From: ebiederm@xmission.com (Eric W. Biederman)
Subject: Re: [PATCH 3/7] sysfs: Keep an nlink count on sysfs directories.
Date: Tue, 12 Jan 2010 07:34:47 -0800
Message-ID: <m1k4vn72w8.fsf@fess.ebiederm.org>
References: <m1d41g9z3q.fsf@fess.ebiederm.org>
	<1263241315-19499-3-git-send-email-ebiederm@xmission.com>
	<20100112134134.5f55e95f@gondolin>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Greg Kroah-Hartman <gregkh@suse.de>,
	Kay Sievers <kay.sievers@vrfy.org>,
	linux-kernel@vger.kernel.org, Tejun Heo <tj@kernel.org>,
	linux-fsdevel@vger.kernel.org,
	Eric Dumazet <eric.dumazet@gmail.com>,
	Benjamin LaHaise <bcrl@lhnet.ca>,
	Serge Hallyn <serue@us.ibm.com>,
	"Eric W. Biederman" <ebiederm@aristanetworks.com>
To: Cornelia Huck <cornelia.huck@de.ibm.com>
Return-path: <linux-kernel-owner+glk-linux-kernel-3=40m.gmane.org-S1753081Ab0ALPfF@vger.kernel.org>
In-Reply-To: <20100112134134.5f55e95f@gondolin> (Cornelia Huck's message of "Tue\, 12 Jan 2010 13\:41\:34 +0100")
Sender: linux-kernel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org

Cornelia Huck <cornelia.huck@de.ibm.com> writes:

> On Mon, 11 Jan 2010 12:21:51 -0800,
> "Eric W. Biederman" <ebiederm@xmission.com> wrote:
>
>> From: Eric W. Biederman <ebiederm@aristanetworks.com>
>> 
>> On large directories sysfs_count_nlinks can be a significant
>> bottleneck, so keep a count in sysfs_dirent.  If we exceed
>> the maximum number of directory entries we can store return
>> nlink of 1.  An nlink of 1 matches what reiserfs does in
>> this case, and it let's find and similar utlities know that
>> we have a the directory nlink can not be used for optimization
>> purposes.
>> 
>> Signed-off-by: Eric W. Biederman <ebiederm@aristanetworks.com>
>> ---
>>  fs/sysfs/dir.c   |   20 ++++++++++++++++++++
>>  fs/sysfs/inode.c |   15 +--------------
>>  fs/sysfs/mount.c |    1 +
>>  fs/sysfs/sysfs.h |    1 +
>>  4 files changed, 23 insertions(+), 14 deletions(-)
>> 
>
>> @@ -420,6 +438,7 @@ int __sysfs_add_one(struct sysfs_addrm_cxt *acxt, struct sysfs_dirent *sd)
>>  		return -EEXIST;
>> 
>>  	sd->s_parent = sysfs_get(acxt->parent_sd);
>> +	sysfs_dir_inc_nlink(sd->s_parent);
>> 
>>  	sysfs_link_sibling(sd);
>> 
>> @@ -512,6 +531,7 @@ void sysfs_remove_one(struct sysfs_addrm_cxt *acxt, struct sysfs_dirent *sd)
>> 
>>  	BUG_ON(sd->s_flags & SYSFS_FLAG_REMOVED);
>> 
>> +	sysfs_dir_dec_nlink(sd->s_parent);
>>  	sysfs_unlink_sibling(sd);
>> 
>>  	/* Update timestamps on the parent */
>
> Shouldn't s_nlink be adjusted in sysfs_rename for the move case as well?

Yes.  I hereby nominate this as my most buggy patch ever.
Ugh.

I thought I actually cared about nlinks when I bothered to keep a
nlink count, but I am starting to think I really don't.

Eric