From mboxrd@z Thu Jan  1 00:00:00 1970
From: Avi Kivity <avi@redhat.com>
Subject: Re: [PATCH v4] KVM: VMX: Enable XSAVE/XRSTORE for guest
Date: Mon, 24 May 2010 16:40:51 +0300
Message-ID: <4BFA81E3.1020205@redhat.com>
References: <1274695425-14185-1-git-send-email-sheng@linux.intel.com> <4BFA80CC.9070800@redhat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Cc: Marcelo Tosatti <mtosatti@redhat.com>, kvm@vger.kernel.org,
	Dexuan Cui <dexuan.cui@intel.com>
To: Sheng Yang <sheng@linux.intel.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:60714 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1753126Ab0EXNlC (ORCPT <rfc822;kvm@vger.kernel.org>);
	Mon, 24 May 2010 09:41:02 -0400
In-Reply-To: <4BFA80CC.9070800@redhat.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On 05/24/2010 04:36 PM, Avi Kivity wrote:
>
> may be simplified if we move xcr0 reload back to guest entry (... :)  
> but make it lazy:
>
>   save_host_state: nothing
>   set cr4.osxsave: nothing
>   clear cr4.osxsave: nothing
>   guest entry: if (gcr4.osxsave && !guest_xcr0_loaded) { 
> guest_xcr0_loaded = true, load gxcr0 }
>   load_host_state: if (guest_xcr0_loaded) { guest_xcr0_loaded = false; 
> load host xcr0 }
>   fpu switching:  if (guest_xcr0_loaded) { guest_xcr0_loaded = false; 
> load host xcr0 }, do fpu stuff
>
> So we delay xcr0 reload as late as possible for both entry and exit.
>

btw, this is similar to how we switch the fpu itself: 
if-needed-and-not-already-loaded on guest entry, if-loaded on vcpu_put.

-- 
Do not meddle in the internals of kernels, for they are subtle and quick to panic.