Messages in this thread | | | Date | Mon, 9 Jun 2008 19:18:52 +0100 (BST) | From | "Maciej W. Rozycki" <> | Subject | Re: [PATCH] NTP: Let update_persistent_clock() sleep |
| |
Hi,
> > Roman does most of the NTP work afaik. I consider Thomas's git-hrt > > tree to be the route via which NTP changes get into linux-next and > > mainline. > > That function is a little misplaced, we already have a driver/rtc dir, > where this should go in the long term and ntp.c only providing the > trigger, that time is stable. There it would also be possible to better > take into account any quirks needed to update the chip.
I posted a separate change to implement a backend using the RTC class device. I think this is the right solution till all the platforms are moved away from legacy RTC drivers.
I think you are right about the long-term implications and apart from any possible quirks I think the interface could get improved as there are RTC chips we support nowadays that provide sub-second resolution.
Maciej
| |