Rewrite the base exception handler that comes with SPIM (described in App A.7) to look for arithmetic overflow exception, and for an arithmetic overflow exception inspect the arithmetic instruction that caused the exception through EPC; modify the contents of the dest register of this instruction as follows. If the sign of the current value in dest register Rd is 1 (remember 1 with an overflow means that the actual result was meant to be positive) then write Rd 0x7fffffff, else Rd 0x80000000. Try out your new exception handler with SPIM on a small program that forces an arithmetic overflow. [15 points]

.ktext 0x80000180

mov $k1, $at # Save $at register

sw $a0, save0 # Handler is not re-entrant and cant use

sw $a1, save1 # stack to save $a0, $a1

# Dont need to save $k0/$k1

mfc0 $k0, $13 # Move Cause into $k0 srl $a0, $k0, 2 # Extract ExcCode field andi $a0, $a0, Oxf bgtz $a0, done # Branch if ExcCode is Int (0) mov $a0, $k0 # Move Cause into $a0

mfco $a1, $14 # Move EPC into $a1

jal print_excp # Print exception error message

done: mfc0 $k0, $14 # Bump EPC

addiu $k0, $k0, 4 # Do not re-execute faulting instruction

mtc0 $k0, $14 # EPC mtc0 $0, $13 # Clear Cause register

mfc0 $k0, $12 # Fix Status register

andi $k0, Oxfffd # Clear EXL bit

ori $k0, Ox1 # Enable interrupts

mtc0 $k0, $12

lw $a0, save0 # Restore registers

lw $a1, save1

mov $at, $k1 eret # Return to EPC

.kdata

save0: .word 0

save1: .word 0