Re: [RFC] Replace the watermark-related union in struct zone with a watermark[] array

[Mel Gorman <mel@csn.ul.ie>]

On Mon, Apr 27, 2009 at 02:04:03PM -0700, David Rientjes wrote:
> On Mon, 27 Apr 2009, Mel Gorman wrote:
> 
> > > I thought the suggestion was for something like
> > > 
> > > 	#define zone_wmark_min(z)	(z->pages_mark[WMARK_MIN])
> > > 	...
> > 
> > Was it the only suggestion? I thought just replacing the union with an
> > array would be an option as well.
> > 
> > The #define approach also requires setter versions like
> > 
> > static inline set_zone_wmark_min(struct zone *z, unsigned long val)
> > {
> > 	z->pages_mark[WMARK_MIN] = val;
> > }
> > 
> > and you need one of those for each watermark if you are to avoid weirdness like
> > 
> > zone_wmark_min(z) = val;
> > 
> > which looks all wrong.
> 
> Agreed, but we only set watermarks in a couple of different locations and 
> they really have no reason to change otherwise, so I don't think it's 
> necessary to care too much about how the setter looks.
> 
> Adding individual get/set functions for each watermark seems like 
> overkill.
> 

I think what you're saying that you'd be ok with

zone_wmark_min(z)
zone_wmark_low(z)
zone_wmark_high(z)

and z->pages_mark[WMARK_MIN] =
and z->pages_mark[WMARK_LOW] =
and z->pages_mark[WMARK_HIGH] =

?

Is that a significant improvement over what the patch currently does? To
me, it seems more verbose.

> I personally had no problem with the union struct aliasing the array, I 
> think ->pages_min, ->pages_low, etc. are already very familiar.
> 

Can the people who do have a problem with the union make some sort of
comment on how they think it should look?

Obviously, I'm pro-the-current-patch :/

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab