I seem to recall the SS1 having a Zilog 8530 Serial Communications
Controller (SCC) for the tty ports.  While the 8530 has a lot of
interesting capabilities, a large FIFO isn't one of them.  I believe it had
a single character FIFO so an interrupt is generated for every character,
not the best design for high speed data... :(

The original Macintosh's (128, 512K, Plus, SE, etc) also used the 8530.
For their high-speed LAN (200+ Kbps), they basically took over the
processor while receiving or transmitting messages.  Since the 8530 could
use HDLC framing, this was possible since the interrupt intensive routines
only kicked in when a message for the particular machine was received.  I
do recall running 56Kbps serial through the modem port but I don't know if
the architecture of the SPARC and the SS1 will support 56K.

Just my US$0.02...


steven_grunza@ieee.org


At 12:51 PM 4/25/00 -0400, Jon Lindgren wrote:
>On Tue, 25 Apr 2000, Tim Walls wrote:
>
>> All,
>> 
>> This weekend I managed to finally get my Sparc Classic demand PPP-ing,
>> routing and NATing happily away - many thanks to everyone who replied
>> to my last (slightly off topic) email about getting ipnat to work :-).
>> 
>> My only remaining question <grin> is of the serial port speed.  Can the
>> serial port of the Sparc Classic (and pppd) be coaxed into working at
>> speeds > 38k4?  After much fiddling I thought I had minicom talking at
>> 57k6 (it could have been lying,) but I can't convince pppd to do
>> the same.
>
>[snip]
>
>Although it may be possible, I've never had any luck with it.  IIRC, the
>lower end models (SS1, SS2, IPC/IPX) don't have great uarts; I believe
>they have small or nonexistant FIFOs in them.  I've never had anywhere
>near blazing performance on my SS1, SS2 or IPC.  YMMV, though, as I gave
>up without too much experimentation...
>
>Hope this helps,
>
>-Jon
> --------------------------------------------------------------------
> "Core dumped.  Your fault."
>
>
>