Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site bbncca.ARPA
Path: utzoo!linus!bbncca!sdyer
From: sdyer@bbncca.ARPA (Steve Dyer)
Newsgroups: net.micro.6809
Subject: Re: Wanted: OS-9 Device Driver Boilerplate
Message-ID: <418@bbncca.ARPA>
Date: Sat, 24-Dec-83 12:39:32 EST
Article-I.D.: bbncca.418
Posted: Sat Dec 24 12:39:32 1983
Date-Received: Sun, 25-Dec-83 00:26:17 EST
References: <410@bbncca.ARPA> <123@mspiggy>
Organization: Bolt, Beranek and Newman, Cambridge, Ma.
Lines: 27

What IS the IRQ signal used for in the CoCo?  I swear, with every
technical turn I take, I am faced with more and more moronic decisions.
I've verified that the R/S Disk Controller uses a WD1793 FDC chip
which does not support side-select, and now I find out that IRQ doesn't
come out of the cartridge slot.  (There's still the open question
of whether the "Deluxe RS232 Pak" even attempts to bring the ACIA
interrupts out onto the bus--probably not, given the cartridge bus.)
I suppose that it should be possible to feed IRQ signals directly
to the 6809's IRQ pin IFF there are not conflicting uses for it.
What does the 60hz line clock use?  If it uses IRQ, it would not
seem incompatible with other devices.  What does the bit-banger
RS232 input line use?

Dan, I congratulate you on your ingenuity in the face of terrible
design decisions.  I take it that your clock-driven /t1 should still
work with protocols like KERMIT or UMODEM even in the face of lost
input characters, simply because they provide a reliable protocol layer.
Right?

My request for boilerplate still stands, however.  A good, canonical
non-proprietary example of an OS-9 driver (not necessarily containing
CoCo vagaries) would be of great interest to me, and probably to the other
members of this group.  Can anyone out there help?
-- 
/Steve Dyer
decvax!bbncca!sdyer
sdyer@bbncca