Add two RLCA then full 64 bytes table can used!

It depends on the remaining free bytes for the kernel area... In this case I would even save two RRA as currently I am doing
AND #78:RRA:RRA:RRA (bits 3-6 contain the bank number)

If you want I can write routine for the loader program which are allocate the memory and fill up the lookup_table.

I am really not sure if eg ZT stores the time from RTC

ZT write the RTC time to the EXOS time variables.

Thanks for the offer! I will write you as soon as the port is in a more advanced state.

Regarding RTCs, are there different ones existing for the EP? Of course the best solution would be to poll the RTC in SymbOS directly like it is done on the CPC and MSX. But the solution with receiving the start time from EXOS is probably ok, too. Sure you will loose some seconds after several floppy disc accesses, but that shouldn't be too bad.

I have a quick question again regarding memory configuration:
- I am now planning to put bank 0 to #f8,#f9,#fa,#ff
- on a standard EP128 bank 1 would be #fb,#fc,#fd,#fe and all additional banks can be anywhere

Now my question:
- what's about EP64 machines, which are expanded to 128K or more? Do they always have RAM at #f8-#fb or can it by anywhere as well, while #f8-#fb isn't available?

I am only asking because the 16K blocknumbers for bank 0 are hardcoded in the code for optimization. If I have to be flexible here as well, I will add a patch routine during the initialization.

Regarding RTCs, are there different ones existing for the EP?

Only one type, using MC146818P compatible RTC chip (which are also used on PC). But can be located on different I/O ports, but this is can be setuped, but I think the autodetect also not to be a hard problem (my clock program write to the CMOS memory and verify it, if ok then the clock chip exist)

Sure you will loose some seconds after several floppy disc accesses, but that shouldn't be too bad.

I think if you also use the Dave 1Hz IRQ for clock, then no problem.

I am now planning to put bank 0 to #f8,#f9,#fa,#ff

At the EXOS compatible mode the FFh needed to be preserved

- what's about EP64 machines, which are expanded to 128K or more? Do they always have RAM at #f8-#fb or can it by anywhere as well, while #f8-#fb isn't available?

Can be at anywhere. Most of expansions developed for the EP128 then don't make F8-FBh area.

Ok, good to know. The 16K blocks of bank 0 can now be located anywhere

Great!

Yes, I already discussed this with Trebmint, the creator of SymStudio (PC-based IDE for SymbOS), and he really wants to have higher resolutions for the EP, too!
As you said for the beginning the two CPC screens will be used, as it's less complicated to have the screen in 16K. And yes, on the PCW I had the same problem with the screen beeing larger than 16K, so I already have some techniques and experiences for switching between 16K during low level graphic operations.
720x288x2 and 360x288x4 sounds really good, though we will have to stay within 8bit for the Y resolution, otherwise it would require huge changes in some of the low level routines to increase it to 16bit. That means we have 720x256x2 and 360x256x4 in the end, which is 22,5KB. I think that's quite ok.
In any case we will have 2 versions of SymbOS EP then, the base version which runs on 128K machines, and the extended version for the higher screen resolutions, which requires 192K or more. This is necessary as the memory map of the OS will be a little bit different, but it's not a problem at all, there are just two different "make" files.