Compare commits
30 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
| c133509368 | |||
| e3122c5b46 | |||
| 7eb7d56b1b | |||
| 8cb6b33809 | |||
| 8ad64bc253 | |||
| eea2145698 | |||
| e684515578 | |||
| ae1a8a7dc7 | |||
| fd5d7947f6 | |||
| a67bdeb2c2 | |||
| f1e12e3b08 | |||
| 93061d1ea1 | |||
| 6d12e7320b | |||
| 78ff2862f6 | |||
| 197e13d93d | |||
| e5f1b22e16 | |||
| b0beca52a3 | |||
| 711190bb67 | |||
| b9a86b040b | |||
| 346c253cd2 | |||
| 76a676883a | |||
| 8d6aefdcc4 | |||
| 3e6d81a008 | |||
| 2e1e725443 | |||
| 907507886d | |||
| 9dcc7bde8b | |||
| 8e56a84566 | |||
| bbd502f67a | |||
| 1492a65454 | |||
| dd12dd4c67 |
+101
-71
@@ -68,79 +68,109 @@ class HomeSettingsFragment : Fragment() {
|
||||
override fun onViewCreated(view: View, savedInstanceState: Bundle?) {
|
||||
mainActivity = requireActivity() as MainActivity
|
||||
|
||||
val optionsList: MutableList<HomeSetting> = mutableListOf(
|
||||
HomeSetting(
|
||||
R.string.advanced_settings,
|
||||
R.string.settings_description,
|
||||
R.drawable.ic_settings
|
||||
) { SettingsActivity.launch(requireContext(), SettingsFile.FILE_NAME_CONFIG, "") },
|
||||
HomeSetting(
|
||||
R.string.open_user_folder,
|
||||
R.string.open_user_folder_description,
|
||||
R.drawable.ic_folder_open
|
||||
) { openFileManager() },
|
||||
HomeSetting(
|
||||
R.string.preferences_theme,
|
||||
R.string.theme_and_color_description,
|
||||
R.drawable.ic_palette
|
||||
) { SettingsActivity.launch(requireContext(), Settings.SECTION_THEME, "") },
|
||||
HomeSetting(
|
||||
R.string.install_gpu_driver,
|
||||
R.string.install_gpu_driver_description,
|
||||
R.drawable.ic_exit
|
||||
) { driverInstaller() },
|
||||
HomeSetting(
|
||||
R.string.install_amiibo_keys,
|
||||
R.string.install_amiibo_keys_description,
|
||||
R.drawable.ic_nfc
|
||||
) { mainActivity.getAmiiboKey.launch(arrayOf("*/*")) },
|
||||
HomeSetting(
|
||||
R.string.install_game_content,
|
||||
R.string.install_game_content_description,
|
||||
R.drawable.ic_system_update_alt
|
||||
) { mainActivity.installGameUpdate.launch(arrayOf("*/*")) },
|
||||
HomeSetting(
|
||||
R.string.select_games_folder,
|
||||
R.string.select_games_folder_description,
|
||||
R.drawable.ic_add
|
||||
) {
|
||||
mainActivity.getGamesDirectory.launch(Intent(Intent.ACTION_OPEN_DOCUMENT_TREE).data)
|
||||
},
|
||||
HomeSetting(
|
||||
R.string.manage_save_data,
|
||||
R.string.import_export_saves_description,
|
||||
R.drawable.ic_save
|
||||
) {
|
||||
ImportExportSavesFragment().show(
|
||||
parentFragmentManager,
|
||||
ImportExportSavesFragment.TAG
|
||||
val optionsList: MutableList<HomeSetting> = mutableListOf<HomeSetting>().apply {
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.advanced_settings,
|
||||
R.string.settings_description,
|
||||
R.drawable.ic_settings
|
||||
) { SettingsActivity.launch(requireContext(), SettingsFile.FILE_NAME_CONFIG, "") }
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.open_user_folder,
|
||||
R.string.open_user_folder_description,
|
||||
R.drawable.ic_folder_open
|
||||
) { openFileManager() }
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.preferences_theme,
|
||||
R.string.theme_and_color_description,
|
||||
R.drawable.ic_palette
|
||||
) { SettingsActivity.launch(requireContext(), Settings.SECTION_THEME, "") }
|
||||
)
|
||||
|
||||
if (GpuDriverHelper.supportsCustomDriverLoading()) {
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.install_gpu_driver,
|
||||
R.string.install_gpu_driver_description,
|
||||
R.drawable.ic_exit
|
||||
) { driverInstaller() }
|
||||
)
|
||||
},
|
||||
HomeSetting(
|
||||
R.string.install_prod_keys,
|
||||
R.string.install_prod_keys_description,
|
||||
R.drawable.ic_unlock
|
||||
) { mainActivity.getProdKey.launch(arrayOf("*/*")) },
|
||||
HomeSetting(
|
||||
R.string.install_firmware,
|
||||
R.string.install_firmware_description,
|
||||
R.drawable.ic_firmware
|
||||
) { mainActivity.getFirmware.launch(arrayOf("application/zip")) },
|
||||
HomeSetting(
|
||||
R.string.share_log,
|
||||
R.string.share_log_description,
|
||||
R.drawable.ic_log
|
||||
) { shareLog() },
|
||||
HomeSetting(
|
||||
R.string.about,
|
||||
R.string.about_description,
|
||||
R.drawable.ic_info_outline
|
||||
) {
|
||||
exitTransition = MaterialSharedAxis(MaterialSharedAxis.X, true)
|
||||
parentFragmentManager.primaryNavigationFragment?.findNavController()
|
||||
?.navigate(R.id.action_homeSettingsFragment_to_aboutFragment)
|
||||
}
|
||||
)
|
||||
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.install_amiibo_keys,
|
||||
R.string.install_amiibo_keys_description,
|
||||
R.drawable.ic_nfc
|
||||
) { mainActivity.getAmiiboKey.launch(arrayOf("*/*")) }
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.install_game_content,
|
||||
R.string.install_game_content_description,
|
||||
R.drawable.ic_system_update_alt
|
||||
) { mainActivity.installGameUpdate.launch(arrayOf("*/*")) }
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.select_games_folder,
|
||||
R.string.select_games_folder_description,
|
||||
R.drawable.ic_add
|
||||
) {
|
||||
mainActivity.getGamesDirectory.launch(
|
||||
Intent(Intent.ACTION_OPEN_DOCUMENT_TREE).data
|
||||
)
|
||||
}
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.manage_save_data,
|
||||
R.string.import_export_saves_description,
|
||||
R.drawable.ic_save
|
||||
) {
|
||||
ImportExportSavesFragment().show(
|
||||
parentFragmentManager,
|
||||
ImportExportSavesFragment.TAG
|
||||
)
|
||||
}
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.install_prod_keys,
|
||||
R.string.install_prod_keys_description,
|
||||
R.drawable.ic_unlock
|
||||
) { mainActivity.getProdKey.launch(arrayOf("*/*")) }
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.install_firmware,
|
||||
R.string.install_firmware_description,
|
||||
R.drawable.ic_firmware
|
||||
) { mainActivity.getFirmware.launch(arrayOf("application/zip")) }
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.share_log,
|
||||
R.string.share_log_description,
|
||||
R.drawable.ic_log
|
||||
) { shareLog() }
|
||||
)
|
||||
add(
|
||||
HomeSetting(
|
||||
R.string.about,
|
||||
R.string.about_description,
|
||||
R.drawable.ic_info_outline
|
||||
) {
|
||||
exitTransition = MaterialSharedAxis(MaterialSharedAxis.X, true)
|
||||
parentFragmentManager.primaryNavigationFragment?.findNavController()
|
||||
?.navigate(R.id.action_homeSettingsFragment_to_aboutFragment)
|
||||
}
|
||||
)
|
||||
}
|
||||
|
||||
if (!BuildConfig.PREMIUM) {
|
||||
optionsList.add(
|
||||
|
||||
@@ -113,6 +113,8 @@ object GpuDriverHelper {
|
||||
initializeDriverParameters(context)
|
||||
}
|
||||
|
||||
external fun supportsCustomDriverLoading(): Boolean
|
||||
|
||||
// Parse the custom driver metadata to retrieve the name.
|
||||
val customDriverName: String?
|
||||
get() {
|
||||
|
||||
@@ -237,6 +237,7 @@ public:
|
||||
m_software_keyboard = android_keyboard.get();
|
||||
m_system.SetShuttingDown(false);
|
||||
m_system.ApplySettings();
|
||||
Settings::LogSettings();
|
||||
m_system.HIDCore().ReloadInputDevices();
|
||||
m_system.SetAppletFrontendSet({
|
||||
nullptr, // Amiibo Settings
|
||||
@@ -560,6 +561,26 @@ void JNICALL Java_org_yuzu_yuzu_1emu_NativeLibrary_initializeGpuDriver(
|
||||
GetJString(env, custom_driver_name), GetJString(env, file_redirect_dir));
|
||||
}
|
||||
|
||||
[[maybe_unused]] static bool CheckKgslPresent() {
|
||||
constexpr auto KgslPath{"/dev/kgsl-3d0"};
|
||||
|
||||
return access(KgslPath, F_OK) == 0;
|
||||
}
|
||||
|
||||
[[maybe_unused]] bool SupportsCustomDriver() {
|
||||
return android_get_device_api_level() >= 28 && CheckKgslPresent();
|
||||
}
|
||||
|
||||
jboolean JNICALL Java_org_yuzu_yuzu_1emu_utils_GpuDriverHelper_supportsCustomDriverLoading(
|
||||
[[maybe_unused]] JNIEnv* env, [[maybe_unused]] jobject instance) {
|
||||
#ifdef ARCHITECTURE_arm64
|
||||
// If the KGSL device exists custom drivers can be loaded using adrenotools
|
||||
return SupportsCustomDriver();
|
||||
#else
|
||||
return false;
|
||||
#endif
|
||||
}
|
||||
|
||||
jboolean Java_org_yuzu_yuzu_1emu_NativeLibrary_reloadKeys(JNIEnv* env,
|
||||
[[maybe_unused]] jclass clazz) {
|
||||
Core::Crypto::KeyManager::Instance().ReloadKeys();
|
||||
|
||||
@@ -7,7 +7,6 @@
|
||||
#include "common/logging/log.h"
|
||||
#include "core/core.h"
|
||||
#include "core/core_timing.h"
|
||||
#include "core/core_timing_util.h"
|
||||
#include "core/memory.h"
|
||||
|
||||
namespace AudioCore::AudioRenderer::ADSP {
|
||||
|
||||
@@ -13,7 +13,6 @@
|
||||
#include "common/thread.h"
|
||||
#include "core/core.h"
|
||||
#include "core/core_timing.h"
|
||||
#include "core/core_timing_util.h"
|
||||
|
||||
MICROPROFILE_DEFINE(Audio_Renderer, "Audio", "DSP", MP_RGB(60, 19, 97));
|
||||
|
||||
@@ -144,6 +143,7 @@ void AudioRenderer::ThreadFunc(std::stop_token stop_token) {
|
||||
|
||||
mailbox->ADSPSendMessage(RenderMessage::AudioRenderer_InitializeOK);
|
||||
|
||||
// 0.12 seconds (2304000 / 19200000)
|
||||
constexpr u64 max_process_time{2'304'000ULL};
|
||||
|
||||
while (!stop_token.stop_requested()) {
|
||||
@@ -184,8 +184,7 @@ void AudioRenderer::ThreadFunc(std::stop_token stop_token) {
|
||||
u64 max_time{max_process_time};
|
||||
if (index == 1 && command_buffer.applet_resource_user_id ==
|
||||
mailbox->GetCommandBuffer(0).applet_resource_user_id) {
|
||||
max_time = max_process_time -
|
||||
Core::Timing::CyclesToNs(render_times_taken[0]).count();
|
||||
max_time = max_process_time - render_times_taken[0];
|
||||
if (render_times_taken[0] > max_process_time) {
|
||||
max_time = 0;
|
||||
}
|
||||
|
||||
@@ -9,7 +9,6 @@
|
||||
#include "common/settings.h"
|
||||
#include "core/core.h"
|
||||
#include "core/core_timing.h"
|
||||
#include "core/core_timing_util.h"
|
||||
#include "core/memory.h"
|
||||
|
||||
namespace AudioCore::AudioRenderer::ADSP {
|
||||
|
||||
@@ -5,7 +5,6 @@
|
||||
#include "audio_core/renderer/command/performance/performance.h"
|
||||
#include "core/core.h"
|
||||
#include "core/core_timing.h"
|
||||
#include "core/core_timing_util.h"
|
||||
|
||||
namespace AudioCore::AudioRenderer {
|
||||
|
||||
@@ -18,20 +17,18 @@ void PerformanceCommand::Process(const ADSP::CommandListProcessor& processor) {
|
||||
auto base{entry_address.translated_address};
|
||||
if (state == PerformanceState::Start) {
|
||||
auto start_time_ptr{reinterpret_cast<u32*>(base + entry_address.entry_start_time_offset)};
|
||||
*start_time_ptr = static_cast<u32>(
|
||||
Core::Timing::CyclesToUs(processor.system->CoreTiming().GetClockTicks() -
|
||||
processor.start_time - processor.current_processing_time)
|
||||
.count());
|
||||
*start_time_ptr =
|
||||
static_cast<u32>(processor.system->CoreTiming().GetClockTicks() - processor.start_time -
|
||||
processor.current_processing_time);
|
||||
} else if (state == PerformanceState::Stop) {
|
||||
auto processed_time_ptr{
|
||||
reinterpret_cast<u32*>(base + entry_address.entry_processed_time_offset)};
|
||||
auto entry_count_ptr{
|
||||
reinterpret_cast<u32*>(base + entry_address.header_entry_count_offset)};
|
||||
|
||||
*processed_time_ptr = static_cast<u32>(
|
||||
Core::Timing::CyclesToUs(processor.system->CoreTiming().GetClockTicks() -
|
||||
processor.start_time - processor.current_processing_time)
|
||||
.count());
|
||||
*processed_time_ptr =
|
||||
static_cast<u32>(processor.system->CoreTiming().GetClockTicks() - processor.start_time -
|
||||
processor.current_processing_time);
|
||||
(*entry_count_ptr)++;
|
||||
}
|
||||
}
|
||||
|
||||
@@ -15,7 +15,6 @@
|
||||
#include "common/settings.h"
|
||||
#include "core/core.h"
|
||||
#include "core/core_timing.h"
|
||||
#include "core/core_timing_util.h"
|
||||
|
||||
namespace AudioCore::Sink {
|
||||
|
||||
|
||||
@@ -172,6 +172,8 @@ if(ARCHITECTURE_x86_64)
|
||||
x64/cpu_wait.h
|
||||
x64/native_clock.cpp
|
||||
x64/native_clock.h
|
||||
x64/rdtsc.cpp
|
||||
x64/rdtsc.h
|
||||
x64/xbyak_abi.h
|
||||
x64/xbyak_util.h
|
||||
)
|
||||
|
||||
@@ -483,6 +483,7 @@ struct Values {
|
||||
AstcRecompression::Uncompressed, AstcRecompression::Uncompressed, AstcRecompression::Bc3,
|
||||
"astc_recompression"};
|
||||
SwitchableSetting<bool> use_video_framerate{false, "use_video_framerate"};
|
||||
SwitchableSetting<bool> barrier_feedback_loops{true, "barrier_feedback_loops"};
|
||||
|
||||
SwitchableSetting<u8> bg_red{0, "bg_red"};
|
||||
SwitchableSetting<u8> bg_green{0, "bg_green"};
|
||||
|
||||
@@ -28,13 +28,12 @@ static s64 GetSystemTimeNS() {
|
||||
// GetSystemTimePreciseAsFileTime returns the file time in 100ns units.
|
||||
static constexpr s64 Multiplier = 100;
|
||||
// Convert Windows epoch to Unix epoch.
|
||||
static constexpr s64 WindowsEpochToUnixEpochNS = 0x19DB1DED53E8000LL;
|
||||
static constexpr s64 WindowsEpochToUnixEpoch = 0x19DB1DED53E8000LL;
|
||||
|
||||
FILETIME filetime;
|
||||
GetSystemTimePreciseAsFileTime(&filetime);
|
||||
return Multiplier * ((static_cast<s64>(filetime.dwHighDateTime) << 32) +
|
||||
static_cast<s64>(filetime.dwLowDateTime)) -
|
||||
WindowsEpochToUnixEpochNS;
|
||||
static_cast<s64>(filetime.dwLowDateTime) - WindowsEpochToUnixEpoch);
|
||||
}
|
||||
#endif
|
||||
|
||||
|
||||
+32
-45
@@ -2,88 +2,75 @@
|
||||
// SPDX-License-Identifier: GPL-2.0-or-later
|
||||
|
||||
#include "common/steady_clock.h"
|
||||
#include "common/uint128.h"
|
||||
#include "common/wall_clock.h"
|
||||
|
||||
#ifdef ARCHITECTURE_x86_64
|
||||
#include "common/x64/cpu_detect.h"
|
||||
#include "common/x64/native_clock.h"
|
||||
#include "common/x64/rdtsc.h"
|
||||
#endif
|
||||
|
||||
namespace Common {
|
||||
|
||||
class StandardWallClock final : public WallClock {
|
||||
public:
|
||||
explicit StandardWallClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequency_)
|
||||
: WallClock{emulated_cpu_frequency_, emulated_clock_frequency_, false},
|
||||
start_time{SteadyClock::Now()} {}
|
||||
explicit StandardWallClock() : start_time{SteadyClock::Now()} {}
|
||||
|
||||
std::chrono::nanoseconds GetTimeNS() override {
|
||||
std::chrono::nanoseconds GetTimeNS() const override {
|
||||
return SteadyClock::Now() - start_time;
|
||||
}
|
||||
|
||||
std::chrono::microseconds GetTimeUS() override {
|
||||
return std::chrono::duration_cast<std::chrono::microseconds>(GetTimeNS());
|
||||
std::chrono::microseconds GetTimeUS() const override {
|
||||
return static_cast<std::chrono::microseconds>(GetHostTicksElapsed() / NsToUsRatio::den);
|
||||
}
|
||||
|
||||
std::chrono::milliseconds GetTimeMS() override {
|
||||
return std::chrono::duration_cast<std::chrono::milliseconds>(GetTimeNS());
|
||||
std::chrono::milliseconds GetTimeMS() const override {
|
||||
return static_cast<std::chrono::milliseconds>(GetHostTicksElapsed() / NsToMsRatio::den);
|
||||
}
|
||||
|
||||
u64 GetClockCycles() override {
|
||||
const u128 temp = Common::Multiply64Into128(GetTimeNS().count(), emulated_clock_frequency);
|
||||
return Common::Divide128On32(temp, NS_RATIO).first;
|
||||
u64 GetCNTPCT() const override {
|
||||
return GetHostTicksElapsed() * NsToCNTPCTRatio::num / NsToCNTPCTRatio::den;
|
||||
}
|
||||
|
||||
u64 GetCPUCycles() override {
|
||||
const u128 temp = Common::Multiply64Into128(GetTimeNS().count(), emulated_cpu_frequency);
|
||||
return Common::Divide128On32(temp, NS_RATIO).first;
|
||||
u64 GetGPUTick() const override {
|
||||
return GetHostTicksElapsed() * NsToGPUTickRatio::num / NsToGPUTickRatio::den;
|
||||
}
|
||||
|
||||
void Pause([[maybe_unused]] bool is_paused) override {
|
||||
// Do nothing in this clock type.
|
||||
u64 GetHostTicksNow() const override {
|
||||
return static_cast<u64>(SteadyClock::Now().time_since_epoch().count());
|
||||
}
|
||||
|
||||
u64 GetHostTicksElapsed() const override {
|
||||
return static_cast<u64>(GetTimeNS().count());
|
||||
}
|
||||
|
||||
bool IsNative() const override {
|
||||
return false;
|
||||
}
|
||||
|
||||
private:
|
||||
SteadyClock::time_point start_time;
|
||||
};
|
||||
|
||||
std::unique_ptr<WallClock> CreateOptimalClock() {
|
||||
#ifdef ARCHITECTURE_x86_64
|
||||
|
||||
std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
|
||||
u64 emulated_clock_frequency) {
|
||||
const auto& caps = GetCPUCaps();
|
||||
u64 rtsc_frequency = 0;
|
||||
if (caps.invariant_tsc) {
|
||||
rtsc_frequency = caps.tsc_frequency ? caps.tsc_frequency : EstimateRDTSCFrequency();
|
||||
}
|
||||
|
||||
// Fallback to StandardWallClock if the hardware TSC does not have the precision greater than:
|
||||
// - A nanosecond
|
||||
// - The emulated CPU frequency
|
||||
// - The emulated clock counter frequency (CNTFRQ)
|
||||
if (rtsc_frequency <= WallClock::NS_RATIO || rtsc_frequency <= emulated_cpu_frequency ||
|
||||
rtsc_frequency <= emulated_clock_frequency) {
|
||||
return std::make_unique<StandardWallClock>(emulated_cpu_frequency,
|
||||
emulated_clock_frequency);
|
||||
if (caps.invariant_tsc && caps.tsc_frequency >= WallClock::GPUTickFreq) {
|
||||
return std::make_unique<X64::NativeClock>(caps.tsc_frequency);
|
||||
} else {
|
||||
return std::make_unique<X64::NativeClock>(emulated_cpu_frequency, emulated_clock_frequency,
|
||||
rtsc_frequency);
|
||||
// Fallback to StandardWallClock if the hardware TSC
|
||||
// - Is not invariant
|
||||
// - Is not more precise than GPUTickFreq
|
||||
return std::make_unique<StandardWallClock>();
|
||||
}
|
||||
}
|
||||
|
||||
#else
|
||||
|
||||
std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
|
||||
u64 emulated_clock_frequency) {
|
||||
return std::make_unique<StandardWallClock>(emulated_cpu_frequency, emulated_clock_frequency);
|
||||
return std::make_unique<StandardWallClock>();
|
||||
#endif
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
std::unique_ptr<WallClock> CreateStandardWallClock(u64 emulated_cpu_frequency,
|
||||
u64 emulated_clock_frequency) {
|
||||
return std::make_unique<StandardWallClock>(emulated_cpu_frequency, emulated_clock_frequency);
|
||||
std::unique_ptr<WallClock> CreateStandardWallClock() {
|
||||
return std::make_unique<StandardWallClock>();
|
||||
}
|
||||
|
||||
} // namespace Common
|
||||
|
||||
+61
-28
@@ -5,6 +5,7 @@
|
||||
|
||||
#include <chrono>
|
||||
#include <memory>
|
||||
#include <ratio>
|
||||
|
||||
#include "common/common_types.h"
|
||||
|
||||
@@ -12,50 +13,82 @@ namespace Common {
|
||||
|
||||
class WallClock {
|
||||
public:
|
||||
static constexpr u64 NS_RATIO = 1'000'000'000;
|
||||
static constexpr u64 US_RATIO = 1'000'000;
|
||||
static constexpr u64 MS_RATIO = 1'000;
|
||||
static constexpr u64 CNTFRQ = 19'200'000; // CNTPCT_EL0 Frequency = 19.2 MHz
|
||||
static constexpr u64 GPUTickFreq = 614'400'000; // GM20B GPU Tick Frequency = 614.4 MHz
|
||||
static constexpr u64 CPUTickFreq = 1'020'000'000; // T210/4 A57 CPU Tick Frequency = 1020.0 MHz
|
||||
|
||||
virtual ~WallClock() = default;
|
||||
|
||||
/// Returns current wall time in nanoseconds
|
||||
[[nodiscard]] virtual std::chrono::nanoseconds GetTimeNS() = 0;
|
||||
/// @returns The time in nanoseconds since the construction of this clock.
|
||||
virtual std::chrono::nanoseconds GetTimeNS() const = 0;
|
||||
|
||||
/// Returns current wall time in microseconds
|
||||
[[nodiscard]] virtual std::chrono::microseconds GetTimeUS() = 0;
|
||||
/// @returns The time in microseconds since the construction of this clock.
|
||||
virtual std::chrono::microseconds GetTimeUS() const = 0;
|
||||
|
||||
/// Returns current wall time in milliseconds
|
||||
[[nodiscard]] virtual std::chrono::milliseconds GetTimeMS() = 0;
|
||||
/// @returns The time in milliseconds since the construction of this clock.
|
||||
virtual std::chrono::milliseconds GetTimeMS() const = 0;
|
||||
|
||||
/// Returns current wall time in emulated clock cycles
|
||||
[[nodiscard]] virtual u64 GetClockCycles() = 0;
|
||||
/// @returns The guest CNTPCT ticks since the construction of this clock.
|
||||
virtual u64 GetCNTPCT() const = 0;
|
||||
|
||||
/// Returns current wall time in emulated cpu cycles
|
||||
[[nodiscard]] virtual u64 GetCPUCycles() = 0;
|
||||
/// @returns The guest GPU ticks since the construction of this clock.
|
||||
virtual u64 GetGPUTick() const = 0;
|
||||
|
||||
virtual void Pause(bool is_paused) = 0;
|
||||
/// @returns The raw host timer ticks since an indeterminate epoch.
|
||||
virtual u64 GetHostTicksNow() const = 0;
|
||||
|
||||
/// Tells if the wall clock, uses the host CPU's hardware clock
|
||||
[[nodiscard]] bool IsNative() const {
|
||||
return is_native;
|
||||
/// @returns The raw host timer ticks since the construction of this clock.
|
||||
virtual u64 GetHostTicksElapsed() const = 0;
|
||||
|
||||
/// @returns Whether the clock directly uses the host's hardware clock.
|
||||
virtual bool IsNative() const = 0;
|
||||
|
||||
static inline u64 NSToCNTPCT(u64 ns) {
|
||||
return ns * NsToCNTPCTRatio::num / NsToCNTPCTRatio::den;
|
||||
}
|
||||
|
||||
static inline u64 NSToGPUTick(u64 ns) {
|
||||
return ns * NsToGPUTickRatio::num / NsToGPUTickRatio::den;
|
||||
}
|
||||
|
||||
// Cycle Timing
|
||||
|
||||
static inline u64 CPUTickToNS(u64 cpu_tick) {
|
||||
return cpu_tick * CPUTickToNsRatio::num / CPUTickToNsRatio::den;
|
||||
}
|
||||
|
||||
static inline u64 CPUTickToUS(u64 cpu_tick) {
|
||||
return cpu_tick * CPUTickToUsRatio::num / CPUTickToUsRatio::den;
|
||||
}
|
||||
|
||||
static inline u64 CPUTickToCNTPCT(u64 cpu_tick) {
|
||||
return cpu_tick * CPUTickToCNTPCTRatio::num / CPUTickToCNTPCTRatio::den;
|
||||
}
|
||||
|
||||
static inline u64 CPUTickToGPUTick(u64 cpu_tick) {
|
||||
return cpu_tick * CPUTickToGPUTickRatio::num / CPUTickToGPUTickRatio::den;
|
||||
}
|
||||
|
||||
protected:
|
||||
explicit WallClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequency_, bool is_native_)
|
||||
: emulated_cpu_frequency{emulated_cpu_frequency_},
|
||||
emulated_clock_frequency{emulated_clock_frequency_}, is_native{is_native_} {}
|
||||
using NsRatio = std::nano;
|
||||
using UsRatio = std::micro;
|
||||
using MsRatio = std::milli;
|
||||
|
||||
u64 emulated_cpu_frequency;
|
||||
u64 emulated_clock_frequency;
|
||||
using NsToUsRatio = std::ratio_divide<std::nano, std::micro>;
|
||||
using NsToMsRatio = std::ratio_divide<std::nano, std::milli>;
|
||||
using NsToCNTPCTRatio = std::ratio<CNTFRQ, std::nano::den>;
|
||||
using NsToGPUTickRatio = std::ratio<GPUTickFreq, std::nano::den>;
|
||||
|
||||
private:
|
||||
bool is_native;
|
||||
// Cycle Timing
|
||||
|
||||
using CPUTickToNsRatio = std::ratio<std::nano::den, CPUTickFreq>;
|
||||
using CPUTickToUsRatio = std::ratio<std::micro::den, CPUTickFreq>;
|
||||
using CPUTickToCNTPCTRatio = std::ratio<CNTFRQ, CPUTickFreq>;
|
||||
using CPUTickToGPUTickRatio = std::ratio<GPUTickFreq, CPUTickFreq>;
|
||||
};
|
||||
|
||||
[[nodiscard]] std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
|
||||
u64 emulated_clock_frequency);
|
||||
std::unique_ptr<WallClock> CreateOptimalClock();
|
||||
|
||||
[[nodiscard]] std::unique_ptr<WallClock> CreateStandardWallClock(u64 emulated_cpu_frequency,
|
||||
u64 emulated_clock_frequency);
|
||||
std::unique_ptr<WallClock> CreateStandardWallClock();
|
||||
|
||||
} // namespace Common
|
||||
|
||||
@@ -14,6 +14,7 @@
|
||||
#include "common/common_types.h"
|
||||
#include "common/logging/log.h"
|
||||
#include "common/x64/cpu_detect.h"
|
||||
#include "common/x64/rdtsc.h"
|
||||
|
||||
#ifdef _WIN32
|
||||
#include <windows.h>
|
||||
@@ -187,6 +188,8 @@ static CPUCaps Detect() {
|
||||
caps.tsc_frequency = static_cast<u64>(caps.crystal_frequency) *
|
||||
caps.tsc_crystal_ratio_numerator /
|
||||
caps.tsc_crystal_ratio_denominator;
|
||||
} else {
|
||||
caps.tsc_frequency = X64::EstimateRDTSCFrequency();
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@@ -9,19 +9,11 @@
|
||||
|
||||
#include "common/x64/cpu_detect.h"
|
||||
#include "common/x64/cpu_wait.h"
|
||||
#include "common/x64/rdtsc.h"
|
||||
|
||||
namespace Common::X64 {
|
||||
|
||||
#ifdef _MSC_VER
|
||||
__forceinline static u64 FencedRDTSC() {
|
||||
_mm_lfence();
|
||||
_ReadWriteBarrier();
|
||||
const u64 result = __rdtsc();
|
||||
_mm_lfence();
|
||||
_ReadWriteBarrier();
|
||||
return result;
|
||||
}
|
||||
|
||||
__forceinline static void TPAUSE() {
|
||||
// 100,000 cycles is a reasonable amount of time to wait to save on CPU resources.
|
||||
// For reference:
|
||||
@@ -32,16 +24,6 @@ __forceinline static void TPAUSE() {
|
||||
_tpause(0, FencedRDTSC() + PauseCycles);
|
||||
}
|
||||
#else
|
||||
static u64 FencedRDTSC() {
|
||||
u64 eax;
|
||||
u64 edx;
|
||||
asm volatile("lfence\n\t"
|
||||
"rdtsc\n\t"
|
||||
"lfence\n\t"
|
||||
: "=a"(eax), "=d"(edx));
|
||||
return (edx << 32) | eax;
|
||||
}
|
||||
|
||||
static void TPAUSE() {
|
||||
// 100,000 cycles is a reasonable amount of time to wait to save on CPU resources.
|
||||
// For reference:
|
||||
|
||||
+26
-140
@@ -1,164 +1,50 @@
|
||||
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
|
||||
// SPDX-License-Identifier: GPL-2.0-or-later
|
||||
|
||||
#include <array>
|
||||
#include <chrono>
|
||||
#include <thread>
|
||||
|
||||
#include "common/atomic_ops.h"
|
||||
#include "common/steady_clock.h"
|
||||
#include "common/uint128.h"
|
||||
#include "common/x64/native_clock.h"
|
||||
#include "common/x64/rdtsc.h"
|
||||
|
||||
#ifdef _MSC_VER
|
||||
#include <intrin.h>
|
||||
#endif
|
||||
namespace Common::X64 {
|
||||
|
||||
namespace Common {
|
||||
NativeClock::NativeClock(u64 rdtsc_frequency_)
|
||||
: start_ticks{FencedRDTSC()}, rdtsc_frequency{rdtsc_frequency_},
|
||||
ns_rdtsc_factor{GetFixedPoint64Factor(NsRatio::den, rdtsc_frequency)},
|
||||
us_rdtsc_factor{GetFixedPoint64Factor(UsRatio::den, rdtsc_frequency)},
|
||||
ms_rdtsc_factor{GetFixedPoint64Factor(MsRatio::den, rdtsc_frequency)},
|
||||
cntpct_rdtsc_factor{GetFixedPoint64Factor(CNTFRQ, rdtsc_frequency)},
|
||||
gputick_rdtsc_factor{GetFixedPoint64Factor(GPUTickFreq, rdtsc_frequency)} {}
|
||||
|
||||
#ifdef _MSC_VER
|
||||
__forceinline static u64 FencedRDTSC() {
|
||||
_mm_lfence();
|
||||
_ReadWriteBarrier();
|
||||
const u64 result = __rdtsc();
|
||||
_mm_lfence();
|
||||
_ReadWriteBarrier();
|
||||
return result;
|
||||
}
|
||||
#else
|
||||
static u64 FencedRDTSC() {
|
||||
u64 eax;
|
||||
u64 edx;
|
||||
asm volatile("lfence\n\t"
|
||||
"rdtsc\n\t"
|
||||
"lfence\n\t"
|
||||
: "=a"(eax), "=d"(edx));
|
||||
return (edx << 32) | eax;
|
||||
}
|
||||
#endif
|
||||
|
||||
template <u64 Nearest>
|
||||
static u64 RoundToNearest(u64 value) {
|
||||
const auto mod = value % Nearest;
|
||||
return mod >= (Nearest / 2) ? (value - mod + Nearest) : (value - mod);
|
||||
std::chrono::nanoseconds NativeClock::GetTimeNS() const {
|
||||
return std::chrono::nanoseconds{MultiplyHigh(GetHostTicksElapsed(), ns_rdtsc_factor)};
|
||||
}
|
||||
|
||||
u64 EstimateRDTSCFrequency() {
|
||||
// Discard the first result measuring the rdtsc.
|
||||
FencedRDTSC();
|
||||
std::this_thread::sleep_for(std::chrono::milliseconds{1});
|
||||
FencedRDTSC();
|
||||
|
||||
// Get the current time.
|
||||
const auto start_time = Common::RealTimeClock::Now();
|
||||
const u64 tsc_start = FencedRDTSC();
|
||||
// Wait for 250 milliseconds.
|
||||
std::this_thread::sleep_for(std::chrono::milliseconds{250});
|
||||
const auto end_time = Common::RealTimeClock::Now();
|
||||
const u64 tsc_end = FencedRDTSC();
|
||||
// Calculate differences.
|
||||
const u64 timer_diff = static_cast<u64>(
|
||||
std::chrono::duration_cast<std::chrono::nanoseconds>(end_time - start_time).count());
|
||||
const u64 tsc_diff = tsc_end - tsc_start;
|
||||
const u64 tsc_freq = MultiplyAndDivide64(tsc_diff, 1000000000ULL, timer_diff);
|
||||
return RoundToNearest<1000>(tsc_freq);
|
||||
std::chrono::microseconds NativeClock::GetTimeUS() const {
|
||||
return std::chrono::microseconds{MultiplyHigh(GetHostTicksElapsed(), us_rdtsc_factor)};
|
||||
}
|
||||
|
||||
namespace X64 {
|
||||
NativeClock::NativeClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequency_,
|
||||
u64 rtsc_frequency_)
|
||||
: WallClock(emulated_cpu_frequency_, emulated_clock_frequency_, true), rtsc_frequency{
|
||||
rtsc_frequency_} {
|
||||
// Thread to re-adjust the RDTSC frequency after 10 seconds has elapsed.
|
||||
time_sync_thread = std::jthread{[this](std::stop_token token) {
|
||||
// Get the current time.
|
||||
const auto start_time = Common::RealTimeClock::Now();
|
||||
const u64 tsc_start = FencedRDTSC();
|
||||
// Wait for 10 seconds.
|
||||
if (!Common::StoppableTimedWait(token, std::chrono::seconds{10})) {
|
||||
return;
|
||||
}
|
||||
const auto end_time = Common::RealTimeClock::Now();
|
||||
const u64 tsc_end = FencedRDTSC();
|
||||
// Calculate differences.
|
||||
const u64 timer_diff = static_cast<u64>(
|
||||
std::chrono::duration_cast<std::chrono::nanoseconds>(end_time - start_time).count());
|
||||
const u64 tsc_diff = tsc_end - tsc_start;
|
||||
const u64 tsc_freq = MultiplyAndDivide64(tsc_diff, 1000000000ULL, timer_diff);
|
||||
rtsc_frequency = tsc_freq;
|
||||
CalculateAndSetFactors();
|
||||
}};
|
||||
|
||||
time_point.inner.last_measure = FencedRDTSC();
|
||||
time_point.inner.accumulated_ticks = 0U;
|
||||
CalculateAndSetFactors();
|
||||
std::chrono::milliseconds NativeClock::GetTimeMS() const {
|
||||
return std::chrono::milliseconds{MultiplyHigh(GetHostTicksElapsed(), ms_rdtsc_factor)};
|
||||
}
|
||||
|
||||
u64 NativeClock::GetRTSC() {
|
||||
TimePoint new_time_point{};
|
||||
TimePoint current_time_point{};
|
||||
|
||||
current_time_point.pack = Common::AtomicLoad128(time_point.pack.data());
|
||||
do {
|
||||
const u64 current_measure = FencedRDTSC();
|
||||
u64 diff = current_measure - current_time_point.inner.last_measure;
|
||||
diff = diff & ~static_cast<u64>(static_cast<s64>(diff) >> 63); // max(diff, 0)
|
||||
new_time_point.inner.last_measure = current_measure > current_time_point.inner.last_measure
|
||||
? current_measure
|
||||
: current_time_point.inner.last_measure;
|
||||
new_time_point.inner.accumulated_ticks = current_time_point.inner.accumulated_ticks + diff;
|
||||
} while (!Common::AtomicCompareAndSwap(time_point.pack.data(), new_time_point.pack,
|
||||
current_time_point.pack, current_time_point.pack));
|
||||
return new_time_point.inner.accumulated_ticks;
|
||||
u64 NativeClock::GetCNTPCT() const {
|
||||
return MultiplyHigh(GetHostTicksElapsed(), cntpct_rdtsc_factor);
|
||||
}
|
||||
|
||||
void NativeClock::Pause(bool is_paused) {
|
||||
if (!is_paused) {
|
||||
TimePoint current_time_point{};
|
||||
TimePoint new_time_point{};
|
||||
|
||||
current_time_point.pack = Common::AtomicLoad128(time_point.pack.data());
|
||||
do {
|
||||
new_time_point.pack = current_time_point.pack;
|
||||
new_time_point.inner.last_measure = FencedRDTSC();
|
||||
} while (!Common::AtomicCompareAndSwap(time_point.pack.data(), new_time_point.pack,
|
||||
current_time_point.pack, current_time_point.pack));
|
||||
}
|
||||
u64 NativeClock::GetGPUTick() const {
|
||||
return MultiplyHigh(GetHostTicksElapsed(), gputick_rdtsc_factor);
|
||||
}
|
||||
|
||||
std::chrono::nanoseconds NativeClock::GetTimeNS() {
|
||||
const u64 rtsc_value = GetRTSC();
|
||||
return std::chrono::nanoseconds{MultiplyHigh(rtsc_value, ns_rtsc_factor)};
|
||||
u64 NativeClock::GetHostTicksNow() const {
|
||||
return FencedRDTSC();
|
||||
}
|
||||
|
||||
std::chrono::microseconds NativeClock::GetTimeUS() {
|
||||
const u64 rtsc_value = GetRTSC();
|
||||
return std::chrono::microseconds{MultiplyHigh(rtsc_value, us_rtsc_factor)};
|
||||
u64 NativeClock::GetHostTicksElapsed() const {
|
||||
return FencedRDTSC() - start_ticks;
|
||||
}
|
||||
|
||||
std::chrono::milliseconds NativeClock::GetTimeMS() {
|
||||
const u64 rtsc_value = GetRTSC();
|
||||
return std::chrono::milliseconds{MultiplyHigh(rtsc_value, ms_rtsc_factor)};
|
||||
bool NativeClock::IsNative() const {
|
||||
return true;
|
||||
}
|
||||
|
||||
u64 NativeClock::GetClockCycles() {
|
||||
const u64 rtsc_value = GetRTSC();
|
||||
return MultiplyHigh(rtsc_value, clock_rtsc_factor);
|
||||
}
|
||||
|
||||
u64 NativeClock::GetCPUCycles() {
|
||||
const u64 rtsc_value = GetRTSC();
|
||||
return MultiplyHigh(rtsc_value, cpu_rtsc_factor);
|
||||
}
|
||||
|
||||
void NativeClock::CalculateAndSetFactors() {
|
||||
ns_rtsc_factor = GetFixedPoint64Factor(NS_RATIO, rtsc_frequency);
|
||||
us_rtsc_factor = GetFixedPoint64Factor(US_RATIO, rtsc_frequency);
|
||||
ms_rtsc_factor = GetFixedPoint64Factor(MS_RATIO, rtsc_frequency);
|
||||
clock_rtsc_factor = GetFixedPoint64Factor(emulated_clock_frequency, rtsc_frequency);
|
||||
cpu_rtsc_factor = GetFixedPoint64Factor(emulated_cpu_frequency, rtsc_frequency);
|
||||
}
|
||||
|
||||
} // namespace X64
|
||||
|
||||
} // namespace Common
|
||||
} // namespace Common::X64
|
||||
|
||||
@@ -3,58 +3,39 @@
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "common/polyfill_thread.h"
|
||||
#include "common/wall_clock.h"
|
||||
|
||||
namespace Common {
|
||||
namespace Common::X64 {
|
||||
|
||||
namespace X64 {
|
||||
class NativeClock final : public WallClock {
|
||||
public:
|
||||
explicit NativeClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequency_,
|
||||
u64 rtsc_frequency_);
|
||||
explicit NativeClock(u64 rdtsc_frequency_);
|
||||
|
||||
std::chrono::nanoseconds GetTimeNS() override;
|
||||
std::chrono::nanoseconds GetTimeNS() const override;
|
||||
|
||||
std::chrono::microseconds GetTimeUS() override;
|
||||
std::chrono::microseconds GetTimeUS() const override;
|
||||
|
||||
std::chrono::milliseconds GetTimeMS() override;
|
||||
std::chrono::milliseconds GetTimeMS() const override;
|
||||
|
||||
u64 GetClockCycles() override;
|
||||
u64 GetCNTPCT() const override;
|
||||
|
||||
u64 GetCPUCycles() override;
|
||||
u64 GetGPUTick() const override;
|
||||
|
||||
void Pause(bool is_paused) override;
|
||||
u64 GetHostTicksNow() const override;
|
||||
|
||||
u64 GetHostTicksElapsed() const override;
|
||||
|
||||
bool IsNative() const override;
|
||||
|
||||
private:
|
||||
u64 GetRTSC();
|
||||
u64 start_ticks;
|
||||
u64 rdtsc_frequency;
|
||||
|
||||
void CalculateAndSetFactors();
|
||||
|
||||
union alignas(16) TimePoint {
|
||||
TimePoint() : pack{} {}
|
||||
u128 pack{};
|
||||
struct Inner {
|
||||
u64 last_measure{};
|
||||
u64 accumulated_ticks{};
|
||||
} inner;
|
||||
};
|
||||
|
||||
TimePoint time_point;
|
||||
|
||||
// factors
|
||||
u64 clock_rtsc_factor{};
|
||||
u64 cpu_rtsc_factor{};
|
||||
u64 ns_rtsc_factor{};
|
||||
u64 us_rtsc_factor{};
|
||||
u64 ms_rtsc_factor{};
|
||||
|
||||
u64 rtsc_frequency;
|
||||
|
||||
std::jthread time_sync_thread;
|
||||
u64 ns_rdtsc_factor;
|
||||
u64 us_rdtsc_factor;
|
||||
u64 ms_rdtsc_factor;
|
||||
u64 cntpct_rdtsc_factor;
|
||||
u64 gputick_rdtsc_factor;
|
||||
};
|
||||
} // namespace X64
|
||||
|
||||
u64 EstimateRDTSCFrequency();
|
||||
|
||||
} // namespace Common
|
||||
} // namespace Common::X64
|
||||
|
||||
@@ -0,0 +1,39 @@
|
||||
// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
|
||||
// SPDX-License-Identifier: GPL-2.0-or-later
|
||||
|
||||
#include <thread>
|
||||
|
||||
#include "common/steady_clock.h"
|
||||
#include "common/uint128.h"
|
||||
#include "common/x64/rdtsc.h"
|
||||
|
||||
namespace Common::X64 {
|
||||
|
||||
template <u64 Nearest>
|
||||
static u64 RoundToNearest(u64 value) {
|
||||
const auto mod = value % Nearest;
|
||||
return mod >= (Nearest / 2) ? (value - mod + Nearest) : (value - mod);
|
||||
}
|
||||
|
||||
u64 EstimateRDTSCFrequency() {
|
||||
// Discard the first result measuring the rdtsc.
|
||||
FencedRDTSC();
|
||||
std::this_thread::sleep_for(std::chrono::milliseconds{1});
|
||||
FencedRDTSC();
|
||||
|
||||
// Get the current time.
|
||||
const auto start_time = RealTimeClock::Now();
|
||||
const u64 tsc_start = FencedRDTSC();
|
||||
// Wait for 100 milliseconds.
|
||||
std::this_thread::sleep_for(std::chrono::milliseconds{100});
|
||||
const auto end_time = RealTimeClock::Now();
|
||||
const u64 tsc_end = FencedRDTSC();
|
||||
// Calculate differences.
|
||||
const u64 timer_diff = static_cast<u64>(
|
||||
std::chrono::duration_cast<std::chrono::nanoseconds>(end_time - start_time).count());
|
||||
const u64 tsc_diff = tsc_end - tsc_start;
|
||||
const u64 tsc_freq = MultiplyAndDivide64(tsc_diff, 1000000000ULL, timer_diff);
|
||||
return RoundToNearest<100'000>(tsc_freq);
|
||||
}
|
||||
|
||||
} // namespace Common::X64
|
||||
@@ -0,0 +1,37 @@
|
||||
// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
|
||||
// SPDX-License-Identifier: GPL-2.0-or-later
|
||||
|
||||
#pragma once
|
||||
|
||||
#ifdef _MSC_VER
|
||||
#include <intrin.h>
|
||||
#endif
|
||||
|
||||
#include "common/common_types.h"
|
||||
|
||||
namespace Common::X64 {
|
||||
|
||||
#ifdef _MSC_VER
|
||||
__forceinline static u64 FencedRDTSC() {
|
||||
_mm_lfence();
|
||||
_ReadWriteBarrier();
|
||||
const u64 result = __rdtsc();
|
||||
_mm_lfence();
|
||||
_ReadWriteBarrier();
|
||||
return result;
|
||||
}
|
||||
#else
|
||||
static inline u64 FencedRDTSC() {
|
||||
u64 eax;
|
||||
u64 edx;
|
||||
asm volatile("lfence\n\t"
|
||||
"rdtsc\n\t"
|
||||
"lfence\n\t"
|
||||
: "=a"(eax), "=d"(edx));
|
||||
return (edx << 32) | eax;
|
||||
}
|
||||
#endif
|
||||
|
||||
u64 EstimateRDTSCFrequency();
|
||||
|
||||
} // namespace Common::X64
|
||||
@@ -14,7 +14,6 @@ add_library(core STATIC
|
||||
core.h
|
||||
core_timing.cpp
|
||||
core_timing.h
|
||||
core_timing_util.h
|
||||
cpu_manager.cpp
|
||||
cpu_manager.h
|
||||
crypto/aes_util.cpp
|
||||
|
||||
+19
-39
@@ -16,12 +16,11 @@
|
||||
|
||||
#include "common/microprofile.h"
|
||||
#include "core/core_timing.h"
|
||||
#include "core/core_timing_util.h"
|
||||
#include "core/hardware_properties.h"
|
||||
|
||||
namespace Core::Timing {
|
||||
|
||||
constexpr s64 MAX_SLICE_LENGTH = 4000;
|
||||
constexpr s64 MAX_SLICE_LENGTH = 10000;
|
||||
|
||||
std::shared_ptr<EventType> CreateEvent(std::string name, TimedCallback&& callback) {
|
||||
return std::make_shared<EventType>(std::move(callback), std::move(name));
|
||||
@@ -45,9 +44,7 @@ struct CoreTiming::Event {
|
||||
}
|
||||
};
|
||||
|
||||
CoreTiming::CoreTiming()
|
||||
: cpu_clock{Common::CreateBestMatchingClock(Hardware::BASE_CLOCK_RATE, Hardware::CNTFREQ)},
|
||||
event_clock{Common::CreateStandardWallClock(Hardware::BASE_CLOCK_RATE, Hardware::CNTFREQ)} {}
|
||||
CoreTiming::CoreTiming() : clock{Common::CreateOptimalClock()} {}
|
||||
|
||||
CoreTiming::~CoreTiming() {
|
||||
Reset();
|
||||
@@ -68,7 +65,7 @@ void CoreTiming::Initialize(std::function<void()>&& on_thread_init_) {
|
||||
on_thread_init = std::move(on_thread_init_);
|
||||
event_fifo_id = 0;
|
||||
shutting_down = false;
|
||||
ticks = 0;
|
||||
cpu_ticks = 0;
|
||||
const auto empty_timed_callback = [](std::uintptr_t, u64, std::chrono::nanoseconds)
|
||||
-> std::optional<std::chrono::nanoseconds> { return std::nullopt; };
|
||||
ev_lost = CreateEvent("_lost_event", empty_timed_callback);
|
||||
@@ -173,38 +170,30 @@ void CoreTiming::UnscheduleEvent(const std::shared_ptr<EventType>& event_type,
|
||||
}
|
||||
|
||||
void CoreTiming::AddTicks(u64 ticks_to_add) {
|
||||
ticks += ticks_to_add;
|
||||
downcount -= static_cast<s64>(ticks);
|
||||
cpu_ticks += ticks_to_add;
|
||||
downcount -= static_cast<s64>(cpu_ticks);
|
||||
}
|
||||
|
||||
void CoreTiming::Idle() {
|
||||
if (!event_queue.empty()) {
|
||||
const u64 next_event_time = event_queue.front().time;
|
||||
const u64 next_ticks = nsToCycles(std::chrono::nanoseconds(next_event_time)) + 10U;
|
||||
if (next_ticks > ticks) {
|
||||
ticks = next_ticks;
|
||||
}
|
||||
return;
|
||||
}
|
||||
ticks += 1000U;
|
||||
cpu_ticks += 1000U;
|
||||
}
|
||||
|
||||
void CoreTiming::ResetTicks() {
|
||||
downcount = MAX_SLICE_LENGTH;
|
||||
}
|
||||
|
||||
u64 CoreTiming::GetCPUTicks() const {
|
||||
if (is_multicore) [[likely]] {
|
||||
return cpu_clock->GetCPUCycles();
|
||||
}
|
||||
return ticks;
|
||||
}
|
||||
|
||||
u64 CoreTiming::GetClockTicks() const {
|
||||
if (is_multicore) [[likely]] {
|
||||
return cpu_clock->GetClockCycles();
|
||||
return clock->GetCNTPCT();
|
||||
}
|
||||
return CpuCyclesToClockCycles(ticks);
|
||||
return Common::WallClock::CPUTickToCNTPCT(cpu_ticks);
|
||||
}
|
||||
|
||||
u64 CoreTiming::GetGPUTicks() const {
|
||||
if (is_multicore) [[likely]] {
|
||||
return clock->GetGPUTick();
|
||||
}
|
||||
return Common::WallClock::CPUTickToGPUTick(cpu_ticks);
|
||||
}
|
||||
|
||||
std::optional<s64> CoreTiming::Advance() {
|
||||
@@ -297,9 +286,7 @@ void CoreTiming::ThreadLoop() {
|
||||
}
|
||||
|
||||
paused_set = true;
|
||||
event_clock->Pause(true);
|
||||
pause_event.Wait();
|
||||
event_clock->Pause(false);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -315,25 +302,18 @@ void CoreTiming::Reset() {
|
||||
has_started = false;
|
||||
}
|
||||
|
||||
std::chrono::nanoseconds CoreTiming::GetCPUTimeNs() const {
|
||||
if (is_multicore) [[likely]] {
|
||||
return cpu_clock->GetTimeNS();
|
||||
}
|
||||
return CyclesToNs(ticks);
|
||||
}
|
||||
|
||||
std::chrono::nanoseconds CoreTiming::GetGlobalTimeNs() const {
|
||||
if (is_multicore) [[likely]] {
|
||||
return event_clock->GetTimeNS();
|
||||
return clock->GetTimeNS();
|
||||
}
|
||||
return CyclesToNs(ticks);
|
||||
return std::chrono::nanoseconds{Common::WallClock::CPUTickToNS(cpu_ticks)};
|
||||
}
|
||||
|
||||
std::chrono::microseconds CoreTiming::GetGlobalTimeUs() const {
|
||||
if (is_multicore) [[likely]] {
|
||||
return event_clock->GetTimeUS();
|
||||
return clock->GetTimeUS();
|
||||
}
|
||||
return CyclesToUs(ticks);
|
||||
return std::chrono::microseconds{Common::WallClock::CPUTickToUS(cpu_ticks)};
|
||||
}
|
||||
|
||||
} // namespace Core::Timing
|
||||
|
||||
@@ -116,14 +116,11 @@ public:
|
||||
return downcount;
|
||||
}
|
||||
|
||||
/// Returns current time in emulated CPU cycles
|
||||
u64 GetCPUTicks() const;
|
||||
|
||||
/// Returns current time in emulated in Clock cycles
|
||||
/// Returns the current CNTPCT tick value.
|
||||
u64 GetClockTicks() const;
|
||||
|
||||
/// Returns current time in nanoseconds.
|
||||
std::chrono::nanoseconds GetCPUTimeNs() const;
|
||||
/// Returns the current GPU tick value.
|
||||
u64 GetGPUTicks() const;
|
||||
|
||||
/// Returns current time in microseconds.
|
||||
std::chrono::microseconds GetGlobalTimeUs() const;
|
||||
@@ -142,8 +139,7 @@ private:
|
||||
|
||||
void Reset();
|
||||
|
||||
std::unique_ptr<Common::WallClock> cpu_clock;
|
||||
std::unique_ptr<Common::WallClock> event_clock;
|
||||
std::unique_ptr<Common::WallClock> clock;
|
||||
|
||||
s64 global_timer = 0;
|
||||
|
||||
@@ -171,7 +167,7 @@ private:
|
||||
s64 pause_end_time{};
|
||||
|
||||
/// Cycle timing
|
||||
u64 ticks{};
|
||||
u64 cpu_ticks{};
|
||||
s64 downcount{};
|
||||
};
|
||||
|
||||
|
||||
@@ -1,58 +0,0 @@
|
||||
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
|
||||
// SPDX-License-Identifier: GPL-2.0-or-later
|
||||
|
||||
#pragma once
|
||||
|
||||
#include <chrono>
|
||||
|
||||
#include "common/common_types.h"
|
||||
#include "core/hardware_properties.h"
|
||||
|
||||
namespace Core::Timing {
|
||||
|
||||
namespace detail {
|
||||
constexpr u64 CNTFREQ_ADJUSTED = Hardware::CNTFREQ / 1000;
|
||||
constexpr u64 BASE_CLOCK_RATE_ADJUSTED = Hardware::BASE_CLOCK_RATE / 1000;
|
||||
} // namespace detail
|
||||
|
||||
[[nodiscard]] constexpr s64 msToCycles(std::chrono::milliseconds ms) {
|
||||
return ms.count() * detail::BASE_CLOCK_RATE_ADJUSTED;
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr s64 usToCycles(std::chrono::microseconds us) {
|
||||
return us.count() * detail::BASE_CLOCK_RATE_ADJUSTED / 1000;
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr s64 nsToCycles(std::chrono::nanoseconds ns) {
|
||||
return ns.count() * detail::BASE_CLOCK_RATE_ADJUSTED / 1000000;
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr u64 msToClockCycles(std::chrono::milliseconds ms) {
|
||||
return static_cast<u64>(ms.count()) * detail::CNTFREQ_ADJUSTED;
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr u64 usToClockCycles(std::chrono::microseconds us) {
|
||||
return us.count() * detail::CNTFREQ_ADJUSTED / 1000;
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr u64 nsToClockCycles(std::chrono::nanoseconds ns) {
|
||||
return ns.count() * detail::CNTFREQ_ADJUSTED / 1000000;
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr u64 CpuCyclesToClockCycles(u64 ticks) {
|
||||
return ticks * detail::CNTFREQ_ADJUSTED / detail::BASE_CLOCK_RATE_ADJUSTED;
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr std::chrono::milliseconds CyclesToMs(s64 cycles) {
|
||||
return std::chrono::milliseconds(cycles / detail::BASE_CLOCK_RATE_ADJUSTED);
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr std::chrono::nanoseconds CyclesToNs(s64 cycles) {
|
||||
return std::chrono::nanoseconds(cycles * 1000000 / detail::BASE_CLOCK_RATE_ADJUSTED);
|
||||
}
|
||||
|
||||
[[nodiscard]] constexpr std::chrono::microseconds CyclesToUs(s64 cycles) {
|
||||
return std::chrono::microseconds(cycles * 1000 / detail::BASE_CLOCK_RATE_ADJUSTED);
|
||||
}
|
||||
|
||||
} // namespace Core::Timing
|
||||
@@ -150,23 +150,29 @@ std::size_t ConcatenatedVfsFile::Read(u8* data, std::size_t length, std::size_t
|
||||
while (cur_length > 0 && it != concatenation_map.end()) {
|
||||
// Check if we can read the file at this position.
|
||||
const auto& file = it->file;
|
||||
const u64 file_offset = it->offset;
|
||||
const u64 map_offset = it->offset;
|
||||
const u64 file_size = file->GetSize();
|
||||
|
||||
if (cur_offset >= file_offset + file_size) {
|
||||
if (cur_offset > map_offset + file_size) {
|
||||
// Entirely out of bounds read.
|
||||
break;
|
||||
}
|
||||
|
||||
// Read the file at this position.
|
||||
const u64 intended_read_size = std::min<u64>(cur_length, file_size);
|
||||
const u64 file_seek = cur_offset - map_offset;
|
||||
const u64 intended_read_size = std::min<u64>(cur_length, file_size - file_seek);
|
||||
const u64 actual_read_size =
|
||||
file->Read(data + (cur_offset - offset), intended_read_size, cur_offset - file_offset);
|
||||
file->Read(data + (cur_offset - offset), intended_read_size, file_seek);
|
||||
|
||||
// Update tracking.
|
||||
cur_offset += actual_read_size;
|
||||
cur_length -= actual_read_size;
|
||||
it++;
|
||||
|
||||
// If we encountered a short read, we're done.
|
||||
if (actual_read_size < intended_read_size) {
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
return cur_offset - offset;
|
||||
|
||||
@@ -184,7 +184,8 @@ u64 KScheduler::UpdateHighestPriorityThread(KThread* highest_thread) {
|
||||
prev_highest_thread != highest_thread) [[likely]] {
|
||||
if (prev_highest_thread != nullptr) [[likely]] {
|
||||
IncrementScheduledCount(prev_highest_thread);
|
||||
prev_highest_thread->SetLastScheduledTick(m_kernel.System().CoreTiming().GetCPUTicks());
|
||||
prev_highest_thread->SetLastScheduledTick(
|
||||
m_kernel.System().CoreTiming().GetClockTicks());
|
||||
}
|
||||
if (m_state.should_count_idle) {
|
||||
if (highest_thread != nullptr) [[likely]] {
|
||||
@@ -351,7 +352,7 @@ void KScheduler::SwitchThread(KThread* next_thread) {
|
||||
|
||||
// Update the CPU time tracking variables.
|
||||
const s64 prev_tick = m_last_context_switch_time;
|
||||
const s64 cur_tick = m_kernel.System().CoreTiming().GetCPUTicks();
|
||||
const s64 cur_tick = m_kernel.System().CoreTiming().GetClockTicks();
|
||||
const s64 tick_diff = cur_tick - prev_tick;
|
||||
cur_thread->AddCpuTime(m_core_id, tick_diff);
|
||||
if (cur_process != nullptr) {
|
||||
|
||||
@@ -199,9 +199,9 @@ Result GetInfo(Core::System& system, u64* result, InfoType info_id_type, Handle
|
||||
if (same_thread && info_sub_id == 0xFFFFFFFFFFFFFFFF) {
|
||||
const u64 thread_ticks = current_thread->GetCpuTime();
|
||||
|
||||
out_ticks = thread_ticks + (core_timing.GetCPUTicks() - prev_ctx_ticks);
|
||||
out_ticks = thread_ticks + (core_timing.GetClockTicks() - prev_ctx_ticks);
|
||||
} else if (same_thread && info_sub_id == system.Kernel().CurrentPhysicalCoreIndex()) {
|
||||
out_ticks = core_timing.GetCPUTicks() - prev_ctx_ticks;
|
||||
out_ticks = core_timing.GetClockTicks() - prev_ctx_ticks;
|
||||
}
|
||||
|
||||
*result = out_ticks;
|
||||
|
||||
@@ -12,16 +12,8 @@ namespace Kernel::Svc {
|
||||
int64_t GetSystemTick(Core::System& system) {
|
||||
LOG_TRACE(Kernel_SVC, "called");
|
||||
|
||||
auto& core_timing = system.CoreTiming();
|
||||
|
||||
// Returns the value of cntpct_el0 (https://switchbrew.org/wiki/SVC#svcGetSystemTick)
|
||||
const u64 result{core_timing.GetClockTicks()};
|
||||
|
||||
if (!system.Kernel().IsMulticore()) {
|
||||
core_timing.AddTicks(400U);
|
||||
}
|
||||
|
||||
return static_cast<int64_t>(result);
|
||||
return static_cast<int64_t>(system.CoreTiming().GetClockTicks());
|
||||
}
|
||||
|
||||
int64_t GetSystemTick64(Core::System& system) {
|
||||
|
||||
@@ -5,7 +5,6 @@
|
||||
#include "common/settings.h"
|
||||
#include "core/core.h"
|
||||
#include "core/core_timing.h"
|
||||
#include "core/core_timing_util.h"
|
||||
#include "core/hid/hid_types.h"
|
||||
#include "core/hle/kernel/k_event.h"
|
||||
#include "core/hle/kernel/k_readable_event.h"
|
||||
|
||||
@@ -51,8 +51,8 @@ void nvdisp_disp0::flip(u32 buffer_handle, u32 offset, android::PixelFormat form
|
||||
stride, format, transform, crop_rect};
|
||||
|
||||
system.GPU().RequestSwapBuffers(&framebuffer, fences, num_fences);
|
||||
system.GetPerfStats().EndSystemFrame();
|
||||
system.SpeedLimiter().DoSpeedLimiting(system.CoreTiming().GetGlobalTimeUs());
|
||||
system.GetPerfStats().EndSystemFrame();
|
||||
system.GetPerfStats().BeginSystemFrame();
|
||||
}
|
||||
|
||||
|
||||
@@ -70,7 +70,8 @@ Nvnflinger::Nvnflinger(Core::System& system_, HosBinderDriverServer& hos_binder_
|
||||
[this](std::uintptr_t, s64 time,
|
||||
std::chrono::nanoseconds ns_late) -> std::optional<std::chrono::nanoseconds> {
|
||||
vsync_signal.store(true);
|
||||
vsync_signal.notify_all();
|
||||
{ const auto lock_guard = Lock(); }
|
||||
vsync_signal.notify_one();
|
||||
return std::chrono::nanoseconds(GetNextTicks());
|
||||
});
|
||||
|
||||
|
||||
@@ -3,6 +3,8 @@
|
||||
|
||||
#pragma once
|
||||
|
||||
#include <ratio>
|
||||
|
||||
#include "common/common_funcs.h"
|
||||
#include "common/common_types.h"
|
||||
#include "common/uuid.h"
|
||||
@@ -74,18 +76,19 @@ static_assert(std::is_trivially_copyable_v<ContinuousAdjustmentTimePoint>,
|
||||
/// https://switchbrew.org/wiki/Glue_services#TimeSpanType
|
||||
struct TimeSpanType {
|
||||
s64 nanoseconds{};
|
||||
static constexpr s64 ns_per_second{1000000000ULL};
|
||||
|
||||
s64 ToSeconds() const {
|
||||
return nanoseconds / ns_per_second;
|
||||
return nanoseconds / std::nano::den;
|
||||
}
|
||||
|
||||
static TimeSpanType FromSeconds(s64 seconds) {
|
||||
return {seconds * ns_per_second};
|
||||
return {seconds * std::nano::den};
|
||||
}
|
||||
|
||||
static TimeSpanType FromTicks(u64 ticks, u64 frequency) {
|
||||
return FromSeconds(static_cast<s64>(ticks) / static_cast<s64>(frequency));
|
||||
template <u64 Frequency>
|
||||
static TimeSpanType FromTicks(u64 ticks) {
|
||||
using TicksToNSRatio = std::ratio<std::nano::den, Frequency>;
|
||||
return {static_cast<s64>(ticks * TicksToNSRatio::num / TicksToNSRatio::den)};
|
||||
}
|
||||
};
|
||||
static_assert(sizeof(TimeSpanType) == 8, "TimeSpanType is incorrect size");
|
||||
|
||||
@@ -10,7 +10,7 @@ namespace Service::Time::Clock {
|
||||
|
||||
TimeSpanType StandardSteadyClockCore::GetCurrentRawTimePoint(Core::System& system) {
|
||||
const TimeSpanType ticks_time_span{
|
||||
TimeSpanType::FromTicks(system.CoreTiming().GetClockTicks(), Core::Hardware::CNTFREQ)};
|
||||
TimeSpanType::FromTicks<Core::Hardware::CNTFREQ>(system.CoreTiming().GetClockTicks())};
|
||||
TimeSpanType raw_time_point{setup_value.nanoseconds + ticks_time_span.nanoseconds};
|
||||
|
||||
if (raw_time_point.nanoseconds < cached_raw_time_point.nanoseconds) {
|
||||
|
||||
@@ -10,7 +10,7 @@ namespace Service::Time::Clock {
|
||||
|
||||
SteadyClockTimePoint TickBasedSteadyClockCore::GetTimePoint(Core::System& system) {
|
||||
const TimeSpanType ticks_time_span{
|
||||
TimeSpanType::FromTicks(system.CoreTiming().GetClockTicks(), Core::Hardware::CNTFREQ)};
|
||||
TimeSpanType::FromTicks<Core::Hardware::CNTFREQ>(system.CoreTiming().GetClockTicks())};
|
||||
|
||||
return {ticks_time_span.ToSeconds(), GetClockSourceId()};
|
||||
}
|
||||
|
||||
@@ -240,8 +240,8 @@ void Module::Interface::CalculateMonotonicSystemClockBaseTimePoint(HLERequestCon
|
||||
const auto current_time_point{steady_clock_core.GetCurrentTimePoint(system)};
|
||||
|
||||
if (current_time_point.clock_source_id == context.steady_time_point.clock_source_id) {
|
||||
const auto ticks{Clock::TimeSpanType::FromTicks(system.CoreTiming().GetClockTicks(),
|
||||
Core::Hardware::CNTFREQ)};
|
||||
const auto ticks{Clock::TimeSpanType::FromTicks<Core::Hardware::CNTFREQ>(
|
||||
system.CoreTiming().GetClockTicks())};
|
||||
const s64 base_time_point{context.offset + current_time_point.time_point -
|
||||
ticks.ToSeconds()};
|
||||
IPC::ResponseBuilder rb{ctx, (sizeof(s64) / 4) + 2};
|
||||
|
||||
@@ -21,8 +21,9 @@ SharedMemory::~SharedMemory() = default;
|
||||
|
||||
void SharedMemory::SetupStandardSteadyClock(const Common::UUID& clock_source_id,
|
||||
Clock::TimeSpanType current_time_point) {
|
||||
const Clock::TimeSpanType ticks_time_span{Clock::TimeSpanType::FromTicks(
|
||||
system.CoreTiming().GetClockTicks(), Core::Hardware::CNTFREQ)};
|
||||
const Clock::TimeSpanType ticks_time_span{
|
||||
Clock::TimeSpanType::FromTicks<Core::Hardware::CNTFREQ>(
|
||||
system.CoreTiming().GetClockTicks())};
|
||||
const Clock::SteadyClockContext context{
|
||||
static_cast<u64>(current_time_point.nanoseconds - ticks_time_span.nanoseconds),
|
||||
clock_source_id};
|
||||
|
||||
@@ -911,9 +911,13 @@ static Result ToCalendarTimeInternal(const TimeZoneRule& rules, s64 time,
|
||||
|
||||
calendar_additional_info.is_dst = rules.ttis[tti_index].is_dst;
|
||||
const char* time_zone{&rules.chars[rules.ttis[tti_index].abbreviation_list_index]};
|
||||
for (int index{}; time_zone[index] != '\0'; ++index) {
|
||||
u32 index;
|
||||
for (index = 0; time_zone[index] != '\0' && time_zone[index] != ',' &&
|
||||
index < calendar_additional_info.timezone_name.size() - 1;
|
||||
++index) {
|
||||
calendar_additional_info.timezone_name[index] = time_zone[index];
|
||||
}
|
||||
calendar_additional_info.timezone_name[index] = '\0';
|
||||
return ResultSuccess;
|
||||
}
|
||||
|
||||
|
||||
@@ -193,18 +193,13 @@ struct GPU::Impl {
|
||||
}
|
||||
|
||||
[[nodiscard]] u64 GetTicks() const {
|
||||
// This values were reversed engineered by fincs from NVN
|
||||
// The gpu clock is reported in units of 385/625 nanoseconds
|
||||
constexpr u64 gpu_ticks_num = 384;
|
||||
constexpr u64 gpu_ticks_den = 625;
|
||||
u64 gpu_tick = system.CoreTiming().GetGPUTicks();
|
||||
|
||||
u64 nanoseconds = system.CoreTiming().GetCPUTimeNs().count();
|
||||
if (Settings::values.use_fast_gpu_time.GetValue()) {
|
||||
nanoseconds /= 256;
|
||||
gpu_tick /= 256;
|
||||
}
|
||||
const u64 nanoseconds_num = nanoseconds / gpu_ticks_den;
|
||||
const u64 nanoseconds_rem = nanoseconds % gpu_ticks_den;
|
||||
return nanoseconds_num * gpu_ticks_num + (nanoseconds_rem * gpu_ticks_num) / gpu_ticks_den;
|
||||
|
||||
return gpu_tick;
|
||||
}
|
||||
|
||||
[[nodiscard]] bool IsAsync() const {
|
||||
|
||||
@@ -111,7 +111,7 @@ GPUVAddr MemoryManager::PageTableOp(GPUVAddr gpu_addr, [[maybe_unused]] VAddr cp
|
||||
[[maybe_unused]] const auto current_entry_type = GetEntry<false>(current_gpu_addr);
|
||||
SetEntry<false>(current_gpu_addr, entry_type);
|
||||
if (current_entry_type != entry_type) {
|
||||
rasterizer->ModifyGPUMemory(unique_identifier, gpu_addr, page_size);
|
||||
rasterizer->ModifyGPUMemory(unique_identifier, current_gpu_addr, page_size);
|
||||
}
|
||||
if constexpr (entry_type == EntryType::Mapped) {
|
||||
const VAddr current_cpu_addr = cpu_addr + offset;
|
||||
@@ -134,7 +134,7 @@ GPUVAddr MemoryManager::BigPageTableOp(GPUVAddr gpu_addr, [[maybe_unused]] VAddr
|
||||
[[maybe_unused]] const auto current_entry_type = GetEntry<true>(current_gpu_addr);
|
||||
SetEntry<true>(current_gpu_addr, entry_type);
|
||||
if (current_entry_type != entry_type) {
|
||||
rasterizer->ModifyGPUMemory(unique_identifier, gpu_addr, big_page_size);
|
||||
rasterizer->ModifyGPUMemory(unique_identifier, current_gpu_addr, big_page_size);
|
||||
}
|
||||
if constexpr (entry_type == EntryType::Mapped) {
|
||||
const VAddr current_cpu_addr = cpu_addr + offset;
|
||||
|
||||
@@ -705,10 +705,7 @@ std::unique_ptr<ComputePipeline> PipelineCache::CreateComputePipeline(
|
||||
std::unique_ptr<ComputePipeline> PipelineCache::CreateComputePipeline(
|
||||
ShaderPools& pools, const ComputePipelineCacheKey& key, Shader::Environment& env,
|
||||
PipelineStatistics* statistics, bool build_in_parallel) try {
|
||||
// TODO: Remove this when Intel fixes their shader compiler.
|
||||
// https://github.com/IGCIT/Intel-GPU-Community-Issue-Tracker-IGCIT/issues/159
|
||||
if (device.GetDriverID() == VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS &&
|
||||
!Settings::values.enable_compute_pipelines.GetValue()) {
|
||||
if (device.HasBrokenCompute()) {
|
||||
LOG_ERROR(Render_Vulkan, "Skipping 0x{:016x}", key.Hash());
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
@@ -186,6 +186,10 @@ void TextureCache<P>::FillComputeImageViews(std::span<ImageViewInOut> views) {
|
||||
|
||||
template <class P>
|
||||
void TextureCache<P>::CheckFeedbackLoop(std::span<const ImageViewInOut> views) {
|
||||
if (!Settings::values.barrier_feedback_loops.GetValue()) {
|
||||
return;
|
||||
}
|
||||
|
||||
const bool requires_barrier = [&] {
|
||||
for (const auto& view : views) {
|
||||
if (!view.id) {
|
||||
@@ -300,7 +304,7 @@ void TextureCache<P>::SynchronizeComputeDescriptors() {
|
||||
}
|
||||
|
||||
template <class P>
|
||||
bool TextureCache<P>::RescaleRenderTargets(bool is_clear) {
|
||||
bool TextureCache<P>::RescaleRenderTargets() {
|
||||
auto& flags = maxwell3d->dirty.flags;
|
||||
u32 scale_rating = 0;
|
||||
bool rescaled = false;
|
||||
@@ -338,13 +342,13 @@ bool TextureCache<P>::RescaleRenderTargets(bool is_clear) {
|
||||
ImageViewId& color_buffer_id = render_targets.color_buffer_ids[index];
|
||||
if (flags[Dirty::ColorBuffer0 + index] || force) {
|
||||
flags[Dirty::ColorBuffer0 + index] = false;
|
||||
BindRenderTarget(&color_buffer_id, FindColorBuffer(index, is_clear));
|
||||
BindRenderTarget(&color_buffer_id, FindColorBuffer(index));
|
||||
}
|
||||
check_rescale(color_buffer_id, tmp_color_images[index]);
|
||||
}
|
||||
if (flags[Dirty::ZetaBuffer] || force) {
|
||||
flags[Dirty::ZetaBuffer] = false;
|
||||
BindRenderTarget(&render_targets.depth_buffer_id, FindDepthBuffer(is_clear));
|
||||
BindRenderTarget(&render_targets.depth_buffer_id, FindDepthBuffer());
|
||||
}
|
||||
check_rescale(render_targets.depth_buffer_id, tmp_depth_image);
|
||||
|
||||
@@ -409,7 +413,7 @@ void TextureCache<P>::UpdateRenderTargets(bool is_clear) {
|
||||
return;
|
||||
}
|
||||
|
||||
const bool rescaled = RescaleRenderTargets(is_clear);
|
||||
const bool rescaled = RescaleRenderTargets();
|
||||
if (is_rescaling != rescaled) {
|
||||
flags[Dirty::RescaleViewports] = true;
|
||||
flags[Dirty::RescaleScissors] = true;
|
||||
@@ -1678,7 +1682,7 @@ SamplerId TextureCache<P>::FindSampler(const TSCEntry& config) {
|
||||
}
|
||||
|
||||
template <class P>
|
||||
ImageViewId TextureCache<P>::FindColorBuffer(size_t index, bool is_clear) {
|
||||
ImageViewId TextureCache<P>::FindColorBuffer(size_t index) {
|
||||
const auto& regs = maxwell3d->regs;
|
||||
if (index >= regs.rt_control.count) {
|
||||
return ImageViewId{};
|
||||
@@ -1692,11 +1696,11 @@ ImageViewId TextureCache<P>::FindColorBuffer(size_t index, bool is_clear) {
|
||||
return ImageViewId{};
|
||||
}
|
||||
const ImageInfo info(regs.rt[index], regs.anti_alias_samples_mode);
|
||||
return FindRenderTargetView(info, gpu_addr, is_clear);
|
||||
return FindRenderTargetView(info, gpu_addr);
|
||||
}
|
||||
|
||||
template <class P>
|
||||
ImageViewId TextureCache<P>::FindDepthBuffer(bool is_clear) {
|
||||
ImageViewId TextureCache<P>::FindDepthBuffer() {
|
||||
const auto& regs = maxwell3d->regs;
|
||||
if (!regs.zeta_enable) {
|
||||
return ImageViewId{};
|
||||
@@ -1706,18 +1710,16 @@ ImageViewId TextureCache<P>::FindDepthBuffer(bool is_clear) {
|
||||
return ImageViewId{};
|
||||
}
|
||||
const ImageInfo info(regs.zeta, regs.zeta_size, regs.anti_alias_samples_mode);
|
||||
return FindRenderTargetView(info, gpu_addr, is_clear);
|
||||
return FindRenderTargetView(info, gpu_addr);
|
||||
}
|
||||
|
||||
template <class P>
|
||||
ImageViewId TextureCache<P>::FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr,
|
||||
bool is_clear) {
|
||||
const auto options = is_clear ? RelaxedOptions::Samples : RelaxedOptions{};
|
||||
ImageViewId TextureCache<P>::FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr) {
|
||||
ImageId image_id{};
|
||||
bool delete_state = has_deleted_images;
|
||||
do {
|
||||
has_deleted_images = false;
|
||||
image_id = FindOrInsertImage(info, gpu_addr, options);
|
||||
image_id = FindOrInsertImage(info, gpu_addr);
|
||||
delete_state |= has_deleted_images;
|
||||
} while (has_deleted_images);
|
||||
has_deleted_images = delete_state;
|
||||
|
||||
@@ -178,9 +178,8 @@ public:
|
||||
void SynchronizeComputeDescriptors();
|
||||
|
||||
/// Updates the Render Targets if they can be rescaled
|
||||
/// @param is_clear True when the render targets are being used for clears
|
||||
/// @retval True if the Render Targets have been rescaled.
|
||||
bool RescaleRenderTargets(bool is_clear);
|
||||
bool RescaleRenderTargets();
|
||||
|
||||
/// Update bound render targets and upload memory if necessary
|
||||
/// @param is_clear True when the render targets are being used for clears
|
||||
@@ -336,14 +335,13 @@ private:
|
||||
[[nodiscard]] SamplerId FindSampler(const TSCEntry& config);
|
||||
|
||||
/// Find or create an image view for the given color buffer index
|
||||
[[nodiscard]] ImageViewId FindColorBuffer(size_t index, bool is_clear);
|
||||
[[nodiscard]] ImageViewId FindColorBuffer(size_t index);
|
||||
|
||||
/// Find or create an image view for the depth buffer
|
||||
[[nodiscard]] ImageViewId FindDepthBuffer(bool is_clear);
|
||||
[[nodiscard]] ImageViewId FindDepthBuffer();
|
||||
|
||||
/// Find or create a view for a render target with the given image parameters
|
||||
[[nodiscard]] ImageViewId FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr,
|
||||
bool is_clear);
|
||||
[[nodiscard]] ImageViewId FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr);
|
||||
|
||||
/// Iterates over all the images in a region calling func
|
||||
template <typename Func>
|
||||
|
||||
@@ -562,6 +562,9 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR
|
||||
LOG_WARNING(Render_Vulkan, "Intel proprietary drivers do not support MSAA image blits");
|
||||
cant_blit_msaa = true;
|
||||
}
|
||||
has_broken_compute =
|
||||
CheckBrokenCompute(properties.driver.driverID, properties.properties.driverVersion) &&
|
||||
!Settings::values.enable_compute_pipelines.GetValue();
|
||||
if (is_intel_anv || (is_qualcomm && !is_s8gen2)) {
|
||||
LOG_WARNING(Render_Vulkan, "Driver does not support native BGR format");
|
||||
must_emulate_bgr565 = true;
|
||||
|
||||
@@ -10,6 +10,7 @@
|
||||
#include <vector>
|
||||
|
||||
#include "common/common_types.h"
|
||||
#include "common/logging/log.h"
|
||||
#include "common/settings.h"
|
||||
#include "video_core/vulkan_common/vulkan_wrapper.h"
|
||||
|
||||
@@ -518,6 +519,11 @@ public:
|
||||
return has_renderdoc || has_nsight_graphics || Settings::values.renderer_debug.GetValue();
|
||||
}
|
||||
|
||||
/// @returns True if compute pipelines can cause crashing.
|
||||
bool HasBrokenCompute() const {
|
||||
return has_broken_compute;
|
||||
}
|
||||
|
||||
/// Returns true when the device does not properly support cube compatibility.
|
||||
bool HasBrokenCubeImageCompability() const {
|
||||
return has_broken_cube_compatibility;
|
||||
@@ -579,6 +585,22 @@ public:
|
||||
return supports_conditional_barriers;
|
||||
}
|
||||
|
||||
[[nodiscard]] static constexpr bool CheckBrokenCompute(VkDriverId driver_id,
|
||||
u32 driver_version) {
|
||||
if (driver_id == VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS) {
|
||||
const u32 major = VK_API_VERSION_MAJOR(driver_version);
|
||||
const u32 minor = VK_API_VERSION_MINOR(driver_version);
|
||||
const u32 patch = VK_API_VERSION_PATCH(driver_version);
|
||||
if (major == 0 && minor == 405 && patch < 286) {
|
||||
LOG_WARNING(
|
||||
Render_Vulkan,
|
||||
"Intel proprietary drivers 0.405.0 until 0.405.286 have broken compute");
|
||||
return true;
|
||||
}
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
private:
|
||||
/// Checks if the physical device is suitable and configures the object state
|
||||
/// with all necessary info about its properties.
|
||||
@@ -672,6 +694,7 @@ private:
|
||||
bool is_integrated{}; ///< Is GPU an iGPU.
|
||||
bool is_virtual{}; ///< Is GPU a virtual GPU.
|
||||
bool is_non_gpu{}; ///< Is SoftwareRasterizer, FPGA, non-GPU device.
|
||||
bool has_broken_compute{}; ///< Compute shaders can cause crashes
|
||||
bool has_broken_cube_compatibility{}; ///< Has broken cube compatibility bit
|
||||
bool has_renderdoc{}; ///< Has RenderDoc attached
|
||||
bool has_nsight_graphics{}; ///< Has Nsight Graphics attached
|
||||
|
||||
@@ -761,6 +761,7 @@ void Config::ReadRendererValues() {
|
||||
ReadGlobalSetting(Settings::values.use_vulkan_driver_pipeline_cache);
|
||||
ReadGlobalSetting(Settings::values.enable_compute_pipelines);
|
||||
ReadGlobalSetting(Settings::values.use_video_framerate);
|
||||
ReadGlobalSetting(Settings::values.barrier_feedback_loops);
|
||||
ReadGlobalSetting(Settings::values.bg_red);
|
||||
ReadGlobalSetting(Settings::values.bg_green);
|
||||
ReadGlobalSetting(Settings::values.bg_blue);
|
||||
@@ -1417,6 +1418,7 @@ void Config::SaveRendererValues() {
|
||||
WriteGlobalSetting(Settings::values.use_vulkan_driver_pipeline_cache);
|
||||
WriteGlobalSetting(Settings::values.enable_compute_pipelines);
|
||||
WriteGlobalSetting(Settings::values.use_video_framerate);
|
||||
WriteGlobalSetting(Settings::values.barrier_feedback_loops);
|
||||
WriteGlobalSetting(Settings::values.bg_red);
|
||||
WriteGlobalSetting(Settings::values.bg_green);
|
||||
WriteGlobalSetting(Settings::values.bg_blue);
|
||||
|
||||
@@ -508,7 +508,7 @@ void ConfigureGraphics::RetrieveVulkanDevices() {
|
||||
vulkan_devices.push_back(QString::fromStdString(record.name));
|
||||
device_present_modes.push_back(record.vsync_support);
|
||||
|
||||
if (record.is_intel_proprietary) {
|
||||
if (record.has_broken_compute) {
|
||||
expose_compute_option();
|
||||
}
|
||||
}
|
||||
|
||||
@@ -43,6 +43,8 @@ void ConfigureGraphicsAdvanced::SetConfiguration() {
|
||||
ui->enable_compute_pipelines_checkbox->setChecked(
|
||||
Settings::values.enable_compute_pipelines.GetValue());
|
||||
ui->use_video_framerate_checkbox->setChecked(Settings::values.use_video_framerate.GetValue());
|
||||
ui->barrier_feedback_loops_checkbox->setChecked(
|
||||
Settings::values.barrier_feedback_loops.GetValue());
|
||||
|
||||
if (Settings::IsConfiguringGlobal()) {
|
||||
ui->gpu_accuracy->setCurrentIndex(
|
||||
@@ -94,6 +96,9 @@ void ConfigureGraphicsAdvanced::ApplyConfiguration() {
|
||||
enable_compute_pipelines);
|
||||
ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_video_framerate,
|
||||
ui->use_video_framerate_checkbox, use_video_framerate);
|
||||
ConfigurationShared::ApplyPerGameSetting(&Settings::values.barrier_feedback_loops,
|
||||
ui->barrier_feedback_loops_checkbox,
|
||||
barrier_feedback_loops);
|
||||
}
|
||||
|
||||
void ConfigureGraphicsAdvanced::changeEvent(QEvent* event) {
|
||||
@@ -130,6 +135,8 @@ void ConfigureGraphicsAdvanced::SetupPerGameUI() {
|
||||
Settings::values.enable_compute_pipelines.UsingGlobal());
|
||||
ui->use_video_framerate_checkbox->setEnabled(
|
||||
Settings::values.use_video_framerate.UsingGlobal());
|
||||
ui->barrier_feedback_loops_checkbox->setEnabled(
|
||||
Settings::values.barrier_feedback_loops.UsingGlobal());
|
||||
|
||||
return;
|
||||
}
|
||||
@@ -157,6 +164,9 @@ void ConfigureGraphicsAdvanced::SetupPerGameUI() {
|
||||
ConfigurationShared::SetColoredTristate(ui->use_video_framerate_checkbox,
|
||||
Settings::values.use_video_framerate,
|
||||
use_video_framerate);
|
||||
ConfigurationShared::SetColoredTristate(ui->barrier_feedback_loops_checkbox,
|
||||
Settings::values.barrier_feedback_loops,
|
||||
barrier_feedback_loops);
|
||||
ConfigurationShared::SetColoredComboBox(
|
||||
ui->gpu_accuracy, ui->label_gpu_accuracy,
|
||||
static_cast<int>(Settings::values.gpu_accuracy.GetValue(true)));
|
||||
|
||||
@@ -48,6 +48,7 @@ private:
|
||||
ConfigurationShared::CheckState use_vulkan_driver_pipeline_cache;
|
||||
ConfigurationShared::CheckState enable_compute_pipelines;
|
||||
ConfigurationShared::CheckState use_video_framerate;
|
||||
ConfigurationShared::CheckState barrier_feedback_loops;
|
||||
|
||||
const Core::System& system;
|
||||
};
|
||||
|
||||
@@ -201,6 +201,16 @@ Compute pipelines are always enabled on all other drivers.</string>
|
||||
</property>
|
||||
</widget>
|
||||
</item>
|
||||
<item>
|
||||
<widget class="QCheckBox" name="barrier_feedback_loops_checkbox">
|
||||
<property name="toolTip">
|
||||
<string>Improves rendering of transparency effects in specific games.</string>
|
||||
</property>
|
||||
<property name="text">
|
||||
<string>Barrier feedback loops</string>
|
||||
</property>
|
||||
</widget>
|
||||
</item>
|
||||
<item>
|
||||
<widget class="QWidget" name="af_layout" native="true">
|
||||
<layout class="QHBoxLayout" name="horizontalLayout_1">
|
||||
|
||||
@@ -5,10 +5,12 @@
|
||||
#include <vector>
|
||||
#include "common/dynamic_library.h"
|
||||
#include "common/logging/log.h"
|
||||
#include "video_core/vulkan_common/vulkan_device.h"
|
||||
#include "video_core/vulkan_common/vulkan_instance.h"
|
||||
#include "video_core/vulkan_common/vulkan_library.h"
|
||||
#include "video_core/vulkan_common/vulkan_surface.h"
|
||||
#include "video_core/vulkan_common/vulkan_wrapper.h"
|
||||
#include "vulkan/vulkan_core.h"
|
||||
#include "yuzu/qt_common.h"
|
||||
#include "yuzu/vk_device_info.h"
|
||||
|
||||
@@ -16,8 +18,8 @@ class QWindow;
|
||||
|
||||
namespace VkDeviceInfo {
|
||||
Record::Record(std::string_view name_, const std::vector<VkPresentModeKHR>& vsync_modes_,
|
||||
bool is_intel_proprietary_)
|
||||
: name{name_}, vsync_support{vsync_modes_}, is_intel_proprietary{is_intel_proprietary_} {}
|
||||
bool has_broken_compute_)
|
||||
: name{name_}, vsync_support{vsync_modes_}, has_broken_compute{has_broken_compute_} {}
|
||||
|
||||
Record::~Record() = default;
|
||||
|
||||
@@ -48,9 +50,10 @@ void PopulateRecords(std::vector<Record>& records, QWindow* window) try {
|
||||
properties.pNext = &driver_properties;
|
||||
dld.vkGetPhysicalDeviceProperties2(physical_device, &properties);
|
||||
|
||||
records.push_back(VkDeviceInfo::Record(name, present_modes,
|
||||
driver_properties.driverID ==
|
||||
VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS));
|
||||
bool has_broken_compute{Vulkan::Device::CheckBrokenCompute(
|
||||
driver_properties.driverID, properties.properties.driverVersion)};
|
||||
|
||||
records.push_back(VkDeviceInfo::Record(name, present_modes, has_broken_compute));
|
||||
}
|
||||
} catch (const Vulkan::vk::Exception& exception) {
|
||||
LOG_ERROR(Frontend, "Failed to enumerate devices with error: {}", exception.what());
|
||||
|
||||
@@ -24,12 +24,12 @@ namespace VkDeviceInfo {
|
||||
class Record {
|
||||
public:
|
||||
explicit Record(std::string_view name, const std::vector<VkPresentModeKHR>& vsync_modes,
|
||||
bool is_intel_proprietary);
|
||||
bool has_broken_compute);
|
||||
~Record();
|
||||
|
||||
const std::string name;
|
||||
const std::vector<VkPresentModeKHR> vsync_support;
|
||||
const bool is_intel_proprietary;
|
||||
const bool has_broken_compute;
|
||||
};
|
||||
|
||||
void PopulateRecords(std::vector<Record>& records, QWindow* window);
|
||||
|
||||
Reference in New Issue
Block a user