On Thu, Oct 19, 2017 at 03:12:33PM +0200, Vlastimil Babka wrote:
> On 10/18/2017 09:59 AM, Mel Gorman wrote:
> > Most callers users of free_hot_cold_page claim the pages being released are
> > cache hot. The exception is the page reclaim paths where it is likely that
> > enough pages will be freed in the near future that the per-cpu lists are
> > going to be recycled and the cache hotness information is lost.
>
> Maybe it would make sense for reclaim to skip pcplists? (out of scope of
> this series, of course).
>
Maybe, but it's a bit risky. The PCP lists are preserved but the number of
zone->lock acquire/releases increases as now every 14 pages reclaimed will
be an acquire/release instead of every pcp->high number of pages reclaimed.
That is a definite cost versus a possibility that the next page allocated no
that CPU will still be cache hot. That in itself may not happen as
scanning lots of pages for reclaim may have filled the cache with
useless information anyway.
> > As no one
> > really cares about the hotness of pages being released to the allocator,
> > just ditch the parameter.
> >
> > The APIs are renamed to indicate that it's no longer about hot/cold pages. It
> > should also be less confusing as there are subtle differences between them.
> > __free_pages drops a reference and frees a page when the refcount reaches
> > zero. free_hot_cold_page handled pages whose refcount was already zero
> > which is non-obvious from the name. free_unref_page should be more obvious.
> >
> > No performance impact is expected as the overhead is marginal. The parameter
> > is removed simply because it is a bit stupid to have a useless parameter
> > copied everywhere.
> >
> > Signed-off-by: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
>
> Acked-by: Vlastimil Babka <vbabka@xxxxxxx>
>
> A comment below, though.
>
> ...
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index 167e163cf733..13582efc57a0 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -2590,7 +2590,7 @@ void mark_free_pages(struct zone *zone)
> > }
> > #endif /* CONFIG_PM */
> >
> > -static bool free_hot_cold_page_prepare(struct page *page, unsigned long pfn)
> > +static bool free_unref_page_prepare(struct page *page, unsigned long pfn)
> > {
> > int migratetype;
> >
> > @@ -2602,8 +2602,7 @@ static bool free_hot_cold_page_prepare(struct page *page, unsigned long pfn)
> > return true;
> > }
> >
> > -static void free_hot_cold_page_commit(struct page *page, unsigned long pfn,
> > - bool cold)
> > +static void free_unref_page_commit(struct page *page, unsigned long pfn)
> > {
> > struct zone *zone = page_zone(page);
> > struct per_cpu_pages *pcp;
> > @@ -2628,10 +2627,7 @@ static void free_hot_cold_page_commit(struct page *page, unsigned long pfn,
> > }
> >
> > pcp = &this_cpu_ptr(zone->pageset)->pcp;
> > - if (!cold)
> > - list_add(&page->lru, &pcp->lists[migratetype]);
> > - else
> > - list_add_tail(&page->lru, &pcp->lists[migratetype]);
> > + list_add_tail(&page->lru, &pcp->lists[migratetype]);
>
> Did you intentionally use the cold version here? Patch 8/8 uses the hot
> version in __rmqueue_pcplist() and that makes more sense to me. It
> should be either negligible or better, not worse.
>
This was unintentional, thanks. The fix is below
---8<---
mm, Remove cold parameter from free_hot_cold_page* -fix
As pointed out by Vlastimil Babka, the pages being freed should be added
to the head, no the tail, of the pcpu list.
Signed-off-by: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
---
mm/page_alloc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 13582efc57a0..06461553a115 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2627,7 +2627,7 @@ static void free_unref_page_commit(struct page *page, unsigned long pfn)
}
pcp = &this_cpu_ptr(zone->pageset)->pcp;
- list_add_tail(&page->lru, &pcp->lists[migratetype]);
+ list_add(&page->lru, &pcp->lists[migratetype]);
pcp->count++;
if (pcp->count >= pcp->high) {
unsigned long batch = READ_ONCE(pcp->batch);
