On 2013-04-11 22:48, Mark Pizzolato - Info Comm wrote:
On Thursday, April 11, 2013 at 1:37 PM, Johnny Billquist wrote:
On 2013-04-11 22:14, Mark Pizzolato - Info Comm wrote:
On Thursday, April 11, 2013 at 7:00 AM, Bob Supnik wrote:
1. Four Massbus channels on the PDP-11 (or VAX). This is straightforward.
The RH simulator already supports two, and the extensions from two to
four use the existing mechanisms:
- add the register blocks definitions to the header file
- define additional RHC, RHD devices with matching unit and register
definitions
- look for places where "2" is hardwired as the maximum number
- modify the autoconfiguration table for the additional device
addresses
This is most of it, but additional devices and units must also be defined in
the pdp11_rp and pdp11_tu which then need to be connected to the new
RHC and RHD devices created above. Maybe a way to dynamically assign
particular rp or tu devices to a particular Massbus is also be necessary
(someone might want 3 disk controllers and 1 tape, while others might want
3 tape and 1 disk controller, etc.). If mixed configurations are possible
something needs to validate that there are no Massbus conflicts.
Isn't there already? I mean you should not be able to attach two different
disk drives with the same unit number. And that is the only kind of conflict
you can have.
From an RSX point of view, even having an RM and an RP disk on the same
Massbus is considered a mixed Massbus. Like I said before. RM and RP are
not the same from the host point of view, and should most likely not be
considered the same from simh point of view either. But that is another
issue.
Conflicts, as such, can only be about the unit number.
Ignoring mixed massbus or not and the details of what unit the OS refers to a
device as, the current pdp11_rp device is structured to have a total of 8 disk
drives. These 8 drives can be a combination of RP or RM drives. The real goal
of this exercise (what I believe the original poster was asking for) was to
expand the number of total RP and RM disks which the guest OS can have access
to. I'm pretty sure that such an expansion would take more than expanding the
unit array sizes in the pdp11_rp code.
Good point. I don't know the innards, so I can't comment on how much
work there would be to add more massbuses.
This is also where I find the flexibility in simh a bit lacking. I would
like to define both buses and devices, and not only devices, with simh
trying to decide which buses various things should be connected to.
But I don't know how hard it would be to change all that.
(There are more things I'd like to fool around with as well, such as
what unit numbers disks attached to certain ports should have, what port
on controllers actually even have anything attached to them, what
addresses controllers respond to, and so on and so on...)
I mean, as long as we're making wish lists, can't I also? :-)
Johnny
_______________________________________________
Simh mailing list
[email protected]
http://mailman.trailing-edge.com/mailman/listinfo/simh
