[LU-10091] o2iblnd fast reg crash on shutdown

• Type: Bug
• Resolution: Cannot Reproduce
• Priority: Major
• Fix Version/s: Lustre 2.11.0
• Severity: 3
• Assignee: Amir Shehata (Inactive)
• Reporter: Amir Shehata (Inactive)

Description

With MLX-5/Centos 7.4 I get the following crash

[ 3029.933315] LNetError: 34858:0:(o2iblnd.c:1460:kiblnd_destroy_fmr_pool()) ASSERTION( fpo->fpo_map_count == 0 ) failed: 
[ 3029.948390] LNetError: 34858:0:(o2iblnd.c:1460:kiblnd_destroy_fmr_pool()) LBUG
[ 3029.959131] Pid: 34858, comm: rmmod
[ 3029.965643] 
[ 3029.965643] Call Trace:
[ 3029.975245]  [<ffffffffc0a8f7ae>] libcfs_call_trace+0x4e/0x60 [libcfs]
[ 3029.984938]  [<ffffffffc0a8f83c>] lbug_with_loc+0x4c/0xb0 [libcfs]
[ 3029.994207]  [<ffffffffc087c644>] kiblnd_destroy_fmr_pool_list+0x3a4/0x3b0 [ko2iblnd]
[ 3030.005138]  [<ffffffffc087c8c2>] kiblnd_net_fini_pools.isra.25+0x72/0xc0 [ko2iblnd]
[ 3030.015842]  [<ffffffffc087f30f>] kiblnd_shutdown+0x19f/0x4b0 [ko2iblnd]
[ 3030.025247]  [<ffffffffc0af7906>] lnet_shutdown_lndni+0x286/0x470 [lnet]
[ 3030.034521]  [<ffffffffc0af8479>] lnet_shutdown_lndnet+0x69/0xe0 [lnet]
[ 3030.043679]  [<ffffffffc0af85fd>] lnet_shutdown_lndnets+0x10d/0x2a0 [lnet]
[ 3030.053011]  [<ffffffffc0afaab8>] LNetNIFini+0x88/0x110 [lnet]
[ 3030.061145]  [<ffffffffc0f7977d>] ptlrpc_ni_fini+0x15d/0x1e0 [ptlrpc]
[ 3030.069900]  [<ffffffffc0f94c57>] ? ptlrpcd_free+0x147/0x2d0 [ptlrpc]
[ 3030.078544]  [<ffffffffc0f79ab3>] ptlrpc_exit_portals+0x13/0x20 [ptlrpc]
[ 3030.087463]  [<ffffffffc1009a0a>] ptlrpc_exit+0x22/0x618 [ptlrpc]
[ 3030.095571]  [<ffffffff810fe55b>] SyS_delete_module+0x19b/0x300
[ 3030.103461]  [<ffffffff816b4fc9>] system_call_fastpath+0x16/0x1b

Issue Links

• is related to: LU-10089 kiblnd_fmr_pool_map() Failed to map mr 10/11 elements