diff options
Diffstat (limited to 'src')
134 files changed, 1363 insertions, 707 deletions
diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt index 6ffc612e7..d1ec8ff08 100644 --- a/src/common/CMakeLists.txt +++ b/src/common/CMakeLists.txt @@ -172,5 +172,5 @@ endif() create_target_directory_groups(common) -target_link_libraries(common PUBLIC Boost::boost fmt microprofile) -target_link_libraries(common PRIVATE lz4_static libzstd_static) +target_link_libraries(common PUBLIC Boost::boost fmt::fmt microprofile) +target_link_libraries(common PRIVATE lz4::lz4 zstd::zstd) diff --git a/src/common/uuid.h b/src/common/uuid.h index f6ad064fb..4d3af8cec 100644 --- a/src/common/uuid.h +++ b/src/common/uuid.h @@ -40,6 +40,11 @@ struct UUID { uuid = INVALID_UUID; } + // TODO(ogniK): Properly generate a Nintendo ID + constexpr u64 GetNintendoID() const { + return uuid[0]; + } + std::string Format() const; std::string FormatSwitch() const; }; diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt index 8546d3602..47418006b 100644 --- a/src/core/CMakeLists.txt +++ b/src/core/CMakeLists.txt @@ -606,11 +606,11 @@ endif() create_target_directory_groups(core) target_link_libraries(core PUBLIC common PRIVATE audio_core video_core) -target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt json-headers mbedtls opus unicorn) +target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt::fmt nlohmann_json::nlohmann_json mbedtls Opus::Opus unicorn) if (YUZU_ENABLE_BOXCAT) target_compile_definitions(core PRIVATE -DYUZU_ENABLE_BOXCAT) - target_link_libraries(core PRIVATE httplib json-headers zip) + target_link_libraries(core PRIVATE httplib nlohmann_json::nlohmann_json zip) endif() if (ENABLE_WEB_SERVICE) diff --git a/src/core/hle/kernel/handle_table.cpp b/src/core/hle/kernel/handle_table.cpp index e441a27fc..35448b576 100644 --- a/src/core/hle/kernel/handle_table.cpp +++ b/src/core/hle/kernel/handle_table.cpp @@ -30,6 +30,7 @@ HandleTable::~HandleTable() = default; ResultCode HandleTable::SetSize(s32 handle_table_size) { if (static_cast<u32>(handle_table_size) > MAX_COUNT) { + LOG_ERROR(Kernel, "Handle table size {} is greater than {}", handle_table_size, MAX_COUNT); return ERR_OUT_OF_MEMORY; } @@ -80,6 +81,7 @@ ResultVal<Handle> HandleTable::Duplicate(Handle handle) { ResultCode HandleTable::Close(Handle handle) { if (!IsValid(handle)) { + LOG_ERROR(Kernel, "Handle is not valid! handle={:08X}", handle); return ERR_INVALID_HANDLE; } diff --git a/src/core/hle/kernel/hle_ipc.cpp b/src/core/hle/kernel/hle_ipc.cpp index 91d94025c..ba0eac4c2 100644 --- a/src/core/hle/kernel/hle_ipc.cpp +++ b/src/core/hle/kernel/hle_ipc.cpp @@ -13,7 +13,6 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "common/logging/log.h" -#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/hle_ipc.h" @@ -57,7 +56,6 @@ std::shared_ptr<WritableEvent> HLERequestContext::SleepClientThread( return true; }); - auto& kernel = Core::System::GetInstance().Kernel(); if (!writable_event) { // Create event if not provided const auto pair = WritableEvent::CreateEventPair(kernel, "HLE Pause Event: " + reason); @@ -79,9 +77,11 @@ std::shared_ptr<WritableEvent> HLERequestContext::SleepClientThread( return writable_event; } -HLERequestContext::HLERequestContext(std::shared_ptr<Kernel::ServerSession> server_session, +HLERequestContext::HLERequestContext(KernelCore& kernel, Core::Memory::Memory& memory, + std::shared_ptr<ServerSession> server_session, std::shared_ptr<Thread> thread) - : server_session(std::move(server_session)), thread(std::move(thread)) { + : server_session(std::move(server_session)), + thread(std::move(thread)), kernel{kernel}, memory{memory} { cmd_buf[0] = 0; } @@ -216,7 +216,6 @@ ResultCode HLERequestContext::PopulateFromIncomingCommandBuffer(const HandleTabl ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(Thread& thread) { auto& owner_process = *thread.GetOwnerProcess(); auto& handle_table = owner_process.GetHandleTable(); - auto& memory = Core::System::GetInstance().Memory(); std::array<u32, IPC::COMMAND_BUFFER_LENGTH> dst_cmdbuf; memory.ReadBlock(owner_process, thread.GetTLSAddress(), dst_cmdbuf.data(), @@ -286,7 +285,6 @@ std::vector<u8> HLERequestContext::ReadBuffer(std::size_t buffer_index) const { std::vector<u8> buffer; const bool is_buffer_a{BufferDescriptorA().size() > buffer_index && BufferDescriptorA()[buffer_index].Size()}; - auto& memory = Core::System::GetInstance().Memory(); if (is_buffer_a) { ASSERT_MSG(BufferDescriptorA().size() > buffer_index, @@ -319,7 +317,6 @@ std::size_t HLERequestContext::WriteBuffer(const void* buffer, std::size_t size, size = buffer_size; // TODO(bunnei): This needs to be HW tested } - auto& memory = Core::System::GetInstance().Memory(); if (is_buffer_b) { ASSERT_MSG(BufferDescriptorB().size() > buffer_index, "BufferDescriptorB invalid buffer_index {}", buffer_index); diff --git a/src/core/hle/kernel/hle_ipc.h b/src/core/hle/kernel/hle_ipc.h index af3330297..b31673928 100644 --- a/src/core/hle/kernel/hle_ipc.h +++ b/src/core/hle/kernel/hle_ipc.h @@ -19,6 +19,10 @@ union ResultCode; +namespace Core::Memory { +class Memory; +} + namespace Service { class ServiceFrameworkBase; } @@ -28,6 +32,7 @@ namespace Kernel { class Domain; class HandleTable; class HLERequestContext; +class KernelCore; class Process; class ServerSession; class Thread; @@ -98,7 +103,8 @@ protected: */ class HLERequestContext { public: - explicit HLERequestContext(std::shared_ptr<ServerSession> session, + explicit HLERequestContext(KernelCore& kernel, Core::Memory::Memory& memory, + std::shared_ptr<ServerSession> session, std::shared_ptr<Thread> thread); ~HLERequestContext(); @@ -305,6 +311,9 @@ private: std::vector<std::shared_ptr<SessionRequestHandler>> domain_request_handlers; bool is_thread_waiting{}; + + KernelCore& kernel; + Core::Memory::Memory& memory; }; } // namespace Kernel diff --git a/src/core/hle/kernel/memory/address_space_info.cpp b/src/core/hle/kernel/memory/address_space_info.cpp index 27fae05e7..a523a2502 100644 --- a/src/core/hle/kernel/memory/address_space_info.cpp +++ b/src/core/hle/kernel/memory/address_space_info.cpp @@ -2,8 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -// This file references various implementation details from Atmosphère, an open-source firmware for -// the Nintendo Switch. Copyright 2018-2020 Atmosphère-NX. +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. #include <array> @@ -49,18 +49,18 @@ constexpr bool IsAllowedIndexForAddress(std::size_t index) { return index < std::size(AddressSpaceInfos) && AddressSpaceInfos[index].GetAddress() != Invalid; } -constexpr std::size_t - AddressSpaceIndices32Bit[static_cast<std::size_t>(AddressSpaceInfo::Type::Count)]{ +constexpr std::array<std::size_t, static_cast<std::size_t>(AddressSpaceInfo::Type::Count)> + AddressSpaceIndices32Bit{ 0, 1, 0, 2, 0, 3, }; -constexpr std::size_t - AddressSpaceIndices36Bit[static_cast<std::size_t>(AddressSpaceInfo::Type::Count)]{ +constexpr std::array<std::size_t, static_cast<std::size_t>(AddressSpaceInfo::Type::Count)> + AddressSpaceIndices36Bit{ 4, 5, 4, 6, 4, 7, }; -constexpr std::size_t - AddressSpaceIndices39Bit[static_cast<std::size_t>(AddressSpaceInfo::Type::Count)]{ +constexpr std::array<std::size_t, static_cast<std::size_t>(AddressSpaceInfo::Type::Count)> + AddressSpaceIndices39Bit{ 9, 8, 8, 10, 12, 11, }; diff --git a/src/core/hle/kernel/memory/address_space_info.h b/src/core/hle/kernel/memory/address_space_info.h index cc9a6421e..c479890be 100644 --- a/src/core/hle/kernel/memory/address_space_info.h +++ b/src/core/hle/kernel/memory/address_space_info.h @@ -2,12 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -// This file references various implementation details from Atmosphère, an open-source firmware for -// the Nintendo Switch. Copyright 2018-2020 Atmosphère-NX. +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. #pragma once -#include "common/common_funcs.h" #include "common/common_types.h" namespace Kernel::Memory { diff --git a/src/core/hle/kernel/memory/memory_block.h b/src/core/hle/kernel/memory/memory_block.h index 9db1f7b39..9d7839d08 100644 --- a/src/core/hle/kernel/memory/memory_block.h +++ b/src/core/hle/kernel/memory/memory_block.h @@ -2,8 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -// This file references various implementation details from Atmosphère, an open-source firmware for -// the Nintendo Switch. Copyright 2018-2020 Atmosphère-NX. +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. #pragma once diff --git a/src/core/hle/kernel/memory/memory_block_manager.cpp b/src/core/hle/kernel/memory/memory_block_manager.cpp index 900395c37..0732fa5a1 100644 --- a/src/core/hle/kernel/memory/memory_block_manager.cpp +++ b/src/core/hle/kernel/memory/memory_block_manager.cpp @@ -67,7 +67,6 @@ void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState p MemoryPermission prev_perm, MemoryAttribute prev_attribute, MemoryState state, MemoryPermission perm, MemoryAttribute attribute) { - const std::size_t prev_count{memory_block_tree.size()}; const VAddr end_addr{addr + num_pages * PageSize}; iterator node{memory_block_tree.begin()}; @@ -109,7 +108,6 @@ void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState p void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState state, MemoryPermission perm, MemoryAttribute attribute) { - const std::size_t prev_count{memory_block_tree.size()}; const VAddr end_addr{addr + num_pages * PageSize}; iterator node{memory_block_tree.begin()}; @@ -145,7 +143,6 @@ void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState s void MemoryBlockManager::UpdateLock(VAddr addr, std::size_t num_pages, LockFunc&& lock_func, MemoryPermission perm) { - const std::size_t prev_count{memory_block_tree.size()}; const VAddr end_addr{addr + num_pages * PageSize}; iterator node{memory_block_tree.begin()}; diff --git a/src/core/hle/kernel/memory/memory_block_manager.h b/src/core/hle/kernel/memory/memory_block_manager.h index 9451b5df6..6e1d41075 100644 --- a/src/core/hle/kernel/memory/memory_block_manager.h +++ b/src/core/hle/kernel/memory/memory_block_manager.h @@ -6,7 +6,6 @@ #include <functional> #include <list> -#include <memory> #include "common/common_types.h" #include "core/hle/kernel/memory/memory_block.h" diff --git a/src/core/hle/kernel/memory/memory_manager.cpp b/src/core/hle/kernel/memory/memory_manager.cpp index 3cd4f9e85..6b432e1b2 100644 --- a/src/core/hle/kernel/memory/memory_manager.cpp +++ b/src/core/hle/kernel/memory/memory_manager.cpp @@ -104,9 +104,9 @@ ResultCode MemoryManager::Allocate(PageLinkedList& page_list, std::size_t num_pa // Ensure that we don't leave anything un-freed auto group_guard = detail::ScopeExit([&] { for (const auto& it : page_list.Nodes()) { - const auto num_pages{std::min( + const auto min_num_pages{std::min( it.GetNumPages(), (chosen_manager.GetEndAddress() - it.GetAddress()) / PageSize)}; - chosen_manager.Free(it.GetAddress(), num_pages); + chosen_manager.Free(it.GetAddress(), min_num_pages); } }); @@ -165,9 +165,9 @@ ResultCode MemoryManager::Free(PageLinkedList& page_list, std::size_t num_pages, // Free all of the pages for (const auto& it : page_list.Nodes()) { - const auto num_pages{std::min( + const auto min_num_pages{std::min( it.GetNumPages(), (chosen_manager.GetEndAddress() - it.GetAddress()) / PageSize)}; - chosen_manager.Free(it.GetAddress(), num_pages); + chosen_manager.Free(it.GetAddress(), min_num_pages); } return RESULT_SUCCESS; diff --git a/src/core/hle/kernel/memory/memory_manager.h b/src/core/hle/kernel/memory/memory_manager.h index b078d7a5e..3cf444857 100644 --- a/src/core/hle/kernel/memory/memory_manager.h +++ b/src/core/hle/kernel/memory/memory_manager.h @@ -7,7 +7,6 @@ #include <array> #include <mutex> -#include "common/common_funcs.h" #include "common/common_types.h" #include "core/hle/kernel/memory/page_heap.h" #include "core/hle/result.h" diff --git a/src/core/hle/kernel/memory/page_heap.cpp b/src/core/hle/kernel/memory/page_heap.cpp index efcbb3cad..0ab1f7205 100644 --- a/src/core/hle/kernel/memory/page_heap.cpp +++ b/src/core/hle/kernel/memory/page_heap.cpp @@ -2,8 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -// This file references various implementation details from Atmosphère, an open-source firmware for -// the Nintendo Switch. Copyright 2018-2020 Atmosphère-NX. +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. #include "core/core.h" #include "core/hle/kernel/memory/page_heap.h" diff --git a/src/core/hle/kernel/memory/page_heap.h b/src/core/hle/kernel/memory/page_heap.h index 380c3f5a1..22b0de860 100644 --- a/src/core/hle/kernel/memory/page_heap.h +++ b/src/core/hle/kernel/memory/page_heap.h @@ -2,8 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -// This file references various implementation details from Atmosphère, an open-source firmware for -// the Nintendo Switch. Copyright 2018-2020 Atmosphère-NX. +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. #pragma once diff --git a/src/core/hle/kernel/memory/page_linked_list.h b/src/core/hle/kernel/memory/page_linked_list.h index 0668d00c6..45dc13eaf 100644 --- a/src/core/hle/kernel/memory/page_linked_list.h +++ b/src/core/hle/kernel/memory/page_linked_list.h @@ -7,7 +7,6 @@ #include <list> #include "common/assert.h" -#include "common/common_funcs.h" #include "common/common_types.h" #include "core/hle/kernel/memory/memory_types.h" #include "core/hle/result.h" diff --git a/src/core/hle/kernel/memory/page_table.cpp b/src/core/hle/kernel/memory/page_table.cpp index 3281611f8..5d6aac00f 100644 --- a/src/core/hle/kernel/memory/page_table.cpp +++ b/src/core/hle/kernel/memory/page_table.cpp @@ -6,7 +6,6 @@ #include "common/assert.h" #include "common/scope_exit.h" #include "core/core.h" -#include "core/device_memory.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/memory/address_space_info.h" diff --git a/src/core/hle/kernel/memory/page_table.h b/src/core/hle/kernel/memory/page_table.h index a867aa050..ce0d38849 100644 --- a/src/core/hle/kernel/memory/page_table.h +++ b/src/core/hle/kernel/memory/page_table.h @@ -4,16 +4,15 @@ #pragma once -#include <list> #include <memory> #include <mutex> -#include "common/common_funcs.h" #include "common/common_types.h" #include "common/page_table.h" #include "core/file_sys/program_metadata.h" #include "core/hle/kernel/memory/memory_block.h" #include "core/hle/kernel/memory/memory_manager.h" +#include "core/hle/result.h" namespace Core { class System; diff --git a/src/core/hle/kernel/memory/slab_heap.h b/src/core/hle/kernel/memory/slab_heap.h index be95fc3f7..465eaddb3 100644 --- a/src/core/hle/kernel/memory/slab_heap.h +++ b/src/core/hle/kernel/memory/slab_heap.h @@ -2,15 +2,14 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -// This file references various implementation details from Atmosphère, an open-source firmware for -// the Nintendo Switch. Copyright 2018-2020 Atmosphère-NX. +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. #pragma once #include <atomic> #include "common/assert.h" -#include "common/common_funcs.h" #include "common/common_types.h" namespace Kernel::Memory { diff --git a/src/core/hle/kernel/memory/system_control.cpp b/src/core/hle/kernel/memory/system_control.cpp index 9cae3c6cb..2f98e9c4c 100644 --- a/src/core/hle/kernel/memory/system_control.cpp +++ b/src/core/hle/kernel/memory/system_control.cpp @@ -2,8 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#pragma once - #include <random> #include "core/hle/kernel/memory/system_control.h" diff --git a/src/core/hle/kernel/mutex.cpp b/src/core/hle/kernel/mutex.cpp index eff4e45b0..7869eb32b 100644 --- a/src/core/hle/kernel/mutex.cpp +++ b/src/core/hle/kernel/mutex.cpp @@ -7,6 +7,7 @@ #include <vector> #include "common/assert.h" +#include "common/logging/log.h" #include "core/core.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" @@ -67,6 +68,7 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle, Handle requesting_thread_handle) { // The mutex address must be 4-byte aligned if ((address % sizeof(u32)) != 0) { + LOG_ERROR(Kernel, "Address is not 4-byte aligned! address={:016X}", address); return ERR_INVALID_ADDRESS; } @@ -88,6 +90,8 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle, } if (holding_thread == nullptr) { + LOG_ERROR(Kernel, "Holding thread does not exist! thread_handle={:08X}", + holding_thread_handle); return ERR_INVALID_HANDLE; } @@ -109,6 +113,7 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle, ResultCode Mutex::Release(VAddr address) { // The mutex address must be 4-byte aligned if ((address % sizeof(u32)) != 0) { + LOG_ERROR(Kernel, "Address is not 4-byte aligned! address={:016X}", address); return ERR_INVALID_ADDRESS; } diff --git a/src/core/hle/kernel/process_capability.cpp b/src/core/hle/kernel/process_capability.cpp index 48e5ae682..63880f13d 100644 --- a/src/core/hle/kernel/process_capability.cpp +++ b/src/core/hle/kernel/process_capability.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include "common/bit_util.h" +#include "common/logging/log.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/memory/page_table.h" @@ -119,22 +120,30 @@ ResultCode ProcessCapabilities::ParseCapabilities(const u32* capabilities, // The MapPhysical type uses two descriptor flags for its parameters. // If there's only one, then there's a problem. if (i >= num_capabilities) { + LOG_ERROR(Kernel, "Invalid combination! i={}", i); return ERR_INVALID_COMBINATION; } const auto size_flags = capabilities[i]; if (GetCapabilityType(size_flags) != CapabilityType::MapPhysical) { + LOG_ERROR(Kernel, "Invalid capability type! size_flags={}", size_flags); return ERR_INVALID_COMBINATION; } const auto result = HandleMapPhysicalFlags(descriptor, size_flags, page_table); if (result.IsError()) { + LOG_ERROR(Kernel, "Failed to map physical flags! descriptor={}, size_flags={}", + descriptor, size_flags); return result; } } else { const auto result = ParseSingleFlagCapability(set_flags, set_svc_bits, descriptor, page_table); if (result.IsError()) { + LOG_ERROR( + Kernel, + "Failed to parse capability flag! set_flags={}, set_svc_bits={}, descriptor={}", + set_flags, set_svc_bits, descriptor); return result; } } @@ -162,6 +171,9 @@ ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& s const u32 flag_length = GetFlagBitOffset(type); const u32 set_flag = 1U << flag_length; if ((set_flag & set_flags & InitializeOnceMask) != 0) { + LOG_ERROR(Kernel, + "Attempted to initialize flags that may only be initialized once. set_flags={}", + set_flags); return ERR_INVALID_COMBINATION; } set_flags |= set_flag; @@ -187,6 +199,7 @@ ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& s break; } + LOG_ERROR(Kernel, "Invalid capability type! type={}", static_cast<u32>(type)); return ERR_INVALID_CAPABILITY_DESCRIPTOR; } @@ -208,23 +221,31 @@ void ProcessCapabilities::Clear() { ResultCode ProcessCapabilities::HandlePriorityCoreNumFlags(u32 flags) { if (priority_mask != 0 || core_mask != 0) { + LOG_ERROR(Kernel, "Core or priority mask are not zero! priority_mask={}, core_mask={}", + priority_mask, core_mask); return ERR_INVALID_CAPABILITY_DESCRIPTOR; } const u32 core_num_min = (flags >> 16) & 0xFF; const u32 core_num_max = (flags >> 24) & 0xFF; if (core_num_min > core_num_max) { + LOG_ERROR(Kernel, "Core min is greater than core max! core_num_min={}, core_num_max={}", + core_num_min, core_num_max); return ERR_INVALID_COMBINATION; } const u32 priority_min = (flags >> 10) & 0x3F; const u32 priority_max = (flags >> 4) & 0x3F; if (priority_min > priority_max) { + LOG_ERROR(Kernel, + "Priority min is greater than priority max! priority_min={}, priority_max={}", + core_num_min, priority_max); return ERR_INVALID_COMBINATION; } // The switch only has 4 usable cores. if (core_num_max >= 4) { + LOG_ERROR(Kernel, "Invalid max cores specified! core_num_max={}", core_num_max); return ERR_INVALID_PROCESSOR_ID; } @@ -259,6 +280,7 @@ ResultCode ProcessCapabilities::HandleSyscallFlags(u32& set_svc_bits, u32 flags) } if (svc_number >= svc_capabilities.size()) { + LOG_ERROR(Kernel, "Process svc capability is out of range! svc_number={}", svc_number); return ERR_OUT_OF_RANGE; } @@ -295,6 +317,8 @@ ResultCode ProcessCapabilities::HandleInterruptFlags(u32 flags) { // emulate that, it's sufficient to mark every interrupt as defined. if (interrupt >= interrupt_capabilities.size()) { + LOG_ERROR(Kernel, "Process interrupt capability is out of range! svc_number={}", + interrupt); return ERR_OUT_OF_RANGE; } @@ -307,6 +331,7 @@ ResultCode ProcessCapabilities::HandleInterruptFlags(u32 flags) { ResultCode ProcessCapabilities::HandleProgramTypeFlags(u32 flags) { const u32 reserved = flags >> 17; if (reserved != 0) { + LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved); return ERR_RESERVED_VALUE; } @@ -324,6 +349,9 @@ ResultCode ProcessCapabilities::HandleKernelVersionFlags(u32 flags) { const u32 major_version = kernel_version >> 19; if (major_version != 0 || flags < 0x80000) { + LOG_ERROR(Kernel, + "Kernel version is non zero or flags are too small! major_version={}, flags={}", + major_version, flags); return ERR_INVALID_CAPABILITY_DESCRIPTOR; } @@ -334,6 +362,7 @@ ResultCode ProcessCapabilities::HandleKernelVersionFlags(u32 flags) { ResultCode ProcessCapabilities::HandleHandleTableFlags(u32 flags) { const u32 reserved = flags >> 26; if (reserved != 0) { + LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved); return ERR_RESERVED_VALUE; } @@ -344,6 +373,7 @@ ResultCode ProcessCapabilities::HandleHandleTableFlags(u32 flags) { ResultCode ProcessCapabilities::HandleDebugFlags(u32 flags) { const u32 reserved = flags >> 19; if (reserved != 0) { + LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved); return ERR_RESERVED_VALUE; } diff --git a/src/core/hle/kernel/readable_event.cpp b/src/core/hle/kernel/readable_event.cpp index 9d3d3a81b..00860fcbd 100644 --- a/src/core/hle/kernel/readable_event.cpp +++ b/src/core/hle/kernel/readable_event.cpp @@ -4,6 +4,7 @@ #include <algorithm> #include "common/assert.h" +#include "common/logging/log.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/readable_event.h" @@ -23,10 +24,12 @@ void ReadableEvent::Acquire(Thread* thread) { } void ReadableEvent::Signal() { - if (!is_signaled) { - is_signaled = true; - SynchronizationObject::Signal(); - }; + if (is_signaled) { + return; + } + + is_signaled = true; + SynchronizationObject::Signal(); } void ReadableEvent::Clear() { @@ -35,6 +38,8 @@ void ReadableEvent::Clear() { ResultCode ReadableEvent::Reset() { if (!is_signaled) { + LOG_ERROR(Kernel, "Handle is not signaled! object_id={}, object_type={}, object_name={}", + GetObjectId(), GetTypeName(), GetName()); return ERR_INVALID_STATE; } diff --git a/src/core/hle/kernel/resource_limit.cpp b/src/core/hle/kernel/resource_limit.cpp index 96e5b9892..d9beaa3a4 100644 --- a/src/core/hle/kernel/resource_limit.cpp +++ b/src/core/hle/kernel/resource_limit.cpp @@ -69,6 +69,8 @@ ResultCode ResourceLimit::SetLimitValue(ResourceType resource, s64 value) { limit[index] = value; return RESULT_SUCCESS; } else { + LOG_ERROR(Kernel, "Limit value is too large! resource={}, value={}, index={}", + static_cast<u32>(resource), value, index); return ERR_INVALID_STATE; } } diff --git a/src/core/hle/kernel/server_session.cpp b/src/core/hle/kernel/server_session.cpp index 0f102ca44..25438b86b 100644 --- a/src/core/hle/kernel/server_session.cpp +++ b/src/core/hle/kernel/server_session.cpp @@ -137,8 +137,8 @@ ResultCode ServerSession::HandleDomainSyncRequest(Kernel::HLERequestContext& con ResultCode ServerSession::QueueSyncRequest(std::shared_ptr<Thread> thread, Core::Memory::Memory& memory) { u32* cmd_buf{reinterpret_cast<u32*>(memory.GetPointer(thread->GetTLSAddress()))}; - std::shared_ptr<Kernel::HLERequestContext> context{ - std::make_shared<Kernel::HLERequestContext>(SharedFrom(this), std::move(thread))}; + auto context = + std::make_shared<HLERequestContext>(kernel, memory, SharedFrom(this), std::move(thread)); context->PopulateFromIncomingCommandBuffer(kernel.CurrentProcess()->GetHandleTable(), cmd_buf); request_queue.Push(std::move(context)); diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index 41ef2caf6..4ae4529f5 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -685,6 +685,8 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha case GetInfoType::TotalPhysicalMemoryAvailableWithoutSystemResource: case GetInfoType::TotalPhysicalMemoryUsedWithoutSystemResource: { if (info_sub_id != 0) { + LOG_ERROR(Kernel_SVC, "Info sub id is non zero! info_id={}, info_sub_id={}", info_id, + info_sub_id); return ERR_INVALID_ENUM_VALUE; } @@ -692,6 +694,8 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha system.Kernel().CurrentProcess()->GetHandleTable(); const auto process = current_process_handle_table.Get<Process>(static_cast<Handle>(handle)); if (!process) { + LOG_ERROR(Kernel_SVC, "Process is not valid! info_id={}, info_sub_id={}, handle={:08X}", + info_id, info_sub_id, handle); return ERR_INVALID_HANDLE; } @@ -783,10 +787,13 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha case GetInfoType::RegisterResourceLimit: { if (handle != 0) { + LOG_ERROR(Kernel, "Handle is non zero! handle={:08X}", handle); return ERR_INVALID_HANDLE; } if (info_sub_id != 0) { + LOG_ERROR(Kernel, "Info sub id is non zero! info_id={}, info_sub_id={}", info_id, + info_sub_id); return ERR_INVALID_COMBINATION; } diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp index a919750a6..db7f379ac 100644 --- a/src/core/hle/kernel/thread.cpp +++ b/src/core/hle/kernel/thread.cpp @@ -423,6 +423,8 @@ ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) { if (new_core == THREADPROCESSORID_DONT_UPDATE) { new_core = use_override ? ideal_core_override : ideal_core; if ((new_affinity_mask & (1ULL << new_core)) == 0) { + LOG_ERROR(Kernel, "New affinity mask is incorrect! new_core={}, new_affinity_mask={}", + new_core, new_affinity_mask); return ERR_INVALID_COMBINATION; } } diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp index cfac8ca9a..630a8b048 100644 --- a/src/core/hle/service/acc/acc.cpp +++ b/src/core/hle/service/acc/acc.cpp @@ -228,7 +228,8 @@ public: class IManagerForApplication final : public ServiceFramework<IManagerForApplication> { public: - IManagerForApplication() : ServiceFramework("IManagerForApplication") { + explicit IManagerForApplication(Common::UUID user_id) + : ServiceFramework("IManagerForApplication"), user_id(user_id) { // clang-format off static const FunctionInfo functions[] = { {0, &IManagerForApplication::CheckAvailability, "CheckAvailability"}, @@ -254,12 +255,14 @@ private: } void GetAccountId(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_ACC, "(STUBBED) called"); - // Should return a nintendo account ID + LOG_DEBUG(Service_ACC, "called"); + IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.PushRaw<u64>(1); + rb.PushRaw<u64>(user_id.GetNintendoID()); } + + Common::UUID user_id; }; void Module::Interface::GetUserCount(Kernel::HLERequestContext& ctx) { @@ -319,46 +322,37 @@ void Module::Interface::IsUserRegistrationRequestPermitted(Kernel::HLERequestCon void Module::Interface::InitializeApplicationInfo(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto pid = rp.Pop<u64>(); - LOG_DEBUG(Service_ACC, "called, process_id={}", pid); + LOG_DEBUG(Service_ACC, "called"); IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(InitializeApplicationInfoBase(pid)); + rb.Push(InitializeApplicationInfoBase()); } void Module::Interface::InitializeApplicationInfoRestricted(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto pid = rp.Pop<u64>(); - LOG_WARNING(Service_ACC, "(Partial implementation) called, process_id={}", pid); + LOG_WARNING(Service_ACC, "(Partial implementation) called"); // TODO(ogniK): We require checking if the user actually owns the title and what not. As of // currently, we assume the user owns the title. InitializeApplicationInfoBase SHOULD be called // first then we do extra checks if the game is a digital copy. IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(InitializeApplicationInfoBase(pid)); + rb.Push(InitializeApplicationInfoBase()); } -ResultCode Module::Interface::InitializeApplicationInfoBase(u64 process_id) { +ResultCode Module::Interface::InitializeApplicationInfoBase() { if (application_info) { LOG_ERROR(Service_ACC, "Application already initialized"); return ERR_ACCOUNTINFO_ALREADY_INITIALIZED; } - const auto& list = system.Kernel().GetProcessList(); - const auto iter = std::find_if(list.begin(), list.end(), [&process_id](const auto& process) { - return process->GetProcessID() == process_id; - }); - - if (iter == list.end()) { - LOG_ERROR(Service_ACC, "Failed to find process ID"); - application_info.application_type = ApplicationType::Unknown; - - return ERR_ACCOUNTINFO_BAD_APPLICATION; - } - - const auto launch_property = system.GetARPManager().GetLaunchProperty((*iter)->GetTitleID()); + // TODO(ogniK): This should be changed to reflect the target process for when we have multiple + // processes emulated. As we don't actually have pid support we should assume we're just using + // our own process + const auto& current_process = system.Kernel().CurrentProcess(); + const auto launch_property = + system.GetARPManager().GetLaunchProperty(current_process->GetTitleID()); if (launch_property.Failed()) { LOG_ERROR(Service_ACC, "Failed to get launch property"); @@ -372,10 +366,12 @@ ResultCode Module::Interface::InitializeApplicationInfoBase(u64 process_id) { case FileSys::StorageId::Host: case FileSys::StorageId::NandUser: case FileSys::StorageId::SdCard: + case FileSys::StorageId::None: // Yuzu specific, differs from hardware application_info.application_type = ApplicationType::Digital; break; default: - LOG_ERROR(Service_ACC, "Invalid game storage ID"); + LOG_ERROR(Service_ACC, "Invalid game storage ID! storage_id={}", + launch_property->base_game_storage_id); return ERR_ACCOUNTINFO_BAD_APPLICATION; } @@ -389,7 +385,7 @@ void Module::Interface::GetBaasAccountManagerForApplication(Kernel::HLERequestCo LOG_DEBUG(Service_ACC, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IManagerForApplication>(); + rb.PushIpcInterface<IManagerForApplication>(profile_manager->GetLastOpenedUser()); } void Module::Interface::IsUserAccountSwitchLocked(Kernel::HLERequestContext& ctx) { @@ -428,6 +424,17 @@ void Module::Interface::GetProfileEditor(Kernel::HLERequestContext& ctx) { rb.PushIpcInterface<IProfileEditor>(user_id, *profile_manager); } +void Module::Interface::ListQualifiedUsers(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_ACC, "called"); + + // All users should be qualified. We don't actually have parental control or anything to do with + // nintendo online currently. We're just going to assume the user running the game has access to + // the game regardless of parental control settings. + ctx.WriteBuffer(profile_manager->GetAllUsers()); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + 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 diff --git a/src/core/hle/service/acc/acc.h b/src/core/hle/service/acc/acc.h index 7a7dc9ec6..74ca39d6e 100644 --- a/src/core/hle/service/acc/acc.h +++ b/src/core/hle/service/acc/acc.h @@ -33,9 +33,10 @@ public: void TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx); void IsUserAccountSwitchLocked(Kernel::HLERequestContext& ctx); void GetProfileEditor(Kernel::HLERequestContext& ctx); + void ListQualifiedUsers(Kernel::HLERequestContext& ctx); private: - ResultCode InitializeApplicationInfoBase(u64 process_id); + ResultCode InitializeApplicationInfoBase(); enum class ApplicationType : u32_le { GameCard = 0, diff --git a/src/core/hle/service/acc/acc_su.cpp b/src/core/hle/service/acc/acc_su.cpp index ae88deda5..2eefc6df5 100644 --- a/src/core/hle/service/acc/acc_su.cpp +++ b/src/core/hle/service/acc/acc_su.cpp @@ -35,7 +35,7 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {113, nullptr, "GetSaveDataThumbnailExistence"}, {120, nullptr, "ListOpenUsersInApplication"}, {130, nullptr, "ActivateOpenContextRetention"}, - {140, nullptr, "ListQualifiedUsers"}, + {140, &ACC_SU::ListQualifiedUsers, "ListQualifiedUsers"}, {150, nullptr, "AuthenticateApplicationAsync"}, {190, nullptr, "GetUserLastOpenedApplication"}, {191, nullptr, "ActivateOpenContextHolder"}, diff --git a/src/core/hle/service/acc/acc_u0.cpp b/src/core/hle/service/acc/acc_u0.cpp index 0ac19f4ff..fb4e7e772 100644 --- a/src/core/hle/service/acc/acc_u0.cpp +++ b/src/core/hle/service/acc/acc_u0.cpp @@ -32,7 +32,7 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {130, nullptr, "LoadOpenContext"}, {131, nullptr, "ListOpenContextStoredUsers"}, {140, &ACC_U0::InitializeApplicationInfoRestricted, "InitializeApplicationInfoRestricted"}, - {141, nullptr, "ListQualifiedUsers"}, + {141, &ACC_U0::ListQualifiedUsers, "ListQualifiedUsers"}, {150, &ACC_U0::IsUserAccountSwitchLocked, "IsUserAccountSwitchLocked"}, }; // clang-format on diff --git a/src/core/hle/service/acc/acc_u1.cpp b/src/core/hle/service/acc/acc_u1.cpp index 2b9c11928..9f29cdc82 100644 --- a/src/core/hle/service/acc/acc_u1.cpp +++ b/src/core/hle/service/acc/acc_u1.cpp @@ -34,7 +34,7 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {112, nullptr, "LoadSaveDataThumbnail"}, {113, nullptr, "GetSaveDataThumbnailExistence"}, {130, nullptr, "ActivateOpenContextRetention"}, - {140, nullptr, "ListQualifiedUsers"}, + {140, &ACC_U1::ListQualifiedUsers, "ListQualifiedUsers"}, {150, nullptr, "AuthenticateApplicationAsync"}, {190, nullptr, "GetUserLastOpenedApplication"}, {191, nullptr, "ActivateOpenContextHolder"}, diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp index 5695d2521..4df74c4f9 100644 --- a/src/core/hle/service/am/am.cpp +++ b/src/core/hle/service/am/am.cpp @@ -52,11 +52,6 @@ enum class LaunchParameterKind : u32 { AccountPreselectedUser = 2, }; -enum class VrMode : u8 { - Disabled = 0, - Enabled = 1, -}; - constexpr u32 LAUNCH_PARAMETER_ACCOUNT_PRESELECTED_USER_MAGIC = 0xC79497CA; struct LaunchParameterAccountPreselectedUser { @@ -685,27 +680,21 @@ void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) { } void ICommonStateGetter::IsVrModeEnabled(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.PushEnum(VrMode::Disabled); + rb.Push(vr_mode_state); } void ICommonStateGetter::SetVrModeEnabled(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto is_vr_mode_enabled = rp.Pop<bool>(); + vr_mode_state = rp.Pop<bool>(); - LOG_WARNING(Service_AM, "(STUBBED) called. is_vr_mode_enabled={}", is_vr_mode_enabled); + LOG_WARNING(Service_AM, "VR Mode is {}", vr_mode_state ? "on" : "off"); IPC::ResponseBuilder rb{ctx, 2}; - if (!is_vr_mode_enabled) { - rb.Push(RESULT_SUCCESS); - } else { - // TODO: Find better error code for this - UNIMPLEMENTED_MSG("is_vr_mode_enabled={}", is_vr_mode_enabled); - rb.Push(RESULT_UNKNOWN); - } + rb.Push(RESULT_SUCCESS); } void ICommonStateGetter::SetLcdBacklighOffEnabled(Kernel::HLERequestContext& ctx) { @@ -1169,7 +1158,7 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {121, nullptr, "ClearUserChannel"}, {122, nullptr, "UnpopToUserChannel"}, {130, &IApplicationFunctions::GetGpuErrorDetectedSystemEvent, "GetGpuErrorDetectedSystemEvent"}, - {140, nullptr, "GetFriendInvitationStorageChannelEvent"}, + {140, &IApplicationFunctions::GetFriendInvitationStorageChannelEvent, "GetFriendInvitationStorageChannelEvent"}, {141, nullptr, "TryPopFromFriendInvitationStorageChannel"}, {150, nullptr, "GetNotificationStorageChannelEvent"}, {151, nullptr, "TryPopFromNotificationStorageChannel"}, @@ -1186,6 +1175,9 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) auto& kernel = system.Kernel(); gpu_error_detected_event = Kernel::WritableEvent::CreateEventPair( kernel, "IApplicationFunctions:GpuErrorDetectedSystemEvent"); + + friend_invitation_storage_channel_event = Kernel::WritableEvent::CreateEventPair( + kernel, "IApplicationFunctions:FriendInvitationStorageChannelEvent"); } IApplicationFunctions::~IApplicationFunctions() = default; @@ -1343,12 +1335,23 @@ void IApplicationFunctions::SetTerminateResult(Kernel::HLERequestContext& ctx) { } void IApplicationFunctions::GetDisplayVersion(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); + + std::array<u8, 0x10> version_string{}; + + FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID()}; + const auto res = pm.GetControlMetadata(); + if (res.first != nullptr) { + const auto& version = res.first->GetVersionString(); + std::copy(version.begin(), version.end(), version_string.begin()); + } else { + constexpr u128 default_version = {1, 0}; + std::memcpy(version_string.data(), default_version.data(), sizeof(u128)); + } IPC::ResponseBuilder rb{ctx, 6}; rb.Push(RESULT_SUCCESS); - rb.Push<u64>(1); - rb.Push<u64>(0); + rb.PushRaw(version_string); } void IApplicationFunctions::GetDesiredLanguage(Kernel::HLERequestContext& ctx) { @@ -1500,6 +1503,14 @@ void IApplicationFunctions::GetGpuErrorDetectedSystemEvent(Kernel::HLERequestCon rb.PushCopyObjects(gpu_error_detected_event.readable); } +void IApplicationFunctions::GetFriendInvitationStorageChannelEvent(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_AM, "called"); + + IPC::ResponseBuilder rb{ctx, 2, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushCopyObjects(friend_invitation_storage_channel_event.readable); +} + void InstallInterfaces(SM::ServiceManager& service_manager, std::shared_ptr<NVFlinger::NVFlinger> nvflinger, Core::System& system) { auto message_queue = std::make_shared<AppletMessageQueue>(system.Kernel()); @@ -1514,14 +1525,15 @@ void InstallInterfaces(SM::ServiceManager& service_manager, std::make_shared<TCAP>()->InstallAsService(service_manager); } -IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions") { +IHomeMenuFunctions::IHomeMenuFunctions(Kernel::KernelCore& kernel) + : ServiceFramework("IHomeMenuFunctions"), kernel(kernel) { // clang-format off static const FunctionInfo functions[] = { {10, &IHomeMenuFunctions::RequestToGetForeground, "RequestToGetForeground"}, {11, nullptr, "LockForeground"}, {12, nullptr, "UnlockForeground"}, {20, nullptr, "PopFromGeneralChannel"}, - {21, nullptr, "GetPopFromGeneralChannelEvent"}, + {21, &IHomeMenuFunctions::GetPopFromGeneralChannelEvent, "GetPopFromGeneralChannelEvent"}, {30, nullptr, "GetHomeButtonWriterLockAccessor"}, {31, nullptr, "GetWriterLockAccessorEx"}, {100, nullptr, "PopRequestLaunchApplicationForDebug"}, @@ -1531,6 +1543,9 @@ IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions" // clang-format on RegisterHandlers(functions); + + pop_from_general_channel_event = Kernel::WritableEvent::CreateEventPair( + kernel, "IHomeMenuFunctions:PopFromGeneralChannelEvent"); } IHomeMenuFunctions::~IHomeMenuFunctions() = default; @@ -1542,6 +1557,14 @@ void IHomeMenuFunctions::RequestToGetForeground(Kernel::HLERequestContext& ctx) rb.Push(RESULT_SUCCESS); } +void IHomeMenuFunctions::GetPopFromGeneralChannelEvent(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AM, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushCopyObjects(pop_from_general_channel_event.readable); +} + IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStateController") { // clang-format off static const FunctionInfo functions[] = { diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h index 53cfce10f..469f7f814 100644 --- a/src/core/hle/service/am/am.h +++ b/src/core/hle/service/am/am.h @@ -191,6 +191,7 @@ private: Core::System& system; std::shared_ptr<AppletMessageQueue> msg_queue; + bool vr_mode_state{}; }; class IStorageImpl { @@ -280,20 +281,26 @@ private: void QueryApplicationPlayStatistics(Kernel::HLERequestContext& ctx); void QueryApplicationPlayStatisticsByUid(Kernel::HLERequestContext& ctx); void GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx); + void GetFriendInvitationStorageChannelEvent(Kernel::HLERequestContext& ctx); bool launch_popped_application_specific = false; bool launch_popped_account_preselect = false; Kernel::EventPair gpu_error_detected_event; + Kernel::EventPair friend_invitation_storage_channel_event; Core::System& system; }; class IHomeMenuFunctions final : public ServiceFramework<IHomeMenuFunctions> { public: - IHomeMenuFunctions(); + explicit IHomeMenuFunctions(Kernel::KernelCore& kernel); ~IHomeMenuFunctions() override; private: void RequestToGetForeground(Kernel::HLERequestContext& ctx); + void GetPopFromGeneralChannelEvent(Kernel::HLERequestContext& ctx); + + Kernel::EventPair pop_from_general_channel_event; + Kernel::KernelCore& kernel; }; class IGlobalStateController final : public ServiceFramework<IGlobalStateController> { diff --git a/src/core/hle/service/am/applet_ae.cpp b/src/core/hle/service/am/applet_ae.cpp index e454b77d8..9df286d17 100644 --- a/src/core/hle/service/am/applet_ae.cpp +++ b/src/core/hle/service/am/applet_ae.cpp @@ -202,7 +202,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IHomeMenuFunctions>(); + rb.PushIpcInterface<IHomeMenuFunctions>(system.Kernel()); } void GetGlobalStateController(Kernel::HLERequestContext& ctx) { diff --git a/src/core/hle/service/audio/audin_u.cpp b/src/core/hle/service/audio/audin_u.cpp index d7f1d348d..3e2299426 100644 --- a/src/core/hle/service/audio/audin_u.cpp +++ b/src/core/hle/service/audio/audin_u.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/hle/ipc_helpers.h" +#include "core/hle/kernel/hle_ipc.h" #include "core/hle/service/audio/audin_u.h" namespace Service::Audio { @@ -36,11 +39,12 @@ public: AudInU::AudInU() : ServiceFramework("audin:u") { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "ListAudioIns"}, - {1, nullptr, "OpenAudioIn"}, - {2, nullptr, "Unknown"}, - {3, nullptr, "OpenAudioInAuto"}, - {4, nullptr, "ListAudioInsAuto"}, + {0, &AudInU::ListAudioIns, "ListAudioIns"}, + {1, &AudInU::OpenAudioIn, "OpenAudioIn"}, + {2, &AudInU::ListAudioIns, "ListAudioInsAuto"}, + {3, &AudInU::OpenAudioIn, "OpenAudioInAuto"}, + {4, &AudInU::ListAudioInsAutoFiltered, "ListAudioInsAutoFiltered"}, + {5, &AudInU::OpenAudioInProtocolSpecified, "OpenAudioInProtocolSpecified"}, }; // clang-format on @@ -49,4 +53,60 @@ AudInU::AudInU() : ServiceFramework("audin:u") { AudInU::~AudInU() = default; +void AudInU::ListAudioIns(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_Audio, "called"); + const std::size_t count = ctx.GetWriteBufferSize() / sizeof(AudioInDeviceName); + + const std::size_t device_count = std::min(count, audio_device_names.size()); + std::vector<AudioInDeviceName> device_names; + device_names.reserve(device_count); + + for (std::size_t i = 0; i < device_count; i++) { + const auto& device_name = audio_device_names[i]; + auto& entry = device_names.emplace_back(); + device_name.copy(entry.data(), device_name.size()); + } + + ctx.WriteBuffer(device_names); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(static_cast<u32>(device_names.size())); +} + +void AudInU::ListAudioInsAutoFiltered(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_Audio, "called"); + constexpr u32 device_count = 0; + + // Since we don't actually use any other audio input devices, we return 0 devices. Filtered + // device listing just omits the default input device + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(static_cast<u32>(device_count)); +} + +void AudInU::OpenInOutImpl(Kernel::HLERequestContext& ctx) { + AudInOutParams params{}; + params.channel_count = 2; + params.sample_format = SampleFormat::PCM16; + params.sample_rate = 48000; + params.state = State::Started; + + IPC::ResponseBuilder rb{ctx, 6, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw<AudInOutParams>(params); + rb.PushIpcInterface<IAudioIn>(); +} + +void AudInU::OpenAudioIn(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_Audio, "(STUBBED) called"); + OpenInOutImpl(ctx); +} + +void AudInU::OpenAudioInProtocolSpecified(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_Audio, "(STUBBED) called"); + OpenInOutImpl(ctx); +} + } // namespace Service::Audio diff --git a/src/core/hle/service/audio/audin_u.h b/src/core/hle/service/audio/audin_u.h index 0538b9560..a599f4a64 100644 --- a/src/core/hle/service/audio/audin_u.h +++ b/src/core/hle/service/audio/audin_u.h @@ -16,6 +16,35 @@ class AudInU final : public ServiceFramework<AudInU> { public: explicit AudInU(); ~AudInU() override; + +private: + enum class SampleFormat : u32_le { + PCM16 = 2, + }; + + enum class State : u32_le { + Started = 0, + Stopped = 1, + }; + + struct AudInOutParams { + u32_le sample_rate{}; + u32_le channel_count{}; + SampleFormat sample_format{}; + State state{}; + }; + static_assert(sizeof(AudInOutParams) == 0x10, "AudInOutParams is an invalid size"); + + using AudioInDeviceName = std::array<char, 256>; + static constexpr std::array<std::string_view, 1> audio_device_names{{ + "BuiltInHeadset", + }}; + + void ListAudioIns(Kernel::HLERequestContext& ctx); + void ListAudioInsAutoFiltered(Kernel::HLERequestContext& ctx); + void OpenInOutImpl(Kernel::HLERequestContext& ctx); + void OpenAudioIn(Kernel::HLERequestContext& ctx); + void OpenAudioInProtocolSpecified(Kernel::HLERequestContext& ctx); }; } // namespace Service::Audio diff --git a/src/core/hle/service/bcat/backend/boxcat.cpp b/src/core/hle/service/bcat/backend/boxcat.cpp index 5febe8fc1..d29e78d7e 100644 --- a/src/core/hle/service/bcat/backend/boxcat.cpp +++ b/src/core/hle/service/bcat/backend/boxcat.cpp @@ -4,8 +4,8 @@ #include <fmt/ostream.h> #include <httplib.h> -#include <json.hpp> #include <mbedtls/sha256.h> +#include <nlohmann/json.hpp> #include "common/hex_util.h" #include "common/logging/backend.h" #include "common/logging/log.h" diff --git a/src/core/hle/service/caps/caps_su.cpp b/src/core/hle/service/caps/caps_su.cpp index 2b4c2d808..e8b0698e8 100644 --- a/src/core/hle/service/caps/caps_su.cpp +++ b/src/core/hle/service/caps/caps_su.cpp @@ -2,6 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/logging/log.h" +#include "core/hle/ipc_helpers.h" #include "core/hle/service/caps/caps_su.h" namespace Service::Capture { @@ -9,8 +11,11 @@ namespace Service::Capture { CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") { // clang-format off static const FunctionInfo functions[] = { + {32, &CAPS_SU::SetShimLibraryVersion, "SetShimLibraryVersion"}, {201, nullptr, "SaveScreenShot"}, {203, nullptr, "SaveScreenShotEx0"}, + {205, nullptr, "SaveScreenShotEx1"}, + {210, nullptr, "SaveScreenShotEx2"}, }; // clang-format on @@ -19,4 +24,11 @@ CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") { CAPS_SU::~CAPS_SU() = default; +void CAPS_SU::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_Capture, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_su.h b/src/core/hle/service/caps/caps_su.h index cb11f7c9a..c494d7c84 100644 --- a/src/core/hle/service/caps/caps_su.h +++ b/src/core/hle/service/caps/caps_su.h @@ -16,6 +16,9 @@ class CAPS_SU final : public ServiceFramework<CAPS_SU> { public: explicit CAPS_SU(); ~CAPS_SU() override; + +private: + void SetShimLibraryVersion(Kernel::HLERequestContext& ctx); }; } // namespace Service::Capture diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp index 6b9b4f3b9..f6503fe2f 100644 --- a/src/core/hle/service/filesystem/fsp_srv.cpp +++ b/src/core/hle/service/filesystem/fsp_srv.cpp @@ -316,8 +316,8 @@ public: {8, &IFileSystem::OpenFile, "OpenFile"}, {9, &IFileSystem::OpenDirectory, "OpenDirectory"}, {10, &IFileSystem::Commit, "Commit"}, - {11, nullptr, "GetFreeSpaceSize"}, - {12, nullptr, "GetTotalSpaceSize"}, + {11, &IFileSystem::GetFreeSpaceSize, "GetFreeSpaceSize"}, + {12, &IFileSystem::GetTotalSpaceSize, "GetTotalSpaceSize"}, {13, &IFileSystem::CleanDirectoryRecursively, "CleanDirectoryRecursively"}, {14, nullptr, "GetFileTimeStampRaw"}, {15, nullptr, "QueryEntry"}, diff --git a/src/core/hle/service/nim/nim.cpp b/src/core/hle/service/nim/nim.cpp index e85f123e2..f19affce7 100644 --- a/src/core/hle/service/nim/nim.cpp +++ b/src/core/hle/service/nim/nim.cpp @@ -15,6 +15,66 @@ namespace Service::NIM { +class IShopServiceAsync final : public ServiceFramework<IShopServiceAsync> { +public: + IShopServiceAsync() : ServiceFramework("IShopServiceAsync") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "Cancel"}, + {1, nullptr, "GetSize"}, + {2, nullptr, "Read"}, + {3, nullptr, "GetErrorCode"}, + {4, nullptr, "Request"}, + {5, nullptr, "Prepare"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class IShopServiceAccessor final : public ServiceFramework<IShopServiceAccessor> { +public: + IShopServiceAccessor() : ServiceFramework("IShopServiceAccessor") { + // clang-format off + static const FunctionInfo functions[] = { + {0, &IShopServiceAccessor::CreateAsyncInterface, "CreateAsyncInterface"}, + }; + // clang-format on + + RegisterHandlers(functions); + } + +private: + void CreateAsyncInterface(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IShopServiceAsync>(); + } +}; + +class IShopServiceAccessServer final : public ServiceFramework<IShopServiceAccessServer> { +public: + IShopServiceAccessServer() : ServiceFramework("IShopServiceAccessServer") { + // clang-format off + static const FunctionInfo functions[] = { + {0, &IShopServiceAccessServer::CreateAccessorInterface, "CreateAccessorInterface"}, + }; + // clang-format on + + RegisterHandlers(functions); + } + +private: + void CreateAccessorInterface(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IShopServiceAccessor>(); + } +}; + class NIM final : public ServiceFramework<NIM> { public: explicit NIM() : ServiceFramework{"nim"} { @@ -78,7 +138,7 @@ public: explicit NIM_ECA() : ServiceFramework{"nim:eca"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "CreateServerInterface"}, + {0, &NIM_ECA::CreateServerInterface, "CreateServerInterface"}, {1, nullptr, "RefreshDebugAvailability"}, {2, nullptr, "ClearDebugResponse"}, {3, nullptr, "RegisterDebugResponse"}, @@ -87,6 +147,14 @@ public: RegisterHandlers(functions); } + +private: + void CreateServerInterface(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IShopServiceAccessServer>(); + } }; class NIM_SHP final : public ServiceFramework<NIM_SHP> { diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h index 642b0a2cb..07b644ec5 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h @@ -159,9 +159,10 @@ private: static_assert(sizeof(IoctlFlushL2) == 8, "IoctlFlushL2 is incorrect size"); struct IoctlGetGpuTime { - u64_le gpu_time; + u64_le gpu_time{}; + INSERT_PADDING_WORDS(2); }; - static_assert(sizeof(IoctlGetGpuTime) == 8, "IoctlGetGpuTime is incorrect size"); + static_assert(sizeof(IoctlGetGpuTime) == 0x10, "IoctlGetGpuTime is incorrect size"); u32 GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, std::vector<u8>& output2, IoctlVersion version); diff --git a/src/core/hle/service/ptm/psm.cpp b/src/core/hle/service/ptm/psm.cpp index c2d5fda94..12d154ecf 100644 --- a/src/core/hle/service/ptm/psm.cpp +++ b/src/core/hle/service/ptm/psm.cpp @@ -12,9 +12,6 @@ namespace Service::PSM { -constexpr u32 BATTERY_FULLY_CHARGED = 100; // 100% Full -constexpr u32 BATTERY_CURRENTLY_CHARGING = 1; // Plugged into an official dock - class PSM final : public ServiceFramework<PSM> { public: explicit PSM() : ServiceFramework{"psm"} { @@ -48,20 +45,30 @@ public: private: void GetBatteryChargePercentage(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_PSM, "(STUBBED) called"); + LOG_DEBUG(Service_PSM, "called"); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(BATTERY_FULLY_CHARGED); + rb.Push<u32>(battery_charge_percentage); } void GetChargerType(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_PSM, "(STUBBED) called"); + LOG_DEBUG(Service_PSM, "called"); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(BATTERY_CURRENTLY_CHARGING); + rb.PushEnum(charger_type); } + + enum class ChargerType : u32 { + Unplugged = 0, + RegularCharger = 1, + LowPowerCharger = 2, + Unknown = 3, + }; + + u32 battery_charge_percentage{100}; // 100% + ChargerType charger_type{ChargerType::RegularCharger}; }; void InstallInterfaces(SM::ServiceManager& sm) { diff --git a/src/core/reporter.cpp b/src/core/reporter.cpp index 558cbe6d7..76cfa5a17 100644 --- a/src/core/reporter.cpp +++ b/src/core/reporter.cpp @@ -4,11 +4,12 @@ #include <ctime> #include <fstream> +#include <iomanip> #include <fmt/chrono.h> #include <fmt/format.h> #include <fmt/ostream.h> -#include <json.hpp> +#include <nlohmann/json.hpp> #include "common/file_util.h" #include "common/hex_util.h" diff --git a/src/core/settings.cpp b/src/core/settings.cpp index cd6c257f5..2b0bdc4d3 100644 --- a/src/core/settings.cpp +++ b/src/core/settings.cpp @@ -96,6 +96,7 @@ void LogSettings() { LogSetting("Renderer_UseAsynchronousGpuEmulation", Settings::values.use_asynchronous_gpu_emulation); LogSetting("Renderer_UseVsync", Settings::values.use_vsync); + LogSetting("Renderer_AnisotropicFilteringLevel", Settings::values.max_anisotropy); LogSetting("Audio_OutputEngine", Settings::values.sink_id); LogSetting("Audio_EnableAudioStretching", Settings::values.enable_audio_stretching); LogSetting("Audio_OutputDevice", Settings::values.audio_device_id); diff --git a/src/input_common/main.cpp b/src/input_common/main.cpp index c98c848cf..95e351e24 100644 --- a/src/input_common/main.cpp +++ b/src/input_common/main.cpp @@ -18,7 +18,9 @@ namespace InputCommon { static std::shared_ptr<Keyboard> keyboard; static std::shared_ptr<MotionEmu> motion_emu; +#ifdef HAVE_SDL2 static std::unique_ptr<SDL::State> sdl; +#endif static std::unique_ptr<CemuhookUDP::State> udp; void Init() { @@ -29,7 +31,9 @@ void Init() { motion_emu = std::make_shared<MotionEmu>(); Input::RegisterFactory<Input::MotionDevice>("motion_emu", motion_emu); +#ifdef HAVE_SDL2 sdl = SDL::Init(); +#endif udp = CemuhookUDP::Init(); } @@ -40,7 +44,9 @@ void Shutdown() { Input::UnregisterFactory<Input::AnalogDevice>("analog_from_button"); Input::UnregisterFactory<Input::MotionDevice>("motion_emu"); motion_emu.reset(); +#ifdef HAVE_SDL2 sdl.reset(); +#endif udp.reset(); } diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index ff53282c9..d23c53843 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -8,6 +8,7 @@ add_library(video_core STATIC dma_pusher.h engines/const_buffer_engine_interface.h engines/const_buffer_info.h + engines/engine_interface.h engines/engine_upload.cpp engines/engine_upload.h engines/fermi_2d.cpp diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 510f11089..56e570994 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -4,7 +4,6 @@ #pragma once -#include <array> #include <list> #include <memory> #include <mutex> @@ -89,10 +88,6 @@ public: map->MarkAsWritten(true); MarkRegionAsWritten(map->GetStart(), map->GetEnd() - 1); } - } else { - if (map->IsWritten()) { - WriteBarrier(); - } } return {ToHandle(block), static_cast<u64>(block->GetOffset(cpu_addr))}; @@ -254,8 +249,6 @@ protected: virtual BufferType ToHandle(const OwnerBuffer& storage) = 0; - virtual void WriteBarrier() = 0; - virtual OwnerBuffer CreateBlock(VAddr cpu_addr, std::size_t size) = 0; virtual void UploadBlockData(const OwnerBuffer& buffer, std::size_t offset, std::size_t size, diff --git a/src/video_core/dma_pusher.cpp b/src/video_core/dma_pusher.cpp index 16311f05e..bdc023d54 100644 --- a/src/video_core/dma_pusher.cpp +++ b/src/video_core/dma_pusher.cpp @@ -27,6 +27,8 @@ void DmaPusher::DispatchCalls() { dma_pushbuffer_subindex = 0; + dma_state.is_last_call = true; + while (system.IsPoweredOn()) { if (!Step()) { break; @@ -82,9 +84,11 @@ bool DmaPusher::Step() { index); CallMultiMethod(&command_header.argument, max_write); dma_state.method_count -= max_write; + dma_state.is_last_call = true; index += max_write; continue; } else { + dma_state.is_last_call = dma_state.method_count <= 1; CallMethod(command_header.argument); } @@ -144,12 +148,22 @@ void DmaPusher::SetState(const CommandHeader& command_header) { } void DmaPusher::CallMethod(u32 argument) const { - gpu.CallMethod({dma_state.method, argument, dma_state.subchannel, dma_state.method_count}); + if (dma_state.method < non_puller_methods) { + gpu.CallMethod({dma_state.method, argument, dma_state.subchannel, dma_state.method_count}); + } else { + subchannels[dma_state.subchannel]->CallMethod(dma_state.method, argument, + dma_state.is_last_call); + } } void DmaPusher::CallMultiMethod(const u32* base_start, u32 num_methods) const { - gpu.CallMultiMethod(dma_state.method, dma_state.subchannel, base_start, num_methods, - dma_state.method_count); + if (dma_state.method < non_puller_methods) { + gpu.CallMultiMethod(dma_state.method, dma_state.subchannel, base_start, num_methods, + dma_state.method_count); + } else { + subchannels[dma_state.subchannel]->CallMultiMethod(dma_state.method, base_start, + num_methods, dma_state.method_count); + } } } // namespace Tegra diff --git a/src/video_core/dma_pusher.h b/src/video_core/dma_pusher.h index 6cef71306..e8b714e94 100644 --- a/src/video_core/dma_pusher.h +++ b/src/video_core/dma_pusher.h @@ -4,11 +4,13 @@ #pragma once +#include <array> #include <vector> #include <queue> #include "common/bit_field.h" #include "common/common_types.h" +#include "video_core/engines/engine_interface.h" namespace Core { class System; @@ -69,7 +71,13 @@ public: void DispatchCalls(); + void BindSubchannel(Tegra::Engines::EngineInterface* engine, u32 subchannel_id) { + subchannels[subchannel_id] = engine; + } + private: + static constexpr u32 non_puller_methods = 0x40; + static constexpr u32 max_subchannels = 8; bool Step(); void SetState(const CommandHeader& command_header); @@ -88,6 +96,7 @@ private: u32 method_count; ///< Current method count u32 length_pending; ///< Large NI command length pending bool non_incrementing; ///< Current command's NI flag + bool is_last_call; }; DmaState dma_state{}; @@ -96,6 +105,8 @@ private: GPUVAddr dma_mget{}; ///< main pushbuffer last read address bool ib_enable{true}; ///< IB mode enabled + std::array<Tegra::Engines::EngineInterface*, max_subchannels> subchannels{}; + GPU& gpu; Core::System& system; }; diff --git a/src/video_core/engines/engine_interface.h b/src/video_core/engines/engine_interface.h new file mode 100644 index 000000000..18a9db7e6 --- /dev/null +++ b/src/video_core/engines/engine_interface.h @@ -0,0 +1,22 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <type_traits> +#include "common/common_types.h" + +namespace Tegra::Engines { + +class EngineInterface { +public: + /// Write the value to the register identified by method. + virtual void CallMethod(u32 method, u32 method_argument, bool is_last_call) = 0; + + /// Write multiple values to the register identified by method. + virtual void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) = 0; +}; + +} // namespace Tegra::Engines diff --git a/src/video_core/engines/fermi_2d.cpp b/src/video_core/engines/fermi_2d.cpp index 8a47614d2..ff10ff40d 100644 --- a/src/video_core/engines/fermi_2d.cpp +++ b/src/video_core/engines/fermi_2d.cpp @@ -12,13 +12,13 @@ namespace Tegra::Engines { Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} {} -void Fermi2D::CallMethod(const GPU::MethodCall& method_call) { - ASSERT_MSG(method_call.method < Regs::NUM_REGS, +void Fermi2D::CallMethod(u32 method, u32 method_argument, bool is_last_call) { + ASSERT_MSG(method < Regs::NUM_REGS, "Invalid Fermi2D register, increase the size of the Regs structure"); - regs.reg_array[method_call.method] = method_call.argument; + regs.reg_array[method] = method_argument; - switch (method_call.method) { + switch (method) { // Trigger the surface copy on the last register write. This is blit_src_y, but this is 64-bit, // so trigger on the second 32-bit write. case FERMI2D_REG_INDEX(blit_src_y) + 1: { @@ -30,7 +30,7 @@ void Fermi2D::CallMethod(const GPU::MethodCall& method_call) { void Fermi2D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending) { for (std::size_t i = 0; i < amount; i++) { - CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)}); + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); } } diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h index 939a5966d..8f37d053f 100644 --- a/src/video_core/engines/fermi_2d.h +++ b/src/video_core/engines/fermi_2d.h @@ -10,6 +10,7 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "common/math_util.h" +#include "video_core/engines/engine_interface.h" #include "video_core/gpu.h" namespace Tegra { @@ -31,16 +32,17 @@ namespace Tegra::Engines { #define FERMI2D_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::Fermi2D::Regs, field_name) / sizeof(u32)) -class Fermi2D final { +class Fermi2D final : public EngineInterface { public: explicit Fermi2D(VideoCore::RasterizerInterface& rasterizer); ~Fermi2D() = default; /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; /// Write multiple values to the register identified by method. - void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending); + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; enum class Origin : u32 { Center = 0, diff --git a/src/video_core/engines/kepler_compute.cpp b/src/video_core/engines/kepler_compute.cpp index 00a12175f..f6237fc6a 100644 --- a/src/video_core/engines/kepler_compute.cpp +++ b/src/video_core/engines/kepler_compute.cpp @@ -24,20 +24,19 @@ KeplerCompute::KeplerCompute(Core::System& system, VideoCore::RasterizerInterfac KeplerCompute::~KeplerCompute() = default; -void KeplerCompute::CallMethod(const GPU::MethodCall& method_call) { - ASSERT_MSG(method_call.method < Regs::NUM_REGS, +void KeplerCompute::CallMethod(u32 method, u32 method_argument, bool is_last_call) { + ASSERT_MSG(method < Regs::NUM_REGS, "Invalid KeplerCompute register, increase the size of the Regs structure"); - regs.reg_array[method_call.method] = method_call.argument; + regs.reg_array[method] = method_argument; - switch (method_call.method) { + switch (method) { case KEPLER_COMPUTE_REG_INDEX(exec_upload): { upload_state.ProcessExec(regs.exec_upload.linear != 0); break; } case KEPLER_COMPUTE_REG_INDEX(data_upload): { - const bool is_last_call = method_call.IsLastCall(); - upload_state.ProcessData(method_call.argument, is_last_call); + upload_state.ProcessData(method_argument, is_last_call); if (is_last_call) { system.GPU().Maxwell3D().OnMemoryWrite(); } @@ -54,7 +53,7 @@ void KeplerCompute::CallMethod(const GPU::MethodCall& method_call) { void KeplerCompute::CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending) { for (std::size_t i = 0; i < amount; i++) { - CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)}); + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); } } diff --git a/src/video_core/engines/kepler_compute.h b/src/video_core/engines/kepler_compute.h index fe55fdfd0..18ceedfaf 100644 --- a/src/video_core/engines/kepler_compute.h +++ b/src/video_core/engines/kepler_compute.h @@ -11,6 +11,7 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "video_core/engines/const_buffer_engine_interface.h" +#include "video_core/engines/engine_interface.h" #include "video_core/engines/engine_upload.h" #include "video_core/engines/shader_type.h" #include "video_core/gpu.h" @@ -39,7 +40,7 @@ namespace Tegra::Engines { #define KEPLER_COMPUTE_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::KeplerCompute::Regs, field_name) / sizeof(u32)) -class KeplerCompute final : public ConstBufferEngineInterface { +class KeplerCompute final : public ConstBufferEngineInterface, public EngineInterface { public: explicit KeplerCompute(Core::System& system, VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager); @@ -200,10 +201,11 @@ public: "KeplerCompute LaunchParams has wrong size"); /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; /// Write multiple values to the register identified by method. - void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending); + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; Texture::FullTextureInfo GetTexture(std::size_t offset) const; diff --git a/src/video_core/engines/kepler_memory.cpp b/src/video_core/engines/kepler_memory.cpp index 586ff15dc..dc71b2eec 100644 --- a/src/video_core/engines/kepler_memory.cpp +++ b/src/video_core/engines/kepler_memory.cpp @@ -19,20 +19,19 @@ KeplerMemory::KeplerMemory(Core::System& system, MemoryManager& memory_manager) KeplerMemory::~KeplerMemory() = default; -void KeplerMemory::CallMethod(const GPU::MethodCall& method_call) { - ASSERT_MSG(method_call.method < Regs::NUM_REGS, +void KeplerMemory::CallMethod(u32 method, u32 method_argument, bool is_last_call) { + ASSERT_MSG(method < Regs::NUM_REGS, "Invalid KeplerMemory register, increase the size of the Regs structure"); - regs.reg_array[method_call.method] = method_call.argument; + regs.reg_array[method] = method_argument; - switch (method_call.method) { + switch (method) { case KEPLERMEMORY_REG_INDEX(exec): { upload_state.ProcessExec(regs.exec.linear != 0); break; } case KEPLERMEMORY_REG_INDEX(data): { - const bool is_last_call = method_call.IsLastCall(); - upload_state.ProcessData(method_call.argument, is_last_call); + upload_state.ProcessData(method_argument, is_last_call); if (is_last_call) { system.GPU().Maxwell3D().OnMemoryWrite(); } @@ -44,7 +43,7 @@ void KeplerMemory::CallMethod(const GPU::MethodCall& method_call) { void KeplerMemory::CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending) { for (std::size_t i = 0; i < amount; i++) { - CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)}); + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); } } diff --git a/src/video_core/engines/kepler_memory.h b/src/video_core/engines/kepler_memory.h index bb26fb030..5b7f71a00 100644 --- a/src/video_core/engines/kepler_memory.h +++ b/src/video_core/engines/kepler_memory.h @@ -10,6 +10,7 @@ #include "common/bit_field.h" #include "common/common_funcs.h" #include "common/common_types.h" +#include "video_core/engines/engine_interface.h" #include "video_core/engines/engine_upload.h" #include "video_core/gpu.h" @@ -32,16 +33,17 @@ namespace Tegra::Engines { #define KEPLERMEMORY_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::KeplerMemory::Regs, field_name) / sizeof(u32)) -class KeplerMemory final { +class KeplerMemory final : public EngineInterface { public: KeplerMemory(Core::System& system, MemoryManager& memory_manager); ~KeplerMemory(); /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; /// Write multiple values to the register identified by method. - void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending); + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; struct Regs { static constexpr size_t NUM_REGS = 0x7F; diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index 39e3b66a2..024c9e43b 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -44,6 +44,12 @@ void Maxwell3D::InitializeRegisterDefaults() { viewport.depth_range_near = 0.0f; viewport.depth_range_far = 1.0f; } + for (auto& viewport : regs.viewport_transform) { + viewport.swizzle.x.Assign(Regs::ViewportSwizzle::PositiveX); + viewport.swizzle.y.Assign(Regs::ViewportSwizzle::PositiveY); + viewport.swizzle.z.Assign(Regs::ViewportSwizzle::PositiveZ); + viewport.swizzle.w.Assign(Regs::ViewportSwizzle::PositiveW); + } // Doom and Bomberman seems to use the uninitialized registers and just enable blend // so initialize blend registers with sane values @@ -125,12 +131,10 @@ void Maxwell3D::CallMacroMethod(u32 method, std::size_t num_parameters, const u3 } } -void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { - const u32 method = method_call.method; - +void Maxwell3D::CallMethod(u32 method, u32 method_argument, bool is_last_call) { if (method == cb_data_state.current) { - regs.reg_array[method] = method_call.argument; - ProcessCBData(method_call.argument); + regs.reg_array[method] = method_argument; + ProcessCBData(method_argument); return; } else if (cb_data_state.current != null_cb_data) { FinishCBData(); @@ -153,10 +157,10 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { executing_macro = method; } - macro_params.push_back(method_call.argument); + macro_params.push_back(method_argument); // Call the macro when there are no more parameters in the command buffer - if (method_call.IsLastCall()) { + if (is_last_call) { CallMacroMethod(executing_macro, macro_params.size(), macro_params.data()); macro_params.clear(); } @@ -166,7 +170,7 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { ASSERT_MSG(method < Regs::NUM_REGS, "Invalid Maxwell3D register, increase the size of the Regs structure"); - u32 arg = method_call.argument; + u32 arg = method_argument; // Keep track of the register value in shadow_state when requested. if (shadow_state.shadow_ram_control == Regs::ShadowRamControl::Track || shadow_state.shadow_ram_control == Regs::ShadowRamControl::TrackWithFilter) { @@ -184,8 +188,12 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { } switch (method) { + case MAXWELL3D_REG_INDEX(wait_for_idle): { + rasterizer.WaitForIdle(); + break; + } case MAXWELL3D_REG_INDEX(shadow_ram_control): { - shadow_state.shadow_ram_control = static_cast<Regs::ShadowRamControl>(method_call.argument); + shadow_state.shadow_ram_control = static_cast<Regs::ShadowRamControl>(method_argument); break; } case MAXWELL3D_REG_INDEX(macros.data): { @@ -268,7 +276,6 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { break; } case MAXWELL3D_REG_INDEX(data_upload): { - const bool is_last_call = method_call.IsLastCall(); upload_state.ProcessData(arg, is_last_call); if (is_last_call) { OnMemoryWrite(); @@ -326,7 +333,7 @@ void Maxwell3D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, } default: { for (std::size_t i = 0; i < amount; i++) { - CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)}); + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); } } } @@ -356,16 +363,15 @@ void Maxwell3D::StepInstance(const MMEDrawMode expected_mode, const u32 count) { StepInstance(expected_mode, count); } -void Maxwell3D::CallMethodFromMME(const GPU::MethodCall& method_call) { - const u32 method = method_call.method; +void Maxwell3D::CallMethodFromMME(u32 method, u32 method_argument) { if (mme_inline[method]) { - regs.reg_array[method] = method_call.argument; + regs.reg_array[method] = method_argument; if (method == MAXWELL3D_REG_INDEX(vertex_buffer.count) || method == MAXWELL3D_REG_INDEX(index_array.count)) { const MMEDrawMode expected_mode = method == MAXWELL3D_REG_INDEX(vertex_buffer.count) ? MMEDrawMode::Array : MMEDrawMode::Indexed; - StepInstance(expected_mode, method_call.argument); + StepInstance(expected_mode, method_argument); } else if (method == MAXWELL3D_REG_INDEX(draw.vertex_begin_gl)) { mme_draw.instance_mode = (regs.draw.instance_next != 0) || (regs.draw.instance_cont != 0); @@ -377,7 +383,7 @@ void Maxwell3D::CallMethodFromMME(const GPU::MethodCall& method_call) { if (mme_draw.current_mode != MMEDrawMode::Undefined) { FlushMMEInlineDraw(); } - CallMethod(method_call); + CallMethod(method, method_argument, true); } } diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index 3dfba8197..05dd6b39b 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -19,6 +19,7 @@ #include "common/math_util.h" #include "video_core/engines/const_buffer_engine_interface.h" #include "video_core/engines/const_buffer_info.h" +#include "video_core/engines/engine_interface.h" #include "video_core/engines/engine_upload.h" #include "video_core/engines/shader_type.h" #include "video_core/gpu.h" @@ -48,7 +49,7 @@ namespace Tegra::Engines { #define MAXWELL3D_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::Maxwell3D::Regs, field_name) / sizeof(u32)) -class Maxwell3D final : public ConstBufferEngineInterface { +class Maxwell3D final : public ConstBufferEngineInterface, public EngineInterface { public: explicit Maxwell3D(Core::System& system, VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager); @@ -575,6 +576,17 @@ public: Replay = 3, }; + enum class ViewportSwizzle : u32 { + PositiveX = 0, + NegativeX = 1, + PositiveY = 2, + NegativeY = 3, + PositiveZ = 4, + NegativeZ = 5, + PositiveW = 6, + NegativeW = 7, + }; + struct RenderTargetConfig { u32 address_high; u32 address_low; @@ -618,7 +630,14 @@ public: f32 translate_x; f32 translate_y; f32 translate_z; - INSERT_UNION_PADDING_WORDS(2); + union { + u32 raw; + BitField<0, 3, ViewportSwizzle> x; + BitField<4, 3, ViewportSwizzle> y; + BitField<8, 3, ViewportSwizzle> z; + BitField<12, 3, ViewportSwizzle> w; + } swizzle; + INSERT_UNION_PADDING_WORDS(1); Common::Rectangle<f32> GetRect() const { return { @@ -709,7 +728,9 @@ public: union { struct { - INSERT_UNION_PADDING_WORDS(0x45); + INSERT_UNION_PADDING_WORDS(0x44); + + u32 wait_for_idle; struct { u32 upload_address; @@ -1179,6 +1200,7 @@ public: BitField<0, 1, u32> depth_range_0_1; BitField<3, 1, u32> depth_clamp_near; BitField<4, 1, u32> depth_clamp_far; + BitField<11, 1, u32> depth_clamp_disabled; } view_volume_clip_control; INSERT_UNION_PADDING_WORDS(0x1F); @@ -1357,13 +1379,14 @@ public: u32 GetRegisterValue(u32 method) const; /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; /// Write multiple values to the register identified by method. - void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending); + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; /// Write the value to the register identified by method. - void CallMethodFromMME(const GPU::MethodCall& method_call); + void CallMethodFromMME(u32 method, u32 method_argument); void FlushMMEInlineDraw(); @@ -1535,6 +1558,7 @@ private: static_assert(offsetof(Maxwell3D::Regs, field_name) == position * 4, \ "Field " #field_name " has invalid position") +ASSERT_REG_POSITION(wait_for_idle, 0x44); ASSERT_REG_POSITION(macros, 0x45); ASSERT_REG_POSITION(shadow_ram_control, 0x49); ASSERT_REG_POSITION(upload, 0x60); diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp index 6630005b0..01d7df405 100644 --- a/src/video_core/engines/maxwell_dma.cpp +++ b/src/video_core/engines/maxwell_dma.cpp @@ -17,16 +17,16 @@ namespace Tegra::Engines { MaxwellDMA::MaxwellDMA(Core::System& system, MemoryManager& memory_manager) : system{system}, memory_manager{memory_manager} {} -void MaxwellDMA::CallMethod(const GPU::MethodCall& method_call) { - ASSERT_MSG(method_call.method < Regs::NUM_REGS, +void MaxwellDMA::CallMethod(u32 method, u32 method_argument, bool is_last_call) { + ASSERT_MSG(method < Regs::NUM_REGS, "Invalid MaxwellDMA register, increase the size of the Regs structure"); - regs.reg_array[method_call.method] = method_call.argument; + regs.reg_array[method] = method_argument; #define MAXWELLDMA_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::MaxwellDMA::Regs, field_name) / sizeof(u32)) - switch (method_call.method) { + switch (method) { case MAXWELLDMA_REG_INDEX(exec): { HandleCopy(); break; @@ -39,7 +39,7 @@ void MaxwellDMA::CallMethod(const GPU::MethodCall& method_call) { void MaxwellDMA::CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending) { for (std::size_t i = 0; i < amount; i++) { - CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)}); + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); } } @@ -90,7 +90,47 @@ void MaxwellDMA::HandleCopy() { ASSERT(regs.exec.enable_2d == 1); if (regs.exec.is_dst_linear && !regs.exec.is_src_linear) { + ASSERT(regs.src_params.BlockDepth() == 0); + // Optimized path for micro copies. + if (regs.dst_pitch * regs.y_count < Texture::GetGOBSize() && regs.dst_pitch <= 64) { + const u32 bytes_per_pixel = regs.dst_pitch / regs.x_count; + const std::size_t src_size = Texture::GetGOBSize(); + const std::size_t dst_size = regs.dst_pitch * regs.y_count; + u32 pos_x = regs.src_params.pos_x; + u32 pos_y = regs.src_params.pos_y; + const u64 offset = + Texture::GetGOBOffset(regs.src_params.size_x, regs.src_params.size_y, pos_x, pos_y, + regs.src_params.BlockDepth(), bytes_per_pixel); + const u32 x_in_gob = 64 / bytes_per_pixel; + pos_x = pos_x % x_in_gob; + pos_y = pos_y % 8; + + if (read_buffer.size() < src_size) { + read_buffer.resize(src_size); + } + + if (write_buffer.size() < dst_size) { + write_buffer.resize(dst_size); + } + + if (Settings::IsGPULevelExtreme()) { + memory_manager.ReadBlock(source + offset, read_buffer.data(), src_size); + memory_manager.ReadBlock(dest, write_buffer.data(), dst_size); + } else { + memory_manager.ReadBlockUnsafe(source + offset, read_buffer.data(), src_size); + memory_manager.ReadBlockUnsafe(dest, write_buffer.data(), dst_size); + } + + Texture::UnswizzleSubrect(regs.x_count, regs.y_count, regs.dst_pitch, + regs.src_params.size_x, bytes_per_pixel, read_buffer.data(), + write_buffer.data(), regs.src_params.BlockHeight(), pos_x, + pos_y); + + memory_manager.WriteBlock(dest, write_buffer.data(), dst_size); + + return; + } // If the input is tiled and the output is linear, deswizzle the input and copy it over. const u32 bytes_per_pixel = regs.dst_pitch / regs.x_count; const std::size_t src_size = Texture::CalculateSize( diff --git a/src/video_core/engines/maxwell_dma.h b/src/video_core/engines/maxwell_dma.h index c43ed8194..502dd8509 100644 --- a/src/video_core/engines/maxwell_dma.h +++ b/src/video_core/engines/maxwell_dma.h @@ -10,6 +10,7 @@ #include "common/bit_field.h" #include "common/common_funcs.h" #include "common/common_types.h" +#include "video_core/engines/engine_interface.h" #include "video_core/gpu.h" namespace Core { @@ -27,16 +28,17 @@ namespace Tegra::Engines { * https://github.com/envytools/envytools/blob/master/rnndb/fifo/gk104_copy.xml */ -class MaxwellDMA final { +class MaxwellDMA final : public EngineInterface { public: explicit MaxwellDMA(Core::System& system, MemoryManager& memory_manager); ~MaxwellDMA() = default; /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; /// Write multiple values to the register identified by method. - void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending); + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; struct Regs { static constexpr std::size_t NUM_REGS = 0x1D6; diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h index cde3a26b9..8dae754d4 100644 --- a/src/video_core/engines/shader_bytecode.h +++ b/src/video_core/engines/shader_bytecode.h @@ -814,6 +814,10 @@ union Instruction { } alu_integer; union { + BitField<43, 1, u64> x; + } iadd; + + union { BitField<39, 1, u64> ftz; BitField<32, 1, u64> saturate; BitField<49, 2, HalfMerge> merge; diff --git a/src/video_core/fence_manager.h b/src/video_core/fence_manager.h index dabd1588c..8b2a6a42c 100644 --- a/src/video_core/fence_manager.h +++ b/src/video_core/fence_manager.h @@ -88,7 +88,8 @@ public: } PopAsyncFlushes(); if (current_fence->IsSemaphore()) { - memory_manager.Write<u32>(current_fence->GetAddress(), current_fence->GetPayload()); + memory_manager.template Write<u32>(current_fence->GetAddress(), + current_fence->GetPayload()); } else { gpu.IncrementSyncPoint(current_fence->GetPayload()); } @@ -134,7 +135,8 @@ private: } PopAsyncFlushes(); if (current_fence->IsSemaphore()) { - memory_manager.Write<u32>(current_fence->GetAddress(), current_fence->GetPayload()); + memory_manager.template Write<u32>(current_fence->GetAddress(), + current_fence->GetPayload()); } else { gpu.IncrementSyncPoint(current_fence->GetPayload()); } diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp index b87fd873d..8eb017f65 100644 --- a/src/video_core/gpu.cpp +++ b/src/video_core/gpu.cpp @@ -299,19 +299,21 @@ void GPU::CallEngineMethod(const MethodCall& method_call) { switch (engine) { case EngineID::FERMI_TWOD_A: - fermi_2d->CallMethod(method_call); + fermi_2d->CallMethod(method_call.method, method_call.argument, method_call.IsLastCall()); break; case EngineID::MAXWELL_B: - maxwell_3d->CallMethod(method_call); + maxwell_3d->CallMethod(method_call.method, method_call.argument, method_call.IsLastCall()); break; case EngineID::KEPLER_COMPUTE_B: - kepler_compute->CallMethod(method_call); + kepler_compute->CallMethod(method_call.method, method_call.argument, + method_call.IsLastCall()); break; case EngineID::MAXWELL_DMA_COPY_A: - maxwell_dma->CallMethod(method_call); + maxwell_dma->CallMethod(method_call.method, method_call.argument, method_call.IsLastCall()); break; case EngineID::KEPLER_INLINE_TO_MEMORY_B: - kepler_memory->CallMethod(method_call); + kepler_memory->CallMethod(method_call.method, method_call.argument, + method_call.IsLastCall()); break; default: UNIMPLEMENTED_MSG("Unimplemented engine"); @@ -347,7 +349,27 @@ void GPU::ProcessBindMethod(const MethodCall& method_call) { // Bind the current subchannel to the desired engine id. LOG_DEBUG(HW_GPU, "Binding subchannel {} to engine {}", method_call.subchannel, method_call.argument); - bound_engines[method_call.subchannel] = static_cast<EngineID>(method_call.argument); + const auto engine_id = static_cast<EngineID>(method_call.argument); + bound_engines[method_call.subchannel] = static_cast<EngineID>(engine_id); + switch (engine_id) { + case EngineID::FERMI_TWOD_A: + dma_pusher->BindSubchannel(fermi_2d.get(), method_call.subchannel); + break; + case EngineID::MAXWELL_B: + dma_pusher->BindSubchannel(maxwell_3d.get(), method_call.subchannel); + break; + case EngineID::KEPLER_COMPUTE_B: + dma_pusher->BindSubchannel(kepler_compute.get(), method_call.subchannel); + break; + case EngineID::MAXWELL_DMA_COPY_A: + dma_pusher->BindSubchannel(maxwell_dma.get(), method_call.subchannel); + break; + case EngineID::KEPLER_INLINE_TO_MEMORY_B: + dma_pusher->BindSubchannel(kepler_memory.get(), method_call.subchannel); + break; + default: + UNIMPLEMENTED_MSG("Unimplemented engine {:04X}", static_cast<u32>(engine_id)); + } } void GPU::ProcessSemaphoreTriggerMethod() { diff --git a/src/video_core/macro_interpreter.cpp b/src/video_core/macro_interpreter.cpp index 42031d80a..947364928 100644 --- a/src/video_core/macro_interpreter.cpp +++ b/src/video_core/macro_interpreter.cpp @@ -328,7 +328,7 @@ void MacroInterpreter::SetMethodAddress(u32 address) { } void MacroInterpreter::Send(u32 value) { - maxwell3d.CallMethodFromMME({method_address.address, value}); + maxwell3d.CallMethodFromMME(method_address.address, value); // Increment the method address by the method increment. method_address.address.Assign(method_address.address.Value() + method_address.increment.Value()); diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h index 603f61952..3cbdac8e7 100644 --- a/src/video_core/rasterizer_interface.h +++ b/src/video_core/rasterizer_interface.h @@ -80,6 +80,9 @@ public: /// and invalidated virtual void FlushAndInvalidateRegion(VAddr addr, u64 size) = 0; + /// Notify the host renderer to wait for previous primitive and compute operations. + virtual void WaitForIdle() = 0; + /// Notify the rasterizer to send all written commands to the host GPU. virtual void FlushCommands() = 0; diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp index 4efce0de7..d2cab50bd 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp @@ -51,10 +51,6 @@ Buffer OGLBufferCache::CreateBlock(VAddr cpu_addr, std::size_t size) { return std::make_shared<CachedBufferBlock>(cpu_addr, size); } -void OGLBufferCache::WriteBarrier() { - glMemoryBarrier(GL_SHADER_STORAGE_BARRIER_BIT); -} - GLuint OGLBufferCache::ToHandle(const Buffer& buffer) { return buffer->GetHandle(); } diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h index a74817857..a9e86cfc7 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.h +++ b/src/video_core/renderer_opengl/gl_buffer_cache.h @@ -59,8 +59,6 @@ protected: GLuint ToHandle(const Buffer& buffer) override; - void WriteBarrier() override; - void UploadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, const u8* data) override; diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index 6fe155bcc..69dcf952f 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -59,14 +59,12 @@ constexpr std::size_t NumSupportedVertexAttributes = 16; template <typename Engine, typename Entry> Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry& entry, ShaderType shader_type, std::size_t index = 0) { - if (entry.IsBindless()) { - const Tegra::Texture::TextureHandle tex_handle = - engine.AccessConstBuffer32(shader_type, entry.GetBuffer(), entry.GetOffset()); + if (entry.is_bindless) { + const auto tex_handle = engine.AccessConstBuffer32(shader_type, entry.buffer, entry.offset); return engine.GetTextureInfo(tex_handle); } const auto& gpu_profile = engine.AccessGuestDriverProfile(); - const u32 offset = - entry.GetOffset() + static_cast<u32>(index * gpu_profile.GetTextureHandlerSize()); + const u32 offset = entry.offset + static_cast<u32>(index * gpu_profile.GetTextureHandlerSize()); if constexpr (std::is_same_v<Engine, Tegra::Engines::Maxwell3D>) { return engine.GetStageTexture(shader_type, offset); } else { @@ -348,7 +346,7 @@ void RasterizerOpenGL::ConfigureFramebuffers() { texture_cache.GuardRenderTargets(true); - View depth_surface = texture_cache.GetDepthBufferSurface(); + View depth_surface = texture_cache.GetDepthBufferSurface(true); const auto& regs = gpu.regs; UNIMPLEMENTED_IF(regs.rt_separate_frag_data == 0); @@ -357,7 +355,7 @@ void RasterizerOpenGL::ConfigureFramebuffers() { FramebufferCacheKey key; const auto colors_count = static_cast<std::size_t>(regs.rt_control.count); for (std::size_t index = 0; index < colors_count; ++index) { - View color_surface{texture_cache.GetColorBufferSurface(index)}; + View color_surface{texture_cache.GetColorBufferSurface(index, true)}; if (!color_surface) { continue; } @@ -381,28 +379,52 @@ void RasterizerOpenGL::ConfigureFramebuffers() { glBindFramebuffer(GL_DRAW_FRAMEBUFFER, framebuffer_cache.GetFramebuffer(key)); } -void RasterizerOpenGL::ConfigureClearFramebuffer(bool using_color_fb, bool using_depth_fb, - bool using_stencil_fb) { +void RasterizerOpenGL::ConfigureClearFramebuffer(bool using_color, bool using_depth_stencil) { auto& gpu = system.GPU().Maxwell3D(); const auto& regs = gpu.regs; texture_cache.GuardRenderTargets(true); View color_surface; - if (using_color_fb) { + + if (using_color) { + // Determine if we have to preserve the contents. + // First we have to make sure all clear masks are enabled. + bool preserve_contents = !regs.clear_buffers.R || !regs.clear_buffers.G || + !regs.clear_buffers.B || !regs.clear_buffers.A; const std::size_t index = regs.clear_buffers.RT; - color_surface = texture_cache.GetColorBufferSurface(index); + if (regs.clear_flags.scissor) { + // Then we have to confirm scissor testing clears the whole image. + const auto& scissor = regs.scissor_test[0]; + preserve_contents |= scissor.min_x > 0; + preserve_contents |= scissor.min_y > 0; + preserve_contents |= scissor.max_x < regs.rt[index].width; + preserve_contents |= scissor.max_y < regs.rt[index].height; + } + + color_surface = texture_cache.GetColorBufferSurface(index, preserve_contents); texture_cache.MarkColorBufferInUse(index); } + View depth_surface; - if (using_depth_fb || using_stencil_fb) { - depth_surface = texture_cache.GetDepthBufferSurface(); + if (using_depth_stencil) { + bool preserve_contents = false; + if (regs.clear_flags.scissor) { + // For depth stencil clears we only have to confirm scissor test covers the whole image. + const auto& scissor = regs.scissor_test[0]; + preserve_contents |= scissor.min_x > 0; + preserve_contents |= scissor.min_y > 0; + preserve_contents |= scissor.max_x < regs.zeta_width; + preserve_contents |= scissor.max_y < regs.zeta_height; + } + + depth_surface = texture_cache.GetDepthBufferSurface(preserve_contents); texture_cache.MarkDepthBufferInUse(); } texture_cache.GuardRenderTargets(false); FramebufferCacheKey key; - key.colors[0] = color_surface; - key.zeta = depth_surface; + key.colors[0] = std::move(color_surface); + key.zeta = std::move(depth_surface); state_tracker.NotifyFramebuffer(); glBindFramebuffer(GL_DRAW_FRAMEBUFFER, framebuffer_cache.GetFramebuffer(key)); @@ -422,8 +444,7 @@ void RasterizerOpenGL::Clear() { if (regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B || regs.clear_buffers.A) { use_color = true; - } - if (use_color) { + state_tracker.NotifyColorMask0(); glColorMaski(0, regs.clear_buffers.R != 0, regs.clear_buffers.G != 0, regs.clear_buffers.B != 0, regs.clear_buffers.A != 0); @@ -461,7 +482,7 @@ void RasterizerOpenGL::Clear() { UNIMPLEMENTED_IF(regs.clear_flags.viewport); - ConfigureClearFramebuffer(use_color, use_depth, use_stencil); + ConfigureClearFramebuffer(use_color, use_depth || use_stencil); if (use_color) { glClearBufferfv(GL_COLOR, 0, regs.clear_color); @@ -725,6 +746,17 @@ void RasterizerOpenGL::FlushAndInvalidateRegion(VAddr addr, u64 size) { InvalidateRegion(addr, size); } +void RasterizerOpenGL::WaitForIdle() { + // Place a barrier on everything that is not framebuffer related. + // This is related to another flag that is not currently implemented. + glMemoryBarrier(GL_VERTEX_ATTRIB_ARRAY_BARRIER_BIT | GL_ELEMENT_ARRAY_BARRIER_BIT | + GL_UNIFORM_BARRIER_BIT | GL_TEXTURE_FETCH_BARRIER_BIT | + GL_SHADER_IMAGE_ACCESS_BARRIER_BIT | GL_COMMAND_BARRIER_BIT | + GL_PIXEL_BUFFER_BARRIER_BIT | GL_TEXTURE_UPDATE_BARRIER_BIT | + GL_BUFFER_UPDATE_BARRIER_BIT | GL_TRANSFORM_FEEDBACK_BARRIER_BIT | + GL_SHADER_STORAGE_BARRIER_BIT | GL_QUERY_BUFFER_BARRIER_BIT); +} + void RasterizerOpenGL::FlushCommands() { // Only flush when we have commands queued to OpenGL. if (num_queued_commands == 0) { @@ -833,9 +865,9 @@ void RasterizerOpenGL::SetupDrawGlobalMemory(std::size_t stage_index, const Shad u32 binding = device.GetBaseBindings(stage_index).shader_storage_buffer; for (const auto& entry : shader->GetEntries().global_memory_entries) { - const auto addr{cbufs.const_buffers[entry.GetCbufIndex()].address + entry.GetCbufOffset()}; - const auto gpu_addr{memory_manager.Read<u64>(addr)}; - const auto size{memory_manager.Read<u32>(addr + 8)}; + const GPUVAddr addr{cbufs.const_buffers[entry.cbuf_index].address + entry.cbuf_offset}; + const GPUVAddr gpu_addr{memory_manager.Read<u64>(addr)}; + const u32 size{memory_manager.Read<u32>(addr + 8)}; SetupGlobalMemory(binding++, entry, gpu_addr, size); } } @@ -847,7 +879,7 @@ void RasterizerOpenGL::SetupComputeGlobalMemory(const Shader& kernel) { u32 binding = 0; for (const auto& entry : kernel->GetEntries().global_memory_entries) { - const auto addr{cbufs[entry.GetCbufIndex()].Address() + entry.GetCbufOffset()}; + const auto addr{cbufs[entry.cbuf_index].Address() + entry.cbuf_offset}; const auto gpu_addr{memory_manager.Read<u64>(addr)}; const auto size{memory_manager.Read<u32>(addr + 8)}; SetupGlobalMemory(binding++, entry, gpu_addr, size); @@ -858,7 +890,7 @@ void RasterizerOpenGL::SetupGlobalMemory(u32 binding, const GlobalMemoryEntry& e GPUVAddr gpu_addr, std::size_t size) { const auto alignment{device.GetShaderStorageBufferAlignment()}; const auto [ssbo, buffer_offset] = - buffer_cache.UploadMemory(gpu_addr, size, alignment, entry.IsWritten()); + buffer_cache.UploadMemory(gpu_addr, size, alignment, entry.is_written); glBindBufferRange(GL_SHADER_STORAGE_BUFFER, binding, ssbo, buffer_offset, static_cast<GLsizeiptr>(size)); } @@ -869,7 +901,7 @@ void RasterizerOpenGL::SetupDrawTextures(std::size_t stage_index, const Shader& u32 binding = device.GetBaseBindings(stage_index).sampler; for (const auto& entry : shader->GetEntries().samplers) { const auto shader_type = static_cast<ShaderType>(stage_index); - for (std::size_t i = 0; i < entry.Size(); ++i) { + for (std::size_t i = 0; i < entry.size; ++i) { const auto texture = GetTextureInfo(maxwell3d, entry, shader_type, i); SetupTexture(binding++, texture, entry); } @@ -881,7 +913,7 @@ void RasterizerOpenGL::SetupComputeTextures(const Shader& kernel) { const auto& compute = system.GPU().KeplerCompute(); u32 binding = 0; for (const auto& entry : kernel->GetEntries().samplers) { - for (std::size_t i = 0; i < entry.Size(); ++i) { + for (std::size_t i = 0; i < entry.size; ++i) { const auto texture = GetTextureInfo(compute, entry, ShaderType::Compute, i); SetupTexture(binding++, texture, entry); } @@ -938,7 +970,7 @@ void RasterizerOpenGL::SetupImage(u32 binding, const Tegra::Texture::TICEntry& t if (!tic.IsBuffer()) { view->ApplySwizzle(tic.x_source, tic.y_source, tic.z_source, tic.w_source); } - if (entry.IsWritten()) { + if (entry.is_written) { view->MarkAsModified(texture_cache.Tick()); } glBindImageTexture(binding, view->GetTexture(), 0, GL_TRUE, 0, GL_READ_WRITE, @@ -987,6 +1019,14 @@ void RasterizerOpenGL::SyncViewport() { const GLdouble near_depth = src.translate_z - src.scale_z * reduce_z; const GLdouble far_depth = src.translate_z + src.scale_z; glDepthRangeIndexed(static_cast<GLuint>(i), near_depth, far_depth); + + if (!GLAD_GL_NV_viewport_swizzle) { + continue; + } + glViewportSwizzleNV(static_cast<GLuint>(i), MaxwellToGL::ViewportSwizzle(src.swizzle.x), + MaxwellToGL::ViewportSwizzle(src.swizzle.y), + MaxwellToGL::ViewportSwizzle(src.swizzle.z), + MaxwellToGL::ViewportSwizzle(src.swizzle.w)); } } } @@ -999,11 +1039,7 @@ void RasterizerOpenGL::SyncDepthClamp() { } flags[Dirty::DepthClampEnabled] = false; - const auto& state = gpu.regs.view_volume_clip_control; - UNIMPLEMENTED_IF_MSG(state.depth_clamp_far != state.depth_clamp_near, - "Unimplemented depth clamp separation!"); - - oglEnable(GL_DEPTH_CLAMP, state.depth_clamp_far || state.depth_clamp_near); + oglEnable(GL_DEPTH_CLAMP, gpu.regs.view_volume_clip_control.depth_clamp_disabled == 0); } void RasterizerOpenGL::SyncClipEnabled(u32 clip_mask) { diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index ebd2173eb..b94c65907 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -75,6 +75,7 @@ public: void SignalSyncPoint(u32 value) override; void ReleaseFences() override; void FlushAndInvalidateRegion(VAddr addr, u64 size) override; + void WaitForIdle() override; void FlushCommands() override; void TickFrame() override; bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, @@ -95,7 +96,8 @@ private: /// Configures the color and depth framebuffer states. void ConfigureFramebuffers(); - void ConfigureClearFramebuffer(bool using_color_fb, bool using_depth_fb, bool using_stencil_fb); + /// Configures the color and depth framebuffer for clearing. + void ConfigureClearFramebuffer(bool using_color, bool using_depth_stencil); /// Configures the current constbuffers to use for the draw command. void SetupDrawConstBuffers(std::size_t stage_index, const Shader& shader); diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index 0cd3ad7e1..99fd4ae2c 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -870,13 +870,13 @@ private: for (const auto& sampler : ir.GetSamplers()) { const std::string name = GetSampler(sampler); const std::string description = fmt::format("layout (binding = {}) uniform", binding); - binding += sampler.IsIndexed() ? sampler.Size() : 1; + binding += sampler.is_indexed ? sampler.size : 1; std::string sampler_type = [&]() { - if (sampler.IsBuffer()) { + if (sampler.is_buffer) { return "samplerBuffer"; } - switch (sampler.GetType()) { + switch (sampler.type) { case Tegra::Shader::TextureType::Texture1D: return "sampler1D"; case Tegra::Shader::TextureType::Texture2D: @@ -890,17 +890,17 @@ private: return "sampler2D"; } }(); - if (sampler.IsArray()) { + if (sampler.is_array) { sampler_type += "Array"; } - if (sampler.IsShadow()) { + if (sampler.is_shadow) { sampler_type += "Shadow"; } - if (!sampler.IsIndexed()) { + if (!sampler.is_indexed) { code.AddLine("{} {} {};", description, sampler_type, name); } else { - code.AddLine("{} {} {}[{}];", description, sampler_type, name, sampler.Size()); + code.AddLine("{} {} {}[{}];", description, sampler_type, name, sampler.size); } } if (!ir.GetSamplers().empty()) { @@ -946,14 +946,14 @@ private: u32 binding = device.GetBaseBindings(stage).image; for (const auto& image : ir.GetImages()) { std::string qualifier = "coherent volatile"; - if (image.IsRead() && !image.IsWritten()) { + if (image.is_read && !image.is_written) { qualifier += " readonly"; - } else if (image.IsWritten() && !image.IsRead()) { + } else if (image.is_written && !image.is_read) { qualifier += " writeonly"; } - const char* format = image.IsAtomic() ? "r32ui, " : ""; - const char* type_declaration = GetImageTypeDeclaration(image.GetType()); + const char* format = image.is_atomic ? "r32ui, " : ""; + const char* type_declaration = GetImageTypeDeclaration(image.type); code.AddLine("layout ({}binding = {}) {} uniform uimage{} {};", format, binding++, qualifier, type_declaration, GetImage(image)); } @@ -1337,8 +1337,8 @@ private: ASSERT(meta); const std::size_t count = operation.GetOperandsCount(); - const bool has_array = meta->sampler.IsArray(); - const bool has_shadow = meta->sampler.IsShadow(); + const bool has_array = meta->sampler.is_array; + const bool has_shadow = meta->sampler.is_shadow; std::string expr = "texture" + function_suffix; if (!meta->aoffi.empty()) { @@ -1346,7 +1346,7 @@ private: } else if (!meta->ptp.empty()) { expr += "Offsets"; } - if (!meta->sampler.IsIndexed()) { + if (!meta->sampler.is_indexed) { expr += '(' + GetSampler(meta->sampler) + ", "; } else { expr += '(' + GetSampler(meta->sampler) + '[' + Visit(meta->index).AsUint() + "], "; @@ -1870,6 +1870,14 @@ private: return GenerateBinaryInfix(operation, ">=", Type::Bool, type, type); } + Expression LogicalAddCarry(Operation operation) { + const std::string carry = code.GenerateTemporary(); + code.AddLine("uint {};", carry); + code.AddLine("uaddCarry({}, {}, {});", VisitOperand(operation, 0).AsUint(), + VisitOperand(operation, 1).AsUint(), carry); + return {fmt::format("({} != 0)", carry), Type::Bool}; + } + Expression LogicalFIsNan(Operation operation) { return GenerateUnary(operation, "isnan", Type::Bool, Type::Float); } @@ -1974,7 +1982,7 @@ private: std::string expr = GenerateTexture( operation, "", {TextureOffset{}, TextureArgument{Type::Float, meta->bias}}); - if (meta->sampler.IsShadow()) { + if (meta->sampler.is_shadow) { expr = "vec4(" + expr + ')'; } return {expr + GetSwizzle(meta->element), Type::Float}; @@ -1986,7 +1994,7 @@ private: std::string expr = GenerateTexture( operation, "Lod", {TextureArgument{Type::Float, meta->lod}, TextureOffset{}}); - if (meta->sampler.IsShadow()) { + if (meta->sampler.is_shadow) { expr = "vec4(" + expr + ')'; } return {expr + GetSwizzle(meta->element), Type::Float}; @@ -1995,11 +2003,11 @@ private: Expression TextureGather(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); - const auto type = meta.sampler.IsShadow() ? Type::Float : Type::Int; - const bool separate_dc = meta.sampler.IsShadow(); + const auto type = meta.sampler.is_shadow ? Type::Float : Type::Int; + const bool separate_dc = meta.sampler.is_shadow; std::vector<TextureIR> ir; - if (meta.sampler.IsShadow()) { + if (meta.sampler.is_shadow) { ir = {TextureOffset{}}; } else { ir = {TextureOffset{}, TextureArgument{type, meta.component}}; @@ -2044,7 +2052,7 @@ private: constexpr std::array constructors = {"int", "ivec2", "ivec3", "ivec4"}; const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); ASSERT(meta); - UNIMPLEMENTED_IF(meta->sampler.IsArray()); + UNIMPLEMENTED_IF(meta->sampler.is_array); const std::size_t count = operation.GetOperandsCount(); std::string expr = "texelFetch("; @@ -2065,7 +2073,7 @@ private: } expr += ')'; - if (meta->lod && !meta->sampler.IsBuffer()) { + if (meta->lod && !meta->sampler.is_buffer) { expr += ", "; expr += Visit(meta->lod).AsInt(); } @@ -2076,12 +2084,10 @@ private: } Expression TextureGradient(Operation operation) { - const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); - ASSERT(meta); - + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); std::string expr = GenerateTexture(operation, "Grad", {TextureDerivates{}, TextureOffset{}}); - return {std::move(expr) + GetSwizzle(meta->element), Type::Float}; + return {std::move(expr) + GetSwizzle(meta.element), Type::Float}; } Expression ImageLoad(Operation operation) { @@ -2441,6 +2447,8 @@ private: &GLSLDecompiler::LogicalNotEqual<Type::Uint>, &GLSLDecompiler::LogicalGreaterEqual<Type::Uint>, + &GLSLDecompiler::LogicalAddCarry, + &GLSLDecompiler::Logical2HLessThan<false>, &GLSLDecompiler::Logical2HEqual<false>, &GLSLDecompiler::Logical2HLessEqual<false>, @@ -2598,11 +2606,11 @@ private: } std::string GetSampler(const Sampler& sampler) const { - return AppendSuffix(static_cast<u32>(sampler.GetIndex()), "sampler"); + return AppendSuffix(sampler.index, "sampler"); } std::string GetImage(const Image& image) const { - return AppendSuffix(static_cast<u32>(image.GetIndex()), "image"); + return AppendSuffix(image.index, "image"); } std::string AppendSuffix(u32 index, std::string_view name) const { diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.h b/src/video_core/renderer_opengl/gl_shader_decompiler.h index e7dbd810c..e8a178764 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.h +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.h @@ -33,36 +33,19 @@ public: } private: - u32 index{}; + u32 index = 0; }; -class GlobalMemoryEntry { -public: - explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset, bool is_read, bool is_written) +struct GlobalMemoryEntry { + constexpr explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset, bool is_read, + bool is_written) : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset}, is_read{is_read}, is_written{ is_written} {} - u32 GetCbufIndex() const { - return cbuf_index; - } - - u32 GetCbufOffset() const { - return cbuf_offset; - } - - bool IsRead() const { - return is_read; - } - - bool IsWritten() const { - return is_written; - } - -private: - u32 cbuf_index{}; - u32 cbuf_offset{}; - bool is_read{}; - bool is_written{}; + u32 cbuf_index = 0; + u32 cbuf_offset = 0; + bool is_read = false; + bool is_written = false; }; struct ShaderEntries { diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index 2c0c77c28..994ae98eb 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -503,5 +503,10 @@ inline GLenum PolygonMode(Maxwell::PolygonMode polygon_mode) { return GL_FILL; } +inline GLenum ViewportSwizzle(Maxwell::ViewportSwizzle swizzle) { + // Enumeration order matches register order. We can convert it arithmetically. + return GL_VIEWPORT_SWIZZLE_POSITIVE_X_NV + static_cast<GLenum>(swizzle); +} + } // namespace MaxwellToGL } // namespace OpenGL diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp index a7f256ff9..6cead3a28 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> #include <cstring> #include <tuple> @@ -81,7 +82,7 @@ void FixedPipelineState::Rasterizer::Fill(const Maxwell& regs) noexcept { primitive_restart_enable.Assign(regs.primitive_restart.enabled != 0 ? 1 : 0); cull_enable.Assign(regs.cull_test_enabled != 0 ? 1 : 0); depth_bias_enable.Assign(enabled_lut[POLYGON_OFFSET_ENABLE_LUT[topology_index]] != 0 ? 1 : 0); - depth_clamp_enable.Assign(clip.depth_clamp_near == 1 || clip.depth_clamp_far == 1 ? 1 : 0); + depth_clamp_disabled.Assign(regs.view_volume_clip_control.depth_clamp_disabled.Value()); ndc_minus_one_to_one.Assign(regs.depth_mode == Maxwell::DepthMode::MinusOneToOne ? 1 : 0); cull_face.Assign(PackCullFace(regs.cull_face)); front_face.Assign(packed_front_face); @@ -101,6 +102,12 @@ void FixedPipelineState::ColorBlending::Fill(const Maxwell& regs) noexcept { } } +void FixedPipelineState::ViewportSwizzles::Fill(const Maxwell& regs) noexcept { + const auto& transform = regs.viewport_transform; + std::transform(transform.begin(), transform.end(), swizzles.begin(), + [](const auto& viewport) { return static_cast<u16>(viewport.swizzle.raw); }); +} + void FixedPipelineState::BlendingAttachment::Fill(const Maxwell& regs, std::size_t index) { const auto& mask = regs.color_mask[regs.color_mask_common ? 0 : index]; @@ -144,6 +151,7 @@ void FixedPipelineState::Fill(const Maxwell& regs) { rasterizer.Fill(regs); depth_stencil.Fill(regs); color_blending.Fill(regs); + viewport_swizzles.Fill(regs); } std::size_t FixedPipelineState::Hash() const noexcept { diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.h b/src/video_core/renderer_vulkan/fixed_pipeline_state.h index 77188b862..cecaee48d 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.h +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.h @@ -153,7 +153,7 @@ struct FixedPipelineState { BitField<4, 1, u32> primitive_restart_enable; BitField<5, 1, u32> cull_enable; BitField<6, 1, u32> depth_bias_enable; - BitField<7, 1, u32> depth_clamp_enable; + BitField<7, 1, u32> depth_clamp_disabled; BitField<8, 1, u32> ndc_minus_one_to_one; BitField<9, 2, u32> cull_face; BitField<11, 1, u32> front_face; @@ -233,10 +233,17 @@ struct FixedPipelineState { void Fill(const Maxwell& regs) noexcept; }; + struct ViewportSwizzles { + std::array<u16, Maxwell::NumViewports> swizzles; + + void Fill(const Maxwell& regs) noexcept; + }; + VertexInput vertex_input; Rasterizer rasterizer; DepthStencil depth_stencil; ColorBlending color_blending; + ViewportSwizzles viewport_swizzles; void Fill(const Maxwell& regs); diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp index cd652dcc9..12be691a5 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp @@ -346,8 +346,6 @@ VkFormat VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttrib break; case Maxwell::VertexAttribute::Type::SignedInt: switch (size) { - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: - return VK_FORMAT_R16G16B16A16_SINT; case Maxwell::VertexAttribute::Size::Size_8: return VK_FORMAT_R8_SINT; case Maxwell::VertexAttribute::Size::Size_8_8: @@ -356,8 +354,22 @@ VkFormat VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttrib return VK_FORMAT_R8G8B8_SINT; case Maxwell::VertexAttribute::Size::Size_8_8_8_8: return VK_FORMAT_R8G8B8A8_SINT; + case Maxwell::VertexAttribute::Size::Size_16: + return VK_FORMAT_R16_SINT; + case Maxwell::VertexAttribute::Size::Size_16_16: + return VK_FORMAT_R16G16_SINT; + case Maxwell::VertexAttribute::Size::Size_16_16_16: + return VK_FORMAT_R16G16B16_SINT; + case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + return VK_FORMAT_R16G16B16A16_SINT; case Maxwell::VertexAttribute::Size::Size_32: return VK_FORMAT_R32_SINT; + case Maxwell::VertexAttribute::Size::Size_32_32: + return VK_FORMAT_R32G32_SINT; + case Maxwell::VertexAttribute::Size::Size_32_32_32: + return VK_FORMAT_R32G32B32_SINT; + case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + return VK_FORMAT_R32G32B32A32_SINT; default: break; } @@ -673,4 +685,27 @@ VkComponentSwizzle SwizzleSource(Tegra::Texture::SwizzleSource swizzle) { return {}; } +VkViewportCoordinateSwizzleNV ViewportSwizzle(Maxwell::ViewportSwizzle swizzle) { + switch (swizzle) { + case Maxwell::ViewportSwizzle::PositiveX: + return VK_VIEWPORT_COORDINATE_SWIZZLE_POSITIVE_X_NV; + case Maxwell::ViewportSwizzle::NegativeX: + return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_X_NV; + case Maxwell::ViewportSwizzle::PositiveY: + return VK_VIEWPORT_COORDINATE_SWIZZLE_POSITIVE_Y_NV; + case Maxwell::ViewportSwizzle::NegativeY: + return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_Y_NV; + case Maxwell::ViewportSwizzle::PositiveZ: + return VK_VIEWPORT_COORDINATE_SWIZZLE_POSITIVE_Z_NV; + case Maxwell::ViewportSwizzle::NegativeZ: + return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_Z_NV; + case Maxwell::ViewportSwizzle::PositiveW: + return VK_VIEWPORT_COORDINATE_SWIZZLE_POSITIVE_W_NV; + case Maxwell::ViewportSwizzle::NegativeW: + return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_W_NV; + } + UNREACHABLE_MSG("Invalid swizzle={}", static_cast<int>(swizzle)); + return {}; +} + } // namespace Vulkan::MaxwellToVK diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.h b/src/video_core/renderer_vulkan/maxwell_to_vk.h index 81bce4c6c..7e213452f 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.h +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.h @@ -59,4 +59,6 @@ VkCullModeFlags CullFace(Maxwell::CullFace cull_face); VkComponentSwizzle SwizzleSource(Tegra::Texture::SwizzleSource swizzle); +VkViewportCoordinateSwizzleNV ViewportSwizzle(Maxwell::ViewportSwizzle swizzle); + } // namespace Vulkan::MaxwellToVK diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.cpp b/src/video_core/renderer_vulkan/renderer_vulkan.cpp index 04532f8f8..59b441943 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.cpp +++ b/src/video_core/renderer_vulkan/renderer_vulkan.cpp @@ -12,15 +12,12 @@ #include <fmt/format.h> -#include "common/assert.h" #include "common/dynamic_library.h" #include "common/logging/log.h" #include "common/telemetry.h" #include "core/core.h" #include "core/core_timing.h" #include "core/frontend/emu_window.h" -#include "core/memory.h" -#include "core/perf_stats.h" #include "core/settings.h" #include "core/telemetry_session.h" #include "video_core/gpu.h" diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.h b/src/video_core/renderer_vulkan/renderer_vulkan.h index 18270909b..522b5bff8 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.h +++ b/src/video_core/renderer_vulkan/renderer_vulkan.h @@ -5,7 +5,6 @@ #pragma once #include <memory> -#include <optional> #include <string> #include <vector> diff --git a/src/video_core/renderer_vulkan/vk_blit_screen.h b/src/video_core/renderer_vulkan/vk_blit_screen.h index 5eb544aea..243640fab 100644 --- a/src/video_core/renderer_vulkan/vk_blit_screen.h +++ b/src/video_core/renderer_vulkan/vk_blit_screen.h @@ -4,7 +4,6 @@ #pragma once -#include <array> #include <memory> #include <tuple> diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp index 81e1de2be..5b494da8c 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp @@ -5,11 +5,7 @@ #include <algorithm> #include <cstring> #include <memory> -#include <optional> -#include <tuple> -#include "common/assert.h" -#include "common/bit_util.h" #include "core/core.h" #include "video_core/renderer_vulkan/vk_buffer_cache.h" #include "video_core/renderer_vulkan/vk_device.h" diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h index 3cd2e2774..a54583e7d 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.h +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h @@ -5,14 +5,11 @@ #pragma once #include <memory> -#include <unordered_map> -#include <vector> #include "common/common_types.h" #include "video_core/buffer_cache/buffer_cache.h" #include "video_core/rasterizer_cache.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" #include "video_core/renderer_vulkan/vk_stream_buffer.h" #include "video_core/renderer_vulkan/wrapper.h" @@ -55,8 +52,6 @@ public: protected: VkBuffer ToHandle(const Buffer& buffer) override; - void WriteBarrier() override {} - Buffer CreateBlock(VAddr cpu_addr, std::size_t size) override; void UploadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.cpp b/src/video_core/renderer_vulkan/vk_compute_pass.cpp index 7b0268033..da71e710c 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pass.cpp @@ -6,7 +6,7 @@ #include <memory> #include <optional> #include <utility> -#include <vector> + #include "common/alignment.h" #include "common/assert.h" #include "common/common_types.h" diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.h b/src/video_core/renderer_vulkan/vk_compute_pass.h index 26bf834de..230b526bc 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.h +++ b/src/video_core/renderer_vulkan/vk_compute_pass.h @@ -6,7 +6,7 @@ #include <optional> #include <utility> -#include <vector> + #include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp index 52566bb79..8e1b46277 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp @@ -2,14 +2,12 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <memory> #include <vector> #include "video_core/renderer_vulkan/vk_compute_pipeline.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.h b/src/video_core/renderer_vulkan/vk_compute_pipeline.h index 33b9af29e..6e2f22a4a 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.h +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.h @@ -4,8 +4,6 @@ #pragma once -#include <memory> - #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" diff --git a/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp b/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp index e9d528aa6..890fd52cf 100644 --- a/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp +++ b/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp @@ -2,7 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <memory> #include <vector> #include "common/common_types.h" diff --git a/src/video_core/renderer_vulkan/vk_descriptor_pool.h b/src/video_core/renderer_vulkan/vk_descriptor_pool.h index ab40c70f0..9efa66bef 100644 --- a/src/video_core/renderer_vulkan/vk_descriptor_pool.h +++ b/src/video_core/renderer_vulkan/vk_descriptor_pool.h @@ -4,10 +4,8 @@ #pragma once -#include <memory> #include <vector> -#include "common/common_types.h" #include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/wrapper.h" diff --git a/src/video_core/renderer_vulkan/vk_device.cpp b/src/video_core/renderer_vulkan/vk_device.cpp index 83e25a467..f0c491d00 100644 --- a/src/video_core/renderer_vulkan/vk_device.cpp +++ b/src/video_core/renderer_vulkan/vk_device.cpp @@ -4,7 +4,6 @@ #include <bitset> #include <chrono> -#include <cstdlib> #include <optional> #include <string_view> #include <thread> @@ -262,6 +261,10 @@ bool VKDevice::Create() { LOG_INFO(Render_Vulkan, "Device doesn't support float16 natively"); } + if (!nv_viewport_swizzle) { + LOG_INFO(Render_Vulkan, "Device doesn't support viewport swizzles"); + } + VkPhysicalDeviceUniformBufferStandardLayoutFeaturesKHR std430_layout; if (khr_uniform_buffer_standard_layout) { std430_layout.sType = @@ -295,6 +298,17 @@ bool VKDevice::Create() { LOG_INFO(Render_Vulkan, "Device doesn't support transform feedbacks"); } + VkPhysicalDeviceCustomBorderColorFeaturesEXT custom_border; + if (ext_custom_border_color) { + custom_border.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_CUSTOM_BORDER_COLOR_FEATURES_EXT; + custom_border.pNext = nullptr; + custom_border.customBorderColors = VK_TRUE; + custom_border.customBorderColorWithoutFormat = VK_TRUE; + SetNext(next, custom_border); + } else { + LOG_INFO(Render_Vulkan, "Device doesn't support custom border colors"); + } + if (!ext_depth_range_unrestricted) { LOG_INFO(Render_Vulkan, "Device doesn't support depth range unrestricted"); } @@ -522,7 +536,9 @@ std::vector<const char*> VKDevice::LoadExtensions() { bool has_khr_shader_float16_int8{}; bool has_ext_subgroup_size_control{}; bool has_ext_transform_feedback{}; + bool has_ext_custom_border_color{}; for (const auto& extension : physical.EnumerateDeviceExtensionProperties()) { + Test(extension, nv_viewport_swizzle, VK_NV_VIEWPORT_SWIZZLE_EXTENSION_NAME, true); Test(extension, khr_uniform_buffer_standard_layout, VK_KHR_UNIFORM_BUFFER_STANDARD_LAYOUT_EXTENSION_NAME, true); Test(extension, has_khr_shader_float16_int8, VK_KHR_SHADER_FLOAT16_INT8_EXTENSION_NAME, @@ -536,6 +552,8 @@ std::vector<const char*> VKDevice::LoadExtensions() { false); Test(extension, has_ext_transform_feedback, VK_EXT_TRANSFORM_FEEDBACK_EXTENSION_NAME, false); + Test(extension, has_ext_custom_border_color, VK_EXT_CUSTOM_BORDER_COLOR_EXTENSION_NAME, + false); if (Settings::values.renderer_debug) { Test(extension, nv_device_diagnostics_config, VK_NV_DEVICE_DIAGNOSTICS_CONFIG_EXTENSION_NAME, true); @@ -608,6 +626,19 @@ std::vector<const char*> VKDevice::LoadExtensions() { } } + if (has_ext_custom_border_color) { + VkPhysicalDeviceCustomBorderColorFeaturesEXT border_features; + border_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_CUSTOM_BORDER_COLOR_FEATURES_EXT; + border_features.pNext = nullptr; + features.pNext = &border_features; + physical.GetFeatures2KHR(features); + + if (border_features.customBorderColors && border_features.customBorderColorWithoutFormat) { + extensions.push_back(VK_EXT_CUSTOM_BORDER_COLOR_EXTENSION_NAME); + ext_custom_border_color = true; + } + } + return extensions; } diff --git a/src/video_core/renderer_vulkan/vk_device.h b/src/video_core/renderer_vulkan/vk_device.h index c8640762d..6b9227b09 100644 --- a/src/video_core/renderer_vulkan/vk_device.h +++ b/src/video_core/renderer_vulkan/vk_device.h @@ -147,6 +147,11 @@ public: return is_formatless_image_load_supported; } + /// Returns true if the device supports VK_NV_viewport_swizzle. + bool IsNvViewportSwizzleSupported() const { + return nv_viewport_swizzle; + } + /// Returns true if the device supports VK_EXT_scalar_block_layout. bool IsKhrUniformBufferStandardLayoutSupported() const { return khr_uniform_buffer_standard_layout; @@ -172,6 +177,11 @@ public: return ext_transform_feedback; } + /// Returns true if the device supports VK_EXT_custom_border_color. + bool IsExtCustomBorderColorSupported() const { + return ext_custom_border_color; + } + /// Returns the vendor name reported from Vulkan. std::string_view GetVendorName() const { return vendor_name; @@ -222,11 +232,13 @@ private: bool is_float16_supported{}; ///< Support for float16 arithmetics. bool is_warp_potentially_bigger{}; ///< Host warp size can be bigger than guest. bool is_formatless_image_load_supported{}; ///< Support for shader image read without format. + bool nv_viewport_swizzle{}; ///< Support for VK_NV_viewport_swizzle. bool khr_uniform_buffer_standard_layout{}; ///< Support for std430 on UBOs. bool ext_index_type_uint8{}; ///< Support for VK_EXT_index_type_uint8. bool ext_depth_range_unrestricted{}; ///< Support for VK_EXT_depth_range_unrestricted. bool ext_shader_viewport_index_layer{}; ///< Support for VK_EXT_shader_viewport_index_layer. bool ext_transform_feedback{}; ///< Support for VK_EXT_transform_feedback. + bool ext_custom_border_color{}; ///< Support for VK_EXT_custom_border_color. bool nv_device_diagnostics_config{}; ///< Support for VK_NV_device_diagnostics_config. // Telemetry parameters diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp index 45bd1fc6c..5beea6a03 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp @@ -2,11 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> #include <array> #include <cstring> #include <vector> -#include "common/assert.h" #include "common/common_types.h" #include "common/microprofile.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" @@ -51,6 +51,23 @@ bool SupportsPrimitiveRestart(VkPrimitiveTopology topology) { topology) == std::end(unsupported_topologies); } +VkViewportSwizzleNV UnpackViewportSwizzle(u16 swizzle) { + union { + u32 raw; + BitField<0, 3, Maxwell::ViewportSwizzle> x; + BitField<4, 3, Maxwell::ViewportSwizzle> y; + BitField<8, 3, Maxwell::ViewportSwizzle> z; + BitField<12, 3, Maxwell::ViewportSwizzle> w; + } const unpacked{swizzle}; + + VkViewportSwizzleNV result; + result.x = MaxwellToVK::ViewportSwizzle(unpacked.x); + result.y = MaxwellToVK::ViewportSwizzle(unpacked.y); + result.z = MaxwellToVK::ViewportSwizzle(unpacked.z); + result.w = MaxwellToVK::ViewportSwizzle(unpacked.w); + return result; +} + } // Anonymous namespace VKGraphicsPipeline::VKGraphicsPipeline(const VKDevice& device, VKScheduler& scheduler, @@ -163,6 +180,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa const auto& ds = fixed_state.depth_stencil; const auto& cd = fixed_state.color_blending; const auto& rs = fixed_state.rasterizer; + const auto& viewport_swizzles = fixed_state.viewport_swizzles.swizzles; std::vector<VkVertexInputBindingDescription> vertex_bindings; std::vector<VkVertexInputBindingDivisorDescriptionEXT> vertex_binding_divisors; @@ -245,11 +263,24 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa viewport_ci.scissorCount = Maxwell::NumViewports; viewport_ci.pScissors = nullptr; + std::array<VkViewportSwizzleNV, Maxwell::NumViewports> swizzles; + std::transform(viewport_swizzles.begin(), viewport_swizzles.end(), swizzles.begin(), + UnpackViewportSwizzle); + VkPipelineViewportSwizzleStateCreateInfoNV swizzle_ci; + swizzle_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_SWIZZLE_STATE_CREATE_INFO_NV; + swizzle_ci.pNext = nullptr; + swizzle_ci.flags = 0; + swizzle_ci.viewportCount = Maxwell::NumViewports; + swizzle_ci.pViewportSwizzles = swizzles.data(); + if (device.IsNvViewportSwizzleSupported()) { + viewport_ci.pNext = &swizzle_ci; + } + VkPipelineRasterizationStateCreateInfo rasterization_ci; rasterization_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO; rasterization_ci.pNext = nullptr; rasterization_ci.flags = 0; - rasterization_ci.depthClampEnable = rs.depth_clamp_enable; + rasterization_ci.depthClampEnable = rs.depth_clamp_disabled == 0 ? VK_TRUE : VK_FALSE; rasterization_ci.rasterizerDiscardEnable = VK_FALSE; rasterization_ci.polygonMode = VK_POLYGON_MODE_FILL; rasterization_ci.cullMode = diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.h b/src/video_core/renderer_vulkan/vk_graphics_pipeline.h index 7aba70960..a1d699a6c 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.h +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.h @@ -5,16 +5,13 @@ #pragma once #include <array> -#include <memory> #include <optional> -#include <unordered_map> #include <vector> #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" #include "video_core/renderer_vulkan/wrapper.h" diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp index 9b703a2f0..fe45ed269 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp @@ -22,7 +22,6 @@ #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" #include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" #include "video_core/renderer_vulkan/wrapper.h" @@ -85,7 +84,7 @@ void AddBindings(std::vector<VkDescriptorSetLayoutBinding>& bindings, u32& bindi u32 count = 1; if constexpr (descriptor_type == VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER) { // Combined image samplers can be arrayed. - count = container[i].Size(); + count = container[i].size; } VkDescriptorSetLayoutBinding& entry = bindings.emplace_back(); entry.binding = binding++; @@ -362,7 +361,7 @@ void AddEntry(std::vector<VkDescriptorUpdateTemplateEntry>& template_entries, u3 if constexpr (descriptor_type == COMBINED_IMAGE_SAMPLER) { for (u32 i = 0; i < count; ++i) { - const u32 num_samplers = container[i].Size(); + const u32 num_samplers = container[i].size; VkDescriptorUpdateTemplateEntry& entry = template_entries.emplace_back(); entry.dstBinding = binding; entry.dstArrayElement = 0; diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.h b/src/video_core/renderer_vulkan/vk_pipeline_cache.h index ebddafb73..0b5796fef 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.h +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.h @@ -21,13 +21,11 @@ #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/vk_graphics_pipeline.h" #include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" #include "video_core/renderer_vulkan/wrapper.h" #include "video_core/shader/memory_util.h" #include "video_core/shader/registry.h" #include "video_core/shader/shader_ir.h" -#include "video_core/surface.h" namespace Core { class System; diff --git a/src/video_core/renderer_vulkan/vk_query_cache.cpp b/src/video_core/renderer_vulkan/vk_query_cache.cpp index 813f7c162..bc91c48cc 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_query_cache.cpp @@ -4,7 +4,6 @@ #include <algorithm> #include <cstddef> -#include <cstdint> #include <utility> #include <vector> diff --git a/src/video_core/renderer_vulkan/vk_query_cache.h b/src/video_core/renderer_vulkan/vk_query_cache.h index b63784f4b..40119e6d3 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.h +++ b/src/video_core/renderer_vulkan/vk_query_cache.h @@ -5,7 +5,6 @@ #pragma once #include <cstddef> -#include <cstdint> #include <memory> #include <utility> #include <vector> diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index c821b1229..8b009fc22 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -9,14 +9,12 @@ #include <vector> #include <boost/container/static_vector.hpp> -#include <boost/functional/hash.hpp> #include "common/alignment.h" #include "common/assert.h" #include "common/logging/log.h" #include "common/microprofile.h" #include "core/core.h" -#include "core/memory.h" #include "core/settings.h" #include "video_core/engines/kepler_compute.h" #include "video_core/engines/maxwell_3d.h" @@ -119,14 +117,13 @@ template <typename Engine, typename Entry> Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry& entry, std::size_t stage, std::size_t index = 0) { const auto stage_type = static_cast<Tegra::Engines::ShaderType>(stage); - if (entry.IsBindless()) { - const Tegra::Texture::TextureHandle tex_handle = - engine.AccessConstBuffer32(stage_type, entry.GetBuffer(), entry.GetOffset()); + if (entry.is_bindless) { + const auto tex_handle = engine.AccessConstBuffer32(stage_type, entry.buffer, entry.offset); return engine.GetTextureInfo(tex_handle); } const auto& gpu_profile = engine.AccessGuestDriverProfile(); const u32 entry_offset = static_cast<u32>(index * gpu_profile.GetTextureHandlerSize()); - const u32 offset = entry.GetOffset() + entry_offset; + const u32 offset = entry.offset + entry_offset; if constexpr (std::is_same_v<Engine, Tegra::Engines::Maxwell3D>) { return engine.GetStageTexture(stage_type, offset); } else { @@ -302,7 +299,7 @@ RasterizerVulkan::RasterizerVulkan(Core::System& system, Core::Frontend::EmuWind buffer_cache(*this, system, device, memory_manager, scheduler, staging_pool), sampler_cache(device), fence_manager(system, *this, device, scheduler, texture_cache, buffer_cache, query_cache), - query_cache(system, *this, device, scheduler) { + query_cache(system, *this, device, scheduler), wfi_event{device.GetLogical().CreateEvent()} { scheduler.SetQueryCache(query_cache); } @@ -576,6 +573,26 @@ void RasterizerVulkan::FlushAndInvalidateRegion(VAddr addr, u64 size) { InvalidateRegion(addr, size); } +void RasterizerVulkan::WaitForIdle() { + // Everything but wait pixel operations. This intentionally includes FRAGMENT_SHADER_BIT because + // fragment shaders can still write storage buffers. + VkPipelineStageFlags flags = + VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT | VK_PIPELINE_STAGE_VERTEX_INPUT_BIT | + VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT | + VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT | + VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT | VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | + VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT | VK_PIPELINE_STAGE_TRANSFER_BIT; + if (device.IsExtTransformFeedbackSupported()) { + flags |= VK_PIPELINE_STAGE_TRANSFORM_FEEDBACK_BIT_EXT; + } + + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([event = *wfi_event, flags](vk::CommandBuffer cmdbuf) { + cmdbuf.SetEvent(event, flags); + cmdbuf.WaitEvents(event, flags, VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT, {}, {}, {}); + }); +} + void RasterizerVulkan::FlushCommands() { if (draw_counter > 0) { draw_counter = 0; @@ -656,7 +673,7 @@ RasterizerVulkan::Texceptions RasterizerVulkan::UpdateAttachments() { Texceptions texceptions; for (std::size_t rt = 0; rt < Maxwell::NumRenderTargets; ++rt) { if (update_rendertargets) { - color_attachments[rt] = texture_cache.GetColorBufferSurface(rt); + color_attachments[rt] = texture_cache.GetColorBufferSurface(rt, true); } if (color_attachments[rt] && WalkAttachmentOverlaps(*color_attachments[rt])) { texceptions[rt] = true; @@ -664,7 +681,7 @@ RasterizerVulkan::Texceptions RasterizerVulkan::UpdateAttachments() { } if (update_rendertargets) { - zeta_attachment = texture_cache.GetDepthBufferSurface(); + zeta_attachment = texture_cache.GetDepthBufferSurface(true); } if (zeta_attachment && WalkAttachmentOverlaps(*zeta_attachment)) { texceptions[ZETA_TEXCEPTION_INDEX] = true; @@ -896,6 +913,9 @@ void RasterizerVulkan::SetupVertexArrays(FixedPipelineState::VertexInput& vertex void RasterizerVulkan::SetupIndexBuffer(BufferBindings& buffer_bindings, DrawParameters& params, bool is_indexed) { + if (params.num_vertices == 0) { + return; + } const auto& regs = system.GPU().Maxwell3D().regs; switch (regs.draw.topology) { case Maxwell::PrimitiveTopology::Quads: { @@ -971,7 +991,7 @@ void RasterizerVulkan::SetupGraphicsTextures(const ShaderEntries& entries, std:: MICROPROFILE_SCOPE(Vulkan_Textures); const auto& gpu = system.GPU().Maxwell3D(); for (const auto& entry : entries.samplers) { - for (std::size_t i = 0; i < entry.Size(); ++i) { + for (std::size_t i = 0; i < entry.size; ++i) { const auto texture = GetTextureInfo(gpu, entry, stage, i); SetupTexture(texture, entry); } @@ -1023,7 +1043,7 @@ void RasterizerVulkan::SetupComputeTextures(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_Textures); const auto& gpu = system.GPU().KeplerCompute(); for (const auto& entry : entries.samplers) { - for (std::size_t i = 0; i < entry.Size(); ++i) { + for (std::size_t i = 0; i < entry.size; ++i) { const auto texture = GetTextureInfo(gpu, entry, ComputeShaderIndex, i); SetupTexture(texture, entry); } @@ -1105,7 +1125,7 @@ void RasterizerVulkan::SetupTexture(const Tegra::Texture::FullTextureInfo& textu void RasterizerVulkan::SetupImage(const Tegra::Texture::TICEntry& tic, const ImageEntry& entry) { auto view = texture_cache.GetImageSurface(tic, entry); - if (entry.IsWritten()) { + if (entry.is_written) { view->MarkAsModified(texture_cache.Tick()); } diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.h b/src/video_core/renderer_vulkan/vk_rasterizer.h index d41a7929e..0ed0e48c6 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.h +++ b/src/video_core/renderer_vulkan/vk_rasterizer.h @@ -14,7 +14,6 @@ #include <boost/functional/hash.hpp> #include "common/common_types.h" -#include "video_core/memory_manager.h" #include "video_core/rasterizer_accelerated.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" @@ -127,6 +126,7 @@ public: void SignalSyncPoint(u32 value) override; void ReleaseFences() override; void FlushAndInvalidateRegion(VAddr addr, u64 size) override; + void WaitForIdle() override; void FlushCommands() override; void TickFrame() override; bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, @@ -276,6 +276,7 @@ private: vk::Buffer default_buffer; VKMemoryCommit default_buffer_commit; + vk::Event wfi_event; std::array<View, Maxwell::NumRenderTargets> color_attachments; View zeta_attachment; diff --git a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp index 07bbcf520..e6f2fa553 100644 --- a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp @@ -2,11 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cstring> -#include <optional> #include <unordered_map> -#include "common/assert.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" #include "video_core/renderer_vulkan/vk_sampler_cache.h" #include "video_core/renderer_vulkan/wrapper.h" @@ -42,9 +39,18 @@ VKSamplerCache::VKSamplerCache(const VKDevice& device) : device{device} {} VKSamplerCache::~VKSamplerCache() = default; vk::Sampler VKSamplerCache::CreateSampler(const Tegra::Texture::TSCEntry& tsc) const { + const bool arbitrary_borders = device.IsExtCustomBorderColorSupported(); + const std::array color = tsc.GetBorderColor(); + + VkSamplerCustomBorderColorCreateInfoEXT border; + border.sType = VK_STRUCTURE_TYPE_SAMPLER_CUSTOM_BORDER_COLOR_CREATE_INFO_EXT; + border.pNext = nullptr; + border.format = VK_FORMAT_UNDEFINED; + std::memcpy(&border.customBorderColor, color.data(), sizeof(color)); + VkSamplerCreateInfo ci; ci.sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO; - ci.pNext = nullptr; + ci.pNext = arbitrary_borders ? &border : nullptr; ci.flags = 0; ci.magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter); ci.minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter); @@ -59,7 +65,7 @@ vk::Sampler VKSamplerCache::CreateSampler(const Tegra::Texture::TSCEntry& tsc) c ci.compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func); ci.minLod = tsc.GetMinLod(); ci.maxLod = tsc.GetMaxLod(); - ci.borderColor = ConvertBorderColor(tsc.GetBorderColor()); + ci.borderColor = arbitrary_borders ? VK_BORDER_COLOR_INT_CUSTOM_EXT : ConvertBorderColor(color); ci.unnormalizedCoordinates = VK_FALSE; return device.GetLogical().CreateSampler(ci); } diff --git a/src/video_core/renderer_vulkan/vk_scheduler.cpp b/src/video_core/renderer_vulkan/vk_scheduler.cpp index ae7ba3eb5..82ec9180e 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.cpp +++ b/src/video_core/renderer_vulkan/vk_scheduler.cpp @@ -8,7 +8,6 @@ #include <thread> #include <utility> -#include "common/assert.h" #include "common/microprofile.h" #include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_query_cache.h" diff --git a/src/video_core/renderer_vulkan/vk_scheduler.h b/src/video_core/renderer_vulkan/vk_scheduler.h index 82a8adc69..970a65566 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.h +++ b/src/video_core/renderer_vulkan/vk_scheduler.h @@ -7,7 +7,6 @@ #include <atomic> #include <condition_variable> #include <memory> -#include <optional> #include <stack> #include <thread> #include <utility> diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp index aaa138f52..18678968c 100644 --- a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp @@ -103,8 +103,8 @@ struct GenericVaryingDescription { }; spv::Dim GetSamplerDim(const Sampler& sampler) { - ASSERT(!sampler.IsBuffer()); - switch (sampler.GetType()) { + ASSERT(!sampler.is_buffer); + switch (sampler.type) { case Tegra::Shader::TextureType::Texture1D: return spv::Dim::Dim1D; case Tegra::Shader::TextureType::Texture2D: @@ -114,13 +114,13 @@ spv::Dim GetSamplerDim(const Sampler& sampler) { case Tegra::Shader::TextureType::TextureCube: return spv::Dim::Cube; default: - UNIMPLEMENTED_MSG("Unimplemented sampler type={}", static_cast<u32>(sampler.GetType())); + UNIMPLEMENTED_MSG("Unimplemented sampler type={}", static_cast<int>(sampler.type)); return spv::Dim::Dim2D; } } std::pair<spv::Dim, bool> GetImageDim(const Image& image) { - switch (image.GetType()) { + switch (image.type) { case Tegra::Shader::ImageType::Texture1D: return {spv::Dim::Dim1D, false}; case Tegra::Shader::ImageType::TextureBuffer: @@ -134,7 +134,7 @@ std::pair<spv::Dim, bool> GetImageDim(const Image& image) { case Tegra::Shader::ImageType::Texture3D: return {spv::Dim::Dim3D, false}; default: - UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<u32>(image.GetType())); + UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<int>(image.type)); return {spv::Dim::Dim2D, false}; } } @@ -879,11 +879,11 @@ private: u32 DeclareTexelBuffers(u32 binding) { for (const auto& sampler : ir.GetSamplers()) { - if (!sampler.IsBuffer()) { + if (!sampler.is_buffer) { continue; } - ASSERT(!sampler.IsArray()); - ASSERT(!sampler.IsShadow()); + ASSERT(!sampler.is_array); + ASSERT(!sampler.is_shadow); constexpr auto dim = spv::Dim::Buffer; constexpr int depth = 0; @@ -894,23 +894,23 @@ private: const Id image_type = TypeImage(t_float, dim, depth, arrayed, ms, sampled, format); const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, image_type); const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant); - AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.GetIndex()))); + AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.index))); Decorate(id, spv::Decoration::Binding, binding++); Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); - texel_buffers.emplace(sampler.GetIndex(), TexelBuffer{image_type, id}); + texel_buffers.emplace(sampler.index, TexelBuffer{image_type, id}); } return binding; } u32 DeclareSamplers(u32 binding) { for (const auto& sampler : ir.GetSamplers()) { - if (sampler.IsBuffer()) { + if (sampler.is_buffer) { continue; } const auto dim = GetSamplerDim(sampler); - const int depth = sampler.IsShadow() ? 1 : 0; - const int arrayed = sampler.IsArray() ? 1 : 0; + const int depth = sampler.is_shadow ? 1 : 0; + const int arrayed = sampler.is_array ? 1 : 0; constexpr bool ms = false; constexpr int sampled = 1; constexpr auto format = spv::ImageFormat::Unknown; @@ -918,17 +918,17 @@ private: const Id sampler_type = TypeSampledImage(image_type); const Id sampler_pointer_type = TypePointer(spv::StorageClass::UniformConstant, sampler_type); - const Id type = sampler.IsIndexed() - ? TypeArray(sampler_type, Constant(t_uint, sampler.Size())) + const Id type = sampler.is_indexed + ? TypeArray(sampler_type, Constant(t_uint, sampler.size)) : sampler_type; const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, type); const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant); - AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.GetIndex()))); + AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.index))); Decorate(id, spv::Decoration::Binding, binding++); Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); - sampled_images.emplace(sampler.GetIndex(), SampledImage{image_type, sampler_type, - sampler_pointer_type, id}); + sampled_images.emplace( + sampler.index, SampledImage{image_type, sampler_type, sampler_pointer_type, id}); } return binding; } @@ -943,17 +943,17 @@ private: const Id image_type = TypeImage(t_uint, dim, depth, arrayed, ms, sampled, format, {}); const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, image_type); const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant); - AddGlobalVariable(Name(id, fmt::format("image_{}", image.GetIndex()))); + AddGlobalVariable(Name(id, fmt::format("image_{}", image.index))); Decorate(id, spv::Decoration::Binding, binding++); Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); - if (image.IsRead() && !image.IsWritten()) { + if (image.is_read && !image.is_written) { Decorate(id, spv::Decoration::NonWritable); - } else if (image.IsWritten() && !image.IsRead()) { + } else if (image.is_written && !image.is_read) { Decorate(id, spv::Decoration::NonReadable); } - images.emplace(static_cast<u32>(image.GetIndex()), StorageImage{image_type, id}); + images.emplace(image.index, StorageImage{image_type, id}); } return binding; } @@ -1584,6 +1584,15 @@ private: return {OpCompositeConstruct(t_half, low, high), Type::HalfFloat}; } + Expression LogicalAddCarry(Operation operation) { + const Id op_a = AsUint(Visit(operation[0])); + const Id op_b = AsUint(Visit(operation[1])); + + const Id result = OpIAddCarry(TypeStruct({t_uint, t_uint}), op_a, op_b); + const Id carry = OpCompositeExtract(t_uint, result, 1); + return {OpINotEqual(t_bool, carry, Constant(t_uint, 0)), Type::Bool}; + } + Expression LogicalAssign(Operation operation) { const Node& dest = operation[0]; const Node& src = operation[1]; @@ -1611,11 +1620,11 @@ private: Id GetTextureSampler(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); - ASSERT(!meta.sampler.IsBuffer()); + ASSERT(!meta.sampler.is_buffer); - const auto& entry = sampled_images.at(meta.sampler.GetIndex()); + const auto& entry = sampled_images.at(meta.sampler.index); Id sampler = entry.variable; - if (meta.sampler.IsIndexed()) { + if (meta.sampler.is_indexed) { const Id index = AsInt(Visit(meta.index)); sampler = OpAccessChain(entry.sampler_pointer_type, sampler, index); } @@ -1624,8 +1633,8 @@ private: Id GetTextureImage(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); - const u32 index = meta.sampler.GetIndex(); - if (meta.sampler.IsBuffer()) { + const u32 index = meta.sampler.index; + if (meta.sampler.is_buffer) { const auto& entry = texel_buffers.at(index); return OpLoad(entry.image_type, entry.image); } else { @@ -1636,7 +1645,7 @@ private: Id GetImage(Operation operation) { const auto& meta = std::get<MetaImage>(operation.GetMeta()); - const auto entry = images.at(meta.image.GetIndex()); + const auto entry = images.at(meta.image.index); return OpLoad(entry.image_type, entry.image); } @@ -1652,7 +1661,7 @@ private: } if (const auto meta = std::get_if<MetaTexture>(&operation.GetMeta())) { // Add array coordinate for textures - if (meta->sampler.IsArray()) { + if (meta->sampler.is_array) { Id array = AsInt(Visit(meta->array)); if (type == Type::Float) { array = OpConvertSToF(t_float, array); @@ -1758,7 +1767,7 @@ private: operands.push_back(GetOffsetCoordinates(operation)); } - if (meta.sampler.IsShadow()) { + if (meta.sampler.is_shadow) { const Id dref = AsFloat(Visit(meta.depth_compare)); return {OpImageSampleDrefExplicitLod(t_float, sampler, coords, dref, mask, operands), Type::Float}; @@ -1773,7 +1782,7 @@ private: const Id coords = GetCoordinates(operation, Type::Float); Id texture{}; - if (meta.sampler.IsShadow()) { + if (meta.sampler.is_shadow) { texture = OpImageDrefGather(t_float4, GetTextureSampler(operation), coords, AsFloat(Visit(meta.depth_compare))); } else { @@ -1800,8 +1809,8 @@ private: } const Id lod = AsUint(Visit(operation[0])); - const std::size_t coords_count = [&]() { - switch (const auto type = meta.sampler.GetType(); type) { + const std::size_t coords_count = [&meta] { + switch (const auto type = meta.sampler.type) { case Tegra::Shader::TextureType::Texture1D: return 1; case Tegra::Shader::TextureType::Texture2D: @@ -1810,7 +1819,7 @@ private: case Tegra::Shader::TextureType::Texture3D: return 3; default: - UNREACHABLE_MSG("Invalid texture type={}", static_cast<u32>(type)); + UNREACHABLE_MSG("Invalid texture type={}", static_cast<int>(type)); return 2; } }(); @@ -1853,7 +1862,7 @@ private: const Id image = GetTextureImage(operation); const Id coords = GetCoordinates(operation, Type::Int); Id fetch; - if (meta.lod && !meta.sampler.IsBuffer()) { + if (meta.lod && !meta.sampler.is_buffer) { fetch = OpImageFetch(t_float4, image, coords, spv::ImageOperandsMask::Lod, AsInt(Visit(meta.lod))); } else { @@ -2518,6 +2527,8 @@ private: &SPIRVDecompiler::Binary<&Module::OpINotEqual, Type::Bool, Type::Uint>, &SPIRVDecompiler::Binary<&Module::OpUGreaterThanEqual, Type::Bool, Type::Uint>, + &SPIRVDecompiler::LogicalAddCarry, + &SPIRVDecompiler::Binary<&Module::OpFOrdLessThan, Type::Bool2, Type::HalfFloat>, &SPIRVDecompiler::Binary<&Module::OpFOrdEqual, Type::Bool2, Type::HalfFloat>, &SPIRVDecompiler::Binary<&Module::OpFOrdLessThanEqual, Type::Bool2, Type::HalfFloat>, @@ -2969,7 +2980,7 @@ ShaderEntries GenerateShaderEntries(const VideoCommon::Shader::ShaderIR& ir) { entries.global_buffers.emplace_back(base.cbuf_index, base.cbuf_offset, usage.is_written); } for (const auto& sampler : ir.GetSamplers()) { - if (sampler.IsBuffer()) { + if (sampler.is_buffer) { entries.texel_buffers.emplace_back(sampler); } else { entries.samplers.emplace_back(sampler); diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.h b/src/video_core/renderer_vulkan/vk_shader_decompiler.h index ffea4709e..f4c05ac3c 100644 --- a/src/video_core/renderer_vulkan/vk_shader_decompiler.h +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.h @@ -5,11 +5,7 @@ #pragma once #include <array> -#include <bitset> -#include <memory> #include <set> -#include <type_traits> -#include <utility> #include <vector> #include "common/common_types.h" diff --git a/src/video_core/renderer_vulkan/vk_shader_util.cpp b/src/video_core/renderer_vulkan/vk_shader_util.cpp index 784839327..112df9c71 100644 --- a/src/video_core/renderer_vulkan/vk_shader_util.cpp +++ b/src/video_core/renderer_vulkan/vk_shader_util.cpp @@ -4,8 +4,7 @@ #include <cstring> #include <memory> -#include <vector> -#include "common/alignment.h" + #include "common/assert.h" #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_device.h" diff --git a/src/video_core/renderer_vulkan/vk_shader_util.h b/src/video_core/renderer_vulkan/vk_shader_util.h index be38d6697..d1d3f3cae 100644 --- a/src/video_core/renderer_vulkan/vk_shader_util.h +++ b/src/video_core/renderer_vulkan/vk_shader_util.h @@ -4,7 +4,6 @@ #pragma once -#include <vector> #include "common/common_types.h" #include "video_core/renderer_vulkan/wrapper.h" diff --git a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h index faf6418fd..3c4901437 100644 --- a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h +++ b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h @@ -5,8 +5,6 @@ #pragma once #include <climits> -#include <unordered_map> -#include <utility> #include <vector> #include "common/common_types.h" diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index de4c23120..55f43e61b 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -10,11 +10,9 @@ #include <variant> #include <vector> -#include "common/alignment.h" #include "common/assert.h" #include "common/common_types.h" #include "core/core.h" -#include "core/memory.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/morton.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" @@ -26,7 +24,6 @@ #include "video_core/renderer_vulkan/vk_texture_cache.h" #include "video_core/renderer_vulkan/wrapper.h" #include "video_core/surface.h" -#include "video_core/textures/convert.h" namespace Vulkan { diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h index 115595f28..f211ccb1e 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.h +++ b/src/video_core/renderer_vulkan/vk_texture_cache.h @@ -7,19 +7,13 @@ #include <memory> #include <unordered_map> -#include "common/assert.h" #include "common/common_types.h" -#include "common/logging/log.h" -#include "common/math_util.h" -#include "video_core/gpu.h" -#include "video_core/rasterizer_cache.h" #include "video_core/renderer_vulkan/vk_image.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/wrapper.h" #include "video_core/texture_cache/surface_base.h" #include "video_core/texture_cache/texture_cache.h" -#include "video_core/textures/decoders.h" namespace Core { class System; diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.h b/src/video_core/renderer_vulkan/vk_update_descriptor.h index 6ba2c9997..cc7e3dff4 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.h +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.h @@ -4,7 +4,6 @@ #pragma once -#include <type_traits> #include <variant> #include <boost/container/static_vector.hpp> diff --git a/src/video_core/renderer_vulkan/wrapper.cpp b/src/video_core/renderer_vulkan/wrapper.cpp index 7f5bc1404..2ce9b0626 100644 --- a/src/video_core/renderer_vulkan/wrapper.cpp +++ b/src/video_core/renderer_vulkan/wrapper.cpp @@ -87,6 +87,7 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkCmdSetStencilReference); X(vkCmdSetStencilWriteMask); X(vkCmdSetViewport); + X(vkCmdWaitEvents); X(vkCreateBuffer); X(vkCreateBufferView); X(vkCreateCommandPool); diff --git a/src/video_core/renderer_vulkan/wrapper.h b/src/video_core/renderer_vulkan/wrapper.h index bda16a2cb..98937a77a 100644 --- a/src/video_core/renderer_vulkan/wrapper.h +++ b/src/video_core/renderer_vulkan/wrapper.h @@ -205,6 +205,7 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkCmdSetStencilReference vkCmdSetStencilReference; PFN_vkCmdSetStencilWriteMask vkCmdSetStencilWriteMask; PFN_vkCmdSetViewport vkCmdSetViewport; + PFN_vkCmdWaitEvents vkCmdWaitEvents; PFN_vkCreateBuffer vkCreateBuffer; PFN_vkCreateBufferView vkCreateBufferView; PFN_vkCreateCommandPool vkCreateCommandPool; @@ -958,6 +959,15 @@ public: dld->vkCmdSetEvent(handle, event, stage_flags); } + void WaitEvents(Span<VkEvent> events, VkPipelineStageFlags src_stage_mask, + VkPipelineStageFlags dst_stage_mask, Span<VkMemoryBarrier> memory_barriers, + Span<VkBufferMemoryBarrier> buffer_barriers, + Span<VkImageMemoryBarrier> image_barriers) const noexcept { + dld->vkCmdWaitEvents(handle, events.size(), events.data(), src_stage_mask, dst_stage_mask, + memory_barriers.size(), memory_barriers.data(), buffer_barriers.size(), + buffer_barriers.data(), image_barriers.size(), image_barriers.data()); + } + void BindTransformFeedbackBuffersEXT(u32 first, u32 count, const VkBuffer* buffers, const VkDeviceSize* offsets, const VkDeviceSize* sizes) const noexcept { diff --git a/src/video_core/shader/decode.cpp b/src/video_core/shader/decode.cpp index 1167ff4ec..a75a5cc63 100644 --- a/src/video_core/shader/decode.cpp +++ b/src/video_core/shader/decode.cpp @@ -32,11 +32,11 @@ void DeduceTextureHandlerSize(VideoCore::GuestDriverProfile& gpu_driver, u32 count{}; std::vector<u32> bound_offsets; for (const auto& sampler : used_samplers) { - if (sampler.IsBindless()) { + if (sampler.is_bindless) { continue; } ++count; - bound_offsets.emplace_back(sampler.GetOffset()); + bound_offsets.emplace_back(sampler.offset); } if (count > 1) { gpu_driver.DeduceTextureHandlerSize(std::move(bound_offsets)); @@ -46,14 +46,14 @@ void DeduceTextureHandlerSize(VideoCore::GuestDriverProfile& gpu_driver, std::optional<u32> TryDeduceSamplerSize(const Sampler& sampler_to_deduce, VideoCore::GuestDriverProfile& gpu_driver, const std::list<Sampler>& used_samplers) { - const u32 base_offset = sampler_to_deduce.GetOffset(); + const u32 base_offset = sampler_to_deduce.offset; u32 max_offset{std::numeric_limits<u32>::max()}; for (const auto& sampler : used_samplers) { - if (sampler.IsBindless()) { + if (sampler.is_bindless) { continue; } - if (sampler.GetOffset() > base_offset) { - max_offset = std::min(sampler.GetOffset(), max_offset); + if (sampler.offset > base_offset) { + max_offset = std::min(sampler.offset, max_offset); } } if (max_offset == std::numeric_limits<u32>::max()) { @@ -353,14 +353,14 @@ void ShaderIR::PostDecode() { return; } for (auto& sampler : used_samplers) { - if (!sampler.IsIndexed()) { + if (!sampler.is_indexed) { continue; } if (const auto size = TryDeduceSamplerSize(sampler, gpu_driver, used_samplers)) { - sampler.SetSize(*size); + sampler.size = *size; } else { LOG_CRITICAL(HW_GPU, "Failed to deduce size of indexed sampler"); - sampler.SetSize(1); + sampler.size = 1; } } } diff --git a/src/video_core/shader/decode/arithmetic_integer.cpp b/src/video_core/shader/decode/arithmetic_integer.cpp index 9af8c606d..a041519b7 100644 --- a/src/video_core/shader/decode/arithmetic_integer.cpp +++ b/src/video_core/shader/decode/arithmetic_integer.cpp @@ -35,15 +35,38 @@ u32 ShaderIR::DecodeArithmeticInteger(NodeBlock& bb, u32 pc) { case OpCode::Id::IADD_C: case OpCode::Id::IADD_R: case OpCode::Id::IADD_IMM: { - UNIMPLEMENTED_IF_MSG(instr.alu.saturate_d, "IADD saturation not implemented"); + UNIMPLEMENTED_IF_MSG(instr.alu.saturate_d, "IADD.SAT"); + UNIMPLEMENTED_IF_MSG(instr.iadd.x && instr.generates_cc, "IADD.X Rd.CC"); op_a = GetOperandAbsNegInteger(op_a, false, instr.alu_integer.negate_a, true); op_b = GetOperandAbsNegInteger(op_b, false, instr.alu_integer.negate_b, true); - const Node value = Operation(OperationCode::IAdd, PRECISE, op_a, op_b); + Node value = Operation(OperationCode::UAdd, op_a, op_b); - SetInternalFlagsFromInteger(bb, value, instr.generates_cc); - SetRegister(bb, instr.gpr0, value); + if (instr.iadd.x) { + Node carry = GetInternalFlag(InternalFlag::Carry); + Node x = Operation(OperationCode::Select, std::move(carry), Immediate(1), Immediate(0)); + value = Operation(OperationCode::UAdd, std::move(value), std::move(x)); + } + + if (instr.generates_cc) { + const Node i0 = Immediate(0); + + Node zero = Operation(OperationCode::LogicalIEqual, value, i0); + Node sign = Operation(OperationCode::LogicalILessThan, value, i0); + Node carry = Operation(OperationCode::LogicalAddCarry, op_a, op_b); + + Node pos_a = Operation(OperationCode::LogicalIGreaterThan, op_a, i0); + Node pos_b = Operation(OperationCode::LogicalIGreaterThan, op_b, i0); + Node pos = Operation(OperationCode::LogicalAnd, std::move(pos_a), std::move(pos_b)); + Node overflow = Operation(OperationCode::LogicalAnd, pos, sign); + + SetInternalFlag(bb, InternalFlag::Zero, std::move(zero)); + SetInternalFlag(bb, InternalFlag::Sign, std::move(sign)); + SetInternalFlag(bb, InternalFlag::Carry, std::move(carry)); + SetInternalFlag(bb, InternalFlag::Overflow, std::move(overflow)); + } + SetRegister(bb, instr.gpr0, std::move(value)); break; } case OpCode::Id::IADD3_C: diff --git a/src/video_core/shader/decode/image.cpp b/src/video_core/shader/decode/image.cpp index 85ee9aa5e..60b6ad72a 100644 --- a/src/video_core/shader/decode/image.cpp +++ b/src/video_core/shader/decode/image.cpp @@ -485,11 +485,10 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { Image& ShaderIR::GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType type) { const auto offset = static_cast<u32>(image.index.Value()); - const auto it = - std::find_if(std::begin(used_images), std::end(used_images), - [offset](const Image& entry) { return entry.GetOffset() == offset; }); + const auto it = std::find_if(std::begin(used_images), std::end(used_images), + [offset](const Image& entry) { return entry.offset == offset; }); if (it != std::end(used_images)) { - ASSERT(!it->IsBindless() && it->GetType() == it->GetType()); + ASSERT(!it->is_bindless && it->type == type); return *it; } @@ -505,13 +504,12 @@ Image& ShaderIR::GetBindlessImage(Tegra::Shader::Register reg, Tegra::Shader::Im const auto buffer = std::get<1>(result); const auto offset = std::get<2>(result); - const auto it = - std::find_if(std::begin(used_images), std::end(used_images), - [buffer = buffer, offset = offset](const Image& entry) { - return entry.GetBuffer() == buffer && entry.GetOffset() == offset; - }); + const auto it = std::find_if(std::begin(used_images), std::end(used_images), + [buffer, offset](const Image& entry) { + return entry.buffer == buffer && entry.offset == offset; + }); if (it != std::end(used_images)) { - ASSERT(it->IsBindless() && it->GetType() == it->GetType()); + ASSERT(it->is_bindless && it->type == type); return *it; } diff --git a/src/video_core/shader/decode/register_set_predicate.cpp b/src/video_core/shader/decode/register_set_predicate.cpp index 8d54cce34..6116c31aa 100644 --- a/src/video_core/shader/decode/register_set_predicate.cpp +++ b/src/video_core/shader/decode/register_set_predicate.cpp @@ -2,6 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <utility> + #include "common/assert.h" #include "common/common_types.h" #include "video_core/engines/shader_bytecode.h" @@ -10,20 +12,20 @@ namespace VideoCommon::Shader { +using std::move; using Tegra::Shader::Instruction; using Tegra::Shader::OpCode; namespace { -constexpr u64 NUM_PROGRAMMABLE_PREDICATES = 7; -} +constexpr u64 NUM_CONDITION_CODES = 4; +constexpr u64 NUM_PREDICATES = 7; +} // namespace u32 ShaderIR::DecodeRegisterSetPredicate(NodeBlock& bb, u32 pc) { const Instruction instr = {program_code[pc]}; const auto opcode = OpCode::Decode(instr); - UNIMPLEMENTED_IF(instr.p2r_r2p.mode != Tegra::Shader::R2pMode::Pr); - - const Node apply_mask = [&] { + Node apply_mask = [this, opcode, instr] { switch (opcode->get().GetId()) { case OpCode::Id::R2P_IMM: case OpCode::Id::P2R_IMM: @@ -34,39 +36,43 @@ u32 ShaderIR::DecodeRegisterSetPredicate(NodeBlock& bb, u32 pc) { } }(); - const auto offset = static_cast<u32>(instr.p2r_r2p.byte) * 8; + const u32 offset = static_cast<u32>(instr.p2r_r2p.byte) * 8; + + const bool cc = instr.p2r_r2p.mode == Tegra::Shader::R2pMode::Cc; + const u64 num_entries = cc ? NUM_CONDITION_CODES : NUM_PREDICATES; + const auto get_entry = [this, cc](u64 entry) { + return cc ? GetInternalFlag(static_cast<InternalFlag>(entry)) : GetPredicate(entry); + }; switch (opcode->get().GetId()) { case OpCode::Id::R2P_IMM: { - const Node mask = GetRegister(instr.gpr8); + Node mask = GetRegister(instr.gpr8); - for (u64 pred = 0; pred < NUM_PROGRAMMABLE_PREDICATES; ++pred) { - const auto shift = static_cast<u32>(pred); + for (u64 entry = 0; entry < num_entries; ++entry) { + const u32 shift = static_cast<u32>(entry); - const Node apply_compare = BitfieldExtract(apply_mask, shift, 1); - const Node condition = - Operation(OperationCode::LogicalUNotEqual, apply_compare, Immediate(0)); + Node apply = BitfieldExtract(apply_mask, shift, 1); + Node condition = Operation(OperationCode::LogicalUNotEqual, apply, Immediate(0)); - const Node value_compare = BitfieldExtract(mask, offset + shift, 1); - const Node value = - Operation(OperationCode::LogicalUNotEqual, value_compare, Immediate(0)); + Node compare = BitfieldExtract(mask, offset + shift, 1); + Node value = Operation(OperationCode::LogicalUNotEqual, move(compare), Immediate(0)); - const Node code = Operation(OperationCode::LogicalAssign, GetPredicate(pred), value); - bb.push_back(Conditional(condition, {code})); + Node code = Operation(OperationCode::LogicalAssign, get_entry(entry), move(value)); + bb.push_back(Conditional(condition, {move(code)})); } break; } case OpCode::Id::P2R_IMM: { Node value = Immediate(0); - for (u64 pred = 0; pred < NUM_PROGRAMMABLE_PREDICATES; ++pred) { - Node bit = Operation(OperationCode::Select, GetPredicate(pred), Immediate(1U << pred), + for (u64 entry = 0; entry < num_entries; ++entry) { + Node bit = Operation(OperationCode::Select, get_entry(entry), Immediate(1U << entry), Immediate(0)); - value = Operation(OperationCode::UBitwiseOr, std::move(value), std::move(bit)); + value = Operation(OperationCode::UBitwiseOr, move(value), move(bit)); } - value = Operation(OperationCode::UBitwiseAnd, std::move(value), apply_mask); - value = BitfieldInsert(GetRegister(instr.gpr8), std::move(value), offset, 8); + value = Operation(OperationCode::UBitwiseAnd, move(value), apply_mask); + value = BitfieldInsert(GetRegister(instr.gpr8), move(value), offset, 8); - SetRegister(bb, instr.gpr0, std::move(value)); + SetRegister(bb, instr.gpr0, move(value)); break; } default: diff --git a/src/video_core/shader/decode/texture.cpp b/src/video_core/shader/decode/texture.cpp index e68f1d305..8f0bb996e 100644 --- a/src/video_core/shader/decode/texture.cpp +++ b/src/video_core/shader/decode/texture.cpp @@ -139,15 +139,15 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { } const Node component = Immediate(static_cast<u32>(instr.tld4s.component)); - const SamplerInfo info{TextureType::Texture2D, false, is_depth_compare, false}; - const Sampler& sampler = *GetSampler(instr.sampler, info); + SamplerInfo info; + info.is_shadow = is_depth_compare; + const std::optional<Sampler> sampler = GetSampler(instr.sampler, info); Node4 values; for (u32 element = 0; element < values.size(); ++element) { - auto coords_copy = coords; - MetaTexture meta{sampler, {}, depth_compare, aoffi, {}, {}, - {}, {}, component, element, {}}; - values[element] = Operation(OperationCode::TextureGather, meta, std::move(coords_copy)); + MetaTexture meta{*sampler, {}, depth_compare, aoffi, {}, {}, + {}, {}, component, element, {}}; + values[element] = Operation(OperationCode::TextureGather, meta, coords); } if (instr.tld4s.fp16_flag) { @@ -165,18 +165,20 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { "AOFFI is not implemented"); const bool is_array = instr.txd.is_array != 0; - u64 base_reg = instr.gpr8.Value(); const auto derivate_reg = instr.gpr20.Value(); const auto texture_type = instr.txd.texture_type.Value(); const auto coord_count = GetCoordCount(texture_type); - Node index_var{}; - const Sampler* sampler = - is_bindless - ? GetBindlessSampler(base_reg, index_var, {{texture_type, is_array, false, false}}) - : GetSampler(instr.sampler, {{texture_type, is_array, false, false}}); + u64 base_reg = instr.gpr8.Value(); + Node index_var; + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + const std::optional<Sampler> sampler = is_bindless + ? GetBindlessSampler(base_reg, info, index_var) + : GetSampler(instr.sampler, info); Node4 values; - if (sampler == nullptr) { - std::generate(values.begin(), values.end(), [] { return Immediate(0); }); + if (!sampler) { + std::generate(values.begin(), values.end(), [this] { return Immediate(0); }); WriteTexInstructionFloat(bb, instr, values); break; } @@ -214,14 +216,12 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { is_bindless = true; [[fallthrough]]; case OpCode::Id::TXQ: { - // TODO: The new commits on the texture refactor, change the way samplers work. - // Sadly, not all texture instructions specify the type of texture their sampler - // uses. This must be fixed at a later instance. - Node index_var{}; - const Sampler* sampler = - is_bindless ? GetBindlessSampler(instr.gpr8, index_var) : GetSampler(instr.sampler); - - if (sampler == nullptr) { + Node index_var; + const std::optional<Sampler> sampler = is_bindless + ? GetBindlessSampler(instr.gpr8, {}, index_var) + : GetSampler(instr.sampler, {}); + + if (!sampler) { u32 indexer = 0; for (u32 element = 0; element < 4; ++element) { if (!instr.txq.IsComponentEnabled(element)) { @@ -267,12 +267,17 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { UNIMPLEMENTED_IF_MSG(instr.tmml.UsesMiscMode(Tegra::Shader::TextureMiscMode::NDV), "NDV is not implemented"); - auto texture_type = instr.tmml.texture_type.Value(); - Node index_var{}; - const Sampler* sampler = - is_bindless ? GetBindlessSampler(instr.gpr20, index_var) : GetSampler(instr.sampler); - - if (sampler == nullptr) { + const auto texture_type = instr.tmml.texture_type.Value(); + const bool is_array = instr.tmml.array != 0; + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + Node index_var; + const std::optional<Sampler> sampler = + is_bindless ? GetBindlessSampler(instr.gpr20, info, index_var) + : GetSampler(instr.sampler, info); + + if (!sampler) { u32 indexer = 0; for (u32 element = 0; element < 2; ++element) { if (!instr.tmml.IsComponentEnabled(element)) { @@ -299,12 +304,11 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { coords.push_back(GetRegister(instr.gpr8.Value() + 1)); break; default: - UNIMPLEMENTED_MSG("Unhandled texture type {}", static_cast<u32>(texture_type)); + UNIMPLEMENTED_MSG("Unhandled texture type {}", static_cast<int>(texture_type)); // Fallback to interpreting as a 2D texture for now coords.push_back(GetRegister(instr.gpr8.Value() + 0)); coords.push_back(GetRegister(instr.gpr8.Value() + 1)); - texture_type = TextureType::Texture2D; } u32 indexer = 0; for (u32 element = 0; element < 2; ++element) { @@ -353,98 +357,103 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { return pc; } -ShaderIR::SamplerInfo ShaderIR::GetSamplerInfo(std::optional<SamplerInfo> sampler_info, u32 offset, +ShaderIR::SamplerInfo ShaderIR::GetSamplerInfo(SamplerInfo info, u32 offset, std::optional<u32> buffer) { - if (sampler_info) { - return *sampler_info; + if (info.IsComplete()) { + return info; } const auto sampler = buffer ? registry.ObtainBindlessSampler(*buffer, offset) : registry.ObtainBoundSampler(offset); if (!sampler) { LOG_WARNING(HW_GPU, "Unknown sampler info"); - return SamplerInfo{TextureType::Texture2D, false, false, false}; - } - return SamplerInfo{sampler->texture_type, sampler->is_array != 0, sampler->is_shadow != 0, - sampler->is_buffer != 0}; + info.type = info.type.value_or(Tegra::Shader::TextureType::Texture2D); + info.is_array = info.is_array.value_or(false); + info.is_shadow = info.is_shadow.value_or(false); + info.is_buffer = info.is_buffer.value_or(false); + return info; + } + info.type = info.type.value_or(sampler->texture_type); + info.is_array = info.is_array.value_or(sampler->is_array != 0); + info.is_shadow = info.is_shadow.value_or(sampler->is_shadow != 0); + info.is_buffer = info.is_buffer.value_or(sampler->is_buffer != 0); + return info; } -const Sampler* ShaderIR::GetSampler(const Tegra::Shader::Sampler& sampler, - std::optional<SamplerInfo> sampler_info) { +std::optional<Sampler> ShaderIR::GetSampler(Tegra::Shader::Sampler sampler, + SamplerInfo sampler_info) { const auto offset = static_cast<u32>(sampler.index.Value()); const auto info = GetSamplerInfo(sampler_info, offset); // If this sampler has already been used, return the existing mapping. - const auto it = - std::find_if(used_samplers.begin(), used_samplers.end(), - [offset](const Sampler& entry) { return entry.GetOffset() == offset; }); + const auto it = std::find_if(used_samplers.begin(), used_samplers.end(), + [offset](const Sampler& entry) { return entry.offset == offset; }); if (it != used_samplers.end()) { - ASSERT(!it->IsBindless() && it->GetType() == info.type && it->IsArray() == info.is_array && - it->IsShadow() == info.is_shadow && it->IsBuffer() == info.is_buffer); - return &*it; + ASSERT(!it->is_bindless && it->type == info.type && it->is_array == info.is_array && + it->is_shadow == info.is_shadow && it->is_buffer == info.is_buffer); + return *it; } // Otherwise create a new mapping for this sampler const auto next_index = static_cast<u32>(used_samplers.size()); - return &used_samplers.emplace_back(next_index, offset, info.type, info.is_array, info.is_shadow, - info.is_buffer, false); + return used_samplers.emplace_back(next_index, offset, *info.type, *info.is_array, + *info.is_shadow, *info.is_buffer, false); } -const Sampler* ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, Node& index_var, - std::optional<SamplerInfo> sampler_info) { +std::optional<Sampler> ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, SamplerInfo info, + Node& index_var) { const Node sampler_register = GetRegister(reg); const auto [base_node, tracked_sampler_info] = TrackBindlessSampler(sampler_register, global_code, static_cast<s64>(global_code.size())); ASSERT(base_node != nullptr); if (base_node == nullptr) { - return nullptr; + return std::nullopt; } if (const auto bindless_sampler_info = std::get_if<BindlessSamplerNode>(&*tracked_sampler_info)) { const u32 buffer = bindless_sampler_info->GetIndex(); const u32 offset = bindless_sampler_info->GetOffset(); - const auto info = GetSamplerInfo(sampler_info, offset, buffer); + info = GetSamplerInfo(info, offset, buffer); // If this sampler has already been used, return the existing mapping. - const auto it = - std::find_if(used_samplers.begin(), used_samplers.end(), - [buffer = buffer, offset = offset](const Sampler& entry) { - return entry.GetBuffer() == buffer && entry.GetOffset() == offset; - }); + const auto it = std::find_if(used_samplers.begin(), used_samplers.end(), + [buffer = buffer, offset = offset](const Sampler& entry) { + return entry.buffer == buffer && entry.offset == offset; + }); if (it != used_samplers.end()) { - ASSERT(it->IsBindless() && it->GetType() == info.type && - it->IsArray() == info.is_array && it->IsShadow() == info.is_shadow); - return &*it; + ASSERT(it->is_bindless && it->type == info.type && it->is_array == info.is_array && + it->is_shadow == info.is_shadow); + return *it; } // Otherwise create a new mapping for this sampler const auto next_index = static_cast<u32>(used_samplers.size()); - return &used_samplers.emplace_back(next_index, offset, buffer, info.type, info.is_array, - info.is_shadow, info.is_buffer, false); - } else if (const auto array_sampler_info = - std::get_if<ArraySamplerNode>(&*tracked_sampler_info)) { + return used_samplers.emplace_back(next_index, offset, buffer, *info.type, *info.is_array, + *info.is_shadow, *info.is_buffer, false); + } + if (const auto array_sampler_info = std::get_if<ArraySamplerNode>(&*tracked_sampler_info)) { const u32 base_offset = array_sampler_info->GetBaseOffset() / 4; index_var = GetCustomVariable(array_sampler_info->GetIndexVar()); - const auto info = GetSamplerInfo(sampler_info, base_offset); + info = GetSamplerInfo(info, base_offset); // If this sampler has already been used, return the existing mapping. const auto it = std::find_if( used_samplers.begin(), used_samplers.end(), - [base_offset](const Sampler& entry) { return entry.GetOffset() == base_offset; }); + [base_offset](const Sampler& entry) { return entry.offset == base_offset; }); if (it != used_samplers.end()) { - ASSERT(!it->IsBindless() && it->GetType() == info.type && - it->IsArray() == info.is_array && it->IsShadow() == info.is_shadow && - it->IsBuffer() == info.is_buffer && it->IsIndexed()); - return &*it; + ASSERT(!it->is_bindless && it->type == info.type && it->is_array == info.is_array && + it->is_shadow == info.is_shadow && it->is_buffer == info.is_buffer && + it->is_indexed); + return *it; } uses_indexed_samplers = true; // Otherwise create a new mapping for this sampler const auto next_index = static_cast<u32>(used_samplers.size()); - return &used_samplers.emplace_back(next_index, base_offset, info.type, info.is_array, - info.is_shadow, info.is_buffer, true); + return used_samplers.emplace_back(next_index, base_offset, *info.type, *info.is_array, + *info.is_shadow, *info.is_buffer, true); } - return nullptr; + return std::nullopt; } void ShaderIR::WriteTexInstructionFloat(NodeBlock& bb, Instruction instr, const Node4& components) { @@ -529,10 +538,16 @@ Node4 ShaderIR::GetTextureCode(Instruction instr, TextureType texture_type, ASSERT_MSG(texture_type != TextureType::Texture3D || !is_array || !is_shadow, "Illegal texture type"); - const SamplerInfo info{texture_type, is_array, is_shadow, false}; + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + info.is_shadow = is_shadow; + info.is_buffer = false; + Node index_var; - const Sampler* sampler = is_bindless ? GetBindlessSampler(*bindless_reg, index_var, info) - : GetSampler(instr.sampler, info); + const std::optional<Sampler> sampler = is_bindless + ? GetBindlessSampler(*bindless_reg, info, index_var) + : GetSampler(instr.sampler, info); if (!sampler) { return {Immediate(0), Immediate(0), Immediate(0), Immediate(0)}; } @@ -683,12 +698,17 @@ Node4 ShaderIR::GetTld4Code(Instruction instr, TextureType texture_type, bool de u64 parameter_register = instr.gpr20.Value(); - const SamplerInfo info{texture_type, is_array, depth_compare, false}; - Node index_var{}; - const Sampler* sampler = is_bindless ? GetBindlessSampler(parameter_register++, index_var, info) - : GetSampler(instr.sampler, info); + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + info.is_shadow = depth_compare; + + Node index_var; + const std::optional<Sampler> sampler = + is_bindless ? GetBindlessSampler(parameter_register++, info, index_var) + : GetSampler(instr.sampler, info); Node4 values; - if (sampler == nullptr) { + if (!sampler) { for (u32 element = 0; element < values.size(); ++element) { values[element] = Immediate(0); } @@ -743,12 +763,12 @@ Node4 ShaderIR::GetTldCode(Tegra::Shader::Instruction instr) { // const Node aoffi_register{is_aoffi ? GetRegister(gpr20_cursor++) : nullptr}; // const Node multisample{is_multisample ? GetRegister(gpr20_cursor++) : nullptr}; - const auto& sampler = *GetSampler(instr.sampler); + const std::optional<Sampler> sampler = GetSampler(instr.sampler, {}); Node4 values; for (u32 element = 0; element < values.size(); ++element) { auto coords_copy = coords; - MetaTexture meta{sampler, array_register, {}, {}, {}, {}, {}, lod, {}, element, {}}; + MetaTexture meta{*sampler, array_register, {}, {}, {}, {}, {}, lod, {}, element, {}}; values[element] = Operation(OperationCode::TexelFetch, meta, std::move(coords_copy)); } @@ -756,7 +776,11 @@ Node4 ShaderIR::GetTldCode(Tegra::Shader::Instruction instr) { } Node4 ShaderIR::GetTldsCode(Instruction instr, TextureType texture_type, bool is_array) { - const Sampler& sampler = *GetSampler(instr.sampler); + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + info.is_shadow = false; + const std::optional<Sampler> sampler = GetSampler(instr.sampler, info); const std::size_t type_coord_count = GetCoordCount(texture_type); const bool lod_enabled = instr.tlds.GetTextureProcessMode() == TextureProcessMode::LL; @@ -784,7 +808,7 @@ Node4 ShaderIR::GetTldsCode(Instruction instr, TextureType texture_type, bool is Node4 values; for (u32 element = 0; element < values.size(); ++element) { auto coords_copy = coords; - MetaTexture meta{sampler, array, {}, {}, {}, {}, {}, lod, {}, element, {}}; + MetaTexture meta{*sampler, array, {}, {}, {}, {}, {}, lod, {}, element, {}}; values[element] = Operation(OperationCode::TexelFetch, meta, std::move(coords_copy)); } return values; diff --git a/src/video_core/shader/node.h b/src/video_core/shader/node.h index 3eee961f5..601c822d2 100644 --- a/src/video_core/shader/node.h +++ b/src/video_core/shader/node.h @@ -132,6 +132,8 @@ enum class OperationCode { LogicalUNotEqual, /// (uint a, uint b) -> bool LogicalUGreaterEqual, /// (uint a, uint b) -> bool + LogicalAddCarry, /// (uint a, uint b) -> bool + Logical2HLessThan, /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2 Logical2HEqual, /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2 Logical2HLessEqual, /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2 @@ -265,76 +267,30 @@ class ArraySamplerNode; using TrackSamplerData = std::variant<BindlessSamplerNode, ArraySamplerNode>; using TrackSampler = std::shared_ptr<TrackSamplerData>; -class Sampler { -public: - /// This constructor is for bound samplers +struct Sampler { + /// Bound samplers constructor constexpr explicit Sampler(u32 index, u32 offset, Tegra::Shader::TextureType type, bool is_array, bool is_shadow, bool is_buffer, bool is_indexed) : index{index}, offset{offset}, type{type}, is_array{is_array}, is_shadow{is_shadow}, is_buffer{is_buffer}, is_indexed{is_indexed} {} - /// This constructor is for bindless samplers + /// Bindless samplers constructor constexpr explicit Sampler(u32 index, u32 offset, u32 buffer, Tegra::Shader::TextureType type, bool is_array, bool is_shadow, bool is_buffer, bool is_indexed) : index{index}, offset{offset}, buffer{buffer}, type{type}, is_array{is_array}, is_shadow{is_shadow}, is_buffer{is_buffer}, is_bindless{true}, is_indexed{is_indexed} {} - constexpr u32 GetIndex() const { - return index; - } - - constexpr u32 GetOffset() const { - return offset; - } - - constexpr u32 GetBuffer() const { - return buffer; - } - - constexpr Tegra::Shader::TextureType GetType() const { - return type; - } - - constexpr bool IsArray() const { - return is_array; - } - - constexpr bool IsShadow() const { - return is_shadow; - } - - constexpr bool IsBuffer() const { - return is_buffer; - } - - constexpr bool IsBindless() const { - return is_bindless; - } - - constexpr bool IsIndexed() const { - return is_indexed; - } - - constexpr u32 Size() const { - return size; - } - - constexpr void SetSize(u32 new_size) { - size = new_size; - } - -private: - u32 index{}; ///< Emulated index given for the this sampler. - u32 offset{}; ///< Offset in the const buffer from where the sampler is being read. - u32 buffer{}; ///< Buffer where the bindless sampler is being read (unused on bound samplers). - u32 size{1}; ///< Size of the sampler. + u32 index = 0; ///< Emulated index given for the this sampler. + u32 offset = 0; ///< Offset in the const buffer from where the sampler is being read. + u32 buffer = 0; ///< Buffer where the bindless sampler is being read (unused on bound samplers). + u32 size = 1; ///< Size of the sampler. Tegra::Shader::TextureType type{}; ///< The type used to sample this texture (Texture2D, etc) - bool is_array{}; ///< Whether the texture is being sampled as an array texture or not. - bool is_shadow{}; ///< Whether the texture is being sampled as a depth texture or not. - bool is_buffer{}; ///< Whether the texture is a texture buffer without sampler. - bool is_bindless{}; ///< Whether this sampler belongs to a bindless texture or not. - bool is_indexed{}; ///< Whether this sampler is an indexed array of textures. + bool is_array = false; ///< Whether the texture is being sampled as an array texture or not. + bool is_shadow = false; ///< Whether the texture is being sampled as a depth texture or not. + bool is_buffer = false; ///< Whether the texture is a texture buffer without sampler. + bool is_bindless = false; ///< Whether this sampler belongs to a bindless texture or not. + bool is_indexed = false; ///< Whether this sampler is an indexed array of textures. }; /// Represents a tracked bindless sampler into a direct const buffer @@ -379,13 +335,13 @@ private: u32 offset; }; -class Image final { +struct Image { public: - /// This constructor is for bound images + /// Bound images constructor constexpr explicit Image(u32 index, u32 offset, Tegra::Shader::ImageType type) : index{index}, offset{offset}, type{type} {} - /// This constructor is for bindless samplers + /// Bindless samplers constructor constexpr explicit Image(u32 index, u32 offset, u32 buffer, Tegra::Shader::ImageType type) : index{index}, offset{offset}, buffer{buffer}, type{type}, is_bindless{true} {} @@ -403,53 +359,20 @@ public: is_atomic = true; } - constexpr u32 GetIndex() const { - return index; - } - - constexpr u32 GetOffset() const { - return offset; - } - - constexpr u32 GetBuffer() const { - return buffer; - } - - constexpr Tegra::Shader::ImageType GetType() const { - return type; - } - - constexpr bool IsBindless() const { - return is_bindless; - } - - constexpr bool IsWritten() const { - return is_written; - } - - constexpr bool IsRead() const { - return is_read; - } - - constexpr bool IsAtomic() const { - return is_atomic; - } - -private: - u32 index{}; - u32 offset{}; - u32 buffer{}; + u32 index = 0; + u32 offset = 0; + u32 buffer = 0; Tegra::Shader::ImageType type{}; - bool is_bindless{}; - bool is_written{}; - bool is_read{}; - bool is_atomic{}; + bool is_bindless = false; + bool is_written = false; + bool is_read = false; + bool is_atomic = false; }; struct GlobalMemoryBase { - u32 cbuf_index{}; - u32 cbuf_offset{}; + u32 cbuf_index = 0; + u32 cbuf_offset = 0; bool operator<(const GlobalMemoryBase& rhs) const { return std::tie(cbuf_index, cbuf_offset) < std::tie(rhs.cbuf_index, rhs.cbuf_offset); @@ -463,7 +386,7 @@ struct MetaArithmetic { /// Parameters describing a texture sampler struct MetaTexture { - const Sampler& sampler; + Sampler sampler; Node array; Node depth_compare; std::vector<Node> aoffi; diff --git a/src/video_core/shader/shader_ir.h b/src/video_core/shader/shader_ir.h index 69de5e68b..15ae152f2 100644 --- a/src/video_core/shader/shader_ir.h +++ b/src/video_core/shader/shader_ir.h @@ -28,12 +28,11 @@ struct ShaderBlock; constexpr u32 MAX_PROGRAM_LENGTH = 0x1000; -class ConstBuffer { -public: - explicit ConstBuffer(u32 max_offset, bool is_indirect) +struct ConstBuffer { + constexpr explicit ConstBuffer(u32 max_offset, bool is_indirect) : max_offset{max_offset}, is_indirect{is_indirect} {} - ConstBuffer() = default; + constexpr ConstBuffer() = default; void MarkAsUsed(u64 offset) { max_offset = std::max(max_offset, static_cast<u32>(offset)); @@ -56,8 +55,8 @@ public: } private: - u32 max_offset{}; - bool is_indirect{}; + u32 max_offset = 0; + bool is_indirect = false; }; struct GlobalMemoryUsage { @@ -191,10 +190,14 @@ private: friend class ASTDecoder; struct SamplerInfo { - Tegra::Shader::TextureType type; - bool is_array; - bool is_shadow; - bool is_buffer; + std::optional<Tegra::Shader::TextureType> type; + std::optional<bool> is_array; + std::optional<bool> is_shadow; + std::optional<bool> is_buffer; + + constexpr bool IsComplete() const noexcept { + return type && is_array && is_shadow && is_buffer; + } }; void Decode(); @@ -327,16 +330,15 @@ private: OperationCode GetPredicateCombiner(Tegra::Shader::PredOperation operation); /// Queries the missing sampler info from the execution context. - SamplerInfo GetSamplerInfo(std::optional<SamplerInfo> sampler_info, u32 offset, + SamplerInfo GetSamplerInfo(SamplerInfo info, u32 offset, std::optional<u32> buffer = std::nullopt); - /// Accesses a texture sampler - const Sampler* GetSampler(const Tegra::Shader::Sampler& sampler, - std::optional<SamplerInfo> sampler_info = std::nullopt); + /// Accesses a texture sampler. + std::optional<Sampler> GetSampler(Tegra::Shader::Sampler sampler, SamplerInfo info); /// Accesses a texture sampler for a bindless texture. - const Sampler* GetBindlessSampler(Tegra::Shader::Register reg, Node& index_var, - std::optional<SamplerInfo> sampler_info = std::nullopt); + std::optional<Sampler> GetBindlessSampler(Tegra::Shader::Register reg, SamplerInfo info, + Node& index_var); /// Accesses an image. Image& GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType type); diff --git a/src/video_core/shader/track.cpp b/src/video_core/shader/track.cpp index 513e9bf49..eb97bfd41 100644 --- a/src/video_core/shader/track.cpp +++ b/src/video_core/shader/track.cpp @@ -153,21 +153,13 @@ std::tuple<Node, u32, u32> ShaderIR::TrackCbuf(Node tracked, const NodeBlock& co if (gpr->GetIndex() == Tegra::Shader::Register::ZeroIndex) { return {}; } - s64 current_cursor = cursor; - while (current_cursor > 0) { - // Reduce the cursor in one to avoid infinite loops when the instruction sets the same - // register that it uses as operand - const auto [source, new_cursor] = TrackRegister(gpr, code, current_cursor - 1); - current_cursor = new_cursor; - if (!source) { - continue; - } - const auto [base_address, index, offset] = TrackCbuf(source, code, current_cursor); - if (base_address != nullptr) { - return {base_address, index, offset}; - } + // Reduce the cursor in one to avoid infinite loops when the instruction sets the same + // register that it uses as operand + const auto [source, new_cursor] = TrackRegister(gpr, code, cursor - 1); + if (!source) { + return {}; } - return {}; + return TrackCbuf(source, code, new_cursor); } if (const auto operation = std::get_if<OperationNode>(&*tracked)) { for (std::size_t i = operation->GetOperandsCount(); i > 0; --i) { diff --git a/src/video_core/texture_cache/surface_params.cpp b/src/video_core/texture_cache/surface_params.cpp index 0de499946..884fabffe 100644 --- a/src/video_core/texture_cache/surface_params.cpp +++ b/src/video_core/texture_cache/surface_params.cpp @@ -81,7 +81,7 @@ SurfaceParams SurfaceParams::CreateForTexture(const FormatLookupTable& lookup_ta params.pixel_format = lookup_table.GetPixelFormat( tic.format, params.srgb_conversion, tic.r_type, tic.g_type, tic.b_type, tic.a_type); params.type = GetFormatType(params.pixel_format); - if (entry.IsShadow() && params.type == SurfaceType::ColorTexture) { + if (entry.is_shadow && params.type == SurfaceType::ColorTexture) { switch (params.pixel_format) { case PixelFormat::R16U: case PixelFormat::R16F: @@ -108,7 +108,7 @@ SurfaceParams SurfaceParams::CreateForTexture(const FormatLookupTable& lookup_ta params.emulated_levels = 1; params.is_layered = false; } else { - params.target = TextureTypeToSurfaceTarget(entry.GetType(), entry.IsArray()); + params.target = TextureTypeToSurfaceTarget(entry.type, entry.is_array); params.width = tic.Width(); params.height = tic.Height(); params.depth = tic.Depth(); @@ -138,7 +138,7 @@ SurfaceParams SurfaceParams::CreateForImage(const FormatLookupTable& lookup_tabl tic.format, params.srgb_conversion, tic.r_type, tic.g_type, tic.b_type, tic.a_type); params.type = GetFormatType(params.pixel_format); params.type = GetFormatType(params.pixel_format); - params.target = ImageTypeToSurfaceTarget(entry.GetType()); + params.target = ImageTypeToSurfaceTarget(entry.type); // TODO: on 1DBuffer we should use the tic info. if (tic.IsBuffer()) { params.target = SurfaceTarget::TextureBuffer; @@ -248,12 +248,12 @@ SurfaceParams SurfaceParams::CreateForFermiCopySurface( VideoCore::Surface::SurfaceTarget SurfaceParams::ExpectedTarget( const VideoCommon::Shader::Sampler& entry) { - return TextureTypeToSurfaceTarget(entry.GetType(), entry.IsArray()); + return TextureTypeToSurfaceTarget(entry.type, entry.is_array); } VideoCore::Surface::SurfaceTarget SurfaceParams::ExpectedTarget( const VideoCommon::Shader::Image& entry) { - return ImageTypeToSurfaceTarget(entry.GetType()); + return ImageTypeToSurfaceTarget(entry.type); } bool SurfaceParams::IsLayered() const { diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index cf6bd005a..d6efc34b2 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -143,7 +143,7 @@ public: } const auto params{SurfaceParams::CreateForTexture(format_lookup_table, tic, entry)}; - const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, false); + const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, true, false); if (guard_samplers) { sampled_textures.push_back(surface); } @@ -163,7 +163,7 @@ public: return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); } const auto params{SurfaceParams::CreateForImage(format_lookup_table, tic, entry)}; - const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, false); + const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, true, false); if (guard_samplers) { sampled_textures.push_back(surface); } @@ -178,7 +178,7 @@ public: return any_rt; } - TView GetDepthBufferSurface() { + TView GetDepthBufferSurface(bool preserve_contents) { std::lock_guard lock{mutex}; auto& maxwell3d = system.GPU().Maxwell3D(); if (!maxwell3d.dirty.flags[VideoCommon::Dirty::ZetaBuffer]) { @@ -199,7 +199,7 @@ public: return {}; } const auto depth_params{SurfaceParams::CreateForDepthBuffer(system)}; - auto surface_view = GetSurface(gpu_addr, *cpu_addr, depth_params, true); + auto surface_view = GetSurface(gpu_addr, *cpu_addr, depth_params, preserve_contents, true); if (depth_buffer.target) depth_buffer.target->MarkAsRenderTarget(false, NO_RT); depth_buffer.target = surface_view.first; @@ -209,7 +209,7 @@ public: return surface_view.second; } - TView GetColorBufferSurface(std::size_t index) { + TView GetColorBufferSurface(std::size_t index, bool preserve_contents) { std::lock_guard lock{mutex}; ASSERT(index < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets); auto& maxwell3d = system.GPU().Maxwell3D(); @@ -239,8 +239,9 @@ public: return {}; } - auto surface_view = GetSurface(gpu_addr, *cpu_addr, - SurfaceParams::CreateForFramebuffer(system, index), true); + auto surface_view = + GetSurface(gpu_addr, *cpu_addr, SurfaceParams::CreateForFramebuffer(system, index), + preserve_contents, true); if (render_targets[index].target) { auto& surface = render_targets[index].target; surface->MarkAsRenderTarget(false, NO_RT); @@ -300,9 +301,9 @@ public: const std::optional<VAddr> src_cpu_addr = system.GPU().MemoryManager().GpuToCpuAddress(src_gpu_addr); std::pair<TSurface, TView> dst_surface = - GetSurface(dst_gpu_addr, *dst_cpu_addr, dst_params, false); + GetSurface(dst_gpu_addr, *dst_cpu_addr, dst_params, true, false); std::pair<TSurface, TView> src_surface = - GetSurface(src_gpu_addr, *src_cpu_addr, src_params, false); + GetSurface(src_gpu_addr, *src_cpu_addr, src_params, true, false); ImageBlit(src_surface.second, dst_surface.second, copy_config); dst_surface.first->MarkAsModified(true, Tick()); } @@ -532,18 +533,22 @@ private: * @param overlaps The overlapping surfaces registered in the cache. * @param params The parameters for the new surface. * @param gpu_addr The starting address of the new surface. + * @param preserve_contents Indicates that the new surface should be loaded from memory or left + * blank. * @param untopological Indicates to the recycler that the texture has no way to match the * overlaps due to topological reasons. **/ std::pair<TSurface, TView> RecycleSurface(std::vector<TSurface>& overlaps, const SurfaceParams& params, const GPUVAddr gpu_addr, + const bool preserve_contents, const MatchTopologyResult untopological) { + const bool do_load = preserve_contents && Settings::IsGPULevelExtreme(); for (auto& surface : overlaps) { Unregister(surface); } switch (PickStrategy(overlaps, params, gpu_addr, untopological)) { case RecycleStrategy::Ignore: { - return InitializeSurface(gpu_addr, params, Settings::IsGPULevelExtreme()); + return InitializeSurface(gpu_addr, params, do_load); } case RecycleStrategy::Flush: { std::sort(overlaps.begin(), overlaps.end(), @@ -553,7 +558,7 @@ private: for (auto& surface : overlaps) { FlushSurface(surface); } - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, preserve_contents); } case RecycleStrategy::BufferCopy: { auto new_surface = GetUncachedSurface(gpu_addr, params); @@ -562,7 +567,7 @@ private: } default: { UNIMPLEMENTED_MSG("Unimplemented Texture Cache Recycling Strategy!"); - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, do_load); } } } @@ -700,11 +705,14 @@ private: * @param params The parameters on the new surface. * @param gpu_addr The starting address of the new surface. * @param cpu_addr The starting address of the new surface on physical memory. + * @param preserve_contents Indicates that the new surface should be loaded from memory or + * left blank. */ std::optional<std::pair<TSurface, TView>> Manage3DSurfaces(std::vector<TSurface>& overlaps, const SurfaceParams& params, const GPUVAddr gpu_addr, - const VAddr cpu_addr) { + const VAddr cpu_addr, + bool preserve_contents) { if (params.target == SurfaceTarget::Texture3D) { bool failed = false; if (params.num_levels > 1) { @@ -754,7 +762,7 @@ private: return std::nullopt; } Unregister(surface); - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, preserve_contents); } return std::nullopt; } @@ -765,7 +773,7 @@ private: return {{surface, surface->GetMainView()}}; } } - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, preserve_contents); } } @@ -788,10 +796,13 @@ private: * * @param gpu_addr The starting address of the candidate surface. * @param params The parameters on the candidate surface. + * @param preserve_contents Indicates that the new surface should be loaded from memory or + * left blank. * @param is_render Whether or not the surface is a render target. **/ std::pair<TSurface, TView> GetSurface(const GPUVAddr gpu_addr, const VAddr cpu_addr, - const SurfaceParams& params, bool is_render) { + const SurfaceParams& params, bool preserve_contents, + bool is_render) { // Step 1 // Check Level 1 Cache for a fast structural match. If candidate surface // matches at certain level we are pretty much done. @@ -800,7 +811,8 @@ private: const auto topological_result = current_surface->MatchesTopology(params); if (topological_result != MatchTopologyResult::FullMatch) { std::vector<TSurface> overlaps{current_surface}; - return RecycleSurface(overlaps, params, gpu_addr, topological_result); + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, + topological_result); } const auto struct_result = current_surface->MatchesStructure(params); @@ -825,7 +837,7 @@ private: // If none are found, we are done. we just load the surface and create it. if (overlaps.empty()) { - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, preserve_contents); } // Step 3 @@ -835,13 +847,15 @@ private: for (const auto& surface : overlaps) { const auto topological_result = surface->MatchesTopology(params); if (topological_result != MatchTopologyResult::FullMatch) { - return RecycleSurface(overlaps, params, gpu_addr, topological_result); + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, + topological_result); } } // Check if it's a 3D texture if (params.block_depth > 0) { - auto surface = Manage3DSurfaces(overlaps, params, gpu_addr, cpu_addr); + auto surface = + Manage3DSurfaces(overlaps, params, gpu_addr, cpu_addr, preserve_contents); if (surface) { return *surface; } @@ -861,7 +875,8 @@ private: return *view; } } - return RecycleSurface(overlaps, params, gpu_addr, MatchTopologyResult::FullMatch); + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, + MatchTopologyResult::FullMatch); } // Now we check if the candidate is a mipmap/layer of the overlap std::optional<TView> view = @@ -885,7 +900,7 @@ private: pair.first->EmplaceView(params, gpu_addr, candidate_size); if (mirage_view) return {pair.first, *mirage_view}; - return RecycleSurface(overlaps, params, gpu_addr, + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, MatchTopologyResult::FullMatch); } return {current_surface, *view}; @@ -901,7 +916,8 @@ private: } } // We failed all the tests, recycle the overlaps into a new texture. - return RecycleSurface(overlaps, params, gpu_addr, MatchTopologyResult::FullMatch); + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, + MatchTopologyResult::FullMatch); } /** @@ -1059,10 +1075,10 @@ private: } std::pair<TSurface, TView> InitializeSurface(GPUVAddr gpu_addr, const SurfaceParams& params, - bool do_load = true) { + bool preserve_contents) { auto new_surface{GetUncachedSurface(gpu_addr, params)}; Register(new_surface); - if (do_load) { + if (preserve_contents) { LoadSurface(new_surface); } return {new_surface, new_surface->GetMainView()}; @@ -1156,7 +1172,7 @@ private: /// Returns true the shader sampler entry is compatible with the TIC texture type. static bool IsTypeCompatible(Tegra::Texture::TextureType tic_type, const VideoCommon::Shader::Sampler& entry) { - const auto shader_type = entry.GetType(); + const auto shader_type = entry.type; switch (tic_type) { case Tegra::Texture::TextureType::Texture1D: case Tegra::Texture::TextureType::Texture1DArray: @@ -1177,7 +1193,7 @@ private: if (shader_type == Tegra::Shader::TextureType::TextureCube) { return true; } - return shader_type == Tegra::Shader::TextureType::Texture2D && entry.IsArray(); + return shader_type == Tegra::Shader::TextureType::Texture2D && entry.is_array; } UNREACHABLE(); return true; diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp index fae8638ec..548e4c3fe 100644 --- a/src/video_core/textures/decoders.cpp +++ b/src/video_core/textures/decoders.cpp @@ -382,4 +382,18 @@ std::size_t CalculateSize(bool tiled, u32 bytes_per_pixel, u32 width, u32 height } } +u64 GetGOBOffset(u32 width, u32 height, u32 dst_x, u32 dst_y, u32 block_height, + u32 bytes_per_pixel) { + auto div_ceil = [](const u32 x, const u32 y) { return ((x + y - 1) / y); }; + const u32 gobs_in_block = 1 << block_height; + const u32 y_blocks = gob_size_y << block_height; + const u32 x_per_gob = gob_size_x / bytes_per_pixel; + const u32 x_blocks = div_ceil(width, x_per_gob); + const u32 block_size = gob_size * gobs_in_block; + const u32 stride = block_size * x_blocks; + const u32 base = (dst_y / y_blocks) * stride + (dst_x / x_per_gob) * block_size; + const u32 relative_y = dst_y % y_blocks; + return base + (relative_y / gob_size_y) * gob_size; +} + } // namespace Tegra::Texture diff --git a/src/video_core/textures/decoders.h b/src/video_core/textures/decoders.h index 9f2d6d308..06f3ebf87 100644 --- a/src/video_core/textures/decoders.h +++ b/src/video_core/textures/decoders.h @@ -59,4 +59,8 @@ void UnswizzleSubrect(u32 subrect_width, u32 subrect_height, u32 dest_pitch, u32 void SwizzleKepler(u32 width, u32 height, u32 dst_x, u32 dst_y, u32 block_height, std::size_t copy_size, const u8* source_data, u8* swizzle_data); +/// Obtains the offset of the gob for positions 'dst_x' & 'dst_y' +u64 GetGOBOffset(u32 width, u32 height, u32 dst_x, u32 dst_y, u32 block_height, + u32 bytes_per_pixel); + } // namespace Tegra::Texture diff --git a/src/web_service/CMakeLists.txt b/src/web_service/CMakeLists.txt index 0c9bb0d55..06ab7c59d 100644 --- a/src/web_service/CMakeLists.txt +++ b/src/web_service/CMakeLists.txt @@ -8,4 +8,4 @@ add_library(web_service STATIC ) create_target_directory_groups(web_service) -target_link_libraries(web_service PRIVATE common json-headers httplib lurlparser) +target_link_libraries(web_service PRIVATE common nlohmann_json::nlohmann_json httplib lurlparser) diff --git a/src/web_service/telemetry_json.cpp b/src/web_service/telemetry_json.cpp index 7538389bf..7a480e33c 100644 --- a/src/web_service/telemetry_json.cpp +++ b/src/web_service/telemetry_json.cpp @@ -2,7 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <json.hpp> +#include <nlohmann/json.hpp> #include "common/detached_tasks.h" #include "common/web_result.h" #include "web_service/telemetry_json.h" diff --git a/src/web_service/verify_login.cpp b/src/web_service/verify_login.cpp index ca4b43b93..bfaa5b70a 100644 --- a/src/web_service/verify_login.cpp +++ b/src/web_service/verify_login.cpp @@ -2,7 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <json.hpp> +#include <nlohmann/json.hpp> #include "common/web_result.h" #include "web_service/verify_login.h" #include "web_service/web_backend.h" diff --git a/src/yuzu/bootmanager.cpp b/src/yuzu/bootmanager.cpp index 1cac2f942..3d759f77b 100644 --- a/src/yuzu/bootmanager.cpp +++ b/src/yuzu/bootmanager.cpp @@ -292,6 +292,8 @@ GRenderWindow::GRenderWindow(GMainWindow* parent_, EmuThread* emu_thread_) setLayout(layout); InputCommon::Init(); + this->setMouseTracking(true); + connect(this, &GRenderWindow::FirstFrameDisplayed, parent_, &GMainWindow::OnLoadComplete); } @@ -385,6 +387,7 @@ void GRenderWindow::mousePressEvent(QMouseEvent* event) { } else if (event->button() == Qt::RightButton) { InputCommon::GetMotionEmu()->BeginTilt(pos.x(), pos.y()); } + QWidget::mousePressEvent(event); } void GRenderWindow::mouseMoveEvent(QMouseEvent* event) { @@ -397,6 +400,7 @@ void GRenderWindow::mouseMoveEvent(QMouseEvent* event) { const auto [x, y] = ScaleTouch(pos); this->TouchMoved(x, y); InputCommon::GetMotionEmu()->Tilt(pos.x(), pos.y()); + QWidget::mouseMoveEvent(event); } void GRenderWindow::mouseReleaseEvent(QMouseEvent* event) { diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp index a44eed047..75c6cf20b 100644 --- a/src/yuzu/configuration/config.cpp +++ b/src/yuzu/configuration/config.cpp @@ -743,6 +743,8 @@ void Config::ReadUIValues() { UISettings::values.profile_index = ReadSetting(QStringLiteral("profileIndex"), 0).toUInt(); UISettings::values.pause_when_in_background = ReadSetting(QStringLiteral("pauseWhenInBackground"), false).toBool(); + UISettings::values.hide_mouse = + ReadSetting(QStringLiteral("hideInactiveMouse"), false).toBool(); ApplyDefaultProfileIfInputInvalid(); @@ -1169,6 +1171,7 @@ void Config::SaveUIValues() { WriteSetting(QStringLiteral("profileIndex"), UISettings::values.profile_index, 0); WriteSetting(QStringLiteral("pauseWhenInBackground"), UISettings::values.pause_when_in_background, false); + WriteSetting(QStringLiteral("hideInactiveMouse"), UISettings::values.hide_mouse, false); qt_config->endGroup(); } diff --git a/src/yuzu/configuration/configure_general.cpp b/src/yuzu/configuration/configure_general.cpp index 5ef927114..cb95423e0 100644 --- a/src/yuzu/configuration/configure_general.cpp +++ b/src/yuzu/configuration/configure_general.cpp @@ -26,6 +26,7 @@ void ConfigureGeneral::SetConfiguration() { ui->toggle_check_exit->setChecked(UISettings::values.confirm_before_closing); ui->toggle_user_on_boot->setChecked(UISettings::values.select_user_on_boot); ui->toggle_background_pause->setChecked(UISettings::values.pause_when_in_background); + ui->toggle_hide_mouse->setChecked(UISettings::values.hide_mouse); ui->toggle_frame_limit->setChecked(Settings::values.use_frame_limit); ui->frame_limit->setEnabled(ui->toggle_frame_limit->isChecked()); @@ -36,6 +37,7 @@ void ConfigureGeneral::ApplyConfiguration() { UISettings::values.confirm_before_closing = ui->toggle_check_exit->isChecked(); UISettings::values.select_user_on_boot = ui->toggle_user_on_boot->isChecked(); UISettings::values.pause_when_in_background = ui->toggle_background_pause->isChecked(); + UISettings::values.hide_mouse = ui->toggle_hide_mouse->isChecked(); Settings::values.use_frame_limit = ui->toggle_frame_limit->isChecked(); Settings::values.frame_limit = ui->frame_limit->value(); diff --git a/src/yuzu/configuration/configure_general.ui b/src/yuzu/configuration/configure_general.ui index 857119bb3..fc3b7e65a 100644 --- a/src/yuzu/configuration/configure_general.ui +++ b/src/yuzu/configuration/configure_general.ui @@ -72,6 +72,13 @@ </property> </widget> </item> + <item> + <widget class="QCheckBox" name="toggle_hide_mouse"> + <property name="text"> + <string>Hide mouse on inactivity</string> + </property> + </widget> + </item> </layout> </item> </layout> diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index b44b4276c..86e8a1d49 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -135,6 +135,8 @@ __declspec(dllexport) int AmdPowerXpressRequestHighPerformance = 1; } #endif +constexpr int default_mouse_timeout = 2500; + constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000; /** @@ -236,6 +238,14 @@ GMainWindow::GMainWindow() // Show one-time "callout" messages to the user ShowTelemetryCallout(); + // make sure menubar has the arrow cursor instead of inheriting from this + ui.menubar->setCursor(QCursor()); + statusBar()->setCursor(QCursor()); + + mouse_hide_timer.setInterval(default_mouse_timeout); + connect(&mouse_hide_timer, &QTimer::timeout, this, &GMainWindow::HideMouseCursor); + connect(ui.menubar, &QMenuBar::hovered, this, &GMainWindow::ShowMouseCursor); + QStringList args = QApplication::arguments(); if (args.length() >= 2) { BootGame(args[1]); @@ -1012,6 +1022,12 @@ void GMainWindow::BootGame(const QString& filename) { async_status_button->setDisabled(true); renderer_status_button->setDisabled(true); + if (UISettings::values.hide_mouse) { + mouse_hide_timer.start(); + setMouseTracking(true); + ui.centralwidget->setMouseTracking(true); + } + const u64 title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID(); std::string title_name; @@ -1080,6 +1096,9 @@ void GMainWindow::ShutdownGame() { game_list->show(); game_list->setFilterFocus(); + setMouseTracking(false); + ui.centralwidget->setMouseTracking(false); + UpdateWindowTitle(); // Disable status bar updates @@ -1837,6 +1856,15 @@ void GMainWindow::OnConfigure() { config->Save(); + if (UISettings::values.hide_mouse && emulation_running) { + setMouseTracking(true); + ui.centralwidget->setMouseTracking(true); + mouse_hide_timer.start(); + } else { + setMouseTracking(false); + ui.centralwidget->setMouseTracking(false); + } + dock_status_button->setChecked(Settings::values.use_docked_mode); async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation); #ifdef HAS_VULKAN @@ -1970,6 +1998,30 @@ void GMainWindow::UpdateStatusBar() { emu_frametime_label->setVisible(true); } +void GMainWindow::HideMouseCursor() { + if (emu_thread == nullptr || UISettings::values.hide_mouse == false) { + mouse_hide_timer.stop(); + ShowMouseCursor(); + return; + } + setCursor(QCursor(Qt::BlankCursor)); +} + +void GMainWindow::ShowMouseCursor() { + unsetCursor(); + if (emu_thread != nullptr && UISettings::values.hide_mouse) { + mouse_hide_timer.start(); + } +} + +void GMainWindow::mouseMoveEvent(QMouseEvent* event) { + ShowMouseCursor(); +} + +void GMainWindow::mousePressEvent(QMouseEvent* event) { + ShowMouseCursor(); +} + void GMainWindow::OnCoreError(Core::System::ResultStatus result, std::string details) { QMessageBox::StandardButton answer; QString status_message; diff --git a/src/yuzu/main.h b/src/yuzu/main.h index 0b750689d..60b17c54a 100644 --- a/src/yuzu/main.h +++ b/src/yuzu/main.h @@ -216,6 +216,8 @@ private: std::optional<u64> SelectRomFSDumpTarget(const FileSys::ContentProvider&, u64 program_id); void UpdateWindowTitle(const QString& title_name = {}); void UpdateStatusBar(); + void HideMouseCursor(); + void ShowMouseCursor(); Ui::MainWindow ui; @@ -244,6 +246,7 @@ private: QString game_path; bool auto_paused = false; + QTimer mouse_hide_timer; // FS std::shared_ptr<FileSys::VfsFilesystem> vfs; @@ -265,4 +268,6 @@ protected: void dropEvent(QDropEvent* event) override; void dragEnterEvent(QDragEnterEvent* event) override; void dragMoveEvent(QDragMoveEvent* event) override; + void mouseMoveEvent(QMouseEvent* event) override; + void mousePressEvent(QMouseEvent* event) override; }; diff --git a/src/yuzu/uisettings.h b/src/yuzu/uisettings.h index a675ecf4d..830932d45 100644 --- a/src/yuzu/uisettings.h +++ b/src/yuzu/uisettings.h @@ -59,6 +59,7 @@ struct Values { bool confirm_before_closing; bool first_start; bool pause_when_in_background; + bool hide_mouse; bool select_user_on_boot; |