Invalidate on fd-based memslot can reuse the code from existing MMU
notifier.

Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com>
Signed-off-by: Chao Peng <chao.p.peng@linux.intel.com>
---
 include/linux/kvm_host.h |  3 +++
 virt/kvm/kvm_main.c      | 35 +++++++++++++++++++++++++++++++++++
 2 files changed, 38 insertions(+)

diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 925c4d9f0a31..f0fd32f6eab3 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -1883,4 +1883,7 @@ static inline void kvm_handle_signal_exit(struct kvm_vcpu *vcpu)
 /* Max number of entries allowed for each kvm dirty ring */
 #define  KVM_DIRTY_RING_MAX_ENTRIES  65536
 
+int kvm_memfd_invalidate_range(struct kvm *kvm, struct inode *inode,
+			       unsigned long start, unsigned long end);
+
 #endif
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index d9a6890dd18a..090afbadb03f 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -811,6 +811,35 @@ static int kvm_init_mmu_notifier(struct kvm *kvm)
 	return mmu_notifier_register(&kvm->mmu_notifier, current->mm);
 }
 
+int kvm_memfd_invalidate_range(struct kvm *kvm, struct inode *inode,
+			       unsigned long start, unsigned long end)
+{
+	int ret;
+	const struct kvm_useraddr_range useraddr_range = {
+		.start		= start,
+		.end		= end,
+		.pte		= __pte(0),
+		.handler	= kvm_unmap_gfn_range,
+		.on_lock	= (void *)kvm_null_fn,
+		.flush_on_ret	= true,
+		.may_block	= false,
+	};
+
+
+	/* Prevent memslot modification */
+	spin_lock(&kvm->mn_invalidate_lock);
+	kvm->mn_active_invalidate_count++;
+	spin_unlock(&kvm->mn_invalidate_lock);
+
+	ret = __kvm_handle_useraddr_range(kvm, &useraddr_range);
+
+	spin_lock(&kvm->mn_invalidate_lock);
+	kvm->mn_active_invalidate_count--;
+	spin_unlock(&kvm->mn_invalidate_lock);
+
+	return ret;
+}
+
 #else  /* !(CONFIG_MMU_NOTIFIER && KVM_ARCH_WANT_MMU_NOTIFIER) */
 
 static int kvm_init_mmu_notifier(struct kvm *kvm)
@@ -818,6 +847,12 @@ static int kvm_init_mmu_notifier(struct kvm *kvm)
 	return 0;
 }
 
+int kvm_memfd_invalidate_range(struct kvm *kvm, struct inode *inode,
+			       unsigned long start, unsigned long end)
+{
+	return 0;
+}
+
 #endif /* CONFIG_MMU_NOTIFIER && KVM_ARCH_WANT_MMU_NOTIFIER */
 
 #ifdef CONFIG_HAVE_KVM_PM_NOTIFIER
-- 
2.17.1

On 11/19/21 14:47, Chao Peng wrote:
> +
> +	/* Prevent memslot modification */
> +	spin_lock(&kvm->mn_invalidate_lock);
> +	kvm->mn_active_invalidate_count++;
> +	spin_unlock(&kvm->mn_invalidate_lock);
> +
> +	ret = __kvm_handle_useraddr_range(kvm, &useraddr_range);
> +
> +	spin_lock(&kvm->mn_invalidate_lock);
> +	kvm->mn_active_invalidate_count--;
> +	spin_unlock(&kvm->mn_invalidate_lock);
> +


You need to follow this with a rcuwait_wake_up as in 
kvm_mmu_notifier_invalidate_range_end.

It's probably best if you move the manipulations of 
mn_active_invalidate_count from kvm_mmu_notifier_invalidate_range_* to 
two separate functions.

Paolo

On Tue, Nov 23, 2021 at 09:46:34AM +0100, Paolo Bonzini wrote:
> On 11/19/21 14:47, Chao Peng wrote:
> > +
> > +	/* Prevent memslot modification */
> > +	spin_lock(&kvm->mn_invalidate_lock);
> > +	kvm->mn_active_invalidate_count++;
> > +	spin_unlock(&kvm->mn_invalidate_lock);
> > +
> > +	ret = __kvm_handle_useraddr_range(kvm, &useraddr_range);
> > +
> > +	spin_lock(&kvm->mn_invalidate_lock);
> > +	kvm->mn_active_invalidate_count--;
> > +	spin_unlock(&kvm->mn_invalidate_lock);
> > +
> 
> 
> You need to follow this with a rcuwait_wake_up as in
> kvm_mmu_notifier_invalidate_range_end.

Oh right.

> 
> It's probably best if you move the manipulations of
> mn_active_invalidate_count from kvm_mmu_notifier_invalidate_range_* to two
> separate functions.

Will do.

> 
> Paolo