RE: Serial Slave 0.7b out

From: Christopher Phillips (
Date: 2002-02-27 13:53:26

(Marko; I hope you do not mind me replying to the list; some of my answers
are also relevant to MagerValp, and I suspect others are also interested).

Marko wrote privately:
> That sounds reasonable.  One 1541 track is at most 22 
> sectors, or 5632 
> bytes, which means that the tool requires at least a 3k 
> memory expansion 
> when it is ported to the VIC-20.

Shouldn't even require that; as I only buffer one sector
at a time (I just don't break the track up into seperate
'packets'; the host requests a track, and does
10*numSectors fread() attempts or until the track has
arrived, with 50ms delay between attempts)

> > I don't have the transfer working in the opposite direction 
> yet, and the
> > code needs a lot of tidying up.
> Which drives does the code support now?

Should be any that support M-W and M-E, has a page I can use
for code at 0x500, a free buffer at 0x300 and lets you trigger
jobs on that buffer by writing to locations 0 5 and 6. (or
whatever they are on 1541).  Oh, the serial lines must be at
$1800 in the same bit layout also, and be drivable by writing
to bits 1 and 3, with zeros in the high nibble and previous
bitpair sometimes incidentally written to bits 0 and 2.

>  You should probably 
> have a look 
> at the VIMM source code, whose IRQ loader works on all 
> Commodore serial 
> bus drives (1541/1570/1571 and 1581).

True.  I will tidy up my own code a little first though.
As well as what I mentioned in my previous email, I should
stop doing a full sync on every byte and just let the c64
skip ahead a few cycles whenever it is falling behind.

> Have you looked at the cbmlink null-modem protocol?  It 
> doesn't use any 
> handshaking, not even XON/XOFF, but it is capable of 
> transferring disk 
> images (well, one sector at a time).

Not yet.  I should also look at SerialSlave.  I am just
happy to have something working at the moment :)

> Did you say POSIX termios? 


> I thought that Apple had some weird Ain and 
> Aout devices, but maybe those were in the older models.

Lack of termios was pre Mac OS X, which 

>  In case your 
> computer really supports POSIX serial routines, then you 
> should be able 
> to compile cbmlink for it.  Have a look at 
> <URL:
> />.  You 
> can only include the c2n232.c and serial.c drivers in the compilation.

I should look at SerialSlave first as I already have the hardware ;-)

> By the way, are you interested in the C2N232 device?  The 
> first batch of 
> circuit boards is ready now, but I'm waiting for the needed 
> SMD capacitors.

I'd be interested in testing it; especially as that is a
'boot free' device so there would be no need to re-insert
test disk if I have killed a transfer by (eg from last
night) running an emulator doing full c64+1541 emulation
of a fastloading demo in the foreground while in the
background trying to do a transfer with my (currently
error-recovery free) protocol!

-Virus scanned and cleared ok

       Message was sent through the cbm-hackers mailing list

Archive generated by hypermail 2.1.1.