Subject: Re: Woops alias Domain/OS keyboard
To: None <port-hp300@NetBSD.ORG>
From: Michael Joosten <joost@ori.cadlab.de>
List: port-hp300
Date: 11/06/1995 16:09:28
>If someone can figure out where the UART (prolly an scn2681) on that 
>keyboard controller lives, I'm certainly willing to make the necessary 
>changes to the console driver to support that keyboard.  Herb is the 
>Apollo export :-), so he'll have to tell me how to detect if a keyboard 
>is actually there, etc.  (Reset, I'd assume, maybe some sort of 
>cookie..)  It's be nice not to have to add an option.

While the 'original Apollo Apollos' (ie, DN3000 - DN4500 (DN5500?)) all have
SCN2681 dual UARTs, the HP400/HP425/etc don't have them 'visually' on the
motherboard. It might well be that they are contained in that mysterious
'Utility Chip at sc 41'. I've found an old IWORKS CD from '93 with
comp.sys.hp* archives, and there some guy from HP reveals :


>>>>>>
The higher speeds, FIFO and CTS/RTS flow control apply to the first (select
code 9) serial port on all Model 400 machines including 425t, but excluding
the new 425e.  This type port, based on the Ns16550, has not been fully
characterized at high speeds on a busy system.

The other two ports (select codes 5 & 6) have the characteristics of the
older Apollo DN-series ports; the 425e has only this type of port.  Although
this port has some buffer, it lacks CTS/RTS and has a fairly low max bps
rate (perhaps 38,400).  I have no further data on it.

Regards,                                              Hewlett-Packard
Bob Niland      Internet: rjn@FC.HP.COM               3404 East Harmony Road
                UUCP: [hplabs|hpfcse]!hpfcrjn!rjn     Ft Collins CO 80525-9599
The 425e's three serial ports are the Apollo DN-series flavor, like the two
extra ports on the other 400s.  The first port (select code 9) is Apollo-
style, and is NOT based on the Ns16550 used in the other 400s.  Each has a
max bit rate of 19200, no RTS, and an in/out FIFO of 1/8 bytes.  They may
not be suitable for "high speed" modems.

> Unusually, I want to get it headless, since it's going to be a server for
> me, rather than an in-office workstation.

The 425e uses only the Apollo "utility" chip for RS-232C, which the 400 does
not support as console.  The 425e does not have the Ns16550 port which the
300s (and 400s) support as console.  Consequently, there is no switch or
boot ROM configuration setting to enable a serial console on the 425e.

<<<<<<<
>From the selftests in a 400t it seems that there are at least three 'Apollo
type' ports, one for the Domain/OS keyboard, the other one connected to the
single RS232, where one has to use a splitter cable to get all of them. As the
Rodime with 10.3.5 made its farewell, I'd like to ask how many /dev/sios are
supported by Domain/OS with this machine? And if one could use the same cable
splitter as was delivered with DN4500/3500? (we still have two of them)?

On the other hand, I *think* I've read that there was a possibility to use
these additional Apollo ports from HPUX, but they had to be configured in the
kernel separately (additional .o module). 

I also have a data sheet for the SCN2681, and it claims that the 40pin version
was well able to do RTS/CTS, even DTR/DCD. The smaller 28pin lacked this. The
DN[34]00 models all have one or two 40pin 2681.

Moreover, Bob Niland speaks about select code 5 & 6. If I use the mapping
sctopa() in NetBSD/arch/hp300/..., I cannot detect anything at the
corresponding addresses. Another wierd one is that the Domain/OS kernel's map
file for the 400t only cites routines for 16550, but not for the 2681 any
more. 

Looks I should post some pleas for help in comp.sys.hp.hardware, perhaps there
is still someone at HP who knows about this stuff. When the 400t appeared,
some Ex-HP guy working at Cayman (Carl Heinzl <carl@Cayman.COM>) gave some
infos about the inner workings of the dual mode boot proms, but that's now 3.5
years ago.

There is still some Apollo hardware here, so I've no problmes using a few
keyboards together, clamping probes to reveal how the protocol works
(especially the raw keystrokes + mouse movements). And I'm still interested in
having this hp400 working with Domain/OS keyboard, NetBSD *and* an X server!

Michael