From mboxrd@z Thu Jan 1 00:00:00 1970 From: Martin Peschke Subject: Re: [Patch] SCSI statistics Date: Tue, 19 Sep 2006 18:48:30 +0200 Message-ID: <45101F5E.2070909@de.ibm.com> References: <1158578439.3060.50.camel@dyn-9-152-230-71.boeblingen.de.ibm.com> <1158614913.3182.78.camel@laptopd505.fenrus.org> <20060918213424.GS2585@parisc-linux.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mtagate2.uk.ibm.com ([195.212.29.135]:48357 "EHLO mtagate2.uk.ibm.com") by vger.kernel.org with ESMTP id S1030253AbWISQsm (ORCPT ); Tue, 19 Sep 2006 12:48:42 -0400 Received: from d06nrmr1407.portsmouth.uk.ibm.com (d06nrmr1407.portsmouth.uk.ibm.com [9.149.38.185]) by mtagate2.uk.ibm.com (8.13.8/8.13.8) with ESMTP id k8JGmfdT166010 for ; Tue, 19 Sep 2006 16:48:41 GMT Received: from d06av04.portsmouth.uk.ibm.com (d06av04.portsmouth.uk.ibm.com [9.149.37.216]) by d06nrmr1407.portsmouth.uk.ibm.com (8.13.6/8.13.6/NCO v8.1.1) with ESMTP id k8JGoshc1900550 for ; Tue, 19 Sep 2006 17:50:54 +0100 Received: from d06av04.portsmouth.uk.ibm.com (loopback [127.0.0.1]) by d06av04.portsmouth.uk.ibm.com (8.12.11.20060308/8.13.3) with ESMTP id k8JGmeJh002937 for ; Tue, 19 Sep 2006 17:48:40 +0100 In-Reply-To: <20060918213424.GS2585@parisc-linux.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Matthew Wilcox Cc: Arjan van de Ven , linux-scsi@vger.kernel.org Matthew Wilcox wrote: > On Mon, Sep 18, 2006 at 11:28:32PM +0200, Arjan van de Ven wrote: >> wouldn't this be nicer as a block level functionality? > > We already have block layer statistics. you mean disk_stats_read() etc. in genhd? > What we don't have are stats > that measure the other ways to send commands to scsi devices like SG and > retries that the scsi system does without notifying the block layer. I > think both are useful (although at first glance they seem redundant). There are no statistics that allow to retrieve data similar to my patch, with the exception of blktrace. But I think blktrace is difficult to use as sole performance analysis tool on a customer's system with high load and applications competing with blktrace postprocessing tools.