[ofa-general] rdma_cm kernel Oops

[Vladimir Sokolovsky]

Hi Sean,
I got the following kernel oops while testing RDS HA (kernel 2.6.20):

rdma_destroy_id+0x124/0x193 corresponds to the line 778 in drivers/infiniband/core/cma.c

    771 static void cma_release_port(struct rdma_id_private *id_priv)
    772 {
    773         struct rdma_bind_list *bind_list = id_priv->bind_list;
    774
    775         if (!bind_list)
    776                 return;
    777
    778         mutex_lock(&lock);
    779         hlist_del(&id_priv->node);
    780         if (hlist_empty(&bind_list->owners)) {
    781                 idr_remove(bind_list->ps, bind_list->port);
    782                 kfree(bind_list);
    783         }
    784         mutex_unlock(&lock);
    785 }


Oops:
Jun[  645.944058] Pid: 7354, comm: rdma_cm_wq Not tainted 2.6.20 #2
  5 09:11:48 sw1[  645.944061] RIP: 0010:[<ffffffff8819aa7c>]  [<ffffffff8819aa7c>] :rdma_cm:rdma_destroy_id+0x124/0x193
23 kernel: [  64[  645.944072] RSP: 0018:ffff81011f223e30  EFLAGS: 00010206
5.816913] rds_sh[  645.944076] RAX: 0000000000100100 RBX: ffff81011d86d340 RCX: ffff8101224d0350
utdown_worker: w[  645.944080] RDX: 0000000000200200 RSI: 0000000000000056 RDI: ffffffff881a2140
as_conn 0 was_co[  645.944084] RBP: ffff8101224d0270 R08: 0000000000000000 R09: 0000000000000000
nning -1
[  645.944087] R10: ffff81011f223d50 R11: 0000000000000048 R12: 0000000000000001
[  645.944091] R13: 0000000000000287 R14: ffffffff8819b445 R15: 0000000000000000
[  645.944095] FS:  0000000000000000(0000) GS:ffffffff8058e000(0000) knlGS:0000000000000000
[  645.944099] CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
[  645.944103] CR2: 0000000000200200 CR3: 000000011e21d000 CR4: 00000000000006e0
[  645.944107] Process rdma_cm_wq (pid: 7354, threadinfo ffff81011f222000, task ffff810117df8830)
[  645.944110] Stack:  ffff8101224d0270 ffff8101224d0270 ffff81011a850a20 ffffffff8819b4a7
[  645.944119]  ffff81011a850a28 ffff81011d89ea48 ffff81011a850a20 ffffffff80239c4e
[  645.944126]  ffff81011d89ea48 ffffffff80239ced ffff8101201c7d98 00000000fffffffc
[  645.944132] Call Trace:
[  645.944143]  [<ffffffff8819b4a7>] :rdma_cm:cma_work_handler+0x62/0x6e
[  645.944153]  [<ffffffff80239c4e>] run_workqueue+0xa5/0x144
[  645.944159]  [<ffffffff80239ced>] worker_thread+0x0/0x165
[  645.944164]  [<ffffffff8023cc58>] keventd_create_kthread+0x0/0x6a
[  645.944169]  [<ffffffff80239e1c>] worker_thread+0x12f/0x165
[  645.944177]  [<ffffffff80225003>] default_wake_function+0x0/0xe
[  645.944184]  [<ffffffff80225003>] default_wake_function+0x0/0xe
[  645.944190]  [<ffffffff8023cc2f>] kthread+0xc8/0xf1
[  645.944198]  [<ffffffff8020a2b8>] child_rip+0xa/0x12
[  645.944203]  [<ffffffff8023cc58>] keventd_create_kthread+0x0/0x6a
[  645.944213]  [<ffffffff8023cb67>] kthread+0x0/0xf1
[  645.944217]  [<ffffffff8020a2ae>] child_rip+0x0/0x12
[  645.944221] 
[  645.944223] 
[  645.944224] Code: 48 89 02 74 04 48 89 50 08 48 c7 85 e0 00 00 00 00 01 10 00 
[  645.944236] RIP  [<ffffffff8819aa7c>] :rdma_cm:rdma_destroy_id+0x124/0x193
[  645.944246]  RSP <ffff81011f223e30>
[  645.944249] CR2: 0000000000200200
[  645.944251]  <4>created cm id ffff8101224d0270 for conn ffff81011c857d48

-- 
Vladimir Sokolovsky <vlad at mellanox.co.il>
Mellanox Technologies Ltd.