From mboxrd@z Thu Jan  1 00:00:00 1970
From: Gleb Natapov <gleb@redhat.com>
Subject: Re: [PATCH v2] KVM: x86: Rework INIT and SIPI handling
Date: Thu, 14 Mar 2013 16:53:10 +0200
Message-ID: <20130314145310.GW11223@redhat.com>
References: <5140662A.7010209@siemens.com>
 <20130314000850.GA4499@amt.cnet>
 <5141A326.3090500@siemens.com>
 <20130314143259.GA4185@amt.cnet>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Jan Kiszka <jan.kiszka@siemens.com>, kvm <kvm@vger.kernel.org>,
	Paolo Bonzini <pbonzini@redhat.com>
To: Marcelo Tosatti <mtosatti@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:9743 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1757315Ab3CNOxP (ORCPT <rfc822;kvm@vger.kernel.org>);
	Thu, 14 Mar 2013 10:53:15 -0400
Content-Disposition: inline
In-Reply-To: <20130314143259.GA4185@amt.cnet>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Thu, Mar 14, 2013 at 11:32:59AM -0300, Marcelo Tosatti wrote:
> > > Also the fact kvm_apic_accept_events() is called from sites is annoying,
> > > why is that necessary again?
> > 
> > - to avoid having pending events in the APIC when delivering mp_state
> >   to user space
> > - to keep mp_state correct after waking up from kvm_vcpu_block (we
> >   could otherwise walk through code paths with the wrong state)
> > - to process pending events when the VCPU was running
> > 
> > Jan
> 
> Ok, its alright (multiple callsites). Perhaps second and third points to
> be collapsed into a single one if moved outside vcpu_enter_guest.

Lets not forget that the point of all this changes was to properly
handle INIT wrt nested guests. I wonder how implementing it will affect
all the call sites. I doubt we want to handle nested vmexit at each one one
them.

--
			Gleb.