* VMX: emulated guest use SYSENTER MSRs
@ 2007-07-06 12:41 Jan Beulich
2007-07-06 12:52 ` Keir Fraser
2007-07-07 7:48 ` Li, Xin B
0 siblings, 2 replies; 6+ messages in thread
From: Jan Beulich @ 2007-07-06 12:41 UTC (permalink / raw)
To: xen-devel
What is the reason for not simply clearing the MSR bitmap bits for these three
MSRs, and rather handling the reads/writes in software?
Jan
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: VMX: emulated guest use SYSENTER MSRs
2007-07-06 12:41 VMX: emulated guest use SYSENTER MSRs Jan Beulich
@ 2007-07-06 12:52 ` Keir Fraser
2007-07-07 7:48 ` Li, Xin B
1 sibling, 0 replies; 6+ messages in thread
From: Keir Fraser @ 2007-07-06 12:52 UTC (permalink / raw)
To: Jan Beulich, xen-devel
Probably because the existing code must remain anyway (since not all VMX
CPUs support the MSR bitmap), and the benefit of avoiding the intercepts on
CPUs that do support it would have entirely negligible performance benefit,
since the MSRs are written only at start of day. We could add calls to avoid
the intercepts anyway though, if you want to provide a patch.
-- Keir
On 6/7/07 13:41, "Jan Beulich" <jbeulich@novell.com> wrote:
> What is the reason for not simply clearing the MSR bitmap bits for these three
> MSRs, and rather handling the reads/writes in software?
>
> Jan
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: VMX: emulated guest use SYSENTER MSRs
2007-07-06 12:41 VMX: emulated guest use SYSENTER MSRs Jan Beulich
2007-07-06 12:52 ` Keir Fraser
@ 2007-07-07 7:48 ` Li, Xin B
2007-07-07 8:52 ` Keir Fraser
1 sibling, 1 reply; 6+ messages in thread
From: Li, Xin B @ 2007-07-07 7:48 UTC (permalink / raw)
To: Jan Beulich, xen-devel
>What is the reason for not simply clearing the MSR bitmap bits
>for these three MSRs, and rather handling the reads/writes in software?
I haven't checked those MSRs, the original concern is we need
save/restore them when context switch, is this needed?
-Xin
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: VMX: emulated guest use SYSENTER MSRs
2007-07-07 7:48 ` Li, Xin B
@ 2007-07-07 8:52 ` Keir Fraser
2007-07-07 8:57 ` Li, Xin B
0 siblings, 1 reply; 6+ messages in thread
From: Keir Fraser @ 2007-07-07 8:52 UTC (permalink / raw)
To: Li, Xin B, Jan Beulich, xen-devel
Since they are included in the VMCS, and read from there on save/restore,
the vmexit appears to be unnecessary.
-- Keir
On 7/7/07 08:48, "Li, Xin B" <xin.b.li@intel.com> wrote:
>
>> What is the reason for not simply clearing the MSR bitmap bits
>> for these three MSRs, and rather handling the reads/writes in software?
>
> I haven't checked those MSRs, the original concern is we need
> save/restore them when context switch, is this needed?
>
> -Xin
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: VMX: emulated guest use SYSENTER MSRs
2007-07-07 8:52 ` Keir Fraser
@ 2007-07-07 8:57 ` Li, Xin B
2007-07-09 9:05 ` Jan Beulich
0 siblings, 1 reply; 6+ messages in thread
From: Li, Xin B @ 2007-07-07 8:57 UTC (permalink / raw)
To: Keir Fraser, Jan Beulich, xen-devel
I will send out the patch soon.
-Xin
>-----Original Message-----
>From: Keir Fraser [mailto:Keir.Fraser@cl.cam.ac.uk]
>Sent: Saturday, July 07, 2007 4:52 PM
>To: Li, Xin B; Jan Beulich; xen-devel@lists.xensource.com
>Subject: Re: [Xen-devel] VMX: emulated guest use SYSENTER MSRs
>
>Since they are included in the VMCS, and read from there on
>save/restore,
>the vmexit appears to be unnecessary.
>
> -- Keir
>
>On 7/7/07 08:48, "Li, Xin B" <xin.b.li@intel.com> wrote:
>
>>
>>> What is the reason for not simply clearing the MSR bitmap bits
>>> for these three MSRs, and rather handling the reads/writes
>in software?
>>
>> I haven't checked those MSRs, the original concern is we need
>> save/restore them when context switch, is this needed?
>>
>> -Xin
>>
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@lists.xensource.com
>> http://lists.xensource.com/xen-devel
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: VMX: emulated guest use SYSENTER MSRs
2007-07-07 8:57 ` Li, Xin B
@ 2007-07-09 9:05 ` Jan Beulich
0 siblings, 0 replies; 6+ messages in thread
From: Jan Beulich @ 2007-07-09 9:05 UTC (permalink / raw)
To: Xin B Li; +Cc: xen-devel
>>> "Li, Xin B" <xin.b.li@intel.com> 07.07.07 10:57 >>>
>I will send out the patch soon.
Not strictly necessary - I have this as part of a larger patch already. But if you
beat me in getting this out, I'll sync up with yours.
Jan
>-----Original Message-----
>From: Keir Fraser [mailto:Keir.Fraser@cl.cam.ac.uk]
>Sent: Saturday, July 07, 2007 4:52 PM
>To: Li, Xin B; Jan Beulich; xen-devel@lists.xensource.com
>Subject: Re: [Xen-devel] VMX: emulated guest use SYSENTER MSRs
>
>Since they are included in the VMCS, and read from there on
>save/restore,
>the vmexit appears to be unnecessary.
>
> -- Keir
>
>On 7/7/07 08:48, "Li, Xin B" <xin.b.li@intel.com> wrote:
>
>>
>>> What is the reason for not simply clearing the MSR bitmap bits
>>> for these three MSRs, and rather handling the reads/writes
>in software?
>>
>> I haven't checked those MSRs, the original concern is we need
>> save/restore them when context switch, is this needed?
>>
>> -Xin
>>
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@lists.xensource.com
>> http://lists.xensource.com/xen-devel
>
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2007-07-09 9:05 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-07-06 12:41 VMX: emulated guest use SYSENTER MSRs Jan Beulich
2007-07-06 12:52 ` Keir Fraser
2007-07-07 7:48 ` Li, Xin B
2007-07-07 8:52 ` Keir Fraser
2007-07-07 8:57 ` Li, Xin B
2007-07-09 9:05 ` Jan Beulich
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.