RE: [Xen-devel][PATCH 2/2] Enhance MTRR/PAT virtualization for EPT & VT-d enabled both

["Xin, Xiaohui" <xiaohui.xin@xxxxxxxxx>]

I knew that. But since at last we should add the parameter in ept_set_entry() 
which will then taint p2m_set_entry() to add an unused parameter, and the 
parameter is meaningless at all in shadow mode. And the flag is used in the 
same way as the flag is_in_uc_mode in hvm_set_uc_mode().
Do you like to add a parameter in set-entry() and then ept_set_entry() and 
p2m_set_entry()?

Thanks
Xiaohui

>-----Original Message-----
>From: Keir Fraser [mailto:keir.fraser@xxxxxxxxxxxxx]
>Sent: 2009年1月22日 19:43
>To: Xin, Xiaohui; xen-devel@xxxxxxxxxxxxxxxxxxx
>Subject: Re: [Xen-devel][PATCH 2/2] Enhance MTRR/PAT virtualization for EPT &
>VT-d enabled both
>
>Looking further, it appears to be an ugly way of getting a parameter to an
>p2m-ept.c-internal function. I don't like that. If you need a parameter,
>define it and pass it down properly.
>
> -- Keir
>
>On 22/01/2009 10:15, "Keir Fraser" <keir.fraser@xxxxxxxxxxxxx> wrote:
>
>> You add a new flag 'is_untaint'. What is tainted/untainted? No comment and
>> not a descriptive field name.
>>
>>  -- Keir
>>
>> On 22/01/2009 09:48, "Xin, Xiaohui" <xiaohui.xin@xxxxxxxxx> wrote:
>>
>>> The patch attached is to set effective memory type for EPT according to the
>>> VT-d snoop control capability, and also includes some cleansup for EPT & 
>>> VT-d
>>> both enabled.
>>>
>>> Signed-off-by: Zhai, Edwin Edwin.Zhai@xxxxxxxxx
>>>
>>> Signed-off-by: Xin, Xiaohui xiaohui.xin@xxxxxxxxx
>>>
>>>
>>>
>>> _______________________________________________
>>> 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