MrPickle5
MrPickle5

Reputation: 522

itoa for negative numbers?

Bounty: +50 reputation points for anyone that provides the code to make this subroutine work with negative numbers.

I wrote a MIPS program to convert Fahrenheit to Celsius. It opens its own output window (i.e. UART) and properly displays the value in Celsius. It does all of this while making calls from C to assembly and vice versa. Entire code posted below.

I am struggling with getting it to work with negative numbers. Just drawing a blank right now for some reason. How do I change my itoa function to write this check?

Anyone have any ideas how to get this working with negative values in MIPS?

.ent itoa
itoa:
  // putting the stack frame together
  addiu sp, sp, -16
  sw    fp, 12(sp)
  move  fp, sp
  sw    a0, 16(fp)
  sw    a1, 20(fp)
  sw    s0, 4(fp)
  sw    s1, 0(sp)

  // there is no divide immediate, so using s1
  li    s1, 10

  itoa_div_begin:
        divu  a0, s1
        mfhi  s0
        mflo  a0

        addiu s0, s0, 0x30
        addiu sp, sp, -1
        sb    s0, 0(sp)

        beq   a0, zero, itoa_div_done
        nop

        j     itoa_div_begin
        nop

  itoa_div_done:
  itoa_copy_begin:
        lb    s0, 0(sp)
        sb    s0, 0(a1)

        addiu sp, sp, 1
        subu  s1, sp, fp
        addiu a1, a1, 1

        beq   s1, zero, itoa_copy_done
        nop

        j     itoa_copy_begin
        nop

  itoa_copy_done:
  li    s0, 0
  sb    s0, 0(a1)

  move  v0, a1

  lw    a0, 16(fp)
  lw    a1, 20(fp)
  lw    s0, 4(fp)
  lw    s1, 0(fp)
  // stack frame
  move  sp, fp
  lw    fp, 12(sp)
  addiu sp, sp, 16

  jr    ra
  nop

.end itoa

Upvotes: 5

Views: 1952

Answers (1)

Patrik
Patrik

Reputation: 2692

You can include this in your code, right above itoa_div_begin:

li t0, 0x40000000
and t0, a0, t0
beqz t0, itoa_div_begin
nop
li t0, 45
sb t0, 0(a1)
addi a1, a1, 1
not a0, a0
addi a0, a0, 1

itoa_div_begin:

What is does:

  • Check for a negative number by checking if the MSB is 1
  • If so, insert - in result, and
  • Use the two's complement of the number for the rest of the function

Upvotes: 6

Related Questions