[LinuxPPS] FW: ntp-4.2.4p0 + ntpd_nmea.patch

[Udo van den Heuvel]

Rodolfo Giometti wrote:
> On Sun, Aug 19, 2007 at 06:31:17PM +0200, Udo van den Heuvel wrote:
> 
>> I am wrestling with:
>> How to deal with io_closeclock in the case of two fd's?
>> (when a NMEA device has two devices open for NMEA and PPS)
>>
>> (this is needed to handle all cases of 1 and two devices with NMEA and PPS)
> 
> I don't understand your problem... looking at file refclock_nmea.c I
> see only this io_closeclock() occurrence:

(....)
>    {
>            register struct nmeaunit *up;
>            struct refclockproc *pp;
> 
>            pp = peer->procptr;
>            up = (struct nmeaunit *)pp->unitptr;
>    #ifdef HAVE_PPSAPI
>            if (up->handle != 0)
>                    time_pps_destroy(up->handle);
>    #endif /* HAVE_PPSAPI */
>            io_closeclock(&pp->io);
>            free(up);
>    }
> 
> and it doesn't touch up->handle at all.

It handles the shuffling of a list of clocks.
I don't have the code in my head for that io_closeclock routine but some
stuff happens that makes me wonder.

So does NMEA with and/or without PPS always appear as one clock?
(yes?)

> However I suggest to you to do something like this during open phase:
> 
>    #ifdef HAVE_PPSAPI
>            /*
>             * Start the PPSAPI interface if it is there. Default to use
>             * the assert edge and do not enable the kernel hardpps.
>             */
>            if (time_pps_create(fd, &up->handle) < 0) {
> 	   	   /* Try the alternare PPS device */

When would this fail?
When no link in /dev is present?
Other situations? (i.e.: when is a PPS device a real enough PPS device
for ntpd?)

> 		   (void) sprintf(device, DEVICEPPS, unit);
> 
> 		   fd = open(device, O_RDWR);
> 		   if (fd < 0)
>                            goto pps_error;
> 
> 		   if (time_pps_create(fd, &up->handle) < 0) 
> 			   goto	pps_error;
>            }
>            return(nmea_ppsapi(peer, 0, 0));
> 
> pps_error:
>            /* No luck, no PPS unit available! */
>            up->handle = 0;
>            msyslog(LOG_ERR,
>                "refclock_nmea: alternate PPS device %s fail : %m", device);
>            return (1);
> 
>    #else
>            return (1);
>    #endif /* HAVE_PPSAPI */
> 
> Please, check the code, I just wrote it without reveiwing. :)

Thanks, I will try to use it.
My sleep reccovery is progressing so I could be doing something
intelligent tomorrow. ;-)

Kind regards,
Udo