When using an IPoIB bond currently only active-backup mode is a valid
use case and this commit strengthens it.

Since commit 2ab82852 ("net/bonding: Enable bonding to enslave
netdevices not supporting set_mac_address()") was introduced till
4.7-rc1, IPoIB didn't support the set_mac_address ndo, and hence the
fail over mac policy always applied to IPoIB bonds.

With the introduction of commit 492a7e67 ("IB/IPoIB: Allow setting
the device address"), that doesn't hold and practically IPoIB bonds are
broken as of that. To fix it, lets go to fail over mac if the device
doesn't support the ndo OR this is IPoIB device.

As a by-product, this commit also prevents a stack corruption which
occurred when trying to copy 20 bytes (IPoIB) device address
to a sockaddr struct that has only 16 bytes of storage.
Signed-off-by: NMark Bloch <markb@mellanox.com>
Signed-off-by: NOr Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Acked-by: NAndy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: NJay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit e826eafa ("bonding: Call netif_carrier_off after
register_netdevice") moved netif_carrier_off() from bond_init() to
bond_create(), but the latter is called only for initial default
devices and ones created through sysfs:

 $ modprobe bonding
 $ echo +bond1 > /sys/class/net/bonding_masters
 $ ip link add bond2 type bond
 $ grep "MII Status" /proc/net/bonding/*
 /proc/net/bonding/bond0:MII Status: down
 /proc/net/bonding/bond1:MII Status: down
 /proc/net/bonding/bond2:MII Status: up

Ensure that carrier is initially off also for devices created through
netlink.
Signed-off-by: NBeniamino Galvani <bgalvani@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently, link notifications are not sent by
bond_set_slave_link_state() upon enslavement if
the slave is enslaved when up.

This happens because slave->link default init value
is 0, which is the same as BOND_LINK_UP, resulting
in bond_set_slave_link_state() ignoring this transition.

This patch sets the default value of slave->link to
BOND_LINK_NOCHANGE, assuring it will count as a state
transition and thus trigger notification logic.
Signed-off-by: NAviv Heller <avivh@mellanox.com>
Reviewed-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

L2 upper device needs to propagate neigh_construct/destroy calls down to
lower devices. Do this by defining default ndo functions and use them in
team, bond, bridge and vlan.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

ether_addr_equal_64bits() requires some care about its arguments,
namely that 8 bytes might be read, even if last 2 byte values are not
used.

KASan detected a violation with null_mac_addr and lacpdu_mcast_addr
in bond_3ad.c

Same problem with mac_bcast[] and mac_v6_allmcast[] in bond_alb.c :
Although the 8-byte alignment was there, KASan would detect out
of bound accesses.

Fixes: 815117ad ("bonding: use ether_addr_equal_unaligned for bond addr compare")
Fixes: bb54e589 ("bonding: Verify RX LACPDU has proper dest mac-addr")
Fixes: 885a136c ("bonding: use compare_ether_addr_64bits() in ALB")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Reported-by: NDmitry Vyukov <dvyukov@google.com>
Acked-by: NDmitry Vyukov <dvyukov@google.com>
Acked-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Acked-by: NDing Tianhong <dingtianhong@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since commit 7bb11dc9 ("bonding: unify all places where
actor-oper key needs to be updated."), the logic in bonding to handle
selection between multiple aggregators has not functioned.

	This affects only configurations wherein the bonding slaves
connect to two discrete aggregators (e.g., two independent switches, each
with LACP enabled), thus creating two separate aggregation groups within a
single bond.

	The cause is a change in 7bb11dc9 to no longer set
AD_PORT_BEGIN on a port after a link state change, which would cause the
port to be reselected for attachment to an aggregator as if were newly
added to the bond.  We cannot restore the prior behavior, as it
contradicts IEEE 802.1AX 5.4.12, which requires ports that "become
inoperable" (lose carrier, setting port_enabled=false as per 802.1AX
5.4.7) to remain selected (i.e., assigned to the aggregator).  As the port
now remains selected, the aggregator selection logic is not invoked.

	A side effect of this change is that aggregators in bonding will
now contain ports that are link down.  The aggregator selection logic
does not currently handle this situation correctly, causing incorrect
aggregator selection.

	This patch makes two changes to repair the aggregator selection
logic in bonding to function as documented and within the confines of the
standard:

	First, the aggregator selection and related logic now utilizes the
number of active ports per aggregator, not the number of selected ports
(as some selected ports may be down).  The ad_select "bandwidth" and
"count" options only consider ports that are link up.

	Second, on any carrier state change of any slave, the aggregator
selection logic is explicitly called to insure the correct aggregator is
active.
Reported-by: NVeli-Matti Lintu <veli-matti.lintu@opinsys.fi>
Fixes: 7bb11dc9 ("bonding: unify all places where actor-oper key needs to be updated.")
Signed-off-by: NJay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

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: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Instead of using a single bit (__QDISC___STATE_RUNNING)
in sch->__state, use a seqcount.

This adds lockdep support, but more importantly it will allow us
to sample qdisc/class statistics without having to grab qdisc root lock.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Cong Wang <xiyou.wangcong@gmail.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

bond_get_stats() can be called from rtnetlink (with RTNL held)
or from /proc/net/dev seq handler (with RCU held)

The logic added in commit 5f0c5f73 ("bonding: make global bonding
stats more reliable") kind of assumed only one cpu could run there.

If multiple threads are reading /proc/net/dev, stats can be really
messed up after a while.

A second problem is that some fields are 32bit, so we need to properly
handle the wrap around problem.

Given that RTNL is not always held, we need to use
bond_for_each_slave_rcu().

Fixes: 5f0c5f73 ("bonding: make global bonding stats more reliable")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Andy Gospodarek <gospo@cumulusnetworks.com>
Cc: Jay Vosburgh <j.vosburgh@gmail.com>
Cc: Veaceslav Falico <vfalico@gmail.com>
Reviewed-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove unnecessary set of flag IFF_MULTICAST, since ether_setup
already does this.
Signed-off-by: NZhang Shengju <zhangshengju@cmss.chinamobile.com>
Reviewed-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NAndy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NDavid Decotigny <decot@googlers.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There is presently a race condition between the bonding periodic
link monitor and the updating of a slave's speed and duplex.  The former
occurs on a periodic basis, and the latter in response to a driver's
calling of netif_carrier_on.

	It is possible for the periodic monitor to run between the
driver call of netif_carrier_on and the receipt of the NETDEV_CHANGE
event that causes bonding to update the slave's speed and duplex.  This
manifests most notably as a report that a slave is up and "0 Mbps full
duplex" after enslavement, but in principle could report an incorrect
speed and duplex after any link up event if the device comes up with a
different speed or duplex.  This affects the 802.3ad aggregator
selection, as the speed and duplex are selection criteria.

	This is fixed by updating the speed and duplex in the periodic
monitor, prior to using that information.

	This was done historically in bonding, but the call to
bond_update_speed_duplex was removed in commit 876254ae ("bonding:
don't call update_speed_duplex() under spinlocks"), as it might sleep
under lock.  Later, the locking was changed to only hold RTNL, and so
after commit 876254ae ("bonding: don't call update_speed_duplex()
under spinlocks") this call is again safe.
Tested-by: N"Tantilov, Emil S" <emil.s.tantilov@intel.com>
Cc: Veaceslav Falico <vfalico@gmail.com>
Cc: dingtianhong <dingtianhong@huawei.com>
Fixes: 876254ae ("bonding: don't call update_speed_duplex() under spinlocks")
Signed-off-by: NJay Vosburgh <jay.vosburgh@canonical.com>
Acked-by: NDing Tianhong <dingtianhong@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The current logic in bond_arp_rcv will accept an incoming ARP for
validation if (a) the receiving slave is either "active" (which includes
the currently active slave, or the current ARP slave) or, (b) there is a
currently active slave, and it has received an ARP since it became active.
For case (b), the receiving slave isn't the currently active slave, and is
receiving the original broadcast ARP request, not an ARP reply from the
target.

	This logic can fail if there is no currently active slave.  In
this situation, the ARP probe logic cycles through all slaves, assigning
each in turn as the "current_arp_slave" for one arp_interval, then setting
that one as "active," and sending an ARP probe from that slave.  The
current logic expects the ARP reply to arrive on the sending
current_arp_slave, however, due to switch FDB updating delays, the reply
may be directed to another slave.

	This can arise if the bonding slaves and switch are working, but
the ARP target is not responding.  When the ARP target recovers, a
condition may result wherein the ARP target host replies faster than the
switch can update its forwarding table, causing each ARP reply to be sent
to the previous current_arp_slave.  This will never pass the logic in
bond_arp_rcv, as neither of the above conditions (a) or (b) are met.

	Some experimentation on a LAN shows ARP reply round trips in the
200 usec range, but my available switches never update their FDB in less
than 4000 usec.

	This patch changes the logic in bond_arp_rcv to additionally
accept an ARP reply for validation on any slave if there is a current ARP
slave and it sent an ARP probe during the previous arp_interval.

Fixes: aeea64ac ("bonding: don't trust arp requests unless active slave really works")
Cc: Veaceslav Falico <vfalico@gmail.com>
Cc: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: NJay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Replace 'goto' with 'return' to remove unnecessary check at label:
err_undo_flags.

The reason is that 'err_undo_flags' do two things for the first slave device:
1.revert bond mac address if it is set by the slave device.
2.revert bond device type if it's not ARPHRD_ETHER.

It's not necessary for the following three places, they changed neither bond
mac address nor type. It's straightforward to return directly.
Signed-off-by: NZhang Shengju <zhangshengju@cmss.chinamobile.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The return value of kzalloc on failure of allocation of memory should
be -ENOMEM and not -1.

Found using Coccinelle. A simplified version of the semantic patch
used is:

//<smpl>
@@
expression *e;
@@

e = kzalloc(...);
if (e == NULL) {
...
return
- -1
+ -ENOMEM
;
}
//</smpl>

The single call site only checks that the return value is not 0,
hence no change is required at the call site.
Signed-off-by: NAmitoj Kaur Chawla <amitoj1606@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

No need to require the bond down while changing these settings, the change
will be reflected immediately and the 3ad mode will sort itself out.
For faster convergence set port->ntt to true in order to generate new
LACPDUs immediately.

CC: Jay Vosburgh <j.vosburgh@gmail.com>
CC: Veaceslav Falico <vfalico@gmail.com>
CC: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently the bonding allows to set ad_actor_system and prio while the
bond device is down, but these are actually applied only if there aren't
any slaves yet (applied to bond device when first slave shows up, and to
slaves at 3ad bind time). After this patch changes are applied immediately
and the new values can be used/seen after the bond's upped so it's not
necessary anymore to release all and enslave again to see the changes.

CC: Jay Vosburgh <j.vosburgh@gmail.com>
CC: Veaceslav Falico <vfalico@gmail.com>
CC: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NJay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

remove some redudant brackets, use sizeof(*) instead of sizeof(struct x).
Signed-off-by: NZhang Shengju <zhangshengju@cmss.chinamobile.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

netdev_dbg() will add bond device name, it will be helpful if we print
slave device name.
Signed-off-by: NZhang Shengju <zhangshengju@cmss.chinamobile.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Sample output with this set applied for an active-backup bond:

$ cat /sys/devices/virtual/net/bond0/lower_p7p1/statistics/rx_nohandler
16568
$ cat /sys/devices/virtual/net/bond0/lower_p5p2/statistics/rx_nohandler
16583
$ cat /sys/devices/virtual/net/bond0/statistics/rx_nohandler
33151

CC: Jay Vosburgh <j.vosburgh@gmail.com>
CC: Veaceslav Falico <vfalico@gmail.com>
CC: Andy Gospodarek <gospo@cumulusnetworks.com>
CC: netdev@vger.kernel.org
Signed-off-by: NJarod Wilson <jarod@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 1f718f0f ("bonding: populate neighbour's private on enslave")
undoes the fix provided by commit c2edacf8 ("bonding / ipv6: no addrconf
for slaves separately from master") by effectively setting the slave flag
after the slave has been opened.  If the slave comes up quickly enough, it
will go through the IPv6 addrconf before the slave flag has been set and
will get a link local IPv6 address.

In order to ensure that addrconf knows to ignore the slave devices on state
change, set IFF_SLAVE before dev_open() during bonding enslavement.

Fixes: 1f718f0f ("bonding: populate neighbour's private on enslave")
Signed-off-by: NKarl Heiss <kheiss@gmail.com>
Signed-off-by: NJay Vosburgh <jay.vosburgh@canonical.com>
Reviewed-by: NJarod Wilson <jarod@redhat.com>
Signed-off-by: NAndy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The spew in /proc/net/bonding/bond0 uses netif_carrier_ok() to determine
mii_status, while /sys/class/net/bond0/bonding/mii_status looks at
curr_active_slave, which doesn't actually seem to be set sometimes when
the bond actually is up. A mode 4 bond configured via ifcfg-foo files on a
Red Hat Enterprise Linux system, after boot, comes up clean and
functional, but the sysfs node shows mii_status of down, while proc shows
up. A simple enough fix here seems to be to use the same method for
determining up or down in both places, and I'd opt for the one that seems
to match reality.

CC: Jay Vosburgh <j.vosburgh@gmail.com>
CC: Veaceslav Falico <vfalico@gmail.com>
CC: Andy Gospodarek <gospo@cumulusnetworks.com>
CC: netdev@vger.kernel.org
Signed-off-by: NJarod Wilson <jarod@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

to_dev is not used anymore so drop it.
Signed-off-by: NGeliang Tang <geliangtang@163.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The name NETIF_F_ALL_CSUM is a misnomer. This does not correspond to the
set of features for offloading all checksums. This is a mask of the
checksum offload related features bits. It is incorrect to set both
NETIF_F_HW_CSUM and NETIF_F_IP_CSUM or NETIF_F_IPV6 at the same time for
features of a device.

This patch:
  - Changes instances of NETIF_F_ALL_CSUM to NETIF_F_CSUM_MASK (where
    NETIF_F_ALL_CSUM is being used as a mask).
  - Changes bonding, sfc/efx, ipvlan, macvlan, vlan, and team drivers to
    use NEITF_F_HW_CSUM in features list instead of NETIF_F_ALL_CSUM.
Signed-off-by: NTom Herbert <tom@herbertland.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It is not necessary to use two brackets. As such, the redudant brackets
are removed.

CC: Jay Vosburgh <j.vosburgh@gmail.com>
CC: Veaceslav Falico <vfalico@gmail.com>
CC: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: NZhu Yanjun <yanjun.zhu@windriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Be correct and symmetric to enslave and set inactive flags during release.
That gives LAG offload drivers - lower state change listeners - possibility
to do proper cleanup.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Let netdev notifier listeners know about link and slave state change.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Similar to state notifications.

We allow caller to indicate if the notification should happen now or later,
depending on if he holds rtnl mutex or not. Introduce bond_slave_link_notify
function (similar to bond_slave_state_notify) which is later on called
with rtnl mutex and goes over slaves and executes delayed notification.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Initialize netdev_lag_upper_info structure by TX type according to
current bonding mode and pass it along via netdev_master_upper_dev_link.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Sometimes the drivers and other code would find it handy to know some
internal information about upper device being changed. So allow upper-code
to pass information down to notifier listeners during linking.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Eliminate netdev_master_upper_dev_link_private and pass priv directly as
a parameter of netdev_master_upper_dev_link.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Similar to other speeds, add 100G to bonding 802.3ad code.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since commit 7d5cd2ce529b, when bond_enslave fails on devices that
are not ARPHRD_ETHER, if needed, it resets the bonding device back to
ARPHRD_ETHER by calling ether_setup.

	Unfortunately, ether_setup clobbers dev->flags, clearing IFF_UP
if the bond device is up, leaving it in a quasi-down state without
having actually gone through dev_close.  For bonding, if any periodic
work queue items are active (miimon, arp_interval, etc), those will
remain running, as they are stopped by bond_close.  At this point, if
the bonding module is unloaded or the bond is deleted, the system will
panic when the work function is called.

	This panic is resolved by calling dev_close on the bond itself
prior to calling ether_setup.

Cc: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NJay Vosburgh <jay.vosburgh@canonical.com>
Fixes: 7d5cd2ce ("bonding: correctly handle bonding type change on enslave failure")
Acked-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Old logic of updating state-machine is not required since
ad_update_actor_keys() does it implicitly. The only loss is
the notification differentiation between speed vs. duplex
change. Now only one unified notification is printed.
Signed-off-by: NMahesh Bandewar <maheshb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

actor_admin, and actor_oper key is changed at multiple locations in
the code. This patch brings all those updates into one location in
an attempt to avoid possible inconsistent updates causing LACP state
machine to go in weird state.

The unified place is ad_update_actor_key() with simple state-machine
logic -
  (a) If port is "duplex" then only it can participate in LACP
  (b) Speed change reinitializes the LACP state-machine.
Signed-off-by: NMahesh Bandewar <maheshb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Eliminate 'else' clause by simply initializing variable
Signed-off-by: NMahesh Bandewar <maheshb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If using a sixtofour device on top of a bonding device,
skb segmentation of TCP traffic is done right before calling
bonding xmit, because bonding only enables TSO for IPv4.

This patch improves single flow performance by about 120 % on my hosts,
because segmentation is deferred right before calling slave xmit.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If skb carries a l4 hash, no need to perform a flow dissection.

Performance is slightly better :

lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.39012e+06
lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.39393e+06
lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.39988e+06

After patch :

lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.43579e+06
lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.44304e+06
lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.44312e+06
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Tom Herbert <tom@herbertland.com>
Cc: Mahesh Bandewar <maheshb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The flags argument will allow control of the dissection process (for
instance whether to parse beyond L3).
Signed-off-by: NTom Herbert <tom@herbertland.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Some consumers of the netdev events API would like to know who is the
active slave when a NETDEV_CHANGEUPPER or NETDEV_BONDING_FAILOVER
events occur. For example, when managing RoCE GIDs, GIDs based on the
bond's ips should only be set on the port which corresponds to active
slave netdevice.
Signed-off-by: NMatan Barak <matanb@mellanox.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>