Re: [PATCH 4/8] KVM: gmem: protect kvm_mmu_invalidate_end()

[Sean Christopherson]

On Mon, Aug 21, 2023, Ashish Kalra wrote:
> Hello Mingwei & Sean,
> 
> On 8/18/2023 9:08 PM, Mingwei Zhang wrote:
> The maximum hits are seen with shmem_fallocate and madvise, which we believe
> are response to shared<->private
> GHCB page-state-chage requests. discard=both handles discard both for
> private and shared memory, so freeing shared memory
> via fallocate(shared_memfd, FALLOC_FL_PUNCH_HOLE, ...) would trigger the
> notifiers when freeing shared pages after guest converts a GPA to
> private.
> 
> Now, as with SNP+guest_memfd, guest private memory is not mapped in host
> anymore, so i added a generic fix (instead of Sean's proposed patch of
> checking for SNP guest inside sev_guest_memory_reclaimed()):
> 
> --- a/virt/kvm/kvm_main.c
> +++ b/virt/kvm/kvm_main.c
> @@ -593,6 +593,9 @@ static __always_inline int __kvm_handle_hva_range(struct
> kvm *kvm,
>                         unsigned long hva_start, hva_end;
> 
>                         slot = container_of(node, struct kvm_memory_slot,
> hva_node[slots->node_idx]);
> +                       if (kvm_slot_can_be_private(slot)) {
> +                               continue;
> +                       }
>                         hva_start = max(range->start, slot->userspace_addr);
>                         hva_end = min(range->end, slot->userspace_addr +
>                                                   (slot->npages <<
> PAGE_SHIFT));

...

> As expected, the SEV hook is not invoked for the guest private memory pages
> (no more invalidation from shmem_fallocate() + madvise()).
> 
> Isn't it better to skip invoking the KVM MMU invalidation notifier when the
> invalidated range belongs to guest private memory ?

Oooh, you're running into problems where KVM blasts both the private and shared
mappings even though invalidations from the mmu_notifier are shared-only by
definition.

The answer is "yes", but simply skipping slots that _can_ be private is wrong,
as KVM still needs to zap any shared mappings.  I have a plan[*], but I completely
spaced on incorporating the idea into the gmem RFC.  I'll add that to the "list
of todos for merging gmem", which I need to get sent out asap.

https://lore.kernel.org/all/ZJX0hk+KpQP0KUyB@xxxxxxxxxx

> > In fact, AFAIC, SNP VM does not track whether each page is previously
> > shared, isn't it? If a page was previously shared and was written by the
> > host kernel or devices before it was changed to private. No one tracks it
> > and dirty caches are there!
> 
> The skipped invalidation here covered the case Mingwei mentioned above,
> where the pages are changed from private->shared and subsequent freeing of
> shared pages triggered the invalidation.
> 
> But, then why are we concerned about this, i thought we have concerns about
> the case where the dirty cache lines contain encrypted guest data ?

Yes, that's my understanding as well (assuming by "this" you mean the case where
the CPU cache has dirty lines for _shared_ addresses).