bonding/802.3ad: fix slave link initialization transition states
authorJarod Wilson <jarod@redhat.com>
Fri, 24 May 2019 13:49:28 +0000 (09:49 -0400)
committerDavid S. Miller <davem@davemloft.net>
Sun, 26 May 2019 20:28:23 +0000 (13:28 -0700)
commit334031219a84b9994594015aab85ed7754c80176
treebaf28492157e9c7da508b26772e922ef6cdbf85e
parentf4bcf14e3997a422169e5a661fa811868ca221d8
bonding/802.3ad: fix slave link initialization transition states

Once in a while, with just the right timing, 802.3ad slaves will fail to
properly initialize, winding up in a weird state, with a partner system
mac address of 00:00:00:00:00:00. This started happening after a fix to
properly track link_failure_count tracking, where an 802.3ad slave that
reported itself as link up in the miimon code, but wasn't able to get a
valid speed/duplex, started getting set to BOND_LINK_FAIL instead of
BOND_LINK_DOWN. That was the proper thing to do for the general "my link
went down" case, but has created a link initialization race that can put
the interface in this odd state.

The simple fix is to instead set the slave link to BOND_LINK_DOWN again,
if the link has never been up (last_link_up == 0), so the link state
doesn't bounce from BOND_LINK_DOWN to BOND_LINK_FAIL -- it hasn't failed
in this case, it simply hasn't been up yet, and this prevents the
unnecessary state change from DOWN to FAIL and getting stuck in an init
failure w/o a partner mac.

Fixes: ea53abfab960 ("bonding/802.3ad: fix link_failure_count tracking")
CC: Jay Vosburgh <j.vosburgh@gmail.com>
CC: Veaceslav Falico <vfalico@gmail.com>
CC: Andy Gospodarek <andy@greyhouse.net>
CC: "David S. Miller" <davem@davemloft.net>
CC: netdev@vger.kernel.org
Tested-by: Heesoon Kim <Heesoon.Kim@stratus.com>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/bonding/bond_main.c