vector_math: Remove AsArray() and Write() functions from Vec[2,3,4]

These are all unused and the Write() ones should arguably not even be in the interface. There are better ways to provide this if we ever need it (like iterators).

README.md

@@ -7,7 +7,7 @@ yuzu is an experimental open-source emulator for the Nintendo Switch from the cr
 
 It is written in C++ with portability in mind, with builds actively maintained for Windows, Linux and macOS. The emulator is currently only useful for homebrew development and research purposes.
 
-yuzu only emulates a subset of Switch hardware and therefore is generally only useful for running/debugging homebrew applications. At this time, yuzu does not run any commercial Switch games. yuzu can boot some games, to varying degrees of success, but does not implement any of the necessary GPU features to render 3D graphics.
+yuzu only emulates a subset of Switch hardware and therefore is generally only useful for running/debugging homebrew applications. At this time, yuzu cannot play any commercial games without major problems. yuzu can boot some games, to varying degrees of success, but does not implement any of the necessary GPU features to render 3D graphics.
 
 yuzu is licensed under the GPLv2 (or any later version). Refer to the license.txt file included.
 

src/common/bit_field.h

@@ -192,11 +192,6 @@ private:
     static_assert(position < 8 * sizeof(T), "Invalid position");
     static_assert(bits <= 8 * sizeof(T), "Invalid number of bits");
     static_assert(bits > 0, "Invalid number of bits");
-    static_assert(std::is_pod<T>::value, "Invalid base type");
+    static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable in a BitField");
 };
 #pragma pack()
-
-#if (__GNUC__ >= 5) || defined(__clang__) || defined(_MSC_VER)
-static_assert(std::is_trivially_copyable<BitField<0, 1, unsigned>>::value,
-              "BitField must be trivially copyable");
-#endif

src/common/vector_math.h

@@ -55,10 +55,6 @@ public:
     T x;
     T y;
 
-    T* AsArray() {
-        return &x;
-    }
-
     Vec2() = default;
     Vec2(const T& _x, const T& _y) : x(_x), y(_y) {}
 
@@ -71,11 +67,6 @@ public:
         return Vec2<T>(f, f);
     }
 
-    void Write(T a[2]) {
-        a[0] = x;
-        a[1] = y;
-    }
-
     Vec2<decltype(T{} + T{})> operator+(const Vec2& other) const {
         return MakeVec(x + other.x, y + other.y);
     }
@@ -205,10 +196,6 @@ public:
     T y;
     T z;
 
-    T* AsArray() {
-        return &x;
-    }
-
     Vec3() = default;
     Vec3(const T& _x, const T& _y, const T& _z) : x(_x), y(_y), z(_z) {}
 
@@ -225,12 +212,6 @@ public:
         return MakeVec(f, f, f);
     }
 
-    void Write(T a[3]) {
-        a[0] = x;
-        a[1] = y;
-        a[2] = z;
-    }
-
     Vec3<decltype(T{} + T{})> operator+(const Vec3& other) const {
         return MakeVec(x + other.x, y + other.y, z + other.z);
     }
@@ -416,10 +397,6 @@ public:
     T z;
     T w;
 
-    T* AsArray() {
-        return &x;
-    }
-
     Vec4() = default;
     Vec4(const T& _x, const T& _y, const T& _z, const T& _w) : x(_x), y(_y), z(_z), w(_w) {}
 
@@ -436,13 +413,6 @@ public:
         return Vec4<T>(f, f, f, f);
     }
 
-    void Write(T a[4]) {
-        a[0] = x;
-        a[1] = y;
-        a[2] = z;
-        a[3] = w;
-    }
-
     Vec4<decltype(T{} + T{})> operator+(const Vec4& other) const {
         return MakeVec(x + other.x, y + other.y, z + other.z, w + other.w);
     }

src/core/hle/kernel/svc.cpp

@@ -861,14 +861,14 @@ static const FunctionDef SVC_Table[] = {
     {0x2B, nullptr, "FlushDataCache"},
     {0x2C, nullptr, "MapPhysicalMemory"},
     {0x2D, nullptr, "UnmapPhysicalMemory"},
-    {0x2E, nullptr, "Unknown"},
+    {0x2E, nullptr, "GetNextThreadInfo"},
     {0x2F, nullptr, "GetLastThreadInfo"},
     {0x30, nullptr, "GetResourceLimitLimitValue"},
     {0x31, nullptr, "GetResourceLimitCurrentValue"},
     {0x32, SvcWrap<SetThreadActivity>, "SetThreadActivity"},
     {0x33, SvcWrap<GetThreadContext>, "GetThreadContext"},
-    {0x34, nullptr, "Unknown"},
-    {0x35, nullptr, "Unknown"},
+    {0x34, nullptr, "WaitForAddress"},
+    {0x35, nullptr, "SignalToAddress"},
     {0x36, nullptr, "Unknown"},
     {0x37, nullptr, "Unknown"},
     {0x38, nullptr, "Unknown"},
@@ -876,7 +876,7 @@ static const FunctionDef SVC_Table[] = {
     {0x3A, nullptr, "Unknown"},
     {0x3B, nullptr, "Unknown"},
     {0x3C, nullptr, "DumpInfo"},
-    {0x3D, nullptr, "Unknown"},
+    {0x3D, nullptr, "DumpInfoNew"},
     {0x3E, nullptr, "Unknown"},
     {0x3F, nullptr, "Unknown"},
     {0x40, nullptr, "CreateSession"},
@@ -887,9 +887,9 @@ static const FunctionDef SVC_Table[] = {
     {0x45, nullptr, "CreateEvent"},
     {0x46, nullptr, "Unknown"},
     {0x47, nullptr, "Unknown"},
-    {0x48, nullptr, "Unknown"},
-    {0x49, nullptr, "Unknown"},
-    {0x4A, nullptr, "Unknown"},
+    {0x48, nullptr, "AllocateUnsafeMemory"},
+    {0x49, nullptr, "FreeUnsafeMemory"},
+    {0x4A, nullptr, "SetUnsafeAllocationLimit"},
     {0x4B, nullptr, "CreateJitMemory"},
     {0x4C, nullptr, "MapJitMemory"},
     {0x4D, nullptr, "SleepSystem"},
@@ -926,7 +926,7 @@ static const FunctionDef SVC_Table[] = {
     {0x6C, nullptr, "SetHardwareBreakPoint"},
     {0x6D, nullptr, "GetDebugThreadParam"},
     {0x6E, nullptr, "Unknown"},
-    {0x6F, nullptr, "Unknown"},
+    {0x6F, nullptr, "GetMemoryInfo"},
     {0x70, nullptr, "CreatePort"},
     {0x71, nullptr, "ManageNamedPort"},
     {0x72, nullptr, "ConnectToPort"},

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

@@ -38,7 +38,10 @@ class IProfile final : public ServiceFramework<IProfile> {
 public:
     IProfile() : ServiceFramework("IProfile") {
         static const FunctionInfo functions[] = {
+            {0, nullptr, "Get"},
             {1, &IProfile::GetBase, "GetBase"},
+            {10, nullptr, "GetImageSize"},
+            {11, nullptr, "LoadImage"},
         };
         RegisterHandlers(functions);
     }
@@ -59,6 +62,11 @@ public:
         static const FunctionInfo functions[] = {
             {0, &IManagerForApplication::CheckAvailability, "CheckAvailability"},
             {1, &IManagerForApplication::GetAccountId, "GetAccountId"},
+            {2, nullptr, "EnsureIdTokenCacheAsync"},
+            {3, nullptr, "LoadIdTokenCache"},
+            {130, nullptr, "GetNintendoAccountUserResourceCacheForApplication"},
+            {150, nullptr, "CreateAuthorizationRequest"},
+            {160, nullptr, "StoreOpenContext"},
         };
         RegisterHandlers(functions);
     }

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

@@ -5,6 +5,7 @@
 #include <algorithm>
 
 #include "common/alignment.h"
+#include "common/microprofile.h"
 #include "common/scope_exit.h"
 #include "core/core.h"
 #include "core/core_timing.h"
@@ -128,6 +129,8 @@ void NVFlinger::Compose() {
         // Search for a queued buffer and acquire it
         auto buffer = buffer_queue->AcquireBuffer();
 
+        MicroProfileFlip();
+
         if (buffer == boost::none) {
             // There was no queued buffer to draw, render previous frame
             Core::System::GetInstance().perf_stats.EndGameFrame();

src/core/hle/service/set/set.cpp

@@ -36,6 +36,7 @@ SET::SET() : ServiceFramework("set") {
         {5, nullptr, "GetAvailableLanguageCodes2"},
         {6, nullptr, "GetAvailableLanguageCodeCount2"},
         {7, nullptr, "GetKeyCodeMap"},
+        {8, nullptr, "GetQuestFlag"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/set/set_cal.cpp

@@ -32,6 +32,15 @@ SET_CAL::SET_CAL() : ServiceFramework("set:cal") {
         {21, nullptr, "GetEticketDeviceKey"},
         {22, nullptr, "GetSpeakerParameter"},
         {23, nullptr, "GetLcdVendorId"},
+        {24, nullptr, "GetEciDeviceCertificate2"},
+        {25, nullptr, "GetEciDeviceKey2"},
+        {26, nullptr, "GetAmiiboKey"},
+        {27, nullptr, "GetAmiiboEcqvCertificate"},
+        {28, nullptr, "GetAmiiboEcdsaCertificate"},
+        {29, nullptr, "GetAmiiboEcqvBlsKey"},
+        {30, nullptr, "GetAmiiboEcqvBlsCertificate"},
+        {31, nullptr, "GetAmiiboEcqvBlsRootCertificate"},
+        {32, nullptr, "GetUnknownId"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/set/set_sys.cpp

@@ -27,6 +27,7 @@ SET_SYS::SET_SYS() : ServiceFramework("set:sys") {
         {2, nullptr, "GetNetworkSettings"},
         {3, nullptr, "GetFirmwareVersion"},
         {4, nullptr, "GetFirmwareVersion2"},
+        {5, nullptr, "GetFirmwareVersionDigest"},
         {7, nullptr, "GetLockScreenFlag"},
         {8, nullptr, "SetLockScreenFlag"},
         {9, nullptr, "GetBacklightSettings"},
@@ -159,6 +160,15 @@ SET_SYS::SET_SYS() : ServiceFramework("set:sys") {
         {138, nullptr, "GetWebInspectorFlag"},
         {139, nullptr, "GetAllowedSslHosts"},
         {140, nullptr, "GetHostFsMountPoint"},
+        {141, nullptr, "GetRequiresRunRepairTimeReviser"},
+        {142, nullptr, "SetRequiresRunRepairTimeReviser"},
+        {143, nullptr, "SetBlePairingSettings"},
+        {144, nullptr, "GetBlePairingSettings"},
+        {145, nullptr, "GetConsoleSixAxisSensorAngularVelocityTimeBias"},
+        {146, nullptr, "SetConsoleSixAxisSensorAngularVelocityTimeBias"},
+        {147, nullptr, "GetConsoleSixAxisSensorAngularAcceleration"},
+        {148, nullptr, "SetConsoleSixAxisSensorAngularAcceleration"},
+        {149, nullptr, "GetRebootlessSystemUpdateVersion"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/sockets/bsd.cpp

@@ -79,9 +79,34 @@ BSD::BSD(const char* name) : ServiceFramework(name) {
         {0, &BSD::RegisterClient, "RegisterClient"},
         {1, &BSD::StartMonitoring, "StartMonitoring"},
         {2, &BSD::Socket, "Socket"},
+        {3, nullptr, "SocketExempt"},
+        {4, nullptr, "Open"},
+        {5, nullptr, "Select"},
+        {6, nullptr, "Poll"},
+        {7, nullptr, "Sysctl"},
+        {8, nullptr, "Recv"},
+        {9, nullptr, "RecvFrom"},
+        {10, nullptr, "Send"},
         {11, &BSD::SendTo, "SendTo"},
+        {12, nullptr, "Accept"},
+        {13, nullptr, "Bind"},
         {14, &BSD::Connect, "Connect"},
+        {15, nullptr, "GetPeerName"},
+        {16, nullptr, "GetSockName"},
+        {17, nullptr, "GetSockOpt"},
+        {18, nullptr, "Listen"},
+        {19, nullptr, "Ioctl"},
+        {20, nullptr, "Fcntl"},
+        {21, nullptr, "SetSockOpt"},
+        {22, nullptr, "Shutdown"},
+        {23, nullptr, "ShutdownAllSockets"},
+        {24, nullptr, "Write"},
+        {25, nullptr, "Read"},
         {26, &BSD::Close, "Close"},
+        {27, nullptr, "DuplicateSocket"},
+        {28, nullptr, "GetResourceStatistics"},
+        {29, nullptr, "RecvMMsg"},
+        {30, nullptr, "SendMMsg"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/sockets/sfdnsres.cpp

@@ -30,6 +30,7 @@ SFDNSRES::SFDNSRES() : ServiceFramework("sfdnsres") {
         {7, nullptr, "GetNameInfo"},
         {8, nullptr, "RequestCancelHandle"},
         {9, nullptr, "CancelSocketCall"},
+        {11, nullptr, "ClearDnsIpServerAddressArray"},
     };
     RegisterHandlers(functions);
 }

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

@@ -33,6 +33,12 @@ SPL::SPL(std::shared_ptr<Module> module) : Module::Interface(std::move(module),
         {23, nullptr, "GetSplWaitEvent"},
         {24, nullptr, "SetSharedData"},
         {25, nullptr, "GetSharedData"},
+        {26, nullptr, "ImportSslRsaKey"},
+        {27, nullptr, "SecureExpModWithSslKey"},
+        {28, nullptr, "ImportEsRsaKey"},
+        {29, nullptr, "SecureExpModWithEsKey"},
+        {30, nullptr, "EncryptManuRsaKeyForImport"},
+        {31, nullptr, "GetPackage2Hash"},
     };
     RegisterHandlers(functions);
 }

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

@@ -20,7 +20,11 @@ public:
     ISystemClock() : ServiceFramework("ISystemClock") {
         static const FunctionInfo functions[] = {
             {0, &ISystemClock::GetCurrentTime, "GetCurrentTime"},
-            {2, &ISystemClock::GetSystemClockContext, "GetSystemClockContext"}};
+            {1, nullptr, "SetCurrentTime"},
+            {2, &ISystemClock::GetSystemClockContext, "GetSystemClockContext"},
+            {3, nullptr, "SetSystemClockContext"},
+
+        };
         RegisterHandlers(functions);
     }
 

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

@@ -14,6 +14,17 @@ TIME_S::TIME_S(std::shared_ptr<Module> time) : Module::Interface(std::move(time)
         {2, &TIME_S::GetStandardSteadyClock, "GetStandardSteadyClock"},
         {3, &TIME_S::GetTimeZoneService, "GetTimeZoneService"},
         {4, &TIME_S::GetStandardLocalSystemClock, "GetStandardLocalSystemClock"},
+        {5, nullptr, "GetEphemeralNetworkSystemClock"},
+        {50, nullptr, "SetStandardSteadyClockInternalOffset"},
+        {100, nullptr, "IsStandardUserSystemClockAutomaticCorrectionEnabled"},
+        {101, nullptr, "SetStandardUserSystemClockAutomaticCorrectionEnabled"},
+        {102, nullptr, "GetStandardUserSystemClockInitialYear"},
+        {200, nullptr, "IsStandardNetworkSystemClockAccuracySufficient"},
+        {300, nullptr, "CalculateMonotonicSystemClockBaseTimePoint"},
+        {400, nullptr, "GetClockSnapshot"},
+        {401, nullptr, "GetClockSnapshotFromSystemClockContext"},
+        {500, nullptr, "CalculateStandardUserSystemClockDifferenceByUser"},
+        {501, nullptr, "CalculateSpanBetween"},
     };
     RegisterHandlers(functions);
 }

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

@@ -14,6 +14,17 @@ TIME_U::TIME_U(std::shared_ptr<Module> time) : Module::Interface(std::move(time)
         {2, &TIME_U::GetStandardSteadyClock, "GetStandardSteadyClock"},
         {3, &TIME_U::GetTimeZoneService, "GetTimeZoneService"},
         {4, &TIME_U::GetStandardLocalSystemClock, "GetStandardLocalSystemClock"},
+        {5, nullptr, "GetEphemeralNetworkSystemClock"},
+        {50, nullptr, "SetStandardSteadyClockInternalOffset"},
+        {100, nullptr, "IsStandardUserSystemClockAutomaticCorrectionEnabled"},
+        {101, nullptr, "SetStandardUserSystemClockAutomaticCorrectionEnabled"},
+        {102, nullptr, "GetStandardUserSystemClockInitialYear"},
+        {200, nullptr, "IsStandardNetworkSystemClockAccuracySufficient"},
+        {300, nullptr, "CalculateMonotonicSystemClockBaseTimePoint"},
+        {400, nullptr, "GetClockSnapshot"},
+        {401, nullptr, "GetClockSnapshotFromSystemClockContext"},
+        {500, nullptr, "CalculateStandardUserSystemClockDifferenceByUser"},
+        {501, nullptr, "CalculateSpanBetween"},
     };
     RegisterHandlers(functions);
 }

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

@@ -580,7 +580,48 @@ public:
     ISystemDisplayService() : ServiceFramework("ISystemDisplayService") {
         static const FunctionInfo functions[] = {
             {1200, nullptr, "GetZOrderCountMin"},
+            {1202, nullptr, "GetZOrderCountMax"},
+            {1203, nullptr, "GetDisplayLogicalResolution"},
+            {1204, nullptr, "SetDisplayMagnification"},
+            {2201, nullptr, "SetLayerPosition"},
+            {2203, nullptr, "SetLayerSize"},
+            {2204, nullptr, "GetLayerZ"},
             {2205, &ISystemDisplayService::SetLayerZ, "SetLayerZ"},
+            {2207, nullptr, "SetLayerVisibility"},
+            {2209, nullptr, "SetLayerAlpha"},
+            {2312, nullptr, "CreateStrayLayer"},
+            {2400, nullptr, "OpenIndirectLayer"},
+            {2401, nullptr, "CloseIndirectLayer"},
+            {2402, nullptr, "FlipIndirectLayer"},
+            {3000, nullptr, "ListDisplayModes"},
+            {3001, nullptr, "ListDisplayRgbRanges"},
+            {3002, nullptr, "ListDisplayContentTypes"},
+            {3200, nullptr, "GetDisplayMode"},
+            {3201, nullptr, "SetDisplayMode"},
+            {3202, nullptr, "GetDisplayUnderscan"},
+            {3203, nullptr, "SetDisplayUnderscan"},
+            {3204, nullptr, "GetDisplayContentType"},
+            {3205, nullptr, "SetDisplayContentType"},
+            {3206, nullptr, "GetDisplayRgbRange"},
+            {3207, nullptr, "SetDisplayRgbRange"},
+            {3208, nullptr, "GetDisplayCmuMode"},
+            {3209, nullptr, "SetDisplayCmuMode"},
+            {3210, nullptr, "GetDisplayContrastRatio"},
+            {3211, nullptr, "SetDisplayContrastRatio"},
+            {3214, nullptr, "GetDisplayGamma"},
+            {3215, nullptr, "SetDisplayGamma"},
+            {3216, nullptr, "GetDisplayCmuLuma"},
+            {3217, nullptr, "SetDisplayCmuLuma"},
+            {8225, nullptr, "GetSharedBufferMemoryHandleId"},
+            {8250, nullptr, "OpenSharedLayer"},
+            {8251, nullptr, "CloseSharedLayer"},
+            {8252, nullptr, "ConnectSharedLayer"},
+            {8253, nullptr, "DisconnectSharedLayer"},
+            {8254, nullptr, "AcquireSharedFrameBuffer"},
+            {8255, nullptr, "PresentSharedFrameBuffer"},
+            {8256, nullptr, "GetSharedFrameBufferAcquirableEvent"},
+            {8257, nullptr, "FillSharedFrameBufferColor"},
+            {8258, nullptr, "CancelSharedFrameBuffer"},
         };
         RegisterHandlers(functions);
     }
@@ -603,10 +644,72 @@ public:
     explicit IManagerDisplayService(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger)
         : ServiceFramework("IManagerDisplayService"), nv_flinger(std::move(nv_flinger)) {
         static const FunctionInfo functions[] = {
+            {200, nullptr, "AllocateProcessHeapBlock"},
+            {201, nullptr, "FreeProcessHeapBlock"},
             {1020, &IManagerDisplayService::CloseDisplay, "CloseDisplay"},
             {1102, nullptr, "GetDisplayResolution"},
             {2010, &IManagerDisplayService::CreateManagedLayer, "CreateManagedLayer"},
+            {2011, nullptr, "DestroyManagedLayer"},
+            {2050, nullptr, "CreateIndirectLayer"},
+            {2051, nullptr, "DestroyIndirectLayer"},
+            {2052, nullptr, "CreateIndirectProducerEndPoint"},
+            {2053, nullptr, "DestroyIndirectProducerEndPoint"},
+            {2054, nullptr, "CreateIndirectConsumerEndPoint"},
+            {2055, nullptr, "DestroyIndirectConsumerEndPoint"},
+            {2300, nullptr, "AcquireLayerTexturePresentingEvent"},
+            {2301, nullptr, "ReleaseLayerTexturePresentingEvent"},
+            {2302, nullptr, "GetDisplayHotplugEvent"},
+            {2402, nullptr, "GetDisplayHotplugState"},
+            {2501, nullptr, "GetCompositorErrorInfo"},
+            {2601, nullptr, "GetDisplayErrorEvent"},
+            {4201, nullptr, "SetDisplayAlpha"},
+            {4203, nullptr, "SetDisplayLayerStack"},
+            {4205, nullptr, "SetDisplayPowerState"},
+            {4206, nullptr, "SetDefaultDisplay"},
             {6000, &IManagerDisplayService::AddToLayerStack, "AddToLayerStack"},
+            {6001, nullptr, "RemoveFromLayerStack"},
+            {6002, nullptr, "SetLayerVisibility"},
+            {6003, nullptr, "SetLayerConfig"},
+            {6004, nullptr, "AttachLayerPresentationTracer"},
+            {6005, nullptr, "DetachLayerPresentationTracer"},
+            {6006, nullptr, "StartLayerPresentationRecording"},
+            {6007, nullptr, "StopLayerPresentationRecording"},
+            {6008, nullptr, "StartLayerPresentationFenceWait"},
+            {6009, nullptr, "StopLayerPresentationFenceWait"},
+            {6010, nullptr, "GetLayerPresentationAllFencesExpiredEvent"},
+            {7000, nullptr, "SetContentVisibility"},
+            {8000, nullptr, "SetConductorLayer"},
+            {8100, nullptr, "SetIndirectProducerFlipOffset"},
+            {8200, nullptr, "CreateSharedBufferStaticStorage"},
+            {8201, nullptr, "CreateSharedBufferTransferMemory"},
+            {8202, nullptr, "DestroySharedBuffer"},
+            {8203, nullptr, "BindSharedLowLevelLayerToManagedLayer"},
+            {8204, nullptr, "BindSharedLowLevelLayerToIndirectLayer"},
+            {8207, nullptr, "UnbindSharedLowLevelLayer"},
+            {8208, nullptr, "ConnectSharedLowLevelLayerToSharedBuffer"},
+            {8209, nullptr, "DisconnectSharedLowLevelLayerFromSharedBuffer"},
+            {8210, nullptr, "CreateSharedLayer"},
+            {8211, nullptr, "DestroySharedLayer"},
+            {8216, nullptr, "AttachSharedLayerToLowLevelLayer"},
+            {8217, nullptr, "ForceDetachSharedLayerFromLowLevelLayer"},
+            {8218, nullptr, "StartDetachSharedLayerFromLowLevelLayer"},
+            {8219, nullptr, "FinishDetachSharedLayerFromLowLevelLayer"},
+            {8220, nullptr, "GetSharedLayerDetachReadyEvent"},
+            {8221, nullptr, "GetSharedLowLevelLayerSynchronizedEvent"},
+            {8222, nullptr, "CheckSharedLowLevelLayerSynchronized"},
+            {8223, nullptr, "RegisterSharedBufferImporterAruid"},
+            {8224, nullptr, "UnregisterSharedBufferImporterAruid"},
+            {8227, nullptr, "CreateSharedBufferProcessHeap"},
+            {8228, nullptr, "GetSharedLayerLayerStacks"},
+            {8229, nullptr, "SetSharedLayerLayerStacks"},
+            {8291, nullptr, "PresentDetachedSharedFrameBufferToLowLevelLayer"},
+            {8292, nullptr, "FillDetachedSharedFrameBufferColor"},
+            {8293, nullptr, "GetDetachedSharedFrameBufferImage"},
+            {8294, nullptr, "SetDetachedSharedFrameBufferImage"},
+            {8295, nullptr, "CopyDetachedSharedFrameBufferImage"},
+            {8296, nullptr, "SetDetachedSharedFrameBufferSubImage"},
+            {8297, nullptr, "GetSharedFrameBufferContentParameter"},
+            {8298, nullptr, "ExpandStartupLogoOnSharedFrameBuffer"},
         };
         RegisterHandlers(functions);
     }
@@ -825,13 +928,21 @@ IApplicationDisplayService::IApplicationDisplayService(
          "GetIndirectDisplayTransactionService"},
         {1000, &IApplicationDisplayService::ListDisplays, "ListDisplays"},
         {1010, &IApplicationDisplayService::OpenDisplay, "OpenDisplay"},
+        {1011, nullptr, "OpenDefaultDisplay"},
         {1020, &IApplicationDisplayService::CloseDisplay, "CloseDisplay"},
+        {1101, nullptr, "SetDisplayEnabled"},
         {1102, &IApplicationDisplayService::GetDisplayResolution, "GetDisplayResolution"},
-        {2101, &IApplicationDisplayService::SetLayerScalingMode, "SetLayerScalingMode"},
         {2020, &IApplicationDisplayService::OpenLayer, "OpenLayer"},
+        {2021, nullptr, "CloseLayer"},
         {2030, &IApplicationDisplayService::CreateStrayLayer, "CreateStrayLayer"},
         {2031, &IApplicationDisplayService::DestroyStrayLayer, "DestroyStrayLayer"},
+        {2101, &IApplicationDisplayService::SetLayerScalingMode, "SetLayerScalingMode"},
+        {2102, nullptr, "ConvertScalingMode"},
+        {2450, nullptr, "GetIndirectLayerImageMap"},
+        {2451, nullptr, "GetIndirectLayerImageCropMap"},
+        {2460, nullptr, "GetIndirectLayerImageRequiredMemoryInfo"},
         {5202, &IApplicationDisplayService::GetDisplayVsyncEvent, "GetDisplayVsyncEvent"},
+        {5203, nullptr, "GetDisplayVsyncEventForDebug"},
     };
     RegisterHandlers(functions);
 }

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

@@ -11,7 +11,6 @@ VI_U::VI_U(std::shared_ptr<Module> module, std::shared_ptr<NVFlinger::NVFlinger>
     : Module::Interface(std::move(module), "vi:u", std::move(nv_flinger)) {
     static const FunctionInfo functions[] = {
         {0, &VI_U::GetDisplayService, "GetDisplayService"},
-        {3, nullptr, "GetDisplayServiceWithProxyNameExchange"},
     };
     RegisterHandlers(functions);
 }

src/video_core/engines/maxwell_3d.cpp

@@ -74,8 +74,6 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
 
     regs.reg_array[method] = value;
 
-#define MAXWELL3D_REG_INDEX(field_name) (offsetof(Regs, field_name) / sizeof(u32))
-
     switch (method) {
     case MAXWELL3D_REG_INDEX(code_address.code_address_high):
     case MAXWELL3D_REG_INDEX(code_address.code_address_low): {
@@ -136,7 +134,7 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
         break;
     }
 
-#undef MAXWELL3D_REG_INDEX
+    VideoCore::g_renderer->Rasterizer()->NotifyMaxwellRegisterChanged(method);
 
     if (debug_context) {
         debug_context->OnEvent(Tegra::DebugContext::Event::MaxwellCommandProcessed, nullptr);
@@ -165,6 +163,7 @@ void Maxwell3D::ProcessQueryGet() {
 void Maxwell3D::DrawArrays() {
     LOG_DEBUG(HW_GPU, "called, topology=%d, count=%d", regs.draw.topology.Value(),
               regs.vertex_buffer.count);
+    ASSERT_MSG(!(regs.index_array.count && regs.vertex_buffer.count), "Both indexed and direct?");
 
     auto debug_context = Core::System::GetInstance().GetGPUDebugContext();
 
@@ -176,7 +175,8 @@ void Maxwell3D::DrawArrays() {
         debug_context->OnEvent(Tegra::DebugContext::Event::FinishedPrimitiveBatch, nullptr);
     }
 
-    VideoCore::g_renderer->Rasterizer()->AccelerateDrawBatch(false /*is_indexed*/);
+    const bool is_indexed{regs.index_array.count && !regs.vertex_buffer.count};
+    VideoCore::g_renderer->Rasterizer()->AccelerateDrawBatch(is_indexed);
 }
 
 void Maxwell3D::ProcessCBBind(Regs::ShaderStage stage) {
@@ -218,10 +218,12 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const {
     Texture::TICEntry tic_entry;
     Memory::ReadBlock(tic_address_cpu, &tic_entry, sizeof(Texture::TICEntry));
 
-    ASSERT_MSG(tic_entry.header_version == Texture::TICHeaderVersion::BlockLinear,
-               "TIC versions other than BlockLinear are unimplemented");
+    ASSERT_MSG(tic_entry.header_version == Texture::TICHeaderVersion::BlockLinear ||
+                   tic_entry.header_version == Texture::TICHeaderVersion::Pitch,
+               "TIC versions other than BlockLinear or Pitch are unimplemented");
 
-    ASSERT_MSG(tic_entry.texture_type == Texture::TextureType::Texture2D,
+    ASSERT_MSG((tic_entry.texture_type == Texture::TextureType::Texture2D) ||
+                   (tic_entry.texture_type == Texture::TextureType::Texture2DNoMipmap),
                "Texture types other than Texture2D are unimplemented");
 
     auto r_type = tic_entry.r_type.Value();

src/video_core/engines/maxwell_3d.h

@@ -20,6 +20,9 @@
 namespace Tegra {
 namespace Engines {
 
+#define MAXWELL3D_REG_INDEX(field_name)                                                            \
+    (offsetof(Tegra::Engines::Maxwell3D::Regs, field_name) / sizeof(u32))
+
 class Maxwell3D final {
 public:
     explicit Maxwell3D(MemoryManager& memory_manager);
@@ -248,6 +251,52 @@ public:
             Patches = 0xe,
         };
 
+        enum class IndexFormat : u32 {
+            UnsignedByte = 0x0,
+            UnsignedShort = 0x1,
+            UnsignedInt = 0x2,
+        };
+
+        struct Blend {
+            enum class Equation : u32 {
+                Add = 1,
+                Subtract = 2,
+                ReverseSubtract = 3,
+                Min = 4,
+                Max = 5,
+            };
+
+            enum class Factor : u32 {
+                Zero = 0x1,
+                One = 0x2,
+                SourceColor = 0x3,
+                OneMinusSourceColor = 0x4,
+                SourceAlpha = 0x5,
+                OneMinusSourceAlpha = 0x6,
+                DestAlpha = 0x7,
+                OneMinusDestAlpha = 0x8,
+                DestColor = 0x9,
+                OneMinusDestColor = 0xa,
+                SourceAlphaSaturate = 0xb,
+                Source1Color = 0x10,
+                OneMinusSource1Color = 0x11,
+                Source1Alpha = 0x12,
+                OneMinusSource1Alpha = 0x13,
+                ConstantColor = 0x61,
+                OneMinusConstantColor = 0x62,
+                ConstantAlpha = 0x63,
+                OneMinusConstantAlpha = 0x64,
+            };
+
+            u32 separate_alpha;
+            Equation equation_rgb;
+            Factor factor_source_rgb;
+            Factor factor_dest_rgb;
+            Equation equation_a;
+            Factor factor_source_a;
+            Factor factor_dest_a;
+        };
+
         union {
             struct {
                 INSERT_PADDING_WORDS(0x200);
@@ -270,7 +319,15 @@ public:
                     }
                 } rt[NumRenderTargets];
 
-                INSERT_PADDING_WORDS(0x80);
+                struct {
+                    f32 scale_x;
+                    f32 scale_y;
+                    f32 scale_z;
+                    u32 translate_x;
+                    u32 translate_y;
+                    u32 translate_z;
+                    INSERT_PADDING_WORDS(2);
+                } viewport_transform[NumViewports];
 
                 struct {
                     union {
@@ -375,7 +432,42 @@ public:
                     };
                 } draw;
 
-                INSERT_PADDING_WORDS(0x139);
+                INSERT_PADDING_WORDS(0x6B);
+
+                struct {
+                    u32 start_addr_high;
+                    u32 start_addr_low;
+                    u32 end_addr_high;
+                    u32 end_addr_low;
+                    IndexFormat format;
+                    u32 first;
+                    u32 count;
+
+                    unsigned FormatSizeInBytes() const {
+                        switch (format) {
+                        case IndexFormat::UnsignedByte:
+                            return 1;
+                        case IndexFormat::UnsignedShort:
+                            return 2;
+                        case IndexFormat::UnsignedInt:
+                            return 4;
+                        }
+                        UNREACHABLE();
+                    }
+
+                    GPUVAddr StartAddress() const {
+                        return static_cast<GPUVAddr>(
+                            (static_cast<GPUVAddr>(start_addr_high) << 32) | start_addr_low);
+                    }
+
+                    GPUVAddr EndAddress() const {
+                        return static_cast<GPUVAddr>((static_cast<GPUVAddr>(end_addr_high) << 32) |
+                                                     end_addr_low);
+                    }
+                } index_array;
+
+                INSERT_PADDING_WORDS(0xC7);
+
                 struct {
                     u32 query_address_high;
                     u32 query_address_low;
@@ -410,7 +502,9 @@ public:
                     }
                 } vertex_array[NumVertexArrays];
 
-                INSERT_PADDING_WORDS(0x40);
+                Blend blend;
+
+                INSERT_PADDING_WORDS(0x39);
 
                 struct {
                     u32 limit_high;
@@ -563,6 +657,7 @@ private:
                   "Field " #field_name " has invalid position")
 
 ASSERT_REG_POSITION(rt, 0x200);
+ASSERT_REG_POSITION(viewport_transform[0], 0x280);
 ASSERT_REG_POSITION(viewport, 0x300);
 ASSERT_REG_POSITION(vertex_buffer, 0x35D);
 ASSERT_REG_POSITION(zeta, 0x3F8);
@@ -572,8 +667,10 @@ ASSERT_REG_POSITION(tsc, 0x557);
 ASSERT_REG_POSITION(tic, 0x55D);
 ASSERT_REG_POSITION(code_address, 0x582);
 ASSERT_REG_POSITION(draw, 0x585);
+ASSERT_REG_POSITION(index_array, 0x5F2);
 ASSERT_REG_POSITION(query, 0x6C0);
 ASSERT_REG_POSITION(vertex_array[0], 0x700);
+ASSERT_REG_POSITION(blend, 0x780);
 ASSERT_REG_POSITION(vertex_array_limit[0], 0x7C0);
 ASSERT_REG_POSITION(shader_config[0], 0x800);
 ASSERT_REG_POSITION(const_buffer, 0x8E0);

src/video_core/engines/shader_bytecode.h

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include <cstring>
 #include <map>
 #include <string>
 #include "common/bit_field.h"
@@ -12,14 +13,10 @@ namespace Tegra {
 namespace Shader {
 
 struct Register {
-    Register() = default;
+    constexpr Register() = default;
 
     constexpr Register(u64 value) : value(value) {}
 
-    constexpr u64 GetIndex() const {
-        return value;
-    }
-
     constexpr operator u64() const {
         return value;
     }
@@ -43,13 +40,13 @@ struct Register {
     }
 
 private:
-    u64 value;
+    u64 value{};
 };
 
 union Attribute {
     Attribute() = default;
 
-    constexpr Attribute(u64 value) : value(value) {}
+    constexpr explicit Attribute(u64 value) : value(value) {}
 
     enum class Index : u64 {
         Position = 7,
@@ -68,7 +65,20 @@ union Attribute {
     } fmt28;
 
     BitField<39, 8, u64> reg;
-    u64 value;
+    u64 value{};
+};
+
+union Sampler {
+    Sampler() = default;
+
+    constexpr explicit Sampler(u64 value) : value(value) {}
+
+    enum class Index : u64 {
+        Sampler_0 = 8,
+    };
+
+    BitField<36, 13, Index> index;
+    u64 value{};
 };
 
 union Uniform {
@@ -80,6 +90,7 @@ union OpCode {
     enum class Id : u64 {
         TEXS = 0x6C,
         IPA = 0xE0,
+        FMUL32_IMM = 0x1E,
         FFMA_IMM = 0x65,
         FFMA_CR = 0x93,
         FFMA_RC = 0xA3,
@@ -132,6 +143,7 @@ union OpCode {
 
         switch (op2) {
         case Id::IPA:
+        case Id::FMUL32_IMM:
             return op2;
         }
 
@@ -225,6 +237,7 @@ union OpCode {
         info_table[Id::FMUL_R] = {Type::Arithmetic, "fmul_r"};
         info_table[Id::FMUL_C] = {Type::Arithmetic, "fmul_c"};
         info_table[Id::FMUL_IMM] = {Type::Arithmetic, "fmul_imm"};
+        info_table[Id::FMUL32_IMM] = {Type::Arithmetic, "fmul32_imm"};
         info_table[Id::FSETP_C] = {Type::Arithmetic, "fsetp_c"};
         info_table[Id::FSETP_R] = {Type::Arithmetic, "fsetp_r"};
         info_table[Id::EXIT] = {Type::Trivial, "exit"};
@@ -238,7 +251,7 @@ union OpCode {
     BitField<55, 9, Id> op3;
     BitField<52, 12, Id> op4;
     BitField<51, 13, Id> op5;
-    u64 value;
+    u64 value{};
 };
 static_assert(sizeof(OpCode) == 0x8, "Incorrect structure size");
 
@@ -280,6 +293,7 @@ enum class SubOp : u64 {
     Lg2 = 0x3,
     Rcp = 0x4,
     Rsq = 0x5,
+    Min = 0x8,
 };
 
 union Instruction {
@@ -295,15 +309,33 @@ union Instruction {
     BitField<20, 8, Register> gpr20;
     BitField<20, 7, SubOp> sub_op;
     BitField<28, 8, Register> gpr28;
-    BitField<36, 13, u64> imm36;
     BitField<39, 8, Register> gpr39;
 
     union {
+        BitField<20, 19, u64> imm20_19;
+        BitField<20, 32, u64> imm20_32;
         BitField<45, 1, u64> negate_b;
         BitField<46, 1, u64> abs_a;
         BitField<48, 1, u64> negate_a;
         BitField<49, 1, u64> abs_b;
         BitField<50, 1, u64> abs_d;
+        BitField<56, 1, u64> negate_imm;
+
+        float GetImm20_19() const {
+            float result{};
+            u32 imm{static_cast<u32>(imm20_19)};
+            imm <<= 12;
+            imm |= negate_imm ? 0x80000000 : 0;
+            std::memcpy(&result, &imm, sizeof(imm));
+            return result;
+        }
+
+        float GetImm20_32() const {
+            float result{};
+            u32 imm{static_cast<u32>(imm20_32)};
+            std::memcpy(&result, &imm, sizeof(imm));
+            return result;
+        }
     } alu;
 
     union {
@@ -311,11 +343,13 @@ union Instruction {
         BitField<49, 1, u64> negate_c;
     } ffma;
 
+    BitField<61, 1, u64> is_b_imm;
     BitField<60, 1, u64> is_b_gpr;
     BitField<59, 1, u64> is_c_gpr;
 
     Attribute attribute;
     Uniform uniform;
+    Sampler sampler;
 
     u64 hex;
 };

src/video_core/gpu.h

@@ -15,7 +15,10 @@ namespace Tegra {
 
 enum class RenderTargetFormat : u32 {
     NONE = 0x0,
+    RGBA16_FLOAT = 0xCA,
+    RGB10_A2_UNORM = 0xD1,
     RGBA8_UNORM = 0xD5,
+    RGBA8_SRGB = 0xD6,
 };
 
 class DebugContext;

src/video_core/rasterizer_interface.h

@@ -19,7 +19,7 @@ public:
     virtual void DrawArrays() = 0;
 
     /// Notify rasterizer that the specified Maxwell register has been changed
-    virtual void NotifyMaxwellRegisterChanged(u32 id) = 0;
+    virtual void NotifyMaxwellRegisterChanged(u32 method) = 0;
 
     /// Notify rasterizer that all caches should be flushed to Switch memory
     virtual void FlushAll() = 0;

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -97,7 +97,6 @@ RasterizerOpenGL::RasterizerOpenGL() {
     state.draw.vertex_buffer = stream_buffer->GetHandle();
 
     shader_program_manager = std::make_unique<GLShader::ProgramManager>();
-
     state.draw.shader_program = 0;
     state.draw.vertex_array = hw_vao.handle;
     state.Apply();
@@ -128,17 +127,6 @@ RasterizerOpenGL::~RasterizerOpenGL() {
     }
 }
 
-void RasterizerOpenGL::AnalyzeVertexArray(bool is_indexed) {
-    const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
-
-    if (is_indexed) {
-        UNREACHABLE();
-    }
-
-    // TODO(bunnei): Add support for 1+ vertex arrays
-    vs_input_size = regs.vertex_buffer.count * regs.vertex_array[0].stride;
-}
-
 void RasterizerOpenGL::SetupVertexArray(u8* array_ptr, GLintptr buffer_offset) {
     MICROPROFILE_SCOPE(OpenGL_VAO);
     const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
@@ -150,6 +138,7 @@ void RasterizerOpenGL::SetupVertexArray(u8* array_ptr, GLintptr buffer_offset) {
 
     // TODO(bunnei): Add support for 1+ vertex arrays
     const auto& vertex_array{regs.vertex_array[0]};
+    const auto& vertex_array_limit{regs.vertex_array_limit[0]};
     ASSERT_MSG(vertex_array.enable, "vertex array 0 is disabled?");
     ASSERT_MSG(!vertex_array.divisor, "vertex array 0 divisor is unimplemented!");
     for (unsigned index = 1; index < Maxwell::NumVertexArrays; ++index) {
@@ -162,6 +151,10 @@ void RasterizerOpenGL::SetupVertexArray(u8* array_ptr, GLintptr buffer_offset) {
     // to avoid OpenGL errors.
     for (unsigned index = 0; index < 16; ++index) {
         auto& attrib = regs.vertex_attrib_format[index];
+        NGLOG_DEBUG(HW_GPU, "vertex attrib {}, count={}, size={}, type={}, offset={}, normalize={}",
+                    index, attrib.ComponentCount(), attrib.SizeString(), attrib.TypeString(),
+                    attrib.offset.Value(), attrib.IsNormalized());
+
         glVertexAttribPointer(index, attrib.ComponentCount(), MaxwellToGL::VertexType(attrib),
                               attrib.IsNormalized() ? GL_TRUE : GL_FALSE, vertex_array.stride,
                               reinterpret_cast<GLvoid*>(buffer_offset + attrib.offset));
@@ -170,7 +163,7 @@ void RasterizerOpenGL::SetupVertexArray(u8* array_ptr, GLintptr buffer_offset) {
     }
 
     // Copy vertex array data
-    const u32 data_size{vertex_array.stride * regs.vertex_buffer.count};
+    const u64 data_size{vertex_array_limit.LimitAddress() - vertex_array.StartAddress() + 1};
     const VAddr data_addr{memory_manager->PhysicalToVirtualAddress(vertex_array.StartAddress())};
     res_cache.FlushRegion(data_addr, data_size, nullptr);
     Memory::ReadBlock(data_addr, array_ptr, data_size);
@@ -333,13 +326,18 @@ void RasterizerOpenGL::DrawArrays() {
 
     // Draw the vertex batch
     const bool is_indexed = accelerate_draw == AccelDraw::Indexed;
-    AnalyzeVertexArray(is_indexed);
+    const u64 index_buffer_size{regs.index_array.count * regs.index_array.FormatSizeInBytes()};
+    const unsigned vertex_num{is_indexed ? regs.index_array.count : regs.vertex_buffer.count};
+
+    // TODO(bunnei): Add support for 1+ vertex arrays
+    vs_input_size = vertex_num * regs.vertex_array[0].stride;
+
     state.draw.vertex_buffer = stream_buffer->GetHandle();
     state.Apply();
 
     size_t buffer_size = static_cast<size_t>(vs_input_size);
     if (is_indexed) {
-        UNREACHABLE();
+        buffer_size = Common::AlignUp(buffer_size, 4) + index_buffer_size;
     }
 
     // Uniform space for the 5 shader stages
@@ -354,9 +352,18 @@ void RasterizerOpenGL::DrawArrays() {
     SetupVertexArray(buffer_ptr, buffer_offset);
     ptr_pos += vs_input_size;
 
+    // If indexed mode, copy the index buffer
     GLintptr index_buffer_offset = 0;
     if (is_indexed) {
-        UNREACHABLE();
+        ptr_pos = Common::AlignUp(ptr_pos, 4);
+
+        const auto& memory_manager = Core::System().GetInstance().GPU().memory_manager;
+        const VAddr index_data_addr{
+            memory_manager->PhysicalToVirtualAddress(regs.index_array.StartAddress())};
+        Memory::ReadBlock(index_data_addr, &buffer_ptr[ptr_pos], index_buffer_size);
+
+        index_buffer_offset = buffer_offset + static_cast<GLintptr>(ptr_pos);
+        ptr_pos += index_buffer_size;
     }
 
     SetupShaders(buffer_ptr, buffer_offset, ptr_pos);
@@ -366,11 +373,16 @@ void RasterizerOpenGL::DrawArrays() {
     shader_program_manager->ApplyTo(state);
     state.Apply();
 
+    const GLenum primitive_mode{MaxwellToGL::PrimitiveTopology(regs.draw.topology)};
     if (is_indexed) {
-        UNREACHABLE();
+        const GLint index_min{static_cast<GLint>(regs.index_array.first)};
+        const GLint index_max{static_cast<GLint>(regs.index_array.first + regs.index_array.count)};
+        glDrawRangeElementsBaseVertex(primitive_mode, index_min, index_max, regs.index_array.count,
+                                      MaxwellToGL::IndexFormat(regs.index_array.format),
+                                      reinterpret_cast<const void*>(index_buffer_offset),
+                                      -index_min);
     } else {
-        glDrawArrays(MaxwellToGL::PrimitiveTopology(regs.draw.topology), 0,
-                     regs.vertex_buffer.count);
+        glDrawArrays(primitive_mode, 0, regs.vertex_buffer.count);
     }
 
     // Disable scissor test
@@ -434,7 +446,32 @@ void RasterizerOpenGL::BindTextures() {
     }
 }
 
-void RasterizerOpenGL::NotifyMaxwellRegisterChanged(u32 id) {}
+void RasterizerOpenGL::NotifyMaxwellRegisterChanged(u32 method) {
+    const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
+    switch (method) {
+    case MAXWELL3D_REG_INDEX(blend.separate_alpha):
+        ASSERT_MSG(false, "unimplemented");
+        break;
+    case MAXWELL3D_REG_INDEX(blend.equation_rgb):
+        state.blend.rgb_equation = MaxwellToGL::BlendEquation(regs.blend.equation_rgb);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.factor_source_rgb):
+        state.blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_rgb);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.factor_dest_rgb):
+        state.blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_rgb);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.equation_a):
+        state.blend.a_equation = MaxwellToGL::BlendEquation(regs.blend.equation_a);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.factor_source_a):
+        state.blend.src_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_a);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.factor_dest_a):
+        state.blend.dst_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_a);
+        break;
+    }
+}
 
 void RasterizerOpenGL::FlushAll() {
     MICROPROFILE_SCOPE(OpenGL_CacheManagement);
@@ -486,9 +523,12 @@ bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& framebu
     src_params.width = std::min(framebuffer.width, pixel_stride);
     src_params.height = framebuffer.height;
     src_params.stride = pixel_stride;
-    src_params.is_tiled = false;
+    src_params.is_tiled = true;
+    src_params.block_height = Tegra::Texture::TICEntry::DefaultBlockHeight;
     src_params.pixel_format =
         SurfaceParams::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format);
+    src_params.component_type =
+        SurfaceParams::ComponentTypeFromGPUPixelFormat(framebuffer.pixel_format);
     src_params.UpdateParams();
 
     MathUtil::Rectangle<u32> src_rect;

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -32,7 +32,7 @@ public:
     ~RasterizerOpenGL() override;
 
     void DrawArrays() override;
-    void NotifyMaxwellRegisterChanged(u32 id) override;
+    void NotifyMaxwellRegisterChanged(u32 method) override;
     void FlushAll() override;
     void FlushRegion(VAddr addr, u64 size) override;
     void InvalidateRegion(VAddr addr, u64 size) override;
@@ -155,7 +155,6 @@ private:
 
     GLsizeiptr vs_input_size;
 
-    void AnalyzeVertexArray(bool is_indexed);
     void SetupVertexArray(u8* array_ptr, GLintptr buffer_offset);
 
     std::array<OGLBuffer, Tegra::Engines::Maxwell3D::Regs::MaxShaderStage> uniform_buffers;

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -36,6 +36,7 @@
 
 using SurfaceType = SurfaceParams::SurfaceType;
 using PixelFormat = SurfaceParams::PixelFormat;
+using ComponentType = SurfaceParams::ComponentType;
 
 struct FormatTuple {
     GLint internal_format;
@@ -47,26 +48,24 @@ struct FormatTuple {
     u32 compression_factor;
 };
 
-static constexpr std::array<FormatTuple, 1> fb_format_tuples = {{
-    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, false, 1}, // RGBA8
+static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, false, 1},                     // ABGR8
+    {GL_RGB, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, false, 1},                        // B5G6R5
+    {GL_COMPRESSED_RGB_S3TC_DXT1_EXT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, true, 16},   // DXT1
+    {GL_COMPRESSED_RGBA_S3TC_DXT3_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, true, 16}, // DXT23
+    {GL_COMPRESSED_RGBA_S3TC_DXT5_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, true, 16}, // DXT45
 }};
 
-static constexpr std::array<FormatTuple, 2> tex_format_tuples = {{
-    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, false, 1},                       // RGBA8
-    {GL_COMPRESSED_RGB_S3TC_DXT1_EXT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, true, 16}, // DXT1
-}};
-
-static const FormatTuple& GetFormatTuple(PixelFormat pixel_format) {
+static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType component_type) {
     const SurfaceType type = SurfaceParams::GetFormatType(pixel_format);
-    if (type == SurfaceType::Color) {
-        ASSERT(static_cast<size_t>(pixel_format) < fb_format_tuples.size());
-        return fb_format_tuples[static_cast<unsigned int>(pixel_format)];
+    if (type == SurfaceType::ColorTexture) {
+        ASSERT(static_cast<size_t>(pixel_format) < tex_format_tuples.size());
+        // For now only UNORM components are supported
+        ASSERT(component_type == ComponentType::UNorm);
+        return tex_format_tuples[static_cast<unsigned int>(pixel_format)];
     } else if (type == SurfaceType::Depth || type == SurfaceType::DepthStencil) {
         // TODO(Subv): Implement depth formats
         ASSERT_MSG(false, "Unimplemented");
-    } else if (type == SurfaceType::Texture) {
-        ASSERT(static_cast<size_t>(pixel_format) < tex_format_tuples.size());
-        return tex_format_tuples[static_cast<unsigned int>(pixel_format)];
     }
 
     UNREACHABLE();
@@ -85,56 +84,42 @@ static u16 GetResolutionScaleFactor() {
 }
 
 template <bool morton_to_gl, PixelFormat format>
-static void MortonCopyTile(u32 stride, u8* tile_buffer, u8* gl_buffer) {
+void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, VAddr base, VAddr start,
+                VAddr end) {
     constexpr u32 bytes_per_pixel = SurfaceParams::GetFormatBpp(format) / 8;
     constexpr u32 gl_bytes_per_pixel = CachedSurface::GetGLBytesPerPixel(format);
-    for (u32 y = 0; y < 8; ++y) {
-        for (u32 x = 0; x < 8; ++x) {
-            u8* tile_ptr = tile_buffer + VideoCore::MortonInterleave(x, y) * bytes_per_pixel;
-            u8* gl_ptr = gl_buffer + ((7 - y) * stride + x) * gl_bytes_per_pixel;
-            if (morton_to_gl) {
-                std::memcpy(gl_ptr, tile_ptr, bytes_per_pixel);
-            } else {
-                std::memcpy(tile_ptr, gl_ptr, bytes_per_pixel);
-            }
-        }
+
+    if (morton_to_gl) {
+        auto data = Tegra::Texture::UnswizzleTexture(
+            base, SurfaceParams::TextureFormatFromPixelFormat(format), stride, height,
+            block_height);
+        std::memcpy(gl_buffer, data.data(), data.size());
+    } else {
+        // TODO(bunnei): Assumes the default rendering GOB size of 16 (128 lines). We should check
+        // the configuration for this and perform more generic un/swizzle
+        LOG_WARNING(Render_OpenGL, "need to use correct swizzle/GOB parameters!");
+        VideoCore::MortonCopyPixels128(stride, height, bytes_per_pixel, gl_bytes_per_pixel,
+                                       Memory::GetPointer(base), gl_buffer, morton_to_gl);
     }
 }
 
-template <bool morton_to_gl, PixelFormat format>
-void MortonCopy(u32 stride, u32 height, u8* gl_buffer, VAddr base, VAddr start, VAddr end) {
-    constexpr u32 bytes_per_pixel = SurfaceParams::GetFormatBpp(format) / 8;
-    constexpr u32 gl_bytes_per_pixel = CachedSurface::GetGLBytesPerPixel(format);
-
-    // TODO(bunnei): Assumes the default rendering GOB size of 16 (128 lines). We should check the
-    // configuration for this and perform more generic un/swizzle
-    LOG_WARNING(Render_OpenGL, "need to use correct swizzle/GOB parameters!");
-    VideoCore::MortonCopyPixels128(stride, height, bytes_per_pixel, gl_bytes_per_pixel,
-                                   Memory::GetPointer(base), gl_buffer, morton_to_gl);
-}
-
-template <>
-void MortonCopy<true, PixelFormat::DXT1>(u32 stride, u32 height, u8* gl_buffer, VAddr base,
-                                         VAddr start, VAddr end) {
-    constexpr u32 bytes_per_pixel = SurfaceParams::GetFormatBpp(PixelFormat::DXT1) / 8;
-    constexpr u32 gl_bytes_per_pixel = CachedSurface::GetGLBytesPerPixel(PixelFormat::DXT1);
-
-    // TODO(bunnei): Assumes the default rendering GOB size of 16 (128 lines). We should check the
-    // configuration for this and perform more generic un/swizzle
-    LOG_WARNING(Render_OpenGL, "need to use correct swizzle/GOB parameters!");
-    auto data =
-        Tegra::Texture::UnswizzleTexture(base, Tegra::Texture::TextureFormat::DXT1, stride, height);
-    std::memcpy(gl_buffer, data.data(), data.size());
-}
-
-static constexpr std::array<void (*)(u32, u32, u8*, VAddr, VAddr, VAddr), 2> morton_to_gl_fns = {
-    MortonCopy<true, PixelFormat::RGBA8>,
-    MortonCopy<true, PixelFormat::DXT1>,
+static constexpr std::array<void (*)(u32, u32, u32, u8*, VAddr, VAddr, VAddr),
+                            SurfaceParams::MaxPixelFormat>
+    morton_to_gl_fns = {
+        MortonCopy<true, PixelFormat::ABGR8>, MortonCopy<true, PixelFormat::B5G6R5>,
+        MortonCopy<true, PixelFormat::DXT1>,  MortonCopy<true, PixelFormat::DXT23>,
+        MortonCopy<true, PixelFormat::DXT45>,
 };
 
-static constexpr std::array<void (*)(u32, u32, u8*, VAddr, VAddr, VAddr), 2> gl_to_morton_fns = {
-    MortonCopy<false, PixelFormat::RGBA8>,
-    MortonCopy<false, PixelFormat::DXT1>,
+static constexpr std::array<void (*)(u32, u32, u32, u8*, VAddr, VAddr, VAddr),
+                            SurfaceParams::MaxPixelFormat>
+    gl_to_morton_fns = {
+        MortonCopy<false, PixelFormat::ABGR8>,
+        MortonCopy<false, PixelFormat::B5G6R5>,
+        // TODO(Subv): Swizzling the DXT1/DXT23/DXT45 formats is not yet supported
+        nullptr,
+        nullptr,
+        nullptr,
 };
 
 // Allocate an uninitialized texture of appropriate size and format for the surface
@@ -183,7 +168,7 @@ static bool BlitTextures(GLuint src_tex, const MathUtil::Rectangle<u32>& src_rec
 
     u32 buffers = 0;
 
-    if (type == SurfaceType::Color || type == SurfaceType::Texture) {
+    if (type == SurfaceType::ColorTexture) {
         glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, src_tex,
                                0);
         glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
@@ -311,15 +296,18 @@ MathUtil::Rectangle<u32> SurfaceParams::GetScaledSubRect(const SurfaceParams& su
 
 bool SurfaceParams::ExactMatch(const SurfaceParams& other_surface) const {
     return std::tie(other_surface.addr, other_surface.width, other_surface.height,
-                    other_surface.stride, other_surface.pixel_format, other_surface.is_tiled) ==
-               std::tie(addr, width, height, stride, pixel_format, is_tiled) &&
+                    other_surface.stride, other_surface.block_height, other_surface.pixel_format,
+                    other_surface.component_type,
+                    other_surface.is_tiled) == std::tie(addr, width, height, stride, block_height,
+                                                        pixel_format, component_type, is_tiled) &&
            pixel_format != PixelFormat::Invalid;
 }
 
 bool SurfaceParams::CanSubRect(const SurfaceParams& sub_surface) const {
     return sub_surface.addr >= addr && sub_surface.end <= end &&
            sub_surface.pixel_format == pixel_format && pixel_format != PixelFormat::Invalid &&
-           sub_surface.is_tiled == is_tiled &&
+           sub_surface.is_tiled == is_tiled && sub_surface.block_height == block_height &&
+           sub_surface.component_type == component_type &&
            (sub_surface.addr - addr) % BytesInPixels(is_tiled ? 64 : 1) == 0 &&
            (sub_surface.stride == stride || sub_surface.height <= (is_tiled ? 8u : 1u)) &&
            GetSubRect(sub_surface).left + sub_surface.width <= stride;
@@ -328,7 +316,8 @@ bool SurfaceParams::CanSubRect(const SurfaceParams& sub_surface) const {
 bool SurfaceParams::CanExpand(const SurfaceParams& expanded_surface) const {
     return pixel_format != PixelFormat::Invalid && pixel_format == expanded_surface.pixel_format &&
            addr <= expanded_surface.end && expanded_surface.addr <= end &&
-           is_tiled == expanded_surface.is_tiled && stride == expanded_surface.stride &&
+           is_tiled == expanded_surface.is_tiled && block_height == expanded_surface.block_height &&
+           component_type == expanded_surface.component_type && stride == expanded_surface.stride &&
            (std::max(expanded_surface.addr, addr) - std::min(expanded_surface.addr, addr)) %
                    BytesInPixels(stride * (is_tiled ? 8 : 1)) ==
                0;
@@ -339,6 +328,10 @@ bool SurfaceParams::CanTexCopy(const SurfaceParams& texcopy_params) const {
         end < texcopy_params.end) {
         return false;
     }
+    if (texcopy_params.block_height != block_height ||
+        texcopy_params.component_type != component_type)
+        return false;
+
     if (texcopy_params.width != texcopy_params.stride) {
         const u32 tile_stride = static_cast<u32>(BytesInPixels(stride * (is_tiled ? 8 : 1)));
         return (texcopy_params.addr - addr) % BytesInPixels(is_tiled ? 64 : 1) == 0 &&
@@ -481,18 +474,13 @@ void CachedSurface::LoadGLBuffer(VAddr load_start, VAddr load_end) {
     const u64 start_offset = load_start - addr;
 
     if (!is_tiled) {
-        ASSERT(type == SurfaceType::Color);
         const u32 bytes_per_pixel{GetFormatBpp() >> 3};
 
-        // TODO(bunnei): Assumes the default rendering GOB size of 16 (128 lines). We should check
-        // the configuration for this and perform more generic un/swizzle
-        LOG_WARNING(Render_OpenGL, "need to use correct swizzle/GOB parameters!");
-        VideoCore::MortonCopyPixels128(width, height, bytes_per_pixel, 4,
-                                       texture_src_data + start_offset, &gl_buffer[start_offset],
-                                       true);
+        std::memcpy(&gl_buffer[start_offset], texture_src_data + start_offset,
+                    bytes_per_pixel * width * height);
     } else {
-        morton_to_gl_fns[static_cast<size_t>(pixel_format)](stride, height, &gl_buffer[0], addr,
-                                                            load_start, load_end);
+        morton_to_gl_fns[static_cast<size_t>(pixel_format)](
+            stride, block_height, height, &gl_buffer[0], addr, load_start, load_end);
     }
 }
 
@@ -533,11 +521,10 @@ void CachedSurface::FlushGLBuffer(VAddr flush_start, VAddr flush_end) {
         if (backup_bytes)
             std::memcpy(&dst_buffer[coarse_start_offset], &backup_data[0], backup_bytes);
     } else if (!is_tiled) {
-        ASSERT(type == SurfaceType::Color);
         std::memcpy(dst_buffer + start_offset, &gl_buffer[start_offset], flush_end - flush_start);
     } else {
-        gl_to_morton_fns[static_cast<size_t>(pixel_format)](stride, height, &gl_buffer[0], addr,
-                                                            flush_start, flush_end);
+        gl_to_morton_fns[static_cast<size_t>(pixel_format)](
+            stride, block_height, height, &gl_buffer[0], addr, flush_start, flush_end);
     }
 }
 
@@ -556,7 +543,7 @@ void CachedSurface::UploadGLTexture(const MathUtil::Rectangle<u32>& rect, GLuint
     GLint y0 = static_cast<GLint>(rect.bottom);
     size_t buffer_offset = (y0 * stride + x0) * GetGLBytesPerPixel(pixel_format);
 
-    const FormatTuple& tuple = GetFormatTuple(pixel_format);
+    const FormatTuple& tuple = GetFormatTuple(pixel_format, component_type);
     GLuint target_tex = texture.handle;
 
     // If not 1x scale, create 1x texture that we will blit from to replace texture subrect in
@@ -629,7 +616,7 @@ void CachedSurface::DownloadGLTexture(const MathUtil::Rectangle<u32>& rect, GLui
     OpenGLState prev_state = state;
     SCOPE_EXIT({ prev_state.Apply(); });
 
-    const FormatTuple& tuple = GetFormatTuple(pixel_format);
+    const FormatTuple& tuple = GetFormatTuple(pixel_format, component_type);
 
     // Ensure no bad interactions with GL_PACK_ALIGNMENT
     ASSERT(stride * GetGLBytesPerPixel(pixel_format) % 4 == 0);
@@ -662,7 +649,7 @@ void CachedSurface::DownloadGLTexture(const MathUtil::Rectangle<u32>& rect, GLui
         state.draw.read_framebuffer = read_fb_handle;
         state.Apply();
 
-        if (type == SurfaceType::Color || type == SurfaceType::Texture) {
+        if (type == SurfaceType::ColorTexture) {
             glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D,
                                    texture.handle, 0);
             glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
@@ -1041,9 +1028,25 @@ Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextu
     params.height = config.tic.Height();
     params.is_tiled = config.tic.IsTiled();
     params.pixel_format = SurfaceParams::PixelFormatFromTextureFormat(config.tic.format);
+
+    // TODO(Subv): Different types per component are not supported.
+    ASSERT(config.tic.r_type.Value() == config.tic.g_type.Value() &&
+           config.tic.r_type.Value() == config.tic.b_type.Value() &&
+           config.tic.r_type.Value() == config.tic.a_type.Value());
+
+    params.component_type = SurfaceParams::ComponentTypeFromTexture(config.tic.r_type.Value());
+
+    if (config.tic.IsTiled()) {
+        params.block_height = config.tic.BlockHeight();
+    } else {
+        // Use the texture-provided stride value if the texture isn't tiled.
+        params.stride = params.PixelsInBytes(config.tic.Pitch());
+    }
+
     params.UpdateParams();
 
-    if (config.tic.Width() % 8 != 0 || config.tic.Height() % 8 != 0) {
+    if (config.tic.Width() % 8 != 0 || config.tic.Height() % 8 != 0 ||
+        params.stride != params.width) {
         Surface src_surface;
         MathUtil::Rectangle<u32> rect;
         std::tie(src_surface, rect) = GetSurfaceSubRect(params, ScaleMatch::Ignore, true);
@@ -1094,10 +1097,13 @@ SurfaceSurfaceRect_Tuple RasterizerCacheOpenGL::GetFramebufferSurfaces(
     color_params.res_scale = resolution_scale_factor;
     color_params.width = config.width;
     color_params.height = config.height;
+    // TODO(Subv): Can framebuffers use a different block height?
+    color_params.block_height = Tegra::Texture::TICEntry::DefaultBlockHeight;
     SurfaceParams depth_params = color_params;
 
     color_params.addr = memory_manager->PhysicalToVirtualAddress(config.Address());
     color_params.pixel_format = SurfaceParams::PixelFormatFromRenderTargetFormat(config.format);
+    color_params.component_type = SurfaceParams::ComponentTypeFromRenderTarget(config.format);
     color_params.UpdateParams();
 
     ASSERT_MSG(!using_depth_fb, "depth buffer is unimplemented");
@@ -1293,7 +1299,6 @@ void RasterizerCacheOpenGL::InvalidateRegion(VAddr addr, u64 size, const Surface
     const SurfaceInterval invalid_interval(addr, addr + size);
 
     if (region_owner != nullptr) {
-        ASSERT(region_owner->type != SurfaceType::Texture);
         ASSERT(addr >= region_owner->addr && addr + size <= region_owner->end);
         // Surfaces can't have a gap
         ASSERT(region_owner->width == region_owner->stride);
@@ -1355,7 +1360,8 @@ Surface RasterizerCacheOpenGL::CreateSurface(const SurfaceParams& params) {
 
     surface->gl_buffer_size = 0;
     surface->invalid_regions.insert(surface->GetInterval());
-    AllocateSurfaceTexture(surface->texture.handle, GetFormatTuple(surface->pixel_format),
+    AllocateSurfaceTexture(surface->texture.handle,
+                           GetFormatTuple(surface->pixel_format, surface->component_type),
                            surface->GetScaledWidth(), surface->GetScaledHeight());
 
     return surface;

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -52,27 +52,45 @@ enum class ScaleMatch {
 
 struct SurfaceParams {
     enum class PixelFormat {
-        RGBA8 = 0,
-        DXT1 = 1,
+        ABGR8 = 0,
+        B5G6R5 = 1,
+        DXT1 = 2,
+        DXT23 = 3,
+        DXT45 = 4,
+
+        Max,
         Invalid = 255,
     };
 
+    static constexpr size_t MaxPixelFormat = static_cast<size_t>(PixelFormat::Max);
+
+    enum class ComponentType {
+        Invalid = 0,
+        SNorm = 1,
+        UNorm = 2,
+        SInt = 3,
+        UInt = 4,
+        Float = 5,
+    };
+
     enum class SurfaceType {
-        Color = 0,
-        Texture = 1,
-        Depth = 2,
-        DepthStencil = 3,
-        Fill = 4,
-        Invalid = 5
+        ColorTexture = 0,
+        Depth = 1,
+        DepthStencil = 2,
+        Fill = 3,
+        Invalid = 4,
     };
 
     static constexpr unsigned int GetFormatBpp(PixelFormat format) {
         if (format == PixelFormat::Invalid)
             return 0;
 
-        constexpr std::array<unsigned int, 2> bpp_table = {
-            32, // RGBA8
-            64, // DXT1
+        constexpr std::array<unsigned int, MaxPixelFormat> bpp_table = {
+            32,  // ABGR8
+            16,  // B5G6R5
+            64,  // DXT1
+            128, // DXT23
+            128, // DXT45
         };
 
         ASSERT(static_cast<size_t>(format) < bpp_table.size());
@@ -85,8 +103,9 @@ struct SurfaceParams {
     static PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
         switch (format) {
         case Tegra::RenderTargetFormat::RGBA8_UNORM:
-            return PixelFormat::RGBA8;
+            return PixelFormat::ABGR8;
         default:
+            NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
         }
     }
@@ -94,8 +113,9 @@ struct SurfaceParams {
     static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
         switch (format) {
         case Tegra::FramebufferConfig::PixelFormat::ABGR8:
-            return PixelFormat::RGBA8;
+            return PixelFormat::ABGR8;
         default:
+            NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
         }
     }
@@ -104,10 +124,69 @@ struct SurfaceParams {
         // TODO(Subv): Properly implement this
         switch (format) {
         case Tegra::Texture::TextureFormat::A8R8G8B8:
-            return PixelFormat::RGBA8;
+            return PixelFormat::ABGR8;
+        case Tegra::Texture::TextureFormat::B5G6R5:
+            return PixelFormat::B5G6R5;
         case Tegra::Texture::TextureFormat::DXT1:
             return PixelFormat::DXT1;
+        case Tegra::Texture::TextureFormat::DXT23:
+            return PixelFormat::DXT23;
+        case Tegra::Texture::TextureFormat::DXT45:
+            return PixelFormat::DXT45;
         default:
+            NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+            UNREACHABLE();
+        }
+    }
+
+    static Tegra::Texture::TextureFormat TextureFormatFromPixelFormat(PixelFormat format) {
+        // TODO(Subv): Properly implement this
+        switch (format) {
+        case PixelFormat::ABGR8:
+            return Tegra::Texture::TextureFormat::A8R8G8B8;
+        case PixelFormat::B5G6R5:
+            return Tegra::Texture::TextureFormat::B5G6R5;
+        case PixelFormat::DXT1:
+            return Tegra::Texture::TextureFormat::DXT1;
+        case PixelFormat::DXT23:
+            return Tegra::Texture::TextureFormat::DXT23;
+        case PixelFormat::DXT45:
+            return Tegra::Texture::TextureFormat::DXT45;
+        default:
+            UNREACHABLE();
+        }
+    }
+
+    static ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
+        // TODO(Subv): Implement more component types
+        switch (type) {
+        case Tegra::Texture::ComponentType::UNORM:
+            return ComponentType::UNorm;
+        default:
+            NGLOG_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::RGB10_A2_UNORM:
+            return ComponentType::UNorm;
+        default:
+            NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+            UNREACHABLE();
+        }
+    }
+
+    static ComponentType ComponentTypeFromGPUPixelFormat(
+        Tegra::FramebufferConfig::PixelFormat format) {
+        switch (format) {
+        case Tegra::FramebufferConfig::PixelFormat::ABGR8:
+            return ComponentType::UNorm;
+        default:
+            NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
         }
     }
@@ -116,8 +195,7 @@ struct SurfaceParams {
         SurfaceType a_type = GetFormatType(pixel_format_a);
         SurfaceType b_type = GetFormatType(pixel_format_b);
 
-        if ((a_type == SurfaceType::Color || a_type == SurfaceType::Texture) &&
-            (b_type == SurfaceType::Color || b_type == SurfaceType::Texture)) {
+        if (a_type == SurfaceType::ColorTexture && b_type == SurfaceType::ColorTexture) {
             return true;
         }
 
@@ -133,12 +211,8 @@ struct SurfaceParams {
     }
 
     static SurfaceType GetFormatType(PixelFormat pixel_format) {
-        if ((unsigned int)pixel_format <= static_cast<unsigned int>(PixelFormat::RGBA8)) {
-            return SurfaceType::Color;
-        }
-
-        if ((unsigned int)pixel_format <= static_cast<unsigned int>(PixelFormat::DXT1)) {
-            return SurfaceType::Texture;
+        if (static_cast<size_t>(pixel_format) < MaxPixelFormat) {
+            return SurfaceType::ColorTexture;
         }
 
         // TODO(Subv): Implement the other formats
@@ -210,11 +284,13 @@ struct SurfaceParams {
     u32 width = 0;
     u32 height = 0;
     u32 stride = 0;
+    u32 block_height = 0;
     u16 res_scale = 1;
 
     bool is_tiled = false;
     PixelFormat pixel_format = PixelFormat::Invalid;
     SurfaceType type = SurfaceType::Invalid;
+    ComponentType component_type = ComponentType::Invalid;
 };
 
 struct CachedSurface : SurfaceParams {

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -17,6 +17,7 @@ using Tegra::Shader::Attribute;
 using Tegra::Shader::Instruction;
 using Tegra::Shader::OpCode;
 using Tegra::Shader::Register;
+using Tegra::Shader::Sampler;
 using Tegra::Shader::SubOp;
 using Tegra::Shader::Uniform;
 
@@ -155,23 +156,27 @@ private:
 
     /// Generates code representing an input attribute register.
     std::string GetInputAttribute(Attribute::Index attribute) {
-        declr_input_attribute.insert(attribute);
+        switch (attribute) {
+        case Attribute::Index::Position:
+            return "position";
+        default:
+            const u32 index{static_cast<u32>(attribute) -
+                            static_cast<u32>(Attribute::Index::Attribute_0)};
+            if (attribute >= Attribute::Index::Attribute_0) {
+                declr_input_attribute.insert(attribute);
+                return "input_attribute_" + std::to_string(index);
+            }
 
-        const u32 index{static_cast<u32>(attribute) -
-                        static_cast<u32>(Attribute::Index::Attribute_0)};
-        if (attribute >= Attribute::Index::Attribute_0) {
-            return "input_attribute_" + std::to_string(index);
+            NGLOG_CRITICAL(HW_GPU, "Unhandled input attribute: {}", index);
+            UNREACHABLE();
         }
-
-        LOG_CRITICAL(HW_GPU, "Unhandled input attribute: 0x%02x", index);
-        UNREACHABLE();
     }
 
     /// Generates code representing an output attribute register.
     std::string GetOutputAttribute(Attribute::Index attribute) {
         switch (attribute) {
         case Attribute::Index::Position:
-            return "gl_Position";
+            return "position";
         default:
             const u32 index{static_cast<u32>(attribute) -
                             static_cast<u32>(Attribute::Index::Attribute_0)};
@@ -180,22 +185,47 @@ private:
                 return "output_attribute_" + std::to_string(index);
             }
 
-            LOG_CRITICAL(HW_GPU, "Unhandled output attribute: 0x%02x", index);
+            NGLOG_CRITICAL(HW_GPU, "Unhandled output attribute: {}", index);
             UNREACHABLE();
         }
     }
 
+    /// Generates code representing a 19-bit immediate value
+    static std::string GetImmediate19(const Instruction& instr) {
+        return std::to_string(instr.alu.GetImm20_19());
+    }
+
+    /// Generates code representing a 32-bit immediate value
+    static std::string GetImmediate32(const Instruction& instr) {
+        return std::to_string(instr.alu.GetImm20_32());
+    }
+
     /// Generates code representing a temporary (GPR) register.
-    std::string GetRegister(const Register& reg) {
-        return *declr_register.insert("register_" + std::to_string(reg)).first;
+    std::string GetRegister(const Register& reg, unsigned elem = 0) {
+        if (stage == Maxwell3D::Regs::ShaderStage::Fragment && reg < 4) {
+            // GPRs 0-3 are output color for the fragment shader
+            return std::string{"color."} + "rgba"[(reg + elem) & 3];
+        }
+
+        return *declr_register.insert("register_" + std::to_string(reg + elem)).first;
     }
 
     /// Generates code representing a uniform (C buffer) register.
     std::string GetUniform(const Uniform& reg) {
-        declr_const_buffers[reg.index].MarkAsUsed(reg.index, reg.offset, stage);
+        declr_const_buffers[reg.index].MarkAsUsed(static_cast<unsigned>(reg.index),
+                                                  static_cast<unsigned>(reg.offset), stage);
         return 'c' + std::to_string(reg.index) + '[' + std::to_string(reg.offset) + ']';
     }
 
+    /// Generates code representing a texture sampler.
+    std::string GetSampler(const Sampler& sampler) const {
+        // TODO(Subv): Support more than just texture sampler 0
+        ASSERT_MSG(sampler.index == Sampler::Index::Sampler_0, "unsupported");
+        const unsigned index{static_cast<unsigned>(sampler.index.Value()) -
+                             static_cast<unsigned>(Sampler::Index::Sampler_0)};
+        return "tex[" + std::to_string(index) + "]";
+    }
+
     /**
      * Adds code that calls a subroutine.
      * @param subroutine the subroutine to call.
@@ -217,12 +247,13 @@ private:
      * @param value the code representing the value to assign.
      */
     void SetDest(u64 elem, const std::string& reg, const std::string& value,
-                 u64 dest_num_components, u64 value_num_components) {
+                 u64 dest_num_components, u64 value_num_components, bool is_abs = false) {
         std::string swizzle = ".";
         swizzle += "xyzw"[elem];
 
         std::string dest = reg + (dest_num_components != 1 ? swizzle : "");
         std::string src = "(" + value + ")" + (value_num_components != 1 ? swizzle : "");
+        src = is_abs ? "abs(" + src + ")" : src;
 
         shader.AddLine(dest + " = " + src + ";");
     }
@@ -240,8 +271,6 @@ private:
 
         switch (OpCode::GetInfo(instr.opcode).type) {
         case OpCode::Type::Arithmetic: {
-            ASSERT(!instr.alu.abs_d);
-
             std::string dest = GetRegister(instr.gpr0);
             std::string op_a = instr.alu.negate_a ? "-" : "";
             op_a += GetRegister(instr.gpr8);
@@ -250,63 +279,114 @@ private:
             }
 
             std::string op_b = instr.alu.negate_b ? "-" : "";
-            if (instr.is_b_gpr) {
-                op_b += GetRegister(instr.gpr20);
+
+            if (instr.is_b_imm) {
+                op_b += GetImmediate19(instr);
             } else {
-                op_b += GetUniform(instr.uniform);
+                if (instr.is_b_gpr) {
+                    op_b += GetRegister(instr.gpr20);
+                } else {
+                    op_b += GetUniform(instr.uniform);
+                }
             }
+
             if (instr.alu.abs_b) {
                 op_b = "abs(" + op_b + ")";
             }
 
             switch (instr.opcode.EffectiveOpCode()) {
             case OpCode::Id::FMUL_C:
-            case OpCode::Id::FMUL_R: {
-                SetDest(0, dest, op_a + " * " + op_b, 1, 1);
+            case OpCode::Id::FMUL_R:
+            case OpCode::Id::FMUL_IMM: {
+                SetDest(0, dest, op_a + " * " + op_b, 1, 1, instr.alu.abs_d);
+                break;
+            }
+            case OpCode::Id::FMUL32_IMM: {
+                // fmul32i doesn't have abs or neg bits.
+                SetDest(0, dest, GetRegister(instr.gpr8) + " * " + GetImmediate32(instr), 1, 1);
                 break;
             }
             case OpCode::Id::FADD_C:
-            case OpCode::Id::FADD_R: {
-                SetDest(0, dest, op_a + " + " + op_b, 1, 1);
+            case OpCode::Id::FADD_R:
+            case OpCode::Id::FADD_IMM: {
+                SetDest(0, dest, op_a + " + " + op_b, 1, 1, instr.alu.abs_d);
+                break;
+            }
+            case OpCode::Id::MUFU: {
+                switch (instr.sub_op) {
+                case SubOp::Cos:
+                    SetDest(0, dest, "cos(" + op_a + ")", 1, 1, instr.alu.abs_d);
+                    break;
+                case SubOp::Sin:
+                    SetDest(0, dest, "sin(" + op_a + ")", 1, 1, instr.alu.abs_d);
+                    break;
+                case SubOp::Ex2:
+                    SetDest(0, dest, "exp2(" + op_a + ")", 1, 1, instr.alu.abs_d);
+                    break;
+                case SubOp::Lg2:
+                    SetDest(0, dest, "log2(" + op_a + ")", 1, 1, instr.alu.abs_d);
+                    break;
+                case SubOp::Rcp:
+                    SetDest(0, dest, "1.0 / " + op_a, 1, 1, instr.alu.abs_d);
+                    break;
+                case SubOp::Rsq:
+                    SetDest(0, dest, "inversesqrt(" + op_a + ")", 1, 1, instr.alu.abs_d);
+                    break;
+                case SubOp::Min:
+                    SetDest(0, dest, "min(" + op_a + "," + op_b + ")", 1, 1, instr.alu.abs_d);
+                    break;
+                default:
+                    NGLOG_CRITICAL(HW_GPU, "Unhandled MUFU sub op: {}",
+                                   static_cast<unsigned>(instr.sub_op.Value()));
+                    UNREACHABLE();
+                }
                 break;
             }
             default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled arithmetic instruction: 0x%02x (%s): 0x%08x",
-                             static_cast<unsigned>(instr.opcode.EffectiveOpCode()),
-                             OpCode::GetInfo(instr.opcode).name.c_str(), instr.hex);
-                throw DecompileFail("Unhandled instruction");
-                break;
+                NGLOG_CRITICAL(HW_GPU, "Unhandled arithmetic instruction: {} ({}): {}",
+                               static_cast<unsigned>(instr.opcode.EffectiveOpCode()),
+                               OpCode::GetInfo(instr.opcode).name, instr.hex);
+                UNREACHABLE();
             }
             }
             break;
         }
         case OpCode::Type::Ffma: {
-            ASSERT_MSG(!instr.ffma.negate_b, "untested");
-            ASSERT_MSG(!instr.ffma.negate_c, "untested");
-
             std::string dest = GetRegister(instr.gpr0);
             std::string op_a = GetRegister(instr.gpr8);
-
             std::string op_b = instr.ffma.negate_b ? "-" : "";
-            op_b += GetUniform(instr.uniform);
-
             std::string op_c = instr.ffma.negate_c ? "-" : "";
-            op_c += GetRegister(instr.gpr39);
 
             switch (instr.opcode.EffectiveOpCode()) {
             case OpCode::Id::FFMA_CR: {
-                SetDest(0, dest, op_a + " * " + op_b + " + " + op_c, 1, 1);
+                op_b += GetUniform(instr.uniform);
+                op_c += GetRegister(instr.gpr39);
                 break;
             }
+            case OpCode::Id::FFMA_RR: {
+                op_b += GetRegister(instr.gpr20);
+                op_c += GetRegister(instr.gpr39);
+                break;
+            }
+            case OpCode::Id::FFMA_RC: {
+                op_b += GetRegister(instr.gpr39);
+                op_c += GetUniform(instr.uniform);
+                break;
+            }
+            case OpCode::Id::FFMA_IMM: {
+                op_b += GetImmediate19(instr);
+                op_c += GetRegister(instr.gpr39);
+                break;
+            }
+            default: {
+                NGLOG_CRITICAL(HW_GPU, "Unhandled FFMA instruction: {} ({}): {}",
+                               static_cast<unsigned>(instr.opcode.EffectiveOpCode()),
+                               OpCode::GetInfo(instr.opcode).name, instr.hex);
+                UNREACHABLE();
+            }
+            }
 
-            default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled arithmetic FFMA instruction: 0x%02x (%s): 0x%08x",
-                             static_cast<unsigned>(instr.opcode.EffectiveOpCode()),
-                             OpCode::GetInfo(instr.opcode).name.c_str(), instr.hex);
-                throw DecompileFail("Unhandled instruction");
-                break;
-            }
-            }
+            SetDest(0, dest, op_a + " * " + op_b + " + " + op_c, 1, 1);
             break;
         }
         case OpCode::Type::Memory: {
@@ -315,22 +395,40 @@ private:
 
             switch (instr.opcode.EffectiveOpCode()) {
             case OpCode::Id::LD_A: {
-                ASSERT(instr.attribute.fmt20.size == 0);
+                ASSERT_MSG(instr.attribute.fmt20.size == 0, "untested");
                 SetDest(instr.attribute.fmt20.element, gpr0, GetInputAttribute(attribute), 1, 4);
                 break;
             }
             case OpCode::Id::ST_A: {
-                ASSERT(instr.attribute.fmt20.size == 0);
+                ASSERT_MSG(instr.attribute.fmt20.size == 0, "untested");
                 SetDest(instr.attribute.fmt20.element, GetOutputAttribute(attribute), gpr0, 4, 1);
                 break;
             }
-            default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: 0x%02x (%s): 0x%08x",
-                             static_cast<unsigned>(instr.opcode.EffectiveOpCode()),
-                             OpCode::GetInfo(instr.opcode).name.c_str(), instr.hex);
-                throw DecompileFail("Unhandled instruction");
+            case OpCode::Id::TEXS: {
+                ASSERT_MSG(instr.attribute.fmt20.size == 4, "untested");
+                const std::string op_a = GetRegister(instr.gpr8);
+                const std::string op_b = GetRegister(instr.gpr20);
+                const std::string sampler = GetSampler(instr.sampler);
+                const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                // Add an extra scope and declare the texture coords inside to prevent overwriting
+                // them in case they are used as outputs of the texs instruction.
+                shader.AddLine("{");
+                ++shader.scope;
+                shader.AddLine(coord);
+                const std::string texture = "texture(" + sampler + ", coords)";
+                for (unsigned elem = 0; elem < instr.attribute.fmt20.size; ++elem) {
+                    SetDest(elem, GetRegister(instr.gpr0, elem), texture, 1, 4);
+                }
+                --shader.scope;
+                shader.AddLine("}");
                 break;
             }
+            default: {
+                NGLOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {} ({}): {}",
+                               static_cast<unsigned>(instr.opcode.EffectiveOpCode()),
+                               OpCode::GetInfo(instr.opcode).name, instr.hex);
+                UNREACHABLE();
+            }
             }
             break;
         }
@@ -342,14 +440,18 @@ private:
                 offset = PROGRAM_END - 1;
                 break;
             }
-
-            default: {
-                LOG_CRITICAL(HW_GPU, "Unhandled instruction: 0x%02x (%s): 0x%08x",
-                             static_cast<unsigned>(instr.opcode.EffectiveOpCode()),
-                             OpCode::GetInfo(instr.opcode).name.c_str(), instr.hex);
-                throw DecompileFail("Unhandled instruction");
+            case OpCode::Id::IPA: {
+                const auto& attribute = instr.attribute.fmt28;
+                std::string dest = GetRegister(instr.gpr0);
+                SetDest(attribute.element, dest, GetInputAttribute(attribute.index), 1, 4);
                 break;
             }
+            default: {
+                NGLOG_CRITICAL(HW_GPU, "Unhandled instruction: {} ({}): {}",
+                               static_cast<unsigned>(instr.opcode.EffectiveOpCode()),
+                               OpCode::GetInfo(instr.opcode).name, instr.hex);
+                UNREACHABLE();
+            }
             }
 
             break;
@@ -501,7 +603,7 @@ private:
     std::set<Attribute::Index> declr_input_attribute;
     std::set<Attribute::Index> declr_output_attribute;
     std::array<ConstBufferEntry, Maxwell3D::Regs::MaxConstBuffers> declr_const_buffers;
-};
+}; // namespace Decompiler
 
 std::string GetCommonDeclarations() {
     return "bool exec_shader();";
@@ -514,7 +616,7 @@ boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code,
         GLSLGenerator generator(subroutines, program_code, main_offset, stage);
         return ProgramResult{generator.GetShaderCode(), generator.GetEntries()};
     } catch (const DecompileFail& exception) {
-        LOG_ERROR(HW_GPU, "Shader decompilation failed: %s", exception.what());
+        NGLOG_ERROR(HW_GPU, "Shader decompilation failed: {}", exception.what());
     }
     return boost::none;
 }

src/video_core/renderer_opengl/gl_shader_gen.cpp

@@ -27,10 +27,19 @@ out gl_PerVertex {
     vec4 gl_Position;
 };
 
+out vec4 position;
+
+layout (std140) uniform vs_config {
+    vec4 viewport_flip;
+};
+
 void main() {
     exec_shader();
-}
 
+    // Viewport can be flipped, which is unsupported by glViewport
+    position.xy *= viewport_flip.xy;
+    gl_Position = position;
+}
 )";
     out += program.first;
     return {out, program.second};
@@ -46,8 +55,13 @@ ProgramResult GenerateFragmentShader(const ShaderSetup& setup, const MaxwellFSCo
                                 .get_value_or({});
     out += R"(
 
+in vec4 position;
 out vec4 color;
 
+layout (std140) uniform fs_config {
+    vec4 viewport_flip;
+};
+
 uniform sampler2D tex[32];
 
 void main() {

src/video_core/renderer_opengl/gl_shader_manager.cpp

@@ -53,6 +53,12 @@ void SetShaderSamplerBindings(GLuint shader) {
 
 } // namespace Impl
 
-void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage) {}
+void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage) {
+    const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
+
+    // TODO(bunnei): Support more than one viewport
+    viewport_flip[0] = regs.viewport_transform[0].scale_x < 0.0 ? -1.0 : 1.0;
+    viewport_flip[1] = regs.viewport_transform[0].scale_y < 0.0 ? -1.0 : 1.0;
+}
 
 } // namespace GLShader

src/video_core/renderer_opengl/gl_shader_manager.h

@@ -30,10 +30,9 @@ void SetShaderSamplerBindings(GLuint shader);
 //       Not following that rule will cause problems on some AMD drivers.
 struct MaxwellUniformData {
     void SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage);
-    // TODO(Subv): Use this for something.
+    alignas(16) GLvec4 viewport_flip;
 };
-// static_assert(sizeof(MaxwellUniformData) == 1024, "MaxwellUniformData structure size is
-// incorrect");
+static_assert(sizeof(MaxwellUniformData) == 16, "MaxwellUniformData structure size is incorrect");
 static_assert(sizeof(MaxwellUniformData) < 16384,
               "MaxwellUniformData structure must be less than 16kb as per the OpenGL spec");
 

src/video_core/renderer_opengl/maxwell_to_gl.h

@@ -31,7 +31,7 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) {
             return GL_UNSIGNED_BYTE;
         }
 
-        LOG_CRITICAL(Render_OpenGL, "Unimplemented vertex size=%s", attrib.SizeString().c_str());
+        NGLOG_CRITICAL(Render_OpenGL, "Unimplemented vertex size={}", attrib.SizeString());
         UNREACHABLE();
         return {};
     }
@@ -40,7 +40,21 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) {
         return GL_FLOAT;
     }
 
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented vertex type=%s", attrib.TypeString().c_str());
+    NGLOG_CRITICAL(Render_OpenGL, "Unimplemented vertex type={}", attrib.TypeString());
+    UNREACHABLE();
+    return {};
+}
+
+inline GLenum IndexFormat(Maxwell::IndexFormat index_format) {
+    switch (index_format) {
+    case Maxwell::IndexFormat::UnsignedByte:
+        return GL_UNSIGNED_BYTE;
+    case Maxwell::IndexFormat::UnsignedShort:
+        return GL_UNSIGNED_SHORT;
+    case Maxwell::IndexFormat::UnsignedInt:
+        return GL_UNSIGNED_INT;
+    }
+    NGLOG_CRITICAL(Render_OpenGL, "Unimplemented index_format={}", static_cast<u32>(index_format));
     UNREACHABLE();
     return {};
 }
@@ -52,7 +66,7 @@ inline GLenum PrimitiveTopology(Maxwell::PrimitiveTopology topology) {
     case Maxwell::PrimitiveTopology::TriangleStrip:
         return GL_TRIANGLE_STRIP;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented primitive topology=%d", topology);
+    NGLOG_CRITICAL(Render_OpenGL, "Unimplemented topology={}", static_cast<u32>(topology));
     UNREACHABLE();
     return {};
 }
@@ -64,18 +78,90 @@ inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode) {
     case Tegra::Texture::TextureFilter::Nearest:
         return GL_NEAREST;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented texture filter mode=%u",
-                 static_cast<u32>(filter_mode));
+    NGLOG_CRITICAL(Render_OpenGL, "Unimplemented texture filter mode={}",
+                   static_cast<u32>(filter_mode));
     UNREACHABLE();
     return {};
 }
 
 inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) {
     switch (wrap_mode) {
+    case Tegra::Texture::WrapMode::Wrap:
+        return GL_REPEAT;
     case Tegra::Texture::WrapMode::ClampToEdge:
         return GL_CLAMP_TO_EDGE;
+    case Tegra::Texture::WrapMode::ClampOGL:
+        // TODO(Subv): GL_CLAMP was removed as of OpenGL 3.1, to implement GL_CLAMP, we can use
+        // GL_CLAMP_TO_BORDER to get the border color of the texture, and then sample the edge to
+        // manually mix them. However the shader part of this is not yet implemented.
+        return GL_CLAMP_TO_BORDER;
     }
-    LOG_CRITICAL(Render_OpenGL, "Unimplemented texture wrap mode=%u", static_cast<u32>(wrap_mode));
+    NGLOG_CRITICAL(Render_OpenGL, "Unimplemented texture wrap mode={}",
+                   static_cast<u32>(wrap_mode));
+    UNREACHABLE();
+    return {};
+}
+
+inline GLenum BlendEquation(Maxwell::Blend::Equation equation) {
+    switch (equation) {
+    case Maxwell::Blend::Equation::Add:
+        return GL_FUNC_ADD;
+    case Maxwell::Blend::Equation::Subtract:
+        return GL_FUNC_SUBTRACT;
+    case Maxwell::Blend::Equation::ReverseSubtract:
+        return GL_FUNC_REVERSE_SUBTRACT;
+    case Maxwell::Blend::Equation::Min:
+        return GL_MIN;
+    case Maxwell::Blend::Equation::Max:
+        return GL_MAX;
+    }
+    NGLOG_CRITICAL(Render_OpenGL, "Unimplemented blend equation={}", static_cast<u32>(equation));
+    UNREACHABLE();
+    return {};
+}
+
+inline GLenum BlendFunc(Maxwell::Blend::Factor factor) {
+    switch (factor) {
+    case Maxwell::Blend::Factor::Zero:
+        return GL_ZERO;
+    case Maxwell::Blend::Factor::One:
+        return GL_ONE;
+    case Maxwell::Blend::Factor::SourceColor:
+        return GL_SRC_COLOR;
+    case Maxwell::Blend::Factor::OneMinusSourceColor:
+        return GL_ONE_MINUS_SRC_COLOR;
+    case Maxwell::Blend::Factor::SourceAlpha:
+        return GL_SRC_ALPHA;
+    case Maxwell::Blend::Factor::OneMinusSourceAlpha:
+        return GL_ONE_MINUS_SRC_ALPHA;
+    case Maxwell::Blend::Factor::DestAlpha:
+        return GL_DST_ALPHA;
+    case Maxwell::Blend::Factor::OneMinusDestAlpha:
+        return GL_ONE_MINUS_DST_ALPHA;
+    case Maxwell::Blend::Factor::DestColor:
+        return GL_DST_COLOR;
+    case Maxwell::Blend::Factor::OneMinusDestColor:
+        return GL_ONE_MINUS_DST_COLOR;
+    case Maxwell::Blend::Factor::SourceAlphaSaturate:
+        return GL_SRC_ALPHA_SATURATE;
+    case Maxwell::Blend::Factor::Source1Color:
+        return GL_SRC1_COLOR;
+    case Maxwell::Blend::Factor::OneMinusSource1Color:
+        return GL_ONE_MINUS_SRC1_COLOR;
+    case Maxwell::Blend::Factor::Source1Alpha:
+        return GL_SRC1_ALPHA;
+    case Maxwell::Blend::Factor::OneMinusSource1Alpha:
+        return GL_ONE_MINUS_SRC1_ALPHA;
+    case Maxwell::Blend::Factor::ConstantColor:
+        return GL_CONSTANT_COLOR;
+    case Maxwell::Blend::Factor::OneMinusConstantColor:
+        return GL_ONE_MINUS_CONSTANT_COLOR;
+    case Maxwell::Blend::Factor::ConstantAlpha:
+        return GL_CONSTANT_ALPHA;
+    case Maxwell::Blend::Factor::OneMinusConstantAlpha:
+        return GL_ONE_MINUS_CONSTANT_ALPHA;
+    }
+    NGLOG_CRITICAL(Render_OpenGL, "Unimplemented blend factor={}", static_cast<u32>(factor));
     UNREACHABLE();
     return {};
 }

src/video_core/textures/decoders.cpp

@@ -48,31 +48,39 @@ u32 BytesPerPixel(TextureFormat format) {
     case TextureFormat::DXT1:
         // In this case a 'pixel' actually refers to a 4x4 tile.
         return 8;
+    case TextureFormat::DXT23:
+    case TextureFormat::DXT45:
+        // In this case a 'pixel' actually refers to a 4x4 tile.
+        return 16;
     case TextureFormat::A8R8G8B8:
         return 4;
+    case TextureFormat::B5G6R5:
+        return 2;
     default:
         UNIMPLEMENTED_MSG("Format not implemented");
         break;
     }
 }
 
-std::vector<u8> UnswizzleTexture(VAddr address, TextureFormat format, u32 width, u32 height) {
+std::vector<u8> UnswizzleTexture(VAddr address, TextureFormat format, u32 width, u32 height,
+                                 u32 block_height) {
     u8* data = Memory::GetPointer(address);
     u32 bytes_per_pixel = BytesPerPixel(format);
 
-    static constexpr u32 DefaultBlockHeight = 16;
-
     std::vector<u8> unswizzled_data(width * height * bytes_per_pixel);
 
     switch (format) {
     case TextureFormat::DXT1:
-        // In the DXT1 format, each 4x4 tile is swizzled instead of just individual pixel values.
+    case TextureFormat::DXT23:
+    case TextureFormat::DXT45:
+        // In the DXT formats, each 4x4 tile is swizzled instead of just individual pixel values.
         CopySwizzledData(width / 4, height / 4, bytes_per_pixel, bytes_per_pixel, data,
-                         unswizzled_data.data(), true, DefaultBlockHeight);
+                         unswizzled_data.data(), true, block_height);
         break;
     case TextureFormat::A8R8G8B8:
+    case TextureFormat::B5G6R5:
         CopySwizzledData(width, height, bytes_per_pixel, bytes_per_pixel, data,
-                         unswizzled_data.data(), true, DefaultBlockHeight);
+                         unswizzled_data.data(), true, block_height);
         break;
     default:
         UNIMPLEMENTED_MSG("Format not implemented");
@@ -89,7 +97,10 @@ std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat
     // TODO(Subv): Implement.
     switch (format) {
     case TextureFormat::DXT1:
+    case TextureFormat::DXT23:
+    case TextureFormat::DXT45:
     case TextureFormat::A8R8G8B8:
+    case TextureFormat::B5G6R5:
         // TODO(Subv): For the time being just forward the same data without any decoding.
         rgba_data = texture_data;
         break;

src/video_core/textures/decoders.h

@@ -14,7 +14,8 @@ namespace Texture {
 /**
  * Unswizzles a swizzled texture without changing its format.
  */
-std::vector<u8> UnswizzleTexture(VAddr address, TextureFormat format, u32 width, u32 height);
+std::vector<u8> UnswizzleTexture(VAddr address, TextureFormat format, u32 width, u32 height,
+                                 u32 block_height = TICEntry::DefaultBlockHeight);
 
 /**
  * Decodes an unswizzled texture into a A8R8G8B8 texture.

src/video_core/textures/texture.h

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include "common/assert.h"
 #include "common/bit_field.h"
 #include "common/common_funcs.h"
 #include "common/common_types.h"
@@ -13,8 +14,11 @@ namespace Tegra {
 namespace Texture {
 
 enum class TextureFormat : u32 {
-    A8R8G8B8 = 8,
+    A8R8G8B8 = 0x8,
+    B5G6R5 = 0x15,
     DXT1 = 0x24,
+    DXT23 = 0x25,
+    DXT45 = 0x26,
 };
 
 enum class TextureType : u32 {
@@ -55,6 +59,8 @@ union TextureHandle {
 static_assert(sizeof(TextureHandle) == 4, "TextureHandle has wrong size");
 
 struct TICEntry {
+    static constexpr u32 DefaultBlockHeight = 16;
+
     union {
         u32 raw;
         BitField<0, 7, TextureFormat> format;
@@ -68,7 +74,12 @@ struct TICEntry {
         BitField<0, 16, u32> address_high;
         BitField<21, 3, TICHeaderVersion> header_version;
     };
-    INSERT_PADDING_BYTES(4);
+    union {
+        BitField<3, 3, u32> block_height;
+
+        // High 16 bits of the pitch value
+        BitField<0, 16, u32> pitch_high;
+    };
     union {
         BitField<0, 16, u32> width_minus_1;
         BitField<23, 4, TextureType> texture_type;
@@ -80,6 +91,13 @@ struct TICEntry {
         return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | address_low);
     }
 
+    u32 Pitch() const {
+        ASSERT(header_version == TICHeaderVersion::Pitch ||
+               header_version == TICHeaderVersion::PitchColorKey);
+        // The pitch value is 21 bits, and is 32B aligned.
+        return pitch_high << 5;
+    }
+
     u32 Width() const {
         return width_minus_1 + 1;
     }
@@ -88,6 +106,13 @@ struct TICEntry {
         return height_minus_1 + 1;
     }
 
+    u32 BlockHeight() const {
+        ASSERT(header_version == TICHeaderVersion::BlockLinear ||
+               header_version == TICHeaderVersion::BlockLinearColorKey);
+        // The block height is stored in log2 format.
+        return 1 << block_height;
+    }
+
     bool IsTiled() const {
         return header_version == TICHeaderVersion::BlockLinear ||
                header_version == TICHeaderVersion::BlockLinearColorKey;

src/yuzu/main.cpp

@@ -44,6 +44,15 @@
 Q_IMPORT_PLUGIN(QWindowsIntegrationPlugin);
 #endif
 
+#ifdef _WIN32
+extern "C" {
+// tells Nvidia and AMD drivers to use the dedicated GPU by default on laptops with switchable
+// graphics
+__declspec(dllexport) unsigned long NvOptimusEnablement = 0x00000001;
+__declspec(dllexport) int AmdPowerXpressRequestHighPerformance = 1;
+}
+#endif
+
 /**
  * "Callouts" are one-time instructional messages shown to the user. In the config settings, there
  * is a bitfield "callout_flags" options, used to track if a message has already been shown to the

src/yuzu_cmd/yuzu.cpp

@@ -37,6 +37,15 @@
 #include "yuzu_cmd/config.h"
 #include "yuzu_cmd/emu_window/emu_window_sdl2.h"
 
+#ifdef _WIN32
+extern "C" {
+// tells Nvidia and AMD drivers to use the dedicated GPU by default on laptops with switchable
+// graphics
+__declspec(dllexport) unsigned long NvOptimusEnablement = 0x00000001;
+__declspec(dllexport) int AmdPowerXpressRequestHighPerformance = 1;
+}
+#endif
+
 static void PrintHelp(const char* argv0) {
     std::cout << "Usage: " << argv0
               << " [options] <filename>\n"