I had a bunch of 8bit packets I had to separate into opcode and operand. I have pulled out the first 2 bits for the opcode and I have the 6 bit operand saved to a Int.
I need to be able to work with negative number in two's compliment. Is there a quick way I can convert the 6 bit operand from twos compliment into a negative Integer?
So it would convert 34 into -30. I can detect the sign at the beginning with a bit mask. but from there I'm lost.
Two's Complement
Code:#include <stdio.h> int six_bit_sign_extend(unsigned char b) { b &= 0x3f; // ensure b is less than 64 return b < 32 ? b : 64 - b; } int main() { unsigned char b = 34; printf("%d\n", six_bit_sign_extend(b)); return 0; }
Nice one thanks... I must have tried about 4 different methods and each one didnt work.
I made a mistake. The sign needs to be reversed:
Code:#include <stdio.h> int six_bit_sign_extend(unsigned char b) { b &= 0x3f; // ensure b is less than 64 return b < 32 ? b : b - 64; } int main() { unsigned char b = 34; printf("%d\n", six_bit_sign_extend(b)); return 0; }
Or for a different approach:
Code:int six_bit_sign_extend(unsigned char b) { b &= 0x3f; return b & 0x20 ? b | ~0x3f : b; }
Last edited by cstryx; 11-23-2016 at 05:20 PM.
Or if you wanted to avoid the conditional:
Looking up the bit-twiddling hack yields this:Code:int six_bit_sign_extend(unsigned char b) { unsigned char s = b & 0x20; // sign bit return (signed char)(s << 1 | s << 2 | (b & 0x3f)); }
Code:int six_bit_sign_extend(unsigned char b) { struct {signed int b: 6;} s; return s.b = b; }
Last edited by algorism; 11-23-2016 at 07:15 PM.
