* team: Bug when macvlans are defined on top of the slaves
@ 2014-08-20 16:08 Erik Hugne
2014-08-22 10:13 ` Jiri Pirko
0 siblings, 1 reply; 2+ messages in thread
From: Erik Hugne @ 2014-08-20 16:08 UTC (permalink / raw)
To: netdev, jiri; +Cc: libteam
When macvlans are defined on top of the team port devices, the following oops
happens immediately when teamd is started.
[ 108.224148] team0: Mode changed to "activebackup"
[ 108.230450] e1000: eth1 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None
[ 108.234107] 8021q: adding VLAN 0 to HW filter on device eth1
[ 108.235593] BUG: unable to handle kernel NULL pointer dereference at 0000000000000018
[ 108.236397] IP: [<ffffffff815ddb7e>] __mutex_lock_slowpath+0x4e/0x310
[ 108.236397] PGD 3fb2e067 PUD 3fb23067 PMD 0
[ 108.236397] Oops: 0002 [#1] SMP
[ 108.236397] Modules linked in: team_mode_activebackup team macvlan
[ 108.236397] CPU: 0 PID: 201 Comm: teamd Not tainted 3.16.0+ #237
[ 108.236397] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
[ 108.236397] task: ffff88003f958ed0 ti: ffff88003d15c000 task.ti: ffff88003d15c000
[ 108.236397] RIP: 0010:[<ffffffff815ddb7e>] [<ffffffff815ddb7e>] __mutex_lock_slowpath+0x4e/0x310
[ 108.236397] RSP: 0018:ffff88003d15f698 EFLAGS: 00010046
[ 108.236397] RAX: 0000000000000100 RBX: 0000000000000010 RCX: 0000000000000001
[ 108.236397] RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000010
[ 108.236397] RBP: ffff88003d15f6e8 R08: ffff88003f850900 R09: ffff88003e400008
[ 108.236397] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000010
[ 108.236397] R13: ffff88003f958ed0 R14: 0000000000000018 R15: 0000000000000246
[ 108.236397] FS: 00007ff10a811740(0000) GS:ffff88003ec00000(0000) knlGS:0000000000000000
[ 108.236397] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 108.236397] CR2: 0000000000000018 CR3: 000000003fa96000 CR4: 00000000000006f0
[ 108.236397] Stack:
[ 108.236397] ffff88003d3d6000 0000000000000081 ffff88003d3d6000 0000000000000001
[ 108.251189] ffff88003d15f710 0000000000000010 0000000000000010 0000000000000001
[ 108.251189] ffff88003d15f7d0 0000000000000000 ffff88003d15f700 ffffffff815dde55
[ 108.251189] Call Trace:
[ 108.251189] [<ffffffff815dde55>] mutex_lock+0x15/0x25
[ 108.251189] [<ffffffffa000a0df>] team_port_change_check+0x1f/0x60 [team]
[ 108.251189] [<ffffffffa000afc8>] team_device_event+0xa8/0x150 [team]
[ 108.251189] [<ffffffff810631ec>] notifier_call_chain+0x4c/0x70
[ 108.251189] [<ffffffff81063301>] raw_notifier_call_chain+0x11/0x20
[ 108.251189] [<ffffffff814b5330>] call_netdevice_notifiers_info+0x30/0x60
[ 108.251189] [<ffffffff814bcdac>] dev_open+0x5c/0x70
[ 108.251189] [<ffffffffa000a582>] team_add_slave+0x2b2/0x6c0 [team]
[ 108.251189] [<ffffffff81564390>] ? inet6_fill_ifla6_attrs+0x360/0x390
[ 108.251189] [<ffffffff814cb73e>] do_setlink+0x9be/0xa20
[ 108.251189] [<ffffffff814c99f9>] ? rtnl_fill_ifinfo+0x899/0xb40
[ 108.251189] [<ffffffff814cbdcf>] rtnl_newlink+0x4ff/0x730
[ 108.251189] [<ffffffff814cb9d8>] ? rtnl_newlink+0x108/0x730
[ 108.251189] [<ffffffff814e5e51>] ? netlink_sendskb+0x11/0x40
[ 108.251189] [<ffffffff814e6033>] ? netlink_unicast+0x1b3/0x250
[ 108.251189] [<ffffffff814ca8e0>] ? rtnl_getlink+0x130/0x1d0
[ 108.251189] [<ffffffff814caa70>] rtnetlink_rcv_msg+0x90/0x250
[ 108.251189] [<ffffffff814ca9e0>] ? rtnetlink_rcv+0x30/0x30
[ 108.251189] [<ffffffff814e6739>] netlink_rcv_skb+0xa9/0xc0
[ 108.251189] [<ffffffff814ca9d3>] rtnetlink_rcv+0x23/0x30
[ 108.251189] [<ffffffff814e5fd8>] netlink_unicast+0x158/0x250
[ 108.251189] [<ffffffff8128912d>] ? memcpy_fromiovec+0x4d/0x90
[ 108.251189] [<ffffffff814e6487>] netlink_sendmsg+0x317/0x410
[ 108.251189] [<ffffffff814a0490>] sock_sendmsg+0xa0/0xc0
[ 108.251189] [<ffffffff814a1d68>] ? move_addr_to_kernel+0x38/0x80
[ 108.251189] [<ffffffff814a1d23>] ___sys_sendmsg+0x373/0x380
[ 108.251189] [<ffffffff8103d034>] ? __do_page_fault+0x284/0x520
[ 108.251189] [<ffffffff814a2bbd>] __sys_sendmsg+0x3d/0x80
[ 108.251189] [<ffffffff814a2c0d>] SyS_sendmsg+0xd/0x20
[ 108.251189] [<ffffffff815dfb52>] system_call_fastpath+0x16/0x1b
[ 108.251189] Code: 35 20 0d 60 00 45 85 f6 75 13 65 8b 04 25 20 b8 00 00 a9 00 ff 1f 00 0f 85 83 02 00 00 9c 41 5f fa b8 00 01 00 00 4d 8d 74 24 08 <f0> 66 41 0f c1 44 24 08 0f b6 d4 38 c2 0f 85 43 02 00 00 44 8b
[ 108.251189] RIP [<ffffffff815ddb7e>] __mutex_lock_slowpath+0x4e/0x310
[ 108.251189] RSP <ffff88003d15f698>
[ 108.251189] CR2: 0000000000000018
[ 108.251189] ---[ end trace 97bf1f31f04db01b ]---
Interface config:
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP mode DEFAULT group default qlen 1000
link/ether 00:0f:ff:10:03:01 brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
link/ether 00:0f:ff:11:04:01 brd ff:ff:ff:ff:ff:ff
4: eth2: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
link/ether 00:0f:ff:11:05:01 brd ff:ff:ff:ff:ff:ff
5: macvlan0@eth1: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN mode DEFAULT group default
link/ether da:ad:a9:57:7a:82 brd ff:ff:ff:ff:ff:ff
6: macvlan1@eth2: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN mode DEFAULT group default
link/ether e6:4d:af:76:55:a2 brd ff:ff:ff:ff:ff:ff
Teamd config:
{
"device": "team0",
"runner": {"name": "activebackup"},
"link_watch": {
"name": "arp_ping",
"interval": 100,
"missed_max": 30,
"source_host": "192.168.123.101",
"target_host": "192.168.123.1"
},
"ports": {
"eth1": {
"prio": -10,
"sticky": true
},
"eth2": {
"prio": 100
}
}
}
teamd version: v1.12
kernel version: recent net-next (33caee3)
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: team: Bug when macvlans are defined on top of the slaves
2014-08-20 16:08 team: Bug when macvlans are defined on top of the slaves Erik Hugne
@ 2014-08-22 10:13 ` Jiri Pirko
0 siblings, 0 replies; 2+ messages in thread
From: Jiri Pirko @ 2014-08-22 10:13 UTC (permalink / raw)
To: Erik Hugne; +Cc: netdev, libteam
Wed, Aug 20, 2014 at 06:08:04PM CEST, erik.hugne@ericsson.com wrote:
>When macvlans are defined on top of the team port devices, the following oops
>happens immediately when teamd is started.
>
>
>[ 108.224148] team0: Mode changed to "activebackup"
>[ 108.230450] e1000: eth1 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None
>[ 108.234107] 8021q: adding VLAN 0 to HW filter on device eth1
>[ 108.235593] BUG: unable to handle kernel NULL pointer dereference at 0000000000000018
>[ 108.236397] IP: [<ffffffff815ddb7e>] __mutex_lock_slowpath+0x4e/0x310
>[ 108.236397] PGD 3fb2e067 PUD 3fb23067 PMD 0
>[ 108.236397] Oops: 0002 [#1] SMP
>[ 108.236397] Modules linked in: team_mode_activebackup team macvlan
>[ 108.236397] CPU: 0 PID: 201 Comm: teamd Not tainted 3.16.0+ #237
>[ 108.236397] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
>[ 108.236397] task: ffff88003f958ed0 ti: ffff88003d15c000 task.ti: ffff88003d15c000
>[ 108.236397] RIP: 0010:[<ffffffff815ddb7e>] [<ffffffff815ddb7e>] __mutex_lock_slowpath+0x4e/0x310
>[ 108.236397] RSP: 0018:ffff88003d15f698 EFLAGS: 00010046
>[ 108.236397] RAX: 0000000000000100 RBX: 0000000000000010 RCX: 0000000000000001
>[ 108.236397] RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000010
>[ 108.236397] RBP: ffff88003d15f6e8 R08: ffff88003f850900 R09: ffff88003e400008
>[ 108.236397] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000010
>[ 108.236397] R13: ffff88003f958ed0 R14: 0000000000000018 R15: 0000000000000246
>[ 108.236397] FS: 00007ff10a811740(0000) GS:ffff88003ec00000(0000) knlGS:0000000000000000
>[ 108.236397] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
>[ 108.236397] CR2: 0000000000000018 CR3: 000000003fa96000 CR4: 00000000000006f0
>[ 108.236397] Stack:
>[ 108.236397] ffff88003d3d6000 0000000000000081 ffff88003d3d6000 0000000000000001
>[ 108.251189] ffff88003d15f710 0000000000000010 0000000000000010 0000000000000001
>[ 108.251189] ffff88003d15f7d0 0000000000000000 ffff88003d15f700 ffffffff815dde55
>[ 108.251189] Call Trace:
>[ 108.251189] [<ffffffff815dde55>] mutex_lock+0x15/0x25
>[ 108.251189] [<ffffffffa000a0df>] team_port_change_check+0x1f/0x60 [team]
>[ 108.251189] [<ffffffffa000afc8>] team_device_event+0xa8/0x150 [team]
>[ 108.251189] [<ffffffff810631ec>] notifier_call_chain+0x4c/0x70
>[ 108.251189] [<ffffffff81063301>] raw_notifier_call_chain+0x11/0x20
>[ 108.251189] [<ffffffff814b5330>] call_netdevice_notifiers_info+0x30/0x60
>[ 108.251189] [<ffffffff814bcdac>] dev_open+0x5c/0x70
>[ 108.251189] [<ffffffffa000a582>] team_add_slave+0x2b2/0x6c0 [team]
I see the problem. team_port_add->team_port_enter sets IFF_TEAM_PORT before
netdev_rx_handler_register is called and therefore team_device_event
thinks that the rx_handler_data pointer is port, but it is macvlan
instead. I'll fix this.
But the thing is, since both macvlan and team are rx_handler users, you
cannot have them both on a single device. If the oops would not appear,
netdev_rx_handler_register would fail anyway. You have to create macvlan on
top of team device instead.
>[ 108.251189] [<ffffffff81564390>] ? inet6_fill_ifla6_attrs+0x360/0x390
>[ 108.251189] [<ffffffff814cb73e>] do_setlink+0x9be/0xa20
>[ 108.251189] [<ffffffff814c99f9>] ? rtnl_fill_ifinfo+0x899/0xb40
>[ 108.251189] [<ffffffff814cbdcf>] rtnl_newlink+0x4ff/0x730
>[ 108.251189] [<ffffffff814cb9d8>] ? rtnl_newlink+0x108/0x730
>[ 108.251189] [<ffffffff814e5e51>] ? netlink_sendskb+0x11/0x40
>[ 108.251189] [<ffffffff814e6033>] ? netlink_unicast+0x1b3/0x250
>[ 108.251189] [<ffffffff814ca8e0>] ? rtnl_getlink+0x130/0x1d0
>[ 108.251189] [<ffffffff814caa70>] rtnetlink_rcv_msg+0x90/0x250
>[ 108.251189] [<ffffffff814ca9e0>] ? rtnetlink_rcv+0x30/0x30
>[ 108.251189] [<ffffffff814e6739>] netlink_rcv_skb+0xa9/0xc0
>[ 108.251189] [<ffffffff814ca9d3>] rtnetlink_rcv+0x23/0x30
>[ 108.251189] [<ffffffff814e5fd8>] netlink_unicast+0x158/0x250
>[ 108.251189] [<ffffffff8128912d>] ? memcpy_fromiovec+0x4d/0x90
>[ 108.251189] [<ffffffff814e6487>] netlink_sendmsg+0x317/0x410
>[ 108.251189] [<ffffffff814a0490>] sock_sendmsg+0xa0/0xc0
>[ 108.251189] [<ffffffff814a1d68>] ? move_addr_to_kernel+0x38/0x80
>[ 108.251189] [<ffffffff814a1d23>] ___sys_sendmsg+0x373/0x380
>[ 108.251189] [<ffffffff8103d034>] ? __do_page_fault+0x284/0x520
>[ 108.251189] [<ffffffff814a2bbd>] __sys_sendmsg+0x3d/0x80
>[ 108.251189] [<ffffffff814a2c0d>] SyS_sendmsg+0xd/0x20
>[ 108.251189] [<ffffffff815dfb52>] system_call_fastpath+0x16/0x1b
>[ 108.251189] Code: 35 20 0d 60 00 45 85 f6 75 13 65 8b 04 25 20 b8 00 00 a9 00 ff 1f 00 0f 85 83 02 00 00 9c 41 5f fa b8 00 01 00 00 4d 8d 74 24 08 <f0> 66 41 0f c1 44 24 08 0f b6 d4 38 c2 0f 85 43 02 00 00 44 8b
>[ 108.251189] RIP [<ffffffff815ddb7e>] __mutex_lock_slowpath+0x4e/0x310
>[ 108.251189] RSP <ffff88003d15f698>
>[ 108.251189] CR2: 0000000000000018
>[ 108.251189] ---[ end trace 97bf1f31f04db01b ]---
>
>
>Interface config:
>
>1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default
> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
>2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP mode DEFAULT group default qlen 1000
> link/ether 00:0f:ff:10:03:01 brd ff:ff:ff:ff:ff:ff
>3: eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
> link/ether 00:0f:ff:11:04:01 brd ff:ff:ff:ff:ff:ff
>4: eth2: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
> link/ether 00:0f:ff:11:05:01 brd ff:ff:ff:ff:ff:ff
>5: macvlan0@eth1: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN mode DEFAULT group default
> link/ether da:ad:a9:57:7a:82 brd ff:ff:ff:ff:ff:ff
>6: macvlan1@eth2: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN mode DEFAULT group default
> link/ether e6:4d:af:76:55:a2 brd ff:ff:ff:ff:ff:ff
>
>
>Teamd config:
>{
> "device": "team0",
> "runner": {"name": "activebackup"},
> "link_watch": {
> "name": "arp_ping",
> "interval": 100,
> "missed_max": 30,
> "source_host": "192.168.123.101",
> "target_host": "192.168.123.1"
> },
> "ports": {
> "eth1": {
> "prio": -10,
> "sticky": true
> },
> "eth2": {
> "prio": 100
> }
> }
>}
>
>teamd version: v1.12
>kernel version: recent net-next (33caee3)
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2014-08-22 10:13 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-08-20 16:08 team: Bug when macvlans are defined on top of the slaves Erik Hugne
2014-08-22 10:13 ` Jiri Pirko
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.