Re: [PATCH 0/8] Fragmentation Avoidance V17

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

On Sat, 15 Oct 2005, Paul Jackson wrote:

> Mel wrote:
> > +#define __GFP_USER       0x80000u  /* User and other easily reclaimed pages */
> > +#define __GFP_KERNRCLM   0x100000u /* Kernel page that is reclaimable */
>
> Sorry, but that __GFP_USER name is still sticking in my craw.
>
> I won't try to reopen my quest to get it named __GFP_REALLY_REALLY_EASY_RCLM
> or whatever it was, but instead will venture on a new quest.
>
> Can we get the 'RCLM' in there.  Especially since this term appears
> naked in such code as:
>

__GFP_USERRCLM is the original name and the motivation for changing it to
__GFP_USER no longer exists. However, __GFP_EASYRCLM would also make sense
as it flags pages that are (surprise surprise) easily reclaimed. Would
__GFP_EASYRCLM be the best choice? __GFP_USERRCLM may imply to some
readers that it is reclaimed by the user somehow. The flags would then be;

__GFP_EASYRCLM - Allocations for pages that are easily reclaimed such as
userspace and buffer pages

__GFP_KERNRCLM - Allocations for pages that may be reclaimed by the kernel
such as caches

No flag - Page cannot be easily reclaimed by any mechanism.

I would be happy with __GFP_USERRCLM but __GFP_EASYRCLM may be more
obvious?

> > -				page = alloc_page(GFP_HIGHUSER);
> > +				page = alloc_page(GFP_HIGHUSER|__GFP_USER);
>
> where it is not at all obvious to the reader of this file (fs/exec.c)
> that the __GFP_USER term is commenting on the reclaim behaviour of
> the page to be allocated.
>
> I'd be happier with:
>
> > +#define __GFP_USERRCLM    0x80000u /* User and other easily reclaimed pages */
> > +#define __GFP_KERNRCLM   0x100000u /* Kernel page that is reclaimable */
>
> and:
>
> > -				page = alloc_page(GFP_HIGHUSER);
> > +				page = alloc_page(GFP_HIGHUSER|__GFP_USERRCLM);
>
> Also the bold assymetry of these two #defines seems to be without motivation,
> one with the 'RCLM', and the other with '    ' four spaces.
>
>

-- 
Mel Gorman
Part-time Phd Student                          Java Applications Developer
University of Limerick                         IBM Dublin Software Lab
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/