Details
-
Bug
-
Resolution: Unresolved
-
Medium
-
None
-
None
-
3
-
9223372036854775807
Description
FMR/FastReg pools teardown should happen after all connections are gone because the pools are tied to TX/conn completion. Execute the pools teardown routine after conns have drained. Otherwise a crash similar to the following may occur
[1208742.386633] LNet: 3551458:0:(lib-ptl.c:988:lnet_clear_lazy_portal()) Active lazy portal 0 on exit [1208742.763986] LNetError: 3551458:0:(o2iblnd.c:1438:kiblnd_destroy_fmr_pool()) ASSERTION( fpo->fpo_map_count == 0 ) failed: [1208742.768564] LNetError: 3551458:0:(o2iblnd.c:1438:kiblnd_destroy_fmr_pool()) LBUG [1208742.779713] Call Trace: [1208742.780998] dump_stack+0x41/0x60 [1208742.782440] lbug_with_loc.cold.8+0x5/0x43 [libcfs] [1208742.784502] kiblnd_destroy_fmr_pool_list+0x2d9/0x310 [ko2iblnd] [1208742.787004] kiblnd_net_fini_pools.isra.28+0x1a7/0x210 [ko2iblnd] [1208742.789533] kiblnd_shutdown+0x10a/0x650 [ko2iblnd] [1208742.791495] ? srso_alias_return_thunk+0x5/0xfcdfd [1208742.793477] ? init_wait_var_entry+0x50/0x50 [1208742.795262] lnet_shutdown_lndni+0x2b0/0x4c0 [lnet] [1208742.797437] lnet_shutdown_lndnet+0x6c/0xb0 [lnet] [1208742.799424] ? lnet_user_add_peer_ni+0x20/0x20 [lnet] [1208742.801725] lnet_shutdown_lndnets+0x11e/0x2f0 [lnet] [1208742.803870] ? lnet_user_add_peer_ni+0x20/0x20 [lnet] [1208742.805847] LNetNIFini+0xcc/0x150 [lnet] [1208742.807593] ptlrpc_exit+0x1e/0x756 [ptlrpc] [1208742.809524] __x64_sys_delete_module+0x141/0x260 [1208742.811420] do_syscall_64+0x5b/0x1a0 [1208742.813109] entry_SYSCALL_64_after_hwframe+0x66/0xcb [1208742.815097] RIP: 0033:0x7f5b80eeac2b
Attachments
Issue Links
- duplicates
-
EX-14204 Loading...