PowerPC Tutorial

Previous Chapter

Chapter 19: Pointers

There will be times where you will need to set a register to an address of a piece of data that resides in your program.

Example:

my_string:
.asciz "Hi this is a string"
.align 2
...
...
...
...
lis r3, my_string@h
ori r3, r3, mystring@l

We've learned about this in the previous chapter in regards to the Assembler Directives. But understand that in the above code, r3 is the Pointer to my_string. It does NOT contain the string itself.

What is a Pointer?

Pointers are well... anything that points to something. That can be a byte of data, a string, or to another pointer. Pointers will always be some sort of Memory Address, this is important to remember!!! Again, a pointer is ALWAYS a Memory Address.

Another example:

pointer_to_byte:
.byte 0xFF
.align 2
...
...
...
...
lis r3, pointer_to_byte@h
ori r3, r3, pointer_to_byte@l

pointer_to_byte is the Label to where the 0xFF byte is located in the Program. We can use the Label as a Pointer.

An important thing we need to discuss is the la instruction

Load Address:
la rD, SIMM (rA) = addi rD, rA, SIMM

Recall back in Chapter 8 that I prefer beginners to use ori over addi. However, you won't ever see the lis+ori used "in the wild". So instead of writing...

lis r3, pointer_to_byte@h
ori r3, r3, pointer_to_byte@l

or even this..

lis r3, pointer_to_byte@ha #Must be ha for the chance that we need sign-extension
addi r3, r3, pointer_to_byte@l

we instead write this....

lis r3, pointer_to_byte@ha
la r3, pointer_to_byte@l

Most code authors will use this style of code. Also, the Assembler prefers this as well. Here's an example of incorporating the la instruction for a pointer that points to a string.

pointer_to_string:
.asciz "Hello PowerPC fanatic!"
.align 2 #Need this so PPC instructions can be aligned
...
...
...
lis r3, pointer_to_string@ha
la r3, pointer_to_string@l (r3) #r3 contains the address (pointer) of the string

There will be times where you will need the Pointer of a Pointer, like this....

pointer_to_bytepointer:
.long bytepointer #.long directive writes out a word-size value. All pointers are word-sized since they are memory addresses

bytepointer:
0xFF
.align 2
...
...
...
lis r3, bytepointer@ha
la r3, bytepointer@l #r3 will contain a pointer that contains the byte pointer

If you wanted just the pointer to bytepointer from the above code, you can do this...

pointer_to_bytepointer:
.long bytepointer

bytepointer:
0xFF
.align 2
...
...
...
lis r3, bytepointer@ha
lwz r3, bytepointer@l #Notice the use of lwz instead of la

Next Chapter

Tutorial Index