diff options
65 files changed, 657 insertions, 431 deletions
diff --git a/src/audio_core/time_stretch.cpp b/src/audio_core/time_stretch.cpp index cee8b12dd..2fe0b3aef 100644 --- a/src/audio_core/time_stretch.cpp +++ b/src/audio_core/time_stretch.cpp @@ -32,10 +32,10 @@ std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out, // We were given actual_samples number of samples, and num_samples were requested from us. double current_ratio = static_cast<double>(num_in) / static_cast<double>(num_out); - const double max_latency = 1.0; // seconds + const double max_latency = 0.25; // seconds const double max_backlog = m_sample_rate * max_latency; const double backlog_fullness = m_sound_touch.numSamples() / max_backlog; - if (backlog_fullness > 5.0) { + if (backlog_fullness > 4.0) { // Too many samples in backlog: Don't push anymore on num_in = 0; } @@ -49,7 +49,7 @@ std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out, // This low-pass filter smoothes out variance in the calculated stretch ratio. // The time-scale determines how responsive this filter is. - constexpr double lpf_time_scale = 2.0; // seconds + constexpr double lpf_time_scale = 0.712; // seconds const double lpf_gain = 1.0 - std::exp(-time_delta / lpf_time_scale); m_stretch_ratio += lpf_gain * (current_ratio - m_stretch_ratio); diff --git a/src/core/core.cpp b/src/core/core.cpp index 7cb86ed92..6d5b5a2d0 100644 --- a/src/core/core.cpp +++ b/src/core/core.cpp @@ -185,7 +185,7 @@ struct System::Impl { LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath); return ResultStatus::ErrorGetLoader; } - std::pair<boost::optional<u32>, Loader::ResultStatus> system_mode = + std::pair<std::optional<u32>, Loader::ResultStatus> system_mode = app_loader->LoadKernelSystemMode(); if (system_mode.second != Loader::ResultStatus::Success) { @@ -312,6 +312,10 @@ Cpu& System::CurrentCpuCore() { return impl->CurrentCpuCore(); } +const Cpu& System::CurrentCpuCore() const { + return impl->CurrentCpuCore(); +} + System::ResultStatus System::RunLoop(bool tight_loop) { return impl->RunLoop(tight_loop); } @@ -342,7 +346,11 @@ PerfStatsResults System::GetAndResetPerfStats() { return impl->GetAndResetPerfStats(); } -Core::TelemetrySession& System::TelemetrySession() const { +TelemetrySession& System::TelemetrySession() { + return *impl->telemetry_session; +} + +const TelemetrySession& System::TelemetrySession() const { return *impl->telemetry_session; } @@ -350,7 +358,11 @@ ARM_Interface& System::CurrentArmInterface() { return CurrentCpuCore().ArmInterface(); } -std::size_t System::CurrentCoreIndex() { +const ARM_Interface& System::CurrentArmInterface() const { + return CurrentCpuCore().ArmInterface(); +} + +std::size_t System::CurrentCoreIndex() const { return CurrentCpuCore().CoreIndex(); } @@ -358,6 +370,10 @@ Kernel::Scheduler& System::CurrentScheduler() { return CurrentCpuCore().Scheduler(); } +const Kernel::Scheduler& System::CurrentScheduler() const { + return CurrentCpuCore().Scheduler(); +} + Kernel::Scheduler& System::Scheduler(std::size_t core_index) { return CpuCore(core_index).Scheduler(); } @@ -378,6 +394,10 @@ ARM_Interface& System::ArmInterface(std::size_t core_index) { return CpuCore(core_index).ArmInterface(); } +const ARM_Interface& System::ArmInterface(std::size_t core_index) const { + return CpuCore(core_index).ArmInterface(); +} + Cpu& System::CpuCore(std::size_t core_index) { ASSERT(core_index < NUM_CPU_CORES); return *impl->cpu_cores[core_index]; @@ -392,6 +412,10 @@ ExclusiveMonitor& System::Monitor() { return *impl->cpu_exclusive_monitor; } +const ExclusiveMonitor& System::Monitor() const { + return *impl->cpu_exclusive_monitor; +} + Tegra::GPU& System::GPU() { return *impl->gpu_core; } diff --git a/src/core/core.h b/src/core/core.h index 173be45f8..cfacceb81 100644 --- a/src/core/core.h +++ b/src/core/core.h @@ -129,11 +129,11 @@ public: */ bool IsPoweredOn() const; - /** - * Returns a reference to the telemetry session for this emulation session. - * @returns Reference to the telemetry session. - */ - Core::TelemetrySession& TelemetrySession() const; + /// Gets a reference to the telemetry session for this emulation session. + Core::TelemetrySession& TelemetrySession(); + + /// Gets a reference to the telemetry session for this emulation session. + const Core::TelemetrySession& TelemetrySession() const; /// Prepare the core emulation for a reschedule void PrepareReschedule(); @@ -144,24 +144,36 @@ public: /// Gets an ARM interface to the CPU core that is currently running ARM_Interface& CurrentArmInterface(); + /// Gets an ARM interface to the CPU core that is currently running + const ARM_Interface& CurrentArmInterface() const; + /// Gets the index of the currently running CPU core - std::size_t CurrentCoreIndex(); + std::size_t CurrentCoreIndex() const; /// Gets the scheduler for the CPU core that is currently running Kernel::Scheduler& CurrentScheduler(); - /// Gets an ARM interface to the CPU core with the specified index + /// Gets the scheduler for the CPU core that is currently running + const Kernel::Scheduler& CurrentScheduler() const; + + /// Gets a reference to an ARM interface for the CPU core with the specified index ARM_Interface& ArmInterface(std::size_t core_index); + /// Gets a const reference to an ARM interface from the CPU core with the specified index + const ARM_Interface& ArmInterface(std::size_t core_index) const; + /// Gets a CPU interface to the CPU core with the specified index Cpu& CpuCore(std::size_t core_index); /// Gets a CPU interface to the CPU core with the specified index const Cpu& CpuCore(std::size_t core_index) const; - /// Gets the exclusive monitor + /// Gets a reference to the exclusive monitor ExclusiveMonitor& Monitor(); + /// Gets a constant reference to the exclusive monitor + const ExclusiveMonitor& Monitor() const; + /// Gets a mutable reference to the GPU interface Tegra::GPU& GPU(); @@ -230,6 +242,9 @@ private: /// Returns the currently running CPU core Cpu& CurrentCpuCore(); + /// Returns the currently running CPU core + const Cpu& CurrentCpuCore() const; + /** * Initialize the emulated system. * @param emu_window Reference to the host-system window used for video output and keyboard diff --git a/src/core/crypto/key_manager.cpp b/src/core/crypto/key_manager.cpp index 89ae79eb3..904afa039 100644 --- a/src/core/crypto/key_manager.cpp +++ b/src/core/crypto/key_manager.cpp @@ -141,28 +141,28 @@ Key128 DeriveKeyblobMACKey(const Key128& keyblob_key, const Key128& mac_source) return mac_key; } -boost::optional<Key128> DeriveSDSeed() { +std::optional<Key128> DeriveSDSeed() { const FileUtil::IOFile save_43(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + "/system/save/8000000000000043", "rb+"); if (!save_43.IsOpen()) - return boost::none; + return {}; const FileUtil::IOFile sd_private( FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir) + "/Nintendo/Contents/private", "rb+"); if (!sd_private.IsOpen()) - return boost::none; + return {}; std::array<u8, 0x10> private_seed{}; if (sd_private.ReadBytes(private_seed.data(), private_seed.size()) != private_seed.size()) { - return boost::none; + return {}; } std::array<u8, 0x10> buffer{}; std::size_t offset = 0; for (; offset + 0x10 < save_43.GetSize(); ++offset) { if (!save_43.Seek(offset, SEEK_SET)) { - return boost::none; + return {}; } save_43.ReadBytes(buffer.data(), buffer.size()); @@ -172,12 +172,12 @@ boost::optional<Key128> DeriveSDSeed() { } if (!save_43.Seek(offset + 0x10, SEEK_SET)) { - return boost::none; + return {}; } Key128 seed{}; if (save_43.ReadBytes(seed.data(), seed.size()) != seed.size()) { - return boost::none; + return {}; } return seed; } @@ -291,26 +291,26 @@ static std::array<u8, target_size> MGF1(const std::array<u8, in_size>& seed) { } template <size_t size> -static boost::optional<u64> FindTicketOffset(const std::array<u8, size>& data) { +static std::optional<u64> FindTicketOffset(const std::array<u8, size>& data) { u64 offset = 0; for (size_t i = 0x20; i < data.size() - 0x10; ++i) { if (data[i] == 0x1) { offset = i + 1; break; } else if (data[i] != 0x0) { - return boost::none; + return {}; } } return offset; } -boost::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket, - const RSAKeyPair<2048>& key) { +std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket, + const RSAKeyPair<2048>& key) { u32 cert_authority; std::memcpy(&cert_authority, ticket.data() + 0x140, sizeof(cert_authority)); if (cert_authority == 0) - return boost::none; + return {}; if (cert_authority != Common::MakeMagic('R', 'o', 'o', 't')) { LOG_INFO(Crypto, "Attempting to parse ticket with non-standard certificate authority {:08X}.", @@ -321,7 +321,7 @@ boost::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket, std::memcpy(rights_id.data(), ticket.data() + 0x2A0, sizeof(Key128)); if (rights_id == Key128{}) - return boost::none; + return {}; Key128 key_temp{}; @@ -356,17 +356,17 @@ boost::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket, std::memcpy(m_2.data(), rsa_step.data() + 0x21, m_2.size()); if (m_0 != 0) - return boost::none; + return {}; m_1 = m_1 ^ MGF1<0x20>(m_2); m_2 = m_2 ^ MGF1<0xDF>(m_1); const auto offset = FindTicketOffset(m_2); - if (offset == boost::none) - return boost::none; - ASSERT(offset.get() > 0); + if (!offset) + return {}; + ASSERT(*offset > 0); - std::memcpy(key_temp.data(), m_2.data() + offset.get(), key_temp.size()); + std::memcpy(key_temp.data(), m_2.data() + *offset, key_temp.size()); return std::make_pair(rights_id, key_temp); } @@ -661,8 +661,8 @@ void KeyManager::DeriveSDSeedLazy() { return; const auto res = DeriveSDSeed(); - if (res != boost::none) - SetKey(S128KeyType::SDSeed, res.get()); + if (res) + SetKey(S128KeyType::SDSeed, *res); } static Key128 CalculateCMAC(const u8* source, size_t size, const Key128& key) { @@ -889,9 +889,9 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) { for (const auto& raw : res) { const auto pair = ParseTicket(raw, rsa_key); - if (pair == boost::none) + if (!pair) continue; - const auto& [rid, key] = pair.value(); + const auto& [rid, key] = *pair; u128 rights_id; std::memcpy(rights_id.data(), rid.data(), rid.size()); SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]); diff --git a/src/core/crypto/key_manager.h b/src/core/crypto/key_manager.h index cccb3c0ae..22f268c65 100644 --- a/src/core/crypto/key_manager.h +++ b/src/core/crypto/key_manager.h @@ -6,9 +6,10 @@ #include <array> #include <map> +#include <optional> #include <string> + #include <boost/container/flat_map.hpp> -#include <boost/optional.hpp> #include <fmt/format.h> #include "common/common_types.h" #include "core/crypto/partition_data_manager.h" @@ -191,14 +192,14 @@ Key128 DeriveMasterKey(const std::array<u8, 0x90>& keyblob, const Key128& master std::array<u8, 0x90> DecryptKeyblob(const std::array<u8, 0xB0>& encrypted_keyblob, const Key128& key); -boost::optional<Key128> DeriveSDSeed(); +std::optional<Key128> DeriveSDSeed(); Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& keys); std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save); // Returns a pair of {rights_id, titlekey}. Fails if the ticket has no certificate authority (offset // 0x140-0x144 is zero) -boost::optional<std::pair<Key128, Key128>> ParseTicket( - const TicketRaw& ticket, const RSAKeyPair<2048>& eticket_extended_key); +std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket, + const RSAKeyPair<2048>& eticket_extended_key); } // namespace Core::Crypto diff --git a/src/core/file_sys/content_archive.cpp b/src/core/file_sys/content_archive.cpp index 77e04704e..b46fe893c 100644 --- a/src/core/file_sys/content_archive.cpp +++ b/src/core/file_sys/content_archive.cpp @@ -4,10 +4,9 @@ #include <algorithm> #include <cstring> +#include <optional> #include <utility> -#include <boost/optional.hpp> - #include "common/logging/log.h" #include "core/crypto/aes_util.h" #include "core/crypto/ctr_encryption_layer.h" @@ -306,18 +305,18 @@ bool NCA::ReadRomFSSection(const NCASectionHeader& section, const NCASectionTabl subsection_buckets.back().entries.push_back({section.bktr.relocation.offset, {0}, ctr_low}); subsection_buckets.back().entries.push_back({size, {0}, 0}); - boost::optional<Core::Crypto::Key128> key = boost::none; + std::optional<Core::Crypto::Key128> key = {}; if (encrypted) { if (has_rights_id) { status = Loader::ResultStatus::Success; key = GetTitlekey(); - if (key == boost::none) { + if (!key) { status = Loader::ResultStatus::ErrorMissingTitlekey; return false; } } else { key = GetKeyAreaKey(NCASectionCryptoType::BKTR); - if (key == boost::none) { + if (!key) { status = Loader::ResultStatus::ErrorMissingKeyAreaKey; return false; } @@ -332,7 +331,7 @@ bool NCA::ReadRomFSSection(const NCASectionHeader& section, const NCASectionTabl auto bktr = std::make_shared<BKTR>( bktr_base_romfs, std::make_shared<OffsetVfsFile>(file, romfs_size, base_offset), relocation_block, relocation_buckets, subsection_block, subsection_buckets, encrypted, - encrypted ? key.get() : Core::Crypto::Key128{}, base_offset, bktr_base_ivfc_offset, + encrypted ? *key : Core::Crypto::Key128{}, base_offset, bktr_base_ivfc_offset, section.raw.section_ctr); // BKTR applies to entire IVFC, so make an offset version to level 6 @@ -388,11 +387,11 @@ u8 NCA::GetCryptoRevision() const { return master_key_id; } -boost::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const { +std::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const { const auto master_key_id = GetCryptoRevision(); if (!keys.HasKey(Core::Crypto::S128KeyType::KeyArea, master_key_id, header.key_index)) - return boost::none; + return {}; std::vector<u8> key_area(header.key_area.begin(), header.key_area.end()); Core::Crypto::AESCipher<Core::Crypto::Key128> cipher( @@ -416,25 +415,25 @@ boost::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType ty return out; } -boost::optional<Core::Crypto::Key128> NCA::GetTitlekey() { +std::optional<Core::Crypto::Key128> NCA::GetTitlekey() { const auto master_key_id = GetCryptoRevision(); u128 rights_id{}; memcpy(rights_id.data(), header.rights_id.data(), 16); if (rights_id == u128{}) { status = Loader::ResultStatus::ErrorInvalidRightsID; - return boost::none; + return {}; } auto titlekey = keys.GetKey(Core::Crypto::S128KeyType::Titlekey, rights_id[1], rights_id[0]); if (titlekey == Core::Crypto::Key128{}) { status = Loader::ResultStatus::ErrorMissingTitlekey; - return boost::none; + return {}; } if (!keys.HasKey(Core::Crypto::S128KeyType::Titlekek, master_key_id)) { status = Loader::ResultStatus::ErrorMissingTitlekek; - return boost::none; + return {}; } Core::Crypto::AESCipher<Core::Crypto::Key128> cipher( @@ -458,25 +457,25 @@ VirtualFile NCA::Decrypt(const NCASectionHeader& s_header, VirtualFile in, u64 s case NCASectionCryptoType::BKTR: LOG_DEBUG(Crypto, "called with mode=CTR, starting_offset={:016X}", starting_offset); { - boost::optional<Core::Crypto::Key128> key = boost::none; + std::optional<Core::Crypto::Key128> key = {}; if (has_rights_id) { status = Loader::ResultStatus::Success; key = GetTitlekey(); - if (key == boost::none) { + if (!key) { if (status == Loader::ResultStatus::Success) status = Loader::ResultStatus::ErrorMissingTitlekey; return nullptr; } } else { key = GetKeyAreaKey(NCASectionCryptoType::CTR); - if (key == boost::none) { + if (!key) { status = Loader::ResultStatus::ErrorMissingKeyAreaKey; return nullptr; } } - auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>( - std::move(in), key.value(), starting_offset); + auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(std::move(in), *key, + starting_offset); std::vector<u8> iv(16); for (u8 i = 0; i < 8; ++i) iv[i] = s_header.raw.section_ctr[0x8 - i - 1]; diff --git a/src/core/file_sys/content_archive.h b/src/core/file_sys/content_archive.h index 211946686..4bba55607 100644 --- a/src/core/file_sys/content_archive.h +++ b/src/core/file_sys/content_archive.h @@ -6,9 +6,10 @@ #include <array> #include <memory> +#include <optional> #include <string> #include <vector> -#include <boost/optional.hpp> + #include "common/common_funcs.h" #include "common/common_types.h" #include "common/swap.h" @@ -111,8 +112,8 @@ private: bool ReadPFS0Section(const NCASectionHeader& section, const NCASectionTableEntry& entry); u8 GetCryptoRevision() const; - boost::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const; - boost::optional<Core::Crypto::Key128> GetTitlekey(); + std::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const; + std::optional<Core::Crypto::Key128> GetTitlekey(); VirtualFile Decrypt(const NCASectionHeader& header, VirtualFile in, u64 starting_offset); std::vector<VirtualDir> dirs; diff --git a/src/core/file_sys/ips_layer.cpp b/src/core/file_sys/ips_layer.cpp index 999939d5a..485c4913a 100644 --- a/src/core/file_sys/ips_layer.cpp +++ b/src/core/file_sys/ips_layer.cpp @@ -103,12 +103,12 @@ VirtualFile PatchIPS(const VirtualFile& in, const VirtualFile& ips) { offset += sizeof(u16); const auto data = ips->ReadByte(offset++); - if (data == boost::none) + if (!data) return nullptr; if (real_offset + rle_size > in_data.size()) rle_size = static_cast<u16>(in_data.size() - real_offset); - std::memset(in_data.data() + real_offset, data.get(), rle_size); + std::memset(in_data.data() + real_offset, *data, rle_size); } else { // Standard Patch auto read = data_size; if (real_offset + read > in_data.size()) diff --git a/src/core/file_sys/patch_manager.cpp b/src/core/file_sys/patch_manager.cpp index cb457b987..0c1156989 100644 --- a/src/core/file_sys/patch_manager.cpp +++ b/src/core/file_sys/patch_manager.cpp @@ -65,7 +65,7 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const { if (update != nullptr && update->GetExeFS() != nullptr && update->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { LOG_INFO(Loader, " ExeFS: Update ({}) applied successfully", - FormatTitleVersion(installed->GetEntryVersion(update_tid).get_value_or(0))); + FormatTitleVersion(installed->GetEntryVersion(update_tid).value_or(0))); exefs = update->GetExeFS(); } @@ -236,7 +236,7 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content if (new_nca->GetStatus() == Loader::ResultStatus::Success && new_nca->GetRomFS() != nullptr) { LOG_INFO(Loader, " RomFS: Update ({}) applied successfully", - FormatTitleVersion(installed->GetEntryVersion(update_tid).get_value_or(0))); + FormatTitleVersion(installed->GetEntryVersion(update_tid).value_or(0))); romfs = new_nca->GetRomFS(); } } else if (update_raw != nullptr) { @@ -280,12 +280,11 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam } else { if (installed->HasEntry(update_tid, ContentRecordType::Program)) { const auto meta_ver = installed->GetEntryVersion(update_tid); - if (meta_ver == boost::none || meta_ver.get() == 0) { + if (meta_ver.value_or(0) == 0) { out.insert_or_assign("Update", ""); } else { out.insert_or_assign( - "Update", - FormatTitleVersion(meta_ver.get(), TitleVersionFormat::ThreeElements)); + "Update", FormatTitleVersion(*meta_ver, TitleVersionFormat::ThreeElements)); } } else if (update_raw != nullptr) { out.insert_or_assign("Update", "PACKED"); diff --git a/src/core/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp index 29b100414..96302a241 100644 --- a/src/core/file_sys/registered_cache.cpp +++ b/src/core/file_sys/registered_cache.cpp @@ -159,28 +159,28 @@ VirtualFile RegisteredCache::GetFileAtID(NcaID id) const { return file; } -static boost::optional<NcaID> CheckMapForContentRecord( +static std::optional<NcaID> CheckMapForContentRecord( const boost::container::flat_map<u64, CNMT>& map, u64 title_id, ContentRecordType type) { if (map.find(title_id) == map.end()) - return boost::none; + return {}; const auto& cnmt = map.at(title_id); const auto iter = std::find_if(cnmt.GetContentRecords().begin(), cnmt.GetContentRecords().end(), [type](const ContentRecord& rec) { return rec.type == type; }); if (iter == cnmt.GetContentRecords().end()) - return boost::none; + return {}; - return boost::make_optional(iter->nca_id); + return std::make_optional(iter->nca_id); } -boost::optional<NcaID> RegisteredCache::GetNcaIDFromMetadata(u64 title_id, - ContentRecordType type) const { +std::optional<NcaID> RegisteredCache::GetNcaIDFromMetadata(u64 title_id, + ContentRecordType type) const { if (type == ContentRecordType::Meta && meta_id.find(title_id) != meta_id.end()) return meta_id.at(title_id); const auto res1 = CheckMapForContentRecord(yuzu_meta, title_id, type); - if (res1 != boost::none) + if (res1) return res1; return CheckMapForContentRecord(meta, title_id, type); } @@ -283,17 +283,14 @@ bool RegisteredCache::HasEntry(RegisteredCacheEntry entry) const { VirtualFile RegisteredCache::GetEntryUnparsed(u64 title_id, ContentRecordType type) const { const auto id = GetNcaIDFromMetadata(title_id, type); - if (id == boost::none) - return nullptr; - - return GetFileAtID(id.get()); + return id ? GetFileAtID(*id) : nullptr; } VirtualFile RegisteredCache::GetEntryUnparsed(RegisteredCacheEntry entry) const { return GetEntryUnparsed(entry.title_id, entry.type); } -boost::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const { +std::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const { const auto meta_iter = meta.find(title_id); if (meta_iter != meta.end()) return meta_iter->second.GetTitleVersion(); @@ -302,15 +299,12 @@ boost::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const { if (yuzu_meta_iter != yuzu_meta.end()) return yuzu_meta_iter->second.GetTitleVersion(); - return boost::none; + return {}; } VirtualFile RegisteredCache::GetEntryRaw(u64 title_id, ContentRecordType type) const { const auto id = GetNcaIDFromMetadata(title_id, type); - if (id == boost::none) - return nullptr; - - return parser(GetFileAtID(id.get()), id.get()); + return id ? parser(GetFileAtID(*id), *id) : nullptr; } VirtualFile RegisteredCache::GetEntryRaw(RegisteredCacheEntry entry) const { @@ -364,8 +358,8 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntries() const { } std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter( - boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type, - boost::optional<u64> title_id) const { + std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type, + std::optional<u64> title_id) const { std::vector<RegisteredCacheEntry> out; IterateAllMetadata<RegisteredCacheEntry>( out, @@ -373,11 +367,11 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter( return RegisteredCacheEntry{c.GetTitleID(), r.type}; }, [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) { - if (title_type != boost::none && title_type.get() != c.GetType()) + if (title_type && *title_type != c.GetType()) return false; - if (record_type != boost::none && record_type.get() != r.type) + if (record_type && *record_type != r.type) return false; - if (title_id != boost::none && title_id.get() != c.GetTitleID()) + if (title_id && *title_id != c.GetTitleID()) return false; return true; }); @@ -459,7 +453,7 @@ InstallResult RegisteredCache::InstallEntry(std::shared_ptr<NCA> nca, TitleType InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy, bool overwrite_if_exists, - boost::optional<NcaID> override_id) { + std::optional<NcaID> override_id) { const auto in = nca->GetBaseFile(); Core::Crypto::SHA256Hash hash{}; @@ -468,12 +462,12 @@ InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const Vfs // game is massive), we're going to cheat and only hash the first MB of the NCA. // Also, for XCIs the NcaID matters, so if the override id isn't none, use that. NcaID id{}; - if (override_id == boost::none) { + if (override_id) { + id = *override_id; + } else { const auto& data = in->ReadBytes(0x100000); mbedtls_sha256(data.data(), data.size(), hash.data(), 0); memcpy(id.data(), hash.data(), 16); - } else { - id = override_id.get(); } std::string path = GetRelativePathFromNcaID(id, false, true); @@ -543,14 +537,14 @@ bool RegisteredCacheUnion::HasEntry(RegisteredCacheEntry entry) const { return HasEntry(entry.title_id, entry.type); } -boost::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const { +std::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const { for (const auto& c : caches) { const auto res = c->GetEntryVersion(title_id); - if (res != boost::none) + if (res) return res; } - return boost::none; + return {}; } VirtualFile RegisteredCacheUnion::GetEntryUnparsed(u64 title_id, ContentRecordType type) const { @@ -609,8 +603,8 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntries() const { } std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter( - boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type, - boost::optional<u64> title_id) const { + std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type, + std::optional<u64> title_id) const { std::vector<RegisteredCacheEntry> out; for (const auto& c : caches) { c->IterateAllMetadata<RegisteredCacheEntry>( @@ -619,11 +613,11 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter( return RegisteredCacheEntry{c.GetTitleID(), r.type}; }, [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) { - if (title_type != boost::none && title_type.get() != c.GetType()) + if (title_type && *title_type != c.GetType()) return false; - if (record_type != boost::none && record_type.get() != r.type) + if (record_type && *record_type != r.type) return false; - if (title_id != boost::none && title_id.get() != c.GetTitleID()) + if (title_id && *title_id != c.GetTitleID()) return false; return true; }); diff --git a/src/core/file_sys/registered_cache.h b/src/core/file_sys/registered_cache.h index 5beceffb3..6cfb16017 100644 --- a/src/core/file_sys/registered_cache.h +++ b/src/core/file_sys/registered_cache.h @@ -84,7 +84,7 @@ public: bool HasEntry(u64 title_id, ContentRecordType type) const; bool HasEntry(RegisteredCacheEntry entry) const; - boost::optional<u32> GetEntryVersion(u64 title_id) const; + std::optional<u32> GetEntryVersion(u64 title_id) const; VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const; VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const; @@ -96,11 +96,10 @@ public: std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const; std::vector<RegisteredCacheEntry> ListEntries() const; - // If a parameter is not boost::none, it will be filtered for from all entries. + // If a parameter is not std::nullopt, it will be filtered for from all entries. std::vector<RegisteredCacheEntry> ListEntriesFilter( - boost::optional<TitleType> title_type = boost::none, - boost::optional<ContentRecordType> record_type = boost::none, - boost::optional<u64> title_id = boost::none) const; + std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {}, + std::optional<u64> title_id = {}) const; // Raw copies all the ncas from the xci/nsp to the csache. Does some quick checks to make sure // there is a meta NCA and all of them are accessible. @@ -125,12 +124,11 @@ private: std::vector<NcaID> AccumulateFiles() const; void ProcessFiles(const std::vector<NcaID>& ids); void AccumulateYuzuMeta(); - boost::optional<NcaID> GetNcaIDFromMetadata(u64 title_id, ContentRecordType type) const; + std::optional<NcaID> GetNcaIDFromMetadata(u64 title_id, ContentRecordType type) const; VirtualFile GetFileAtID(NcaID id) const; VirtualFile OpenFileOrDirectoryConcat(const VirtualDir& dir, std::string_view path) const; InstallResult RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy, - bool overwrite_if_exists, - boost::optional<NcaID> override_id = boost::none); + bool overwrite_if_exists, std::optional<NcaID> override_id = {}); bool RawInstallYuzuMeta(const CNMT& cnmt); VirtualDir dir; @@ -153,7 +151,7 @@ public: bool HasEntry(u64 title_id, ContentRecordType type) const; bool HasEntry(RegisteredCacheEntry entry) const; - boost::optional<u32> GetEntryVersion(u64 title_id) const; + std::optional<u32> GetEntryVersion(u64 title_id) const; VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const; VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const; @@ -165,11 +163,10 @@ public: std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const; std::vector<RegisteredCacheEntry> ListEntries() const; - // If a parameter is not boost::none, it will be filtered for from all entries. + // If a parameter is not std::nullopt, it will be filtered for from all entries. std::vector<RegisteredCacheEntry> ListEntriesFilter( - boost::optional<TitleType> title_type = boost::none, - boost::optional<ContentRecordType> record_type = boost::none, - boost::optional<u64> title_id = boost::none) const; + std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {}, + std::optional<u64> title_id = {}) const; private: std::vector<RegisteredCache*> caches; diff --git a/src/core/file_sys/vfs.cpp b/src/core/file_sys/vfs.cpp index 3824c74e0..7b584de7f 100644 --- a/src/core/file_sys/vfs.cpp +++ b/src/core/file_sys/vfs.cpp @@ -167,13 +167,13 @@ std::string VfsFile::GetExtension() const { VfsDirectory::~VfsDirectory() = default; -boost::optional<u8> VfsFile::ReadByte(std::size_t offset) const { +std::optional<u8> VfsFile::ReadByte(std::size_t offset) const { u8 out{}; std::size_t size = Read(&out, 1, offset); if (size == 1) return out; - return boost::none; + return {}; } std::vector<u8> VfsFile::ReadBytes(std::size_t size, std::size_t offset) const { diff --git a/src/core/file_sys/vfs.h b/src/core/file_sys/vfs.h index 09dc9f288..002f99d4e 100644 --- a/src/core/file_sys/vfs.h +++ b/src/core/file_sys/vfs.h @@ -4,13 +4,15 @@ #pragma once +#include <functional> #include <map> #include <memory> +#include <optional> #include <string> #include <string_view> #include <type_traits> #include <vector> -#include <boost/optional.hpp> + #include "common/common_types.h" #include "core/file_sys/vfs_types.h" @@ -103,8 +105,8 @@ public: // into file. Returns number of bytes successfully written. virtual std::size_t Write(const u8* data, std::size_t length, std::size_t offset = 0) = 0; - // Reads exactly one byte at the offset provided, returning boost::none on error. - virtual boost::optional<u8> ReadByte(std::size_t offset = 0) const; + // Reads exactly one byte at the offset provided, returning std::nullopt on error. + virtual std::optional<u8> ReadByte(std::size_t offset = 0) const; // Reads size bytes starting at offset in file into a vector. virtual std::vector<u8> ReadBytes(std::size_t size, std::size_t offset = 0) const; // Reads all the bytes from the file into a vector. Equivalent to 'file->Read(file->GetSize(), diff --git a/src/core/file_sys/vfs_offset.cpp b/src/core/file_sys/vfs_offset.cpp index a4c6719a0..c96f88488 100644 --- a/src/core/file_sys/vfs_offset.cpp +++ b/src/core/file_sys/vfs_offset.cpp @@ -57,11 +57,11 @@ std::size_t OffsetVfsFile::Write(const u8* data, std::size_t length, std::size_t return file->Write(data, TrimToFit(length, r_offset), offset + r_offset); } -boost::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const { +std::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const { if (r_offset < size) return file->ReadByte(offset + r_offset); - return boost::none; + return {}; } std::vector<u8> OffsetVfsFile::ReadBytes(std::size_t r_size, std::size_t r_offset) const { diff --git a/src/core/file_sys/vfs_offset.h b/src/core/file_sys/vfs_offset.h index 8062702a7..f7b7a3256 100644 --- a/src/core/file_sys/vfs_offset.h +++ b/src/core/file_sys/vfs_offset.h @@ -29,7 +29,7 @@ public: bool IsReadable() const override; std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override; std::size_t Write(const u8* data, std::size_t length, std::size_t offset) override; - boost::optional<u8> ReadByte(std::size_t offset) const override; + std::optional<u8> ReadByte(std::size_t offset) const override; std::vector<u8> ReadBytes(std::size_t size, std::size_t offset) const override; std::vector<u8> ReadAllBytes() const override; bool WriteByte(u8 data, std::size_t offset) override; diff --git a/src/core/file_sys/vfs_static.h b/src/core/file_sys/vfs_static.h index 44fab51d1..9f5a90b1b 100644 --- a/src/core/file_sys/vfs_static.h +++ b/src/core/file_sys/vfs_static.h @@ -53,10 +53,10 @@ public: return 0; } - boost::optional<u8> ReadByte(std::size_t offset) const override { + std::optional<u8> ReadByte(std::size_t offset) const override { if (offset < size) return value; - return boost::none; + return {}; } std::vector<u8> ReadBytes(std::size_t length, std::size_t offset) const override { diff --git a/src/core/hle/ipc_helpers.h b/src/core/hle/ipc_helpers.h index a4bfe2eb0..0a7142ada 100644 --- a/src/core/hle/ipc_helpers.h +++ b/src/core/hle/ipc_helpers.h @@ -117,8 +117,7 @@ public: AlignWithPadding(); - const bool request_has_domain_header{context.GetDomainMessageHeader() != nullptr}; - if (context.Session()->IsDomain() && request_has_domain_header) { + if (context.Session()->IsDomain() && context.HasDomainMessageHeader()) { IPC::DomainMessageHeader domain_header{}; domain_header.num_objects = num_domain_objects; PushRaw(domain_header); diff --git a/src/core/hle/kernel/hle_ipc.h b/src/core/hle/kernel/hle_ipc.h index f01491daa..a38e34b74 100644 --- a/src/core/hle/kernel/hle_ipc.h +++ b/src/core/hle/kernel/hle_ipc.h @@ -161,8 +161,12 @@ public: return buffer_c_desciptors; } - const std::shared_ptr<IPC::DomainMessageHeader>& GetDomainMessageHeader() const { - return domain_message_header; + const IPC::DomainMessageHeader* GetDomainMessageHeader() const { + return domain_message_header.get(); + } + + bool HasDomainMessageHeader() const { + return domain_message_header != nullptr; } /// Helper function to read a buffer using the appropriate buffer descriptor diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp index 4b6b32dd5..1fd4ba5d2 100644 --- a/src/core/hle/kernel/kernel.cpp +++ b/src/core/hle/kernel/kernel.cpp @@ -32,7 +32,7 @@ namespace Kernel { */ static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_late) { const auto proper_handle = static_cast<Handle>(thread_handle); - auto& system = Core::System::GetInstance(); + const auto& system = Core::System::GetInstance(); // Lock the global kernel mutex when we enter the kernel HLE. std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock); @@ -90,7 +90,7 @@ static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_ /// The timer callback event, called when a timer is fired static void TimerCallback(u64 timer_handle, int cycles_late) { const auto proper_handle = static_cast<Handle>(timer_handle); - auto& system = Core::System::GetInstance(); + const auto& system = Core::System::GetInstance(); SharedPtr<Timer> timer = system.Kernel().RetrieveTimerFromCallbackHandleTable(proper_handle); if (timer == nullptr) { diff --git a/src/core/hle/kernel/server_session.cpp b/src/core/hle/kernel/server_session.cpp index 5fc320403..80897f3a4 100644 --- a/src/core/hle/kernel/server_session.cpp +++ b/src/core/hle/kernel/server_session.cpp @@ -63,7 +63,7 @@ void ServerSession::Acquire(Thread* thread) { } ResultCode ServerSession::HandleDomainSyncRequest(Kernel::HLERequestContext& context) { - auto& domain_message_header = context.GetDomainMessageHeader(); + auto* const domain_message_header = context.GetDomainMessageHeader(); if (domain_message_header) { // Set domain handlers in HLE context, used for domain objects (IPC interfaces) as inputs context.SetDomainRequestHandlers(domain_request_handlers); @@ -111,7 +111,7 @@ ResultCode ServerSession::HandleSyncRequest(SharedPtr<Thread> thread) { ResultCode result = RESULT_SUCCESS; // If the session has been converted to a domain, handle the domain request - if (IsDomain() && context.GetDomainMessageHeader()) { + if (IsDomain() && context.HasDomainMessageHeader()) { result = HandleDomainSyncRequest(context); // If there is no domain header, the regular session handler is used } else if (hle_handler != nullptr) { diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index 4e490e2b5..c7c579aaf 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -572,7 +572,7 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) return ERR_INVALID_HANDLE; } - auto& system = Core::System::GetInstance(); + const auto& system = Core::System::GetInstance(); const auto& scheduler = system.CurrentScheduler(); const auto* const current_thread = scheduler.GetCurrentThread(); const bool same_thread = current_thread == thread; diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp index 59bc9e0af..dd5cd9ced 100644 --- a/src/core/hle/kernel/thread.cpp +++ b/src/core/hle/kernel/thread.cpp @@ -4,9 +4,9 @@ #include <algorithm> #include <cinttypes> +#include <optional> #include <vector> -#include <boost/optional.hpp> #include <boost/range/algorithm_ext/erase.hpp> #include "common/assert.h" @@ -94,7 +94,7 @@ void Thread::CancelWakeupTimer() { CoreTiming::UnscheduleEventThreadsafe(kernel.ThreadWakeupCallbackEventType(), callback_handle); } -static boost::optional<s32> GetNextProcessorId(u64 mask) { +static std::optional<s32> GetNextProcessorId(u64 mask) { for (s32 index = 0; index < Core::NUM_CPU_CORES; ++index) { if (mask & (1ULL << index)) { if (!Core::System::GetInstance().Scheduler(index).GetCurrentThread()) { @@ -142,7 +142,7 @@ void Thread::ResumeFromWait() { status = ThreadStatus::Ready; - boost::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask); + std::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask); if (!new_processor_id) { new_processor_id = processor_id; } @@ -369,7 +369,7 @@ void Thread::ChangeCore(u32 core, u64 mask) { return; } - boost::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)}; + std::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)}; if (!new_processor_id) { new_processor_id = processor_id; diff --git a/src/core/hle/service/acc/profile_manager.cpp b/src/core/hle/service/acc/profile_manager.cpp index 3cac1b4ff..c08394e4c 100644 --- a/src/core/hle/service/acc/profile_manager.cpp +++ b/src/core/hle/service/acc/profile_manager.cpp @@ -195,7 +195,7 @@ std::size_t ProfileManager::GetOpenUserCount() const { /// Checks if a user id exists in our profile manager bool ProfileManager::UserExists(UUID uuid) const { - return GetUserIndex(uuid) != std::nullopt; + return GetUserIndex(uuid).has_value(); } bool ProfileManager::UserExistsIndex(std::size_t index) const { diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp index 59aafd616..ac3ff9f20 100644 --- a/src/core/hle/service/am/am.cpp +++ b/src/core/hle/service/am/am.cpp @@ -743,7 +743,7 @@ void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) { Account::ProfileManager profile_manager{}; const auto uuid = profile_manager.GetUser(Settings::values.current_user); - ASSERT(uuid != std::nullopt); + ASSERT(uuid); params.current_user = uuid->uuid; IPC::ResponseBuilder rb{ctx, 2, 0, 1}; diff --git a/src/core/hle/service/nvflinger/buffer_queue.cpp b/src/core/hle/service/nvflinger/buffer_queue.cpp index fd98d541d..630ebbfc7 100644 --- a/src/core/hle/service/nvflinger/buffer_queue.cpp +++ b/src/core/hle/service/nvflinger/buffer_queue.cpp @@ -31,7 +31,7 @@ void BufferQueue::SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer) buffer_wait_event->Signal(); } -boost::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) { +std::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) { auto itr = std::find_if(queue.begin(), queue.end(), [&](const Buffer& buffer) { // Only consider free buffers. Buffers become free once again after they've been Acquired // and Released by the compositor, see the NVFlinger::Compose method. @@ -44,7 +44,7 @@ boost::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) { }); if (itr == queue.end()) { - return boost::none; + return {}; } itr->status = Buffer::Status::Dequeued; @@ -70,12 +70,12 @@ void BufferQueue::QueueBuffer(u32 slot, BufferTransformFlags transform, itr->crop_rect = crop_rect; } -boost::optional<const BufferQueue::Buffer&> BufferQueue::AcquireBuffer() { +std::optional<std::reference_wrapper<const BufferQueue::Buffer>> BufferQueue::AcquireBuffer() { auto itr = std::find_if(queue.begin(), queue.end(), [](const Buffer& buffer) { return buffer.status == Buffer::Status::Queued; }); if (itr == queue.end()) - return boost::none; + return {}; itr->status = Buffer::Status::Acquired; return *itr; } diff --git a/src/core/hle/service/nvflinger/buffer_queue.h b/src/core/hle/service/nvflinger/buffer_queue.h index 50b767732..2fe81a560 100644 --- a/src/core/hle/service/nvflinger/buffer_queue.h +++ b/src/core/hle/service/nvflinger/buffer_queue.h @@ -4,8 +4,9 @@ #pragma once +#include <optional> #include <vector> -#include <boost/optional.hpp> + #include "common/common_funcs.h" #include "common/math_util.h" #include "common/swap.h" @@ -73,11 +74,11 @@ public: }; void SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer); - boost::optional<u32> DequeueBuffer(u32 width, u32 height); + std::optional<u32> DequeueBuffer(u32 width, u32 height); const IGBPBuffer& RequestBuffer(u32 slot) const; void QueueBuffer(u32 slot, BufferTransformFlags transform, const MathUtil::Rectangle<int>& crop_rect); - boost::optional<const Buffer&> AcquireBuffer(); + std::optional<std::reference_wrapper<const Buffer>> AcquireBuffer(); void ReleaseBuffer(u32 slot); u32 Query(QueryType type); diff --git a/src/core/hle/service/nvflinger/nvflinger.cpp b/src/core/hle/service/nvflinger/nvflinger.cpp index d47b6f659..214e6d1b3 100644 --- a/src/core/hle/service/nvflinger/nvflinger.cpp +++ b/src/core/hle/service/nvflinger/nvflinger.cpp @@ -3,7 +3,7 @@ // Refer to the license.txt file included. #include <algorithm> -#include <boost/optional.hpp> +#include <optional> #include "common/alignment.h" #include "common/assert.h" @@ -134,7 +134,7 @@ void NVFlinger::Compose() { MicroProfileFlip(); - if (buffer == boost::none) { + if (!buffer) { auto& system_instance = Core::System::GetInstance(); // There was no queued buffer to draw, render previous frame @@ -143,7 +143,7 @@ void NVFlinger::Compose() { continue; } - auto& igbp_buffer = buffer->igbp_buffer; + auto& igbp_buffer = buffer->get().igbp_buffer; // Now send the buffer to the GPU for drawing. // TODO(Subv): Support more than just disp0. The display device selection is probably based @@ -152,10 +152,10 @@ void NVFlinger::Compose() { ASSERT(nvdisp); nvdisp->flip(igbp_buffer.gpu_buffer_id, igbp_buffer.offset, igbp_buffer.format, - igbp_buffer.width, igbp_buffer.height, igbp_buffer.stride, buffer->transform, - buffer->crop_rect); + igbp_buffer.width, igbp_buffer.height, igbp_buffer.stride, + buffer->get().transform, buffer->get().crop_rect); - buffer_queue->ReleaseBuffer(buffer->slot); + buffer_queue->ReleaseBuffer(buffer->get().slot); } } diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp index 184537daa..d764b2406 100644 --- a/src/core/hle/service/vi/vi.cpp +++ b/src/core/hle/service/vi/vi.cpp @@ -6,9 +6,10 @@ #include <array> #include <cstring> #include <memory> +#include <optional> #include <type_traits> #include <utility> -#include <boost/optional.hpp> + #include "common/alignment.h" #include "common/assert.h" #include "common/common_funcs.h" @@ -506,9 +507,9 @@ private: IGBPDequeueBufferRequestParcel request{ctx.ReadBuffer()}; const u32 width{request.data.width}; const u32 height{request.data.height}; - boost::optional<u32> slot = buffer_queue->DequeueBuffer(width, height); + std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height); - if (slot != boost::none) { + if (slot) { // Buffer is available IGBPDequeueBufferResponseParcel response{*slot}; ctx.WriteBuffer(response.Serialize()); @@ -520,7 +521,7 @@ private: Kernel::ThreadWakeupReason reason) { // Repeat TransactParcel DequeueBuffer when a buffer is available auto buffer_queue = nv_flinger->GetBufferQueue(id); - boost::optional<u32> slot = buffer_queue->DequeueBuffer(width, height); + std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height); IGBPDequeueBufferResponseParcel response{*slot}; ctx.WriteBuffer(response.Serialize()); IPC::ResponseBuilder rb{ctx, 2}; diff --git a/src/core/loader/loader.h b/src/core/loader/loader.h index e562b3a04..7686634bf 100644 --- a/src/core/loader/loader.h +++ b/src/core/loader/loader.h @@ -6,10 +6,11 @@ #include <iosfwd> #include <memory> +#include <optional> #include <string> #include <utility> #include <vector> -#include <boost/optional.hpp> + #include "common/common_types.h" #include "core/file_sys/vfs.h" @@ -145,7 +146,7 @@ public: * information. * @returns A pair with the optional system mode, and and the status. */ - virtual std::pair<boost::optional<u32>, ResultStatus> LoadKernelSystemMode() { + virtual std::pair<std::optional<u32>, ResultStatus> LoadKernelSystemMode() { // 96MB allocated to the application. return std::make_pair(2, ResultStatus::Success); } diff --git a/src/core/memory.cpp b/src/core/memory.cpp index 014298ed6..70abd856a 100644 --- a/src/core/memory.cpp +++ b/src/core/memory.cpp @@ -4,9 +4,9 @@ #include <algorithm> #include <cstring> +#include <optional> #include <utility> -#include <boost/optional.hpp> #include "common/assert.h" #include "common/common_types.h" #include "common/logging/log.h" diff --git a/src/core/memory_hook.h b/src/core/memory_hook.h index 0269c7ff1..940777107 100644 --- a/src/core/memory_hook.h +++ b/src/core/memory_hook.h @@ -5,7 +5,8 @@ #pragma once #include <memory> -#include <boost/optional.hpp> +#include <optional> + #include "common/common_types.h" namespace Memory { @@ -18,19 +19,19 @@ namespace Memory { * * A hook may be mapped to multiple regions of memory. * - * If a boost::none or false is returned from a function, the read/write request is passed through + * If a std::nullopt or false is returned from a function, the read/write request is passed through * to the underlying memory region. */ class MemoryHook { public: virtual ~MemoryHook(); - virtual boost::optional<bool> IsValidAddress(VAddr addr) = 0; + virtual std::optional<bool> IsValidAddress(VAddr addr) = 0; - virtual boost::optional<u8> Read8(VAddr addr) = 0; - virtual boost::optional<u16> Read16(VAddr addr) = 0; - virtual boost::optional<u32> Read32(VAddr addr) = 0; - virtual boost::optional<u64> Read64(VAddr addr) = 0; + virtual std::optional<u8> Read8(VAddr addr) = 0; + virtual std::optional<u16> Read16(VAddr addr) = 0; + virtual std::optional<u32> Read32(VAddr addr) = 0; + virtual std::optional<u64> Read64(VAddr addr) = 0; virtual bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) = 0; diff --git a/src/tests/core/arm/arm_test_common.cpp b/src/tests/core/arm/arm_test_common.cpp index 37e15bad0..9b8a44fa1 100644 --- a/src/tests/core/arm/arm_test_common.cpp +++ b/src/tests/core/arm/arm_test_common.cpp @@ -64,11 +64,11 @@ void TestEnvironment::ClearWriteRecords() { TestEnvironment::TestMemory::~TestMemory() {} -boost::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) { +std::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) { return true; } -boost::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) { +std::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) { const auto iter = data.find(addr); if (iter == data.end()) { @@ -79,15 +79,15 @@ boost::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) { return iter->second; } -boost::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) { +std::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) { return *Read8(addr) | static_cast<u16>(*Read8(addr + 1)) << 8; } -boost::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) { +std::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) { return *Read16(addr) | static_cast<u32>(*Read16(addr + 2)) << 16; } -boost::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) { +std::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) { return *Read32(addr) | static_cast<u64>(*Read32(addr + 4)) << 32; } diff --git a/src/tests/core/arm/arm_test_common.h b/src/tests/core/arm/arm_test_common.h index 5de8dab4e..0b7539601 100644 --- a/src/tests/core/arm/arm_test_common.h +++ b/src/tests/core/arm/arm_test_common.h @@ -64,12 +64,12 @@ private: ~TestMemory() override; - boost::optional<bool> IsValidAddress(VAddr addr) override; + std::optional<bool> IsValidAddress(VAddr addr) override; - boost::optional<u8> Read8(VAddr addr) override; - boost::optional<u16> Read16(VAddr addr) override; - boost::optional<u32> Read32(VAddr addr) override; - boost::optional<u64> Read64(VAddr addr) override; + std::optional<u8> Read8(VAddr addr) override; + std::optional<u16> Read16(VAddr addr) override; + std::optional<u32> Read32(VAddr addr) override; + std::optional<u64> Read64(VAddr addr) override; bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) override; diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index 09ecc5bad..c5f7128ec 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -51,6 +51,8 @@ add_library(video_core STATIC renderer_opengl/maxwell_to_gl.h renderer_opengl/renderer_opengl.cpp renderer_opengl/renderer_opengl.h + renderer_opengl/utils.cpp + renderer_opengl/utils.h textures/astc.cpp textures/astc.h textures/decoders.cpp diff --git a/src/video_core/command_processor.cpp b/src/video_core/command_processor.cpp index f1aa6091b..28e8c13aa 100644 --- a/src/video_core/command_processor.cpp +++ b/src/video_core/command_processor.cpp @@ -81,7 +81,7 @@ void GPU::ProcessCommandLists(const std::vector<CommandListHeader>& commands) { for (auto entry : commands) { Tegra::GPUVAddr address = entry.Address(); u32 size = entry.sz; - const boost::optional<VAddr> head_address = memory_manager->GpuToCpuAddress(address); + const std::optional<VAddr> head_address = memory_manager->GpuToCpuAddress(address); VAddr current_addr = *head_address; while (current_addr < *head_address + size * sizeof(CommandHeader)) { const CommandHeader header = {Memory::Read32(current_addr)}; diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index 27ef865a2..7357d20d1 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -167,7 +167,7 @@ void Maxwell3D::ProcessQueryGet() { GPUVAddr sequence_address = regs.query.QueryAddress(); // Since the sequence address is given as a GPU VAddr, we have to convert it to an application // VAddr before writing. - boost::optional<VAddr> address = memory_manager.GpuToCpuAddress(sequence_address); + std::optional<VAddr> address = memory_manager.GpuToCpuAddress(sequence_address); // TODO(Subv): Support the other query units. ASSERT_MSG(regs.query.query_get.unit == Regs::QueryUnit::Crop, @@ -285,7 +285,7 @@ void Maxwell3D::ProcessCBData(u32 value) { // Don't allow writing past the end of the buffer. ASSERT(regs.const_buffer.cb_pos + sizeof(u32) <= regs.const_buffer.cb_size); - boost::optional<VAddr> address = + std::optional<VAddr> address = memory_manager.GpuToCpuAddress(buffer_address + regs.const_buffer.cb_pos); Memory::Write32(*address, value); @@ -298,7 +298,7 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const { GPUVAddr tic_base_address = regs.tic.TICAddress(); GPUVAddr tic_address_gpu = tic_base_address + tic_index * sizeof(Texture::TICEntry); - boost::optional<VAddr> tic_address_cpu = memory_manager.GpuToCpuAddress(tic_address_gpu); + std::optional<VAddr> tic_address_cpu = memory_manager.GpuToCpuAddress(tic_address_gpu); Texture::TICEntry tic_entry; Memory::ReadBlock(*tic_address_cpu, &tic_entry, sizeof(Texture::TICEntry)); @@ -322,7 +322,7 @@ Texture::TSCEntry Maxwell3D::GetTSCEntry(u32 tsc_index) const { GPUVAddr tsc_base_address = regs.tsc.TSCAddress(); GPUVAddr tsc_address_gpu = tsc_base_address + tsc_index * sizeof(Texture::TSCEntry); - boost::optional<VAddr> tsc_address_cpu = memory_manager.GpuToCpuAddress(tsc_address_gpu); + std::optional<VAddr> tsc_address_cpu = memory_manager.GpuToCpuAddress(tsc_address_gpu); Texture::TSCEntry tsc_entry; Memory::ReadBlock(*tsc_address_cpu, &tsc_entry, sizeof(Texture::TSCEntry)); @@ -386,7 +386,7 @@ Texture::FullTextureInfo Maxwell3D::GetStageTexture(Regs::ShaderStage stage, ASSERT(tex_info_address < tex_info_buffer.address + tex_info_buffer.size); - boost::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address); + std::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address); Texture::TextureHandle tex_handle{Memory::Read32(*tex_address_cpu)}; Texture::FullTextureInfo tex_info{}; diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h index 141b9159b..b84da512f 100644 --- a/src/video_core/engines/shader_bytecode.h +++ b/src/video_core/engines/shader_bytecode.h @@ -5,12 +5,11 @@ #pragma once #include <bitset> +#include <optional> #include <string> #include <tuple> #include <vector> -#include <boost/optional.hpp> - #include "common/assert.h" #include "common/bit_field.h" #include "common/common_types.h" @@ -1456,7 +1455,7 @@ public: Type type; }; - static boost::optional<const Matcher&> Decode(Instruction instr) { + static std::optional<std::reference_wrapper<const Matcher>> Decode(Instruction instr) { static const auto table{GetDecodeTable()}; const auto matches_instruction = [instr](const auto& matcher) { @@ -1464,7 +1463,8 @@ public: }; auto iter = std::find_if(table.begin(), table.end(), matches_instruction); - return iter != table.end() ? boost::optional<const Matcher&>(*iter) : boost::none; + return iter != table.end() ? std::optional<std::reference_wrapper<const Matcher>>(*iter) + : std::nullopt; } private: diff --git a/src/video_core/macro_interpreter.cpp b/src/video_core/macro_interpreter.cpp index 377bd66ab..f6af132fb 100644 --- a/src/video_core/macro_interpreter.cpp +++ b/src/video_core/macro_interpreter.cpp @@ -29,7 +29,7 @@ void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> pa void MacroInterpreter::Reset() { registers = {}; pc = 0; - delayed_pc = boost::none; + delayed_pc = {}; method_address.raw = 0; parameters.clear(); // The next parameter index starts at 1, because $r1 already has the value of the first @@ -44,10 +44,10 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) { pc += 4; // Update the program counter if we were delayed - if (delayed_pc != boost::none) { + if (delayed_pc) { ASSERT(is_delay_slot); pc = *delayed_pc; - delayed_pc = boost::none; + delayed_pc = {}; } switch (opcode.operation) { diff --git a/src/video_core/macro_interpreter.h b/src/video_core/macro_interpreter.h index cee0baaf3..773684bde 100644 --- a/src/video_core/macro_interpreter.h +++ b/src/video_core/macro_interpreter.h @@ -5,8 +5,9 @@ #pragma once #include <array> +#include <optional> #include <vector> -#include <boost/optional.hpp> + #include "common/bit_field.h" #include "common/common_types.h" @@ -149,7 +150,7 @@ private: Engines::Maxwell3D& maxwell3d; u32 pc; ///< Current program counter - boost::optional<u32> + std::optional<u32> delayed_pc; ///< Program counter to execute at after the delay slot is executed. static constexpr std::size_t NumMacroRegisters = 8; diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp index 022d4ab74..90a8e825d 100644 --- a/src/video_core/memory_manager.cpp +++ b/src/video_core/memory_manager.cpp @@ -9,7 +9,7 @@ namespace Tegra { GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) { - boost::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, align); + std::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, align); ASSERT(gpu_addr); for (u64 offset = 0; offset < size; offset += PAGE_SIZE) { @@ -34,7 +34,7 @@ GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) { } GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) { - boost::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, PAGE_SIZE); + std::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, PAGE_SIZE); ASSERT(gpu_addr); for (u64 offset = 0; offset < size; offset += PAGE_SIZE) { @@ -97,7 +97,7 @@ GPUVAddr MemoryManager::GetRegionEnd(GPUVAddr region_start) const { return {}; } -boost::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) { +std::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) { GPUVAddr gpu_addr = 0; u64 free_space = 0; align = (align + PAGE_MASK) & ~PAGE_MASK; @@ -118,7 +118,7 @@ boost::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) { return {}; } -boost::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) { +std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) { VAddr base_addr = PageSlot(gpu_addr); if (base_addr == static_cast<u64>(PageStatus::Allocated) || diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h index caf80093f..b1255fd56 100644 --- a/src/video_core/memory_manager.h +++ b/src/video_core/memory_manager.h @@ -6,10 +6,9 @@ #include <array> #include <memory> +#include <optional> #include <vector> -#include <boost/optional.hpp> - #include "common/common_types.h" namespace Tegra { @@ -27,7 +26,7 @@ public: GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size); GPUVAddr UnmapBuffer(GPUVAddr gpu_addr, u64 size); GPUVAddr GetRegionEnd(GPUVAddr region_start) const; - boost::optional<VAddr> GpuToCpuAddress(GPUVAddr gpu_addr); + std::optional<VAddr> GpuToCpuAddress(GPUVAddr gpu_addr); std::vector<GPUVAddr> CpuToGpuAddress(VAddr cpu_addr) const; static constexpr u64 PAGE_BITS = 16; @@ -35,7 +34,7 @@ public: static constexpr u64 PAGE_MASK = PAGE_SIZE - 1; private: - boost::optional<GPUVAddr> FindFreeBlock(u64 size, u64 align = 1); + std::optional<GPUVAddr> FindFreeBlock(u64 size, u64 align = 1); bool IsPageMapped(GPUVAddr gpu_addr); VAddr& PageSlot(GPUVAddr gpu_addr); diff --git a/src/video_core/renderer_base.h b/src/video_core/renderer_base.h index 2cd0738ff..669e26e15 100644 --- a/src/video_core/renderer_base.h +++ b/src/video_core/renderer_base.h @@ -6,7 +6,8 @@ #include <atomic> #include <memory> -#include <boost/optional.hpp> +#include <optional> + #include "common/common_types.h" #include "video_core/gpu.h" #include "video_core/rasterizer_interface.h" @@ -28,7 +29,8 @@ public: virtual ~RendererBase(); /// Swap buffers (render frame) - virtual void SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) = 0; + virtual void SwapBuffers( + std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) = 0; /// Initialize the renderer virtual bool Init() = 0; diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp index c142095c5..41a54b3e7 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp @@ -17,7 +17,7 @@ OGLBufferCache::OGLBufferCache(std::size_t size) : stream_buffer(GL_ARRAY_BUFFER GLintptr OGLBufferCache::UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size, std::size_t alignment, bool cache) { auto& memory_manager = Core::System::GetInstance().GPU().MemoryManager(); - const boost::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)}; + const std::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)}; // Cache management is a big overhead, so only cache entries with a given size. // TODO: Figure out which size is the best for given games. diff --git a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp index ee1d9601b..741f14bc3 100644 --- a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp +++ b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp @@ -45,7 +45,7 @@ GLintptr PrimitiveAssembler::MakeQuadIndexed(Tegra::GPUVAddr gpu_addr, std::size auto [dst_pointer, index_offset] = buffer_cache.ReserveMemory(map_size); auto& memory_manager = Core::System::GetInstance().GPU().MemoryManager(); - const boost::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)}; + const std::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)}; const u8* source{Memory::GetPointer(*cpu_addr)}; for (u32 primitive = 0; primitive < count / 4; ++primitive) { diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index cb180b93c..bf381271e 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -401,7 +401,7 @@ void RasterizerOpenGL::UpdatePagesCachedCount(VAddr addr, u64 size, int delta) { void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_depth_fb, bool preserve_contents, - boost::optional<std::size_t> single_color_target) { + std::optional<std::size_t> single_color_target) { MICROPROFILE_SCOPE(OpenGL_Framebuffer); const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; @@ -731,11 +731,15 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr if (mag_filter != config.mag_filter) { mag_filter = config.mag_filter; - glSamplerParameteri(s, GL_TEXTURE_MAG_FILTER, MaxwellToGL::TextureFilterMode(mag_filter)); + glSamplerParameteri( + s, GL_TEXTURE_MAG_FILTER, + MaxwellToGL::TextureFilterMode(mag_filter, Tegra::Texture::TextureMipmapFilter::None)); } - if (min_filter != config.min_filter) { + if (min_filter != config.min_filter || mip_filter != config.mip_filter) { min_filter = config.min_filter; - glSamplerParameteri(s, GL_TEXTURE_MIN_FILTER, MaxwellToGL::TextureFilterMode(min_filter)); + mip_filter = config.mip_filter; + glSamplerParameteri(s, GL_TEXTURE_MIN_FILTER, + MaxwellToGL::TextureFilterMode(min_filter, mip_filter)); } if (wrap_u != config.wrap_u) { diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index 5020a5392..47097c569 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -8,12 +8,12 @@ #include <cstddef> #include <map> #include <memory> +#include <optional> #include <tuple> #include <utility> #include <vector> #include <boost/icl/interval_map.hpp> -#include <boost/optional.hpp> #include <boost/range/iterator_range.hpp> #include <glad/glad.h> @@ -93,6 +93,7 @@ private: private: Tegra::Texture::TextureFilter mag_filter; Tegra::Texture::TextureFilter min_filter; + Tegra::Texture::TextureMipmapFilter mip_filter; Tegra::Texture::WrapMode wrap_u; Tegra::Texture::WrapMode wrap_v; Tegra::Texture::WrapMode wrap_p; @@ -110,7 +111,7 @@ private: */ void ConfigureFramebuffers(bool use_color_fb = true, bool using_depth_fb = true, bool preserve_contents = true, - boost::optional<std::size_t> single_color_target = {}); + std::optional<std::size_t> single_color_target = {}); /* * Configures the current constbuffers to use for the draw command. diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp index b057e2efa..1d43a419d 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp @@ -16,6 +16,7 @@ #include "core/settings.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_opengl/gl_rasterizer_cache.h" +#include "video_core/renderer_opengl/utils.h" #include "video_core/textures/astc.h" #include "video_core/textures/decoders.h" #include "video_core/utils.h" @@ -90,27 +91,36 @@ void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) { } } -std::size_t SurfaceParams::InnerMemorySize(bool layer_only) const { +std::size_t SurfaceParams::InnerMipmapMemorySize(u32 mip_level, bool force_gl, bool layer_only, + bool uncompressed) const { const u32 compression_factor{GetCompressionFactor(pixel_format)}; const u32 bytes_per_pixel{GetBytesPerPixel(pixel_format)}; u32 m_depth = (layer_only ? 1U : depth); - u32 m_width = std::max(1U, width / compression_factor); - u32 m_height = std::max(1U, height / compression_factor); - std::size_t size = Tegra::Texture::CalculateSize(is_tiled, bytes_per_pixel, m_width, m_height, - m_depth, block_height, block_depth); - u32 m_block_height = block_height; - u32 m_block_depth = block_depth; - std::size_t block_size_bytes = 512 * block_height * block_depth; // 512 is GOB size - for (u32 i = 1; i < max_mip_level; i++) { - m_width = std::max(1U, m_width / 2); - m_height = std::max(1U, m_height / 2); - m_depth = std::max(1U, m_depth / 2); - m_block_height = std::max(1U, m_block_height / 2); - m_block_depth = std::max(1U, m_block_depth / 2); - size += Tegra::Texture::CalculateSize(is_tiled, bytes_per_pixel, m_width, m_height, m_depth, - m_block_height, m_block_depth); + u32 m_width = MipWidth(mip_level); + u32 m_height = MipHeight(mip_level); + m_width = uncompressed ? m_width + : std::max(1U, (m_width + compression_factor - 1) / compression_factor); + m_height = uncompressed + ? m_height + : std::max(1U, (m_height + compression_factor - 1) / compression_factor); + m_depth = std::max(1U, m_depth >> mip_level); + u32 m_block_height = MipBlockHeight(mip_level); + u32 m_block_depth = MipBlockDepth(mip_level); + return Tegra::Texture::CalculateSize(force_gl ? false : is_tiled, bytes_per_pixel, m_width, + m_height, m_depth, m_block_height, m_block_depth); +} + +std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only, + bool uncompressed) const { + std::size_t block_size_bytes = Tegra::Texture::GetGOBSize() * block_height * block_depth; + std::size_t size = 0; + for (u32 i = 0; i < max_mip_level; i++) { + size += InnerMipmapMemorySize(i, force_gl, layer_only, uncompressed); + } + if (!force_gl && is_tiled) { + size = Common::AlignUp(size, block_size_bytes); } - return is_tiled ? Common::AlignUp(size, block_size_bytes) : size; + return size; } /*static*/ SurfaceParams SurfaceParams::CreateForTexture( @@ -188,7 +198,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool layer_only) const { params.unaligned_height = config.height; params.target = SurfaceTarget::Texture2D; params.depth = 1; - params.max_mip_level = 0; + params.max_mip_level = 1; params.is_layered = false; // Render target specific parameters, not used for caching @@ -222,7 +232,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool layer_only) const { params.unaligned_height = zeta_height; params.target = SurfaceTarget::Texture2D; params.depth = 1; - params.max_mip_level = 0; + params.max_mip_level = 1; params.is_layered = false; params.rt = {}; @@ -249,7 +259,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool layer_only) const { params.unaligned_height = config.height; params.target = SurfaceTarget::Texture2D; params.depth = 1; - params.max_mip_level = 0; + params.max_mip_level = 1; params.rt = {}; params.InitCacheParameters(config.Address()); @@ -273,7 +283,7 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form {GL_R11F_G11F_B10F, GL_RGB, GL_UNSIGNED_INT_10F_11F_11F_REV, ComponentType::Float, false}, // R11FG11FB10F {GL_RGBA32UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false}, // RGBA32UI - {GL_COMPRESSED_RGB_S3TC_DXT1_EXT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, + {GL_COMPRESSED_RGBA_S3TC_DXT1_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, true}, // DXT1 {GL_COMPRESSED_RGBA_S3TC_DXT3_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, true}, // DXT23 @@ -318,7 +328,7 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X4 {GL_SRGB8_ALPHA8, GL_BGRA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // BGRA8 // Compressed sRGB formats - {GL_COMPRESSED_SRGB_S3TC_DXT1_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, + {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, true}, // DXT1_SRGB {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, true}, // DXT23_SRGB @@ -373,13 +383,13 @@ static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType return format; } -MathUtil::Rectangle<u32> SurfaceParams::GetRect() const { - u32 actual_height{unaligned_height}; +MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const { + u32 actual_height{std::max(1U, unaligned_height >> mip_level)}; if (IsPixelFormatASTC(pixel_format)) { // ASTC formats must stop at the ATSC block size boundary actual_height = Common::AlignDown(actual_height, GetASTCBlockSize(pixel_format).second); } - return {0, actual_height, width, 0}; + return {0, actual_height, MipWidth(mip_level), 0}; } /// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN @@ -563,28 +573,31 @@ static constexpr GLConversionArray gl_to_morton_fns = { }; void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params, - std::vector<u8>& gl_buffer) { - u32 depth = params.depth; + std::vector<u8>& gl_buffer, u32 mip_level) { + u32 depth = params.MipDepth(mip_level); if (params.target == SurfaceParams::SurfaceTarget::Texture2D) { // TODO(Blinkhawk): Eliminate this condition once all texture types are implemented. depth = 1U; } if (params.is_layered) { - u64 offset = 0; + u64 offset = params.GetMipmapLevelOffset(mip_level); u64 offset_gl = 0; u64 layer_size = params.LayerMemorySize(); - u64 gl_size = params.LayerSizeGL(); - for (u32 i = 0; i < depth; i++) { + u64 gl_size = params.LayerSizeGL(mip_level); + for (u32 i = 0; i < params.depth; i++) { functions[static_cast<std::size_t>(params.pixel_format)]( - params.width, params.block_height, params.height, params.block_depth, 1, + params.MipWidth(mip_level), params.MipBlockHeight(mip_level), + params.MipHeight(mip_level), params.MipBlockDepth(mip_level), 1, gl_buffer.data() + offset_gl, gl_size, params.addr + offset); offset += layer_size; offset_gl += gl_size; } } else { + u64 offset = params.GetMipmapLevelOffset(mip_level); functions[static_cast<std::size_t>(params.pixel_format)]( - params.width, params.block_height, params.height, params.block_depth, depth, - gl_buffer.data(), gl_buffer.size(), params.addr); + params.MipWidth(mip_level), params.MipBlockHeight(mip_level), + params.MipHeight(mip_level), params.MipBlockDepth(mip_level), depth, gl_buffer.data(), + gl_buffer.size(), params.addr + offset); } } @@ -839,34 +852,41 @@ CachedSurface::CachedSurface(const SurfaceParams& params) // Only pre-create the texture for non-compressed textures. switch (params.target) { case SurfaceParams::SurfaceTarget::Texture1D: - glTexStorage1D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format, - rect.GetWidth()); + glTexStorage1D(SurfaceTargetToGL(params.target), params.max_mip_level, + format_tuple.internal_format, rect.GetWidth()); break; case SurfaceParams::SurfaceTarget::Texture2D: case SurfaceParams::SurfaceTarget::TextureCubemap: - glTexStorage2D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format, - rect.GetWidth(), rect.GetHeight()); + glTexStorage2D(SurfaceTargetToGL(params.target), params.max_mip_level, + format_tuple.internal_format, rect.GetWidth(), rect.GetHeight()); break; case SurfaceParams::SurfaceTarget::Texture3D: case SurfaceParams::SurfaceTarget::Texture2DArray: - glTexStorage3D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format, - rect.GetWidth(), rect.GetHeight(), params.depth); + glTexStorage3D(SurfaceTargetToGL(params.target), params.max_mip_level, + format_tuple.internal_format, rect.GetWidth(), rect.GetHeight(), + params.depth); break; default: LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}", static_cast<u32>(params.target)); UNREACHABLE(); - glTexStorage2D(GL_TEXTURE_2D, 1, format_tuple.internal_format, rect.GetWidth(), - rect.GetHeight()); + glTexStorage2D(GL_TEXTURE_2D, params.max_mip_level, format_tuple.internal_format, + rect.GetWidth(), rect.GetHeight()); } } glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_MIN_FILTER, GL_LINEAR); + glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_MAG_FILTER, GL_LINEAR); glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE); glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE); + glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_MAX_LEVEL, + params.max_mip_level - 1); + if (params.max_mip_level == 1) { + glTexParameterf(SurfaceTargetToGL(params.target), GL_TEXTURE_LOD_BIAS, 1000.0); + } - VideoCore::LabelGLObject(GL_TEXTURE, texture.handle, params.addr, - SurfaceParams::SurfaceTargetName(params.target)); + LabelGLObject(GL_TEXTURE, texture.handle, params.addr, + SurfaceParams::SurfaceTargetName(params.target)); // Clamp size to mapped GPU memory region // TODO(bunnei): Super Mario Odyssey maps a 0x40000 byte region and then uses it for a 0x80000 @@ -992,20 +1012,22 @@ static void ConvertFormatAsNeeded_FlushGLBuffer(std::vector<u8>& data, PixelForm MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 64, 192)); void CachedSurface::LoadGLBuffer() { MICROPROFILE_SCOPE(OpenGL_SurfaceLoad); - - gl_buffer.resize(params.size_in_bytes_gl); + gl_buffer.resize(params.max_mip_level); + for (u32 i = 0; i < params.max_mip_level; i++) + gl_buffer[i].resize(params.GetMipmapSizeGL(i)); if (params.is_tiled) { ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}", params.block_width, static_cast<u32>(params.target)); - - SwizzleFunc(morton_to_gl_fns, params, gl_buffer); + for (u32 i = 0; i < params.max_mip_level; i++) + SwizzleFunc(morton_to_gl_fns, params, gl_buffer[i], i); } else { const auto texture_src_data{Memory::GetPointer(params.addr)}; const auto texture_src_data_end{texture_src_data + params.size_in_bytes_gl}; - gl_buffer.assign(texture_src_data, texture_src_data_end); + gl_buffer[0].assign(texture_src_data, texture_src_data_end); } - - ConvertFormatAsNeeded_LoadGLBuffer(gl_buffer, params.pixel_format, params.width, params.height); + for (u32 i = 0; i < params.max_mip_level; i++) + ConvertFormatAsNeeded_LoadGLBuffer(gl_buffer[i], params.pixel_format, params.MipWidth(i), + params.MipHeight(i)); } MICROPROFILE_DEFINE(OpenGL_SurfaceFlush, "OpenGL", "Surface Flush", MP_RGB(128, 192, 64)); @@ -1015,7 +1037,8 @@ void CachedSurface::FlushGLBuffer() { ASSERT_MSG(!IsPixelFormatASTC(params.pixel_format), "Unimplemented"); // OpenGL temporary buffer needs to be big enough to store raw texture size - gl_buffer.resize(GetSizeInBytes()); + gl_buffer.resize(1); + gl_buffer[0].resize(GetSizeInBytes()); const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type); // Ensure no bad interactions with GL_UNPACK_ALIGNMENT @@ -1024,9 +1047,9 @@ void CachedSurface::FlushGLBuffer() { ASSERT(!tuple.compressed); glBindBuffer(GL_PIXEL_PACK_BUFFER, 0); glGetTextureImage(texture.handle, 0, tuple.format, tuple.type, - static_cast<GLsizei>(gl_buffer.size()), gl_buffer.data()); + static_cast<GLsizei>(gl_buffer[0].size()), gl_buffer[0].data()); glPixelStorei(GL_PACK_ROW_LENGTH, 0); - ConvertFormatAsNeeded_FlushGLBuffer(gl_buffer, params.pixel_format, params.width, + ConvertFormatAsNeeded_FlushGLBuffer(gl_buffer[0], params.pixel_format, params.width, params.height); ASSERT(params.type != SurfaceType::Fill); const u8* const texture_src_data = Memory::GetPointer(params.addr); @@ -1035,26 +1058,21 @@ void CachedSurface::FlushGLBuffer() { ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}", params.block_width, static_cast<u32>(params.target)); - SwizzleFunc(gl_to_morton_fns, params, gl_buffer); + SwizzleFunc(gl_to_morton_fns, params, gl_buffer[0], 0); } else { - std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer.data(), GetSizeInBytes()); + std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer[0].data(), GetSizeInBytes()); } } -MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 64, 192)); -void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle) { - if (params.type == SurfaceType::Fill) - return; - - MICROPROFILE_SCOPE(OpenGL_TextureUL); - - const auto& rect{params.GetRect()}; +void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, + GLuint draw_fb_handle) { + const auto& rect{params.GetRect(mip_map)}; // Load data from memory to the surface const GLint x0 = static_cast<GLint>(rect.left); const GLint y0 = static_cast<GLint>(rect.bottom); std::size_t buffer_offset = - static_cast<std::size_t>(static_cast<std::size_t>(y0) * params.width + + static_cast<std::size_t>(static_cast<std::size_t>(y0) * params.MipWidth(mip_map) + static_cast<std::size_t>(x0)) * SurfaceParams::GetBytesPerPixel(params.pixel_format); @@ -1072,88 +1090,117 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle cur_state.Apply(); // Ensure no bad interactions with GL_UNPACK_ALIGNMENT - ASSERT(params.width * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 == 0); - glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.width)); + ASSERT(params.MipWidth(mip_map) * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 == + 0); + glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.MipWidth(mip_map))); + GLsizei image_size = static_cast<GLsizei>(params.GetMipmapSizeGL(mip_map, false)); glActiveTexture(GL_TEXTURE0); if (tuple.compressed) { switch (params.target) { case SurfaceParams::SurfaceTarget::Texture2D: - glCompressedTexImage2D( - SurfaceTargetToGL(params.target), 0, tuple.internal_format, - static_cast<GLsizei>(params.width), static_cast<GLsizei>(params.height), 0, - static_cast<GLsizei>(params.size_in_bytes_gl), &gl_buffer[buffer_offset]); + glCompressedTexImage2D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format, + static_cast<GLsizei>(params.MipWidth(mip_map)), + static_cast<GLsizei>(params.MipHeight(mip_map)), 0, image_size, + &gl_buffer[mip_map][buffer_offset]); break; case SurfaceParams::SurfaceTarget::Texture3D: + glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format, + static_cast<GLsizei>(params.MipWidth(mip_map)), + static_cast<GLsizei>(params.MipHeight(mip_map)), + static_cast<GLsizei>(params.MipDepth(mip_map)), 0, image_size, + &gl_buffer[mip_map][buffer_offset]); + break; case SurfaceParams::SurfaceTarget::Texture2DArray: - glCompressedTexImage3D( - SurfaceTargetToGL(params.target), 0, tuple.internal_format, - static_cast<GLsizei>(params.width), static_cast<GLsizei>(params.height), - static_cast<GLsizei>(params.depth), 0, - static_cast<GLsizei>(params.size_in_bytes_gl), &gl_buffer[buffer_offset]); + glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format, + static_cast<GLsizei>(params.MipWidth(mip_map)), + static_cast<GLsizei>(params.MipHeight(mip_map)), + static_cast<GLsizei>(params.depth), 0, image_size, + &gl_buffer[mip_map][buffer_offset]); break; - case SurfaceParams::SurfaceTarget::TextureCubemap: + case SurfaceParams::SurfaceTarget::TextureCubemap: { + GLsizei layer_size = static_cast<GLsizei>(params.LayerSizeGL(mip_map)); for (std::size_t face = 0; face < params.depth; ++face) { glCompressedTexImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), - 0, tuple.internal_format, static_cast<GLsizei>(params.width), - static_cast<GLsizei>(params.height), 0, - static_cast<GLsizei>(params.SizeInBytesCubeFaceGL()), - &gl_buffer[buffer_offset]); - buffer_offset += params.SizeInBytesCubeFace(); + mip_map, tuple.internal_format, + static_cast<GLsizei>(params.MipWidth(mip_map)), + static_cast<GLsizei>(params.MipHeight(mip_map)), 0, + layer_size, &gl_buffer[mip_map][buffer_offset]); + buffer_offset += layer_size; } break; + } default: LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}", static_cast<u32>(params.target)); UNREACHABLE(); - glCompressedTexImage2D( - GL_TEXTURE_2D, 0, tuple.internal_format, static_cast<GLsizei>(params.width), - static_cast<GLsizei>(params.height), 0, - static_cast<GLsizei>(params.size_in_bytes_gl), &gl_buffer[buffer_offset]); + glCompressedTexImage2D(GL_TEXTURE_2D, mip_map, tuple.internal_format, + static_cast<GLsizei>(params.MipWidth(mip_map)), + static_cast<GLsizei>(params.MipHeight(mip_map)), 0, + static_cast<GLsizei>(params.size_in_bytes_gl), + &gl_buffer[mip_map][buffer_offset]); } } else { switch (params.target) { case SurfaceParams::SurfaceTarget::Texture1D: - glTexSubImage1D(SurfaceTargetToGL(params.target), 0, x0, + glTexSubImage1D(SurfaceTargetToGL(params.target), mip_map, x0, static_cast<GLsizei>(rect.GetWidth()), tuple.format, tuple.type, - &gl_buffer[buffer_offset]); + &gl_buffer[mip_map][buffer_offset]); break; case SurfaceParams::SurfaceTarget::Texture2D: - glTexSubImage2D(SurfaceTargetToGL(params.target), 0, x0, y0, + glTexSubImage2D(SurfaceTargetToGL(params.target), mip_map, x0, y0, static_cast<GLsizei>(rect.GetWidth()), static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type, - &gl_buffer[buffer_offset]); + &gl_buffer[mip_map][buffer_offset]); break; case SurfaceParams::SurfaceTarget::Texture3D: + glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0, + static_cast<GLsizei>(rect.GetWidth()), + static_cast<GLsizei>(rect.GetHeight()), params.MipDepth(mip_map), + tuple.format, tuple.type, &gl_buffer[mip_map][buffer_offset]); + break; case SurfaceParams::SurfaceTarget::Texture2DArray: - glTexSubImage3D(SurfaceTargetToGL(params.target), 0, x0, y0, 0, + glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0, static_cast<GLsizei>(rect.GetWidth()), static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format, - tuple.type, &gl_buffer[buffer_offset]); + tuple.type, &gl_buffer[mip_map][buffer_offset]); break; - case SurfaceParams::SurfaceTarget::TextureCubemap: + case SurfaceParams::SurfaceTarget::TextureCubemap: { + std::size_t start = buffer_offset; for (std::size_t face = 0; face < params.depth; ++face) { - glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), 0, x0, - y0, static_cast<GLsizei>(rect.GetWidth()), + glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), mip_map, + x0, y0, static_cast<GLsizei>(rect.GetWidth()), static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type, - &gl_buffer[buffer_offset]); - buffer_offset += params.SizeInBytesCubeFace(); + &gl_buffer[mip_map][buffer_offset]); + buffer_offset += params.LayerSizeGL(mip_map); } break; + } default: LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}", static_cast<u32>(params.target)); UNREACHABLE(); - glTexSubImage2D(GL_TEXTURE_2D, 0, x0, y0, static_cast<GLsizei>(rect.GetWidth()), + glTexSubImage2D(GL_TEXTURE_2D, mip_map, x0, y0, static_cast<GLsizei>(rect.GetWidth()), static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type, - &gl_buffer[buffer_offset]); + &gl_buffer[mip_map][buffer_offset]); } } glPixelStorei(GL_UNPACK_ROW_LENGTH, 0); } +MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 64, 192)); +void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle) { + if (params.type == SurfaceType::Fill) + return; + + MICROPROFILE_SCOPE(OpenGL_TextureUL); + + for (u32 i = 0; i < params.max_mip_level; i++) + UploadGLMipmapTexture(i, read_fb_handle, draw_fb_handle); +} + RasterizerCacheOpenGL::RasterizerCacheOpenGL() { read_framebuffer.Create(); draw_framebuffer.Create(); diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h index b4701a616..e72f4f2d2 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h @@ -395,7 +395,7 @@ struct SurfaceParams { 64, // RG32UI 32, // R32UI 16, // ASTC_2D_8X8 - 32, // ASTC_2D_8X5 + 16, // ASTC_2D_8X5 32, // ASTC_2D_5X4 32, // BGRA8_SRGB 64, // DXT1_SRGB @@ -404,7 +404,7 @@ struct SurfaceParams { 128, // BC7U 32, // ASTC_2D_4X4_SRGB 16, // ASTC_2D_8X8_SRGB - 32, // ASTC_2D_8X5_SRGB + 16, // ASTC_2D_8X5_SRGB 32, // ASTC_2D_5X4_SRGB 32, // Z32F 16, // Z16 @@ -834,7 +834,7 @@ struct SurfaceParams { } /// Returns the rectangle corresponding to this surface - MathUtil::Rectangle<u32> GetRect() const; + MathUtil::Rectangle<u32> GetRect(u32 mip_level = 0) const; /// Returns the total size of this surface in bytes, adjusted for compression std::size_t SizeInBytesRaw(bool ignore_tiled = false) const { @@ -865,7 +865,7 @@ struct SurfaceParams { /// Returns the exact size of memory occupied by the texture in VRAM, including mipmaps. std::size_t MemorySize() const { - std::size_t size = InnerMemorySize(is_layered); + std::size_t size = InnerMemorySize(false, is_layered); if (is_layered) return size * depth; return size; @@ -874,12 +874,78 @@ struct SurfaceParams { /// Returns the exact size of the memory occupied by a layer in a texture in VRAM, including /// mipmaps. std::size_t LayerMemorySize() const { - return InnerMemorySize(true); + return InnerMemorySize(false, true); } /// Returns the size of a layer of this surface in OpenGL. - std::size_t LayerSizeGL() const { - return SizeInBytesRaw(true) / depth; + std::size_t LayerSizeGL(u32 mip_level) const { + return InnerMipmapMemorySize(mip_level, true, is_layered, false); + } + + std::size_t GetMipmapSizeGL(u32 mip_level, bool ignore_compressed = true) const { + std::size_t size = InnerMipmapMemorySize(mip_level, true, is_layered, ignore_compressed); + if (is_layered) + return size * depth; + return size; + } + + std::size_t GetMipmapLevelOffset(u32 mip_level) const { + std::size_t offset = 0; + for (u32 i = 0; i < mip_level; i++) + offset += InnerMipmapMemorySize(i, false, is_layered); + return offset; + } + + std::size_t GetMipmapLevelOffsetGL(u32 mip_level) const { + std::size_t offset = 0; + for (u32 i = 0; i < mip_level; i++) + offset += InnerMipmapMemorySize(i, true, is_layered); + return offset; + } + + u32 MipWidth(u32 mip_level) const { + return std::max(1U, width >> mip_level); + } + + u32 MipHeight(u32 mip_level) const { + return std::max(1U, height >> mip_level); + } + + u32 MipDepth(u32 mip_level) const { + return std::max(1U, depth >> mip_level); + } + + // Auto block resizing algorithm from: + // https://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/drivers/nouveau/nv50/nv50_miptree.c + u32 MipBlockHeight(u32 mip_level) const { + if (mip_level == 0) + return block_height; + u32 alt_height = MipHeight(mip_level); + u32 h = GetDefaultBlockHeight(pixel_format); + u32 blocks_in_y = (alt_height + h - 1) / h; + u32 bh = 16; + while (bh > 1 && blocks_in_y <= bh * 4) { + bh >>= 1; + } + return bh; + } + + u32 MipBlockDepth(u32 mip_level) const { + if (mip_level == 0) + return block_depth; + if (is_layered) + return 1; + u32 depth = MipDepth(mip_level); + u32 bd = 32; + while (bd > 1 && depth * 2 <= bd) { + bd >>= 1; + } + if (bd == 32) { + u32 bh = MipBlockHeight(mip_level); + if (bh >= 4) + return 16; + } + return bd; } /// Creates SurfaceParams from a texture configuration @@ -940,7 +1006,10 @@ struct SurfaceParams { } rt; private: - std::size_t InnerMemorySize(bool layer_only = false) const; + std::size_t InnerMipmapMemorySize(u32 mip_level, bool force_gl = false, bool layer_only = false, + bool uncompressed = false) const; + std::size_t InnerMemorySize(bool force_gl = false, bool layer_only = false, + bool uncompressed = false) const; }; }; // namespace OpenGL @@ -1002,8 +1071,10 @@ public: void UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle); private: + void UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, GLuint draw_fb_handle); + OGLTexture texture; - std::vector<u8> gl_buffer; + std::vector<std::vector<u8>> gl_buffer; SurfaceParams params; GLenum gl_target; std::size_t cached_size_in_bytes; diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 1a03a677f..9522fd344 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -8,6 +8,7 @@ #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_opengl/gl_shader_cache.h" #include "video_core/renderer_opengl/gl_shader_manager.h" +#include "video_core/renderer_opengl/utils.h" #include "video_core/utils.h" namespace OpenGL { @@ -89,7 +90,7 @@ CachedShader::CachedShader(VAddr addr, Maxwell::ShaderProgram program_type) shader.Create(program_result.first.c_str(), gl_type); program.Create(true, shader.handle); SetShaderUniformBlockBindings(program.handle); - VideoCore::LabelGLObject(GL_PROGRAM, program.handle, addr); + LabelGLObject(GL_PROGRAM, program.handle, addr); } else { // Store shader's code to lazily build it on draw geometry_programs.code = program_result.first; @@ -130,7 +131,7 @@ GLuint CachedShader::LazyGeometryProgram(OGLProgram& target_program, shader.Create(source.c_str(), GL_GEOMETRY_SHADER); target_program.Create(true, shader.handle); SetShaderUniformBlockBindings(target_program.handle); - VideoCore::LabelGLObject(GL_PROGRAM, target_program.handle, addr, debug_name); + LabelGLObject(GL_PROGRAM, target_program.handle, addr, debug_name); return target_program.handle; }; diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index dcf6941b0..d1f6ffe40 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -3,12 +3,12 @@ // Refer to the license.txt file included. #include <map> +#include <optional> #include <set> #include <string> #include <string_view> #include <unordered_set> -#include <boost/optional.hpp> #include <fmt/format.h> #include "common/assert.h" @@ -144,7 +144,7 @@ private: for (u32 offset = begin; offset != end && offset != PROGRAM_END; ++offset) { const Instruction instr = {program_code[offset]}; if (const auto opcode = OpCode::Decode(instr)) { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::EXIT: { // The EXIT instruction can be predicated, which means that the shader can // conditionally end on this instruction. We have to consider the case where the @@ -430,7 +430,7 @@ public: */ void SetRegisterToInputAttibute(const Register& reg, u64 elem, Attribute::Index attribute, const Tegra::Shader::IpaMode& input_mode, - boost::optional<Register> vertex = {}) { + std::optional<Register> vertex = {}) { const std::string dest = GetRegisterAsFloat(reg); const std::string src = GetInputAttribute(attribute, input_mode, vertex) + GetSwizzle(elem); shader.AddLine(dest + " = " + src + ';'); @@ -807,10 +807,10 @@ private: /// Generates code representing an input attribute register. std::string GetInputAttribute(Attribute::Index attribute, const Tegra::Shader::IpaMode& input_mode, - boost::optional<Register> vertex = {}) { + std::optional<Register> vertex = {}) { auto GeometryPass = [&](const std::string& name) { if (stage == Maxwell3D::Regs::ShaderStage::Geometry && vertex) { - return "gs_" + name + '[' + GetRegisterAsInteger(vertex.value(), 0, false) + ']'; + return "gs_" + name + '[' + GetRegisterAsInteger(*vertex, 0, false) + ']'; } return name; }; @@ -1465,7 +1465,7 @@ private: } shader.AddLine( - fmt::format("// {}: {} (0x{:016x})", offset, opcode->GetName(), instr.value)); + fmt::format("// {}: {} (0x{:016x})", offset, opcode->get().GetName(), instr.value)); using Tegra::Shader::Pred; ASSERT_MSG(instr.pred.full_pred != Pred::NeverExecute, @@ -1473,7 +1473,7 @@ private: // Some instructions (like SSY) don't have a predicate field, they are always // unconditionally executed. - bool can_be_predicated = OpCode::IsPredicatedInstruction(opcode->GetId()); + bool can_be_predicated = OpCode::IsPredicatedInstruction(opcode->get().GetId()); if (can_be_predicated && instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex)) { shader.AddLine("if (" + @@ -1483,7 +1483,7 @@ private: ++shader.scope; } - switch (opcode->GetType()) { + switch (opcode->get().GetType()) { case OpCode::Type::Arithmetic: { std::string op_a = regs.GetRegisterAsFloat(instr.gpr8); @@ -1500,7 +1500,7 @@ private: } } - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::MOV_C: case OpCode::Id::MOV_R: { // MOV does not have neither 'abs' nor 'neg' bits. @@ -1600,14 +1600,15 @@ private: break; } default: { - LOG_CRITICAL(HW_GPU, "Unhandled arithmetic instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled arithmetic instruction: {}", + opcode->get().GetName()); UNREACHABLE(); } } break; } case OpCode::Type::ArithmeticImmediate: { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::MOV32_IMM: { regs.SetRegisterToFloat(instr.gpr0, 0, GetImmediate32(instr), 1, 1); break; @@ -1651,7 +1652,7 @@ private: std::string op_a = instr.bfe.negate_a ? "-" : ""; op_a += regs.GetRegisterAsInteger(instr.gpr8); - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::BFE_IMM: { std::string inner_shift = '(' + op_a + " << " + std::to_string(instr.bfe.GetLeftShiftValue()) + ')'; @@ -1663,7 +1664,7 @@ private: break; } default: { - LOG_CRITICAL(HW_GPU, "Unhandled BFE instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled BFE instruction: {}", opcode->get().GetName()); UNREACHABLE(); } } @@ -1685,7 +1686,7 @@ private: } } - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::SHR_C: case OpCode::Id::SHR_R: case OpCode::Id::SHR_IMM: { @@ -1705,7 +1706,7 @@ private: regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " << " + op_b, 1, 1); break; default: { - LOG_CRITICAL(HW_GPU, "Unhandled shift instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled shift instruction: {}", opcode->get().GetName()); UNREACHABLE(); } } @@ -1715,7 +1716,7 @@ private: std::string op_a = regs.GetRegisterAsInteger(instr.gpr8); std::string op_b = std::to_string(instr.alu.imm20_32.Value()); - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::IADD32I: if (instr.iadd32i.negate_a) op_a = "-(" + op_a + ')'; @@ -1737,7 +1738,7 @@ private: } default: { LOG_CRITICAL(HW_GPU, "Unhandled ArithmeticIntegerImmediate instruction: {}", - opcode->GetName()); + opcode->get().GetName()); UNREACHABLE(); } } @@ -1757,7 +1758,7 @@ private: } } - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::IADD_C: case OpCode::Id::IADD_R: case OpCode::Id::IADD_IMM: { @@ -1793,7 +1794,7 @@ private: } }; - if (opcode->GetId() == OpCode::Id::IADD3_R) { + if (opcode->get().GetId() == OpCode::Id::IADD3_R) { apply_height(instr.iadd3.height_a, op_a); apply_height(instr.iadd3.height_b, op_b); apply_height(instr.iadd3.height_c, op_c); @@ -1809,7 +1810,7 @@ private: op_c = "-(" + op_c + ')'; std::string result; - if (opcode->GetId() == OpCode::Id::IADD3_R) { + if (opcode->get().GetId() == OpCode::Id::IADD3_R) { switch (instr.iadd3.mode) { case Tegra::Shader::IAdd3Mode::RightShift: // TODO(tech4me): According to @@ -1884,7 +1885,7 @@ private: const std::string op_c = regs.GetRegisterAsInteger(instr.gpr39); std::string lut; - if (opcode->GetId() == OpCode::Id::LOP3_R) { + if (opcode->get().GetId() == OpCode::Id::LOP3_R) { lut = '(' + std::to_string(instr.alu.lop3.GetImmLut28()) + ')'; } else { lut = '(' + std::to_string(instr.alu.lop3.GetImmLut48()) + ')'; @@ -1914,7 +1915,7 @@ private: case OpCode::Id::LEA_HI: { std::string op_c; - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::LEA_R2: { op_a = regs.GetRegisterAsInteger(instr.gpr20); op_b = regs.GetRegisterAsInteger(instr.gpr39); @@ -1959,7 +1960,8 @@ private: op_b = regs.GetRegisterAsInteger(instr.gpr8); op_a = std::to_string(instr.lea.imm.entry_a); op_c = std::to_string(instr.lea.imm.entry_b); - LOG_CRITICAL(HW_GPU, "Unhandled LEA subinstruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled LEA subinstruction: {}", + opcode->get().GetName()); UNREACHABLE(); } } @@ -1974,7 +1976,7 @@ private: } default: { LOG_CRITICAL(HW_GPU, "Unhandled ArithmeticInteger instruction: {}", - opcode->GetName()); + opcode->get().GetName()); UNREACHABLE(); } } @@ -1982,20 +1984,21 @@ private: break; } case OpCode::Type::ArithmeticHalf: { - if (opcode->GetId() == OpCode::Id::HADD2_C || opcode->GetId() == OpCode::Id::HADD2_R) { + if (opcode->get().GetId() == OpCode::Id::HADD2_C || + opcode->get().GetId() == OpCode::Id::HADD2_R) { ASSERT_MSG(instr.alu_half.ftz == 0, "Unimplemented"); } const bool negate_a = - opcode->GetId() != OpCode::Id::HMUL2_R && instr.alu_half.negate_a != 0; + opcode->get().GetId() != OpCode::Id::HMUL2_R && instr.alu_half.negate_a != 0; const bool negate_b = - opcode->GetId() != OpCode::Id::HMUL2_C && instr.alu_half.negate_b != 0; + opcode->get().GetId() != OpCode::Id::HMUL2_C && instr.alu_half.negate_b != 0; const std::string op_a = GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr8, 0, false), instr.alu_half.type_a, instr.alu_half.abs_a != 0, negate_a); std::string op_b; - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::HADD2_C: case OpCode::Id::HMUL2_C: op_b = regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset, @@ -2013,7 +2016,7 @@ private: op_b = GetHalfFloat(op_b, instr.alu_half.type_b, instr.alu_half.abs_b != 0, negate_b); const std::string result = [&]() { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::HADD2_C: case OpCode::Id::HADD2_R: return '(' + op_a + " + " + op_b + ')'; @@ -2021,7 +2024,8 @@ private: case OpCode::Id::HMUL2_R: return '(' + op_a + " * " + op_b + ')'; default: - LOG_CRITICAL(HW_GPU, "Unhandled half float instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled half float instruction: {}", + opcode->get().GetName()); UNREACHABLE(); return std::string("0"); } @@ -2032,7 +2036,7 @@ private: break; } case OpCode::Type::ArithmeticHalfImmediate: { - if (opcode->GetId() == OpCode::Id::HADD2_IMM) { + if (opcode->get().GetId() == OpCode::Id::HADD2_IMM) { ASSERT_MSG(instr.alu_half_imm.ftz == 0, "Unimplemented"); } else { ASSERT_MSG(instr.alu_half_imm.precision == Tegra::Shader::HalfPrecision::None, @@ -2046,7 +2050,7 @@ private: const std::string op_b = UnpackHalfImmediate(instr, true); const std::string result = [&]() { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::HADD2_IMM: return op_a + " + " + op_b; case OpCode::Id::HMUL2_IMM: @@ -2072,7 +2076,7 @@ private: ASSERT_MSG(instr.ffma.tab5980_1 == 0, "FFMA tab5980_1({}) not implemented", instr.ffma.tab5980_1.Value()); - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::FFMA_CR: { op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset, GLSLRegister::Type::Float); @@ -2096,7 +2100,7 @@ private: break; } default: { - LOG_CRITICAL(HW_GPU, "Unhandled FFMA instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled FFMA instruction: {}", opcode->get().GetName()); UNREACHABLE(); } } @@ -2107,14 +2111,14 @@ private: break; } case OpCode::Type::Hfma2: { - if (opcode->GetId() == OpCode::Id::HFMA2_RR) { + if (opcode->get().GetId() == OpCode::Id::HFMA2_RR) { ASSERT_MSG(instr.hfma2.rr.precision == Tegra::Shader::HalfPrecision::None, "Unimplemented"); } else { ASSERT_MSG(instr.hfma2.precision == Tegra::Shader::HalfPrecision::None, "Unimplemented"); } - const bool saturate = opcode->GetId() == OpCode::Id::HFMA2_RR + const bool saturate = opcode->get().GetId() == OpCode::Id::HFMA2_RR ? instr.hfma2.rr.saturate != 0 : instr.hfma2.saturate != 0; @@ -2122,7 +2126,7 @@ private: GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr8, 0, false), instr.hfma2.type_a); std::string op_b, op_c; - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::HFMA2_CR: op_b = GetHalfFloat(regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset, GLSLRegister::Type::UnsignedInteger), @@ -2160,7 +2164,7 @@ private: break; } case OpCode::Type::Conversion: { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::I2I_R: { ASSERT_MSG(!instr.conversion.selector, "Unimplemented"); @@ -2298,14 +2302,15 @@ private: break; } default: { - LOG_CRITICAL(HW_GPU, "Unhandled conversion instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled conversion instruction: {}", + opcode->get().GetName()); UNREACHABLE(); } } break; } case OpCode::Type::Memory: { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::LD_A: { // Note: Shouldn't this be interp mode flat? As in no interpolation made. ASSERT_MSG(instr.gpr8.Value() == Register::ZeroIndex, @@ -2949,7 +2954,7 @@ private: break; } default: { - LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->get().GetName()); UNREACHABLE(); } } @@ -3043,7 +3048,7 @@ private: instr.hsetp2.abs_a, instr.hsetp2.negate_a); const std::string op_b = [&]() { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::HSETP2_R: return GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr20, 0, false), instr.hsetp2.type_b, instr.hsetp2.abs_a, @@ -3105,7 +3110,7 @@ private: break; } case OpCode::Type::PredicateSetPredicate: { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::PSETP: { const std::string op_a = GetPredicateCondition(instr.psetp.pred12, instr.psetp.neg_pred12 != 0); @@ -3151,7 +3156,8 @@ private: break; } default: { - LOG_CRITICAL(HW_GPU, "Unhandled predicate instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled predicate instruction: {}", + opcode->get().GetName()); UNREACHABLE(); } } @@ -3239,7 +3245,7 @@ private: instr.hset2.abs_a != 0, instr.hset2.negate_a != 0); const std::string op_b = [&]() { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::HSET2_R: return GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr20, 0, false), instr.hset2.type_b, instr.hset2.abs_b != 0, @@ -3288,7 +3294,7 @@ private: const bool is_signed{instr.xmad.sign_a == 1}; bool is_merge{}; - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::XMAD_CR: { is_merge = instr.xmad.merge_56; op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset, @@ -3317,7 +3323,7 @@ private: break; } default: { - LOG_CRITICAL(HW_GPU, "Unhandled XMAD instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled XMAD instruction: {}", opcode->get().GetName()); UNREACHABLE(); } } @@ -3369,7 +3375,7 @@ private: break; } default: { - switch (opcode->GetId()) { + switch (opcode->get().GetId()) { case OpCode::Id::EXIT: { if (stage == Maxwell3D::Regs::ShaderStage::Fragment) { EmitFragmentOutputsWrite(); @@ -3564,7 +3570,7 @@ private: break; } default: { - LOG_CRITICAL(HW_GPU, "Unhandled instruction: {}", opcode->GetName()); + LOG_CRITICAL(HW_GPU, "Unhandled instruction: {}", opcode->get().GetName()); UNREACHABLE(); } } @@ -3705,9 +3711,9 @@ std::string GetCommonDeclarations() { RasterizerOpenGL::MaxConstbufferSize / sizeof(GLvec4)); } -boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset, - Maxwell3D::Regs::ShaderStage stage, - const std::string& suffix) { +std::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset, + Maxwell3D::Regs::ShaderStage stage, + const std::string& suffix) { try { const auto subroutines = ControlFlowAnalyzer(program_code, main_offset, suffix).GetSubroutines(); @@ -3716,7 +3722,7 @@ boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, } catch (const DecompileFail& exception) { LOG_ERROR(HW_GPU, "Shader decompilation failed: {}", exception.what()); } - return boost::none; + return {}; } } // namespace OpenGL::GLShader::Decompiler diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.h b/src/video_core/renderer_opengl/gl_shader_decompiler.h index b20cc4bfa..d01a4a7ee 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.h +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.h @@ -6,8 +6,8 @@ #include <array> #include <functional> +#include <optional> #include <string> -#include <boost/optional.hpp> #include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_opengl/gl_shader_gen.h" @@ -18,8 +18,8 @@ using Tegra::Engines::Maxwell3D; std::string GetCommonDeclarations(); -boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset, - Maxwell3D::Regs::ShaderStage stage, - const std::string& suffix); +std::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset, + Maxwell3D::Regs::ShaderStage stage, + const std::string& suffix); } // namespace OpenGL::GLShader::Decompiler diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp index dfb562706..9d17edd63 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.cpp +++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp @@ -37,7 +37,7 @@ layout(std140) uniform vs_config { ProgramResult program = Decompiler::DecompileProgram(setup.program.code, PROGRAM_OFFSET, Maxwell3D::Regs::ShaderStage::Vertex, "vertex") - .get_value_or({}); + .value_or(ProgramResult()); out += program.first; @@ -45,7 +45,7 @@ layout(std140) uniform vs_config { ProgramResult program_b = Decompiler::DecompileProgram(setup.program.code_b, PROGRAM_OFFSET, Maxwell3D::Regs::ShaderStage::Vertex, "vertex_b") - .get_value_or({}); + .value_or(ProgramResult()); out += program_b.first; } @@ -90,7 +90,7 @@ ProgramResult GenerateGeometryShader(const ShaderSetup& setup) { ProgramResult program = Decompiler::DecompileProgram(setup.program.code, PROGRAM_OFFSET, Maxwell3D::Regs::ShaderStage::Geometry, "geometry") - .get_value_or({}); + .value_or(ProgramResult()); out += R"( out gl_PerVertex { vec4 gl_Position; @@ -124,7 +124,7 @@ ProgramResult GenerateFragmentShader(const ShaderSetup& setup) { ProgramResult program = Decompiler::DecompileProgram(setup.program.code, PROGRAM_OFFSET, Maxwell3D::Regs::ShaderStage::Fragment, "fragment") - .get_value_or({}); + .value_or(ProgramResult()); out += R"( layout(location = 0) out vec4 FragColor0; layout(location = 1) out vec4 FragColor1; diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index 0f6dcab2b..87d511c38 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -135,12 +135,29 @@ inline GLenum PrimitiveTopology(Maxwell::PrimitiveTopology topology) { return {}; } -inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode) { +inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode, + Tegra::Texture::TextureMipmapFilter mip_filter_mode) { switch (filter_mode) { - case Tegra::Texture::TextureFilter::Linear: - return GL_LINEAR; - case Tegra::Texture::TextureFilter::Nearest: - return GL_NEAREST; + case Tegra::Texture::TextureFilter::Linear: { + switch (mip_filter_mode) { + case Tegra::Texture::TextureMipmapFilter::None: + return GL_LINEAR; + case Tegra::Texture::TextureMipmapFilter::Nearest: + return GL_NEAREST_MIPMAP_LINEAR; + case Tegra::Texture::TextureMipmapFilter::Linear: + return GL_LINEAR_MIPMAP_LINEAR; + } + } + case Tegra::Texture::TextureFilter::Nearest: { + switch (mip_filter_mode) { + case Tegra::Texture::TextureMipmapFilter::None: + return GL_NEAREST; + case Tegra::Texture::TextureMipmapFilter::Nearest: + return GL_NEAREST_MIPMAP_NEAREST; + case Tegra::Texture::TextureMipmapFilter::Linear: + return GL_LINEAR_MIPMAP_NEAREST; + } + } } LOG_CRITICAL(Render_OpenGL, "Unimplemented texture filter mode={}", static_cast<u32>(filter_mode)); diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp index 90b68943d..ea38da932 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.cpp +++ b/src/video_core/renderer_opengl/renderer_opengl.cpp @@ -115,7 +115,8 @@ RendererOpenGL::RendererOpenGL(Core::Frontend::EmuWindow& window) RendererOpenGL::~RendererOpenGL() = default; /// Swap buffers (render frame) -void RendererOpenGL::SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) { +void RendererOpenGL::SwapBuffers( + std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) { ScopeAcquireGLContext acquire_context{render_window}; Core::System::GetInstance().GetPerfStats().EndSystemFrame(); @@ -124,11 +125,11 @@ void RendererOpenGL::SwapBuffers(boost::optional<const Tegra::FramebufferConfig& OpenGLState prev_state = OpenGLState::GetCurState(); state.Apply(); - if (framebuffer != boost::none) { + if (framebuffer) { // If framebuffer is provided, reload it from memory to a texture - if (screen_info.texture.width != (GLsizei)framebuffer->width || - screen_info.texture.height != (GLsizei)framebuffer->height || - screen_info.texture.pixel_format != framebuffer->pixel_format) { + if (screen_info.texture.width != (GLsizei)framebuffer->get().width || + screen_info.texture.height != (GLsizei)framebuffer->get().height || + screen_info.texture.pixel_format != framebuffer->get().pixel_format) { // Reallocate texture if the framebuffer size has changed. // This is expected to not happen very often and hence should not be a // performance problem. diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h index 961467a62..c0868c0e4 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.h +++ b/src/video_core/renderer_opengl/renderer_opengl.h @@ -51,7 +51,8 @@ public: ~RendererOpenGL() override; /// Swap buffers (render frame) - void SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) override; + void SwapBuffers( + std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) override; /// Initialize the renderer bool Init() override; diff --git a/src/video_core/renderer_opengl/utils.cpp b/src/video_core/renderer_opengl/utils.cpp new file mode 100644 index 000000000..d84634cb3 --- /dev/null +++ b/src/video_core/renderer_opengl/utils.cpp @@ -0,0 +1,38 @@ +// Copyright 2014 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <string> +#include <fmt/format.h> +#include <glad/glad.h> +#include "common/common_types.h" +#include "video_core/renderer_opengl/utils.h" + +namespace OpenGL { + +void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string extra_info) { + if (!GLAD_GL_KHR_debug) { + return; // We don't need to throw an error as this is just for debugging + } + const std::string nice_addr = fmt::format("0x{:016x}", addr); + std::string object_label; + + if (extra_info.empty()) { + switch (identifier) { + case GL_TEXTURE: + object_label = "Texture@" + nice_addr; + break; + case GL_PROGRAM: + object_label = "Shader@" + nice_addr; + break; + default: + object_label = fmt::format("Object(0x{:x})@{}", identifier, nice_addr); + break; + } + } else { + object_label = extra_info + '@' + nice_addr; + } + glObjectLabel(identifier, handle, -1, static_cast<const GLchar*>(object_label.c_str())); +} + +} // namespace OpenGL
\ No newline at end of file diff --git a/src/video_core/renderer_opengl/utils.h b/src/video_core/renderer_opengl/utils.h new file mode 100644 index 000000000..1fcb6fc11 --- /dev/null +++ b/src/video_core/renderer_opengl/utils.h @@ -0,0 +1,15 @@ +// Copyright 2014 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <string> +#include <glad/glad.h> +#include "common/common_types.h" + +namespace OpenGL { + +void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string extra_info = ""); + +} // namespace OpenGL
\ No newline at end of file diff --git a/src/video_core/textures/decoders.h b/src/video_core/textures/decoders.h index 4726f54a5..b390219e4 100644 --- a/src/video_core/textures/decoders.h +++ b/src/video_core/textures/decoders.h @@ -10,6 +10,12 @@ namespace Tegra::Texture { +// GOBSize constant. Calculated by 64 bytes in x multiplied by 8 y coords, represents +// an small rect of (64/bytes_per_pixel)X8. +inline std::size_t GetGOBSize() { + return 512; +} + /** * Unswizzles a swizzled texture without changing its format. */ diff --git a/src/video_core/utils.h b/src/video_core/utils.h index 237cc1307..e0a14d48f 100644 --- a/src/video_core/utils.h +++ b/src/video_core/utils.h @@ -161,30 +161,4 @@ static inline void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixe } } -static void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, - std::string extra_info = "") { - if (!GLAD_GL_KHR_debug) { - return; // We don't need to throw an error as this is just for debugging - } - const std::string nice_addr = fmt::format("0x{:016x}", addr); - std::string object_label; - - if (extra_info.empty()) { - switch (identifier) { - case GL_TEXTURE: - object_label = "Texture@" + nice_addr; - break; - case GL_PROGRAM: - object_label = "Shader@" + nice_addr; - break; - default: - object_label = fmt::format("Object(0x{:x})@{}", identifier, nice_addr); - break; - } - } else { - object_label = extra_info + '@' + nice_addr; - } - glObjectLabel(identifier, handle, -1, static_cast<const GLchar*>(object_label.c_str())); -} - } // namespace VideoCore diff --git a/src/yuzu/configuration/configure_input.cpp b/src/yuzu/configuration/configure_input.cpp index 94789c064..42a7beac6 100644 --- a/src/yuzu/configuration/configure_input.cpp +++ b/src/yuzu/configuration/configure_input.cpp @@ -322,7 +322,7 @@ void ConfigureInput::setPollingResult(const Common::ParamPackage& params, bool a } updateButtonLabels(); - input_setter = boost::none; + input_setter = {}; } void ConfigureInput::keyPressEvent(QKeyEvent* event) { diff --git a/src/yuzu/configuration/configure_input.h b/src/yuzu/configuration/configure_input.h index d1198db81..32c7183f9 100644 --- a/src/yuzu/configuration/configure_input.h +++ b/src/yuzu/configuration/configure_input.h @@ -7,11 +7,13 @@ #include <array> #include <functional> #include <memory> +#include <optional> #include <string> #include <unordered_map> + #include <QKeyEvent> #include <QWidget> -#include <boost/optional.hpp> + #include "common/param_package.h" #include "core/settings.h" #include "input_common/main.h" @@ -41,7 +43,7 @@ private: std::unique_ptr<QTimer> poll_timer; /// This will be the the setting function when an input is awaiting configuration. - boost::optional<std::function<void(const Common::ParamPackage&)>> input_setter; + std::optional<std::function<void(const Common::ParamPackage&)>> input_setter; std::array<Common::ParamPackage, Settings::NativeButton::NumButtons> buttons_param; std::array<Common::ParamPackage, Settings::NativeAnalog::NumAnalogs> analogs_param; diff --git a/src/yuzu/configuration/configure_system.cpp b/src/yuzu/configuration/configure_system.cpp index acb458001..1b8aa7de2 100644 --- a/src/yuzu/configuration/configure_system.cpp +++ b/src/yuzu/configuration/configure_system.cpp @@ -173,7 +173,7 @@ void ConfigureSystem::UpdateCurrentUser() { ui->pm_add->setEnabled(profile_manager->GetUserCount() < Service::Account::MAX_USERS); const auto& current_user = profile_manager->GetUser(Settings::values.current_user); - ASSERT(current_user != std::nullopt); + ASSERT(current_user); const auto username = GetAccountUsername(*profile_manager, *current_user); scene->clear(); @@ -261,7 +261,7 @@ void ConfigureSystem::AddUser() { void ConfigureSystem::RenameUser() { const auto user = tree_view->currentIndex().row(); const auto uuid = profile_manager->GetUser(user); - ASSERT(uuid != std::nullopt); + ASSERT(uuid); Service::Account::ProfileBase profile; if (!profile_manager->GetProfileBase(*uuid, profile)) @@ -297,7 +297,7 @@ void ConfigureSystem::RenameUser() { void ConfigureSystem::DeleteUser() { const auto index = tree_view->currentIndex().row(); const auto uuid = profile_manager->GetUser(index); - ASSERT(uuid != std::nullopt); + ASSERT(uuid); const auto username = GetAccountUsername(*profile_manager, *uuid); const auto confirm = QMessageBox::question( @@ -324,7 +324,7 @@ void ConfigureSystem::DeleteUser() { void ConfigureSystem::SetUserImage() { const auto index = tree_view->currentIndex().row(); const auto uuid = profile_manager->GetUser(index); - ASSERT(uuid != std::nullopt); + ASSERT(uuid); const auto file = QFileDialog::getOpenFileName(this, tr("Select User Image"), QString(), tr("JPEG Images (*.jpg *.jpeg)")); diff --git a/src/yuzu/debugger/graphics/graphics_surface.cpp b/src/yuzu/debugger/graphics/graphics_surface.cpp index 44d423da2..0adbab27d 100644 --- a/src/yuzu/debugger/graphics/graphics_surface.cpp +++ b/src/yuzu/debugger/graphics/graphics_surface.cpp @@ -382,7 +382,7 @@ void GraphicsSurfaceWidget::OnUpdate() { // TODO: Implement a good way to visualize alpha components! QImage decoded_image(surface_width, surface_height, QImage::Format_ARGB32); - boost::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address); + std::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address); // 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. @@ -444,7 +444,7 @@ void GraphicsSurfaceWidget::SaveSurface() { pixmap->save(&file, "PNG"); } else if (selectedFilter == bin_filter) { auto& gpu = Core::System::GetInstance().GPU(); - boost::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address); + std::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address); const u8* buffer = Memory::GetPointer(*address); ASSERT_MSG(buffer != nullptr, "Memory not accessible"); diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index b5bfa6741..c5a56cbfd 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -786,7 +786,7 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target ASSERT(index != -1 && index < 8); const auto user_id = manager.GetUser(index); - ASSERT(user_id != std::nullopt); + ASSERT(user_id); path = nand_dir + FileSys::SaveDataFactory::GetFullPath(FileSys::SaveDataSpaceId::NandUser, FileSys::SaveDataType::SaveData, program_id, user_id->uuid, 0); @@ -1560,7 +1560,7 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) { } } -boost::optional<u64> GMainWindow::SelectRomFSDumpTarget( +std::optional<u64> GMainWindow::SelectRomFSDumpTarget( const FileSys::RegisteredCacheUnion& installed, u64 program_id) { const auto dlc_entries = installed.ListEntriesFilter(FileSys::TitleType::AOC, FileSys::ContentRecordType::Data); @@ -1587,7 +1587,7 @@ boost::optional<u64> GMainWindow::SelectRomFSDumpTarget( this, tr("Select RomFS Dump Target"), tr("Please select which RomFS you would like to dump."), list, 0, false, &ok); if (!ok) { - return boost::none; + return {}; } return romfs_tids[list.indexOf(res)]; diff --git a/src/yuzu/main.h b/src/yuzu/main.h index 7c7c223e1..af637d89e 100644 --- a/src/yuzu/main.h +++ b/src/yuzu/main.h @@ -5,12 +5,12 @@ #pragma once #include <memory> +#include <optional> #include <unordered_map> #include <QMainWindow> #include <QTimer> -#include <boost/optional.hpp> #include "common/common_types.h" #include "core/core.h" #include "ui_main.h" @@ -178,8 +178,7 @@ private slots: void OnReinitializeKeys(ReinitializeKeyBehavior behavior); private: - boost::optional<u64> SelectRomFSDumpTarget(const FileSys::RegisteredCacheUnion&, - u64 program_id); + std::optional<u64> SelectRomFSDumpTarget(const FileSys::RegisteredCacheUnion&, u64 program_id); void UpdateStatusBar(); Ui::MainWindow ui; |