WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-devel

Re: [Xen-devel] [PATCH 3/10] Add HVM support

To: "Tian, Kevin" <kevin.tian@xxxxxxxxx>
Subject: Re: [Xen-devel] [PATCH 3/10] Add HVM support
From: Keir Fraser <keir@xxxxxxxxxxxxx>
Date: Tue, 10 Jul 2007 14:52:23 +0100
Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Tue, 10 Jul 2007 06:50:25 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <D470B4E54465E3469E2ABBC5AFAC390F013B1FB6@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: Ace4v8AvEHvNElKqQi6Qqle8QqzaiAKGA7QQAAVN3AAAAXlToAABp7KO
Thread-topic: [Xen-devel] [PATCH 3/10] Add HVM support
User-agent: Microsoft-Entourage/11.3.3.061214


On 10/7/07 14:12, "Tian, Kevin" <kevin.tian@xxxxxxxxx> wrote:

> Seems that I misunderstood your question. Normally such information
> is described in ICH spec and I didn't see VMX related warn by far for
> deep C states. Some deep C state is described with implicit cache flush
> which, I assume, also applies to the internal VMCS cache too. Or else
> all the side effects will be explicitly described such as LAPIC timer stop
> in some deep state. In that case, software needs to tackle that effect
> correspondingly.

Okay.

Anyway, back to your patch 3/10. With a view to cleanly adding VMXOFF on
suspend, and to allow efficient VMCLEARing if we need it in future, e.g.,
for deep-C states, I think you should change the suspend_domain() hook into
suspend_cpu():
 1. This is then symmetric with the resume_cpu() hook.
 2. It's a natural place to put VMXOFF (unlike suspend_domain()).

Of course, the question then is: how do you find the active VMCS's that need
clearing? I suggest you add a list_head to arch_vmx_struct, have a per-cpu
list of active VMCS's, enqueue on vmx_load_vmcs() and dequeue on
__vmx_clear_vmcs().

Could you revise patch 3/10 and resend, please?

 -- Keir


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel