Files
allwpilib/wpilibc/src/main/native/cpp/system/Notifier.cpp

178 lines
4.8 KiB
C++
Raw Normal View History

// Copyright (c) FIRST and other WPILib contributors.
// Open Source Software; you can modify and/or share it under the terms of
// the WPILib BSD license file in the root directory of this project.
2025-11-07 19:56:21 -05:00
#include "wpi/system/Notifier.hpp"
#include <utility>
2025-11-07 19:56:21 -05:00
#include "wpi/hal/DriverStation.h"
#include "wpi/hal/Notifier.hpp"
2025-11-07 19:56:21 -05:00
#include "wpi/hal/Threads.h"
#include "wpi/system/Errors.hpp"
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
#include "wpi/util/Synchronization.h"
2025-11-07 20:00:05 -05:00
using namespace wpi;
Notifier::Notifier(std::function<void()> callback) {
if (!callback) {
2025-11-07 20:00:43 -05:00
throw WPILIB_MakeError(err::NullParameter, "callback");
}
m_callback = callback;
int32_t status = 0;
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
m_notifier = HAL_CreateNotifier(&status);
WPILIB_CheckErrorStatus(status, "CreateNotifier");
m_thread = std::thread([=, this] {
for (;;) {
int32_t status = 0;
HAL_NotifierHandle notifier = m_notifier.load();
if (notifier == 0) {
break;
}
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
if (WPI_WaitForObject(notifier) == 0) {
break;
}
std::function<void()> callback;
{
std::scoped_lock lock(m_processMutex);
callback = m_callback;
}
// Call callback
if (callback) {
callback();
}
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
// Ack notifier
HAL_AcknowledgeNotifierAlarm(notifier, &status);
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
WPILIB_CheckErrorStatus(status, "AcknowledgeNotifier");
}
});
}
Notifier::Notifier(int priority, std::function<void()> callback) {
if (!callback) {
2025-11-07 20:00:43 -05:00
throw WPILIB_MakeError(err::NullParameter, "callback");
}
m_callback = callback;
int32_t status = 0;
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
m_notifier = HAL_CreateNotifier(&status);
2025-11-07 20:00:43 -05:00
WPILIB_CheckErrorStatus(status, "InitializeNotifier");
m_thread = std::thread([=, this] {
int32_t status = 0;
HAL_SetCurrentThreadPriority(true, priority, &status);
for (;;) {
HAL_NotifierHandle notifier = m_notifier.load();
if (notifier == 0) {
break;
}
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
if (WPI_WaitForObject(notifier) == 0) {
break;
}
std::function<void()> callback;
{
std::scoped_lock lock(m_processMutex);
callback = m_callback;
}
// call callback
if (callback) {
try {
callback();
2025-11-07 20:00:05 -05:00
} catch (const wpi::RuntimeError& e) {
e.Report();
2025-11-07 20:00:43 -05:00
WPILIB_ReportError(
err::Error,
"Error in Notifier thread."
" The above stacktrace can help determine where the error "
"occurred.\n"
" See https://wpilib.org/stacktrace for more information.\n");
throw;
} catch (const std::exception& e) {
HAL_SendError(1, err::Error, 0, e.what(), "", "", 1);
throw;
}
}
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
// Ack notifier
HAL_AcknowledgeNotifierAlarm(notifier, &status);
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
WPILIB_CheckErrorStatus(status, "AcknowledgeNotifier");
}
});
}
Notifier::~Notifier() {
// atomically set handle to 0, then clean
HAL_NotifierHandle handle = m_notifier.exchange(0);
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
HAL_DestroyNotifier(handle);
// Join the thread to ensure the callback has exited.
if (m_thread.joinable()) {
m_thread.join();
}
}
Notifier::Notifier(Notifier&& rhs)
: m_thread(std::move(rhs.m_thread)),
m_notifier(rhs.m_notifier.load()),
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
m_callback(std::move(rhs.m_callback)) {
rhs.m_notifier = HAL_kInvalidHandle;
}
Notifier& Notifier::operator=(Notifier&& rhs) {
m_thread = std::move(rhs.m_thread);
m_notifier = rhs.m_notifier.load();
rhs.m_notifier = HAL_kInvalidHandle;
m_callback = std::move(rhs.m_callback);
return *this;
}
void Notifier::SetName(std::string_view name) {
2019-11-09 11:41:58 -08:00
int32_t status = 0;
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
HAL_SetNotifierName(m_notifier, name, &status);
2019-11-09 11:41:58 -08:00
}
void Notifier::SetCallback(std::function<void()> callback) {
std::scoped_lock lock(m_processMutex);
m_callback = callback;
}
2025-11-07 20:00:05 -05:00
void Notifier::StartSingle(wpi::units::second_t delay) {
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
int32_t status = 0;
HAL_SetNotifierAlarm(m_notifier, static_cast<uint64_t>(delay * 1e6), 0, false,
false, &status);
}
2025-11-07 20:00:05 -05:00
void Notifier::StartPeriodic(wpi::units::second_t period) {
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
int32_t status = 0;
HAL_SetNotifierAlarm(m_notifier, static_cast<uint64_t>(period * 1e6),
static_cast<uint64_t>(period * 1e6), false, false,
&status);
}
2025-11-07 20:00:05 -05:00
void Notifier::StartPeriodic(wpi::units::hertz_t frequency) {
StartPeriodic(1 / frequency);
}
void Notifier::Stop() {
int32_t status = 0;
HAL_CancelNotifierAlarm(m_notifier, false, &status);
2025-11-07 20:00:43 -05:00
WPILIB_CheckErrorStatus(status, "CancelNotifierAlarm");
}
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
int32_t Notifier::GetOverrun() const {
int32_t status = 0;
[hal] Revamp notifiers (#8424) This changes the HAL notifier interface to: - Use wpiutil signal objects. This means waiting is done through the `WPI_WaitObject` API instead of a dedicated function and allows for higher level code to simultaneously wait on notifiers and other events. - Interval timers are supported at the HAL layer - Handlers are now required to acknowledge notifications. This is invisible to users unless they're directly using the HAL API. - For interval timers, an overrun count is maintained to detect if the handler didn't acknowledge The underlying implementation still uses condition variables for the actual waiting. In basic testing using this approach seemed to be lower jitter than timerfd. Currently, the simulation and systemcore implementations are nearly identical except for a few additional sim hook bits. This could be refactored, but keeping them separate may make sense to keep the systemcore implementation easy to read and reason about, or if we ever choose to use a different underlying timer implementation on systemcore. The simulation side API is unchanged in form but does change in function--waiting for notifiers now only waits for currently running (or newly signaled) notifiers to acknowledge. To avoid a race condition in sim stepTiming, users of the low level API must make any alarm updates (especially for one-shot alarms) prior to acknowledging the previous alarm. The only current use of the interval timer feature is the `Notifier` class. The `TimedRobot` implementation still uses a single notifier and its own interval timing logic to ensure consistent callback order. Using separate notifiers for each user-level interval would substantially increase complexity. `Watchdog` also doesn't use the interval timer, as it's looking for an amount of time since the last `set` call rather than a recurring interval time. To reduce flicker, the sim GUI uses a fade out when a timeout goes from set to unset. This fixes tsan for wpilib and commands, and also fixes some spurious test failures.
2025-11-29 11:00:18 -08:00
int32_t overrun = HAL_GetNotifierOverrun(m_notifier, &status);
WPILIB_CheckErrorStatus(status, "GetNotifierOverrun");
return overrun;
}
bool Notifier::SetHALThreadPriority(bool realTime, int32_t priority) {
int32_t status = 0;
return HAL_SetNotifierThreadPriority(realTime, priority, &status);
}