Re: [PATCH 12/54] KVM: x86/mmu: Drop the intermediate "transient" __kvm_sync_page()

[Paolo Bonzini]

On 22/06/21 19:56, Sean Christopherson wrote:
> @@ -2008,10 +2001,19 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
>   			goto trace_get_page;
>   
>   		if (sp->unsync) {
> -			/* The page is good, but __kvm_sync_page might still end
> -			 * up zapping it.  If so, break in order to rebuild it.
> +			/*
> +			 * The page is good, but is stale.  "Sync" the page to
> +			 * get the latest guest state, but don't write-protect
> +			 * the page and don't mark it synchronized!  KVM needs
> +			 * to ensure the mapping is valid, but doesn't need to
> +			 * fully sync (write-protect) the page until the guest
> +			 * invalidates the TLB mapping.  This allows multiple
> +			 * SPs for a single gfn to be unsync.
> +			 *
> +			 * If the sync fails, the page is zapped.  If so, break
> +			 * If so, break in order to rebuild it.
>   			 */

This should be a separate patch I think.  In addition it should point out the
place where write protection does happen, which is mmu_unsync_children:

                        /*
                         * The page is good, but is stale.  kvm_sync_page does
                         * get the latest guest state, but (unlike mmu_unsync_children)
                         * it doesn't write-protect the page or mark it synchronized!
                         * This way the validity of the mapping is ensured, but the
                         * overhead of write protection is not incurred until the
                         * guest invalidates the TLB mapping.  This allows multiple
                         * SPs for a single gfn to be unsync.
                         *
                         * If the sync fails, the page is zapped.  If so, break
                         * in order to rebuild it.
                         */

Paolo