zrlio / urdma

Verbs on DPDK
100 stars 24 forks source link

kni release abnormally #47

Open ZiYang-oyxy opened 6 years ago

ZiYang-oyxy commented 6 years ago
[  +0.000001]  siw_cm(QP3007734)
[  +0.000001]  siw_cm(QP3007734): SIW QP state: ERROR => ERROR
[  +0.000001]  siw_cm NOP: State: 5
[  +0.000001]  siw_cm(QP3007734): Exit with 0
[  +0.000001]  siw_obj siw_cm(QP3007734): Put Reference
[  +0.000001]  siw_obj(QP3007734): Old refcount: 2
[  +0.000003]  siw_cm(QP3007734): SIW QP state=5, cep=0x          (null)
[  +0.000001]  siw_cm(QP3007734)
[  +0.000001]  siw_cm(QP3007734): SIW QP state: ERROR => ERROR
[  +0.000001]  siw_cm NOP: State: 5
[  +0.000001]  siw_obj(CQ2314293): Old refcount: 3
[  +0.000001]  siw_obj(CQ2314293): Old refcount: 2
[  +0.000001]  siw_obj(PD11587634): Old refcount: 2
[  +0.000001]  siw_obj(QP3007734): Old refcount: 1
[  +0.000001]  siw_obj siw_cm(QP3007734): Free Object
[  +0.000001]  siw_obj(OBJ3007734): IDR Remove Object
[  +0.000002]  siw_obj(OBJ2314293): IDR Remove Object
[  +0.000001]  siw_obj(CQ2314293): Old refcount: 1
[  +0.000001]  siw_obj(CQ2314293): Free Object
[  +0.000002]  siw_obj(OBJ11587634): IDR Remove Object
[  +0.000001]  siw_obj(PD11587634): Old refcount: 1
[  +0.000001]  siw_obj(PD11587634): Free Object
[  +1.926792] chardev release: remove rtr wait event for cep ffff881fc795a800
[  +0.000004]  siw_obj siw_cm(CEP 0xffff881fc795a800): New refcount: 0
[  +0.000001]  siw_obj siw_cm(CEP 0xffff881fc795a800): Free Object
[  +0.000016]  siw_dm (dev=kni0): Event 9
[  +0.000001]  siw_dm sdev=urdma_0 got unhandled netdev event 9 for netdev=kni0
[  +0.000064]  siw_dm (dev=kni0): Event 2
[  +0.000002]  siw_eh: reporting 10
[  +0.000072]  siw_dm (dev=kni0): Event 6
[  +0.000016]  siw_dm: deregister urdma_0 netdev=kni0
[  +0.000003] ib_srpt srpt_remove_one(urdma_0): nothing to do.
[  +0.000074]  siw_dm: destroy siw device urdma_0 netdev=kni0
[  +0.009839]  siw_dm (dev=kni0): Event 17
[  +0.000002]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.004661]  siw_dm (dev=kni0): Event 6
[  +0.006046]  siw_dm (dev=kni0): Event 17
[  +0.000003]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005060]  siw_dm (dev=kni0): Event 6
[  +0.009052]  siw_dm (dev=kni0): Event 17
[  +0.000002]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005061]  siw_dm (dev=kni0): Event 6
[  +0.013020]  siw_dm (dev=kni0): Event 17
[  +0.000002]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005094]  siw_dm (dev=kni0): Event 6
[  +0.006046]  siw_dm (dev=kni0): Event 17
[  +0.000003]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005060]  siw_dm (dev=kni0): Event 6
[  +0.009051]  siw_dm (dev=kni0): Event 17
[  +0.000002]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005062]  siw_dm (dev=kni0): Event 6
[  +0.006041]  siw_dm (dev=kni0): Event 17
[  +0.000002]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005065]  siw_dm (dev=kni0): Event 6
[  +0.006045]  siw_dm (dev=kni0): Event 17
[  +0.000002]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005059]  siw_dm (dev=kni0): Event 6
[  +0.009050]  siw_dm (dev=kni0): Event 17
[  +0.000003]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005060]  siw_dm (dev=kni0): Event 6
[  +0.006047]  siw_dm (dev=kni0): Event 17
[  +0.000002]  siw_dm sdev=<NONE> got unhandled netdev event 17 for netdev=kni0
[  +1.005053] unregister_netdevice: waiting for kni0 to become free. Usage count = 1
patrickmacarthur commented 6 years ago

Did you have an RDMA process running when urdmad exited? Do you have the logs from urdmad? Did urdmad exit due to a segfault or were you attempting to stop it normally when this happened?

ZiYang-oyxy commented 6 years ago

It doesn't happen just bychance when I send TERM signal to the daemon and there is no RDMA secondary process. I will keep track of it

ZiYang-oyxy commented 5 years ago
do_read_established_event
 -> fetch_dest_hwaddr
     -> n = dst_neigh_lookup(&rt->dst, &dst_ipv4_addr);

Cause this issue if the lookup failed.