netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH net-next-2.6] net: Consistent skb timestamping
@ 2010-05-06 12:01 Eric Dumazet
  2010-05-06 15:12 ` Tom Herbert
  0 siblings, 1 reply; 7+ messages in thread
From: Eric Dumazet @ 2010-05-06 12:01 UTC (permalink / raw)
  To: David Miller; +Cc: netdev, Tom Herbert

With RPS inclusion, skb timestamping is not consistent in RX path.

If netif_receive_skb() is used, its deferred after RPS dispatch.

If netif_rx() is used, its done before RPS dispatch.

This can give strange tcpdump timestamps results.

I think timestamping should be done as soon as possible in the receive
path, to get meaningful values (ie timestamps taken at the time packet
was delivered by NIC driver to our stack), even if NAPI already can
defer timestamping a bit (RPS can help to reduce the gap)

Remove timestamping from __netif_receive_skb, and add it to
netif_receive_skb(), before RPS.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
 net/core/dev.c |   46 ++++++++++++++++++++++++++--------------------
 1 file changed, 26 insertions(+), 20 deletions(-)

diff --git a/net/core/dev.c b/net/core/dev.c
index 36d53be..3278003 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -1454,7 +1454,7 @@ void net_disable_timestamp(void)
 }
 EXPORT_SYMBOL(net_disable_timestamp);
 
-static inline void net_timestamp(struct sk_buff *skb)
+static inline void net_timestamp_set(struct sk_buff *skb)
 {
 	if (atomic_read(&netstamp_needed))
 		__net_timestamp(skb);
@@ -1462,6 +1462,12 @@ static inline void net_timestamp(struct sk_buff *skb)
 		skb->tstamp.tv64 = 0;
 }
 
+static inline void net_timestamp_check(struct sk_buff *skb)
+{
+	if (!skb->tstamp.tv64 && atomic_read(&netstamp_needed))
+		__net_timestamp(skb);
+}
+
 /**
  * dev_forward_skb - loopback an skb to another netif
  *
@@ -1509,9 +1515,9 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev)
 
 #ifdef CONFIG_NET_CLS_ACT
 	if (!(skb->tstamp.tv64 && (G_TC_FROM(skb->tc_verd) & AT_INGRESS)))
-		net_timestamp(skb);
+		net_timestamp_set(skb);
 #else
-	net_timestamp(skb);
+	net_timestamp_set(skb);
 #endif
 
 	rcu_read_lock();
@@ -2458,8 +2464,7 @@ int netif_rx(struct sk_buff *skb)
 	if (netpoll_rx(skb))
 		return NET_RX_DROP;
 
-	if (!skb->tstamp.tv64)
-		net_timestamp(skb);
+	net_timestamp_check(skb);
 
 #ifdef CONFIG_RPS
 	{
@@ -2780,9 +2785,6 @@ static int __netif_receive_skb(struct sk_buff *skb)
 	int ret = NET_RX_DROP;
 	__be16 type;
 
-	if (!skb->tstamp.tv64)
-		net_timestamp(skb);
-
 	if (vlan_tx_tag_present(skb) && vlan_hwaccel_do_receive(skb))
 		return NET_RX_SUCCESS;
 
@@ -2899,23 +2901,27 @@ out:
  */
 int netif_receive_skb(struct sk_buff *skb)
 {
+	net_timestamp_check(skb);
+
 #ifdef CONFIG_RPS
-	struct rps_dev_flow voidflow, *rflow = &voidflow;
-	int cpu, ret;
+	{
+		struct rps_dev_flow voidflow, *rflow = &voidflow;
+		int cpu, ret;
 
-	rcu_read_lock();
+		rcu_read_lock();
 
-	cpu = get_rps_cpu(skb->dev, skb, &rflow);
+		cpu = get_rps_cpu(skb->dev, skb, &rflow);
 
-	if (cpu >= 0) {
-		ret = enqueue_to_backlog(skb, cpu, &rflow->last_qtail);
-		rcu_read_unlock();
-	} else {
-		rcu_read_unlock();
-		ret = __netif_receive_skb(skb);
-	}
+		if (cpu >= 0) {
+			ret = enqueue_to_backlog(skb, cpu, &rflow->last_qtail);
+			rcu_read_unlock();
+		} else {
+			rcu_read_unlock();
+			ret = __netif_receive_skb(skb);
+		}
 
-	return ret;
+		return ret;
+	}
 #else
 	return __netif_receive_skb(skb);
 #endif



^ permalink raw reply related	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2010-05-16 18:31 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-05-06 12:01 [PATCH net-next-2.6] net: Consistent skb timestamping Eric Dumazet
2010-05-06 15:12 ` Tom Herbert
2010-05-06 15:37   ` Eric Dumazet
2010-05-06 16:14     ` Eric Dumazet
2010-05-16  6:57       ` David Miller
2010-05-16  6:56   ` David Miller
2010-05-16 18:30     ` Dimitris Michailidis

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).