I have written the following piece of code which MISRA does not like:
UartPtr->C &= ((uint8_t)(~SIO_C2_SBK));
with
#define SIO_C2_SBK ((uint8_t)0x01u)
and UartPtr
is defined as
UartPtr = (UartStruct*) 0x12345678; /* I know that this is also a violation of MISRA */
with the underlying datastructure:
typedef volatile struct UartStructTag
{
uint8_t BDH;
uint8_t BDL;
uint8_t C1;
uint8_t C2;
} UartStruct;
My Misra checker complains about the first line and states, that
An integer constant expression with negative value is being converted to an unsigned type.
However, the following line does not yield into a problem with MISRA:
UartPtr->C |= ((uint8_t)(SIO_C2_SBK));
So the problem comes from the bitwise negation. But as all operations are directly casted to uint8_t, i do not get the violation of the MISRA standard. Who wants to help me here?
In any arithmetic expression, values of types smaller than int
are implicitly converted to int
before they are processed. The C language cannot do arithmetic on types smaller than int
. Thus, your code actually behaves like this:
UartPtr->C &= ((uint8_t)(~(int)(uint8_t)0x01u));
which is just
UartPtr->C &= ((uint8_t)(~1));
where ~1
has the value -2
on two's complement architectures.
To fix this issue, convert to unsigned
or any other unsigned type larger than int
before applying bitwise not:
UartPtr->C &= ((uint8_t)(~(unsigned)SIO_C2_SBK));