stirringhalo
stirringhalo

Reputation: 53

Passing Arguments C -> NASM -> C

admittedly this is some homework help, but a specific problem I can't seem to get past.

I'm trying to write a program which takes a string in of hex characters, calls an assembler function which gives me the decimal value of the hex string. That assembler function calls a "checker" function in C which makes sure each character is a legal HEX value.

My question is, how do I take an EBX register in assembler and properly pass it to a C function expecting a character. I can't seem to properly pass from assembler back to C. Am I accidentally passing a pointer here? I also can't seem for the life of me to get an individual character out of EBX even by breaking it up into bytes.

Note, -1 is returned when the character is invalid.

What I'm hoping for:

Please enter a maximal 4 digit hex integer using a string of hex digits: FBE You entered: FBE FBE - F - 15

What I get: Please enter a maximal 4 digit hex integer using a string of hex digits: FBE You entered: FBE FBE - M - -1

EDIT: The check digit function according to the assignment must only take individual characters. So I will be breaking up the string in the main NASM function for the full functionality. Still trying to get it to work with once character at a time..

C:

#include <stdio.h>
#include <string.h>

int main(void)
{  
    char  input[255];
int   dec_value;

while (1)
{
    printf ("Please enter a maximal 4 digit hex integer using a string of hex digits: ");
    scanf ("%s",input);
    if (strlen(input) <= 4)
    {
        break;
    }
    printf ("The string is too long!\n");
}

printf ("You entered: ");
printf ("%s\n",input);
extern int hex2dec(char[]);
dec_value = hex2dec(input);
printf ("%i",dec_value);
if (dec_value == -1) {
    printf ("There's an invalid character!\n");
}
else {
    printf ("Decimal value of character %s is:%d \n", input, dec_value); 
}       
return 0;
}

int checkdigit (char  hex)
{
    printf (" - %c - ", hex);
    if ( (hex <= 70 && hex >= 65) || (hex >= 48 && hex <= 57) ) {
        if ( hex >= 65 ) {
            printf ("Letter");
            return ( (int) (hex-'A'+10 ));
        }
        else {
            printf ("Number");
            return hex - 48;
        }
    }
    return -1;
}

NASM:

segment .data
segment .text
global  hex2dec
extern  checkdigit, printf

hex2dec:        
    push    EBP
    mov     EBP,ESP
    push    EDX
    push    EBX

    mov     EDX,0D         ; 0 EDX
    xor     EBX,EBX
    mov     EBX, DWORD [EBP+8]    ; copy the string to EDX

    push    EBX 

    call    printf      ; print whole string
    call    checkdigit     ; pass character to interpret

    add     ESP,4              ;on return clear the stack,                           
    ;the value is in EAX
    pop     EBX     ;restore EBX        
    pop     EDX     ;restore EDX
    pop     EBP 
    ret

Upvotes: 3

Views: 5409

Answers (2)

Multimedia Mike
Multimedia Mike

Reputation: 13216

Chris Dodd is correct-- send a char (8-bit byte) instead of a pointer (32-bit quantity).

So far, you don't seem to be doing anything with EDX except clearing it. Also, you don't need to clear EBX to 0 before loading its value from the stack (same as writing "a=12;a=65;"-- the first assignment is irrelevant because it is thrown away immediately).

Anyway, so you have loaded a pointer to the string into EBX. Now load the 8-bit byte that EBX points to. The syntax for this is [EBX], as such:

mov EDX, [EBX]

But doing that will load 4 bytes (because EDX is a 32-bit register). You only want the first byte, so specify a target register of the lower 8 bits of EDX (DL):

mov DL, [EBX]

It's a good thing you already cleared EDX to 0 (because the above instruction only overwrites the bottom 8 bits). At this point, EDX contains the byte you want to process, so push EDX on the stack instead of EBX.

I hope this has expanded your general understanding of x86 assembly.

Upvotes: 1

Chris Dodd
Chris Dodd

Reputation: 126140

You're passing the argument to hex2dig (which is a char *) to checkdigit (which expects a char). You need to actually load a character into a register and then push that register onto the stack to pass a char to checkdigit

Upvotes: 3

Related Questions