KVM: Add barriers to allow mmu_notifier_retry to be used locklessly

This adds an smp_wmb in kvm_mmu_notifier_invalidate_range_end() and an smp_rmb in mmu_notifier_retry() so that mmu_notifier_retry() will give the correct answer when called without kvm->mmu_lock being held. PowerPC Book3S HV KVM wants to use a bitlock per guest page rather than a single global spinlock in order to improve the scalability of updates to the guest MMU hashed page table, and so needs this. Signed-off-by: Paul Mackerras <paulus@samba.org> Acked-by: Avi Kivity <avi@redhat.com> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>

KVM: Add barriers to allow mmu_notifier_retry to be used locklessly
This adds an smp_wmb in kvm_mmu_notifier_invalidate_range_end() and an smp_rmb in mmu_notifier_retry() so that mmu_notifier_retry() will give the correct answer when called without kvm->mmu_lock being held. PowerPC Book3S HV KVM wants to use a bitlock per guest page rather than a single global spinlock in order to improve the scalability of updates to the guest MMU hashed page table, and so needs this. Signed-off-by: Paul Mackerras <paulus@samba.org> Acked-by: Avi Kivity <avi@redhat.com> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Paul Mackerras · Avi Kivity
1 parent 342d3db763
Showing 2 changed files with 12 additions and 8 deletions Side-by-side Diff
include/linux/kvm_host.h
virt/kvm/kvm_main.c
@@ -702,12 +702,16 @@
 	if (unlikely(vcpu->kvm->mmu_notifier_count))
 		return 1;
 	/*
-	 * Both reads happen under the mmu_lock and both values are
-	 * modified under mmu_lock, so there's no need of smb_rmb()
-	 * here in between, otherwise mmu_notifier_count should be
-	 * read before mmu_notifier_seq, see
-	 * mmu_notifier_invalidate_range_end write side.
+	 * Ensure the read of mmu_notifier_count happens before the read
+	 * of mmu_notifier_seq.  This interacts with the smp_wmb() in
+	 * mmu_notifier_invalidate_range_end to make sure that the caller
+	 * either sees the old (non-zero) value of mmu_notifier_count or
+	 * the new (incremented) value of mmu_notifier_seq.
+	 * PowerPC Book3s HV KVM calls this under a per-page lock
+	 * rather than under kvm->mmu_lock, for scalability, so
+	 * can't rely on kvm->mmu_lock to keep things ordered.
 	 */
+	smp_rmb();
 	if (vcpu->kvm->mmu_notifier_seq != mmu_seq)
 		return 1;
 	return 0;
@@ -357,11 +357,11 @@
 	 * been freed.
 	 */
 	kvm->mmu_notifier_seq++;
+	smp_wmb();
 	/*
 	 * The above sequence increase must be visible before the
-	 * below count decrease but both values are read by the kvm
-	 * page fault under mmu_lock spinlock so we don't need to add
-	 * a smb_wmb() here in between the two.
+	 * below count decrease, which is ensured by the smp_wmb above
+	 * in conjunction with the smp_rmb in mmu_notifier_retry().
 	 */
 	kvm->mmu_notifier_count--;
 	spin_unlock(&kvm->mmu_lock);
...	...	@@ -702,12 +702,16 @@
702	702	if (unlikely(vcpu->kvm->mmu_notifier_count))
703	703	return 1;
704	704	/*
705		- * Both reads happen under the mmu_lock and both values are
706		- * modified under mmu_lock, so there's no need of smb_rmb()
707		- * here in between, otherwise mmu_notifier_count should be
708		- * read before mmu_notifier_seq, see
709		- * mmu_notifier_invalidate_range_end write side.
	705	+ * Ensure the read of mmu_notifier_count happens before the read
	706	+ * of mmu_notifier_seq. This interacts with the smp_wmb() in
	707	+ * mmu_notifier_invalidate_range_end to make sure that the caller
	708	+ * either sees the old (non-zero) value of mmu_notifier_count or
	709	+ * the new (incremented) value of mmu_notifier_seq.
	710	+ * PowerPC Book3s HV KVM calls this under a per-page lock
	711	+ * rather than under kvm->mmu_lock, for scalability, so
	712	+ * can't rely on kvm->mmu_lock to keep things ordered.
710	713	*/
	714	+ smp_rmb();
711	715	if (vcpu->kvm->mmu_notifier_seq != mmu_seq)
712	716	return 1;
713	717	return 0;
...	...	@@ -357,11 +357,11 @@
357	357	* been freed.
358	358	*/
359	359	kvm->mmu_notifier_seq++;
	360	+ smp_wmb();
360	361	/*
361	362	* The above sequence increase must be visible before the
362		- * below count decrease but both values are read by the kvm
363		- * page fault under mmu_lock spinlock so we don't need to add
364		- * a smb_wmb() here in between the two.
	363	+ * below count decrease, which is ensured by the smp_wmb above
	364	+ * in conjunction with the smp_rmb in mmu_notifier_retry().
365	365	*/
366	366	kvm->mmu_notifier_count--;
367	367	spin_unlock(&kvm->mmu_lock);