From: Wu Fengguang <fengguang.wu@intel.com>
To: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Peter Zijlstra <peterz@infradead.org>, Jan Kara <jack@suse.cz>,
Steve Rago <sar@nec-labs.com>,
"linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>,
"jens.axboe" <jens.axboe@oracle.com>,
Peter Staubach <staubach@redhat.com>,
Arjan van de Ven <arjan@infradead.org>,
Ingo Molnar <mingo@elte.hu>,
"linux-fsdevel@vger.kernel.org" <linux-fsdevel@vger.kernel.org>
Subject: Re: [PATCH 2/6] VM/NFS: The VM must tell the filesystem when to free reclaimable pages
Date: Thu, 7 Jan 2010 10:29:03 +0800 [thread overview]
Message-ID: <20100107022903.GD9475@localhost> (raw)
In-Reply-To: <20100106205110.22547.81645.stgit@localhost.localdomain>
On Thu, Jan 07, 2010 at 04:51:10AM +0800, Trond Myklebust wrote:
> balance_dirty_pages() should really tell the filesystem whether or not it
> has an excess of actual dirty pages, or whether it would be more useful to
> start freeing up the unstable writes.
>
> Assume that if the number of unstable writes is more than 1/2 the number of
> reclaimable pages, then we should force NFS to free up the former.
>
> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
> Acked-by: Jan Kara <jack@suse.cz>
> ---
>
> fs/nfs/write.c | 2 +-
> include/linux/writeback.h | 5 +++++
> mm/page-writeback.c | 9 ++++++++-
> 3 files changed, 14 insertions(+), 2 deletions(-)
>
> diff --git a/fs/nfs/write.c b/fs/nfs/write.c
> index 910be28..ee3daf4 100644
> --- a/fs/nfs/write.c
> +++ b/fs/nfs/write.c
> @@ -1417,7 +1417,7 @@ int nfs_commit_unstable_pages(struct address_space *mapping,
> /* Don't commit yet if this is a non-blocking flush and there are
> * outstanding writes for this mapping.
> */
> - if (wbc->sync_mode != WB_SYNC_ALL &&
> + if (!wbc->force_commit && wbc->sync_mode != WB_SYNC_ALL &&
> radix_tree_tagged(&NFS_I(inode)->nfs_page_tree,
> NFS_PAGE_TAG_LOCKED)) {
> mark_inode_unstable_pages(inode);
> diff --git a/include/linux/writeback.h b/include/linux/writeback.h
> index 76e8903..3fd5c3e 100644
> --- a/include/linux/writeback.h
> +++ b/include/linux/writeback.h
> @@ -62,6 +62,11 @@ struct writeback_control {
> * so we use a single control to update them
> */
> unsigned no_nrwrite_index_update:1;
> + /*
> + * The following is used by balance_dirty_pages() to
> + * force NFS to commit unstable pages.
> + */
In fact it may be too late to force commit at balance_dirty_pages()
time: commit takes time and the application has already been blocked.
If not convenient for now, I can make the change -- I'll remove the
writeback_inodes_wbc() call altogether from balance_dirty_pages().
> + unsigned force_commit:1;
> };
nfs_commit may be a more newbie friendly name?
Thanks,
Fengguang
> /*
> diff --git a/mm/page-writeback.c b/mm/page-writeback.c
> index 0b19943..ede5356 100644
> --- a/mm/page-writeback.c
> +++ b/mm/page-writeback.c
> @@ -485,6 +485,7 @@ static void balance_dirty_pages(struct address_space *mapping,
> {
> long nr_reclaimable, bdi_nr_reclaimable;
> long nr_writeback, bdi_nr_writeback;
> + long nr_unstable_nfs;
> unsigned long background_thresh;
> unsigned long dirty_thresh;
> unsigned long bdi_thresh;
> @@ -505,8 +506,9 @@ static void balance_dirty_pages(struct address_space *mapping,
> get_dirty_limits(&background_thresh, &dirty_thresh,
> &bdi_thresh, bdi);
>
> + nr_unstable_nfs = global_page_state(NR_UNSTABLE_NFS);
> nr_reclaimable = global_page_state(NR_FILE_DIRTY) +
> - global_page_state(NR_UNSTABLE_NFS);
> + nr_unstable_nfs;
> nr_writeback = global_page_state(NR_WRITEBACK);
>
> bdi_nr_reclaimable = bdi_stat(bdi, BDI_RECLAIMABLE);
> @@ -537,6 +539,11 @@ static void balance_dirty_pages(struct address_space *mapping,
> * up.
> */
> if (bdi_nr_reclaimable > bdi_thresh) {
> + wbc.force_commit = 0;
> + /* Force NFS to also free up unstable writes. */
> + if (nr_unstable_nfs > nr_reclaimable / 2)
> + wbc.force_commit = 1;
> +
> writeback_inodes_wbc(&wbc);
> pages_written += write_chunk - wbc.nr_to_write;
> get_dirty_limits(&background_thresh, &dirty_thresh,
>
next prev parent reply other threads:[~2010-01-07 2:29 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <1261015420.1947.54.camel@serenity>
[not found] ` <1261037877.27920.36.camel@laptop>
[not found] ` <20091219122033.GA11360@localhost>
[not found] ` <1261232747.1947.194.camel@serenity>
2009-12-22 1:59 ` [PATCH] improve the performance of large sequential write NFS workloads Wu Fengguang
2009-12-22 12:35 ` Jan Kara
[not found] ` <20091222123538.GB604-jyMamyUUXNJG4ohzP4jBZS1Fcj925eT/@public.gmane.org>
2009-12-23 8:43 ` Christoph Hellwig
[not found] ` <20091223084302.GA14912-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2009-12-23 13:32 ` Jan Kara
[not found] ` <20091223133244.GB3159-+0h/O2h83AeN3ZZ/Hiejyg@public.gmane.org>
2009-12-24 5:25 ` Wu Fengguang
2009-12-24 1:26 ` Wu Fengguang
2009-12-22 16:41 ` Steve Rago
2009-12-24 1:21 ` Wu Fengguang
2009-12-24 14:49 ` Steve Rago
2009-12-25 7:37 ` Wu Fengguang
2009-12-23 14:21 ` Trond Myklebust
2009-12-23 18:05 ` Jan Kara
[not found] ` <20091223180551.GD3159-+0h/O2h83AeN3ZZ/Hiejyg@public.gmane.org>
2009-12-23 19:12 ` Trond Myklebust
2009-12-24 2:52 ` Wu Fengguang
2009-12-24 12:04 ` Trond Myklebust
2009-12-25 5:56 ` Wu Fengguang
2009-12-30 16:22 ` Trond Myklebust
2009-12-31 5:04 ` Wu Fengguang
2009-12-31 19:13 ` Trond Myklebust
2010-01-06 3:03 ` Wu Fengguang
2010-01-06 16:56 ` Trond Myklebust
2010-01-06 18:26 ` Trond Myklebust
2010-01-06 18:37 ` Peter Zijlstra
2010-01-06 18:52 ` Trond Myklebust
2010-01-06 19:07 ` Peter Zijlstra
2010-01-06 19:21 ` Trond Myklebust
2010-01-06 19:53 ` Trond Myklebust
2010-01-06 20:09 ` Jan Kara
[not found] ` <20100106200928.GB22781-+0h/O2h83AeN3ZZ/Hiejyg@public.gmane.org>
2010-01-06 20:51 ` [PATCH 0/6] " Trond Myklebust
[not found] ` <20100106205110.22547.85345.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-06 20:51 ` [PATCH 6/6] NFS: Run COMMIT as an asynchronous RPC call when wbc->for_background is set Trond Myklebust
[not found] ` <20100106205110.22547.31434.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-07 2:32 ` Wu Fengguang
2010-01-06 20:51 ` [PATCH 1/6] VFS: Ensure that writeback_single_inode() commits unstable writes Trond Myklebust
[not found] ` <20100106205110.22547.17971.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-06 21:38 ` Jan Kara
[not found] ` <20100106213843.GD22781-+0h/O2h83AeN3ZZ/Hiejyg@public.gmane.org>
2010-01-06 21:48 ` Trond Myklebust
2010-01-07 2:18 ` Wu Fengguang
[not found] ` <1262839082.2185.15.camel@localhost>
2010-01-07 4:48 ` Wu Fengguang
2010-01-07 4:53 ` [PATCH 0/5] Re: [PATCH] improve the performance of large sequential write NFS workloads Trond Myklebust
[not found] ` <20100107045330.5986.55090.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-07 4:53 ` [PATCH 3/5] VM: Don't call bdi_stat(BDI_UNSTABLE) on non-nfs backing-devices Trond Myklebust
2010-01-07 4:53 ` [PATCH 1/5] VFS: Ensure that writeback_single_inode() commits unstable writes Trond Myklebust
2010-01-07 4:53 ` [PATCH 5/5] NFS: Run COMMIT as an asynchronous RPC call when wbc->for_background is set Trond Myklebust
2010-01-07 4:53 ` [PATCH 2/5] VM: Split out the accounting of unstable writes from BDI_RECLAIMABLE Trond Myklebust
2010-01-07 4:53 ` [PATCH 4/5] VM/NFS: The VM must tell the filesystem when to free reclaimable pages Trond Myklebust
2010-01-07 14:56 ` [PATCH 1/6] VFS: Ensure that writeback_single_inode() commits unstable writes Wu Fengguang
2010-01-07 15:10 ` Trond Myklebust
2010-01-08 1:17 ` Wu Fengguang
2010-01-08 1:37 ` Trond Myklebust
2010-01-08 1:53 ` Wu Fengguang
2010-01-08 9:25 ` Christoph Hellwig
2010-01-08 13:46 ` Trond Myklebust
2010-01-08 13:54 ` Christoph Hellwig
2010-01-08 14:15 ` Trond Myklebust
2010-01-06 20:51 ` [PATCH 4/6] VM: Don't call bdi_stat(BDI_UNSTABLE) on non-nfs backing-devices Trond Myklebust
2010-01-07 1:56 ` Wu Fengguang
2010-01-06 20:51 ` [PATCH 5/6] VM: Use per-bdi unstable accounting to improve use of wbc->force_commit Trond Myklebust
[not found] ` <20100106205110.22547.32584.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-07 2:34 ` Wu Fengguang
2010-01-06 20:51 ` [PATCH 2/6] VM/NFS: The VM must tell the filesystem when to free reclaimable pages Trond Myklebust
2010-01-07 2:29 ` Wu Fengguang [this message]
2010-01-07 4:49 ` Trond Myklebust
2010-01-07 5:03 ` Wu Fengguang
2010-01-07 5:30 ` Trond Myklebust
2010-01-07 14:37 ` Wu Fengguang
2010-01-06 20:51 ` [PATCH 3/6] VM: Split out the accounting of unstable writes from BDI_RECLAIMABLE Trond Myklebust
[not found] ` <20100106205110.22547.93554.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-07 1:48 ` Wu Fengguang
2010-01-06 21:44 ` [PATCH 0/6] Re: [PATCH] improve the performance of large sequential write NFS workloads Jan Kara
2010-01-06 22:03 ` Trond Myklebust
2010-01-07 8:16 ` Peter Zijlstra
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=20100107022903.GD9475@localhost \
--to=fengguang.wu@intel.com \
--cc=Trond.Myklebust@netapp.com \
--cc=arjan@infradead.org \
--cc=jack@suse.cz \
--cc=jens.axboe@oracle.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-nfs@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=peterz@infradead.org \
--cc=sar@nec-labs.com \
--cc=staubach@redhat.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).