Kernel is crashing with the following stacktrace:
  BUG: unable to handle kernel NULL pointer dereference at
    
00000000000005bc
  IP: lpfc_nvme_register_port+0x1a8/0x3a0 [lpfc]
  ...
  Call Trace:
  lpfc_nlp_state_cleanup+0x2b2/0x500 [lpfc]
  lpfc_nlp_set_state+0xd7/0x1a0 [lpfc]
  lpfc_cmpl_prli_prli_issue+0x1f7/0x450 [lpfc]
  lpfc_disc_state_machine+0x7a/0x1e0 [lpfc]
  lpfc_cmpl_els_prli+0x16f/0x1e0 [lpfc]
  lpfc_sli_sp_handle_rspiocb+0x5b2/0x690 [lpfc]
  lpfc_sli_handle_slow_ring_event_s4+0x182/0x230 [lpfc]
  lpfc_do_work+0x87f/0x1570 [lpfc]
  kthread+0x10d/0x130
  ret_from_fork+0x35/0x40
During target side fault injections, it is possible to hit the
NLP_WAIT_FOR_UNREG case in lpfc_nvme_remoteport_delete. A prior commit
fixed a rebind and delete race condition, but called lpfc_nlp_put
unconditionally. This triggered a deletion and the crash.
Fix by movng nlp_put to inside the NLP_WAIT_FOR_UNREG case, where the nlp
will be being unregistered/removed. Leave the reference if the flag isn't
set.
Link: https://lore.kernel.org/r/20200322181304.37655-8-jsmart2021@gmail.com
Fixes: b15bd3e6212e ("scsi: lpfc: Fix nvme remoteport registration race conditions")
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
 
        if (ndlp->upcall_flags & NLP_WAIT_FOR_UNREG) {
                ndlp->nrport = NULL;
                ndlp->upcall_flags &= ~NLP_WAIT_FOR_UNREG;
-       }
-       spin_unlock_irq(&vport->phba->hbalock);
+               spin_unlock_irq(&vport->phba->hbalock);
 
-       /* Remove original register reference. The host transport
-        * won't reference this rport/remoteport any further.
-        */
-       lpfc_nlp_put(ndlp);
+               /* Remove original register reference. The host transport
+                * won't reference this rport/remoteport any further.
+                */
+               lpfc_nlp_put(ndlp);
+       } else {
+               spin_unlock_irq(&vport->phba->hbalock);
+       }
 
  rport_err:
        return;