Johnny Billquist wrote:
On 2015-11-24
16:35, Al Kossow wrote:
On
11/23/15 11:46 PM, Johnny Billquist wrote:
Your native interface have the additional problem
that in addition to
requiring people to write their own device driver for any OS usage, it
will be rather difficult to get booting from it, since that
require special support.
There is no reason you can't have two simulated controllers, one small
enough and early enough to boot a range of operating systems (RL02?),
then
another which exports a simple block-level interface which would be
simple enough to easily write drivers against.
True. But having your system on an RL02 is still no fun. It's a rather
small disk, which cause some headache for larger systems.
Actually, by V05.05 of RT-11, it was impossible to have all of the
Source Code
files on a single RL02 even just for the RT-11 portion. If you
collected all of
the binary distributions for V05.07 of RT-11 and the layered products, that
would also exceed a single RL02.
RL02 is also
interesting because there was a 22 bit version for qbus.
That's one definition of "interesting". :-)
I'm trying to remember if DSD had extended
block length or partioning
for their controllers
Not sure about it, but what I normally observed was that 3rd party
controllers using large disks for some DEC disk usually put several
logical disks on one physical disk. You could call that partitioning.
For example, the DSD 880/30 (from Data Systems Design of course) emulates
3 RL02 disk drives using a single internal (non-removable) hard drive.
The box
also holds a single RX03 floppy disk drive (8" floppy disk drive which
supports
using single-sided media specified by DEC as an RX02 floppy in addition to
media which have the same physical interface, but which are double-sided).
For a Qbus system, the dual module controller was the interface to both the
three RL02 hard drives and the single RX03 floppy drive. I don't know if
DSD also made a separate controller for the Unibus for the DSD 880/30.
With regard to the address support by the controller for the Qbus, the
floppy
drive definitely supported only an 18-bit address. That 18-bit ONLY support
by DSD was identical to the 18-bit support that DEC provided for its Qbus
controller for the RX02, so both DEC and DSD needed a bounce buffer
managed by software to support the RX02 floppy disk for systems with more
than 256 KB of physical memory.
As for DSD support for the RL02 for a 22-bit buffer address, a quick look
at the DSD manual was not able to say one way or the other. However,
it seems more likely the the DSD controller for the RL02 supported ONLY
an 18-bit address. I have all the DSD hardware, but it is not operational
at this point. If anyone else has experience with the DSD controller for
the emulated RL02, let us know if there was 22-bit address support for
its emulated RL02 drive.
Jerome Fine
