config: Removed setting icon config

rasterizer_interface: Remove renderer-specific ScreenInfo type from AccelerateDraw() in RasterizerInterface

src/common/hex_util.cpp

@@ -4,6 +4,8 @@
 
 #include "common/hex_util.h"
 
+namespace Common {
+
 u8 ToHexNibble(char c1) {
     if (c1 >= 65 && c1 <= 70)
         return c1 - 55;
@@ -25,3 +27,5 @@ std::array<u8, 32> operator""_array32(const char* str, size_t len) {
         throw std::logic_error("Not of correct size.");
     return HexStringToArray<32>(str);
 }
+
+} // namespace Common

src/common/hex_util.h

@@ -10,6 +10,8 @@
 #include <fmt/format.h>
 #include "common/common_types.h"
 
+namespace Common {
+
 u8 ToHexNibble(char c1);
 
 template <size_t Size, bool le = false>
@@ -35,3 +37,5 @@ std::string HexArrayToString(std::array<u8, Size> array, bool upper = true) {
 
 std::array<u8, 0x10> operator"" _array16(const char* str, size_t len);
 std::array<u8, 0x20> operator"" _array32(const char* str, size_t len);
+
+} // namespace Common

src/common/telemetry.cpp

@@ -3,8 +3,15 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
+#include <cstring>
+#include "common/assert.h"
+#include "common/scm_rev.h"
 #include "common/telemetry.h"
 
+#ifdef ARCHITECTURE_x86_64
+#include "common/x64/cpu_detect.h"
+#endif
+
 namespace Telemetry {
 
 void FieldCollection::Accept(VisitorInterface& visitor) const {
@@ -37,4 +44,62 @@ template class Field<std::string>;
 template class Field<const char*>;
 template class Field<std::chrono::microseconds>;
 
+#ifdef ARCHITECTURE_x86_64
+static const char* CpuVendorToStr(Common::CPUVendor vendor) {
+    switch (vendor) {
+    case Common::CPUVendor::INTEL:
+        return "Intel";
+    case Common::CPUVendor::AMD:
+        return "Amd";
+    case Common::CPUVendor::OTHER:
+        return "Other";
+    }
+    UNREACHABLE();
+}
+#endif
+
+void AppendBuildInfo(FieldCollection& fc) {
+    const bool is_git_dirty{std::strstr(Common::g_scm_desc, "dirty") != nullptr};
+    fc.AddField(FieldType::App, "Git_IsDirty", is_git_dirty);
+    fc.AddField(FieldType::App, "Git_Branch", Common::g_scm_branch);
+    fc.AddField(FieldType::App, "Git_Revision", Common::g_scm_rev);
+    fc.AddField(FieldType::App, "BuildDate", Common::g_build_date);
+    fc.AddField(FieldType::App, "BuildName", Common::g_build_name);
+}
+
+void AppendCPUInfo(FieldCollection& fc) {
+#ifdef ARCHITECTURE_x86_64
+    fc.AddField(FieldType::UserSystem, "CPU_Model", Common::GetCPUCaps().cpu_string);
+    fc.AddField(FieldType::UserSystem, "CPU_BrandString", Common::GetCPUCaps().brand_string);
+    fc.AddField(FieldType::UserSystem, "CPU_Vendor", CpuVendorToStr(Common::GetCPUCaps().vendor));
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_AES", Common::GetCPUCaps().aes);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_AVX", Common::GetCPUCaps().avx);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_AVX2", Common::GetCPUCaps().avx2);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_BMI1", Common::GetCPUCaps().bmi1);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_BMI2", Common::GetCPUCaps().bmi2);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_FMA", Common::GetCPUCaps().fma);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_FMA4", Common::GetCPUCaps().fma4);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_SSE", Common::GetCPUCaps().sse);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_SSE2", Common::GetCPUCaps().sse2);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_SSE3", Common::GetCPUCaps().sse3);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_SSSE3", Common::GetCPUCaps().ssse3);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_SSE41", Common::GetCPUCaps().sse4_1);
+    fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_SSE42", Common::GetCPUCaps().sse4_2);
+#else
+    fc.AddField(FieldType::UserSystem, "CPU_Model", "Other");
+#endif
+}
+
+void AppendOSInfo(FieldCollection& fc) {
+#ifdef __APPLE__
+    fc.AddField(FieldType::UserSystem, "OsPlatform", "Apple");
+#elif defined(_WIN32)
+    fc.AddField(FieldType::UserSystem, "OsPlatform", "Windows");
+#elif defined(__linux__) || defined(linux) || defined(__linux)
+    fc.AddField(FieldType::UserSystem, "OsPlatform", "Linux");
+#else
+    fc.AddField(FieldType::UserSystem, "OsPlatform", "Unknown");
+#endif
+}
+
 } // namespace Telemetry

src/common/telemetry.h

@@ -180,4 +180,16 @@ struct NullVisitor : public VisitorInterface {
     void Complete() override {}
 };
 
+/// Appends build-specific information to the given FieldCollection,
+/// such as branch name, revision hash, etc.
+void AppendBuildInfo(FieldCollection& fc);
+
+/// Appends CPU-specific information to the given FieldCollection,
+/// such as instruction set extensions, etc.
+void AppendCPUInfo(FieldCollection& fc);
+
+/// Appends OS-specific information to the given FieldCollection,
+/// such as platform name, etc.
+void AppendOSInfo(FieldCollection& fc);
+
 } // namespace Telemetry

src/core/CMakeLists.txt

@@ -122,6 +122,8 @@ add_library(core STATIC
     hle/service/acc/acc_u0.h
     hle/service/acc/acc_u1.cpp
     hle/service/acc/acc_u1.h
+    hle/service/acc/profile_manager.cpp
+    hle/service/acc/profile_manager.h
     hle/service/am/am.cpp
     hle/service/am/am.h
     hle/service/am/applet_ae.cpp

src/core/arm/dynarmic/arm_dynarmic.cpp

@@ -134,6 +134,9 @@ std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const {
     config.dczid_el0 = 4;
     config.ctr_el0 = 0x8444c004;
 
+    // Unpredictable instructions
+    config.define_unpredictable_behaviour = true;
+
     return std::make_unique<Dynarmic::A64::Jit>(config);
 }
 

src/core/crypto/key_manager.cpp

@@ -52,20 +52,20 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {
         out[1].erase(std::remove(out[1].begin(), out[1].end(), ' '), out[1].end());
 
         if (is_title_keys) {
-            auto rights_id_raw = HexStringToArray<16>(out[0]);
+            auto rights_id_raw = Common::HexStringToArray<16>(out[0]);
             u128 rights_id{};
             std::memcpy(rights_id.data(), rights_id_raw.data(), rights_id_raw.size());
-            Key128 key = HexStringToArray<16>(out[1]);
+            Key128 key = Common::HexStringToArray<16>(out[1]);
             SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
         } else {
             std::transform(out[0].begin(), out[0].end(), out[0].begin(), ::tolower);
             if (s128_file_id.find(out[0]) != s128_file_id.end()) {
                 const auto index = s128_file_id.at(out[0]);
-                Key128 key = HexStringToArray<16>(out[1]);
+                Key128 key = Common::HexStringToArray<16>(out[1]);
                 SetKey(index.type, key, index.field1, index.field2);
             } else if (s256_file_id.find(out[0]) != s256_file_id.end()) {
                 const auto index = s256_file_id.at(out[0]);
-                Key256 key = HexStringToArray<32>(out[1]);
+                Key256 key = Common::HexStringToArray<32>(out[1]);
                 SetKey(index.type, key, index.field1, index.field2);
             }
         }

src/core/file_sys/content_archive.h

@@ -27,6 +27,7 @@ enum class NCAContentType : u8 {
     Control = 2,
     Manual = 3,
     Data = 4,
+    Data_Unknown5 = 5, ///< Seems to be used on some system archives
 };
 
 enum class NCASectionCryptoType : u8 {

src/core/file_sys/nca_metadata.h

@@ -7,6 +7,7 @@
 #include <cstring>
 #include <memory>
 #include <vector>
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/swap.h"
 #include "core/file_sys/vfs.h"

src/core/file_sys/registered_cache.cpp

@@ -37,11 +37,12 @@ static bool FollowsNcaIdFormat(std::string_view name) {
 static std::string GetRelativePathFromNcaID(const std::array<u8, 16>& nca_id, bool second_hex_upper,
                                             bool within_two_digit) {
     if (!within_two_digit)
-        return fmt::format("/{}.nca", HexArrayToString(nca_id, second_hex_upper));
+        return fmt::format("/{}.nca", Common::HexArrayToString(nca_id, second_hex_upper));
 
     Core::Crypto::SHA256Hash hash{};
     mbedtls_sha256(nca_id.data(), nca_id.size(), hash.data(), 0);
-    return fmt::format("/000000{:02X}/{}.nca", hash[0], HexArrayToString(nca_id, second_hex_upper));
+    return fmt::format("/000000{:02X}/{}.nca", hash[0],
+                       Common::HexArrayToString(nca_id, second_hex_upper));
 }
 
 static std::string GetCNMTName(TitleType type, u64 title_id) {
@@ -76,12 +77,13 @@ static ContentRecordType GetCRTypeFromNCAType(NCAContentType type) {
     case NCAContentType::Control:
         return ContentRecordType::Control;
     case NCAContentType::Data:
+    case NCAContentType::Data_Unknown5:
         return ContentRecordType::Data;
     case NCAContentType::Manual:
         // TODO(DarkLordZach): Peek at NCA contents to differentiate Manual and Legal.
         return ContentRecordType::Manual;
     default:
-        UNREACHABLE();
+        UNREACHABLE_MSG("Invalid NCAContentType={:02X}", static_cast<u8>(type));
     }
 }
 
@@ -170,7 +172,7 @@ std::vector<NcaID> RegisteredCache::AccumulateFiles() const {
     std::vector<NcaID> ids;
     for (const auto& d2_dir : dir->GetSubdirectories()) {
         if (FollowsNcaIdFormat(d2_dir->GetName())) {
-            ids.push_back(HexStringToArray<0x10, true>(d2_dir->GetName().substr(0, 0x20)));
+            ids.push_back(Common::HexStringToArray<0x10, true>(d2_dir->GetName().substr(0, 0x20)));
             continue;
         }
 
@@ -181,20 +183,21 @@ std::vector<NcaID> RegisteredCache::AccumulateFiles() const {
             if (!FollowsNcaIdFormat(nca_dir->GetName()))
                 continue;
 
-            ids.push_back(HexStringToArray<0x10, true>(nca_dir->GetName().substr(0, 0x20)));
+            ids.push_back(Common::HexStringToArray<0x10, true>(nca_dir->GetName().substr(0, 0x20)));
         }
 
         for (const auto& nca_file : d2_dir->GetFiles()) {
             if (!FollowsNcaIdFormat(nca_file->GetName()))
                 continue;
 
-            ids.push_back(HexStringToArray<0x10, true>(nca_file->GetName().substr(0, 0x20)));
+            ids.push_back(
+                Common::HexStringToArray<0x10, true>(nca_file->GetName().substr(0, 0x20)));
         }
     }
 
     for (const auto& d2_file : dir->GetFiles()) {
         if (FollowsNcaIdFormat(d2_file->GetName()))
-            ids.push_back(HexStringToArray<0x10, true>(d2_file->GetName().substr(0, 0x20)));
+            ids.push_back(Common::HexStringToArray<0x10, true>(d2_file->GetName().substr(0, 0x20)));
     }
     return ids;
 }
@@ -339,7 +342,7 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter(
 }
 
 static std::shared_ptr<NCA> GetNCAFromXCIForID(std::shared_ptr<XCI> xci, const NcaID& id) {
-    const auto filename = fmt::format("{}.nca", HexArrayToString(id, false));
+    const auto filename = fmt::format("{}.nca", Common::HexArrayToString(id, false));
     const auto iter =
         std::find_if(xci->GetNCAs().begin(), xci->GetNCAs().end(),
                      [&filename](std::shared_ptr<NCA> nca) { return nca->GetName() == filename; });
@@ -361,7 +364,7 @@ InstallResult RegisteredCache::InstallEntry(std::shared_ptr<XCI> xci, bool overw
 
     // Install Metadata File
     const auto meta_id_raw = (*meta_iter)->GetName().substr(0, 32);
-    const auto meta_id = HexStringToArray<16>(meta_id_raw);
+    const auto meta_id = Common::HexStringToArray<16>(meta_id_raw);
 
     const auto res = RawInstallNCA(*meta_iter, copy, overwrite_if_exists, meta_id);
     if (res != InstallResult::Success)

src/core/file_sys/romfs_factory.cpp

@@ -6,20 +6,57 @@
 #include <memory>
 #include "common/common_types.h"
 #include "common/logging/log.h"
+#include "core/file_sys/nca_metadata.h"
+#include "core/file_sys/registered_cache.h"
 #include "core/file_sys/romfs_factory.h"
+#include "core/hle/kernel/process.h"
+#include "core/hle/service/filesystem/filesystem.h"
+#include "core/loader/loader.h"
 
 namespace FileSys {
 
 RomFSFactory::RomFSFactory(Loader::AppLoader& app_loader) {
     // Load the RomFS from the app
-    if (Loader::ResultStatus::Success != app_loader.ReadRomFS(file)) {
+    if (app_loader.ReadRomFS(file) != Loader::ResultStatus::Success) {
         LOG_ERROR(Service_FS, "Unable to read RomFS!");
     }
 }
 
-ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id) {
-    // TODO(DarkLordZach): Use title id.
+ResultVal<VirtualFile> RomFSFactory::OpenCurrentProcess() {
     return MakeResult<VirtualFile>(file);
 }
 
+ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage, ContentRecordType type) {
+    switch (storage) {
+    case StorageId::NandSystem: {
+        const auto res = Service::FileSystem::GetSystemNANDContents()->GetEntry(title_id, type);
+        if (res == nullptr) {
+            // TODO(DarkLordZach): Find the right error code to use here
+            return ResultCode(-1);
+        }
+        const auto romfs = res->GetRomFS();
+        if (romfs == nullptr) {
+            // TODO(DarkLordZach): Find the right error code to use here
+            return ResultCode(-1);
+        }
+        return MakeResult<VirtualFile>(romfs);
+    }
+    case StorageId::NandUser: {
+        const auto res = Service::FileSystem::GetUserNANDContents()->GetEntry(title_id, type);
+        if (res == nullptr) {
+            // TODO(DarkLordZach): Find the right error code to use here
+            return ResultCode(-1);
+        }
+        const auto romfs = res->GetRomFS();
+        if (romfs == nullptr) {
+            // TODO(DarkLordZach): Find the right error code to use here
+            return ResultCode(-1);
+        }
+        return MakeResult<VirtualFile>(romfs);
+    }
+    default:
+        UNIMPLEMENTED_MSG("Unimplemented storage_id={:02X}", static_cast<u8>(storage));
+    }
+}
+
 } // namespace FileSys

src/core/file_sys/romfs_factory.h

@@ -6,17 +6,33 @@
 
 #include <memory>
 #include "common/common_types.h"
+#include "core/file_sys/vfs.h"
 #include "core/hle/result.h"
-#include "core/loader/loader.h"
+
+namespace Loader {
+class AppLoader;
+} // namespace Loader
 
 namespace FileSys {
 
+enum class ContentRecordType : u8;
+
+enum class StorageId : u8 {
+    None = 0,
+    Host = 1,
+    GameCard = 2,
+    NandSystem = 3,
+    NandUser = 4,
+    SdCard = 5,
+};
+
 /// File system interface to the RomFS archive
 class RomFSFactory {
 public:
     explicit RomFSFactory(Loader::AppLoader& app_loader);
 
-    ResultVal<VirtualFile> Open(u64 title_id);
+    ResultVal<VirtualFile> OpenCurrentProcess();
+    ResultVal<VirtualFile> Open(u64 title_id, StorageId storage, ContentRecordType type);
 
 private:
     VirtualFile file;

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

@@ -3,17 +3,19 @@
 // Refer to the license.txt file included.
 
 #include <array>
+#include "common/common_types.h"
 #include "common/logging/log.h"
+#include "common/swap.h"
+#include "core/core_timing.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/service/acc/acc.h"
 #include "core/hle/service/acc/acc_aa.h"
 #include "core/hle/service/acc/acc_su.h"
 #include "core/hle/service/acc/acc_u0.h"
 #include "core/hle/service/acc/acc_u1.h"
-#include "core/settings.h"
+#include "core/hle/service/acc/profile_manager.h"
 
 namespace Service::Account {
-
 // TODO: RE this structure
 struct UserData {
     INSERT_PADDING_WORDS(1);
@@ -25,19 +27,10 @@ struct UserData {
 };
 static_assert(sizeof(UserData) == 0x80, "UserData structure has incorrect size");
 
-struct ProfileBase {
-    u128 user_id;
-    u64 timestamp;
-    std::array<u8, 0x20> username;
-};
-static_assert(sizeof(ProfileBase) == 0x38, "ProfileBase structure has incorrect size");
-
-// TODO(ogniK): Generate a real user id based on username, md5(username) maybe?
-static constexpr u128 DEFAULT_USER_ID{1ull, 0ull};
-
 class IProfile final : public ServiceFramework<IProfile> {
 public:
-    explicit IProfile(u128 user_id) : ServiceFramework("IProfile"), user_id(user_id) {
+    explicit IProfile(UUID user_id, ProfileManager& profile_manager)
+        : ServiceFramework("IProfile"), profile_manager(profile_manager), user_id(user_id) {
         static const FunctionInfo functions[] = {
             {0, &IProfile::Get, "Get"},
             {1, &IProfile::GetBase, "GetBase"},
@@ -49,46 +42,42 @@ public:
 
 private:
     void Get(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_ACC, "(STUBBED) called");
+        LOG_INFO(Service_ACC, "called user_id={}", user_id.Format());
         ProfileBase profile_base{};
-        profile_base.user_id = user_id;
-        if (Settings::values.username.size() > profile_base.username.size()) {
-            std::copy_n(Settings::values.username.begin(), profile_base.username.size(),
-                        profile_base.username.begin());
+        std::array<u8, MAX_DATA> data{};
+        if (profile_manager.GetProfileBaseAndData(user_id, profile_base, data)) {
+            ctx.WriteBuffer(data);
+            IPC::ResponseBuilder rb{ctx, 16};
+            rb.Push(RESULT_SUCCESS);
+            rb.PushRaw(profile_base);
         } else {
-            std::copy(Settings::values.username.begin(), Settings::values.username.end(),
-                      profile_base.username.begin());
+            LOG_ERROR(Service_ACC, "Failed to get profile base and data for user={}",
+                      user_id.Format());
+            IPC::ResponseBuilder rb{ctx, 2};
+            rb.Push(ResultCode(-1)); // TODO(ogniK): Get actual error code
         }
-
-        IPC::ResponseBuilder rb{ctx, 16};
-        rb.Push(RESULT_SUCCESS);
-        rb.PushRaw(profile_base);
     }
 
     void GetBase(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_ACC, "(STUBBED) called");
-
-        // TODO(Subv): Retrieve this information from somewhere.
+        LOG_INFO(Service_ACC, "called user_id={}", user_id.Format());
         ProfileBase profile_base{};
-        profile_base.user_id = user_id;
-        if (Settings::values.username.size() > profile_base.username.size()) {
-            std::copy_n(Settings::values.username.begin(), profile_base.username.size(),
-                        profile_base.username.begin());
+        if (profile_manager.GetProfileBase(user_id, profile_base)) {
+            IPC::ResponseBuilder rb{ctx, 16};
+            rb.Push(RESULT_SUCCESS);
+            rb.PushRaw(profile_base);
         } else {
-            std::copy(Settings::values.username.begin(), Settings::values.username.end(),
-                      profile_base.username.begin());
+            LOG_ERROR(Service_ACC, "Failed to get profile base for user={}", user_id.Format());
+            IPC::ResponseBuilder rb{ctx, 2};
+            rb.Push(ResultCode(-1)); // TODO(ogniK): Get actual error code
         }
-        IPC::ResponseBuilder rb{ctx, 16};
-        rb.Push(RESULT_SUCCESS);
-        rb.PushRaw(profile_base);
     }
 
     void LoadImage(Kernel::HLERequestContext& ctx) {
         LOG_WARNING(Service_ACC, "(STUBBED) called");
         // smallest jpeg https://github.com/mathiasbynens/small/blob/master/jpeg.jpg
         // TODO(mailwl): load actual profile image from disk, width 256px, max size 0x20000
-        const u32 jpeg_size = 107;
-        static const std::array<u8, jpeg_size> jpeg{
+        constexpr u32 jpeg_size = 107;
+        static constexpr std::array<u8, jpeg_size> jpeg{
             0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03,
             0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04,
             0x08, 0x06, 0x06, 0x05, 0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a,
@@ -98,13 +87,14 @@ private:
             0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08, 0x01, 0x01,
             0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
         };
-        ctx.WriteBuffer(jpeg.data(), jpeg_size);
+        ctx.WriteBuffer(jpeg);
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push<u32>(jpeg_size);
     }
 
-    u128 user_id; ///< The user id this profile refers to.
+    const ProfileManager& profile_manager;
+    UUID user_id; ///< The user id this profile refers to.
 };
 
 class IManagerForApplication final : public ServiceFramework<IManagerForApplication> {
@@ -141,44 +131,57 @@ private:
 };
 
 void Module::Interface::GetUserCount(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_ACC, "(STUBBED) called");
+    LOG_INFO(Service_ACC, "called");
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
-    rb.Push<u32>(1);
+    rb.Push<u32>(static_cast<u32>(profile_manager->GetUserCount()));
 }
 
 void Module::Interface::GetUserExistence(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_ACC, "(STUBBED) called");
+    IPC::RequestParser rp{ctx};
+    UUID user_id = rp.PopRaw<UUID>();
+    LOG_INFO(Service_ACC, "called user_id={}", user_id.Format());
+
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
-    rb.Push(true); // TODO: Check when this is supposed to return true and when not
+    rb.Push(profile_manager->UserExists(user_id));
 }
 
 void Module::Interface::ListAllUsers(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_ACC, "(STUBBED) called");
-    // TODO(Subv): There is only one user for now.
-    const std::vector<u128> user_ids = {DEFAULT_USER_ID};
-    ctx.WriteBuffer(user_ids);
+    LOG_INFO(Service_ACC, "called");
+    ctx.WriteBuffer(profile_manager->GetAllUsers());
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(RESULT_SUCCESS);
 }
 
 void Module::Interface::ListOpenUsers(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_ACC, "(STUBBED) called");
-    // TODO(Subv): There is only one user for now.
-    const std::vector<u128> user_ids = {DEFAULT_USER_ID};
-    ctx.WriteBuffer(user_ids);
+    LOG_INFO(Service_ACC, "called");
+    ctx.WriteBuffer(profile_manager->GetOpenUsers());
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(RESULT_SUCCESS);
 }
 
+void Module::Interface::GetLastOpenedUser(Kernel::HLERequestContext& ctx) {
+    LOG_INFO(Service_ACC, "called");
+    IPC::ResponseBuilder rb{ctx, 6};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushRaw<UUID>(profile_manager->GetLastOpenedUser());
+}
+
 void Module::Interface::GetProfile(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
-    u128 user_id = rp.PopRaw<u128>();
+    UUID user_id = rp.PopRaw<UUID>();
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<IProfile>(user_id);
-    LOG_DEBUG(Service_ACC, "called user_id=0x{:016X}{:016X}", user_id[1], user_id[0]);
+    rb.PushIpcInterface<IProfile>(user_id, *profile_manager);
+    LOG_DEBUG(Service_ACC, "called user_id={}", user_id.Format());
+}
+
+void Module::Interface::IsUserRegistrationRequestPermitted(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_ACC, "(STUBBED) called");
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push(profile_manager->CanSystemRegisterUser());
 }
 
 void Module::Interface::InitializeApplicationInfo(Kernel::HLERequestContext& ctx) {
@@ -194,22 +197,20 @@ void Module::Interface::GetBaasAccountManagerForApplication(Kernel::HLERequestCo
     LOG_DEBUG(Service_ACC, "called");
 }
 
-void Module::Interface::GetLastOpenedUser(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_ACC, "(STUBBED) called");
-    IPC::ResponseBuilder rb{ctx, 6};
-    rb.Push(RESULT_SUCCESS);
-    rb.PushRaw(DEFAULT_USER_ID);
-}
+Module::Interface::Interface(std::shared_ptr<Module> module,
+                             std::shared_ptr<ProfileManager> profile_manager, const char* name)
+    : ServiceFramework(name), module(std::move(module)),
+      profile_manager(std::move(profile_manager)) {}
 
-Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
-    : ServiceFramework(name), module(std::move(module)) {}
+Module::Interface::~Interface() = default;
 
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     auto module = std::make_shared<Module>();
-    std::make_shared<ACC_AA>(module)->InstallAsService(service_manager);
-    std::make_shared<ACC_SU>(module)->InstallAsService(service_manager);
-    std::make_shared<ACC_U0>(module)->InstallAsService(service_manager);
-    std::make_shared<ACC_U1>(module)->InstallAsService(service_manager);
+    auto profile_manager = std::make_shared<ProfileManager>();
+    std::make_shared<ACC_AA>(module, profile_manager)->InstallAsService(service_manager);
+    std::make_shared<ACC_SU>(module, profile_manager)->InstallAsService(service_manager);
+    std::make_shared<ACC_U0>(module, profile_manager)->InstallAsService(service_manager);
+    std::make_shared<ACC_U1>(module, profile_manager)->InstallAsService(service_manager);
 }
 
 } // namespace Service::Account

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

@@ -8,11 +8,15 @@
 
 namespace Service::Account {
 
+class ProfileManager;
+
 class Module final {
 public:
     class Interface : public ServiceFramework<Interface> {
     public:
-        explicit Interface(std::shared_ptr<Module> module, const char* name);
+        explicit Interface(std::shared_ptr<Module> module,
+                           std::shared_ptr<ProfileManager> profile_manager, const char* name);
+        ~Interface() override;
 
         void GetUserCount(Kernel::HLERequestContext& ctx);
         void GetUserExistence(Kernel::HLERequestContext& ctx);
@@ -22,9 +26,11 @@ public:
         void GetProfile(Kernel::HLERequestContext& ctx);
         void InitializeApplicationInfo(Kernel::HLERequestContext& ctx);
         void GetBaasAccountManagerForApplication(Kernel::HLERequestContext& ctx);
+        void IsUserRegistrationRequestPermitted(Kernel::HLERequestContext& ctx);
 
     protected:
         std::shared_ptr<Module> module;
+        std::shared_ptr<ProfileManager> profile_manager;
     };
 };
 

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

@@ -6,7 +6,8 @@
 
 namespace Service::Account {
 
-ACC_AA::ACC_AA(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "acc:aa") {
+ACC_AA::ACC_AA(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> profile_manager)
+    : Module::Interface(std::move(module), std::move(profile_manager), "acc:aa") {
     static const FunctionInfo functions[] = {
         {0, nullptr, "EnsureCacheAsync"},
         {1, nullptr, "LoadCache"},

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

@@ -10,7 +10,8 @@ namespace Service::Account {
 
 class ACC_AA final : public Module::Interface {
 public:
-    explicit ACC_AA(std::shared_ptr<Module> module);
+    explicit ACC_AA(std::shared_ptr<Module> module,
+                    std::shared_ptr<ProfileManager> profile_manager);
 };
 
 } // namespace Service::Account

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

@@ -6,7 +6,8 @@
 
 namespace Service::Account {
 
-ACC_SU::ACC_SU(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "acc:su") {
+ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> profile_manager)
+    : Module::Interface(std::move(module), std::move(profile_manager), "acc:su") {
     static const FunctionInfo functions[] = {
         {0, &ACC_SU::GetUserCount, "GetUserCount"},
         {1, &ACC_SU::GetUserExistence, "GetUserExistence"},
@@ -15,7 +16,7 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module) : Module::Interface(std::move(mod
         {4, &ACC_SU::GetLastOpenedUser, "GetLastOpenedUser"},
         {5, &ACC_SU::GetProfile, "GetProfile"},
         {6, nullptr, "GetProfileDigest"},
-        {50, nullptr, "IsUserRegistrationRequestPermitted"},
+        {50, &ACC_SU::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"},
         {51, nullptr, "TrySelectUserWithoutInteraction"},
         {60, nullptr, "ListOpenContextStoredUsers"},
         {100, nullptr, "GetUserRegistrationNotifier"},

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

@@ -11,7 +11,8 @@ namespace Account {
 
 class ACC_SU final : public Module::Interface {
 public:
-    explicit ACC_SU(std::shared_ptr<Module> module);
+    explicit ACC_SU(std::shared_ptr<Module> module,
+                    std::shared_ptr<ProfileManager> profile_manager);
 };
 
 } // namespace Account

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

@@ -6,7 +6,8 @@
 
 namespace Service::Account {
 
-ACC_U0::ACC_U0(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "acc:u0") {
+ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> profile_manager)
+    : Module::Interface(std::move(module), std::move(profile_manager), "acc:u0") {
     static const FunctionInfo functions[] = {
         {0, &ACC_U0::GetUserCount, "GetUserCount"},
         {1, &ACC_U0::GetUserExistence, "GetUserExistence"},
@@ -15,7 +16,7 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module) : Module::Interface(std::move(mod
         {4, &ACC_U0::GetLastOpenedUser, "GetLastOpenedUser"},
         {5, &ACC_U0::GetProfile, "GetProfile"},
         {6, nullptr, "GetProfileDigest"},
-        {50, nullptr, "IsUserRegistrationRequestPermitted"},
+        {50, &ACC_U0::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"},
         {51, nullptr, "TrySelectUserWithoutInteraction"},
         {60, nullptr, "ListOpenContextStoredUsers"},
         {100, &ACC_U0::InitializeApplicationInfo, "InitializeApplicationInfo"},

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

@@ -10,7 +10,8 @@ namespace Service::Account {
 
 class ACC_U0 final : public Module::Interface {
 public:
-    explicit ACC_U0(std::shared_ptr<Module> module);
+    explicit ACC_U0(std::shared_ptr<Module> module,
+                    std::shared_ptr<ProfileManager> profile_manager);
 };
 
 } // namespace Service::Account

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

@@ -6,7 +6,8 @@
 
 namespace Service::Account {
 
-ACC_U1::ACC_U1(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "acc:u1") {
+ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> profile_manager)
+    : Module::Interface(std::move(module), std::move(profile_manager), "acc:u1") {
     static const FunctionInfo functions[] = {
         {0, &ACC_U1::GetUserCount, "GetUserCount"},
         {1, &ACC_U1::GetUserExistence, "GetUserExistence"},
@@ -15,7 +16,7 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module) : Module::Interface(std::move(mod
         {4, &ACC_U1::GetLastOpenedUser, "GetLastOpenedUser"},
         {5, &ACC_U1::GetProfile, "GetProfile"},
         {6, nullptr, "GetProfileDigest"},
-        {50, nullptr, "IsUserRegistrationRequestPermitted"},
+        {50, &ACC_U1::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"},
         {51, nullptr, "TrySelectUserWithoutInteraction"},
         {60, nullptr, "ListOpenContextStoredUsers"},
         {100, nullptr, "GetUserRegistrationNotifier"},

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

@@ -10,7 +10,8 @@ namespace Service::Account {
 
 class ACC_U1 final : public Module::Interface {
 public:
-    explicit ACC_U1(std::shared_ptr<Module> module);
+    explicit ACC_U1(std::shared_ptr<Module> module,
+                    std::shared_ptr<ProfileManager> profile_manager);
 };
 
 } // namespace Service::Account

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

@@ -0,0 +1,236 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <random>
+#include <boost/optional.hpp>
+#include "core/hle/service/acc/profile_manager.h"
+#include "core/settings.h"
+
+namespace Service::Account {
+// TODO(ogniK): Get actual error codes
+constexpr ResultCode ERROR_TOO_MANY_USERS(ErrorModule::Account, -1);
+constexpr ResultCode ERROR_USER_ALREADY_EXISTS(ErrorModule::Account, -2);
+constexpr ResultCode ERROR_ARGUMENT_IS_NULL(ErrorModule::Account, 20);
+
+const UUID& UUID::Generate() {
+    std::random_device device;
+    std::mt19937 gen(device());
+    std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max());
+    uuid[0] = distribution(gen);
+    uuid[1] = distribution(gen);
+    return *this;
+}
+
+ProfileManager::ProfileManager() {
+    // TODO(ogniK): Create the default user we have for now until loading/saving users is added
+    auto user_uuid = UUID{1, 0};
+    CreateNewUser(user_uuid, Settings::values.username);
+    OpenUser(user_uuid);
+}
+
+/// After a users creation it needs to be "registered" to the system. AddToProfiles handles the
+/// internal management of the users profiles
+boost::optional<size_t> ProfileManager::AddToProfiles(const ProfileInfo& user) {
+    if (user_count >= MAX_USERS) {
+        return boost::none;
+    }
+    profiles[user_count] = user;
+    return user_count++;
+}
+
+/// Deletes a specific profile based on it's profile index
+bool ProfileManager::RemoveProfileAtIndex(size_t index) {
+    if (index >= MAX_USERS || index >= user_count) {
+        return false;
+    }
+    if (index < user_count - 1) {
+        std::rotate(profiles.begin() + index, profiles.begin() + index + 1, profiles.end());
+    }
+    profiles.back() = {};
+    user_count--;
+    return true;
+}
+
+/// Helper function to register a user to the system
+ResultCode ProfileManager::AddUser(const ProfileInfo& user) {
+    if (AddToProfiles(user) == boost::none) {
+        return ERROR_TOO_MANY_USERS;
+    }
+    return RESULT_SUCCESS;
+}
+
+/// Create a new user on the system. If the uuid of the user already exists, the user is not
+/// created.
+ResultCode ProfileManager::CreateNewUser(UUID uuid, const ProfileUsername& username) {
+    if (user_count == MAX_USERS) {
+        return ERROR_TOO_MANY_USERS;
+    }
+    if (!uuid) {
+        return ERROR_ARGUMENT_IS_NULL;
+    }
+    if (username[0] == 0x0) {
+        return ERROR_ARGUMENT_IS_NULL;
+    }
+    if (std::any_of(profiles.begin(), profiles.end(),
+                    [&uuid](const ProfileInfo& profile) { return uuid == profile.user_uuid; })) {
+        return ERROR_USER_ALREADY_EXISTS;
+    }
+    ProfileInfo profile;
+    profile.user_uuid = uuid;
+    profile.username = username;
+    profile.data = {};
+    profile.creation_time = 0x0;
+    profile.is_open = false;
+    return AddUser(profile);
+}
+
+/// Creates a new user on the system. This function allows a much simpler method of registration
+/// specifically by allowing an std::string for the username. This is required specifically since
+/// we're loading a string straight from the config
+ResultCode ProfileManager::CreateNewUser(UUID uuid, const std::string& username) {
+    ProfileUsername username_output;
+    if (username.size() > username_output.size()) {
+        std::copy_n(username.begin(), username_output.size(), username_output.begin());
+    } else {
+        std::copy(username.begin(), username.end(), username_output.begin());
+    }
+    return CreateNewUser(uuid, username_output);
+}
+
+/// Returns a users profile index based on their user id.
+boost::optional<size_t> ProfileManager::GetUserIndex(const UUID& uuid) const {
+    if (!uuid) {
+        return boost::none;
+    }
+    auto iter = std::find_if(profiles.begin(), profiles.end(),
+                             [&uuid](const ProfileInfo& p) { return p.user_uuid == uuid; });
+    if (iter == profiles.end()) {
+        return boost::none;
+    }
+    return static_cast<size_t>(std::distance(profiles.begin(), iter));
+}
+
+/// Returns a users profile index based on their profile
+boost::optional<size_t> ProfileManager::GetUserIndex(const ProfileInfo& user) const {
+    return GetUserIndex(user.user_uuid);
+}
+
+/// Returns the data structure used by the switch when GetProfileBase is called on acc:*
+bool ProfileManager::GetProfileBase(boost::optional<size_t> index, ProfileBase& profile) const {
+    if (index == boost::none || index >= MAX_USERS) {
+        return false;
+    }
+    const auto& prof_info = profiles[index.get()];
+    profile.user_uuid = prof_info.user_uuid;
+    profile.username = prof_info.username;
+    profile.timestamp = prof_info.creation_time;
+    return true;
+}
+
+/// Returns the data structure used by the switch when GetProfileBase is called on acc:*
+bool ProfileManager::GetProfileBase(UUID uuid, ProfileBase& profile) const {
+    auto idx = GetUserIndex(uuid);
+    return GetProfileBase(idx, profile);
+}
+
+/// Returns the data structure used by the switch when GetProfileBase is called on acc:*
+bool ProfileManager::GetProfileBase(const ProfileInfo& user, ProfileBase& profile) const {
+    return GetProfileBase(user.user_uuid, profile);
+}
+
+/// Returns the current user count on the system. We keep a variable which tracks the count so we
+/// don't have to loop the internal profile array every call.
+size_t ProfileManager::GetUserCount() const {
+    return user_count;
+}
+
+/// Lists the current "opened" users on the system. Users are typically not open until they sign
+/// into something or pick a profile. As of right now users should all be open until qlaunch is
+/// booting
+size_t ProfileManager::GetOpenUserCount() const {
+    return std::count_if(profiles.begin(), profiles.end(),
+                         [](const ProfileInfo& p) { return p.is_open; });
+}
+
+/// Checks if a user id exists in our profile manager
+bool ProfileManager::UserExists(UUID uuid) const {
+    return (GetUserIndex(uuid) != boost::none);
+}
+
+/// Opens a specific user
+void ProfileManager::OpenUser(UUID uuid) {
+    auto idx = GetUserIndex(uuid);
+    if (idx == boost::none) {
+        return;
+    }
+    profiles[idx.get()].is_open = true;
+    last_opened_user = uuid;
+}
+
+/// Closes a specific user
+void ProfileManager::CloseUser(UUID uuid) {
+    auto idx = GetUserIndex(uuid);
+    if (idx == boost::none) {
+        return;
+    }
+    profiles[idx.get()].is_open = false;
+}
+
+/// Gets all valid user ids on the system
+UserIDArray ProfileManager::GetAllUsers() const {
+    UserIDArray output;
+    std::transform(profiles.begin(), profiles.end(), output.begin(),
+                   [](const ProfileInfo& p) { return p.user_uuid; });
+    return output;
+}
+
+/// Get all the open users on the system and zero out the rest of the data. This is specifically
+/// needed for GetOpenUsers and we need to ensure the rest of the output buffer is zero'd out
+UserIDArray ProfileManager::GetOpenUsers() const {
+    UserIDArray output;
+    std::transform(profiles.begin(), profiles.end(), output.begin(), [](const ProfileInfo& p) {
+        if (p.is_open)
+            return p.user_uuid;
+        return UUID{};
+    });
+    std::stable_partition(output.begin(), output.end(), [](const UUID& uuid) { return uuid; });
+    return output;
+}
+
+/// Returns the last user which was opened
+UUID ProfileManager::GetLastOpenedUser() const {
+    return last_opened_user;
+}
+
+/// Return the users profile base and the unknown arbitary data.
+bool ProfileManager::GetProfileBaseAndData(boost::optional<size_t> index, ProfileBase& profile,
+                                           ProfileData& data) const {
+    if (GetProfileBase(index, profile)) {
+        data = profiles[index.get()].data;
+        return true;
+    }
+    return false;
+}
+
+/// Return the users profile base and the unknown arbitary data.
+bool ProfileManager::GetProfileBaseAndData(UUID uuid, ProfileBase& profile,
+                                           ProfileData& data) const {
+    auto idx = GetUserIndex(uuid);
+    return GetProfileBaseAndData(idx, profile, data);
+}
+
+/// Return the users profile base and the unknown arbitary data.
+bool ProfileManager::GetProfileBaseAndData(const ProfileInfo& user, ProfileBase& profile,
+                                           ProfileData& data) const {
+    return GetProfileBaseAndData(user.user_uuid, profile, data);
+}
+
+/// Returns if the system is allowing user registrations or not
+bool ProfileManager::CanSystemRegisterUser() const {
+    return false; // TODO(ogniK): Games shouldn't have
+                  // access to user registration, when we
+    // emulate qlaunch. Update this to dynamically change.
+}
+
+}; // namespace Service::Account

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

@@ -0,0 +1,117 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+
+#include "boost/optional.hpp"
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/hle/result.h"
+
+namespace Service::Account {
+constexpr size_t MAX_USERS = 8;
+constexpr size_t MAX_DATA = 128;
+constexpr u128 INVALID_UUID{{0, 0}};
+
+struct UUID {
+    // UUIDs which are 0 are considered invalid!
+    u128 uuid = INVALID_UUID;
+    UUID() = default;
+    explicit UUID(const u128& id) : uuid{id} {}
+    explicit UUID(const u64 lo, const u64 hi) : uuid{{lo, hi}} {}
+
+    explicit operator bool() const {
+        return uuid != INVALID_UUID;
+    }
+
+    bool operator==(const UUID& rhs) const {
+        return uuid == rhs.uuid;
+    }
+
+    bool operator!=(const UUID& rhs) const {
+        return !operator==(rhs);
+    }
+
+    // TODO(ogniK): Properly generate uuids based on RFC-4122
+    const UUID& Generate();
+
+    // Set the UUID to {0,0} to be considered an invalid user
+    void Invalidate() {
+        uuid = INVALID_UUID;
+    }
+    std::string Format() const {
+        return fmt::format("0x{:016X}{:016X}", uuid[1], uuid[0]);
+    }
+};
+static_assert(sizeof(UUID) == 16, "UUID is an invalid size!");
+
+using ProfileUsername = std::array<u8, 0x20>;
+using ProfileData = std::array<u8, MAX_DATA>;
+using UserIDArray = std::array<UUID, MAX_USERS>;
+
+/// This holds general information about a users profile. This is where we store all the information
+/// based on a specific user
+struct ProfileInfo {
+    UUID user_uuid;
+    ProfileUsername username;
+    u64 creation_time;
+    ProfileData data; // TODO(ognik): Work out what this is
+    bool is_open;
+};
+
+struct ProfileBase {
+    UUID user_uuid;
+    u64_le timestamp;
+    ProfileUsername username;
+
+    // Zero out all the fields to make the profile slot considered "Empty"
+    void Invalidate() {
+        user_uuid.Invalidate();
+        timestamp = 0;
+        username.fill(0);
+    }
+};
+static_assert(sizeof(ProfileBase) == 0x38, "ProfileBase is an invalid size");
+
+/// The profile manager is used for handling multiple user profiles at once. It keeps track of open
+/// users, all the accounts registered on the "system" as well as fetching individual "ProfileInfo"
+/// objects
+class ProfileManager {
+public:
+    ProfileManager(); // TODO(ogniK): Load from system save
+    ResultCode AddUser(const ProfileInfo& user);
+    ResultCode CreateNewUser(UUID uuid, const ProfileUsername& username);
+    ResultCode CreateNewUser(UUID uuid, const std::string& username);
+    boost::optional<size_t> GetUserIndex(const UUID& uuid) const;
+    boost::optional<size_t> GetUserIndex(const ProfileInfo& user) const;
+    bool GetProfileBase(boost::optional<size_t> index, ProfileBase& profile) const;
+    bool GetProfileBase(UUID uuid, ProfileBase& profile) const;
+    bool GetProfileBase(const ProfileInfo& user, ProfileBase& profile) const;
+    bool GetProfileBaseAndData(boost::optional<size_t> index, ProfileBase& profile,
+                               ProfileData& data) const;
+    bool GetProfileBaseAndData(UUID uuid, ProfileBase& profile, ProfileData& data) const;
+    bool GetProfileBaseAndData(const ProfileInfo& user, ProfileBase& profile,
+                               ProfileData& data) const;
+    size_t GetUserCount() const;
+    size_t GetOpenUserCount() const;
+    bool UserExists(UUID uuid) const;
+    void OpenUser(UUID uuid);
+    void CloseUser(UUID uuid);
+    UserIDArray GetOpenUsers() const;
+    UserIDArray GetAllUsers() const;
+    UUID GetLastOpenedUser() const;
+
+    bool CanSystemRegisterUser() const;
+
+private:
+    std::array<ProfileInfo, MAX_USERS> profiles{};
+    size_t user_count = 0;
+    boost::optional<size_t> AddToProfiles(const ProfileInfo& profile);
+    bool RemoveProfileAtIndex(size_t index);
+    UUID last_opened_user{INVALID_UUID};
+};
+
+}; // namespace Service::Account

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

@@ -145,8 +145,8 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
         {51, nullptr, "ApproveToDisplay"},
         {60, nullptr, "OverrideAutoSleepTimeAndDimmingTime"},
         {61, nullptr, "SetMediaPlaybackState"},
-        {62, nullptr, "SetIdleTimeDetectionExtension"},
-        {63, nullptr, "GetIdleTimeDetectionExtension"},
+        {62, &ISelfController::SetIdleTimeDetectionExtension, "SetIdleTimeDetectionExtension"},
+        {63, &ISelfController::GetIdleTimeDetectionExtension, "GetIdleTimeDetectionExtension"},
         {64, nullptr, "SetInputDetectionSourceSet"},
         {65, nullptr, "ReportUserIsActive"},
         {66, nullptr, "GetCurrentIlluminance"},
@@ -281,6 +281,23 @@ void ISelfController::SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx)
     LOG_WARNING(Service_AM, "(STUBBED) called");
 }
 
+void ISelfController::SetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    idle_time_detection_extension = rp.Pop<u32>();
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
+void ISelfController::GetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push<u32>(idle_time_detection_extension);
+
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
 ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter") {
     static const FunctionInfo functions[] = {
         {0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"},
@@ -306,7 +323,8 @@ ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter"
         {52, nullptr, "SwitchLcdBacklight"},
         {55, nullptr, "IsInControllerFirmwareUpdateSection"},
         {60, nullptr, "GetDefaultDisplayResolution"},
-        {61, nullptr, "GetDefaultDisplayResolutionChangeEvent"},
+        {61, &ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent,
+         "GetDefaultDisplayResolutionChangeEvent"},
         {62, nullptr, "GetHdcpAuthenticationState"},
         {63, nullptr, "GetHdcpAuthenticationStateChangeEvent"},
     };
@@ -341,6 +359,16 @@ void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_AM, "(STUBBED) called");
 }
 
+void ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent(Kernel::HLERequestContext& ctx) {
+    event->Signal();
+
+    IPC::ResponseBuilder rb{ctx, 2, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushCopyObjects(event);
+
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+}
+
 void ICommonStateGetter::GetOperationMode(Kernel::HLERequestContext& ctx) {
     const bool use_docked_mode{Settings::values.use_docked_mode};
     IPC::ResponseBuilder rb{ctx, 3};

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

@@ -87,9 +87,12 @@ private:
     void CreateManagedDisplayLayer(Kernel::HLERequestContext& ctx);
     void SetScreenShotPermission(Kernel::HLERequestContext& ctx);
     void SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx);
+    void SetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx);
+    void GetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
     Kernel::SharedPtr<Kernel::Event> launchable_event;
+    u32 idle_time_detection_extension = 0;
 };
 
 class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> {
@@ -110,6 +113,7 @@ private:
     void GetEventHandle(Kernel::HLERequestContext& ctx);
     void ReceiveMessage(Kernel::HLERequestContext& ctx);
     void GetCurrentFocusState(Kernel::HLERequestContext& ctx);
+    void GetDefaultDisplayResolutionChangeEvent(Kernel::HLERequestContext& ctx);
     void GetOperationMode(Kernel::HLERequestContext& ctx);
     void GetPerformanceMode(Kernel::HLERequestContext& ctx);
 

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

@@ -7,7 +7,9 @@
 #include "common/assert.h"
 #include "common/file_util.h"
 #include "core/core.h"
+#include "core/file_sys/bis_factory.h"
 #include "core/file_sys/errors.h"
+#include "core/file_sys/romfs_factory.h"
 #include "core/file_sys/savedata_factory.h"
 #include "core/file_sys/sdmc_factory.h"
 #include "core/file_sys/vfs.h"
@@ -256,15 +258,28 @@ ResultCode RegisterBIS(std::unique_ptr<FileSys::BISFactory>&& factory) {
     return RESULT_SUCCESS;
 }
 
-ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id) {
-    LOG_TRACE(Service_FS, "Opening RomFS for title_id={:016X}", title_id);
+ResultVal<FileSys::VirtualFile> OpenRomFSCurrentProcess() {
+    LOG_TRACE(Service_FS, "Opening RomFS for current process");
 
     if (romfs_factory == nullptr) {
         // TODO(bunnei): Find a better error code for this
         return ResultCode(-1);
     }
 
-    return romfs_factory->Open(title_id);
+    return romfs_factory->OpenCurrentProcess();
+}
+
+ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId storage_id,
+                                          FileSys::ContentRecordType type) {
+    LOG_TRACE(Service_FS, "Opening RomFS for title_id={:016X}, storage_id={:02X}, type={:02X}",
+              title_id, static_cast<u8>(storage_id), static_cast<u8>(type));
+
+    if (romfs_factory == nullptr) {
+        // TODO(bunnei): Find a better error code for this
+        return ResultCode(-1);
+    }
+
+    return romfs_factory->Open(title_id, storage_id, type);
 }
 
 ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space,

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

@@ -6,14 +6,24 @@
 
 #include <memory>
 #include "common/common_types.h"
-#include "core/file_sys/bis_factory.h"
 #include "core/file_sys/directory.h"
 #include "core/file_sys/mode.h"
-#include "core/file_sys/romfs_factory.h"
-#include "core/file_sys/savedata_factory.h"
-#include "core/file_sys/sdmc_factory.h"
 #include "core/hle/result.h"
 
+namespace FileSys {
+class BISFactory;
+class RegisteredCache;
+class RomFSFactory;
+class SaveDataFactory;
+class SDMCFactory;
+
+enum class ContentRecordType : u8;
+enum class SaveDataSpaceId : u8;
+enum class StorageId : u8;
+
+struct SaveDataDescriptor;
+} // namespace FileSys
+
 namespace Service {
 
 namespace SM {
@@ -27,7 +37,9 @@ ResultCode RegisterSaveData(std::unique_ptr<FileSys::SaveDataFactory>&& factory)
 ResultCode RegisterSDMC(std::unique_ptr<FileSys::SDMCFactory>&& factory);
 ResultCode RegisterBIS(std::unique_ptr<FileSys::BISFactory>&& factory);
 
-ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id);
+ResultVal<FileSys::VirtualFile> OpenRomFSCurrentProcess();
+ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId storage_id,
+                                          FileSys::ContentRecordType type);
 ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space,
                                             FileSys::SaveDataDescriptor save_struct);
 ResultVal<FileSys::VirtualDir> OpenSDMC();

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

@@ -13,9 +13,11 @@
 #include "common/common_types.h"
 #include "common/logging/log.h"
 #include "common/string_util.h"
-#include "core/core.h"
 #include "core/file_sys/directory.h"
 #include "core/file_sys/errors.h"
+#include "core/file_sys/nca_metadata.h"
+#include "core/file_sys/savedata_factory.h"
+#include "core/file_sys/vfs.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/service/filesystem/filesystem.h"
@@ -23,15 +25,6 @@
 
 namespace Service::FileSystem {
 
-enum class StorageId : u8 {
-    None = 0,
-    Host = 1,
-    GameCard = 2,
-    NandSystem = 3,
-    NandUser = 4,
-    SdCard = 5,
-};
-
 class IStorage final : public ServiceFramework<IStorage> {
 public:
     explicit IStorage(FileSys::VirtualFile backend_)
@@ -467,7 +460,7 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") {
         {110, nullptr, "OpenContentStorageFileSystem"},
         {200, &FSP_SRV::OpenDataStorageByCurrentProcess, "OpenDataStorageByCurrentProcess"},
         {201, nullptr, "OpenDataStorageByProgramId"},
-        {202, nullptr, "OpenDataStorageByDataId"},
+        {202, &FSP_SRV::OpenDataStorageByDataId, "OpenDataStorageByDataId"},
         {203, &FSP_SRV::OpenRomStorage, "OpenRomStorage"},
         {400, nullptr, "OpenDeviceOperator"},
         {500, nullptr, "OpenSdCardDetectionEventNotifier"},
@@ -580,7 +573,7 @@ void FSP_SRV::GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) {
 void FSP_SRV::OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_FS, "called");
 
-    auto romfs = OpenRomFS(Core::System::GetInstance().CurrentProcess()->program_id);
+    auto romfs = OpenRomFSCurrentProcess();
     if (romfs.Failed()) {
         // TODO (bunnei): Find the right error code to use here
         LOG_CRITICAL(Service_FS, "no file system interface available!");
@@ -596,10 +589,37 @@ void FSP_SRV::OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx) {
     rb.PushIpcInterface<IStorage>(std::move(storage));
 }
 
+void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto storage_id = rp.PopRaw<FileSys::StorageId>();
+    const auto unknown = rp.PopRaw<u32>();
+    const auto title_id = rp.PopRaw<u64>();
+
+    LOG_DEBUG(Service_FS, "called with storage_id={:02X}, unknown={:08X}, title_id={:016X}",
+              static_cast<u8>(storage_id), unknown, title_id);
+
+    auto data = OpenRomFS(title_id, storage_id, FileSys::ContentRecordType::Data);
+    if (data.Failed()) {
+        // TODO(DarkLordZach): Find the right error code to use here
+        LOG_ERROR(Service_FS,
+                  "could not open data storage with title_id={:016X}, storage_id={:02X}", title_id,
+                  static_cast<u8>(storage_id));
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(ResultCode(-1));
+        return;
+    }
+
+    IStorage storage(std::move(data.Unwrap()));
+
+    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushIpcInterface<IStorage>(std::move(storage));
+}
+
 void FSP_SRV::OpenRomStorage(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
 
-    auto storage_id = rp.PopRaw<StorageId>();
+    auto storage_id = rp.PopRaw<FileSys::StorageId>();
     auto title_id = rp.PopRaw<u64>();
 
     LOG_DEBUG(Service_FS, "called with storage_id={:02X}, title_id={:016X}",

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

@@ -25,6 +25,7 @@ private:
     void MountSaveData(Kernel::HLERequestContext& ctx);
     void GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx);
     void OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx);
+    void OpenDataStorageByDataId(Kernel::HLERequestContext& ctx);
     void OpenRomStorage(Kernel::HLERequestContext& ctx);
 
     FileSys::VirtualFile romfs;

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

@@ -14,7 +14,8 @@ public:
     IParentalControlService() : ServiceFramework("IParentalControlService") {
         static const FunctionInfo functions[] = {
             {1, &IParentalControlService::Initialize, "Initialize"},
-            {1001, nullptr, "CheckFreeCommunicationPermission"},
+            {1001, &IParentalControlService::CheckFreeCommunicationPermission,
+             "CheckFreeCommunicationPermission"},
             {1002, nullptr, "ConfirmLaunchApplicationPermission"},
             {1003, nullptr, "ConfirmResumeApplicationPermission"},
             {1004, nullptr, "ConfirmSnsPostPermission"},
@@ -116,6 +117,12 @@ private:
         IPC::ResponseBuilder rb{ctx, 2, 0, 0};
         rb.Push(RESULT_SUCCESS);
     }
+
+    void CheckFreeCommunicationPermission(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_PCTL, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
 };
 
 void Module::Interface::CreateService(Kernel::HLERequestContext& ctx) {

src/core/loader/deconstructed_rom_directory.cpp

@@ -8,6 +8,7 @@
 #include "common/logging/log.h"
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/control_metadata.h"
+#include "core/file_sys/romfs_factory.h"
 #include "core/gdbstub/gdbstub.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/resource_limit.h"

src/core/loader/nca.cpp

@@ -3,28 +3,23 @@
 // Refer to the license.txt file included.
 
 #include <utility>
-#include <vector>
 
 #include "common/file_util.h"
 #include "common/logging/log.h"
-#include "common/string_util.h"
-#include "common/swap.h"
-#include "core/core.h"
 #include "core/file_sys/content_archive.h"
-#include "core/file_sys/program_metadata.h"
-#include "core/gdbstub/gdbstub.h"
+#include "core/file_sys/romfs_factory.h"
 #include "core/hle/kernel/process.h"
-#include "core/hle/kernel/resource_limit.h"
 #include "core/hle/service/filesystem/filesystem.h"
+#include "core/loader/deconstructed_rom_directory.h"
 #include "core/loader/nca.h"
-#include "core/loader/nso.h"
-#include "core/memory.h"
 
 namespace Loader {
 
 AppLoader_NCA::AppLoader_NCA(FileSys::VirtualFile file_)
     : AppLoader(std::move(file_)), nca(std::make_unique<FileSys::NCA>(file)) {}
 
+AppLoader_NCA::~AppLoader_NCA() = default;
+
 FileType AppLoader_NCA::IdentifyType(const FileSys::VirtualFile& file) {
     FileSys::NCA nca(file);
 
@@ -83,6 +78,4 @@ ResultStatus AppLoader_NCA::ReadProgramId(u64& out_program_id) {
     return ResultStatus::Success;
 }
 
-AppLoader_NCA::~AppLoader_NCA() = default;
-
 } // namespace Loader

src/core/loader/nca.h

@@ -4,20 +4,24 @@
 
 #pragma once
 
-#include <string>
 #include "common/common_types.h"
-#include "core/file_sys/content_archive.h"
-#include "core/file_sys/program_metadata.h"
+#include "core/file_sys/vfs.h"
 #include "core/hle/kernel/object.h"
 #include "core/loader/loader.h"
-#include "deconstructed_rom_directory.h"
+
+namespace FileSys {
+class NCA;
+}
 
 namespace Loader {
 
+class AppLoader_DeconstructedRomDirectory;
+
 /// Loads an NCA file
 class AppLoader_NCA final : public AppLoader {
 public:
     explicit AppLoader_NCA(FileSys::VirtualFile file);
+    ~AppLoader_NCA() override;
 
     /**
      * Returns the type of the file
@@ -35,12 +39,7 @@ public:
     ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
     ResultStatus ReadProgramId(u64& out_program_id) override;
 
-    ~AppLoader_NCA();
-
 private:
-    FileSys::ProgramMetadata metadata;
-
-    FileSys::NCAHeader header;
     std::unique_ptr<FileSys::NCA> nca;
     std::unique_ptr<AppLoader_DeconstructedRomDirectory> directory_loader;
 };

src/core/loader/xci.cpp

@@ -4,22 +4,14 @@
 
 #include <vector>
 
-#include "common/file_util.h"
-#include "common/logging/log.h"
-#include "common/string_util.h"
-#include "common/swap.h"
-#include "core/core.h"
+#include "common/common_types.h"
+#include "core/file_sys/card_image.h"
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/control_metadata.h"
-#include "core/file_sys/program_metadata.h"
 #include "core/file_sys/romfs.h"
-#include "core/gdbstub/gdbstub.h"
 #include "core/hle/kernel/process.h"
-#include "core/hle/kernel/resource_limit.h"
-#include "core/hle/service/filesystem/filesystem.h"
-#include "core/loader/nso.h"
+#include "core/loader/nca.h"
 #include "core/loader/xci.h"
-#include "core/memory.h"
 
 namespace Loader {
 

src/core/loader/xci.h

@@ -6,12 +6,18 @@
 
 #include <memory>
 #include "common/common_types.h"
-#include "core/file_sys/card_image.h"
+#include "core/file_sys/vfs.h"
 #include "core/loader/loader.h"
-#include "core/loader/nca.h"
+
+namespace FileSys {
+class NACP;
+class XCI;
+} // namespace FileSys
 
 namespace Loader {
 
+class AppLoader_NCA;
+
 /// Loads an XCI file
 class AppLoader_XCI final : public AppLoader {
 public:
@@ -37,8 +43,6 @@ public:
     ResultStatus ReadTitle(std::string& title) override;
 
 private:
-    FileSys::ProgramMetadata metadata;
-
     std::unique_ptr<FileSys::XCI> xci;
     std::unique_ptr<AppLoader_NCA> nca_loader;
 

src/core/telemetry_session.cpp

@@ -2,34 +2,16 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <cstring>
-
 #include "common/assert.h"
+#include "common/common_types.h"
 #include "common/file_util.h"
-#include "common/scm_rev.h"
-#ifdef ARCHITECTURE_x86_64
-#include "common/x64/cpu_detect.h"
-#endif
+
 #include "core/core.h"
 #include "core/settings.h"
 #include "core/telemetry_session.h"
 
 namespace Core {
 
-#ifdef ARCHITECTURE_x86_64
-static const char* CpuVendorToStr(Common::CPUVendor vendor) {
-    switch (vendor) {
-    case Common::CPUVendor::INTEL:
-        return "Intel";
-    case Common::CPUVendor::AMD:
-        return "Amd";
-    case Common::CPUVendor::OTHER:
-        return "Other";
-    }
-    UNREACHABLE();
-}
-#endif
-
 static u64 GenerateTelemetryId() {
     u64 telemetry_id{};
     return telemetry_id;
@@ -37,8 +19,8 @@ static u64 GenerateTelemetryId() {
 
 u64 GetTelemetryId() {
     u64 telemetry_id{};
-    static const std::string& filename{FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) +
-                                       "telemetry_id"};
+    const std::string filename{FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) +
+                               "telemetry_id"};
 
     if (FileUtil::Exists(filename)) {
         FileUtil::IOFile file(filename, "rb");
@@ -62,8 +44,8 @@ u64 GetTelemetryId() {
 
 u64 RegenerateTelemetryId() {
     const u64 new_telemetry_id{GenerateTelemetryId()};
-    static const std::string& filename{FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) +
-                                       "telemetry_id"};
+    const std::string filename{FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) +
+                               "telemetry_id"};
 
     FileUtil::IOFile file(filename, "wb");
     if (!file.IsOpen()) {
@@ -112,48 +94,11 @@ TelemetrySession::TelemetrySession() {
     }
 
     // Log application information
-    const bool is_git_dirty{std::strstr(Common::g_scm_desc, "dirty") != nullptr};
-    AddField(Telemetry::FieldType::App, "Git_IsDirty", is_git_dirty);
-    AddField(Telemetry::FieldType::App, "Git_Branch", Common::g_scm_branch);
-    AddField(Telemetry::FieldType::App, "Git_Revision", Common::g_scm_rev);
-    AddField(Telemetry::FieldType::App, "BuildDate", Common::g_build_date);
-    AddField(Telemetry::FieldType::App, "BuildName", Common::g_build_name);
+    Telemetry::AppendBuildInfo(field_collection);
 
-// Log user system information
-#ifdef ARCHITECTURE_x86_64
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Model", Common::GetCPUCaps().cpu_string);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_BrandString",
-             Common::GetCPUCaps().brand_string);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Vendor",
-             CpuVendorToStr(Common::GetCPUCaps().vendor));
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_AES", Common::GetCPUCaps().aes);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_AVX", Common::GetCPUCaps().avx);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_AVX2", Common::GetCPUCaps().avx2);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_BMI1", Common::GetCPUCaps().bmi1);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_BMI2", Common::GetCPUCaps().bmi2);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_FMA", Common::GetCPUCaps().fma);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_FMA4", Common::GetCPUCaps().fma4);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_SSE", Common::GetCPUCaps().sse);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_SSE2", Common::GetCPUCaps().sse2);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_SSE3", Common::GetCPUCaps().sse3);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_SSSE3",
-             Common::GetCPUCaps().ssse3);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_SSE41",
-             Common::GetCPUCaps().sse4_1);
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Extension_x64_SSE42",
-             Common::GetCPUCaps().sse4_2);
-#else
-    AddField(Telemetry::FieldType::UserSystem, "CPU_Model", "Other");
-#endif
-#ifdef __APPLE__
-    AddField(Telemetry::FieldType::UserSystem, "OsPlatform", "Apple");
-#elif defined(_WIN32)
-    AddField(Telemetry::FieldType::UserSystem, "OsPlatform", "Windows");
-#elif defined(__linux__) || defined(linux) || defined(__linux)
-    AddField(Telemetry::FieldType::UserSystem, "OsPlatform", "Linux");
-#else
-    AddField(Telemetry::FieldType::UserSystem, "OsPlatform", "Unknown");
-#endif
+    // Log user system information
+    Telemetry::AppendCPUInfo(field_collection);
+    Telemetry::AppendOSInfo(field_collection);
 
     // Log user configuration information
     AddField(Telemetry::FieldType::UserConfig, "Core_UseCpuJit", Settings::values.use_cpu_jit);

src/video_core/engines/maxwell_3d.cpp

@@ -222,6 +222,18 @@ void Maxwell3D::DrawArrays() {
         debug_context->OnEvent(Tegra::DebugContext::Event::FinishedPrimitiveBatch, nullptr);
     }
 
+    // Both instance configuration registers can not be set at the same time.
+    ASSERT_MSG(!regs.draw.instance_next || !regs.draw.instance_cont,
+               "Illegal combination of instancing parameters");
+
+    if (regs.draw.instance_next) {
+        // Increment the current instance *before* drawing.
+        state.current_instance += 1;
+    } else if (!regs.draw.instance_cont) {
+        // Reset the current instance to 0.
+        state.current_instance = 0;
+    }
+
     const bool is_indexed{regs.index_array.count && !regs.vertex_buffer.count};
     rasterizer.AccelerateDrawBatch(is_indexed);
 

src/video_core/engines/maxwell_3d.h

@@ -638,6 +638,8 @@ public:
                     union {
                         u32 vertex_begin_gl;
                         BitField<0, 16, PrimitiveTopology> topology;
+                        BitField<26, 1, u32> instance_next;
+                        BitField<27, 1, u32> instance_cont;
                     };
                 } draw;
 
@@ -677,7 +679,19 @@ public:
 
                 INSERT_PADDING_WORDS(0x7);
 
-                INSERT_PADDING_WORDS(0x46);
+                INSERT_PADDING_WORDS(0x20);
+
+                struct {
+                    u32 is_instanced[NumVertexArrays];
+
+                    /// Returns whether the vertex array specified by index is supposed to be
+                    /// accessed per instance or not.
+                    bool IsInstancingEnabled(u32 index) const {
+                        return is_instanced[index];
+                    }
+                } instanced_arrays;
+
+                INSERT_PADDING_WORDS(0x6);
 
                 Cull cull;
 
@@ -830,6 +844,7 @@ public:
         };
 
         std::array<ShaderStageInfo, Regs::MaxShaderStage> shader_stages;
+        u32 current_instance = 0; ///< Current instance to be used to simulate instanced rendering.
     };
 
     State state{};
@@ -925,6 +940,7 @@ ASSERT_REG_POSITION(point_coord_replace, 0x581);
 ASSERT_REG_POSITION(code_address, 0x582);
 ASSERT_REG_POSITION(draw, 0x585);
 ASSERT_REG_POSITION(index_array, 0x5F2);
+ASSERT_REG_POSITION(instanced_arrays, 0x620);
 ASSERT_REG_POSITION(cull, 0x646);
 ASSERT_REG_POSITION(clear_buffers, 0x674);
 ASSERT_REG_POSITION(query, 0x6C0);

src/video_core/engines/shader_bytecode.h

@@ -12,6 +12,7 @@
 
 #include <boost/optional.hpp>
 
+#include "common/assert.h"
 #include "common/bit_field.h"
 #include "common/common_types.h"
 
@@ -79,6 +80,9 @@ union Attribute {
         // shader, and a tuple of (TessCoord.x, TessCoord.y, TessCoord.z, ~) when inside a Tess Eval
         // shader.
         TessCoordInstanceIDVertexID = 47,
+        // This attribute contains a tuple of (Unk, Unk, Unk, gl_FrontFacing) when inside a fragment
+        // shader. It is unknown what the other values contain.
+        FrontFacing = 63,
     };
 
     union {
@@ -141,6 +145,7 @@ enum class PredCondition : u64 {
     NotEqual = 5,
     GreaterEqual = 6,
     LessThanWithNan = 9,
+    GreaterThanWithNan = 12,
     NotEqualWithNan = 13,
     // TODO(Subv): Other condition types
 };
@@ -213,6 +218,18 @@ enum class FlowCondition : u64 {
     Fcsm_Tr = 0x1C, // TODO(bunnei): What is this used for?
 };
 
+enum class PredicateResultMode : u64 {
+    None = 0x0,
+    NotZero = 0x3,
+};
+
+enum class TextureType : u64 {
+    Texture1D = 0,
+    Texture2D = 1,
+    Texture3D = 2,
+    TextureCube = 3,
+};
+
 union Instruction {
     Instruction& operator=(const Instruction& instr) {
         value = instr.value;
@@ -253,7 +270,7 @@ union Instruction {
             BitField<39, 1, u64> invert_a;
             BitField<40, 1, u64> invert_b;
             BitField<41, 2, LogicOperation> operation;
-            BitField<44, 2, u64> unk44;
+            BitField<44, 2, PredicateResultMode> pred_result_mode;
             BitField<48, 3, Pred> pred48;
         } lop;
 
@@ -282,6 +299,10 @@ union Instruction {
         }
     } alu;
 
+    union {
+        BitField<48, 1, u64> negate_b;
+    } fmul;
+
     union {
         BitField<48, 1, u64> is_signed;
     } shift;
@@ -420,6 +441,8 @@ union Instruction {
     } conversion;
 
     union {
+        BitField<28, 1, u64> array;
+        BitField<29, 2, TextureType> texture_type;
         BitField<31, 4, u64> component_mask;
 
         bool IsComponentEnabled(size_t component) const {
@@ -428,28 +451,91 @@ union Instruction {
     } tex;
 
     union {
-        BitField<50, 3, u64> component_mask_selector;
+        BitField<28, 1, u64> array;
+        BitField<29, 2, TextureType> texture_type;
+        BitField<56, 2, u64> component;
+    } tld4;
+
+    union {
+        BitField<52, 2, u64> component;
+    } tld4s;
+
+    union {
         BitField<0, 8, Register> gpr0;
         BitField<28, 8, Register> gpr28;
+        BitField<50, 3, u64> component_mask_selector;
+        BitField<53, 4, u64> texture_info;
+
+        TextureType GetTextureType() const {
+            // The TEXS instruction has a weird encoding for the texture type.
+            if (texture_info == 0)
+                return TextureType::Texture1D;
+            if (texture_info >= 1 && texture_info <= 9)
+                return TextureType::Texture2D;
+            if (texture_info >= 10 && texture_info <= 11)
+                return TextureType::Texture3D;
+            if (texture_info >= 12 && texture_info <= 13)
+                return TextureType::TextureCube;
+
+            LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}",
+                         static_cast<u32>(texture_info.Value()));
+            UNREACHABLE();
+        }
+
+        bool IsArrayTexture() const {
+            // TEXS only supports Texture2D arrays.
+            return texture_info >= 7 && texture_info <= 9;
+        }
 
         bool HasTwoDestinations() const {
             return gpr28.Value() != Register::ZeroIndex;
         }
 
         bool IsComponentEnabled(size_t component) const {
-            static constexpr std::array<std::array<u32, 8>, 4> mask_lut{
-                {{},
-                 {0x1, 0x2, 0x4, 0x8, 0x3},
-                 {0x1, 0x2, 0x4, 0x8, 0x3, 0x9, 0xa, 0xc},
-                 {0x7, 0xb, 0xd, 0xe, 0xf}}};
+            static constexpr std::array<std::array<u32, 8>, 4> mask_lut{{
+                {},
+                {0x1, 0x2, 0x4, 0x8, 0x3, 0x9, 0xa, 0xc},
+                {0x1, 0x2, 0x4, 0x8, 0x3, 0x9, 0xa, 0xc},
+                {0x7, 0xb, 0xd, 0xe, 0xf},
+            }};
 
             size_t index{gpr0.Value() != Register::ZeroIndex ? 1U : 0U};
             index |= gpr28.Value() != Register::ZeroIndex ? 2 : 0;
 
-            return ((1ull << component) & mask_lut[index][component_mask_selector]) != 0;
+            u32 mask = mask_lut[index][component_mask_selector];
+            // A mask of 0 means this instruction uses an unimplemented mask.
+            ASSERT(mask != 0);
+            return ((1ull << component) & mask) != 0;
         }
     } texs;
 
+    union {
+        BitField<53, 4, u64> texture_info;
+
+        TextureType GetTextureType() const {
+            // The TLDS instruction has a weird encoding for the texture type.
+            if (texture_info >= 0 && texture_info <= 1) {
+                return TextureType::Texture1D;
+            }
+            if (texture_info == 2 || texture_info == 8 || texture_info == 12 ||
+                texture_info >= 4 && texture_info <= 6) {
+                return TextureType::Texture2D;
+            }
+            if (texture_info == 7) {
+                return TextureType::Texture3D;
+            }
+
+            LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}",
+                         static_cast<u32>(texture_info.Value()));
+            UNREACHABLE();
+        }
+
+        bool IsArrayTexture() const {
+            // TEXS only supports Texture2D arrays.
+            return texture_info == 8;
+        }
+    } tlds;
+
     union {
         BitField<20, 24, u64> target;
         BitField<5, 1, u64> constant_buffer;
@@ -512,10 +598,14 @@ public:
         LD_A,
         LD_C,
         ST_A,
+        LDG, // Load from global memory
+        STG, // Store in global memory
         TEX,
-        TEXQ, // Texture Query
-        TEXS, // Texture Fetch with scalar/non-vec4 source/destinations
-        TLDS, // Texture Load with scalar/non-vec4 source/destinations
+        TEXQ,  // Texture Query
+        TEXS,  // Texture Fetch with scalar/non-vec4 source/destinations
+        TLDS,  // Texture Load with scalar/non-vec4 source/destinations
+        TLD4,  // Texture Load 4
+        TLD4S, // Texture Load 4 with scalar / non - vec4 source / destinations
         EXIT,
         IPA,
         FFMA_IMM, // Fused Multiply and Add
@@ -723,10 +813,14 @@ private:
             INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
             INST("1110111110010---", Id::LD_C, Type::Memory, "LD_C"),
             INST("1110111111110---", Id::ST_A, Type::Memory, "ST_A"),
+            INST("1110111011010---", Id::LDG, Type::Memory, "LDG"),
+            INST("1110111011011---", Id::STG, Type::Memory, "STG"),
             INST("110000----111---", Id::TEX, Type::Memory, "TEX"),
             INST("1101111101001---", Id::TEXQ, Type::Memory, "TEXQ"),
             INST("1101100---------", Id::TEXS, Type::Memory, "TEXS"),
             INST("1101101---------", Id::TLDS, Type::Memory, "TLDS"),
+            INST("110010----111---", Id::TLD4, Type::Memory, "TLD4"),
+            INST("1101111100------", Id::TLD4S, Type::Memory, "TLD4S"),
             INST("111000110000----", Id::EXIT, Type::Trivial, "EXIT"),
             INST("11100000--------", Id::IPA, Type::Trivial, "IPA"),
             INST("0011001-1-------", Id::FFMA_IMM, Type::Ffma, "FFMA_IMM"),

src/video_core/gpu.cpp

@@ -55,6 +55,7 @@ u32 RenderTargetBytesPerPixel(RenderTargetFormat format) {
     case RenderTargetFormat::RGBA8_UNORM:
     case RenderTargetFormat::RGBA8_SNORM:
     case RenderTargetFormat::RGBA8_SRGB:
+    case RenderTargetFormat::RGBA8_UINT:
     case RenderTargetFormat::RGB10_A2_UNORM:
     case RenderTargetFormat::BGRA8_UNORM:
     case RenderTargetFormat::RG16_UNORM:

src/video_core/gpu.h

@@ -30,6 +30,7 @@ enum class RenderTargetFormat : u32 {
     RGBA8_UNORM = 0xD5,
     RGBA8_SRGB = 0xD6,
     RGBA8_SNORM = 0xD7,
+    RGBA8_UINT = 0xD9,
     RG16_UNORM = 0xDA,
     RG16_SNORM = 0xDB,
     RG16_SINT = 0xDC,

src/video_core/rasterizer_interface.h

@@ -8,8 +8,6 @@
 #include "video_core/gpu.h"
 #include "video_core/memory_manager.h"
 
-struct ScreenInfo;
-
 namespace VideoCore {
 
 class RasterizerInterface {
@@ -55,7 +53,7 @@ public:
 
     /// Attempt to use a faster method to display the framebuffer to screen
     virtual bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
-                                   u32 pixel_stride, ScreenInfo& screen_info) {
+                                   u32 pixel_stride) {
         return false;
     }
 

src/video_core/renderer_base.cpp

@@ -2,7 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <memory>
 #include "core/frontend/emu_window.h"
 #include "core/settings.h"
 #include "video_core/renderer_base.h"
@@ -17,18 +16,11 @@ RendererBase::RendererBase(Core::Frontend::EmuWindow& window) : render_window{wi
 RendererBase::~RendererBase() = default;
 
 void RendererBase::RefreshBaseSettings() {
-    RefreshRasterizerSetting();
     UpdateCurrentFramebufferLayout();
 
     renderer_settings.use_framelimiter = Settings::values.toggle_framelimit;
 }
 
-void RendererBase::RefreshRasterizerSetting() {
-    if (rasterizer == nullptr) {
-        rasterizer = std::make_unique<RasterizerOpenGL>(render_window);
-    }
-}
-
 void RendererBase::UpdateCurrentFramebufferLayout() {
     const Layout::FramebufferLayout& layout = render_window.GetFramebufferLayout();
 

src/video_core/renderer_base.h

@@ -58,9 +58,6 @@ public:
     void RefreshBaseSettings();
 
 protected:
-    /// Refreshes settings specific to the rasterizer.
-    void RefreshRasterizerSetting();
-
     Core::Frontend::EmuWindow& render_window; ///< Reference to the render window handle.
     std::unique_ptr<RasterizerInterface> rasterizer;
     f32 m_current_fps = 0.0f; ///< Current framerate, should be set by the renderer

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -36,8 +36,8 @@ MICROPROFILE_DEFINE(OpenGL_Drawing, "OpenGL", "Drawing", MP_RGB(128, 128, 192));
 MICROPROFILE_DEFINE(OpenGL_Blits, "OpenGL", "Blits", MP_RGB(100, 100, 255));
 MICROPROFILE_DEFINE(OpenGL_CacheManagement, "OpenGL", "Cache Mgmt", MP_RGB(100, 255, 100));
 
-RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window)
-    : emu_window{window}, stream_buffer(GL_ARRAY_BUFFER, STREAM_BUFFER_SIZE) {
+RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo& info)
+    : emu_window{window}, screen_info{info}, stream_buffer(GL_ARRAY_BUFFER, STREAM_BUFFER_SIZE) {
     // Create sampler objects
     for (size_t i = 0; i < texture_samplers.size(); ++i) {
         texture_samplers[i].Create();
@@ -98,7 +98,8 @@ RasterizerOpenGL::~RasterizerOpenGL() {}
 std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr,
                                                              GLintptr buffer_offset) {
     MICROPROFILE_SCOPE(OpenGL_VAO);
-    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+    const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+    const auto& regs = gpu.regs;
 
     state.draw.vertex_array = hw_vao.handle;
     state.draw.vertex_buffer = stream_buffer.GetHandle();
@@ -110,9 +111,13 @@ std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr,
         if (!vertex_array.IsEnabled())
             continue;
 
-        const Tegra::GPUVAddr start = vertex_array.StartAddress();
+        Tegra::GPUVAddr start = vertex_array.StartAddress();
         const Tegra::GPUVAddr end = regs.vertex_array_limit[index].LimitAddress();
 
+        if (regs.instanced_arrays.IsInstancingEnabled(index) && vertex_array.divisor != 0) {
+            start += vertex_array.stride * (gpu.state.current_instance / vertex_array.divisor);
+        }
+
         ASSERT(end > start);
         u64 size = end - start + 1;
 
@@ -124,7 +129,15 @@ std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr,
         glBindVertexBuffer(index, stream_buffer.GetHandle(), vertex_buffer_offset,
                            vertex_array.stride);
 
-        ASSERT_MSG(vertex_array.divisor == 0, "Vertex buffer divisor unimplemented");
+        if (regs.instanced_arrays.IsInstancingEnabled(index) && vertex_array.divisor != 0) {
+            // Tell OpenGL that this is an instanced vertex buffer to prevent accessing different
+            // indexes on each vertex. We do the instance indexing manually by incrementing the
+            // start address of the vertex buffer.
+            glVertexBindingDivisor(index, 1);
+        } else {
+            // Disable the vertex buffer instancing.
+            glVertexBindingDivisor(index, 0);
+        }
     }
 
     // Use the vertex array as-is, assumes that the data is formatted correctly for OpenGL.
@@ -291,7 +304,8 @@ bool RasterizerOpenGL::AccelerateDrawBatch(bool is_indexed) {
 }
 
 std::pair<Surface, Surface> RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb,
-                                                                    bool using_depth_fb) {
+                                                                    bool using_depth_fb,
+                                                                    bool preserve_contents) {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
     if (regs.rt[0].format == Tegra::RenderTargetFormat::NONE) {
@@ -314,7 +328,7 @@ std::pair<Surface, Surface> RasterizerOpenGL::ConfigureFramebuffers(bool using_c
     Surface depth_surface;
     MathUtil::Rectangle<u32> surfaces_rect;
     std::tie(color_surface, depth_surface, surfaces_rect) =
-        res_cache.GetFramebufferSurfaces(using_color_fb, using_depth_fb);
+        res_cache.GetFramebufferSurfaces(using_color_fb, using_depth_fb, preserve_contents);
 
     const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()};
     const MathUtil::Rectangle<u32> draw_rect{
@@ -377,7 +391,7 @@ void RasterizerOpenGL::Clear() {
     ScopeAcquireGLContext acquire_context{emu_window};
 
     auto [dirty_color_surface, dirty_depth_surface] =
-        ConfigureFramebuffers(use_color_fb, use_depth_fb);
+        ConfigureFramebuffers(use_color_fb, use_depth_fb, false);
 
     // TODO(Subv): Support clearing only partial colors.
     glClearColor(regs.clear_color[0], regs.clear_color[1], regs.clear_color[2],
@@ -432,7 +446,7 @@ void RasterizerOpenGL::DrawArrays() {
     ScopeAcquireGLContext acquire_context{emu_window};
 
     auto [dirty_color_surface, dirty_depth_surface] =
-        ConfigureFramebuffers(true, regs.zeta.Address() != 0 && regs.zeta_enable != 0);
+        ConfigureFramebuffers(true, regs.zeta.Address() != 0 && regs.zeta_enable != 0, true);
 
     SyncDepthTestState();
     SyncBlendState();
@@ -561,8 +575,7 @@ bool RasterizerOpenGL::AccelerateFill(const void* config) {
 }
 
 bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config,
-                                         VAddr framebuffer_addr, u32 pixel_stride,
-                                         ScreenInfo& screen_info) {
+                                         VAddr framebuffer_addr, u32 pixel_stride) {
     if (!framebuffer_addr) {
         return {};
     }

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -30,7 +30,7 @@ class EmuWindow;
 
 class RasterizerOpenGL : public VideoCore::RasterizerInterface {
 public:
-    explicit RasterizerOpenGL(Core::Frontend::EmuWindow& renderer);
+    explicit RasterizerOpenGL(Core::Frontend::EmuWindow& renderer, ScreenInfo& info);
     ~RasterizerOpenGL() override;
 
     void DrawArrays() override;
@@ -43,8 +43,8 @@ public:
     bool AccelerateDisplayTransfer(const void* config) override;
     bool AccelerateTextureCopy(const void* config) override;
     bool AccelerateFill(const void* config) override;
-    bool AccelerateDisplay(const Tegra::FramebufferConfig& framebuffer, VAddr framebuffer_addr,
-                           u32 pixel_stride, ScreenInfo& screen_info) override;
+    bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
+                           u32 pixel_stride) override;
     bool AccelerateDrawBatch(bool is_indexed) override;
 
     /// OpenGL shader generated for a given Maxwell register state
@@ -87,7 +87,8 @@ private:
 
     /// Configures the color and depth framebuffer states and returns the dirty <Color, Depth>
     /// surfaces if writing was enabled.
-    std::pair<Surface, Surface> ConfigureFramebuffers(bool using_color_fb, bool using_depth_fb);
+    std::pair<Surface, Surface> ConfigureFramebuffers(bool using_color_fb, bool using_depth_fb,
+                                                      bool preserve_contents);
 
     /// Binds the framebuffer color and depth surface
     void BindFramebufferSurfaces(const Surface& color_surface, const Surface& depth_surface,
@@ -151,6 +152,8 @@ private:
 
     Core::Frontend::EmuWindow& emu_window;
 
+    ScreenInfo& screen_info;
+
     std::unique_ptr<GLShader::ProgramManager> shader_program_manager;
     OGLVertexArray sw_vao;
     OGLVertexArray hw_vao;

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -94,6 +94,7 @@ struct FormatTuple {
 static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{
     {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // ABGR8U
     {GL_RGBA8, GL_RGBA, GL_BYTE, ComponentType::SNorm, false},                     // ABGR8S
+    {GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE, ComponentType::UInt, false},   // ABGR8UI
     {GL_RGB, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, ComponentType::UNorm, false},    // B5G6R5U
     {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV, ComponentType::UNorm,
      false}, // A2B10G10R10U
@@ -142,14 +143,16 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
     {GL_RG32UI, GL_RG_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false},              // RG32UI
     {GL_R32UI, GL_RED_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false},              // R32UI
 
+    // Depth formats
+    {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F
+    {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT, ComponentType::UNorm,
+     false}, // Z16
+
     // DepthStencil formats
     {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8, ComponentType::UNorm,
      false}, // Z24S8
     {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8, ComponentType::UNorm,
-     false},                                                                            // S8Z24
-    {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F
-    {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT, ComponentType::UNorm,
-     false}, // Z16
+     false}, // S8Z24
     {GL_DEPTH32F_STENCIL8, GL_DEPTH_STENCIL, GL_FLOAT_32_UNSIGNED_INT_24_8_REV,
      ComponentType::Float, false}, // Z32FS8
 }};
@@ -243,6 +246,7 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPU
         // 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>,
@@ -283,10 +287,10 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPU
         MortonCopy<true, PixelFormat::RG8S>,
         MortonCopy<true, PixelFormat::RG32UI>,
         MortonCopy<true, PixelFormat::R32UI>,
-        MortonCopy<true, PixelFormat::Z24S8>,
-        MortonCopy<true, PixelFormat::S8Z24>,
         MortonCopy<true, PixelFormat::Z32F>,
         MortonCopy<true, PixelFormat::Z16>,
+        MortonCopy<true, PixelFormat::Z24S8>,
+        MortonCopy<true, PixelFormat::S8Z24>,
         MortonCopy<true, PixelFormat::Z32FS8>,
         // clang-format on
 };
@@ -297,6 +301,7 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPU
         // 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>,
@@ -339,10 +344,10 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPU
         MortonCopy<false, PixelFormat::RG8S>,
         MortonCopy<false, PixelFormat::RG32UI>,
         MortonCopy<false, PixelFormat::R32UI>,
-        MortonCopy<false, PixelFormat::Z24S8>,
-        MortonCopy<false, PixelFormat::S8Z24>,
         MortonCopy<false, PixelFormat::Z32F>,
         MortonCopy<false, PixelFormat::Z16>,
+        MortonCopy<false, PixelFormat::Z24S8>,
+        MortonCopy<false, PixelFormat::S8Z24>,
         MortonCopy<false, PixelFormat::Z32FS8>,
         // clang-format on
 };
@@ -681,7 +686,8 @@ Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextu
 }
 
 SurfaceSurfaceRect_Tuple RasterizerCacheOpenGL::GetFramebufferSurfaces(bool using_color_fb,
-                                                                       bool using_depth_fb) {
+                                                                       bool using_depth_fb,
+                                                                       bool preserve_contents) {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
     // TODO(bunnei): This is hard corded to use just the first render buffer
@@ -703,7 +709,7 @@ SurfaceSurfaceRect_Tuple RasterizerCacheOpenGL::GetFramebufferSurfaces(bool usin
     MathUtil::Rectangle<u32> color_rect{};
     Surface color_surface;
     if (using_color_fb) {
-        color_surface = GetSurface(color_params);
+        color_surface = GetSurface(color_params, preserve_contents);
         if (color_surface) {
             color_rect = color_surface->GetSurfaceParams().GetRect();
         }
@@ -712,7 +718,7 @@ SurfaceSurfaceRect_Tuple RasterizerCacheOpenGL::GetFramebufferSurfaces(bool usin
     MathUtil::Rectangle<u32> depth_rect{};
     Surface depth_surface;
     if (using_depth_fb) {
-        depth_surface = GetSurface(depth_params);
+        depth_surface = GetSurface(depth_params, preserve_contents);
         if (depth_surface) {
             depth_rect = depth_surface->GetSurfaceParams().GetRect();
         }
@@ -747,7 +753,7 @@ void RasterizerCacheOpenGL::FlushSurface(const Surface& surface) {
     surface->FlushGLBuffer();
 }
 
-Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params) {
+Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool preserve_contents) {
     if (params.addr == 0 || params.height * params.width == 0) {
         return {};
     }
@@ -769,9 +775,13 @@ Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params) {
         } else if (surface->GetSurfaceParams().IsCompatibleSurface(params)) {
             // Use the cached surface as-is
             return surface;
-        } else {
-            // If surface parameters changed, recreate the surface from the old one
+        } else if (preserve_contents) {
+            // If surface parameters changed and we care about keeping the previous data, recreate
+            // the surface from the old one
             return RecreateSurface(surface, params);
+        } else {
+            // Delete the old surface before creating a new one to prevent collisions.
+            UnregisterSurface(surface);
         }
     }
 
@@ -788,14 +798,58 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& surface,
     // Verify surface is compatible for blitting
     const auto& params{surface->GetSurfaceParams()};
     ASSERT(params.type == new_params.type);
-    ASSERT(params.pixel_format == new_params.pixel_format);
-    ASSERT(params.component_type == new_params.component_type);
+    ASSERT_MSG(params.GetCompressionFactor(params.pixel_format) == 1,
+               "Compressed texture reinterpretation is not supported");
 
     // Create a new surface with the new parameters, and blit the previous surface to it
     Surface new_surface{std::make_shared<CachedSurface>(new_params)};
-    BlitTextures(surface->Texture().handle, params.GetRect(), new_surface->Texture().handle,
-                 new_surface->GetSurfaceParams().GetRect(), params.type, read_framebuffer.handle,
-                 draw_framebuffer.handle);
+
+    auto source_format = GetFormatTuple(params.pixel_format, params.component_type);
+    auto dest_format = GetFormatTuple(new_params.pixel_format, new_params.component_type);
+
+    size_t buffer_size = std::max(params.SizeInBytes(), new_params.SizeInBytes());
+
+    // Use a Pixel Buffer Object to download the previous texture and then upload it to the new one
+    // using the new format.
+    OGLBuffer pbo;
+    pbo.Create();
+
+    glBindBuffer(GL_PIXEL_PACK_BUFFER, pbo.handle);
+    glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW_ARB);
+    glGetTextureImage(surface->Texture().handle, 0, source_format.format, source_format.type,
+                      params.SizeInBytes(), nullptr);
+
+    // If the new texture is bigger than the previous one, we need to fill in the rest with data
+    // from the CPU.
+    if (params.SizeInBytes() < new_params.SizeInBytes()) {
+        // Upload the rest of the memory.
+        if (new_params.is_tiled) {
+            // TODO(Subv): We might have to de-tile the subtexture and re-tile it with the rest of
+            // the data in this case. Games like Super Mario Odyssey seem to hit this case when
+            // drawing, it re-uses the memory of a previous texture as a bigger framebuffer but it
+            // doesn't clear it beforehand, the texture is already full of zeros.
+            LOG_CRITICAL(HW_GPU, "Trying to upload extra texture data from the CPU during "
+                                 "reinterpretation but the texture is tiled.");
+        }
+        size_t remaining_size = new_params.SizeInBytes() - params.SizeInBytes();
+        auto address = Core::System::GetInstance().GPU().memory_manager->GpuToCpuAddress(
+            new_params.addr + params.SizeInBytes());
+        std::vector<u8> data(remaining_size);
+        Memory::ReadBlock(*address, data.data(), data.size());
+        glBufferSubData(GL_PIXEL_PACK_BUFFER, params.SizeInBytes(), remaining_size, data.data());
+    }
+
+    glBindBuffer(GL_PIXEL_PACK_BUFFER, 0);
+
+    const auto& dest_rect{new_params.GetRect()};
+
+    glBindBuffer(GL_PIXEL_UNPACK_BUFFER, pbo.handle);
+    glTextureSubImage2D(
+        new_surface->Texture().handle, 0, 0, 0, static_cast<GLsizei>(dest_rect.GetWidth()),
+        static_cast<GLsizei>(dest_rect.GetHeight()), dest_format.format, dest_format.type, nullptr);
+    glBindBuffer(GL_PIXEL_UNPACK_BUFFER, 0);
+
+    pbo.Release();
 
     // Update cache accordingly
     UnregisterSurface(surface);

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -25,55 +25,60 @@ struct SurfaceParams {
     enum class PixelFormat {
         ABGR8U = 0,
         ABGR8S = 1,
-        B5G6R5U = 2,
-        A2B10G10R10U = 3,
-        A1B5G5R5U = 4,
-        R8U = 5,
-        R8UI = 6,
-        RGBA16F = 7,
-        RGBA16U = 8,
-        RGBA16UI = 9,
-        R11FG11FB10F = 10,
-        RGBA32UI = 11,
-        DXT1 = 12,
-        DXT23 = 13,
-        DXT45 = 14,
-        DXN1 = 15, // This is also known as BC4
-        DXN2UNORM = 16,
-        DXN2SNORM = 17,
-        BC7U = 18,
-        ASTC_2D_4X4 = 19,
-        G8R8U = 20,
-        G8R8S = 21,
-        BGRA8 = 22,
-        RGBA32F = 23,
-        RG32F = 24,
-        R32F = 25,
-        R16F = 26,
-        R16U = 27,
-        R16S = 28,
-        R16UI = 29,
-        R16I = 30,
-        RG16 = 31,
-        RG16F = 32,
-        RG16UI = 33,
-        RG16I = 34,
-        RG16S = 35,
-        RGB32F = 36,
-        SRGBA8 = 37,
-        RG8U = 38,
-        RG8S = 39,
-        RG32UI = 40,
-        R32UI = 41,
+        ABGR8UI = 2,
+        B5G6R5U = 3,
+        A2B10G10R10U = 4,
+        A1B5G5R5U = 5,
+        R8U = 6,
+        R8UI = 7,
+        RGBA16F = 8,
+        RGBA16U = 9,
+        RGBA16UI = 10,
+        R11FG11FB10F = 11,
+        RGBA32UI = 12,
+        DXT1 = 13,
+        DXT23 = 14,
+        DXT45 = 15,
+        DXN1 = 16, // This is also known as BC4
+        DXN2UNORM = 17,
+        DXN2SNORM = 18,
+        BC7U = 19,
+        ASTC_2D_4X4 = 20,
+        G8R8U = 21,
+        G8R8S = 22,
+        BGRA8 = 23,
+        RGBA32F = 24,
+        RG32F = 25,
+        R32F = 26,
+        R16F = 27,
+        R16U = 28,
+        R16S = 29,
+        R16UI = 30,
+        R16I = 31,
+        RG16 = 32,
+        RG16F = 33,
+        RG16UI = 34,
+        RG16I = 35,
+        RG16S = 36,
+        RGB32F = 37,
+        SRGBA8 = 38,
+        RG8U = 39,
+        RG8S = 40,
+        RG32UI = 41,
+        R32UI = 42,
 
         MaxColorFormat,
 
+        // Depth formats
+        Z32F = 43,
+        Z16 = 44,
+
+        MaxDepthFormat,
+
         // DepthStencil formats
-        Z24S8 = 42,
-        S8Z24 = 43,
-        Z32F = 44,
-        Z16 = 45,
-        Z32FS8 = 46,
+        Z24S8 = 45,
+        S8Z24 = 46,
+        Z32FS8 = 47,
 
         MaxDepthStencilFormat,
 
@@ -113,6 +118,7 @@ struct SurfaceParams {
         constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{
             1, // ABGR8U
             1, // ABGR8S
+            1, // ABGR8UI
             1, // B5G6R5U
             1, // A2B10G10R10U
             1, // A1B5G5R5U
@@ -153,10 +159,10 @@ struct SurfaceParams {
             1, // RG8S
             1, // RG32UI
             1, // R32UI
-            1, // Z24S8
-            1, // S8Z24
             1, // Z32F
             1, // Z16
+            1, // Z24S8
+            1, // S8Z24
             1, // Z32FS8
         }};
 
@@ -171,6 +177,7 @@ struct SurfaceParams {
         constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
             32,  // ABGR8U
             32,  // ABGR8S
+            32,  // ABGR8UI
             16,  // B5G6R5U
             32,  // A2B10G10R10U
             16,  // A1B5G5R5U
@@ -211,10 +218,10 @@ struct SurfaceParams {
             16,  // RG8S
             64,  // RG32UI
             32,  // R32UI
-            32,  // Z24S8
-            32,  // S8Z24
             32,  // Z32F
             16,  // Z16
+            32,  // Z24S8
+            32,  // S8Z24
             64,  // Z32FS8
         }};
 
@@ -253,6 +260,8 @@ struct SurfaceParams {
             return PixelFormat::ABGR8U;
         case Tegra::RenderTargetFormat::RGBA8_SNORM:
             return PixelFormat::ABGR8S;
+        case Tegra::RenderTargetFormat::RGBA8_UINT:
+            return PixelFormat::ABGR8UI;
         case Tegra::RenderTargetFormat::BGRA8_UNORM:
             return PixelFormat::BGRA8;
         case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
@@ -323,6 +332,8 @@ struct SurfaceParams {
                 return PixelFormat::ABGR8U;
             case Tegra::Texture::ComponentType::SNORM:
                 return PixelFormat::ABGR8S;
+            case Tegra::Texture::ComponentType::UINT:
+                return PixelFormat::ABGR8UI;
             }
             LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
                          static_cast<u32>(component_type));
@@ -547,6 +558,7 @@ struct SurfaceParams {
         case Tegra::RenderTargetFormat::R16_UINT:
         case Tegra::RenderTargetFormat::RG32_UINT:
         case Tegra::RenderTargetFormat::R32_UINT:
+        case Tegra::RenderTargetFormat::RGBA8_UINT:
             return ComponentType::UInt;
         case Tegra::RenderTargetFormat::RG16_SINT:
         case Tegra::RenderTargetFormat::R16_SINT:
@@ -587,6 +599,10 @@ struct SurfaceParams {
             return SurfaceType::ColorTexture;
         }
 
+        if (static_cast<size_t>(pixel_format) < static_cast<size_t>(PixelFormat::MaxDepthFormat)) {
+            return SurfaceType::Depth;
+        }
+
         if (static_cast<size_t>(pixel_format) <
             static_cast<size_t>(PixelFormat::MaxDepthStencilFormat)) {
             return SurfaceType::DepthStencil;
@@ -706,7 +722,8 @@ public:
     Surface GetTextureSurface(const Tegra::Texture::FullTextureInfo& config);
 
     /// Get the color and depth surfaces based on the framebuffer configuration
-    SurfaceSurfaceRect_Tuple GetFramebufferSurfaces(bool using_color_fb, bool using_depth_fb);
+    SurfaceSurfaceRect_Tuple GetFramebufferSurfaces(bool using_color_fb, bool using_depth_fb,
+                                                    bool preserve_contents);
 
     /// Flushes the surface to Switch memory
     void FlushSurface(const Surface& surface);
@@ -722,7 +739,7 @@ public:
 
 private:
     void LoadSurface(const Surface& surface);
-    Surface GetSurface(const SurfaceParams& params);
+    Surface GetSurface(const SurfaceParams& params, bool preserve_contents = true);
 
     /// Recreates a surface with new parameters
     Surface RecreateSurface(const Surface& surface, const SurfaceParams& new_params);

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -26,6 +26,7 @@ using Tegra::Shader::Sampler;
 using Tegra::Shader::SubOp;
 
 constexpr u32 PROGRAM_END = MAX_PROGRAM_CODE_LENGTH;
+constexpr u32 PROGRAM_HEADER_SIZE = 0x50;
 
 class DecompileFail : public std::runtime_error {
 public:
@@ -439,13 +440,12 @@ public:
         }
         declarations.AddNewLine();
 
-        // Append the sampler2D array for the used textures.
-        size_t num_samplers = GetSamplers().size();
-        if (num_samplers > 0) {
-            declarations.AddLine("uniform sampler2D " + SamplerEntry::GetArrayName(stage) + '[' +
-                                 std::to_string(num_samplers) + "];");
-            declarations.AddNewLine();
+        const auto& samplers = GetSamplers();
+        for (const auto& sampler : samplers) {
+            declarations.AddLine("uniform " + sampler.GetTypeString() + ' ' + sampler.GetName() +
+                                 ';');
         }
+        declarations.AddNewLine();
     }
 
     /// Returns a list of constant buffer declarations
@@ -457,13 +457,14 @@ public:
     }
 
     /// Returns a list of samplers used in the shader
-    std::vector<SamplerEntry> GetSamplers() const {
+    const std::vector<SamplerEntry>& GetSamplers() const {
         return used_samplers;
     }
 
     /// Returns the GLSL sampler used for the input shader sampler, and creates a new one if
     /// necessary.
-    std::string AccessSampler(const Sampler& sampler) {
+    std::string AccessSampler(const Sampler& sampler, Tegra::Shader::TextureType type,
+                              bool is_array) {
         size_t offset = static_cast<size_t>(sampler.index.Value());
 
         // If this sampler has already been used, return the existing mapping.
@@ -472,12 +473,13 @@ public:
                          [&](const SamplerEntry& entry) { return entry.GetOffset() == offset; });
 
         if (itr != used_samplers.end()) {
+            ASSERT(itr->GetType() == type && itr->IsArray() == is_array);
             return itr->GetName();
         }
 
         // Otherwise create a new mapping for this sampler
         size_t next_index = used_samplers.size();
-        SamplerEntry entry{stage, offset, next_index};
+        SamplerEntry entry{stage, offset, next_index, type, is_array};
         used_samplers.emplace_back(entry);
         return entry.GetName();
     }
@@ -541,7 +543,11 @@ private:
             // vertex shader, and what's the value of the fourth element when inside a Tess Eval
             // shader.
             ASSERT(stage == Maxwell3D::Regs::ShaderStage::Vertex);
-            return "vec4(0, 0, uintBitsToFloat(gl_InstanceID), uintBitsToFloat(gl_VertexID))";
+            return "vec4(0, 0, uintBitsToFloat(instance_id.x), uintBitsToFloat(gl_VertexID))";
+        case Attribute::Index::FrontFacing:
+            // TODO(Subv): Find out what the values are for the other elements.
+            ASSERT(stage == Maxwell3D::Regs::ShaderStage::Fragment);
+            return "vec4(0, 0, 0, uintBitsToFloat(gl_FrontFacing ? 1 : 0))";
         default:
             const u32 index{static_cast<u32>(attribute) -
                             static_cast<u32>(Attribute::Index::Attribute_0)};
@@ -616,6 +622,23 @@ public:
     }
 
 private:
+    // Shader program header for a Fragment Shader.
+    struct FragmentHeader {
+        INSERT_PADDING_WORDS(5);
+        INSERT_PADDING_WORDS(13);
+        u32 enabled_color_outputs;
+        union {
+            BitField<0, 1, u32> writes_samplemask;
+            BitField<1, 1, u32> writes_depth;
+        };
+
+        bool IsColorComponentOutputEnabled(u32 render_target, u32 component) const {
+            u32 bit = render_target * 4 + component;
+            return enabled_color_outputs & (1 << bit);
+        }
+    };
+    static_assert(sizeof(FragmentHeader) == PROGRAM_HEADER_SIZE, "FragmentHeader size is wrong");
+
     /// Gets the Subroutine object corresponding to the specified address.
     const Subroutine& GetSubroutine(u32 begin, u32 end) const {
         auto iter = subroutines.find(Subroutine{begin, end, suffix});
@@ -634,8 +657,8 @@ private:
     }
 
     /// Generates code representing a texture sampler.
-    std::string GetSampler(const Sampler& sampler) {
-        return regs.AccessSampler(sampler);
+    std::string GetSampler(const Sampler& sampler, Tegra::Shader::TextureType type, bool is_array) {
+        return regs.AccessSampler(sampler, type, is_array);
     }
 
     /**
@@ -703,10 +726,11 @@ private:
                                        const std::string& op_a, const std::string& op_b) const {
         using Tegra::Shader::PredCondition;
         static const std::unordered_map<PredCondition, const char*> PredicateComparisonStrings = {
-            {PredCondition::LessThan, "<"},        {PredCondition::Equal, "=="},
-            {PredCondition::LessEqual, "<="},      {PredCondition::GreaterThan, ">"},
-            {PredCondition::NotEqual, "!="},       {PredCondition::GreaterEqual, ">="},
-            {PredCondition::LessThanWithNan, "<"}, {PredCondition::NotEqualWithNan, "!="},
+            {PredCondition::LessThan, "<"},           {PredCondition::Equal, "=="},
+            {PredCondition::LessEqual, "<="},         {PredCondition::GreaterThan, ">"},
+            {PredCondition::NotEqual, "!="},          {PredCondition::GreaterEqual, ">="},
+            {PredCondition::LessThanWithNan, "<"},    {PredCondition::NotEqualWithNan, "!="},
+            {PredCondition::GreaterThanWithNan, ">"},
         };
 
         const auto& comparison{PredicateComparisonStrings.find(condition)};
@@ -715,7 +739,8 @@ private:
 
         std::string predicate{'(' + op_a + ") " + comparison->second + " (" + op_b + ')'};
         if (condition == PredCondition::LessThanWithNan ||
-            condition == PredCondition::NotEqualWithNan) {
+            condition == PredCondition::NotEqualWithNan ||
+            condition == PredCondition::GreaterThanWithNan) {
             predicate += " || isnan(" + op_a + ") || isnan(" + op_b + ')';
         }
 
@@ -741,6 +766,30 @@ private:
         return op->second;
     }
 
+    /**
+     * Transforms the input string GLSL operand into one that applies the abs() function and negates
+     * the output if necessary. When both abs and neg are true, the negation will be applied after
+     * taking the absolute value.
+     * @param operand The input operand to take the abs() of, negate, or both.
+     * @param abs Whether to apply the abs() function to the input operand.
+     * @param neg Whether to negate the input operand.
+     * @returns String corresponding to the operand after being transformed by the abs() and
+     * negation operations.
+     */
+    static std::string GetOperandAbsNeg(const std::string& operand, bool abs, bool neg) {
+        std::string result = operand;
+
+        if (abs) {
+            result = "abs(" + result + ')';
+        }
+
+        if (neg) {
+            result = "-(" + result + ')';
+        }
+
+        return result;
+    }
+
     /*
      * Returns whether the instruction at the specified offset is a 'sched' instruction.
      * Sched instructions always appear before a sequence of 3 instructions.
@@ -754,28 +803,51 @@ private:
     }
 
     void WriteLogicOperation(Register dest, LogicOperation logic_op, const std::string& op_a,
-                             const std::string& op_b) {
+                             const std::string& op_b,
+                             Tegra::Shader::PredicateResultMode predicate_mode,
+                             Tegra::Shader::Pred predicate) {
+        std::string result{};
         switch (logic_op) {
         case LogicOperation::And: {
-            regs.SetRegisterToInteger(dest, true, 0, '(' + op_a + " & " + op_b + ')', 1, 1);
+            result = '(' + op_a + " & " + op_b + ')';
             break;
         }
         case LogicOperation::Or: {
-            regs.SetRegisterToInteger(dest, true, 0, '(' + op_a + " | " + op_b + ')', 1, 1);
+            result = '(' + op_a + " | " + op_b + ')';
             break;
         }
         case LogicOperation::Xor: {
-            regs.SetRegisterToInteger(dest, true, 0, '(' + op_a + " ^ " + op_b + ')', 1, 1);
+            result = '(' + op_a + " ^ " + op_b + ')';
             break;
         }
         case LogicOperation::PassB: {
-            regs.SetRegisterToInteger(dest, true, 0, op_b, 1, 1);
+            result = op_b;
             break;
         }
         default:
             LOG_CRITICAL(HW_GPU, "Unimplemented logic operation: {}", static_cast<u32>(logic_op));
             UNREACHABLE();
         }
+
+        if (dest != Tegra::Shader::Register::ZeroIndex) {
+            regs.SetRegisterToInteger(dest, true, 0, result, 1, 1);
+        }
+
+        using Tegra::Shader::PredicateResultMode;
+        // Write the predicate value depending on the predicate mode.
+        switch (predicate_mode) {
+        case PredicateResultMode::None:
+            // Do nothing.
+            return;
+        case PredicateResultMode::NotZero:
+            // Set the predicate to true if the result is not zero.
+            SetPredicate(static_cast<u64>(predicate), '(' + result + ") != 0");
+            break;
+        default:
+            LOG_CRITICAL(HW_GPU, "Unimplemented predicate result mode: {}",
+                         static_cast<u32>(predicate_mode));
+            UNREACHABLE();
+        }
     }
 
     void WriteTexsInstruction(const Instruction& instr, const std::string& coord,
@@ -786,33 +858,90 @@ private:
         ++shader.scope;
         shader.AddLine(coord);
 
-        // TEXS has two destination registers. RG goes into gpr0+0 and gpr0+1, and BA
-        // goes into gpr28+0 and gpr28+1
-        size_t texs_offset{};
+        // TEXS has two destination registers and a swizzle. The first two elements in the swizzle
+        // go into gpr0+0 and gpr0+1, and the rest goes into gpr28+0 and gpr28+1
 
-        size_t src_elem{};
-        for (const auto& dest : {instr.gpr0.Value(), instr.gpr28.Value()}) {
-            size_t dest_elem{};
-            for (unsigned elem = 0; elem < 2; ++elem) {
-                if (!instr.texs.IsComponentEnabled(src_elem++)) {
-                    // Skip disabled components
-                    continue;
-                }
-                regs.SetRegisterToFloat(dest, elem + texs_offset, texture, 1, 4, false,
-                                        dest_elem++);
+        size_t written_components = 0;
+        for (u32 component = 0; component < 4; ++component) {
+            if (!instr.texs.IsComponentEnabled(component)) {
+                continue;
             }
 
-            if (!instr.texs.HasTwoDestinations()) {
-                // Skip the second destination
-                break;
+            if (written_components < 2) {
+                // Write the first two swizzle components to gpr0 and gpr0+1
+                regs.SetRegisterToFloat(instr.gpr0, component, texture, 1, 4, false,
+                                        written_components % 2);
+            } else {
+                ASSERT(instr.texs.HasTwoDestinations());
+                // Write the rest of the swizzle components to gpr28 and gpr28+1
+                regs.SetRegisterToFloat(instr.gpr28, component, texture, 1, 4, false,
+                                        written_components % 2);
             }
 
-            texs_offset += 2;
+            ++written_components;
         }
+
         --shader.scope;
         shader.AddLine('}');
     }
 
+    /*
+     * Emits code to push the input target address to the SSY address stack, incrementing the stack
+     * top.
+     */
+    void EmitPushToSSYStack(u32 target) {
+        shader.AddLine('{');
+        ++shader.scope;
+        shader.AddLine("ssy_stack[ssy_stack_top] = " + std::to_string(target) + "u;");
+        shader.AddLine("ssy_stack_top++;");
+        --shader.scope;
+        shader.AddLine('}');
+    }
+
+    /*
+     * Emits code to pop an address from the SSY address stack, setting the jump address to the
+     * popped address and decrementing the stack top.
+     */
+    void EmitPopFromSSYStack() {
+        shader.AddLine('{');
+        ++shader.scope;
+        shader.AddLine("ssy_stack_top--;");
+        shader.AddLine("jmp_to = ssy_stack[ssy_stack_top];");
+        shader.AddLine("break;");
+        --shader.scope;
+        shader.AddLine('}');
+    }
+
+    /// Writes the output values from a fragment shader to the corresponding GLSL output variables.
+    void EmitFragmentOutputsWrite() {
+        ASSERT(stage == Maxwell3D::Regs::ShaderStage::Fragment);
+        FragmentHeader header;
+        std::memcpy(&header, program_code.data(), PROGRAM_HEADER_SIZE);
+
+        ASSERT_MSG(header.writes_samplemask == 0, "Samplemask write is unimplemented");
+
+        // Write the color outputs using the data in the shader registers, disabled
+        // rendertargets/components are skipped in the register assignment.
+        u32 current_reg = 0;
+        for (u32 render_target = 0; render_target < Maxwell3D::Regs::NumRenderTargets;
+             ++render_target) {
+            // TODO(Subv): Figure out how dual-source blending is configured in the Switch.
+            for (u32 component = 0; component < 4; ++component) {
+                if (header.IsColorComponentOutputEnabled(render_target, component)) {
+                    shader.AddLine(fmt::format("color[{}][{}] = {};", render_target, component,
+                                               regs.GetRegisterAsFloat(current_reg)));
+                    ++current_reg;
+                }
+            }
+        }
+
+        if (header.writes_depth) {
+            // The depth output is always 2 registers after the last color output, and current_reg
+            // already contains one past the last color register.
+            shader.AddLine("gl_FragDepth = " + regs.GetRegisterAsFloat(current_reg + 1) + ';');
+        }
+    }
+
     /**
      * Compiles a single instruction from Tegra to GLSL.
      * @param offset the offset of the Tegra shader instruction.
@@ -857,13 +986,6 @@ private:
         switch (opcode->GetType()) {
         case OpCode::Type::Arithmetic: {
             std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
-            if (instr.alu.abs_a) {
-                op_a = "abs(" + op_a + ')';
-            }
-
-            if (instr.alu.negate_a) {
-                op_a = "-(" + op_a + ')';
-            }
 
             std::string op_b;
 
@@ -878,17 +1000,10 @@ private:
                 }
             }
 
-            if (instr.alu.abs_b) {
-                op_b = "abs(" + op_b + ')';
-            }
-
-            if (instr.alu.negate_b) {
-                op_b = "-(" + op_b + ')';
-            }
-
             switch (opcode->GetId()) {
             case OpCode::Id::MOV_C:
             case OpCode::Id::MOV_R: {
+                // MOV does not have neither 'abs' nor 'neg' bits.
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_b, 1, 1);
                 break;
             }
@@ -896,6 +1011,8 @@ private:
             case OpCode::Id::FMUL_C:
             case OpCode::Id::FMUL_R:
             case OpCode::Id::FMUL_IMM: {
+                // FMUL does not have 'abs' bits and only the second operand has a 'neg' bit.
+                op_b = GetOperandAbsNeg(op_b, false, instr.fmul.negate_b);
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b, 1, 1,
                                         instr.alu.saturate_d);
                 break;
@@ -903,11 +1020,14 @@ private:
             case OpCode::Id::FADD_C:
             case OpCode::Id::FADD_R:
             case OpCode::Id::FADD_IMM: {
+                op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a);
+                op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b);
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1,
                                         instr.alu.saturate_d);
                 break;
             }
             case OpCode::Id::MUFU: {
+                op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a);
                 switch (instr.sub_op) {
                 case SubOp::Cos:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "cos(" + op_a + ')', 1, 1,
@@ -947,6 +1067,9 @@ private:
             case OpCode::Id::FMNMX_C:
             case OpCode::Id::FMNMX_R:
             case OpCode::Id::FMNMX_IMM: {
+                op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a);
+                op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b);
+
                 std::string condition =
                     GetPredicateCondition(instr.alu.fmnmx.pred, instr.alu.fmnmx.negate_pred != 0);
                 std::string parameters = op_a + ',' + op_b;
@@ -960,7 +1083,7 @@ private:
             case OpCode::Id::RRO_R:
             case OpCode::Id::RRO_IMM: {
                 // Currently RRO is only implemented as a register move.
-                // Usage of `abs_b` and `negate_b` here should also be correct.
+                op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b);
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_b, 1, 1);
                 LOG_WARNING(HW_GPU, "RRO instruction is incomplete");
                 break;
@@ -1097,7 +1220,9 @@ private:
                 if (instr.alu.lop32i.invert_b)
                     op_b = "~(" + op_b + ')';
 
-                WriteLogicOperation(instr.gpr0, instr.alu.lop32i.operation, op_a, op_b);
+                WriteLogicOperation(instr.gpr0, instr.alu.lop32i.operation, op_a, op_b,
+                                    Tegra::Shader::PredicateResultMode::None,
+                                    Tegra::Shader::Pred::UnusedIndex);
                 break;
             }
             default: {
@@ -1163,16 +1288,14 @@ private:
             case OpCode::Id::LOP_C:
             case OpCode::Id::LOP_R:
             case OpCode::Id::LOP_IMM: {
-                ASSERT_MSG(!instr.alu.lop.unk44, "Unimplemented");
-                ASSERT_MSG(instr.alu.lop.pred48 == Pred::UnusedIndex, "Unimplemented");
-
                 if (instr.alu.lop.invert_a)
                     op_a = "~(" + op_a + ')';
 
                 if (instr.alu.lop.invert_b)
                     op_b = "~(" + op_b + ')';
 
-                WriteLogicOperation(instr.gpr0, instr.alu.lop.operation, op_a, op_b);
+                WriteLogicOperation(instr.gpr0, instr.alu.lop.operation, op_a, op_b,
+                                    instr.alu.lop.pred_result_mode, instr.alu.lop.pred48);
                 break;
             }
             case OpCode::Id::IMNMX_C:
@@ -1237,8 +1360,6 @@ private:
             break;
         }
         case OpCode::Type::Conversion: {
-            ASSERT_MSG(!instr.conversion.negate_a, "Unimplemented");
-
             switch (opcode->GetId()) {
             case OpCode::Id::I2I_R: {
                 ASSERT_MSG(!instr.conversion.selector, "Unimplemented");
@@ -1435,10 +1556,39 @@ private:
                 break;
             }
             case OpCode::Id::TEX: {
-                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
-                const std::string sampler = GetSampler(instr.sampler);
-                const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                ASSERT_MSG(instr.tex.array == 0, "TEX arrays unimplemented");
+                std::string coord{};
+
+                switch (instr.tex.texture_type) {
+                case Tegra::Shader::TextureType::Texture2D: {
+                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                    break;
+                }
+                case Tegra::Shader::TextureType::Texture3D: {
+                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    std::string z = regs.GetRegisterAsFloat(instr.gpr20);
+                    coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                    break;
+                }
+                case Tegra::Shader::TextureType::TextureCube: {
+                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    std::string z = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
+                    ASSERT(instr.gpr20.Value() == Register::ZeroIndex);
+                    coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                    break;
+                }
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+                                 static_cast<u32>(instr.tex.texture_type.Value()));
+                    UNREACHABLE();
+                }
+
+                const std::string sampler =
+                    GetSampler(instr.sampler, instr.tex.texture_type, instr.tex.array);
                 // Add an extra scope and declare the texture coords inside to prevent
                 // overwriting them in case they are used as outputs of the texs instruction.
                 shader.AddLine("{");
@@ -1460,24 +1610,129 @@ private:
                 break;
             }
             case OpCode::Id::TEXS: {
-                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20);
-                const std::string sampler = GetSampler(instr.sampler);
-                const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                std::string coord{};
+
+                switch (instr.texs.GetTextureType()) {
+                case Tegra::Shader::TextureType::Texture2D: {
+                    if (instr.texs.IsArrayTexture()) {
+                        std::string index = regs.GetRegisterAsInteger(instr.gpr8);
+                        std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                        std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+                        coord = "vec3 coords = vec3(" + x + ", " + y + ", " + index + ");";
+                    } else {
+                        std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                        std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+                        coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                    }
+                    break;
+                }
+                case Tegra::Shader::TextureType::Texture3D: {
+                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+                    std::string z = regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
+                    coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                    break;
+                }
+                case Tegra::Shader::TextureType::TextureCube: {
+                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    std::string z = regs.GetRegisterAsFloat(instr.gpr20);
+                    coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                    break;
+                }
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+                                 static_cast<u32>(instr.texs.GetTextureType()));
+                    UNREACHABLE();
+                }
+                const std::string sampler = GetSampler(instr.sampler, instr.texs.GetTextureType(),
+                                                       instr.texs.IsArrayTexture());
 
                 const std::string texture = "texture(" + sampler + ", coords)";
                 WriteTexsInstruction(instr, coord, texture);
                 break;
             }
             case OpCode::Id::TLDS: {
-                const std::string op_a = regs.GetRegisterAsInteger(instr.gpr8);
-                const std::string op_b = regs.GetRegisterAsInteger(instr.gpr20);
-                const std::string sampler = GetSampler(instr.sampler);
-                const std::string coord = "ivec2 coords = ivec2(" + op_a + ", " + op_b + ");";
+                ASSERT(instr.tlds.GetTextureType() == Tegra::Shader::TextureType::Texture2D);
+                ASSERT(instr.tlds.IsArrayTexture() == false);
+                std::string coord{};
+
+                switch (instr.tlds.GetTextureType()) {
+                case Tegra::Shader::TextureType::Texture2D: {
+                    if (instr.tlds.IsArrayTexture()) {
+                        LOG_CRITICAL(HW_GPU, "Unhandled 2d array texture");
+                        UNREACHABLE();
+                    } else {
+                        std::string x = regs.GetRegisterAsInteger(instr.gpr8);
+                        std::string y = regs.GetRegisterAsInteger(instr.gpr20);
+                        coord = "ivec2 coords = ivec2(" + x + ", " + y + ");";
+                    }
+                    break;
+                }
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+                                 static_cast<u32>(instr.tlds.GetTextureType()));
+                    UNREACHABLE();
+                }
+                const std::string sampler = GetSampler(instr.sampler, instr.tlds.GetTextureType(),
+                                                       instr.tlds.IsArrayTexture());
                 const std::string texture = "texelFetch(" + sampler + ", coords, 0)";
                 WriteTexsInstruction(instr, coord, texture);
                 break;
             }
+            case OpCode::Id::TLD4: {
+                ASSERT(instr.tld4.texture_type == Tegra::Shader::TextureType::Texture2D);
+                ASSERT(instr.tld4.array == 0);
+                std::string coord{};
+
+                switch (instr.tld4.texture_type) {
+                case Tegra::Shader::TextureType::Texture2D: {
+                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+                    break;
+                }
+                default:
+                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+                                 static_cast<u32>(instr.tld4.texture_type.Value()));
+                    UNREACHABLE();
+                }
+
+                const std::string sampler =
+                    GetSampler(instr.sampler, instr.tld4.texture_type, instr.tld4.array);
+                // Add an extra scope and declare the texture coords inside to prevent
+                // overwriting them in case they are used as outputs of the texs instruction.
+                shader.AddLine("{");
+                ++shader.scope;
+                shader.AddLine(coord);
+                const std::string texture = "textureGather(" + sampler + ", coords, " +
+                                            std::to_string(instr.tld4.component) + ')';
+
+                size_t dest_elem{};
+                for (size_t elem = 0; elem < 4; ++elem) {
+                    if (!instr.tex.IsComponentEnabled(elem)) {
+                        // Skip disabled components
+                        continue;
+                    }
+                    regs.SetRegisterToFloat(instr.gpr0, elem, texture, 1, 4, false, dest_elem);
+                    ++dest_elem;
+                }
+                --shader.scope;
+                shader.AddLine("}");
+                break;
+            }
+            case OpCode::Id::TLD4S: {
+                const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
+                const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20);
+                // TODO(Subv): Figure out how the sampler type is encoded in the TLD4S instruction.
+                const std::string sampler =
+                    GetSampler(instr.sampler, Tegra::Shader::TextureType::Texture2D, false);
+                const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+                const std::string texture = "textureGather(" + sampler + ", coords, " +
+                                            std::to_string(instr.tld4s.component) + ')';
+                WriteTexsInstruction(instr, coord, texture);
+                break;
+            }
             default: {
                 LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->GetName());
                 UNREACHABLE();
@@ -1777,12 +2032,8 @@ private:
         default: {
             switch (opcode->GetId()) {
             case OpCode::Id::EXIT: {
-                // Final color output is currently hardcoded to GPR0-3 for fragment shaders
                 if (stage == Maxwell3D::Regs::ShaderStage::Fragment) {
-                    shader.AddLine("color.r = " + regs.GetRegisterAsFloat(0) + ';');
-                    shader.AddLine("color.g = " + regs.GetRegisterAsFloat(1) + ';');
-                    shader.AddLine("color.b = " + regs.GetRegisterAsFloat(2) + ';');
-                    shader.AddLine("color.a = " + regs.GetRegisterAsFloat(3) + ';');
+                    EmitFragmentOutputsWrite();
                 }
 
                 switch (instr.flow.cond) {
@@ -1841,13 +2092,13 @@ private:
                 ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer SSY is not supported");
 
                 u32 target = offset + instr.bra.GetBranchTarget();
-                shader.AddLine("ssy_target = " + std::to_string(target) + "u;");
+                EmitPushToSSYStack(target);
                 break;
             }
             case OpCode::Id::SYNC: {
                 // The SYNC opcode jumps to the address previously set by the SSY opcode
                 ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
-                shader.AddLine("{ jmp_to = ssy_target; break; }");
+                EmitPopFromSSYStack();
                 break;
             }
             case OpCode::Id::DEPBAR: {
@@ -1918,7 +2169,13 @@ private:
             } else {
                 labels.insert(subroutine.begin);
                 shader.AddLine("uint jmp_to = " + std::to_string(subroutine.begin) + "u;");
-                shader.AddLine("uint ssy_target = 0u;");
+
+                // TODO(Subv): Figure out the actual depth of the SSY stack, for now it seems
+                // unlikely that shaders will use 20 nested SSYs.
+                constexpr u32 SSY_STACK_SIZE = 20;
+                shader.AddLine("uint ssy_stack[" + std::to_string(SSY_STACK_SIZE) + "];");
+                shader.AddLine("uint ssy_stack_top = 0u;");
+
                 shader.AddLine("while (true) {");
                 ++shader.scope;
 

src/video_core/renderer_opengl/gl_shader_gen.cpp

@@ -38,6 +38,7 @@ out vec4 position;
 
 layout (std140) uniform vs_config {
     vec4 viewport_flip;
+    uvec4 instance_id;
 };
 
 void main() {
@@ -86,10 +87,11 @@ ProgramResult GenerateFragmentShader(const ShaderSetup& setup, const MaxwellFSCo
             .get_value_or({});
     out += R"(
 in vec4 position;
-out vec4 color;
+layout(location = 0) out vec4 color[8];
 
 layout (std140) uniform fs_config {
     vec4 viewport_flip;
+    uvec4 instance_id;
 };
 
 void main() {

src/video_core/renderer_opengl/gl_shader_gen.h

@@ -11,6 +11,7 @@
 #include <vector>
 #include "common/common_types.h"
 #include "common/hash.h"
+#include "video_core/engines/shader_bytecode.h"
 
 namespace GLShader {
 
@@ -72,8 +73,9 @@ class SamplerEntry {
     using Maxwell = Tegra::Engines::Maxwell3D::Regs;
 
 public:
-    SamplerEntry(Maxwell::ShaderStage stage, size_t offset, size_t index)
-        : offset(offset), stage(stage), sampler_index(index) {}
+    SamplerEntry(Maxwell::ShaderStage stage, size_t offset, size_t index,
+                 Tegra::Shader::TextureType type, bool is_array)
+        : offset(offset), stage(stage), sampler_index(index), type(type), is_array(is_array) {}
 
     size_t GetOffset() const {
         return offset;
@@ -88,8 +90,41 @@ public:
     }
 
     std::string GetName() const {
-        return std::string(TextureSamplerNames[static_cast<size_t>(stage)]) + '[' +
-               std::to_string(sampler_index) + ']';
+        return std::string(TextureSamplerNames[static_cast<size_t>(stage)]) + '_' +
+               std::to_string(sampler_index);
+    }
+
+    std::string GetTypeString() const {
+        using Tegra::Shader::TextureType;
+        std::string glsl_type;
+
+        switch (type) {
+        case TextureType::Texture1D:
+            glsl_type = "sampler1D";
+            break;
+        case TextureType::Texture2D:
+            glsl_type = "sampler2D";
+            break;
+        case TextureType::Texture3D:
+            glsl_type = "sampler3D";
+            break;
+        case TextureType::TextureCube:
+            glsl_type = "samplerCube";
+            break;
+        default:
+            UNIMPLEMENTED();
+        }
+        if (is_array)
+            glsl_type += "Array";
+        return glsl_type;
+    }
+
+    Tegra::Shader::TextureType GetType() const {
+        return type;
+    }
+
+    bool IsArray() const {
+        return is_array;
     }
 
     static std::string GetArrayName(Maxwell::ShaderStage stage) {
@@ -100,11 +135,14 @@ private:
     static constexpr std::array<const char*, Maxwell::MaxShaderStage> TextureSamplerNames = {
         "tex_vs", "tex_tessc", "tex_tesse", "tex_gs", "tex_fs",
     };
+
     /// Offset in TSC memory from which to read the sampler object, as specified by the sampling
     /// instruction.
     size_t offset;
-    Maxwell::ShaderStage stage; ///< Shader stage where this sampler was used.
-    size_t sampler_index;       ///< Value used to index into the generated GLSL sampler array.
+    Maxwell::ShaderStage stage;      ///< Shader stage where this sampler was used.
+    size_t sampler_index;            ///< Value used to index into the generated GLSL sampler array.
+    Tegra::Shader::TextureType type; ///< The type used to sample this texture (Texture2D, etc)
+    bool is_array; ///< Whether the texture is being sampled as an array texture or not.
 };
 
 struct ShaderEntries {

src/video_core/renderer_opengl/gl_shader_manager.cpp

@@ -37,11 +37,16 @@ void SetShaderUniformBlockBindings(GLuint shader) {
 } // namespace Impl
 
 void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage) {
-    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+    const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+    const auto& regs = gpu.regs;
+    const auto& state = gpu.state;
 
     // TODO(bunnei): Support more than one viewport
     viewport_flip[0] = regs.viewport_transform[0].scale_x < 0.0 ? -1.0f : 1.0f;
     viewport_flip[1] = regs.viewport_transform[0].scale_y < 0.0 ? -1.0f : 1.0f;
+
+    // We only assign the instance to the first component of the vector, the rest is just padding.
+    instance_id[0] = state.current_instance;
 }
 
 } // namespace GLShader

src/video_core/renderer_opengl/gl_shader_manager.h

@@ -24,14 +24,15 @@ void SetShaderUniformBlockBindings(GLuint shader);
 } // namespace Impl
 
 /// Uniform structure for the Uniform Buffer Object, all vectors must be 16-byte aligned
-// NOTE: Always keep a vec4 at the end. The GL spec is not clear wether the alignment at
+// NOTE: Always keep a vec4 at the end. The GL spec is not clear whether the alignment at
 //       the end of a uniform block is included in UNIFORM_BLOCK_DATA_SIZE or not.
 //       Not following that rule will cause problems on some AMD drivers.
 struct MaxwellUniformData {
     void SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage);
     alignas(16) GLvec4 viewport_flip;
+    alignas(16) GLuvec4 instance_id;
 };
-static_assert(sizeof(MaxwellUniformData) == 16, "MaxwellUniformData structure size is incorrect");
+static_assert(sizeof(MaxwellUniformData) == 32, "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_stream_buffer.h

@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#pragma once
+
 #include <tuple>
 #include <glad/glad.h>
 #include "common/common_types.h"

src/video_core/renderer_opengl/maxwell_to_gl.h

@@ -147,6 +147,8 @@ inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) {
         // GL_CLAMP_TO_BORDER to get the border color of the texture, and then sample the edge to
         // manually mix them. However the shader part of this is not yet implemented.
         return GL_CLAMP_TO_BORDER;
+    case Tegra::Texture::WrapMode::MirrorOnceClampToEdge:
+        return GL_MIRROR_CLAMP_TO_EDGE;
     }
     LOG_CRITICAL(Render_OpenGL, "Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode));
     UNREACHABLE();

src/video_core/renderer_opengl/renderer_opengl.cpp

@@ -16,6 +16,7 @@
 #include "core/memory.h"
 #include "core/settings.h"
 #include "core/tracer/recorder.h"
+#include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/renderer_opengl.h"
 #include "video_core/utils.h"
 
@@ -130,7 +131,7 @@ void RendererOpenGL::SwapBuffers(boost::optional<const Tegra::FramebufferConfig&
         }
 
         // Load the framebuffer from memory, draw it to the screen, and swap buffers
-        LoadFBToScreenInfo(*framebuffer, screen_info);
+        LoadFBToScreenInfo(*framebuffer);
         DrawScreen();
         render_window.SwapBuffers();
     }
@@ -142,14 +143,12 @@ void RendererOpenGL::SwapBuffers(boost::optional<const Tegra::FramebufferConfig&
 
     // Restore the rasterizer state
     prev_state.Apply();
-    RefreshRasterizerSetting();
 }
 
 /**
  * Loads framebuffer from emulated memory into the active OpenGL texture.
  */
-void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuffer,
-                                        ScreenInfo& screen_info) {
+void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuffer) {
     const u32 bytes_per_pixel{Tegra::FramebufferConfig::BytesPerPixel(framebuffer.pixel_format)};
     const u64 size_in_bytes{framebuffer.stride * framebuffer.height * bytes_per_pixel};
     const VAddr framebuffer_addr{framebuffer.address + framebuffer.offset};
@@ -162,8 +161,7 @@ void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuf
     // only allows rows to have a memory alignement of 4.
     ASSERT(framebuffer.stride % 4 == 0);
 
-    if (!rasterizer->AccelerateDisplay(framebuffer, framebuffer_addr, framebuffer.stride,
-                                       screen_info)) {
+    if (!rasterizer->AccelerateDisplay(framebuffer, framebuffer_addr, framebuffer.stride)) {
         // Reset the screen info's display texture to its own permanent texture
         screen_info.display_texture = screen_info.texture.resource.handle;
 
@@ -276,6 +274,14 @@ void RendererOpenGL::InitOpenGLObjects() {
     LoadColorToActiveGLTexture(0, 0, 0, 0, screen_info.texture);
 }
 
+void RendererOpenGL::CreateRasterizer() {
+    if (rasterizer) {
+        return;
+    }
+
+    rasterizer = std::make_unique<RasterizerOpenGL>(render_window, screen_info);
+}
+
 void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture,
                                                  const Tegra::FramebufferConfig& framebuffer) {
 
@@ -425,14 +431,14 @@ static void APIENTRY DebugHandler(GLenum source, GLenum type, GLuint id, GLenum
 
     switch (severity) {
     case GL_DEBUG_SEVERITY_HIGH:
-        LOG_ERROR(Render_OpenGL, format, str_source, str_type, id, message);
+        LOG_CRITICAL(Render_OpenGL, format, str_source, str_type, id, message);
         break;
     case GL_DEBUG_SEVERITY_MEDIUM:
         LOG_WARNING(Render_OpenGL, format, str_source, str_type, id, message);
         break;
     case GL_DEBUG_SEVERITY_NOTIFICATION:
     case GL_DEBUG_SEVERITY_LOW:
-        LOG_TRACE(Render_OpenGL, format, str_source, str_type, id, message);
+        LOG_DEBUG(Render_OpenGL, format, str_source, str_type, id, message);
         break;
     }
 }
@@ -463,8 +469,7 @@ bool RendererOpenGL::Init() {
     }
 
     InitOpenGLObjects();
-
-    RefreshRasterizerSetting();
+    CreateRasterizer();
 
     return true;
 }

src/video_core/renderer_opengl/renderer_opengl.h

@@ -59,6 +59,8 @@ public:
 
 private:
     void InitOpenGLObjects();
+    void CreateRasterizer();
+
     void ConfigureFramebufferTexture(TextureInfo& texture,
                                      const Tegra::FramebufferConfig& framebuffer);
     void DrawScreen();
@@ -66,7 +68,7 @@ private:
     void UpdateFramerate();
 
     // Loads framebuffer from emulated memory into the display information structure
-    void LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuffer, ScreenInfo& screen_info);
+    void LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuffer);
     // Fills active OpenGL texture with the given RGBA color.
     void LoadColorToActiveGLTexture(u8 color_r, u8 color_g, u8 color_b, u8 color_a,
                                     const TextureInfo& texture);

src/yuzu/configuration/configure_gamelist.cpp

@@ -56,7 +56,6 @@ void ConfigureGameList::setConfiguration() {
 
 void ConfigureGameList::applyConfiguration() {
     UISettings::values.show_unknown = ui->show_unknown->isChecked();
-    UISettings::values.icon_size = ui->icon_size_combobox->currentData().toUInt();
     UISettings::values.row_1_text_id = ui->row_1_text_combobox->currentData().toUInt();
     UISettings::values.row_2_text_id = ui->row_2_text_combobox->currentData().toUInt();
     Settings::Apply();

src/yuzu/game_list.cpp

@@ -16,6 +16,7 @@
 #include "common/string_util.h"
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/control_metadata.h"
+#include "core/file_sys/registered_cache.h"
 #include "core/file_sys/romfs.h"
 #include "core/file_sys/vfs_real.h"
 #include "core/loader/loader.h"
@@ -438,7 +439,7 @@ void GameListWorker::AddInstalledTitlesToGameList() {
 
         std::vector<u8> icon;
         std::string name;
-        u64 program_id;
+        u64 program_id = 0;
         loader->ReadProgramId(program_id);
 
         const auto& control =
@@ -509,7 +510,7 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign
             std::vector<u8> icon;
             const auto res1 = loader->ReadIcon(icon);
 
-            u64 program_id;
+            u64 program_id = 0;
             const auto res2 = loader->ReadProgramId(program_id);
 
             std::string name = " ";

src/yuzu/game_list_p.h

@@ -6,12 +6,15 @@
 
 #include <array>
 #include <atomic>
+#include <map>
+#include <memory>
 #include <utility>
 #include <QImage>
 #include <QRunnable>
 #include <QStandardItem>
 #include <QString>
 #include "common/string_util.h"
+#include "core/file_sys/content_archive.h"
 #include "ui_settings.h"
 #include "yuzu/util/util.h"
 

src/yuzu/main.cpp

@@ -27,7 +27,9 @@
 #include "common/string_util.h"
 #include "core/core.h"
 #include "core/crypto/key_manager.h"
+#include "core/file_sys/bis_factory.h"
 #include "core/file_sys/card_image.h"
+#include "core/file_sys/registered_cache.h"
 #include "core/file_sys/vfs_real.h"
 #include "core/gdbstub/gdbstub.h"
 #include "core/loader/loader.h"
@@ -377,6 +379,8 @@ bool GMainWindow::SupportsRequiredGLExtensions() {
         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");
 
     // Extensions required to support some texture formats.
     if (!GLAD_GL_EXT_texture_compression_s3tc)
@@ -628,24 +632,33 @@ void GMainWindow::OnMenuInstallToNAND() {
     QString filename = QFileDialog::getOpenFileName(this, tr("Install File"),
                                                     UISettings::values.roms_path, file_filter);
 
+    if (filename.isEmpty()) {
+        return;
+    }
+
     const auto qt_raw_copy = [this](FileSys::VirtualFile src, FileSys::VirtualFile dest) {
         if (src == nullptr || dest == nullptr)
             return false;
         if (!dest->Resize(src->GetSize()))
             return false;
 
-        QProgressDialog progress(fmt::format("Installing file \"{}\"...", src->GetName()).c_str(),
-                                 "Cancel", 0, src->GetSize() / 0x1000, this);
+        std::array<u8, 0x1000> buffer{};
+        const int progress_maximum = static_cast<int>(src->GetSize() / buffer.size());
+
+        QProgressDialog progress(
+            tr("Installing file \"%1\"...").arg(QString::fromStdString(src->GetName())),
+            tr("Cancel"), 0, progress_maximum, this);
         progress.setWindowModality(Qt::WindowModal);
 
-        std::array<u8, 0x1000> buffer{};
-        for (size_t i = 0; i < src->GetSize(); i += 0x1000) {
+        for (size_t i = 0; i < src->GetSize(); i += buffer.size()) {
             if (progress.wasCanceled()) {
                 dest->Resize(0);
                 return false;
             }
 
-            progress.setValue(i / 0x1000);
+            const int progress_value = static_cast<int>(i / buffer.size());
+            progress.setValue(progress_value);
+
             const auto read = src->Read(buffer.data(), buffer.size(), i);
             dest->Write(buffer.data(), read, i);
         }
@@ -668,92 +681,88 @@ void GMainWindow::OnMenuInstallToNAND() {
     };
 
     const auto overwrite = [this]() {
-        return QMessageBox::question(this, "Failed to Install",
-                                     "The file you are attempting to install already exists "
-                                     "in the cache. Would you like to overwrite it?") ==
+        return QMessageBox::question(this, tr("Failed to Install"),
+                                     tr("The file you are attempting to install already exists "
+                                        "in the cache. Would you like to overwrite it?")) ==
                QMessageBox::Yes;
     };
 
-    if (!filename.isEmpty()) {
-        if (filename.endsWith("xci", Qt::CaseInsensitive)) {
-            const auto xci = std::make_shared<FileSys::XCI>(
-                vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read));
-            if (xci->GetStatus() != Loader::ResultStatus::Success) {
-                failed();
-                return;
-            }
-            const auto res =
-                Service::FileSystem::GetUserNANDContents()->InstallEntry(xci, false, qt_raw_copy);
-            if (res == FileSys::InstallResult::Success) {
-                success();
-            } else {
-                if (res == FileSys::InstallResult::ErrorAlreadyExists) {
-                    if (overwrite()) {
-                        const auto res2 = Service::FileSystem::GetUserNANDContents()->InstallEntry(
-                            xci, true, qt_raw_copy);
-                        if (res2 == FileSys::InstallResult::Success) {
-                            success();
-                        } else {
-                            failed();
-                        }
+    if (filename.endsWith("xci", Qt::CaseInsensitive)) {
+        const auto xci = std::make_shared<FileSys::XCI>(
+            vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read));
+        if (xci->GetStatus() != Loader::ResultStatus::Success) {
+            failed();
+            return;
+        }
+        const auto res =
+            Service::FileSystem::GetUserNANDContents()->InstallEntry(xci, false, qt_raw_copy);
+        if (res == FileSys::InstallResult::Success) {
+            success();
+        } else {
+            if (res == FileSys::InstallResult::ErrorAlreadyExists) {
+                if (overwrite()) {
+                    const auto res2 = Service::FileSystem::GetUserNANDContents()->InstallEntry(
+                        xci, true, qt_raw_copy);
+                    if (res2 == FileSys::InstallResult::Success) {
+                        success();
+                    } else {
+                        failed();
                     }
+                }
+            } else {
+                failed();
+            }
+        }
+    } else {
+        const auto nca = std::make_shared<FileSys::NCA>(
+            vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read));
+        if (nca->GetStatus() != Loader::ResultStatus::Success) {
+            failed();
+            return;
+        }
+
+        const QStringList tt_options{tr("System Application"),
+                                     tr("System Archive"),
+                                     tr("System Application Update"),
+                                     tr("Firmware Package (Type A)"),
+                                     tr("Firmware Package (Type B)"),
+                                     tr("Game"),
+                                     tr("Game Update"),
+                                     tr("Game DLC"),
+                                     tr("Delta Title")};
+        bool ok;
+        const auto item = QInputDialog::getItem(
+            this, tr("Select NCA Install Type..."),
+            tr("Please select the type of title you would like to install this NCA as:\n(In "
+               "most instances, the default 'Game' is fine.)"),
+            tt_options, 5, false, &ok);
+
+        auto index = tt_options.indexOf(item);
+        if (!ok || index == -1) {
+            QMessageBox::warning(this, tr("Failed to Install"),
+                                 tr("The title type you selected for the NCA is invalid."));
+            return;
+        }
+
+        if (index >= 5)
+            index += 0x7B;
+
+        const auto res = Service::FileSystem::GetUserNANDContents()->InstallEntry(
+            nca, static_cast<FileSys::TitleType>(index), false, qt_raw_copy);
+        if (res == FileSys::InstallResult::Success) {
+            success();
+        } else if (res == FileSys::InstallResult::ErrorAlreadyExists) {
+            if (overwrite()) {
+                const auto res2 = Service::FileSystem::GetUserNANDContents()->InstallEntry(
+                    nca, static_cast<FileSys::TitleType>(index), true, qt_raw_copy);
+                if (res2 == FileSys::InstallResult::Success) {
+                    success();
                 } else {
                     failed();
                 }
             }
         } else {
-            const auto nca = std::make_shared<FileSys::NCA>(
-                vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read));
-            if (nca->GetStatus() != Loader::ResultStatus::Success) {
-                failed();
-                return;
-            }
-
-            static const QStringList tt_options{"System Application",
-                                                "System Archive",
-                                                "System Application Update",
-                                                "Firmware Package (Type A)",
-                                                "Firmware Package (Type B)",
-                                                "Game",
-                                                "Game Update",
-                                                "Game DLC",
-                                                "Delta Title"};
-            bool ok;
-            const auto item = QInputDialog::getItem(
-                this, tr("Select NCA Install Type..."),
-                tr("Please select the type of title you would like to install this NCA as:\n(In "
-                   "most instances, the default 'Game' is fine.)"),
-                tt_options, 5, false, &ok);
-
-            auto index = tt_options.indexOf(item);
-            if (!ok || index == -1) {
-                QMessageBox::warning(this, tr("Failed to Install"),
-                                     tr("The title type you selected for the NCA is invalid."));
-                return;
-            }
-
-            if (index >= 5)
-                index += 0x7B;
-
-            const auto res = Service::FileSystem::GetUserNANDContents()->InstallEntry(
-                nca, static_cast<FileSys::TitleType>(index), false, qt_raw_copy);
-            if (res == FileSys::InstallResult::Success) {
-                success();
-            } else {
-                if (res == FileSys::InstallResult::ErrorAlreadyExists) {
-                    if (overwrite()) {
-                        const auto res2 = Service::FileSystem::GetUserNANDContents()->InstallEntry(
-                            nca, static_cast<FileSys::TitleType>(index), true, qt_raw_copy);
-                        if (res2 == FileSys::InstallResult::Success) {
-                            success();
-                        } else {
-                            failed();
-                        }
-                    }
-                } else {
-                    failed();
-                }
-            }
+            failed();
         }
     }
 }

src/yuzu_cmd/emu_window/emu_window_sdl2.cpp

@@ -89,6 +89,8 @@ bool EmuWindow_SDL2::SupportsRequiredGLExtensions() {
         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");
 
     // Extensions required to support some texture formats.
     if (!GLAD_GL_EXT_texture_compression_s3tc)