Bank switching problem

[flashjazzcat]

Right - I've stared at this code long enough. Can someone give me a sense check on this because I think I'm going crazy. I have a loop which calls three routines designed to set up certain locations in extended RAM banks (tab line, pointers, etc). I have a table called BANKTAB which is clearly established to contain the PORTB mask for the MAIN (unbanked) RAM at $4000-$7FFF, followed by the banking values for the extended RAM. BANKTAB begins:

 

F3, CF, CB,C7, C3...

 

These are pre-masked values designed to be written straight to PORTB (OS ROM switched in, BASIC ROM off). This initialisation loop needs to work on all the banks in the list (10 in this case), EXCEPT the main bank (BANKTAB+0) which contains program code. So I have:

 

 LDY BANKS
 BEQ *+3
 DEY
 STY BANK
BANKLP
 ldy bank
 beq notmainbank
 lda banktab,y
 sta portb
 JSR ERASEALL
 JSR DEFTABS
 JSR DEFNAME
 LDA #0
 STA EDITLINE
 STA EDITLINE+1
 DEC BANK
 bne banklp ; don't set up main bank - it's got code in it
;
notmainbank
 lda banktab
 sta portb

 

Trouble is, the banked RAM never gets written to. And no matter where I set breakpoints up here, portb always contains $F3. I just wanted to make sure I wasn't going insane...

[ivop]

 LDY BANKS
 BEQ *+3
 DEY
 STY BANK
BANKLP

 

 

BANKS is a memory location or a constant (ldy #BANKS)? And is it >= 2 ?

Edited February 2, 2010 by ivop

[flashjazzcat]

BANKS is a memory location or a constant (ldy #BANKS)? And is it >= 2 ?

Sorry, BANKS is a memory location. It contains $0A in the tests.

 

Thankfully, the time it took me to write up the plea for help allowed my mind to clear sufficiently that I could diagnose the problem. The ERASEALL routine was calling STA_LINK_Y, which reads thus:

 

sta_link_y
 	pha
 	lda textbankmask
 	sta portb
 	pla
 	sta (link),y
 	pha
 	lda banktab
 	sta portb
 	pla
 	rts

 

It's expected that textbankmask is pre-initialized to the value at BANKTAB+CURRENT_BANK. Of course I'd missed the fact that ERASEALL called this routine, which resets PORTB to the main bank on exit. So that's why DEFTABS and DEFNAME never operated on the extended bank, and portb always contained $F3 at the end of the loop.

 

Thanks for your help there, anyway. It's appreciated.

Edited February 2, 2010 by flashjazzcat