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] vmx: fix debugctl handling

To: "Jan Beulich" <jbeulich@xxxxxxxxxx>, "Haitao Shan" <maillists.shan@xxxxxxxxx>
Subject: RE: [Xen-devel] [PATCH] vmx: fix debugctl handling
From: "Shan, Haitao" <haitao.shan@xxxxxxxxx>
Date: Tue, 18 Mar 2008 10:29:03 +0800
Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Mon, 17 Mar 2008 19:41:06 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <47DE9B9E.76E4.0078.0@xxxxxxxxxx>
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>
References: <47DA8BD4.76E4.0078.0@xxxxxxxxxx><823A93EED437D048963A3697DB0E35DE0131BCC6@xxxxxxxxxxxxxxxxxxxxxxxxxxxx><47DE3DF1.76E4.0078.0@xxxxxxxxxx><481ad8630803170638g7a4fb317t21417e0a9bdb1d1f@xxxxxxxxxxxxxx> <47DE9B9E.76E4.0078.0@xxxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: AciIRDNJiCyhzrQOTnOaLdscBrismQAWccWw
Thread-topic: [Xen-devel] [PATCH] vmx: fix debugctl handling
Oh, yes. Sorry for my bad memory. The spec does say that there is no
branch trace activities during VMExit/VMEntry. It should work.
But anyhow, I think it is better to try and see whether it works.:)
For PEBS, this actually can happen during VMExit/VMEntry. I have been
debugging this for a long time when I made the patch to enable PMU usage
in HVM guests. If you programme the PMCs and enable PEBS and DS in
guest, while in host these features are disabled, events happened during
VMExit/VMEntry will be lost. And it is hard to emulate the lost one in
VMM. So I disabled these features through IA32_MISC_ENABLE. You can see
in current Xen's code.

Best Regards
Haitao Shan

Jan Beulich wrote:
>>>> "Haitao Shan" <maillists.shan@xxxxxxxxx> 17.03.08 14:38 >>>
>> Yes, I agree. Once allowing guest to modify this MSR, you have to
>> implement some save/load functions. I am wondering what you are
>> trying to enable? As far as I know, DS area 
> 
> Yes, I'm considering adding support for DS, since Linux 2.6.25-rc
> briefly had support for this (it's now disabled again), so it's
> likely that pretty soon the functionality in the kernel will be
> there. Hence the hypervisor ought to support it.
> 
>> (both PEBS and BTS) is not working properly now. You should assume
>> that any events which will trigger DS write might happen right at
>> the moment that CPU is in VMExit/VMEntry. Only using HW to switch
>> related MSRs is not enough. 
> 
> That would be very bad - it would mean that you cannot reliably
> virtualize this. But - are you sure here? For PEBS, I could see this
> to be true (although it should be properly dealt with in hardware),
> but 
> BTS shouldn't have any activity during vmexit/vmentry as there are
> no branches during that time.
> 
> Jan
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel

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

<Prev in Thread] Current Thread [Next in Thread>