Re: [PATCH] vm - swap_prefetch-15

[Con Kolivas <kernel@kolivas.org>]

On Fri, 7 Oct 2005 21:31, Pekka Enberg wrote:
> Hi,
>
> On 10/7/05, Con Kolivas <kernel@kolivas.org> wrote:
> > Good point, thanks! Any and all feedback is appreciated.
>
> Well, since you asked :-)
>
> > +/*
> > + * How many pages to prefetch at a time. We prefetch SWAP_CLUSTER_MAX *
> > + * swap_prefetch per PREFETCH_INTERVAL, but prefetch ten times as much
> > at a + * time in laptop_mode to minimise the time we keep the disk
> > spinning. + */
> > +#define PREFETCH_PAGES()     (SWAP_CLUSTER_MAX * swap_prefetch * \
> > +                                     (1 + 9 * laptop_mode))
>
> This looks strange. Please either drop the parenthesis from PREFETCH_PAGES
> or make it a real static inline function.

I have seen this sort of macro style before in the kernel where () just makes 
it clear that it is a function but a real static inline is ok with me.

> > +/*
> > + * Find the zone with the most free pages, recheck the watermarks and
> > + * then directly allocate the ram. We don't want prefetch to use
> > + * __alloc_pages and go calling on reclaim.
> > + */
> > +static struct page *prefetch_get_page(void)
> > +{
>
> Should this be put in mm/page_alloc.c? It is, after all, a special-purpose
> page allocator. That way you wouldn't have to export zone_statistics and
> buffered_rmqueue.

Makes sense but it is only used in the CONFIG_SWAP_PREFETCH case so it would 
end up as a static inline in swap.h to avoid ending being #ifdefed in 
page_alloc.c. Do you think that's preferable to having it in 
swap_prefetch.c ?

>
> > +/*
> > + * trickle_swap is the main function that initiates the swap
> > prefetching. It + * first checks to see if the busy flag is set, and does
> > not prefetch if it + * is, as the flag implied we are low on memory or
> > swapping in currently. + * Otherwise it runs till PREFETCH_PAGES() are
> > prefetched.
> > + * This function returns 1 if it succeeds in a cycle of prefetching, 0
> > if it + * is interrupted or -1 if there is nothing left to prefetch.
> > + */
> > +static int trickle_swap(void)
> > +{
>
> This could perhaps use a three-state enum as return value. I find return
> value checks in kprefetchd() slightly confusing.

Good idea.

Thanks!
Con
-
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/