Re: [Xen-devel] [RFC PATCH V2 1/26] VIOMMU: Add vIOMMU helper functions to create, destroy and query capabilities

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 02.06.17 at 09:31, <julien.grall@xxxxxxx> wrote:
> On 06/02/2017 08:10 AM, Lan Tianyu wrote:
>> On 2017年05月30日 23:42, Jan Beulich wrote:
>>>>>> On 30.05.17 at 17:36,<wei.liu2@xxxxxxxxxx>  wrote:
>>>> On Thu, May 18, 2017 at 01:34:31AM -0400, Lan Tianyu wrote:
>>>>> --- a/xen/common/Kconfig
>>>>> +++ b/xen/common/Kconfig
>>>>> @@ -73,6 +73,17 @@ config TMEM
>>>>>   
>>>>>             If unsure, say Y.
>>>>>   
>>>>> +config VIOMMU
>>>>> + def_bool y
>>>>> + prompt "Xen vIOMMU Support" if EXPERT = "y"
>>>>> + depends on X86
>>>>> + ---help---
>>>>> +  Virtual IOMMU provides interrupt remapping function for guest and
>>>>> +  it allows guest to boot up more than 255 vcpus which requires interrupt
>>>>> +  remapping function.
>>>>> +
>>>>> +   If unsure, say Y.
>>>> Indentation. And this should be disabled by default.
>>> It's actually a question whether in our current scheme a Kconfig
>>> option is appropriate here in the first place. I'd rather see this be
>>> an always built feature which needs enabling on the command line
>>> for the time being.
>> 
>>            In the RFC V1, we made vIOMMU always built-in feature. But 
>> ARM or other arches doesn't have vIOMMU support.
>> 
>> Julien suggested to introduce a new Kconfig and only built vIOMMU on 
>> x86. Both two ways won't affect vIOMMU function.
>> 
>> https://www.mail-archive.com/xen-devel@xxxxxxxxxxxxx/msg101421.html 
>> 
>> Jan & Julien, we need to make a choice here
> We should definitely not compiled in code that are not used for an 
> architecture. This would be dead code or potential bug if not disabled 
> correctly.

I agree, but imo this should be a prompt-less Kconfig option,
selected under suitable conditions.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel