From: Rik van Riel <riel@redhat.com>
To: Minchan Kim <minchan.kim@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
linux-mm <linux-mm@kvack.org>,
LKML <linux-kernel@vger.kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
Johannes Weiner <hannes@cmpxchg.org>,
Nick Piggin <npiggin@kernel.dk>
Subject: Re: [RFC 1/2] deactive invalidated pages
Date: Sun, 21 Nov 2010 20:17:29 -0500 [thread overview]
Message-ID: <4CE9C4A9.3040406@redhat.com> (raw)
In-Reply-To: <bdd6628e81c06f6871983c971d91160fca3f8b5e.1290349672.git.minchan.kim@gmail.com>
On 11/21/2010 09:30 AM, Minchan Kim wrote:
> Recently, there are reported problem about thrashing.
> (http://marc.info/?l=rsync&m=128885034930933&w=2)
> It happens by backup workloads(ex, nightly rsync).
> That's because the workload makes just use-once pages
> and touches pages twice. It promotes the page into
> active list so that it results in working set page eviction.
>
> Some app developer want to support POSIX_FADV_NOREUSE.
> But other OSes don't support it, either.
> (http://marc.info/?l=linux-mm&m=128928979512086&w=2)
>
> By Other approach, app developer uses POSIX_FADV_DONTNEED.
> But it has a problem. If kernel meets page is writing
> during invalidate_mapping_pages, it can't work.
> It is very hard for application programmer to use it.
> Because they always have to sync data before calling
> fadivse(..POSIX_FADV_DONTNEED) to make sure the pages could
> be discardable. At last, they can't use deferred write of kernel
> so that they could see performance loss.
> (http://insights.oetiker.ch/linux/fadvise.html)
This looks promising...
Acked-by: Rik van Riel <riel@redhat.com>
--
All rights reversed
WARNING: multiple messages have this Message-ID (diff)
From: Rik van Riel <riel@redhat.com>
To: Minchan Kim <minchan.kim@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
linux-mm <linux-mm@kvack.org>,
LKML <linux-kernel@vger.kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
Johannes Weiner <hannes@cmpxchg.org>,
Nick Piggin <npiggin@kernel.dk>
Subject: Re: [RFC 1/2] deactive invalidated pages
Date: Sun, 21 Nov 2010 20:17:29 -0500 [thread overview]
Message-ID: <4CE9C4A9.3040406@redhat.com> (raw)
In-Reply-To: <bdd6628e81c06f6871983c971d91160fca3f8b5e.1290349672.git.minchan.kim@gmail.com>
On 11/21/2010 09:30 AM, Minchan Kim wrote:
> Recently, there are reported problem about thrashing.
> (http://marc.info/?l=rsync&m=128885034930933&w=2)
> It happens by backup workloads(ex, nightly rsync).
> That's because the workload makes just use-once pages
> and touches pages twice. It promotes the page into
> active list so that it results in working set page eviction.
>
> Some app developer want to support POSIX_FADV_NOREUSE.
> But other OSes don't support it, either.
> (http://marc.info/?l=linux-mm&m=128928979512086&w=2)
>
> By Other approach, app developer uses POSIX_FADV_DONTNEED.
> But it has a problem. If kernel meets page is writing
> during invalidate_mapping_pages, it can't work.
> It is very hard for application programmer to use it.
> Because they always have to sync data before calling
> fadivse(..POSIX_FADV_DONTNEED) to make sure the pages could
> be discardable. At last, they can't use deferred write of kernel
> so that they could see performance loss.
> (http://insights.oetiker.ch/linux/fadvise.html)
This looks promising...
Acked-by: Rik van Riel <riel@redhat.com>
--
All rights reversed
--
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 policy in Canada: sign http://dissolvethecrtc.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2010-11-22 1:18 UTC|newest]
Thread overview: 108+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-11-21 14:30 [RFC 1/2] deactive invalidated pages Minchan Kim
2010-11-21 14:30 ` Minchan Kim
2010-11-21 14:30 ` [RFC 2/2] Prevent promotion of page in madvise_dontneed Minchan Kim
2010-11-21 14:30 ` Minchan Kim
2010-11-21 14:38 ` Minchan Kim
2010-11-21 14:38 ` Minchan Kim
2010-11-21 16:34 ` Ben Gamari
2010-11-21 16:34 ` Ben Gamari
2010-11-22 0:31 ` Minchan Kim
2010-11-22 0:31 ` Minchan Kim
2010-11-22 22:21 ` Andrew Morton
2010-11-22 22:21 ` Andrew Morton
2010-11-23 4:57 ` Minchan Kim
2010-11-23 4:57 ` Minchan Kim
2010-11-23 9:50 ` Mel Gorman
2010-11-23 9:50 ` Mel Gorman
2010-11-23 23:49 ` Minchan Kim
2010-11-23 23:49 ` Minchan Kim
2010-11-21 15:21 ` [RFC 1/2] deactive invalidated pages Ben Gamari
2010-11-21 15:21 ` Ben Gamari
2010-11-23 7:16 ` KOSAKI Motohiro
2010-11-23 7:16 ` KOSAKI Motohiro
2010-11-23 13:48 ` Ben Gamari
2010-11-23 13:48 ` Ben Gamari
2010-11-23 23:48 ` Minchan Kim
2010-11-23 23:48 ` Minchan Kim
2010-11-23 14:49 ` [RFC PATCH] fadvise support in rsync Ben Gamari
2010-11-23 14:49 ` Ben Gamari
2010-11-23 15:35 ` Pádraig Brady
2010-11-23 15:35 ` Pádraig Brady
2010-11-24 0:17 ` KOSAKI Motohiro
2010-11-24 0:17 ` KOSAKI Motohiro
2010-11-23 14:49 ` [PATCH 1/3] Add fadvise interface wrapper Ben Gamari
2010-11-23 14:49 ` Ben Gamari
2010-11-23 14:49 ` [PATCH 2/3] Inform kernel of FADV_DONTNEED hint in sender Ben Gamari
2010-11-23 14:49 ` Ben Gamari
2010-11-23 14:49 ` [PATCH 3/3] Inform kernel of FADV_DONTNEED hint in receiver Ben Gamari
2010-11-23 14:49 ` Ben Gamari
2010-11-22 1:17 ` Rik van Riel [this message]
2010-11-22 1:17 ` [RFC 1/2] deactive invalidated pages Rik van Riel
2010-11-22 22:14 ` Andrew Morton
2010-11-22 22:14 ` Andrew Morton
2010-11-23 4:52 ` Minchan Kim
2010-11-23 4:52 ` Minchan Kim
2010-11-23 5:01 ` Andrew Morton
2010-11-23 5:01 ` Andrew Morton
2010-11-23 5:23 ` Minchan Kim
2010-11-23 5:23 ` Minchan Kim
2010-11-23 5:22 ` Andrew Morton
2010-11-23 5:22 ` Andrew Morton
2010-11-23 5:45 ` Minchan Kim
2010-11-23 5:45 ` Minchan Kim
2010-11-23 5:48 ` Andrew Morton
2010-11-23 5:48 ` Andrew Morton
2010-11-23 6:05 ` Minchan Kim
2010-11-23 6:05 ` Minchan Kim
2010-11-23 7:15 ` Andrew Morton
2010-11-23 7:15 ` Andrew Morton
2010-11-23 7:44 ` Minchan Kim
2010-11-23 7:44 ` Minchan Kim
2010-11-23 7:53 ` Andrew Morton
2010-11-23 7:53 ` Andrew Morton
2010-11-23 8:02 ` Minchan Kim
2010-11-23 8:02 ` Minchan Kim
2010-11-23 9:43 ` Mel Gorman
2010-11-23 9:43 ` Mel Gorman
2010-11-23 23:32 ` Minchan Kim
2010-11-23 23:32 ` Minchan Kim
2010-11-23 9:38 ` Mel Gorman
2010-11-23 9:38 ` Mel Gorman
2010-11-23 14:55 ` Ben Gamari
2010-11-23 14:55 ` Ben Gamari
2010-11-23 14:58 ` Mel Gorman
2010-11-23 14:58 ` Mel Gorman
2010-11-23 20:35 ` Andrew Morton
2010-11-23 20:35 ` Andrew Morton
2010-11-23 22:10 ` Mel Gorman
2010-11-23 22:10 ` Mel Gorman
2010-11-23 23:45 ` Minchan Kim
2010-11-23 23:45 ` Minchan Kim
2010-11-24 18:01 ` Mel Gorman
2010-11-24 18:01 ` Mel Gorman
2010-11-23 7:16 ` KOSAKI Motohiro
2010-11-23 7:16 ` KOSAKI Motohiro
2010-11-23 7:40 ` Minchan Kim
2010-11-23 7:40 ` Minchan Kim
2010-11-23 7:42 ` Andrew Morton
2010-11-23 7:42 ` Andrew Morton
2010-11-23 8:01 ` KOSAKI Motohiro
2010-11-23 8:01 ` KOSAKI Motohiro
2010-11-23 8:44 ` Minchan Kim
2010-11-23 8:44 ` Minchan Kim
2010-11-23 9:02 ` KOSAKI Motohiro
2010-11-23 9:02 ` KOSAKI Motohiro
2010-11-23 9:05 ` Minchan Kim
2010-11-23 9:05 ` Minchan Kim
2010-11-23 9:07 ` Minchan Kim
2010-11-23 9:07 ` Minchan Kim
2010-11-23 14:57 ` Ben Gamari
2010-11-23 14:57 ` Ben Gamari
2010-11-24 0:13 ` KOSAKI Motohiro
2010-11-24 0:13 ` KOSAKI Motohiro
2010-11-23 9:28 ` Mel Gorman
2010-11-23 9:28 ` Mel Gorman
2010-11-23 23:24 ` Minchan Kim
2010-11-23 23:24 ` Minchan Kim
2010-11-24 10:02 ` Mel Gorman
2010-11-24 10:02 ` 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=4CE9C4A9.3040406@redhat.com \
--to=riel@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=hannes@cmpxchg.org \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=minchan.kim@gmail.com \
--cc=npiggin@kernel.dk \
--cc=peterz@infradead.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.