Patchwork [v2] bonding: send IPv6 neighbor advertisement on failover

login
register
mail settings
Submitter Brian Haley
Date Oct. 10, 2008, 12:52 a.m.
Message ID <48EEA752.6030505@hp.com>
Download mbox | patch
Permalink /patch/3668/
State Changes Requested
Delegated to: David Miller
Headers show

Comments

Brian Haley - Oct. 10, 2008, 12:52 a.m.
Updated to address Vlad's comment about storing the link-local IPv6 
address in the bond/vlan structure, I no longer overwrite the existing 
address with a newer one.  Also added missed locking for the inet6_dev.

I don't see any way to address David Steven's comment since there is 
currently no NA tunable in the IPv6 code.

---

This patch adds better IPv6 failover support for bonding devices,
especially when in active-backup mode and there are only IPv6 addresses
configured, as reported by Alex Sidorenko.

- Creates a new file, net/drivers/bonding/bond_ipv6.c, for the
   IPv6-specific routines.  Both regular bonds and VLANs over bonds
   are supported.

- Adds a new tunable, num_unsol_na, to limit the number of unsolicited
   IPv6 Neighbor Advertisements that are sent on a failover event.
   Default is 1.

- Creates two new IPv6 neighbor discovery functions:

   ndisc_build_skb()
   ndisc_send_skb()

   These were required to support VLANs since we have to be able to
   add the VLAN id to the skb since ndisc_send_na() and friends
   shouldn't be asked to do this.  These two routines are basically
   __ndisc_send() split into two pieces, in a slightly different order.

- Updates Documentation/networking/bonding.txt and bumps the rev of bond
   support to 3.4.0.

On failover, this new code will generate one packet:

- An unsolicited IPv6 Neighbor Advertisement, which helps the switch
   learn that the address has moved to the new slave.

Testing has shown that sending just the NA results in pretty good
behavior when in active-back mode, I saw no lost ping packets for example.

-Brian

Signed-off-by: Brian Haley <brian.haley@hp.com>
---
David Stevens - Oct. 10, 2008, 2:23 a.m.
> I don't see any way to address David Steven's comment since there is 
> currently no NA tunable in the IPv6 code.

Brian,
        How do you feel about doubling up with dad_transmits for that
part?
        At least the switch/cache updating portion is the same
in both cases.

                                                        +-DLS

--
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
Brian Haley - Oct. 10, 2008, 2:34 p.m.
David Stevens wrote:
>> I don't see any way to address David Steven's comment since there is 
>> currently no NA tunable in the IPv6 code.
> 
> Brian,
>         How do you feel about doubling up with dad_transmits for that
> part?

I don't really want to since this is bonding-specific behavior, and 
we're not performing DAD for the address.  This is just another sysfs 
entry: /sys/class/net/bond*/bonding/num_unsol_na, not a sysctl.

>         At least the switch/cache updating portion is the same
> in both cases.

I don't think they are.  In the DAD case we're probing for another node 
that might have the address configured, in the bond failover case we 
want to update the switch quickly so we don't drop packets.

-Brian
--
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
David Stevens - Oct. 10, 2008, 3:03 p.m.
Brian Haley <brian.haley@hp.com> wrote on 10/10/2008 07:34:58 AM:

> I don't really want to since this is bonding-specific behavior, and 
> we're not performing DAD for the address.  This is just another sysfs 
> entry: /sys/class/net/bond*/bonding/num_unsol_na, not a sysctl.

        I think they really are the same case, and doing DAD
would solve the problem just as well. But I can hold my nose a
little bit and live with it. :-) Getting the problem solved is
more important than the details.

                                                        +-DLS

--
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
Vlad Yasevich - Oct. 10, 2008, 3:27 p.m.
Brian Haley wrote:
> Updated to address Vlad's comment about storing the link-local IPv6
> address in the bond/vlan structure, I no longer overwrite the existing
> address with a newer one.  Also added missed locking for the inet6_dev.
> 
> I don't see any way to address David Steven's comment since there is
> currently no NA tunable in the IPv6 code.
> 
> ---
> 
> This patch adds better IPv6 failover support for bonding devices,
> especially when in active-backup mode and there are only IPv6 addresses
> configured, as reported by Alex Sidorenko.
> 
> - Creates a new file, net/drivers/bonding/bond_ipv6.c, for the
>   IPv6-specific routines.  Both regular bonds and VLANs over bonds
>   are supported.
> 
> - Adds a new tunable, num_unsol_na, to limit the number of unsolicited
>   IPv6 Neighbor Advertisements that are sent on a failover event.
>   Default is 1.
> 
> - Creates two new IPv6 neighbor discovery functions:
> 
>   ndisc_build_skb()
>   ndisc_send_skb()
> 
>   These were required to support VLANs since we have to be able to
>   add the VLAN id to the skb since ndisc_send_na() and friends
>   shouldn't be asked to do this.  These two routines are basically
>   __ndisc_send() split into two pieces, in a slightly different order.
> 
> - Updates Documentation/networking/bonding.txt and bumps the rev of bond
>   support to 3.4.0.
> 
> On failover, this new code will generate one packet:
> 
> - An unsolicited IPv6 Neighbor Advertisement, which helps the switch
>   learn that the address has moved to the new slave.
> 
> Testing has shown that sending just the NA results in pretty good
> behavior when in active-back mode, I saw no lost ping packets for example.
> 
> -Brian
> 
> Signed-off-by: Brian Haley <brian.haley@hp.com>
> ---
> 

Looks good to me.  This provides a nice base that we can build on.

-vlad
--
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
Jay Vosburgh - Oct. 10, 2008, 3:53 p.m.
David Stevens <dlstevens@us.ibm.com> wrote:

>Brian Haley <brian.haley@hp.com> wrote on 10/10/2008 07:34:58 AM:
>
>> I don't really want to since this is bonding-specific behavior, and 
>> we're not performing DAD for the address.  This is just another sysfs 
>> entry: /sys/class/net/bond*/bonding/num_unsol_na, not a sysctl.
>
>        I think they really are the same case, and doing DAD
>would solve the problem just as well. But I can hold my nose a
>little bit and live with it. :-) Getting the problem solved is
>more important than the details.

	If I'm reading things correctly, DAD sends neighbor
solicitations, and we're sending neighbor advertisements, and not
running the DAD logic.

	As a semi-related question, what does IPv6 do if it receives a
gratutitous NA, and finds a duplicate?

	I agree that doing DAD would update the switches, peers, etc,
but, if I'm reading the IPv6 code correctly, the delay between probes is
one second (nd_tbl.retrans_time), and it looks like there's an initial
delay of up to 1 second as well (in addrconf_dad_kick, the
rtr_solicit_delay).  For failover purposes, we want to issue the
gratuitous ARP or NA packets immediately with a minimal delay between
probes.

	Is my understanding of the DAD behavior correct?

	-J

---
	-Jay Vosburgh, IBM Linux Technology Center, fubar@us.ibm.com
--
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
Brian Haley - Oct. 10, 2008, 4:04 p.m.
Jay Vosburgh wrote:
> 	As a semi-related question, what does IPv6 do if it receives a
> gratutitous NA, and finds a duplicate?

If a node has an IPv6 neighbor entry and receives an unsolicited NA it 
will change it's state to stale, forcing a re-lookup on the next 
transmit.  An un-solicited NA will change the state to reachable.

> 	I agree that doing DAD would update the switches, peers, etc,
> but, if I'm reading the IPv6 code correctly, the delay between probes is
> one second (nd_tbl.retrans_time), and it looks like there's an initial
> delay of up to 1 second as well (in addrconf_dad_kick, the
> rtr_solicit_delay).  For failover purposes, we want to issue the
> gratuitous ARP or NA packets immediately with a minimal delay between
> probes.

Right, and since a bond failover event should be rare, I think sending 
the NA immediately is OK.  All those random delays are meant to cover 
the case, for example, when a router sends an advertisement with a new 
prefix - you don't want everyone that receives it to do DAD immediately 
since it could overwhelm the network.

-Brian
--
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
Vlad Yasevich - Oct. 10, 2008, 4:29 p.m.
Brian Haley wrote:
> Jay Vosburgh wrote:
>>     As a semi-related question, what does IPv6 do if it receives a
>> gratutitous NA, and finds a duplicate?
> 
> If a node has an IPv6 neighbor entry and receives an unsolicited NA it
> will change it's state to stale, forcing a re-lookup on the next
> transmit.  An un-solicited NA will change the state to reachable.
                ^^^^^^^^^^^^

You probably meant to say "solicited".  Unsolicited NAs can only change
the state to STALE.

Also, the re-lookup will happen on a delay after the transmit.

-vlad

> 
>>     I agree that doing DAD would update the switches, peers, etc,
>> but, if I'm reading the IPv6 code correctly, the delay between probes is
>> one second (nd_tbl.retrans_time), and it looks like there's an initial
>> delay of up to 1 second as well (in addrconf_dad_kick, the
>> rtr_solicit_delay).  For failover purposes, we want to issue the
>> gratuitous ARP or NA packets immediately with a minimal delay between
>> probes.
> 
> Right, and since a bond failover event should be rare, I think sending
> the NA immediately is OK.  All those random delays are meant to cover
> the case, for example, when a router sends an advertisement with a new
> prefix - you don't want everyone that receives it to do DAD immediately
> since it could overwhelm the network.
> 
> -Brian
> -- 
> 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
> 

--
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
Sridhar Samudrala - Oct. 10, 2008, 4:56 p.m.
On Fri, 2008-10-10 at 12:29 -0400, Vlad Yasevich wrote:
> Brian Haley wrote:
> > Jay Vosburgh wrote:
> >>     As a semi-related question, what does IPv6 do if it receives a
> >> gratutitous NA, and finds a duplicate?

I think Jay was asking about the case where the target address in the NA 
matches the address of the receiving interface. The RFCs don't describe
how to handle such a case and leave it to the implementation. Linux logs
a warning message and ignores such NA.

Thanks
Sridhar

> > 
> > If a node has an IPv6 neighbor entry and receives an unsolicited NA it
> > will change it's state to stale, forcing a re-lookup on the next
> > transmit.  An un-solicited NA will change the state to reachable.
>                 ^^^^^^^^^^^^
> 
> You probably meant to say "solicited".  Unsolicited NAs can only change
> the state to STALE.
> 
> Also, the re-lookup will happen on a delay after the transmit.
> 
> -vlad


--
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
Vlad Yasevich - Oct. 10, 2008, 5:15 p.m.
Sridhar Samudrala wrote:
> On Fri, 2008-10-10 at 12:29 -0400, Vlad Yasevich wrote:
>> Brian Haley wrote:
>>> Jay Vosburgh wrote:
>>>>     As a semi-related question, what does IPv6 do if it receives a
>>>> gratutitous NA, and finds a duplicate?
> 
> I think Jay was asking about the case where the target address in the NA 
> matches the address of the receiving interface. The RFCs don't describe
> how to handle such a case and leave it to the implementation. Linux logs
> a warning message and ignores such NA.
> 

Yes, but in this case, you have duplicate addresses configured.  This
can happen when subnets merge and isn't really related to bonding driver.

In such a case, if we do an NS, it will trigger an NA and we'll end up
logging such a warning.  If we do an NA, the other end, if it's linux, will
log this warning, or deal with it in its own manner.

So, it's really a draw.

-vlad

> Thanks
> Sridhar
> 
>>> If a node has an IPv6 neighbor entry and receives an unsolicited NA it
>>> will change it's state to stale, forcing a re-lookup on the next
>>> transmit.  An un-solicited NA will change the state to reachable.
>>                 ^^^^^^^^^^^^
>>
>> You probably meant to say "solicited".  Unsolicited NAs can only change
>> the state to STALE.
>>
>> Also, the re-lookup will happen on a delay after the transmit.
>>
>> -vlad
> 
> 

--
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
Brian Haley - Oct. 27, 2008, 8:07 p.m.
Hi Jay,

Did you ever get a chance to look at this patch?  Now that net-next is 
open I'd like to resubmit it.

Thanks,

-Brian

Brian Haley wrote:
> Updated to address Vlad's comment about storing the link-local IPv6 
> address in the bond/vlan structure, I no longer overwrite the existing 
> address with a newer one.  Also added missed locking for the inet6_dev.
> 
> I don't see any way to address David Steven's comment since there is 
> currently no NA tunable in the IPv6 code.
> 
> ---
> 
> This patch adds better IPv6 failover support for bonding devices,
> especially when in active-backup mode and there are only IPv6 addresses
> configured, as reported by Alex Sidorenko.
> 
> - Creates a new file, net/drivers/bonding/bond_ipv6.c, for the
>   IPv6-specific routines.  Both regular bonds and VLANs over bonds
>   are supported.
> 
> - Adds a new tunable, num_unsol_na, to limit the number of unsolicited
>   IPv6 Neighbor Advertisements that are sent on a failover event.
>   Default is 1.
> 
> - Creates two new IPv6 neighbor discovery functions:
> 
>   ndisc_build_skb()
>   ndisc_send_skb()
> 
>   These were required to support VLANs since we have to be able to
>   add the VLAN id to the skb since ndisc_send_na() and friends
>   shouldn't be asked to do this.  These two routines are basically
>   __ndisc_send() split into two pieces, in a slightly different order.
> 
> - Updates Documentation/networking/bonding.txt and bumps the rev of bond
>   support to 3.4.0.
> 
> On failover, this new code will generate one packet:
> 
> - An unsolicited IPv6 Neighbor Advertisement, which helps the switch
>   learn that the address has moved to the new slave.
> 
> Testing has shown that sending just the NA results in pretty good
> behavior when in active-back mode, I saw no lost ping packets for example.
> 
> -Brian
> 
> Signed-off-by: Brian Haley <brian.haley@hp.com>
> ---
> 
--
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
Jay Vosburgh - Oct. 28, 2008, 12:24 a.m.
Brian Haley <brian.haley@hp.com> wrote:

>Hi Jay,
>
>Did you ever get a chance to look at this patch?  Now that net-next is
>open I'd like to resubmit it.

	I'm good with it; I'm merging it with a couple of other feature
things and I'll post it with that set in a day or two when things open
(unless I missed a mail, I don't think net-next is open as I type this).

	-J

---
	-Jay Vosburgh, IBM Linux Technology Center, fubar@us.ibm.com


>Thanks,
>
>-Brian
>
>Brian Haley wrote:
>> Updated to address Vlad's comment about storing the link-local IPv6
>> address in the bond/vlan structure, I no longer overwrite the existing
>> address with a newer one.  Also added missed locking for the inet6_dev.
>>
>> I don't see any way to address David Steven's comment since there is
>> currently no NA tunable in the IPv6 code.
>>
>> ---
>>
>> This patch adds better IPv6 failover support for bonding devices,
>> especially when in active-backup mode and there are only IPv6 addresses
>> configured, as reported by Alex Sidorenko.
>>
>> - Creates a new file, net/drivers/bonding/bond_ipv6.c, for the
>>   IPv6-specific routines.  Both regular bonds and VLANs over bonds
>>   are supported.
>>
>> - Adds a new tunable, num_unsol_na, to limit the number of unsolicited
>>   IPv6 Neighbor Advertisements that are sent on a failover event.
>>   Default is 1.
>>
>> - Creates two new IPv6 neighbor discovery functions:
>>
>>   ndisc_build_skb()
>>   ndisc_send_skb()
>>
>>   These were required to support VLANs since we have to be able to
>>   add the VLAN id to the skb since ndisc_send_na() and friends
>>   shouldn't be asked to do this.  These two routines are basically
>>   __ndisc_send() split into two pieces, in a slightly different order.
>>
>> - Updates Documentation/networking/bonding.txt and bumps the rev of bond
>>   support to 3.4.0.
>>
>> On failover, this new code will generate one packet:
>>
>> - An unsolicited IPv6 Neighbor Advertisement, which helps the switch
>>   learn that the address has moved to the new slave.
>>
>> Testing has shown that sending just the NA results in pretty good
>> behavior when in active-back mode, I saw no lost ping packets for example.
>>
>> -Brian
>>
>> Signed-off-by: Brian Haley <brian.haley@hp.com>
>> ---
>>
>--
>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
--
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

Patch

diff --git a/Documentation/networking/bonding.txt b/Documentation/networking/bonding.txt
index 688dfe1..4025565 100644
--- a/Documentation/networking/bonding.txt
+++ b/Documentation/networking/bonding.txt
@@ -551,6 +551,16 @@  num_grat_arp
 	affects only the active-backup mode.  This option was added for
 	bonding version 3.3.0.
 
+num_unsol_na
+
+	Specifies the number of unsolicited IPv6 Neighbor Advertisements
+	to be issued after a failover event.  One unsolicited NA is issued
+	immediately after the failover.
+
+	The valid range is 0 - 255; the default value is 1.  This option
+	affects only the active-backup mode.  This option was added for
+	bonding version 3.4.0.
+
 primary
 
 	A string (eth0, eth2, etc) specifying which slave is the
diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 99ae44e..0f97249 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -61,6 +61,7 @@  config DUMMY
 config BONDING
 	tristate "Bonding driver support"
 	depends on INET
+	depends on IPV6 || IPV6=n
 	---help---
 	  Say 'Y' or 'M' if you wish to be able to 'bond' multiple Ethernet
 	  Channels together. This is called 'Etherchannel' by Cisco,
diff --git a/drivers/net/bonding/Makefile b/drivers/net/bonding/Makefile
index 5cdae2b..6f9c6fa 100644
--- a/drivers/net/bonding/Makefile
+++ b/drivers/net/bonding/Makefile
@@ -6,3 +6,6 @@  obj-$(CONFIG_BONDING) += bonding.o
 
 bonding-objs := bond_main.o bond_3ad.o bond_alb.o bond_sysfs.o
 
+ipv6-$(subst m,y,$(CONFIG_IPV6)) += bond_ipv6.o
+bonding-objs += $(ipv6-y)
+
diff --git a/drivers/net/bonding/bond_ipv6.c b/drivers/net/bonding/bond_ipv6.c
new file mode 100644
index 0000000..6fd49af
--- /dev/null
+++ b/drivers/net/bonding/bond_ipv6.c
@@ -0,0 +1,218 @@ 
+/*
+ * Copyright(c) 2008 Hewlett-Packard Development Company, L.P.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+ * or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+ * for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
+ *
+ * The full GNU General Public License is included in this distribution in the
+ * file called LICENSE.
+ *
+ */
+
+//#define BONDING_DEBUG 1
+
+#include <linux/types.h>
+#include <linux/if_vlan.h>
+#include <net/ipv6.h>
+#include <net/ndisc.h>
+#include <net/addrconf.h>
+#include "bonding.h"
+
+/*
+ * Assign bond->master_ipv6 to the next IPv6 address in the list, or
+ * zero it out if there are none.
+ */
+static void bond_glean_dev_ipv6(struct net_device *dev, struct in6_addr *addr)
+{
+	struct inet6_dev *idev;
+	struct inet6_ifaddr *ifa;
+
+	if (!dev)
+		return;
+
+	idev = in6_dev_get(dev);
+	if (!idev)
+		return;
+
+	read_lock_bh(&idev->lock);
+	ifa = idev->addr_list;
+	if (ifa)
+		ipv6_addr_copy(addr, &ifa->addr);
+	else
+		ipv6_addr_set(addr, 0, 0, 0, 0);
+
+	read_unlock_bh(&idev->lock);
+
+	in6_dev_put(idev);
+}
+
+static void bond_na_send(struct net_device *slave_dev,
+			 struct in6_addr *daddr,
+			 int router,
+			 unsigned short vlan_id)
+{
+	struct in6_addr mcaddr;
+	struct icmp6hdr icmp6h = {
+		.icmp6_type = NDISC_NEIGHBOUR_ADVERTISEMENT,
+	};
+	struct sk_buff *skb;
+
+	icmp6h.icmp6_router = router;
+	icmp6h.icmp6_solicited = 0;
+	icmp6h.icmp6_override = 1;
+
+	addrconf_addr_solict_mult(daddr, &mcaddr);
+
+	dprintk("ipv6 na on slave %s: dest %s, src %s\n" NIP6_FMT NIP6_FMT,
+	       slave->name, NIP6(&mcaddr), NIP6(daddr));
+
+	skb = ndisc_build_skb(slave_dev, &mcaddr, daddr, &icmp6h, daddr,
+			      ND_OPT_TARGET_LL_ADDR);
+
+	if (!skb) {
+		printk(KERN_ERR DRV_NAME ": NA packet allocation failed\n");
+		return;
+	}
+
+	if (vlan_id) {
+		skb = vlan_put_tag(skb, vlan_id);
+		if (!skb) {
+			printk(KERN_ERR DRV_NAME ": failed to insert VLAN tag\n");
+			return;
+		}
+	}
+
+	ndisc_send_skb(skb, slave_dev, NULL, &mcaddr, daddr, &icmp6h);
+}
+
+/*
+ * Kick out an unsolicited Neighbor Advertisement for an IPv6 address on
+ * the bonding master.  This will help the switch learn our address
+ * if in active-backup mode.
+ *
+ * Caller must hold curr_slave_lock for read or better
+ */
+void bond_send_unsolicited_na(struct bonding *bond)
+{
+	struct slave *slave = bond->curr_active_slave;
+	struct vlan_entry *vlan;
+	struct inet6_dev *idev;
+	int is_router;
+
+	dprintk("bond_send_unsol_na: bond %s slave %s\n", bond->dev->name,
+				slave ? slave->dev->name : "NULL");
+
+	if (!slave || !bond->send_unsol_na ||
+	    test_bit(__LINK_STATE_LINKWATCH_PENDING, &slave->dev->state))
+		return;
+
+	bond->send_unsol_na--;
+
+	idev = in6_dev_get(bond->dev);
+	if (!idev)
+		return;
+
+	is_router = !!idev->cnf.forwarding;
+
+	in6_dev_put(idev);
+
+	if (!ipv6_addr_any(&bond->master_ipv6))
+		bond_na_send(slave->dev, &bond->master_ipv6, is_router, 0);
+
+	list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
+		if (!ipv6_addr_any(&vlan->vlan_ipv6)) {
+			bond_na_send(slave->dev, &vlan->vlan_ipv6, is_router,
+				     vlan->vlan_id);
+		}
+	}
+}
+
+/*
+ * bond_inet6addr_event: handle inet6addr notifier chain events.
+ *
+ * We keep track of device IPv6 addresses primarily to use as source
+ * addresses in NS probes.
+ *
+ * We track one IPv6 for the main device (if it has one).
+ */
+static int bond_inet6addr_event(struct notifier_block *this,
+				unsigned long event,
+				void *ptr)
+{
+	struct inet6_ifaddr *ifa = ptr;
+	struct net_device *vlan_dev, *event_dev = ifa->idev->dev;
+	struct bonding *bond;
+	struct vlan_entry *vlan;
+
+	if (dev_net(event_dev) != &init_net)
+		return NOTIFY_DONE;
+
+	list_for_each_entry(bond, &bond_dev_list, bond_list) {
+		if (bond->dev == event_dev) {
+			switch (event) {
+			case NETDEV_UP:
+				if (ipv6_addr_any(&bond->master_ipv6))
+					ipv6_addr_copy(&bond->master_ipv6,
+						       &ifa->addr);
+				return NOTIFY_OK;
+			case NETDEV_DOWN:
+				if (ipv6_addr_equal(&bond->master_ipv6,
+						    &ifa->addr))
+					bond_glean_dev_ipv6(bond->dev,
+							    &bond->master_ipv6);
+				return NOTIFY_OK;
+			default:
+				return NOTIFY_DONE;
+			}
+		}
+
+		list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
+			vlan_dev = vlan_group_get_device(bond->vlgrp,
+							 vlan->vlan_id);
+			if (vlan_dev == event_dev) {
+				switch (event) {
+				case NETDEV_UP:
+					if (ipv6_addr_any(&vlan->vlan_ipv6))
+						ipv6_addr_copy(&vlan->vlan_ipv6,
+							       &ifa->addr);
+					return NOTIFY_OK;
+				case NETDEV_DOWN:
+					if (ipv6_addr_equal(&vlan->vlan_ipv6,
+							    &ifa->addr))
+						bond_glean_dev_ipv6(vlan_dev,
+								    &vlan->vlan_ipv6);
+					return NOTIFY_OK;
+				default:
+					return NOTIFY_DONE;
+				}
+			}
+		}
+	}
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block bond_inet6addr_notifier = {
+	.notifier_call = bond_inet6addr_event,
+};
+
+void bond_register_ipv6_notifier(void)
+{
+	register_inet6addr_notifier(&bond_inet6addr_notifier);
+}
+
+void bond_unregister_ipv6_notifier(void)
+{
+	unregister_inet6addr_notifier(&bond_inet6addr_notifier);
+}
+
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 8e2be24..fba0f0c 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -89,6 +89,7 @@ 
 
 static int max_bonds	= BOND_DEFAULT_MAX_BONDS;
 static int num_grat_arp = 1;
+static int num_unsol_na = 1;
 static int miimon	= BOND_LINK_MON_INTERV;
 static int updelay	= 0;
 static int downdelay	= 0;
@@ -107,6 +108,8 @@  module_param(max_bonds, int, 0);
 MODULE_PARM_DESC(max_bonds, "Max number of bonded devices");
 module_param(num_grat_arp, int, 0644);
 MODULE_PARM_DESC(num_grat_arp, "Number of gratuitous ARP packets to send on failover event");
+module_param(num_unsol_na, int, 0644);
+MODULE_PARM_DESC(num_unsol_na, "Number of unsolicited IPv6 Neighbor Advertisements packets to send on failover event");
 module_param(miimon, int, 0);
 MODULE_PARM_DESC(miimon, "Link check interval in milliseconds");
 module_param(updelay, int, 0);
@@ -242,14 +245,13 @@  static int bond_add_vlan(struct bonding *bond, unsigned short vlan_id)
 	dprintk("bond: %s, vlan id %d\n",
 		(bond ? bond->dev->name: "None"), vlan_id);
 
-	vlan = kmalloc(sizeof(struct vlan_entry), GFP_KERNEL);
+	vlan = kzalloc(sizeof(struct vlan_entry), GFP_KERNEL);
 	if (!vlan) {
 		return -ENOMEM;
 	}
 
 	INIT_LIST_HEAD(&vlan->vlan_list);
 	vlan->vlan_id = vlan_id;
-	vlan->vlan_ip = 0;
 
 	write_lock_bh(&bond->lock);
 
@@ -1208,6 +1210,9 @@  void bond_change_active_slave(struct bonding *bond, struct slave *new_active)
 			bond->send_grat_arp = bond->params.num_grat_arp;
 			bond_send_gratuitous_arp(bond);
 
+			bond->send_unsol_na = bond->params.num_unsol_na;
+			bond_send_unsolicited_na(bond);
+
 			write_unlock_bh(&bond->curr_slave_lock);
 			read_unlock(&bond->lock);
 
@@ -2441,6 +2446,12 @@  void bond_mii_monitor(struct work_struct *work)
 		read_unlock(&bond->curr_slave_lock);
 	}
 
+	if (bond->send_unsol_na) {
+		read_lock(&bond->curr_slave_lock);
+		bond_send_unsolicited_na(bond);
+		read_unlock(&bond->curr_slave_lock);
+	}
+
 	if (bond_miimon_inspect(bond)) {
 		read_unlock(&bond->lock);
 		rtnl_lock();
@@ -3138,6 +3149,12 @@  void bond_activebackup_arp_mon(struct work_struct *work)
 		read_unlock(&bond->curr_slave_lock);
 	}
 
+	if (bond->send_unsol_na) {
+		read_lock(&bond->curr_slave_lock);
+		bond_send_unsolicited_na(bond);
+		read_unlock(&bond->curr_slave_lock);
+	}
+
 	if (bond_ab_arp_inspect(bond, delta_in_ticks)) {
 		read_unlock(&bond->lock);
 		rtnl_lock();
@@ -3813,6 +3830,7 @@  static int bond_close(struct net_device *bond_dev)
 	write_lock_bh(&bond->lock);
 
 	bond->send_grat_arp = 0;
+	bond->send_unsol_na = 0;
 
 	/* signal timers not to re-arm */
 	bond->kill_timers = 1;
@@ -4528,6 +4546,7 @@  static int bond_init(struct net_device *bond_dev, struct bond_params *params)
 	bond->primary_slave = NULL;
 	bond->dev = bond_dev;
 	bond->send_grat_arp = 0;
+	bond->send_unsol_na = 0;
 	bond->setup_by_slave = 0;
 	INIT_LIST_HEAD(&bond->vlan_list);
 
@@ -4776,6 +4795,13 @@  static int bond_check_params(struct bond_params *params)
 		num_grat_arp = 1;
 	}
 
+	if (num_unsol_na < 0 || num_unsol_na > 255) {
+		printk(KERN_WARNING DRV_NAME
+		       ": Warning: num_unsol_na (%d) not in range 0-255 so it "
+		       "was reset to 1 \n", num_unsol_na);
+		num_unsol_na = 1;
+	}
+
 	/* reset values for 802.3ad */
 	if (bond_mode == BOND_MODE_8023AD) {
 		if (!miimon) {
@@ -4977,6 +5003,7 @@  static int bond_check_params(struct bond_params *params)
 	params->xmit_policy = xmit_hashtype;
 	params->miimon = miimon;
 	params->num_grat_arp = num_grat_arp;
+	params->num_unsol_na = num_unsol_na;
 	params->arp_interval = arp_interval;
 	params->arp_validate = arp_validate_value;
 	params->updelay = updelay;
@@ -5129,6 +5156,7 @@  static int __init bonding_init(void)
 
 	register_netdevice_notifier(&bond_netdev_notifier);
 	register_inetaddr_notifier(&bond_inetaddr_notifier);
+	bond_register_ipv6_notifier();
 
 	goto out;
 err:
@@ -5151,6 +5179,7 @@  static void __exit bonding_exit(void)
 {
 	unregister_netdevice_notifier(&bond_netdev_notifier);
 	unregister_inetaddr_notifier(&bond_inetaddr_notifier);
+	bond_unregister_ipv6_notifier();
 
 	bond_destroy_sysfs();
 
diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
index 3bdb473..70ae376 100644
--- a/drivers/net/bonding/bond_sysfs.c
+++ b/drivers/net/bonding/bond_sysfs.c
@@ -981,6 +981,47 @@  out:
 	return ret;
 }
 static DEVICE_ATTR(num_grat_arp, S_IRUGO | S_IWUSR, bonding_show_n_grat_arp, bonding_store_n_grat_arp);
+
+/*
+ * Show and set the number of unsolicted NA's to send after a failover event.
+ */
+static ssize_t bonding_show_n_unsol_na(struct device *d,
+				       struct device_attribute *attr,
+				       char *buf)
+{
+	struct bonding *bond = to_bond(d);
+
+	return sprintf(buf, "%d\n", bond->params.num_unsol_na);
+}
+
+static ssize_t bonding_store_n_unsol_na(struct device *d,
+					struct device_attribute *attr,
+					const char *buf, size_t count)
+{
+	int new_value, ret = count;
+	struct bonding *bond = to_bond(d);
+
+	if (sscanf(buf, "%d", &new_value) != 1) {
+		printk(KERN_ERR DRV_NAME
+		       ": %s: no num_unsol_na value specified.\n",
+		       bond->dev->name);
+		ret = -EINVAL;
+		goto out;
+	}
+	if (new_value < 0 || new_value > 255) {
+		printk(KERN_ERR DRV_NAME
+		       ": %s: Invalid num_unsol_na value %d not in range 0-255; rejected.\n",
+		       bond->dev->name, new_value);
+		ret = -EINVAL;
+		goto out;
+	} else {
+		bond->params.num_unsol_na = new_value;
+	}
+out:
+	return ret;
+}
+static DEVICE_ATTR(num_unsol_na, S_IRUGO | S_IWUSR, bonding_show_n_unsol_na, bonding_store_n_unsol_na);
+
 /*
  * Show and set the MII monitor interval.  There are two tricky bits
  * here.  First, if MII monitoring is activated, then we must disable
@@ -1419,6 +1460,7 @@  static struct attribute *per_bond_attrs[] = {
 	&dev_attr_lacp_rate.attr,
 	&dev_attr_xmit_hash_policy.attr,
 	&dev_attr_num_grat_arp.attr,
+	&dev_attr_num_unsol_na.attr,
 	&dev_attr_miimon.attr,
 	&dev_attr_primary.attr,
 	&dev_attr_use_carrier.attr,
diff --git a/drivers/net/bonding/bonding.h b/drivers/net/bonding/bonding.h
index ffb668d..0491c7c 100644
--- a/drivers/net/bonding/bonding.h
+++ b/drivers/net/bonding/bonding.h
@@ -19,16 +19,19 @@ 
 #include <linux/proc_fs.h>
 #include <linux/if_bonding.h>
 #include <linux/kobject.h>
+#include <linux/in6.h>
 #include "bond_3ad.h"
 #include "bond_alb.h"
 
-#define DRV_VERSION	"3.3.0"
-#define DRV_RELDATE	"June 10, 2008"
+#define DRV_VERSION	"3.4.0"
+#define DRV_RELDATE	"October 7, 2008"
 #define DRV_NAME	"bonding"
 #define DRV_DESCRIPTION	"Ethernet Channel Bonding Driver"
 
 #define BOND_MAX_ARP_TARGETS	16
 
+extern struct list_head bond_dev_list;
+
 #ifdef BONDING_DEBUG
 #define dprintk(fmt, args...) \
 	printk(KERN_DEBUG     \
@@ -126,6 +129,7 @@  struct bond_params {
 	int xmit_policy;
 	int miimon;
 	int num_grat_arp;
+	int num_unsol_na;
 	int arp_interval;
 	int arp_validate;
 	int use_carrier;
@@ -148,6 +152,9 @@  struct vlan_entry {
 	struct list_head vlan_list;
 	__be32 vlan_ip;
 	unsigned short vlan_id;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	struct in6_addr vlan_ipv6;
+#endif
 };
 
 struct slave {
@@ -195,6 +202,7 @@  struct bonding {
 	rwlock_t curr_slave_lock;
 	s8       kill_timers;
 	s8	 send_grat_arp;
+	s8	 send_unsol_na;
 	s8	 setup_by_slave;
 	struct   net_device_stats stats;
 #ifdef CONFIG_PROC_FS
@@ -218,6 +226,9 @@  struct bonding {
 	struct   delayed_work arp_work;
 	struct   delayed_work alb_work;
 	struct   delayed_work ad_work;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	struct   in6_addr master_ipv6;
+#endif
 };
 
 /**
@@ -341,5 +352,24 @@  extern struct bond_parm_tbl xmit_hashtype_tbl[];
 extern struct bond_parm_tbl arp_validate_tbl[];
 extern struct bond_parm_tbl fail_over_mac_tbl[];
 
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+void bond_send_unsolicited_na(struct bonding *bond);
+void bond_register_ipv6_notifier(void);
+void bond_unregister_ipv6_notifier(void);
+#else
+static inline void bond_send_unsolicited_na(struct bonding *bond)
+{
+	return;
+}
+static inline void bond_register_ipv6_notifier(void)
+{
+	return;
+}
+static inline void bond_unregister_ipv6_notifier(void)
+{
+	return;
+}
+#endif
+
 #endif /* _LINUX_BONDING_H */
 
diff --git a/include/net/ndisc.h b/include/net/ndisc.h
index a01b7c4..1b0da3d 100644
--- a/include/net/ndisc.h
+++ b/include/net/ndisc.h
@@ -108,6 +108,20 @@  extern void			ndisc_send_redirect(struct sk_buff *skb,
 
 extern int			ndisc_mc_map(struct in6_addr *addr, char *buf, struct net_device *dev, int dir);
 
+extern struct sk_buff		*ndisc_build_skb(struct net_device *dev,
+						 const struct in6_addr *daddr,
+						 const struct in6_addr *saddr,
+						 struct icmp6hdr *icmp6h,
+						 const struct in6_addr *target,
+						 int llinfo);
+
+extern void			ndisc_send_skb(struct sk_buff *skb,
+					       struct net_device *dev,
+					       struct neighbour *neigh,
+					       const struct in6_addr *daddr,
+					       const struct in6_addr *saddr,
+					       struct icmp6hdr *icmp6h);
+
 
 
 /*
diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
index 840b157..afd76ac 100644
--- a/net/ipv6/ndisc.c
+++ b/net/ipv6/ndisc.c
@@ -437,38 +437,20 @@  static void pndisc_destructor(struct pneigh_entry *n)
 	ipv6_dev_mc_dec(dev, &maddr);
 }
 
-/*
- *	Send a Neighbour Advertisement
- */
-static void __ndisc_send(struct net_device *dev,
-			 struct neighbour *neigh,
-			 const struct in6_addr *daddr,
-			 const struct in6_addr *saddr,
-			 struct icmp6hdr *icmp6h, const struct in6_addr *target,
-			 int llinfo)
+struct sk_buff *ndisc_build_skb(struct net_device *dev,
+				const struct in6_addr *daddr,
+				const struct in6_addr *saddr,
+				struct icmp6hdr *icmp6h,
+				const struct in6_addr *target,
+				int llinfo)
 {
-	struct flowi fl;
-	struct dst_entry *dst;
 	struct net *net = dev_net(dev);
 	struct sock *sk = net->ipv6.ndisc_sk;
 	struct sk_buff *skb;
 	struct icmp6hdr *hdr;
-	struct inet6_dev *idev;
 	int len;
 	int err;
-	u8 *opt, type;
-
-	type = icmp6h->icmp6_type;
-
-	icmpv6_flow_init(sk, &fl, type, saddr, daddr, dev->ifindex);
-
-	dst = icmp6_dst_alloc(dev, neigh, daddr);
-	if (!dst)
-		return;
-
-	err = xfrm_lookup(&dst, &fl, NULL, 0);
-	if (err < 0)
-		return;
+	u8 *opt;
 
 	if (!dev->addr_len)
 		llinfo = 0;
@@ -485,8 +467,7 @@  static void __ndisc_send(struct net_device *dev,
 		ND_PRINTK0(KERN_ERR
 			   "ICMPv6 ND: %s() failed to allocate an skb.\n",
 			   __func__);
-		dst_release(dst);
-		return;
+		return NULL;
 	}
 
 	skb_reserve(skb, LL_RESERVED_SPACE(dev));
@@ -513,6 +494,42 @@  static void __ndisc_send(struct net_device *dev,
 					   csum_partial((__u8 *) hdr,
 							len, 0));
 
+	return skb;
+}
+
+EXPORT_SYMBOL(ndisc_build_skb);
+
+void ndisc_send_skb(struct sk_buff *skb,
+		    struct net_device *dev,
+		    struct neighbour *neigh,
+		    const struct in6_addr *daddr,
+		    const struct in6_addr *saddr,
+		    struct icmp6hdr *icmp6h)
+{
+	struct flowi fl;
+	struct dst_entry *dst;
+	struct net *net = dev_net(dev);
+	struct sock *sk = net->ipv6.ndisc_sk;
+	struct inet6_dev *idev;
+	int err;
+	u8 type;
+
+	type = icmp6h->icmp6_type;
+
+	icmpv6_flow_init(sk, &fl, type, saddr, daddr, dev->ifindex);
+
+	dst = icmp6_dst_alloc(dev, neigh, daddr);
+	if (!dst) {
+		kfree_skb(skb);
+		return;
+	}
+
+	err = xfrm_lookup(&dst, &fl, NULL, 0);
+	if (err < 0) {
+		kfree_skb(skb);
+		return;
+	}
+
 	skb->dst = dst;
 
 	idev = in6_dev_get(dst->dev);
@@ -529,6 +546,27 @@  static void __ndisc_send(struct net_device *dev,
 		in6_dev_put(idev);
 }
 
+EXPORT_SYMBOL(ndisc_send_skb);
+
+/*
+ *	Send a Neighbour Discover packet
+ */
+static void __ndisc_send(struct net_device *dev,
+			 struct neighbour *neigh,
+			 const struct in6_addr *daddr,
+			 const struct in6_addr *saddr,
+			 struct icmp6hdr *icmp6h, const struct in6_addr *target,
+			 int llinfo)
+{
+	struct sk_buff *skb;
+
+	skb = ndisc_build_skb(dev, daddr, saddr, icmp6h, target, llinfo);
+	if (!skb)
+		return;
+
+	ndisc_send_skb(skb, dev, neigh, daddr, saddr, icmp6h);
+}
+
 static void ndisc_send_na(struct net_device *dev, struct neighbour *neigh,
 			  const struct in6_addr *daddr,
 			  const struct in6_addr *solicited_addr,