gpsd-dev
[Top][All Lists]
Advanced

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

Re: [gpsd-dev] SiRF IV on gpsmon


From: Gary E. Miller
Subject: Re: [gpsd-dev] SiRF IV on gpsmon
Date: Mon, 18 Nov 2013 10:58:36 -0800

Yo Eric!

On Mon, 18 Nov 2013 04:21:06 -0500
"Eric S. Raymond" <address@hidden> wrote:

> Gary E. Miller <address@hidden>:
> > I read the SiRF IV doc.  It specifically says that once you
> > initialize a SiRCC IV that you must wait for each ACK before
> > proceeding.  I found that by lengthening the delays between
> > initialization messages that things work a bit better.  But the doc
> > says that waits up to 4 seconds may be required annd that sending
> > another message before the ACK will result in problems.  Looks
> > pretty much like what I amm seeing.
> 
> OK, looking at code...I see your workaround. Has the problem that the
> settle-down waits you put in potentially block message traffic from
> the device, especially during the big sequence of configuration sends
> at activation time.

I did not put those in, they were there, just not consistent.  All I
did was make them consistent.

And yes, the sirf driver code has basically allowed no output from the
GPS while doing a large nnumber of control writes.

> There's a slightly better way to do it, I think.  The model is the
> way that driver probes are spaced out over the first nine seconds of
> NMEA activation
> - look in drivers.c:nmea_event_hook(). We can pull a similar trick in
> the SiRF configure_event hook to space the 9 sends it wants to do
> over the first 36 seconds, guaranteeing a 4-second interval between
> them.

The problem is, and I have tested this, if you always wait 4 seconds
then the GPS output buffer overflows.

> Then we put a guard in sirf_write() itself that stalls until 4 seconds
> have elapsed since the *last* send, then sends and updates the
> last-send timestamp.  That way control sends will never fail or screw
> up the device unless SiRF is being optimistic about the settling time.

But the doc says to wait for the ACK.
 
RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97701
        address@hidden  Tel:+1(541)382-8588

Attachment: signature.asc
Description: PGP signature


reply via email to

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