[PATCH v2] tests: posix: use cpu clock for sleep

[Yury Khrustalev]

On Tue, Feb 17, 2026 at 01:09:56PM -0500, DJ Delorie wrote:
> Yury Khrustalev <yury.khrustalev@arm.com> writes:
> >  so I don't see what's wrong here. Please elaborate.
> 
> Perhaps I'm miscommunicating.
> 
> I don't care which clock you use to determine if 2 seconds has passed.
> 
> I'm looking at these lines:
> 
> +  while (ts.tv_sec < end || (ts.tv_sec == end && ts.tv_nsec < end_ns))
> +    clock_gettime (CLOCK_PROCESS_CPUTIME_ID, &ts);
> 
> This is a busy-wait loop.  It would be better on most platforms to do
> *something* to pause the program, even if only a little.  Example:
> 
> +  while (ts.tv_sec < end || (ts.tv_sec == end && ts.tv_nsec < end_ns))
> +    {
> +      sleep(1);
> +      clock_gettime (CLOCK_PROCESS_CPUTIME_ID, &ts);
> +    }

I see, thanks for clarification!

> 
> On your target, the sleep() would do nothing,

The problem is that it doesn't to nothing exactly: it advances "time" by
whatever delta we're waiting for but does so instantly so the current
thread continues but its perception of the result of time() has changed
in a "correct" way.

If busy waiting is not ideal, then I guess I have to keep this patch
downstream. Let me think more about this.