bonding: fix active-backup transition after link failure
authorMahesh Bandewar <maheshb@google.com>
Fri, 6 Dec 2019 23:44:55 +0000 (15:44 -0800)
committerJakub Kicinski <jakub.kicinski@netronome.com>
Sun, 15 Dec 2019 00:22:34 +0000 (16:22 -0800)
After the recent fix in commit 1899bb325149 ("bonding: fix state
transition issue in link monitoring"), the active-backup mode with
miimon initially come-up fine but after a link-failure, both members
transition into backup state.

Following steps to reproduce the scenario (eth1 and eth2 are the
slaves of the bond):

    ip link set eth1 up
    ip link set eth2 down
    sleep 1
    ip link set eth2 up
    ip link set eth1 down
    cat /sys/class/net/eth1/bonding_slave/state
    cat /sys/class/net/eth2/bonding_slave/state

Fixes: 1899bb325149 ("bonding: fix state transition issue in link monitoring")
CC: Jay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: Mahesh Bandewar <maheshb@google.com>
Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
drivers/net/bonding/bond_main.c

index 041aa9649dfc7ab19a3736168f60b6cd1e467734..48d5ec770b94242fddb71a02effab6fc1ee988e8 100644 (file)
@@ -2272,9 +2272,6 @@ static void bond_miimon_commit(struct bonding *bond)
                        } else if (BOND_MODE(bond) != BOND_MODE_ACTIVEBACKUP) {
                                /* make it immediately active */
                                bond_set_active_slave(slave);
-                       } else if (slave != primary) {
-                               /* prevent it from being the active one */
-                               bond_set_backup_slave(slave);
                        }
 
                        slave_info(bond->dev, slave->dev, "link status definitely up, %u Mbps %s duplex\n",