diff --git a/include/linux/net_dropmon.h b/include/linux/net_dropmon.h index 0217fb81a63011ae6f8d610e622c9a56855a0472..e8a8b5c50ed0d0862254017f86454c9fb53d909c 100644 --- a/include/linux/net_dropmon.h +++ b/include/linux/net_dropmon.h @@ -2,12 +2,20 @@ #define __NET_DROPMON_H #include +#include 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 #define NET_DM_CFG_ALERT_DELAY 2 diff --git a/include/trace/napi.h b/include/trace/napi.h new file mode 100644 index 0000000000000000000000000000000000000000..a8989c4547e7e7b790140957dcaaa3c7917952cd --- /dev/null +++ b/include/trace/napi.h @@ -0,0 +1,11 @@ +#ifndef _TRACE_NAPI_H_ +#define _TRACE_NAPI_H_ + +#include +#include + +DECLARE_TRACE(napi_poll, + TP_PROTO(struct napi_struct *napi), + TP_ARGS(napi)); + +#endif diff --git a/net/core/dev.c b/net/core/dev.c index 92ebeca29901fd803daba7f2cc979a6a68b890d2..3942266d1f6c7a0ca3c23beb1b9d1be7cc26c9cf 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -126,6 +126,7 @@ #include #include #include +#include #include "net-sysfs.h" @@ -2771,8 +2772,10 @@ static void net_rx_action(struct softirq_action *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); diff --git a/net/core/drop_monitor.c b/net/core/drop_monitor.c index 2797b711a97839c215914fb4c1ebfe1c3ae50689..a6c2ac2828fb22269594621a1892c7d631dd2fee 100644 --- a/net/core/drop_monitor.c +++ b/net/core/drop_monitor.c @@ -22,8 +22,10 @@ #include #include #include +#include #include +#include #include @@ -38,7 +40,8 @@ static void send_dm_alert(struct work_struct *unused); * 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; @@ -47,6 +50,13 @@ struct per_cpu_dm_data { 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 = { .id = GENL_ID_GENERATE, .hdrsize = 0, @@ -59,7 +69,8 @@ static DEFINE_PER_CPU(struct per_cpu_dm_data, dm_cpu_data); 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) { @@ -115,7 +126,7 @@ static void sched_send_work(unsigned long unused) 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; @@ -159,24 +170,80 @@ static void trace_kfree_skb_hit(struct sk_buff *skb, void *location) 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 rc; @@ -204,6 +271,44 @@ static int net_dm_cmd_trace(struct sk_buff *skb, 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[] = { { @@ -220,6 +325,10 @@ static struct genl_ops dropmon_ops[] = { }, }; +static struct notifier_block dropmon_net_notifier = { + .notifier_call = dropmon_net_event +}; + static int __init init_net_drop_monitor(void) { int cpu; @@ -243,12 +352,18 @@ 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; for_each_present_cpu(cpu) { @@ -259,6 +374,7 @@ static int __init init_net_drop_monitor(void) data->send_timer.data = cpu; data->send_timer.function = sched_send_work; } + goto out; out_unreg: diff --git a/net/core/net-traces.c b/net/core/net-traces.c index c8fb45665e4f4af4cf6ced67fe90fff64f74e3ed..b07b25bd2cde44315313a24d9c8d9a10ef54fb71 100644 --- a/net/core/net-traces.c +++ b/net/core/net-traces.c @@ -20,6 +20,7 @@ #include #include #include +#include #include #include @@ -27,3 +28,6 @@ DEFINE_TRACE(kfree_skb); EXPORT_TRACEPOINT_SYMBOL_GPL(kfree_skb); + +DEFINE_TRACE(napi_poll); +EXPORT_TRACEPOINT_SYMBOL_GPL(napi_poll); diff --git a/net/core/netpoll.c b/net/core/netpoll.c index 64f51eec6576290e3b2df361316ee5448396642d..00b14e2c50ed3fa1cb39dd0d9c834a6f51bdbab6 100644 --- a/net/core/netpoll.c +++ b/net/core/netpoll.c @@ -24,6 +24,7 @@ #include #include #include +#include /* * We maintain a small pool of fully-sized skbs, to make sure the @@ -137,6 +138,7 @@ static int poll_one_napi(struct netpoll_info *npinfo, 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);