Re: [PATCH] tools/nolibc: add support for clock_nanosleep() and nanosleep()

["Thomas Weißschuh" <linux@weissschuh.net>]

On 2025-07-06 08:26:33+0200, Willy Tarreau wrote:
> On Fri, Jul 04, 2025 at 04:19:48PM +0200, Thomas Weißschuh wrote:
> > +static __attribute__((unused))
> > +int sys_clock_nanosleep(clockid_t clockid, int flags, const struct timespec *rqtp,
> > +			struct timespec *rmtp)
> > +{
> > +#if defined(__NR_clock_nanosleep)
> > +	return my_syscall4(__NR_clock_nanosleep, clockid, flags, rqtp, rmtp);
> > +#elif defined(__NR_clock_nanosleep_time64)
> > +	struct __kernel_timespec krqtp, krmtp;
> > +	int ret;
> > +
> > +	__nolibc_timespec_user_to_kernel(rqtp, &krqtp);
> > +	ret = my_syscall4(__NR_clock_nanosleep_time64, clockid, flags, &krqtp, &krmtp);
> > +	if (rmtp)
> > +		__nolibc_timespec_kernel_to_user(&krmtp, rmtp);
> > +	return ret;
> > +#else
> > +	return __nolibc_enosys(__func__, clockid, flags, rqtp, rmtp);
> > +#endif
> 
> I don't know which archs do not have clock_nanosleep, but if it becomes
> needed on some of them, we could probably fall back to pslelect() if
> available, of course, and ignore the clockid.

No architecture should ever run into the #else.
It is mostly for completeness and consistency with other architectures.
As for falling back to pselect(), the clockid is indeed important :-)

> Acked-by: Willy Tarreau <w@1wt.eu>

Thanks!