[LinuxPPS] LinuxPPS 5.0.0-rc1

[Rodolfo Giometti]

On Fri, Aug 10, 2007 at 05:44:21PM +0200, Udo van den Heuvel wrote:
> Rodolfo Giometti wrote:
> > On Fri, Aug 10, 2007 at 04:42:03PM +0200, Udo van den Heuvel wrote:
> >> Please review and give feedback so we all can learn.
> >>
> > 
> >> --- refclock_nmea.c.origineel	2007-08-10 16:16:16.000000000 +0200
> >> +++ refclock_nmea.c	2007-08-10 16:38:01.000000000 +0200
> (...)
> >> @@ -91,6 +93,7 @@
> >>  	pps_params_t pps_params; /* pps parameters */
> >>  	pps_info_t pps_info;	/* last pps data */
> >>  	pps_handle_t handle;	/* pps handlebars */
> >> +	int handle_created;	/* pps handle created flag */
> > 
> > Now this is not needed anymore since pps_handle_t is becomed an "int"
> > again... however this is not an error, but your patch may result
> > smaller if you don't use it! :)
> 
> pps_handle_t is not in my patch, but I will change it to int.

Sorry, was my fault... I mean that you shouldn't use "handle_created"
anymore since "pps_handle_t" is becomed an "int" so you can set handle
to "-1" in case of error.

> >>  	/*
> >>  	 * Open serial port. Use CLK line discipline, if available.
> >>  	 */
> >> -	(void)sprintf(device, DEVICE, unit);
> >> +	(void)sprintf(gpsdevice, GPSDEVICE, unit);
> >> +	(void)sprintf(ppsdevice, PPSDEVICE, unit);
> >>  
> >> -	fd = refclock_open(device, SPEED232, LDISC_CLK);
> >> -	if (fd <= 0) {
> >> +	gps_fd = refclock_open(gpsdevice, SPEED232, LDISC_CLK);
> >> +	pps_fd = refclock_open(ppsdevice, SPEED232, LDISC_CLK);
> >> +	if (gps_fd <= 0) {
> > 
> > Make attention here! Refclock should work in both ways:
> > 
> > 1) the GPS source and PPS source are the same device, or
> > 
> > 2) the	GPS source and PPS source are __different__ devices.
> > 
> > Your code assumes case 2 only. I suggest you in using bit 7 as PARSE
> > refclock in order to know what to do (please see my preovious letter
> > on this topic for a code example).
> 
> This is what puzzles me. As in the wiki you write:
> 
> (under  Modifing a reference clock to work with LinuxPPS)
> 
> So your programs should check if the GPS data source (the serial port
> for instance) is a PPS source too, otherwise they should provide the
> possibility to open another device as PPS source.
> 
> So I just wait for the 4th octet of the fake IP that ntpd uses?
> I browsed the parse reflock source and I found in ntp_request.h there is
> some definition of the mode parameter, but how do I fill this struct?

Please, look as PARSE does:

	/*
	 * PPS via PPSAPI
	 */
	#if defined(HAVE_PPSAPI)
	                parse->hardppsstate = PARSE_HARDPPS_DISABLE;
	                if (CLK_PPS(parse->peer))
	                {
	                  if (time_pps_create(parse->ppsfd, &parse->ppshandle) < 0)
	                    {
	                      msyslog(LOG_NOTICE, "PARSE receiver #%d: parse_start: could not set up PPS: %m", CLK_UNIT(parse->peer));
	                    }
	                  else
	                    {
	                      parse_ppsapi(parse);
	                    }
	                }
	#endif

you should do something similar, or you can just try the PPS source
anyway, for example:

	gps_fd = open(gpsdevice, ...);

	#if defined(HAVE_PPSAPI)
	handle = -1;
	pps_fd = -1;
	if (time_pps_create(gps_fd, &handle) {
		/* use handle */
	} else {
		/* try "ppsdevice" */
		pps_fd = open(ppsdevice, ...);
		if (time_pps_create(pps_fd, &handle) {
			/* use	handle */
		else {
			/* error: no valid PPS sources! *
		}
	}
	#endif

	...

	/* At the end we should close all opened handlers */
	if (pps_fd >= 0)
		time_pps_destroy(&handle)
	if (pps_fd >= 0)
		close(pps_fd);
	close(gps_fd);

Ciao,

Rodolfo

-- 

GNU/Linux Solutions                  e-mail:    giometti at enneenne.com
Linux Device Driver                             giometti at gnudd.com
Embedded Systems                     		giometti at linux.it
UNIX programming                     phone:     +39 349 2432127