ixbge (OVH HOST-32L) netdev kernel warnings

[hvisage]

HI there,

Using an OVH HOST-32L, I'm seeing these warnings on a continual basis.

ANy advice/knowledge how to fix these?

[27741.221495] ------------[ cut here ]------------
[27741.222022] WARNING: CPU: 7 PID: 3067 at net/core/dev.c:2392 skb_warn_bad_offload+0xcf/0xe0()
[27741.222559] ixgbe: caps=(0x0000080644114bb3, 0x0000000000000000) len=7389 data_len=7335 gso_size=1452 gso_type=5 ip_summed=0
[27741.223100] Modules linked in: veth ip_set ip6table_filter ip6_tables iptable_filter ip_tables softdog x_tables nfsd auth_rpcgss nfs_acl nfs lockd grace fscache sunrpc ib_iser rdma_cm iw_cm ib_cm ib_sa ib_mad ib_core ib_addr iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi nfnetlink_log nfnetlink zfs(PO) zunicode(PO) zcommon(PO) znvpair(PO) spl(O) zavl(PO) intel_rapl x86_pkg_temp_thermal intel_powerclamp ast coretemp ttm kvm_intel drm_kms_helper kvm drm i2c_algo_bit snd_pcm syscopyarea crct10dif_pclmul crc32_pclmul sysfillrect sysimgblt snd_timer snd soundcore aesni_intel joydev aes_x86_64 input_leds lrw gf128mul glue_helper mxm_wmi ablk_helper cryptd pcspkr ioatdma sb_edac edac_core i2c_i801 mei_me shpchp 8250_fintek ipmi_ssif mei acpi_pad lpc_ich mac_hid wmi vhost_net vhost macvtap macvlan
[27741.226864] ipmi_si ipmi_poweroff ipmi_devintf ipmi_msghandler autofs4 btrfs xor raid6_pq raid1 hid_generic usbmouse usbkbd usbhid hid ahci libahci ixgbe(O) dca vxlan ip6_udp_tunnel udp_tunnel ptp pps_core
[27741.228230] CPU: 7 PID: 3067 Comm: vhost-3053 Tainted: P W O 4.2.8-1-pve #1
[27741.228979] Hardware name: Supermicro Super Server/X10SDV-4C-TLN2F, BIOS 1.0b 09/09/2015
[27741.229738] 0000000000000000 000000000b8d832a ffff88087fdc38c8 ffffffff81802a68
[27741.230484] 0000000000000000 ffff88087fdc3920 ffff88087fdc3908 ffffffff8107bbda
[27741.231198] ffff88087fdc3938 ffff88030c0c0e00 ffff880855080000 0000000000000005
[27741.231929] Call Trace:
[27741.232684] <IRQ> [<ffffffff81802a68>] dump_stack+0x45/0x57
[27741.233422] [<ffffffff8107bbda>] warn_slowpath_common+0x8a/0xc0
[27741.234159] [<ffffffff8107bc65>] warn_slowpath_fmt+0x55/0x70
[27741.234899] [<ffffffff816e4d8f>] skb_warn_bad_offload+0xcf/0xe0
[27741.235613] [<ffffffff816eb0ee>] __skb_gso_segment+0x7e/0xd0
[27741.236348] [<ffffffff816eb493>] validate_xmit_skb.isra.97.part.98+0x133/0x2c0
[27741.237105] [<ffffffff816eba1f>] validate_xmit_skb_list+0x3f/0x60
[27741.237835] [<ffffffff8170f761>] sch_direct_xmit+0x171/0x210
[27741.238553] [<ffffffff816ebd71>] __dev_queue_xmit+0x251/0x570
[27741.239300] [<ffffffff816ec0a3>] dev_queue_xmit_sk+0x13/0x20
[27741.240050] [<ffffffff817d7d3d>] br_dev_queue_push_xmit+0x8d/0x160
[27741.240781] [<ffffffff817d7e49>] br_forward_finish+0x39/0xa0
[27741.241508] [<ffffffff817d7cb0>] ? deliver_clone+0x60/0x60
[27741.242257] [<ffffffff817d7f47>] __br_forward+0x97/0x130
[27741.242976] [<ffffffff810aba4f>] ? sched_clock_cpu+0x8f/0xa0
[27741.243698] [<ffffffff817d7e10>] ? br_dev_queue_push_xmit+0x160/0x160
[27741.244452] [<ffffffff817d84e7>] br_forward+0x87/0xa0
[27741.245242] [<ffffffff817d95df>] br_handle_frame_finish+0x35f/0x670
[27741.245955] [<ffffffff810a645c>] ? try_to_wake_up+0x20c/0x340
[27741.246693] [<ffffffff817d9a65>] br_handle_frame+0x175/0x290
[27741.247433] [<ffffffff817d9280>] ? br_handle_local_finish+0xa0/0xa0
[27741.248159] [<ffffffff816e9272>] __netif_receive_skb_core+0x262/0xa10
[27741.248891] [<ffffffffc005ef2a>] ? ixgbe_poll+0x43a/0x660 [ixgbe]
[27741.249633] [<ffffffff816e9a3a>] __netif_receive_skb+0x1a/0x70
[27741.250352] [<ffffffff816ea7d8>] process_backlog+0xa8/0x150
[27741.251065] [<ffffffff816e9fc0>] net_rx_action+0x220/0x340
[27741.251803] [<ffffffff81080075>] __do_softirq+0x105/0x260
[27741.252529] [<ffffffff8180b55c>] do_softirq_own_stack+0x1c/0x30
[27741.253251] <EOI> [<ffffffff8107f980>] do_softirq.part.20+0x40/0x50
[27741.253996] [<ffffffff8108023d>] do_softirq+0x1d/0x20
[27741.254727] [<ffffffff816e8713>] netif_rx_ni+0x33/0x70
[27741.255431] [<ffffffff815c26ce>] tun_get_user+0x52e/0x930
[27741.256140] [<ffffffff815c2b22>] tun_sendmsg+0x52/0x80
[27741.256848] [<ffffffffc00201bd>] handle_tx+0x2fd/0x510 [vhost_net]
[27741.257511] [<ffffffffc0020405>] handle_tx_kick+0x15/0x20 [vhost_net]
[27741.258157] [<ffffffffc013a61e>] vhost_worker+0x10e/0x1b0 [vhost]
[27741.258803] [<ffffffffc013a510>] ? vhost_dev_reset_owner+0x50/0x50 [vhost]
[27741.259390] [<ffffffff8109b1fa>] kthread+0xea/0x100
[27741.260001] [<ffffffff8109b110>] ? kthread_create_on_node+0x1f0/0x1f0
[27741.260582] [<ffffffff81809e5f>] ret_from_fork+0x3f/0x70
[27741.261165] [<ffffffff8109b110>] ? kthread_create_on_node+0x1f0/0x1f0
[27741.261704] ---[ end trace fbe791f2ca97f3fd ]---

 

[michelv]

I know this is a old thread but I had the same error recently on my proxmox 5.0-32 (kernel 4.10.17-3-pve).
So hopefully my answer may help some others with this error.

I don't know since when I had this error but this makes my logs grow a lot and filled up my root filesystem.

Here is what I did to make this warning/error to stop:

[27741.222022] WARNING: CPU: 7 PID: 3067 at net/core/dev.c:2392 skb_warn_bad_offload+0xcf/0xe0()
...
[27741.228230] CPU: 7 PID: 3067 Comm: vhost-3053 Tainted: P W O 4.2.8-1-pve #1

Notice theses two log lines which tell us this is an offload network error from the PID 3067 with CMD vhost-3053.
Here "vhost-3053" means this is a VM initiated by the pid 3053.
Then get the vm id by searching pid 3053:

$ ps -ef | grep 3053

Now we know from which vm this error comes from.
Connect to the vm and turn off all network offload option.
In my linux, I did:

$ ethtool -K eth0 sg off

Then no more of this error in my proxmox syslog.

PS: I tried to disable network offload option directly from host network interfaces but it didn't work for me, but it works when disabled from vm os network interface.

HTH,
Michel.

 

[fabian]

see https://bugzilla.proxmox.com/show_bug.cgi?id=1505 , updated kernel is already on pve-no-subscription.