netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v3 net-next 0/5] net: better lockdep annotations
@ 2016-06-09 14:45 Eric Dumazet
  2016-06-09 14:45 ` [PATCH v3 net-next 1/5] net: sched: fix qdisc->running " Eric Dumazet
                   ` (5 more replies)
  0 siblings, 6 replies; 8+ messages in thread
From: Eric Dumazet @ 2016-06-09 14:45 UTC (permalink / raw)
  To: David S . Miller; +Cc: netdev, Eric Dumazet, David Ahern, Eric Dumazet

Introduction of qdisc->running seqcount added lockdep false positives.

While chasing the bug, it came to me that we had a lot of copies of the
same stuff in virtual drivers.

This patch series has the qdisc->running fix (considers that a trylock
is attempted in lockdep terminology), and adds a generic helper so
that we no longer have to patch many virtual drivers when a new per-device
or per-qdisc lock is added.

Thanks to David Ahern for reporting the issue and testing my patches :)


Eric Dumazet (5):
  net: sched: fix qdisc->running lockdep annotations
  net: add netdev_lockdep_set_classes() helper
  net: vrf: call netdev_lockdep_set_classes()
  net: macvlan: call netdev_lockdep_set_classes()
  net: ipvlan: call netdev_lockdep_set_classes()

 drivers/net/bonding/bond_main.c  | 24 +-----------------------
 drivers/net/ipvlan/ipvlan_main.c | 22 +---------------------
 drivers/net/macvlan.c            | 11 +----------
 drivers/net/ppp/ppp_generic.c    |  6 +-----
 drivers/net/team/team.c          | 21 +--------------------
 drivers/net/vrf.c                |  2 +-
 include/linux/netdevice.h        | 17 +++++++++++++++++
 include/net/sch_generic.h        |  6 +++++-
 net/bluetooth/6lowpan.c          | 15 +--------------
 net/ieee802154/6lowpan/core.c    | 16 +---------------
 net/l2tp/l2tp_eth.c              |  6 +-----
 net/sched/sch_generic.c          |  4 ++--
 12 files changed, 33 insertions(+), 117 deletions(-)

-- 
2.8.0.rc3.226.g39d4020

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [PATCH v3 net-next 1/5] net: sched: fix qdisc->running lockdep annotations
  2016-06-09 14:45 [PATCH v3 net-next 0/5] net: better lockdep annotations Eric Dumazet
@ 2016-06-09 14:45 ` Eric Dumazet
  2016-06-09 14:45 ` [PATCH v3 net-next 2/5] net: add netdev_lockdep_set_classes() helper Eric Dumazet
                   ` (4 subsequent siblings)
  5 siblings, 0 replies; 8+ messages in thread
From: Eric Dumazet @ 2016-06-09 14:45 UTC (permalink / raw)
  To: David S . Miller; +Cc: netdev, Eric Dumazet, David Ahern, Eric Dumazet

1) qdisc_run_begin() is really using the equivalent of a trylock.
  Instead of using write_seqcount_begin(), use a combination of
  raw_write_seqcount_begin() and correct lockdep annotation.

2) sch_direct_xmit() should use regular spin_lock(root_lock)

Fixes: f9eb8aea2a1e ("net_sched: transform qdisc running bit into a seqcount")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: David Ahern <dsa@cumulusnetworks.com>
---
 include/net/sch_generic.h | 6 +++++-
 net/sched/sch_generic.c   | 4 ++--
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index c4f5749342ec..d53ecfc1d60e 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -104,7 +104,11 @@ static inline bool qdisc_run_begin(struct Qdisc *qdisc)
 {
 	if (qdisc_is_running(qdisc))
 		return false;
-	write_seqcount_begin(&qdisc->running);
+	/* Variant of write_seqcount_begin() telling lockdep a trylock
+	 * was attempted.
+	 */
+	raw_write_seqcount_begin(&qdisc->running);
+	seqcount_acquire(&qdisc->running.dep_map, 0, 1, _RET_IP_);
 	return true;
 }
 
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index cebea73e70ac..cad810b45e31 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -137,10 +137,10 @@ int sch_direct_xmit(struct sk_buff *skb, struct Qdisc *q,
 
 		HARD_TX_UNLOCK(dev, txq);
 	} else {
-		spin_lock_nested(root_lock, SINGLE_DEPTH_NESTING);
+		spin_lock(root_lock);
 		return qdisc_qlen(q);
 	}
-	spin_lock_nested(root_lock, SINGLE_DEPTH_NESTING);
+	spin_lock(root_lock);
 
 	if (dev_xmit_complete(ret)) {
 		/* Driver sent out skb successfully or skb was consumed */
-- 
2.8.0.rc3.226.g39d4020

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* [PATCH v3 net-next 2/5] net: add netdev_lockdep_set_classes() helper
  2016-06-09 14:45 [PATCH v3 net-next 0/5] net: better lockdep annotations Eric Dumazet
  2016-06-09 14:45 ` [PATCH v3 net-next 1/5] net: sched: fix qdisc->running " Eric Dumazet
@ 2016-06-09 14:45 ` Eric Dumazet
  2016-06-09 14:45 ` [PATCH v3 net-next 3/5] net: vrf: call netdev_lockdep_set_classes() Eric Dumazet
                   ` (3 subsequent siblings)
  5 siblings, 0 replies; 8+ messages in thread
From: Eric Dumazet @ 2016-06-09 14:45 UTC (permalink / raw)
  To: David S . Miller; +Cc: netdev, Eric Dumazet, David Ahern, Eric Dumazet

It is time to add netdev_lockdep_set_classes() helper
so that lockdep annotations per device type are easier to manage.

This removes a lot of copies and missing annotations.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 drivers/net/bonding/bond_main.c | 24 +-----------------------
 drivers/net/ppp/ppp_generic.c   |  6 +-----
 drivers/net/team/team.c         | 21 +--------------------
 include/linux/netdevice.h       | 17 +++++++++++++++++
 net/bluetooth/6lowpan.c         | 15 +--------------
 net/ieee802154/6lowpan/core.c   | 16 +---------------
 net/l2tp/l2tp_eth.c             |  6 +-----
 7 files changed, 23 insertions(+), 82 deletions(-)

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 681af31a60ed..90157e20357e 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -4607,28 +4607,6 @@ static int bond_check_params(struct bond_params *params)
 	return 0;
 }
 
-static struct lock_class_key bonding_netdev_xmit_lock_key;
-static struct lock_class_key bonding_netdev_addr_lock_key;
-static struct lock_class_key bonding_tx_busylock_key;
-static struct lock_class_key bonding_qdisc_running_key;
-
-static void bond_set_lockdep_class_one(struct net_device *dev,
-				       struct netdev_queue *txq,
-				       void *_unused)
-{
-	lockdep_set_class(&txq->_xmit_lock,
-			  &bonding_netdev_xmit_lock_key);
-}
-
-static void bond_set_lockdep_class(struct net_device *dev)
-{
-	lockdep_set_class(&dev->addr_list_lock,
-			  &bonding_netdev_addr_lock_key);
-	netdev_for_each_tx_queue(dev, bond_set_lockdep_class_one, NULL);
-	dev->qdisc_tx_busylock = &bonding_tx_busylock_key;
-	dev->qdisc_running_key = &bonding_qdisc_running_key;
-}
-
 /* Called from registration process */
 static int bond_init(struct net_device *bond_dev)
 {
@@ -4641,7 +4619,7 @@ static int bond_init(struct net_device *bond_dev)
 	if (!bond->wq)
 		return -ENOMEM;
 
-	bond_set_lockdep_class(bond_dev);
+	netdev_lockdep_set_classes(bond_dev);
 
 	list_add_tail(&bond->bond_list, &bn->dev_list);
 
diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
index aeabaa42317f..17953ab15000 100644
--- a/drivers/net/ppp/ppp_generic.c
+++ b/drivers/net/ppp/ppp_generic.c
@@ -1312,13 +1312,9 @@ ppp_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats64)
 	return stats64;
 }
 
-static struct lock_class_key ppp_tx_busylock;
-static struct lock_class_key ppp_qdisc_running_key;
-
 static int ppp_dev_init(struct net_device *dev)
 {
-	dev->qdisc_tx_busylock = &ppp_tx_busylock;
-	dev->qdisc_running_key = &ppp_qdisc_running_key;
+	netdev_lockdep_set_classes(dev);
 	return 0;
 }
 
diff --git a/drivers/net/team/team.c b/drivers/net/team/team.c
index 00eb38956a2c..0a1bb8387d96 100644
--- a/drivers/net/team/team.c
+++ b/drivers/net/team/team.c
@@ -1574,25 +1574,6 @@ static const struct team_option team_options[] = {
 	},
 };
 
-static struct lock_class_key team_netdev_xmit_lock_key;
-static struct lock_class_key team_netdev_addr_lock_key;
-static struct lock_class_key team_tx_busylock_key;
-static struct lock_class_key team_qdisc_running_key;
-
-static void team_set_lockdep_class_one(struct net_device *dev,
-				       struct netdev_queue *txq,
-				       void *unused)
-{
-	lockdep_set_class(&txq->_xmit_lock, &team_netdev_xmit_lock_key);
-}
-
-static void team_set_lockdep_class(struct net_device *dev)
-{
-	lockdep_set_class(&dev->addr_list_lock, &team_netdev_addr_lock_key);
-	netdev_for_each_tx_queue(dev, team_set_lockdep_class_one, NULL);
-	dev->qdisc_tx_busylock = &team_tx_busylock_key;
-	dev->qdisc_running_key = &team_qdisc_running_key;
-}
 
 static int team_init(struct net_device *dev)
 {
@@ -1628,7 +1609,7 @@ static int team_init(struct net_device *dev)
 		goto err_options_register;
 	netif_carrier_off(dev);
 
-	team_set_lockdep_class(dev);
+	netdev_lockdep_set_classes(dev);
 
 	return 0;
 
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 541562333ba5..4f234b102892 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -1946,6 +1946,23 @@ static inline void netdev_for_each_tx_queue(struct net_device *dev,
 		f(dev, &dev->_tx[i], arg);
 }
 
+#define netdev_lockdep_set_classes(dev)				\
+{								\
+	static struct lock_class_key qdisc_tx_busylock_key;	\
+	static struct lock_class_key qdisc_running_key;		\
+	static struct lock_class_key qdisc_xmit_lock_key;	\
+	static struct lock_class_key dev_addr_list_lock_key;	\
+	unsigned int i;						\
+								\
+	(dev)->qdisc_tx_busylock = &qdisc_tx_busylock_key;	\
+	(dev)->qdisc_running_key = &qdisc_running_key;		\
+	lockdep_set_class(&(dev)->addr_list_lock,		\
+			  &dev_addr_list_lock_key); 		\
+	for (i = 0; i < (dev)->num_tx_queues; i++)		\
+		lockdep_set_class(&(dev)->_tx[i]._xmit_lock,	\
+				  &qdisc_xmit_lock_key);	\
+}
+
 struct netdev_queue *netdev_pick_tx(struct net_device *dev,
 				    struct sk_buff *skb,
 				    void *accel_priv);
diff --git a/net/bluetooth/6lowpan.c b/net/bluetooth/6lowpan.c
index 977a11e418d0..d020299baba4 100644
--- a/net/bluetooth/6lowpan.c
+++ b/net/bluetooth/6lowpan.c
@@ -627,22 +627,9 @@ static netdev_tx_t bt_xmit(struct sk_buff *skb, struct net_device *netdev)
 	return err < 0 ? NET_XMIT_DROP : err;
 }
 
-static struct lock_class_key bt_tx_busylock;
-static struct lock_class_key bt_netdev_xmit_lock_key;
-static struct lock_class_key bt_qdisc_running_key;
-
-static void bt_set_lockdep_class_one(struct net_device *dev,
-				     struct netdev_queue *txq,
-				     void *_unused)
-{
-	lockdep_set_class(&txq->_xmit_lock, &bt_netdev_xmit_lock_key);
-}
-
 static int bt_dev_init(struct net_device *dev)
 {
-	netdev_for_each_tx_queue(dev, bt_set_lockdep_class_one, NULL);
-	dev->qdisc_tx_busylock = &bt_tx_busylock;
-	dev->qdisc_running_key = &bt_qdisc_running_key;
+	netdev_lockdep_set_classes(dev);
 
 	return 0;
 }
diff --git a/net/ieee802154/6lowpan/core.c b/net/ieee802154/6lowpan/core.c
index 14aa5effd29a..4e2b30894224 100644
--- a/net/ieee802154/6lowpan/core.c
+++ b/net/ieee802154/6lowpan/core.c
@@ -58,23 +58,9 @@ static struct header_ops lowpan_header_ops = {
 	.create	= lowpan_header_create,
 };
 
-static struct lock_class_key lowpan_tx_busylock;
-static struct lock_class_key lowpan_netdev_xmit_lock_key;
-static struct lock_class_key lowpan_qdisc_running_key;
-
-static void lowpan_set_lockdep_class_one(struct net_device *ldev,
-					 struct netdev_queue *txq,
-					 void *_unused)
-{
-	lockdep_set_class(&txq->_xmit_lock,
-			  &lowpan_netdev_xmit_lock_key);
-}
-
 static int lowpan_dev_init(struct net_device *ldev)
 {
-	netdev_for_each_tx_queue(ldev, lowpan_set_lockdep_class_one, NULL);
-	ldev->qdisc_tx_busylock = &lowpan_tx_busylock;
-	ldev->qdisc_running_key = &lowpan_qdisc_running_key;
+	netdev_lockdep_set_classes(ldev);
 
 	return 0;
 }
diff --git a/net/l2tp/l2tp_eth.c b/net/l2tp/l2tp_eth.c
index c00d72d182fa..57fc5a46ce06 100644
--- a/net/l2tp/l2tp_eth.c
+++ b/net/l2tp/l2tp_eth.c
@@ -67,9 +67,6 @@ static inline struct l2tp_eth_net *l2tp_eth_pernet(struct net *net)
 	return net_generic(net, l2tp_eth_net_id);
 }
 
-static struct lock_class_key l2tp_eth_tx_busylock;
-static struct lock_class_key l2tp_qdisc_running_key;
-
 static int l2tp_eth_dev_init(struct net_device *dev)
 {
 	struct l2tp_eth *priv = netdev_priv(dev);
@@ -77,8 +74,7 @@ static int l2tp_eth_dev_init(struct net_device *dev)
 	priv->dev = dev;
 	eth_hw_addr_random(dev);
 	eth_broadcast_addr(dev->broadcast);
-	dev->qdisc_tx_busylock = &l2tp_eth_tx_busylock;
-	dev->qdisc_running_key = &l2tp_qdisc_running_key;
+	netdev_lockdep_set_classes(dev);
 
 	return 0;
 }
-- 
2.8.0.rc3.226.g39d4020

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* [PATCH v3 net-next 3/5] net: vrf: call netdev_lockdep_set_classes()
  2016-06-09 14:45 [PATCH v3 net-next 0/5] net: better lockdep annotations Eric Dumazet
  2016-06-09 14:45 ` [PATCH v3 net-next 1/5] net: sched: fix qdisc->running " Eric Dumazet
  2016-06-09 14:45 ` [PATCH v3 net-next 2/5] net: add netdev_lockdep_set_classes() helper Eric Dumazet
@ 2016-06-09 14:45 ` Eric Dumazet
  2016-06-09 15:05   ` David Ahern
  2016-06-09 14:45 ` [PATCH v3 net-next 4/5] net: macvlan: " Eric Dumazet
                   ` (2 subsequent siblings)
  5 siblings, 1 reply; 8+ messages in thread
From: Eric Dumazet @ 2016-06-09 14:45 UTC (permalink / raw)
  To: David S . Miller; +Cc: netdev, Eric Dumazet, David Ahern, Eric Dumazet

In case a qdisc is used on a vrf device, we need to use different
lockdep classes to avoid false positives.

Use the new netdev_lockdep_set_classes() generic helper.

Reported-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 drivers/net/vrf.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/vrf.c b/drivers/net/vrf.c
index e3e693bf0dec..b82e3527924e 100644
--- a/drivers/net/vrf.c
+++ b/drivers/net/vrf.c
@@ -661,7 +661,7 @@ static int vrf_dev_init(struct net_device *dev)
 
 	/* similarly, oper state is irrelevant; set to up to avoid confusion */
 	dev->operstate = IF_OPER_UP;
-
+	netdev_lockdep_set_classes(dev);
 	return 0;
 
 out_rth:
-- 
2.8.0.rc3.226.g39d4020

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* [PATCH v3 net-next 4/5] net: macvlan: call netdev_lockdep_set_classes()
  2016-06-09 14:45 [PATCH v3 net-next 0/5] net: better lockdep annotations Eric Dumazet
                   ` (2 preceding siblings ...)
  2016-06-09 14:45 ` [PATCH v3 net-next 3/5] net: vrf: call netdev_lockdep_set_classes() Eric Dumazet
@ 2016-06-09 14:45 ` Eric Dumazet
  2016-06-09 14:45 ` [PATCH v3 net-next 5/5] net: ipvlan: " Eric Dumazet
  2016-06-09 20:29 ` [PATCH v3 net-next 0/5] net: better lockdep annotations David Miller
  5 siblings, 0 replies; 8+ messages in thread
From: Eric Dumazet @ 2016-06-09 14:45 UTC (permalink / raw)
  To: David S . Miller; +Cc: netdev, Eric Dumazet, David Ahern, Eric Dumazet

In case a qdisc is used on a macvlan device, we need to use different
lockdep classes to avoid false positives.

Use the new netdev_lockdep_set_classes() generic helper.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 drivers/net/macvlan.c | 11 +----------
 1 file changed, 1 insertion(+), 10 deletions(-)

diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c
index 0c65bd914aed..cd9b53834bf6 100644
--- a/drivers/net/macvlan.c
+++ b/drivers/net/macvlan.c
@@ -788,7 +788,6 @@ static int macvlan_change_mtu(struct net_device *dev, int new_mtu)
  * "super class" of normal network devices; split their locks off into a
  * separate class since they always nest.
  */
-static struct lock_class_key macvlan_netdev_xmit_lock_key;
 static struct lock_class_key macvlan_netdev_addr_lock_key;
 
 #define ALWAYS_ON_FEATURES \
@@ -809,20 +808,12 @@ static int macvlan_get_nest_level(struct net_device *dev)
 	return ((struct macvlan_dev *)netdev_priv(dev))->nest_level;
 }
 
-static void macvlan_set_lockdep_class_one(struct net_device *dev,
-					  struct netdev_queue *txq,
-					  void *_unused)
-{
-	lockdep_set_class(&txq->_xmit_lock,
-			  &macvlan_netdev_xmit_lock_key);
-}
-
 static void macvlan_set_lockdep_class(struct net_device *dev)
 {
+	netdev_lockdep_set_classes(dev);
 	lockdep_set_class_and_subclass(&dev->addr_list_lock,
 				       &macvlan_netdev_addr_lock_key,
 				       macvlan_get_nest_level(dev));
-	netdev_for_each_tx_queue(dev, macvlan_set_lockdep_class_one, NULL);
 }
 
 static int macvlan_init(struct net_device *dev)
-- 
2.8.0.rc3.226.g39d4020

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* [PATCH v3 net-next 5/5] net: ipvlan: call netdev_lockdep_set_classes()
  2016-06-09 14:45 [PATCH v3 net-next 0/5] net: better lockdep annotations Eric Dumazet
                   ` (3 preceding siblings ...)
  2016-06-09 14:45 ` [PATCH v3 net-next 4/5] net: macvlan: " Eric Dumazet
@ 2016-06-09 14:45 ` Eric Dumazet
  2016-06-09 20:29 ` [PATCH v3 net-next 0/5] net: better lockdep annotations David Miller
  5 siblings, 0 replies; 8+ messages in thread
From: Eric Dumazet @ 2016-06-09 14:45 UTC (permalink / raw)
  To: David S . Miller; +Cc: netdev, Eric Dumazet, David Ahern, Eric Dumazet

In case a qdisc is used on a ipvlan device, we need to use different
lockdep classes to avoid false positives.

Use the new netdev_lockdep_set_classes() generic helper.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 drivers/net/ipvlan/ipvlan_main.c | 22 +---------------------
 1 file changed, 1 insertion(+), 21 deletions(-)

diff --git a/drivers/net/ipvlan/ipvlan_main.c b/drivers/net/ipvlan/ipvlan_main.c
index 1c4d395fbd49..18b4e8c7f68a 100644
--- a/drivers/net/ipvlan/ipvlan_main.c
+++ b/drivers/net/ipvlan/ipvlan_main.c
@@ -80,13 +80,6 @@ static void ipvlan_port_destroy(struct net_device *dev)
 	kfree_rcu(port, rcu);
 }
 
-/* ipvlan network devices have devices nesting below it and are a special
- * "super class" of normal network devices; split their locks off into a
- * separate class since they always nest.
- */
-static struct lock_class_key ipvlan_netdev_xmit_lock_key;
-static struct lock_class_key ipvlan_netdev_addr_lock_key;
-
 #define IPVLAN_FEATURES \
 	(NETIF_F_SG | NETIF_F_HW_CSUM | NETIF_F_HIGHDMA | NETIF_F_FRAGLIST | \
 	 NETIF_F_GSO | NETIF_F_TSO | NETIF_F_UFO | NETIF_F_GSO_ROBUST | \
@@ -96,19 +89,6 @@ static struct lock_class_key ipvlan_netdev_addr_lock_key;
 #define IPVLAN_STATE_MASK \
 	((1<<__LINK_STATE_NOCARRIER) | (1<<__LINK_STATE_DORMANT))
 
-static void ipvlan_set_lockdep_class_one(struct net_device *dev,
-					 struct netdev_queue *txq,
-					 void *_unused)
-{
-	lockdep_set_class(&txq->_xmit_lock, &ipvlan_netdev_xmit_lock_key);
-}
-
-static void ipvlan_set_lockdep_class(struct net_device *dev)
-{
-	lockdep_set_class(&dev->addr_list_lock, &ipvlan_netdev_addr_lock_key);
-	netdev_for_each_tx_queue(dev, ipvlan_set_lockdep_class_one, NULL);
-}
-
 static int ipvlan_init(struct net_device *dev)
 {
 	struct ipvl_dev *ipvlan = netdev_priv(dev);
@@ -123,7 +103,7 @@ static int ipvlan_init(struct net_device *dev)
 	dev->gso_max_segs = phy_dev->gso_max_segs;
 	dev->hard_header_len = phy_dev->hard_header_len;
 
-	ipvlan_set_lockdep_class(dev);
+	netdev_lockdep_set_classes(dev);
 
 	ipvlan->pcpu_stats = alloc_percpu(struct ipvl_pcpu_stats);
 	if (!ipvlan->pcpu_stats)
-- 
2.8.0.rc3.226.g39d4020

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* Re: [PATCH v3 net-next 3/5] net: vrf: call netdev_lockdep_set_classes()
  2016-06-09 14:45 ` [PATCH v3 net-next 3/5] net: vrf: call netdev_lockdep_set_classes() Eric Dumazet
@ 2016-06-09 15:05   ` David Ahern
  0 siblings, 0 replies; 8+ messages in thread
From: David Ahern @ 2016-06-09 15:05 UTC (permalink / raw)
  To: Eric Dumazet, David S . Miller; +Cc: netdev, Eric Dumazet

On 6/9/16 8:45 AM, Eric Dumazet wrote:
> In case a qdisc is used on a vrf device, we need to use different
> lockdep classes to avoid false positives.
>
> Use the new netdev_lockdep_set_classes() generic helper.
>
> Reported-by: David Ahern <dsa@cumulusnetworks.com>
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> ---
>  drivers/net/vrf.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/vrf.c b/drivers/net/vrf.c
> index e3e693bf0dec..b82e3527924e 100644
> --- a/drivers/net/vrf.c
> +++ b/drivers/net/vrf.c
> @@ -661,7 +661,7 @@ static int vrf_dev_init(struct net_device *dev)
>
>  	/* similarly, oper state is irrelevant; set to up to avoid confusion */
>  	dev->operstate = IF_OPER_UP;
> -
> +	netdev_lockdep_set_classes(dev);
>  	return 0;
>
>  out_rth:
>

Works for me. Thanks, Eric.

Tested-by: David Ahern <dsa@cumulusnetworks.com>

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v3 net-next 0/5] net: better lockdep annotations
  2016-06-09 14:45 [PATCH v3 net-next 0/5] net: better lockdep annotations Eric Dumazet
                   ` (4 preceding siblings ...)
  2016-06-09 14:45 ` [PATCH v3 net-next 5/5] net: ipvlan: " Eric Dumazet
@ 2016-06-09 20:29 ` David Miller
  5 siblings, 0 replies; 8+ messages in thread
From: David Miller @ 2016-06-09 20:29 UTC (permalink / raw)
  To: edumazet; +Cc: netdev, dsa, eric.dumazet

From: Eric Dumazet <edumazet@google.com>
Date: Thu,  9 Jun 2016 07:45:10 -0700

> Introduction of qdisc->running seqcount added lockdep false positives.
> 
> While chasing the bug, it came to me that we had a lot of copies of the
> same stuff in virtual drivers.
> 
> This patch series has the qdisc->running fix (considers that a trylock
> is attempted in lockdep terminology), and adds a generic helper so
> that we no longer have to patch many virtual drivers when a new per-device
> or per-qdisc lock is added.
> 
> Thanks to David Ahern for reporting the issue and testing my patches :)

Indeed :)

Applied, thanks Eric.

^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2016-06-09 20:29 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-06-09 14:45 [PATCH v3 net-next 0/5] net: better lockdep annotations Eric Dumazet
2016-06-09 14:45 ` [PATCH v3 net-next 1/5] net: sched: fix qdisc->running " Eric Dumazet
2016-06-09 14:45 ` [PATCH v3 net-next 2/5] net: add netdev_lockdep_set_classes() helper Eric Dumazet
2016-06-09 14:45 ` [PATCH v3 net-next 3/5] net: vrf: call netdev_lockdep_set_classes() Eric Dumazet
2016-06-09 15:05   ` David Ahern
2016-06-09 14:45 ` [PATCH v3 net-next 4/5] net: macvlan: " Eric Dumazet
2016-06-09 14:45 ` [PATCH v3 net-next 5/5] net: ipvlan: " Eric Dumazet
2016-06-09 20:29 ` [PATCH v3 net-next 0/5] net: better lockdep annotations David Miller

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).