rpms/kernel/devel linux-2.6-dropwatch-protocol.patch, NONE, 1.1 config-generic, 1.262, 1.263 kernel.spec, 1.1467, 1.1468

[Neil Horman]

Author: nhorman

Update of /cvs/extras/rpms/kernel/devel
In directory cvs1.fedora.phx.redhat.com:/tmp/cvs-serv4589

Modified Files:
	config-generic kernel.spec 
Added Files:
	linux-2.6-dropwatch-protocol.patch 
Log Message:
Adding dropmonitor netlink protocol bits from 2.6.30 net-next tree

linux-2.6-dropwatch-protocol.patch:

--- NEW FILE linux-2.6-dropwatch-protocol.patch ---
diff -up linux-2.6.29.noarch/include/linux/Kbuild.orig linux-2.6.29.noarch/include/linux/Kbuild
--- linux-2.6.29.noarch/include/linux/Kbuild.orig	2009-03-24 10:31:15.000000000 -0400
+++ linux-2.6.29.noarch/include/linux/Kbuild	2009-03-24 10:30:07.000000000 -0400
@@ -115,6 +115,7 @@ header-y += mqueue.h
 header-y += mtio.h
 header-y += ncp_no.h
 header-y += neighbour.h
+header-y += net_dropmon.h
 header-y += netfilter_arp.h
 header-y += netrom.h
 header-y += nfs2.h
diff -up /dev/null linux-2.6.29.noarch/include/linux/net_dropmon.h
--- /dev/null	2009-03-19 08:40:03.289190608 -0400
+++ linux-2.6.29.noarch/include/linux/net_dropmon.h	2009-03-24 10:28:36.000000000 -0400
@@ -0,0 +1,56 @@
+#ifndef __NET_DROPMON_H
+#define __NET_DROPMON_H
+
+#include <linux/netlink.h>
+
+struct net_dm_drop_point {
+	__u8 pc[8];
+	__u32 count;
+};
+
+#define NET_DM_CFG_VERSION  0
+#define NET_DM_CFG_ALERT_COUNT  1
+#define NET_DM_CFG_ALERT_DELAY 2
+#define NET_DM_CFG_MAX 3
+
+struct net_dm_config_entry {
+	__u32 type;
+	__u64 data __attribute__((aligned(8)));
+};
+
+struct net_dm_config_msg {
+	__u32 entries;
+	struct net_dm_config_entry options[0];
+};
+
+struct net_dm_alert_msg {
+	__u32 entries;
+	struct net_dm_drop_point points[0];
+};
+
+struct net_dm_user_msg {
+	union {
+		struct net_dm_config_msg user;
+		struct net_dm_alert_msg alert;
+	} u;
+};
+
+
+/* These are the netlink message types for this protocol */
+
+enum {
+	NET_DM_CMD_UNSPEC = 0,
+	NET_DM_CMD_ALERT,
+	NET_DM_CMD_CONFIG,
+	NET_DM_CMD_START,
+	NET_DM_CMD_STOP,
+	_NET_DM_CMD_MAX,
+};
+
+#define NET_DM_CMD_MAX (_NET_DM_CMD_MAX - 1)
+
+/*
+ * Our group identifiers
+ */
+#define NET_DM_GRP_ALERT 1
+#endif
diff -up linux-2.6.29.noarch/include/linux/skbuff.h.orig linux-2.6.29.noarch/include/linux/skbuff.h
--- linux-2.6.29.noarch/include/linux/skbuff.h.orig	2009-03-24 10:31:15.000000000 -0400
+++ linux-2.6.29.noarch/include/linux/skbuff.h	2009-03-24 10:04:38.000000000 -0400
@@ -373,6 +373,7 @@ extern void skb_dma_unmap(struct device 
 #endif
 
 extern void kfree_skb(struct sk_buff *skb);
+extern void consume_skb(struct sk_buff *skb);
 extern void	       __kfree_skb(struct sk_buff *skb);
 extern struct sk_buff *__alloc_skb(unsigned int size,
 				   gfp_t priority, int fclone, int node);
@@ -411,7 +412,8 @@ extern int	       skb_to_sgvec(struct sk
 extern int	       skb_cow_data(struct sk_buff *skb, int tailbits,
 				    struct sk_buff **trailer);
 extern int	       skb_pad(struct sk_buff *skb, int pad);
-#define dev_kfree_skb(a)	kfree_skb(a)
+#define dev_kfree_skb(a)	consume_skb(a)
+#define dev_consume_skb(a)	kfree_skb_clean(a)
 extern void	      skb_over_panic(struct sk_buff *skb, int len,
 				     void *here);
 extern void	      skb_under_panic(struct sk_buff *skb, int len,
diff -up /dev/null linux-2.6.29.noarch/include/trace/skb.h
--- /dev/null	2009-03-19 08:40:03.289190608 -0400
+++ linux-2.6.29.noarch/include/trace/skb.h	2009-03-24 10:04:34.000000000 -0400
@@ -0,0 +1,8 @@
+#ifndef _TRACE_SKB_H_
+#define _TRACE_SKB_H_
+
+DECLARE_TRACE(kfree_skb,
+	TPPROTO(struct sk_buff *skb, void *location),
+	TPARGS(skb, location));
+
+#endif
diff -up linux-2.6.29.noarch/net/core/datagram.c.orig linux-2.6.29.noarch/net/core/datagram.c
--- linux-2.6.29.noarch/net/core/datagram.c.orig	2009-03-24 10:31:15.000000000 -0400
+++ linux-2.6.29.noarch/net/core/datagram.c	2009-03-24 10:04:38.000000000 -0400
@@ -208,7 +208,7 @@ struct sk_buff *skb_recv_datagram(struct
 
 void skb_free_datagram(struct sock *sk, struct sk_buff *skb)
 {
-	kfree_skb(skb);
+	consume_skb(skb);
 	sk_mem_reclaim_partial(sk);
 }
 
diff -up /dev/null linux-2.6.29.noarch/net/core/drop_monitor.c
--- /dev/null	2009-03-19 08:40:03.289190608 -0400
+++ linux-2.6.29.noarch/net/core/drop_monitor.c	2009-03-24 10:28:36.000000000 -0400
@@ -0,0 +1,263 @@
+/*
+ * Monitoring code for network dropped packet alerts
+ *
+ * Copyright (C) 2009 Neil Horman <nhorman at tuxdriver.com>
+ */
+
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/string.h>
+#include <linux/if_arp.h>
+#include <linux/inetdevice.h>
+#include <linux/inet.h>
+#include <linux/interrupt.h>
+#include <linux/netpoll.h>
+#include <linux/sched.h>
+#include <linux/delay.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+#include <linux/netlink.h>
+#include <linux/net_dropmon.h>
+#include <linux/percpu.h>
+#include <linux/timer.h>
+#include <linux/bitops.h>
+#include <net/genetlink.h>
+
+#include <trace/skb.h>
+
+#include <asm/unaligned.h>
+
+#define TRACE_ON 1
+#define TRACE_OFF 0
+
+static void send_dm_alert(struct work_struct *unused);
+
+
+/*
+ * Globals, our netlink socket pointer
+ * and the work handle that will send up
+ * netlink alerts
+ */
+struct sock *dm_sock;
+
+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;
+};
+
+static struct genl_family net_drop_monitor_family = {
+	.id             = GENL_ID_GENERATE,
+	.hdrsize        = 0,
+	.name           = "NET_DM",
+	.version        = 1,
+	.maxattr        = NET_DM_CMD_MAX,
+};
+
+static DEFINE_PER_CPU(struct per_cpu_dm_data, dm_cpu_data);
+
+static int dm_hit_limit = 64;
+static int dm_delay = 1;
+
+
+static void reset_per_cpu_data(struct per_cpu_dm_data *data)
+{
+	size_t al;
+	struct net_dm_alert_msg *msg;
+
+	al = sizeof(struct net_dm_alert_msg);
+	al += dm_hit_limit * sizeof(struct net_dm_drop_point);
+	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));
+	memset(msg, 0, al);
+	atomic_set(&data->dm_hit_count, dm_hit_limit);
+}
+
+static void send_dm_alert(struct work_struct *unused)
+{
+	struct sk_buff *skb;
+	struct per_cpu_dm_data *data = &__get_cpu_var(dm_cpu_data);
+
+	/*
+	 * Grab the skb we're about to send
+	 */
+	skb = data->skb;
+
+	/*
+	 * Replace it with a new one
+	 */
+	reset_per_cpu_data(data);
+
+	/*
+	 * Ship it!
+	 */
+	genlmsg_multicast(skb, 0, NET_DM_GRP_ALERT, GFP_KERNEL);
+
+}
+
+/*
+ * This is the timer function to delay the sending of an alert
+ * in the event that more drops will arrive during the
+ * hysteresis period.  Note that it operates under the timer interrupt
+ * so we don't need to disable preemption here
+ */
+static void sched_send_work(unsigned long unused)
+{
+	struct per_cpu_dm_data *data =  &__get_cpu_var(dm_cpu_data);
+
+	schedule_work(&data->dm_alert_work);
+}
+
+static void trace_kfree_skb_hit(struct sk_buff *skb, void *location)
+{
+	struct net_dm_alert_msg *msg;
+	struct nlmsghdr *nlh;
+	int i;
+	struct per_cpu_dm_data *data = &__get_cpu_var(dm_cpu_data);
+
+
+	if (!atomic_add_unless(&data->dm_hit_count, -1, 0)) {
+		/*
+		 * we're already at zero, discard this hit
+		 */
+		goto out;
+	}
+
+	nlh = (struct nlmsghdr *)data->skb->data;
+	msg = genlmsg_data(nlmsg_data(nlh));
+	for (i = 0; i < msg->entries; i++) {
+		if (!memcmp(&location, msg->points[i].pc, sizeof(void *))) {
+			msg->points[i].count++;
+			goto out;
+		}
+	}
+
+	/*
+	 * We need to create a new entry
+	 */
+	__nla_reserve_nohdr(data->skb, sizeof(struct net_dm_drop_point));
+	memcpy(msg->points[msg->entries].pc, &location, sizeof(void *));
+	msg->points[msg->entries].count = 1;
+	msg->entries++;
+
+	if (!timer_pending(&data->send_timer)) {
+		data->send_timer.expires = jiffies + dm_delay * HZ;
+		add_timer_on(&data->send_timer, smp_processor_id());
+	}
+
+out:
+	return;
+}
+
+static int set_all_monitor_traces(int state)
+{
+	int rc = 0;
+
+	switch (state) {
+	case TRACE_ON:
+		rc |= register_trace_kfree_skb(trace_kfree_skb_hit);
+		break;
+	case TRACE_OFF:
+		rc |= unregister_trace_kfree_skb(trace_kfree_skb_hit);
+
+		tracepoint_synchronize_unregister();
+		break;
+	default:
+		rc = 1;
+		break;
+	}
+
+	if (rc)
+		return -EINPROGRESS;
+	return rc;
+}
+
+
+static int net_dm_cmd_config(struct sk_buff *skb,
+			struct genl_info *info)
+{
+	return -ENOTSUPP;
+}
+
+static int net_dm_cmd_trace(struct sk_buff *skb,
+			struct genl_info *info)
+{
+	switch (info->genlhdr->cmd) {
+	case NET_DM_CMD_START:
+		return set_all_monitor_traces(TRACE_ON);
+		break;
+	case NET_DM_CMD_STOP:
+		return set_all_monitor_traces(TRACE_OFF);
+		break;
+	}
+
+	return -ENOTSUPP;
+}
+
+
+static struct genl_ops dropmon_ops[] = {
+	{
+		.cmd = NET_DM_CMD_CONFIG,
+		.doit = net_dm_cmd_config,
+	},
+	{
+		.cmd = NET_DM_CMD_START,
+		.doit = net_dm_cmd_trace,
+	},
+	{
+		.cmd = NET_DM_CMD_STOP,
+		.doit = net_dm_cmd_trace,
+	},
+};
+
+static int __init init_net_drop_monitor(void)
+{
+	int cpu;
+	int rc, i, ret;
+	struct per_cpu_dm_data *data;
+	printk(KERN_INFO "Initalizing network drop monitor service\n");
+
+	if (sizeof(void *) > 8) {
+		printk(KERN_ERR "Unable to store program counters on this arch, Drop monitor failed\n");
+		return -ENOSPC;
+	}
+
+	if (genl_register_family(&net_drop_monitor_family) < 0) {
+		printk(KERN_ERR "Could not create drop monitor netlink family\n");
+		return -EFAULT;
+	}
+
+	rc = -EFAULT;
+
+	for (i = 0; i < ARRAY_SIZE(dropmon_ops); i++) {
+		ret = genl_register_ops(&net_drop_monitor_family,
+					&dropmon_ops[i]);
+		if (ret) {
+			printk(KERN_CRIT "failed to register operation %d\n",
+				dropmon_ops[i].cmd);
+			goto out_unreg;
+		}
+	}
+
+	rc = 0;
+
+	for_each_present_cpu(cpu) {
+		data = &per_cpu(dm_cpu_data, cpu);
+		reset_per_cpu_data(data);
+		INIT_WORK(&data->dm_alert_work, send_dm_alert);
+		init_timer(&data->send_timer);
+		data->send_timer.data = cpu;
+		data->send_timer.function = sched_send_work;
+	}
+	goto out;
+
+out_unreg:
+	genl_unregister_family(&net_drop_monitor_family);
+out:
+	return rc;
+}
+
+late_initcall(init_net_drop_monitor);
diff -up linux-2.6.29.noarch/net/core/Makefile.orig linux-2.6.29.noarch/net/core/Makefile
--- linux-2.6.29.noarch/net/core/Makefile.orig	2009-03-24 10:31:14.000000000 -0400
+++ linux-2.6.29.noarch/net/core/Makefile	2009-03-24 10:30:07.000000000 -0400
@@ -17,3 +17,6 @@ obj-$(CONFIG_NET_PKTGEN) += pktgen.o
 obj-$(CONFIG_NETPOLL) += netpoll.o
 obj-$(CONFIG_NET_DMA) += user_dma.o
 obj-$(CONFIG_FIB_RULES) += fib_rules.o
+obj-$(CONFIG_TRACEPOINTS) += net-traces.o
+obj-$(CONFIG_NET_DROP_MONITOR) += drop_monitor.o
+
diff -up /dev/null linux-2.6.29.noarch/net/core/net-traces.c
--- /dev/null	2009-03-19 08:40:03.289190608 -0400
+++ linux-2.6.29.noarch/net/core/net-traces.c	2009-03-24 10:04:34.000000000 -0400
@@ -0,0 +1,29 @@
+/*
+ * consolidates trace point definitions
+ *
+ * Copyright (C) 2009 Neil Horman <nhorman at tuxdriver.com>
+ */
+
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/string.h>
+#include <linux/if_arp.h>
+#include <linux/inetdevice.h>
+#include <linux/inet.h>
+#include <linux/interrupt.h>
+#include <linux/netpoll.h>
+#include <linux/sched.h>
+#include <linux/delay.h>
+#include <linux/rcupdate.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+#include <linux/netlink.h>
+#include <linux/net_dropmon.h>
+#include <trace/skb.h>
+
+#include <asm/unaligned.h>
+#include <asm/bitops.h>
+
+
+DEFINE_TRACE(kfree_skb);
+EXPORT_TRACEPOINT_SYMBOL_GPL(kfree_skb);
diff -up linux-2.6.29.noarch/net/core/skbuff.c.orig linux-2.6.29.noarch/net/core/skbuff.c
--- linux-2.6.29.noarch/net/core/skbuff.c.orig	2009-03-24 10:31:15.000000000 -0400
+++ linux-2.6.29.noarch/net/core/skbuff.c	2009-03-24 10:28:10.000000000 -0400
@@ -64,6 +64,7 @@
 
 #include <asm/uaccess.h>
 #include <asm/system.h>
+#include <trace/skb.h>
 
 #include "kmap_skb.h"
 
@@ -434,6 +435,26 @@ void kfree_skb(struct sk_buff *skb)
 		smp_rmb();
 	else if (likely(!atomic_dec_and_test(&skb->users)))
 		return;
+	trace_kfree_skb(skb, __builtin_return_address(0));
+	__kfree_skb(skb);
+}
+
+/**
+ *    consume_skb - free an skbuff
+ *    @skb: buffer to free
+ *
+ *    Drop a ref to the buffer and free it if the usage count has hit zero
+ *    Functions identically to kfree_skb, but kfree_skb assumes that the frame
+ *    is being dropped after a failure and notes that
+ */
+void consume_skb(struct sk_buff *skb)
+{
+	if (unlikely(!skb))
+		return;
+	if (likely(atomic_read(&skb->users) == 1))
+		smp_rmb();
+	else if (likely(!atomic_dec_and_test(&skb->users)))
+		return;
 	__kfree_skb(skb);
 }
 
@@ -2895,6 +2916,7 @@ void __skb_warn_lro_forwarding(const str
 EXPORT_SYMBOL(___pskb_trim);
 EXPORT_SYMBOL(__kfree_skb);
 EXPORT_SYMBOL(kfree_skb);
+EXPORT_SYMBOL(consume_skb);
 EXPORT_SYMBOL(__pskb_pull_tail);
 EXPORT_SYMBOL(__alloc_skb);
 EXPORT_SYMBOL(__netdev_alloc_skb);
diff -up linux-2.6.29.noarch/net/ipv4/arp.c.orig linux-2.6.29.noarch/net/ipv4/arp.c
--- linux-2.6.29.noarch/net/ipv4/arp.c.orig	2009-03-24 10:31:15.000000000 -0400
+++ linux-2.6.29.noarch/net/ipv4/arp.c	2009-03-24 10:04:38.000000000 -0400
@@ -892,7 +892,7 @@ static int arp_process(struct sk_buff *s
 out:
 	if (in_dev)
 		in_dev_put(in_dev);
-	kfree_skb(skb);
+	consume_skb(skb);
 	return 0;
 }
 
diff -up linux-2.6.29.noarch/net/ipv4/udp.c.orig linux-2.6.29.noarch/net/ipv4/udp.c
--- linux-2.6.29.noarch/net/ipv4/udp.c.orig	2009-03-24 10:31:15.000000000 -0400
+++ linux-2.6.29.noarch/net/ipv4/udp.c	2009-03-24 10:04:38.000000000 -0400
@@ -1180,7 +1180,7 @@ static int __udp4_lib_mcast_deliver(stru
 			sk = sknext;
 		} while (sknext);
 	} else
-		kfree_skb(skb);
+		consume_skb(skb);
 	spin_unlock(&hslot->lock);
 	return 0;
 }
diff -up linux-2.6.29.noarch/net/Kconfig.orig linux-2.6.29.noarch/net/Kconfig
--- linux-2.6.29.noarch/net/Kconfig.orig	2009-03-24 10:31:15.000000000 -0400
+++ linux-2.6.29.noarch/net/Kconfig	2009-03-24 10:30:07.000000000 -0400
@@ -220,6 +220,17 @@ config NET_TCPPROBE
 	To compile this code as a module, choose M here: the
 	module will be called tcp_probe.
 
+config NET_DROP_MONITOR
+	boolean "Network packet drop alerting service"
+	depends on INET && EXPERIMENTAL && TRACEPOINTS
+	---help---
+	This feature provides an alerting service to userspace in the
+	event that packets are discarded in the network stack.  Alerts
+	are broadcast via netlink socket to any listening user space
+	process.  If you don't need network drop alerts, or if you are ok
+	just checking the various proc files and other utilities for
+	drop statistics, say N here.
+
 endmenu
 
 endmenu
diff -up linux-2.6.29.noarch/net/packet/af_packet.c.orig linux-2.6.29.noarch/net/packet/af_packet.c
--- linux-2.6.29.noarch/net/packet/af_packet.c.orig	2009-03-24 10:31:15.000000000 -0400
+++ linux-2.6.29.noarch/net/packet/af_packet.c	2009-03-24 10:04:38.000000000 -0400
@@ -584,7 +584,7 @@ drop_n_restore:
 		skb->len = skb_len;
 	}
 drop:
-	kfree_skb(skb);
+	consume_skb(skb);
 	return 0;
 }
 


Index: config-generic
===================================================================
RCS file: /cvs/extras/rpms/kernel/devel/config-generic,v
retrieving revision 1.262
retrieving revision 1.263
diff -u -r1.262 -r1.263
--- config-generic	24 Mar 2009 01:31:28 -0000	1.262
+++ config-generic	25 Mar 2009 18:35:42 -0000	1.263
@@ -1068,6 +1068,7 @@
 #
 CONFIG_NET_PKTGEN=m
 # CONFIG_NET_TCPPROBE is not set
+CONFIG_NET_DROP_MONITOR=y
 CONFIG_NETDEVICES=y
 
 #


Index: kernel.spec
===================================================================
RCS file: /cvs/extras/rpms/kernel/devel/kernel.spec,v
retrieving revision 1.1467
retrieving revision 1.1468
diff -u -r1.1467 -r1.1468
--- kernel.spec	25 Mar 2009 16:18:21 -0000	1.1467
+++ kernel.spec	25 Mar 2009 18:35:42 -0000	1.1468
@@ -686,6 +686,9 @@
 
 Patch9002: cpufreq-add-atom-to-p4-clockmod.patch
 
+#Adding dropwatch into rawhide until we get to 2.6.30
+Patch9003: linux-2.6-dropwatch-protocol.patch
+
 %endif
 
 BuildRoot: %{_tmppath}/kernel-%{KVERREL}-root
@@ -1229,6 +1232,8 @@
 
 ApplyPatch cpufreq-add-atom-to-p4-clockmod.patch
 
+ApplyPatch linux-2.6-dropwatch-protocol.patch
+
 # END OF PATCH APPLICATIONS
 
 %endif
@@ -1812,6 +1817,9 @@
 # and build.
 
 %changelog
+* Wed Mar 25 2009 Neil Horman <nhorman at redhat.com>
+- Add dropmonitor/dropwatch protocol from 2.6.30
+
 * Wed Mar 25 2009 Kyle McMartin <kyle at redhat.com>
 - alsa-rewrite-hw_ptr-updaters.patch: snd_pcm_update_hw_ptr() tries to
   detect the unexpected hwptr jumps more strictly to avoid the position