Re: [RFC][patch 02/12] remove clocksource inline functions

[Martin Schwidefsky <schwidefsky@de.ibm.com>]

On Wed, 29 Jul 2009 18:27:54 +0200
Martin Schwidefsky <schwidefsky@de.ibm.com> wrote:

> On Wed, 29 Jul 2009 16:36:46 +0100
> Will Newton <will.newton@gmail.com> wrote:
> 
> > On Wed, Jul 29, 2009 at 4:32 PM, Martin
> > Schwidefsky<schwidefsky@de.ibm.com> wrote:
> > 
> > > Hmm, you have an object of type struct clocksource and you do
> > > cs->read(cs). If that is not clear enough then I don't know what is. We
> > > do that all over the place in the linux kernel. And I personally find
> > > these useless wrappers rather annoying. I don't like to have to jump to
> > > another place to find out that it just calls the read function of the
> > > object.
> > 
> > An argument for the helper is that it eases grepability. Sure you can
> > search for "->read" but that's going to turn up all kinds of
> > non-clocksource code as well. grep clocksource_read will get you
> > exactly what you want.
> 
> That would make sense if the clocksource_read calls are littered all
> over the kernel source. But they are not, the only user is
> timekeeping.c
 
And What I find odd as well is e.g. this:

offset = (clocksource_read(clock) - clock->cycle_last) & clock->mask

A mixture of a wrapper to get something out of the clock and a direct
access to other fields of the clocksource. Either it is wrapped or it
isn't, no? A proper wrapper would do all of the above and not just the
call over the read function pointer.


-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.