From: Michal Hocko <mhocko@kernel.org>
To: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>,
linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Rik van Riel <riel@redhat.com>,
David Rientjes <rientjes@google.com>,
Johannes Weiner <hannes@cmpxchg.org>,
Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>,
linux-kernel@vger.kernel.org,
Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [RFC 13/13] mm, compaction: fix and improve watermark handling
Date: Wed, 18 May 2016 16:27:53 +0200 [thread overview]
Message-ID: <20160518142753.GJ21654@dhcp22.suse.cz> (raw)
In-Reply-To: <20160518135004.GE2527@techsingularity.net>
On Wed 18-05-16 14:50:04, Mel Gorman wrote:
> On Mon, May 16, 2016 at 11:25:05AM +0200, Michal Hocko wrote:
> > On Tue 10-05-16 09:36:03, Vlastimil Babka wrote:
> > > Compaction has been using watermark checks when deciding whether it was
> > > successful, and whether compaction is at all suitable. There are few problems
> > > with these checks.
> > >
> > > - __compact_finished() uses low watermark in a check that has to pass if
> > > the direct compaction is to finish and allocation should succeed. This is
> > > too pessimistic, as the allocation will typically use min watermark. It
> > > may happen that during compaction, we drop below the low watermark (due to
> > > parallel activity), but still form the target high-order page. By checking
> > > against low watermark, we might needlessly continue compaction. After this
> > > patch, the check uses direct compactor's alloc_flags to determine the
> > > watermark, which is effectively the min watermark.
> >
> > OK, this makes some sense. It would be great if we could have at least
> > some clarification why the low wmark has been used previously. Probably
> > Mel can remember?
> >
>
> Two reasons -- it was a very rough estimate of whether enough pages are free
> for compaction to have any chance. Secondly, it was to minimise the risk
> that compaction would isolate so many pages that the zone was completely
> depleted. This was a concern during the initial prototype of compaction.
>
> > > - __compaction_suitable() then checks the low watermark plus a (2 << order) gap
> > > to decide if there's enough free memory to perform compaction. This check
> >
> > And this was a real head scratcher when I started looking into the
> > compaction recently. Why do we need to be above low watermark to even
> > start compaction. Compaction uses additional memory only for a short
> > period of time and then releases the already migrated pages.
> >
>
> Simply minimising the risk that compaction would deplete the entire
> zone. Sure, it hands pages back shortly afterwards. At the time of the
> initial prototype, page migration was severely broken and the system was
> constantly crashing. The cautious checks were left in place after page
> migration was fixed as there wasn't a compelling reason to remove them
> at the time.
OK, then moving to min_wmark + bias from low_wmark should work, right?
This would at least remove the discrepancy between the reclaim and
compaction thresholds to some degree. Which is good IMHO.
Thanks!
--
Michal Hocko
SUSE Labs
next prev parent reply other threads:[~2016-05-18 14:27 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-05-10 7:35 [RFC 00/13] make direct compaction more deterministic Vlastimil Babka
2016-05-10 7:35 ` [RFC 01/13] mm, compaction: don't isolate PageWriteback pages in MIGRATE_SYNC_LIGHT mode Vlastimil Babka
2016-05-11 12:40 ` Michal Hocko
2016-05-10 7:35 ` [RFC 02/13] mm, page_alloc: set alloc_flags only once in slowpath Vlastimil Babka
2016-05-10 11:28 ` Tetsuo Handa
2016-05-10 12:30 ` Vlastimil Babka
2016-05-12 12:41 ` Michal Hocko
2016-05-31 6:20 ` Joonsoo Kim
2016-05-31 7:59 ` Vlastimil Babka
2016-06-02 1:50 ` Joonsoo Kim
2016-05-10 7:35 ` [RFC 03/13] mm, page_alloc: don't retry initial attempt " Vlastimil Babka
2016-05-12 12:48 ` Michal Hocko
2016-05-31 6:25 ` Joonsoo Kim
2016-05-31 12:03 ` Vlastimil Babka
2016-05-10 7:35 ` [RFC 04/13] mm, page_alloc: restructure direct compaction handling " Vlastimil Babka
2016-05-12 13:29 ` Michal Hocko
2016-05-13 8:10 ` Vlastimil Babka
2016-05-13 8:31 ` Michal Hocko
2016-05-10 7:35 ` [RFC 05/13] mm, page_alloc: make THP-specific decisions more generic Vlastimil Babka
2016-05-12 13:43 ` Michal Hocko
2016-05-10 7:35 ` [RFC 06/13] mm, thp: remove __GFP_NORETRY from khugepaged and madvised allocations Vlastimil Babka
2016-05-12 16:20 ` Michal Hocko
2016-05-13 8:23 ` Vlastimil Babka
2016-05-13 12:05 ` Michal Hocko
2016-05-18 11:59 ` Vlastimil Babka
2016-05-18 15:24 ` Michal Hocko
2016-05-20 13:57 ` Vlastimil Babka
2016-05-23 8:39 ` Michal Hocko
2016-05-10 7:35 ` [RFC 07/13] mm, compaction: introduce direct compaction priority Vlastimil Babka
2016-05-13 12:37 ` Michal Hocko
2016-05-10 7:35 ` [RFC 08/13] mm, compaction: simplify contended compaction handling Vlastimil Babka
2016-05-13 13:09 ` Michal Hocko
2016-05-16 7:10 ` Vlastimil Babka
2016-05-10 7:35 ` [RFC 09/13] mm, compaction: make whole_zone flag ignore cached scanner positions Vlastimil Babka
2016-05-13 13:23 ` Michal Hocko
2016-05-10 7:36 ` [RFC 10/13] mm, compaction: cleanup unused functions Vlastimil Babka
2016-05-10 7:36 ` [RFC 11/13] mm, compaction: add the ultimate direct compaction priority Vlastimil Babka
2016-05-13 13:38 ` Michal Hocko
2016-05-16 7:17 ` Vlastimil Babka
2016-05-16 8:11 ` Michal Hocko
2016-05-18 12:46 ` Vlastimil Babka
2016-05-10 7:36 ` [RFC 12/13] mm, compaction: more reliably increase " Vlastimil Babka
2016-05-10 12:55 ` Vlastimil Babka
2016-05-13 14:15 ` Michal Hocko
2016-05-16 7:31 ` Vlastimil Babka
2016-05-16 8:14 ` Michal Hocko
2016-05-16 9:27 ` Vlastimil Babka
2016-05-16 9:52 ` Michal Hocko
2016-05-31 6:37 ` Joonsoo Kim
2016-05-31 12:07 ` Vlastimil Babka
2016-05-31 12:29 ` Vlastimil Babka
2016-06-02 2:50 ` Joonsoo Kim
2016-05-10 7:36 ` [RFC 13/13] mm, compaction: fix and improve watermark handling Vlastimil Babka
2016-05-16 9:25 ` Michal Hocko
2016-05-16 9:50 ` Vlastimil Babka
2016-05-16 12:30 ` Michal Hocko
2016-05-18 13:50 ` Mel Gorman
2016-05-18 14:27 ` Michal Hocko [this message]
2016-05-18 14:40 ` Mel Gorman
2016-05-17 20:01 ` [RFC 00/13] make direct compaction more deterministic Michal Hocko
2016-05-18 7:19 ` Vlastimil Babka
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20160518142753.GJ21654@dhcp22.suse.cz \
--to=mhocko@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=hannes@cmpxchg.org \
--cc=iamjoonsoo.kim@lge.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@techsingularity.net \
--cc=penguin-kernel@i-love.sakura.ne.jp \
--cc=riel@redhat.com \
--cc=rientjes@google.com \
--cc=torvalds@linux-foundation.org \
--cc=vbabka@suse.cz \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).