Clock not ticking during S3
Ian Smith
smithi at nimnet.asn.au
Wed Oct 2 10:40:10 UTC 2013
On Tue, 1 Oct 2013 13:12:38 +0200, Dominic Fandrey wrote:
> The following is in my ntpd log.
>
> ...
> 27 Sep 23:06:40 ntpd[3045]: Listening on interface #67 wlan0, fe80::21c:bfff:fe58:3a87#123 Enabled
> 27 Sep 23:06:49 ntpd[3045]: Listening on interface #68 wlan0, 172.17.2.154#123 Enabled
>
> The system is sent to S3 at this point and woken 4 days later.
>
> This is how it comes up:
>
> 27 Sep 23:07:03 ntpd[3045]: no servers reachable
> 27 Sep 23:19:54 ntpd[3045]: synchronized to 83.170.1.225, stratum 2
> 27 Sep 23:19:54 ntpd[3045]: time correction of 306709 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time.
>
> Roughly 3 and a half days of time missing. I've never seen anything like
> it before.
>
> This is my system.
> FreeBSD mobileKamikaze.norad 9.2-PRERELEASE FreeBSD 9.2-PRERELEASE #0 r254957: Tue Aug 27 19:07:40 CEST 2013 root at mobileKamikaze.norad:/usr/obj/HP6510b-9/amd64/usr/src/sys/HP6510b-9 amd64
My 9.1 system is (physically) broken at the moment so I only have 8.2
sources to hand, but in any case it looks like the RTC - the only source
of clock time available on resume, at least on i386 - was either stopped
on suspend when the RTC was updated from system time, or - perhaps more
likely? - couldn't be properly read back to restore time on resume.
Is there logging of this suspend and resume cycle in /var/log/meesages ?
ccing Alexander, resident master of clocks last time I tried following
RTC suspend/resume with a view to maybe implementing resume-on-alarm.
cheers, Ian
More information about the freebsd-stable
mailing list