From: Al Viro <viro@ZenIV.linux.org.uk>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>,
Andi Kleen <andi@firstfloor.org>,
linux-kernel@vger.kernel.org, linux-nfs@vger.kernel.org
Subject: Re: [RFC PATCH 0/2] Fix up the NFS mmap code
Date: Sat, 9 Jan 2010 02:30:35 +0000 [thread overview]
Message-ID: <20100109023035.GH30528@ZenIV.linux.org.uk> (raw)
In-Reply-To: <alpine.LFD.2.00.1001081814240.7821-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
On Fri, Jan 08, 2010 at 06:22:01PM -0800, Linus Torvalds wrote:
>
>
> On Sat, 9 Jan 2010, Al Viro wrote:
> >
> > Well, sure - it steps on i_mutex-before-mmmap_sem first from ls somewhere and
> > records the ordering for posterity. Then NFS steps into mmap() (on a
> > different inode) and gets conflicting ordering.
>
> Look closer: the inodes for directories and for non-directories have
> i_mutex in different lockdep classes.
>
> So that "on a different inode" thing should have made it a non-issue,
> since there is no actual chain back. There is "mmap_sem ->
> i_mutex_regular_file" (for mmap) and there is "i_mutex_directory ->
> mmap_sem" (for filldir), but that isn't an ABBA.
>
> The problem _seems_ to be (if I read Andi's chain correctly) that a
> directory hasn't gone through the i_mutex_dir_key change, so filldir ends
> up being counted against the default i_mutex_key.
Interesting... There's nfs_update_inode(), but it ought to scream bloody
murder on the type change (and its return value is ignore by nfs_fhget(),
BTW).
WARNING: multiple messages have this Message-ID (diff)
From: Al Viro <viro@ZenIV.linux.org.uk>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>,
Andi Kleen <andi@firstfloor.org>,
linux-kernel@vger.kernel.org, linux-nfs@vger.kernel.org
Subject: Re: [RFC PATCH 0/2] Fix up the NFS mmap code
Date: Sat, 9 Jan 2010 02:30:35 +0000 [thread overview]
Message-ID: <20100109023035.GH30528@ZenIV.linux.org.uk> (raw)
In-Reply-To: <alpine.LFD.2.00.1001081814240.7821@localhost.localdomain>
On Fri, Jan 08, 2010 at 06:22:01PM -0800, Linus Torvalds wrote:
>
>
> On Sat, 9 Jan 2010, Al Viro wrote:
> >
> > Well, sure - it steps on i_mutex-before-mmmap_sem first from ls somewhere and
> > records the ordering for posterity. Then NFS steps into mmap() (on a
> > different inode) and gets conflicting ordering.
>
> Look closer: the inodes for directories and for non-directories have
> i_mutex in different lockdep classes.
>
> So that "on a different inode" thing should have made it a non-issue,
> since there is no actual chain back. There is "mmap_sem ->
> i_mutex_regular_file" (for mmap) and there is "i_mutex_directory ->
> mmap_sem" (for filldir), but that isn't an ABBA.
>
> The problem _seems_ to be (if I read Andi's chain correctly) that a
> directory hasn't gone through the i_mutex_dir_key change, so filldir ends
> up being counted against the default i_mutex_key.
Interesting... There's nfs_update_inode(), but it ought to scream bloody
murder on the type change (and its return value is ignore by nfs_fhget(),
BTW).
next prev parent reply other threads:[~2010-01-09 2:30 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-01-07 20:29 [GIT PULL] Please pull NFS client bugfixes Trond Myklebust
2010-01-07 21:00 ` Andi Kleen
2010-01-07 21:23 ` Peter Staubach
2010-01-07 21:35 ` Andi Kleen
2010-01-07 21:53 ` Trond Myklebust
2010-01-07 23:51 ` Andi Kleen
2010-01-08 0:14 ` Trond Myklebust
2010-01-08 0:34 ` Linus Torvalds
2010-01-08 0:45 ` Andi Kleen
2010-01-08 1:03 ` Trond Myklebust
2010-01-08 1:03 ` Trond Myklebust
2010-01-08 1:12 ` Linus Torvalds
2010-01-08 1:22 ` Trond Myklebust
2010-01-08 1:26 ` Trond Myklebust
2010-01-09 0:56 ` [RFC PATCH 0/2] Fix up the NFS mmap code Trond Myklebust
2010-01-09 0:56 ` [RFC PATCH 2/2] NFS: Fix a potential deadlock in nfs_file_mmap() Trond Myklebust
[not found] ` <20100109005624.7473.15560.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-09 1:54 ` Al Viro
2010-01-09 1:54 ` Al Viro
[not found] ` <20100109005624.7473.33215.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-09 0:56 ` [RFC PATCH 1/2] VFS: Add a mmap_file() callback to struct file_operations Trond Myklebust
2010-01-09 0:56 ` Trond Myklebust
2010-01-09 1:17 ` [RFC PATCH 0/2] Fix up the NFS mmap code Linus Torvalds
2010-01-09 1:17 ` Linus Torvalds
[not found] ` <alpine.LFD.2.00.1001081709470.7821-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-09 1:38 ` Al Viro
2010-01-09 1:38 ` Al Viro
2010-01-09 1:46 ` Al Viro
2010-01-09 1:57 ` Linus Torvalds
[not found] ` <alpine.LFD.2.00.1001081750080.7821-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-09 2:11 ` Al Viro
2010-01-09 2:11 ` Al Viro
2010-01-09 2:22 ` Linus Torvalds
[not found] ` <alpine.LFD.2.00.1001081814240.7821-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
2010-01-09 2:30 ` Al Viro [this message]
2010-01-09 2:30 ` Al Viro
2010-01-09 2:40 ` Al Viro
2010-01-09 2:43 ` Al Viro
2010-01-10 2:00 ` Andi Kleen
2010-01-10 2:00 ` Andi Kleen
2010-01-08 1:30 ` [GIT PULL] Please pull NFS client bugfixes Linus Torvalds
2010-01-08 1:35 ` Linus Torvalds
2010-01-08 2:00 ` Linus Torvalds
2010-01-14 13:18 ` Peter Zijlstra
2010-01-08 5:19 ` Andi Kleen
2010-01-08 1:22 ` Linus Torvalds
2010-01-08 0:43 ` Andi Kleen
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=20100109023035.GH30528@ZenIV.linux.org.uk \
--to=viro@zeniv.linux.org.uk \
--cc=Trond.Myklebust@netapp.com \
--cc=andi@firstfloor.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nfs@vger.kernel.org \
--cc=torvalds@linux-foundation.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.