From mboxrd@z Thu Jan 1 00:00:00 1970 From: Xiao Guangrong Subject: Re: [PATCH v5 0/3] KVM: perf: kvm events analysis tool Date: Tue, 06 Mar 2012 18:42:21 +0800 Message-ID: <4F55EA0D.4040408@linux.vnet.ibm.com> References: <4F55D110.2020609@linux.vnet.ibm.com> <20120306090730.GB27238@elte.hu> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: Avi Kivity , Arnaldo Carvalho de Melo , Marcelo Tosatti , David Ahern , LKML , KVM To: Ingo Molnar Return-path: In-Reply-To: <20120306090730.GB27238@elte.hu> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org Thanks for your review, Ingo! On 03/06/2012 05:07 PM, Ingo Molnar wrote: > So, your new tool has a similar workflow to: > > perf kvm record > perf kvm report > > but differs from it in terms of events used and in terms of > reported output. > > To me it appears that your tool is basically pretty similar to > 'perf stat', adapted to KVM, right? > > So, could your new tool's workflow be simplified like this: > > perf kvm stat .. > > ? > > To automatically stat all vcpus in the system, the well-known > -a/--all-cpus system-wide method could be used: > > perf kvm stat -a ... > > with stat output following immediately after it has finished. > Actually, the stat information has already been included in the report. > It should also be possible to use those new events in a > recording fashion - a new, rather logical command sub-space > could be used for that: > > perf kvm stat record ... > perf kvm stat report ... > > [ This could be expanded to regular 'perf stat' as well: 'perf > stat record' and 'perf stat report' would be useful - but I > suspect that's outside the scope of your patches. ] > I totally agree with you except i prefer 'perf kvm events' to 'perf kvm stat' :) : it records some specified kvm events and smartly analyze it. I think it matches its doing better.