From ad8d75fff811a6a230f7f43b05a6483099349533 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 14 Apr 2009 19:39:12 -0400 Subject: tracing/events: move trace point headers into include/trace/events Impact: clean up Create a sub directory in include/trace called events to keep the trace point headers in their own separate directory. Only headers that declare trace points should be defined in this directory. Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Neil Horman Cc: Zhao Lei Cc: Eduard - Gabriel Munteanu Cc: Pekka Enberg Signed-off-by: Steven Rostedt --- net/core/drop_monitor.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core/drop_monitor.c') diff --git a/net/core/drop_monitor.c b/net/core/drop_monitor.c index 9fd0dc3cca99..b75b6cea49da 100644 --- a/net/core/drop_monitor.c +++ b/net/core/drop_monitor.c @@ -23,7 +23,7 @@ #include #include -#include +#include #include -- cgit v1.2.3 From 683703a26e4677db437a1480682851e27c7a154f Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Mon, 27 Apr 2009 03:17:31 -0700 Subject: drop_monitor: Update netlink protocol to include netlink attribute header in alert message When I initially implemented this protocol, I disregarded the use of netlink attribute headers, thinking for my purposes they weren't needed. I've come to find out that, as I'm starting to work with sending down messages with associated data (like config messages), the kernel code spits out warnings about trailing data in a netlink skb that doesn't have an associated header on it. As such, I'm going to start including attribute headers in my netlink transaction, and so for completeness, I should likely include them on messages bound from the kernel to user space. This patch adds that header to the kernel, and bumps the protocol version accordingly Signed-off-by: Neil Horman Signed-off-by: David S. Miller --- net/core/drop_monitor.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) (limited to 'net/core/drop_monitor.c') diff --git a/net/core/drop_monitor.c b/net/core/drop_monitor.c index 9fd0dc3cca99..2797b711a978 100644 --- a/net/core/drop_monitor.c +++ b/net/core/drop_monitor.c @@ -51,7 +51,7 @@ static struct genl_family net_drop_monitor_family = { .id = GENL_ID_GENERATE, .hdrsize = 0, .name = "NET_DM", - .version = 1, + .version = 2, .maxattr = NET_DM_CMD_MAX, }; @@ -65,13 +65,17 @@ static void reset_per_cpu_data(struct per_cpu_dm_data *data) { size_t al; struct net_dm_alert_msg *msg; + struct nlattr *nla; al = sizeof(struct net_dm_alert_msg); al += dm_hit_limit * sizeof(struct net_dm_drop_point); + al += sizeof(struct nlattr); + data->skb = genlmsg_new(al, GFP_KERNEL); genlmsg_put(data->skb, 0, 0, &net_drop_monitor_family, 0, NET_DM_CMD_ALERT); - msg = __nla_reserve_nohdr(data->skb, sizeof(struct net_dm_alert_msg)); + nla = nla_reserve(data->skb, NLA_UNSPEC, sizeof(struct net_dm_alert_msg)); + msg = nla_data(nla); memset(msg, 0, al); atomic_set(&data->dm_hit_count, dm_hit_limit); } @@ -115,6 +119,7 @@ static void trace_kfree_skb_hit(struct sk_buff *skb, void *location) { struct net_dm_alert_msg *msg; struct nlmsghdr *nlh; + struct nlattr *nla; int i; struct per_cpu_dm_data *data = &__get_cpu_var(dm_cpu_data); @@ -127,7 +132,8 @@ static void trace_kfree_skb_hit(struct sk_buff *skb, void *location) } nlh = (struct nlmsghdr *)data->skb->data; - msg = genlmsg_data(nlmsg_data(nlh)); + nla = genlmsg_data(nlmsg_data(nlh)); + msg = nla_data(nla); for (i = 0; i < msg->entries; i++) { if (!memcmp(&location, msg->points[i].pc, sizeof(void *))) { msg->points[i].count++; @@ -139,6 +145,7 @@ static void trace_kfree_skb_hit(struct sk_buff *skb, void *location) * We need to create a new entry */ __nla_reserve_nohdr(data->skb, sizeof(struct net_dm_drop_point)); + nla->nla_len += NLA_ALIGN(sizeof(struct net_dm_drop_point)); memcpy(msg->points[msg->entries].pc, &location, sizeof(void *)); msg->points[msg->entries].count = 1; msg->entries++; -- cgit v1.2.3 From 4ea7e38696c7e798c47ebbecadfd392f23f814f9 Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Thu, 21 May 2009 07:36:08 +0000 Subject: 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 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 --- 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(-) create mode 100644 include/trace/napi.h (limited to 'net/core/drop_monitor.c') diff --git a/include/linux/net_dropmon.h b/include/linux/net_dropmon.h index 0217fb81a630..e8a8b5c50ed0 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 000000000000..a8989c4547e7 --- /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 92ebeca29901..3942266d1f6c 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 2797b711a978..a6c2ac2828fb 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 @@ out: 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 c8fb45665e4f..b07b25bd2cde 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 64f51eec6576..00b14e2c50ed 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); -- cgit v1.2.3