Re: [PATCH v3 3/3] KVM: x86/mmu: Defer allocation of shadow MMU's hashed page list

[Sean Christopherson <seanjc@google.com>]

On Sat, May 17, 2025, Paolo Bonzini wrote:
> On 5/16/25 23:54, Sean Christopherson wrote:
> > +	/*
> > +	 * Write mmu_page_hash exactly once as there may be concurrent readers,
> > +	 * e.g. to check for shadowed PTEs in mmu_try_to_unsync_pages().  Note,
> > +	 * mmu_lock must be held for write to add (or remove) shadow pages, and
> > +	 * so readers are guaranteed to see an empty list for their current
> > +	 * mmu_lock critical section.
> > +	 */
> > +	WRITE_ONCE(kvm->arch.mmu_page_hash, h);
> 
> Use smp_store_release here (unlike READ_ONCE(), it's technically incorrect
> to use WRITE_ONCE() here!),

Can you elaborate why?  Due to my x86-centric life, my memory ordering knowledge
is woefully inadequate.

> with a remark that it pairs with kvm_get_mmu_page_hash().  That's both more
> accurate and leads to a better comment than "write exactly once".