Intel XL710 (i40e) reports a NETDEV WATCHDOG timeout (tx_timeout) when added to a bond

Issue

• An i40e interface triggers the NETDEV WATCHDOG and reports a tx_timeout when it becomes the active interface in a bond. This might also happen constantly in LACP bonding.

  Apr 08 14:02:45 localhost kernel: bond0: making interface enp1s0f0 the new active one
  Apr 08 14:02:52 localhost kernel: ------------[ cut here ]------------
  Apr 08 14:02:52 localhost kernel: WARNING: CPU: 2 PID: 0 at net/sched/sch_generic.c:356 dev_watchdog+0x248/0x260
  Apr 08 14:02:52 localhost kernel: NETDEV WATCHDOG: enp1s0f0 (i40e): transmit queue 6 timed out
  Apr 08 14:02:52 localhost kernel: Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache bonding sunrpc skx_edac intel_powerclamp coretemp intel_rapl iosf_mbi kvm_intel     kvm irqbypass crc32_pclmul ghash_clmulni_intel iTCO_wdt iTCO_vendor_support aesni_intel lrw gf128mul glue_helper ablk_helper cryptd ipmi_ssif dm_mod pcspkr cdc_ether usbnet sg mii mei_me lpc_ich mei i2c_i801 wmi ipmi_si ipmi_devintf ipmi_msghandler acpi_power_meter acpi_pad ip_tables ext4 mbcache jbd2 sd_mod crc_t10dif crct10dif_generic crct10dif_pclmul crct10dif_common crc32c_intel mgag200 i2c_algo_bit drm_kms_helper i40e syscopyarea sysfillrect sysimgblt fb_sys_fops ttm tg3 drm ahci ptp libahci pps_core megaraid_sas libata drm_panel_orientation_quirks nfit libnvdimm
  Apr 08 14:02:52 localhost kernel: CPU: 2 PID: 0 Comm: swapper/2 Kdump: loaded Not tainted 3.10.0-957.10.1.el7.x86_64 #1
  Apr 08 14:02:52 localhost kernel: Hardware name:
  Apr 08 14:02:52 localhost kernel: Call Trace:
  Apr 08 14:02:52 localhost kernel:  <IRQ>  [<ffffffff9fd68c49>] dump_stack+0x19/0x1b
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f697948>] __warn+0xd8/0x100
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f6979cf>] warn_slowpath_fmt+0x5f/0x80
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9fc6c2b8>] dev_watchdog+0x248/0x260
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9fc6c070>] ? dev_deactivate_queue.constprop.27+0x60/0x60
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f6a8278>] call_timer_fn+0x38/0x110
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9fc6c070>] ? dev_deactivate_queue.constprop.27+0x60/0x60
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f6aa6dd>] run_timer_softirq+0x24d/0x300
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f6a1225>] __do_softirq+0xf5/0x280
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9fd7f2ec>] call_softirq+0x1c/0x30
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f62e675>] do_softirq+0x65/0xa0
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f6a15a5>] irq_exit+0x105/0x110
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9fd80698>] smp_apic_timer_interrupt+0x48/0x60
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9fd7cdb2>] apic_timer_interrupt+0x162/0x170
  Apr 08 14:02:52 localhost kernel:  <EOI>  [<ffffffff9fbb3457>] ? cpuidle_enter_state+0x57/0xd0
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9fbb35ae>] cpuidle_idle_call+0xde/0x230
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f63670e>] arch_cpu_idle+0xe/0xc0
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f6fc93a>] cpu_startup_entry+0x14a/0x1e0
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f657e17>] start_secondary+0x1f7/0x270
  Apr 08 14:02:52 localhost kernel:  [<ffffffff9f6000d5>] start_cpu+0x5/0x14
  Apr 08 14:02:52 localhost kernel: ---[ end trace a7b83f01706b526c ]--- 
  Apr 08 14:02:52 localhost kernel: i40e 0000:d8:00.0 enp1s0f0: tx_timeout: VSI_seid: 390, Q 6, NTC: 0x0, HWB: 0x0, NTU: 0x1, TAIL: 0x1, INT: 0x1
  Apr 08 14:02:52 localhost kernel: i40e 0000:d8:00.0 enp1s0f0: tx_timeout recovery level 1, hung_queue 6
  Apr 08 14:02:52 localhost kernel: i40e 0000:d8:00.0: VSI seid 390 Tx ring 0 disable timeout
  Apr 08 14:02:52 localhost kernel: bond0: link status definitely down for interface enp1s0f0, disabling it
  Apr 08 14:02:52 localhost kernel: bond0: making interface eno1 the new active one
  Apr 08 14:02:52 localhost kernel: Non-contiguous TC - Disabling DCB
  Apr 08 14:02:52 localhost kernel: bond0: link status definitely up for interface enp1s0f0, 10000 Mbps full duplex