Subject: Re: slattach -s 38400
To: None <nomad@castle.org, port-pc532@NetBSD.ORG>
From: Phil Budne <budd@cs.bu.edu>
List: port-pc532
Date: 01/25/1995 17:05:44
	To: port-pc532@NetBSD.ORG
	Subject: slattach -s 38400
	Date: Tue, 24 Jan 1995 22:04:44 -0800
	From: Lee Damon <nomad@castle.org>
	X-Loop: port-pc532@NetBSD.ORG

	Has anyone been able to get slattach to accept 38400 as a valid speed?

	dragon# /sbin/slattach -h -s 38400 /dev/tty02
	tcsetattr: Invalid argument
	dragon# /sbin/slattach -h -s 19200 /dev/tty02

	It kind of annoys me that I'm talking to my pc532 at half the speed I
	should be able to use.

Is your default rate 19200?  Try opening tty03 at 9600, and holding it
there, and THEN try the slattach!!

My hacked up version of the scn driver uses the counter timer to
generate a single fixed rate that is always available so that opening
one port at that rate doesn't lock the other side of the same DUART
into using that "side" of the baud rate generator table.

I've also experimented with using the C/T to generate 57600, but I'm
unable to read the data out quickly enough.  I tried using the RxRdy
interrupt to service JUST the input side, and stuff the data into a
ring buffer that is serviced from a timeout(), but is seems like 100Hz
isn't fast enough to get the data out at that rate, and I need to try
using a direct software interrupt for just the SCN driver.  I also
suspect I was being burned by the chip either NOT performing automatic
RTS/CTS flow, like I requested (I wrote code for this, and modem
events, (including /dev/cuNN devices, if you're willing to hack
tty.c)), OR that the chip doesn't drop CTS until after the FIFO is
full, and RTS isn't honored once a transmit starts!!

-phil