From mboxrd@z Thu Jan  1 00:00:00 1970
From: Christoph Lameter <cl@linux.com>
Subject: Re: Poll about irqsafe_cpu_add and others
Date: Thu, 17 Mar 2011 14:21:33 -0500 (CDT)
Message-ID: <alpine.DEB.2.00.1103171419280.18529@router.home>
References: <1300371834.6315.93.camel@edumazet-laptop>  <20110317.081420.71114992.davem@davemloft.net>  <alpine.DEB.2.00.1103171016270.12540@router.home>  <1300380801.6315.306.camel@edumazet-laptop>  <alpine.DEB.2.00.1103171245430.16620@router.home>
 <1300386569.6315.404.camel@edumazet-laptop>  <alpine.DEB.2.00.1103171338350.18529@router.home> <1300388139.6315.418.camel@edumazet-laptop>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Return-path: <linux-arch-owner@vger.kernel.org>
Received: from smtp110.prem.mail.ac4.yahoo.com ([76.13.13.93]:20788 "HELO
	smtp110.prem.mail.ac4.yahoo.com" rhost-flags-OK-OK-OK-OK)
	by vger.kernel.org with SMTP id S1755285Ab1CQTVi (ORCPT
	<rfc822;linux-arch@vger.kernel.org>);
	Thu, 17 Mar 2011 15:21:38 -0400
In-Reply-To: <1300388139.6315.418.camel@edumazet-laptop>
Sender: linux-arch-owner@vger.kernel.org
List-ID: <linux-arch.vger.kernel.org>
To: Eric Dumazet <eric.dumazet@gmail.com>
Cc: David Miller <davem@davemloft.net>, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, netdev@vger.kernel.org, netfilter-devel@vger.kernel.org

On Thu, 17 Mar 2011, Eric Dumazet wrote:

> > > I wonder why we dont use :
> > >
> > > 	addl	$0x2,%fs:xt_u64
> > > 	addcl	$0x0,%fs:xt_u64+4
> >
> > The compiler is fed the following
> >
> > 	*__this_cpu_ptr(xt_u64) += 2
> >
> > __this_cpu_ptr makes it:
> >
> > 	*(xt_u64 + __my_cpu_offset) += 2
> >
> > So the compiler calculates the address first and then increments it.
> >
> > The compiler could optimize this I think. Wonder why that does not happen.
>
> Compiler is really forced to compute addr, thats why.
>
> Hmm, we should not fallback to generic ops I think, but tweak
>
> percpu_add_op() {

percpu_add_op() is not used. This is a 64 bit operation on a 32 bit
machine thus we fall back to this_cpu_generic_to_op()

#define __this_cpu_generic_to_op(pcp, val, op)                          \
do {                                                                    \
        *__this_cpu_ptr(&(pcp)) op val;                                 \
} while (0)