renderer_opengl: Replace Memory::GetPointer with Read/WriteBlock.

Only trivial usages are rewritten.

src/audio_core/stream.cpp

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

src/common/logging/backend.cpp

@@ -12,7 +12,8 @@
 #include <thread>
 #include <vector>
 #ifdef _WIN32
-#include <share.h> // For _SH_DENYWR
+#include <share.h>   // For _SH_DENYWR
+#include <windows.h> // For OutputDebugStringA
 #else
 #define _SH_DENYWR 0
 #endif
@@ -139,12 +140,18 @@ void FileBackend::Write(const Entry& entry) {
     if (!file.IsOpen() || bytes_written > MAX_BYTES_WRITTEN) {
         return;
     }
-    bytes_written += file.WriteString(FormatLogMessage(entry) + '\n');
+    bytes_written += file.WriteString(FormatLogMessage(entry).append(1, '\n'));
     if (entry.log_level >= Level::Error) {
         file.Flush();
     }
 }
 
+void DebuggerBackend::Write(const Entry& entry) {
+#ifdef _WIN32
+    ::OutputDebugStringA(FormatLogMessage(entry).append(1, '\n').c_str());
+#endif
+}
+
 /// Macro listing all log classes. Code should define CLS and SUB as desired before invoking this.
 #define ALL_LOG_CLASSES()                                                                          \
     CLS(Log)                                                                                       \

src/common/logging/backend.h

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

src/common/telemetry.h

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

src/core/hle/kernel/svc.cpp

@@ -34,6 +34,7 @@
 #include "core/hle/lock.h"
 #include "core/hle/result.h"
 #include "core/hle/service/service.h"
+#include "core/settings.h"
 
 namespace Kernel {
 namespace {
@@ -395,16 +396,42 @@ struct BreakReason {
 /// Break program execution
 static void Break(u32 reason, u64 info1, u64 info2) {
     BreakReason break_reason{reason};
+    bool has_dumped_buffer{};
 
+    const auto handle_debug_buffer = [&](VAddr addr, u64 sz) {
+        if (sz == 0 || addr == 0 || has_dumped_buffer) {
+            return;
+        }
+
+        // This typically is an error code so we're going to assume this is the case
+        if (sz == sizeof(u32)) {
+            LOG_CRITICAL(Debug_Emulated, "debug_buffer_err_code={:X}", Memory::Read32(addr));
+        } else {
+            // We don't know what's in here so we'll hexdump it
+            std::vector<u8> debug_buffer(sz);
+            Memory::ReadBlock(addr, debug_buffer.data(), sz);
+            std::string hexdump;
+            for (std::size_t i = 0; i < debug_buffer.size(); i++) {
+                hexdump += fmt::format("{:02X} ", debug_buffer[i]);
+                if (i != 0 && i % 16 == 0) {
+                    hexdump += '\n';
+                }
+            }
+            LOG_CRITICAL(Debug_Emulated, "debug_buffer=\n{}", hexdump);
+        }
+        has_dumped_buffer = true;
+    };
     switch (break_reason.break_type) {
     case BreakType::Panic:
         LOG_CRITICAL(Debug_Emulated, "Signalling debugger, PANIC! info1=0x{:016X}, info2=0x{:016X}",
                      info1, info2);
+        handle_debug_buffer(info1, info2);
         break;
     case BreakType::AssertionFailed:
         LOG_CRITICAL(Debug_Emulated,
                      "Signalling debugger, Assertion failed! info1=0x{:016X}, info2=0x{:016X}",
                      info1, info2);
+        handle_debug_buffer(info1, info2);
         break;
     case BreakType::PreNROLoad:
         LOG_WARNING(
@@ -433,6 +460,7 @@ static void Break(u32 reason, u64 info1, u64 info2) {
             Debug_Emulated,
             "Signalling debugger, Unknown break reason {}, info1=0x{:016X}, info2=0x{:016X}",
             static_cast<u32>(break_reason.break_type.Value()), info1, info2);
+        handle_debug_buffer(info1, info2);
         break;
     }
 
@@ -441,6 +469,7 @@ static void Break(u32 reason, u64 info1, u64 info2) {
             Debug_Emulated,
             "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}",
             reason, info1, info2);
+        handle_debug_buffer(info1, info2);
         ASSERT(false);
 
         Core::CurrentProcess()->PrepareForTermination();
@@ -530,7 +559,7 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
         *result = 0;
         break;
     case GetInfoType::RandomEntropy:
-        *result = 0;
+        *result = Settings::values.rng_seed.value_or(0);
         break;
     case GetInfoType::ASLRRegionBaseAddr:
         *result = vm_manager.GetASLRRegionBaseAddress();

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

@@ -242,6 +242,28 @@ void Module::Interface::GetBaasAccountManagerForApplication(Kernel::HLERequestCo
     LOG_DEBUG(Service_ACC, "called");
 }
 
+void Module::Interface::TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_ACC, "called");
+    // A u8 is passed into this function which we can safely ignore. It's to determine if we have
+    // access to use the network or not by the looks of it
+    IPC::ResponseBuilder rb{ctx, 6};
+    if (profile_manager->GetUserCount() != 1) {
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u128>(INVALID_UUID);
+        return;
+    }
+    auto user_list = profile_manager->GetAllUsers();
+    if (user_list.empty()) {
+        rb.Push(ResultCode(-1)); // TODO(ogniK): Find the correct error code
+        rb.PushRaw<u128>(INVALID_UUID);
+        return;
+    }
+
+    // Select the first user we have
+    rb.Push(RESULT_SUCCESS);
+    rb.PushRaw<u128>(profile_manager->GetUser(0)->uuid);
+}
+
 Module::Interface::Interface(std::shared_ptr<Module> module,
                              std::shared_ptr<ProfileManager> profile_manager, const char* name)
     : ServiceFramework(name), module(std::move(module)),

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

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

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

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

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

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

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

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

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

@@ -57,7 +57,8 @@ struct UUID {
 };
 static_assert(sizeof(UUID) == 16, "UUID is an invalid size!");
 
-using ProfileUsername = std::array<u8, 0x20>;
+constexpr std::size_t profile_username_size = 32;
+using ProfileUsername = std::array<u8, profile_username_size>;
 using ProfileData = std::array<u8, MAX_DATA>;
 using UserIDArray = std::array<UUID, MAX_USERS>;
 

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

@@ -338,7 +338,54 @@ void ISelfController::GetIdleTimeDetectionExtension(Kernel::HLERequestContext& c
     LOG_WARNING(Service_AM, "(STUBBED) called");
 }
 
-ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter") {
+AppletMessageQueue::AppletMessageQueue() {
+    auto& kernel = Core::System::GetInstance().Kernel();
+    on_new_message = Kernel::Event::Create(kernel, Kernel::ResetType::Sticky,
+                                           "AMMessageQueue:OnMessageRecieved");
+    on_operation_mode_changed = Kernel::Event::Create(kernel, Kernel::ResetType::OneShot,
+                                                      "AMMessageQueue:OperationModeChanged");
+}
+
+AppletMessageQueue::~AppletMessageQueue() = default;
+
+const Kernel::SharedPtr<Kernel::Event>& AppletMessageQueue::GetMesssageRecieveEvent() const {
+    return on_new_message;
+}
+
+const Kernel::SharedPtr<Kernel::Event>& AppletMessageQueue::GetOperationModeChangedEvent() const {
+    return on_operation_mode_changed;
+}
+
+void AppletMessageQueue::PushMessage(AppletMessage msg) {
+    messages.push(msg);
+    on_new_message->Signal();
+}
+
+AppletMessageQueue::AppletMessage AppletMessageQueue::PopMessage() {
+    if (messages.empty()) {
+        on_new_message->Clear();
+        return AppletMessage::NoMessage;
+    }
+    auto msg = messages.front();
+    messages.pop();
+    if (messages.empty()) {
+        on_new_message->Clear();
+    }
+    return msg;
+}
+
+std::size_t AppletMessageQueue::GetMessageCount() const {
+    return messages.size();
+}
+
+void AppletMessageQueue::OperationModeChanged() {
+    PushMessage(AppletMessage::OperationModeChanged);
+    PushMessage(AppletMessage::PerformanceModeChanged);
+    on_operation_mode_changed->Signal();
+}
+
+ICommonStateGetter::ICommonStateGetter(std::shared_ptr<AppletMessageQueue> msg_queue)
+    : ServiceFramework("ICommonStateGetter"), msg_queue(std::move(msg_queue)) {
     // clang-format off
     static const FunctionInfo functions[] = {
         {0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"},
@@ -388,21 +435,19 @@ void ICommonStateGetter::GetBootMode(Kernel::HLERequestContext& ctx) {
 }
 
 void ICommonStateGetter::GetEventHandle(Kernel::HLERequestContext& ctx) {
-    event->Signal();
-
     IPC::ResponseBuilder rb{ctx, 2, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushCopyObjects(event);
+    rb.PushCopyObjects(msg_queue->GetMesssageRecieveEvent());
 
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    LOG_DEBUG(Service_AM, "called");
 }
 
 void ICommonStateGetter::ReceiveMessage(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
-    rb.Push<u32>(15);
+    rb.PushEnum<AppletMessageQueue::AppletMessage>(msg_queue->PopMessage());
 
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    LOG_DEBUG(Service_AM, "called");
 }
 
 void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) {
@@ -414,13 +459,11 @@ void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) {
 }
 
 void ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent(Kernel::HLERequestContext& ctx) {
-    event->Signal();
-
     IPC::ResponseBuilder rb{ctx, 2, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushCopyObjects(event);
+    rb.PushCopyObjects(msg_queue->GetOperationModeChangedEvent());
 
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    LOG_DEBUG(Service_AM, "called");
 }
 
 void ICommonStateGetter::GetDefaultDisplayResolution(Kernel::HLERequestContext& ctx) {
@@ -444,7 +487,7 @@ void ICommonStateGetter::GetOperationMode(Kernel::HLERequestContext& ctx) {
     rb.Push(RESULT_SUCCESS);
     rb.Push(static_cast<u8>(use_docked_mode ? OperationMode::Docked : OperationMode::Handheld));
 
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    LOG_DEBUG(Service_AM, "called");
 }
 
 void ICommonStateGetter::GetPerformanceMode(Kernel::HLERequestContext& ctx) {
@@ -454,7 +497,7 @@ void ICommonStateGetter::GetPerformanceMode(Kernel::HLERequestContext& ctx) {
     rb.Push(static_cast<u32>(use_docked_mode ? APM::PerformanceMode::Docked
                                              : APM::PerformanceMode::Handheld));
 
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    LOG_DEBUG(Service_AM, "called");
 }
 
 class IStorageAccessor final : public ServiceFramework<IStorageAccessor> {
@@ -840,8 +883,12 @@ void IApplicationFunctions::GetPseudoDeviceId(Kernel::HLERequestContext& ctx) {
 
 void InstallInterfaces(SM::ServiceManager& service_manager,
                        std::shared_ptr<NVFlinger::NVFlinger> nvflinger) {
-    std::make_shared<AppletAE>(nvflinger)->InstallAsService(service_manager);
-    std::make_shared<AppletOE>(nvflinger)->InstallAsService(service_manager);
+    auto message_queue = std::make_shared<AppletMessageQueue>();
+    message_queue->PushMessage(
+        AppletMessageQueue::AppletMessage::FocusStateChanged); // Needed on game boot
+
+    std::make_shared<AppletAE>(nvflinger, message_queue)->InstallAsService(service_manager);
+    std::make_shared<AppletOE>(nvflinger, message_queue)->InstallAsService(service_manager);
     std::make_shared<IdleSys>()->InstallAsService(service_manager);
     std::make_shared<OMM>()->InstallAsService(service_manager);
     std::make_shared<SPSM>()->InstallAsService(service_manager);

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

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <queue>
 #include "core/hle/service/service.h"
 
 namespace Kernel {
@@ -39,6 +40,31 @@ enum SystemLanguage {
     TraditionalChinese = 16,
 };
 
+class AppletMessageQueue {
+public:
+    enum class AppletMessage : u32 {
+        NoMessage = 0,
+        FocusStateChanged = 15,
+        OperationModeChanged = 30,
+        PerformanceModeChanged = 31,
+    };
+
+    AppletMessageQueue();
+    ~AppletMessageQueue();
+
+    const Kernel::SharedPtr<Kernel::Event>& GetMesssageRecieveEvent() const;
+    const Kernel::SharedPtr<Kernel::Event>& GetOperationModeChangedEvent() const;
+    void PushMessage(AppletMessage msg);
+    AppletMessage PopMessage();
+    std::size_t GetMessageCount() const;
+    void OperationModeChanged();
+
+private:
+    std::queue<AppletMessage> messages;
+    Kernel::SharedPtr<Kernel::Event> on_new_message;
+    Kernel::SharedPtr<Kernel::Event> on_operation_mode_changed;
+};
+
 class IWindowController final : public ServiceFramework<IWindowController> {
 public:
     IWindowController();
@@ -102,7 +128,7 @@ private:
 
 class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> {
 public:
-    ICommonStateGetter();
+    explicit ICommonStateGetter(std::shared_ptr<AppletMessageQueue> msg_queue);
     ~ICommonStateGetter() override;
 
 private:
@@ -126,6 +152,7 @@ private:
     void GetDefaultDisplayResolution(Kernel::HLERequestContext& ctx);
 
     Kernel::SharedPtr<Kernel::Event> event;
+    std::shared_ptr<AppletMessageQueue> msg_queue;
 };
 
 class ILibraryAppletCreator final : public ServiceFramework<ILibraryAppletCreator> {

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

@@ -12,8 +12,10 @@ namespace Service::AM {
 
 class ILibraryAppletProxy final : public ServiceFramework<ILibraryAppletProxy> {
 public:
-    explicit ILibraryAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
-        : ServiceFramework("ILibraryAppletProxy"), nvflinger(std::move(nvflinger)) {
+    explicit ILibraryAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger,
+                                 std::shared_ptr<AppletMessageQueue> msg_queue)
+        : ServiceFramework("ILibraryAppletProxy"), nvflinger(std::move(nvflinger)),
+          msg_queue(std::move(msg_queue)) {
         static const FunctionInfo functions[] = {
             {0, &ILibraryAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"},
             {1, &ILibraryAppletProxy::GetSelfController, "GetSelfController"},
@@ -32,7 +34,7 @@ private:
     void GetCommonStateGetter(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 0, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushIpcInterface<ICommonStateGetter>();
+        rb.PushIpcInterface<ICommonStateGetter>(msg_queue);
         LOG_DEBUG(Service_AM, "called");
     }
 
@@ -93,12 +95,15 @@ private:
     }
 
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
+    std::shared_ptr<AppletMessageQueue> msg_queue;
 };
 
 class ISystemAppletProxy final : public ServiceFramework<ISystemAppletProxy> {
 public:
-    explicit ISystemAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
-        : ServiceFramework("ISystemAppletProxy"), nvflinger(std::move(nvflinger)) {
+    explicit ISystemAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger,
+                                std::shared_ptr<AppletMessageQueue> msg_queue)
+        : ServiceFramework("ISystemAppletProxy"), nvflinger(std::move(nvflinger)),
+          msg_queue(std::move(msg_queue)) {
         static const FunctionInfo functions[] = {
             {0, &ISystemAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"},
             {1, &ISystemAppletProxy::GetSelfController, "GetSelfController"},
@@ -119,7 +124,7 @@ private:
     void GetCommonStateGetter(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 0, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushIpcInterface<ICommonStateGetter>();
+        rb.PushIpcInterface<ICommonStateGetter>(msg_queue);
         LOG_DEBUG(Service_AM, "called");
     }
 
@@ -186,31 +191,34 @@ private:
         LOG_DEBUG(Service_AM, "called");
     }
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
+    std::shared_ptr<AppletMessageQueue> msg_queue;
 };
 
 void AppletAE::OpenSystemAppletProxy(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<ISystemAppletProxy>(nvflinger);
+    rb.PushIpcInterface<ISystemAppletProxy>(nvflinger, msg_queue);
     LOG_DEBUG(Service_AM, "called");
 }
 
 void AppletAE::OpenLibraryAppletProxy(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger);
+    rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger, msg_queue);
     LOG_DEBUG(Service_AM, "called");
 }
 
 void AppletAE::OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger);
+    rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger, msg_queue);
     LOG_DEBUG(Service_AM, "called");
 }
 
-AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
-    : ServiceFramework("appletAE"), nvflinger(std::move(nvflinger)) {
+AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger,
+                   std::shared_ptr<AppletMessageQueue> msg_queue)
+    : ServiceFramework("appletAE"), nvflinger(std::move(nvflinger)),
+      msg_queue(std::move(msg_queue)) {
     // clang-format off
     static const FunctionInfo functions[] = {
         {100, &AppletAE::OpenSystemAppletProxy, "OpenSystemAppletProxy"},
@@ -228,4 +236,8 @@ AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
 
 AppletAE::~AppletAE() = default;
 
+const std::shared_ptr<AppletMessageQueue>& AppletAE::GetMessageQueue() const {
+    return msg_queue;
+}
+
 } // namespace Service::AM

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

@@ -17,15 +17,19 @@ namespace AM {
 
 class AppletAE final : public ServiceFramework<AppletAE> {
 public:
-    explicit AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger);
+    explicit AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger,
+                      std::shared_ptr<AppletMessageQueue> msg_queue);
     ~AppletAE() override;
 
+    const std::shared_ptr<AppletMessageQueue>& GetMessageQueue() const;
+
 private:
     void OpenSystemAppletProxy(Kernel::HLERequestContext& ctx);
     void OpenLibraryAppletProxy(Kernel::HLERequestContext& ctx);
     void OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
+    std::shared_ptr<AppletMessageQueue> msg_queue;
 };
 
 } // namespace AM

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

@@ -12,8 +12,10 @@ namespace Service::AM {
 
 class IApplicationProxy final : public ServiceFramework<IApplicationProxy> {
 public:
-    explicit IApplicationProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
-        : ServiceFramework("IApplicationProxy"), nvflinger(std::move(nvflinger)) {
+    explicit IApplicationProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger,
+                               std::shared_ptr<AppletMessageQueue> msg_queue)
+        : ServiceFramework("IApplicationProxy"), nvflinger(std::move(nvflinger)),
+          msg_queue(std::move(msg_queue)) {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IApplicationProxy::GetCommonStateGetter, "GetCommonStateGetter"},
@@ -70,7 +72,7 @@ private:
     void GetCommonStateGetter(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 0, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushIpcInterface<ICommonStateGetter>();
+        rb.PushIpcInterface<ICommonStateGetter>(msg_queue);
         LOG_DEBUG(Service_AM, "called");
     }
 
@@ -89,17 +91,20 @@ private:
     }
 
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
+    std::shared_ptr<AppletMessageQueue> msg_queue;
 };
 
 void AppletOE::OpenApplicationProxy(Kernel::HLERequestContext& ctx) {
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<IApplicationProxy>(nvflinger);
+    rb.PushIpcInterface<IApplicationProxy>(nvflinger, msg_queue);
     LOG_DEBUG(Service_AM, "called");
 }
 
-AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
-    : ServiceFramework("appletOE"), nvflinger(std::move(nvflinger)) {
+AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger,
+                   std::shared_ptr<AppletMessageQueue> msg_queue)
+    : ServiceFramework("appletOE"), nvflinger(std::move(nvflinger)),
+      msg_queue(std::move(msg_queue)) {
     static const FunctionInfo functions[] = {
         {0, &AppletOE::OpenApplicationProxy, "OpenApplicationProxy"},
     };
@@ -108,4 +113,8 @@ AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
 
 AppletOE::~AppletOE() = default;
 
+const std::shared_ptr<AppletMessageQueue>& AppletOE::GetMessageQueue() const {
+    return msg_queue;
+}
+
 } // namespace Service::AM

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

@@ -17,13 +17,17 @@ namespace AM {
 
 class AppletOE final : public ServiceFramework<AppletOE> {
 public:
-    explicit AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger);
+    explicit AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger,
+                      std::shared_ptr<AppletMessageQueue> msg_queue);
     ~AppletOE() override;
 
+    const std::shared_ptr<AppletMessageQueue>& GetMessageQueue() const;
+
 private:
     void OpenApplicationProxy(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
+    std::shared_ptr<AppletMessageQueue> msg_queue;
 };
 
 } // namespace AM

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

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

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

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

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

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

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

@@ -58,9 +58,9 @@ public:
         /// Rotate source image 90 degrees clockwise
         Rotate90 = 0x04,
         /// Rotate source image 180 degrees
-        Roate180 = 0x03,
+        Rotate180 = 0x03,
         /// Rotate source image 270 degrees clockwise
-        Roate270 = 0x07,
+        Rotate270 = 0x07,
     };
 
     struct Buffer {

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

@@ -3,18 +3,23 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
+#include <chrono>
 #include <cstdlib>
+#include <ctime>
+#include <functional>
 #include <vector>
 #include "common/logging/log.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/service/spl/csrng.h"
 #include "core/hle/service/spl/module.h"
 #include "core/hle/service/spl/spl.h"
+#include "core/settings.h"
 
 namespace Service::SPL {
 
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
-    : ServiceFramework(name), module(std::move(module)) {}
+    : ServiceFramework(name), module(std::move(module)),
+      rng(Settings::values.rng_seed.value_or(std::time(nullptr))) {}
 
 Module::Interface::~Interface() = default;
 
@@ -24,7 +29,7 @@ void Module::Interface::GetRandomBytes(Kernel::HLERequestContext& ctx) {
     std::size_t size = ctx.GetWriteBufferSize();
 
     std::vector<u8> data(size);
-    std::generate(data.begin(), data.end(), std::rand);
+    std::generate(data.begin(), data.end(), rng);
 
     ctx.WriteBuffer(data);
 

src/core/hle/service/spl/module.h

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include <random>
 #include "core/hle/service/service.h"
 
 namespace Service::SPL {
@@ -19,6 +20,9 @@ public:
 
     protected:
         std::shared_ptr<Module> module;
+
+    private:
+        std::mt19937 rng;
     };
 };
 

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

@@ -21,7 +21,7 @@ Time::Time(std::shared_ptr<Module> time, const char* name)
         {102, nullptr, "GetStandardUserSystemClockInitialYear"},
         {200, nullptr, "IsStandardNetworkSystemClockAccuracySufficient"},
         {300, nullptr, "CalculateMonotonicSystemClockBaseTimePoint"},
-        {400, nullptr, "GetClockSnapshot"},
+        {400, &Time::GetClockSnapshot, "GetClockSnapshot"},
         {401, nullptr, "GetClockSnapshotFromSystemClockContext"},
         {500, nullptr, "CalculateStandardUserSystemClockDifferenceByUser"},
         {501, nullptr, "CalculateSpanBetween"},

src/core/hle/service/time/time.cpp

@@ -15,6 +15,44 @@
 
 namespace Service::Time {
 
+static void PosixToCalendar(u64 posix_time, CalendarTime& calendar_time,
+                            CalendarAdditionalInfo& additional_info,
+                            [[maybe_unused]] const TimeZoneRule& /*rule*/) {
+    const std::time_t time(posix_time);
+    const std::tm* tm = std::localtime(&time);
+    if (tm == nullptr) {
+        calendar_time = {};
+        additional_info = {};
+        return;
+    }
+    calendar_time.year = tm->tm_year + 1900;
+    calendar_time.month = tm->tm_mon + 1;
+    calendar_time.day = tm->tm_mday;
+    calendar_time.hour = tm->tm_hour;
+    calendar_time.minute = tm->tm_min;
+    calendar_time.second = tm->tm_sec;
+
+    additional_info.day_of_week = tm->tm_wday;
+    additional_info.day_of_year = tm->tm_yday;
+    std::memcpy(additional_info.name.data(), "UTC", sizeof("UTC"));
+    additional_info.utc_offset = 0;
+}
+
+static u64 CalendarToPosix(const CalendarTime& calendar_time,
+                           [[maybe_unused]] const TimeZoneRule& /*rule*/) {
+    std::tm time{};
+    time.tm_year = calendar_time.year - 1900;
+    time.tm_mon = calendar_time.month - 1;
+    time.tm_mday = calendar_time.day;
+
+    time.tm_hour = calendar_time.hour;
+    time.tm_min = calendar_time.minute;
+    time.tm_sec = calendar_time.second;
+
+    std::time_t epoch_time = std::mktime(&time);
+    return static_cast<u64>(epoch_time);
+}
+
 class ISystemClock final : public ServiceFramework<ISystemClock> {
 public:
     ISystemClock() : ServiceFramework("ISystemClock") {
@@ -80,8 +118,8 @@ public:
             {5, nullptr, "GetTimeZoneRuleVersion"},
             {100, &ITimeZoneService::ToCalendarTime, "ToCalendarTime"},
             {101, &ITimeZoneService::ToCalendarTimeWithMyRule, "ToCalendarTimeWithMyRule"},
-            {201, nullptr, "ToPosixTime"},
-            {202, nullptr, "ToPosixTimeWithMyRule"},
+            {201, &ITimeZoneService::ToPosixTime, "ToPosixTime"},
+            {202, &ITimeZoneService::ToPosixTimeWithMyRule, "ToPosixTimeWithMyRule"},
         };
         RegisterHandlers(functions);
     }
@@ -151,24 +189,29 @@ private:
         rb.PushRaw(additional_info);
     }
 
-    void PosixToCalendar(u64 posix_time, CalendarTime& calendar_time,
-                         CalendarAdditionalInfo& additional_info, const TimeZoneRule& /*rule*/) {
-        std::time_t t(posix_time);
-        std::tm* tm = std::localtime(&t);
-        if (!tm) {
-            return;
-        }
-        calendar_time.year = tm->tm_year + 1900;
-        calendar_time.month = tm->tm_mon + 1;
-        calendar_time.day = tm->tm_mday;
-        calendar_time.hour = tm->tm_hour;
-        calendar_time.minute = tm->tm_min;
-        calendar_time.second = tm->tm_sec;
+    void ToPosixTime(Kernel::HLERequestContext& ctx) {
+        // TODO(ogniK): Figure out how to handle multiple times
+        LOG_WARNING(Service_Time, "(STUBBED) called");
+        IPC::RequestParser rp{ctx};
+        auto calendar_time = rp.PopRaw<CalendarTime>();
+        auto posix_time = CalendarToPosix(calendar_time, {});
 
-        additional_info.day_of_week = tm->tm_wday;
-        additional_info.day_of_year = tm->tm_yday;
-        std::memcpy(additional_info.name.data(), "UTC", sizeof("UTC"));
-        additional_info.utc_offset = 0;
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(1); // Amount of times we're returning
+        ctx.WriteBuffer(&posix_time, sizeof(u64));
+    }
+
+    void ToPosixTimeWithMyRule(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_Time, "(STUBBED) called");
+        IPC::RequestParser rp{ctx};
+        auto calendar_time = rp.PopRaw<CalendarTime>();
+        auto posix_time = CalendarToPosix(calendar_time, {});
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(1); // Amount of times we're returning
+        ctx.WriteBuffer(&posix_time, sizeof(u64));
     }
 };
 
@@ -207,6 +250,55 @@ void Module::Interface::GetStandardLocalSystemClock(Kernel::HLERequestContext& c
     LOG_DEBUG(Service_Time, "called");
 }
 
+void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_Time, "called");
+
+    IPC::RequestParser rp{ctx};
+    auto unknown_u8 = rp.PopRaw<u8>();
+
+    ClockSnapshot clock_snapshot{};
+
+    const s64 time_since_epoch{std::chrono::duration_cast<std::chrono::seconds>(
+                                   std::chrono::system_clock::now().time_since_epoch())
+                                   .count()};
+    CalendarTime calendar_time{};
+    const std::time_t time(time_since_epoch);
+    const std::tm* tm = std::localtime(&time);
+    if (tm == nullptr) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(ResultCode(-1)); // TODO(ogniK): Find appropriate error code
+        return;
+    }
+    SteadyClockTimePoint steady_clock_time_point{CoreTiming::cyclesToMs(CoreTiming::GetTicks()) /
+                                                 1000};
+
+    LocationName location_name{"UTC"};
+    calendar_time.year = tm->tm_year + 1900;
+    calendar_time.month = tm->tm_mon + 1;
+    calendar_time.day = tm->tm_mday;
+    calendar_time.hour = tm->tm_hour;
+    calendar_time.minute = tm->tm_min;
+    calendar_time.second = tm->tm_sec;
+    clock_snapshot.system_posix_time = time_since_epoch;
+    clock_snapshot.network_posix_time = time_since_epoch;
+    clock_snapshot.system_calendar_time = calendar_time;
+    clock_snapshot.network_calendar_time = calendar_time;
+
+    CalendarAdditionalInfo additional_info{};
+    PosixToCalendar(time_since_epoch, calendar_time, additional_info, {});
+
+    clock_snapshot.system_calendar_info = additional_info;
+    clock_snapshot.network_calendar_info = additional_info;
+
+    clock_snapshot.steady_clock_timepoint = steady_clock_time_point;
+    clock_snapshot.location_name = location_name;
+    clock_snapshot.clock_auto_adjustment_enabled = 1;
+    clock_snapshot.ipc_u8 = unknown_u8;
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    ctx.WriteBuffer(&clock_snapshot, sizeof(ClockSnapshot));
+}
+
 Module::Interface::Interface(std::shared_ptr<Module> time, const char* name)
     : ServiceFramework(name), time(std::move(time)) {}
 

src/core/hle/service/time/time.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <array>
+#include "common/common_funcs.h"
 #include "core/hle/service/service.h"
 
 namespace Service::Time {
@@ -53,6 +54,23 @@ struct SystemClockContext {
 static_assert(sizeof(SystemClockContext) == 0x20,
               "SystemClockContext structure has incorrect size");
 
+struct ClockSnapshot {
+    SystemClockContext user_clock_context;
+    SystemClockContext network_clock_context;
+    s64_le system_posix_time;
+    s64_le network_posix_time;
+    CalendarTime system_calendar_time;
+    CalendarTime network_calendar_time;
+    CalendarAdditionalInfo system_calendar_info;
+    CalendarAdditionalInfo network_calendar_info;
+    SteadyClockTimePoint steady_clock_timepoint;
+    LocationName location_name;
+    u8 clock_auto_adjustment_enabled;
+    u8 ipc_u8;
+    INSERT_PADDING_BYTES(2);
+};
+static_assert(sizeof(ClockSnapshot) == 0xd0, "ClockSnapshot is an invalid size");
+
 class Module final {
 public:
     class Interface : public ServiceFramework<Interface> {
@@ -65,6 +83,7 @@ public:
         void GetStandardSteadyClock(Kernel::HLERequestContext& ctx);
         void GetTimeZoneService(Kernel::HLERequestContext& ctx);
         void GetStandardLocalSystemClock(Kernel::HLERequestContext& ctx);
+        void GetClockSnapshot(Kernel::HLERequestContext& ctx);
 
     protected:
         std::shared_ptr<Module> time;

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

@@ -132,11 +132,11 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "BindNoticeEvent"},
-            {1, nullptr, "Unknown1"},
+            {1, nullptr, "UnbindNoticeEvent"},
             {2, nullptr, "GetStatus"},
             {3, nullptr, "GetNotice"},
-            {4, nullptr, "Unknown2"},
-            {5, nullptr, "Unknown3"},
+            {4, nullptr, "EnablePowerRequestNotice"},
+            {5, nullptr, "DisablePowerRequestNotice"},
             {6, nullptr, "ReplyPowerRequest"},
         };
         // clang-format on

src/core/settings.h

@@ -6,6 +6,7 @@
 
 #include <array>
 #include <atomic>
+#include <optional>
 #include <string>
 #include "common/common_types.h"
 
@@ -114,8 +115,9 @@ struct Values {
     // System
     bool use_docked_mode;
     bool enable_nfc;
-    int current_user;
-    int language_index;
+    std::optional<u64> rng_seed;
+    s32 current_user;
+    s32 language_index;
 
     // Controls
     std::array<std::string, NativeButton::NumButtons> buttons;

src/core/telemetry_session.cpp

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

src/core/telemetry_session.h

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

src/video_core/CMakeLists.txt

@@ -21,6 +21,7 @@ add_library(video_core STATIC
     macro_interpreter.h
     memory_manager.cpp
     memory_manager.h
+    rasterizer_cache.cpp
     rasterizer_cache.h
     rasterizer_interface.h
     renderer_base.cpp
@@ -33,6 +34,7 @@ add_library(video_core STATIC
     renderer_opengl/gl_rasterizer.h
     renderer_opengl/gl_rasterizer_cache.cpp
     renderer_opengl/gl_rasterizer_cache.h
+    renderer_opengl/gl_resource_manager.cpp
     renderer_opengl/gl_resource_manager.h
     renderer_opengl/gl_shader_cache.cpp
     renderer_opengl/gl_shader_cache.h
@@ -53,6 +55,8 @@ add_library(video_core STATIC
     renderer_opengl/renderer_opengl.h
     renderer_opengl/utils.cpp
     renderer_opengl/utils.h
+    surface.cpp
+    surface.h
     textures/astc.cpp
     textures/astc.h
     textures/decoders.cpp

src/video_core/engines/maxwell_3d.cpp

@@ -37,21 +37,52 @@ void Maxwell3D::InitializeRegisterDefaults() {
         regs.viewport[viewport].depth_range_near = 0.0f;
         regs.viewport[viewport].depth_range_far = 1.0f;
     }
+    // Doom and Bomberman seems to use the uninitialized registers and just enable blend
+    // so initialize blend registers with sane values
+    regs.blend.equation_rgb = Regs::Blend::Equation::Add;
+    regs.blend.factor_source_rgb = Regs::Blend::Factor::One;
+    regs.blend.factor_dest_rgb = Regs::Blend::Factor::Zero;
+    regs.blend.equation_a = Regs::Blend::Equation::Add;
+    regs.blend.factor_source_a = Regs::Blend::Factor::One;
+    regs.blend.factor_dest_a = Regs::Blend::Factor::Zero;
+    for (std::size_t blend_index = 0; blend_index < Regs::NumRenderTargets; blend_index++) {
+        regs.independent_blend[blend_index].equation_rgb = Regs::Blend::Equation::Add;
+        regs.independent_blend[blend_index].factor_source_rgb = Regs::Blend::Factor::One;
+        regs.independent_blend[blend_index].factor_dest_rgb = Regs::Blend::Factor::Zero;
+        regs.independent_blend[blend_index].equation_a = Regs::Blend::Equation::Add;
+        regs.independent_blend[blend_index].factor_source_a = Regs::Blend::Factor::One;
+        regs.independent_blend[blend_index].factor_dest_a = Regs::Blend::Factor::Zero;
+    }
+    regs.stencil_front_op_fail = Regs::StencilOp::Keep;
+    regs.stencil_front_op_zfail = Regs::StencilOp::Keep;
+    regs.stencil_front_op_zpass = Regs::StencilOp::Keep;
+    regs.stencil_front_func_func = Regs::ComparisonOp::Always;
+    regs.stencil_front_func_mask = 0xFFFFFFFF;
+    regs.stencil_front_mask = 0xFFFFFFFF;
+    regs.stencil_two_side_enable = 1;
+    regs.stencil_back_op_fail = Regs::StencilOp::Keep;
+    regs.stencil_back_op_zfail = Regs::StencilOp::Keep;
+    regs.stencil_back_op_zpass = Regs::StencilOp::Keep;
+    regs.stencil_back_func_func = Regs::ComparisonOp::Always;
+    regs.stencil_back_func_mask = 0xFFFFFFFF;
+    regs.stencil_back_mask = 0xFFFFFFFF;
 }
 
 void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
     // Reset the current macro.
     executing_macro = 0;
 
-    // The requested macro must have been uploaded already.
-    auto macro_code = uploaded_macros.find(method);
-    if (macro_code == uploaded_macros.end()) {
-        LOG_ERROR(HW_GPU, "Macro {:04X} was not uploaded", method);
+    // Lookup the macro offset
+    const u32 entry{(method - MacroRegistersStart) >> 1};
+    const auto& search{macro_offsets.find(entry)};
+    if (search == macro_offsets.end()) {
+        LOG_CRITICAL(HW_GPU, "macro not found for method 0x{:X}!", method);
+        UNREACHABLE();
         return;
     }
 
     // Execute the current macro.
-    macro_interpreter.Execute(macro_code->second, std::move(parameters));
+    macro_interpreter.Execute(search->second, std::move(parameters));
 }
 
 void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
@@ -90,13 +121,25 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
         debug_context->OnEvent(Tegra::DebugContext::Event::MaxwellCommandLoaded, nullptr);
     }
 
+    u32 old = regs.reg_array[method];
     regs.reg_array[method] = value;
 
+    if (value != old) {
+        if (method >= MAXWELL3D_REG_INDEX(vertex_attrib_format) &&
+            method < MAXWELL3D_REG_INDEX(vertex_attrib_format) + regs.vertex_attrib_format.size()) {
+            dirty_flags.vertex_attrib_format = true;
+        }
+    }
+
     switch (method) {
     case MAXWELL3D_REG_INDEX(macros.data): {
         ProcessMacroUpload(value);
         break;
     }
+    case MAXWELL3D_REG_INDEX(macros.bind): {
+        ProcessMacroBind(value);
+        break;
+    }
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]):
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]):
     case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]):
@@ -158,9 +201,13 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
 }
 
 void Maxwell3D::ProcessMacroUpload(u32 data) {
-    // Store the uploaded macro code to interpret them when they're called.
-    auto& macro = uploaded_macros[regs.macros.entry * 2 + MacroRegistersStart];
-    macro.push_back(data);
+    ASSERT_MSG(regs.macros.upload_address < macro_memory.size(),
+               "upload_address exceeded macro_memory size!");
+    macro_memory[regs.macros.upload_address++] = data;
+}
+
+void Maxwell3D::ProcessMacroBind(u32 data) {
+    macro_offsets[regs.macros.entry] = data;
 }
 
 void Maxwell3D::ProcessQueryGet() {

src/video_core/engines/maxwell_3d.h

@@ -345,6 +345,14 @@ public:
             Invert = 6,
             IncrWrap = 7,
             DecrWrap = 8,
+            KeepOGL = 0x1E00,
+            ZeroOGL = 0,
+            ReplaceOGL = 0x1E01,
+            IncrOGL = 0x1E02,
+            DecrOGL = 0x1E03,
+            InvertOGL = 0x150A,
+            IncrWrapOGL = 0x8507,
+            DecrWrapOGL = 0x8508,
         };
 
         enum class MemoryLayout : u32 {
@@ -462,6 +470,16 @@ public:
             }
         };
 
+        struct ColorMask {
+            union {
+                u32 raw;
+                BitField<0, 4, u32> R;
+                BitField<4, 4, u32> G;
+                BitField<8, 4, u32> B;
+                BitField<12, 4, u32> A;
+            };
+        };
+
         bool IsShaderConfigEnabled(std::size_t index) const {
             // The VertexB is always enabled.
             if (index == static_cast<std::size_t>(Regs::ShaderProgram::VertexB)) {
@@ -475,12 +493,13 @@ public:
                 INSERT_PADDING_WORDS(0x45);
 
                 struct {
-                    INSERT_PADDING_WORDS(1);
+                    u32 upload_address;
                     u32 data;
                     u32 entry;
+                    u32 bind;
                 } macros;
 
-                INSERT_PADDING_WORDS(0x189);
+                INSERT_PADDING_WORDS(0x188);
 
                 u32 tfb_enabled;
 
@@ -570,7 +589,11 @@ public:
                 u32 stencil_back_mask;
                 u32 stencil_back_func_mask;
 
-                INSERT_PADDING_WORDS(0x13);
+                INSERT_PADDING_WORDS(0xC);
+
+                u32 color_mask_common;
+
+                INSERT_PADDING_WORDS(0x6);
 
                 u32 rt_separate_frag_data;
 
@@ -645,8 +668,14 @@ public:
                 ComparisonOp depth_test_func;
                 float alpha_test_ref;
                 ComparisonOp alpha_test_func;
-
-                INSERT_PADDING_WORDS(0x9);
+                u32 draw_tfb_stride;
+                struct {
+                    float r;
+                    float g;
+                    float b;
+                    float a;
+                } blend_color;
+                INSERT_PADDING_WORDS(0x4);
 
                 struct {
                     u32 separate_alpha;
@@ -840,8 +869,9 @@ public:
                     BitField<6, 4, u32> RT;
                     BitField<10, 11, u32> layer;
                 } clear_buffers;
-
-                INSERT_PADDING_WORDS(0x4B);
+                INSERT_PADDING_WORDS(0xB);
+                std::array<ColorMask, NumRenderTargets> color_mask;
+                INSERT_PADDING_WORDS(0x38);
 
                 struct {
                     u32 query_address_high;
@@ -982,6 +1012,12 @@ public:
     State state{};
     MemoryManager& memory_manager;
 
+    struct DirtyFlags {
+        bool vertex_attrib_format = true;
+    };
+
+    DirtyFlags dirty_flags;
+
     /// Reads a register value located at the input method address
     u32 GetRegisterValue(u32 method) const;
 
@@ -994,12 +1030,25 @@ public:
     /// Returns the texture information for a specific texture in a specific shader stage.
     Texture::FullTextureInfo GetStageTexture(Regs::ShaderStage stage, std::size_t offset) const;
 
+    /// Memory for macro code - it's undetermined how big this is, however 1MB is much larger than
+    /// we've seen used.
+    using MacroMemory = std::array<u32, 0x40000>;
+
+    /// Gets a reference to macro memory.
+    const MacroMemory& GetMacroMemory() const {
+        return macro_memory;
+    }
+
 private:
     void InitializeRegisterDefaults();
 
     VideoCore::RasterizerInterface& rasterizer;
 
-    std::unordered_map<u32, std::vector<u32>> uploaded_macros;
+    /// Start offsets of each macro in macro_memory
+    std::unordered_map<u32, u32> macro_offsets;
+
+    /// Memory for macro code
+    MacroMemory macro_memory;
 
     /// Macro method that is currently being executed / being fed parameters.
     u32 executing_macro = 0;
@@ -1022,9 +1071,12 @@ private:
      */
     void CallMacroMethod(u32 method, std::vector<u32> parameters);
 
-    /// Handles writes to the macro uploading registers.
+    /// Handles writes to the macro uploading register.
     void ProcessMacroUpload(u32 data);
 
+    /// Handles writes to the macro bind register.
+    void ProcessMacroBind(u32 data);
+
     /// Handles a write to the CLEAR_BUFFERS register.
     void ProcessClearBuffers();
 
@@ -1058,6 +1110,7 @@ ASSERT_REG_POSITION(scissor_test, 0x380);
 ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5);
 ASSERT_REG_POSITION(stencil_back_mask, 0x3D6);
 ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7);
+ASSERT_REG_POSITION(color_mask_common, 0x3E4);
 ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB);
 ASSERT_REG_POSITION(zeta, 0x3F8);
 ASSERT_REG_POSITION(vertex_attrib_format, 0x458);
@@ -1070,6 +1123,10 @@ ASSERT_REG_POSITION(depth_write_enabled, 0x4BA);
 ASSERT_REG_POSITION(alpha_test_enabled, 0x4BB);
 ASSERT_REG_POSITION(d3d_cull_mode, 0x4C2);
 ASSERT_REG_POSITION(depth_test_func, 0x4C3);
+ASSERT_REG_POSITION(alpha_test_ref, 0x4C4);
+ASSERT_REG_POSITION(alpha_test_func, 0x4C5);
+ASSERT_REG_POSITION(draw_tfb_stride, 0x4C6);
+ASSERT_REG_POSITION(blend_color, 0x4C7);
 ASSERT_REG_POSITION(blend, 0x4CF);
 ASSERT_REG_POSITION(stencil_enable, 0x4E0);
 ASSERT_REG_POSITION(stencil_front_op_fail, 0x4E1);
@@ -1100,6 +1157,7 @@ ASSERT_REG_POSITION(instanced_arrays, 0x620);
 ASSERT_REG_POSITION(cull, 0x646);
 ASSERT_REG_POSITION(logic_op, 0x671);
 ASSERT_REG_POSITION(clear_buffers, 0x674);
+ASSERT_REG_POSITION(color_mask, 0x680);
 ASSERT_REG_POSITION(query, 0x6C0);
 ASSERT_REG_POSITION(vertex_array[0], 0x700);
 ASSERT_REG_POSITION(independent_blend, 0x780);

src/video_core/engines/shader_bytecode.h

@@ -577,6 +577,10 @@ union Instruction {
         BitField<55, 1, u64> saturate;
     } fmul32;
 
+    union {
+        BitField<52, 1, u64> generates_cc;
+    } op_32;
+
     union {
         BitField<48, 1, u64> is_signed;
     } shift;
@@ -1231,6 +1235,7 @@ union Instruction {
     BitField<60, 1, u64> is_b_gpr;
     BitField<59, 1, u64> is_c_gpr;
     BitField<20, 24, s64> smem_imm;
+    BitField<0, 5, ControlCode> flow_control_code;
 
     Attribute attribute;
     Sampler sampler;
@@ -1658,4 +1663,4 @@ private:
     }
 };
 
-} // namespace Tegra::Shader
+} // namespace Tegra::Shader

src/video_core/macro_interpreter.cpp

@@ -11,7 +11,7 @@ namespace Tegra {
 
 MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {}
 
-void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> parameters) {
+void MacroInterpreter::Execute(u32 offset, std::vector<u32> parameters) {
     Reset();
     registers[1] = parameters[0];
     this->parameters = std::move(parameters);
@@ -19,7 +19,7 @@ void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> pa
     // Execute the code until we hit an exit condition.
     bool keep_executing = true;
     while (keep_executing) {
-        keep_executing = Step(code, false);
+        keep_executing = Step(offset, false);
     }
 
     // Assert the the macro used all the input parameters
@@ -37,10 +37,10 @@ void MacroInterpreter::Reset() {
     next_parameter_index = 1;
 }
 
-bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
+bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
     u32 base_address = pc;
 
-    Opcode opcode = GetOpcode(code);
+    Opcode opcode = GetOpcode(offset);
     pc += 4;
 
     // Update the program counter if we were delayed
@@ -108,7 +108,7 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
 
             delayed_pc = base_address + opcode.GetBranchTarget();
             // Execute one more instruction due to the delay slot.
-            return Step(code, true);
+            return Step(offset, true);
         }
         break;
     }
@@ -121,17 +121,18 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
         // Exit has a delay slot, execute the next instruction
         // Note: Executing an exit during a branch delay slot will cause the instruction at the
         // branch target to be executed before exiting.
-        Step(code, true);
+        Step(offset, true);
         return false;
     }
 
     return true;
 }
 
-MacroInterpreter::Opcode MacroInterpreter::GetOpcode(const std::vector<u32>& code) const {
+MacroInterpreter::Opcode MacroInterpreter::GetOpcode(u32 offset) const {
+    const auto& macro_memory{maxwell3d.GetMacroMemory()};
     ASSERT((pc % sizeof(u32)) == 0);
-    ASSERT(pc < code.size() * sizeof(u32));
-    return {code[pc / sizeof(u32)]};
+    ASSERT((pc + offset) < macro_memory.size() * sizeof(u32));
+    return {macro_memory[offset + pc / sizeof(u32)]};
 }
 
 u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const {

src/video_core/macro_interpreter.h

@@ -22,10 +22,10 @@ public:
 
     /**
      * Executes the macro code with the specified input parameters.
-     * @param code The macro byte code to execute
-     * @param parameters The parameters of the macro
+     * @param offset Offset to start execution at.
+     * @param parameters The parameters of the macro.
      */
-    void Execute(const std::vector<u32>& code, std::vector<u32> parameters);
+    void Execute(u32 offset, std::vector<u32> parameters);
 
 private:
     enum class Operation : u32 {
@@ -110,11 +110,11 @@ private:
     /**
      * Executes a single macro instruction located at the current program counter. Returns whether
      * the interpreter should keep running.
-     * @param code The macro code to execute.
+     * @param offset Offset to start execution at.
      * @param is_delay_slot Whether the current step is being executed due to a delay slot in a
      * previous instruction.
      */
-    bool Step(const std::vector<u32>& code, bool is_delay_slot);
+    bool Step(u32 offset, bool is_delay_slot);
 
     /// Calculates the result of an ALU operation. src_a OP src_b;
     u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const;
@@ -127,7 +127,7 @@ private:
     bool EvaluateBranchCondition(BranchCondition cond, u32 value) const;
 
     /// Reads an opcode at the current program counter location.
-    Opcode GetOpcode(const std::vector<u32>& code) const;
+    Opcode GetOpcode(u32 offset) const;
 
     /// Returns the specified register's value. Register 0 is hardcoded to always return 0.
     u32 GetRegister(u32 register_id) const;

src/video_core/memory_manager.cpp

@@ -4,18 +4,21 @@
 
 #include "common/alignment.h"
 #include "common/assert.h"
+#include "common/logging/log.h"
 #include "video_core/memory_manager.h"
 
 namespace Tegra {
 
 GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
-    std::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, align);
-    ASSERT(gpu_addr);
+    const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, align, PageStatus::Unmapped)};
 
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(*gpu_addr + offset);
+    ASSERT_MSG(gpu_addr, "unable to find available GPU memory");
+
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(*gpu_addr + offset)};
 
         ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
+
         slot = static_cast<u64>(PageStatus::Allocated);
     }
 
@@ -23,10 +26,11 @@ GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
 }
 
 GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) {
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(gpu_addr + offset);
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(gpu_addr + offset)};
 
         ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
+
         slot = static_cast<u64>(PageStatus::Allocated);
     }
 
@@ -34,17 +38,19 @@ GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) {
 }
 
 GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) {
-    std::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, PAGE_SIZE);
-    ASSERT(gpu_addr);
+    const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, PAGE_SIZE, PageStatus::Unmapped)};
 
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(*gpu_addr + offset);
+    ASSERT_MSG(gpu_addr, "unable to find available GPU memory");
+
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(*gpu_addr + offset)};
 
         ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
+
         slot = cpu_addr + offset;
     }
 
-    MappedRegion region{cpu_addr, *gpu_addr, size};
+    const MappedRegion region{cpu_addr, *gpu_addr, size};
     mapped_regions.push_back(region);
 
     return *gpu_addr;
@@ -53,14 +59,31 @@ GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) {
 GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size) {
     ASSERT((gpu_addr & PAGE_MASK) == 0);
 
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(gpu_addr + offset);
+    if (PageSlot(gpu_addr) != static_cast<u64>(PageStatus::Allocated)) {
+        // Page has been already mapped. In this case, we must find a new area of memory to use that
+        // is different than the specified one. Super Mario Odyssey hits this scenario when changing
+        // areas, but we do not want to overwrite the old pages.
+        // TODO(bunnei): We need to write a hardware test to confirm this behavior.
+
+        LOG_ERROR(HW_GPU, "attempting to map addr 0x{:016X}, which is not available!", gpu_addr);
+
+        const std::optional<GPUVAddr> new_gpu_addr{
+            FindFreeBlock(gpu_addr, size, PAGE_SIZE, PageStatus::Allocated)};
+
+        ASSERT_MSG(new_gpu_addr, "unable to find available GPU memory");
+
+        gpu_addr = *new_gpu_addr;
+    }
+
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(gpu_addr + offset)};
 
         ASSERT(slot == static_cast<u64>(PageStatus::Allocated));
+
         slot = cpu_addr + offset;
     }
 
-    MappedRegion region{cpu_addr, gpu_addr, size};
+    const MappedRegion region{cpu_addr, gpu_addr, size};
     mapped_regions.push_back(region);
 
     return gpu_addr;
@@ -69,11 +92,12 @@ GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size)
 GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) {
     ASSERT((gpu_addr & PAGE_MASK) == 0);
 
-    for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
-        VAddr& slot = PageSlot(gpu_addr + offset);
+    for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
+        VAddr& slot{PageSlot(gpu_addr + offset)};
 
         ASSERT(slot != static_cast<u64>(PageStatus::Allocated) &&
                slot != static_cast<u64>(PageStatus::Unmapped));
+
         slot = static_cast<u64>(PageStatus::Unmapped);
     }
 
@@ -97,13 +121,14 @@ GPUVAddr MemoryManager::GetRegionEnd(GPUVAddr region_start) const {
     return {};
 }
 
-std::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
-    GPUVAddr gpu_addr = 0;
-    u64 free_space = 0;
+std::optional<GPUVAddr> MemoryManager::FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,
+                                                     PageStatus status) {
+    GPUVAddr gpu_addr{region_start};
+    u64 free_space{};
     align = (align + PAGE_MASK) & ~PAGE_MASK;
 
     while (gpu_addr + free_space < MAX_ADDRESS) {
-        if (!IsPageMapped(gpu_addr + free_space)) {
+        if (PageSlot(gpu_addr + free_space) == static_cast<u64>(status)) {
             free_space += PAGE_SIZE;
             if (free_space >= size) {
                 return gpu_addr;
@@ -119,7 +144,7 @@ std::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
 }
 
 std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) {
-    VAddr base_addr = PageSlot(gpu_addr);
+    const VAddr base_addr{PageSlot(gpu_addr)};
 
     if (base_addr == static_cast<u64>(PageStatus::Allocated) ||
         base_addr == static_cast<u64>(PageStatus::Unmapped)) {
@@ -133,19 +158,15 @@ std::vector<GPUVAddr> MemoryManager::CpuToGpuAddress(VAddr cpu_addr) const {
     std::vector<GPUVAddr> results;
     for (const auto& region : mapped_regions) {
         if (cpu_addr >= region.cpu_addr && cpu_addr < (region.cpu_addr + region.size)) {
-            u64 offset = cpu_addr - region.cpu_addr;
+            const u64 offset{cpu_addr - region.cpu_addr};
             results.push_back(region.gpu_addr + offset);
         }
     }
     return results;
 }
 
-bool MemoryManager::IsPageMapped(GPUVAddr gpu_addr) {
-    return PageSlot(gpu_addr) != static_cast<u64>(PageStatus::Unmapped);
-}
-
 VAddr& MemoryManager::PageSlot(GPUVAddr gpu_addr) {
-    auto& block = page_table[(gpu_addr >> (PAGE_BITS + PAGE_TABLE_BITS)) & PAGE_TABLE_MASK];
+    auto& block{page_table[(gpu_addr >> (PAGE_BITS + PAGE_TABLE_BITS)) & PAGE_TABLE_MASK]};
     if (!block) {
         block = std::make_unique<PageBlock>();
         block->fill(static_cast<VAddr>(PageStatus::Unmapped));

src/video_core/memory_manager.h

@@ -34,15 +34,15 @@ public:
     static constexpr u64 PAGE_MASK = PAGE_SIZE - 1;
 
 private:
-    std::optional<GPUVAddr> FindFreeBlock(u64 size, u64 align = 1);
-    bool IsPageMapped(GPUVAddr gpu_addr);
-    VAddr& PageSlot(GPUVAddr gpu_addr);
-
     enum class PageStatus : u64 {
         Unmapped = 0xFFFFFFFFFFFFFFFFULL,
         Allocated = 0xFFFFFFFFFFFFFFFEULL,
     };
 
+    std::optional<GPUVAddr> FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,
+                                          PageStatus status);
+    VAddr& PageSlot(GPUVAddr gpu_addr);
+
     static constexpr u64 MAX_ADDRESS{0x10000000000ULL};
     static constexpr u64 PAGE_TABLE_BITS{10};
     static constexpr u64 PAGE_TABLE_SIZE{1 << PAGE_TABLE_BITS};

src/video_core/rasterizer_cache.cpp

@@ -0,0 +1,7 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "video_core/rasterizer_cache.h"
+
+RasterizerCacheObject::~RasterizerCacheObject() = default;

src/video_core/rasterizer_cache.h

@@ -10,13 +10,13 @@
 #include <boost/range/iterator_range_core.hpp>
 
 #include "common/common_types.h"
-#include "core/core.h"
 #include "core/settings.h"
 #include "video_core/rasterizer_interface.h"
-#include "video_core/renderer_base.h"
 
 class RasterizerCacheObject {
 public:
+    virtual ~RasterizerCacheObject();
+
     /// Gets the address of the shader in guest memory, required for cache management
     virtual VAddr GetAddr() const = 0;
 
@@ -64,6 +64,8 @@ class RasterizerCache : NonCopyable {
     friend class RasterizerCacheObject;
 
 public:
+    explicit RasterizerCache(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} {}
+
     /// Write any cached resources overlapping the specified region back to memory
     void FlushRegion(Tegra::GPUVAddr addr, size_t size) {
         const auto& objects{GetSortedObjectsFromRegion(addr, size)};
@@ -109,14 +111,12 @@ protected:
     void Register(const T& object) {
         object->SetIsRegistered(true);
         object_cache.add({GetInterval(object), ObjectSet{object}});
-        auto& rasterizer = Core::System::GetInstance().Renderer().Rasterizer();
         rasterizer.UpdatePagesCachedCount(object->GetAddr(), object->GetSizeInBytes(), 1);
     }
 
     /// Unregisters an object from the cache
     void Unregister(const T& object) {
         object->SetIsRegistered(false);
-        auto& rasterizer = Core::System::GetInstance().Renderer().Rasterizer();
         rasterizer.UpdatePagesCachedCount(object->GetAddr(), object->GetSizeInBytes(), -1);
 
         // Only flush if use_accurate_gpu_emulation is enabled, as it incurs a performance hit
@@ -177,4 +177,5 @@ private:
 
     ObjectCache object_cache; ///< Cache of objects
     u64 modified_ticks{};     ///< Counter of cache state ticks, used for in-order flushing
+    VideoCore::RasterizerInterface& rasterizer;
 };

src/video_core/renderer_opengl/gl_buffer_cache.cpp

@@ -9,10 +9,12 @@
 #include "core/core.h"
 #include "core/memory.h"
 #include "video_core/renderer_opengl/gl_buffer_cache.h"
+#include "video_core/renderer_opengl/gl_rasterizer.h"
 
 namespace OpenGL {
 
-OGLBufferCache::OGLBufferCache(std::size_t size) : stream_buffer(GL_ARRAY_BUFFER, size) {}
+OGLBufferCache::OGLBufferCache(RasterizerOpenGL& rasterizer, std::size_t size)
+    : RasterizerCache{rasterizer}, stream_buffer(GL_ARRAY_BUFFER, size) {}
 
 GLintptr OGLBufferCache::UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size,
                                       std::size_t alignment, bool cache) {

src/video_core/renderer_opengl/gl_buffer_cache.h

@@ -15,6 +15,8 @@
 
 namespace OpenGL {
 
+class RasterizerOpenGL;
+
 struct CachedBufferEntry final : public RasterizerCacheObject {
     VAddr GetAddr() const override {
         return addr;
@@ -35,7 +37,7 @@ struct CachedBufferEntry final : public RasterizerCacheObject {
 
 class OGLBufferCache final : public RasterizerCache<std::shared_ptr<CachedBufferEntry>> {
 public:
-    explicit OGLBufferCache(std::size_t size);
+    explicit OGLBufferCache(RasterizerOpenGL& rasterizer, std::size_t size);
 
     /// Uploads data from a guest GPU address. Returns host's buffer offset where it's been
     /// allocated.

src/video_core/renderer_opengl/gl_primitive_assembler.cpp

@@ -6,6 +6,7 @@
 #include <array>
 #include "common/assert.h"
 #include "common/common_types.h"
+#include "core/core.h"
 #include "core/memory.h"
 #include "video_core/renderer_opengl/gl_buffer_cache.h"
 #include "video_core/renderer_opengl/gl_primitive_assembler.h"
@@ -46,12 +47,13 @@ GLintptr PrimitiveAssembler::MakeQuadIndexed(Tegra::GPUVAddr gpu_addr, std::size
 
     auto& memory_manager = Core::System::GetInstance().GPU().MemoryManager();
     const std::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)};
-    const u8* source{Memory::GetPointer(*cpu_addr)};
+    index_cache.resize(count);
+    Memory::ReadBlock(*cpu_addr, index_cache.data(), count);
 
     for (u32 primitive = 0; primitive < count / 4; ++primitive) {
         for (std::size_t i = 0; i < TRIANGLES_PER_QUAD; ++i) {
             const u32 index = primitive * 4 + QUAD_MAP[i];
-            const u8* src_offset = source + (index * index_size);
+            const u8* src_offset = index_cache.data() + (index * index_size);
 
             std::memcpy(dst_pointer, src_offset, index_size);
             dst_pointer += index_size;
@@ -61,4 +63,4 @@ GLintptr PrimitiveAssembler::MakeQuadIndexed(Tegra::GPUVAddr gpu_addr, std::size
     return index_offset;
 }
 
-} // namespace OpenGL
+} // namespace OpenGL

src/video_core/renderer_opengl/gl_primitive_assembler.h

@@ -28,6 +28,7 @@ public:
 
 private:
     OGLBufferCache& buffer_cache;
+    std::vector<u8> index_cache;
 };
 
-} // namespace OpenGL
+} // namespace OpenGL

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -30,10 +30,11 @@
 namespace OpenGL {
 
 using Maxwell = Tegra::Engines::Maxwell3D::Regs;
-using PixelFormat = SurfaceParams::PixelFormat;
-using SurfaceType = SurfaceParams::SurfaceType;
+using PixelFormat = VideoCore::Surface::PixelFormat;
+using SurfaceType = VideoCore::Surface::SurfaceType;
 
-MICROPROFILE_DEFINE(OpenGL_VAO, "OpenGL", "Vertex Array Setup", MP_RGB(128, 128, 192));
+MICROPROFILE_DEFINE(OpenGL_VAO, "OpenGL", "Vertex Format Setup", MP_RGB(128, 128, 192));
+MICROPROFILE_DEFINE(OpenGL_VB, "OpenGL", "Vertex Buffer Setup", MP_RGB(128, 128, 192));
 MICROPROFILE_DEFINE(OpenGL_Shader, "OpenGL", "Shader Setup", MP_RGB(128, 128, 192));
 MICROPROFILE_DEFINE(OpenGL_UBO, "OpenGL", "Const Buffer Setup", MP_RGB(128, 128, 192));
 MICROPROFILE_DEFINE(OpenGL_Index, "OpenGL", "Index Buffer Setup", MP_RGB(128, 128, 192));
@@ -79,7 +80,8 @@ struct DrawParameters {
 };
 
 RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo& info)
-    : emu_window{window}, screen_info{info}, buffer_cache(STREAM_BUFFER_SIZE) {
+    : res_cache{*this}, shader_cache{*this}, emu_window{window}, screen_info{info},
+      buffer_cache(*this, STREAM_BUFFER_SIZE) {
     // Create sampler objects
     for (std::size_t i = 0; i < texture_samplers.size(); ++i) {
         texture_samplers[i].Create();
@@ -122,18 +124,23 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
 
 RasterizerOpenGL::~RasterizerOpenGL() {}
 
-void RasterizerOpenGL::SetupVertexArrays() {
-    MICROPROFILE_SCOPE(OpenGL_VAO);
-    const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+void RasterizerOpenGL::SetupVertexFormat() {
+    auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
     const auto& regs = gpu.regs;
 
+    if (!gpu.dirty_flags.vertex_attrib_format)
+        return;
+    gpu.dirty_flags.vertex_attrib_format = false;
+
+    MICROPROFILE_SCOPE(OpenGL_VAO);
+
     auto [iter, is_cache_miss] = vertex_array_cache.try_emplace(regs.vertex_attrib_format);
     auto& VAO = iter->second;
 
     if (is_cache_miss) {
         VAO.Create();
         state.draw.vertex_array = VAO.handle;
-        state.Apply();
+        state.ApplyVertexBufferState();
 
         // The index buffer binding is stored within the VAO. Stupid OpenGL, but easy to work
         // around.
@@ -175,8 +182,13 @@ void RasterizerOpenGL::SetupVertexArrays() {
         }
     }
     state.draw.vertex_array = VAO.handle;
-    state.draw.vertex_buffer = buffer_cache.GetHandle();
-    state.Apply();
+    state.ApplyVertexBufferState();
+}
+
+void RasterizerOpenGL::SetupVertexBuffer() {
+    MICROPROFILE_SCOPE(OpenGL_VB);
+    const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+    const auto& regs = gpu.regs;
 
     // Upload all guest vertex arrays sequentially to our buffer
     for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) {
@@ -203,6 +215,9 @@ void RasterizerOpenGL::SetupVertexArrays() {
             glVertexBindingDivisor(index, 0);
         }
     }
+
+    // Implicit set by glBindVertexBuffer. Stupid glstate handling...
+    state.draw.vertex_buffer = buffer_cache.GetHandle();
 }
 
 DrawParameters RasterizerOpenGL::SetupDraw() {
@@ -327,8 +342,6 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) {
             index++;
         }
     }
-
-    state.Apply();
 }
 
 std::size_t RasterizerOpenGL::CalculateVertexArraysSize() const {
@@ -397,8 +410,8 @@ void RasterizerOpenGL::UpdatePagesCachedCount(VAddr addr, u64 size, int delta) {
         cached_pages.add({pages_interval, delta});
 }
 
-void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_depth_fb,
-                                             bool preserve_contents,
+void RasterizerOpenGL::ConfigureFramebuffers(OpenGLState& current_state, bool using_color_fb,
+                                             bool using_depth_fb, bool preserve_contents,
                                              std::optional<std::size_t> single_color_target) {
     MICROPROFILE_SCOPE(OpenGL_Framebuffer);
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
@@ -414,9 +427,9 @@ void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_dep
     ASSERT_MSG(regs.rt_separate_frag_data == 0, "Unimplemented");
 
     // Bind the framebuffer surfaces
-    state.draw.draw_framebuffer = framebuffer.handle;
-    state.Apply();
-    state.framebuffer_srgb.enabled = regs.framebuffer_srgb != 0;
+    current_state.draw.draw_framebuffer = framebuffer.handle;
+    current_state.ApplyFramebufferState();
+    current_state.framebuffer_srgb.enabled = regs.framebuffer_srgb != 0;
 
     if (using_color_fb) {
         if (single_color_target) {
@@ -494,10 +507,7 @@ void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_dep
         glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
                                0);
     }
-
-    SyncViewport();
-
-    state.Apply();
+    SyncViewport(current_state);
 }
 
 void RasterizerOpenGL::Clear() {
@@ -510,22 +520,23 @@ void RasterizerOpenGL::Clear() {
     bool use_stencil{};
 
     OpenGLState clear_state;
-    clear_state.draw.draw_framebuffer = framebuffer.handle;
-    clear_state.color_mask.red_enabled = regs.clear_buffers.R ? GL_TRUE : GL_FALSE;
-    clear_state.color_mask.green_enabled = regs.clear_buffers.G ? GL_TRUE : GL_FALSE;
-    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;
-
     if (regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B ||
         regs.clear_buffers.A) {
         use_color = true;
     }
+    if (use_color) {
+        clear_state.color_mask[0].red_enabled = regs.clear_buffers.R ? GL_TRUE : GL_FALSE;
+        clear_state.color_mask[0].green_enabled = regs.clear_buffers.G ? GL_TRUE : GL_FALSE;
+        clear_state.color_mask[0].blue_enabled = regs.clear_buffers.B ? GL_TRUE : GL_FALSE;
+        clear_state.color_mask[0].alpha_enabled = regs.clear_buffers.A ? GL_TRUE : GL_FALSE;
+    }
     if (regs.clear_buffers.Z) {
         ASSERT_MSG(regs.zeta_enable != 0, "Tried to clear Z but buffer is not enabled!");
         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.
+        // ignored when clearing the buffer in the Switch, but OpenGL obeys it so we set it to
+        // true.
         clear_state.depth.test_enabled = true;
         clear_state.depth.test_func = GL_ALWAYS;
     }
@@ -542,11 +553,8 @@ void RasterizerOpenGL::Clear() {
 
     ScopeAcquireGLContext acquire_context{emu_window};
 
-    ConfigureFramebuffers(use_color, use_depth || use_stencil, false,
+    ConfigureFramebuffers(clear_state, use_color, use_depth || use_stencil, false,
                           regs.clear_buffers.RT.Value());
-    // Copy the sRGB setting to the clear state to avoid problem with
-    // specific driver implementations
-    clear_state.framebuffer_srgb.enabled = state.framebuffer_srgb.enabled;
     clear_state.Apply();
 
     if (use_color) {
@@ -572,15 +580,14 @@ void RasterizerOpenGL::DrawArrays() {
 
     ScopeAcquireGLContext acquire_context{emu_window};
 
-    ConfigureFramebuffers();
-
+    ConfigureFramebuffers(state);
+    SyncColorMask();
     SyncDepthTestState();
     SyncStencilTestState();
     SyncBlendState();
     SyncLogicOpState();
     SyncCullMode();
     SyncPrimitiveRestart();
-    SyncDepthRange();
     SyncScissorTest();
     // Alpha Testing is synced on shaders.
     SyncTransformFeedback();
@@ -594,7 +601,7 @@ void RasterizerOpenGL::DrawArrays() {
     const bool is_indexed = accelerate_draw == AccelDraw::Indexed;
 
     state.draw.vertex_buffer = buffer_cache.GetHandle();
-    state.Apply();
+    state.ApplyVertexBufferState();
 
     std::size_t buffer_size = CalculateVertexArraysSize();
 
@@ -621,7 +628,8 @@ void RasterizerOpenGL::DrawArrays() {
 
     buffer_cache.Map(buffer_size);
 
-    SetupVertexArrays();
+    SetupVertexFormat();
+    SetupVertexBuffer();
     DrawParameters params = SetupDraw();
     SetupShaders(params.primitive_mode);
 
@@ -701,7 +709,8 @@ bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config,
 
     // Verify that the cached surface is the same size and format as the requested framebuffer
     const auto& params{surface->GetSurfaceParams()};
-    const auto& pixel_format{SurfaceParams::PixelFormatFromGPUPixelFormat(config.pixel_format)};
+    const auto& pixel_format{
+        VideoCore::Surface::PixelFormatFromGPUPixelFormat(config.pixel_format)};
     ASSERT_MSG(params.width == config.width, "Framebuffer width is different");
     ASSERT_MSG(params.height == config.height, "Framebuffer height is different");
     ASSERT_MSG(params.pixel_format == pixel_format, "Framebuffer pixel_format is different");
@@ -724,9 +733,9 @@ void RasterizerOpenGL::SamplerInfo::Create() {
     glSamplerParameteri(sampler.handle, GL_TEXTURE_COMPARE_FUNC, GL_NEVER);
 }
 
-void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntry& config) {
+void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::FullTextureInfo& info) {
     const GLuint s = sampler.handle;
-
+    const Tegra::Texture::TSCEntry& config = info.tsc;
     if (mag_filter != config.mag_filter) {
         mag_filter = config.mag_filter;
         glSamplerParameteri(
@@ -777,6 +786,22 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr
             glSamplerParameterfv(s, GL_TEXTURE_BORDER_COLOR, border_color.data());
         }
     }
+    if (info.tic.use_header_opt_control == 0) {
+        if (GLAD_GL_ARB_texture_filter_anisotropic) {
+            glSamplerParameterf(s, GL_TEXTURE_MAX_ANISOTROPY,
+                                static_cast<float>(1 << info.tic.max_anisotropy.Value()));
+        } else if (GLAD_GL_EXT_texture_filter_anisotropic) {
+            glSamplerParameterf(s, GL_TEXTURE_MAX_ANISOTROPY_EXT,
+                                static_cast<float>(1 << info.tic.max_anisotropy.Value()));
+        }
+        glSamplerParameterf(s, GL_TEXTURE_MIN_LOD,
+                            static_cast<float>(info.tic.res_min_mip_level.Value()));
+        glSamplerParameterf(s, GL_TEXTURE_MAX_LOD,
+                            static_cast<float>(info.tic.res_max_mip_level.Value() == 0
+                                                   ? 16
+                                                   : info.tic.res_max_mip_level.Value()));
+        glSamplerParameterf(s, GL_TEXTURE_LOD_BIAS, info.tic.mip_lod_bias.Value() / 256.f);
+    }
 }
 
 u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, Shader& shader,
@@ -874,7 +899,7 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
             continue;
         }
 
-        texture_samplers[current_bindpoint].SyncWithConfig(texture.tsc);
+        texture_samplers[current_bindpoint].SyncWithConfig(texture);
         Surface surface = res_cache.GetTextureSurface(texture, entry);
         if (surface != nullptr) {
             state.texture_units[current_bindpoint].texture = surface->Texture().handle;
@@ -896,14 +921,18 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
     return current_unit + static_cast<u32>(entries.size());
 }
 
-void RasterizerOpenGL::SyncViewport() {
+void RasterizerOpenGL::SyncViewport(OpenGLState& current_state) {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
-    const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()};
-
-    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());
+    for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
+        const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()};
+        auto& viewport = current_state.viewports[i];
+        viewport.x = viewport_rect.left;
+        viewport.y = viewport_rect.bottom;
+        viewport.width = static_cast<GLfloat>(viewport_rect.GetWidth());
+        viewport.height = static_cast<GLfloat>(viewport_rect.GetHeight());
+        viewport.depth_range_far = regs.viewport[i].depth_range_far;
+        viewport.depth_range_near = regs.viewport[i].depth_range_near;
+    }
 }
 
 void RasterizerOpenGL::SyncClipEnabled() {
@@ -945,13 +974,6 @@ void RasterizerOpenGL::SyncPrimitiveRestart() {
     state.primitive_restart.index = regs.primitive_restart.index;
 }
 
-void RasterizerOpenGL::SyncDepthRange() {
-    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
-
-    state.depth.depth_range_near = regs.viewport->depth_range_near;
-    state.depth.depth_range_far = regs.viewport->depth_range_far;
-}
-
 void RasterizerOpenGL::SyncDepthTestState() {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
@@ -972,9 +994,6 @@ void RasterizerOpenGL::SyncStencilTestState() {
         return;
     }
 
-    // TODO(bunnei): Verify behavior when this is not set
-    ASSERT(regs.stencil_two_side_enable);
-
     state.stencil.front.test_func = MaxwellToGL::ComparisonOp(regs.stencil_front_func_func);
     state.stencil.front.test_ref = regs.stencil_front_func_ref;
     state.stencil.front.test_mask = regs.stencil_front_func_mask;
@@ -982,36 +1001,79 @@ void RasterizerOpenGL::SyncStencilTestState() {
     state.stencil.front.action_depth_fail = MaxwellToGL::StencilOp(regs.stencil_front_op_zfail);
     state.stencil.front.action_depth_pass = MaxwellToGL::StencilOp(regs.stencil_front_op_zpass);
     state.stencil.front.write_mask = regs.stencil_front_mask;
+    if (regs.stencil_two_side_enable) {
+        state.stencil.back.test_func = MaxwellToGL::ComparisonOp(regs.stencil_back_func_func);
+        state.stencil.back.test_ref = regs.stencil_back_func_ref;
+        state.stencil.back.test_mask = regs.stencil_back_func_mask;
+        state.stencil.back.action_stencil_fail = MaxwellToGL::StencilOp(regs.stencil_back_op_fail);
+        state.stencil.back.action_depth_fail = MaxwellToGL::StencilOp(regs.stencil_back_op_zfail);
+        state.stencil.back.action_depth_pass = MaxwellToGL::StencilOp(regs.stencil_back_op_zpass);
+        state.stencil.back.write_mask = regs.stencil_back_mask;
+    } else {
+        state.stencil.back.test_func = GL_ALWAYS;
+        state.stencil.back.test_ref = 0;
+        state.stencil.back.test_mask = 0xFFFFFFFF;
+        state.stencil.back.write_mask = 0xFFFFFFFF;
+        state.stencil.back.action_stencil_fail = GL_KEEP;
+        state.stencil.back.action_depth_fail = GL_KEEP;
+        state.stencil.back.action_depth_pass = GL_KEEP;
+    }
+}
 
-    state.stencil.back.test_func = MaxwellToGL::ComparisonOp(regs.stencil_back_func_func);
-    state.stencil.back.test_ref = regs.stencil_back_func_ref;
-    state.stencil.back.test_mask = regs.stencil_back_func_mask;
-    state.stencil.back.action_stencil_fail = MaxwellToGL::StencilOp(regs.stencil_back_op_fail);
-    state.stencil.back.action_depth_fail = MaxwellToGL::StencilOp(regs.stencil_back_op_zfail);
-    state.stencil.back.action_depth_pass = MaxwellToGL::StencilOp(regs.stencil_back_op_zpass);
-    state.stencil.back.write_mask = regs.stencil_back_mask;
+void RasterizerOpenGL::SyncColorMask() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+    for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
+        const auto& source = regs.color_mask[regs.color_mask_common ? 0 : i];
+        auto& dest = state.color_mask[i];
+        dest.red_enabled = (source.R == 0) ? GL_FALSE : GL_TRUE;
+        dest.green_enabled = (source.G == 0) ? GL_FALSE : GL_TRUE;
+        dest.blue_enabled = (source.B == 0) ? GL_FALSE : GL_TRUE;
+        dest.alpha_enabled = (source.A == 0) ? GL_FALSE : GL_TRUE;
+    }
 }
 
 void RasterizerOpenGL::SyncBlendState() {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
-    // TODO(Subv): Support more than just render target 0.
-    state.blend.enabled = regs.blend.enable[0] != 0;
+    state.blend_color.red = regs.blend_color.r;
+    state.blend_color.green = regs.blend_color.g;
+    state.blend_color.blue = regs.blend_color.b;
+    state.blend_color.alpha = regs.blend_color.a;
 
-    if (!state.blend.enabled)
+    state.independant_blend.enabled = regs.independent_blend_enable;
+    if (!state.independant_blend.enabled) {
+        auto& blend = state.blend[0];
+        blend.enabled = regs.blend.enable[0] != 0;
+        blend.separate_alpha = regs.blend.separate_alpha;
+        blend.rgb_equation = MaxwellToGL::BlendEquation(regs.blend.equation_rgb);
+        blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_rgb);
+        blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_rgb);
+        if (blend.separate_alpha) {
+            blend.a_equation = MaxwellToGL::BlendEquation(regs.blend.equation_a);
+            blend.src_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_a);
+            blend.dst_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_a);
+        }
+        for (size_t i = 1; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
+            state.blend[i].enabled = false;
+        }
         return;
+    }
 
-    ASSERT_MSG(regs.logic_op.enable == 0,
-               "Blending and logic op can't be enabled at the same time.");
-
-    ASSERT_MSG(regs.independent_blend_enable == 1, "Only independent blending is implemented");
-    ASSERT_MSG(!regs.independent_blend[0].separate_alpha, "Unimplemented");
-    state.blend.rgb_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_rgb);
-    state.blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_rgb);
-    state.blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_rgb);
-    state.blend.a_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_a);
-    state.blend.src_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_a);
-    state.blend.dst_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_a);
+    for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
+        auto& blend = state.blend[i];
+        blend.enabled = regs.blend.enable[i] != 0;
+        if (!blend.enabled)
+            continue;
+        blend.separate_alpha = regs.independent_blend[i].separate_alpha;
+        blend.rgb_equation = MaxwellToGL::BlendEquation(regs.independent_blend[i].equation_rgb);
+        blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_source_rgb);
+        blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_dest_rgb);
+        if (blend.separate_alpha) {
+            blend.a_equation = MaxwellToGL::BlendEquation(regs.independent_blend[i].equation_a);
+            blend.src_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_source_a);
+            blend.dst_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_dest_a);
+        }
+    }
 }
 
 void RasterizerOpenGL::SyncLogicOpState() {
@@ -1030,19 +1092,19 @@ void RasterizerOpenGL::SyncLogicOpState() {
 }
 
 void RasterizerOpenGL::SyncScissorTest() {
+    // TODO: what is the correct behavior here, a single scissor for all targets
+    // or scissor disabled for the rest of the targets?
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
-
     state.scissor.enabled = (regs.scissor_test.enable != 0);
-    // TODO(Blinkhawk): Figure if the hardware supports scissor testing per viewport and how it's
-    // implemented.
-    if (regs.scissor_test.enable != 0) {
-        const u32 width = regs.scissor_test.max_x - regs.scissor_test.min_x;
-        const u32 height = regs.scissor_test.max_y - regs.scissor_test.min_y;
-        state.scissor.x = regs.scissor_test.min_x;
-        state.scissor.y = regs.scissor_test.min_y;
-        state.scissor.width = width;
-        state.scissor.height = height;
+    if (regs.scissor_test.enable == 0) {
+        return;
     }
+    const u32 width = regs.scissor_test.max_x - regs.scissor_test.min_x;
+    const u32 height = regs.scissor_test.max_y - regs.scissor_test.min_y;
+    state.scissor.x = regs.scissor_test.min_x;
+    state.scissor.y = regs.scissor_test.min_y;
+    state.scissor.width = width;
+    state.scissor.height = height;
 }
 
 void RasterizerOpenGL::SyncTransformFeedback() {
@@ -1067,9 +1129,8 @@ void RasterizerOpenGL::CheckAlphaTests() {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
     if (regs.alpha_test_enabled != 0 && regs.rt_control.count > 1) {
-        LOG_CRITICAL(
-            Render_OpenGL,
-            "Alpha Testing is enabled with Multiple Render Targets, this behavior is undefined.");
+        LOG_CRITICAL(Render_OpenGL, "Alpha Testing is enabled with Multiple Render Targets, "
+                                    "this behavior is undefined.");
         UNREACHABLE();
     }
 }

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -88,7 +88,7 @@ private:
         /// SamplerInfo struct.
         void Create();
         /// Syncs the sampler object with the config, updating any necessary state.
-        void SyncWithConfig(const Tegra::Texture::TSCEntry& config);
+        void SyncWithConfig(const Tegra::Texture::FullTextureInfo& info);
 
     private:
         Tegra::Texture::TextureFilter mag_filter;
@@ -109,8 +109,8 @@ private:
      * @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,
+    void ConfigureFramebuffers(OpenGLState& current_state, bool use_color_fb = true,
+                               bool using_depth_fb = true, bool preserve_contents = true,
                                std::optional<std::size_t> single_color_target = {});
 
     /*
@@ -133,8 +133,8 @@ private:
     u32 SetupTextures(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, Shader& shader,
                       GLenum primitive_mode, u32 current_unit);
 
-    /// Syncs the viewport to match the guest state
-    void SyncViewport();
+    /// Syncs the viewport and depth range to match the guest state
+    void SyncViewport(OpenGLState& current_state);
 
     /// Syncs the clip enabled status to match the guest state
     void SyncClipEnabled();
@@ -148,9 +148,6 @@ private:
     /// Syncs the primitve restart to match the guest state
     void SyncPrimitiveRestart();
 
-    /// Syncs the depth range to match the guest state
-    void SyncDepthRange();
-
     /// Syncs the depth test state to match the guest state
     void SyncDepthTestState();
 
@@ -172,6 +169,9 @@ private:
     /// Syncs the point state to match the guest state
     void SyncPointState();
 
+    /// Syncs Color Mask
+    void SyncColorMask();
+
     /// Check asserts for alpha testing.
     void CheckAlphaTests();
 
@@ -207,7 +207,8 @@ private:
 
     std::size_t CalculateIndexBufferSize() const;
 
-    void SetupVertexArrays();
+    void SetupVertexFormat();
+    void SetupVertexBuffer();
 
     DrawParameters SetupDraw();
 

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -15,17 +15,24 @@
 #include "core/memory.h"
 #include "core/settings.h"
 #include "video_core/engines/maxwell_3d.h"
+#include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/gl_rasterizer_cache.h"
+#include "video_core/renderer_opengl/gl_state.h"
 #include "video_core/renderer_opengl/utils.h"
+#include "video_core/surface.h"
 #include "video_core/textures/astc.h"
 #include "video_core/textures/decoders.h"
 #include "video_core/utils.h"
 
 namespace OpenGL {
 
-using SurfaceType = SurfaceParams::SurfaceType;
-using PixelFormat = SurfaceParams::PixelFormat;
-using ComponentType = SurfaceParams::ComponentType;
+using VideoCore::Surface::ComponentTypeFromDepthFormat;
+using VideoCore::Surface::ComponentTypeFromRenderTarget;
+using VideoCore::Surface::ComponentTypeFromTexture;
+using VideoCore::Surface::PixelFormatFromDepthFormat;
+using VideoCore::Surface::PixelFormatFromRenderTargetFormat;
+using VideoCore::Surface::PixelFormatFromTextureFormat;
+using VideoCore::Surface::SurfaceTargetFromTextureType;
 
 struct FormatTuple {
     GLint internal_format;
@@ -35,46 +42,6 @@ struct FormatTuple {
     bool compressed;
 };
 
-static bool IsPixelFormatASTC(PixelFormat format) {
-    switch (format) {
-    case PixelFormat::ASTC_2D_4X4:
-    case PixelFormat::ASTC_2D_5X4:
-    case PixelFormat::ASTC_2D_8X8:
-    case PixelFormat::ASTC_2D_8X5:
-    case PixelFormat::ASTC_2D_4X4_SRGB:
-    case PixelFormat::ASTC_2D_5X4_SRGB:
-    case PixelFormat::ASTC_2D_8X8_SRGB:
-    case PixelFormat::ASTC_2D_8X5_SRGB:
-        return true;
-    default:
-        return false;
-    }
-}
-
-static std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) {
-    switch (format) {
-    case PixelFormat::ASTC_2D_4X4:
-        return {4, 4};
-    case PixelFormat::ASTC_2D_5X4:
-        return {5, 4};
-    case PixelFormat::ASTC_2D_8X8:
-        return {8, 8};
-    case PixelFormat::ASTC_2D_8X5:
-        return {8, 5};
-    case PixelFormat::ASTC_2D_4X4_SRGB:
-        return {4, 4};
-    case PixelFormat::ASTC_2D_5X4_SRGB:
-        return {5, 4};
-    case PixelFormat::ASTC_2D_8X8_SRGB:
-        return {8, 8};
-    case PixelFormat::ASTC_2D_8X5_SRGB:
-        return {8, 5};
-    default:
-        LOG_CRITICAL(HW_GPU, "Unhandled format: {}", static_cast<u32>(format));
-        UNREACHABLE();
-    }
-}
-
 void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
     auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
     const auto cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr_)};
@@ -93,16 +60,14 @@ void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
 
 std::size_t SurfaceParams::InnerMipmapMemorySize(u32 mip_level, bool force_gl, bool layer_only,
                                                  bool uncompressed) const {
-    const u32 compression_factor{GetCompressionFactor(pixel_format)};
+    const u32 tile_x{GetDefaultBlockWidth(pixel_format)};
+    const u32 tile_y{GetDefaultBlockHeight(pixel_format)};
     const u32 bytes_per_pixel{GetBytesPerPixel(pixel_format)};
     u32 m_depth = (layer_only ? 1U : depth);
     u32 m_width = MipWidth(mip_level);
     u32 m_height = MipHeight(mip_level);
-    m_width = uncompressed ? m_width
-                           : std::max(1U, (m_width + compression_factor - 1) / compression_factor);
-    m_height = uncompressed
-                   ? m_height
-                   : std::max(1U, (m_height + compression_factor - 1) / compression_factor);
+    m_width = uncompressed ? m_width : std::max(1U, (m_width + tile_x - 1) / tile_x);
+    m_height = uncompressed ? m_height : std::max(1U, (m_height + tile_y - 1) / tile_y);
     m_depth = std::max(1U, m_depth >> mip_level);
     u32 m_block_height = MipBlockHeight(mip_level);
     u32 m_block_depth = MipBlockDepth(mip_level);
@@ -163,6 +128,13 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
             params.target = SurfaceTarget::Texture2D;
         }
         break;
+    case SurfaceTarget::TextureCubeArray:
+        params.depth = config.tic.Depth() * 6;
+        if (!entry.IsArray()) {
+            ASSERT(params.depth == 6);
+            params.target = SurfaceTarget::TextureCubemap;
+        }
+        break;
     default:
         LOG_CRITICAL(HW_GPU, "Unknown depth for target={}", static_cast<u32>(params.target));
         UNREACHABLE();
@@ -267,7 +239,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
     return params;
 }
 
-static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{
+static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format_tuples = {{
     {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // ABGR8U
     {GL_RGBA8, GL_RGBA, GL_BYTE, ComponentType::SNorm, false},                     // ABGR8S
     {GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE, ComponentType::UInt, false},   // ABGR8UI
@@ -340,6 +312,10 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
     {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X8_SRGB
     {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X5_SRGB
     {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X4_SRGB
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},        // ASTC_2D_5X5
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X5_SRGB
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},        // ASTC_2D_10X8
+    {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_10X8_SRGB
 
     // Depth formats
     {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F
@@ -355,20 +331,22 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
      ComponentType::Float, false}, // Z32FS8
 }};
 
-static GLenum SurfaceTargetToGL(SurfaceParams::SurfaceTarget target) {
+static GLenum SurfaceTargetToGL(SurfaceTarget target) {
     switch (target) {
-    case SurfaceParams::SurfaceTarget::Texture1D:
+    case SurfaceTarget::Texture1D:
         return GL_TEXTURE_1D;
-    case SurfaceParams::SurfaceTarget::Texture2D:
+    case SurfaceTarget::Texture2D:
         return GL_TEXTURE_2D;
-    case SurfaceParams::SurfaceTarget::Texture3D:
+    case SurfaceTarget::Texture3D:
         return GL_TEXTURE_3D;
-    case SurfaceParams::SurfaceTarget::Texture1DArray:
+    case SurfaceTarget::Texture1DArray:
         return GL_TEXTURE_1D_ARRAY;
-    case SurfaceParams::SurfaceTarget::Texture2DArray:
+    case SurfaceTarget::Texture2DArray:
         return GL_TEXTURE_2D_ARRAY;
-    case SurfaceParams::SurfaceTarget::TextureCubemap:
+    case SurfaceTarget::TextureCubemap:
         return GL_TEXTURE_CUBE_MAP;
+    case SurfaceTarget::TextureCubeArray:
+        return GL_TEXTURE_CUBE_MAP_ARRAY_ARB;
     }
     LOG_CRITICAL(Render_OpenGL, "Unimplemented texture target={}", static_cast<u32>(target));
     UNREACHABLE();
@@ -392,50 +370,32 @@ MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
     return {0, actual_height, MipWidth(mip_level), 0};
 }
 
-/// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN
-static bool IsFormatBCn(PixelFormat format) {
-    switch (format) {
-    case PixelFormat::DXT1:
-    case PixelFormat::DXT23:
-    case PixelFormat::DXT45:
-    case PixelFormat::DXN1:
-    case PixelFormat::DXN2SNORM:
-    case PixelFormat::DXN2UNORM:
-    case PixelFormat::BC7U:
-    case PixelFormat::BC6H_UF16:
-    case PixelFormat::BC6H_SF16:
-    case PixelFormat::DXT1_SRGB:
-    case PixelFormat::DXT23_SRGB:
-    case PixelFormat::DXT45_SRGB:
-    case PixelFormat::BC7U_SRGB:
-        return true;
-    }
-    return false;
-}
-
 template <bool morton_to_gl, PixelFormat format>
 void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth, u8* gl_buffer,
                 std::size_t gl_buffer_size, VAddr addr) {
-    constexpr u32 bytes_per_pixel = SurfaceParams::GetBytesPerPixel(format);
+    constexpr u32 bytes_per_pixel = GetBytesPerPixel(format);
 
     // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
     // pixel values.
-    const u32 tile_size{IsFormatBCn(format) ? 4U : 1U};
+    const u32 tile_size_x{GetDefaultBlockWidth(format)};
+    const u32 tile_size_y{GetDefaultBlockHeight(format)};
 
     if (morton_to_gl) {
-        const std::vector<u8> data = Tegra::Texture::UnswizzleTexture(
-            addr, tile_size, bytes_per_pixel, stride, height, depth, block_height, block_depth);
+        const std::vector<u8> data =
+            Tegra::Texture::UnswizzleTexture(addr, tile_size_x, tile_size_y, bytes_per_pixel,
+                                             stride, height, depth, block_height, block_depth);
         const std::size_t size_to_copy{std::min(gl_buffer_size, data.size())};
         memcpy(gl_buffer, data.data(), size_to_copy);
     } else {
-        Tegra::Texture::CopySwizzledData(stride / tile_size, height / tile_size, depth,
+        Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x,
+                                         (height + tile_size_y - 1) / tile_size_y, depth,
                                          bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr),
                                          gl_buffer, false, block_height, block_depth);
     }
 }
 
 using GLConversionArray = std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
-                                     SurfaceParams::MaxPixelFormat>;
+                                     VideoCore::Surface::MaxPixelFormat>;
 
 static constexpr GLConversionArray morton_to_gl_fns = {
     // clang-format off
@@ -496,6 +456,10 @@ static constexpr GLConversionArray morton_to_gl_fns = {
         MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>,
         MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>,
         MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X5>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_10X8>,
+        MortonCopy<true, PixelFormat::ASTC_2D_10X8_SRGB>,
         MortonCopy<true, PixelFormat::Z32F>,
         MortonCopy<true, PixelFormat::Z16>,
         MortonCopy<true, PixelFormat::Z24S8>,
@@ -564,6 +528,10 @@ static constexpr GLConversionArray gl_to_morton_fns = {
         nullptr,
         nullptr,
         nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
         MortonCopy<false, PixelFormat::Z32F>,
         MortonCopy<false, PixelFormat::Z16>,
         MortonCopy<false, PixelFormat::Z24S8>,
@@ -575,7 +543,7 @@ static constexpr GLConversionArray gl_to_morton_fns = {
 void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params,
                  std::vector<u8>& gl_buffer, u32 mip_level) {
     u32 depth = params.MipDepth(mip_level);
-    if (params.target == SurfaceParams::SurfaceTarget::Texture2D) {
+    if (params.target == SurfaceTarget::Texture2D) {
         // TODO(Blinkhawk): Eliminate this condition once all texture types are implemented.
         depth = 1U;
     }
@@ -601,9 +569,11 @@ void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params
     }
 }
 
+MICROPROFILE_DEFINE(OpenGL_BlitSurface, "OpenGL", "BlitSurface", MP_RGB(128, 192, 64));
 static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
                         GLuint read_fb_handle, GLuint draw_fb_handle, GLenum src_attachment = 0,
                         GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
+    MICROPROFILE_SCOPE(OpenGL_BlitSurface);
 
     const auto& src_params{src_surface->GetSurfaceParams()};
     const auto& dst_params{dst_surface->GetSurfaceParams()};
@@ -616,19 +586,19 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
     state.draw.draw_framebuffer = draw_fb_handle;
     // Set sRGB enabled if the destination surfaces need it
     state.framebuffer_srgb.enabled = dst_params.srgb_conversion;
-    state.Apply();
+    state.ApplyFramebufferState();
 
     u32 buffers{};
 
     if (src_params.type == SurfaceType::ColorTexture) {
         switch (src_params.target) {
-        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture2D:
             glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
                                    GL_TEXTURE_2D, src_surface->Texture().handle, 0);
             glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
                                    0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::TextureCubemap:
             glFramebufferTexture2D(
                 GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
                 static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
@@ -637,12 +607,12 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
                 GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
                 static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture2DArray:
             glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
                                       src_surface->Texture().handle, 0, 0);
             glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture3D:
             glFramebufferTexture3D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
                                    SurfaceTargetToGL(src_params.target),
                                    src_surface->Texture().handle, 0, 0);
@@ -658,13 +628,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
         }
 
         switch (dst_params.target) {
-        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture2D:
             glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
                                    GL_TEXTURE_2D, dst_surface->Texture().handle, 0);
             glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
                                    0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::TextureCubemap:
             glFramebufferTexture2D(
                 GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
                 static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
@@ -673,13 +643,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
                 GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
                 static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture2DArray:
             glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
                                       dst_surface->Texture().handle, 0, 0);
             glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
             break;
 
-        case SurfaceParams::SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture3D:
             glFramebufferTexture3D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
                                    SurfaceTargetToGL(dst_params.target),
                                    dst_surface->Texture().handle, 0, 0);
@@ -743,9 +713,11 @@ static void FastCopySurface(const Surface& src_surface, const Surface& dst_surfa
                        0, 0, width, height, 1);
 }
 
+MICROPROFILE_DEFINE(OpenGL_CopySurface, "OpenGL", "CopySurface", MP_RGB(128, 192, 64));
 static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
                         GLuint copy_pbo_handle, GLenum src_attachment = 0,
                         GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
+    MICROPROFILE_SCOPE(OpenGL_CopySurface);
     ASSERT_MSG(dst_attachment == 0, "Unimplemented");
 
     const auto& src_params{src_surface->GetSurfaceParams()};
@@ -780,8 +752,7 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
         }
         std::size_t remaining_size = dst_params.size_in_bytes - src_params.size_in_bytes;
         std::vector<u8> data(remaining_size);
-        std::memcpy(data.data(), Memory::GetPointer(dst_params.addr + src_params.size_in_bytes),
-                    data.size());
+        Memory::ReadBlock(dst_params.addr + src_params.size_in_bytes, data.data(), data.size());
 
         glBufferSubData(GL_PIXEL_PACK_BUFFER, src_params.size_in_bytes, remaining_size,
                         data.data());
@@ -800,21 +771,22 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
         UNREACHABLE();
     } else {
         switch (dst_params.target) {
-        case SurfaceParams::SurfaceTarget::Texture1D:
+        case SurfaceTarget::Texture1D:
             glTextureSubImage1D(dst_surface->Texture().handle, 0, 0, width, dest_format.format,
                                 dest_format.type, nullptr);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture2D:
             glTextureSubImage2D(dst_surface->Texture().handle, 0, 0, 0, width, height,
                                 dest_format.format, dest_format.type, nullptr);
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::TextureCubeArray:
             glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0, 0, width, height,
                                 static_cast<GLsizei>(dst_params.depth), dest_format.format,
                                 dest_format.type, nullptr);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::TextureCubemap:
             glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0,
                                 static_cast<GLint>(cubemap_face), width, height, 1,
                                 dest_format.format, dest_format.type, nullptr);
@@ -851,17 +823,18 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
     if (!format_tuple.compressed) {
         // Only pre-create the texture for non-compressed textures.
         switch (params.target) {
-        case SurfaceParams::SurfaceTarget::Texture1D:
+        case SurfaceTarget::Texture1D:
             glTexStorage1D(SurfaceTargetToGL(params.target), params.max_mip_level,
                            format_tuple.internal_format, rect.GetWidth());
             break;
-        case SurfaceParams::SurfaceTarget::Texture2D:
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
+        case SurfaceTarget::Texture2D:
+        case SurfaceTarget::TextureCubemap:
             glTexStorage2D(SurfaceTargetToGL(params.target), params.max_mip_level,
                            format_tuple.internal_format, rect.GetWidth(), rect.GetHeight());
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::TextureCubeArray:
             glTexStorage3D(SurfaceTargetToGL(params.target), params.max_mip_level,
                            format_tuple.internal_format, rect.GetWidth(), rect.GetHeight(),
                            params.depth);
@@ -916,7 +889,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height, bo
 
     S8Z24 s8z24_pixel{};
     Z24S8 z24s8_pixel{};
-    constexpr auto bpp{SurfaceParams::GetBytesPerPixel(PixelFormat::S8Z24)};
+    constexpr auto bpp{GetBytesPerPixel(PixelFormat::S8Z24)};
     for (std::size_t y = 0; y < height; ++y) {
         for (std::size_t x = 0; x < width; ++x) {
             const std::size_t offset{bpp * (y * width + x)};
@@ -936,7 +909,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height, bo
 }
 
 static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
-    constexpr auto bpp{SurfaceParams::GetBytesPerPixel(PixelFormat::G8R8U)};
+    constexpr auto bpp{GetBytesPerPixel(PixelFormat::G8R8U)};
     for (std::size_t y = 0; y < height; ++y) {
         for (std::size_t x = 0; x < width; ++x) {
             const std::size_t offset{bpp * (y * width + x)};
@@ -953,21 +926,26 @@ static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
  * typical desktop GPUs.
  */
 static void ConvertFormatAsNeeded_LoadGLBuffer(std::vector<u8>& data, PixelFormat pixel_format,
-                                               u32 width, u32 height) {
+                                               u32 width, u32 height, u32 depth) {
     switch (pixel_format) {
     case PixelFormat::ASTC_2D_4X4:
     case PixelFormat::ASTC_2D_8X8:
     case PixelFormat::ASTC_2D_8X5:
     case PixelFormat::ASTC_2D_5X4:
+    case PixelFormat::ASTC_2D_5X5:
     case PixelFormat::ASTC_2D_4X4_SRGB:
     case PixelFormat::ASTC_2D_8X8_SRGB:
     case PixelFormat::ASTC_2D_8X5_SRGB:
-    case PixelFormat::ASTC_2D_5X4_SRGB: {
+    case PixelFormat::ASTC_2D_5X4_SRGB:
+    case PixelFormat::ASTC_2D_5X5_SRGB:
+    case PixelFormat::ASTC_2D_10X8:
+    case PixelFormat::ASTC_2D_10X8_SRGB: {
         // Convert ASTC pixel formats to RGBA8, as most desktop GPUs do not support ASTC.
         u32 block_width{};
         u32 block_height{};
         std::tie(block_width, block_height) = GetASTCBlockSize(pixel_format);
-        data = Tegra::Texture::ASTC::Decompress(data, width, height, block_width, block_height);
+        data =
+            Tegra::Texture::ASTC::Decompress(data, width, height, depth, block_width, block_height);
         break;
     }
     case PixelFormat::S8Z24:
@@ -996,7 +974,11 @@ static void ConvertFormatAsNeeded_FlushGLBuffer(std::vector<u8>& data, PixelForm
     case PixelFormat::ASTC_2D_4X4:
     case PixelFormat::ASTC_2D_8X8:
     case PixelFormat::ASTC_2D_4X4_SRGB:
-    case PixelFormat::ASTC_2D_8X8_SRGB: {
+    case PixelFormat::ASTC_2D_8X8_SRGB:
+    case PixelFormat::ASTC_2D_5X5:
+    case PixelFormat::ASTC_2D_5X5_SRGB:
+    case PixelFormat::ASTC_2D_10X8:
+    case PixelFormat::ASTC_2D_10X8_SRGB: {
         LOG_CRITICAL(HW_GPU, "Conversion of format {} after texture flushing is not implemented",
                      static_cast<u32>(pixel_format));
         UNREACHABLE();
@@ -1009,7 +991,7 @@ static void ConvertFormatAsNeeded_FlushGLBuffer(std::vector<u8>& data, PixelForm
     }
 }
 
-MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 64, 192));
+MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 192, 64));
 void CachedSurface::LoadGLBuffer() {
     MICROPROFILE_SCOPE(OpenGL_SurfaceLoad);
     gl_buffer.resize(params.max_mip_level);
@@ -1021,13 +1003,12 @@ void CachedSurface::LoadGLBuffer() {
         for (u32 i = 0; i < params.max_mip_level; i++)
             SwizzleFunc(morton_to_gl_fns, params, gl_buffer[i], i);
     } else {
-        const auto texture_src_data{Memory::GetPointer(params.addr)};
-        const auto texture_src_data_end{texture_src_data + params.size_in_bytes_gl};
-        gl_buffer[0].assign(texture_src_data, texture_src_data_end);
+        gl_buffer[0].resize(params.size_in_bytes_gl);
+        Memory::ReadBlock(params.addr, gl_buffer[0].data(), gl_buffer[0].size());
     }
     for (u32 i = 0; i < params.max_mip_level; i++)
         ConvertFormatAsNeeded_LoadGLBuffer(gl_buffer[i], params.pixel_format, params.MipWidth(i),
-                                           params.MipHeight(i));
+                                           params.MipHeight(i), params.MipDepth(i));
 }
 
 MICROPROFILE_DEFINE(OpenGL_SurfaceFlush, "OpenGL", "Surface Flush", MP_RGB(128, 192, 64));
@@ -1042,7 +1023,7 @@ void CachedSurface::FlushGLBuffer() {
 
     const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
     // Ensure no bad interactions with GL_UNPACK_ALIGNMENT
-    ASSERT(params.width * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 == 0);
+    ASSERT(params.width * GetBytesPerPixel(params.pixel_format) % 4 == 0);
     glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(params.width));
     ASSERT(!tuple.compressed);
     glBindBuffer(GL_PIXEL_PACK_BUFFER, 0);
@@ -1052,15 +1033,14 @@ void CachedSurface::FlushGLBuffer() {
     ConvertFormatAsNeeded_FlushGLBuffer(gl_buffer[0], params.pixel_format, params.width,
                                         params.height);
     ASSERT(params.type != SurfaceType::Fill);
-    const u8* const texture_src_data = Memory::GetPointer(params.addr);
-    ASSERT(texture_src_data);
+    ASSERT(/*texture_src_data*/ Memory::GetPointer(params.addr));
     if (params.is_tiled) {
         ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
                    params.block_width, static_cast<u32>(params.target));
 
         SwizzleFunc(gl_to_morton_fns, params, gl_buffer[0], 0);
     } else {
-        std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer[0].data(), GetSizeInBytes());
+        Memory::WriteBlock(GetAddr(), gl_buffer[0].data(), GetSizeInBytes());
     }
 }
 
@@ -1074,7 +1054,7 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
     std::size_t buffer_offset =
         static_cast<std::size_t>(static_cast<std::size_t>(y0) * params.MipWidth(mip_map) +
                                  static_cast<std::size_t>(x0)) *
-        SurfaceParams::GetBytesPerPixel(params.pixel_format);
+        GetBytesPerPixel(params.pixel_format);
 
     const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
     const GLuint target_tex = texture.handle;
@@ -1090,35 +1070,35 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
     cur_state.Apply();
 
     // Ensure no bad interactions with GL_UNPACK_ALIGNMENT
-    ASSERT(params.MipWidth(mip_map) * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 ==
-           0);
+    ASSERT(params.MipWidth(mip_map) * GetBytesPerPixel(params.pixel_format) % 4 == 0);
     glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.MipWidth(mip_map)));
 
     GLsizei image_size = static_cast<GLsizei>(params.GetMipmapSizeGL(mip_map, false));
     glActiveTexture(GL_TEXTURE0);
     if (tuple.compressed) {
         switch (params.target) {
-        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture2D:
             glCompressedTexImage2D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
                                    static_cast<GLsizei>(params.MipWidth(mip_map)),
                                    static_cast<GLsizei>(params.MipHeight(mip_map)), 0, image_size,
                                    &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture3D:
             glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
                                    static_cast<GLsizei>(params.MipWidth(mip_map)),
                                    static_cast<GLsizei>(params.MipHeight(mip_map)),
                                    static_cast<GLsizei>(params.MipDepth(mip_map)), 0, image_size,
                                    &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::TextureCubeArray:
             glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
                                    static_cast<GLsizei>(params.MipWidth(mip_map)),
                                    static_cast<GLsizei>(params.MipHeight(mip_map)),
                                    static_cast<GLsizei>(params.depth), 0, image_size,
                                    &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap: {
+        case SurfaceTarget::TextureCubemap: {
             GLsizei layer_size = static_cast<GLsizei>(params.LayerSizeGL(mip_map));
             for (std::size_t face = 0; face < params.depth; ++face) {
                 glCompressedTexImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face),
@@ -1143,30 +1123,31 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
     } else {
 
         switch (params.target) {
-        case SurfaceParams::SurfaceTarget::Texture1D:
+        case SurfaceTarget::Texture1D:
             glTexSubImage1D(SurfaceTargetToGL(params.target), mip_map, x0,
                             static_cast<GLsizei>(rect.GetWidth()), tuple.format, tuple.type,
                             &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture2D:
             glTexSubImage2D(SurfaceTargetToGL(params.target), mip_map, x0, y0,
                             static_cast<GLsizei>(rect.GetWidth()),
                             static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
                             &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::Texture3D:
+        case SurfaceTarget::Texture3D:
             glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0,
                             static_cast<GLsizei>(rect.GetWidth()),
                             static_cast<GLsizei>(rect.GetHeight()), params.MipDepth(mip_map),
                             tuple.format, tuple.type, &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::TextureCubeArray:
             glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0,
                             static_cast<GLsizei>(rect.GetWidth()),
                             static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format,
                             tuple.type, &gl_buffer[mip_map][buffer_offset]);
             break;
-        case SurfaceParams::SurfaceTarget::TextureCubemap: {
+        case SurfaceTarget::TextureCubemap: {
             std::size_t start = buffer_offset;
             for (std::size_t face = 0; face < params.depth; ++face) {
                 glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), mip_map,
@@ -1190,7 +1171,7 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
     glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
 }
 
-MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 64, 192));
+MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 192, 64));
 void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle) {
     if (params.type == SurfaceType::Fill)
         return;
@@ -1201,7 +1182,8 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
         UploadGLMipmapTexture(i, read_fb_handle, draw_fb_handle);
 }
 
-RasterizerCacheOpenGL::RasterizerCacheOpenGL() {
+RasterizerCacheOpenGL::RasterizerCacheOpenGL(RasterizerOpenGL& rasterizer)
+    : RasterizerCache{rasterizer} {
     read_framebuffer.Create();
     draw_framebuffer.Create();
     copy_pbo.Create();
@@ -1341,8 +1323,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
     // For compatible surfaces, we can just do fast glCopyImageSubData based copy
     if (old_params.target == new_params.target && old_params.type == new_params.type &&
         old_params.depth == new_params.depth && old_params.depth == 1 &&
-        SurfaceParams::GetFormatBpp(old_params.pixel_format) ==
-            SurfaceParams::GetFormatBpp(new_params.pixel_format)) {
+        GetFormatBpp(old_params.pixel_format) == GetFormatBpp(new_params.pixel_format)) {
         FastCopySurface(old_surface, new_surface);
         return new_surface;
     }
@@ -1355,15 +1336,17 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
     const bool is_blit{old_params.pixel_format == new_params.pixel_format};
 
     switch (new_params.target) {
-    case SurfaceParams::SurfaceTarget::Texture2D:
+    case SurfaceTarget::Texture2D:
         if (is_blit) {
             BlitSurface(old_surface, new_surface, read_framebuffer.handle, draw_framebuffer.handle);
         } else {
             CopySurface(old_surface, new_surface, copy_pbo.handle);
         }
         break;
-    case SurfaceParams::SurfaceTarget::TextureCubemap:
-    case SurfaceParams::SurfaceTarget::Texture3D:
+    case SurfaceTarget::TextureCubemap:
+    case SurfaceTarget::Texture3D:
+    case SurfaceTarget::Texture2DArray:
+    case SurfaceTarget::TextureCubeArray:
         AccurateCopySurface(old_surface, new_surface);
         break;
     default:
@@ -1373,7 +1356,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
     }
 
     return new_surface;
-} // namespace OpenGL
+}
 
 Surface RasterizerCacheOpenGL::TryFindFramebufferSurface(VAddr addr) const {
     return TryGet(addr);

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -7,6 +7,7 @@
 #include <array>
 #include <map>
 #include <memory>
+#include <string>
 #include <vector>
 
 #include "common/alignment.h"
@@ -18,6 +19,7 @@
 #include "video_core/rasterizer_cache.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
 #include "video_core/renderer_opengl/gl_shader_gen.h"
+#include "video_core/surface.h"
 #include "video_core/textures/decoders.h"
 #include "video_core/textures/texture.h"
 
@@ -27,135 +29,12 @@ class CachedSurface;
 using Surface = std::shared_ptr<CachedSurface>;
 using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, MathUtil::Rectangle<u32>>;
 
+using SurfaceTarget = VideoCore::Surface::SurfaceTarget;
+using SurfaceType = VideoCore::Surface::SurfaceType;
+using PixelFormat = VideoCore::Surface::PixelFormat;
+using ComponentType = VideoCore::Surface::ComponentType;
+
 struct SurfaceParams {
-    enum class PixelFormat {
-        ABGR8U = 0,
-        ABGR8S = 1,
-        ABGR8UI = 2,
-        B5G6R5U = 3,
-        A2B10G10R10U = 4,
-        A1B5G5R5U = 5,
-        R8U = 6,
-        R8UI = 7,
-        RGBA16F = 8,
-        RGBA16U = 9,
-        RGBA16UI = 10,
-        R11FG11FB10F = 11,
-        RGBA32UI = 12,
-        DXT1 = 13,
-        DXT23 = 14,
-        DXT45 = 15,
-        DXN1 = 16, // This is also known as BC4
-        DXN2UNORM = 17,
-        DXN2SNORM = 18,
-        BC7U = 19,
-        BC6H_UF16 = 20,
-        BC6H_SF16 = 21,
-        ASTC_2D_4X4 = 22,
-        G8R8U = 23,
-        G8R8S = 24,
-        BGRA8 = 25,
-        RGBA32F = 26,
-        RG32F = 27,
-        R32F = 28,
-        R16F = 29,
-        R16U = 30,
-        R16S = 31,
-        R16UI = 32,
-        R16I = 33,
-        RG16 = 34,
-        RG16F = 35,
-        RG16UI = 36,
-        RG16I = 37,
-        RG16S = 38,
-        RGB32F = 39,
-        RGBA8_SRGB = 40,
-        RG8U = 41,
-        RG8S = 42,
-        RG32UI = 43,
-        R32UI = 44,
-        ASTC_2D_8X8 = 45,
-        ASTC_2D_8X5 = 46,
-        ASTC_2D_5X4 = 47,
-        BGRA8_SRGB = 48,
-        DXT1_SRGB = 49,
-        DXT23_SRGB = 50,
-        DXT45_SRGB = 51,
-        BC7U_SRGB = 52,
-        ASTC_2D_4X4_SRGB = 53,
-        ASTC_2D_8X8_SRGB = 54,
-        ASTC_2D_8X5_SRGB = 55,
-        ASTC_2D_5X4_SRGB = 56,
-
-        MaxColorFormat,
-
-        // Depth formats
-        Z32F = 57,
-        Z16 = 58,
-
-        MaxDepthFormat,
-
-        // DepthStencil formats
-        Z24S8 = 59,
-        S8Z24 = 60,
-        Z32FS8 = 61,
-
-        MaxDepthStencilFormat,
-
-        Max = MaxDepthStencilFormat,
-        Invalid = 255,
-    };
-
-    static constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max);
-
-    enum class ComponentType {
-        Invalid = 0,
-        SNorm = 1,
-        UNorm = 2,
-        SInt = 3,
-        UInt = 4,
-        Float = 5,
-    };
-
-    enum class SurfaceType {
-        ColorTexture = 0,
-        Depth = 1,
-        DepthStencil = 2,
-        Fill = 3,
-        Invalid = 4,
-    };
-
-    enum class SurfaceTarget {
-        Texture1D,
-        Texture2D,
-        Texture3D,
-        Texture1DArray,
-        Texture2DArray,
-        TextureCubemap,
-    };
-
-    static SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type) {
-        switch (texture_type) {
-        case Tegra::Texture::TextureType::Texture1D:
-            return SurfaceTarget::Texture1D;
-        case Tegra::Texture::TextureType::Texture2D:
-        case Tegra::Texture::TextureType::Texture2DNoMipmap:
-            return SurfaceTarget::Texture2D;
-        case Tegra::Texture::TextureType::Texture3D:
-            return SurfaceTarget::Texture3D;
-        case Tegra::Texture::TextureType::TextureCubemap:
-            return SurfaceTarget::TextureCubemap;
-        case Tegra::Texture::TextureType::Texture1DArray:
-            return SurfaceTarget::Texture1DArray;
-        case Tegra::Texture::TextureType::Texture2DArray:
-            return SurfaceTarget::Texture2DArray;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", static_cast<u32>(texture_type));
-            UNREACHABLE();
-            return SurfaceTarget::Texture2D;
-        }
-    }
-
     static std::string SurfaceTargetName(SurfaceTarget target) {
         switch (target) {
         case SurfaceTarget::Texture1D:
@@ -170,6 +49,8 @@ struct SurfaceParams {
             return "Texture2DArray";
         case SurfaceTarget::TextureCubemap:
             return "TextureCubemap";
+        case SurfaceTarget::TextureCubeArray:
+            return "TextureCubeArray";
         default:
             LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
             UNREACHABLE();
@@ -177,660 +58,8 @@ struct SurfaceParams {
         }
     }
 
-    static bool SurfaceTargetIsLayered(SurfaceTarget target) {
-        switch (target) {
-        case SurfaceTarget::Texture1D:
-        case SurfaceTarget::Texture2D:
-        case SurfaceTarget::Texture3D:
-            return false;
-        case SurfaceTarget::Texture1DArray:
-        case SurfaceTarget::Texture2DArray:
-        case SurfaceTarget::TextureCubemap:
-            return true;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
-            UNREACHABLE();
-            return false;
-        }
-    }
-
-    /**
-     * Gets the compression factor for the specified PixelFormat. This applies to just the
-     * "compressed width" and "compressed height", not the overall compression factor of a
-     * compressed image. This is used for maintaining proper surface sizes for compressed
-     * texture formats.
-     */
-    static constexpr u32 GetCompressionFactor(PixelFormat format) {
-        if (format == PixelFormat::Invalid)
-            return 0;
-
-        constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{
-            1, // ABGR8U
-            1, // ABGR8S
-            1, // ABGR8UI
-            1, // B5G6R5U
-            1, // A2B10G10R10U
-            1, // A1B5G5R5U
-            1, // R8U
-            1, // R8UI
-            1, // RGBA16F
-            1, // RGBA16U
-            1, // RGBA16UI
-            1, // R11FG11FB10F
-            1, // RGBA32UI
-            4, // DXT1
-            4, // DXT23
-            4, // DXT45
-            4, // DXN1
-            4, // DXN2UNORM
-            4, // DXN2SNORM
-            4, // BC7U
-            4, // BC6H_UF16
-            4, // BC6H_SF16
-            4, // ASTC_2D_4X4
-            1, // G8R8U
-            1, // G8R8S
-            1, // BGRA8
-            1, // RGBA32F
-            1, // RG32F
-            1, // R32F
-            1, // R16F
-            1, // R16U
-            1, // R16S
-            1, // R16UI
-            1, // R16I
-            1, // RG16
-            1, // RG16F
-            1, // RG16UI
-            1, // RG16I
-            1, // RG16S
-            1, // RGB32F
-            1, // RGBA8_SRGB
-            1, // RG8U
-            1, // RG8S
-            1, // RG32UI
-            1, // R32UI
-            4, // ASTC_2D_8X8
-            4, // ASTC_2D_8X5
-            4, // ASTC_2D_5X4
-            1, // BGRA8_SRGB
-            4, // DXT1_SRGB
-            4, // DXT23_SRGB
-            4, // DXT45_SRGB
-            4, // BC7U_SRGB
-            4, // ASTC_2D_4X4_SRGB
-            4, // ASTC_2D_8X8_SRGB
-            4, // ASTC_2D_8X5_SRGB
-            4, // ASTC_2D_5X4_SRGB
-            1, // Z32F
-            1, // Z16
-            1, // Z24S8
-            1, // S8Z24
-            1, // Z32FS8
-        }};
-
-        ASSERT(static_cast<std::size_t>(format) < compression_factor_table.size());
-        return compression_factor_table[static_cast<std::size_t>(format)];
-    }
-
-    static constexpr u32 GetDefaultBlockHeight(PixelFormat format) {
-        if (format == PixelFormat::Invalid)
-            return 0;
-        constexpr std::array<u32, MaxPixelFormat> block_height_table = {{
-            1, // ABGR8U
-            1, // ABGR8S
-            1, // ABGR8UI
-            1, // B5G6R5U
-            1, // A2B10G10R10U
-            1, // A1B5G5R5U
-            1, // R8U
-            1, // R8UI
-            1, // RGBA16F
-            1, // RGBA16U
-            1, // RGBA16UI
-            1, // R11FG11FB10F
-            1, // RGBA32UI
-            4, // DXT1
-            4, // DXT23
-            4, // DXT45
-            4, // DXN1
-            4, // DXN2UNORM
-            4, // DXN2SNORM
-            4, // BC7U
-            4, // BC6H_UF16
-            4, // BC6H_SF16
-            4, // ASTC_2D_4X4
-            1, // G8R8U
-            1, // G8R8S
-            1, // BGRA8
-            1, // RGBA32F
-            1, // RG32F
-            1, // R32F
-            1, // R16F
-            1, // R16U
-            1, // R16S
-            1, // R16UI
-            1, // R16I
-            1, // RG16
-            1, // RG16F
-            1, // RG16UI
-            1, // RG16I
-            1, // RG16S
-            1, // RGB32F
-            1, // RGBA8_SRGB
-            1, // RG8U
-            1, // RG8S
-            1, // RG32UI
-            1, // R32UI
-            8, // ASTC_2D_8X8
-            5, // ASTC_2D_8X5
-            4, // ASTC_2D_5X4
-            1, // BGRA8_SRGB
-            4, // DXT1_SRGB
-            4, // DXT23_SRGB
-            4, // DXT45_SRGB
-            4, // BC7U_SRGB
-            4, // ASTC_2D_4X4_SRGB
-            8, // ASTC_2D_8X8_SRGB
-            5, // ASTC_2D_8X5_SRGB
-            4, // ASTC_2D_5X4_SRGB
-            1, // Z32F
-            1, // Z16
-            1, // Z24S8
-            1, // S8Z24
-            1, // Z32FS8
-        }};
-        ASSERT(static_cast<std::size_t>(format) < block_height_table.size());
-        return block_height_table[static_cast<std::size_t>(format)];
-    }
-
-    static constexpr u32 GetFormatBpp(PixelFormat format) {
-        if (format == PixelFormat::Invalid)
-            return 0;
-
-        constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
-            32,  // ABGR8U
-            32,  // ABGR8S
-            32,  // ABGR8UI
-            16,  // B5G6R5U
-            32,  // A2B10G10R10U
-            16,  // A1B5G5R5U
-            8,   // R8U
-            8,   // R8UI
-            64,  // RGBA16F
-            64,  // RGBA16U
-            64,  // RGBA16UI
-            32,  // R11FG11FB10F
-            128, // RGBA32UI
-            64,  // DXT1
-            128, // DXT23
-            128, // DXT45
-            64,  // DXN1
-            128, // DXN2UNORM
-            128, // DXN2SNORM
-            128, // BC7U
-            128, // BC6H_UF16
-            128, // BC6H_SF16
-            32,  // ASTC_2D_4X4
-            16,  // G8R8U
-            16,  // G8R8S
-            32,  // BGRA8
-            128, // RGBA32F
-            64,  // RG32F
-            32,  // R32F
-            16,  // R16F
-            16,  // R16U
-            16,  // R16S
-            16,  // R16UI
-            16,  // R16I
-            32,  // RG16
-            32,  // RG16F
-            32,  // RG16UI
-            32,  // RG16I
-            32,  // RG16S
-            96,  // RGB32F
-            32,  // RGBA8_SRGB
-            16,  // RG8U
-            16,  // RG8S
-            64,  // RG32UI
-            32,  // R32UI
-            16,  // ASTC_2D_8X8
-            16,  // ASTC_2D_8X5
-            32,  // ASTC_2D_5X4
-            32,  // BGRA8_SRGB
-            64,  // DXT1_SRGB
-            128, // DXT23_SRGB
-            128, // DXT45_SRGB
-            128, // BC7U
-            32,  // ASTC_2D_4X4_SRGB
-            16,  // ASTC_2D_8X8_SRGB
-            16,  // ASTC_2D_8X5_SRGB
-            32,  // ASTC_2D_5X4_SRGB
-            32,  // Z32F
-            16,  // Z16
-            32,  // Z24S8
-            32,  // S8Z24
-            64,  // Z32FS8
-        }};
-
-        ASSERT(static_cast<std::size_t>(format) < bpp_table.size());
-        return bpp_table[static_cast<std::size_t>(format)];
-    }
-
     u32 GetFormatBpp() const {
-        return GetFormatBpp(pixel_format);
-    }
-
-    static PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
-        switch (format) {
-        case Tegra::DepthFormat::S8_Z24_UNORM:
-            return PixelFormat::S8Z24;
-        case Tegra::DepthFormat::Z24_S8_UNORM:
-            return PixelFormat::Z24S8;
-        case Tegra::DepthFormat::Z32_FLOAT:
-            return PixelFormat::Z32F;
-        case Tegra::DepthFormat::Z16_UNORM:
-            return PixelFormat::Z16;
-        case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
-            return PixelFormat::Z32FS8;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
-    }
-
-    static PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
-        switch (format) {
-        // TODO (Hexagon12): Converting SRGBA to RGBA is a hack and doesn't completely correct the
-        // gamma.
-        case Tegra::RenderTargetFormat::RGBA8_SRGB:
-            return PixelFormat::RGBA8_SRGB;
-        case Tegra::RenderTargetFormat::RGBA8_UNORM:
-            return PixelFormat::ABGR8U;
-        case Tegra::RenderTargetFormat::RGBA8_SNORM:
-            return PixelFormat::ABGR8S;
-        case Tegra::RenderTargetFormat::RGBA8_UINT:
-            return PixelFormat::ABGR8UI;
-        case Tegra::RenderTargetFormat::BGRA8_SRGB:
-            return PixelFormat::BGRA8_SRGB;
-        case Tegra::RenderTargetFormat::BGRA8_UNORM:
-            return PixelFormat::BGRA8;
-        case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
-            return PixelFormat::A2B10G10R10U;
-        case Tegra::RenderTargetFormat::RGBA16_FLOAT:
-            return PixelFormat::RGBA16F;
-        case Tegra::RenderTargetFormat::RGBA16_UNORM:
-            return PixelFormat::RGBA16U;
-        case Tegra::RenderTargetFormat::RGBA16_UINT:
-            return PixelFormat::RGBA16UI;
-        case Tegra::RenderTargetFormat::RGBA32_FLOAT:
-            return PixelFormat::RGBA32F;
-        case Tegra::RenderTargetFormat::RG32_FLOAT:
-            return PixelFormat::RG32F;
-        case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
-            return PixelFormat::R11FG11FB10F;
-        case Tegra::RenderTargetFormat::B5G6R5_UNORM:
-            return PixelFormat::B5G6R5U;
-        case Tegra::RenderTargetFormat::BGR5A1_UNORM:
-            return PixelFormat::A1B5G5R5U;
-        case Tegra::RenderTargetFormat::RGBA32_UINT:
-            return PixelFormat::RGBA32UI;
-        case Tegra::RenderTargetFormat::R8_UNORM:
-            return PixelFormat::R8U;
-        case Tegra::RenderTargetFormat::R8_UINT:
-            return PixelFormat::R8UI;
-        case Tegra::RenderTargetFormat::RG16_FLOAT:
-            return PixelFormat::RG16F;
-        case Tegra::RenderTargetFormat::RG16_UINT:
-            return PixelFormat::RG16UI;
-        case Tegra::RenderTargetFormat::RG16_SINT:
-            return PixelFormat::RG16I;
-        case Tegra::RenderTargetFormat::RG16_UNORM:
-            return PixelFormat::RG16;
-        case Tegra::RenderTargetFormat::RG16_SNORM:
-            return PixelFormat::RG16S;
-        case Tegra::RenderTargetFormat::RG8_UNORM:
-            return PixelFormat::RG8U;
-        case Tegra::RenderTargetFormat::RG8_SNORM:
-            return PixelFormat::RG8S;
-        case Tegra::RenderTargetFormat::R16_FLOAT:
-            return PixelFormat::R16F;
-        case Tegra::RenderTargetFormat::R16_UNORM:
-            return PixelFormat::R16U;
-        case Tegra::RenderTargetFormat::R16_SNORM:
-            return PixelFormat::R16S;
-        case Tegra::RenderTargetFormat::R16_UINT:
-            return PixelFormat::R16UI;
-        case Tegra::RenderTargetFormat::R16_SINT:
-            return PixelFormat::R16I;
-        case Tegra::RenderTargetFormat::R32_FLOAT:
-            return PixelFormat::R32F;
-        case Tegra::RenderTargetFormat::R32_UINT:
-            return PixelFormat::R32UI;
-        case Tegra::RenderTargetFormat::RG32_UINT:
-            return PixelFormat::RG32UI;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
-    }
-
-    static PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
-                                                    Tegra::Texture::ComponentType component_type,
-                                                    bool is_srgb) {
-        // TODO(Subv): Properly implement this
-        switch (format) {
-        case Tegra::Texture::TextureFormat::A8R8G8B8:
-            if (is_srgb) {
-                return PixelFormat::RGBA8_SRGB;
-            }
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::ABGR8U;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::ABGR8S;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::ABGR8UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::B5G6R5:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::B5G6R5U;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::A2B10G10R10:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::A2B10G10R10U;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::A1B5G5R5:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::A1B5G5R5U;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R8:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::R8U;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::R8UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::G8R8:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::G8R8U;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::G8R8S;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R16_G16_B16_A16:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::RGBA16U;
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RGBA16F;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::BF10GF11RF11:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::R11FG11FB10F;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R32_G32_B32_A32:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RGBA32F;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::RGBA32UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R32_G32:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RG32F;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::RG32UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R32_G32_B32:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RGB32F;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R16:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::R16F;
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::R16U;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::R16S;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::R16UI;
-            case Tegra::Texture::ComponentType::SINT:
-                return PixelFormat::R16I;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::R32:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::R32F;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::R32UI;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::ZF32:
-            return PixelFormat::Z32F;
-        case Tegra::Texture::TextureFormat::Z16:
-            return PixelFormat::Z16;
-        case Tegra::Texture::TextureFormat::Z24S8:
-            return PixelFormat::Z24S8;
-        case Tegra::Texture::TextureFormat::DXT1:
-            return is_srgb ? PixelFormat::DXT1_SRGB : PixelFormat::DXT1;
-        case Tegra::Texture::TextureFormat::DXT23:
-            return is_srgb ? PixelFormat::DXT23_SRGB : PixelFormat::DXT23;
-        case Tegra::Texture::TextureFormat::DXT45:
-            return is_srgb ? PixelFormat::DXT45_SRGB : PixelFormat::DXT45;
-        case Tegra::Texture::TextureFormat::DXN1:
-            return PixelFormat::DXN1;
-        case Tegra::Texture::TextureFormat::DXN2:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::DXN2UNORM;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::DXN2SNORM;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        case Tegra::Texture::TextureFormat::BC7U:
-            return is_srgb ? PixelFormat::BC7U_SRGB : PixelFormat::BC7U;
-        case Tegra::Texture::TextureFormat::BC6H_UF16:
-            return PixelFormat::BC6H_UF16;
-        case Tegra::Texture::TextureFormat::BC6H_SF16:
-            return PixelFormat::BC6H_SF16;
-        case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
-            return is_srgb ? PixelFormat::ASTC_2D_4X4_SRGB : PixelFormat::ASTC_2D_4X4;
-        case Tegra::Texture::TextureFormat::ASTC_2D_5X4:
-            return is_srgb ? PixelFormat::ASTC_2D_5X4_SRGB : PixelFormat::ASTC_2D_5X4;
-        case Tegra::Texture::TextureFormat::ASTC_2D_8X8:
-            return is_srgb ? PixelFormat::ASTC_2D_8X8_SRGB : PixelFormat::ASTC_2D_8X8;
-        case Tegra::Texture::TextureFormat::ASTC_2D_8X5:
-            return is_srgb ? PixelFormat::ASTC_2D_8X5_SRGB : PixelFormat::ASTC_2D_8X5;
-        case Tegra::Texture::TextureFormat::R16_G16:
-            switch (component_type) {
-            case Tegra::Texture::ComponentType::FLOAT:
-                return PixelFormat::RG16F;
-            case Tegra::Texture::ComponentType::UNORM:
-                return PixelFormat::RG16;
-            case Tegra::Texture::ComponentType::SNORM:
-                return PixelFormat::RG16S;
-            case Tegra::Texture::ComponentType::UINT:
-                return PixelFormat::RG16UI;
-            case Tegra::Texture::ComponentType::SINT:
-                return PixelFormat::RG16I;
-            }
-            LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
-                         static_cast<u32>(component_type));
-            UNREACHABLE();
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}, component_type={}",
-                         static_cast<u32>(format), static_cast<u32>(component_type));
-            UNREACHABLE();
-        }
-    }
-
-    static ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
-        // TODO(Subv): Implement more component types
-        switch (type) {
-        case Tegra::Texture::ComponentType::UNORM:
-            return ComponentType::UNorm;
-        case Tegra::Texture::ComponentType::FLOAT:
-            return ComponentType::Float;
-        case Tegra::Texture::ComponentType::SNORM:
-            return ComponentType::SNorm;
-        case Tegra::Texture::ComponentType::UINT:
-            return ComponentType::UInt;
-        case Tegra::Texture::ComponentType::SINT:
-            return ComponentType::SInt;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented component type={}", static_cast<u32>(type));
-            UNREACHABLE();
-        }
-    }
-
-    static ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format) {
-        // TODO(Subv): Implement more render targets
-        switch (format) {
-        case Tegra::RenderTargetFormat::RGBA8_UNORM:
-        case Tegra::RenderTargetFormat::RGBA8_SRGB:
-        case Tegra::RenderTargetFormat::BGRA8_UNORM:
-        case Tegra::RenderTargetFormat::BGRA8_SRGB:
-        case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
-        case Tegra::RenderTargetFormat::R8_UNORM:
-        case Tegra::RenderTargetFormat::RG16_UNORM:
-        case Tegra::RenderTargetFormat::R16_UNORM:
-        case Tegra::RenderTargetFormat::B5G6R5_UNORM:
-        case Tegra::RenderTargetFormat::BGR5A1_UNORM:
-        case Tegra::RenderTargetFormat::RG8_UNORM:
-        case Tegra::RenderTargetFormat::RGBA16_UNORM:
-            return ComponentType::UNorm;
-        case Tegra::RenderTargetFormat::RGBA8_SNORM:
-        case Tegra::RenderTargetFormat::RG16_SNORM:
-        case Tegra::RenderTargetFormat::R16_SNORM:
-        case Tegra::RenderTargetFormat::RG8_SNORM:
-            return ComponentType::SNorm;
-        case Tegra::RenderTargetFormat::RGBA16_FLOAT:
-        case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
-        case Tegra::RenderTargetFormat::RGBA32_FLOAT:
-        case Tegra::RenderTargetFormat::RG32_FLOAT:
-        case Tegra::RenderTargetFormat::RG16_FLOAT:
-        case Tegra::RenderTargetFormat::R16_FLOAT:
-        case Tegra::RenderTargetFormat::R32_FLOAT:
-            return ComponentType::Float;
-        case Tegra::RenderTargetFormat::RGBA32_UINT:
-        case Tegra::RenderTargetFormat::RGBA16_UINT:
-        case Tegra::RenderTargetFormat::RG16_UINT:
-        case Tegra::RenderTargetFormat::R8_UINT:
-        case Tegra::RenderTargetFormat::R16_UINT:
-        case Tegra::RenderTargetFormat::RG32_UINT:
-        case Tegra::RenderTargetFormat::R32_UINT:
-        case Tegra::RenderTargetFormat::RGBA8_UINT:
-            return ComponentType::UInt;
-        case Tegra::RenderTargetFormat::RG16_SINT:
-        case Tegra::RenderTargetFormat::R16_SINT:
-            return ComponentType::SInt;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
-    }
-
-    static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
-        switch (format) {
-        case Tegra::FramebufferConfig::PixelFormat::ABGR8:
-            return PixelFormat::ABGR8U;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
-    }
-
-    static ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format) {
-        switch (format) {
-        case Tegra::DepthFormat::Z16_UNORM:
-        case Tegra::DepthFormat::S8_Z24_UNORM:
-        case Tegra::DepthFormat::Z24_S8_UNORM:
-            return ComponentType::UNorm;
-        case Tegra::DepthFormat::Z32_FLOAT:
-        case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
-            return ComponentType::Float;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-            UNREACHABLE();
-        }
-    }
-
-    static SurfaceType GetFormatType(PixelFormat pixel_format) {
-        if (static_cast<std::size_t>(pixel_format) <
-            static_cast<std::size_t>(PixelFormat::MaxColorFormat)) {
-            return SurfaceType::ColorTexture;
-        }
-
-        if (static_cast<std::size_t>(pixel_format) <
-            static_cast<std::size_t>(PixelFormat::MaxDepthFormat)) {
-            return SurfaceType::Depth;
-        }
-
-        if (static_cast<std::size_t>(pixel_format) <
-            static_cast<std::size_t>(PixelFormat::MaxDepthStencilFormat)) {
-            return SurfaceType::DepthStencil;
-        }
-
-        // TODO(Subv): Implement the other formats
-        ASSERT(false);
-
-        return SurfaceType::Invalid;
-    }
-
-    /// Returns the sizer in bytes of the specified pixel format
-    static constexpr u32 GetBytesPerPixel(PixelFormat pixel_format) {
-        if (pixel_format == SurfaceParams::PixelFormat::Invalid) {
-            return 0;
-        }
-        return GetFormatBpp(pixel_format) / CHAR_BIT;
+        return VideoCore::Surface::GetFormatBpp(pixel_format);
     }
 
     /// Returns the rectangle corresponding to this surface
@@ -912,7 +141,7 @@ struct SurfaceParams {
     }
 
     u32 MipDepth(u32 mip_level) const {
-        return std::max(1U, depth >> mip_level);
+        return is_layered ? depth : std::max(1U, depth >> mip_level);
     }
 
     // Auto block resizing algorithm from:
@@ -1035,6 +264,8 @@ struct hash<SurfaceReserveKey> {
 
 namespace OpenGL {
 
+class RasterizerOpenGL;
+
 class CachedSurface final : public RasterizerCacheObject {
 public:
     CachedSurface(const SurfaceParams& params);
@@ -1082,7 +313,7 @@ private:
 
 class RasterizerCacheOpenGL final : public RasterizerCache<Surface> {
 public:
-    RasterizerCacheOpenGL();
+    explicit RasterizerCacheOpenGL(RasterizerOpenGL& rasterizer);
 
     /// Get a surface based on the texture configuration
     Surface GetTextureSurface(const Tegra::Texture::FullTextureInfo& config,

src/video_core/renderer_opengl/gl_resource_manager.cpp

@@ -0,0 +1,186 @@
+// Copyright 2015 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <utility>
+#include <glad/glad.h>
+#include "common/common_types.h"
+#include "common/microprofile.h"
+#include "video_core/renderer_opengl/gl_resource_manager.h"
+#include "video_core/renderer_opengl/gl_shader_util.h"
+#include "video_core/renderer_opengl/gl_state.h"
+
+MICROPROFILE_DEFINE(OpenGL_ResourceCreation, "OpenGL", "Resource Creation", MP_RGB(128, 128, 192));
+MICROPROFILE_DEFINE(OpenGL_ResourceDeletion, "OpenGL", "Resource Deletion", MP_RGB(128, 128, 192));
+
+namespace OpenGL {
+
+void OGLTexture::Create() {
+    if (handle != 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    glGenTextures(1, &handle);
+}
+
+void OGLTexture::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteTextures(1, &handle);
+    OpenGLState::GetCurState().UnbindTexture(handle).Apply();
+    handle = 0;
+}
+
+void OGLSampler::Create() {
+    if (handle != 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    glGenSamplers(1, &handle);
+}
+
+void OGLSampler::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteSamplers(1, &handle);
+    OpenGLState::GetCurState().ResetSampler(handle).Apply();
+    handle = 0;
+}
+
+void OGLShader::Create(const char* source, GLenum type) {
+    if (handle != 0)
+        return;
+    if (source == nullptr)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    handle = GLShader::LoadShader(source, type);
+}
+
+void OGLShader::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteShader(handle);
+    handle = 0;
+}
+
+void OGLProgram::CreateFromSource(const char* vert_shader, const char* geo_shader,
+                                  const char* frag_shader, bool separable_program) {
+    OGLShader vert, geo, frag;
+    if (vert_shader)
+        vert.Create(vert_shader, GL_VERTEX_SHADER);
+    if (geo_shader)
+        geo.Create(geo_shader, GL_GEOMETRY_SHADER);
+    if (frag_shader)
+        frag.Create(frag_shader, GL_FRAGMENT_SHADER);
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    Create(separable_program, vert.handle, geo.handle, frag.handle);
+}
+
+void OGLProgram::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteProgram(handle);
+    OpenGLState::GetCurState().ResetProgram(handle).Apply();
+    handle = 0;
+}
+
+void OGLPipeline::Create() {
+    if (handle != 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    glGenProgramPipelines(1, &handle);
+}
+
+void OGLPipeline::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteProgramPipelines(1, &handle);
+    OpenGLState::GetCurState().ResetPipeline(handle).Apply();
+    handle = 0;
+}
+
+void OGLBuffer::Create() {
+    if (handle != 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    glGenBuffers(1, &handle);
+}
+
+void OGLBuffer::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteBuffers(1, &handle);
+    OpenGLState::GetCurState().ResetBuffer(handle).Apply();
+    handle = 0;
+}
+
+void OGLSync::Create() {
+    if (handle != 0)
+        return;
+
+    // Don't profile here, this one is expected to happen ingame.
+    handle = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0);
+}
+
+void OGLSync::Release() {
+    if (handle == 0)
+        return;
+
+    // Don't profile here, this one is expected to happen ingame.
+    glDeleteSync(handle);
+    handle = 0;
+}
+
+void OGLVertexArray::Create() {
+    if (handle != 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    glGenVertexArrays(1, &handle);
+}
+
+void OGLVertexArray::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteVertexArrays(1, &handle);
+    OpenGLState::GetCurState().ResetVertexArray(handle).Apply();
+    handle = 0;
+}
+
+void OGLFramebuffer::Create() {
+    if (handle != 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
+    glGenFramebuffers(1, &handle);
+}
+
+void OGLFramebuffer::Release() {
+    if (handle == 0)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
+    glDeleteFramebuffers(1, &handle);
+    OpenGLState::GetCurState().ResetFramebuffer(handle).Apply();
+    handle = 0;
+}
+
+} // namespace OpenGL

src/video_core/renderer_opengl/gl_resource_manager.h

@@ -8,7 +8,6 @@
 #include <glad/glad.h>
 #include "common/common_types.h"
 #include "video_core/renderer_opengl/gl_shader_util.h"
-#include "video_core/renderer_opengl/gl_state.h"
 
 namespace OpenGL {
 
@@ -29,20 +28,10 @@ public:
     }
 
     /// Creates a new internal OpenGL resource and stores the handle
-    void Create() {
-        if (handle != 0)
-            return;
-        glGenTextures(1, &handle);
-    }
+    void Create();
 
     /// Deletes the internal OpenGL resource
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteTextures(1, &handle);
-        OpenGLState::GetCurState().UnbindTexture(handle).Apply();
-        handle = 0;
-    }
+    void Release();
 
     GLuint handle = 0;
 };
@@ -64,20 +53,10 @@ public:
     }
 
     /// Creates a new internal OpenGL resource and stores the handle
-    void Create() {
-        if (handle != 0)
-            return;
-        glGenSamplers(1, &handle);
-    }
+    void Create();
 
     /// Deletes the internal OpenGL resource
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteSamplers(1, &handle);
-        OpenGLState::GetCurState().ResetSampler(handle).Apply();
-        handle = 0;
-    }
+    void Release();
 
     GLuint handle = 0;
 };
@@ -98,20 +77,9 @@ public:
         return *this;
     }
 
-    void Create(const char* source, GLenum type) {
-        if (handle != 0)
-            return;
-        if (source == nullptr)
-            return;
-        handle = GLShader::LoadShader(source, type);
-    }
+    void Create(const char* source, GLenum type);
 
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteShader(handle);
-        handle = 0;
-    }
+    void Release();
 
     GLuint handle = 0;
 };
@@ -141,25 +109,10 @@ public:
 
     /// Creates a new internal OpenGL resource and stores the handle
     void CreateFromSource(const char* vert_shader, const char* geo_shader, const char* frag_shader,
-                          bool separable_program = false) {
-        OGLShader vert, geo, frag;
-        if (vert_shader)
-            vert.Create(vert_shader, GL_VERTEX_SHADER);
-        if (geo_shader)
-            geo.Create(geo_shader, GL_GEOMETRY_SHADER);
-        if (frag_shader)
-            frag.Create(frag_shader, GL_FRAGMENT_SHADER);
-        Create(separable_program, vert.handle, geo.handle, frag.handle);
-    }
+                          bool separable_program = false);
 
     /// Deletes the internal OpenGL resource
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteProgram(handle);
-        OpenGLState::GetCurState().ResetProgram(handle).Apply();
-        handle = 0;
-    }
+    void Release();
 
     GLuint handle = 0;
 };
@@ -178,20 +131,10 @@ public:
     }
 
     /// Creates a new internal OpenGL resource and stores the handle
-    void Create() {
-        if (handle != 0)
-            return;
-        glGenProgramPipelines(1, &handle);
-    }
+    void Create();
 
     /// Deletes the internal OpenGL resource
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteProgramPipelines(1, &handle);
-        OpenGLState::GetCurState().ResetPipeline(handle).Apply();
-        handle = 0;
-    }
+    void Release();
 
     GLuint handle = 0;
 };
@@ -213,20 +156,10 @@ public:
     }
 
     /// Creates a new internal OpenGL resource and stores the handle
-    void Create() {
-        if (handle != 0)
-            return;
-        glGenBuffers(1, &handle);
-    }
+    void Create();
 
     /// Deletes the internal OpenGL resource
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteBuffers(1, &handle);
-        OpenGLState::GetCurState().ResetBuffer(handle).Apply();
-        handle = 0;
-    }
+    void Release();
 
     GLuint handle = 0;
 };
@@ -247,19 +180,10 @@ public:
     }
 
     /// Creates a new internal OpenGL resource and stores the handle
-    void Create() {
-        if (handle != 0)
-            return;
-        handle = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0);
-    }
+    void Create();
 
     /// Deletes the internal OpenGL resource
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteSync(handle);
-        handle = 0;
-    }
+    void Release();
 
     GLsync handle = 0;
 };
@@ -281,20 +205,10 @@ public:
     }
 
     /// Creates a new internal OpenGL resource and stores the handle
-    void Create() {
-        if (handle != 0)
-            return;
-        glGenVertexArrays(1, &handle);
-    }
+    void Create();
 
     /// Deletes the internal OpenGL resource
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteVertexArrays(1, &handle);
-        OpenGLState::GetCurState().ResetVertexArray(handle).Apply();
-        handle = 0;
-    }
+    void Release();
 
     GLuint handle = 0;
 };
@@ -316,20 +230,10 @@ public:
     }
 
     /// Creates a new internal OpenGL resource and stores the handle
-    void Create() {
-        if (handle != 0)
-            return;
-        glGenFramebuffers(1, &handle);
-    }
+    void Create();
 
     /// Deletes the internal OpenGL resource
-    void Release() {
-        if (handle == 0)
-            return;
-        glDeleteFramebuffers(1, &handle);
-        OpenGLState::GetCurState().ResetFramebuffer(handle).Apply();
-        handle = 0;
-    }
+    void Release();
 
     GLuint handle = 0;
 };

src/video_core/renderer_opengl/gl_shader_cache.cpp

@@ -6,10 +6,10 @@
 #include "core/core.h"
 #include "core/memory.h"
 #include "video_core/engines/maxwell_3d.h"
+#include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/gl_shader_cache.h"
 #include "video_core/renderer_opengl/gl_shader_manager.h"
 #include "video_core/renderer_opengl/utils.h"
-#include "video_core/utils.h"
 
 namespace OpenGL {
 
@@ -121,12 +121,16 @@ GLint CachedShader::GetUniformLocation(const GLShader::SamplerEntry& sampler) {
 }
 
 GLuint CachedShader::LazyGeometryProgram(OGLProgram& target_program,
-                                         const std::string& glsl_topology,
+                                         const std::string& glsl_topology, u32 max_vertices,
                                          const std::string& debug_name) {
     if (target_program.handle != 0) {
         return target_program.handle;
     }
-    const std::string source{geometry_programs.code + "layout (" + glsl_topology + ") in;\n"};
+    std::string source = "#version 430 core\n";
+    source += "layout (" + glsl_topology + ") in;\n";
+    source += "#define MAX_VERTEX_INPUT " + std::to_string(max_vertices) + '\n';
+    source += geometry_programs.code;
+
     OGLShader shader;
     shader.Create(source.c_str(), GL_GEOMETRY_SHADER);
     target_program.Create(true, shader.handle);
@@ -135,6 +139,8 @@ GLuint CachedShader::LazyGeometryProgram(OGLProgram& target_program,
     return target_program.handle;
 };
 
+ShaderCacheOpenGL::ShaderCacheOpenGL(RasterizerOpenGL& rasterizer) : RasterizerCache{rasterizer} {}
+
 Shader ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) {
     const VAddr program_addr{GetShaderAddress(program)};
 

src/video_core/renderer_opengl/gl_shader_cache.h

@@ -16,6 +16,8 @@
 namespace OpenGL {
 
 class CachedShader;
+class RasterizerOpenGL;
+
 using Shader = std::shared_ptr<CachedShader>;
 using Maxwell = Tegra::Engines::Maxwell3D::Regs;
 
@@ -46,22 +48,23 @@ public:
         }
         switch (primitive_mode) {
         case GL_POINTS:
-            return LazyGeometryProgram(geometry_programs.points, "points", "ShaderPoints");
+            return LazyGeometryProgram(geometry_programs.points, "points", 1, "ShaderPoints");
         case GL_LINES:
         case GL_LINE_STRIP:
-            return LazyGeometryProgram(geometry_programs.lines, "lines", "ShaderLines");
+            return LazyGeometryProgram(geometry_programs.lines, "lines", 2, "ShaderLines");
         case GL_LINES_ADJACENCY:
         case GL_LINE_STRIP_ADJACENCY:
-            return LazyGeometryProgram(geometry_programs.lines_adjacency, "lines_adjacency",
+            return LazyGeometryProgram(geometry_programs.lines_adjacency, "lines_adjacency", 4,
                                        "ShaderLinesAdjacency");
         case GL_TRIANGLES:
         case GL_TRIANGLE_STRIP:
         case GL_TRIANGLE_FAN:
-            return LazyGeometryProgram(geometry_programs.triangles, "triangles", "ShaderTriangles");
+            return LazyGeometryProgram(geometry_programs.triangles, "triangles", 3,
+                                       "ShaderTriangles");
         case GL_TRIANGLES_ADJACENCY:
         case GL_TRIANGLE_STRIP_ADJACENCY:
             return LazyGeometryProgram(geometry_programs.triangles_adjacency, "triangles_adjacency",
-                                       "ShaderLines");
+                                       6, "ShaderTrianglesAdjacency");
         default:
             UNREACHABLE_MSG("Unknown primitive mode.");
         }
@@ -76,7 +79,7 @@ public:
 private:
     /// Generates a geometry shader or returns one that already exists.
     GLuint LazyGeometryProgram(OGLProgram& target_program, const std::string& glsl_topology,
-                               const std::string& debug_name);
+                               u32 max_vertices, const std::string& debug_name);
 
     VAddr addr;
     Maxwell::ShaderProgram program_type;
@@ -104,6 +107,8 @@ private:
 
 class ShaderCacheOpenGL final : public RasterizerCache<Shader> {
 public:
+    explicit ShaderCacheOpenGL(RasterizerOpenGL& rasterizer);
+
     /// Gets the current specified shader stage program
     Shader GetStageProgram(Maxwell::ShaderProgram program);
 };

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -373,6 +373,7 @@ public:
         if (sets_cc) {
             const std::string zero_condition = "( " + ConvertIntegerSize(value, size) + " == 0 )";
             SetInternalFlag(InternalFlag::ZeroFlag, zero_condition);
+            LOG_WARNING(HW_GPU, "Control Codes Imcomplete.");
         }
     }
 
@@ -493,10 +494,10 @@ public:
             // instruction for now.
             if (stage == Maxwell3D::Regs::ShaderStage::Geometry) {
                 // TODO(Rodrigo): nouveau sets some attributes after setting emitting a geometry
-                // shader. These instructions use a dirty register as buffer index. To avoid some
-                // drivers from complaining for the out of boundary writes, guard them.
-                const std::string buf_index{"min(" + GetRegisterAsInteger(buf_reg) + ", " +
-                                            std::to_string(MAX_GEOMETRY_BUFFERS - 1) + ')'};
+                // shader. These instructions use a dirty register as buffer index, to avoid some
+                // drivers from complaining about out of boundary writes, guard them.
+                const std::string buf_index{"((" + GetRegisterAsInteger(buf_reg) + ") % " +
+                                            std::to_string(MAX_GEOMETRY_BUFFERS) + ')'};
                 shader.AddLine("amem[" + buf_index + "][" +
                                std::to_string(static_cast<u32>(attribute)) + ']' +
                                GetSwizzle(elem) + " = " + src + ';');
@@ -810,7 +811,11 @@ private:
                                   std::optional<Register> vertex = {}) {
         auto GeometryPass = [&](const std::string& name) {
             if (stage == Maxwell3D::Regs::ShaderStage::Geometry && vertex) {
-                return "gs_" + name + '[' + GetRegisterAsInteger(*vertex, 0, false) + ']';
+                // TODO(Rodrigo): Guard geometry inputs against out of bound reads. Some games set
+                // an 0x80000000 index for those and the shader fails to build. Find out why this
+                // happens and what's its intent.
+                return "gs_" + name + '[' + GetRegisterAsInteger(*vertex, 0, false) +
+                       " % MAX_VERTEX_INPUT]";
             }
             return name;
         };
@@ -1525,6 +1530,10 @@ private:
 
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b, 1, 1,
                                         instr.alu.saturate_d, 0, true);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FMUL Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::FADD_C:
@@ -1535,6 +1544,10 @@ private:
 
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1,
                                         instr.alu.saturate_d, 0, true);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FADD Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::MUFU: {
@@ -1588,6 +1601,10 @@ private:
                                         '(' + condition + ") ? min(" + parameters + ") : max(" +
                                             parameters + ')',
                                         1, 1, false, 0, true);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FMNMX Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::RRO_C:
@@ -1618,6 +1635,10 @@ private:
                                         regs.GetRegisterAsFloat(instr.gpr8) + " * " +
                                             GetImmediate32(instr),
                                         1, 1, instr.fmul32.saturate, 0, true);
+                if (instr.op_32.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FMUL32 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::FADD32I: {
@@ -1641,6 +1662,10 @@ private:
                 }
 
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1, false, 0, true);
+                if (instr.op_32.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "FADD32 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             }
@@ -1661,6 +1686,10 @@ private:
                     std::to_string(instr.bfe.GetLeftShiftValue() + instr.bfe.shift_position) + ')';
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, outer_shift, 1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "BFE Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             default: {
@@ -1698,12 +1727,20 @@ private:
                 // Cast to int is superfluous for arithmetic shift, it's only for a logical shift
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, "int(" + op_a + " >> " + op_b + ')',
                                           1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "SHR Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::SHL_C:
             case OpCode::Id::SHL_R:
             case OpCode::Id::SHL_IMM:
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " << " + op_b, 1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "SHL Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             default: {
                 LOG_CRITICAL(HW_GPU, "Unhandled shift instruction: {}", opcode->get().GetName());
@@ -1723,6 +1760,10 @@ private:
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1,
                                           instr.iadd32i.saturate != 0);
+                if (instr.op_32.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "IADD32 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             case OpCode::Id::LOP32I: {
                 if (instr.alu.lop32i.invert_a)
@@ -1734,6 +1775,10 @@ private:
                 WriteLogicOperation(instr.gpr0, instr.alu.lop32i.operation, op_a, op_b,
                                     Tegra::Shader::PredicateResultMode::None,
                                     Tegra::Shader::Pred::UnusedIndex);
+                if (instr.op_32.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "LOP32I Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             default: {
@@ -1770,6 +1815,10 @@ private:
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1,
                                           instr.alu.saturate_d);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "IADD Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::IADD3_C:
@@ -1831,6 +1880,11 @@ private:
                 }
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0, result, 1, 1);
+
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "IADD3 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::ISCADD_C:
@@ -1846,6 +1900,10 @@ private:
 
                 regs.SetRegisterToInteger(instr.gpr0, true, 0,
                                           "((" + op_a + " << " + shift + ") + " + op_b + ')', 1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "ISCADD Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::POPC_C:
@@ -1877,6 +1935,10 @@ private:
 
                 WriteLogicOperation(instr.gpr0, instr.alu.lop.operation, op_a, op_b,
                                     instr.alu.lop.pred_result_mode, instr.alu.lop.pred48);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "LOP Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::LOP3_C:
@@ -1892,6 +1954,10 @@ private:
                 }
 
                 WriteLop3Instruction(instr.gpr0, op_a, op_b, op_c, lut);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "LOP3 Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::IMNMX_C:
@@ -1906,6 +1972,10 @@ private:
                                           '(' + condition + ") ? min(" + parameters + ") : max(" +
                                               parameters + ')',
                                           1, 1);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "IMNMX Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::LEA_R2:
@@ -2107,6 +2177,10 @@ private:
 
             regs.SetRegisterToFloat(instr.gpr0, 0, "fma(" + op_a + ", " + op_b + ", " + op_c + ')',
                                     1, 1, instr.alu.saturate_d, 0, true);
+            if (instr.generates_cc) {
+                LOG_CRITICAL(HW_GPU, "FFMA Generates an unhandled Control Code");
+                UNREACHABLE();
+            }
 
             break;
         }
@@ -2212,6 +2286,11 @@ private:
                 }
 
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
+
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "I2F Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::F2F_R: {
@@ -2250,6 +2329,11 @@ private:
                 }
 
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1, instr.alu.saturate_d);
+
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "F2F Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             case OpCode::Id::F2I_R:
@@ -2299,6 +2383,10 @@ private:
 
                 regs.SetRegisterToInteger(instr.gpr0, instr.conversion.is_output_signed, 0, op_a, 1,
                                           1, false, 0, instr.conversion.dest_size);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "F2I Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
                 break;
             }
             default: {
@@ -2658,12 +2746,12 @@ private:
                 }
                 case 3: {
                     if (is_array) {
-                        UNIMPLEMENTED_MSG("3-coordinate arrays not fully implemented");
-                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
-                        const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
-                        coord = "vec2 coords = vec2(" + x + ", " + y + ");";
-                        texture_type = Tegra::Shader::TextureType::Texture2D;
-                        is_array = false;
+                        const std::string index = regs.GetRegisterAsInteger(instr.gpr8);
+                        const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                        const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
+                        const std::string z = regs.GetRegisterAsFloat(instr.gpr20);
+                        coord =
+                            "vec4 coords = vec4(" + x + ", " + y + ", " + z + ", " + index + ");";
                     } else {
                         const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
                         const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
@@ -2693,7 +2781,11 @@ private:
                     break;
                 }
                 case Tegra::Shader::TextureProcessMode::LZ: {
-                    texture = "textureLod(" + sampler + ", coords, 0.0)";
+                    if (depth_compare && is_array) {
+                        texture = "texture(" + sampler + ", coords)";
+                    } else {
+                        texture = "textureLod(" + sampler + ", coords, 0.0)";
+                    }
                     break;
                 }
                 case Tegra::Shader::TextureProcessMode::LL: {
@@ -3107,6 +3199,11 @@ private:
                 regs.SetRegisterToFloat(instr.gpr0, 0, value, 1, 1);
             }
 
+            if (instr.generates_cc) {
+                LOG_CRITICAL(HW_GPU, "PSET Generates an unhandled Control Code");
+                UNREACHABLE();
+            }
+
             break;
         }
         case OpCode::Type::PredicateSetPredicate: {
@@ -3372,6 +3469,10 @@ private:
             }
 
             regs.SetRegisterToInteger(instr.gpr0, is_signed, 0, sum, 1, 1);
+            if (instr.generates_cc) {
+                LOG_CRITICAL(HW_GPU, "XMAD Generates an unhandled Control Code");
+                UNREACHABLE();
+            }
             break;
         }
         default: {
@@ -3381,6 +3482,12 @@ private:
                     EmitFragmentOutputsWrite();
                 }
 
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "EXIT Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
+
                 switch (instr.flow.cond) {
                 case Tegra::Shader::FlowCondition::Always:
                     shader.AddLine("return true;");
@@ -3410,6 +3517,11 @@ private:
 
                 // Enclose "discard" in a conditional, so that GLSL compilation does not complain
                 // about unexecuted instructions that may follow this.
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "KIL Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
                 shader.AddLine("if (true) {");
                 ++shader.scope;
                 shader.AddLine("discard;");
@@ -3467,6 +3579,11 @@ private:
             case OpCode::Id::BRA: {
                 ASSERT_MSG(instr.bra.constant_buffer == 0,
                            "BRA with constant buffers are not implemented");
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "BRA Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
                 const u32 target = offset + instr.bra.GetBranchTarget();
                 shader.AddLine("{ jmp_to = " + std::to_string(target) + "u; break; }");
                 break;
@@ -3507,13 +3624,21 @@ private:
             }
             case OpCode::Id::SYNC: {
                 // The SYNC opcode jumps to the address previously set by the SSY opcode
-                ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "SYNC Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
                 EmitPopFromFlowStack();
                 break;
             }
             case OpCode::Id::BRK: {
                 // The BRK opcode jumps to the address previously set by the PBK opcode
-                ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
+                const Tegra::Shader::ControlCode cc = instr.flow_control_code;
+                if (cc != Tegra::Shader::ControlCode::T) {
+                    LOG_CRITICAL(HW_GPU, "BRK Control Code used: {}", static_cast<u32>(cc));
+                    UNREACHABLE();
+                }
                 EmitPopFromFlowStack();
                 break;
             }
@@ -3543,6 +3668,11 @@ private:
                 regs.SetRegisterToInteger(instr.gpr0, result_signed, 1, result, 1, 1,
                                           instr.vmad.saturate == 1, 0, Register::Size::Word,
                                           instr.vmad.cc);
+                if (instr.generates_cc) {
+                    LOG_CRITICAL(HW_GPU, "VMAD Generates an unhandled Control Code");
+                    UNREACHABLE();
+                }
+
                 break;
             }
             case OpCode::Id::VSETP: {

src/video_core/renderer_opengl/gl_shader_gen.cpp

@@ -82,8 +82,8 @@ void main() {
 }
 
 ProgramResult GenerateGeometryShader(const ShaderSetup& setup) {
-    std::string out = "#version 430 core\n";
-    out += "#extension GL_ARB_separate_shader_objects : enable\n\n";
+    // Version is intentionally skipped in shader generation, it's added by the lazy compilation.
+    std::string out = "#extension GL_ARB_separate_shader_objects : enable\n\n";
     out += Decompiler::GetCommonDeclarations();
     out += "bool exec_geometry();\n";
 

src/video_core/renderer_opengl/gl_shader_manager.h

@@ -7,6 +7,7 @@
 #include <glad/glad.h>
 
 #include "video_core/renderer_opengl/gl_resource_manager.h"
+#include "video_core/renderer_opengl/gl_state.h"
 #include "video_core/renderer_opengl/maxwell_to_gl.h"
 
 namespace OpenGL::GLShader {

src/video_core/renderer_opengl/gl_state.cpp

@@ -22,17 +22,15 @@ OpenGLState::OpenGLState() {
     depth.test_enabled = false;
     depth.test_func = GL_LESS;
     depth.write_mask = GL_TRUE;
-    depth.depth_range_near = 0.0f;
-    depth.depth_range_far = 1.0f;
 
     primitive_restart.enabled = false;
     primitive_restart.index = 0;
-
-    color_mask.red_enabled = GL_TRUE;
-    color_mask.green_enabled = GL_TRUE;
-    color_mask.blue_enabled = GL_TRUE;
-    color_mask.alpha_enabled = GL_TRUE;
-
+    for (auto& item : color_mask) {
+        item.red_enabled = GL_TRUE;
+        item.green_enabled = GL_TRUE;
+        item.blue_enabled = GL_TRUE;
+        item.alpha_enabled = GL_TRUE;
+    }
     stencil.test_enabled = false;
     auto reset_stencil = [](auto& config) {
         config.test_func = GL_ALWAYS;
@@ -45,19 +43,33 @@ OpenGLState::OpenGLState() {
     };
     reset_stencil(stencil.front);
     reset_stencil(stencil.back);
-
-    blend.enabled = true;
-    blend.rgb_equation = GL_FUNC_ADD;
-    blend.a_equation = GL_FUNC_ADD;
-    blend.src_rgb_func = GL_ONE;
-    blend.dst_rgb_func = GL_ZERO;
-    blend.src_a_func = GL_ONE;
-    blend.dst_a_func = GL_ZERO;
-    blend.color.red = 0.0f;
-    blend.color.green = 0.0f;
-    blend.color.blue = 0.0f;
-    blend.color.alpha = 0.0f;
-
+    for (auto& item : viewports) {
+        item.x = 0;
+        item.y = 0;
+        item.width = 0;
+        item.height = 0;
+        item.depth_range_near = 0.0f;
+        item.depth_range_far = 1.0f;
+    }
+    scissor.enabled = false;
+    scissor.x = 0;
+    scissor.y = 0;
+    scissor.width = 0;
+    scissor.height = 0;
+    for (auto& item : blend) {
+        item.enabled = true;
+        item.rgb_equation = GL_FUNC_ADD;
+        item.a_equation = GL_FUNC_ADD;
+        item.src_rgb_func = GL_ONE;
+        item.dst_rgb_func = GL_ZERO;
+        item.src_a_func = GL_ONE;
+        item.dst_a_func = GL_ZERO;
+    }
+    independant_blend.enabled = false;
+    blend_color.red = 0.0f;
+    blend_color.green = 0.0f;
+    blend_color.blue = 0.0f;
+    blend_color.alpha = 0.0f;
     logic_op.enabled = false;
     logic_op.operation = GL_COPY;
 
@@ -73,17 +85,6 @@ OpenGLState::OpenGLState() {
     draw.shader_program = 0;
     draw.program_pipeline = 0;
 
-    scissor.enabled = false;
-    scissor.x = 0;
-    scissor.y = 0;
-    scissor.width = 0;
-    scissor.height = 0;
-
-    viewport.x = 0;
-    viewport.y = 0;
-    viewport.width = 0;
-    viewport.height = 0;
-
     clip_distance = {};
 
     point.size = 1;
@@ -134,6 +135,32 @@ void OpenGLState::ApplyCulling() const {
     }
 }
 
+void OpenGLState::ApplyColorMask() const {
+    if (GLAD_GL_ARB_viewport_array) {
+        for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
+            const auto& updated = color_mask[i];
+            const auto& current = cur_state.color_mask[i];
+            if (updated.red_enabled != current.red_enabled ||
+                updated.green_enabled != current.green_enabled ||
+                updated.blue_enabled != current.blue_enabled ||
+                updated.alpha_enabled != current.alpha_enabled) {
+                glColorMaski(static_cast<GLuint>(i), updated.red_enabled, updated.green_enabled,
+                             updated.blue_enabled, updated.alpha_enabled);
+            }
+        }
+    } else {
+        const auto& updated = color_mask[0];
+        const auto& current = cur_state.color_mask[0];
+        if (updated.red_enabled != current.red_enabled ||
+            updated.green_enabled != current.green_enabled ||
+            updated.blue_enabled != current.blue_enabled ||
+            updated.alpha_enabled != current.alpha_enabled) {
+            glColorMask(updated.red_enabled, updated.green_enabled, updated.blue_enabled,
+                        updated.alpha_enabled);
+        }
+    }
+}
+
 void OpenGLState::ApplyDepth() const {
     // Depth test
     const bool depth_test_changed = depth.test_enabled != cur_state.depth.test_enabled;
@@ -152,11 +179,6 @@ void OpenGLState::ApplyDepth() const {
     if (depth.write_mask != cur_state.depth.write_mask) {
         glDepthMask(depth.write_mask);
     }
-    // Depth range
-    if (depth.depth_range_near != cur_state.depth.depth_range_near ||
-        depth.depth_range_far != cur_state.depth.depth_range_far) {
-        glDepthRange(depth.depth_range_near, depth.depth_range_far);
-    }
 }
 
 void OpenGLState::ApplyPrimitiveRestart() const {
@@ -208,7 +230,7 @@ void OpenGLState::ApplyStencilTest() const {
     }
 }
 
-void OpenGLState::ApplyScissorTest() const {
+void OpenGLState::ApplyScissor() const {
     const bool scissor_changed = scissor.enabled != cur_state.scissor.enabled;
     if (scissor_changed) {
         if (scissor.enabled) {
@@ -217,51 +239,141 @@ void OpenGLState::ApplyScissorTest() const {
             glDisable(GL_SCISSOR_TEST);
         }
     }
-    if (scissor_changed || scissor_changed || scissor.x != cur_state.scissor.x ||
-        scissor.y != cur_state.scissor.y || scissor.width != cur_state.scissor.width ||
-        scissor.height != cur_state.scissor.height) {
+    if (scissor.enabled &&
+        (scissor_changed || scissor.x != cur_state.scissor.x || scissor.y != cur_state.scissor.y ||
+         scissor.width != cur_state.scissor.width || scissor.height != cur_state.scissor.height)) {
         glScissor(scissor.x, scissor.y, scissor.width, scissor.height);
     }
 }
 
-void OpenGLState::ApplyBlending() const {
-    const bool blend_changed = blend.enabled != cur_state.blend.enabled;
+void OpenGLState::ApplyViewport() const {
+    if (GLAD_GL_ARB_viewport_array) {
+        for (GLuint i = 0;
+             i < static_cast<GLuint>(Tegra::Engines::Maxwell3D::Regs::NumRenderTargets); i++) {
+            const auto& current = cur_state.viewports[i];
+            const auto& updated = viewports[i];
+            if (updated.x != current.x || updated.y != current.y ||
+                updated.width != current.width || updated.height != current.height) {
+                glViewportIndexedf(i, updated.x, updated.y, updated.width, updated.height);
+            }
+            if (updated.depth_range_near != current.depth_range_near ||
+                updated.depth_range_far != current.depth_range_far) {
+                glDepthRangeIndexed(i, updated.depth_range_near, updated.depth_range_far);
+            }
+        }
+    } else {
+        const auto& current = cur_state.viewports[0];
+        const auto& updated = viewports[0];
+        if (updated.x != current.x || updated.y != current.y || updated.width != current.width ||
+            updated.height != current.height) {
+            glViewport(updated.x, updated.y, updated.width, updated.height);
+        }
+        if (updated.depth_range_near != current.depth_range_near ||
+            updated.depth_range_far != current.depth_range_far) {
+            glDepthRange(updated.depth_range_near, updated.depth_range_far);
+        }
+    }
+}
+
+void OpenGLState::ApplyGlobalBlending() const {
+    const Blend& current = cur_state.blend[0];
+    const Blend& updated = blend[0];
+    const bool blend_changed = updated.enabled != current.enabled;
     if (blend_changed) {
-        if (blend.enabled) {
-            ASSERT(!logic_op.enabled);
+        if (updated.enabled) {
             glEnable(GL_BLEND);
         } else {
             glDisable(GL_BLEND);
         }
     }
-    if (blend.enabled) {
-        if (blend_changed || blend.color.red != cur_state.blend.color.red ||
-            blend.color.green != cur_state.blend.color.green ||
-            blend.color.blue != cur_state.blend.color.blue ||
-            blend.color.alpha != cur_state.blend.color.alpha) {
-            glBlendColor(blend.color.red, blend.color.green, blend.color.blue, blend.color.alpha);
+    if (!updated.enabled) {
+        return;
+    }
+    if (updated.separate_alpha) {
+        if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
+            updated.dst_rgb_func != current.dst_rgb_func ||
+            updated.src_a_func != current.src_a_func || updated.dst_a_func != current.dst_a_func) {
+            glBlendFuncSeparate(updated.src_rgb_func, updated.dst_rgb_func, updated.src_a_func,
+                                updated.dst_a_func);
         }
 
-        if (blend_changed || blend.src_rgb_func != cur_state.blend.src_rgb_func ||
-            blend.dst_rgb_func != cur_state.blend.dst_rgb_func ||
-            blend.src_a_func != cur_state.blend.src_a_func ||
-            blend.dst_a_func != cur_state.blend.dst_a_func) {
-            glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func, blend.src_a_func,
-                                blend.dst_a_func);
+        if (blend_changed || updated.rgb_equation != current.rgb_equation ||
+            updated.a_equation != current.a_equation) {
+            glBlendEquationSeparate(updated.rgb_equation, updated.a_equation);
+        }
+    } else {
+        if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
+            updated.dst_rgb_func != current.dst_rgb_func) {
+            glBlendFunc(updated.src_rgb_func, updated.dst_rgb_func);
         }
 
-        if (blend_changed || blend.rgb_equation != cur_state.blend.rgb_equation ||
-            blend.a_equation != cur_state.blend.a_equation) {
-            glBlendEquationSeparate(blend.rgb_equation, blend.a_equation);
+        if (blend_changed || updated.rgb_equation != current.rgb_equation) {
+            glBlendEquation(updated.rgb_equation);
         }
     }
 }
 
+void OpenGLState::ApplyTargetBlending(int target, bool force) const {
+    const Blend& updated = blend[target];
+    const Blend& current = cur_state.blend[target];
+    const bool blend_changed = updated.enabled != current.enabled || force;
+    if (blend_changed) {
+        if (updated.enabled) {
+            glEnablei(GL_BLEND, static_cast<GLuint>(target));
+        } else {
+            glDisablei(GL_BLEND, static_cast<GLuint>(target));
+        }
+    }
+    if (!updated.enabled) {
+        return;
+    }
+    if (updated.separate_alpha) {
+        if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
+            updated.dst_rgb_func != current.dst_rgb_func ||
+            updated.src_a_func != current.src_a_func || updated.dst_a_func != current.dst_a_func) {
+            glBlendFuncSeparateiARB(static_cast<GLuint>(target), updated.src_rgb_func,
+                                    updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func);
+        }
+
+        if (blend_changed || updated.rgb_equation != current.rgb_equation ||
+            updated.a_equation != current.a_equation) {
+            glBlendEquationSeparateiARB(static_cast<GLuint>(target), updated.rgb_equation,
+                                        updated.a_equation);
+        }
+    } else {
+        if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
+            updated.dst_rgb_func != current.dst_rgb_func) {
+            glBlendFunciARB(static_cast<GLuint>(target), updated.src_rgb_func,
+                            updated.dst_rgb_func);
+        }
+
+        if (blend_changed || updated.rgb_equation != current.rgb_equation) {
+            glBlendEquationiARB(static_cast<GLuint>(target), updated.rgb_equation);
+        }
+    }
+}
+
+void OpenGLState::ApplyBlending() const {
+    if (independant_blend.enabled) {
+        for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
+            ApplyTargetBlending(i,
+                                independant_blend.enabled != cur_state.independant_blend.enabled);
+        }
+    } else {
+        ApplyGlobalBlending();
+    }
+    if (blend_color.red != cur_state.blend_color.red ||
+        blend_color.green != cur_state.blend_color.green ||
+        blend_color.blue != cur_state.blend_color.blue ||
+        blend_color.alpha != cur_state.blend_color.alpha) {
+        glBlendColor(blend_color.red, blend_color.green, blend_color.blue, blend_color.alpha);
+    }
+}
+
 void OpenGLState::ApplyLogicOp() const {
     const bool logic_op_changed = logic_op.enabled != cur_state.logic_op.enabled;
     if (logic_op_changed) {
         if (logic_op.enabled) {
-            ASSERT(!blend.enabled);
             glEnable(GL_COLOR_LOGIC_OP);
         } else {
             glDisable(GL_COLOR_LOGIC_OP);
@@ -315,7 +427,7 @@ void OpenGLState::ApplySamplers() const {
     }
 }
 
-void OpenGLState::Apply() const {
+void OpenGLState::ApplyFramebufferState() const {
     // Framebuffer
     if (draw.read_framebuffer != cur_state.draw.read_framebuffer) {
         glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer);
@@ -323,7 +435,9 @@ void OpenGLState::Apply() const {
     if (draw.draw_framebuffer != cur_state.draw.draw_framebuffer) {
         glBindFramebuffer(GL_DRAW_FRAMEBUFFER, draw.draw_framebuffer);
     }
+}
 
+void OpenGLState::ApplyVertexBufferState() const {
     // Vertex array
     if (draw.vertex_array != cur_state.draw.vertex_array) {
         glBindVertexArray(draw.vertex_array);
@@ -333,7 +447,11 @@ void OpenGLState::Apply() const {
     if (draw.vertex_buffer != cur_state.draw.vertex_buffer) {
         glBindBuffer(GL_ARRAY_BUFFER, draw.vertex_buffer);
     }
+}
 
+void OpenGLState::Apply() const {
+    ApplyFramebufferState();
+    ApplyVertexBufferState();
     // Uniform buffer
     if (draw.uniform_buffer != cur_state.draw.uniform_buffer) {
         glBindBuffer(GL_UNIFORM_BUFFER, draw.uniform_buffer);
@@ -348,12 +466,6 @@ void OpenGLState::Apply() const {
     if (draw.program_pipeline != cur_state.draw.program_pipeline) {
         glBindProgramPipeline(draw.program_pipeline);
     }
-    // Viewport
-    if (viewport.x != cur_state.viewport.x || viewport.y != cur_state.viewport.y ||
-        viewport.width != cur_state.viewport.width ||
-        viewport.height != cur_state.viewport.height) {
-        glViewport(viewport.x, viewport.y, viewport.width, viewport.height);
-    }
     // Clip distance
     for (std::size_t i = 0; i < clip_distance.size(); ++i) {
         if (clip_distance[i] != cur_state.clip_distance[i]) {
@@ -364,19 +476,13 @@ void OpenGLState::Apply() const {
             }
         }
     }
-    // Color mask
-    if (color_mask.red_enabled != cur_state.color_mask.red_enabled ||
-        color_mask.green_enabled != cur_state.color_mask.green_enabled ||
-        color_mask.blue_enabled != cur_state.color_mask.blue_enabled ||
-        color_mask.alpha_enabled != cur_state.color_mask.alpha_enabled) {
-        glColorMask(color_mask.red_enabled, color_mask.green_enabled, color_mask.blue_enabled,
-                    color_mask.alpha_enabled);
-    }
     // Point
     if (point.size != cur_state.point.size) {
         glPointSize(point.size);
     }
-    ApplyScissorTest();
+    ApplyColorMask();
+    ApplyViewport();
+    ApplyScissor();
     ApplyStencilTest();
     ApplySRgb();
     ApplyCulling();

src/video_core/renderer_opengl/gl_state.h

@@ -46,11 +46,9 @@ public:
     } cull;
 
     struct {
-        bool test_enabled;        // GL_DEPTH_TEST
-        GLenum test_func;         // GL_DEPTH_FUNC
-        GLboolean write_mask;     // GL_DEPTH_WRITEMASK
-        GLfloat depth_range_near; // GL_DEPTH_RANGE
-        GLfloat depth_range_far;  // GL_DEPTH_RANGE
+        bool test_enabled;    // GL_DEPTH_TEST
+        GLenum test_func;     // GL_DEPTH_FUNC
+        GLboolean write_mask; // GL_DEPTH_WRITEMASK
     } depth;
 
     struct {
@@ -58,13 +56,14 @@ public:
         GLuint index;
     } primitive_restart; // GL_PRIMITIVE_RESTART
 
-    struct {
+    struct ColorMask {
         GLboolean red_enabled;
         GLboolean green_enabled;
         GLboolean blue_enabled;
         GLboolean alpha_enabled;
-    } color_mask; // GL_COLOR_WRITEMASK
-
+    };
+    std::array<ColorMask, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets>
+        color_mask; // GL_COLOR_WRITEMASK
     struct {
         bool test_enabled; // GL_STENCIL_TEST
         struct {
@@ -78,22 +77,28 @@ public:
         } front, back;
     } stencil;
 
-    struct {
+    struct Blend {
         bool enabled;        // GL_BLEND
+        bool separate_alpha; // Independent blend enabled
         GLenum rgb_equation; // GL_BLEND_EQUATION_RGB
         GLenum a_equation;   // GL_BLEND_EQUATION_ALPHA
         GLenum src_rgb_func; // GL_BLEND_SRC_RGB
         GLenum dst_rgb_func; // GL_BLEND_DST_RGB
         GLenum src_a_func;   // GL_BLEND_SRC_ALPHA
         GLenum dst_a_func;   // GL_BLEND_DST_ALPHA
+    };
+    std::array<Blend, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> blend;
 
-        struct {
-            GLclampf red;
-            GLclampf green;
-            GLclampf blue;
-            GLclampf alpha;
-        } color; // GL_BLEND_COLOR
-    } blend;
+    struct {
+        bool enabled;
+    } independant_blend;
+
+    struct {
+        GLclampf red;
+        GLclampf green;
+        GLclampf blue;
+        GLclampf alpha;
+    } blend_color; // GL_BLEND_COLOR
 
     struct {
         bool enabled; // GL_LOGIC_OP_MODE
@@ -138,6 +143,16 @@ public:
         GLuint program_pipeline; // GL_PROGRAM_PIPELINE_BINDING
     } draw;
 
+    struct viewport {
+        GLfloat x;
+        GLfloat y;
+        GLfloat width;
+        GLfloat height;
+        GLfloat depth_range_near; // GL_DEPTH_RANGE
+        GLfloat depth_range_far;  // GL_DEPTH_RANGE
+    };
+    std::array<viewport, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> viewports;
+
     struct {
         bool enabled; // GL_SCISSOR_TEST
         GLint x;
@@ -146,13 +161,6 @@ public:
         GLsizei height;
     } scissor;
 
-    struct {
-        GLint x;
-        GLint y;
-        GLsizei width;
-        GLsizei height;
-    } viewport;
-
     struct {
         float size; // GL_POINT_SIZE
     } point;
@@ -173,6 +181,10 @@ public:
     }
     /// Apply this state as the current OpenGL state
     void Apply() const;
+    /// Apply only the state afecting the framebuffer
+    void ApplyFramebufferState() const;
+    /// Apply only the state afecting the vertex buffer
+    void ApplyVertexBufferState() const;
     /// Set the initial OpenGL state
     static void ApplyDefaultState();
     /// Resets any references to the given resource
@@ -191,14 +203,18 @@ private:
     static bool s_rgb_used;
     void ApplySRgb() const;
     void ApplyCulling() const;
+    void ApplyColorMask() const;
     void ApplyDepth() const;
     void ApplyPrimitiveRestart() const;
     void ApplyStencilTest() const;
-    void ApplyScissorTest() const;
+    void ApplyViewport() const;
+    void ApplyTargetBlending(int target, bool force) const;
+    void ApplyGlobalBlending() const;
     void ApplyBlending() const;
     void ApplyLogicOp() const;
     void ApplyTextures() const;
     void ApplySamplers() const;
+    void ApplyScissor() const;
 };
 
 } // namespace OpenGL

src/video_core/renderer_opengl/gl_stream_buffer.cpp

@@ -6,9 +6,13 @@
 #include <vector>
 #include "common/alignment.h"
 #include "common/assert.h"
+#include "common/microprofile.h"
 #include "video_core/renderer_opengl/gl_state.h"
 #include "video_core/renderer_opengl/gl_stream_buffer.h"
 
+MICROPROFILE_DEFINE(OpenGL_StreamBuffer, "OpenGL", "Stream Buffer Orphaning",
+                    MP_RGB(128, 128, 192));
+
 namespace OpenGL {
 
 OGLStreamBuffer::OGLStreamBuffer(GLenum target, GLsizeiptr size, bool prefer_coherent)
@@ -75,6 +79,7 @@ std::tuple<u8*, GLintptr, bool> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr a
     }
 
     if (invalidate || !persistent) {
+        MICROPROFILE_SCOPE(OpenGL_StreamBuffer);
         GLbitfield flags = GL_MAP_WRITE_BIT | (persistent ? GL_MAP_PERSISTENT_BIT : 0) |
                            (coherent ? GL_MAP_COHERENT_BIT : GL_MAP_FLUSH_EXPLICIT_BIT) |
                            (invalidate ? GL_MAP_INVALIDATE_BUFFER_BIT : GL_MAP_UNSYNCHRONIZED_BIT);

src/video_core/renderer_opengl/maxwell_to_gl.h

@@ -159,10 +159,8 @@ inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode,
         }
     }
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented texture filter mode={}",
-                 static_cast<u32>(filter_mode));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented texture filter mode={}", static_cast<u32>(filter_mode));
+    return GL_LINEAR;
 }
 
 inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) {
@@ -183,9 +181,8 @@ inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) {
     case Tegra::Texture::WrapMode::MirrorOnceClampToEdge:
         return GL_MIRROR_CLAMP_TO_EDGE;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode));
+    return GL_REPEAT;
 }
 
 inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) {
@@ -207,10 +204,9 @@ inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) {
     case Tegra::Texture::DepthCompareFunc::Always:
         return GL_ALWAYS;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented texture depth compare function ={}",
-                 static_cast<u32>(func));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented texture depth compare function ={}",
+              static_cast<u32>(func));
+    return GL_GREATER;
 }
 
 inline GLenum BlendEquation(Maxwell::Blend::Equation equation) {
@@ -226,9 +222,8 @@ inline GLenum BlendEquation(Maxwell::Blend::Equation equation) {
     case Maxwell::Blend::Equation::Max:
         return GL_MAX;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented blend equation={}", static_cast<u32>(equation));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented blend equation={}", static_cast<u32>(equation));
+    return GL_FUNC_ADD;
 }
 
 inline GLenum BlendFunc(Maxwell::Blend::Factor factor) {
@@ -291,9 +286,8 @@ inline GLenum BlendFunc(Maxwell::Blend::Factor factor) {
     case Maxwell::Blend::Factor::OneMinusConstantAlphaGL:
         return GL_ONE_MINUS_CONSTANT_ALPHA;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented blend factor={}", static_cast<u32>(factor));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented blend factor={}", static_cast<u32>(factor));
+    return GL_ZERO;
 }
 
 inline GLenum SwizzleSource(Tegra::Texture::SwizzleSource source) {
@@ -312,9 +306,8 @@ inline GLenum SwizzleSource(Tegra::Texture::SwizzleSource source) {
     case Tegra::Texture::SwizzleSource::OneFloat:
         return GL_ONE;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented swizzle source={}", static_cast<u32>(source));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented swizzle source={}", static_cast<u32>(source));
+    return GL_ZERO;
 }
 
 inline GLenum ComparisonOp(Maxwell::ComparisonOp comparison) {
@@ -344,33 +337,39 @@ inline GLenum ComparisonOp(Maxwell::ComparisonOp comparison) {
     case Maxwell::ComparisonOp::AlwaysOld:
         return GL_ALWAYS;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented comparison op={}", static_cast<u32>(comparison));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented comparison op={}", static_cast<u32>(comparison));
+    return GL_ALWAYS;
 }
 
 inline GLenum StencilOp(Maxwell::StencilOp stencil) {
     switch (stencil) {
     case Maxwell::StencilOp::Keep:
+    case Maxwell::StencilOp::KeepOGL:
         return GL_KEEP;
     case Maxwell::StencilOp::Zero:
+    case Maxwell::StencilOp::ZeroOGL:
         return GL_ZERO;
     case Maxwell::StencilOp::Replace:
+    case Maxwell::StencilOp::ReplaceOGL:
         return GL_REPLACE;
     case Maxwell::StencilOp::Incr:
+    case Maxwell::StencilOp::IncrOGL:
         return GL_INCR;
     case Maxwell::StencilOp::Decr:
+    case Maxwell::StencilOp::DecrOGL:
         return GL_DECR;
     case Maxwell::StencilOp::Invert:
+    case Maxwell::StencilOp::InvertOGL:
         return GL_INVERT;
     case Maxwell::StencilOp::IncrWrap:
+    case Maxwell::StencilOp::IncrWrapOGL:
         return GL_INCR_WRAP;
     case Maxwell::StencilOp::DecrWrap:
+    case Maxwell::StencilOp::DecrWrapOGL:
         return GL_DECR_WRAP;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented stencil op={}", static_cast<u32>(stencil));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented stencil op={}", static_cast<u32>(stencil));
+    return GL_KEEP;
 }
 
 inline GLenum FrontFace(Maxwell::Cull::FrontFace front_face) {
@@ -380,9 +379,8 @@ inline GLenum FrontFace(Maxwell::Cull::FrontFace front_face) {
     case Maxwell::Cull::FrontFace::CounterClockWise:
         return GL_CCW;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented front face cull={}", static_cast<u32>(front_face));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented front face cull={}", static_cast<u32>(front_face));
+    return GL_CCW;
 }
 
 inline GLenum CullFace(Maxwell::Cull::CullFace cull_face) {
@@ -394,9 +392,8 @@ inline GLenum CullFace(Maxwell::Cull::CullFace cull_face) {
     case Maxwell::Cull::CullFace::FrontAndBack:
         return GL_FRONT_AND_BACK;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented cull face={}", static_cast<u32>(cull_face));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented cull face={}", static_cast<u32>(cull_face));
+    return GL_BACK;
 }
 
 inline GLenum LogicOp(Maxwell::LogicOperation operation) {
@@ -434,9 +431,8 @@ inline GLenum LogicOp(Maxwell::LogicOperation operation) {
     case Maxwell::LogicOperation::Set:
         return GL_SET;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented logic operation={}", static_cast<u32>(operation));
-    UNREACHABLE();
-    return {};
+    LOG_ERROR(Render_OpenGL, "Unimplemented logic operation={}", static_cast<u32>(operation));
+    return GL_COPY;
 }
 
 } // namespace MaxwellToGL

src/video_core/surface.cpp

@@ -0,0 +1,490 @@
+// Copyright 2014 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/common_types.h"
+#include "common/math_util.h"
+#include "video_core/surface.h"
+
+namespace VideoCore::Surface {
+
+SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type) {
+    switch (texture_type) {
+    case Tegra::Texture::TextureType::Texture1D:
+        return SurfaceTarget::Texture1D;
+    case Tegra::Texture::TextureType::Texture2D:
+    case Tegra::Texture::TextureType::Texture2DNoMipmap:
+        return SurfaceTarget::Texture2D;
+    case Tegra::Texture::TextureType::Texture3D:
+        return SurfaceTarget::Texture3D;
+    case Tegra::Texture::TextureType::TextureCubemap:
+        return SurfaceTarget::TextureCubemap;
+    case Tegra::Texture::TextureType::TextureCubeArray:
+        return SurfaceTarget::TextureCubeArray;
+    case Tegra::Texture::TextureType::Texture1DArray:
+        return SurfaceTarget::Texture1DArray;
+    case Tegra::Texture::TextureType::Texture2DArray:
+        return SurfaceTarget::Texture2DArray;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", static_cast<u32>(texture_type));
+        UNREACHABLE();
+        return SurfaceTarget::Texture2D;
+    }
+}
+
+bool SurfaceTargetIsLayered(SurfaceTarget target) {
+    switch (target) {
+    case SurfaceTarget::Texture1D:
+    case SurfaceTarget::Texture2D:
+    case SurfaceTarget::Texture3D:
+        return false;
+    case SurfaceTarget::Texture1DArray:
+    case SurfaceTarget::Texture2DArray:
+    case SurfaceTarget::TextureCubemap:
+    case SurfaceTarget::TextureCubeArray:
+        return true;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
+        UNREACHABLE();
+        return false;
+    }
+}
+
+PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
+    switch (format) {
+    case Tegra::DepthFormat::S8_Z24_UNORM:
+        return PixelFormat::S8Z24;
+    case Tegra::DepthFormat::Z24_S8_UNORM:
+        return PixelFormat::Z24S8;
+    case Tegra::DepthFormat::Z32_FLOAT:
+        return PixelFormat::Z32F;
+    case Tegra::DepthFormat::Z16_UNORM:
+        return PixelFormat::Z16;
+    case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
+        return PixelFormat::Z32FS8;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
+    switch (format) {
+        // TODO (Hexagon12): Converting SRGBA to RGBA is a hack and doesn't completely correct the
+        // gamma.
+    case Tegra::RenderTargetFormat::RGBA8_SRGB:
+        return PixelFormat::RGBA8_SRGB;
+    case Tegra::RenderTargetFormat::RGBA8_UNORM:
+        return PixelFormat::ABGR8U;
+    case Tegra::RenderTargetFormat::RGBA8_SNORM:
+        return PixelFormat::ABGR8S;
+    case Tegra::RenderTargetFormat::RGBA8_UINT:
+        return PixelFormat::ABGR8UI;
+    case Tegra::RenderTargetFormat::BGRA8_SRGB:
+        return PixelFormat::BGRA8_SRGB;
+    case Tegra::RenderTargetFormat::BGRA8_UNORM:
+        return PixelFormat::BGRA8;
+    case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
+        return PixelFormat::A2B10G10R10U;
+    case Tegra::RenderTargetFormat::RGBA16_FLOAT:
+        return PixelFormat::RGBA16F;
+    case Tegra::RenderTargetFormat::RGBA16_UNORM:
+        return PixelFormat::RGBA16U;
+    case Tegra::RenderTargetFormat::RGBA16_UINT:
+        return PixelFormat::RGBA16UI;
+    case Tegra::RenderTargetFormat::RGBA32_FLOAT:
+        return PixelFormat::RGBA32F;
+    case Tegra::RenderTargetFormat::RG32_FLOAT:
+        return PixelFormat::RG32F;
+    case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
+        return PixelFormat::R11FG11FB10F;
+    case Tegra::RenderTargetFormat::B5G6R5_UNORM:
+        return PixelFormat::B5G6R5U;
+    case Tegra::RenderTargetFormat::BGR5A1_UNORM:
+        return PixelFormat::A1B5G5R5U;
+    case Tegra::RenderTargetFormat::RGBA32_UINT:
+        return PixelFormat::RGBA32UI;
+    case Tegra::RenderTargetFormat::R8_UNORM:
+        return PixelFormat::R8U;
+    case Tegra::RenderTargetFormat::R8_UINT:
+        return PixelFormat::R8UI;
+    case Tegra::RenderTargetFormat::RG16_FLOAT:
+        return PixelFormat::RG16F;
+    case Tegra::RenderTargetFormat::RG16_UINT:
+        return PixelFormat::RG16UI;
+    case Tegra::RenderTargetFormat::RG16_SINT:
+        return PixelFormat::RG16I;
+    case Tegra::RenderTargetFormat::RG16_UNORM:
+        return PixelFormat::RG16;
+    case Tegra::RenderTargetFormat::RG16_SNORM:
+        return PixelFormat::RG16S;
+    case Tegra::RenderTargetFormat::RG8_UNORM:
+        return PixelFormat::RG8U;
+    case Tegra::RenderTargetFormat::RG8_SNORM:
+        return PixelFormat::RG8S;
+    case Tegra::RenderTargetFormat::R16_FLOAT:
+        return PixelFormat::R16F;
+    case Tegra::RenderTargetFormat::R16_UNORM:
+        return PixelFormat::R16U;
+    case Tegra::RenderTargetFormat::R16_SNORM:
+        return PixelFormat::R16S;
+    case Tegra::RenderTargetFormat::R16_UINT:
+        return PixelFormat::R16UI;
+    case Tegra::RenderTargetFormat::R16_SINT:
+        return PixelFormat::R16I;
+    case Tegra::RenderTargetFormat::R32_FLOAT:
+        return PixelFormat::R32F;
+    case Tegra::RenderTargetFormat::R32_UINT:
+        return PixelFormat::R32UI;
+    case Tegra::RenderTargetFormat::RG32_UINT:
+        return PixelFormat::RG32UI;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
+                                         Tegra::Texture::ComponentType component_type,
+                                         bool is_srgb) {
+    // TODO(Subv): Properly implement this
+    switch (format) {
+    case Tegra::Texture::TextureFormat::A8R8G8B8:
+        if (is_srgb) {
+            return PixelFormat::RGBA8_SRGB;
+        }
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::ABGR8U;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::ABGR8S;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::ABGR8UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::B5G6R5:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::B5G6R5U;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::A2B10G10R10:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::A2B10G10R10U;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::A1B5G5R5:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::A1B5G5R5U;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R8:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::R8U;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::R8UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::G8R8:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::G8R8U;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::G8R8S;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R16_G16_B16_A16:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::RGBA16U;
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RGBA16F;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::BF10GF11RF11:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::R11FG11FB10F;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R32_G32_B32_A32:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RGBA32F;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::RGBA32UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R32_G32:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RG32F;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::RG32UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R32_G32_B32:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RGB32F;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R16:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::R16F;
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::R16U;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::R16S;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::R16UI;
+        case Tegra::Texture::ComponentType::SINT:
+            return PixelFormat::R16I;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::R32:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::R32F;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::R32UI;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::ZF32:
+        return PixelFormat::Z32F;
+    case Tegra::Texture::TextureFormat::Z16:
+        return PixelFormat::Z16;
+    case Tegra::Texture::TextureFormat::Z24S8:
+        return PixelFormat::Z24S8;
+    case Tegra::Texture::TextureFormat::DXT1:
+        return is_srgb ? PixelFormat::DXT1_SRGB : PixelFormat::DXT1;
+    case Tegra::Texture::TextureFormat::DXT23:
+        return is_srgb ? PixelFormat::DXT23_SRGB : PixelFormat::DXT23;
+    case Tegra::Texture::TextureFormat::DXT45:
+        return is_srgb ? PixelFormat::DXT45_SRGB : PixelFormat::DXT45;
+    case Tegra::Texture::TextureFormat::DXN1:
+        return PixelFormat::DXN1;
+    case Tegra::Texture::TextureFormat::DXN2:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::DXN2UNORM;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::DXN2SNORM;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    case Tegra::Texture::TextureFormat::BC7U:
+        return is_srgb ? PixelFormat::BC7U_SRGB : PixelFormat::BC7U;
+    case Tegra::Texture::TextureFormat::BC6H_UF16:
+        return PixelFormat::BC6H_UF16;
+    case Tegra::Texture::TextureFormat::BC6H_SF16:
+        return PixelFormat::BC6H_SF16;
+    case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
+        return is_srgb ? PixelFormat::ASTC_2D_4X4_SRGB : PixelFormat::ASTC_2D_4X4;
+    case Tegra::Texture::TextureFormat::ASTC_2D_5X4:
+        return is_srgb ? PixelFormat::ASTC_2D_5X4_SRGB : PixelFormat::ASTC_2D_5X4;
+    case Tegra::Texture::TextureFormat::ASTC_2D_5X5:
+        return is_srgb ? PixelFormat::ASTC_2D_5X5_SRGB : PixelFormat::ASTC_2D_5X5;
+    case Tegra::Texture::TextureFormat::ASTC_2D_8X8:
+        return is_srgb ? PixelFormat::ASTC_2D_8X8_SRGB : PixelFormat::ASTC_2D_8X8;
+    case Tegra::Texture::TextureFormat::ASTC_2D_8X5:
+        return is_srgb ? PixelFormat::ASTC_2D_8X5_SRGB : PixelFormat::ASTC_2D_8X5;
+    case Tegra::Texture::TextureFormat::ASTC_2D_10X8:
+        return is_srgb ? PixelFormat::ASTC_2D_10X8_SRGB : PixelFormat::ASTC_2D_10X8;
+    case Tegra::Texture::TextureFormat::R16_G16:
+        switch (component_type) {
+        case Tegra::Texture::ComponentType::FLOAT:
+            return PixelFormat::RG16F;
+        case Tegra::Texture::ComponentType::UNORM:
+            return PixelFormat::RG16;
+        case Tegra::Texture::ComponentType::SNORM:
+            return PixelFormat::RG16S;
+        case Tegra::Texture::ComponentType::UINT:
+            return PixelFormat::RG16UI;
+        case Tegra::Texture::ComponentType::SINT:
+            return PixelFormat::RG16I;
+        }
+        LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
+        UNREACHABLE();
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}, component_type={}", static_cast<u32>(format),
+                     static_cast<u32>(component_type));
+        UNREACHABLE();
+    }
+}
+
+ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
+    // TODO(Subv): Implement more component types
+    switch (type) {
+    case Tegra::Texture::ComponentType::UNORM:
+        return ComponentType::UNorm;
+    case Tegra::Texture::ComponentType::FLOAT:
+        return ComponentType::Float;
+    case Tegra::Texture::ComponentType::SNORM:
+        return ComponentType::SNorm;
+    case Tegra::Texture::ComponentType::UINT:
+        return ComponentType::UInt;
+    case Tegra::Texture::ComponentType::SINT:
+        return ComponentType::SInt;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented component type={}", static_cast<u32>(type));
+        UNREACHABLE();
+    }
+}
+
+ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format) {
+    // TODO(Subv): Implement more render targets
+    switch (format) {
+    case Tegra::RenderTargetFormat::RGBA8_UNORM:
+    case Tegra::RenderTargetFormat::RGBA8_SRGB:
+    case Tegra::RenderTargetFormat::BGRA8_UNORM:
+    case Tegra::RenderTargetFormat::BGRA8_SRGB:
+    case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
+    case Tegra::RenderTargetFormat::R8_UNORM:
+    case Tegra::RenderTargetFormat::RG16_UNORM:
+    case Tegra::RenderTargetFormat::R16_UNORM:
+    case Tegra::RenderTargetFormat::B5G6R5_UNORM:
+    case Tegra::RenderTargetFormat::BGR5A1_UNORM:
+    case Tegra::RenderTargetFormat::RG8_UNORM:
+    case Tegra::RenderTargetFormat::RGBA16_UNORM:
+        return ComponentType::UNorm;
+    case Tegra::RenderTargetFormat::RGBA8_SNORM:
+    case Tegra::RenderTargetFormat::RG16_SNORM:
+    case Tegra::RenderTargetFormat::R16_SNORM:
+    case Tegra::RenderTargetFormat::RG8_SNORM:
+        return ComponentType::SNorm;
+    case Tegra::RenderTargetFormat::RGBA16_FLOAT:
+    case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
+    case Tegra::RenderTargetFormat::RGBA32_FLOAT:
+    case Tegra::RenderTargetFormat::RG32_FLOAT:
+    case Tegra::RenderTargetFormat::RG16_FLOAT:
+    case Tegra::RenderTargetFormat::R16_FLOAT:
+    case Tegra::RenderTargetFormat::R32_FLOAT:
+        return ComponentType::Float;
+    case Tegra::RenderTargetFormat::RGBA32_UINT:
+    case Tegra::RenderTargetFormat::RGBA16_UINT:
+    case Tegra::RenderTargetFormat::RG16_UINT:
+    case Tegra::RenderTargetFormat::R8_UINT:
+    case Tegra::RenderTargetFormat::R16_UINT:
+    case Tegra::RenderTargetFormat::RG32_UINT:
+    case Tegra::RenderTargetFormat::R32_UINT:
+    case Tegra::RenderTargetFormat::RGBA8_UINT:
+        return ComponentType::UInt;
+    case Tegra::RenderTargetFormat::RG16_SINT:
+    case Tegra::RenderTargetFormat::R16_SINT:
+        return ComponentType::SInt;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
+    switch (format) {
+    case Tegra::FramebufferConfig::PixelFormat::ABGR8:
+        return PixelFormat::ABGR8U;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format) {
+    switch (format) {
+    case Tegra::DepthFormat::Z16_UNORM:
+    case Tegra::DepthFormat::S8_Z24_UNORM:
+    case Tegra::DepthFormat::Z24_S8_UNORM:
+        return ComponentType::UNorm;
+    case Tegra::DepthFormat::Z32_FLOAT:
+    case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
+        return ComponentType::Float;
+    default:
+        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+        UNREACHABLE();
+    }
+}
+
+SurfaceType GetFormatType(PixelFormat pixel_format) {
+    if (static_cast<std::size_t>(pixel_format) <
+        static_cast<std::size_t>(PixelFormat::MaxColorFormat)) {
+        return SurfaceType::ColorTexture;
+    }
+
+    if (static_cast<std::size_t>(pixel_format) <
+        static_cast<std::size_t>(PixelFormat::MaxDepthFormat)) {
+        return SurfaceType::Depth;
+    }
+
+    if (static_cast<std::size_t>(pixel_format) <
+        static_cast<std::size_t>(PixelFormat::MaxDepthStencilFormat)) {
+        return SurfaceType::DepthStencil;
+    }
+
+    // TODO(Subv): Implement the other formats
+    ASSERT(false);
+
+    return SurfaceType::Invalid;
+}
+
+bool IsPixelFormatASTC(PixelFormat format) {
+    switch (format) {
+    case PixelFormat::ASTC_2D_4X4:
+    case PixelFormat::ASTC_2D_5X4:
+    case PixelFormat::ASTC_2D_5X5:
+    case PixelFormat::ASTC_2D_8X8:
+    case PixelFormat::ASTC_2D_8X5:
+    case PixelFormat::ASTC_2D_4X4_SRGB:
+    case PixelFormat::ASTC_2D_5X4_SRGB:
+    case PixelFormat::ASTC_2D_5X5_SRGB:
+    case PixelFormat::ASTC_2D_8X8_SRGB:
+    case PixelFormat::ASTC_2D_8X5_SRGB:
+    case PixelFormat::ASTC_2D_10X8:
+    case PixelFormat::ASTC_2D_10X8_SRGB:
+        return true;
+    default:
+        return false;
+    }
+}
+
+std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) {
+    return {GetDefaultBlockWidth(format), GetDefaultBlockHeight(format)};
+}
+
+bool IsFormatBCn(PixelFormat format) {
+    switch (format) {
+    case PixelFormat::DXT1:
+    case PixelFormat::DXT23:
+    case PixelFormat::DXT45:
+    case PixelFormat::DXN1:
+    case PixelFormat::DXN2SNORM:
+    case PixelFormat::DXN2UNORM:
+    case PixelFormat::BC7U:
+    case PixelFormat::BC6H_UF16:
+    case PixelFormat::BC6H_SF16:
+    case PixelFormat::DXT1_SRGB:
+    case PixelFormat::DXT23_SRGB:
+    case PixelFormat::DXT45_SRGB:
+    case PixelFormat::BC7U_SRGB:
+        return true;
+    }
+    return false;
+}
+
+} // namespace VideoCore::Surface

src/video_core/surface.h

@@ -0,0 +1,477 @@
+// Copyright 2014 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <climits>
+#include <utility>
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "common/logging/log.h"
+#include "video_core/gpu.h"
+#include "video_core/textures/texture.h"
+
+namespace VideoCore::Surface {
+
+enum class PixelFormat {
+    ABGR8U = 0,
+    ABGR8S = 1,
+    ABGR8UI = 2,
+    B5G6R5U = 3,
+    A2B10G10R10U = 4,
+    A1B5G5R5U = 5,
+    R8U = 6,
+    R8UI = 7,
+    RGBA16F = 8,
+    RGBA16U = 9,
+    RGBA16UI = 10,
+    R11FG11FB10F = 11,
+    RGBA32UI = 12,
+    DXT1 = 13,
+    DXT23 = 14,
+    DXT45 = 15,
+    DXN1 = 16, // This is also known as BC4
+    DXN2UNORM = 17,
+    DXN2SNORM = 18,
+    BC7U = 19,
+    BC6H_UF16 = 20,
+    BC6H_SF16 = 21,
+    ASTC_2D_4X4 = 22,
+    G8R8U = 23,
+    G8R8S = 24,
+    BGRA8 = 25,
+    RGBA32F = 26,
+    RG32F = 27,
+    R32F = 28,
+    R16F = 29,
+    R16U = 30,
+    R16S = 31,
+    R16UI = 32,
+    R16I = 33,
+    RG16 = 34,
+    RG16F = 35,
+    RG16UI = 36,
+    RG16I = 37,
+    RG16S = 38,
+    RGB32F = 39,
+    RGBA8_SRGB = 40,
+    RG8U = 41,
+    RG8S = 42,
+    RG32UI = 43,
+    R32UI = 44,
+    ASTC_2D_8X8 = 45,
+    ASTC_2D_8X5 = 46,
+    ASTC_2D_5X4 = 47,
+    BGRA8_SRGB = 48,
+    DXT1_SRGB = 49,
+    DXT23_SRGB = 50,
+    DXT45_SRGB = 51,
+    BC7U_SRGB = 52,
+    ASTC_2D_4X4_SRGB = 53,
+    ASTC_2D_8X8_SRGB = 54,
+    ASTC_2D_8X5_SRGB = 55,
+    ASTC_2D_5X4_SRGB = 56,
+    ASTC_2D_5X5 = 57,
+    ASTC_2D_5X5_SRGB = 58,
+    ASTC_2D_10X8 = 59,
+    ASTC_2D_10X8_SRGB = 60,
+
+    MaxColorFormat,
+
+    // Depth formats
+    Z32F = 61,
+    Z16 = 62,
+
+    MaxDepthFormat,
+
+    // DepthStencil formats
+    Z24S8 = 63,
+    S8Z24 = 64,
+    Z32FS8 = 65,
+
+    MaxDepthStencilFormat,
+
+    Max = MaxDepthStencilFormat,
+    Invalid = 255,
+};
+
+static constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max);
+
+enum class ComponentType {
+    Invalid = 0,
+    SNorm = 1,
+    UNorm = 2,
+    SInt = 3,
+    UInt = 4,
+    Float = 5,
+};
+
+enum class SurfaceType {
+    ColorTexture = 0,
+    Depth = 1,
+    DepthStencil = 2,
+    Fill = 3,
+    Invalid = 4,
+};
+
+enum class SurfaceTarget {
+    Texture1D,
+    Texture2D,
+    Texture3D,
+    Texture1DArray,
+    Texture2DArray,
+    TextureCubemap,
+    TextureCubeArray,
+};
+
+/**
+ * Gets the compression factor for the specified PixelFormat. This applies to just the
+ * "compressed width" and "compressed height", not the overall compression factor of a
+ * compressed image. This is used for maintaining proper surface sizes for compressed
+ * texture formats.
+ */
+static constexpr u32 GetCompressionFactor(PixelFormat format) {
+    if (format == PixelFormat::Invalid)
+        return 0;
+
+    constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{
+        1, // ABGR8U
+        1, // ABGR8S
+        1, // ABGR8UI
+        1, // B5G6R5U
+        1, // A2B10G10R10U
+        1, // A1B5G5R5U
+        1, // R8U
+        1, // R8UI
+        1, // RGBA16F
+        1, // RGBA16U
+        1, // RGBA16UI
+        1, // R11FG11FB10F
+        1, // RGBA32UI
+        4, // DXT1
+        4, // DXT23
+        4, // DXT45
+        4, // DXN1
+        4, // DXN2UNORM
+        4, // DXN2SNORM
+        4, // BC7U
+        4, // BC6H_UF16
+        4, // BC6H_SF16
+        4, // ASTC_2D_4X4
+        1, // G8R8U
+        1, // G8R8S
+        1, // BGRA8
+        1, // RGBA32F
+        1, // RG32F
+        1, // R32F
+        1, // R16F
+        1, // R16U
+        1, // R16S
+        1, // R16UI
+        1, // R16I
+        1, // RG16
+        1, // RG16F
+        1, // RG16UI
+        1, // RG16I
+        1, // RG16S
+        1, // RGB32F
+        1, // RGBA8_SRGB
+        1, // RG8U
+        1, // RG8S
+        1, // RG32UI
+        1, // R32UI
+        4, // ASTC_2D_8X8
+        4, // ASTC_2D_8X5
+        4, // ASTC_2D_5X4
+        1, // BGRA8_SRGB
+        4, // DXT1_SRGB
+        4, // DXT23_SRGB
+        4, // DXT45_SRGB
+        4, // BC7U_SRGB
+        4, // ASTC_2D_4X4_SRGB
+        4, // ASTC_2D_8X8_SRGB
+        4, // ASTC_2D_8X5_SRGB
+        4, // ASTC_2D_5X4_SRGB
+        4, // ASTC_2D_5X5
+        4, // ASTC_2D_5X5_SRGB
+        4, // ASTC_2D_10X8
+        4, // ASTC_2D_10X8_SRGB
+        1, // Z32F
+        1, // Z16
+        1, // Z24S8
+        1, // S8Z24
+        1, // Z32FS8
+    }};
+
+    ASSERT(static_cast<std::size_t>(format) < compression_factor_table.size());
+    return compression_factor_table[static_cast<std::size_t>(format)];
+}
+
+static constexpr u32 GetDefaultBlockWidth(PixelFormat format) {
+    if (format == PixelFormat::Invalid)
+        return 0;
+    constexpr std::array<u32, MaxPixelFormat> block_width_table = {{
+        1,  // ABGR8U
+        1,  // ABGR8S
+        1,  // ABGR8UI
+        1,  // B5G6R5U
+        1,  // A2B10G10R10U
+        1,  // A1B5G5R5U
+        1,  // R8U
+        1,  // R8UI
+        1,  // RGBA16F
+        1,  // RGBA16U
+        1,  // RGBA16UI
+        1,  // R11FG11FB10F
+        1,  // RGBA32UI
+        4,  // DXT1
+        4,  // DXT23
+        4,  // DXT45
+        4,  // DXN1
+        4,  // DXN2UNORM
+        4,  // DXN2SNORM
+        4,  // BC7U
+        4,  // BC6H_UF16
+        4,  // BC6H_SF16
+        4,  // ASTC_2D_4X4
+        1,  // G8R8U
+        1,  // G8R8S
+        1,  // BGRA8
+        1,  // RGBA32F
+        1,  // RG32F
+        1,  // R32F
+        1,  // R16F
+        1,  // R16U
+        1,  // R16S
+        1,  // R16UI
+        1,  // R16I
+        1,  // RG16
+        1,  // RG16F
+        1,  // RG16UI
+        1,  // RG16I
+        1,  // RG16S
+        1,  // RGB32F
+        1,  // RGBA8_SRGB
+        1,  // RG8U
+        1,  // RG8S
+        1,  // RG32UI
+        1,  // R32UI
+        8,  // ASTC_2D_8X8
+        8,  // ASTC_2D_8X5
+        5,  // ASTC_2D_5X4
+        1,  // BGRA8_SRGB
+        4,  // DXT1_SRGB
+        4,  // DXT23_SRGB
+        4,  // DXT45_SRGB
+        4,  // BC7U_SRGB
+        4,  // ASTC_2D_4X4_SRGB
+        8,  // ASTC_2D_8X8_SRGB
+        8,  // ASTC_2D_8X5_SRGB
+        5,  // ASTC_2D_5X4_SRGB
+        5,  // ASTC_2D_5X5
+        5,  // ASTC_2D_5X5_SRGB
+        10, // ASTC_2D_10X8
+        10, // ASTC_2D_10X8_SRGB
+        1,  // Z32F
+        1,  // Z16
+        1,  // Z24S8
+        1,  // S8Z24
+        1,  // Z32FS8
+    }};
+    ASSERT(static_cast<std::size_t>(format) < block_width_table.size());
+    return block_width_table[static_cast<std::size_t>(format)];
+}
+
+static constexpr u32 GetDefaultBlockHeight(PixelFormat format) {
+    if (format == PixelFormat::Invalid)
+        return 0;
+
+    constexpr std::array<u32, MaxPixelFormat> block_height_table = {{
+        1, // ABGR8U
+        1, // ABGR8S
+        1, // ABGR8UI
+        1, // B5G6R5U
+        1, // A2B10G10R10U
+        1, // A1B5G5R5U
+        1, // R8U
+        1, // R8UI
+        1, // RGBA16F
+        1, // RGBA16U
+        1, // RGBA16UI
+        1, // R11FG11FB10F
+        1, // RGBA32UI
+        4, // DXT1
+        4, // DXT23
+        4, // DXT45
+        4, // DXN1
+        4, // DXN2UNORM
+        4, // DXN2SNORM
+        4, // BC7U
+        4, // BC6H_UF16
+        4, // BC6H_SF16
+        4, // ASTC_2D_4X4
+        1, // G8R8U
+        1, // G8R8S
+        1, // BGRA8
+        1, // RGBA32F
+        1, // RG32F
+        1, // R32F
+        1, // R16F
+        1, // R16U
+        1, // R16S
+        1, // R16UI
+        1, // R16I
+        1, // RG16
+        1, // RG16F
+        1, // RG16UI
+        1, // RG16I
+        1, // RG16S
+        1, // RGB32F
+        1, // RGBA8_SRGB
+        1, // RG8U
+        1, // RG8S
+        1, // RG32UI
+        1, // R32UI
+        8, // ASTC_2D_8X8
+        5, // ASTC_2D_8X5
+        4, // ASTC_2D_5X4
+        1, // BGRA8_SRGB
+        4, // DXT1_SRGB
+        4, // DXT23_SRGB
+        4, // DXT45_SRGB
+        4, // BC7U_SRGB
+        4, // ASTC_2D_4X4_SRGB
+        8, // ASTC_2D_8X8_SRGB
+        5, // ASTC_2D_8X5_SRGB
+        4, // ASTC_2D_5X4_SRGB
+        5, // ASTC_2D_5X5
+        5, // ASTC_2D_5X5_SRGB
+        8, // ASTC_2D_10X8
+        8, // ASTC_2D_10X8_SRGB
+        1, // Z32F
+        1, // Z16
+        1, // Z24S8
+        1, // S8Z24
+        1, // Z32FS8
+    }};
+
+    ASSERT(static_cast<std::size_t>(format) < block_height_table.size());
+    return block_height_table[static_cast<std::size_t>(format)];
+}
+
+static constexpr u32 GetFormatBpp(PixelFormat format) {
+    if (format == PixelFormat::Invalid)
+        return 0;
+
+    constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
+        32,  // ABGR8U
+        32,  // ABGR8S
+        32,  // ABGR8UI
+        16,  // B5G6R5U
+        32,  // A2B10G10R10U
+        16,  // A1B5G5R5U
+        8,   // R8U
+        8,   // R8UI
+        64,  // RGBA16F
+        64,  // RGBA16U
+        64,  // RGBA16UI
+        32,  // R11FG11FB10F
+        128, // RGBA32UI
+        64,  // DXT1
+        128, // DXT23
+        128, // DXT45
+        64,  // DXN1
+        128, // DXN2UNORM
+        128, // DXN2SNORM
+        128, // BC7U
+        128, // BC6H_UF16
+        128, // BC6H_SF16
+        128, // ASTC_2D_4X4
+        16,  // G8R8U
+        16,  // G8R8S
+        32,  // BGRA8
+        128, // RGBA32F
+        64,  // RG32F
+        32,  // R32F
+        16,  // R16F
+        16,  // R16U
+        16,  // R16S
+        16,  // R16UI
+        16,  // R16I
+        32,  // RG16
+        32,  // RG16F
+        32,  // RG16UI
+        32,  // RG16I
+        32,  // RG16S
+        96,  // RGB32F
+        32,  // RGBA8_SRGB
+        16,  // RG8U
+        16,  // RG8S
+        64,  // RG32UI
+        32,  // R32UI
+        128, // ASTC_2D_8X8
+        128, // ASTC_2D_8X5
+        128, // ASTC_2D_5X4
+        32,  // BGRA8_SRGB
+        64,  // DXT1_SRGB
+        128, // DXT23_SRGB
+        128, // DXT45_SRGB
+        128, // BC7U
+        128, // ASTC_2D_4X4_SRGB
+        128, // ASTC_2D_8X8_SRGB
+        128, // ASTC_2D_8X5_SRGB
+        128, // ASTC_2D_5X4_SRGB
+        128, // ASTC_2D_5X5
+        128, // ASTC_2D_5X5_SRGB
+        128, // ASTC_2D_10X8
+        128, // ASTC_2D_10X8_SRGB
+        32,  // Z32F
+        16,  // Z16
+        32,  // Z24S8
+        32,  // S8Z24
+        64,  // Z32FS8
+    }};
+
+    ASSERT(static_cast<std::size_t>(format) < bpp_table.size());
+    return bpp_table[static_cast<std::size_t>(format)];
+}
+
+/// Returns the sizer in bytes of the specified pixel format
+static constexpr u32 GetBytesPerPixel(PixelFormat pixel_format) {
+    if (pixel_format == PixelFormat::Invalid) {
+        return 0;
+    }
+    return GetFormatBpp(pixel_format) / CHAR_BIT;
+}
+
+SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type);
+
+bool SurfaceTargetIsLayered(SurfaceTarget target);
+
+PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format);
+
+PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format);
+
+PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
+                                         Tegra::Texture::ComponentType component_type,
+                                         bool is_srgb);
+
+ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type);
+
+ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format);
+
+PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format);
+
+ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format);
+
+SurfaceType GetFormatType(PixelFormat pixel_format);
+
+bool IsPixelFormatASTC(PixelFormat format);
+
+std::pair<u32, u32> GetASTCBlockSize(PixelFormat format);
+
+/// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN
+bool IsFormatBCn(PixelFormat format);
+
+} // namespace VideoCore::Surface

src/video_core/textures/astc.cpp

@@ -1598,27 +1598,29 @@ static void DecompressBlock(uint8_t inBuf[16], const uint32_t blockWidth,
 namespace Tegra::Texture::ASTC {
 
 std::vector<uint8_t> Decompress(std::vector<uint8_t>& data, uint32_t width, uint32_t height,
-                                uint32_t block_width, uint32_t block_height) {
+                                uint32_t depth, uint32_t block_width, uint32_t block_height) {
     uint32_t blockIdx = 0;
-    std::vector<uint8_t> outData(height * width * 4);
-    for (uint32_t j = 0; j < height; j += block_height) {
-        for (uint32_t i = 0; i < width; i += block_width) {
+    std::vector<uint8_t> outData(height * width * depth * 4);
+    for (uint32_t k = 0; k < depth; k++) {
+        for (uint32_t j = 0; j < height; j += block_height) {
+            for (uint32_t i = 0; i < width; i += block_width) {
 
-            uint8_t* blockPtr = data.data() + blockIdx * 16;
+                uint8_t* blockPtr = data.data() + blockIdx * 16;
 
-            // Blocks can be at most 12x12
-            uint32_t uncompData[144];
-            ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData);
+                // Blocks can be at most 12x12
+                uint32_t uncompData[144];
+                ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData);
 
-            uint32_t decompWidth = std::min(block_width, width - i);
-            uint32_t decompHeight = std::min(block_height, height - j);
+                uint32_t decompWidth = std::min(block_width, width - i);
+                uint32_t decompHeight = std::min(block_height, height - j);
 
-            uint8_t* outRow = outData.data() + (j * width + i) * 4;
-            for (uint32_t jj = 0; jj < decompHeight; jj++) {
-                memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4);
+                uint8_t* outRow = outData.data() + (j * width + i) * 4;
+                for (uint32_t jj = 0; jj < decompHeight; jj++) {
+                    memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4);
+                }
+
+                blockIdx++;
             }
-
-            blockIdx++;
         }
     }
 

src/video_core/textures/astc.h

@@ -10,6 +10,6 @@
 namespace Tegra::Texture::ASTC {
 
 std::vector<uint8_t> Decompress(std::vector<uint8_t>& data, uint32_t width, uint32_t height,
-                                uint32_t block_width, uint32_t block_height);
+                                uint32_t depth, uint32_t block_width, uint32_t block_height);
 
 } // namespace Tegra::Texture::ASTC

src/video_core/textures/decoders.cpp

@@ -202,6 +202,8 @@ u32 BytesPerPixel(TextureFormat format) {
     case TextureFormat::ASTC_2D_5X4:
     case TextureFormat::ASTC_2D_8X8:
     case TextureFormat::ASTC_2D_8X5:
+    case TextureFormat::ASTC_2D_10X8:
+    case TextureFormat::ASTC_2D_5X5:
     case TextureFormat::A8R8G8B8:
     case TextureFormat::A2B10G10R10:
     case TextureFormat::BF10GF11RF11:
@@ -227,12 +229,14 @@ u32 BytesPerPixel(TextureFormat format) {
     }
 }
 
-std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
-                                 u32 height, u32 depth, u32 block_height, u32 block_depth) {
+std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size_x, u32 tile_size_y,
+                                 u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
+                                 u32 block_height, u32 block_depth) {
     std::vector<u8> unswizzled_data(width * height * depth * bytes_per_pixel);
-    CopySwizzledData(width / tile_size, height / tile_size, depth, bytes_per_pixel, bytes_per_pixel,
-                     Memory::GetPointer(address), unswizzled_data.data(), true, block_height,
-                     block_depth);
+    CopySwizzledData((width + tile_size_x - 1) / tile_size_x,
+                     (height + tile_size_y - 1) / tile_size_y, depth, bytes_per_pixel,
+                     bytes_per_pixel, Memory::GetPointer(address), unswizzled_data.data(), true,
+                     block_height, block_depth);
     return unswizzled_data;
 }
 
@@ -292,6 +296,8 @@ std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat
     case TextureFormat::BC6H_SF16:
     case TextureFormat::ASTC_2D_4X4:
     case TextureFormat::ASTC_2D_8X8:
+    case TextureFormat::ASTC_2D_5X5:
+    case TextureFormat::ASTC_2D_10X8:
     case TextureFormat::A8R8G8B8:
     case TextureFormat::A2B10G10R10:
     case TextureFormat::A1B5G5R5:

src/video_core/textures/decoders.h

@@ -19,8 +19,8 @@ inline std::size_t GetGOBSize() {
 /**
  * Unswizzles a swizzled texture without changing its format.
  */
-std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
-                                 u32 height, u32 depth,
+std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size_x, u32 tile_size_y,
+                                 u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
                                  u32 block_height = TICEntry::DefaultBlockHeight,
                                  u32 block_depth = TICEntry::DefaultBlockHeight);
 

src/video_core/textures/texture.h

@@ -168,20 +168,29 @@ struct TICEntry {
 
         // High 16 bits of the pitch value
         BitField<0, 16, u32> pitch_high;
-
+        BitField<26, 1, u32> use_header_opt_control;
+        BitField<27, 1, u32> depth_texture;
         BitField<28, 4, u32> max_mip_level;
     };
     union {
         BitField<0, 16, u32> width_minus_1;
         BitField<22, 1, u32> srgb_conversion;
         BitField<23, 4, TextureType> texture_type;
+        BitField<29, 3, u32> border_size;
     };
     union {
         BitField<0, 16, u32> height_minus_1;
         BitField<16, 15, u32> depth_minus_1;
     };
+    union {
+        BitField<6, 13, u32> mip_lod_bias;
+        BitField<27, 3, u32> max_anisotropy;
+    };
 
-    INSERT_PADDING_BYTES(8);
+    union {
+        BitField<0, 4, u32> res_min_mip_level;
+        BitField<4, 4, u32> res_max_mip_level;
+    };
 
     GPUVAddr Address() const {
         return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | address_low);

src/web_service/telemetry_json.cpp

@@ -102,16 +102,27 @@ void TelemetryJson::Complete() {
     impl->SerializeSection(Telemetry::FieldType::App, "App");
     impl->SerializeSection(Telemetry::FieldType::Session, "Session");
     impl->SerializeSection(Telemetry::FieldType::Performance, "Performance");
-    impl->SerializeSection(Telemetry::FieldType::UserFeedback, "UserFeedback");
     impl->SerializeSection(Telemetry::FieldType::UserConfig, "UserConfig");
     impl->SerializeSection(Telemetry::FieldType::UserSystem, "UserSystem");
 
     auto content = impl->TopSection().dump();
     // Send the telemetry async but don't handle the errors since they were written to the log
-    Common::DetachedTasks::AddTask(
-        [host{impl->host}, username{impl->username}, token{impl->token}, content]() {
-            Client{host, username, token}.PostJson("/telemetry", content, true);
-        });
+    Common::DetachedTasks::AddTask([host{impl->host}, content]() {
+        Client{host, "", ""}.PostJson("/telemetry", content, true);
+    });
+}
+
+bool TelemetryJson::SubmitTestcase() {
+    impl->SerializeSection(Telemetry::FieldType::App, "App");
+    impl->SerializeSection(Telemetry::FieldType::Session, "Session");
+    impl->SerializeSection(Telemetry::FieldType::UserFeedback, "UserFeedback");
+    impl->SerializeSection(Telemetry::FieldType::UserSystem, "UserSystem");
+
+    auto content = impl->TopSection().dump();
+    Client client(impl->host, impl->username, impl->token);
+    auto value = client.PostJson("/gamedb/testcase", content, false);
+
+    return value.result_code == Common::WebResult::Code::Success;
 }
 
 } // namespace WebService

src/web_service/telemetry_json.h

@@ -35,6 +35,7 @@ public:
     void Visit(const Telemetry::Field<std::chrono::microseconds>& field) override;
 
     void Complete() override;
+    bool SubmitTestcase() override;
 
 private:
     struct Impl;

src/yuzu/CMakeLists.txt

@@ -56,6 +56,8 @@ add_executable(yuzu
     main.h
     ui_settings.cpp
     ui_settings.h
+    util/limitable_input_dialog.cpp
+    util/limitable_input_dialog.h
     util/spinbox.cpp
     util/spinbox.h
     util/util.cpp

src/yuzu/compatdb.cpp

@@ -5,6 +5,7 @@
 #include <QButtonGroup>
 #include <QMessageBox>
 #include <QPushButton>
+#include <QtConcurrent/qtconcurrentrun.h>
 #include "common/logging/log.h"
 #include "common/telemetry.h"
 #include "core/core.h"
@@ -23,6 +24,8 @@ CompatDB::CompatDB(QWidget* parent)
     connect(ui->radioButton_IntroMenu, &QRadioButton::clicked, this, &CompatDB::EnableNext);
     connect(ui->radioButton_WontBoot, &QRadioButton::clicked, this, &CompatDB::EnableNext);
     connect(button(NextButton), &QPushButton::clicked, this, &CompatDB::Submit);
+    connect(&testcase_watcher, &QFutureWatcher<bool>::finished, this,
+            &CompatDB::OnTestcaseSubmitted);
 }
 
 CompatDB::~CompatDB() = default;
@@ -48,18 +51,38 @@ void CompatDB::Submit() {
         }
         break;
     case CompatDBPage::Final:
+        back();
         LOG_DEBUG(Frontend, "Compatibility Rating: {}", compatibility->checkedId());
         Core::Telemetry().AddField(Telemetry::FieldType::UserFeedback, "Compatibility",
                                    compatibility->checkedId());
-        // older versions of QT don't support the "NoCancelButtonOnLastPage" option, this is a
-        // workaround
+
+        button(NextButton)->setEnabled(false);
+        button(NextButton)->setText(tr("Submitting"));
         button(QWizard::CancelButton)->setVisible(false);
+
+        testcase_watcher.setFuture(QtConcurrent::run(
+            [this]() { return Core::System::GetInstance().TelemetrySession().SubmitTestcase(); }));
         break;
     default:
         LOG_ERROR(Frontend, "Unexpected page: {}", currentId());
     }
 }
 
+void CompatDB::OnTestcaseSubmitted() {
+    if (!testcase_watcher.result()) {
+        QMessageBox::critical(this, tr("Communication error"),
+                              tr("An error occured while sending the Testcase"));
+        button(NextButton)->setEnabled(true);
+        button(NextButton)->setText(tr("Next"));
+        button(QWizard::CancelButton)->setVisible(true);
+    } else {
+        next();
+        // older versions of QT don't support the "NoCancelButtonOnLastPage" option, this is a
+        // workaround
+        button(QWizard::CancelButton)->setVisible(false);
+    }
+}
+
 void CompatDB::EnableNext() {
     button(NextButton)->setEnabled(true);
 }

src/yuzu/compatdb.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <memory>
+#include <QFutureWatcher>
 #include <QWizard>
 
 namespace Ui {
@@ -19,8 +20,11 @@ public:
     ~CompatDB();
 
 private:
+    QFutureWatcher<bool> testcase_watcher;
+
     std::unique_ptr<Ui::CompatDB> ui;
 
     void Submit();
+    void OnTestcaseSubmitted();
     void EnableNext();
 };

src/yuzu/configuration/config.cpp

@@ -134,6 +134,14 @@ void Config::ReadValues() {
                                                     Service::Account::MAX_USERS - 1);
 
     Settings::values.language_index = qt_config->value("language_index", 1).toInt();
+
+    const auto enabled = qt_config->value("rng_seed_enabled", false).toBool();
+    if (enabled) {
+        Settings::values.rng_seed = qt_config->value("rng_seed", 0).toULongLong();
+    } else {
+        Settings::values.rng_seed = std::nullopt;
+    }
+
     qt_config->endGroup();
 
     qt_config->beginGroup("Miscellaneous");
@@ -272,6 +280,10 @@ void Config::SaveValues() {
     qt_config->setValue("current_user", Settings::values.current_user);
 
     qt_config->setValue("language_index", Settings::values.language_index);
+
+    qt_config->setValue("rng_seed_enabled", Settings::values.rng_seed.has_value());
+    qt_config->setValue("rng_seed", Settings::values.rng_seed.value_or(0));
+
     qt_config->endGroup();
 
     qt_config->beginGroup("Miscellaneous");

src/yuzu/configuration/configure_general.cpp

@@ -3,6 +3,10 @@
 // Refer to the license.txt file included.
 
 #include "core/core.h"
+#include "core/hle/service/am/am.h"
+#include "core/hle/service/am/applet_ae.h"
+#include "core/hle/service/am/applet_oe.h"
+#include "core/hle/service/sm/sm.h"
 #include "core/settings.h"
 #include "ui_configure_general.h"
 #include "yuzu/configuration/configure_general.h"
@@ -20,7 +24,6 @@ ConfigureGeneral::ConfigureGeneral(QWidget* parent)
     this->setConfiguration();
 
     ui->use_cpu_jit->setEnabled(!Core::System::GetInstance().IsPoweredOn());
-    ui->use_docked_mode->setEnabled(!Core::System::GetInstance().IsPoweredOn());
 }
 
 ConfigureGeneral::~ConfigureGeneral() = default;
@@ -38,6 +41,30 @@ void ConfigureGeneral::PopulateHotkeyList(const HotkeyRegistry& registry) {
     ui->widget->Populate(registry);
 }
 
+void ConfigureGeneral::OnDockedModeChanged(bool last_state, bool new_state) {
+    if (last_state == new_state) {
+        return;
+    }
+
+    Core::System& system{Core::System::GetInstance()};
+    Service::SM::ServiceManager& sm = system.ServiceManager();
+
+    // Message queue is shared between these services, we just need to signal an operation
+    // change to one and it will handle both automatically
+    auto applet_oe = sm.GetService<Service::AM::AppletOE>("appletOE");
+    auto applet_ae = sm.GetService<Service::AM::AppletAE>("appletAE");
+    bool has_signalled = false;
+
+    if (applet_oe != nullptr) {
+        applet_oe->GetMessageQueue()->OperationModeChanged();
+        has_signalled = true;
+    }
+
+    if (applet_ae != nullptr && !has_signalled) {
+        applet_ae->GetMessageQueue()->OperationModeChanged();
+    }
+}
+
 void ConfigureGeneral::applyConfiguration() {
     UISettings::values.gamedir_deepscan = ui->toggle_deepscan->isChecked();
     UISettings::values.confirm_before_closing = ui->toggle_check_exit->isChecked();
@@ -45,6 +72,9 @@ void ConfigureGeneral::applyConfiguration() {
         ui->theme_combobox->itemData(ui->theme_combobox->currentIndex()).toString();
 
     Settings::values.use_cpu_jit = ui->use_cpu_jit->isChecked();
+    const bool pre_docked_mode = Settings::values.use_docked_mode;
     Settings::values.use_docked_mode = ui->use_docked_mode->isChecked();
+    OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode);
+
     Settings::values.enable_nfc = ui->enable_nfc->isChecked();
 }

src/yuzu/configuration/configure_general.h

@@ -25,6 +25,7 @@ public:
 
 private:
     void setConfiguration();
+    void OnDockedModeChanged(bool last_state, bool new_state);
 
     std::unique_ptr<Ui::ConfigureGeneral> ui;
 };

src/yuzu/configuration/configure_system.cpp

@@ -6,20 +6,20 @@
 #include <QFileDialog>
 #include <QGraphicsItem>
 #include <QGraphicsScene>
-#include <QInputDialog>
+#include <QHeaderView>
 #include <QMessageBox>
 #include <QStandardItemModel>
 #include <QTreeView>
 #include <QVBoxLayout>
-#include "common/common_paths.h"
-#include "common/logging/backend.h"
+#include "common/assert.h"
+#include "common/file_util.h"
 #include "common/string_util.h"
 #include "core/core.h"
 #include "core/hle/service/acc/profile_manager.h"
 #include "core/settings.h"
 #include "ui_configure_system.h"
 #include "yuzu/configuration/configure_system.h"
-#include "yuzu/main.h"
+#include "yuzu/util/limitable_input_dialog.h"
 
 namespace {
 constexpr std::array<int, 12> days_in_month = {{
@@ -83,6 +83,12 @@ QPixmap GetIcon(Service::Account::UUID uuid) {
 
     return icon.scaled(64, 64, Qt::IgnoreAspectRatio, Qt::SmoothTransformation);
 }
+
+QString GetProfileUsernameFromUser(QWidget* parent, const QString& description_text) {
+    return LimitableInputDialog::GetText(parent, ConfigureSystem::tr("Enter Username"),
+                                         description_text, 1,
+                                         static_cast<int>(Service::Account::profile_username_size));
+}
 } // Anonymous namespace
 
 ConfigureSystem::ConfigureSystem(QWidget* parent)
@@ -131,6 +137,12 @@ ConfigureSystem::ConfigureSystem(QWidget* parent)
     connect(ui->pm_remove, &QPushButton::pressed, this, &ConfigureSystem::DeleteUser);
     connect(ui->pm_set_image, &QPushButton::pressed, this, &ConfigureSystem::SetUserImage);
 
+    connect(ui->rng_seed_checkbox, &QCheckBox::stateChanged, this, [this](bool checked) {
+        ui->rng_seed_edit->setEnabled(checked);
+        if (!checked)
+            ui->rng_seed_edit->setText(QStringLiteral("0000000000000000"));
+    });
+
     scene = new QGraphicsScene;
     ui->current_user_icon->setScene(scene);
 
@@ -149,6 +161,14 @@ void ConfigureSystem::setConfiguration() {
 
     PopulateUserList();
     UpdateCurrentUser();
+
+    ui->rng_seed_checkbox->setChecked(Settings::values.rng_seed.has_value());
+    ui->rng_seed_edit->setEnabled(Settings::values.rng_seed.has_value());
+
+    const auto rng_seed = QString("%1")
+                              .arg(Settings::values.rng_seed.value_or(0), 16, 16, QLatin1Char{'0'})
+                              .toUpper();
+    ui->rng_seed_edit->setText(rng_seed);
 }
 
 void ConfigureSystem::PopulateUserList() {
@@ -189,6 +209,12 @@ void ConfigureSystem::applyConfiguration() {
         return;
 
     Settings::values.language_index = ui->combo_language->currentIndex();
+
+    if (ui->rng_seed_checkbox->isChecked())
+        Settings::values.rng_seed = ui->rng_seed_edit->text().toULongLong(nullptr, 16);
+    else
+        Settings::values.rng_seed = std::nullopt;
+
     Settings::Apply();
 }
 
@@ -234,7 +260,7 @@ void ConfigureSystem::RefreshConsoleID() {
 
 void ConfigureSystem::SelectUser(const QModelIndex& index) {
     Settings::values.current_user =
-        std::clamp<std::size_t>(index.row(), 0, profile_manager->GetUserCount() - 1);
+        std::clamp<s32>(index.row(), 0, static_cast<s32>(profile_manager->GetUserCount() - 1));
 
     UpdateCurrentUser();
 
@@ -244,15 +270,13 @@ void ConfigureSystem::SelectUser(const QModelIndex& index) {
 }
 
 void ConfigureSystem::AddUser() {
-    const auto uuid = Service::Account::UUID::Generate();
-
-    bool ok = false;
     const auto username =
-        QInputDialog::getText(this, tr("Enter Username"), tr("Enter a username for the new user:"),
-                              QLineEdit::Normal, QString(), &ok);
-    if (!ok)
+        GetProfileUsernameFromUser(this, tr("Enter a username for the new user:"));
+    if (username.isEmpty()) {
         return;
+    }
 
+    const auto uuid = Service::Account::UUID::Generate();
     profile_manager->CreateNewUser(uuid, username.toStdString());
 
     item_model->appendRow(new QStandardItem{GetIcon(uuid), FormatUserEntryText(username, uuid)});
@@ -267,24 +291,15 @@ void ConfigureSystem::RenameUser() {
     if (!profile_manager->GetProfileBase(*uuid, profile))
         return;
 
-    bool ok = false;
-    const auto old_username = GetAccountUsername(*profile_manager, *uuid);
-    const auto new_username =
-        QInputDialog::getText(this, tr("Enter Username"), tr("Enter a new username:"),
-                              QLineEdit::Normal, old_username, &ok);
-
-    if (!ok)
+    const auto new_username = GetProfileUsernameFromUser(this, tr("Enter a new username:"));
+    if (new_username.isEmpty()) {
         return;
-
-    std::fill(profile.username.begin(), profile.username.end(), '\0');
-    const auto username_std = new_username.toStdString();
-    if (username_std.size() > profile.username.size()) {
-        std::copy_n(username_std.begin(), std::min(profile.username.size(), username_std.size()),
-                    profile.username.begin());
-    } else {
-        std::copy(username_std.begin(), username_std.end(), profile.username.begin());
     }
 
+    const auto username_std = new_username.toStdString();
+    std::fill(profile.username.begin(), profile.username.end(), '\0');
+    std::copy(username_std.begin(), username_std.end(), profile.username.begin());
+
     profile_manager->SetProfileBase(*uuid, profile);
 
     item_model->setItem(

src/yuzu/configuration/configure_system.ui

@@ -6,7 +6,7 @@
    <rect>
     <x>0</x>
     <y>0</y>
-    <width>360</width>
+    <width>366</width>
     <height>483</height>
    </rect>
   </property>
@@ -22,98 +22,6 @@
         <string>System Settings</string>
        </property>
        <layout class="QGridLayout" name="gridLayout">
-        <item row="1" column="0">
-         <widget class="QLabel" name="label_language">
-          <property name="text">
-           <string>Language</string>
-          </property>
-         </widget>
-        </item>
-        <item row="0" column="0">
-         <widget class="QLabel" name="label_birthday">
-          <property name="text">
-           <string>Birthday</string>
-          </property>
-         </widget>
-        </item>
-        <item row="3" column="0">
-         <widget class="QLabel" name="label_console_id">
-          <property name="text">
-           <string>Console ID:</string>
-          </property>
-         </widget>
-        </item>
-        <item row="0" column="1">
-         <layout class="QHBoxLayout" name="horizontalLayout_birthday2">
-          <item>
-           <widget class="QComboBox" name="combo_birthmonth">
-            <item>
-             <property name="text">
-              <string>January</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>February</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>March</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>April</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>May</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>June</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>July</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>August</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>September</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>October</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>November</string>
-             </property>
-            </item>
-            <item>
-             <property name="text">
-              <string>December</string>
-             </property>
-            </item>
-           </widget>
-          </item>
-          <item>
-           <widget class="QComboBox" name="combo_birthday"/>
-          </item>
-         </layout>
-        </item>
         <item row="1" column="1">
          <widget class="QComboBox" name="combo_language">
           <property name="toolTip">
@@ -206,6 +114,13 @@
           </item>
          </widget>
         </item>
+        <item row="3" column="0">
+         <widget class="QLabel" name="label_console_id">
+          <property name="text">
+           <string>Console ID:</string>
+          </property>
+         </widget>
+        </item>
         <item row="2" column="0">
          <widget class="QLabel" name="label_sound">
           <property name="text">
@@ -213,6 +128,100 @@
           </property>
          </widget>
         </item>
+        <item row="0" column="0">
+         <widget class="QLabel" name="label_birthday">
+          <property name="text">
+           <string>Birthday</string>
+          </property>
+         </widget>
+        </item>
+        <item row="0" column="1">
+         <layout class="QHBoxLayout" name="horizontalLayout_birthday2">
+          <item>
+           <widget class="QComboBox" name="combo_birthmonth">
+            <item>
+             <property name="text">
+              <string>January</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>February</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>March</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>April</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>May</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>June</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>July</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>August</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>September</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>October</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>November</string>
+             </property>
+            </item>
+            <item>
+             <property name="text">
+              <string>December</string>
+             </property>
+            </item>
+           </widget>
+          </item>
+          <item>
+           <widget class="QComboBox" name="combo_birthday"/>
+          </item>
+         </layout>
+        </item>
+        <item row="3" column="1">
+         <widget class="QPushButton" name="button_regenerate_console_id">
+          <property name="sizePolicy">
+           <sizepolicy hsizetype="Fixed" vsizetype="Fixed">
+            <horstretch>0</horstretch>
+            <verstretch>0</verstretch>
+           </sizepolicy>
+          </property>
+          <property name="layoutDirection">
+           <enum>Qt::RightToLeft</enum>
+          </property>
+          <property name="text">
+           <string>Regenerate</string>
+          </property>
+         </widget>
+        </item>
         <item row="2" column="1">
          <widget class="QComboBox" name="combo_sound">
           <item>
@@ -232,19 +241,38 @@
           </item>
          </widget>
         </item>
-        <item row="3" column="1">
-         <widget class="QPushButton" name="button_regenerate_console_id">
+        <item row="1" column="0">
+         <widget class="QLabel" name="label_language">
+          <property name="text">
+           <string>Language</string>
+          </property>
+         </widget>
+        </item>
+        <item row="4" column="0">
+         <widget class="QCheckBox" name="rng_seed_checkbox">
+          <property name="text">
+           <string>RNG Seed</string>
+          </property>
+         </widget>
+        </item>
+        <item row="4" column="1">
+         <widget class="QLineEdit" name="rng_seed_edit">
           <property name="sizePolicy">
-           <sizepolicy hsizetype="Fixed" vsizetype="Fixed">
+           <sizepolicy hsizetype="Minimum" vsizetype="Fixed">
             <horstretch>0</horstretch>
             <verstretch>0</verstretch>
            </sizepolicy>
           </property>
-          <property name="layoutDirection">
-           <enum>Qt::RightToLeft</enum>
+          <property name="font">
+           <font>
+            <family>Lucida Console</family>
+           </font>
           </property>
-          <property name="text">
-           <string>Regenerate</string>
+          <property name="inputMask">
+           <string>HHHHHHHHHHHHHHHH</string>
+          </property>
+          <property name="maxLength">
+           <number>16</number>
           </property>
          </widget>
         </item>

src/yuzu/debugger/graphics/graphics_surface.cpp

@@ -386,9 +386,9 @@ void GraphicsSurfaceWidget::OnUpdate() {
 
     // TODO(bunnei): Will not work with BCn formats that swizzle 4x4 tiles.
     // Needs to be fixed if we plan to use this feature more, otherwise we may remove it.
-    auto unswizzled_data =
-        Tegra::Texture::UnswizzleTexture(*address, 1, Tegra::Texture::BytesPerPixel(surface_format),
-                                         surface_width, surface_height, 1U);
+    auto unswizzled_data = Tegra::Texture::UnswizzleTexture(
+        *address, 1, 1, Tegra::Texture::BytesPerPixel(surface_format), surface_width,
+        surface_height, 1U);
 
     auto texture_data = Tegra::Texture::DecodeTexture(unswizzled_data, surface_format,
                                                       surface_width, surface_height);

src/yuzu/main.cpp

@@ -142,6 +142,9 @@ static void InitializeLogging() {
     const std::string& log_dir = FileUtil::GetUserPath(FileUtil::UserPath::LogDir);
     FileUtil::CreateFullPath(log_dir);
     Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir + LOG_FILE));
+#ifdef _WIN32
+    Log::AddBackend(std::make_unique<Log::DebuggerBackend>());
+#endif
 }
 
 GMainWindow::GMainWindow()
@@ -454,6 +457,7 @@ void GMainWindow::ConnectMenuEvents() {
     connect(ui.action_Fullscreen, &QAction::triggered, this, &GMainWindow::ToggleFullscreen);
 
     // Help
+    connect(ui.action_Open_yuzu_Folder, &QAction::triggered, this, &GMainWindow::OnOpenYuzuFolder);
     connect(ui.action_Rederive, &QAction::triggered, this,
             std::bind(&GMainWindow::OnReinitializeKeys, this, ReinitializeKeyBehavior::Warning));
     connect(ui.action_About, &QAction::triggered, this, &GMainWindow::OnAbout);
@@ -929,7 +933,8 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa
     const auto full = res == "Full";
     const auto entry_size = CalculateRomFSEntrySize(extracted, full);
 
-    QProgressDialog progress(tr("Extracting RomFS..."), tr("Cancel"), 0, entry_size, this);
+    QProgressDialog progress(tr("Extracting RomFS..."), tr("Cancel"), 0,
+                             static_cast<s32>(entry_size), this);
     progress.setWindowModality(Qt::WindowModal);
     progress.setMinimumDuration(100);
 
@@ -1374,6 +1379,11 @@ void GMainWindow::OnLoadAmiibo() {
     }
 }
 
+void GMainWindow::OnOpenYuzuFolder() {
+    QDesktopServices::openUrl(QUrl::fromLocalFile(
+        QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::UserDir))));
+}
+
 void GMainWindow::OnAbout() {
     AboutDialog aboutDialog(this);
     aboutDialog.exec();
@@ -1532,7 +1542,7 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) {
                    "derivation. It will be attempted but may not complete.<br><br>") +
                     errors +
                     tr("<br><br>You can get all of these and dump all of your games easily by "
-                       "following <a href='https://yuzu-emu.org/help/quickstart/quickstart/'>the "
+                       "following <a href='https://yuzu-emu.org/help/quickstart/'>the "
                        "quickstart guide</a>. Alternatively, you can use another method of dumping "
                        "to obtain all of your keys."));
         }
@@ -1612,7 +1622,7 @@ void GMainWindow::closeEvent(QCloseEvent* event) {
         return;
     }
 
-    if (ui.action_Fullscreen->isChecked()) {
+    if (!ui.action_Fullscreen->isChecked()) {
         UISettings::values.geometry = saveGeometry();
         UISettings::values.renderwindow_geometry = render_window->saveGeometry();
     }

src/yuzu/main.h

@@ -167,6 +167,7 @@ private slots:
     void OnMenuRecentFile();
     void OnConfigure();
     void OnLoadAmiibo();
+    void OnOpenYuzuFolder();
     void OnAbout();
     void OnToggleFilterBar();
     void OnDisplayTitleBars(bool);

src/yuzu/main.ui

@@ -110,6 +110,7 @@
      <string>&amp;Help</string>
     </property>
     <addaction name="action_Report_Compatibility"/>
+    <addaction name="action_Open_yuzu_Folder" />
     <addaction name="separator"/>
     <addaction name="action_About"/>
    </widget>
@@ -277,6 +278,11 @@
        <bool>false</bool>
      </property>
    </action>
+   <action name="action_Open_yuzu_Folder">
+     <property name="text">
+       <string>Open yuzu Folder</string>
+     </property>
+   </action>
   </widget>
  <resources/>
  <connections/>

src/yuzu/util/limitable_input_dialog.cpp

@@ -0,0 +1,59 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <QDialogButtonBox>
+#include <QLabel>
+#include <QLineEdit>
+#include <QPushButton>
+#include <QVBoxLayout>
+#include "yuzu/util/limitable_input_dialog.h"
+
+LimitableInputDialog::LimitableInputDialog(QWidget* parent) : QDialog{parent} {
+    CreateUI();
+    ConnectEvents();
+}
+
+LimitableInputDialog::~LimitableInputDialog() = default;
+
+void LimitableInputDialog::CreateUI() {
+    setWindowFlags(windowFlags() & ~Qt::WindowContextHelpButtonHint);
+
+    text_label = new QLabel(this);
+    text_entry = new QLineEdit(this);
+    buttons = new QDialogButtonBox(QDialogButtonBox::Ok | QDialogButtonBox::Cancel, this);
+
+    auto* const layout = new QVBoxLayout;
+    layout->addWidget(text_label);
+    layout->addWidget(text_entry);
+    layout->addWidget(buttons);
+
+    setLayout(layout);
+}
+
+void LimitableInputDialog::ConnectEvents() {
+    connect(buttons, &QDialogButtonBox::accepted, this, &QDialog::accept);
+    connect(buttons, &QDialogButtonBox::rejected, this, &QDialog::reject);
+}
+
+QString LimitableInputDialog::GetText(QWidget* parent, const QString& title, const QString& text,
+                                      int min_character_limit, int max_character_limit) {
+    Q_ASSERT(min_character_limit <= max_character_limit);
+
+    LimitableInputDialog dialog{parent};
+    dialog.setWindowTitle(title);
+    dialog.text_label->setText(text);
+    dialog.text_entry->setMaxLength(max_character_limit);
+
+    auto* const ok_button = dialog.buttons->button(QDialogButtonBox::Ok);
+    ok_button->setEnabled(false);
+    connect(dialog.text_entry, &QLineEdit::textEdited, [&](const QString& new_text) {
+        ok_button->setEnabled(new_text.length() >= min_character_limit);
+    });
+
+    if (dialog.exec() != QDialog::Accepted) {
+        return {};
+    }
+
+    return dialog.text_entry->text();
+}

src/yuzu/util/limitable_input_dialog.h

@@ -0,0 +1,31 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <QDialog>
+
+class QDialogButtonBox;
+class QLabel;
+class QLineEdit;
+
+/// A QDialog that functions similarly to QInputDialog, however, it allows
+/// restricting the minimum and total number of characters that can be entered.
+class LimitableInputDialog final : public QDialog {
+    Q_OBJECT
+public:
+    explicit LimitableInputDialog(QWidget* parent = nullptr);
+    ~LimitableInputDialog() override;
+
+    static QString GetText(QWidget* parent, const QString& title, const QString& text,
+                           int min_character_limit, int max_character_limit);
+
+private:
+    void CreateUI();
+    void ConnectEvents();
+
+    QLabel* text_label;
+    QLineEdit* text_entry;
+    QDialogButtonBox* buttons;
+};

src/yuzu_cmd/config.cpp

@@ -132,6 +132,13 @@ void Config::ReadValues() {
     Settings::values.current_user = std::clamp<int>(
         sdl2_config->GetInteger("System", "current_user", 0), 0, Service::Account::MAX_USERS - 1);
 
+    const auto enabled = sdl2_config->GetBoolean("System", "rng_seed_enabled", false);
+    if (enabled) {
+        Settings::values.rng_seed = sdl2_config->GetInteger("System", "rng_seed", 0);
+    } else {
+        Settings::values.rng_seed = std::nullopt;
+    }
+
     // Miscellaneous
     Settings::values.log_filter = sdl2_config->Get("Miscellaneous", "log_filter", "*:Trace");
     Settings::values.use_dev_keys = sdl2_config->GetBoolean("Miscellaneous", "use_dev_keys", false);

src/yuzu_cmd/default_ini.h

@@ -178,6 +178,11 @@ use_docked_mode =
 # 1 (default): Yes, 0 : No
 enable_nfc =
 
+# Sets the seed for the RNG generator built into the switch
+# rng_seed will be ignored and randomly generated if rng_seed_enabled is false
+rng_seed_enabled =
+rng_seed =
+
 # Sets the account username, max length is 32 characters
 # yuzu (default)
 username = yuzu

src/yuzu_cmd/yuzu.cpp

@@ -76,6 +76,9 @@ static void InitializeLogging() {
     const std::string& log_dir = FileUtil::GetUserPath(FileUtil::UserPath::LogDir);
     FileUtil::CreateFullPath(log_dir);
     Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir + LOG_FILE));
+#ifdef _WIN32
+    Log::AddBackend(std::make_unique<Log::DebuggerBackend>());
+#endif
 }
 
 /// Application entry point