Re: [PATCH][delayacct] Fix the timespec_sub() interface (was Re: [Patch 1/8] Setup)

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

On Wed, May 10, 2006 at 03:24:49AM -0700, Andrew Morton wrote:
> Balbir Singh <balbir@in.ibm.com> wrote:
> >
> > Please find the updated patch, which changes the interface of timespec_sub()
> > as suggested in the review comments
> > 
> > ...
> >
> >  /*
> > - * sub = end - start, in normalized form
> > + * sub = lhs - rhs, in normalized form
> >   */
> > -static inline void timespec_sub(struct timespec *start, struct timespec *end,
> > -				struct timespec *sub)
> > +static inline struct timespec timespec_sub(struct timespec *lhs,
> > +						struct timespec *rhs)
> >  {
> 
> I'd have thought that it would be more consistent and a saner interface to
> use pass-by-value:
> 
> static inline struct timespec timespec_sub(struct timespec lhs,
> 						struct timespec rhs)
> 
> It should generate the same code.
> 
> I mentioned this last time - did you choose to not do this for some reason,
> or did it just slip past?

Sorry, that definitely slip past.

I'll send another update

	Balbir Singh,
	Linux Technology Center,
	IBM Software Labs