Re: [gpsd-dev] ntpshmmon

[Gary E. Miller]

Yo Nuno!

On Tue, 14 Jun 2016 22:07:51 +0100
Nuno Gonçalves <address@hidden> wrote:

> Epoch 97.0, processed by gpsd at 97.131, printed by ntpshmmon at 98.10
> Epoch 98.0, processed by gpsd at 98.129, printed by ntpshmmon at 98.60.

And the next one:

sample NTP0 1465896598.600334941 1465896598.129354454 1465896598.000000000 0  
-1 s

Yup, very wrong thing.  Either GPS problem or gpsd problem.  One of
those is off 0.500 seconds!  Not acceptable.

Except when you say 'printed by ntpshmmon', that time has nothing to do
with when ntpshmmon printed anything.

You are misinterpreting the output, see below.

> I don't know if the confusion is in regard to the first value, but it
> is the print time at ntpshmmon. It is not part of the shm message. If
> you kill gpsd that will the only value that continues to be updated.

Uh, no, both times come from the packet.  ntpshmmon add no timing 
information.

Here is the C code:

        printf("sample %s %ld.%09ld %ld.%09ld %ld.%09ld %d %3d\n",
               ntp_name(i),
               (long)shm_stat.tvc.tv_sec, shm_stat.tvc.tv_nsec,
               (long)shm_stat.tvr.tv_sec, shm_stat.tvr.tv_nsec,
               (long)shm_stat.tvt.tv_sec, shm_stat.tvt.tv_nsec,
               shm_stat.leap, shm_stat.precision);

tvc, tvr and tvt are not set by ntpshmmon, only read by ntpshmmon.
All the times are from the packet, nothing about when ntpshmmon saw
anything.

Here are the definitions of tvc, tvr and tvt:

    struct timespec tvc;    /* System time when SHM read, for debug only */
    struct timespec tvr;    /* System time at GPS time */
    struct timespec tvt;    /* GPS time */


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

pgp_Q595TLSqA.pgp
Description: OpenPGP digital signature