Re: [PATCH 03/15] KVM: x86/mmu: Ensure MMU pages are available when allocating roots

[Sean Christopherson]

On Tue, Mar 02, 2021, Ben Gardon wrote:
> > @@ -3241,16 +3237,10 @@ static int mmu_alloc_direct_roots(struct kvm_vcpu *vcpu)
> >
> >         if (is_tdp_mmu_enabled(vcpu->kvm)) {
> >                 root = kvm_tdp_mmu_get_vcpu_root_hpa(vcpu);
> > -
> > -               if (!VALID_PAGE(root))
> > -                       return -ENOSPC;
> >                 vcpu->arch.mmu->root_hpa = root;
> >         } else if (shadow_root_level >= PT64_ROOT_4LEVEL) {
> >                 root = mmu_alloc_root(vcpu, 0, 0, shadow_root_level,
> >                                       true);
> > -
> > -               if (!VALID_PAGE(root))
> > -                       return -ENOSPC;
> 
> There's so much going on in mmu_alloc_root that removing this check
> makes me nervous, but I think it should be safe.

Just think of it as a variant of kvm_mmu_get_page(), then all your fears will
melt away. ;-)

> I checked though the function because I was worried it might yield
> somewhere in there, which could result in the page cache being emptied
> and the allocation failing, but I don't think mmu_alloc_root this
> function will yield.

Ugh, mmu_alloc_root() won't yield, but get_zeroed_page() used to allocate pae_root
and lm_root on-demand in mmu_alloc_shadow_roots() will.  The two options are
(a) allocate the fake roots before taking the lock and (b) allocate them from
vcpu->arch.mmu_shadow_page_cache.  I probably prefer (a).  (b) will slide
directly into the existing code, but would require bumping the min number of
objects for mmu_shadow_page_cache in mmu_topup_memory_caches().  I'd prefer not
to have yet more code that has to deal with this insanity.