Different length for instructions and data? about customasm HOT 5 CLOSED

Hi! I'm glad you're finding it useful!

Could you please post an example program, showing the problem and your workaround? I'd like to make sure I'm getting the problem right. I'm not sure how your program bank integrates with the data bank -- there are two address spaces?

from customasm.

In order to make the most out of 8 bit addresses, there are 4 address spaces: instruction low, instruction high, data and stack (the stack isn't important since you don't program it directly).
This means that even though addresses are technically 2 byte long, the PC ony gets increased by one when jumping to the next one.


Note that the output position in the file (outp column) is correct. The only problem is the wrong address, that causes jumps to go to twice the correct position,

Using #bits 16 means that the data bank stops working completely, so I used #bits 8 and came up with this solution:
Instead of defining a jump like this: J {Addr8: u8} -> { 0xA0 @ Addr8[7:0] }
I define it like this: J {Addr8: u9} -> { 0xA0 @ Addr8[8:1] }, so the address gets divided by 2.
Also in bankdef for the program I have to set the #size as twice the actual size.
The ony drawback is that when you type the address instead of using a label it doesn't work (because the number you type will get divided by 2), but you shouldn't be doing that anyways.

Thank you for your time and hard work 👍

from customasm.

The issue seems to have been solved in v0.11 with the ability to set #bits for each bank. However, it looks like it doesn't work unless I type the directive every time I want to change banks:

#bank data
#bits 8
...

#bank program
#bits 16
...

Is this a bug or intended behaviour? I think the #bank directive should set the #bits automatically if you already specified it in the bankdef

from customasm.

Oops, sorry! I've fixed this in v0.11.1. Could you let me know if it's now working for you?

from customasm.

Yes, now it's working perfectly. Thank you!

from customasm.