Re: Address of second TMSCP controller

[Johnny Billquist <bqt%update.uu.se@localhost>]

On 2021-01-29 15:21, Paul Koning wrote:
> > On Jan 28, 2021, at 3:27 PM, Johnny Billquist <bqt%Update.UU.SE@localhost> wrote:
> >
> > On 2021-01-26 14:06, Nigel Johnson wrote:
> >
> > > > Thanks, Mouse.  Yes, as an FE I used to have to figure out those floating assignment from charts - some DEC systems (RSTS, RT11) could not be told where to look - they had to be in the right place. ...
> >
> > ...
> > I think RSTS/E made more use of the probing at runtime. Not sure about RT-11.
>
> On RSTS/E:
>
> 1. At boot time it would scan the bus looking for addresses that respond, and assign them to devices according to the floating address rules.
>
> 2. You could always explicitly set the address, overriding the float handling for that device.  So no, it is not correct that RSTS "could not be told where to look".

Thanks for clarifying that, Paul. I didn't know one way or another, so I 
tried to stay very neutral.

> 3. There also exist "floating vector" rules, which in a similar fashion would guide what interrupt vector to assign to a device.  I don't know if any OS actually used those.  RSTS definitely did not.  Instead, at boot time it would "poke" each identified device to cause it to interrupt, which would tell it what the vector is rather than having to assume.  Two exceptions: programmable interrupt devices would simply be assigned a vector that was available, and the CR11 card reader would be assumed to be at its standard vector since there is no way to make it interrupt (short of attempting to read a card, which isn't a good answer).

Yeah. Same with RSX. Except it only probed during (again) SYSGEN. After 
that, it was assumed that they didn't change. So they are all defined in 
the system image, and are just assumed to be there.
The only probing that happens is that in RSX-11M-PLUS, at boot, it is 
checked if the CSRs respond, and if not, the device is never 
transitioned to online when devices in general are attempted to be 
brought online. And you can change both CSR address and vector of a 
device at runtime. But that requires that you first set the device 
offline, then change, and then bring it online again.

11M is more primitive. It automatically tries to bring devices online, 
and if they don't respond, they are just permanently offline. No tool 
available to change things dynamically.

  Johnny

--
Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt%softjar.se@localhost             ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol