From mboxrd@z Thu Jan 1 00:00:00 1970 From: Brian Haley Subject: Re: [PATCH net-next-2.6] net: optimize INET input path further Date: Wed, 01 Dec 2010 13:00:21 -0500 Message-ID: <4CF68D35.8020701@hp.com> References: <1291179847.2856.452.camel@edumazet-laptop> <4CF687F3.7030107@hp.com> <1291225358.2856.1035.camel@edumazet-laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: David Miller , netdev To: Eric Dumazet Return-path: Received: from g5t0009.atlanta.hp.com ([15.192.0.46]:12750 "EHLO g5t0009.atlanta.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753899Ab0LASAY (ORCPT ); Wed, 1 Dec 2010 13:00:24 -0500 In-Reply-To: <1291225358.2856.1035.camel@edumazet-laptop> Sender: netdev-owner@vger.kernel.org List-ID: On 12/01/2010 12:42 PM, Eric Dumazet wrote: >> Putting IPv4 addresses in sock_common doesn't make it so common anymore :) >> >> Is it possible to make this a union so other address families like IPv6 >> can benefit from this as well, or will that blow the whole cache line >> effect you were trying to achieve? > > This might be OK, depending on cache line size and/or arch. > > On x86_32 for example, that might even be a good thing, because refcnt > might still be in the first 64bytes of socket. > > By the way, ipv6 sock includes inet, so includes ipv4 addresses too, I > only moved them in the 'whole structure' Yes, all that IPv4 address baggage is still there in an IPv6 sock, even if not used. I haven't even looked close enough to see if it is possible to move the IPv6 addresses since I think there are times when both are in-use. -Brian