shader_bytecode: Add Control Code enum 0xf

Control Code 0xf means to unconditionally execute the instruction. This
value is passed to most BRA, EXIT and SYNC instructions (among others)
but this may not always be the case.

src/core/core.cpp

@@ -148,7 +148,7 @@ struct System::Impl {
         telemetry_session = std::make_unique<Core::TelemetrySession>();
         service_manager = std::make_shared<Service::SM::ServiceManager>();
 
-        Service::Init(service_manager, virtual_filesystem);
+        Service::Init(service_manager, *virtual_filesystem);
         GDBStub::Init();
 
         renderer = VideoCore::CreateRenderer(emu_window);

src/core/crypto/key_manager.cpp

@@ -98,7 +98,7 @@ std::array<u8, 144> DecryptKeyblob(const std::array<u8, 176>& encrypted_keyblob,
     return keyblob;
 }
 
-void KeyManager::DeriveGeneralPurposeKeys(u8 crypto_revision) {
+void KeyManager::DeriveGeneralPurposeKeys(std::size_t crypto_revision) {
     const auto kek_generation_source =
         GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration));
     const auto key_generation_source =
@@ -147,31 +147,38 @@ boost::optional<Key128> DeriveSDSeed() {
                                    "rb+");
     if (!save_43.IsOpen())
         return boost::none;
+
     const FileUtil::IOFile sd_private(
         FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir) + "/Nintendo/Contents/private", "rb+");
     if (!sd_private.IsOpen())
         return boost::none;
 
-    sd_private.Seek(0, SEEK_SET);
     std::array<u8, 0x10> private_seed{};
-    if (sd_private.ReadBytes(private_seed.data(), private_seed.size()) != 0x10)
+    if (sd_private.ReadBytes(private_seed.data(), private_seed.size()) != private_seed.size()) {
         return boost::none;
+    }
 
     std::array<u8, 0x10> buffer{};
     std::size_t offset = 0;
     for (; offset + 0x10 < save_43.GetSize(); ++offset) {
-        save_43.Seek(offset, SEEK_SET);
+        if (!save_43.Seek(offset, SEEK_SET)) {
+            return boost::none;
+        }
+
         save_43.ReadBytes(buffer.data(), buffer.size());
-        if (buffer == private_seed)
+        if (buffer == private_seed) {
             break;
+        }
     }
 
-    if (offset + 0x10 >= save_43.GetSize())
+    if (!save_43.Seek(offset + 0x10, SEEK_SET)) {
         return boost::none;
+    }
 
     Key128 seed{};
-    save_43.Seek(offset + 0x10, SEEK_SET);
-    save_43.ReadBytes(seed.data(), seed.size());
+    if (save_43.ReadBytes(seed.data(), seed.size()) != seed.size()) {
+        return boost::none;
+    }
     return seed;
 }
 
@@ -234,7 +241,9 @@ std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save) {
         return {};
 
     std::vector<u8> buffer(ticket_save.GetSize());
-    ticket_save.ReadBytes(buffer.data(), buffer.size());
+    if (ticket_save.ReadBytes(buffer.data(), buffer.size()) != buffer.size()) {
+        return {};
+    }
 
     std::vector<TicketRaw> out;
     u32 magic{};
@@ -261,6 +270,9 @@ static std::array<u8, size> operator^(const std::array<u8, size>& lhs,
 
 template <size_t target_size, size_t in_size>
 static std::array<u8, target_size> MGF1(const std::array<u8, in_size>& seed) {
+    // Avoids truncation overflow within the loop below.
+    static_assert(target_size <= 0xFF);
+
     std::array<u8, in_size + 4> seed_exp{};
     std::memcpy(seed_exp.data(), seed.data(), in_size);
 
@@ -268,7 +280,7 @@ static std::array<u8, target_size> MGF1(const std::array<u8, in_size>& seed) {
     size_t i = 0;
     while (out.size() < target_size) {
         out.resize(out.size() + 0x20);
-        seed_exp[in_size + 3] = i;
+        seed_exp[in_size + 3] = static_cast<u8>(i);
         mbedtls_sha256(seed_exp.data(), seed_exp.size(), out.data() + out.size() - 0x20, 0);
         ++i;
     }
@@ -299,10 +311,11 @@ boost::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
     std::memcpy(&cert_authority, ticket.data() + 0x140, sizeof(cert_authority));
     if (cert_authority == 0)
         return boost::none;
-    if (cert_authority != Common::MakeMagic('R', 'o', 'o', 't'))
+    if (cert_authority != Common::MakeMagic('R', 'o', 'o', 't')) {
         LOG_INFO(Crypto,
                  "Attempting to parse ticket with non-standard certificate authority {:08X}.",
                  cert_authority);
+    }
 
     Key128 rights_id;
     std::memcpy(rights_id.data(), ticket.data() + 0x2A0, sizeof(Key128));
@@ -871,9 +884,9 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) {
                                      "/system/save/80000000000000e2",
                                  "rb+");
 
+    const auto blob2 = GetTicketblob(save2);
     auto res = GetTicketblob(save1);
-    const auto res2 = GetTicketblob(save2);
-    std::copy(res2.begin(), res2.end(), std::back_inserter(res));
+    res.insert(res.end(), blob2.begin(), blob2.end());
 
     for (const auto& raw : res) {
         const auto pair = ParseTicket(raw, rsa_key);

src/core/crypto/key_manager.h

@@ -175,7 +175,7 @@ private:
     void WriteKeyToFile(KeyCategory category, std::string_view keyname,
                         const std::array<u8, Size>& key);
 
-    void DeriveGeneralPurposeKeys(u8 crypto_revision);
+    void DeriveGeneralPurposeKeys(std::size_t crypto_revision);
 
     void SetKeyWrapped(S128KeyType id, Key128 key, u64 field1 = 0, u64 field2 = 0);
     void SetKeyWrapped(S256KeyType id, Key256 key, u64 field1 = 0, u64 field2 = 0);

src/core/crypto/partition_data_manager.cpp

@@ -11,7 +11,6 @@
 #include <array>
 #include <cctype>
 #include <cstring>
-#include <boost/optional/optional.hpp>
 #include <mbedtls/sha256.h>
 #include "common/assert.h"
 #include "common/common_funcs.h"
@@ -19,7 +18,7 @@
 #include "common/hex_util.h"
 #include "common/logging/log.h"
 #include "common/string_util.h"
-#include "core/crypto/ctr_encryption_layer.h"
+#include "common/swap.h"
 #include "core/crypto/key_manager.h"
 #include "core/crypto/partition_data_manager.h"
 #include "core/crypto/xts_encryption_layer.h"
@@ -302,7 +301,7 @@ FileSys::VirtualFile FindFileInDirWithNames(const FileSys::VirtualDir& dir,
     return nullptr;
 }
 
-PartitionDataManager::PartitionDataManager(FileSys::VirtualDir sysdata_dir)
+PartitionDataManager::PartitionDataManager(const FileSys::VirtualDir& sysdata_dir)
     : boot0(FindFileInDirWithNames(sysdata_dir, "BOOT0")),
       fuses(FindFileInDirWithNames(sysdata_dir, "fuse")),
       kfuses(FindFileInDirWithNames(sysdata_dir, "kfuse")),
@@ -314,13 +313,14 @@ PartitionDataManager::PartitionDataManager(FileSys::VirtualDir sysdata_dir)
           FindFileInDirWithNames(sysdata_dir, "BCPKG2-5-Repair-Main"),
           FindFileInDirWithNames(sysdata_dir, "BCPKG2-6-Repair-Sub"),
       }),
+      prodinfo(FindFileInDirWithNames(sysdata_dir, "PRODINFO")),
       secure_monitor(FindFileInDirWithNames(sysdata_dir, "secmon")),
       package1_decrypted(FindFileInDirWithNames(sysdata_dir, "pkg1_decr")),
       secure_monitor_bytes(secure_monitor == nullptr ? std::vector<u8>{}
                                                      : secure_monitor->ReadAllBytes()),
       package1_decrypted_bytes(package1_decrypted == nullptr ? std::vector<u8>{}
-                                                             : package1_decrypted->ReadAllBytes()),
-      prodinfo(FindFileInDirWithNames(sysdata_dir, "PRODINFO")) {}
+                                                             : package1_decrypted->ReadAllBytes()) {
+}
 
 PartitionDataManager::~PartitionDataManager() = default;
 
@@ -332,18 +332,19 @@ FileSys::VirtualFile PartitionDataManager::GetBoot0Raw() const {
     return boot0;
 }
 
-std::array<u8, 176> PartitionDataManager::GetEncryptedKeyblob(u8 index) const {
-    if (HasBoot0() && index < 32)
+PartitionDataManager::EncryptedKeyBlob PartitionDataManager::GetEncryptedKeyblob(
+    std::size_t index) const {
+    if (HasBoot0() && index < NUM_ENCRYPTED_KEYBLOBS)
         return GetEncryptedKeyblobs()[index];
     return {};
 }
 
-std::array<std::array<u8, 176>, 32> PartitionDataManager::GetEncryptedKeyblobs() const {
+PartitionDataManager::EncryptedKeyBlobs PartitionDataManager::GetEncryptedKeyblobs() const {
     if (!HasBoot0())
         return {};
 
-    std::array<std::array<u8, 176>, 32> out{};
-    for (size_t i = 0; i < 0x20; ++i)
+    EncryptedKeyBlobs out{};
+    for (size_t i = 0; i < out.size(); ++i)
         boot0->Read(out[i].data(), out[i].size(), 0x180000 + i * 0x200);
     return out;
 }
@@ -389,7 +390,7 @@ std::array<u8, 16> PartitionDataManager::GetKeyblobMACKeySource() const {
     return FindKeyFromHex(package1_decrypted_bytes, source_hashes[0]);
 }
 
-std::array<u8, 16> PartitionDataManager::GetKeyblobKeySource(u8 revision) const {
+std::array<u8, 16> PartitionDataManager::GetKeyblobKeySource(std::size_t revision) const {
     if (keyblob_source_hashes[revision] == SHA256Hash{}) {
         LOG_WARNING(Crypto,
                     "No keyblob source hash for crypto revision {:02X}! Cannot derive keys...",
@@ -446,7 +447,7 @@ bool AttemptDecrypt(const std::array<u8, 16>& key, Package2Header& header) {
     return false;
 }
 
-void PartitionDataManager::DecryptPackage2(std::array<std::array<u8, 16>, 0x20> package2_keys,
+void PartitionDataManager::DecryptPackage2(const std::array<Key128, 0x20>& package2_keys,
                                            Package2Type type) {
     FileSys::VirtualFile file = std::make_shared<FileSys::OffsetVfsFile>(
         package2[static_cast<size_t>(type)],
@@ -456,43 +457,38 @@ void PartitionDataManager::DecryptPackage2(std::array<std::array<u8, 16>, 0x20>
     if (file->ReadObject(&header) != sizeof(Package2Header))
         return;
 
-    u8 revision = 0xFF;
+    std::size_t revision = 0xFF;
     if (header.magic != Common::MakeMagic('P', 'K', '2', '1')) {
-        for (size_t i = 0; i < package2_keys.size(); ++i) {
-            if (AttemptDecrypt(package2_keys[i], header))
+        for (std::size_t i = 0; i < package2_keys.size(); ++i) {
+            if (AttemptDecrypt(package2_keys[i], header)) {
                 revision = i;
+            }
         }
     }
 
     if (header.magic != Common::MakeMagic('P', 'K', '2', '1'))
         return;
 
-    const std::vector<u8> s1_iv(header.section_ctr[1].begin(), header.section_ctr[1].end());
-
     const auto a = std::make_shared<FileSys::OffsetVfsFile>(
         file, header.section_size[1], header.section_size[0] + sizeof(Package2Header));
 
     auto c = a->ReadAllBytes();
 
     AESCipher<Key128> cipher(package2_keys[revision], Mode::CTR);
-    cipher.SetIV(s1_iv);
+    cipher.SetIV({header.section_ctr[1].begin(), header.section_ctr[1].end()});
     cipher.Transcode(c.data(), c.size(), c.data(), Op::Decrypt);
 
-    // package2_decrypted[static_cast<size_t>(type)] = s1;
-
     INIHeader ini;
     std::memcpy(&ini, c.data(), sizeof(INIHeader));
     if (ini.magic != Common::MakeMagic('I', 'N', 'I', '1'))
         return;
 
-    std::map<u64, KIPHeader> kips{};
     u64 offset = sizeof(INIHeader);
     for (size_t i = 0; i < ini.process_count; ++i) {
         KIPHeader kip;
         std::memcpy(&kip, c.data() + offset, sizeof(KIPHeader));
         if (kip.magic != Common::MakeMagic('K', 'I', 'P', '1'))
             return;
-        kips.emplace(offset, kip);
 
         const auto name =
             Common::StringFromFixedZeroTerminatedBuffer(kip.name.data(), kip.name.size());
@@ -503,33 +499,29 @@ void PartitionDataManager::DecryptPackage2(std::array<std::array<u8, 16>, 0x20>
             continue;
         }
 
-        std::vector<u8> text(kip.sections[0].size_compressed);
-        std::vector<u8> rodata(kip.sections[1].size_compressed);
-        std::vector<u8> data(kip.sections[2].size_compressed);
+        const u64 initial_offset = sizeof(KIPHeader) + offset;
+        const auto text_begin = c.cbegin() + initial_offset;
+        const auto text_end = text_begin + kip.sections[0].size_compressed;
+        const std::vector<u8> text = DecompressBLZ({text_begin, text_end});
 
-        u64 offset_sec = sizeof(KIPHeader) + offset;
-        std::memcpy(text.data(), c.data() + offset_sec, text.size());
-        offset_sec += text.size();
-        std::memcpy(rodata.data(), c.data() + offset_sec, rodata.size());
-        offset_sec += rodata.size();
-        std::memcpy(data.data(), c.data() + offset_sec, data.size());
+        const auto rodata_end = text_end + kip.sections[1].size_compressed;
+        const std::vector<u8> rodata = DecompressBLZ({text_end, rodata_end});
 
-        offset += sizeof(KIPHeader) + kip.sections[0].size_compressed +
-                  kip.sections[1].size_compressed + kip.sections[2].size_compressed;
+        const auto data_end = rodata_end + kip.sections[2].size_compressed;
+        const std::vector<u8> data = DecompressBLZ({rodata_end, data_end});
 
-        text = DecompressBLZ(text);
-        rodata = DecompressBLZ(rodata);
-        data = DecompressBLZ(data);
+        std::vector<u8> out;
+        out.reserve(text.size() + rodata.size() + data.size());
+        out.insert(out.end(), text.begin(), text.end());
+        out.insert(out.end(), rodata.begin(), rodata.end());
+        out.insert(out.end(), data.begin(), data.end());
 
-        std::vector<u8> out(text.size() + rodata.size() + data.size());
-        std::memcpy(out.data(), text.data(), text.size());
-        std::memcpy(out.data() + text.size(), rodata.data(), rodata.size());
-        std::memcpy(out.data() + text.size() + rodata.size(), data.data(), data.size());
+        offset += sizeof(KIPHeader) + out.size();
 
         if (name == "FS")
-            package2_fs[static_cast<size_t>(type)] = out;
+            package2_fs[static_cast<size_t>(type)] = std::move(out);
         else if (name == "spl")
-            package2_spl[static_cast<size_t>(type)] = out;
+            package2_spl[static_cast<size_t>(type)] = std::move(out);
     }
 }
 

src/core/crypto/partition_data_manager.h

@@ -5,9 +5,7 @@
 #pragma once
 
 #include <vector>
-#include "common/common_funcs.h"
 #include "common/common_types.h"
-#include "common/swap.h"
 #include "core/file_sys/vfs_types.h"
 
 namespace Core::Crypto {
@@ -24,15 +22,20 @@ enum class Package2Type {
 class PartitionDataManager {
 public:
     static const u8 MAX_KEYBLOB_SOURCE_HASH;
+    static constexpr std::size_t NUM_ENCRYPTED_KEYBLOBS = 32;
+    static constexpr std::size_t ENCRYPTED_KEYBLOB_SIZE = 0xB0;
 
-    explicit PartitionDataManager(FileSys::VirtualDir sysdata_dir);
+    using EncryptedKeyBlob = std::array<u8, ENCRYPTED_KEYBLOB_SIZE>;
+    using EncryptedKeyBlobs = std::array<EncryptedKeyBlob, NUM_ENCRYPTED_KEYBLOBS>;
+
+    explicit PartitionDataManager(const FileSys::VirtualDir& sysdata_dir);
     ~PartitionDataManager();
 
     // BOOT0
     bool HasBoot0() const;
     FileSys::VirtualFile GetBoot0Raw() const;
-    std::array<u8, 0xB0> GetEncryptedKeyblob(u8 index) const;
-    std::array<std::array<u8, 0xB0>, 0x20> GetEncryptedKeyblobs() const;
+    EncryptedKeyBlob GetEncryptedKeyblob(std::size_t index) const;
+    EncryptedKeyBlobs GetEncryptedKeyblobs() const;
     std::vector<u8> GetSecureMonitor() const;
     std::array<u8, 0x10> GetPackage2KeySource() const;
     std::array<u8, 0x10> GetAESKekGenerationSource() const;
@@ -43,7 +46,7 @@ public:
     std::vector<u8> GetPackage1Decrypted() const;
     std::array<u8, 0x10> GetMasterKeySource() const;
     std::array<u8, 0x10> GetKeyblobMACKeySource() const;
-    std::array<u8, 0x10> GetKeyblobKeySource(u8 revision) const;
+    std::array<u8, 0x10> GetKeyblobKeySource(std::size_t revision) const;
 
     // Fuses
     bool HasFuses() const;
@@ -57,7 +60,8 @@ public:
     // Package2
     bool HasPackage2(Package2Type type = Package2Type::NormalMain) const;
     FileSys::VirtualFile GetPackage2Raw(Package2Type type = Package2Type::NormalMain) const;
-    void DecryptPackage2(std::array<std::array<u8, 16>, 0x20> package2, Package2Type type);
+    void DecryptPackage2(const std::array<std::array<u8, 16>, 0x20>& package2_keys,
+                         Package2Type type);
     const std::vector<u8>& GetPackage2FSDecompressed(
         Package2Type type = Package2Type::NormalMain) const;
     std::array<u8, 0x10> GetKeyAreaKeyApplicationSource(

src/core/hle/kernel/process.h

@@ -24,6 +24,7 @@ class ProgramMetadata;
 namespace Kernel {
 
 class KernelCore;
+class ResourceLimit;
 
 struct AddressMapping {
     // Address and size must be page-aligned
@@ -57,9 +58,23 @@ union ProcessFlags {
     BitField<12, 1, u16> loaded_high; ///< Application loaded high (not at 0x00100000).
 };
 
-enum class ProcessStatus { Created, Running, Exited };
-
-class ResourceLimit;
+/**
+ * Indicates the status of a Process instance.
+ *
+ * @note These match the values as used by kernel,
+ *       so new entries should only be added if RE
+ *       shows that a new value has been introduced.
+ */
+enum class ProcessStatus {
+    Created,
+    CreatedWithDebuggerAttached,
+    Running,
+    WaitingForDebuggerToAttach,
+    DebuggerAttached,
+    Exiting,
+    Exited,
+    DebugBreak,
+};
 
 struct CodeSet final {
     struct Segment {

src/core/hle/kernel/svc.cpp

@@ -389,6 +389,12 @@ static void Break(u32 reason, u64 info1, u64 info2) {
             "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}",
             reason, info1, info2);
         ASSERT(false);
+
+        Core::CurrentProcess()->PrepareForTermination();
+
+        // Kill the current thread
+        GetCurrentThread()->Stop();
+        Core::System::GetInstance().PrepareReschedule();
     }
 }
 
@@ -1092,6 +1098,29 @@ static ResultCode ClearEvent(Handle handle) {
     return RESULT_SUCCESS;
 }
 
+static ResultCode GetProcessInfo(u64* out, Handle process_handle, u32 type) {
+    LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, type=0x{:X}", process_handle, type);
+
+    // This function currently only allows retrieving a process' status.
+    enum class InfoType {
+        Status,
+    };
+
+    const auto& kernel = Core::System::GetInstance().Kernel();
+    const auto process = kernel.HandleTable().Get<Process>(process_handle);
+    if (!process) {
+        return ERR_INVALID_HANDLE;
+    }
+
+    const auto info_type = static_cast<InfoType>(type);
+    if (info_type != InfoType::Status) {
+        return ERR_INVALID_ENUM_VALUE;
+    }
+
+    *out = static_cast<u64>(process->GetStatus());
+    return RESULT_SUCCESS;
+}
+
 namespace {
 struct FunctionDef {
     using Func = void();
@@ -1227,7 +1256,7 @@ static const FunctionDef SVC_Table[] = {
     {0x79, nullptr, "CreateProcess"},
     {0x7A, nullptr, "StartProcess"},
     {0x7B, nullptr, "TerminateProcess"},
-    {0x7C, nullptr, "GetProcessInfo"},
+    {0x7C, SvcWrap<GetProcessInfo>, "GetProcessInfo"},
     {0x7D, nullptr, "CreateResourceLimit"},
     {0x7E, nullptr, "SetResourceLimitLimitValue"},
     {0x7F, nullptr, "CallSecureMonitor"},

src/core/hle/kernel/svc_wrap.h

@@ -77,6 +77,14 @@ void SvcWrap() {
     FuncReturn(retval);
 }
 
+template <ResultCode func(u64*, u32, u32)>
+void SvcWrap() {
+    u64 param_1 = 0;
+    u32 retval = func(&param_1, static_cast<u32>(Param(1)), static_cast<u32>(Param(2))).raw;
+    Core::CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(retval);
+}
+
 template <ResultCode func(u32, u64)>
 void SvcWrap() {
     FuncReturn(func(static_cast<u32>(Param(0)), Param(1)).raw);

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

@@ -361,19 +361,19 @@ FileSys::VirtualDir GetModificationLoadRoot(u64 title_id) {
     return bis_factory->GetModificationLoadRoot(title_id);
 }
 
-void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite) {
+void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite) {
     if (overwrite) {
         bis_factory = nullptr;
         save_data_factory = nullptr;
         sdmc_factory = nullptr;
     }
 
-    auto nand_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir),
-                                             FileSys::Mode::ReadWrite);
-    auto sd_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir),
-                                           FileSys::Mode::ReadWrite);
-    auto load_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir),
-                                             FileSys::Mode::ReadWrite);
+    auto nand_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir),
+                                            FileSys::Mode::ReadWrite);
+    auto sd_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir),
+                                          FileSys::Mode::ReadWrite);
+    auto load_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir),
+                                            FileSys::Mode::ReadWrite);
 
     if (bis_factory == nullptr)
         bis_factory = std::make_unique<FileSys::BISFactory>(nand_directory, load_directory);
@@ -383,7 +383,7 @@ void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite) {
         sdmc_factory = std::make_unique<FileSys::SDMCFactory>(std::move(sd_directory));
 }
 
-void InstallInterfaces(SM::ServiceManager& service_manager, const FileSys::VirtualFilesystem& vfs) {
+void InstallInterfaces(SM::ServiceManager& service_manager, FileSys::VfsFilesystem& vfs) {
     romfs_factory = nullptr;
     CreateFactories(vfs, false);
     std::make_shared<FSP_LDR>()->InstallAsService(service_manager);

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

@@ -57,9 +57,9 @@ FileSys::VirtualDir GetModificationLoadRoot(u64 title_id);
 
 // Creates the SaveData, SDMC, and BIS Factories. Should be called once and before any function
 // above is called.
-void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite = true);
+void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite = true);
 
-void InstallInterfaces(SM::ServiceManager& service_manager, const FileSys::VirtualFilesystem& vfs);
+void InstallInterfaces(SM::ServiceManager& service_manager, FileSys::VfsFilesystem& vfs);
 
 // A class that wraps a VfsDirectory with methods that return ResultVal and ResultCode instead of
 // pointers and booleans. This makes using a VfsDirectory with switch services much easier and

src/core/hle/service/service.cpp

@@ -197,7 +197,7 @@ ResultCode ServiceFrameworkBase::HandleSyncRequest(Kernel::HLERequestContext& co
 // Module interface
 
 /// Initialize ServiceManager
-void Init(std::shared_ptr<SM::ServiceManager>& sm, const FileSys::VirtualFilesystem& rfs) {
+void Init(std::shared_ptr<SM::ServiceManager>& sm, FileSys::VfsFilesystem& vfs) {
     // NVFlinger needs to be accessed by several services like Vi and AppletOE so we instantiate it
     // here and pass it into the respective InstallInterfaces functions.
     auto nv_flinger = std::make_shared<NVFlinger::NVFlinger>();
@@ -220,7 +220,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, const FileSys::VirtualFilesys
     EUPLD::InstallInterfaces(*sm);
     Fatal::InstallInterfaces(*sm);
     FGM::InstallInterfaces(*sm);
-    FileSystem::InstallInterfaces(*sm, rfs);
+    FileSystem::InstallInterfaces(*sm, vfs);
     Friend::InstallInterfaces(*sm);
     GRC::InstallInterfaces(*sm);
     HID::InstallInterfaces(*sm);

src/core/hle/service/service.h

@@ -180,8 +180,7 @@ private:
 };
 
 /// Initialize ServiceManager
-void Init(std::shared_ptr<SM::ServiceManager>& sm,
-          const std::shared_ptr<FileSys::VfsFilesystem>& vfs);
+void Init(std::shared_ptr<SM::ServiceManager>& sm, FileSys::VfsFilesystem& vfs);
 
 /// Shutdown ServiceManager
 void Shutdown();

src/video_core/engines/fermi_2d.cpp

@@ -62,14 +62,16 @@ void Fermi2D::HandleSurfaceCopy() {
         u8* dst_buffer = Memory::GetPointer(dest_cpu);
         if (!regs.src.linear && regs.dst.linear) {
             // If the input is tiled and the output is linear, deswizzle the input and copy it over.
-            Texture::CopySwizzledData(regs.src.width, regs.src.height, src_bytes_per_pixel,
-                                      dst_bytes_per_pixel, src_buffer, dst_buffer, true,
-                                      regs.src.BlockHeight());
+            Texture::CopySwizzledData(regs.src.width, regs.src.height, regs.src.depth,
+                                      src_bytes_per_pixel, dst_bytes_per_pixel, src_buffer,
+                                      dst_buffer, true, regs.src.BlockHeight(),
+                                      regs.src.BlockDepth());
         } else {
             // If the input is linear and the output is tiled, swizzle the input and copy it over.
-            Texture::CopySwizzledData(regs.src.width, regs.src.height, src_bytes_per_pixel,
-                                      dst_bytes_per_pixel, dst_buffer, src_buffer, false,
-                                      regs.dst.BlockHeight());
+            Texture::CopySwizzledData(regs.src.width, regs.src.height, regs.src.depth,
+                                      src_bytes_per_pixel, dst_bytes_per_pixel, dst_buffer,
+                                      src_buffer, false, regs.dst.BlockHeight(),
+                                      regs.dst.BlockDepth());
         }
     }
 }

src/video_core/engines/maxwell_dma.cpp

@@ -68,12 +68,14 @@ void MaxwellDMA::HandleCopy() {
 
     if (regs.exec.is_dst_linear && !regs.exec.is_src_linear) {
         // If the input is tiled and the output is linear, deswizzle the input and copy it over.
-        Texture::CopySwizzledData(regs.src_params.size_x, regs.src_params.size_y, 1, 1, src_buffer,
-                                  dst_buffer, true, regs.src_params.BlockHeight());
+        Texture::CopySwizzledData(regs.src_params.size_x, regs.src_params.size_y,
+                                  regs.src_params.size_z, 1, 1, src_buffer, dst_buffer, true,
+                                  regs.src_params.BlockHeight(), regs.src_params.BlockDepth());
     } else {
         // If the input is linear and the output is tiled, swizzle the input and copy it over.
-        Texture::CopySwizzledData(regs.dst_params.size_x, regs.dst_params.size_y, 1, 1, dst_buffer,
-                                  src_buffer, false, regs.dst_params.BlockHeight());
+        Texture::CopySwizzledData(regs.dst_params.size_x, regs.dst_params.size_y,
+                                  regs.dst_params.size_z, 1, 1, dst_buffer, src_buffer, false,
+                                  regs.dst_params.BlockHeight(), regs.dst_params.BlockDepth());
     }
 }
 

src/video_core/engines/maxwell_dma.h

@@ -43,6 +43,10 @@ public:
             u32 BlockHeight() const {
                 return 1 << block_height;
             }
+
+            u32 BlockDepth() const {
+                return 1 << block_depth;
+            }
         };
 
         static_assert(sizeof(Parameters) == 24, "Parameters has wrong size");

src/video_core/engines/shader_bytecode.h

@@ -267,7 +267,7 @@ enum class ControlCode : u64 {
     GTU = 12,
     NEU = 13,
     GEU = 14,
-    //
+    T = 15,
     OFF = 16,
     LO = 17,
     SFF = 18,

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -286,7 +286,8 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) {
             &ubo, sizeof(ubo), static_cast<std::size_t>(uniform_buffer_alignment));
 
         // Bind the buffer
-        glBindBufferRange(GL_UNIFORM_BUFFER, stage, buffer_cache.GetHandle(), offset, sizeof(ubo));
+        glBindBufferRange(GL_UNIFORM_BUFFER, static_cast<GLuint>(stage), buffer_cache.GetHandle(),
+                          offset, static_cast<GLsizeiptr>(sizeof(ubo)));
 
         Shader shader{shader_cache.GetStageProgram(program)};
 

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -231,6 +231,8 @@ 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
     {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X8
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X5
+    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X4
 
     // Depth formats
     {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F
@@ -277,7 +279,9 @@ static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType
 static bool IsPixelFormatASTC(PixelFormat format) {
     switch (format) {
     case PixelFormat::ASTC_2D_4X4:
+    case PixelFormat::ASTC_2D_5X4:
     case PixelFormat::ASTC_2D_8X8:
+    case PixelFormat::ASTC_2D_8X5:
         return true;
     default:
         return false;
@@ -288,8 +292,12 @@ static std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) {
     switch (format) {
     case PixelFormat::ASTC_2D_4X4:
         return {4, 4};
+    case PixelFormat::ASTC_2D_5X4:
+        return {5, 4};
     case PixelFormat::ASTC_2D_8X8:
         return {8, 8};
+    case PixelFormat::ASTC_2D_8X5:
+        return {8, 5};
     default:
         LOG_CRITICAL(HW_GPU, "Unhandled format: {}", static_cast<u32>(format));
         UNREACHABLE();
@@ -323,8 +331,8 @@ static bool IsFormatBCn(PixelFormat format) {
 }
 
 template <bool morton_to_gl, PixelFormat format>
-void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, std::size_t gl_buffer_size,
-                VAddr addr) {
+void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth, u8* gl_buffer,
+                std::size_t gl_buffer_size, VAddr addr) {
     constexpr u32 bytes_per_pixel = SurfaceParams::GetFormatBpp(format) / CHAR_BIT;
     constexpr u32 gl_bytes_per_pixel = CachedSurface::GetGLBytesPerPixel(format);
 
@@ -333,7 +341,7 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, std::si
         // pixel values.
         const u32 tile_size{IsFormatBCn(format) ? 4U : 1U};
         const std::vector<u8> data = Tegra::Texture::UnswizzleTexture(
-            addr, tile_size, bytes_per_pixel, stride, height, block_height);
+            addr, tile_size, bytes_per_pixel, stride, height, depth, block_height, block_depth);
         const std::size_t size_to_copy{std::min(gl_buffer_size, data.size())};
         memcpy(gl_buffer, data.data(), size_to_copy);
     } else {
@@ -345,7 +353,7 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, u8* gl_buffer, std::si
     }
 }
 
-static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
+static constexpr std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
                             SurfaceParams::MaxPixelFormat>
     morton_to_gl_fns = {
         // clang-format off
@@ -395,6 +403,8 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
         MortonCopy<true, PixelFormat::RG32UI>,
         MortonCopy<true, PixelFormat::R32UI>,
         MortonCopy<true, PixelFormat::ASTC_2D_8X8>,
+        MortonCopy<true, PixelFormat::ASTC_2D_8X5>,
+        MortonCopy<true, PixelFormat::ASTC_2D_5X4>,
         MortonCopy<true, PixelFormat::Z32F>,
         MortonCopy<true, PixelFormat::Z16>,
         MortonCopy<true, PixelFormat::Z24S8>,
@@ -403,7 +413,7 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
         // clang-format on
 };
 
-static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
+static constexpr std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
                             SurfaceParams::MaxPixelFormat>
     gl_to_morton_fns = {
         // clang-format off
@@ -455,6 +465,8 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
         MortonCopy<false, PixelFormat::RG32UI>,
         MortonCopy<false, PixelFormat::R32UI>,
         nullptr,
+        nullptr,
+        nullptr,
         MortonCopy<false, PixelFormat::Z32F>,
         MortonCopy<false, PixelFormat::Z16>,
         MortonCopy<false, PixelFormat::Z24S8>,
@@ -790,7 +802,9 @@ static void ConvertFormatAsNeeded_LoadGLBuffer(std::vector<u8>& data, PixelForma
                                                u32 width, u32 height) {
     switch (pixel_format) {
     case PixelFormat::ASTC_2D_4X4:
-    case PixelFormat::ASTC_2D_8X8: {
+    case PixelFormat::ASTC_2D_8X8:
+    case PixelFormat::ASTC_2D_8X5:
+    case PixelFormat::ASTC_2D_5X4: {
         // Convert ASTC pixel formats to RGBA8, as most desktop GPUs do not support ASTC.
         u32 block_width{};
         u32 block_height{};
@@ -827,36 +841,23 @@ void CachedSurface::LoadGLBuffer() {
 
     if (params.is_tiled) {
         gl_buffer.resize(total_size);
+        u32 depth = params.depth;
+        u32 block_depth = params.block_depth;
 
         ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
                    params.block_width, static_cast<u32>(params.target));
-        ASSERT_MSG(params.block_depth == 1, "Block depth is defined as {} on texture type {}",
-                   params.block_depth, static_cast<u32>(params.target));
 
-        // TODO(bunnei): This only unswizzles and copies a 2D texture - we do not yet know how to do
-        // this for 3D textures, etc.
-        switch (params.target) {
-        case SurfaceParams::SurfaceTarget::Texture2D:
-            // Pass impl. to the fallback code below
-            break;
-        case SurfaceParams::SurfaceTarget::Texture2DArray:
-        case SurfaceParams::SurfaceTarget::TextureCubemap:
-            for (std::size_t index = 0; index < params.depth; ++index) {
-                const std::size_t offset{index * copy_size};
-                morton_to_gl_fns[static_cast<std::size_t>(params.pixel_format)](
-                    params.width, params.block_height, params.height, gl_buffer.data() + offset,
-                    copy_size, params.addr + offset);
-            }
-            break;
-        default:
-            LOG_CRITICAL(HW_GPU, "Unimplemented tiled load for target={}",
-                         static_cast<u32>(params.target));
-            UNREACHABLE();
+        if (params.target == SurfaceParams::SurfaceTarget::Texture2D) {
+            // TODO(Blinkhawk): Eliminate this condition once all texture types are implemented.
+            depth = 1U;
+            block_depth = 1U;
         }
 
+        const std::size_t size = copy_size * depth;
+
         morton_to_gl_fns[static_cast<std::size_t>(params.pixel_format)](
-            params.width, params.block_height, params.height, gl_buffer.data(), copy_size,
-            params.addr);
+            params.width, params.block_height, params.height, block_depth, depth, gl_buffer.data(),
+            size, params.addr);
     } else {
         const u8* const texture_src_data_end{texture_src_data + total_size};
         gl_buffer.assign(texture_src_data, texture_src_data_end);

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -74,19 +74,21 @@ struct SurfaceParams {
         RG32UI = 43,
         R32UI = 44,
         ASTC_2D_8X8 = 45,
+        ASTC_2D_8X5 = 46,
+        ASTC_2D_5X4 = 47,
 
         MaxColorFormat,
 
         // Depth formats
-        Z32F = 46,
-        Z16 = 47,
+        Z32F = 48,
+        Z16 = 49,
 
         MaxDepthFormat,
 
         // DepthStencil formats
-        Z24S8 = 48,
-        S8Z24 = 49,
-        Z32FS8 = 50,
+        Z24S8 = 50,
+        S8Z24 = 51,
+        Z32FS8 = 52,
 
         MaxDepthStencilFormat,
 
@@ -220,6 +222,8 @@ struct SurfaceParams {
             1, // RG32UI
             1, // R32UI
             4, // ASTC_2D_8X8
+            4, // ASTC_2D_8X5
+            4, // ASTC_2D_5X4
             1, // Z32F
             1, // Z16
             1, // Z24S8
@@ -282,6 +286,8 @@ struct SurfaceParams {
             64,  // RG32UI
             32,  // R32UI
             16,  // ASTC_2D_8X8
+            32,  // ASTC_2D_8X5
+            32,  // ASTC_2D_5X4
             32,  // Z32F
             16,  // Z16
             32,  // Z24S8
@@ -553,8 +559,12 @@ struct SurfaceParams {
             return PixelFormat::BC6H_SF16;
         case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
             return PixelFormat::ASTC_2D_4X4;
+        case Tegra::Texture::TextureFormat::ASTC_2D_5X4:
+            return PixelFormat::ASTC_2D_5X4;
         case Tegra::Texture::TextureFormat::ASTC_2D_8X8:
             return PixelFormat::ASTC_2D_8X8;
+        case Tegra::Texture::TextureFormat::ASTC_2D_8X5:
+            return PixelFormat::ASTC_2D_8X5;
         case Tegra::Texture::TextureFormat::R16_G16:
             switch (component_type) {
             case Tegra::Texture::ComponentType::FLOAT:

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -1436,7 +1436,6 @@ private:
 
             break;
         }
-
         case OpCode::Type::Shift: {
             std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, true);
             std::string op_b;
@@ -1478,7 +1477,6 @@ private:
             }
             break;
         }
-
         case OpCode::Type::ArithmeticIntegerImmediate: {
             std::string op_a = regs.GetRegisterAsInteger(instr.gpr8);
             std::string op_b = std::to_string(instr.alu.imm20_32.Value());
@@ -2254,8 +2252,6 @@ private:
                 ASSERT_MSG(!instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::MZ),
                            "MZ is not implemented");
 
-                u32 op_c_offset = 0;
-
                 switch (texture_type) {
                 case Tegra::Shader::TextureType::Texture1D: {
                     const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
@@ -2270,7 +2266,6 @@ private:
                         const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
                         const std::string y = regs.GetRegisterAsInteger(instr.gpr20);
                         coord = "ivec2 coords = ivec2(" + x + ", " + y + ");";
-                        op_c_offset = 1;
                     }
                     break;
                 }
@@ -2282,14 +2277,13 @@ private:
                 const std::string sampler =
                     GetSampler(instr.sampler, texture_type, is_array, false);
                 std::string texture = "texelFetch(" + sampler + ", coords, 0)";
+                const std::string op_c = regs.GetRegisterAsInteger(instr.gpr20.Value() + 1);
                 switch (instr.tlds.GetTextureProcessMode()) {
                 case Tegra::Shader::TextureProcessMode::LZ: {
                     texture = "texelFetch(" + sampler + ", coords, 0)";
                     break;
                 }
                 case Tegra::Shader::TextureProcessMode::LL: {
-                    const std::string op_c =
-                        regs.GetRegisterAsInteger(instr.gpr20.Value() + op_c_offset);
                     texture = "texelFetch(" + sampler + ", coords, " + op_c + ')';
                     break;
                 }
@@ -2630,14 +2624,14 @@ private:
                 const std::string pred =
                     GetPredicateCondition(instr.csetp.pred39, instr.csetp.neg_pred39 != 0);
                 const std::string combiner = GetPredicateCombiner(instr.csetp.op);
-                const std::string controlCode = regs.GetControlCode(instr.csetp.cc);
+                const std::string control_code = regs.GetControlCode(instr.csetp.cc);
                 if (instr.csetp.pred3 != static_cast<u64>(Pred::UnusedIndex)) {
                     SetPredicate(instr.csetp.pred3,
-                                 '(' + controlCode + ") " + combiner + " (" + pred + ')');
+                                 '(' + control_code + ") " + combiner + " (" + pred + ')');
                 }
                 if (instr.csetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) {
                     SetPredicate(instr.csetp.pred0,
-                                 "!(" + controlCode + ") " + combiner + " (" + pred + ')');
+                                 "!(" + control_code + ") " + combiner + " (" + pred + ')');
                 }
                 break;
             }

src/video_core/textures/decoders.cpp

@@ -40,72 +40,146 @@ struct alignas(64) SwizzleTable {
 constexpr auto legacy_swizzle_table = SwizzleTable<8, 64, 1>();
 constexpr auto fast_swizzle_table = SwizzleTable<8, 4, 16>();
 
-static void LegacySwizzleData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                              u8* swizzled_data, u8* unswizzled_data, bool unswizzle,
-                              u32 block_height) {
+/**
+ * This function manages ALL the GOBs(Group of Bytes) Inside a single block.
+ * Instead of going gob by gob, we map the coordinates inside a block and manage from
+ * those. Block_Width is assumed to be 1.
+ */
+void PreciseProcessBlock(u8* swizzled_data, u8* unswizzled_data, const bool unswizzle,
+                         const u32 x_start, const u32 y_start, const u32 z_start, const u32 x_end,
+                         const u32 y_end, const u32 z_end, const u32 tile_offset,
+                         const u32 xy_block_size, const u32 layer_z, const u32 stride_x,
+                         const u32 bytes_per_pixel, const u32 out_bytes_per_pixel) {
     std::array<u8*, 2> data_ptrs;
-    const std::size_t stride = width * bytes_per_pixel;
-    const std::size_t gobs_in_x = 64;
-    const std::size_t gobs_in_y = 8;
-    const std::size_t gobs_size = gobs_in_x * gobs_in_y;
-    const std::size_t image_width_in_gobs{(stride + gobs_in_x - 1) / gobs_in_x};
-    for (std::size_t y = 0; y < height; ++y) {
-        const std::size_t gob_y_address =
-            (y / (gobs_in_y * block_height)) * gobs_size * block_height * image_width_in_gobs +
-            (y % (gobs_in_y * block_height) / gobs_in_y) * gobs_size;
-        const auto& table = legacy_swizzle_table[y % gobs_in_y];
-        for (std::size_t x = 0; x < width; ++x) {
-            const std::size_t gob_address =
-                gob_y_address + (x * bytes_per_pixel / gobs_in_x) * gobs_size * block_height;
-            const std::size_t x2 = x * bytes_per_pixel;
-            const std::size_t swizzle_offset = gob_address + table[x2 % gobs_in_x];
-            const std::size_t pixel_index = (x + y * width) * out_bytes_per_pixel;
+    u32 z_address = tile_offset;
+    const u32 gob_size_x = 64;
+    const u32 gob_size_y = 8;
+    const u32 gob_size_z = 1;
+    const u32 gob_size = gob_size_x * gob_size_y * gob_size_z;
+    for (u32 z = z_start; z < z_end; z++) {
+        u32 y_address = z_address;
+        u32 pixel_base = layer_z * z + y_start * stride_x;
+        for (u32 y = y_start; y < y_end; y++) {
+            const auto& table = legacy_swizzle_table[y % gob_size_y];
+            for (u32 x = x_start; x < x_end; x++) {
+                const u32 swizzle_offset{y_address + table[x * bytes_per_pixel % gob_size_x]};
+                const u32 pixel_index{x * out_bytes_per_pixel + pixel_base};
+                data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
+                data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
+                std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
+            }
+            pixel_base += stride_x;
+            if ((y + 1) % gob_size_y == 0)
+                y_address += gob_size;
+        }
+        z_address += xy_block_size;
+    }
+}
 
-            data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
-            data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
+/**
+ * This function manages ALL the GOBs(Group of Bytes) Inside a single block.
+ * Instead of going gob by gob, we map the coordinates inside a block and manage from
+ * those. Block_Width is assumed to be 1.
+ */
+void FastProcessBlock(u8* swizzled_data, u8* unswizzled_data, const bool unswizzle,
+                      const u32 x_start, const u32 y_start, const u32 z_start, const u32 x_end,
+                      const u32 y_end, const u32 z_end, const u32 tile_offset,
+                      const u32 xy_block_size, const u32 layer_z, const u32 stride_x,
+                      const u32 bytes_per_pixel, const u32 out_bytes_per_pixel) {
+    std::array<u8*, 2> data_ptrs;
+    u32 z_address = tile_offset;
+    const u32 x_startb = x_start * bytes_per_pixel;
+    const u32 x_endb = x_end * bytes_per_pixel;
+    const u32 copy_size = 16;
+    const u32 gob_size_x = 64;
+    const u32 gob_size_y = 8;
+    const u32 gob_size_z = 1;
+    const u32 gob_size = gob_size_x * gob_size_y * gob_size_z;
+    for (u32 z = z_start; z < z_end; z++) {
+        u32 y_address = z_address;
+        u32 pixel_base = layer_z * z + y_start * stride_x;
+        for (u32 y = y_start; y < y_end; y++) {
+            const auto& table = fast_swizzle_table[y % gob_size_y];
+            for (u32 xb = x_startb; xb < x_endb; xb += copy_size) {
+                const u32 swizzle_offset{y_address + table[(xb / copy_size) % 4]};
+                const u32 out_x = xb * out_bytes_per_pixel / bytes_per_pixel;
+                const u32 pixel_index{out_x + pixel_base};
+                data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
+                data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
+                std::memcpy(data_ptrs[0], data_ptrs[1], copy_size);
+            }
+            pixel_base += stride_x;
+            if ((y + 1) % gob_size_y == 0)
+                y_address += gob_size;
+        }
+        z_address += xy_block_size;
+    }
+}
 
-            std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
+/**
+ * This function unswizzles or swizzles a texture by mapping Linear to BlockLinear Textue.
+ * The body of this function takes care of splitting the swizzled texture into blocks,
+ * and managing the extents of it. Once all the parameters of a single block are obtained,
+ * the function calls 'ProcessBlock' to process that particular Block.
+ *
+ * Documentation for the memory layout and decoding can be found at:
+ *  https://envytools.readthedocs.io/en/latest/hw/memory/g80-surface.html#blocklinear-surfaces
+ */
+template <bool fast>
+void SwizzledData(u8* swizzled_data, u8* unswizzled_data, const bool unswizzle, const u32 width,
+                  const u32 height, const u32 depth, const u32 bytes_per_pixel,
+                  const u32 out_bytes_per_pixel, const u32 block_height, const u32 block_depth) {
+    auto div_ceil = [](const u32 x, const u32 y) { return ((x + y - 1) / y); };
+    const u32 stride_x = width * out_bytes_per_pixel;
+    const u32 layer_z = height * stride_x;
+    const u32 gob_x_bytes = 64;
+    const u32 gob_elements_x = gob_x_bytes / bytes_per_pixel;
+    const u32 gob_elements_y = 8;
+    const u32 gob_elements_z = 1;
+    const u32 block_x_elements = gob_elements_x;
+    const u32 block_y_elements = gob_elements_y * block_height;
+    const u32 block_z_elements = gob_elements_z * block_depth;
+    const u32 blocks_on_x = div_ceil(width, block_x_elements);
+    const u32 blocks_on_y = div_ceil(height, block_y_elements);
+    const u32 blocks_on_z = div_ceil(depth, block_z_elements);
+    const u32 blocks = blocks_on_x * blocks_on_y * blocks_on_z;
+    const u32 gob_size = gob_x_bytes * gob_elements_y * gob_elements_z;
+    const u32 xy_block_size = gob_size * block_height;
+    const u32 block_size = xy_block_size * block_depth;
+    u32 tile_offset = 0;
+    for (u32 zb = 0; zb < blocks_on_z; zb++) {
+        const u32 z_start = zb * block_z_elements;
+        const u32 z_end = std::min(depth, z_start + block_z_elements);
+        for (u32 yb = 0; yb < blocks_on_y; yb++) {
+            const u32 y_start = yb * block_y_elements;
+            const u32 y_end = std::min(height, y_start + block_y_elements);
+            for (u32 xb = 0; xb < blocks_on_x; xb++) {
+                const u32 x_start = xb * block_x_elements;
+                const u32 x_end = std::min(width, x_start + block_x_elements);
+                if (fast) {
+                    FastProcessBlock(swizzled_data, unswizzled_data, unswizzle, x_start, y_start,
+                                     z_start, x_end, y_end, z_end, tile_offset, xy_block_size,
+                                     layer_z, stride_x, bytes_per_pixel, out_bytes_per_pixel);
+                } else {
+                    PreciseProcessBlock(swizzled_data, unswizzled_data, unswizzle, x_start, y_start,
+                                        z_start, x_end, y_end, z_end, tile_offset, xy_block_size,
+                                        layer_z, stride_x, bytes_per_pixel, out_bytes_per_pixel);
+                }
+                tile_offset += block_size;
+            }
         }
     }
 }
 
-static void FastSwizzleData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                            u8* swizzled_data, u8* unswizzled_data, bool unswizzle,
-                            u32 block_height) {
-    std::array<u8*, 2> data_ptrs;
-    const std::size_t stride{width * bytes_per_pixel};
-    const std::size_t gobs_in_x = 64;
-    const std::size_t gobs_in_y = 8;
-    const std::size_t gobs_size = gobs_in_x * gobs_in_y;
-    const std::size_t image_width_in_gobs{(stride + gobs_in_x - 1) / gobs_in_x};
-    const std::size_t copy_size{16};
-    for (std::size_t y = 0; y < height; ++y) {
-        const std::size_t initial_gob =
-            (y / (gobs_in_y * block_height)) * gobs_size * block_height * image_width_in_gobs +
-            (y % (gobs_in_y * block_height) / gobs_in_y) * gobs_size;
-        const std::size_t pixel_base{y * width * out_bytes_per_pixel};
-        const auto& table = fast_swizzle_table[y % gobs_in_y];
-        for (std::size_t xb = 0; xb < stride; xb += copy_size) {
-            const std::size_t gob_address{initial_gob +
-                                          (xb / gobs_in_x) * gobs_size * block_height};
-            const std::size_t swizzle_offset{gob_address + table[(xb / 16) % 4]};
-            const std::size_t out_x = xb * out_bytes_per_pixel / bytes_per_pixel;
-            const std::size_t pixel_index{out_x + pixel_base};
-            data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
-            data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
-            std::memcpy(data_ptrs[0], data_ptrs[1], copy_size);
-        }
-    }
-}
-
-void CopySwizzledData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                      u8* swizzled_data, u8* unswizzled_data, bool unswizzle, u32 block_height) {
+void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel,
+                      u32 out_bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data,
+                      bool unswizzle, u32 block_height, u32 block_depth) {
     if (bytes_per_pixel % 3 != 0 && (width * bytes_per_pixel) % 16 == 0) {
-        FastSwizzleData(width, height, bytes_per_pixel, out_bytes_per_pixel, swizzled_data,
-                        unswizzled_data, unswizzle, block_height);
+        SwizzledData<true>(swizzled_data, unswizzled_data, unswizzle, width, height, depth,
+                           bytes_per_pixel, out_bytes_per_pixel, block_height, block_depth);
     } else {
-        LegacySwizzleData(width, height, bytes_per_pixel, out_bytes_per_pixel, swizzled_data,
-                          unswizzled_data, unswizzle, block_height);
+        SwizzledData<false>(swizzled_data, unswizzled_data, unswizzle, width, height, depth,
+                            bytes_per_pixel, out_bytes_per_pixel, block_height, block_depth);
     }
 }
 
@@ -126,7 +200,9 @@ u32 BytesPerPixel(TextureFormat format) {
     case TextureFormat::R32_G32_B32:
         return 12;
     case TextureFormat::ASTC_2D_4X4:
+    case TextureFormat::ASTC_2D_5X4:
     case TextureFormat::ASTC_2D_8X8:
+    case TextureFormat::ASTC_2D_8X5:
     case TextureFormat::A8R8G8B8:
     case TextureFormat::A2B10G10R10:
     case TextureFormat::BF10GF11RF11:
@@ -153,10 +229,11 @@ u32 BytesPerPixel(TextureFormat format) {
 }
 
 std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
-                                 u32 height, u32 block_height) {
-    std::vector<u8> unswizzled_data(width * height * bytes_per_pixel);
-    CopySwizzledData(width / tile_size, height / tile_size, bytes_per_pixel, bytes_per_pixel,
-                     Memory::GetPointer(address), unswizzled_data.data(), true, block_height);
+                                 u32 height, u32 depth, u32 block_height, u32 block_depth) {
+    std::vector<u8> unswizzled_data(width * height * depth * bytes_per_pixel);
+    CopySwizzledData(width / tile_size, height / tile_size, depth, bytes_per_pixel, bytes_per_pixel,
+                     Memory::GetPointer(address), unswizzled_data.data(), true, block_height,
+                     block_depth);
     return unswizzled_data;
 }
 

src/video_core/textures/decoders.h

@@ -14,17 +14,14 @@ namespace Tegra::Texture {
  * Unswizzles a swizzled texture without changing its format.
  */
 std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
-                                 u32 height, u32 block_height = TICEntry::DefaultBlockHeight);
-
-/**
- * Unswizzles a swizzled depth texture without changing its format.
- */
-std::vector<u8> UnswizzleDepthTexture(VAddr address, DepthFormat format, u32 width, u32 height,
-                                      u32 block_height = TICEntry::DefaultBlockHeight);
+                                 u32 height, u32 depth,
+                                 u32 block_height = TICEntry::DefaultBlockHeight,
+                                 u32 block_depth = TICEntry::DefaultBlockHeight);
 
 /// Copies texture data from a buffer and performs swizzling/unswizzling as necessary.
-void CopySwizzledData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                      u8* swizzled_data, u8* unswizzled_data, bool unswizzle, u32 block_height);
+void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel,
+                      u32 out_bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data,
+                      bool unswizzle, u32 block_height, u32 block_depth);
 
 /**
  * Decodes an unswizzled texture into a A8R8G8B8 texture.

src/video_core/textures/texture.h

@@ -141,6 +141,7 @@ static_assert(sizeof(TextureHandle) == 4, "TextureHandle has wrong size");
 
 struct TICEntry {
     static constexpr u32 DefaultBlockHeight = 16;
+    static constexpr u32 DefaultBlockDepth = 1;
 
     union {
         u32 raw;

src/yuzu/debugger/graphics/graphics_surface.cpp

@@ -386,8 +386,9 @@ void GraphicsSurfaceWidget::OnUpdate() {
 
     // TODO(bunnei): Will not work with BCn formats that swizzle 4x4 tiles.
     // Needs to be fixed if we plan to use this feature more, otherwise we may remove it.
-    auto unswizzled_data = Tegra::Texture::UnswizzleTexture(
-        *address, 1, Tegra::Texture::BytesPerPixel(surface_format), surface_width, surface_height);
+    auto unswizzled_data =
+        Tegra::Texture::UnswizzleTexture(*address, 1, Tegra::Texture::BytesPerPixel(surface_format),
+                                         surface_width, surface_height, 1U);
 
     auto texture_data = Tegra::Texture::DecodeTexture(unswizzled_data, surface_format,
                                                       surface_width, surface_height);

src/yuzu/main.cpp

@@ -176,7 +176,7 @@ GMainWindow::GMainWindow()
     OnReinitializeKeys(ReinitializeKeyBehavior::NoWarning);
 
     // Necessary to load titles from nand in gamelist.
-    Service::FileSystem::CreateFactories(vfs);
+    Service::FileSystem::CreateFactories(*vfs);
     game_list->LoadCompatibilityList();
     game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan);
 
@@ -908,22 +908,20 @@ void GMainWindow::OnGameListNavigateToGamedbEntry(u64 program_id,
 }
 
 void GMainWindow::OnMenuLoadFile() {
-    QString extensions;
-    for (const auto& piece : game_list->supported_file_extensions)
-        extensions += "*." + piece + " ";
+    const QString extensions =
+        QString("*.").append(GameList::supported_file_extensions.join(" *.")).append(" main");
+    const QString file_filter = tr("Switch Executable (%1);;All Files (*.*)",
+                                   "%1 is an identifier for the Switch executable file extensions.")
+                                    .arg(extensions);
+    const QString filename = QFileDialog::getOpenFileName(
+        this, tr("Load File"), UISettings::values.roms_path, file_filter);
 
-    extensions += "main ";
-
-    QString file_filter = tr("Switch Executable") + " (" + extensions + ")";
-    file_filter += ";;" + tr("All Files (*.*)");
-
-    QString filename = QFileDialog::getOpenFileName(this, tr("Load File"),
-                                                    UISettings::values.roms_path, file_filter);
-    if (!filename.isEmpty()) {
-        UISettings::values.roms_path = QFileInfo(filename).path();
-
-        BootGame(filename);
+    if (filename.isEmpty()) {
+        return;
     }
+
+    UISettings::values.roms_path = QFileInfo(filename).path();
+    BootGame(filename);
 }
 
 void GMainWindow::OnMenuLoadFolder() {
@@ -1139,7 +1137,7 @@ void GMainWindow::OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget target)
         FileUtil::GetUserPath(target == EmulatedDirectoryTarget::SDMC ? FileUtil::UserPath::SDMCDir
                                                                       : FileUtil::UserPath::NANDDir,
                               dir_path.toStdString());
-        Service::FileSystem::CreateFactories(vfs);
+        Service::FileSystem::CreateFactories(*vfs);
         game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan);
     }
 }
@@ -1410,7 +1408,7 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) {
 
         const auto function = [this, &keys, &pdm] {
             keys.PopulateFromPartitionData(pdm);
-            Service::FileSystem::CreateFactories(vfs);
+            Service::FileSystem::CreateFactories(*vfs);
             keys.DeriveETicket(pdm);
         };
 
@@ -1450,7 +1448,7 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) {
         prog.close();
     }
 
-    Service::FileSystem::CreateFactories(vfs);
+    Service::FileSystem::CreateFactories(*vfs);
 
     if (behavior == ReinitializeKeyBehavior::Warning) {
         game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan);

src/yuzu_cmd/yuzu.cpp

@@ -175,7 +175,7 @@ int main(int argc, char** argv) {
 
     Core::System& system{Core::System::GetInstance()};
     system.SetFilesystem(std::make_shared<FileSys::RealVfsFilesystem>());
-    Service::FileSystem::CreateFactories(system.GetFilesystem());
+    Service::FileSystem::CreateFactories(*system.GetFilesystem());
 
     SCOPE_EXIT({ system.Shutdown(); });