From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 10 Jan 2001 11:20:02 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 10 Jan 2001 11:19:42 -0500 Received: from pcep-jamie.cern.ch ([137.138.38.126]:43535 "EHLO pcep-jamie.cern.ch") by vger.kernel.org with ESMTP id ; Wed, 10 Jan 2001 11:19:41 -0500 Date: Wed, 10 Jan 2001 17:18:28 +0100 From: Jamie Lokier To: Ingo Molnar Cc: Manfred Spraul , linux-kernel@vger.kernel.org Subject: Re: [PLEASE-TESTME] Zerocopy networking patch, 2.4.0-1 Message-ID: <20010110171828.I30055@pcep-jamie.cern.ch> In-Reply-To: <3A5C4FAC.CA6E46A9@colorfullife.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from mingo@elte.hu on Wed, Jan 10, 2001 at 01:07:53PM +0100 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Ingo Molnar wrote: > > > well, this is a performance problem if you are using threads. For normal > > > processes there is no need for a SMP cross-call, there TLB flushes are > > > local only. > > > > > But that would be ugly as hell: > > so apache 2.0 would become slower with MSG_NOCOPY, whereas samba 2.2 > > would become faster. > > there *is* a cost of having a shared VM - and this is i suspect > unavoidable. Is it possible to avoid the SMP cross-call in the case that the other threads have neither accessed nor dirtied the page in question? One way to implement this is to share VMs but not the page tables, or to share parts of the page tables that don't contain writable pages. Just a sudden inspired thought... I don't know if it is possible or worthwhile. enjoy, -- Jamie - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/