Re: [Patch 1/8] Setup

[Balbir Singh <balbir@in.ibm.com>]

On Mon, May 08, 2006 at 02:17:13PM -0700, Andrew Morton wrote:
> Balbir Singh <balbir@in.ibm.com> wrote:
> >
> >  /*
> > + * sub = end - start, in normalized form
> > + */
> > +static inline void timespec_sub(struct timespec *start, struct timespec *end,
> > +				struct timespec *sub)
> > +{
> > +	set_normalized_timespec(sub, end->tv_sec - start->tv_sec,
> > +				end->tv_nsec - start->tv_nsec);
> > +}
> 
> The interface might not be right here.
> 
> - I think "lhs" and "rhs" would be better names than "start" and "end". 
>   After all, we don't _know_ that the caller is using the two times as a
>   start and an end.  The caller might be taking the difference between two
>   differences, for example.
> 
> - The existing timespec and timeval funtions tend to do return-by-value. 
>   So this would become
> 
> 	static inline struct timespec timespec_sub(struct timespec lhs,
> 							struct timespec rhs)
> 
>   (and given that it's inlined, the added overhead of passing the
>   arguments by value will be zero)

Agreed, I will make these changes.

> 
> - If we don't want to do that then at least let's get the arguments in a
>   sane order:
> 
> 	static inline void timespec_sub(struct timespec *result,
> 				struct timespec lhs, struct timespec rhs)
>

-- 

	Balbir Singh,
	Linux Technology Center,
	IBM Software Labs