On Wed, May 29, 2024 at 2:55 PM Sean Christopherson <seanjc@xxxxxxxxxx> wrote:
>
> On Wed, May 29, 2024, James Houghton wrote:
> > For implementation mmu_notifier_{test,clear}_young, the KVM memslot
> > walker used to take the MMU lock for us. Now make the architectures
> > take it themselves.
>
> Hmm, *forcing* architectures to take mmu_lock is a step backwards. Rather than
> add all of this churn, what about adding CONFIG_KVM_MMU_NOTIFIER_LOCKLESS, e.g.
>
> static __always_inline int kvm_handle_hva_range_no_flush(struct mmu_notifier *mn,
> unsigned long start,
> unsigned long end,
> gfn_handler_t handler)
> {
> struct kvm *kvm = mmu_notifier_to_kvm(mn);
> const struct kvm_mmu_notifier_range range = {
> .start = start,
> .end = end,
> .handler = handler,
> .on_lock = (void *)kvm_null_fn,
> .flush_on_ret = false,
> .may_block = false,
> .lockless = IS_ENABLED(CONFIG_KVM_MMU_NOTIFIER_LOCKLESS),
> };
>
> return __kvm_handle_hva_range(kvm, &range).ret;
> }
Thanks Sean, yes this is a lot better. I will do this for v5.
