scsi: lpfc: Fix NMI crash during rmmod due to circular hbalock dependency

Remove hbalock dependency for lpfc_abts_els_sgl_list and
lpfc_abts_nvmet_ctx_list.  The lists are adaquately synchronized with the
sgl_list_lock and abts_nvmet_buf_list_lock.

Link: https://lore.kernel.org/r/20210412013127.2387-5-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1 files changed, 11 insertions, 15 deletions

diff --git a/drivers/scsi/lpfc/lpfc_nvmet.c b/drivers/scsi/lpfc/lpfc_nvmet.c
index c84da8e6b65d..f2d9a3580887 100644
--- a/drivers/scsi/lpfc/lpfc_nvmet.c
+++ b/drivers/scsi/lpfc/lpfc_nvmet.c
@@ -1440,7 +1440,10 @@ __lpfc_nvmet_clean_io_for_cpu(struct lpfc_hba *phba,
 		list_del_init(&ctx_buf->list);
 		spin_unlock(&phba->sli4_hba.abts_nvmet_buf_list_lock);
 
+		spin_lock(&phba->hbalock);
 		__lpfc_clear_active_sglq(phba, ctx_buf->sglq->sli4_lxritag);
+		spin_unlock(&phba->hbalock);
+
 		ctx_buf->sglq->state = SGL_FREED;
 		ctx_buf->sglq->ndlp = NULL;
 
@@ -1787,8 +1790,7 @@ lpfc_sli4_nvmet_xri_aborted(struct lpfc_hba *phba,
 		atomic_inc(&tgtp->xmt_fcp_xri_abort_cqe);
 	}
 
-	spin_lock_irqsave(&phba->hbalock, iflag);
-	spin_lock(&phba->sli4_hba.abts_nvmet_buf_list_lock);
+	spin_lock_irqsave(&phba->sli4_hba.abts_nvmet_buf_list_lock, iflag);
 	list_for_each_entry_safe(ctxp, next_ctxp,
 				 &phba->sli4_hba.lpfc_abts_nvmet_ctx_list,
 				 list) {
@@ -1806,10 +1808,10 @@ lpfc_sli4_nvmet_xri_aborted(struct lpfc_hba *phba,
 		}
 		ctxp->flag &= ~LPFC_NVME_XBUSY;
 		spin_unlock(&ctxp->ctxlock);
-		spin_unlock(&phba->sli4_hba.abts_nvmet_buf_list_lock);
+		spin_unlock_irqrestore(&phba->sli4_hba.abts_nvmet_buf_list_lock,
+				       iflag);
 
 		rrq_empty = list_empty(&phba->active_rrq_list);
-		spin_unlock_irqrestore(&phba->hbalock, iflag);
 		ndlp = lpfc_findnode_did(phba->pport, ctxp->sid);
 		if (ndlp &&
 		    (ndlp->nlp_state == NLP_STE_UNMAPPED_NODE ||
@@ -1830,9 +1832,7 @@ lpfc_sli4_nvmet_xri_aborted(struct lpfc_hba *phba,
 			lpfc_worker_wake_up(phba);
 		return;
 	}
-	spin_unlock(&phba->sli4_hba.abts_nvmet_buf_list_lock);
-	spin_unlock_irqrestore(&phba->hbalock, iflag);
-
+	spin_unlock_irqrestore(&phba->sli4_hba.abts_nvmet_buf_list_lock, iflag);
 	ctxp = lpfc_nvmet_get_ctx_for_xri(phba, xri);
 	if (ctxp) {
 		/*
@@ -1876,8 +1876,7 @@ lpfc_nvmet_rcv_unsol_abort(struct lpfc_vport *vport,
 	sid = sli4_sid_from_fc_hdr(fc_hdr);
 	oxid = be16_to_cpu(fc_hdr->fh_ox_id);
 
-	spin_lock_irqsave(&phba->hbalock, iflag);
-	spin_lock(&phba->sli4_hba.abts_nvmet_buf_list_lock);
+	spin_lock_irqsave(&phba->sli4_hba.abts_nvmet_buf_list_lock, iflag);
 	list_for_each_entry_safe(ctxp, next_ctxp,
 				 &phba->sli4_hba.lpfc_abts_nvmet_ctx_list,
 				 list) {
@@ -1886,9 +1885,8 @@ lpfc_nvmet_rcv_unsol_abort(struct lpfc_vport *vport,
 
 		xri = ctxp->ctxbuf->sglq->sli4_xritag;
 
-		spin_unlock(&phba->sli4_hba.abts_nvmet_buf_list_lock);
-		spin_unlock_irqrestore(&phba->hbalock, iflag);
-
+		spin_unlock_irqrestore(&phba->sli4_hba.abts_nvmet_buf_list_lock,
+				       iflag);
 		spin_lock_irqsave(&ctxp->ctxlock, iflag);
 		ctxp->flag |= LPFC_NVME_ABTS_RCV;
 		spin_unlock_irqrestore(&ctxp->ctxlock, iflag);
@@ -1907,9 +1905,7 @@ lpfc_nvmet_rcv_unsol_abort(struct lpfc_vport *vport,
 		lpfc_sli4_seq_abort_rsp(vport, fc_hdr, 1);
 		return 0;
 	}
-	spin_unlock(&phba->sli4_hba.abts_nvmet_buf_list_lock);
-	spin_unlock_irqrestore(&phba->hbalock, iflag);
-
+	spin_unlock_irqrestore(&phba->sli4_hba.abts_nvmet_buf_list_lock, iflag);
 	/* check the wait list */
 	if (phba->sli4_hba.nvmet_io_wait_cnt) {
 		struct rqb_dmabuf *nvmebuf;