Date: Mon, 13 Jul 2015 20:56:46 +0100
From: Al Viro <viro@ZenIV.linux.org.uk>
To: Ben Myers <bpm@sgi.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
        "J. Bruce Fields" <bfields@fieldses.org>, linux-kernel@vger.kernel.org,
        linux-fsdevel@vger.kernel.org
Subject: Re: [RFC] freeing unlinked file indefinitely delayed
Message-ID: <20150713195645.GR17109@ZenIV.linux.org.uk>
References: <20150708014237.GC17109@ZenIV.linux.org.uk>
 <20150708154143.GG4015@sgi.com>
 <20150712150035.GJ17109@ZenIV.linux.org.uk>
 <20150713181751.GZ4568@sgi.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20150713181751.GZ4568@sgi.com>
User-Agent: Mutt/1.5.21 (2010-09-15)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2049
Lines: 43

On Mon, Jul 13, 2015 at 01:17:51PM -0500, Ben Myers wrote:
> > For one thing, this patch does *not* check for i_nlink at all.
> 
> I agree that no checking of i_nlink has the advantage of brevity.
> Anyone who is using dentry.d_fsdata with an open_by_handle workload (if
> there are any) will be affected.

Translate, please.  What does d_fsdata have to anything above?

> > For another, there's no such thing as 'filesystems internal lock' for
> > i_nlink protection - that's handled by i_mutex...  And what does
> > iget() have to do with any of that?
> 
> i_mutex is good enough only for local filesystems.
> Network/clustered/distributed filesystems need to take an internal lock
> to provide exclusion for this .unlink with a .link on another host.
> That's where I'm coming from with iget().  
> 
> Maybe plumbing i_op.unlink with another argument to return i_nlink is
> something to consider?  A helper for the few filesystems that need to do
> this might be good enough in the near term.

????

a) iget() had been gone since way back
b) it never had been called by VFS - it's a filesystem's responsibility
c) again, what the hell does iget() or its replacements have to do with
dentry eviction?  It does *NOT* affect dentry refcount.  Never had.
d) checks for _inode_ retention in icache are done by filesystem code, which
is certainly free to use its locks.  Incidentally, for normal filesystems
no locks are needed at all - everything that changes ->i_nlink is holding
a referfence to in-core inode, so in a situation when its refcount is zero
and ->i_lock is held (providing an exclusion with icache lookups), ->i_nlink
is guaranteed to be stable.
e) why would VFS possibly want to know if there are links remaining after
successful ->unlink()?

I'm sorry, but you are not making any sense...
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/