diff options
75 files changed, 639 insertions, 220 deletions
diff --git a/.lgtm.yml b/.lgtm.yml deleted file mode 100644 index 7cd3f9926..000000000 --- a/.lgtm.yml +++ /dev/null @@ -1,13 +0,0 @@ -# SPDX-FileCopyrightText: 2020 yuzu Emulator Project -# SPDX-License-Identifier: GPL-2.0-or-later - -path_classifiers: - library: "externals" -extraction: - cpp: - prepare: - packages: - - "libsdl2-dev" - - "qtmultimedia5-dev" - - "libtbb-dev" - - "libjack-jackd2-dev" diff --git a/externals/dynarmic b/externals/dynarmic -Subproject c08c5a9362bb224dc343c2f616c24df027dfdf1 +Subproject f9e6a3df5c84bcc74be46c289a74a78e5e28d62 diff --git a/src/audio_core/renderer/system.cpp b/src/audio_core/renderer/system.cpp index ad869facb..53b258c4f 100644 --- a/src/audio_core/renderer/system.cpp +++ b/src/audio_core/renderer/system.cpp @@ -436,10 +436,7 @@ void System::Stop() { } if (execution_mode == ExecutionMode::Auto) { - // Should wait for the system to terminate here, but core timing (should have) already - // stopped, so this isn't needed. Find a way to make this definite. - - // terminate_event.Wait(); + terminate_event.Wait(); } } diff --git a/src/audio_core/sink/sdl2_sink.cpp b/src/audio_core/sink/sdl2_sink.cpp index ee1a0652f..c1529d1f9 100644 --- a/src/audio_core/sink/sdl2_sink.cpp +++ b/src/audio_core/sink/sdl2_sink.cpp @@ -3,6 +3,7 @@ #include <span> #include <vector> +#include <SDL.h> #include "audio_core/common/common.h" #include "audio_core/sink/sdl2_sink.h" @@ -10,16 +11,6 @@ #include "common/logging/log.h" #include "core/core.h" -// Ignore -Wimplicit-fallthrough due to https://github.com/libsdl-org/SDL/issues/4307 -#ifdef __clang__ -#pragma clang diagnostic push -#pragma clang diagnostic ignored "-Wimplicit-fallthrough" -#endif -#include <SDL.h> -#ifdef __clang__ -#pragma clang diagnostic pop -#endif - namespace AudioCore::Sink { /** * SDL sink stream, responsible for sinking samples to hardware. diff --git a/src/common/address_space.inc b/src/common/address_space.inc index 2195dabd5..c97dc8651 100644 --- a/src/common/address_space.inc +++ b/src/common/address_space.inc @@ -72,7 +72,7 @@ MAP_MEMBER(void)::MapLocked(VaType virt, PaType phys, VaType size, ExtraBlockInf } }()}; - if (block_end_predecessor->virt >= virt) { + if (block_end_predecessor != blocks.begin() && block_end_predecessor->virt >= virt) { // If this block's start would be overlapped by the map then reuse it as a tail // block block_end_predecessor->virt = virt_end; diff --git a/src/common/input.h b/src/common/input.h index 51b277c1f..66fb15f0a 100644 --- a/src/common/input.h +++ b/src/common/input.h @@ -111,6 +111,8 @@ struct AnalogProperties { float offset{}; // Invert direction of the sensor data bool inverted{}; + // Invert the state if it's converted to a button + bool inverted_button{}; // Press once to activate, press again to release bool toggle{}; }; diff --git a/src/common/settings.cpp b/src/common/settings.cpp index 92794f4a2..f1ee42ab2 100644 --- a/src/common/settings.cpp +++ b/src/common/settings.cpp @@ -45,6 +45,7 @@ void LogSettings() { log_setting("System_LanguageIndex", values.language_index.GetValue()); log_setting("System_RegionIndex", values.region_index.GetValue()); log_setting("System_TimeZoneIndex", values.time_zone_index.GetValue()); + log_setting("System_UnsafeMemoryLayout", values.use_unsafe_extended_memory_layout.GetValue()); log_setting("Core_UseMultiCore", values.use_multi_core.GetValue()); log_setting("CPU_Accuracy", values.cpu_accuracy.GetValue()); log_setting("Renderer_UseResolutionScaling", values.resolution_setup.GetValue()); @@ -191,7 +192,7 @@ void RestoreGlobalState(bool is_powered_on) { // Core values.use_multi_core.SetGlobal(true); - values.use_extended_memory_layout.SetGlobal(true); + values.use_unsafe_extended_memory_layout.SetGlobal(true); // CPU values.cpu_accuracy.SetGlobal(true); @@ -225,7 +226,6 @@ void RestoreGlobalState(bool is_powered_on) { values.shader_backend.SetGlobal(true); values.use_asynchronous_shaders.SetGlobal(true); values.use_fast_gpu_time.SetGlobal(true); - values.use_pessimistic_flushes.SetGlobal(true); values.use_vulkan_driver_pipeline_cache.SetGlobal(true); values.bg_red.SetGlobal(true); values.bg_green.SetGlobal(true); diff --git a/src/common/settings.h b/src/common/settings.h index c0faa7406..2bf191cef 100644 --- a/src/common/settings.h +++ b/src/common/settings.h @@ -395,7 +395,8 @@ struct Values { // Core SwitchableSetting<bool> use_multi_core{true, "use_multi_core"}; - SwitchableSetting<bool> use_extended_memory_layout{false, "use_extended_memory_layout"}; + SwitchableSetting<bool> use_unsafe_extended_memory_layout{false, + "use_unsafe_extended_memory_layout"}; // Cpu SwitchableSetting<CPUAccuracy, true> cpu_accuracy{CPUAccuracy::Auto, CPUAccuracy::Auto, @@ -468,7 +469,6 @@ struct Values { ShaderBackend::SPIRV, "shader_backend"}; SwitchableSetting<bool> use_asynchronous_shaders{false, "use_asynchronous_shaders"}; SwitchableSetting<bool> use_fast_gpu_time{true, "use_fast_gpu_time"}; - SwitchableSetting<bool> use_pessimistic_flushes{false, "use_pessimistic_flushes"}; SwitchableSetting<bool> use_vulkan_driver_pipeline_cache{true, "use_vulkan_driver_pipeline_cache"}; diff --git a/src/common/vector_math.h b/src/common/vector_math.h index 0e2095c45..b4885835d 100644 --- a/src/common/vector_math.h +++ b/src/common/vector_math.h @@ -259,6 +259,20 @@ public: return *this; } + void RotateFromOrigin(float roll, float pitch, float yaw) { + float temp = y; + y = std::cos(roll) * y - std::sin(roll) * z; + z = std::sin(roll) * temp + std::cos(roll) * z; + + temp = x; + x = std::cos(pitch) * x + std::sin(pitch) * z; + z = -std::sin(pitch) * temp + std::cos(pitch) * z; + + temp = x; + x = std::cos(yaw) * x - std::sin(yaw) * y; + y = std::sin(yaw) * temp + std::cos(yaw) * y; + } + [[nodiscard]] constexpr T Length2() const { return x * x + y * y + z * z; } diff --git a/src/core/core.cpp b/src/core/core.cpp index caa6a77be..06fba4ce5 100644 --- a/src/core/core.cpp +++ b/src/core/core.cpp @@ -137,7 +137,7 @@ struct System::Impl { device_memory = std::make_unique<Core::DeviceMemory>(); is_multicore = Settings::values.use_multi_core.GetValue(); - extended_memory_layout = Settings::values.use_extended_memory_layout.GetValue(); + extended_memory_layout = Settings::values.use_unsafe_extended_memory_layout.GetValue(); core_timing.SetMulticore(is_multicore); core_timing.Initialize([&system]() { system.RegisterHostThread(); }); @@ -169,7 +169,7 @@ struct System::Impl { void ReinitializeIfNecessary(System& system) { const bool must_reinitialize = is_multicore != Settings::values.use_multi_core.GetValue() || - extended_memory_layout != Settings::values.use_extended_memory_layout.GetValue(); + extended_memory_layout != Settings::values.use_unsafe_extended_memory_layout.GetValue(); if (!must_reinitialize) { return; @@ -178,7 +178,7 @@ struct System::Impl { LOG_DEBUG(Kernel, "Re-initializing"); is_multicore = Settings::values.use_multi_core.GetValue(); - extended_memory_layout = Settings::values.use_extended_memory_layout.GetValue(); + extended_memory_layout = Settings::values.use_unsafe_extended_memory_layout.GetValue(); Initialize(system); } @@ -293,6 +293,7 @@ struct System::Impl { ASSERT(Kernel::KProcess::Initialize(main_process, system, "main", Kernel::KProcess::ProcessType::Userland, resource_limit) .IsSuccess()); + Kernel::KProcess::Register(system.Kernel(), main_process); kernel.MakeApplicationProcess(main_process); const auto [load_result, load_parameters] = app_loader->Load(*main_process, system); if (load_result != Loader::ResultStatus::Success) { diff --git a/src/core/hid/emulated_controller.cpp b/src/core/hid/emulated_controller.cpp index a70f8807c..ecab85893 100644 --- a/src/core/hid/emulated_controller.cpp +++ b/src/core/hid/emulated_controller.cpp @@ -376,6 +376,7 @@ void EmulatedController::ReloadInput() { motion.accel = emulated_motion.GetAcceleration(); motion.gyro = emulated_motion.GetGyroscope(); motion.rotation = emulated_motion.GetRotations(); + motion.euler = emulated_motion.GetEulerAngles(); motion.orientation = emulated_motion.GetOrientation(); motion.is_at_rest = !emulated_motion.IsMoving(motion_sensitivity); } @@ -551,6 +552,8 @@ void EmulatedController::EnableSystemButtons() { void EmulatedController::DisableSystemButtons() { std::scoped_lock lock{mutex}; system_buttons_enabled = false; + controller.home_button_state.raw = 0; + controller.capture_button_state.raw = 0; } void EmulatedController::ResetSystemButtons() { @@ -734,6 +737,8 @@ void EmulatedController::SetButton(const Common::Input::CallbackStatus& callback if (is_configuring) { controller.npad_button_state.raw = NpadButton::None; controller.debug_pad_button_state.raw = 0; + controller.home_button_state.raw = 0; + controller.capture_button_state.raw = 0; lock.unlock(); TriggerOnChange(ControllerTriggerType::Button, false); return; @@ -976,14 +981,11 @@ void EmulatedController::SetMotion(const Common::Input::CallbackStatus& callback emulated.UpdateOrientation(raw_status.delta_timestamp); force_update_motion = raw_status.force_update; - if (is_configuring) { - return; - } - auto& motion = controller.motion_state[index]; motion.accel = emulated.GetAcceleration(); motion.gyro = emulated.GetGyroscope(); motion.rotation = emulated.GetRotations(); + motion.euler = emulated.GetEulerAngles(); motion.orientation = emulated.GetOrientation(); motion.is_at_rest = !emulated.IsMoving(motion_sensitivity); } diff --git a/src/core/hid/emulated_controller.h b/src/core/hid/emulated_controller.h index 429655355..6e01f4e12 100644 --- a/src/core/hid/emulated_controller.h +++ b/src/core/hid/emulated_controller.h @@ -106,6 +106,7 @@ struct ControllerMotion { Common::Vec3f accel{}; Common::Vec3f gyro{}; Common::Vec3f rotation{}; + Common::Vec3f euler{}; std::array<Common::Vec3f, 3> orientation{}; bool is_at_rest{}; }; diff --git a/src/core/hid/input_converter.cpp b/src/core/hid/input_converter.cpp index 7cee39a53..53b00b1f9 100644 --- a/src/core/hid/input_converter.cpp +++ b/src/core/hid/input_converter.cpp @@ -54,6 +54,7 @@ Common::Input::ButtonStatus TransformToButton(const Common::Input::CallbackStatu case Common::Input::InputType::Analog: status.value = TransformToTrigger(callback).pressed.value; status.toggle = callback.analog_status.properties.toggle; + status.inverted = callback.analog_status.properties.inverted_button; break; case Common::Input::InputType::Trigger: status.value = TransformToTrigger(callback).pressed.value; @@ -61,6 +62,9 @@ Common::Input::ButtonStatus TransformToButton(const Common::Input::CallbackStatu case Common::Input::InputType::Button: status = callback.button_status; break; + case Common::Input::InputType::Motion: + status.value = std::abs(callback.motion_status.gyro.x.raw_value) > 1.0f; + break; default: LOG_ERROR(Input, "Conversion from type {} to button not implemented", callback.type); break; @@ -226,6 +230,10 @@ Common::Input::TriggerStatus TransformToTrigger(const Common::Input::CallbackSta status = callback.trigger_status; calculate_button_value = false; break; + case Common::Input::InputType::Motion: + status.analog.properties.range = 1.0f; + raw_value = callback.motion_status.accel.x.raw_value; + break; default: LOG_ERROR(Input, "Conversion from type {} to trigger not implemented", callback.type); break; diff --git a/src/core/hid/motion_input.cpp b/src/core/hid/motion_input.cpp index 0dd66c1cc..b60478dbb 100644 --- a/src/core/hid/motion_input.cpp +++ b/src/core/hid/motion_input.cpp @@ -1,6 +1,8 @@ // SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project // SPDX-License-Identifier: GPL-2.0-or-later +#include <cmath> + #include "common/math_util.h" #include "core/hid/motion_input.h" @@ -51,6 +53,20 @@ void MotionInput::SetQuaternion(const Common::Quaternion<f32>& quaternion) { quat = quaternion; } +void MotionInput::SetEulerAngles(const Common::Vec3f& euler_angles) { + const float cr = std::cos(euler_angles.x * 0.5f); + const float sr = std::sin(euler_angles.x * 0.5f); + const float cp = std::cos(euler_angles.y * 0.5f); + const float sp = std::sin(euler_angles.y * 0.5f); + const float cy = std::cos(euler_angles.z * 0.5f); + const float sy = std::sin(euler_angles.z * 0.5f); + + quat.w = cr * cp * cy + sr * sp * sy; + quat.xyz.x = sr * cp * cy - cr * sp * sy; + quat.xyz.y = cr * sp * cy + sr * cp * sy; + quat.xyz.z = cr * cp * sy - sr * sp * cy; +} + void MotionInput::SetGyroBias(const Common::Vec3f& bias) { gyro_bias = bias; } @@ -222,6 +238,26 @@ Common::Vec3f MotionInput::GetRotations() const { return rotations; } +Common::Vec3f MotionInput::GetEulerAngles() const { + // roll (x-axis rotation) + const float sinr_cosp = 2 * (quat.w * quat.xyz.x + quat.xyz.y * quat.xyz.z); + const float cosr_cosp = 1 - 2 * (quat.xyz.x * quat.xyz.x + quat.xyz.y * quat.xyz.y); + + // pitch (y-axis rotation) + const float sinp = std::sqrt(1 + 2 * (quat.w * quat.xyz.y - quat.xyz.x * quat.xyz.z)); + const float cosp = std::sqrt(1 - 2 * (quat.w * quat.xyz.y - quat.xyz.x * quat.xyz.z)); + + // yaw (z-axis rotation) + const float siny_cosp = 2 * (quat.w * quat.xyz.z + quat.xyz.x * quat.xyz.y); + const float cosy_cosp = 1 - 2 * (quat.xyz.y * quat.xyz.y + quat.xyz.z * quat.xyz.z); + + return { + std::atan2(sinr_cosp, cosr_cosp), + 2 * std::atan2(sinp, cosp) - Common::PI / 2, + std::atan2(siny_cosp, cosy_cosp), + }; +} + void MotionInput::ResetOrientation() { if (!reset_enabled || only_accelerometer) { return; diff --git a/src/core/hid/motion_input.h b/src/core/hid/motion_input.h index 9f3fc1cf7..482719359 100644 --- a/src/core/hid/motion_input.h +++ b/src/core/hid/motion_input.h @@ -35,6 +35,7 @@ public: void SetAcceleration(const Common::Vec3f& acceleration); void SetGyroscope(const Common::Vec3f& gyroscope); void SetQuaternion(const Common::Quaternion<f32>& quaternion); + void SetEulerAngles(const Common::Vec3f& euler_angles); void SetGyroBias(const Common::Vec3f& bias); void SetGyroThreshold(f32 threshold); @@ -54,6 +55,7 @@ public: [[nodiscard]] Common::Vec3f GetGyroBias() const; [[nodiscard]] Common::Vec3f GetRotations() const; [[nodiscard]] Common::Quaternion<f32> GetQuaternion() const; + [[nodiscard]] Common::Vec3f GetEulerAngles() const; [[nodiscard]] bool IsMoving(f32 sensitivity) const; [[nodiscard]] bool IsCalibrated(f32 sensitivity) const; diff --git a/src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp b/src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp index 36d0d20d2..49bdc671e 100644 --- a/src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp +++ b/src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp @@ -35,12 +35,13 @@ namespace { using namespace Common::Literals; u32 GetMemorySizeForInit() { - return Settings::values.use_extended_memory_layout ? Smc::MemorySize_8GB : Smc::MemorySize_4GB; + return Settings::values.use_unsafe_extended_memory_layout ? Smc::MemorySize_8GB + : Smc::MemorySize_4GB; } Smc::MemoryArrangement GetMemoryArrangeForInit() { - return Settings::values.use_extended_memory_layout ? Smc::MemoryArrangement_8GB - : Smc::MemoryArrangement_4GB; + return Settings::values.use_unsafe_extended_memory_layout ? Smc::MemoryArrangement_8GB + : Smc::MemoryArrangement_4GB; } } // namespace diff --git a/src/core/hle/kernel/k_auto_object.h b/src/core/hle/kernel/k_auto_object.h index 9b71fe371..f384b1568 100644 --- a/src/core/hle/kernel/k_auto_object.h +++ b/src/core/hle/kernel/k_auto_object.h @@ -182,8 +182,8 @@ public: explicit KAutoObjectWithList(KernelCore& kernel) : KAutoObject(kernel) {} static int Compare(const KAutoObjectWithList& lhs, const KAutoObjectWithList& rhs) { - const u64 lid = lhs.GetId(); - const u64 rid = rhs.GetId(); + const uintptr_t lid = reinterpret_cast<uintptr_t>(std::addressof(lhs)); + const uintptr_t rid = reinterpret_cast<uintptr_t>(std::addressof(rhs)); if (lid < rid) { return -1; diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp index 4f3366c9d..f33600ca5 100644 --- a/src/core/hle/kernel/kernel.cpp +++ b/src/core/hle/kernel/kernel.cpp @@ -95,7 +95,7 @@ struct KernelCore::Impl { pt_heap_region.GetSize()); } - InitializeHackSharedMemory(); + InitializeHackSharedMemory(kernel); RegisterHostThread(nullptr); } @@ -216,10 +216,12 @@ struct KernelCore::Impl { auto* main_thread{Kernel::KThread::Create(system.Kernel())}; main_thread->SetCurrentCore(core); ASSERT(Kernel::KThread::InitializeMainThread(system, main_thread, core).IsSuccess()); + KThread::Register(system.Kernel(), main_thread); auto* idle_thread{Kernel::KThread::Create(system.Kernel())}; idle_thread->SetCurrentCore(core); ASSERT(Kernel::KThread::InitializeIdleThread(system, idle_thread, core).IsSuccess()); + KThread::Register(system.Kernel(), idle_thread); schedulers[i]->Initialize(main_thread, idle_thread, core); } @@ -230,6 +232,7 @@ struct KernelCore::Impl { const Core::Timing::CoreTiming& core_timing) { system_resource_limit = KResourceLimit::Create(system.Kernel()); system_resource_limit->Initialize(&core_timing); + KResourceLimit::Register(kernel, system_resource_limit); const auto sizes{memory_layout->GetTotalAndKernelMemorySizes()}; const auto total_size{sizes.first}; @@ -355,6 +358,7 @@ struct KernelCore::Impl { ASSERT(KThread::InitializeHighPriorityThread(system, shutdown_threads[core_id], {}, {}, core_id) .IsSuccess()); + KThread::Register(system.Kernel(), shutdown_threads[core_id]); } } @@ -729,7 +733,7 @@ struct KernelCore::Impl { memory_manager->Initialize(management_region.GetAddress(), management_region.GetSize()); } - void InitializeHackSharedMemory() { + void InitializeHackSharedMemory(KernelCore& kernel) { // Setup memory regions for emulated processes // TODO(bunnei): These should not be hardcoded regions initialized within the kernel constexpr std::size_t hid_size{0x40000}; @@ -746,14 +750,23 @@ struct KernelCore::Impl { hid_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None, Svc::MemoryPermission::Read, hid_size); + KSharedMemory::Register(kernel, hid_shared_mem); + font_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None, Svc::MemoryPermission::Read, font_size); + KSharedMemory::Register(kernel, font_shared_mem); + irs_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None, Svc::MemoryPermission::Read, irs_size); + KSharedMemory::Register(kernel, irs_shared_mem); + time_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None, Svc::MemoryPermission::Read, time_size); + KSharedMemory::Register(kernel, time_shared_mem); + hidbus_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None, Svc::MemoryPermission::Read, hidbus_size); + KSharedMemory::Register(kernel, hidbus_shared_mem); } std::mutex registered_objects_lock; @@ -1072,12 +1085,15 @@ static std::jthread RunHostThreadFunc(KernelCore& kernel, KProcess* process, // Commit the thread reservation. thread_reservation.Commit(); + // Register the thread. + KThread::Register(kernel, thread); + return std::jthread( [&kernel, thread, thread_name{std::move(thread_name)}, func{std::move(func)}] { // Set the thread name. Common::SetCurrentThreadName(thread_name.c_str()); - // Register the thread. + // Set the thread as current. kernel.RegisterHostThread(thread); // Run the callback. @@ -1099,6 +1115,9 @@ std::jthread KernelCore::RunOnHostCoreProcess(std::string&& process_name, // Ensure that we don't hold onto any extra references. SCOPE_EXIT({ process->Close(); }); + // Register the new process. + KProcess::Register(*this, process); + // Run the host thread. return RunHostThreadFunc(*this, process, std::move(process_name), std::move(func)); } @@ -1124,6 +1143,9 @@ void KernelCore::RunOnGuestCoreProcess(std::string&& process_name, std::function // Ensure that we don't hold onto any extra references. SCOPE_EXIT({ process->Close(); }); + // Register the new process. + KProcess::Register(*this, process); + // Reserve a new thread from the process resource limit. KScopedResourceReservation thread_reservation(process, LimitableResource::ThreadCountMax); ASSERT(thread_reservation.Succeeded()); @@ -1136,6 +1158,9 @@ void KernelCore::RunOnGuestCoreProcess(std::string&& process_name, std::function // Commit the thread reservation. thread_reservation.Commit(); + // Register the new thread. + KThread::Register(*this, thread); + // Begin running the thread. ASSERT(R_SUCCEEDED(thread->Run())); } diff --git a/src/core/hle/service/ipc_helpers.h b/src/core/hle/service/ipc_helpers.h index e4cb4e1f2..0e222362e 100644 --- a/src/core/hle/service/ipc_helpers.h +++ b/src/core/hle/service/ipc_helpers.h @@ -156,6 +156,7 @@ public: auto* session = Kernel::KSession::Create(kernel); session->Initialize(nullptr, 0); + Kernel::KSession::Register(kernel, session); auto next_manager = std::make_shared<Service::SessionRequestManager>( kernel, manager->GetServerManager()); diff --git a/src/core/hle/service/kernel_helpers.cpp b/src/core/hle/service/kernel_helpers.cpp index a39ce5212..6a313a03b 100644 --- a/src/core/hle/service/kernel_helpers.cpp +++ b/src/core/hle/service/kernel_helpers.cpp @@ -25,6 +25,9 @@ ServiceContext::ServiceContext(Core::System& system_, std::string name_) Kernel::KProcess::ProcessType::KernelInternal, kernel.GetSystemResourceLimit()) .IsSuccess()); + + // Register the process. + Kernel::KProcess::Register(kernel, process); process_created = true; } diff --git a/src/core/hle/service/mutex.cpp b/src/core/hle/service/mutex.cpp index 07589a0f0..b0ff71d1b 100644 --- a/src/core/hle/service/mutex.cpp +++ b/src/core/hle/service/mutex.cpp @@ -12,6 +12,9 @@ Mutex::Mutex(Core::System& system) : m_system(system) { m_event = Kernel::KEvent::Create(system.Kernel()); m_event->Initialize(nullptr); + // Register the event. + Kernel::KEvent::Register(system.Kernel(), m_event); + ASSERT(R_SUCCEEDED(m_event->Signal())); } diff --git a/src/core/hle/service/server_manager.cpp b/src/core/hle/service/server_manager.cpp index 6b4a1291e..156bc27d8 100644 --- a/src/core/hle/service/server_manager.cpp +++ b/src/core/hle/service/server_manager.cpp @@ -33,6 +33,9 @@ ServerManager::ServerManager(Core::System& system) : m_system{system}, m_serve_m // Initialize event. m_event = Kernel::KEvent::Create(system.Kernel()); m_event->Initialize(nullptr); + + // Register event. + Kernel::KEvent::Register(system.Kernel(), m_event); } ServerManager::~ServerManager() { @@ -160,6 +163,9 @@ Result ServerManager::ManageDeferral(Kernel::KEvent** out_event) { // Initialize the event. m_deferral_event->Initialize(nullptr); + // Register the event. + Kernel::KEvent::Register(m_system.Kernel(), m_deferral_event); + // Set the output. *out_event = m_deferral_event; diff --git a/src/core/hle/service/sm/sm.cpp b/src/core/hle/service/sm/sm.cpp index c45be5726..1608fa24c 100644 --- a/src/core/hle/service/sm/sm.cpp +++ b/src/core/hle/service/sm/sm.cpp @@ -64,6 +64,9 @@ Result ServiceManager::RegisterService(std::string name, u32 max_sessions, auto* port = Kernel::KPort::Create(kernel); port->Initialize(ServerSessionCountMax, false, 0); + // Register the port. + Kernel::KPort::Register(kernel, port); + service_ports.emplace(name, port); registered_services.emplace(name, handler); if (deferral_event) { diff --git a/src/core/hle/service/sm/sm_controller.cpp b/src/core/hle/service/sm/sm_controller.cpp index 419c1df2b..7dce28fe0 100644 --- a/src/core/hle/service/sm/sm_controller.cpp +++ b/src/core/hle/service/sm/sm_controller.cpp @@ -49,6 +49,9 @@ void Controller::CloneCurrentObject(HLERequestContext& ctx) { // Commit the session reservation. session_reservation.Commit(); + // Register the session. + Kernel::KSession::Register(system.Kernel(), session); + // Register with server manager. session_manager->GetServerManager().RegisterSession(&session->GetServerSession(), session_manager); diff --git a/src/core/memory.cpp b/src/core/memory.cpp index 432310632..a9667463f 100644 --- a/src/core/memory.cpp +++ b/src/core/memory.cpp @@ -462,7 +462,7 @@ struct Memory::Impl { } if (Settings::IsFastmemEnabled()) { - const bool is_read_enable = Settings::IsGPULevelHigh() || !cached; + const bool is_read_enable = !Settings::IsGPULevelExtreme() || !cached; system.DeviceMemory().buffer.Protect(vaddr, size, is_read_enable, !cached); } diff --git a/src/input_common/input_engine.cpp b/src/input_common/input_engine.cpp index 91aa96aa7..49f5e7f54 100644 --- a/src/input_common/input_engine.cpp +++ b/src/input_common/input_engine.cpp @@ -58,6 +58,8 @@ void InputEngine::SetHatButton(const PadIdentifier& identifier, int button, u8 v } void InputEngine::SetAxis(const PadIdentifier& identifier, int axis, f32 value) { + value /= 2.0f; + value -= 0.5f; { std::scoped_lock lock{mutex}; ControllerData& controller = controller_list.at(identifier); diff --git a/src/input_common/input_mapping.cpp b/src/input_common/input_mapping.cpp index 9361b00c5..8c2ee4eb3 100644 --- a/src/input_common/input_mapping.cpp +++ b/src/input_common/input_mapping.cpp @@ -82,6 +82,9 @@ void MappingFactory::RegisterButton(const MappingData& data) { new_input.Set("axis", data.index); new_input.Set("threshold", 0.5f); break; + case EngineInputType::Motion: + new_input.Set("motion", data.index); + break; default: return; } diff --git a/src/input_common/input_poller.cpp b/src/input_common/input_poller.cpp index 8c6a6521a..5c2c4a463 100644 --- a/src/input_common/input_poller.cpp +++ b/src/input_common/input_poller.cpp @@ -939,6 +939,7 @@ std::unique_ptr<Common::Input::InputDevice> InputFactory::CreateAnalogDevice( .threshold = std::clamp(params.Get("threshold", 0.5f), 0.0f, 1.0f), .offset = std::clamp(params.Get("offset", 0.0f), -1.0f, 1.0f), .inverted = params.Get("invert", "+") == "-", + .inverted_button = params.Get("inverted", false) != 0, .toggle = params.Get("toggle", false) != 0, }; input_engine->PreSetController(identifier); diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 7975564b5..e534e1e9c 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -1426,7 +1426,7 @@ bool BufferCache<P>::SynchronizeBufferNoModified(Buffer& buffer, VAddr cpu_addr, .size = sub_size, }); total_size_bytes += sub_size; - largest_copy = std::max(largest_copy, sub_size); + largest_copy = std::max<u64>(largest_copy, sub_size); } const std::span<BufferCopy> copies_span(copies.data(), copies.size()); UploadMemory(buffer, total_size_bytes, largest_copy, copies_span); diff --git a/src/video_core/buffer_cache/memory_tracker_base.h b/src/video_core/buffer_cache/memory_tracker_base.h index 4bc59017f..dc4ebfcaa 100644 --- a/src/video_core/buffer_cache/memory_tracker_base.h +++ b/src/video_core/buffer_cache/memory_tracker_base.h @@ -170,7 +170,8 @@ private: std::size_t page_index{cpu_address >> HIGHER_PAGE_BITS}; u64 page_offset{cpu_address & HIGHER_PAGE_MASK}; while (remaining_size > 0) { - const std::size_t copy_amount{std::min(HIGHER_PAGE_SIZE - page_offset, remaining_size)}; + const std::size_t copy_amount{ + std::min<std::size_t>(HIGHER_PAGE_SIZE - page_offset, remaining_size)}; auto* manager{top_tier[page_index]}; if (manager) { if constexpr (BOOL_BREAK) { @@ -206,7 +207,8 @@ private: u64 begin = std::numeric_limits<u64>::max(); u64 end = 0; while (remaining_size > 0) { - const std::size_t copy_amount{std::min(HIGHER_PAGE_SIZE - page_offset, remaining_size)}; + const std::size_t copy_amount{ + std::min<std::size_t>(HIGHER_PAGE_SIZE - page_offset, remaining_size)}; auto* manager{top_tier[page_index]}; const auto execute = [&] { auto [new_begin, new_end] = func(manager, page_offset, copy_amount); diff --git a/src/video_core/compatible_formats.cpp b/src/video_core/compatible_formats.cpp index 4e75f33ca..ab4f4d407 100644 --- a/src/video_core/compatible_formats.cpp +++ b/src/video_core/compatible_formats.cpp @@ -126,15 +126,14 @@ constexpr std::array VIEW_CLASS_ASTC_8x8_RGBA{ PixelFormat::ASTC_2D_8X8_SRGB, }; -// Missing formats: -// PixelFormat::ASTC_2D_10X5_UNORM -// PixelFormat::ASTC_2D_10X5_SRGB - -// Missing formats: -// PixelFormat::ASTC_2D_10X6_SRGB +constexpr std::array VIEW_CLASS_ASTC_10x5_RGBA{ + PixelFormat::ASTC_2D_10X5_UNORM, + PixelFormat::ASTC_2D_10X5_SRGB, +}; constexpr std::array VIEW_CLASS_ASTC_10x6_RGBA{ PixelFormat::ASTC_2D_10X6_UNORM, + PixelFormat::ASTC_2D_10X6_SRGB, }; constexpr std::array VIEW_CLASS_ASTC_10x8_RGBA{ @@ -147,9 +146,10 @@ constexpr std::array VIEW_CLASS_ASTC_10x10_RGBA{ PixelFormat::ASTC_2D_10X10_SRGB, }; -// Missing formats -// ASTC_2D_12X10_UNORM, -// ASTC_2D_12X10_SRGB, +constexpr std::array VIEW_CLASS_ASTC_12x10_RGBA{ + PixelFormat::ASTC_2D_12X10_UNORM, + PixelFormat::ASTC_2D_12X10_SRGB, +}; constexpr std::array VIEW_CLASS_ASTC_12x12_RGBA{ PixelFormat::ASTC_2D_12X12_UNORM, @@ -229,9 +229,11 @@ constexpr Table MakeViewTable() { EnableRange(view, VIEW_CLASS_ASTC_6x6_RGBA); EnableRange(view, VIEW_CLASS_ASTC_8x5_RGBA); EnableRange(view, VIEW_CLASS_ASTC_8x8_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_10x5_RGBA); EnableRange(view, VIEW_CLASS_ASTC_10x6_RGBA); EnableRange(view, VIEW_CLASS_ASTC_10x8_RGBA); EnableRange(view, VIEW_CLASS_ASTC_10x10_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_12x10_RGBA); EnableRange(view, VIEW_CLASS_ASTC_12x12_RGBA); return view; } diff --git a/src/video_core/fence_manager.h b/src/video_core/fence_manager.h index c390ac91b..3b2f6aab6 100644 --- a/src/video_core/fence_manager.h +++ b/src/video_core/fence_manager.h @@ -4,13 +4,20 @@ #pragma once #include <algorithm> +#include <condition_variable> #include <cstring> #include <deque> #include <functional> #include <memory> +#include <mutex> +#include <thread> #include <queue> #include "common/common_types.h" +#include "common/microprofile.h" +#include "common/scope_exit.h" +#include "common/settings.h" +#include "common/thread.h" #include "video_core/delayed_destruction_ring.h" #include "video_core/gpu.h" #include "video_core/host1x/host1x.h" @@ -23,15 +30,26 @@ class FenceBase { public: explicit FenceBase(bool is_stubbed_) : is_stubbed{is_stubbed_} {} + bool IsStubbed() const { + return is_stubbed; + } + protected: bool is_stubbed; }; -template <typename TFence, typename TTextureCache, typename TTBufferCache, typename TQueryCache> +template <typename Traits> class FenceManager { + using TFence = typename Traits::FenceType; + using TTextureCache = typename Traits::TextureCacheType; + using TBufferCache = typename Traits::BufferCacheType; + using TQueryCache = typename Traits::QueryCacheType; + static constexpr bool can_async_check = Traits::HAS_ASYNC_CHECK; + public: /// Notify the fence manager about a new frame void TickFrame() { + std::unique_lock lock(ring_guard); delayed_destruction_ring.Tick(); } @@ -46,17 +64,33 @@ public: } void SignalFence(std::function<void()>&& func) { - TryReleasePendingFences(); + rasterizer.InvalidateGPUCache(); + bool delay_fence = Settings::IsGPULevelHigh(); + if constexpr (!can_async_check) { + TryReleasePendingFences<false>(); + } const bool should_flush = ShouldFlush(); CommitAsyncFlushes(); - uncommitted_operations.emplace_back(std::move(func)); - CommitOperations(); TFence new_fence = CreateFence(!should_flush); - fences.push(new_fence); + if constexpr (can_async_check) { + guard.lock(); + } + if (delay_fence) { + uncommitted_operations.emplace_back(std::move(func)); + } + pending_operations.emplace_back(std::move(uncommitted_operations)); QueueFence(new_fence); + if (!delay_fence) { + func(); + } + fences.push(std::move(new_fence)); if (should_flush) { rasterizer.FlushCommands(); } + if constexpr (can_async_check) { + guard.unlock(); + cv.notify_all(); + } } void SignalSyncPoint(u32 value) { @@ -66,29 +100,30 @@ public: } void WaitPendingFences() { - while (!fences.empty()) { - TFence& current_fence = fences.front(); - if (ShouldWait()) { - WaitFence(current_fence); - } - PopAsyncFlushes(); - auto operations = std::move(pending_operations.front()); - pending_operations.pop_front(); - for (auto& operation : operations) { - operation(); - } - PopFence(); + if constexpr (!can_async_check) { + TryReleasePendingFences<true>(); } } protected: explicit FenceManager(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_, - TTextureCache& texture_cache_, TTBufferCache& buffer_cache_, + TTextureCache& texture_cache_, TBufferCache& buffer_cache_, TQueryCache& query_cache_) : rasterizer{rasterizer_}, gpu{gpu_}, syncpoint_manager{gpu.Host1x().GetSyncpointManager()}, - texture_cache{texture_cache_}, buffer_cache{buffer_cache_}, query_cache{query_cache_} {} + texture_cache{texture_cache_}, buffer_cache{buffer_cache_}, query_cache{query_cache_} { + if constexpr (can_async_check) { + fence_thread = + std::jthread([this](std::stop_token token) { ReleaseThreadFunc(token); }); + } + } - virtual ~FenceManager() = default; + virtual ~FenceManager() { + if constexpr (can_async_check) { + fence_thread.request_stop(); + cv.notify_all(); + fence_thread.join(); + } + } /// Creates a Fence Interface, does not create a backend fence if 'is_stubbed' is /// true @@ -104,15 +139,20 @@ protected: Tegra::GPU& gpu; Tegra::Host1x::SyncpointManager& syncpoint_manager; TTextureCache& texture_cache; - TTBufferCache& buffer_cache; + TBufferCache& buffer_cache; TQueryCache& query_cache; private: + template <bool force_wait> void TryReleasePendingFences() { while (!fences.empty()) { TFence& current_fence = fences.front(); if (ShouldWait() && !IsFenceSignaled(current_fence)) { - return; + if constexpr (force_wait) { + WaitFence(current_fence); + } else { + return; + } } PopAsyncFlushes(); auto operations = std::move(pending_operations.front()); @@ -120,7 +160,49 @@ private: for (auto& operation : operations) { operation(); } - PopFence(); + { + std::unique_lock lock(ring_guard); + delayed_destruction_ring.Push(std::move(current_fence)); + } + fences.pop(); + } + } + + void ReleaseThreadFunc(std::stop_token stop_token) { + std::string name = "GPUFencingThread"; + MicroProfileOnThreadCreate(name.c_str()); + + // Cleanup + SCOPE_EXIT({ MicroProfileOnThreadExit(); }); + + Common::SetCurrentThreadName(name.c_str()); + Common::SetCurrentThreadPriority(Common::ThreadPriority::High); + + TFence current_fence; + std::deque<std::function<void()>> current_operations; + while (!stop_token.stop_requested()) { + { + std::unique_lock lock(guard); + cv.wait(lock, [&] { return stop_token.stop_requested() || !fences.empty(); }); + if (stop_token.stop_requested()) [[unlikely]] { + return; + } + current_fence = std::move(fences.front()); + current_operations = std::move(pending_operations.front()); + fences.pop(); + pending_operations.pop_front(); + } + if (!current_fence->IsStubbed()) { + WaitFence(current_fence); + } + PopAsyncFlushes(); + for (auto& operation : current_operations) { + operation(); + } + { + std::unique_lock lock(ring_guard); + delayed_destruction_ring.Push(std::move(current_fence)); + } } } @@ -154,19 +236,16 @@ private: query_cache.CommitAsyncFlushes(); } - void PopFence() { - delayed_destruction_ring.Push(std::move(fences.front())); - fences.pop(); - } - - void CommitOperations() { - pending_operations.emplace_back(std::move(uncommitted_operations)); - } - std::queue<TFence> fences; std::deque<std::function<void()>> uncommitted_operations; std::deque<std::deque<std::function<void()>>> pending_operations; + std::mutex guard; + std::mutex ring_guard; + std::condition_variable cv; + + std::jthread fence_thread; + DelayedDestructionRing<TFence, 6> delayed_destruction_ring; }; diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp index 01fb5b546..7b2cde7a7 100644 --- a/src/video_core/memory_manager.cpp +++ b/src/video_core/memory_manager.cpp @@ -82,6 +82,7 @@ void MemoryManager::SetEntry(size_t position, MemoryManager::EntryType entry) { } PTEKind MemoryManager::GetPageKind(GPUVAddr gpu_addr) const { + std::unique_lock<std::mutex> lock(guard); return kind_map.GetValueAt(gpu_addr); } @@ -160,7 +161,10 @@ GPUVAddr MemoryManager::BigPageTableOp(GPUVAddr gpu_addr, [[maybe_unused]] VAddr } remaining_size -= big_page_size; } - kind_map.Map(gpu_addr, gpu_addr + size, kind); + { + std::unique_lock<std::mutex> lock(guard); + kind_map.Map(gpu_addr, gpu_addr + size, kind); + } return gpu_addr; } @@ -553,6 +557,7 @@ size_t MemoryManager::MaxContinuousRange(GPUVAddr gpu_addr, size_t size) const { } size_t MemoryManager::GetMemoryLayoutSize(GPUVAddr gpu_addr, size_t max_size) const { + std::unique_lock<std::mutex> lock(guard); return kind_map.GetContinuousSizeFrom(gpu_addr); } @@ -745,10 +750,10 @@ void MemoryManager::FlushCaching() { return; } accumulator->Callback([this](GPUVAddr addr, size_t size) { - GetSubmappedRangeImpl<false>(addr, size, page_stash); + GetSubmappedRangeImpl<false>(addr, size, page_stash2); }); - rasterizer->InnerInvalidation(page_stash); - page_stash.clear(); + rasterizer->InnerInvalidation(page_stash2); + page_stash2.clear(); accumulator->Clear(); } diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h index fbbe856c4..794535122 100644 --- a/src/video_core/memory_manager.h +++ b/src/video_core/memory_manager.h @@ -5,6 +5,7 @@ #include <atomic> #include <map> +#include <mutex> #include <optional> #include <vector> @@ -215,6 +216,9 @@ private: std::vector<u64> big_page_continuous; std::vector<std::pair<VAddr, std::size_t>> page_stash{}; + std::vector<std::pair<VAddr, std::size_t>> page_stash2{}; + + mutable std::mutex guard; static constexpr size_t continuous_bits = 64; diff --git a/src/video_core/query_cache.h b/src/video_core/query_cache.h index 8906ba6d8..941de95c1 100644 --- a/src/video_core/query_cache.h +++ b/src/video_core/query_cache.h @@ -6,6 +6,7 @@ #include <algorithm> #include <array> #include <cstring> +#include <functional> #include <iterator> #include <list> #include <memory> @@ -17,13 +18,19 @@ #include "common/assert.h" #include "common/settings.h" +#include "core/memory.h" #include "video_core/control/channel_state_cache.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" +#include "video_core/texture_cache/slot_vector.h" namespace VideoCommon { +using AsyncJobId = SlotId; + +static constexpr AsyncJobId NULL_ASYNC_JOB_ID{0}; + template <class QueryCache, class HostCounter> class CounterStreamBase { public: @@ -93,9 +100,13 @@ private: template <class QueryCache, class CachedQuery, class CounterStream, class HostCounter> class QueryCacheBase : public VideoCommon::ChannelSetupCaches<VideoCommon::ChannelInfo> { public: - explicit QueryCacheBase(VideoCore::RasterizerInterface& rasterizer_) - : rasterizer{rasterizer_}, streams{{CounterStream{static_cast<QueryCache&>(*this), - VideoCore::QueryType::SamplesPassed}}} {} + explicit QueryCacheBase(VideoCore::RasterizerInterface& rasterizer_, + Core::Memory::Memory& cpu_memory_) + : rasterizer{rasterizer_}, + cpu_memory{cpu_memory_}, streams{{CounterStream{static_cast<QueryCache&>(*this), + VideoCore::QueryType::SamplesPassed}}} { + (void)slot_async_jobs.insert(); // Null value + } void InvalidateRegion(VAddr addr, std::size_t size) { std::unique_lock lock{mutex}; @@ -126,10 +137,15 @@ public: query = Register(type, *cpu_addr, host_ptr, timestamp.has_value()); } - query->BindCounter(Stream(type).Current(), timestamp); - if (Settings::values.use_asynchronous_gpu_emulation.GetValue()) { - AsyncFlushQuery(*cpu_addr); + auto result = query->BindCounter(Stream(type).Current(), timestamp); + if (result) { + auto async_job_id = query->GetAsyncJob(); + auto& async_job = slot_async_jobs[async_job_id]; + async_job.collected = true; + async_job.value = *result; + query->SetAsyncJob(NULL_ASYNC_JOB_ID); } + AsyncFlushQuery(query, timestamp, lock); } /// Updates counters from GPU state. Expected to be called once per draw, clear or dispatch. @@ -173,15 +189,18 @@ public: } void CommitAsyncFlushes() { + std::unique_lock lock{mutex}; committed_flushes.push_back(uncommitted_flushes); uncommitted_flushes.reset(); } bool HasUncommittedFlushes() const { + std::unique_lock lock{mutex}; return uncommitted_flushes != nullptr; } bool ShouldWaitAsyncFlushes() const { + std::unique_lock lock{mutex}; if (committed_flushes.empty()) { return false; } @@ -189,6 +208,7 @@ public: } void PopAsyncFlushes() { + std::unique_lock lock{mutex}; if (committed_flushes.empty()) { return; } @@ -197,15 +217,25 @@ public: committed_flushes.pop_front(); return; } - for (VAddr query_address : *flush_list) { - FlushAndRemoveRegion(query_address, 4); + for (AsyncJobId async_job_id : *flush_list) { + AsyncJob& async_job = slot_async_jobs[async_job_id]; + if (!async_job.collected) { + FlushAndRemoveRegion(async_job.query_location, 2, true); + } } committed_flushes.pop_front(); } private: + struct AsyncJob { + bool collected = false; + u64 value = 0; + VAddr query_location = 0; + std::optional<u64> timestamp{}; + }; + /// Flushes a memory range to guest memory and removes it from the cache. - void FlushAndRemoveRegion(VAddr addr, std::size_t size) { + void FlushAndRemoveRegion(VAddr addr, std::size_t size, bool async = false) { const u64 addr_begin = addr; const u64 addr_end = addr_begin + size; const auto in_range = [addr_begin, addr_end](const CachedQuery& query) { @@ -226,7 +256,16 @@ private: continue; } rasterizer.UpdatePagesCachedCount(query.GetCpuAddr(), query.SizeInBytes(), -1); - query.Flush(); + AsyncJobId async_job_id = query.GetAsyncJob(); + auto flush_result = query.Flush(async); + if (async_job_id == NULL_ASYNC_JOB_ID) { + ASSERT_MSG(false, "This should not be reachable at all"); + continue; + } + AsyncJob& async_job = slot_async_jobs[async_job_id]; + async_job.collected = true; + async_job.value = flush_result; + query.SetAsyncJob(NULL_ASYNC_JOB_ID); } std::erase_if(contents, in_range); } @@ -253,26 +292,60 @@ private: return found != std::end(contents) ? &*found : nullptr; } - void AsyncFlushQuery(VAddr addr) { - if (!uncommitted_flushes) { - uncommitted_flushes = std::make_shared<std::vector<VAddr>>(); + void AsyncFlushQuery(CachedQuery* query, std::optional<u64> timestamp, + std::unique_lock<std::recursive_mutex>& lock) { + const AsyncJobId new_async_job_id = slot_async_jobs.insert(); + { + AsyncJob& async_job = slot_async_jobs[new_async_job_id]; + query->SetAsyncJob(new_async_job_id); + async_job.query_location = query->GetCpuAddr(); + async_job.collected = false; + + if (!uncommitted_flushes) { + uncommitted_flushes = std::make_shared<std::vector<AsyncJobId>>(); + } + uncommitted_flushes->push_back(new_async_job_id); } - uncommitted_flushes->push_back(addr); + lock.unlock(); + std::function<void()> operation([this, new_async_job_id, timestamp] { + std::unique_lock local_lock{mutex}; + AsyncJob& async_job = slot_async_jobs[new_async_job_id]; + u64 value = async_job.value; + VAddr address = async_job.query_location; + slot_async_jobs.erase(new_async_job_id); + local_lock.unlock(); + if (timestamp) { + u64 timestamp_value = *timestamp; + cpu_memory.WriteBlockUnsafe(address + sizeof(u64), ×tamp_value, sizeof(u64)); + cpu_memory.WriteBlockUnsafe(address, &value, sizeof(u64)); + rasterizer.InvalidateRegion(address, sizeof(u64) * 2, + VideoCommon::CacheType::NoQueryCache); + } else { + u32 small_value = static_cast<u32>(value); + cpu_memory.WriteBlockUnsafe(address, &small_value, sizeof(u32)); + rasterizer.InvalidateRegion(address, sizeof(u32), + VideoCommon::CacheType::NoQueryCache); + } + }); + rasterizer.SyncOperation(std::move(operation)); } static constexpr std::uintptr_t YUZU_PAGESIZE = 4096; static constexpr unsigned YUZU_PAGEBITS = 12; + SlotVector<AsyncJob> slot_async_jobs; + VideoCore::RasterizerInterface& rasterizer; + Core::Memory::Memory& cpu_memory; - std::recursive_mutex mutex; + mutable std::recursive_mutex mutex; std::unordered_map<u64, std::vector<CachedQuery>> cached_queries; std::array<CounterStream, VideoCore::NumQueryTypes> streams; - std::shared_ptr<std::vector<VAddr>> uncommitted_flushes{}; - std::list<std::shared_ptr<std::vector<VAddr>>> committed_flushes; + std::shared_ptr<std::vector<AsyncJobId>> uncommitted_flushes{}; + std::list<std::shared_ptr<std::vector<AsyncJobId>>> committed_flushes; }; template <class QueryCache, class HostCounter> @@ -291,12 +364,12 @@ public: virtual ~HostCounterBase() = default; /// Returns the current value of the query. - u64 Query() { + u64 Query(bool async = false) { if (result) { return *result; } - u64 value = BlockingQuery() + base_result; + u64 value = BlockingQuery(async) + base_result; if (dependency) { value += dependency->Query(); dependency = nullptr; @@ -317,7 +390,7 @@ public: protected: /// Returns the value of query from the backend API blocking as needed. - virtual u64 BlockingQuery() const = 0; + virtual u64 BlockingQuery(bool async = false) const = 0; private: std::shared_ptr<HostCounter> dependency; ///< Counter to add to this value. @@ -340,26 +413,33 @@ public: CachedQueryBase& operator=(const CachedQueryBase&) = delete; /// Flushes the query to guest memory. - virtual void Flush() { + virtual u64 Flush(bool async = false) { // When counter is nullptr it means that it's just been reset. We are supposed to write a // zero in these cases. - const u64 value = counter ? counter->Query() : 0; + const u64 value = counter ? counter->Query(async) : 0; + if (async) { + return value; + } std::memcpy(host_ptr, &value, sizeof(u64)); if (timestamp) { std::memcpy(host_ptr + TIMESTAMP_OFFSET, &*timestamp, sizeof(u64)); } + return value; } /// Binds a counter to this query. - void BindCounter(std::shared_ptr<HostCounter> counter_, std::optional<u64> timestamp_) { + std::optional<u64> BindCounter(std::shared_ptr<HostCounter> counter_, + std::optional<u64> timestamp_) { + std::optional<u64> result{}; if (counter) { // If there's an old counter set it means the query is being rewritten by the game. // To avoid losing the data forever, flush here. - Flush(); + result = std::make_optional(Flush()); } counter = std::move(counter_); timestamp = timestamp_; + return result; } VAddr GetCpuAddr() const noexcept { @@ -374,6 +454,14 @@ public: return with_timestamp ? LARGE_QUERY_SIZE : SMALL_QUERY_SIZE; } + void SetAsyncJob(AsyncJobId assigned_async_job_) { + assigned_async_job = assigned_async_job_; + } + + AsyncJobId GetAsyncJob() const { + return assigned_async_job; + } + protected: /// Returns true when querying the counter may potentially block. bool WaitPending() const noexcept { @@ -389,6 +477,7 @@ private: u8* host_ptr; ///< Writable host pointer. std::shared_ptr<HostCounter> counter; ///< Host counter to query, owns the dependency tree. std::optional<u64> timestamp; ///< Timestamp to flush to guest memory. + AsyncJobId assigned_async_job; }; } // namespace VideoCommon diff --git a/src/video_core/renderer_opengl/gl_fence_manager.h b/src/video_core/renderer_opengl/gl_fence_manager.h index f1446e732..e21b19dcc 100644 --- a/src/video_core/renderer_opengl/gl_fence_manager.h +++ b/src/video_core/renderer_opengl/gl_fence_manager.h @@ -30,7 +30,17 @@ private: }; using Fence = std::shared_ptr<GLInnerFence>; -using GenericFenceManager = VideoCommon::FenceManager<Fence, TextureCache, BufferCache, QueryCache>; + +struct FenceManagerParams { + using FenceType = Fence; + using BufferCacheType = BufferCache; + using TextureCacheType = TextureCache; + using QueryCacheType = QueryCache; + + static constexpr bool HAS_ASYNC_CHECK = false; +}; + +using GenericFenceManager = VideoCommon::FenceManager<FenceManagerParams>; class FenceManagerOpenGL final : public GenericFenceManager { public: diff --git a/src/video_core/renderer_opengl/gl_query_cache.cpp b/src/video_core/renderer_opengl/gl_query_cache.cpp index 5070db441..99d7347f5 100644 --- a/src/video_core/renderer_opengl/gl_query_cache.cpp +++ b/src/video_core/renderer_opengl/gl_query_cache.cpp @@ -26,8 +26,8 @@ constexpr GLenum GetTarget(VideoCore::QueryType type) { } // Anonymous namespace -QueryCache::QueryCache(RasterizerOpenGL& rasterizer_) - : QueryCacheBase(rasterizer_), gl_rasterizer{rasterizer_} {} +QueryCache::QueryCache(RasterizerOpenGL& rasterizer_, Core::Memory::Memory& cpu_memory_) + : QueryCacheBase(rasterizer_, cpu_memory_), gl_rasterizer{rasterizer_} {} QueryCache::~QueryCache() = default; @@ -74,7 +74,7 @@ void HostCounter::EndQuery() { glEndQuery(GetTarget(type)); } -u64 HostCounter::BlockingQuery() const { +u64 HostCounter::BlockingQuery([[maybe_unused]] bool async) const { GLint64 value; glGetQueryObjecti64v(query.handle, GL_QUERY_RESULT, &value); return static_cast<u64>(value); @@ -96,7 +96,7 @@ CachedQuery& CachedQuery::operator=(CachedQuery&& rhs) noexcept { return *this; } -void CachedQuery::Flush() { +u64 CachedQuery::Flush([[maybe_unused]] bool async) { // Waiting for a query while another query of the same target is enabled locks Nvidia's driver. // To avoid this disable and re-enable keeping the dependency stream. // But we only have to do this if we have pending waits to be done. @@ -106,11 +106,13 @@ void CachedQuery::Flush() { stream.Update(false); } - VideoCommon::CachedQueryBase<HostCounter>::Flush(); + auto result = VideoCommon::CachedQueryBase<HostCounter>::Flush(); if (slice_counter) { stream.Update(true); } + + return result; } } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_query_cache.h b/src/video_core/renderer_opengl/gl_query_cache.h index 14ce59990..872513f22 100644 --- a/src/video_core/renderer_opengl/gl_query_cache.h +++ b/src/video_core/renderer_opengl/gl_query_cache.h @@ -28,7 +28,7 @@ using CounterStream = VideoCommon::CounterStreamBase<QueryCache, HostCounter>; class QueryCache final : public VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, HostCounter> { public: - explicit QueryCache(RasterizerOpenGL& rasterizer_); + explicit QueryCache(RasterizerOpenGL& rasterizer_, Core::Memory::Memory& cpu_memory_); ~QueryCache(); OGLQuery AllocateQuery(VideoCore::QueryType type); @@ -51,7 +51,7 @@ public: void EndQuery(); private: - u64 BlockingQuery() const override; + u64 BlockingQuery(bool async = false) const override; QueryCache& cache; const VideoCore::QueryType type; @@ -70,7 +70,7 @@ public: CachedQuery(const CachedQuery&) = delete; CachedQuery& operator=(const CachedQuery&) = delete; - void Flush() override; + u64 Flush(bool async = false) override; private: QueryCache* cache; diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index 4993d4709..0089b4b27 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -63,7 +63,7 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window_, Tegra buffer_cache(*this, cpu_memory_, buffer_cache_runtime), shader_cache(*this, emu_window_, device, texture_cache, buffer_cache, program_manager, state_tracker, gpu.ShaderNotify()), - query_cache(*this), accelerate_dma(buffer_cache, texture_cache), + query_cache(*this, cpu_memory_), accelerate_dma(buffer_cache, texture_cache), fence_manager(*this, gpu, texture_cache, buffer_cache, query_cache), blit_image(program_manager_) {} diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index 032a8ebc5..47cccd0e5 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -861,9 +861,12 @@ GLuint Image::StorageHandle() noexcept { case PixelFormat::ASTC_2D_8X5_SRGB: case PixelFormat::ASTC_2D_5X4_SRGB: case PixelFormat::ASTC_2D_5X5_SRGB: + case PixelFormat::ASTC_2D_10X5_SRGB: + case PixelFormat::ASTC_2D_10X6_SRGB: case PixelFormat::ASTC_2D_10X8_SRGB: case PixelFormat::ASTC_2D_6X6_SRGB: case PixelFormat::ASTC_2D_10X10_SRGB: + case PixelFormat::ASTC_2D_12X10_SRGB: case PixelFormat::ASTC_2D_12X12_SRGB: case PixelFormat::ASTC_2D_8X6_SRGB: case PixelFormat::ASTC_2D_6X5_SRGB: diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index ef1190e1f..c7dc7e0a1 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -100,10 +100,13 @@ constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> FORMAT_TAB {GL_COMPRESSED_RGBA_ASTC_6x6_KHR}, // ASTC_2D_6X6_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x6_KHR}, // ASTC_2D_6X6_SRGB {GL_COMPRESSED_RGBA_ASTC_10x6_KHR}, // ASTC_2D_10X6_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x6_KHR}, // ASTC_2D_10X6_SRGB {GL_COMPRESSED_RGBA_ASTC_10x5_KHR}, // ASTC_2D_10X5_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x5_KHR}, // ASTC_2D_10X5_SRGB {GL_COMPRESSED_RGBA_ASTC_10x10_KHR}, // ASTC_2D_10X10_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x10_KHR}, // ASTC_2D_10X10_SRGB + {GL_COMPRESSED_RGBA_ASTC_12x10_KHR}, // ASTC_2D_12X10_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x10_KHR}, // ASTC_2D_12X10_SRGB {GL_COMPRESSED_RGBA_ASTC_12x12_KHR}, // ASTC_2D_12X12_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x12_KHR}, // ASTC_2D_12X12_SRGB {GL_COMPRESSED_RGBA_ASTC_8x6_KHR}, // ASTC_2D_8X6_UNORM diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp index 5dce51be8..8853cf0f7 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp @@ -197,10 +197,13 @@ struct FormatTuple { {VK_FORMAT_ASTC_6x6_UNORM_BLOCK}, // ASTC_2D_6X6_UNORM {VK_FORMAT_ASTC_6x6_SRGB_BLOCK}, // ASTC_2D_6X6_SRGB {VK_FORMAT_ASTC_10x6_UNORM_BLOCK}, // ASTC_2D_10X6_UNORM + {VK_FORMAT_ASTC_10x6_SRGB_BLOCK}, // ASTC_2D_10X6_SRGB {VK_FORMAT_ASTC_10x5_UNORM_BLOCK}, // ASTC_2D_10X5_UNORM {VK_FORMAT_ASTC_10x5_SRGB_BLOCK}, // ASTC_2D_10X5_SRGB {VK_FORMAT_ASTC_10x10_UNORM_BLOCK}, // ASTC_2D_10X10_UNORM {VK_FORMAT_ASTC_10x10_SRGB_BLOCK}, // ASTC_2D_10X10_SRGB + {VK_FORMAT_ASTC_12x10_UNORM_BLOCK}, // ASTC_2D_12X10_UNORM + {VK_FORMAT_ASTC_12x10_SRGB_BLOCK}, // ASTC_2D_12X10_SRGB {VK_FORMAT_ASTC_12x12_UNORM_BLOCK}, // ASTC_2D_12X12_UNORM {VK_FORMAT_ASTC_12x12_SRGB_BLOCK}, // ASTC_2D_12X12_SRGB {VK_FORMAT_ASTC_8x6_UNORM_BLOCK}, // ASTC_2D_8X6_UNORM diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.cpp b/src/video_core/renderer_vulkan/renderer_vulkan.cpp index 94ae2a2f8..8e31eba34 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.cpp +++ b/src/video_core/renderer_vulkan/renderer_vulkan.cpp @@ -134,7 +134,7 @@ void RendererVulkan::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { Frame* frame = present_manager.GetRenderFrame(); blit_screen.DrawToSwapchain(frame, *framebuffer, use_accelerated, is_srgb); scheduler.Flush(*frame->render_ready); - scheduler.Record([this, frame](vk::CommandBuffer) { present_manager.PushFrame(frame); }); + present_manager.Present(frame); gpu.RendererFrameEndNotify(); rasterizer.TickFrame(); diff --git a/src/video_core/renderer_vulkan/vk_fence_manager.cpp b/src/video_core/renderer_vulkan/vk_fence_manager.cpp index 0214b103a..fad9e3832 100644 --- a/src/video_core/renderer_vulkan/vk_fence_manager.cpp +++ b/src/video_core/renderer_vulkan/vk_fence_manager.cpp @@ -5,6 +5,7 @@ #include "video_core/renderer_vulkan/vk_buffer_cache.h" #include "video_core/renderer_vulkan/vk_fence_manager.h" +#include "video_core/renderer_vulkan/vk_query_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_texture_cache.h" #include "video_core/vulkan_common/vulkan_device.h" diff --git a/src/video_core/renderer_vulkan/vk_fence_manager.h b/src/video_core/renderer_vulkan/vk_fence_manager.h index 7fe2afcd9..145359d4e 100644 --- a/src/video_core/renderer_vulkan/vk_fence_manager.h +++ b/src/video_core/renderer_vulkan/vk_fence_manager.h @@ -40,7 +40,16 @@ private: }; using Fence = std::shared_ptr<InnerFence>; -using GenericFenceManager = VideoCommon::FenceManager<Fence, TextureCache, BufferCache, QueryCache>; +struct FenceManagerParams { + using FenceType = Fence; + using BufferCacheType = BufferCache; + using TextureCacheType = TextureCache; + using QueryCacheType = QueryCache; + + static constexpr bool HAS_ASYNC_CHECK = true; +}; + +using GenericFenceManager = VideoCommon::FenceManager<FenceManagerParams>; class FenceManager final : public GenericFenceManager { public: diff --git a/src/video_core/renderer_vulkan/vk_present_manager.cpp b/src/video_core/renderer_vulkan/vk_present_manager.cpp index a137c66f2..c49583013 100644 --- a/src/video_core/renderer_vulkan/vk_present_manager.cpp +++ b/src/video_core/renderer_vulkan/vk_present_manager.cpp @@ -153,16 +153,19 @@ Frame* PresentManager::GetRenderFrame() { return frame; } -void PresentManager::PushFrame(Frame* frame) { +void PresentManager::Present(Frame* frame) { if (!use_present_thread) { + scheduler.WaitWorker(); CopyToSwapchain(frame); free_queue.push(frame); return; } - std::unique_lock lock{queue_mutex}; - present_queue.push(frame); - frame_cv.notify_one(); + scheduler.Record([this, frame](vk::CommandBuffer) { + std::unique_lock lock{queue_mutex}; + present_queue.push(frame); + frame_cv.notify_one(); + }); } void PresentManager::RecreateFrame(Frame* frame, u32 width, u32 height, bool is_srgb, diff --git a/src/video_core/renderer_vulkan/vk_present_manager.h b/src/video_core/renderer_vulkan/vk_present_manager.h index 9885fd7c6..420a775e2 100644 --- a/src/video_core/renderer_vulkan/vk_present_manager.h +++ b/src/video_core/renderer_vulkan/vk_present_manager.h @@ -45,7 +45,7 @@ public: Frame* GetRenderFrame(); /// Pushes a frame for presentation - void PushFrame(Frame* frame); + void Present(Frame* frame); /// Recreates the present frame to match the provided parameters void RecreateFrame(Frame* frame, u32 width, u32 height, bool is_srgb, diff --git a/src/video_core/renderer_vulkan/vk_query_cache.cpp b/src/video_core/renderer_vulkan/vk_query_cache.cpp index 929c8ece6..d67490449 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_query_cache.cpp @@ -66,9 +66,10 @@ void QueryPool::Reserve(std::pair<VkQueryPool, u32> query) { } } -QueryCache::QueryCache(VideoCore::RasterizerInterface& rasterizer_, const Device& device_, +QueryCache::QueryCache(VideoCore::RasterizerInterface& rasterizer_, + Core::Memory::Memory& cpu_memory_, const Device& device_, Scheduler& scheduler_) - : QueryCacheBase{rasterizer_}, device{device_}, scheduler{scheduler_}, + : QueryCacheBase{rasterizer_, cpu_memory_}, device{device_}, scheduler{scheduler_}, query_pools{ QueryPool{device_, scheduler_, QueryType::SamplesPassed}, } {} @@ -98,8 +99,10 @@ HostCounter::HostCounter(QueryCache& cache_, std::shared_ptr<HostCounter> depend query{cache_.AllocateQuery(type_)}, tick{cache_.GetScheduler().CurrentTick()} { const vk::Device* logical = &cache.GetDevice().GetLogical(); cache.GetScheduler().Record([logical, query = query](vk::CommandBuffer cmdbuf) { + const bool use_precise = Settings::IsGPULevelHigh(); logical->ResetQueryPool(query.first, query.second, 1); - cmdbuf.BeginQuery(query.first, query.second, VK_QUERY_CONTROL_PRECISE_BIT); + cmdbuf.BeginQuery(query.first, query.second, + use_precise ? VK_QUERY_CONTROL_PRECISE_BIT : 0); }); } @@ -112,8 +115,10 @@ void HostCounter::EndQuery() { [query = query](vk::CommandBuffer cmdbuf) { cmdbuf.EndQuery(query.first, query.second); }); } -u64 HostCounter::BlockingQuery() const { - cache.GetScheduler().Wait(tick); +u64 HostCounter::BlockingQuery(bool async) const { + if (!async) { + cache.GetScheduler().Wait(tick); + } u64 data; const VkResult query_result = cache.GetDevice().GetLogical().GetQueryResults( query.first, query.second, 1, sizeof(data), &data, sizeof(data), diff --git a/src/video_core/renderer_vulkan/vk_query_cache.h b/src/video_core/renderer_vulkan/vk_query_cache.h index 26762ee09..c1b9552eb 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.h +++ b/src/video_core/renderer_vulkan/vk_query_cache.h @@ -52,7 +52,8 @@ private: class QueryCache final : public VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, HostCounter> { public: - explicit QueryCache(VideoCore::RasterizerInterface& rasterizer_, const Device& device_, + explicit QueryCache(VideoCore::RasterizerInterface& rasterizer_, + Core::Memory::Memory& cpu_memory_, const Device& device_, Scheduler& scheduler_); ~QueryCache(); @@ -83,7 +84,7 @@ public: void EndQuery(); private: - u64 BlockingQuery() const override; + u64 BlockingQuery(bool async = false) const override; QueryCache& cache; const VideoCore::QueryType type; diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index 2559a3aa7..d1489fc95 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -172,7 +172,8 @@ RasterizerVulkan::RasterizerVulkan(Core::Frontend::EmuWindow& emu_window_, Tegra buffer_cache(*this, cpu_memory_, buffer_cache_runtime), pipeline_cache(*this, device, scheduler, descriptor_pool, update_descriptor_queue, render_pass_cache, buffer_cache, texture_cache, gpu.ShaderNotify()), - query_cache{*this, device, scheduler}, accelerate_dma(buffer_cache, texture_cache, scheduler), + query_cache{*this, cpu_memory_, device, scheduler}, + accelerate_dma(buffer_cache, texture_cache, scheduler), fence_manager(*this, gpu, texture_cache, buffer_cache, query_cache, device, scheduler), wfi_event(device.GetLogical().CreateEvent()) { scheduler.SetQueryCache(query_cache); @@ -675,7 +676,8 @@ bool RasterizerVulkan::AccelerateConditionalRendering() { const GPUVAddr condition_address{maxwell3d->regs.render_enable.Address()}; Maxwell::ReportSemaphore::Compare cmp; if (gpu_memory->IsMemoryDirty(condition_address, sizeof(cmp), - VideoCommon::CacheType::BufferCache)) { + VideoCommon::CacheType::BufferCache | + VideoCommon::CacheType::QueryCache)) { return true; } return false; diff --git a/src/video_core/shader_cache.cpp b/src/video_core/shader_cache.cpp index d9482371b..c5213875b 100644 --- a/src/video_core/shader_cache.cpp +++ b/src/video_core/shader_cache.cpp @@ -228,14 +228,14 @@ const ShaderInfo* ShaderCache::MakeShaderInfo(GenericEnvironment& env, VAddr cpu auto info = std::make_unique<ShaderInfo>(); if (const std::optional<u64> cached_hash{env.Analyze()}) { info->unique_hash = *cached_hash; - info->size_bytes = env.CachedSize(); + info->size_bytes = env.CachedSizeBytes(); } else { // Slow path, not really hit on commercial games // Build a control flow graph to get the real shader size Shader::ObjectPool<Shader::Maxwell::Flow::Block> flow_block; Shader::Maxwell::Flow::CFG cfg{env, flow_block, env.StartAddress()}; info->unique_hash = env.CalculateHash(); - info->size_bytes = env.ReadSize(); + info->size_bytes = env.ReadSizeBytes(); } const size_t size_bytes{info->size_bytes}; const ShaderInfo* const result{info.get()}; diff --git a/src/video_core/shader_environment.cpp b/src/video_core/shader_environment.cpp index 574760f80..c7cb56243 100644 --- a/src/video_core/shader_environment.cpp +++ b/src/video_core/shader_environment.cpp @@ -170,15 +170,19 @@ std::optional<u64> GenericEnvironment::Analyze() { void GenericEnvironment::SetCachedSize(size_t size_bytes) { cached_lowest = start_address; cached_highest = start_address + static_cast<u32>(size_bytes); - code.resize(CachedSize()); + code.resize(CachedSizeWords()); gpu_memory->ReadBlock(program_base + cached_lowest, code.data(), code.size() * sizeof(u64)); } -size_t GenericEnvironment::CachedSize() const noexcept { - return cached_highest - cached_lowest + INST_SIZE; +size_t GenericEnvironment::CachedSizeWords() const noexcept { + return CachedSizeBytes() / INST_SIZE; } -size_t GenericEnvironment::ReadSize() const noexcept { +size_t GenericEnvironment::CachedSizeBytes() const noexcept { + return static_cast<size_t>(cached_highest) - cached_lowest + INST_SIZE; +} + +size_t GenericEnvironment::ReadSizeBytes() const noexcept { return read_highest - read_lowest + INST_SIZE; } @@ -187,7 +191,7 @@ bool GenericEnvironment::CanBeSerialized() const noexcept { } u64 GenericEnvironment::CalculateHash() const { - const size_t size{ReadSize()}; + const size_t size{ReadSizeBytes()}; const auto data{std::make_unique<char[]>(size)}; gpu_memory->ReadBlock(program_base + read_lowest, data.get(), size); return Common::CityHash64(data.get(), size); @@ -198,7 +202,7 @@ void GenericEnvironment::Dump(u64 hash) { } void GenericEnvironment::Serialize(std::ofstream& file) const { - const u64 code_size{static_cast<u64>(CachedSize())}; + const u64 code_size{static_cast<u64>(CachedSizeBytes())}; const u64 num_texture_types{static_cast<u64>(texture_types.size())}; const u64 num_texture_pixel_formats{static_cast<u64>(texture_pixel_formats.size())}; const u64 num_cbuf_values{static_cast<u64>(cbuf_values.size())}; diff --git a/src/video_core/shader_environment.h b/src/video_core/shader_environment.h index d75987a52..a0f61cbda 100644 --- a/src/video_core/shader_environment.h +++ b/src/video_core/shader_environment.h @@ -48,9 +48,11 @@ public: void SetCachedSize(size_t size_bytes); - [[nodiscard]] size_t CachedSize() const noexcept; + [[nodiscard]] size_t CachedSizeWords() const noexcept; - [[nodiscard]] size_t ReadSize() const noexcept; + [[nodiscard]] size_t CachedSizeBytes() const noexcept; + + [[nodiscard]] size_t ReadSizeBytes() const noexcept; [[nodiscard]] bool CanBeSerialized() const noexcept; diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp index 1a76d4178..cb51529e4 100644 --- a/src/video_core/surface.cpp +++ b/src/video_core/surface.cpp @@ -250,10 +250,13 @@ bool IsPixelFormatASTC(PixelFormat format) { case PixelFormat::ASTC_2D_6X6_UNORM: case PixelFormat::ASTC_2D_6X6_SRGB: case PixelFormat::ASTC_2D_10X6_UNORM: + case PixelFormat::ASTC_2D_10X6_SRGB: case PixelFormat::ASTC_2D_10X5_UNORM: case PixelFormat::ASTC_2D_10X5_SRGB: case PixelFormat::ASTC_2D_10X10_UNORM: case PixelFormat::ASTC_2D_10X10_SRGB: + case PixelFormat::ASTC_2D_12X10_UNORM: + case PixelFormat::ASTC_2D_12X10_SRGB: case PixelFormat::ASTC_2D_12X12_UNORM: case PixelFormat::ASTC_2D_12X12_SRGB: case PixelFormat::ASTC_2D_8X6_UNORM: @@ -279,11 +282,13 @@ bool IsPixelFormatSRGB(PixelFormat format) { case PixelFormat::ASTC_2D_8X5_SRGB: case PixelFormat::ASTC_2D_5X4_SRGB: case PixelFormat::ASTC_2D_5X5_SRGB: + case PixelFormat::ASTC_2D_10X6_SRGB: case PixelFormat::ASTC_2D_10X8_SRGB: case PixelFormat::ASTC_2D_6X6_SRGB: case PixelFormat::ASTC_2D_10X5_SRGB: case PixelFormat::ASTC_2D_10X10_SRGB: case PixelFormat::ASTC_2D_12X12_SRGB: + case PixelFormat::ASTC_2D_12X10_SRGB: case PixelFormat::ASTC_2D_8X6_SRGB: case PixelFormat::ASTC_2D_6X5_SRGB: return true; diff --git a/src/video_core/surface.h b/src/video_core/surface.h index 44b79af20..0225d3287 100644 --- a/src/video_core/surface.h +++ b/src/video_core/surface.h @@ -95,10 +95,13 @@ enum class PixelFormat { ASTC_2D_6X6_UNORM, ASTC_2D_6X6_SRGB, ASTC_2D_10X6_UNORM, + ASTC_2D_10X6_SRGB, ASTC_2D_10X5_UNORM, ASTC_2D_10X5_SRGB, ASTC_2D_10X10_UNORM, ASTC_2D_10X10_SRGB, + ASTC_2D_12X10_UNORM, + ASTC_2D_12X10_SRGB, ASTC_2D_12X12_UNORM, ASTC_2D_12X12_SRGB, ASTC_2D_8X6_UNORM, @@ -232,10 +235,13 @@ constexpr std::array<u8, MaxPixelFormat> BLOCK_WIDTH_TABLE = {{ 6, // ASTC_2D_6X6_UNORM 6, // ASTC_2D_6X6_SRGB 10, // ASTC_2D_10X6_UNORM + 10, // ASTC_2D_10X6_SRGB 10, // ASTC_2D_10X5_UNORM 10, // ASTC_2D_10X5_SRGB 10, // ASTC_2D_10X10_UNORM 10, // ASTC_2D_10X10_SRGB + 12, // ASTC_2D_12X10_UNORM + 12, // ASTC_2D_12X10_SRGB 12, // ASTC_2D_12X12_UNORM 12, // ASTC_2D_12X12_SRGB 8, // ASTC_2D_8X6_UNORM @@ -338,10 +344,13 @@ constexpr std::array<u8, MaxPixelFormat> BLOCK_HEIGHT_TABLE = {{ 6, // ASTC_2D_6X6_UNORM 6, // ASTC_2D_6X6_SRGB 6, // ASTC_2D_10X6_UNORM + 6, // ASTC_2D_10X6_SRGB 5, // ASTC_2D_10X5_UNORM 5, // ASTC_2D_10X5_SRGB 10, // ASTC_2D_10X10_UNORM 10, // ASTC_2D_10X10_SRGB + 10, // ASTC_2D_12X10_UNORM + 10, // ASTC_2D_12X10_SRGB 12, // ASTC_2D_12X12_UNORM 12, // ASTC_2D_12X12_SRGB 6, // ASTC_2D_8X6_UNORM @@ -444,10 +453,13 @@ constexpr std::array<u8, MaxPixelFormat> BITS_PER_BLOCK_TABLE = {{ 128, // ASTC_2D_6X6_UNORM 128, // ASTC_2D_6X6_SRGB 128, // ASTC_2D_10X6_UNORM + 128, // ASTC_2D_10X6_SRGB 128, // ASTC_2D_10X5_UNORM 128, // ASTC_2D_10X5_SRGB 128, // ASTC_2D_10X10_UNORM 128, // ASTC_2D_10X10_SRGB + 128, // ASTC_2D_12X10_UNORM + 128, // ASTC_2D_12X10_SRGB 128, // ASTC_2D_12X12_UNORM 128, // ASTC_2D_12X12_SRGB 128, // ASTC_2D_8X6_UNORM diff --git a/src/video_core/texture_cache/format_lookup_table.cpp b/src/video_core/texture_cache/format_lookup_table.cpp index 5fc2b2fec..11ced6c38 100644 --- a/src/video_core/texture_cache/format_lookup_table.cpp +++ b/src/video_core/texture_cache/format_lookup_table.cpp @@ -210,6 +210,8 @@ PixelFormat PixelFormatFromTextureInfo(TextureFormat format, ComponentType red, return PixelFormat::ASTC_2D_6X6_SRGB; case Hash(TextureFormat::ASTC_2D_10X6, UNORM, LINEAR): return PixelFormat::ASTC_2D_10X6_UNORM; + case Hash(TextureFormat::ASTC_2D_10X6, UNORM, SRGB): + return PixelFormat::ASTC_2D_10X6_SRGB; case Hash(TextureFormat::ASTC_2D_10X5, UNORM, LINEAR): return PixelFormat::ASTC_2D_10X5_UNORM; case Hash(TextureFormat::ASTC_2D_10X5, UNORM, SRGB): @@ -218,6 +220,10 @@ PixelFormat PixelFormatFromTextureInfo(TextureFormat format, ComponentType red, return PixelFormat::ASTC_2D_10X10_UNORM; case Hash(TextureFormat::ASTC_2D_10X10, UNORM, SRGB): return PixelFormat::ASTC_2D_10X10_SRGB; + case Hash(TextureFormat::ASTC_2D_12X10, UNORM, LINEAR): + return PixelFormat::ASTC_2D_12X10_UNORM; + case Hash(TextureFormat::ASTC_2D_12X10, UNORM, SRGB): + return PixelFormat::ASTC_2D_12X10_SRGB; case Hash(TextureFormat::ASTC_2D_12X12, UNORM, LINEAR): return PixelFormat::ASTC_2D_12X12_UNORM; case Hash(TextureFormat::ASTC_2D_12X12, UNORM, SRGB): diff --git a/src/video_core/texture_cache/formatter.h b/src/video_core/texture_cache/formatter.h index f1f0a057b..b97147797 100644 --- a/src/video_core/texture_cache/formatter.h +++ b/src/video_core/texture_cache/formatter.h @@ -179,6 +179,8 @@ struct fmt::formatter<VideoCore::Surface::PixelFormat> : fmt::formatter<fmt::str return "ASTC_2D_6X6_SRGB"; case PixelFormat::ASTC_2D_10X6_UNORM: return "ASTC_2D_10X6_UNORM"; + case PixelFormat::ASTC_2D_10X6_SRGB: + return "ASTC_2D_10X6_SRGB"; case PixelFormat::ASTC_2D_10X5_UNORM: return "ASTC_2D_10X5_UNORM"; case PixelFormat::ASTC_2D_10X5_SRGB: @@ -187,6 +189,10 @@ struct fmt::formatter<VideoCore::Surface::PixelFormat> : fmt::formatter<fmt::str return "ASTC_2D_10X10_UNORM"; case PixelFormat::ASTC_2D_10X10_SRGB: return "ASTC_2D_10X10_SRGB"; + case PixelFormat::ASTC_2D_12X10_UNORM: + return "ASTC_2D_12X10_UNORM"; + case PixelFormat::ASTC_2D_12X10_SRGB: + return "ASTC_2D_12X10_SRGB"; case PixelFormat::ASTC_2D_12X12_UNORM: return "ASTC_2D_12X12_UNORM"; case PixelFormat::ASTC_2D_12X12_SRGB: diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index e601f8446..f335009d0 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -888,7 +888,7 @@ void TextureCache<P>::DownloadImageIntoBuffer(typename TextureCache<P>::Image* i buffer, download_map.buffer, }; - std::array buffer_offsets{ + std::array<u64, 2> buffer_offsets{ buffer_offset, download_map.offset, }; diff --git a/src/video_core/vulkan_common/vulkan_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp index 6f288b3f8..6ffca2af2 100644 --- a/src/video_core/vulkan_common/vulkan_device.cpp +++ b/src/video_core/vulkan_common/vulkan_device.cpp @@ -617,7 +617,9 @@ bool Device::ShouldBoostClocks() const { const bool is_steam_deck = vendor_id == 0x1002 && device_id == 0x163F; - return validated_driver && !is_steam_deck; + const bool is_debugging = this->HasDebuggingToolAttached(); + + return validated_driver && !is_steam_deck && !is_debugging; } bool Device::GetSuitability(bool requires_swapchain) { diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp index 4a8436e5c..a85eb4687 100644 --- a/src/yuzu/configuration/config.cpp +++ b/src/yuzu/configuration/config.cpp @@ -498,7 +498,7 @@ void Config::ReadCoreValues() { qt_config->beginGroup(QStringLiteral("Core")); ReadGlobalSetting(Settings::values.use_multi_core); - ReadGlobalSetting(Settings::values.use_extended_memory_layout); + ReadGlobalSetting(Settings::values.use_unsafe_extended_memory_layout); qt_config->endGroup(); } @@ -713,7 +713,6 @@ void Config::ReadRendererValues() { ReadGlobalSetting(Settings::values.shader_backend); ReadGlobalSetting(Settings::values.use_asynchronous_shaders); ReadGlobalSetting(Settings::values.use_fast_gpu_time); - ReadGlobalSetting(Settings::values.use_pessimistic_flushes); ReadGlobalSetting(Settings::values.use_vulkan_driver_pipeline_cache); ReadGlobalSetting(Settings::values.bg_red); ReadGlobalSetting(Settings::values.bg_green); @@ -1166,7 +1165,7 @@ void Config::SaveCoreValues() { qt_config->beginGroup(QStringLiteral("Core")); WriteGlobalSetting(Settings::values.use_multi_core); - WriteGlobalSetting(Settings::values.use_extended_memory_layout); + WriteGlobalSetting(Settings::values.use_unsafe_extended_memory_layout); qt_config->endGroup(); } @@ -1362,7 +1361,6 @@ void Config::SaveRendererValues() { Settings::values.shader_backend.UsingGlobal()); WriteGlobalSetting(Settings::values.use_asynchronous_shaders); WriteGlobalSetting(Settings::values.use_fast_gpu_time); - WriteGlobalSetting(Settings::values.use_pessimistic_flushes); WriteGlobalSetting(Settings::values.use_vulkan_driver_pipeline_cache); WriteGlobalSetting(Settings::values.bg_red); WriteGlobalSetting(Settings::values.bg_green); diff --git a/src/yuzu/configuration/configure_general.cpp b/src/yuzu/configuration/configure_general.cpp index 207bcdc4d..26258d744 100644 --- a/src/yuzu/configuration/configure_general.cpp +++ b/src/yuzu/configuration/configure_general.cpp @@ -35,9 +35,6 @@ void ConfigureGeneral::SetConfiguration() { ui->use_multi_core->setEnabled(runtime_lock); ui->use_multi_core->setChecked(Settings::values.use_multi_core.GetValue()); - ui->use_extended_memory_layout->setEnabled(runtime_lock); - ui->use_extended_memory_layout->setChecked( - Settings::values.use_extended_memory_layout.GetValue()); ui->toggle_check_exit->setChecked(UISettings::values.confirm_before_closing.GetValue()); ui->toggle_user_on_boot->setChecked(UISettings::values.select_user_on_boot.GetValue()); @@ -79,9 +76,6 @@ void ConfigureGeneral::ResetDefaults() { void ConfigureGeneral::ApplyConfiguration() { ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_multi_core, ui->use_multi_core, use_multi_core); - ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_extended_memory_layout, - ui->use_extended_memory_layout, - use_extended_memory_layout); if (Settings::IsConfiguringGlobal()) { UISettings::values.confirm_before_closing = ui->toggle_check_exit->isChecked(); @@ -141,9 +135,6 @@ void ConfigureGeneral::SetupPerGameUI() { Settings::values.use_speed_limit, use_speed_limit); ConfigurationShared::SetColoredTristate(ui->use_multi_core, Settings::values.use_multi_core, use_multi_core); - ConfigurationShared::SetColoredTristate(ui->use_extended_memory_layout, - Settings::values.use_extended_memory_layout, - use_extended_memory_layout); connect(ui->toggle_speed_limit, &QCheckBox::clicked, ui->speed_limit, [this]() { ui->speed_limit->setEnabled(ui->toggle_speed_limit->isChecked() && diff --git a/src/yuzu/configuration/configure_general.h b/src/yuzu/configuration/configure_general.h index a090c1a3f..7ff63f425 100644 --- a/src/yuzu/configuration/configure_general.h +++ b/src/yuzu/configuration/configure_general.h @@ -47,7 +47,6 @@ private: ConfigurationShared::CheckState use_speed_limit; ConfigurationShared::CheckState use_multi_core; - ConfigurationShared::CheckState use_extended_memory_layout; const Core::System& system; }; diff --git a/src/yuzu/configuration/configure_general.ui b/src/yuzu/configuration/configure_general.ui index add110bb0..986a1625b 100644 --- a/src/yuzu/configuration/configure_general.ui +++ b/src/yuzu/configuration/configure_general.ui @@ -62,13 +62,6 @@ </widget> </item> <item> - <widget class="QCheckBox" name="use_extended_memory_layout"> - <property name="text"> - <string>Extended memory layout (8GB DRAM)</string> - </property> - </widget> - </item> - <item> <widget class="QCheckBox" name="toggle_check_exit"> <property name="text"> <string>Confirm exit while emulation is running</string> diff --git a/src/yuzu/configuration/configure_graphics_advanced.cpp b/src/yuzu/configuration/configure_graphics_advanced.cpp index 4072ce145..005b022ca 100644 --- a/src/yuzu/configuration/configure_graphics_advanced.cpp +++ b/src/yuzu/configuration/configure_graphics_advanced.cpp @@ -32,7 +32,6 @@ void ConfigureGraphicsAdvanced::SetConfiguration() { ui->async_astc->setChecked(Settings::values.async_astc.GetValue()); ui->use_asynchronous_shaders->setChecked(Settings::values.use_asynchronous_shaders.GetValue()); ui->use_fast_gpu_time->setChecked(Settings::values.use_fast_gpu_time.GetValue()); - ui->use_pessimistic_flushes->setChecked(Settings::values.use_pessimistic_flushes.GetValue()); ui->use_vulkan_driver_pipeline_cache->setChecked( Settings::values.use_vulkan_driver_pipeline_cache.GetValue()); @@ -68,8 +67,6 @@ void ConfigureGraphicsAdvanced::ApplyConfiguration() { use_asynchronous_shaders); ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_fast_gpu_time, ui->use_fast_gpu_time, use_fast_gpu_time); - ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_pessimistic_flushes, - ui->use_pessimistic_flushes, use_pessimistic_flushes); ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_vulkan_driver_pipeline_cache, ui->use_vulkan_driver_pipeline_cache, use_vulkan_driver_pipeline_cache); @@ -98,8 +95,6 @@ void ConfigureGraphicsAdvanced::SetupPerGameUI() { ui->use_asynchronous_shaders->setEnabled( Settings::values.use_asynchronous_shaders.UsingGlobal()); ui->use_fast_gpu_time->setEnabled(Settings::values.use_fast_gpu_time.UsingGlobal()); - ui->use_pessimistic_flushes->setEnabled( - Settings::values.use_pessimistic_flushes.UsingGlobal()); ui->use_vulkan_driver_pipeline_cache->setEnabled( Settings::values.use_vulkan_driver_pipeline_cache.UsingGlobal()); ui->anisotropic_filtering_combobox->setEnabled( @@ -120,9 +115,6 @@ void ConfigureGraphicsAdvanced::SetupPerGameUI() { use_asynchronous_shaders); ConfigurationShared::SetColoredTristate(ui->use_fast_gpu_time, Settings::values.use_fast_gpu_time, use_fast_gpu_time); - ConfigurationShared::SetColoredTristate(ui->use_pessimistic_flushes, - Settings::values.use_pessimistic_flushes, - use_pessimistic_flushes); ConfigurationShared::SetColoredTristate(ui->use_vulkan_driver_pipeline_cache, Settings::values.use_vulkan_driver_pipeline_cache, use_vulkan_driver_pipeline_cache); diff --git a/src/yuzu/configuration/configure_graphics_advanced.h b/src/yuzu/configuration/configure_graphics_advanced.h index 5394ed40a..ff5060957 100644 --- a/src/yuzu/configuration/configure_graphics_advanced.h +++ b/src/yuzu/configuration/configure_graphics_advanced.h @@ -42,7 +42,6 @@ private: ConfigurationShared::CheckState async_astc; ConfigurationShared::CheckState use_asynchronous_shaders; ConfigurationShared::CheckState use_fast_gpu_time; - ConfigurationShared::CheckState use_pessimistic_flushes; ConfigurationShared::CheckState use_vulkan_driver_pipeline_cache; const Core::System& system; diff --git a/src/yuzu/configuration/configure_graphics_advanced.ui b/src/yuzu/configuration/configure_graphics_advanced.ui index 134023032..d073fe9b1 100644 --- a/src/yuzu/configuration/configure_graphics_advanced.ui +++ b/src/yuzu/configuration/configure_graphics_advanced.ui @@ -117,16 +117,6 @@ </widget> </item> <item> - <widget class="QCheckBox" name="use_pessimistic_flushes"> - <property name="toolTip"> - <string>Enables pessimistic buffer flushes. This option will force unmodified buffers to be flushed, which can cost performance.</string> - </property> - <property name="text"> - <string>Use pessimistic buffer flushes (Hack)</string> - </property> - </widget> - </item> - <item> <widget class="QCheckBox" name="use_vulkan_driver_pipeline_cache"> <property name="toolTip"> <string>Enables GPU vendor-specific pipeline cache. This option can improve shader loading time significantly in cases where the Vulkan driver does not store pipeline cache files internally.</string> diff --git a/src/yuzu/configuration/configure_input_player.cpp b/src/yuzu/configuration/configure_input_player.cpp index a21a3eaa9..561a08dc5 100644 --- a/src/yuzu/configuration/configure_input_player.cpp +++ b/src/yuzu/configuration/configure_input_player.cpp @@ -207,7 +207,7 @@ QString ConfigureInputPlayer::ButtonToText(const Common::ParamPackage& param) { } if (param.Has("axis")) { const QString axis = QString::fromStdString(param.Get("axis", "")); - return QObject::tr("%1%2Axis %3").arg(toggle, invert, axis); + return QObject::tr("%1%2%3Axis %4").arg(toggle, inverted, invert, axis); } if (param.Has("axis_x") && param.Has("axis_y") && param.Has("axis_z")) { const QString axis_x = QString::fromStdString(param.Get("axis_x", "")); @@ -230,7 +230,7 @@ QString ConfigureInputPlayer::ButtonToText(const Common::ParamPackage& param) { return QObject::tr("%1%2%3Hat %4").arg(turbo, toggle, inverted, button_name); } if (param.Has("axis")) { - return QObject::tr("%1%2Axis %3").arg(toggle, inverted, button_name); + return QObject::tr("%1%2%3Axis %4").arg(toggle, inverted, invert, button_name); } if (param.Has("motion")) { return QObject::tr("%1%2Axis %3").arg(toggle, inverted, button_name); @@ -411,6 +411,12 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i button_map[button_id]->setText(ButtonToText(param)); emulated_controller->SetButtonParam(button_id, param); }); + context_menu.addAction(tr("Invert button"), [&] { + const bool invert_value = !param.Get("inverted", false); + param.Set("inverted", invert_value); + button_map[button_id]->setText(ButtonToText(param)); + emulated_controller->SetButtonParam(button_id, param); + }); context_menu.addAction(tr("Set threshold"), [&] { const int button_threshold = static_cast<int>(param.Get("threshold", 0.5f) * 100.0f); diff --git a/src/yuzu/configuration/configure_input_player_widget.cpp b/src/yuzu/configuration/configure_input_player_widget.cpp index c287220fc..fe1ee2289 100644 --- a/src/yuzu/configuration/configure_input_player_widget.cpp +++ b/src/yuzu/configuration/configure_input_player_widget.cpp @@ -180,6 +180,10 @@ void PlayerControlPreview::ControllerUpdate(Core::HID::ControllerTriggerType typ battery_values = controller->GetBatteryValues(); needs_redraw = true; break; + case Core::HID::ControllerTriggerType::Motion: + motion_values = controller->GetMotions(); + needs_redraw = true; + break; default: break; } @@ -313,6 +317,15 @@ void PlayerControlPreview::DrawLeftController(QPainter& p, const QPointF center) DrawRawJoystick(p, center + QPointF(-140, 90), QPointF(0, 0)); } + { + // Draw motion cubes + using namespace Settings::NativeMotion; + p.setPen(colors.outline); + p.setBrush(colors.transparent); + Draw3dCube(p, center + QPointF(-140, 90), + motion_values[Settings::NativeMotion::MotionLeft].euler, 20.0f); + } + using namespace Settings::NativeButton; // D-pad constants @@ -435,6 +448,15 @@ void PlayerControlPreview::DrawRightController(QPainter& p, const QPointF center DrawRawJoystick(p, QPointF(0, 0), center + QPointF(140, 90)); } + { + // Draw motion cubes + using namespace Settings::NativeMotion; + p.setPen(colors.outline); + p.setBrush(colors.transparent); + Draw3dCube(p, center + QPointF(140, 90), + motion_values[Settings::NativeMotion::MotionRight].euler, 20.0f); + } + using namespace Settings::NativeButton; // Face buttons constants @@ -555,6 +577,17 @@ void PlayerControlPreview::DrawDualController(QPainter& p, const QPointF center) DrawRawJoystick(p, center + QPointF(-180, 90), center + QPointF(180, 90)); } + { + // Draw motion cubes + using namespace Settings::NativeMotion; + p.setPen(colors.outline); + p.setBrush(colors.transparent); + Draw3dCube(p, center + QPointF(-180, -5), + motion_values[Settings::NativeMotion::MotionLeft].euler, 20.0f); + Draw3dCube(p, center + QPointF(180, -5), + motion_values[Settings::NativeMotion::MotionRight].euler, 20.0f); + } + using namespace Settings::NativeButton; // Face buttons constants @@ -647,6 +680,15 @@ void PlayerControlPreview::DrawHandheldController(QPainter& p, const QPointF cen DrawRawJoystick(p, center + QPointF(-50, 0), center + QPointF(50, 0)); } + { + // Draw motion cubes + using namespace Settings::NativeMotion; + p.setPen(colors.outline); + p.setBrush(colors.transparent); + Draw3dCube(p, center + QPointF(0, -115), + motion_values[Settings::NativeMotion::MotionLeft].euler, 15.0f); + } + using namespace Settings::NativeButton; // Face buttons constants @@ -750,6 +792,15 @@ void PlayerControlPreview::DrawProController(QPainter& p, const QPointF center) DrawRawJoystick(p, center + QPointF(-50, 105), center + QPointF(50, 105)); } + { + // Draw motion cubes + using namespace Settings::NativeMotion; + p.setPen(colors.button); + p.setBrush(colors.transparent); + Draw3dCube(p, center + QPointF(0, -100), + motion_values[Settings::NativeMotion::MotionLeft].euler, 15.0f); + } + using namespace Settings::NativeButton; // Face buttons constants @@ -2871,6 +2922,46 @@ void PlayerControlPreview::DrawArrow(QPainter& p, const QPointF center, const Di DrawPolygon(p, arrow_symbol); } +// Draw motion functions +void PlayerControlPreview::Draw3dCube(QPainter& p, QPointF center, const Common::Vec3f& euler, + float size) { + std::array<Common::Vec3f, 8> cube{ + Common::Vec3f{-1, -1, -1}, + {-1, 1, -1}, + {1, 1, -1}, + {1, -1, -1}, + {-1, -1, 1}, + {-1, 1, 1}, + {1, 1, 1}, + {1, -1, 1}, + }; + + for (Common::Vec3f& point : cube) { + point.RotateFromOrigin(euler.x, euler.y, euler.z); + point *= size; + } + + const std::array<QPointF, 4> front_face{ + center + QPointF{cube[0].x, cube[0].y}, + center + QPointF{cube[1].x, cube[1].y}, + center + QPointF{cube[2].x, cube[2].y}, + center + QPointF{cube[3].x, cube[3].y}, + }; + const std::array<QPointF, 4> back_face{ + center + QPointF{cube[4].x, cube[4].y}, + center + QPointF{cube[5].x, cube[5].y}, + center + QPointF{cube[6].x, cube[6].y}, + center + QPointF{cube[7].x, cube[7].y}, + }; + + DrawPolygon(p, front_face); + DrawPolygon(p, back_face); + p.drawLine(center + QPointF{cube[0].x, cube[0].y}, center + QPointF{cube[4].x, cube[4].y}); + p.drawLine(center + QPointF{cube[1].x, cube[1].y}, center + QPointF{cube[5].x, cube[5].y}); + p.drawLine(center + QPointF{cube[2].x, cube[2].y}, center + QPointF{cube[6].x, cube[6].y}); + p.drawLine(center + QPointF{cube[3].x, cube[3].y}, center + QPointF{cube[7].x, cube[7].y}); +} + template <size_t N> void PlayerControlPreview::DrawPolygon(QPainter& p, const std::array<QPointF, N>& polygon) { p.drawPolygon(polygon.data(), static_cast<int>(polygon.size())); diff --git a/src/yuzu/configuration/configure_input_player_widget.h b/src/yuzu/configuration/configure_input_player_widget.h index 267d134de..a16943c3c 100644 --- a/src/yuzu/configuration/configure_input_player_widget.h +++ b/src/yuzu/configuration/configure_input_player_widget.h @@ -9,6 +9,7 @@ #include "common/input.h" #include "common/settings_input.h" +#include "common/vector_math.h" #include "core/hid/emulated_controller.h" #include "core/hid/hid_types.h" @@ -193,6 +194,9 @@ private: void DrawSymbol(QPainter& p, QPointF center, Symbol symbol, float icon_size); void DrawArrow(QPainter& p, QPointF center, Direction direction, float size); + // Draw motion functions + void Draw3dCube(QPainter& p, QPointF center, const Common::Vec3f& euler, float size); + // Draw primitive types template <size_t N> void DrawPolygon(QPainter& p, const std::array<QPointF, N>& polygon); @@ -222,4 +226,5 @@ private: Core::HID::SticksValues stick_values{}; Core::HID::TriggerValues trigger_values{}; Core::HID::BatteryValues battery_values{}; + Core::HID::MotionState motion_values{}; }; diff --git a/src/yuzu/configuration/configure_system.cpp b/src/yuzu/configuration/configure_system.cpp index 6af34f793..286ccc5cd 100644 --- a/src/yuzu/configuration/configure_system.cpp +++ b/src/yuzu/configuration/configure_system.cpp @@ -111,6 +111,9 @@ void ConfigureSystem::SetConfiguration() { ui->custom_rtc_edit->setDateTime(QDateTime::fromSecsSinceEpoch(rtc_time)); ui->device_name_edit->setText( QString::fromUtf8(Settings::values.device_name.GetValue().c_str())); + ui->use_unsafe_extended_memory_layout->setEnabled(enabled); + ui->use_unsafe_extended_memory_layout->setChecked( + Settings::values.use_unsafe_extended_memory_layout.GetValue()); if (Settings::IsConfiguringGlobal()) { ui->combo_language->setCurrentIndex(Settings::values.language_index.GetValue()); @@ -160,6 +163,9 @@ void ConfigureSystem::ApplyConfiguration() { ConfigurationShared::ApplyPerGameSetting(&Settings::values.region_index, ui->combo_region); ConfigurationShared::ApplyPerGameSetting(&Settings::values.time_zone_index, ui->combo_time_zone); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_unsafe_extended_memory_layout, + ui->use_unsafe_extended_memory_layout, + use_unsafe_extended_memory_layout); if (Settings::IsConfiguringGlobal()) { // Guard if during game and set to game-specific value @@ -215,6 +221,10 @@ void ConfigureSystem::SetupPerGameUI() { Settings::values.rng_seed.GetValue().has_value(), Settings::values.rng_seed.GetValue(true).has_value(), use_rng_seed); + ConfigurationShared::SetColoredTristate(ui->use_unsafe_extended_memory_layout, + Settings::values.use_unsafe_extended_memory_layout, + use_unsafe_extended_memory_layout); + ui->custom_rtc_checkbox->setVisible(false); ui->custom_rtc_edit->setVisible(false); } diff --git a/src/yuzu/configuration/configure_system.h b/src/yuzu/configuration/configure_system.h index ec28724a1..ce1a91601 100644 --- a/src/yuzu/configuration/configure_system.h +++ b/src/yuzu/configuration/configure_system.h @@ -41,6 +41,7 @@ private: bool enabled = false; ConfigurationShared::CheckState use_rng_seed; + ConfigurationShared::CheckState use_unsafe_extended_memory_layout; Core::System& system; }; diff --git a/src/yuzu/configuration/configure_system.ui b/src/yuzu/configuration/configure_system.ui index 9e7bc3b93..e0caecd5e 100644 --- a/src/yuzu/configuration/configure_system.ui +++ b/src/yuzu/configuration/configure_system.ui @@ -478,6 +478,13 @@ </property> </widget> </item> + <item row="7" column="0"> + <widget class="QCheckBox" name="use_unsafe_extended_memory_layout"> + <property name="text"> + <string>Unsafe extended memory layout (8GB DRAM)</string> + </property> + </widget> + </item> </layout> </item> </layout> diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp index a692ef809..a6418e693 100644 --- a/src/yuzu_cmd/config.cpp +++ b/src/yuzu_cmd/config.cpp @@ -4,18 +4,8 @@ #include <memory> #include <optional> #include <sstream> - -// Ignore -Wimplicit-fallthrough due to https://github.com/libsdl-org/SDL/issues/4307 -#ifdef __clang__ -#pragma clang diagnostic push -#pragma clang diagnostic ignored "-Wimplicit-fallthrough" -#endif -#include <SDL.h> -#ifdef __clang__ -#pragma clang diagnostic pop -#endif - #include <INIReader.h> +#include <SDL.h> #include "common/fs/file.h" #include "common/fs/fs.h" #include "common/fs/path_util.h" @@ -274,7 +264,7 @@ void Config::ReadValues() { // Core ReadSetting("Core", Settings::values.use_multi_core); - ReadSetting("Core", Settings::values.use_extended_memory_layout); + ReadSetting("Core", Settings::values.use_unsafe_extended_memory_layout); // Cpu ReadSetting("Cpu", Settings::values.cpu_accuracy); @@ -327,7 +317,6 @@ void Config::ReadValues() { ReadSetting("Renderer", Settings::values.accelerate_astc); ReadSetting("Renderer", Settings::values.async_astc); ReadSetting("Renderer", Settings::values.use_fast_gpu_time); - ReadSetting("Renderer", Settings::values.use_pessimistic_flushes); ReadSetting("Renderer", Settings::values.use_vulkan_driver_pipeline_cache); ReadSetting("Renderer", Settings::values.bg_red); diff --git a/src/yuzu_cmd/default_ini.h b/src/yuzu_cmd/default_ini.h index 1990c0707..086ed4cfa 100644 --- a/src/yuzu_cmd/default_ini.h +++ b/src/yuzu_cmd/default_ini.h @@ -163,9 +163,9 @@ keyboard_enabled = # 0: Disabled, 1 (default): Enabled use_multi_core = -# Enable extended guest system memory layout (8GB DRAM) +# Enable unsafe extended guest system memory layout (8GB DRAM) # 0 (default): Disabled, 1: Enabled -use_extended_memory_layout = +use_unsafe_extended_memory_layout = [Cpu] # Adjusts various optimizations. @@ -380,10 +380,6 @@ use_asynchronous_gpu_emulation = # 0: Off, 1 (default): On use_fast_gpu_time = -# Force unmodified buffers to be flushed, which can cost performance. -# 0: Off (default), 1: On -use_pessimistic_flushes = - # Whether to use garbage collection or not for GPU caches. # 0 (default): Off, 1: On use_caches_gc = diff --git a/vcpkg.json b/vcpkg.json index 0352dab77..19f99e89e 100644 --- a/vcpkg.json +++ b/vcpkg.json @@ -49,7 +49,7 @@ "overrides": [ { "name": "catch2", - "version": "3.0.1" + "version": "3.3.1" }, { "name": "fmt", |