A bit is just an eighth of a byte. Which bits serve what purpose is implementation specific and should be included in the documentation for whatever platform with which you are working (you may want to take a look at the IA-32 manuals from Intel for some examples, x86's have several control and status registers).
I assume you want to learn about bit manipulation techniques (how to set a bit, test a bit, clear a bit, etc).
There's really nothing to it, you basically just use an appropriate bitwise operator and bitmask for your situation.
For example, if you have a a byte, and you want to set bit 5 (I am referring to bit 0 as the least significant bit and bit 7 as the most significant bit) while leaving the other bits unchanged, you simply OR the byte with the bitmask 0x20 (0b00100000). Since there are 0's everywhere but the 5th bit, this operation will leave all the other bits unchanged (1 OR 0 is 1, 0 OR 0 is 0) while setting bit 5 (1 OR 1 is 1 and 0 OR 1 is 1).
It is generally considered bad practice to use "magic numbers" as bitmasks. You usually see people construct bitmasks using bitshift operators. For example, the above bitmask would be written as (1 << 5) (0b00000001 shifted left 5 times). Any modern compiler will optimize this back to a magic number in the machine code.
So here's how you would perform each of the four bit operations in C:
Code:
unsigned char myByte;
//set bit 5
myByte |= (1 << 5);
//clear bit 5 (Note how we need the negation of the previous bitmask here in order to clear bit 5 while leaving the rest unchanged.
//We need ones everywhere except for the bit we are trying to clear, because 1 AND 1 = 1 and 1 AND 0 = 0.)
myByte &= ~(1 << 5);
//Toggle bit 5
myByte ^= (1 << 5);
//Test bit 5
if(myByte & (1 << 5)) { printf("bit 5 set!"); }
That's pretty much all there is to it. C also has bitfields which lets you name specific sets of bits in a larger structure.
15 Jul 2011, 5:24 pm
