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, 2 Apr 2003 00:06:42 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 2 Apr 2003 00:06:42 -0500 Received: from 210-86-35-88.dialup.xtra.co.nz ([210.86.35.88]:260 "EHLO riven.neverborn.ORG") by vger.kernel.org with ESMTP id ; Wed, 2 Apr 2003 00:06:40 -0500 Date: Wed, 2 Apr 2003 17:17:55 +1200 From: "leon j. breedt" To: linux-kernel@vger.kernel.org Subject: ASUS P4PE spurious hdparm -tT results...UPDATE Message-ID: <20030402051755.GA3281@riven.neverborn.ORG> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.4i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org hi, i'm still getting the same weird negative test results for the WDC WD800JB-00CRA1 drive on 2.4.21-pre6: /dev/hda: Timing buffer-cache reads: -3008 MB in 0.00 seconds = -inf kB/sec Timing buffered disk reads: -1504 MB in 0.00 seconds = -inf kB/sec Hmm.. suspicious results: probably not enough free memory for a proper test. however...i noticed that when i booted up with kernel 2.4.20-xfs_pre6 (the Gentoo 1.4rc3 LiveCD kernel), i didn't get strange test results. the difference between the two seemed to be that the 2.4.20-xfs_pre6 kernel initialized the drive in mdma2 mode, and didn't support udma5 for the IDE layer in that version of the kernel. so maybe this is a bug that's crept in since 2.4.21 and better ICH4 support days? the IDE layer in that version of the kernel is based on a version from www.linuxdiskcert.org, i'm not sure how different that is from the version Alan & c have been working on. any ideas what may cause this, for just the WDC drive and not the IBM? both are on the same channel.. thanks, leon.