Re: [PATCH] introduce setup_timer() helper

[Arjan van de Ven <arjanv@redhat.com>]

On Sun, Sep 18, 2005 at 07:51:20PM +0400, Oleg Nesterov wrote:
> Arjan van de Ven wrote:
> > 
> > >                               unsigned long data)
> > > +{
> > > +     timer->function = function;
> > > +     timer->data = data;
> > > +     init_timer(timer);
> > > +}
> > 
> > are you sure you want to do this in this order???
> > I'd expect the init_timer to be first...
> 
> I think it does not matter from correctness point of view.

right now.. it probably doesn't.
However I think conceptually, touching a timer before init_timer() is just
wrong. For one... it would prevent init_timer() from being able to use
memset() on the timer. Which it doesn't today but it's the kind of thing
that you don't want to prevent happening in the future.

> 	setup_timer(timer, expr1(), expr2())
> 
> it is better to initialize ->func and ->data first, otherwise
> the compiler should save the results from expr{1,2}, then call
> init_timer(), then copy these results to *timer.

I don't see how that is different....