From mboxrd@z Thu Jan  1 00:00:00 1970
From: Gleb Natapov <gleb@redhat.com>
Subject: Re: KVM: x86: Racy mp_state manipulations
Date: Tue, 5 Mar 2013 09:28:20 +0200
Message-ID: <20130305072820.GU23616@redhat.com>
References: <51337EDD.40303@web.de>
 <5134ABDD.4090908@redhat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Jan Kiszka <jan.kiszka@web.de>,
	Marcelo Tosatti <mtosatti@redhat.com>,
	kvm <kvm@vger.kernel.org>
To: Paolo Bonzini <pbonzini@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:17378 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1752773Ab3CEH2W (ORCPT <rfc822;kvm@vger.kernel.org>);
	Tue, 5 Mar 2013 02:28:22 -0500
Content-Disposition: inline
In-Reply-To: <5134ABDD.4090908@redhat.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Mon, Mar 04, 2013 at 03:12:45PM +0100, Paolo Bonzini wrote:
> The latter makes sense since it's not a fast path, but the only
> transition that is acceptable to KVM_MP_STATE_HALTED is from
> KVM_MP_STATE_RUNNABLE:
> 
> from \ to    RUNNABLE     UNINIT      INIT     HALTED       SIPI
> RUNNABLE       n/a          yes        yes       yes         NO
> UNINIT         NO           n/a        yes       NO          NO
> INIT           NO           yes        n/a       NO          yes
> HALTED         yes          yes        yes       n/a         NO
> SIPI           yes          yes        yes       NO          n/a
> 
About UNINIT column. It should be NO for all of them. vcpu enters UNINIT
state only after creation and never returns to it.

--
			Gleb.