Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261373AbTIYQNy (ORCPT ); Thu, 25 Sep 2003 12:13:54 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261356AbTIYQNy (ORCPT ); Thu, 25 Sep 2003 12:13:54 -0400 Received: from jurassic.park.msu.ru ([195.208.223.243]:62986 "EHLO jurassic.park.msu.ru") by vger.kernel.org with ESMTP id S261328AbTIYQNv (ORCPT ); Thu, 25 Sep 2003 12:13:51 -0400 Date: Thu, 25 Sep 2003 20:11:27 +0400 From: Ivan Kokshaysky To: "David S. Miller" Cc: davidm@hpl.hp.com, davidm@napali.hpl.hp.com, kevin.vanmaren@unisys.com, peter@chubb.wattle.id.au, bcrl@kvack.org, ak@suse.de, iod00d@hp.com, peterc@gelato.unsw.edu.au, linux-ns83820@kvack.org, linux-ia64@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: NS83820 2.6.0-test5 driver seems unstable on IA64 Message-ID: <20030925201127.A11889@jurassic.park.msu.ru> References: <20030923105712.552dbb1e.davem@redhat.com> <16240.36993.148535.613568@napali.hpl.hp.com> <20030923114744.137d5dac.davem@redhat.com> <16240.40001.632466.644215@napali.hpl.hp.com> <20030923121044.483d3a5c.davem@redhat.com> <16240.42563.834328.584444@napali.hpl.hp.com> <20030923200832.1a58a215.davem@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20030923200832.1a58a215.davem@redhat.com>; from davem@redhat.com on Tue, Sep 23, 2003 at 08:08:32PM -0700 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1266 Lines: 25 On Tue, Sep 23, 2003 at 08:08:32PM -0700, David S. Miller wrote: > It's not even decided that the copying version is faster. On Alpha it is - I did some experimentation yesterday with the tulip driver and "rx_copybreak=0" vs default "rx_copybreak=1518". That's true, copying the large packet is very expensive due to constant cache misses. On UP1500 it costs more than 2500 cycles per 1.5Kb copy on large ftp transfers. OTOH, minimal time to handle unaligned load is ~200 cycles (ev6 cpu), but only when the trap handler itself and its data are in L1 cache (like doing unaligned load in a tight loop). In the real life average time is ~400 cycles due to cache misses. Further, with "rx_copybreak=0" I've got 20 (!) unaligned traps per TCP packet, i.e. penalty was 8000 vs 2500 cycles... Note that my network setup is as simple as possible - no routing/filtering etc., just a networked workstation. I guess one with a more complex routing/firewalling configuration would have a LOT more unaligned traps per packet. Ivan. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/