>> >> > static inline unsigned int
>tbase_get_deferrable(tvec_base_t *base)
>> >> > {
>> >> > - return ((unsigned int)(unsigned long)base &
>> >TBASE_DEFERRABLE_FLAG);
>> >> > + return (unsigned int)((unsigned long)base &
>> >TBASE_DEFERRABLE_FLAG);
>> >> > }
>> >...
>> >> The change makes sense, but does it actually "fix" anything?
>> >>
>> >
>> >Yes - this first place fixes logical error, so it's a sin
>> >- even if not punishable in practice. (It's also unnecessary
>> >test for long to int conversion.)
>> >
>>
>> I am sorry, I don't understand. What is the logical error in
>the first
>> one?
>>
>> Actually, your change makes it different from what was originally
>> indended.
>> Original intention was to type convert base to a 32 bit value and
>> bitwise& with FLAG.
>
>But that is not what the original code is doing. If you wanted to
>typecast "base" to "a 32 bit value" then you should've used u32
>instead.
>
>Anyway, if you originally intended to actually typecast "base" to
>unsigned int, then you could do that directly without typecasting it
>first to unsigned long (unnecessarily) and then to unsigned int. Of
>course, if your system implements a pointer as something bigger than
>unsigned int (which is what you eventually convert "base" to), then
>you're screwed anyway and the intermediate typecast to unsigned long
>doesn't buy you anything at all.
On a 64 bit system, converting pointer to int causes unnecessary
compiler
warning, and intermediate long conversion was to avoid that. I will have