Re: [PATCH v8 07/46] KVM: Rename memory attribute APIs to prepare for in-place gmem conversion

[Ackerley Tng]

Binbin Wu <binbin.wu@xxxxxxxxxxxxxxx> writes:

>
> [...snip...]
>
>> +static inline bool kvm_mem_range_is_private(struct kvm *kvm, gfn_t start,
>> +					    gfn_t end)
>> +{
>> +	return kvm_range_has_vm_memory_attributes(kvm, start, end,
>> +						  KVM_MEMORY_ATTRIBUTE_PRIVATE,
>> +						  KVM_MEMORY_ATTRIBUTE_PRIVATE);
>>  }
>
> This function is added, but never used in this patch series.
> Is it intended to be called only when CONFIG_KVM_VM_MEMORY_ATTRIBUTES is
> enabled?
>

Thank you for catching this! I think in some earlier revision this was
meant to be used from the guest_memfd populate flow.

I think the version of kvm_gmem_range_is_private in this revision is
good because it is symmetric. If conversion is enabled, call the gmem
range-has-attributes function, and if conversion is disabled, use the VM
range-has-attributes function.

Sean, if no new revision is needed would you be able to drop
kvm_mem_range_is_private() while you're pulling it in?

>>
>> [...snip...]
>>