From: Eric Dumazet <dada1@cosmosbay.com>
To: paulmck@us.ibm.com
Cc: Andrew Morton <akpm@osdl.org>, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] shrinks dentry struct
Date: Tue, 13 Dec 2005 19:24:57 +0100 [thread overview]
Message-ID: <439F11F9.8070300@cosmosbay.com> (raw)
In-Reply-To: <20051213180315.GB14158@us.ibm.com>
Paul E. McKenney a écrit :
> On Tue, Nov 29, 2005 at 04:22:00PM +0100, Eric Dumazet wrote:
>
>>Hi Andrew
>>
>>Could you add this patch to mm ?
>>
>>Thank you
>>
>>[PATCH] shrinks dentry struct
>>
>>Some long time ago, dentry struct was carefully tuned so that on 32 bits
>>UP, sizeof(struct dentry) was exactly 128, ie a power of 2, and a multiple
>>of memory cache lines.
>>
>>Then RCU was added and dentry struct enlarged by two pointers, with nice
>>results for SMP, but not so good on UP, because breaking the above tuning
>>(128 + 8 = 136 bytes)
>>
>>This patch reverts this unwanted side effect, by using an union (d_u),
>>where d_rcu and d_child are placed so that these two fields can share their
>>memory needs.
>>
>>At the time d_free() is called (and d_rcu is really used), d_child is known
>>to be empty and not touched by the dentry freeing.
>>
>>Lockless lookups only access d_name, d_parent, d_lock, d_op, d_flags (so
>>the previous content of d_child is not needed if said dentry was unhashed
>>but still accessed by a CPU because of RCU constraints)
>>
>>As dentry cache easily contains millions of entries, a size reduction is
>>worth the extra complexity of the ugly C union.
>
>
> Looks sound to me! Some opportunities for simplification below.
>
> (Please accept my apologies for the delay -- some diversions turned out
> to be more consuming than I had expected.)
>
> Thanx, Paul
>
Hi Paul
My patch only address the layout of dentry structure, basically a 'global
substitute' on various places.
Adding some 'optimizations' or simplifications was not the goal, so please
submit a patch if you feel the need for it :)
Thank you
Eric
prev parent reply other threads:[~2005-12-13 18:25 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-11-29 7:17 [PATCH] race condition in procfs Grzegorz Nosek
2005-11-29 8:09 ` Andrew Morton
2005-11-29 8:38 ` Grzegorz Nosek
2005-11-29 13:25 ` Grzegorz Nosek
2005-11-29 14:04 ` Grzegorz Nosek
2005-11-29 14:28 ` Steven Rostedt
2005-11-29 14:39 ` Grzegorz Nosek
2005-11-29 14:49 ` Steven Rostedt
2005-11-30 14:41 ` Grzegorz Nosek
2005-11-30 15:14 ` Steven Rostedt
2005-11-30 15:29 ` Grzegorz Nosek
2005-11-30 16:25 ` Steven Rostedt
2005-11-30 17:23 ` Grzegorz Nosek
2005-12-01 20:38 ` Grzegorz Nosek
2005-11-29 15:22 ` [PATCH] shrinks dentry struct Eric Dumazet
2005-11-30 2:06 ` Paul Jackson
2005-11-30 2:14 ` Andrew Morton
2005-11-30 2:43 ` Paul Jackson
2005-11-30 6:56 ` Hugh Dickins
2005-12-03 1:15 ` [PATCH] remove unused blkp field in percpu_data Eric Dumazet
2005-12-13 18:03 ` [PATCH] shrinks dentry struct Paul E. McKenney
2005-12-13 18:24 ` Eric Dumazet [this message]
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=439F11F9.8070300@cosmosbay.com \
--to=dada1@cosmosbay.com \
--cc=akpm@osdl.org \
--cc=linux-kernel@vger.kernel.org \
--cc=paulmck@us.ibm.com \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox