Merge branch 'vhost_net-rx-batch-dequeuing'

Jason Wang says:

====================
vhost_net rx batch dequeuing

This series tries to implement rx batching for vhost-net. This is done
by batching the dequeuing from skb_array which was exported by
underlayer socket and pass the sbk back through msg_control to finish
userspace copying. This is also the requirement for more batching
implemention on rx path.

Tests shows at most 7.56% improvment bon rx pps on top of batch
zeroing and no obvious changes for TCP_STREAM/TCP_RR result.

Please review.

Thanks

Changes from V4:
- drop batch zeroing patch
- renew the performance numbers
- move skb pointer array out of vhost_net structure

Changes from V3:
- add batch zeroing patch to fix the build warnings

Changes from V2:
- rebase to net-next HEAD
- use unconsume helpers to put skb back on releasing
- introduce and use vhost_net internal buffer helpers
- renew performance numbers on top of batch zeroing

Changes from V1:
- switch to use for() in __ptr_ring_consume_batched()
- rename peek_head_len_batched() to fetch_skbs()
- use skb_array_consume_batched() instead of
  skb_array_consume_batched_bh() since no consumer run in bh
- drop the lockless peeking patch since skb_array could be resized, so
  it's not safe to call lockless one
====================

Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

4 files changed, 161 insertions, 0 deletions

diff --git a/include/linux/if_tap.h b/include/linux/if_tap.h
index 3482c3c2037d..4837157da0dc 100644
--- a/include/linux/if_tap.h
+++ b/include/linux/if_tap.h
@@ -3,6 +3,7 @@
 
 #if IS_ENABLED(CONFIG_TAP)
 struct socket *tap_get_socket(struct file *);
+struct skb_array *tap_get_skb_array(struct file *file);
 #else
 #include <linux/err.h>
 #include <linux/errno.h>
@@ -12,6 +13,10 @@ static inline struct socket *tap_get_socket(struct file *f)
 {
 	return ERR_PTR(-EINVAL);
 }
+static inline struct skb_array *tap_get_skb_array(struct file *f)
+{
+	return ERR_PTR(-EINVAL);
+}
 #endif /* CONFIG_TAP */
 
 #include <net/sock.h>
diff --git a/include/linux/if_tun.h b/include/linux/if_tun.h
index ed6da2e6df90..bf9bdf42d577 100644
--- a/include/linux/if_tun.h
+++ b/include/linux/if_tun.h
@@ -19,6 +19,7 @@
 
 #if defined(CONFIG_TUN) || defined(CONFIG_TUN_MODULE)
 struct socket *tun_get_socket(struct file *);
+struct skb_array *tun_get_skb_array(struct file *file);
 #else
 #include <linux/err.h>
 #include <linux/errno.h>
@@ -28,5 +29,9 @@ static inline struct socket *tun_get_socket(struct file *f)
 {
 	return ERR_PTR(-EINVAL);
 }
+static inline struct skb_array *tun_get_skb_array(struct file *f)
+{
+	return ERR_PTR(-EINVAL);
+}
 #endif /* CONFIG_TUN */
 #endif /* __IF_TUN_H */
diff --git a/include/linux/ptr_ring.h b/include/linux/ptr_ring.h
index 6b2e0dd88569..d8c97ec8a8e6 100644
--- a/include/linux/ptr_ring.h
+++ b/include/linux/ptr_ring.h
@@ -278,6 +278,22 @@ static inline void *__ptr_ring_consume(struct ptr_ring *r)
 	return ptr;
 }
 
+static inline int __ptr_ring_consume_batched(struct ptr_ring *r,
+					     void **array, int n)
+{
+	void *ptr;
+	int i;
+
+	for (i = 0; i < n; i++) {
+		ptr = __ptr_ring_consume(r);
+		if (!ptr)
+			break;
+		array[i] = ptr;
+	}
+
+	return i;
+}
+
 /*
  * Note: resize (below) nests producer lock within consumer lock, so if you
  * call this in interrupt or BH context, you must disable interrupts/BH when
@@ -328,6 +344,55 @@ static inline void *ptr_ring_consume_bh(struct ptr_ring *r)
 	return ptr;
 }
 
+static inline int ptr_ring_consume_batched(struct ptr_ring *r,
+					   void **array, int n)
+{
+	int ret;
+
+	spin_lock(&r->consumer_lock);
+	ret = __ptr_ring_consume_batched(r, array, n);
+	spin_unlock(&r->consumer_lock);
+
+	return ret;
+}
+
+static inline int ptr_ring_consume_batched_irq(struct ptr_ring *r,
+					       void **array, int n)
+{
+	int ret;
+
+	spin_lock_irq(&r->consumer_lock);
+	ret = __ptr_ring_consume_batched(r, array, n);
+	spin_unlock_irq(&r->consumer_lock);
+
+	return ret;
+}
+
+static inline int ptr_ring_consume_batched_any(struct ptr_ring *r,
+					       void **array, int n)
+{
+	unsigned long flags;
+	int ret;
+
+	spin_lock_irqsave(&r->consumer_lock, flags);
+	ret = __ptr_ring_consume_batched(r, array, n);
+	spin_unlock_irqrestore(&r->consumer_lock, flags);
+
+	return ret;
+}
+
+static inline int ptr_ring_consume_batched_bh(struct ptr_ring *r,
+					      void **array, int n)
+{
+	int ret;
+
+	spin_lock_bh(&r->consumer_lock);
+	ret = __ptr_ring_consume_batched(r, array, n);
+	spin_unlock_bh(&r->consumer_lock);
+
+	return ret;
+}
+
 /* Cast to structure type and call a function without discarding from FIFO.
  * Function must return a value.
  * Callers must take consumer_lock.
@@ -403,6 +468,61 @@ static inline int ptr_ring_init(struct ptr_ring *r, int size, gfp_t gfp)
 	return 0;
 }
 
+/*
+ * Return entries into ring. Destroy entries that don't fit.
+ *
+ * Note: this is expected to be a rare slow path operation.
+ *
+ * Note: producer lock is nested within consumer lock, so if you
+ * resize you must make sure all uses nest correctly.
+ * In particular if you consume ring in interrupt or BH context, you must
+ * disable interrupts/BH when doing so.
+ */
+static inline void ptr_ring_unconsume(struct ptr_ring *r, void **batch, int n,
+				      void (*destroy)(void *))
+{
+	unsigned long flags;
+	int head;
+
+	spin_lock_irqsave(&r->consumer_lock, flags);
+	spin_lock(&r->producer_lock);
+
+	if (!r->size)
+		goto done;
+
+	/*
+	 * Clean out buffered entries (for simplicity). This way following code
+	 * can test entries for NULL and if not assume they are valid.
+	 */
+	head = r->consumer_head - 1;
+	while (likely(head >= r->consumer_tail))
+		r->queue[head--] = NULL;
+	r->consumer_tail = r->consumer_head;
+
+	/*
+	 * Go over entries in batch, start moving head back and copy entries.
+	 * Stop when we run into previously unconsumed entries.
+	 */
+	while (n) {
+		head = r->consumer_head - 1;
+		if (head < 0)
+			head = r->size - 1;
+		if (r->queue[head]) {
+			/* This batch entry will have to be destroyed. */
+			goto done;
+		}
+		r->queue[head] = batch[--n];
+		r->consumer_tail = r->consumer_head = head;
+	}
+
+done:
+	/* Destroy all entries left in the batch. */
+	while (n)
+		destroy(batch[--n]);
+	spin_unlock(&r->producer_lock);
+	spin_unlock_irqrestore(&r->consumer_lock, flags);
+}
+
 static inline void **__ptr_ring_swap_queue(struct ptr_ring *r, void **queue,
 					   int size, gfp_t gfp,
 					   void (*destroy)(void *))
diff --git a/include/linux/skb_array.h b/include/linux/skb_array.h
index f4dfade428f0..35226cd4efb0 100644
--- a/include/linux/skb_array.h
+++ b/include/linux/skb_array.h
@@ -97,21 +97,46 @@ static inline struct sk_buff *skb_array_consume(struct skb_array *a)
 	return ptr_ring_consume(&a->ring);
 }
 
+static inline int skb_array_consume_batched(struct skb_array *a,
+					    struct sk_buff **array, int n)
+{
+	return ptr_ring_consume_batched(&a->ring, (void **)array, n);
+}
+
 static inline struct sk_buff *skb_array_consume_irq(struct skb_array *a)
 {
 	return ptr_ring_consume_irq(&a->ring);
 }
 
+static inline int skb_array_consume_batched_irq(struct skb_array *a,
+						struct sk_buff **array, int n)
+{
+	return ptr_ring_consume_batched_irq(&a->ring, (void **)array, n);
+}
+
 static inline struct sk_buff *skb_array_consume_any(struct skb_array *a)
 {
 	return ptr_ring_consume_any(&a->ring);
 }
 
+static inline int skb_array_consume_batched_any(struct skb_array *a,
+						struct sk_buff **array, int n)
+{
+	return ptr_ring_consume_batched_any(&a->ring, (void **)array, n);
+}
+
+
 static inline struct sk_buff *skb_array_consume_bh(struct skb_array *a)
 {
 	return ptr_ring_consume_bh(&a->ring);
 }
 
+static inline int skb_array_consume_batched_bh(struct skb_array *a,
+					       struct sk_buff **array, int n)
+{
+	return ptr_ring_consume_batched_bh(&a->ring, (void **)array, n);
+}
+
 static inline int __skb_array_len_with_tag(struct sk_buff *skb)
 {
 	if (likely(skb)) {
@@ -156,6 +181,12 @@ static void __skb_array_destroy_skb(void *ptr)
 	kfree_skb(ptr);
 }
 
+static inline void skb_array_unconsume(struct skb_array *a,
+				       struct sk_buff **skbs, int n)
+{
+	ptr_ring_unconsume(&a->ring, (void **)skbs, n, __skb_array_destroy_skb);
+}
+
 static inline int skb_array_resize(struct skb_array *a, int size, gfp_t gfp)
 {
 	return ptr_ring_resize(&a->ring, size, gfp, __skb_array_destroy_skb);