asm
section .text
    global _sum, _do_the_job

section .data
    mas_size equ 4

_do_the_job:
    push ebp
    mov ebp, esp

    ; eax - pointer to array start.
    ; array of integers, step is 4

    mov edi, [ebp+8]
    mov ebx, [ebp+12]

    ; find minimum
    mov ecx, ebx
    mov esi,0
    mov eax,[edi + esi]

    outer:
        push ecx
        mov ecx, ebx
    inner:
        cmp eax,[edi + esi]
        jle no_need
        mov eax,[edi + esi]
    no_need:
        add esi, 4
        loop inner

        pop ecx
        loop outer

    ; subtract
    mov ecx, ebx
    mov esi,0
    _outer:
        push ecx
        mov ecx, ebx
    _inner:
        sub [edi + esi],eax

        add esi, 4
        loop _inner

        pop ecx
        loop _outer

    mov eax,edi
    ; epilogue
    mov esp, ebp
    pop ebp

    ret


c
#include <stdio.h>
#include <stdlib.h>
#include <time.h>

extern void do_the_job(int* arr, int len);

#define LEN 7

void main()
{
    srand(time(NULL));
    int* a = calloc(LEN * LEN, sizeof(int));
    int i;
    int j;

    for (i = 0; i < LEN; i += 1) {
        for (j = 0; j < LEN; ++j) {
            a[i * LEN + j] = rand() % 98 + 1;
        }
    }

    printf("BeFoRe:\n");

    for (i = 0; i < LEN; i += 1) {
        for (j = 0; j < LEN; ++j) {
            printf("%2d ", a[i * LEN + j]);
        }
        printf("\n");
    }

    do_the_job(a, LEN);

    printf("AfTeR:\n");

    for (i = 0; i < LEN; i += 1) {
        for (j = 0; j < LEN; ++j) {
            printf("%2d ", a[i * LEN + j]);
        }
        printf("\n");
    }
    free(a);
}


sh
clear
../nasm -f elf -o 6.o 6.asm
gcc -c main.c -o main.o
gcc -o main main.o 6.o
./main

run in cygwin x32