struct timex and Linux adjtimex()

[Konstantin Belousov]

On Thu, Dec 03, 2020 at 10:13:34PM +0000, Poul-Henning Kamp wrote:
> --------
> Cy Schubert writes:
> 
> > I will go back 
> > with my initial proposal of a timespec add/subtract syscall takes a 
> > timespec as input increments or decrements the clock by the timespec and 
> > returns a timespec with the time.
> 
> I would be tempted by the clock_settime(2) "clock_id" argument.
> 
> The functionality required has a LOT more commonality with
> clock_settime(2) than with ntp_adjtime(2), and absconding with a
> couple of the top bits of clock_id for "CLOCK_ADD_ADJUSTMENT" and
> "CLOCK_SUB_ADJUSTMENT" would be be a pretty clean solution.

The adjustments are only applicable to CLOCK_REALTIME, am I wrong ?
If only CLOCK_REALTIME is affected (or even if things like CLOCK_MONOTONIC
get parallel adjustment), it makes more sense to allocate designated clock
ids like CLOCK_BSD_ADD_ADJ instead of carving bits in id numbers.

I mean, we do not need/want to adjust CLOCK_VIRTUAL, CLOCK_PROF and other
non-wall clock related clocks.