Re: [PATCH v3 5/6] arm64/xen: introduce CONFIG_XEN and hypercall.S on ARM64

[Christopher Covington <cov@codeaurora.org>]

Hi Will,

On 06/05/2013 08:50 AM, Will Deacon wrote:
> On Wed, Jun 05, 2013 at 01:44:55PM +0100, Arnd Bergmann wrote:
>> On Wednesday 05 June 2013 13:15:29 Stefano Stabellini wrote:
>>> diff --git a/arch/arm64/Makefile b/arch/arm64/Makefile
>>> index c95c5cb..79dd13d 100644
>>> --- a/arch/arm64/Makefile
>>> +++ b/arch/arm64/Makefile
>>> @@ -37,6 +37,7 @@ TEXT_OFFSET := 0x00080000
>>>  export TEXT_OFFSET GZFLAGS
>>>  
>>>  core-y         += arch/arm64/kernel/ arch/arm64/mm/
>>> +core-$(CONFIG_XEN)             += arch/arm64/xen/
>>>  libs-y         := arch/arm64/lib/ $(libs-y)
>>>  libs-y         += $(LIBGCC)
>>>  
>>> diff --git a/arch/arm64/xen/Makefile b/arch/arm64/xen/Makefile
>>> new file mode 100644
>>> index 0000000..be24040
>>> --- /dev/null
>>> +++ b/arch/arm64/xen/Makefile
>>> @@ -0,0 +1,2 @@
>>> +xen-arm-y      += $(addprefix ../../arm/xen/, enlighten.o grant-table.o)
>>> +obj-y          := xen-arm.o hypercall.o
>>
>> I think it would be nicer to redirect the entire directory, not just
>> the enlighten.o and grant-table.o files. You could do in arch/arm64/Makefile:
>>
>> core-(CONFIG_XEN) += arch/arm/xen/
>>
>> That leaves a small difference in hypercall.o, which I think you can
>> handle with an #ifdef.
>>
>> I believe the reason why KVM does the more elaborate variant is that
>> they want to be able to build their code as a loadable module that
>> also includes code from virt/kvm, which you don't need.
> 
> I thought we scrapped the idea of KVM as a loadable module on ARM, mainly
> due to the complexities with retrospective initialisation of HYP mode/EL2?

What if Hyp/EL2 support were dubbed regular kernel code and the rest of KVM
was made loadable?

Christopher

-- 
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
hosted by the Linux Foundation.