Re: [PATCH v5 3/4] KVM: arm64/mmu: use gfn_to_pfn_page

[Sean Christopherson]

On Mon, Nov 29, 2021, David Stevens wrote:
> @@ -1142,14 +1146,17 @@ static int user_mem_abort(struct kvm_vcpu *vcpu, phys_addr_t fault_ipa,
>  
>  	/* Mark the page dirty only if the fault is handled successfully */
>  	if (writable && !ret) {
> -		kvm_set_pfn_dirty(pfn);
> +		if (page)
> +			kvm_set_pfn_dirty(pfn);

If kvm_set_page_dirty() is changed to be less dumb:

		if (page)
			kvm_set_page_dirty(page);

>  		mark_page_dirty_in_slot(kvm, memslot, gfn);
>  	}
>  
>  out_unlock:
>  	spin_unlock(&kvm->mmu_lock);
> -	kvm_set_pfn_accessed(pfn);
> -	kvm_release_pfn_clean(pfn);
> +	if (page) {
> +		kvm_set_pfn_accessed(pfn);
> +		put_page(page);

Oof, KVM's helpers are stupid.  Take a page, convert it to a pfn, then convert it
back to a page, just to mark it dirty or put a ref.  Can you fold the below 
(completely untested) patch in before the x86/arm64 patches?  That way this code
can be:

	if (page)
		kvm_release_page_accessed(page);

and x86 can do:

	if (fault->page)
		kvm_release_page_clean(page);

instead of open-coding put_page().