gpsd-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gpsd-users] Current Leap Seconds


From: Gary E. Miller
Subject: Re: [gpsd-users] Current Leap Seconds
Date: Mon, 2 May 2016 18:19:31 -0700

Yo Parker!

On Wed, 27 Apr 2016 15:39:59 +0000
Parker Mackenzie <address@hidden> wrote:

> Will gpsd update the kernel with the current number of active leap
> seconds?

gpsd never tells the Linux kernel anything.  If you pair gpsd with
a time daemon, like ntpd, NTPsec, or chonryd, then the time daemon
will do the right thing.

But, note that the kernel has no concept of leap seconds.  Kernel time
is seconds, ignoring leap seconds, since: 1970-01-01 00:00:00 +0000 (UTC).

> My GPS (ublox) provides the current/previous leap seconds as
> well as an indication of a pending leap second. Before I plumb gpsd,
> pps, and NTP SHM into my application I wanted to make sure when the
> device booted (not connected to a network) the atomic clock would be
> set appropriately.

You make no mention of using a time daemon.  You need a time daemon to
get the kernel time synced, but ntpd and NTPsec, do not work at all when
not connected to the network.

gpsd does in fact pass your GPS leap second on the the chronyd daemon.

You can see that in timehint.c around line 316 to 350.

  I found the code in gpsd which provides the
> information about a pending leap second; but, did not see where
> current/previous leap seconds would be accounted for.

There is no use for previous leap seconds.  We only need to know i
a leap second is coming because the kernel has no concept of leap seconds.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        address@hidden  Tel:+1 541 382 8588

Attachment: pgpp9KdMWa9pm.pgp
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]