lkml.org 
[lkml]   [2014]   [Feb]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [patch v2] mm, compaction: avoid isolating pinned pages
On Thu, 6 Feb 2014, Vlastimil Babka wrote:

> > Page migration will fail for memory that is pinned in memory with, for
> > example, get_user_pages(). In this case, it is unnecessary to take
> > zone->lru_lock or isolating the page and passing it to page migration
> > which will ultimately fail.
> >
> > This is a racy check, the page can still change from under us, but in
> > that case we'll just fail later when attempting to move the page.
> >
> > This avoids very expensive memory compaction when faulting transparent
> > hugepages after pinning a lot of memory with a Mellanox driver.
> >
> > On a 128GB machine and pinning ~120GB of memory, before this patch we
> > see the enormous disparity in the number of page migration failures
> > because of the pinning (from /proc/vmstat):
> >
> > compact_pages_moved 8450
> > compact_pagemigrate_failed 15614415
> >
> > 0.05% of pages isolated are successfully migrated and explicitly
> > triggering memory compaction takes 102 seconds. After the patch:
> >
> > compact_pages_moved 9197
> > compact_pagemigrate_failed 7
> >
> > 99.9% of pages isolated are now successfully migrated in this
> > configuration and memory compaction takes less than one second.
> >
> > Signed-off-by: David Rientjes <rientjes@google.com>
> > ---
> > v2: address page count issue per Joonsoo
> >
> > mm/compaction.c | 9 +++++++++
> > 1 file changed, 9 insertions(+)
> >
> > diff --git a/mm/compaction.c b/mm/compaction.c
> > --- a/mm/compaction.c
> > +++ b/mm/compaction.c
> > @@ -578,6 +578,15 @@ isolate_migratepages_range(struct zone *zone, struct
> > compact_control *cc,
> > continue;
> > }
> > + /*
> > + * Migration will fail if an anonymous page is pinned in
> > memory,
> > + * so avoid taking lru_lock and isolating it unnecessarily in
> > an
> > + * admittedly racy check.
> > + */
> > + if (!page_mapping(page) &&
> > + page_count(page) > page_mapcount(page))
> > + continue;
> > +
>
> Hm this page_count() seems it could substantially increase the chance of race
> with prep_compound_page that your patch "mm, page_alloc: make first_page
> visible before PageTail" tries to fix :)
>

That's why I sent the fix for page_count().

The "racy check" the comment eludes to above concerns the fact that
page_count() and page_mapcount() can change out from under us before
isolation and if we had not avoided isolating them that they would have
been migratable later. We accept that as a consequence of doing this in a
lockless way without page references.


\
 
 \ /
  Last update: 2014-02-06 22:41    [W:0.066 / U:0.220 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site