ARMv8-A Architecture

Question 1

Which type of Architecture is ARMv8? (RISC or CISC)

Answer 1

RISC

Question 2

Is the ARMv8 a Load/Store Machine?

Answer 2

Yes

Question 3

How many registers does the Register file contain?

Answer 3

Register file contains 31 64-bit registers

Question 4

Which data is manipulated by most instructions in the registers?

Answer 4

32-bit or 64-bit data

Question 5

What are the 2 execution states of ARMv8 architecture?

Answer 5

AArch64 and AArch32

Question 6

What does the AArch64 execution state use?

Answer 6

Uses the A64 Instruction set and 64-bit registers

Question 7

What does the AArch32 execution state use?

Answer 7

Uses the A32 or T32 Instruction sets (provides computability with older ARM and THUMB instruction sets using 32-bit registers)

Question 8

What are the ARMv8’s 4 exception levels

Answer 8

EL0 - for normal user apps with limited privileges EL1 - for the OS kernel EL2 - for a Hypervisor EL3 - Low-Level firmware

Question 9

EL0

Answer 9

• Restricted access to a limited set of instructions and registers - Most programs work at this level

Question 10

EL1

Answer 10

• Privileged access to instructions, registers and memory - Accessed indirectly by user programs using system calls - Lots of System registers accessible only in EL1 and used in OS kernel code

Question 11

EL2

Answer 11

• Support virtualization, where the computer hosts multiple guest OSs, each on its own virtual machine

Question 12

EL3

Answer 12

• Include the Secure Monitor

Question 13

How many General-purpose registers does AArch64 have? What are they numbered as? When using all 64 bits, use? When using only the lower-order 32 bits, use?

Answer 13

31 64-bit wide general-purpose registers Numbered from 0 to 30 x or X (eXtended) w or W (Word)

Question 14

Register uses: x0-x7 x8 x9-x15 x16-x17

Answer 14

x0 - x7: to pass args into a procedure and return results x8: indirect result location register x9 - x15: temp registers x16 - x17: intra-procedure-call temp registers (IP0, IP1)

Question 15

Register uses: x18 x19-x28 x29 x30

Answer 15

x18: platform register x19 - x28: callee-savee registers (value is preserved by any function you call) x29: frame pointer (FP) register x30: procedure link register (LR)

Question 16

Special-Purpose Registers: (Use/Types)

Answer 16

Stack Pointer - Points to top of run-time stack SP: 64-bit WSP: 32-bit Zero Register - Gives 0 value when read from and discards value when written to XZR: 64-bit WZR: 32-bit Program Counter - Holds address of currently executing instruction / Can’t be accessed directly as a named register / Is changed indirectly by branch and other instructions / Is used implicitly by PC-relative load/stores PC: 64-bit

Question 17

How many floating-point registers are there?

Answer 17

32 128-bit wide floating-point registers

Question 18

What is an immediate value and how can it be used?

Answer 18

An immediate value (a constant) may be used as final source operand or some instructions

Question 19

Which symbol is optional when using gcc?

Answer 19

A # symbol can prefix the immediate, but is optional when using gcc

Question 20

How many constants are allowed

Answer 20

It depends on the particular instruction and how many bits are available within the machine instruction

Question 21

Prefixes to prevent immediates to be assumed as decimals

Answer 21

Hexadecimal: 0x Octal: 0 Binary: 0b

Question 22

Why use aliases?

Answer 22

mov x29, sp is alias for add x29, sp, 0 provided for readability and programmer convenience

Question 23

Commonly used instructions

Answer 23

• Move immediate mov Wd, #imm32 [-(2^31) to +(2^32) - 1] mov Xd, #imm64 [-(2^63) to +(2^64) - 1] - Move register mov Wd, Wm [alias for orr Wd, wzr, Wm] mov Xd, Xm [alias for orr Xd, xzr, Xm]

Question 24

What is Branch and Link instruction used for? In what are args and return value?

Answer 24

To call a function (library or your own) Arguments put into x0-x7 before call Return value is in x0

Question 25

Main routine of a program structure

Answer 25

Question 26

What does .global main do?

Answer 26

Makes the label "main" visible to the linker The main() routine is where execution always starts

Question 27

What does ... [sp, -16] do?

Answer 27

Allocates 16 bytes in stack memory (in RAM) Does this by pre-incrementing the SP register by 16

Question 28

What does stp x29, x30, ... do?

Answer 28

**_St_**ores the contents of the **_p_**air of registers to the stack Saves the state of the registers used by calling code

Question 29

What does SP do after stp?

Answer 29

SP points to the location in RAM where we write to

Question 30

What does mov x29, sp do? What is the FP used as?

Answer 30

Updates FP to the current SP FP may be used as a base address in the routine

Question 31

What does ldp x29, x30, ... do?

Answer 31

**_L_**oa**_d**_s the _**p_**airs of registers from RAM Restores the state of the FP and LR registers

Question 32

What does the SP do after ldp?

Answer 32

SP points to the location in RAM where we read from

Question 33

What does ..., [sp], 16 do?

Answer 33

Deallocates 16 bytes of stack memory Does so by post-incrementing SP by +16

Question 34

What does ret do?

Answer 34

Returns control to calling code (in OS) Uses the address in LR

Question 35

Addition registers/operands

Answer 35

1 DR, 2SO

Question 36

Subtraction registers/operands

Answer 36

1 DR, 2SO

Question 37

Multiplication registers/operands/immediates Aliases

Answer 37

1 DR, 2 or 3 SR No immediates allowed mul Wd, Wn, Wm Alias for madd, Wd, Wn, Wm, wzr

Question 38

Multiply-add registers/operands

Answer 38

madd Wd, Wn, Wm, Wa

Question 39

Multiply-subtract registers/operands

Answer 39

msub Wd, Wn, Wm, Wa Calculates: Wd = Wa – (Wn \* Wm)

Question 40

Multiply-negate registers/operands

Answer 40

mneg Wd, Wn, Wm, Wa Calculates -(Wm \* Wa)

Question 41

Division registers/operands/immediates Forms

Answer 41

1DR, 2SR Does **integer** division No immediates allowed Signed-form sdiv Wd, Wn, Wm Operands are signed integers Unsigned-form udiv Wd, Wn, Wm Operands are unsigned integers

Question 42

To get remainder or Modulus, do?

Answer 42

numerator - (quotient \* denominator) Use msub

Question 43

What does dividing by 0 do?

Answer 43

Does not generate an exception but instead writes 0 to DR

Question 44

Printing to Standard Output

Answer 44

printf() Invoked with 1 or more args First is the format string (usually literal) Rest corresponds to placeholders in the string

Question 45

What does a branch instruction do? What happens to PC?

Answer 45

A branch instruction transfers control to another part of the program PC is not incremented as usual but is set ot th ecomputed address of an instruction (corresponds to the value of its label) b *label* is an unconditional branch that's always taken

Question 46

What are condition flags used for? SIze? What does 0 and 1 mean?

Answer 46

To store information about the result of an instruction Are single bits in the CPU 0 = false, 1 = true

Question 47

What are the 4 condition flags and when are they true?

Answer 47

* Z: true if result is **_z_**ero * N: true if result is a **_n_**egative * V: true if result o**_v_**erflows * C: true if result generates a **_c_**arry out

Question 48

How are condition flags set?

Answer 48

Condition flags are set by instructions that end in "s" (**_s_**et flags) e.g. sub**_s_**, add**_s_**

Question 49

What is cmp an Alias for?

Answer 49

Form (64-bit): cmp Xn, Xm Is an alias for: subs wzr, Xn, Xm

Question 50

What are *Conditional branch* instructions used for?

Answer 50

Conditional branch instructions use the condition flags to make a decision e.g b.eq If Z is true, branches Otherwise drops through to the following instruction

Question 51

Diagram for signed integers b.cc where cc is the condition code

Answer 51

Condition codes for signed integers are

Question 52

How are loops formed?

Answer 52

By branching from the bottom of the loop to the top

Question 53

What type of loop is a do loop and its format

Answer 53

A do loop is a post-test loop (loop body executed at least once)

Question 54

How can you optimize code for a pre-test loop?

Answer 54

Move the test to the end of the loop It reduces the loop by one instruction Branch does not use complemented logic (idk) Must branch to the test the first time though

Question 55

How can a for loop be formed?

Answer 55

Convert it to a while loop

Question 56

How can you make the If construct

Answer 56

Is formed by branching over the statement body if the condition is not true Uses logical complement b. lt \<-\> b.ge b. le \<-\> b.gt b. eq \<-\> b.ne

Question 57

If-else construct

Answer 57

Question 58

Using gdb debugger Start a program under debugger control Set breakpoint Run program Continue to next breakpoint Single Step ni display/i $pc p$reg q

Answer 58

gdb myprogram b *label* * r* * c* * si* * ni* (also executes next instruction but if a function call, proceeds until the function returns) shows current instruction when single stepping (do before running program) prints contents of register (Can append a *format* character: Signed decimal: p/d Hexadecimal: p/x Binary: p/t) q