Thread: incorrect result in program

I have this simple program below:

Code:

#include <stdio.h>
#include <limits.h>

unsigned int rightrot(unsigned int x, unsigned int n)

{

   /* calculate number of bits in type */

   size_t s = sizeof(x) * CHAR_BIT;

   size_t p;



   /* limit shift to range 0 - (s - 1) */

   if(n < s)

       p = n; 

   else

       p = n % s;



   /* if either is zero then the original value is unchanged */

   if((0 == x) || (0 == p))

       return x;



   return (x >> p) | (x << (s - p));

}





int main(void)

{

   unsigned int val;

   unsigned int pos;

   val = 0xFF94;
   pos = 5;

   printf("%u\n", rightrot(val, pos));

}

The result it prints is 2684356604 on my 32-bit computer. The result I expect is as follows:

0xFF94 is 0000000000000000 1111111110010100 in binary.

Shift it right by 5:
0000000000000000 0000011111111100

Then take that result in shift it right by 27 (s is 32 and p is 5, so the difference is 27):
1111111110000000 0000000000000000

Now we use bitwise or:
0000000000000000 0000011111111100 | 1111111110000000 0000000000000000 = 1111111110000000 0000011111111100

That in decimal is 4286580732. So how does it come up with 2684356604?

Surely you want to isolate the bits being shifted out (with a bitwise &) and only | those with the intermediate result? (1U << n) - 1 is probably your mask