dropmon: add ability to detect when hardware dropsrxpackets

Patch to add the ability to detect drops in hardware interfaces via dropwatch.
Adds a tracepoint to net_rx_action to signal everytime a napi instance is
polled. The dropmon code then periodically checks to see if the rx_frames
counter has changed, and if so, adds a drop notification to the netlink
protocol, using the reserved all-0's vector to indicate the drop location was in
hardware, rather than somewhere in the code.

Signed-off-by: Neil Horman <[e-mail]>

include/linux/net_dropmon.h | 8 ++
include/trace/napi.h | 11 +++
net/core/dev.c | 5 +
net/core/drop_monitor.c | 124 ++++++++++++++++++++++++++++++++++++++++++--
net/core/net-traces.c | 4 +
net/core/netpoll.c | 2
6 files changed, 149 insertions(+), 5 deletions(-)
Signed-off-by: David S. Miller <[e-mail]>

include/linux/net_dropmon.h

···
 #define __NET_DROPMON_H
 
 #include <linux/netlink.h>
+#include <linux/types.h>
 
 struct net_dm_drop_point {
 	__u8 pc[8];
 	__u32 count;
 };
+
+#define is_drop_point_hw(x) do {\
+	int ____i, ____j;\
+	for (____i = 0; ____i < 8; i ____i++)\
+		____j |= x[____i];\
+	____j;\
+} while (0)
 
 #define NET_DM_CFG_VERSION  0
 #define NET_DM_CFG_ALERT_COUNT  1

include/trace/napi.h

···
+#ifndef _TRACE_NAPI_H_
+#define _TRACE_NAPI_H_
+
+#include <linux/netdevice.h>
+#include <linux/tracepoint.h>
+
+DECLARE_TRACE(napi_poll,
+	TP_PROTO(struct napi_struct *napi),
+	TP_ARGS(napi));
+
+#endif

net/core/dev.c

···
 #include <linux/in.h>
 #include <linux/jhash.h>
 #include <linux/random.h>
+#include <trace/napi.h>
 
 #include "net-sysfs.h"
 
···
 		 * accidently calling ->poll() when NAPI is not scheduled.
 		 */
 		work = 0;
-		if (test_bit(NAPI_STATE_SCHED, &n->state))
+		if (test_bit(NAPI_STATE_SCHED, &n->state)) {
 			work = n->poll(n, weight);
+			trace_napi_poll(n);
+		}
 
 		WARN_ON_ONCE(work > weight);
 

net/core/drop_monitor.c

···
 #include <linux/timer.h>
 #include <linux/bitops.h>
 #include <net/genetlink.h>
+#include <net/netevent.h>
 
 #include <trace/skb.h>
+#include <trace/napi.h>
 
 #include <asm/unaligned.h>
 
···
  * and the work handle that will send up
  * netlink alerts
  */
-struct sock *dm_sock;
+static int trace_state = TRACE_OFF;
+static spinlock_t trace_state_lock = SPIN_LOCK_UNLOCKED;
 
 struct per_cpu_dm_data {
 	struct work_struct dm_alert_work;
 	struct sk_buff *skb;
 	atomic_t dm_hit_count;
 	struct timer_list send_timer;
+};
+
+struct dm_hw_stat_delta {
+	struct net_device *dev;
+	struct list_head list;
+	struct rcu_head rcu;
+	unsigned long last_drop_val;
 };
 
 static struct genl_family net_drop_monitor_family = {
···
 
 static int dm_hit_limit = 64;
 static int dm_delay = 1;
-
+static unsigned long dm_hw_check_delta = 2*HZ;
+static LIST_HEAD(hw_stats_list);
 
 static void reset_per_cpu_data(struct per_cpu_dm_data *data)
 {
···
 	schedule_work(&data->dm_alert_work);
 }
 
-static void trace_kfree_skb_hit(struct sk_buff *skb, void *location)
+static void trace_drop_common(struct sk_buff *skb, void *location)
 {
 	struct net_dm_alert_msg *msg;
 	struct nlmsghdr *nlh;
···
 	return;
 }
 
+static void trace_kfree_skb_hit(struct sk_buff *skb, void *location)
+{
+	trace_drop_common(skb, location);
+}
+
+static void trace_napi_poll_hit(struct napi_struct *napi)
+{
+	struct dm_hw_stat_delta *new_stat;
+
+	/*
+	 * Ratelimit our check time to dm_hw_check_delta jiffies
+	 */
+	if (!time_after(jiffies, napi->dev->last_rx + dm_hw_check_delta))
+		return;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(new_stat, &hw_stats_list, list) {
+		if ((new_stat->dev == napi->dev)  &&
+		    (napi->dev->stats.rx_dropped != new_stat->last_drop_val)) {
+			trace_drop_common(NULL, NULL);
+			new_stat->last_drop_val = napi->dev->stats.rx_dropped;
+			break;
+		}
+	}
+	rcu_read_unlock();
+}
+
+
+static void free_dm_hw_stat(struct rcu_head *head)
+{
+	struct dm_hw_stat_delta *n;
+	n = container_of(head, struct dm_hw_stat_delta, rcu);
+	kfree(n);
+}
+
 static int set_all_monitor_traces(int state)
 {
 	int rc = 0;
+	struct dm_hw_stat_delta *new_stat = NULL;
+	struct dm_hw_stat_delta *temp;
+
+	spin_lock(&trace_state_lock);
 
 	switch (state) {
 	case TRACE_ON:
 		rc |= register_trace_kfree_skb(trace_kfree_skb_hit);
+		rc |= register_trace_napi_poll(trace_napi_poll_hit);
 		break;
 	case TRACE_OFF:
 		rc |= unregister_trace_kfree_skb(trace_kfree_skb_hit);
+		rc |= unregister_trace_napi_poll(trace_napi_poll_hit);
 
 		tracepoint_synchronize_unregister();
+
+		/*
+		 * Clean the device list
+		 */
+		list_for_each_entry_safe(new_stat, temp, &hw_stats_list, list) {
+			if (new_stat->dev == NULL) {
+				list_del_rcu(&new_stat->list);
+				call_rcu(&new_stat->rcu, free_dm_hw_stat);
+			}
+		}
 		break;
 	default:
 		rc = 1;
 		break;
 	}
+
+	if (!rc)
+		trace_state = state;
+
+	spin_unlock(&trace_state_lock);
 
 	if (rc)
 		return -EINPROGRESS;
···
 	return -ENOTSUPP;
 }
 
+static int dropmon_net_event(struct notifier_block *ev_block,
+			unsigned long event, void *ptr)
+{
+	struct net_device *dev = ptr;
+	struct dm_hw_stat_delta *new_stat = NULL;
+	struct dm_hw_stat_delta *tmp;
+
+	switch (event) {
+	case NETDEV_REGISTER:
+		new_stat = kzalloc(sizeof(struct dm_hw_stat_delta), GFP_KERNEL);
+
+		if (!new_stat)
+			goto out;
+
+		new_stat->dev = dev;
+		INIT_RCU_HEAD(&new_stat->rcu);
+		spin_lock(&trace_state_lock);
+		list_add_rcu(&new_stat->list, &hw_stats_list);
+		spin_unlock(&trace_state_lock);
+		break;
+	case NETDEV_UNREGISTER:
+		spin_lock(&trace_state_lock);
+		list_for_each_entry_safe(new_stat, tmp, &hw_stats_list, list) {
+			if (new_stat->dev == dev) {
+				new_stat->dev = NULL;
+				if (trace_state == TRACE_OFF) {
+					list_del_rcu(&new_stat->list);
+					call_rcu(&new_stat->rcu, free_dm_hw_stat);
+					break;
+				}
+			}
+		}
+		spin_unlock(&trace_state_lock);
+		break;
+	}
+out:
+	return NOTIFY_DONE;
+}
 
 static struct genl_ops dropmon_ops[] = {
 	{
···
 		.cmd = NET_DM_CMD_STOP,
 		.doit = net_dm_cmd_trace,
 	},
+};
+
+static struct notifier_block dropmon_net_notifier = {
+	.notifier_call = dropmon_net_event
 };
 
 static int __init init_net_drop_monitor(void)
···
 		ret = genl_register_ops(&net_drop_monitor_family,
 					&dropmon_ops[i]);
 		if (ret) {
-			printk(KERN_CRIT "failed to register operation %d\n",
+			printk(KERN_CRIT "Failed to register operation %d\n",
 				dropmon_ops[i].cmd);
 			goto out_unreg;
 		}
+	}
+
+	rc = register_netdevice_notifier(&dropmon_net_notifier);
+	if (rc < 0) {
+		printk(KERN_CRIT "Failed to register netdevice notifier\n");
+		goto out_unreg;
 	}
 
 	rc = 0;
···
 		data->send_timer.data = cpu;
 		data->send_timer.function = sched_send_work;
 	}
+
 	goto out;
 
 out_unreg:

net/core/net-traces.c

···
 #include <linux/netlink.h>
 #include <linux/net_dropmon.h>
 #include <trace/skb.h>
+#include <trace/napi.h>
 
 #include <asm/unaligned.h>
 #include <asm/bitops.h>
···
 
 DEFINE_TRACE(kfree_skb);
 EXPORT_TRACEPOINT_SYMBOL_GPL(kfree_skb);
+
+DEFINE_TRACE(napi_poll);
+EXPORT_TRACEPOINT_SYMBOL_GPL(napi_poll);

net/core/netpoll.c

···
 #include <net/tcp.h>
 #include <net/udp.h>
 #include <asm/unaligned.h>
+#include <trace/napi.h>
 
 /*
  * We maintain a small pool of fully-sized skbs, to make sure the
···
 	set_bit(NAPI_STATE_NPSVC, &napi->state);
 
 	work = napi->poll(napi, budget);
+	trace_napi_poll(napi->dev);
 
 	clear_bit(NAPI_STATE_NPSVC, &napi->state);
 	atomic_dec(&trapped);