From: Alan Cox <alan@lxorguk.ukuu.org.uk>
To: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Andi Kleen <ak@suse.de>,
Linux Memory Management <linux-mm@kvack.org>,
linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 2.6.13] lockless pagecache 2/7
Date: Sun, 04 Sep 2005 09:20:18 +0100 [thread overview]
Message-ID: <1125822018.23858.2.camel@localhost.localdomain> (raw)
In-Reply-To: <431A4767.4030403@yahoo.com.au>
On Sul, 2005-09-04 at 11:01 +1000, Nick Piggin wrote:
> I would be surprised if it was a big loss... but I'm assuming
> a locked cmpxchg isn't outlandishly expensive. Basically:
>
> read_lock_irqsave(cacheline1);
> atomic_inc_return(cacheline2);
> read_unlock_irqrestore(cacheline1);
>
> Turns into
>
> atomic_cmpxchg();
>
> I'll do some microbenchmarks and get back to you. I'm quite
> interested now ;) What sort of AMDs did you have in mind,
Athlon or higher give very different atomic numbers to P4. If you are
losing the read_lock/unlock then the atomic_cmpxchg should be faster on
all I agree.
One question however - atomic_foo operations are not store barriers so
you might need mb() and friends for PPC ?
Alan
WARNING: multiple messages have this Message-ID (diff)
From: Alan Cox <alan@lxorguk.ukuu.org.uk>
To: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Andi Kleen <ak@suse.de>,
Linux Memory Management <linux-mm@kvack.org>,
linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 2.6.13] lockless pagecache 2/7
Date: Sun, 04 Sep 2005 09:20:18 +0100 [thread overview]
Message-ID: <1125822018.23858.2.camel@localhost.localdomain> (raw)
In-Reply-To: <431A4767.4030403@yahoo.com.au>
On Sul, 2005-09-04 at 11:01 +1000, Nick Piggin wrote:
> I would be surprised if it was a big loss... but I'm assuming
> a locked cmpxchg isn't outlandishly expensive. Basically:
>
> read_lock_irqsave(cacheline1);
> atomic_inc_return(cacheline2);
> read_unlock_irqrestore(cacheline1);
>
> Turns into
>
> atomic_cmpxchg();
>
> I'll do some microbenchmarks and get back to you. I'm quite
> interested now ;) What sort of AMDs did you have in mind,
Athlon or higher give very different atomic numbers to P4. If you are
losing the read_lock/unlock then the atomic_cmpxchg should be faster on
all I agree.
One question however - atomic_foo operations are not store barriers so
you might need mb() and friends for PPC ?
Alan
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2005-09-05 16:32 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-09-02 6:25 New lockless pagecache Nick Piggin
2005-09-02 6:25 ` Nick Piggin
2005-09-02 6:28 ` [PATCH 2.6.13] lockless pagecache 1/7 Nick Piggin
2005-09-02 6:29 ` [PATCH 2.6.13] lockless pagecache 2/7 Nick Piggin
2005-09-02 6:30 ` [PATCH 2.6.13] lockless pagecache 3/7 Nick Piggin
2005-09-02 6:30 ` [PATCH 2.6.13] lockless pagecache 4/7 Nick Piggin
2005-09-02 6:31 ` [PATCH 2.6.13] lockless pagecache 5/7 Nick Piggin
2005-09-02 6:32 ` [PATCH 2.6.13] lockless pagecache 6/7 Nick Piggin
2005-09-02 6:32 ` [PATCH 2.6.13] lockless pagecache 7/7 Nick Piggin
2005-09-09 13:00 ` Christoph Lameter
2005-09-09 13:00 ` Christoph Lameter
2005-09-09 15:23 ` Nick Piggin
2005-09-09 15:23 ` Nick Piggin
2005-09-09 5:36 ` [PATCH 2.6.13] lockless pagecache 5/7 Christoph Lameter
2005-09-09 5:36 ` Christoph Lameter
2005-09-09 6:22 ` Nick Piggin
2005-09-09 6:22 ` Nick Piggin
2005-09-02 13:08 ` [PATCH 2.6.13] lockless pagecache 2/7 Alan Cox
2005-09-02 13:08 ` Alan Cox
2005-09-02 20:41 ` Andi Kleen
2005-09-02 20:41 ` Andi Kleen
2005-09-02 21:12 ` David S. Miller
2005-09-02 21:12 ` David S. Miller, Andi Kleen
2005-09-02 21:43 ` Nick Piggin
2005-09-02 21:43 ` Nick Piggin
2005-09-02 21:22 ` Nick Piggin
2005-09-02 21:22 ` Nick Piggin
2005-09-02 21:31 ` David S. Miller
2005-09-02 21:31 ` David S. Miller, Nick Piggin
2005-09-02 21:47 ` Nick Piggin
2005-09-02 21:47 ` Nick Piggin
2005-09-02 21:57 ` David S. Miller
2005-09-02 21:57 ` David S. Miller, Nick Piggin
2005-09-02 23:57 ` Alan Cox
2005-09-02 23:57 ` Alan Cox
2005-09-03 1:40 ` Nick Piggin
2005-09-03 1:40 ` Nick Piggin
2005-09-03 17:31 ` Alan Cox
2005-09-03 17:31 ` Alan Cox
2005-09-04 1:01 ` Nick Piggin
2005-09-04 1:01 ` Nick Piggin
2005-09-04 8:20 ` Alan Cox [this message]
2005-09-04 8:20 ` Alan Cox
2005-09-06 1:03 ` Nick Piggin
2005-09-06 1:03 ` Nick Piggin
2005-09-02 18:26 ` Christoph Lameter
2005-09-02 18:26 ` Christoph Lameter
2005-09-02 21:26 ` Nick Piggin
2005-09-02 21:26 ` Nick Piggin
2005-09-03 1:33 ` Christoph Lameter
2005-09-03 1:33 ` Christoph Lameter
2005-09-02 6:45 ` New lockless pagecache Nick Piggin
2005-09-02 6:45 ` Nick Piggin
2005-09-15 19:50 ` Alok kataria
2005-09-16 3:12 ` Nick Piggin
2005-09-16 3:12 ` Nick Piggin
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=1125822018.23858.2.camel@localhost.localdomain \
--to=alan@lxorguk.ukuu.org.uk \
--cc=ak@suse.de \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=nickpiggin@yahoo.com.au \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.