---

The ARM processor from the point of view of an assembler programmer

---

• Introduction
  • Recall the structure of the CPU (= Central Processing Unit, or processor) :

  • When you write assembler programs, you will need to use/access the registers within the CPU (processor):

    ---

  • Each register has a unique name

    Unfortunately, each type of CPU (processor) has different names for its registers

    Example:

    CPU Name of some of the registers ------------------- ------------------------------- Intel processor rax, rbx, rcx, rdx, ... ARM processor r0, r1, r2, r3, ...

    When you write an assembler program, you will need to use the registers in the CPU inside your assembler program

    Because the registers' names are different for different processor type, the assembler programs written for one type of processor cannot be compiled and run on another type of processor

    Fortunately:

    Assembler programming techiques are similar regardless of CPU type So the assembler programming techiques that you will learn in CS255 can be apply to any assembler language When you need to write assembler programs for another processor, all you need to find out are: The registers of that processor           The assembler nmemonics of that assembler language

    ---

    ---

  • So the first step in learning assembler programming in the ARM assembler language, you will need to know the CPU structure of the ARM processor (CPU)

  ---

  ---

  ---

  ---

• CPU structure of the ARM processor
  • The ARM processor (CPU) has the following internal structure:

    The ARM processor has 12 general purpose registers with the following register names:

    R0 (or R0 the register names are not case senstive) R1 R2 ... R10 R11 -- the register R11 is designed as the frame pointer (FP) register which will be explained (very) later in this course

    The ARM processor has the following special purpose registers with the following register names:

    R12 or the Intra Procedure (IP) Call register The IP register is used as temporary (sratch) register. We won't need to use this register in CS255. R13 or the Stack Pointer (IP) register The stack pointer (SP) register contains the address of the top of the system stack We will learn to use the SP register when we discuss how to write function calls in assembler programming R14 or the Link Register (LR) The Link Register (SP) contains the return address of the calling program We will learn to use the LR register when we discuss how to write function calls in assembler programming R15 or the Program Counter (PC) The Program Counter (PC) contains the address of the next program instruction We have explained the use of the Program Counter in this webpage (lecture): click here) CPSR (= Current Program Status Register) The CPSR register contains indications ("flags") of the status of the current instruction execution For the assembler programmer, the following flags (indications) are especially important: N flag: this flag (bit) is set (= 1) if the result of the computation is a negative value              Otherwise, this flag (bit) is reset (= 0) Z flag: this flag (bit) is set (= 1) if the result of the computation is equal to 0 (zero)              Otherwise, this flag (bit) is reset (= 0) V flag: this flag (bit) is set (= 1) if the result of the computation has caused overflow (later)              Otherwise, this flag (bit) is reset (= 0) C flag: this flag (bit) is set (= 1) if the result of the computation produced a carry              Otherwise, this flag (bit) is reset (= 0) As an assembler programmer, you do not need to keep track of the status of these 4 flags. These 4 flags are used when we compare 2 values To compare the values x and y, we compute the subtraction x − y. The setting of these 4 flags will tell us if: x > y x >= y x < y x <= y x == y or x != y This topic will be discussed later when we discuss how if-statements in Java is executed in assembler language

  ---

  ---

  ---

  ---

• Recommended (or commonly used) register names for assembler programming
  • Although you can use the registers names R0, R1, R2, ...., R15 to identify the registers in the ARM processor, the UNIX assembler (= compiler) also recognize the following commonly used register names:

    ARM register Commonly used register name in assembler programming --------------- ------------------------------------------------------ R0 r0 or R0 R1 r1 or R1 R2 r2 or R2 R3 r3 or R3 R4 r4 or R4 R5 r5 or R5 R6 r6 or R6 R7 r7 or R7 R8 r8 or R8 R9 r9 or R9 R10 r10 or R10 R11 fp (frame pointer) R12 ip (intra procedure) R13 sp (stack pointer) R14 lr (link register) R15 pc (program counter)

    Assembler programs will not use the CPSR register directly and will therefore not refer to the CPSR register by a register name.

  ---

  ---

  ---

  ---

• EGTAPI and the registers in the ARM processor
  • The EGTAPI tool is specially designed to help you experience assembler programming using the ARM processor

    Part of the "experience" is showing all the information that you need to see the assembler programming executing inside the (ARM) computer.

    The values in the ARM registers are part of these information

    If you look at the EGTAPI window:

    The values in the (general and special purpose) registers are shown in the left panel:

    You can see all the ARM registers are displayed in the left panel For the CPSR register (shown as PSR), only the N,Z,V,C flags are relevant for us.

  ---

  ---

  ---

  ---

• How to use the general purpose registers (of any CPU)
  • The general purpose registers are primarily used like a scratch pad to store values used in computations

    Because the ALU (Arithmetic and Logic Unit) is the computation circuit of the CPU and the ALU takes inputs from the general purpose registers, a rule of thumb in assembler programming is:

    When you want to perform computation on some value, you must first get that value in one of the general purpose registers !!! For example: In Java, you would write: z = x + y; when you want to add the values in x and y and store the sum in z This Java statement is excuted by the computer using the following assembler instructions (I give them now in English-description, later, we will learn how to write them in ARM assembler code): Get the value of x in register R0 Get the value of y in register R1 Add R0 and R1 and store result in R0 Put the value in R0 in variable z

  ---

  ---

  ---

  ---

• Closer look at the general purpose registers
  • Each general purpose register in the ARM processor consists of 4 bytes:

    ---

  • Recall that high level programming language has different data types, e.g.:

    byte (8 bites) short (16 bits) int (32 bits) float (32 bits)

    Since a high level programming language has different data types (and data type can have different operand sizes), the ARM processor must support operations on different data types with its registers

    Part of this support is storing the values of the different types in its registers (for computation).

    Because of the different type sizes, some part of a register must not be used.

    Here is how the ARM processor store the values depending on the size of the data type: