TIP: Using Unions for Manipulating Bits
This is simple custom data type that allows you to manipulate bits, bytes, or whatever comes into a 32 bit integer. It uses unions for this purpose because all data types within a union share the same memory. This can be very useful for low level purposes such as setting flags without using masking.
Here is the custom type union comprising one DWORD.
typedef struct _BITS8
{
unsigned char bit0 : 1;
unsigned char bit1 : 1;
unsigned char bit2 : 1;
unsigned char bit3 : 1;
unsigned char bit4 : 1;
unsigned char bit5 : 1;
unsigned char bit6 : 1;
unsigned char bit7 : 1;
}BITS8;
typedef struct _QUATERS
{
unsigned char quater0 : 4;
unsigned char quater1 : 4;
}QUATERS;
typedef union _BITSQUATERS
{
BITS8 bits;
QUATERS quaters;
unsigned char byte;
}BITSQUATERS;
typedef union _BITS32
{
//all 4 bytes - represented as 8 bits and 2 quaters
BITSQUATERS byte[4];
short word[2]; //two 16 bit int
int dword; //32 bit int
}BITS32;
The structure BITS8 represents single byte as 8 separate bits you can access! If you have a variable of type char and you wish to change some of the bits within it, then all you have to do is take its address, cast it to BITS8 * and then you can change whichever bit you like.
char c = 0; BITS8 *pBits8 = (BITS8 *)&c pBits8->bit4 = 1;//set bit4 of variable c
| bit7 | bit6 | bit5 | bit4 | bit3 | bit2 | bit1 | bit0 |
The way struct BITS8 represents one byte in memory
If you would want to change the the high an low quater of varialble c, then cast its address to QUATERS * just like this:
char c=0; QUATERS *pQuaters=(QUATERS *)&c pQuaters->quater0 = 0xC; //set the low quater of c pQuaters->quater1 = 0xD; //set the high quater of c
| quater1 | quater0 |
The way struct QUATERS represents one byte in memory
If you want to set bits, quaters or the whole byte, then use the BITSQUTERS * pointer this way:
char c=0; BITSQUATERS *pBitsQuaters =(BITSQUATERS *)&c pBitsQuaters->byte = 0xFF; //set the whole byte pBitsQuaters->quaters.quater0 = 0xC; //set the low quater of c pBitsQuaters->quaters.quater1 = 0xD; //set the high quater of c pBitsQuaters->bits.bit7 = 0; //set bit7
| BYTE | |||||||
| quater1 | quater0 | ||||||
| bit7 | bit6 | bit5 | bit4 | bit3 | bit2 | bit1 | bit0 |
The way union BITSQUATERS represents one byte in memory
Here is how to set up dword, low and high word, separate bytes and bits as well — using a BITS32 * pointer:
int i = 0; BITS32 * pbits32 = (BITS32 *) &i pbits32->dword = 0x33221100; //set all 32 bits pbits32->word[0] = 0xcc11; //set low int16 of dword pbits32->byte[0].byte = 0xff; //set first byte pbits32->byte[0].quaters.quater0 = 0xF;//set first byte low quarter pbits32->byte[3].bits.bit7 = 1; //set third byte bit7
| DWORD | |||||||||||||||||||||||||||||||
| WORD1 | WORD0 | ||||||||||||||||||||||||||||||
| byte3 | byte2 | byte1 | byte0 | ||||||||||||||||||||||||||||
| quater1 | quater0 | quater1 | quater0 | quater1 | quater0 | quater1 | quater0 | ||||||||||||||||||||||||
| bit7 | bit6 | bit5 | bit4 | bit3 | bit2 | bit1 | bit0 | bit7 | bit6 | bit5 | bit4 | bit3 | bit2 | bit1 | bit0 | bit7 | bit6 | bit5 | bit4 | bit3 | bit2 | bit1 | bit0 | bit7 | bit6 | bit5 | bit4 | bit3 | bit2 | bit1 | bit0 |
The way union BITS32 represents one double word in memory
Hope that will come in handy! Enjoy

Comments
Endian considerations
Posted by foxmuldr on 02/24/2010 01:02amDepending on the machine, big-endian or little-endian compatibility issues could also arise. As long as you're dealing with a single machine it should be okay, but if you're designing code to interact with other systems it will likely break.
ReplyJust be aware of this
Posted by S_M_A on 07/02/2009 03:34pmThe packing order for bitfields is not standardized i.e. this type of code is not very portable. To further complicate things, using unsigned chars in bitfields isn't according to the standard either so changing compiler might also make the code break.
ReplySome code examples would be nice to have
Posted by CBasicNet on 06/09/2009 09:51pm