[CCing the new bug number] Am 09.10.2014 um 02:20 schrieb Ben Hutchings: > On Thu, 2014-10-09 at 01:22 +0200, Michael Biebl wrote: >> Does ntp and hwclock --hctosys not play well together? > > Yes, that's it! > > If you use NTP and tell the kernel to adjust the system time, that > causes the kernel to periodically write the system time to the RTC too, > because the system time will be more accurate. > > For hysterical raisins, the first call to settimeofday() that specifies > a 'time zone' (local time offset) implicitly sets whether the RTC is > supposed to hold local time or UTC: if the time value is unspecified and > the time offset is non-zero then the RTC holds local time, otherwise it > holds UTC. There is no way to change this later! > > 'hwclock --hctosys' specifies both time value and time offset, and > therefore implicitly configures the RTC to hold UTC (but only when NTP > is used and it is written periodically by the kernel). > > So we absolutely have to run 'hwclock --systz' first. We could *also* > run 'hwclock --hctosys' straight after that.
Both variants seem to work here, i.e. systz only or hctosys *after* systz. Tested both, with and without ntp enabled, under systemd and sysvinit. -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
signature.asc
Description: OpenPGP digital signature
