From mboxrd@z Thu Jan  1 00:00:00 1970
From: Glauber Costa <glommer@redhat.com>
Subject: Re: [RFC 5/8] change kernel accounting to include steal time
Date: Mon, 30 Aug 2010 14:26:14 -0300
Message-ID: <20100830172614.GM17084@mothafucka.localdomain>
References: <1283184391-7785-2-git-send-email-glommer@redhat.com>
 <1283184391-7785-3-git-send-email-glommer@redhat.com>
 <1283184391-7785-4-git-send-email-glommer@redhat.com>
 <1283184391-7785-5-git-send-email-glommer@redhat.com>
 <1283184391-7785-6-git-send-email-glommer@redhat.com>
 <1283184391-7785-7-git-send-email-glommer@redhat.com>
 <1283184391-7785-8-git-send-email-glommer@redhat.com>
 <1283184391-7785-9-git-send-email-glommer@redhat.com>
 <1283184391-7785-10-git-send-email-glommer@redhat.com>
 <1283186785.1820.1174.camel@laptop>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: kvm@vger.kernel.org, avi@redhat.com, zamsden@redhat.com,
	mtosatti@redhat.com, riel@redhat.com, mingo@elte.hu,
	jeremy@goop.org
To: Peter Zijlstra <peterz@infradead.org>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:57489 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1752833Ab0H3R0c (ORCPT <rfc822;kvm@vger.kernel.org>);
	Mon, 30 Aug 2010 13:26:32 -0400
Content-Disposition: inline
In-Reply-To: <1283186785.1820.1174.camel@laptop>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Mon, Aug 30, 2010 at 06:46:25PM +0200, Peter Zijlstra wrote:
> On Mon, 2010-08-30 at 12:06 -0400, Glauber Costa wrote:
> > This patch proposes a common steal time implementation. When no
> > steal time is accounted, we just add a branch to the current
> > accounting code, that shouldn't add much overhead.
> 
> But why not use alternative code and avoid that conditional
> all-together?
> 
> >   * Account user cpu time to a process.
> >   * @p: the process that the cpu time gets accounted to
> > @@ -3169,6 +3179,12 @@ void account_user_time(struct task_struct *p, cputime_t cputime,
> >  	struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
> >  	cputime64_t tmp;
> >  
> > +	tmp = get_steal_time_from_hypervisor();
> > +	if (tmp) {
> > +		account_steal_time(tmp);
> > +		return;
> > +	}
> > +
> >  	/* Add user time to process. */
> >  	p->utime = cputime_add(p->utime, cputime);
> >  	p->utimescaled = cputime_add(p->utimescaled, cputime_scaled);
> > @@ -3234,6 +3250,12 @@ void account_system_time(struct task_struct *p, int hardirq_offset,
> >  		return;
> >  	}
> >  
> > +	tmp = get_steal_time_from_hypervisor();
> > +	if (tmp) {
> > +		account_steal_time(tmp);
> > +		return;
> > +	}
> > +
> >  	/* Add system time to process. */
> >  	p->stime = cputime_add(p->stime, cputime);
> >  	p->stimescaled = cputime_add(p->stimescaled, cputime_scaled);
> 
> Why replicate that code and not restructure account_process_tick()?
For both: because this way is easier, and this is just an rfc, so I
am mainly interested in getting convergence on the design, or leaving
this behind altogether. I do agree with both your comments, btw