Push instruction assembly arm

Suppose the Program Stack is as follows before we push a new value:

The stack pointer SP points at the top of the program stack

All memory cells that lies below the memory location pointed to by SP are used (= reserved/allocated)

For simplicity , we will only store int typed values on the program stack

(The ARM processor can only push the value in a whole register (= 32 bits) on the stack.

If you push a data that consists of less than 32 bits , then you (as an assembler programmer) need to manage that yourself .

push // Push the value in "regName" onto the program stack

The push instruction in ARM can push multiple registers with a single push instruction , but you must specify the list of registers in a specific order .

main: mov r0, #4 push // Push value in r0 on stack mov r0, #9 push // Push value in r0 on stack

Result after the first push instruction (snapshot from EGTAPI 's stack area)

Notice that the stack pointer register SP (= 0x7affc) is pointing to the top of the program stack

Result after the second push instruction (snapshot from EGTAPI 's stack area)

Notice that the stack pointer register SP (= 0x7aff8) is updated and still points to the top of the program stack

• Prog file: /home/cs255001/demo/asm/8-sub/push.s

How to run the program:

ARM (assembler) instruction to pop an int typed value from the (top of the) program stack :

pop // Pop the int typed value from the top // of the program stack and store it in register "regName"

main: mov r0, #4 push mov r0, #9 push  pop // r1 = 9 pop // r2 = 4

• Prog file: /home/cs255001/demo/asm/8-sub/push.s

How to run the program:

When memory used to stored values of a variable is no longer needed , the memory that was reserved for the variable must be de-allocated (= un-reserved )

Suppose the following is the state of the Program Stack after we have reserved some memory for the variable x :

If we want to de-allocate the memory space used for variable x and discard the value in the de-alocated variable , the resulting stack will look like this:

In this example, we assume that the value in the variable x is no longer needed nor used in future computation.

I.e.: we discard the variable stored in the variable x

• If we only want to de-allocate some space on the syatem stack , there is a more efficient way to do this task (rather than using the pop instruction !)

Example: to de-allocate one int typed variable from the (top of the) program stack , we use:

add sp, sp, #4 // Move program stacktop down 4 bytes // This will remove the reservation of 4 bytes // that was used to stored the int typed variable

Example: to de-allocate two int typed variable from the (top of the) program stack , we use:

add sp, sp, #8 // Move program stacktop down 8 bytes // This will remove the reservation of 8 bytes // that was used to stored TWO int typed variable

The stack pointer SP has 2 important functions :

• SP points to the location in the stack that the push and pop operations will operate

+----------+ | | | | AVAILABLE . (not used) | | SP -->+----------+ | | ^ | | | USED . | | | | Program | | (= reserved . ) | stack | | | | | | | v +----------+ 

When you perform a push operation (= increase the stack size ), you will: