From: Alasdair G Kergon <agk@redhat.com>
To: Stefan Bader <sbader3@googlemail.com>
Cc: Tejun Heo <htejun@gmail.com>, David Chinner <dgc@sgi.com>,
linux-kernel@vger.kernel.org, linux-raid@vger.kernel.org,
device-mapper development <dm-devel@redhat.com>,
Jens Axboe <jens.axboe@oracle.com>,
linux-fsdevel@vger.kernel.org,
Andreas Dilger <adilger@clusterfs.com>
Subject: Re: Re: [RFD] BIO_RW_BARRIER - what it means for devices, filesystems, and dm/md.
Date: Tue, 29 May 2007 23:05:00 +0100 [thread overview]
Message-ID: <20070529220500.GA6513@agk.fab.redhat.com> (raw)
In-Reply-To: <5201e28f0705290225v14fdac44hb0382a4137a84d01@mail.gmail.com>
On Tue, May 29, 2007 at 11:25:42AM +0200, Stefan Bader wrote:
> doing a sort of suspend, issuing the
> barrier request, calling flush to all mapped devices and then wait for
> in-flight I/O to go to zero?
Something like that is needed for some dm targets to support barriers.
(We needn't always wait for *all* in-flight I/O.)
When faced with -EOPNOTSUP, do all callers fall back to a sync in
the places a barrier would have been used, or are there any more
sophisticated strategies attempting to optimise code without barriers?
> I am not a hundred percent sure about
> that but I think that just passing the barrier flag on to mapped
> devices might in some (maybe they are rare) cases cause a layer above
> to think all data is on-disk while this isn't necessarily true (see my
> previous post). What do you think?
An efficient I/O barrier implementation would not normally involve
flushing AFAIK: dm surely wouldn't "cause" a higher layer to assume
stronger semantics than are provided.
Alasdair
--
agk@redhat.com
WARNING: multiple messages have this Message-ID (diff)
From: Alasdair G Kergon <agk@redhat.com>
To: Stefan Bader <sbader3@googlemail.com>
Cc: device-mapper development <dm-devel@redhat.com>,
linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-raid@vger.kernel.org, Jens Axboe <jens.axboe@oracle.com>,
David Chinner <dgc@sgi.com>, Phillip Susi <psusi@cfl.rr.com>,
Stefan Bader <Stefan.Bader@de.ibm.com>,
Andreas Dilger <adilger@clusterfs.com>,
Tejun Heo <htejun@gmail.com>
Subject: Re: [dm-devel] Re: [RFD] BIO_RW_BARRIER - what it means for devices, filesystems, and dm/md.
Date: Tue, 29 May 2007 23:05:00 +0100 [thread overview]
Message-ID: <20070529220500.GA6513@agk.fab.redhat.com> (raw)
In-Reply-To: <5201e28f0705290225v14fdac44hb0382a4137a84d01@mail.gmail.com>
On Tue, May 29, 2007 at 11:25:42AM +0200, Stefan Bader wrote:
> doing a sort of suspend, issuing the
> barrier request, calling flush to all mapped devices and then wait for
> in-flight I/O to go to zero?
Something like that is needed for some dm targets to support barriers.
(We needn't always wait for *all* in-flight I/O.)
When faced with -EOPNOTSUP, do all callers fall back to a sync in
the places a barrier would have been used, or are there any more
sophisticated strategies attempting to optimise code without barriers?
> I am not a hundred percent sure about
> that but I think that just passing the barrier flag on to mapped
> devices might in some (maybe they are rare) cases cause a layer above
> to think all data is on-disk while this isn't necessarily true (see my
> previous post). What do you think?
An efficient I/O barrier implementation would not normally involve
flushing AFAIK: dm surely wouldn't "cause" a higher layer to assume
stronger semantics than are provided.
Alasdair
--
agk@redhat.com
next prev parent reply other threads:[~2007-05-29 22:05 UTC|newest]
Thread overview: 134+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-05-25 7:58 [RFD] BIO_RW_BARRIER - what it means for devices, filesystems, and dm/md Neil Brown
2007-05-25 7:58 ` Neil Brown
2007-05-25 11:15 ` David Chinner
2007-05-25 11:49 ` Jens Axboe
2007-05-25 14:49 ` Phillip Susi
2007-05-25 14:49 ` [dm-devel] " Phillip Susi
2007-05-28 18:32 ` Jens Axboe
2007-05-25 13:52 ` Stefan Bader
2007-05-28 1:37 ` Neil Brown
2007-05-29 9:12 ` Stefan Bader
2007-05-25 15:11 ` Phillip Susi
2007-05-25 15:11 ` [dm-devel] " Phillip Susi
2007-05-26 1:03 ` Andreas Dilger
2007-05-26 10:27 ` Tejun Heo
2007-05-28 1:30 ` Neil Brown
2007-05-28 2:45 ` David Chinner
2007-05-28 2:57 ` Neil Brown
2007-05-28 4:29 ` David Chinner
2007-05-28 4:29 ` David Chinner
2007-05-31 0:46 ` Neil Brown
2007-05-31 0:57 ` Alasdair G Kergon
2007-05-31 0:57 ` [dm-devel] " Alasdair G Kergon
2007-05-31 1:07 ` Alasdair G Kergon
2007-05-31 1:07 ` [dm-devel] " Alasdair G Kergon
2007-05-31 1:11 ` David Chinner
2007-05-31 1:11 ` [dm-devel] " David Chinner
2007-05-28 4:48 ` Timothy Shimmin
2007-05-29 6:45 ` Jeremy Higdon
2007-05-29 20:03 ` Phillip Susi
2007-05-29 23:48 ` David Chinner
2007-05-30 0:01 ` david
2007-05-30 6:17 ` David Chinner
2007-05-30 8:55 ` Stefan Bader
2007-05-30 8:55 ` Stefan Bader
2007-05-30 16:52 ` david
2007-05-30 16:52 ` david
2007-05-31 0:20 ` David Chinner
2007-05-31 6:26 ` Jens Axboe
2007-05-31 7:03 ` David Chinner
2007-05-31 7:06 ` Jens Axboe
2007-05-31 13:30 ` Bill Davidsen
2007-05-31 13:36 ` Jens Axboe
2007-06-01 16:04 ` Bill Davidsen
2007-06-02 14:51 ` Jens Axboe
2007-06-02 19:55 ` Bill Davidsen
2007-06-01 3:16 ` Tejun Heo
2007-06-01 3:16 ` Tejun Heo
2007-06-01 8:21 ` Jens Axboe
2007-06-01 8:21 ` Jens Axboe
2007-06-02 9:20 ` Tejun Heo
2007-06-02 14:34 ` Jens Axboe
2007-06-02 14:34 ` Jens Axboe
2007-06-02 22:57 ` Guy Watkins
2007-06-02 22:57 ` Guy Watkins
2007-06-04 7:39 ` Tejun Heo
2007-05-31 18:31 ` Phillip Susi
2007-05-31 19:00 ` Jens Axboe
2007-05-31 19:21 ` david
2007-05-31 19:21 ` david
2007-05-31 19:40 ` Jens Axboe
2007-05-31 23:34 ` David Chinner
2007-06-01 5:59 ` Neil Brown
2007-06-01 6:11 ` Jens Axboe
2007-06-01 7:53 ` David Chinner
2007-06-01 23:56 ` Bill Davidsen
2007-05-31 18:24 ` Phillip Susi
2007-05-31 18:24 ` Phillip Susi
2007-05-30 16:45 ` Phillip Susi
2007-05-30 20:27 ` [dm-devel] " Phillip Susi
2007-05-31 6:24 ` Jens Axboe
2007-05-31 6:24 ` [dm-devel] " Jens Axboe
2007-05-31 18:37 ` Phillip Susi
2007-05-31 18:58 ` Jens Axboe
2007-06-02 0:04 ` Bill Davidsen
2007-05-28 9:29 ` Tejun Heo
2007-05-28 9:43 ` Alasdair G Kergon
2007-05-28 9:43 ` [dm-devel] " Alasdair G Kergon
2007-05-29 9:25 ` Stefan Bader
2007-05-29 22:05 ` Alasdair G Kergon [this message]
2007-05-29 22:05 ` Alasdair G Kergon
2007-05-30 9:12 ` Stefan Bader
2007-05-30 10:41 ` Alasdair G Kergon
2007-05-30 10:41 ` [dm-devel] " Alasdair G Kergon
2007-05-30 16:55 ` Phillip Susi
2007-05-30 16:55 ` [dm-devel] " Phillip Susi
2007-05-31 11:14 ` Stefan Bader
2007-06-01 3:25 ` Tejun Heo
2007-06-01 3:25 ` [dm-devel] " Tejun Heo
2007-06-01 5:55 ` david
2007-06-01 5:55 ` [dm-devel] " david
2007-06-01 7:16 ` Tejun Heo
2007-06-01 17:07 ` Valdis.Kletnieks
2007-06-01 18:09 ` Tejun Heo
2007-07-10 18:39 ` Ric Wheeler
2007-07-10 23:40 ` Valdis.Kletnieks
2007-07-11 2:49 ` Tejun Heo
2007-07-11 22:44 ` Ric Wheeler
2007-07-12 17:34 ` Valdis.Kletnieks
2007-07-12 19:43 ` Ric Wheeler
2007-07-12 23:10 ` Guy Watkins
2007-07-12 23:10 ` Guy Watkins
2007-07-13 11:30 ` Ric Wheeler
2007-07-11 2:51 ` Tejun Heo
2007-05-29 19:59 ` Phillip Susi
2007-05-31 0:22 ` Neil Brown
2007-05-30 9:35 ` Jens Axboe
2007-07-05 12:28 ` Tejun Heo
2007-07-09 12:27 ` Jens Axboe
2007-07-18 10:56 ` [PATCH] block: cosmetic changes Tejun Heo
2007-07-18 10:56 ` Tejun Heo
2007-07-18 10:59 ` [PATCH] block: factor out bio_check_eod() Tejun Heo
2007-07-18 10:59 ` Tejun Heo
2007-07-18 11:06 ` Jens Axboe
2007-07-18 11:18 ` Tejun Heo
2007-07-18 11:18 ` Tejun Heo
2007-07-18 11:31 ` Jens Axboe
2007-07-18 11:31 ` Jens Axboe
2007-07-18 11:33 ` Tejun Heo
2007-07-18 11:33 ` Tejun Heo
2007-07-18 11:34 ` Jens Axboe
2007-07-18 11:34 ` Jens Axboe
2007-07-18 11:41 ` Tejun Heo
2007-07-18 11:41 ` Tejun Heo
2007-07-18 11:45 ` Jens Axboe
2007-07-18 11:49 ` Jens Axboe
2007-07-18 12:34 ` Tejun Heo
2007-07-18 12:31 ` Jens Axboe
2007-05-28 11:17 ` [RFD] BIO_RW_BARRIER - what it means for devices, filesystems, and dm/md Nikita Danilov
2007-05-31 3:31 ` Neil Brown
2007-05-28 14:43 ` Bill Davidsen
2007-05-28 14:43 ` Bill Davidsen
2007-05-31 0:37 ` Neil Brown
2007-05-31 12:28 ` Bill Davidsen
2007-05-31 12:28 ` Bill Davidsen
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=20070529220500.GA6513@agk.fab.redhat.com \
--to=agk@redhat.com \
--cc=adilger@clusterfs.com \
--cc=dgc@sgi.com \
--cc=dm-devel@redhat.com \
--cc=htejun@gmail.com \
--cc=jens.axboe@oracle.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-raid@vger.kernel.org \
--cc=sbader3@googlemail.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 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.