[ntp:hackers] u-blox reference clock driver

[Gary E. Miller]

Yo Udo!

On Sun, 12 Aug 2018 11:05:20 +0200
Udo van den Heuvel <udovdh at xs4all.nl> wrote:

> On 11-08-18 23:03, Gary E. Miller via devel wrote:
> > I've never found a good list of which products are FTS.  I think the
> > LEA-M8F, LEA-M8T and NEO-M8T are FTS.  Likely a few others, but not
> > the basic NEO-M8L, M, N, Q or S variants.  
> 
> Interesting info!
> Also: how could we interface these clocks to x86* hardware in the
> best way?

Theoretical best way that has never happened anywhere yet?

Or best practical way in current use?

gpsd knows about the NEO-M8T, but currently the results are no better
than with the NEO-M8N.

The key new feature in the TIMEMARK driver is haing the Linux kernel
output a signal back to the NEO-M8T so that the GPS can commpute the
CPU interrupt latency.  This requires an FTS device.  Then the new
ntpd/ublox driver uses that number to currect the timestamp sent to the
CPU.

This was on ARM.  To do the same on x96, or amd64, would not be a big
deal.  If you are handy at modifying kernel drivers.  Instead of
toggling a RasPi GPIO pin as Lucas does, toggle a serial port line.

But all theoretical until we get a copy of Lucas's kernel mods and
ntpd driver.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem at rellim.com  Tel:+1 541 382 8588

	    Veritas liberabit vos. -- Quid est veritas?
    "If you can’t measure it, you can’t improve it." - Lord Kelvin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 851 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ntpsec.org/pipermail/devel/attachments/20180812/17517652/attachment.bin>