scsi: lpfc: nvmet: Avoid hang / use-after-free again when destroying …

…targetport

We cannot wait on a completion object in the lpfc_nvme_targetport structure
in the _destroy_targetport() code path because the NVMe/fc transport will
free that structure immediately after the .targetport_delete() callback.
This results in a use-after-free, and a crash if slub_debug=FZPU is
enabled.

An earlier fix put put the completion on the stack, but commit 2a0fb34
("scsi: lpfc: Correct localport timeout duration error") subsequently
changed the code to reference the completion through a pointer in the
object rather than the local stack variable.  Fix this by using the stack
variable directly.

Link: https://lore.kernel.org/r/20200729231011.13240-1-emilne@redhat.com
Fixes: 2a0fb34 ("scsi: lpfc: Correct localport timeout duration error")
Reviewed-by: James Smart <james.smart@broadcom.com>
Signed-off-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

drivers/scsi/lpfc/lpfc_nvmet.c

@@ -2110,7 +2110,7 @@ lpfc_nvmet_destroy_targetport(struct lpfc_hba *phba)
 		}
 		tgtp->tport_unreg_cmp = &tport_unreg_cmp;
 		nvmet_fc_unregister_targetport(phba->targetport);
-		if (!wait_for_completion_timeout(tgtp->tport_unreg_cmp,
+		if (!wait_for_completion_timeout(&tport_unreg_cmp,
 					msecs_to_jiffies(LPFC_NVMET_WAIT_TMO)))
 			lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
 					"6179 Unreg targetport x%px timeout "