net: add skb_mstamp infrastructure

ktime_get() is too expensive on some cases, and we'd like to get
usec resolution timestamps in TCP stack.

This patch adds a light weight facility using a combination of
local_clock() and jiffies samples.

Instead of :

        u64 t0, t1;

        t0 = ktime_get();
        // stuff
        t1 = ktime_get();
        delta_us = ktime_us_delta(t1, t0);

use :
        struct skb_mstamp t0, t1;

        skb_mstamp_get(&t0);
        // stuff
        skb_mstamp_get(&t1);
        delta_us = skb_mstamp_us_delta(&t1, &t0);

Note : local_clock() might have a (bounded) drift between cpus.

Do not use this infra in place of ktime_get() without understanding the
issues.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Larry Brakmo <brakmo@google.com>
Cc: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>

include/linux/skbuff.h

@@ -32,6 +32,7 @@
 #include <linux/hrtimer.h>
 #include <linux/dma-mapping.h>
 #include <linux/netdev_features.h>
+#include <linux/sched.h>
 #include <net/flow_keys.h>
 
 /* A. Checksumming of received packets by device.
@@ -356,11 +357,62 @@ typedef unsigned int sk_buff_data_t;
 typedef unsigned char *sk_buff_data_t;
 #endif
 
+/**
+ * struct skb_mstamp - multi resolution time stamps
+ * @stamp_us: timestamp in us resolution
+ * @stamp_jiffies: timestamp in jiffies
+ */
+struct skb_mstamp {
+	union {
+		u64		v64;
+		struct {
+			u32	stamp_us;
+			u32	stamp_jiffies;
+		};
+	};
+};
+
+/**
+ * skb_mstamp_get - get current timestamp
+ * @cl: place to store timestamps
+ */
+static inline void skb_mstamp_get(struct skb_mstamp *cl)
+{
+	u64 val = local_clock();
+
+	do_div(val, NSEC_PER_USEC);
+	cl->stamp_us = (u32)val;
+	cl->stamp_jiffies = (u32)jiffies;
+}
+
+/**
+ * skb_mstamp_delta - compute the difference in usec between two skb_mstamp
+ * @t1: pointer to newest sample
+ * @t0: pointer to oldest sample
+ */
+static inline u32 skb_mstamp_us_delta(const struct skb_mstamp *t1,
+				      const struct skb_mstamp *t0)
+{
+	s32 delta_us = t1->stamp_us - t0->stamp_us;
+	u32 delta_jiffies = t1->stamp_jiffies - t0->stamp_jiffies;
+
+	/* If delta_us is negative, this might be because interval is too big,
+	 * or local_clock() drift is too big : fallback using jiffies.
+	 */
+	if (delta_us <= 0 ||
+	    delta_jiffies >= (INT_MAX / (USEC_PER_SEC / HZ)))
+
+		delta_us = jiffies_to_usecs(delta_jiffies);
+
+	return delta_us;
+}
+
+
 /** 
  *	struct sk_buff - socket buffer
  *	@next: Next buffer in list
  *	@prev: Previous buffer in list
- *	@tstamp: Time we arrived
+ *	@tstamp: Time we arrived/left
  *	@sk: Socket we are owned by
  *	@dev: Device we arrived on/are leaving by
  *	@cb: Control buffer. Free for use by every layer. Put private vars here
@@ -429,7 +481,10 @@ struct sk_buff {
 	struct sk_buff		*next;
 	struct sk_buff		*prev;
 
+	union {
 		ktime_t		tstamp;
+		struct skb_mstamp skb_mstamp;
+	};
 
 	struct sock		*sk;
 	struct net_device	*dev;