Re: [PATCH v7 02/24] KVM: arm64: Donate MMIO to the hypervisor

[Mostafa Saleh <smostafa@google.com>]

On Wed, Jul 15, 2026 at 06:26:39PM +0100, Vincent Donnefort wrote:
> On Wed, Jul 15, 2026 at 11:58:43AM +0000, Mostafa Saleh wrote:
> > Add a function to donate MMIO to the hypervisor so IOMMU hypervisor
> > drivers can protect and access the MMIO of IOMMUs.
> > 
> > As donating MMIO is very rare, and we don’t need to encode the full
> > state, it’s reasonable to have a separate function to do this.
> > It will init the host s2 page table with an invalid leaf with the owner ID
> > to prevent the host from mapping the page on faults.
> > 
> > Also, prevent kvm_pgtable_stage2_unmap() from removing owner ID from
> > stage-2 PTEs, as this can be triggered from recycle logic under memory
> > pressure. There is no code relying on this, as all ownership changes is
> > done via kvm_pgtable_stage2_set_owner()
> > 
> > For the error path in IOMMU drivers, add a function to donate MMIO
> > back from hyp to host. However, that leaks the hypervisor virtual
> > address range which should be acceptable as this is quite rare and
> > it matches the behaviour of fix_map/block.
> > 
> > Signed-off-by: Mostafa Saleh <smostafa@google.com>
> > ---
> >  arch/arm64/kvm/hyp/include/nvhe/mem_protect.h |  7 ++
> >  arch/arm64/kvm/hyp/nvhe/mem_protect.c         | 91 ++++++++++++++++++-
> >  arch/arm64/kvm/hyp/pgtable.c                  | 11 +--
> >  3 files changed, 102 insertions(+), 7 deletions(-)
> > 
[...]
> > +	u64 offset;
> > +	int ret;
> > +
> > +	/* Only before de-privilege. */
> > +	if (static_branch_unlikely(&kvm_protected_mode_initialized))
> > +		return -EPERM;
> > +
> > +	if (!PAGE_ALIGNED(addr | size) ||
> 
> You wouldn't need that with u64 pfn, u64 nr_pages :) 

Well that need to be added somewhere :) but I am ok changing the args.

> 
> > +	    !pfn_range_is_valid(hyp_phys_to_pfn(addr), size >> PAGE_SHIFT))
> > +		return -EINVAL;
> > +
> > +	ret = __pkvm_create_private_mapping(addr, size, PAGE_HYP_DEVICE, haddr);
> > +	if (ret)
> > +		return ret;
> > +
> > +	host_lock_component();
> > +	for (offset = 0; offset < size; offset += PAGE_SIZE) {
> > +		if (addr_is_memory(addr + offset)) {
> > +			ret = -EINVAL;
> > +			goto unlock;
> > +		}
> > +		ret = kvm_pgtable_get_leaf(&host_mmu.pgt, addr + offset, &pte, NULL);
> > +		if (ret)
> > +			goto unlock;
> 
> Is that called for big regions? because that looks quite inefficient even for
> something that is init and IIRC we want to use this for GIC hardening too?
> 

It is 128KB, and that's called once at boot (where we walk the whole
addres space anyway) so I don't think it has any impact, and we do
that in other hot paths as __pkvm_guest_share_host().

> Perhaps it wouldn't be complicated to walk hyp_memory[] to make sure this
> doesn't overlap any memory region. And then to reuse check_page_state_range() to
> verify the host stage-2? 

tbh, I think that current implementation is simpler, as we need to
write a special callback for check_page_state_range()

> 
> > +		if (pte && !kvm_pte_valid(pte)) {
> > +			ret = -EPERM;
> > +			goto unlock;
> > +		}
> > +	}
> > +	/*
> > +	 * We set HYP as the owner of the MMIO pages in the host stage-2, for:
> > +	 * - host aborts: host_stage2_adjust_range() would fail for invalid non zero PTEs.
> > +	 * - recycle under memory pressure: host_stage2_unmap_dev_all() would call
> > +	 *   kvm_pgtable_stage2_unmap() which will not clear non zero invalid ptes (counted).
> > +	 * - other MMIO donation: Would fail as we check that the PTE is valid or empty.
> > +	 */
> 
> Not sure we want that level of detail, it will probably become stall very
> quickly. Perhaps we can simply say "the annotation is refcounted and protects
> this region until it is released with __pkvm_hyp_donate_host_mmio()" or
> something like that?
> 
> Also that makes me think, do we want to add this to the ownership_selftest?

Sure, I can look into that.

> 
> > +	ret = host_stage2_try(kvm_pgtable_stage2_annotate, &host_mmu.pgt,
> > +			      addr, size, &host_s2_pool,
> > +			      KVM_HOST_INVALID_PTE_TYPE_DONATION,
> > +			      FIELD_PREP(KVM_HOST_DONATION_PTE_OWNER_MASK, PKVM_ID_HYP));
> > +unlock:
> > +	host_unlock_component();
> > +	return ret;
> > +}
> > +
> > +int __pkvm_hyp_donate_host_mmio(phys_addr_t addr, size_t size)
> > +{
> > +	kvm_pte_t pte;
> > +	u64 offset;
> > +	int ret = 0;
> > +
> > +	if (static_branch_unlikely(&kvm_protected_mode_initialized))
> > +		return -EPERM;
> > +
> > +	if (!PAGE_ALIGNED(addr | size) ||
> > +	    !pfn_range_is_valid(hyp_phys_to_pfn(addr), size >> PAGE_SHIFT))
> > +		return -EINVAL;
> > +
> > +	host_lock_component();
> > +	for (offset = 0; offset < size; offset += PAGE_SIZE) {
> > +		if (addr_is_memory(addr + offset)) {
> > +			ret = -EINVAL;
> > +			goto unlock;
> > +		}
> > +		ret = kvm_pgtable_get_leaf(&host_mmu.pgt, addr + offset, &pte, NULL);
> > +		if (ret)
> > +			goto unlock;
> > +		if (!pte || kvm_pte_valid(pte)) {
> > +			ret = -EINVAL;
> > +			goto unlock;
> > +		}
> > +		if (FIELD_GET(KVM_HOST_DONATION_PTE_OWNER_MASK, pte) != PKVM_ID_HYP) {
> > +			ret = -EPERM;
> > +			goto unlock;
> > +		}
> > +	}
> > +	WARN_ON(host_stage2_idmap_locked(addr, size, PKVM_HOST_MMIO_PROT));
> > +unlock:
> > +	host_unlock_component();
> > +	return ret;
> > +}
> > +
> >  int __pkvm_hyp_donate_host(u64 pfn, u64 nr_pages)
> >  {
> >  	u64 phys = hyp_pfn_to_phys(pfn);
> > diff --git a/arch/arm64/kvm/hyp/pgtable.c b/arch/arm64/kvm/hyp/pgtable.c
> > index 91a7dfad6686..3073184cf6ad 100644
> > --- a/arch/arm64/kvm/hyp/pgtable.c
> > +++ b/arch/arm64/kvm/hyp/pgtable.c
> > @@ -1161,13 +1161,12 @@ static int stage2_unmap_walker(const struct kvm_pgtable_visit_ctx *ctx,
> >  	kvm_pte_t *childp = NULL;
> >  	bool need_flush = false;
> >  
> > -	if (!kvm_pte_valid(ctx->old)) {
> > -		if (stage2_pte_is_counted(ctx->old)) {
> > -			kvm_clear_pte(ctx->ptep);
> > -			mm_ops->put_page(ctx->ptep);
> > -		}
> > +	/*
> > +	 * That also ignores stage2_pte_is_counted() instead of clearing
> > +	 * the PTE as the MMIO can be owned by the hypervisor.
> 
> Not sure we should reference something pKVM specific in that code. Perhaps just
> say we don't touch refcounted PTEs?

Sure, I can simplify it.

Thanks,
Mostafa