Compare commits

..

6 Commits

Author SHA1 Message Date
Zach Hilman
bdaa76c0db ns: Implement command 400: GetApplicationControlData
Returns the raw NACP bytes and the raw icon bytes into a title-provided buffer. Pulls from Registration Cache for control data, returning all zeros should it not exist.
2018-10-29 16:20:16 -04:00
Zach Hilman
5ee19add1b fsp_srv: Implement ISaveDataInfoReader
An object to read SaveDataInfo objects, which describe a unique save on the system. This implementation iterates through all the directories in the save data space and uses the paths to reconstruct the metadata.
2018-10-29 13:54:39 -04:00
Zach Hilman
2e8177f0c9 fsp_srv: Implement command 61: OpenSaveDataInfoReaderBySaveDataSpaceId
Needed by Checkpoint. Returns an object that can iterate through all savedata on the system.
2018-10-29 13:54:39 -04:00
Zach Hilman
df264d2ccb savedata_factory: Expose accessors for SaveDataSpace 2018-10-29 13:54:38 -04:00
Zach Hilman
f2f679bf3f loader/nro: Call RegisterRomFS from Load
Allows NRO homebrew to use the RomFS in the ASET section.
2018-10-29 13:54:38 -04:00
Zach Hilman
b04e39107f control_metadata: Add GetRawBytes function to NACP
Returns the raw bytes of the NACP file. Needed for GetApplicationControlData which returns the raw, unprocessed NACP to the game.
2018-10-29 13:54:38 -04:00
110 changed files with 2062 additions and 2900 deletions

View File

@@ -11,6 +11,7 @@
#include "audio_core/stream.h"
#include "common/assert.h"
#include "common/logging/log.h"
#include "common/microprofile.h"
#include "core/core_timing.h"
#include "core/core_timing_util.h"
#include "core/settings.h"
@@ -103,7 +104,10 @@ void Stream::PlayNextBuffer() {
CoreTiming::ScheduleEventThreadsafe(GetBufferReleaseCycles(*active_buffer), release_event, {});
}
MICROPROFILE_DEFINE(AudioOutput, "Audio", "ReleaseActiveBuffer", MP_RGB(100, 100, 255));
void Stream::ReleaseActiveBuffer() {
MICROPROFILE_SCOPE(AudioOutput);
ASSERT(active_buffer);
released_buffers.push(std::move(active_buffer));
release_callback();

View File

@@ -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 = 0.25; // seconds
const double max_latency = 1.0; // seconds
const double max_backlog = m_sample_rate * max_latency;
const double backlog_fullness = m_sound_touch.numSamples() / max_backlog;
if (backlog_fullness > 4.0) {
if (backlog_fullness > 5.0) {
// 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 = 0.712; // seconds
constexpr double lpf_time_scale = 2.0; // seconds
const double lpf_gain = 1.0 - std::exp(-time_delta / lpf_time_scale);
m_stretch_ratio += lpf_gain * (current_ratio - m_stretch_ratio);

View File

@@ -12,8 +12,7 @@
#include <thread>
#include <vector>
#ifdef _WIN32
#include <share.h> // For _SH_DENYWR
#include <windows.h> // For OutputDebugStringA
#include <share.h> // For _SH_DENYWR
#else
#define _SH_DENYWR 0
#endif
@@ -140,18 +139,12 @@ void FileBackend::Write(const Entry& entry) {
if (!file.IsOpen() || bytes_written > MAX_BYTES_WRITTEN) {
return;
}
bytes_written += file.WriteString(FormatLogMessage(entry).append(1, '\n'));
bytes_written += file.WriteString(FormatLogMessage(entry) + '\n');
if (entry.log_level >= Level::Error) {
file.Flush();
}
}
void DebuggerBackend::Write(const Entry& entry) {
#ifdef _WIN32
::OutputDebugStringA(FormatLogMessage(entry).append(1, '\n').c_str());
#endif
}
/// Macro listing all log classes. Code should define CLS and SUB as desired before invoking this.
#define ALL_LOG_CLASSES() \
CLS(Log) \

View File

@@ -103,20 +103,6 @@ private:
std::size_t bytes_written;
};
/**
* Backend that writes to Visual Studio's output window
*/
class DebuggerBackend : public Backend {
public:
static const char* Name() {
return "debugger";
}
const char* GetName() const override {
return Name();
}
void Write(const Entry& entry) override;
};
void AddBackend(std::unique_ptr<Backend> backend);
void RemoveBackend(std::string_view backend_name);

View File

@@ -153,7 +153,6 @@ struct VisitorInterface : NonCopyable {
/// Completion method, called once all fields have been visited
virtual void Complete() = 0;
virtual bool SubmitTestcase() = 0;
};
/**
@@ -179,9 +178,6 @@ struct NullVisitor : public VisitorInterface {
void Visit(const Field<std::chrono::microseconds>& /*field*/) override {}
void Complete() override {}
bool SubmitTestcase() override {
return false;
}
};
/// Appends build-specific information to the given FieldCollection,

View File

@@ -185,7 +185,7 @@ struct System::Impl {
LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath);
return ResultStatus::ErrorGetLoader;
}
std::pair<std::optional<u32>, Loader::ResultStatus> system_mode =
std::pair<boost::optional<u32>, Loader::ResultStatus> system_mode =
app_loader->LoadKernelSystemMode();
if (system_mode.second != Loader::ResultStatus::Success) {
@@ -312,10 +312,6 @@ 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);
}
@@ -346,11 +342,7 @@ PerfStatsResults System::GetAndResetPerfStats() {
return impl->GetAndResetPerfStats();
}
TelemetrySession& System::TelemetrySession() {
return *impl->telemetry_session;
}
const TelemetrySession& System::TelemetrySession() const {
Core::TelemetrySession& System::TelemetrySession() const {
return *impl->telemetry_session;
}
@@ -358,11 +350,7 @@ ARM_Interface& System::CurrentArmInterface() {
return CurrentCpuCore().ArmInterface();
}
const ARM_Interface& System::CurrentArmInterface() const {
return CurrentCpuCore().ArmInterface();
}
std::size_t System::CurrentCoreIndex() const {
std::size_t System::CurrentCoreIndex() {
return CurrentCpuCore().CoreIndex();
}
@@ -370,10 +358,6 @@ 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();
}
@@ -394,10 +378,6 @@ 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];
@@ -412,10 +392,6 @@ 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;
}

View File

@@ -129,11 +129,11 @@ public:
*/
bool IsPoweredOn() 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;
/**
* Returns a reference to the telemetry session for this emulation session.
* @returns Reference to the telemetry session.
*/
Core::TelemetrySession& TelemetrySession() const;
/// Prepare the core emulation for a reschedule
void PrepareReschedule();
@@ -144,36 +144,24 @@ 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() const;
std::size_t CurrentCoreIndex();
/// Gets the scheduler for the CPU core that is currently running
Kernel::Scheduler& CurrentScheduler();
/// 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
/// Gets an ARM interface to 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 a reference to the exclusive monitor
/// Gets 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();
@@ -242,9 +230,6 @@ 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

View File

@@ -141,28 +141,28 @@ Key128 DeriveKeyblobMACKey(const Key128& keyblob_key, const Key128& mac_source)
return mac_key;
}
std::optional<Key128> DeriveSDSeed() {
boost::optional<Key128> DeriveSDSeed() {
const FileUtil::IOFile save_43(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
"/system/save/8000000000000043",
"rb+");
if (!save_43.IsOpen())
return {};
return boost::none;
const FileUtil::IOFile sd_private(
FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir) + "/Nintendo/Contents/private", "rb+");
if (!sd_private.IsOpen())
return {};
return boost::none;
std::array<u8, 0x10> private_seed{};
if (sd_private.ReadBytes(private_seed.data(), private_seed.size()) != private_seed.size()) {
return {};
return boost::none;
}
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 {};
return boost::none;
}
save_43.ReadBytes(buffer.data(), buffer.size());
@@ -172,12 +172,12 @@ std::optional<Key128> DeriveSDSeed() {
}
if (!save_43.Seek(offset + 0x10, SEEK_SET)) {
return {};
return boost::none;
}
Key128 seed{};
if (save_43.ReadBytes(seed.data(), seed.size()) != seed.size()) {
return {};
return boost::none;
}
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 std::optional<u64> FindTicketOffset(const std::array<u8, size>& data) {
static boost::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 {};
return boost::none;
}
}
return offset;
}
std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
const RSAKeyPair<2048>& key) {
boost::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 {};
return boost::none;
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 @@ std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
std::memcpy(rights_id.data(), ticket.data() + 0x2A0, sizeof(Key128));
if (rights_id == Key128{})
return {};
return boost::none;
Key128 key_temp{};
@@ -356,17 +356,17 @@ std::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 {};
return boost::none;
m_1 = m_1 ^ MGF1<0x20>(m_2);
m_2 = m_2 ^ MGF1<0xDF>(m_1);
const auto offset = FindTicketOffset(m_2);
if (!offset)
return {};
ASSERT(*offset > 0);
if (offset == boost::none)
return boost::none;
ASSERT(offset.get() > 0);
std::memcpy(key_temp.data(), m_2.data() + *offset, key_temp.size());
std::memcpy(key_temp.data(), m_2.data() + offset.get(), 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)
SetKey(S128KeyType::SDSeed, *res);
if (res != boost::none)
SetKey(S128KeyType::SDSeed, res.get());
}
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)
if (pair == boost::none)
continue;
const auto& [rid, key] = *pair;
const auto& [rid, key] = pair.value();
u128 rights_id;
std::memcpy(rights_id.data(), rid.data(), rid.size());
SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);

View File

@@ -6,10 +6,9 @@
#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"
@@ -192,14 +191,14 @@ 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);
std::optional<Key128> DeriveSDSeed();
boost::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)
std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
const RSAKeyPair<2048>& eticket_extended_key);
boost::optional<std::pair<Key128, Key128>> ParseTicket(
const TicketRaw& ticket, const RSAKeyPair<2048>& eticket_extended_key);
} // namespace Core::Crypto

View File

@@ -4,9 +4,10 @@
#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"
@@ -305,18 +306,18 @@ 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});
std::optional<Core::Crypto::Key128> key = {};
boost::optional<Core::Crypto::Key128> key = boost::none;
if (encrypted) {
if (has_rights_id) {
status = Loader::ResultStatus::Success;
key = GetTitlekey();
if (!key) {
if (key == boost::none) {
status = Loader::ResultStatus::ErrorMissingTitlekey;
return false;
}
} else {
key = GetKeyAreaKey(NCASectionCryptoType::BKTR);
if (!key) {
if (key == boost::none) {
status = Loader::ResultStatus::ErrorMissingKeyAreaKey;
return false;
}
@@ -331,7 +332,7 @@ 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 : Core::Crypto::Key128{}, base_offset, bktr_base_ivfc_offset,
encrypted ? key.get() : 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
@@ -387,11 +388,11 @@ u8 NCA::GetCryptoRevision() const {
return master_key_id;
}
std::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const {
boost::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 {};
return boost::none;
std::vector<u8> key_area(header.key_area.begin(), header.key_area.end());
Core::Crypto::AESCipher<Core::Crypto::Key128> cipher(
@@ -415,25 +416,25 @@ std::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type
return out;
}
std::optional<Core::Crypto::Key128> NCA::GetTitlekey() {
boost::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 {};
return boost::none;
}
auto titlekey = keys.GetKey(Core::Crypto::S128KeyType::Titlekey, rights_id[1], rights_id[0]);
if (titlekey == Core::Crypto::Key128{}) {
status = Loader::ResultStatus::ErrorMissingTitlekey;
return {};
return boost::none;
}
if (!keys.HasKey(Core::Crypto::S128KeyType::Titlekek, master_key_id)) {
status = Loader::ResultStatus::ErrorMissingTitlekek;
return {};
return boost::none;
}
Core::Crypto::AESCipher<Core::Crypto::Key128> cipher(
@@ -457,25 +458,25 @@ 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);
{
std::optional<Core::Crypto::Key128> key = {};
boost::optional<Core::Crypto::Key128> key = boost::none;
if (has_rights_id) {
status = Loader::ResultStatus::Success;
key = GetTitlekey();
if (!key) {
if (key == boost::none) {
if (status == Loader::ResultStatus::Success)
status = Loader::ResultStatus::ErrorMissingTitlekey;
return nullptr;
}
} else {
key = GetKeyAreaKey(NCASectionCryptoType::CTR);
if (!key) {
if (key == boost::none) {
status = Loader::ResultStatus::ErrorMissingKeyAreaKey;
return nullptr;
}
}
auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(std::move(in), *key,
starting_offset);
auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(
std::move(in), key.value(), starting_offset);
std::vector<u8> iv(16);
for (u8 i = 0; i < 8; ++i)
iv[i] = s_header.raw.section_ctr[0x8 - i - 1];

View File

@@ -6,10 +6,9 @@
#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"
@@ -112,8 +111,8 @@ private:
bool ReadPFS0Section(const NCASectionHeader& section, const NCASectionTableEntry& entry);
u8 GetCryptoRevision() const;
std::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const;
std::optional<Core::Crypto::Key128> GetTitlekey();
boost::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const;
boost::optional<Core::Crypto::Key128> GetTitlekey();
VirtualFile Decrypt(const NCASectionHeader& header, VirtualFile in, u64 starting_offset);
std::vector<VirtualDir> dirs;

View File

@@ -66,4 +66,10 @@ std::string NACP::GetVersionString() const {
return Common::StringFromFixedZeroTerminatedBuffer(raw->version_string.data(),
raw->version_string.size());
}
std::vector<u8> NACP::GetRawBytes() const {
std::vector<u8> out(sizeof(RawNACP));
std::memcpy(out.data(), raw.get(), sizeof(RawNACP));
return out;
}
} // namespace FileSys

View File

@@ -81,6 +81,7 @@ public:
u64 GetTitleId() const;
u64 GetDLCBaseTitleId() const;
std::string GetVersionString() const;
std::vector<u8> GetRawBytes() const;
private:
std::unique_ptr<RawNACP> raw;

View File

@@ -27,6 +27,7 @@
#include <map>
#include <memory>
#include <string>
#include <boost/detail/container_fwd.hpp>
#include "common/common_types.h"
#include "core/file_sys/vfs.h"

View File

@@ -103,12 +103,12 @@ VirtualFile PatchIPS(const VirtualFile& in, const VirtualFile& ips) {
offset += sizeof(u16);
const auto data = ips->ReadByte(offset++);
if (!data)
if (data == boost::none)
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, rle_size);
std::memset(in_data.data() + real_offset, data.get(), rle_size);
} else { // Standard Patch
auto read = data_size;
if (real_offset + read > in_data.size())

View File

@@ -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).value_or(0)));
FormatTitleVersion(installed->GetEntryVersion(update_tid).get_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).value_or(0)));
FormatTitleVersion(installed->GetEntryVersion(update_tid).get_value_or(0)));
romfs = new_nca->GetRomFS();
}
} else if (update_raw != nullptr) {
@@ -280,11 +280,12 @@ 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.value_or(0) == 0) {
if (meta_ver == boost::none || meta_ver.get() == 0) {
out.insert_or_assign("Update", "");
} else {
out.insert_or_assign(
"Update", FormatTitleVersion(*meta_ver, TitleVersionFormat::ThreeElements));
"Update",
FormatTitleVersion(meta_ver.get(), TitleVersionFormat::ThreeElements));
}
} else if (update_raw != nullptr) {
out.insert_or_assign("Update", "PACKED");

View File

@@ -159,28 +159,28 @@ VirtualFile RegisteredCache::GetFileAtID(NcaID id) const {
return file;
}
static std::optional<NcaID> CheckMapForContentRecord(
static boost::optional<NcaID> CheckMapForContentRecord(
const boost::container::flat_map<u64, CNMT>& map, u64 title_id, ContentRecordType type) {
if (map.find(title_id) == map.end())
return {};
return boost::none;
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 {};
return boost::none;
return std::make_optional(iter->nca_id);
return boost::make_optional(iter->nca_id);
}
std::optional<NcaID> RegisteredCache::GetNcaIDFromMetadata(u64 title_id,
ContentRecordType type) const {
boost::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)
if (res1 != boost::none)
return res1;
return CheckMapForContentRecord(meta, title_id, type);
}
@@ -283,14 +283,17 @@ bool RegisteredCache::HasEntry(RegisteredCacheEntry entry) const {
VirtualFile RegisteredCache::GetEntryUnparsed(u64 title_id, ContentRecordType type) const {
const auto id = GetNcaIDFromMetadata(title_id, type);
return id ? GetFileAtID(*id) : nullptr;
if (id == boost::none)
return nullptr;
return GetFileAtID(id.get());
}
VirtualFile RegisteredCache::GetEntryUnparsed(RegisteredCacheEntry entry) const {
return GetEntryUnparsed(entry.title_id, entry.type);
}
std::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const {
boost::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();
@@ -299,12 +302,15 @@ std::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const {
if (yuzu_meta_iter != yuzu_meta.end())
return yuzu_meta_iter->second.GetTitleVersion();
return {};
return boost::none;
}
VirtualFile RegisteredCache::GetEntryRaw(u64 title_id, ContentRecordType type) const {
const auto id = GetNcaIDFromMetadata(title_id, type);
return id ? parser(GetFileAtID(*id), *id) : nullptr;
if (id == boost::none)
return nullptr;
return parser(GetFileAtID(id.get()), id.get());
}
VirtualFile RegisteredCache::GetEntryRaw(RegisteredCacheEntry entry) const {
@@ -358,8 +364,8 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntries() const {
}
std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter(
std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type,
std::optional<u64> title_id) const {
boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type,
boost::optional<u64> title_id) const {
std::vector<RegisteredCacheEntry> out;
IterateAllMetadata<RegisteredCacheEntry>(
out,
@@ -367,11 +373,11 @@ 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 && *title_type != c.GetType())
if (title_type != boost::none && title_type.get() != c.GetType())
return false;
if (record_type && *record_type != r.type)
if (record_type != boost::none && record_type.get() != r.type)
return false;
if (title_id && *title_id != c.GetTitleID())
if (title_id != boost::none && title_id.get() != c.GetTitleID())
return false;
return true;
});
@@ -453,7 +459,7 @@ InstallResult RegisteredCache::InstallEntry(std::shared_ptr<NCA> nca, TitleType
InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy,
bool overwrite_if_exists,
std::optional<NcaID> override_id) {
boost::optional<NcaID> override_id) {
const auto in = nca->GetBaseFile();
Core::Crypto::SHA256Hash hash{};
@@ -462,12 +468,12 @@ 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) {
id = *override_id;
} else {
if (override_id == boost::none) {
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);
@@ -537,14 +543,14 @@ bool RegisteredCacheUnion::HasEntry(RegisteredCacheEntry entry) const {
return HasEntry(entry.title_id, entry.type);
}
std::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const {
boost::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const {
for (const auto& c : caches) {
const auto res = c->GetEntryVersion(title_id);
if (res)
if (res != boost::none)
return res;
}
return {};
return boost::none;
}
VirtualFile RegisteredCacheUnion::GetEntryUnparsed(u64 title_id, ContentRecordType type) const {
@@ -603,8 +609,8 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntries() const {
}
std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter(
std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type,
std::optional<u64> title_id) const {
boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type,
boost::optional<u64> title_id) const {
std::vector<RegisteredCacheEntry> out;
for (const auto& c : caches) {
c->IterateAllMetadata<RegisteredCacheEntry>(
@@ -613,11 +619,11 @@ 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 && *title_type != c.GetType())
if (title_type != boost::none && title_type.get() != c.GetType())
return false;
if (record_type && *record_type != r.type)
if (record_type != boost::none && record_type.get() != r.type)
return false;
if (title_id && *title_id != c.GetTitleID())
if (title_id != boost::none && title_id.get() != c.GetTitleID())
return false;
return true;
});

View File

@@ -84,7 +84,7 @@ public:
bool HasEntry(u64 title_id, ContentRecordType type) const;
bool HasEntry(RegisteredCacheEntry entry) const;
std::optional<u32> GetEntryVersion(u64 title_id) const;
boost::optional<u32> GetEntryVersion(u64 title_id) const;
VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const;
VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const;
@@ -96,10 +96,11 @@ public:
std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const;
std::vector<RegisteredCacheEntry> ListEntries() const;
// If a parameter is not std::nullopt, it will be filtered for from all entries.
// If a parameter is not boost::none, it will be filtered for from all entries.
std::vector<RegisteredCacheEntry> ListEntriesFilter(
std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {},
std::optional<u64> title_id = {}) const;
boost::optional<TitleType> title_type = boost::none,
boost::optional<ContentRecordType> record_type = boost::none,
boost::optional<u64> title_id = boost::none) 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.
@@ -124,11 +125,12 @@ private:
std::vector<NcaID> AccumulateFiles() const;
void ProcessFiles(const std::vector<NcaID>& ids);
void AccumulateYuzuMeta();
std::optional<NcaID> GetNcaIDFromMetadata(u64 title_id, ContentRecordType type) const;
boost::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, std::optional<NcaID> override_id = {});
bool overwrite_if_exists,
boost::optional<NcaID> override_id = boost::none);
bool RawInstallYuzuMeta(const CNMT& cnmt);
VirtualDir dir;
@@ -151,7 +153,7 @@ public:
bool HasEntry(u64 title_id, ContentRecordType type) const;
bool HasEntry(RegisteredCacheEntry entry) const;
std::optional<u32> GetEntryVersion(u64 title_id) const;
boost::optional<u32> GetEntryVersion(u64 title_id) const;
VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const;
VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const;
@@ -163,10 +165,11 @@ public:
std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const;
std::vector<RegisteredCacheEntry> ListEntries() const;
// If a parameter is not std::nullopt, it will be filtered for from all entries.
// If a parameter is not boost::none, it will be filtered for from all entries.
std::vector<RegisteredCacheEntry> ListEntriesFilter(
std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {},
std::optional<u64> title_id = {}) const;
boost::optional<TitleType> title_type = boost::none,
boost::optional<ContentRecordType> record_type = boost::none,
boost::optional<u64> title_id = boost::none) const;
private:
std::vector<RegisteredCache*> caches;

View File

@@ -83,6 +83,24 @@ ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, SaveDataDescr
return MakeResult<VirtualDir>(std::move(out));
}
VirtualDir SaveDataFactory::GetSaveDataSpaceDirectory(SaveDataSpaceId space) const {
return dir->GetDirectoryRelative(GetSaveDataSpaceIdPath(space));
}
std::string SaveDataFactory::GetSaveDataSpaceIdPath(SaveDataSpaceId space) {
switch (space) {
case SaveDataSpaceId::NandSystem:
return "/system/";
case SaveDataSpaceId::NandUser:
return "/user/";
case SaveDataSpaceId::TemporaryStorage:
return "/temp/";
default:
ASSERT_MSG(false, "Unrecognized SaveDataSpaceId: {:02X}", static_cast<u8>(space));
return "/unrecognized/"; ///< To prevent corruption when ignoring asserts.
}
}
std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType type, u64 title_id,
u128 user_id, u64 save_id) {
// According to switchbrew, if a save is of type SaveData and the title id field is 0, it should
@@ -90,21 +108,7 @@ std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType typ
if (type == SaveDataType::SaveData && title_id == 0)
title_id = Core::CurrentProcess()->GetTitleID();
std::string out;
switch (space) {
case SaveDataSpaceId::NandSystem:
out = "/system/";
break;
case SaveDataSpaceId::NandUser:
out = "/user/";
break;
case SaveDataSpaceId::TemporaryStorage:
out = "/temp/";
break;
default:
ASSERT_MSG(false, "Unrecognized SaveDataSpaceId: {:02X}", static_cast<u8>(space));
}
std::string out = GetSaveDataSpaceIdPath(space);
switch (type) {
case SaveDataType::SystemSaveData:

View File

@@ -52,6 +52,9 @@ public:
ResultVal<VirtualDir> Open(SaveDataSpaceId space, SaveDataDescriptor meta);
VirtualDir GetSaveDataSpaceDirectory(SaveDataSpaceId space) const;
static std::string GetSaveDataSpaceIdPath(SaveDataSpaceId space);
static std::string GetFullPath(SaveDataSpaceId space, SaveDataType type, u64 title_id,
u128 user_id, u64 save_id);

View File

@@ -167,13 +167,13 @@ std::string VfsFile::GetExtension() const {
VfsDirectory::~VfsDirectory() = default;
std::optional<u8> VfsFile::ReadByte(std::size_t offset) const {
boost::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 {};
return boost::none;
}
std::vector<u8> VfsFile::ReadBytes(std::size_t size, std::size_t offset) const {

View File

@@ -4,15 +4,13 @@
#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"
@@ -105,8 +103,8 @@ 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 std::nullopt on error.
virtual std::optional<u8> ReadByte(std::size_t offset = 0) const;
// 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 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(),

View File

@@ -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);
}
std::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const {
boost::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const {
if (r_offset < size)
return file->ReadByte(offset + r_offset);
return {};
return boost::none;
}
std::vector<u8> OffsetVfsFile::ReadBytes(std::size_t r_size, std::size_t r_offset) const {

View File

@@ -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;
std::optional<u8> ReadByte(std::size_t offset) const override;
boost::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;

View File

@@ -53,10 +53,10 @@ public:
return 0;
}
std::optional<u8> ReadByte(std::size_t offset) const override {
boost::optional<u8> ReadByte(std::size_t offset) const override {
if (offset < size)
return value;
return {};
return boost::none;
}
std::vector<u8> ReadBytes(std::size_t length, std::size_t offset) const override {

View File

@@ -117,7 +117,8 @@ public:
AlignWithPadding();
if (context.Session()->IsDomain() && context.HasDomainMessageHeader()) {
const bool request_has_domain_header{context.GetDomainMessageHeader() != nullptr};
if (context.Session()->IsDomain() && request_has_domain_header) {
IPC::DomainMessageHeader domain_header{};
domain_header.num_objects = num_domain_objects;
PushRaw(domain_header);

View File

@@ -161,12 +161,8 @@ public:
return buffer_c_desciptors;
}
const IPC::DomainMessageHeader* GetDomainMessageHeader() const {
return domain_message_header.get();
}
bool HasDomainMessageHeader() const {
return domain_message_header != nullptr;
const std::shared_ptr<IPC::DomainMessageHeader>& GetDomainMessageHeader() const {
return domain_message_header;
}
/// Helper function to read a buffer using the appropriate buffer descriptor

View File

@@ -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);
const auto& system = Core::System::GetInstance();
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);
const auto& system = Core::System::GetInstance();
auto& system = Core::System::GetInstance();
SharedPtr<Timer> timer = system.Kernel().RetrieveTimerFromCallbackHandleTable(proper_handle);
if (timer == nullptr) {

View File

@@ -6,6 +6,8 @@
#include <utility>
#include <vector>
#include <boost/range/algorithm_ext/erase.hpp>
#include "common/assert.h"
#include "core/core.h"
#include "core/hle/kernel/errors.h"

View File

@@ -63,7 +63,7 @@ void ServerSession::Acquire(Thread* thread) {
}
ResultCode ServerSession::HandleDomainSyncRequest(Kernel::HLERequestContext& context) {
auto* const domain_message_header = context.GetDomainMessageHeader();
auto& 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.HasDomainMessageHeader()) {
if (IsDomain() && context.GetDomainMessageHeader()) {
result = HandleDomainSyncRequest(context);
// If there is no domain header, the regular session handler is used
} else if (hle_handler != nullptr) {

View File

@@ -572,7 +572,7 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
return ERR_INVALID_HANDLE;
}
const auto& system = Core::System::GetInstance();
auto& system = Core::System::GetInstance();
const auto& scheduler = system.CurrentScheduler();
const auto* const current_thread = scheduler.GetCurrentThread();
const bool same_thread = current_thread == thread;

View File

@@ -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 std::optional<s32> GetNextProcessorId(u64 mask) {
static boost::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;
std::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask);
boost::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;
}
std::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
boost::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
if (!new_processor_id) {
new_processor_id = processor_id;

View File

@@ -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).has_value();
return GetUserIndex(uuid) != std::nullopt;
}
bool ProfileManager::UserExistsIndex(std::size_t index) const {

View File

@@ -57,8 +57,7 @@ struct UUID {
};
static_assert(sizeof(UUID) == 16, "UUID is an invalid size!");
constexpr std::size_t profile_username_size = 32;
using ProfileUsername = std::array<u8, profile_username_size>;
using ProfileUsername = std::array<u8, 0x20>;
using ProfileData = std::array<u8, MAX_DATA>;
using UserIDArray = std::array<UUID, MAX_USERS>;

View File

@@ -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);
ASSERT(uuid != std::nullopt);
params.current_user = uuid->uuid;
IPC::ResponseBuilder rb{ctx, 2, 0, 1};

View File

@@ -161,7 +161,7 @@ void HwOpus::OpenOpusDecoder(Kernel::HLERequestContext& ctx) {
ASSERT_MSG(channel_count == 1 || channel_count == 2, "Invalid channel count");
std::size_t worker_sz = WorkerBufferSize(channel_count);
ASSERT_MSG(buffer_sz >= worker_sz, "Worker buffer too large");
ASSERT_MSG(buffer_sz < worker_sz, "Worker buffer too large");
std::unique_ptr<OpusDecoder, OpusDeleter> decoder{
static_cast<OpusDecoder*>(operator new(worker_sz))};
if (opus_decoder_init(decoder.get(), sample_rate, channel_count)) {

View File

@@ -309,6 +309,16 @@ ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space,
return save_data_factory->Open(space, save_struct);
}
ResultVal<FileSys::VirtualDir> OpenSaveDataSpace(FileSys::SaveDataSpaceId space) {
LOG_TRACE(Service_FS, "Opening Save Data Space for space_id={:01X}", static_cast<u8>(space));
if (save_data_factory == nullptr) {
return ResultCode(ErrorModule::FS, FileSys::ErrCodes::TitleNotFound);
}
return MakeResult(save_data_factory->GetSaveDataSpaceDirectory(space));
}
ResultVal<FileSys::VirtualDir> OpenSDMC() {
LOG_TRACE(Service_FS, "Opening SDMC");

View File

@@ -45,6 +45,7 @@ ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId stora
FileSys::ContentRecordType type);
ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space,
FileSys::SaveDataDescriptor save_struct);
ResultVal<FileSys::VirtualDir> OpenSaveDataSpace(FileSys::SaveDataSpaceId space);
ResultVal<FileSys::VirtualDir> OpenSDMC();
std::unique_ptr<FileSys::RegisteredCacheUnion> GetUnionContents();

View File

@@ -11,6 +11,7 @@
#include "common/assert.h"
#include "common/common_types.h"
#include "common/hex_util.h"
#include "common/logging/log.h"
#include "common/string_util.h"
#include "core/file_sys/directory.h"
@@ -451,7 +452,147 @@ private:
VfsDirectoryServiceWrapper backend;
};
class ISaveDataInfoReader final : public ServiceFramework<ISaveDataInfoReader> {
public:
explicit ISaveDataInfoReader(FileSys::SaveDataSpaceId space)
: ServiceFramework("ISaveDataInfoReader") {
static const FunctionInfo functions[] = {
{0, &ISaveDataInfoReader::ReadSaveDataInfo, "ReadSaveDataInfo"},
};
RegisterHandlers(functions);
FindAllSaves(space);
}
void ReadSaveDataInfo(Kernel::HLERequestContext& ctx) {
// Calculate how many entries we can fit in the output buffer
const u64 count_entries = ctx.GetWriteBufferSize() / sizeof(SaveDataInfo);
// Cap at total number of entries.
const u64 actual_entries = std::min(count_entries, info.size() - next_entry_index);
// Determine data start and end
const auto* begin = reinterpret_cast<u8*>(info.data() + next_entry_index);
const auto* end = reinterpret_cast<u8*>(info.data() + next_entry_index + actual_entries);
const auto range_size = static_cast<std::size_t>(std::distance(begin, end));
next_entry_index += actual_entries;
// Write the data to memory
ctx.WriteBuffer(begin, range_size);
IPC::ResponseBuilder rb{ctx, 3};
rb.Push(RESULT_SUCCESS);
rb.Push<u32>(static_cast<u32>(actual_entries));
}
private:
static u64 stoull_be(std::string_view str) {
if (str.size() != 16)
return 0;
const auto bytes = Common::HexStringToArray<0x8>(str);
u64 out{};
std::memcpy(&out, bytes.data(), sizeof(u64));
return Common::swap64(out);
}
void FindAllSaves(FileSys::SaveDataSpaceId space) {
const auto save_root = OpenSaveDataSpace(space);
ASSERT(save_root.Succeeded());
for (const auto& type : (*save_root)->GetSubdirectories()) {
if (type->GetName() == "save") {
for (const auto& save_id : type->GetSubdirectories()) {
for (const auto& user_id : save_id->GetSubdirectories()) {
const auto save_id_numeric = stoull_be(save_id->GetName());
auto user_id_numeric = Common::HexStringToArray<0x10>(user_id->GetName());
std::reverse(user_id_numeric.begin(), user_id_numeric.end());
if (save_id_numeric != 0) {
// System Save Data
info.emplace_back(SaveDataInfo{
0,
space,
FileSys::SaveDataType::SystemSaveData,
{},
user_id_numeric,
save_id_numeric,
0,
user_id->GetSize(),
{},
});
continue;
}
for (const auto& title_id : user_id->GetSubdirectories()) {
const auto device =
std::all_of(user_id_numeric.begin(), user_id_numeric.end(),
[](u8 val) { return val == 0; });
info.emplace_back(SaveDataInfo{
0,
space,
device ? FileSys::SaveDataType::DeviceSaveData
: FileSys::SaveDataType::SaveData,
{},
user_id_numeric,
save_id_numeric,
stoull_be(title_id->GetName()),
title_id->GetSize(),
{},
});
}
}
}
} else if (space == FileSys::SaveDataSpaceId::TemporaryStorage) {
// Temporary Storage
for (const auto& user_id : type->GetSubdirectories()) {
for (const auto& title_id : user_id->GetSubdirectories()) {
if (!title_id->GetFiles().empty() ||
!title_id->GetSubdirectories().empty()) {
auto user_id_numeric =
Common::HexStringToArray<0x10>(user_id->GetName());
std::reverse(user_id_numeric.begin(), user_id_numeric.end());
info.emplace_back(SaveDataInfo{
0,
space,
FileSys::SaveDataType::TemporaryStorage,
{},
user_id_numeric,
stoull_be(type->GetName()),
stoull_be(title_id->GetName()),
title_id->GetSize(),
{},
});
}
}
}
}
}
}
struct SaveDataInfo {
u64_le save_id_unknown;
FileSys::SaveDataSpaceId space;
FileSys::SaveDataType type;
INSERT_PADDING_BYTES(0x6);
std::array<u8, 0x10> user_id;
u64_le save_id;
u64_le title_id;
u64_le save_image_size;
INSERT_PADDING_BYTES(0x28);
};
static_assert(sizeof(SaveDataInfo) == 0x60, "SaveDataInfo has incorrect size.");
std::vector<SaveDataInfo> info;
u64 next_entry_index = 0;
};
FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") {
// clang-format off
static const FunctionInfo functions[] = {
{0, nullptr, "MountContent"},
{1, &FSP_SRV::Initialize, "Initialize"},
@@ -485,7 +626,7 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") {
{58, nullptr, "ReadSaveDataFileSystemExtraData"},
{59, nullptr, "WriteSaveDataFileSystemExtraData"},
{60, nullptr, "OpenSaveDataInfoReader"},
{61, nullptr, "OpenSaveDataInfoReaderBySaveDataSpaceId"},
{61, &FSP_SRV::OpenSaveDataInfoReaderBySaveDataSpaceId, "OpenSaveDataInfoReaderBySaveDataSpaceId"},
{62, nullptr, "OpenCacheStorageList"},
{64, nullptr, "OpenSaveDataInternalStorageFileSystem"},
{65, nullptr, "UpdateSaveDataMacForDebug"},
@@ -544,6 +685,7 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") {
{1009, nullptr, "GetAndClearMemoryReportInfo"},
{1100, nullptr, "OverrideSaveDataTransferTokenSignVerificationKey"},
};
// clang-format on
RegisterHandlers(functions);
}
@@ -618,6 +760,15 @@ void FSP_SRV::OpenReadOnlySaveDataFileSystem(Kernel::HLERequestContext& ctx) {
MountSaveData(ctx);
}
void FSP_SRV::OpenSaveDataInfoReaderBySaveDataSpaceId(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto space = rp.PopRaw<FileSys::SaveDataSpaceId>();
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(RESULT_SUCCESS);
rb.PushIpcInterface<ISaveDataInfoReader>(std::make_shared<ISaveDataInfoReader>(space));
}
void FSP_SRV::GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) {
LOG_WARNING(Service_FS, "(STUBBED) called");

View File

@@ -25,6 +25,7 @@ private:
void CreateSaveData(Kernel::HLERequestContext& ctx);
void MountSaveData(Kernel::HLERequestContext& ctx);
void OpenReadOnlySaveDataFileSystem(Kernel::HLERequestContext& ctx);
void OpenSaveDataInfoReaderBySaveDataSpaceId(Kernel::HLERequestContext& ctx);
void GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx);
void OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx);
void OpenDataStorageByDataId(Kernel::HLERequestContext& ctx);

View File

@@ -427,9 +427,6 @@ void Controller_NPad::VibrateController(const std::vector<u32>& controller_ids,
}
Kernel::SharedPtr<Kernel::Event> Controller_NPad::GetStyleSetChangedEvent() const {
// TODO(ogniK): Figure out the best time to signal this event. This event seems that it should
// be signalled at least once, and signaled after a new controller is connected?
styleset_changed_event->Signal();
return styleset_changed_event;
}

View File

@@ -96,8 +96,6 @@ public:
// TODO(shinyquagsire23): Other update callbacks? (accel, gyro?)
CoreTiming::ScheduleEvent(pad_update_ticks, pad_update_event);
ReloadInputDevices();
}
void ActivateController(HidController controller) {

View File

@@ -2,6 +2,9 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include "common/logging/log.h"
#include "core/file_sys/control_metadata.h"
#include "core/file_sys/patch_manager.h"
#include "core/hle/ipc_helpers.h"
#include "core/hle/kernel/hle_ipc.h"
#include "core/hle/service/ns/ns.h"
@@ -118,7 +121,7 @@ public:
{305, nullptr, "TerminateSystemApplet"},
{306, nullptr, "LaunchOverlayApplet"},
{307, nullptr, "TerminateOverlayApplet"},
{400, nullptr, "GetApplicationControlData"},
{400, &IApplicationManagerInterface::GetApplicationControlData, "GetApplicationControlData"},
{401, nullptr, "InvalidateAllApplicationControlCache"},
{402, nullptr, "RequestDownloadApplicationControlData"},
{403, nullptr, "GetMaxApplicationControlCacheCount"},
@@ -243,6 +246,65 @@ public:
RegisterHandlers(functions);
}
void GetApplicationControlData(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto flag = rp.PopRaw<u64>();
LOG_DEBUG(Service_NS, "called with flag={:016X}", flag);
const auto title_id = rp.PopRaw<u64>();
const auto size = ctx.GetWriteBufferSize();
const FileSys::PatchManager pm{title_id};
const auto control = pm.GetControlMetadata();
std::vector<u8> out;
if (control.first != nullptr) {
if (size < 0x4000) {
LOG_ERROR(Service_NS,
"output buffer is too small! (actual={:016X}, expected_min=0x4000)",
size);
IPC::ResponseBuilder rb{ctx, 2};
// TODO(DarkLordZach): Find a better error code for this.
rb.Push(ResultCode(-1));
return;
}
out.resize(0x4000);
const auto bytes = control.first->GetRawBytes();
std::memcpy(out.data(), bytes.data(), bytes.size());
} else {
LOG_WARNING(Service_NS, "missing NACP data for title_id={:016X}, defaulting to zeros.",
title_id);
out.resize(std::min<u64>(0x4000, size));
}
if (control.second != nullptr) {
if (size < 0x4000 + control.second->GetSize()) {
LOG_ERROR(Service_NS,
"output buffer is too small! (actual={:016X}, expected_min={:016X})",
size, 0x4000 + control.second->GetSize());
IPC::ResponseBuilder rb{ctx, 2};
// TODO(DarkLordZach): Find a better error code for this.
rb.Push(ResultCode(-1));
return;
}
out.resize(0x4000 + control.second->GetSize());
control.second->Read(out.data() + 0x4000, control.second->GetSize());
} else {
LOG_WARNING(Service_NS, "missing icon data for title_id={:016X}, defaulting to zeros.",
title_id);
}
ctx.WriteBuffer(out);
IPC::ResponseBuilder rb{ctx, 3};
rb.Push(RESULT_SUCCESS);
rb.Push<u32>(static_cast<u32>(out.size()));
}
};
class IApplicationVersionInterface final : public ServiceFramework<IApplicationVersionInterface> {

View File

@@ -31,7 +31,7 @@ void BufferQueue::SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer)
buffer_wait_event->Signal();
}
std::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) {
boost::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 @@ std::optional<u32> BufferQueue::DequeueBuffer(u32 width, u32 height) {
});
if (itr == queue.end()) {
return {};
return boost::none;
}
itr->status = Buffer::Status::Dequeued;
@@ -70,12 +70,12 @@ void BufferQueue::QueueBuffer(u32 slot, BufferTransformFlags transform,
itr->crop_rect = crop_rect;
}
std::optional<std::reference_wrapper<const BufferQueue::Buffer>> BufferQueue::AcquireBuffer() {
boost::optional<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 {};
return boost::none;
itr->status = Buffer::Status::Acquired;
return *itr;
}

View File

@@ -4,9 +4,8 @@
#pragma once
#include <optional>
#include <vector>
#include <boost/optional.hpp>
#include "common/common_funcs.h"
#include "common/math_util.h"
#include "common/swap.h"
@@ -58,9 +57,9 @@ public:
/// Rotate source image 90 degrees clockwise
Rotate90 = 0x04,
/// Rotate source image 180 degrees
Rotate180 = 0x03,
Roate180 = 0x03,
/// Rotate source image 270 degrees clockwise
Rotate270 = 0x07,
Roate270 = 0x07,
};
struct Buffer {
@@ -74,11 +73,11 @@ public:
};
void SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer);
std::optional<u32> DequeueBuffer(u32 width, u32 height);
boost::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);
std::optional<std::reference_wrapper<const Buffer>> AcquireBuffer();
boost::optional<const Buffer&> AcquireBuffer();
void ReleaseBuffer(u32 slot);
u32 Query(QueryType type);

View File

@@ -3,7 +3,7 @@
// Refer to the license.txt file included.
#include <algorithm>
#include <optional>
#include <boost/optional.hpp>
#include "common/alignment.h"
#include "common/assert.h"
@@ -134,7 +134,7 @@ void NVFlinger::Compose() {
MicroProfileFlip();
if (!buffer) {
if (buffer == boost::none) {
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->get().igbp_buffer;
auto& igbp_buffer = buffer->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->get().transform, buffer->get().crop_rect);
igbp_buffer.width, igbp_buffer.height, igbp_buffer.stride, buffer->transform,
buffer->crop_rect);
buffer_queue->ReleaseBuffer(buffer->get().slot);
buffer_queue->ReleaseBuffer(buffer->slot);
}
}

View File

@@ -132,11 +132,11 @@ public:
// clang-format off
static const FunctionInfo functions[] = {
{0, nullptr, "BindNoticeEvent"},
{1, nullptr, "UnbindNoticeEvent"},
{1, nullptr, "Unknown1"},
{2, nullptr, "GetStatus"},
{3, nullptr, "GetNotice"},
{4, nullptr, "EnablePowerRequestNotice"},
{5, nullptr, "DisablePowerRequestNotice"},
{4, nullptr, "Unknown2"},
{5, nullptr, "Unknown3"},
{6, nullptr, "ReplyPowerRequest"},
};
// clang-format on

View File

@@ -6,10 +6,9 @@
#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"
@@ -507,9 +506,9 @@ private:
IGBPDequeueBufferRequestParcel request{ctx.ReadBuffer()};
const u32 width{request.data.width};
const u32 height{request.data.height};
std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
boost::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
if (slot) {
if (slot != boost::none) {
// Buffer is available
IGBPDequeueBufferResponseParcel response{*slot};
ctx.WriteBuffer(response.Serialize());
@@ -521,7 +520,7 @@ private:
Kernel::ThreadWakeupReason reason) {
// Repeat TransactParcel DequeueBuffer when a buffer is available
auto buffer_queue = nv_flinger->GetBufferQueue(id);
std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
boost::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
IGBPDequeueBufferResponseParcel response{*slot};
ctx.WriteBuffer(response.Serialize());
IPC::ResponseBuilder rb{ctx, 2};

View File

@@ -6,11 +6,10 @@
#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"
@@ -146,7 +145,7 @@ public:
* information.
* @returns A pair with the optional system mode, and and the status.
*/
virtual std::pair<std::optional<u32>, ResultStatus> LoadKernelSystemMode() {
virtual std::pair<boost::optional<u32>, ResultStatus> LoadKernelSystemMode() {
// 96MB allocated to the application.
return std::make_pair(2, ResultStatus::Success);
}

View File

@@ -12,10 +12,12 @@
#include "common/swap.h"
#include "core/core.h"
#include "core/file_sys/control_metadata.h"
#include "core/file_sys/romfs_factory.h"
#include "core/file_sys/vfs_offset.h"
#include "core/gdbstub/gdbstub.h"
#include "core/hle/kernel/process.h"
#include "core/hle/kernel/vm_manager.h"
#include "core/hle/service/filesystem/filesystem.h"
#include "core/loader/nro.h"
#include "core/loader/nso.h"
#include "core/memory.h"
@@ -208,6 +210,9 @@ ResultStatus AppLoader_NRO::Load(Kernel::Process& process) {
return ResultStatus::ErrorLoadingNRO;
}
if (romfs != nullptr)
Service::FileSystem::RegisterRomFS(std::make_unique<FileSys::RomFSFactory>(*this));
process.Run(base_address, Kernel::THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE);
is_loaded = true;

View File

@@ -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"

View File

@@ -5,8 +5,7 @@
#pragma once
#include <memory>
#include <optional>
#include <boost/optional.hpp>
#include "common/common_types.h"
namespace Memory {
@@ -19,19 +18,19 @@ namespace Memory {
*
* A hook may be mapped to multiple regions of memory.
*
* If a std::nullopt or false is returned from a function, the read/write request is passed through
* If a boost::none 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 std::optional<bool> IsValidAddress(VAddr addr) = 0;
virtual boost::optional<bool> IsValidAddress(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 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 bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) = 0;

View File

@@ -184,13 +184,4 @@ TelemetrySession::~TelemetrySession() {
backend = nullptr;
}
bool TelemetrySession::SubmitTestcase() {
#ifdef ENABLE_WEB_SERVICE
field_collection.Accept(*backend);
return backend->SubmitTestcase();
#else
return false;
#endif
}
} // namespace Core

View File

@@ -31,12 +31,6 @@ public:
field_collection.AddField(type, name, std::move(value));
}
/**
* Submits a Testcase.
* @returns A bool indicating whether the submission succeeded
*/
bool SubmitTestcase();
private:
Telemetry::FieldCollection field_collection; ///< Tracks all added fields for the session
std::unique_ptr<Telemetry::VisitorInterface> backend; ///< Backend interface that logs fields

View File

@@ -64,11 +64,11 @@ void TestEnvironment::ClearWriteRecords() {
TestEnvironment::TestMemory::~TestMemory() {}
std::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) {
boost::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) {
return true;
}
std::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
boost::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
const auto iter = data.find(addr);
if (iter == data.end()) {
@@ -79,15 +79,15 @@ std::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) {
return iter->second;
}
std::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) {
boost::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) {
return *Read8(addr) | static_cast<u16>(*Read8(addr + 1)) << 8;
}
std::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) {
boost::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) {
return *Read16(addr) | static_cast<u32>(*Read16(addr + 2)) << 16;
}
std::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) {
boost::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) {
return *Read32(addr) | static_cast<u64>(*Read32(addr + 4)) << 32;
}

View File

@@ -64,12 +64,12 @@ private:
~TestMemory() override;
std::optional<bool> IsValidAddress(VAddr addr) override;
boost::optional<bool> IsValidAddress(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;
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;
bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) override;

View File

@@ -33,7 +33,6 @@ add_library(video_core STATIC
renderer_opengl/gl_rasterizer.h
renderer_opengl/gl_rasterizer_cache.cpp
renderer_opengl/gl_rasterizer_cache.h
renderer_opengl/gl_resource_manager.cpp
renderer_opengl/gl_resource_manager.h
renderer_opengl/gl_shader_cache.cpp
renderer_opengl/gl_shader_cache.h
@@ -52,10 +51,6 @@ 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
surface.cpp
surface.h
textures/astc.cpp
textures/astc.h
textures/decoders.cpp

View File

@@ -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 std::optional<VAddr> head_address = memory_manager->GpuToCpuAddress(address);
const boost::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)};

View File

@@ -37,39 +37,21 @@ void Maxwell3D::InitializeRegisterDefaults() {
regs.viewport[viewport].depth_range_near = 0.0f;
regs.viewport[viewport].depth_range_far = 1.0f;
}
// Doom and Bomberman seems to use the uninitialized registers and just enable blend
// so initialize blend registers with sane values
regs.blend.equation_rgb = Regs::Blend::Equation::Add;
regs.blend.factor_source_rgb = Regs::Blend::Factor::One;
regs.blend.factor_dest_rgb = Regs::Blend::Factor::Zero;
regs.blend.equation_a = Regs::Blend::Equation::Add;
regs.blend.factor_source_a = Regs::Blend::Factor::One;
regs.blend.factor_dest_a = Regs::Blend::Factor::Zero;
for (std::size_t blend_index = 0; blend_index < Regs::NumRenderTargets; blend_index++) {
regs.independent_blend[blend_index].equation_rgb = Regs::Blend::Equation::Add;
regs.independent_blend[blend_index].factor_source_rgb = Regs::Blend::Factor::One;
regs.independent_blend[blend_index].factor_dest_rgb = Regs::Blend::Factor::Zero;
regs.independent_blend[blend_index].equation_a = Regs::Blend::Equation::Add;
regs.independent_blend[blend_index].factor_source_a = Regs::Blend::Factor::One;
regs.independent_blend[blend_index].factor_dest_a = Regs::Blend::Factor::Zero;
}
}
void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
// Reset the current macro.
executing_macro = 0;
// Lookup the macro offset
const u32 entry{(method - MacroRegistersStart) >> 1};
const auto& search{macro_offsets.find(entry)};
if (search == macro_offsets.end()) {
LOG_CRITICAL(HW_GPU, "macro not found for method 0x{:X}!", method);
UNREACHABLE();
// The requested macro must have been uploaded already.
auto macro_code = uploaded_macros.find(method);
if (macro_code == uploaded_macros.end()) {
LOG_ERROR(HW_GPU, "Macro {:04X} was not uploaded", method);
return;
}
// Execute the current macro.
macro_interpreter.Execute(search->second, std::move(parameters));
macro_interpreter.Execute(macro_code->second, std::move(parameters));
}
void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
@@ -115,10 +97,6 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
ProcessMacroUpload(value);
break;
}
case MAXWELL3D_REG_INDEX(macros.bind): {
ProcessMacroBind(value);
break;
}
case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]):
case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]):
case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]):
@@ -180,20 +158,16 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) {
}
void Maxwell3D::ProcessMacroUpload(u32 data) {
ASSERT_MSG(regs.macros.upload_address < macro_memory.size(),
"upload_address exceeded macro_memory size!");
macro_memory[regs.macros.upload_address++] = data;
}
void Maxwell3D::ProcessMacroBind(u32 data) {
macro_offsets[regs.macros.entry] = data;
// Store the uploaded macro code to interpret them when they're called.
auto& macro = uploaded_macros[regs.macros.entry * 2 + MacroRegistersStart];
macro.push_back(data);
}
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.
std::optional<VAddr> address = memory_manager.GpuToCpuAddress(sequence_address);
boost::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,
@@ -311,7 +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);
std::optional<VAddr> address =
boost::optional<VAddr> address =
memory_manager.GpuToCpuAddress(buffer_address + regs.const_buffer.cb_pos);
Memory::Write32(*address, value);
@@ -324,7 +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);
std::optional<VAddr> tic_address_cpu = memory_manager.GpuToCpuAddress(tic_address_gpu);
boost::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));
@@ -348,7 +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);
std::optional<VAddr> tsc_address_cpu = memory_manager.GpuToCpuAddress(tsc_address_gpu);
boost::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));
@@ -412,7 +386,7 @@ Texture::FullTextureInfo Maxwell3D::GetStageTexture(Regs::ShaderStage stage,
ASSERT(tex_info_address < tex_info_buffer.address + tex_info_buffer.size);
std::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address);
boost::optional<VAddr> tex_address_cpu = memory_manager.GpuToCpuAddress(tex_info_address);
Texture::TextureHandle tex_handle{Memory::Read32(*tex_address_cpu)};
Texture::FullTextureInfo tex_info{};

View File

@@ -462,16 +462,6 @@ public:
}
};
struct ColorMask {
union {
u32 raw;
BitField<0, 4, u32> R;
BitField<4, 4, u32> G;
BitField<8, 4, u32> B;
BitField<12, 4, u32> A;
};
};
bool IsShaderConfigEnabled(std::size_t index) const {
// The VertexB is always enabled.
if (index == static_cast<std::size_t>(Regs::ShaderProgram::VertexB)) {
@@ -485,13 +475,12 @@ public:
INSERT_PADDING_WORDS(0x45);
struct {
u32 upload_address;
INSERT_PADDING_WORDS(1);
u32 data;
u32 entry;
u32 bind;
} macros;
INSERT_PADDING_WORDS(0x188);
INSERT_PADDING_WORDS(0x189);
u32 tfb_enabled;
@@ -581,11 +570,7 @@ public:
u32 stencil_back_mask;
u32 stencil_back_func_mask;
INSERT_PADDING_WORDS(0xC);
u32 color_mask_common;
INSERT_PADDING_WORDS(0x6);
INSERT_PADDING_WORDS(0x13);
u32 rt_separate_frag_data;
@@ -660,14 +645,8 @@ public:
ComparisonOp depth_test_func;
float alpha_test_ref;
ComparisonOp alpha_test_func;
u32 draw_tfb_stride;
struct {
float r;
float g;
float b;
float a;
} blend_color;
INSERT_PADDING_WORDS(0x4);
INSERT_PADDING_WORDS(0x9);
struct {
u32 separate_alpha;
@@ -861,9 +840,8 @@ public:
BitField<6, 4, u32> RT;
BitField<10, 11, u32> layer;
} clear_buffers;
INSERT_PADDING_WORDS(0xB);
std::array<ColorMask, NumRenderTargets> color_mask;
INSERT_PADDING_WORDS(0x38);
INSERT_PADDING_WORDS(0x4B);
struct {
u32 query_address_high;
@@ -1016,25 +994,12 @@ public:
/// Returns the texture information for a specific texture in a specific shader stage.
Texture::FullTextureInfo GetStageTexture(Regs::ShaderStage stage, std::size_t offset) const;
/// Memory for macro code - it's undetermined how big this is, however 1MB is much larger than
/// we've seen used.
using MacroMemory = std::array<u32, 0x40000>;
/// Gets a reference to macro memory.
const MacroMemory& GetMacroMemory() const {
return macro_memory;
}
private:
void InitializeRegisterDefaults();
VideoCore::RasterizerInterface& rasterizer;
/// Start offsets of each macro in macro_memory
std::unordered_map<u32, u32> macro_offsets;
/// Memory for macro code
MacroMemory macro_memory;
std::unordered_map<u32, std::vector<u32>> uploaded_macros;
/// Macro method that is currently being executed / being fed parameters.
u32 executing_macro = 0;
@@ -1057,12 +1022,9 @@ private:
*/
void CallMacroMethod(u32 method, std::vector<u32> parameters);
/// Handles writes to the macro uploading register.
/// Handles writes to the macro uploading registers.
void ProcessMacroUpload(u32 data);
/// Handles writes to the macro bind register.
void ProcessMacroBind(u32 data);
/// Handles a write to the CLEAR_BUFFERS register.
void ProcessClearBuffers();
@@ -1096,7 +1058,6 @@ ASSERT_REG_POSITION(scissor_test, 0x380);
ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5);
ASSERT_REG_POSITION(stencil_back_mask, 0x3D6);
ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7);
ASSERT_REG_POSITION(color_mask_common, 0x3E4);
ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB);
ASSERT_REG_POSITION(zeta, 0x3F8);
ASSERT_REG_POSITION(vertex_attrib_format, 0x458);
@@ -1109,10 +1070,6 @@ ASSERT_REG_POSITION(depth_write_enabled, 0x4BA);
ASSERT_REG_POSITION(alpha_test_enabled, 0x4BB);
ASSERT_REG_POSITION(d3d_cull_mode, 0x4C2);
ASSERT_REG_POSITION(depth_test_func, 0x4C3);
ASSERT_REG_POSITION(alpha_test_ref, 0x4C4);
ASSERT_REG_POSITION(alpha_test_func, 0x4C5);
ASSERT_REG_POSITION(draw_tfb_stride, 0x4C6);
ASSERT_REG_POSITION(blend_color, 0x4C7);
ASSERT_REG_POSITION(blend, 0x4CF);
ASSERT_REG_POSITION(stencil_enable, 0x4E0);
ASSERT_REG_POSITION(stencil_front_op_fail, 0x4E1);
@@ -1143,7 +1100,6 @@ ASSERT_REG_POSITION(instanced_arrays, 0x620);
ASSERT_REG_POSITION(cull, 0x646);
ASSERT_REG_POSITION(logic_op, 0x671);
ASSERT_REG_POSITION(clear_buffers, 0x674);
ASSERT_REG_POSITION(color_mask, 0x680);
ASSERT_REG_POSITION(query, 0x6C0);
ASSERT_REG_POSITION(vertex_array[0], 0x700);
ASSERT_REG_POSITION(independent_blend, 0x780);

View File

@@ -5,11 +5,12 @@
#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"
@@ -577,10 +578,6 @@ union Instruction {
BitField<55, 1, u64> saturate;
} fmul32;
union {
BitField<52, 1, u64> generates_cc;
} op_32;
union {
BitField<48, 1, u64> is_signed;
} shift;
@@ -1235,7 +1232,6 @@ union Instruction {
BitField<60, 1, u64> is_b_gpr;
BitField<59, 1, u64> is_c_gpr;
BitField<20, 24, s64> smem_imm;
BitField<0, 5, ControlCode> flow_control_code;
Attribute attribute;
Sampler sampler;
@@ -1460,7 +1456,7 @@ public:
Type type;
};
static std::optional<std::reference_wrapper<const Matcher>> Decode(Instruction instr) {
static boost::optional<const Matcher&> Decode(Instruction instr) {
static const auto table{GetDecodeTable()};
const auto matches_instruction = [instr](const auto& matcher) {
@@ -1468,8 +1464,7 @@ public:
};
auto iter = std::find_if(table.begin(), table.end(), matches_instruction);
return iter != table.end() ? std::optional<std::reference_wrapper<const Matcher>>(*iter)
: std::nullopt;
return iter != table.end() ? boost::optional<const Matcher&>(*iter) : boost::none;
}
private:
@@ -1663,4 +1658,4 @@ private:
}
};
} // namespace Tegra::Shader
} // namespace Tegra::Shader

View File

@@ -11,7 +11,7 @@ namespace Tegra {
MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {}
void MacroInterpreter::Execute(u32 offset, std::vector<u32> parameters) {
void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> parameters) {
Reset();
registers[1] = parameters[0];
this->parameters = std::move(parameters);
@@ -19,7 +19,7 @@ void MacroInterpreter::Execute(u32 offset, std::vector<u32> parameters) {
// Execute the code until we hit an exit condition.
bool keep_executing = true;
while (keep_executing) {
keep_executing = Step(offset, false);
keep_executing = Step(code, false);
}
// Assert the the macro used all the input parameters
@@ -29,7 +29,7 @@ void MacroInterpreter::Execute(u32 offset, std::vector<u32> parameters) {
void MacroInterpreter::Reset() {
registers = {};
pc = 0;
delayed_pc = {};
delayed_pc = boost::none;
method_address.raw = 0;
parameters.clear();
// The next parameter index starts at 1, because $r1 already has the value of the first
@@ -37,17 +37,17 @@ void MacroInterpreter::Reset() {
next_parameter_index = 1;
}
bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) {
u32 base_address = pc;
Opcode opcode = GetOpcode(offset);
Opcode opcode = GetOpcode(code);
pc += 4;
// Update the program counter if we were delayed
if (delayed_pc) {
if (delayed_pc != boost::none) {
ASSERT(is_delay_slot);
pc = *delayed_pc;
delayed_pc = {};
delayed_pc = boost::none;
}
switch (opcode.operation) {
@@ -108,7 +108,7 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
delayed_pc = base_address + opcode.GetBranchTarget();
// Execute one more instruction due to the delay slot.
return Step(offset, true);
return Step(code, true);
}
break;
}
@@ -121,18 +121,17 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
// Exit has a delay slot, execute the next instruction
// Note: Executing an exit during a branch delay slot will cause the instruction at the
// branch target to be executed before exiting.
Step(offset, true);
Step(code, true);
return false;
}
return true;
}
MacroInterpreter::Opcode MacroInterpreter::GetOpcode(u32 offset) const {
const auto& macro_memory{maxwell3d.GetMacroMemory()};
MacroInterpreter::Opcode MacroInterpreter::GetOpcode(const std::vector<u32>& code) const {
ASSERT((pc % sizeof(u32)) == 0);
ASSERT((pc + offset) < macro_memory.size() * sizeof(u32));
return {macro_memory[offset + pc / sizeof(u32)]};
ASSERT(pc < code.size() * sizeof(u32));
return {code[pc / sizeof(u32)]};
}
u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const {

View File

@@ -5,9 +5,8 @@
#pragma once
#include <array>
#include <optional>
#include <vector>
#include <boost/optional.hpp>
#include "common/bit_field.h"
#include "common/common_types.h"
@@ -22,10 +21,10 @@ public:
/**
* Executes the macro code with the specified input parameters.
* @param offset Offset to start execution at.
* @param parameters The parameters of the macro.
* @param code The macro byte code to execute
* @param parameters The parameters of the macro
*/
void Execute(u32 offset, std::vector<u32> parameters);
void Execute(const std::vector<u32>& code, std::vector<u32> parameters);
private:
enum class Operation : u32 {
@@ -110,11 +109,11 @@ private:
/**
* Executes a single macro instruction located at the current program counter. Returns whether
* the interpreter should keep running.
* @param offset Offset to start execution at.
* @param code The macro code to execute.
* @param is_delay_slot Whether the current step is being executed due to a delay slot in a
* previous instruction.
*/
bool Step(u32 offset, bool is_delay_slot);
bool Step(const std::vector<u32>& code, bool is_delay_slot);
/// Calculates the result of an ALU operation. src_a OP src_b;
u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const;
@@ -127,7 +126,7 @@ private:
bool EvaluateBranchCondition(BranchCondition cond, u32 value) const;
/// Reads an opcode at the current program counter location.
Opcode GetOpcode(u32 offset) const;
Opcode GetOpcode(const std::vector<u32>& code) const;
/// Returns the specified register's value. Register 0 is hardcoded to always return 0.
u32 GetRegister(u32 register_id) const;
@@ -150,7 +149,7 @@ private:
Engines::Maxwell3D& maxwell3d;
u32 pc; ///< Current program counter
std::optional<u32>
boost::optional<u32>
delayed_pc; ///< Program counter to execute at after the delay slot is executed.
static constexpr std::size_t NumMacroRegisters = 8;

View File

@@ -4,21 +4,18 @@
#include "common/alignment.h"
#include "common/assert.h"
#include "common/logging/log.h"
#include "video_core/memory_manager.h"
namespace Tegra {
GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, align, PageStatus::Unmapped)};
boost::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, align);
ASSERT(gpu_addr);
ASSERT_MSG(gpu_addr, "unable to find available GPU memory");
for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
VAddr& slot{PageSlot(*gpu_addr + offset)};
for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
VAddr& slot = PageSlot(*gpu_addr + offset);
ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
slot = static_cast<u64>(PageStatus::Allocated);
}
@@ -26,11 +23,10 @@ GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
}
GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) {
for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
VAddr& slot{PageSlot(gpu_addr + offset)};
for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
VAddr& slot = PageSlot(gpu_addr + offset);
ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
slot = static_cast<u64>(PageStatus::Allocated);
}
@@ -38,19 +34,17 @@ GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) {
}
GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) {
const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, PAGE_SIZE, PageStatus::Unmapped)};
boost::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, PAGE_SIZE);
ASSERT(gpu_addr);
ASSERT_MSG(gpu_addr, "unable to find available GPU memory");
for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
VAddr& slot{PageSlot(*gpu_addr + offset)};
for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
VAddr& slot = PageSlot(*gpu_addr + offset);
ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
slot = cpu_addr + offset;
}
const MappedRegion region{cpu_addr, *gpu_addr, size};
MappedRegion region{cpu_addr, *gpu_addr, size};
mapped_regions.push_back(region);
return *gpu_addr;
@@ -59,31 +53,14 @@ GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) {
GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size) {
ASSERT((gpu_addr & PAGE_MASK) == 0);
if (PageSlot(gpu_addr) != static_cast<u64>(PageStatus::Allocated)) {
// Page has been already mapped. In this case, we must find a new area of memory to use that
// is different than the specified one. Super Mario Odyssey hits this scenario when changing
// areas, but we do not want to overwrite the old pages.
// TODO(bunnei): We need to write a hardware test to confirm this behavior.
LOG_ERROR(HW_GPU, "attempting to map addr 0x{:016X}, which is not available!", gpu_addr);
const std::optional<GPUVAddr> new_gpu_addr{
FindFreeBlock(gpu_addr, size, PAGE_SIZE, PageStatus::Allocated)};
ASSERT_MSG(new_gpu_addr, "unable to find available GPU memory");
gpu_addr = *new_gpu_addr;
}
for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
VAddr& slot{PageSlot(gpu_addr + offset)};
for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
VAddr& slot = PageSlot(gpu_addr + offset);
ASSERT(slot == static_cast<u64>(PageStatus::Allocated));
slot = cpu_addr + offset;
}
const MappedRegion region{cpu_addr, gpu_addr, size};
MappedRegion region{cpu_addr, gpu_addr, size};
mapped_regions.push_back(region);
return gpu_addr;
@@ -92,12 +69,11 @@ GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size)
GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) {
ASSERT((gpu_addr & PAGE_MASK) == 0);
for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
VAddr& slot{PageSlot(gpu_addr + offset)};
for (u64 offset = 0; offset < size; offset += PAGE_SIZE) {
VAddr& slot = PageSlot(gpu_addr + offset);
ASSERT(slot != static_cast<u64>(PageStatus::Allocated) &&
slot != static_cast<u64>(PageStatus::Unmapped));
slot = static_cast<u64>(PageStatus::Unmapped);
}
@@ -121,14 +97,13 @@ GPUVAddr MemoryManager::GetRegionEnd(GPUVAddr region_start) const {
return {};
}
std::optional<GPUVAddr> MemoryManager::FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,
PageStatus status) {
GPUVAddr gpu_addr{region_start};
u64 free_space{};
boost::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) {
GPUVAddr gpu_addr = 0;
u64 free_space = 0;
align = (align + PAGE_MASK) & ~PAGE_MASK;
while (gpu_addr + free_space < MAX_ADDRESS) {
if (PageSlot(gpu_addr + free_space) == static_cast<u64>(status)) {
if (!IsPageMapped(gpu_addr + free_space)) {
free_space += PAGE_SIZE;
if (free_space >= size) {
return gpu_addr;
@@ -143,8 +118,8 @@ std::optional<GPUVAddr> MemoryManager::FindFreeBlock(GPUVAddr region_start, u64
return {};
}
std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) {
const VAddr base_addr{PageSlot(gpu_addr)};
boost::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) {
VAddr base_addr = PageSlot(gpu_addr);
if (base_addr == static_cast<u64>(PageStatus::Allocated) ||
base_addr == static_cast<u64>(PageStatus::Unmapped)) {
@@ -158,15 +133,19 @@ std::vector<GPUVAddr> MemoryManager::CpuToGpuAddress(VAddr cpu_addr) const {
std::vector<GPUVAddr> results;
for (const auto& region : mapped_regions) {
if (cpu_addr >= region.cpu_addr && cpu_addr < (region.cpu_addr + region.size)) {
const u64 offset{cpu_addr - region.cpu_addr};
u64 offset = cpu_addr - region.cpu_addr;
results.push_back(region.gpu_addr + offset);
}
}
return results;
}
bool MemoryManager::IsPageMapped(GPUVAddr gpu_addr) {
return PageSlot(gpu_addr) != static_cast<u64>(PageStatus::Unmapped);
}
VAddr& MemoryManager::PageSlot(GPUVAddr gpu_addr) {
auto& block{page_table[(gpu_addr >> (PAGE_BITS + PAGE_TABLE_BITS)) & PAGE_TABLE_MASK]};
auto& block = page_table[(gpu_addr >> (PAGE_BITS + PAGE_TABLE_BITS)) & PAGE_TABLE_MASK];
if (!block) {
block = std::make_unique<PageBlock>();
block->fill(static_cast<VAddr>(PageStatus::Unmapped));

View File

@@ -6,9 +6,10 @@
#include <array>
#include <memory>
#include <optional>
#include <vector>
#include <boost/optional.hpp>
#include "common/common_types.h"
namespace Tegra {
@@ -26,7 +27,7 @@ public:
GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size);
GPUVAddr UnmapBuffer(GPUVAddr gpu_addr, u64 size);
GPUVAddr GetRegionEnd(GPUVAddr region_start) const;
std::optional<VAddr> GpuToCpuAddress(GPUVAddr gpu_addr);
boost::optional<VAddr> GpuToCpuAddress(GPUVAddr gpu_addr);
std::vector<GPUVAddr> CpuToGpuAddress(VAddr cpu_addr) const;
static constexpr u64 PAGE_BITS = 16;
@@ -34,15 +35,15 @@ public:
static constexpr u64 PAGE_MASK = PAGE_SIZE - 1;
private:
boost::optional<GPUVAddr> FindFreeBlock(u64 size, u64 align = 1);
bool IsPageMapped(GPUVAddr gpu_addr);
VAddr& PageSlot(GPUVAddr gpu_addr);
enum class PageStatus : u64 {
Unmapped = 0xFFFFFFFFFFFFFFFFULL,
Allocated = 0xFFFFFFFFFFFFFFFEULL,
};
std::optional<GPUVAddr> FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,
PageStatus status);
VAddr& PageSlot(GPUVAddr gpu_addr);
static constexpr u64 MAX_ADDRESS{0x10000000000ULL};
static constexpr u64 PAGE_TABLE_BITS{10};
static constexpr u64 PAGE_TABLE_SIZE{1 << PAGE_TABLE_BITS};

View File

@@ -6,8 +6,7 @@
#include <atomic>
#include <memory>
#include <optional>
#include <boost/optional.hpp>
#include "common/common_types.h"
#include "video_core/gpu.h"
#include "video_core/rasterizer_interface.h"
@@ -29,8 +28,7 @@ public:
virtual ~RendererBase();
/// Swap buffers (render frame)
virtual void SwapBuffers(
std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) = 0;
virtual void SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) = 0;
/// Initialize the renderer
virtual bool Init() = 0;

View File

@@ -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 std::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)};
const boost::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.

View File

@@ -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 std::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)};
const boost::optional<VAddr> cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr)};
const u8* source{Memory::GetPointer(*cpu_addr)};
for (u32 primitive = 0; primitive < count / 4; ++primitive) {

View File

@@ -30,8 +30,8 @@
namespace OpenGL {
using Maxwell = Tegra::Engines::Maxwell3D::Regs;
using PixelFormat = VideoCore::Surface::PixelFormat;
using SurfaceType = VideoCore::Surface::SurfaceType;
using PixelFormat = SurfaceParams::PixelFormat;
using SurfaceType = SurfaceParams::SurfaceType;
MICROPROFILE_DEFINE(OpenGL_VAO, "OpenGL", "Vertex Array Setup", MP_RGB(128, 128, 192));
MICROPROFILE_DEFINE(OpenGL_Shader, "OpenGL", "Shader Setup", MP_RGB(128, 128, 192));
@@ -104,7 +104,7 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
}
ASSERT_MSG(has_ARB_separate_shader_objects, "has_ARB_separate_shader_objects is unsupported");
OpenGLState::ApplyDefaultState();
// Clipping plane 0 is always enabled for PICA fixed clip plane z <= 0
state.clip_distance[0] = true;
@@ -115,6 +115,8 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
state.draw.shader_program = 0;
state.Apply();
glEnable(GL_BLEND);
glGetIntegerv(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT, &uniform_buffer_alignment);
LOG_CRITICAL(Render_OpenGL, "Sync fixed function OpenGL state here!");
@@ -399,7 +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,
std::optional<std::size_t> single_color_target) {
boost::optional<std::size_t> single_color_target) {
MICROPROFILE_SCOPE(OpenGL_Framebuffer);
const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
@@ -511,10 +513,10 @@ void RasterizerOpenGL::Clear() {
OpenGLState clear_state;
clear_state.draw.draw_framebuffer = framebuffer.handle;
clear_state.color_mask[0].red_enabled = regs.clear_buffers.R ? GL_TRUE : GL_FALSE;
clear_state.color_mask[0].green_enabled = regs.clear_buffers.G ? GL_TRUE : GL_FALSE;
clear_state.color_mask[0].blue_enabled = regs.clear_buffers.B ? GL_TRUE : GL_FALSE;
clear_state.color_mask[0].alpha_enabled = regs.clear_buffers.A ? GL_TRUE : GL_FALSE;
clear_state.color_mask.red_enabled = regs.clear_buffers.R ? GL_TRUE : GL_FALSE;
clear_state.color_mask.green_enabled = regs.clear_buffers.G ? GL_TRUE : GL_FALSE;
clear_state.color_mask.blue_enabled = regs.clear_buffers.B ? GL_TRUE : GL_FALSE;
clear_state.color_mask.alpha_enabled = regs.clear_buffers.A ? GL_TRUE : GL_FALSE;
if (regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B ||
regs.clear_buffers.A) {
@@ -573,13 +575,14 @@ void RasterizerOpenGL::DrawArrays() {
ScopeAcquireGLContext acquire_context{emu_window};
ConfigureFramebuffers();
SyncColorMask();
SyncDepthTestState();
SyncStencilTestState();
SyncBlendState();
SyncLogicOpState();
SyncCullMode();
SyncPrimitiveRestart();
SyncDepthRange();
SyncScissorTest();
// Alpha Testing is synced on shaders.
SyncTransformFeedback();
@@ -700,8 +703,7 @@ bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config,
// Verify that the cached surface is the same size and format as the requested framebuffer
const auto& params{surface->GetSurfaceParams()};
const auto& pixel_format{
VideoCore::Surface::PixelFormatFromGPUPixelFormat(config.pixel_format)};
const auto& pixel_format{SurfaceParams::PixelFormatFromGPUPixelFormat(config.pixel_format)};
ASSERT_MSG(params.width == config.width, "Framebuffer width is different");
ASSERT_MSG(params.height == config.height, "Framebuffer height is different");
ASSERT_MSG(params.pixel_format == pixel_format, "Framebuffer pixel_format is different");
@@ -729,15 +731,11 @@ 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, Tegra::Texture::TextureMipmapFilter::None));
glSamplerParameteri(s, GL_TEXTURE_MAG_FILTER, MaxwellToGL::TextureFilterMode(mag_filter));
}
if (min_filter != config.min_filter || mip_filter != config.mip_filter) {
if (min_filter != config.min_filter) {
min_filter = config.min_filter;
mip_filter = config.mip_filter;
glSamplerParameteri(s, GL_TEXTURE_MIN_FILTER,
MaxwellToGL::TextureFilterMode(min_filter, mip_filter));
glSamplerParameteri(s, GL_TEXTURE_MIN_FILTER, MaxwellToGL::TextureFilterMode(min_filter));
}
if (wrap_u != config.wrap_u) {
@@ -898,16 +896,12 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
void RasterizerOpenGL::SyncViewport() {
const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()};
auto& viewport = state.viewports[i];
viewport.x = viewport_rect.left;
viewport.y = viewport_rect.bottom;
viewport.width = static_cast<GLfloat>(viewport_rect.GetWidth());
viewport.height = static_cast<GLfloat>(viewport_rect.GetHeight());
viewport.depth_range_far = regs.viewport[i].depth_range_far;
viewport.depth_range_near = regs.viewport[i].depth_range_near;
}
const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()};
state.viewport.x = viewport_rect.left;
state.viewport.y = viewport_rect.bottom;
state.viewport.width = static_cast<GLsizei>(viewport_rect.GetWidth());
state.viewport.height = static_cast<GLsizei>(viewport_rect.GetHeight());
}
void RasterizerOpenGL::SyncClipEnabled() {
@@ -949,6 +943,13 @@ void RasterizerOpenGL::SyncPrimitiveRestart() {
state.primitive_restart.index = regs.primitive_restart.index;
}
void RasterizerOpenGL::SyncDepthRange() {
const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
state.depth.depth_range_near = regs.viewport->depth_range_near;
state.depth.depth_range_far = regs.viewport->depth_range_far;
}
void RasterizerOpenGL::SyncDepthTestState() {
const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
@@ -989,60 +990,26 @@ void RasterizerOpenGL::SyncStencilTestState() {
state.stencil.back.write_mask = regs.stencil_back_mask;
}
void RasterizerOpenGL::SyncColorMask() {
const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
const auto& source = regs.color_mask[regs.color_mask_common ? 0 : i];
auto& dest = state.color_mask[i];
dest.red_enabled = (source.R == 0) ? GL_FALSE : GL_TRUE;
dest.green_enabled = (source.G == 0) ? GL_FALSE : GL_TRUE;
dest.blue_enabled = (source.B == 0) ? GL_FALSE : GL_TRUE;
dest.alpha_enabled = (source.A == 0) ? GL_FALSE : GL_TRUE;
}
}
void RasterizerOpenGL::SyncBlendState() {
const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
state.blend_color.red = regs.blend_color.r;
state.blend_color.green = regs.blend_color.g;
state.blend_color.blue = regs.blend_color.b;
state.blend_color.alpha = regs.blend_color.a;
// TODO(Subv): Support more than just render target 0.
state.blend.enabled = regs.blend.enable[0] != 0;
state.independant_blend.enabled = regs.independent_blend_enable;
if (!state.independant_blend.enabled) {
auto& blend = state.blend[0];
blend.enabled = regs.blend.enable[0] != 0;
blend.separate_alpha = regs.blend.separate_alpha;
blend.rgb_equation = MaxwellToGL::BlendEquation(regs.blend.equation_rgb);
blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_rgb);
blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_rgb);
if (blend.separate_alpha) {
blend.a_equation = MaxwellToGL::BlendEquation(regs.blend.equation_a);
blend.src_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_a);
blend.dst_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_a);
}
for (size_t i = 1; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
state.blend[i].enabled = false;
}
if (!state.blend.enabled)
return;
}
for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
auto& blend = state.blend[i];
blend.enabled = regs.blend.enable[i] != 0;
if (!blend.enabled)
continue;
blend.separate_alpha = regs.independent_blend[i].separate_alpha;
blend.rgb_equation = MaxwellToGL::BlendEquation(regs.independent_blend[i].equation_rgb);
blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_source_rgb);
blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_dest_rgb);
if (blend.separate_alpha) {
blend.a_equation = MaxwellToGL::BlendEquation(regs.independent_blend[i].equation_a);
blend.src_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_source_a);
blend.dst_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_dest_a);
}
}
ASSERT_MSG(regs.logic_op.enable == 0,
"Blending and logic op can't be enabled at the same time.");
ASSERT_MSG(regs.independent_blend_enable == 1, "Only independent blending is implemented");
ASSERT_MSG(!regs.independent_blend[0].separate_alpha, "Unimplemented");
state.blend.rgb_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_rgb);
state.blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_rgb);
state.blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_rgb);
state.blend.a_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_a);
state.blend.src_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_a);
state.blend.dst_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_a);
}
void RasterizerOpenGL::SyncLogicOpState() {
@@ -1061,19 +1028,19 @@ void RasterizerOpenGL::SyncLogicOpState() {
}
void RasterizerOpenGL::SyncScissorTest() {
// TODO: what is the correct behavior here, a single scissor for all targets
// or scissor disabled for the rest of the targets?
const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
state.scissor.enabled = (regs.scissor_test.enable != 0);
if (regs.scissor_test.enable == 0) {
return;
// TODO(Blinkhawk): Figure if the hardware supports scissor testing per viewport and how it's
// implemented.
if (regs.scissor_test.enable != 0) {
const u32 width = regs.scissor_test.max_x - regs.scissor_test.min_x;
const u32 height = regs.scissor_test.max_y - regs.scissor_test.min_y;
state.scissor.x = regs.scissor_test.min_x;
state.scissor.y = regs.scissor_test.min_y;
state.scissor.width = width;
state.scissor.height = height;
}
const u32 width = regs.scissor_test.max_x - regs.scissor_test.min_x;
const u32 height = regs.scissor_test.max_y - regs.scissor_test.min_y;
state.scissor.x = regs.scissor_test.min_x;
state.scissor.y = regs.scissor_test.min_y;
state.scissor.width = width;
state.scissor.height = height;
}
void RasterizerOpenGL::SyncTransformFeedback() {

View File

@@ -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>
@@ -93,7 +93,6 @@ 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;
@@ -111,7 +110,7 @@ private:
*/
void ConfigureFramebuffers(bool use_color_fb = true, bool using_depth_fb = true,
bool preserve_contents = true,
std::optional<std::size_t> single_color_target = {});
boost::optional<std::size_t> single_color_target = {});
/*
* Configures the current constbuffers to use for the draw command.
@@ -133,7 +132,7 @@ private:
u32 SetupTextures(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, Shader& shader,
GLenum primitive_mode, u32 current_unit);
/// Syncs the viewport and depth range to match the guest state
/// Syncs the viewport to match the guest state
void SyncViewport();
/// Syncs the clip enabled status to match the guest state
@@ -148,6 +147,9 @@ private:
/// Syncs the primitve restart to match the guest state
void SyncPrimitiveRestart();
/// Syncs the depth range to match the guest state
void SyncDepthRange();
/// Syncs the depth test state to match the guest state
void SyncDepthTestState();
@@ -169,9 +171,6 @@ private:
/// Syncs the point state to match the guest state
void SyncPointState();
/// Syncs Color Mask
void SyncColorMask();
/// Check asserts for alpha testing.
void CheckAlphaTests();

View File

@@ -16,22 +16,15 @@
#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/gl_state.h"
#include "video_core/renderer_opengl/utils.h"
#include "video_core/surface.h"
#include "video_core/textures/astc.h"
#include "video_core/textures/decoders.h"
#include "video_core/utils.h"
namespace OpenGL {
using VideoCore::Surface::ComponentTypeFromDepthFormat;
using VideoCore::Surface::ComponentTypeFromRenderTarget;
using VideoCore::Surface::ComponentTypeFromTexture;
using VideoCore::Surface::PixelFormatFromDepthFormat;
using VideoCore::Surface::PixelFormatFromRenderTargetFormat;
using VideoCore::Surface::PixelFormatFromTextureFormat;
using VideoCore::Surface::SurfaceTargetFromTextureType;
using SurfaceType = SurfaceParams::SurfaceType;
using PixelFormat = SurfaceParams::PixelFormat;
using ComponentType = SurfaceParams::ComponentType;
struct FormatTuple {
GLint internal_format;
@@ -41,6 +34,46 @@ struct FormatTuple {
bool compressed;
};
static bool IsPixelFormatASTC(PixelFormat format) {
switch (format) {
case PixelFormat::ASTC_2D_4X4:
case PixelFormat::ASTC_2D_5X4:
case PixelFormat::ASTC_2D_8X8:
case PixelFormat::ASTC_2D_8X5:
case PixelFormat::ASTC_2D_4X4_SRGB:
case PixelFormat::ASTC_2D_5X4_SRGB:
case PixelFormat::ASTC_2D_8X8_SRGB:
case PixelFormat::ASTC_2D_8X5_SRGB:
return true;
default:
return false;
}
}
static std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) {
switch (format) {
case PixelFormat::ASTC_2D_4X4:
return {4, 4};
case PixelFormat::ASTC_2D_5X4:
return {5, 4};
case PixelFormat::ASTC_2D_8X8:
return {8, 8};
case PixelFormat::ASTC_2D_8X5:
return {8, 5};
case PixelFormat::ASTC_2D_4X4_SRGB:
return {4, 4};
case PixelFormat::ASTC_2D_5X4_SRGB:
return {5, 4};
case PixelFormat::ASTC_2D_8X8_SRGB:
return {8, 8};
case PixelFormat::ASTC_2D_8X5_SRGB:
return {8, 5};
default:
LOG_CRITICAL(HW_GPU, "Unhandled format: {}", static_cast<u32>(format));
UNREACHABLE();
}
}
void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
const auto cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr_)};
@@ -57,34 +90,27 @@ void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
}
}
std::size_t SurfaceParams::InnerMipmapMemorySize(u32 mip_level, bool force_gl, bool layer_only,
bool uncompressed) const {
const u32 tile_x{GetDefaultBlockWidth(pixel_format)};
const u32 tile_y{GetDefaultBlockHeight(pixel_format)};
std::size_t SurfaceParams::InnerMemorySize(bool layer_only) 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 = MipWidth(mip_level);
u32 m_height = MipHeight(mip_level);
m_width = uncompressed ? m_width : std::max(1U, (m_width + tile_x - 1) / tile_x);
m_height = uncompressed ? m_height : std::max(1U, (m_height + tile_y - 1) / tile_y);
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);
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);
}
if (!force_gl && is_tiled) {
size = Common::AlignUp(size, block_size_bytes);
}
return size;
return is_tiled ? Common::AlignUp(size, block_size_bytes) : size;
}
/*static*/ SurfaceParams SurfaceParams::CreateForTexture(
@@ -127,13 +153,6 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
params.target = SurfaceTarget::Texture2D;
}
break;
case SurfaceTarget::TextureCubeArray:
params.depth = config.tic.Depth() * 6;
if (!entry.IsArray()) {
ASSERT(params.depth == 6);
params.target = SurfaceTarget::TextureCubemap;
}
break;
default:
LOG_CRITICAL(HW_GPU, "Unknown depth for target={}", static_cast<u32>(params.target));
UNREACHABLE();
@@ -169,7 +188,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
params.unaligned_height = config.height;
params.target = SurfaceTarget::Texture2D;
params.depth = 1;
params.max_mip_level = 1;
params.max_mip_level = 0;
params.is_layered = false;
// Render target specific parameters, not used for caching
@@ -203,7 +222,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
params.unaligned_height = zeta_height;
params.target = SurfaceTarget::Texture2D;
params.depth = 1;
params.max_mip_level = 1;
params.max_mip_level = 0;
params.is_layered = false;
params.rt = {};
@@ -230,7 +249,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
params.unaligned_height = config.height;
params.target = SurfaceTarget::Texture2D;
params.depth = 1;
params.max_mip_level = 1;
params.max_mip_level = 0;
params.rt = {};
params.InitCacheParameters(config.Address());
@@ -238,7 +257,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
return params;
}
static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format_tuples = {{
static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{
{GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // ABGR8U
{GL_RGBA8, GL_RGBA, GL_BYTE, ComponentType::SNorm, false}, // ABGR8S
{GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE, ComponentType::UInt, false}, // ABGR8UI
@@ -311,8 +330,6 @@ static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex
{GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X8_SRGB
{GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X5_SRGB
{GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X4_SRGB
{GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X5
{GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X5_SRGB
// Depth formats
{GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F
@@ -328,22 +345,20 @@ static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex
ComponentType::Float, false}, // Z32FS8
}};
static GLenum SurfaceTargetToGL(SurfaceTarget target) {
static GLenum SurfaceTargetToGL(SurfaceParams::SurfaceTarget target) {
switch (target) {
case SurfaceTarget::Texture1D:
case SurfaceParams::SurfaceTarget::Texture1D:
return GL_TEXTURE_1D;
case SurfaceTarget::Texture2D:
case SurfaceParams::SurfaceTarget::Texture2D:
return GL_TEXTURE_2D;
case SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture3D:
return GL_TEXTURE_3D;
case SurfaceTarget::Texture1DArray:
case SurfaceParams::SurfaceTarget::Texture1DArray:
return GL_TEXTURE_1D_ARRAY;
case SurfaceTarget::Texture2DArray:
case SurfaceParams::SurfaceTarget::Texture2DArray:
return GL_TEXTURE_2D_ARRAY;
case SurfaceTarget::TextureCubemap:
case SurfaceParams::SurfaceTarget::TextureCubemap:
return GL_TEXTURE_CUBE_MAP;
case SurfaceTarget::TextureCubeArray:
return GL_TEXTURE_CUBE_MAP_ARRAY_ARB;
}
LOG_CRITICAL(Render_OpenGL, "Unimplemented texture target={}", static_cast<u32>(target));
UNREACHABLE();
@@ -358,41 +373,59 @@ static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType
return format;
}
MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
u32 actual_height{std::max(1U, unaligned_height >> mip_level)};
MathUtil::Rectangle<u32> SurfaceParams::GetRect() const {
u32 actual_height{unaligned_height};
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, MipWidth(mip_level), 0};
return {0, actual_height, width, 0};
}
/// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN
static bool IsFormatBCn(PixelFormat format) {
switch (format) {
case PixelFormat::DXT1:
case PixelFormat::DXT23:
case PixelFormat::DXT45:
case PixelFormat::DXN1:
case PixelFormat::DXN2SNORM:
case PixelFormat::DXN2UNORM:
case PixelFormat::BC7U:
case PixelFormat::BC6H_UF16:
case PixelFormat::BC6H_SF16:
case PixelFormat::DXT1_SRGB:
case PixelFormat::DXT23_SRGB:
case PixelFormat::DXT45_SRGB:
case PixelFormat::BC7U_SRGB:
return true;
}
return false;
}
template <bool morton_to_gl, PixelFormat format>
void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth, u8* gl_buffer,
std::size_t gl_buffer_size, VAddr addr) {
constexpr u32 bytes_per_pixel = GetBytesPerPixel(format);
constexpr u32 bytes_per_pixel = SurfaceParams::GetBytesPerPixel(format);
// With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
// pixel values.
const u32 tile_size_x{GetDefaultBlockWidth(format)};
const u32 tile_size_y{GetDefaultBlockHeight(format)};
const u32 tile_size{IsFormatBCn(format) ? 4U : 1U};
if (morton_to_gl) {
const std::vector<u8> data =
Tegra::Texture::UnswizzleTexture(addr, tile_size_x, tile_size_y, bytes_per_pixel,
stride, height, depth, block_height, block_depth);
const std::vector<u8> data = Tegra::Texture::UnswizzleTexture(
addr, tile_size, bytes_per_pixel, stride, height, depth, block_height, block_depth);
const std::size_t size_to_copy{std::min(gl_buffer_size, data.size())};
memcpy(gl_buffer, data.data(), size_to_copy);
} else {
Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x,
(height + tile_size_y - 1) / tile_size_y, depth,
Tegra::Texture::CopySwizzledData(stride / tile_size, height / tile_size, depth,
bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr),
gl_buffer, false, block_height, block_depth);
}
}
using GLConversionArray = std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
VideoCore::Surface::MaxPixelFormat>;
SurfaceParams::MaxPixelFormat>;
static constexpr GLConversionArray morton_to_gl_fns = {
// clang-format off
@@ -453,8 +486,6 @@ static constexpr GLConversionArray morton_to_gl_fns = {
MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>,
MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>,
MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>,
MortonCopy<true, PixelFormat::ASTC_2D_5X5>,
MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>,
MortonCopy<true, PixelFormat::Z32F>,
MortonCopy<true, PixelFormat::Z16>,
MortonCopy<true, PixelFormat::Z24S8>,
@@ -523,8 +554,6 @@ static constexpr GLConversionArray gl_to_morton_fns = {
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
MortonCopy<false, PixelFormat::Z32F>,
MortonCopy<false, PixelFormat::Z16>,
MortonCopy<false, PixelFormat::Z24S8>,
@@ -534,39 +563,34 @@ static constexpr GLConversionArray gl_to_morton_fns = {
};
void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params,
std::vector<u8>& gl_buffer, u32 mip_level) {
u32 depth = params.MipDepth(mip_level);
if (params.target == SurfaceTarget::Texture2D) {
std::vector<u8>& gl_buffer) {
u32 depth = params.depth;
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 = params.GetMipmapLevelOffset(mip_level);
u64 offset = 0;
u64 offset_gl = 0;
u64 layer_size = params.LayerMemorySize();
u64 gl_size = params.LayerSizeGL(mip_level);
for (u32 i = 0; i < params.depth; i++) {
u64 gl_size = params.LayerSizeGL();
for (u32 i = 0; i < depth; i++) {
functions[static_cast<std::size_t>(params.pixel_format)](
params.MipWidth(mip_level), params.MipBlockHeight(mip_level),
params.MipHeight(mip_level), params.MipBlockDepth(mip_level), 1,
params.width, params.block_height, params.height, params.block_depth, 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.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);
params.width, params.block_height, params.height, params.block_depth, depth,
gl_buffer.data(), gl_buffer.size(), params.addr);
}
}
MICROPROFILE_DEFINE(OpenGL_BlitSurface, "OpenGL", "BlitSurface", MP_RGB(128, 192, 64));
static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
GLuint read_fb_handle, GLuint draw_fb_handle, GLenum src_attachment = 0,
GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
MICROPROFILE_SCOPE(OpenGL_BlitSurface);
const auto& src_params{src_surface->GetSurfaceParams()};
const auto& dst_params{dst_surface->GetSurfaceParams()};
@@ -585,13 +609,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
if (src_params.type == SurfaceType::ColorTexture) {
switch (src_params.target) {
case SurfaceTarget::Texture2D:
case SurfaceParams::SurfaceTarget::Texture2D:
glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
GL_TEXTURE_2D, src_surface->Texture().handle, 0);
glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
0, 0);
break;
case SurfaceTarget::TextureCubemap:
case SurfaceParams::SurfaceTarget::TextureCubemap:
glFramebufferTexture2D(
GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
@@ -600,12 +624,12 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
break;
case SurfaceTarget::Texture2DArray:
case SurfaceParams::SurfaceTarget::Texture2DArray:
glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
src_surface->Texture().handle, 0, 0);
glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
break;
case SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture3D:
glFramebufferTexture3D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
SurfaceTargetToGL(src_params.target),
src_surface->Texture().handle, 0, 0);
@@ -621,13 +645,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
}
switch (dst_params.target) {
case SurfaceTarget::Texture2D:
case SurfaceParams::SurfaceTarget::Texture2D:
glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
GL_TEXTURE_2D, dst_surface->Texture().handle, 0);
glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
0, 0);
break;
case SurfaceTarget::TextureCubemap:
case SurfaceParams::SurfaceTarget::TextureCubemap:
glFramebufferTexture2D(
GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
@@ -636,13 +660,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
break;
case SurfaceTarget::Texture2DArray:
case SurfaceParams::SurfaceTarget::Texture2DArray:
glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
dst_surface->Texture().handle, 0, 0);
glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
break;
case SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture3D:
glFramebufferTexture3D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
SurfaceTargetToGL(dst_params.target),
dst_surface->Texture().handle, 0, 0);
@@ -706,11 +730,9 @@ static void FastCopySurface(const Surface& src_surface, const Surface& dst_surfa
0, 0, width, height, 1);
}
MICROPROFILE_DEFINE(OpenGL_CopySurface, "OpenGL", "CopySurface", MP_RGB(128, 192, 64));
static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
GLuint copy_pbo_handle, GLenum src_attachment = 0,
GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
MICROPROFILE_SCOPE(OpenGL_CopySurface);
ASSERT_MSG(dst_attachment == 0, "Unimplemented");
const auto& src_params{src_surface->GetSurfaceParams()};
@@ -765,22 +787,21 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
UNREACHABLE();
} else {
switch (dst_params.target) {
case SurfaceTarget::Texture1D:
case SurfaceParams::SurfaceTarget::Texture1D:
glTextureSubImage1D(dst_surface->Texture().handle, 0, 0, width, dest_format.format,
dest_format.type, nullptr);
break;
case SurfaceTarget::Texture2D:
case SurfaceParams::SurfaceTarget::Texture2D:
glTextureSubImage2D(dst_surface->Texture().handle, 0, 0, 0, width, height,
dest_format.format, dest_format.type, nullptr);
break;
case SurfaceTarget::Texture3D:
case SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubeArray:
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture2DArray:
glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0, 0, width, height,
static_cast<GLsizei>(dst_params.depth), dest_format.format,
dest_format.type, nullptr);
break;
case SurfaceTarget::TextureCubemap:
case SurfaceParams::SurfaceTarget::TextureCubemap:
glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0,
static_cast<GLint>(cubemap_face), width, height, 1,
dest_format.format, dest_format.type, nullptr);
@@ -817,43 +838,35 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
if (!format_tuple.compressed) {
// Only pre-create the texture for non-compressed textures.
switch (params.target) {
case SurfaceTarget::Texture1D:
glTexStorage1D(SurfaceTargetToGL(params.target), params.max_mip_level,
format_tuple.internal_format, rect.GetWidth());
case SurfaceParams::SurfaceTarget::Texture1D:
glTexStorage1D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format,
rect.GetWidth());
break;
case SurfaceTarget::Texture2D:
case SurfaceTarget::TextureCubemap:
glTexStorage2D(SurfaceTargetToGL(params.target), params.max_mip_level,
format_tuple.internal_format, rect.GetWidth(), rect.GetHeight());
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceParams::SurfaceTarget::TextureCubemap:
glTexStorage2D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format,
rect.GetWidth(), rect.GetHeight());
break;
case SurfaceTarget::Texture3D:
case SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubeArray:
glTexStorage3D(SurfaceTargetToGL(params.target), params.max_mip_level,
format_tuple.internal_format, rect.GetWidth(), rect.GetHeight(),
params.depth);
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture2DArray:
glTexStorage3D(SurfaceTargetToGL(params.target), 1, 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, params.max_mip_level, format_tuple.internal_format,
rect.GetWidth(), rect.GetHeight());
glTexStorage2D(GL_TEXTURE_2D, 1, 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);
}
LabelGLObject(GL_TEXTURE, texture.handle, params.addr,
SurfaceParams::SurfaceTargetName(params.target));
VideoCore::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
@@ -883,7 +896,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height, bo
S8Z24 s8z24_pixel{};
Z24S8 z24s8_pixel{};
constexpr auto bpp{GetBytesPerPixel(PixelFormat::S8Z24)};
constexpr auto bpp{SurfaceParams::GetBytesPerPixel(PixelFormat::S8Z24)};
for (std::size_t y = 0; y < height; ++y) {
for (std::size_t x = 0; x < width; ++x) {
const std::size_t offset{bpp * (y * width + x)};
@@ -903,7 +916,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height, bo
}
static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
constexpr auto bpp{GetBytesPerPixel(PixelFormat::G8R8U)};
constexpr auto bpp{SurfaceParams::GetBytesPerPixel(PixelFormat::G8R8U)};
for (std::size_t y = 0; y < height; ++y) {
for (std::size_t x = 0; x < width; ++x) {
const std::size_t offset{bpp * (y * width + x)};
@@ -920,24 +933,21 @@ static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
* typical desktop GPUs.
*/
static void ConvertFormatAsNeeded_LoadGLBuffer(std::vector<u8>& data, PixelFormat pixel_format,
u32 width, u32 height, u32 depth) {
u32 width, u32 height) {
switch (pixel_format) {
case PixelFormat::ASTC_2D_4X4:
case PixelFormat::ASTC_2D_8X8:
case PixelFormat::ASTC_2D_8X5:
case PixelFormat::ASTC_2D_5X4:
case PixelFormat::ASTC_2D_5X5:
case PixelFormat::ASTC_2D_4X4_SRGB:
case PixelFormat::ASTC_2D_8X8_SRGB:
case PixelFormat::ASTC_2D_8X5_SRGB:
case PixelFormat::ASTC_2D_5X4_SRGB:
case PixelFormat::ASTC_2D_5X5_SRGB: {
case PixelFormat::ASTC_2D_5X4_SRGB: {
// Convert ASTC pixel formats to RGBA8, as most desktop GPUs do not support ASTC.
u32 block_width{};
u32 block_height{};
std::tie(block_width, block_height) = GetASTCBlockSize(pixel_format);
data =
Tegra::Texture::ASTC::Decompress(data, width, height, depth, block_width, block_height);
data = Tegra::Texture::ASTC::Decompress(data, width, height, block_width, block_height);
break;
}
case PixelFormat::S8Z24:
@@ -979,25 +989,23 @@ static void ConvertFormatAsNeeded_FlushGLBuffer(std::vector<u8>& data, PixelForm
}
}
MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 192, 64));
MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 64, 192));
void CachedSurface::LoadGLBuffer() {
MICROPROFILE_SCOPE(OpenGL_SurfaceLoad);
gl_buffer.resize(params.max_mip_level);
for (u32 i = 0; i < params.max_mip_level; i++)
gl_buffer[i].resize(params.GetMipmapSizeGL(i));
gl_buffer.resize(params.size_in_bytes_gl);
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));
for (u32 i = 0; i < params.max_mip_level; i++)
SwizzleFunc(morton_to_gl_fns, params, gl_buffer[i], i);
SwizzleFunc(morton_to_gl_fns, params, gl_buffer);
} 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[0].assign(texture_src_data, texture_src_data_end);
gl_buffer.assign(texture_src_data, texture_src_data_end);
}
for (u32 i = 0; i < params.max_mip_level; i++)
ConvertFormatAsNeeded_LoadGLBuffer(gl_buffer[i], params.pixel_format, params.MipWidth(i),
params.MipHeight(i), params.MipDepth(i));
ConvertFormatAsNeeded_LoadGLBuffer(gl_buffer, params.pixel_format, params.width, params.height);
}
MICROPROFILE_DEFINE(OpenGL_SurfaceFlush, "OpenGL", "Surface Flush", MP_RGB(128, 192, 64));
@@ -1007,19 +1015,18 @@ 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(1);
gl_buffer[0].resize(GetSizeInBytes());
gl_buffer.resize(GetSizeInBytes());
const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
// Ensure no bad interactions with GL_UNPACK_ALIGNMENT
ASSERT(params.width * GetBytesPerPixel(params.pixel_format) % 4 == 0);
ASSERT(params.width * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 == 0);
glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(params.width));
ASSERT(!tuple.compressed);
glBindBuffer(GL_PIXEL_PACK_BUFFER, 0);
glGetTextureImage(texture.handle, 0, tuple.format, tuple.type,
static_cast<GLsizei>(gl_buffer[0].size()), gl_buffer[0].data());
static_cast<GLsizei>(gl_buffer.size()), gl_buffer.data());
glPixelStorei(GL_PACK_ROW_LENGTH, 0);
ConvertFormatAsNeeded_FlushGLBuffer(gl_buffer[0], params.pixel_format, params.width,
ConvertFormatAsNeeded_FlushGLBuffer(gl_buffer, params.pixel_format, params.width,
params.height);
ASSERT(params.type != SurfaceType::Fill);
const u8* const texture_src_data = Memory::GetPointer(params.addr);
@@ -1028,23 +1035,28 @@ void CachedSurface::FlushGLBuffer() {
ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
params.block_width, static_cast<u32>(params.target));
SwizzleFunc(gl_to_morton_fns, params, gl_buffer[0], 0);
SwizzleFunc(gl_to_morton_fns, params, gl_buffer);
} else {
std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer[0].data(), GetSizeInBytes());
std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer.data(), GetSizeInBytes());
}
}
void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
GLuint draw_fb_handle) {
const auto& rect{params.GetRect(mip_map)};
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()};
// 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.MipWidth(mip_map) +
static_cast<std::size_t>(static_cast<std::size_t>(y0) * params.width +
static_cast<std::size_t>(x0)) *
GetBytesPerPixel(params.pixel_format);
SurfaceParams::GetBytesPerPixel(params.pixel_format);
const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
const GLuint target_tex = texture.handle;
@@ -1060,118 +1072,88 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
cur_state.Apply();
// Ensure no bad interactions with GL_UNPACK_ALIGNMENT
ASSERT(params.MipWidth(mip_map) * GetBytesPerPixel(params.pixel_format) % 4 == 0);
glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.MipWidth(mip_map)));
ASSERT(params.width * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 == 0);
glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.width));
GLsizei image_size = static_cast<GLsizei>(params.GetMipmapSizeGL(mip_map, false));
glActiveTexture(GL_TEXTURE0);
if (tuple.compressed) {
switch (params.target) {
case SurfaceTarget::Texture2D:
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]);
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]);
break;
case 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]);
case SurfaceParams::SurfaceTarget::Texture3D:
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]);
break;
case SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubeArray:
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 SurfaceTarget::TextureCubemap: {
GLsizei layer_size = static_cast<GLsizei>(params.LayerSizeGL(mip_map));
case SurfaceParams::SurfaceTarget::TextureCubemap:
for (std::size_t face = 0; face < params.depth; ++face) {
glCompressedTexImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face),
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;
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();
}
break;
}
default:
LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
static_cast<u32>(params.target));
UNREACHABLE();
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]);
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]);
}
} else {
switch (params.target) {
case SurfaceTarget::Texture1D:
glTexSubImage1D(SurfaceTargetToGL(params.target), mip_map, x0,
case SurfaceParams::SurfaceTarget::Texture1D:
glTexSubImage1D(SurfaceTargetToGL(params.target), 0, x0,
static_cast<GLsizei>(rect.GetWidth()), tuple.format, tuple.type,
&gl_buffer[mip_map][buffer_offset]);
&gl_buffer[buffer_offset]);
break;
case SurfaceTarget::Texture2D:
glTexSubImage2D(SurfaceTargetToGL(params.target), mip_map, x0, y0,
case SurfaceParams::SurfaceTarget::Texture2D:
glTexSubImage2D(SurfaceTargetToGL(params.target), 0, x0, y0,
static_cast<GLsizei>(rect.GetWidth()),
static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
&gl_buffer[mip_map][buffer_offset]);
&gl_buffer[buffer_offset]);
break;
case 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 SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubeArray:
glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0,
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture2DArray:
glTexSubImage3D(SurfaceTargetToGL(params.target), 0, x0, y0, 0,
static_cast<GLsizei>(rect.GetWidth()),
static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format,
tuple.type, &gl_buffer[mip_map][buffer_offset]);
tuple.type, &gl_buffer[buffer_offset]);
break;
case SurfaceTarget::TextureCubemap: {
std::size_t start = buffer_offset;
case SurfaceParams::SurfaceTarget::TextureCubemap:
for (std::size_t face = 0; face < params.depth; ++face) {
glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), mip_map,
x0, y0, static_cast<GLsizei>(rect.GetWidth()),
glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), 0, x0,
y0, static_cast<GLsizei>(rect.GetWidth()),
static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
&gl_buffer[mip_map][buffer_offset]);
buffer_offset += params.LayerSizeGL(mip_map);
&gl_buffer[buffer_offset]);
buffer_offset += params.SizeInBytesCubeFace();
}
break;
}
default:
LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
static_cast<u32>(params.target));
UNREACHABLE();
glTexSubImage2D(GL_TEXTURE_2D, mip_map, x0, y0, static_cast<GLsizei>(rect.GetWidth()),
glTexSubImage2D(GL_TEXTURE_2D, 0, x0, y0, static_cast<GLsizei>(rect.GetWidth()),
static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
&gl_buffer[mip_map][buffer_offset]);
&gl_buffer[buffer_offset]);
}
}
glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
}
MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 192, 64));
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();
@@ -1312,7 +1294,8 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
// For compatible surfaces, we can just do fast glCopyImageSubData based copy
if (old_params.target == new_params.target && old_params.type == new_params.type &&
old_params.depth == new_params.depth && old_params.depth == 1 &&
GetFormatBpp(old_params.pixel_format) == GetFormatBpp(new_params.pixel_format)) {
SurfaceParams::GetFormatBpp(old_params.pixel_format) ==
SurfaceParams::GetFormatBpp(new_params.pixel_format)) {
FastCopySurface(old_surface, new_surface);
return new_surface;
}
@@ -1325,16 +1308,15 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
const bool is_blit{old_params.pixel_format == new_params.pixel_format};
switch (new_params.target) {
case SurfaceTarget::Texture2D:
case SurfaceParams::SurfaceTarget::Texture2D:
if (is_blit) {
BlitSurface(old_surface, new_surface, read_framebuffer.handle, draw_framebuffer.handle);
} else {
CopySurface(old_surface, new_surface, copy_pbo.handle);
}
break;
case SurfaceTarget::TextureCubemap:
case SurfaceTarget::Texture3D:
case SurfaceTarget::TextureCubeArray:
case SurfaceParams::SurfaceTarget::TextureCubemap:
case SurfaceParams::SurfaceTarget::Texture3D:
AccurateCopySurface(old_surface, new_surface);
break;
default:
@@ -1344,7 +1326,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
}
return new_surface;
}
} // namespace OpenGL
Surface RasterizerCacheOpenGL::TryFindFramebufferSurface(VAddr addr) const {
return TryGet(addr);

View File

@@ -7,7 +7,6 @@
#include <array>
#include <map>
#include <memory>
#include <string>
#include <vector>
#include "common/alignment.h"
@@ -19,7 +18,6 @@
#include "video_core/rasterizer_cache.h"
#include "video_core/renderer_opengl/gl_resource_manager.h"
#include "video_core/renderer_opengl/gl_shader_gen.h"
#include "video_core/surface.h"
#include "video_core/textures/decoders.h"
#include "video_core/textures/texture.h"
@@ -29,12 +27,135 @@ class CachedSurface;
using Surface = std::shared_ptr<CachedSurface>;
using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, MathUtil::Rectangle<u32>>;
using SurfaceTarget = VideoCore::Surface::SurfaceTarget;
using SurfaceType = VideoCore::Surface::SurfaceType;
using PixelFormat = VideoCore::Surface::PixelFormat;
using ComponentType = VideoCore::Surface::ComponentType;
struct SurfaceParams {
enum class PixelFormat {
ABGR8U = 0,
ABGR8S = 1,
ABGR8UI = 2,
B5G6R5U = 3,
A2B10G10R10U = 4,
A1B5G5R5U = 5,
R8U = 6,
R8UI = 7,
RGBA16F = 8,
RGBA16U = 9,
RGBA16UI = 10,
R11FG11FB10F = 11,
RGBA32UI = 12,
DXT1 = 13,
DXT23 = 14,
DXT45 = 15,
DXN1 = 16, // This is also known as BC4
DXN2UNORM = 17,
DXN2SNORM = 18,
BC7U = 19,
BC6H_UF16 = 20,
BC6H_SF16 = 21,
ASTC_2D_4X4 = 22,
G8R8U = 23,
G8R8S = 24,
BGRA8 = 25,
RGBA32F = 26,
RG32F = 27,
R32F = 28,
R16F = 29,
R16U = 30,
R16S = 31,
R16UI = 32,
R16I = 33,
RG16 = 34,
RG16F = 35,
RG16UI = 36,
RG16I = 37,
RG16S = 38,
RGB32F = 39,
RGBA8_SRGB = 40,
RG8U = 41,
RG8S = 42,
RG32UI = 43,
R32UI = 44,
ASTC_2D_8X8 = 45,
ASTC_2D_8X5 = 46,
ASTC_2D_5X4 = 47,
BGRA8_SRGB = 48,
DXT1_SRGB = 49,
DXT23_SRGB = 50,
DXT45_SRGB = 51,
BC7U_SRGB = 52,
ASTC_2D_4X4_SRGB = 53,
ASTC_2D_8X8_SRGB = 54,
ASTC_2D_8X5_SRGB = 55,
ASTC_2D_5X4_SRGB = 56,
MaxColorFormat,
// Depth formats
Z32F = 57,
Z16 = 58,
MaxDepthFormat,
// DepthStencil formats
Z24S8 = 59,
S8Z24 = 60,
Z32FS8 = 61,
MaxDepthStencilFormat,
Max = MaxDepthStencilFormat,
Invalid = 255,
};
static constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max);
enum class ComponentType {
Invalid = 0,
SNorm = 1,
UNorm = 2,
SInt = 3,
UInt = 4,
Float = 5,
};
enum class SurfaceType {
ColorTexture = 0,
Depth = 1,
DepthStencil = 2,
Fill = 3,
Invalid = 4,
};
enum class SurfaceTarget {
Texture1D,
Texture2D,
Texture3D,
Texture1DArray,
Texture2DArray,
TextureCubemap,
};
static SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type) {
switch (texture_type) {
case Tegra::Texture::TextureType::Texture1D:
return SurfaceTarget::Texture1D;
case Tegra::Texture::TextureType::Texture2D:
case Tegra::Texture::TextureType::Texture2DNoMipmap:
return SurfaceTarget::Texture2D;
case Tegra::Texture::TextureType::Texture3D:
return SurfaceTarget::Texture3D;
case Tegra::Texture::TextureType::TextureCubemap:
return SurfaceTarget::TextureCubemap;
case Tegra::Texture::TextureType::Texture1DArray:
return SurfaceTarget::Texture1DArray;
case Tegra::Texture::TextureType::Texture2DArray:
return SurfaceTarget::Texture2DArray;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", static_cast<u32>(texture_type));
UNREACHABLE();
return SurfaceTarget::Texture2D;
}
}
static std::string SurfaceTargetName(SurfaceTarget target) {
switch (target) {
case SurfaceTarget::Texture1D:
@@ -49,8 +170,6 @@ struct SurfaceParams {
return "Texture2DArray";
case SurfaceTarget::TextureCubemap:
return "TextureCubemap";
case SurfaceTarget::TextureCubeArray:
return "TextureCubeArray";
default:
LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
UNREACHABLE();
@@ -58,12 +177,664 @@ struct SurfaceParams {
}
}
static bool SurfaceTargetIsLayered(SurfaceTarget target) {
switch (target) {
case SurfaceTarget::Texture1D:
case SurfaceTarget::Texture2D:
case SurfaceTarget::Texture3D:
return false;
case SurfaceTarget::Texture1DArray:
case SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubemap:
return true;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
UNREACHABLE();
return false;
}
}
/**
* Gets the compression factor for the specified PixelFormat. This applies to just the
* "compressed width" and "compressed height", not the overall compression factor of a
* compressed image. This is used for maintaining proper surface sizes for compressed
* texture formats.
*/
static constexpr u32 GetCompressionFactor(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{
1, // ABGR8U
1, // ABGR8S
1, // ABGR8UI
1, // B5G6R5U
1, // A2B10G10R10U
1, // A1B5G5R5U
1, // R8U
1, // R8UI
1, // RGBA16F
1, // RGBA16U
1, // RGBA16UI
1, // R11FG11FB10F
1, // RGBA32UI
4, // DXT1
4, // DXT23
4, // DXT45
4, // DXN1
4, // DXN2UNORM
4, // DXN2SNORM
4, // BC7U
4, // BC6H_UF16
4, // BC6H_SF16
4, // ASTC_2D_4X4
1, // G8R8U
1, // G8R8S
1, // BGRA8
1, // RGBA32F
1, // RG32F
1, // R32F
1, // R16F
1, // R16U
1, // R16S
1, // R16UI
1, // R16I
1, // RG16
1, // RG16F
1, // RG16UI
1, // RG16I
1, // RG16S
1, // RGB32F
1, // RGBA8_SRGB
1, // RG8U
1, // RG8S
1, // RG32UI
1, // R32UI
4, // ASTC_2D_8X8
4, // ASTC_2D_8X5
4, // ASTC_2D_5X4
1, // BGRA8_SRGB
4, // DXT1_SRGB
4, // DXT23_SRGB
4, // DXT45_SRGB
4, // BC7U_SRGB
4, // ASTC_2D_4X4_SRGB
4, // ASTC_2D_8X8_SRGB
4, // ASTC_2D_8X5_SRGB
4, // ASTC_2D_5X4_SRGB
1, // Z32F
1, // Z16
1, // Z24S8
1, // S8Z24
1, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < compression_factor_table.size());
return compression_factor_table[static_cast<std::size_t>(format)];
}
static constexpr u32 GetDefaultBlockHeight(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> block_height_table = {{
1, // ABGR8U
1, // ABGR8S
1, // ABGR8UI
1, // B5G6R5U
1, // A2B10G10R10U
1, // A1B5G5R5U
1, // R8U
1, // R8UI
1, // RGBA16F
1, // RGBA16U
1, // RGBA16UI
1, // R11FG11FB10F
1, // RGBA32UI
4, // DXT1
4, // DXT23
4, // DXT45
4, // DXN1
4, // DXN2UNORM
4, // DXN2SNORM
4, // BC7U
4, // BC6H_UF16
4, // BC6H_SF16
4, // ASTC_2D_4X4
1, // G8R8U
1, // G8R8S
1, // BGRA8
1, // RGBA32F
1, // RG32F
1, // R32F
1, // R16F
1, // R16U
1, // R16S
1, // R16UI
1, // R16I
1, // RG16
1, // RG16F
1, // RG16UI
1, // RG16I
1, // RG16S
1, // RGB32F
1, // RGBA8_SRGB
1, // RG8U
1, // RG8S
1, // RG32UI
1, // R32UI
8, // ASTC_2D_8X8
5, // ASTC_2D_8X5
4, // ASTC_2D_5X4
1, // BGRA8_SRGB
4, // DXT1_SRGB
4, // DXT23_SRGB
4, // DXT45_SRGB
4, // BC7U_SRGB
4, // ASTC_2D_4X4_SRGB
8, // ASTC_2D_8X8_SRGB
5, // ASTC_2D_8X5_SRGB
4, // ASTC_2D_5X4_SRGB
1, // Z32F
1, // Z16
1, // Z24S8
1, // S8Z24
1, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < block_height_table.size());
return block_height_table[static_cast<std::size_t>(format)];
}
static constexpr u32 GetFormatBpp(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
32, // ABGR8U
32, // ABGR8S
32, // ABGR8UI
16, // B5G6R5U
32, // A2B10G10R10U
16, // A1B5G5R5U
8, // R8U
8, // R8UI
64, // RGBA16F
64, // RGBA16U
64, // RGBA16UI
32, // R11FG11FB10F
128, // RGBA32UI
64, // DXT1
128, // DXT23
128, // DXT45
64, // DXN1
128, // DXN2UNORM
128, // DXN2SNORM
128, // BC7U
128, // BC6H_UF16
128, // BC6H_SF16
32, // ASTC_2D_4X4
16, // G8R8U
16, // G8R8S
32, // BGRA8
128, // RGBA32F
64, // RG32F
32, // R32F
16, // R16F
16, // R16U
16, // R16S
16, // R16UI
16, // R16I
32, // RG16
32, // RG16F
32, // RG16UI
32, // RG16I
32, // RG16S
96, // RGB32F
32, // RGBA8_SRGB
16, // RG8U
16, // RG8S
64, // RG32UI
32, // R32UI
16, // ASTC_2D_8X8
16, // ASTC_2D_8X5
32, // ASTC_2D_5X4
32, // BGRA8_SRGB
64, // DXT1_SRGB
128, // DXT23_SRGB
128, // DXT45_SRGB
128, // BC7U
32, // ASTC_2D_4X4_SRGB
16, // ASTC_2D_8X8_SRGB
16, // ASTC_2D_8X5_SRGB
32, // ASTC_2D_5X4_SRGB
32, // Z32F
16, // Z16
32, // Z24S8
32, // S8Z24
64, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < bpp_table.size());
return bpp_table[static_cast<std::size_t>(format)];
}
u32 GetFormatBpp() const {
return VideoCore::Surface::GetFormatBpp(pixel_format);
return GetFormatBpp(pixel_format);
}
static PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
switch (format) {
case Tegra::DepthFormat::S8_Z24_UNORM:
return PixelFormat::S8Z24;
case Tegra::DepthFormat::Z24_S8_UNORM:
return PixelFormat::Z24S8;
case Tegra::DepthFormat::Z32_FLOAT:
return PixelFormat::Z32F;
case Tegra::DepthFormat::Z16_UNORM:
return PixelFormat::Z16;
case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
return PixelFormat::Z32FS8;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
switch (format) {
// TODO (Hexagon12): Converting SRGBA to RGBA is a hack and doesn't completely correct the
// gamma.
case Tegra::RenderTargetFormat::RGBA8_SRGB:
return PixelFormat::RGBA8_SRGB;
case Tegra::RenderTargetFormat::RGBA8_UNORM:
return PixelFormat::ABGR8U;
case Tegra::RenderTargetFormat::RGBA8_SNORM:
return PixelFormat::ABGR8S;
case Tegra::RenderTargetFormat::RGBA8_UINT:
return PixelFormat::ABGR8UI;
case Tegra::RenderTargetFormat::BGRA8_SRGB:
return PixelFormat::BGRA8_SRGB;
case Tegra::RenderTargetFormat::BGRA8_UNORM:
return PixelFormat::BGRA8;
case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
return PixelFormat::A2B10G10R10U;
case Tegra::RenderTargetFormat::RGBA16_FLOAT:
return PixelFormat::RGBA16F;
case Tegra::RenderTargetFormat::RGBA16_UNORM:
return PixelFormat::RGBA16U;
case Tegra::RenderTargetFormat::RGBA16_UINT:
return PixelFormat::RGBA16UI;
case Tegra::RenderTargetFormat::RGBA32_FLOAT:
return PixelFormat::RGBA32F;
case Tegra::RenderTargetFormat::RG32_FLOAT:
return PixelFormat::RG32F;
case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
return PixelFormat::R11FG11FB10F;
case Tegra::RenderTargetFormat::B5G6R5_UNORM:
return PixelFormat::B5G6R5U;
case Tegra::RenderTargetFormat::BGR5A1_UNORM:
return PixelFormat::A1B5G5R5U;
case Tegra::RenderTargetFormat::RGBA32_UINT:
return PixelFormat::RGBA32UI;
case Tegra::RenderTargetFormat::R8_UNORM:
return PixelFormat::R8U;
case Tegra::RenderTargetFormat::R8_UINT:
return PixelFormat::R8UI;
case Tegra::RenderTargetFormat::RG16_FLOAT:
return PixelFormat::RG16F;
case Tegra::RenderTargetFormat::RG16_UINT:
return PixelFormat::RG16UI;
case Tegra::RenderTargetFormat::RG16_SINT:
return PixelFormat::RG16I;
case Tegra::RenderTargetFormat::RG16_UNORM:
return PixelFormat::RG16;
case Tegra::RenderTargetFormat::RG16_SNORM:
return PixelFormat::RG16S;
case Tegra::RenderTargetFormat::RG8_UNORM:
return PixelFormat::RG8U;
case Tegra::RenderTargetFormat::RG8_SNORM:
return PixelFormat::RG8S;
case Tegra::RenderTargetFormat::R16_FLOAT:
return PixelFormat::R16F;
case Tegra::RenderTargetFormat::R16_UNORM:
return PixelFormat::R16U;
case Tegra::RenderTargetFormat::R16_SNORM:
return PixelFormat::R16S;
case Tegra::RenderTargetFormat::R16_UINT:
return PixelFormat::R16UI;
case Tegra::RenderTargetFormat::R16_SINT:
return PixelFormat::R16I;
case Tegra::RenderTargetFormat::R32_FLOAT:
return PixelFormat::R32F;
case Tegra::RenderTargetFormat::R32_UINT:
return PixelFormat::R32UI;
case Tegra::RenderTargetFormat::RG32_UINT:
return PixelFormat::RG32UI;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
Tegra::Texture::ComponentType component_type,
bool is_srgb) {
// TODO(Subv): Properly implement this
switch (format) {
case Tegra::Texture::TextureFormat::A8R8G8B8:
if (is_srgb) {
return PixelFormat::RGBA8_SRGB;
}
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::ABGR8U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::ABGR8S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::ABGR8UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::B5G6R5:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::B5G6R5U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::A2B10G10R10:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::A2B10G10R10U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::A1B5G5R5:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::A1B5G5R5U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R8:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::R8U;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R8UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::G8R8:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::G8R8U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::G8R8S;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R16_G16_B16_A16:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::RGBA16U;
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGBA16F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::BF10GF11RF11:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R11FG11FB10F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32_B32_A32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGBA32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RGBA32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RG32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RG32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32_B32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGB32F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R16:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R16F;
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::R16U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::R16S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R16UI;
case Tegra::Texture::ComponentType::SINT:
return PixelFormat::R16I;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::ZF32:
return PixelFormat::Z32F;
case Tegra::Texture::TextureFormat::Z16:
return PixelFormat::Z16;
case Tegra::Texture::TextureFormat::Z24S8:
return PixelFormat::Z24S8;
case Tegra::Texture::TextureFormat::DXT1:
return is_srgb ? PixelFormat::DXT1_SRGB : PixelFormat::DXT1;
case Tegra::Texture::TextureFormat::DXT23:
return is_srgb ? PixelFormat::DXT23_SRGB : PixelFormat::DXT23;
case Tegra::Texture::TextureFormat::DXT45:
return is_srgb ? PixelFormat::DXT45_SRGB : PixelFormat::DXT45;
case Tegra::Texture::TextureFormat::DXN1:
return PixelFormat::DXN1;
case Tegra::Texture::TextureFormat::DXN2:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::DXN2UNORM;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::DXN2SNORM;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::BC7U:
return is_srgb ? PixelFormat::BC7U_SRGB : PixelFormat::BC7U;
case Tegra::Texture::TextureFormat::BC6H_UF16:
return PixelFormat::BC6H_UF16;
case Tegra::Texture::TextureFormat::BC6H_SF16:
return PixelFormat::BC6H_SF16;
case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
return is_srgb ? PixelFormat::ASTC_2D_4X4_SRGB : PixelFormat::ASTC_2D_4X4;
case Tegra::Texture::TextureFormat::ASTC_2D_5X4:
return is_srgb ? PixelFormat::ASTC_2D_5X4_SRGB : PixelFormat::ASTC_2D_5X4;
case Tegra::Texture::TextureFormat::ASTC_2D_8X8:
return is_srgb ? PixelFormat::ASTC_2D_8X8_SRGB : PixelFormat::ASTC_2D_8X8;
case Tegra::Texture::TextureFormat::ASTC_2D_8X5:
return is_srgb ? PixelFormat::ASTC_2D_8X5_SRGB : PixelFormat::ASTC_2D_8X5;
case Tegra::Texture::TextureFormat::R16_G16:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RG16F;
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::RG16;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::RG16S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RG16UI;
case Tegra::Texture::ComponentType::SINT:
return PixelFormat::RG16I;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}, component_type={}",
static_cast<u32>(format), static_cast<u32>(component_type));
UNREACHABLE();
}
}
static ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
// TODO(Subv): Implement more component types
switch (type) {
case Tegra::Texture::ComponentType::UNORM:
return ComponentType::UNorm;
case Tegra::Texture::ComponentType::FLOAT:
return ComponentType::Float;
case Tegra::Texture::ComponentType::SNORM:
return ComponentType::SNorm;
case Tegra::Texture::ComponentType::UINT:
return ComponentType::UInt;
case Tegra::Texture::ComponentType::SINT:
return ComponentType::SInt;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented component type={}", static_cast<u32>(type));
UNREACHABLE();
}
}
static ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format) {
// TODO(Subv): Implement more render targets
switch (format) {
case Tegra::RenderTargetFormat::RGBA8_UNORM:
case Tegra::RenderTargetFormat::RGBA8_SRGB:
case Tegra::RenderTargetFormat::BGRA8_UNORM:
case Tegra::RenderTargetFormat::BGRA8_SRGB:
case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
case Tegra::RenderTargetFormat::R8_UNORM:
case Tegra::RenderTargetFormat::RG16_UNORM:
case Tegra::RenderTargetFormat::R16_UNORM:
case Tegra::RenderTargetFormat::B5G6R5_UNORM:
case Tegra::RenderTargetFormat::BGR5A1_UNORM:
case Tegra::RenderTargetFormat::RG8_UNORM:
case Tegra::RenderTargetFormat::RGBA16_UNORM:
return ComponentType::UNorm;
case Tegra::RenderTargetFormat::RGBA8_SNORM:
case Tegra::RenderTargetFormat::RG16_SNORM:
case Tegra::RenderTargetFormat::R16_SNORM:
case Tegra::RenderTargetFormat::RG8_SNORM:
return ComponentType::SNorm;
case Tegra::RenderTargetFormat::RGBA16_FLOAT:
case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
case Tegra::RenderTargetFormat::RGBA32_FLOAT:
case Tegra::RenderTargetFormat::RG32_FLOAT:
case Tegra::RenderTargetFormat::RG16_FLOAT:
case Tegra::RenderTargetFormat::R16_FLOAT:
case Tegra::RenderTargetFormat::R32_FLOAT:
return ComponentType::Float;
case Tegra::RenderTargetFormat::RGBA32_UINT:
case Tegra::RenderTargetFormat::RGBA16_UINT:
case Tegra::RenderTargetFormat::RG16_UINT:
case Tegra::RenderTargetFormat::R8_UINT:
case Tegra::RenderTargetFormat::R16_UINT:
case Tegra::RenderTargetFormat::RG32_UINT:
case Tegra::RenderTargetFormat::R32_UINT:
case Tegra::RenderTargetFormat::RGBA8_UINT:
return ComponentType::UInt;
case Tegra::RenderTargetFormat::RG16_SINT:
case Tegra::RenderTargetFormat::R16_SINT:
return ComponentType::SInt;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
switch (format) {
case Tegra::FramebufferConfig::PixelFormat::ABGR8:
return PixelFormat::ABGR8U;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format) {
switch (format) {
case Tegra::DepthFormat::Z16_UNORM:
case Tegra::DepthFormat::S8_Z24_UNORM:
case Tegra::DepthFormat::Z24_S8_UNORM:
return ComponentType::UNorm;
case Tegra::DepthFormat::Z32_FLOAT:
case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
return ComponentType::Float;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static SurfaceType GetFormatType(PixelFormat pixel_format) {
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxColorFormat)) {
return SurfaceType::ColorTexture;
}
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxDepthFormat)) {
return SurfaceType::Depth;
}
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxDepthStencilFormat)) {
return SurfaceType::DepthStencil;
}
// TODO(Subv): Implement the other formats
ASSERT(false);
return SurfaceType::Invalid;
}
/// Returns the sizer in bytes of the specified pixel format
static constexpr u32 GetBytesPerPixel(PixelFormat pixel_format) {
if (pixel_format == SurfaceParams::PixelFormat::Invalid) {
return 0;
}
return GetFormatBpp(pixel_format) / CHAR_BIT;
}
/// Returns the rectangle corresponding to this surface
MathUtil::Rectangle<u32> GetRect(u32 mip_level = 0) const;
MathUtil::Rectangle<u32> GetRect() const;
/// Returns the total size of this surface in bytes, adjusted for compression
std::size_t SizeInBytesRaw(bool ignore_tiled = false) const {
@@ -94,7 +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(false, is_layered);
std::size_t size = InnerMemorySize(is_layered);
if (is_layered)
return size * depth;
return size;
@@ -103,78 +874,12 @@ 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(false, true);
return InnerMemorySize(true);
}
/// Returns the size of a layer of this surface in OpenGL.
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 is_layered ? depth : 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;
std::size_t LayerSizeGL() const {
return SizeInBytesRaw(true) / depth;
}
/// Creates SurfaceParams from a texture configuration
@@ -235,10 +940,7 @@ struct SurfaceParams {
} rt;
private:
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;
std::size_t InnerMemorySize(bool layer_only = false) const;
};
}; // namespace OpenGL
@@ -300,10 +1002,8 @@ 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<std::vector<u8>> gl_buffer;
std::vector<u8> gl_buffer;
SurfaceParams params;
GLenum gl_target;
std::size_t cached_size_in_bytes;

View File

@@ -1,188 +0,0 @@
// Copyright 2015 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <utility>
#include <glad/glad.h>
#include "common/common_types.h"
#include "common/microprofile.h"
#include "video_core/renderer_opengl/gl_resource_manager.h"
#include "video_core/renderer_opengl/gl_shader_util.h"
#include "video_core/renderer_opengl/gl_state.h"
MICROPROFILE_DEFINE(OpenGL_ResourceCreation, "OpenGL", "Resource Creation",
MP_RGB(128, 128, 192));
MICROPROFILE_DEFINE(OpenGL_ResourceDeletion, "OpenGL", "Resource Deletion",
MP_RGB(128, 128, 192));
namespace OpenGL {
void OGLTexture::Create() {
if (handle != 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
glGenTextures(1, &handle);
}
void OGLTexture::Release() {
if (handle == 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
glDeleteTextures(1, &handle);
OpenGLState::GetCurState().UnbindTexture(handle).Apply();
handle = 0;
}
void OGLSampler::Create() {
if (handle != 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
glGenSamplers(1, &handle);
}
void OGLSampler::Release() {
if (handle == 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
glDeleteSamplers(1, &handle);
OpenGLState::GetCurState().ResetSampler(handle).Apply();
handle = 0;
}
void OGLShader::Create(const char* source, GLenum type) {
if (handle != 0)
return;
if (source == nullptr)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
handle = GLShader::LoadShader(source, type);
}
void OGLShader::Release() {
if (handle == 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
glDeleteShader(handle);
handle = 0;
}
void OGLProgram::CreateFromSource(const char* vert_shader, const char* geo_shader,
const char* frag_shader, bool separable_program) {
OGLShader vert, geo, frag;
if (vert_shader)
vert.Create(vert_shader, GL_VERTEX_SHADER);
if (geo_shader)
geo.Create(geo_shader, GL_GEOMETRY_SHADER);
if (frag_shader)
frag.Create(frag_shader, GL_FRAGMENT_SHADER);
MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
Create(separable_program, vert.handle, geo.handle, frag.handle);
}
void OGLProgram::Release() {
if (handle == 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
glDeleteProgram(handle);
OpenGLState::GetCurState().ResetProgram(handle).Apply();
handle = 0;
}
void OGLPipeline::Create() {
if (handle != 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
glGenProgramPipelines(1, &handle);
}
void OGLPipeline::Release() {
if (handle == 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
glDeleteProgramPipelines(1, &handle);
OpenGLState::GetCurState().ResetPipeline(handle).Apply();
handle = 0;
}
void OGLBuffer::Create() {
if (handle != 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
glGenBuffers(1, &handle);
}
void OGLBuffer::Release() {
if (handle == 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
glDeleteBuffers(1, &handle);
OpenGLState::GetCurState().ResetBuffer(handle).Apply();
handle = 0;
}
void OGLSync::Create() {
if (handle != 0)
return;
// Don't profile here, this one is expected to happen ingame.
handle = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0);
}
void OGLSync::Release() {
if (handle == 0)
return;
// Don't profile here, this one is expected to happen ingame.
glDeleteSync(handle);
handle = 0;
}
void OGLVertexArray::Create() {
if (handle != 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
glGenVertexArrays(1, &handle);
}
void OGLVertexArray::Release() {
if (handle == 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
glDeleteVertexArrays(1, &handle);
OpenGLState::GetCurState().ResetVertexArray(handle).Apply();
handle = 0;
}
void OGLFramebuffer::Create() {
if (handle != 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceCreation);
glGenFramebuffers(1, &handle);
}
void OGLFramebuffer::Release() {
if (handle == 0)
return;
MICROPROFILE_SCOPE(OpenGL_ResourceDeletion);
glDeleteFramebuffers(1, &handle);
OpenGLState::GetCurState().ResetFramebuffer(handle).Apply();
handle = 0;
}
} // namespace OpenGL

View File

@@ -8,6 +8,7 @@
#include <glad/glad.h>
#include "common/common_types.h"
#include "video_core/renderer_opengl/gl_shader_util.h"
#include "video_core/renderer_opengl/gl_state.h"
namespace OpenGL {
@@ -28,10 +29,20 @@ public:
}
/// Creates a new internal OpenGL resource and stores the handle
void Create();
void Create() {
if (handle != 0)
return;
glGenTextures(1, &handle);
}
/// Deletes the internal OpenGL resource
void Release();
void Release() {
if (handle == 0)
return;
glDeleteTextures(1, &handle);
OpenGLState::GetCurState().UnbindTexture(handle).Apply();
handle = 0;
}
GLuint handle = 0;
};
@@ -53,10 +64,20 @@ public:
}
/// Creates a new internal OpenGL resource and stores the handle
void Create();
void Create() {
if (handle != 0)
return;
glGenSamplers(1, &handle);
}
/// Deletes the internal OpenGL resource
void Release();
void Release() {
if (handle == 0)
return;
glDeleteSamplers(1, &handle);
OpenGLState::GetCurState().ResetSampler(handle).Apply();
handle = 0;
}
GLuint handle = 0;
};
@@ -77,9 +98,20 @@ public:
return *this;
}
void Create(const char* source, GLenum type);
void Create(const char* source, GLenum type) {
if (handle != 0)
return;
if (source == nullptr)
return;
handle = GLShader::LoadShader(source, type);
}
void Release();
void Release() {
if (handle == 0)
return;
glDeleteShader(handle);
handle = 0;
}
GLuint handle = 0;
};
@@ -109,10 +141,25 @@ public:
/// Creates a new internal OpenGL resource and stores the handle
void CreateFromSource(const char* vert_shader, const char* geo_shader, const char* frag_shader,
bool separable_program = false);
bool separable_program = false) {
OGLShader vert, geo, frag;
if (vert_shader)
vert.Create(vert_shader, GL_VERTEX_SHADER);
if (geo_shader)
geo.Create(geo_shader, GL_GEOMETRY_SHADER);
if (frag_shader)
frag.Create(frag_shader, GL_FRAGMENT_SHADER);
Create(separable_program, vert.handle, geo.handle, frag.handle);
}
/// Deletes the internal OpenGL resource
void Release();
void Release() {
if (handle == 0)
return;
glDeleteProgram(handle);
OpenGLState::GetCurState().ResetProgram(handle).Apply();
handle = 0;
}
GLuint handle = 0;
};
@@ -131,10 +178,20 @@ public:
}
/// Creates a new internal OpenGL resource and stores the handle
void Create();
void Create() {
if (handle != 0)
return;
glGenProgramPipelines(1, &handle);
}
/// Deletes the internal OpenGL resource
void Release();
void Release() {
if (handle == 0)
return;
glDeleteProgramPipelines(1, &handle);
OpenGLState::GetCurState().ResetPipeline(handle).Apply();
handle = 0;
}
GLuint handle = 0;
};
@@ -156,10 +213,20 @@ public:
}
/// Creates a new internal OpenGL resource and stores the handle
void Create();
void Create() {
if (handle != 0)
return;
glGenBuffers(1, &handle);
}
/// Deletes the internal OpenGL resource
void Release();
void Release() {
if (handle == 0)
return;
glDeleteBuffers(1, &handle);
OpenGLState::GetCurState().ResetBuffer(handle).Apply();
handle = 0;
}
GLuint handle = 0;
};
@@ -180,10 +247,19 @@ public:
}
/// Creates a new internal OpenGL resource and stores the handle
void Create();
void Create() {
if (handle != 0)
return;
handle = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0);
}
/// Deletes the internal OpenGL resource
void Release();
void Release() {
if (handle == 0)
return;
glDeleteSync(handle);
handle = 0;
}
GLsync handle = 0;
};
@@ -205,10 +281,20 @@ public:
}
/// Creates a new internal OpenGL resource and stores the handle
void Create();
void Create() {
if (handle != 0)
return;
glGenVertexArrays(1, &handle);
}
/// Deletes the internal OpenGL resource
void Release();
void Release() {
if (handle == 0)
return;
glDeleteVertexArrays(1, &handle);
OpenGLState::GetCurState().ResetVertexArray(handle).Apply();
handle = 0;
}
GLuint handle = 0;
};
@@ -230,10 +316,20 @@ public:
}
/// Creates a new internal OpenGL resource and stores the handle
void Create();
void Create() {
if (handle != 0)
return;
glGenFramebuffers(1, &handle);
}
/// Deletes the internal OpenGL resource
void Release();
void Release() {
if (handle == 0)
return;
glDeleteFramebuffers(1, &handle);
OpenGLState::GetCurState().ResetFramebuffer(handle).Apply();
handle = 0;
}
GLuint handle = 0;
};

View File

@@ -8,7 +8,6 @@
#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 {
@@ -90,7 +89,7 @@ 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);
LabelGLObject(GL_PROGRAM, program.handle, addr);
VideoCore::LabelGLObject(GL_PROGRAM, program.handle, addr);
} else {
// Store shader's code to lazily build it on draw
geometry_programs.code = program_result.first;
@@ -131,7 +130,7 @@ GLuint CachedShader::LazyGeometryProgram(OGLProgram& target_program,
shader.Create(source.c_str(), GL_GEOMETRY_SHADER);
target_program.Create(true, shader.handle);
SetShaderUniformBlockBindings(target_program.handle);
LabelGLObject(GL_PROGRAM, target_program.handle, addr, debug_name);
VideoCore::LabelGLObject(GL_PROGRAM, target_program.handle, addr, debug_name);
return target_program.handle;
};

View File

@@ -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->get().GetId()) {
switch (opcode->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
@@ -373,7 +373,6 @@ public:
if (sets_cc) {
const std::string zero_condition = "( " + ConvertIntegerSize(value, size) + " == 0 )";
SetInternalFlag(InternalFlag::ZeroFlag, zero_condition);
LOG_WARNING(HW_GPU, "Control Codes Imcomplete.");
}
}
@@ -431,7 +430,7 @@ public:
*/
void SetRegisterToInputAttibute(const Register& reg, u64 elem, Attribute::Index attribute,
const Tegra::Shader::IpaMode& input_mode,
std::optional<Register> vertex = {}) {
boost::optional<Register> vertex = {}) {
const std::string dest = GetRegisterAsFloat(reg);
const std::string src = GetInputAttribute(attribute, input_mode, vertex) + GetSwizzle(elem);
shader.AddLine(dest + " = " + src + ';');
@@ -808,10 +807,10 @@ private:
/// Generates code representing an input attribute register.
std::string GetInputAttribute(Attribute::Index attribute,
const Tegra::Shader::IpaMode& input_mode,
std::optional<Register> vertex = {}) {
boost::optional<Register> vertex = {}) {
auto GeometryPass = [&](const std::string& name) {
if (stage == Maxwell3D::Regs::ShaderStage::Geometry && vertex) {
return "gs_" + name + '[' + GetRegisterAsInteger(*vertex, 0, false) + ']';
return "gs_" + name + '[' + GetRegisterAsInteger(vertex.value(), 0, false) + ']';
}
return name;
};
@@ -1466,7 +1465,7 @@ private:
}
shader.AddLine(
fmt::format("// {}: {} (0x{:016x})", offset, opcode->get().GetName(), instr.value));
fmt::format("// {}: {} (0x{:016x})", offset, opcode->GetName(), instr.value));
using Tegra::Shader::Pred;
ASSERT_MSG(instr.pred.full_pred != Pred::NeverExecute,
@@ -1474,7 +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->get().GetId());
bool can_be_predicated = OpCode::IsPredicatedInstruction(opcode->GetId());
if (can_be_predicated && instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex)) {
shader.AddLine("if (" +
@@ -1484,7 +1483,7 @@ private:
++shader.scope;
}
switch (opcode->get().GetType()) {
switch (opcode->GetType()) {
case OpCode::Type::Arithmetic: {
std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
@@ -1501,7 +1500,7 @@ private:
}
}
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::MOV_C:
case OpCode::Id::MOV_R: {
// MOV does not have neither 'abs' nor 'neg' bits.
@@ -1526,10 +1525,6 @@ private:
regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b, 1, 1,
instr.alu.saturate_d, 0, true);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "FMUL Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::FADD_C:
@@ -1540,10 +1535,6 @@ private:
regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1,
instr.alu.saturate_d, 0, true);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "FADD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::MUFU: {
@@ -1597,10 +1588,6 @@ private:
'(' + condition + ") ? min(" + parameters + ") : max(" +
parameters + ')',
1, 1, false, 0, true);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "FMNMX Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::RRO_C:
@@ -1613,15 +1600,14 @@ private:
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled arithmetic instruction: {}",
opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled arithmetic instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
break;
}
case OpCode::Type::ArithmeticImmediate: {
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::MOV32_IMM: {
regs.SetRegisterToFloat(instr.gpr0, 0, GetImmediate32(instr), 1, 1);
break;
@@ -1631,10 +1617,6 @@ private:
regs.GetRegisterAsFloat(instr.gpr8) + " * " +
GetImmediate32(instr),
1, 1, instr.fmul32.saturate, 0, true);
if (instr.op_32.generates_cc) {
LOG_CRITICAL(HW_GPU, "FMUL32 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::FADD32I: {
@@ -1658,10 +1640,6 @@ private:
}
regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1, false, 0, true);
if (instr.op_32.generates_cc) {
LOG_CRITICAL(HW_GPU, "FADD32 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
}
@@ -1673,7 +1651,7 @@ private:
std::string op_a = instr.bfe.negate_a ? "-" : "";
op_a += regs.GetRegisterAsInteger(instr.gpr8);
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::BFE_IMM: {
std::string inner_shift =
'(' + op_a + " << " + std::to_string(instr.bfe.GetLeftShiftValue()) + ')';
@@ -1682,14 +1660,10 @@ private:
std::to_string(instr.bfe.GetLeftShiftValue() + instr.bfe.shift_position) + ')';
regs.SetRegisterToInteger(instr.gpr0, true, 0, outer_shift, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "BFE Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled BFE instruction: {}", opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled BFE instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
@@ -1711,7 +1685,7 @@ private:
}
}
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::SHR_C:
case OpCode::Id::SHR_R:
case OpCode::Id::SHR_IMM: {
@@ -1723,23 +1697,15 @@ private:
// Cast to int is superfluous for arithmetic shift, it's only for a logical shift
regs.SetRegisterToInteger(instr.gpr0, true, 0, "int(" + op_a + " >> " + op_b + ')',
1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "SHR Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::SHL_C:
case OpCode::Id::SHL_R:
case OpCode::Id::SHL_IMM:
regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " << " + op_b, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "SHL Generates an unhandled Control Code");
UNREACHABLE();
}
break;
default: {
LOG_CRITICAL(HW_GPU, "Unhandled shift instruction: {}", opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled shift instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
@@ -1749,17 +1715,13 @@ private:
std::string op_a = regs.GetRegisterAsInteger(instr.gpr8);
std::string op_b = std::to_string(instr.alu.imm20_32.Value());
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::IADD32I:
if (instr.iadd32i.negate_a)
op_a = "-(" + op_a + ')';
regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1,
instr.iadd32i.saturate != 0);
if (instr.op_32.generates_cc) {
LOG_CRITICAL(HW_GPU, "IADD32 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
case OpCode::Id::LOP32I: {
if (instr.alu.lop32i.invert_a)
@@ -1771,15 +1733,11 @@ private:
WriteLogicOperation(instr.gpr0, instr.alu.lop32i.operation, op_a, op_b,
Tegra::Shader::PredicateResultMode::None,
Tegra::Shader::Pred::UnusedIndex);
if (instr.op_32.generates_cc) {
LOG_CRITICAL(HW_GPU, "LOP32I Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled ArithmeticIntegerImmediate instruction: {}",
opcode->get().GetName());
opcode->GetName());
UNREACHABLE();
}
}
@@ -1799,7 +1757,7 @@ private:
}
}
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::IADD_C:
case OpCode::Id::IADD_R:
case OpCode::Id::IADD_IMM: {
@@ -1811,10 +1769,6 @@ private:
regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1,
instr.alu.saturate_d);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "IADD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::IADD3_C:
@@ -1839,7 +1793,7 @@ private:
}
};
if (opcode->get().GetId() == OpCode::Id::IADD3_R) {
if (opcode->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);
@@ -1855,7 +1809,7 @@ private:
op_c = "-(" + op_c + ')';
std::string result;
if (opcode->get().GetId() == OpCode::Id::IADD3_R) {
if (opcode->GetId() == OpCode::Id::IADD3_R) {
switch (instr.iadd3.mode) {
case Tegra::Shader::IAdd3Mode::RightShift:
// TODO(tech4me): According to
@@ -1876,11 +1830,6 @@ private:
}
regs.SetRegisterToInteger(instr.gpr0, true, 0, result, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "IADD3 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::ISCADD_C:
@@ -1896,10 +1845,6 @@ private:
regs.SetRegisterToInteger(instr.gpr0, true, 0,
"((" + op_a + " << " + shift + ") + " + op_b + ')', 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "ISCADD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::POPC_C:
@@ -1931,10 +1876,6 @@ private:
WriteLogicOperation(instr.gpr0, instr.alu.lop.operation, op_a, op_b,
instr.alu.lop.pred_result_mode, instr.alu.lop.pred48);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "LOP Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::LOP3_C:
@@ -1943,17 +1884,13 @@ private:
const std::string op_c = regs.GetRegisterAsInteger(instr.gpr39);
std::string lut;
if (opcode->get().GetId() == OpCode::Id::LOP3_R) {
if (opcode->GetId() == OpCode::Id::LOP3_R) {
lut = '(' + std::to_string(instr.alu.lop3.GetImmLut28()) + ')';
} else {
lut = '(' + std::to_string(instr.alu.lop3.GetImmLut48()) + ')';
}
WriteLop3Instruction(instr.gpr0, op_a, op_b, op_c, lut);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "LOP3 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::IMNMX_C:
@@ -1968,10 +1905,6 @@ private:
'(' + condition + ") ? min(" + parameters + ") : max(" +
parameters + ')',
1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "IMNMX Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::LEA_R2:
@@ -1981,7 +1914,7 @@ private:
case OpCode::Id::LEA_HI: {
std::string op_c;
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::LEA_R2: {
op_a = regs.GetRegisterAsInteger(instr.gpr20);
op_b = regs.GetRegisterAsInteger(instr.gpr39);
@@ -2026,8 +1959,7 @@ 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->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled LEA subinstruction: {}", opcode->GetName());
UNREACHABLE();
}
}
@@ -2042,7 +1974,7 @@ private:
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled ArithmeticInteger instruction: {}",
opcode->get().GetName());
opcode->GetName());
UNREACHABLE();
}
}
@@ -2050,21 +1982,20 @@ private:
break;
}
case OpCode::Type::ArithmeticHalf: {
if (opcode->get().GetId() == OpCode::Id::HADD2_C ||
opcode->get().GetId() == OpCode::Id::HADD2_R) {
if (opcode->GetId() == OpCode::Id::HADD2_C || opcode->GetId() == OpCode::Id::HADD2_R) {
ASSERT_MSG(instr.alu_half.ftz == 0, "Unimplemented");
}
const bool negate_a =
opcode->get().GetId() != OpCode::Id::HMUL2_R && instr.alu_half.negate_a != 0;
opcode->GetId() != OpCode::Id::HMUL2_R && instr.alu_half.negate_a != 0;
const bool negate_b =
opcode->get().GetId() != OpCode::Id::HMUL2_C && instr.alu_half.negate_b != 0;
opcode->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->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::HADD2_C:
case OpCode::Id::HMUL2_C:
op_b = regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
@@ -2082,7 +2013,7 @@ 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->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::HADD2_C:
case OpCode::Id::HADD2_R:
return '(' + op_a + " + " + op_b + ')';
@@ -2090,8 +2021,7 @@ private:
case OpCode::Id::HMUL2_R:
return '(' + op_a + " * " + op_b + ')';
default:
LOG_CRITICAL(HW_GPU, "Unhandled half float instruction: {}",
opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled half float instruction: {}", opcode->GetName());
UNREACHABLE();
return std::string("0");
}
@@ -2102,7 +2032,7 @@ private:
break;
}
case OpCode::Type::ArithmeticHalfImmediate: {
if (opcode->get().GetId() == OpCode::Id::HADD2_IMM) {
if (opcode->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,
@@ -2116,7 +2046,7 @@ private:
const std::string op_b = UnpackHalfImmediate(instr, true);
const std::string result = [&]() {
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::HADD2_IMM:
return op_a + " + " + op_b;
case OpCode::Id::HMUL2_IMM:
@@ -2142,7 +2072,7 @@ private:
ASSERT_MSG(instr.ffma.tab5980_1 == 0, "FFMA tab5980_1({}) not implemented",
instr.ffma.tab5980_1.Value());
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::FFMA_CR: {
op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
GLSLRegister::Type::Float);
@@ -2166,29 +2096,25 @@ private:
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled FFMA instruction: {}", opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled FFMA instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
regs.SetRegisterToFloat(instr.gpr0, 0, "fma(" + op_a + ", " + op_b + ", " + op_c + ')',
1, 1, instr.alu.saturate_d, 0, true);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "FFMA Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Type::Hfma2: {
if (opcode->get().GetId() == OpCode::Id::HFMA2_RR) {
if (opcode->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->get().GetId() == OpCode::Id::HFMA2_RR
const bool saturate = opcode->GetId() == OpCode::Id::HFMA2_RR
? instr.hfma2.rr.saturate != 0
: instr.hfma2.saturate != 0;
@@ -2196,7 +2122,7 @@ private:
GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr8, 0, false), instr.hfma2.type_a);
std::string op_b, op_c;
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::HFMA2_CR:
op_b = GetHalfFloat(regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
GLSLRegister::Type::UnsignedInteger),
@@ -2234,7 +2160,7 @@ private:
break;
}
case OpCode::Type::Conversion: {
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::I2I_R: {
ASSERT_MSG(!instr.conversion.selector, "Unimplemented");
@@ -2282,11 +2208,6 @@ private:
}
regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "I2F Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::F2F_R: {
@@ -2325,11 +2246,6 @@ private:
}
regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1, instr.alu.saturate_d);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "F2F Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::F2I_R:
@@ -2379,22 +2295,17 @@ private:
regs.SetRegisterToInteger(instr.gpr0, instr.conversion.is_output_signed, 0, op_a, 1,
1, false, 0, instr.conversion.dest_size);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "F2I Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled conversion instruction: {}",
opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled conversion instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
break;
}
case OpCode::Type::Memory: {
switch (opcode->get().GetId()) {
switch (opcode->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,
@@ -3038,7 +2949,7 @@ private:
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
@@ -3132,7 +3043,7 @@ private:
instr.hsetp2.abs_a, instr.hsetp2.negate_a);
const std::string op_b = [&]() {
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::HSETP2_R:
return GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr20, 0, false),
instr.hsetp2.type_b, instr.hsetp2.abs_a,
@@ -3191,15 +3102,10 @@ private:
regs.SetRegisterToFloat(instr.gpr0, 0, value, 1, 1);
}
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "PSET Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Type::PredicateSetPredicate: {
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::PSETP: {
const std::string op_a =
GetPredicateCondition(instr.psetp.pred12, instr.psetp.neg_pred12 != 0);
@@ -3245,8 +3151,7 @@ private:
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled predicate instruction: {}",
opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled predicate instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
@@ -3334,7 +3239,7 @@ private:
instr.hset2.abs_a != 0, instr.hset2.negate_a != 0);
const std::string op_b = [&]() {
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::HSET2_R:
return GetHalfFloat(regs.GetRegisterAsInteger(instr.gpr20, 0, false),
instr.hset2.type_b, instr.hset2.abs_b != 0,
@@ -3383,7 +3288,7 @@ private:
const bool is_signed{instr.xmad.sign_a == 1};
bool is_merge{};
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::XMAD_CR: {
is_merge = instr.xmad.merge_56;
op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
@@ -3412,7 +3317,7 @@ private:
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled XMAD instruction: {}", opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled XMAD instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
@@ -3461,25 +3366,15 @@ private:
}
regs.SetRegisterToInteger(instr.gpr0, is_signed, 0, sum, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "XMAD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
default: {
switch (opcode->get().GetId()) {
switch (opcode->GetId()) {
case OpCode::Id::EXIT: {
if (stage == Maxwell3D::Regs::ShaderStage::Fragment) {
EmitFragmentOutputsWrite();
}
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "EXIT Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
switch (instr.flow.cond) {
case Tegra::Shader::FlowCondition::Always:
shader.AddLine("return true;");
@@ -3509,11 +3404,6 @@ private:
// Enclose "discard" in a conditional, so that GLSL compilation does not complain
// about unexecuted instructions that may follow this.
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "KIL Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
shader.AddLine("if (true) {");
++shader.scope;
shader.AddLine("discard;");
@@ -3571,11 +3461,6 @@ private:
case OpCode::Id::BRA: {
ASSERT_MSG(instr.bra.constant_buffer == 0,
"BRA with constant buffers are not implemented");
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "BRA Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
const u32 target = offset + instr.bra.GetBranchTarget();
shader.AddLine("{ jmp_to = " + std::to_string(target) + "u; break; }");
break;
@@ -3616,21 +3501,13 @@ private:
}
case OpCode::Id::SYNC: {
// The SYNC opcode jumps to the address previously set by the SSY opcode
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "SYNC Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
EmitPopFromFlowStack();
break;
}
case OpCode::Id::BRK: {
// The BRK opcode jumps to the address previously set by the PBK opcode
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "BRK Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
EmitPopFromFlowStack();
break;
}
@@ -3660,11 +3537,6 @@ private:
regs.SetRegisterToInteger(instr.gpr0, result_signed, 1, result, 1, 1,
instr.vmad.saturate == 1, 0, Register::Size::Word,
instr.vmad.cc);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "VMAD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::VSETP: {
@@ -3692,7 +3564,7 @@ private:
break;
}
default: {
LOG_CRITICAL(HW_GPU, "Unhandled instruction: {}", opcode->get().GetName());
LOG_CRITICAL(HW_GPU, "Unhandled instruction: {}", opcode->GetName());
UNREACHABLE();
}
}
@@ -3833,9 +3705,9 @@ std::string GetCommonDeclarations() {
RasterizerOpenGL::MaxConstbufferSize / sizeof(GLvec4));
}
std::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,
Maxwell3D::Regs::ShaderStage stage,
const std::string& suffix) {
boost::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();
@@ -3844,7 +3716,7 @@ std::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u
} catch (const DecompileFail& exception) {
LOG_ERROR(HW_GPU, "Shader decompilation failed: {}", exception.what());
}
return {};
return boost::none;
}
} // namespace OpenGL::GLShader::Decompiler

View File

@@ -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();
std::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,
Maxwell3D::Regs::ShaderStage stage,
const std::string& suffix);
boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,
Maxwell3D::Regs::ShaderStage stage,
const std::string& suffix);
} // namespace OpenGL::GLShader::Decompiler

View File

@@ -37,7 +37,7 @@ layout(std140) uniform vs_config {
ProgramResult program =
Decompiler::DecompileProgram(setup.program.code, PROGRAM_OFFSET,
Maxwell3D::Regs::ShaderStage::Vertex, "vertex")
.value_or(ProgramResult());
.get_value_or({});
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")
.value_or(ProgramResult());
.get_value_or({});
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")
.value_or(ProgramResult());
.get_value_or({});
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")
.value_or(ProgramResult());
.get_value_or({});
out += R"(
layout(location = 0) out vec4 FragColor0;
layout(location = 1) out vec4 FragColor1;

View File

@@ -7,7 +7,6 @@
#include <glad/glad.h>
#include "video_core/renderer_opengl/gl_resource_manager.h"
#include "video_core/renderer_opengl/gl_state.h"
#include "video_core/renderer_opengl/maxwell_to_gl.h"
namespace OpenGL::GLShader {

View File

@@ -22,15 +22,17 @@ OpenGLState::OpenGLState() {
depth.test_enabled = false;
depth.test_func = GL_LESS;
depth.write_mask = GL_TRUE;
depth.depth_range_near = 0.0f;
depth.depth_range_far = 1.0f;
primitive_restart.enabled = false;
primitive_restart.index = 0;
for (auto& item : color_mask) {
item.red_enabled = GL_TRUE;
item.green_enabled = GL_TRUE;
item.blue_enabled = GL_TRUE;
item.alpha_enabled = GL_TRUE;
}
color_mask.red_enabled = GL_TRUE;
color_mask.green_enabled = GL_TRUE;
color_mask.blue_enabled = GL_TRUE;
color_mask.alpha_enabled = GL_TRUE;
stencil.test_enabled = false;
auto reset_stencil = [](auto& config) {
config.test_func = GL_ALWAYS;
@@ -43,33 +45,19 @@ OpenGLState::OpenGLState() {
};
reset_stencil(stencil.front);
reset_stencil(stencil.back);
for (auto& item : viewports) {
item.x = 0;
item.y = 0;
item.width = 0;
item.height = 0;
item.depth_range_near = 0.0f;
item.depth_range_far = 1.0f;
}
scissor.enabled = false;
scissor.x = 0;
scissor.y = 0;
scissor.width = 0;
scissor.height = 0;
for (auto& item : blend) {
item.enabled = true;
item.rgb_equation = GL_FUNC_ADD;
item.a_equation = GL_FUNC_ADD;
item.src_rgb_func = GL_ONE;
item.dst_rgb_func = GL_ZERO;
item.src_a_func = GL_ONE;
item.dst_a_func = GL_ZERO;
}
independant_blend.enabled = false;
blend_color.red = 0.0f;
blend_color.green = 0.0f;
blend_color.blue = 0.0f;
blend_color.alpha = 0.0f;
blend.enabled = true;
blend.rgb_equation = GL_FUNC_ADD;
blend.a_equation = GL_FUNC_ADD;
blend.src_rgb_func = GL_ONE;
blend.dst_rgb_func = GL_ZERO;
blend.src_a_func = GL_ONE;
blend.dst_a_func = GL_ZERO;
blend.color.red = 0.0f;
blend.color.green = 0.0f;
blend.color.blue = 0.0f;
blend.color.alpha = 0.0f;
logic_op.enabled = false;
logic_op.operation = GL_COPY;
@@ -85,23 +73,23 @@ OpenGLState::OpenGLState() {
draw.shader_program = 0;
draw.program_pipeline = 0;
scissor.enabled = false;
scissor.x = 0;
scissor.y = 0;
scissor.width = 0;
scissor.height = 0;
viewport.x = 0;
viewport.y = 0;
viewport.width = 0;
viewport.height = 0;
clip_distance = {};
point.size = 1;
}
void OpenGLState::ApplyDefaultState() {
glDisable(GL_FRAMEBUFFER_SRGB);
glDisable(GL_CULL_FACE);
glDisable(GL_DEPTH_TEST);
glDisable(GL_PRIMITIVE_RESTART);
glDisable(GL_STENCIL_TEST);
glEnable(GL_BLEND);
glDisable(GL_COLOR_LOGIC_OP);
glDisable(GL_SCISSOR_TEST);
}
void OpenGLState::ApplySRgb() const {
void OpenGLState::Apply() const {
// sRGB
if (framebuffer_srgb.enabled != cur_state.framebuffer_srgb.enabled) {
if (framebuffer_srgb.enabled) {
@@ -112,281 +100,139 @@ void OpenGLState::ApplySRgb() const {
glDisable(GL_FRAMEBUFFER_SRGB);
}
}
}
void OpenGLState::ApplyCulling() const {
// Culling
const bool cull_changed = cull.enabled != cur_state.cull.enabled;
if (cull_changed) {
if (cull.enabled != cur_state.cull.enabled) {
if (cull.enabled) {
glEnable(GL_CULL_FACE);
} else {
glDisable(GL_CULL_FACE);
}
}
if (cull.enabled) {
if (cull_changed || cull.mode != cur_state.cull.mode) {
glCullFace(cull.mode);
}
if (cull_changed || cull.front_face != cur_state.cull.front_face) {
glFrontFace(cull.front_face);
}
if (cull.mode != cur_state.cull.mode) {
glCullFace(cull.mode);
}
}
void OpenGLState::ApplyColorMask() const {
if (GLAD_GL_ARB_viewport_array) {
for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
const auto& updated = color_mask[i];
const auto& current = cur_state.color_mask[i];
if (updated.red_enabled != current.red_enabled ||
updated.green_enabled != current.green_enabled ||
updated.blue_enabled != current.blue_enabled ||
updated.alpha_enabled != current.alpha_enabled) {
glColorMaski(static_cast<GLuint>(i), updated.red_enabled, updated.green_enabled,
updated.blue_enabled, updated.alpha_enabled);
}
}
} else {
const auto& updated = color_mask[0];
const auto& current = cur_state.color_mask[0];
if (updated.red_enabled != current.red_enabled ||
updated.green_enabled != current.green_enabled ||
updated.blue_enabled != current.blue_enabled ||
updated.alpha_enabled != current.alpha_enabled) {
glColorMask(updated.red_enabled, updated.green_enabled, updated.blue_enabled,
updated.alpha_enabled);
}
if (cull.front_face != cur_state.cull.front_face) {
glFrontFace(cull.front_face);
}
}
void OpenGLState::ApplyDepth() const {
// Depth test
const bool depth_test_changed = depth.test_enabled != cur_state.depth.test_enabled;
if (depth_test_changed) {
if (depth.test_enabled != cur_state.depth.test_enabled) {
if (depth.test_enabled) {
glEnable(GL_DEPTH_TEST);
} else {
glDisable(GL_DEPTH_TEST);
}
}
if (depth.test_enabled &&
(depth_test_changed || depth.test_func != cur_state.depth.test_func)) {
if (depth.test_func != cur_state.depth.test_func) {
glDepthFunc(depth.test_func);
}
// Depth mask
if (depth.write_mask != cur_state.depth.write_mask) {
glDepthMask(depth.write_mask);
}
}
void OpenGLState::ApplyPrimitiveRestart() const {
const bool primitive_restart_changed =
primitive_restart.enabled != cur_state.primitive_restart.enabled;
if (primitive_restart_changed) {
// Depth range
if (depth.depth_range_near != cur_state.depth.depth_range_near ||
depth.depth_range_far != cur_state.depth.depth_range_far) {
glDepthRange(depth.depth_range_near, depth.depth_range_far);
}
// Primitive restart
if (primitive_restart.enabled != cur_state.primitive_restart.enabled) {
if (primitive_restart.enabled) {
glEnable(GL_PRIMITIVE_RESTART);
} else {
glDisable(GL_PRIMITIVE_RESTART);
}
}
if (primitive_restart_changed ||
(primitive_restart.enabled &&
primitive_restart.index != cur_state.primitive_restart.index)) {
if (primitive_restart.index != cur_state.primitive_restart.index) {
glPrimitiveRestartIndex(primitive_restart.index);
}
}
void OpenGLState::ApplyStencilTest() const {
const bool stencil_test_changed = stencil.test_enabled != cur_state.stencil.test_enabled;
if (stencil_test_changed) {
// Color mask
if (color_mask.red_enabled != cur_state.color_mask.red_enabled ||
color_mask.green_enabled != cur_state.color_mask.green_enabled ||
color_mask.blue_enabled != cur_state.color_mask.blue_enabled ||
color_mask.alpha_enabled != cur_state.color_mask.alpha_enabled) {
glColorMask(color_mask.red_enabled, color_mask.green_enabled, color_mask.blue_enabled,
color_mask.alpha_enabled);
}
// Stencil test
if (stencil.test_enabled != cur_state.stencil.test_enabled) {
if (stencil.test_enabled) {
glEnable(GL_STENCIL_TEST);
} else {
glDisable(GL_STENCIL_TEST);
}
}
if (stencil.test_enabled) {
auto config_stencil = [stencil_test_changed](GLenum face, const auto& config,
const auto& prev_config) {
if (stencil_test_changed || config.test_func != prev_config.test_func ||
config.test_ref != prev_config.test_ref ||
config.test_mask != prev_config.test_mask) {
glStencilFuncSeparate(face, config.test_func, config.test_ref, config.test_mask);
}
if (stencil_test_changed || config.action_depth_fail != prev_config.action_depth_fail ||
config.action_depth_pass != prev_config.action_depth_pass ||
config.action_stencil_fail != prev_config.action_stencil_fail) {
glStencilOpSeparate(face, config.action_stencil_fail, config.action_depth_fail,
config.action_depth_pass);
}
if (config.write_mask != prev_config.write_mask) {
glStencilMaskSeparate(face, config.write_mask);
}
};
config_stencil(GL_FRONT, stencil.front, cur_state.stencil.front);
config_stencil(GL_BACK, stencil.back, cur_state.stencil.back);
}
}
auto config_stencil = [](GLenum face, const auto& config, const auto& prev_config) {
if (config.test_func != prev_config.test_func || config.test_ref != prev_config.test_ref ||
config.test_mask != prev_config.test_mask) {
glStencilFuncSeparate(face, config.test_func, config.test_ref, config.test_mask);
}
if (config.action_depth_fail != prev_config.action_depth_fail ||
config.action_depth_pass != prev_config.action_depth_pass ||
config.action_stencil_fail != prev_config.action_stencil_fail) {
glStencilOpSeparate(face, config.action_stencil_fail, config.action_depth_fail,
config.action_depth_pass);
}
if (config.write_mask != prev_config.write_mask) {
glStencilMaskSeparate(face, config.write_mask);
}
};
config_stencil(GL_FRONT, stencil.front, cur_state.stencil.front);
config_stencil(GL_BACK, stencil.back, cur_state.stencil.back);
void OpenGLState::ApplyScissor() const {
const bool scissor_changed = scissor.enabled != cur_state.scissor.enabled;
if (scissor_changed) {
if (scissor.enabled) {
glEnable(GL_SCISSOR_TEST);
} else {
glDisable(GL_SCISSOR_TEST);
}
}
if (scissor.enabled &&
(scissor_changed || scissor.x != cur_state.scissor.x || scissor.y != cur_state.scissor.y ||
scissor.width != cur_state.scissor.width || scissor.height != cur_state.scissor.height)) {
glScissor(scissor.x, scissor.y, scissor.width, scissor.height);
}
}
void OpenGLState::ApplyViewport() const {
if (GLAD_GL_ARB_viewport_array) {
for (GLuint i = 0;
i < static_cast<GLuint>(Tegra::Engines::Maxwell3D::Regs::NumRenderTargets); i++) {
const auto& current = cur_state.viewports[i];
const auto& updated = viewports[i];
if (updated.x != current.x || updated.y != current.y ||
updated.width != current.width || updated.height != current.height) {
glViewportIndexedf(i, updated.x, updated.y, updated.width, updated.height);
}
if (updated.depth_range_near != current.depth_range_near ||
updated.depth_range_far != current.depth_range_far) {
glDepthRangeIndexed(i, updated.depth_range_near, updated.depth_range_far);
}
}
} else {
const auto& current = cur_state.viewports[0];
const auto& updated = viewports[0];
if (updated.x != current.x || updated.y != current.y || updated.width != current.width ||
updated.height != current.height) {
glViewport(updated.x, updated.y, updated.width, updated.height);
}
if (updated.depth_range_near != current.depth_range_near ||
updated.depth_range_far != current.depth_range_far) {
glDepthRange(updated.depth_range_near, updated.depth_range_far);
}
}
}
void OpenGLState::ApplyGlobalBlending() const {
const Blend& current = cur_state.blend[0];
const Blend& updated = blend[0];
const bool blend_changed = updated.enabled != current.enabled;
if (blend_changed) {
if (updated.enabled) {
// Blending
if (blend.enabled != cur_state.blend.enabled) {
if (blend.enabled) {
ASSERT(!logic_op.enabled);
glEnable(GL_BLEND);
} else {
glDisable(GL_BLEND);
}
}
if (!updated.enabled) {
return;
}
if (updated.separate_alpha) {
if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
updated.dst_rgb_func != current.dst_rgb_func ||
updated.src_a_func != current.src_a_func || updated.dst_a_func != current.dst_a_func) {
glBlendFuncSeparate(updated.src_rgb_func, updated.dst_rgb_func, updated.src_a_func,
updated.dst_a_func);
}
if (blend_changed || updated.rgb_equation != current.rgb_equation ||
updated.a_equation != current.a_equation) {
glBlendEquationSeparate(updated.rgb_equation, updated.a_equation);
}
} else {
if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
updated.dst_rgb_func != current.dst_rgb_func) {
glBlendFunc(updated.src_rgb_func, updated.dst_rgb_func);
}
if (blend_changed || updated.rgb_equation != current.rgb_equation) {
glBlendEquation(updated.rgb_equation);
}
if (blend.color.red != cur_state.blend.color.red ||
blend.color.green != cur_state.blend.color.green ||
blend.color.blue != cur_state.blend.color.blue ||
blend.color.alpha != cur_state.blend.color.alpha) {
glBlendColor(blend.color.red, blend.color.green, blend.color.blue, blend.color.alpha);
}
}
void OpenGLState::ApplyTargetBlending(int target, bool force) const {
const Blend& updated = blend[target];
const Blend& current = cur_state.blend[target];
const bool blend_changed = updated.enabled != current.enabled || force;
if (blend_changed) {
if (updated.enabled) {
glEnablei(GL_BLEND, static_cast<GLuint>(target));
} else {
glDisablei(GL_BLEND, static_cast<GLuint>(target));
}
if (blend.src_rgb_func != cur_state.blend.src_rgb_func ||
blend.dst_rgb_func != cur_state.blend.dst_rgb_func ||
blend.src_a_func != cur_state.blend.src_a_func ||
blend.dst_a_func != cur_state.blend.dst_a_func) {
glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func, blend.src_a_func,
blend.dst_a_func);
}
if (!updated.enabled) {
return;
}
if (updated.separate_alpha) {
if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
updated.dst_rgb_func != current.dst_rgb_func ||
updated.src_a_func != current.src_a_func || updated.dst_a_func != current.dst_a_func) {
glBlendFuncSeparateiARB(static_cast<GLuint>(target), updated.src_rgb_func,
updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func);
}
if (blend_changed || updated.rgb_equation != current.rgb_equation ||
updated.a_equation != current.a_equation) {
glBlendEquationSeparateiARB(static_cast<GLuint>(target), updated.rgb_equation,
updated.a_equation);
}
} else {
if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
updated.dst_rgb_func != current.dst_rgb_func) {
glBlendFunciARB(static_cast<GLuint>(target), updated.src_rgb_func,
updated.dst_rgb_func);
}
if (blend_changed || updated.rgb_equation != current.rgb_equation) {
glBlendEquationiARB(static_cast<GLuint>(target), updated.rgb_equation);
}
if (blend.rgb_equation != cur_state.blend.rgb_equation ||
blend.a_equation != cur_state.blend.a_equation) {
glBlendEquationSeparate(blend.rgb_equation, blend.a_equation);
}
}
void OpenGLState::ApplyBlending() const {
if (independant_blend.enabled) {
for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) {
ApplyTargetBlending(i,
independant_blend.enabled != cur_state.independant_blend.enabled);
}
} else {
ApplyGlobalBlending();
}
if (blend_color.red != cur_state.blend_color.red ||
blend_color.green != cur_state.blend_color.green ||
blend_color.blue != cur_state.blend_color.blue ||
blend_color.alpha != cur_state.blend_color.alpha) {
glBlendColor(blend_color.red, blend_color.green, blend_color.blue, blend_color.alpha);
}
}
void OpenGLState::ApplyLogicOp() const {
const bool logic_op_changed = logic_op.enabled != cur_state.logic_op.enabled;
if (logic_op_changed) {
// Logic Operation
if (logic_op.enabled != cur_state.logic_op.enabled) {
if (logic_op.enabled) {
ASSERT(!blend.enabled);
glEnable(GL_COLOR_LOGIC_OP);
} else {
glDisable(GL_COLOR_LOGIC_OP);
}
}
if (logic_op.enabled &&
(logic_op_changed || logic_op.operation != cur_state.logic_op.operation)) {
if (logic_op.operation != cur_state.logic_op.operation) {
glLogicOp(logic_op.operation);
}
}
void OpenGLState::ApplyTextures() const {
// Textures
for (std::size_t i = 0; i < std::size(texture_units); ++i) {
const auto& texture_unit = texture_units[i];
const auto& cur_state_texture_unit = cur_state.texture_units[i];
@@ -405,29 +251,28 @@ void OpenGLState::ApplyTextures() const {
glTexParameteriv(texture_unit.target, GL_TEXTURE_SWIZZLE_RGBA, mask.data());
}
}
}
void OpenGLState::ApplySamplers() const {
bool has_delta{};
std::size_t first{}, last{};
std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> samplers;
for (std::size_t i = 0; i < std::size(samplers); ++i) {
samplers[i] = texture_units[i].sampler;
if (samplers[i] != cur_state.texture_units[i].sampler) {
if (!has_delta) {
first = i;
has_delta = true;
// Samplers
{
bool has_delta{};
std::size_t first{}, last{};
std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> samplers;
for (std::size_t i = 0; i < std::size(samplers); ++i) {
samplers[i] = texture_units[i].sampler;
if (samplers[i] != cur_state.texture_units[i].sampler) {
if (!has_delta) {
first = i;
has_delta = true;
}
last = i;
}
last = i;
}
if (has_delta) {
glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
samplers.data());
}
}
if (has_delta) {
glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
samplers.data());
}
}
void OpenGLState::Apply() const {
// Framebuffer
if (draw.read_framebuffer != cur_state.draw.read_framebuffer) {
glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer);
@@ -460,6 +305,27 @@ void OpenGLState::Apply() const {
if (draw.program_pipeline != cur_state.draw.program_pipeline) {
glBindProgramPipeline(draw.program_pipeline);
}
// Scissor test
if (scissor.enabled != cur_state.scissor.enabled) {
if (scissor.enabled) {
glEnable(GL_SCISSOR_TEST);
} else {
glDisable(GL_SCISSOR_TEST);
}
}
if (scissor.x != cur_state.scissor.x || scissor.y != cur_state.scissor.y ||
scissor.width != cur_state.scissor.width || scissor.height != cur_state.scissor.height) {
glScissor(scissor.x, scissor.y, scissor.width, scissor.height);
}
if (viewport.x != cur_state.viewport.x || viewport.y != cur_state.viewport.y ||
viewport.width != cur_state.viewport.width ||
viewport.height != cur_state.viewport.height) {
glViewport(viewport.x, viewport.y, viewport.width, viewport.height);
}
// Clip distance
for (std::size_t i = 0; i < clip_distance.size(); ++i) {
if (clip_distance[i] != cur_state.clip_distance[i]) {
@@ -470,22 +336,12 @@ void OpenGLState::Apply() const {
}
}
}
// Point
if (point.size != cur_state.point.size) {
glPointSize(point.size);
}
ApplyColorMask();
ApplyViewport();
ApplyScissor();
ApplyStencilTest();
ApplySRgb();
ApplyCulling();
ApplyDepth();
ApplyPrimitiveRestart();
ApplyBlending();
ApplyLogicOp();
ApplyTextures();
ApplySamplers();
cur_state = *this;
}

View File

@@ -46,9 +46,11 @@ public:
} cull;
struct {
bool test_enabled; // GL_DEPTH_TEST
GLenum test_func; // GL_DEPTH_FUNC
GLboolean write_mask; // GL_DEPTH_WRITEMASK
bool test_enabled; // GL_DEPTH_TEST
GLenum test_func; // GL_DEPTH_FUNC
GLboolean write_mask; // GL_DEPTH_WRITEMASK
GLfloat depth_range_near; // GL_DEPTH_RANGE
GLfloat depth_range_far; // GL_DEPTH_RANGE
} depth;
struct {
@@ -56,14 +58,13 @@ public:
GLuint index;
} primitive_restart; // GL_PRIMITIVE_RESTART
struct ColorMask {
struct {
GLboolean red_enabled;
GLboolean green_enabled;
GLboolean blue_enabled;
GLboolean alpha_enabled;
};
std::array<ColorMask, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets>
color_mask; // GL_COLOR_WRITEMASK
} color_mask; // GL_COLOR_WRITEMASK
struct {
bool test_enabled; // GL_STENCIL_TEST
struct {
@@ -77,28 +78,22 @@ public:
} front, back;
} stencil;
struct Blend {
struct {
bool enabled; // GL_BLEND
bool separate_alpha; // Independent blend enabled
GLenum rgb_equation; // GL_BLEND_EQUATION_RGB
GLenum a_equation; // GL_BLEND_EQUATION_ALPHA
GLenum src_rgb_func; // GL_BLEND_SRC_RGB
GLenum dst_rgb_func; // GL_BLEND_DST_RGB
GLenum src_a_func; // GL_BLEND_SRC_ALPHA
GLenum dst_a_func; // GL_BLEND_DST_ALPHA
};
std::array<Blend, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> blend;
struct {
bool enabled;
} independant_blend;
struct {
GLclampf red;
GLclampf green;
GLclampf blue;
GLclampf alpha;
} blend_color; // GL_BLEND_COLOR
struct {
GLclampf red;
GLclampf green;
GLclampf blue;
GLclampf alpha;
} color; // GL_BLEND_COLOR
} blend;
struct {
bool enabled; // GL_LOGIC_OP_MODE
@@ -143,16 +138,6 @@ public:
GLuint program_pipeline; // GL_PROGRAM_PIPELINE_BINDING
} draw;
struct viewport {
GLfloat x;
GLfloat y;
GLfloat width;
GLfloat height;
GLfloat depth_range_near; // GL_DEPTH_RANGE
GLfloat depth_range_far; // GL_DEPTH_RANGE
};
std::array<viewport, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> viewports;
struct {
bool enabled; // GL_SCISSOR_TEST
GLint x;
@@ -161,6 +146,13 @@ public:
GLsizei height;
} scissor;
struct {
GLint x;
GLint y;
GLsizei width;
GLsizei height;
} viewport;
struct {
float size; // GL_POINT_SIZE
} point;
@@ -181,8 +173,7 @@ public:
}
/// Apply this state as the current OpenGL state
void Apply() const;
/// Set the initial OpenGL state
static void ApplyDefaultState();
/// Resets any references to the given resource
OpenGLState& UnbindTexture(GLuint handle);
OpenGLState& ResetSampler(GLuint handle);
@@ -197,20 +188,6 @@ private:
// Workaround for sRGB problems caused by
// QT not supporting srgb output
static bool s_rgb_used;
void ApplySRgb() const;
void ApplyCulling() const;
void ApplyColorMask() const;
void ApplyDepth() const;
void ApplyPrimitiveRestart() const;
void ApplyStencilTest() const;
void ApplyViewport() const;
void ApplyTargetBlending(int target, bool force) const;
void ApplyGlobalBlending() const;
void ApplyBlending() const;
void ApplyLogicOp() const;
void ApplyTextures() const;
void ApplySamplers() const;
void ApplyScissor() const;
};
} // namespace OpenGL

View File

@@ -6,13 +6,9 @@
#include <vector>
#include "common/alignment.h"
#include "common/assert.h"
#include "common/microprofile.h"
#include "video_core/renderer_opengl/gl_state.h"
#include "video_core/renderer_opengl/gl_stream_buffer.h"
MICROPROFILE_DEFINE(OpenGL_StreamBuffer, "OpenGL", "Stream Buffer Orphaning",
MP_RGB(128, 128, 192));
namespace OpenGL {
OGLStreamBuffer::OGLStreamBuffer(GLenum target, GLsizeiptr size, bool prefer_coherent)
@@ -79,7 +75,6 @@ std::tuple<u8*, GLintptr, bool> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr a
}
if (invalidate || !persistent) {
MICROPROFILE_SCOPE(OpenGL_StreamBuffer);
GLbitfield flags = GL_MAP_WRITE_BIT | (persistent ? GL_MAP_PERSISTENT_BIT : 0) |
(coherent ? GL_MAP_COHERENT_BIT : GL_MAP_FLUSH_EXPLICIT_BIT) |
(invalidate ? GL_MAP_INVALIDATE_BUFFER_BIT : GL_MAP_UNSYNCHRONIZED_BIT);

View File

@@ -135,29 +135,12 @@ inline GLenum PrimitiveTopology(Maxwell::PrimitiveTopology topology) {
return {};
}
inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode,
Tegra::Texture::TextureMipmapFilter mip_filter_mode) {
inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode) {
switch (filter_mode) {
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;
}
}
case Tegra::Texture::TextureFilter::Linear:
return GL_LINEAR;
case Tegra::Texture::TextureFilter::Nearest:
return GL_NEAREST;
}
LOG_CRITICAL(Render_OpenGL, "Unimplemented texture filter mode={}",
static_cast<u32>(filter_mode));

View File

@@ -115,8 +115,7 @@ RendererOpenGL::RendererOpenGL(Core::Frontend::EmuWindow& window)
RendererOpenGL::~RendererOpenGL() = default;
/// Swap buffers (render frame)
void RendererOpenGL::SwapBuffers(
std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
void RendererOpenGL::SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) {
ScopeAcquireGLContext acquire_context{render_window};
Core::System::GetInstance().GetPerfStats().EndSystemFrame();
@@ -125,11 +124,11 @@ void RendererOpenGL::SwapBuffers(
OpenGLState prev_state = OpenGLState::GetCurState();
state.Apply();
if (framebuffer) {
if (framebuffer != boost::none) {
// If framebuffer is provided, reload it from memory to a texture
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) {
if (screen_info.texture.width != (GLsizei)framebuffer->width ||
screen_info.texture.height != (GLsizei)framebuffer->height ||
screen_info.texture.pixel_format != framebuffer->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.

View File

@@ -51,8 +51,7 @@ public:
~RendererOpenGL() override;
/// Swap buffers (render frame)
void SwapBuffers(
std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) override;
void SwapBuffers(boost::optional<const Tegra::FramebufferConfig&> framebuffer) override;
/// Initialize the renderer
bool Init() override;

View File

@@ -1,38 +0,0 @@
// 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

View File

@@ -1,15 +0,0 @@
// 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

View File

@@ -1,486 +0,0 @@
// Copyright 2014 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include "common/common_types.h"
#include "common/math_util.h"
#include "video_core/surface.h"
namespace VideoCore::Surface {
SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type) {
switch (texture_type) {
case Tegra::Texture::TextureType::Texture1D:
return SurfaceTarget::Texture1D;
case Tegra::Texture::TextureType::Texture2D:
case Tegra::Texture::TextureType::Texture2DNoMipmap:
return SurfaceTarget::Texture2D;
case Tegra::Texture::TextureType::Texture3D:
return SurfaceTarget::Texture3D;
case Tegra::Texture::TextureType::TextureCubemap:
return SurfaceTarget::TextureCubemap;
case Tegra::Texture::TextureType::TextureCubeArray:
return SurfaceTarget::TextureCubeArray;
case Tegra::Texture::TextureType::Texture1DArray:
return SurfaceTarget::Texture1DArray;
case Tegra::Texture::TextureType::Texture2DArray:
return SurfaceTarget::Texture2DArray;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", static_cast<u32>(texture_type));
UNREACHABLE();
return SurfaceTarget::Texture2D;
}
}
bool SurfaceTargetIsLayered(SurfaceTarget target) {
switch (target) {
case SurfaceTarget::Texture1D:
case SurfaceTarget::Texture2D:
case SurfaceTarget::Texture3D:
return false;
case SurfaceTarget::Texture1DArray:
case SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubemap:
case SurfaceTarget::TextureCubeArray:
return true;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
UNREACHABLE();
return false;
}
}
PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
switch (format) {
case Tegra::DepthFormat::S8_Z24_UNORM:
return PixelFormat::S8Z24;
case Tegra::DepthFormat::Z24_S8_UNORM:
return PixelFormat::Z24S8;
case Tegra::DepthFormat::Z32_FLOAT:
return PixelFormat::Z32F;
case Tegra::DepthFormat::Z16_UNORM:
return PixelFormat::Z16;
case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
return PixelFormat::Z32FS8;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
switch (format) {
// TODO (Hexagon12): Converting SRGBA to RGBA is a hack and doesn't completely correct the
// gamma.
case Tegra::RenderTargetFormat::RGBA8_SRGB:
return PixelFormat::RGBA8_SRGB;
case Tegra::RenderTargetFormat::RGBA8_UNORM:
return PixelFormat::ABGR8U;
case Tegra::RenderTargetFormat::RGBA8_SNORM:
return PixelFormat::ABGR8S;
case Tegra::RenderTargetFormat::RGBA8_UINT:
return PixelFormat::ABGR8UI;
case Tegra::RenderTargetFormat::BGRA8_SRGB:
return PixelFormat::BGRA8_SRGB;
case Tegra::RenderTargetFormat::BGRA8_UNORM:
return PixelFormat::BGRA8;
case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
return PixelFormat::A2B10G10R10U;
case Tegra::RenderTargetFormat::RGBA16_FLOAT:
return PixelFormat::RGBA16F;
case Tegra::RenderTargetFormat::RGBA16_UNORM:
return PixelFormat::RGBA16U;
case Tegra::RenderTargetFormat::RGBA16_UINT:
return PixelFormat::RGBA16UI;
case Tegra::RenderTargetFormat::RGBA32_FLOAT:
return PixelFormat::RGBA32F;
case Tegra::RenderTargetFormat::RG32_FLOAT:
return PixelFormat::RG32F;
case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
return PixelFormat::R11FG11FB10F;
case Tegra::RenderTargetFormat::B5G6R5_UNORM:
return PixelFormat::B5G6R5U;
case Tegra::RenderTargetFormat::BGR5A1_UNORM:
return PixelFormat::A1B5G5R5U;
case Tegra::RenderTargetFormat::RGBA32_UINT:
return PixelFormat::RGBA32UI;
case Tegra::RenderTargetFormat::R8_UNORM:
return PixelFormat::R8U;
case Tegra::RenderTargetFormat::R8_UINT:
return PixelFormat::R8UI;
case Tegra::RenderTargetFormat::RG16_FLOAT:
return PixelFormat::RG16F;
case Tegra::RenderTargetFormat::RG16_UINT:
return PixelFormat::RG16UI;
case Tegra::RenderTargetFormat::RG16_SINT:
return PixelFormat::RG16I;
case Tegra::RenderTargetFormat::RG16_UNORM:
return PixelFormat::RG16;
case Tegra::RenderTargetFormat::RG16_SNORM:
return PixelFormat::RG16S;
case Tegra::RenderTargetFormat::RG8_UNORM:
return PixelFormat::RG8U;
case Tegra::RenderTargetFormat::RG8_SNORM:
return PixelFormat::RG8S;
case Tegra::RenderTargetFormat::R16_FLOAT:
return PixelFormat::R16F;
case Tegra::RenderTargetFormat::R16_UNORM:
return PixelFormat::R16U;
case Tegra::RenderTargetFormat::R16_SNORM:
return PixelFormat::R16S;
case Tegra::RenderTargetFormat::R16_UINT:
return PixelFormat::R16UI;
case Tegra::RenderTargetFormat::R16_SINT:
return PixelFormat::R16I;
case Tegra::RenderTargetFormat::R32_FLOAT:
return PixelFormat::R32F;
case Tegra::RenderTargetFormat::R32_UINT:
return PixelFormat::R32UI;
case Tegra::RenderTargetFormat::RG32_UINT:
return PixelFormat::RG32UI;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
Tegra::Texture::ComponentType component_type,
bool is_srgb) {
// TODO(Subv): Properly implement this
switch (format) {
case Tegra::Texture::TextureFormat::A8R8G8B8:
if (is_srgb) {
return PixelFormat::RGBA8_SRGB;
}
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::ABGR8U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::ABGR8S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::ABGR8UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::B5G6R5:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::B5G6R5U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::A2B10G10R10:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::A2B10G10R10U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::A1B5G5R5:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::A1B5G5R5U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R8:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::R8U;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R8UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::G8R8:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::G8R8U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::G8R8S;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R16_G16_B16_A16:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::RGBA16U;
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGBA16F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::BF10GF11RF11:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R11FG11FB10F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32_B32_A32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGBA32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RGBA32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RG32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RG32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32_B32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGB32F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R16:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R16F;
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::R16U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::R16S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R16UI;
case Tegra::Texture::ComponentType::SINT:
return PixelFormat::R16I;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::ZF32:
return PixelFormat::Z32F;
case Tegra::Texture::TextureFormat::Z16:
return PixelFormat::Z16;
case Tegra::Texture::TextureFormat::Z24S8:
return PixelFormat::Z24S8;
case Tegra::Texture::TextureFormat::DXT1:
return is_srgb ? PixelFormat::DXT1_SRGB : PixelFormat::DXT1;
case Tegra::Texture::TextureFormat::DXT23:
return is_srgb ? PixelFormat::DXT23_SRGB : PixelFormat::DXT23;
case Tegra::Texture::TextureFormat::DXT45:
return is_srgb ? PixelFormat::DXT45_SRGB : PixelFormat::DXT45;
case Tegra::Texture::TextureFormat::DXN1:
return PixelFormat::DXN1;
case Tegra::Texture::TextureFormat::DXN2:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::DXN2UNORM;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::DXN2SNORM;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::BC7U:
return is_srgb ? PixelFormat::BC7U_SRGB : PixelFormat::BC7U;
case Tegra::Texture::TextureFormat::BC6H_UF16:
return PixelFormat::BC6H_UF16;
case Tegra::Texture::TextureFormat::BC6H_SF16:
return PixelFormat::BC6H_SF16;
case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
return is_srgb ? PixelFormat::ASTC_2D_4X4_SRGB : PixelFormat::ASTC_2D_4X4;
case Tegra::Texture::TextureFormat::ASTC_2D_5X4:
return is_srgb ? PixelFormat::ASTC_2D_5X4_SRGB : PixelFormat::ASTC_2D_5X4;
case Tegra::Texture::TextureFormat::ASTC_2D_5X5:
return is_srgb ? PixelFormat::ASTC_2D_5X5_SRGB : PixelFormat::ASTC_2D_5X5;
case Tegra::Texture::TextureFormat::ASTC_2D_8X8:
return is_srgb ? PixelFormat::ASTC_2D_8X8_SRGB : PixelFormat::ASTC_2D_8X8;
case Tegra::Texture::TextureFormat::ASTC_2D_8X5:
return is_srgb ? PixelFormat::ASTC_2D_8X5_SRGB : PixelFormat::ASTC_2D_8X5;
case Tegra::Texture::TextureFormat::R16_G16:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RG16F;
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::RG16;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::RG16S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RG16UI;
case Tegra::Texture::ComponentType::SINT:
return PixelFormat::RG16I;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type));
UNREACHABLE();
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}, component_type={}", static_cast<u32>(format),
static_cast<u32>(component_type));
UNREACHABLE();
}
}
ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
// TODO(Subv): Implement more component types
switch (type) {
case Tegra::Texture::ComponentType::UNORM:
return ComponentType::UNorm;
case Tegra::Texture::ComponentType::FLOAT:
return ComponentType::Float;
case Tegra::Texture::ComponentType::SNORM:
return ComponentType::SNorm;
case Tegra::Texture::ComponentType::UINT:
return ComponentType::UInt;
case Tegra::Texture::ComponentType::SINT:
return ComponentType::SInt;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented component type={}", static_cast<u32>(type));
UNREACHABLE();
}
}
ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format) {
// TODO(Subv): Implement more render targets
switch (format) {
case Tegra::RenderTargetFormat::RGBA8_UNORM:
case Tegra::RenderTargetFormat::RGBA8_SRGB:
case Tegra::RenderTargetFormat::BGRA8_UNORM:
case Tegra::RenderTargetFormat::BGRA8_SRGB:
case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
case Tegra::RenderTargetFormat::R8_UNORM:
case Tegra::RenderTargetFormat::RG16_UNORM:
case Tegra::RenderTargetFormat::R16_UNORM:
case Tegra::RenderTargetFormat::B5G6R5_UNORM:
case Tegra::RenderTargetFormat::BGR5A1_UNORM:
case Tegra::RenderTargetFormat::RG8_UNORM:
case Tegra::RenderTargetFormat::RGBA16_UNORM:
return ComponentType::UNorm;
case Tegra::RenderTargetFormat::RGBA8_SNORM:
case Tegra::RenderTargetFormat::RG16_SNORM:
case Tegra::RenderTargetFormat::R16_SNORM:
case Tegra::RenderTargetFormat::RG8_SNORM:
return ComponentType::SNorm;
case Tegra::RenderTargetFormat::RGBA16_FLOAT:
case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
case Tegra::RenderTargetFormat::RGBA32_FLOAT:
case Tegra::RenderTargetFormat::RG32_FLOAT:
case Tegra::RenderTargetFormat::RG16_FLOAT:
case Tegra::RenderTargetFormat::R16_FLOAT:
case Tegra::RenderTargetFormat::R32_FLOAT:
return ComponentType::Float;
case Tegra::RenderTargetFormat::RGBA32_UINT:
case Tegra::RenderTargetFormat::RGBA16_UINT:
case Tegra::RenderTargetFormat::RG16_UINT:
case Tegra::RenderTargetFormat::R8_UINT:
case Tegra::RenderTargetFormat::R16_UINT:
case Tegra::RenderTargetFormat::RG32_UINT:
case Tegra::RenderTargetFormat::R32_UINT:
case Tegra::RenderTargetFormat::RGBA8_UINT:
return ComponentType::UInt;
case Tegra::RenderTargetFormat::RG16_SINT:
case Tegra::RenderTargetFormat::R16_SINT:
return ComponentType::SInt;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
switch (format) {
case Tegra::FramebufferConfig::PixelFormat::ABGR8:
return PixelFormat::ABGR8U;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format) {
switch (format) {
case Tegra::DepthFormat::Z16_UNORM:
case Tegra::DepthFormat::S8_Z24_UNORM:
case Tegra::DepthFormat::Z24_S8_UNORM:
return ComponentType::UNorm;
case Tegra::DepthFormat::Z32_FLOAT:
case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
return ComponentType::Float;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
SurfaceType GetFormatType(PixelFormat pixel_format) {
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxColorFormat)) {
return SurfaceType::ColorTexture;
}
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxDepthFormat)) {
return SurfaceType::Depth;
}
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxDepthStencilFormat)) {
return SurfaceType::DepthStencil;
}
// TODO(Subv): Implement the other formats
ASSERT(false);
return SurfaceType::Invalid;
}
bool IsPixelFormatASTC(PixelFormat format) {
switch (format) {
case PixelFormat::ASTC_2D_4X4:
case PixelFormat::ASTC_2D_5X4:
case PixelFormat::ASTC_2D_5X5:
case PixelFormat::ASTC_2D_8X8:
case PixelFormat::ASTC_2D_8X5:
case PixelFormat::ASTC_2D_4X4_SRGB:
case PixelFormat::ASTC_2D_5X4_SRGB:
case PixelFormat::ASTC_2D_5X5_SRGB:
case PixelFormat::ASTC_2D_8X8_SRGB:
case PixelFormat::ASTC_2D_8X5_SRGB:
return true;
default:
return false;
}
}
std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) {
return {GetDefaultBlockWidth(format), GetDefaultBlockHeight(format)};
}
bool IsFormatBCn(PixelFormat format) {
switch (format) {
case PixelFormat::DXT1:
case PixelFormat::DXT23:
case PixelFormat::DXT45:
case PixelFormat::DXN1:
case PixelFormat::DXN2SNORM:
case PixelFormat::DXN2UNORM:
case PixelFormat::BC7U:
case PixelFormat::BC6H_UF16:
case PixelFormat::BC6H_SF16:
case PixelFormat::DXT1_SRGB:
case PixelFormat::DXT23_SRGB:
case PixelFormat::DXT45_SRGB:
case PixelFormat::BC7U_SRGB:
return true;
}
return false;
}
} // namespace VideoCore::Surface

View File

@@ -1,467 +0,0 @@
// Copyright 2014 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <climits>
#include <utility>
#include "common/assert.h"
#include "common/common_types.h"
#include "common/logging/log.h"
#include "video_core/gpu.h"
#include "video_core/textures/texture.h"
namespace VideoCore::Surface {
enum class PixelFormat {
ABGR8U = 0,
ABGR8S = 1,
ABGR8UI = 2,
B5G6R5U = 3,
A2B10G10R10U = 4,
A1B5G5R5U = 5,
R8U = 6,
R8UI = 7,
RGBA16F = 8,
RGBA16U = 9,
RGBA16UI = 10,
R11FG11FB10F = 11,
RGBA32UI = 12,
DXT1 = 13,
DXT23 = 14,
DXT45 = 15,
DXN1 = 16, // This is also known as BC4
DXN2UNORM = 17,
DXN2SNORM = 18,
BC7U = 19,
BC6H_UF16 = 20,
BC6H_SF16 = 21,
ASTC_2D_4X4 = 22,
G8R8U = 23,
G8R8S = 24,
BGRA8 = 25,
RGBA32F = 26,
RG32F = 27,
R32F = 28,
R16F = 29,
R16U = 30,
R16S = 31,
R16UI = 32,
R16I = 33,
RG16 = 34,
RG16F = 35,
RG16UI = 36,
RG16I = 37,
RG16S = 38,
RGB32F = 39,
RGBA8_SRGB = 40,
RG8U = 41,
RG8S = 42,
RG32UI = 43,
R32UI = 44,
ASTC_2D_8X8 = 45,
ASTC_2D_8X5 = 46,
ASTC_2D_5X4 = 47,
BGRA8_SRGB = 48,
DXT1_SRGB = 49,
DXT23_SRGB = 50,
DXT45_SRGB = 51,
BC7U_SRGB = 52,
ASTC_2D_4X4_SRGB = 53,
ASTC_2D_8X8_SRGB = 54,
ASTC_2D_8X5_SRGB = 55,
ASTC_2D_5X4_SRGB = 56,
ASTC_2D_5X5 = 57,
ASTC_2D_5X5_SRGB = 58,
MaxColorFormat,
// Depth formats
Z32F = 59,
Z16 = 60,
MaxDepthFormat,
// DepthStencil formats
Z24S8 = 61,
S8Z24 = 62,
Z32FS8 = 63,
MaxDepthStencilFormat,
Max = MaxDepthStencilFormat,
Invalid = 255,
};
static constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max);
enum class ComponentType {
Invalid = 0,
SNorm = 1,
UNorm = 2,
SInt = 3,
UInt = 4,
Float = 5,
};
enum class SurfaceType {
ColorTexture = 0,
Depth = 1,
DepthStencil = 2,
Fill = 3,
Invalid = 4,
};
enum class SurfaceTarget {
Texture1D,
Texture2D,
Texture3D,
Texture1DArray,
Texture2DArray,
TextureCubemap,
TextureCubeArray,
};
/**
* Gets the compression factor for the specified PixelFormat. This applies to just the
* "compressed width" and "compressed height", not the overall compression factor of a
* compressed image. This is used for maintaining proper surface sizes for compressed
* texture formats.
*/
static constexpr u32 GetCompressionFactor(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{
1, // ABGR8U
1, // ABGR8S
1, // ABGR8UI
1, // B5G6R5U
1, // A2B10G10R10U
1, // A1B5G5R5U
1, // R8U
1, // R8UI
1, // RGBA16F
1, // RGBA16U
1, // RGBA16UI
1, // R11FG11FB10F
1, // RGBA32UI
4, // DXT1
4, // DXT23
4, // DXT45
4, // DXN1
4, // DXN2UNORM
4, // DXN2SNORM
4, // BC7U
4, // BC6H_UF16
4, // BC6H_SF16
4, // ASTC_2D_4X4
1, // G8R8U
1, // G8R8S
1, // BGRA8
1, // RGBA32F
1, // RG32F
1, // R32F
1, // R16F
1, // R16U
1, // R16S
1, // R16UI
1, // R16I
1, // RG16
1, // RG16F
1, // RG16UI
1, // RG16I
1, // RG16S
1, // RGB32F
1, // RGBA8_SRGB
1, // RG8U
1, // RG8S
1, // RG32UI
1, // R32UI
4, // ASTC_2D_8X8
4, // ASTC_2D_8X5
4, // ASTC_2D_5X4
1, // BGRA8_SRGB
4, // DXT1_SRGB
4, // DXT23_SRGB
4, // DXT45_SRGB
4, // BC7U_SRGB
4, // ASTC_2D_4X4_SRGB
4, // ASTC_2D_8X8_SRGB
4, // ASTC_2D_8X5_SRGB
4, // ASTC_2D_5X4_SRGB
4, // ASTC_2D_5X5
4, // ASTC_2D_5X5_SRGB
1, // Z32F
1, // Z16
1, // Z24S8
1, // S8Z24
1, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < compression_factor_table.size());
return compression_factor_table[static_cast<std::size_t>(format)];
}
static constexpr u32 GetDefaultBlockWidth(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> block_width_table = {{
1, // ABGR8U
1, // ABGR8S
1, // ABGR8UI
1, // B5G6R5U
1, // A2B10G10R10U
1, // A1B5G5R5U
1, // R8U
1, // R8UI
1, // RGBA16F
1, // RGBA16U
1, // RGBA16UI
1, // R11FG11FB10F
1, // RGBA32UI
4, // DXT1
4, // DXT23
4, // DXT45
4, // DXN1
4, // DXN2UNORM
4, // DXN2SNORM
4, // BC7U
4, // BC6H_UF16
4, // BC6H_SF16
4, // ASTC_2D_4X4
1, // G8R8U
1, // G8R8S
1, // BGRA8
1, // RGBA32F
1, // RG32F
1, // R32F
1, // R16F
1, // R16U
1, // R16S
1, // R16UI
1, // R16I
1, // RG16
1, // RG16F
1, // RG16UI
1, // RG16I
1, // RG16S
1, // RGB32F
1, // RGBA8_SRGB
1, // RG8U
1, // RG8S
1, // RG32UI
1, // R32UI
8, // ASTC_2D_8X8
8, // ASTC_2D_8X5
5, // ASTC_2D_5X4
1, // BGRA8_SRGB
4, // DXT1_SRGB
4, // DXT23_SRGB
4, // DXT45_SRGB
4, // BC7U_SRGB
4, // ASTC_2D_4X4_SRGB
8, // ASTC_2D_8X8_SRGB
8, // ASTC_2D_8X5_SRGB
5, // ASTC_2D_5X4_SRGB
5, // ASTC_2D_5X5
5, // ASTC_2D_5X5_SRGB
1, // Z32F
1, // Z16
1, // Z24S8
1, // S8Z24
1, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < block_width_table.size());
return block_width_table[static_cast<std::size_t>(format)];
}
static constexpr u32 GetDefaultBlockHeight(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> block_height_table = {{
1, // ABGR8U
1, // ABGR8S
1, // ABGR8UI
1, // B5G6R5U
1, // A2B10G10R10U
1, // A1B5G5R5U
1, // R8U
1, // R8UI
1, // RGBA16F
1, // RGBA16U
1, // RGBA16UI
1, // R11FG11FB10F
1, // RGBA32UI
4, // DXT1
4, // DXT23
4, // DXT45
4, // DXN1
4, // DXN2UNORM
4, // DXN2SNORM
4, // BC7U
4, // BC6H_UF16
4, // BC6H_SF16
4, // ASTC_2D_4X4
1, // G8R8U
1, // G8R8S
1, // BGRA8
1, // RGBA32F
1, // RG32F
1, // R32F
1, // R16F
1, // R16U
1, // R16S
1, // R16UI
1, // R16I
1, // RG16
1, // RG16F
1, // RG16UI
1, // RG16I
1, // RG16S
1, // RGB32F
1, // RGBA8_SRGB
1, // RG8U
1, // RG8S
1, // RG32UI
1, // R32UI
8, // ASTC_2D_8X8
5, // ASTC_2D_8X5
4, // ASTC_2D_5X4
1, // BGRA8_SRGB
4, // DXT1_SRGB
4, // DXT23_SRGB
4, // DXT45_SRGB
4, // BC7U_SRGB
4, // ASTC_2D_4X4_SRGB
8, // ASTC_2D_8X8_SRGB
5, // ASTC_2D_8X5_SRGB
4, // ASTC_2D_5X4_SRGB
5, // ASTC_2D_5X5
5, // ASTC_2D_5X5_SRGB
1, // Z32F
1, // Z16
1, // Z24S8
1, // S8Z24
1, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < block_height_table.size());
return block_height_table[static_cast<std::size_t>(format)];
}
static constexpr u32 GetFormatBpp(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
32, // ABGR8U
32, // ABGR8S
32, // ABGR8UI
16, // B5G6R5U
32, // A2B10G10R10U
16, // A1B5G5R5U
8, // R8U
8, // R8UI
64, // RGBA16F
64, // RGBA16U
64, // RGBA16UI
32, // R11FG11FB10F
128, // RGBA32UI
64, // DXT1
128, // DXT23
128, // DXT45
64, // DXN1
128, // DXN2UNORM
128, // DXN2SNORM
128, // BC7U
128, // BC6H_UF16
128, // BC6H_SF16
128, // ASTC_2D_4X4
16, // G8R8U
16, // G8R8S
32, // BGRA8
128, // RGBA32F
64, // RG32F
32, // R32F
16, // R16F
16, // R16U
16, // R16S
16, // R16UI
16, // R16I
32, // RG16
32, // RG16F
32, // RG16UI
32, // RG16I
32, // RG16S
96, // RGB32F
32, // RGBA8_SRGB
16, // RG8U
16, // RG8S
64, // RG32UI
32, // R32UI
128, // ASTC_2D_8X8
128, // ASTC_2D_8X5
128, // ASTC_2D_5X4
32, // BGRA8_SRGB
64, // DXT1_SRGB
128, // DXT23_SRGB
128, // DXT45_SRGB
128, // BC7U
128, // ASTC_2D_4X4_SRGB
128, // ASTC_2D_8X8_SRGB
128, // ASTC_2D_8X5_SRGB
128, // ASTC_2D_5X4_SRGB
128, // ASTC_2D_5X5
128, // ASTC_2D_5X5_SRGB
32, // Z32F
16, // Z16
32, // Z24S8
32, // S8Z24
64, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < bpp_table.size());
return bpp_table[static_cast<std::size_t>(format)];
}
/// Returns the sizer in bytes of the specified pixel format
static constexpr u32 GetBytesPerPixel(PixelFormat pixel_format) {
if (pixel_format == PixelFormat::Invalid) {
return 0;
}
return GetFormatBpp(pixel_format) / CHAR_BIT;
}
SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type);
bool SurfaceTargetIsLayered(SurfaceTarget target);
PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format);
PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format);
PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
Tegra::Texture::ComponentType component_type,
bool is_srgb);
ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type);
ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format);
PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format);
ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format);
SurfaceType GetFormatType(PixelFormat pixel_format);
bool IsPixelFormatASTC(PixelFormat format);
std::pair<u32, u32> GetASTCBlockSize(PixelFormat format);
/// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN
bool IsFormatBCn(PixelFormat format);
} // namespace VideoCore::Surface

View File

@@ -1598,29 +1598,27 @@ static void DecompressBlock(uint8_t inBuf[16], const uint32_t blockWidth,
namespace Tegra::Texture::ASTC {
std::vector<uint8_t> Decompress(std::vector<uint8_t>& data, uint32_t width, uint32_t height,
uint32_t depth, uint32_t block_width, uint32_t block_height) {
uint32_t block_width, uint32_t block_height) {
uint32_t blockIdx = 0;
std::vector<uint8_t> outData(height * width * depth * 4);
for (uint32_t k = 0; k < depth; k++) {
for (uint32_t j = 0; j < height; j += block_height) {
for (uint32_t i = 0; i < width; i += block_width) {
std::vector<uint8_t> outData(height * width * 4);
for (uint32_t j = 0; j < height; j += block_height) {
for (uint32_t i = 0; i < width; i += block_width) {
uint8_t* blockPtr = data.data() + blockIdx * 16;
uint8_t* blockPtr = data.data() + blockIdx * 16;
// Blocks can be at most 12x12
uint32_t uncompData[144];
ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData);
// Blocks can be at most 12x12
uint32_t uncompData[144];
ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData);
uint32_t decompWidth = std::min(block_width, width - i);
uint32_t decompHeight = std::min(block_height, height - j);
uint32_t decompWidth = std::min(block_width, width - i);
uint32_t decompHeight = std::min(block_height, height - j);
uint8_t* outRow = outData.data() + (j * width + i) * 4;
for (uint32_t jj = 0; jj < decompHeight; jj++) {
memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4);
}
blockIdx++;
uint8_t* outRow = outData.data() + (j * width + i) * 4;
for (uint32_t jj = 0; jj < decompHeight; jj++) {
memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4);
}
blockIdx++;
}
}

View File

@@ -10,6 +10,6 @@
namespace Tegra::Texture::ASTC {
std::vector<uint8_t> Decompress(std::vector<uint8_t>& data, uint32_t width, uint32_t height,
uint32_t depth, uint32_t block_width, uint32_t block_height);
uint32_t block_width, uint32_t block_height);
} // namespace Tegra::Texture::ASTC

View File

@@ -227,14 +227,12 @@ u32 BytesPerPixel(TextureFormat format) {
}
}
std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size_x, u32 tile_size_y,
u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
u32 block_height, u32 block_depth) {
std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
u32 height, u32 depth, u32 block_height, u32 block_depth) {
std::vector<u8> unswizzled_data(width * height * depth * bytes_per_pixel);
CopySwizzledData((width + tile_size_x - 1) / tile_size_x,
(height + tile_size_y - 1) / tile_size_y, depth, bytes_per_pixel,
bytes_per_pixel, Memory::GetPointer(address), unswizzled_data.data(), true,
block_height, block_depth);
CopySwizzledData(width / tile_size, height / tile_size, depth, bytes_per_pixel, bytes_per_pixel,
Memory::GetPointer(address), unswizzled_data.data(), true, block_height,
block_depth);
return unswizzled_data;
}

View File

@@ -10,17 +10,11 @@
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.
*/
std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size_x, u32 tile_size_y,
u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
u32 height, u32 depth,
u32 block_height = TICEntry::DefaultBlockHeight,
u32 block_depth = TICEntry::DefaultBlockHeight);

View File

@@ -161,4 +161,30 @@ 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

View File

@@ -102,27 +102,16 @@ void TelemetryJson::Complete() {
impl->SerializeSection(Telemetry::FieldType::App, "App");
impl->SerializeSection(Telemetry::FieldType::Session, "Session");
impl->SerializeSection(Telemetry::FieldType::Performance, "Performance");
impl->SerializeSection(Telemetry::FieldType::UserFeedback, "UserFeedback");
impl->SerializeSection(Telemetry::FieldType::UserConfig, "UserConfig");
impl->SerializeSection(Telemetry::FieldType::UserSystem, "UserSystem");
auto content = impl->TopSection().dump();
// Send the telemetry async but don't handle the errors since they were written to the log
Common::DetachedTasks::AddTask([host{impl->host}, content]() {
Client{host, "", ""}.PostJson("/telemetry", content, true);
});
}
bool TelemetryJson::SubmitTestcase() {
impl->SerializeSection(Telemetry::FieldType::App, "App");
impl->SerializeSection(Telemetry::FieldType::Session, "Session");
impl->SerializeSection(Telemetry::FieldType::UserFeedback, "UserFeedback");
impl->SerializeSection(Telemetry::FieldType::UserSystem, "UserSystem");
auto content = impl->TopSection().dump();
Client client(impl->host, impl->username, impl->token);
auto value = client.PostJson("/gamedb/testcase", content, false);
return value.result_code == Common::WebResult::Code::Success;
Common::DetachedTasks::AddTask(
[host{impl->host}, username{impl->username}, token{impl->token}, content]() {
Client{host, username, token}.PostJson("/telemetry", content, true);
});
}
} // namespace WebService

View File

@@ -35,7 +35,6 @@ public:
void Visit(const Telemetry::Field<std::chrono::microseconds>& field) override;
void Complete() override;
bool SubmitTestcase() override;
private:
struct Impl;

View File

@@ -56,8 +56,6 @@ add_executable(yuzu
main.h
ui_settings.cpp
ui_settings.h
util/limitable_input_dialog.cpp
util/limitable_input_dialog.h
util/spinbox.cpp
util/spinbox.h
util/util.cpp

View File

@@ -5,7 +5,6 @@
#include <QButtonGroup>
#include <QMessageBox>
#include <QPushButton>
#include <QtConcurrent/qtconcurrentrun.h>
#include "common/logging/log.h"
#include "common/telemetry.h"
#include "core/core.h"
@@ -24,8 +23,6 @@ CompatDB::CompatDB(QWidget* parent)
connect(ui->radioButton_IntroMenu, &QRadioButton::clicked, this, &CompatDB::EnableNext);
connect(ui->radioButton_WontBoot, &QRadioButton::clicked, this, &CompatDB::EnableNext);
connect(button(NextButton), &QPushButton::clicked, this, &CompatDB::Submit);
connect(&testcase_watcher, &QFutureWatcher<bool>::finished, this,
&CompatDB::OnTestcaseSubmitted);
}
CompatDB::~CompatDB() = default;
@@ -51,38 +48,18 @@ void CompatDB::Submit() {
}
break;
case CompatDBPage::Final:
back();
LOG_DEBUG(Frontend, "Compatibility Rating: {}", compatibility->checkedId());
Core::Telemetry().AddField(Telemetry::FieldType::UserFeedback, "Compatibility",
compatibility->checkedId());
button(NextButton)->setEnabled(false);
button(NextButton)->setText(tr("Submitting"));
// older versions of QT don't support the "NoCancelButtonOnLastPage" option, this is a
// workaround
button(QWizard::CancelButton)->setVisible(false);
testcase_watcher.setFuture(QtConcurrent::run(
[this]() { return Core::System::GetInstance().TelemetrySession().SubmitTestcase(); }));
break;
default:
LOG_ERROR(Frontend, "Unexpected page: {}", currentId());
}
}
void CompatDB::OnTestcaseSubmitted() {
if (!testcase_watcher.result()) {
QMessageBox::critical(this, tr("Communication error"),
tr("An error occured while sending the Testcase"));
button(NextButton)->setEnabled(true);
button(NextButton)->setText(tr("Next"));
button(QWizard::CancelButton)->setVisible(true);
} else {
next();
// older versions of QT don't support the "NoCancelButtonOnLastPage" option, this is a
// workaround
button(QWizard::CancelButton)->setVisible(false);
}
}
void CompatDB::EnableNext() {
button(NextButton)->setEnabled(true);
}

View File

@@ -5,7 +5,6 @@
#pragma once
#include <memory>
#include <QFutureWatcher>
#include <QWizard>
namespace Ui {
@@ -20,11 +19,8 @@ public:
~CompatDB();
private:
QFutureWatcher<bool> testcase_watcher;
std::unique_ptr<Ui::CompatDB> ui;
void Submit();
void OnTestcaseSubmitted();
void EnableNext();
};

View File

@@ -322,7 +322,7 @@ void ConfigureInput::setPollingResult(const Common::ParamPackage& params, bool a
}
updateButtonLabels();
input_setter = {};
input_setter = boost::none;
}
void ConfigureInput::keyPressEvent(QKeyEvent* event) {

Some files were not shown because too many files have changed in this diff Show More