[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <m1hbqs2k1k.fsf@fess.ebiederm.org>
Date: Mon, 11 Jan 2010 17:23:19 -0800
From: ebiederm@...ssion.com (Eric W. Biederman)
To: Benjamin LaHaise <bcrl@...et.ca>
Cc: Tejun Heo <tj@...nel.org>, Greg Kroah-Hartman <gregkh@...e.de>,
Kay Sievers <kay.sievers@...y.org>,
linux-kernel@...r.kernel.org,
Cornelia Huck <cornelia.huck@...ibm.com>,
linux-fsdevel@...r.kernel.org,
Eric Dumazet <eric.dumazet@...il.com>,
Serge Hallyn <serue@...ibm.com>,
"Eric W. Biederman" <ebiederm@...stanetworks.com>
Subject: Re: [PATCH 3/7] sysfs: Keep an nlink count on sysfs directories.
Benjamin LaHaise <bcrl@...et.ca> writes:
> On Mon, Jan 11, 2010 at 05:06:53PM -0800, Eric W. Biederman wrote:
>> I don't see the link count as interesting enough to store more than
>> 16bits for it. Even with 32bits of storage for nlink sysfs would have
>> to have to handle the rollover case as I am doing now. So I don't
>> see any advantage to storing more bits.
>
> I'm not terribly concerned with what value gets returned, but rather about
> how long is spent calculating it. Ideally, new sysfs directory entries
> can be inserted in an O(1) or other reasonable order operation. I'll try
> to find some time to re-run my interface scaling tests with your latest
> changes.
Sounds good. My changes so far have been the easy low hanging fruit.
readdir should be at O(N) and stat etc should all be constant time.
Lookups by name are still O(N), and I'm not ready to look at going
better until after I have network namespace support into sysfs.
Eric
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists