Setting the current time in RTC may be a blocking operation (driver
needs to wait for oscillator stabilization after reset and so on).
This may cause the unwanted effect of clock_settime blocking the
code execution for a considerable amount of time.
The solution is to plan a low priority work that takes care
of setting the time in RTC and let clock_settime continue. We don't
have to check if the work is available, just cancel it if there
is a new time set request.
This is used only if CONFIG_SCHED_LPWORK is enabled.
Signed-off-by: Michal Lenc <michallenc@seznam.cz>
debug.h is a NuttX-specific, non-POSIX header. Placing it in the
top-level include/ directory creates naming conflicts with external
projects that define their own debug.h.
This commit moves the canonical header to include/nuttx/debug.h,
following the NuttX convention for non-POSIX/non-standard headers,
and updates all in-tree references.
A backward-compatibility shim is left at include/debug.h that
emits a deprecation #warning and re-includes <nuttx/debug.h>,
allowing out-of-tree code to continue building while migrating.
Signed-off-by: Piyush Patle <piyushpatle228@gmail.com>
Fix the issue where clock_gettime(CLOCK_MONOTONIC) always returns 0 in
tickless mode, caused by the scheduler tick counter (g_system_ticks) not
being updated with the actual timer ticks.
1. Add clock_update_sched_ticks() function to sched/clock/clock_sched_ticks.c:
a. This function directly sets the system tick counter to a specific value
(vs incrementing in clock_increase_sched_ticks), with full documentation
matching the existing code style.
b. Uses seqlock to ensure thread-safe access to g_system_ticks.
2. Call clock_update_sched_ticks() in nxsched_process_timer() (sched_processtickless.c):
a. Syncs the scheduler tick counter with the actual timer ticks retrieved via up_timer_gettick().
With this fix, g_system_ticks is properly updated in tickless mode,
and clock_gettime(CLOCK_MONOTONIC) returns the correct non-zero monotonic time.
Signed-off-by: chao an <anchao.archer@bytedance.com>
Changes the default value for CONFIG_BOARD_LOOPSPERMSEC to -1, which is
invalid. All boards that forget to configure this value will encounter a
static assertion at compile time, enforcing that configurations
upstreamed to NuttX have calibrated values for correct delay timings.
Boards which implement ALARM_ARCH or TIMER_ARCH do not rely on the
busy-loop delay implementation and thus only have a run-time check to
ensure proper delay-timings (in the case where delays are used before
the alarm/timer driver is registered).
Some boards already in the NuttX upstream do not have calibrated values,
but there are no users who currently own the board to submit a
calibrated value for the configurations to use. In this scenario, the
value is temporarily set to 0 and a warning is displayed so that users
of these boards are informed of the calibration process.
Signed-off-by: Matteo Golin <matteo.golin@gmail.com>
The current implementation retrieves time from the same source for both
CLOCK_BOOTTIME and CLOCK_MONOTONIC, which is incorrect. CLOCK_BOOTTIME
includes time spent in system suspend, whereas CLOCK_MONOTONIC does not.
This patch fixes the incorrect clock source handling and ensures the
two clocks are distinguished properly.
Signed-off-by: Chengdong Wang <wangchengdong@lixiang.com>
Replace CONFIG_SYSTEM_TIME64 conditional atomic64 operations with a unified
seqlock-based approach for managing the system tick counter. This simplifies
code by eliminating architecture-specific branches and provides more robust
thread-safe access patterns.
Signed-off-by: hujun5 <hujun5@xiaomi.com>
This change enables performance monitoring unit (PMU) access from userspace
on ARMv7-R architecture by adding CONFIG_ARCH_HAVE_PERF_EVENTS_USER_ACCESS
support and building PMU code for userspace when needed.
Signed-off-by: hujun5 <hujun5@xiaomi.com>
Implement Linux-compatible TFD_TIMER_CANCEL_ON_SET flag for timerfd to
allow applications to detect discontinuous changes to CLOCK_REALTIME.
Background:
According to Linux timerfd_create(2) man page, when a timerfd is created
with CLOCK_REALTIME and TFD_TIMER_CANCEL_ON_SET flag is specified, the
read() operation should fail with ECANCELED if the real-time clock
undergoes a discontinuous change. This allows applications to detect
and respond to system time changes.
Implementation:
1. Add clock notifier infrastructure (clock_notifier.h/c) to notify
interested parties when system time changes
2. Implement TFD_TIMER_CANCEL_ON_SET flag support in timerfd
3. Register timerfd with clock notifier when flag is set
4. Cancel timer and return ECANCELED when clock change is detected
5. Notify clock changes in:
- clock_settime()
- clock_initialize()
- clock_timekeeping_set_wall_time()
- rpmsg_rtc time synchronization
Changes include:
- New files: include/nuttx/clock_notifier.h, sched/clock/clock_notifier.c
- Modified: fs/vfs/fs_timerfd.c to handle TFD_TIMER_CANCEL_ON_SET
- Modified: clock subsystem to call notifier chain on time changes
- Modified: rpmsg_rtc to notify time changes during sync
Use case example:
Applications using timerfd with absolute time can now detect when
system time is adjusted (e.g., NTP sync, manual time change) and
take appropriate action such as recalculating timeouts or updating
scheduled events.
Reference: https://man7.org/linux/man-pages/man2/timerfd_create.2.html
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
Add ARCH_HAVE_PERF_EVENTS_USER_ACCESS capability to allow applications
to directly access hardware perf counters via perf_gettime() from
userspace, enabling performance monitoring and profiling without syscalls.
Signed-off-by: Yanfeng Liu <p-liuyanfeng9@xiaomi.com>
Remove perf_init() function and automatic initialization call, relying
on static initialization instead to reduce code complexity and avoid
global variable access in perf_update callback.
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
dd ARCH_PERF_COUNT_BITWIDTH configuration to support architectures
with different perf counter bit widths (TriCore 31-bit, others 32-bit),
enabling proper overflow correction across diverse platforms.
Signed-off-by: zhangyuan29 <zhangyuan29@xiaomi.com>
Move hardware perf counter read after spinlock acquisition to prevent
reading stale values due to concurrent overflow updates.
Signed-off-by: yinshengkai <yinshengkai@xiaomi.com>
This patch implements clock_adjtime() with CLOCKFD support, enabling
precise time and frequency adjustments for PTP clocks through the
standard POSIX clock API.
Key changes include:
1. New clock_adjtime() system call:
- Added sched/clock/clock_adjtime.c implementation
- Provides POSIX-compliant time adjustment interface
- Supports both standard clocks and CLOCKFD-based dynamic clocks
- Added CONFIG_CLOCK_ADJTIME Kconfig option
2. CLOCKFD support in clock_adjtime():
- Detects CLOCKFD-encoded clockids via CLOCKFD_TO_FD() check
- Extracts file descriptor and validates through fs_getfilep()
- Issues PTP_CLOCK_ADJTIME ioctl to underlying PTP clock device
- Returns clock status and adjustment results via struct timex
3. Enhanced struct timex support:
- Extended include/sys/timex.h with additional ADJ_* flags
- Added ADJ_OFFSET, ADJ_FREQUENCY, ADJ_MAXERROR, ADJ_ESTERROR
- Added ADJ_STATUS, ADJ_TIMECONST, ADJ_TAI, ADJ_SETOFFSET
- Added ADJ_MICRO, ADJ_NANO for time unit selection
- Added STA_* status flags for clock state reporting
4. Clock operations supported:
- ADJ_SETOFFSET: Apply time offset adjustment
- ADJ_FREQUENCY: Adjust clock frequency (PPM)
- ADJ_MAXERROR: Set maximum error estimate
- ADJ_ESTERROR: Set estimated error
- ADJ_STATUS: Modify clock status bits
- ADJ_NANO: Use nanosecond resolution
- ADJ_SETOFFSET: Set absolute time offset
5. API declarations:
- Added clock_adjtime() prototype in include/nuttx/clock.h
- Enabled by CONFIG_CLOCK_ADJTIME configuration
- Compatible with Linux clock_adjtime() interface
Usage example:
int fd = open("/dev/ptp0", O_RDWR);
struct timex tx = {0};
tx.modes = ADJ_FREQUENCY;
tx.freq = 10000000; /* Adjust frequency by 10 PPM */
clock_adjtime(CLOCKFD(fd), &tx);
close(fd);
This completes the PTP clock framework's POSIX clock API integration,
providing comprehensive time control capabilities for precision timing
applications including PTP synchronization daemons (ptp4l, timemaster).
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
This patch extends CLOCKFD support to clock_getres(), allowing
applications to query the resolution of PTP clocks through file
descriptors using the standard POSIX clock API.
Key changes include:
1. Move clock_getres() from libc to kernel:
- Relocated from libs/libc/sched/clock_getres.c to sched/clock/
- Enables direct kernel-level file descriptor handling
- Maintains POSIX compliance while adding CLOCKFD support
2. CLOCKFD support in clock_getres():
- Detects CLOCKFD-encoded clockids using CLOCKFD_TO_FD() check
- Extracts file descriptor from clockid parameter
- Validates file descriptor through fs_getfilep()
- Issues PTP_CLOCK_GETRES ioctl to query clock resolution
3. PTP clock resolution query:
- Retrieves resolution from PTP clock device via ioctl
- Returns nanosecond-precision timing resolution
- Enables applications to determine clock accuracy capabilities
- Falls back to standard clock resolution for non-CLOCKFD clockids
4. Error handling:
- Returns EINVAL for invalid CLOCKFD file descriptors
- Properly manages file reference counting with fs_putfilep()
- Maintains backward compatibility with existing clock types
5. Build system updates:
- Updated libs/libc/sched/CMakeLists.txt and Make.defs
- Updated sched/clock/CMakeLists.txt to include clock_getres.c
- Ensures proper compilation in kernel context
Usage example:
int fd = open("/dev/ptp0", O_RDONLY);
struct timespec res;
clock_getres(CLOCKFD(fd), &res); /* Query PTP clock resolution */
printf("Resolution: %ld.%09ld sec\n", res.tv_sec, res.tv_nsec);
close(fd);
This completes the basic POSIX clock API support for dynamic PTP clocks,
providing applications with comprehensive clock information access.
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
This patch implements POSIX-compliant CLOCKFD support, enabling user
applications to access dynamic PTP clocks through file descriptors
combined with clock_gettime() and clock_settime() system calls.
Key changes include:
1. CLOCKFD macro support:
- Added CLOCKFD() macro in include/nuttx/clock.h
- Allows encoding file descriptor into clockid_t: CLOCKFD(fd)
- Enables accessing PTP clocks via: clock_gettime(CLOCKFD(fd), &ts)
2. clock_gettime() enhancements:
- Modified sched/clock/clock_gettime.c to detect CLOCKFD clockids
- Extracts file descriptor from clockid using CLOCKFD_TO_FD()
- Validates file descriptor and calls file_ioctl() with PTP_CLOCK_GETTIME
- Falls back to standard clock handling for non-CLOCKFD clockids
3. clock_settime() enhancements:
- Modified sched/clock/clock_settime.c to support CLOCKFD clockids
- Extracts file descriptor and validates accessibility
- Issues PTP_CLOCK_SETTIME ioctl to underlying PTP clock device
- Maintains backward compatibility with standard POSIX clocks
4. File descriptor validation:
- Checks file descriptor validity using fs_getfilep()
- Ensures proper reference counting with fs_putfilep()
- Returns appropriate error codes (EINVAL, EBADF) on failures
5. Integration with PTP clock framework:
- Seamlessly integrates with PTP clock driver's ioctl interface
- Enables standard POSIX clock API usage for dynamic PTP clocks
- No changes required to existing applications using standard clocks
Usage example:
int fd = open("/dev/ptp0", O_RDONLY);
struct timespec ts;
clock_gettime(CLOCKFD(fd), &ts); /* Get PTP clock time */
clock_settime(CLOCKFD(fd), &ts); /* Set PTP clock time */
close(fd);
This implementation follows the Linux kernel's PTP clock model, providing
a familiar interface for developers working with PTP hardware.
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
The current implementation only calls
clock_inittime() when RTC is enabled,
which causes it to be skipped on other
architectures. This patch ensures
clock_inittime() is invoked during clock
initialization for all cases.
Signed-off-by: Chengdong Wang <wangchengdong@lixiang.com>
This reverts the removal of ndelay_accurate from #14450, since as
mentioned in #17011, this fails to consider the `sim` architecture
where CONFIG_BOARD_LOOPSPERMSEC was set to 0 because of reliance on the
accurate implementations of the up_delay functions. All the commit did
was remove a more accurate implementation in favour of a less accurate
one.
Signed-off-by: Matteo Golin <matteo.golin@gmail.com>
Make g_system_ticks a static variable, as it is only accessed by
clock_get_sched_ticks().
This change improves code modularity and enhances safety, since
clock_get_sched_ticks() performs proper synchronization when
reading g_system_ticks.
Signed-off-by: Chengdong Wang wangchengdong@lixiang.com
Replace the spinlock/sched_lock pair in the sched clock module with
spin_lock_irqsave_nopreempt() to improve code clarity and consistency.
Signed-off-by: Chengdong Wang wangchengdong@lixiang.com
Improve nxclock_gettime(), to get the system time excluding
the time that the system is suspended, when the clockid is
CLOCK_MONOTONIC
Signed-off-by: Chengdong Wang wangchengdong@lixiang.com
Replace direct usage of g_system_ticks with clock_increase_sched_ticks()
and clock_get_sched_ticks()
Signed-off-by: Chengdong Wang wangchengdong@lixiang.com
In the current NuttX kernel implementation, there is no function
to represent pure scheduler ticks.
The existing clock_systime_ticks() cannot serve this purpose,
because it keeps increasing even when the scheduler is not running.
This happens since the returned value reflects real hardware time:
as long as the timer or alarm hardware is active,
clock_systime_ticks() will continuously increase.
This patch introduces two new functions that provide access to
scheduler ticks — a counter that only increases when the scheduler
itself is running.
Signed-off-by: Chengdong Wang wangchengdong@lixiang.com
clock_systime_timespec() always returns 0, so there is no need to
check the return value in the caller code, let us remove the return
value directly.
Signed-off-by: chao an <anchao.archer@bytedance.com>
Many different architectures re-implemented the exact same code for
`up_*delay` because it was originally written as architecture dependent
code. Busy-waiting can be done regardless of architecture, so this
commit moves that duplicated implementation to a common file with weak
definitions so that individual architectures (see tc32) are still able
to override the definition if needed/desired.
Default implementation is not included if ARCH_TIMER is enabled, since
it is more accurate and provides its own weak definitions to override.
Signed-off-by: Matteo Golin <matteo.golin@gmail.com>
This commit replaced periodical timer with the wd_start_next to improve the timing accuracy.
Signed-off-by: ouyangxiangzhen <ouyangxiangzhen@xiaomi.com>
This commit resolves a timing error caused by the round-up behavior in clock_time2ticks. In rare cases, this could lead to a two-tick increment within a single tick interval. To fix this, we introduced clock_time2ticks_floor, which guarantees the correct semantics for obtaining current system ticks.
Signed-off-by: ouyangxiangzhen <ouyangxiangzhen@xiaomi.com>
The missing <nuttx/spinlock.h> header caused following compile errors:
CC: clock/clock_adjtime.c clock/clock_adjtime.c: In function 'adjtime_wdog_callback':
clock/clock_adjtime.c:67:11: error: implicit declaration of function 'spin_lock_irqsave' [-Wimplicit-function-declaration]
67 | flags = spin_lock_irqsave(&g_adjtime_lock);
| ^~~~~~~~~~~~~~~~~
clock/clock_adjtime.c:78:3: error: implicit declaration of function 'spin_unlock_irqrestore' [-Wimplicit-function-declaration]
78 | spin_unlock_irqrestore(&g_adjtime_lock, flags);
Signed-off-by: Michal Lenc <michallenc@seznam.cz>
When using alarm_arch implementation, 64-bit time can be returned. Using unsign long will cause precision loss.
Signed-off-by: yinshengkai <yinshengkai@xiaomi.com>
LD: nuttx.elf
ld: /home/ligd/platform/mainline/nuttx/staging/libsched.a(mq_sndinternal.o): in function `wd_start_realtime':
/home/ligd/platform/mainline/nuttx/include/nuttx/wdog.h:274: undefined reference to `clock_realtime2absticks'
/home/ligd/platform/mainline/nuttx/include/nuttx/wdog.h:274:(.text+0xad): relocation truncated to fit: R_X86_64_PLT32 against undefined symbol `clock_realtime2absticks'
ld: /home/ligd/platform/mainline/nuttx/staging/libsched.a(mq_rcvinternal.o): in function `wd_start_realtime':
/home/ligd/platform/mainline/nuttx/include/nuttx/wdog.h:274: undefined reference to `clock_realtime2absticks'
/home/ligd/platform/mainline/nuttx/include/nuttx/wdog.h:274:(.text+0xad): relocation truncated to fit: R_X86_64_PLT32 against undefined symbol `clock_realtime2absticks'
make[1]: *** [Makefile:128: nuttx.elf] Error 1
make: *** [tools/Unix.mk:551: nuttx.elf] Error 2
Signed-off-by: ligd <liguiding1@xiaomi.com>
There will be a large performance loss after SCHED_CRITMONITOR is enabled.
By isolating thread running time-related functions, CPU load can be run with less overhead.
Signed-off-by: yinshengkai <yinshengkai@xiaomi.com>
Signed-off-by: buxiasen <buxiasen@xiaomi.com>
