[sebhc] H17 ROM and disk info

[Walter Moore]

> I was already thinking about having a custom driver for a large capacity
> pc-hard-disk.  But using one to operate on H17 disk images is intriguing!

As long as you don't physically boot from the H-17, there is no problem
mounting/reading/writing them.  Just get the interleave correct!  As for
your "pc-hard-disk", I was going to write a quick block device driver which
would transfer sectors between the PC and my H-8 (or H89).  I don't really
care about speed; I just want to have the PC be another disk.  I was going
to do this over one of the serial ports.  9600 baud would be good enough for
me (see note at bottom).  It would be very easy to have the disk be any size
that fits within HDOS' limits for maximum tracks, sectors, sides.  I was
even thinking of making every sector just another file, part of some big
tree.  It would be easy to have a PC-side app which could combine "sectors"
into files, and turn files into "sectors" on the disk.  Another approach
would be to simply have a "cluster" which would be HDOS' max, so that each
directory entry could map to a real file in the PC space.  Lots of ways to
do it.

> >What is the goal of having "H-17" emulation in the emulator?
>
> Well, to be able to boot and run original, unmodified software.  :-)

99.99% of software should run unmodified.  If the software was written to
run on an H-8, H-89 or H-90 with any of the controllers (including second
source products), it will run.  The only thing that cannot run would be an
original HDOS distribution disk.  You could get one to run by trapping the
actual device vectors and performing the appropriate action.  Then you
wouldn't even have to write a small device driver.  Just trap the read
vector and do a read.  If control never gets into the H-17 ROM code, you
don't have to emulate the device.  I like the thought of trapping the
dispatch vectors - it's even easier.

> >Now, I can tell you from personal experience that HDOS 2.0 will run
> without
> >an H-17 installed in the system.
>
> How did you boot it?  Or are you just talking about a system with a
> DIFFERENT controller, like a -47?

Exactly.  The H-47 doesn't depend on an H-17 and it doesn't supply ROM
routines either.  I believe HDOS patched the vectors to those routines like
multiply (not positive) so that they were there.

> >...I need to power the system back up and look at the files supplied with
> >HDOS 2.0
>
> Or .. check the HDOS 2.0 sources in the "sources" section of the archives.
> :-)


I have the source listings; they just are not convenient right now.  Plus I
want to see what came on the distribution disks.

> >This can be done the quickest with a small system device
> >driver were the emulator can recognize a couple of addresses to trap on
> and
> >perform a sector read/write, or even through a couple of unused I/O port
> >address which are emulated.
>
> So you're still emulating a controller ... just a very smart "virtual"
> one.


Yep, I guess so - One a lot easier to emulate.

..walt


A long time ago I was trying to run a couple of serial ports at I believe
19.2K.  I was having problems missing interrupts.  Turns out that even if
you turn off front panel updating, all the tick handling on the H-8 took
something on the order of 700 clock cycles.  If the front panel update was
enabled, every 32nd tick took 2000 (more?) cycles.  No way you could do fast
serial I/O.  I wrote a stub to do the minimum possible for the tick counter
which was pretty small.  Note that this disabled the front panel.  Just a
note for anyone wanting really fast serial I/O





--
Delivered by the SEBHC Mailing List