From: Mel Gorman <mgorman@techsingularity.net>
To: Vlastimil Babka <vbabka@suse.cz>
Cc: Linux-MM <linux-mm@kvack.org>,
David Rientjes <rientjes@google.com>,
Andrea Arcangeli <aarcange@redhat.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
Michal Hocko <mhocko@kernel.org>,
ying.huang@intel.com, kirill@shutemov.name,
Andrew Morton <akpm@linux-foundation.org>,
Linux List Kernel Mailing <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 09/14] mm, compaction: Ignore the fragmentation avoidance boost for isolation and compaction
Date: Tue, 18 Dec 2018 14:29:52 +0000 [thread overview]
Message-ID: <20181218142952.GL29005@techsingularity.net> (raw)
In-Reply-To: <adae728e-0e62-abb3-901e-0696930bb7dd@suse.cz>
On Tue, Dec 18, 2018 at 02:58:33PM +0100, Vlastimil Babka wrote:
> On 12/18/18 2:51 PM, Mel Gorman wrote:
> > On Tue, Dec 18, 2018 at 01:36:42PM +0100, Vlastimil Babka wrote:
> >> On 12/15/18 12:03 AM, Mel Gorman wrote:
> >>> When pageblocks get fragmented, watermarks are artifically boosted to pages
> >>> are reclaimed to avoid further fragmentation events. However, compaction
> >>> is often either fragmentation-neutral or moving movable pages away from
> >>> unmovable/reclaimable pages. As the actual watermarks are preserved,
> >>> allow compaction to ignore the boost factor.
> >>
> >> Right, I should have realized that when reviewing the boost patch. I
> >> think it would be useful to do the same change in
> >> __compaction_suitable() as well. Compaction has its own "gap".
> >>
> >
> > That gap is somewhat static though so I'm a bit more wary of it. However,
>
> Well, watermark boost is dynamic, but based on allocations stealing from
> other migratetypes, not reflecting compaction chances of success.
>
True.
> > the check in __isolate_free_page looks too agressive. We isolate in
> > units of COMPACT_CLUSTER_MAX yet the watermark check there is based on
> > the allocation request. That means for THP that we check if 512 pages
> > can be allocated when only somewhere between 1 and 32 is needed for that
> > compaction cycle to complete. Adjusting that might be more appropriate?
>
> AFAIU the code in __isolate_free_page() reflects that if there's less
> than 512 free pages gap, we might form a high-order page for THP but
> won't be able to allocate it afterwards due to watermark.
Yeah but it used to be a lot more important when watermark checking for
high-orders was very different. Now, if the watermark is met for order-0
and a large enough free page is allocated, the allocation succeeds so
it's a lot less relevant than it used to be. kswapd will still run in
the background for order-0 if necessary so a heavy watermark check there
doesn't really help.
--
Mel Gorman
SUSE Labs
next prev parent reply other threads:[~2018-12-18 14:29 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-12-14 23:02 [RFC PATCH 00/14] Increase success rates and reduce latency of compaction v1 Mel Gorman
2018-12-14 23:02 ` [PATCH 01/14] mm, compaction: Shrink compact_control Mel Gorman
2018-12-17 12:27 ` Vlastimil Babka
2018-12-14 23:02 ` [PATCH 02/14] mm, compaction: Rearrange compact_control Mel Gorman
2018-12-17 13:20 ` Vlastimil Babka
2018-12-14 23:02 ` [PATCH 03/14] mm, compaction: Remove last_migrated_pfn from compact_control Mel Gorman
2018-12-17 13:50 ` Vlastimil Babka
2018-12-14 23:03 ` [PATCH 04/14] mm, compaction: Rename map_pages to split_map_pages Mel Gorman
2018-12-17 14:06 ` Vlastimil Babka
2018-12-17 14:30 ` Mel Gorman
2018-12-14 23:03 ` [PATCH 05/14] mm, compaction: Skip pageblocks with reserved pages Mel Gorman
2018-12-18 8:08 ` Vlastimil Babka
2018-12-18 8:38 ` Mel Gorman
2018-12-14 23:03 ` [PATCH 06/14] mm, migrate: Immediately fail migration of a page with no migration handler Mel Gorman
2018-12-18 9:06 ` Vlastimil Babka
2018-12-18 9:55 ` Mel Gorman
2018-12-20 19:44 ` Yang Shi
2018-12-20 20:31 ` Mel Gorman
2018-12-14 23:03 ` [PATCH 07/14] mm, compaction: Always finish scanning of a full pageblock Mel Gorman
2018-12-18 9:23 ` Vlastimil Babka
2018-12-14 23:03 ` [PATCH 08/14] mm, compaction: Use the page allocator bulk-free helper for lists of pages Mel Gorman
2018-12-18 9:55 ` Vlastimil Babka
2018-12-19 16:04 ` Mel Gorman
2018-12-14 23:03 ` [PATCH 09/14] mm, compaction: Ignore the fragmentation avoidance boost for isolation and compaction Mel Gorman
2018-12-18 12:36 ` Vlastimil Babka
2018-12-18 13:51 ` Mel Gorman
2018-12-18 13:58 ` Vlastimil Babka
2018-12-18 14:29 ` Mel Gorman [this message]
2018-12-14 23:04 ` [PATCH 10/14] mm, compaction: Use free lists to quickly locate a migration source Mel Gorman
2018-12-14 23:05 ` [PATCH 11/14] mm, compaction: Keep migration source private to a single compaction instance Mel Gorman
2018-12-14 23:05 ` [PATCH 12/14] mm, compaction: Use free lists to quickly locate a migration target Mel Gorman
2018-12-14 23:05 ` [PATCH 13/14] mm, compaction: Capture a page under direct compaction Mel Gorman
2018-12-14 23:06 ` [PATCH 14/14] mm, compaction: Do not direct compact remote memory Mel Gorman
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=20181218142952.GL29005@techsingularity.net \
--to=mgorman@techsingularity.net \
--cc=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=kirill@shutemov.name \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@kernel.org \
--cc=rientjes@google.com \
--cc=torvalds@linux-foundation.org \
--cc=vbabka@suse.cz \
--cc=ying.huang@intel.com \
/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).