Re: [PATCH] KVM: arm64: Fix memory check in host_stage2_set_owner_locked()

[Mostafa Saleh <smostafa@google.com>]

On Tue, May 06, 2025 at 09:32:22AM +0100, Marc Zyngier wrote:
> On Thu, 01 May 2025 17:24:50 +0100,
> Mostafa Saleh <smostafa@google.com> wrote:
> > 
> > I found this simple bug while preparing some patches for pKVM.
> > AFAICT, it should be harmless (besides crashing the kernel if it
> > was misbehaving)
> > 
> > Fixes: e94a7dea2972 ("KVM: arm64: Move host page ownership tracking to the hyp vmemmap")
> > Signed-off-by: Mostafa Saleh <smostafa@google.com>
> > ---
> >  arch/arm64/kvm/hyp/nvhe/mem_protect.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/arch/arm64/kvm/hyp/nvhe/mem_protect.c b/arch/arm64/kvm/hyp/nvhe/mem_protect.c
> > index 2a5284f749b4..e80f3ebd3e2a 100644
> > --- a/arch/arm64/kvm/hyp/nvhe/mem_protect.c
> > +++ b/arch/arm64/kvm/hyp/nvhe/mem_protect.c
> > @@ -503,7 +503,7 @@ int host_stage2_set_owner_locked(phys_addr_t addr, u64 size, u8 owner_id)
> >  {
> >  	int ret;
> >  
> > -	if (!addr_is_memory(addr))
> > +	if (!range_is_memory(addr, addr + size))
> >  		return -EPERM;
> >  
> >  	ret = host_stage2_try(kvm_pgtable_stage2_set_owner, &host_mmu.pgt,
> 
> I vaguely seem to remember that there was an assumption around
> addr/size representing a single page, and therefore addr_is_memory()
> was doing the right thing.
> 
> Has this assumption changed? Or is this only a figment of my imagination?

From what I see, this might have size exceeding page size in the
following cases:
1- When creating a VM with 4K kernel as the struct pkvm_hyp_vm size
   is at least 0x1130 (without vCPUs)
2- When creating a VM and PGD size is bigger than one page (for cases
   with concatenation as 52 bits and 4K would have 64K PGD)

Thanks,
Mostafa

> 
> Thanks,
> 
> 	M.
> 
> -- 
> Without deviation from the norm, progress is not possible.