Wednesday 7 August 2013

Compile 32 Bit Binaries On 64 Bit Machine

Well I had this special need if you recall my previous blog post since my friend had 64 bit machine. Sometimes, there might be this necessity to compile 32 bit binaries on your 64 bit machine. This post describes how to do so.

First make sure the necessary x86 libraries are installed. We require 32-bit shared libraries for AMD64 to compile binaries in 32 bit format. The command below installs the i386 version of libc6-dev:

$ sudo apt-get install libc6-dev-i386

Now you can compile your code in 32 bit binary format using the -m32 flag where 32 represents the x86 processor (-m64 would mean x64 processor).

$ gcc -m32 -o test test.c

I hope this helps :)


Pointers Without Pointer Variables

Since a pointer variable is nothing but a variable holding 4 bytes memory address (at least on 32-bit addressing), I had a thought that non-pointer variables which can hold 4-bytes of data can be used in place of pointer variables. This post shows how this can be achieved.
The code example below uses an unsigned integer variable in order to store memory addresses to point the integer array.


int main(int argc, char **argv)
        int num[] = {1, 2, 3, 4, 5};
        unsigned int ptr;
        int i;
        ptr = (unsigned int) num;
        for (i = 0; i < 5; i++)
                printf("%p - %d\n\n", (void *) ptr, *(int *)(ptr));
                ptr = ptr + 4;
        return 0;

The same concept can be used to use non-pointer variable for pointing other datatypes. After all, its about correct type-casting and since 4 bytes datatype can hold memory addresses, pointer is not always necessary. It must be noted that the increment would be different for different datatypes. Since integer requires 4 bytes, ptr is incremented in this example. If we had character array, then ptr would have to be increased by 1 since char type requires 1 byte.

However, pointers are there to make our life easy. It was just for fun :)