From: "J. Bruce Fields" <bfields@fieldses.org>
To: Nick Piggin <npiggin@gmail.com>
Cc: linux-nfs@vger.kernel.org, linux-fsdevel@vger.kernel.org
Subject: Re: lifetime of DCACHE_DISCONECTED dentries
Date: Mon, 29 Nov 2010 14:32:48 -0500 [thread overview]
Message-ID: <20101129193248.GA9897@fieldses.org> (raw)
In-Reply-To: <AANLkTinMq1W1vx-jZnf7cb1cAnh9OyoVRYtVS7zpujfs@mail.gmail.com>
On Mon, Nov 29, 2010 at 02:56:22PM +1100, Nick Piggin wrote:
> On Tue, Nov 16, 2010 at 5:45 PM, Nick Piggin <npiggin@gmail.com> wrote:
> > On Tue, Nov 16, 2010 at 4:48 AM, J. Bruce Fields <bfields@fieldses.org> wrote:
> >> On Sat, Nov 13, 2010 at 10:53:12PM +1100, Nick Piggin wrote:
> >>> On Sat, Nov 13, 2010 at 5:43 AM, J. Bruce Fields <bfields@fieldses.org> wrote:
> >
> >>> > - putfh: look up the filehandle. The only alias found for the
> >>> > inode will be DCACHE_UNHASHED alias referenced by the filp
> >>> > associated with the nfsd open. d_obtain_alias() doesn't like
> >>> > this, so it creates a new DCACHE_DISCONECTED dentry and
> >>> > returns that instead.
> >>>
> >>> This seems to be where the thing goes wrong. It isn't a hashed dentry at
> >>> this point here, so d_obtain_alias should not be making one.
> >>
> >> Sounds sensible. (But can you think of any actual bugs that will result
> >> from trying to add a new hashed dentry in this case?)
> >
> > Well, this one? :)
> >
> >
> >>> I think the inode i_nlink games are much more appropriate on this side of
> >>> the equation, rather than the dput side (after all, d_obtain_alias is setting
> >>> up an alias for the inode).
> >>>
> >>> Can you even put the link check into __d_find_alias?
> >>>
> >>> - if (S_ISDIR(inode->i_mode) || !d_unhashed(alias)) {
> >>> + if (S_ISDIR(inode->i_mode) || !inode->i_nlink ||
> >>> !d_unhashed(alias)) {
> >>>
> >>> Something like that?
> >>
> >> The immediate result of that would be for the close rpc (or any rpc's
> >> sent after the file was unlinked) to fail with ESTALE.
> >
> > Why is that? Seems like it would be a bug, because a hashed dentry may
> > be unhashed at any time concurrently to nfsd operation, so it should be
> > able to tolerate that so long as it has a ref on the inode?
>
> Ping? Did you work out why nfs fails with ESTALE in that case? It seems
> to work in my testing (and do the right thing with freeing the inode).
Bah, sorry, I read too quickly, got the sense of the test backwards, and
thought you were suggesting __d_find_alias() shouldn't return an alias
in the i_nlink == 0 case!
Yes, agreed, that should solve my problem.
But what's the reason for the d_unhashed() check now? Could we get rid
of it entirely?
--b.
next prev parent reply other threads:[~2010-11-29 19:32 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-11-12 18:43 lifetime of DCACHE_DISCONECTED dentries J. Bruce Fields
2010-11-13 11:53 ` Nick Piggin
2010-11-15 17:48 ` J. Bruce Fields
2010-11-16 6:45 ` Nick Piggin
2010-11-29 3:56 ` Nick Piggin
2010-11-29 19:32 ` J. Bruce Fields [this message]
2010-11-30 1:00 ` Nick Piggin
2010-11-30 18:39 ` J. Bruce Fields
2010-12-03 22:33 ` [PATCH] nfsd4: allow __d_obtain_alias() to return unhashed dentries J. Bruce Fields
2010-12-13 5:19 ` Nick Piggin
2010-12-14 22:01 ` J. Bruce Fields
2010-12-17 17:53 ` [PATCH] fs/dcache: use standard list macro for d_find_alias J. Bruce Fields
2010-12-17 18:00 ` [PATCH 2/2] fs/dcache: allow __d_obtain_alias() to return unhashed dentries J. Bruce Fields
2010-12-18 2:01 ` Nick Piggin
2010-12-18 16:16 ` J. Bruce Fields
2010-12-19 14:53 ` Nick Piggin
2010-12-27 23:46 ` [PATCH] " J. Bruce Fields
2011-01-18 20:45 ` J. Bruce Fields
2011-01-18 22:02 ` Nick Piggin
2011-01-18 22:08 ` J. Bruce Fields
2011-03-08 18:13 ` J. Bruce Fields
2011-03-10 10:58 ` Al Viro
2011-03-11 4:07 ` NeilBrown
2012-02-14 17:03 ` J. Bruce Fields
2012-02-15 16:56 ` J. Bruce Fields
2012-02-16 3:06 ` NeilBrown
2012-02-16 11:51 ` J. Bruce Fields
2012-02-16 16:08 ` J. Bruce Fields
2012-02-16 22:30 ` J. Bruce Fields
2012-02-17 16:34 ` Peng Tao
2012-03-13 20:55 ` J. Bruce Fields
2012-03-13 20:58 ` [PATCH 1/2] vfs: stop d_splice_alias creating directory aliases J. Bruce Fields
2012-03-13 20:58 ` [PATCH 2/2] vfs: remove unused __d_splice_alias argument J. Bruce Fields
2012-02-20 2:55 ` [PATCH] fs/dcache: allow __d_obtain_alias() to return unhashed dentries NeilBrown
2012-02-29 23:10 ` J. Bruce Fields
2012-06-28 13:59 ` J. Bruce Fields
2012-06-29 20:10 ` J. Bruce Fields
2012-06-29 20:29 ` J. Bruce Fields
2012-07-01 23:15 ` NeilBrown
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=20101129193248.GA9897@fieldses.org \
--to=bfields@fieldses.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-nfs@vger.kernel.org \
--cc=npiggin@gmail.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).