W dniu 30.10.2018 o 01:11, Paweł Staszewski pisze:
Sorry not complete - followed by hw csum:

[  342.190831] vlan1490: hw csum failure
[  342.190835] CPU: 52 PID: 0 Comm: swapper/52 Not tainted 4.19.0+ #1
[  342.190836] Call Trace:
[  342.190839]  <IRQ>
[  342.190849]  dump_stack+0x46/0x5b
[  342.190856]  __skb_checksum_complete+0x9a/0xa0
[  342.190859]  tcp_v4_rcv+0xef/0x960
[  342.190864]  ip_local_deliver_finish+0x49/0xd0
[  342.190866]  ip_local_deliver+0x5e/0xe0
[  342.190869]  ? ip_sublist_rcv_finish+0x50/0x50
[  342.190870]  ip_rcv+0x41/0xc0
[  342.190874]  __netif_receive_skb_one_core+0x4b/0x70
[  342.190877]  netif_receive_skb_internal+0x2f/0xd0
[  342.190879]  napi_gro_receive+0xb7/0xe0
[  342.190884]  mlx5e_handle_rx_cqe+0x7a/0xd0
[  342.190886]  mlx5e_poll_rx_cq+0xc6/0x930
[  342.190888]  mlx5e_napi_poll+0xab/0xc90
[  342.190893]  ? kmem_cache_free_bulk+0x1e4/0x280
[  342.190895]  net_rx_action+0x1f1/0x320
[  342.190901]  __do_softirq+0xec/0x2b7
[  342.190908]  irq_exit+0x7b/0x80
[  342.190910]  do_IRQ+0x45/0xc0
[  342.190912]  common_interrupt+0xf/0xf
[  342.190914]  </IRQ>
[  342.190916] RIP: 0010:mwait_idle+0x5f/0x1b0
[  342.190917] Code: a8 01 0f 85 3f 01 00 00 31 d2 65 48 8b 04 25 80 4c 01 00 48 89 d1 0f 01 c8 48 8b 00 a8 08 0f 85 40 01 00 00 31 c0 fb 0f 01 c9 <65> 8b 2d 2a c9 6a 7e 0f 1f 44 00 00 65 48 8b 04 25 80 4c 01 00 f0 [  342.190918] RSP: 0018:ffffc900034e7eb8 EFLAGS: 00000246 ORIG_RAX: ffffffffffffffdd [  342.190920] RAX: 0000000000000000 RBX: 0000000000000034 RCX: 0000000000000000 [  342.190921] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 [  342.190922] RBP: 0000000000000034 R08: 0000000000000057 R09: ffff88086fa1fbc0 [  342.190923] R10: 0000000000000000 R11: 00000001000028cc R12: ffff88086d180000 [  342.190923] R13: ffff88086d180000 R14: 0000000000000000 R15: 0000000000000000
[  342.190929]  do_idle+0x1a3/0x1c0
[  342.190931]  cpu_startup_entry+0x14/0x20
[  342.190934]  start_secondary+0x165/0x190
[  342.190939]  secondary_startup_64+0xa4/0xb0


W dniu 30.10.2018 o 01:10, Paweł Staszewski pisze:
Hi


Just checked in test lab latest kernel and have weird traces:

[  219.888673] CPU: 52 PID: 0 Comm: swapper/52 Not tainted 4.19.0+ #1
[  219.888674] Call Trace:
[  219.888676]  <IRQ>
[  219.888685]  dump_stack+0x46/0x5b
[  219.888691]  __skb_checksum_complete+0x9a/0xa0
[  219.888694]  tcp_v4_rcv+0xef/0x960
[  219.888698]  ip_local_deliver_finish+0x49/0xd0
[  219.888700]  ip_local_deliver+0x5e/0xe0
[  219.888702]  ? ip_sublist_rcv_finish+0x50/0x50
[  219.888703]  ip_rcv+0x41/0xc0
[  219.888706]  __netif_receive_skb_one_core+0x4b/0x70
[  219.888708]  netif_receive_skb_internal+0x2f/0xd0
[  219.888710]  napi_gro_receive+0xb7/0xe0
[  219.888714]  mlx5e_handle_rx_cqe+0x7a/0xd0
[  219.888716]  mlx5e_poll_rx_cq+0xc6/0x930
[  219.888717]  mlx5e_napi_poll+0xab/0xc90
[  219.888722]  ? enqueue_task_fair+0x286/0xc40
[  219.888723]  ? enqueue_task_fair+0x1d6/0xc40
[  219.888725]  net_rx_action+0x1f1/0x320
[  219.888730]  __do_softirq+0xec/0x2b7
[  219.888736]  irq_exit+0x7b/0x80
[  219.888737]  do_IRQ+0x45/0xc0
[  219.888740]  common_interrupt+0xf/0xf
[  219.888742]  </IRQ>
[  219.888743] RIP: 0010:mwait_idle+0x5f/0x1b0
[  219.888745] Code: a8 01 0f 85 3f 01 00 00 31 d2 65 48 8b 04 25 80 4c 01 00 48 89 d1 0f 01 c8 48 8b 00 a8 08 0f 85 40 01 00 00 31 c0 fb 0f 01 c9 <65> 8b 2d 2a c9 6a 7e 0f 1f 44 00 00 65 48 8b 04 25 80 4c 01 00 f0 [  219.888746] RSP: 0018:ffffc900034e7eb8 EFLAGS: 00000246 ORIG_RAX: ffffffffffffffde [  219.888749] RAX: 0000000000000000 RBX: 0000000000000034 RCX: 0000000000000000 [  219.888749] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 [  219.888750] RBP: 0000000000000034 R08: 000000000000003b R09: ffff88086fa1fbc0 [  219.888751] R10: 0000000000000000 R11: 00000000ffffb15d R12: ffff88086d180000 [  219.888752] R13: ffff88086d180000 R14: 0000000000000000 R15: 0000000000000000
[  219.888754]  do_idle+0x1a3/0x1c0
[  219.888757]  cpu_startup_entry+0x14/0x20
[  219.888760]  start_secondary+0x165/0x190




Also some perf top attacked to this - 14G rx traffic on vlans (pktgen generated random destination ip's and forwarded by test server)

   PerfTop:   45296 irqs/sec  kernel:99.3%  exact:  0.0% [4000Hz cycles],  (all, 56 CPUs)
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

     7.43%  [kernel]       [k] mlx5e_skb_from_cqe_linear
     5.17%  [kernel]       [k] mlx5e_sq_xmit
     3.83%  [kernel]       [k] fib_table_lookup
     3.41%  [kernel]       [k] irq_entries_start
     2.91%  [kernel]       [k] build_skb
     2.50%  [kernel]       [k] mlx5_eq_int
     2.29%  [kernel]       [k] _raw_spin_lock
     2.27%  [kernel]       [k] tasklet_action_common.isra.21
     1.99%  [kernel]       [k] _raw_spin_lock_irqsave
     1.91%  [kernel]       [k] memcpy_erms
     1.77%  [kernel]       [k] __build_skb
     1.70%  [kernel]       [k] vlan_do_receive
     1.59%  [kernel]       [k] get_page_from_freelist
     1.56%  [kernel]       [k] mlx5e_poll_tx_cq
     1.53%  [kernel]       [k] __dev_queue_xmit
     1.40%  [kernel]       [k] pfifo_fast_dequeue
     1.37%  [kernel]       [k] dev_gro_receive
     1.34%  [kernel]       [k] ipt_do_table
     1.30%  [kernel]       [k] mlx5e_poll_rx_cq
     1.28%  [kernel]       [k] mlx5e_post_rx_wqes
     1.27%  [kernel]       [k] ip_finish_output2
     1.09%  [kernel]       [k] inet_gro_receive
     1.08%  [kernel]       [k] _raw_spin_lock_irq
     1.04%  [kernel]       [k] __sched_text_start
     0.99%  [kernel]       [k] tcp_gro_receive
     0.98%  [kernel]       [k] find_busiest_group
     0.97%  [kernel]       [k] __netif_receive_skb_core
     0.85%  [kernel]       [k] ip_route_input_rcu
     0.85%  [kernel]       [k] free_one_page
     0.84%  [kernel]       [k] mlx5e_handle_rx_cqe
     0.76%  [kernel]       [k] do_idle
     0.72%  [kernel]       [k] mlx5e_xmit
     0.71%  [kernel]       [k] cmd_exec
     0.71%  [kernel]       [k] __page_pool_put_page
     0.69%  [kernel]       [k] kmem_cache_alloc
     0.68%  [kernel]       [k] mlx5_cmd_comp_handler
     0.68%  [kernel]       [k] queued_spin_lock_slowpath
     0.68%  [kernel]       [k] cmd_work_handler
     0.68%  [kernel]       [k] pfifo_fast_enqueue
     0.67%  [kernel]       [k] try_to_wake_up
     0.66%  [kernel]       [k] _raw_spin_trylock
     0.62%  [kernel]       [k] dev_hard_start_xmit
     0.62%  [kernel]       [k] ip_forward
     0.62%  [kernel]       [k] swiotlb_map_page
     0.61%  [kernel]       [k] page_frag_free
     0.60%  [kernel]       [k] mlx5e_build_rx_skb
     0.60%  [kernel]       [k] skb_release_data
     0.57%  [kernel]       [k] netif_skb_features
     0.52%  [kernel]       [k] vlan_dev_hard_start_xmit
     0.50%  [kernel]       [k] kmem_cache_free_bulk
     0.49%  [kernel]       [k] enqueue_task_fair
     0.49%  [kernel]       [k] validate_xmit_skb.isra.142
     0.49%  [kernel]       [k] skb_gro_receive
     0.49%  [kernel]       [k] __qdisc_run

Reply via email to