Subject: Re: NTP pulse-per-second timestamp diff
To: Ty Sarna <tsarna@endicor.com>
From: Ignatios Souvatzis <is@jocelyn.rhein.de>
List: tech-kern
Date: 03/27/1998 22:24:57
> In article <199803271938.OAA14149@ginger.cmf.nrl.navy.mil> you write:
> > AFAIK, the com driver (which I believe is the only supported UART
> > that has a hardware FIFO) already does this for lower baud rates;
> > anything 1200 baud or lower, it sets the fifo trigger to 1.
> 
> The zs chips have a hardware FIFO, but the zs driver (at least - dunno
> about other drivers we may have for the zilog chips) always uses a trigger
> depth of 1 (which is all that's supported on the 8530).  Some machines
> (alpha?) have the '230 which can interrupt at depth 4 (for a larger 8
> byte FIFO).  The zs driver didn't use this feature though, last I
> looked. 
> 
> There may be other chips that do, but com and zs seem to cover the vast
> majority of NetBSD machines...

The Amiga A2232 board uses a 6502 cpu to handle lots of things, including an
internal fifo for the 7 lines. This could count as a hardware fifo from the
POV of NetBSD.

Regards,
	Ignatios