[time-nuts] Leap second glitches on NTP using Z3801A
Scott Mace
smace at intt.net
Mon Nov 17 04:11:09 UTC 2008
Here's the diff based on 4.2.4p4
--- ntp-4.2.4p4.orig/ntpd/refclock_hpgps.c 2006-06-06 15:16:51.000000000 -0500
+++ ntp-4.2.4p4/ntpd/refclock_hpgps.c 2008-08-25 09:56:29.000000000 -0500
@@ -535,7 +535,8 @@
switch (leapchar) {
case '+':
- pp->leap = LEAP_ADDSECOND;
+ if ((month == 6) || (month == 12))
+ pp->leap = LEAP_ADDSECOND;
break;
case '0':
@@ -543,7 +544,8 @@
break;
case '-':
- pp->leap = LEAP_DELSECOND;
+ if ((month == 6) || (month == 12))
+ pp->leap = LEAP_DELSECOND;
break;
default:
Scott Mace wrote:
> Yes, I noticed this as well and modified the refclock driver to filter
> it as it does in the oncore refclock.
>
> Scott
>
> Hal Murray wrote:
>> Is anybody running ntpd with their Z3801A?
>>
>> If so, please check your log files and tell me if you see a bogus leap second
>> at the end of the past several months. I've seen them for Aug, Sep, and Oct.
>> I think they are coming from my Z3801A, but it might be something else.
>>
>> The GPS satellites are now announcing a leap second that will happen at the
>> end of the year. The refclock driver passes that to ntpd and ntpd passes it
>> to the kernel and magic happens.
>>
>> I think the refclock-ntpd interface assumes the leap second will happen at
>> the end of the current month. NIST only announces leap seconds a month ahead
>> on WWVB and ACTS.
>>
>> The Oncore refclock driver has a filter to wait until the current month to
>> pass the info to ntpd. I'm working on something similar for the HP driver.
>>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
>
More information about the Time-nuts_lists.febo.com
mailing list