                const bool high = type == Tegra::Shader::HalfType::H1_H1;
+                const char unpack_index = "xy"[high ? 1 : 0];
+                return "vec2(unpackHalf2x16(" + operand + ")." + unpack_index + ')';
+            }
+            default:
+                UNREACHABLE();
+                return std::string("vec2(0)");
+            }
+        }();
+
+        return GetOperandAbsNeg(value, abs, neg);
+    }
+
+    /*
      * Returns whether the instruction at the specified offset is a 'sched' instruction.
      * Sched instructions always appear before a sequence of 3 instructions.
      */
@@ -1002,28 +1288,47 @@ private:
         shader.AddLine('}');
     }
 
+    static u32 TextureCoordinates(Tegra::Shader::TextureType texture_type) {
+        switch (texture_type) {
+        case Tegra::Shader::TextureType::Texture1D: {
+            return 1;
+        }
+        case Tegra::Shader::TextureType::Texture2D: {
+            return 2;
+        }
+        case Tegra::Shader::TextureType::Texture3D:
+        case Tegra::Shader::TextureType::TextureCube: {
+            return 3;
+        }
+        default:
+            LOG_CRITICAL(HW_GPU, "Unhandled texture type {}", static_cast<u32>(texture_type));
+            UNREACHABLE();
+            return 0;
+        }
+    }
+
     /*
-     * Emits code to push the input target address to the SSY address stack, incrementing the stack
+     * Emits code to push the input target address to the flow address stack, incrementing the stack
      * top.
      */
-    void EmitPushToSSYStack(u32 target) {
+    void EmitPushToFlowStack(u32 target) {
         shader.AddLine('{');
         ++shader.scope;
-        shader.AddLine("ssy_stack[ssy_stack_top] = " + std::to_string(target) + "u;");
-        shader.AddLine("ssy_stack_top++;");
+        shader.AddLine("flow_stack[flow_stack_top] = " + std::to_string(target) + "u;");
+        shader.AddLine("flow_stack_top++;");
         --shader.scope;
         shader.AddLine('}');
     }
 
     /*
-     * Emits code to pop an address from the SSY address stack, setting the jump address to the
+     * Emits code to pop an address from the flow address stack, setting the jump address to the
      * popped address and decrementing the stack top.
      */
-    void EmitPopFromSSYStack() {
+    void EmitPopFromFlowStack() {
         shader.AddLine('{');
         ++shader.scope;
-        shader.AddLine("ssy_stack_top--;");
-        shader.AddLine("jmp_to = ssy_stack[ssy_stack_top];");
+        shader.AddLine("flow_stack_top--;");
+        shader.AddLine("jmp_to = flow_stack[flow_stack_top];");
         shader.AddLine("break;");
         --shader.scope;
         shader.AddLine('}');
@@ -1035,9 +1340,29 @@ private:
 
         ASSERT_MSG(header.ps.omap.sample_mask == 0, "Samplemask write is unimplemented");
 
+        shader.AddLine("if (alpha_test[0] != 0) {");
+        ++shader.scope;
+        // We start on the register containing the alpha value in the first RT.
+        u32 current_reg = 3;
+        for (u32 render_target = 0; render_target < Maxwell3D::Regs::NumRenderTargets;
+             ++render_target) {
+            // TODO(Blinkhawk): verify the behavior of alpha testing on hardware when
+            // multiple render targets are used.
+            if (header.ps.IsColorComponentOutputEnabled(render_target, 0) ||
+                header.ps.IsColorComponentOutputEnabled(render_target, 1) ||
+                header.ps.IsColorComponentOutputEnabled(render_target, 2) ||
+                header.ps.IsColorComponentOutputEnabled(render_target, 3)) {
+                shader.AddLine(fmt::format("if (!AlphaFunc({})) discard;",
+                                           regs.GetRegisterAsFloat(current_reg)));
+                current_reg += 4;
+            }
+        }
+        --shader.scope;
+        shader.AddLine('}');
+
         // Write the color outputs using the data in the shader registers, disabled
         // rendertargets/components are skipped in the register assignment.
-        u32 current_reg = 0;
+        current_reg = 0;
         for (u32 render_target = 0; render_target < Maxwell3D::Regs::NumRenderTargets;
              ++render_target) {
             // TODO(Subv): Figure out how dual-source blending is configured in the Switch.
@@ -1061,6 +1386,63 @@ private:
         }
     }
 
+    /// Unpacks a video instruction operand (e.g. VMAD).
+    std::string GetVideoOperand(const std::string& op, bool is_chunk, bool is_signed,
+                                Tegra::Shader::VideoType type, u64 byte_height) {
+        const std::string value = [&]() {
+            if (!is_chunk) {
+                const auto offset = static_cast<u32>(byte_height * 8);
+                return "((" + op + " >> " + std::to_string(offset) + ") & 0xff)";
+            }
+            const std::string zero = "0";
+
+            switch (type) {
+            case Tegra::Shader::VideoType::Size16_Low:
+                return '(' + op + " & 0xffff)";
+            case Tegra::Shader::VideoType::Size16_High:
+                return '(' + op + " >> 16)";
+            case Tegra::Shader::VideoType::Size32:
+                // TODO(Rodrigo): From my hardware tests it becomes a bit "mad" when
+                // this type is used (1 * 1 + 0 == 0x5b800000). Until a better
+                // explanation is found: assert.
+                UNIMPLEMENTED();
+                return zero;
+            case Tegra::Shader::VideoType::Invalid:
+                UNREACHABLE_MSG("Invalid instruction encoding");
+                return zero;
+            default:
+                UNREACHABLE();
+                return zero;
+            }
+        }();
+
+        if (is_signed) {
+            return "int(" + value + ')';
+        }
+        return value;
+    };
+
+    /// Gets the A operand for a video instruction.
+    std::string GetVideoOperandA(Instruction instr) {
+        return GetVideoOperand(regs.GetRegisterAsInteger(instr.gpr8, 0, false),
+                               instr.video.is_byte_chunk_a != 0, instr.video.signed_a,
+                               instr.video.type_a, instr.video.byte_height_a);
+    }
+
+    /// Gets the B operand for a video instruction.
+    std::string GetVideoOperandB(Instruction instr) {
+        if (instr.video.use_register_b) {
+            return GetVideoOperand(regs.GetRegisterAsInteger(instr.gpr20, 0, false),
+                                   instr.video.is_byte_chunk_b != 0, instr.video.signed_b,
+                                   instr.video.type_b, instr.video.byte_height_b);
+        } else {
+            return '(' +
+                   std::to_string(instr.video.signed_b ? static_cast<s16>(instr.alu.GetImm20_16())
+                                                       : instr.alu.GetImm20_16()) +
+                   ')';
+        }
+    }
+
     /**
      * Compiles a single instruction from Tegra to GLSL.
      * @param offset the offset of the Tegra shader instruction.
@@ -1083,8 +1465,8 @@ private:
             return offset + 1;
         }
 
-        shader.AddLine("// " + std::to_string(offset) + ": " + opcode->GetName() + " (" +
-                       std::to_string(instr.value) + ')');
+        shader.AddLine(
+            fmt::format("// {}: {} (0x{:016x})", offset, opcode->get().GetName(), instr.value));
 
         using Tegra::Shader::Pred;
         ASSERT_MSG(instr.pred.full_pred != Pred::NeverExecute,
@@ -1092,7 +1474,7 @@ private:
 
         // Some instructions (like SSY) don't have a predicate field, they are always
         // unconditionally executed.
-        bool can_be_predicated = OpCode::IsPredicatedInstruction(opcode->GetId());
+        bool can_be_predicated = OpCode::IsPredicatedInstruction(opcode->get().GetId());
 
         if (can_be_predicated && instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex)) {
             shader.AddLine("if (" +
@@ -1102,7 +1484,7 @@ private:
             ++shader.scope;
         }
 
-        switch (opcode->GetType()) {
+        switch (opcode->get().GetType()) {
         case OpCode::Type::Arithmetic: {
             std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
 
@@ -1119,7 +1501,7 @@ private:
                 }
             }
 
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::MOV_C:
             case OpCode::Id::MOV_R: {
                 // MOV does not have neither 'abs' nor 'neg' bits.
@@ -1141,8 +1523,13 @@ private:
                 ASSERT_MSG(instr.fmul.cc == 0, "FMUL cc is not implemented");
 
                 op_b = GetOperandAbsNeg(op_b, false, instr.fmul.negate_b);
+
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b, 1, 1,
-                                        instr.alu.saturate_d);
+                                        instr.alu.saturate_d, 0, true);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FMUL Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::FADD_C:
@@ -1150,8 +1537,13 @@ private:
             case OpCode::Id::FADD_IMM: {
                 op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a);
                 op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b);
+
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1,
-                                        instr.alu.saturate_d);
+                                        instr.alu.saturate_d, 0, true);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FADD Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::MUFU: {
@@ -1159,31 +1551,31 @@ private:
                 switch (instr.sub_op) {
                 case SubOp::Cos:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "cos(" + op_a + ')', 1, 1,
-                                            instr.alu.saturate_d);
+                                            instr.alu.saturate_d, 0, true);
                     break;
                 case SubOp::Sin:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "sin(" + op_a + ')', 1, 1,
-                                            instr.alu.saturate_d);
+                                            instr.alu.saturate_d, 0, true);
                     break;
                 case SubOp::Ex2:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "exp2(" + op_a + ')', 1, 1,
-                                            instr.alu.saturate_d);
+                                            instr.alu.saturate_d, 0, true);
                     break;
                 case SubOp::Lg2:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "log2(" + op_a + ')', 1, 1,
-                                            instr.alu.saturate_d);
+                                            instr.alu.saturate_d, 0, true);
                     break;
                 case SubOp::Rcp:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "1.0 / " + op_a, 1, 1,
-                                            instr.alu.saturate_d);
+                                            instr.alu.saturate_d, 0, true);
                     break;
                 case SubOp::Rsq:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "inversesqrt(" + op_a + ')', 1, 1,
-                                            instr.alu.saturate_d);
+                                            instr.alu.saturate_d, 0, true);
                     break;
                 case SubOp::Sqrt:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "sqrt(" + op_a + ')', 1, 1,
-                                            instr.alu.saturate_d);
+                                            instr.alu.saturate_d, 0, true);
                     break;
                 default:
                     LOG_CRITICAL(HW_GPU, "Unhandled MUFU sub op: {0:x}",
@@ -1204,7 +1596,11 @@ private:
                 regs.SetRegisterToFloat(instr.gpr0, 0,
                                         '(' + condition + ") ? min(" + parameters + ") : max(" +
                                             parameters + ')',
-                                        1, 1);
+                                        1, 1, false, 0, true);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FMNMX Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::RRO_C:
@@ -1217,22 +1613,28 @@ private:
                 break;
             }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled arithmetic instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled arithmetic instruction: {}",
+                             opcode->get().GetName());
                 UNREACHABLE();
             }
             }
             break;
         }
         case OpCode::Type::ArithmeticImmediate: {
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::MOV32_IMM: {
                 regs.SetRegisterToFloat(instr.gpr0, 0, GetImmediate32(instr), 1, 1);
                 break;
             }
             case OpCode::Id::FMUL32_IMM: {
-                regs.SetRegisterToFloat(
-                    instr.gpr0, 0,
-                    regs.GetRegisterAsFloat(instr.gpr8) + " * " + GetImmediate32(instr), 1, 1);
+                regs.SetRegisterToFloat(instr.gpr0, 0,
+                                        regs.GetRegisterAsFloat(instr.gpr8) + " * " +
+                                            GetImmediate32(instr),
+                                        1, 1, instr.fmul32.saturate, 0, true);
+                if (instr.op_32.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FMUL32 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::FADD32I: {
@@ -1255,7 +1657,11 @@ private:
                     op_b = "-(" + op_b + ')';
                 }
 
-                regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1);
+                regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1, false, 0, true);
+                if (instr.op_32.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FADD32 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             }
@@ -1267,7 +1673,7 @@ private:
             std::string op_a = instr.bfe.negate_a ? "-" : "";
             op_a += regs.GetRegisterAsInteger(instr.gpr8);
 
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::BFE_IMM: {
                 std::string inner_shift =
                     '(' + op_a + " << " + std::to_string(instr.bfe.GetLeftShiftValue()) + ')';
@@ -1276,17 +1682,20 @@ private:
                     std::to_string(instr.bfe.GetLeftShiftValue() + instr.bfe.shift_position) + ')';
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, outer_shift, 1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "BFE Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled BFE instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled BFE instruction: {}", opcode->get().GetName());
                 UNREACHABLE();
             }
             }
 
             break;
         }
-
         case OpCode::Type::Shift: {
             std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, true);
             std::string op_b;
@@ -1302,7 +1711,7 @@ private:
                 }
             }
 
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::SHR_C:
             case OpCode::Id::SHR_R:
             case OpCode::Id::SHR_IMM: {
@@ -1314,32 +1723,43 @@ private:
                 // Cast to int is superfluous for arithmetic shift, it's only for a logical shift
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, "int(" + op_a + " >> " + op_b + ')',
                                           1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "SHR Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::SHL_C:
             case OpCode::Id::SHL_R:
             case OpCode::Id::SHL_IMM:
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " << " + op_b, 1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "SHL Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled shift instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled shift instruction: {}", opcode->get().GetName());
                 UNREACHABLE();
             }
             }
             break;
         }
-
         case OpCode::Type::ArithmeticIntegerImmediate: {
             std::string op_a = regs.GetRegisterAsInteger(instr.gpr8);
             std::string op_b = std::to_string(instr.alu.imm20_32.Value());
 
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::IADD32I:
                 if (instr.iadd32i.negate_a)
                     op_a = "-(" + op_a + ')';
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1,
                                           instr.iadd32i.saturate != 0);
+                if (instr.op_32.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "IADD32 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             case OpCode::Id::LOP32I: {
                 if (instr.alu.lop32i.invert_a)
@@ -1351,11 +1771,15 @@ private:
                 WriteLogicOperation(instr.gpr0, instr.alu.lop32i.operation, op_a, op_b,
                                     Tegra::Shader::PredicateResultMode::None,
                                     Tegra::Shader::Pred::UnusedIndex);
+                if (instr.op_32.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "LOP32I Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             default: {
                 LOG_CRITICAL(HW_GPU, "Unhandled ArithmeticIntegerImmediate instruction: {}",
-                             opcode->GetName());
+                             opcode->get().GetName());
                 UNREACHABLE();
             }
             }
@@ -1375,7 +1799,7 @@ private:
                 }
             }
 
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::IADD_C:
             case OpCode::Id::IADD_R:
             case OpCode::Id::IADD_IMM: {
@@ -1387,6 +1811,10 @@ private:
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1,
                                           instr.alu.saturate_d);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "IADD Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::IADD3_C:
@@ -1411,7 +1839,7 @@ private:
                     }
                 };
 
-                if (opcode->GetId() == OpCode::Id::IADD3_R) {
+                if (opcode->get().GetId() == OpCode::Id::IADD3_R) {
                     apply_height(instr.iadd3.height_a, op_a);
                     apply_height(instr.iadd3.height_b, op_b);
                     apply_height(instr.iadd3.height_c, op_c);
@@ -1427,7 +1855,7 @@ private:
                     op_c = "-(" + op_c + ')';
 
                 std::string result;
-                if (opcode->GetId() == OpCode::Id::IADD3_R) {
+                if (opcode->get().GetId() == OpCode::Id::IADD3_R) {
                     switch (instr.iadd3.mode) {
                     case Tegra::Shader::IAdd3Mode::RightShift:
                         // TODO(tech4me): According to
@@ -1448,6 +1876,11 @@ private:
                 }
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, result, 1, 1);
+
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "IADD3 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::ISCADD_C:
@@ -1463,6 +1896,10 @@ private:
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0,
                                           "((" + op_a + " << " + shift + ") + " + op_b + ')', 1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "ISCADD Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::POPC_C:
@@ -1494,6 +1931,10 @@ private:
 
                 WriteLogicOperation(instr.gpr0, instr.alu.lop.operation, op_a, op_b,
                                     instr.alu.lop.pred_result_mode, instr.alu.lop.pred48);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "LOP Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::LOP3_C:
@@ -1502,13 +1943,17 @@ private:
                 const std::string op_c = regs.GetRegisterAsInteger(instr.gpr39);
                 std::string lut;
 
-                if (opcode->GetId() == OpCode::Id::LOP3_R) {
+                if (opcode->get().GetId() == OpCode::Id::LOP3_R) {
                     lut = '(' + std::to_string(instr.alu.lop3.GetImmLut28()) + ')';
                 } else {
                     lut = '(' + std::to_string(instr.alu.lop3.GetImmLut48()) + ')';
                 }
 
                 WriteLop3Instruction(instr.gpr0, op_a, op_b, op_c, lut);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "LOP3 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::IMNMX_C:
@@ -1523,6 +1968,10 @@ private:
                                           '(' + condition + ") ? min(" + parameters + ") : max(" +
                                               parameters + ')',
                                           1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "IMNMX Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::LEA_R2:
@@ -1532,7 +1981,7 @@ private:
             case OpCode::Id::LEA_HI: {
                 std::string op_c;
 
-                switch (opcode->GetId()) {
+                switch (opcode->get().GetId()) {
                 case OpCode::Id::LEA_R2: {
                     op_a = regs.GetRegisterAsInteger(instr.gpr20);
                     op_b = regs.GetRegisterAsInteger(instr.gpr39);
@@ -1577,7 +2026,8 @@ private:
                     op_b = regs.GetRegisterAsInteger(instr.gpr8);
                     op_a = std::to_string(instr.lea.imm.entry_a);
                     op_c = std::to_string(instr.lea.imm.entry_b);
-                    LOG_CRITICAL(HW_GPU, "Unhandled LEA subinstruction: {}", opcode->GetName());
+                    LOG_CRITICAL(HW_GPU, "Unhandled LEA subinstruction: {}",
+                                 opcode->get().GetName());
                     UNREACHABLE();
                 }
                 }
@@ -1592,13 +2042,95 @@ private:
             }
             default: {
                 LOG_CRITICAL(HW_GPU, "Unhandled ArithmeticInteger instruction: {}",
-                             opcode->GetName());
+                             opcode->get().GetName());
                 UNREACHABLE();
             }
             }
 
             break;
         }
+        case OpCode::Type::ArithmeticHalf: {
+            if (opcode->get().GetId() == OpCode::Id::HADD2_C ||
+                opcode->get().GetId() == OpCode::Id::HADD2_R) {
+                ASSERT_MSG(instr.alu_half.ftz == 0, "Unimplemented");
+            }
+            const bool negate_a =
+                opcode->get().GetId() != OpCode::Id::HMUL2_R && instr.alu_half.negate_a != 0;
+            const bool negate_b =
+                opcode->get().GetId() != OpCode::Id::HMUL2_C && instr.alu_half.negate_b != 0;
+
+            const std::string op_a =
+                GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr8, 0, false), instr.alu_half.type_a,
+                             instr.alu_half.abs_a != 0, negate_a);
+
+            std::string op_b;
+            switch (opcode->get().GetId()) {
+            case OpCode::Id::HADD2_C:
+            case OpCode::Id::HMUL2_C:
+                op_b = regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
+                                       GLSLRegister::Type::UnsignedInteger);
+                break;
+            case OpCode::Id::HADD2_R:
+            case OpCode::Id::HMUL2_R:
+                op_b = regs.GetRegisterAsInteger(instr.gpr20, 0, false);
+                break;
+            default:
+                UNREACHABLE();
+                op_b = "0";
+                break;
+            }
+            op_b = GetHalfFloat(op_b, instr.alu_half.type_b, instr.alu_half.abs_b != 0, negate_b);
+
+            const std::string result = [&]() {
+                switch (opcode->get().GetId()) {
+                case OpCode::Id::HADD2_C:
+                case OpCode::Id::HADD2_R:
+                    return '(' + op_a + " + " + op_b + ')';
+                case OpCode::Id::HMUL2_C:
+                case OpCode::Id::HMUL2_R:
+                    return '(' + op_a + " * " + op_b + ')';
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled half float instruction: {}",
+                                 opcode->get().GetName());
+                    UNREACHABLE();
+                    return std::string("0");
+                }
+            }();
+
+            regs.SetRegisterToHalfFloat(instr.gpr0, 0, result, instr.alu_half.merge, 1, 1,
+                                        instr.alu_half.saturate != 0);
+            break;
+        }
+        case OpCode::Type::ArithmeticHalfImmediate: {
+            if (opcode->get().GetId() == OpCode::Id::HADD2_IMM) {
+                ASSERT_MSG(instr.alu_half_imm.ftz == 0, "Unimplemented");
+            } else {
+                ASSERT_MSG(instr.alu_half_imm.precision == Tegra::Shader::HalfPrecision::None,
+                           "Unimplemented");
+            }
+
+            const std::string op_a = GetHalfFloat(
+                regs.GetRegisterAsInteger(instr.gpr8, 0, false), instr.alu_half_imm.type_a,
+                instr.alu_half_imm.abs_a != 0, instr.alu_half_imm.negate_a != 0);
+
+            const std::string op_b = UnpackHalfImmediate(instr, true);
+
+            const std::string result = [&]() {
+                switch (opcode->get().GetId()) {
+                case OpCode::Id::HADD2_IMM:
+                    return op_a + " + " + op_b;
+                case OpCode::Id::HMUL2_IMM:
+                    return op_a + " * " + op_b;
+                default:
+                    UNREACHABLE();
+                    return std::string("0");
+                }
+            }();
+
+            regs.SetRegisterToHalfFloat(instr.gpr0, 0, result, instr.alu_half_imm.merge, 1, 1,
+                                        instr.alu_half_imm.saturate != 0);
+            break;
+        }
         case OpCode::Type::Ffma: {
             const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
             std::string op_b = instr.ffma.negate_b ? "-" : "";
@@ -1610,7 +2142,7 @@ private:
             ASSERT_MSG(instr.ffma.tab5980_1 == 0, "FFMA tab5980_1({}) not implemented",
                        instr.ffma.tab5980_1.Value());
 
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::FFMA_CR: {
                 op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
                                         GLSLRegister::Type::Float);
@@ -1634,17 +2166,75 @@ private:
                 break;
             }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled FFMA instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled FFMA instruction: {}", opcode->get().GetName());
                 UNREACHABLE();
             }
             }
 
-            regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b + " + " + op_c, 1, 1,
-                                    instr.alu.saturate_d);
+            regs.SetRegisterToFloat(instr.gpr0, 0, "fma(" + op_a + ", " + op_b + ", " + op_c + ')',
+                                    1, 1, instr.alu.saturate_d, 0, true);
+            if (instr.generates_cc) {
+                LOG_CRITICAL(HW_GPU, "FFMA Generates an unhandled Control Code");
+                UNREACHABLE();
+            }
+
+            break;
+        }
+        case OpCode::Type::Hfma2: {
+            if (opcode->get().GetId() == OpCode::Id::HFMA2_RR) {
+                ASSERT_MSG(instr.hfma2.rr.precision == Tegra::Shader::HalfPrecision::None,
+                           "Unimplemented");
+            } else {
+                ASSERT_MSG(instr.hfma2.precision == Tegra::Shader::HalfPrecision::None,
+                           "Unimplemented");
+            }
+            const bool saturate = opcode->get().GetId() == OpCode::Id::HFMA2_RR
+                                      ? instr.hfma2.rr.saturate != 0
+                                      : instr.hfma2.saturate != 0;
+
+            const std::string op_a =
+                GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr8, 0, false), instr.hfma2.type_a);
+            std::string op_b, op_c;
+
+            switch (opcode->get().GetId()) {
+            case OpCode::Id::HFMA2_CR:
+                op_b = GetHalfFloat(regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
+                                                    GLSLRegister::Type::UnsignedInteger),
+                                    instr.hfma2.type_b, false, instr.hfma2.negate_b);
+                op_c = GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr39, 0, false),
+                                    instr.hfma2.type_reg39, false, instr.hfma2.negate_c);
+                break;
+            case OpCode::Id::HFMA2_RC:
+                op_b = GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr39, 0, false),
+                                    instr.hfma2.type_reg39, false, instr.hfma2.negate_b);
+                op_c = GetHalfFloat(regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
+                                                    GLSLRegister::Type::UnsignedInteger),
+                                    instr.hfma2.type_b, false, instr.hfma2.negate_c);
+                break;
+            case OpCode::Id::HFMA2_RR:
+                op_b = GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr20, 0, false),
+                                    instr.hfma2.type_b, false, instr.hfma2.negate_b);
+                op_c = GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr39, 0, false),
+                                    instr.hfma2.rr.type_c, false, instr.hfma2.rr.negate_c);
+                break;
+            case OpCode::Id::HFMA2_IMM_R:
+                op_b = UnpackHalfImmediate(instr, true);
+                op_c = GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr39, 0, false),
+                                    instr.hfma2.type_reg39, false, instr.hfma2.negate_c);
+                break;
+            default:
+                UNREACHABLE();
+                op_c = op_b = "vec2(0)";
+                break;
+            }
+
+            const std::string result = '(' + op_a + " * " + op_b + " + " + op_c + ')';
+
+            regs.SetRegisterToHalfFloat(instr.gpr0, 0, result, instr.hfma2.merge, 1, 1, saturate);
             break;
         }
         case OpCode::Type::Conversion: {
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::I2I_R: {
                 ASSERT_MSG(!instr.conversion.selector, "Unimplemented");
 
@@ -1692,6 +2282,11 @@ private:
                 }
 
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
+
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "I2F Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::F2F_R: {
@@ -1730,6 +2325,11 @@ private:
                 }
 
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1, instr.alu.saturate_d);
+
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "F2F Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::F2I_R:
@@ -1779,17 +2379,22 @@ private:
 
                 regs.SetRegisterToInteger(instr.gpr0, instr.conversion.is_output_signed, 0, op_a, 1,
                                           1, false, 0, instr.conversion.dest_size);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "F2I Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled conversion instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled conversion instruction: {}",
+                             opcode->get().GetName());
                 UNREACHABLE();
             }
             }
             break;
         }
         case OpCode::Type::Memory: {
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::LD_A: {
                 // Note: Shouldn't this be interp mode flat? As in no interpolation made.
                 ASSERT_MSG(instr.gpr8.Value() == Register::ZeroIndex,
@@ -1806,7 +2411,7 @@ private:
                 const auto LoadNextElement = [&](u32 reg_offset) {
                     regs.SetRegisterToInputAttibute(instr.gpr0.Value() + reg_offset, next_element,
                                                     static_cast<Attribute::Index>(next_index),
-                                                    input_mode);
+                                                    input_mode, instr.gpr39.Value());
 
                     // Load the next attribute element into the following register. If the element
                     // to load goes beyond the vec4 size, load the first element of the next
@@ -1859,6 +2464,39 @@ private:
                 shader.AddLine("}");
                 break;
             }
+            case OpCode::Id::LD_L: {
+                // Add an extra scope and declare the index register inside to prevent
+                // overwriting it in case it is used as an output of the LD instruction.
+                shader.AddLine('{');
+                ++shader.scope;
+
+                std::string op = '(' + regs.GetRegisterAsInteger(instr.gpr8, 0, false) + " + " +
+                                 std::to_string(instr.smem_imm.Value()) + ')';
+
+                shader.AddLine("uint index = (" + op + " / 4);");
+
+                const std::string op_a = regs.GetLocalMemoryAsFloat("index");
+
+                if (instr.ld_l.unknown != 1) {
+                    LOG_CRITICAL(HW_GPU, "LD_L Unhandled mode: {}",
+                                 static_cast<unsigned>(instr.ld_l.unknown.Value()));
+                    UNREACHABLE();
+                }
+
+                switch (instr.ldst_sl.type.Value()) {
+                case Tegra::Shader::StoreType::Bytes32:
+                    regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
+                    break;
+                default:
+                    LOG_CRITICAL(HW_GPU, "LD_L Unhandled type: {}",
+                                 static_cast<unsigned>(instr.ldst_sl.type.Value()));
+                    UNREACHABLE();
+                }
+
+                --shader.scope;
+                shader.AddLine('}');
+                break;
+            }
             case OpCode::Id::ST_A: {
                 ASSERT_MSG(instr.gpr8.Value() == Register::ZeroIndex,
                            "Indirect attribute loads are not supported");
@@ -1870,8 +2508,8 @@ private:
 
                 const auto StoreNextElement = [&](u32 reg_offset) {
                     regs.SetOutputAttributeToRegister(static_cast<Attribute::Index>(next_index),
-                                                      next_element,
-                                                      instr.gpr0.Value() + reg_offset);
+                                                      next_element, instr.gpr0.Value() + reg_offset,
+                                                      instr.gpr39.Value());
 
                     // Load the next attribute element into the following register. If the element
                     // to load goes beyond the vec4 size, load the first element of the next
@@ -1887,33 +2525,113 @@ private:
 
                 break;
             }
+            case OpCode::Id::ST_L: {
+                // Add an extra scope and declare the index register inside to prevent
+                // overwriting it in case it is used as an output of the LD instruction.
+                shader.AddLine('{');
+                ++shader.scope;
+
+                std::string op = '(' + regs.GetRegisterAsInteger(instr.gpr8, 0, false) + " + " +
+                                 std::to_string(instr.smem_imm.Value()) + ')';
+
+                shader.AddLine("uint index = (" + op + " / 4);");
+
+                if (instr.st_l.unknown != 0) {
+                    LOG_CRITICAL(HW_GPU, "ST_L Unhandled mode: {}",
+                                 static_cast<unsigned>(instr.st_l.unknown.Value()));
+                    UNREACHABLE();
+                }
+
+                switch (instr.ldst_sl.type.Value()) {
+                case Tegra::Shader::StoreType::Bytes32:
+                    regs.SetLocalMemoryAsFloat("index", regs.GetRegisterAsFloat(instr.gpr0));
+                    break;
+                default:
+                    LOG_CRITICAL(HW_GPU, "ST_L Unhandled type: {}",
+                                 static_cast<unsigned>(instr.ldst_sl.type.Value()));
+                    UNREACHABLE();
+                }
+
+                --shader.scope;
+                shader.AddLine('}');
+                break;
+            }
             case OpCode::Id::TEX: {
-                ASSERT_MSG(instr.tex.array == 0, "TEX arrays unimplemented");
                 Tegra::Shader::TextureType texture_type{instr.tex.texture_type};
                 std::string coord;
+                const bool is_array = instr.tex.array != 0;
 
                 ASSERT_MSG(!instr.tex.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP),
                            "NODEP is not implemented");
                 ASSERT_MSG(!instr.tex.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI),
                            "AOFFI is not implemented");
-                ASSERT_MSG(!instr.tex.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC),
-                           "DC is not implemented");
 
-                switch (texture_type) {
-                case Tegra::Shader::TextureType::Texture1D: {
-                    const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    coord = "float coords = " + x + ';';
+                const bool depth_compare =
+                    instr.tex.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC);
+                u32 num_coordinates = TextureCoordinates(texture_type);
+                if (depth_compare)
+                    num_coordinates += 1;
+
+                switch (num_coordinates) {
+                case 1: {
+                    if (is_array) {
+                        const std::string index = regs.GetRegisterAsInteger(instr.gpr8);
+                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                        coord = "vec2 coords = vec2(" + x + ", " + index + ");";
+                    } else {
+                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                        coord = "float coords = " + x + ';';
+                    }
                     break;
                 }
-                case Tegra::Shader::TextureType::Texture2D: {
-                    const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
-                    coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                case 2: {
+                    if (is_array) {
+                        const std::string index = regs.GetRegisterAsInteger(instr.gpr8);
+                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                        const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
+                        coord = "vec3 coords = vec3(" + x + ", " + y + ", " + index + ");";
+                    } else {
+                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                        const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                        coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                    }
+                    break;
+                }
+                case 3: {
+                    if (depth_compare) {
+                        if (is_array) {
+                            const std::string index = regs.GetRegisterAsInteger(instr.gpr8);
+                            const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                            const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+                            const std::string z = regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
+                            coord = "vec4 coords = vec4(" + x + ", " + y + ", " + z + ", " + index +
+                                    ");";
+                        } else {
+                            const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                            const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                            const std::string z = regs.GetRegisterAsFloat(instr.gpr20);
+                            coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                        }
+                    } else {
+                        if (is_array) {
+                            const std::string index = regs.GetRegisterAsInteger(instr.gpr8);
+                            const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                            const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
+                            const std::string z = regs.GetRegisterAsFloat(instr.gpr8.Value() + 3);
+                            coord = "vec4 coords = vec4(" + x + ", " + y + ", " + z + ", " + index +
+                                    ");";
+                        } else {
+                            const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                            const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                            const std::string z = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
+                            coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                        }
+                    }
                     break;
                 }
                 default:
-                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
-                                 static_cast<u32>(texture_type));
+                    LOG_CRITICAL(HW_GPU, "Unhandled coordinates number {}",
+                                 static_cast<u32>(num_coordinates));
                     UNREACHABLE();
 
                     // Fallback to interpreting as a 2D texture for now
@@ -1924,9 +2642,10 @@ private:
                 }
                 // TODO: make sure coordinates are always indexed to gpr8 and gpr20 is always bias
                 // or lod.
-                const std::string op_c = regs.GetRegisterAsFloat(instr.gpr20);
+                std::string op_c;
 
-                const std::string sampler = GetSampler(instr.sampler, texture_type, false);
+                const std::string sampler =
+                    GetSampler(instr.sampler, texture_type, is_array, depth_compare);
                 // Add an extra scope and declare the texture coords inside to prevent
                 // overwriting them in case they are used as outputs of the texs instruction.
 
@@ -1935,7 +2654,7 @@ private:
                 shader.AddLine(coord);
                 std::string texture;
 
-                switch (instr.tex.process_mode) {
+                switch (instr.tex.GetTextureProcessMode()) {
                 case Tegra::Shader::TextureProcessMode::None: {
                     texture = "texture(" + sampler + ", coords)";
                     break;
@@ -1946,12 +2665,25 @@ private:
                 }
                 case Tegra::Shader::TextureProcessMode::LB:
                 case Tegra::Shader::TextureProcessMode::LBA: {
+                    if (depth_compare) {
+                        if (is_array)
+                            op_c = regs.GetRegisterAsFloat(instr.gpr20.Value() + 2);
+                        else
+                            op_c = regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
+                    } else {
+                        op_c = regs.GetRegisterAsFloat(instr.gpr20);
+                    }
                     // TODO: Figure if A suffix changes the equation at all.
                     texture = "texture(" + sampler + ", coords, " + op_c + ')';
                     break;
                 }
                 case Tegra::Shader::TextureProcessMode::LL:
                 case Tegra::Shader::TextureProcessMode::LLA: {
+                    if (num_coordinates <= 2) {
+                        op_c = regs.GetRegisterAsFloat(instr.gpr20);
+                    } else {
+                        op_c = regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
+                    }
                     // TODO: Figure if A suffix changes the equation at all.
                     texture = "textureLod(" + sampler + ", coords, " + op_c + ')';
                     break;
@@ -1959,18 +2691,22 @@ private:
                 default: {
                     texture = "texture(" + sampler + ", coords)";
                     LOG_CRITICAL(HW_GPU, "Unhandled texture process mode {}",
-                                 static_cast<u32>(instr.tex.process_mode.Value()));
+                                 static_cast<u32>(instr.tex.GetTextureProcessMode()));
                     UNREACHABLE();
                 }
                 }
-                std::size_t dest_elem{};
-                for (std::size_t elem = 0; elem < 4; ++elem) {
-                    if (!instr.tex.IsComponentEnabled(elem)) {
-                        // Skip disabled components
-                        continue;
+                if (!depth_compare) {
+                    std::size_t dest_elem{};
+                    for (std::size_t elem = 0; elem < 4; ++elem) {
+                        if (!instr.tex.IsComponentEnabled(elem)) {
+                            // Skip disabled components
+                            continue;
+                        }
+                        regs.SetRegisterToFloat(instr.gpr0, elem, texture, 1, 4, false, dest_elem);
+                        ++dest_elem;
                     }
-                    regs.SetRegisterToFloat(instr.gpr0, elem, texture, 1, 4, false, dest_elem);
-                    ++dest_elem;
+                } else {
+                    regs.SetRegisterToFloat(instr.gpr0, 0, texture, 1, 1, false);
                 }
                 --shader.scope;
                 shader.AddLine("}");
@@ -1983,11 +2719,15 @@ private:
 
                 ASSERT_MSG(!instr.texs.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP),
                            "NODEP is not implemented");
-                ASSERT_MSG(!instr.texs.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC),
-                           "DC is not implemented");
 
-                switch (texture_type) {
-                case Tegra::Shader::TextureType::Texture2D: {
+                const bool depth_compare =
+                    instr.texs.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC);
+                u32 num_coordinates = TextureCoordinates(texture_type);
+                if (depth_compare)
+                    num_coordinates += 1;
+
+                switch (num_coordinates) {
+                case 2: {
                     if (is_array) {
                         const std::string index = regs.GetRegisterAsInteger(instr.gpr8);
                         const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
@@ -2000,17 +2740,25 @@ private:
                     }
                     break;
                 }
-                case Tegra::Shader::TextureType::TextureCube: {
-                    ASSERT_MSG(!is_array, "Unimplemented");
-                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
-                    std::string z = regs.GetRegisterAsFloat(instr.gpr20);
-                    coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                case 3: {
+                    if (is_array) {
+                        UNIMPLEMENTED_MSG("3-coordinate arrays not fully implemented");
+                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                        const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+                        coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                        texture_type = Tegra::Shader::TextureType::Texture2D;
+                        is_array = false;
+                    } else {
+                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                        const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                        const std::string z = regs.GetRegisterAsFloat(instr.gpr20);
+                        coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                    }
                     break;
                 }
                 default:
-                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
-                                 static_cast<u32>(texture_type));
+                    LOG_CRITICAL(HW_GPU, "Unhandled coordinates number {}",
+                                 static_cast<u32>(num_coordinates));
                     UNREACHABLE();
 
                     // Fallback to interpreting as a 2D texture for now
@@ -2020,9 +2768,35 @@ private:
                     texture_type = Tegra::Shader::TextureType::Texture2D;
                     is_array = false;
                 }
-                const std::string sampler = GetSampler(instr.sampler, texture_type, is_array);
-                const std::string texture = "texture(" + sampler + ", coords)";
-                WriteTexsInstruction(instr, coord, texture);
+                const std::string sampler =
+                    GetSampler(instr.sampler, texture_type, is_array, depth_compare);
+                std::string texture;
+                switch (instr.texs.GetTextureProcessMode()) {
+                case Tegra::Shader::TextureProcessMode::None: {
+                    texture = "texture(" + sampler + ", coords)";
+                    break;
+                }
+                case Tegra::Shader::TextureProcessMode::LZ: {
+                    texture = "textureLod(" + sampler + ", coords, 0.0)";
+                    break;
+                }
+                case Tegra::Shader::TextureProcessMode::LL: {
+                    const std::string op_c = regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
+                    texture = "textureLod(" + sampler + ", coords, " + op_c + ')';
+                    break;
+                }
+                default: {
+                    texture = "texture(" + sampler + ", coords)";
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture process mode {}",
+                                 static_cast<u32>(instr.texs.GetTextureProcessMode()));
+                    UNREACHABLE();
+                }
+                }
+                if (!depth_compare) {
+                    WriteTexsInstruction(instr, coord, texture);
+                } else {
+                    WriteTexsInstruction(instr, coord, "vec4(" + texture + ')');
+                }
                 break;
             }
             case OpCode::Id::TLDS: {
@@ -2040,6 +2814,8 @@ private:
                 ASSERT_MSG(!instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::MZ),
                            "MZ is not implemented");
 
+                u32 op_c_offset = 0;
+
                 switch (texture_type) {
                 case Tegra::Shader::TextureType::Texture1D: {
                     const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
@@ -2054,6 +2830,7 @@ private:
                         const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
                         const std::string y = regs.GetRegisterAsInteger(instr.gpr20);
                         coord = "ivec2 coords = ivec2(" + x + ", " + y + ");";
+                        op_c_offset = 1;
                     }
                     break;
                 }
@@ -2062,9 +2839,27 @@ private:
                                  static_cast<u32>(texture_type));
                     UNREACHABLE();
                 }
-
-                const std::string sampler = GetSampler(instr.sampler, texture_type, is_array);
-                const std::string texture = "texelFetch(" + sampler + ", coords, 0)";
+                const std::string sampler =
+                    GetSampler(instr.sampler, texture_type, is_array, false);
+                std::string texture = "texelFetch(" + sampler + ", coords, 0)";
+                switch (instr.tlds.GetTextureProcessMode()) {
+                case Tegra::Shader::TextureProcessMode::LZ: {
+                    texture = "texelFetch(" + sampler + ", coords, 0)";
+                    break;
+                }
+                case Tegra::Shader::TextureProcessMode::LL: {
+                    const std::string op_c =
+                        regs.GetRegisterAsInteger(instr.gpr20.Value() + op_c_offset);
+                    texture = "texelFetch(" + sampler + ", coords, " + op_c + ')';
+                    break;
+                }
+                default: {
+                    texture = "texelFetch(" + sampler + ", coords, 0)";
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture process mode {}",
+                                 static_cast<u32>(instr.tlds.GetTextureProcessMode()));
+                    UNREACHABLE();
+                }
+                }
                 WriteTexsInstruction(instr, coord, texture);
                 break;
             }
@@ -2077,28 +2872,43 @@ private:
                            "NODEP is not implemented");
                 ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI),
                            "AOFFI is not implemented");
-                ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC),
-                           "DC is not implemented");
                 ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::NDV),
                            "NDV is not implemented");
                 ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::PTP),
                            "PTP is not implemented");
-
-                switch (instr.tld4.texture_type) {
-                case Tegra::Shader::TextureType::Texture2D: {
+                const bool depth_compare =
+                    instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC);
+                auto texture_type = instr.tld4.texture_type.Value();
+                u32 num_coordinates = TextureCoordinates(texture_type);
+                if (depth_compare)
+                    num_coordinates += 1;
+
+                switch (num_coordinates) {
+                case 2: {
                     const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
                     const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
                     coord = "vec2 coords = vec2(" + x + ", " + y + ");";
                     break;
                 }
+                case 3: {
+                    const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    const std::string z = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
+                    coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                    break;
+                }
                 default:
-                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
-                                 static_cast<u32>(instr.tld4.texture_type.Value()));
+                    LOG_CRITICAL(HW_GPU, "Unhandled coordinates number {}",
+                                 static_cast<u32>(num_coordinates));
                     UNREACHABLE();
+                    const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                    texture_type = Tegra::Shader::TextureType::Texture2D;
                 }
 
                 const std::string sampler =
-                    GetSampler(instr.sampler, instr.tld4.texture_type, false);
+                    GetSampler(instr.sampler, texture_type, false, depth_compare);
                 // Add an extra scope and declare the texture coords inside to prevent
                 // overwriting them in case they are used as outputs of the texs instruction.
                 shader.AddLine("{");
@@ -2106,15 +2916,18 @@ private:
                 shader.AddLine(coord);
                 const std::string texture = "textureGather(" + sampler + ", coords, " +
                                             std::to_string(instr.tld4.component) + ')';
-
-                std::size_t dest_elem{};
-                for (std::size_t elem = 0; elem < 4; ++elem) {
-                    if (!instr.tex.IsComponentEnabled(elem)) {
-                        // Skip disabled components
-                        continue;
+                if (!depth_compare) {
+                    std::size_t dest_elem{};
+                    for (std::size_t elem = 0; elem < 4; ++elem) {
+                        if (!instr.tex.IsComponentEnabled(elem)) {
+                            // Skip disabled components
+                            continue;
+                        }
+                        regs.SetRegisterToFloat(instr.gpr0, elem, texture, 1, 4, false, dest_elem);
+                        ++dest_elem;
                     }
-                    regs.SetRegisterToFloat(instr.gpr0, elem, texture, 1, 4, false, dest_elem);
-                    ++dest_elem;
+                } else {
+                    regs.SetRegisterToFloat(instr.gpr0, 0, texture, 1, 1, false);
                 }
                 --shader.scope;
                 shader.AddLine("}");
@@ -2125,18 +2938,30 @@ private:
                            "NODEP is not implemented");
                 ASSERT_MSG(!instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI),
                            "AOFFI is not implemented");
-                ASSERT_MSG(!instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC),
-                           "DC is not implemented");
 
+                const bool depth_compare =
+                    instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC);
                 const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
                 const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20);
                 // TODO(Subv): Figure out how the sampler type is encoded in the TLD4S instruction.
-                const std::string sampler =
-                    GetSampler(instr.sampler, Tegra::Shader::TextureType::Texture2D, false);
-                const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                const std::string sampler = GetSampler(
+                    instr.sampler, Tegra::Shader::TextureType::Texture2D, false, depth_compare);
+                std::string coord;
+                if (!depth_compare) {
+                    coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                } else {
+                    // Note: TLD4S coordinate encoding works just like TEXS's
+                    const std::string op_c = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    coord = "vec3 coords = vec3(" + op_a + ", " + op_c + ", " + op_b + ");";
+                }
                 const std::string texture = "textureGather(" + sampler + ", coords, " +
                                             std::to_string(instr.tld4s.component) + ')';
-                WriteTexsInstruction(instr, coord, texture);
+
+                if (!depth_compare) {
+                    WriteTexsInstruction(instr, coord, texture);
+                } else {
+                    WriteTexsInstruction(instr, coord, "vec4(" + texture + ')');
+                }
                 break;
             }
             case OpCode::Id::TXQ: {
@@ -2147,7 +2972,7 @@ private:
                 // Sadly, not all texture instructions specify the type of texture their sampler
                 // uses. This must be fixed at a later instance.
                 const std::string sampler =
-                    GetSampler(instr.sampler, Tegra::Shader::TextureType::Texture2D, false);
+                    GetSampler(instr.sampler, Tegra::Shader::TextureType::Texture2D, false, false);
                 switch (instr.txq.query_type) {
                 case Tegra::Shader::TextureQueryType::Dimension: {
                     const std::string texture = "textureQueryLevels(" + sampler + ')';
@@ -2168,24 +2993,22 @@ private:
                 ASSERT_MSG(!instr.tmml.UsesMiscMode(Tegra::Shader::TextureMiscMode::NDV),
                            "NDV is not implemented");
 
-                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
                 const bool is_array = instr.tmml.array != 0;
                 auto texture_type = instr.tmml.texture_type.Value();
-                const std::string sampler = GetSampler(instr.sampler, texture_type, is_array);
+                const std::string sampler =
+                    GetSampler(instr.sampler, texture_type, is_array, false);
 
                 // TODO: add coordinates for different samplers once other texture types are
                 // implemented.
                 std::string coord;
                 switch (texture_type) {
                 case Tegra::Shader::TextureType::Texture1D: {
-                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
                     coord = "float coords = " + x + ';';
                     break;
                 }
                 case Tegra::Shader::TextureType::Texture2D: {
-                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
                     coord = "vec2 coords = vec2(" + x + ", " + y + ");";
                     break;
                 }
@@ -2195,8 +3018,7 @@ private:
                     UNREACHABLE();
 
                     // Fallback to interpreting as a 2D texture for now
-                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
                     coord = "vec2 coords = vec2(" + x + ", " + y + ");";
                     texture_type = Tegra::Shader::TextureType::Texture2D;
                 }
@@ -2216,27 +3038,20 @@ private:
                 break;
             }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->get().GetName());
                 UNREACHABLE();
             }
             }
             break;
         }
         case OpCode::Type::FloatSetPredicate: {
-            std::string op_a = instr.fsetp.neg_a ? "-" : "";
-            op_a += regs.GetRegisterAsFloat(instr.gpr8);
-
-            if (instr.fsetp.abs_a) {
-                op_a = "abs(" + op_a + ')';
-            }
+            const std::string op_a =
+                GetOperandAbsNeg(regs.GetRegisterAsFloat(instr.gpr8), instr.fsetp.abs_a != 0,
+                                 instr.fsetp.neg_a != 0);
 
-            std::string op_b{};
+            std::string op_b;
 
             if (instr.is_b_imm) {
-                if (instr.fsetp.neg_b) {
-                    // Only the immediate version of fsetp has a neg_b bit.
-                    op_b += '-';
-                }
                 op_b += '(' + GetImmediate19(instr) + ')';
             } else {
                 if (instr.is_b_gpr) {
@@ -2309,6 +3124,51 @@ private:
             }
             break;
         }
+        case OpCode::Type::HalfSetPredicate: {
+            ASSERT_MSG(instr.hsetp2.ftz == 0, "Unimplemented");
+
+            const std::string op_a =
+                GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr8, 0, false), instr.hsetp2.type_a,
+                             instr.hsetp2.abs_a, instr.hsetp2.negate_a);
+
+            const std::string op_b = [&]() {
+                switch (opcode->get().GetId()) {
+                case OpCode::Id::HSETP2_R:
+                    return GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr20, 0, false),
+                                        instr.hsetp2.type_b, instr.hsetp2.abs_a,
+                                        instr.hsetp2.negate_b);
+                default:
+                    UNREACHABLE();
+                    return std::string("vec2(0)");
+                }
+            }();
+
+            // We can't use the constant predicate as destination.
+            ASSERT(instr.hsetp2.pred3 != static_cast<u64>(Pred::UnusedIndex));
+
+            const std::string second_pred =
+                GetPredicateCondition(instr.hsetp2.pred39, instr.hsetp2.neg_pred != 0);
+
+            const std::string combiner = GetPredicateCombiner(instr.hsetp2.op);
+
+            const std::string component_combiner = instr.hsetp2.h_and ? "&&" : "||";
+            const std::string predicate =
+                '(' + GetPredicateComparison(instr.hsetp2.cond, op_a + ".x", op_b + ".x") + ' ' +
+                component_combiner + ' ' +
+                GetPredicateComparison(instr.hsetp2.cond, op_a + ".y", op_b + ".y") + ')';
+
+            // Set the primary predicate to the result of Predicate OP SecondPredicate
+            SetPredicate(instr.hsetp2.pred3,
+                         '(' + predicate + ") " + combiner + " (" + second_pred + ')');
+
+            if (instr.hsetp2.pred0 != static_cast<u64>(Pred::UnusedIndex)) {
+                // Set the secondary predicate to the result of !Predicate OP SecondPredicate,
+                // if enabled
+                SetPredicate(instr.hsetp2.pred0,
+                             "!(" + predicate + ") " + combiner + " (" + second_pred + ')');
+            }
+            break;
+        }
         case OpCode::Type::PredicateSetRegister: {
             const std::string op_a =
                 GetPredicateCondition(instr.pset.pred12, instr.pset.neg_pred12 != 0);
@@ -2331,10 +3191,15 @@ private:
                 regs.SetRegisterToFloat(instr.gpr0, 0, value, 1, 1);
             }
 
+            if (instr.generates_cc) {
+                LOG_CRITICAL(HW_GPU, "PSET Generates an unhandled Control Code");
+                UNREACHABLE();
+            }
+
             break;
         }
         case OpCode::Type::PredicateSetPredicate: {
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::PSETP: {
                 const std::string op_a =
                     GetPredicateCondition(instr.psetp.pred12, instr.psetp.neg_pred12 != 0);
@@ -2368,52 +3233,44 @@ private:
                 const std::string pred =
                     GetPredicateCondition(instr.csetp.pred39, instr.csetp.neg_pred39 != 0);
                 const std::string combiner = GetPredicateCombiner(instr.csetp.op);
-                const std::string controlCode = regs.GetControlCode(instr.csetp.cc);
+                const std::string control_code = regs.GetControlCode(instr.csetp.cc);
                 if (instr.csetp.pred3 != static_cast<u64>(Pred::UnusedIndex)) {
                     SetPredicate(instr.csetp.pred3,
-                                 '(' + controlCode + ") " + combiner + " (" + pred + ')');
+                                 '(' + control_code + ") " + combiner + " (" + pred + ')');
                 }
                 if (instr.csetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) {
                     SetPredicate(instr.csetp.pred0,
-                                 "!(" + controlCode + ") " + combiner + " (" + pred + ')');
+                                 "!(" + control_code + ") " + combiner + " (" + pred + ')');
                 }
                 break;
             }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled predicate instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled predicate instruction: {}",
+                             opcode->get().GetName());
                 UNREACHABLE();
             }
             }
             break;
         }
         case OpCode::Type::FloatSet: {
-            std::string op_a = instr.fset.neg_a ? "-" : "";
-            op_a += regs.GetRegisterAsFloat(instr.gpr8);
-
-            if (instr.fset.abs_a) {
-                op_a = "abs(" + op_a + ')';
-            }
+            const std::string op_a = GetOperandAbsNeg(regs.GetRegisterAsFloat(instr.gpr8),
+                                                      instr.fset.abs_a != 0, instr.fset.neg_a != 0);
 
-            std::string op_b = instr.fset.neg_b ? "-" : "";
+            std::string op_b;
 
             if (instr.is_b_imm) {
                 const std::string imm = GetImmediate19(instr);
-                if (instr.fset.neg_imm)
-                    op_b += "(-" + imm + ')';
-                else
-                    op_b += imm;
+                op_b = imm;
             } else {
                 if (instr.is_b_gpr) {
-                    op_b += regs.GetRegisterAsFloat(instr.gpr20);
+                    op_b = regs.GetRegisterAsFloat(instr.gpr20);
                 } else {
-                    op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
-                                            GLSLRegister::Type::Float);
+                    op_b = regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
+                                           GLSLRegister::Type::Float);
                 }
             }
 
-            if (instr.fset.abs_b) {
-                op_b = "abs(" + op_b + ')';
-            }
+            op_b = GetOperandAbsNeg(op_b, instr.fset.abs_b != 0, instr.fset.neg_b != 0);
 
             // The fset instruction sets a register to 1.0 or -1 (depending on the bf bit) if the
             // condition is true, and to 0 otherwise.
@@ -2469,6 +3326,50 @@ private:
             }
             break;
         }
+        case OpCode::Type::HalfSet: {
+            ASSERT_MSG(instr.hset2.ftz == 0, "Unimplemented");
+
+            const std::string op_a =
+                GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr8, 0, false), instr.hset2.type_a,
+                             instr.hset2.abs_a != 0, instr.hset2.negate_a != 0);
+
+            const std::string op_b = [&]() {
+                switch (opcode->get().GetId()) {
+                case OpCode::Id::HSET2_R:
+                    return GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr20, 0, false),
+                                        instr.hset2.type_b, instr.hset2.abs_b != 0,
+                                        instr.hset2.negate_b != 0);
+                default:
+                    UNREACHABLE();
+                    return std::string("vec2(0)");
+                }
+            }();
+
+            const std::string second_pred =
+                GetPredicateCondition(instr.hset2.pred39, instr.hset2.neg_pred != 0);
+
+            const std::string combiner = GetPredicateCombiner(instr.hset2.op);
+
+            // HSET2 operates on each half float in the pack.
+            std::string result;
+            for (int i = 0; i < 2; ++i) {
+                const std::string float_value = i == 0 ? "0x00003c00" : "0x3c000000";
+                const std::string integer_value = i == 0 ? "0x0000ffff" : "0xffff0000";
+                const std::string value = instr.hset2.bf == 1 ? float_value : integer_value;
+
+                const std::string comp = std::string(".") + "xy"[i];
+                const std::string predicate =
+                    "((" + GetPredicateComparison(instr.hset2.cond, op_a + comp, op_b + comp) +
+                    ") " + combiner + " (" + second_pred + "))";
+
+                result += '(' + predicate + " ? " + value + " : 0)";
+                if (i == 0) {
+                    result += " | ";
+                }
+            }
+            regs.SetRegisterToInteger(instr.gpr0, false, 0, '(' + result + ')', 1, 1);
+            break;
+        }
         case OpCode::Type::Xmad: {
             ASSERT_MSG(!instr.xmad.sign_a, "Unimplemented");
             ASSERT_MSG(!instr.xmad.sign_b, "Unimplemented");
@@ -2482,7 +3383,7 @@ private:
             const bool is_signed{instr.xmad.sign_a == 1};
 
             bool is_merge{};
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::XMAD_CR: {
                 is_merge = instr.xmad.merge_56;
                 op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
@@ -2511,7 +3412,7 @@ private:
                 break;
             }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled XMAD instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled XMAD instruction: {}", opcode->get().GetName());
                 UNREACHABLE();
             }
             }
@@ -2560,15 +3461,25 @@ private:
             }
 
             regs.SetRegisterToInteger(instr.gpr0, is_signed, 0, sum, 1, 1);
+            if (instr.generates_cc) {
+                LOG_CRITICAL(HW_GPU, "XMAD Generates an unhandled Control Code");
+                UNREACHABLE();
+            }
             break;
         }
         default: {
-            switch (opcode->GetId()) {
+            switch (opcode->get().GetId()) {
             case OpCode::Id::EXIT: {
                 if (stage == Maxwell3D::Regs::ShaderStage::Fragment) {
                     EmitFragmentOutputsWrite();
                 }
 
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "EXIT Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
+
                 switch (instr.flow.cond) {
                 case Tegra::Shader::FlowCondition::Always:
                     shader.AddLine("return true;");
@@ -2598,6 +3509,11 @@ private:
 
                 // Enclose "discard" in a conditional, so that GLSL compilation does not complain
                 // about unexecuted instructions that may follow this.
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "KIL Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
                 shader.AddLine("if (true) {");
                 ++shader.scope;
                 shader.AddLine("discard;");
@@ -2606,9 +3522,60 @@ private:
 
                 break;
             }
+            case OpCode::Id::OUT_R: {
+                ASSERT(instr.gpr20.Value() == Register::ZeroIndex);
+                ASSERT_MSG(stage == Maxwell3D::Regs::ShaderStage::Geometry,
+                           "OUT is expected to be used in a geometry shader.");
+
+                if (instr.out.emit) {
+                    // gpr0 is used to store the next address. Hardware returns a pointer but
+                    // we just return the next index with a cyclic cap.
+                    const std::string current{regs.GetRegisterAsInteger(instr.gpr8, 0, false)};
+                    const std::string next = "((" + current + " + 1" + ") % " +
+                                             std::to_string(MAX_GEOMETRY_BUFFERS) + ')';
+                    shader.AddLine("emit_vertex(" + current + ");");
+                    regs.SetRegisterToInteger(instr.gpr0, false, 0, next, 1, 1);
+                }
+                if (instr.out.cut) {
+                    shader.AddLine("EndPrimitive();");
+                }
+
+                break;
+            }
+            case OpCode::Id::MOV_SYS: {
+                switch (instr.sys20) {
+                case Tegra::Shader::SystemVariable::InvocationInfo: {
+                    LOG_WARNING(HW_GPU, "MOV_SYS instruction with InvocationInfo is incomplete");
+                    regs.SetRegisterToInteger(instr.gpr0, false, 0, "0u", 1, 1);
+                    break;
+                }
+                default: {
+                    LOG_CRITICAL(HW_GPU, "Unhandled system move: {}",
+                                 static_cast<u32>(instr.sys20.Value()));
+                    UNREACHABLE();
+                }
+                }
+                break;
+            }
+            case OpCode::Id::ISBERD: {
+                ASSERT(instr.isberd.o == 0);
+                ASSERT(instr.isberd.skew == 0);
+                ASSERT(instr.isberd.shift == Tegra::Shader::IsberdShift::None);
+                ASSERT(instr.isberd.mode == Tegra::Shader::IsberdMode::None);
+                ASSERT_MSG(stage == Maxwell3D::Regs::ShaderStage::Geometry,
+                           "ISBERD is expected to be used in a geometry shader.");
+                LOG_WARNING(HW_GPU, "ISBERD instruction is incomplete");
+                regs.SetRegisterToFloat(instr.gpr0, 0, regs.GetRegisterAsFloat(instr.gpr8), 1, 1);
+                break;
+            }
             case OpCode::Id::BRA: {
                 ASSERT_MSG(instr.bra.constant_buffer == 0,
                            "BRA with constant buffers are not implemented");
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "BRA Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
                 const u32 target = offset + instr.bra.GetBranchTarget();
                 shader.AddLine("{ jmp_to = " + std::to_string(target) + "u; break; }");
                 break;
@@ -2631,16 +3598,40 @@ private:
                 // The SSY opcode tells the GPU where to re-converge divergent execution paths, it
                 // sets the target of the jump that the SYNC instruction will make. The SSY opcode
                 // has a similar structure to the BRA opcode.
-                ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer SSY is not supported");
+                ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer flow is not supported");
+
+                const u32 target = offset + instr.bra.GetBranchTarget();
+                EmitPushToFlowStack(target);
+                break;
+            }
+            case OpCode::Id::PBK: {
+                // PBK pushes to a stack the address where BRK will jump to. This shares stack with
+                // SSY but using SYNC on a PBK address will kill the shader execution. We don't
+                // emulate this because it's very unlikely a driver will emit such invalid shader.
+                ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer PBK is not supported");
 
                 const u32 target = offset + instr.bra.GetBranchTarget();
-                EmitPushToSSYStack(target);
+                EmitPushToFlowStack(target);
                 break;
             }
             case OpCode::Id::SYNC: {
                 // The SYNC opcode jumps to the address previously set by the SSY opcode
-                ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
-                EmitPopFromSSYStack();
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "SYNC Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
+                EmitPopFromFlowStack();
+                break;
+            }
+            case OpCode::Id::BRK: {
+                // The BRK opcode jumps to the address previously set by the PBK opcode
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "BRK Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
+                EmitPopFromFlowStack();
                 break;
             }
             case OpCode::Id::DEPBAR: {
@@ -2649,8 +3640,59 @@ private:
                 LOG_WARNING(HW_GPU, "DEPBAR instruction is stubbed");
                 break;
             }
+            case OpCode::Id::VMAD: {
+                const bool result_signed = instr.video.signed_a == 1 || instr.video.signed_b == 1;
+                const std::string op_a = GetVideoOperandA(instr);
+                const std::string op_b = GetVideoOperandB(instr);
+                const std::string op_c = regs.GetRegisterAsInteger(instr.gpr39, 0, result_signed);
+
+                std::string result = '(' + op_a + " * " + op_b + " + " + op_c + ')';
+
+                switch (instr.vmad.shr) {
+                case Tegra::Shader::VmadShr::Shr7:
+                    result = '(' + result + " >> 7)";
+                    break;
+                case Tegra::Shader::VmadShr::Shr15:
+                    result = '(' + result + " >> 15)";
+                    break;
+                }
+
+                regs.SetRegisterToInteger(instr.gpr0, result_signed, 1, result, 1, 1,
+                                          instr.vmad.saturate == 1, 0, Register::Size::Word,
+                                          instr.vmad.cc);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "VMAD Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
+
+                break;
+            }
+            case OpCode::Id::VSETP: {
+                const std::string op_a = GetVideoOperandA(instr);
+                const std::string op_b = GetVideoOperandB(instr);
+
+                // We can't use the constant predicate as destination.
+                ASSERT(instr.vsetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
+
+                const std::string second_pred = GetPredicateCondition(instr.vsetp.pred39, false);
+
+                const std::string combiner = GetPredicateCombiner(instr.vsetp.op);
+
+                const std::string predicate = GetPredicateComparison(instr.vsetp.cond, op_a, op_b);
+                // Set the primary predicate to the result of Predicate OP SecondPredicate
+                SetPredicate(instr.vsetp.pred3,
+                             '(' + predicate + ") " + combiner + " (" + second_pred + ')');
+
+                if (instr.vsetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) {
+                    // Set the secondary predicate to the result of !Predicate OP SecondPredicate,
+                    // if enabled
+                    SetPredicate(instr.vsetp.pred0,
+                                 "!(" + predicate + ") " + combiner + " (" + second_pred + ')');
+                }
+                break;
+            }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled instruction: {}", opcode->GetName());
+                LOG_CRITICAL(HW_GPU, "Unhandled instruction: {}", opcode->get().GetName());
                 UNREACHABLE();
             }
             }
@@ -2712,11 +3754,11 @@ private:
                 labels.insert(subroutine.begin);
                 shader.AddLine("uint jmp_to = " + std::to_string(subroutine.begin) + "u;");
 
-                // TODO(Subv): Figure out the actual depth of the SSY stack, for now it seems
-                // unlikely that shaders will use 20 nested SSYs.
-                constexpr u32 SSY_STACK_SIZE = 20;
-                shader.AddLine("uint ssy_stack[" + std::to_string(SSY_STACK_SIZE) + "];");
-                shader.AddLine("uint ssy_stack_top = 0u;");
+                // TODO(Subv): Figure out the actual depth of the flow stack, for now it seems
+                // unlikely that shaders will use 20 nested SSYs and PBKs.
+                constexpr u32 FLOW_STACK_SIZE = 20;
+                shader.AddLine("uint flow_stack[" + std::to_string(FLOW_STACK_SIZE) + "];");
+                shader.AddLine("uint flow_stack_top = 0u;");
 
                 shader.AddLine("while (true) {");
                 ++shader.scope;
@@ -2776,23 +3818,24 @@ private:
     const u32 main_offset;
     Maxwell3D::Regs::ShaderStage stage;
     const std::string& suffix;
+    u64 local_memory_size;
 
     ShaderWriter shader;
     ShaderWriter declarations;
-    GLSLRegisterManager regs{shader, declarations, stage, suffix};
+    GLSLRegisterManager regs{shader, declarations, stage, suffix, header};
 
     // Declarations
     std::set<std::string> declr_predicates;
-}; // namespace Decompiler
+}; // namespace OpenGL::GLShader::Decompiler
 
 std::string GetCommonDeclarations() {
     return fmt::format("#define MAX_CONSTBUFFER_ELEMENTS {}\n",
                        RasterizerOpenGL::MaxConstbufferSize / sizeof(GLvec4));
 }
 
-boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,
-                                                Maxwell3D::Regs::ShaderStage stage,
-                                                const std::string& suffix) {
+std::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,
+                                              Maxwell3D::Regs::ShaderStage stage,
+                                              const std::string& suffix) {
     try {
         const auto subroutines =
             ControlFlowAnalyzer(program_code, main_offset, suffix).GetSubroutines();
@@ -2801,7 +3844,7 @@ boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code,
     } catch (const DecompileFail& exception) {
         LOG_ERROR(HW_GPU, "Shader decompilation failed: {}", exception.what());
     }
-    return boost::none;
+    return {};
 }
 
 } // namespace OpenGL::GLShader::Decompiler
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.h b/src/video_core/renderer_opengl/gl_shader_decompiler.h
index b20cc4bfa..d01a4a7ee 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.h
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.h
@@ -6,8 +6,8 @@
 
 #include <array>
 #include <functional>
+#include <optional>
 #include <string>
-#include <boost/optional.hpp>
 #include "common/common_types.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_opengl/gl_shader_gen.h"
@@ -18,8 +18,8 @@ using Tegra::Engines::Maxwell3D;
 
 std::string GetCommonDeclarations();
 
-boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,
-                                                Maxwell3D::Regs::ShaderStage stage,
-                                                const std::string& suffix);
+std::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,
+                                              Maxwell3D::Regs::ShaderStage stage,
+                                              const std::string& suffix);
 
 } // namespace OpenGL::GLShader::Decompiler
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp
index b0466c18f..9d17edd63 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp
@@ -17,7 +17,18 @@ ProgramResult GenerateVertexShader(const ShaderSetup& setup) {
     std::string out = "#version 430 core\n";
     out += "#extension GL_ARB_separate_shader_objects : enable\n\n";
     out += Decompiler::GetCommonDeclarations();
-    out += "bool exec_vertex();\n";
+
+    out += R"(
+
+layout (location = 0) out vec4 position;
+
+layout(std140) uniform vs_config {
+    vec4 viewport_flip;
+    uvec4 instance_id;
+    uvec4 flip_stage;
+    uvec4 alpha_test;
+};
+)";
 
     if (setup.IsDualProgram()) {
         out += "bool exec_vertex_b();\n";
@@ -26,20 +37,19 @@ ProgramResult GenerateVertexShader(const ShaderSetup& setup) {
     ProgramResult program =
         Decompiler::DecompileProgram(setup.program.code, PROGRAM_OFFSET,
                                      Maxwell3D::Regs::ShaderStage::Vertex, "vertex")
-            .get_value_or({});
+            .value_or(ProgramResult());
 
-    out += R"(
-
-out gl_PerVertex {
-    vec4 gl_Position;
-};
+    out += program.first;
 
-out vec4 position;
+    if (setup.IsDualProgram()) {
+        ProgramResult program_b =
+            Decompiler::DecompileProgram(setup.program.code_b, PROGRAM_OFFSET,
+                                         Maxwell3D::Regs::ShaderStage::Vertex, "vertex_b")
+                .value_or(ProgramResult());
+        out += program_b.first;
+    }
 
-layout (std140) uniform vs_config {
-    vec4 viewport_flip;
-    uvec4 instance_id;
-};
+    out += R"(
 
 void main() {
     position = vec4(0.0, 0.0, 0.0, 0.0);
@@ -52,27 +62,56 @@ void main() {
 
     out += R"(
 
-    // Viewport can be flipped, which is unsupported by glViewport
-    position.xy *= viewport_flip.xy;
+    // Check if the flip stage is VertexB
+    if (flip_stage[0] == 1) {
+        // Viewport can be flipped, which is unsupported by glViewport
+        position.xy *= viewport_flip.xy;
+    }
     gl_Position = position;
 
     // TODO(bunnei): This is likely a hack, position.w should be interpolated as 1.0
     // For now, this is here to bring order in lieu of proper emulation
-    position.w = 1.0;
+    if (flip_stage[0] == 1) {
+        position.w = 1.0;
+    }
 }
 
 )";
 
-    out += program.first;
+    return {out, program.second};
+}
 
-    if (setup.IsDualProgram()) {
-        ProgramResult program_b =
-            Decompiler::DecompileProgram(setup.program.code_b, PROGRAM_OFFSET,
-                                         Maxwell3D::Regs::ShaderStage::Vertex, "vertex_b")
-                .get_value_or({});
-        out += program_b.first;
-    }
+ProgramResult GenerateGeometryShader(const ShaderSetup& setup) {
+    std::string out = "#version 430 core\n";
+    out += "#extension GL_ARB_separate_shader_objects : enable\n\n";
+    out += Decompiler::GetCommonDeclarations();
+    out += "bool exec_geometry();\n";
+
+    ProgramResult program =
+        Decompiler::DecompileProgram(setup.program.code, PROGRAM_OFFSET,
+                                     Maxwell3D::Regs::ShaderStage::Geometry, "geometry")
+            .value_or(ProgramResult());
+    out += R"(
+out gl_PerVertex {
+    vec4 gl_Position;
+};
 
+layout (location = 0) in vec4 gs_position[];
+layout (location = 0) out vec4 position;
+
+layout (std140) uniform gs_config {
+    vec4 viewport_flip;
+    uvec4 instance_id;
+    uvec4 flip_stage;
+    uvec4 alpha_test;
+};
+
+void main() {
+    exec_geometry();
+}
+
+)";
+    out += program.first;
     return {out, program.second};
 }
 
@@ -85,9 +124,8 @@ ProgramResult GenerateFragmentShader(const ShaderSetup& setup) {
     ProgramResult program =
         Decompiler::DecompileProgram(setup.program.code, PROGRAM_OFFSET,
                                      Maxwell3D::Regs::ShaderStage::Fragment, "fragment")
-            .get_value_or({});
+            .value_or(ProgramResult());
     out += R"(
-in vec4 position;
 layout(location = 0) out vec4 FragColor0;
 layout(location = 1) out vec4 FragColor1;
 layout(location = 2) out vec4 FragColor2;
@@ -97,11 +135,39 @@ layout(location = 5) out vec4 FragColor5;
 layout(location = 6) out vec4 FragColor6;
 layout(location = 7) out vec4 FragColor7;
 
+layout (location = 0) in vec4 position;
+
 layout (std140) uniform fs_config {
     vec4 viewport_flip;
     uvec4 instance_id;
+    uvec4 flip_stage;
+    uvec4 alpha_test;
 };
 
+bool AlphaFunc(in float value) {
+    float ref = uintBitsToFloat(alpha_test[2]);
+    switch (alpha_test[1]) {
+        case 1:
+            return false;
+        case 2:
+            return value < ref;
+        case 3:
+            return value == ref;
+        case 4:
+            return value <= ref;
+        case 5:
+            return value > ref;
+        case 6:
+            return value != ref;
+        case 7:
+            return value >= ref;
+        case 8:
+            return true;
+        default:
+            return false;
+    }
+}
+
 void main() {
     exec_fragment();
 }
@@ -110,5 +176,4 @@ void main() {
     out += program.first;
     return {out, program.second};
 }
-
 } // namespace OpenGL::GLShader
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.h b/src/video_core/renderer_opengl/gl_shader_gen.h
index d53b93ad5..520b9d4e3 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.h
+++ b/src/video_core/renderer_opengl/gl_shader_gen.h
@@ -16,6 +16,8 @@ namespace OpenGL::GLShader {
 constexpr std::size_t MAX_PROGRAM_CODE_LENGTH{0x1000};
 using ProgramCode = std::vector<u64>;
 
+enum : u32 { POSITION_VARYING_LOCATION = 0, GENERIC_VARYING_START_LOCATION = 1 };
+
 class ConstBufferEntry {
     using Maxwell = Tegra::Engines::Maxwell3D::Regs;
 
@@ -75,8 +77,9 @@ class SamplerEntry {
 
 public:
     SamplerEntry(Maxwell::ShaderStage stage, std::size_t offset, std::size_t index,
-                 Tegra::Shader::TextureType type, bool is_array)
-        : offset(offset), stage(stage), sampler_index(index), type(type), is_array(is_array) {}
+                 Tegra::Shader::TextureType type, bool is_array, bool is_shadow)
+        : offset(offset), stage(stage), sampler_index(index), type(type), is_array(is_array),
+          is_shadow(is_shadow) {}
 
     std::size_t GetOffset() const {
         return offset;
@@ -117,6 +120,8 @@ public:
         }
         if (is_array)
             glsl_type += "Array";
+        if (is_shadow)
+            glsl_type += "Shadow";
         return glsl_type;
     }
 
@@ -128,6 +133,10 @@ public:
         return is_array;
     }
 
+    bool IsShadow() const {
+        return is_shadow;
+    }
+
     u32 GetHash() const {
         return (static_cast<u32>(stage) << 16) | static_cast<u32>(sampler_index);
     }
@@ -147,7 +156,8 @@ private:
     Maxwell::ShaderStage stage;      ///< Shader stage where this sampler was used.
     std::size_t sampler_index;       ///< Value used to index into the generated GLSL sampler array.
     Tegra::Shader::TextureType type; ///< The type used to sample this texture (Texture2D, etc)
-    bool is_array; ///< Whether the texture is being sampled as an array texture or not.
+    bool is_array;  ///< Whether the texture is being sampled as an array texture or not.
+    bool is_shadow; ///< Whether the texture is being sampled as a depth texture or not.
 };
 
 struct ShaderEntries {
@@ -188,6 +198,12 @@ private:
 ProgramResult GenerateVertexShader(const ShaderSetup& setup);
 
 /**
+ * Generates the GLSL geometry shader program source code for the given GS program
+ * @returns String of the shader source code
+ */
+ProgramResult GenerateGeometryShader(const ShaderSetup& setup);
+
+/**
  * Generates the GLSL fragment shader program source code for the given FS program
  * @returns String of the shader source code
  */
diff --git a/src/video_core/renderer_opengl/gl_shader_manager.cpp b/src/video_core/renderer_opengl/gl_shader_manager.cpp
index 022d32a86..8b8869ecb 100644
--- a/src/video_core/renderer_opengl/gl_shader_manager.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_manager.cpp
@@ -16,8 +16,27 @@ void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& sh
     viewport_flip[0] = regs.viewport_transform[0].scale_x < 0.0 ? -1.0f : 1.0f;
     viewport_flip[1] = regs.viewport_transform[0].scale_y < 0.0 ? -1.0f : 1.0f;
 
+    u32 func = static_cast<u32>(regs.alpha_test_func);
+    // Normalize the gl variants of opCompare to be the same as the normal variants
+    u32 op_gl_variant_base = static_cast<u32>(Tegra::Engines::Maxwell3D::Regs::ComparisonOp::Never);
+    if (func >= op_gl_variant_base) {
+        func = func - op_gl_variant_base + 1U;
+    }
+
+    alpha_test.enabled = regs.alpha_test_enabled;
+    alpha_test.func = func;
+    alpha_test.ref = regs.alpha_test_ref;
+
     // We only assign the instance to the first component of the vector, the rest is just padding.
     instance_id[0] = state.current_instance;
+
+    // Assign in which stage the position has to be flipped
+    // (the last stage before the fragment shader).
+    if (gpu.regs.shader_config[static_cast<u32>(Maxwell3D::Regs::ShaderProgram::Geometry)].enable) {
+        flip_stage[0] = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::Geometry);
+    } else {
+        flip_stage[0] = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::VertexB);
+    }
 }
 
 } // namespace OpenGL::GLShader
diff --git a/src/video_core/renderer_opengl/gl_shader_manager.h b/src/video_core/renderer_opengl/gl_shader_manager.h
index 3de15ba9b..36fe1f04c 100644
--- a/src/video_core/renderer_opengl/gl_shader_manager.h
+++ b/src/video_core/renderer_opengl/gl_shader_manager.h
@@ -21,8 +21,15 @@ struct MaxwellUniformData {
     void SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage);
     alignas(16) GLvec4 viewport_flip;
     alignas(16) GLuvec4 instance_id;
+    alignas(16) GLuvec4 flip_stage;
+    struct alignas(16) {
+        GLuint enabled;
+        GLuint func;
+        GLfloat ref;
+        GLuint padding;
+    } alpha_test;
 };
-static_assert(sizeof(MaxwellUniformData) == 32, "MaxwellUniformData structure size is incorrect");
+static_assert(sizeof(MaxwellUniformData) == 64, "MaxwellUniformData structure size is incorrect");
 static_assert(sizeof(MaxwellUniformData) < 16384,
               "MaxwellUniformData structure must be less than 16kb as per the OpenGL spec");
 
@@ -36,6 +43,10 @@ public:
         vs = program;
     }
 
+    void UseProgrammableGeometryShader(GLuint program) {
+        gs = program;
+    }
+
     void UseProgrammableFragmentShader(GLuint program) {
         fs = program;
     }
diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp
index 1fe26a2a9..b6b426f34 100644
--- a/src/video_core/renderer_opengl/gl_state.cpp
+++ b/src/video_core/renderer_opengl/gl_state.cpp
@@ -11,9 +11,10 @@
 namespace OpenGL {
 
 OpenGLState OpenGLState::cur_state;
-
+bool OpenGLState::s_rgb_used;
 OpenGLState::OpenGLState() {
     // These all match default OpenGL values
+    framebuffer_srgb.enabled = false;
     cull.enabled = false;
     cull.mode = GL_BACK;
     cull.front_face = GL_CCW;
@@ -21,6 +22,11 @@ OpenGLState::OpenGLState() {
     depth.test_enabled = false;
     depth.test_func = GL_LESS;
     depth.write_mask = GL_TRUE;
+    depth.depth_range_near = 0.0f;
+    depth.depth_range_far = 1.0f;
+
+    primitive_restart.enabled = false;
+    primitive_restart.index = 0;
 
     color_mask.red_enabled = GL_TRUE;
     color_mask.green_enabled = GL_TRUE;
@@ -83,79 +89,144 @@ OpenGLState::OpenGLState() {
     point.size = 1;
 }
 
-void OpenGLState::Apply() const {
+void OpenGLState::ApplyDefaultState() {
+    glDisable(GL_FRAMEBUFFER_SRGB);
+    glDisable(GL_CULL_FACE);
+    glDisable(GL_DEPTH_TEST);
+    glDisable(GL_PRIMITIVE_RESTART);
+    glDisable(GL_STENCIL_TEST);
+    glEnable(GL_BLEND);
+    glDisable(GL_COLOR_LOGIC_OP);
+    glDisable(GL_SCISSOR_TEST);
+}
+
+void OpenGLState::ApplySRgb() const {
+    // sRGB
+    if (framebuffer_srgb.enabled != cur_state.framebuffer_srgb.enabled) {
+        if (framebuffer_srgb.enabled) {
+            // Track if sRGB is used
+            s_rgb_used = true;
+            glEnable(GL_FRAMEBUFFER_SRGB);
+        } else {
+            glDisable(GL_FRAMEBUFFER_SRGB);
+        }
+    }
+}
+
+void OpenGLState::ApplyCulling() const {
     // Culling
-    if (cull.enabled != cur_state.cull.enabled) {
+    const bool cull_changed = cull.enabled != cur_state.cull.enabled;
+    if (cull_changed) {
         if (cull.enabled) {
             glEnable(GL_CULL_FACE);
         } else {
             glDisable(GL_CULL_FACE);
         }
     }
+    if (cull.enabled) {
+        if (cull_changed || cull.mode != cur_state.cull.mode) {
+            glCullFace(cull.mode);
+        }
 
-    if (cull.mode != cur_state.cull.mode) {
-        glCullFace(cull.mode);
-    }
-
-    if (cull.front_face != cur_state.cull.front_face) {
-        glFrontFace(cull.front_face);
+        if (cull_changed || cull.front_face != cur_state.cull.front_face) {
+            glFrontFace(cull.front_face);
+        }
     }
+}
 
+void OpenGLState::ApplyDepth() const {
     // Depth test
-    if (depth.test_enabled != cur_state.depth.test_enabled) {
+    const bool depth_test_changed = depth.test_enabled != cur_state.depth.test_enabled;
+    if (depth_test_changed) {
         if (depth.test_enabled) {
             glEnable(GL_DEPTH_TEST);
         } else {
             glDisable(GL_DEPTH_TEST);
         }
     }
-
-    if (depth.test_func != cur_state.depth.test_func) {
+    if (depth.test_enabled &&
+        (depth_test_changed || depth.test_func != cur_state.depth.test_func)) {
         glDepthFunc(depth.test_func);
     }
-
     // Depth mask
     if (depth.write_mask != cur_state.depth.write_mask) {
         glDepthMask(depth.write_mask);
     }
+    // Depth range
+    if (depth.depth_range_near != cur_state.depth.depth_range_near ||
+        depth.depth_range_far != cur_state.depth.depth_range_far) {
+        glDepthRange(depth.depth_range_near, depth.depth_range_far);
+    }
+}
 
-    // Color mask
-    if (color_mask.red_enabled != cur_state.color_mask.red_enabled ||
-        color_mask.green_enabled != cur_state.color_mask.green_enabled ||
-        color_mask.blue_enabled != cur_state.color_mask.blue_enabled ||
-        color_mask.alpha_enabled != cur_state.color_mask.alpha_enabled) {
-        glColorMask(color_mask.red_enabled, color_mask.green_enabled, color_mask.blue_enabled,
-                    color_mask.alpha_enabled);
+void OpenGLState::ApplyPrimitiveRestart() const {
+    const bool primitive_restart_changed =
+        primitive_restart.enabled != cur_state.primitive_restart.enabled;
+    if (primitive_restart_changed) {
+        if (primitive_restart.enabled) {
+            glEnable(GL_PRIMITIVE_RESTART);
+        } else {
+            glDisable(GL_PRIMITIVE_RESTART);
+        }
+    }
+    if (primitive_restart_changed ||
+        (primitive_restart.enabled &&
+         primitive_restart.index != cur_state.primitive_restart.index)) {
+        glPrimitiveRestartIndex(primitive_restart.index);
     }
+}
 
-    // Stencil test
-    if (stencil.test_enabled != cur_state.stencil.test_enabled) {
+void OpenGLState::ApplyStencilTest() const {
+    const bool stencil_test_changed = stencil.test_enabled != cur_state.stencil.test_enabled;
+    if (stencil_test_changed) {
         if (stencil.test_enabled) {
             glEnable(GL_STENCIL_TEST);
         } else {
             glDisable(GL_STENCIL_TEST);
         }
     }
-    auto config_stencil = [](GLenum face, const auto& config, const auto& prev_config) {
-        if (config.test_func != prev_config.test_func || config.test_ref != prev_config.test_ref ||
-            config.test_mask != prev_config.test_mask) {
-            glStencilFuncSeparate(face, config.test_func, config.test_ref, config.test_mask);
-        }
-        if (config.action_depth_fail != prev_config.action_depth_fail ||
-            config.action_depth_pass != prev_config.action_depth_pass ||
-            config.action_stencil_fail != prev_config.action_stencil_fail) {
-            glStencilOpSeparate(face, config.action_stencil_fail, config.action_depth_fail,
-                                config.action_depth_pass);
-        }
-        if (config.write_mask != prev_config.write_mask) {
-            glStencilMaskSeparate(face, config.write_mask);
+    if (stencil.test_enabled) {
+        auto config_stencil = [stencil_test_changed](GLenum face, const auto& config,
+                                                     const auto& prev_config) {
+            if (stencil_test_changed || config.test_func != prev_config.test_func ||
+                config.test_ref != prev_config.test_ref ||
+                config.test_mask != prev_config.test_mask) {
+                glStencilFuncSeparate(face, config.test_func, config.test_ref, config.test_mask);
+            }
+            if (stencil_test_changed || config.action_depth_fail != prev_config.action_depth_fail ||
+                config.action_depth_pass != prev_config.action_depth_pass ||
+                config.action_stencil_fail != prev_config.action_stencil_fail) {
+                glStencilOpSeparate(face, config.action_stencil_fail, config.action_depth_fail,
+                                    config.action_depth_pass);
+            }
+            if (config.write_mask != prev_config.write_mask) {
+                glStencilMaskSeparate(face, config.write_mask);
+            }
+        };
+        config_stencil(GL_FRONT, stencil.front, cur_state.stencil.front);
+        config_stencil(GL_BACK, stencil.back, cur_state.stencil.back);
+    }
+}
+
+void OpenGLState::ApplyScissorTest() const {
+    const bool scissor_changed = scissor.enabled != cur_state.scissor.enabled;
+    if (scissor_changed) {
+        if (scissor.enabled) {
+            glEnable(GL_SCISSOR_TEST);
+        } else {
+            glDisable(GL_SCISSOR_TEST);
         }
-    };
-    config_stencil(GL_FRONT, stencil.front, cur_state.stencil.front);
-    config_stencil(GL_BACK, stencil.back, cur_state.stencil.back);
+    }
+    if (scissor_changed || scissor_changed || scissor.x != cur_state.scissor.x ||
+        scissor.y != cur_state.scissor.y || scissor.width != cur_state.scissor.width ||
+        scissor.height != cur_state.scissor.height) {
+        glScissor(scissor.x, scissor.y, scissor.width, scissor.height);
+    }
+}
 
-    // Blending
-    if (blend.enabled != cur_state.blend.enabled) {
+void OpenGLState::ApplyBlending() const {
+    const bool blend_changed = blend.enabled != cur_state.blend.enabled;
+    if (blend_changed) {
         if (blend.enabled) {
             ASSERT(!logic_op.enabled);
             glEnable(GL_BLEND);
@@ -163,29 +234,32 @@ void OpenGLState::Apply() const {
             glDisable(GL_BLEND);
         }
     }
+    if (blend.enabled) {
+        if (blend_changed || blend.color.red != cur_state.blend.color.red ||
+            blend.color.green != cur_state.blend.color.green ||
+            blend.color.blue != cur_state.blend.color.blue ||
+            blend.color.alpha != cur_state.blend.color.alpha) {
+            glBlendColor(blend.color.red, blend.color.green, blend.color.blue, blend.color.alpha);
+        }
 
-    if (blend.color.red != cur_state.blend.color.red ||
-        blend.color.green != cur_state.blend.color.green ||
-        blend.color.blue != cur_state.blend.color.blue ||
-        blend.color.alpha != cur_state.blend.color.alpha) {
-        glBlendColor(blend.color.red, blend.color.green, blend.color.blue, blend.color.alpha);
-    }
-
-    if (blend.src_rgb_func != cur_state.blend.src_rgb_func ||
-        blend.dst_rgb_func != cur_state.blend.dst_rgb_func ||
-        blend.src_a_func != cur_state.blend.src_a_func ||
-        blend.dst_a_func != cur_state.blend.dst_a_func) {
-        glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func, blend.src_a_func,
-                            blend.dst_a_func);
-    }
+        if (blend_changed || blend.src_rgb_func != cur_state.blend.src_rgb_func ||
+            blend.dst_rgb_func != cur_state.blend.dst_rgb_func ||
+            blend.src_a_func != cur_state.blend.src_a_func ||
+            blend.dst_a_func != cur_state.blend.dst_a_func) {
+            glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func, blend.src_a_func,
+                                blend.dst_a_func);
+        }
 
-    if (blend.rgb_equation != cur_state.blend.rgb_equation ||
-        blend.a_equation != cur_state.blend.a_equation) {
-        glBlendEquationSeparate(blend.rgb_equation, blend.a_equation);
+        if (blend_changed || blend.rgb_equation != cur_state.blend.rgb_equation ||
+            blend.a_equation != cur_state.blend.a_equation) {
+            glBlendEquationSeparate(blend.rgb_equation, blend.a_equation);
+        }
     }
+}
 
-    // Logic Operation
-    if (logic_op.enabled != cur_state.logic_op.enabled) {
+void OpenGLState::ApplyLogicOp() const {
+    const bool logic_op_changed = logic_op.enabled != cur_state.logic_op.enabled;
+    if (logic_op_changed) {
         if (logic_op.enabled) {
             ASSERT(!blend.enabled);
             glEnable(GL_COLOR_LOGIC_OP);
@@ -194,11 +268,13 @@ void OpenGLState::Apply() const {
         }
     }
 
-    if (logic_op.operation != cur_state.logic_op.operation) {
+    if (logic_op.enabled &&
+        (logic_op_changed || logic_op.operation != cur_state.logic_op.operation)) {
         glLogicOp(logic_op.operation);
     }
+}
 
-    // Textures
+void OpenGLState::ApplyTextures() const {
     for (std::size_t i = 0; i < std::size(texture_units); ++i) {
         const auto& texture_unit = texture_units[i];
         const auto& cur_state_texture_unit = cur_state.texture_units[i];
@@ -217,28 +293,29 @@ void OpenGLState::Apply() const {
             glTexParameteriv(texture_unit.target, GL_TEXTURE_SWIZZLE_RGBA, mask.data());
         }
     }
+}
 
-    // Samplers
-    {
-        bool has_delta{};
-        std::size_t first{}, last{};
-        std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> samplers;
-        for (std::size_t i = 0; i < std::size(samplers); ++i) {
-            samplers[i] = texture_units[i].sampler;
-            if (samplers[i] != cur_state.texture_units[i].sampler) {
-                if (!has_delta) {
-                    first = i;
-                    has_delta = true;
-                }
-                last = i;
+void OpenGLState::ApplySamplers() const {
+    bool has_delta{};
+    std::size_t first{}, last{};
+    std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> samplers;
+    for (std::size_t i = 0; i < std::size(samplers); ++i) {
+        samplers[i] = texture_units[i].sampler;
+        if (samplers[i] != cur_state.texture_units[i].sampler) {
+            if (!has_delta) {
+                first = i;
+                has_delta = true;
             }
-        }
-        if (has_delta) {
-            glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
-                           samplers.data());
+            last = i;
         }
     }
+    if (has_delta) {
+        glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
+                       samplers.data());
+    }
+}
 
+void OpenGLState::Apply() const {
     // Framebuffer
     if (draw.read_framebuffer != cur_state.draw.read_framebuffer) {
         glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer);
@@ -271,27 +348,12 @@ void OpenGLState::Apply() const {
     if (draw.program_pipeline != cur_state.draw.program_pipeline) {
         glBindProgramPipeline(draw.program_pipeline);
     }
-
-    // Scissor test
-    if (scissor.enabled != cur_state.scissor.enabled) {
-        if (scissor.enabled) {
-            glEnable(GL_SCISSOR_TEST);
-        } else {
-            glDisable(GL_SCISSOR_TEST);
-        }
-    }
-
-    if (scissor.x != cur_state.scissor.x || scissor.y != cur_state.scissor.y ||
-        scissor.width != cur_state.scissor.width || scissor.height != cur_state.scissor.height) {
-        glScissor(scissor.x, scissor.y, scissor.width, scissor.height);
-    }
-
+    // Viewport
     if (viewport.x != cur_state.viewport.x || viewport.y != cur_state.viewport.y ||
         viewport.width != cur_state.viewport.width ||
         viewport.height != cur_state.viewport.height) {
         glViewport(viewport.x, viewport.y, viewport.width, viewport.height);
     }
-
     // Clip distance
     for (std::size_t i = 0; i < clip_distance.size(); ++i) {
         if (clip_distance[i] != cur_state.clip_distance[i]) {
@@ -302,12 +364,28 @@ void OpenGLState::Apply() const {
             }
         }
     }
-
+    // Color mask
+    if (color_mask.red_enabled != cur_state.color_mask.red_enabled ||
+        color_mask.green_enabled != cur_state.color_mask.green_enabled ||
+        color_mask.blue_enabled != cur_state.color_mask.blue_enabled ||
+        color_mask.alpha_enabled != cur_state.color_mask.alpha_enabled) {
+        glColorMask(color_mask.red_enabled, color_mask.green_enabled, color_mask.blue_enabled,
+                    color_mask.alpha_enabled);
+    }
     // Point
     if (point.size != cur_state.point.size) {
         glPointSize(point.size);
     }
-
+    ApplyScissorTest();
+    ApplyStencilTest();
+    ApplySRgb();
+    ApplyCulling();
+    ApplyDepth();
+    ApplyPrimitiveRestart();
+    ApplyBlending();
+    ApplyLogicOp();
+    ApplyTextures();
+    ApplySamplers();
     cur_state = *this;
 }
 
diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h
index dc21a2ee3..fe648aff6 100644
--- a/src/video_core/renderer_opengl/gl_state.h
+++ b/src/video_core/renderer_opengl/gl_state.h
@@ -36,18 +36,29 @@ constexpr TextureUnit ProcTexDiffLUT{9};
 class OpenGLState {
 public:
     struct {
+        bool enabled; // GL_FRAMEBUFFER_SRGB
+    } framebuffer_srgb;
+
+    struct {
         bool enabled;      // GL_CULL_FACE
         GLenum mode;       // GL_CULL_FACE_MODE
         GLenum front_face; // GL_FRONT_FACE
     } cull;
 
     struct {
-        bool test_enabled;    // GL_DEPTH_TEST
-        GLenum test_func;     // GL_DEPTH_FUNC
-        GLboolean write_mask; // GL_DEPTH_WRITEMASK
+        bool test_enabled;        // GL_DEPTH_TEST
+        GLenum test_func;         // GL_DEPTH_FUNC
+        GLboolean write_mask;     // GL_DEPTH_WRITEMASK
+        GLfloat depth_range_near; // GL_DEPTH_RANGE
+        GLfloat depth_range_far;  // GL_DEPTH_RANGE
     } depth;
 
     struct {
+        bool enabled;
+        GLuint index;
+    } primitive_restart; // GL_PRIMITIVE_RESTART
+
+    struct {
         GLboolean red_enabled;
         GLboolean green_enabled;
         GLboolean blue_enabled;
@@ -154,10 +165,16 @@ public:
     static OpenGLState GetCurState() {
         return cur_state;
     }
-
+    static bool GetsRGBUsed() {
+        return s_rgb_used;
+    }
+    static void ClearsRGBUsed() {
+        s_rgb_used = false;
+    }
     /// Apply this state as the current OpenGL state
     void Apply() const;
-
+    /// Set the initial OpenGL state
+    static void ApplyDefaultState();
     /// Resets any references to the given resource
     OpenGLState& UnbindTexture(GLuint handle);
     OpenGLState& ResetSampler(GLuint handle);
@@ -169,6 +186,19 @@ public:
 
 private:
     static OpenGLState cur_state;
+    // Workaround for sRGB problems caused by
+    // QT not supporting srgb output
+    static bool s_rgb_used;
+    void ApplySRgb() const;
+    void ApplyCulling() const;
+    void ApplyDepth() const;
+    void ApplyPrimitiveRestart() const;
+    void ApplyStencilTest() const;
+    void ApplyScissorTest() const;
+    void ApplyBlending() const;
+    void ApplyLogicOp() const;
+    void ApplyTextures() const;
+    void ApplySamplers() const;
 };
 
 } // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h
index 67273e164..87d511c38 100644
--- a/src/video_core/renderer_opengl/maxwell_to_gl.h
+++ b/src/video_core/renderer_opengl/maxwell_to_gl.h
@@ -82,8 +82,20 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) {
         return {};
     }
 
-    case Maxwell::VertexAttribute::Type::Float:
-        return GL_FLOAT;
+    case Maxwell::VertexAttribute::Type::Float: {
+        switch (attrib.size) {
+        case Maxwell::VertexAttribute::Size::Size_16:
+        case Maxwell::VertexAttribute::Size::Size_16_16:
+        case Maxwell::VertexAttribute::Size::Size_16_16_16:
+        case Maxwell::VertexAttribute::Size::Size_16_16_16_16:
+            return GL_HALF_FLOAT;
+        case Maxwell::VertexAttribute::Size::Size_32:
+        case Maxwell::VertexAttribute::Size::Size_32_32:
+        case Maxwell::VertexAttribute::Size::Size_32_32_32:
+        case Maxwell::VertexAttribute::Size::Size_32_32_32_32:
+            return GL_FLOAT;
+        }
+    }
     }
 
     LOG_CRITICAL(Render_OpenGL, "Unimplemented vertex type={}", attrib.TypeString());
@@ -123,12 +135,29 @@ inline GLenum PrimitiveTopology(Maxwell::PrimitiveTopology topology) {
     return {};
 }
 
-inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode) {
+inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode,
+                                Tegra::Texture::TextureMipmapFilter mip_filter_mode) {
     switch (filter_mode) {
-    case Tegra::Texture::TextureFilter::Linear:
-        return GL_LINEAR;
-    case Tegra::Texture::TextureFilter::Nearest:
-        return GL_NEAREST;
+    case Tegra::Texture::TextureFilter::Linear: {
+        switch (mip_filter_mode) {
+        case Tegra::Texture::TextureMipmapFilter::None:
+            return GL_LINEAR;
+        case Tegra::Texture::TextureMipmapFilter::Nearest:
+            return GL_NEAREST_MIPMAP_LINEAR;
+        case Tegra::Texture::TextureMipmapFilter::Linear:
+            return GL_LINEAR_MIPMAP_LINEAR;
+        }
+    }
+    case Tegra::Texture::TextureFilter::Nearest: {
+        switch (mip_filter_mode) {
+        case Tegra::Texture::TextureMipmapFilter::None:
+            return GL_NEAREST;
+        case Tegra::Texture::TextureMipmapFilter::Nearest:
+            return GL_NEAREST_MIPMAP_NEAREST;
+        case Tegra::Texture::TextureMipmapFilter::Linear:
+            return GL_LINEAR_MIPMAP_NEAREST;
+        }
+    }
     }
     LOG_CRITICAL(Render_OpenGL, "Unimplemented texture filter mode={}",
                  static_cast<u32>(filter_mode));
@@ -159,6 +188,31 @@ inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) {
     return {};
 }
 
+inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) {
+    switch (func) {
+    case Tegra::Texture::DepthCompareFunc::Never:
+        return GL_NEVER;
+    case Tegra::Texture::DepthCompareFunc::Less:
+        return GL_LESS;
+    case Tegra::Texture::DepthCompareFunc::LessEqual:
+        return GL_LEQUAL;
+    case Tegra::Texture::DepthCompareFunc::Equal:
+        return GL_EQUAL;
+    case Tegra::Texture::DepthCompareFunc::NotEqual:
+        return GL_NOTEQUAL;
+    case Tegra::Texture::DepthCompareFunc::Greater:
+        return GL_GREATER;
+    case Tegra::Texture::DepthCompareFunc::GreaterEqual:
+        return GL_GEQUAL;
+    case Tegra::Texture::DepthCompareFunc::Always:
+        return GL_ALWAYS;
+    }
+    LOG_CRITICAL(Render_OpenGL, "Unimplemented texture depth compare function ={}",
+                 static_cast<u32>(func));
+    UNREACHABLE();
+    return {};
+}
+
 inline GLenum BlendEquation(Maxwell::Blend::Equation equation) {
     switch (equation) {
     case Maxwell::Blend::Equation::Add:
diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp
index 96d916b07..ea38da932 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.cpp
+++ b/src/video_core/renderer_opengl/renderer_opengl.cpp
@@ -115,7 +115,8 @@ RendererOpenGL::RendererOpenGL(Core::Frontend::EmuWindow& window)
 RendererOpenGL::~RendererOpenGL() = default;
 
 /// Swap buffers (render frame)
-void RendererOpenGL::SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) {
+void RendererOpenGL::SwapBuffers(
+    std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
     ScopeAcquireGLContext acquire_context{render_window};
 
     Core::System::GetInstance().GetPerfStats().EndSystemFrame();
@@ -124,11 +125,11 @@ void RendererOpenGL::SwapBuffers(boost::optional<const Tegra::FramebufferConfig&
     OpenGLState prev_state = OpenGLState::GetCurState();
     state.Apply();
 
-    if (framebuffer != boost::none) {
+    if (framebuffer) {
         // If framebuffer is provided, reload it from memory to a texture
-        if (screen_info.texture.width != (GLsizei)framebuffer->width ||
-            screen_info.texture.height != (GLsizei)framebuffer->height ||
-            screen_info.texture.pixel_format != framebuffer->pixel_format) {
+        if (screen_info.texture.width != (GLsizei)framebuffer->get().width ||
+            screen_info.texture.height != (GLsizei)framebuffer->get().height ||
+            screen_info.texture.pixel_format != framebuffer->get().pixel_format) {
             // Reallocate texture if the framebuffer size has changed.
             // This is expected to not happen very often and hence should not be a
             // performance problem.
@@ -283,7 +284,8 @@ void RendererOpenGL::CreateRasterizer() {
     if (rasterizer) {
         return;
     }
-
+    // Initialize sRGB Usage
+    OpenGLState::ClearsRGBUsed();
     rasterizer = std::make_unique<RasterizerOpenGL>(render_window, screen_info);
 }
 
@@ -356,13 +358,20 @@ void RendererOpenGL::DrawScreenTriangles(const ScreenInfo& screen_info, float x,
 
     state.texture_units[0].texture = screen_info.display_texture;
     state.texture_units[0].swizzle = {GL_RED, GL_GREEN, GL_BLUE, GL_ALPHA};
+    // Workaround brigthness problems in SMO by enabling sRGB in the final output
+    // if it has been used in the frame
+    // Needed because of this bug in QT
+    // QTBUG-50987
+    state.framebuffer_srgb.enabled = OpenGLState::GetsRGBUsed();
     state.Apply();
-
     glBufferSubData(GL_ARRAY_BUFFER, 0, sizeof(vertices), vertices.data());
     glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
-
+    // restore default state
+    state.framebuffer_srgb.enabled = false;
     state.texture_units[0].texture = 0;
     state.Apply();
+    // Clear sRGB state for the next frame
+    OpenGLState::ClearsRGBUsed();
 }
 
 /**
diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h
index 961467a62..c0868c0e4 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.h
+++ b/src/video_core/renderer_opengl/renderer_opengl.h
@@ -51,7 +51,8 @@ public:
     ~RendererOpenGL() override;
 
     /// Swap buffers (render frame)
-    void SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) override;
+    void SwapBuffers(
+        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) override;
 
     /// Initialize the renderer
     bool Init() override;
diff --git a/src/video_core/renderer_opengl/utils.cpp b/src/video_core/renderer_opengl/utils.cpp
new file mode 100644
index 000000000..d84634cb3
--- /dev/null
+++ b/src/video_core/renderer_opengl/utils.cpp
@@ -0,0 +1,38 @@
+// Copyright 2014 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <string>
+#include <fmt/format.h>
+#include <glad/glad.h>
+#include "common/common_types.h"
+#include "video_core/renderer_opengl/utils.h"
+
+namespace OpenGL {
+
+void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string extra_info) {
+    if (!GLAD_GL_KHR_debug) {
+        return; // We don't need to throw an error as this is just for debugging
+    }
+    const std::string nice_addr = fmt::format("0x{:016x}", addr);
+    std::string object_label;
+
+    if (extra_info.empty()) {
+        switch (identifier) {
+        case GL_TEXTURE:
+            object_label = "Texture@" + nice_addr;
+            break;
+        case GL_PROGRAM:
+            object_label = "Shader@" + nice_addr;
+            break;
+        default:
+            object_label = fmt::format("Object(0x{:x})@{}", identifier, nice_addr);
+            break;
+        }
+    } else {
+        object_label = extra_info + '@' + nice_addr;
+    }
+    glObjectLabel(identifier, handle, -1, static_cast<const GLchar*>(object_label.c_str()));
+}
+
+} // namespace OpenGL
\ No newline at end of file
diff --git a/src/video_core/renderer_opengl/utils.h b/src/video_core/renderer_opengl/utils.h
new file mode 100644
index 000000000..1fcb6fc11
--- /dev/null
+++ b/src/video_core/renderer_opengl/utils.h
@@ -0,0 +1,15 @@
+// Copyright 2014 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <string>
+#include <glad/glad.h>
+#include "common/common_types.h"
+
+namespace OpenGL {
+
+void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string extra_info = "");
+
+} // namespace OpenGL
\ No newline at end of file
diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp
new file mode 100644
index 000000000..e284a4604
--- /dev/null
+++ b/src/video_core/surface.cpp
@@ -0,0 +1,483 @@
+// Copyright 2014 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/common_types.h"
+#include "common/math_util.h"
+#include "video_core/surface.h"
+
+namespace VideoCore::Surface {
+
+SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type) {
+    switch (texture_type) {
+    case Tegra::Texture::TextureType::Texture1D:
+        return SurfaceTarget::Texture1D;
+    case Tegra::Texture::TextureType::Texture2D:
+    case Tegra::Texture::TextureType::Texture2DNoMipmap:
+        return SurfaceTarget::Texture2D;
+    case Tegra::Texture::TextureType::Texture3D:
+        return SurfaceTarget::Texture3D;
+    case Tegra::Texture::TextureType::TextureCubemap:
+        return SurfaceTarget::TextureCubemap;
+    case Tegra::Texture::TextureType::Texture1DArray:
+        return SurfaceTarget::Texture1DArray;
+    case Tegra::Texture::TextureType::Texture2DArray:
+        return SurfaceTarget::Texture2DArray;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", static_cast<u32>(texture_type));
+        UNREACHABLE();
+        return SurfaceTarget::Texture2D;
+    }
+}
+
+bool SurfaceTargetIsLayered(SurfaceTarget target) {
+    switch (target) {
+    case SurfaceTarget::Texture1D:
+    case SurfaceTarget::Texture2D:
+    case SurfaceTarget::Texture3D:
+        return false;
+    case SurfaceTarget::Texture1DArray:
+    case SurfaceTarget::Texture2DArray:
+    case SurfaceTarget::TextureCubemap:
+        return true;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
+        UNREACHABLE();
+        return false;
+    }
+}
+
+PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
+    switch (format) {
+    case Tegra::DepthFormat::S8_Z24_UNORM:
+        return PixelFormat::S8Z24;
+    case Tegra::DepthFormat::Z24_S8_UNORM:
+        return PixelFormat::Z24S8;
+    case Tegra::DepthFormat::Z32_FLOAT:
+        return PixelFormat::Z32F;
+    case Tegra::DepthFormat::Z16_UNORM:
+        return PixelFormat::Z16;
+    case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
+        return PixelFormat::Z32FS8;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
+    switch (format) {
+        // TODO (Hexagon12): Converting SRGBA to RGBA is a hack and doesn't completely correct the
+        // gamma.
+    case Tegra::RenderTargetFormat::RGBA8_SRGB:
+        return PixelFormat::RGBA8_SRGB;
+    case Tegra::RenderTargetFormat::RGBA8_UNORM:
+        return PixelFormat::ABGR8U;
+    case Tegra::RenderTargetFormat::RGBA8_SNORM:
+        return PixelFormat::ABGR8S;
+    case Tegra::RenderTargetFormat::RGBA8_UINT:
+        return PixelFormat::ABGR8UI;
+    case Tegra::RenderTargetFormat::BGRA8_SRGB:
+        return PixelFormat::BGRA8_SRGB;
+    case Tegra::RenderTargetFormat::BGRA8_UNORM:
+        return PixelFormat::BGRA8;
+    case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
+        return PixelFormat::A2B10G10R10U;
+    case Tegra::RenderTargetFormat::RGBA16_FLOAT:
+        return PixelFormat::RGBA16F;
+    case Tegra::RenderTargetFormat::RGBA16_UNORM:
+        return PixelFormat::RGBA16U;
+    case Tegra::RenderTargetFormat::RGBA16_UINT:
+        return PixelFormat::RGBA16UI;
+    case Tegra::RenderTargetFormat::RGBA32_FLOAT:
+        return PixelFormat::RGBA32F;
+    case Tegra::RenderTargetFormat::RG32_FLOAT:
+        return PixelFormat::RG32F;
+    case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
+        return PixelFormat::R11FG11FB10F;
+    case Tegra::RenderTargetFormat::B5G6R5_UNORM:
+        return PixelFormat::B5G6R5U;
+    case Tegra::RenderTargetFormat::BGR5A1_UNORM:
+        return PixelFormat::A1B5G5R5U;
+    case Tegra::RenderTargetFormat::RGBA32_UINT:
+        return PixelFormat::RGBA32UI;
+    case Tegra::RenderTargetFormat::R8_UNORM:
+        return PixelFormat::R8U;
+    case Tegra::RenderTargetFormat::R8_UINT:
+        return PixelFormat::R8UI;
+    case Tegra::RenderTargetFormat::RG16_FLOAT:
+        return PixelFormat::RG16F;
+    case Tegra::RenderTargetFormat::RG16_UINT:
+        return PixelFormat::RG16UI;
+    case Tegra::RenderTargetFormat::RG16_SINT:
+        return PixelFormat::RG16I;
+    case Tegra::RenderTargetFormat::RG16_UNORM:
+        return PixelFormat::RG16;
+    case Tegra::RenderTargetFormat::RG16_SNORM:
+        return PixelFormat::RG16S;
+    case Tegra::RenderTargetFormat::RG8_UNORM:
+        return PixelFormat::RG8U;
+    case Tegra::RenderTargetFormat::RG8_SNORM:
+        return PixelFormat::RG8S;
+    case Tegra::RenderTargetFormat::R16_FLOAT:
+        return PixelFormat::R16F;
+    case Tegra::RenderTargetFormat::R16_UNORM:
+        return PixelFormat::R16U;
+    case Tegra::RenderTargetFormat::R16_SNORM:
+        return PixelFormat::R16S;
+    case Tegra::RenderTargetFormat::R16_UINT:
+        return PixelFormat::R16UI;
+    case Tegra::RenderTargetFormat::R16_SINT:
+        return PixelFormat::R16I;
+    case Tegra::RenderTargetFormat::R32_FLOAT:
+        return PixelFormat::R32F;
+    case Tegra::RenderTargetFormat::R32_UINT:
+        return PixelFormat::R32UI;
+    case Tegra::RenderTargetFormat::RG32_UINT:
+        return PixelFormat::RG32UI;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
+                                         Tegra::Texture::ComponentType component_type,
+                                         bool is_srgb) {
+    // TODO(Subv): Properly implement this
+    switch (format) {
+    case Tegra::Texture::TextureFormat::A8R8G8B8:
+        if (is_srgb) {
+            return PixelFormat::RGBA8_SRGB;
+        }
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::ABGR8U;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::ABGR8S;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::ABGR8UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::B5G6R5:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::B5G6R5U;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::A2B10G10R10:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::A2B10G10R10U;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::A1B5G5R5:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::A1B5G5R5U;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R8:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::R8U;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::R8UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::G8R8:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::G8R8U;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::G8R8S;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R16_G16_B16_A16:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::RGBA16U;
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RGBA16F;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::BF10GF11RF11:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::R11FG11FB10F;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R32_G32_B32_A32:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RGBA32F;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::RGBA32UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R32_G32:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RG32F;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::RG32UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R32_G32_B32:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RGB32F;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R16:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::R16F;
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::R16U;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::R16S;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::R16UI;
+        case Tegra::Texture::ComponentType::SINT:
+            return PixelFormat::R16I;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R32:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::R32F;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::R32UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::ZF32:
+        return PixelFormat::Z32F;
+    case Tegra::Texture::TextureFormat::Z16:
+        return PixelFormat::Z16;
+    case Tegra::Texture::TextureFormat::Z24S8:
+        return PixelFormat::Z24S8;
+    case Tegra::Texture::TextureFormat::DXT1:
+        return is_srgb ? PixelFormat::DXT1_SRGB : PixelFormat::DXT1;
+    case Tegra::Texture::TextureFormat::DXT23:
+        return is_srgb ? PixelFormat::DXT23_SRGB : PixelFormat::DXT23;
+    case Tegra::Texture::TextureFormat::DXT45:
+        return is_srgb ? PixelFormat::DXT45_SRGB : PixelFormat::DXT45;
+    case Tegra::Texture::TextureFormat::DXN1:
+        return PixelFormat::DXN1;
+    case Tegra::Texture::TextureFormat::DXN2:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::DXN2UNORM;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::DXN2SNORM;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::BC7U:
+        return is_srgb ? PixelFormat::BC7U_SRGB : PixelFormat::BC7U;
+    case Tegra::Texture::TextureFormat::BC6H_UF16:
+        return PixelFormat::BC6H_UF16;
+    case Tegra::Texture::TextureFormat::BC6H_SF16:
+        return PixelFormat::BC6H_SF16;
+    case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
+        return is_srgb ? PixelFormat::ASTC_2D_4X4_SRGB : PixelFormat::ASTC_2D_4X4;
+    case Tegra::Texture::TextureFormat::ASTC_2D_5X4:
+        return is_srgb ? PixelFormat::ASTC_2D_5X4_SRGB : PixelFormat::ASTC_2D_5X4;
+    case Tegra::Texture::TextureFormat::ASTC_2D_5X5:
+        return is_srgb ? PixelFormat::ASTC_2D_5X5_SRGB : PixelFormat::ASTC_2D_5X5;
+    case Tegra::Texture::TextureFormat::ASTC_2D_8X8:
+        return is_srgb ? PixelFormat::ASTC_2D_8X8_SRGB : PixelFormat::ASTC_2D_8X8;
+    case Tegra::Texture::TextureFormat::ASTC_2D_8X5:
+        return is_srgb ? PixelFormat::ASTC_2D_8X5_SRGB : PixelFormat::ASTC_2D_8X5;
+    case Tegra::Texture::TextureFormat::R16_G16:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RG16F;
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::RG16;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::RG16S;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::RG16UI;
+        case Tegra::Texture::ComponentType::SINT:
+            return PixelFormat::RG16I;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}, component_type={}", static_cast<u32>(format),
+                     static_cast<u32>(component_type));
+        UNREACHABLE();
+    }
+}
+
+ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
+    // TODO(Subv): Implement more component types
+    switch (type) {
+    case Tegra::Texture::ComponentType::UNORM:
+        return ComponentType::UNorm;
+    case Tegra::Texture::ComponentType::FLOAT:
+        return ComponentType::Float;
+    case Tegra::Texture::ComponentType::SNORM:
+        return ComponentType::SNorm;
+    case Tegra::Texture::ComponentType::UINT:
+        return ComponentType::UInt;
+    case Tegra::Texture::ComponentType::SINT:
+        return ComponentType::SInt;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented component type={}", static_cast<u32>(type));
+        UNREACHABLE();
+    }
+}
+
+ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format) {
+    // TODO(Subv): Implement more render targets
+    switch (format) {
+    case Tegra::RenderTargetFormat::RGBA8_UNORM:
+    case Tegra::RenderTargetFormat::RGBA8_SRGB:
+    case Tegra::RenderTargetFormat::BGRA8_UNORM:
+    case Tegra::RenderTargetFormat::BGRA8_SRGB:
+    case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
+    case Tegra::RenderTargetFormat::R8_UNORM:
+    case Tegra::RenderTargetFormat::RG16_UNORM:
+    case Tegra::RenderTargetFormat::R16_UNORM:
+    case Tegra::RenderTargetFormat::B5G6R5_UNORM:
+    case Tegra::RenderTargetFormat::BGR5A1_UNORM:
+    case Tegra::RenderTargetFormat::RG8_UNORM:
+    case Tegra::RenderTargetFormat::RGBA16_UNORM:
+        return ComponentType::UNorm;
+    case Tegra::RenderTargetFormat::RGBA8_SNORM:
+    case Tegra::RenderTargetFormat::RG16_SNORM:
+    case Tegra::RenderTargetFormat::R16_SNORM:
+    case Tegra::RenderTargetFormat::RG8_SNORM:
+        return ComponentType::SNorm;
+    case Tegra::RenderTargetFormat::RGBA16_FLOAT:
+    case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
+    case Tegra::RenderTargetFormat::RGBA32_FLOAT:
+    case Tegra::RenderTargetFormat::RG32_FLOAT:
+    case Tegra::RenderTargetFormat::RG16_FLOAT:
+    case Tegra::RenderTargetFormat::R16_FLOAT:
+    case Tegra::RenderTargetFormat::R32_FLOAT:
+        return ComponentType::Float;
+    case Tegra::RenderTargetFormat::RGBA32_UINT:
+    case Tegra::RenderTargetFormat::RGBA16_UINT:
+    case Tegra::RenderTargetFormat::RG16_UINT:
+    case Tegra::RenderTargetFormat::R8_UINT:
+    case Tegra::RenderTargetFormat::R16_UINT:
+    case Tegra::RenderTargetFormat::RG32_UINT:
+    case Tegra::RenderTargetFormat::R32_UINT:
+    case Tegra::RenderTargetFormat::RGBA8_UINT:
+        return ComponentType::UInt;
+    case Tegra::RenderTargetFormat::RG16_SINT:
+    case Tegra::RenderTargetFormat::R16_SINT:
+        return ComponentType::SInt;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
+    switch (format) {
+    case Tegra::FramebufferConfig::PixelFormat::ABGR8:
+        return PixelFormat::ABGR8U;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format) {
+    switch (format) {
+    case Tegra::DepthFormat::Z16_UNORM:
+    case Tegra::DepthFormat::S8_Z24_UNORM:
+    case Tegra::DepthFormat::Z24_S8_UNORM:
+        return ComponentType::UNorm;
+    case Tegra::DepthFormat::Z32_FLOAT:
+    case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
+        return ComponentType::Float;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+SurfaceType GetFormatType(PixelFormat pixel_format) {
+    if (static_cast<std::size_t>(pixel_format) <
+        static_cast<std::size_t>(PixelFormat::MaxColorFormat)) {
+        return SurfaceType::ColorTexture;
+    }
+
+    if (static_cast<std::size_t>(pixel_format) <
+        static_cast<std::size_t>(PixelFormat::MaxDepthFormat)) {
+        return SurfaceType::Depth;
+    }
+
+    if (static_cast<std::size_t>(pixel_format) <
+        static_cast<std::size_t>(PixelFormat::MaxDepthStencilFormat)) {
+        return SurfaceType::DepthStencil;
+    }
+
+    // TODO(Subv): Implement the other formats
+    ASSERT(false);
+
+    return SurfaceType::Invalid;
+}
+
+bool IsPixelFormatASTC(PixelFormat format) {
+    switch (format) {
+    case PixelFormat::ASTC_2D_4X4:
+    case PixelFormat::ASTC_2D_5X4:
+    case PixelFormat::ASTC_2D_5X5:
+    case PixelFormat::ASTC_2D_8X8:
+    case PixelFormat::ASTC_2D_8X5:
+    case PixelFormat::ASTC_2D_4X4_SRGB:
+    case PixelFormat::ASTC_2D_5X4_SRGB:
+    case PixelFormat::ASTC_2D_5X5_SRGB:
+    case PixelFormat::ASTC_2D_8X8_SRGB:
+    case PixelFormat::ASTC_2D_8X5_SRGB:
+        return true;
+    default:
+        return false;
+    }
+}
+
+std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) {
+    return {GetDefaultBlockWidth(format), GetDefaultBlockHeight(format)};
+}
+
+bool IsFormatBCn(PixelFormat format) {
+    switch (format) {
+    case PixelFormat::DXT1:
+    case PixelFormat::DXT23:
+    case PixelFormat::DXT45:
+    case PixelFormat::DXN1:
+    case PixelFormat::DXN2SNORM:
+    case PixelFormat::DXN2UNORM:
+    case PixelFormat::BC7U:
+    case PixelFormat::BC6H_UF16:
+    case PixelFormat::BC6H_SF16:
+    case PixelFormat::DXT1_SRGB:
+    case PixelFormat::DXT23_SRGB:
+    case PixelFormat::DXT45_SRGB:
+    case PixelFormat::BC7U_SRGB:
+        return true;
+    }
+    return false;
+}
+
+} // namespace VideoCore::Surface
diff --git a/src/video_core/surface.h b/src/video_core/surface.h
new file mode 100644
index 000000000..0ef7849a4
--- /dev/null
+++ b/src/video_core/surface.h
@@ -0,0 +1,466 @@
+// Copyright 2014 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <climits>
+#include <utility>
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "common/logging/log.h"
+#include "video_core/gpu.h"
+#include "video_core/textures/texture.h"
+
+namespace VideoCore::Surface {
+
+enum class PixelFormat {
+    ABGR8U = 0,
+    ABGR8S = 1,
+    ABGR8UI = 2,
+    B5G6R5U = 3,
+    A2B10G10R10U = 4,
+    A1B5G5R5U = 5,
+    R8U = 6,
+    R8UI = 7,
+    RGBA16F = 8,
+    RGBA16U = 9,
+    RGBA16UI = 10,
+    R11FG11FB10F = 11,
+    RGBA32UI = 12,
+    DXT1 = 13,
+    DXT23 = 14,
+    DXT45 = 15,
+    DXN1 = 16, // This is also known as BC4
+    DXN2UNORM = 17,
+    DXN2SNORM = 18,
+    BC7U = 19,
+    BC6H_UF16 = 20,
+    BC6H_SF16 = 21,
+    ASTC_2D_4X4 = 22,
+    G8R8U = 23,
+    G8R8S = 24,
+    BGRA8 = 25,
+    RGBA32F = 26,
+    RG32F = 27,
+    R32F = 28,
+    R16F = 29,
+    R16U = 30,
+    R16S = 31,
+    R16UI = 32,
+    R16I = 33,
+    RG16 = 34,
+    RG16F = 35,
+    RG16UI = 36,
+    RG16I = 37,
+    RG16S = 38,
+    RGB32F = 39,
+    RGBA8_SRGB = 40,
+    RG8U = 41,
+    RG8S = 42,
+    RG32UI = 43,
+    R32UI = 44,
+    ASTC_2D_8X8 = 45,
+    ASTC_2D_8X5 = 46,
+    ASTC_2D_5X4 = 47,
+    BGRA8_SRGB = 48,
+    DXT1_SRGB = 49,
+    DXT23_SRGB = 50,
+    DXT45_SRGB = 51,
+    BC7U_SRGB = 52,
+    ASTC_2D_4X4_SRGB = 53,
+    ASTC_2D_8X8_SRGB = 54,
+    ASTC_2D_8X5_SRGB = 55,
+    ASTC_2D_5X4_SRGB = 56,
+    ASTC_2D_5X5 = 57,
+    ASTC_2D_5X5_SRGB = 58,
+
+    MaxColorFormat,
+
+    // Depth formats
+    Z32F = 59,
+    Z16 = 60,
+
+    MaxDepthFormat,
+
+    // DepthStencil formats
+    Z24S8 = 61,
+    S8Z24 = 62,
+    Z32FS8 = 63,
+
+    MaxDepthStencilFormat,
+
+    Max = MaxDepthStencilFormat,
+    Invalid = 255,
+};
+
+static constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max);
+
+enum class ComponentType {
+    Invalid = 0,
+    SNorm = 1,
+    UNorm = 2,
+    SInt = 3,
+    UInt = 4,
+    Float = 5,
+};
+
+enum class SurfaceType {
+    ColorTexture = 0,
+    Depth = 1,
+    DepthStencil = 2,
+    Fill = 3,
+    Invalid = 4,
+};
+
+enum class SurfaceTarget {
+    Texture1D,
+    Texture2D,
+    Texture3D,
+    Texture1DArray,
+    Texture2DArray,
+    TextureCubemap,
+};
+
+/**
+ * Gets the compression factor for the specified PixelFormat. This applies to just the
+ * "compressed width" and "compressed height", not the overall compression factor of a
+ * compressed image. This is used for maintaining proper surface sizes for compressed
+ * texture formats.
+ */
+static constexpr u32 GetCompressionFactor(PixelFormat format) {
+    if (format == PixelFormat::Invalid)
+        return 0;
+
+    constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{
+        1, // ABGR8U
+        1, // ABGR8S
+        1, // ABGR8UI
+        1, // B5G6R5U
+        1, // A2B10G10R10U
+        1, // A1B5G5R5U
+        1, // R8U
+        1, // R8UI
+        1, // RGBA16F
+        1, // RGBA16U
+        1, // RGBA16UI
+        1, // R11FG11FB10F
+        1, // RGBA32UI
+        4, // DXT1
+        4, // DXT23
+        4, // DXT45
+        4, // DXN1
+        4, // DXN2UNORM
+        4, // DXN2SNORM
+        4, // BC7U
+        4, // BC6H_UF16
+        4, // BC6H_SF16
+        4, // ASTC_2D_4X4
+        1, // G8R8U
+        1, // G8R8S
+        1, // BGRA8
+        1, // RGBA32F
+        1, // RG32F
+        1, // R32F
+        1, // R16F
+        1, // R16U
+        1, // R16S
+        1, // R16UI
+        1, // R16I
+        1, // RG16
+        1, // RG16F
+        1, // RG16UI
+        1, // RG16I
+        1, // RG16S
+        1, // RGB32F
+        1, // RGBA8_SRGB
+        1, // RG8U
+        1, // RG8S
+        1, // RG32UI
+        1, // R32UI
+        4, // ASTC_2D_8X8
+        4, // ASTC_2D_8X5
+        4, // ASTC_2D_5X4
+        1, // BGRA8_SRGB
+        4, // DXT1_SRGB
+        4, // DXT23_SRGB
+        4, // DXT45_SRGB
+        4, // BC7U_SRGB
+        4, // ASTC_2D_4X4_SRGB
+        4, // ASTC_2D_8X8_SRGB
+        4, // ASTC_2D_8X5_SRGB
+        4, // ASTC_2D_5X4_SRGB
+        4, // ASTC_2D_5X5
+        4, // ASTC_2D_5X5_SRGB
+        1, // Z32F
+        1, // Z16
+        1, // Z24S8
+        1, // S8Z24
+        1, // Z32FS8
+    }};
+
+    ASSERT(static_cast<std::size_t>(format) < compression_factor_table.size());
+    return compression_factor_table[static_cast<std::size_t>(format)];
+}
+
+static constexpr u32 GetDefaultBlockWidth(PixelFormat format) {
+    if (format == PixelFormat::Invalid)
+        return 0;
+    constexpr std::array<u32, MaxPixelFormat> block_width_table = {{
+        1, // ABGR8U
+        1, // ABGR8S
+        1, // ABGR8UI
+        1, // B5G6R5U
+        1, // A2B10G10R10U
+        1, // A1B5G5R5U
+        1, // R8U
+        1, // R8UI
+        1, // RGBA16F
+        1, // RGBA16U
+        1, // RGBA16UI
+        1, // R11FG11FB10F
+        1, // RGBA32UI
+        4, // DXT1
+        4, // DXT23
+        4, // DXT45
+        4, // DXN1
+        4, // DXN2UNORM
+        4, // DXN2SNORM
+        4, // BC7U
+        4, // BC6H_UF16
+        4, // BC6H_SF16
+        4, // ASTC_2D_4X4
+        1, // G8R8U
+        1, // G8R8S
+        1, // BGRA8
+        1, // RGBA32F
+        1, // RG32F
+        1, // R32F
+        1, // R16F
+        1, // R16U
+        1, // R16S
+        1, // R16UI
+        1, // R16I
+        1, // RG16
+        1, // RG16F
+        1, // RG16UI
+        1, // RG16I
+        1, // RG16S
+        1, // RGB32F
+        1, // RGBA8_SRGB
+        1, // RG8U
+        1, // RG8S
+        1, // RG32UI
+        1, // R32UI
+        8, // ASTC_2D_8X8
+        8, // ASTC_2D_8X5
+        5, // ASTC_2D_5X4
+        1, // BGRA8_SRGB
+        4, // DXT1_SRGB
+        4, // DXT23_SRGB
+        4, // DXT45_SRGB
+        4, // BC7U_SRGB
+        4, // ASTC_2D_4X4_SRGB
+        8, // ASTC_2D_8X8_SRGB
+        8, // ASTC_2D_8X5_SRGB
+        5, // ASTC_2D_5X4_SRGB
+        5, // ASTC_2D_5X5
+        5, // ASTC_2D_5X5_SRGB
+        1, // Z32F
+        1, // Z16
+        1, // Z24S8
+        1, // S8Z24
+        1, // Z32FS8
+    }};
+    ASSERT(static_cast<std::size_t>(format) < block_width_table.size());
+    return block_width_table[static_cast<std::size_t>(format)];
+}
+
+static constexpr u32 GetDefaultBlockHeight(PixelFormat format) {
+    if (format == PixelFormat::Invalid)
+        return 0;
+
+    constexpr std::array<u32, MaxPixelFormat> block_height_table = {{
+        1, // ABGR8U
+        1, // ABGR8S
+        1, // ABGR8UI
+        1, // B5G6R5U
+        1, // A2B10G10R10U
+        1, // A1B5G5R5U
+        1, // R8U
+        1, // R8UI
+        1, // RGBA16F
+        1, // RGBA16U
+        1, // RGBA16UI
+        1, // R11FG11FB10F
+        1, // RGBA32UI
+        4, // DXT1
+        4, // DXT23
+        4, // DXT45
+        4, // DXN1
+        4, // DXN2UNORM
+        4, // DXN2SNORM
+        4, // BC7U
+        4, // BC6H_UF16
+        4, // BC6H_SF16
+        4, // ASTC_2D_4X4
+        1, // G8R8U
+        1, // G8R8S
+        1, // BGRA8
+        1, // RGBA32F
+        1, // RG32F
+        1, // R32F
+        1, // R16F
+        1, // R16U
+        1, // R16S
+        1, // R16UI
+        1, // R16I
+        1, // RG16
+        1, // RG16F
+        1, // RG16UI
+        1, // RG16I
+        1, // RG16S
+        1, // RGB32F
+        1, // RGBA8_SRGB
+        1, // RG8U
+        1, // RG8S
+        1, // RG32UI
+        1, // R32UI
+        8, // ASTC_2D_8X8
+        5, // ASTC_2D_8X5
+        4, // ASTC_2D_5X4
+        1, // BGRA8_SRGB
+        4, // DXT1_SRGB
+        4, // DXT23_SRGB
+        4, // DXT45_SRGB
+        4, // BC7U_SRGB
+        4, // ASTC_2D_4X4_SRGB
+        8, // ASTC_2D_8X8_SRGB
+        5, // ASTC_2D_8X5_SRGB
+        4, // ASTC_2D_5X4_SRGB
+        5, // ASTC_2D_5X5
+        5, // ASTC_2D_5X5_SRGB
+        1, // Z32F
+        1, // Z16
+        1, // Z24S8
+        1, // S8Z24
+        1, // Z32FS8
+    }};
+
+    ASSERT(static_cast<std::size_t>(format) < block_height_table.size());
+    return block_height_table[static_cast<std::size_t>(format)];
+}
+
+static constexpr u32 GetFormatBpp(PixelFormat format) {
+    if (format == PixelFormat::Invalid)
+        return 0;
+
+    constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
+        32,  // ABGR8U
+        32,  // ABGR8S
+        32,  // ABGR8UI
+        16,  // B5G6R5U
+        32,  // A2B10G10R10U
+        16,  // A1B5G5R5U
+        8,   // R8U
+        8,   // R8UI
+        64,  // RGBA16F
+        64,  // RGBA16U
+        64,  // RGBA16UI
+        32,  // R11FG11FB10F
+        128, // RGBA32UI
+        64,  // DXT1
+        128, // DXT23
+        128, // DXT45
+        64,  // DXN1
+        128, // DXN2UNORM
+        128, // DXN2SNORM
+        128, // BC7U
+        128, // BC6H_UF16
+        128, // BC6H_SF16
+        128, // ASTC_2D_4X4
+        16,  // G8R8U
+        16,  // G8R8S
+        32,  // BGRA8
+        128, // RGBA32F
+        64,  // RG32F
+        32,  // R32F
+        16,  // R16F
+        16,  // R16U
+        16,  // R16S
+        16,  // R16UI
+        16,  // R16I
+        32,  // RG16
+        32,  // RG16F
+        32,  // RG16UI
+        32,  // RG16I
+        32,  // RG16S
+        96,  // RGB32F
+        32,  // RGBA8_SRGB
+        16,  // RG8U
+        16,  // RG8S
+        64,  // RG32UI
+        32,  // R32UI
+        128, // ASTC_2D_8X8
+        128, // ASTC_2D_8X5
+        128, // ASTC_2D_5X4
+        32,  // BGRA8_SRGB
+        64,  // DXT1_SRGB
+        128, // DXT23_SRGB
+        128, // DXT45_SRGB
+        128, // BC7U
+        128, // ASTC_2D_4X4_SRGB
+        128, // ASTC_2D_8X8_SRGB
+        128, // ASTC_2D_8X5_SRGB
+        128, // ASTC_2D_5X4_SRGB
+        128, // ASTC_2D_5X5
+        128, // ASTC_2D_5X5_SRGB
+        32,  // Z32F
+        16,  // Z16
+        32,  // Z24S8
+        32,  // S8Z24
+        64,  // Z32FS8
+    }};
+
+    ASSERT(static_cast<std::size_t>(format) < bpp_table.size());
+    return bpp_table[static_cast<std::size_t>(format)];
+}
+
+/// Returns the sizer in bytes of the specified pixel format
+static constexpr u32 GetBytesPerPixel(PixelFormat pixel_format) {
+    if (pixel_format == PixelFormat::Invalid) {
+        return 0;
+    }
+    return GetFormatBpp(pixel_format) / CHAR_BIT;
+}
+
+SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type);
+
+bool SurfaceTargetIsLayered(SurfaceTarget target);
+
+PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format);
+
+PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format);
+
+PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
+                                         Tegra::Texture::ComponentType component_type,
+                                         bool is_srgb);
+
+ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type);
+
+ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format);
+
+PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format);
+
+ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format);
+
+SurfaceType GetFormatType(PixelFormat pixel_format);
+
+bool IsPixelFormatASTC(PixelFormat format);
+
+std::pair<u32, u32> GetASTCBlockSize(PixelFormat format);
+
+/// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN
+bool IsFormatBCn(PixelFormat format);
+
+} // namespace VideoCore::Surface
diff --git a/src/video_core/textures/astc.cpp b/src/video_core/textures/astc.cpp
index b1feacae9..bc50a4876 100644
--- a/src/video_core/textures/astc.cpp
+++ b/src/video_core/textures/astc.cpp
@@ -1598,27 +1598,29 @@ static void DecompressBlock(uint8_t inBuf[16], const uint32_t blockWidth,
 namespace Tegra::Texture::ASTC {
 
 std::vector<uint8_t> Decompress(std::vector<uint8_t>& data, uint32_t width, uint32_t height,
-                                uint32_t block_width, uint32_t block_height) {
+                                uint32_t depth, uint32_t block_width, uint32_t block_height) {
     uint32_t blockIdx = 0;
-    std::vector<uint8_t> outData(height * width * 4);
-    for (uint32_t j = 0; j < height; j += block_height) {
-        for (uint32_t i = 0; i < width; i += block_width) {
+    std::vector<uint8_t> outData(height * width * depth * 4);
+    for (uint32_t k = 0; k < depth; k++) {
+        for (uint32_t j = 0; j < height; j += block_height) {
+            for (uint32_t i = 0; i < width; i += block_width) {
 
-            uint8_t* blockPtr = data.data() + blockIdx * 16;
+                uint8_t* blockPtr = data.data() + blockIdx * 16;
 
-            // Blocks can be at most 12x12
-            uint32_t uncompData[144];
-            ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData);
+                // Blocks can be at most 12x12
+                uint32_t uncompData[144];
+                ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData);
 
-            uint32_t decompWidth = std::min(block_width, width - i);
-            uint32_t decompHeight = std::min(block_height, height - j);
+                uint32_t decompWidth = std::min(block_width, width - i);
+                uint32_t decompHeight = std::min(block_height, height - j);
 
-            uint8_t* outRow = outData.data() + (j * width + i) * 4;
-            for (uint32_t jj = 0; jj < decompHeight; jj++) {
-                memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4);
-            }
+                uint8_t* outRow = outData.data() + (j * width + i) * 4;
+                for (uint32_t jj = 0; jj < decompHeight; jj++) {
+                    memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4);
+                }
 
-            blockIdx++;
+                blockIdx++;
+            }
         }
     }
 
diff --git a/src/video_core/textures/astc.h b/src/video_core/textures/astc.h
index f0d7c0e56..d419dd025 100644
--- a/src/video_core/textures/astc.h
+++ b/src/video_core/textures/astc.h
@@ -10,6 +10,6 @@
 namespace Tegra::Texture::ASTC {
 
 std::vector<uint8_t> Decompress(std::vector<uint8_t>& data, uint32_t width, uint32_t height,
-                                uint32_t block_width, uint32_t block_height);
+                                uint32_t depth, uint32_t block_width, uint32_t block_height);
 
 } // namespace Tegra::Texture::ASTC
diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp
index 3d5476e5d..3066abf61 100644
--- a/src/video_core/textures/decoders.cpp
+++ b/src/video_core/textures/decoders.cpp
@@ -4,6 +4,7 @@
 
 #include <cmath>
 #include <cstring>
+#include "common/alignment.h"
 #include "common/assert.h"
 #include "core/memory.h"
 #include "video_core/gpu.h"
@@ -39,72 +40,145 @@ struct alignas(64) SwizzleTable {
 constexpr auto legacy_swizzle_table = SwizzleTable<8, 64, 1>();
 constexpr auto fast_swizzle_table = SwizzleTable<8, 4, 16>();
 
-static void LegacySwizzleData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                              u8* swizzled_data, u8* unswizzled_data, bool unswizzle,
-                              u32 block_height) {
+/**
+ * This function manages ALL the GOBs(Group of Bytes) Inside a single block.
+ * Instead of going gob by gob, we map the coordinates inside a block and manage from
+ * those. Block_Width is assumed to be 1.
+ */
+void PreciseProcessBlock(u8* swizzled_data, u8* unswizzled_data, const bool unswizzle,
+                         const u32 x_start, const u32 y_start, const u32 z_start, const u32 x_end,
+                         const u32 y_end, const u32 z_end, const u32 tile_offset,
+                         const u32 xy_block_size, const u32 layer_z, const u32 stride_x,
+                         const u32 bytes_per_pixel, const u32 out_bytes_per_pixel) {
     std::array<u8*, 2> data_ptrs;
-    const std::size_t stride = width * bytes_per_pixel;
-    const std::size_t gobs_in_x = 64;
-    const std::size_t gobs_in_y = 8;
-    const std::size_t gobs_size = gobs_in_x * gobs_in_y;
-    const std::size_t image_width_in_gobs{(stride + gobs_in_x - 1) / gobs_in_x};
-    for (std::size_t y = 0; y < height; ++y) {
-        const std::size_t gob_y_address =
-            (y / (gobs_in_y * block_height)) * gobs_size * block_height * image_width_in_gobs +
-            (y % (gobs_in_y * block_height) / gobs_in_y) * gobs_size;
-        const auto& table = legacy_swizzle_table[y % gobs_in_y];
-        for (std::size_t x = 0; x < width; ++x) {
-            const std::size_t gob_address =
-                gob_y_address + (x * bytes_per_pixel / gobs_in_x) * gobs_size * block_height;
-            const std::size_t x2 = x * bytes_per_pixel;
-            const std::size_t swizzle_offset = gob_address + table[x2 % gobs_in_x];
-            const std::size_t pixel_index = (x + y * width) * out_bytes_per_pixel;
-
-            data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
-            data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
-
-            std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
+    u32 z_address = tile_offset;
+    const u32 gob_size_x = 64;
+    const u32 gob_size_y = 8;
+    const u32 gob_size_z = 1;
+    const u32 gob_size = gob_size_x * gob_size_y * gob_size_z;
+    for (u32 z = z_start; z < z_end; z++) {
+        u32 y_address = z_address;
+        u32 pixel_base = layer_z * z + y_start * stride_x;
+        for (u32 y = y_start; y < y_end; y++) {
+            const auto& table = legacy_swizzle_table[y % gob_size_y];
+            for (u32 x = x_start; x < x_end; x++) {
+                const u32 swizzle_offset{y_address + table[x * bytes_per_pixel % gob_size_x]};
+                const u32 pixel_index{x * out_bytes_per_pixel + pixel_base};
+                data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
+                data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
+                std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
+            }
+            pixel_base += stride_x;
+            if ((y + 1) % gob_size_y == 0)
+                y_address += gob_size;
         }
+        z_address += xy_block_size;
     }
 }
 
-static void FastSwizzleData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                            u8* swizzled_data, u8* unswizzled_data, bool unswizzle,
-                            u32 block_height) {
+/**
+ * This function manages ALL the GOBs(Group of Bytes) Inside a single block.
+ * Instead of going gob by gob, we map the coordinates inside a block and manage from
+ * those. Block_Width is assumed to be 1.
+ */
+void FastProcessBlock(u8* swizzled_data, u8* unswizzled_data, const bool unswizzle,
+                      const u32 x_start, const u32 y_start, const u32 z_start, const u32 x_end,
+                      const u32 y_end, const u32 z_end, const u32 tile_offset,
+                      const u32 xy_block_size, const u32 layer_z, const u32 stride_x,
+                      const u32 bytes_per_pixel, const u32 out_bytes_per_pixel) {
     std::array<u8*, 2> data_ptrs;
-    const std::size_t stride{width * bytes_per_pixel};
-    const std::size_t gobs_in_x = 64;
-    const std::size_t gobs_in_y = 8;
-    const std::size_t gobs_size = gobs_in_x * gobs_in_y;
-    const std::size_t image_width_in_gobs{(stride + gobs_in_x - 1) / gobs_in_x};
-    const std::size_t copy_size{16};
-    for (std::size_t y = 0; y < height; ++y) {
-        const std::size_t initial_gob =
-            (y / (gobs_in_y * block_height)) * gobs_size * block_height * image_width_in_gobs +
-            (y % (gobs_in_y * block_height) / gobs_in_y) * gobs_size;
-        const std::size_t pixel_base{y * width * out_bytes_per_pixel};
-        const auto& table = fast_swizzle_table[y % gobs_in_y];
-        for (std::size_t xb = 0; xb < stride; xb += copy_size) {
-            const std::size_t gob_address{initial_gob +
-                                          (xb / gobs_in_x) * gobs_size * block_height};
-            const std::size_t swizzle_offset{gob_address + table[(xb / 16) % 4]};
-            const std::size_t out_x = xb * out_bytes_per_pixel / bytes_per_pixel;
-            const std::size_t pixel_index{out_x + pixel_base};
-            data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
-            data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
-            std::memcpy(data_ptrs[0], data_ptrs[1], copy_size);
+    u32 z_address = tile_offset;
+    const u32 x_startb = x_start * bytes_per_pixel;
+    const u32 x_endb = x_end * bytes_per_pixel;
+    const u32 copy_size = 16;
+    const u32 gob_size_x = 64;
+    const u32 gob_size_y = 8;
+    const u32 gob_size_z = 1;
+    const u32 gob_size = gob_size_x * gob_size_y * gob_size_z;
+    for (u32 z = z_start; z < z_end; z++) {
+        u32 y_address = z_address;
+        u32 pixel_base = layer_z * z + y_start * stride_x;
+        for (u32 y = y_start; y < y_end; y++) {
+            const auto& table = fast_swizzle_table[y % gob_size_y];
+            for (u32 xb = x_startb; xb < x_endb; xb += copy_size) {
+                const u32 swizzle_offset{y_address + table[(xb / copy_size) % 4]};
+                const u32 out_x = xb * out_bytes_per_pixel / bytes_per_pixel;
+                const u32 pixel_index{out_x + pixel_base};
+                data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
+                data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
+                std::memcpy(data_ptrs[0], data_ptrs[1], copy_size);
+            }
+            pixel_base += stride_x;
+            if ((y + 1) % gob_size_y == 0)
+                y_address += gob_size;
+        }
+        z_address += xy_block_size;
+    }
+}
+
+/**
+ * This function unswizzles or swizzles a texture by mapping Linear to BlockLinear Textue.
+ * The body of this function takes care of splitting the swizzled texture into blocks,
+ * and managing the extents of it. Once all the parameters of a single block are obtained,
+ * the function calls 'ProcessBlock' to process that particular Block.
+ *
+ * Documentation for the memory layout and decoding can be found at:
+ *  https://envytools.readthedocs.io/en/latest/hw/memory/g80-surface.html#blocklinear-surfaces
+ */
+template <bool fast>
+void SwizzledData(u8* swizzled_data, u8* unswizzled_data, const bool unswizzle, const u32 width,
+                  const u32 height, const u32 depth, const u32 bytes_per_pixel,
+                  const u32 out_bytes_per_pixel, const u32 block_height, const u32 block_depth) {
+    auto div_ceil = [](const u32 x, const u32 y) { return ((x + y - 1) / y); };
+    const u32 stride_x = width * out_bytes_per_pixel;
+    const u32 layer_z = height * stride_x;
+    const u32 gob_x_bytes = 64;
+    const u32 gob_elements_x = gob_x_bytes / bytes_per_pixel;
+    const u32 gob_elements_y = 8;
+    const u32 gob_elements_z = 1;
+    const u32 block_x_elements = gob_elements_x;
+    const u32 block_y_elements = gob_elements_y * block_height;
+    const u32 block_z_elements = gob_elements_z * block_depth;
+    const u32 blocks_on_x = div_ceil(width, block_x_elements);
+    const u32 blocks_on_y = div_ceil(height, block_y_elements);
+    const u32 blocks_on_z = div_ceil(depth, block_z_elements);
+    const u32 gob_size = gob_x_bytes * gob_elements_y * gob_elements_z;
+    const u32 xy_block_size = gob_size * block_height;
+    const u32 block_size = xy_block_size * block_depth;
+    u32 tile_offset = 0;
+    for (u32 zb = 0; zb < blocks_on_z; zb++) {
+        const u32 z_start = zb * block_z_elements;
+        const u32 z_end = std::min(depth, z_start + block_z_elements);
+        for (u32 yb = 0; yb < blocks_on_y; yb++) {
+            const u32 y_start = yb * block_y_elements;
+            const u32 y_end = std::min(height, y_start + block_y_elements);
+            for (u32 xb = 0; xb < blocks_on_x; xb++) {
+                const u32 x_start = xb * block_x_elements;
+                const u32 x_end = std::min(width, x_start + block_x_elements);
+                if (fast) {
+                    FastProcessBlock(swizzled_data, unswizzled_data, unswizzle, x_start, y_start,
+                                     z_start, x_end, y_end, z_end, tile_offset, xy_block_size,
+                                     layer_z, stride_x, bytes_per_pixel, out_bytes_per_pixel);
+                } else {
+                    PreciseProcessBlock(swizzled_data, unswizzled_data, unswizzle, x_start, y_start,
+                                        z_start, x_end, y_end, z_end, tile_offset, xy_block_size,
+                                        layer_z, stride_x, bytes_per_pixel, out_bytes_per_pixel);
+                }
+                tile_offset += block_size;
+            }
         }
     }
 }
 
-void CopySwizzledData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                      u8* swizzled_data, u8* unswizzled_data, bool unswizzle, u32 block_height) {
+void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel,
+                      u32 out_bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data,
+                      bool unswizzle, u32 block_height, u32 block_depth) {
     if (bytes_per_pixel % 3 != 0 && (width * bytes_per_pixel) % 16 == 0) {
-        FastSwizzleData(width, height, bytes_per_pixel, out_bytes_per_pixel, swizzled_data,
-                        unswizzled_data, unswizzle, block_height);
+        SwizzledData<true>(swizzled_data, unswizzled_data, unswizzle, width, height, depth,
+                           bytes_per_pixel, out_bytes_per_pixel, block_height, block_depth);
     } else {
-        LegacySwizzleData(width, height, bytes_per_pixel, out_bytes_per_pixel, swizzled_data,
-                          unswizzled_data, unswizzle, block_height);
+        SwizzledData<false>(swizzled_data, unswizzled_data, unswizzle, width, height, depth,
+                            bytes_per_pixel, out_bytes_per_pixel, block_height, block_depth);
     }
 }
 
@@ -125,7 +199,9 @@ u32 BytesPerPixel(TextureFormat format) {
     case TextureFormat::R32_G32_B32:
         return 12;
     case TextureFormat::ASTC_2D_4X4:
+    case TextureFormat::ASTC_2D_5X4:
     case TextureFormat::ASTC_2D_8X8:
+    case TextureFormat::ASTC_2D_8X5:
     case TextureFormat::A8R8G8B8:
     case TextureFormat::A2B10G10R10:
     case TextureFormat::BF10GF11RF11:
@@ -151,14 +227,57 @@ u32 BytesPerPixel(TextureFormat format) {
     }
 }
 
-std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
-                                 u32 height, u32 block_height) {
-    std::vector<u8> unswizzled_data(width * height * bytes_per_pixel);
-    CopySwizzledData(width / tile_size, height / tile_size, bytes_per_pixel, bytes_per_pixel,
-                     Memory::GetPointer(address), unswizzled_data.data(), true, block_height);
+std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size_x, u32 tile_size_y,
+                                 u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
+                                 u32 block_height, u32 block_depth) {
+    std::vector<u8> unswizzled_data(width * height * depth * bytes_per_pixel);
+    CopySwizzledData((width + tile_size_x - 1) / tile_size_x,
+                     (height + tile_size_y - 1) / tile_size_y, depth, bytes_per_pixel,
+                     bytes_per_pixel, Memory::GetPointer(address), unswizzled_data.data(), true,
+                     block_height, block_depth);
     return unswizzled_data;
 }
 
+void SwizzleSubrect(u32 subrect_width, u32 subrect_height, u32 source_pitch, u32 swizzled_width,
+                    u32 bytes_per_pixel, VAddr swizzled_data, VAddr unswizzled_data,
+                    u32 block_height) {
+    const u32 image_width_in_gobs{(swizzled_width * bytes_per_pixel + 63) / 64};
+    for (u32 line = 0; line < subrect_height; ++line) {
+        const u32 gob_address_y =
+            (line / (8 * block_height)) * 512 * block_height * image_width_in_gobs +
+            (line % (8 * block_height) / 8) * 512;
+        const auto& table = legacy_swizzle_table[line % 8];
+        for (u32 x = 0; x < subrect_width; ++x) {
+            const u32 gob_address = gob_address_y + (x * bytes_per_pixel / 64) * 512 * block_height;
+            const u32 swizzled_offset = gob_address + table[(x * bytes_per_pixel) % 64];
+            const VAddr source_line = unswizzled_data + line * source_pitch + x * bytes_per_pixel;
+            const VAddr dest_addr = swizzled_data + swizzled_offset;
+
+            Memory::CopyBlock(dest_addr, source_line, bytes_per_pixel);
+        }
+    }
+}
+
+void UnswizzleSubrect(u32 subrect_width, u32 subrect_height, u32 dest_pitch, u32 swizzled_width,
+                      u32 bytes_per_pixel, VAddr swizzled_data, VAddr unswizzled_data,
+                      u32 block_height, u32 offset_x, u32 offset_y) {
+    for (u32 line = 0; line < subrect_height; ++line) {
+        const u32 y2 = line + offset_y;
+        const u32 gob_address_y =
+            (y2 / (8 * block_height)) * 512 * block_height + (y2 % (8 * block_height) / 8) * 512;
+        const auto& table = legacy_swizzle_table[y2 % 8];
+        for (u32 x = 0; x < subrect_width; ++x) {
+            const u32 x2 = (x + offset_x) * bytes_per_pixel;
+            const u32 gob_address = gob_address_y + (x2 / 64) * 512 * block_height;
+            const u32 swizzled_offset = gob_address + table[x2 % 64];
+            const VAddr dest_line = unswizzled_data + line * dest_pitch + x * bytes_per_pixel;
+            const VAddr source_addr = swizzled_data + swizzled_offset;
+
+            Memory::CopyBlock(dest_line, source_addr, bytes_per_pixel);
+        }
+    }
+}
+
 std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat format, u32 width,
                               u32 height) {
     std::vector<u8> rgba_data;
@@ -199,4 +318,19 @@ std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat
     return rgba_data;
 }
 
+std::size_t CalculateSize(bool tiled, u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
+                          u32 block_height, u32 block_depth) {
+    if (tiled) {
+        const u32 gobs_in_x = 64;
+        const u32 gobs_in_y = 8;
+        const u32 gobs_in_z = 1;
+        const u32 aligned_width = Common::AlignUp(width * bytes_per_pixel, gobs_in_x);
+        const u32 aligned_height = Common::AlignUp(height, gobs_in_y * block_height);
+        const u32 aligned_depth = Common::AlignUp(depth, gobs_in_z * block_depth);
+        return aligned_width * aligned_height * aligned_depth;
+    } else {
+        return width * height * depth * bytes_per_pixel;
+    }
+}
+
 } // namespace Tegra::Texture
diff --git a/src/video_core/textures/decoders.h b/src/video_core/textures/decoders.h
index 1f7b731be..ba065510b 100644
--- a/src/video_core/textures/decoders.h
+++ b/src/video_core/textures/decoders.h
@@ -10,21 +10,24 @@
 
 namespace Tegra::Texture {
 
-/**
- * Unswizzles a swizzled texture without changing its format.
- */
-std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
-                                 u32 height, u32 block_height = TICEntry::DefaultBlockHeight);
+// GOBSize constant. Calculated by 64 bytes in x multiplied by 8 y coords, represents
+// an small rect of (64/bytes_per_pixel)X8.
+inline std::size_t GetGOBSize() {
+    return 512;
+}
 
 /**
- * Unswizzles a swizzled depth texture without changing its format.
+ * Unswizzles a swizzled texture without changing its format.
  */
-std::vector<u8> UnswizzleDepthTexture(VAddr address, DepthFormat format, u32 width, u32 height,
-                                      u32 block_height = TICEntry::DefaultBlockHeight);
+std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size_x, u32 tile_size_y,
+                                 u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
+                                 u32 block_height = TICEntry::DefaultBlockHeight,
+                                 u32 block_depth = TICEntry::DefaultBlockHeight);
 
 /// Copies texture data from a buffer and performs swizzling/unswizzling as necessary.
-void CopySwizzledData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                      u8* swizzled_data, u8* unswizzled_data, bool unswizzle, u32 block_height);
+void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel,
+                      u32 out_bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data,
+                      bool unswizzle, u32 block_height, u32 block_depth);
 
 /**
  * Decodes an unswizzled texture into a A8R8G8B8 texture.
@@ -32,4 +35,19 @@ void CopySwizzledData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_
 std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat format, u32 width,
                               u32 height);
 
+/**
+ * This function calculates the correct size of a texture depending if it's tiled or not.
+ */
+std::size_t CalculateSize(bool tiled, u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
+                          u32 block_height, u32 block_depth);
+
+/// Copies an untiled subrectangle into a tiled surface.
+void SwizzleSubrect(u32 subrect_width, u32 subrect_height, u32 source_pitch, u32 swizzled_width,
+                    u32 bytes_per_pixel, VAddr swizzled_data, VAddr unswizzled_data,
+                    u32 block_height);
+/// Copies a tiled subrectangle into a linear surface.
+void UnswizzleSubrect(u32 subrect_width, u32 subrect_height, u32 dest_pitch, u32 swizzled_width,
+                      u32 bytes_per_pixel, VAddr swizzled_data, VAddr unswizzled_data,
+                      u32 block_height, u32 offset_x, u32 offset_y);
+
 } // namespace Tegra::Texture
diff --git a/src/video_core/textures/texture.h b/src/video_core/textures/texture.h
index 14aea4838..d12d2ecb8 100644
--- a/src/video_core/textures/texture.h
+++ b/src/video_core/textures/texture.h
@@ -141,6 +141,7 @@ static_assert(sizeof(TextureHandle) == 4, "TextureHandle has wrong size");
 
 struct TICEntry {
     static constexpr u32 DefaultBlockHeight = 16;
+    static constexpr u32 DefaultBlockDepth = 1;
 
     union {
         u32 raw;
@@ -161,7 +162,9 @@ struct TICEntry {
         BitField<21, 3, TICHeaderVersion> header_version;
     };
     union {
+        BitField<0, 3, u32> block_width;
         BitField<3, 3, u32> block_height;
+        BitField<6, 3, u32> block_depth;
 
         // High 16 bits of the pitch value
         BitField<0, 16, u32> pitch_high;
@@ -170,6 +173,7 @@ struct TICEntry {
     };
     union {
         BitField<0, 16, u32> width_minus_1;
+        BitField<22, 1, u32> srgb_conversion;
         BitField<23, 4, TextureType> texture_type;
     };
     union {
@@ -202,17 +206,32 @@ struct TICEntry {
         return depth_minus_1 + 1;
     }
 
+    u32 BlockWidth() const {
+        ASSERT(IsTiled());
+        // The block height is stored in log2 format.
+        return 1 << block_width;
+    }
+
     u32 BlockHeight() const {
-        ASSERT(header_version == TICHeaderVersion::BlockLinear ||
-               header_version == TICHeaderVersion::BlockLinearColorKey);
+        ASSERT(IsTiled());
         // The block height is stored in log2 format.
         return 1 << block_height;
     }
 
+    u32 BlockDepth() const {
+        ASSERT(IsTiled());
+        // The block height is stored in log2 format.
+        return 1 << block_depth;
+    }
+
     bool IsTiled() const {
         return header_version == TICHeaderVersion::BlockLinear ||
                header_version == TICHeaderVersion::BlockLinearColorKey;
     }
+
+    bool IsSrgbConversionEnabled() const {
+        return srgb_conversion != 0;
+    }
 };
 static_assert(sizeof(TICEntry) == 0x20, "TICEntry has wrong size");
 
@@ -227,6 +246,17 @@ enum class WrapMode : u32 {
     MirrorOnceClampOGL = 7,
 };
 
+enum class DepthCompareFunc : u32 {
+    Never = 0,
+    Less = 1,
+    Equal = 2,
+    LessEqual = 3,
+    Greater = 4,
+    NotEqual = 5,
+    GreaterEqual = 6,
+    Always = 7,
+};
+
 enum class TextureFilter : u32 {
     Nearest = 1,
     Linear = 2,
@@ -244,7 +274,7 @@ struct TSCEntry {
         BitField<3, 3, WrapMode> wrap_v;
         BitField<6, 3, WrapMode> wrap_p;
         BitField<9, 1, u32> depth_compare_enabled;
-        BitField<10, 3, u32> depth_compare_func;
+        BitField<10, 3, DepthCompareFunc> depth_compare_func;
     };
     union {
         BitField<0, 2, TextureFilter> mag_filter;
diff --git a/src/video_core/utils.h b/src/video_core/utils.h
index 681919ae3..e0a14d48f 100644
--- a/src/video_core/utils.h
+++ b/src/video_core/utils.h
@@ -161,26 +161,4 @@ static inline void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixe
     }
 }
 
-static void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr,
-                          std::string extra_info = "") {
-    if (!GLAD_GL_KHR_debug) {
-        return; // We don't need to throw an error as this is just for debugging
-    }
-    const std::string nice_addr = fmt::format("0x{:016x}", addr);
-    std::string object_label;
-
-    switch (identifier) {
-    case GL_TEXTURE:
-        object_label = extra_info + "@" + nice_addr;
-        break;
-    case GL_PROGRAM:
-        object_label = "ShaderProgram@" + nice_addr;
-        break;
-    default:
-        object_label = fmt::format("Object(0x{:x})@{}", identifier, nice_addr);
-        break;
-    }
-    glObjectLabel(identifier, handle, -1, static_cast<const GLchar*>(object_label.c_str()));
-}
-
 } // namespace VideoCore
diff --git a/src/web_service/CMakeLists.txt b/src/web_service/CMakeLists.txt
index 1c83e9c34..01f2d129d 100644
--- a/src/web_service/CMakeLists.txt
+++ b/src/web_service/CMakeLists.txt
@@ -10,7 +10,7 @@ add_library(web_service STATIC
 create_target_directory_groups(web_service)
 
 get_directory_property(OPENSSL_LIBS
-        DIRECTORY ${CMAKE_SOURCE_DIR}/externals/libressl
+        DIRECTORY ${PROJECT_SOURCE_DIR}/externals/libressl
         DEFINITION OPENSSL_LIBS)
-target_compile_definitions(web_service PUBLIC -DCPPHTTPLIB_OPENSSL_SUPPORT)
+target_compile_definitions(web_service PRIVATE -DCPPHTTPLIB_OPENSSL_SUPPORT)
 target_link_libraries(web_service PRIVATE common json-headers ${OPENSSL_LIBS} httplib lurlparser)
diff --git a/src/web_service/telemetry_json.cpp b/src/web_service/telemetry_json.cpp
index 033ea1ea4..9156ce802 100644
--- a/src/web_service/telemetry_json.cpp
+++ b/src/web_service/telemetry_json.cpp
@@ -2,98 +2,127 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <thread>
-#include "common/assert.h"
+#include <json.hpp>
 #include "common/detached_tasks.h"
+#include "common/web_result.h"
 #include "web_service/telemetry_json.h"
 #include "web_service/web_backend.h"
 
 namespace WebService {
 
-TelemetryJson::TelemetryJson(const std::string& host, const std::string& username,
-                             const std::string& token)
-    : host(std::move(host)), username(std::move(username)), token(std::move(token)) {}
-TelemetryJson::~TelemetryJson() = default;
+struct TelemetryJson::Impl {
+    Impl(std::string host, std::string username, std::string token)
+        : host{std::move(host)}, username{std::move(username)}, token{std::move(token)} {}
 
-template <class T>
-void TelemetryJson::Serialize(Telemetry::FieldType type, const std::string& name, T value) {
-    sections[static_cast<u8>(type)][name] = value;
-}
+    nlohmann::json& TopSection() {
+        return sections[static_cast<u8>(Telemetry::FieldType::None)];
+    }
 
-void TelemetryJson::SerializeSection(Telemetry::FieldType type, const std::string& name) {
-    TopSection()[name] = sections[static_cast<unsigned>(type)];
-}
+    const nlohmann::json& TopSection() const {
+        return sections[static_cast<u8>(Telemetry::FieldType::None)];
+    }
+
+    template <class T>
+    void Serialize(Telemetry::FieldType type, const std::string& name, T value) {
+        sections[static_cast<u8>(type)][name] = value;
+    }
+
+    void SerializeSection(Telemetry::FieldType type, const std::string& name) {
+        TopSection()[name] = sections[static_cast<unsigned>(type)];
+    }
+
+    nlohmann::json output;
+    std::array<nlohmann::json, 7> sections;
+    std::string host;
+    std::string username;
+    std::string token;
+};
+
+TelemetryJson::TelemetryJson(std::string host, std::string username, std::string token)
+    : impl{std::make_unique<Impl>(std::move(host), std::move(username), std::move(token))} {}
+TelemetryJson::~TelemetryJson() = default;
 
 void TelemetryJson::Visit(const Telemetry::Field<bool>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<double>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<float>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<u8>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<u16>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<u32>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<u64>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<s8>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<s16>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<s32>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<s64>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<std::string>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue());
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<const char*>& field) {
-    Serialize(field.GetType(), field.GetName(), std::string(field.GetValue()));
+    impl->Serialize(field.GetType(), field.GetName(), std::string(field.GetValue()));
 }
 
 void TelemetryJson::Visit(const Telemetry::Field<std::chrono::microseconds>& field) {
-    Serialize(field.GetType(), field.GetName(), field.GetValue().count());
+    impl->Serialize(field.GetType(), field.GetName(), field.GetValue().count());
 }
 
 void TelemetryJson::Complete() {
-    SerializeSection(Telemetry::FieldType::App, "App");
-    SerializeSection(Telemetry::FieldType::Session, "Session");
-    SerializeSection(Telemetry::FieldType::Performance, "Performance");
-    SerializeSection(Telemetry::FieldType::UserFeedback, "UserFeedback");
-    SerializeSection(Telemetry::FieldType::UserConfig, "UserConfig");
-    SerializeSection(Telemetry::FieldType::UserSystem, "UserSystem");
-
-    auto content = TopSection().dump();
+    impl->SerializeSection(Telemetry::FieldType::App, "App");
+    impl->SerializeSection(Telemetry::FieldType::Session, "Session");
+    impl->SerializeSection(Telemetry::FieldType::Performance, "Performance");
+    impl->SerializeSection(Telemetry::FieldType::UserConfig, "UserConfig");
+    impl->SerializeSection(Telemetry::FieldType::UserSystem, "UserSystem");
+
+    auto content = impl->TopSection().dump();
     // Send the telemetry async but don't handle the errors since they were written to the log
-    Common::DetachedTasks::AddTask(
-        [host{this->host}, username{this->username}, token{this->token}, content]() {
-            Client{host, username, token}.PostJson("/telemetry", content, true);
-        });
+    Common::DetachedTasks::AddTask([host{impl->host}, content]() {
+        Client{host, "", ""}.PostJson("/telemetry", content, true);
+    });
+}
+
+bool TelemetryJson::SubmitTestcase() {
+    impl->SerializeSection(Telemetry::FieldType::App, "App");
+    impl->SerializeSection(Telemetry::FieldType::Session, "Session");
+    impl->SerializeSection(Telemetry::FieldType::UserFeedback, "UserFeedback");
+    impl->SerializeSection(Telemetry::FieldType::UserSystem, "UserSystem");
+
+    auto content = impl->TopSection().dump();
+    Client client(impl->host, impl->username, impl->token);
+    auto value = client.PostJson("/gamedb/testcase", content, false);
+
+    return value.result_code == Common::WebResult::Code::Success;
 }
 
 } // namespace WebService
diff --git a/src/web_service/telemetry_json.h b/src/web_service/telemetry_json.h
index 0fe6f9a3e..dfd202829 100644
--- a/src/web_service/telemetry_json.h
+++ b/src/web_service/telemetry_json.h
@@ -4,11 +4,9 @@
 
 #pragma once
 
-#include <array>
+#include <chrono>
 #include <string>
-#include <json.hpp>
 #include "common/telemetry.h"
-#include "common/web_result.h"
 
 namespace WebService {
 
@@ -18,8 +16,8 @@ namespace WebService {
  */
 class TelemetryJson : public Telemetry::VisitorInterface {
 public:
-    TelemetryJson(const std::string& host, const std::string& username, const std::string& token);
-    ~TelemetryJson();
+    TelemetryJson(std::string host, std::string username, std::string token);
+    ~TelemetryJson() override;
 
     void Visit(const Telemetry::Field<bool>& field) override;
     void Visit(const Telemetry::Field<double>& field) override;
@@ -37,22 +35,11 @@ public:
     void Visit(const Telemetry::Field<std::chrono::microseconds>& field) override;
 
     void Complete() override;
+    bool SubmitTestcase() override;
 
 private:
-    nlohmann::json& TopSection() {
-        return sections[static_cast<u8>(Telemetry::FieldType::None)];
-    }
-
-    template <class T>
-    void Serialize(Telemetry::FieldType type, const std::string& name, T value);
-
-    void SerializeSection(Telemetry::FieldType type, const std::string& name);
-
-    nlohmann::json output;
-    std::array<nlohmann::json, 7> sections;
-    std::string host;
-    std::string username;
-    std::string token;
+    struct Impl;
+    std::unique_ptr<Impl> impl;
 };
 
 } // namespace WebService
diff --git a/src/web_service/verify_login.cpp b/src/web_service/verify_login.cpp
index 124aa3863..ca4b43b93 100644
--- a/src/web_service/verify_login.cpp
+++ b/src/web_service/verify_login.cpp
@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <json.hpp>
+#include "common/web_result.h"
 #include "web_service/verify_login.h"
 #include "web_service/web_backend.h"
 
diff --git a/src/web_service/web_backend.cpp b/src/web_service/web_backend.cpp
index 787b0fbcb..b7737b615 100644
--- a/src/web_service/web_backend.cpp
+++ b/src/web_service/web_backend.cpp
@@ -3,9 +3,11 @@
 // Refer to the license.txt file included.
 
 #include <cstdlib>
+#include <mutex>
 #include <string>
-#include <thread>
 #include <LUrlParser.h>
+#include <httplib.h>
+#include "common/common_types.h"
 #include "common/logging/log.h"
 #include "common/web_result.h"
 #include "core/settings.h"
@@ -20,99 +22,132 @@ constexpr u32 HTTPS_PORT = 443;
 
 constexpr u32 TIMEOUT_SECONDS = 30;
 
-Client::JWTCache Client::jwt_cache{};
+struct Client::Impl {
+    Impl(std::string host, std::string username, std::string token)
+        : host{std::move(host)}, username{std::move(username)}, token{std::move(token)} {
+        std::lock_guard<std::mutex> lock(jwt_cache.mutex);
+        if (this->username == jwt_cache.username && this->token == jwt_cache.token) {
+            jwt = jwt_cache.jwt;
+        }
+    }
+
+    /// A generic function handles POST, GET and DELETE request together
+    Common::WebResult GenericJson(const std::string& method, const std::string& path,
+                                  const std::string& data, bool allow_anonymous) {
+        if (jwt.empty()) {
+            UpdateJWT();
+        }
+
+        if (jwt.empty() && !allow_anonymous) {
+            LOG_ERROR(WebService, "Credentials must be provided for authenticated requests");
+            return Common::WebResult{Common::WebResult::Code::CredentialsMissing,
+                                     "Credentials needed"};
+        }
+
+        auto result = GenericJson(method, path, data, jwt);
+        if (result.result_string == "401") {
+            // Try again with new JWT
+            UpdateJWT();
+            result = GenericJson(method, path, data, jwt);
+        }
 
-Client::Client(const std::string& host, const std::string& username, const std::string& token)
-    : host(host), username(username), token(token) {
-    std::lock_guard<std::mutex> lock(jwt_cache.mutex);
-    if (username == jwt_cache.username && token == jwt_cache.token) {
-        jwt = jwt_cache.jwt;
+        return result;
     }
-}
 
-Common::WebResult Client::GenericJson(const std::string& method, const std::string& path,
-                                      const std::string& data, const std::string& jwt,
-                                      const std::string& username, const std::string& token) {
-    if (cli == nullptr) {
-        auto parsedUrl = LUrlParser::clParseURL::ParseURL(host);
-        int port;
-        if (parsedUrl.m_Scheme == "http") {
-            if (!parsedUrl.GetPort(&port)) {
-                port = HTTP_PORT;
-            }
-            cli =
-                std::make_unique<httplib::Client>(parsedUrl.m_Host.c_str(), port, TIMEOUT_SECONDS);
-        } else if (parsedUrl.m_Scheme == "https") {
-            if (!parsedUrl.GetPort(&port)) {
-                port = HTTPS_PORT;
+    /**
+     * A generic function with explicit authentication method specified
+     * JWT is used if the jwt parameter is not empty
+     * username + token is used if jwt is empty but username and token are not empty
+     * anonymous if all of jwt, username and token are empty
+     */
+    Common::WebResult GenericJson(const std::string& method, const std::string& path,
+                                  const std::string& data, const std::string& jwt = "",
+                                  const std::string& username = "", const std::string& token = "") {
+        if (cli == nullptr) {
+            auto parsedUrl = LUrlParser::clParseURL::ParseURL(host);
+            int port;
+            if (parsedUrl.m_Scheme == "http") {
+                if (!parsedUrl.GetPort(&port)) {
+                    port = HTTP_PORT;
+                }
+                cli = std::make_unique<httplib::Client>(parsedUrl.m_Host.c_str(), port,
+                                                        TIMEOUT_SECONDS);
+            } else if (parsedUrl.m_Scheme == "https") {
+                if (!parsedUrl.GetPort(&port)) {
+                    port = HTTPS_PORT;
+                }
+                cli = std::make_unique<httplib::SSLClient>(parsedUrl.m_Host.c_str(), port,
+                                                           TIMEOUT_SECONDS);
+            } else {
+                LOG_ERROR(WebService, "Bad URL scheme {}", parsedUrl.m_Scheme);
+                return Common::WebResult{Common::WebResult::Code::InvalidURL, "Bad URL scheme"};
             }
-            cli = std::make_unique<httplib::SSLClient>(parsedUrl.m_Host.c_str(), port,
-                                                       TIMEOUT_SECONDS);
-        } else {
-            LOG_ERROR(WebService, "Bad URL scheme {}", parsedUrl.m_Scheme);
-            return Common::WebResult{Common::WebResult::Code::InvalidURL, "Bad URL scheme"};
         }
-    }
-    if (cli == nullptr) {
-        LOG_ERROR(WebService, "Invalid URL {}", host + path);
-        return Common::WebResult{Common::WebResult::Code::InvalidURL, "Invalid URL"};
-    }
+        if (cli == nullptr) {
+            LOG_ERROR(WebService, "Invalid URL {}", host + path);
+            return Common::WebResult{Common::WebResult::Code::InvalidURL, "Invalid URL"};
+        }
 
-    httplib::Headers params;
-    if (!jwt.empty()) {
-        params = {
-            {std::string("Authorization"), fmt::format("Bearer {}", jwt)},
-        };
-    } else if (!username.empty()) {
-        params = {
-            {std::string("x-username"), username},
-            {std::string("x-token"), token},
+        httplib::Headers params;
+        if (!jwt.empty()) {
+            params = {
+                {std::string("Authorization"), fmt::format("Bearer {}", jwt)},
+            };
+        } else if (!username.empty()) {
+            params = {
+                {std::string("x-username"), username},
+                {std::string("x-token"), token},
+            };
+        }
+
+        params.emplace(std::string("api-version"),
+                       std::string(API_VERSION.begin(), API_VERSION.end()));
+        if (method != "GET") {
+            params.emplace(std::string("Content-Type"), std::string("application/json"));
         };
-    }
 
-    params.emplace(std::string("api-version"), std::string(API_VERSION.begin(), API_VERSION.end()));
-    if (method != "GET") {
-        params.emplace(std::string("Content-Type"), std::string("application/json"));
-    };
+        httplib::Request request;
+        request.method = method;
+        request.path = path;
+        request.headers = params;
+        request.body = data;
 
-    httplib::Request request;
-    request.method = method;
-    request.path = path;
-    request.headers = params;
-    request.body = data;
+        httplib::Response response;
 
-    httplib::Response response;
+        if (!cli->send(request, response)) {
+            LOG_ERROR(WebService, "{} to {} returned null", method, host + path);
+            return Common::WebResult{Common::WebResult::Code::LibError, "Null response"};
+        }
 
-    if (!cli->send(request, response)) {
-        LOG_ERROR(WebService, "{} to {} returned null", method, host + path);
-        return Common::WebResult{Common::WebResult::Code::LibError, "Null response"};
-    }
+        if (response.status >= 400) {
+            LOG_ERROR(WebService, "{} to {} returned error status code: {}", method, host + path,
+                      response.status);
+            return Common::WebResult{Common::WebResult::Code::HttpError,
+                                     std::to_string(response.status)};
+        }
 
-    if (response.status >= 400) {
-        LOG_ERROR(WebService, "{} to {} returned error status code: {}", method, host + path,
-                  response.status);
-        return Common::WebResult{Common::WebResult::Code::HttpError,
-                                 std::to_string(response.status)};
-    }
+        auto content_type = response.headers.find("content-type");
 
-    auto content_type = response.headers.find("content-type");
+        if (content_type == response.headers.end()) {
+            LOG_ERROR(WebService, "{} to {} returned no content", method, host + path);
+            return Common::WebResult{Common::WebResult::Code::WrongContent, ""};
+        }
 
-    if (content_type == response.headers.end()) {
-        LOG_ERROR(WebService, "{} to {} returned no content", method, host + path);
-        return Common::WebResult{Common::WebResult::Code::WrongContent, ""};
+        if (content_type->second.find("application/json") == std::string::npos &&
+            content_type->second.find("text/html; charset=utf-8") == std::string::npos) {
+            LOG_ERROR(WebService, "{} to {} returned wrong content: {}", method, host + path,
+                      content_type->second);
+            return Common::WebResult{Common::WebResult::Code::WrongContent, "Wrong content"};
+        }
+        return Common::WebResult{Common::WebResult::Code::Success, "", response.body};
     }
 
-    if (content_type->second.find("application/json") == std::string::npos &&
-        content_type->second.find("text/html; charset=utf-8") == std::string::npos) {
-        LOG_ERROR(WebService, "{} to {} returned wrong content: {}", method, host + path,
-                  content_type->second);
-        return Common::WebResult{Common::WebResult::Code::WrongContent, "Wrong content"};
-    }
-    return Common::WebResult{Common::WebResult::Code::Success, "", response.body};
-}
+    // Retrieve a new JWT from given username and token
+    void UpdateJWT() {
+        if (username.empty() || token.empty()) {
+            return;
+        }
 
-void Client::UpdateJWT() {
-    if (!username.empty() && !token.empty()) {
         auto result = GenericJson("POST", "/jwt/internal", "", "", username, token);
         if (result.result_code != Common::WebResult::Code::Success) {
             LOG_ERROR(WebService, "UpdateJWT failed");
@@ -123,27 +158,39 @@ void Client::UpdateJWT() {
             jwt_cache.jwt = jwt = result.returned_data;
         }
     }
-}
 
-Common::WebResult Client::GenericJson(const std::string& method, const std::string& path,
-                                      const std::string& data, bool allow_anonymous) {
-    if (jwt.empty()) {
-        UpdateJWT();
-    }
+    std::string host;
+    std::string username;
+    std::string token;
+    std::string jwt;
+    std::unique_ptr<httplib::Client> cli;
+
+    struct JWTCache {
+        std::mutex mutex;
+        std::string username;
+        std::string token;
+        std::string jwt;
+    };
+    static inline JWTCache jwt_cache;
+};
 
-    if (jwt.empty() && !allow_anonymous) {
-        LOG_ERROR(WebService, "Credentials must be provided for authenticated requests");
-        return Common::WebResult{Common::WebResult::Code::CredentialsMissing, "Credentials needed"};
-    }
+Client::Client(std::string host, std::string username, std::string token)
+    : impl{std::make_unique<Impl>(std::move(host), std::move(username), std::move(token))} {}
 
-    auto result = GenericJson(method, path, data, jwt);
-    if (result.result_string == "401") {
-        // Try again with new JWT
-        UpdateJWT();
-        result = GenericJson(method, path, data, jwt);
-    }
+Client::~Client() = default;
+
+Common::WebResult Client::PostJson(const std::string& path, const std::string& data,
+                                   bool allow_anonymous) {
+    return impl->GenericJson("POST", path, data, allow_anonymous);
+}
+
+Common::WebResult Client::GetJson(const std::string& path, bool allow_anonymous) {
+    return impl->GenericJson("GET", path, "", allow_anonymous);
+}
 
-    return result;
+Common::WebResult Client::DeleteJson(const std::string& path, const std::string& data,
+                                     bool allow_anonymous) {
+    return impl->GenericJson("DELETE", path, data, allow_anonymous);
 }
 
 } // namespace WebService
diff --git a/src/web_service/web_backend.h b/src/web_service/web_backend.h
index d75fbcc15..c637e09df 100644
--- a/src/web_service/web_backend.h
+++ b/src/web_service/web_backend.h
@@ -4,23 +4,19 @@
 
 #pragma once
 
-#include <functional>
-#include <mutex>
+#include <memory>
 #include <string>
-#include <tuple>
-#include <httplib.h>
-#include "common/common_types.h"
-#include "common/web_result.h"
 
-namespace httplib {
-class Client;
+namespace Common {
+struct WebResult;
 }
 
 namespace WebService {
 
 class Client {
 public:
-    Client(const std::string& host, const std::string& username, const std::string& token);
+    Client(std::string host, std::string username, std::string token);
+    ~Client();
 
     /**
      * Posts JSON to the specified path.
@@ -30,9 +26,7 @@ public:
      * @return the result of the request.
      */
     Common::WebResult PostJson(const std::string& path, const std::string& data,
-                               bool allow_anonymous) {
-        return GenericJson("POST", path, data, allow_anonymous);
-    }
+                               bool allow_anonymous);
 
     /**
      * Gets JSON from the specified path.
@@ -40,9 +34,7 @@ public:
      * @param allow_anonymous If true, allow anonymous unauthenticated requests.
      * @return the result of the request.
      */
-    Common::WebResult GetJson(const std::string& path, bool allow_anonymous) {
-        return GenericJson("GET", path, "", allow_anonymous);
-    }
+    Common::WebResult GetJson(const std::string& path, bool allow_anonymous);
 
     /**
      * Deletes JSON to the specified path.
@@ -52,41 +44,11 @@ public:
      * @return the result of the request.
      */
     Common::WebResult DeleteJson(const std::string& path, const std::string& data,
-                                 bool allow_anonymous) {
-        return GenericJson("DELETE", path, data, allow_anonymous);
-    }
+                                 bool allow_anonymous);
 
 private:
-    /// A generic function handles POST, GET and DELETE request together
-    Common::WebResult GenericJson(const std::string& method, const std::string& path,
-                                  const std::string& data, bool allow_anonymous);
-
-    /**
-     * A generic function with explicit authentication method specified
-     * JWT is used if the jwt parameter is not empty
-     * username + token is used if jwt is empty but username and token are not empty
-     * anonymous if all of jwt, username and token are empty
-     */
-    Common::WebResult GenericJson(const std::string& method, const std::string& path,
-                                  const std::string& data, const std::string& jwt = "",
-                                  const std::string& username = "", const std::string& token = "");
-
-    // Retrieve a new JWT from given username and token
-    void UpdateJWT();
-
-    std::string host;
-    std::string username;
-    std::string token;
-    std::string jwt;
-    std::unique_ptr<httplib::Client> cli;
-
-    struct JWTCache {
-        std::mutex mutex;
-        std::string username;
-        std::string token;
-        std::string jwt;
-    };
-    static JWTCache jwt_cache;
+    struct Impl;
+    std::unique_ptr<Impl> impl;
 };
 
 } // namespace WebService
diff --git a/src/yuzu/CMakeLists.txt b/src/yuzu/CMakeLists.txt
index 04464ad5e..f9ca2948e 100644
--- a/src/yuzu/CMakeLists.txt
+++ b/src/yuzu/CMakeLists.txt
@@ -56,6 +56,8 @@ add_executable(yuzu
     main.h
     ui_settings.cpp
     ui_settings.h
+    util/limitable_input_dialog.cpp
+    util/limitable_input_dialog.h
     util/spinbox.cpp
     util/spinbox.h
     util/util.cpp
@@ -82,10 +84,10 @@ set(UIS
 )
 
 file(GLOB COMPAT_LIST
-     ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.qrc
-     ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
-file(GLOB_RECURSE ICONS ${CMAKE_SOURCE_DIR}/dist/icons/*)
-file(GLOB_RECURSE THEMES ${CMAKE_SOURCE_DIR}/dist/qt_themes/*)
+     ${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.qrc
+     ${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
+file(GLOB_RECURSE ICONS ${PROJECT_SOURCE_DIR}/dist/icons/*)
+file(GLOB_RECURSE THEMES ${PROJECT_SOURCE_DIR}/dist/qt_themes/*)
 
 qt5_wrap_ui(UI_HDRS ${UIS})
 
@@ -121,7 +123,7 @@ target_link_libraries(yuzu PRIVATE Boost::boost glad Qt5::OpenGL Qt5::Widgets)
 target_link_libraries(yuzu PRIVATE ${PLATFORM_LIBRARIES} Threads::Threads)
 
 if (YUZU_ENABLE_COMPATIBILITY_REPORTING)
-    add_definitions(-DYUZU_ENABLE_COMPATIBILITY_REPORTING)
+    target_compile_definitions(yuzu PRIVATE -DYUZU_ENABLE_COMPATIBILITY_REPORTING)
 endif()
 
 if (USE_DISCORD_PRESENCE)
diff --git a/src/yuzu/bootmanager.cpp b/src/yuzu/bootmanager.cpp
index 4e4c108ab..39eef8858 100644
--- a/src/yuzu/bootmanager.cpp
+++ b/src/yuzu/bootmanager.cpp
@@ -8,7 +8,6 @@
 
 #include "common/microprofile.h"
 #include "common/scm_rev.h"
-#include "common/string_util.h"
 #include "core/core.h"
 #include "core/frontend/framebuffer_layout.h"
 #include "core/settings.h"
@@ -107,9 +106,9 @@ private:
 GRenderWindow::GRenderWindow(QWidget* parent, EmuThread* emu_thread)
     : QWidget(parent), child(nullptr), emu_thread(emu_thread) {
 
-    std::string window_title = fmt::format("yuzu {} | {}-{}", Common::g_build_name,
-                                           Common::g_scm_branch, Common::g_scm_desc);
-    setWindowTitle(QString::fromStdString(window_title));
+    setWindowTitle(QStringLiteral("yuzu %1 | %2-%3")
+                       .arg(Common::g_build_name, Common::g_scm_branch, Common::g_scm_desc));
+    setAttribute(Qt::WA_AcceptTouchEvents);
 
     InputCommon::Init();
     InputCommon::StartJoystickEventHandler();
@@ -190,11 +189,17 @@ QByteArray GRenderWindow::saveGeometry() {
         return geometry;
 }
 
-qreal GRenderWindow::windowPixelRatio() {
+qreal GRenderWindow::windowPixelRatio() const {
     // windowHandle() might not be accessible until the window is displayed to screen.
     return windowHandle() ? windowHandle()->screen()->devicePixelRatio() : 1.0f;
 }
 
+std::pair<unsigned, unsigned> GRenderWindow::ScaleTouch(const QPointF pos) const {
+    const qreal pixel_ratio = windowPixelRatio();
+    return {static_cast<unsigned>(std::max(std::round(pos.x() * pixel_ratio), qreal{0.0})),
+            static_cast<unsigned>(std::max(std::round(pos.y() * pixel_ratio), qreal{0.0}))};
+}
+
 void GRenderWindow::closeEvent(QCloseEvent* event) {
     emit Closed();
     QWidget::closeEvent(event);
@@ -209,31 +214,81 @@ void GRenderWindow::keyReleaseEvent(QKeyEvent* event) {
 }
 
 void GRenderWindow::mousePressEvent(QMouseEvent* event) {
+    if (event->source() == Qt::MouseEventSynthesizedBySystem)
+        return; // touch input is handled in TouchBeginEvent
+
     auto pos = event->pos();
     if (event->button() == Qt::LeftButton) {
-        qreal pixelRatio = windowPixelRatio();
-        this->TouchPressed(static_cast<unsigned>(pos.x() * pixelRatio),
-                           static_cast<unsigned>(pos.y() * pixelRatio));
+        const auto [x, y] = ScaleTouch(pos);
+        this->TouchPressed(x, y);
     } else if (event->button() == Qt::RightButton) {
         InputCommon::GetMotionEmu()->BeginTilt(pos.x(), pos.y());
     }
 }
 
 void GRenderWindow::mouseMoveEvent(QMouseEvent* event) {
+    if (event->source() == Qt::MouseEventSynthesizedBySystem)
+        return; // touch input is handled in TouchUpdateEvent
+
     auto pos = event->pos();
-    qreal pixelRatio = windowPixelRatio();
-    this->TouchMoved(std::max(static_cast<unsigned>(pos.x() * pixelRatio), 0u),
-                     std::max(static_cast<unsigned>(pos.y() * pixelRatio), 0u));
+    const auto [x, y] = ScaleTouch(pos);
+    this->TouchMoved(x, y);
     InputCommon::GetMotionEmu()->Tilt(pos.x(), pos.y());
 }
 
 void GRenderWindow::mouseReleaseEvent(QMouseEvent* event) {
+    if (event->source() == Qt::MouseEventSynthesizedBySystem)
+        return; // touch input is handled in TouchEndEvent
+
     if (event->button() == Qt::LeftButton)
         this->TouchReleased();
     else if (event->button() == Qt::RightButton)
         InputCommon::GetMotionEmu()->EndTilt();
 }
 
+void GRenderWindow::TouchBeginEvent(const QTouchEvent* event) {
+    // TouchBegin always has exactly one touch point, so take the .first()
+    const auto [x, y] = ScaleTouch(event->touchPoints().first().pos());
+    this->TouchPressed(x, y);
+}
+
+void GRenderWindow::TouchUpdateEvent(const QTouchEvent* event) {
+    QPointF pos;
+    int active_points = 0;
+
+    // average all active touch points
+    for (const auto tp : event->touchPoints()) {
+        if (tp.state() & (Qt::TouchPointPressed | Qt::TouchPointMoved | Qt::TouchPointStationary)) {
+            active_points++;
+            pos += tp.pos();
+        }
+    }
+
+    pos /= active_points;
+
+    const auto [x, y] = ScaleTouch(pos);
+    this->TouchMoved(x, y);
+}
+
+void GRenderWindow::TouchEndEvent() {
+    this->TouchReleased();
+}
+
+bool GRenderWindow::event(QEvent* event) {
+    if (event->type() == QEvent::TouchBegin) {
+        TouchBeginEvent(static_cast<QTouchEvent*>(event));
+        return true;
+    } else if (event->type() == QEvent::TouchUpdate) {
+        TouchUpdateEvent(static_cast<QTouchEvent*>(event));
+        return true;
+    } else if (event->type() == QEvent::TouchEnd || event->type() == QEvent::TouchCancel) {
+        TouchEndEvent();
+        return true;
+    }
+
+    return QWidget::event(event);
+}
+
 void GRenderWindow::focusOutEvent(QFocusEvent* event) {
     QWidget::focusOutEvent(event);
     InputCommon::GetKeyboard()->ReleaseAllKeys();
diff --git a/src/yuzu/bootmanager.h b/src/yuzu/bootmanager.h
index f133bfadf..873985564 100644
--- a/src/yuzu/bootmanager.h
+++ b/src/yuzu/bootmanager.h
@@ -15,6 +15,7 @@
 
 class QKeyEvent;
 class QScreen;
+class QTouchEvent;
 
 class GGLWidgetInternal;
 class GMainWindow;
@@ -119,7 +120,7 @@ public:
     void restoreGeometry(const QByteArray& geometry); // overridden
     QByteArray saveGeometry();                        // overridden
 
-    qreal windowPixelRatio();
+    qreal windowPixelRatio() const;
 
     void closeEvent(QCloseEvent* event) override;
 
@@ -130,6 +131,8 @@ public:
     void mouseMoveEvent(QMouseEvent* event) override;
     void mouseReleaseEvent(QMouseEvent* event) override;
 
+    bool event(QEvent* event) override;
+
     void focusOutEvent(QFocusEvent* event) override;
 
     void OnClientAreaResized(unsigned width, unsigned height);
@@ -148,6 +151,11 @@ signals:
     void Closed();
 
 private:
+    std::pair<unsigned, unsigned> ScaleTouch(const QPointF pos) const;
+    void TouchBeginEvent(const QTouchEvent* event);
+    void TouchUpdateEvent(const QTouchEvent* event);
+    void TouchEndEvent();
+
     void OnMinimalClientAreaChangeRequest(
         const std::pair<unsigned, unsigned>& minimal_size) override;
 
diff --git a/src/yuzu/compatdb.cpp b/src/yuzu/compatdb.cpp
index 91e754274..5f0896f84 100644
--- a/src/yuzu/compatdb.cpp
+++ b/src/yuzu/compatdb.cpp
@@ -5,6 +5,7 @@
 #include <QButtonGroup>
 #include <QMessageBox>
 #include <QPushButton>
+#include <QtConcurrent/qtconcurrentrun.h>
 #include "common/logging/log.h"
 #include "common/telemetry.h"
 #include "core/core.h"
@@ -23,6 +24,8 @@ CompatDB::CompatDB(QWidget* parent)
     connect(ui->radioButton_IntroMenu, &QRadioButton::clicked, this, &CompatDB::EnableNext);
     connect(ui->radioButton_WontBoot, &QRadioButton::clicked, this, &CompatDB::EnableNext);
     connect(button(NextButton), &QPushButton::clicked, this, &CompatDB::Submit);
+    connect(&testcase_watcher, &QFutureWatcher<bool>::finished, this,
+            &CompatDB::OnTestcaseSubmitted);
 }
 
 CompatDB::~CompatDB() = default;
@@ -48,18 +51,38 @@ void CompatDB::Submit() {
         }
         break;
     case CompatDBPage::Final:
+        back();
         LOG_DEBUG(Frontend, "Compatibility Rating: {}", compatibility->checkedId());
         Core::Telemetry().AddField(Telemetry::FieldType::UserFeedback, "Compatibility",
                                    compatibility->checkedId());
-        // older versions of QT don't support the "NoCancelButtonOnLastPage" option, this is a
-        // workaround
+
+        button(NextButton)->setEnabled(false);
+        button(NextButton)->setText(tr("Submitting"));
         button(QWizard::CancelButton)->setVisible(false);
+
+        testcase_watcher.setFuture(QtConcurrent::run(
+            [this]() { return Core::System::GetInstance().TelemetrySession().SubmitTestcase(); }));
         break;
     default:
         LOG_ERROR(Frontend, "Unexpected page: {}", currentId());
     }
 }
 
+void CompatDB::OnTestcaseSubmitted() {
+    if (!testcase_watcher.result()) {
+        QMessageBox::critical(this, tr("Communication error"),
+                              tr("An error occured while sending the Testcase"));
+        button(NextButton)->setEnabled(true);
+        button(NextButton)->setText(tr("Next"));
+        button(QWizard::CancelButton)->setVisible(true);
+    } else {
+        next();
+        // older versions of QT don't support the "NoCancelButtonOnLastPage" option, this is a
+        // workaround
+        button(QWizard::CancelButton)->setVisible(false);
+    }
+}
+
 void CompatDB::EnableNext() {
     button(NextButton)->setEnabled(true);
 }
diff --git a/src/yuzu/compatdb.h b/src/yuzu/compatdb.h
index ca0dd11d6..5381f67f7 100644
--- a/src/yuzu/compatdb.h
+++ b/src/yuzu/compatdb.h
@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <QFutureWatcher>
 #include <QWizard>
 
 namespace Ui {
@@ -19,8 +20,11 @@ public:
     ~CompatDB();
 
 private:
+    QFutureWatcher<bool> testcase_watcher;
+
     std::unique_ptr<Ui::CompatDB> ui;
 
     void Submit();
+    void OnTestcaseSubmitted();
     void EnableNext();
 };
diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp
index 650dd03c0..d4fd60a73 100644
--- a/src/yuzu/configuration/config.cpp
+++ b/src/yuzu/configuration/config.cpp
@@ -4,6 +4,7 @@
 
 #include <QSettings>
 #include "common/file_util.h"
+#include "core/hle/service/acc/profile_manager.h"
 #include "input_common/main.h"
 #include "yuzu/configuration/config.h"
 #include "yuzu/ui_settings.h"
@@ -12,11 +13,16 @@ Config::Config() {
     // TODO: Don't hardcode the path; let the frontend decide where to put the config files.
     qt_config_loc = FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) + "qt-config.ini";
     FileUtil::CreateFullPath(qt_config_loc);
-    qt_config = new QSettings(QString::fromStdString(qt_config_loc), QSettings::IniFormat);
+    qt_config =
+        std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), QSettings::IniFormat);
 
     Reload();
 }
 
+Config::~Config() {
+    Save();
+}
+
 const std::array<int, Settings::NativeButton::NumButtons> Config::default_buttons = {
     Qt::Key_A, Qt::Key_S, Qt::Key_Z,    Qt::Key_X,  Qt::Key_3,     Qt::Key_4,    Qt::Key_Q,
     Qt::Key_W, Qt::Key_1, Qt::Key_2,    Qt::Key_N,  Qt::Key_M,     Qt::Key_F,    Qt::Key_T,
@@ -85,8 +91,8 @@ void Config::ReadValues() {
     Settings::values.resolution_factor = qt_config->value("resolution_factor", 1.0).toFloat();
     Settings::values.use_frame_limit = qt_config->value("use_frame_limit", true).toBool();
     Settings::values.frame_limit = qt_config->value("frame_limit", 100).toInt();
-    Settings::values.use_accurate_framebuffers =
-        qt_config->value("use_accurate_framebuffers", false).toBool();
+    Settings::values.use_accurate_gpu_emulation =
+        qt_config->value("use_accurate_gpu_emulation", false).toBool();
 
     Settings::values.bg_red = qt_config->value("bg_red", 0.0).toFloat();
     Settings::values.bg_green = qt_config->value("bg_green", 0.0).toFloat();
@@ -122,7 +128,11 @@ void Config::ReadValues() {
 
     qt_config->beginGroup("System");
     Settings::values.use_docked_mode = qt_config->value("use_docked_mode", false).toBool();
-    Settings::values.username = qt_config->value("username", "yuzu").toString().toStdString();
+    Settings::values.enable_nfc = qt_config->value("enable_nfc", true).toBool();
+
+    Settings::values.current_user = std::clamp<int>(qt_config->value("current_user", 0).toInt(), 0,
+                                                    Service::Account::MAX_USERS - 1);
+
     Settings::values.language_index = qt_config->value("language_index", 1).toInt();
     qt_config->endGroup();
 
@@ -134,6 +144,7 @@ void Config::ReadValues() {
     qt_config->beginGroup("Debugging");
     Settings::values.use_gdbstub = qt_config->value("use_gdbstub", false).toBool();
     Settings::values.gdbstub_port = qt_config->value("gdbstub_port", 24689).toInt();
+    Settings::values.program_args = qt_config->value("program_args", "").toString().toStdString();
     qt_config->endGroup();
 
     qt_config->beginGroup("WebService");
@@ -232,7 +243,7 @@ void Config::SaveValues() {
     qt_config->setValue("resolution_factor", (double)Settings::values.resolution_factor);
     qt_config->setValue("use_frame_limit", Settings::values.use_frame_limit);
     qt_config->setValue("frame_limit", Settings::values.frame_limit);
-    qt_config->setValue("use_accurate_framebuffers", Settings::values.use_accurate_framebuffers);
+    qt_config->setValue("use_accurate_gpu_emulation", Settings::values.use_accurate_gpu_emulation);
 
     // Cast to double because Qt's written float values are not human-readable
     qt_config->setValue("bg_red", (double)Settings::values.bg_red);
@@ -257,7 +268,9 @@ void Config::SaveValues() {
 
     qt_config->beginGroup("System");
     qt_config->setValue("use_docked_mode", Settings::values.use_docked_mode);
-    qt_config->setValue("username", QString::fromStdString(Settings::values.username));
+    qt_config->setValue("enable_nfc", Settings::values.enable_nfc);
+    qt_config->setValue("current_user", Settings::values.current_user);
+
     qt_config->setValue("language_index", Settings::values.language_index);
     qt_config->endGroup();
 
@@ -269,6 +282,7 @@ void Config::SaveValues() {
     qt_config->beginGroup("Debugging");
     qt_config->setValue("use_gdbstub", Settings::values.use_gdbstub);
     qt_config->setValue("gdbstub_port", Settings::values.gdbstub_port);
+    qt_config->setValue("program_args", QString::fromStdString(Settings::values.program_args));
     qt_config->endGroup();
 
     qt_config->beginGroup("WebService");
@@ -333,9 +347,3 @@ void Config::Reload() {
 void Config::Save() {
     SaveValues();
 }
-
-Config::~Config() {
-    Save();
-
-    delete qt_config;
-}
diff --git a/src/yuzu/configuration/config.h b/src/yuzu/configuration/config.h
index cbf745ea2..9c99c1b75 100644
--- a/src/yuzu/configuration/config.h
+++ b/src/yuzu/configuration/config.h
@@ -5,6 +5,7 @@
 #pragma once
 
 #include <array>
+#include <memory>
 #include <string>
 #include <QVariant>
 #include "core/settings.h"
@@ -12,12 +13,6 @@
 class QSettings;
 
 class Config {
-    QSettings* qt_config;
-    std::string qt_config_loc;
-
-    void ReadValues();
-    void SaveValues();
-
 public:
     Config();
     ~Config();
@@ -27,4 +22,11 @@ public:
 
     static const std::array<int, Settings::NativeButton::NumButtons> default_buttons;
     static const std::array<std::array<int, 5>, Settings::NativeAnalog::NumAnalogs> default_analogs;
+
+private:
+    void ReadValues();
+    void SaveValues();
+
+    std::unique_ptr<QSettings> qt_config;
+    std::string qt_config_loc;
 };
diff --git a/src/yuzu/configuration/configure_debug.cpp b/src/yuzu/configuration/configure_debug.cpp
index 45d84f19a..9e765fc93 100644
--- a/src/yuzu/configuration/configure_debug.cpp
+++ b/src/yuzu/configuration/configure_debug.cpp
@@ -33,6 +33,7 @@ void ConfigureDebug::setConfiguration() {
     ui->toggle_console->setEnabled(!Core::System::GetInstance().IsPoweredOn());
     ui->toggle_console->setChecked(UISettings::values.show_console);
     ui->log_filter_edit->setText(QString::fromStdString(Settings::values.log_filter));
+    ui->homebrew_args_edit->setText(QString::fromStdString(Settings::values.program_args));
 }
 
 void ConfigureDebug::applyConfiguration() {
@@ -40,6 +41,7 @@ void ConfigureDebug::applyConfiguration() {
     Settings::values.gdbstub_port = ui->gdbport_spinbox->value();
     UISettings::values.show_console = ui->toggle_console->isChecked();
     Settings::values.log_filter = ui->log_filter_edit->text().toStdString();
+    Settings::values.program_args = ui->homebrew_args_edit->text().toStdString();
     Debugger::ToggleConsole();
     Log::Filter filter;
     filter.ParseFilterString(Settings::values.log_filter);
diff --git a/src/yuzu/configuration/configure_debug.ui b/src/yuzu/configuration/configure_debug.ui
index 5ae7276bd..ff4987604 100644
--- a/src/yuzu/configuration/configure_debug.ui
+++ b/src/yuzu/configuration/configure_debug.ui
@@ -107,6 +107,29 @@
     </widget>
    </item>
    <item>
+    <widget class="QGroupBox" name="groupBox_3">
+     <property name="title">
+      <string>Homebrew</string>
+     </property>
+     <layout class="QVBoxLayout" name="verticalLayout">
+      <item>
+       <layout class="QHBoxLayout" name="horizontalLayout">
+        <item>
+         <widget class="QLabel" name="label">
+          <property name="text">
+           <string>Arguments String</string>
+          </property>
+         </widget>
+        </item>
+        <item>
+         <widget class="QLineEdit" name="homebrew_args_edit"/>
+        </item>
+       </layout>
+      </item>
+     </layout>
+    </widget>
+   </item>
+   <item>
     <spacer name="verticalSpacer">
      <property name="orientation">
       <enum>Qt::Vertical</enum>
diff --git a/src/yuzu/configuration/configure_general.cpp b/src/yuzu/configuration/configure_general.cpp
index f5db9e55b..537d6e576 100644
--- a/src/yuzu/configuration/configure_general.cpp
+++ b/src/yuzu/configuration/configure_general.cpp
@@ -31,6 +31,7 @@ void ConfigureGeneral::setConfiguration() {
     ui->theme_combobox->setCurrentIndex(ui->theme_combobox->findData(UISettings::values.theme));
     ui->use_cpu_jit->setChecked(Settings::values.use_cpu_jit);
     ui->use_docked_mode->setChecked(Settings::values.use_docked_mode);
+    ui->enable_nfc->setChecked(Settings::values.enable_nfc);
 }
 
 void ConfigureGeneral::PopulateHotkeyList(const HotkeyRegistry& registry) {
@@ -45,4 +46,5 @@ void ConfigureGeneral::applyConfiguration() {
 
     Settings::values.use_cpu_jit = ui->use_cpu_jit->isChecked();
     Settings::values.use_docked_mode = ui->use_docked_mode->isChecked();
+    Settings::values.enable_nfc = ui->enable_nfc->isChecked();
 }
diff --git a/src/yuzu/configuration/configure_general.ui b/src/yuzu/configuration/configure_general.ui
index 1775c4d40..b82fffde8 100644
--- a/src/yuzu/configuration/configure_general.ui
+++ b/src/yuzu/configuration/configure_general.ui
@@ -68,19 +68,26 @@
        <property name="title">
         <string>Emulation</string>
        </property>
-        <layout class="QHBoxLayout" name="EmulationHorizontalLayout">
+       <layout class="QHBoxLayout" name="EmulationHorizontalLayout">
+        <item>
+         <layout class="QVBoxLayout" name="EmulationVerticalLayout">
+          <item>
+           <widget class="QCheckBox" name="use_docked_mode">
+            <property name="text">
+             <string>Enable docked mode</string>
+            </property>
+           </widget>
+          </item>
           <item>
-            <layout class="QVBoxLayout" name="EmulationVerticalLayout">
-              <item>
-                <widget class="QCheckBox" name="use_docked_mode">
-                  <property name="text">
-                    <string>Enable docked mode</string>
-                  </property>
-                </widget>
-              </item>
-            </layout>
+           <widget class="QCheckBox" name="enable_nfc">
+            <property name="text">
+             <string>Enable NFC</string>
+            </property>
+           </widget>
           </item>
-        </layout>
+         </layout>
+        </item>
+       </layout>
       </widget>
      </item>
      <item>
diff --git a/src/yuzu/configuration/configure_graphics.cpp b/src/yuzu/configuration/configure_graphics.cpp
index cd1549462..8290b4384 100644
--- a/src/yuzu/configuration/configure_graphics.cpp
+++ b/src/yuzu/configuration/configure_graphics.cpp
@@ -75,7 +75,7 @@ void ConfigureGraphics::setConfiguration() {
         static_cast<int>(FromResolutionFactor(Settings::values.resolution_factor)));
     ui->toggle_frame_limit->setChecked(Settings::values.use_frame_limit);
     ui->frame_limit->setValue(Settings::values.frame_limit);
-    ui->use_accurate_framebuffers->setChecked(Settings::values.use_accurate_framebuffers);
+    ui->use_accurate_gpu_emulation->setChecked(Settings::values.use_accurate_gpu_emulation);
     bg_color = QColor::fromRgbF(Settings::values.bg_red, Settings::values.bg_green,
                                 Settings::values.bg_blue);
     ui->bg_button->setStyleSheet(
@@ -87,7 +87,7 @@ void ConfigureGraphics::applyConfiguration() {
         ToResolutionFactor(static_cast<Resolution>(ui->resolution_factor_combobox->currentIndex()));
     Settings::values.use_frame_limit = ui->toggle_frame_limit->isChecked();
     Settings::values.frame_limit = ui->frame_limit->value();
-    Settings::values.use_accurate_framebuffers = ui->use_accurate_framebuffers->isChecked();
+    Settings::values.use_accurate_gpu_emulation = ui->use_accurate_gpu_emulation->isChecked();
     Settings::values.bg_red = static_cast<float>(bg_color.redF());
     Settings::values.bg_green = static_cast<float>(bg_color.greenF());
     Settings::values.bg_blue = static_cast<float>(bg_color.blueF());
diff --git a/src/yuzu/configuration/configure_graphics.ui b/src/yuzu/configuration/configure_graphics.ui
index 8fc00af1b..91fcad994 100644
--- a/src/yuzu/configuration/configure_graphics.ui
+++ b/src/yuzu/configuration/configure_graphics.ui
@@ -50,9 +50,9 @@
           </layout>
         </item>
         <item>
-         <widget class="QCheckBox" name="use_accurate_framebuffers">
+         <widget class="QCheckBox" name="use_accurate_gpu_emulation">
           <property name="text">
-           <string>Use accurate framebuffers (slow)</string>
+           <string>Use accurate GPU emulation (slow)</string>
           </property>
          </widget>
         </item>
diff --git a/src/yuzu/configuration/configure_input.cpp b/src/yuzu/configuration/configure_input.cpp
index 473937ea9..42a7beac6 100644
--- a/src/yuzu/configuration/configure_input.cpp
+++ b/src/yuzu/configuration/configure_input.cpp
@@ -5,6 +5,7 @@
 #include <algorithm>
 #include <memory>
 #include <utility>
+#include <QMenu>
 #include <QMessageBox>
 #include <QTimer>
 #include "common/param_package.h"
@@ -128,28 +129,63 @@ ConfigureInput::ConfigureInput(QWidget* parent)
     analog_map_stick = {ui->buttonLStickAnalog, ui->buttonRStickAnalog};
 
     for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; button_id++) {
-        if (button_map[button_id])
-            connect(button_map[button_id], &QPushButton::released, [=]() {
-                handleClick(
-                    button_map[button_id],
-                    [=](const Common::ParamPackage& params) { buttons_param[button_id] = params; },
-                    InputCommon::Polling::DeviceType::Button);
-            });
+        if (!button_map[button_id])
+            continue;
+        button_map[button_id]->setContextMenuPolicy(Qt::CustomContextMenu);
+        connect(button_map[button_id], &QPushButton::released, [=]() {
+            handleClick(
+                button_map[button_id],
+                [=](const Common::ParamPackage& params) { buttons_param[button_id] = params; },
+                InputCommon::Polling::DeviceType::Button);
+        });
+        connect(button_map[button_id], &QPushButton::customContextMenuRequested,
+                [=](const QPoint& menu_location) {
+                    QMenu context_menu;
+                    context_menu.addAction(tr("Clear"), [&] {
+                        buttons_param[button_id].Clear();
+                        button_map[button_id]->setText(tr("[not set]"));
+                    });
+                    context_menu.addAction(tr("Restore Default"), [&] {
+                        buttons_param[button_id] = Common::ParamPackage{
+                            InputCommon::GenerateKeyboardParam(Config::default_buttons[button_id])};
+                        button_map[button_id]->setText(ButtonToText(buttons_param[button_id]));
+                    });
+                    context_menu.exec(button_map[button_id]->mapToGlobal(menu_location));
+                });
     }
 
     for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; analog_id++) {
         for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; sub_button_id++) {
-            if (analog_map_buttons[analog_id][sub_button_id] != nullptr) {
-                connect(analog_map_buttons[analog_id][sub_button_id], &QPushButton::released,
-                        [=]() {
-                            handleClick(analog_map_buttons[analog_id][sub_button_id],
-                                        [=](const Common::ParamPackage& params) {
-                                            SetAnalogButton(params, analogs_param[analog_id],
-                                                            analog_sub_buttons[sub_button_id]);
-                                        },
-                                        InputCommon::Polling::DeviceType::Button);
+            if (!analog_map_buttons[analog_id][sub_button_id])
+                continue;
+            analog_map_buttons[analog_id][sub_button_id]->setContextMenuPolicy(
+                Qt::CustomContextMenu);
+            connect(analog_map_buttons[analog_id][sub_button_id], &QPushButton::released, [=]() {
+                handleClick(analog_map_buttons[analog_id][sub_button_id],
+                            [=](const Common::ParamPackage& params) {
+                                SetAnalogButton(params, analogs_param[analog_id],
+                                                analog_sub_buttons[sub_button_id]);
+                            },
+                            InputCommon::Polling::DeviceType::Button);
+            });
+            connect(analog_map_buttons[analog_id][sub_button_id],
+                    &QPushButton::customContextMenuRequested, [=](const QPoint& menu_location) {
+                        QMenu context_menu;
+                        context_menu.addAction(tr("Clear"), [&] {
+                            analogs_param[analog_id].Erase(analog_sub_buttons[sub_button_id]);
+                            analog_map_buttons[analog_id][sub_button_id]->setText(tr("[not set]"));
                         });
-            }
+                        context_menu.addAction(tr("Restore Default"), [&] {
+                            Common::ParamPackage params{InputCommon::GenerateKeyboardParam(
+                                Config::default_analogs[analog_id][sub_button_id])};
+                            SetAnalogButton(params, analogs_param[analog_id],
+                                            analog_sub_buttons[sub_button_id]);
+                            analog_map_buttons[analog_id][sub_button_id]->setText(AnalogToText(
+                                analogs_param[analog_id], analog_sub_buttons[sub_button_id]));
+                        });
+                        context_menu.exec(analog_map_buttons[analog_id][sub_button_id]->mapToGlobal(
+                            menu_location));
+                    });
         }
         connect(analog_map_stick[analog_id], &QPushButton::released, [=]() {
             QMessageBox::information(this, tr("Information"),
@@ -162,6 +198,7 @@ ConfigureInput::ConfigureInput(QWidget* parent)
         });
     }
 
+    connect(ui->buttonClearAll, &QPushButton::released, [this] { ClearAll(); });
     connect(ui->buttonRestoreDefaults, &QPushButton::released, [this]() { restoreDefaults(); });
 
     timeout_timer->setSingleShot(true);
@@ -215,7 +252,21 @@ void ConfigureInput::restoreDefaults() {
         }
     }
     updateButtonLabels();
-    applyConfiguration();
+}
+
+void ConfigureInput::ClearAll() {
+    for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; button_id++) {
+        if (button_map[button_id] && button_map[button_id]->isEnabled())
+            buttons_param[button_id].Clear();
+    }
+    for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; analog_id++) {
+        for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; sub_button_id++) {
+            if (analog_map_buttons[analog_id][sub_button_id] &&
+                analog_map_buttons[analog_id][sub_button_id]->isEnabled())
+                analogs_param[analog_id].Erase(analog_sub_buttons[sub_button_id]);
+        }
+    }
+    updateButtonLabels();
 }
 
 void ConfigureInput::updateButtonLabels() {
@@ -271,7 +322,7 @@ void ConfigureInput::setPollingResult(const Common::ParamPackage& params, bool a
     }
 
     updateButtonLabels();
-    input_setter = boost::none;
+    input_setter = {};
 }
 
 void ConfigureInput::keyPressEvent(QKeyEvent* event) {
diff --git a/src/yuzu/configuration/configure_input.h b/src/yuzu/configuration/configure_input.h
index a0bef86d5..32c7183f9 100644
--- a/src/yuzu/configuration/configure_input.h
+++ b/src/yuzu/configuration/configure_input.h
@@ -7,11 +7,13 @@
 #include <array>
 #include <functional>
 #include <memory>
+#include <optional>
 #include <string>
 #include <unordered_map>
+
 #include <QKeyEvent>
 #include <QWidget>
-#include <boost/optional.hpp>
+
 #include "common/param_package.h"
 #include "core/settings.h"
 #include "input_common/main.h"
@@ -41,7 +43,7 @@ private:
     std::unique_ptr<QTimer> poll_timer;
 
     /// This will be the the setting function when an input is awaiting configuration.
-    boost::optional<std::function<void(const Common::ParamPackage&)>> input_setter;
+    std::optional<std::function<void(const Common::ParamPackage&)>> input_setter;
 
     std::array<Common::ParamPackage, Settings::NativeButton::NumButtons> buttons_param;
     std::array<Common::ParamPackage, Settings::NativeAnalog::NumAnalogs> analogs_param;
@@ -72,6 +74,9 @@ private:
     void loadConfiguration();
     /// Restore all buttons to their default values.
     void restoreDefaults();
+    /// Clear all input configuration
+    void ClearAll();
+
     /// Update UI to reflect current configuration.
     void updateButtonLabels();
 
diff --git a/src/yuzu/configuration/configure_input.ui b/src/yuzu/configuration/configure_input.ui
index 8bfa5df62..8a019a693 100644
--- a/src/yuzu/configuration/configure_input.ui
+++ b/src/yuzu/configuration/configure_input.ui
@@ -695,6 +695,34 @@ Capture:</string>
       </spacer>
      </item>
      <item>
+      <widget class="QPushButton" name="buttonClearAll">
+       <property name="sizePolicy">
+        <sizepolicy hsizetype="Minimum" vsizetype="Fixed">
+         <horstretch>0</horstretch>
+         <verstretch>0</verstretch>
+        </sizepolicy>
+       </property>
+       <property name="sizeIncrement">
+        <size>
+         <width>0</width>
+         <height>0</height>
+        </size>
+       </property>
+       <property name="baseSize">
+        <size>
+         <width>0</width>
+         <height>0</height>
+        </size>
+       </property>
+       <property name="layoutDirection">
+        <enum>Qt::LeftToRight</enum>
+       </property>
+       <property name="text">
+        <string>Clear All</string>
+       </property>
+      </widget>
+     </item>
+     <item>
       <widget class="QPushButton" name="buttonRestoreDefaults">
        <property name="sizePolicy">
         <sizepolicy hsizetype="Minimum" vsizetype="Fixed">
diff --git a/src/yuzu/configuration/configure_system.cpp b/src/yuzu/configuration/configure_system.cpp
index e9ed9c38f..b4b4a4a56 100644
--- a/src/yuzu/configuration/configure_system.cpp
+++ b/src/yuzu/configuration/configure_system.cpp
@@ -2,14 +2,27 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
+#include <QFileDialog>
+#include <QGraphicsItem>
+#include <QGraphicsScene>
+#include <QHeaderView>
 #include <QMessageBox>
+#include <QStandardItemModel>
+#include <QTreeView>
+#include <QVBoxLayout>
+#include "common/assert.h"
+#include "common/file_util.h"
+#include "common/string_util.h"
 #include "core/core.h"
+#include "core/hle/service/acc/profile_manager.h"
 #include "core/settings.h"
 #include "ui_configure_system.h"
 #include "yuzu/configuration/configure_system.h"
-#include "yuzu/main.h"
+#include "yuzu/util/limitable_input_dialog.h"
 
-static const std::array<int, 12> days_in_month = {{
+namespace {
+constexpr std::array<int, 12> days_in_month = {{
     31,
     29,
     31,
@@ -24,13 +37,108 @@ static const std::array<int, 12> days_in_month = {{
     31,
 }};
 
-ConfigureSystem::ConfigureSystem(QWidget* parent) : QWidget(parent), ui(new Ui::ConfigureSystem) {
+// Same backup JPEG used by acc IProfile::GetImage if no jpeg found
+constexpr std::array<u8, 107> backup_jpeg{
+    0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x02, 0x02,
+    0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04, 0x08, 0x06, 0x06, 0x05,
+    0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a, 0x0c, 0x0f, 0x0c, 0x0a, 0x0b, 0x0e,
+    0x0b, 0x09, 0x09, 0x0d, 0x11, 0x0d, 0x0e, 0x0f, 0x10, 0x10, 0x11, 0x10, 0x0a, 0x0c, 0x12, 0x13,
+    0x12, 0x10, 0x13, 0x0f, 0x10, 0x10, 0x10, 0xff, 0xc9, 0x00, 0x0b, 0x08, 0x00, 0x01, 0x00, 0x01,
+    0x01, 0x01, 0x11, 0x00, 0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08,
+    0x01, 0x01, 0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
+};
+
+QString GetImagePath(Service::Account::UUID uuid) {
+    const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                      "/system/save/8000000000000010/su/avators/" + uuid.FormatSwitch() + ".jpg";
+    return QString::fromStdString(path);
+}
+
+QString GetAccountUsername(const Service::Account::ProfileManager& manager,
+                           Service::Account::UUID uuid) {
+    Service::Account::ProfileBase profile;
+    if (!manager.GetProfileBase(uuid, profile)) {
+        return {};
+    }
+
+    const auto text = Common::StringFromFixedZeroTerminatedBuffer(
+        reinterpret_cast<const char*>(profile.username.data()), profile.username.size());
+    return QString::fromStdString(text);
+}
+
+QString FormatUserEntryText(const QString& username, Service::Account::UUID uuid) {
+    return ConfigureSystem::tr("%1\n%2",
+                               "%1 is the profile username, %2 is the formatted UUID (e.g. "
+                               "00112233-4455-6677-8899-AABBCCDDEEFF))")
+        .arg(username, QString::fromStdString(uuid.FormatSwitch()));
+}
+
+QPixmap GetIcon(Service::Account::UUID uuid) {
+    QPixmap icon{GetImagePath(uuid)};
+
+    if (!icon) {
+        icon.fill(Qt::black);
+        icon.loadFromData(backup_jpeg.data(), static_cast<u32>(backup_jpeg.size()));
+    }
+
+    return icon.scaled(64, 64, Qt::IgnoreAspectRatio, Qt::SmoothTransformation);
+}
+
+QString GetProfileUsernameFromUser(QWidget* parent, const QString& description_text) {
+    return LimitableInputDialog::GetText(parent, ConfigureSystem::tr("Enter Username"),
+                                         description_text, 1,
+                                         static_cast<int>(Service::Account::profile_username_size));
+}
+} // Anonymous namespace
+
+ConfigureSystem::ConfigureSystem(QWidget* parent)
+    : QWidget(parent), ui(new Ui::ConfigureSystem),
+      profile_manager(std::make_unique<Service::Account::ProfileManager>()) {
     ui->setupUi(this);
     connect(ui->combo_birthmonth,
             static_cast<void (QComboBox::*)(int)>(&QComboBox::currentIndexChanged), this,
-            &ConfigureSystem::updateBirthdayComboBox);
+            &ConfigureSystem::UpdateBirthdayComboBox);
     connect(ui->button_regenerate_console_id, &QPushButton::clicked, this,
-            &ConfigureSystem::refreshConsoleID);
+            &ConfigureSystem::RefreshConsoleID);
+
+    layout = new QVBoxLayout;
+    tree_view = new QTreeView;
+    item_model = new QStandardItemModel(tree_view);
+    tree_view->setModel(item_model);
+
+    tree_view->setAlternatingRowColors(true);
+    tree_view->setSelectionMode(QHeaderView::SingleSelection);
+    tree_view->setSelectionBehavior(QHeaderView::SelectRows);
+    tree_view->setVerticalScrollMode(QHeaderView::ScrollPerPixel);
+    tree_view->setHorizontalScrollMode(QHeaderView::ScrollPerPixel);
+    tree_view->setSortingEnabled(true);
+    tree_view->setEditTriggers(QHeaderView::NoEditTriggers);
+    tree_view->setUniformRowHeights(true);
+    tree_view->setIconSize({64, 64});
+    tree_view->setContextMenuPolicy(Qt::NoContextMenu);
+
+    item_model->insertColumns(0, 1);
+    item_model->setHeaderData(0, Qt::Horizontal, "Users");
+
+    // We must register all custom types with the Qt Automoc system so that we are able to use it
+    // with signals/slots. In this case, QList falls under the umbrells of custom types.
+    qRegisterMetaType<QList<QStandardItem*>>("QList<QStandardItem*>");
+
+    layout->setContentsMargins(0, 0, 0, 0);
+    layout->setSpacing(0);
+    layout->addWidget(tree_view);
+
+    ui->scrollArea->setLayout(layout);
+
+    connect(tree_view, &QTreeView::clicked, this, &ConfigureSystem::SelectUser);
+
+    connect(ui->pm_add, &QPushButton::pressed, this, &ConfigureSystem::AddUser);
+    connect(ui->pm_rename, &QPushButton::pressed, this, &ConfigureSystem::RenameUser);
+    connect(ui->pm_remove, &QPushButton::pressed, this, &ConfigureSystem::DeleteUser);
+    connect(ui->pm_set_image, &QPushButton::pressed, this, &ConfigureSystem::SetUserImage);
+
+    scene = new QGraphicsScene;
+    ui->current_user_icon->setScene(scene);
 
     this->setConfiguration();
 }
@@ -39,8 +147,45 @@ ConfigureSystem::~ConfigureSystem() = default;
 
 void ConfigureSystem::setConfiguration() {
     enabled = !Core::System::GetInstance().IsPoweredOn();
-    ui->edit_username->setText(QString::fromStdString(Settings::values.username));
+
     ui->combo_language->setCurrentIndex(Settings::values.language_index);
+
+    item_model->removeRows(0, item_model->rowCount());
+    list_items.clear();
+
+    PopulateUserList();
+    UpdateCurrentUser();
+}
+
+void ConfigureSystem::PopulateUserList() {
+    const auto& profiles = profile_manager->GetAllUsers();
+    for (const auto& user : profiles) {
+        Service::Account::ProfileBase profile;
+        if (!profile_manager->GetProfileBase(user, profile))
+            continue;
+
+        const auto username = Common::StringFromFixedZeroTerminatedBuffer(
+            reinterpret_cast<const char*>(profile.username.data()), profile.username.size());
+
+        list_items.push_back(QList<QStandardItem*>{new QStandardItem{
+            GetIcon(user), FormatUserEntryText(QString::fromStdString(username), user)}});
+    }
+
+    for (const auto& item : list_items)
+        item_model->appendRow(item);
+}
+
+void ConfigureSystem::UpdateCurrentUser() {
+    ui->pm_add->setEnabled(profile_manager->GetUserCount() < Service::Account::MAX_USERS);
+
+    const auto& current_user = profile_manager->GetUser(Settings::values.current_user);
+    ASSERT(current_user);
+    const auto username = GetAccountUsername(*profile_manager, *current_user);
+
+    scene->clear();
+    scene->addPixmap(
+        GetIcon(*current_user).scaled(48, 48, Qt::IgnoreAspectRatio, Qt::SmoothTransformation));
+    ui->current_user_username->setText(username);
 }
 
 void ConfigureSystem::ReadSystemSettings() {}
@@ -48,12 +193,12 @@ void ConfigureSystem::ReadSystemSettings() {}
 void ConfigureSystem::applyConfiguration() {
     if (!enabled)
         return;
-    Settings::values.username = ui->edit_username->text().toStdString();
+
     Settings::values.language_index = ui->combo_language->currentIndex();
     Settings::Apply();
 }
 
-void ConfigureSystem::updateBirthdayComboBox(int birthmonth_index) {
+void ConfigureSystem::UpdateBirthdayComboBox(int birthmonth_index) {
     if (birthmonth_index < 0 || birthmonth_index >= 12)
         return;
 
@@ -78,7 +223,7 @@ void ConfigureSystem::updateBirthdayComboBox(int birthmonth_index) {
     ui->combo_birthday->setCurrentIndex(birthday_index);
 }
 
-void ConfigureSystem::refreshConsoleID() {
+void ConfigureSystem::RefreshConsoleID() {
     QMessageBox::StandardButton reply;
     QString warning_text = tr("This will replace your current virtual Switch with a new one. "
                               "Your current virtual Switch will not be recoverable. "
@@ -92,3 +237,130 @@ void ConfigureSystem::refreshConsoleID() {
     ui->label_console_id->setText(
         tr("Console ID: 0x%1").arg(QString::number(console_id, 16).toUpper()));
 }
+
+void ConfigureSystem::SelectUser(const QModelIndex& index) {
+    Settings::values.current_user =
+        std::clamp<std::size_t>(index.row(), 0, profile_manager->GetUserCount() - 1);
+
+    UpdateCurrentUser();
+
+    ui->pm_remove->setEnabled(profile_manager->GetUserCount() >= 2);
+    ui->pm_rename->setEnabled(true);
+    ui->pm_set_image->setEnabled(true);
+}
+
+void ConfigureSystem::AddUser() {
+    const auto username =
+        GetProfileUsernameFromUser(this, tr("Enter a username for the new user:"));
+    if (username.isEmpty()) {
+        return;
+    }
+
+    const auto uuid = Service::Account::UUID::Generate();
+    profile_manager->CreateNewUser(uuid, username.toStdString());
+
+    item_model->appendRow(new QStandardItem{GetIcon(uuid), FormatUserEntryText(username, uuid)});
+}
+
+void ConfigureSystem::RenameUser() {
+    const auto user = tree_view->currentIndex().row();
+    const auto uuid = profile_manager->GetUser(user);
+    ASSERT(uuid);
+
+    Service::Account::ProfileBase profile;
+    if (!profile_manager->GetProfileBase(*uuid, profile))
+        return;
+
+    const auto new_username = GetProfileUsernameFromUser(this, tr("Enter a new username:"));
+    if (new_username.isEmpty()) {
+        return;
+    }
+
+    const auto username_std = new_username.toStdString();
+    std::fill(profile.username.begin(), profile.username.end(), '\0');
+    std::copy(username_std.begin(), username_std.end(), profile.username.begin());
+
+    profile_manager->SetProfileBase(*uuid, profile);
+
+    item_model->setItem(
+        user, 0,
+        new QStandardItem{GetIcon(*uuid),
+                          FormatUserEntryText(QString::fromStdString(username_std), *uuid)});
+    UpdateCurrentUser();
+}
+
+void ConfigureSystem::DeleteUser() {
+    const auto index = tree_view->currentIndex().row();
+    const auto uuid = profile_manager->GetUser(index);
+    ASSERT(uuid);
+    const auto username = GetAccountUsername(*profile_manager, *uuid);
+
+    const auto confirm = QMessageBox::question(
+        this, tr("Confirm Delete"),
+        tr("You are about to delete user with name \"%1\". Are you sure?").arg(username));
+
+    if (confirm == QMessageBox::No)
+        return;
+
+    if (Settings::values.current_user == tree_view->currentIndex().row())
+        Settings::values.current_user = 0;
+    UpdateCurrentUser();
+
+    if (!profile_manager->RemoveUser(*uuid))
+        return;
+
+    item_model->removeRows(tree_view->currentIndex().row(), 1);
+    tree_view->clearSelection();
+
+    ui->pm_remove->setEnabled(false);
+    ui->pm_rename->setEnabled(false);
+}
+
+void ConfigureSystem::SetUserImage() {
+    const auto index = tree_view->currentIndex().row();
+    const auto uuid = profile_manager->GetUser(index);
+    ASSERT(uuid);
+
+    const auto file = QFileDialog::getOpenFileName(this, tr("Select User Image"), QString(),
+                                                   tr("JPEG Images (*.jpg *.jpeg)"));
+
+    if (file.isEmpty()) {
+        return;
+    }
+
+    const auto image_path = GetImagePath(*uuid);
+    if (QFile::exists(image_path) && !QFile::remove(image_path)) {
+        QMessageBox::warning(
+            this, tr("Error deleting image"),
+            tr("Error occurred attempting to overwrite previous image at: %1.").arg(image_path));
+        return;
+    }
+
+    const auto raw_path = QString::fromStdString(
+        FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + "/system/save/8000000000000010");
+    const QFileInfo raw_info{raw_path};
+    if (raw_info.exists() && !raw_info.isDir() && !QFile::remove(raw_path)) {
+        QMessageBox::warning(this, tr("Error deleting file"),
+                             tr("Unable to delete existing file: %1.").arg(raw_path));
+        return;
+    }
+
+    const QString absolute_dst_path = QFileInfo{image_path}.absolutePath();
+    if (!QDir{raw_path}.mkpath(absolute_dst_path)) {
+        QMessageBox::warning(
+            this, tr("Error creating user image directory"),
+            tr("Unable to create directory %1 for storing user images.").arg(absolute_dst_path));
+        return;
+    }
+
+    if (!QFile::copy(file, image_path)) {
+        QMessageBox::warning(this, tr("Error copying user image"),
+                             tr("Unable to copy image from %1 to %2").arg(file, image_path));
+        return;
+    }
+
+    const auto username = GetAccountUsername(*profile_manager, *uuid);
+    item_model->setItem(index, 0,
+                        new QStandardItem{GetIcon(*uuid), FormatUserEntryText(username, *uuid)});
+    UpdateCurrentUser();
+}
diff --git a/src/yuzu/configuration/configure_system.h b/src/yuzu/configuration/configure_system.h
index f13de17d4..07764e1f7 100644
--- a/src/yuzu/configuration/configure_system.h
+++ b/src/yuzu/configuration/configure_system.h
@@ -5,8 +5,20 @@
 #pragma once
 
 #include <memory>
+
+#include <QList>
 #include <QWidget>
 
+class QGraphicsScene;
+class QStandardItem;
+class QStandardItemModel;
+class QTreeView;
+class QVBoxLayout;
+
+namespace Service::Account {
+class ProfileManager;
+}
+
 namespace Ui {
 class ConfigureSystem;
 }
@@ -16,23 +28,39 @@ class ConfigureSystem : public QWidget {
 
 public:
     explicit ConfigureSystem(QWidget* parent = nullptr);
-    ~ConfigureSystem();
+    ~ConfigureSystem() override;
 
     void applyConfiguration();
     void setConfiguration();
 
-public slots:
-    void updateBirthdayComboBox(int birthmonth_index);
-    void refreshConsoleID();
-
 private:
     void ReadSystemSettings();
 
+    void UpdateBirthdayComboBox(int birthmonth_index);
+    void RefreshConsoleID();
+
+    void PopulateUserList();
+    void UpdateCurrentUser();
+    void SelectUser(const QModelIndex& index);
+    void AddUser();
+    void RenameUser();
+    void DeleteUser();
+    void SetUserImage();
+
+    QVBoxLayout* layout;
+    QTreeView* tree_view;
+    QStandardItemModel* item_model;
+    QGraphicsScene* scene;
+
+    std::vector<QList<QStandardItem*>> list_items;
+
     std::unique_ptr<Ui::ConfigureSystem> ui;
-    bool enabled;
+    bool enabled = false;
+
+    int birthmonth = 0;
+    int birthday = 0;
+    int language_index = 0;
+    int sound_index = 0;
 
-    std::u16string username;
-    int birthmonth, birthday;
-    int language_index;
-    int sound_index;
+    std::unique_ptr<Service::Account::ProfileManager> profile_manager;
 };
diff --git a/src/yuzu/configuration/configure_system.ui b/src/yuzu/configuration/configure_system.ui
index f3f8db038..020b32a37 100644
--- a/src/yuzu/configuration/configure_system.ui
+++ b/src/yuzu/configuration/configure_system.ui
@@ -7,7 +7,7 @@
     <x>0</x>
     <y>0</y>
     <width>360</width>
-    <height>377</height>
+    <height>483</height>
    </rect>
   </property>
   <property name="windowTitle">
@@ -22,34 +22,28 @@
         <string>System Settings</string>
        </property>
        <layout class="QGridLayout" name="gridLayout">
-        <item row="0" column="0">
-         <widget class="QLabel" name="label_username">
+        <item row="1" column="0">
+         <widget class="QLabel" name="label_language">
           <property name="text">
-           <string>Username</string>
+           <string>Language</string>
           </property>
          </widget>
         </item>
-        <item row="0" column="1">
-         <widget class="QLineEdit" name="edit_username">
-          <property name="sizePolicy">
-           <sizepolicy hsizetype="Preferred" vsizetype="Fixed">
-            <horstretch>0</horstretch>
-            <verstretch>0</verstretch>
-           </sizepolicy>
-          </property>
-          <property name="maxLength">
-           <number>32</number>
+        <item row="0" column="0">
+         <widget class="QLabel" name="label_birthday">
+          <property name="text">
+           <string>Birthday</string>
           </property>
          </widget>
         </item>
-        <item row="1" column="0">
-         <widget class="QLabel" name="label_birthday">
+        <item row="3" column="0">
+         <widget class="QLabel" name="label_console_id">
           <property name="text">
-           <string>Birthday</string>
+           <string>Console ID:</string>
           </property>
          </widget>
         </item>
-        <item row="1" column="1">
+        <item row="0" column="1">
          <layout class="QHBoxLayout" name="horizontalLayout_birthday2">
           <item>
            <widget class="QComboBox" name="combo_birthmonth">
@@ -120,14 +114,7 @@
           </item>
          </layout>
         </item>
-        <item row="2" column="0">
-         <widget class="QLabel" name="label_language">
-          <property name="text">
-           <string>Language</string>
-          </property>
-         </widget>
-        </item>
-        <item row="2" column="1">
+        <item row="1" column="1">
          <widget class="QComboBox" name="combo_language">
           <property name="toolTip">
            <string>Note: this can be overridden when region setting is auto-select</string>
@@ -187,31 +174,31 @@
             <string>Russian (Русский)</string>
            </property>
           </item>
-           <item>
-             <property name="text">
-               <string>Taiwanese</string>
-             </property>
-           </item>
-           <item>
-             <property name="text">
-               <string>British English</string>
-             </property>
-           </item>
-           <item>
-             <property name="text">
-               <string>Canadian French</string>
-             </property>
-           </item>
-           <item>
-             <property name="text">
-               <string>Latin American Spanish</string>
-             </property>
-           </item>
-           <item>
-             <property name="text">
-               <string>Simplified Chinese</string>
-             </property>
-           </item>
+          <item>
+           <property name="text">
+            <string>Taiwanese</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>British English</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>Canadian French</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>Latin American Spanish</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>Simplified Chinese</string>
+           </property>
+          </item>
           <item>
            <property name="text">
             <string>Traditional Chinese (正體中文)</string>
@@ -219,14 +206,14 @@
           </item>
          </widget>
         </item>
-        <item row="3" column="0">
+        <item row="2" column="0">
          <widget class="QLabel" name="label_sound">
           <property name="text">
            <string>Sound output mode</string>
           </property>
          </widget>
         </item>
-        <item row="3" column="1">
+        <item row="2" column="1">
          <widget class="QComboBox" name="combo_sound">
           <item>
            <property name="text">
@@ -245,14 +232,7 @@
           </item>
          </widget>
         </item>
-        <item row="4" column="0">
-         <widget class="QLabel" name="label_console_id">
-          <property name="text">
-           <string>Console ID:</string>
-          </property>
-         </widget>
-        </item>
-        <item row="4" column="1">
+        <item row="3" column="1">
          <widget class="QPushButton" name="button_regenerate_console_id">
           <property name="sizePolicy">
            <sizepolicy hsizetype="Fixed" vsizetype="Fixed">
@@ -272,6 +252,143 @@
       </widget>
      </item>
      <item>
+      <widget class="QGroupBox" name="gridGroupBox">
+       <property name="title">
+        <string>Profile Manager</string>
+       </property>
+       <layout class="QGridLayout" name="gridLayout_2">
+        <property name="sizeConstraint">
+         <enum>QLayout::SetNoConstraint</enum>
+        </property>
+        <item row="0" column="0">
+         <layout class="QHBoxLayout" name="horizontalLayout_2">
+          <item>
+           <widget class="QLabel" name="label">
+            <property name="sizePolicy">
+             <sizepolicy hsizetype="Preferred" vsizetype="Minimum">
+              <horstretch>0</horstretch>
+              <verstretch>0</verstretch>
+             </sizepolicy>
+            </property>
+            <property name="text">
+             <string>Current User</string>
+            </property>
+           </widget>
+          </item>
+          <item>
+           <widget class="QGraphicsView" name="current_user_icon">
+            <property name="minimumSize">
+             <size>
+              <width>48</width>
+              <height>48</height>
+             </size>
+            </property>
+            <property name="maximumSize">
+             <size>
+              <width>48</width>
+              <height>48</height>
+             </size>
+            </property>
+            <property name="verticalScrollBarPolicy">
+             <enum>Qt::ScrollBarAlwaysOff</enum>
+            </property>
+            <property name="horizontalScrollBarPolicy">
+             <enum>Qt::ScrollBarAlwaysOff</enum>
+            </property>
+            <property name="interactive">
+             <bool>false</bool>
+            </property>
+           </widget>
+          </item>
+          <item>
+           <widget class="QLabel" name="current_user_username">
+            <property name="sizePolicy">
+             <sizepolicy hsizetype="Preferred" vsizetype="Minimum">
+              <horstretch>0</horstretch>
+              <verstretch>0</verstretch>
+             </sizepolicy>
+            </property>
+            <property name="text">
+             <string>Username</string>
+            </property>
+           </widget>
+          </item>
+         </layout>
+        </item>
+        <item row="1" column="0">
+         <widget class="QScrollArea" name="scrollArea">
+          <property name="sizePolicy">
+           <sizepolicy hsizetype="Expanding" vsizetype="Expanding">
+            <horstretch>0</horstretch>
+            <verstretch>0</verstretch>
+           </sizepolicy>
+          </property>
+          <property name="frameShape">
+           <enum>QFrame::StyledPanel</enum>
+          </property>
+          <property name="widgetResizable">
+           <bool>false</bool>
+          </property>
+         </widget>
+        </item>
+        <item row="2" column="0">
+         <layout class="QHBoxLayout" name="horizontalLayout_3">
+          <item>
+           <widget class="QPushButton" name="pm_set_image">
+            <property name="enabled">
+             <bool>false</bool>
+            </property>
+            <property name="text">
+             <string>Set Image</string>
+            </property>
+           </widget>
+          </item>
+          <item>
+           <spacer name="horizontalSpacer">
+            <property name="orientation">
+             <enum>Qt::Horizontal</enum>
+            </property>
+            <property name="sizeHint" stdset="0">
+             <size>
+              <width>40</width>
+              <height>20</height>
+             </size>
+            </property>
+           </spacer>
+          </item>
+          <item>
+           <widget class="QPushButton" name="pm_add">
+            <property name="text">
+             <string>Add</string>
+            </property>
+           </widget>
+          </item>
+          <item>
+           <widget class="QPushButton" name="pm_rename">
+            <property name="enabled">
+             <bool>false</bool>
+            </property>
+            <property name="text">
+             <string>Rename</string>
+            </property>
+           </widget>
+          </item>
+          <item>
+           <widget class="QPushButton" name="pm_remove">
+            <property name="enabled">
+             <bool>false</bool>
+            </property>
+            <property name="text">
+             <string>Remove</string>
+            </property>
+           </widget>
+          </item>
+         </layout>
+        </item>
+       </layout>
+      </widget>
+     </item>
+     <item>
       <widget class="QLabel" name="label_disable_info">
        <property name="text">
         <string>System settings are available only when game is not running.</string>
@@ -281,19 +398,6 @@
        </property>
       </widget>
      </item>
-     <item>
-      <spacer name="verticalSpacer">
-       <property name="orientation">
-        <enum>Qt::Vertical</enum>
-       </property>
-       <property name="sizeHint" stdset="0">
-        <size>
-         <width>20</width>
-         <height>40</height>
-        </size>
-       </property>
-      </spacer>
-     </item>
     </layout>
    </item>
   </layout>
diff --git a/src/yuzu/debugger/graphics/graphics_breakpoints.cpp b/src/yuzu/debugger/graphics/graphics_breakpoints.cpp
index b5c88f944..67ed0ba6d 100644
--- a/src/yuzu/debugger/graphics/graphics_breakpoints.cpp
+++ b/src/yuzu/debugger/graphics/graphics_breakpoints.cpp
@@ -2,7 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <map>
 #include <QLabel>
 #include <QMetaType>
 #include <QPushButton>
@@ -32,21 +31,8 @@ QVariant BreakPointModel::data(const QModelIndex& index, int role) const {
     switch (role) {
     case Qt::DisplayRole: {
         if (index.column() == 0) {
-            static const std::map<Tegra::DebugContext::Event, QString> map = {
-                {Tegra::DebugContext::Event::MaxwellCommandLoaded, tr("Maxwell command loaded")},
-                {Tegra::DebugContext::Event::MaxwellCommandProcessed,
-                 tr("Maxwell command processed")},
-                {Tegra::DebugContext::Event::IncomingPrimitiveBatch,
-                 tr("Incoming primitive batch")},
-                {Tegra::DebugContext::Event::FinishedPrimitiveBatch,
-                 tr("Finished primitive batch")},
-            };
-
-            DEBUG_ASSERT(map.size() ==
-                         static_cast<std::size_t>(Tegra::DebugContext::Event::NumEvents));
-            return (map.find(event) != map.end()) ? map.at(event) : QString();
+            return DebugContextEventToString(event);
         }
-
         break;
     }
 
@@ -128,6 +114,23 @@ void BreakPointModel::OnResumed() {
     active_breakpoint = context->active_breakpoint;
 }
 
+QString BreakPointModel::DebugContextEventToString(Tegra::DebugContext::Event event) {
+    switch (event) {
+    case Tegra::DebugContext::Event::MaxwellCommandLoaded:
+        return tr("Maxwell command loaded");
+    case Tegra::DebugContext::Event::MaxwellCommandProcessed:
+        return tr("Maxwell command processed");
+    case Tegra::DebugContext::Event::IncomingPrimitiveBatch:
+        return tr("Incoming primitive batch");
+    case Tegra::DebugContext::Event::FinishedPrimitiveBatch:
+        return tr("Finished primitive batch");
+    case Tegra::DebugContext::Event::NumEvents:
+        break;
+    }
+
+    return tr("Unknown debug context event");
+}
+
 GraphicsBreakPointsWidget::GraphicsBreakPointsWidget(
     std::shared_ptr<Tegra::DebugContext> debug_context, QWidget* parent)
     : QDockWidget(tr("Maxwell Breakpoints"), parent), Tegra::DebugContext::BreakPointObserver(
diff --git a/src/yuzu/debugger/graphics/graphics_breakpoints_p.h b/src/yuzu/debugger/graphics/graphics_breakpoints_p.h
index 7112b87e6..fb488e38f 100644
--- a/src/yuzu/debugger/graphics/graphics_breakpoints_p.h
+++ b/src/yuzu/debugger/graphics/graphics_breakpoints_p.h
@@ -29,6 +29,8 @@ public:
     void OnResumed();
 
 private:
+    static QString DebugContextEventToString(Tegra::DebugContext::Event event);
+
     std::weak_ptr<Tegra::DebugContext> context_weak;
     bool at_breakpoint;
     Tegra::DebugContext::Event active_breakpoint;
diff --git a/src/yuzu/debugger/graphics/graphics_surface.cpp b/src/yuzu/debugger/graphics/graphics_surface.cpp
index cbcd5dd5f..707747422 100644
--- a/src/yuzu/debugger/graphics/graphics_surface.cpp
+++ b/src/yuzu/debugger/graphics/graphics_surface.cpp
@@ -382,12 +382,13 @@ void GraphicsSurfaceWidget::OnUpdate() {
     // TODO: Implement a good way to visualize alpha components!
 
     QImage decoded_image(surface_width, surface_height, QImage::Format_ARGB32);
-    boost::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address);
+    std::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address);
 
     // TODO(bunnei): Will not work with BCn formats that swizzle 4x4 tiles.
     // Needs to be fixed if we plan to use this feature more, otherwise we may remove it.
     auto unswizzled_data = Tegra::Texture::UnswizzleTexture(
-        *address, 1, Tegra::Texture::BytesPerPixel(surface_format), surface_width, surface_height);
+        *address, 1, 1, Tegra::Texture::BytesPerPixel(surface_format), surface_width,
+        surface_height, 1U);
 
     auto texture_data = Tegra::Texture::DecodeTexture(unswizzled_data, surface_format,
                                                       surface_width, surface_height);
@@ -443,7 +444,7 @@ void GraphicsSurfaceWidget::SaveSurface() {
             pixmap->save(&file, "PNG");
     } else if (selectedFilter == bin_filter) {
         auto& gpu = Core::System::GetInstance().GPU();
-        boost::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address);
+        std::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address);
 
         const u8* buffer = Memory::GetPointer(*address);
         ASSERT_MSG(buffer != nullptr, "Memory not accessible");
diff --git a/src/yuzu/debugger/wait_tree.cpp b/src/yuzu/debugger/wait_tree.cpp
index 4a09da685..0c831c9f4 100644
--- a/src/yuzu/debugger/wait_tree.cpp
+++ b/src/yuzu/debugger/wait_tree.cpp
@@ -9,8 +9,8 @@
 #include "core/core.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/handle_table.h"
-#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/mutex.h"
+#include "core/hle/kernel/process.h"
 #include "core/hle/kernel/scheduler.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/timer.h"
@@ -66,10 +66,11 @@ std::vector<std::unique_ptr<WaitTreeThread>> WaitTreeItem::MakeThreadItemList()
         }
     };
 
-    add_threads(Core::System::GetInstance().Scheduler(0)->GetThreadList());
-    add_threads(Core::System::GetInstance().Scheduler(1)->GetThreadList());
-    add_threads(Core::System::GetInstance().Scheduler(2)->GetThreadList());
-    add_threads(Core::System::GetInstance().Scheduler(3)->GetThreadList());
+    const auto& system = Core::System::GetInstance();
+    add_threads(system.Scheduler(0).GetThreadList());
+    add_threads(system.Scheduler(1).GetThreadList());
+    add_threads(system.Scheduler(2).GetThreadList());
+    add_threads(system.Scheduler(3).GetThreadList());
 
     return item_list;
 }
@@ -82,7 +83,7 @@ QString WaitTreeText::GetText() const {
 }
 
 WaitTreeMutexInfo::WaitTreeMutexInfo(VAddr mutex_address) : mutex_address(mutex_address) {
-    auto& handle_table = Core::System::GetInstance().Kernel().HandleTable();
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
 
     mutex_value = Memory::Read32(mutex_address);
     owner_handle = static_cast<Kernel::Handle>(mutex_value & Kernel::Mutex::MutexOwnerMask);
diff --git a/src/yuzu/debugger/wait_tree.h b/src/yuzu/debugger/wait_tree.h
index defbf734f..331f89885 100644
--- a/src/yuzu/debugger/wait_tree.h
+++ b/src/yuzu/debugger/wait_tree.h
@@ -11,7 +11,6 @@
 #include <QAbstractItemModel>
 #include <QDockWidget>
 #include <QTreeView>
-#include <boost/container/flat_set.hpp>
 #include "common/common_types.h"
 #include "core/hle/kernel/object.h"
 
diff --git a/src/yuzu/game_list.cpp b/src/yuzu/game_list.cpp
index 67890455a..a5a4aa432 100644
--- a/src/yuzu/game_list.cpp
+++ b/src/yuzu/game_list.cpp
@@ -16,7 +16,6 @@
 #include <fmt/format.h>
 #include "common/common_paths.h"
 #include "common/common_types.h"
-#include "common/file_util.h"
 #include "common/logging/log.h"
 #include "core/file_sys/patch_manager.h"
 #include "yuzu/compatibility_list.h"
@@ -217,11 +216,11 @@ GameList::GameList(FileSys::VirtualFilesystem vfs, GMainWindow* parent)
     tree_view->setContextMenuPolicy(Qt::CustomContextMenu);
 
     item_model->insertColumns(0, COLUMN_COUNT);
-    item_model->setHeaderData(COLUMN_NAME, Qt::Horizontal, "Name");
-    item_model->setHeaderData(COLUMN_COMPATIBILITY, Qt::Horizontal, "Compatibility");
-    item_model->setHeaderData(COLUMN_ADD_ONS, Qt::Horizontal, "Add-ons");
-    item_model->setHeaderData(COLUMN_FILE_TYPE, Qt::Horizontal, "File type");
-    item_model->setHeaderData(COLUMN_SIZE, Qt::Horizontal, "Size");
+    item_model->setHeaderData(COLUMN_NAME, Qt::Horizontal, tr("Name"));
+    item_model->setHeaderData(COLUMN_COMPATIBILITY, Qt::Horizontal, tr("Compatibility"));
+    item_model->setHeaderData(COLUMN_ADD_ONS, Qt::Horizontal, tr("Add-ons"));
+    item_model->setHeaderData(COLUMN_FILE_TYPE, Qt::Horizontal, tr("File type"));
+    item_model->setHeaderData(COLUMN_SIZE, Qt::Horizontal, tr("Size"));
 
     connect(tree_view, &QTreeView::activated, this, &GameList::ValidateEntry);
     connect(tree_view, &QTreeView::customContextMenuRequested, this, &GameList::PopupContextMenu);
@@ -387,9 +386,9 @@ void GameList::LoadCompatibilityList() {
 }
 
 void GameList::PopulateAsync(const QString& dir_path, bool deep_scan) {
-    if (!FileUtil::Exists(dir_path.toStdString()) ||
-        !FileUtil::IsDirectory(dir_path.toStdString())) {
-        LOG_ERROR(Frontend, "Could not find game list folder at {}", dir_path.toLocal8Bit().data());
+    const QFileInfo dir_info{dir_path};
+    if (!dir_info.exists() || !dir_info.isDir()) {
+        LOG_ERROR(Frontend, "Could not find game list folder at {}", dir_path.toStdString());
         search_field->setFilterResult(0, 0);
         return;
     }
diff --git a/src/yuzu/game_list_worker.cpp b/src/yuzu/game_list_worker.cpp
index 1947bdb93..3d865a12d 100644
--- a/src/yuzu/game_list_worker.cpp
+++ b/src/yuzu/game_list_worker.cpp
@@ -27,9 +27,8 @@
 #include "yuzu/ui_settings.h"
 
 namespace {
-void GetMetadataFromControlNCA(const FileSys::PatchManager& patch_manager,
-                               const std::shared_ptr<FileSys::NCA>& nca, std::vector<u8>& icon,
-                               std::string& name) {
+void GetMetadataFromControlNCA(const FileSys::PatchManager& patch_manager, const FileSys::NCA& nca,
+                               std::vector<u8>& icon, std::string& name) {
     auto [nacp, icon_file] = patch_manager.ParseControlNCA(nca);
     if (icon_file != nullptr)
         icon = icon_file->ReadAllBytes();
@@ -57,16 +56,30 @@ QString FormatGameName(const std::string& physical_name) {
     return physical_name_as_qstring;
 }
 
-QString FormatPatchNameVersions(const FileSys::PatchManager& patch_manager, bool updatable = true) {
+QString FormatPatchNameVersions(const FileSys::PatchManager& patch_manager,
+                                Loader::AppLoader& loader, bool updatable = true) {
     QString out;
-    for (const auto& kv : patch_manager.GetPatchVersionNames()) {
-        if (!updatable && kv.first == "Update")
+    FileSys::VirtualFile update_raw;
+    loader.ReadUpdateRaw(update_raw);
+    for (const auto& kv : patch_manager.GetPatchVersionNames(update_raw)) {
+        const bool is_update = kv.first == "Update";
+        if (!updatable && is_update) {
             continue;
+        }
+
+        const QString type = QString::fromStdString(kv.first);
 
         if (kv.second.empty()) {
-            out.append(fmt::format("{}\n", kv.first).c_str());
+            out.append(QStringLiteral("%1\n").arg(type));
         } else {
-            out.append(fmt::format("{} ({})\n", kv.first, kv.second).c_str());
+            auto ver = kv.second;
+
+            // Display container name for packed updates
+            if (is_update && ver == "PACKED") {
+                ver = Loader::GetFileTypeString(loader.GetFileType());
+            }
+
+            out.append(QStringLiteral("%1 (%2)\n").arg(type, QString::fromStdString(ver)));
         }
     }
 
@@ -88,7 +101,7 @@ void GameListWorker::AddInstalledTitlesToGameList() {
                                                           FileSys::ContentRecordType::Program);
 
     for (const auto& game : installed_games) {
-        const auto& file = cache->GetEntryUnparsed(game);
+        const auto file = cache->GetEntryUnparsed(game);
         std::unique_ptr<Loader::AppLoader> loader = Loader::GetLoader(file);
         if (!loader)
             continue;
@@ -99,9 +112,9 @@ void GameListWorker::AddInstalledTitlesToGameList() {
         loader->ReadProgramId(program_id);
 
         const FileSys::PatchManager patch{program_id};
-        const auto& control = cache->GetEntry(game.title_id, FileSys::ContentRecordType::Control);
+        const auto control = cache->GetEntry(game.title_id, FileSys::ContentRecordType::Control);
         if (control != nullptr)
-            GetMetadataFromControlNCA(patch, control, icon, name);
+            GetMetadataFromControlNCA(patch, *control, icon, name);
 
         auto it = FindMatchingCompatibilityEntry(compatibility_list, program_id);
 
@@ -116,7 +129,7 @@ void GameListWorker::AddInstalledTitlesToGameList() {
                 QString::fromStdString(Loader::GetFileTypeString(loader->GetFileType())),
                 program_id),
             new GameListItemCompat(compatibility),
-            new GameListItem(FormatPatchNameVersions(patch)),
+            new GameListItem(FormatPatchNameVersions(patch, *loader)),
             new GameListItem(
                 QString::fromStdString(Loader::GetFileTypeString(loader->GetFileType()))),
             new GameListItemSize(file->GetSize()),
@@ -127,9 +140,10 @@ void GameListWorker::AddInstalledTitlesToGameList() {
                                                        FileSys::ContentRecordType::Control);
 
     for (const auto& entry : control_data) {
-        const auto nca = cache->GetEntry(entry);
-        if (nca != nullptr)
-            nca_control_map.insert_or_assign(entry.title_id, nca);
+        auto nca = cache->GetEntry(entry);
+        if (nca != nullptr) {
+            nca_control_map.insert_or_assign(entry.title_id, std::move(nca));
+        }
     }
 }
 
@@ -145,9 +159,11 @@ void GameListWorker::FillControlMap(const std::string& dir_path) {
         QFileInfo file_info(physical_name.c_str());
         if (!is_dir && file_info.suffix().toStdString() == "nca") {
             auto nca =
-                std::make_shared<FileSys::NCA>(vfs->OpenFile(physical_name, FileSys::Mode::Read));
-            if (nca->GetType() == FileSys::NCAContentType::Control)
-                nca_control_map.insert_or_assign(nca->GetTitleId(), nca);
+                std::make_unique<FileSys::NCA>(vfs->OpenFile(physical_name, FileSys::Mode::Read));
+            if (nca->GetType() == FileSys::NCAContentType::Control) {
+                const u64 title_id = nca->GetTitleId();
+                nca_control_map.insert_or_assign(title_id, std::move(nca));
+            }
         }
         return true;
     };
@@ -188,8 +204,8 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign
                 res2 == Loader::ResultStatus::Success) {
                 // Use from metadata pool.
                 if (nca_control_map.find(program_id) != nca_control_map.end()) {
-                    const auto nca = nca_control_map[program_id];
-                    GetMetadataFromControlNCA(patch, nca, icon, name);
+                    const auto& nca = nca_control_map[program_id];
+                    GetMetadataFromControlNCA(patch, *nca, icon, name);
                 }
             }
 
@@ -206,7 +222,8 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign
                     QString::fromStdString(Loader::GetFileTypeString(loader->GetFileType())),
                     program_id),
                 new GameListItemCompat(compatibility),
-                new GameListItem(FormatPatchNameVersions(patch, loader->IsRomFSUpdatable())),
+                new GameListItem(
+                    FormatPatchNameVersions(patch, *loader, loader->IsRomFSUpdatable())),
                 new GameListItem(
                     QString::fromStdString(Loader::GetFileTypeString(loader->GetFileType()))),
                 new GameListItemSize(FileUtil::GetSize(physical_name)),
diff --git a/src/yuzu/game_list_worker.h b/src/yuzu/game_list_worker.h
index 09d20c42f..0e42d0bde 100644
--- a/src/yuzu/game_list_worker.h
+++ b/src/yuzu/game_list_worker.h
@@ -63,7 +63,7 @@ private:
     void AddFstEntriesToGameList(const std::string& dir_path, unsigned int recursion = 0);
 
     std::shared_ptr<FileSys::VfsFilesystem> vfs;
-    std::map<u64, std::shared_ptr<FileSys::NCA>> nca_control_map;
+    std::map<u64, std::unique_ptr<FileSys::NCA>> nca_control_map;
     QStringList watch_list;
     QString dir_path;
     bool deep_scan;
diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp
index 1d06d6c95..74a44be37 100644
--- a/src/yuzu/main.cpp
+++ b/src/yuzu/main.cpp
@@ -10,6 +10,7 @@
 // VFS includes must be before glad as they will conflict with Windows file api, which uses defines.
 #include "core/file_sys/vfs.h"
 #include "core/file_sys/vfs_real.h"
+#include "core/hle/service/acc/profile_manager.h"
 
 // These are wrappers to avoid the calls to CreateDirectory and CreateFile becuase of the Windows
 // defines.
@@ -29,8 +30,10 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual
 #define QT_NO_OPENGL
 #include <QDesktopWidget>
 #include <QDialogButtonBox>
+#include <QFile>
 #include <QFileDialog>
 #include <QMessageBox>
+#include <QtConcurrent/QtConcurrent>
 #include <QtGui>
 #include <QtWidgets>
 #include <fmt/format.h>
@@ -59,6 +62,8 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual
 #include "core/hle/kernel/process.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/hle/service/filesystem/fsp_ldr.h"
+#include "core/hle/service/nfp/nfp.h"
+#include "core/hle/service/sm/sm.h"
 #include "core/loader/loader.h"
 #include "core/perf_stats.h"
 #include "core/settings.h"
@@ -99,6 +104,8 @@ __declspec(dllexport) int AmdPowerXpressRequestHighPerformance = 1;
 }
 #endif
 
+constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000;
+
 /**
  * "Callouts" are one-time instructional messages shown to the user. In the config settings, there
  * is a bitfield "callout_flags" options, used to track if a message has already been shown to the
@@ -174,8 +181,11 @@ GMainWindow::GMainWindow()
                        .arg(Common::g_build_fullname, Common::g_scm_branch, Common::g_scm_desc));
     show();
 
+    // Gen keys if necessary
+    OnReinitializeKeys(ReinitializeKeyBehavior::NoWarning);
+
     // Necessary to load titles from nand in gamelist.
-    Service::FileSystem::CreateFactories(vfs);
+    Service::FileSystem::CreateFactories(*vfs);
     game_list->LoadCompatibilityList();
     game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan);
 
@@ -421,6 +431,7 @@ void GMainWindow::ConnectMenuEvents() {
     connect(ui.action_Select_SDMC_Directory, &QAction::triggered, this,
             [this] { OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget::SDMC); });
     connect(ui.action_Exit, &QAction::triggered, this, &QMainWindow::close);
+    connect(ui.action_Load_Amiibo, &QAction::triggered, this, &GMainWindow::OnLoadAmiibo);
 
     // Emulation
     connect(ui.action_Start, &QAction::triggered, this, &GMainWindow::OnStartGame);
@@ -446,6 +457,9 @@ void GMainWindow::ConnectMenuEvents() {
     connect(ui.action_Fullscreen, &QAction::triggered, this, &GMainWindow::ToggleFullscreen);
 
     // Help
+    connect(ui.action_Open_yuzu_Folder, &QAction::triggered, this, &GMainWindow::OnOpenYuzuFolder);
+    connect(ui.action_Rederive, &QAction::triggered, this,
+            std::bind(&GMainWindow::OnReinitializeKeys, this, ReinitializeKeyBehavior::Warning));
     connect(ui.action_About, &QAction::triggered, this, &GMainWindow::OnAbout);
 }
 
@@ -488,6 +502,8 @@ QStringList GMainWindow::GetUnsupportedGLExtensions() {
         unsupported_ext.append("ARB_texture_storage");
     if (!GLAD_GL_ARB_multi_bind)
         unsupported_ext.append("ARB_multi_bind");
+    if (!GLAD_GL_ARB_copy_image)
+        unsupported_ext.append("ARB_copy_image");
 
     // Extensions required to support some texture formats.
     if (!GLAD_GL_EXT_texture_compression_s3tc)
@@ -685,6 +701,7 @@ void GMainWindow::ShutdownGame() {
     ui.action_Stop->setEnabled(false);
     ui.action_Restart->setEnabled(false);
     ui.action_Report_Compatibility->setEnabled(false);
+    ui.action_Load_Amiibo->setEnabled(false);
     render_window->hide();
     game_list->show();
     game_list->setFilterFocus();
@@ -746,12 +763,43 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target
         open_target = "Save Data";
         const std::string nand_dir = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir);
         ASSERT(program_id != 0);
-        // TODO(tech4me): Update this to work with arbitrary user profile
-        // Refer to core/hle/service/acc/profile_manager.cpp ProfileManager constructor
-        constexpr u128 user_id = {1, 0};
+
+        Service::Account::ProfileManager manager{};
+        const auto user_ids = manager.GetAllUsers();
+        QStringList list;
+        for (const auto& user_id : user_ids) {
+            if (user_id == Service::Account::UUID{})
+                continue;
+            Service::Account::ProfileBase base;
+            if (!manager.GetProfileBase(user_id, base))
+                continue;
+
+            list.push_back(QString::fromStdString(Common::StringFromFixedZeroTerminatedBuffer(
+                reinterpret_cast<const char*>(base.username.data()), base.username.size())));
+        }
+
+        bool ok = false;
+        const auto index_string =
+            QInputDialog::getItem(this, tr("Select User"),
+                                  tr("Please select the user's save data you would like to open."),
+                                  list, Settings::values.current_user, false, &ok);
+        if (!ok)
+            return;
+
+        const auto index = list.indexOf(index_string);
+        ASSERT(index != -1 && index < 8);
+
+        const auto user_id = manager.GetUser(index);
+        ASSERT(user_id);
         path = nand_dir + FileSys::SaveDataFactory::GetFullPath(FileSys::SaveDataSpaceId::NandUser,
                                                                 FileSys::SaveDataType::SaveData,
-                                                                program_id, user_id, 0);
+                                                                program_id, user_id->uuid, 0);
+
+        if (!FileUtil::Exists(path)) {
+            FileUtil::CreateFullPath(path);
+            FileUtil::CreateDir(path);
+        }
+
         break;
     }
     case GameListOpenTarget::ModData: {
@@ -818,14 +866,10 @@ static bool RomFSRawCopy(QProgressDialog& dialog, const FileSys::VirtualDir& src
 }
 
 void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_path) {
-    const auto path = fmt::format("{}{:016X}/romfs",
-                                  FileUtil::GetUserPath(FileUtil::UserPath::DumpDir), program_id);
-
-    const auto failed = [this, &path] {
+    const auto failed = [this] {
         QMessageBox::warning(this, tr("RomFS Extraction Failed!"),
                              tr("There was an error copying the RomFS files or the user "
                                 "cancelled the operation."));
-        vfs->DeleteDirectory(path);
     };
 
     const auto loader = Loader::GetLoader(vfs->OpenFile(game_path, FileSys::Mode::Read));
@@ -840,10 +884,24 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa
         return;
     }
 
-    const auto romfs =
-        loader->IsRomFSUpdatable()
-            ? FileSys::PatchManager(program_id).PatchRomFS(file, loader->ReadRomFSIVFCOffset())
-            : file;
+    const auto installed = Service::FileSystem::GetUnionContents();
+    auto romfs_title_id = SelectRomFSDumpTarget(*installed, program_id);
+
+    if (!romfs_title_id) {
+        failed();
+        return;
+    }
+
+    const auto path = fmt::format(
+        "{}{:016X}/romfs", FileUtil::GetUserPath(FileUtil::UserPath::DumpDir), *romfs_title_id);
+
+    FileSys::VirtualFile romfs;
+
+    if (*romfs_title_id == program_id) {
+        romfs = file;
+    } else {
+        romfs = installed->GetEntry(*romfs_title_id, FileSys::ContentRecordType::Data)->GetRomFS();
+    }
 
     const auto extracted = FileSys::ExtractRomFS(romfs, FileSys::RomFSExtractionType::Full);
     if (extracted == nullptr) {
@@ -855,6 +913,7 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa
 
     if (out == nullptr) {
         failed();
+        vfs->DeleteDirectory(path);
         return;
     }
 
@@ -865,8 +924,11 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa
            "files into the new directory while <br>skeleton will only create the directory "
            "structure."),
         {"Full", "Skeleton"}, 0, false, &ok);
-    if (!ok)
+    if (!ok) {
         failed();
+        vfs->DeleteDirectory(path);
+        return;
+    }
 
     const auto full = res == "Full";
     const auto entry_size = CalculateRomFSEntrySize(extracted, full);
@@ -883,6 +945,7 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa
     } else {
         progress.close();
         failed();
+        vfs->DeleteDirectory(path);
     }
 }
 
@@ -903,22 +966,20 @@ void GMainWindow::OnGameListNavigateToGamedbEntry(u64 program_id,
 }
 
 void GMainWindow::OnMenuLoadFile() {
-    QString extensions;
-    for (const auto& piece : game_list->supported_file_extensions)
-        extensions += "*." + piece + " ";
-
-    extensions += "main ";
-
-    QString file_filter = tr("Switch Executable") + " (" + extensions + ")";
-    file_filter += ";;" + tr("All Files (*.*)");
-
-    QString filename = QFileDialog::getOpenFileName(this, tr("Load File"),
-                                                    UISettings::values.roms_path, file_filter);
-    if (!filename.isEmpty()) {
-        UISettings::values.roms_path = QFileInfo(filename).path();
+    const QString extensions =
+        QString("*.").append(GameList::supported_file_extensions.join(" *.")).append(" main");
+    const QString file_filter = tr("Switch Executable (%1);;All Files (*.*)",
+                                   "%1 is an identifier for the Switch executable file extensions.")
+                                    .arg(extensions);
+    const QString filename = QFileDialog::getOpenFileName(
+        this, tr("Load File"), UISettings::values.roms_path, file_filter);
 
-        BootGame(filename);
+    if (filename.isEmpty()) {
+        return;
     }
+
+    UISettings::values.roms_path = QFileInfo(filename).path();
+    BootGame(filename);
 }
 
 void GMainWindow::OnMenuLoadFolder() {
@@ -1134,7 +1195,7 @@ void GMainWindow::OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget target)
         FileUtil::GetUserPath(target == EmulatedDirectoryTarget::SDMC ? FileUtil::UserPath::SDMCDir
                                                                       : FileUtil::UserPath::NANDDir,
                               dir_path.toStdString());
-        Service::FileSystem::CreateFactories(vfs);
+        Service::FileSystem::CreateFactories(*vfs);
         game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan);
     }
 }
@@ -1171,6 +1232,7 @@ void GMainWindow::OnStartGame() {
     ui.action_Report_Compatibility->setEnabled(true);
 
     discord_rpc->Update();
+    ui.action_Load_Amiibo->setEnabled(true);
 }
 
 void GMainWindow::OnPauseGame() {
@@ -1275,6 +1337,52 @@ void GMainWindow::OnConfigure() {
     }
 }
 
+void GMainWindow::OnLoadAmiibo() {
+    const QString extensions{"*.bin"};
+    const QString file_filter = tr("Amiibo File (%1);; All Files (*.*)").arg(extensions);
+    const QString filename = QFileDialog::getOpenFileName(this, tr("Load Amiibo"), "", file_filter);
+
+    if (filename.isEmpty()) {
+        return;
+    }
+
+    Core::System& system{Core::System::GetInstance()};
+    Service::SM::ServiceManager& sm = system.ServiceManager();
+    auto nfc = sm.GetService<Service::NFP::Module::Interface>("nfp:user");
+    if (nfc == nullptr) {
+        return;
+    }
+
+    QFile nfc_file{filename};
+    if (!nfc_file.open(QIODevice::ReadOnly)) {
+        QMessageBox::warning(this, tr("Error opening Amiibo data file"),
+                             tr("Unable to open Amiibo file \"%1\" for reading.").arg(filename));
+        return;
+    }
+
+    const u64 nfc_file_size = nfc_file.size();
+    std::vector<u8> buffer(nfc_file_size);
+    const u64 read_size = nfc_file.read(reinterpret_cast<char*>(buffer.data()), nfc_file_size);
+    if (nfc_file_size != read_size) {
+        QMessageBox::warning(this, tr("Error reading Amiibo data file"),
+                             tr("Unable to fully read Amiibo data. Expected to read %1 bytes, but "
+                                "was only able to read %2 bytes.")
+                                 .arg(nfc_file_size)
+                                 .arg(read_size));
+        return;
+    }
+
+    if (!nfc->LoadAmiibo(buffer)) {
+        QMessageBox::warning(this, tr("Error loading Amiibo data"),
+                             tr("Unable to load Amiibo data."));
+    }
+}
+
+void GMainWindow::OnOpenYuzuFolder() {
+    QDesktopServices::openUrl(QUrl::fromLocalFile(
+        QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::UserDir))));
+}
+
 void GMainWindow::OnAbout() {
     AboutDialog aboutDialog(this);
     aboutDialog.exec();
@@ -1315,15 +1423,17 @@ void GMainWindow::UpdateStatusBar() {
 void GMainWindow::OnCoreError(Core::System::ResultStatus result, std::string details) {
     QMessageBox::StandardButton answer;
     QString status_message;
-    const QString common_message = tr(
-        "The game you are trying to load requires additional files from your Switch to be dumped "
-        "before playing.<br/><br/>For more information on dumping these files, please see the "
-        "following wiki page: <a "
-        "href='https://yuzu-emu.org/wiki/"
-        "dumping-system-archives-and-the-shared-fonts-from-a-switch-console/'>Dumping System "
-        "Archives and the Shared Fonts from a Switch Console</a>.<br/><br/>Would you like to quit "
-        "back to the game list? Continuing emulation may result in crashes, corrupted save "
-        "data, or other bugs.");
+    const QString common_message =
+        tr("The game you are trying to load requires additional files from your Switch to be "
+           "dumped "
+           "before playing.<br/><br/>For more information on dumping these files, please see the "
+           "following wiki page: <a "
+           "href='https://yuzu-emu.org/wiki/"
+           "dumping-system-archives-and-the-shared-fonts-from-a-switch-console/'>Dumping System "
+           "Archives and the Shared Fonts from a Switch Console</a>.<br/><br/>Would you like to "
+           "quit "
+           "back to the game list? Continuing emulation may result in crashes, corrupted save "
+           "data, or other bugs.");
     switch (result) {
     case Core::System::ResultStatus::ErrorSystemFiles: {
         QString message = "yuzu was unable to locate a Switch system archive";
@@ -1354,9 +1464,12 @@ void GMainWindow::OnCoreError(Core::System::ResultStatus result, std::string det
             this, tr("Fatal Error"),
             tr("yuzu has encountered a fatal error, please see the log for more details. "
                "For more information on accessing the log, please see the following page: "
-               "<a href='https://community.citra-emu.org/t/how-to-upload-the-log-file/296'>How to "
-               "Upload the Log File</a>.<br/><br/>Would you like to quit back to the game list? "
-               "Continuing emulation may result in crashes, corrupted save data, or other bugs."),
+               "<a href='https://community.citra-emu.org/t/how-to-upload-the-log-file/296'>How "
+               "to "
+               "Upload the Log File</a>.<br/><br/>Would you like to quit back to the game "
+               "list? "
+               "Continuing emulation may result in crashes, corrupted save data, or other "
+               "bugs."),
             QMessageBox::Yes | QMessageBox::No, QMessageBox::No);
         status_message = "Fatal Error encountered";
         break;
@@ -1376,6 +1489,122 @@ void GMainWindow::OnCoreError(Core::System::ResultStatus result, std::string det
     }
 }
 
+void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) {
+    if (behavior == ReinitializeKeyBehavior::Warning) {
+        const auto res = QMessageBox::information(
+            this, tr("Confirm Key Rederivation"),
+            tr("You are about to force rederive all of your keys. \nIf you do not know what this "
+               "means or what you are doing, \nthis is a potentially destructive action. \nPlease "
+               "make "
+               "sure this is what you want \nand optionally make backups.\n\nThis will delete your "
+               "autogenerated key files and re-run the key derivation module."),
+            QMessageBox::StandardButtons{QMessageBox::Ok, QMessageBox::Cancel});
+
+        if (res == QMessageBox::Cancel)
+            return;
+
+        FileUtil::Delete(FileUtil::GetUserPath(FileUtil::UserPath::KeysDir) +
+                         "prod.keys_autogenerated");
+        FileUtil::Delete(FileUtil::GetUserPath(FileUtil::UserPath::KeysDir) +
+                         "console.keys_autogenerated");
+        FileUtil::Delete(FileUtil::GetUserPath(FileUtil::UserPath::KeysDir) +
+                         "title.keys_autogenerated");
+    }
+
+    Core::Crypto::KeyManager keys{};
+    if (keys.BaseDeriveNecessary()) {
+        Core::Crypto::PartitionDataManager pdm{vfs->OpenDirectory(
+            FileUtil::GetUserPath(FileUtil::UserPath::SysDataDir), FileSys::Mode::Read)};
+
+        const auto function = [this, &keys, &pdm] {
+            keys.PopulateFromPartitionData(pdm);
+            Service::FileSystem::CreateFactories(*vfs);
+            keys.DeriveETicket(pdm);
+        };
+
+        QString errors;
+
+        if (!pdm.HasFuses())
+            errors += tr("- Missing fuses - Cannot derive SBK\n");
+        if (!pdm.HasBoot0())
+            errors += tr("- Missing BOOT0 - Cannot derive master keys\n");
+        if (!pdm.HasPackage2())
+            errors += tr("- Missing BCPKG2-1-Normal-Main - Cannot derive general keys\n");
+        if (!pdm.HasProdInfo())
+            errors += tr("- Missing PRODINFO - Cannot derive title keys\n");
+
+        if (!errors.isEmpty()) {
+
+            QMessageBox::warning(
+                this, tr("Warning Missing Derivation Components"),
+                tr("The following are missing from your configuration that may hinder key "
+                   "derivation. It will be attempted but may not complete.<br><br>") +
+                    errors +
+                    tr("<br><br>You can get all of these and dump all of your games easily by "
+                       "following <a href='https://yuzu-emu.org/help/quickstart/'>the "
+                       "quickstart guide</a>. Alternatively, you can use another method of dumping "
+                       "to obtain all of your keys."));
+        }
+
+        QProgressDialog prog;
+        prog.setRange(0, 0);
+        prog.setLabelText(tr("Deriving keys...\nThis may take up to a minute depending \non your "
+                             "system's performance."));
+        prog.setWindowTitle(tr("Deriving Keys"));
+
+        prog.show();
+
+        auto future = QtConcurrent::run(function);
+        while (!future.isFinished()) {
+            QCoreApplication::processEvents();
+        }
+
+        prog.close();
+    }
+
+    Service::FileSystem::CreateFactories(*vfs);
+
+    if (behavior == ReinitializeKeyBehavior::Warning) {
+        game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan);
+    }
+}
+
+std::optional<u64> GMainWindow::SelectRomFSDumpTarget(
+    const FileSys::RegisteredCacheUnion& installed, u64 program_id) {
+    const auto dlc_entries =
+        installed.ListEntriesFilter(FileSys::TitleType::AOC, FileSys::ContentRecordType::Data);
+    std::vector<FileSys::RegisteredCacheEntry> dlc_match;
+    dlc_match.reserve(dlc_entries.size());
+    std::copy_if(dlc_entries.begin(), dlc_entries.end(), std::back_inserter(dlc_match),
+                 [&program_id, &installed](const FileSys::RegisteredCacheEntry& entry) {
+                     return (entry.title_id & DLC_BASE_TITLE_ID_MASK) == program_id &&
+                            installed.GetEntry(entry)->GetStatus() == Loader::ResultStatus::Success;
+                 });
+
+    std::vector<u64> romfs_tids;
+    romfs_tids.push_back(program_id);
+    for (const auto& entry : dlc_match)
+        romfs_tids.push_back(entry.title_id);
+
+    if (romfs_tids.size() > 1) {
+        QStringList list{"Base"};
+        for (std::size_t i = 1; i < romfs_tids.size(); ++i)
+            list.push_back(QStringLiteral("DLC %1").arg(romfs_tids[i] & 0x7FF));
+
+        bool ok;
+        const auto res = QInputDialog::getItem(
+            this, tr("Select RomFS Dump Target"),
+            tr("Please select which RomFS you would like to dump."), list, 0, false, &ok);
+        if (!ok) {
+            return {};
+        }
+
+        return romfs_tids[list.indexOf(res)];
+    }
+
+    return program_id;
+}
+
 bool GMainWindow::ConfirmClose() {
     if (emu_thread == nullptr || !UISettings::values.confirm_before_closing)
         return true;
@@ -1484,7 +1713,7 @@ void GMainWindow::UpdateUITheme() {
     emit UpdateThemedIcons();
 }
 
-void GMainWindow::SetDiscordEnabled(bool state) {
+void GMainWindow::SetDiscordEnabled([[maybe_unused]] bool state) {
 #ifdef USE_DISCORD_PRESENCE
     if (state) {
         discord_rpc = std::make_unique<DiscordRPC::DiscordImpl>();
diff --git a/src/yuzu/main.h b/src/yuzu/main.h
index fe0e9a50a..929250e8c 100644
--- a/src/yuzu/main.h
+++ b/src/yuzu/main.h
@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <optional>
 #include <unordered_map>
 
 #include <QMainWindow>
@@ -29,8 +30,9 @@ class WaitTreeWidget;
 enum class GameListOpenTarget;
 
 namespace FileSys {
+class RegisteredCacheUnion;
 class VfsFilesystem;
-}
+} // namespace FileSys
 
 namespace Tegra {
 class DebugContext;
@@ -41,6 +43,11 @@ enum class EmulatedDirectoryTarget {
     SDMC,
 };
 
+enum class ReinitializeKeyBehavior {
+    NoWarning,
+    Warning,
+};
+
 namespace DiscordRPC {
 class DiscordInterface;
 }
@@ -159,6 +166,8 @@ private slots:
     void OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget target);
     void OnMenuRecentFile();
     void OnConfigure();
+    void OnLoadAmiibo();
+    void OnOpenYuzuFolder();
     void OnAbout();
     void OnToggleFilterBar();
     void OnDisplayTitleBars(bool);
@@ -167,8 +176,10 @@ private slots:
     void HideFullscreen();
     void ToggleWindowMode();
     void OnCoreError(Core::System::ResultStatus, std::string);
+    void OnReinitializeKeys(ReinitializeKeyBehavior behavior);
 
 private:
+    std::optional<u64> SelectRomFSDumpTarget(const FileSys::RegisteredCacheUnion&, u64 program_id);
     void UpdateStatusBar();
 
     Ui::MainWindow ui;
diff --git a/src/yuzu/main.ui b/src/yuzu/main.ui
index cb1664b21..28cf269e7 100644
--- a/src/yuzu/main.ui
+++ b/src/yuzu/main.ui
@@ -57,8 +57,8 @@
       <string>Recent Files</string>
      </property>
     </widget>
-     <addaction name="action_Install_File_NAND" />
-     <addaction name="separator"/>
+    <addaction name="action_Install_File_NAND"/>
+    <addaction name="separator"/>
     <addaction name="action_Load_File"/>
     <addaction name="action_Load_Folder"/>
     <addaction name="separator"/>
@@ -68,6 +68,8 @@
     <addaction name="action_Select_NAND_Directory"/>
     <addaction name="action_Select_SDMC_Directory"/>
     <addaction name="separator"/>
+    <addaction name="action_Load_Amiibo"/>
+    <addaction name="separator"/>
     <addaction name="action_Exit"/>
    </widget>
    <widget class="QMenu" name="menu_Emulation">
@@ -97,24 +99,35 @@
     <addaction name="action_Show_Status_Bar"/>
     <addaction name="menu_View_Debugging"/>
    </widget>
+   <widget class ="QMenu" name="menu_Tools">
+    <property name="title">
+     <string>Tools</string>
+    </property>
+    <addaction name="action_Rederive" />
+   </widget>
    <widget class="QMenu" name="menu_Help">
     <property name="title">
      <string>&amp;Help</string>
     </property>
     <addaction name="action_Report_Compatibility"/>
+    <addaction name="action_Open_yuzu_Folder" />
     <addaction name="separator"/>
     <addaction name="action_About"/>
    </widget>
    <addaction name="menu_File"/>
    <addaction name="menu_Emulation"/>
    <addaction name="menu_View"/>
+   <addaction name="menu_Tools" />
    <addaction name="menu_Help"/>
   </widget>
-   <action name="action_Install_File_NAND">
-     <property name="text">
-       <string>Install File to NAND...</string>
-     </property>
-   </action>
+  <action name="action_Install_File_NAND">
+   <property name="enabled">
+    <bool>true</bool>
+   </property>
+   <property name="text">
+    <string>Install File to NAND...</string>
+   </property>
+  </action>
   <action name="action_Load_File">
    <property name="text">
     <string>Load File...</string>
@@ -159,6 +172,11 @@
     <string>&amp;Stop</string>
    </property>
   </action>
+   <action name="action_Rederive">
+     <property name="text">
+       <string>Reinitialize keys...</string>
+     </property>
+   </action>
   <action name="action_About">
    <property name="text">
     <string>About yuzu</string>
@@ -241,6 +259,14 @@
        <string>Restart</string>
      </property>
    </action>
+  <action name="action_Load_Amiibo">
+    <property name="enabled">
+      <bool>false</bool>
+    </property>
+    <property name="text">
+      <string>Load Amiibo...</string>
+    </property>
+  </action>
    <action name="action_Report_Compatibility">
      <property name="enabled">
        <bool>false</bool>
@@ -252,6 +278,11 @@
        <bool>false</bool>
      </property>
    </action>
+   <action name="action_Open_yuzu_Folder">
+     <property name="text">
+       <string>Open yuzu Folder</string>
+     </property>
+   </action>
   </widget>
  <resources/>
  <connections/>
diff --git a/src/yuzu/util/limitable_input_dialog.cpp b/src/yuzu/util/limitable_input_dialog.cpp
new file mode 100644
index 000000000..edd78e579
--- /dev/null
+++ b/src/yuzu/util/limitable_input_dialog.cpp
@@ -0,0 +1,59 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <QDialogButtonBox>
+#include <QLabel>
+#include <QLineEdit>
+#include <QPushButton>
+#include <QVBoxLayout>
+#include "yuzu/util/limitable_input_dialog.h"
+
+LimitableInputDialog::LimitableInputDialog(QWidget* parent) : QDialog{parent} {
+    CreateUI();
+    ConnectEvents();
+}
+
+LimitableInputDialog::~LimitableInputDialog() = default;
+
+void LimitableInputDialog::CreateUI() {
+    setWindowFlags(windowFlags() & ~Qt::WindowContextHelpButtonHint);
+
+    text_label = new QLabel(this);
+    text_entry = new QLineEdit(this);
+    buttons = new QDialogButtonBox(QDialogButtonBox::Ok | QDialogButtonBox::Cancel, this);
+
+    auto* const layout = new QVBoxLayout;
+    layout->addWidget(text_label);
+    layout->addWidget(text_entry);
+    layout->addWidget(buttons);
+
+    setLayout(layout);
+}
+
+void LimitableInputDialog::ConnectEvents() {
+    connect(buttons, &QDialogButtonBox::accepted, this, &QDialog::accept);
+    connect(buttons, &QDialogButtonBox::rejected, this, &QDialog::reject);
+}
+
+QString LimitableInputDialog::GetText(QWidget* parent, const QString& title, const QString& text,
+                                      int min_character_limit, int max_character_limit) {
+    Q_ASSERT(min_character_limit <= max_character_limit);
+
+    LimitableInputDialog dialog{parent};
+    dialog.setWindowTitle(title);
+    dialog.text_label->setText(text);
+    dialog.text_entry->setMaxLength(max_character_limit);
+
+    auto* const ok_button = dialog.buttons->button(QDialogButtonBox::Ok);
+    ok_button->setEnabled(false);
+    connect(dialog.text_entry, &QLineEdit::textEdited, [&](const QString& new_text) {
+        ok_button->setEnabled(new_text.length() >= min_character_limit);
+    });
+
+    if (dialog.exec() != QDialog::Accepted) {
+        return {};
+    }
+
+    return dialog.text_entry->text();
+}
diff --git a/src/yuzu/util/limitable_input_dialog.h b/src/yuzu/util/limitable_input_dialog.h
new file mode 100644
index 000000000..164ad7301
--- /dev/null
+++ b/src/yuzu/util/limitable_input_dialog.h
@@ -0,0 +1,31 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <QDialog>
+
+class QDialogButtonBox;
+class QLabel;
+class QLineEdit;
+
+/// A QDialog that functions similarly to QInputDialog, however, it allows
+/// restricting the minimum and total number of characters that can be entered.
+class LimitableInputDialog final : public QDialog {
+    Q_OBJECT
+public:
+    explicit LimitableInputDialog(QWidget* parent = nullptr);
+    ~LimitableInputDialog() override;
+
+    static QString GetText(QWidget* parent, const QString& title, const QString& text,
+                           int min_character_limit, int max_character_limit);
+
+private:
+    void CreateUI();
+    void ConnectEvents();
+
+    QLabel* text_label;
+    QLineEdit* text_entry;
+    QDialogButtonBox* buttons;
+};
diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp
index 9d934e220..b456266a6 100644
--- a/src/yuzu_cmd/config.cpp
+++ b/src/yuzu_cmd/config.cpp
@@ -8,6 +8,7 @@
 #include "common/file_util.h"
 #include "common/logging/log.h"
 #include "common/param_package.h"
+#include "core/hle/service/acc/profile_manager.h"
 #include "core/settings.h"
 #include "input_common/main.h"
 #include "yuzu_cmd/config.h"
@@ -99,8 +100,8 @@ void Config::ReadValues() {
     Settings::values.use_frame_limit = sdl2_config->GetBoolean("Renderer", "use_frame_limit", true);
     Settings::values.frame_limit =
         static_cast<u16>(sdl2_config->GetInteger("Renderer", "frame_limit", 100));
-    Settings::values.use_accurate_framebuffers =
-        sdl2_config->GetBoolean("Renderer", "use_accurate_framebuffers", false);
+    Settings::values.use_accurate_gpu_emulation =
+        sdl2_config->GetBoolean("Renderer", "use_accurate_gpu_emulation", false);
 
     Settings::values.bg_red = (float)sdl2_config->GetReal("Renderer", "bg_red", 0.0);
     Settings::values.bg_green = (float)sdl2_config->GetReal("Renderer", "bg_green", 0.0);
@@ -125,10 +126,11 @@ void Config::ReadValues() {
 
     // System
     Settings::values.use_docked_mode = sdl2_config->GetBoolean("System", "use_docked_mode", false);
-    Settings::values.username = sdl2_config->Get("System", "username", "yuzu");
-    if (Settings::values.username.empty()) {
-        Settings::values.username = "yuzu";
-    }
+    Settings::values.enable_nfc = sdl2_config->GetBoolean("System", "enable_nfc", true);
+    const auto size = sdl2_config->GetInteger("System", "users_size", 0);
+
+    Settings::values.current_user = std::clamp<int>(
+        sdl2_config->GetInteger("System", "current_user", 0), 0, Service::Account::MAX_USERS - 1);
 
     // Miscellaneous
     Settings::values.log_filter = sdl2_config->Get("Miscellaneous", "log_filter", "*:Trace");
@@ -138,6 +140,7 @@ void Config::ReadValues() {
     Settings::values.use_gdbstub = sdl2_config->GetBoolean("Debugging", "use_gdbstub", false);
     Settings::values.gdbstub_port =
         static_cast<u16>(sdl2_config->GetInteger("Debugging", "gdbstub_port", 24689));
+    Settings::values.program_args = sdl2_config->Get("Debugging", "program_args", "");
 
     // Web Service
     Settings::values.enable_telemetry =
diff --git a/src/yuzu_cmd/default_ini.h b/src/yuzu_cmd/default_ini.h
index 762396e3b..e0b223cd6 100644
--- a/src/yuzu_cmd/default_ini.h
+++ b/src/yuzu_cmd/default_ini.h
@@ -110,9 +110,9 @@ use_frame_limit =
 # 1 - 9999: Speed limit as a percentage of target game speed. 100 (default)
 frame_limit =
 
-# Whether to use accurate framebuffers
+# Whether to use accurate GPU emulation
 # 0 (default): Off (fast), 1 : On (slow)
-use_accurate_framebuffers =
+use_accurate_gpu_emulation =
 
 # The clear color for the renderer. What shows up on the sides of the bottom screen.
 # Must be in range of 0.0-1.0. Defaults to 1.0 for all.
@@ -174,6 +174,10 @@ use_virtual_sd =
 # 1: Yes, 0 (default): No
 use_docked_mode =
 
+# Allow the use of NFC in games
+# 1 (default): Yes, 0 : No
+enable_nfc =
+
 # Sets the account username, max length is 32 characters
 # yuzu (default)
 username = yuzu
diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp
index 0733301b2..a9ad92a80 100644
--- a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp
+++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp
@@ -40,6 +40,35 @@ void EmuWindow_SDL2::OnMouseButton(u32 button, u8 state, s32 x, s32 y) {
     }
 }
 
+std::pair<unsigned, unsigned> EmuWindow_SDL2::TouchToPixelPos(float touch_x, float touch_y) const {
+    int w, h;
+    SDL_GetWindowSize(render_window, &w, &h);
+
+    touch_x *= w;
+    touch_y *= h;
+
+    return {static_cast<unsigned>(std::max(std::round(touch_x), 0.0f)),
+            static_cast<unsigned>(std::max(std::round(touch_y), 0.0f))};
+}
+
+void EmuWindow_SDL2::OnFingerDown(float x, float y) {
+    // TODO(NeatNit): keep track of multitouch using the fingerID and a dictionary of some kind
+    // This isn't critical because the best we can do when we have that is to average them, like the
+    // 3DS does
+
+    const auto [px, py] = TouchToPixelPos(x, y);
+    TouchPressed(px, py);
+}
+
+void EmuWindow_SDL2::OnFingerMotion(float x, float y) {
+    const auto [px, py] = TouchToPixelPos(x, y);
+    TouchMoved(px, py);
+}
+
+void EmuWindow_SDL2::OnFingerUp() {
+    TouchReleased();
+}
+
 void EmuWindow_SDL2::OnKeyEvent(int key, u8 state) {
     if (state == SDL_PRESSED) {
         InputCommon::GetKeyboard()->PressKey(key);
@@ -98,6 +127,8 @@ bool EmuWindow_SDL2::SupportsRequiredGLExtensions() {
         unsupported_ext.push_back("ARB_texture_storage");
     if (!GLAD_GL_ARB_multi_bind)
         unsupported_ext.push_back("ARB_multi_bind");
+    if (!GLAD_GL_ARB_copy_image)
+        unsupported_ext.push_back("ARB_copy_image");
 
     // Extensions required to support some texture formats.
     if (!GLAD_GL_EXT_texture_compression_s3tc)
@@ -217,11 +248,26 @@ void EmuWindow_SDL2::PollEvents() {
             OnKeyEvent(static_cast<int>(event.key.keysym.scancode), event.key.state);
             break;
         case SDL_MOUSEMOTION:
-            OnMouseMotion(event.motion.x, event.motion.y);
+            // ignore if it came from touch
+            if (event.button.which != SDL_TOUCH_MOUSEID)
+                OnMouseMotion(event.motion.x, event.motion.y);
             break;
         case SDL_MOUSEBUTTONDOWN:
         case SDL_MOUSEBUTTONUP:
-            OnMouseButton(event.button.button, event.button.state, event.button.x, event.button.y);
+            // ignore if it came from touch
+            if (event.button.which != SDL_TOUCH_MOUSEID) {
+                OnMouseButton(event.button.button, event.button.state, event.button.x,
+                              event.button.y);
+            }
+            break;
+        case SDL_FINGERDOWN:
+            OnFingerDown(event.tfinger.x, event.tfinger.y);
+            break;
+        case SDL_FINGERMOTION:
+            OnFingerMotion(event.tfinger.x, event.tfinger.y);
+            break;
+        case SDL_FINGERUP:
+            OnFingerUp();
             break;
         case SDL_QUIT:
             is_open = false;
diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.h b/src/yuzu_cmd/emu_window/emu_window_sdl2.h
index d34902109..b0d4116cc 100644
--- a/src/yuzu_cmd/emu_window/emu_window_sdl2.h
+++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.h
@@ -40,6 +40,18 @@ private:
     /// Called by PollEvents when a mouse button is pressed or released
     void OnMouseButton(u32 button, u8 state, s32 x, s32 y);
 
+    /// Translates pixel position (0..1) to pixel positions
+    std::pair<unsigned, unsigned> TouchToPixelPos(float touch_x, float touch_y) const;
+
+    /// Called by PollEvents when a finger starts touching the touchscreen
+    void OnFingerDown(float x, float y);
+
+    /// Called by PollEvents when a finger moves while touching the touchscreen
+    void OnFingerMotion(float x, float y);
+
+    /// Called by PollEvents when a finger stops touching the touchscreen
+    void OnFingerUp();
+
     /// Called by PollEvents when any event that may cause the window to be resized occurs
     void OnResize();
 
diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp
index bab465c1d..806127b12 100644
--- a/src/yuzu_cmd/yuzu.cpp
+++ b/src/yuzu_cmd/yuzu.cpp
@@ -56,9 +56,10 @@ static void PrintHelp(const char* argv0) {
     std::cout << "Usage: " << argv0
               << " [options] <filename>\n"
                  "-g, --gdbport=NUMBER  Enable gdb stub on port NUMBER\n"
-                 "-f, --fullscreen     Start in fullscreen mode\n"
+                 "-f, --fullscreen      Start in fullscreen mode\n"
                  "-h, --help            Display this help and exit\n"
-                 "-v, --version         Output version information and exit\n";
+                 "-v, --version         Output version information and exit\n"
+                 "-p, --program         Pass following string as arguments to executable\n";
 }
 
 static void PrintVersion() {
@@ -106,15 +107,13 @@ int main(int argc, char** argv) {
     bool fullscreen = false;
 
     static struct option long_options[] = {
-        {"gdbport", required_argument, 0, 'g'},
-        {"fullscreen", no_argument, 0, 'f'},
-        {"help", no_argument, 0, 'h'},
-        {"version", no_argument, 0, 'v'},
-        {0, 0, 0, 0},
+        {"gdbport", required_argument, 0, 'g'}, {"fullscreen", no_argument, 0, 'f'},
+        {"help", no_argument, 0, 'h'},          {"version", no_argument, 0, 'v'},
+        {"program", optional_argument, 0, 'p'}, {0, 0, 0, 0},
     };
 
     while (optind < argc) {
-        char arg = getopt_long(argc, argv, "g:fhv", long_options, &option_index);
+        char arg = getopt_long(argc, argv, "g:fhvp::", long_options, &option_index);
         if (arg != -1) {
             switch (arg) {
             case 'g':
@@ -138,6 +137,10 @@ int main(int argc, char** argv) {
             case 'v':
                 PrintVersion();
                 return 0;
+            case 'p':
+                Settings::values.program_args = argv[optind];
+                ++optind;
+                break;
             }
         } else {
 #ifdef _WIN32
@@ -175,7 +178,7 @@ int main(int argc, char** argv) {
 
     Core::System& system{Core::System::GetInstance()};
     system.SetFilesystem(std::make_shared<FileSys::RealVfsFilesystem>());
-    Service::FileSystem::CreateFactories(system.GetFilesystem());
+    Service::FileSystem::CreateFactories(*system.GetFilesystem());
 
     SCOPE_EXIT({ system.Shutdown(); });