linux-stable.git/drivers/net/bonding, branch linux-3.18.y

bonding: fix arp_validate toggling in active-backup mode

2019-05-16T07:17:24+00:00

[ Upstream commit a9b8a2b39ce65df45687cf9ef648885c2a99fe75 ]

There's currently a problem with toggling arp_validate on and off with an
active-backup bond. At the moment, you can start up a bond, like so:

modprobe bonding mode=1 arp_interval=100 arp_validate=0 arp_ip_targets=192.168.1.1
ip link set bond0 down
echo "ens4f0" > /sys/class/net/bond0/bonding/slaves
echo "ens4f1" > /sys/class/net/bond0/bonding/slaves
ip link set bond0 up
ip addr add 192.168.1.2/24 dev bond0

Pings to 192.168.1.1 work just fine. Now turn on arp_validate:

echo 1 > /sys/class/net/bond0/bonding/arp_validate

Pings to 192.168.1.1 continue to work just fine. Now when you go to turn
arp_validate off again, the link falls flat on it's face:

echo 0 > /sys/class/net/bond0/bonding/arp_validate
dmesg
...
[133191.911987] bond0: Setting arp_validate to none (0)
[133194.257793] bond0: bond_should_notify_peers: slave ens4f0
[133194.258031] bond0: link status definitely down for interface ens4f0, disabling it
[133194.259000] bond0: making interface ens4f1 the new active one
[133197.330130] bond0: link status definitely down for interface ens4f1, disabling it
[133197.331191] bond0: now running without any active interface!

The problem lies in bond_options.c, where passing in arp_validate=0
results in bond->recv_probe getting set to NULL. This flies directly in
the face of commit 3fe68df97c7f, which says we need to set recv_probe =
bond_arp_recv, even if we're not using arp_validate. Said commit fixed
this in bond_option_arp_interval_set, but missed that we can get to that
same state in bond_option_arp_validate_set as well.

One solution would be to universally set recv_probe = bond_arp_recv here
as well, but I don't think bond_option_arp_validate_set has any business
touching recv_probe at all, and that should be left to the arp_interval
code, so we can just make things much tidier here.

Fixes: 3fe68df97c7f ("bonding: always set recv_probe to bond_arp_rcv in arp monitor")
CC: Jay Vosburgh 
CC: Veaceslav Falico 
CC: Andy Gospodarek 
CC: "David S. Miller" 
CC: netdev@vger.kernel.org
Signed-off-by: Jarod Wilson 
Signed-off-by: Jay Vosburgh 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

bonding: show full hw address in sysfs for slave entries

2019-05-16T07:17:19+00:00

[ Upstream commit 18bebc6dd3281955240062655a4df35eef2c46b3 ]

Bond expects ethernet hwaddr for its slave, but it can be longer than 6
bytes - infiniband interface for example.

 # cat /sys/devices//net/ib0/address
 80:00:02:08:fe:80:00:00:00:00:00:00:7c:fe:90:03:00:be:5d:e1

 # cat /sys/devices//net/ib0/bonding_slave/perm_hwaddr
 80:00:02:08:fe:80

So print full hwaddr in sysfs "bonding_slave/perm_hwaddr" as well.

Signed-off-by: Konstantin Khorenko 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin

bonding: fix event handling for stacked bonds

2019-04-27T07:30:29+00:00

[ Upstream commit 92480b3977fd3884649d404cbbaf839b70035699 ]

When a bond is enslaved to another bond, bond_netdev_event() only
handles the event as if the bond is a master, and skips treating the
bond as a slave.

This leads to a refcount leak on the slave, since we don't remove the
adjacency to its master and the master holds a reference on the slave.

Reproducer:
  ip link add bondL type bond
  ip link add bondU type bond
  ip link set bondL master bondU
  ip link del bondL

No "Fixes:" tag, this code is older than git history.

Signed-off-by: Sabrina Dubroca 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

bonding: fix 802.3ad state sent to partner when unbinding slave

2018-12-21T13:08:48+00:00

[ Upstream commit 3b5b3a3331d141e8f2a7aaae3a94dfa1e61ecbe4 ]

Previously when unbinding a slave the 802.3ad implementation only told
partner that the port is not suitable for aggregation by setting the port
aggregation state from aggregatable to individual. This is not enough. If the
physical layer still stays up and we only unbinded this port from the bond there
is nothing in the aggregation status alone to prevent the partner from sending
traffic towards us. To ensure that the partner doesn't consider this
port at all anymore we should also disable collecting and distributing to
signal that this actor is going away. Also clear AD_STATE_SYNCHRONIZATION to
ensure partner exits collecting + distributing state.

I have tested this behaviour againts Arista EOS switches with mlx5 cards
(physical link stays up even when interface is down) and simulated
the same situation virtually Linux <-> Linux with two network namespaces
running two veth device pairs. In both cases setting aggregation to
individual doesn't alone prevent traffic from being to sent towards this
port given that the link stays up in partners end. Partner still keeps
it's end in collecting + distributing state and continues until timeout is
reached. In most cases this means we are losing the traffic partner sends
towards our port while we wait for timeout. This is most visible with slow
periodic time (LACP rate slow).

Other open source implementations like Open VSwitch and libreswitch, and
vendor implementations like Arista EOS, seem to disable collecting +
distributing to when doing similar port disabling/detaching/removing change.
With this patch kernel implementation would behave the same way and ensure
partner doesn't consider our actor viable anymore.

Signed-off-by: Toni Peltonen 
Signed-off-by: Jay Vosburgh 
Acked-by: Jonathan Toppins 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin

bonding: do not allow rlb updates to invalid mac

2018-05-25T08:54:52+00:00

[ Upstream commit 4fa8667ca3989ce14cf66301fa251544fbddbdd0 ]

Make sure multicast, broadcast, and zero mac's cannot be the output of rlb
updates, which should all be directed arps. Receive load balancing will be
collapsed if any of these happen, as the switch will broadcast.

Signed-off-by: Debabrata Banerjee 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

bonding: do not set slave_dev npinfo before slave_enable_netpoll in bond_enslave

2018-04-29T05:45:28+00:00

[ Upstream commit ddea788c63094f7c483783265563dd5b50052e28 ]

After Commit 8a8efa22f51b ("bonding: sync netpoll code with bridge"), it
would set slave_dev npinfo in slave_enable_netpoll when enslaving a dev
if bond->dev->npinfo was set.

However now slave_dev npinfo is set with bond->dev->npinfo before calling
slave_enable_netpoll. With slave_dev npinfo set, __netpoll_setup called
in slave_enable_netpoll will not call slave dev's .ndo_netpoll_setup().
It causes that the lower dev of this slave dev can't set its npinfo.

One way to reproduce it:

  # modprobe bonding
  # brctl addbr br0
  # brctl addif br0 eth1
  # ifconfig bond0 192.168.122.1/24 up
  # ifenslave bond0 eth2
  # systemctl restart netconsole
  # ifenslave bond0 br0
  # ifconfig eth2 down
  # systemctl restart netconsole

The netpoll won't really work.

This patch is to remove that slave_dev npinfo setting in bond_enslave().

Fixes: 8a8efa22f51b ("bonding: sync netpoll code with bridge")
Signed-off-by: Xin Long 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

bonding: process the err returned by dev_set_allmulti properly in bond_enslave

2018-04-13T17:52:24+00:00

[ Upstream commit 9f5a90c107741b864398f4ac0014711a8c1d8474 ]

When dev_set_promiscuity(1) succeeds but dev_set_allmulti(1) fails,
dev_set_promiscuity(-1) should be done before going to the err path.
Otherwise, dev->promiscuity will leak.

Fixes: 7e1a1ac1fbaa ("bonding: Check return of dev_set_promiscuity/allmulti")
Signed-off-by: Xin Long 
Acked-by: Andy Gospodarek 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

bonding: move dev_mc_sync after master_upper_dev_link in bond_enslave

2018-04-13T17:52:24+00:00

[ Upstream commit ae42cc62a9f07f1f6979054ed92606b9c30f4a2e ]

Beniamino found a crash when adding vlan as slave of bond which is also
the parent link:

  ip link add bond1 type bond
  ip link set bond1 up
  ip link add link bond1 vlan1 type vlan id 80
  ip link set vlan1 master bond1

The call trace is as below:

  [] queued_spin_lock_slowpath+0xb/0xf
  [] _raw_spin_lock+0x20/0x30
  [] dev_mc_sync+0x37/0x80
  [] vlan_dev_set_rx_mode+0x1c/0x30 [8021q]
  [] __dev_set_rx_mode+0x5a/0xa0
  [] dev_mc_sync_multiple+0x78/0x80
  [] bond_enslave+0x67c/0x1190 [bonding]
  [] do_setlink+0x9c9/0xe50
  [] rtnl_newlink+0x522/0x880
  [] rtnetlink_rcv_msg+0xa7/0x260
  [] netlink_rcv_skb+0xab/0xc0
  [] rtnetlink_rcv+0x28/0x30
  [] netlink_unicast+0x170/0x210
  [] netlink_sendmsg+0x308/0x420
  [] sock_sendmsg+0xb6/0xf0

This is actually a dead lock caused by sync slave hwaddr from master when
the master is the slave's 'slave'. This dead loop check is actually done
by netdev_master_upper_dev_link. However, Commit 1f718f0f4f97 ("bonding:
populate neighbour's private on enslave") moved it after dev_mc_sync.

This patch is to fix it by moving dev_mc_sync after master_upper_dev_link,
so that this loop check would be earlier than dev_mc_sync. It also moves
if (mode == BOND_MODE_8023AD) into if (!bond_uses_primary) clause as an
improvement.

Note team driver also has this issue, I will fix it in another patch.

Fixes: 1f718f0f4f97 ("bonding: populate neighbour's private on enslave")
Reported-by: Beniamino Galvani 
Signed-off-by: Xin Long 
Acked-by: Andy Gospodarek 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

bonding: fix the err path for dev hwaddr sync in bond_enslave

2018-04-13T17:52:24+00:00

[ Upstream commit 5c78f6bfae2b10ff70e21d343e64584ea6280c26 ]

vlan_vids_add_by_dev is called right after dev hwaddr sync, so on
the err path it should unsync dev hwaddr. Otherwise, the slave
dev's hwaddr will never be unsync when this err happens.

Fixes: 1ff412ad7714 ("bonding: change the bond's vlan syncing functions with the standard ones")
Signed-off-by: Xin Long 
Reviewed-by: Nikolay Aleksandrov 
Acked-by: Andy Gospodarek 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

bonding: Don't update slave->link until ready to commit

2018-04-13T17:52:15+00:00

[ Upstream commit 797a93647a48d6cb8a20641a86a71713a947f786 ]

In the loadbalance arp monitoring scheme, when a slave link change is
detected, the slave->link is immediately updated and slave_state_changed
is set. Later down the function, the rtnl_lock is acquired and the
changes are committed, updating the bond link state.

However, the acquisition of the rtnl_lock can fail. The next time the
monitor runs, since slave->link is already updated, it determines that
link is unchanged. This results in the bond link state permanently out
of sync with the slave link.

This patch modifies bond_loadbalance_arp_mon() to handle link changes
identical to bond_ab_arp_{inspect/commit}(). The new link state is
maintained in slave->new_link until we're ready to commit at which point
it's copied into slave->link.

NOTE: miimon_{inspect/commit}() has a more complex state machine
requiring the use of the bond_{propose,commit}_link_state() functions
which maintains the intermediate state in slave->link_new_state. The arp
monitors don't require that.

Testing: This bug is very easy to reproduce with the following steps.
1. In a loop, toggle a slave link of a bond slave interface.
2. In a separate loop, do ifconfig up/down of an unrelated interface to
create contention for rtnl_lock.
Within a few iterations, the bond link goes out of sync with the slave
link.

Signed-off-by: Nithin Nayak Sujir 
Cc: Mahesh Bandewar 
Cc: Jay Vosburgh 
Acked-by: Mahesh Bandewar 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman