Re: PROBLEM: skew message does not handle negative ns skew

[Feng Tang]

On Tue, Jun 06, 2023 at 02:09:08PM +0100, Chris Bainbridge wrote:
> On Tue, 6 Jun 2023 at 13:50, Feng Tang <feng.tang@xxxxxxxxx> wrote:
> >
> > And I'm have no idea if there is a real hardware/firmware issue
> > or just a false alarm.
> 
> Is a negative reported skew valid? I don't know, I had assumed so, so
> the problem was the conversion from -878159 ns to 18446744073708 ms.

I think it's valid. The related code is from kernel/time/clocksource.c: 

	"
	cs_wd_msec = div_u64_rem(cs_nsec - wd_nsec, 1000U * 1000U, &wd_rem);
	wd_msec = div_u64_rem(wd_nsec, 1000U * 1000U, &wd_rem);
	pr_warn("                      Clocksource '%s' skewed %lld ns (%lld ms) over watchdog '%s' interval of %lld ns (%lld ms)\n",
		cs->name, cs_nsec - wd_nsec, cs_wd_msec, watchdog->name, wd_nsec, wd_msec);
	"

The negative value just means the watchdog is running faster than
TSC in the 512 ms checking interval. The 18446744073708 ms is just
a conversion from s64 value in ns (-878159) to a u64 ns, then a
u64 ms. 

Thanks,
Feng