diff options
116 files changed, 3692 insertions, 2004 deletions
diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h index 046417da3..71ba4be40 100644 --- a/src/audio_core/audio_renderer.h +++ b/src/audio_core/audio_renderer.h @@ -143,7 +143,7 @@ struct AuxInfo { std::array<u8, 24> output_mix_buffers; u32_le mix_buffer_count; u32_le sample_rate; // Stored in the aux buffer currently - u32_le sampe_count; + u32_le sample_count; u64_le send_buffer_info; u64_le send_buffer_base; diff --git a/src/audio_core/stream.cpp b/src/audio_core/stream.cpp index 742a5e0a0..f35628e45 100644 --- a/src/audio_core/stream.cpp +++ b/src/audio_core/stream.cpp @@ -11,7 +11,6 @@ #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" @@ -104,10 +103,7 @@ 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(); diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp index 6d5218465..5753b871a 100644 --- a/src/common/logging/backend.cpp +++ b/src/common/logging/backend.cpp @@ -12,7 +12,8 @@ #include <thread> #include <vector> #ifdef _WIN32 -#include <share.h> // For _SH_DENYWR +#include <share.h> // For _SH_DENYWR +#include <windows.h> // For OutputDebugStringA #else #define _SH_DENYWR 0 #endif @@ -139,12 +140,18 @@ void FileBackend::Write(const Entry& entry) { if (!file.IsOpen() || bytes_written > MAX_BYTES_WRITTEN) { return; } - bytes_written += file.WriteString(FormatLogMessage(entry) + '\n'); + bytes_written += file.WriteString(FormatLogMessage(entry).append(1, '\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) \ diff --git a/src/common/logging/backend.h b/src/common/logging/backend.h index 11edbf1b6..91bb0c309 100644 --- a/src/common/logging/backend.h +++ b/src/common/logging/backend.h @@ -103,6 +103,20 @@ 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); diff --git a/src/common/string_util.cpp b/src/common/string_util.cpp index 731d1db34..14f7037d8 100644 --- a/src/common/string_util.cpp +++ b/src/common/string_util.cpp @@ -4,11 +4,10 @@ #include <algorithm> #include <cctype> -#include <cerrno> #include <codecvt> -#include <cstdio> #include <cstdlib> -#include <cstring> +#include <locale> +#include <sstream> #include "common/common_paths.h" #include "common/logging/log.h" #include "common/string_util.h" @@ -33,24 +32,6 @@ std::string ToUpper(std::string str) { return str; } -// For Debugging. Read out an u8 array. -std::string ArrayToString(const u8* data, std::size_t size, int line_len, bool spaces) { - std::ostringstream oss; - oss << std::setfill('0') << std::hex; - - for (int line = 0; size; ++data, --size) { - oss << std::setw(2) << (int)*data; - - if (line_len == ++line) { - oss << '\n'; - line = 0; - } else if (spaces) - oss << ' '; - } - - return oss.str(); -} - std::string StringFromBuffer(const std::vector<u8>& data) { return std::string(data.begin(), std::find(data.begin(), data.end(), '\0')); } @@ -75,40 +56,6 @@ std::string StripQuotes(const std::string& s) { return s; } -bool TryParse(const std::string& str, u32* const output) { - char* endptr = nullptr; - - // Reset errno to a value other than ERANGE - errno = 0; - - unsigned long value = strtoul(str.c_str(), &endptr, 0); - - if (!endptr || *endptr) - return false; - - if (errno == ERANGE) - return false; - -#if ULONG_MAX > UINT_MAX - if (value >= 0x100000000ull && value <= 0xFFFFFFFF00000000ull) - return false; -#endif - - *output = static_cast<u32>(value); - return true; -} - -bool TryParse(const std::string& str, bool* const output) { - if ("1" == str || "true" == ToLower(str)) - *output = true; - else if ("0" == str || "false" == ToLower(str)) - *output = false; - else - return false; - - return true; -} - std::string StringFromBool(bool value) { return value ? "True" : "False"; } diff --git a/src/common/string_util.h b/src/common/string_util.h index 32bf6a19c..08f96533b 100644 --- a/src/common/string_util.h +++ b/src/common/string_util.h @@ -5,8 +5,6 @@ #pragma once #include <cstddef> -#include <iomanip> -#include <sstream> #include <string> #include <vector> #include "common/common_types.h" @@ -19,44 +17,13 @@ std::string ToLower(std::string str); /// Make a string uppercase std::string ToUpper(std::string str); -std::string ArrayToString(const u8* data, std::size_t size, int line_len = 20, bool spaces = true); - std::string StringFromBuffer(const std::vector<u8>& data); std::string StripSpaces(const std::string& s); std::string StripQuotes(const std::string& s); -// Thousand separator. Turns 12345678 into 12,345,678 -template <typename I> -std::string ThousandSeparate(I value, int spaces = 0) { - std::ostringstream oss; - -// std::locale("") seems to be broken on many platforms -#if defined _WIN32 || (defined __linux__ && !defined __clang__) - oss.imbue(std::locale("")); -#endif - oss << std::setw(spaces) << value; - - return oss.str(); -} - std::string StringFromBool(bool value); -bool TryParse(const std::string& str, bool* output); -bool TryParse(const std::string& str, u32* output); - -template <typename N> -static bool TryParse(const std::string& str, N* const output) { - std::istringstream iss(str); - - N tmp = 0; - if (iss >> tmp) { - *output = tmp; - return true; - } else - return false; -} - std::string TabsToSpaces(int tab_size, std::string in); void SplitString(const std::string& str, char delim, std::vector<std::string>& output); diff --git a/src/common/telemetry.h b/src/common/telemetry.h index 8d6ab986b..854a73fae 100644 --- a/src/common/telemetry.h +++ b/src/common/telemetry.h @@ -153,6 +153,7 @@ struct VisitorInterface : NonCopyable { /// Completion method, called once all fields have been visited virtual void Complete() = 0; + virtual bool SubmitTestcase() = 0; }; /** @@ -178,6 +179,9 @@ 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, diff --git a/src/core/file_sys/bis_factory.cpp b/src/core/file_sys/bis_factory.cpp index 76a2b7e86..e29f70b3a 100644 --- a/src/core/file_sys/bis_factory.cpp +++ b/src/core/file_sys/bis_factory.cpp @@ -8,8 +8,9 @@ namespace FileSys { -BISFactory::BISFactory(VirtualDir nand_root_, VirtualDir load_root_) +BISFactory::BISFactory(VirtualDir nand_root_, VirtualDir load_root_, VirtualDir dump_root_) : nand_root(std::move(nand_root_)), load_root(std::move(load_root_)), + dump_root(std::move(dump_root_)), sysnand_cache(std::make_unique<RegisteredCache>( GetOrCreateDirectoryRelative(nand_root, "/system/Contents/registered"))), usrnand_cache(std::make_unique<RegisteredCache>( @@ -32,4 +33,10 @@ VirtualDir BISFactory::GetModificationLoadRoot(u64 title_id) const { return GetOrCreateDirectoryRelative(load_root, fmt::format("/{:016X}", title_id)); } +VirtualDir BISFactory::GetModificationDumpRoot(u64 title_id) const { + if (title_id == 0) + return nullptr; + return GetOrCreateDirectoryRelative(dump_root, fmt::format("/{:016X}", title_id)); +} + } // namespace FileSys diff --git a/src/core/file_sys/bis_factory.h b/src/core/file_sys/bis_factory.h index 364d309bd..453c11ad2 100644 --- a/src/core/file_sys/bis_factory.h +++ b/src/core/file_sys/bis_factory.h @@ -17,17 +17,19 @@ class RegisteredCache; /// registered caches. class BISFactory { public: - explicit BISFactory(VirtualDir nand_root, VirtualDir load_root); + explicit BISFactory(VirtualDir nand_root, VirtualDir load_root, VirtualDir dump_root); ~BISFactory(); RegisteredCache* GetSystemNANDContents() const; RegisteredCache* GetUserNANDContents() const; VirtualDir GetModificationLoadRoot(u64 title_id) const; + VirtualDir GetModificationDumpRoot(u64 title_id) const; private: VirtualDir nand_root; VirtualDir load_root; + VirtualDir dump_root; std::unique_ptr<RegisteredCache> sysnand_cache; std::unique_ptr<RegisteredCache> usrnand_cache; diff --git a/src/core/file_sys/control_metadata.cpp b/src/core/file_sys/control_metadata.cpp index a012c2be9..c8fa912bf 100644 --- a/src/core/file_sys/control_metadata.cpp +++ b/src/core/file_sys/control_metadata.cpp @@ -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 diff --git a/src/core/file_sys/control_metadata.h b/src/core/file_sys/control_metadata.h index 141f7e056..bfaad46b4 100644 --- a/src/core/file_sys/control_metadata.h +++ b/src/core/file_sys/control_metadata.h @@ -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; diff --git a/src/core/file_sys/errors.h b/src/core/file_sys/errors.h index fea0593c7..e4a4ee4ab 100644 --- a/src/core/file_sys/errors.h +++ b/src/core/file_sys/errors.h @@ -8,25 +8,10 @@ namespace FileSys { -namespace ErrCodes { -enum { - NotFound = 1, - TitleNotFound = 1002, - SdCardNotFound = 2001, - RomFSNotFound = 2520, -}; -} - -constexpr ResultCode ERROR_PATH_NOT_FOUND(ErrorModule::FS, ErrCodes::NotFound); - -// TODO(bunnei): Replace these with correct errors for Switch OS -constexpr ResultCode ERROR_INVALID_PATH(-1); -constexpr ResultCode ERROR_UNSUPPORTED_OPEN_FLAGS(-1); -constexpr ResultCode ERROR_INVALID_OPEN_FLAGS(-1); -constexpr ResultCode ERROR_FILE_NOT_FOUND(-1); -constexpr ResultCode ERROR_UNEXPECTED_FILE_OR_DIRECTORY(-1); -constexpr ResultCode ERROR_DIRECTORY_ALREADY_EXISTS(-1); -constexpr ResultCode ERROR_FILE_ALREADY_EXISTS(-1); -constexpr ResultCode ERROR_DIRECTORY_NOT_EMPTY(-1); +constexpr ResultCode ERROR_PATH_NOT_FOUND{ErrorModule::FS, 1}; +constexpr ResultCode ERROR_ENTITY_NOT_FOUND{ErrorModule::FS, 1002}; +constexpr ResultCode ERROR_SD_CARD_NOT_FOUND{ErrorModule::FS, 2001}; +constexpr ResultCode ERROR_INVALID_OFFSET{ErrorModule::FS, 6061}; +constexpr ResultCode ERROR_INVALID_SIZE{ErrorModule::FS, 6062}; } // namespace FileSys diff --git a/src/core/file_sys/patch_manager.cpp b/src/core/file_sys/patch_manager.cpp index 0c1156989..8d062eb3e 100644 --- a/src/core/file_sys/patch_manager.cpp +++ b/src/core/file_sys/patch_manager.cpp @@ -19,6 +19,7 @@ #include "core/file_sys/vfs_vector.h" #include "core/hle/service/filesystem/filesystem.h" #include "core/loader/loader.h" +#include "core/settings.h" namespace FileSys { @@ -119,6 +120,18 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso) const { const auto build_id_raw = Common::HexArrayToString(header.build_id); const auto build_id = build_id_raw.substr(0, build_id_raw.find_last_not_of('0') + 1); + if (Settings::values.dump_nso) { + LOG_INFO(Loader, "Dumping NSO for build_id={}, title_id={:016X}", build_id, title_id); + const auto dump_dir = Service::FileSystem::GetModificationDumpRoot(title_id); + if (dump_dir != nullptr) { + const auto nso_dir = GetOrCreateDirectoryRelative(dump_dir, "/nso"); + const auto file = nso_dir->CreateFile(fmt::format("{}.nso", build_id)); + + file->Resize(nso.size()); + file->WriteBytes(nso); + } + } + LOG_INFO(Loader, "Patching NSO for build_id={}", build_id); const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id); diff --git a/src/core/file_sys/savedata_factory.cpp b/src/core/file_sys/savedata_factory.cpp index ef1aaebbb..5434f2149 100644 --- a/src/core/file_sys/savedata_factory.cpp +++ b/src/core/file_sys/savedata_factory.cpp @@ -83,28 +83,32 @@ ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, SaveDataDescr return MakeResult<VirtualDir>(std::move(out)); } -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 - // be interpreted as the title id of the current process. - if (type == SaveDataType::SaveData && title_id == 0) - title_id = Core::CurrentProcess()->GetTitleID(); - - std::string out; +VirtualDir SaveDataFactory::GetSaveDataSpaceDirectory(SaveDataSpaceId space) const { + return dir->GetDirectoryRelative(GetSaveDataSpaceIdPath(space)); +} +std::string SaveDataFactory::GetSaveDataSpaceIdPath(SaveDataSpaceId space) { switch (space) { case SaveDataSpaceId::NandSystem: - out = "/system/"; - break; + return "/system/"; case SaveDataSpaceId::NandUser: - out = "/user/"; - break; + return "/user/"; case SaveDataSpaceId::TemporaryStorage: - out = "/temp/"; - break; + 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 + // be interpreted as the title id of the current process. + if (type == SaveDataType::SaveData && title_id == 0) + title_id = Core::CurrentProcess()->GetTitleID(); + + std::string out = GetSaveDataSpaceIdPath(space); switch (type) { case SaveDataType::SystemSaveData: diff --git a/src/core/file_sys/savedata_factory.h b/src/core/file_sys/savedata_factory.h index d69ef6741..2a0088040 100644 --- a/src/core/file_sys/savedata_factory.h +++ b/src/core/file_sys/savedata_factory.h @@ -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); diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp index 420218d59..f06b6bb55 100644 --- a/src/core/hle/kernel/process.cpp +++ b/src/core/hle/kernel/process.cpp @@ -5,11 +5,9 @@ #include <algorithm> #include <memory> #include "common/assert.h" -#include "common/common_funcs.h" #include "common/logging/log.h" #include "core/core.h" #include "core/file_sys/program_metadata.h" -#include "core/hle/kernel/errors.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/resource_limit.h" @@ -17,6 +15,7 @@ #include "core/hle/kernel/thread.h" #include "core/hle/kernel/vm_manager.h" #include "core/memory.h" +#include "core/settings.h" namespace Kernel { @@ -35,6 +34,11 @@ SharedPtr<Process> Process::Create(KernelCore& kernel, std::string&& name) { process->process_id = kernel.CreateNewProcessID(); process->svc_access_mask.set(); + std::mt19937 rng(Settings::values.rng_seed.value_or(0)); + std::uniform_int_distribution<u64> distribution; + std::generate(process->random_entropy.begin(), process->random_entropy.end(), + [&] { return distribution(rng); }); + kernel.AppendNewProcess(process); return process; } @@ -241,83 +245,15 @@ void Process::LoadModule(CodeSet module_, VAddr base_addr) { } ResultVal<VAddr> Process::HeapAllocate(VAddr target, u64 size, VMAPermission perms) { - if (target < vm_manager.GetHeapRegionBaseAddress() || - target + size > vm_manager.GetHeapRegionEndAddress() || target + size < target) { - return ERR_INVALID_ADDRESS; - } - - if (heap_memory == nullptr) { - // Initialize heap - heap_memory = std::make_shared<std::vector<u8>>(); - heap_start = heap_end = target; - } else { - vm_manager.UnmapRange(heap_start, heap_end - heap_start); - } - - // If necessary, expand backing vector to cover new heap extents. - if (target < heap_start) { - heap_memory->insert(begin(*heap_memory), heap_start - target, 0); - heap_start = target; - vm_manager.RefreshMemoryBlockMappings(heap_memory.get()); - } - if (target + size > heap_end) { - heap_memory->insert(end(*heap_memory), (target + size) - heap_end, 0); - heap_end = target + size; - vm_manager.RefreshMemoryBlockMappings(heap_memory.get()); - } - ASSERT(heap_end - heap_start == heap_memory->size()); - - CASCADE_RESULT(auto vma, vm_manager.MapMemoryBlock(target, heap_memory, target - heap_start, - size, MemoryState::Heap)); - vm_manager.Reprotect(vma, perms); - - heap_used = size; - - return MakeResult<VAddr>(heap_end - size); + return vm_manager.HeapAllocate(target, size, perms); } ResultCode Process::HeapFree(VAddr target, u32 size) { - if (target < vm_manager.GetHeapRegionBaseAddress() || - target + size > vm_manager.GetHeapRegionEndAddress() || target + size < target) { - return ERR_INVALID_ADDRESS; - } - - if (size == 0) { - return RESULT_SUCCESS; - } - - ResultCode result = vm_manager.UnmapRange(target, size); - if (result.IsError()) - return result; - - heap_used -= size; - - return RESULT_SUCCESS; + return vm_manager.HeapFree(target, size); } ResultCode Process::MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size) { - auto vma = vm_manager.FindVMA(src_addr); - - ASSERT_MSG(vma != vm_manager.vma_map.end(), "Invalid memory address"); - ASSERT_MSG(vma->second.backing_block, "Backing block doesn't exist for address"); - - // The returned VMA might be a bigger one encompassing the desired address. - auto vma_offset = src_addr - vma->first; - ASSERT_MSG(vma_offset + size <= vma->second.size, - "Shared memory exceeds bounds of mapped block"); - - const std::shared_ptr<std::vector<u8>>& backing_block = vma->second.backing_block; - std::size_t backing_block_offset = vma->second.offset + vma_offset; - - CASCADE_RESULT(auto new_vma, - vm_manager.MapMemoryBlock(dst_addr, backing_block, backing_block_offset, size, - MemoryState::Mapped)); - // Protect mirror with permissions from old region - vm_manager.Reprotect(new_vma, vma->second.permissions); - // Remove permissions from old region - vm_manager.Reprotect(vma, VMAPermission::None); - - return RESULT_SUCCESS; + return vm_manager.MirrorMemory(dst_addr, src_addr, size); } ResultCode Process::UnmapMemory(VAddr dst_addr, VAddr /*src_addr*/, u64 size) { diff --git a/src/core/hle/kernel/process.h b/src/core/hle/kernel/process.h index 8d2616c79..cf48787ce 100644 --- a/src/core/hle/kernel/process.h +++ b/src/core/hle/kernel/process.h @@ -8,6 +8,7 @@ #include <bitset> #include <cstddef> #include <memory> +#include <random> #include <string> #include <vector> #include <boost/container/static_vector.hpp> @@ -119,6 +120,8 @@ struct CodeSet final { class Process final : public Object { public: + static constexpr std::size_t RANDOM_ENTROPY_SIZE = 4; + static SharedPtr<Process> Create(KernelCore& kernel, std::string&& name); std::string GetTypeName() const override { @@ -212,6 +215,11 @@ public: total_process_running_time_ticks += ticks; } + /// Gets 8 bytes of random data for svcGetInfo RandomEntropy + u64 GetRandomEntropy(std::size_t index) const { + return random_entropy.at(index); + } + /** * Loads process-specifics configuration info with metadata provided * by an executable. @@ -292,17 +300,6 @@ private: u32 allowed_thread_priority_mask = 0xFFFFFFFF; u32 is_virtual_address_memory_enabled = 0; - // Memory used to back the allocations in the regular heap. A single vector is used to cover - // the entire virtual address space extents that bound the allocations, including any holes. - // This makes deallocation and reallocation of holes fast and keeps process memory contiguous - // in the emulator address space, allowing Memory::GetPointer to be reasonably safe. - std::shared_ptr<std::vector<u8>> heap_memory; - - // The left/right bounds of the address space covered by heap_memory. - VAddr heap_start = 0; - VAddr heap_end = 0; - u64 heap_used = 0; - /// The Thread Local Storage area is allocated as processes create threads, /// each TLS area is 0x200 bytes, so one page (0x1000) is split up in 8 parts, and each part /// holds the TLS for a specific thread. This vector contains which parts are in use for each @@ -321,6 +318,9 @@ private: /// Per-process handle table for storing created object handles in. HandleTable handle_table; + /// Random values for svcGetInfo RandomEntropy + std::array<u64, RANDOM_ENTROPY_SIZE> random_entropy; + std::string name; }; diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index c7c579aaf..b0b6508d9 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -34,6 +34,7 @@ #include "core/hle/lock.h" #include "core/hle/result.h" #include "core/hle/service/service.h" +#include "core/settings.h" namespace Kernel { namespace { @@ -395,16 +396,42 @@ struct BreakReason { /// Break program execution static void Break(u32 reason, u64 info1, u64 info2) { BreakReason break_reason{reason}; + bool has_dumped_buffer{}; + const auto handle_debug_buffer = [&](VAddr addr, u64 sz) { + if (sz == 0 || addr == 0 || has_dumped_buffer) { + return; + } + + // This typically is an error code so we're going to assume this is the case + if (sz == sizeof(u32)) { + LOG_CRITICAL(Debug_Emulated, "debug_buffer_err_code={:X}", Memory::Read32(addr)); + } else { + // We don't know what's in here so we'll hexdump it + std::vector<u8> debug_buffer(sz); + Memory::ReadBlock(addr, debug_buffer.data(), sz); + std::string hexdump; + for (std::size_t i = 0; i < debug_buffer.size(); i++) { + hexdump += fmt::format("{:02X} ", debug_buffer[i]); + if (i != 0 && i % 16 == 0) { + hexdump += '\n'; + } + } + LOG_CRITICAL(Debug_Emulated, "debug_buffer=\n{}", hexdump); + } + has_dumped_buffer = true; + }; switch (break_reason.break_type) { case BreakType::Panic: LOG_CRITICAL(Debug_Emulated, "Signalling debugger, PANIC! info1=0x{:016X}, info2=0x{:016X}", info1, info2); + handle_debug_buffer(info1, info2); break; case BreakType::AssertionFailed: LOG_CRITICAL(Debug_Emulated, "Signalling debugger, Assertion failed! info1=0x{:016X}, info2=0x{:016X}", info1, info2); + handle_debug_buffer(info1, info2); break; case BreakType::PreNROLoad: LOG_WARNING( @@ -433,6 +460,7 @@ static void Break(u32 reason, u64 info1, u64 info2) { Debug_Emulated, "Signalling debugger, Unknown break reason {}, info1=0x{:016X}, info2=0x{:016X}", static_cast<u32>(break_reason.break_type.Value()), info1, info2); + handle_debug_buffer(info1, info2); break; } @@ -441,6 +469,7 @@ static void Break(u32 reason, u64 info1, u64 info2) { Debug_Emulated, "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}", reason, info1, info2); + handle_debug_buffer(info1, info2); ASSERT(false); Core::CurrentProcess()->PrepareForTermination(); @@ -530,7 +559,16 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) *result = 0; break; case GetInfoType::RandomEntropy: - *result = 0; + if (handle != 0) { + return ERR_INVALID_HANDLE; + } + + if (info_sub_id >= Process::RANDOM_ENTROPY_SIZE) { + return ERR_INVALID_COMBINATION_KERNEL; + } + + *result = current_process->GetRandomEntropy(info_sub_id); + return RESULT_SUCCESS; break; case GetInfoType::ASLRRegionBaseAddr: *result = vm_manager.GetASLRRegionBaseAddress(); diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp index dd5cd9ced..4ffb76818 100644 --- a/src/core/hle/kernel/thread.cpp +++ b/src/core/hle/kernel/thread.cpp @@ -142,36 +142,7 @@ void Thread::ResumeFromWait() { status = ThreadStatus::Ready; - std::optional<s32> new_processor_id = GetNextProcessorId(affinity_mask); - if (!new_processor_id) { - new_processor_id = processor_id; - } - if (ideal_core != -1 && - Core::System::GetInstance().Scheduler(ideal_core).GetCurrentThread() == nullptr) { - new_processor_id = ideal_core; - } - - ASSERT(*new_processor_id < 4); - - // Add thread to new core's scheduler - auto* next_scheduler = &Core::System::GetInstance().Scheduler(*new_processor_id); - - if (*new_processor_id != processor_id) { - // Remove thread from previous core's scheduler - scheduler->RemoveThread(this); - next_scheduler->AddThread(this, current_priority); - } - - processor_id = *new_processor_id; - - // If the thread was ready, unschedule from the previous core and schedule on the new core - scheduler->UnscheduleThread(this, current_priority); - next_scheduler->ScheduleThread(this, current_priority); - - // Change thread's scheduler - scheduler = next_scheduler; - - Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule(); + ChangeScheduler(); } /** @@ -364,42 +335,45 @@ void Thread::UpdatePriority() { void Thread::ChangeCore(u32 core, u64 mask) { ideal_core = core; affinity_mask = mask; + ChangeScheduler(); +} +void Thread::ChangeScheduler() { if (status != ThreadStatus::Ready) { return; } + auto& system = Core::System::GetInstance(); std::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)}; if (!new_processor_id) { new_processor_id = processor_id; } - if (ideal_core != -1 && - Core::System::GetInstance().Scheduler(ideal_core).GetCurrentThread() == nullptr) { + if (ideal_core != -1 && system.Scheduler(ideal_core).GetCurrentThread() == nullptr) { new_processor_id = ideal_core; } ASSERT(*new_processor_id < 4); // Add thread to new core's scheduler - auto* next_scheduler = &Core::System::GetInstance().Scheduler(*new_processor_id); + auto& next_scheduler = system.Scheduler(*new_processor_id); if (*new_processor_id != processor_id) { // Remove thread from previous core's scheduler scheduler->RemoveThread(this); - next_scheduler->AddThread(this, current_priority); + next_scheduler.AddThread(this, current_priority); } processor_id = *new_processor_id; // If the thread was ready, unschedule from the previous core and schedule on the new core scheduler->UnscheduleThread(this, current_priority); - next_scheduler->ScheduleThread(this, current_priority); + next_scheduler.ScheduleThread(this, current_priority); // Change thread's scheduler - scheduler = next_scheduler; + scheduler = &next_scheduler; - Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule(); + system.CpuCore(processor_id).PrepareReschedule(); } bool Thread::AllWaitObjectsReady() { diff --git a/src/core/hle/kernel/thread.h b/src/core/hle/kernel/thread.h index 4a6e11239..d384d50db 100644 --- a/src/core/hle/kernel/thread.h +++ b/src/core/hle/kernel/thread.h @@ -374,6 +374,8 @@ private: explicit Thread(KernelCore& kernel); ~Thread() override; + void ChangeScheduler(); + Core::ARM_Interface::ThreadContext context{}; u32 thread_id = 0; diff --git a/src/core/hle/kernel/vm_manager.cpp b/src/core/hle/kernel/vm_manager.cpp index 1a92c8f70..ec7fd6150 100644 --- a/src/core/hle/kernel/vm_manager.cpp +++ b/src/core/hle/kernel/vm_manager.cpp @@ -243,6 +243,85 @@ ResultCode VMManager::ReprotectRange(VAddr target, u64 size, VMAPermission new_p return RESULT_SUCCESS; } +ResultVal<VAddr> VMManager::HeapAllocate(VAddr target, u64 size, VMAPermission perms) { + if (target < GetHeapRegionBaseAddress() || target + size > GetHeapRegionEndAddress() || + target + size < target) { + return ERR_INVALID_ADDRESS; + } + + if (heap_memory == nullptr) { + // Initialize heap + heap_memory = std::make_shared<std::vector<u8>>(); + heap_start = heap_end = target; + } else { + UnmapRange(heap_start, heap_end - heap_start); + } + + // If necessary, expand backing vector to cover new heap extents. + if (target < heap_start) { + heap_memory->insert(begin(*heap_memory), heap_start - target, 0); + heap_start = target; + RefreshMemoryBlockMappings(heap_memory.get()); + } + if (target + size > heap_end) { + heap_memory->insert(end(*heap_memory), (target + size) - heap_end, 0); + heap_end = target + size; + RefreshMemoryBlockMappings(heap_memory.get()); + } + ASSERT(heap_end - heap_start == heap_memory->size()); + + CASCADE_RESULT(auto vma, MapMemoryBlock(target, heap_memory, target - heap_start, size, + MemoryState::Heap)); + Reprotect(vma, perms); + + heap_used = size; + + return MakeResult<VAddr>(heap_end - size); +} + +ResultCode VMManager::HeapFree(VAddr target, u64 size) { + if (target < GetHeapRegionBaseAddress() || target + size > GetHeapRegionEndAddress() || + target + size < target) { + return ERR_INVALID_ADDRESS; + } + + if (size == 0) { + return RESULT_SUCCESS; + } + + const ResultCode result = UnmapRange(target, size); + if (result.IsError()) { + return result; + } + + heap_used -= size; + return RESULT_SUCCESS; +} + +ResultCode VMManager::MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size) { + const auto vma = FindVMA(src_addr); + + ASSERT_MSG(vma != vma_map.end(), "Invalid memory address"); + ASSERT_MSG(vma->second.backing_block, "Backing block doesn't exist for address"); + + // The returned VMA might be a bigger one encompassing the desired address. + const auto vma_offset = src_addr - vma->first; + ASSERT_MSG(vma_offset + size <= vma->second.size, + "Shared memory exceeds bounds of mapped block"); + + const std::shared_ptr<std::vector<u8>>& backing_block = vma->second.backing_block; + const std::size_t backing_block_offset = vma->second.offset + vma_offset; + + CASCADE_RESULT(auto new_vma, MapMemoryBlock(dst_addr, backing_block, backing_block_offset, size, + MemoryState::Mapped)); + // Protect mirror with permissions from old region + Reprotect(new_vma, vma->second.permissions); + // Remove permissions from old region + Reprotect(vma, VMAPermission::None); + + return RESULT_SUCCESS; +} + void VMManager::RefreshMemoryBlockMappings(const std::vector<u8>* block) { // If this ever proves to have a noticeable performance impact, allow users of the function to // specify a specific range of addresses to limit the scan to. @@ -495,8 +574,7 @@ u64 VMManager::GetTotalMemoryUsage() const { } u64 VMManager::GetTotalHeapUsage() const { - LOG_WARNING(Kernel, "(STUBBED) called"); - return 0x0; + return heap_used; } VAddr VMManager::GetAddressSpaceBaseAddress() const { diff --git a/src/core/hle/kernel/vm_manager.h b/src/core/hle/kernel/vm_manager.h index 2447cbb8f..248cc46dc 100644 --- a/src/core/hle/kernel/vm_manager.h +++ b/src/core/hle/kernel/vm_manager.h @@ -186,6 +186,11 @@ public: /// Changes the permissions of a range of addresses, splitting VMAs as necessary. ResultCode ReprotectRange(VAddr target, u64 size, VMAPermission new_perms); + ResultVal<VAddr> HeapAllocate(VAddr target, u64 size, VMAPermission perms); + ResultCode HeapFree(VAddr target, u64 size); + + ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size); + /** * Scans all VMAs and updates the page table range of any that use the given vector as backing * memory. This should be called after any operation that causes reallocation of the vector. @@ -343,5 +348,15 @@ private: VAddr tls_io_region_base = 0; VAddr tls_io_region_end = 0; + + // Memory used to back the allocations in the regular heap. A single vector is used to cover + // the entire virtual address space extents that bound the allocations, including any holes. + // This makes deallocation and reallocation of holes fast and keeps process memory contiguous + // in the emulator address space, allowing Memory::GetPointer to be reasonably safe. + std::shared_ptr<std::vector<u8>> heap_memory; + // The left/right bounds of the address space covered by heap_memory. + VAddr heap_start = 0; + VAddr heap_end = 0; + u64 heap_used = 0; }; } // namespace Kernel diff --git a/src/core/hle/result.h b/src/core/hle/result.h index c6b18cfba..bfb77cc31 100644 --- a/src/core/hle/result.h +++ b/src/core/hle/result.h @@ -19,8 +19,6 @@ enum class ErrorDescription : u32 { Success = 0, RemoteProcessDead = 301, - InvalidOffset = 6061, - InvalidLength = 6062, }; /** diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp index c6437a671..c629f9357 100644 --- a/src/core/hle/service/acc/acc.cpp +++ b/src/core/hle/service/acc/acc.cpp @@ -242,6 +242,30 @@ void Module::Interface::GetBaasAccountManagerForApplication(Kernel::HLERequestCo LOG_DEBUG(Service_ACC, "called"); } +void Module::Interface::TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_ACC, "called"); + // A u8 is passed into this function which we can safely ignore. It's to determine if we have + // access to use the network or not by the looks of it + IPC::ResponseBuilder rb{ctx, 6}; + if (profile_manager->GetUserCount() != 1) { + rb.Push(RESULT_SUCCESS); + rb.PushRaw<u128>(INVALID_UUID); + return; + } + + const auto user_list = profile_manager->GetAllUsers(); + if (std::all_of(user_list.begin(), user_list.end(), + [](const auto& user) { return user.uuid == INVALID_UUID; })) { + rb.Push(ResultCode(-1)); // TODO(ogniK): Find the correct error code + rb.PushRaw<u128>(INVALID_UUID); + return; + } + + // Select the first user we have + rb.Push(RESULT_SUCCESS); + rb.PushRaw<u128>(profile_manager->GetUser(0)->uuid); +} + Module::Interface::Interface(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> profile_manager, const char* name) : ServiceFramework(name), module(std::move(module)), diff --git a/src/core/hle/service/acc/acc.h b/src/core/hle/service/acc/acc.h index c7ed74351..89b2104fa 100644 --- a/src/core/hle/service/acc/acc.h +++ b/src/core/hle/service/acc/acc.h @@ -27,6 +27,7 @@ public: void InitializeApplicationInfo(Kernel::HLERequestContext& ctx); void GetBaasAccountManagerForApplication(Kernel::HLERequestContext& ctx); void IsUserRegistrationRequestPermitted(Kernel::HLERequestContext& ctx); + void TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx); protected: std::shared_ptr<Module> module; diff --git a/src/core/hle/service/acc/acc_su.cpp b/src/core/hle/service/acc/acc_su.cpp index ad455c3a7..5e2030355 100644 --- a/src/core/hle/service/acc/acc_su.cpp +++ b/src/core/hle/service/acc/acc_su.cpp @@ -17,7 +17,7 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {5, &ACC_SU::GetProfile, "GetProfile"}, {6, nullptr, "GetProfileDigest"}, {50, &ACC_SU::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, - {51, nullptr, "TrySelectUserWithoutInteraction"}, + {51, &ACC_SU::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, {60, nullptr, "ListOpenContextStoredUsers"}, {100, nullptr, "GetUserRegistrationNotifier"}, {101, nullptr, "GetUserStateChangeNotifier"}, diff --git a/src/core/hle/service/acc/acc_u0.cpp b/src/core/hle/service/acc/acc_u0.cpp index 72d4adf35..a4d705b45 100644 --- a/src/core/hle/service/acc/acc_u0.cpp +++ b/src/core/hle/service/acc/acc_u0.cpp @@ -17,7 +17,7 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {5, &ACC_U0::GetProfile, "GetProfile"}, {6, nullptr, "GetProfileDigest"}, {50, &ACC_U0::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, - {51, nullptr, "TrySelectUserWithoutInteraction"}, + {51, &ACC_U0::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, {60, nullptr, "ListOpenContextStoredUsers"}, {100, &ACC_U0::InitializeApplicationInfo, "InitializeApplicationInfo"}, {101, &ACC_U0::GetBaasAccountManagerForApplication, "GetBaasAccountManagerForApplication"}, diff --git a/src/core/hle/service/acc/acc_u1.cpp b/src/core/hle/service/acc/acc_u1.cpp index d480f08e5..8fffc93b5 100644 --- a/src/core/hle/service/acc/acc_u1.cpp +++ b/src/core/hle/service/acc/acc_u1.cpp @@ -17,7 +17,7 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {5, &ACC_U1::GetProfile, "GetProfile"}, {6, nullptr, "GetProfileDigest"}, {50, &ACC_U1::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, - {51, nullptr, "TrySelectUserWithoutInteraction"}, + {51, &ACC_U1::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, {60, nullptr, "ListOpenContextStoredUsers"}, {100, nullptr, "GetUserRegistrationNotifier"}, {101, nullptr, "GetUserStateChangeNotifier"}, diff --git a/src/core/hle/service/acc/profile_manager.cpp b/src/core/hle/service/acc/profile_manager.cpp index c08394e4c..968263846 100644 --- a/src/core/hle/service/acc/profile_manager.cpp +++ b/src/core/hle/service/acc/profile_manager.cpp @@ -2,8 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <cstring> #include <random> +#include <fmt/format.h> + #include "common/file_util.h" #include "core/hle/service/acc/profile_manager.h" #include "core/settings.h" @@ -39,6 +42,19 @@ UUID UUID::Generate() { return UUID{distribution(gen), distribution(gen)}; } +std::string UUID::Format() const { + return fmt::format("0x{:016X}{:016X}", uuid[1], uuid[0]); +} + +std::string UUID::FormatSwitch() const { + std::array<u8, 16> s{}; + std::memcpy(s.data(), uuid.data(), sizeof(u128)); + return fmt::format("{:02x}{:02x}{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{" + ":02x}{:02x}{:02x}{:02x}{:02x}", + s[0], s[1], s[2], s[3], s[4], s[5], s[6], s[7], s[8], s[9], s[10], s[11], + s[12], s[13], s[14], s[15]); +} + ProfileManager::ProfileManager() { ParseUserSaveFile(); @@ -325,11 +341,12 @@ void ProfileManager::ParseUserSaveFile() { return; } - for (std::size_t i = 0; i < MAX_USERS; ++i) { - const auto& user = data.users[i]; + for (const auto& user : data.users) { + if (user.uuid == UUID(INVALID_UUID)) { + continue; + } - if (user.uuid != UUID(INVALID_UUID)) - AddUser({user.uuid, user.username, user.timestamp, {}, false}); + AddUser({user.uuid, user.username, user.timestamp, {}, false}); } std::stable_partition(profiles.begin(), profiles.end(), diff --git a/src/core/hle/service/acc/profile_manager.h b/src/core/hle/service/acc/profile_manager.h index 1cd2e51b2..d2d8e6c6b 100644 --- a/src/core/hle/service/acc/profile_manager.h +++ b/src/core/hle/service/acc/profile_manager.h @@ -42,22 +42,14 @@ struct UUID { void Invalidate() { uuid = INVALID_UUID; } - std::string Format() const { - return fmt::format("0x{:016X}{:016X}", uuid[1], uuid[0]); - } - std::string FormatSwitch() const { - std::array<u8, 16> s{}; - std::memcpy(s.data(), uuid.data(), sizeof(u128)); - return fmt::format("{:02x}{:02x}{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{" - ":02x}{:02x}{:02x}{:02x}{:02x}", - s[0], s[1], s[2], s[3], s[4], s[5], s[6], s[7], s[8], s[9], s[10], s[11], - s[12], s[13], s[14], s[15]); - } + std::string Format() const; + std::string FormatSwitch() const; }; static_assert(sizeof(UUID) == 16, "UUID is an invalid size!"); -using ProfileUsername = std::array<u8, 0x20>; +constexpr std::size_t profile_username_size = 32; +using ProfileUsername = std::array<u8, profile_username_size>; using ProfileData = std::array<u8, MAX_DATA>; using UserIDArray = std::array<UUID, MAX_USERS>; diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp index ac3ff9f20..0477ce66e 100644 --- a/src/core/hle/service/am/am.cpp +++ b/src/core/hle/service/am/am.cpp @@ -338,7 +338,54 @@ void ISelfController::GetIdleTimeDetectionExtension(Kernel::HLERequestContext& c LOG_WARNING(Service_AM, "(STUBBED) called"); } -ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter") { +AppletMessageQueue::AppletMessageQueue() { + auto& kernel = Core::System::GetInstance().Kernel(); + on_new_message = Kernel::Event::Create(kernel, Kernel::ResetType::Sticky, + "AMMessageQueue:OnMessageRecieved"); + on_operation_mode_changed = Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, + "AMMessageQueue:OperationModeChanged"); +} + +AppletMessageQueue::~AppletMessageQueue() = default; + +const Kernel::SharedPtr<Kernel::Event>& AppletMessageQueue::GetMesssageRecieveEvent() const { + return on_new_message; +} + +const Kernel::SharedPtr<Kernel::Event>& AppletMessageQueue::GetOperationModeChangedEvent() const { + return on_operation_mode_changed; +} + +void AppletMessageQueue::PushMessage(AppletMessage msg) { + messages.push(msg); + on_new_message->Signal(); +} + +AppletMessageQueue::AppletMessage AppletMessageQueue::PopMessage() { + if (messages.empty()) { + on_new_message->Clear(); + return AppletMessage::NoMessage; + } + auto msg = messages.front(); + messages.pop(); + if (messages.empty()) { + on_new_message->Clear(); + } + return msg; +} + +std::size_t AppletMessageQueue::GetMessageCount() const { + return messages.size(); +} + +void AppletMessageQueue::OperationModeChanged() { + PushMessage(AppletMessage::OperationModeChanged); + PushMessage(AppletMessage::PerformanceModeChanged); + on_operation_mode_changed->Signal(); +} + +ICommonStateGetter::ICommonStateGetter(std::shared_ptr<AppletMessageQueue> msg_queue) + : ServiceFramework("ICommonStateGetter"), msg_queue(std::move(msg_queue)) { // clang-format off static const FunctionInfo functions[] = { {0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"}, @@ -388,21 +435,19 @@ void ICommonStateGetter::GetBootMode(Kernel::HLERequestContext& ctx) { } void ICommonStateGetter::GetEventHandle(Kernel::HLERequestContext& ctx) { - event->Signal(); - IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); - rb.PushCopyObjects(event); + rb.PushCopyObjects(msg_queue->GetMesssageRecieveEvent()); - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); } void ICommonStateGetter::ReceiveMessage(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(15); + rb.PushEnum<AppletMessageQueue::AppletMessage>(msg_queue->PopMessage()); - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); } void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) { @@ -414,13 +459,11 @@ void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) { } void ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent(Kernel::HLERequestContext& ctx) { - event->Signal(); - IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); - rb.PushCopyObjects(event); + rb.PushCopyObjects(msg_queue->GetOperationModeChangedEvent()); - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); } void ICommonStateGetter::GetDefaultDisplayResolution(Kernel::HLERequestContext& ctx) { @@ -444,7 +487,7 @@ void ICommonStateGetter::GetOperationMode(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); rb.Push(static_cast<u8>(use_docked_mode ? OperationMode::Docked : OperationMode::Handheld)); - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); } void ICommonStateGetter::GetPerformanceMode(Kernel::HLERequestContext& ctx) { @@ -454,7 +497,7 @@ void ICommonStateGetter::GetPerformanceMode(Kernel::HLERequestContext& ctx) { rb.Push(static_cast<u32>(use_docked_mode ? APM::PerformanceMode::Docked : APM::PerformanceMode::Handheld)); - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); } class IStorageAccessor final : public ServiceFramework<IStorageAccessor> { @@ -840,8 +883,12 @@ void IApplicationFunctions::GetPseudoDeviceId(Kernel::HLERequestContext& ctx) { void InstallInterfaces(SM::ServiceManager& service_manager, std::shared_ptr<NVFlinger::NVFlinger> nvflinger) { - std::make_shared<AppletAE>(nvflinger)->InstallAsService(service_manager); - std::make_shared<AppletOE>(nvflinger)->InstallAsService(service_manager); + auto message_queue = std::make_shared<AppletMessageQueue>(); + message_queue->PushMessage( + AppletMessageQueue::AppletMessage::FocusStateChanged); // Needed on game boot + + std::make_shared<AppletAE>(nvflinger, message_queue)->InstallAsService(service_manager); + std::make_shared<AppletOE>(nvflinger, message_queue)->InstallAsService(service_manager); std::make_shared<IdleSys>()->InstallAsService(service_manager); std::make_shared<OMM>()->InstallAsService(service_manager); std::make_shared<SPSM>()->InstallAsService(service_manager); diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h index 095f94851..2f1c20bce 100644 --- a/src/core/hle/service/am/am.h +++ b/src/core/hle/service/am/am.h @@ -5,6 +5,7 @@ #pragma once #include <memory> +#include <queue> #include "core/hle/service/service.h" namespace Kernel { @@ -39,6 +40,31 @@ enum SystemLanguage { TraditionalChinese = 16, }; +class AppletMessageQueue { +public: + enum class AppletMessage : u32 { + NoMessage = 0, + FocusStateChanged = 15, + OperationModeChanged = 30, + PerformanceModeChanged = 31, + }; + + AppletMessageQueue(); + ~AppletMessageQueue(); + + const Kernel::SharedPtr<Kernel::Event>& GetMesssageRecieveEvent() const; + const Kernel::SharedPtr<Kernel::Event>& GetOperationModeChangedEvent() const; + void PushMessage(AppletMessage msg); + AppletMessage PopMessage(); + std::size_t GetMessageCount() const; + void OperationModeChanged(); + +private: + std::queue<AppletMessage> messages; + Kernel::SharedPtr<Kernel::Event> on_new_message; + Kernel::SharedPtr<Kernel::Event> on_operation_mode_changed; +}; + class IWindowController final : public ServiceFramework<IWindowController> { public: IWindowController(); @@ -102,7 +128,7 @@ private: class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> { public: - ICommonStateGetter(); + explicit ICommonStateGetter(std::shared_ptr<AppletMessageQueue> msg_queue); ~ICommonStateGetter() override; private: @@ -126,6 +152,7 @@ private: void GetDefaultDisplayResolution(Kernel::HLERequestContext& ctx); Kernel::SharedPtr<Kernel::Event> event; + std::shared_ptr<AppletMessageQueue> msg_queue; }; class ILibraryAppletCreator final : public ServiceFramework<ILibraryAppletCreator> { diff --git a/src/core/hle/service/am/applet_ae.cpp b/src/core/hle/service/am/applet_ae.cpp index 68ea778e8..ec93e3529 100644 --- a/src/core/hle/service/am/applet_ae.cpp +++ b/src/core/hle/service/am/applet_ae.cpp @@ -12,8 +12,10 @@ namespace Service::AM { class ILibraryAppletProxy final : public ServiceFramework<ILibraryAppletProxy> { public: - explicit ILibraryAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger) - : ServiceFramework("ILibraryAppletProxy"), nvflinger(std::move(nvflinger)) { + explicit ILibraryAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, + std::shared_ptr<AppletMessageQueue> msg_queue) + : ServiceFramework("ILibraryAppletProxy"), nvflinger(std::move(nvflinger)), + msg_queue(std::move(msg_queue)) { static const FunctionInfo functions[] = { {0, &ILibraryAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"}, {1, &ILibraryAppletProxy::GetSelfController, "GetSelfController"}, @@ -32,7 +34,7 @@ private: void GetCommonStateGetter(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ICommonStateGetter>(); + rb.PushIpcInterface<ICommonStateGetter>(msg_queue); LOG_DEBUG(Service_AM, "called"); } @@ -93,12 +95,15 @@ private: } std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + std::shared_ptr<AppletMessageQueue> msg_queue; }; class ISystemAppletProxy final : public ServiceFramework<ISystemAppletProxy> { public: - explicit ISystemAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger) - : ServiceFramework("ISystemAppletProxy"), nvflinger(std::move(nvflinger)) { + explicit ISystemAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, + std::shared_ptr<AppletMessageQueue> msg_queue) + : ServiceFramework("ISystemAppletProxy"), nvflinger(std::move(nvflinger)), + msg_queue(std::move(msg_queue)) { static const FunctionInfo functions[] = { {0, &ISystemAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"}, {1, &ISystemAppletProxy::GetSelfController, "GetSelfController"}, @@ -119,7 +124,7 @@ private: void GetCommonStateGetter(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ICommonStateGetter>(); + rb.PushIpcInterface<ICommonStateGetter>(msg_queue); LOG_DEBUG(Service_AM, "called"); } @@ -186,31 +191,34 @@ private: LOG_DEBUG(Service_AM, "called"); } std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + std::shared_ptr<AppletMessageQueue> msg_queue; }; void AppletAE::OpenSystemAppletProxy(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemAppletProxy>(nvflinger); + rb.PushIpcInterface<ISystemAppletProxy>(nvflinger, msg_queue); LOG_DEBUG(Service_AM, "called"); } void AppletAE::OpenLibraryAppletProxy(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger); + rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger, msg_queue); LOG_DEBUG(Service_AM, "called"); } void AppletAE::OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger); + rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger, msg_queue); LOG_DEBUG(Service_AM, "called"); } -AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger) - : ServiceFramework("appletAE"), nvflinger(std::move(nvflinger)) { +AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, + std::shared_ptr<AppletMessageQueue> msg_queue) + : ServiceFramework("appletAE"), nvflinger(std::move(nvflinger)), + msg_queue(std::move(msg_queue)) { // clang-format off static const FunctionInfo functions[] = { {100, &AppletAE::OpenSystemAppletProxy, "OpenSystemAppletProxy"}, @@ -228,4 +236,8 @@ AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger) AppletAE::~AppletAE() = default; +const std::shared_ptr<AppletMessageQueue>& AppletAE::GetMessageQueue() const { + return msg_queue; +} + } // namespace Service::AM diff --git a/src/core/hle/service/am/applet_ae.h b/src/core/hle/service/am/applet_ae.h index 1ed77baa4..902db2665 100644 --- a/src/core/hle/service/am/applet_ae.h +++ b/src/core/hle/service/am/applet_ae.h @@ -17,15 +17,19 @@ namespace AM { class AppletAE final : public ServiceFramework<AppletAE> { public: - explicit AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger); + explicit AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, + std::shared_ptr<AppletMessageQueue> msg_queue); ~AppletAE() override; + const std::shared_ptr<AppletMessageQueue>& GetMessageQueue() const; + private: void OpenSystemAppletProxy(Kernel::HLERequestContext& ctx); void OpenLibraryAppletProxy(Kernel::HLERequestContext& ctx); void OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx); std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + std::shared_ptr<AppletMessageQueue> msg_queue; }; } // namespace AM diff --git a/src/core/hle/service/am/applet_oe.cpp b/src/core/hle/service/am/applet_oe.cpp index 60717afd9..20c8d5fff 100644 --- a/src/core/hle/service/am/applet_oe.cpp +++ b/src/core/hle/service/am/applet_oe.cpp @@ -12,8 +12,10 @@ namespace Service::AM { class IApplicationProxy final : public ServiceFramework<IApplicationProxy> { public: - explicit IApplicationProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger) - : ServiceFramework("IApplicationProxy"), nvflinger(std::move(nvflinger)) { + explicit IApplicationProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, + std::shared_ptr<AppletMessageQueue> msg_queue) + : ServiceFramework("IApplicationProxy"), nvflinger(std::move(nvflinger)), + msg_queue(std::move(msg_queue)) { // clang-format off static const FunctionInfo functions[] = { {0, &IApplicationProxy::GetCommonStateGetter, "GetCommonStateGetter"}, @@ -70,7 +72,7 @@ private: void GetCommonStateGetter(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ICommonStateGetter>(); + rb.PushIpcInterface<ICommonStateGetter>(msg_queue); LOG_DEBUG(Service_AM, "called"); } @@ -89,17 +91,20 @@ private: } std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + std::shared_ptr<AppletMessageQueue> msg_queue; }; void AppletOE::OpenApplicationProxy(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IApplicationProxy>(nvflinger); + rb.PushIpcInterface<IApplicationProxy>(nvflinger, msg_queue); LOG_DEBUG(Service_AM, "called"); } -AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger) - : ServiceFramework("appletOE"), nvflinger(std::move(nvflinger)) { +AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, + std::shared_ptr<AppletMessageQueue> msg_queue) + : ServiceFramework("appletOE"), nvflinger(std::move(nvflinger)), + msg_queue(std::move(msg_queue)) { static const FunctionInfo functions[] = { {0, &AppletOE::OpenApplicationProxy, "OpenApplicationProxy"}, }; @@ -108,4 +113,8 @@ AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger) AppletOE::~AppletOE() = default; +const std::shared_ptr<AppletMessageQueue>& AppletOE::GetMessageQueue() const { + return msg_queue; +} + } // namespace Service::AM diff --git a/src/core/hle/service/am/applet_oe.h b/src/core/hle/service/am/applet_oe.h index 60cfdfd9d..bbd0108ef 100644 --- a/src/core/hle/service/am/applet_oe.h +++ b/src/core/hle/service/am/applet_oe.h @@ -17,13 +17,17 @@ namespace AM { class AppletOE final : public ServiceFramework<AppletOE> { public: - explicit AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger); + explicit AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, + std::shared_ptr<AppletMessageQueue> msg_queue); ~AppletOE() override; + const std::shared_ptr<AppletMessageQueue>& GetMessageQueue() const; + private: void OpenApplicationProxy(Kernel::HLERequestContext& ctx); std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + std::shared_ptr<AppletMessageQueue> msg_queue; }; } // namespace AM diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp index fac6785a5..d3ea57ea7 100644 --- a/src/core/hle/service/audio/audren_u.cpp +++ b/src/core/hle/service/audio/audren_u.cpp @@ -28,13 +28,13 @@ public: {1, &IAudioRenderer::GetSampleCount, "GetSampleCount"}, {2, &IAudioRenderer::GetMixBufferCount, "GetMixBufferCount"}, {3, &IAudioRenderer::GetState, "GetState"}, - {4, &IAudioRenderer::RequestUpdate, "RequestUpdate"}, + {4, &IAudioRenderer::RequestUpdateImpl, "RequestUpdate"}, {5, &IAudioRenderer::Start, "Start"}, {6, &IAudioRenderer::Stop, "Stop"}, {7, &IAudioRenderer::QuerySystemEvent, "QuerySystemEvent"}, - {8, nullptr, "SetRenderingTimeLimit"}, - {9, nullptr, "GetRenderingTimeLimit"}, - {10, nullptr, "RequestUpdateAuto"}, + {8, &IAudioRenderer::SetRenderingTimeLimit, "SetRenderingTimeLimit"}, + {9, &IAudioRenderer::GetRenderingTimeLimit, "GetRenderingTimeLimit"}, + {10, &IAudioRenderer::RequestUpdateImpl, "RequestUpdateAuto"}, {11, nullptr, "ExecuteAudioRendererRendering"}, }; // clang-format on @@ -79,7 +79,7 @@ private: LOG_DEBUG(Service_Audio, "called"); } - void RequestUpdate(Kernel::HLERequestContext& ctx) { + void RequestUpdateImpl(Kernel::HLERequestContext& ctx) { ctx.WriteBuffer(renderer->UpdateAudioRenderer(ctx.ReadBuffer())); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -110,8 +110,29 @@ private: LOG_WARNING(Service_Audio, "(STUBBED) called"); } + void SetRenderingTimeLimit(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + rendering_time_limit_percent = rp.Pop<u32>(); + ASSERT(rendering_time_limit_percent >= 0 && rendering_time_limit_percent <= 100); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + + LOG_DEBUG(Service_Audio, "called. rendering_time_limit_percent={}", + rendering_time_limit_percent); + } + + void GetRenderingTimeLimit(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_Audio, "called"); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(rendering_time_limit_percent); + } + Kernel::SharedPtr<Kernel::Event> system_event; std::unique_ptr<AudioCore::AudioRenderer> renderer; + u32 rendering_time_limit_percent = 100; }; class IAudioDevice final : public ServiceFramework<IAudioDevice> { diff --git a/src/core/hle/service/audio/hwopus.cpp b/src/core/hle/service/audio/hwopus.cpp index 7168c6a10..783c39503 100644 --- a/src/core/hle/service/audio/hwopus.cpp +++ b/src/core/hle/service/audio/hwopus.cpp @@ -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)) { diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp index ad865751f..5d6294016 100644 --- a/src/core/hle/service/filesystem/filesystem.cpp +++ b/src/core/hle/service/filesystem/filesystem.cpp @@ -303,17 +303,27 @@ ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space, static_cast<u8>(space), save_struct.DebugInfo()); if (save_data_factory == nullptr) { - return ResultCode(ErrorModule::FS, FileSys::ErrCodes::TitleNotFound); + return FileSys::ERROR_ENTITY_NOT_FOUND; } 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 FileSys::ERROR_ENTITY_NOT_FOUND; + } + + return MakeResult(save_data_factory->GetSaveDataSpaceDirectory(space)); +} + ResultVal<FileSys::VirtualDir> OpenSDMC() { LOG_TRACE(Service_FS, "Opening SDMC"); if (sdmc_factory == nullptr) { - return ResultCode(ErrorModule::FS, FileSys::ErrCodes::SdCardNotFound); + return FileSys::ERROR_SD_CARD_NOT_FOUND; } return sdmc_factory->Open(); @@ -367,6 +377,15 @@ FileSys::VirtualDir GetModificationLoadRoot(u64 title_id) { return bis_factory->GetModificationLoadRoot(title_id); } +FileSys::VirtualDir GetModificationDumpRoot(u64 title_id) { + LOG_TRACE(Service_FS, "Opening mod dump root for tid={:016X}", title_id); + + if (bis_factory == nullptr) + return nullptr; + + return bis_factory->GetModificationDumpRoot(title_id); +} + void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite) { if (overwrite) { bis_factory = nullptr; @@ -380,13 +399,21 @@ void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite) { FileSys::Mode::ReadWrite); auto load_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir), FileSys::Mode::ReadWrite); + auto dump_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::DumpDir), + FileSys::Mode::ReadWrite); - if (bis_factory == nullptr) - bis_factory = std::make_unique<FileSys::BISFactory>(nand_directory, load_directory); - if (save_data_factory == nullptr) + if (bis_factory == nullptr) { + bis_factory = + std::make_unique<FileSys::BISFactory>(nand_directory, load_directory, dump_directory); + } + + if (save_data_factory == nullptr) { save_data_factory = std::make_unique<FileSys::SaveDataFactory>(std::move(nand_directory)); - if (sdmc_factory == nullptr) + } + + if (sdmc_factory == nullptr) { sdmc_factory = std::make_unique<FileSys::SDMCFactory>(std::move(sd_directory)); + } } void InstallInterfaces(SM::ServiceManager& service_manager, FileSys::VfsFilesystem& vfs) { diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h index 6f3e0cd03..ff9182e84 100644 --- a/src/core/hle/service/filesystem/filesystem.h +++ b/src/core/hle/service/filesystem/filesystem.h @@ -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::shared_ptr<FileSys::RegisteredCacheUnion> GetUnionContents(); @@ -54,6 +55,7 @@ FileSys::RegisteredCache* GetUserNANDContents(); FileSys::RegisteredCache* GetSDMCContents(); FileSys::VirtualDir GetModificationLoadRoot(u64 title_id); +FileSys::VirtualDir GetModificationDumpRoot(u64 title_id); // Creates the SaveData, SDMC, and BIS Factories. Should be called once and before any function // above is called. diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp index c1c83a11d..038dc80b1 100644 --- a/src/core/hle/service/filesystem/fsp_srv.cpp +++ b/src/core/hle/service/filesystem/fsp_srv.cpp @@ -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" @@ -62,12 +63,12 @@ private: // Error checking if (length < 0) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ResultCode(ErrorModule::FS, ErrorDescription::InvalidLength)); + rb.Push(FileSys::ERROR_INVALID_SIZE); return; } if (offset < 0) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ResultCode(ErrorModule::FS, ErrorDescription::InvalidOffset)); + rb.Push(FileSys::ERROR_INVALID_OFFSET); return; } @@ -107,12 +108,12 @@ private: // Error checking if (length < 0) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ResultCode(ErrorModule::FS, ErrorDescription::InvalidLength)); + rb.Push(FileSys::ERROR_INVALID_SIZE); return; } if (offset < 0) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ResultCode(ErrorModule::FS, ErrorDescription::InvalidOffset)); + rb.Push(FileSys::ERROR_INVALID_OFFSET); return; } @@ -138,12 +139,12 @@ private: // Error checking if (length < 0) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ResultCode(ErrorModule::FS, ErrorDescription::InvalidLength)); + rb.Push(FileSys::ERROR_INVALID_SIZE); return; } if (offset < 0) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ResultCode(ErrorModule::FS, ErrorDescription::InvalidOffset)); + rb.Push(FileSys::ERROR_INVALID_OFFSET); return; } @@ -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); } @@ -602,7 +744,7 @@ void FSP_SRV::MountSaveData(Kernel::HLERequestContext& ctx) { if (dir.Failed()) { IPC::ResponseBuilder rb{ctx, 2, 0, 0}; - rb.Push(ResultCode(ErrorModule::FS, FileSys::ErrCodes::TitleNotFound)); + rb.Push(FileSys::ERROR_ENTITY_NOT_FOUND); return; } @@ -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"); @@ -685,7 +836,7 @@ void FSP_SRV::OpenRomStorage(Kernel::HLERequestContext& ctx) { static_cast<u8>(storage_id), title_id); IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ResultCode(ErrorModule::FS, FileSys::ErrCodes::TitleNotFound)); + rb.Push(FileSys::ERROR_ENTITY_NOT_FOUND); } } // namespace Service::FileSystem diff --git a/src/core/hle/service/filesystem/fsp_srv.h b/src/core/hle/service/filesystem/fsp_srv.h index 4aa0358cb..e7abec0a3 100644 --- a/src/core/hle/service/filesystem/fsp_srv.h +++ b/src/core/hle/service/filesystem/fsp_srv.h @@ -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); diff --git a/src/core/hle/service/hid/controllers/npad.cpp b/src/core/hle/service/hid/controllers/npad.cpp index 4b4d1324f..205e4fd14 100644 --- a/src/core/hle/service/hid/controllers/npad.cpp +++ b/src/core/hle/service/hid/controllers/npad.cpp @@ -40,6 +40,29 @@ enum class JoystickId : std::size_t { Joystick_Right, }; +static std::size_t NPadIdToIndex(u32 npad_id) { + switch (npad_id) { + case 0: + case 1: + case 2: + case 3: + case 4: + case 5: + case 6: + case 7: + return npad_id; + case 8: + case NPAD_HANDHELD: + return 8; + case 9: + case NPAD_UNKNOWN: + return 9; + default: + UNIMPLEMENTED_MSG("Unknown npad id {}", npad_id); + return 0; + } +} + Controller_NPad::Controller_NPad() = default; Controller_NPad::~Controller_NPad() = default; @@ -288,10 +311,11 @@ void Controller_NPad::OnUpdate(u8* data, std::size_t data_len) { switch (controller_type) { case NPadControllerType::Handheld: handheld_entry.connection_status.raw = 0; - handheld_entry.connection_status.IsConnected.Assign(1); - if (!Settings::values.use_docked_mode) { - handheld_entry.connection_status.IsWired.Assign(1); - } + handheld_entry.connection_status.IsWired.Assign(1); + handheld_entry.connection_status.IsLeftJoyConnected.Assign(1); + handheld_entry.connection_status.IsRightJoyConnected.Assign(1); + handheld_entry.connection_status.IsLeftJoyWired.Assign(1); + handheld_entry.connection_status.IsRightJoyWired.Assign(1); handheld_entry.pad_states.raw = pad_state.raw; handheld_entry.l_stick = lstick_entry; handheld_entry.r_stick = rstick_entry; @@ -310,6 +334,7 @@ void Controller_NPad::OnUpdate(u8* data, std::size_t data_len) { dual_entry.pad_states.raw = pad_state.raw; dual_entry.l_stick = lstick_entry; dual_entry.r_stick = rstick_entry; + break; case NPadControllerType::JoyLeft: left_entry.connection_status.raw = 0; @@ -370,16 +395,30 @@ void Controller_NPad::SetSupportedNPadIdTypes(u8* data, std::size_t length) { supported_npad_id_types.clear(); supported_npad_id_types.resize(length / sizeof(u32)); std::memcpy(supported_npad_id_types.data(), data, length); + bool had_controller_update = false; for (std::size_t i = 0; i < connected_controllers.size(); i++) { auto& controller = connected_controllers[i]; if (!controller.is_connected) { continue; } if (!IsControllerSupported(PREFERRED_CONTROLLER)) { - controller.type = DecideBestController(PREFERRED_CONTROLLER); - InitNewlyAddedControler(i); + const auto best_type = DecideBestController(PREFERRED_CONTROLLER); + const bool is_handheld = (best_type == NPadControllerType::Handheld || + PREFERRED_CONTROLLER == NPadControllerType::Handheld); + if (is_handheld) { + controller.type = NPadControllerType::None; + controller.is_connected = false; + AddNewController(best_type); + } else { + controller.type = best_type; + InitNewlyAddedControler(i); + } + had_controller_update = true; } } + if (had_controller_update) { + styleset_changed_event->Signal(); + } } void Controller_NPad::GetSupportedNpadIdTypes(u32* data, std::size_t max_length) { @@ -392,8 +431,10 @@ std::size_t Controller_NPad::GetSupportedNPadIdTypesSize() const { } void Controller_NPad::SetHoldType(NpadHoldType joy_hold_type) { + styleset_changed_event->Signal(); hold_type = joy_hold_type; } + Controller_NPad::NpadHoldType Controller_NPad::GetHoldType() const { return hold_type; } @@ -427,6 +468,9 @@ 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; } @@ -452,15 +496,11 @@ void Controller_NPad::AddNewController(NPadControllerType controller) { } void Controller_NPad::ConnectNPad(u32 npad_id) { - if (npad_id >= connected_controllers.size()) - return; - connected_controllers[npad_id].is_connected = true; + connected_controllers[NPadIdToIndex(npad_id)].is_connected = true; } void Controller_NPad::DisconnectNPad(u32 npad_id) { - if (npad_id >= connected_controllers.size()) - return; - connected_controllers[npad_id].is_connected = false; + connected_controllers[NPadIdToIndex(npad_id)].is_connected = false; } Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) { diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp index a9aa9ec78..39631b14f 100644 --- a/src/core/hle/service/hid/hid.cpp +++ b/src/core/hle/service/hid/hid.cpp @@ -96,6 +96,8 @@ public: // TODO(shinyquagsire23): Other update callbacks? (accel, gyro?) CoreTiming::ScheduleEvent(pad_update_ticks, pad_update_event); + + ReloadInputDevices(); } void ActivateController(HidController controller) { @@ -284,10 +286,10 @@ public: {519, nullptr, "GetPalmaOperationResult"}, {520, nullptr, "ReadPalmaPlayLog"}, {521, nullptr, "ResetPalmaPlayLog"}, - {522, nullptr, "SetIsPalmaAllConnectable"}, + {522, &Hid::SetIsPalmaAllConnectable, "SetIsPalmaAllConnectable"}, {523, nullptr, "SetIsPalmaPairedConnectable"}, {524, nullptr, "PairPalma"}, - {525, nullptr, "SetPalmaBoostMode"}, + {525, &Hid::SetPalmaBoostMode, "SetPalmaBoostMode"}, {1000, nullptr, "SetNpadCommunicationMode"}, {1001, nullptr, "GetNpadCommunicationMode"}, }; @@ -594,6 +596,18 @@ private: rb.Push(RESULT_SUCCESS); LOG_WARNING(Service_HID, "(STUBBED) called"); } + + void SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + LOG_WARNING(Service_HID, "(STUBBED) called"); + } + + void SetPalmaBoostMode(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + LOG_WARNING(Service_HID, "(STUBBED) called"); + } }; class HidDbg final : public ServiceFramework<HidDbg> { diff --git a/src/core/hle/service/nfp/nfp.cpp b/src/core/hle/service/nfp/nfp.cpp index c1af878fe..1d6e7756f 100644 --- a/src/core/hle/service/nfp/nfp.cpp +++ b/src/core/hle/service/nfp/nfp.cpp @@ -212,7 +212,7 @@ private: IPC::ResponseBuilder rb{ctx, 2}; auto amiibo = nfp_interface.GetAmiiboBuffer(); TagInfo tag_info{}; - std::memcpy(tag_info.uuid.data(), amiibo.uuid.data(), sizeof(tag_info.uuid.size())); + tag_info.uuid = amiibo.uuid; tag_info.uuid_length = static_cast<u8>(tag_info.uuid.size()); tag_info.protocol = 1; // TODO(ogniK): Figure out actual values diff --git a/src/core/hle/service/ns/ns.cpp b/src/core/hle/service/ns/ns.cpp index 07c1381fe..1d2978f24 100644 --- a/src/core/hle/service/ns/ns.cpp +++ b/src/core/hle/service/ns/ns.cpp @@ -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> { diff --git a/src/core/hle/service/nvflinger/buffer_queue.h b/src/core/hle/service/nvflinger/buffer_queue.h index 2fe81a560..8cff5eb71 100644 --- a/src/core/hle/service/nvflinger/buffer_queue.h +++ b/src/core/hle/service/nvflinger/buffer_queue.h @@ -58,9 +58,9 @@ public: /// Rotate source image 90 degrees clockwise Rotate90 = 0x04, /// Rotate source image 180 degrees - Roate180 = 0x03, + Rotate180 = 0x03, /// Rotate source image 270 degrees clockwise - Roate270 = 0x07, + Rotate270 = 0x07, }; struct Buffer { diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp index a4cf45267..1ec340466 100644 --- a/src/core/hle/service/service.cpp +++ b/src/core/hle/service/service.cpp @@ -80,8 +80,8 @@ namespace Service { * Creates a function string for logging, complete with the name (or header code, depending * on what's passed in) the port name, and all the cmd_buff arguments. */ -static std::string MakeFunctionString(const char* name, const char* port_name, - const u32* cmd_buff) { +[[maybe_unused]] static std::string MakeFunctionString(const char* name, const char* port_name, + const u32* cmd_buff) { // Number of params == bits 0-5 + bits 6-11 int num_params = (cmd_buff[0] & 0x3F) + ((cmd_buff[0] >> 6) & 0x3F); diff --git a/src/core/hle/service/spl/module.cpp b/src/core/hle/service/spl/module.cpp index 44a6717d0..b2de2a818 100644 --- a/src/core/hle/service/spl/module.cpp +++ b/src/core/hle/service/spl/module.cpp @@ -3,18 +3,23 @@ // Refer to the license.txt file included. #include <algorithm> +#include <chrono> #include <cstdlib> +#include <ctime> +#include <functional> #include <vector> #include "common/logging/log.h" #include "core/hle/ipc_helpers.h" #include "core/hle/service/spl/csrng.h" #include "core/hle/service/spl/module.h" #include "core/hle/service/spl/spl.h" +#include "core/settings.h" namespace Service::SPL { Module::Interface::Interface(std::shared_ptr<Module> module, const char* name) - : ServiceFramework(name), module(std::move(module)) {} + : ServiceFramework(name), module(std::move(module)), + rng(Settings::values.rng_seed.value_or(std::time(nullptr))) {} Module::Interface::~Interface() = default; @@ -23,8 +28,9 @@ void Module::Interface::GetRandomBytes(Kernel::HLERequestContext& ctx) { std::size_t size = ctx.GetWriteBufferSize(); + std::uniform_int_distribution<u16> distribution(0, std::numeric_limits<u8>::max()); std::vector<u8> data(size); - std::generate(data.begin(), data.end(), std::rand); + std::generate(data.begin(), data.end(), [&] { return static_cast<u8>(distribution(rng)); }); ctx.WriteBuffer(data); diff --git a/src/core/hle/service/spl/module.h b/src/core/hle/service/spl/module.h index 48fda6099..afa1f0295 100644 --- a/src/core/hle/service/spl/module.h +++ b/src/core/hle/service/spl/module.h @@ -4,6 +4,7 @@ #pragma once +#include <random> #include "core/hle/service/service.h" namespace Service::SPL { @@ -19,6 +20,9 @@ public: protected: std::shared_ptr<Module> module; + + private: + std::mt19937 rng; }; }; diff --git a/src/core/hle/service/time/interface.cpp b/src/core/hle/service/time/interface.cpp index 18a5d71d5..e3cbd7004 100644 --- a/src/core/hle/service/time/interface.cpp +++ b/src/core/hle/service/time/interface.cpp @@ -21,7 +21,7 @@ Time::Time(std::shared_ptr<Module> time, const char* name) {102, nullptr, "GetStandardUserSystemClockInitialYear"}, {200, nullptr, "IsStandardNetworkSystemClockAccuracySufficient"}, {300, nullptr, "CalculateMonotonicSystemClockBaseTimePoint"}, - {400, nullptr, "GetClockSnapshot"}, + {400, &Time::GetClockSnapshot, "GetClockSnapshot"}, {401, nullptr, "GetClockSnapshotFromSystemClockContext"}, {500, nullptr, "CalculateStandardUserSystemClockDifferenceByUser"}, {501, nullptr, "CalculateSpanBetween"}, diff --git a/src/core/hle/service/time/time.cpp b/src/core/hle/service/time/time.cpp index 28fd8debc..85e7b1195 100644 --- a/src/core/hle/service/time/time.cpp +++ b/src/core/hle/service/time/time.cpp @@ -15,6 +15,44 @@ namespace Service::Time { +static void PosixToCalendar(u64 posix_time, CalendarTime& calendar_time, + CalendarAdditionalInfo& additional_info, + [[maybe_unused]] const TimeZoneRule& /*rule*/) { + const std::time_t time(posix_time); + const std::tm* tm = std::localtime(&time); + if (tm == nullptr) { + calendar_time = {}; + additional_info = {}; + return; + } + calendar_time.year = tm->tm_year + 1900; + calendar_time.month = tm->tm_mon + 1; + calendar_time.day = tm->tm_mday; + calendar_time.hour = tm->tm_hour; + calendar_time.minute = tm->tm_min; + calendar_time.second = tm->tm_sec; + + additional_info.day_of_week = tm->tm_wday; + additional_info.day_of_year = tm->tm_yday; + std::memcpy(additional_info.name.data(), "UTC", sizeof("UTC")); + additional_info.utc_offset = 0; +} + +static u64 CalendarToPosix(const CalendarTime& calendar_time, + [[maybe_unused]] const TimeZoneRule& /*rule*/) { + std::tm time{}; + time.tm_year = calendar_time.year - 1900; + time.tm_mon = calendar_time.month - 1; + time.tm_mday = calendar_time.day; + + time.tm_hour = calendar_time.hour; + time.tm_min = calendar_time.minute; + time.tm_sec = calendar_time.second; + + std::time_t epoch_time = std::mktime(&time); + return static_cast<u64>(epoch_time); +} + class ISystemClock final : public ServiceFramework<ISystemClock> { public: ISystemClock() : ServiceFramework("ISystemClock") { @@ -80,8 +118,8 @@ public: {5, nullptr, "GetTimeZoneRuleVersion"}, {100, &ITimeZoneService::ToCalendarTime, "ToCalendarTime"}, {101, &ITimeZoneService::ToCalendarTimeWithMyRule, "ToCalendarTimeWithMyRule"}, - {201, nullptr, "ToPosixTime"}, - {202, nullptr, "ToPosixTimeWithMyRule"}, + {201, &ITimeZoneService::ToPosixTime, "ToPosixTime"}, + {202, &ITimeZoneService::ToPosixTimeWithMyRule, "ToPosixTimeWithMyRule"}, }; RegisterHandlers(functions); } @@ -151,24 +189,29 @@ private: rb.PushRaw(additional_info); } - void PosixToCalendar(u64 posix_time, CalendarTime& calendar_time, - CalendarAdditionalInfo& additional_info, const TimeZoneRule& /*rule*/) { - std::time_t t(posix_time); - std::tm* tm = std::localtime(&t); - if (!tm) { - return; - } - calendar_time.year = tm->tm_year + 1900; - calendar_time.month = tm->tm_mon + 1; - calendar_time.day = tm->tm_mday; - calendar_time.hour = tm->tm_hour; - calendar_time.minute = tm->tm_min; - calendar_time.second = tm->tm_sec; - - additional_info.day_of_week = tm->tm_wday; - additional_info.day_of_year = tm->tm_yday; - std::memcpy(additional_info.name.data(), "UTC", sizeof("UTC")); - additional_info.utc_offset = 0; + void ToPosixTime(Kernel::HLERequestContext& ctx) { + // TODO(ogniK): Figure out how to handle multiple times + LOG_WARNING(Service_Time, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + auto calendar_time = rp.PopRaw<CalendarTime>(); + auto posix_time = CalendarToPosix(calendar_time, {}); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw<u32>(1); // Amount of times we're returning + ctx.WriteBuffer(&posix_time, sizeof(u64)); + } + + void ToPosixTimeWithMyRule(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_Time, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + auto calendar_time = rp.PopRaw<CalendarTime>(); + auto posix_time = CalendarToPosix(calendar_time, {}); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw<u32>(1); // Amount of times we're returning + ctx.WriteBuffer(&posix_time, sizeof(u64)); } }; @@ -207,6 +250,55 @@ void Module::Interface::GetStandardLocalSystemClock(Kernel::HLERequestContext& c LOG_DEBUG(Service_Time, "called"); } +void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_Time, "called"); + + IPC::RequestParser rp{ctx}; + auto unknown_u8 = rp.PopRaw<u8>(); + + ClockSnapshot clock_snapshot{}; + + const s64 time_since_epoch{std::chrono::duration_cast<std::chrono::seconds>( + std::chrono::system_clock::now().time_since_epoch()) + .count()}; + CalendarTime calendar_time{}; + const std::time_t time(time_since_epoch); + const std::tm* tm = std::localtime(&time); + if (tm == nullptr) { + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(ResultCode(-1)); // TODO(ogniK): Find appropriate error code + return; + } + SteadyClockTimePoint steady_clock_time_point{CoreTiming::cyclesToMs(CoreTiming::GetTicks()) / + 1000}; + + LocationName location_name{"UTC"}; + calendar_time.year = tm->tm_year + 1900; + calendar_time.month = tm->tm_mon + 1; + calendar_time.day = tm->tm_mday; + calendar_time.hour = tm->tm_hour; + calendar_time.minute = tm->tm_min; + calendar_time.second = tm->tm_sec; + clock_snapshot.system_posix_time = time_since_epoch; + clock_snapshot.network_posix_time = time_since_epoch; + clock_snapshot.system_calendar_time = calendar_time; + clock_snapshot.network_calendar_time = calendar_time; + + CalendarAdditionalInfo additional_info{}; + PosixToCalendar(time_since_epoch, calendar_time, additional_info, {}); + + clock_snapshot.system_calendar_info = additional_info; + clock_snapshot.network_calendar_info = additional_info; + + clock_snapshot.steady_clock_timepoint = steady_clock_time_point; + clock_snapshot.location_name = location_name; + clock_snapshot.clock_auto_adjustment_enabled = 1; + clock_snapshot.ipc_u8 = unknown_u8; + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + ctx.WriteBuffer(&clock_snapshot, sizeof(ClockSnapshot)); +} + Module::Interface::Interface(std::shared_ptr<Module> time, const char* name) : ServiceFramework(name), time(std::move(time)) {} diff --git a/src/core/hle/service/time/time.h b/src/core/hle/service/time/time.h index 5659ecad3..77871ae07 100644 --- a/src/core/hle/service/time/time.h +++ b/src/core/hle/service/time/time.h @@ -5,6 +5,7 @@ #pragma once #include <array> +#include "common/common_funcs.h" #include "core/hle/service/service.h" namespace Service::Time { @@ -53,6 +54,23 @@ struct SystemClockContext { static_assert(sizeof(SystemClockContext) == 0x20, "SystemClockContext structure has incorrect size"); +struct ClockSnapshot { + SystemClockContext user_clock_context; + SystemClockContext network_clock_context; + s64_le system_posix_time; + s64_le network_posix_time; + CalendarTime system_calendar_time; + CalendarTime network_calendar_time; + CalendarAdditionalInfo system_calendar_info; + CalendarAdditionalInfo network_calendar_info; + SteadyClockTimePoint steady_clock_timepoint; + LocationName location_name; + u8 clock_auto_adjustment_enabled; + u8 ipc_u8; + INSERT_PADDING_BYTES(2); +}; +static_assert(sizeof(ClockSnapshot) == 0xd0, "ClockSnapshot is an invalid size"); + class Module final { public: class Interface : public ServiceFramework<Interface> { @@ -65,6 +83,7 @@ public: void GetStandardSteadyClock(Kernel::HLERequestContext& ctx); void GetTimeZoneService(Kernel::HLERequestContext& ctx); void GetStandardLocalSystemClock(Kernel::HLERequestContext& ctx); + void GetClockSnapshot(Kernel::HLERequestContext& ctx); protected: std::shared_ptr<Module> time; diff --git a/src/core/hle/service/usb/usb.cpp b/src/core/hle/service/usb/usb.cpp index c489da071..f0a831d45 100644 --- a/src/core/hle/service/usb/usb.cpp +++ b/src/core/hle/service/usb/usb.cpp @@ -132,11 +132,11 @@ public: // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "BindNoticeEvent"}, - {1, nullptr, "Unknown1"}, + {1, nullptr, "UnbindNoticeEvent"}, {2, nullptr, "GetStatus"}, {3, nullptr, "GetNotice"}, - {4, nullptr, "Unknown2"}, - {5, nullptr, "Unknown3"}, + {4, nullptr, "EnablePowerRequestNotice"}, + {5, nullptr, "DisablePowerRequestNotice"}, {6, nullptr, "ReplyPowerRequest"}, }; // clang-format on diff --git a/src/core/loader/nro.cpp b/src/core/loader/nro.cpp index bc8e402a8..c8e491fec 100644 --- a/src/core/loader/nro.cpp +++ b/src/core/loader/nro.cpp @@ -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; diff --git a/src/core/loader/nso.cpp b/src/core/loader/nso.cpp index 68efca5c0..aaf006309 100644 --- a/src/core/loader/nso.cpp +++ b/src/core/loader/nso.cpp @@ -154,7 +154,7 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(const FileSys::VfsFile& file, VAd program_image.resize(image_size); // Apply patches if necessary - if (pm && pm->HasNSOPatch(nso_header.build_id)) { + if (pm && (pm->HasNSOPatch(nso_header.build_id) || Settings::values.dump_nso)) { std::vector<u8> pi_header(program_image.size() + 0x100); std::memcpy(pi_header.data(), &nso_header, sizeof(NsoHeader)); std::memcpy(pi_header.data() + 0x100, program_image.data(), program_image.size()); diff --git a/src/core/settings.h b/src/core/settings.h index b5aeff29b..e424479f2 100644 --- a/src/core/settings.h +++ b/src/core/settings.h @@ -6,6 +6,7 @@ #include <array> #include <atomic> +#include <optional> #include <string> #include "common/common_types.h" @@ -114,8 +115,9 @@ struct Values { // System bool use_docked_mode; bool enable_nfc; - int current_user; - int language_index; + std::optional<u32> rng_seed; + s32 current_user; + s32 language_index; // Controls std::array<std::string, NativeButton::NumButtons> buttons; @@ -157,6 +159,7 @@ struct Values { bool use_gdbstub; u16 gdbstub_port; std::string program_args; + bool dump_nso; // WebService bool enable_telemetry; diff --git a/src/core/telemetry_session.cpp b/src/core/telemetry_session.cpp index 0de13edd3..a3b08c740 100644 --- a/src/core/telemetry_session.cpp +++ b/src/core/telemetry_session.cpp @@ -184,4 +184,13 @@ TelemetrySession::~TelemetrySession() { backend = nullptr; } +bool TelemetrySession::SubmitTestcase() { +#ifdef ENABLE_WEB_SERVICE + field_collection.Accept(*backend); + return backend->SubmitTestcase(); +#else + return false; +#endif +} + } // namespace Core diff --git a/src/core/telemetry_session.h b/src/core/telemetry_session.h index 2a4845797..023612b79 100644 --- a/src/core/telemetry_session.h +++ b/src/core/telemetry_session.h @@ -31,6 +31,12 @@ 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 diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index c5f7128ec..a780215c1 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -21,6 +21,7 @@ add_library(video_core STATIC macro_interpreter.h memory_manager.cpp memory_manager.h + rasterizer_cache.cpp rasterizer_cache.h rasterizer_interface.h renderer_base.cpp @@ -33,6 +34,7 @@ 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 @@ -53,6 +55,8 @@ add_library(video_core STATIC 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 diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index 7357d20d1..6de07ea56 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -37,21 +37,52 @@ 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; + } + regs.stencil_front_op_fail = Regs::StencilOp::Keep; + regs.stencil_front_op_zfail = Regs::StencilOp::Keep; + regs.stencil_front_op_zpass = Regs::StencilOp::Keep; + regs.stencil_front_func_func = Regs::ComparisonOp::Always; + regs.stencil_front_func_mask = 0xFFFFFFFF; + regs.stencil_front_mask = 0xFFFFFFFF; + regs.stencil_two_side_enable = 1; + regs.stencil_back_op_fail = Regs::StencilOp::Keep; + regs.stencil_back_op_zfail = Regs::StencilOp::Keep; + regs.stencil_back_op_zpass = Regs::StencilOp::Keep; + regs.stencil_back_func_func = Regs::ComparisonOp::Always; + regs.stencil_back_func_mask = 0xFFFFFFFF; + regs.stencil_back_mask = 0xFFFFFFFF; } void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) { // Reset the current macro. executing_macro = 0; - // 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); + // 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(); return; } // Execute the current macro. - macro_interpreter.Execute(macro_code->second, std::move(parameters)); + macro_interpreter.Execute(search->second, std::move(parameters)); } void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) { @@ -90,13 +121,23 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) { debug_context->OnEvent(Tegra::DebugContext::Event::MaxwellCommandLoaded, nullptr); } - regs.reg_array[method] = value; + if (regs.reg_array[method] != value) { + regs.reg_array[method] = value; + if (method >= MAXWELL3D_REG_INDEX(vertex_attrib_format) && + method < MAXWELL3D_REG_INDEX(vertex_attrib_format) + regs.vertex_attrib_format.size()) { + dirty_flags.vertex_attrib_format = true; + } + } switch (method) { case MAXWELL3D_REG_INDEX(macros.data): { 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]): @@ -158,9 +199,13 @@ void Maxwell3D::WriteReg(u32 method, u32 value, u32 remaining_params) { } void Maxwell3D::ProcessMacroUpload(u32 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); + 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; } void Maxwell3D::ProcessQueryGet() { diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index 443affc36..91ca57883 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -345,6 +345,14 @@ public: Invert = 6, IncrWrap = 7, DecrWrap = 8, + KeepOGL = 0x1E00, + ZeroOGL = 0, + ReplaceOGL = 0x1E01, + IncrOGL = 0x1E02, + DecrOGL = 0x1E03, + InvertOGL = 0x150A, + IncrWrapOGL = 0x8507, + DecrWrapOGL = 0x8508, }; enum class MemoryLayout : u32 { @@ -462,6 +470,16 @@ 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)) { @@ -475,12 +493,13 @@ public: INSERT_PADDING_WORDS(0x45); struct { - INSERT_PADDING_WORDS(1); + u32 upload_address; u32 data; u32 entry; + u32 bind; } macros; - INSERT_PADDING_WORDS(0x189); + INSERT_PADDING_WORDS(0x188); u32 tfb_enabled; @@ -570,7 +589,11 @@ public: u32 stencil_back_mask; u32 stencil_back_func_mask; - INSERT_PADDING_WORDS(0x13); + INSERT_PADDING_WORDS(0xC); + + u32 color_mask_common; + + INSERT_PADDING_WORDS(0x6); u32 rt_separate_frag_data; @@ -645,8 +668,14 @@ public: ComparisonOp depth_test_func; float alpha_test_ref; ComparisonOp alpha_test_func; - - INSERT_PADDING_WORDS(0x9); + u32 draw_tfb_stride; + struct { + float r; + float g; + float b; + float a; + } blend_color; + INSERT_PADDING_WORDS(0x4); struct { u32 separate_alpha; @@ -840,8 +869,9 @@ public: BitField<6, 4, u32> RT; BitField<10, 11, u32> layer; } clear_buffers; - - INSERT_PADDING_WORDS(0x4B); + INSERT_PADDING_WORDS(0xB); + std::array<ColorMask, NumRenderTargets> color_mask; + INSERT_PADDING_WORDS(0x38); struct { u32 query_address_high; @@ -982,6 +1012,12 @@ public: State state{}; MemoryManager& memory_manager; + struct DirtyFlags { + bool vertex_attrib_format = true; + }; + + DirtyFlags dirty_flags; + /// Reads a register value located at the input method address u32 GetRegisterValue(u32 method) const; @@ -994,12 +1030,25 @@ 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; - std::unordered_map<u32, std::vector<u32>> uploaded_macros; + /// Start offsets of each macro in macro_memory + std::unordered_map<u32, u32> macro_offsets; + + /// Memory for macro code + MacroMemory macro_memory; /// Macro method that is currently being executed / being fed parameters. u32 executing_macro = 0; @@ -1022,9 +1071,12 @@ private: */ void CallMacroMethod(u32 method, std::vector<u32> parameters); - /// Handles writes to the macro uploading registers. + /// Handles writes to the macro uploading register. 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(); @@ -1058,6 +1110,7 @@ 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); @@ -1070,6 +1123,10 @@ 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); @@ -1100,6 +1157,7 @@ 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); diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h index b84da512f..83a6fd875 100644 --- a/src/video_core/engines/shader_bytecode.h +++ b/src/video_core/engines/shader_bytecode.h @@ -578,6 +578,10 @@ union Instruction { } fmul32; union { + BitField<52, 1, u64> generates_cc; + } op_32; + + union { BitField<48, 1, u64> is_signed; } shift; @@ -1231,6 +1235,7 @@ union Instruction { BitField<60, 1, u64> is_b_gpr; BitField<59, 1, u64> is_c_gpr; BitField<20, 24, s64> smem_imm; + BitField<0, 5, ControlCode> flow_control_code; Attribute attribute; Sampler sampler; @@ -1658,4 +1663,4 @@ private: } }; -} // namespace Tegra::Shader
\ No newline at end of file +} // namespace Tegra::Shader diff --git a/src/video_core/macro_interpreter.cpp b/src/video_core/macro_interpreter.cpp index f6af132fb..335a8d407 100644 --- a/src/video_core/macro_interpreter.cpp +++ b/src/video_core/macro_interpreter.cpp @@ -11,7 +11,7 @@ namespace Tegra { MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {} -void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> parameters) { +void MacroInterpreter::Execute(u32 offset, std::vector<u32> parameters) { Reset(); registers[1] = parameters[0]; this->parameters = std::move(parameters); @@ -19,7 +19,7 @@ void MacroInterpreter::Execute(const std::vector<u32>& code, std::vector<u32> pa // Execute the code until we hit an exit condition. bool keep_executing = true; while (keep_executing) { - keep_executing = Step(code, false); + keep_executing = Step(offset, false); } // Assert the the macro used all the input parameters @@ -37,10 +37,10 @@ void MacroInterpreter::Reset() { next_parameter_index = 1; } -bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) { +bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) { u32 base_address = pc; - Opcode opcode = GetOpcode(code); + Opcode opcode = GetOpcode(offset); pc += 4; // Update the program counter if we were delayed @@ -108,7 +108,7 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, bool is_delay_slot) { delayed_pc = base_address + opcode.GetBranchTarget(); // Execute one more instruction due to the delay slot. - return Step(code, true); + return Step(offset, true); } break; } @@ -121,17 +121,18 @@ bool MacroInterpreter::Step(const std::vector<u32>& code, 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(code, true); + Step(offset, true); return false; } return true; } -MacroInterpreter::Opcode MacroInterpreter::GetOpcode(const std::vector<u32>& code) const { +MacroInterpreter::Opcode MacroInterpreter::GetOpcode(u32 offset) const { + const auto& macro_memory{maxwell3d.GetMacroMemory()}; ASSERT((pc % sizeof(u32)) == 0); - ASSERT(pc < code.size() * sizeof(u32)); - return {code[pc / sizeof(u32)]}; + ASSERT((pc + offset) < macro_memory.size() * sizeof(u32)); + return {macro_memory[offset + pc / sizeof(u32)]}; } u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const { diff --git a/src/video_core/macro_interpreter.h b/src/video_core/macro_interpreter.h index 773684bde..62d1ce289 100644 --- a/src/video_core/macro_interpreter.h +++ b/src/video_core/macro_interpreter.h @@ -22,10 +22,10 @@ public: /** * Executes the macro code with the specified input parameters. - * @param code The macro byte code to execute - * @param parameters The parameters of the macro + * @param offset Offset to start execution at. + * @param parameters The parameters of the macro. */ - void Execute(const std::vector<u32>& code, std::vector<u32> parameters); + void Execute(u32 offset, std::vector<u32> parameters); private: enum class Operation : u32 { @@ -110,11 +110,11 @@ private: /** * Executes a single macro instruction located at the current program counter. Returns whether * the interpreter should keep running. - * @param code The macro code to execute. + * @param offset Offset to start execution at. * @param is_delay_slot Whether the current step is being executed due to a delay slot in a * previous instruction. */ - bool Step(const std::vector<u32>& code, bool is_delay_slot); + bool Step(u32 offset, bool is_delay_slot); /// Calculates the result of an ALU operation. src_a OP src_b; u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const; @@ -127,7 +127,7 @@ private: bool EvaluateBranchCondition(BranchCondition cond, u32 value) const; /// Reads an opcode at the current program counter location. - Opcode GetOpcode(const std::vector<u32>& code) const; + Opcode GetOpcode(u32 offset) const; /// Returns the specified register's value. Register 0 is hardcoded to always return 0. u32 GetRegister(u32 register_id) const; diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp index 90a8e825d..77a20bb84 100644 --- a/src/video_core/memory_manager.cpp +++ b/src/video_core/memory_manager.cpp @@ -4,18 +4,21 @@ #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) { - std::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, align); - ASSERT(gpu_addr); + const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, align, PageStatus::Unmapped)}; - for (u64 offset = 0; offset < size; offset += PAGE_SIZE) { - VAddr& slot = PageSlot(*gpu_addr + offset); + ASSERT_MSG(gpu_addr, "unable to find available GPU memory"); + + for (u64 offset{}; offset < size; offset += PAGE_SIZE) { + VAddr& slot{PageSlot(*gpu_addr + offset)}; ASSERT(slot == static_cast<u64>(PageStatus::Unmapped)); + slot = static_cast<u64>(PageStatus::Allocated); } @@ -23,10 +26,11 @@ GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) { } GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) { - for (u64 offset = 0; offset < size; offset += PAGE_SIZE) { - VAddr& slot = PageSlot(gpu_addr + offset); + for (u64 offset{}; offset < size; offset += PAGE_SIZE) { + VAddr& slot{PageSlot(gpu_addr + offset)}; ASSERT(slot == static_cast<u64>(PageStatus::Unmapped)); + slot = static_cast<u64>(PageStatus::Allocated); } @@ -34,17 +38,19 @@ GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) { } GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) { - std::optional<GPUVAddr> gpu_addr = FindFreeBlock(size, PAGE_SIZE); - ASSERT(gpu_addr); + const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, PAGE_SIZE, PageStatus::Unmapped)}; + + ASSERT_MSG(gpu_addr, "unable to find available GPU memory"); - for (u64 offset = 0; offset < size; offset += PAGE_SIZE) { - VAddr& slot = PageSlot(*gpu_addr + offset); + for (u64 offset{}; offset < size; offset += PAGE_SIZE) { + VAddr& slot{PageSlot(*gpu_addr + offset)}; ASSERT(slot == static_cast<u64>(PageStatus::Unmapped)); + slot = cpu_addr + offset; } - MappedRegion region{cpu_addr, *gpu_addr, size}; + const MappedRegion region{cpu_addr, *gpu_addr, size}; mapped_regions.push_back(region); return *gpu_addr; @@ -53,14 +59,31 @@ 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); - for (u64 offset = 0; offset < size; offset += PAGE_SIZE) { - VAddr& slot = PageSlot(gpu_addr + offset); + 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)}; ASSERT(slot == static_cast<u64>(PageStatus::Allocated)); + slot = cpu_addr + offset; } - MappedRegion region{cpu_addr, gpu_addr, size}; + const MappedRegion region{cpu_addr, gpu_addr, size}; mapped_regions.push_back(region); return gpu_addr; @@ -69,11 +92,12 @@ 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 = 0; offset < size; offset += PAGE_SIZE) { - VAddr& slot = PageSlot(gpu_addr + offset); + for (u64 offset{}; 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); } @@ -97,13 +121,14 @@ GPUVAddr MemoryManager::GetRegionEnd(GPUVAddr region_start) const { return {}; } -std::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) { - GPUVAddr gpu_addr = 0; - u64 free_space = 0; +std::optional<GPUVAddr> MemoryManager::FindFreeBlock(GPUVAddr region_start, u64 size, u64 align, + PageStatus status) { + GPUVAddr gpu_addr{region_start}; + u64 free_space{}; align = (align + PAGE_MASK) & ~PAGE_MASK; while (gpu_addr + free_space < MAX_ADDRESS) { - if (!IsPageMapped(gpu_addr + free_space)) { + if (PageSlot(gpu_addr + free_space) == static_cast<u64>(status)) { free_space += PAGE_SIZE; if (free_space >= size) { return gpu_addr; @@ -119,7 +144,7 @@ std::optional<GPUVAddr> MemoryManager::FindFreeBlock(u64 size, u64 align) { } std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) { - VAddr base_addr = PageSlot(gpu_addr); + const VAddr base_addr{PageSlot(gpu_addr)}; if (base_addr == static_cast<u64>(PageStatus::Allocated) || base_addr == static_cast<u64>(PageStatus::Unmapped)) { @@ -133,19 +158,15 @@ 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)) { - u64 offset = cpu_addr - region.cpu_addr; + const 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)); diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h index b1255fd56..4eb338aa2 100644 --- a/src/video_core/memory_manager.h +++ b/src/video_core/memory_manager.h @@ -34,15 +34,15 @@ public: static constexpr u64 PAGE_MASK = PAGE_SIZE - 1; private: - std::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}; diff --git a/src/video_core/rasterizer_cache.cpp b/src/video_core/rasterizer_cache.cpp new file mode 100644 index 000000000..093b2cdf4 --- /dev/null +++ b/src/video_core/rasterizer_cache.cpp @@ -0,0 +1,7 @@ +// Copyright 2018 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "video_core/rasterizer_cache.h" + +RasterizerCacheObject::~RasterizerCacheObject() = default; diff --git a/src/video_core/rasterizer_cache.h b/src/video_core/rasterizer_cache.h index 0a3b3951e..bcf0c15a4 100644 --- a/src/video_core/rasterizer_cache.h +++ b/src/video_core/rasterizer_cache.h @@ -5,18 +5,19 @@ #pragma once #include <set> +#include <unordered_map> #include <boost/icl/interval_map.hpp> #include <boost/range/iterator_range_core.hpp> #include "common/common_types.h" -#include "core/core.h" #include "core/settings.h" #include "video_core/rasterizer_interface.h" -#include "video_core/renderer_base.h" class RasterizerCacheObject { public: + virtual ~RasterizerCacheObject(); + /// Gets the address of the shader in guest memory, required for cache management virtual VAddr GetAddr() const = 0; @@ -64,6 +65,8 @@ class RasterizerCache : NonCopyable { friend class RasterizerCacheObject; public: + explicit RasterizerCache(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} {} + /// Write any cached resources overlapping the specified region back to memory void FlushRegion(Tegra::GPUVAddr addr, size_t size) { const auto& objects{GetSortedObjectsFromRegion(addr, size)}; @@ -86,45 +89,39 @@ public: /// Invalidates everything in the cache void InvalidateAll() { - while (object_cache.begin() != object_cache.end()) { - Unregister(*object_cache.begin()->second.begin()); + while (interval_cache.begin() != interval_cache.end()) { + Unregister(*interval_cache.begin()->second.begin()); } } protected: /// Tries to get an object from the cache with the specified address T TryGet(VAddr addr) const { - const ObjectInterval interval{addr}; - for (auto& pair : boost::make_iterator_range(object_cache.equal_range(interval))) { - for (auto& cached_object : pair.second) { - if (cached_object->GetAddr() == addr) { - return cached_object; - } - } - } + const auto iter = map_cache.find(addr); + if (iter != map_cache.end()) + return iter->second; return nullptr; } /// Register an object into the cache void Register(const T& object) { object->SetIsRegistered(true); - object_cache.add({GetInterval(object), ObjectSet{object}}); - auto& rasterizer = Core::System::GetInstance().Renderer().Rasterizer(); + interval_cache.add({GetInterval(object), ObjectSet{object}}); + map_cache.insert({object->GetAddr(), object}); rasterizer.UpdatePagesCachedCount(object->GetAddr(), object->GetSizeInBytes(), 1); } /// Unregisters an object from the cache void Unregister(const T& object) { object->SetIsRegistered(false); - auto& rasterizer = Core::System::GetInstance().Renderer().Rasterizer(); rasterizer.UpdatePagesCachedCount(object->GetAddr(), object->GetSizeInBytes(), -1); - // Only flush if use_accurate_gpu_emulation is enabled, as it incurs a performance hit if (Settings::values.use_accurate_gpu_emulation) { FlushObject(object); } - object_cache.subtract({GetInterval(object), ObjectSet{object}}); + interval_cache.subtract({GetInterval(object), ObjectSet{object}}); + map_cache.erase(object->GetAddr()); } /// Returns a ticks counter used for tracking when cached objects were last modified @@ -141,7 +138,7 @@ private: std::vector<T> objects; const ObjectInterval interval{addr, addr + size}; - for (auto& pair : boost::make_iterator_range(object_cache.equal_range(interval))) { + for (auto& pair : boost::make_iterator_range(interval_cache.equal_range(interval))) { for (auto& cached_object : pair.second) { if (!cached_object) { continue; @@ -167,14 +164,17 @@ private: } using ObjectSet = std::set<T>; - using ObjectCache = boost::icl::interval_map<VAddr, ObjectSet>; - using ObjectInterval = typename ObjectCache::interval_type; + using ObjectCache = std::unordered_map<VAddr, T>; + using IntervalCache = boost::icl::interval_map<VAddr, ObjectSet>; + using ObjectInterval = typename IntervalCache::interval_type; static auto GetInterval(const T& object) { return ObjectInterval::right_open(object->GetAddr(), object->GetAddr() + object->GetSizeInBytes()); } - ObjectCache object_cache; ///< Cache of objects - u64 modified_ticks{}; ///< Counter of cache state ticks, used for in-order flushing + ObjectCache map_cache; + IntervalCache interval_cache; ///< Cache of objects + u64 modified_ticks{}; ///< Counter of cache state ticks, used for in-order flushing + VideoCore::RasterizerInterface& rasterizer; }; diff --git a/src/video_core/renderer_base.cpp b/src/video_core/renderer_base.cpp index 0df3725c2..1482cdb40 100644 --- a/src/video_core/renderer_base.cpp +++ b/src/video_core/renderer_base.cpp @@ -5,7 +5,6 @@ #include "core/frontend/emu_window.h" #include "core/settings.h" #include "video_core/renderer_base.h" -#include "video_core/renderer_opengl/gl_rasterizer.h" namespace VideoCore { diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp index 41a54b3e7..075192c3f 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp @@ -9,10 +9,12 @@ #include "core/core.h" #include "core/memory.h" #include "video_core/renderer_opengl/gl_buffer_cache.h" +#include "video_core/renderer_opengl/gl_rasterizer.h" namespace OpenGL { -OGLBufferCache::OGLBufferCache(std::size_t size) : stream_buffer(GL_ARRAY_BUFFER, size) {} +OGLBufferCache::OGLBufferCache(RasterizerOpenGL& rasterizer, std::size_t size) + : RasterizerCache{rasterizer}, stream_buffer(GL_ARRAY_BUFFER, size) {} GLintptr OGLBufferCache::UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size, std::size_t alignment, bool cache) { diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h index be29dc8be..91fca3f6c 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.h +++ b/src/video_core/renderer_opengl/gl_buffer_cache.h @@ -15,6 +15,8 @@ namespace OpenGL { +class RasterizerOpenGL; + struct CachedBufferEntry final : public RasterizerCacheObject { VAddr GetAddr() const override { return addr; @@ -35,7 +37,7 @@ struct CachedBufferEntry final : public RasterizerCacheObject { class OGLBufferCache final : public RasterizerCache<std::shared_ptr<CachedBufferEntry>> { public: - explicit OGLBufferCache(std::size_t size); + explicit OGLBufferCache(RasterizerOpenGL& rasterizer, std::size_t size); /// Uploads data from a guest GPU address. Returns host's buffer offset where it's been /// allocated. diff --git a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp index 741f14bc3..d9ed08437 100644 --- a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp +++ b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp @@ -6,6 +6,7 @@ #include <array> #include "common/assert.h" #include "common/common_types.h" +#include "core/core.h" #include "core/memory.h" #include "video_core/renderer_opengl/gl_buffer_cache.h" #include "video_core/renderer_opengl/gl_primitive_assembler.h" diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index bf381271e..54cc47a9b 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -30,10 +30,11 @@ namespace OpenGL { using Maxwell = Tegra::Engines::Maxwell3D::Regs; -using PixelFormat = SurfaceParams::PixelFormat; -using SurfaceType = SurfaceParams::SurfaceType; +using PixelFormat = VideoCore::Surface::PixelFormat; +using SurfaceType = VideoCore::Surface::SurfaceType; -MICROPROFILE_DEFINE(OpenGL_VAO, "OpenGL", "Vertex Array Setup", MP_RGB(128, 128, 192)); +MICROPROFILE_DEFINE(OpenGL_VAO, "OpenGL", "Vertex Format Setup", MP_RGB(128, 128, 192)); +MICROPROFILE_DEFINE(OpenGL_VB, "OpenGL", "Vertex Buffer Setup", MP_RGB(128, 128, 192)); MICROPROFILE_DEFINE(OpenGL_Shader, "OpenGL", "Shader Setup", MP_RGB(128, 128, 192)); MICROPROFILE_DEFINE(OpenGL_UBO, "OpenGL", "Const Buffer Setup", MP_RGB(128, 128, 192)); MICROPROFILE_DEFINE(OpenGL_Index, "OpenGL", "Index Buffer Setup", MP_RGB(128, 128, 192)); @@ -79,7 +80,8 @@ struct DrawParameters { }; RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo& info) - : emu_window{window}, screen_info{info}, buffer_cache(STREAM_BUFFER_SIZE) { + : res_cache{*this}, shader_cache{*this}, emu_window{window}, screen_info{info}, + buffer_cache(*this, STREAM_BUFFER_SIZE) { // Create sampler objects for (std::size_t i = 0; i < texture_samplers.size(); ++i) { texture_samplers[i].Create(); @@ -104,7 +106,7 @@ 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; @@ -115,8 +117,6 @@ 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!"); @@ -124,18 +124,23 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo RasterizerOpenGL::~RasterizerOpenGL() {} -void RasterizerOpenGL::SetupVertexArrays() { - MICROPROFILE_SCOPE(OpenGL_VAO); - const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D(); +void RasterizerOpenGL::SetupVertexFormat() { + auto& gpu = Core::System::GetInstance().GPU().Maxwell3D(); const auto& regs = gpu.regs; + if (!gpu.dirty_flags.vertex_attrib_format) + return; + gpu.dirty_flags.vertex_attrib_format = false; + + MICROPROFILE_SCOPE(OpenGL_VAO); + auto [iter, is_cache_miss] = vertex_array_cache.try_emplace(regs.vertex_attrib_format); auto& VAO = iter->second; if (is_cache_miss) { VAO.Create(); state.draw.vertex_array = VAO.handle; - state.Apply(); + state.ApplyVertexBufferState(); // The index buffer binding is stored within the VAO. Stupid OpenGL, but easy to work // around. @@ -177,8 +182,13 @@ void RasterizerOpenGL::SetupVertexArrays() { } } state.draw.vertex_array = VAO.handle; - state.draw.vertex_buffer = buffer_cache.GetHandle(); - state.Apply(); + state.ApplyVertexBufferState(); +} + +void RasterizerOpenGL::SetupVertexBuffer() { + MICROPROFILE_SCOPE(OpenGL_VB); + const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D(); + const auto& regs = gpu.regs; // Upload all guest vertex arrays sequentially to our buffer for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) { @@ -205,6 +215,9 @@ void RasterizerOpenGL::SetupVertexArrays() { glVertexBindingDivisor(index, 0); } } + + // Implicit set by glBindVertexBuffer. Stupid glstate handling... + state.draw.vertex_buffer = buffer_cache.GetHandle(); } DrawParameters RasterizerOpenGL::SetupDraw() { @@ -329,8 +342,6 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { index++; } } - - state.Apply(); } std::size_t RasterizerOpenGL::CalculateVertexArraysSize() const { @@ -399,8 +410,8 @@ void RasterizerOpenGL::UpdatePagesCachedCount(VAddr addr, u64 size, int delta) { cached_pages.add({pages_interval, delta}); } -void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_depth_fb, - bool preserve_contents, +void RasterizerOpenGL::ConfigureFramebuffers(OpenGLState& current_state, bool using_color_fb, + bool using_depth_fb, bool preserve_contents, std::optional<std::size_t> single_color_target) { MICROPROFILE_SCOPE(OpenGL_Framebuffer); const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; @@ -416,9 +427,9 @@ void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_dep ASSERT_MSG(regs.rt_separate_frag_data == 0, "Unimplemented"); // Bind the framebuffer surfaces - state.draw.draw_framebuffer = framebuffer.handle; - state.Apply(); - state.framebuffer_srgb.enabled = regs.framebuffer_srgb != 0; + current_state.draw.draw_framebuffer = framebuffer.handle; + current_state.ApplyFramebufferState(); + current_state.framebuffer_srgb.enabled = regs.framebuffer_srgb != 0; if (using_color_fb) { if (single_color_target) { @@ -496,10 +507,7 @@ void RasterizerOpenGL::ConfigureFramebuffers(bool using_color_fb, bool using_dep glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0); } - - SyncViewport(); - - state.Apply(); + SyncViewport(current_state); } void RasterizerOpenGL::Clear() { @@ -512,22 +520,23 @@ void RasterizerOpenGL::Clear() { bool use_stencil{}; OpenGLState clear_state; - clear_state.draw.draw_framebuffer = framebuffer.handle; - 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) { use_color = true; } + if (use_color) { + 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; + } if (regs.clear_buffers.Z) { ASSERT_MSG(regs.zeta_enable != 0, "Tried to clear Z but buffer is not enabled!"); use_depth = true; // Always enable the depth write when clearing the depth buffer. The depth write mask is - // ignored when clearing the buffer in the Switch, but OpenGL obeys it so we set it to true. + // ignored when clearing the buffer in the Switch, but OpenGL obeys it so we set it to + // true. clear_state.depth.test_enabled = true; clear_state.depth.test_func = GL_ALWAYS; } @@ -544,11 +553,8 @@ void RasterizerOpenGL::Clear() { ScopeAcquireGLContext acquire_context{emu_window}; - ConfigureFramebuffers(use_color, use_depth || use_stencil, false, + ConfigureFramebuffers(clear_state, use_color, use_depth || use_stencil, false, regs.clear_buffers.RT.Value()); - // Copy the sRGB setting to the clear state to avoid problem with - // specific driver implementations - clear_state.framebuffer_srgb.enabled = state.framebuffer_srgb.enabled; clear_state.Apply(); if (use_color) { @@ -574,15 +580,14 @@ void RasterizerOpenGL::DrawArrays() { ScopeAcquireGLContext acquire_context{emu_window}; - ConfigureFramebuffers(); - + ConfigureFramebuffers(state); + SyncColorMask(); SyncDepthTestState(); SyncStencilTestState(); SyncBlendState(); SyncLogicOpState(); SyncCullMode(); SyncPrimitiveRestart(); - SyncDepthRange(); SyncScissorTest(); // Alpha Testing is synced on shaders. SyncTransformFeedback(); @@ -596,7 +601,7 @@ void RasterizerOpenGL::DrawArrays() { const bool is_indexed = accelerate_draw == AccelDraw::Indexed; state.draw.vertex_buffer = buffer_cache.GetHandle(); - state.Apply(); + state.ApplyVertexBufferState(); std::size_t buffer_size = CalculateVertexArraysSize(); @@ -623,7 +628,8 @@ void RasterizerOpenGL::DrawArrays() { buffer_cache.Map(buffer_size); - SetupVertexArrays(); + SetupVertexFormat(); + SetupVertexBuffer(); DrawParameters params = SetupDraw(); SetupShaders(params.primitive_mode); @@ -703,7 +709,8 @@ 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{SurfaceParams::PixelFormatFromGPUPixelFormat(config.pixel_format)}; + const auto& pixel_format{ + VideoCore::Surface::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"); @@ -726,9 +733,9 @@ void RasterizerOpenGL::SamplerInfo::Create() { glSamplerParameteri(sampler.handle, GL_TEXTURE_COMPARE_FUNC, GL_NEVER); } -void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntry& config) { +void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::FullTextureInfo& info) { const GLuint s = sampler.handle; - + const Tegra::Texture::TSCEntry& config = info.tsc; if (mag_filter != config.mag_filter) { mag_filter = config.mag_filter; glSamplerParameteri( @@ -779,6 +786,22 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr glSamplerParameterfv(s, GL_TEXTURE_BORDER_COLOR, border_color.data()); } } + if (info.tic.use_header_opt_control == 0) { + if (GLAD_GL_ARB_texture_filter_anisotropic) { + glSamplerParameterf(s, GL_TEXTURE_MAX_ANISOTROPY, + static_cast<float>(1 << info.tic.max_anisotropy.Value())); + } else if (GLAD_GL_EXT_texture_filter_anisotropic) { + glSamplerParameterf(s, GL_TEXTURE_MAX_ANISOTROPY_EXT, + static_cast<float>(1 << info.tic.max_anisotropy.Value())); + } + glSamplerParameterf(s, GL_TEXTURE_MIN_LOD, + static_cast<float>(info.tic.res_min_mip_level.Value())); + glSamplerParameterf(s, GL_TEXTURE_MAX_LOD, + static_cast<float>(info.tic.res_max_mip_level.Value() == 0 + ? 16 + : info.tic.res_max_mip_level.Value())); + glSamplerParameterf(s, GL_TEXTURE_LOD_BIAS, info.tic.mip_lod_bias.Value() / 256.f); + } } u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, Shader& shader, @@ -876,7 +899,7 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader, continue; } - texture_samplers[current_bindpoint].SyncWithConfig(texture.tsc); + texture_samplers[current_bindpoint].SyncWithConfig(texture); Surface surface = res_cache.GetTextureSurface(texture, entry); if (surface != nullptr) { state.texture_units[current_bindpoint].texture = surface->Texture().handle; @@ -898,14 +921,18 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader, return current_unit + static_cast<u32>(entries.size()); } -void RasterizerOpenGL::SyncViewport() { +void RasterizerOpenGL::SyncViewport(OpenGLState& current_state) { const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; - 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()); + 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 = current_state.viewports[i]; + viewport.x = viewport_rect.left; + viewport.y = viewport_rect.bottom; + viewport.width = static_cast<GLfloat>(viewport_rect.GetWidth()); + viewport.height = static_cast<GLfloat>(viewport_rect.GetHeight()); + viewport.depth_range_far = regs.viewport[i].depth_range_far; + viewport.depth_range_near = regs.viewport[i].depth_range_near; + } } void RasterizerOpenGL::SyncClipEnabled() { @@ -947,13 +974,6 @@ 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; @@ -974,9 +994,6 @@ void RasterizerOpenGL::SyncStencilTestState() { return; } - // TODO(bunnei): Verify behavior when this is not set - ASSERT(regs.stencil_two_side_enable); - state.stencil.front.test_func = MaxwellToGL::ComparisonOp(regs.stencil_front_func_func); state.stencil.front.test_ref = regs.stencil_front_func_ref; state.stencil.front.test_mask = regs.stencil_front_func_mask; @@ -984,36 +1001,79 @@ void RasterizerOpenGL::SyncStencilTestState() { state.stencil.front.action_depth_fail = MaxwellToGL::StencilOp(regs.stencil_front_op_zfail); state.stencil.front.action_depth_pass = MaxwellToGL::StencilOp(regs.stencil_front_op_zpass); state.stencil.front.write_mask = regs.stencil_front_mask; + if (regs.stencil_two_side_enable) { + state.stencil.back.test_func = MaxwellToGL::ComparisonOp(regs.stencil_back_func_func); + state.stencil.back.test_ref = regs.stencil_back_func_ref; + state.stencil.back.test_mask = regs.stencil_back_func_mask; + state.stencil.back.action_stencil_fail = MaxwellToGL::StencilOp(regs.stencil_back_op_fail); + state.stencil.back.action_depth_fail = MaxwellToGL::StencilOp(regs.stencil_back_op_zfail); + state.stencil.back.action_depth_pass = MaxwellToGL::StencilOp(regs.stencil_back_op_zpass); + state.stencil.back.write_mask = regs.stencil_back_mask; + } else { + state.stencil.back.test_func = GL_ALWAYS; + state.stencil.back.test_ref = 0; + state.stencil.back.test_mask = 0xFFFFFFFF; + state.stencil.back.write_mask = 0xFFFFFFFF; + state.stencil.back.action_stencil_fail = GL_KEEP; + state.stencil.back.action_depth_fail = GL_KEEP; + state.stencil.back.action_depth_pass = GL_KEEP; + } +} - state.stencil.back.test_func = MaxwellToGL::ComparisonOp(regs.stencil_back_func_func); - state.stencil.back.test_ref = regs.stencil_back_func_ref; - state.stencil.back.test_mask = regs.stencil_back_func_mask; - state.stencil.back.action_stencil_fail = MaxwellToGL::StencilOp(regs.stencil_back_op_fail); - state.stencil.back.action_depth_fail = MaxwellToGL::StencilOp(regs.stencil_back_op_zfail); - state.stencil.back.action_depth_pass = MaxwellToGL::StencilOp(regs.stencil_back_op_zpass); - 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; - // TODO(Subv): Support more than just render target 0. - state.blend.enabled = regs.blend.enable[0] != 0; - - if (!state.blend.enabled) + 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; + + 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; + } return; + } - 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); + 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); + } + } } void RasterizerOpenGL::SyncLogicOpState() { @@ -1032,19 +1092,19 @@ 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); - // 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; + if (regs.scissor_test.enable == 0) { + return; } + 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() { @@ -1069,9 +1129,8 @@ void RasterizerOpenGL::CheckAlphaTests() { const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; if (regs.alpha_test_enabled != 0 && regs.rt_control.count > 1) { - LOG_CRITICAL( - Render_OpenGL, - "Alpha Testing is enabled with Multiple Render Targets, this behavior is undefined."); + LOG_CRITICAL(Render_OpenGL, "Alpha Testing is enabled with Multiple Render Targets, " + "this behavior is undefined."); UNREACHABLE(); } } diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index 47097c569..8ef0f6c12 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -88,7 +88,7 @@ private: /// SamplerInfo struct. void Create(); /// Syncs the sampler object with the config, updating any necessary state. - void SyncWithConfig(const Tegra::Texture::TSCEntry& config); + void SyncWithConfig(const Tegra::Texture::FullTextureInfo& info); private: Tegra::Texture::TextureFilter mag_filter; @@ -109,8 +109,8 @@ private: * @param preserve_contents If true, tries to preserve data from a previously used framebuffer. * @param single_color_target Specifies if a single color buffer target should be used. */ - void ConfigureFramebuffers(bool use_color_fb = true, bool using_depth_fb = true, - bool preserve_contents = true, + void ConfigureFramebuffers(OpenGLState& current_state, bool use_color_fb = true, + bool using_depth_fb = true, bool preserve_contents = true, std::optional<std::size_t> single_color_target = {}); /* @@ -133,8 +133,8 @@ private: u32 SetupTextures(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, Shader& shader, GLenum primitive_mode, u32 current_unit); - /// Syncs the viewport to match the guest state - void SyncViewport(); + /// Syncs the viewport and depth range to match the guest state + void SyncViewport(OpenGLState& current_state); /// Syncs the clip enabled status to match the guest state void SyncClipEnabled(); @@ -148,9 +148,6 @@ 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(); @@ -172,6 +169,9 @@ private: /// Syncs the point state to match the guest state void SyncPointState(); + /// Syncs Color Mask + void SyncColorMask(); + /// Check asserts for alpha testing. void CheckAlphaTests(); @@ -207,7 +207,8 @@ private: std::size_t CalculateIndexBufferSize() const; - void SetupVertexArrays(); + void SetupVertexFormat(); + void SetupVertexBuffer(); DrawParameters SetupDraw(); diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp index 1d43a419d..26711e6f7 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp @@ -15,17 +15,24 @@ #include "core/memory.h" #include "core/settings.h" #include "video_core/engines/maxwell_3d.h" +#include "video_core/renderer_opengl/gl_rasterizer.h" #include "video_core/renderer_opengl/gl_rasterizer_cache.h" +#include "video_core/renderer_opengl/gl_state.h" #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 SurfaceType = SurfaceParams::SurfaceType; -using PixelFormat = SurfaceParams::PixelFormat; -using ComponentType = SurfaceParams::ComponentType; +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; struct FormatTuple { GLint internal_format; @@ -35,46 +42,6 @@ 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_)}; @@ -93,16 +60,14 @@ 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 compression_factor{GetCompressionFactor(pixel_format)}; + const u32 tile_x{GetDefaultBlockWidth(pixel_format)}; + const u32 tile_y{GetDefaultBlockHeight(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 + compression_factor - 1) / compression_factor); - m_height = uncompressed - ? m_height - : std::max(1U, (m_height + compression_factor - 1) / compression_factor); + 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); @@ -163,6 +128,13 @@ 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(); @@ -267,7 +239,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only, return params; } -static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{ +static constexpr std::array<FormatTuple, VideoCore::Surface::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 @@ -340,6 +312,10 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form {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 + {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_10X8 + {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_10X8_SRGB // Depth formats {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F @@ -355,20 +331,22 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form ComponentType::Float, false}, // Z32FS8 }}; -static GLenum SurfaceTargetToGL(SurfaceParams::SurfaceTarget target) { +static GLenum SurfaceTargetToGL(SurfaceTarget target) { switch (target) { - case SurfaceParams::SurfaceTarget::Texture1D: + case SurfaceTarget::Texture1D: return GL_TEXTURE_1D; - case SurfaceParams::SurfaceTarget::Texture2D: + case SurfaceTarget::Texture2D: return GL_TEXTURE_2D; - case SurfaceParams::SurfaceTarget::Texture3D: + case SurfaceTarget::Texture3D: return GL_TEXTURE_3D; - case SurfaceParams::SurfaceTarget::Texture1DArray: + case SurfaceTarget::Texture1DArray: return GL_TEXTURE_1D_ARRAY; - case SurfaceParams::SurfaceTarget::Texture2DArray: + case SurfaceTarget::Texture2DArray: return GL_TEXTURE_2D_ARRAY; - case SurfaceParams::SurfaceTarget::TextureCubemap: + case 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(); @@ -392,50 +370,32 @@ MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const { return {0, actual_height, MipWidth(mip_level), 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 = SurfaceParams::GetBytesPerPixel(format); + constexpr u32 bytes_per_pixel = GetBytesPerPixel(format); // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual // pixel values. - const u32 tile_size{IsFormatBCn(format) ? 4U : 1U}; + const u32 tile_size_x{GetDefaultBlockWidth(format)}; + const u32 tile_size_y{GetDefaultBlockHeight(format)}; if (morton_to_gl) { - const std::vector<u8> data = Tegra::Texture::UnswizzleTexture( - addr, tile_size, bytes_per_pixel, stride, height, depth, block_height, block_depth); + 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::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, height / tile_size, depth, + Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x, + (height + tile_size_y - 1) / tile_size_y, depth, bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr), gl_buffer, false, block_height, block_depth); } } using GLConversionArray = std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr), - SurfaceParams::MaxPixelFormat>; + VideoCore::Surface::MaxPixelFormat>; static constexpr GLConversionArray morton_to_gl_fns = { // clang-format off @@ -496,6 +456,10 @@ 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::ASTC_2D_10X8>, + MortonCopy<true, PixelFormat::ASTC_2D_10X8_SRGB>, MortonCopy<true, PixelFormat::Z32F>, MortonCopy<true, PixelFormat::Z16>, MortonCopy<true, PixelFormat::Z24S8>, @@ -564,6 +528,10 @@ static constexpr GLConversionArray gl_to_morton_fns = { nullptr, nullptr, nullptr, + nullptr, + nullptr, + nullptr, + nullptr, MortonCopy<false, PixelFormat::Z32F>, MortonCopy<false, PixelFormat::Z16>, MortonCopy<false, PixelFormat::Z24S8>, @@ -575,7 +543,7 @@ 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 == SurfaceParams::SurfaceTarget::Texture2D) { + if (params.target == SurfaceTarget::Texture2D) { // TODO(Blinkhawk): Eliminate this condition once all texture types are implemented. depth = 1U; } @@ -601,9 +569,11 @@ void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params } } +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()}; @@ -616,19 +586,19 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface, state.draw.draw_framebuffer = draw_fb_handle; // Set sRGB enabled if the destination surfaces need it state.framebuffer_srgb.enabled = dst_params.srgb_conversion; - state.Apply(); + state.ApplyFramebufferState(); u32 buffers{}; if (src_params.type == SurfaceType::ColorTexture) { switch (src_params.target) { - case SurfaceParams::SurfaceTarget::Texture2D: + case 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 SurfaceParams::SurfaceTarget::TextureCubemap: + case SurfaceTarget::TextureCubemap: glFramebufferTexture2D( GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment, static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), @@ -637,12 +607,12 @@ 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 SurfaceParams::SurfaceTarget::Texture2DArray: + case 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 SurfaceParams::SurfaceTarget::Texture3D: + case SurfaceTarget::Texture3D: glFramebufferTexture3D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment, SurfaceTargetToGL(src_params.target), src_surface->Texture().handle, 0, 0); @@ -658,13 +628,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface, } switch (dst_params.target) { - case SurfaceParams::SurfaceTarget::Texture2D: + case 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 SurfaceParams::SurfaceTarget::TextureCubemap: + case SurfaceTarget::TextureCubemap: glFramebufferTexture2D( GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment, static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), @@ -673,13 +643,13 @@ 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 SurfaceParams::SurfaceTarget::Texture2DArray: + case 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 SurfaceParams::SurfaceTarget::Texture3D: + case SurfaceTarget::Texture3D: glFramebufferTexture3D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment, SurfaceTargetToGL(dst_params.target), dst_surface->Texture().handle, 0, 0); @@ -743,18 +713,20 @@ 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) { + const GLuint copy_pbo_handle, const GLenum src_attachment = 0, + const GLenum dst_attachment = 0, const std::size_t cubemap_face = 0) { + MICROPROFILE_SCOPE(OpenGL_CopySurface); ASSERT_MSG(dst_attachment == 0, "Unimplemented"); const auto& src_params{src_surface->GetSurfaceParams()}; const auto& dst_params{dst_surface->GetSurfaceParams()}; - auto source_format = GetFormatTuple(src_params.pixel_format, src_params.component_type); - auto dest_format = GetFormatTuple(dst_params.pixel_format, dst_params.component_type); + const auto source_format = GetFormatTuple(src_params.pixel_format, src_params.component_type); + const auto dest_format = GetFormatTuple(dst_params.pixel_format, dst_params.component_type); - std::size_t buffer_size = std::max(src_params.size_in_bytes, dst_params.size_in_bytes); + const std::size_t buffer_size = std::max(src_params.size_in_bytes, dst_params.size_in_bytes); glBindBuffer(GL_PIXEL_PACK_BUFFER, copy_pbo_handle); glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW_ARB); @@ -778,13 +750,10 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface, LOG_DEBUG(HW_GPU, "Trying to upload extra texture data from the CPU during " "reinterpretation but the texture is tiled."); } - std::size_t remaining_size = dst_params.size_in_bytes - src_params.size_in_bytes; - std::vector<u8> data(remaining_size); - std::memcpy(data.data(), Memory::GetPointer(dst_params.addr + src_params.size_in_bytes), - data.size()); + const std::size_t remaining_size = dst_params.size_in_bytes - src_params.size_in_bytes; glBufferSubData(GL_PIXEL_PACK_BUFFER, src_params.size_in_bytes, remaining_size, - data.data()); + Memory::GetPointer(dst_params.addr + src_params.size_in_bytes)); } glBindBuffer(GL_PIXEL_PACK_BUFFER, 0); @@ -800,21 +769,22 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface, UNREACHABLE(); } else { switch (dst_params.target) { - case SurfaceParams::SurfaceTarget::Texture1D: + case SurfaceTarget::Texture1D: glTextureSubImage1D(dst_surface->Texture().handle, 0, 0, width, dest_format.format, dest_format.type, nullptr); break; - case SurfaceParams::SurfaceTarget::Texture2D: + case SurfaceTarget::Texture2D: glTextureSubImage2D(dst_surface->Texture().handle, 0, 0, 0, width, height, dest_format.format, dest_format.type, nullptr); break; - case SurfaceParams::SurfaceTarget::Texture3D: - case SurfaceParams::SurfaceTarget::Texture2DArray: + case SurfaceTarget::Texture3D: + case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubeArray: 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 SurfaceParams::SurfaceTarget::TextureCubemap: + case 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); @@ -851,17 +821,18 @@ CachedSurface::CachedSurface(const SurfaceParams& params) if (!format_tuple.compressed) { // Only pre-create the texture for non-compressed textures. switch (params.target) { - case SurfaceParams::SurfaceTarget::Texture1D: + case SurfaceTarget::Texture1D: glTexStorage1D(SurfaceTargetToGL(params.target), params.max_mip_level, format_tuple.internal_format, rect.GetWidth()); break; - case SurfaceParams::SurfaceTarget::Texture2D: - case SurfaceParams::SurfaceTarget::TextureCubemap: + case SurfaceTarget::Texture2D: + case SurfaceTarget::TextureCubemap: glTexStorage2D(SurfaceTargetToGL(params.target), params.max_mip_level, format_tuple.internal_format, rect.GetWidth(), rect.GetHeight()); break; - case SurfaceParams::SurfaceTarget::Texture3D: - case SurfaceParams::SurfaceTarget::Texture2DArray: + 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); @@ -916,7 +887,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height, bo S8Z24 s8z24_pixel{}; Z24S8 z24s8_pixel{}; - constexpr auto bpp{SurfaceParams::GetBytesPerPixel(PixelFormat::S8Z24)}; + constexpr auto bpp{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)}; @@ -936,7 +907,7 @@ 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{SurfaceParams::GetBytesPerPixel(PixelFormat::G8R8U)}; + constexpr auto bpp{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)}; @@ -953,21 +924,26 @@ 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 width, u32 height, u32 depth) { 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_5X4_SRGB: + case PixelFormat::ASTC_2D_5X5_SRGB: + case PixelFormat::ASTC_2D_10X8: + case PixelFormat::ASTC_2D_10X8_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, block_width, block_height); + data = + Tegra::Texture::ASTC::Decompress(data, width, height, depth, block_width, block_height); break; } case PixelFormat::S8Z24: @@ -996,7 +972,11 @@ static void ConvertFormatAsNeeded_FlushGLBuffer(std::vector<u8>& data, PixelForm case PixelFormat::ASTC_2D_4X4: case PixelFormat::ASTC_2D_8X8: case PixelFormat::ASTC_2D_4X4_SRGB: - case PixelFormat::ASTC_2D_8X8_SRGB: { + case PixelFormat::ASTC_2D_8X8_SRGB: + case PixelFormat::ASTC_2D_5X5: + case PixelFormat::ASTC_2D_5X5_SRGB: + case PixelFormat::ASTC_2D_10X8: + case PixelFormat::ASTC_2D_10X8_SRGB: { LOG_CRITICAL(HW_GPU, "Conversion of format {} after texture flushing is not implemented", static_cast<u32>(pixel_format)); UNREACHABLE(); @@ -1009,7 +989,7 @@ static void ConvertFormatAsNeeded_FlushGLBuffer(std::vector<u8>& data, PixelForm } } -MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 64, 192)); +MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 192, 64)); void CachedSurface::LoadGLBuffer() { MICROPROFILE_SCOPE(OpenGL_SurfaceLoad); gl_buffer.resize(params.max_mip_level); @@ -1027,7 +1007,7 @@ void CachedSurface::LoadGLBuffer() { } 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.MipHeight(i), params.MipDepth(i)); } MICROPROFILE_DEFINE(OpenGL_SurfaceFlush, "OpenGL", "Surface Flush", MP_RGB(128, 192, 64)); @@ -1042,7 +1022,7 @@ void CachedSurface::FlushGLBuffer() { const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type); // Ensure no bad interactions with GL_UNPACK_ALIGNMENT - ASSERT(params.width * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 == 0); + ASSERT(params.width * 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); @@ -1074,7 +1054,7 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, 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>(x0)) * - SurfaceParams::GetBytesPerPixel(params.pixel_format); + GetBytesPerPixel(params.pixel_format); const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type); const GLuint target_tex = texture.handle; @@ -1090,35 +1070,35 @@ 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) * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 == - 0); + ASSERT(params.MipWidth(mip_map) * GetBytesPerPixel(params.pixel_format) % 4 == 0); glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.MipWidth(mip_map))); GLsizei image_size = static_cast<GLsizei>(params.GetMipmapSizeGL(mip_map, false)); glActiveTexture(GL_TEXTURE0); if (tuple.compressed) { switch (params.target) { - case SurfaceParams::SurfaceTarget::Texture2D: + 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]); break; - case SurfaceParams::SurfaceTarget::Texture3D: + 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]); break; - case SurfaceParams::SurfaceTarget::Texture2DArray: + 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 SurfaceParams::SurfaceTarget::TextureCubemap: { + case SurfaceTarget::TextureCubemap: { GLsizei layer_size = static_cast<GLsizei>(params.LayerSizeGL(mip_map)); for (std::size_t face = 0; face < params.depth; ++face) { glCompressedTexImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), @@ -1143,30 +1123,31 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, } else { switch (params.target) { - case SurfaceParams::SurfaceTarget::Texture1D: + case SurfaceTarget::Texture1D: glTexSubImage1D(SurfaceTargetToGL(params.target), mip_map, x0, static_cast<GLsizei>(rect.GetWidth()), tuple.format, tuple.type, &gl_buffer[mip_map][buffer_offset]); break; - case SurfaceParams::SurfaceTarget::Texture2D: + case SurfaceTarget::Texture2D: glTexSubImage2D(SurfaceTargetToGL(params.target), mip_map, x0, y0, static_cast<GLsizei>(rect.GetWidth()), static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type, &gl_buffer[mip_map][buffer_offset]); break; - case SurfaceParams::SurfaceTarget::Texture3D: + 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 SurfaceParams::SurfaceTarget::Texture2DArray: + case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubeArray: glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0, static_cast<GLsizei>(rect.GetWidth()), static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format, tuple.type, &gl_buffer[mip_map][buffer_offset]); break; - case SurfaceParams::SurfaceTarget::TextureCubemap: { + case SurfaceTarget::TextureCubemap: { std::size_t start = buffer_offset; for (std::size_t face = 0; face < params.depth; ++face) { glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), mip_map, @@ -1190,7 +1171,7 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, glPixelStorei(GL_UNPACK_ROW_LENGTH, 0); } -MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 64, 192)); +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; @@ -1201,7 +1182,8 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle UploadGLMipmapTexture(i, read_fb_handle, draw_fb_handle); } -RasterizerCacheOpenGL::RasterizerCacheOpenGL() { +RasterizerCacheOpenGL::RasterizerCacheOpenGL(RasterizerOpenGL& rasterizer) + : RasterizerCache{rasterizer} { read_framebuffer.Create(); draw_framebuffer.Create(); copy_pbo.Create(); @@ -1341,8 +1323,7 @@ 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 && - SurfaceParams::GetFormatBpp(old_params.pixel_format) == - SurfaceParams::GetFormatBpp(new_params.pixel_format)) { + GetFormatBpp(old_params.pixel_format) == GetFormatBpp(new_params.pixel_format)) { FastCopySurface(old_surface, new_surface); return new_surface; } @@ -1355,15 +1336,17 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface, const bool is_blit{old_params.pixel_format == new_params.pixel_format}; switch (new_params.target) { - case SurfaceParams::SurfaceTarget::Texture2D: + case 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 SurfaceParams::SurfaceTarget::TextureCubemap: - case SurfaceParams::SurfaceTarget::Texture3D: + case SurfaceTarget::TextureCubemap: + case SurfaceTarget::Texture3D: + case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubeArray: AccurateCopySurface(old_surface, new_surface); break; default: @@ -1373,7 +1356,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface, } return new_surface; -} // namespace OpenGL +} Surface RasterizerCacheOpenGL::TryFindFramebufferSurface(VAddr addr) const { return TryGet(addr); diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h index e72f4f2d2..494f6b903 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h @@ -7,6 +7,7 @@ #include <array> #include <map> #include <memory> +#include <string> #include <vector> #include "common/alignment.h" @@ -18,6 +19,7 @@ #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" @@ -27,135 +29,12 @@ class CachedSurface; using Surface = std::shared_ptr<CachedSurface>; using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, MathUtil::Rectangle<u32>>; -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; - } - } +using SurfaceTarget = VideoCore::Surface::SurfaceTarget; +using SurfaceType = VideoCore::Surface::SurfaceType; +using PixelFormat = VideoCore::Surface::PixelFormat; +using ComponentType = VideoCore::Surface::ComponentType; +struct SurfaceParams { static std::string SurfaceTargetName(SurfaceTarget target) { switch (target) { case SurfaceTarget::Texture1D: @@ -170,6 +49,8 @@ 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(); @@ -177,660 +58,8 @@ 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 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; + return VideoCore::Surface::GetFormatBpp(pixel_format); } /// Returns the rectangle corresponding to this surface @@ -912,7 +141,7 @@ struct SurfaceParams { } u32 MipDepth(u32 mip_level) const { - return std::max(1U, depth >> mip_level); + return is_layered ? depth : std::max(1U, depth >> mip_level); } // Auto block resizing algorithm from: @@ -1035,6 +264,8 @@ struct hash<SurfaceReserveKey> { namespace OpenGL { +class RasterizerOpenGL; + class CachedSurface final : public RasterizerCacheObject { public: CachedSurface(const SurfaceParams& params); @@ -1082,7 +313,7 @@ private: class RasterizerCacheOpenGL final : public RasterizerCache<Surface> { public: - RasterizerCacheOpenGL(); + explicit RasterizerCacheOpenGL(RasterizerOpenGL& rasterizer); /// Get a surface based on the texture configuration Surface GetTextureSurface(const Tegra::Texture::FullTextureInfo& config, diff --git a/src/video_core/renderer_opengl/gl_resource_manager.cpp b/src/video_core/renderer_opengl/gl_resource_manager.cpp new file mode 100644 index 000000000..c17d5ac00 --- /dev/null +++ b/src/video_core/renderer_opengl/gl_resource_manager.cpp @@ -0,0 +1,186 @@ +// 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 diff --git a/src/video_core/renderer_opengl/gl_resource_manager.h b/src/video_core/renderer_opengl/gl_resource_manager.h index 3bc1b83b5..e33f1e973 100644 --- a/src/video_core/renderer_opengl/gl_resource_manager.h +++ b/src/video_core/renderer_opengl/gl_resource_manager.h @@ -8,7 +8,6 @@ #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 { @@ -29,20 +28,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenTextures(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteTextures(1, &handle); - OpenGLState::GetCurState().UnbindTexture(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -64,20 +53,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenSamplers(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteSamplers(1, &handle); - OpenGLState::GetCurState().ResetSampler(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -98,20 +77,9 @@ public: return *this; } - void Create(const char* source, GLenum type) { - if (handle != 0) - return; - if (source == nullptr) - return; - handle = GLShader::LoadShader(source, type); - } + void Create(const char* source, GLenum type); - void Release() { - if (handle == 0) - return; - glDeleteShader(handle); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -141,25 +109,10 @@ 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) { - 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); - } + bool separable_program = false); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteProgram(handle); - OpenGLState::GetCurState().ResetProgram(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -178,20 +131,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenProgramPipelines(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteProgramPipelines(1, &handle); - OpenGLState::GetCurState().ResetPipeline(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -213,20 +156,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenBuffers(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteBuffers(1, &handle); - OpenGLState::GetCurState().ResetBuffer(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -247,19 +180,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - handle = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteSync(handle); - handle = 0; - } + void Release(); GLsync handle = 0; }; @@ -281,20 +205,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenVertexArrays(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteVertexArrays(1, &handle); - OpenGLState::GetCurState().ResetVertexArray(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -316,20 +230,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenFramebuffers(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteFramebuffers(1, &handle); - OpenGLState::GetCurState().ResetFramebuffer(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 9522fd344..a85a7c0c5 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -6,10 +6,10 @@ #include "core/core.h" #include "core/memory.h" #include "video_core/engines/maxwell_3d.h" +#include "video_core/renderer_opengl/gl_rasterizer.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 { @@ -121,12 +121,16 @@ GLint CachedShader::GetUniformLocation(const GLShader::SamplerEntry& sampler) { } GLuint CachedShader::LazyGeometryProgram(OGLProgram& target_program, - const std::string& glsl_topology, + const std::string& glsl_topology, u32 max_vertices, const std::string& debug_name) { if (target_program.handle != 0) { return target_program.handle; } - const std::string source{geometry_programs.code + "layout (" + glsl_topology + ") in;\n"}; + std::string source = "#version 430 core\n"; + source += "layout (" + glsl_topology + ") in;\n"; + source += "#define MAX_VERTEX_INPUT " + std::to_string(max_vertices) + '\n'; + source += geometry_programs.code; + OGLShader shader; shader.Create(source.c_str(), GL_GEOMETRY_SHADER); target_program.Create(true, shader.handle); @@ -135,6 +139,8 @@ GLuint CachedShader::LazyGeometryProgram(OGLProgram& target_program, return target_program.handle; }; +ShaderCacheOpenGL::ShaderCacheOpenGL(RasterizerOpenGL& rasterizer) : RasterizerCache{rasterizer} {} + Shader ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) { const VAddr program_addr{GetShaderAddress(program)}; diff --git a/src/video_core/renderer_opengl/gl_shader_cache.h b/src/video_core/renderer_opengl/gl_shader_cache.h index a210f1731..ffbf21831 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.h +++ b/src/video_core/renderer_opengl/gl_shader_cache.h @@ -16,6 +16,8 @@ namespace OpenGL { class CachedShader; +class RasterizerOpenGL; + using Shader = std::shared_ptr<CachedShader>; using Maxwell = Tegra::Engines::Maxwell3D::Regs; @@ -46,22 +48,23 @@ public: } switch (primitive_mode) { case GL_POINTS: - return LazyGeometryProgram(geometry_programs.points, "points", "ShaderPoints"); + return LazyGeometryProgram(geometry_programs.points, "points", 1, "ShaderPoints"); case GL_LINES: case GL_LINE_STRIP: - return LazyGeometryProgram(geometry_programs.lines, "lines", "ShaderLines"); + return LazyGeometryProgram(geometry_programs.lines, "lines", 2, "ShaderLines"); case GL_LINES_ADJACENCY: case GL_LINE_STRIP_ADJACENCY: - return LazyGeometryProgram(geometry_programs.lines_adjacency, "lines_adjacency", + return LazyGeometryProgram(geometry_programs.lines_adjacency, "lines_adjacency", 4, "ShaderLinesAdjacency"); case GL_TRIANGLES: case GL_TRIANGLE_STRIP: case GL_TRIANGLE_FAN: - return LazyGeometryProgram(geometry_programs.triangles, "triangles", "ShaderTriangles"); + return LazyGeometryProgram(geometry_programs.triangles, "triangles", 3, + "ShaderTriangles"); case GL_TRIANGLES_ADJACENCY: case GL_TRIANGLE_STRIP_ADJACENCY: return LazyGeometryProgram(geometry_programs.triangles_adjacency, "triangles_adjacency", - "ShaderLines"); + 6, "ShaderTrianglesAdjacency"); default: UNREACHABLE_MSG("Unknown primitive mode."); } @@ -76,7 +79,7 @@ public: private: /// Generates a geometry shader or returns one that already exists. GLuint LazyGeometryProgram(OGLProgram& target_program, const std::string& glsl_topology, - const std::string& debug_name); + u32 max_vertices, const std::string& debug_name); VAddr addr; Maxwell::ShaderProgram program_type; @@ -104,6 +107,8 @@ private: class ShaderCacheOpenGL final : public RasterizerCache<Shader> { public: + explicit ShaderCacheOpenGL(RasterizerOpenGL& rasterizer); + /// Gets the current specified shader stage program Shader GetStageProgram(Maxwell::ShaderProgram program); }; diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index d1f6ffe40..5fde22ad4 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -373,6 +373,7 @@ 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."); } } @@ -493,10 +494,10 @@ public: // instruction for now. if (stage == Maxwell3D::Regs::ShaderStage::Geometry) { // TODO(Rodrigo): nouveau sets some attributes after setting emitting a geometry - // shader. These instructions use a dirty register as buffer index. To avoid some - // drivers from complaining for the out of boundary writes, guard them. - const std::string buf_index{"min(" + GetRegisterAsInteger(buf_reg) + ", " + - std::to_string(MAX_GEOMETRY_BUFFERS - 1) + ')'}; + // shader. These instructions use a dirty register as buffer index, to avoid some + // drivers from complaining about out of boundary writes, guard them. + const std::string buf_index{"((" + GetRegisterAsInteger(buf_reg) + ") % " + + std::to_string(MAX_GEOMETRY_BUFFERS) + ')'}; shader.AddLine("amem[" + buf_index + "][" + std::to_string(static_cast<u32>(attribute)) + ']' + GetSwizzle(elem) + " = " + src + ';'); @@ -810,7 +811,11 @@ private: std::optional<Register> vertex = {}) { auto GeometryPass = [&](const std::string& name) { if (stage == Maxwell3D::Regs::ShaderStage::Geometry && vertex) { - return "gs_" + name + '[' + GetRegisterAsInteger(*vertex, 0, false) + ']'; + // TODO(Rodrigo): Guard geometry inputs against out of bound reads. Some games set + // an 0x80000000 index for those and the shader fails to build. Find out why this + // happens and what's its intent. + return "gs_" + name + '[' + GetRegisterAsInteger(*vertex, 0, false) + + " % MAX_VERTEX_INPUT]"; } return name; }; @@ -1525,6 +1530,10 @@ 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: @@ -1535,6 +1544,10 @@ 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: { @@ -1588,6 +1601,10 @@ 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: @@ -1618,6 +1635,10 @@ 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: { @@ -1641,6 +1662,10 @@ 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; } } @@ -1661,6 +1686,10 @@ 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: { @@ -1698,12 +1727,20 @@ 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()); @@ -1723,6 +1760,10 @@ private: 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) @@ -1734,6 +1775,10 @@ 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: { @@ -1770,6 +1815,10 @@ 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: @@ -1831,6 +1880,11 @@ 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: @@ -1846,6 +1900,10 @@ 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: @@ -1877,6 +1935,10 @@ 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: @@ -1892,6 +1954,10 @@ private: } 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: @@ -1906,6 +1972,10 @@ 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: @@ -2107,6 +2177,10 @@ private: 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; } @@ -2212,6 +2286,11 @@ 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: { @@ -2250,6 +2329,11 @@ 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: @@ -2299,6 +2383,10 @@ 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: { @@ -2658,12 +2746,12 @@ private: } case 3: { if (is_array) { - UNIMPLEMENTED_MSG("3-coordinate arrays not fully implemented"); - const std::string x = regs.GetRegisterAsFloat(instr.gpr8); - const std::string y = regs.GetRegisterAsFloat(instr.gpr20); - coord = "vec2 coords = vec2(" + x + ", " + y + ");"; - texture_type = Tegra::Shader::TextureType::Texture2D; - is_array = false; + const std::string index = regs.GetRegisterAsInteger(instr.gpr8); + const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1); + const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2); + const std::string z = regs.GetRegisterAsFloat(instr.gpr20); + coord = + "vec4 coords = vec4(" + x + ", " + y + ", " + z + ", " + index + ");"; } else { const std::string x = regs.GetRegisterAsFloat(instr.gpr8); const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1); @@ -2693,7 +2781,11 @@ private: break; } case Tegra::Shader::TextureProcessMode::LZ: { - texture = "textureLod(" + sampler + ", coords, 0.0)"; + if (depth_compare && is_array) { + texture = "texture(" + sampler + ", coords)"; + } else { + texture = "textureLod(" + sampler + ", coords, 0.0)"; + } break; } case Tegra::Shader::TextureProcessMode::LL: { @@ -3107,6 +3199,11 @@ 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: { @@ -3372,6 +3469,10 @@ 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: { @@ -3381,6 +3482,12 @@ private: 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;"); @@ -3410,6 +3517,11 @@ 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;"); @@ -3467,6 +3579,11 @@ 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; @@ -3507,13 +3624,21 @@ private: } case OpCode::Id::SYNC: { // The SYNC opcode jumps to the address previously set by the SSY opcode - ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always); + 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(); + } EmitPopFromFlowStack(); break; } case OpCode::Id::BRK: { // The BRK opcode jumps to the address previously set by the PBK opcode - ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always); + 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(); + } EmitPopFromFlowStack(); break; } @@ -3543,6 +3668,11 @@ 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: { diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp index 9d17edd63..eea090e52 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.cpp +++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp @@ -82,8 +82,8 @@ void main() { } ProgramResult GenerateGeometryShader(const ShaderSetup& setup) { - std::string out = "#version 430 core\n"; - out += "#extension GL_ARB_separate_shader_objects : enable\n\n"; + // Version is intentionally skipped in shader generation, it's added by the lazy compilation. + std::string out = "#extension GL_ARB_separate_shader_objects : enable\n\n"; out += Decompiler::GetCommonDeclarations(); out += "bool exec_geometry();\n"; diff --git a/src/video_core/renderer_opengl/gl_shader_manager.h b/src/video_core/renderer_opengl/gl_shader_manager.h index 36fe1f04c..2a069cdd8 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.h +++ b/src/video_core/renderer_opengl/gl_shader_manager.h @@ -7,6 +7,7 @@ #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 { diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp index d8a43cc94..98622a058 100644 --- a/src/video_core/renderer_opengl/gl_state.cpp +++ b/src/video_core/renderer_opengl/gl_state.cpp @@ -22,17 +22,15 @@ 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; - - color_mask.red_enabled = GL_TRUE; - color_mask.green_enabled = GL_TRUE; - color_mask.blue_enabled = GL_TRUE; - color_mask.alpha_enabled = GL_TRUE; - + 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; + } stencil.test_enabled = false; auto reset_stencil = [](auto& config) { config.test_func = GL_ALWAYS; @@ -45,19 +43,33 @@ OpenGLState::OpenGLState() { }; reset_stencil(stencil.front); reset_stencil(stencil.back); - - 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; - + 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; logic_op.enabled = false; logic_op.operation = GL_COPY; @@ -73,23 +85,23 @@ 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::Apply() const { +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 { // sRGB if (framebuffer_srgb.enabled != cur_state.framebuffer_srgb.enabled) { if (framebuffer_srgb.enabled) { @@ -100,139 +112,282 @@ void OpenGLState::Apply() const { glDisable(GL_FRAMEBUFFER_SRGB); } } +} + +void OpenGLState::ApplyCulling() const { // Culling - if (cull.enabled != cur_state.cull.enabled) { + const bool cull_changed = cull.enabled != cur_state.cull.enabled; + if (cull_changed) { 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.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.front_face != cur_state.cull.front_face) { - glFrontFace(cull.front_face); +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); + } } +} +void OpenGLState::ApplyDepth() const { // Depth test - if (depth.test_enabled != cur_state.depth.test_enabled) { + const bool depth_test_changed = depth.test_enabled != cur_state.depth.test_enabled; + if (depth_test_changed) { if (depth.test_enabled) { glEnable(GL_DEPTH_TEST); } else { glDisable(GL_DEPTH_TEST); } } - - if (depth.test_func != cur_state.depth.test_func) { + if (depth.test_enabled && + (depth_test_changed || 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); } +} - // 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) { +void OpenGLState::ApplyPrimitiveRestart() const { + const bool primitive_restart_changed = + primitive_restart.enabled != cur_state.primitive_restart.enabled; + if (primitive_restart_changed) { if (primitive_restart.enabled) { glEnable(GL_PRIMITIVE_RESTART); } else { glDisable(GL_PRIMITIVE_RESTART); } } - if (primitive_restart.index != cur_state.primitive_restart.index) { + if (primitive_restart_changed || + (primitive_restart.enabled && + primitive_restart.index != cur_state.primitive_restart.index)) { glPrimitiveRestartIndex(primitive_restart.index); } +} - // 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) { +void OpenGLState::ApplyStencilTest() const { + const bool stencil_test_changed = stencil.test_enabled != cur_state.stencil.test_enabled; + if (stencil_test_changed) { if (stencil.test_enabled) { glEnable(GL_STENCIL_TEST); } else { glDisable(GL_STENCIL_TEST); } } - 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 (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); + } +} + +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 (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 (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); + } } - if (config.write_mask != prev_config.write_mask) { - glStencilMaskSeparate(face, config.write_mask); + } 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(static_cast<GLint>(updated.x), static_cast<GLint>(updated.y), + static_cast<GLsizei>(updated.width), static_cast<GLsizei>(updated.height)); } - }; - config_stencil(GL_FRONT, stencil.front, cur_state.stencil.front); - config_stencil(GL_BACK, stencil.back, cur_state.stencil.back); + 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); + } + } +} - // Blending - if (blend.enabled != cur_state.blend.enabled) { - if (blend.enabled) { - ASSERT(!logic_op.enabled); +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) { 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.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); + if (blend_changed || updated.rgb_equation != current.rgb_equation) { + glBlendEquation(updated.rgb_equation); + } } +} - 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); +void OpenGLState::ApplyTargetBlending(std::size_t 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 (!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.rgb_equation != cur_state.blend.rgb_equation || - blend.a_equation != cur_state.blend.a_equation) { - glBlendEquationSeparate(blend.rgb_equation, blend.a_equation); + 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); + } } +} - // Logic Operation - if (logic_op.enabled != cur_state.logic_op.enabled) { +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) { if (logic_op.enabled) { - ASSERT(!blend.enabled); glEnable(GL_COLOR_LOGIC_OP); } else { glDisable(GL_COLOR_LOGIC_OP); } } - if (logic_op.operation != cur_state.logic_op.operation) { + if (logic_op.enabled && + (logic_op_changed || logic_op.operation != cur_state.logic_op.operation)) { glLogicOp(logic_op.operation); } +} - // Textures +void OpenGLState::ApplyTextures() const { 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]; @@ -251,28 +406,29 @@ void OpenGLState::Apply() const { glTexParameteriv(texture_unit.target, GL_TEXTURE_SWIZZLE_RGBA, mask.data()); } } +} - // 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; +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; } + 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::ApplyFramebufferState() const { // Framebuffer if (draw.read_framebuffer != cur_state.draw.read_framebuffer) { glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer); @@ -280,7 +436,9 @@ void OpenGLState::Apply() const { if (draw.draw_framebuffer != cur_state.draw.draw_framebuffer) { glBindFramebuffer(GL_DRAW_FRAMEBUFFER, draw.draw_framebuffer); } +} +void OpenGLState::ApplyVertexBufferState() const { // Vertex array if (draw.vertex_array != cur_state.draw.vertex_array) { glBindVertexArray(draw.vertex_array); @@ -290,7 +448,11 @@ void OpenGLState::Apply() const { if (draw.vertex_buffer != cur_state.draw.vertex_buffer) { glBindBuffer(GL_ARRAY_BUFFER, draw.vertex_buffer); } +} +void OpenGLState::Apply() const { + ApplyFramebufferState(); + ApplyVertexBufferState(); // Uniform buffer if (draw.uniform_buffer != cur_state.draw.uniform_buffer) { glBindBuffer(GL_UNIFORM_BUFFER, draw.uniform_buffer); @@ -305,27 +467,6 @@ 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]) { @@ -336,12 +477,22 @@ 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; } diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h index 9e2c573b5..e5d1baae6 100644 --- a/src/video_core/renderer_opengl/gl_state.h +++ b/src/video_core/renderer_opengl/gl_state.h @@ -46,11 +46,9 @@ public: } cull; struct { - 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 + bool test_enabled; // GL_DEPTH_TEST + GLenum test_func; // GL_DEPTH_FUNC + GLboolean write_mask; // GL_DEPTH_WRITEMASK } depth; struct { @@ -58,13 +56,14 @@ public: GLuint index; } primitive_restart; // GL_PRIMITIVE_RESTART - struct { + struct ColorMask { GLboolean red_enabled; GLboolean green_enabled; GLboolean blue_enabled; GLboolean alpha_enabled; - } color_mask; // GL_COLOR_WRITEMASK - + }; + std::array<ColorMask, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> + color_mask; // GL_COLOR_WRITEMASK struct { bool test_enabled; // GL_STENCIL_TEST struct { @@ -78,22 +77,28 @@ public: } front, back; } stencil; - struct { + struct Blend { 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 { - GLclampf red; - GLclampf green; - GLclampf blue; - GLclampf alpha; - } color; // GL_BLEND_COLOR - } blend; + struct { + bool enabled; + } independant_blend; + + struct { + GLclampf red; + GLclampf green; + GLclampf blue; + GLclampf alpha; + } blend_color; // GL_BLEND_COLOR struct { bool enabled; // GL_LOGIC_OP_MODE @@ -138,6 +143,16 @@ 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; @@ -147,13 +162,6 @@ public: } scissor; struct { - GLint x; - GLint y; - GLsizei width; - GLsizei height; - } viewport; - - struct { float size; // GL_POINT_SIZE } point; @@ -173,7 +181,12 @@ public: } /// Apply this state as the current OpenGL state void Apply() const; - + /// Apply only the state afecting the framebuffer + void ApplyFramebufferState() const; + /// Apply only the state afecting the vertex buffer + void ApplyVertexBufferState() const; + /// Set the initial OpenGL state + static void ApplyDefaultState(); /// Resets any references to the given resource OpenGLState& UnbindTexture(GLuint handle); OpenGLState& ResetSampler(GLuint handle); @@ -188,6 +201,20 @@ 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(std::size_t target, bool force) const; + void ApplyGlobalBlending() const; + void ApplyBlending() const; + void ApplyLogicOp() const; + void ApplyTextures() const; + void ApplySamplers() const; + void ApplyScissor() const; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.cpp b/src/video_core/renderer_opengl/gl_stream_buffer.cpp index e409228cc..b97b895a4 100644 --- a/src/video_core/renderer_opengl/gl_stream_buffer.cpp +++ b/src/video_core/renderer_opengl/gl_stream_buffer.cpp @@ -6,9 +6,13 @@ #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) @@ -75,6 +79,7 @@ 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); diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index 87d511c38..3ce2cc6d2 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -159,10 +159,8 @@ inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode, } } } - LOG_CRITICAL(Render_OpenGL, "Unimplemented texture filter mode={}", - static_cast<u32>(filter_mode)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented texture filter mode={}", static_cast<u32>(filter_mode)); + return GL_LINEAR; } inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) { @@ -183,9 +181,8 @@ inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) { case Tegra::Texture::WrapMode::MirrorOnceClampToEdge: return GL_MIRROR_CLAMP_TO_EDGE; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode)); + return GL_REPEAT; } inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) { @@ -207,10 +204,9 @@ inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) { case Tegra::Texture::DepthCompareFunc::Always: return GL_ALWAYS; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented texture depth compare function ={}", - static_cast<u32>(func)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented texture depth compare function ={}", + static_cast<u32>(func)); + return GL_GREATER; } inline GLenum BlendEquation(Maxwell::Blend::Equation equation) { @@ -226,9 +222,8 @@ inline GLenum BlendEquation(Maxwell::Blend::Equation equation) { case Maxwell::Blend::Equation::Max: return GL_MAX; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented blend equation={}", static_cast<u32>(equation)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented blend equation={}", static_cast<u32>(equation)); + return GL_FUNC_ADD; } inline GLenum BlendFunc(Maxwell::Blend::Factor factor) { @@ -291,9 +286,8 @@ inline GLenum BlendFunc(Maxwell::Blend::Factor factor) { case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: return GL_ONE_MINUS_CONSTANT_ALPHA; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented blend factor={}", static_cast<u32>(factor)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented blend factor={}", static_cast<u32>(factor)); + return GL_ZERO; } inline GLenum SwizzleSource(Tegra::Texture::SwizzleSource source) { @@ -312,9 +306,8 @@ inline GLenum SwizzleSource(Tegra::Texture::SwizzleSource source) { case Tegra::Texture::SwizzleSource::OneFloat: return GL_ONE; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented swizzle source={}", static_cast<u32>(source)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented swizzle source={}", static_cast<u32>(source)); + return GL_ZERO; } inline GLenum ComparisonOp(Maxwell::ComparisonOp comparison) { @@ -344,33 +337,39 @@ inline GLenum ComparisonOp(Maxwell::ComparisonOp comparison) { case Maxwell::ComparisonOp::AlwaysOld: return GL_ALWAYS; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented comparison op={}", static_cast<u32>(comparison)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented comparison op={}", static_cast<u32>(comparison)); + return GL_ALWAYS; } inline GLenum StencilOp(Maxwell::StencilOp stencil) { switch (stencil) { case Maxwell::StencilOp::Keep: + case Maxwell::StencilOp::KeepOGL: return GL_KEEP; case Maxwell::StencilOp::Zero: + case Maxwell::StencilOp::ZeroOGL: return GL_ZERO; case Maxwell::StencilOp::Replace: + case Maxwell::StencilOp::ReplaceOGL: return GL_REPLACE; case Maxwell::StencilOp::Incr: + case Maxwell::StencilOp::IncrOGL: return GL_INCR; case Maxwell::StencilOp::Decr: + case Maxwell::StencilOp::DecrOGL: return GL_DECR; case Maxwell::StencilOp::Invert: + case Maxwell::StencilOp::InvertOGL: return GL_INVERT; case Maxwell::StencilOp::IncrWrap: + case Maxwell::StencilOp::IncrWrapOGL: return GL_INCR_WRAP; case Maxwell::StencilOp::DecrWrap: + case Maxwell::StencilOp::DecrWrapOGL: return GL_DECR_WRAP; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented stencil op={}", static_cast<u32>(stencil)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented stencil op={}", static_cast<u32>(stencil)); + return GL_KEEP; } inline GLenum FrontFace(Maxwell::Cull::FrontFace front_face) { @@ -380,9 +379,8 @@ inline GLenum FrontFace(Maxwell::Cull::FrontFace front_face) { case Maxwell::Cull::FrontFace::CounterClockWise: return GL_CCW; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented front face cull={}", static_cast<u32>(front_face)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented front face cull={}", static_cast<u32>(front_face)); + return GL_CCW; } inline GLenum CullFace(Maxwell::Cull::CullFace cull_face) { @@ -394,9 +392,8 @@ inline GLenum CullFace(Maxwell::Cull::CullFace cull_face) { case Maxwell::Cull::CullFace::FrontAndBack: return GL_FRONT_AND_BACK; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented cull face={}", static_cast<u32>(cull_face)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented cull face={}", static_cast<u32>(cull_face)); + return GL_BACK; } inline GLenum LogicOp(Maxwell::LogicOperation operation) { @@ -434,9 +431,8 @@ inline GLenum LogicOp(Maxwell::LogicOperation operation) { case Maxwell::LogicOperation::Set: return GL_SET; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented logic operation={}", static_cast<u32>(operation)); - UNREACHABLE(); - return {}; + LOG_ERROR(Render_OpenGL, "Unimplemented logic operation={}", static_cast<u32>(operation)); + return GL_COPY; } } // namespace MaxwellToGL diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp new file mode 100644 index 000000000..9582dd2ca --- /dev/null +++ b/src/video_core/surface.cpp @@ -0,0 +1,490 @@ +// 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::ASTC_2D_10X8: + return is_srgb ? PixelFormat::ASTC_2D_10X8_SRGB : PixelFormat::ASTC_2D_10X8; + 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: + case PixelFormat::ASTC_2D_10X8: + case PixelFormat::ASTC_2D_10X8_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 diff --git a/src/video_core/surface.h b/src/video_core/surface.h new file mode 100644 index 000000000..0dd3eb2e4 --- /dev/null +++ b/src/video_core/surface.h @@ -0,0 +1,477 @@ +// 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, + ASTC_2D_10X8 = 59, + ASTC_2D_10X8_SRGB = 60, + + MaxColorFormat, + + // Depth formats + Z32F = 61, + Z16 = 62, + + MaxDepthFormat, + + // DepthStencil formats + Z24S8 = 63, + S8Z24 = 64, + Z32FS8 = 65, + + 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 + 4, // ASTC_2D_10X8 + 4, // ASTC_2D_10X8_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 + 10, // ASTC_2D_10X8 + 10, // ASTC_2D_10X8_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 + 8, // ASTC_2D_10X8 + 8, // ASTC_2D_10X8_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 + 128, // ASTC_2D_10X8 + 128, // ASTC_2D_10X8_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 diff --git a/src/video_core/textures/astc.cpp b/src/video_core/textures/astc.cpp index b1feacae9..bc50a4876 100644 --- a/src/video_core/textures/astc.cpp +++ b/src/video_core/textures/astc.cpp @@ -1598,27 +1598,29 @@ 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 block_width, uint32_t block_height) { + uint32_t depth, uint32_t block_width, uint32_t block_height) { uint32_t blockIdx = 0; - 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) { + 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) { - 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); - } + 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++; + blockIdx++; + } } } diff --git a/src/video_core/textures/astc.h b/src/video_core/textures/astc.h index f0d7c0e56..d419dd025 100644 --- a/src/video_core/textures/astc.h +++ b/src/video_core/textures/astc.h @@ -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 block_width, uint32_t block_height); + uint32_t depth, uint32_t block_width, uint32_t block_height); } // namespace Tegra::Texture::ASTC diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp index 550ca856c..a9d134d14 100644 --- a/src/video_core/textures/decoders.cpp +++ b/src/video_core/textures/decoders.cpp @@ -202,6 +202,8 @@ u32 BytesPerPixel(TextureFormat format) { case TextureFormat::ASTC_2D_5X4: case TextureFormat::ASTC_2D_8X8: case TextureFormat::ASTC_2D_8X5: + case TextureFormat::ASTC_2D_10X8: + case TextureFormat::ASTC_2D_5X5: case TextureFormat::A8R8G8B8: case TextureFormat::A2B10G10R10: case TextureFormat::BF10GF11RF11: @@ -227,12 +229,14 @@ u32 BytesPerPixel(TextureFormat format) { } } -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> 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> unswizzled_data(width * height * depth * bytes_per_pixel); - CopySwizzledData(width / tile_size, height / tile_size, depth, bytes_per_pixel, bytes_per_pixel, - Memory::GetPointer(address), unswizzled_data.data(), true, block_height, - block_depth); + 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); return unswizzled_data; } @@ -292,6 +296,8 @@ std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat case TextureFormat::BC6H_SF16: case TextureFormat::ASTC_2D_4X4: case TextureFormat::ASTC_2D_8X8: + case TextureFormat::ASTC_2D_5X5: + case TextureFormat::ASTC_2D_10X8: case TextureFormat::A8R8G8B8: case TextureFormat::A2B10G10R10: case TextureFormat::A1B5G5R5: diff --git a/src/video_core/textures/decoders.h b/src/video_core/textures/decoders.h index b390219e4..ba065510b 100644 --- a/src/video_core/textures/decoders.h +++ b/src/video_core/textures/decoders.h @@ -19,8 +19,8 @@ inline std::size_t GetGOBSize() { /** * Unswizzles a swizzled texture without changing its format. */ -std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width, - u32 height, u32 depth, +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 = TICEntry::DefaultBlockHeight, u32 block_depth = TICEntry::DefaultBlockHeight); diff --git a/src/video_core/textures/texture.h b/src/video_core/textures/texture.h index d12d2ecb8..e199d019a 100644 --- a/src/video_core/textures/texture.h +++ b/src/video_core/textures/texture.h @@ -168,20 +168,29 @@ struct TICEntry { // High 16 bits of the pitch value BitField<0, 16, u32> pitch_high; - + BitField<26, 1, u32> use_header_opt_control; + BitField<27, 1, u32> depth_texture; BitField<28, 4, u32> max_mip_level; }; union { BitField<0, 16, u32> width_minus_1; BitField<22, 1, u32> srgb_conversion; BitField<23, 4, TextureType> texture_type; + BitField<29, 3, u32> border_size; }; union { BitField<0, 16, u32> height_minus_1; BitField<16, 15, u32> depth_minus_1; }; + union { + BitField<6, 13, u32> mip_lod_bias; + BitField<27, 3, u32> max_anisotropy; + }; - INSERT_PADDING_BYTES(8); + union { + BitField<0, 4, u32> res_min_mip_level; + BitField<4, 4, u32> res_max_mip_level; + }; GPUVAddr Address() const { return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | address_low); diff --git a/src/web_service/telemetry_json.cpp b/src/web_service/telemetry_json.cpp index 0a8f2bd9e..9156ce802 100644 --- a/src/web_service/telemetry_json.cpp +++ b/src/web_service/telemetry_json.cpp @@ -102,16 +102,27 @@ 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}, username{impl->username}, token{impl->token}, content]() { - Client{host, username, token}.PostJson("/telemetry", content, true); - }); + 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; } } // namespace WebService diff --git a/src/web_service/telemetry_json.h b/src/web_service/telemetry_json.h index 93371414a..dfd202829 100644 --- a/src/web_service/telemetry_json.h +++ b/src/web_service/telemetry_json.h @@ -35,6 +35,7 @@ public: void Visit(const Telemetry::Field<std::chrono::microseconds>& field) override; void Complete() override; + bool SubmitTestcase() override; private: struct Impl; diff --git a/src/yuzu/CMakeLists.txt b/src/yuzu/CMakeLists.txt index 9379d9110..f9ca2948e 100644 --- a/src/yuzu/CMakeLists.txt +++ b/src/yuzu/CMakeLists.txt @@ -56,6 +56,8 @@ 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 diff --git a/src/yuzu/compatdb.cpp b/src/yuzu/compatdb.cpp index 91e754274..5f0896f84 100644 --- a/src/yuzu/compatdb.cpp +++ b/src/yuzu/compatdb.cpp @@ -5,6 +5,7 @@ #include <QButtonGroup> #include <QMessageBox> #include <QPushButton> +#include <QtConcurrent/qtconcurrentrun.h> #include "common/logging/log.h" #include "common/telemetry.h" #include "core/core.h" @@ -23,6 +24,8 @@ 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; @@ -48,18 +51,38 @@ void CompatDB::Submit() { } break; case CompatDBPage::Final: + back(); LOG_DEBUG(Frontend, "Compatibility Rating: {}", compatibility->checkedId()); Core::Telemetry().AddField(Telemetry::FieldType::UserFeedback, "Compatibility", compatibility->checkedId()); - // older versions of QT don't support the "NoCancelButtonOnLastPage" option, this is a - // workaround + + button(NextButton)->setEnabled(false); + button(NextButton)->setText(tr("Submitting")); 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); } diff --git a/src/yuzu/compatdb.h b/src/yuzu/compatdb.h index ca0dd11d6..5381f67f7 100644 --- a/src/yuzu/compatdb.h +++ b/src/yuzu/compatdb.h @@ -5,6 +5,7 @@ #pragma once #include <memory> +#include <QFutureWatcher> #include <QWizard> namespace Ui { @@ -19,8 +20,11 @@ public: ~CompatDB(); private: + QFutureWatcher<bool> testcase_watcher; + std::unique_ptr<Ui::CompatDB> ui; void Submit(); + void OnTestcaseSubmitted(); void EnableNext(); }; diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp index 526cf6c10..be69fb831 100644 --- a/src/yuzu/configuration/config.cpp +++ b/src/yuzu/configuration/config.cpp @@ -134,6 +134,14 @@ void Config::ReadValues() { Service::Account::MAX_USERS - 1); Settings::values.language_index = qt_config->value("language_index", 1).toInt(); + + const auto enabled = qt_config->value("rng_seed_enabled", false).toBool(); + if (enabled) { + Settings::values.rng_seed = qt_config->value("rng_seed", 0).toULongLong(); + } else { + Settings::values.rng_seed = std::nullopt; + } + qt_config->endGroup(); qt_config->beginGroup("Miscellaneous"); @@ -145,6 +153,7 @@ void Config::ReadValues() { Settings::values.use_gdbstub = qt_config->value("use_gdbstub", false).toBool(); Settings::values.gdbstub_port = qt_config->value("gdbstub_port", 24689).toInt(); Settings::values.program_args = qt_config->value("program_args", "").toString().toStdString(); + Settings::values.dump_nso = qt_config->value("dump_nso", false).toBool(); qt_config->endGroup(); qt_config->beginGroup("WebService"); @@ -273,6 +282,10 @@ void Config::SaveValues() { qt_config->setValue("current_user", Settings::values.current_user); qt_config->setValue("language_index", Settings::values.language_index); + + qt_config->setValue("rng_seed_enabled", Settings::values.rng_seed.has_value()); + qt_config->setValue("rng_seed", Settings::values.rng_seed.value_or(0)); + qt_config->endGroup(); qt_config->beginGroup("Miscellaneous"); @@ -284,6 +297,7 @@ void Config::SaveValues() { qt_config->setValue("use_gdbstub", Settings::values.use_gdbstub); qt_config->setValue("gdbstub_port", Settings::values.gdbstub_port); qt_config->setValue("program_args", QString::fromStdString(Settings::values.program_args)); + qt_config->setValue("dump_nso", Settings::values.dump_nso); qt_config->endGroup(); qt_config->beginGroup("WebService"); diff --git a/src/yuzu/configuration/configure_debug.cpp b/src/yuzu/configuration/configure_debug.cpp index 9e765fc93..fd5876b41 100644 --- a/src/yuzu/configuration/configure_debug.cpp +++ b/src/yuzu/configuration/configure_debug.cpp @@ -34,6 +34,7 @@ void ConfigureDebug::setConfiguration() { ui->toggle_console->setChecked(UISettings::values.show_console); ui->log_filter_edit->setText(QString::fromStdString(Settings::values.log_filter)); ui->homebrew_args_edit->setText(QString::fromStdString(Settings::values.program_args)); + ui->dump_decompressed_nso->setChecked(Settings::values.dump_nso); } void ConfigureDebug::applyConfiguration() { @@ -42,6 +43,7 @@ void ConfigureDebug::applyConfiguration() { UISettings::values.show_console = ui->toggle_console->isChecked(); Settings::values.log_filter = ui->log_filter_edit->text().toStdString(); Settings::values.program_args = ui->homebrew_args_edit->text().toStdString(); + Settings::values.dump_nso = ui->dump_decompressed_nso->isChecked(); Debugger::ToggleConsole(); Log::Filter filter; filter.ParseFilterString(Settings::values.log_filter); diff --git a/src/yuzu/configuration/configure_debug.ui b/src/yuzu/configuration/configure_debug.ui index ff4987604..9c5b702f8 100644 --- a/src/yuzu/configuration/configure_debug.ui +++ b/src/yuzu/configuration/configure_debug.ui @@ -7,7 +7,7 @@ <x>0</x> <y>0</y> <width>400</width> - <height>300</height> + <height>357</height> </rect> </property> <property name="windowTitle"> @@ -130,6 +130,25 @@ </widget> </item> <item> + <widget class="QGroupBox" name="groupBox_4"> + <property name="title"> + <string>Dump</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout_4"> + <item> + <widget class="QCheckBox" name="dump_decompressed_nso"> + <property name="whatsThis"> + <string>When checked, any NSO yuzu tries to load or patch will be copied decompressed to the yuzu/dump directory.</string> + </property> + <property name="text"> + <string>Dump Decompressed NSOs</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> <spacer name="verticalSpacer"> <property name="orientation"> <enum>Qt::Vertical</enum> diff --git a/src/yuzu/configuration/configure_general.cpp b/src/yuzu/configuration/configure_general.cpp index 537d6e576..314f51203 100644 --- a/src/yuzu/configuration/configure_general.cpp +++ b/src/yuzu/configuration/configure_general.cpp @@ -3,6 +3,10 @@ // Refer to the license.txt file included. #include "core/core.h" +#include "core/hle/service/am/am.h" +#include "core/hle/service/am/applet_ae.h" +#include "core/hle/service/am/applet_oe.h" +#include "core/hle/service/sm/sm.h" #include "core/settings.h" #include "ui_configure_general.h" #include "yuzu/configuration/configure_general.h" @@ -20,7 +24,6 @@ ConfigureGeneral::ConfigureGeneral(QWidget* parent) this->setConfiguration(); ui->use_cpu_jit->setEnabled(!Core::System::GetInstance().IsPoweredOn()); - ui->use_docked_mode->setEnabled(!Core::System::GetInstance().IsPoweredOn()); } ConfigureGeneral::~ConfigureGeneral() = default; @@ -38,6 +41,33 @@ void ConfigureGeneral::PopulateHotkeyList(const HotkeyRegistry& registry) { ui->widget->Populate(registry); } +void ConfigureGeneral::OnDockedModeChanged(bool last_state, bool new_state) { + if (last_state == new_state) { + return; + } + + Core::System& system{Core::System::GetInstance()}; + if (!system.IsPoweredOn()) { + return; + } + Service::SM::ServiceManager& sm = system.ServiceManager(); + + // Message queue is shared between these services, we just need to signal an operation + // change to one and it will handle both automatically + auto applet_oe = sm.GetService<Service::AM::AppletOE>("appletOE"); + auto applet_ae = sm.GetService<Service::AM::AppletAE>("appletAE"); + bool has_signalled = false; + + if (applet_oe != nullptr) { + applet_oe->GetMessageQueue()->OperationModeChanged(); + has_signalled = true; + } + + if (applet_ae != nullptr && !has_signalled) { + applet_ae->GetMessageQueue()->OperationModeChanged(); + } +} + void ConfigureGeneral::applyConfiguration() { UISettings::values.gamedir_deepscan = ui->toggle_deepscan->isChecked(); UISettings::values.confirm_before_closing = ui->toggle_check_exit->isChecked(); @@ -45,6 +75,9 @@ void ConfigureGeneral::applyConfiguration() { ui->theme_combobox->itemData(ui->theme_combobox->currentIndex()).toString(); Settings::values.use_cpu_jit = ui->use_cpu_jit->isChecked(); + const bool pre_docked_mode = Settings::values.use_docked_mode; Settings::values.use_docked_mode = ui->use_docked_mode->isChecked(); + OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode); + Settings::values.enable_nfc = ui->enable_nfc->isChecked(); } diff --git a/src/yuzu/configuration/configure_general.h b/src/yuzu/configuration/configure_general.h index 4770034cc..2210d48da 100644 --- a/src/yuzu/configuration/configure_general.h +++ b/src/yuzu/configuration/configure_general.h @@ -25,6 +25,7 @@ public: private: void setConfiguration(); + void OnDockedModeChanged(bool last_state, bool new_state); std::unique_ptr<Ui::ConfigureGeneral> ui; }; diff --git a/src/yuzu/configuration/configure_system.cpp b/src/yuzu/configuration/configure_system.cpp index 1b8aa7de2..ab5d46492 100644 --- a/src/yuzu/configuration/configure_system.cpp +++ b/src/yuzu/configuration/configure_system.cpp @@ -6,20 +6,20 @@ #include <QFileDialog> #include <QGraphicsItem> #include <QGraphicsScene> -#include <QInputDialog> +#include <QHeaderView> #include <QMessageBox> #include <QStandardItemModel> #include <QTreeView> #include <QVBoxLayout> -#include "common/common_paths.h" -#include "common/logging/backend.h" +#include "common/assert.h" +#include "common/file_util.h" #include "common/string_util.h" #include "core/core.h" #include "core/hle/service/acc/profile_manager.h" #include "core/settings.h" #include "ui_configure_system.h" #include "yuzu/configuration/configure_system.h" -#include "yuzu/main.h" +#include "yuzu/util/limitable_input_dialog.h" namespace { constexpr std::array<int, 12> days_in_month = {{ @@ -83,6 +83,12 @@ QPixmap GetIcon(Service::Account::UUID uuid) { return icon.scaled(64, 64, Qt::IgnoreAspectRatio, Qt::SmoothTransformation); } + +QString GetProfileUsernameFromUser(QWidget* parent, const QString& description_text) { + return LimitableInputDialog::GetText(parent, ConfigureSystem::tr("Enter Username"), + description_text, 1, + static_cast<int>(Service::Account::profile_username_size)); +} } // Anonymous namespace ConfigureSystem::ConfigureSystem(QWidget* parent) @@ -131,6 +137,12 @@ ConfigureSystem::ConfigureSystem(QWidget* parent) connect(ui->pm_remove, &QPushButton::pressed, this, &ConfigureSystem::DeleteUser); connect(ui->pm_set_image, &QPushButton::pressed, this, &ConfigureSystem::SetUserImage); + connect(ui->rng_seed_checkbox, &QCheckBox::stateChanged, this, [this](bool checked) { + ui->rng_seed_edit->setEnabled(checked); + if (!checked) + ui->rng_seed_edit->setText(QStringLiteral("00000000")); + }); + scene = new QGraphicsScene; ui->current_user_icon->setScene(scene); @@ -149,6 +161,13 @@ void ConfigureSystem::setConfiguration() { PopulateUserList(); UpdateCurrentUser(); + + ui->rng_seed_checkbox->setChecked(Settings::values.rng_seed.has_value()); + ui->rng_seed_edit->setEnabled(Settings::values.rng_seed.has_value()); + + const auto rng_seed = + QString("%1").arg(Settings::values.rng_seed.value_or(0), 8, 16, QLatin1Char{'0'}).toUpper(); + ui->rng_seed_edit->setText(rng_seed); } void ConfigureSystem::PopulateUserList() { @@ -189,6 +208,12 @@ void ConfigureSystem::applyConfiguration() { return; Settings::values.language_index = ui->combo_language->currentIndex(); + + if (ui->rng_seed_checkbox->isChecked()) + Settings::values.rng_seed = ui->rng_seed_edit->text().toULongLong(nullptr, 16); + else + Settings::values.rng_seed = std::nullopt; + Settings::Apply(); } @@ -234,7 +259,7 @@ void ConfigureSystem::RefreshConsoleID() { void ConfigureSystem::SelectUser(const QModelIndex& index) { Settings::values.current_user = - std::clamp<std::size_t>(index.row(), 0, profile_manager->GetUserCount() - 1); + std::clamp<s32>(index.row(), 0, static_cast<s32>(profile_manager->GetUserCount() - 1)); UpdateCurrentUser(); @@ -244,15 +269,13 @@ void ConfigureSystem::SelectUser(const QModelIndex& index) { } void ConfigureSystem::AddUser() { - const auto uuid = Service::Account::UUID::Generate(); - - bool ok = false; const auto username = - QInputDialog::getText(this, tr("Enter Username"), tr("Enter a username for the new user:"), - QLineEdit::Normal, QString(), &ok); - if (!ok) + GetProfileUsernameFromUser(this, tr("Enter a username for the new user:")); + if (username.isEmpty()) { return; + } + const auto uuid = Service::Account::UUID::Generate(); profile_manager->CreateNewUser(uuid, username.toStdString()); item_model->appendRow(new QStandardItem{GetIcon(uuid), FormatUserEntryText(username, uuid)}); @@ -267,23 +290,14 @@ void ConfigureSystem::RenameUser() { if (!profile_manager->GetProfileBase(*uuid, profile)) return; - bool ok = false; - const auto old_username = GetAccountUsername(*profile_manager, *uuid); - const auto new_username = - QInputDialog::getText(this, tr("Enter Username"), tr("Enter a new username:"), - QLineEdit::Normal, old_username, &ok); - - if (!ok) + const auto new_username = GetProfileUsernameFromUser(this, tr("Enter a new username:")); + if (new_username.isEmpty()) { return; + } - std::fill(profile.username.begin(), profile.username.end(), '\0'); const auto username_std = new_username.toStdString(); - if (username_std.size() > profile.username.size()) { - std::copy_n(username_std.begin(), std::min(profile.username.size(), username_std.size()), - profile.username.begin()); - } else { - std::copy(username_std.begin(), username_std.end(), profile.username.begin()); - } + std::fill(profile.username.begin(), profile.username.end(), '\0'); + std::copy(username_std.begin(), username_std.end(), profile.username.begin()); profile_manager->SetProfileBase(*uuid, profile); diff --git a/src/yuzu/configuration/configure_system.ui b/src/yuzu/configuration/configure_system.ui index 020b32a37..a91580893 100644 --- a/src/yuzu/configuration/configure_system.ui +++ b/src/yuzu/configuration/configure_system.ui @@ -6,7 +6,7 @@ <rect> <x>0</x> <y>0</y> - <width>360</width> + <width>366</width> <height>483</height> </rect> </property> @@ -22,98 +22,6 @@ <string>System Settings</string> </property> <layout class="QGridLayout" name="gridLayout"> - <item row="1" column="0"> - <widget class="QLabel" name="label_language"> - <property name="text"> - <string>Language</string> - </property> - </widget> - </item> - <item row="0" column="0"> - <widget class="QLabel" name="label_birthday"> - <property name="text"> - <string>Birthday</string> - </property> - </widget> - </item> - <item row="3" column="0"> - <widget class="QLabel" name="label_console_id"> - <property name="text"> - <string>Console ID:</string> - </property> - </widget> - </item> - <item row="0" column="1"> - <layout class="QHBoxLayout" name="horizontalLayout_birthday2"> - <item> - <widget class="QComboBox" name="combo_birthmonth"> - <item> - <property name="text"> - <string>January</string> - </property> - </item> - <item> - <property name="text"> - <string>February</string> - </property> - </item> - <item> - <property name="text"> - <string>March</string> - </property> - </item> - <item> - <property name="text"> - <string>April</string> - </property> - </item> - <item> - <property name="text"> - <string>May</string> - </property> - </item> - <item> - <property name="text"> - <string>June</string> - </property> - </item> - <item> - <property name="text"> - <string>July</string> - </property> - </item> - <item> - <property name="text"> - <string>August</string> - </property> - </item> - <item> - <property name="text"> - <string>September</string> - </property> - </item> - <item> - <property name="text"> - <string>October</string> - </property> - </item> - <item> - <property name="text"> - <string>November</string> - </property> - </item> - <item> - <property name="text"> - <string>December</string> - </property> - </item> - </widget> - </item> - <item> - <widget class="QComboBox" name="combo_birthday"/> - </item> - </layout> - </item> <item row="1" column="1"> <widget class="QComboBox" name="combo_language"> <property name="toolTip"> @@ -206,6 +114,13 @@ </item> </widget> </item> + <item row="3" column="0"> + <widget class="QLabel" name="label_console_id"> + <property name="text"> + <string>Console ID:</string> + </property> + </widget> + </item> <item row="2" column="0"> <widget class="QLabel" name="label_sound"> <property name="text"> @@ -213,6 +128,100 @@ </property> </widget> </item> + <item row="0" column="0"> + <widget class="QLabel" name="label_birthday"> + <property name="text"> + <string>Birthday</string> + </property> + </widget> + </item> + <item row="0" column="1"> + <layout class="QHBoxLayout" name="horizontalLayout_birthday2"> + <item> + <widget class="QComboBox" name="combo_birthmonth"> + <item> + <property name="text"> + <string>January</string> + </property> + </item> + <item> + <property name="text"> + <string>February</string> + </property> + </item> + <item> + <property name="text"> + <string>March</string> + </property> + </item> + <item> + <property name="text"> + <string>April</string> + </property> + </item> + <item> + <property name="text"> + <string>May</string> + </property> + </item> + <item> + <property name="text"> + <string>June</string> + </property> + </item> + <item> + <property name="text"> + <string>July</string> + </property> + </item> + <item> + <property name="text"> + <string>August</string> + </property> + </item> + <item> + <property name="text"> + <string>September</string> + </property> + </item> + <item> + <property name="text"> + <string>October</string> + </property> + </item> + <item> + <property name="text"> + <string>November</string> + </property> + </item> + <item> + <property name="text"> + <string>December</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="combo_birthday"/> + </item> + </layout> + </item> + <item row="3" column="1"> + <widget class="QPushButton" name="button_regenerate_console_id"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="layoutDirection"> + <enum>Qt::RightToLeft</enum> + </property> + <property name="text"> + <string>Regenerate</string> + </property> + </widget> + </item> <item row="2" column="1"> <widget class="QComboBox" name="combo_sound"> <item> @@ -232,19 +241,38 @@ </item> </widget> </item> - <item row="3" column="1"> - <widget class="QPushButton" name="button_regenerate_console_id"> + <item row="1" column="0"> + <widget class="QLabel" name="label_language"> + <property name="text"> + <string>Language</string> + </property> + </widget> + </item> + <item row="4" column="0"> + <widget class="QCheckBox" name="rng_seed_checkbox"> + <property name="text"> + <string>RNG Seed</string> + </property> + </widget> + </item> + <item row="4" column="1"> + <widget class="QLineEdit" name="rng_seed_edit"> <property name="sizePolicy"> - <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> <horstretch>0</horstretch> <verstretch>0</verstretch> </sizepolicy> </property> - <property name="layoutDirection"> - <enum>Qt::RightToLeft</enum> + <property name="font"> + <font> + <family>Lucida Console</family> + </font> </property> - <property name="text"> - <string>Regenerate</string> + <property name="inputMask"> + <string notr="true">HHHHHHHH</string> + </property> + <property name="maxLength"> + <number>8</number> </property> </widget> </item> diff --git a/src/yuzu/debugger/graphics/graphics_surface.cpp b/src/yuzu/debugger/graphics/graphics_surface.cpp index 0adbab27d..707747422 100644 --- a/src/yuzu/debugger/graphics/graphics_surface.cpp +++ b/src/yuzu/debugger/graphics/graphics_surface.cpp @@ -386,9 +386,9 @@ void GraphicsSurfaceWidget::OnUpdate() { // TODO(bunnei): Will not work with BCn formats that swizzle 4x4 tiles. // Needs to be fixed if we plan to use this feature more, otherwise we may remove it. - auto unswizzled_data = - Tegra::Texture::UnswizzleTexture(*address, 1, Tegra::Texture::BytesPerPixel(surface_format), - surface_width, surface_height, 1U); + auto unswizzled_data = Tegra::Texture::UnswizzleTexture( + *address, 1, 1, Tegra::Texture::BytesPerPixel(surface_format), surface_width, + surface_height, 1U); auto texture_data = Tegra::Texture::DecodeTexture(unswizzled_data, surface_format, surface_width, surface_height); diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index c5a56cbfd..131ad19de 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -142,6 +142,9 @@ static void InitializeLogging() { const std::string& log_dir = FileUtil::GetUserPath(FileUtil::UserPath::LogDir); FileUtil::CreateFullPath(log_dir); Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir + LOG_FILE)); +#ifdef _WIN32 + Log::AddBackend(std::make_unique<Log::DebuggerBackend>()); +#endif } GMainWindow::GMainWindow() @@ -454,6 +457,7 @@ void GMainWindow::ConnectMenuEvents() { connect(ui.action_Fullscreen, &QAction::triggered, this, &GMainWindow::ToggleFullscreen); // Help + connect(ui.action_Open_yuzu_Folder, &QAction::triggered, this, &GMainWindow::OnOpenYuzuFolder); connect(ui.action_Rederive, &QAction::triggered, this, std::bind(&GMainWindow::OnReinitializeKeys, this, ReinitializeKeyBehavior::Warning)); connect(ui.action_About, &QAction::triggered, this, &GMainWindow::OnAbout); @@ -929,7 +933,8 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa const auto full = res == "Full"; const auto entry_size = CalculateRomFSEntrySize(extracted, full); - QProgressDialog progress(tr("Extracting RomFS..."), tr("Cancel"), 0, entry_size, this); + QProgressDialog progress(tr("Extracting RomFS..."), tr("Cancel"), 0, + static_cast<s32>(entry_size), this); progress.setWindowModality(Qt::WindowModal); progress.setMinimumDuration(100); @@ -1374,6 +1379,11 @@ void GMainWindow::OnLoadAmiibo() { } } +void GMainWindow::OnOpenYuzuFolder() { + QDesktopServices::openUrl(QUrl::fromLocalFile( + QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::UserDir)))); +} + void GMainWindow::OnAbout() { AboutDialog aboutDialog(this); aboutDialog.exec(); @@ -1532,7 +1542,7 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) { "derivation. It will be attempted but may not complete.<br><br>") + errors + tr("<br><br>You can get all of these and dump all of your games easily by " - "following <a href='https://yuzu-emu.org/help/quickstart/quickstart/'>the " + "following <a href='https://yuzu-emu.org/help/quickstart/'>the " "quickstart guide</a>. Alternatively, you can use another method of dumping " "to obtain all of your keys.")); } @@ -1612,7 +1622,7 @@ void GMainWindow::closeEvent(QCloseEvent* event) { return; } - if (ui.action_Fullscreen->isChecked()) { + if (!ui.action_Fullscreen->isChecked()) { UISettings::values.geometry = saveGeometry(); UISettings::values.renderwindow_geometry = render_window->saveGeometry(); } diff --git a/src/yuzu/main.h b/src/yuzu/main.h index af637d89e..929250e8c 100644 --- a/src/yuzu/main.h +++ b/src/yuzu/main.h @@ -167,6 +167,7 @@ private slots: void OnMenuRecentFile(); void OnConfigure(); void OnLoadAmiibo(); + void OnOpenYuzuFolder(); void OnAbout(); void OnToggleFilterBar(); void OnDisplayTitleBars(bool); diff --git a/src/yuzu/main.ui b/src/yuzu/main.ui index 48d099591..75e96387f 100644 --- a/src/yuzu/main.ui +++ b/src/yuzu/main.ui @@ -70,6 +70,8 @@ <addaction name="separator"/> <addaction name="action_Load_Amiibo"/> <addaction name="separator"/> + <addaction name="action_Open_yuzu_Folder"/> + <addaction name="separator"/> <addaction name="action_Exit"/> </widget> <widget class="QMenu" name="menu_Emulation"> @@ -277,6 +279,11 @@ <bool>false</bool> </property> </action> + <action name="action_Open_yuzu_Folder"> + <property name="text"> + <string>Open yuzu Folder</string> + </property> + </action> </widget> <resources/> <connections/> diff --git a/src/yuzu/util/limitable_input_dialog.cpp b/src/yuzu/util/limitable_input_dialog.cpp new file mode 100644 index 000000000..edd78e579 --- /dev/null +++ b/src/yuzu/util/limitable_input_dialog.cpp @@ -0,0 +1,59 @@ +// Copyright 2018 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QDialogButtonBox> +#include <QLabel> +#include <QLineEdit> +#include <QPushButton> +#include <QVBoxLayout> +#include "yuzu/util/limitable_input_dialog.h" + +LimitableInputDialog::LimitableInputDialog(QWidget* parent) : QDialog{parent} { + CreateUI(); + ConnectEvents(); +} + +LimitableInputDialog::~LimitableInputDialog() = default; + +void LimitableInputDialog::CreateUI() { + setWindowFlags(windowFlags() & ~Qt::WindowContextHelpButtonHint); + + text_label = new QLabel(this); + text_entry = new QLineEdit(this); + buttons = new QDialogButtonBox(QDialogButtonBox::Ok | QDialogButtonBox::Cancel, this); + + auto* const layout = new QVBoxLayout; + layout->addWidget(text_label); + layout->addWidget(text_entry); + layout->addWidget(buttons); + + setLayout(layout); +} + +void LimitableInputDialog::ConnectEvents() { + connect(buttons, &QDialogButtonBox::accepted, this, &QDialog::accept); + connect(buttons, &QDialogButtonBox::rejected, this, &QDialog::reject); +} + +QString LimitableInputDialog::GetText(QWidget* parent, const QString& title, const QString& text, + int min_character_limit, int max_character_limit) { + Q_ASSERT(min_character_limit <= max_character_limit); + + LimitableInputDialog dialog{parent}; + dialog.setWindowTitle(title); + dialog.text_label->setText(text); + dialog.text_entry->setMaxLength(max_character_limit); + + auto* const ok_button = dialog.buttons->button(QDialogButtonBox::Ok); + ok_button->setEnabled(false); + connect(dialog.text_entry, &QLineEdit::textEdited, [&](const QString& new_text) { + ok_button->setEnabled(new_text.length() >= min_character_limit); + }); + + if (dialog.exec() != QDialog::Accepted) { + return {}; + } + + return dialog.text_entry->text(); +} diff --git a/src/yuzu/util/limitable_input_dialog.h b/src/yuzu/util/limitable_input_dialog.h new file mode 100644 index 000000000..164ad7301 --- /dev/null +++ b/src/yuzu/util/limitable_input_dialog.h @@ -0,0 +1,31 @@ +// Copyright 2018 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <QDialog> + +class QDialogButtonBox; +class QLabel; +class QLineEdit; + +/// A QDialog that functions similarly to QInputDialog, however, it allows +/// restricting the minimum and total number of characters that can be entered. +class LimitableInputDialog final : public QDialog { + Q_OBJECT +public: + explicit LimitableInputDialog(QWidget* parent = nullptr); + ~LimitableInputDialog() override; + + static QString GetText(QWidget* parent, const QString& title, const QString& text, + int min_character_limit, int max_character_limit); + +private: + void CreateUI(); + void ConnectEvents(); + + QLabel* text_label; + QLineEdit* text_entry; + QDialogButtonBox* buttons; +}; diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp index b456266a6..9cc409fd5 100644 --- a/src/yuzu_cmd/config.cpp +++ b/src/yuzu_cmd/config.cpp @@ -132,6 +132,13 @@ void Config::ReadValues() { Settings::values.current_user = std::clamp<int>( sdl2_config->GetInteger("System", "current_user", 0), 0, Service::Account::MAX_USERS - 1); + const auto enabled = sdl2_config->GetBoolean("System", "rng_seed_enabled", false); + if (enabled) { + Settings::values.rng_seed = sdl2_config->GetInteger("System", "rng_seed", 0); + } else { + Settings::values.rng_seed = std::nullopt; + } + // Miscellaneous Settings::values.log_filter = sdl2_config->Get("Miscellaneous", "log_filter", "*:Trace"); Settings::values.use_dev_keys = sdl2_config->GetBoolean("Miscellaneous", "use_dev_keys", false); @@ -141,6 +148,7 @@ void Config::ReadValues() { Settings::values.gdbstub_port = static_cast<u16>(sdl2_config->GetInteger("Debugging", "gdbstub_port", 24689)); Settings::values.program_args = sdl2_config->Get("Debugging", "program_args", ""); + Settings::values.dump_nso = sdl2_config->GetBoolean("Debugging", "dump_nso", false); // Web Service Settings::values.enable_telemetry = diff --git a/src/yuzu_cmd/default_ini.h b/src/yuzu_cmd/default_ini.h index e0b223cd6..ecf625e7b 100644 --- a/src/yuzu_cmd/default_ini.h +++ b/src/yuzu_cmd/default_ini.h @@ -178,6 +178,11 @@ use_docked_mode = # 1 (default): Yes, 0 : No enable_nfc = +# Sets the seed for the RNG generator built into the switch +# rng_seed will be ignored and randomly generated if rng_seed_enabled is false +rng_seed_enabled = +rng_seed = + # Sets the account username, max length is 32 characters # yuzu (default) username = yuzu @@ -201,6 +206,8 @@ log_filter = *:Trace # Port for listening to GDB connections. use_gdbstub=false gdbstub_port=24689 +# Determines whether or not yuzu will dump all NSOs it attempts to load while loading them +dump_nso=false [WebService] # Whether or not to enable telemetry diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp index c8b93b85b..806127b12 100644 --- a/src/yuzu_cmd/yuzu.cpp +++ b/src/yuzu_cmd/yuzu.cpp @@ -76,6 +76,9 @@ static void InitializeLogging() { const std::string& log_dir = FileUtil::GetUserPath(FileUtil::UserPath::LogDir); FileUtil::CreateFullPath(log_dir); Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir + LOG_FILE)); +#ifdef _WIN32 + Log::AddBackend(std::make_unique<Log::DebuggerBackend>()); +#endif } /// Application entry point |