What are pages? (In the context of PIC programming) / Computer Science / Forums

Forums

4hv.org :: Forums :: Computer Science

« Previous topic | Next topic »

What are pages? (In the context of PIC programming)

Move Thread

LAN_403

The Zander

Wed Nov 22 2006, 11:46PM

Registered Member #295 Joined: Tue Mar 07 2006, 06:59PM
Location:
Posts: 23

i'm debugging a large lookup table based program and it has been brought to my attention that if my program if longer than the 8-bits contained in the program counter (PCL), a simple "addwf PCL" type lookup table may not function correctly in all cases, and it has somthing to do with pages (which i assume are simply higher order bits of the PCL). i would be greatful if someone would help me understand this concept.

the lookup tables i am reffering to are of this form

addwf pcl, 1
retlw 'data'
retlw 'data'
.
.
.

and i am working with a PIC16F877A

Thanx
-TheZander

Bjørn

Thu Nov 23 2006, 01:53AM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

Yes, the most significant 8 bit of the program counter is held in a separate latch called PCLATH, when you update PCL it automatically reads PCLATH at the same time so it is safe to update PCLATH first since it will not affect the program counter.

So if your table crosses a 256 byte page boundary or if you don't know if it does then you need code like this to calculate the value for PCLATH:

;**** Look up 7 segment data in the table and output it to the display  
        movwf	offset7seg
        movlw 	LOW ledLUT 	 ;Get low 8 bits of address
        addwf 	offset7seg	  ;Do an 8-bit add operation
        movlw 	HIGH ledLUT 	;Get high 5 bits of address
        btfsc 	STATUS,C 	   ;Page crossed?
        addlw 	.1 		     ;Yes then increment high address
        movwf 	PCLATH 		 ;Load high address in latch
        movf 	offset7seg,W 	;Load computed offset in w reg

ledLUT  addwf   pcl, 1
        retlw   'data'
        retlw   'data'

The Zander

Thu Nov 23 2006, 04:57AM

Registered Member #295 Joined: Tue Mar 07 2006, 06:59PM
Location:
Posts: 23

thank you for the code snippet, i think iunderstand the instructions, let me insure my assumptions are correct. offset7seg is just the (address-1) of the first entry in LUT. it would also make sense that LOW ledLUT, and HIGH ledLUT are, combined, the adress of the entry in the LUT (1 being the first entry). am i on the right path so far? anywayz, my question is how do i go about determining the addresses, should i just count the lines starting from the org? can MPLAB help me with this?

Thanx
-TheZander

Bjørn

Thu Nov 23 2006, 06:48AM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

You are on the right track.

offset7seg is the offset into the table, 0 points to the first data item.
ledLUT marks the start of the table, it is at the addwf pcl instruction instead of the first data item because the PC automatically increments by 1, so if offset7seg = 0 then you will have the first item returned when the instruction after the add is run.

LOW ledLUT returns the 8 least significant bits of the address of the table, HIGH ledLUT returns the most significant byte, so all is atomatic, no counting needed. Remember that labels are 16 bit constants calculated at assembly time, HIGH and LOW are just assembler directives.

Remember that this code must be entered by using a CALL instruction or the table would not know where to return to. Just load it in a simulator and single step through it, that should make everything clear.

Also using instruction of this type "addwf pcl, 1" is not very intuitive. "addwf pcl" is the form that was originally though of by the designer before Microchip started messing with it.

The Zander

Fri Nov 24 2006, 12:44AM

Registered Member #295 Joined: Tue Mar 07 2006, 06:59PM
Location:
Posts: 23

very cool, i had no idea that "high" and "low" were compiler directives. i assume they work in combination with any operand that would normally have a label as a parameter. i also didn't realize that i could use lables as literal values (without explicitly declaring them using "equ").

i have never used the MPLAB simulator before, but i'll check it out, looks as if it might become an indispensable tool.

one more question, can i use the call instruction to call functions from pages other than the one the "call" instruction is located in. and does "lcall" have the same effect as call (i'm thinking obsolete instruction)?

thank you for the insight
-TheZander

Bjørn

Fri Nov 24 2006, 04:17AM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

Yes, you can use the HIGH and LOW directives everywhere you have constants. It is very useful for example when you calculate constants at assembly time.

The datasheet for 16F870 says that the call instruction uses 11 bits for the address so it will reach 2048 instructions or 8 pages. If you need to reach longer you need to use macros like lcall or manually update PCLATH. Remember that different PICS are slightly different so read the datasheet for your exact model.

I have made a simple version of FORTH using assembler directives:
http://sciencezero.org/computing/download/pforth10.inc

I have figured out how to make a 16 bit optimizing FORTH compiler using only assembler directives but I have not bothered to make it yet. So the assembler is pretty powerful when you manage to figure out how everything works.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob