[RFC PATCH 06/13] workaround: detect time stamp when command flags are expected

[Patrick Ohly <patrick.ohly@intel.com>]

This happens when IP_MULTICAST_LOOP is on. Apparently the time
stamped packet goes through the loop device's start_hard_xmit?!
TODO: find a clean solution.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
---
 net/core/skbuff.c |    7 +++++++
 1 files changed, 7 insertions(+), 0 deletions(-)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 7d714b8..7d9f1dd 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -2574,6 +2574,13 @@ void skb_hwtstamp_tx(struct sk_buff *orig_skb,
 	struct sk_buff *skb;
 	int err = -ENOMEM;
 
+	/* sanity check: extra bits set => might be a real time stamp */
+	if (orig_skb->tstamp.tv64 & ~(SKB_TSTAMP_TX_HARDWARE|SKB_TSTAMP_TX_HARDWARE_IN_PROGRESS|SKB_TSTAMP_TX_SOFTWARE)) {
+		printk(KERN_DEBUG
+			"skb_hwtstamp_tx: invalid command flags\n");
+		return;
+	}
+
 	if (!sk)
 		return;
 
-- 
1.6.0.4