scsi: lpfc: Fix RQ empty firmware trap

When nvme target deferred receive logic waits for exchange resources,
the corresponding receive buffer is not replenished with the hardware.
This can result in a lack of asynchronous receive buffer resources in
the hardware, resulting in a "2885 Port Status Event: ... error
1=0x52004a01 ..." message.

Correct by replenishing the buffer whenenver the deferred logic kicks
in.  Update corresponding debug messages and statistics as well.

[mkp: applied by hand]

Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

drivers/scsi/lpfc/lpfc_attr.c

@@ -259,6 +259,12 @@ lpfc_nvme_info_show(struct device *dev, struct device_attribute *attr,
 				atomic_read(&tgtp->xmt_abort_rsp),
 				atomic_read(&tgtp->xmt_abort_rsp_error));
 
+		len += snprintf(buf + len, PAGE_SIZE - len,
+				"DELAY: ctx %08x  fod %08x wqfull %08x\n",
+				atomic_read(&tgtp->defer_ctx),
+				atomic_read(&tgtp->defer_fod),
+				atomic_read(&tgtp->defer_wqfull));
+
 		/* Calculate outstanding IOs */
 		tot = atomic_read(&tgtp->rcv_fcp_cmd_drop);
 		tot += atomic_read(&tgtp->xmt_fcp_release);

drivers/scsi/lpfc/lpfc_mem.c

@@ -753,12 +753,16 @@ lpfc_rq_buf_free(struct lpfc_hba *phba, struct lpfc_dmabuf *mp)
 	drqe.address_hi = putPaddrHigh(rqb_entry->dbuf.phys);
 	rc = lpfc_sli4_rq_put(rqb_entry->hrq, rqb_entry->drq, &hrqe, &drqe);
 	if (rc < 0) {
+		(rqbp->rqb_free_buffer)(phba, rqb_entry);
 		lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
-				"6409 Cannot post to RQ %d: %x %x\n",
+				"6409 Cannot post to HRQ %d: %x %x %x "
+				"DRQ %x %x\n",
 				rqb_entry->hrq->queue_id,
 				rqb_entry->hrq->host_index,
-				rqb_entry->hrq->hba_index);
+				rqb_entry->hrq->hba_index,
-		(rqbp->rqb_free_buffer)(phba, rqb_entry);
+				rqb_entry->hrq->entry_count,
+				rqb_entry->drq->host_index,
+				rqb_entry->drq->hba_index);
 	} else {
 		list_add_tail(&rqb_entry->hbuf.list, &rqbp->rqb_buffer_list);
 		rqbp->buffer_count++;

drivers/scsi/lpfc/lpfc_nvmet.c

@@ -270,8 +270,6 @@ lpfc_nvmet_ctxbuf_post(struct lpfc_hba *phba, struct lpfc_nvmet_ctxbuf *ctx_buf)
 					 "NVMET RCV BUSY: xri x%x sz %d "
 					 "from %06x\n",
 					 oxid, size, sid);
-			/* defer repost rcv buffer till .defer_rcv callback */
-			ctxp->flag &= ~LPFC_NVMET_DEFER_RCV_REPOST;
 			atomic_inc(&tgtp->rcv_fcp_cmd_out);
 			return;
 		}
@@ -837,6 +835,7 @@ lpfc_nvmet_xmt_fcp_op(struct nvmet_fc_target_port *tgtport,
 		list_add_tail(&nvmewqeq->list, &wq->wqfull_list);
 		wq->q_flag |= HBA_NVMET_WQFULL;
 		spin_unlock_irqrestore(&pring->ring_lock, iflags);
+		atomic_inc(&lpfc_nvmep->defer_wqfull);
 		return 0;
 	}
 
@@ -975,11 +974,9 @@ lpfc_nvmet_defer_rcv(struct nvmet_fc_target_port *tgtport,
 
 	tgtp = phba->targetport->private;
 	atomic_inc(&tgtp->rcv_fcp_cmd_defer);
-	if (ctxp->flag & LPFC_NVMET_DEFER_RCV_REPOST)
+
-		lpfc_rq_buf_free(phba, &nvmebuf->hbuf); /* repost */
+	/* Free the nvmebuf since a new buffer already replaced it */
-	else
+	nvmebuf->hrq->rqbp->rqb_free_buffer(phba, nvmebuf);
-		nvmebuf->hrq->rqbp->rqb_free_buffer(phba, nvmebuf);
-	ctxp->flag &= ~LPFC_NVMET_DEFER_RCV_REPOST;
 }
 
 static struct nvmet_fc_target_template lpfc_tgttemplate = {
@@ -1309,6 +1306,9 @@ lpfc_nvmet_create_targetport(struct lpfc_hba *phba)
 		atomic_set(&tgtp->xmt_abort_sol, 0);
 		atomic_set(&tgtp->xmt_abort_rsp, 0);
 		atomic_set(&tgtp->xmt_abort_rsp_error, 0);
+		atomic_set(&tgtp->defer_ctx, 0);
+		atomic_set(&tgtp->defer_fod, 0);
+		atomic_set(&tgtp->defer_wqfull, 0);
 	}
 	return error;
 }
@@ -1810,6 +1810,8 @@ lpfc_nvmet_unsol_fcp_buffer(struct lpfc_hba *phba,
 	lpfc_nvmeio_data(phba, "NVMET FCP  RCV: xri x%x sz %d CPU %02x\n",
 			 oxid, size, smp_processor_id());
 
+	tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private;
+
 	if (!ctx_buf) {
 		/* Queue this NVME IO to process later */
 		spin_lock_irqsave(&phba->sli4_hba.nvmet_io_wait_lock, iflag);
@@ -1825,10 +1827,11 @@ lpfc_nvmet_unsol_fcp_buffer(struct lpfc_hba *phba,
 		lpfc_post_rq_buffer(
 			phba, phba->sli4_hba.nvmet_mrq_hdr[qno],
 			phba->sli4_hba.nvmet_mrq_data[qno], 1, qno);
+
+		atomic_inc(&tgtp->defer_ctx);
 		return;
 	}
 
-	tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private;
 	payload = (uint32_t *)(nvmebuf->dbuf.virt);
 	sid = sli4_sid_from_fc_hdr(fc_hdr);
 
@@ -1892,12 +1895,20 @@ lpfc_nvmet_unsol_fcp_buffer(struct lpfc_hba *phba,
 
 	/* Processing of FCP command is deferred */
 	if (rc == -EOVERFLOW) {
+		/*
+		 * Post a brand new DMA buffer to RQ and defer
+		 * freeing rcv buffer till .defer_rcv callback
+		 */
+		qno = nvmebuf->idx;
+		lpfc_post_rq_buffer(
+			phba, phba->sli4_hba.nvmet_mrq_hdr[qno],
+			phba->sli4_hba.nvmet_mrq_data[qno], 1, qno);
+
 		lpfc_nvmeio_data(phba,
 				 "NVMET RCV BUSY: xri x%x sz %d from %06x\n",
 				 oxid, size, sid);
-		/* defer reposting rcv buffer till .defer_rcv callback */
-		ctxp->flag |= LPFC_NVMET_DEFER_RCV_REPOST;
 		atomic_inc(&tgtp->rcv_fcp_cmd_out);
+		atomic_inc(&tgtp->defer_fod);
 		return;
 	}
 	ctxp->rqb_buffer = nvmebuf;

drivers/scsi/lpfc/lpfc_nvmet.h

@@ -72,7 +72,6 @@ struct lpfc_nvmet_tgtport {
 	atomic_t xmt_fcp_rsp_aborted;
 	atomic_t xmt_fcp_rsp_drop;
 
-
 	/* Stats counters - lpfc_nvmet_xmt_fcp_abort */
 	atomic_t xmt_fcp_xri_abort_cqe;
 	atomic_t xmt_fcp_abort;
@@ -81,6 +80,11 @@ struct lpfc_nvmet_tgtport {
 	atomic_t xmt_abort_unsol;
 	atomic_t xmt_abort_rsp;
 	atomic_t xmt_abort_rsp_error;
+
+	/* Stats counters - defer IO */
+	atomic_t defer_ctx;
+	atomic_t defer_fod;
+	atomic_t defer_wqfull;
 };
 
 struct lpfc_nvmet_ctx_info {
@@ -131,7 +135,6 @@ struct lpfc_nvmet_rcv_ctx {
 #define LPFC_NVMET_XBUSY		0x4  /* XB bit set on IO cmpl */
 #define LPFC_NVMET_CTX_RLS		0x8  /* ctx free requested */
 #define LPFC_NVMET_ABTS_RCV		0x10  /* ABTS received on exchange */
-#define LPFC_NVMET_DEFER_RCV_REPOST	0x20  /* repost to RQ on defer rcv */
 #define LPFC_NVMET_DEFER_WQFULL		0x40  /* Waiting on a free WQE */
 	struct rqb_dmabuf *rqb_buffer;
 	struct lpfc_nvmet_ctxbuf *ctxbuf;

drivers/scsi/lpfc/lpfc_sli.c

@@ -6535,9 +6535,11 @@ lpfc_post_rq_buffer(struct lpfc_hba *phba, struct lpfc_queue *hrq,
 	struct lpfc_rqe hrqe;
 	struct lpfc_rqe drqe;
 	struct lpfc_rqb *rqbp;
+	unsigned long flags;
 	struct rqb_dmabuf *rqb_buffer;
 	LIST_HEAD(rqb_buf_list);
 
+	spin_lock_irqsave(&phba->hbalock, flags);
 	rqbp = hrq->rqbp;
 	for (i = 0; i < count; i++) {
 		/* IF RQ is already full, don't bother */
@@ -6561,6 +6563,15 @@ lpfc_post_rq_buffer(struct lpfc_hba *phba, struct lpfc_queue *hrq,
 		drqe.address_hi = putPaddrHigh(rqb_buffer->dbuf.phys);
 		rc = lpfc_sli4_rq_put(hrq, drq, &hrqe, &drqe);
 		if (rc < 0) {
+			lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
+					"6421 Cannot post to HRQ %d: %x %x %x "
+					"DRQ %x %x\n",
+					hrq->queue_id,
+					hrq->host_index,
+					hrq->hba_index,
+					hrq->entry_count,
+					drq->host_index,
+					drq->hba_index);
 			rqbp->rqb_free_buffer(phba, rqb_buffer);
 		} else {
 			list_add_tail(&rqb_buffer->hbuf.list,
@@ -6568,6 +6579,7 @@ lpfc_post_rq_buffer(struct lpfc_hba *phba, struct lpfc_queue *hrq,
 			rqbp->buffer_count++;
 		}
 	}
+	spin_unlock_irqrestore(&phba->hbalock, flags);
 	return 1;
 }