From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1757802Ab2FUGoJ (ORCPT <rfc822;w@1wt.eu>);
	Thu, 21 Jun 2012 02:44:09 -0400
Received: from mx1.redhat.com ([209.132.183.28]:40493 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1754643Ab2FUGoG (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Thu, 21 Jun 2012 02:44:06 -0400
Date: Thu, 21 Jun 2012 09:43:58 +0300
From: Gleb Natapov <gleb@redhat.com>
To: Rik van Riel <riel@redhat.com>
Cc: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>,
        Avi Kivity <avi@redhat.com>, Marcelo Tosatti <mtosatti@redhat.com>,
        Srikar <srikar@linux.vnet.ibm.com>,
        Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com>,
        Peter Zijlstra <peterz@infradead.org>,
        "Nikunj A. Dadhania" <nikunj@linux.vnet.ibm.com>,
        KVM <kvm@vger.kernel.org>, Ingo Molnar <mingo@redhat.com>,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] kvm: handle last_boosted_vcpu = 0 case
Message-ID: <20120621064358.GS6533@redhat.com>
References: <20120619202047.26191.40429.sendpatchset@codeblue>
 <20120619165104.2a4574f8@annuminas.surriel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20120619165104.2a4574f8@annuminas.surriel.com>
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Tue, Jun 19, 2012 at 04:51:04PM -0400, Rik van Riel wrote:
> On Wed, 20 Jun 2012 01:50:50 +0530
> Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com> wrote:
> 
> > 
> > In ple handler code, last_boosted_vcpu (lbv) variable is
> > serving as reference point to start when we enter.
> 
> > Also statistical analysis (below) is showing lbv is not very well
> > distributed with current approach.
> 
> You are the second person to spot this bug today (yes, today).
> 
> Due to time zones, the first person has not had a chance yet to
> test the patch below, which might fix the issue...
> 
> Please let me know how it goes.
> 
> ====8<====
> 
> If last_boosted_vcpu == 0, then we fall through all test cases and
> may end up with all VCPUs pouncing on vcpu 0.  With a large enough
> guest, this can result in enormous runqueue lock contention, which
> can prevent vcpu0 from running, leading to a livelock.
> 
> Changing < to <= makes sure we properly handle that case.
> 
> Signed-off-by: Rik van Riel <riel@redhat.com>
> ---
>  virt/kvm/kvm_main.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> index 7e14068..1da542b 100644
> --- a/virt/kvm/kvm_main.c
> +++ b/virt/kvm/kvm_main.c
> @@ -1586,7 +1586,7 @@ void kvm_vcpu_on_spin(struct kvm_vcpu *me)
>  	 */
>  	for (pass = 0; pass < 2 && !yielded; pass++) {
>  		kvm_for_each_vcpu(i, vcpu, kvm) {
> -			if (!pass && i < last_boosted_vcpu) {
> +			if (!pass && i <= last_boosted_vcpu) {
>  				i = last_boosted_vcpu;
>  				continue;
>  			} else if (pass && i > last_boosted_vcpu)
> 
Looks correct. We can simplify this by introducing something like:

#define kvm_for_each_vcpu_from(idx, n, vcpup, kvm) \
        for (n = atomic_read(&kvm->online_vcpus); \
             n && (vcpup = kvm_get_vcpu(kvm, idx)) != NULL; \
             n--, idx = (idx+1) % atomic_read(&kvm->online_vcpus))

--
			Gleb.