From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756012Ab1C3PYu (ORCPT ); Wed, 30 Mar 2011 11:24:50 -0400 Received: from smtpbg52.qq.com ([64.71.138.43]:40172 "HELO smtpbg52.qq.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1753728Ab1C3PYt (ORCPT ); Wed, 30 Mar 2011 11:24:49 -0400 X-QQ-mid: esmtp1t1301498606t619t04275 X-QQ-SSF: 00000000000000F0FG400F000000000 Date: Wed, 30 Mar 2011 23:23:30 +0800 From: "Lina Lu" To: "Vivek Goyal" Cc: "linux kernel mailing list" Subject: cfq-iosched.c:Use cfqq->nr_sectors in charge the vdisktime Message-ID: <201103302323256090890@foxmail.com> X-mailer: Foxmail 6, 15, 201, 23 [cn] Mime-Version: 1.0 Content-Type: text/plain; charset="gb2312" Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Vivek, I find the weight policy can be more accuracy with cfqq->nr_sectors instead of cfqq->slice_dispatch. Today, I try to modify cfq_group_served(), and use "charge = cfqq->nr_sectors; " instead of "charge = cfqq->slice_dispatch; " . The test result seens more accuracy. Why you choose slice_dispatch here? Is the nr_sectors will lower the total performance? And in iops mod, if I try to apply weight policy on two IO processes with different avgrq-sz, the test results will not exact match the weight value. Thanks Lina