[ntp:questions] Start of new GPS 1024 week epoch

unruh unruh at invalid.ca
Wed Aug 14 20:03:52 UTC 2013

On 2013-08-14, David Taylor <david-taylor at blueyonder.co.uk.invalid> wrote:
> On 14/08/2013 12:54, unruh wrote:
> []
>> Except of course if the rd_timestamp.l_i is way out (that is why one
>> would want to use a gps clock to fix it-- eg on bootup with the
>> Raspberry Pi say),
> []
> Could you not use something like the timestamp of some file (e.g. the 
> drift file) or other system file to get the approximate year?  I haven't 
> studied the code (I find C not easy to read or navigate) so perhaps it 
> already does this.  Then you would only need to set the real time once?

Sure, you can kludge your way around the problem if you know there is a
problem. But ntpd code does not have any such routine in it, so you have
to do it by hand. Knowing it needs to be done, it can be done, but of
course most people have no idea it needs to be done. 

> As it happens, all of my RPi pack will have an Internet source as well, 
> and they all have recent GPS receivers which (I hope) don't have this 
> issue....

Many people have their RPi in some remote location with no internet
connection, collecting data, or whatever. GPS is the only source of
time. One would like an "install and go" solution without having to
worry about things like whether the time is in the right epoch,
whatever. (power goes out on that buoy out on the antlantic for a day,
and the thing reboots finally when the sun comes out again. You do not
want to have to swim out to it to reset the clock.)
I agree that this is a problem only under the "perfect storm"
situation-- old broken gps receiver, computer with no RTC, no internet
connection, etc. so the set of affected people could well be of size
much much less than 1.

More information about the questions mailing list