Re: [PATCH v2 06/10] vmscan: make isolate_lru_page with filteraware

From: Minchan Kim
Date: Tue May 31 2011 - 09:51:41 EST


On Tue, May 31, 2011 at 03:46:09PM +0200, Johannes Weiner wrote:
> On Mon, May 30, 2011 at 03:13:45AM +0900, Minchan Kim wrote:
> > In __zone_reclaim case, we don't want to shrink mapped page.
> > Nonetheless, we have isolated mapped page and re-add it into
> > LRU's head. It's unnecessary CPU overhead and makes LRU churning.
> >
> > Of course, when we isolate the page, the page might be mapped but
> > when we try to migrate the page, the page would be not mapped.
> > So it could be migrated. But race is rare and although it happens,
> > it's no big deal.
> >
> > Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
> > Acked-by: Johannes Weiner <hannes@xxxxxxxxxxx>
> > Cc: Mel Gorman <mgorman@xxxxxxx>
> > Cc: Rik van Riel <riel@xxxxxxxxxx>
> > Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx>
> > Signed-off-by: Minchan Kim <minchan.kim@xxxxxxxxx>
> > ---
> > mm/vmscan.c | 29 +++++++++++++++++++++--------
> > 1 files changed, 21 insertions(+), 8 deletions(-)
> >
> > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > index 9972356..39941c7 100644
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -1395,6 +1395,7 @@ shrink_inactive_list(unsigned long nr_to_scan, struct zone *zone,
> > unsigned long nr_taken;
> > unsigned long nr_anon;
> > unsigned long nr_file;
> > + enum ISOLATE_PAGE_MODE mode = ISOLATE_NONE;
> >
> > while (unlikely(too_many_isolated(zone, file, sc))) {
> > congestion_wait(BLK_RW_ASYNC, HZ/10);
> > @@ -1406,13 +1407,20 @@ shrink_inactive_list(unsigned long nr_to_scan, struct zone *zone,
> >
> > set_reclaim_mode(priority, sc, false);
> > lru_add_drain();
> > +
> > + if (!sc->may_unmap)
> > + mode |= ISOLATE_UNMAPPED;
> > + if (!sc->may_writepage)
> > + mode |= ISOLATE_CLEAN;
> > + mode |= sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM ?
> > + ISOLATE_BOTH : ISOLATE_INACTIVE;
>
> Hmm, it would probably be cleaner to fully convert the isolation mode
> into independent flags. INACTIVE, ACTIVE, BOTH is currently a
> tri-state among flags, which is a bit ugly.
>
> mode = ISOLATE_INACTIVE;
> if (!sc->may_unmap)
> mode |= ISOLATE_UNMAPPED;
> if (!sc->may_writepage)
> mode |= ISOLATE_CLEAN;
> if (sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM)
> mode |= ISOLATE_ACTIVE;
>
> What do you think?

It's good point.
Actually, I am trying it for unevictable page migration.
I removed BOTH and insert ISOLATE_UNEVICTABLE.
But it's in my queue and doesn't published yet.
The summary is that I am going on that way.
I will clean up it in v3, too.

==
Subject: [PATCH 1/2] Cleanup ISOLATE_BOTH

Before 2.6.38, we just had two lru list(active/inactive).
Now we have added one more lru type list. ie, unevictable.
So ISOLATE_BOTH is not clear naming.
This patch removes ISOLATE_BOTH and instead of it,
it require to use more explicit word.

This patch should not change old behavir and it's used by
next patch series.
==

--
Kind regards
Minchan Kim
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/