210 files changed, 2076 insertions, 957 deletions

diff --git a/src/audio_core/audio_renderer.cpp b/src/audio_core/audio_renderer.cpp
index 397b107f5..a75cd3be5 100644
--- a/src/audio_core/audio_renderer.cpp
+++ b/src/audio_core/audio_renderer.cpp
@@ -17,10 +17,10 @@ AudioRenderer::AudioRenderer(AudioRendererParameter params,
                              Kernel::SharedPtr<Kernel::Event> buffer_event)
     : worker_params{params}, buffer_event{buffer_event}, voices(params.voice_count) {
 
-    audio_core = std::make_unique<AudioCore::AudioOut>();
-    stream = audio_core->OpenStream(STREAM_SAMPLE_RATE, STREAM_NUM_CHANNELS, "AudioRenderer",
-                                    [=]() { buffer_event->Signal(); });
-    audio_core->StartStream(stream);
+    audio_out = std::make_unique<AudioCore::AudioOut>();
+    stream = audio_out->OpenStream(STREAM_SAMPLE_RATE, STREAM_NUM_CHANNELS, "AudioRenderer",
+                                   [=]() { buffer_event->Signal(); });
+    audio_out->StartStream(stream);
 
     QueueMixedBuffer(0);
     QueueMixedBuffer(1);
@@ -236,11 +236,11 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) {
             }
         }
     }
-    audio_core->QueueBuffer(stream, tag, std::move(buffer));
+    audio_out->QueueBuffer(stream, tag, std::move(buffer));
 }
 
 void AudioRenderer::ReleaseAndQueueBuffers() {
-    const auto released_buffers{audio_core->GetTagsAndReleaseBuffers(stream, 2)};
+    const auto released_buffers{audio_out->GetTagsAndReleaseBuffers(stream, 2)};
     for (const auto& tag : released_buffers) {
         QueueMixedBuffer(tag);
     }
diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h
index eba67f28e..6d069d693 100644
--- a/src/audio_core/audio_renderer.h
+++ b/src/audio_core/audio_renderer.h
@@ -204,7 +204,7 @@ private:
     AudioRendererParameter worker_params;
     Kernel::SharedPtr<Kernel::Event> buffer_event;
     std::vector<VoiceState> voices;
-    std::unique_ptr<AudioCore::AudioOut> audio_core;
+    std::unique_ptr<AudioCore::AudioOut> audio_out;
     AudioCore::StreamPtr stream;
 };
 
diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt
index 7ddc87539..26f727d96 100644
--- a/src/core/CMakeLists.txt
+++ b/src/core/CMakeLists.txt
@@ -388,7 +388,7 @@ add_library(core STATIC
 create_target_directory_groups(core)
 
 target_link_libraries(core PUBLIC common PRIVATE audio_core video_core)
-target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt lz4_static mbedtls opus unicorn)
+target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt lz4_static mbedtls opus unicorn open_source_archives)
 
 if (ARCHITECTURE_x86_64)
     target_sources(core PRIVATE
diff --git a/src/core/hle/kernel/errors.h b/src/core/hle/kernel/errors.h
index 4054d5db6..ad39c8271 100644
--- a/src/core/hle/kernel/errors.h
+++ b/src/core/hle/kernel/errors.h
@@ -21,6 +21,7 @@ enum {
     HandleTableFull = 105,
     InvalidMemoryState = 106,
     InvalidMemoryPermissions = 108,
+    InvalidThreadPriority = 112,
     InvalidProcessorId = 113,
     InvalidHandle = 114,
     InvalidCombination = 116,
@@ -36,7 +37,7 @@ 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 these with correct errors for Switch OS
+// 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_PORT_NAME_TOO_LONG(ErrorModule::Kernel, ErrCodes::TooLarge);
@@ -53,15 +54,16 @@ constexpr ResultCode ERR_INVALID_ADDRESS_STATE(ErrorModule::Kernel, ErrCodes::In
 constexpr ResultCode ERR_INVALID_MEMORY_PERMISSIONS(ErrorModule::Kernel,
                                                     ErrCodes::InvalidMemoryPermissions);
 constexpr ResultCode ERR_INVALID_HANDLE(ErrorModule::Kernel, ErrCodes::InvalidHandle);
+constexpr ResultCode ERR_INVALID_PROCESSOR_ID(ErrorModule::Kernel, ErrCodes::InvalidProcessorId);
 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_NOT_FOUND(-1);
-constexpr ResultCode ERR_OUT_OF_RANGE(-1);
-constexpr ResultCode ERR_OUT_OF_RANGE_KERNEL(-1);
 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);
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index 1c9373ed8..f500fd2e7 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -273,7 +273,11 @@ static void Break(u64 reason, u64 info1, u64 info2) {
 }
 
 /// Used to output a message on a debug hardware unit - does nothing on a retail unit
-static void OutputDebugString(VAddr address, s32 len) {
+static void OutputDebugString(VAddr address, u64 len) {
+    if (len == 0) {
+        return;
+    }
+
     std::string str(len, '\0');
     Memory::ReadBlock(address, str.data(), str.size());
     LOG_DEBUG(Debug_Emulated, "{}", str);
@@ -378,7 +382,7 @@ static ResultCode GetThreadPriority(u32* priority, Handle handle) {
 /// Sets the priority for the specified thread
 static ResultCode SetThreadPriority(Handle handle, u32 priority) {
     if (priority > THREADPRIO_LOWEST) {
-        return ERR_OUT_OF_RANGE;
+        return ERR_INVALID_THREAD_PRIORITY;
     }
 
     auto& kernel = Core::System::GetInstance().Kernel();
@@ -523,7 +527,7 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
     std::string name = fmt::format("unknown-{:X}", entry_point);
 
     if (priority > THREADPRIO_LOWEST) {
-        return ERR_OUT_OF_RANGE;
+        return ERR_INVALID_THREAD_PRIORITY;
     }
 
     SharedPtr<ResourceLimit>& resource_limit = Core::CurrentProcess()->resource_limit;
@@ -544,8 +548,8 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
     case THREADPROCESSORID_3:
         break;
     default:
-        ASSERT_MSG(false, "Unsupported thread processor ID: {}", processor_id);
-        break;
+        LOG_ERROR(Kernel_SVC, "Invalid thread processor ID: {}", processor_id);
+        return ERR_INVALID_PROCESSOR_ID;
     }
 
     auto& kernel = Core::System::GetInstance().Kernel();
diff --git a/src/core/hle/kernel/svc_wrap.h b/src/core/hle/kernel/svc_wrap.h
index 79c3fe31b..1eda5f879 100644
--- a/src/core/hle/kernel/svc_wrap.h
+++ b/src/core/hle/kernel/svc_wrap.h
@@ -222,9 +222,9 @@ void SvcWrap() {
     func((s64)PARAM(0));
 }
 
-template <void func(u64, s32 len)>
+template <void func(u64, u64 len)>
 void SvcWrap() {
-    func(PARAM(0), (s32)(PARAM(1) & 0xFFFFFFFF));
+    func(PARAM(0), PARAM(1));
 }
 
 template <void func(u64, u64, u64)>
diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp
index 3d10d9af2..3f12a84dc 100644
--- a/src/core/hle/kernel/thread.cpp
+++ b/src/core/hle/kernel/thread.cpp
@@ -227,12 +227,12 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
     // Check if priority is in ranged. Lowest priority -> highest priority id.
     if (priority > THREADPRIO_LOWEST) {
         LOG_ERROR(Kernel_SVC, "Invalid thread priority: {}", priority);
-        return ERR_OUT_OF_RANGE;
+        return ERR_INVALID_THREAD_PRIORITY;
     }
 
     if (processor_id > THREADPROCESSORID_MAX) {
         LOG_ERROR(Kernel_SVC, "Invalid processor id: {}", processor_id);
-        return ERR_OUT_OF_RANGE_KERNEL;
+        return ERR_INVALID_PROCESSOR_ID;
     }
 
     // TODO(yuriks): Other checks, returning 0xD9001BEA
diff --git a/src/core/hle/service/acc/acc_aa.cpp b/src/core/hle/service/acc/acc_aa.cpp
index 9bd595a37..e84d9f7cf 100644
--- a/src/core/hle/service/acc/acc_aa.cpp
+++ b/src/core/hle/service/acc/acc_aa.cpp
@@ -18,4 +18,6 @@ ACC_AA::ACC_AA(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
     RegisterHandlers(functions);
 }
 
+ACC_AA::~ACC_AA() = default;
+
 } // namespace Service::Account
diff --git a/src/core/hle/service/acc/acc_aa.h b/src/core/hle/service/acc/acc_aa.h
index 2e08c781a..9edb0421b 100644
--- a/src/core/hle/service/acc/acc_aa.h
+++ b/src/core/hle/service/acc/acc_aa.h
@@ -12,6 +12,7 @@ class ACC_AA final : public Module::Interface {
 public:
     explicit ACC_AA(std::shared_ptr<Module> module,
                     std::shared_ptr<ProfileManager> profile_manager);
+    ~ACC_AA() override;
 };
 
 } // namespace Service::Account
diff --git a/src/core/hle/service/acc/acc_su.cpp b/src/core/hle/service/acc/acc_su.cpp
index 0218ee859..ad455c3a7 100644
--- a/src/core/hle/service/acc/acc_su.cpp
+++ b/src/core/hle/service/acc/acc_su.cpp
@@ -51,4 +51,6 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
     RegisterHandlers(functions);
 }
 
+ACC_SU::~ACC_SU() = default;
+
 } // namespace Service::Account
diff --git a/src/core/hle/service/acc/acc_su.h b/src/core/hle/service/acc/acc_su.h
index 79a47d88d..a3eb885bf 100644
--- a/src/core/hle/service/acc/acc_su.h
+++ b/src/core/hle/service/acc/acc_su.h
@@ -13,6 +13,7 @@ class ACC_SU final : public Module::Interface {
 public:
     explicit ACC_SU(std::shared_ptr<Module> module,
                     std::shared_ptr<ProfileManager> profile_manager);
+    ~ACC_SU() override;
 };
 
 } // namespace Account
diff --git a/src/core/hle/service/acc/acc_u0.cpp b/src/core/hle/service/acc/acc_u0.cpp
index 84a4d05b8..72d4adf35 100644
--- a/src/core/hle/service/acc/acc_u0.cpp
+++ b/src/core/hle/service/acc/acc_u0.cpp
@@ -31,4 +31,6 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
     RegisterHandlers(functions);
 }
 
+ACC_U0::~ACC_U0() = default;
+
 } // namespace Service::Account
diff --git a/src/core/hle/service/acc/acc_u0.h b/src/core/hle/service/acc/acc_u0.h
index e8a114f99..a1290e0bd 100644
--- a/src/core/hle/service/acc/acc_u0.h
+++ b/src/core/hle/service/acc/acc_u0.h
@@ -12,6 +12,7 @@ class ACC_U0 final : public Module::Interface {
 public:
     explicit ACC_U0(std::shared_ptr<Module> module,
                     std::shared_ptr<ProfileManager> profile_manager);
+    ~ACC_U0() override;
 };
 
 } // namespace Service::Account
diff --git a/src/core/hle/service/acc/acc_u1.cpp b/src/core/hle/service/acc/acc_u1.cpp
index 495693949..d480f08e5 100644
--- a/src/core/hle/service/acc/acc_u1.cpp
+++ b/src/core/hle/service/acc/acc_u1.cpp
@@ -38,4 +38,6 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
     RegisterHandlers(functions);
 }
 
+ACC_U1::~ACC_U1() = default;
+
 } // namespace Service::Account
diff --git a/src/core/hle/service/acc/acc_u1.h b/src/core/hle/service/acc/acc_u1.h
index a77520e6f..9e79daee3 100644
--- a/src/core/hle/service/acc/acc_u1.h
+++ b/src/core/hle/service/acc/acc_u1.h
@@ -12,6 +12,7 @@ class ACC_U1 final : public Module::Interface {
 public:
     explicit ACC_U1(std::shared_ptr<Module> module,
                     std::shared_ptr<ProfileManager> profile_manager);
+    ~ACC_U1() override;
 };
 
 } // namespace Service::Account
diff --git a/src/core/hle/service/acc/profile_manager.cpp b/src/core/hle/service/acc/profile_manager.cpp
index e0b03d763..4ccebef23 100644
--- a/src/core/hle/service/acc/profile_manager.cpp
+++ b/src/core/hle/service/acc/profile_manager.cpp
@@ -29,6 +29,8 @@ ProfileManager::ProfileManager() {
     OpenUser(user_uuid);
 }
 
+ProfileManager::~ProfileManager() = default;
+
 /// After a users creation it needs to be "registered" to the system. AddToProfiles handles the
 /// internal management of the users profiles
 boost::optional<size_t> ProfileManager::AddToProfiles(const ProfileInfo& user) {
diff --git a/src/core/hle/service/acc/profile_manager.h b/src/core/hle/service/acc/profile_manager.h
index 52967844d..cd8df93a5 100644
--- a/src/core/hle/service/acc/profile_manager.h
+++ b/src/core/hle/service/acc/profile_manager.h
@@ -82,6 +82,8 @@ static_assert(sizeof(ProfileBase) == 0x38, "ProfileBase is an invalid size");
 class ProfileManager {
 public:
     ProfileManager(); // TODO(ogniK): Load from system save
+    ~ProfileManager();
+
     ResultCode AddUser(const ProfileInfo& user);
     ResultCode CreateNewUser(UUID uuid, const ProfileUsername& username);
     ResultCode CreateNewUser(UUID uuid, const std::string& username);
diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp
index 818c03e0f..a57ed3042 100644
--- a/src/core/hle/service/am/am.cpp
+++ b/src/core/hle/service/am/am.cpp
@@ -35,6 +35,8 @@ IWindowController::IWindowController() : ServiceFramework("IWindowController") {
     RegisterHandlers(functions);
 }
 
+IWindowController::~IWindowController() = default;
+
 void IWindowController::GetAppletResourceUserId(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_AM, "(STUBBED) called");
     IPC::ResponseBuilder rb{ctx, 4};
@@ -61,6 +63,8 @@ IAudioController::IAudioController() : ServiceFramework("IAudioController") {
     RegisterHandlers(functions);
 }
 
+IAudioController::~IAudioController() = default;
+
 void IAudioController::SetExpectedMasterVolume(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_AM, "(STUBBED) called");
     IPC::ResponseBuilder rb{ctx, 2};
@@ -116,7 +120,10 @@ IDisplayController::IDisplayController() : ServiceFramework("IDisplayController"
     RegisterHandlers(functions);
 }
 
+IDisplayController::~IDisplayController() = default;
+
 IDebugFunctions::IDebugFunctions() : ServiceFramework("IDebugFunctions") {}
+IDebugFunctions::~IDebugFunctions() = default;
 
 ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     : ServiceFramework("ISelfController"), nvflinger(std::move(nvflinger)) {
@@ -165,6 +172,8 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
         Kernel::Event::Create(kernel, Kernel::ResetType::Sticky, "ISelfController:LaunchableEvent");
 }
 
+ISelfController::~ISelfController() = default;
+
 void ISelfController::SetFocusHandlingMode(Kernel::HLERequestContext& ctx) {
     // Takes 3 input u8s with each field located immediately after the previous u8, these are
     // bool flags. No output.
@@ -337,6 +346,8 @@ ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter"
     event = Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "ICommonStateGetter:Event");
 }
 
+ICommonStateGetter::~ICommonStateGetter() = default;
+
 void ICommonStateGetter::GetBootMode(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
@@ -573,6 +584,8 @@ ILibraryAppletCreator::ILibraryAppletCreator() : ServiceFramework("ILibraryApple
     RegisterHandlers(functions);
 }
 
+ILibraryAppletCreator::~ILibraryAppletCreator() = default;
+
 void ILibraryAppletCreator::CreateLibraryApplet(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
 
@@ -638,6 +651,8 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
     RegisterHandlers(functions);
 }
 
+IApplicationFunctions::~IApplicationFunctions() = default;
+
 void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) {
     constexpr std::array<u8, 0x88> data{{
         0xca, 0x97, 0x94, 0xc7, // Magic
@@ -760,6 +775,8 @@ IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions"
     RegisterHandlers(functions);
 }
 
+IHomeMenuFunctions::~IHomeMenuFunctions() = default;
+
 void IHomeMenuFunctions::RequestToGetForeground(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(RESULT_SUCCESS);
@@ -783,6 +800,8 @@ IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStat
     RegisterHandlers(functions);
 }
 
+IGlobalStateController::~IGlobalStateController() = default;
+
 IApplicationCreator::IApplicationCreator() : ServiceFramework("IApplicationCreator") {
     static const FunctionInfo functions[] = {
         {0, nullptr, "CreateApplication"},
@@ -793,6 +812,8 @@ IApplicationCreator::IApplicationCreator() : ServiceFramework("IApplicationCreat
     RegisterHandlers(functions);
 }
 
+IApplicationCreator::~IApplicationCreator() = default;
+
 IProcessWindingController::IProcessWindingController()
     : ServiceFramework("IProcessWindingController") {
     static const FunctionInfo functions[] = {
@@ -807,4 +828,6 @@ IProcessWindingController::IProcessWindingController()
     };
     RegisterHandlers(functions);
 }
+
+IProcessWindingController::~IProcessWindingController() = default;
 } // namespace Service::AM
diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h
index 9e8bb4e43..fd9ae296b 100644
--- a/src/core/hle/service/am/am.h
+++ b/src/core/hle/service/am/am.h
@@ -42,6 +42,7 @@ enum SystemLanguage {
 class IWindowController final : public ServiceFramework<IWindowController> {
 public:
     IWindowController();
+    ~IWindowController() override;
 
 private:
     void GetAppletResourceUserId(Kernel::HLERequestContext& ctx);
@@ -51,6 +52,7 @@ private:
 class IAudioController final : public ServiceFramework<IAudioController> {
 public:
     IAudioController();
+    ~IAudioController() override;
 
 private:
     void SetExpectedMasterVolume(Kernel::HLERequestContext& ctx);
@@ -63,16 +65,19 @@ private:
 class IDisplayController final : public ServiceFramework<IDisplayController> {
 public:
     IDisplayController();
+    ~IDisplayController() override;
 };
 
 class IDebugFunctions final : public ServiceFramework<IDebugFunctions> {
 public:
     IDebugFunctions();
+    ~IDebugFunctions() override;
 };
 
 class ISelfController final : public ServiceFramework<ISelfController> {
 public:
     explicit ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger);
+    ~ISelfController() override;
 
 private:
     void SetFocusHandlingMode(Kernel::HLERequestContext& ctx);
@@ -98,6 +103,7 @@ private:
 class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> {
 public:
     ICommonStateGetter();
+    ~ICommonStateGetter() override;
 
 private:
     enum class FocusState : u8 {
@@ -124,6 +130,7 @@ private:
 class ILibraryAppletCreator final : public ServiceFramework<ILibraryAppletCreator> {
 public:
     ILibraryAppletCreator();
+    ~ILibraryAppletCreator() override;
 
 private:
     void CreateLibraryApplet(Kernel::HLERequestContext& ctx);
@@ -133,6 +140,7 @@ private:
 class IApplicationFunctions final : public ServiceFramework<IApplicationFunctions> {
 public:
     IApplicationFunctions();
+    ~IApplicationFunctions() override;
 
 private:
     void PopLaunchParameter(Kernel::HLERequestContext& ctx);
@@ -150,6 +158,7 @@ private:
 class IHomeMenuFunctions final : public ServiceFramework<IHomeMenuFunctions> {
 public:
     IHomeMenuFunctions();
+    ~IHomeMenuFunctions() override;
 
 private:
     void RequestToGetForeground(Kernel::HLERequestContext& ctx);
@@ -158,16 +167,19 @@ private:
 class IGlobalStateController final : public ServiceFramework<IGlobalStateController> {
 public:
     IGlobalStateController();
+    ~IGlobalStateController() override;
 };
 
 class IApplicationCreator final : public ServiceFramework<IApplicationCreator> {
 public:
     IApplicationCreator();
+    ~IApplicationCreator() override;
 };
 
 class IProcessWindingController final : public ServiceFramework<IProcessWindingController> {
 public:
     IProcessWindingController();
+    ~IProcessWindingController() override;
 };
 
 /// Registers all AM services with the specified service manager.
diff --git a/src/core/hle/service/am/applet_ae.cpp b/src/core/hle/service/am/applet_ae.cpp
index 7cebc918a..4296c255e 100644
--- a/src/core/hle/service/am/applet_ae.cpp
+++ b/src/core/hle/service/am/applet_ae.cpp
@@ -222,4 +222,6 @@ AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     RegisterHandlers(functions);
 }
 
+AppletAE::~AppletAE() = default;
+
 } // namespace Service::AM
diff --git a/src/core/hle/service/am/applet_ae.h b/src/core/hle/service/am/applet_ae.h
index bdc57b9bc..1ed77baa4 100644
--- a/src/core/hle/service/am/applet_ae.h
+++ b/src/core/hle/service/am/applet_ae.h
@@ -18,7 +18,7 @@ namespace AM {
 class AppletAE final : public ServiceFramework<AppletAE> {
 public:
     explicit AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger);
-    ~AppletAE() = default;
+    ~AppletAE() override;
 
 private:
     void OpenSystemAppletProxy(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/am/applet_oe.cpp b/src/core/hle/service/am/applet_oe.cpp
index beea7d19b..e45cf6e20 100644
--- a/src/core/hle/service/am/applet_oe.cpp
+++ b/src/core/hle/service/am/applet_oe.cpp
@@ -103,4 +103,6 @@ AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     RegisterHandlers(functions);
 }
 
+AppletOE::~AppletOE() = default;
+
 } // namespace Service::AM
diff --git a/src/core/hle/service/am/applet_oe.h b/src/core/hle/service/am/applet_oe.h
index c52e2a322..60cfdfd9d 100644
--- a/src/core/hle/service/am/applet_oe.h
+++ b/src/core/hle/service/am/applet_oe.h
@@ -18,7 +18,7 @@ namespace AM {
 class AppletOE final : public ServiceFramework<AppletOE> {
 public:
     explicit AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger);
-    ~AppletOE() = default;
+    ~AppletOE() override;
 
 private:
     void OpenApplicationProxy(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/am/idle.cpp b/src/core/hle/service/am/idle.cpp
index af46e9494..0e3088bc8 100644
--- a/src/core/hle/service/am/idle.cpp
+++ b/src/core/hle/service/am/idle.cpp
@@ -21,4 +21,6 @@ IdleSys::IdleSys() : ServiceFramework{"idle:sys"} {
     RegisterHandlers(functions);
 }
 
+IdleSys::~IdleSys() = default;
+
 } // namespace Service::AM
diff --git a/src/core/hle/service/am/idle.h b/src/core/hle/service/am/idle.h
index 1eb68d2c9..c44e856b1 100644
--- a/src/core/hle/service/am/idle.h
+++ b/src/core/hle/service/am/idle.h
@@ -11,6 +11,7 @@ namespace Service::AM {
 class IdleSys final : public ServiceFramework<IdleSys> {
 public:
     explicit IdleSys();
+    ~IdleSys() override;
 };
 
 } // namespace Service::AM
diff --git a/src/core/hle/service/am/omm.cpp b/src/core/hle/service/am/omm.cpp
index 447fe8669..1c37f849f 100644
--- a/src/core/hle/service/am/omm.cpp
+++ b/src/core/hle/service/am/omm.cpp
@@ -39,4 +39,6 @@ OMM::OMM() : ServiceFramework{"omm"} {
     RegisterHandlers(functions);
 }
 
+OMM::~OMM() = default;
+
 } // namespace Service::AM
diff --git a/src/core/hle/service/am/omm.h b/src/core/hle/service/am/omm.h
index 49e5d331c..59dc91b72 100644
--- a/src/core/hle/service/am/omm.h
+++ b/src/core/hle/service/am/omm.h
@@ -11,6 +11,7 @@ namespace Service::AM {
 class OMM final : public ServiceFramework<OMM> {
 public:
     explicit OMM();
+    ~OMM() override;
 };
 
 } // namespace Service::AM
diff --git a/src/core/hle/service/am/spsm.cpp b/src/core/hle/service/am/spsm.cpp
index a05d433d0..003ee8667 100644
--- a/src/core/hle/service/am/spsm.cpp
+++ b/src/core/hle/service/am/spsm.cpp
@@ -27,4 +27,6 @@ SPSM::SPSM() : ServiceFramework{"spsm"} {
     RegisterHandlers(functions);
 }
 
+SPSM::~SPSM() = default;
+
 } // namespace Service::AM
diff --git a/src/core/hle/service/am/spsm.h b/src/core/hle/service/am/spsm.h
index 57dde62e1..3a0b979fa 100644
--- a/src/core/hle/service/am/spsm.h
+++ b/src/core/hle/service/am/spsm.h
@@ -11,6 +11,7 @@ namespace Service::AM {
 class SPSM final : public ServiceFramework<SPSM> {
 public:
     explicit SPSM();
+    ~SPSM() 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 6e7438580..d9eeac9ec 100644
--- a/src/core/hle/service/aoc/aoc_u.cpp
+++ b/src/core/hle/service/aoc/aoc_u.cpp
@@ -23,6 +23,8 @@ AOC_U::AOC_U() : ServiceFramework("aoc:u") {
     RegisterHandlers(functions);
 }
 
+AOC_U::~AOC_U() = default;
+
 void AOC_U::CountAddOnContent(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 4};
     rb.Push(RESULT_SUCCESS);
diff --git a/src/core/hle/service/aoc/aoc_u.h b/src/core/hle/service/aoc/aoc_u.h
index 17d48ef30..29ce8f488 100644
--- a/src/core/hle/service/aoc/aoc_u.h
+++ b/src/core/hle/service/aoc/aoc_u.h
@@ -11,7 +11,7 @@ namespace Service::AOC {
 class AOC_U final : public ServiceFramework<AOC_U> {
 public:
     AOC_U();
-    ~AOC_U() = default;
+    ~AOC_U() override;
 
 private:
     void CountAddOnContent(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/apm/apm.cpp b/src/core/hle/service/apm/apm.cpp
index 4109cb7f7..f3c09bbb1 100644
--- a/src/core/hle/service/apm/apm.cpp
+++ b/src/core/hle/service/apm/apm.cpp
@@ -9,6 +9,9 @@
 
 namespace Service::APM {
 
+Module::Module() = default;
+Module::~Module() = default;
+
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     auto module_ = std::make_shared<Module>();
     std::make_shared<APM>(module_, "apm")->InstallAsService(service_manager);
diff --git a/src/core/hle/service/apm/apm.h b/src/core/hle/service/apm/apm.h
index 90a80d51b..4d7d5bb7c 100644
--- a/src/core/hle/service/apm/apm.h
+++ b/src/core/hle/service/apm/apm.h
@@ -15,8 +15,8 @@ enum class PerformanceMode : u8 {
 
 class Module final {
 public:
-    Module() = default;
-    ~Module() = default;
+    Module();
+    ~Module();
 };
 
 /// Registers all AM services with the specified service manager.
diff --git a/src/core/hle/service/apm/interface.cpp b/src/core/hle/service/apm/interface.cpp
index 4cd8132f5..c22bd3859 100644
--- a/src/core/hle/service/apm/interface.cpp
+++ b/src/core/hle/service/apm/interface.cpp
@@ -70,6 +70,8 @@ APM::APM(std::shared_ptr<Module> apm, const char* name)
     RegisterHandlers(functions);
 }
 
+APM::~APM() = default;
+
 void APM::OpenSession(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
@@ -93,6 +95,8 @@ APM_Sys::APM_Sys() : ServiceFramework{"apm:sys"} {
     RegisterHandlers(functions);
 }
 
+APM_Sys::~APM_Sys() = default;
+
 void APM_Sys::GetPerformanceEvent(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
diff --git a/src/core/hle/service/apm/interface.h b/src/core/hle/service/apm/interface.h
index d14264ad7..773541aa4 100644
--- a/src/core/hle/service/apm/interface.h
+++ b/src/core/hle/service/apm/interface.h
@@ -11,7 +11,7 @@ namespace Service::APM {
 class APM final : public ServiceFramework<APM> {
 public:
     explicit APM(std::shared_ptr<Module> apm, const char* name);
-    ~APM() = default;
+    ~APM() override;
 
 private:
     void OpenSession(Kernel::HLERequestContext& ctx);
@@ -22,6 +22,7 @@ private:
 class APM_Sys final : public ServiceFramework<APM_Sys> {
 public:
     explicit APM_Sys();
+    ~APM_Sys() override;
 
 private:
     void GetPerformanceEvent(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/audio/audctl.cpp b/src/core/hle/service/audio/audctl.cpp
index 37c3fdcac..b6b71f966 100644
--- a/src/core/hle/service/audio/audctl.cpp
+++ b/src/core/hle/service/audio/audctl.cpp
@@ -42,4 +42,6 @@ AudCtl::AudCtl() : ServiceFramework{"audctl"} {
     RegisterHandlers(functions);
 }
 
+AudCtl::~AudCtl() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audctl.h b/src/core/hle/service/audio/audctl.h
index ed837bdf2..9d2d9e83b 100644
--- a/src/core/hle/service/audio/audctl.h
+++ b/src/core/hle/service/audio/audctl.h
@@ -11,6 +11,7 @@ namespace Service::Audio {
 class AudCtl final : public ServiceFramework<AudCtl> {
 public:
     explicit AudCtl();
+    ~AudCtl() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/auddbg.cpp b/src/core/hle/service/audio/auddbg.cpp
index b08c21a20..8fff3e4b4 100644
--- a/src/core/hle/service/audio/auddbg.cpp
+++ b/src/core/hle/service/audio/auddbg.cpp
@@ -17,4 +17,6 @@ AudDbg::AudDbg(const char* name) : ServiceFramework{name} {
     RegisterHandlers(functions);
 }
 
+AudDbg::~AudDbg() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/auddbg.h b/src/core/hle/service/audio/auddbg.h
index a2f540b75..6689f4759 100644
--- a/src/core/hle/service/audio/auddbg.h
+++ b/src/core/hle/service/audio/auddbg.h
@@ -11,6 +11,7 @@ namespace Service::Audio {
 class AudDbg final : public ServiceFramework<AudDbg> {
 public:
     explicit AudDbg(const char* name);
+    ~AudDbg() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audin_a.cpp b/src/core/hle/service/audio/audin_a.cpp
index a70d5bca4..ddd12f35e 100644
--- a/src/core/hle/service/audio/audin_a.cpp
+++ b/src/core/hle/service/audio/audin_a.cpp
@@ -19,4 +19,6 @@ AudInA::AudInA() : ServiceFramework{"audin:a"} {
     RegisterHandlers(functions);
 }
 
+AudInA::~AudInA() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audin_a.h b/src/core/hle/service/audio/audin_a.h
index e4c75510f..e7623bc29 100644
--- a/src/core/hle/service/audio/audin_a.h
+++ b/src/core/hle/service/audio/audin_a.h
@@ -11,6 +11,7 @@ namespace Service::Audio {
 class AudInA final : public ServiceFramework<AudInA> {
 public:
     explicit AudInA();
+    ~AudInA() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audin_u.cpp b/src/core/hle/service/audio/audin_u.cpp
index cbc49e55e..657010312 100644
--- a/src/core/hle/service/audio/audin_u.cpp
+++ b/src/core/hle/service/audio/audin_u.cpp
@@ -41,4 +41,6 @@ AudInU::AudInU() : ServiceFramework("audin:u") {
     RegisterHandlers(functions);
 }
 
+AudInU::~AudInU() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audin_u.h b/src/core/hle/service/audio/audin_u.h
index 2e65efb5b..0538b9560 100644
--- a/src/core/hle/service/audio/audin_u.h
+++ b/src/core/hle/service/audio/audin_u.h
@@ -15,7 +15,7 @@ namespace Service::Audio {
 class AudInU final : public ServiceFramework<AudInU> {
 public:
     explicit AudInU();
-    ~AudInU() = default;
+    ~AudInU() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audio.cpp b/src/core/hle/service/audio/audio.cpp
index 6b5e15633..128df7db5 100644
--- a/src/core/hle/service/audio/audio.cpp
+++ b/src/core/hle/service/audio/audio.cpp
@@ -15,6 +15,7 @@
 #include "core/hle/service/audio/audren_u.h"
 #include "core/hle/service/audio/codecctl.h"
 #include "core/hle/service/audio/hwopus.h"
+#include "core/hle/service/service.h"
 
 namespace Service::Audio {
 
diff --git a/src/core/hle/service/audio/audio.h b/src/core/hle/service/audio/audio.h
index 95e5691f7..f5bd3bf5f 100644
--- a/src/core/hle/service/audio/audio.h
+++ b/src/core/hle/service/audio/audio.h
@@ -4,7 +4,9 @@
 
 #pragma once
 
-#include "core/hle/service/service.h"
+namespace Service::SM {
+class ServiceManager;
+}
 
 namespace Service::Audio {
 
diff --git a/src/core/hle/service/audio/audout_a.cpp b/src/core/hle/service/audio/audout_a.cpp
index bf8d40157..85febbca3 100644
--- a/src/core/hle/service/audio/audout_a.cpp
+++ b/src/core/hle/service/audio/audout_a.cpp
@@ -21,4 +21,6 @@ AudOutA::AudOutA() : ServiceFramework{"audout:a"} {
     RegisterHandlers(functions);
 }
 
+AudOutA::~AudOutA() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audout_a.h b/src/core/hle/service/audio/audout_a.h
index 91a069152..d65b66e8e 100644
--- a/src/core/hle/service/audio/audout_a.h
+++ b/src/core/hle/service/audio/audout_a.h
@@ -11,6 +11,7 @@ namespace Service::Audio {
 class AudOutA final : public ServiceFramework<AudOutA> {
 public:
     explicit AudOutA();
+    ~AudOutA() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp
index 5f370bbdf..80a002322 100644
--- a/src/core/hle/service/audio/audout_u.cpp
+++ b/src/core/hle/service/audio/audout_u.cpp
@@ -3,15 +3,20 @@
 // Refer to the license.txt file included.
 
 #include <array>
+#include <cstring>
 #include <vector>
 
+#include "audio_core/audio_out.h"
 #include "audio_core/codec.h"
+#include "common/common_funcs.h"
 #include "common/logging/log.h"
+#include "common/swap.h"
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/hle_ipc.h"
 #include "core/hle/service/audio/audout_u.h"
+#include "core/memory.h"
 
 namespace Service::Audio {
 
@@ -25,6 +30,18 @@ enum {
 constexpr std::array<char, 10> DefaultDevice{{"DeviceOut"}};
 constexpr int DefaultSampleRate{48000};
 
+struct AudoutParams {
+    s32_le sample_rate;
+    u16_le channel_count;
+    INSERT_PADDING_BYTES(2);
+};
+static_assert(sizeof(AudoutParams) == 0x8, "AudoutParams is an invalid size");
+
+enum class AudioState : u32 {
+    Started,
+    Stopped,
+};
+
 class IAudioOut final : public ServiceFramework<IAudioOut> {
 public:
     IAudioOut(AudoutParams audio_params, AudioCore::AudioOut& audio_core)
@@ -218,4 +235,6 @@ AudOutU::AudOutU() : ServiceFramework("audout:u") {
     audio_core = std::make_unique<AudioCore::AudioOut>();
 }
 
+AudOutU::~AudOutU() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audout_u.h b/src/core/hle/service/audio/audout_u.h
index fd491f65d..dcaf64708 100644
--- a/src/core/hle/service/audio/audout_u.h
+++ b/src/core/hle/service/audio/audout_u.h
@@ -4,33 +4,24 @@
 
 #pragma once
 
-#include "audio_core/audio_out.h"
 #include "core/hle/service/service.h"
 
+namespace AudioCore {
+class AudioOut;
+}
+
 namespace Kernel {
 class HLERequestContext;
 }
 
 namespace Service::Audio {
 
-struct AudoutParams {
-    s32_le sample_rate;
-    u16_le channel_count;
-    INSERT_PADDING_BYTES(2);
-};
-static_assert(sizeof(AudoutParams) == 0x8, "AudoutParams is an invalid size");
-
-enum class AudioState : u32 {
-    Started,
-    Stopped,
-};
-
 class IAudioOut;
 
 class AudOutU final : public ServiceFramework<AudOutU> {
 public:
     AudOutU();
-    ~AudOutU() = default;
+    ~AudOutU() override;
 
 private:
     std::shared_ptr<IAudioOut> audio_out_interface;
diff --git a/src/core/hle/service/audio/audrec_a.cpp b/src/core/hle/service/audio/audrec_a.cpp
index 016eabf53..ce1bfb48d 100644
--- a/src/core/hle/service/audio/audrec_a.cpp
+++ b/src/core/hle/service/audio/audrec_a.cpp
@@ -17,4 +17,6 @@ AudRecA::AudRecA() : ServiceFramework{"audrec:a"} {
     RegisterHandlers(functions);
 }
 
+AudRecA::~AudRecA() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audrec_a.h b/src/core/hle/service/audio/audrec_a.h
index 9685047f2..384d24c69 100644
--- a/src/core/hle/service/audio/audrec_a.h
+++ b/src/core/hle/service/audio/audrec_a.h
@@ -11,6 +11,7 @@ namespace Service::Audio {
 class AudRecA final : public ServiceFramework<AudRecA> {
 public:
     explicit AudRecA();
+    ~AudRecA() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audrec_u.cpp b/src/core/hle/service/audio/audrec_u.cpp
index 74909415c..34974afa9 100644
--- a/src/core/hle/service/audio/audrec_u.cpp
+++ b/src/core/hle/service/audio/audrec_u.cpp
@@ -36,4 +36,6 @@ AudRecU::AudRecU() : ServiceFramework("audrec:u") {
     RegisterHandlers(functions);
 }
 
+AudRecU::~AudRecU() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audrec_u.h b/src/core/hle/service/audio/audrec_u.h
index 46daa33a4..ca3d638e8 100644
--- a/src/core/hle/service/audio/audrec_u.h
+++ b/src/core/hle/service/audio/audrec_u.h
@@ -15,7 +15,7 @@ namespace Service::Audio {
 class AudRecU final : public ServiceFramework<AudRecU> {
 public:
     explicit AudRecU();
-    ~AudRecU() = default;
+    ~AudRecU() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audren_a.cpp b/src/core/hle/service/audio/audren_a.cpp
index 616ff3dc4..edb66d985 100644
--- a/src/core/hle/service/audio/audren_a.cpp
+++ b/src/core/hle/service/audio/audren_a.cpp
@@ -23,4 +23,6 @@ AudRenA::AudRenA() : ServiceFramework{"audren:a"} {
     RegisterHandlers(functions);
 }
 
+AudRenA::~AudRenA() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audren_a.h b/src/core/hle/service/audio/audren_a.h
index 5ecf2e184..81fef0ffe 100644
--- a/src/core/hle/service/audio/audren_a.h
+++ b/src/core/hle/service/audio/audren_a.h
@@ -11,6 +11,7 @@ namespace Service::Audio {
 class AudRenA final : public ServiceFramework<AudRenA> {
 public:
     explicit AudRenA();
+    ~AudRenA() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp
index 016db7c82..e84c4fa2b 100644
--- a/src/core/hle/service/audio/audren_u.cpp
+++ b/src/core/hle/service/audio/audren_u.cpp
@@ -2,12 +2,14 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include <array>
+#include <memory>
 
+#include "audio_core/audio_renderer.h"
 #include "common/alignment.h"
+#include "common/common_funcs.h"
 #include "common/logging/log.h"
-#include "core/core_timing.h"
-#include "core/core_timing_util.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/hle_ipc.h"
@@ -198,6 +200,8 @@ AudRenU::AudRenU() : ServiceFramework("audren:u") {
     RegisterHandlers(functions);
 }
 
+AudRenU::~AudRenU() = default;
+
 void AudRenU::OpenAudioRenderer(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     auto params = rp.PopRaw<AudioCore::AudioRendererParameter>();
diff --git a/src/core/hle/service/audio/audren_u.h b/src/core/hle/service/audio/audren_u.h
index 8600ac6e4..c6bc3a90a 100644
--- a/src/core/hle/service/audio/audren_u.h
+++ b/src/core/hle/service/audio/audren_u.h
@@ -4,7 +4,6 @@
 
 #pragma once
 
-#include "audio_core/audio_renderer.h"
 #include "core/hle/service/service.h"
 
 namespace Kernel {
@@ -16,7 +15,7 @@ namespace Service::Audio {
 class AudRenU final : public ServiceFramework<AudRenU> {
 public:
     explicit AudRenU();
-    ~AudRenU() = default;
+    ~AudRenU() override;
 
 private:
     void OpenAudioRenderer(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/audio/codecctl.cpp b/src/core/hle/service/audio/codecctl.cpp
index 212c8d448..c6864146d 100644
--- a/src/core/hle/service/audio/codecctl.cpp
+++ b/src/core/hle/service/audio/codecctl.cpp
@@ -28,4 +28,6 @@ CodecCtl::CodecCtl() : ServiceFramework("codecctl") {
     RegisterHandlers(functions);
 }
 
+CodecCtl::~CodecCtl() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/codecctl.h b/src/core/hle/service/audio/codecctl.h
index d9ac29b67..2fe75b6e2 100644
--- a/src/core/hle/service/audio/codecctl.h
+++ b/src/core/hle/service/audio/codecctl.h
@@ -15,7 +15,7 @@ namespace Service::Audio {
 class CodecCtl final : public ServiceFramework<CodecCtl> {
 public:
     explicit CodecCtl();
-    ~CodecCtl() = default;
+    ~CodecCtl() override;
 };
 
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/hwopus.cpp b/src/core/hle/service/audio/hwopus.cpp
index 371cd4997..668fef145 100644
--- a/src/core/hle/service/audio/hwopus.cpp
+++ b/src/core/hle/service/audio/hwopus.cpp
@@ -3,7 +3,12 @@
 // Refer to the license.txt file included.
 
 #include <cstring>
+#include <memory>
+#include <vector>
+
 #include <opus.h>
+
+#include "common/common_funcs.h"
 #include "common/logging/log.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/hle_ipc.h"
@@ -151,4 +156,6 @@ HwOpus::HwOpus() : ServiceFramework("hwopus") {
     RegisterHandlers(functions);
 }
 
+HwOpus::~HwOpus() = default;
+
 } // namespace Service::Audio
diff --git a/src/core/hle/service/audio/hwopus.h b/src/core/hle/service/audio/hwopus.h
index 5258d59f3..602ede8ba 100644
--- a/src/core/hle/service/audio/hwopus.h
+++ b/src/core/hle/service/audio/hwopus.h
@@ -11,7 +11,7 @@ namespace Service::Audio {
 class HwOpus final : public ServiceFramework<HwOpus> {
 public:
     explicit HwOpus();
-    ~HwOpus() = default;
+    ~HwOpus() override;
 
 private:
     void OpenOpusDecoder(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/bcat/bcat.cpp b/src/core/hle/service/bcat/bcat.cpp
index 20ce692dc..179aa4949 100644
--- a/src/core/hle/service/bcat/bcat.cpp
+++ b/src/core/hle/service/bcat/bcat.cpp
@@ -13,4 +13,6 @@ BCAT::BCAT(std::shared_ptr<Module> module, const char* name)
     };
     RegisterHandlers(functions);
 }
+
+BCAT::~BCAT() = default;
 } // namespace Service::BCAT
diff --git a/src/core/hle/service/bcat/bcat.h b/src/core/hle/service/bcat/bcat.h
index 6632996a0..802bd689a 100644
--- a/src/core/hle/service/bcat/bcat.h
+++ b/src/core/hle/service/bcat/bcat.h
@@ -11,6 +11,7 @@ namespace Service::BCAT {
 class BCAT final : public Module::Interface {
 public:
     explicit BCAT(std::shared_ptr<Module> module, const char* name);
+    ~BCAT() override;
 };
 
 } // namespace Service::BCAT
diff --git a/src/core/hle/service/bcat/module.cpp b/src/core/hle/service/bcat/module.cpp
index 35e024c3d..6e7b795fb 100644
--- a/src/core/hle/service/bcat/module.cpp
+++ b/src/core/hle/service/bcat/module.cpp
@@ -42,6 +42,8 @@ void Module::Interface::CreateBcatService(Kernel::HLERequestContext& ctx) {
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
     : ServiceFramework(name), module(std::move(module)) {}
 
+Module::Interface::~Interface() = default;
+
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     auto module = std::make_shared<Module>();
     std::make_shared<BCAT>(module, "bcat:a")->InstallAsService(service_manager);
diff --git a/src/core/hle/service/bcat/module.h b/src/core/hle/service/bcat/module.h
index 62f6f5f9d..f0d63cab0 100644
--- a/src/core/hle/service/bcat/module.h
+++ b/src/core/hle/service/bcat/module.h
@@ -13,6 +13,7 @@ public:
     class Interface : public ServiceFramework<Interface> {
     public:
         explicit Interface(std::shared_ptr<Module> module, const char* name);
+        ~Interface() override;
 
         void CreateBcatService(Kernel::HLERequestContext& ctx);
 
diff --git a/src/core/hle/service/fatal/fatal.cpp b/src/core/hle/service/fatal/fatal.cpp
index 299b9474f..b436ce4e6 100644
--- a/src/core/hle/service/fatal/fatal.cpp
+++ b/src/core/hle/service/fatal/fatal.cpp
@@ -13,6 +13,8 @@ namespace Service::Fatal {
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
     : ServiceFramework(name), module(std::move(module)) {}
 
+Module::Interface::~Interface() = default;
+
 void Module::Interface::ThrowFatalWithPolicy(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp(ctx);
     u32 error_code = rp.Pop<u32>();
diff --git a/src/core/hle/service/fatal/fatal.h b/src/core/hle/service/fatal/fatal.h
index ca607e236..4d9a5be52 100644
--- a/src/core/hle/service/fatal/fatal.h
+++ b/src/core/hle/service/fatal/fatal.h
@@ -13,6 +13,7 @@ public:
     class Interface : public ServiceFramework<Interface> {
     public:
         explicit Interface(std::shared_ptr<Module> module, const char* name);
+        ~Interface() override;
 
         void ThrowFatalWithPolicy(Kernel::HLERequestContext& ctx);
         void ThrowFatalWithCpuContext(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/fatal/fatal_p.cpp b/src/core/hle/service/fatal/fatal_p.cpp
index a5254ac2f..9e5f872ff 100644
--- a/src/core/hle/service/fatal/fatal_p.cpp
+++ b/src/core/hle/service/fatal/fatal_p.cpp
@@ -9,4 +9,6 @@ namespace Service::Fatal {
 Fatal_P::Fatal_P(std::shared_ptr<Module> module)
     : Module::Interface(std::move(module), "fatal:p") {}
 
+Fatal_P::~Fatal_P() = default;
+
 } // namespace Service::Fatal
diff --git a/src/core/hle/service/fatal/fatal_p.h b/src/core/hle/service/fatal/fatal_p.h
index bfd8c8b74..6e9c5979f 100644
--- a/src/core/hle/service/fatal/fatal_p.h
+++ b/src/core/hle/service/fatal/fatal_p.h
@@ -11,6 +11,7 @@ namespace Service::Fatal {
 class Fatal_P final : public Module::Interface {
 public:
     explicit Fatal_P(std::shared_ptr<Module> module);
+    ~Fatal_P() override;
 };
 
 } // namespace Service::Fatal
diff --git a/src/core/hle/service/fatal/fatal_u.cpp b/src/core/hle/service/fatal/fatal_u.cpp
index f0631329e..befc307cf 100644
--- a/src/core/hle/service/fatal/fatal_u.cpp
+++ b/src/core/hle/service/fatal/fatal_u.cpp
@@ -15,4 +15,6 @@ Fatal_U::Fatal_U(std::shared_ptr<Module> module) : Module::Interface(std::move(m
     RegisterHandlers(functions);
 }
 
+Fatal_U::~Fatal_U() = default;
+
 } // namespace Service::Fatal
diff --git a/src/core/hle/service/fatal/fatal_u.h b/src/core/hle/service/fatal/fatal_u.h
index 9b1a9e97a..72cb6d076 100644
--- a/src/core/hle/service/fatal/fatal_u.h
+++ b/src/core/hle/service/fatal/fatal_u.h
@@ -11,6 +11,7 @@ namespace Service::Fatal {
 class Fatal_U final : public Module::Interface {
 public:
     explicit Fatal_U(std::shared_ptr<Module> module);
+    ~Fatal_U() override;
 };
 
 } // namespace Service::Fatal
diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp
index 04c9d750f..5c4971724 100644
--- a/src/core/hle/service/filesystem/filesystem.cpp
+++ b/src/core/hle/service/filesystem/filesystem.cpp
@@ -40,6 +40,8 @@ static FileSys::VirtualDir GetDirectoryRelativeWrapped(FileSys::VirtualDir base,
 VfsDirectoryServiceWrapper::VfsDirectoryServiceWrapper(FileSys::VirtualDir backing_)
     : backing(std::move(backing_)) {}
 
+VfsDirectoryServiceWrapper::~VfsDirectoryServiceWrapper() = default;
+
 std::string VfsDirectoryServiceWrapper::GetName() const {
     return backing->GetName();
 }
diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h
index 793a7b06f..aab65a2b8 100644
--- a/src/core/hle/service/filesystem/filesystem.h
+++ b/src/core/hle/service/filesystem/filesystem.h
@@ -64,6 +64,7 @@ void InstallInterfaces(SM::ServiceManager& service_manager, const FileSys::Virtu
 class VfsDirectoryServiceWrapper {
 public:
     explicit VfsDirectoryServiceWrapper(FileSys::VirtualDir backing);
+    ~VfsDirectoryServiceWrapper();
 
     /**
      * Get a descriptive name for the archive (e.g. "RomFS", "SaveData", etc.)
diff --git a/src/core/hle/service/filesystem/fsp_ldr.cpp b/src/core/hle/service/filesystem/fsp_ldr.cpp
index 0ab9c2606..fb487d5bc 100644
--- a/src/core/hle/service/filesystem/fsp_ldr.cpp
+++ b/src/core/hle/service/filesystem/fsp_ldr.cpp
@@ -19,4 +19,6 @@ FSP_LDR::FSP_LDR() : ServiceFramework{"fsp:ldr"} {
     RegisterHandlers(functions);
 }
 
+FSP_LDR::~FSP_LDR() = default;
+
 } // namespace Service::FileSystem
diff --git a/src/core/hle/service/filesystem/fsp_ldr.h b/src/core/hle/service/filesystem/fsp_ldr.h
index fa8e11b4c..8210b7729 100644
--- a/src/core/hle/service/filesystem/fsp_ldr.h
+++ b/src/core/hle/service/filesystem/fsp_ldr.h
@@ -11,6 +11,7 @@ namespace Service::FileSystem {
 class FSP_LDR final : public ServiceFramework<FSP_LDR> {
 public:
     explicit FSP_LDR();
+    ~FSP_LDR() override;
 };
 
 } // namespace Service::FileSystem
diff --git a/src/core/hle/service/filesystem/fsp_pr.cpp b/src/core/hle/service/filesystem/fsp_pr.cpp
index 32b0ae454..378201610 100644
--- a/src/core/hle/service/filesystem/fsp_pr.cpp
+++ b/src/core/hle/service/filesystem/fsp_pr.cpp
@@ -20,4 +20,6 @@ FSP_PR::FSP_PR() : ServiceFramework{"fsp:pr"} {
     RegisterHandlers(functions);
 }
 
+FSP_PR::~FSP_PR() = default;
+
 } // namespace Service::FileSystem
diff --git a/src/core/hle/service/filesystem/fsp_pr.h b/src/core/hle/service/filesystem/fsp_pr.h
index 62edcd08a..556ae5ce9 100644
--- a/src/core/hle/service/filesystem/fsp_pr.h
+++ b/src/core/hle/service/filesystem/fsp_pr.h
@@ -11,6 +11,7 @@ namespace Service::FileSystem {
 class FSP_PR final : public ServiceFramework<FSP_PR> {
 public:
     explicit FSP_PR();
+    ~FSP_PR() override;
 };
 
 } // namespace Service::FileSystem
diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp
index 3f8ff67e8..cabaf5a55 100644
--- a/src/core/hle/service/filesystem/fsp_srv.cpp
+++ b/src/core/hle/service/filesystem/fsp_srv.cpp
@@ -520,6 +520,8 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") {
     RegisterHandlers(functions);
 }
 
+FSP_SRV::~FSP_SRV() = default;
+
 void FSP_SRV::Initialize(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_FS, "(STUBBED) called");
 
diff --git a/src/core/hle/service/filesystem/fsp_srv.h b/src/core/hle/service/filesystem/fsp_srv.h
index 2b5c21abb..4aa0358cb 100644
--- a/src/core/hle/service/filesystem/fsp_srv.h
+++ b/src/core/hle/service/filesystem/fsp_srv.h
@@ -16,7 +16,7 @@ namespace Service::FileSystem {
 class FSP_SRV final : public ServiceFramework<FSP_SRV> {
 public:
     explicit FSP_SRV();
-    ~FSP_SRV() = default;
+    ~FSP_SRV() override;
 
 private:
     void Initialize(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/friend/friend.cpp b/src/core/hle/service/friend/friend.cpp
index f2b0e509a..d9225d624 100644
--- a/src/core/hle/service/friend/friend.cpp
+++ b/src/core/hle/service/friend/friend.cpp
@@ -118,6 +118,8 @@ void Module::Interface::CreateFriendService(Kernel::HLERequestContext& ctx) {
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
     : ServiceFramework(name), module(std::move(module)) {}
 
+Module::Interface::~Interface() = default;
+
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     auto module = std::make_shared<Module>();
     std::make_shared<Friend>(module, "friend:a")->InstallAsService(service_manager);
diff --git a/src/core/hle/service/friend/friend.h b/src/core/hle/service/friend/friend.h
index c1b36518a..e762840cb 100644
--- a/src/core/hle/service/friend/friend.h
+++ b/src/core/hle/service/friend/friend.h
@@ -13,6 +13,7 @@ public:
     class Interface : public ServiceFramework<Interface> {
     public:
         explicit Interface(std::shared_ptr<Module> module, const char* name);
+        ~Interface() override;
 
         void CreateFriendService(Kernel::HLERequestContext& ctx);
 
diff --git a/src/core/hle/service/friend/interface.cpp b/src/core/hle/service/friend/interface.cpp
index 27c6a09e2..5a6840af5 100644
--- a/src/core/hle/service/friend/interface.cpp
+++ b/src/core/hle/service/friend/interface.cpp
@@ -16,4 +16,6 @@ Friend::Friend(std::shared_ptr<Module> module, const char* name)
     RegisterHandlers(functions);
 }
 
+Friend::~Friend() = default;
+
 } // namespace Service::Friend
diff --git a/src/core/hle/service/friend/interface.h b/src/core/hle/service/friend/interface.h
index 89dae8471..1963def39 100644
--- a/src/core/hle/service/friend/interface.h
+++ b/src/core/hle/service/friend/interface.h
@@ -11,6 +11,7 @@ namespace Service::Friend {
 class Friend final : public Module::Interface {
 public:
     explicit Friend(std::shared_ptr<Module> module, const char* name);
+    ~Friend() override;
 };
 
 } // namespace Service::Friend
diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp
index 0d31abe8b..a8e0c869f 100644
--- a/src/core/hle/service/hid/hid.cpp
+++ b/src/core/hle/service/hid/hid.cpp
@@ -2,7 +2,6 @@
 // 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/core_timing.h"
@@ -78,7 +77,7 @@ private:
         SharedMemory mem{};
         std::memcpy(&mem, shared_mem->GetPointer(), sizeof(SharedMemory));
 
-        if (is_device_reload_pending.exchange(false))
+        if (Settings::values.is_device_reload_pending.exchange(false))
             LoadInputDevices();
 
         // Set up controllers as neon red+blue Joy-Con attached to console
@@ -267,7 +266,6 @@ private:
     CoreTiming::EventType* pad_update_event;
 
     // Stored input state info
-    std::atomic<bool> is_device_reload_pending{true};
     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;
@@ -797,7 +795,9 @@ public:
     }
 };
 
-void ReloadInputDevices() {}
+void ReloadInputDevices() {
+    Settings::values.is_device_reload_pending.store(true);
+}
 
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     std::make_shared<Hid>()->InstallAsService(service_manager);
diff --git a/src/core/hle/service/hid/irs.cpp b/src/core/hle/service/hid/irs.cpp
index aaf311912..e587ad0d8 100644
--- a/src/core/hle/service/hid/irs.cpp
+++ b/src/core/hle/service/hid/irs.cpp
@@ -33,6 +33,8 @@ IRS::IRS() : ServiceFramework{"irs"} {
     RegisterHandlers(functions);
 }
 
+IRS::~IRS() = default;
+
 IRS_SYS::IRS_SYS() : ServiceFramework{"irs:sys"} {
     // clang-format off
     static const FunctionInfo functions[] = {
@@ -46,4 +48,6 @@ IRS_SYS::IRS_SYS() : ServiceFramework{"irs:sys"} {
     RegisterHandlers(functions);
 }
 
+IRS_SYS::~IRS_SYS() = default;
+
 } // namespace Service::HID
diff --git a/src/core/hle/service/hid/irs.h b/src/core/hle/service/hid/irs.h
index a8be701c7..6fb16b45d 100644
--- a/src/core/hle/service/hid/irs.h
+++ b/src/core/hle/service/hid/irs.h
@@ -11,11 +11,13 @@ namespace Service::HID {
 class IRS final : public ServiceFramework<IRS> {
 public:
     explicit IRS();
+    ~IRS() override;
 };
 
 class IRS_SYS final : public ServiceFramework<IRS_SYS> {
 public:
     explicit IRS_SYS();
+    ~IRS_SYS() override;
 };
 
 } // namespace Service::HID
diff --git a/src/core/hle/service/hid/xcd.cpp b/src/core/hle/service/hid/xcd.cpp
index 49f733f60..c8e9125f6 100644
--- a/src/core/hle/service/hid/xcd.cpp
+++ b/src/core/hle/service/hid/xcd.cpp
@@ -34,4 +34,6 @@ XCD_SYS::XCD_SYS() : ServiceFramework{"xcd:sys"} {
     RegisterHandlers(functions);
 }
 
+XCD_SYS::~XCD_SYS() = default;
+
 } // namespace Service::HID
diff --git a/src/core/hle/service/hid/xcd.h b/src/core/hle/service/hid/xcd.h
index 232a044df..fd506d303 100644
--- a/src/core/hle/service/hid/xcd.h
+++ b/src/core/hle/service/hid/xcd.h
@@ -11,6 +11,7 @@ namespace Service::HID {
 class XCD_SYS final : public ServiceFramework<XCD_SYS> {
 public:
     explicit XCD_SYS();
+    ~XCD_SYS() override;
 };
 
 } // namespace Service::HID
diff --git a/src/core/hle/service/nfp/nfp.cpp b/src/core/hle/service/nfp/nfp.cpp
index 4f7543af5..f8d2127d9 100644
--- a/src/core/hle/service/nfp/nfp.cpp
+++ b/src/core/hle/service/nfp/nfp.cpp
@@ -14,6 +14,8 @@ namespace Service::NFP {
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
     : ServiceFramework(name), module(std::move(module)) {}
 
+Module::Interface::~Interface() = default;
+
 class IUser final : public ServiceFramework<IUser> {
 public:
     IUser() : ServiceFramework("IUser") {
diff --git a/src/core/hle/service/nfp/nfp.h b/src/core/hle/service/nfp/nfp.h
index 0cd7be3d5..77df343c4 100644
--- a/src/core/hle/service/nfp/nfp.h
+++ b/src/core/hle/service/nfp/nfp.h
@@ -13,6 +13,7 @@ public:
     class Interface : public ServiceFramework<Interface> {
     public:
         explicit Interface(std::shared_ptr<Module> module, const char* name);
+        ~Interface() override;
 
         void CreateUserInterface(Kernel::HLERequestContext& ctx);
 
diff --git a/src/core/hle/service/nfp/nfp_user.cpp b/src/core/hle/service/nfp/nfp_user.cpp
index b608fe693..784a87c1b 100644
--- a/src/core/hle/service/nfp/nfp_user.cpp
+++ b/src/core/hle/service/nfp/nfp_user.cpp
@@ -14,4 +14,6 @@ NFP_User::NFP_User(std::shared_ptr<Module> module)
     RegisterHandlers(functions);
 }
 
+NFP_User::~NFP_User() = default;
+
 } // namespace Service::NFP
diff --git a/src/core/hle/service/nfp/nfp_user.h b/src/core/hle/service/nfp/nfp_user.h
index 700043114..65d9aaf48 100644
--- a/src/core/hle/service/nfp/nfp_user.h
+++ b/src/core/hle/service/nfp/nfp_user.h
@@ -11,6 +11,7 @@ namespace Service::NFP {
 class NFP_User final : public Module::Interface {
 public:
     explicit NFP_User(std::shared_ptr<Module> module);
+    ~NFP_User() override;
 };
 
 } // namespace Service::NFP
diff --git a/src/core/hle/service/ns/pl_u.cpp b/src/core/hle/service/ns/pl_u.cpp
index 878bbe439..ac0eaaa8f 100644
--- a/src/core/hle/service/ns/pl_u.cpp
+++ b/src/core/hle/service/ns/pl_u.cpp
@@ -2,14 +2,30 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
+#include <cstring>
+#include <vector>
+
+#include <FontChineseSimplified.h>
+#include <FontChineseTraditional.h>
+#include <FontExtendedChineseSimplified.h>
+#include <FontKorean.h>
+#include <FontNintendoExtended.h>
+#include <FontStandard.h>
+
+#include "common/assert.h"
 #include "common/common_paths.h"
+#include "common/common_types.h"
 #include "common/file_util.h"
+#include "common/logging/log.h"
+#include "common/swap.h"
 #include "core/core.h"
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/nca_metadata.h"
 #include "core/file_sys/registered_cache.h"
 #include "core/file_sys/romfs.h"
 #include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/shared_memory.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/hle/service/ns/pl_u.h"
 
@@ -28,49 +44,41 @@ struct FontRegion {
     u32 size;
 };
 
-static constexpr std::array<std::pair<FontArchives, const char*>, 7> SHARED_FONTS{
+constexpr std::array<std::pair<FontArchives, const char*>, 7> SHARED_FONTS{
     std::make_pair(FontArchives::Standard, "nintendo_udsg-r_std_003.bfttf"),
     std::make_pair(FontArchives::ChineseSimple, "nintendo_udsg-r_org_zh-cn_003.bfttf"),
     std::make_pair(FontArchives::ChineseSimple, "nintendo_udsg-r_ext_zh-cn_003.bfttf"),
     std::make_pair(FontArchives::ChineseTraditional, "nintendo_udjxh-db_zh-tw_003.bfttf"),
     std::make_pair(FontArchives::Korean, "nintendo_udsg-r_ko_003.bfttf"),
     std::make_pair(FontArchives::Extension, "nintendo_ext_003.bfttf"),
-    std::make_pair(FontArchives::Extension, "nintendo_ext2_003.bfttf")};
+    std::make_pair(FontArchives::Extension, "nintendo_ext2_003.bfttf"),
+};
 
-static constexpr std::array<const char*, 7> SHARED_FONTS_TTF{"FontStandard.ttf",
-                                                             "FontChineseSimplified.ttf",
-                                                             "FontExtendedChineseSimplified.ttf",
-                                                             "FontChineseTraditional.ttf",
-                                                             "FontKorean.ttf",
-                                                             "FontNintendoExtended.ttf",
-                                                             "FontNintendoExtended2.ttf"};
+constexpr std::array<const char*, 7> SHARED_FONTS_TTF{
+    "FontStandard.ttf",
+    "FontChineseSimplified.ttf",
+    "FontExtendedChineseSimplified.ttf",
+    "FontChineseTraditional.ttf",
+    "FontKorean.ttf",
+    "FontNintendoExtended.ttf",
+    "FontNintendoExtended2.ttf",
+};
 
 // The below data is specific to shared font data dumped from Switch on f/w 2.2
 // Virtual address and offsets/sizes likely will vary by dump
-static constexpr VAddr SHARED_FONT_MEM_VADDR{0x00000009d3016000ULL};
-static constexpr u32 EXPECTED_RESULT{
-    0x7f9a0218}; // What we expect the decrypted bfttf first 4 bytes to be
-static constexpr u32 EXPECTED_MAGIC{
-    0x36f81a1e}; // What we expect the encrypted bfttf first 4 bytes to be
-static constexpr u64 SHARED_FONT_MEM_SIZE{0x1100000};
-static constexpr FontRegion EMPTY_REGION{0, 0};
-std::vector<FontRegion>
-    SHARED_FONT_REGIONS{}; // Automatically populated based on shared_fonts dump or system archives
-
-const FontRegion& GetSharedFontRegion(size_t index) {
-    if (index >= SHARED_FONT_REGIONS.size() || SHARED_FONT_REGIONS.empty()) {
-        // No font fallback
-        return EMPTY_REGION;
-    }
-    return SHARED_FONT_REGIONS.at(index);
-}
+constexpr VAddr SHARED_FONT_MEM_VADDR{0x00000009d3016000ULL};
+constexpr u32 EXPECTED_RESULT{0x7f9a0218}; // What we expect the decrypted bfttf first 4 bytes to be
+constexpr u32 EXPECTED_MAGIC{0x36f81a1e};  // What we expect the encrypted bfttf first 4 bytes to be
+constexpr u64 SHARED_FONT_MEM_SIZE{0x1100000};
+constexpr FontRegion EMPTY_REGION{0, 0};
 
 enum class LoadState : u32 {
     Loading = 0,
     Done = 1,
 };
 
-void DecryptSharedFont(const std::vector<u32>& input, std::vector<u8>& output, size_t& offset) {
+static void DecryptSharedFont(const std::vector<u32>& input, std::vector<u8>& output,
+                              size_t& offset) {
     ASSERT_MSG(offset + (input.size() * sizeof(u32)) < SHARED_FONT_MEM_SIZE,
                "Shared fonts exceeds 17mb!");
     ASSERT_MSG(input[0] == EXPECTED_MAGIC, "Failed to derive key, unexpected magic number");
@@ -97,28 +105,52 @@ static void EncryptSharedFont(const std::vector<u8>& input, std::vector<u8>& out
     offset += input.size() + (sizeof(u32) * 2);
 }
 
+// Helper function to make BuildSharedFontsRawRegions a bit nicer
 static u32 GetU32Swapped(const u8* data) {
     u32 value;
     std::memcpy(&value, data, sizeof(value));
-    return Common::swap32(value); // Helper function to make BuildSharedFontsRawRegions a bit nicer
+    return Common::swap32(value);
 }
 
-void BuildSharedFontsRawRegions(const std::vector<u8>& input) {
-    unsigned cur_offset = 0; // As we can derive the xor key we can just populate the offsets based
-                             // on the shared memory dump
-    for (size_t i = 0; i < SHARED_FONTS.size(); i++) {
-        // Out of shared fonts/Invalid font
-        if (GetU32Swapped(input.data() + cur_offset) != EXPECTED_RESULT)
-            break;
-        const u32 KEY = GetU32Swapped(input.data() + cur_offset) ^
-                        EXPECTED_MAGIC; // Derive key withing inverse xor
-        const u32 SIZE = GetU32Swapped(input.data() + cur_offset + 4) ^ KEY;
-        SHARED_FONT_REGIONS.push_back(FontRegion{cur_offset + 8, SIZE});
-        cur_offset += SIZE + 8;
+struct PL_U::Impl {
+    const FontRegion& GetSharedFontRegion(size_t index) const {
+        if (index >= shared_font_regions.size() || shared_font_regions.empty()) {
+            // No font fallback
+            return EMPTY_REGION;
+        }
+        return shared_font_regions.at(index);
     }
-}
 
-PL_U::PL_U() : ServiceFramework("pl:u") {
+    void BuildSharedFontsRawRegions(const std::vector<u8>& input) {
+        // As we can derive the xor key we can just populate the offsets
+        // based on the shared memory dump
+        unsigned cur_offset = 0;
+
+        for (size_t i = 0; i < SHARED_FONTS.size(); i++) {
+            // Out of shared fonts/invalid font
+            if (GetU32Swapped(input.data() + cur_offset) != EXPECTED_RESULT) {
+                break;
+            }
+
+            // Derive key withing inverse xor
+            const u32 KEY = GetU32Swapped(input.data() + cur_offset) ^ EXPECTED_MAGIC;
+            const u32 SIZE = GetU32Swapped(input.data() + cur_offset + 4) ^ KEY;
+            shared_font_regions.push_back(FontRegion{cur_offset + 8, SIZE});
+            cur_offset += SIZE + 8;
+        }
+    }
+
+    /// Handle to shared memory region designated for a shared font
+    Kernel::SharedPtr<Kernel::SharedMemory> shared_font_mem;
+
+    /// Backing memory for the shared font data
+    std::shared_ptr<std::vector<u8>> shared_font;
+
+    // Automatically populated based on shared_fonts dump or system archives.
+    std::vector<FontRegion> shared_font_regions;
+};
+
+PL_U::PL_U() : ServiceFramework("pl:u"), impl{std::make_unique<Impl>()} {
     static const FunctionInfo functions[] = {
         {0, &PL_U::RequestLoad, "RequestLoad"},
         {1, &PL_U::GetLoadState, "GetLoadState"},
@@ -134,7 +166,7 @@ PL_U::PL_U() : ServiceFramework("pl:u") {
     // Rebuild shared fonts from data ncas
     if (nand->HasEntry(static_cast<u64>(FontArchives::Standard),
                        FileSys::ContentRecordType::Data)) {
-        shared_font = std::make_shared<std::vector<u8>>(SHARED_FONT_MEM_SIZE);
+        impl->shared_font = std::make_shared<std::vector<u8>>(SHARED_FONT_MEM_SIZE);
         for (auto font : SHARED_FONTS) {
             const auto nca =
                 nand->GetEntry(static_cast<u64>(font.first), FileSys::ContentRecordType::Data);
@@ -170,12 +202,12 @@ PL_U::PL_U() : ServiceFramework("pl:u") {
                 static_cast<u32>(offset + 8),
                 static_cast<u32>((font_data_u32.size() * sizeof(u32)) -
                                  8)}; // Font offset and size do not account for the header
-            DecryptSharedFont(font_data_u32, *shared_font, offset);
-            SHARED_FONT_REGIONS.push_back(region);
+            DecryptSharedFont(font_data_u32, *impl->shared_font, offset);
+            impl->shared_font_regions.push_back(region);
         }
 
     } else {
-        shared_font = std::make_shared<std::vector<u8>>(
+        impl->shared_font = std::make_shared<std::vector<u8>>(
             SHARED_FONT_MEM_SIZE); // Shared memory needs to always be allocated and a fixed size
 
         const std::string user_path = FileUtil::GetUserPath(FileUtil::UserPath::SysDataDir);
@@ -199,8 +231,8 @@ PL_U::PL_U() : ServiceFramework("pl:u") {
                         static_cast<u32>(offset + 8),
                         static_cast<u32>(ttf_bytes.size())}; // Font offset and size do not account
                                                              // for the header
-                    EncryptSharedFont(ttf_bytes, *shared_font, offset);
-                    SHARED_FONT_REGIONS.push_back(region);
+                    EncryptSharedFont(ttf_bytes, *impl->shared_font, offset);
+                    impl->shared_font_regions.push_back(region);
                 } else {
                     LOG_WARNING(Service_NS, "Unable to load font: {}", font_ttf);
                 }
@@ -215,14 +247,33 @@ PL_U::PL_U() : ServiceFramework("pl:u") {
         if (file.IsOpen()) {
             // Read shared font data
             ASSERT(file.GetSize() == SHARED_FONT_MEM_SIZE);
-            file.ReadBytes(shared_font->data(), shared_font->size());
-            BuildSharedFontsRawRegions(*shared_font);
+            file.ReadBytes(impl->shared_font->data(), impl->shared_font->size());
+            impl->BuildSharedFontsRawRegions(*impl->shared_font);
         } else {
-            LOG_WARNING(Service_NS, "Unable to load shared font: {}", filepath);
+            LOG_WARNING(Service_NS,
+                        "Shared Font file missing. Loading open source replacement from memory");
+
+            const std::vector<std::vector<u8>> open_source_shared_fonts_ttf = {
+                {std::begin(FontChineseSimplified), std::end(FontChineseSimplified)},
+                {std::begin(FontChineseTraditional), std::end(FontChineseTraditional)},
+                {std::begin(FontExtendedChineseSimplified),
+                 std::end(FontExtendedChineseSimplified)},
+                {std::begin(FontNintendoExtended), std::end(FontNintendoExtended)},
+                {std::begin(FontStandard), std::end(FontStandard)},
+            };
+
+            for (const std::vector<u8>& font_ttf : open_source_shared_fonts_ttf) {
+                const FontRegion region{static_cast<u32>(offset + 8),
+                                        static_cast<u32>(font_ttf.size())};
+                EncryptSharedFont(font_ttf, *impl->shared_font, offset);
+                impl->shared_font_regions.push_back(region);
+            }
         }
     }
 }
 
+PL_U::~PL_U() = default;
+
 void PL_U::RequestLoad(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const u32 shared_font_type{rp.Pop<u32>()};
@@ -249,7 +300,7 @@ void PL_U::GetSize(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_NS, "called, font_id={}", font_id);
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
-    rb.Push<u32>(GetSharedFontRegion(font_id).size);
+    rb.Push<u32>(impl->GetSharedFontRegion(font_id).size);
 }
 
 void PL_U::GetSharedMemoryAddressOffset(Kernel::HLERequestContext& ctx) {
@@ -259,17 +310,18 @@ void PL_U::GetSharedMemoryAddressOffset(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_NS, "called, font_id={}", font_id);
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
-    rb.Push<u32>(GetSharedFontRegion(font_id).offset);
+    rb.Push<u32>(impl->GetSharedFontRegion(font_id).offset);
 }
 
 void PL_U::GetSharedMemoryNativeHandle(Kernel::HLERequestContext& ctx) {
     // Map backing memory for the font data
-    Core::CurrentProcess()->vm_manager.MapMemoryBlock(
-        SHARED_FONT_MEM_VADDR, shared_font, 0, SHARED_FONT_MEM_SIZE, Kernel::MemoryState::Shared);
+    Core::CurrentProcess()->vm_manager.MapMemoryBlock(SHARED_FONT_MEM_VADDR, impl->shared_font, 0,
+                                                      SHARED_FONT_MEM_SIZE,
+                                                      Kernel::MemoryState::Shared);
 
     // Create shared font memory object
     auto& kernel = Core::System::GetInstance().Kernel();
-    shared_font_mem = Kernel::SharedMemory::Create(
+    impl->shared_font_mem = Kernel::SharedMemory::Create(
         kernel, Core::CurrentProcess(), SHARED_FONT_MEM_SIZE, Kernel::MemoryPermission::ReadWrite,
         Kernel::MemoryPermission::Read, SHARED_FONT_MEM_VADDR, Kernel::MemoryRegion::BASE,
         "PL_U:shared_font_mem");
@@ -277,7 +329,7 @@ void PL_U::GetSharedMemoryNativeHandle(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_NS, "called");
     IPC::ResponseBuilder rb{ctx, 2, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushCopyObjects(shared_font_mem);
+    rb.PushCopyObjects(impl->shared_font_mem);
 }
 
 void PL_U::GetSharedFontInOrderOfPriority(Kernel::HLERequestContext& ctx) {
@@ -290,9 +342,9 @@ void PL_U::GetSharedFontInOrderOfPriority(Kernel::HLERequestContext& ctx) {
     std::vector<u32> font_sizes;
 
     // TODO(ogniK): Have actual priority order
-    for (size_t i = 0; i < SHARED_FONT_REGIONS.size(); i++) {
+    for (size_t i = 0; i < impl->shared_font_regions.size(); i++) {
         font_codes.push_back(static_cast<u32>(i));
-        auto region = GetSharedFontRegion(i);
+        auto region = impl->GetSharedFontRegion(i);
         font_offsets.push_back(region.offset);
         font_sizes.push_back(region.size);
     }
diff --git a/src/core/hle/service/ns/pl_u.h b/src/core/hle/service/ns/pl_u.h
index fcc2acab7..253f26a2a 100644
--- a/src/core/hle/service/ns/pl_u.h
+++ b/src/core/hle/service/ns/pl_u.h
@@ -5,7 +5,6 @@
 #pragma once
 
 #include <memory>
-#include "core/hle/kernel/shared_memory.h"
 #include "core/hle/service/service.h"
 
 namespace Service::NS {
@@ -13,7 +12,7 @@ namespace Service::NS {
 class PL_U final : public ServiceFramework<PL_U> {
 public:
     PL_U();
-    ~PL_U() = default;
+    ~PL_U() override;
 
 private:
     void RequestLoad(Kernel::HLERequestContext& ctx);
@@ -23,11 +22,8 @@ private:
     void GetSharedMemoryNativeHandle(Kernel::HLERequestContext& ctx);
     void GetSharedFontInOrderOfPriority(Kernel::HLERequestContext& ctx);
 
-    /// Handle to shared memory region designated for a shared font
-    Kernel::SharedPtr<Kernel::SharedMemory> shared_font_mem;
-
-    /// Backing memory for the shared font data
-    std::shared_ptr<std::vector<u8>> shared_font;
+    struct Impl;
+    std::unique_ptr<Impl> impl;
 };
 
 } // namespace Service::NS
diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp
index 0b37098e1..92acc57b1 100644
--- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp
@@ -13,6 +13,9 @@
 
 namespace Service::Nvidia::Devices {
 
+nvdisp_disp0::nvdisp_disp0(std::shared_ptr<nvmap> nvmap_dev) : nvmap_dev(std::move(nvmap_dev)) {}
+nvdisp_disp0 ::~nvdisp_disp0() = default;
+
 u32 nvdisp_disp0::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h
index 6f0697b58..a45086e45 100644
--- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h
+++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h
@@ -17,8 +17,8 @@ class nvmap;
 
 class nvdisp_disp0 final : public nvdevice {
 public:
-    explicit nvdisp_disp0(std::shared_ptr<nvmap> nvmap_dev) : nvmap_dev(std::move(nvmap_dev)) {}
-    ~nvdisp_disp0() = default;
+    explicit nvdisp_disp0(std::shared_ptr<nvmap> nvmap_dev);
+    ~nvdisp_disp0();
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 
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 75487c4e8..25d5a93fa 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
@@ -3,6 +3,8 @@
 // Refer to the license.txt file included.
 
 #include <cstring>
+#include <utility>
+
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/core.h"
@@ -14,6 +16,9 @@
 
 namespace Service::Nvidia::Devices {
 
+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;
+
 u32 nvhost_as_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}",
               command.raw, input.size(), output.size());
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h
index 9f8999d9c..eb14b1da8 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h
+++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h
@@ -6,7 +6,6 @@
 
 #include <memory>
 #include <unordered_map>
-#include <utility>
 #include <vector>
 #include "common/common_types.h"
 #include "common/swap.h"
@@ -18,8 +17,8 @@ class nvmap;
 
 class nvhost_as_gpu final : public nvdevice {
 public:
-    explicit nvhost_as_gpu(std::shared_ptr<nvmap> nvmap_dev) : nvmap_dev(std::move(nvmap_dev)) {}
-    ~nvhost_as_gpu() override = default;
+    explicit nvhost_as_gpu(std::shared_ptr<nvmap> nvmap_dev);
+    ~nvhost_as_gpu() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp
index 5685eb2be..b39fb9ef9 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp
@@ -11,6 +11,9 @@
 
 namespace Service::Nvidia::Devices {
 
+nvhost_ctrl::nvhost_ctrl() = default;
+nvhost_ctrl::~nvhost_ctrl() = default;
+
 u32 nvhost_ctrl::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}",
               command.raw, input.size(), output.size());
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h
index 6b496e9fe..6d0de2212 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h
+++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h
@@ -13,8 +13,8 @@ namespace Service::Nvidia::Devices {
 
 class nvhost_ctrl final : public nvdevice {
 public:
-    nvhost_ctrl() = default;
-    ~nvhost_ctrl() override = default;
+    nvhost_ctrl();
+    ~nvhost_ctrl() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp
index ae421247d..7a88ae029 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp
@@ -9,6 +9,9 @@
 
 namespace Service::Nvidia::Devices {
 
+nvhost_ctrl_gpu::nvhost_ctrl_gpu() = default;
+nvhost_ctrl_gpu::~nvhost_ctrl_gpu() = default;
+
 u32 nvhost_ctrl_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}",
               command.raw, input.size(), output.size());
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h
index f09113e67..3bbf028ad 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h
+++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h
@@ -13,8 +13,8 @@ namespace Service::Nvidia::Devices {
 
 class nvhost_ctrl_gpu final : public nvdevice {
 public:
-    nvhost_ctrl_gpu() = default;
-    ~nvhost_ctrl_gpu() override = default;
+    nvhost_ctrl_gpu();
+    ~nvhost_ctrl_gpu() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp
index 4cdf7f613..874d5e1c3 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp
@@ -8,11 +8,15 @@
 #include "core/core.h"
 #include "core/hle/service/nvdrv/devices/nvhost_gpu.h"
 #include "core/memory.h"
+#include "video_core/command_processor.h"
 #include "video_core/gpu.h"
 #include "video_core/memory_manager.h"
 
 namespace Service::Nvidia::Devices {
 
+nvhost_gpu::nvhost_gpu(std::shared_ptr<nvmap> nvmap_dev) : nvmap_dev(std::move(nvmap_dev)) {}
+nvhost_gpu::~nvhost_gpu() = default;
+
 u32 nvhost_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}",
               command.raw, input.size(), output.size());
@@ -134,17 +138,16 @@ u32 nvhost_gpu::SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& outp
     LOG_WARNING(Service_NVDRV, "(STUBBED) called, gpfifo={:X}, num_entries={:X}, flags={:X}",
                 params.address, params.num_entries, params.flags);
 
-    ASSERT_MSG(input.size() ==
-                   sizeof(IoctlSubmitGpfifo) + params.num_entries * sizeof(IoctlGpfifoEntry),
+    ASSERT_MSG(input.size() == sizeof(IoctlSubmitGpfifo) +
+                                   params.num_entries * sizeof(Tegra::CommandListHeader),
                "Incorrect input size");
 
-    std::vector<IoctlGpfifoEntry> entries(params.num_entries);
+    std::vector<Tegra::CommandListHeader> entries(params.num_entries);
     std::memcpy(entries.data(), &input[sizeof(IoctlSubmitGpfifo)],
-                params.num_entries * sizeof(IoctlGpfifoEntry));
-    for (auto entry : entries) {
-        Tegra::GPUVAddr va_addr = entry.Address();
-        Core::System::GetInstance().GPU().ProcessCommandList(va_addr, entry.sz);
-    }
+                params.num_entries * sizeof(Tegra::CommandListHeader));
+
+    Core::System::GetInstance().GPU().ProcessCommandLists(entries);
+
     params.fence_out.id = 0;
     params.fence_out.value = 0;
     std::memcpy(output.data(), &params, sizeof(IoctlSubmitGpfifo));
@@ -160,14 +163,12 @@ u32 nvhost_gpu::KickoffPB(const std::vector<u8>& input, std::vector<u8>& output)
     LOG_WARNING(Service_NVDRV, "(STUBBED) called, gpfifo={:X}, num_entries={:X}, flags={:X}",
                 params.address, params.num_entries, params.flags);
 
-    std::vector<IoctlGpfifoEntry> entries(params.num_entries);
+    std::vector<Tegra::CommandListHeader> entries(params.num_entries);
     Memory::ReadBlock(params.address, entries.data(),
-                      params.num_entries * sizeof(IoctlGpfifoEntry));
+                      params.num_entries * sizeof(Tegra::CommandListHeader));
+
+    Core::System::GetInstance().GPU().ProcessCommandLists(entries);
 
-    for (auto entry : entries) {
-        Tegra::GPUVAddr va_addr = entry.Address();
-        Core::System::GetInstance().GPU().ProcessCommandList(va_addr, entry.sz);
-    }
     params.fence_out.id = 0;
     params.fence_out.value = 0;
     std::memcpy(output.data(), &params, output.size());
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h
index 03b7356d0..62beb5c0c 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h
+++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h
@@ -10,7 +10,6 @@
 #include "common/common_types.h"
 #include "common/swap.h"
 #include "core/hle/service/nvdrv/devices/nvdevice.h"
-#include "video_core/memory_manager.h"
 
 namespace Service::Nvidia::Devices {
 
@@ -21,8 +20,8 @@ constexpr u32 NVGPU_IOCTL_CHANNEL_KICKOFF_PB(0x1b);
 
 class nvhost_gpu final : public nvdevice {
 public:
-    explicit nvhost_gpu(std::shared_ptr<nvmap> nvmap_dev) : nvmap_dev(std::move(nvmap_dev)) {}
-    ~nvhost_gpu() override = default;
+    explicit nvhost_gpu(std::shared_ptr<nvmap> nvmap_dev);
+    ~nvhost_gpu() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 
@@ -151,22 +150,6 @@ private:
     };
     static_assert(sizeof(IoctlAllocObjCtx) == 16, "IoctlAllocObjCtx is incorrect size");
 
-    struct IoctlGpfifoEntry {
-        u32_le entry0; // gpu_va_lo
-        union {
-            u32_le entry1; // gpu_va_hi | (unk_0x02 << 0x08) | (size << 0x0A) | (unk_0x01 << 0x1F)
-            BitField<0, 8, u32_le> gpu_va_hi;
-            BitField<8, 2, u32_le> unk1;
-            BitField<10, 21, u32_le> sz;
-            BitField<31, 1, u32_le> unk2;
-        };
-
-        Tegra::GPUVAddr Address() const {
-            return (static_cast<Tegra::GPUVAddr>(gpu_va_hi) << 32) | entry0;
-        }
-    };
-    static_assert(sizeof(IoctlGpfifoEntry) == 8, "IoctlGpfifoEntry is incorrect size");
-
     struct IoctlSubmitGpfifo {
         u64_le address;     // pointer to gpfifo entry structs
         u32_le num_entries; // number of fence objects being submitted
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
index 364619e67..46dbbc37c 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
@@ -10,6 +10,9 @@
 
 namespace Service::Nvidia::Devices {
 
+nvhost_nvdec::nvhost_nvdec() = default;
+nvhost_nvdec::~nvhost_nvdec() = default;
+
 u32 nvhost_nvdec::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}",
               command.raw, input.size(), output.size());
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h
index 6ad74421b..0e7b284f8 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h
+++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h
@@ -13,8 +13,8 @@ namespace Service::Nvidia::Devices {
 
 class nvhost_nvdec final : public nvdevice {
 public:
-    nvhost_nvdec() = default;
-    ~nvhost_nvdec() override = default;
+    nvhost_nvdec();
+    ~nvhost_nvdec() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp
index 51f01077b..c67f934f6 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp
@@ -10,6 +10,9 @@
 
 namespace Service::Nvidia::Devices {
 
+nvhost_nvjpg::nvhost_nvjpg() = default;
+nvhost_nvjpg::~nvhost_nvjpg() = default;
+
 u32 nvhost_nvjpg::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}",
               command.raw, input.size(), output.size());
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h
index 2b0eb43ee..89fd5e95e 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h
+++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h
@@ -13,8 +13,8 @@ namespace Service::Nvidia::Devices {
 
 class nvhost_nvjpg final : public nvdevice {
 public:
-    nvhost_nvjpg() = default;
-    ~nvhost_nvjpg() override = default;
+    nvhost_nvjpg();
+    ~nvhost_nvjpg() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
index fcb488d50..727b9fee4 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
@@ -10,6 +10,9 @@
 
 namespace Service::Nvidia::Devices {
 
+nvhost_vic::nvhost_vic() = default;
+nvhost_vic::~nvhost_vic() = default;
+
 u32 nvhost_vic::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}",
               command.raw, input.size(), output.size());
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.h b/src/core/hle/service/nvdrv/devices/nvhost_vic.h
index c7d681e52..fc24c3f9c 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_vic.h
+++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.h
@@ -13,8 +13,8 @@ namespace Service::Nvidia::Devices {
 
 class nvhost_vic final : public nvdevice {
 public:
-    nvhost_vic() = default;
-    ~nvhost_vic() override = default;
+    nvhost_vic();
+    ~nvhost_vic() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 
diff --git a/src/core/hle/service/nvdrv/devices/nvmap.cpp b/src/core/hle/service/nvdrv/devices/nvmap.cpp
index e9305bfb3..a2287cc1b 100644
--- a/src/core/hle/service/nvdrv/devices/nvmap.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvmap.cpp
@@ -11,6 +11,9 @@
 
 namespace Service::Nvidia::Devices {
 
+nvmap::nvmap() = default;
+nvmap::~nvmap() = default;
+
 VAddr nvmap::GetObjectAddress(u32 handle) const {
     auto object = GetObject(handle);
     ASSERT(object);
diff --git a/src/core/hle/service/nvdrv/devices/nvmap.h b/src/core/hle/service/nvdrv/devices/nvmap.h
index f2eec6409..396230c19 100644
--- a/src/core/hle/service/nvdrv/devices/nvmap.h
+++ b/src/core/hle/service/nvdrv/devices/nvmap.h
@@ -16,8 +16,8 @@ namespace Service::Nvidia::Devices {
 
 class nvmap final : public nvdevice {
 public:
-    nvmap() = default;
-    ~nvmap() override = default;
+    nvmap();
+    ~nvmap() override;
 
     /// Returns the allocated address of an nvmap object given its handle.
     VAddr GetObjectAddress(u32 handle) const;
diff --git a/src/core/hle/service/nvdrv/interface.cpp b/src/core/hle/service/nvdrv/interface.cpp
index 634ab9196..ac3859353 100644
--- a/src/core/hle/service/nvdrv/interface.cpp
+++ b/src/core/hle/service/nvdrv/interface.cpp
@@ -112,4 +112,6 @@ NVDRV::NVDRV(std::shared_ptr<Module> nvdrv, const char* name)
     query_event = Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "NVDRV::query_event");
 }
 
+NVDRV::~NVDRV() = default;
+
 } // namespace Service::Nvidia
diff --git a/src/core/hle/service/nvdrv/interface.h b/src/core/hle/service/nvdrv/interface.h
index 1c3529bb6..d340893c2 100644
--- a/src/core/hle/service/nvdrv/interface.h
+++ b/src/core/hle/service/nvdrv/interface.h
@@ -14,7 +14,7 @@ namespace Service::Nvidia {
 class NVDRV final : public ServiceFramework<NVDRV> {
 public:
     NVDRV(std::shared_ptr<Module> nvdrv, const char* name);
-    ~NVDRV() = default;
+    ~NVDRV();
 
 private:
     void Open(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/nvdrv/nvdrv.cpp b/src/core/hle/service/nvdrv/nvdrv.cpp
index 2de39822f..6e4b8f2c6 100644
--- a/src/core/hle/service/nvdrv/nvdrv.cpp
+++ b/src/core/hle/service/nvdrv/nvdrv.cpp
@@ -45,6 +45,8 @@ Module::Module() {
     devices["/dev/nvhost-vic"] = std::make_shared<Devices::nvhost_vic>();
 }
 
+Module::~Module() = default;
+
 u32 Module::Open(const std::string& device_name) {
     ASSERT_MSG(devices.find(device_name) != devices.end(), "Trying to open unknown device {}",
                device_name);
diff --git a/src/core/hle/service/nvdrv/nvdrv.h b/src/core/hle/service/nvdrv/nvdrv.h
index 99eb1128a..53564f696 100644
--- a/src/core/hle/service/nvdrv/nvdrv.h
+++ b/src/core/hle/service/nvdrv/nvdrv.h
@@ -30,7 +30,7 @@ static_assert(sizeof(IoctlFence) == 8, "IoctlFence has wrong size");
 class Module final {
 public:
     Module();
-    ~Module() = default;
+    ~Module();
 
     /// Returns a pointer to one of the available devices, identified by its name.
     template <typename T>
diff --git a/src/core/hle/service/nvdrv/nvmemp.cpp b/src/core/hle/service/nvdrv/nvmemp.cpp
index 0e8e21bad..b7b8b7a1b 100644
--- a/src/core/hle/service/nvdrv/nvmemp.cpp
+++ b/src/core/hle/service/nvdrv/nvmemp.cpp
@@ -16,6 +16,8 @@ NVMEMP::NVMEMP() : ServiceFramework("nvmemp") {
     RegisterHandlers(functions);
 }
 
+NVMEMP::~NVMEMP() = default;
+
 void NVMEMP::Cmd0(Kernel::HLERequestContext& ctx) {
     UNIMPLEMENTED();
 }
diff --git a/src/core/hle/service/nvdrv/nvmemp.h b/src/core/hle/service/nvdrv/nvmemp.h
index dfdcabf4a..5a4dfc1f9 100644
--- a/src/core/hle/service/nvdrv/nvmemp.h
+++ b/src/core/hle/service/nvdrv/nvmemp.h
@@ -11,7 +11,7 @@ namespace Service::Nvidia {
 class NVMEMP final : public ServiceFramework<NVMEMP> {
 public:
     NVMEMP();
-    ~NVMEMP() = default;
+    ~NVMEMP();
 
 private:
     void Cmd0(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/nvflinger/buffer_queue.cpp b/src/core/hle/service/nvflinger/buffer_queue.cpp
index 8d8962276..34f98fe5a 100644
--- a/src/core/hle/service/nvflinger/buffer_queue.cpp
+++ b/src/core/hle/service/nvflinger/buffer_queue.cpp
@@ -18,6 +18,8 @@ BufferQueue::BufferQueue(u32 id, u64 layer_id) : id(id), layer_id(layer_id) {
         Kernel::Event::Create(kernel, Kernel::ResetType::Sticky, "BufferQueue NativeHandle");
 }
 
+BufferQueue::~BufferQueue() = default;
+
 void BufferQueue::SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer) {
     Buffer buffer{};
     buffer.slot = slot;
diff --git a/src/core/hle/service/nvflinger/buffer_queue.h b/src/core/hle/service/nvflinger/buffer_queue.h
index db2e17c0c..17c81928a 100644
--- a/src/core/hle/service/nvflinger/buffer_queue.h
+++ b/src/core/hle/service/nvflinger/buffer_queue.h
@@ -46,7 +46,7 @@ public:
     };
 
     BufferQueue(u32 id, u64 layer_id);
-    ~BufferQueue() = default;
+    ~BufferQueue();
 
     enum class BufferTransformFlags : u32 {
         /// No transform flags are set
diff --git a/src/core/hle/service/nvflinger/nvflinger.cpp b/src/core/hle/service/nvflinger/nvflinger.cpp
index 06040da6f..7455ddd19 100644
--- a/src/core/hle/service/nvflinger/nvflinger.cpp
+++ b/src/core/hle/service/nvflinger/nvflinger.cpp
@@ -160,10 +160,13 @@ void NVFlinger::Compose() {
 }
 
 Layer::Layer(u64 id, std::shared_ptr<BufferQueue> queue) : id(id), buffer_queue(std::move(queue)) {}
+Layer::~Layer() = default;
 
 Display::Display(u64 id, std::string name) : id(id), name(std::move(name)) {
     auto& kernel = Core::System::GetInstance().Kernel();
     vsync_event = Kernel::Event::Create(kernel, Kernel::ResetType::Pulse, "Display VSync Event");
 }
 
+Display::~Display() = default;
+
 } // namespace Service::NVFlinger
diff --git a/src/core/hle/service/nvflinger/nvflinger.h b/src/core/hle/service/nvflinger/nvflinger.h
index f7112949f..3dc69e69b 100644
--- a/src/core/hle/service/nvflinger/nvflinger.h
+++ b/src/core/hle/service/nvflinger/nvflinger.h
@@ -26,7 +26,7 @@ class BufferQueue;
 
 struct Layer {
     Layer(u64 id, std::shared_ptr<BufferQueue> queue);
-    ~Layer() = default;
+    ~Layer();
 
     u64 id;
     std::shared_ptr<BufferQueue> buffer_queue;
@@ -34,7 +34,7 @@ struct Layer {
 
 struct Display {
     Display(u64 id, std::string name);
-    ~Display() = default;
+    ~Display();
 
     u64 id;
     std::string name;
diff --git a/src/core/hle/service/pctl/module.cpp b/src/core/hle/service/pctl/module.cpp
index 6cc3b1992..4fd185f69 100644
--- a/src/core/hle/service/pctl/module.cpp
+++ b/src/core/hle/service/pctl/module.cpp
@@ -142,6 +142,8 @@ void Module::Interface::CreateServiceWithoutInitialize(Kernel::HLERequestContext
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
     : ServiceFramework(name), module(std::move(module)) {}
 
+Module::Interface::~Interface() = default;
+
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     auto module = std::make_shared<Module>();
     std::make_shared<PCTL>(module, "pctl")->InstallAsService(service_manager);
diff --git a/src/core/hle/service/pctl/module.h b/src/core/hle/service/pctl/module.h
index e7d492760..3e449110d 100644
--- a/src/core/hle/service/pctl/module.h
+++ b/src/core/hle/service/pctl/module.h
@@ -13,6 +13,7 @@ public:
     class Interface : public ServiceFramework<Interface> {
     public:
         explicit Interface(std::shared_ptr<Module> module, const char* name);
+        ~Interface() override;
 
         void CreateService(Kernel::HLERequestContext& ctx);
         void CreateServiceWithoutInitialize(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/pctl/pctl.cpp b/src/core/hle/service/pctl/pctl.cpp
index de2741d66..af9d1433a 100644
--- a/src/core/hle/service/pctl/pctl.cpp
+++ b/src/core/hle/service/pctl/pctl.cpp
@@ -14,4 +14,6 @@ PCTL::PCTL(std::shared_ptr<Module> module, const char* name)
     };
     RegisterHandlers(functions);
 }
+
+PCTL::~PCTL() = default;
 } // namespace Service::PCTL
diff --git a/src/core/hle/service/pctl/pctl.h b/src/core/hle/service/pctl/pctl.h
index 8ddf69128..c33ea80b6 100644
--- a/src/core/hle/service/pctl/pctl.h
+++ b/src/core/hle/service/pctl/pctl.h
@@ -11,6 +11,7 @@ namespace Service::PCTL {
 class PCTL final : public Module::Interface {
 public:
     explicit PCTL(std::shared_ptr<Module> module, const char* name);
+    ~PCTL() override;
 };
 
 } // namespace Service::PCTL
diff --git a/src/core/hle/service/prepo/prepo.cpp b/src/core/hle/service/prepo/prepo.cpp
index 3c43b8d8c..6a9eccfb5 100644
--- a/src/core/hle/service/prepo/prepo.cpp
+++ b/src/core/hle/service/prepo/prepo.cpp
@@ -1,36 +1,47 @@
-#include <cinttypes>
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
 #include "common/logging/log.h"
 #include "core/hle/ipc_helpers.h"
-#include "core/hle/kernel/event.h"
 #include "core/hle/service/prepo/prepo.h"
+#include "core/hle/service/service.h"
 
 namespace Service::PlayReport {
-PlayReport::PlayReport(const char* name) : ServiceFramework(name) {
-    static const FunctionInfo functions[] = {
-        {10100, nullptr, "SaveReport"},
-        {10101, &PlayReport::SaveReportWithUser, "SaveReportWithUser"},
-        {10200, nullptr, "RequestImmediateTransmission"},
-        {10300, nullptr, "GetTransmissionStatus"},
-        {20100, nullptr, "SaveSystemReport"},
-        {20200, nullptr, "SetOperationMode"},
-        {20101, nullptr, "SaveSystemReportWithUser"},
-        {30100, nullptr, "ClearStorage"},
-        {40100, nullptr, "IsUserAgreementCheckEnabled"},
-        {40101, nullptr, "SetUserAgreementCheckEnabled"},
-        {90100, nullptr, "GetStorageUsage"},
-        {90200, nullptr, "GetStatistics"},
-        {90201, nullptr, "GetThroughputHistory"},
-        {90300, nullptr, "GetLastUploadError"},
-    };
-    RegisterHandlers(functions);
-};
 
-void PlayReport::SaveReportWithUser(Kernel::HLERequestContext& ctx) {
-    // TODO(ogniK): Do we want to add play report?
-    LOG_WARNING(Service_PREPO, "(STUBBED) called");
+class PlayReport final : public ServiceFramework<PlayReport> {
+public:
+    explicit PlayReport(const char* name) : ServiceFramework{name} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {10100, nullptr, "SaveReport"},
+            {10101, &PlayReport::SaveReportWithUser, "SaveReportWithUser"},
+            {10200, nullptr, "RequestImmediateTransmission"},
+            {10300, nullptr, "GetTransmissionStatus"},
+            {20100, nullptr, "SaveSystemReport"},
+            {20200, nullptr, "SetOperationMode"},
+            {20101, nullptr, "SaveSystemReportWithUser"},
+            {30100, nullptr, "ClearStorage"},
+            {40100, nullptr, "IsUserAgreementCheckEnabled"},
+            {40101, nullptr, "SetUserAgreementCheckEnabled"},
+            {90100, nullptr, "GetStorageUsage"},
+            {90200, nullptr, "GetStatistics"},
+            {90201, nullptr, "GetThroughputHistory"},
+            {90300, nullptr, "GetLastUploadError"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+
+private:
+    void SaveReportWithUser(Kernel::HLERequestContext& ctx) {
+        // TODO(ogniK): Do we want to add play report?
+        LOG_WARNING(Service_PREPO, "(STUBBED) called");
 
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
 };
 
 void InstallInterfaces(SM::ServiceManager& service_manager) {
diff --git a/src/core/hle/service/prepo/prepo.h b/src/core/hle/service/prepo/prepo.h
index f5a6aba6d..0e7b01331 100644
--- a/src/core/hle/service/prepo/prepo.h
+++ b/src/core/hle/service/prepo/prepo.h
@@ -4,22 +4,12 @@
 
 #pragma once
 
-#include <memory>
-#include <string>
-#include "core/hle/kernel/event.h"
-#include "core/hle/service/service.h"
+namespace Service::SM {
+class ServiceManager;
+}
 
 namespace Service::PlayReport {
 
-class PlayReport final : public ServiceFramework<PlayReport> {
-public:
-    explicit PlayReport(const char* name);
-    ~PlayReport() = default;
-
-private:
-    void SaveReportWithUser(Kernel::HLERequestContext& ctx);
-};
-
 void InstallInterfaces(SM::ServiceManager& service_manager);
 
 } // namespace Service::PlayReport
diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp
index 9d804652e..9bb7c7b26 100644
--- a/src/core/hle/service/service.cpp
+++ b/src/core/hle/service/service.cpp
@@ -74,8 +74,6 @@ using Kernel::SharedPtr;
 
 namespace Service {
 
-std::unordered_map<std::string, SharedPtr<ClientPort>> g_kernel_named_ports;
-
 /**
  * Creates a function string for logging, complete with the name (or header code, depending
  * on what's passed in) the port name, and all the cmd_buff arguments.
diff --git a/src/core/hle/service/set/set.cpp b/src/core/hle/service/set/set.cpp
index 92b0640e8..59eb20155 100644
--- a/src/core/hle/service/set/set.cpp
+++ b/src/core/hle/service/set/set.cpp
@@ -112,4 +112,6 @@ SET::SET() : ServiceFramework("set") {
     RegisterHandlers(functions);
 }
 
+SET::~SET() = default;
+
 } // namespace Service::Set
diff --git a/src/core/hle/service/set/set.h b/src/core/hle/service/set/set.h
index 669e740b7..5f0214359 100644
--- a/src/core/hle/service/set/set.h
+++ b/src/core/hle/service/set/set.h
@@ -33,7 +33,7 @@ LanguageCode GetLanguageCodeFromIndex(size_t idx);
 class SET final : public ServiceFramework<SET> {
 public:
     explicit SET();
-    ~SET() = default;
+    ~SET() override;
 
 private:
     void GetLanguageCode(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/set/set_cal.cpp b/src/core/hle/service/set/set_cal.cpp
index 7066ef725..5af356d10 100644
--- a/src/core/hle/service/set/set_cal.cpp
+++ b/src/core/hle/service/set/set_cal.cpp
@@ -44,4 +44,6 @@ SET_CAL::SET_CAL() : ServiceFramework("set:cal") {
     RegisterHandlers(functions);
 }
 
+SET_CAL::~SET_CAL() = default;
+
 } // namespace Service::Set
diff --git a/src/core/hle/service/set/set_cal.h b/src/core/hle/service/set/set_cal.h
index bb50336aa..583036eac 100644
--- a/src/core/hle/service/set/set_cal.h
+++ b/src/core/hle/service/set/set_cal.h
@@ -11,7 +11,7 @@ namespace Service::Set {
 class SET_CAL final : public ServiceFramework<SET_CAL> {
 public:
     explicit SET_CAL();
-    ~SET_CAL() = default;
+    ~SET_CAL();
 };
 
 } // namespace Service::Set
diff --git a/src/core/hle/service/set/set_fd.cpp b/src/core/hle/service/set/set_fd.cpp
index c9f938716..cac6af86d 100644
--- a/src/core/hle/service/set/set_fd.cpp
+++ b/src/core/hle/service/set/set_fd.cpp
@@ -20,4 +20,6 @@ SET_FD::SET_FD() : ServiceFramework("set:fd") {
     RegisterHandlers(functions);
 }
 
+SET_FD::~SET_FD() = default;
+
 } // namespace Service::Set
diff --git a/src/core/hle/service/set/set_fd.h b/src/core/hle/service/set/set_fd.h
index dbd850bc7..216e65f1f 100644
--- a/src/core/hle/service/set/set_fd.h
+++ b/src/core/hle/service/set/set_fd.h
@@ -11,7 +11,7 @@ namespace Service::Set {
 class SET_FD final : public ServiceFramework<SET_FD> {
 public:
     explicit SET_FD();
-    ~SET_FD() = default;
+    ~SET_FD() override;
 };
 
 } // namespace Service::Set
diff --git a/src/core/hle/service/sockets/bsd.cpp b/src/core/hle/service/sockets/bsd.cpp
index 3211a8346..4342f3b2d 100644
--- a/src/core/hle/service/sockets/bsd.cpp
+++ b/src/core/hle/service/sockets/bsd.cpp
@@ -109,6 +109,8 @@ BSD::BSD(const char* name) : ServiceFramework(name) {
     RegisterHandlers(functions);
 }
 
+BSD::~BSD() = default;
+
 BSDCFG::BSDCFG() : ServiceFramework{"bsdcfg"} {
     // clang-format off
     static const FunctionInfo functions[] = {
@@ -131,4 +133,6 @@ BSDCFG::BSDCFG() : ServiceFramework{"bsdcfg"} {
     RegisterHandlers(functions);
 }
 
+BSDCFG::~BSDCFG() = default;
+
 } // namespace Service::Sockets
diff --git a/src/core/hle/service/sockets/bsd.h b/src/core/hle/service/sockets/bsd.h
index c1da59b24..0fe0e65c6 100644
--- a/src/core/hle/service/sockets/bsd.h
+++ b/src/core/hle/service/sockets/bsd.h
@@ -12,7 +12,7 @@ namespace Service::Sockets {
 class BSD final : public ServiceFramework<BSD> {
 public:
     explicit BSD(const char* name);
-    ~BSD() = default;
+    ~BSD() override;
 
 private:
     void RegisterClient(Kernel::HLERequestContext& ctx);
@@ -29,6 +29,7 @@ private:
 class BSDCFG final : public ServiceFramework<BSDCFG> {
 public:
     explicit BSDCFG();
+    ~BSDCFG() override;
 };
 
 } // namespace Service::Sockets
diff --git a/src/core/hle/service/sockets/ethc.cpp b/src/core/hle/service/sockets/ethc.cpp
index d53c25eec..abbeb4c50 100644
--- a/src/core/hle/service/sockets/ethc.cpp
+++ b/src/core/hle/service/sockets/ethc.cpp
@@ -21,6 +21,8 @@ ETHC_C::ETHC_C() : ServiceFramework{"ethc:c"} {
     RegisterHandlers(functions);
 }
 
+ETHC_C::~ETHC_C() = default;
+
 ETHC_I::ETHC_I() : ServiceFramework{"ethc:i"} {
     // clang-format off
     static const FunctionInfo functions[] = {
@@ -35,4 +37,6 @@ ETHC_I::ETHC_I() : ServiceFramework{"ethc:i"} {
     RegisterHandlers(functions);
 }
 
+ETHC_I::~ETHC_I() = default;
+
 } // namespace Service::Sockets
diff --git a/src/core/hle/service/sockets/ethc.h b/src/core/hle/service/sockets/ethc.h
index 9a3c88100..da2c7f741 100644
--- a/src/core/hle/service/sockets/ethc.h
+++ b/src/core/hle/service/sockets/ethc.h
@@ -11,11 +11,13 @@ namespace Service::Sockets {
 class ETHC_C final : public ServiceFramework<ETHC_C> {
 public:
     explicit ETHC_C();
+    ~ETHC_C() override;
 };
 
 class ETHC_I final : public ServiceFramework<ETHC_I> {
 public:
     explicit ETHC_I();
+    ~ETHC_I() override;
 };
 
 } // namespace Service::Sockets
diff --git a/src/core/hle/service/sockets/nsd.cpp b/src/core/hle/service/sockets/nsd.cpp
index 8682dc2e0..e6d73065e 100644
--- a/src/core/hle/service/sockets/nsd.cpp
+++ b/src/core/hle/service/sockets/nsd.cpp
@@ -29,4 +29,6 @@ NSD::NSD(const char* name) : ServiceFramework(name) {
     RegisterHandlers(functions);
 }
 
+NSD::~NSD() = default;
+
 } // namespace Service::Sockets
diff --git a/src/core/hle/service/sockets/nsd.h b/src/core/hle/service/sockets/nsd.h
index 3b7edfc43..d842e3232 100644
--- a/src/core/hle/service/sockets/nsd.h
+++ b/src/core/hle/service/sockets/nsd.h
@@ -12,7 +12,7 @@ namespace Service::Sockets {
 class NSD final : public ServiceFramework<NSD> {
 public:
     explicit NSD(const char* name);
-    ~NSD() = default;
+    ~NSD() override;
 };
 
 } // namespace Service::Sockets
diff --git a/src/core/hle/service/sockets/sfdnsres.cpp b/src/core/hle/service/sockets/sfdnsres.cpp
index d235c4cfd..13ab1d31e 100644
--- a/src/core/hle/service/sockets/sfdnsres.cpp
+++ b/src/core/hle/service/sockets/sfdnsres.cpp
@@ -34,4 +34,6 @@ SFDNSRES::SFDNSRES() : ServiceFramework("sfdnsres") {
     RegisterHandlers(functions);
 }
 
+SFDNSRES::~SFDNSRES() = default;
+
 } // namespace Service::Sockets
diff --git a/src/core/hle/service/sockets/sfdnsres.h b/src/core/hle/service/sockets/sfdnsres.h
index 62c7e35bf..eda432903 100644
--- a/src/core/hle/service/sockets/sfdnsres.h
+++ b/src/core/hle/service/sockets/sfdnsres.h
@@ -12,7 +12,7 @@ namespace Service::Sockets {
 class SFDNSRES final : public ServiceFramework<SFDNSRES> {
 public:
     explicit SFDNSRES();
-    ~SFDNSRES() = default;
+    ~SFDNSRES() override;
 
 private:
     void GetAddrInfo(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/spl/csrng.cpp b/src/core/hle/service/spl/csrng.cpp
index b9e6b799d..674928798 100644
--- a/src/core/hle/service/spl/csrng.cpp
+++ b/src/core/hle/service/spl/csrng.cpp
@@ -13,4 +13,6 @@ CSRNG::CSRNG(std::shared_ptr<Module> module) : Module::Interface(std::move(modul
     RegisterHandlers(functions);
 }
 
+CSRNG::~CSRNG() = default;
+
 } // namespace Service::SPL
diff --git a/src/core/hle/service/spl/csrng.h b/src/core/hle/service/spl/csrng.h
index 3f849b5a7..764d5ceb0 100644
--- a/src/core/hle/service/spl/csrng.h
+++ b/src/core/hle/service/spl/csrng.h
@@ -11,6 +11,7 @@ namespace Service::SPL {
 class CSRNG final : public Module::Interface {
 public:
     explicit CSRNG(std::shared_ptr<Module> module);
+    ~CSRNG() override;
 };
 
 } // namespace Service::SPL
diff --git a/src/core/hle/service/spl/module.cpp b/src/core/hle/service/spl/module.cpp
index 3f5a342a7..0d8441fb1 100644
--- a/src/core/hle/service/spl/module.cpp
+++ b/src/core/hle/service/spl/module.cpp
@@ -16,6 +16,8 @@ namespace Service::SPL {
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
     : ServiceFramework(name), module(std::move(module)) {}
 
+Module::Interface::~Interface() = default;
+
 void Module::Interface::GetRandomBytes(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
 
diff --git a/src/core/hle/service/spl/module.h b/src/core/hle/service/spl/module.h
index f24d998e8..48fda6099 100644
--- a/src/core/hle/service/spl/module.h
+++ b/src/core/hle/service/spl/module.h
@@ -13,6 +13,7 @@ public:
     class Interface : public ServiceFramework<Interface> {
     public:
         explicit Interface(std::shared_ptr<Module> module, const char* name);
+        ~Interface() override;
 
         void GetRandomBytes(Kernel::HLERequestContext& ctx);
 
diff --git a/src/core/hle/service/spl/spl.cpp b/src/core/hle/service/spl/spl.cpp
index bb1e03342..70cb41905 100644
--- a/src/core/hle/service/spl/spl.cpp
+++ b/src/core/hle/service/spl/spl.cpp
@@ -42,4 +42,6 @@ SPL::SPL(std::shared_ptr<Module> module) : Module::Interface(std::move(module),
     RegisterHandlers(functions);
 }
 
+SPL::~SPL() = default;
+
 } // namespace Service::SPL
diff --git a/src/core/hle/service/spl/spl.h b/src/core/hle/service/spl/spl.h
index 69c4c1747..3637d1623 100644
--- a/src/core/hle/service/spl/spl.h
+++ b/src/core/hle/service/spl/spl.h
@@ -11,6 +11,7 @@ namespace Service::SPL {
 class SPL final : public Module::Interface {
 public:
     explicit SPL(std::shared_ptr<Module> module);
+    ~SPL() override;
 };
 
 } // namespace Service::SPL
diff --git a/src/core/hle/service/time/interface.cpp b/src/core/hle/service/time/interface.cpp
index 048d5b077..18a5d71d5 100644
--- a/src/core/hle/service/time/interface.cpp
+++ b/src/core/hle/service/time/interface.cpp
@@ -29,4 +29,6 @@ Time::Time(std::shared_ptr<Module> time, const char* name)
     RegisterHandlers(functions);
 }
 
+Time::~Time() = default;
+
 } // namespace Service::Time
diff --git a/src/core/hle/service/time/interface.h b/src/core/hle/service/time/interface.h
index 183a53db1..cd6b44dec 100644
--- a/src/core/hle/service/time/interface.h
+++ b/src/core/hle/service/time/interface.h
@@ -11,6 +11,7 @@ namespace Service::Time {
 class Time final : public Module::Interface {
 public:
     explicit Time(std::shared_ptr<Module> time, const char* name);
+    ~Time() override;
 };
 
 } // namespace Service::Time
diff --git a/src/core/hle/service/time/time.cpp b/src/core/hle/service/time/time.cpp
index 2172c681b..28fd8debc 100644
--- a/src/core/hle/service/time/time.cpp
+++ b/src/core/hle/service/time/time.cpp
@@ -210,6 +210,8 @@ void Module::Interface::GetStandardLocalSystemClock(Kernel::HLERequestContext& c
 Module::Interface::Interface(std::shared_ptr<Module> time, const char* name)
     : ServiceFramework(name), time(std::move(time)) {}
 
+Module::Interface::~Interface() = default;
+
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     auto time = std::make_shared<Module>();
     std::make_shared<Time>(time, "time:a")->InstallAsService(service_manager);
diff --git a/src/core/hle/service/time/time.h b/src/core/hle/service/time/time.h
index 8dde28a94..5659ecad3 100644
--- a/src/core/hle/service/time/time.h
+++ b/src/core/hle/service/time/time.h
@@ -58,6 +58,7 @@ public:
     class Interface : public ServiceFramework<Interface> {
     public:
         explicit Interface(std::shared_ptr<Module> time, const char* name);
+        ~Interface() override;
 
         void GetStandardUserSystemClock(Kernel::HLERequestContext& ctx);
         void GetStandardNetworkSystemClock(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp
index 993f1e65a..85244ac3b 100644
--- a/src/core/hle/service/vi/vi.cpp
+++ b/src/core/hle/service/vi/vi.cpp
@@ -984,6 +984,8 @@ Module::Interface::Interface(std::shared_ptr<Module> module, const char* name,
                              std::shared_ptr<NVFlinger::NVFlinger> nv_flinger)
     : ServiceFramework(name), module(std::move(module)), nv_flinger(std::move(nv_flinger)) {}
 
+Module::Interface::~Interface() = default;
+
 void Module::Interface::GetDisplayService(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_VI, "(STUBBED) called");
 
diff --git a/src/core/hle/service/vi/vi.h b/src/core/hle/service/vi/vi.h
index 92f5b6059..c2dc83605 100644
--- a/src/core/hle/service/vi/vi.h
+++ b/src/core/hle/service/vi/vi.h
@@ -26,6 +26,7 @@ public:
     public:
         explicit Interface(std::shared_ptr<Module> module, const char* name,
                            std::shared_ptr<NVFlinger::NVFlinger> nv_flinger);
+        ~Interface() override;
 
         void GetDisplayService(Kernel::HLERequestContext& ctx);
 
diff --git a/src/core/hle/service/vi/vi_m.cpp b/src/core/hle/service/vi/vi_m.cpp
index d47da565b..207c06b16 100644
--- a/src/core/hle/service/vi/vi_m.cpp
+++ b/src/core/hle/service/vi/vi_m.cpp
@@ -15,4 +15,6 @@ VI_M::VI_M(std::shared_ptr<Module> module, std::shared_ptr<NVFlinger::NVFlinger>
     RegisterHandlers(functions);
 }
 
+VI_M::~VI_M() = default;
+
 } // namespace Service::VI
diff --git a/src/core/hle/service/vi/vi_m.h b/src/core/hle/service/vi/vi_m.h
index 6abb9b3a3..487d58d50 100644
--- a/src/core/hle/service/vi/vi_m.h
+++ b/src/core/hle/service/vi/vi_m.h
@@ -11,6 +11,7 @@ namespace Service::VI {
 class VI_M final : public Module::Interface {
 public:
     explicit VI_M(std::shared_ptr<Module> module, std::shared_ptr<NVFlinger::NVFlinger> nv_flinger);
+    ~VI_M() override;
 };
 
 } // namespace Service::VI
diff --git a/src/core/hle/service/vi/vi_s.cpp b/src/core/hle/service/vi/vi_s.cpp
index 8f82e797f..920e6a1f6 100644
--- a/src/core/hle/service/vi/vi_s.cpp
+++ b/src/core/hle/service/vi/vi_s.cpp
@@ -15,4 +15,6 @@ VI_S::VI_S(std::shared_ptr<Module> module, std::shared_ptr<NVFlinger::NVFlinger>
     RegisterHandlers(functions);
 }
 
+VI_S::~VI_S() = default;
+
 } // namespace Service::VI
diff --git a/src/core/hle/service/vi/vi_s.h b/src/core/hle/service/vi/vi_s.h
index 8f16f804f..bbc31148f 100644
--- a/src/core/hle/service/vi/vi_s.h
+++ b/src/core/hle/service/vi/vi_s.h
@@ -11,6 +11,7 @@ namespace Service::VI {
 class VI_S final : public Module::Interface {
 public:
     explicit VI_S(std::shared_ptr<Module> module, std::shared_ptr<NVFlinger::NVFlinger> nv_flinger);
+    ~VI_S() override;
 };
 
 } // namespace Service::VI
diff --git a/src/core/hle/service/vi/vi_u.cpp b/src/core/hle/service/vi/vi_u.cpp
index b84aed1d5..d81e410d6 100644
--- a/src/core/hle/service/vi/vi_u.cpp
+++ b/src/core/hle/service/vi/vi_u.cpp
@@ -14,4 +14,6 @@ VI_U::VI_U(std::shared_ptr<Module> module, std::shared_ptr<NVFlinger::NVFlinger>
     RegisterHandlers(functions);
 }
 
+VI_U::~VI_U() = default;
+
 } // namespace Service::VI
diff --git a/src/core/hle/service/vi/vi_u.h b/src/core/hle/service/vi/vi_u.h
index e9b4f76b2..b92f28c92 100644
--- a/src/core/hle/service/vi/vi_u.h
+++ b/src/core/hle/service/vi/vi_u.h
@@ -11,6 +11,7 @@ namespace Service::VI {
 class VI_U final : public Module::Interface {
 public:
     explicit VI_U(std::shared_ptr<Module> module, std::shared_ptr<NVFlinger::NVFlinger> nv_flinger);
+    ~VI_U() override;
 };
 
 } // namespace Service::VI
diff --git a/src/core/settings.h b/src/core/settings.h
index c25f8ba70..0318d019c 100644
--- a/src/core/settings.h
+++ b/src/core/settings.h
@@ -5,6 +5,7 @@
 #pragma once
 
 #include <array>
+#include <atomic>
 #include <string>
 #include "common/common_types.h"
 
@@ -120,6 +121,7 @@ struct Values {
     std::array<std::string, NativeAnalog::NumAnalogs> analogs;
     std::string motion_device;
     std::string touch_device;
+    std::atomic_bool is_device_reload_pending{true};
 
     // Core
     bool use_cpu_jit;
diff --git a/src/input_common/main.cpp b/src/input_common/main.cpp
index b12623d55..37f572853 100644
--- a/src/input_common/main.cpp
+++ b/src/input_common/main.cpp
@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <memory>
+#include <thread>
 #include "common/param_package.h"
 #include "input_common/analog_from_button.h"
 #include "input_common/keyboard.h"
@@ -17,6 +18,10 @@ namespace InputCommon {
 static std::shared_ptr<Keyboard> keyboard;
 static std::shared_ptr<MotionEmu> motion_emu;
 
+#ifdef HAVE_SDL2
+static std::thread poll_thread;
+#endif
+
 void Init() {
     keyboard = std::make_shared<Keyboard>();
     Input::RegisterFactory<Input::ButtonDevice>("keyboard", keyboard);
@@ -30,6 +35,12 @@ void Init() {
 #endif
 }
 
+void StartJoystickEventHandler() {
+#ifdef HAVE_SDL2
+    poll_thread = std::thread(SDL::PollLoop);
+#endif
+}
+
 void Shutdown() {
     Input::UnregisterFactory<Input::ButtonDevice>("keyboard");
     keyboard.reset();
@@ -39,6 +50,7 @@ void Shutdown() {
 
 #ifdef HAVE_SDL2
     SDL::Shutdown();
+    poll_thread.join();
 #endif
 }
 
diff --git a/src/input_common/main.h b/src/input_common/main.h
index 77a0ce90b..9eb13106e 100644
--- a/src/input_common/main.h
+++ b/src/input_common/main.h
@@ -20,6 +20,8 @@ void Init();
 /// Deregisters all built-in input device factories and shuts them down.
 void Shutdown();
 
+void StartJoystickEventHandler();
+
 class Keyboard;
 
 /// Gets the keyboard button device factory.
diff --git a/src/input_common/sdl/sdl.cpp b/src/input_common/sdl/sdl.cpp
index d1b960fd7..faf3c1fa3 100644
--- a/src/input_common/sdl/sdl.cpp
+++ b/src/input_common/sdl/sdl.cpp
@@ -2,15 +2,24 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
+#include <atomic>
 #include <cmath>
+#include <functional>
+#include <iterator>
+#include <mutex>
 #include <string>
+#include <thread>
 #include <tuple>
 #include <unordered_map>
 #include <utility>
+#include <vector>
 #include <SDL.h>
+#include "common/assert.h"
 #include "common/logging/log.h"
 #include "common/math_util.h"
 #include "common/param_package.h"
+#include "common/threadsafe_queue.h"
 #include "input_common/main.h"
 #include "input_common/sdl/sdl.h"
 
@@ -21,33 +30,51 @@ namespace SDL {
 class SDLJoystick;
 class SDLButtonFactory;
 class SDLAnalogFactory;
-static std::unordered_map<int, std::weak_ptr<SDLJoystick>> joystick_list;
+
+/// Map of GUID of a list of corresponding virtual Joysticks
+static std::unordered_map<std::string, std::vector<std::shared_ptr<SDLJoystick>>> joystick_map;
+static std::mutex joystick_map_mutex;
+
 static std::shared_ptr<SDLButtonFactory> button_factory;
 static std::shared_ptr<SDLAnalogFactory> analog_factory;
 
-static bool initialized = false;
+/// Used by the Pollers during config
+static std::atomic<bool> polling;
+static Common::SPSCQueue<SDL_Event> event_queue;
+
+static std::atomic<bool> initialized = false;
+
+static std::string GetGUID(SDL_Joystick* joystick) {
+    SDL_JoystickGUID guid = SDL_JoystickGetGUID(joystick);
+    char guid_str[33];
+    SDL_JoystickGetGUIDString(guid, guid_str, sizeof(guid_str));
+    return guid_str;
+}
 
 class SDLJoystick {
 public:
-    explicit SDLJoystick(int joystick_index)
-        : joystick{SDL_JoystickOpen(joystick_index), SDL_JoystickClose} {
-        if (!joystick) {
-            LOG_ERROR(Input, "failed to open joystick {}", joystick_index);
-        }
+    SDLJoystick(std::string guid_, int port_, SDL_Joystick* joystick,
+                decltype(&SDL_JoystickClose) deleter = &SDL_JoystickClose)
+        : guid{std::move(guid_)}, port{port_}, sdl_joystick{joystick, deleter} {}
+
+    void SetButton(int button, bool value) {
+        std::lock_guard<std::mutex> lock(mutex);
+        state.buttons[button] = value;
     }
 
     bool GetButton(int button) const {
-        if (!joystick)
-            return {};
-        SDL_JoystickUpdate();
-        return SDL_JoystickGetButton(joystick.get(), button) == 1;
+        std::lock_guard<std::mutex> lock(mutex);
+        return state.buttons.at(button);
+    }
+
+    void SetAxis(int axis, Sint16 value) {
+        std::lock_guard<std::mutex> lock(mutex);
+        state.axes[axis] = value;
     }
 
     float GetAxis(int axis) const {
-        if (!joystick)
-            return {};
-        SDL_JoystickUpdate();
-        return SDL_JoystickGetAxis(joystick.get(), axis) / 32767.0f;
+        std::lock_guard<std::mutex> lock(mutex);
+        return state.axes.at(axis) / 32767.0f;
     }
 
     std::tuple<float, float> GetAnalog(int axis_x, int axis_y) const {
@@ -67,18 +94,213 @@ public:
         return std::make_tuple(x, y);
     }
 
+    void SetHat(int hat, Uint8 direction) {
+        std::lock_guard<std::mutex> lock(mutex);
+        state.hats[hat] = direction;
+    }
+
     bool GetHatDirection(int hat, Uint8 direction) const {
-        return (SDL_JoystickGetHat(joystick.get(), hat) & direction) != 0;
+        std::lock_guard<std::mutex> lock(mutex);
+        return (state.hats.at(hat) & direction) != 0;
+    }
+    /**
+     * The guid of the joystick
+     */
+    const std::string& GetGUID() const {
+        return guid;
+    }
+
+    /**
+     * The number of joystick from the same type that were connected before this joystick
+     */
+    int GetPort() const {
+        return port;
+    }
+
+    SDL_Joystick* GetSDLJoystick() const {
+        return sdl_joystick.get();
     }
 
-    SDL_JoystickID GetJoystickID() const {
-        return SDL_JoystickInstanceID(joystick.get());
+    void SetSDLJoystick(SDL_Joystick* joystick,
+                        decltype(&SDL_JoystickClose) deleter = &SDL_JoystickClose) {
+        sdl_joystick =
+            std::unique_ptr<SDL_Joystick, decltype(&SDL_JoystickClose)>(joystick, deleter);
     }
 
 private:
-    std::unique_ptr<SDL_Joystick, decltype(&SDL_JoystickClose)> joystick;
+    struct State {
+        std::unordered_map<int, bool> buttons;
+        std::unordered_map<int, Sint16> axes;
+        std::unordered_map<int, Uint8> hats;
+    } state;
+    std::string guid;
+    int port;
+    std::unique_ptr<SDL_Joystick, decltype(&SDL_JoystickClose)> sdl_joystick;
+    mutable std::mutex mutex;
 };
 
+/**
+ * Get the nth joystick with the corresponding GUID
+ */
+static std::shared_ptr<SDLJoystick> GetSDLJoystickByGUID(const std::string& guid, int port) {
+    std::lock_guard<std::mutex> lock(joystick_map_mutex);
+    const auto it = joystick_map.find(guid);
+    if (it != joystick_map.end()) {
+        while (it->second.size() <= port) {
+            auto joystick = std::make_shared<SDLJoystick>(guid, it->second.size(), nullptr,
+                                                          [](SDL_Joystick*) {});
+            it->second.emplace_back(std::move(joystick));
+        }
+        return it->second[port];
+    }
+    auto joystick = std::make_shared<SDLJoystick>(guid, 0, nullptr, [](SDL_Joystick*) {});
+    return joystick_map[guid].emplace_back(std::move(joystick));
+}
+
+/**
+ * Check how many identical joysticks (by guid) were connected before the one with sdl_id and so tie
+ * it to a SDLJoystick with the same guid and that port
+ */
+static std::shared_ptr<SDLJoystick> GetSDLJoystickBySDLID(SDL_JoystickID sdl_id) {
+    std::lock_guard<std::mutex> lock(joystick_map_mutex);
+    auto sdl_joystick = SDL_JoystickFromInstanceID(sdl_id);
+    const std::string guid = GetGUID(sdl_joystick);
+    auto map_it = joystick_map.find(guid);
+    if (map_it != joystick_map.end()) {
+        auto vec_it = std::find_if(map_it->second.begin(), map_it->second.end(),
+                                   [&sdl_joystick](const std::shared_ptr<SDLJoystick>& joystick) {
+                                       return sdl_joystick == joystick->GetSDLJoystick();
+                                   });
+        if (vec_it != map_it->second.end()) {
+            // This is the common case: There is already an existing SDL_Joystick maped to a
+            // SDLJoystick. return the SDLJoystick
+            return *vec_it;
+        }
+        // Search for a SDLJoystick without a mapped SDL_Joystick...
+        auto nullptr_it = std::find_if(map_it->second.begin(), map_it->second.end(),
+                                       [](const std::shared_ptr<SDLJoystick>& joystick) {
+                                           return !joystick->GetSDLJoystick();
+                                       });
+        if (nullptr_it != map_it->second.end()) {
+            // ... and map it
+            (*nullptr_it)->SetSDLJoystick(sdl_joystick);
+            return *nullptr_it;
+        }
+        // There is no SDLJoystick without a mapped SDL_Joystick
+        // Create a new SDLJoystick
+        auto joystick = std::make_shared<SDLJoystick>(guid, map_it->second.size(), sdl_joystick);
+        return map_it->second.emplace_back(std::move(joystick));
+    }
+    auto joystick = std::make_shared<SDLJoystick>(guid, 0, sdl_joystick);
+    return joystick_map[guid].emplace_back(std::move(joystick));
+}
+
+void InitJoystick(int joystick_index) {
+    std::lock_guard<std::mutex> lock(joystick_map_mutex);
+    SDL_Joystick* sdl_joystick = SDL_JoystickOpen(joystick_index);
+    if (!sdl_joystick) {
+        LOG_ERROR(Input, "failed to open joystick {}", joystick_index);
+        return;
+    }
+    std::string guid = GetGUID(sdl_joystick);
+    if (joystick_map.find(guid) == joystick_map.end()) {
+        auto joystick = std::make_shared<SDLJoystick>(guid, 0, sdl_joystick);
+        joystick_map[guid].emplace_back(std::move(joystick));
+        return;
+    }
+    auto& joystick_guid_list = joystick_map[guid];
+    const auto it = std::find_if(
+        joystick_guid_list.begin(), joystick_guid_list.end(),
+        [](const std::shared_ptr<SDLJoystick>& joystick) { return !joystick->GetSDLJoystick(); });
+    if (it != joystick_guid_list.end()) {
+        (*it)->SetSDLJoystick(sdl_joystick);
+        return;
+    }
+    auto joystick = std::make_shared<SDLJoystick>(guid, joystick_guid_list.size(), sdl_joystick);
+    joystick_guid_list.emplace_back(std::move(joystick));
+}
+
+void CloseJoystick(SDL_Joystick* sdl_joystick) {
+    std::lock_guard<std::mutex> lock(joystick_map_mutex);
+    std::string guid = GetGUID(sdl_joystick);
+    // This call to guid is save since the joystick is guranteed to be in that map
+    auto& joystick_guid_list = joystick_map[guid];
+    const auto joystick_it =
+        std::find_if(joystick_guid_list.begin(), joystick_guid_list.end(),
+                     [&sdl_joystick](const std::shared_ptr<SDLJoystick>& joystick) {
+                         return joystick->GetSDLJoystick() == sdl_joystick;
+                     });
+    (*joystick_it)->SetSDLJoystick(nullptr, [](SDL_Joystick*) {});
+}
+
+void HandleGameControllerEvent(const SDL_Event& event) {
+    switch (event.type) {
+    case SDL_JOYBUTTONUP: {
+        auto joystick = GetSDLJoystickBySDLID(event.jbutton.which);
+        if (joystick) {
+            joystick->SetButton(event.jbutton.button, false);
+        }
+        break;
+    }
+    case SDL_JOYBUTTONDOWN: {
+        auto joystick = GetSDLJoystickBySDLID(event.jbutton.which);
+        if (joystick) {
+            joystick->SetButton(event.jbutton.button, true);
+        }
+        break;
+    }
+    case SDL_JOYHATMOTION: {
+        auto joystick = GetSDLJoystickBySDLID(event.jhat.which);
+        if (joystick) {
+            joystick->SetHat(event.jhat.hat, event.jhat.value);
+        }
+        break;
+    }
+    case SDL_JOYAXISMOTION: {
+        auto joystick = GetSDLJoystickBySDLID(event.jaxis.which);
+        if (joystick) {
+            joystick->SetAxis(event.jaxis.axis, event.jaxis.value);
+        }
+        break;
+    }
+    case SDL_JOYDEVICEREMOVED:
+        LOG_DEBUG(Input, "Controller removed with Instance_ID {}", event.jdevice.which);
+        CloseJoystick(SDL_JoystickFromInstanceID(event.jdevice.which));
+        break;
+    case SDL_JOYDEVICEADDED:
+        LOG_DEBUG(Input, "Controller connected with device index {}", event.jdevice.which);
+        InitJoystick(event.jdevice.which);
+        break;
+    }
+}
+
+void CloseSDLJoysticks() {
+    std::lock_guard<std::mutex> lock(joystick_map_mutex);
+    joystick_map.clear();
+}
+
+void PollLoop() {
+    if (SDL_Init(SDL_INIT_JOYSTICK) < 0) {
+        LOG_CRITICAL(Input, "SDL_Init(SDL_INIT_JOYSTICK) failed with: {}", SDL_GetError());
+        return;
+    }
+
+    SDL_Event event;
+    while (initialized) {
+        // Wait for 10 ms or until an event happens
+        if (SDL_WaitEventTimeout(&event, 10)) {
+            // Don't handle the event if we are configuring
+            if (polling) {
+                event_queue.Push(event);
+            } else {
+                HandleGameControllerEvent(event);
+            }
+        }
+    }
+    CloseSDLJoysticks();
+    SDL_QuitSubSystem(SDL_INIT_JOYSTICK);
+}
+
 class SDLButton final : public Input::ButtonDevice {
 public:
     explicit SDLButton(std::shared_ptr<SDLJoystick> joystick_, int button_)
@@ -144,22 +366,14 @@ private:
     int axis_y;
 };
 
-static std::shared_ptr<SDLJoystick> GetJoystick(int joystick_index) {
-    std::shared_ptr<SDLJoystick> joystick = joystick_list[joystick_index].lock();
-    if (!joystick) {
-        joystick = std::make_shared<SDLJoystick>(joystick_index);
-        joystick_list[joystick_index] = joystick;
-    }
-    return joystick;
-}
-
 /// A button device factory that creates button devices from SDL joystick
 class SDLButtonFactory final : public Input::Factory<Input::ButtonDevice> {
 public:
     /**
      * Creates a button device from a joystick button
      * @param params contains parameters for creating the device:
-     *     - "joystick": the index of the joystick to bind
+     *     - "guid": the guid of the joystick to bind
+     *     - "port": the nth joystick of the same type to bind
      *     - "button"(optional): the index of the button to bind
      *     - "hat"(optional): the index of the hat to bind as direction buttons
      *     - "axis"(optional): the index of the axis to bind
@@ -167,12 +381,15 @@ public:
      *         "down", "left" or "right"
      *     - "threshold"(only used for axis): a float value in (-1.0, 1.0) which the button is
      *         triggered if the axis value crosses
-     *     - "direction"(only used for axis): "+" means the button is triggered when the axis value
-     *         is greater than the threshold; "-" means the button is triggered when the axis value
-     *         is smaller than the threshold
+     *     - "direction"(only used for axis): "+" means the button is triggered when the axis
+     * value is greater than the threshold; "-" means the button is triggered when the axis
+     * value is smaller than the threshold
      */
     std::unique_ptr<Input::ButtonDevice> Create(const Common::ParamPackage& params) override {
-        const int joystick_index = params.Get("joystick", 0);
+        const std::string guid = params.Get("guid", "0");
+        const int port = params.Get("port", 0);
+
+        auto joystick = GetSDLJoystickByGUID(guid, port);
 
         if (params.Has("hat")) {
             const int hat = params.Get("hat", 0);
@@ -189,8 +406,9 @@ public:
             } else {
                 direction = 0;
             }
-            return std::make_unique<SDLDirectionButton>(GetJoystick(joystick_index), hat,
-                                                        direction);
+            // This is necessary so accessing GetHat with hat won't crash
+            joystick->SetHat(hat, SDL_HAT_CENTERED);
+            return std::make_unique<SDLDirectionButton>(joystick, hat, direction);
         }
 
         if (params.Has("axis")) {
@@ -206,12 +424,15 @@ public:
                 trigger_if_greater = true;
                 LOG_ERROR(Input, "Unknown direction '{}'", direction_name);
             }
-            return std::make_unique<SDLAxisButton>(GetJoystick(joystick_index), axis, threshold,
-                                                   trigger_if_greater);
+            // This is necessary so accessing GetAxis with axis won't crash
+            joystick->SetAxis(axis, 0);
+            return std::make_unique<SDLAxisButton>(joystick, axis, threshold, trigger_if_greater);
         }
 
         const int button = params.Get("button", 0);
-        return std::make_unique<SDLButton>(GetJoystick(joystick_index), button);
+        // This is necessary so accessing GetButton with button won't crash
+        joystick->SetButton(button, false);
+        return std::make_unique<SDLButton>(joystick, button);
     }
 };
 
@@ -221,27 +442,32 @@ public:
     /**
      * Creates analog device from joystick axes
      * @param params contains parameters for creating the device:
-     *     - "joystick": the index of the joystick to bind
+     *     - "guid": the guid of the joystick to bind
+     *     - "port": the nth joystick of the same type
      *     - "axis_x": the index of the axis to be bind as x-axis
      *     - "axis_y": the index of the axis to be bind as y-axis
      */
     std::unique_ptr<Input::AnalogDevice> Create(const Common::ParamPackage& params) override {
-        const int joystick_index = params.Get("joystick", 0);
+        const std::string guid = params.Get("guid", "0");
+        const int port = params.Get("port", 0);
         const int axis_x = params.Get("axis_x", 0);
         const int axis_y = params.Get("axis_y", 1);
-        return std::make_unique<SDLAnalog>(GetJoystick(joystick_index), axis_x, axis_y);
+
+        auto joystick = GetSDLJoystickByGUID(guid, port);
+
+        // This is necessary so accessing GetAxis with axis_x and axis_y won't crash
+        joystick->SetAxis(axis_x, 0);
+        joystick->SetAxis(axis_y, 0);
+        return std::make_unique<SDLAnalog>(joystick, axis_x, axis_y);
     }
 };
 
 void Init() {
-    if (SDL_Init(SDL_INIT_JOYSTICK) < 0) {
-        LOG_CRITICAL(Input, "SDL_Init(SDL_INIT_JOYSTICK) failed with: {}", SDL_GetError());
-    } else {
-        using namespace Input;
-        RegisterFactory<ButtonDevice>("sdl", std::make_shared<SDLButtonFactory>());
-        RegisterFactory<AnalogDevice>("sdl", std::make_shared<SDLAnalogFactory>());
-        initialized = true;
-    }
+    using namespace Input;
+    RegisterFactory<ButtonDevice>("sdl", std::make_shared<SDLButtonFactory>());
+    RegisterFactory<AnalogDevice>("sdl", std::make_shared<SDLAnalogFactory>());
+    polling = false;
+    initialized = true;
 }
 
 void Shutdown() {
@@ -249,30 +475,17 @@ void Shutdown() {
         using namespace Input;
         UnregisterFactory<ButtonDevice>("sdl");
         UnregisterFactory<AnalogDevice>("sdl");
-        SDL_QuitSubSystem(SDL_INIT_JOYSTICK);
+        initialized = false;
     }
 }
 
-/**
- * This function converts a joystick ID used in SDL events to the device index. This is necessary
- * because Citra opens joysticks using their indices, not their IDs.
- */
-static int JoystickIDToDeviceIndex(SDL_JoystickID id) {
-    int num_joysticks = SDL_NumJoysticks();
-    for (int i = 0; i < num_joysticks; i++) {
-        auto joystick = GetJoystick(i);
-        if (joystick->GetJoystickID() == id) {
-            return i;
-        }
-    }
-    return -1;
-}
-
 Common::ParamPackage SDLEventToButtonParamPackage(const SDL_Event& event) {
     Common::ParamPackage params({{"engine", "sdl"}});
     switch (event.type) {
-    case SDL_JOYAXISMOTION:
-        params.Set("joystick", JoystickIDToDeviceIndex(event.jaxis.which));
+    case SDL_JOYAXISMOTION: {
+        auto joystick = GetSDLJoystickBySDLID(event.jaxis.which);
+        params.Set("port", joystick->GetPort());
+        params.Set("guid", joystick->GetGUID());
         params.Set("axis", event.jaxis.axis);
         if (event.jaxis.value > 0) {
             params.Set("direction", "+");
@@ -282,12 +495,18 @@ Common::ParamPackage SDLEventToButtonParamPackage(const SDL_Event& event) {
             params.Set("threshold", "-0.5");
         }
         break;
-    case SDL_JOYBUTTONUP:
-        params.Set("joystick", JoystickIDToDeviceIndex(event.jbutton.which));
+    }
+    case SDL_JOYBUTTONUP: {
+        auto joystick = GetSDLJoystickBySDLID(event.jbutton.which);
+        params.Set("port", joystick->GetPort());
+        params.Set("guid", joystick->GetGUID());
         params.Set("button", event.jbutton.button);
         break;
-    case SDL_JOYHATMOTION:
-        params.Set("joystick", JoystickIDToDeviceIndex(event.jhat.which));
+    }
+    case SDL_JOYHATMOTION: {
+        auto joystick = GetSDLJoystickBySDLID(event.jhat.which);
+        params.Set("port", joystick->GetPort());
+        params.Set("guid", joystick->GetGUID());
         params.Set("hat", event.jhat.hat);
         switch (event.jhat.value) {
         case SDL_HAT_UP:
@@ -307,6 +526,7 @@ Common::ParamPackage SDLEventToButtonParamPackage(const SDL_Event& event) {
         }
         break;
     }
+    }
     return params;
 }
 
@@ -315,31 +535,20 @@ namespace Polling {
 class SDLPoller : public InputCommon::Polling::DevicePoller {
 public:
     void Start() override {
-        // SDL joysticks must be opened, otherwise they don't generate events
-        SDL_JoystickUpdate();
-        int num_joysticks = SDL_NumJoysticks();
-        for (int i = 0; i < num_joysticks; i++) {
-            joysticks_opened.emplace_back(GetJoystick(i));
-        }
-        // Empty event queue to get rid of old events. citra-qt doesn't use the queue
-        SDL_Event dummy;
-        while (SDL_PollEvent(&dummy)) {
-        }
+        event_queue.Clear();
+        polling = true;
     }
 
     void Stop() override {
-        joysticks_opened.clear();
+        polling = false;
     }
-
-private:
-    std::vector<std::shared_ptr<SDLJoystick>> joysticks_opened;
 };
 
 class SDLButtonPoller final : public SDLPoller {
 public:
     Common::ParamPackage GetNextInput() override {
         SDL_Event event;
-        while (SDL_PollEvent(&event)) {
+        while (event_queue.Pop(event)) {
             switch (event.type) {
             case SDL_JOYAXISMOTION:
                 if (std::abs(event.jaxis.value / 32767.0) < 0.5) {
@@ -367,7 +576,7 @@ public:
 
     Common::ParamPackage GetNextInput() override {
         SDL_Event event;
-        while (SDL_PollEvent(&event)) {
+        while (event_queue.Pop(event)) {
             if (event.type != SDL_JOYAXISMOTION || std::abs(event.jaxis.value / 32767.0) < 0.5) {
                 continue;
             }
@@ -384,8 +593,10 @@ public:
         }
         Common::ParamPackage params;
         if (analog_xaxis != -1 && analog_yaxis != -1) {
+            auto joystick = GetSDLJoystickBySDLID(event.jaxis.which);
             params.Set("engine", "sdl");
-            params.Set("joystick", JoystickIDToDeviceIndex(analog_axes_joystick));
+            params.Set("port", joystick->GetPort());
+            params.Set("guid", joystick->GetGUID());
             params.Set("axis_x", analog_xaxis);
             params.Set("axis_y", analog_yaxis);
             analog_xaxis = -1;
diff --git a/src/input_common/sdl/sdl.h b/src/input_common/sdl/sdl.h
index 7934099d4..0206860d3 100644
--- a/src/input_common/sdl/sdl.h
+++ b/src/input_common/sdl/sdl.h
@@ -28,6 +28,15 @@ void Init();
 /// Unresisters SDL device factories and shut them down.
 void Shutdown();
 
+/// Needs to be called before SDL_QuitSubSystem.
+void CloseSDLJoysticks();
+
+/// Handle SDL_Events for joysticks from SDL_PollEvent
+void HandleGameControllerEvent(const SDL_Event& event);
+
+/// A Loop that calls HandleGameControllerEvent until Shutdown is called
+void PollLoop();
+
 /// Creates a ParamPackage from an SDL_Event that can directly be used to create a ButtonDevice
 Common::ParamPackage SDLEventToButtonParamPackage(const SDL_Event& event);
 
diff --git a/src/video_core/command_processor.cpp b/src/video_core/command_processor.cpp
index d5831e752..2625ddfdc 100644
--- a/src/video_core/command_processor.cpp
+++ b/src/video_core/command_processor.cpp
@@ -28,98 +28,106 @@ enum class BufferMethods {
     CountBufferMethods = 0x40,
 };
 
-void GPU::WriteReg(u32 method, u32 subchannel, u32 value, u32 remaining_params) {
-    LOG_TRACE(HW_GPU,
-              "Processing method {:08X} on subchannel {} value "
-              "{:08X} remaining params {}",
-              method, subchannel, value, remaining_params);
-
-    ASSERT(subchannel < bound_engines.size());
-
-    if (method == static_cast<u32>(BufferMethods::BindObject)) {
-        // Bind the current subchannel to the desired engine id.
-        LOG_DEBUG(HW_GPU, "Binding subchannel {} to engine {}", subchannel, value);
-        bound_engines[subchannel] = static_cast<EngineID>(value);
-        return;
-    }
+MICROPROFILE_DEFINE(ProcessCommandLists, "GPU", "Execute command buffer", MP_RGB(128, 128, 192));
 
-    if (method < static_cast<u32>(BufferMethods::CountBufferMethods)) {
-        // TODO(Subv): Research and implement these methods.
-        LOG_ERROR(HW_GPU, "Special buffer methods other than Bind are not implemented");
-        return;
-    }
+void GPU::ProcessCommandLists(const std::vector<CommandListHeader>& commands) {
+    MICROPROFILE_SCOPE(ProcessCommandLists);
 
-    const EngineID engine = bound_engines[subchannel];
-
-    switch (engine) {
-    case EngineID::FERMI_TWOD_A:
-        fermi_2d->WriteReg(method, value);
-        break;
-    case EngineID::MAXWELL_B:
-        maxwell_3d->WriteReg(method, value, remaining_params);
-        break;
-    case EngineID::MAXWELL_COMPUTE_B:
-        maxwell_compute->WriteReg(method, value);
-        break;
-    case EngineID::MAXWELL_DMA_COPY_A:
-        maxwell_dma->WriteReg(method, value);
-        break;
-    default:
-        UNIMPLEMENTED_MSG("Unimplemented engine");
-    }
-}
+    auto WriteReg = [this](u32 method, u32 subchannel, u32 value, u32 remaining_params) {
+        LOG_TRACE(HW_GPU,
+                  "Processing method {:08X} on subchannel {} value "
+                  "{:08X} remaining params {}",
+                  method, subchannel, value, remaining_params);
 
-void GPU::ProcessCommandList(GPUVAddr address, u32 size) {
-    const boost::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)};
-        current_addr += sizeof(u32);
-
-        switch (header.mode.Value()) {
-        case SubmissionMode::IncreasingOld:
-        case SubmissionMode::Increasing: {
-            // Increase the method value with each argument.
-            for (unsigned i = 0; i < header.arg_count; ++i) {
-                WriteReg(header.method + i, header.subchannel, Memory::Read32(current_addr),
-                         header.arg_count - i - 1);
-                current_addr += sizeof(u32);
-            }
-            break;
+        ASSERT(subchannel < bound_engines.size());
+
+        if (method == static_cast<u32>(BufferMethods::BindObject)) {
+            // Bind the current subchannel to the desired engine id.
+            LOG_DEBUG(HW_GPU, "Binding subchannel {} to engine {}", subchannel, value);
+            bound_engines[subchannel] = static_cast<EngineID>(value);
+            return;
         }
-        case SubmissionMode::NonIncreasingOld:
-        case SubmissionMode::NonIncreasing: {
-            // Use the same method value for all arguments.
-            for (unsigned i = 0; i < header.arg_count; ++i) {
-                WriteReg(header.method, header.subchannel, Memory::Read32(current_addr),
-                         header.arg_count - i - 1);
-                current_addr += sizeof(u32);
-            }
+
+        if (method < static_cast<u32>(BufferMethods::CountBufferMethods)) {
+            // TODO(Subv): Research and implement these methods.
+            LOG_ERROR(HW_GPU, "Special buffer methods other than Bind are not implemented");
+            return;
+        }
+
+        const EngineID engine = bound_engines[subchannel];
+
+        switch (engine) {
+        case EngineID::FERMI_TWOD_A:
+            fermi_2d->WriteReg(method, value);
+            break;
+        case EngineID::MAXWELL_B:
+            maxwell_3d->WriteReg(method, value, remaining_params);
             break;
+        case EngineID::MAXWELL_COMPUTE_B:
+            maxwell_compute->WriteReg(method, value);
+            break;
+        case EngineID::MAXWELL_DMA_COPY_A:
+            maxwell_dma->WriteReg(method, value);
+            break;
+        default:
+            UNIMPLEMENTED_MSG("Unimplemented engine");
         }
-        case SubmissionMode::IncreaseOnce: {
-            ASSERT(header.arg_count.Value() >= 1);
+    };
 
-            // Use the original method for the first argument and then the next method for all other
-            // arguments.
-            WriteReg(header.method, header.subchannel, Memory::Read32(current_addr),
-                     header.arg_count - 1);
+    for (auto entry : commands) {
+        Tegra::GPUVAddr address = entry.Address();
+        u32 size = entry.sz;
+        const boost::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)};
             current_addr += sizeof(u32);
 
-            for (unsigned i = 1; i < header.arg_count; ++i) {
-                WriteReg(header.method + 1, header.subchannel, Memory::Read32(current_addr),
-                         header.arg_count - i - 1);
+            switch (header.mode.Value()) {
+            case SubmissionMode::IncreasingOld:
+            case SubmissionMode::Increasing: {
+                // Increase the method value with each argument.
+                for (unsigned i = 0; i < header.arg_count; ++i) {
+                    WriteReg(header.method + i, header.subchannel, Memory::Read32(current_addr),
+                             header.arg_count - i - 1);
+                    current_addr += sizeof(u32);
+                }
+                break;
+            }
+            case SubmissionMode::NonIncreasingOld:
+            case SubmissionMode::NonIncreasing: {
+                // Use the same method value for all arguments.
+                for (unsigned i = 0; i < header.arg_count; ++i) {
+                    WriteReg(header.method, header.subchannel, Memory::Read32(current_addr),
+                             header.arg_count - i - 1);
+                    current_addr += sizeof(u32);
+                }
+                break;
+            }
+            case SubmissionMode::IncreaseOnce: {
+                ASSERT(header.arg_count.Value() >= 1);
+
+                // Use the original method for the first argument and then the next method for all
+                // other arguments.
+                WriteReg(header.method, header.subchannel, Memory::Read32(current_addr),
+                         header.arg_count - 1);
                 current_addr += sizeof(u32);
+
+                for (unsigned i = 1; i < header.arg_count; ++i) {
+                    WriteReg(header.method + 1, header.subchannel, Memory::Read32(current_addr),
+                             header.arg_count - i - 1);
+                    current_addr += sizeof(u32);
+                }
+                break;
+            }
+            case SubmissionMode::Inline: {
+                // The register value is stored in the bits 16-28 as an immediate
+                WriteReg(header.method, header.subchannel, header.inline_data, 0);
+                break;
+            }
+            default:
+                UNIMPLEMENTED();
             }
-            break;
-        }
-        case SubmissionMode::Inline: {
-            // The register value is stored in the bits 16-28 as an immediate
-            WriteReg(header.method, header.subchannel, header.inline_data, 0);
-            break;
-        }
-        default:
-            UNIMPLEMENTED();
         }
     }
 }
diff --git a/src/video_core/command_processor.h b/src/video_core/command_processor.h
index a01153e0b..bd766e77a 100644
--- a/src/video_core/command_processor.h
+++ b/src/video_core/command_processor.h
@@ -7,6 +7,7 @@
 #include <type_traits>
 #include "common/bit_field.h"
 #include "common/common_types.h"
+#include "video_core/memory_manager.h"
 
 namespace Tegra {
 
@@ -19,6 +20,22 @@ enum class SubmissionMode : u32 {
     IncreaseOnce = 5
 };
 
+struct CommandListHeader {
+    u32 entry0; // gpu_va_lo
+    union {
+        u32 entry1; // gpu_va_hi | (unk_0x02 << 0x08) | (size << 0x0A) | (unk_0x01 << 0x1F)
+        BitField<0, 8, u32> gpu_va_hi;
+        BitField<8, 2, u32> unk1;
+        BitField<10, 21, u32> sz;
+        BitField<31, 1, u32> unk2;
+    };
+
+    GPUVAddr Address() const {
+        return (static_cast<GPUVAddr>(gpu_va_hi) << 32) | entry0;
+    }
+};
+static_assert(sizeof(CommandListHeader) == 8, "CommandListHeader is incorrect size");
+
 union CommandHeader {
     u32 hex;
 
diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp
index e63ad4d46..329079ddd 100644
--- a/src/video_core/engines/maxwell_3d.cpp
+++ b/src/video_core/engines/maxwell_3d.cpp
@@ -135,8 +135,6 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
         break;
     }
 
-    rasterizer.NotifyMaxwellRegisterChanged(method);
-
     if (debug_context) {
         debug_context->OnEvent(Tegra::DebugContext::Event::MaxwellCommandProcessed, nullptr);
     }
@@ -293,10 +291,6 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const {
                    tic_entry.header_version == Texture::TICHeaderVersion::Pitch,
                "TIC versions other than BlockLinear or Pitch are unimplemented");
 
-    ASSERT_MSG((tic_entry.texture_type == Texture::TextureType::Texture2D) ||
-                   (tic_entry.texture_type == Texture::TextureType::Texture2DNoMipmap),
-               "Texture types other than Texture2D are unimplemented");
-
     auto r_type = tic_entry.r_type.Value();
     auto g_type = tic_entry.g_type.Value();
     auto b_type = tic_entry.b_type.Value();
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index f59d01738..d3be900a4 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -533,7 +533,11 @@ public:
                 u32 stencil_back_mask;
                 u32 stencil_back_func_mask;
 
-                INSERT_PADDING_WORDS(0x20);
+                INSERT_PADDING_WORDS(0x13);
+
+                u32 rt_separate_frag_data;
+
+                INSERT_PADDING_WORDS(0xC);
 
                 struct {
                     u32 address_high;
@@ -557,7 +561,22 @@ public:
                 struct {
                     union {
                         BitField<0, 4, u32> count;
+                        BitField<4, 3, u32> map_0;
+                        BitField<7, 3, u32> map_1;
+                        BitField<10, 3, u32> map_2;
+                        BitField<13, 3, u32> map_3;
+                        BitField<16, 3, u32> map_4;
+                        BitField<19, 3, u32> map_5;
+                        BitField<22, 3, u32> map_6;
+                        BitField<25, 3, u32> map_7;
                     };
+
+                    u32 GetMap(size_t index) const {
+                        const std::array<u32, NumRenderTargets> maps{map_0, map_1, map_2, map_3,
+                                                                     map_4, map_5, map_6, map_7};
+                        ASSERT(index < maps.size());
+                        return maps[index];
+                    }
                 } rt_control;
 
                 INSERT_PADDING_WORDS(0x2);
@@ -968,6 +987,7 @@ ASSERT_REG_POSITION(clear_stencil, 0x368);
 ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5);
 ASSERT_REG_POSITION(stencil_back_mask, 0x3D6);
 ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7);
+ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB);
 ASSERT_REG_POSITION(zeta, 0x3F8);
 ASSERT_REG_POSITION(vertex_attrib_format, 0x458);
 ASSERT_REG_POSITION(rt_control, 0x487);
diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp
index 6e740713f..c24d33d5c 100644
--- a/src/video_core/engines/maxwell_dma.cpp
+++ b/src/video_core/engines/maxwell_dma.cpp
@@ -41,7 +41,6 @@ void MaxwellDMA::HandleCopy() {
 
     // TODO(Subv): Perform more research and implement all features of this engine.
     ASSERT(regs.exec.enable_swizzle == 0);
-    ASSERT(regs.exec.enable_2d == 1);
     ASSERT(regs.exec.query_mode == Regs::QueryMode::None);
     ASSERT(regs.exec.query_intr == Regs::QueryIntr::None);
     ASSERT(regs.exec.copy_mode == Regs::CopyMode::Unk2);
@@ -51,10 +50,19 @@ void MaxwellDMA::HandleCopy() {
     ASSERT(regs.dst_params.pos_y == 0);
 
     if (regs.exec.is_dst_linear == regs.exec.is_src_linear) {
-        Memory::CopyBlock(dest_cpu, source_cpu, regs.x_count * regs.y_count);
+        size_t copy_size = regs.x_count;
+
+        // 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;
+        }
+
+        Memory::CopyBlock(dest_cpu, source_cpu, copy_size);
         return;
     }
 
+    ASSERT(regs.exec.enable_2d == 1);
     u8* src_buffer = Memory::GetPointer(source_cpu);
     u8* dst_buffer = Memory::GetPointer(dest_cpu);
 
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index d2388673e..58f2904ce 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -244,6 +244,25 @@ enum class TextureType : u64 {
     TextureCube = 3,
 };
 
+enum class TextureQueryType : u64 {
+    Dimension = 1,
+    TextureType = 2,
+    SamplePosition = 5,
+    Filter = 16,
+    LevelOfDetail = 18,
+    Wrap = 20,
+    BorderColor = 22,
+};
+
+enum class TextureProcessMode : u64 {
+    None = 0,
+    LZ = 1,  // Unknown, appears to be the same as none.
+    LB = 2,  // Load Bias.
+    LL = 3,  // Load LOD (LevelOfDetail)
+    LBA = 6, // Load Bias. The A is unknown, does not appear to differ with LB
+    LLA = 7  // Load LOD. The A is unknown, does not appear to differ with LL
+};
+
 enum class IpaInterpMode : u64 { Linear = 0, Perspective = 1, Flat = 2, Sc = 3 };
 enum class IpaSampleMode : u64 { Default = 0, Centroid = 1, Offset = 2 };
 
@@ -414,6 +433,45 @@ union Instruction {
     } bfe;
 
     union {
+        BitField<48, 3, u64> pred48;
+
+        union {
+            BitField<20, 20, u64> entry_a;
+            BitField<39, 5, u64> entry_b;
+            BitField<45, 1, u64> neg;
+            BitField<46, 1, u64> uses_cc;
+        } imm;
+
+        union {
+            BitField<20, 14, u64> cb_index;
+            BitField<34, 5, u64> cb_offset;
+            BitField<56, 1, u64> neg;
+            BitField<57, 1, u64> uses_cc;
+        } hi;
+
+        union {
+            BitField<20, 14, u64> cb_index;
+            BitField<34, 5, u64> cb_offset;
+            BitField<39, 5, u64> entry_a;
+            BitField<45, 1, u64> neg;
+            BitField<46, 1, u64> uses_cc;
+        } rz;
+
+        union {
+            BitField<39, 5, u64> entry_a;
+            BitField<45, 1, u64> neg;
+            BitField<46, 1, u64> uses_cc;
+        } r1;
+
+        union {
+            BitField<28, 8, u64> entry_a;
+            BitField<37, 1, u64> neg;
+            BitField<38, 1, u64> uses_cc;
+        } r2;
+
+    } lea;
+
+    union {
         BitField<0, 5, FlowCondition> cond;
     } flow;
 
@@ -468,6 +526,18 @@ union Instruction {
     } psetp;
 
     union {
+        BitField<12, 3, u64> pred12;
+        BitField<15, 1, u64> neg_pred12;
+        BitField<24, 2, PredOperation> cond;
+        BitField<29, 3, u64> pred29;
+        BitField<32, 1, u64> neg_pred29;
+        BitField<39, 3, u64> pred39;
+        BitField<42, 1, u64> neg_pred39;
+        BitField<44, 1, u64> bf;
+        BitField<45, 2, PredOperation> op;
+    } pset;
+
+    union {
         BitField<39, 3, u64> pred39;
         BitField<42, 1, u64> neg_pred;
         BitField<43, 1, u64> neg_a;
@@ -512,6 +582,7 @@ union Instruction {
         BitField<28, 1, u64> array;
         BitField<29, 2, TextureType> texture_type;
         BitField<31, 4, u64> component_mask;
+        BitField<55, 3, TextureProcessMode> process_mode;
 
         bool IsComponentEnabled(size_t component) const {
             return ((1ull << component) & component_mask) != 0;
@@ -519,6 +590,21 @@ union Instruction {
     } tex;
 
     union {
+        BitField<22, 6, TextureQueryType> query_type;
+        BitField<31, 4, u64> component_mask;
+    } txq;
+
+    union {
+        BitField<28, 1, u64> array;
+        BitField<29, 2, TextureType> texture_type;
+        BitField<31, 4, u64> component_mask;
+
+        bool IsComponentEnabled(size_t component) const {
+            return ((1ull << component) & component_mask) != 0;
+        }
+    } tmml;
+
+    union {
         BitField<28, 1, u64> array;
         BitField<29, 2, TextureType> texture_type;
         BitField<56, 2, u64> component;
@@ -670,11 +756,13 @@ public:
         LDG, // Load from global memory
         STG, // Store in global memory
         TEX,
-        TEXQ,  // Texture Query
-        TEXS,  // Texture Fetch with scalar/non-vec4 source/destinations
-        TLDS,  // Texture Load with scalar/non-vec4 source/destinations
-        TLD4,  // Texture Load 4
-        TLD4S, // Texture Load 4 with scalar / non - vec4 source / destinations
+        TXQ,    // Texture Query
+        TEXS,   // Texture Fetch with scalar/non-vec4 source/destinations
+        TLDS,   // Texture Load with scalar/non-vec4 source/destinations
+        TLD4,   // Texture Load 4
+        TLD4S,  // Texture Load 4 with scalar / non - vec4 source / destinations
+        TMML_B, // Texture Mip Map Level
+        TMML,   // Texture Mip Map Level
         EXIT,
         IPA,
         FFMA_IMM, // Fused Multiply and Add
@@ -699,6 +787,11 @@ public:
         ISCADD_C, // Scale and Add
         ISCADD_R,
         ISCADD_IMM,
+        LEA_R1,
+        LEA_R2,
+        LEA_RZ,
+        LEA_IMM,
+        LEA_HI,
         POPC_C,
         POPC_R,
         POPC_IMM,
@@ -757,6 +850,7 @@ public:
         ISET_C,
         ISET_IMM,
         PSETP,
+        PSET,
         XMAD_IMM,
         XMAD_CR,
         XMAD_RC,
@@ -780,6 +874,7 @@ public:
         IntegerSet,
         IntegerSetPredicate,
         PredicateSetPredicate,
+        PredicateSetRegister,
         Conversion,
         Xmad,
         Unknown,
@@ -894,11 +989,13 @@ private:
             INST("1110111011010---", Id::LDG, Type::Memory, "LDG"),
             INST("1110111011011---", Id::STG, Type::Memory, "STG"),
             INST("110000----111---", Id::TEX, Type::Memory, "TEX"),
-            INST("1101111101001---", Id::TEXQ, Type::Memory, "TEXQ"),
+            INST("1101111101001---", Id::TXQ, Type::Memory, "TXQ"),
             INST("1101100---------", Id::TEXS, Type::Memory, "TEXS"),
             INST("1101101---------", Id::TLDS, Type::Memory, "TLDS"),
             INST("110010----111---", Id::TLD4, Type::Memory, "TLD4"),
             INST("1101111100------", Id::TLD4S, Type::Memory, "TLD4S"),
+            INST("110111110110----", Id::TMML_B, Type::Memory, "TMML_B"),
+            INST("1101111101011---", Id::TMML, Type::Memory, "TMML"),
             INST("111000110000----", Id::EXIT, Type::Trivial, "EXIT"),
             INST("11100000--------", Id::IPA, Type::Trivial, "IPA"),
             INST("0011001-1-------", Id::FFMA_IMM, Type::Ffma, "FFMA_IMM"),
@@ -929,6 +1026,11 @@ private:
             INST("0100110010100---", Id::SEL_C, Type::ArithmeticInteger, "SEL_C"),
             INST("0101110010100---", Id::SEL_R, Type::ArithmeticInteger, "SEL_R"),
             INST("0011100-10100---", Id::SEL_IMM, Type::ArithmeticInteger, "SEL_IMM"),
+            INST("0101101111011---", Id::LEA_R2, Type::ArithmeticInteger, "LEA_R2"),
+            INST("0101101111010---", Id::LEA_R1, Type::ArithmeticInteger, "LEA_R1"),
+            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("0101000010000---", Id::MUFU, Type::Arithmetic, "MUFU"),
             INST("0100110010010---", Id::RRO_C, Type::Arithmetic, "RRO_C"),
             INST("0101110010010---", Id::RRO_R, Type::Arithmetic, "RRO_R"),
@@ -983,6 +1085,7 @@ private:
             INST("010110110101----", Id::ISET_R, Type::IntegerSet, "ISET_R"),
             INST("010010110101----", Id::ISET_C, Type::IntegerSet, "ISET_C"),
             INST("0011011-0101----", Id::ISET_IMM, Type::IntegerSet, "ISET_IMM"),
+            INST("0101000010001---", Id::PSET, Type::PredicateSetRegister, "PSET"),
             INST("0101000010010---", Id::PSETP, Type::PredicateSetPredicate, "PSETP"),
             INST("0011011-00------", Id::XMAD_IMM, Type::Xmad, "XMAD_IMM"),
             INST("0100111---------", Id::XMAD_CR, Type::Xmad, "XMAD_CR"),
diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp
index e6d8e65c6..86a809f86 100644
--- a/src/video_core/gpu.cpp
+++ b/src/video_core/gpu.cpp
@@ -66,6 +66,7 @@ u32 RenderTargetBytesPerPixel(RenderTargetFormat format) {
     case RenderTargetFormat::RGBA8_UINT:
     case RenderTargetFormat::RGB10_A2_UNORM:
     case RenderTargetFormat::BGRA8_UNORM:
+    case RenderTargetFormat::BGRA8_SRGB:
     case RenderTargetFormat::RG16_UNORM:
     case RenderTargetFormat::RG16_SNORM:
     case RenderTargetFormat::RG16_UINT:
diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h
index d29f31f52..589a59b4f 100644
--- a/src/video_core/gpu.h
+++ b/src/video_core/gpu.h
@@ -6,6 +6,7 @@
 
 #include <array>
 #include <memory>
+#include <vector>
 #include "common/common_types.h"
 #include "core/hle/service/nvflinger/buffer_queue.h"
 #include "video_core/memory_manager.h"
@@ -26,6 +27,7 @@ enum class RenderTargetFormat : u32 {
     RG32_FLOAT = 0xCB,
     RG32_UINT = 0xCD,
     BGRA8_UNORM = 0xCF,
+    BGRA8_SRGB = 0xD0,
     RGB10_A2_UNORM = 0xD1,
     RGBA8_UNORM = 0xD5,
     RGBA8_SRGB = 0xD6,
@@ -67,6 +69,7 @@ u32 RenderTargetBytesPerPixel(RenderTargetFormat format);
 /// Returns the number of bytes per pixel of each depth format.
 u32 DepthFormatBytesPerPixel(DepthFormat format);
 
+struct CommandListHeader;
 class DebugContext;
 
 /**
@@ -115,7 +118,7 @@ public:
     ~GPU();
 
     /// Processes a command list stored at the specified address in GPU memory.
-    void ProcessCommandList(GPUVAddr address, u32 size);
+    void ProcessCommandLists(const std::vector<CommandListHeader>& commands);
 
     /// Returns a reference to the Maxwell3D GPU engine.
     Engines::Maxwell3D& Maxwell3D();
@@ -130,9 +133,6 @@ public:
     const Tegra::MemoryManager& MemoryManager() const;
 
 private:
-    /// Writes a single register in the engine bound to the specified subchannel
-    void WriteReg(u32 method, u32 subchannel, u32 value, u32 remaining_params);
-
     std::unique_ptr<Tegra::MemoryManager> memory_manager;
 
     /// Mapping of command subchannels to their bound engine ids.
diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h
index 9d78e8b6b..cd819d69f 100644
--- a/src/video_core/rasterizer_interface.h
+++ b/src/video_core/rasterizer_interface.h
@@ -20,9 +20,6 @@ public:
     /// Clear the current framebuffer
     virtual void Clear() = 0;
 
-    /// Notify rasterizer that the specified Maxwell register has been changed
-    virtual void NotifyMaxwellRegisterChanged(u32 method) = 0;
-
     /// Notify rasterizer that all caches should be flushed to Switch memory
     virtual void FlushAll() = 0;
 
diff --git a/src/video_core/renderer_base.cpp b/src/video_core/renderer_base.cpp
index be17a2b9c..0df3725c2 100644
--- a/src/video_core/renderer_base.cpp
+++ b/src/video_core/renderer_base.cpp
@@ -19,6 +19,7 @@ void RendererBase::RefreshBaseSettings() {
     UpdateCurrentFramebufferLayout();
 
     renderer_settings.use_framelimiter = Settings::values.use_frame_limit;
+    renderer_settings.set_background_color = true;
 }
 
 void RendererBase::UpdateCurrentFramebufferLayout() {
diff --git a/src/video_core/renderer_base.h b/src/video_core/renderer_base.h
index 2a357f9d0..2cd0738ff 100644
--- a/src/video_core/renderer_base.h
+++ b/src/video_core/renderer_base.h
@@ -19,6 +19,7 @@ namespace VideoCore {
 
 struct RendererSettings {
     std::atomic_bool use_framelimiter{false};
+    std::atomic_bool set_background_color{false};
 };
 
 class RendererBase : NonCopyable {
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 0c3bbc475..c59f3af1b 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -294,61 +294,80 @@ void RasterizerOpenGL::UpdatePagesCachedCount(VAddr addr, u64 size, int delta) {
         cached_pages.add({pages_interval, delta});
 }
 
-std::pair<Surface, Surface> RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb,
-                                                                    bool using_depth_fb,
-                                                                    bool preserve_contents) {
+void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_depth_fb,
+                                             bool preserve_contents,
+                                             boost::optional<size_t> single_color_target) {
     MICROPROFILE_SCOPE(OpenGL_Framebuffer);
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
-    if (regs.rt[0].format == Tegra::RenderTargetFormat::NONE) {
-        LOG_ERROR(HW_GPU, "RenderTargetFormat is not configured");
-        using_color_fb = false;
+    Surface depth_surface;
+    if (using_depth_fb) {
+        depth_surface = res_cache.GetDepthBufferSurface(preserve_contents);
     }
 
-    const bool has_stencil = regs.stencil_enable;
-    const bool write_color_fb =
-        state.color_mask.red_enabled == GL_TRUE || state.color_mask.green_enabled == GL_TRUE ||
-        state.color_mask.blue_enabled == GL_TRUE || state.color_mask.alpha_enabled == GL_TRUE;
-
-    const bool write_depth_fb =
-        (state.depth.test_enabled && state.depth.write_mask == GL_TRUE) ||
-        (has_stencil && (state.stencil.front.write_mask || state.stencil.back.write_mask));
+    // TODO(bunnei): Figure out how the below register works. According to envytools, this should be
+    // used to enable multiple render targets. However, it is left unset on all games that I have
+    // tested.
+    ASSERT_MSG(regs.rt_separate_frag_data == 0, "Unimplemented");
 
-    Surface color_surface;
-    Surface depth_surface;
-    MathUtil::Rectangle<u32> surfaces_rect;
-    std::tie(color_surface, depth_surface, surfaces_rect) =
-        res_cache.GetFramebufferSurfaces(using_color_fb, using_depth_fb, preserve_contents);
+    // Bind the framebuffer surfaces
+    state.draw.draw_framebuffer = framebuffer.handle;
+    state.Apply();
 
-    const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()};
-    const MathUtil::Rectangle<u32> draw_rect{
-        static_cast<u32>(std::clamp<s32>(static_cast<s32>(surfaces_rect.left) + viewport_rect.left,
-                                         surfaces_rect.left, surfaces_rect.right)), // Left
-        static_cast<u32>(std::clamp<s32>(static_cast<s32>(surfaces_rect.bottom) + viewport_rect.top,
-                                         surfaces_rect.bottom, surfaces_rect.top)), // Top
-        static_cast<u32>(std::clamp<s32>(static_cast<s32>(surfaces_rect.left) + viewport_rect.right,
-                                         surfaces_rect.left, surfaces_rect.right)), // Right
-        static_cast<u32>(
-            std::clamp<s32>(static_cast<s32>(surfaces_rect.bottom) + viewport_rect.bottom,
-                            surfaces_rect.bottom, surfaces_rect.top))}; // Bottom
+    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);
+            glFramebufferTexture2D(
+                GL_DRAW_FRAMEBUFFER,
+                GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(*single_color_target), GL_TEXTURE_2D,
+                color_surface != nullptr ? color_surface->Texture().handle : 0, 0);
+            glDrawBuffer(GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(*single_color_target));
+        } else {
+            // Multiple color attachments are enabled
+            std::array<GLenum, Maxwell::NumRenderTargets> buffers;
+            for (size_t index = 0; index < Maxwell::NumRenderTargets; ++index) {
+                Surface color_surface = res_cache.GetColorBufferSurface(index, preserve_contents);
+                buffers[index] = GL_COLOR_ATTACHMENT0 + regs.rt_control.GetMap(index);
+                glFramebufferTexture2D(
+                    GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(index),
+                    GL_TEXTURE_2D, color_surface != nullptr ? color_surface->Texture().handle : 0,
+                    0);
+            }
+            glDrawBuffers(regs.rt_control.count, buffers.data());
+        }
+    } else {
+        // No color attachments are enabled - zero out all of them
+        for (size_t index = 0; index < Maxwell::NumRenderTargets; ++index) {
+            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER,
+                                   GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(index), GL_TEXTURE_2D,
+                                   0, 0);
+        }
+        glDrawBuffer(GL_NONE);
+    }
 
-    // Bind the framebuffer surfaces
-    BindFramebufferSurfaces(color_surface, depth_surface, has_stencil);
+    if (depth_surface) {
+        if (regs.stencil_enable) {
+            // Attach both depth and stencil
+            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
+                                   depth_surface->Texture().handle, 0);
+        } else {
+            // Attach depth
+            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D,
+                                   depth_surface->Texture().handle, 0);
+            // Clear stencil attachment
+            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
+        }
+    } else {
+        // Clear both depth and stencil attachment
+        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
+                               0);
+    }
 
-    SyncViewport(surfaces_rect);
+    SyncViewport();
 
-    // Viewport can have negative offsets or larger dimensions than our framebuffer sub-rect. Enable
-    // scissor test to prevent drawing outside of the framebuffer region
-    state.scissor.enabled = true;
-    state.scissor.x = draw_rect.left;
-    state.scissor.y = draw_rect.bottom;
-    state.scissor.width = draw_rect.GetWidth();
-    state.scissor.height = draw_rect.GetHeight();
     state.Apply();
-
-    // Only return the surface to be marked as dirty if writing to it is enabled.
-    return std::make_pair(write_color_fb ? color_surface : nullptr,
-                          write_depth_fb ? depth_surface : nullptr);
 }
 
 void RasterizerOpenGL::Clear() {
@@ -356,8 +375,9 @@ void RasterizerOpenGL::Clear() {
     SCOPE_EXIT({ prev_state.Apply(); });
 
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
-    bool use_color_fb = false;
-    bool use_depth_fb = false;
+    bool use_color{};
+    bool use_depth{};
+    bool use_stencil{};
 
     OpenGLState clear_state;
     clear_state.draw.draw_framebuffer = state.draw.draw_framebuffer;
@@ -366,22 +386,13 @@ void RasterizerOpenGL::Clear() {
     clear_state.color_mask.blue_enabled = regs.clear_buffers.B ? GL_TRUE : GL_FALSE;
     clear_state.color_mask.alpha_enabled = regs.clear_buffers.A ? GL_TRUE : GL_FALSE;
 
-    GLbitfield clear_mask{};
     if (regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B ||
         regs.clear_buffers.A) {
-        if (regs.clear_buffers.RT == 0) {
-            // We only support clearing the first color attachment for now
-            clear_mask |= GL_COLOR_BUFFER_BIT;
-            use_color_fb = true;
-        } else {
-            // TODO(subv): Add support for the other color attachments
-            LOG_CRITICAL(HW_GPU, "Clear unimplemented for RT {}", regs.clear_buffers.RT);
-        }
+        use_color = true;
     }
     if (regs.clear_buffers.Z) {
         ASSERT_MSG(regs.zeta_enable != 0, "Tried to clear Z but buffer is not enabled!");
-        use_depth_fb = true;
-        clear_mask |= GL_DEPTH_BUFFER_BIT;
+        use_depth = true;
 
         // Always enable the depth write when clearing the depth buffer. The depth write mask is
         // ignored when clearing the buffer in the Switch, but OpenGL obeys it so we set it to true.
@@ -390,34 +401,33 @@ void RasterizerOpenGL::Clear() {
     }
     if (regs.clear_buffers.S) {
         ASSERT_MSG(regs.zeta_enable != 0, "Tried to clear stencil but buffer is not enabled!");
-        use_depth_fb = true;
-        clear_mask |= GL_STENCIL_BUFFER_BIT;
+        use_stencil = true;
         clear_state.stencil.test_enabled = true;
     }
 
-    if (!use_color_fb && !use_depth_fb) {
+    if (!use_color && !use_depth && !use_stencil) {
         // No color surface nor depth/stencil surface are enabled
         return;
     }
 
-    if (clear_mask == 0) {
-        // No clear mask is enabled
-        return;
-    }
-
     ScopeAcquireGLContext acquire_context{emu_window};
 
-    auto [dirty_color_surface, dirty_depth_surface] =
-        ConfigureFramebuffers(use_color_fb, use_depth_fb, false);
+    ConfigureFramebuffers(use_color, use_depth || use_stencil, false,
+                          regs.clear_buffers.RT.Value());
 
     clear_state.Apply();
 
-    glClearColor(regs.clear_color[0], regs.clear_color[1], regs.clear_color[2],
-                 regs.clear_color[3]);
-    glClearDepth(regs.clear_depth);
-    glClearStencil(regs.clear_stencil);
+    if (use_color) {
+        glClearBufferfv(GL_COLOR, regs.clear_buffers.RT, regs.clear_color);
+    }
 
-    glClear(clear_mask);
+    if (use_depth && use_stencil) {
+        glClearBufferfi(GL_DEPTH_STENCIL, 0, regs.clear_depth, regs.clear_stencil);
+    } else if (use_depth) {
+        glClearBufferfv(GL_DEPTH, 0, &regs.clear_depth);
+    } else if (use_stencil) {
+        glClearBufferiv(GL_STENCIL, 0, &regs.clear_stencil);
+    }
 }
 
 void RasterizerOpenGL::DrawArrays() {
@@ -430,8 +440,7 @@ void RasterizerOpenGL::DrawArrays() {
 
     ScopeAcquireGLContext acquire_context{emu_window};
 
-    const auto [dirty_color_surface, dirty_depth_surface] =
-        ConfigureFramebuffers(true, regs.zeta.Address() != 0 && regs.zeta_enable != 0, true);
+    ConfigureFramebuffers();
 
     SyncDepthTestState();
     SyncStencilTestState();
@@ -525,8 +534,6 @@ void RasterizerOpenGL::DrawArrays() {
     state.Apply();
 }
 
-void RasterizerOpenGL::NotifyMaxwellRegisterChanged(u32 method) {}
-
 void RasterizerOpenGL::FlushAll() {}
 
 void RasterizerOpenGL::FlushRegion(VAddr addr, u64 size) {}
@@ -586,7 +593,7 @@ bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config,
 void RasterizerOpenGL::SamplerInfo::Create() {
     sampler.Create();
     mag_filter = min_filter = Tegra::Texture::TextureFilter::Linear;
-    wrap_u = wrap_v = Tegra::Texture::WrapMode::Wrap;
+    wrap_u = wrap_v = wrap_p = Tegra::Texture::WrapMode::Wrap;
 
     // default is GL_LINEAR_MIPMAP_LINEAR
     glSamplerParameteri(sampler.handle, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
@@ -613,8 +620,13 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr
         wrap_v = config.wrap_v;
         glSamplerParameteri(s, GL_TEXTURE_WRAP_T, MaxwellToGL::WrapMode(wrap_v));
     }
+    if (wrap_p != config.wrap_p) {
+        wrap_p = config.wrap_p;
+        glSamplerParameteri(s, GL_TEXTURE_WRAP_R, MaxwellToGL::WrapMode(wrap_p));
+    }
 
-    if (wrap_u == Tegra::Texture::WrapMode::Border || wrap_v == Tegra::Texture::WrapMode::Border) {
+    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,
                                           config.border_color_b, config.border_color_a}};
         if (border_color != new_border_color) {
@@ -698,14 +710,15 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
         const auto texture = maxwell3d.GetStageTexture(entry.GetStage(), entry.GetOffset());
 
         if (!texture.enabled) {
-            state.texture_units[current_bindpoint].texture_2d = 0;
+            state.texture_units[current_bindpoint].texture = 0;
             continue;
         }
 
         texture_samplers[current_bindpoint].SyncWithConfig(texture.tsc);
         Surface surface = res_cache.GetTextureSurface(texture);
         if (surface != nullptr) {
-            state.texture_units[current_bindpoint].texture_2d = surface->Texture().handle;
+            state.texture_units[current_bindpoint].texture = surface->Texture().handle;
+            state.texture_units[current_bindpoint].target = surface->Target();
             state.texture_units[current_bindpoint].swizzle.r =
                 MaxwellToGL::SwizzleSource(texture.tic.x_source);
             state.texture_units[current_bindpoint].swizzle.g =
@@ -716,45 +729,19 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
                 MaxwellToGL::SwizzleSource(texture.tic.w_source);
         } else {
             // Can occur when texture addr is null or its memory is unmapped/invalid
-            state.texture_units[current_bindpoint].texture_2d = 0;
+            state.texture_units[current_bindpoint].texture = 0;
         }
     }
 
     return current_unit + static_cast<u32>(entries.size());
 }
 
-void RasterizerOpenGL::BindFramebufferSurfaces(const Surface& color_surface,
-                                               const Surface& depth_surface, bool has_stencil) {
-    state.draw.draw_framebuffer = framebuffer.handle;
-    state.Apply();
-
-    glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D,
-                           color_surface != nullptr ? color_surface->Texture().handle : 0, 0);
-    if (depth_surface != nullptr) {
-        if (has_stencil) {
-            // attach both depth and stencil
-            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
-                                   depth_surface->Texture().handle, 0);
-        } else {
-            // attach depth
-            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D,
-                                   depth_surface->Texture().handle, 0);
-            // clear stencil attachment
-            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
-        }
-    } else {
-        // clear both depth and stencil attachment
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
-                               0);
-    }
-}
-
-void RasterizerOpenGL::SyncViewport(const MathUtil::Rectangle<u32>& surfaces_rect) {
+void RasterizerOpenGL::SyncViewport() {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
     const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()};
 
-    state.viewport.x = static_cast<GLint>(surfaces_rect.left) + viewport_rect.left;
-    state.viewport.y = static_cast<GLint>(surfaces_rect.bottom) + viewport_rect.bottom;
+    state.viewport.x = viewport_rect.left;
+    state.viewport.y = viewport_rect.bottom;
     state.viewport.width = static_cast<GLsizei>(viewport_rect.GetWidth());
     state.viewport.height = static_cast<GLsizei>(viewport_rect.GetHeight());
 }
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index 9c30dc0e8..745c3dc0c 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -13,6 +13,7 @@
 #include <vector>
 
 #include <boost/icl/interval_map.hpp>
+#include <boost/optional.hpp>
 #include <boost/range/iterator_range.hpp>
 #include <glad/glad.h>
 
@@ -45,7 +46,6 @@ public:
 
     void DrawArrays() override;
     void Clear() override;
-    void NotifyMaxwellRegisterChanged(u32 method) override;
     void FlushAll() override;
     void FlushRegion(VAddr addr, u64 size) override;
     void InvalidateRegion(VAddr addr, u64 size) override;
@@ -93,17 +93,20 @@ private:
         Tegra::Texture::TextureFilter min_filter;
         Tegra::Texture::WrapMode wrap_u;
         Tegra::Texture::WrapMode wrap_v;
+        Tegra::Texture::WrapMode wrap_p;
         GLvec4 border_color;
     };
 
-    /// Configures the color and depth framebuffer states and returns the dirty <Color, Depth>
-    /// surfaces if writing was enabled.
-    std::pair<Surface, Surface> ConfigureFramebuffers(bool using_color_fb, bool using_depth_fb,
-                                                      bool preserve_contents);
-
-    /// Binds the framebuffer color and depth surface
-    void BindFramebufferSurfaces(const Surface& color_surface, const Surface& depth_surface,
-                                 bool has_stencil);
+    /**
+     * Configures the color and depth framebuffer states.
+     * @param use_color_fb If true, configure color framebuffers.
+     * @param using_depth_fb If true, configure the depth/stencil framebuffer.
+     * @param preserve_contents If true, tries to preserve data from a previously used framebuffer.
+     * @param single_color_target Specifies if a single color buffer target should be used.
+     */
+    void ConfigureFramebuffers(bool use_color_fb = true, bool using_depth_fb = true,
+                               bool preserve_contents = true,
+                               boost::optional<size_t> single_color_target = {});
 
     /*
      * Configures the current constbuffers to use for the draw command.
@@ -126,7 +129,7 @@ private:
                       u32 current_unit);
 
     /// Syncs the viewport to match the guest state
-    void SyncViewport(const MathUtil::Rectangle<u32>& surfaces_rect);
+    void SyncViewport();
 
     /// Syncs the clip enabled status to match the guest state
     void SyncClipEnabled();
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index f6b2c5a86..32001e44b 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -7,6 +7,7 @@
 
 #include "common/alignment.h"
 #include "common/assert.h"
+#include "common/logging/log.h"
 #include "common/microprofile.h"
 #include "common/scope_exit.h"
 #include "core/core.h"
@@ -52,14 +53,30 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
     params.width = Common::AlignUp(config.tic.Width(), GetCompressionFactor(params.pixel_format));
     params.height = Common::AlignUp(config.tic.Height(), GetCompressionFactor(params.pixel_format));
     params.unaligned_height = config.tic.Height();
+    params.target = SurfaceTargetFromTextureType(config.tic.texture_type);
+
+    switch (params.target) {
+    case SurfaceTarget::Texture1D:
+    case SurfaceTarget::Texture2D:
+        params.depth = 1;
+        break;
+    case SurfaceTarget::Texture3D:
+    case SurfaceTarget::Texture2DArray:
+        params.depth = config.tic.Depth();
+        break;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unknown depth for target={}", static_cast<u32>(params.target));
+        UNREACHABLE();
+        params.depth = 1;
+        break;
+    }
+
     params.size_in_bytes = params.SizeInBytes();
-    params.cache_width = Common::AlignUp(params.width, 16);
-    params.cache_height = Common::AlignUp(params.height, 16);
     return params;
 }
 
-/*static*/ SurfaceParams SurfaceParams::CreateForFramebuffer(
-    const Tegra::Engines::Maxwell3D::Regs::RenderTargetConfig& config) {
+/*static*/ SurfaceParams SurfaceParams::CreateForFramebuffer(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;
@@ -70,9 +87,9 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
     params.width = config.width;
     params.height = config.height;
     params.unaligned_height = config.height;
+    params.target = SurfaceTarget::Texture2D;
+    params.depth = 1;
     params.size_in_bytes = params.SizeInBytes();
-    params.cache_width = Common::AlignUp(params.width, 16);
-    params.cache_height = Common::AlignUp(params.height, 16);
     return params;
 }
 
@@ -86,13 +103,12 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
     params.pixel_format = PixelFormatFromDepthFormat(format);
     params.component_type = ComponentTypeFromDepthFormat(format);
     params.type = GetFormatType(params.pixel_format);
-    params.size_in_bytes = params.SizeInBytes();
     params.width = zeta_width;
     params.height = zeta_height;
     params.unaligned_height = zeta_height;
+    params.target = SurfaceTarget::Texture2D;
+    params.depth = 1;
     params.size_in_bytes = params.SizeInBytes();
-    params.cache_width = Common::AlignUp(params.width, 16);
-    params.cache_height = Common::AlignUp(params.height, 16);
     return params;
 }
 
@@ -100,7 +116,7 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
     {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
-    {GL_RGB, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, ComponentType::UNorm, false},    // B5G6R5U
+    {GL_RGB8, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, ComponentType::UNorm, false},   // B5G6R5U
     {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV, ComponentType::UNorm,
      false}, // A2B10G10R10U
     {GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_1_5_5_5_REV, ComponentType::UNorm, false}, // A1B5G5R5U
@@ -166,6 +182,26 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
      ComponentType::Float, false}, // Z32FS8
 }};
 
+static GLenum SurfaceTargetToGL(SurfaceParams::SurfaceTarget target) {
+    switch (target) {
+    case SurfaceParams::SurfaceTarget::Texture1D:
+        return GL_TEXTURE_1D;
+    case SurfaceParams::SurfaceTarget::Texture2D:
+        return GL_TEXTURE_2D;
+    case SurfaceParams::SurfaceTarget::Texture3D:
+        return GL_TEXTURE_3D;
+    case SurfaceParams::SurfaceTarget::Texture1DArray:
+        return GL_TEXTURE_1D_ARRAY;
+    case SurfaceParams::SurfaceTarget::Texture2DArray:
+        return GL_TEXTURE_2D_ARRAY;
+    case SurfaceParams::SurfaceTarget::TextureCubemap:
+        return GL_TEXTURE_CUBE_MAP;
+    }
+    LOG_CRITICAL(Render_OpenGL, "Unimplemented texture target={}", static_cast<u32>(target));
+    UNREACHABLE();
+    return {};
+}
+
 static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType component_type) {
     ASSERT(static_cast<size_t>(pixel_format) < tex_format_tuples.size());
     auto& format = tex_format_tuples[static_cast<unsigned int>(pixel_format)];
@@ -220,7 +256,8 @@ static bool IsFormatBCn(PixelFormat format) {
 }
 
 template <bool morton_to_gl, PixelFormat format>
-void MortonCopy(u32 stride, u32 block_height, u32 height, std::vector<u8>& gl_buffer, VAddr addr) {
+void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, 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);
 
@@ -230,18 +267,18 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, std::vector<u8>& gl_bu
         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 size_t size_to_copy{std::min(gl_buffer.size(), data.size())};
-        gl_buffer.assign(data.begin(), data.begin() + size_to_copy);
+        const 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.data(), morton_to_gl);
+                                       Memory::GetPointer(addr), gl_buffer, morton_to_gl);
     }
 }
 
-static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, VAddr),
+static constexpr std::array<void (*)(u32, u32, u32, u8*, size_t, VAddr),
                             SurfaceParams::MaxPixelFormat>
     morton_to_gl_fns = {
         // clang-format off
@@ -298,7 +335,7 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, VAddr),
         // clang-format on
 };
 
-static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, VAddr),
+static constexpr std::array<void (*)(u32, u32, u32, u8*, size_t, VAddr),
                             SurfaceParams::MaxPixelFormat>
     gl_to_morton_fns = {
         // clang-format off
@@ -357,33 +394,6 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, VAddr),
         // clang-format on
 };
 
-// Allocate an uninitialized texture of appropriate size and format for the surface
-static void AllocateSurfaceTexture(GLuint texture, const FormatTuple& format_tuple, u32 width,
-                                   u32 height) {
-    OpenGLState cur_state = OpenGLState::GetCurState();
-
-    // Keep track of previous texture bindings
-    GLuint old_tex = cur_state.texture_units[0].texture_2d;
-    cur_state.texture_units[0].texture_2d = texture;
-    cur_state.Apply();
-    glActiveTexture(GL_TEXTURE0);
-
-    if (!format_tuple.compressed) {
-        // Only pre-create the texture for non-compressed textures.
-        glTexImage2D(GL_TEXTURE_2D, 0, format_tuple.internal_format, width, height, 0,
-                     format_tuple.format, format_tuple.type, nullptr);
-    }
-
-    glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAX_LEVEL, 0);
-    glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
-    glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
-
-    // Restore previous texture bindings
-    cur_state.texture_units[0].texture_2d = old_tex;
-    cur_state.Apply();
-}
-
 static bool BlitTextures(GLuint src_tex, const MathUtil::Rectangle<u32>& src_rect, GLuint dst_tex,
                          const MathUtil::Rectangle<u32>& dst_rect, SurfaceType type,
                          GLuint read_fb_handle, GLuint draw_fb_handle) {
@@ -438,12 +448,53 @@ static bool BlitTextures(GLuint src_tex, const MathUtil::Rectangle<u32>& src_rec
     return true;
 }
 
-CachedSurface::CachedSurface(const SurfaceParams& params) : params(params) {
+CachedSurface::CachedSurface(const SurfaceParams& params)
+    : params(params), gl_target(SurfaceTargetToGL(params.target)) {
     texture.Create();
     const auto& rect{params.GetRect()};
-    AllocateSurfaceTexture(texture.handle,
-                           GetFormatTuple(params.pixel_format, params.component_type),
+
+    // Keep track of previous texture bindings
+    OpenGLState cur_state = OpenGLState::GetCurState();
+    const auto& old_tex = cur_state.texture_units[0];
+    SCOPE_EXIT({
+        cur_state.texture_units[0] = old_tex;
+        cur_state.Apply();
+    });
+
+    cur_state.texture_units[0].texture = texture.handle;
+    cur_state.texture_units[0].target = SurfaceTargetToGL(params.target);
+    cur_state.Apply();
+    glActiveTexture(GL_TEXTURE0);
+
+    const auto& format_tuple = GetFormatTuple(params.pixel_format, params.component_type);
+    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());
+            break;
+        case SurfaceParams::SurfaceTarget::Texture2D:
+            glTexStorage2D(SurfaceTargetToGL(params.target), 1, 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);
+            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());
+        }
+    }
+
+    glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_MIN_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);
 }
 
 static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) {
@@ -461,7 +512,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) {
 
     S8Z24 input_pixel{};
     Z24S8 output_pixel{};
-    const auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::S8Z24)};
+    constexpr auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::S8Z24)};
     for (size_t y = 0; y < height; ++y) {
         for (size_t x = 0; x < width; ++x) {
             const size_t offset{bpp * (y * width + x)};
@@ -474,7 +525,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) {
 }
 
 static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
-    const auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::G8R8U)};
+    constexpr auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::G8R8U)};
     for (size_t y = 0; y < height; ++y) {
         for (size_t x = 0; x < width; ++x) {
             const size_t offset{bpp * (y * width + x)};
@@ -514,23 +565,6 @@ static void ConvertFormatAsNeeded_LoadGLBuffer(std::vector<u8>& data, PixelForma
     }
 }
 
-/**
- * Helper function to perform software conversion (as needed) when flushing a buffer 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::ASTC_2D_4X4:
-    case PixelFormat::S8Z24:
-        LOG_CRITICAL(Render_OpenGL, "Unimplemented pixel_format={}",
-                     static_cast<u32>(pixel_format));
-        UNREACHABLE();
-        break;
-    }
-}
-
 MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 64, 192));
 void CachedSurface::LoadGLBuffer() {
     ASSERT(params.type != SurfaceType::Fill);
@@ -545,13 +579,25 @@ void CachedSurface::LoadGLBuffer() {
     MICROPROFILE_SCOPE(OpenGL_SurfaceLoad);
 
     if (params.is_tiled) {
-        gl_buffer.resize(copy_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;
+        default:
+            LOG_CRITICAL(HW_GPU, "Unimplemented tiled load for target={}",
+                         static_cast<u32>(params.target));
+            UNREACHABLE();
+        }
 
+        gl_buffer.resize(static_cast<size_t>(params.depth) * copy_size);
         morton_to_gl_fns[static_cast<size_t>(params.pixel_format)](
-            params.width, params.block_height, params.height, gl_buffer, params.addr);
+            params.width, params.block_height, params.height, gl_buffer.data(), copy_size,
+            params.addr);
     } else {
-        const u8* const texture_src_data_end = texture_src_data + copy_size;
-
+        const u8* const texture_src_data_end{texture_src_data +
+                                             (static_cast<size_t>(params.depth) * copy_size)};
         gl_buffer.assign(texture_src_data, texture_src_data_end);
     }
 
@@ -560,23 +606,7 @@ void CachedSurface::LoadGLBuffer() {
 
 MICROPROFILE_DEFINE(OpenGL_SurfaceFlush, "OpenGL", "Surface Flush", MP_RGB(128, 192, 64));
 void CachedSurface::FlushGLBuffer() {
-    u8* const dst_buffer = Memory::GetPointer(params.addr);
-
-    ASSERT(dst_buffer);
-    ASSERT(gl_buffer.size() ==
-           params.width * params.height * GetGLBytesPerPixel(params.pixel_format));
-
-    MICROPROFILE_SCOPE(OpenGL_SurfaceFlush);
-
-    ConvertFormatAsNeeded_FlushGLBuffer(gl_buffer, params.pixel_format, params.width,
-                                        params.height);
-
-    if (!params.is_tiled) {
-        std::memcpy(dst_buffer, gl_buffer.data(), params.size_in_bytes);
-    } else {
-        gl_to_morton_fns[static_cast<size_t>(params.pixel_format)](
-            params.width, params.block_height, params.height, gl_buffer, params.addr);
-    }
+    ASSERT_MSG(false, "Unimplemented");
 }
 
 MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 64, 192));
@@ -586,22 +616,29 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
 
     MICROPROFILE_SCOPE(OpenGL_TextureUL);
 
-    ASSERT(gl_buffer.size() ==
-           params.width * params.height * GetGLBytesPerPixel(params.pixel_format));
+    ASSERT(gl_buffer.size() == static_cast<size_t>(params.width) * params.height *
+                                   GetGLBytesPerPixel(params.pixel_format) * params.depth);
 
     const auto& rect{params.GetRect()};
 
     // Load data from memory to the surface
-    GLint x0 = static_cast<GLint>(rect.left);
-    GLint y0 = static_cast<GLint>(rect.bottom);
-    size_t buffer_offset = (y0 * params.width + x0) * GetGLBytesPerPixel(params.pixel_format);
+    const GLint x0 = static_cast<GLint>(rect.left);
+    const GLint y0 = static_cast<GLint>(rect.bottom);
+    const size_t buffer_offset =
+        static_cast<size_t>(static_cast<size_t>(y0) * params.width + static_cast<size_t>(x0)) *
+        GetGLBytesPerPixel(params.pixel_format);
 
     const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
-    GLuint target_tex = texture.handle;
+    const GLuint target_tex = texture.handle;
     OpenGLState cur_state = OpenGLState::GetCurState();
 
-    GLuint old_tex = cur_state.texture_units[0].texture_2d;
-    cur_state.texture_units[0].texture_2d = target_tex;
+    const auto& old_tex = cur_state.texture_units[0];
+    SCOPE_EXIT({
+        cur_state.texture_units[0] = old_tex;
+        cur_state.Apply();
+    });
+    cur_state.texture_units[0].texture = target_tex;
+    cur_state.texture_units[0].target = SurfaceTargetToGL(params.target);
     cur_state.Apply();
 
     // Ensure no bad interactions with GL_UNPACK_ALIGNMENT
@@ -610,136 +647,102 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
 
     glActiveTexture(GL_TEXTURE0);
     if (tuple.compressed) {
-        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),
-            &gl_buffer[buffer_offset]);
+        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), &gl_buffer[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),
+                &gl_buffer[buffer_offset]);
+            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),
+                &gl_buffer[buffer_offset]);
+        }
     } else {
-        glTexSubImage2D(GL_TEXTURE_2D, 0, x0, y0, static_cast<GLsizei>(rect.GetWidth()),
-                        static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
-                        &gl_buffer[buffer_offset]);
-    }
-
-    glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
-
-    cur_state.texture_units[0].texture_2d = old_tex;
-    cur_state.Apply();
-}
-
-MICROPROFILE_DEFINE(OpenGL_TextureDL, "OpenGL", "Texture Download", MP_RGB(128, 192, 64));
-void CachedSurface::DownloadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle) {
-    if (params.type == SurfaceType::Fill)
-        return;
-
-    MICROPROFILE_SCOPE(OpenGL_TextureDL);
-
-    gl_buffer.resize(params.width * params.height * GetGLBytesPerPixel(params.pixel_format));
-
-    OpenGLState state = OpenGLState::GetCurState();
-    OpenGLState prev_state = state;
-    SCOPE_EXIT({ prev_state.Apply(); });
-
-    const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
-
-    // Ensure no bad interactions with GL_PACK_ALIGNMENT
-    ASSERT(params.width * GetGLBytesPerPixel(params.pixel_format) % 4 == 0);
-    glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(params.width));
-
-    const auto& rect{params.GetRect()};
-    size_t buffer_offset =
-        (rect.bottom * params.width + rect.left) * GetGLBytesPerPixel(params.pixel_format);
-
-    state.UnbindTexture(texture.handle);
-    state.draw.read_framebuffer = read_fb_handle;
-    state.Apply();
 
-    if (params.type == SurfaceType::ColorTexture) {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D,
-                               texture.handle, 0);
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
-                               0);
-    } else if (params.type == SurfaceType::Depth) {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D,
-                               texture.handle, 0);
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
-    } else {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
-                               texture.handle, 0);
+        switch (params.target) {
+        case SurfaceParams::SurfaceTarget::Texture1D:
+            glTexSubImage1D(SurfaceTargetToGL(params.target), 0, x0,
+                            static_cast<GLsizei>(rect.GetWidth()), tuple.format, tuple.type,
+                            &gl_buffer[buffer_offset]);
+            break;
+        case SurfaceParams::SurfaceTarget::Texture2D:
+            glTexSubImage2D(SurfaceTargetToGL(params.target), 0, x0, y0,
+                            static_cast<GLsizei>(rect.GetWidth()),
+                            static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
+                            &gl_buffer[buffer_offset]);
+            break;
+        case SurfaceParams::SurfaceTarget::Texture3D:
+        case SurfaceParams::SurfaceTarget::Texture2DArray:
+            glTexSubImage3D(SurfaceTargetToGL(params.target), 0, x0, y0, 0,
+                            static_cast<GLsizei>(rect.GetWidth()),
+                            static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format,
+                            tuple.type, &gl_buffer[buffer_offset]);
+            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()),
+                            static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
+                            &gl_buffer[buffer_offset]);
+        }
     }
-    glReadPixels(static_cast<GLint>(rect.left), static_cast<GLint>(rect.bottom),
-                 static_cast<GLsizei>(rect.GetWidth()), static_cast<GLsizei>(rect.GetHeight()),
-                 tuple.format, tuple.type, &gl_buffer[buffer_offset]);
 
-    glPixelStorei(GL_PACK_ROW_LENGTH, 0);
+    glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
 }
 
 RasterizerCacheOpenGL::RasterizerCacheOpenGL() {
     read_framebuffer.Create();
     draw_framebuffer.Create();
+    copy_pbo.Create();
 }
 
 Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextureInfo& config) {
     return GetSurface(SurfaceParams::CreateForTexture(config));
 }
 
-SurfaceSurfaceRect_Tuple RasterizerCacheOpenGL::GetFramebufferSurfaces(bool using_color_fb,
-                                                                       bool using_depth_fb,
-                                                                       bool preserve_contents) {
-    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+Surface RasterizerCacheOpenGL::GetDepthBufferSurface(bool preserve_contents) {
+    const auto& regs{Core::System::GetInstance().GPU().Maxwell3D().regs};
+    if (!regs.zeta.Address() || !regs.zeta_enable) {
+        return {};
+    }
 
-    // TODO(bunnei): This is hard corded to use just the first render buffer
-    LOG_TRACE(Render_OpenGL, "hard-coded for render target 0!");
+    SurfaceParams depth_params{SurfaceParams::CreateForDepthBuffer(
+        regs.zeta_width, regs.zeta_height, regs.zeta.Address(), regs.zeta.format)};
 
-    // get color and depth surfaces
-    SurfaceParams color_params{};
-    SurfaceParams depth_params{};
+    return GetSurface(depth_params, preserve_contents);
+}
 
-    if (using_color_fb) {
-        color_params = SurfaceParams::CreateForFramebuffer(regs.rt[0]);
-    }
+Surface RasterizerCacheOpenGL::GetColorBufferSurface(size_t index, bool preserve_contents) {
+    const auto& regs{Core::System::GetInstance().GPU().Maxwell3D().regs};
 
-    if (using_depth_fb) {
-        depth_params = SurfaceParams::CreateForDepthBuffer(regs.zeta_width, regs.zeta_height,
-                                                           regs.zeta.Address(), regs.zeta.format);
-    }
+    ASSERT(index < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets);
 
-    MathUtil::Rectangle<u32> color_rect{};
-    Surface color_surface;
-    if (using_color_fb) {
-        color_surface = GetSurface(color_params, preserve_contents);
-        if (color_surface) {
-            color_rect = color_surface->GetSurfaceParams().GetRect();
-        }
+    if (index >= regs.rt_control.count) {
+        return {};
     }
 
-    MathUtil::Rectangle<u32> depth_rect{};
-    Surface depth_surface;
-    if (using_depth_fb) {
-        depth_surface = GetSurface(depth_params, preserve_contents);
-        if (depth_surface) {
-            depth_rect = depth_surface->GetSurfaceParams().GetRect();
-        }
+    if (regs.rt[index].Address() == 0 || regs.rt[index].format == Tegra::RenderTargetFormat::NONE) {
+        return {};
     }
 
-    MathUtil::Rectangle<u32> fb_rect{};
-    if (color_surface && depth_surface) {
-        fb_rect = color_rect;
-        // Color and Depth surfaces must have the same dimensions and offsets
-        if (color_rect.bottom != depth_rect.bottom || color_rect.top != depth_rect.top ||
-            color_rect.left != depth_rect.left || color_rect.right != depth_rect.right) {
-            color_surface = GetSurface(color_params);
-            depth_surface = GetSurface(depth_params);
-            fb_rect = color_surface->GetSurfaceParams().GetRect();
-        }
-    } else if (color_surface) {
-        fb_rect = color_rect;
-    } else if (depth_surface) {
-        fb_rect = depth_rect;
-    }
+    const SurfaceParams color_params{SurfaceParams::CreateForFramebuffer(index)};
 
-    return std::make_tuple(color_surface, depth_surface, fb_rect);
+    return GetSurface(color_params, preserve_contents);
 }
 
 void RasterizerCacheOpenGL::LoadSurface(const Surface& surface) {
@@ -748,7 +751,6 @@ void RasterizerCacheOpenGL::LoadSurface(const Surface& surface) {
 }
 
 void RasterizerCacheOpenGL::FlushSurface(const Surface& surface) {
-    surface->DownloadGLTexture(read_framebuffer.handle, draw_framebuffer.handle);
     surface->FlushGLBuffer();
 }
 
@@ -806,27 +808,26 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& surface,
     // Get a new surface with the new parameters, and blit the previous surface to it
     Surface new_surface{GetUncachedSurface(new_params)};
 
-    // If format is unchanged, we can do a faster blit without reinterpreting pixel data
-    if (params.pixel_format == new_params.pixel_format) {
+    if (params.pixel_format == new_params.pixel_format ||
+        !Settings::values.use_accurate_framebuffers) {
+        // 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.
+
         BlitTextures(surface->Texture().handle, params.GetRect(), new_surface->Texture().handle,
-                     new_surface->GetSurfaceParams().GetRect(), params.type,
-                     read_framebuffer.handle, draw_framebuffer.handle);
-        return new_surface;
-    }
+                     params.GetRect(), params.type, read_framebuffer.handle,
+                     draw_framebuffer.handle);
+    } else {
+        // When use_accurate_framebuffers 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.
 
-    // When using accurate framebuffers, always copy old data to new surface, regardless of format
-    if (Settings::values.use_accurate_framebuffers) {
         auto source_format = GetFormatTuple(params.pixel_format, params.component_type);
         auto dest_format = GetFormatTuple(new_params.pixel_format, new_params.component_type);
 
         size_t buffer_size = std::max(params.SizeInBytes(), new_params.SizeInBytes());
 
-        // Use a Pixel Buffer Object to download the previous texture and then upload it to the new
-        // one using the new format.
-        OGLBuffer pbo;
-        pbo.Create();
-
-        glBindBuffer(GL_PIXEL_PACK_BUFFER, pbo.handle);
+        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(surface->Texture().handle, 0,
@@ -845,8 +846,8 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& surface,
                 // of the data in this case. Games like Super Mario Odyssey seem to hit this case
                 // when drawing, it re-uses the memory of a previous texture as a bigger framebuffer
                 // but it doesn't clear it beforehand, the texture is already full of zeros.
-                LOG_CRITICAL(HW_GPU, "Trying to upload extra texture data from the CPU during "
-                                     "reinterpretation but the texture is tiled.");
+                LOG_DEBUG(HW_GPU, "Trying to upload extra texture data from the CPU during "
+                                  "reinterpretation but the texture is tiled.");
             }
             size_t remaining_size = new_params.SizeInBytes() - params.SizeInBytes();
             std::vector<u8> data(remaining_size);
@@ -859,21 +860,38 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& surface,
 
         const auto& dest_rect{new_params.GetRect()};
 
-        glBindBuffer(GL_PIXEL_UNPACK_BUFFER, pbo.handle);
+        glBindBuffer(GL_PIXEL_UNPACK_BUFFER, copy_pbo.handle);
         if (dest_format.compressed) {
-            glCompressedTexSubImage2D(
-                GL_TEXTURE_2D, 0, 0, 0, static_cast<GLsizei>(dest_rect.GetWidth()),
-                static_cast<GLsizei>(dest_rect.GetHeight()), dest_format.format,
-                static_cast<GLsizei>(new_params.SizeInBytes()), nullptr);
+            LOG_CRITICAL(HW_GPU, "Compressed copy is unimplemented!");
+            UNREACHABLE();
         } else {
-            glTextureSubImage2D(new_surface->Texture().handle, 0, 0, 0,
-                                static_cast<GLsizei>(dest_rect.GetWidth()),
-                                static_cast<GLsizei>(dest_rect.GetHeight()), dest_format.format,
-                                dest_format.type, nullptr);
+            switch (new_params.target) {
+            case SurfaceParams::SurfaceTarget::Texture1D:
+                glTextureSubImage1D(new_surface->Texture().handle, 0, 0,
+                                    static_cast<GLsizei>(dest_rect.GetWidth()), dest_format.format,
+                                    dest_format.type, nullptr);
+                break;
+            case SurfaceParams::SurfaceTarget::Texture2D:
+                glTextureSubImage2D(new_surface->Texture().handle, 0, 0, 0,
+                                    static_cast<GLsizei>(dest_rect.GetWidth()),
+                                    static_cast<GLsizei>(dest_rect.GetHeight()), dest_format.format,
+                                    dest_format.type, nullptr);
+                break;
+            case SurfaceParams::SurfaceTarget::Texture3D:
+            case SurfaceParams::SurfaceTarget::Texture2DArray:
+                glTextureSubImage3D(new_surface->Texture().handle, 0, 0, 0, 0,
+                                    static_cast<GLsizei>(dest_rect.GetWidth()),
+                                    static_cast<GLsizei>(dest_rect.GetHeight()),
+                                    static_cast<GLsizei>(new_params.depth), dest_format.format,
+                                    dest_format.type, nullptr);
+                break;
+            default:
+                LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
+                             static_cast<u32>(params.target));
+                UNREACHABLE();
+            }
         }
         glBindBuffer(GL_PIXEL_UNPACK_BUFFER, 0);
-
-        pbo.Release();
     }
 
     return new_surface;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
index aad75f200..57ea8593b 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
@@ -109,6 +109,33 @@ struct SurfaceParams {
         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::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;
+        }
+    }
+
     /**
      * 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
@@ -270,6 +297,7 @@ struct SurfaceParams {
             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:
@@ -542,6 +570,7 @@ struct SurfaceParams {
         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:
@@ -635,15 +664,14 @@ struct SurfaceParams {
         ASSERT(width % compression_factor == 0);
         ASSERT(height % compression_factor == 0);
         return (width / compression_factor) * (height / compression_factor) *
-               GetFormatBpp(pixel_format) / CHAR_BIT;
+               GetFormatBpp(pixel_format) * depth / CHAR_BIT;
     }
 
     /// Creates SurfaceParams from a texture configuration
     static SurfaceParams CreateForTexture(const Tegra::Texture::FullTextureInfo& config);
 
     /// Creates SurfaceParams from a framebuffer configuration
-    static SurfaceParams CreateForFramebuffer(
-        const Tegra::Engines::Maxwell3D::Regs::RenderTargetConfig& config);
+    static SurfaceParams CreateForFramebuffer(size_t index);
 
     /// Creates SurfaceParams for a depth buffer configuration
     static SurfaceParams CreateForDepthBuffer(u32 zeta_width, u32 zeta_height,
@@ -652,8 +680,8 @@ struct SurfaceParams {
 
     /// Checks if surfaces are compatible for caching
     bool IsCompatibleSurface(const SurfaceParams& other) const {
-        return std::tie(pixel_format, type, cache_width, cache_height) ==
-               std::tie(other.pixel_format, other.type, other.cache_width, other.cache_height);
+        return std::tie(pixel_format, type, width, height) ==
+               std::tie(other.pixel_format, other.type, other.width, other.height);
     }
 
     VAddr addr;
@@ -664,12 +692,10 @@ struct SurfaceParams {
     SurfaceType type;
     u32 width;
     u32 height;
+    u32 depth;
     u32 unaligned_height;
     size_t size_in_bytes;
-
-    // Parameters used for caching only
-    u32 cache_width;
-    u32 cache_height;
+    SurfaceTarget target;
 };
 
 }; // namespace OpenGL
@@ -709,6 +735,10 @@ public:
         return texture;
     }
 
+    GLenum Target() const {
+        return gl_target;
+    }
+
     static constexpr unsigned int GetGLBytesPerPixel(SurfaceParams::PixelFormat format) {
         if (format == SurfaceParams::PixelFormat::Invalid)
             return 0;
@@ -724,14 +754,14 @@ public:
     void LoadGLBuffer();
     void FlushGLBuffer();
 
-    // Upload/Download data in gl_buffer in/to this surface's texture
+    // Upload data in gl_buffer to this surface's texture
     void UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle);
-    void DownloadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle);
 
 private:
     OGLTexture texture;
     std::vector<u8> gl_buffer;
     SurfaceParams params;
+    GLenum gl_target;
 };
 
 class RasterizerCacheOpenGL final : public RasterizerCache<Surface> {
@@ -741,9 +771,11 @@ public:
     /// Get a surface based on the texture configuration
     Surface GetTextureSurface(const Tegra::Texture::FullTextureInfo& config);
 
-    /// Get the color and depth surfaces based on the framebuffer configuration
-    SurfaceSurfaceRect_Tuple GetFramebufferSurfaces(bool using_color_fb, bool using_depth_fb,
-                                                    bool preserve_contents);
+    /// Get the depth surface based on the framebuffer configuration
+    Surface GetDepthBufferSurface(bool preserve_contents);
+
+    /// Get the color surface based on the framebuffer configuration and the specified render target
+    Surface GetColorBufferSurface(size_t index, bool preserve_contents);
 
     /// Flushes the surface to Switch memory
     void FlushSurface(const Surface& surface);
@@ -774,6 +806,10 @@ private:
 
     OGLFramebuffer read_framebuffer;
     OGLFramebuffer draw_framebuffer;
+
+    /// Use a Pixel Buffer Object to download the previous texture and then upload it to the new one
+    /// using the new format.
+    OGLBuffer copy_pbo;
 };
 
 } // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 7e4b85ac3..61080f5cc 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -13,8 +13,8 @@ namespace OpenGL {
 
 /// Gets the address for the specified shader stage program
 static VAddr GetShaderAddress(Maxwell::ShaderProgram program) {
-    auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
-    auto& shader_config = gpu.regs.shader_config[static_cast<size_t>(program)];
+    const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+    const auto& shader_config = gpu.regs.shader_config[static_cast<size_t>(program)];
     return *gpu.memory_manager.GpuToCpuAddress(gpu.regs.code_address.CodeAddress() +
                                                shader_config.offset);
 }
@@ -86,7 +86,7 @@ CachedShader::CachedShader(VAddr addr, Maxwell::ShaderProgram program_type)
 }
 
 GLuint CachedShader::GetProgramResourceIndex(const GLShader::ConstBufferEntry& buffer) {
-    auto search{resource_cache.find(buffer.GetHash())};
+    const auto search{resource_cache.find(buffer.GetHash())};
     if (search == resource_cache.end()) {
         const GLuint index{
             glGetProgramResourceIndex(program.handle, GL_UNIFORM_BLOCK, buffer.GetName().c_str())};
@@ -98,7 +98,7 @@ GLuint CachedShader::GetProgramResourceIndex(const GLShader::ConstBufferEntry& b
 }
 
 GLint CachedShader::GetUniformLocation(const GLShader::SamplerEntry& sampler) {
-    auto search{uniform_cache.find(sampler.GetHash())};
+    const auto search{uniform_cache.find(sampler.GetHash())};
     if (search == uniform_cache.end()) {
         const GLint index{glGetUniformLocation(program.handle, sampler.GetName().c_str())};
         uniform_cache[sampler.GetHash()] = index;
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index 841647ebe..2d56370c7 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -113,7 +113,7 @@ private:
 
     /// Scans a range of code for labels and determines the exit method.
     ExitMethod Scan(u32 begin, u32 end, std::set<u32>& labels) {
-        auto [iter, inserted] =
+        const auto [iter, inserted] =
             exit_method_map.emplace(std::make_pair(begin, end), ExitMethod::Undetermined);
         ExitMethod& exit_method = iter->second;
         if (!inserted)
@@ -131,22 +131,22 @@ private:
                     if (instr.pred.pred_index == static_cast<u64>(Pred::UnusedIndex)) {
                         return exit_method = ExitMethod::AlwaysEnd;
                     } else {
-                        ExitMethod not_met = Scan(offset + 1, end, labels);
+                        const ExitMethod not_met = Scan(offset + 1, end, labels);
                         return exit_method = ParallelExit(ExitMethod::AlwaysEnd, not_met);
                     }
                 }
                 case OpCode::Id::BRA: {
-                    u32 target = offset + instr.bra.GetBranchTarget();
+                    const u32 target = offset + instr.bra.GetBranchTarget();
                     labels.insert(target);
-                    ExitMethod no_jmp = Scan(offset + 1, end, labels);
-                    ExitMethod jmp = Scan(target, end, labels);
+                    const ExitMethod no_jmp = Scan(offset + 1, end, labels);
+                    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.
                     ASSERT_MSG(instr.bra.constant_buffer == 0,
                                "Constant buffer SSY is not supported");
-                    u32 target = offset + instr.bra.GetBranchTarget();
+                    const u32 target = offset + instr.bra.GetBranchTarget();
                     labels.insert(target);
                     // Continue scanning for an exit method.
                     break;
@@ -346,8 +346,8 @@ public:
      */
     void SetRegisterToInputAttibute(const Register& reg, u64 elem, Attribute::Index attribute,
                                     const Tegra::Shader::IpaMode& input_mode) {
-        std::string dest = GetRegisterAsFloat(reg);
-        std::string src = GetInputAttribute(attribute, input_mode) + GetSwizzle(elem);
+        const std::string dest = GetRegisterAsFloat(reg);
+        const std::string src = GetInputAttribute(attribute, input_mode) + GetSwizzle(elem);
         shader.AddLine(dest + " = " + src + ';');
     }
 
@@ -359,8 +359,8 @@ public:
      * @param reg The register to use as the source value.
      */
     void SetOutputAttributeToRegister(Attribute::Index attribute, u64 elem, const Register& reg) {
-        std::string dest = GetOutputAttribute(attribute);
-        std::string src = GetRegisterAsFloat(reg);
+        const std::string dest = GetOutputAttribute(attribute);
+        const std::string src = GetRegisterAsFloat(reg);
 
         if (!dest.empty()) {
             // Can happen with unknown/unimplemented output attributes, in which case we ignore the
@@ -393,9 +393,9 @@ public:
                                    GLSLRegister::Type type) {
         declr_const_buffers[cbuf_index].MarkAsUsedIndirect(cbuf_index, stage);
 
-        std::string final_offset = fmt::format("({} + {})", index_str, offset / 4);
-        std::string value = 'c' + std::to_string(cbuf_index) + '[' + final_offset + " / 4][" +
-                            final_offset + " % 4]";
+        const std::string final_offset = fmt::format("({} + {})", index_str, offset / 4);
+        const std::string value = 'c' + std::to_string(cbuf_index) + '[' + final_offset + " / 4][" +
+                                  final_offset + " % 4]";
 
         if (type == GLSLRegister::Type::Float) {
             return value;
@@ -443,13 +443,12 @@ public:
         }
         declarations.AddNewLine();
 
-        // Append the sampler2D array for the used textures.
-        const size_t num_samplers = used_samplers.size();
-        if (num_samplers > 0) {
-            declarations.AddLine("uniform sampler2D " + SamplerEntry::GetArrayName(stage) + '[' +
-                                 std::to_string(num_samplers) + "];");
-            declarations.AddNewLine();
+        const auto& samplers = GetSamplers();
+        for (const auto& sampler : samplers) {
+            declarations.AddLine("uniform " + sampler.GetTypeString() + ' ' + sampler.GetName() +
+                                 ';');
         }
+        declarations.AddNewLine();
     }
 
     /// Returns a list of constant buffer declarations
@@ -461,27 +460,29 @@ public:
     }
 
     /// Returns a list of samplers used in the shader
-    std::vector<SamplerEntry> GetSamplers() const {
+    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) {
-        size_t offset = static_cast<size_t>(sampler.index.Value());
+    std::string AccessSampler(const Sampler& sampler, Tegra::Shader::TextureType type,
+                              bool is_array) {
+        const size_t offset = static_cast<size_t>(sampler.index.Value());
 
         // If this sampler has already been used, return the existing mapping.
-        auto itr =
+        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);
             return itr->GetName();
         }
 
         // Otherwise create a new mapping for this sampler
-        size_t next_index = used_samplers.size();
-        SamplerEntry entry{stage, offset, next_index};
+        const 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();
     }
@@ -698,7 +699,7 @@ private:
         };
 
         bool IsColorComponentOutputEnabled(u32 render_target, u32 component) const {
-            u32 bit = render_target * 4 + component;
+            const u32 bit = render_target * 4 + component;
             return enabled_color_outputs & (1 << bit);
         }
     };
@@ -706,7 +707,7 @@ private:
 
     /// Gets the Subroutine object corresponding to the specified address.
     const Subroutine& GetSubroutine(u32 begin, u32 end) const {
-        auto iter = subroutines.find(Subroutine{begin, end, suffix});
+        const auto iter = subroutines.find(Subroutine{begin, end, suffix});
         ASSERT(iter != subroutines.end());
         return *iter;
     }
@@ -722,8 +723,8 @@ private:
     }
 
     /// Generates code representing a texture sampler.
-    std::string GetSampler(const Sampler& sampler) {
-        return regs.AccessSampler(sampler);
+    std::string GetSampler(const Sampler& sampler, Tegra::Shader::TextureType type, bool is_array) {
+        return regs.AccessSampler(sampler, type, is_array);
     }
 
     /**
@@ -751,7 +752,7 @@ private:
         // Can't assign to the constant predicate.
         ASSERT(pred != static_cast<u64>(Pred::UnusedIndex));
 
-        std::string variable = 'p' + std::to_string(pred) + '_' + suffix;
+        const std::string variable = 'p' + std::to_string(pred) + '_' + suffix;
         shader.AddLine(variable + " = " + value + ';');
         declr_predicates.insert(std::move(variable));
     }
@@ -1022,7 +1023,7 @@ private:
             // TODO(Subv): Figure out how dual-source blending is configured in the Switch.
             for (u32 component = 0; component < 4; ++component) {
                 if (header.IsColorComponentOutputEnabled(render_target, component)) {
-                    shader.AddLine(fmt::format("color[{}][{}] = {};", render_target, component,
+                    shader.AddLine(fmt::format("FragColor{}[{}] = {};", render_target, component,
                                                regs.GetRegisterAsFloat(current_reg)));
                     ++current_reg;
                 }
@@ -1032,7 +1033,11 @@ private:
         if (header.writes_depth) {
             // The depth output is always 2 registers after the last color output, and current_reg
             // already contains one past the last color register.
-            shader.AddLine("gl_FragDepth = " + regs.GetRegisterAsFloat(current_reg + 1) + ';');
+
+            shader.AddLine(
+                "gl_FragDepth = " +
+                regs.GetRegisterAsFloat(static_cast<Tegra::Shader::Register>(current_reg) + 1) +
+                ';');
         }
     }
 
@@ -1434,7 +1439,7 @@ private:
                 if (instr.alu_integer.negate_b)
                     op_b = "-(" + op_b + ')';
 
-                std::string shift = std::to_string(instr.alu_integer.shift_amount.Value());
+                const std::string shift = std::to_string(instr.alu_integer.shift_amount.Value());
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0,
                                           "((" + op_a + " << " + shift + ") + " + op_b + ')', 1, 1);
@@ -1452,7 +1457,7 @@ private:
             case OpCode::Id::SEL_C:
             case OpCode::Id::SEL_R:
             case OpCode::Id::SEL_IMM: {
-                std::string condition =
+                const std::string condition =
                     GetPredicateCondition(instr.sel.pred, instr.sel.neg_pred != 0);
                 regs.SetRegisterToInteger(instr.gpr0, true, 0,
                                           '(' + condition + ") ? " + op_a + " : " + op_b, 1, 1);
@@ -1474,8 +1479,9 @@ private:
             case OpCode::Id::LOP3_C:
             case OpCode::Id::LOP3_R:
             case OpCode::Id::LOP3_IMM: {
-                std::string op_c = regs.GetRegisterAsInteger(instr.gpr39);
+                const std::string op_c = regs.GetRegisterAsInteger(instr.gpr39);
                 std::string lut;
+
                 if (opcode->GetId() == OpCode::Id::LOP3_R) {
                     lut = '(' + std::to_string(instr.alu.lop3.GetImmLut28()) + ')';
                 } else {
@@ -1490,15 +1496,82 @@ private:
             case OpCode::Id::IMNMX_IMM: {
                 ASSERT_MSG(instr.imnmx.exchange == Tegra::Shader::IMinMaxExchange::None,
                            "Unimplemented");
-                std::string condition =
+                const std::string condition =
                     GetPredicateCondition(instr.imnmx.pred, instr.imnmx.negate_pred != 0);
-                std::string parameters = op_a + ',' + op_b;
+                const std::string parameters = op_a + ',' + op_b;
                 regs.SetRegisterToInteger(instr.gpr0, instr.imnmx.is_signed, 0,
                                           '(' + condition + ") ? min(" + parameters + ") : max(" +
                                               parameters + ')',
                                           1, 1);
                 break;
             }
+            case OpCode::Id::LEA_R2:
+            case OpCode::Id::LEA_R1:
+            case OpCode::Id::LEA_IMM:
+            case OpCode::Id::LEA_RZ:
+            case OpCode::Id::LEA_HI: {
+                std::string op_a;
+                std::string op_b;
+                std::string op_c;
+
+                switch (opcode->GetId()) {
+                case OpCode::Id::LEA_R2: {
+                    op_a = regs.GetRegisterAsInteger(instr.gpr20);
+                    op_b = regs.GetRegisterAsInteger(instr.gpr39);
+                    op_c = std::to_string(instr.lea.r2.entry_a);
+                    break;
+                }
+
+                case OpCode::Id::LEA_R1: {
+                    const bool neg = instr.lea.r1.neg != 0;
+                    op_a = regs.GetRegisterAsInteger(instr.gpr8);
+                    if (neg)
+                        op_a = "-(" + op_a + ')';
+                    op_b = regs.GetRegisterAsInteger(instr.gpr20);
+                    op_c = std::to_string(instr.lea.r1.entry_a);
+                    break;
+                }
+
+                case OpCode::Id::LEA_IMM: {
+                    const bool neg = instr.lea.imm.neg != 0;
+                    op_b = regs.GetRegisterAsInteger(instr.gpr8);
+                    if (neg)
+                        op_b = "-(" + op_b + ')';
+                    op_a = std::to_string(instr.lea.imm.entry_a);
+                    op_c = std::to_string(instr.lea.imm.entry_b);
+                    break;
+                }
+
+                case OpCode::Id::LEA_RZ: {
+                    const bool neg = instr.lea.rz.neg != 0;
+                    op_b = regs.GetRegisterAsInteger(instr.gpr8);
+                    if (neg)
+                        op_b = "-(" + op_b + ')';
+                    op_a = regs.GetUniform(instr.lea.rz.cb_index, instr.lea.rz.cb_offset,
+                                           GLSLRegister::Type::Integer);
+                    op_c = std::to_string(instr.lea.rz.entry_a);
+
+                    break;
+                }
+
+                case OpCode::Id::LEA_HI:
+                default: {
+                    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());
+                    UNREACHABLE();
+                }
+                }
+                if (instr.lea.pred48 != static_cast<u64>(Pred::UnusedIndex)) {
+                    LOG_ERROR(HW_GPU, "Unhandled LEA Predicate");
+                    UNREACHABLE();
+                }
+                const std::string value = '(' + op_a + " + (" + op_b + "*(1 << " + op_c + ")))";
+                regs.SetRegisterToInteger(instr.gpr0, true, 0, value, 1, 1);
+
+                break;
+            }
             default: {
                 LOG_CRITICAL(HW_GPU, "Unhandled ArithmeticInteger instruction: {}",
                              opcode->GetName());
@@ -1509,7 +1582,7 @@ private:
             break;
         }
         case OpCode::Type::Ffma: {
-            std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
+            const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
             std::string op_b = instr.ffma.negate_b ? "-" : "";
             std::string op_c = instr.ffma.negate_c ? "-" : "";
 
@@ -1719,7 +1792,7 @@ private:
                 shader.AddLine("uint index = (" + regs.GetRegisterAsInteger(instr.gpr8, 0, false) +
                                " / 4) & (MAX_CONSTBUFFER_ELEMENTS - 1);");
 
-                std::string op_a =
+                const std::string op_a =
                     regs.GetUniformIndirect(instr.cbuf36.index, instr.cbuf36.offset + 0, "index",
                                             GLSLRegister::Type::Float);
 
@@ -1729,7 +1802,7 @@ private:
                     break;
 
                 case Tegra::Shader::UniformType::Double: {
-                    std::string op_b =
+                    const std::string op_b =
                         regs.GetUniformIndirect(instr.cbuf36.index, instr.cbuf36.offset + 4,
                                                 "index", GLSLRegister::Type::Float);
                     regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
@@ -1753,17 +1826,74 @@ private:
                 break;
             }
             case OpCode::Id::TEX: {
-                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
-                const std::string sampler = GetSampler(instr.sampler);
-                const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                ASSERT_MSG(instr.tex.array == 0, "TEX arrays unimplemented");
+                Tegra::Shader::TextureType texture_type{instr.tex.texture_type};
+                std::string coord;
+
+                switch (texture_type) {
+                case Tegra::Shader::TextureType::Texture1D: {
+                    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 + ");";
+                    break;
+                }
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+                                 static_cast<u32>(texture_type));
+                    UNREACHABLE();
+
+                    // Fallback to interpreting as a 2D texture for now
+                    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;
+                }
+                // 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);
+
+                const std::string sampler = GetSampler(instr.sampler, texture_type, false);
                 // 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("{");
                 ++shader.scope;
                 shader.AddLine(coord);
-                const std::string texture = "texture(" + sampler + ", coords)";
+                std::string texture;
 
+                switch (instr.tex.process_mode) {
+                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::LB:
+                case Tegra::Shader::TextureProcessMode::LBA: {
+                    // 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: {
+                    // TODO: Figure if A suffix changes the equation at all.
+                    texture = "textureLod(" + sampler + ", coords, " + op_c + ')';
+                    break;
+                }
+                default: {
+                    texture = "texture(" + sampler + ", coords)";
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture process mode {}",
+                                 static_cast<u32>(instr.tex.process_mode.Value()));
+                    UNREACHABLE();
+                }
+                }
                 size_t dest_elem{};
                 for (size_t elem = 0; elem < 4; ++elem) {
                     if (!instr.tex.IsComponentEnabled(elem)) {
@@ -1778,20 +1908,65 @@ private:
                 break;
             }
             case OpCode::Id::TEXS: {
-                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20);
-                const std::string sampler = GetSampler(instr.sampler);
-                const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                std::string coord;
+                Tegra::Shader::TextureType texture_type{instr.texs.GetTextureType()};
+                bool is_array{instr.texs.IsArrayTexture()};
+
+                switch (texture_type) {
+                case Tegra::Shader::TextureType::Texture2D: {
+                    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);
+                        coord = "vec3 coords = vec3(" + x + ", " + y + ", " + index + ");";
+                    } else {
+                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                        const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+                        coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                    }
+                    break;
+                }
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+                                 static_cast<u32>(texture_type));
+                    UNREACHABLE();
 
+                    // Fallback to interpreting as a 2D texture for now
+                    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;
+                }
+                const std::string sampler = GetSampler(instr.sampler, texture_type, is_array);
                 const std::string texture = "texture(" + sampler + ", coords)";
                 WriteTexsInstruction(instr, coord, texture);
                 break;
             }
             case OpCode::Id::TLDS: {
-                const std::string op_a = regs.GetRegisterAsInteger(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsInteger(instr.gpr20);
-                const std::string sampler = GetSampler(instr.sampler);
-                const std::string coord = "ivec2 coords = ivec2(" + op_a + ", " + op_b + ");";
+                ASSERT(instr.tlds.GetTextureType() == Tegra::Shader::TextureType::Texture2D);
+                ASSERT(instr.tlds.IsArrayTexture() == false);
+                std::string coord;
+
+                switch (instr.tlds.GetTextureType()) {
+                case Tegra::Shader::TextureType::Texture2D: {
+                    if (instr.tlds.IsArrayTexture()) {
+                        LOG_CRITICAL(HW_GPU, "Unhandled 2d array texture");
+                        UNREACHABLE();
+                    } else {
+                        const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
+                        const std::string y = regs.GetRegisterAsInteger(instr.gpr20);
+                        coord = "ivec2 coords = ivec2(" + x + ", " + y + ");";
+                    }
+                    break;
+                }
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+                                 static_cast<u32>(instr.tlds.GetTextureType()));
+                    UNREACHABLE();
+                }
+                const std::string sampler = GetSampler(instr.sampler, instr.tlds.GetTextureType(),
+                                                       instr.tlds.IsArrayTexture());
                 const std::string texture = "texelFetch(" + sampler + ", coords, 0)";
                 WriteTexsInstruction(instr, coord, texture);
                 break;
@@ -1799,12 +1974,12 @@ private:
             case OpCode::Id::TLD4: {
                 ASSERT(instr.tld4.texture_type == Tegra::Shader::TextureType::Texture2D);
                 ASSERT(instr.tld4.array == 0);
-                std::string coord{};
+                std::string coord;
 
                 switch (instr.tld4.texture_type) {
                 case Tegra::Shader::TextureType::Texture2D: {
-                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    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;
                 }
@@ -1814,7 +1989,8 @@ private:
                     UNREACHABLE();
                 }
 
-                const std::string sampler = GetSampler(instr.sampler);
+                const std::string sampler =
+                    GetSampler(instr.sampler, instr.tld4.texture_type, false);
                 // 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("{");
@@ -1840,13 +2016,82 @@ private:
                 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);
+                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 texture = "textureGather(" + sampler + ", coords, " +
                                             std::to_string(instr.tld4s.component) + ')';
                 WriteTexsInstruction(instr, coord, texture);
                 break;
             }
+            case OpCode::Id::TXQ: {
+                // TODO: the new commits on the texture refactor, change the way samplers work.
+                // 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);
+                switch (instr.txq.query_type) {
+                case Tegra::Shader::TextureQueryType::Dimension: {
+                    const std::string texture = "textureQueryLevels(" + sampler + ')';
+                    regs.SetRegisterToInteger(instr.gpr0, true, 0, texture, 1, 1);
+                    break;
+                }
+                default: {
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture query type: {}",
+                                 static_cast<u32>(instr.txq.query_type.Value()));
+                    UNREACHABLE();
+                }
+                }
+                break;
+            }
+            case OpCode::Id::TMML: {
+                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
+                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                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);
+
+                // 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);
+                    coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                    break;
+                }
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+                                 static_cast<u32>(texture_type));
+                    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);
+                    coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                    texture_type = Tegra::Shader::TextureType::Texture2D;
+                }
+                // 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('{');
+                ++shader.scope;
+                shader.AddLine(coord);
+                const std::string texture = "textureQueryLod(" + sampler + ", coords)";
+                const std::string tmp = "vec2 tmp = " + texture + "*vec2(256.0, 256.0);";
+                shader.AddLine(tmp);
+
+                regs.SetRegisterToInteger(instr.gpr0, true, 0, "int(tmp.y)", 1, 1);
+                regs.SetRegisterToInteger(instr.gpr0.Value() + 1, false, 0, "uint(tmp.x)", 1, 1);
+                --shader.scope;
+                shader.AddLine('}');
+                break;
+            }
             default: {
                 LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->GetName());
                 UNREACHABLE();
@@ -1886,12 +2131,12 @@ private:
             // We can't use the constant predicate as destination.
             ASSERT(instr.fsetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
 
-            std::string second_pred =
+            const std::string second_pred =
                 GetPredicateCondition(instr.fsetp.pred39, instr.fsetp.neg_pred != 0);
 
-            std::string combiner = GetPredicateCombiner(instr.fsetp.op);
+            const std::string combiner = GetPredicateCombiner(instr.fsetp.op);
 
-            std::string predicate = GetPredicateComparison(instr.fsetp.cond, op_a, op_b);
+            const std::string predicate = GetPredicateComparison(instr.fsetp.cond, op_a, op_b);
             // Set the primary predicate to the result of Predicate OP SecondPredicate
             SetPredicate(instr.fsetp.pred3,
                          '(' + predicate + ") " + combiner + " (" + second_pred + ')');
@@ -1905,7 +2150,8 @@ private:
             break;
         }
         case OpCode::Type::IntegerSetPredicate: {
-            std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, instr.isetp.is_signed);
+            const std::string op_a =
+                regs.GetRegisterAsInteger(instr.gpr8, 0, instr.isetp.is_signed);
             std::string op_b;
 
             if (instr.is_b_imm) {
@@ -1922,12 +2168,12 @@ private:
             // We can't use the constant predicate as destination.
             ASSERT(instr.isetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
 
-            std::string second_pred =
+            const std::string second_pred =
                 GetPredicateCondition(instr.isetp.pred39, instr.isetp.neg_pred != 0);
 
-            std::string combiner = GetPredicateCombiner(instr.isetp.op);
+            const std::string combiner = GetPredicateCombiner(instr.isetp.op);
 
-            std::string predicate = GetPredicateComparison(instr.isetp.cond, op_a, op_b);
+            const std::string predicate = GetPredicateComparison(instr.isetp.cond, op_a, op_b);
             // Set the primary predicate to the result of Predicate OP SecondPredicate
             SetPredicate(instr.isetp.pred3,
                          '(' + predicate + ") " + combiner + " (" + second_pred + ')');
@@ -1940,21 +2186,45 @@ private:
             }
             break;
         }
+        case OpCode::Type::PredicateSetRegister: {
+            const std::string op_a =
+                GetPredicateCondition(instr.pset.pred12, instr.pset.neg_pred12 != 0);
+            const std::string op_b =
+                GetPredicateCondition(instr.pset.pred29, instr.pset.neg_pred29 != 0);
+
+            const std::string second_pred =
+                GetPredicateCondition(instr.pset.pred39, instr.pset.neg_pred39 != 0);
+
+            const std::string combiner = GetPredicateCombiner(instr.pset.op);
+
+            const std::string predicate =
+                '(' + op_a + ") " + GetPredicateCombiner(instr.pset.cond) + " (" + op_b + ')';
+            const std::string result = '(' + predicate + ") " + combiner + " (" + second_pred + ')';
+            if (instr.pset.bf == 0) {
+                const std::string value = '(' + result + ") ? 0xFFFFFFFF : 0";
+                regs.SetRegisterToInteger(instr.gpr0, false, 0, value, 1, 1);
+            } else {
+                const std::string value = '(' + result + ") ? 1.0 : 0.0";
+                regs.SetRegisterToFloat(instr.gpr0, 0, value, 1, 1);
+            }
+
+            break;
+        }
         case OpCode::Type::PredicateSetPredicate: {
-            std::string op_a =
+            const std::string op_a =
                 GetPredicateCondition(instr.psetp.pred12, instr.psetp.neg_pred12 != 0);
-            std::string op_b =
+            const std::string op_b =
                 GetPredicateCondition(instr.psetp.pred29, instr.psetp.neg_pred29 != 0);
 
             // We can't use the constant predicate as destination.
             ASSERT(instr.psetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
 
-            std::string second_pred =
+            const std::string second_pred =
                 GetPredicateCondition(instr.psetp.pred39, instr.psetp.neg_pred39 != 0);
 
-            std::string combiner = GetPredicateCombiner(instr.psetp.op);
+            const std::string combiner = GetPredicateCombiner(instr.psetp.op);
 
-            std::string predicate =
+            const std::string predicate =
                 '(' + op_a + ") " + GetPredicateCombiner(instr.psetp.cond) + " (" + op_b + ')';
 
             // Set the primary predicate to the result of Predicate OP SecondPredicate
@@ -1980,7 +2250,7 @@ private:
             std::string op_b = instr.fset.neg_b ? "-" : "";
 
             if (instr.is_b_imm) {
-                std::string imm = GetImmediate19(instr);
+                const std::string imm = GetImmediate19(instr);
                 if (instr.fset.neg_imm)
                     op_b += "(-" + imm + ')';
                 else
@@ -2000,13 +2270,14 @@ private:
 
             // 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.
-            std::string second_pred =
+            const std::string second_pred =
                 GetPredicateCondition(instr.fset.pred39, instr.fset.neg_pred != 0);
 
-            std::string combiner = GetPredicateCombiner(instr.fset.op);
+            const std::string combiner = GetPredicateCombiner(instr.fset.op);
 
-            std::string predicate = "((" + GetPredicateComparison(instr.fset.cond, op_a, op_b) +
-                                    ") " + combiner + " (" + second_pred + "))";
+            const std::string predicate = "((" +
+                                          GetPredicateComparison(instr.fset.cond, op_a, op_b) +
+                                          ") " + combiner + " (" + second_pred + "))";
 
             if (instr.fset.bf) {
                 regs.SetRegisterToFloat(instr.gpr0, 0, predicate + " ? 1.0 : 0.0", 1, 1);
@@ -2017,7 +2288,7 @@ private:
             break;
         }
         case OpCode::Type::IntegerSet: {
-            std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, instr.iset.is_signed);
+            const std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, instr.iset.is_signed);
 
             std::string op_b;
 
@@ -2034,13 +2305,14 @@ private:
 
             // The iset instruction sets a register to 1.0 or -1 (depending on the bf bit) if the
             // condition is true, and to 0 otherwise.
-            std::string second_pred =
+            const std::string second_pred =
                 GetPredicateCondition(instr.iset.pred39, instr.iset.neg_pred != 0);
 
-            std::string combiner = GetPredicateCombiner(instr.iset.op);
+            const std::string combiner = GetPredicateCombiner(instr.iset.op);
 
-            std::string predicate = "((" + GetPredicateComparison(instr.iset.cond, op_a, op_b) +
-                                    ") " + combiner + " (" + second_pred + "))";
+            const std::string predicate = "((" +
+                                          GetPredicateComparison(instr.iset.cond, op_a, op_b) +
+                                          ") " + combiner + " (" + second_pred + "))";
 
             if (instr.iset.bf) {
                 regs.SetRegisterToFloat(instr.gpr0, 0, predicate + " ? 1.0 : 0.0", 1, 1);
@@ -2190,7 +2462,7 @@ private:
             case OpCode::Id::BRA: {
                 ASSERT_MSG(instr.bra.constant_buffer == 0,
                            "BRA with constant buffers are not implemented");
-                u32 target = offset + instr.bra.GetBranchTarget();
+                const u32 target = offset + instr.bra.GetBranchTarget();
                 shader.AddLine("{ jmp_to = " + std::to_string(target) + "u; break; }");
                 break;
             }
@@ -2214,7 +2486,7 @@ private:
                 // has a similar structure to the BRA opcode.
                 ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer SSY is not supported");
 
-                u32 target = offset + instr.bra.GetBranchTarget();
+                const u32 target = offset + instr.bra.GetBranchTarget();
                 EmitPushToSSYStack(target);
                 break;
             }
@@ -2308,10 +2580,10 @@ private:
                     shader.AddLine("case " + std::to_string(label) + "u: {");
                     ++shader.scope;
 
-                    auto next_it = labels.lower_bound(label + 1);
-                    u32 next_label = next_it == labels.end() ? subroutine.end : *next_it;
+                    const auto next_it = labels.lower_bound(label + 1);
+                    const u32 next_label = next_it == labels.end() ? subroutine.end : *next_it;
 
-                    u32 compile_end = CompileRange(label, next_label);
+                    const u32 compile_end = CompileRange(label, next_label);
                     if (compile_end > next_label && compile_end != PROGRAM_END) {
                         // This happens only when there is a label inside a IF/LOOP block
                         shader.AddLine(" jmp_to = " + std::to_string(compile_end) + "u; break; }");
@@ -2374,7 +2646,8 @@ boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code,
                                                 Maxwell3D::Regs::ShaderStage stage,
                                                 const std::string& suffix) {
     try {
-        auto subroutines = ControlFlowAnalyzer(program_code, main_offset, suffix).GetSubroutines();
+        const auto subroutines =
+            ControlFlowAnalyzer(program_code, main_offset, suffix).GetSubroutines();
         GLSLGenerator generator(subroutines, program_code, main_offset, stage, suffix);
         return ProgramResult{generator.GetShaderCode(), generator.GetEntries()};
     } catch (const DecompileFail& exception) {
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp
index e1b1a9d73..b0466c18f 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp
@@ -88,7 +88,14 @@ ProgramResult GenerateFragmentShader(const ShaderSetup& setup) {
             .get_value_or({});
     out += R"(
 in vec4 position;
-layout(location = 0) out vec4 color[8];
+layout(location = 0) out vec4 FragColor0;
+layout(location = 1) out vec4 FragColor1;
+layout(location = 2) out vec4 FragColor2;
+layout(location = 3) out vec4 FragColor3;
+layout(location = 4) out vec4 FragColor4;
+layout(location = 5) out vec4 FragColor5;
+layout(location = 6) out vec4 FragColor6;
+layout(location = 7) out vec4 FragColor7;
 
 layout (std140) uniform fs_config {
     vec4 viewport_flip;
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.h b/src/video_core/renderer_opengl/gl_shader_gen.h
index cbb2090ea..a43e2997b 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.h
+++ b/src/video_core/renderer_opengl/gl_shader_gen.h
@@ -9,6 +9,7 @@
 #include <vector>
 
 #include "common/common_types.h"
+#include "video_core/engines/shader_bytecode.h"
 
 namespace OpenGL::GLShader {
 
@@ -73,8 +74,9 @@ class SamplerEntry {
     using Maxwell = Tegra::Engines::Maxwell3D::Regs;
 
 public:
-    SamplerEntry(Maxwell::ShaderStage stage, size_t offset, size_t index)
-        : offset(offset), stage(stage), sampler_index(index) {}
+    SamplerEntry(Maxwell::ShaderStage stage, size_t offset, size_t index,
+                 Tegra::Shader::TextureType type, bool is_array)
+        : offset(offset), stage(stage), sampler_index(index), type(type), is_array(is_array) {}
 
     size_t GetOffset() const {
         return offset;
@@ -89,8 +91,41 @@ public:
     }
 
     std::string GetName() const {
-        return std::string(TextureSamplerNames[static_cast<size_t>(stage)]) + '[' +
-               std::to_string(sampler_index) + ']';
+        return std::string(TextureSamplerNames[static_cast<size_t>(stage)]) + '_' +
+               std::to_string(sampler_index);
+    }
+
+    std::string GetTypeString() const {
+        using Tegra::Shader::TextureType;
+        std::string glsl_type;
+
+        switch (type) {
+        case TextureType::Texture1D:
+            glsl_type = "sampler1D";
+            break;
+        case TextureType::Texture2D:
+            glsl_type = "sampler2D";
+            break;
+        case TextureType::Texture3D:
+            glsl_type = "sampler3D";
+            break;
+        case TextureType::TextureCube:
+            glsl_type = "samplerCube";
+            break;
+        default:
+            UNIMPLEMENTED();
+        }
+        if (is_array)
+            glsl_type += "Array";
+        return glsl_type;
+    }
+
+    Tegra::Shader::TextureType GetType() const {
+        return type;
+    }
+
+    bool IsArray() const {
+        return is_array;
     }
 
     u32 GetHash() const {
@@ -105,11 +140,14 @@ private:
     static constexpr std::array<const char*, Maxwell::MaxShaderStage> TextureSamplerNames = {
         "tex_vs", "tex_tessc", "tex_tesse", "tex_gs", "tex_fs",
     };
+
     /// Offset in TSC memory from which to read the sampler object, as specified by the sampling
     /// instruction.
     size_t offset;
-    Maxwell::ShaderStage stage; ///< Shader stage where this sampler was used.
-    size_t sampler_index;       ///< Value used to index into the generated GLSL sampler array.
+    Maxwell::ShaderStage stage;      ///< Shader stage where this sampler was used.
+    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.
 };
 
 struct ShaderEntries {
diff --git a/src/video_core/renderer_opengl/gl_shader_util.cpp b/src/video_core/renderer_opengl/gl_shader_util.cpp
index 5781d9d16..5f3fe067e 100644
--- a/src/video_core/renderer_opengl/gl_shader_util.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_util.cpp
@@ -25,7 +25,7 @@ GLuint LoadShader(const char* source, GLenum type) {
     default:
         UNREACHABLE();
     }
-    GLuint shader_id = glCreateShader(type);
+    const GLuint shader_id = glCreateShader(type);
     glShaderSource(shader_id, 1, &source, nullptr);
     LOG_DEBUG(Render_OpenGL, "Compiling {} shader...", debug_type);
     glCompileShader(shader_id);
diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp
index 60a4defd1..6f70deb96 100644
--- a/src/video_core/renderer_opengl/gl_state.cpp
+++ b/src/video_core/renderer_opengl/gl_state.cpp
@@ -200,9 +200,9 @@ void OpenGLState::Apply() const {
         const auto& texture_unit = texture_units[i];
         const auto& cur_state_texture_unit = cur_state.texture_units[i];
 
-        if (texture_unit.texture_2d != cur_state_texture_unit.texture_2d) {
+        if (texture_unit.texture != cur_state_texture_unit.texture) {
             glActiveTexture(TextureUnits::MaxwellTexture(static_cast<int>(i)).Enum());
-            glBindTexture(GL_TEXTURE_2D, texture_unit.texture_2d);
+            glBindTexture(texture_unit.target, texture_unit.texture);
         }
         if (texture_unit.sampler != cur_state_texture_unit.sampler) {
             glBindSampler(static_cast<GLuint>(i), texture_unit.sampler);
@@ -214,7 +214,7 @@ void OpenGLState::Apply() const {
             texture_unit.swizzle.a != cur_state_texture_unit.swizzle.a) {
             std::array<GLint, 4> mask = {texture_unit.swizzle.r, texture_unit.swizzle.g,
                                          texture_unit.swizzle.b, texture_unit.swizzle.a};
-            glTexParameteriv(GL_TEXTURE_2D, GL_TEXTURE_SWIZZLE_RGBA, mask.data());
+            glTexParameteriv(texture_unit.target, GL_TEXTURE_SWIZZLE_RGBA, mask.data());
         }
     }
 
@@ -287,7 +287,7 @@ void OpenGLState::Apply() const {
 
 OpenGLState& OpenGLState::UnbindTexture(GLuint handle) {
     for (auto& unit : texture_units) {
-        if (unit.texture_2d == handle) {
+        if (unit.texture == handle) {
             unit.Unbind();
         }
     }
diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h
index 46e96a97d..e3e24b9e7 100644
--- a/src/video_core/renderer_opengl/gl_state.h
+++ b/src/video_core/renderer_opengl/gl_state.h
@@ -94,8 +94,9 @@ public:
 
     // 3 texture units - one for each that is used in PICA fragment shader emulation
     struct TextureUnit {
-        GLuint texture_2d; // GL_TEXTURE_BINDING_2D
-        GLuint sampler;    // GL_SAMPLER_BINDING
+        GLuint texture; // GL_TEXTURE_BINDING_2D
+        GLuint sampler; // GL_SAMPLER_BINDING
+        GLenum target;
         struct {
             GLint r; // GL_TEXTURE_SWIZZLE_R
             GLint g; // GL_TEXTURE_SWIZZLE_G
@@ -104,7 +105,7 @@ public:
         } swizzle;
 
         void Unbind() {
-            texture_2d = 0;
+            texture = 0;
             swizzle.r = GL_RED;
             swizzle.g = GL_GREEN;
             swizzle.b = GL_BLUE;
@@ -114,6 +115,7 @@ public:
         void Reset() {
             Unbind();
             sampler = 0;
+            target = GL_TEXTURE_2D;
         }
     };
     std::array<TextureUnit, 32> texture_units;
diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.cpp b/src/video_core/renderer_opengl/gl_stream_buffer.cpp
index e565afcee..aadf68f16 100644
--- a/src/video_core/renderer_opengl/gl_stream_buffer.cpp
+++ b/src/video_core/renderer_opengl/gl_stream_buffer.cpp
@@ -29,7 +29,7 @@ OGLStreamBuffer::OGLStreamBuffer(GLenum target, GLsizeiptr size, bool prefer_coh
     if (GLAD_GL_ARB_buffer_storage) {
         persistent = true;
         coherent = prefer_coherent;
-        GLbitfield flags =
+        const GLbitfield flags =
             GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT | (coherent ? GL_MAP_COHERENT_BIT : 0);
         glBufferStorage(gl_target, allocate_size, nullptr, flags);
         mapped_ptr = static_cast<u8*>(glMapBufferRange(
diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp
index 411a73d50..96d916b07 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.cpp
+++ b/src/video_core/renderer_opengl/renderer_opengl.cpp
@@ -177,7 +177,7 @@ void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuf
                                        Memory::GetPointer(framebuffer_addr),
                                        gl_framebuffer_data.data(), true);
 
-        state.texture_units[0].texture_2d = screen_info.texture.resource.handle;
+        state.texture_units[0].texture = screen_info.texture.resource.handle;
         state.Apply();
 
         glActiveTexture(GL_TEXTURE0);
@@ -194,7 +194,7 @@ void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuf
 
         glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
 
-        state.texture_units[0].texture_2d = 0;
+        state.texture_units[0].texture = 0;
         state.Apply();
     }
 }
@@ -205,7 +205,7 @@ void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuf
  */
 void RendererOpenGL::LoadColorToActiveGLTexture(u8 color_r, u8 color_g, u8 color_b, u8 color_a,
                                                 const TextureInfo& texture) {
-    state.texture_units[0].texture_2d = texture.resource.handle;
+    state.texture_units[0].texture = texture.resource.handle;
     state.Apply();
 
     glActiveTexture(GL_TEXTURE0);
@@ -214,7 +214,7 @@ void RendererOpenGL::LoadColorToActiveGLTexture(u8 color_r, u8 color_g, u8 color
     // Update existing texture
     glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA, 1, 1, 0, GL_RGBA, GL_UNSIGNED_BYTE, framebuffer_data);
 
-    state.texture_units[0].texture_2d = 0;
+    state.texture_units[0].texture = 0;
     state.Apply();
 }
 
@@ -260,7 +260,7 @@ void RendererOpenGL::InitOpenGLObjects() {
     // Allocation of storage is deferred until the first frame, when we
     // know the framebuffer size.
 
-    state.texture_units[0].texture_2d = screen_info.texture.resource.handle;
+    state.texture_units[0].texture = screen_info.texture.resource.handle;
     state.Apply();
 
     glActiveTexture(GL_TEXTURE0);
@@ -272,7 +272,7 @@ void RendererOpenGL::InitOpenGLObjects() {
 
     screen_info.display_texture = screen_info.texture.resource.handle;
 
-    state.texture_units[0].texture_2d = 0;
+    state.texture_units[0].texture = 0;
     state.Apply();
 
     // Clear screen to black
@@ -305,14 +305,14 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture,
         UNREACHABLE();
     }
 
-    state.texture_units[0].texture_2d = texture.resource.handle;
+    state.texture_units[0].texture = texture.resource.handle;
     state.Apply();
 
     glActiveTexture(GL_TEXTURE0);
     glTexImage2D(GL_TEXTURE_2D, 0, internal_format, texture.width, texture.height, 0,
                  texture.gl_format, texture.gl_type, nullptr);
 
-    state.texture_units[0].texture_2d = 0;
+    state.texture_units[0].texture = 0;
     state.Apply();
 }
 
@@ -354,14 +354,14 @@ void RendererOpenGL::DrawScreenTriangles(const ScreenInfo& screen_info, float x,
         ScreenRectVertex(x + w, y + h, texcoords.bottom * scale_u, right * scale_v),
     }};
 
-    state.texture_units[0].texture_2d = screen_info.display_texture;
+    state.texture_units[0].texture = screen_info.display_texture;
     state.texture_units[0].swizzle = {GL_RED, GL_GREEN, GL_BLUE, GL_ALPHA};
     state.Apply();
 
     glBufferSubData(GL_ARRAY_BUFFER, 0, sizeof(vertices), vertices.data());
     glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
 
-    state.texture_units[0].texture_2d = 0;
+    state.texture_units[0].texture = 0;
     state.Apply();
 }
 
@@ -369,6 +369,12 @@ void RendererOpenGL::DrawScreenTriangles(const ScreenInfo& screen_info, float x,
  * Draws the emulated screens to the emulator window.
  */
 void RendererOpenGL::DrawScreen() {
+    if (renderer_settings.set_background_color) {
+        // Update background color before drawing
+        glClearColor(Settings::values.bg_red, Settings::values.bg_green, Settings::values.bg_blue,
+                     0.0f);
+    }
+
     const auto& layout = render_window.GetFramebufferLayout();
     const auto& screen = layout.screen;
 
diff --git a/src/video_core/textures/texture.h b/src/video_core/textures/texture.h
index c6bd2f4b9..c2fb824b2 100644
--- a/src/video_core/textures/texture.h
+++ b/src/video_core/textures/texture.h
@@ -170,8 +170,12 @@ struct TICEntry {
         BitField<0, 16, u32> width_minus_1;
         BitField<23, 4, TextureType> texture_type;
     };
-    u16 height_minus_1;
-    INSERT_PADDING_BYTES(10);
+    union {
+        BitField<0, 16, u32> height_minus_1;
+        BitField<16, 15, u32> depth_minus_1;
+    };
+
+    INSERT_PADDING_BYTES(8);
 
     GPUVAddr Address() const {
         return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | address_low);
@@ -192,6 +196,10 @@ struct TICEntry {
         return height_minus_1 + 1;
     }
 
+    u32 Depth() const {
+        return depth_minus_1 + 1;
+    }
+
     u32 BlockHeight() const {
         ASSERT(header_version == TICHeaderVersion::BlockLinear ||
                header_version == TICHeaderVersion::BlockLinearColorKey);
diff --git a/src/yuzu/CMakeLists.txt b/src/yuzu/CMakeLists.txt
index a2b6e984e..f48b69809 100644
--- a/src/yuzu/CMakeLists.txt
+++ b/src/yuzu/CMakeLists.txt
@@ -9,6 +9,8 @@ add_executable(yuzu
     about_dialog.h
     bootmanager.cpp
     bootmanager.h
+    compatibility_list.cpp
+    compatibility_list.h
     configuration/config.cpp
     configuration/config.h
     configuration/configure_audio.cpp
diff --git a/src/yuzu/bootmanager.cpp b/src/yuzu/bootmanager.cpp
index 4a60f450a..4e4c108ab 100644
--- a/src/yuzu/bootmanager.cpp
+++ b/src/yuzu/bootmanager.cpp
@@ -112,6 +112,7 @@ GRenderWindow::GRenderWindow(QWidget* parent, EmuThread* emu_thread)
     setWindowTitle(QString::fromStdString(window_title));
 
     InputCommon::Init();
+    InputCommon::StartJoystickEventHandler();
 }
 
 GRenderWindow::~GRenderWindow() {
diff --git a/src/yuzu/compatibility_list.cpp b/src/yuzu/compatibility_list.cpp
new file mode 100644
index 000000000..2d2cfd03c
--- /dev/null
+++ b/src/yuzu/compatibility_list.cpp
@@ -0,0 +1,18 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <algorithm>
+
+#include <fmt/format.h>
+
+#include "yuzu/compatibility_list.h"
+
+CompatibilityList::const_iterator FindMatchingCompatibilityEntry(
+    const CompatibilityList& compatibility_list, u64 program_id) {
+    return std::find_if(compatibility_list.begin(), compatibility_list.end(),
+                        [program_id](const auto& element) {
+                            std::string pid = fmt::format("{:016X}", program_id);
+                            return element.first == pid;
+                        });
+}
diff --git a/src/yuzu/compatibility_list.h b/src/yuzu/compatibility_list.h
new file mode 100644
index 000000000..bc0175bd3
--- /dev/null
+++ b/src/yuzu/compatibility_list.h
@@ -0,0 +1,17 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <string>
+#include <unordered_map>
+
+#include <QString>
+
+#include "common/common_types.h"
+
+using CompatibilityList = std::unordered_map<std::string, std::pair<QString, QString>>;
+
+CompatibilityList::const_iterator FindMatchingCompatibilityEntry(
+    const CompatibilityList& compatibility_list, u64 program_id);
diff --git a/src/yuzu/configuration/configure_general.cpp b/src/yuzu/configuration/configure_general.cpp
index d8caee1e8..9292d9a42 100644
--- a/src/yuzu/configuration/configure_general.cpp
+++ b/src/yuzu/configuration/configure_general.cpp
@@ -20,7 +20,6 @@ ConfigureGeneral::ConfigureGeneral(QWidget* parent)
     this->setConfiguration();
 
     ui->use_cpu_jit->setEnabled(!Core::System::GetInstance().IsPoweredOn());
-    ui->use_multi_core->setEnabled(!Core::System::GetInstance().IsPoweredOn());
     ui->use_docked_mode->setEnabled(!Core::System::GetInstance().IsPoweredOn());
 }
 
@@ -31,7 +30,6 @@ void ConfigureGeneral::setConfiguration() {
     ui->toggle_check_exit->setChecked(UISettings::values.confirm_before_closing);
     ui->theme_combobox->setCurrentIndex(ui->theme_combobox->findData(UISettings::values.theme));
     ui->use_cpu_jit->setChecked(Settings::values.use_cpu_jit);
-    ui->use_multi_core->setChecked(Settings::values.use_multi_core);
     ui->use_docked_mode->setChecked(Settings::values.use_docked_mode);
 }
 
@@ -46,6 +44,5 @@ void ConfigureGeneral::applyConfiguration() {
         ui->theme_combobox->itemData(ui->theme_combobox->currentIndex()).toString();
 
     Settings::values.use_cpu_jit = ui->use_cpu_jit->isChecked();
-    Settings::values.use_multi_core = ui->use_multi_core->isChecked();
     Settings::values.use_docked_mode = ui->use_docked_mode->isChecked();
 }
diff --git a/src/yuzu/configuration/configure_general.ui b/src/yuzu/configuration/configure_general.ui
index 233adbe27..1775c4d40 100644
--- a/src/yuzu/configuration/configure_general.ui
+++ b/src/yuzu/configuration/configure_general.ui
@@ -58,13 +58,6 @@
             </property>
            </widget>
           </item>
-          <item>
-           <widget class="QCheckBox" name="use_multi_core">
-            <property name="text">
-             <string>Enable multi-core</string>
-            </property>
-           </widget>
-          </item>
          </layout>
         </item>
        </layout>
diff --git a/src/yuzu/configuration/configure_graphics.cpp b/src/yuzu/configuration/configure_graphics.cpp
index ee1287028..839d58f59 100644
--- a/src/yuzu/configuration/configure_graphics.cpp
+++ b/src/yuzu/configuration/configure_graphics.cpp
@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <QColorDialog>
 #include "core/core.h"
 #include "core/settings.h"
 #include "ui_configure_graphics.h"
@@ -16,6 +17,14 @@ ConfigureGraphics::ConfigureGraphics(QWidget* parent)
     ui->frame_limit->setEnabled(Settings::values.use_frame_limit);
     connect(ui->toggle_frame_limit, &QCheckBox::stateChanged, ui->frame_limit,
             &QSpinBox::setEnabled);
+    connect(ui->bg_button, &QPushButton::clicked, this, [this] {
+        const QColor new_bg_color = QColorDialog::getColor(bg_color);
+        if (!new_bg_color.isValid())
+            return;
+        bg_color = new_bg_color;
+        ui->bg_button->setStyleSheet(
+            QString("QPushButton { background-color: %1 }").arg(bg_color.name()));
+    });
 }
 
 ConfigureGraphics::~ConfigureGraphics() = default;
@@ -65,6 +74,10 @@ void ConfigureGraphics::setConfiguration() {
     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);
+    bg_color = QColor::fromRgbF(Settings::values.bg_red, Settings::values.bg_green,
+                                Settings::values.bg_blue);
+    ui->bg_button->setStyleSheet(
+        QString("QPushButton { background-color: %1 }").arg(bg_color.name()));
 }
 
 void ConfigureGraphics::applyConfiguration() {
@@ -73,4 +86,7 @@ void ConfigureGraphics::applyConfiguration() {
     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.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.h b/src/yuzu/configuration/configure_graphics.h
index 5497a55f7..9bda26fd6 100644
--- a/src/yuzu/configuration/configure_graphics.h
+++ b/src/yuzu/configuration/configure_graphics.h
@@ -25,4 +25,5 @@ private:
 
 private:
     std::unique_ptr<Ui::ConfigureGraphics> ui;
+    QColor bg_color;
 };
diff --git a/src/yuzu/configuration/configure_graphics.ui b/src/yuzu/configuration/configure_graphics.ui
index 3bc18c26e..8fc00af1b 100644
--- a/src/yuzu/configuration/configure_graphics.ui
+++ b/src/yuzu/configuration/configure_graphics.ui
@@ -96,6 +96,27 @@
           </item>
          </layout>
         </item>
+         <item>
+           <layout class="QHBoxLayout" name="horizontalLayout_6">
+             <item>
+               <widget class="QLabel" name="bg_label">
+                 <property name="text">
+                   <string>Background Color:</string>
+                 </property>
+               </widget>
+             </item>
+             <item>
+               <widget class="QPushButton" name="bg_button">
+                 <property name="maximumSize">
+                   <size>
+                     <width>40</width>
+                     <height>16777215</height>
+                   </size>
+                 </property>
+               </widget>
+             </item>
+           </layout>
+         </item>
        </layout>
       </widget>
      </item>
diff --git a/src/yuzu/configuration/configure_input.cpp b/src/yuzu/configuration/configure_input.cpp
index 5e7badedf..d29abb74b 100644
--- a/src/yuzu/configuration/configure_input.cpp
+++ b/src/yuzu/configuration/configure_input.cpp
@@ -1,4 +1,4 @@
-// Copyright 2016 Citra Emulator Project
+// Copyright 2016 Citra Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
@@ -53,19 +53,18 @@ static QString ButtonToText(const Common::ParamPackage& param) {
     } else if (param.Get("engine", "") == "keyboard") {
         return getKeyName(param.Get("code", 0));
     } else if (param.Get("engine", "") == "sdl") {
-        QString text = QString(QObject::tr("Joystick %1")).arg(param.Get("joystick", "").c_str());
         if (param.Has("hat")) {
-            text += QString(QObject::tr(" Hat %1 %2"))
-                        .arg(param.Get("hat", "").c_str(), param.Get("direction", "").c_str());
+            return QString(QObject::tr("Hat %1 %2"))
+                .arg(param.Get("hat", "").c_str(), param.Get("direction", "").c_str());
         }
         if (param.Has("axis")) {
-            text += QString(QObject::tr(" Axis %1%2"))
-                        .arg(param.Get("axis", "").c_str(), param.Get("direction", "").c_str());
+            return QString(QObject::tr("Axis %1%2"))
+                .arg(param.Get("axis", "").c_str(), param.Get("direction", "").c_str());
         }
         if (param.Has("button")) {
-            text += QString(QObject::tr(" Button %1")).arg(param.Get("button", "").c_str());
+            return QString(QObject::tr("Button %1")).arg(param.Get("button", "").c_str());
         }
-        return text;
+        return QString();
     } else {
         return QObject::tr("[unknown]");
     }
@@ -81,13 +80,12 @@ static QString AnalogToText(const Common::ParamPackage& param, const std::string
             return QString(QObject::tr("[unused]"));
         }
 
-        QString text = QString(QObject::tr("Joystick %1")).arg(param.Get("joystick", "").c_str());
         if (dir == "left" || dir == "right") {
-            text += QString(QObject::tr(" Axis %1")).arg(param.Get("axis_x", "").c_str());
+            return QString(QObject::tr("Axis %1")).arg(param.Get("axis_x", "").c_str());
         } else if (dir == "up" || dir == "down") {
-            text += QString(QObject::tr(" Axis %1")).arg(param.Get("axis_y", "").c_str());
+            return QString(QObject::tr("Axis %1")).arg(param.Get("axis_y", "").c_str());
         }
-        return text;
+        return QString();
     } else {
         return QObject::tr("[unknown]");
     }
diff --git a/src/yuzu/game_list.cpp b/src/yuzu/game_list.cpp
index 86532e4a9..3b3b551bb 100644
--- a/src/yuzu/game_list.cpp
+++ b/src/yuzu/game_list.cpp
@@ -19,6 +19,7 @@
 #include "common/file_util.h"
 #include "common/logging/log.h"
 #include "core/file_sys/patch_manager.h"
+#include "yuzu/compatibility_list.h"
 #include "yuzu/game_list.h"
 #include "yuzu/game_list_p.h"
 #include "yuzu/game_list_worker.h"
@@ -365,7 +366,7 @@ void GameList::LoadCompatibilityList() {
     QJsonDocument json = QJsonDocument::fromJson(string_content.toUtf8());
     QJsonArray arr = json.array();
 
-    for (const QJsonValue& value : arr) {
+    for (const QJsonValueRef& value : arr) {
         QJsonObject game = value.toObject();
 
         if (game.contains("compatibility") && game["compatibility"].isDouble()) {
@@ -373,9 +374,9 @@ void GameList::LoadCompatibilityList() {
             QString directory = game["directory"].toString();
             QJsonArray ids = game["releases"].toArray();
 
-            for (const QJsonValue& value : ids) {
-                QJsonObject object = value.toObject();
-                QString id = object["id"].toString();
+            for (const QJsonValueRef& id_ref : ids) {
+                QJsonObject id_object = id_ref.toObject();
+                QString id = id_object["id"].toString();
                 compatibility_list.emplace(
                     id.toUpper().toStdString(),
                     std::make_pair(QString::number(compatibility), directory));
diff --git a/src/yuzu/game_list.h b/src/yuzu/game_list.h
index 3fcb298ed..2713e7b54 100644
--- a/src/yuzu/game_list.h
+++ b/src/yuzu/game_list.h
@@ -4,8 +4,6 @@
 
 #pragma once
 
-#include <unordered_map>
-
 #include <QFileSystemWatcher>
 #include <QHBoxLayout>
 #include <QLabel>
@@ -21,6 +19,7 @@
 #include <QWidget>
 
 #include "common/common_types.h"
+#include "yuzu/compatibility_list.h"
 
 class GameListWorker;
 class GMainWindow;
@@ -90,9 +89,8 @@ signals:
     void GameChosen(QString game_path);
     void ShouldCancelWorker();
     void OpenFolderRequested(u64 program_id, GameListOpenTarget target);
-    void NavigateToGamedbEntryRequested(
-        u64 program_id,
-        std::unordered_map<std::string, std::pair<QString, QString>>& compatibility_list);
+    void NavigateToGamedbEntryRequested(u64 program_id,
+                                        const CompatibilityList& compatibility_list);
 
 private slots:
     void onTextChanged(const QString& newText);
@@ -114,7 +112,7 @@ private:
     QStandardItemModel* item_model = nullptr;
     GameListWorker* current_worker = nullptr;
     QFileSystemWatcher* watcher = nullptr;
-    std::unordered_map<std::string, std::pair<QString, QString>> compatibility_list;
+    CompatibilityList compatibility_list;
 };
 
 Q_DECLARE_METATYPE(GameListOpenTarget);
diff --git a/src/yuzu/game_list_p.h b/src/yuzu/game_list_p.h
index 2720bf143..f22e422e5 100644
--- a/src/yuzu/game_list_p.h
+++ b/src/yuzu/game_list_p.h
@@ -176,14 +176,3 @@ public:
         return data(SizeRole).toULongLong() < other.data(SizeRole).toULongLong();
     }
 };
-
-inline auto FindMatchingCompatibilityEntry(
-    const std::unordered_map<std::string, std::pair<QString, QString>>& compatibility_list,
-    u64 program_id) {
-    return std::find_if(
-        compatibility_list.begin(), compatibility_list.end(),
-        [program_id](const std::pair<std::string, std::pair<QString, QString>>& element) {
-            std::string pid = fmt::format("{:016X}", program_id);
-            return element.first == pid;
-        });
-}
diff --git a/src/yuzu/game_list_worker.cpp b/src/yuzu/game_list_worker.cpp
index 9f26935d6..e228d61bd 100644
--- a/src/yuzu/game_list_worker.cpp
+++ b/src/yuzu/game_list_worker.cpp
@@ -20,6 +20,7 @@
 #include "core/file_sys/registered_cache.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/loader.h"
+#include "yuzu/compatibility_list.h"
 #include "yuzu/game_list.h"
 #include "yuzu/game_list_p.h"
 #include "yuzu/game_list_worker.h"
@@ -75,9 +76,8 @@ QString FormatPatchNameVersions(const FileSys::PatchManager& patch_manager, bool
 }
 } // Anonymous namespace
 
-GameListWorker::GameListWorker(
-    FileSys::VirtualFilesystem vfs, QString dir_path, bool deep_scan,
-    const std::unordered_map<std::string, std::pair<QString, QString>>& compatibility_list)
+GameListWorker::GameListWorker(FileSys::VirtualFilesystem vfs, QString dir_path, bool deep_scan,
+                               const CompatibilityList& compatibility_list)
     : vfs(std::move(vfs)), dir_path(std::move(dir_path)), deep_scan(deep_scan),
       compatibility_list(compatibility_list) {}
 
diff --git a/src/yuzu/game_list_worker.h b/src/yuzu/game_list_worker.h
index 42c93fc31..09d20c42f 100644
--- a/src/yuzu/game_list_worker.h
+++ b/src/yuzu/game_list_worker.h
@@ -16,6 +16,7 @@
 #include <QString>
 
 #include "common/common_types.h"
+#include "yuzu/compatibility_list.h"
 
 class QStandardItem;
 
@@ -32,9 +33,8 @@ class GameListWorker : public QObject, public QRunnable {
     Q_OBJECT
 
 public:
-    GameListWorker(
-        std::shared_ptr<FileSys::VfsFilesystem> vfs, QString dir_path, bool deep_scan,
-        const std::unordered_map<std::string, std::pair<QString, QString>>& compatibility_list);
+    GameListWorker(std::shared_ptr<FileSys::VfsFilesystem> vfs, QString dir_path, bool deep_scan,
+                   const CompatibilityList& compatibility_list);
     ~GameListWorker() override;
 
     /// Starts the processing of directory tree information.
@@ -67,6 +67,6 @@ private:
     QStringList watch_list;
     QString dir_path;
     bool deep_scan;
-    const std::unordered_map<std::string, std::pair<QString, QString>>& compatibility_list;
+    const CompatibilityList& compatibility_list;
     std::atomic_bool stop_processing;
 };
diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp
index 2cd282a51..05a4a55e8 100644
--- a/src/yuzu/main.cpp
+++ b/src/yuzu/main.cpp
@@ -47,6 +47,7 @@
 #include "video_core/debug_utils/debug_utils.h"
 #include "yuzu/about_dialog.h"
 #include "yuzu/bootmanager.h"
+#include "yuzu/compatibility_list.h"
 #include "yuzu/configuration/config.h"
 #include "yuzu/configuration/configure_dialog.h"
 #include "yuzu/debugger/console.h"
@@ -446,6 +447,8 @@ QStringList GMainWindow::GetUnsupportedGLExtensions() {
         unsupported_ext.append("ARB_texture_mirror_clamp_to_edge");
     if (!GLAD_GL_ARB_base_instance)
         unsupported_ext.append("ARB_base_instance");
+    if (!GLAD_GL_ARB_texture_storage)
+        unsupported_ext.append("ARB_texture_storage");
 
     // Extensions required to support some texture formats.
     if (!GLAD_GL_EXT_texture_compression_s3tc)
@@ -608,7 +611,7 @@ void GMainWindow::BootGame(const QString& filename) {
     }
 
     setWindowTitle(QString("yuzu %1| %4 | %2-%3")
-                       .arg(Common::g_build_name, Common::g_scm_branch, Common::g_scm_desc,
+                       .arg(Common::g_build_fullname, Common::g_scm_branch, Common::g_scm_desc,
                             QString::fromStdString(title_name)));
 
     render_window->show();
@@ -643,7 +646,7 @@ void GMainWindow::ShutdownGame() {
     game_list->show();
     game_list->setFilterFocus();
     setWindowTitle(QString("yuzu %1| %2-%3")
-                       .arg(Common::g_build_name, Common::g_scm_branch, Common::g_scm_desc));
+                       .arg(Common::g_build_fullname, Common::g_scm_branch, Common::g_scm_desc));
 
     // Disable status bar updates
     status_bar_update_timer.stop();
@@ -725,14 +728,11 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target
     QDesktopServices::openUrl(QUrl::fromLocalFile(qpath));
 }
 
-void GMainWindow::OnGameListNavigateToGamedbEntry(
-    u64 program_id,
-    std::unordered_map<std::string, std::pair<QString, QString>>& compatibility_list) {
-
-    auto it = FindMatchingCompatibilityEntry(compatibility_list, program_id);
+void GMainWindow::OnGameListNavigateToGamedbEntry(u64 program_id,
+                                                  const CompatibilityList& compatibility_list) {
+    const auto it = FindMatchingCompatibilityEntry(compatibility_list, program_id);
 
     QString directory;
-
     if (it != compatibility_list.end())
         directory = it->second.second;
 
diff --git a/src/yuzu/main.h b/src/yuzu/main.h
index 089ea2445..552e3e61c 100644
--- a/src/yuzu/main.h
+++ b/src/yuzu/main.h
@@ -13,6 +13,7 @@
 #include "common/common_types.h"
 #include "core/core.h"
 #include "ui_main.h"
+#include "yuzu/compatibility_list.h"
 #include "yuzu/hotkeys.h"
 
 class Config;
@@ -137,9 +138,8 @@ private slots:
     /// Called whenever a user selects a game in the game list widget.
     void OnGameListLoadFile(QString game_path);
     void OnGameListOpenFolder(u64 program_id, GameListOpenTarget target);
-    void OnGameListNavigateToGamedbEntry(
-        u64 program_id,
-        std::unordered_map<std::string, std::pair<QString, QString>>& compatibility_list);
+    void OnGameListNavigateToGamedbEntry(u64 program_id,
+                                         const CompatibilityList& compatibility_list);
     void OnMenuLoadFile();
     void OnMenuLoadFolder();
     void OnMenuInstallToNAND();
diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp
index 2f7916256..d213929bd 100644
--- a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp
+++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp
@@ -16,6 +16,7 @@
 #include "input_common/keyboard.h"
 #include "input_common/main.h"
 #include "input_common/motion_emu.h"
+#include "input_common/sdl/sdl.h"
 #include "yuzu_cmd/emu_window/emu_window_sdl2.h"
 
 void EmuWindow_SDL2::OnMouseMotion(s32 x, s32 y) {
@@ -93,6 +94,8 @@ bool EmuWindow_SDL2::SupportsRequiredGLExtensions() {
         unsupported_ext.push_back("ARB_texture_mirror_clamp_to_edge");
     if (!GLAD_GL_ARB_base_instance)
         unsupported_ext.push_back("ARB_base_instance");
+    if (!GLAD_GL_ARB_texture_storage)
+        unsupported_ext.push_back("ARB_texture_storage");
 
     // Extensions required to support some texture formats.
     if (!GLAD_GL_EXT_texture_compression_s3tc)
@@ -116,7 +119,7 @@ EmuWindow_SDL2::EmuWindow_SDL2(bool fullscreen) {
     SDL_SetMainReady();
 
     // Initialize the window
-    if (SDL_Init(SDL_INIT_VIDEO) < 0) {
+    if (SDL_Init(SDL_INIT_VIDEO | SDL_INIT_JOYSTICK) < 0) {
         LOG_CRITICAL(Frontend, "Failed to initialize SDL2! Exiting...");
         exit(1);
     }
@@ -176,6 +179,7 @@ EmuWindow_SDL2::EmuWindow_SDL2(bool fullscreen) {
 }
 
 EmuWindow_SDL2::~EmuWindow_SDL2() {
+    InputCommon::SDL::CloseSDLJoysticks();
     SDL_GL_DeleteContext(gl_context);
     SDL_Quit();
 
@@ -220,6 +224,9 @@ void EmuWindow_SDL2::PollEvents() {
         case SDL_QUIT:
             is_open = false;
             break;
+        default:
+            InputCommon::SDL::HandleGameControllerEvent(event);
+            break;
         }
     }
 }
diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp
index 173ac0e0f..b1c364fbb 100644
--- a/src/yuzu_cmd/yuzu.cpp
+++ b/src/yuzu_cmd/yuzu.cpp
@@ -82,6 +82,9 @@ int main(int argc, char** argv) {
     int option_index = 0;
     bool use_gdbstub = Settings::values.use_gdbstub;
     u32 gdb_port = static_cast<u32>(Settings::values.gdbstub_port);
+
+    InitializeLogging();
+
     char* endarg;
 #ifdef _WIN32
     int argc_w;
@@ -144,8 +147,6 @@ int main(int argc, char** argv) {
     LocalFree(argv_w);
 #endif
 
-    InitializeLogging();
-
     MicroProfileOnThreadCreate("EmuThread");
     SCOPE_EXIT({ MicroProfileShutdown(); });