Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 11 Feb 2022 02:20:17 +0000
From:   Sean Christopherson <seanjc@google.com>
To:     Paolo Bonzini <pbonzini@redhat.com>
Cc:     linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        vkuznets@redhat.com, mlevitsk@redhat.com, dmatlack@google.com
Subject: Re: [PATCH 08/12] KVM: MMU: do not consult levels when freeing roots
Message-ID: <YgXH4eM8inTvDdgT@google.com>
References: <20220209170020.1775368-1-pbonzini@redhat.com>
 <20220209170020.1775368-9-pbonzini@redhat.com>
 <YgWwrG+EQgTwyt8v@google.com>
 <YgWzyBbAZe89ljqO@google.com>
 <ba9e1a56-f769-01c1-607f-3630a62a1b5d@redhat.com>
 <YgW9bqM1M/zJEzqy@google.com>
 <YgW/ZiURGlh5+nUr@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YgW/ZiURGlh5+nUr@google.com>
Precedence: bulk

On Fri, Feb 11, 2022, Sean Christopherson wrote:
> On Fri, Feb 11, 2022, Sean Christopherson wrote:
> > On Fri, Feb 11, 2022, Paolo Bonzini wrote:
> > > On 2/11/22 01:54, Sean Christopherson wrote:
> > > > > > @@ -3242,8 +3245,7 @@ void kvm_mmu_free_roots(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
> > > > > >   					   &invalid_list);
> > > > > >   	if (free_active_root) {
> > > > > > -		if (mmu->shadow_root_level >= PT64_ROOT_4LEVEL &&
> > > > > > -		    (mmu->root_level >= PT64_ROOT_4LEVEL || mmu->direct_map)) {
> > > > > > +		if (to_shadow_page(mmu->root.hpa)) {
> > > > > >   			mmu_free_root_page(kvm, &mmu->root.hpa, &invalid_list);
> > > > > >   		} else if (mmu->pae_root) {
> > > > 
> > > > Gah, this is technically wrong.  It shouldn't truly matter, but it's wrong.  root.hpa
> > > > will not be backed by shadow page if the root is pml4_root or pml5_root, in which
> > > > case freeing the PAE root is wrong.  They should obviously be invalid already, but
> > > > it's a little confusing because KVM wanders down a path that may not be relevant
> > > > to the current mode.
> > > 
> > > pml4_root and pml5_root are dummy, and the first "real" level of page tables
> > > is stored in pae_root for that case too, so I think that should DTRT.
> > 
> > Ugh, completely forgot that detail.  You're correct.

Mostly correct.  The first "real" level will be PML4 in the hCR4.LA57=1, gCR4.LA57=0
nested NPT case.  Ditto for shadowing PAE NPT with 4/5-level NPT, though in that
case KVM still allocates pae_root entries, it just happens to be a "real" level.

And now I realize why I'm so confused, mmu_alloc_shadow_roots() is also broken
with respect to 5-level shadowing 4-level.  I believe the part that got fixed
was 5-level with a 32-bit guest.  Ugh.

For the stuff that actually works in KVM, this will do just fine.  5-level nNPT
can be punted to the future.