long bit field problem

[BasePointer]

Hi,

I'm using CW 5.1 with MC9S08LC60.

Code:

typedef struct { unsigned char R0:1; unsigned char R1:1; unsigned char R2:1; unsigned char R3:1; unsigned char R4:1; unsigned char R5:1; unsigned char R6:1; unsigned char R7:1; unsigned char R8:1; unsigned char R9:1; unsigned char R10:1; unsigned char R11:1; unsigned char R12:1; unsigned char R13:1; unsigned char R14:1; unsigned char R15:1;  unsigned char R16:1; unsigned char R17:1; unsigned char R18:1; unsigned char R19:1; unsigned char R20:1; unsigned char R21:1; unsigned char R22:1; unsigned char R23:1; unsigned char R24:1; unsigned char R25:1; unsigned char R26:1; unsigned char R27:1;  unsigned char R28:1; unsigned char R29:1; unsigned char R30:1; unsigned char R31:1;} TMaskBits;typedef union { unsigned long Val; TMaskBits Bit;} TMASK;

 

I'm declaring a variable:

Code:

TMASK tmp;tmp.Val = 531968; // 0b10000001111000000000

If I test tmp.Bit.R12, I see it is not 1.

If I test tmp.Bit.R19, I see it is not 0.
But all they should be.

What do you thing about the problem here? Is this a bug?

Here is the result:

Code:

R0 -> 0R1 -> 0R2 -> 0R3 -> 0R4 -> 0R5 -> 0R6 -> 0R7 -> 0R8 -> 0R9 -> 0R10-> 0R11-> 1R12-> 0R13-> 0R14-> 0R15-> 0R16-> 0R17-> 1R18-> 1R19-> 1R20-> 1R21-> 0R22-> 0R23-> 0R24-> 0R25-> 0R26-> 0R27-> 0R28-> 0R29-> 0R30-> 0R31-> 0

 

[Lundin · Accepted solution]

You can't use the type long for bitfields.

Problems with and suggestions about bitfields can be found here:

http://forums.freescale.com/freescale/board/message?board.id=CWCOMM&message.id=3751

Message Edited by Lundin on 2007-02-2008:52 AM
Alban fixed link.

Message Edited by Alban on 2007-02-20 08:39 AM

[BasePointer]

Hi Lundin,

This is a endianism problem. I changed the byte order of bitfield. And I saw it worked like I expected.

Code:

typedef struct { unsigned char R24:1; unsigned char R25:1; unsigned char R26:1; unsigned char R27:1;  unsigned char R28:1; unsigned char R29:1; unsigned char R30:1; unsigned char R31:1;

 unsigned char R16:1; unsigned char R17:1; unsigned char R18:1; unsigned char R19:1; unsigned char R20:1; unsigned char R21:1; unsigned char R22:1; unsigned char R23:1;

 unsigned char R8:1; unsigned char R9:1; unsigned char R10:1; unsigned char R11:1; unsigned char R12:1; unsigned char R13:1; unsigned char R14:1; unsigned char R15:1;

 unsigned char R0:1; unsigned char R1:1; unsigned char R2:1; unsigned char R3:1; unsigned char R4:1; unsigned char R5:1; unsigned char R6:1; unsigned char R7:1;} TMaskBits;

 

[Lundin]

Yes, as mentioned in my post, the endianess of bitfields is not specified in the standard, nor is the result you get from using long. That code will not be portable.