Hi Meindert,
Here's what I've got to learn about Bank-switching (not too much, indeed):
You probably know already, but I'd like to remark that Bank-switching has nothing to do with CPU, they're NOP's. Thus, the only device involved in BS operations is the ROM PAC itself, it MUST produce it's own ENBANK (1-4) OP code, it must decode it, verify that this specific code points to the address page(s)** where this PAC is placed at, and perform the necessary hardware switching.
** When a Bank-Switch OP code is found, and its target page is determined (say for example H'9000) the port's companion page (say H'8000) is also switched.
Now to your question:
Is the Bank-Switching reset to BANK 1 on STAND-BY or SLEEP?
Not exactly, enabled bank is not affected by SLEEP/STAND BY transitions... but there's not need of such reset. In fact, looking at the HP *official* BS scheme as for the ADV PAC (other schemes exist i.e. RAM-Box and HEPAX) I've always found that it turns back to Bank 1 after Bank 2 operation completion:
Example...
Bank 1................Bank 2
Code...
.
.
Code
BS2 Code ->...... Code
........................ Code...
Code......... <-BS1 Code
Code..
.
.
End
So you'll find Bank 1 enabled at such ROM PAC's whenever in Sleep or Stand-by.
Other schemes are manufacturer dependant and may (they do indeed) react in a different way. The RAM-Box referred above is a good example of unbalanced Switching scheme, and this *suggest* (haven't got a *real* RAM-Box to apply my probes to... :-(... and verify this, but the code is pretty straight forward and CAT 2 behavior -read below- is determinant) that the enabled bank 1 or 2 will also remain enabled while in Sleep/Stand-by. So, after performing a "PG10" command you'll see "RAMBOX b" on "CAT 2", even if you turn your 41 off and back on.
I have no info regarding HEPAX BS but apparently it follows the HP "standard" expanding it to 4 banks.
That means that you must implement different scheme behaviors depending on the the module(s) you're gonna replycate...
Please take what stated above as "to the best of my knowledge". I've collected this info as empirical observation on the modules behavior, during Clonix developement process.
A more "documented" advice will be of a great help to all of us :-)
Cheers from the Canaries and...
Meindert, keep the MLDL2000 wheel turning!!
Diego.