The 05/26/2020 20:55, Richard Cochran wrote: > On Thu, May 14, 2020 at 12:28:08PM +0200, Olivier Dautricourt wrote: > > The required time adjustment is written in the Timestamp Update registers > > while the Sub-second increment register is programmed with the period > > of the clock, which is the precision of our correction. > > I don't see in this patch where the "required time adjustment is > written in the Timestamp Update registers". > > What am I missing? > > Thanks, > Richard
This routine already exists in the driver (adjust_systime function). This one is called on ADJ_SETOFFSET in both functionning modes. So in both modes the phase jump is done by setting the target time in those update registers, set a flag and waiting for this flag to be cleared, that would mean that the correction is effective. Regards, -- Olivier Dautricourt
