Re: [PATCH RFC] NFSD: fix cannot umounting mount points under pseudo root

From: NeilBrown <neilb@suse.de>
To: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Kinglong Mee <kinglongmee@gmail.com>,
	"linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH RFC] NFSD: fix cannot umounting mount points under pseudo root
Date: Thu, 23 Apr 2015 09:44:31 +1000	[thread overview]
Message-ID: <20150423094431.1a8aa68b@notabene.brown> (raw)
In-Reply-To: <20150422150703.GA1247@fieldses.org>

[-- Attachment #1: Type: text/plain, Size: 1232 bytes --]

On Wed, 22 Apr 2015 11:07:03 -0400 "J. Bruce Fields" <bfields@fieldses.org>
wrote:

> > Reference of dentry/mnt is like a cache, avoids re-finding of them,
> > it is necessary to store them in svc_export.
> > 
> > Neil points out another way of 'fs_pin', I will check that.
> 
> Yes, that'd be interesting.  On a very quick look--I don't understand
> what it's meant to do at all.  But if it does provide a way to get a
> callback on umount, that'd certainly be interesting....

Yeah, on a quick look it isn't really obvious at all.

But I didn't read the code.  I read
 https://lwn.net/Articles/636730/

which says:

    In its place is a mechanism by which an object can be added to a vfsmount
    structure (which represents a mounted filesystem); that object supports 
    only one method: kill(). These "pin" objects hang around until the final
    reference to the vfsmount goes away, at which point each one's kill() function
    is called. It thus is a clean mechanism for performing cleanup when a filesystem
    goes away.

This is used to close "BSD process accounting" files on umount, and sound
like the perfect interface for flushing things from the sunrpc cache on
umount.

NeilBrown

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 811 bytes --]