npad: Use NPadIdToIndex to prevent invalid array access

Add support for viewport_transfom_enable register

.gitmodules

@@ -13,9 +13,6 @@
 [submodule "dynarmic"]
     path = externals/dynarmic
     url = https://github.com/MerryMage/dynarmic.git
-[submodule "xbyak"]
-    path = externals/xbyak
-    url = https://github.com/herumi/xbyak.git
 [submodule "fmt"]
     path = externals/fmt
     url = https://github.com/fmtlib/fmt.git

CMakeLists.txt

@@ -377,7 +377,7 @@ if (CLANG_FORMAT)
     set(CCOMMENT "Running clang format against all the .h and .cpp files in src/")
     if (WIN32)
         add_custom_target(clang-format
-            COMMAND powershell.exe -Command "${CLANG_FORMAT} -i @(Get-ChildItem -Recurse ${SRCS}/* -Include \'*.h\', \'*.cpp\')"
+            COMMAND powershell.exe -Command "Get-ChildItem ${SRCS}/* -Include *.cpp,*.h -Recurse | Foreach {${CLANG_FORMAT} -i $_.fullname}"
             COMMENT ${CCOMMENT})
     elseif(MINGW)
         add_custom_target(clang-format

externals/CMakeLists.txt

@@ -9,7 +9,6 @@ target_include_directories(catch-single-include INTERFACE catch/single_include)
 
 # Dynarmic
 if (ARCHITECTURE_x86_64)
-    add_library(xbyak INTERFACE)
     set(DYNARMIC_TESTS OFF)
     set(DYNARMIC_NO_BUNDLED_FMT ON)
     add_subdirectory(dynarmic)
@@ -53,14 +52,6 @@ target_include_directories(unicorn-headers INTERFACE ./unicorn/include)
 # SoundTouch
 add_subdirectory(soundtouch)
 
-# Xbyak
-if (ARCHITECTURE_x86_64)
-    # Defined before "dynarmic" above
-    # add_library(xbyak INTERFACE)
-    target_include_directories(xbyak INTERFACE ./xbyak/xbyak)
-    target_compile_definitions(xbyak INTERFACE XBYAK_NO_OP_NAMES)
-endif()
-
 # Opus
 add_subdirectory(opus)
 target_include_directories(opus INTERFACE ./opus/include)

src/common/CMakeLists.txt

@@ -44,7 +44,6 @@ add_library(common STATIC
     detached_tasks.cpp
     detached_tasks.h
     bit_field.h
-    bit_set.h
     cityhash.cpp
     cityhash.h
     color.h
@@ -95,14 +94,9 @@ if(ARCHITECTURE_x86_64)
         PRIVATE
             x64/cpu_detect.cpp
             x64/cpu_detect.h
-            x64/xbyak_abi.h
-            x64/xbyak_util.h
     )
 endif()
 
 create_target_directory_groups(common)
 
 target_link_libraries(common PUBLIC Boost::boost fmt microprofile)
-if (ARCHITECTURE_x86_64)
-    target_link_libraries(common PRIVATE xbyak)
-endif()

src/common/assert.h

@@ -52,5 +52,8 @@ __declspec(noinline, noreturn)
 #define DEBUG_ASSERT_MSG(_a_, _desc_, ...)
 #endif
 
-#define UNIMPLEMENTED() LOG_CRITICAL(Debug, "Unimplemented code!")
+#define UNIMPLEMENTED() ASSERT_MSG(false, "Unimplemented code!")
 #define UNIMPLEMENTED_MSG(...) ASSERT_MSG(false, __VA_ARGS__)
+
+#define UNIMPLEMENTED_IF(cond) ASSERT_MSG(!(cond), "Unimplemented code!")
+#define UNIMPLEMENTED_IF_MSG(cond, ...) ASSERT_MSG(!(cond), __VA_ARGS__)

src/common/bit_set.h

@@ -1,244 +0,0 @@
-// This file is under the public domain.
-
-#pragma once
-
-#include <cstddef>
-#ifdef _WIN32
-#include <intrin.h>
-#endif
-#include <initializer_list>
-#include <new>
-#include <type_traits>
-#include "common/common_types.h"
-
-// namespace avoids conflict with OS X Carbon; don't use BitSet<T> directly
-namespace Common {
-
-// Helper functions:
-
-#ifdef _MSC_VER
-template <typename T>
-static inline int CountSetBits(T v) {
-    // from https://graphics.stanford.edu/~seander/bithacks.html
-    // GCC has this built in, but MSVC's intrinsic will only emit the actual
-    // POPCNT instruction, which we're not depending on
-    v = v - ((v >> 1) & (T) ~(T)0 / 3);
-    v = (v & (T) ~(T)0 / 15 * 3) + ((v >> 2) & (T) ~(T)0 / 15 * 3);
-    v = (v + (v >> 4)) & (T) ~(T)0 / 255 * 15;
-    return (T)(v * ((T) ~(T)0 / 255)) >> (sizeof(T) - 1) * 8;
-}
-static inline int LeastSignificantSetBit(u8 val) {
-    unsigned long index;
-    _BitScanForward(&index, val);
-    return (int)index;
-}
-static inline int LeastSignificantSetBit(u16 val) {
-    unsigned long index;
-    _BitScanForward(&index, val);
-    return (int)index;
-}
-static inline int LeastSignificantSetBit(u32 val) {
-    unsigned long index;
-    _BitScanForward(&index, val);
-    return (int)index;
-}
-static inline int LeastSignificantSetBit(u64 val) {
-    unsigned long index;
-    _BitScanForward64(&index, val);
-    return (int)index;
-}
-#else
-static inline int CountSetBits(u8 val) {
-    return __builtin_popcount(val);
-}
-static inline int CountSetBits(u16 val) {
-    return __builtin_popcount(val);
-}
-static inline int CountSetBits(u32 val) {
-    return __builtin_popcount(val);
-}
-static inline int CountSetBits(u64 val) {
-    return __builtin_popcountll(val);
-}
-static inline int LeastSignificantSetBit(u8 val) {
-    return __builtin_ctz(val);
-}
-static inline int LeastSignificantSetBit(u16 val) {
-    return __builtin_ctz(val);
-}
-static inline int LeastSignificantSetBit(u32 val) {
-    return __builtin_ctz(val);
-}
-static inline int LeastSignificantSetBit(u64 val) {
-    return __builtin_ctzll(val);
-}
-#endif
-
-// Similar to std::bitset, this is a class which encapsulates a bitset, i.e.
-// using the set bits of an integer to represent a set of integers.  Like that
-// class, it acts like an array of bools:
-//     BitSet32 bs;
-//     bs[1] = true;
-// but also like the underlying integer ([0] = least significant bit):
-//     BitSet32 bs2 = ...;
-//     bs = (bs ^ bs2) & BitSet32(0xffff);
-// The following additional functionality is provided:
-// - Construction using an initializer list.
-//     BitSet bs { 1, 2, 4, 8 };
-// - Efficiently iterating through the set bits:
-//     for (int i : bs)
-//         [i is the *index* of a set bit]
-//   (This uses the appropriate CPU instruction to find the next set bit in one
-//   operation.)
-// - Counting set bits using .Count() - see comment on that method.
-
-// TODO: use constexpr when MSVC gets out of the Dark Ages
-
-template <typename IntTy>
-class BitSet {
-    static_assert(!std::is_signed_v<IntTy>, "BitSet should not be used with signed types");
-
-public:
-    // A reference to a particular bit, returned from operator[].
-    class Ref {
-    public:
-        Ref(Ref&& other) : m_bs(other.m_bs), m_mask(other.m_mask) {}
-        Ref(BitSet* bs, IntTy mask) : m_bs(bs), m_mask(mask) {}
-        operator bool() const {
-            return (m_bs->m_val & m_mask) != 0;
-        }
-        bool operator=(bool set) {
-            m_bs->m_val = (m_bs->m_val & ~m_mask) | (set ? m_mask : 0);
-            return set;
-        }
-
-    private:
-        BitSet* m_bs;
-        IntTy m_mask;
-    };
-
-    // A STL-like iterator is required to be able to use range-based for loops.
-    class Iterator {
-    public:
-        Iterator(const Iterator& other) : m_val(other.m_val), m_bit(other.m_bit) {}
-        Iterator(IntTy val) : m_val(val), m_bit(0) {}
-        Iterator& operator=(Iterator other) {
-            new (this) Iterator(other);
-            return *this;
-        }
-        int operator*() {
-            return m_bit + ComputeLsb();
-        }
-        Iterator& operator++() {
-            int lsb = ComputeLsb();
-            m_val >>= lsb + 1;
-            m_bit += lsb + 1;
-            m_has_lsb = false;
-            return *this;
-        }
-        Iterator operator++(int _) {
-            Iterator other(*this);
-            ++*this;
-            return other;
-        }
-        bool operator==(Iterator other) const {
-            return m_val == other.m_val;
-        }
-        bool operator!=(Iterator other) const {
-            return m_val != other.m_val;
-        }
-
-    private:
-        int ComputeLsb() {
-            if (!m_has_lsb) {
-                m_lsb = LeastSignificantSetBit(m_val);
-                m_has_lsb = true;
-            }
-            return m_lsb;
-        }
-        IntTy m_val;
-        int m_bit;
-        int m_lsb = -1;
-        bool m_has_lsb = false;
-    };
-
-    BitSet() : m_val(0) {}
-    explicit BitSet(IntTy val) : m_val(val) {}
-    BitSet(std::initializer_list<int> init) {
-        m_val = 0;
-        for (int bit : init)
-            m_val |= (IntTy)1 << bit;
-    }
-
-    static BitSet AllTrue(std::size_t count) {
-        return BitSet(count == sizeof(IntTy) * 8 ? ~(IntTy)0 : (((IntTy)1 << count) - 1));
-    }
-
-    Ref operator[](std::size_t bit) {
-        return Ref(this, (IntTy)1 << bit);
-    }
-    const Ref operator[](std::size_t bit) const {
-        return (*const_cast<BitSet*>(this))[bit];
-    }
-    bool operator==(BitSet other) const {
-        return m_val == other.m_val;
-    }
-    bool operator!=(BitSet other) const {
-        return m_val != other.m_val;
-    }
-    bool operator<(BitSet other) const {
-        return m_val < other.m_val;
-    }
-    bool operator>(BitSet other) const {
-        return m_val > other.m_val;
-    }
-    BitSet operator|(BitSet other) const {
-        return BitSet(m_val | other.m_val);
-    }
-    BitSet operator&(BitSet other) const {
-        return BitSet(m_val & other.m_val);
-    }
-    BitSet operator^(BitSet other) const {
-        return BitSet(m_val ^ other.m_val);
-    }
-    BitSet operator~() const {
-        return BitSet(~m_val);
-    }
-    BitSet& operator|=(BitSet other) {
-        return *this = *this | other;
-    }
-    BitSet& operator&=(BitSet other) {
-        return *this = *this & other;
-    }
-    BitSet& operator^=(BitSet other) {
-        return *this = *this ^ other;
-    }
-    operator u32() = delete;
-    operator bool() {
-        return m_val != 0;
-    }
-
-    // Warning: Even though on modern CPUs this is a single fast instruction,
-    // Dolphin's official builds do not currently assume POPCNT support on x86,
-    // so slower explicit bit twiddling is generated.  Still should generally
-    // be faster than a loop.
-    unsigned int Count() const {
-        return CountSetBits(m_val);
-    }
-
-    Iterator begin() const {
-        return Iterator(m_val);
-    }
-    Iterator end() const {
-        return Iterator(0);
-    }
-
-    IntTy m_val;
-};
-
-} // namespace Common
-
-typedef Common::BitSet<u8> BitSet8;
-typedef Common::BitSet<u16> BitSet16;
-typedef Common::BitSet<u32> BitSet32;
-typedef Common::BitSet<u64> BitSet64;

src/common/math_util.h

@@ -4,18 +4,12 @@
 
 #pragma once
 
-#include <algorithm>
 #include <cstdlib>
 #include <type_traits>
 
 namespace MathUtil {
 
-static constexpr float PI = 3.14159265f;
-
-inline bool IntervalsIntersect(unsigned start0, unsigned length0, unsigned start1,
-                               unsigned length1) {
-    return (std::max(start0, start1) < std::min(start0 + length0, start1 + length1));
-}
+constexpr float PI = 3.14159265f;
 
 template <class T>
 struct Rectangle {
@@ -24,16 +18,16 @@ struct Rectangle {
     T right{};
     T bottom{};
 
-    Rectangle() = default;
+    constexpr Rectangle() = default;
 
-    Rectangle(T left, T top, T right, T bottom)
+    constexpr Rectangle(T left, T top, T right, T bottom)
         : left(left), top(top), right(right), bottom(bottom) {}
 
     T GetWidth() const {
-        return std::abs(static_cast<typename std::make_signed<T>::type>(right - left));
+        return std::abs(static_cast<std::make_signed_t<T>>(right - left));
     }
     T GetHeight() const {
-        return std::abs(static_cast<typename std::make_signed<T>::type>(bottom - top));
+        return std::abs(static_cast<std::make_signed_t<T>>(bottom - top));
     }
     Rectangle<T> TranslateX(const T x) const {
         return Rectangle{left + x, top, right + x, bottom};

src/common/thread.cpp

@@ -25,23 +25,6 @@
 
 namespace Common {
 
-int CurrentThreadId() {
-#ifdef _MSC_VER
-    return GetCurrentThreadId();
-#elif defined __APPLE__
-    return mach_thread_self();
-#else
-    return 0;
-#endif
-}
-
-#ifdef _WIN32
-// Supporting functions
-void SleepCurrentThread(int ms) {
-    Sleep(ms);
-}
-#endif
-
 #ifdef _MSC_VER
 
 void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask) {
@@ -62,7 +45,7 @@ void SwitchCurrentThread() {
 
 // This is implemented much nicer in upcoming msvc++, see:
 // http://msdn.microsoft.com/en-us/library/xcb2z8hs(VS.100).aspx
-void SetCurrentThreadName(const char* szThreadName) {
+void SetCurrentThreadName(const char* name) {
     static const DWORD MS_VC_EXCEPTION = 0x406D1388;
 
 #pragma pack(push, 8)
@@ -75,7 +58,7 @@ void SetCurrentThreadName(const char* szThreadName) {
 #pragma pack(pop)
 
     info.dwType = 0x1000;
-    info.szName = szThreadName;
+    info.szName = name;
     info.dwThreadID = -1; // dwThreadID;
     info.dwFlags = 0;
 
@@ -107,10 +90,6 @@ void SetCurrentThreadAffinity(u32 mask) {
 }
 
 #ifndef _WIN32
-void SleepCurrentThread(int ms) {
-    usleep(1000 * ms);
-}
-
 void SwitchCurrentThread() {
     usleep(1000 * 1);
 }
@@ -118,15 +97,15 @@ void SwitchCurrentThread() {
 
 // MinGW with the POSIX threading model does not support pthread_setname_np
 #if !defined(_WIN32) || defined(_MSC_VER)
-void SetCurrentThreadName(const char* szThreadName) {
+void SetCurrentThreadName(const char* name) {
 #ifdef __APPLE__
-    pthread_setname_np(szThreadName);
+    pthread_setname_np(name);
 #elif defined(__Bitrig__) || defined(__DragonFly__) || defined(__FreeBSD__) || defined(__OpenBSD__)
-    pthread_set_name_np(pthread_self(), szThreadName);
+    pthread_set_name_np(pthread_self(), name);
 #elif defined(__NetBSD__)
-    pthread_setname_np(pthread_self(), "%s", (void*)szThreadName);
+    pthread_setname_np(pthread_self(), "%s", (void*)name);
 #else
-    pthread_setname_np(pthread_self(), szThreadName);
+    pthread_setname_np(pthread_self(), name);
 #endif
 }
 #endif

src/common/thread.h

@@ -13,15 +13,8 @@
 
 namespace Common {
 
-int CurrentThreadId();
-
-void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask);
-void SetCurrentThreadAffinity(u32 mask);
-
 class Event {
 public:
-    Event() : is_set(false) {}
-
     void Set() {
         std::lock_guard<std::mutex> lk(mutex);
         if (!is_set) {
@@ -53,14 +46,14 @@ public:
     }
 
 private:
-    bool is_set;
+    bool is_set = false;
     std::condition_variable condvar;
     std::mutex mutex;
 };
 
 class Barrier {
 public:
-    explicit Barrier(std::size_t count_) : count(count_), waiting(0), generation(0) {}
+    explicit Barrier(std::size_t count_) : count(count_) {}
 
     /// Blocks until all "count" threads have called Sync()
     void Sync() {
@@ -80,12 +73,13 @@ public:
 private:
     std::condition_variable condvar;
     std::mutex mutex;
-    const std::size_t count;
-    std::size_t waiting;
-    std::size_t generation; // Incremented once each time the barrier is used
+    std::size_t count;
+    std::size_t waiting = 0;
+    std::size_t generation = 0; // Incremented once each time the barrier is used
 };
 
-void SleepCurrentThread(int ms);
+void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask);
+void SetCurrentThreadAffinity(u32 mask);
 void SwitchCurrentThread(); // On Linux, this is equal to sleep 1ms
 void SetCurrentThreadName(const char* name);
 

src/common/x64/xbyak_abi.h

@@ -1,222 +0,0 @@
-// Copyright 2016 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <initializer_list>
-#include <xbyak.h>
-#include "common/assert.h"
-#include "common/bit_set.h"
-
-namespace Common::X64 {
-
-inline int RegToIndex(const Xbyak::Reg& reg) {
-    using Kind = Xbyak::Reg::Kind;
-    ASSERT_MSG((reg.getKind() & (Kind::REG | Kind::XMM)) != 0,
-               "RegSet only support GPRs and XMM registers.");
-    ASSERT_MSG(reg.getIdx() < 16, "RegSet only supports XXM0-15.");
-    return reg.getIdx() + (reg.getKind() == Kind::REG ? 0 : 16);
-}
-
-inline Xbyak::Reg64 IndexToReg64(int reg_index) {
-    ASSERT(reg_index < 16);
-    return Xbyak::Reg64(reg_index);
-}
-
-inline Xbyak::Xmm IndexToXmm(int reg_index) {
-    ASSERT(reg_index >= 16 && reg_index < 32);
-    return Xbyak::Xmm(reg_index - 16);
-}
-
-inline Xbyak::Reg IndexToReg(int reg_index) {
-    if (reg_index < 16) {
-        return IndexToReg64(reg_index);
-    } else {
-        return IndexToXmm(reg_index);
-    }
-}
-
-inline BitSet32 BuildRegSet(std::initializer_list<Xbyak::Reg> regs) {
-    BitSet32 bits;
-    for (const Xbyak::Reg& reg : regs) {
-        bits[RegToIndex(reg)] = true;
-    }
-    return bits;
-}
-
-const BitSet32 ABI_ALL_GPRS(0x0000FFFF);
-const BitSet32 ABI_ALL_XMMS(0xFFFF0000);
-
-#ifdef _WIN32
-
-// Microsoft x64 ABI
-const Xbyak::Reg ABI_RETURN = Xbyak::util::rax;
-const Xbyak::Reg ABI_PARAM1 = Xbyak::util::rcx;
-const Xbyak::Reg ABI_PARAM2 = Xbyak::util::rdx;
-const Xbyak::Reg ABI_PARAM3 = Xbyak::util::r8;
-const Xbyak::Reg ABI_PARAM4 = Xbyak::util::r9;
-
-const BitSet32 ABI_ALL_CALLER_SAVED = BuildRegSet({
-    // GPRs
-    Xbyak::util::rcx,
-    Xbyak::util::rdx,
-    Xbyak::util::r8,
-    Xbyak::util::r9,
-    Xbyak::util::r10,
-    Xbyak::util::r11,
-    // XMMs
-    Xbyak::util::xmm0,
-    Xbyak::util::xmm1,
-    Xbyak::util::xmm2,
-    Xbyak::util::xmm3,
-    Xbyak::util::xmm4,
-    Xbyak::util::xmm5,
-});
-
-const BitSet32 ABI_ALL_CALLEE_SAVED = BuildRegSet({
-    // GPRs
-    Xbyak::util::rbx,
-    Xbyak::util::rsi,
-    Xbyak::util::rdi,
-    Xbyak::util::rbp,
-    Xbyak::util::r12,
-    Xbyak::util::r13,
-    Xbyak::util::r14,
-    Xbyak::util::r15,
-    // XMMs
-    Xbyak::util::xmm6,
-    Xbyak::util::xmm7,
-    Xbyak::util::xmm8,
-    Xbyak::util::xmm9,
-    Xbyak::util::xmm10,
-    Xbyak::util::xmm11,
-    Xbyak::util::xmm12,
-    Xbyak::util::xmm13,
-    Xbyak::util::xmm14,
-    Xbyak::util::xmm15,
-});
-
-constexpr std::size_t ABI_SHADOW_SPACE = 0x20;
-
-#else
-
-// System V x86-64 ABI
-const Xbyak::Reg ABI_RETURN = Xbyak::util::rax;
-const Xbyak::Reg ABI_PARAM1 = Xbyak::util::rdi;
-const Xbyak::Reg ABI_PARAM2 = Xbyak::util::rsi;
-const Xbyak::Reg ABI_PARAM3 = Xbyak::util::rdx;
-const Xbyak::Reg ABI_PARAM4 = Xbyak::util::rcx;
-
-const BitSet32 ABI_ALL_CALLER_SAVED = BuildRegSet({
-    // GPRs
-    Xbyak::util::rcx,
-    Xbyak::util::rdx,
-    Xbyak::util::rdi,
-    Xbyak::util::rsi,
-    Xbyak::util::r8,
-    Xbyak::util::r9,
-    Xbyak::util::r10,
-    Xbyak::util::r11,
-    // XMMs
-    Xbyak::util::xmm0,
-    Xbyak::util::xmm1,
-    Xbyak::util::xmm2,
-    Xbyak::util::xmm3,
-    Xbyak::util::xmm4,
-    Xbyak::util::xmm5,
-    Xbyak::util::xmm6,
-    Xbyak::util::xmm7,
-    Xbyak::util::xmm8,
-    Xbyak::util::xmm9,
-    Xbyak::util::xmm10,
-    Xbyak::util::xmm11,
-    Xbyak::util::xmm12,
-    Xbyak::util::xmm13,
-    Xbyak::util::xmm14,
-    Xbyak::util::xmm15,
-});
-
-const BitSet32 ABI_ALL_CALLEE_SAVED = BuildRegSet({
-    // GPRs
-    Xbyak::util::rbx,
-    Xbyak::util::rbp,
-    Xbyak::util::r12,
-    Xbyak::util::r13,
-    Xbyak::util::r14,
-    Xbyak::util::r15,
-});
-
-constexpr std::size_t ABI_SHADOW_SPACE = 0;
-
-#endif
-
-inline void ABI_CalculateFrameSize(BitSet32 regs, std::size_t rsp_alignment,
-                                   std::size_t needed_frame_size, s32* out_subtraction,
-                                   s32* out_xmm_offset) {
-    int count = (regs & ABI_ALL_GPRS).Count();
-    rsp_alignment -= count * 8;
-    std::size_t subtraction = 0;
-    int xmm_count = (regs & ABI_ALL_XMMS).Count();
-    if (xmm_count) {
-        // If we have any XMMs to save, we must align the stack here.
-        subtraction = rsp_alignment & 0xF;
-    }
-    subtraction += 0x10 * xmm_count;
-    std::size_t xmm_base_subtraction = subtraction;
-    subtraction += needed_frame_size;
-    subtraction += ABI_SHADOW_SPACE;
-    // Final alignment.
-    rsp_alignment -= subtraction;
-    subtraction += rsp_alignment & 0xF;
-
-    *out_subtraction = (s32)subtraction;
-    *out_xmm_offset = (s32)(subtraction - xmm_base_subtraction);
-}
-
-inline std::size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs,
-                                                   std::size_t rsp_alignment,
-                                                   std::size_t needed_frame_size = 0) {
-    s32 subtraction, xmm_offset;
-    ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size, &subtraction, &xmm_offset);
-
-    for (int reg_index : (regs & ABI_ALL_GPRS)) {
-        code.push(IndexToReg64(reg_index));
-    }
-
-    if (subtraction != 0) {
-        code.sub(code.rsp, subtraction);
-    }
-
-    for (int reg_index : (regs & ABI_ALL_XMMS)) {
-        code.movaps(code.xword[code.rsp + xmm_offset], IndexToXmm(reg_index));
-        xmm_offset += 0x10;
-    }
-
-    return ABI_SHADOW_SPACE;
-}
-
-inline void ABI_PopRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs,
-                                           std::size_t rsp_alignment,
-                                           std::size_t needed_frame_size = 0) {
-    s32 subtraction, xmm_offset;
-    ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size, &subtraction, &xmm_offset);
-
-    for (int reg_index : (regs & ABI_ALL_XMMS)) {
-        code.movaps(IndexToXmm(reg_index), code.xword[code.rsp + xmm_offset]);
-        xmm_offset += 0x10;
-    }
-
-    if (subtraction != 0) {
-        code.add(code.rsp, subtraction);
-    }
-
-    // GPRs need to be popped in reverse order
-    for (int reg_index = 15; reg_index >= 0; reg_index--) {
-        if (regs[reg_index]) {
-            code.pop(IndexToReg64(reg_index));
-        }
-    }
-}
-
-} // namespace Common::X64

src/common/x64/xbyak_util.h

@@ -1,47 +0,0 @@
-// Copyright 2016 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <type_traits>
-#include <xbyak.h>
-#include "common/x64/xbyak_abi.h"
-
-namespace Common::X64 {
-
-// Constants for use with cmpps/cmpss
-enum {
-    CMP_EQ = 0,
-    CMP_LT = 1,
-    CMP_LE = 2,
-    CMP_UNORD = 3,
-    CMP_NEQ = 4,
-    CMP_NLT = 5,
-    CMP_NLE = 6,
-    CMP_ORD = 7,
-};
-
-inline bool IsWithin2G(uintptr_t ref, uintptr_t target) {
-    u64 distance = target - (ref + 5);
-    return !(distance >= 0x8000'0000ULL && distance <= ~0x8000'0000ULL);
-}
-
-inline bool IsWithin2G(const Xbyak::CodeGenerator& code, uintptr_t target) {
-    return IsWithin2G(reinterpret_cast<uintptr_t>(code.getCurr()), target);
-}
-
-template <typename T>
-inline void CallFarFunction(Xbyak::CodeGenerator& code, const T f) {
-    static_assert(std::is_pointer_v<T>, "Argument must be a (function) pointer.");
-    std::size_t addr = reinterpret_cast<std::size_t>(f);
-    if (IsWithin2G(code, addr)) {
-        code.call(f);
-    } else {
-        // ABI_RETURN is a safe temp register to use before a call
-        code.mov(ABI_RETURN, addr);
-        code.call(ABI_RETURN);
-    }
-}
-
-} // namespace Common::X64

src/core/CMakeLists.txt

@@ -12,6 +12,8 @@ add_library(core STATIC
     core_timing.h
     core_timing_util.cpp
     core_timing_util.h
+    cpu_core_manager.cpp
+    cpu_core_manager.h
     crypto/aes_util.cpp
     crypto/aes_util.h
     crypto/encryption_layer.cpp
@@ -156,6 +158,8 @@ add_library(core STATIC
     hle/service/am/applets/applets.h
     hle/service/am/applets/software_keyboard.cpp
     hle/service/am/applets/software_keyboard.h
+    hle/service/am/applets/stub_applet.cpp
+    hle/service/am/applets/stub_applet.h
     hle/service/am/idle.cpp
     hle/service/am/idle.h
     hle/service/am/omm.cpp

src/core/core.cpp

@@ -14,6 +14,7 @@
 #include "core/core.h"
 #include "core/core_cpu.h"
 #include "core/core_timing.h"
+#include "core/cpu_core_manager.h"
 #include "core/file_sys/mode.h"
 #include "core/file_sys/vfs_concat.h"
 #include "core/file_sys/vfs_real.h"
@@ -28,7 +29,6 @@
 #include "core/hle/service/sm/sm.h"
 #include "core/loader/loader.h"
 #include "core/perf_stats.h"
-#include "core/settings.h"
 #include "core/telemetry_session.h"
 #include "frontend/applets/software_keyboard.h"
 #include "video_core/debug_utils/debug_utils.h"
@@ -71,64 +71,22 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
 
     return vfs->OpenFile(path, FileSys::Mode::Read);
 }
-
-/// Runs a CPU core while the system is powered on
-void RunCpuCore(Cpu& cpu_state) {
-    while (Core::System::GetInstance().IsPoweredOn()) {
-        cpu_state.RunLoop(true);
-    }
-}
 } // Anonymous namespace
 
 struct System::Impl {
     Cpu& CurrentCpuCore() {
-        if (Settings::values.use_multi_core) {
-            const auto& search = thread_to_cpu.find(std::this_thread::get_id());
-            ASSERT(search != thread_to_cpu.end());
-            ASSERT(search->second);
-            return *search->second;
-        }
-
-        // Otherwise, use single-threaded mode active_core variable
-        return *cpu_cores[active_core];
+        return cpu_core_manager.GetCurrentCore();
     }
 
     ResultStatus RunLoop(bool tight_loop) {
         status = ResultStatus::Success;
 
-        // Update thread_to_cpu in case Core 0 is run from a different host thread
-        thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0].get();
-
-        if (GDBStub::IsServerEnabled()) {
-            GDBStub::HandlePacket();
-
-            // If the loop is halted and we want to step, use a tiny (1) number of instructions to
-            // execute. Otherwise, get out of the loop function.
-            if (GDBStub::GetCpuHaltFlag()) {
-                if (GDBStub::GetCpuStepFlag()) {
-                    tight_loop = false;
-                } else {
-                    return ResultStatus::Success;
-                }
-            }
-        }
-
-        for (active_core = 0; active_core < NUM_CPU_CORES; ++active_core) {
-            cpu_cores[active_core]->RunLoop(tight_loop);
-            if (Settings::values.use_multi_core) {
-                // Cores 1-3 are run on other threads in this mode
-                break;
-            }
-        }
-
-        if (GDBStub::IsServerEnabled()) {
-            GDBStub::SetCpuStepFlag(false);
-        }
+        cpu_core_manager.RunLoop(tight_loop);
 
         return status;
     }
 
-    ResultStatus Init(Frontend::EmuWindow& emu_window) {
+    ResultStatus Init(System& system, Frontend::EmuWindow& emu_window) {
         LOG_DEBUG(HW_Memory, "initialized OK");
 
         CoreTiming::Init();
@@ -145,12 +103,6 @@ struct System::Impl {
         auto main_process = Kernel::Process::Create(kernel, "main");
         kernel.MakeCurrentProcess(main_process.get());
 
-        cpu_barrier = std::make_unique<CpuBarrier>();
-        cpu_exclusive_monitor = Cpu::MakeExclusiveMonitor(cpu_cores.size());
-        for (std::size_t index = 0; index < cpu_cores.size(); ++index) {
-            cpu_cores[index] = std::make_unique<Cpu>(*cpu_exclusive_monitor, *cpu_barrier, index);
-        }
-
         telemetry_session = std::make_unique<Core::TelemetrySession>();
         service_manager = std::make_shared<Service::SM::ServiceManager>();
 
@@ -164,17 +116,8 @@ struct System::Impl {
 
         gpu_core = std::make_unique<Tegra::GPU>(renderer->Rasterizer());
 
-        // Create threads for CPU cores 1-3, and build thread_to_cpu map
-        // CPU core 0 is run on the main thread
-        thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0].get();
-        if (Settings::values.use_multi_core) {
-            for (std::size_t index = 0; index < cpu_core_threads.size(); ++index) {
-                cpu_core_threads[index] =
-                    std::make_unique<std::thread>(RunCpuCore, std::ref(*cpu_cores[index + 1]));
-                thread_to_cpu[cpu_core_threads[index]->get_id()] = cpu_cores[index + 1].get();
-            }
-        }
-
+        cpu_core_manager.Initialize(system);
+        is_powered_on = true;
         LOG_DEBUG(Core, "Initialized OK");
 
         // Reset counters and set time origin to current frame
@@ -184,7 +127,8 @@ struct System::Impl {
         return ResultStatus::Success;
     }
 
-    ResultStatus Load(Frontend::EmuWindow& emu_window, const std::string& filepath) {
+    ResultStatus Load(System& system, Frontend::EmuWindow& emu_window,
+                      const std::string& filepath) {
         app_loader = Loader::GetLoader(GetGameFileFromPath(virtual_filesystem, filepath));
 
         if (!app_loader) {
@@ -201,7 +145,7 @@ struct System::Impl {
             return ResultStatus::ErrorSystemMode;
         }
 
-        ResultStatus init_result{Init(emu_window)};
+        ResultStatus init_result{Init(system, emu_window)};
         if (init_result != ResultStatus::Success) {
             LOG_CRITICAL(Core, "Failed to initialize system (Error {})!",
                          static_cast<int>(init_result));
@@ -231,6 +175,8 @@ struct System::Impl {
         Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_Frametime",
                              perf_results.frametime * 1000.0);
 
+        is_powered_on = false;
+
         // Shutdown emulation session
         renderer.reset();
         GDBStub::Shutdown();
@@ -240,19 +186,7 @@ struct System::Impl {
         gpu_core.reset();
 
         // Close all CPU/threading state
-        cpu_barrier->NotifyEnd();
-        if (Settings::values.use_multi_core) {
-            for (auto& thread : cpu_core_threads) {
-                thread->join();
-                thread.reset();
-            }
-        }
-        thread_to_cpu.clear();
-        for (auto& cpu_core : cpu_cores) {
-            cpu_core.reset();
-        }
-        cpu_exclusive_monitor.reset();
-        cpu_barrier.reset();
+        cpu_core_manager.Shutdown();
 
         // Shutdown kernel and core timing
         kernel.Shutdown();
@@ -289,11 +223,8 @@ struct System::Impl {
     std::unique_ptr<VideoCore::RendererBase> renderer;
     std::unique_ptr<Tegra::GPU> gpu_core;
     std::shared_ptr<Tegra::DebugContext> debug_context;
-    std::unique_ptr<ExclusiveMonitor> cpu_exclusive_monitor;
-    std::unique_ptr<CpuBarrier> cpu_barrier;
-    std::array<std::unique_ptr<Cpu>, NUM_CPU_CORES> cpu_cores;
-    std::array<std::unique_ptr<std::thread>, NUM_CPU_CORES - 1> cpu_core_threads;
-    std::size_t active_core{}; ///< Active core, only used in single thread mode
+    CpuCoreManager cpu_core_manager;
+    bool is_powered_on = false;
 
     /// Frontend applets
     std::unique_ptr<Core::Frontend::SoftwareKeyboardApplet> software_keyboard;
@@ -307,9 +238,6 @@ struct System::Impl {
     ResultStatus status = ResultStatus::Success;
     std::string status_details = "";
 
-    /// Map of guest threads to CPU cores
-    std::map<std::thread::id, Cpu*> thread_to_cpu;
-
     Core::PerfStats perf_stats;
     Core::FrameLimiter frame_limiter;
 };
@@ -334,17 +262,15 @@ System::ResultStatus System::SingleStep() {
 }
 
 void System::InvalidateCpuInstructionCaches() {
-    for (auto& cpu : impl->cpu_cores) {
-        cpu->ArmInterface().ClearInstructionCache();
-    }
+    impl->cpu_core_manager.InvalidateAllInstructionCaches();
 }
 
 System::ResultStatus System::Load(Frontend::EmuWindow& emu_window, const std::string& filepath) {
-    return impl->Load(emu_window, filepath);
+    return impl->Load(*this, emu_window, filepath);
 }
 
 bool System::IsPoweredOn() const {
-    return impl->cpu_barrier && impl->cpu_barrier->IsAlive();
+    return impl->is_powered_on;
 }
 
 void System::PrepareReschedule() {
@@ -408,21 +334,20 @@ const ARM_Interface& System::ArmInterface(std::size_t core_index) const {
 }
 
 Cpu& System::CpuCore(std::size_t core_index) {
-    ASSERT(core_index < NUM_CPU_CORES);
-    return *impl->cpu_cores[core_index];
+    return impl->cpu_core_manager.GetCore(core_index);
 }
 
 const Cpu& System::CpuCore(std::size_t core_index) const {
     ASSERT(core_index < NUM_CPU_CORES);
-    return *impl->cpu_cores[core_index];
+    return impl->cpu_core_manager.GetCore(core_index);
 }
 
 ExclusiveMonitor& System::Monitor() {
-    return *impl->cpu_exclusive_monitor;
+    return impl->cpu_core_manager.GetExclusiveMonitor();
 }
 
 const ExclusiveMonitor& System::Monitor() const {
-    return *impl->cpu_exclusive_monitor;
+    return impl->cpu_core_manager.GetExclusiveMonitor();
 }
 
 Tegra::GPU& System::GPU() {
@@ -506,7 +431,7 @@ const Core::Frontend::SoftwareKeyboardApplet& System::GetSoftwareKeyboard() cons
 }
 
 System::ResultStatus System::Init(Frontend::EmuWindow& emu_window) {
-    return impl->Init(emu_window);
+    return impl->Init(*this, emu_window);
 }
 
 void System::Shutdown() {

src/core/cpu_core_manager.cpp

@@ -0,0 +1,142 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/assert.h"
+#include "core/arm/exclusive_monitor.h"
+#include "core/core.h"
+#include "core/core_cpu.h"
+#include "core/cpu_core_manager.h"
+#include "core/gdbstub/gdbstub.h"
+#include "core/settings.h"
+
+namespace Core {
+namespace {
+void RunCpuCore(const System& system, Cpu& cpu_state) {
+    while (system.IsPoweredOn()) {
+        cpu_state.RunLoop(true);
+    }
+}
+} // Anonymous namespace
+
+CpuCoreManager::CpuCoreManager() = default;
+CpuCoreManager::~CpuCoreManager() = default;
+
+void CpuCoreManager::Initialize(System& system) {
+    barrier = std::make_unique<CpuBarrier>();
+    exclusive_monitor = Cpu::MakeExclusiveMonitor(cores.size());
+
+    for (std::size_t index = 0; index < cores.size(); ++index) {
+        cores[index] = std::make_unique<Cpu>(*exclusive_monitor, *barrier, index);
+    }
+
+    // Create threads for CPU cores 1-3, and build thread_to_cpu map
+    // CPU core 0 is run on the main thread
+    thread_to_cpu[std::this_thread::get_id()] = cores[0].get();
+    if (!Settings::values.use_multi_core) {
+        return;
+    }
+
+    for (std::size_t index = 0; index < core_threads.size(); ++index) {
+        core_threads[index] = std::make_unique<std::thread>(RunCpuCore, std::cref(system),
+                                                            std::ref(*cores[index + 1]));
+        thread_to_cpu[core_threads[index]->get_id()] = cores[index + 1].get();
+    }
+}
+
+void CpuCoreManager::Shutdown() {
+    barrier->NotifyEnd();
+    if (Settings::values.use_multi_core) {
+        for (auto& thread : core_threads) {
+            thread->join();
+            thread.reset();
+        }
+    }
+
+    thread_to_cpu.clear();
+    for (auto& cpu_core : cores) {
+        cpu_core.reset();
+    }
+
+    exclusive_monitor.reset();
+    barrier.reset();
+}
+
+Cpu& CpuCoreManager::GetCore(std::size_t index) {
+    return *cores.at(index);
+}
+
+const Cpu& CpuCoreManager::GetCore(std::size_t index) const {
+    return *cores.at(index);
+}
+
+ExclusiveMonitor& CpuCoreManager::GetExclusiveMonitor() {
+    return *exclusive_monitor;
+}
+
+const ExclusiveMonitor& CpuCoreManager::GetExclusiveMonitor() const {
+    return *exclusive_monitor;
+}
+
+Cpu& CpuCoreManager::GetCurrentCore() {
+    if (Settings::values.use_multi_core) {
+        const auto& search = thread_to_cpu.find(std::this_thread::get_id());
+        ASSERT(search != thread_to_cpu.end());
+        ASSERT(search->second);
+        return *search->second;
+    }
+
+    // Otherwise, use single-threaded mode active_core variable
+    return *cores[active_core];
+}
+
+const Cpu& CpuCoreManager::GetCurrentCore() const {
+    if (Settings::values.use_multi_core) {
+        const auto& search = thread_to_cpu.find(std::this_thread::get_id());
+        ASSERT(search != thread_to_cpu.end());
+        ASSERT(search->second);
+        return *search->second;
+    }
+
+    // Otherwise, use single-threaded mode active_core variable
+    return *cores[active_core];
+}
+
+void CpuCoreManager::RunLoop(bool tight_loop) {
+    // Update thread_to_cpu in case Core 0 is run from a different host thread
+    thread_to_cpu[std::this_thread::get_id()] = cores[0].get();
+
+    if (GDBStub::IsServerEnabled()) {
+        GDBStub::HandlePacket();
+
+        // If the loop is halted and we want to step, use a tiny (1) number of instructions to
+        // execute. Otherwise, get out of the loop function.
+        if (GDBStub::GetCpuHaltFlag()) {
+            if (GDBStub::GetCpuStepFlag()) {
+                tight_loop = false;
+            } else {
+                return;
+            }
+        }
+    }
+
+    for (active_core = 0; active_core < NUM_CPU_CORES; ++active_core) {
+        cores[active_core]->RunLoop(tight_loop);
+        if (Settings::values.use_multi_core) {
+            // Cores 1-3 are run on other threads in this mode
+            break;
+        }
+    }
+
+    if (GDBStub::IsServerEnabled()) {
+        GDBStub::SetCpuStepFlag(false);
+    }
+}
+
+void CpuCoreManager::InvalidateAllInstructionCaches() {
+    for (auto& cpu : cores) {
+        cpu->ArmInterface().ClearInstructionCache();
+    }
+}
+
+} // namespace Core

src/core/cpu_core_manager.h

@@ -0,0 +1,59 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <map>
+#include <memory>
+#include <thread>
+
+namespace Core {
+
+class Cpu;
+class CpuBarrier;
+class ExclusiveMonitor;
+class System;
+
+class CpuCoreManager {
+public:
+    CpuCoreManager();
+    CpuCoreManager(const CpuCoreManager&) = delete;
+    CpuCoreManager(CpuCoreManager&&) = delete;
+
+    ~CpuCoreManager();
+
+    CpuCoreManager& operator=(const CpuCoreManager&) = delete;
+    CpuCoreManager& operator=(CpuCoreManager&&) = delete;
+
+    void Initialize(System& system);
+    void Shutdown();
+
+    Cpu& GetCore(std::size_t index);
+    const Cpu& GetCore(std::size_t index) const;
+
+    Cpu& GetCurrentCore();
+    const Cpu& GetCurrentCore() const;
+
+    ExclusiveMonitor& GetExclusiveMonitor();
+    const ExclusiveMonitor& GetExclusiveMonitor() const;
+
+    void RunLoop(bool tight_loop);
+
+    void InvalidateAllInstructionCaches();
+
+private:
+    static constexpr std::size_t NUM_CPU_CORES = 4;
+
+    std::unique_ptr<ExclusiveMonitor> exclusive_monitor;
+    std::unique_ptr<CpuBarrier> barrier;
+    std::array<std::unique_ptr<Cpu>, NUM_CPU_CORES> cores;
+    std::array<std::unique_ptr<std::thread>, NUM_CPU_CORES - 1> core_threads;
+    std::size_t active_core{}; ///< Active core, only used in single thread mode
+
+    /// Map of guest threads to CPU cores
+    std::map<std::thread::id, Cpu*> thread_to_cpu;
+};
+
+} // namespace Core

src/core/file_sys/card_image.h

@@ -32,7 +32,18 @@ enum class GamecardSize : u8 {
 };
 
 struct GamecardInfo {
-    std::array<u8, 0x70> data;
+    u64_le firmware_version;
+    u32_le access_control_flags;
+    u32_le read_wait_time1;
+    u32_le read_wait_time2;
+    u32_le write_wait_time1;
+    u32_le write_wait_time2;
+    u32_le firmware_mode;
+    u32_le cup_version;
+    std::array<u8, 4> reserved1;
+    u64_le update_partition_hash;
+    u64_le cup_id;
+    std::array<u8, 0x38> reserved2;
 };
 static_assert(sizeof(GamecardInfo) == 0x70, "GamecardInfo has incorrect size.");
 

src/core/file_sys/patch_manager.cpp

@@ -26,6 +26,11 @@ namespace FileSys {
 constexpr u64 SINGLE_BYTE_MODULUS = 0x100;
 constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000;
 
+constexpr std::array<const char*, 14> EXEFS_FILE_NAMES{
+    "main",    "main.npdm", "rtld",    "sdk",     "subsdk0", "subsdk1", "subsdk2",
+    "subsdk3", "subsdk4",   "subsdk5", "subsdk6", "subsdk7", "subsdk8", "subsdk9",
+};
+
 struct NSOBuildHeader {
     u32_le magic;
     INSERT_PADDING_BYTES(0x3C);
@@ -57,6 +62,15 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
     if (exefs == nullptr)
         return exefs;
 
+    if (Settings::values.dump_exefs) {
+        LOG_INFO(Loader, "Dumping ExeFS for title_id={:016X}", title_id);
+        const auto dump_dir = Service::FileSystem::GetModificationDumpRoot(title_id);
+        if (dump_dir != nullptr) {
+            const auto exefs_dir = GetOrCreateDirectoryRelative(dump_dir, "/exefs");
+            VfsRawCopyD(exefs, exefs_dir);
+        }
+    }
+
     const auto installed = Service::FileSystem::GetUnionContents();
 
     // Game Updates
@@ -70,6 +84,30 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
         exefs = update->GetExeFS();
     }
 
+    // LayeredExeFS
+    const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+    if (load_dir != nullptr && load_dir->GetSize() > 0) {
+        auto patch_dirs = load_dir->GetSubdirectories();
+        std::sort(
+            patch_dirs.begin(), patch_dirs.end(),
+            [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
+
+        std::vector<VirtualDir> layers;
+        layers.reserve(patch_dirs.size() + 1);
+        for (const auto& subdir : patch_dirs) {
+            auto exefs_dir = subdir->GetSubdirectory("exefs");
+            if (exefs_dir != nullptr)
+                layers.push_back(std::move(exefs_dir));
+        }
+        layers.push_back(exefs);
+
+        auto layered = LayeredVfsDirectory::MakeLayeredDirectory(std::move(layers));
+        if (layered != nullptr) {
+            LOG_INFO(Loader, "    ExeFS: LayeredExeFS patches applied successfully");
+            exefs = std::move(layered);
+        }
+    }
+
     return exefs;
 }
 
@@ -314,18 +352,25 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
             if (IsDirValidAndNonEmpty(exefs_dir)) {
                 bool ips = false;
                 bool ipswitch = false;
+                bool layeredfs = false;
 
                 for (const auto& file : exefs_dir->GetFiles()) {
-                    if (file->GetExtension() == "ips")
+                    if (file->GetExtension() == "ips") {
                         ips = true;
-                    else if (file->GetExtension() == "pchtxt")
+                    } else if (file->GetExtension() == "pchtxt") {
                         ipswitch = true;
+                    } else if (std::find(EXEFS_FILE_NAMES.begin(), EXEFS_FILE_NAMES.end(),
+                                         file->GetName()) != EXEFS_FILE_NAMES.end()) {
+                        layeredfs = true;
+                    }
                 }
 
                 if (ips)
                     AppendCommaIfNotEmpty(types, "IPS");
                 if (ipswitch)
                     AppendCommaIfNotEmpty(types, "IPSwitch");
+                if (layeredfs)
+                    AppendCommaIfNotEmpty(types, "LayeredExeFS");
             }
             if (IsDirValidAndNonEmpty(mod->GetSubdirectory("romfs")))
                 AppendCommaIfNotEmpty(types, "LayeredFS");

src/core/gdbstub/gdbstub.cpp

@@ -71,10 +71,6 @@ constexpr u32 PSTATE_REGISTER = 33;
 constexpr u32 UC_ARM64_REG_Q0 = 34;
 constexpr u32 FPCR_REGISTER = 66;
 
-// TODO/WiP - Used while working on support for FPU
-constexpr u32 TODO_DUMMY_REG_997 = 997;
-constexpr u32 TODO_DUMMY_REG_998 = 998;
-
 // For sample XML files see the GDB source /gdb/features
 // GDB also wants the l character at the start
 // This XML defines what the registers are for this specific ARM device
@@ -260,6 +256,36 @@ static void RegWrite(std::size_t id, u64 val, Kernel::Thread* thread = nullptr)
     }
 }
 
+static u128 FpuRead(std::size_t id, Kernel::Thread* thread = nullptr) {
+    if (!thread) {
+        return u128{0};
+    }
+
+    auto& thread_context = thread->GetContext();
+
+    if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) {
+        return thread_context.vector_registers[id - UC_ARM64_REG_Q0];
+    } else if (id == FPCR_REGISTER) {
+        return u128{thread_context.fpcr, 0};
+    } else {
+        return u128{0};
+    }
+}
+
+static void FpuWrite(std::size_t id, u128 val, Kernel::Thread* thread = nullptr) {
+    if (!thread) {
+        return;
+    }
+
+    auto& thread_context = thread->GetContext();
+
+    if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) {
+        thread_context.vector_registers[id - UC_ARM64_REG_Q0] = val;
+    } else if (id == FPCR_REGISTER) {
+        thread_context.fpcr = val[0];
+    }
+}
+
 /**
  * Turns hex string character into the equivalent byte.
  *
@@ -409,6 +435,27 @@ static u64 GdbHexToLong(const u8* src) {
     return output;
 }
 
+/**
+ * Convert a gdb-formatted hex string into a u128.
+ *
+ * @param src Pointer to hex string.
+ */
+static u128 GdbHexToU128(const u8* src) {
+    u128 output;
+
+    for (int i = 0; i < 16; i += 2) {
+        output[0] = (output[0] << 4) | HexCharToValue(src[15 - i - 1]);
+        output[0] = (output[0] << 4) | HexCharToValue(src[15 - i]);
+    }
+
+    for (int i = 0; i < 16; i += 2) {
+        output[1] = (output[1] << 4) | HexCharToValue(src[16 + 15 - i - 1]);
+        output[1] = (output[1] << 4) | HexCharToValue(src[16 + 15 - i]);
+    }
+
+    return output;
+}
+
 /// Read a byte from the gdb client.
 static u8 ReadByte() {
     u8 c;
@@ -599,8 +646,7 @@ static void HandleQuery() {
         for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
             const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
             for (const auto& thread : threads) {
-                val += fmt::format("{:x}", thread->GetThreadID());
-                val += ",";
+                val += fmt::format("{:x},", thread->GetThreadID());
             }
         }
         val.pop_back();
@@ -791,11 +837,15 @@ static void ReadRegister() {
     } else if (id == PSTATE_REGISTER) {
         IntToGdbHex(reply, static_cast<u32>(RegRead(id, current_thread)));
     } else if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) {
-        LongToGdbHex(reply, RegRead(id, current_thread));
+        u128 r = FpuRead(id, current_thread);
+        LongToGdbHex(reply, r[0]);
+        LongToGdbHex(reply + 16, r[1]);
     } else if (id == FPCR_REGISTER) {
-        LongToGdbHex(reply, RegRead(TODO_DUMMY_REG_998, current_thread));
-    } else {
-        LongToGdbHex(reply, RegRead(TODO_DUMMY_REG_997, current_thread));
+        u128 r = FpuRead(id, current_thread);
+        IntToGdbHex(reply, static_cast<u32>(r[0]));
+    } else if (id == FPCR_REGISTER + 1) {
+        u128 r = FpuRead(id, current_thread);
+        IntToGdbHex(reply, static_cast<u32>(r[0] >> 32));
     }
 
     SendReply(reinterpret_cast<char*>(reply));
@@ -822,13 +872,18 @@ static void ReadRegisters() {
 
     bufptr += 8;
 
-    for (u32 reg = UC_ARM64_REG_Q0; reg <= UC_ARM64_REG_Q0 + 31; reg++) {
-        LongToGdbHex(bufptr + reg * 16, RegRead(reg, current_thread));
+    u128 r;
+
+    for (u32 reg = UC_ARM64_REG_Q0; reg < FPCR_REGISTER; reg++) {
+        r = FpuRead(reg, current_thread);
+        LongToGdbHex(bufptr + reg * 32, r[0]);
+        LongToGdbHex(bufptr + reg * 32 + 16, r[1]);
     }
 
     bufptr += 32 * 32;
 
-    LongToGdbHex(bufptr, RegRead(TODO_DUMMY_REG_998, current_thread));
+    r = FpuRead(FPCR_REGISTER, current_thread);
+    IntToGdbHex(bufptr, static_cast<u32>(r[0]));
 
     bufptr += 8;
 
@@ -853,14 +908,12 @@ static void WriteRegister() {
     } else if (id == PSTATE_REGISTER) {
         RegWrite(id, GdbHexToInt(buffer_ptr), current_thread);
     } else if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) {
-        RegWrite(id, GdbHexToLong(buffer_ptr), current_thread);
+        FpuWrite(id, GdbHexToU128(buffer_ptr), current_thread);
     } else if (id == FPCR_REGISTER) {
-        RegWrite(TODO_DUMMY_REG_998, GdbHexToLong(buffer_ptr), current_thread);
-    } else {
-        RegWrite(TODO_DUMMY_REG_997, GdbHexToLong(buffer_ptr), current_thread);
+    } else if (id == FPCR_REGISTER + 1) {
     }
 
-    // Update Unicorn context skipping scheduler, no running threads at this point
+    // Update ARM context, skipping scheduler - no running threads at this point
     Core::System::GetInstance()
         .ArmInterface(current_core)
         .LoadContext(current_thread->GetContext());
@@ -885,13 +938,13 @@ static void WriteRegisters() {
         } else if (reg >= UC_ARM64_REG_Q0 && reg < FPCR_REGISTER) {
             RegWrite(reg, GdbHexToLong(buffer_ptr + i * 16), current_thread);
         } else if (reg == FPCR_REGISTER) {
-            RegWrite(TODO_DUMMY_REG_998, GdbHexToLong(buffer_ptr + i * 16), current_thread);
-        } else {
-            UNIMPLEMENTED();
+            RegWrite(FPCR_REGISTER, GdbHexToLong(buffer_ptr + i * 16), current_thread);
+        } else if (reg == FPCR_REGISTER + 1) {
+            RegWrite(FPCR_REGISTER, GdbHexToLong(buffer_ptr + i * 16), current_thread);
         }
     }
 
-    // Update Unicorn context skipping scheduler, no running threads at this point
+    // Update ARM context, skipping scheduler - no running threads at this point
     Core::System::GetInstance()
         .ArmInterface(current_core)
         .LoadContext(current_thread->GetContext());
@@ -917,12 +970,6 @@ static void ReadMemory() {
         SendReply("E01");
     }
 
-    const auto& vm_manager = Core::CurrentProcess()->VMManager();
-    if (addr < vm_manager.GetCodeRegionBaseAddress() ||
-        addr >= vm_manager.GetMapRegionEndAddress()) {
-        return SendReply("E00");
-    }
-
     if (!Memory::IsValidVirtualAddress(addr)) {
         return SendReply("E00");
     }
@@ -967,7 +1014,7 @@ void Break(bool is_memory_break) {
 static void Step() {
     if (command_length > 1) {
         RegWrite(PC_REGISTER, GdbHexToLong(command_buffer + 1), current_thread);
-        // Update Unicorn context skipping scheduler, no running threads at this point
+        // Update ARM context, skipping scheduler - no running threads at this point
         Core::System::GetInstance()
             .ArmInterface(current_core)
             .LoadContext(current_thread->GetContext());
@@ -1010,7 +1057,7 @@ static bool CommitBreakpoint(BreakpointType type, VAddr addr, u64 len) {
     breakpoint.addr = addr;
     breakpoint.len = len;
     Memory::ReadBlock(addr, breakpoint.inst.data(), breakpoint.inst.size());
-    static constexpr std::array<u8, 4> btrap{{0x00, 0x7d, 0x20, 0xd4}};
+    static constexpr std::array<u8, 4> btrap{0x00, 0x7d, 0x20, 0xd4};
     Memory::WriteBlock(addr, btrap.data(), btrap.size());
     Core::System::GetInstance().InvalidateCpuInstructionCaches();
     p.insert({addr, breakpoint});
@@ -1321,13 +1368,15 @@ void SetCpuStepFlag(bool is_step) {
 }
 
 void SendTrap(Kernel::Thread* thread, int trap) {
-    if (send_trap) {
-        if (!halt_loop || current_thread == thread) {
-            current_thread = thread;
-            SendSignal(thread, trap);
-        }
-        halt_loop = true;
-        send_trap = false;
+    if (!send_trap) {
+        return;
     }
+
+    if (!halt_loop || current_thread == thread) {
+        current_thread = thread;
+        SendSignal(thread, trap);
+    }
+    halt_loop = true;
+    send_trap = false;
 }
 }; // namespace GDBStub

src/core/hle/kernel/handle_table.cpp

@@ -12,12 +12,23 @@
 #include "core/hle/kernel/thread.h"
 
 namespace Kernel {
+namespace {
+constexpr u16 GetSlot(Handle handle) {
+    return handle >> 15;
+}
+
+constexpr u16 GetGeneration(Handle handle) {
+    return handle & 0x7FFF;
+}
+} // Anonymous namespace
 
 HandleTable::HandleTable() {
     next_generation = 1;
     Clear();
 }
 
+HandleTable::~HandleTable() = default;
+
 ResultVal<Handle> HandleTable::Create(SharedPtr<Object> obj) {
     DEBUG_ASSERT(obj != nullptr);
 

src/core/hle/kernel/handle_table.h

@@ -43,6 +43,7 @@ enum KernelHandle : Handle {
 class HandleTable final : NonCopyable {
 public:
     HandleTable();
+    ~HandleTable();
 
     /**
      * Allocates a handle for the given object.
@@ -89,18 +90,8 @@ public:
     void Clear();
 
 private:
-    /**
-     * This is the maximum limit of handles allowed per process in CTR-OS. It can be further
-     * reduced by ExHeader values, but this is not emulated here.
-     */
-    static const std::size_t MAX_COUNT = 4096;
-
-    static u16 GetSlot(Handle handle) {
-        return handle >> 15;
-    }
-    static u16 GetGeneration(Handle handle) {
-        return handle & 0x7FFF;
-    }
+    /// This is the maximum limit of handles allowed per process in Horizon
+    static constexpr std::size_t MAX_COUNT = 1024;
 
     /// Stores the Object referenced by the handle or null if the slot is empty.
     std::array<SharedPtr<Object>, MAX_COUNT> objects;

src/core/hle/kernel/process.cpp

@@ -4,6 +4,7 @@
 
 #include <algorithm>
 #include <memory>
+#include <random>
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/core.h"

src/core/hle/kernel/process.h

@@ -8,7 +8,6 @@
 #include <bitset>
 #include <cstddef>
 #include <memory>
-#include <random>
 #include <string>
 #include <vector>
 #include <boost/container/static_vector.hpp>

src/core/hle/kernel/shared_memory.cpp

@@ -61,7 +61,7 @@ SharedPtr<SharedMemory> SharedMemory::Create(KernelCore& kernel, SharedPtr<Proce
 }
 
 SharedPtr<SharedMemory> SharedMemory::CreateForApplet(
-    KernelCore& kernel, std::shared_ptr<std::vector<u8>> heap_block, u32 offset, u32 size,
+    KernelCore& kernel, std::shared_ptr<std::vector<u8>> heap_block, std::size_t offset, u64 size,
     MemoryPermission permissions, MemoryPermission other_permissions, std::string name) {
     SharedPtr<SharedMemory> shared_memory(new SharedMemory(kernel));
 
@@ -78,10 +78,10 @@ SharedPtr<SharedMemory> SharedMemory::CreateForApplet(
     return shared_memory;
 }
 
-ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermission permissions,
+ResultCode SharedMemory::Map(Process& target_process, VAddr address, MemoryPermission permissions,
                              MemoryPermission other_permissions) {
     const MemoryPermission own_other_permissions =
-        target_process == owner_process ? this->permissions : this->other_permissions;
+        &target_process == owner_process ? this->permissions : this->other_permissions;
 
     // Automatically allocated memory blocks can only be mapped with other_permissions = DontCare
     if (base_address == 0 && other_permissions != MemoryPermission::DontCare) {
@@ -106,7 +106,7 @@ ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermi
     VAddr target_address = address;
 
     // Map the memory block into the target process
-    auto result = target_process->VMManager().MapMemoryBlock(
+    auto result = target_process.VMManager().MapMemoryBlock(
         target_address, backing_block, backing_block_offset, size, MemoryState::Shared);
     if (result.Failed()) {
         LOG_ERROR(
@@ -116,14 +116,14 @@ ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermi
         return result.Code();
     }
 
-    return target_process->VMManager().ReprotectRange(target_address, size,
-                                                      ConvertPermissions(permissions));
+    return target_process.VMManager().ReprotectRange(target_address, size,
+                                                     ConvertPermissions(permissions));
 }
 
-ResultCode SharedMemory::Unmap(Process* target_process, VAddr address) {
+ResultCode SharedMemory::Unmap(Process& target_process, VAddr address) {
     // TODO(Subv): Verify what happens if the application tries to unmap an address that is not
     // mapped to a SharedMemory.
-    return target_process->VMManager().UnmapRange(address, size);
+    return target_process.VMManager().UnmapRange(address, size);
 }
 
 VMAPermission SharedMemory::ConvertPermissions(MemoryPermission permission) {
@@ -132,7 +132,11 @@ VMAPermission SharedMemory::ConvertPermissions(MemoryPermission permission) {
     return static_cast<VMAPermission>(masked_permissions);
 }
 
-u8* SharedMemory::GetPointer(u32 offset) {
+u8* SharedMemory::GetPointer(std::size_t offset) {
+    return backing_block->data() + backing_block_offset + offset;
+}
+
+const u8* SharedMemory::GetPointer(std::size_t offset) const {
     return backing_block->data() + backing_block_offset + offset;
 }
 

src/core/hle/kernel/shared_memory.h

@@ -64,7 +64,7 @@ public:
      */
     static SharedPtr<SharedMemory> CreateForApplet(KernelCore& kernel,
                                                    std::shared_ptr<std::vector<u8>> heap_block,
-                                                   u32 offset, u32 size,
+                                                   std::size_t offset, u64 size,
                                                    MemoryPermission permissions,
                                                    MemoryPermission other_permissions,
                                                    std::string name = "Unknown Applet");
@@ -81,6 +81,11 @@ public:
         return HANDLE_TYPE;
     }
 
+    /// Gets the size of the underlying memory block in bytes.
+    u64 GetSize() const {
+        return size;
+    }
+
     /**
      * Converts the specified MemoryPermission into the equivalent VMAPermission.
      * @param permission The MemoryPermission to convert.
@@ -94,44 +99,51 @@ public:
      * @param permissions Memory block map permissions (specified by SVC field)
      * @param other_permissions Memory block map other permissions (specified by SVC field)
      */
-    ResultCode Map(Process* target_process, VAddr address, MemoryPermission permissions,
+    ResultCode Map(Process& target_process, VAddr address, MemoryPermission permissions,
                    MemoryPermission other_permissions);
 
     /**
      * Unmaps a shared memory block from the specified address in system memory
-     * @param target_process Process from which to umap the memory block.
+     * @param target_process Process from which to unmap the memory block.
      * @param address Address in system memory where the shared memory block is mapped
      * @return Result code of the unmap operation
      */
-    ResultCode Unmap(Process* target_process, VAddr address);
+    ResultCode Unmap(Process& target_process, VAddr address);
 
     /**
      * Gets a pointer to the shared memory block
      * @param offset Offset from the start of the shared memory block to get pointer
-     * @return Pointer to the shared memory block from the specified offset
+     * @return A pointer to the shared memory block from the specified offset
      */
-    u8* GetPointer(u32 offset = 0);
+    u8* GetPointer(std::size_t offset = 0);
 
-    /// Process that created this shared memory block.
-    SharedPtr<Process> owner_process;
-    /// Address of shared memory block in the owner process if specified.
-    VAddr base_address;
-    /// Backing memory for this shared memory block.
-    std::shared_ptr<std::vector<u8>> backing_block;
-    /// Offset into the backing block for this shared memory.
-    std::size_t backing_block_offset;
-    /// Size of the memory block. Page-aligned.
-    u64 size;
-    /// Permission restrictions applied to the process which created the block.
-    MemoryPermission permissions;
-    /// Permission restrictions applied to other processes mapping the block.
-    MemoryPermission other_permissions;
-    /// Name of shared memory object.
-    std::string name;
+    /**
+     * Gets a constant pointer to the shared memory block
+     * @param offset Offset from the start of the shared memory block to get pointer
+     * @return A constant pointer to the shared memory block from the specified offset
+     */
+    const u8* GetPointer(std::size_t offset = 0) const;
 
 private:
     explicit SharedMemory(KernelCore& kernel);
     ~SharedMemory() override;
+
+    /// Backing memory for this shared memory block.
+    std::shared_ptr<std::vector<u8>> backing_block;
+    /// Offset into the backing block for this shared memory.
+    std::size_t backing_block_offset = 0;
+    /// Size of the memory block. Page-aligned.
+    u64 size = 0;
+    /// Permission restrictions applied to the process which created the block.
+    MemoryPermission permissions{};
+    /// Permission restrictions applied to other processes mapping the block.
+    MemoryPermission other_permissions{};
+    /// Process that created this shared memory block.
+    SharedPtr<Process> owner_process;
+    /// Address of shared memory block in the owner process if specified.
+    VAddr base_address = 0;
+    /// Name of shared memory object.
+    std::string name;
 };
 
 } // namespace Kernel

src/core/hle/kernel/svc.cpp

@@ -671,7 +671,8 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
         break;
     }
     default:
-        UNIMPLEMENTED();
+        LOG_WARNING(Kernel_SVC, "(STUBBED) Unimplemented svcGetInfo id=0x{:016X}", info_id);
+        return ERR_INVALID_ENUM_VALUE;
     }
 
     return RESULT_SUCCESS;
@@ -789,7 +790,7 @@ static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 s
         return ERR_INVALID_MEMORY_RANGE;
     }
 
-    return shared_memory->Map(current_process, addr, permissions_type, MemoryPermission::DontCare);
+    return shared_memory->Map(*current_process, addr, permissions_type, MemoryPermission::DontCare);
 }
 
 static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 size) {
@@ -819,7 +820,7 @@ static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64
         return ERR_INVALID_MEMORY_RANGE;
     }
 
-    return shared_memory->Unmap(current_process, addr);
+    return shared_memory->Unmap(*current_process, addr);
 }
 
 /// Query process memory

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

@@ -6,8 +6,6 @@
 #include <cinttypes>
 #include <cstring>
 #include <stack>
-#include "applets/applets.h"
-#include "applets/software_keyboard.h"
 #include "audio_core/audio_renderer.h"
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
@@ -18,6 +16,9 @@
 #include "core/hle/service/am/am.h"
 #include "core/hle/service/am/applet_ae.h"
 #include "core/hle/service/am/applet_oe.h"
+#include "core/hle/service/am/applets/applets.h"
+#include "core/hle/service/am/applets/software_keyboard.h"
+#include "core/hle/service/am/applets/stub_applet.h"
 #include "core/hle/service/am/idle.h"
 #include "core/hle/service/am/omm.h"
 #include "core/hle/service/am/spsm.h"
@@ -482,11 +483,15 @@ void ICommonStateGetter::GetDefaultDisplayResolution(Kernel::HLERequestContext&
     rb.Push(RESULT_SUCCESS);
 
     if (Settings::values.use_docked_mode) {
-        rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth));
-        rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight));
+        rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth) *
+                static_cast<u32>(Settings::values.resolution_factor));
+        rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight) *
+                static_cast<u32>(Settings::values.resolution_factor));
     } else {
-        rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedWidth));
-        rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedHeight));
+        rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedWidth) *
+                static_cast<u32>(Settings::values.resolution_factor));
+        rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedHeight) *
+                static_cast<u32>(Settings::values.resolution_factor));
     }
 
     LOG_DEBUG(Service_AM, "called");
@@ -532,8 +537,7 @@ void ICommonStateGetter::GetPerformanceMode(Kernel::HLERequestContext& ctx) {
 class ILibraryAppletAccessor final : public ServiceFramework<ILibraryAppletAccessor> {
 public:
     explicit ILibraryAppletAccessor(std::shared_ptr<Applets::Applet> applet)
-        : ServiceFramework("ILibraryAppletAccessor"), applet(std::move(applet)),
-          broker(std::make_shared<Applets::AppletDataBroker>()) {
+        : ServiceFramework("ILibraryAppletAccessor"), applet(std::move(applet)) {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, &ILibraryAppletAccessor::GetAppletStateChangedEvent, "GetAppletStateChangedEvent"},
@@ -562,7 +566,7 @@ public:
 
 private:
     void GetAppletStateChangedEvent(Kernel::HLERequestContext& ctx) {
-        const auto event = broker->GetStateChangedEvent();
+        const auto event = applet->GetBroker().GetStateChangedEvent();
         event->Signal();
 
         IPC::ResponseBuilder rb{ctx, 2, 1};
@@ -590,7 +594,7 @@ private:
     void Start(Kernel::HLERequestContext& ctx) {
         ASSERT(applet != nullptr);
 
-        applet->Initialize(broker);
+        applet->Initialize();
         applet->Execute();
 
         IPC::ResponseBuilder rb{ctx, 2};
@@ -601,7 +605,7 @@ private:
 
     void PushInData(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
-        broker->PushNormalDataFromGame(*rp.PopIpcInterface<IStorage>());
+        applet->GetBroker().PushNormalDataFromGame(*rp.PopIpcInterface<IStorage>());
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
@@ -612,7 +616,7 @@ private:
     void PopOutData(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 0, 1};
 
-        const auto storage = broker->PopNormalDataToGame();
+        const auto storage = applet->GetBroker().PopNormalDataToGame();
         if (storage == nullptr) {
             rb.Push(ERR_NO_DATA_IN_CHANNEL);
             return;
@@ -626,7 +630,7 @@ private:
 
     void PushInteractiveInData(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
-        broker->PushInteractiveDataFromGame(*rp.PopIpcInterface<IStorage>());
+        applet->GetBroker().PushInteractiveDataFromGame(*rp.PopIpcInterface<IStorage>());
 
         ASSERT(applet->IsInitialized());
         applet->ExecuteInteractive();
@@ -641,7 +645,7 @@ private:
     void PopInteractiveOutData(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 0, 1};
 
-        const auto storage = broker->PopInteractiveDataToGame();
+        const auto storage = applet->GetBroker().PopInteractiveDataToGame();
         if (storage == nullptr) {
             rb.Push(ERR_NO_DATA_IN_CHANNEL);
             return;
@@ -656,7 +660,7 @@ private:
     void GetPopOutDataEvent(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushCopyObjects(broker->GetNormalDataEvent());
+        rb.PushCopyObjects(applet->GetBroker().GetNormalDataEvent());
 
         LOG_DEBUG(Service_AM, "called");
     }
@@ -664,13 +668,12 @@ private:
     void GetPopInteractiveOutDataEvent(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushCopyObjects(broker->GetInteractiveDataEvent());
+        rb.PushCopyObjects(applet->GetBroker().GetInteractiveDataEvent());
 
         LOG_DEBUG(Service_AM, "called");
     }
 
     std::shared_ptr<Applets::Applet> applet;
-    std::shared_ptr<Applets::AppletDataBroker> broker;
 };
 
 void IStorage::Open(Kernel::HLERequestContext& ctx) {
@@ -763,8 +766,9 @@ static std::shared_ptr<Applets::Applet> GetAppletFromId(AppletId id) {
     case AppletId::SoftwareKeyboard:
         return std::make_shared<Applets::SoftwareKeyboard>();
     default:
-        UNREACHABLE_MSG("Unimplemented AppletId [{:08X}]!", static_cast<u32>(id));
-        return nullptr;
+        LOG_ERROR(Service_AM, "Unimplemented AppletId [{:08X}]! -- Falling back to stub!",
+                  static_cast<u32>(id));
+        return std::make_shared<Applets::StubApplet>();
     }
 }
 
@@ -820,8 +824,8 @@ void ILibraryAppletCreator::CreateTransferMemoryStorage(Kernel::HLERequestContex
         return;
     }
 
-    const auto mem_begin = shared_mem->backing_block->begin() + shared_mem->backing_block_offset;
-    const auto mem_end = mem_begin + shared_mem->size;
+    const u8* mem_begin = shared_mem->GetPointer();
+    const u8* mem_end = mem_begin + shared_mem->GetSize();
     std::vector<u8> memory{mem_begin, mem_end};
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};

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

@@ -98,10 +98,8 @@ Applet::Applet() = default;
 
 Applet::~Applet() = default;
 
-void Applet::Initialize(std::shared_ptr<AppletDataBroker> broker_) {
-    broker = std::move(broker_);
-
-    const auto common = broker->PopNormalDataToApplet();
+void Applet::Initialize() {
+    const auto common = broker.PopNormalDataToApplet();
     ASSERT(common != nullptr);
 
     const auto common_data = common->GetData();

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

@@ -4,14 +4,17 @@
 
 #pragma once
 
-#include <functional>
 #include <memory>
 #include <queue>
 #include "common/swap.h"
-#include "core/hle/kernel/event.h"
+#include "core/hle/kernel/kernel.h"
 
 union ResultCode;
 
+namespace Kernel {
+class Event;
+}
+
 namespace Service::AM {
 
 class IStorage;
@@ -43,19 +46,26 @@ public:
 
 private:
     // Queues are named from applet's perspective
-    std::queue<std::unique_ptr<IStorage>>
-        in_channel; // PopNormalDataToApplet and PushNormalDataFromGame
-    std::queue<std::unique_ptr<IStorage>>
-        out_channel; // PopNormalDataToGame and PushNormalDataFromApplet
-    std::queue<std::unique_ptr<IStorage>>
-        in_interactive_channel; // PopInteractiveDataToApplet and PushInteractiveDataFromGame
-    std::queue<std::unique_ptr<IStorage>>
-        out_interactive_channel; // PopInteractiveDataToGame and PushInteractiveDataFromApplet
+
+    // PopNormalDataToApplet and PushNormalDataFromGame
+    std::queue<std::unique_ptr<IStorage>> in_channel;
+
+    // PopNormalDataToGame and PushNormalDataFromApplet
+    std::queue<std::unique_ptr<IStorage>> out_channel;
+
+    // PopInteractiveDataToApplet and PushInteractiveDataFromGame
+    std::queue<std::unique_ptr<IStorage>> in_interactive_channel;
+
+    // PopInteractiveDataToGame and PushInteractiveDataFromApplet
+    std::queue<std::unique_ptr<IStorage>> out_interactive_channel;
 
     Kernel::SharedPtr<Kernel::Event> state_changed_event;
-    Kernel::SharedPtr<Kernel::Event> pop_out_data_event; // Signaled on PushNormalDataFromApplet
-    Kernel::SharedPtr<Kernel::Event>
-        pop_interactive_out_data_event; // Signaled on PushInteractiveDataFromApplet
+
+    // Signaled on PushNormalDataFromApplet
+    Kernel::SharedPtr<Kernel::Event> pop_out_data_event;
+
+    // Signaled on PushInteractiveDataFromApplet
+    Kernel::SharedPtr<Kernel::Event> pop_interactive_out_data_event;
 };
 
 class Applet {
@@ -63,7 +73,7 @@ public:
     Applet();
     virtual ~Applet();
 
-    virtual void Initialize(std::shared_ptr<AppletDataBroker> broker);
+    virtual void Initialize();
 
     virtual bool TransactionComplete() const = 0;
     virtual ResultCode GetStatus() const = 0;
@@ -74,6 +84,14 @@ public:
         return initialized;
     }
 
+    AppletDataBroker& GetBroker() {
+        return broker;
+    }
+
+    const AppletDataBroker& GetBroker() const {
+        return broker;
+    }
+
 protected:
     struct CommonArguments {
         u32_le arguments_version;
@@ -85,8 +103,8 @@ protected:
     };
     static_assert(sizeof(CommonArguments) == 0x20, "CommonArguments has incorrect size.");
 
-    CommonArguments common_args;
-    std::shared_ptr<AppletDataBroker> broker;
+    CommonArguments common_args{};
+    AppletDataBroker broker;
     bool initialized = false;
 };
 

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

@@ -42,21 +42,21 @@ SoftwareKeyboard::SoftwareKeyboard() = default;
 
 SoftwareKeyboard::~SoftwareKeyboard() = default;
 
-void SoftwareKeyboard::Initialize(std::shared_ptr<AppletDataBroker> broker_) {
+void SoftwareKeyboard::Initialize() {
     complete = false;
     initial_text.clear();
     final_data.clear();
 
-    Applet::Initialize(std::move(broker_));
+    Applet::Initialize();
 
-    const auto keyboard_config_storage = broker->PopNormalDataToApplet();
+    const auto keyboard_config_storage = broker.PopNormalDataToApplet();
     ASSERT(keyboard_config_storage != nullptr);
     const auto& keyboard_config = keyboard_config_storage->GetData();
 
     ASSERT(keyboard_config.size() >= sizeof(KeyboardConfig));
     std::memcpy(&config, keyboard_config.data(), sizeof(KeyboardConfig));
 
-    const auto work_buffer_storage = broker->PopNormalDataToApplet();
+    const auto work_buffer_storage = broker.PopNormalDataToApplet();
     ASSERT(work_buffer_storage != nullptr);
     const auto& work_buffer = work_buffer_storage->GetData();
 
@@ -81,7 +81,7 @@ void SoftwareKeyboard::ExecuteInteractive() {
     if (complete)
         return;
 
-    const auto storage = broker->PopInteractiveDataToApplet();
+    const auto storage = broker.PopInteractiveDataToApplet();
     ASSERT(storage != nullptr);
     const auto data = storage->GetData();
     const auto status = static_cast<bool>(data[0]);
@@ -95,13 +95,13 @@ void SoftwareKeyboard::ExecuteInteractive() {
         std::memcpy(string.data(), data.data() + 4, string.size() * 2);
         frontend.SendTextCheckDialog(
             Common::UTF16StringFromFixedZeroTerminatedBuffer(string.data(), string.size()),
-            [this] { broker->SignalStateChanged(); });
+            [this] { broker.SignalStateChanged(); });
     }
 }
 
 void SoftwareKeyboard::Execute() {
     if (complete) {
-        broker->PushNormalDataFromApplet(IStorage{final_data});
+        broker.PushNormalDataFromApplet(IStorage{final_data});
         return;
     }
 
@@ -145,17 +145,17 @@ void SoftwareKeyboard::WriteText(std::optional<std::u16string> text) {
         final_data = output_main;
 
         if (complete) {
-            broker->PushNormalDataFromApplet(IStorage{output_main});
+            broker.PushNormalDataFromApplet(IStorage{output_main});
         } else {
-            broker->PushInteractiveDataFromApplet(IStorage{output_sub});
+            broker.PushInteractiveDataFromApplet(IStorage{output_sub});
         }
 
-        broker->SignalStateChanged();
+        broker.SignalStateChanged();
     } else {
         output_main[0] = 1;
         complete = true;
-        broker->PushNormalDataFromApplet(IStorage{output_main});
-        broker->SignalStateChanged();
+        broker.PushNormalDataFromApplet(IStorage{output_main});
+        broker.SignalStateChanged();
     }
 }
 } // namespace Service::AM::Applets

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

@@ -4,7 +4,12 @@
 
 #pragma once
 
+#include <array>
+#include <string>
+#include <vector>
+
 #include "common/common_funcs.h"
+#include "common/swap.h"
 #include "core/hle/service/am/am.h"
 #include "core/hle/service/am/applets/applets.h"
 
@@ -50,7 +55,7 @@ public:
     SoftwareKeyboard();
     ~SoftwareKeyboard() override;
 
-    void Initialize(std::shared_ptr<AppletDataBroker> broker) override;
+    void Initialize() override;
 
     bool TransactionComplete() const override;
     ResultCode GetStatus() const override;

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

@@ -0,0 +1,70 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <string>
+
+#include "common/hex_util.h"
+#include "common/logging/log.h"
+#include "core/hle/result.h"
+#include "core/hle/service/am/am.h"
+#include "core/hle/service/am/applets/stub_applet.h"
+
+namespace Service::AM::Applets {
+
+static void LogCurrentStorage(AppletDataBroker& broker, std::string prefix) {
+    std::unique_ptr<IStorage> storage = broker.PopNormalDataToApplet();
+    for (; storage != nullptr; storage = broker.PopNormalDataToApplet()) {
+        const auto data = storage->GetData();
+        LOG_INFO(Service_AM,
+                 "called (STUBBED), during {} recieved normal data with size={:08X}, data={}",
+                 prefix, data.size(), Common::HexVectorToString(data));
+    }
+
+    storage = broker.PopInteractiveDataToApplet();
+    for (; storage != nullptr; storage = broker.PopInteractiveDataToApplet()) {
+        const auto data = storage->GetData();
+        LOG_INFO(Service_AM,
+                 "called (STUBBED), during {} recieved interactive data with size={:08X}, data={}",
+                 prefix, data.size(), Common::HexVectorToString(data));
+    }
+}
+
+StubApplet::StubApplet() = default;
+
+StubApplet::~StubApplet() = default;
+
+void StubApplet::Initialize() {
+    LOG_WARNING(Service_AM, "called (STUBBED)");
+    Applet::Initialize();
+    LogCurrentStorage(broker, "Initialize");
+}
+
+bool StubApplet::TransactionComplete() const {
+    LOG_WARNING(Service_AM, "called (STUBBED)");
+    return true;
+}
+
+ResultCode StubApplet::GetStatus() const {
+    LOG_WARNING(Service_AM, "called (STUBBED)");
+    return RESULT_SUCCESS;
+}
+
+void StubApplet::ExecuteInteractive() {
+    LOG_WARNING(Service_AM, "called (STUBBED)");
+    LogCurrentStorage(broker, "ExecuteInteractive");
+
+    broker.PushNormalDataFromApplet(IStorage{std::vector<u8>(0x1000)});
+    broker.PushInteractiveDataFromApplet(IStorage{std::vector<u8>(0x1000)});
+    broker.SignalStateChanged();
+}
+
+void StubApplet::Execute() {
+    LOG_WARNING(Service_AM, "called (STUBBED)");
+    LogCurrentStorage(broker, "Execute");
+
+    broker.PushNormalDataFromApplet(IStorage{std::vector<u8>(0x1000)});
+    broker.PushInteractiveDataFromApplet(IStorage{std::vector<u8>(0x1000)});
+    broker.SignalStateChanged();
+}
+} // namespace Service::AM::Applets

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

@@ -0,0 +1,24 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/service/am/applets/applets.h"
+
+namespace Service::AM::Applets {
+
+class StubApplet final : public Applet {
+public:
+    StubApplet();
+    ~StubApplet() override;
+
+    void Initialize() override;
+
+    bool TransactionComplete() const override;
+    ResultCode GetStatus() const override;
+    void ExecuteInteractive() override;
+    void Execute() override;
+};
+
+} // namespace Service::AM::Applets

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

@@ -44,8 +44,10 @@ enum class AudioState : u32 {
 
 class IAudioOut final : public ServiceFramework<IAudioOut> {
 public:
-    IAudioOut(AudoutParams audio_params, AudioCore::AudioOut& audio_core)
-        : ServiceFramework("IAudioOut"), audio_core(audio_core), audio_params(audio_params) {
+    IAudioOut(AudoutParams audio_params, AudioCore::AudioOut& audio_core, std::string&& device_name,
+              std::string&& unique_name)
+        : ServiceFramework("IAudioOut"), audio_core(audio_core), audio_params(audio_params),
+          device_name(std::move(device_name)) {
 
         static const FunctionInfo functions[] = {
             {0, &IAudioOut::GetAudioOutState, "GetAudioOutState"},
@@ -69,7 +71,7 @@ public:
             Kernel::Event::Create(kernel, Kernel::ResetType::Sticky, "IAudioOutBufferReleased");
 
         stream = audio_core.OpenStream(audio_params.sample_rate, audio_params.channel_count,
-                                       "IAudioOut", [=]() { buffer_event->Signal(); });
+                                       std::move(unique_name), [=]() { buffer_event->Signal(); });
     }
 
 private:
@@ -177,6 +179,7 @@ private:
 
     AudioCore::AudioOut& audio_core;
     AudioCore::StreamPtr stream;
+    std::string device_name;
 
     AudoutParams audio_params{};
 
@@ -199,7 +202,15 @@ void AudOutU::ListAudioOutsImpl(Kernel::HLERequestContext& ctx) {
 void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_Audio, "called");
 
-    ctx.WriteBuffer(DefaultDevice);
+    const auto device_name_data{ctx.ReadBuffer()};
+    std::string device_name;
+    if (device_name_data[0] != '\0') {
+        device_name.assign(device_name_data.begin(), device_name_data.end());
+    } else {
+        device_name.assign(DefaultDevice.begin(), DefaultDevice.end());
+    }
+    ctx.WriteBuffer(device_name);
+
     IPC::RequestParser rp{ctx};
     auto params{rp.PopRaw<AudoutParams>()};
     if (params.channel_count <= 2) {
@@ -212,10 +223,9 @@ void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {
         params.sample_rate = DefaultSampleRate;
     }
 
-    // TODO(bunnei): Support more than one IAudioOut interface. When we add this, ListAudioOutsImpl
-    // will likely need to be updated as well.
-    ASSERT_MSG(!audio_out_interface, "Unimplemented");
-    audio_out_interface = std::make_shared<IAudioOut>(params, *audio_core);
+    std::string unique_name{fmt::format("{}-{}", device_name, audio_out_interfaces.size())};
+    auto audio_out_interface = std::make_shared<IAudioOut>(
+        params, *audio_core, std::move(device_name), std::move(unique_name));
 
     IPC::ResponseBuilder rb{ctx, 6, 0, 1};
     rb.Push(RESULT_SUCCESS);
@@ -224,6 +234,8 @@ void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {
     rb.Push<u32>(static_cast<u32>(AudioCore::Codec::PcmFormat::Int16));
     rb.Push<u32>(static_cast<u32>(AudioState::Stopped));
     rb.PushIpcInterface<Audio::IAudioOut>(audio_out_interface);
+
+    audio_out_interfaces.push_back(std::move(audio_out_interface));
 }
 
 AudOutU::AudOutU() : ServiceFramework("audout:u") {

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

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include <vector>
 #include "core/hle/service/service.h"
 
 namespace AudioCore {
@@ -24,7 +25,7 @@ public:
     ~AudOutU() override;
 
 private:
-    std::shared_ptr<IAudioOut> audio_out_interface;
+    std::vector<std::shared_ptr<IAudioOut>> audio_out_interfaces;
     std::unique_ptr<AudioCore::AudioOut> audio_core;
 
     void ListAudioOutsImpl(Kernel::HLERequestContext& ctx);

src/core/hle/service/filesystem/filesystem.cpp

@@ -341,6 +341,10 @@ std::shared_ptr<FileSys::RegisteredCacheUnion> GetUnionContents() {
     return registered_cache_union;
 }
 
+void ClearUnionContents() {
+    registered_cache_union = nullptr;
+}
+
 FileSys::RegisteredCache* GetSystemNANDContents() {
     LOG_TRACE(Service_FS, "Opening System NAND Contents");
 
@@ -391,6 +395,7 @@ void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite) {
         bis_factory = nullptr;
         save_data_factory = nullptr;
         sdmc_factory = nullptr;
+        ClearUnionContents();
     }
 
     auto nand_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir),

src/core/hle/service/filesystem/filesystem.h

@@ -49,6 +49,7 @@ ResultVal<FileSys::VirtualDir> OpenSaveDataSpace(FileSys::SaveDataSpaceId space)
 ResultVal<FileSys::VirtualDir> OpenSDMC();
 
 std::shared_ptr<FileSys::RegisteredCacheUnion> GetUnionContents();
+void ClearUnionContents();
 
 FileSys::RegisteredCache* GetSystemNANDContents();
 FileSys::RegisteredCache* GetUserNANDContents();

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

@@ -71,8 +71,9 @@ void Controller_DebugPad::OnUpdate(u8* data, std::size_t size) {
 
 void Controller_DebugPad::OnLoadInputDevices() {
     std::transform(Settings::values.debug_pad_buttons.begin(),
-                   Settings::values.debug_pad_buttons.end(), buttons.begin(),
-                   Input::CreateDevice<Input::ButtonDevice>);
+                   Settings::values.debug_pad_buttons.begin() +
+                       Settings::NativeButton::NUM_BUTTONS_HID,
+                   buttons.begin(), Input::CreateDevice<Input::ButtonDevice>);
     std::transform(Settings::values.debug_pad_analogs.begin(),
                    Settings::values.debug_pad_analogs.end(), analogs.begin(),
                    Input::CreateDevice<Input::AnalogDevice>);

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

@@ -575,8 +575,8 @@ void Controller_NPad::AddNewControllerAt(NPadControllerType controller, u32 npad
         return;
     }
 
-    connected_controllers[npad_id] = {controller, true};
-    InitNewlyAddedControler(npad_id);
+    connected_controllers[NPadIdToIndex(npad_id)] = {controller, true};
+    InitNewlyAddedControler(NPadIdToIndex(npad_id));
 }
 
 void Controller_NPad::ConnectNPad(u32 npad_id) {

src/core/hle/service/ldr/ldr.cpp

@@ -16,35 +16,18 @@
 
 namespace Service::LDR {
 
-namespace ErrCodes {
-enum {
-    InvalidMemoryState = 51,
-    InvalidNRO = 52,
-    InvalidNRR = 53,
-    MissingNRRHash = 54,
-    MaximumNRO = 55,
-    MaximumNRR = 56,
-    AlreadyLoaded = 57,
-    InvalidAlignment = 81,
-    InvalidSize = 82,
-    InvalidNROAddress = 84,
-    InvalidNRRAddress = 85,
-    NotInitialized = 87,
-};
-}
-
-constexpr ResultCode ERROR_INVALID_MEMORY_STATE(ErrorModule::Loader, ErrCodes::InvalidMemoryState);
-constexpr ResultCode ERROR_INVALID_NRO(ErrorModule::Loader, ErrCodes::InvalidNRO);
-constexpr ResultCode ERROR_INVALID_NRR(ErrorModule::Loader, ErrCodes::InvalidNRR);
-constexpr ResultCode ERROR_MISSING_NRR_HASH(ErrorModule::Loader, ErrCodes::MissingNRRHash);
-constexpr ResultCode ERROR_MAXIMUM_NRO(ErrorModule::Loader, ErrCodes::MaximumNRO);
-constexpr ResultCode ERROR_MAXIMUM_NRR(ErrorModule::Loader, ErrCodes::MaximumNRR);
-constexpr ResultCode ERROR_ALREADY_LOADED(ErrorModule::Loader, ErrCodes::AlreadyLoaded);
-constexpr ResultCode ERROR_INVALID_ALIGNMENT(ErrorModule::Loader, ErrCodes::InvalidAlignment);
-constexpr ResultCode ERROR_INVALID_SIZE(ErrorModule::Loader, ErrCodes::InvalidSize);
-constexpr ResultCode ERROR_INVALID_NRO_ADDRESS(ErrorModule::Loader, ErrCodes::InvalidNROAddress);
-constexpr ResultCode ERROR_INVALID_NRR_ADDRESS(ErrorModule::Loader, ErrCodes::InvalidNRRAddress);
-constexpr ResultCode ERROR_NOT_INITIALIZED(ErrorModule::Loader, ErrCodes::NotInitialized);
+constexpr ResultCode ERROR_INVALID_MEMORY_STATE{ErrorModule::Loader, 51};
+constexpr ResultCode ERROR_INVALID_NRO{ErrorModule::Loader, 52};
+constexpr ResultCode ERROR_INVALID_NRR{ErrorModule::Loader, 53};
+constexpr ResultCode ERROR_MISSING_NRR_HASH{ErrorModule::Loader, 54};
+constexpr ResultCode ERROR_MAXIMUM_NRO{ErrorModule::Loader, 55};
+constexpr ResultCode ERROR_MAXIMUM_NRR{ErrorModule::Loader, 56};
+constexpr ResultCode ERROR_ALREADY_LOADED{ErrorModule::Loader, 57};
+constexpr ResultCode ERROR_INVALID_ALIGNMENT{ErrorModule::Loader, 81};
+constexpr ResultCode ERROR_INVALID_SIZE{ErrorModule::Loader, 82};
+constexpr ResultCode ERROR_INVALID_NRO_ADDRESS{ErrorModule::Loader, 84};
+constexpr ResultCode ERROR_INVALID_NRR_ADDRESS{ErrorModule::Loader, 85};
+constexpr ResultCode ERROR_NOT_INITIALIZED{ErrorModule::Loader, 87};
 
 constexpr u64 MAXIMUM_LOADED_RO = 0x40;
 

src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp

@@ -5,6 +5,8 @@
 #include <cstring>
 #include "common/assert.h"
 #include "common/logging/log.h"
+#include "core/core_timing.h"
+#include "core/core_timing_util.h"
 #include "core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h"
 
 namespace Service::Nvidia::Devices {
@@ -33,6 +35,8 @@ u32 nvhost_ctrl_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vec
         return ZBCQueryTable(input, output);
     case IoctlCommand::IocFlushL2:
         return FlushL2(input, output);
+    case IoctlCommand::IocGetGpuTime:
+        return GetGpuTime(input, output);
     }
     UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
@@ -169,4 +173,13 @@ u32 nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>& outp
     return 0;
 }
 
+u32 nvhost_ctrl_gpu::GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output) {
+    LOG_DEBUG(Service_NVDRV, "called");
+    IoctlGetGpuTime params{};
+    std::memcpy(&params, input.data(), input.size());
+    params.gpu_time = CoreTiming::cyclesToNs(CoreTiming::GetTicks());
+    std::memcpy(output.data(), &params, output.size());
+    return 0;
+}
+
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h

@@ -156,6 +156,11 @@ private:
     };
     static_assert(sizeof(IoctlFlushL2) == 8, "IoctlFlushL2 is incorrect size");
 
+    struct IoctlGetGpuTime {
+        u64_le gpu_time;
+    };
+    static_assert(sizeof(IoctlGetGpuTime) == 8, "IoctlGetGpuTime is incorrect size");
+
     u32 GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output);
     u32 GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output);
     u32 GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output);
@@ -164,6 +169,7 @@ private:
     u32 ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output);
     u32 ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output);
     u32 FlushL2(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output);
 };
 
 } // namespace Service::Nvidia::Devices

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

@@ -88,6 +88,20 @@ void NVDRV::FinishInitialize(Kernel::HLERequestContext& ctx) {
     rb.Push(RESULT_SUCCESS);
 }
 
+void NVDRV::GetStatus(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_NVDRV, "(STUBBED) called");
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void NVDRV::DumpGraphicsMemoryInfo(Kernel::HLERequestContext& ctx) {
+    // According to SwitchBrew, this has no inputs and no outputs, so effectively does nothing on
+    // retail hardware.
+    LOG_DEBUG(Service_NVDRV, "called");
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 NVDRV::NVDRV(std::shared_ptr<Module> nvdrv, const char* name)
     : ServiceFramework(name), nvdrv(std::move(nvdrv)) {
     static const FunctionInfo functions[] = {
@@ -97,10 +111,10 @@ NVDRV::NVDRV(std::shared_ptr<Module> nvdrv, const char* name)
         {3, &NVDRV::Initialize, "Initialize"},
         {4, &NVDRV::QueryEvent, "QueryEvent"},
         {5, nullptr, "MapSharedMem"},
-        {6, nullptr, "GetStatus"},
+        {6, &NVDRV::GetStatus, "GetStatus"},
         {7, nullptr, "ForceSetClientPID"},
         {8, &NVDRV::SetClientPID, "SetClientPID"},
-        {9, nullptr, "DumpGraphicsMemoryInfo"},
+        {9, &NVDRV::DumpGraphicsMemoryInfo, "DumpGraphicsMemoryInfo"},
         {10, nullptr, "InitializeDevtools"},
         {11, &NVDRV::Ioctl, "Ioctl2"},
         {12, nullptr, "Ioctl3"},

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

@@ -24,6 +24,8 @@ private:
     void QueryEvent(Kernel::HLERequestContext& ctx);
     void SetClientPID(Kernel::HLERequestContext& ctx);
     void FinishInitialize(Kernel::HLERequestContext& ctx);
+    void GetStatus(Kernel::HLERequestContext& ctx);
+    void DumpGraphicsMemoryInfo(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<Module> nvdrv;
 

src/core/hle/service/sm/sm.cpp

@@ -63,6 +63,17 @@ ResultVal<Kernel::SharedPtr<Kernel::ServerPort>> ServiceManager::RegisterService
     return MakeResult<Kernel::SharedPtr<Kernel::ServerPort>>(std::move(server_port));
 }
 
+ResultCode ServiceManager::UnregisterService(const std::string& name) {
+    CASCADE_CODE(ValidateServiceName(name));
+
+    const auto iter = registered_services.find(name);
+    if (iter == registered_services.end())
+        return ERR_SERVICE_NOT_REGISTERED;
+
+    registered_services.erase(iter);
+    return RESULT_SUCCESS;
+}
+
 ResultVal<Kernel::SharedPtr<Kernel::ClientPort>> ServiceManager::GetServicePort(
     const std::string& name) {
 
@@ -127,13 +138,52 @@ void SM::GetService(Kernel::HLERequestContext& ctx) {
     }
 }
 
+void SM::RegisterService(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+
+    const auto name_buf = rp.PopRaw<std::array<char, 8>>();
+    const auto end = std::find(name_buf.begin(), name_buf.end(), '\0');
+
+    const std::string name(name_buf.begin(), end);
+
+    const auto unk_bool = static_cast<bool>(rp.PopRaw<u32>());
+    const auto session_count = rp.PopRaw<u32>();
+
+    LOG_DEBUG(Service_SM, "called with unk_bool={}", unk_bool);
+
+    auto handle = service_manager->RegisterService(name, session_count);
+    if (handle.Failed()) {
+        LOG_ERROR(Service_SM, "failed to register service with error_code={:08X}",
+                  handle.Code().raw);
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(handle.Code());
+        return;
+    }
+
+    IPC::ResponseBuilder rb{ctx, 2, 0, 1, IPC::ResponseBuilder::Flags::AlwaysMoveHandles};
+    rb.Push(handle.Code());
+    rb.PushMoveObjects(std::move(handle).Unwrap());
+}
+
+void SM::UnregisterService(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+
+    const auto name_buf = rp.PopRaw<std::array<char, 8>>();
+    const auto end = std::find(name_buf.begin(), name_buf.end(), '\0');
+
+    const std::string name(name_buf.begin(), end);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(service_manager->UnregisterService(name));
+}
+
 SM::SM(std::shared_ptr<ServiceManager> service_manager)
     : ServiceFramework("sm:", 4), service_manager(std::move(service_manager)) {
     static const FunctionInfo functions[] = {
         {0x00000000, &SM::Initialize, "Initialize"},
         {0x00000001, &SM::GetService, "GetService"},
-        {0x00000002, nullptr, "RegisterService"},
-        {0x00000003, nullptr, "UnregisterService"},
+        {0x00000002, &SM::RegisterService, "RegisterService"},
+        {0x00000003, &SM::UnregisterService, "UnregisterService"},
     };
     RegisterHandlers(functions);
 }

src/core/hle/service/sm/sm.h

@@ -35,6 +35,8 @@ public:
 private:
     void Initialize(Kernel::HLERequestContext& ctx);
     void GetService(Kernel::HLERequestContext& ctx);
+    void RegisterService(Kernel::HLERequestContext& ctx);
+    void UnregisterService(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<ServiceManager> service_manager;
 };
@@ -48,6 +50,7 @@ public:
 
     ResultVal<Kernel::SharedPtr<Kernel::ServerPort>> RegisterService(std::string name,
                                                                      unsigned int max_sessions);
+    ResultCode UnregisterService(const std::string& name);
     ResultVal<Kernel::SharedPtr<Kernel::ClientPort>> GetServicePort(const std::string& name);
     ResultVal<Kernel::SharedPtr<Kernel::ClientSession>> ConnectToService(const std::string& name);
 

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

@@ -510,7 +510,11 @@ private:
 
         if (transaction == TransactionId::Connect) {
             IGBPConnectRequestParcel request{ctx.ReadBuffer()};
-            IGBPConnectResponseParcel response{1280, 720};
+            IGBPConnectResponseParcel response{
+                static_cast<u32>(static_cast<u32>(DisplayResolution::UndockedWidth) *
+                                 Settings::values.resolution_factor),
+                static_cast<u32>(static_cast<u32>(DisplayResolution::UndockedHeight) *
+                                 Settings::values.resolution_factor)};
             ctx.WriteBuffer(response.Serialize());
         } else if (transaction == TransactionId::SetPreallocatedBuffer) {
             IGBPSetPreallocatedBufferRequestParcel request{ctx.ReadBuffer()};
@@ -692,11 +696,15 @@ private:
         rb.Push(RESULT_SUCCESS);
 
         if (Settings::values.use_docked_mode) {
-            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth));
-            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight));
+            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth) *
+                    static_cast<u32>(Settings::values.resolution_factor));
+            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight) *
+                    static_cast<u32>(Settings::values.resolution_factor));
         } else {
-            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedWidth));
-            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedHeight));
+            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedWidth) *
+                    static_cast<u32>(Settings::values.resolution_factor));
+            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedHeight) *
+                    static_cast<u32>(Settings::values.resolution_factor));
         }
 
         rb.PushRaw<float>(60.0f);
@@ -901,11 +909,15 @@ private:
         rb.Push(RESULT_SUCCESS);
 
         if (Settings::values.use_docked_mode) {
-            rb.Push(static_cast<u64>(DisplayResolution::DockedWidth));
-            rb.Push(static_cast<u64>(DisplayResolution::DockedHeight));
+            rb.Push(static_cast<u64>(DisplayResolution::DockedWidth) *
+                    static_cast<u32>(Settings::values.resolution_factor));
+            rb.Push(static_cast<u64>(DisplayResolution::DockedHeight) *
+                    static_cast<u32>(Settings::values.resolution_factor));
         } else {
-            rb.Push(static_cast<u64>(DisplayResolution::UndockedWidth));
-            rb.Push(static_cast<u64>(DisplayResolution::UndockedHeight));
+            rb.Push(static_cast<u64>(DisplayResolution::UndockedWidth) *
+                    static_cast<u32>(Settings::values.resolution_factor));
+            rb.Push(static_cast<u64>(DisplayResolution::UndockedHeight) *
+                    static_cast<u32>(Settings::values.resolution_factor));
         }
     }
 
@@ -922,6 +934,8 @@ private:
     void ListDisplays(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
         DisplayInfo display_info;
+        display_info.width *= static_cast<u64>(Settings::values.resolution_factor);
+        display_info.height *= static_cast<u64>(Settings::values.resolution_factor);
         ctx.WriteBuffer(&display_info, sizeof(DisplayInfo));
         IPC::ResponseBuilder rb{ctx, 4};
         rb.Push(RESULT_SUCCESS);

src/core/settings.h

@@ -403,6 +403,7 @@ struct Values {
     bool use_gdbstub;
     u16 gdbstub_port;
     std::string program_args;
+    bool dump_exefs;
     bool dump_nso;
 
     // WebService

src/video_core/CMakeLists.txt

@@ -21,6 +21,8 @@ add_library(video_core STATIC
     macro_interpreter.h
     memory_manager.cpp
     memory_manager.h
+    morton.cpp
+    morton.h
     rasterizer_cache.cpp
     rasterizer_cache.h
     rasterizer_interface.h
@@ -62,7 +64,6 @@ add_library(video_core STATIC
     textures/decoders.cpp
     textures/decoders.h
     textures/texture.h
-    utils.h
     video_core.cpp
     video_core.h
 )

src/video_core/command_processor.cpp

@@ -34,6 +34,9 @@ MICROPROFILE_DEFINE(ProcessCommandLists, "GPU", "Execute command buffer", MP_RGB
 void GPU::ProcessCommandLists(const std::vector<CommandListHeader>& commands) {
     MICROPROFILE_SCOPE(ProcessCommandLists);
 
+    // On entering GPU code, assume all memory may be touched by the ARM core.
+    maxwell_3d->dirty_flags.OnMemoryWrite();
+
     auto WriteReg = [this](u32 method, u32 subchannel, u32 value, u32 remaining_params) {
         LOG_TRACE(HW_GPU,
                   "Processing method {:08X} on subchannel {} value "

src/video_core/engines/fermi_2d.cpp

@@ -2,8 +2,10 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include "core/core.h"
 #include "core/memory.h"
 #include "video_core/engines/fermi_2d.h"
+#include "video_core/engines/maxwell_3d.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/textures/decoders.h"
 
@@ -47,6 +49,9 @@ void Fermi2D::HandleSurfaceCopy() {
     u32 dst_bytes_per_pixel = RenderTargetBytesPerPixel(regs.dst.format);
 
     if (!rasterizer.AccelerateSurfaceCopy(regs.src, regs.dst)) {
+        // All copies here update the main memory, so mark all rasterizer states as invalid.
+        Core::System::GetInstance().GPU().Maxwell3D().dirty_flags.OnMemoryWrite();
+
         rasterizer.FlushRegion(source_cpu, src_bytes_per_pixel * regs.src.width * regs.src.height);
         // We have to invalidate the destination region to evict any outdated surfaces from the
         // cache. We do this before actually writing the new data because the destination address

src/video_core/engines/kepler_memory.cpp

@@ -3,8 +3,10 @@
 // Refer to the license.txt file included.
 
 #include "common/logging/log.h"
+#include "core/core.h"
 #include "core/memory.h"
 #include "video_core/engines/kepler_memory.h"
+#include "video_core/engines/maxwell_3d.h"
 #include "video_core/rasterizer_interface.h"
 
 namespace Tegra::Engines {
@@ -47,6 +49,7 @@ void KeplerMemory::ProcessData(u32 data) {
     rasterizer.InvalidateRegion(dest_address, sizeof(u32));
 
     Memory::Write32(dest_address, data);
+    Core::System::GetInstance().GPU().Maxwell3D().dirty_flags.OnMemoryWrite();
 
     state.write_offset++;
 }

src/video_core/engines/maxwell_3d.cpp

@@ -69,6 +69,15 @@ void Maxwell3D::InitializeRegisterDefaults() {
     // TODO(Rodrigo): Most games do not set a point size. I think this is a case of a
     // register carrying a default value. Assume it's OpenGL's default (1).
     regs.point_size = 1.0f;
+
+    // TODO(bunnei): Some games do not initialize the color masks (e.g. Sonic Mania). Assuming a
+    // default of enabled fixes rendering here.
+    for (std::size_t color_mask = 0; color_mask < Regs::NumRenderTargets; color_mask++) {
+        regs.color_mask[color_mask].R.Assign(1);
+        regs.color_mask[color_mask].G.Assign(1);
+        regs.color_mask[color_mask].B.Assign(1);
+        regs.color_mask[color_mask].A.Assign(1);
+    }
 }
 
 void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
@@ -126,10 +135,24 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
 
     if (regs.reg_array[method] != value) {
         regs.reg_array[method] = value;
+        // Vertex format
         if (method >= MAXWELL3D_REG_INDEX(vertex_attrib_format) &&
             method < MAXWELL3D_REG_INDEX(vertex_attrib_format) + regs.vertex_attrib_format.size()) {
             dirty_flags.vertex_attrib_format = true;
         }
+
+        // Vertex buffer
+        if (method >= MAXWELL3D_REG_INDEX(vertex_array) &&
+            method < MAXWELL3D_REG_INDEX(vertex_array) + 4 * 32) {
+            dirty_flags.vertex_array |= 1u << ((method - MAXWELL3D_REG_INDEX(vertex_array)) >> 2);
+        } else if (method >= MAXWELL3D_REG_INDEX(vertex_array_limit) &&
+                   method < MAXWELL3D_REG_INDEX(vertex_array_limit) + 2 * 32) {
+            dirty_flags.vertex_array |=
+                1u << ((method - MAXWELL3D_REG_INDEX(vertex_array_limit)) >> 1);
+        } else if (method >= MAXWELL3D_REG_INDEX(instanced_arrays) &&
+                   method < MAXWELL3D_REG_INDEX(instanced_arrays) + 32) {
+            dirty_flags.vertex_array |= 1u << (method - MAXWELL3D_REG_INDEX(instanced_arrays));
+        }
     }
 
     switch (method) {
@@ -261,6 +284,7 @@ void Maxwell3D::ProcessQueryGet() {
             query_result.timestamp = CoreTiming::GetTicks();
             Memory::WriteBlock(*address, &query_result, sizeof(query_result));
         }
+        dirty_flags.OnMemoryWrite();
         break;
     }
     default:
@@ -337,6 +361,7 @@ void Maxwell3D::ProcessCBData(u32 value) {
         memory_manager.GpuToCpuAddress(buffer_address + regs.const_buffer.cb_pos);
 
     Memory::Write32(*address, value);
+    dirty_flags.OnMemoryWrite();
 
     // Increment the current buffer position.
     regs.const_buffer.cb_pos = regs.const_buffer.cb_pos + 4;

src/video_core/engines/maxwell_3d.h

@@ -389,6 +389,13 @@ public:
                 ReverseSubtract = 3,
                 Min = 4,
                 Max = 5,
+
+                // These values are used by Nouveau and some games.
+                AddGL = 0x8006,
+                SubtractGL = 0x8007,
+                ReverseSubtractGL = 0x8008,
+                MinGL = 0x800a,
+                MaxGL = 0x800b
             };
 
             enum class Factor : u32 {
@@ -583,10 +590,18 @@ public:
 
                 float clear_color[4];
                 float clear_depth;
+
                 INSERT_PADDING_WORDS(0x3);
+
                 s32 clear_stencil;
 
-                INSERT_PADDING_WORDS(0x17);
+                INSERT_PADDING_WORDS(0x7);
+
+                u32 polygon_offset_point_enable;
+                u32 polygon_offset_line_enable;
+                u32 polygon_offset_fill_enable;
+
+                INSERT_PADDING_WORDS(0xD);
 
                 std::array<ScissorTest, NumViewports> scissor_test;
 
@@ -624,7 +639,16 @@ public:
                     }
                 } zeta;
 
-                INSERT_PADDING_WORDS(0x5B);
+                INSERT_PADDING_WORDS(0x41);
+
+                union {
+                    BitField<0, 4, u32> stencil;
+                    BitField<4, 4, u32> unknown;
+                    BitField<8, 4, u32> scissor;
+                    BitField<12, 4, u32> viewport;
+                } clear_flags;
+
+                INSERT_PADDING_WORDS(0x19);
 
                 std::array<VertexAttribute, NumVertexAttributes> vertex_attrib_format;
 
@@ -712,6 +736,7 @@ public:
                 u32 frag_color_clamp;
 
                 union {
+                    BitField<0, 1, u32> y_negate;
                     BitField<4, 1, u32> triangle_rast_flip;
                 } screen_y_control;
 
@@ -745,7 +770,11 @@ public:
                     }
                 } tsc;
 
-                INSERT_PADDING_WORDS(0x3);
+                INSERT_PADDING_WORDS(0x1);
+
+                float polygon_offset_factor;
+
+                INSERT_PADDING_WORDS(0x1);
 
                 struct {
                     u32 tic_address_high;
@@ -770,7 +799,9 @@ public:
 
                 u32 framebuffer_srgb;
 
-                INSERT_PADDING_WORDS(0x12);
+                float polygon_offset_units;
+
+                INSERT_PADDING_WORDS(0x11);
 
                 union {
                     BitField<2, 1, u32> coord_origin;
@@ -847,7 +878,9 @@ public:
 
                 INSERT_PADDING_WORDS(0x7);
 
-                INSERT_PADDING_WORDS(0x20);
+                INSERT_PADDING_WORDS(0x1F);
+
+                float polygon_offset_clamp;
 
                 struct {
                     u32 is_instanced[NumVertexArrays];
@@ -863,7 +896,13 @@ public:
 
                 Cull cull;
 
-                INSERT_PADDING_WORDS(0x28);
+                u32 pixel_center_integer;
+
+                INSERT_PADDING_WORDS(0x1);
+
+                u32 viewport_transform_enabled;
+
+                INSERT_PADDING_WORDS(0x25);
 
                 struct {
                     u32 enable;
@@ -1028,6 +1067,11 @@ public:
 
     struct DirtyFlags {
         bool vertex_attrib_format = true;
+        u32 vertex_array = 0xFFFFFFFF;
+
+        void OnMemoryWrite() {
+            vertex_array = 0xFFFFFFFF;
+        }
     };
 
     DirtyFlags dirty_flags;
@@ -1120,6 +1164,9 @@ ASSERT_REG_POSITION(vertex_buffer, 0x35D);
 ASSERT_REG_POSITION(clear_color[0], 0x360);
 ASSERT_REG_POSITION(clear_depth, 0x364);
 ASSERT_REG_POSITION(clear_stencil, 0x368);
+ASSERT_REG_POSITION(polygon_offset_point_enable, 0x370);
+ASSERT_REG_POSITION(polygon_offset_line_enable, 0x371);
+ASSERT_REG_POSITION(polygon_offset_fill_enable, 0x372);
 ASSERT_REG_POSITION(scissor_test, 0x380);
 ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5);
 ASSERT_REG_POSITION(stencil_back_mask, 0x3D6);
@@ -1127,6 +1174,7 @@ ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7);
 ASSERT_REG_POSITION(color_mask_common, 0x3E4);
 ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB);
 ASSERT_REG_POSITION(zeta, 0x3F8);
+ASSERT_REG_POSITION(clear_flags, 0x43E);
 ASSERT_REG_POSITION(vertex_attrib_format, 0x458);
 ASSERT_REG_POSITION(rt_control, 0x487);
 ASSERT_REG_POSITION(zeta_width, 0x48a);
@@ -1157,6 +1205,7 @@ ASSERT_REG_POSITION(point_size, 0x546);
 ASSERT_REG_POSITION(zeta_enable, 0x54E);
 ASSERT_REG_POSITION(multisample_control, 0x54F);
 ASSERT_REG_POSITION(tsc, 0x557);
+ASSERT_REG_POSITION(polygon_offset_factor, 0x55b);
 ASSERT_REG_POSITION(tic, 0x55D);
 ASSERT_REG_POSITION(stencil_two_side_enable, 0x565);
 ASSERT_REG_POSITION(stencil_back_op_fail, 0x566);
@@ -1164,13 +1213,17 @@ ASSERT_REG_POSITION(stencil_back_op_zfail, 0x567);
 ASSERT_REG_POSITION(stencil_back_op_zpass, 0x568);
 ASSERT_REG_POSITION(stencil_back_func_func, 0x569);
 ASSERT_REG_POSITION(framebuffer_srgb, 0x56E);
+ASSERT_REG_POSITION(polygon_offset_units, 0x56F);
 ASSERT_REG_POSITION(point_coord_replace, 0x581);
 ASSERT_REG_POSITION(code_address, 0x582);
 ASSERT_REG_POSITION(draw, 0x585);
 ASSERT_REG_POSITION(primitive_restart, 0x591);
 ASSERT_REG_POSITION(index_array, 0x5F2);
+ASSERT_REG_POSITION(polygon_offset_clamp, 0x61F);
 ASSERT_REG_POSITION(instanced_arrays, 0x620);
 ASSERT_REG_POSITION(cull, 0x646);
+ASSERT_REG_POSITION(pixel_center_integer, 0x649);
+ASSERT_REG_POSITION(viewport_transform_enabled, 0x64B);
 ASSERT_REG_POSITION(logic_op, 0x671);
 ASSERT_REG_POSITION(clear_buffers, 0x674);
 ASSERT_REG_POSITION(color_mask, 0x680);

src/video_core/engines/maxwell_dma.cpp

@@ -2,7 +2,9 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include "core/core.h"
 #include "core/memory.h"
+#include "video_core/engines/maxwell_3d.h"
 #include "video_core/engines/maxwell_dma.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/textures/decoders.h"
@@ -54,6 +56,9 @@ void MaxwellDMA::HandleCopy() {
         return;
     }
 
+    // All copies here update the main memory, so mark all rasterizer states as invalid.
+    Core::System::GetInstance().GPU().Maxwell3D().dirty_flags.OnMemoryWrite();
+
     if (regs.exec.is_dst_linear && regs.exec.is_src_linear) {
         // When the enable_2d bit is disabled, the copy is performed as if we were copying a 1D
         // buffer of length `x_count`, otherwise we copy a 2D image of dimensions (x_count,

src/video_core/engines/shader_bytecode.h

@@ -82,6 +82,8 @@ union Attribute {
         Position = 7,
         Attribute_0 = 8,
         Attribute_31 = 39,
+        ClipDistances0123 = 44,
+        ClipDistances4567 = 45,
         PointCoord = 46,
         // This attribute contains a tuple of (~, ~, InstanceId, VertexId) when inside a vertex
         // shader, and a tuple of (TessCoord.x, TessCoord.y, TessCoord.z, ~) when inside a Tess Eval
@@ -153,6 +155,7 @@ enum class PredCondition : u64 {
     NotEqual = 5,
     GreaterEqual = 6,
     LessThanWithNan = 9,
+    LessEqualWithNan = 11,
     GreaterThanWithNan = 12,
     NotEqualWithNan = 13,
     GreaterEqualWithNan = 14,
@@ -261,7 +264,7 @@ enum class FlowCondition : u64 {
     Fcsm_Tr = 0x1C, // TODO(bunnei): What is this used for?
 };
 
-enum class ControlCode : u64 {
+enum class ConditionCode : u64 {
     F = 0,
     LT = 1,
     EQ = 2,
@@ -365,6 +368,11 @@ enum class HalfPrecision : u64 {
     FMZ = 2,
 };
 
+enum class R2pMode : u64 {
+    Pr = 0,
+    Cc = 1,
+};
+
 enum class IpaInterpMode : u64 {
     Linear = 0,
     Perspective = 1,
@@ -569,7 +577,6 @@ union Instruction {
         BitField<39, 2, u64> tab5cb8_2;
         BitField<41, 3, u64> tab5c68_1;
         BitField<44, 2, u64> tab5c68_0;
-        BitField<47, 1, u64> cc;
         BitField<48, 1, u64> negate_b;
     } fmul;
 
@@ -831,7 +838,7 @@ union Instruction {
     union {
         BitField<0, 3, u64> pred0;
         BitField<3, 3, u64> pred3;
-        BitField<8, 5, ControlCode> cc; // flag in cc
+        BitField<8, 5, ConditionCode> cc; // flag in cc
         BitField<39, 3, u64> pred39;
         BitField<42, 1, u64> neg_pred39;
         BitField<45, 4, PredOperation> op; // op with pred39
@@ -854,6 +861,12 @@ union Instruction {
         BitField<39, 3, u64> pred39;
     } hsetp2;
 
+    union {
+        BitField<40, 1, R2pMode> mode;
+        BitField<41, 2, u64> byte;
+        BitField<20, 7, u64> immediate_mask;
+    } r2p;
+
     union {
         BitField<39, 3, u64> pred39;
         BitField<42, 1, u64> neg_pred;
@@ -1235,7 +1248,7 @@ union Instruction {
     BitField<60, 1, u64> is_b_gpr;
     BitField<59, 1, u64> is_c_gpr;
     BitField<20, 24, s64> smem_imm;
-    BitField<0, 5, ControlCode> flow_control_code;
+    BitField<0, 5, ConditionCode> flow_condition_code;
 
     Attribute attribute;
     Sampler sampler;
@@ -1256,6 +1269,7 @@ public:
         BFE_C,
         BFE_R,
         BFE_IMM,
+        BFI_IMM_R,
         BRA,
         PBK,
         LD_A,
@@ -1381,6 +1395,7 @@ public:
         PSETP,
         PSET,
         CSETP,
+        R2P_IMM,
         XMAD_IMM,
         XMAD_CR,
         XMAD_RC,
@@ -1396,6 +1411,7 @@ public:
         ArithmeticHalf,
         ArithmeticHalfImmediate,
         Bfe,
+        Bfi,
         Shift,
         Ffma,
         Hfma2,
@@ -1410,6 +1426,7 @@ public:
         HalfSetPredicate,
         PredicateSetPredicate,
         PredicateSetRegister,
+        RegisterSetPredicate,
         Conversion,
         Xmad,
         Unknown,
@@ -1613,6 +1630,7 @@ private:
             INST("0100110000000---", Id::BFE_C, Type::Bfe, "BFE_C"),
             INST("0101110000000---", Id::BFE_R, Type::Bfe, "BFE_R"),
             INST("0011100-00000---", Id::BFE_IMM, Type::Bfe, "BFE_IMM"),
+            INST("0011011-11110---", Id::BFI_IMM_R, Type::Bfi, "BFI_IMM_R"),
             INST("0100110001000---", Id::LOP_C, Type::ArithmeticInteger, "LOP_C"),
             INST("0101110001000---", Id::LOP_R, Type::ArithmeticInteger, "LOP_R"),
             INST("0011100001000---", Id::LOP_IMM, Type::ArithmeticInteger, "LOP_IMM"),
@@ -1647,6 +1665,7 @@ private:
             INST("0101000010001---", Id::PSET, Type::PredicateSetRegister, "PSET"),
             INST("0101000010010---", Id::PSETP, Type::PredicateSetPredicate, "PSETP"),
             INST("010100001010----", Id::CSETP, Type::PredicateSetPredicate, "CSETP"),
+            INST("0011100-11110---", Id::R2P_IMM, Type::RegisterSetPredicate, "R2P_IMM"),
             INST("0011011-00------", Id::XMAD_IMM, Type::Xmad, "XMAD_IMM"),
             INST("0100111---------", Id::XMAD_CR, Type::Xmad, "XMAD_CR"),
             INST("010100010-------", Id::XMAD_RC, Type::Xmad, "XMAD_RC"),

src/video_core/engines/shader_header.h

@@ -62,7 +62,16 @@ struct Header {
             INSERT_PADDING_BYTES(1);  // ImapSystemValuesB
             INSERT_PADDING_BYTES(16); // ImapGenericVector[32]
             INSERT_PADDING_BYTES(2);  // ImapColor
-            INSERT_PADDING_BYTES(2);  // ImapSystemValuesC
+            union {
+                BitField<0, 8, u16> clip_distances;
+                BitField<8, 1, u16> point_sprite_s;
+                BitField<9, 1, u16> point_sprite_t;
+                BitField<10, 1, u16> fog_coordinate;
+                BitField<12, 1, u16> tessellation_eval_point_u;
+                BitField<13, 1, u16> tessellation_eval_point_v;
+                BitField<14, 1, u16> instance_id;
+                BitField<15, 1, u16> vertex_id;
+            };
             INSERT_PADDING_BYTES(5);  // ImapFixedFncTexture[10]
             INSERT_PADDING_BYTES(1);  // ImapReserved
             INSERT_PADDING_BYTES(3);  // OmapSystemValuesA

src/video_core/macro_interpreter.cpp

@@ -35,6 +35,7 @@ void MacroInterpreter::Reset() {
     // The next parameter index starts at 1, because $r1 already has the value of the first
     // parameter.
     next_parameter_index = 1;
+    carry_flag = false;
 }
 
 bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
@@ -135,14 +136,28 @@ MacroInterpreter::Opcode MacroInterpreter::GetOpcode(u32 offset) const {
     return {macro_memory[offset + pc / sizeof(u32)]};
 }
 
-u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const {
+u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) {
     switch (operation) {
-    case ALUOperation::Add:
-        return src_a + src_b;
-    // TODO(Subv): Implement AddWithCarry
-    case ALUOperation::Subtract:
-        return src_a - src_b;
-    // TODO(Subv): Implement SubtractWithBorrow
+    case ALUOperation::Add: {
+        const u64 result{static_cast<u64>(src_a) + src_b};
+        carry_flag = result > 0xffffffff;
+        return static_cast<u32>(result);
+    }
+    case ALUOperation::AddWithCarry: {
+        const u64 result{static_cast<u64>(src_a) + src_b + (carry_flag ? 1ULL : 0ULL)};
+        carry_flag = result > 0xffffffff;
+        return static_cast<u32>(result);
+    }
+    case ALUOperation::Subtract: {
+        const u64 result{static_cast<u64>(src_a) - src_b};
+        carry_flag = result < 0x100000000;
+        return static_cast<u32>(result);
+    }
+    case ALUOperation::SubtractWithBorrow: {
+        const u64 result{static_cast<u64>(src_a) - src_b - (carry_flag ? 0ULL : 1ULL)};
+        carry_flag = result < 0x100000000;
+        return static_cast<u32>(result);
+    }
     case ALUOperation::Xor:
         return src_a ^ src_b;
     case ALUOperation::Or:

src/video_core/macro_interpreter.h

@@ -117,7 +117,7 @@ private:
     bool Step(u32 offset, bool is_delay_slot);
 
     /// Calculates the result of an ALU operation. src_a OP src_b;
-    u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const;
+    u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b);
 
     /// Performs the result operation on the input result and stores it in the specified register
     /// (if necessary).
@@ -165,5 +165,7 @@ private:
     std::vector<u32> parameters;
     /// Index of the next parameter that will be fetched by the 'parm' instruction.
     u32 next_parameter_index = 0;
+
+    bool carry_flag{};
 };
 } // namespace Tegra

src/video_core/memory_manager.cpp

@@ -9,6 +9,13 @@
 
 namespace Tegra {
 
+MemoryManager::MemoryManager() {
+    // Mark the first page as reserved, so that 0 is not a valid GPUVAddr. Otherwise, games might
+    // try to use 0 as a valid address, which is also used to mean nullptr. This fixes a bug with
+    // Undertale using 0 for a render target.
+    PageSlot(0) = static_cast<u64>(PageStatus::Reserved);
+}
+
 GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
     const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, align, PageStatus::Unmapped)};
 

src/video_core/memory_manager.h

@@ -18,7 +18,7 @@ using GPUVAddr = u64;
 
 class MemoryManager final {
 public:
-    MemoryManager() = default;
+    MemoryManager();
 
     GPUVAddr AllocateSpace(u64 size, u64 align);
     GPUVAddr AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align);
@@ -37,6 +37,7 @@ private:
     enum class PageStatus : u64 {
         Unmapped = 0xFFFFFFFFFFFFFFFFULL,
         Allocated = 0xFFFFFFFFFFFFFFFEULL,
+        Reserved = 0xFFFFFFFFFFFFFFFDULL,
     };
 
     std::optional<GPUVAddr> FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,

src/video_core/morton.cpp

@@ -0,0 +1,353 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <array>
+#include <cstring>
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "core/memory.h"
+#include "video_core/morton.h"
+#include "video_core/surface.h"
+#include "video_core/textures/decoders.h"
+
+namespace VideoCore {
+
+using Surface::GetBytesPerPixel;
+using Surface::PixelFormat;
+
+using MortonCopyFn = void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr);
+using ConversionArray = std::array<MortonCopyFn, Surface::MaxPixelFormat>;
+
+template <bool morton_to_linear, PixelFormat format>
+static void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth,
+                       u8* buffer, std::size_t buffer_size, VAddr addr) {
+    constexpr u32 bytes_per_pixel = GetBytesPerPixel(format);
+
+    // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
+    // pixel values.
+    const u32 tile_size_x{GetDefaultBlockWidth(format)};
+    const u32 tile_size_y{GetDefaultBlockHeight(format)};
+
+    if constexpr (morton_to_linear) {
+        Tegra::Texture::UnswizzleTexture(buffer, addr, tile_size_x, tile_size_y, bytes_per_pixel,
+                                         stride, height, depth, block_height, block_depth);
+    } else {
+        Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x,
+                                         (height + tile_size_y - 1) / tile_size_y, depth,
+                                         bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr),
+                                         buffer, false, block_height, block_depth);
+    }
+}
+
+static constexpr ConversionArray morton_to_linear_fns = {
+    // clang-format off
+        MortonCopy<true, PixelFormat::ABGR8U>,
+        MortonCopy<true, PixelFormat::ABGR8S>,
+        MortonCopy<true, PixelFormat::ABGR8UI>,
+        MortonCopy<true, PixelFormat::B5G6R5U>,
+        MortonCopy<true, PixelFormat::A2B10G10R10U>,
+        MortonCopy<true, PixelFormat::A1B5G5R5U>,
+        MortonCopy<true, PixelFormat::R8U>,
+        MortonCopy<true, PixelFormat::R8UI>,
+        MortonCopy<true, PixelFormat::RGBA16F>,
+        MortonCopy<true, PixelFormat::RGBA16U>,
+        MortonCopy<true, PixelFormat::RGBA16UI>,
+        MortonCopy<true, PixelFormat::R11FG11FB10F>,
+        MortonCopy<true, PixelFormat::RGBA32UI>,
+        MortonCopy<true, PixelFormat::DXT1>,
+        MortonCopy<true, PixelFormat::DXT23>,
+        MortonCopy<true, PixelFormat::DXT45>,
+        MortonCopy<true, PixelFormat::DXN1>,
+        MortonCopy<true, PixelFormat::DXN2UNORM>,
+        MortonCopy<true, PixelFormat::DXN2SNORM>,
+        MortonCopy<true, PixelFormat::BC7U>,
+        MortonCopy<true, PixelFormat::BC6H_UF16>,
+        MortonCopy<true, PixelFormat::BC6H_SF16>,
+        MortonCopy<true, PixelFormat::ASTC_2D_4X4>,
+        MortonCopy<true, PixelFormat::G8R8U>,
+        MortonCopy<true, PixelFormat::G8R8S>,
+        MortonCopy<true, PixelFormat::BGRA8>,
+        MortonCopy<true, PixelFormat::RGBA32F>,
+        MortonCopy<true, PixelFormat::RG32F>,
+        MortonCopy<true, PixelFormat::R32F>,
+        MortonCopy<true, PixelFormat::R16F>,
+        MortonCopy<true, PixelFormat::R16U>,
+        MortonCopy<true, PixelFormat::R16S>,
+        MortonCopy<true, PixelFormat::R16UI>,
+        MortonCopy<true, PixelFormat::R16I>,
+        MortonCopy<true, PixelFormat::RG16>,
+        MortonCopy<true, PixelFormat::RG16F>,
+        MortonCopy<true, PixelFormat::RG16UI>,
+        MortonCopy<true, PixelFormat::RG16I>,
+        MortonCopy<true, PixelFormat::RG16S>,
+        MortonCopy<true, PixelFormat::RGB32F>,
+        MortonCopy<true, PixelFormat::RGBA8_SRGB>,
+        MortonCopy<true, PixelFormat::RG8U>,
+        MortonCopy<true, PixelFormat::RG8S>,
+        MortonCopy<true, PixelFormat::RG32UI>,
+        MortonCopy<true, PixelFormat::R32UI>,
+        MortonCopy<true, PixelFormat::ASTC_2D_8X8>,
+        MortonCopy<true, PixelFormat::ASTC_2D_8X5>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X4>,
+        MortonCopy<true, PixelFormat::BGRA8_SRGB>,
+        MortonCopy<true, PixelFormat::DXT1_SRGB>,
+        MortonCopy<true, PixelFormat::DXT23_SRGB>,
+        MortonCopy<true, PixelFormat::DXT45_SRGB>,
+        MortonCopy<true, PixelFormat::BC7U_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_4X4_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X5>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>,
+        MortonCopy<true, PixelFormat::ASTC_2D_10X8>,
+        MortonCopy<true, PixelFormat::ASTC_2D_10X8_SRGB>,
+        MortonCopy<true, PixelFormat::Z32F>,
+        MortonCopy<true, PixelFormat::Z16>,
+        MortonCopy<true, PixelFormat::Z24S8>,
+        MortonCopy<true, PixelFormat::S8Z24>,
+        MortonCopy<true, PixelFormat::Z32FS8>,
+    // clang-format on
+};
+
+static constexpr ConversionArray linear_to_morton_fns = {
+    // clang-format off
+        MortonCopy<false, PixelFormat::ABGR8U>,
+        MortonCopy<false, PixelFormat::ABGR8S>,
+        MortonCopy<false, PixelFormat::ABGR8UI>,
+        MortonCopy<false, PixelFormat::B5G6R5U>,
+        MortonCopy<false, PixelFormat::A2B10G10R10U>,
+        MortonCopy<false, PixelFormat::A1B5G5R5U>,
+        MortonCopy<false, PixelFormat::R8U>,
+        MortonCopy<false, PixelFormat::R8UI>,
+        MortonCopy<false, PixelFormat::RGBA16F>,
+        MortonCopy<false, PixelFormat::RGBA16U>,
+        MortonCopy<false, PixelFormat::RGBA16UI>,
+        MortonCopy<false, PixelFormat::R11FG11FB10F>,
+        MortonCopy<false, PixelFormat::RGBA32UI>,
+        MortonCopy<false, PixelFormat::DXT1>,
+        MortonCopy<false, PixelFormat::DXT23>,
+        MortonCopy<false, PixelFormat::DXT45>,
+        MortonCopy<false, PixelFormat::DXN1>,
+        MortonCopy<false, PixelFormat::DXN2UNORM>,
+        MortonCopy<false, PixelFormat::DXN2SNORM>,
+        MortonCopy<false, PixelFormat::BC7U>,
+        MortonCopy<false, PixelFormat::BC6H_UF16>,
+        MortonCopy<false, PixelFormat::BC6H_SF16>,
+        // TODO(Subv): Swizzling ASTC formats are not supported
+        nullptr,
+        MortonCopy<false, PixelFormat::G8R8U>,
+        MortonCopy<false, PixelFormat::G8R8S>,
+        MortonCopy<false, PixelFormat::BGRA8>,
+        MortonCopy<false, PixelFormat::RGBA32F>,
+        MortonCopy<false, PixelFormat::RG32F>,
+        MortonCopy<false, PixelFormat::R32F>,
+        MortonCopy<false, PixelFormat::R16F>,
+        MortonCopy<false, PixelFormat::R16U>,
+        MortonCopy<false, PixelFormat::R16S>,
+        MortonCopy<false, PixelFormat::R16UI>,
+        MortonCopy<false, PixelFormat::R16I>,
+        MortonCopy<false, PixelFormat::RG16>,
+        MortonCopy<false, PixelFormat::RG16F>,
+        MortonCopy<false, PixelFormat::RG16UI>,
+        MortonCopy<false, PixelFormat::RG16I>,
+        MortonCopy<false, PixelFormat::RG16S>,
+        MortonCopy<false, PixelFormat::RGB32F>,
+        MortonCopy<false, PixelFormat::RGBA8_SRGB>,
+        MortonCopy<false, PixelFormat::RG8U>,
+        MortonCopy<false, PixelFormat::RG8S>,
+        MortonCopy<false, PixelFormat::RG32UI>,
+        MortonCopy<false, PixelFormat::R32UI>,
+        nullptr,
+        nullptr,
+        nullptr,
+        MortonCopy<false, PixelFormat::BGRA8_SRGB>,
+        MortonCopy<false, PixelFormat::DXT1_SRGB>,
+        MortonCopy<false, PixelFormat::DXT23_SRGB>,
+        MortonCopy<false, PixelFormat::DXT45_SRGB>,
+        MortonCopy<false, PixelFormat::BC7U_SRGB>,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
+        nullptr,
+        MortonCopy<false, PixelFormat::Z32F>,
+        MortonCopy<false, PixelFormat::Z16>,
+        MortonCopy<false, PixelFormat::Z24S8>,
+        MortonCopy<false, PixelFormat::S8Z24>,
+        MortonCopy<false, PixelFormat::Z32FS8>,
+    // clang-format on
+};
+
+constexpr MortonCopyFn GetSwizzleFunction(MortonSwizzleMode mode, Surface::PixelFormat format) {
+    switch (mode) {
+    case MortonSwizzleMode::MortonToLinear:
+        return morton_to_linear_fns[static_cast<std::size_t>(format)];
+    case MortonSwizzleMode::LinearToMorton:
+        return linear_to_morton_fns[static_cast<std::size_t>(format)];
+    }
+}
+
+/// 8x8 Z-Order coordinate from 2D coordinates
+static u32 MortonInterleave(u32 x, u32 y) {
+    static const u32 xlut[] = {0x00, 0x01, 0x04, 0x05, 0x10, 0x11, 0x14, 0x15};
+    static const u32 ylut[] = {0x00, 0x02, 0x08, 0x0a, 0x20, 0x22, 0x28, 0x2a};
+    return xlut[x % 8] + ylut[y % 8];
+}
+
+/// Calculates the offset of the position of the pixel in Morton order
+static u32 GetMortonOffset(u32 x, u32 y, u32 bytes_per_pixel) {
+    // Images are split into 8x8 tiles. Each tile is composed of four 4x4 subtiles each
+    // of which is composed of four 2x2 subtiles each of which is composed of four texels.
+    // Each structure is embedded into the next-bigger one in a diagonal pattern, e.g.
+    // texels are laid out in a 2x2 subtile like this:
+    // 2 3
+    // 0 1
+    //
+    // The full 8x8 tile has the texels arranged like this:
+    //
+    // 42 43 46 47 58 59 62 63
+    // 40 41 44 45 56 57 60 61
+    // 34 35 38 39 50 51 54 55
+    // 32 33 36 37 48 49 52 53
+    // 10 11 14 15 26 27 30 31
+    // 08 09 12 13 24 25 28 29
+    // 02 03 06 07 18 19 22 23
+    // 00 01 04 05 16 17 20 21
+    //
+    // This pattern is what's called Z-order curve, or Morton order.
+
+    const unsigned int block_height = 8;
+    const unsigned int coarse_x = x & ~7;
+
+    u32 i = MortonInterleave(x, y);
+
+    const unsigned int offset = coarse_x * block_height;
+
+    return (i + offset) * bytes_per_pixel;
+}
+
+static u32 MortonInterleave128(u32 x, u32 y) {
+    // 128x128 Z-Order coordinate from 2D coordinates
+    static constexpr u32 xlut[] = {
+        0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042,
+        0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809,
+        0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000,
+        0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043,
+        0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a,
+        0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001,
+        0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048,
+        0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b,
+        0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002,
+        0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049,
+        0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840,
+        0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003,
+        0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a,
+        0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841,
+        0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008,
+        0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b,
+        0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842,
+        0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009,
+        0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800,
+        0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843,
+        0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a,
+        0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801,
+        0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848,
+        0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b,
+        0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802,
+        0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849,
+        0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040,
+        0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803,
+        0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a,
+        0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041,
+        0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808,
+        0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b,
+        0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042,
+        0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809,
+        0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b,
+    };
+    static constexpr u32 ylut[] = {
+        0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090,
+        0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124,
+        0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200,
+        0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294,
+        0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330,
+        0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404,
+        0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0,
+        0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534,
+        0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610,
+        0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4,
+        0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780,
+        0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014,
+        0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0,
+        0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184,
+        0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220,
+        0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4,
+        0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390,
+        0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424,
+        0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500,
+        0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594,
+        0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630,
+        0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704,
+        0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0,
+        0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034,
+        0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110,
+        0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4,
+        0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280,
+        0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314,
+        0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0,
+        0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484,
+        0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520,
+        0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4,
+        0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690,
+        0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724,
+        0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4,
+    };
+    return xlut[x % 128] + ylut[y % 128];
+}
+
+static u32 GetMortonOffset128(u32 x, u32 y, u32 bytes_per_pixel) {
+    // Calculates the offset of the position of the pixel in Morton order
+    // Framebuffer images are split into 128x128 tiles.
+
+    constexpr u32 block_height = 128;
+    const u32 coarse_x = x & ~127;
+
+    const u32 i = MortonInterleave128(x, y);
+
+    const u32 offset = coarse_x * block_height;
+
+    return (i + offset) * bytes_per_pixel;
+}
+
+void MortonSwizzle(MortonSwizzleMode mode, Surface::PixelFormat format, u32 stride,
+                   u32 block_height, u32 height, u32 block_depth, u32 depth, u8* buffer,
+                   std::size_t buffer_size, VAddr addr) {
+
+    GetSwizzleFunction(mode, format)(stride, block_height, height, block_depth, depth, buffer,
+                                     buffer_size, addr);
+}
+
+void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixel, u32 linear_bytes_per_pixel,
+                         u8* morton_data, u8* linear_data, bool morton_to_linear) {
+    u8* data_ptrs[2];
+    for (u32 y = 0; y < height; ++y) {
+        for (u32 x = 0; x < width; ++x) {
+            const u32 coarse_y = y & ~127;
+            const u32 morton_offset =
+                GetMortonOffset128(x, y, bytes_per_pixel) + coarse_y * width * bytes_per_pixel;
+            const u32 linear_pixel_index = (x + y * width) * linear_bytes_per_pixel;
+
+            data_ptrs[morton_to_linear ? 1 : 0] = morton_data + morton_offset;
+            data_ptrs[morton_to_linear ? 0 : 1] = &linear_data[linear_pixel_index];
+
+            std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
+        }
+    }
+}
+
+} // namespace VideoCore

src/video_core/morton.h

@@ -0,0 +1,21 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+#include "video_core/surface.h"
+
+namespace VideoCore {
+
+enum class MortonSwizzleMode { MortonToLinear, LinearToMorton };
+
+void MortonSwizzle(MortonSwizzleMode mode, VideoCore::Surface::PixelFormat format, u32 stride,
+                   u32 block_height, u32 height, u32 block_depth, u32 depth, u8* buffer,
+                   std::size_t buffer_size, VAddr addr);
+
+void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixel, u32 linear_bytes_per_pixel,
+                         u8* morton_data, u8* linear_data, bool morton_to_linear);
+
+} // namespace VideoCore

src/video_core/renderer_opengl/gl_buffer_cache.cpp

@@ -76,7 +76,7 @@ std::tuple<u8*, GLintptr> OGLBufferCache::ReserveMemory(std::size_t size, std::s
     return std::make_tuple(uploaded_ptr, uploaded_offset);
 }
 
-void OGLBufferCache::Map(std::size_t max_size) {
+bool OGLBufferCache::Map(std::size_t max_size) {
     bool invalidate;
     std::tie(buffer_ptr, buffer_offset_base, invalidate) =
         stream_buffer.Map(static_cast<GLsizeiptr>(max_size), 4);
@@ -85,6 +85,7 @@ void OGLBufferCache::Map(std::size_t max_size) {
     if (invalidate) {
         InvalidateAll();
     }
+    return invalidate;
 }
 
 void OGLBufferCache::Unmap() {

src/video_core/renderer_opengl/gl_buffer_cache.h

@@ -50,7 +50,7 @@ public:
     /// Reserves memory to be used by host's CPU. Returns mapped address and offset.
     std::tuple<u8*, GLintptr> ReserveMemory(std::size_t size, std::size_t alignment = 4);
 
-    void Map(std::size_t max_size);
+    bool Map(std::size_t max_size);
     void Unmap();
 
     GLuint GetHandle() const;

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -98,14 +98,9 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
             has_ARB_direct_state_access = true;
         } else if (extension == "GL_ARB_multi_bind") {
             has_ARB_multi_bind = true;
-        } else if (extension == "GL_ARB_separate_shader_objects") {
-            has_ARB_separate_shader_objects = true;
-        } else if (extension == "GL_ARB_vertex_attrib_binding") {
-            has_ARB_vertex_attrib_binding = true;
         }
     }
 
-    ASSERT_MSG(has_ARB_separate_shader_objects, "has_ARB_separate_shader_objects is unsupported");
     OpenGLState::ApplyDefaultState();
 
     // Create render framebuffer
@@ -181,15 +176,25 @@ void RasterizerOpenGL::SetupVertexFormat() {
     }
     state.draw.vertex_array = VAO.handle;
     state.ApplyVertexBufferState();
+
+    // Rebinding the VAO invalidates the vertex buffer bindings.
+    gpu.dirty_flags.vertex_array = 0xFFFFFFFF;
 }
 
 void RasterizerOpenGL::SetupVertexBuffer() {
-    MICROPROFILE_SCOPE(OpenGL_VB);
-    const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+    auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
     const auto& regs = gpu.regs;
 
+    if (!gpu.dirty_flags.vertex_array)
+        return;
+
+    MICROPROFILE_SCOPE(OpenGL_VB);
+
     // Upload all guest vertex arrays sequentially to our buffer
     for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) {
+        if (~gpu.dirty_flags.vertex_array & (1u << index))
+            continue;
+
         const auto& vertex_array = regs.vertex_array[index];
         if (!vertex_array.IsEnabled())
             continue;
@@ -216,6 +221,8 @@ void RasterizerOpenGL::SetupVertexBuffer() {
 
     // Implicit set by glBindVertexBuffer. Stupid glstate handling...
     state.draw.vertex_buffer = buffer_cache.GetHandle();
+
+    gpu.dirty_flags.vertex_array = 0;
 }
 
 DrawParameters RasterizerOpenGL::SetupDraw() {
@@ -542,6 +549,30 @@ void RasterizerOpenGL::Clear() {
         ASSERT_MSG(regs.zeta_enable != 0, "Tried to clear stencil but buffer is not enabled!");
         use_stencil = true;
         clear_state.stencil.test_enabled = true;
+        if (regs.clear_flags.stencil) {
+            // Stencil affects the clear so fill it with the used masks
+            clear_state.stencil.front.test_func = GL_ALWAYS;
+            clear_state.stencil.front.test_mask = regs.stencil_front_func_mask;
+            clear_state.stencil.front.action_stencil_fail = GL_KEEP;
+            clear_state.stencil.front.action_depth_fail = GL_KEEP;
+            clear_state.stencil.front.action_depth_pass = GL_KEEP;
+            clear_state.stencil.front.write_mask = regs.stencil_front_mask;
+            if (regs.stencil_two_side_enable) {
+                clear_state.stencil.back.test_func = GL_ALWAYS;
+                clear_state.stencil.back.test_mask = regs.stencil_back_func_mask;
+                clear_state.stencil.back.action_stencil_fail = GL_KEEP;
+                clear_state.stencil.back.action_depth_fail = GL_KEEP;
+                clear_state.stencil.back.action_depth_pass = GL_KEEP;
+                clear_state.stencil.back.write_mask = regs.stencil_back_mask;
+            } else {
+                clear_state.stencil.back.test_func = GL_ALWAYS;
+                clear_state.stencil.back.test_mask = 0xFFFFFFFF;
+                clear_state.stencil.back.write_mask = 0xFFFFFFFF;
+                clear_state.stencil.back.action_stencil_fail = GL_KEEP;
+                clear_state.stencil.back.action_depth_fail = GL_KEEP;
+                clear_state.stencil.back.action_depth_pass = GL_KEEP;
+            }
+        }
     }
 
     if (!use_color && !use_depth && !use_stencil) {
@@ -553,6 +584,14 @@ void RasterizerOpenGL::Clear() {
 
     ConfigureFramebuffers(clear_state, use_color, use_depth || use_stencil, false,
                           regs.clear_buffers.RT.Value());
+    if (regs.clear_flags.scissor) {
+        SyncScissorTest(clear_state);
+    }
+
+    if (regs.clear_flags.viewport) {
+        clear_state.EmulateViewportWithScissor();
+    }
+
     clear_state.Apply();
 
     if (use_color) {
@@ -573,7 +612,7 @@ void RasterizerOpenGL::DrawArrays() {
         return;
 
     MICROPROFILE_SCOPE(OpenGL_Drawing);
-    const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+    auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
     const auto& regs = gpu.regs;
 
     ScopeAcquireGLContext acquire_context{emu_window};
@@ -588,12 +627,12 @@ void RasterizerOpenGL::DrawArrays() {
     SyncLogicOpState();
     SyncCullMode();
     SyncPrimitiveRestart();
-    SyncScissorTest();
+    SyncScissorTest(state);
     // Alpha Testing is synced on shaders.
     SyncTransformFeedback();
     SyncPointState();
     CheckAlphaTests();
-
+    SyncPolygonOffset();
     // TODO(bunnei): Sync framebuffer_scale uniform here
     // TODO(bunnei): Sync scissorbox uniform(s) here
 
@@ -626,7 +665,11 @@ void RasterizerOpenGL::DrawArrays() {
     // Add space for at least 18 constant buffers
     buffer_size += Maxwell::MaxConstBuffers * (MaxConstbufferSize + uniform_buffer_alignment);
 
-    buffer_cache.Map(buffer_size);
+    bool invalidate = buffer_cache.Map(buffer_size);
+    if (invalidate) {
+        // As all cached buffers are invalidated, we need to recheck their state.
+        gpu.dirty_flags.vertex_attrib_format = 0xFFFFFFFF;
+    }
 
     SetupVertexFormat();
     SetupVertexBuffer();
@@ -815,7 +858,7 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr
     }
     const u32 bias = config.mip_lod_bias.Value();
     // Sign extend the 13-bit value.
-    const u32 mask = 1U << (13 - 1);
+    constexpr u32 mask = 1U << (13 - 1);
     const float bias_lod = static_cast<s32>((bias ^ mask) - mask) / 256.f;
     if (lod_bias != bias_lod) {
         lod_bias = bias_lod;
@@ -942,13 +985,25 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
 
 void RasterizerOpenGL::SyncViewport(OpenGLState& current_state) {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
-    for (std::size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumViewports; i++) {
-        const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()};
+    const bool geometry_shaders_enabled =
+        regs.IsShaderConfigEnabled(static_cast<size_t>(Maxwell::ShaderProgram::Geometry));
+    const std::size_t viewport_count =
+        geometry_shaders_enabled ? Tegra::Engines::Maxwell3D::Regs::NumViewports : 1;
+    for (std::size_t i = 0; i < viewport_count; i++) {
         auto& viewport = current_state.viewports[i];
-        viewport.x = viewport_rect.left;
-        viewport.y = viewport_rect.bottom;
-        viewport.width = static_cast<GLfloat>(viewport_rect.GetWidth());
-        viewport.height = static_cast<GLfloat>(viewport_rect.GetHeight());
+        const auto& src = regs.viewports[i];
+        if (regs.viewport_transform_enabled) {
+            const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()};
+            viewport.x = viewport_rect.left;
+            viewport.y = viewport_rect.bottom;
+            viewport.width = viewport_rect.GetWidth();
+            viewport.height = viewport_rect.GetHeight();
+        } else {
+            viewport.x = src.x;
+            viewport.y = src.y;
+            viewport.width = src.width;
+            viewport.height = src.height;
+        }
         viewport.depth_range_far = regs.viewports[i].depth_range_far;
         viewport.depth_range_near = regs.viewports[i].depth_range_near;
     }
@@ -1120,11 +1175,15 @@ void RasterizerOpenGL::SyncLogicOpState() {
     state.logic_op.operation = MaxwellToGL::LogicOp(regs.logic_op.operation);
 }
 
-void RasterizerOpenGL::SyncScissorTest() {
+void RasterizerOpenGL::SyncScissorTest(OpenGLState& current_state) {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
-    for (std::size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumViewports; i++) {
+    const bool geometry_shaders_enabled =
+        regs.IsShaderConfigEnabled(static_cast<size_t>(Maxwell::ShaderProgram::Geometry));
+    const std::size_t viewport_count =
+        geometry_shaders_enabled ? Tegra::Engines::Maxwell3D::Regs::NumViewports : 1;
+    for (std::size_t i = 0; i < viewport_count; i++) {
         const auto& src = regs.scissor_test[i];
-        auto& dst = state.viewports[i].scissor;
+        auto& dst = current_state.viewports[i].scissor;
         dst.enabled = (src.enable != 0);
         if (dst.enabled == 0) {
             return;
@@ -1152,6 +1211,16 @@ void RasterizerOpenGL::SyncPointState() {
     state.point.size = regs.point_size;
 }
 
+void RasterizerOpenGL::SyncPolygonOffset() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+    state.polygon_offset.fill_enable = regs.polygon_offset_fill_enable != 0;
+    state.polygon_offset.line_enable = regs.polygon_offset_line_enable != 0;
+    state.polygon_offset.point_enable = regs.polygon_offset_point_enable != 0;
+    state.polygon_offset.units = regs.polygon_offset_units;
+    state.polygon_offset.factor = regs.polygon_offset_factor;
+    state.polygon_offset.clamp = regs.polygon_offset_clamp;
+}
+
 void RasterizerOpenGL::CheckAlphaTests() {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -91,19 +91,20 @@ private:
         void SyncWithConfig(const Tegra::Texture::TSCEntry& info);
 
     private:
-        Tegra::Texture::TextureFilter mag_filter;
-        Tegra::Texture::TextureFilter min_filter;
-        Tegra::Texture::TextureMipmapFilter mip_filter;
-        Tegra::Texture::WrapMode wrap_u;
-        Tegra::Texture::WrapMode wrap_v;
-        Tegra::Texture::WrapMode wrap_p;
-        bool uses_depth_compare;
-        Tegra::Texture::DepthCompareFunc depth_compare_func;
-        GLvec4 border_color;
-        float min_lod;
-        float max_lod;
-        float lod_bias;
-        float max_anisotropic;
+        Tegra::Texture::TextureFilter mag_filter = Tegra::Texture::TextureFilter::Nearest;
+        Tegra::Texture::TextureFilter min_filter = Tegra::Texture::TextureFilter::Nearest;
+        Tegra::Texture::TextureMipmapFilter mip_filter = Tegra::Texture::TextureMipmapFilter::None;
+        Tegra::Texture::WrapMode wrap_u = Tegra::Texture::WrapMode::ClampToEdge;
+        Tegra::Texture::WrapMode wrap_v = Tegra::Texture::WrapMode::ClampToEdge;
+        Tegra::Texture::WrapMode wrap_p = Tegra::Texture::WrapMode::ClampToEdge;
+        bool uses_depth_compare = false;
+        Tegra::Texture::DepthCompareFunc depth_compare_func =
+            Tegra::Texture::DepthCompareFunc::Always;
+        GLvec4 border_color = {};
+        float min_lod = 0.0f;
+        float max_lod = 16.0f;
+        float lod_bias = 0.0f;
+        float max_anisotropic = 1.0f;
     };
 
     /**
@@ -171,7 +172,7 @@ private:
     void SyncMultiSampleState();
 
     /// Syncs the scissor test state to match the guest state
-    void SyncScissorTest();
+    void SyncScissorTest(OpenGLState& current_state);
 
     /// Syncs the transform feedback state to match the guest state
     void SyncTransformFeedback();
@@ -182,13 +183,14 @@ private:
     /// Syncs Color Mask
     void SyncColorMask();
 
+    /// Syncs the polygon offsets
+    void SyncPolygonOffset();
+
     /// Check asserts for alpha testing.
     void CheckAlphaTests();
 
     bool has_ARB_direct_state_access = false;
     bool has_ARB_multi_bind = false;
-    bool has_ARB_separate_shader_objects = false;
-    bool has_ARB_vertex_attrib_binding = false;
 
     OpenGLState state;
 

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -15,6 +15,7 @@
 #include "core/memory.h"
 #include "core/settings.h"
 #include "video_core/engines/maxwell_3d.h"
+#include "video_core/morton.h"
 #include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/gl_rasterizer_cache.h"
 #include "video_core/renderer_opengl/gl_state.h"
@@ -22,10 +23,11 @@
 #include "video_core/surface.h"
 #include "video_core/textures/astc.h"
 #include "video_core/textures/decoders.h"
-#include "video_core/utils.h"
 
 namespace OpenGL {
 
+using VideoCore::MortonSwizzle;
+using VideoCore::MortonSwizzleMode;
 using VideoCore::Surface::ComponentTypeFromDepthFormat;
 using VideoCore::Surface::ComponentTypeFromRenderTarget;
 using VideoCore::Surface::ComponentTypeFromTexture;
@@ -265,11 +267,11 @@ static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex
     {GL_COMPRESSED_RG_RGTC2, GL_RG, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
      true},                                                                     // DXN2UNORM
     {GL_COMPRESSED_SIGNED_RG_RGTC2, GL_RG, GL_INT, ComponentType::SNorm, true}, // DXN2SNORM
-    {GL_COMPRESSED_RGBA_BPTC_UNORM_ARB, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+    {GL_COMPRESSED_RGBA_BPTC_UNORM, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
      true}, // BC7U
-    {GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT_ARB, GL_RGB, GL_UNSIGNED_INT_8_8_8_8,
-     ComponentType::Float, true}, // BC6H_UF16
-    {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT_ARB, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::Float,
+    {GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::Float,
+     true}, // BC6H_UF16
+    {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::Float,
      true},                                                                    // BC6H_SF16
     {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},        // ASTC_2D_4X4
     {GL_RG8, GL_RG, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},            // G8R8U
@@ -306,8 +308,8 @@ static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex
      true}, // DXT23_SRGB
     {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
      true}, // DXT45_SRGB
-    {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM_ARB, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8,
-     ComponentType::UNorm, true},                                              // BC7U_SRGB
+    {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+     true},                                                                    // BC7U_SRGB
     {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_4X4_SRGB
     {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X8_SRGB
     {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X5_SRGB
@@ -346,7 +348,7 @@ static GLenum SurfaceTargetToGL(SurfaceTarget target) {
     case SurfaceTarget::TextureCubemap:
         return GL_TEXTURE_CUBE_MAP;
     case SurfaceTarget::TextureCubeArray:
-        return GL_TEXTURE_CUBE_MAP_ARRAY_ARB;
+        return GL_TEXTURE_CUBE_MAP_ARRAY;
     }
     LOG_CRITICAL(Render_OpenGL, "Unimplemented texture target={}", static_cast<u32>(target));
     UNREACHABLE();
@@ -370,174 +372,7 @@ MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
     return {0, actual_height, MipWidth(mip_level), 0};
 }
 
-template <bool morton_to_gl, PixelFormat format>
-void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth, u8* gl_buffer,
-                std::size_t gl_buffer_size, VAddr addr) {
-    constexpr u32 bytes_per_pixel = GetBytesPerPixel(format);
-
-    // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
-    // pixel values.
-    const u32 tile_size_x{GetDefaultBlockWidth(format)};
-    const u32 tile_size_y{GetDefaultBlockHeight(format)};
-
-    if (morton_to_gl) {
-        Tegra::Texture::UnswizzleTexture(gl_buffer, addr, tile_size_x, tile_size_y, bytes_per_pixel,
-                                         stride, height, depth, block_height, block_depth);
-    } else {
-        Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x,
-                                         (height + tile_size_y - 1) / tile_size_y, depth,
-                                         bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr),
-                                         gl_buffer, false, block_height, block_depth);
-    }
-}
-
-using GLConversionArray = std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
-                                     VideoCore::Surface::MaxPixelFormat>;
-
-static constexpr GLConversionArray morton_to_gl_fns = {
-    // clang-format off
-        MortonCopy<true, PixelFormat::ABGR8U>,
-        MortonCopy<true, PixelFormat::ABGR8S>,
-        MortonCopy<true, PixelFormat::ABGR8UI>,
-        MortonCopy<true, PixelFormat::B5G6R5U>,
-        MortonCopy<true, PixelFormat::A2B10G10R10U>,
-        MortonCopy<true, PixelFormat::A1B5G5R5U>,
-        MortonCopy<true, PixelFormat::R8U>,
-        MortonCopy<true, PixelFormat::R8UI>,
-        MortonCopy<true, PixelFormat::RGBA16F>,
-        MortonCopy<true, PixelFormat::RGBA16U>,
-        MortonCopy<true, PixelFormat::RGBA16UI>,
-        MortonCopy<true, PixelFormat::R11FG11FB10F>,
-        MortonCopy<true, PixelFormat::RGBA32UI>,
-        MortonCopy<true, PixelFormat::DXT1>,
-        MortonCopy<true, PixelFormat::DXT23>,
-        MortonCopy<true, PixelFormat::DXT45>,
-        MortonCopy<true, PixelFormat::DXN1>,
-        MortonCopy<true, PixelFormat::DXN2UNORM>,
-        MortonCopy<true, PixelFormat::DXN2SNORM>,
-        MortonCopy<true, PixelFormat::BC7U>,
-        MortonCopy<true, PixelFormat::BC6H_UF16>,
-        MortonCopy<true, PixelFormat::BC6H_SF16>,
-        MortonCopy<true, PixelFormat::ASTC_2D_4X4>,
-        MortonCopy<true, PixelFormat::G8R8U>,
-        MortonCopy<true, PixelFormat::G8R8S>,
-        MortonCopy<true, PixelFormat::BGRA8>,
-        MortonCopy<true, PixelFormat::RGBA32F>,
-        MortonCopy<true, PixelFormat::RG32F>,
-        MortonCopy<true, PixelFormat::R32F>,
-        MortonCopy<true, PixelFormat::R16F>,
-        MortonCopy<true, PixelFormat::R16U>,
-        MortonCopy<true, PixelFormat::R16S>,
-        MortonCopy<true, PixelFormat::R16UI>,
-        MortonCopy<true, PixelFormat::R16I>,
-        MortonCopy<true, PixelFormat::RG16>,
-        MortonCopy<true, PixelFormat::RG16F>,
-        MortonCopy<true, PixelFormat::RG16UI>,
-        MortonCopy<true, PixelFormat::RG16I>,
-        MortonCopy<true, PixelFormat::RG16S>,
-        MortonCopy<true, PixelFormat::RGB32F>,
-        MortonCopy<true, PixelFormat::RGBA8_SRGB>,
-        MortonCopy<true, PixelFormat::RG8U>,
-        MortonCopy<true, PixelFormat::RG8S>,
-        MortonCopy<true, PixelFormat::RG32UI>,
-        MortonCopy<true, PixelFormat::R32UI>,
-        MortonCopy<true, PixelFormat::ASTC_2D_8X8>,
-        MortonCopy<true, PixelFormat::ASTC_2D_8X5>,
-        MortonCopy<true, PixelFormat::ASTC_2D_5X4>,
-        MortonCopy<true, PixelFormat::BGRA8_SRGB>,
-        MortonCopy<true, PixelFormat::DXT1_SRGB>,
-        MortonCopy<true, PixelFormat::DXT23_SRGB>,
-        MortonCopy<true, PixelFormat::DXT45_SRGB>,
-        MortonCopy<true, PixelFormat::BC7U_SRGB>,
-        MortonCopy<true, PixelFormat::ASTC_2D_4X4_SRGB>,
-        MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>,
-        MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>,
-        MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>,
-        MortonCopy<true, PixelFormat::ASTC_2D_5X5>,
-        MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>,
-        MortonCopy<true, PixelFormat::ASTC_2D_10X8>,
-        MortonCopy<true, PixelFormat::ASTC_2D_10X8_SRGB>,
-        MortonCopy<true, PixelFormat::Z32F>,
-        MortonCopy<true, PixelFormat::Z16>,
-        MortonCopy<true, PixelFormat::Z24S8>,
-        MortonCopy<true, PixelFormat::S8Z24>,
-        MortonCopy<true, PixelFormat::Z32FS8>,
-    // clang-format on
-};
-
-static constexpr GLConversionArray gl_to_morton_fns = {
-    // clang-format off
-        MortonCopy<false, PixelFormat::ABGR8U>,
-        MortonCopy<false, PixelFormat::ABGR8S>,
-        MortonCopy<false, PixelFormat::ABGR8UI>,
-        MortonCopy<false, PixelFormat::B5G6R5U>,
-        MortonCopy<false, PixelFormat::A2B10G10R10U>,
-        MortonCopy<false, PixelFormat::A1B5G5R5U>,
-        MortonCopy<false, PixelFormat::R8U>,
-        MortonCopy<false, PixelFormat::R8UI>,
-        MortonCopy<false, PixelFormat::RGBA16F>,
-        MortonCopy<false, PixelFormat::RGBA16U>,
-        MortonCopy<false, PixelFormat::RGBA16UI>,
-        MortonCopy<false, PixelFormat::R11FG11FB10F>,
-        MortonCopy<false, PixelFormat::RGBA32UI>,
-        MortonCopy<false, PixelFormat::DXT1>,
-        MortonCopy<false, PixelFormat::DXT23>,
-        MortonCopy<false, PixelFormat::DXT45>,
-        MortonCopy<false, PixelFormat::DXN1>,
-        MortonCopy<false, PixelFormat::DXN2UNORM>,
-        MortonCopy<false, PixelFormat::DXN2SNORM>,
-        MortonCopy<false, PixelFormat::BC7U>,
-        MortonCopy<false, PixelFormat::BC6H_UF16>,
-        MortonCopy<false, PixelFormat::BC6H_SF16>,
-        // TODO(Subv): Swizzling ASTC formats are not supported
-        nullptr,
-        MortonCopy<false, PixelFormat::G8R8U>,
-        MortonCopy<false, PixelFormat::G8R8S>,
-        MortonCopy<false, PixelFormat::BGRA8>,
-        MortonCopy<false, PixelFormat::RGBA32F>,
-        MortonCopy<false, PixelFormat::RG32F>,
-        MortonCopy<false, PixelFormat::R32F>,
-        MortonCopy<false, PixelFormat::R16F>,
-        MortonCopy<false, PixelFormat::R16U>,
-        MortonCopy<false, PixelFormat::R16S>,
-        MortonCopy<false, PixelFormat::R16UI>,
-        MortonCopy<false, PixelFormat::R16I>,
-        MortonCopy<false, PixelFormat::RG16>,
-        MortonCopy<false, PixelFormat::RG16F>,
-        MortonCopy<false, PixelFormat::RG16UI>,
-        MortonCopy<false, PixelFormat::RG16I>,
-        MortonCopy<false, PixelFormat::RG16S>,
-        MortonCopy<false, PixelFormat::RGB32F>,
-        MortonCopy<false, PixelFormat::RGBA8_SRGB>,
-        MortonCopy<false, PixelFormat::RG8U>,
-        MortonCopy<false, PixelFormat::RG8S>,
-        MortonCopy<false, PixelFormat::RG32UI>,
-        MortonCopy<false, PixelFormat::R32UI>,
-        nullptr,
-        nullptr,
-        nullptr,
-        MortonCopy<false, PixelFormat::BGRA8_SRGB>,
-        MortonCopy<false, PixelFormat::DXT1_SRGB>,
-        MortonCopy<false, PixelFormat::DXT23_SRGB>,
-        MortonCopy<false, PixelFormat::DXT45_SRGB>,
-        MortonCopy<false, PixelFormat::BC7U_SRGB>,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        nullptr,
-        MortonCopy<false, PixelFormat::Z32F>,
-        MortonCopy<false, PixelFormat::Z16>,
-        MortonCopy<false, PixelFormat::Z24S8>,
-        MortonCopy<false, PixelFormat::S8Z24>,
-        MortonCopy<false, PixelFormat::Z32FS8>,
-    // clang-format on
-};
-
-void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params,
+void SwizzleFunc(const MortonSwizzleMode& mode, const SurfaceParams& params,
                  std::vector<u8>& gl_buffer, u32 mip_level) {
     u32 depth = params.MipDepth(mip_level);
     if (params.target == SurfaceTarget::Texture2D) {
@@ -550,19 +385,19 @@ void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params
         const u64 layer_size = params.LayerMemorySize();
         const u64 gl_size = params.LayerSizeGL(mip_level);
         for (u32 i = 0; i < params.depth; i++) {
-            functions[static_cast<std::size_t>(params.pixel_format)](
-                params.MipWidth(mip_level), params.MipBlockHeight(mip_level),
-                params.MipHeight(mip_level), params.MipBlockDepth(mip_level), 1,
-                gl_buffer.data() + offset_gl, gl_size, params.addr + offset);
+            MortonSwizzle(mode, params.pixel_format, params.MipWidth(mip_level),
+                          params.MipBlockHeight(mip_level), params.MipHeight(mip_level),
+                          params.MipBlockDepth(mip_level), 1, gl_buffer.data() + offset_gl, gl_size,
+                          params.addr + offset);
             offset += layer_size;
             offset_gl += gl_size;
         }
     } else {
         const u64 offset = params.GetMipmapLevelOffset(mip_level);
-        functions[static_cast<std::size_t>(params.pixel_format)](
-            params.MipWidth(mip_level), params.MipBlockHeight(mip_level),
-            params.MipHeight(mip_level), params.MipBlockDepth(mip_level), depth, gl_buffer.data(),
-            gl_buffer.size(), params.addr + offset);
+        MortonSwizzle(mode, params.pixel_format, params.MipWidth(mip_level),
+                      params.MipBlockHeight(mip_level), params.MipHeight(mip_level),
+                      params.MipBlockDepth(mip_level), depth, gl_buffer.data(), gl_buffer.size(),
+                      params.addr + offset);
     }
 }
 
@@ -726,7 +561,7 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
     const std::size_t buffer_size = std::max(src_params.size_in_bytes, dst_params.size_in_bytes);
 
     glBindBuffer(GL_PIXEL_PACK_BUFFER, copy_pbo_handle);
-    glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW_ARB);
+    glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW);
     if (source_format.compressed) {
         glGetCompressedTextureImage(src_surface->Texture().handle, src_attachment,
                                     static_cast<GLsizei>(src_params.size_in_bytes), nullptr);
@@ -996,7 +831,7 @@ void CachedSurface::LoadGLBuffer() {
         ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
                    params.block_width, static_cast<u32>(params.target));
         for (u32 i = 0; i < params.max_mip_level; i++)
-            SwizzleFunc(morton_to_gl_fns, params, gl_buffer[i], i);
+            SwizzleFunc(MortonSwizzleMode::MortonToLinear, params, gl_buffer[i], i);
     } else {
         const auto texture_src_data{Memory::GetPointer(params.addr)};
         const auto texture_src_data_end{texture_src_data + params.size_in_bytes_gl};
@@ -1035,7 +870,7 @@ void CachedSurface::FlushGLBuffer() {
         ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
                    params.block_width, static_cast<u32>(params.target));
 
-        SwizzleFunc(gl_to_morton_fns, params, gl_buffer[0], 0);
+        SwizzleFunc(MortonSwizzleMode::LinearToMorton, params, gl_buffer[0], 0);
     } else {
         std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer[0].data(), GetSizeInBytes());
     }
@@ -1275,6 +1110,31 @@ Surface RasterizerCacheOpenGL::GetUncachedSurface(const SurfaceParams& params) {
     return surface;
 }
 
+void RasterizerCacheOpenGL::FastLayeredCopySurface(const Surface& src_surface,
+                                                   const Surface& dst_surface) {
+    const auto& init_params{src_surface->GetSurfaceParams()};
+    const auto& dst_params{dst_surface->GetSurfaceParams()};
+    VAddr address = init_params.addr;
+    const std::size_t layer_size = dst_params.LayerMemorySize();
+    for (u32 layer = 0; layer < dst_params.depth; layer++) {
+        for (u32 mipmap = 0; mipmap < dst_params.max_mip_level; mipmap++) {
+            const VAddr sub_address = address + dst_params.GetMipmapLevelOffset(mipmap);
+            const Surface& copy = TryGet(sub_address);
+            if (!copy)
+                continue;
+            const auto& src_params{copy->GetSurfaceParams()};
+            const u32 width{std::min(src_params.width, dst_params.MipWidth(mipmap))};
+            const u32 height{std::min(src_params.height, dst_params.MipHeight(mipmap))};
+
+            glCopyImageSubData(copy->Texture().handle, SurfaceTargetToGL(src_params.target), 0, 0,
+                               0, 0, dst_surface->Texture().handle,
+                               SurfaceTargetToGL(dst_params.target), mipmap, 0, 0, layer, width,
+                               height, 1);
+        }
+        address += layer_size;
+    }
+}
+
 void RasterizerCacheOpenGL::FermiCopySurface(
     const Tegra::Engines::Fermi2D::Regs::Surface& src_config,
     const Tegra::Engines::Fermi2D::Regs::Surface& dst_config) {
@@ -1340,11 +1200,13 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
             CopySurface(old_surface, new_surface, copy_pbo.handle);
         }
         break;
-    case SurfaceTarget::TextureCubemap:
     case SurfaceTarget::Texture3D:
+        AccurateCopySurface(old_surface, new_surface);
+        break;
+    case SurfaceTarget::TextureCubemap:
     case SurfaceTarget::Texture2DArray:
     case SurfaceTarget::TextureCubeArray:
-        AccurateCopySurface(old_surface, new_surface);
+        FastLayeredCopySurface(old_surface, new_surface);
         break;
     default:
         LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -350,6 +350,7 @@ private:
 
     /// Performs a slow but accurate surface copy, flushing to RAM and reinterpreting the data
     void AccurateCopySurface(const Surface& src_surface, const Surface& dst_surface);
+    void FastLayeredCopySurface(const Surface& src_surface, const Surface& dst_surface);
 
     /// The surface reserve is a "backup" cache, this is where we put unique surfaces that have
     /// previously been used. This is to prevent surfaces from being constantly created and

src/video_core/renderer_opengl/gl_shader_cache.cpp

@@ -84,6 +84,7 @@ CachedShader::CachedShader(VAddr addr, Maxwell::ShaderProgram program_type)
     }
 
     entries = program_result.second;
+    shader_length = entries.shader_length;
 
     if (program_type != Maxwell::ShaderProgram::Geometry) {
         OGLShader shader;

src/video_core/renderer_opengl/gl_shader_cache.h

@@ -30,7 +30,7 @@ public:
     }
 
     std::size_t GetSizeInBytes() const override {
-        return GLShader::MAX_PROGRAM_CODE_LENGTH * sizeof(u64);
+        return shader_length;
     }
 
     // We do not have to flush this cache as things in it are never modified by us.
@@ -82,6 +82,7 @@ private:
                                u32 max_vertices, const std::string& debug_name);
 
     VAddr addr;
+    std::size_t shader_length;
     Maxwell::ShaderProgram program_type;
     GLShader::ShaderSetup setup;
     GLShader::ShaderEntries entries;

src/video_core/renderer_opengl/gl_shader_gen.cpp

@@ -24,8 +24,7 @@ layout (location = 0) out vec4 position;
 
 layout(std140) uniform vs_config {
     vec4 viewport_flip;
-    uvec4 instance_id;
-    uvec4 flip_stage;
+    uvec4 config_pack; // instance_id, flip_stage, y_direction, padding
     uvec4 alpha_test;
 };
 )";
@@ -63,7 +62,8 @@ void main() {
     out += R"(
 
     // Check if the flip stage is VertexB
-    if (flip_stage[0] == 1) {
+    // Config pack's second value is flip_stage
+    if (config_pack[1] == 1) {
         // Viewport can be flipped, which is unsupported by glViewport
         position.xy *= viewport_flip.xy;
     }
@@ -71,7 +71,7 @@ void main() {
 
     // TODO(bunnei): This is likely a hack, position.w should be interpolated as 1.0
     // For now, this is here to bring order in lieu of proper emulation
-    if (flip_stage[0] == 1) {
+    if (config_pack[1] == 1) {
         position.w = 1.0;
     }
 }
@@ -101,8 +101,7 @@ layout (location = 0) out vec4 position;
 
 layout (std140) uniform gs_config {
     vec4 viewport_flip;
-    uvec4 instance_id;
-    uvec4 flip_stage;
+    uvec4 config_pack; // instance_id, flip_stage, y_direction, padding
     uvec4 alpha_test;
 };
 
@@ -139,8 +138,7 @@ layout (location = 0) in vec4 position;
 
 layout (std140) uniform fs_config {
     vec4 viewport_flip;
-    uvec4 instance_id;
-    uvec4 flip_stage;
+    uvec4 config_pack; // instance_id, flip_stage, y_direction, padding
     uvec4 alpha_test;
 };
 

src/video_core/renderer_opengl/gl_shader_gen.h

@@ -163,6 +163,7 @@ private:
 struct ShaderEntries {
     std::vector<ConstBufferEntry> const_buffer_entries;
     std::vector<SamplerEntry> texture_samplers;
+    std::size_t shader_length;
 };
 
 using ProgramResult = std::pair<std::string, ShaderEntries>;

src/video_core/renderer_opengl/gl_shader_manager.cpp

@@ -27,16 +27,18 @@ void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& sh
     alpha_test.func = func;
     alpha_test.ref = regs.alpha_test_ref;
 
-    // We only assign the instance to the first component of the vector, the rest is just padding.
-    instance_id[0] = state.current_instance;
+    instance_id = state.current_instance;
 
     // Assign in which stage the position has to be flipped
     // (the last stage before the fragment shader).
     if (gpu.regs.shader_config[static_cast<u32>(Maxwell3D::Regs::ShaderProgram::Geometry)].enable) {
-        flip_stage[0] = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::Geometry);
+        flip_stage = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::Geometry);
     } else {
-        flip_stage[0] = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::VertexB);
+        flip_stage = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::VertexB);
     }
+
+    // Y_NEGATE controls what value S2R returns for the Y_DIRECTION system value.
+    y_direction = regs.screen_y_control.y_negate == 0 ? 1.f : -1.f;
 }
 
 } // namespace OpenGL::GLShader

src/video_core/renderer_opengl/gl_shader_manager.h

@@ -21,8 +21,11 @@ using Tegra::Engines::Maxwell3D;
 struct MaxwellUniformData {
     void SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage);
     alignas(16) GLvec4 viewport_flip;
-    alignas(16) GLuvec4 instance_id;
-    alignas(16) GLuvec4 flip_stage;
+    struct alignas(16) {
+        GLuint instance_id;
+        GLuint flip_stage;
+        GLfloat y_direction;
+    };
     struct alignas(16) {
         GLuint enabled;
         GLuint func;
@@ -30,7 +33,7 @@ struct MaxwellUniformData {
         GLuint padding;
     } alpha_test;
 };
-static_assert(sizeof(MaxwellUniformData) == 64, "MaxwellUniformData structure size is incorrect");
+static_assert(sizeof(MaxwellUniformData) == 48, "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/gl_state.cpp

@@ -92,6 +92,13 @@ OpenGLState::OpenGLState() {
 
     point.size = 1;
     fragment_color_clamp.enabled = false;
+
+    polygon_offset.fill_enable = false;
+    polygon_offset.line_enable = false;
+    polygon_offset.point_enable = false;
+    polygon_offset.factor = 0.0f;
+    polygon_offset.units = 0.0f;
+    polygon_offset.clamp = 0.0f;
 }
 
 void OpenGLState::ApplyDefaultState() {
@@ -233,6 +240,28 @@ void OpenGLState::ApplyStencilTest() const {
         config_stencil(GL_BACK, stencil.back, cur_state.stencil.back);
     }
 }
+// Viewport does not affects glClearBuffer so emulate viewport using scissor test
+void OpenGLState::EmulateViewportWithScissor() {
+    auto& current = viewports[0];
+    if (current.scissor.enabled) {
+        const GLint left = std::max(current.x, current.scissor.x);
+        const GLint right =
+            std::max(current.x + current.width, current.scissor.x + current.scissor.width);
+        const GLint bottom = std::max(current.y, current.scissor.y);
+        const GLint top =
+            std::max(current.y + current.height, current.scissor.y + current.scissor.height);
+        current.scissor.x = std::max(left, 0);
+        current.scissor.y = std::max(bottom, 0);
+        current.scissor.width = std::max(right - left, 0);
+        current.scissor.height = std::max(top - bottom, 0);
+    } else {
+        current.scissor.enabled = true;
+        current.scissor.x = current.x;
+        current.scissor.y = current.y;
+        current.scissor.width = current.width;
+        current.scissor.height = current.height;
+    }
+}
 
 void OpenGLState::ApplyViewport() const {
     if (GLAD_GL_ARB_viewport_array && geometry_shaders.enabled) {
@@ -242,7 +271,9 @@ void OpenGLState::ApplyViewport() const {
             const auto& updated = viewports[i];
             if (updated.x != current.x || updated.y != current.y ||
                 updated.width != current.width || updated.height != current.height) {
-                glViewportIndexedf(i, updated.x, updated.y, updated.width, updated.height);
+                glViewportIndexedf(
+                    i, static_cast<GLfloat>(updated.x), static_cast<GLfloat>(updated.y),
+                    static_cast<GLfloat>(updated.width), static_cast<GLfloat>(updated.height));
             }
             if (updated.depth_range_near != current.depth_range_near ||
                 updated.depth_range_far != current.depth_range_far) {
@@ -270,8 +301,7 @@ void OpenGLState::ApplyViewport() const {
         const auto& updated = viewports[0];
         if (updated.x != current.x || updated.y != current.y || updated.width != current.width ||
             updated.height != current.height) {
-            glViewport(static_cast<GLint>(updated.x), static_cast<GLint>(updated.y),
-                       static_cast<GLsizei>(updated.width), static_cast<GLsizei>(updated.height));
+            glViewport(updated.x, updated.y, updated.width, updated.height);
         }
         if (updated.depth_range_near != current.depth_range_near ||
             updated.depth_range_far != current.depth_range_far) {
@@ -339,14 +369,14 @@ void OpenGLState::ApplyTargetBlending(std::size_t target, bool force) const {
     if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
         updated.dst_rgb_func != current.dst_rgb_func || updated.src_a_func != current.src_a_func ||
         updated.dst_a_func != current.dst_a_func) {
-        glBlendFuncSeparateiARB(static_cast<GLuint>(target), updated.src_rgb_func,
-                                updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func);
+        glBlendFuncSeparatei(static_cast<GLuint>(target), updated.src_rgb_func,
+                             updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func);
     }
 
     if (blend_changed || updated.rgb_equation != current.rgb_equation ||
         updated.a_equation != current.a_equation) {
-        glBlendEquationSeparateiARB(static_cast<GLuint>(target), updated.rgb_equation,
-                                    updated.a_equation);
+        glBlendEquationSeparatei(static_cast<GLuint>(target), updated.rgb_equation,
+                                 updated.a_equation);
     }
 }
 
@@ -383,6 +413,55 @@ void OpenGLState::ApplyLogicOp() const {
     }
 }
 
+void OpenGLState::ApplyPolygonOffset() const {
+
+    const bool fill_enable_changed =
+        polygon_offset.fill_enable != cur_state.polygon_offset.fill_enable;
+    const bool line_enable_changed =
+        polygon_offset.line_enable != cur_state.polygon_offset.line_enable;
+    const bool point_enable_changed =
+        polygon_offset.point_enable != cur_state.polygon_offset.point_enable;
+    const bool factor_changed = polygon_offset.factor != cur_state.polygon_offset.factor;
+    const bool units_changed = polygon_offset.units != cur_state.polygon_offset.units;
+    const bool clamp_changed = polygon_offset.clamp != cur_state.polygon_offset.clamp;
+
+    if (fill_enable_changed) {
+        if (polygon_offset.fill_enable) {
+            glEnable(GL_POLYGON_OFFSET_FILL);
+        } else {
+            glDisable(GL_POLYGON_OFFSET_FILL);
+        }
+    }
+
+    if (line_enable_changed) {
+        if (polygon_offset.line_enable) {
+            glEnable(GL_POLYGON_OFFSET_LINE);
+        } else {
+            glDisable(GL_POLYGON_OFFSET_LINE);
+        }
+    }
+
+    if (point_enable_changed) {
+        if (polygon_offset.point_enable) {
+            glEnable(GL_POLYGON_OFFSET_POINT);
+        } else {
+            glDisable(GL_POLYGON_OFFSET_POINT);
+        }
+    }
+
+    if ((polygon_offset.fill_enable || polygon_offset.line_enable || polygon_offset.point_enable) &&
+        (factor_changed || units_changed || clamp_changed)) {
+
+        if (GLAD_GL_EXT_polygon_offset_clamp && polygon_offset.clamp != 0) {
+            glPolygonOffsetClamp(polygon_offset.factor, polygon_offset.units, polygon_offset.clamp);
+        } else {
+            glPolygonOffset(polygon_offset.factor, polygon_offset.units);
+            UNIMPLEMENTED_IF_MSG(polygon_offset.clamp != 0,
+                                 "Unimplemented Depth polygon offset clamp.");
+        }
+    }
+}
+
 void OpenGLState::ApplyTextures() const {
     for (std::size_t i = 0; i < std::size(texture_units); ++i) {
         const auto& texture_unit = texture_units[i];
@@ -509,6 +588,7 @@ void OpenGLState::Apply() const {
     ApplyLogicOp();
     ApplyTextures();
     ApplySamplers();
+    ApplyPolygonOffset();
     cur_state = *this;
 }
 

src/video_core/renderer_opengl/gl_state.h

@@ -156,10 +156,10 @@ public:
     } draw;
 
     struct viewport {
-        GLfloat x;
-        GLfloat y;
-        GLfloat width;
-        GLfloat height;
+        GLint x;
+        GLint y;
+        GLint width;
+        GLint height;
         GLfloat depth_range_near; // GL_DEPTH_RANGE
         GLfloat depth_range_far;  // GL_DEPTH_RANGE
         struct {
@@ -176,6 +176,15 @@ public:
         float size; // GL_POINT_SIZE
     } point;
 
+    struct {
+        bool point_enable;
+        bool line_enable;
+        bool fill_enable;
+        GLfloat units;
+        GLfloat factor;
+        GLfloat clamp;
+    } polygon_offset;
+
     std::array<bool, 2> clip_distance; // GL_CLIP_DISTANCE
 
     OpenGLState();
@@ -206,6 +215,7 @@ public:
     OpenGLState& ResetBuffer(GLuint handle);
     OpenGLState& ResetVertexArray(GLuint handle);
     OpenGLState& ResetFramebuffer(GLuint handle);
+    void EmulateViewportWithScissor();
 
 private:
     static OpenGLState cur_state;
@@ -225,6 +235,7 @@ private:
     void ApplyLogicOp() const;
     void ApplyTextures() const;
     void ApplySamplers() const;
+    void ApplyPolygonOffset() const;
 };
 
 } // namespace OpenGL

src/video_core/renderer_opengl/maxwell_to_gl.h

@@ -218,14 +218,19 @@ inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) {
 inline GLenum BlendEquation(Maxwell::Blend::Equation equation) {
     switch (equation) {
     case Maxwell::Blend::Equation::Add:
+    case Maxwell::Blend::Equation::AddGL:
         return GL_FUNC_ADD;
     case Maxwell::Blend::Equation::Subtract:
+    case Maxwell::Blend::Equation::SubtractGL:
         return GL_FUNC_SUBTRACT;
     case Maxwell::Blend::Equation::ReverseSubtract:
+    case Maxwell::Blend::Equation::ReverseSubtractGL:
         return GL_FUNC_REVERSE_SUBTRACT;
     case Maxwell::Blend::Equation::Min:
+    case Maxwell::Blend::Equation::MinGL:
         return GL_MIN;
     case Maxwell::Blend::Equation::Max:
+    case Maxwell::Blend::Equation::MaxGL:
         return GL_MAX;
     }
     LOG_ERROR(Render_OpenGL, "Unimplemented blend equation={}", static_cast<u32>(equation));

src/video_core/renderer_opengl/renderer_opengl.cpp

@@ -19,9 +19,9 @@
 #include "core/settings.h"
 #include "core/telemetry_session.h"
 #include "core/tracer/recorder.h"
+#include "video_core/morton.h"
 #include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/renderer_opengl.h"
-#include "video_core/utils.h"
 
 namespace OpenGL {
 
@@ -490,7 +490,7 @@ bool RendererOpenGL::Init() {
     Core::Telemetry().AddField(Telemetry::FieldType::UserSystem, "GPU_Model", gpu_model);
     Core::Telemetry().AddField(Telemetry::FieldType::UserSystem, "GPU_OpenGL_Version", gl_version);
 
-    if (!GLAD_GL_VERSION_3_3) {
+    if (!GLAD_GL_VERSION_4_3) {
         return false;
     }
 

src/video_core/utils.h

@@ -1,164 +0,0 @@
-// Copyright 2014 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include "common/common_types.h"
-
-namespace VideoCore {
-
-// 8x8 Z-Order coordinate from 2D coordinates
-static inline u32 MortonInterleave(u32 x, u32 y) {
-    static const u32 xlut[] = {0x00, 0x01, 0x04, 0x05, 0x10, 0x11, 0x14, 0x15};
-    static const u32 ylut[] = {0x00, 0x02, 0x08, 0x0a, 0x20, 0x22, 0x28, 0x2a};
-    return xlut[x % 8] + ylut[y % 8];
-}
-
-/**
- * Calculates the offset of the position of the pixel in Morton order
- */
-static inline u32 GetMortonOffset(u32 x, u32 y, u32 bytes_per_pixel) {
-    // Images are split into 8x8 tiles. Each tile is composed of four 4x4 subtiles each
-    // of which is composed of four 2x2 subtiles each of which is composed of four texels.
-    // Each structure is embedded into the next-bigger one in a diagonal pattern, e.g.
-    // texels are laid out in a 2x2 subtile like this:
-    // 2 3
-    // 0 1
-    //
-    // The full 8x8 tile has the texels arranged like this:
-    //
-    // 42 43 46 47 58 59 62 63
-    // 40 41 44 45 56 57 60 61
-    // 34 35 38 39 50 51 54 55
-    // 32 33 36 37 48 49 52 53
-    // 10 11 14 15 26 27 30 31
-    // 08 09 12 13 24 25 28 29
-    // 02 03 06 07 18 19 22 23
-    // 00 01 04 05 16 17 20 21
-    //
-    // This pattern is what's called Z-order curve, or Morton order.
-
-    const unsigned int block_height = 8;
-    const unsigned int coarse_x = x & ~7;
-
-    u32 i = VideoCore::MortonInterleave(x, y);
-
-    const unsigned int offset = coarse_x * block_height;
-
-    return (i + offset) * bytes_per_pixel;
-}
-
-static inline u32 MortonInterleave128(u32 x, u32 y) {
-    // 128x128 Z-Order coordinate from 2D coordinates
-    static constexpr u32 xlut[] = {
-        0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042,
-        0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809,
-        0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000,
-        0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043,
-        0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a,
-        0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001,
-        0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048,
-        0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b,
-        0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002,
-        0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049,
-        0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840,
-        0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003,
-        0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a,
-        0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841,
-        0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008,
-        0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b,
-        0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842,
-        0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009,
-        0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800,
-        0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843,
-        0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a,
-        0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801,
-        0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848,
-        0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b,
-        0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802,
-        0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849,
-        0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040,
-        0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803,
-        0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a,
-        0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041,
-        0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808,
-        0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b,
-        0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042,
-        0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809,
-        0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b,
-    };
-    static constexpr u32 ylut[] = {
-        0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090,
-        0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124,
-        0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200,
-        0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294,
-        0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330,
-        0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404,
-        0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0,
-        0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534,
-        0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610,
-        0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4,
-        0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780,
-        0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014,
-        0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0,
-        0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184,
-        0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220,
-        0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4,
-        0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390,
-        0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424,
-        0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500,
-        0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594,
-        0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630,
-        0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704,
-        0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0,
-        0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034,
-        0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110,
-        0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4,
-        0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280,
-        0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314,
-        0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0,
-        0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484,
-        0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520,
-        0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4,
-        0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690,
-        0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724,
-        0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4,
-    };
-    return xlut[x % 128] + ylut[y % 128];
-}
-
-static inline u32 GetMortonOffset128(u32 x, u32 y, u32 bytes_per_pixel) {
-    // Calculates the offset of the position of the pixel in Morton order
-    // Framebuffer images are split into 128x128 tiles.
-
-    const unsigned int block_height = 128;
-    const unsigned int coarse_x = x & ~127;
-
-    u32 i = MortonInterleave128(x, y);
-
-    const unsigned int offset = coarse_x * block_height;
-
-    return (i + offset) * bytes_per_pixel;
-}
-
-static inline void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixel,
-                                       u32 gl_bytes_per_pixel, u8* morton_data, u8* gl_data,
-                                       bool morton_to_gl) {
-    u8* data_ptrs[2];
-    for (unsigned y = 0; y < height; ++y) {
-        for (unsigned x = 0; x < width; ++x) {
-            const u32 coarse_y = y & ~127;
-            u32 morton_offset =
-                GetMortonOffset128(x, y, bytes_per_pixel) + coarse_y * width * bytes_per_pixel;
-            u32 gl_pixel_index = (x + y * width) * gl_bytes_per_pixel;
-
-            data_ptrs[morton_to_gl] = morton_data + morton_offset;
-            data_ptrs[!morton_to_gl] = &gl_data[gl_pixel_index];
-
-            memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
-        }
-    }
-}
-
-} // namespace VideoCore

src/yuzu/applets/software_keyboard.cpp

@@ -82,8 +82,8 @@ QtSoftwareKeyboardDialog::QtSoftwareKeyboardDialog(
                            : QString::fromStdU16String(parameters.submit_text),
                        QDialogButtonBox::AcceptRole);
 
-    connect(buttons, &QDialogButtonBox::accepted, this, &QtSoftwareKeyboardDialog::Submit);
-    connect(buttons, &QDialogButtonBox::rejected, this, &QtSoftwareKeyboardDialog::Reject);
+    connect(buttons, &QDialogButtonBox::accepted, this, &QtSoftwareKeyboardDialog::accept);
+    connect(buttons, &QDialogButtonBox::rejected, this, &QtSoftwareKeyboardDialog::reject);
     layout->addWidget(header_label);
     layout->addWidget(sub_label);
     layout->addWidget(guide_label);
@@ -96,16 +96,16 @@ QtSoftwareKeyboardDialog::QtSoftwareKeyboardDialog(
 
 QtSoftwareKeyboardDialog::~QtSoftwareKeyboardDialog() = default;
 
-void QtSoftwareKeyboardDialog::Submit() {
+void QtSoftwareKeyboardDialog::accept() {
     ok = true;
     text = line_edit->text().toStdU16String();
-    accept();
+    QDialog::accept();
 }
 
-void QtSoftwareKeyboardDialog::Reject() {
+void QtSoftwareKeyboardDialog::reject() {
     ok = false;
     text.clear();
-    accept();
+    QDialog::reject();
 }
 
 std::u16string QtSoftwareKeyboardDialog::GetText() const {
@@ -129,13 +129,13 @@ QtSoftwareKeyboard::~QtSoftwareKeyboard() = default;
 
 void QtSoftwareKeyboard::RequestText(std::function<void(std::optional<std::u16string>)> out,
                                      Core::Frontend::SoftwareKeyboardParameters parameters) const {
-    text_output = out;
+    text_output = std::move(out);
     emit MainWindowGetText(parameters);
 }
 
 void QtSoftwareKeyboard::SendTextCheckDialog(std::u16string error_message,
                                              std::function<void()> finished_check) const {
-    this->finished_check = finished_check;
+    this->finished_check = std::move(finished_check);
     emit MainWindowTextCheckDialog(error_message);
 }
 

src/yuzu/applets/software_keyboard.h

@@ -33,8 +33,8 @@ public:
                              Core::Frontend::SoftwareKeyboardParameters parameters);
     ~QtSoftwareKeyboardDialog() override;
 
-    void Submit();
-    void Reject();
+    void accept() override;
+    void reject() override;
 
     std::u16string GetText() const;
     bool GetStatus() const;
@@ -70,11 +70,10 @@ signals:
     void MainWindowGetText(Core::Frontend::SoftwareKeyboardParameters parameters) const;
     void MainWindowTextCheckDialog(std::u16string error_message) const;
 
-public slots:
+private:
     void MainWindowFinishedText(std::optional<std::u16string> text);
     void MainWindowFinishedCheckDialog();
 
-private:
     mutable std::function<void(std::optional<std::u16string>)> text_output;
     mutable std::function<void()> finished_check;
 };

src/yuzu/bootmanager.cpp

@@ -310,7 +310,7 @@ void GRenderWindow::InitRenderTarget() {
     // TODO: One of these flags might be interesting: WA_OpaquePaintEvent, WA_NoBackground,
     // WA_DontShowOnScreen, WA_DeleteOnClose
     QGLFormat fmt;
-    fmt.setVersion(3, 3);
+    fmt.setVersion(4, 3);
     fmt.setProfile(QGLFormat::CoreProfile);
     fmt.setSwapInterval(false);
 

src/yuzu/configuration/config.cpp

@@ -432,6 +432,7 @@ void Config::ReadValues() {
     Settings::values.use_gdbstub = qt_config->value("use_gdbstub", false).toBool();
     Settings::values.gdbstub_port = qt_config->value("gdbstub_port", 24689).toInt();
     Settings::values.program_args = qt_config->value("program_args", "").toString().toStdString();
+    Settings::values.dump_exefs = qt_config->value("dump_exefs", false).toBool();
     Settings::values.dump_nso = qt_config->value("dump_nso", false).toBool();
     qt_config->endGroup();
 
@@ -638,6 +639,7 @@ void Config::SaveValues() {
     qt_config->setValue("use_gdbstub", Settings::values.use_gdbstub);
     qt_config->setValue("gdbstub_port", Settings::values.gdbstub_port);
     qt_config->setValue("program_args", QString::fromStdString(Settings::values.program_args));
+    qt_config->setValue("dump_exefs", Settings::values.dump_exefs);
     qt_config->setValue("dump_nso", Settings::values.dump_nso);
     qt_config->endGroup();
 

src/yuzu/configuration/configure_debug.cpp

@@ -34,6 +34,7 @@ void ConfigureDebug::setConfiguration() {
     ui->toggle_console->setChecked(UISettings::values.show_console);
     ui->log_filter_edit->setText(QString::fromStdString(Settings::values.log_filter));
     ui->homebrew_args_edit->setText(QString::fromStdString(Settings::values.program_args));
+    ui->dump_exefs->setChecked(Settings::values.dump_exefs);
     ui->dump_decompressed_nso->setChecked(Settings::values.dump_nso);
 }
 
@@ -43,6 +44,7 @@ void ConfigureDebug::applyConfiguration() {
     UISettings::values.show_console = ui->toggle_console->isChecked();
     Settings::values.log_filter = ui->log_filter_edit->text().toStdString();
     Settings::values.program_args = ui->homebrew_args_edit->text().toStdString();
+    Settings::values.dump_exefs = ui->dump_exefs->isChecked();
     Settings::values.dump_nso = ui->dump_decompressed_nso->isChecked();
     Debugger::ToggleConsole();
     Log::Filter filter;

src/yuzu/configuration/configure_debug.ui

@@ -145,6 +145,16 @@
         </property>
        </widget>
       </item>
+      <item>
+       <widget class="QCheckBox" name="dump_exefs">
+        <property name="whatsThis">
+         <string>When checked, any game that yuzu loads will have its ExeFS dumped to the yuzu/dump directory.</string>
+        </property>
+        <property name="text">
+         <string>Dump ExeFS</string>
+        </property>
+       </widget>
+      </item>
      </layout>
     </widget>
    </item>

src/yuzu/configuration/configure_graphics.ui

@@ -23,31 +23,31 @@
        </property>
        <layout class="QVBoxLayout" name="verticalLayout_2">
         <item>
-          <layout class="QHBoxLayout" name="horizontalLayout_2">
-            <item>
-              <widget class="QCheckBox" name="toggle_frame_limit">
-                <property name="text">
-                  <string>Limit Speed Percent</string>
-                </property>
-              </widget>
-            </item>
-            <item>
-              <widget class="QSpinBox" name="frame_limit">
-                <property name="suffix">
-                  <string>%</string>
-                </property>
-                <property name="minimum">
-                  <number>1</number>
-                </property>
-                <property name="maximum">
-                  <number>9999</number>
-                </property>
-                <property name="value">
-                  <number>100</number>
-                </property>
-              </widget>
-            </item>
-          </layout>
+         <layout class="QHBoxLayout" name="horizontalLayout_2">
+          <item>
+           <widget class="QCheckBox" name="toggle_frame_limit">
+            <property name="text">
+             <string>Limit Speed Percent</string>
+            </property>
+           </widget>
+          </item>
+          <item>
+           <widget class="QSpinBox" name="frame_limit">
+            <property name="suffix">
+             <string>%</string>
+            </property>
+            <property name="minimum">
+             <number>1</number>
+            </property>
+            <property name="maximum">
+             <number>9999</number>
+            </property>
+            <property name="value">
+             <number>100</number>
+            </property>
+           </widget>
+          </item>
+         </layout>
         </item>
         <item>
          <widget class="QCheckBox" name="use_accurate_gpu_emulation">
@@ -61,7 +61,7 @@
           <item>
            <widget class="QLabel" name="label">
             <property name="text">
-             <string>Internal Resolution:(Currently does nothing.)</string>
+             <string>Internal Resolution</string>
             </property>
            </widget>
           </item>
@@ -96,27 +96,27 @@
           </item>
          </layout>
         </item>
-         <item>
-           <layout class="QHBoxLayout" name="horizontalLayout_6">
-             <item>
-               <widget class="QLabel" name="bg_label">
-                 <property name="text">
-                   <string>Background Color:</string>
-                 </property>
-               </widget>
-             </item>
-             <item>
-               <widget class="QPushButton" name="bg_button">
-                 <property name="maximumSize">
-                   <size>
-                     <width>40</width>
-                     <height>16777215</height>
-                   </size>
-                 </property>
-               </widget>
-             </item>
-           </layout>
-         </item>
+        <item>
+         <layout class="QHBoxLayout" name="horizontalLayout_6">
+          <item>
+           <widget class="QLabel" name="bg_label">
+            <property name="text">
+             <string>Background Color:</string>
+            </property>
+           </widget>
+          </item>
+          <item>
+           <widget class="QPushButton" name="bg_button">
+            <property name="maximumSize">
+             <size>
+              <width>40</width>
+              <height>16777215</height>
+             </size>
+            </property>
+           </widget>
+          </item>
+         </layout>
+        </item>
        </layout>
       </widget>
      </item>

src/yuzu/main.cpp

@@ -518,32 +518,18 @@ void GMainWindow::OnDisplayTitleBars(bool show) {
 QStringList GMainWindow::GetUnsupportedGLExtensions() {
     QStringList unsupported_ext;
 
-    if (!GLAD_GL_ARB_program_interface_query)
-        unsupported_ext.append("ARB_program_interface_query");
-    if (!GLAD_GL_ARB_separate_shader_objects)
-        unsupported_ext.append("ARB_separate_shader_objects");
-    if (!GLAD_GL_ARB_vertex_attrib_binding)
-        unsupported_ext.append("ARB_vertex_attrib_binding");
     if (!GLAD_GL_ARB_vertex_type_10f_11f_11f_rev)
         unsupported_ext.append("ARB_vertex_type_10f_11f_11f_rev");
     if (!GLAD_GL_ARB_texture_mirror_clamp_to_edge)
         unsupported_ext.append("ARB_texture_mirror_clamp_to_edge");
-    if (!GLAD_GL_ARB_base_instance)
-        unsupported_ext.append("ARB_base_instance");
-    if (!GLAD_GL_ARB_texture_storage)
-        unsupported_ext.append("ARB_texture_storage");
     if (!GLAD_GL_ARB_multi_bind)
         unsupported_ext.append("ARB_multi_bind");
-    if (!GLAD_GL_ARB_copy_image)
-        unsupported_ext.append("ARB_copy_image");
 
     // Extensions required to support some texture formats.
     if (!GLAD_GL_EXT_texture_compression_s3tc)
         unsupported_ext.append("EXT_texture_compression_s3tc");
     if (!GLAD_GL_ARB_texture_compression_rgtc)
         unsupported_ext.append("ARB_texture_compression_rgtc");
-    if (!GLAD_GL_ARB_texture_compression_bptc)
-        unsupported_ext.append("ARB_texture_compression_bptc");
     if (!GLAD_GL_ARB_depth_buffer_float)
         unsupported_ext.append("ARB_depth_buffer_float");
 
@@ -562,8 +548,8 @@ bool GMainWindow::LoadROM(const QString& filename) {
     render_window->MakeCurrent();
 
     if (!gladLoadGL()) {
-        QMessageBox::critical(this, tr("Error while initializing OpenGL 3.3 Core!"),
-                              tr("Your GPU may not support OpenGL 3.3, or you do not "
+        QMessageBox::critical(this, tr("Error while initializing OpenGL 4.3 Core!"),
+                              tr("Your GPU may not support OpenGL 4.3, or you do not "
                                  "have the latest graphics driver."));
         return false;
     }

src/yuzu_cmd/config.cpp

@@ -366,6 +366,7 @@ void Config::ReadValues() {
     Settings::values.gdbstub_port =
         static_cast<u16>(sdl2_config->GetInteger("Debugging", "gdbstub_port", 24689));
     Settings::values.program_args = sdl2_config->Get("Debugging", "program_args", "");
+    Settings::values.dump_exefs = sdl2_config->GetBoolean("Debugging", "dump_exefs", false);
     Settings::values.dump_nso = sdl2_config->GetBoolean("Debugging", "dump_nso", false);
 
     // Web Service

src/yuzu_cmd/default_ini.h

@@ -206,6 +206,8 @@ log_filter = *:Trace
 # Port for listening to GDB connections.
 use_gdbstub=false
 gdbstub_port=24689
+# Determines whether or not yuzu will dump the ExeFS of all games it attempts to load while loading them
+dump_exefs=false
 # Determines whether or not yuzu will dump all NSOs it attempts to load while loading them
 dump_nso=false
 

src/yuzu_cmd/emu_window/emu_window_sdl2.cpp

@@ -111,32 +111,18 @@ void EmuWindow_SDL2::Fullscreen() {
 bool EmuWindow_SDL2::SupportsRequiredGLExtensions() {
     std::vector<std::string> unsupported_ext;
 
-    if (!GLAD_GL_ARB_program_interface_query)
-        unsupported_ext.push_back("ARB_program_interface_query");
-    if (!GLAD_GL_ARB_separate_shader_objects)
-        unsupported_ext.push_back("ARB_separate_shader_objects");
-    if (!GLAD_GL_ARB_vertex_attrib_binding)
-        unsupported_ext.push_back("ARB_vertex_attrib_binding");
     if (!GLAD_GL_ARB_vertex_type_10f_11f_11f_rev)
         unsupported_ext.push_back("ARB_vertex_type_10f_11f_11f_rev");
     if (!GLAD_GL_ARB_texture_mirror_clamp_to_edge)
         unsupported_ext.push_back("ARB_texture_mirror_clamp_to_edge");
-    if (!GLAD_GL_ARB_base_instance)
-        unsupported_ext.push_back("ARB_base_instance");
-    if (!GLAD_GL_ARB_texture_storage)
-        unsupported_ext.push_back("ARB_texture_storage");
     if (!GLAD_GL_ARB_multi_bind)
         unsupported_ext.push_back("ARB_multi_bind");
-    if (!GLAD_GL_ARB_copy_image)
-        unsupported_ext.push_back("ARB_copy_image");
 
     // Extensions required to support some texture formats.
     if (!GLAD_GL_EXT_texture_compression_s3tc)
         unsupported_ext.push_back("EXT_texture_compression_s3tc");
     if (!GLAD_GL_ARB_texture_compression_rgtc)
         unsupported_ext.push_back("ARB_texture_compression_rgtc");
-    if (!GLAD_GL_ARB_texture_compression_bptc)
-        unsupported_ext.push_back("ARB_texture_compression_bptc");
     if (!GLAD_GL_ARB_depth_buffer_float)
         unsupported_ext.push_back("ARB_depth_buffer_float");
 
@@ -157,7 +143,7 @@ EmuWindow_SDL2::EmuWindow_SDL2(bool fullscreen) {
         exit(1);
     }
 
-    SDL_GL_SetAttribute(SDL_GL_CONTEXT_MAJOR_VERSION, 3);
+    SDL_GL_SetAttribute(SDL_GL_CONTEXT_MAJOR_VERSION, 4);
     SDL_GL_SetAttribute(SDL_GL_CONTEXT_MINOR_VERSION, 3);
     SDL_GL_SetAttribute(SDL_GL_CONTEXT_PROFILE_MASK, SDL_GL_CONTEXT_PROFILE_CORE);
     SDL_GL_SetAttribute(SDL_GL_DOUBLEBUFFER, 1);