bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2009-12-13 21:06:07 -07:00
|
|
|
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
|
|
|
|
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/if_vlan.h>
|
|
|
|
#include <net/ipv6.h>
|
|
|
|
#include <net/ndisc.h>
|
|
|
|
#include <net/addrconf.h>
|
2009-10-29 08:18:26 -06:00
|
|
|
#include <net/netns/generic.h>
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
#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;
|
|
|
|
|
|
|
|
if (!dev)
|
|
|
|
return;
|
|
|
|
|
|
|
|
idev = in6_dev_get(dev);
|
|
|
|
if (!idev)
|
|
|
|
return;
|
|
|
|
|
|
|
|
read_lock_bh(&idev->lock);
|
2010-03-17 14:31:13 -06:00
|
|
|
if (!list_empty(&idev->addr_list)) {
|
|
|
|
struct inet6_ifaddr *ifa
|
|
|
|
= list_first_entry(&idev->addr_list,
|
|
|
|
struct inet6_ifaddr, if_list);
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
ipv6_addr_copy(addr, &ifa->addr);
|
2010-03-17 14:31:13 -06:00
|
|
|
} else
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
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);
|
|
|
|
|
2008-12-10 00:09:22 -07:00
|
|
|
pr_debug("ipv6 na on slave %s: dest %pI6, src %pI6\n",
|
2009-12-13 21:06:07 -07:00
|
|
|
slave_dev->name, &mcaddr, daddr);
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
|
|
|
|
skb = ndisc_build_skb(slave_dev, &mcaddr, daddr, &icmp6h, daddr,
|
|
|
|
ND_OPT_TARGET_LL_ADDR);
|
|
|
|
|
|
|
|
if (!skb) {
|
2009-12-13 21:06:07 -07:00
|
|
|
pr_err("NA packet allocation failed\n");
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (vlan_id) {
|
2010-12-13 01:20:24 -07:00
|
|
|
/* The Ethernet header is not present yet, so it is
|
|
|
|
* too early to insert a VLAN tag. Force use of an
|
|
|
|
* out-of-line tag here and let dev_hard_start_xmit()
|
|
|
|
* insert it if the slave hardware can't.
|
|
|
|
*/
|
|
|
|
skb = __vlan_hwaccel_put_tag(skb, vlan_id);
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
if (!skb) {
|
2009-12-13 21:06:07 -07:00
|
|
|
pr_err("failed to insert VLAN tag\n");
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
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;
|
|
|
|
|
2009-12-13 21:06:07 -07:00
|
|
|
pr_debug("%s: bond %s slave %s\n", bond->dev->name,
|
|
|
|
__func__, slave ? slave->dev->name : "NULL");
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
|
|
|
|
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;
|
2009-10-29 08:18:26 -06:00
|
|
|
struct bond_net *bn = net_generic(dev_net(event_dev), bond_net_id);
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
|
2009-10-29 08:18:26 -06:00
|
|
|
list_for_each_entry(bond, &bn->dev_list, bond_list) {
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
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) {
|
2010-07-18 21:24:14 -06:00
|
|
|
if (!bond->vlgrp)
|
|
|
|
continue;
|
bonding: send IPv6 neighbor advertisement on failover
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.
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-11-04 18:51:14 -07:00
|
|
|
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);
|
|
|
|
}
|
|
|
|
|