From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261725AbTILOtB (ORCPT ); Fri, 12 Sep 2003 10:49:01 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261724AbTILOtB (ORCPT ); Fri, 12 Sep 2003 10:49:01 -0400 Received: from pc1-cwma1-5-cust4.swan.cable.ntl.com ([80.5.120.4]:51351 "EHLO dhcp23.swansea.linux.org.uk") by vger.kernel.org with ESMTP id S261723AbTILOs7 (ORCPT ); Fri, 12 Sep 2003 10:48:59 -0400 Subject: Re: [PATCH] NCR53c406a.c warning From: Alan Cox To: "Maciej W. Rozycki" Cc: James Bottomley , Geert Uytterhoeven , SCSI Mailing List , Linux Kernel Development In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1063378054.5785.1.camel@dhcp23.swansea.linux.org.uk> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.4 (1.4.4-5) Date: Fri, 12 Sep 2003 15:47:35 +0100 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Gwe, 2003-09-12 at 15:16, Maciej W. Rozycki wrote: > I've encountered an ISA adapter using this chip in polled mode (no ISA > IRQ line routed to the chip) quite recently. But I can't say if the guy > using it won't throw it away before final 2.6. ;-) For a lot of these dumb controllers polled I/O seems to materially outperform interrupt driven I/O on SMP boxes - its true of the NCR5380 for example. We trade a ton of locking jamming the box up for a CPU thats spending some of its time pretending to be half a scsi chip.