From: Andreas Dilger <adilger@clusterfs.com>
To: Sage Weil <sage@newdream.net>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>,
Christoph Hellwig <hch@infradead.org>,
Brad Boyer <flar@allandria.com>,
Anton Altaparmakov <aia21@cam.ac.uk>,
Gary Grider <ggrider@lanl.gov>,
linux-fsdevel@vger.kernel.org
Subject: Re: NFSv4/pNFS possible POSIX I/O API standards
Date: Mon, 4 Dec 2006 00:32:00 -0700 [thread overview]
Message-ID: <20061204073200.GB5637@schatzie.adilger.int> (raw)
In-Reply-To: <Pine.LNX.4.62.0612021830030.31986@wtf.di.newdream.net>
On Dec 03, 2006 08:10 -0800, Sage Weil wrote:
> My only concern is the "at least as recent as the opendir()" part,
> in contrast to statlite(), which has undefined "recentness" of its
> result for fields not specified in the mask.
>
> Ideally, I'd like to see readdirplus() also take a statlite() style mask,
> so that you can choose between either "vaguely recent" and "at least as
> recent as opendir()".
In my opinion (which may not be that of the original statlite() authors)
is that the flags should really be called "valid", and any bit set in
the flag can be considered valid, and any unset bit means "this field
has no valid data". Having it mean "it might be out of date" gives the
false impression that it might contain valid (if slightly out of date)
data.
> As you mentioned, by the time you look at the result of any call (in the
> absence of locking) it may be out of date. But simply establishing an
> ordering is useful, especially in a clustered environment where some nodes
> are waiting for other nodes (via barriers or whatever) and then want to
> see the effects of previously completed fs operations.
Ah, OK. I didn't understand what you were getting at before. I agree
that it makes sense to have the same semantics as readdir() in this regard.
> I'm wondering if a corresponding opendirplus() (or similar) would also be
> appropriate to inform the kernel/filesystem that readdirplus() will
> follow, and stat information should be gathered/buffered. Or do most
> implementations wait for the first readdir() before doing any actual work
> anyway?
I'm not sure what some filesystems might do here. I suppose NFS has weak
enough cache semantics that it _might_ return stale cached data from the
client in order to fill the readdirplus() data, but it is just as likely
that it ships the whole thing to the server and returns everything in
one shot. That would imply everything would be at least as up-to-date
as the opendir().
Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.
next prev parent reply other threads:[~2006-12-04 7:32 UTC|newest]
Thread overview: 124+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-11-28 4:34 NFSv4/pNFS possible POSIX I/O API standards Gary Grider
2006-11-28 5:54 ` Christoph Hellwig
2006-11-28 10:54 ` Andreas Dilger
2006-11-28 11:28 ` Anton Altaparmakov
2006-11-28 20:17 ` Russell Cattelan
2006-11-28 23:28 ` Wendy Cheng
2006-11-29 9:12 ` Christoph Hellwig
2006-11-29 9:04 ` Christoph Hellwig
2006-11-29 9:14 ` Christoph Hellwig
2006-11-29 9:48 ` Andreas Dilger
2006-11-29 10:18 ` Anton Altaparmakov
2006-11-29 8:26 ` Brad Boyer
2006-11-30 9:25 ` Christoph Hellwig
2006-11-30 17:49 ` Sage Weil
2006-12-01 5:26 ` Trond Myklebust
2006-12-01 7:08 ` Sage Weil
2006-12-01 14:41 ` Trond Myklebust
2006-12-01 16:47 ` Sage Weil
2006-12-01 18:07 ` Trond Myklebust
2006-12-01 18:42 ` Sage Weil
2006-12-01 19:13 ` Trond Myklebust
2006-12-01 20:32 ` Sage Weil
2006-12-04 18:02 ` Peter Staubach
2006-12-05 23:20 ` readdirplus() as possible POSIX I/O API Sage Weil
2006-12-06 15:48 ` Peter Staubach
2006-12-03 1:57 ` NFSv4/pNFS possible POSIX I/O API standards Andreas Dilger
2006-12-03 7:34 ` Kari Hurtta
2006-12-03 1:52 ` Andreas Dilger
2006-12-03 16:10 ` Sage Weil
2006-12-04 7:32 ` Andreas Dilger [this message]
2006-12-04 15:15 ` Trond Myklebust
2006-12-05 0:59 ` Rob Ross
2006-12-05 4:44 ` Gary Grider
2006-12-05 10:05 ` Christoph Hellwig
2006-12-05 5:56 ` Trond Myklebust
2006-12-05 10:07 ` Christoph Hellwig
2006-12-05 14:20 ` Matthew Wilcox
2006-12-06 15:04 ` Rob Ross
2006-12-06 15:44 ` Matthew Wilcox
2006-12-06 16:15 ` Rob Ross
2006-12-05 14:55 ` Trond Myklebust
2006-12-05 22:11 ` Rob Ross
2006-12-05 23:24 ` Trond Myklebust
2006-12-06 16:42 ` Rob Ross
2006-12-06 12:22 ` Ragnar Kjørstad
2006-12-06 15:14 ` Trond Myklebust
2006-12-05 16:55 ` Latchesar Ionkov
2006-12-05 22:12 ` Christoph Hellwig
2006-12-06 23:12 ` Latchesar Ionkov
2006-12-06 23:33 ` Trond Myklebust
2006-12-05 21:50 ` Rob Ross
2006-12-05 22:05 ` Christoph Hellwig
2006-12-05 23:18 ` Sage Weil
2006-12-05 23:55 ` Ulrich Drepper
2006-12-06 10:06 ` Andreas Dilger
2006-12-06 17:19 ` Ulrich Drepper
2006-12-06 17:27 ` Rob Ross
2006-12-06 17:42 ` Ulrich Drepper
2006-12-06 18:01 ` Ragnar Kjørstad
2006-12-06 18:13 ` Ulrich Drepper
2006-12-17 14:41 ` Ragnar Kjørstad
2006-12-17 19:07 ` Ulrich Drepper
2006-12-17 19:38 ` Matthew Wilcox
2006-12-17 21:51 ` Ulrich Drepper
2006-12-18 2:57 ` Ragnar Kjørstad
2006-12-18 3:54 ` Gary Grider
2006-12-07 5:57 ` Andreas Dilger
2006-12-15 22:37 ` Ulrich Drepper
2006-12-16 18:13 ` Andreas Dilger
2006-12-16 19:08 ` Ulrich Drepper
2006-12-14 23:58 ` statlite() Rob Ross
2006-12-07 23:39 ` NFSv4/pNFS possible POSIX I/O API standards Nikita Danilov
2006-12-05 14:37 ` Peter Staubach
2006-12-05 10:26 ` readdirplus() as possible POSIX I/O API Andreas Dilger
2006-12-05 15:23 ` Trond Myklebust
2006-12-06 10:28 ` Andreas Dilger
2006-12-06 15:10 ` Trond Myklebust
2006-12-05 17:06 ` Latchesar Ionkov
2006-12-05 22:48 ` Rob Ross
2006-11-29 10:25 ` NFSv4/pNFS possible POSIX I/O API standards Steven Whitehouse
2006-11-30 12:29 ` Christoph Hellwig
2006-12-01 15:52 ` Ric Wheeler
2006-11-29 12:23 ` Matthew Wilcox
2006-11-29 12:35 ` Matthew Wilcox
2006-11-29 16:26 ` Gary Grider
2006-11-29 17:18 ` Christoph Hellwig
2006-11-29 12:39 ` Christoph Hellwig
2006-12-01 22:29 ` Rob Ross
2006-12-02 2:35 ` Latchesar Ionkov
2006-12-05 0:37 ` Rob Ross
2006-12-05 10:02 ` Christoph Hellwig
2006-12-05 16:47 ` Latchesar Ionkov
2006-12-05 17:01 ` Matthew Wilcox
[not found] ` <f158dc670612050909m366594c5ubaa87d9a9ecc8c2a@mail.gmail.com>
2006-12-05 17:10 ` Latchesar Ionkov
2006-12-05 17:39 ` Matthew Wilcox
2006-12-05 21:55 ` Rob Ross
2006-12-05 21:50 ` Peter Staubach
2006-12-05 21:44 ` Rob Ross
2006-12-06 11:01 ` openg Christoph Hellwig
2006-12-06 15:41 ` openg Trond Myklebust
2006-12-06 15:42 ` openg Rob Ross
2006-12-06 23:32 ` openg Christoph Hellwig
2006-12-14 23:36 ` openg Rob Ross
2006-12-06 23:25 ` Re: NFSv4/pNFS possible POSIX I/O API standards Latchesar Ionkov
2006-12-06 9:48 ` David Chinner
2006-12-06 15:53 ` openg and path_to_handle Rob Ross
2006-12-06 16:04 ` Matthew Wilcox
2006-12-06 16:20 ` Rob Ross
2006-12-06 20:57 ` David Chinner
2006-12-06 20:40 ` David Chinner
2006-12-06 20:50 ` Matthew Wilcox
2006-12-06 21:09 ` David Chinner
2006-12-06 22:09 ` Andreas Dilger
2006-12-06 22:17 ` Matthew Wilcox
2006-12-06 22:41 ` Andreas Dilger
2006-12-06 23:39 ` Christoph Hellwig
2006-12-14 22:52 ` Rob Ross
2006-12-06 20:50 ` Rob Ross
2006-12-06 21:01 ` David Chinner
2006-12-06 23:19 ` Latchesar Ionkov
2006-12-14 21:00 ` Rob Ross
2006-12-14 21:20 ` Matthew Wilcox
2006-12-14 23:02 ` Rob Ross
2006-11-28 15:08 ` NFSv4/pNFS possible POSIX I/O API standards Matthew Wilcox
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=20061204073200.GB5637@schatzie.adilger.int \
--to=adilger@clusterfs.com \
--cc=aia21@cam.ac.uk \
--cc=flar@allandria.com \
--cc=ggrider@lanl.gov \
--cc=hch@infradead.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=sage@newdream.net \
--cc=trond.myklebust@fys.uio.no \
/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).