[PATCH 09/15] tick/sched: Move dyntick-idle cputime accounting to cputime code

[Frederic Weisbecker]

Le Mon, Jan 19, 2026 at 03:35:52PM +0100, Peter Zijlstra a écrit :
> On Fri, Jan 16, 2026 at 03:52:02PM +0100, Frederic Weisbecker wrote:
> 
> > +static void kcpustat_idle_stop(struct kernel_cpustat *kc, ktime_t now)
> >  {
> > +	u64 *cpustat = kc->cpustat;
> > +	ktime_t delta;
> > +
> > +	if (!kc->idle_elapse)
> > +		return;
> > +
> > +	delta = ktime_sub(now, kc->idle_entrytime);
> > +
> > +	write_seqcount_begin(&kc->idle_sleeptime_seq);
> > +	if (nr_iowait_cpu(smp_processor_id()) > 0)
> > +		cpustat[CPUTIME_IOWAIT] = ktime_add(cpustat[CPUTIME_IOWAIT], delta);
> > +	else
> > +		cpustat[CPUTIME_IDLE] = ktime_add(cpustat[CPUTIME_IDLE], delta);
> > +
> > +	kc->idle_entrytime = now;
> > +	kc->idle_elapse = false;
> > +	write_seqcount_end(&kc->idle_sleeptime_seq);
> >  }
> 
> I realize this is mostly code movement; but do we really want to
> preserve ktime_{sub,add}() and all that?
> 
> I mean, we killed that 32bit ktime nonsense ages ago.

Good point, this should just be u64.

Thanks!

-- 
Frederic Weisbecker
SUSE Labs