From: Rik van Riel <riel@redhat.com>
To: Mel Gorman <mgorman@suse.de>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>,
linux-mm@kvack.org, Minchan Kim <minchan@kernel.org>,
Marek Szyprowski <m.szyprowski@samsung.com>,
Kyungmin Park <kyungmin.park@samsung.com>
Subject: Re: [PATCH v3] mm: compaction: handle incorrect Unmovable type pageblocks
Date: Thu, 26 Apr 2012 11:53:47 -0400 [thread overview]
Message-ID: <4F996F8B.1020207@redhat.com> (raw)
In-Reply-To: <20120426143620.GF15299@suse.de>
On 04/26/2012 10:36 AM, Mel Gorman wrote:
> Hmm, at what point does COMPACT_ASYNC_FULL get used? I see it gets
> used for the proc interface but it's not used via the page allocator at
> all.
He is using COMPACT_SYNC for the proc interface, and
COMPACT_ASYNC_FULL from kswapd.
> Minimally I was expecting to see if being used from the page allocator.
Makes sense, especially if we get the CPU overhead
saving stuff that we talked about at LSF to work :)
> A better option might be to track the number of MIGRATE_UNMOVABLE blocks that
> were skipped over during COMPACT_ASYNC_PARTIAL and if it was a high
> percentage and it looked like compaction failed then to retry with
> COMPACT_ASYNC_FULL. If you took this option, try_to_compact_pages()
> would still only take sync as a parameter and keep the decision within
> compaction.c
This I don't get.
If we have a small number of MIGRATE_UNMOVABLE blocks,
is it worth skipping over them?
If we have really large number of MIGRATE_UNMOVABLE blocks,
did we let things get out of hand? By giving the page
allocator this many unmovable blocks to choose from, we
could have ended up with actually non-compactable memory.
If we have a medium number of MIGRATE_UNMOVABLE blocks,
is it worth doing a restart and scanning all the movable
blocks again?
In other words, could it be better to always try to
rescue the unmovable blocks?
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2012-04-26 15:53 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-04-26 8:15 [PATCH v3] mm: compaction: handle incorrect Unmovable type pageblocks Bartlomiej Zolnierkiewicz
2012-04-26 14:36 ` Mel Gorman
2012-04-26 15:53 ` Rik van Riel [this message]
2012-04-26 16:47 ` Mel Gorman
2012-04-26 18:52 ` Rik van Riel
2012-04-27 9:45 ` Mel Gorman
2012-04-27 0:58 ` Minchan Kim
2012-04-27 9:56 ` Mel Gorman
2012-04-30 2:44 ` Minchan Kim
2012-04-30 8:31 ` Mel Gorman
2012-04-30 8:55 ` Minchan Kim
2012-04-30 9:16 ` Mel Gorman
2012-04-26 15:42 ` Rik van Riel
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=4F996F8B.1020207@redhat.com \
--to=riel@redhat.com \
--cc=b.zolnierkie@samsung.com \
--cc=kyungmin.park@samsung.com \
--cc=linux-mm@kvack.org \
--cc=m.szyprowski@samsung.com \
--cc=mgorman@suse.de \
--cc=minchan@kernel.org \
/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).