CPSC 2310 - Exam 3

How many general purpose registers are there in x86-64?

16

How many general purpose registers did the 8086 have?

8

Why are transistors foundational for computers?

They are small, tough, power efficient, and can turn on and off billions of times per second

What does compilation from C do?

Translates C to assembly, then assembly to machine code

What is the program counter?

Holds the memory address of the next instruction to be executed

What is the program counter called in assembly?

%rip

What is the integer register file?

16 named locations containing 64 bit values, holding the register addresses used for C pointers or integer data

What is the function of the integer register?

Keeps track of the program state and hold temporary data (function arguments, local variables, and return values)

What do condition code registers do?

Hold status information about the most recently executed arithmetic or logical instructions

What is the range of size for x86-64 instructions?

1 - 15 bytes

Disassemblers need access to source code or assembly code of the program. True or false?

False

What is the 2 byte (16 bit) data type called in x86-64?

Word

What is the 4 byte (32 bit) data type?

Double word ("l")

What is the 8 byte (64 bit) data type?

Quad word ("q")

What are the 2 formats for floating point data in x86-64?

Single (4 bytes) and double (8 bytes) precision

What is the fastest type of memory?

Registers

What is %rbp?

The base pointer to the bottom of the stack

What is %rsp?

The stack pointer

What is a calling convention?

Governs how functions on a particular architecture and OS interact, ensuring functions compiled by different compilers can interoperate, allowing OS to run code from different programming languages and compilers

What do calling conventions regulate?

- Placement of function arguments

- Where return values go

- What registers a function may use

- How local variables are allocated

What is the caller?

A function that calls another function

What is the callee?

The function being called

How many registers are set aside in x86-64 for function arguments? What happens if you have more arguments than registers?

6; they are pushed onto the stack

What is the return register in x86-64?

%rax

How do calling conventions handle function parameters being structures?

They use the ABI (application binary interface) to determine how to handle it

What are the two most used ABIs?

- 64-bit Microsoft ABI

- x86-64 System V ABI (Linux, macOS, BSD)

How does Microsoft ABI handle function parameters being structures?

All structures are passed by value and pushed onto the stack.

How does System V ABI handle function parameters being structures?

An algorithm is used to determine how to handle it. If the struct cannot be stored in the registers, it'll be pushed to the stack

What are the steps required to call a function sometimes called?

The entry sequence

What are the steps required to return sometimes called?

The exit sequence

What does the caller do in the entry sequence?

- Store the first 6 arguments in registers

- If there are more than 6 arguments, push the rest to the stack.

- Save any caller-saved registers

- Execute the callq function

What does the callee do in the exit sequence?

- Place the return value in %rax

- Restore the stack pointer (%rsp) to its value at entry

- Execute 'retq' (like pop %rip), which removes the return address from the stack and jumps to that address

- Clean up exit space it prepared for arguments and restore caller-saved registers

What are the three operand possibilities?

- Immediate

- Register

- Memory

What are immediates?

Constant values written using $, followed by an integer or number (e.g. $-577 or $0x1F)

What are registers (as an operand)?

An arbitrary register to be referenced in the instruction

What is memory (as an operand)?

An effective (or computed) address that must be accessed in the instruction

What does mov do?

Copy data from one location to another

x86-64 allows move instructions between memory and memory. True or false?

False, memory-memory is not allowed

How does x86-64 handle memory to memory movement?

It uses two instructions:

- Load the source value into a register

- Write the registers value to the destination

What does memory-memory movement look like in x86-64?

1. movq -4(%rbp), %rax

2. movq %rax, -16(%rbp)

What type of move instruction is?:

movl $0x4050, %eax

Immediate - Register

What type of move instruction is?:

movw %bp, %sp

Register - Register

What type of move instruction is?:

movb (%rdi, %rcx), %al

Memory - Register

What type of move instruction is?:

movb $-17, (%r10)

Immediate - Memory

What type of move instruction is?:

movq %rax, -12(%rbp)

Register - Memory

What is movabsq used for?

64-bit immediate data

What is special about movabsq?

You can only have an immediate as a source operand

What is the suffix for this move instruction?

mov_ %eax, (%rsp)

l

3 multiple choice options

What is the suffix for this move instruction?

mov_ (%rax), %dx

w

3 multiple choice options

What is the suffix for this move instruction?

mov_ $0xFF, %bl

b

3 multiple choice options

What is the suffix for this move instruction?

mov_ (%rsp, %rdx, 4), %dl

b

3 multiple choice options

What is the suffix for this move instruction?

mov_ (%rdx), %rax

q

3 multiple choice options

What is the suffix for this move instruction?

mov_ %dx, (%rax)

w

3 multiple choice options

Where is this moving?

mov_ %eax, (%rsp)

Register - Memory

Where is this moving?

mov_ (%rax), %dx

Memory - Register

Where is this moving?

mov_ $0xFF, %bl

Immediate - Register

Where is this moving?

mov_ (%rsp, %rdx, 4), %dl

Memory - Register

Where is this moving?

mov_ (%rdx), %rax

Memory - Register

Where is this moving?

mov_ %dx, (%rax)

Register - Memory

What is wrong with this move instruction?

movb $0xF, (%ebx)

%ebx cannot be an address (it is 32 bits)

What is wrong with this move instruction?

movl %rax, (%rsp)

Moving 64 bits - 64 bits is not a movl, it is a movq

What is wrong with this move instruction?

movw (%rax), 4(%rsp)

This is memory-memory, which is not allowed

What is wrong with this move instruction?

movb %al, %sl

%sl doesn't exist

What is wrong with this move instruction?

movq %rax, $0x123

Immediates cannot be a destination

What is wrong with this move instruction?

movl %eax, %dx

This should be movw because of %dx

What is wrong with this move instruction?

movb %si, 8(%rbp)

%si is 2 bytes (16 bits), so this should be a movw

What type of extension is used when assigning long = unsigned char?

Zero extension, since we go from 1 byte -> 8 bytes and the right hand side is unsigned

What type of extension is used when assigning unsigned long = short?

Signed extension, since we go from 2 bytes -> 8 bytes and the left hand side is signed

What type of extension is used when assigning short = char?

Signed extension, since we go from 1 byte -> 2 bytes and both sides are signed

What does the movz instruction do?

Move with zero-extension

What does the movs instruction do?

Move with signed extension

What is unique about the stack in assembly?

It grows downward

What does %rsp hold?

The address of the top of the stack element (the most recent item pushed onto the stack)

What do you need to do to push a new quad word onto the stack?

- Decrement the stack pointer by 8

- Write the new value at the new top-of-stack address

What does push1 %rax do?

- subq $8, %rsp - decrement the stack pointer

- movq %rax, (%rsp) - store the value of %rax on the stack

What do you need to do to pop a quad word off the stack?

- Read from the top-of-stack address

- Increment the stack pointer by 8

What does popq %rdx do?

- movq (%rsp), %rdx - read from %rsp and store in %rdx

- addq $8, %rsp - increment stack pointer

What is leaq?

It is a variant of movq, where it reads from memory to a register, and doesn't dereference memory. It copies the effective memory address to the destination

Given %rax = x, %rcx = y:

What is the result of leaq 6(%rax), %rdx?

%rdx = 6 + x

Given %rax = x, %rcx = y:

What is the result of leaq (%rax, %rcx), %rdx?

%rdx = x + y

Given %rax = x, %rcx = y:

What is the result of leaq (%rax, %rcx, 4), %rdx?

%rdx = x + 4y

Given %rax = x, %rcx = y:

What is the result of leaq 7(%rax, %rax, 8), %rdx?

%rdx = 7 + x + 8x = 7 + 9x

Given %rax = x, %rcx = y:

What is the result of leaq 0xA(, %rcx, 4), %rdx?

%rdx = 10 + 4y

Given %rax = x, %rcx = y:

What is the result of leaq 9(%rax, %rcx, 2), %rdx?

%rdx = 9 + x + 2y

What does x86-64 do for multiplying 8 bytes?

The result is 128 bits, which is an oct word. The higher order 64 bits will be stored in %rdx, and the low order 64 bits wil be stored in %rax.

What does x86-64 use for unsigned 8 byte multiplication?

mulq

What does x86-64 use for signed 8 byte multiplication?

imulq

Where does x86-64 store the result of dividing 8 bytes?

The quotient is stored in %rax, and the remainder is stored in %rdx

What does x86-64 use for unsigned 8 byte division?

divq

What does x86-64 use for signed 8 byte division?

idivq