[PATCH net V1 0/2] net/bonding: fixes related to neigh

[PATCH net V1 0/2] net/bonding: fixes related to neigh

[Or Gerlitz]

From: Shlomo Pongratz <shlomop@mellanox.com>

Dave, 

This small patch set fixes some issues we stepped on when preparing 
for some IPoIB changes. 

With the 1st patch, the kernel networking code of arp_netdev_event calls 
neigh_changeaddr which further flashes the boding neighbours in some more 
cases which weren't covered by commit 7d26bb103 "bonding: emit event when 
bonding changes MAC"

The 2nd patch fixes a bug where under bonding, the IPoIB neigh cleanup
function wasn't called.

Or.

changes from V0: applied feedback from Jay Vosburgh
  - emit the address change in bond_release but not in bond_release_all
    where its unneeded.
  - rephrase the comment in bond_neigh_init on why we need to pre-proxy 
    the cleanup function. Also removed a comment which isn't needed there.
  

Shlomo Pongratz (2):
  net/bonding: emit address change event also in bond_release
  net/bonding: correctly proxy slave neigh param setup ndo function

 drivers/net/bonding/bond_main.c |   54 +++++++++++++++++++++++++++++++++------
 1 files changed, 46 insertions(+), 8 deletions(-)

[PATCH net V1 1/2] net/bonding: emit address change event also in bond_release

[Or Gerlitz]

From: Shlomo Pongratz <shlomop@mellanox.com>

commit 7d26bb103c4 "bonding: emit event when bonding changes MAC" didn't
take care to emit the NETDEV_CHANGEADDR event in bond_release, where bonding
actually changes the mac address (to all zeroes). As a result the neighbours
aren't deleted by the core networking code (which does so upon getting that
event).

Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>
---
 drivers/net/bonding/bond_main.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index a20b585..d38f635 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -2035,6 +2035,9 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev)
 	write_unlock_bh(&bond->lock);
 	unblock_netpoll_tx();
 
+	if (bond->slave_cnt == 0)
+		call_netdevice_notifiers(NETDEV_CHANGEADDR, bond->dev);
+
 	bond_compute_features(bond);
 	if (!(bond_dev->features & NETIF_F_VLAN_CHALLENGED) &&
 	    (old_features & NETIF_F_VLAN_CHALLENGED))
-- 
1.7.1

Re: [PATCH net V1 1/2] net/bonding: emit address change event also in bond_release

[Jay Vosburgh]

Or Gerlitz <ogerlitz@mellanox.com> wrote:

>From: Shlomo Pongratz <shlomop@mellanox.com>
>
>commit 7d26bb103c4 "bonding: emit event when bonding changes MAC" didn't
>take care to emit the NETDEV_CHANGEADDR event in bond_release, where bonding
>actually changes the mac address (to all zeroes). As a result the neighbours
>aren't deleted by the core networking code (which does so upon getting that
>event).
>
>Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>

Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>

>---
> drivers/net/bonding/bond_main.c |    3 +++
> 1 files changed, 3 insertions(+), 0 deletions(-)
>
>diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
>index a20b585..d38f635 100644
>--- a/drivers/net/bonding/bond_main.c
>+++ b/drivers/net/bonding/bond_main.c
>@@ -2035,6 +2035,9 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev)
> 	write_unlock_bh(&bond->lock);
> 	unblock_netpoll_tx();
>
>+	if (bond->slave_cnt == 0)
>+		call_netdevice_notifiers(NETDEV_CHANGEADDR, bond->dev);
>+
> 	bond_compute_features(bond);
> 	if (!(bond_dev->features & NETIF_F_VLAN_CHALLENGED) &&
> 	    (old_features & NETIF_F_VLAN_CHALLENGED))
>-- 
>1.7.1
>

Re: [PATCH net V1 1/2] net/bonding: emit address change event also in bond_release

[David Miller]

From: Or Gerlitz <ogerlitz@mellanox.com>
Date: Wed,  4 Apr 2012 11:56:19 +0300

> From: Shlomo Pongratz <shlomop@mellanox.com>
> 
> commit 7d26bb103c4 "bonding: emit event when bonding changes MAC" didn't
> take care to emit the NETDEV_CHANGEADDR event in bond_release, where bonding
> actually changes the mac address (to all zeroes). As a result the neighbours
> aren't deleted by the core networking code (which does so upon getting that
> event).
> 
> Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>

Applied.

[PATCH net V1 2/2] net/bonding: correctly proxy slave neigh param setup ndo function

[Or Gerlitz]

From: Shlomo Pongratz <shlomop@mellanox.com>

The current implemenation was buggy for slaves who use ndo_neigh_setup,
since the networking stack invokes the bonding device ndo entry (from
neigh_params_alloc) before any devices are enslaved, and the bonding
driver can't further delegate the call at that point in time. As a
result when bonding IPoIB devices, the neigh_cleanup hasn't been called.

Fix that by deferring the actual call into the slave ndo_neigh_setup
from the time the bonding neigh_setup is called.

Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>
---
 drivers/net/bonding/bond_main.c |   51 ++++++++++++++++++++++++++++++++------
 1 files changed, 43 insertions(+), 8 deletions(-)

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index d38f635..c8dc1a5 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -3705,17 +3705,52 @@ static void bond_set_multicast_list(struct net_device *bond_dev)
 	read_unlock(&bond->lock);
 }
 
-static int bond_neigh_setup(struct net_device *dev, struct neigh_parms *parms)
+static int bond_neigh_init(struct neighbour *n)
 {
-	struct bonding *bond = netdev_priv(dev);
+	struct bonding *bond = netdev_priv(n->dev);
 	struct slave *slave = bond->first_slave;
+	const struct net_device_ops *slave_ops;
+	struct neigh_parms parms;
+	int ret;
+
+	if (!slave)
+		return 0;
+
+	slave_ops = slave->dev->netdev_ops;
+
+	if (!slave_ops->ndo_neigh_setup)
+		return 0;
+
+	parms.neigh_setup = NULL;
+	parms.neigh_cleanup = NULL;
+	ret = slave_ops->ndo_neigh_setup(slave->dev, &parms);
+	if (ret)
+		return ret;
+
+	/*
+	 * Assign slave's neigh_cleanup to neighbour in case cleanup is called
+	 * after the last slave has been detached.  Assumes that all slaves
+	 * utilize the same neigh_cleanup (true at this writing as only user
+	 * is ipoib).
+	 */
+	n->parms->neigh_cleanup = parms.neigh_cleanup;
+
+	if (!parms.neigh_setup)
+		return 0;
+
+	return parms.neigh_setup(n);
+}
+
+/*
+ * The bonding ndo_neigh_setup is called at init time beofre any
+ * slave exists. So we must declare proxy setup function which will
+ * be used at run time to resolve the actual slave neigh param setup.
+ */
+static int bond_neigh_setup(struct net_device *dev,
+			    struct neigh_parms *parms)
+{
+	parms->neigh_setup   = bond_neigh_init;
 
-	if (slave) {
-		const struct net_device_ops *slave_ops
-			= slave->dev->netdev_ops;
-		if (slave_ops->ndo_neigh_setup)
-			return slave_ops->ndo_neigh_setup(slave->dev, parms);
-	}
 	return 0;
 }
 
-- 
1.7.1

Re: [PATCH net V1 2/2] net/bonding: correctly proxy slave neigh param setup ndo function

[Jay Vosburgh]

Or Gerlitz <ogerlitz@mellanox.com> wrote:

>From: Shlomo Pongratz <shlomop@mellanox.com>
>
>The current implemenation was buggy for slaves who use ndo_neigh_setup,
>since the networking stack invokes the bonding device ndo entry (from
>neigh_params_alloc) before any devices are enslaved, and the bonding
>driver can't further delegate the call at that point in time. As a
>result when bonding IPoIB devices, the neigh_cleanup hasn't been called.
>
>Fix that by deferring the actual call into the slave ndo_neigh_setup
>from the time the bonding neigh_setup is called.
>
>Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>

Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>

>---
> drivers/net/bonding/bond_main.c |   51 ++++++++++++++++++++++++++++++++------
> 1 files changed, 43 insertions(+), 8 deletions(-)
>
>diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
>index d38f635..c8dc1a5 100644
>--- a/drivers/net/bonding/bond_main.c
>+++ b/drivers/net/bonding/bond_main.c
>@@ -3705,17 +3705,52 @@ static void bond_set_multicast_list(struct net_device *bond_dev)
> 	read_unlock(&bond->lock);
> }
>
>-static int bond_neigh_setup(struct net_device *dev, struct neigh_parms *parms)
>+static int bond_neigh_init(struct neighbour *n)
> {
>-	struct bonding *bond = netdev_priv(dev);
>+	struct bonding *bond = netdev_priv(n->dev);
> 	struct slave *slave = bond->first_slave;
>+	const struct net_device_ops *slave_ops;
>+	struct neigh_parms parms;
>+	int ret;
>+
>+	if (!slave)
>+		return 0;
>+
>+	slave_ops = slave->dev->netdev_ops;
>+
>+	if (!slave_ops->ndo_neigh_setup)
>+		return 0;
>+
>+	parms.neigh_setup = NULL;
>+	parms.neigh_cleanup = NULL;
>+	ret = slave_ops->ndo_neigh_setup(slave->dev, &parms);
>+	if (ret)
>+		return ret;
>+
>+	/*
>+	 * Assign slave's neigh_cleanup to neighbour in case cleanup is called
>+	 * after the last slave has been detached.  Assumes that all slaves
>+	 * utilize the same neigh_cleanup (true at this writing as only user
>+	 * is ipoib).
>+	 */
>+	n->parms->neigh_cleanup = parms.neigh_cleanup;
>+
>+	if (!parms.neigh_setup)
>+		return 0;
>+
>+	return parms.neigh_setup(n);
>+}
>+
>+/*
>+ * The bonding ndo_neigh_setup is called at init time beofre any
>+ * slave exists. So we must declare proxy setup function which will
>+ * be used at run time to resolve the actual slave neigh param setup.
>+ */
>+static int bond_neigh_setup(struct net_device *dev,
>+			    struct neigh_parms *parms)
>+{
>+	parms->neigh_setup   = bond_neigh_init;
>
>-	if (slave) {
>-		const struct net_device_ops *slave_ops
>-			= slave->dev->netdev_ops;
>-		if (slave_ops->ndo_neigh_setup)
>-			return slave_ops->ndo_neigh_setup(slave->dev, parms);
>-	}
> 	return 0;
> }
>
>-- 
>1.7.1
>
>--
>To unsubscribe from this list: send the line "unsubscribe netdev" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH net V1 2/2] net/bonding: correctly proxy slave neigh param setup ndo function

[David Miller]

From: Or Gerlitz <ogerlitz@mellanox.com>
Date: Wed,  4 Apr 2012 11:56:20 +0300

> From: Shlomo Pongratz <shlomop@mellanox.com>
> 
> The current implemenation was buggy for slaves who use ndo_neigh_setup,
> since the networking stack invokes the bonding device ndo entry (from
> neigh_params_alloc) before any devices are enslaved, and the bonding
> driver can't further delegate the call at that point in time. As a
> result when bonding IPoIB devices, the neigh_cleanup hasn't been called.
> 
> Fix that by deferring the actual call into the slave ndo_neigh_setup
> from the time the bonding neigh_setup is called.
> 
> Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>

Applied.