Re: [PATCH 2/2] KVM: SEV: Fall back to __vmalloc() for SEV-ES scratch area if necessary

[Paolo Bonzini]

On 11/9/21 23:23, Sean Christopherson wrote:
> The buffer is purely a KVM software construct, i.e. there's
> no need for it to be physically contiguous, and at a max allowed size of
> 16kb it's just large enough that kzalloc() could feasibly fail due to
> memory fragmentation.

That isn't entirely correct, as kzalloc() won't really fail for requests 
up to 32kb, i.e. order <= PAGE_ALLOC_COSTLY_ORDER.  However, the 
reasoning is correct so I queued the patches.

Paolo