scsi: lpfc: Merge per-protocol WQ/CQ pairs into single per-cpu pair

Currently, each hardware queue, typically allocated per-cpu, consists of a
WQ/CQ pair per protocol. Meaning if both SCSI and NVMe are supported 2
WQ/CQ pairs will exist for the hardware queue. Separate queues are
unnecessary. The current implementation wastes memory backing the 2nd set
of queues, and the use of double the SLI-4 WQ/CQ's means less hardware
queues can be supported which means there may not always be enough to have
a pair per cpu. If there is only 1 pair per cpu, more cpu's may get their
own WQ/CQ.

Rework the implementation to use a single WQ/CQ pair by both protocols.

Signed-off-by: Dick Kennedy <[email protected]>
Signed-off-by: James Smart <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>

Author: jsmart-gh

drivers/scsi/lpfc/lpfc.h

@@ -734,14 +734,13 @@ struct lpfc_hba {
 #define HBA_AER_ENABLED		0x1000 /* AER enabled with HBA */
 #define HBA_DEVLOSS_TMO         0x2000 /* HBA in devloss timeout */
 #define HBA_RRQ_ACTIVE		0x4000 /* process the rrq active list */
-#define HBA_FCP_IOQ_FLUSH	0x8000 /* FCP I/O queues being flushed */
+#define HBA_IOQ_FLUSH		0x8000 /* FCP/NVME I/O queues being flushed */
 #define HBA_FW_DUMP_OP		0x10000 /* Skips fn reset before FW dump */
 #define HBA_RECOVERABLE_UE	0x20000 /* Firmware supports recoverable UE */
 #define HBA_FORCED_LINK_SPEED	0x40000 /*
 					 * Firmware supports Forced Link Speed
 					 * capability
 					 */
-#define HBA_NVME_IOQ_FLUSH      0x80000 /* NVME IO queues flushed. */
 #define HBA_FLOGI_ISSUED	0x100000 /* FLOGI was issued */
 
 	uint32_t fcp_ring_in_use; /* When polling test if intr-hndlr active*/

drivers/scsi/lpfc/lpfc_crtn.h

@@ -326,7 +326,7 @@ void lpfc_sli_bemem_bcopy(void *, void *, uint32_t);
 void lpfc_sli_abort_iocb_ring(struct lpfc_hba *, struct lpfc_sli_ring *);
 void lpfc_sli_abort_fcp_rings(struct lpfc_hba *phba);
 void lpfc_sli_hba_iocb_abort(struct lpfc_hba *);
-void lpfc_sli_flush_fcp_rings(struct lpfc_hba *);
+void lpfc_sli_flush_io_rings(struct lpfc_hba *phba);
 int lpfc_sli_ringpostbuf_put(struct lpfc_hba *, struct lpfc_sli_ring *,
 			     struct lpfc_dmabuf *);
 struct lpfc_dmabuf *lpfc_sli_ringpostbuf_get(struct lpfc_hba *,

drivers/scsi/lpfc/lpfc_debugfs.c

@@ -416,8 +416,7 @@ lpfc_debugfs_commonxripools_data(struct lpfc_hba *phba, char *buf, int size)
 		qp = &phba->sli4_hba.hdwq[lpfc_debugfs_last_xripool];
 
 		len += scnprintf(buf + len, size - len, "HdwQ %d Info ", i);
-		spin_lock_irqsave(&qp->abts_scsi_buf_list_lock, iflag);
-		spin_lock(&qp->abts_nvme_buf_list_lock);
+		spin_lock_irqsave(&qp->abts_io_buf_list_lock, iflag);
 		spin_lock(&qp->io_buf_list_get_lock);
 		spin_lock(&qp->io_buf_list_put_lock);
 		out = qp->total_io_bufs - (qp->get_io_bufs + qp->put_io_bufs +
@@ -430,8 +429,7 @@ lpfc_debugfs_commonxripools_data(struct lpfc_hba *phba, char *buf, int size)
 			qp->abts_nvme_io_bufs, out);
 		spin_unlock(&qp->io_buf_list_put_lock);
 		spin_unlock(&qp->io_buf_list_get_lock);
-		spin_unlock(&qp->abts_nvme_buf_list_lock);
-		spin_unlock_irqrestore(&qp->abts_scsi_buf_list_lock, iflag);
+		spin_unlock_irqrestore(&qp->abts_io_buf_list_lock, iflag);
 
 		lpfc_debugfs_last_xripool++;
 		if (lpfc_debugfs_last_xripool >= phba->cfg_hdw_queue)
@@ -533,9 +531,7 @@ lpfc_debugfs_multixripools_data(struct lpfc_hba *phba, char *buf, int size)
 			continue;
 		pbl_pool = &multixri_pool->pbl_pool;
 		pvt_pool = &multixri_pool->pvt_pool;
-		txcmplq_cnt = qp->fcp_wq->pring->txcmplq_cnt;
-		if (qp->nvme_wq)
-			txcmplq_cnt += qp->nvme_wq->pring->txcmplq_cnt;
+		txcmplq_cnt = qp->io_wq->pring->txcmplq_cnt;
 
 		scnprintf(tmp, sizeof(tmp),
 			  "%03d: %4d %4d %4d %4d | %10d %10d ",
@@ -3786,23 +3782,13 @@ lpfc_idiag_wqs_for_cq(struct lpfc_hba *phba, char *wqtype, char *pbuffer,
 	int qidx;
 
 	for (qidx = 0; qidx < phba->cfg_hdw_queue; qidx++) {
-		qp = phba->sli4_hba.hdwq[qidx].fcp_wq;
+		qp = phba->sli4_hba.hdwq[qidx].io_wq;
 		if (qp->assoc_qid != cq_id)
 			continue;
 		*len = __lpfc_idiag_print_wq(qp, wqtype, pbuffer, *len);
 		if (*len >= max_cnt)
 			return 1;
 	}
-	if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) {
-		for (qidx = 0; qidx < phba->cfg_hdw_queue; qidx++) {
-			qp = phba->sli4_hba.hdwq[qidx].nvme_wq;
-			if (qp->assoc_qid != cq_id)
-				continue;
-			*len = __lpfc_idiag_print_wq(qp, wqtype, pbuffer, *len);
-			if (*len >= max_cnt)
-				return 1;
-		}
-	}
 	return 0;
 }
 
@@ -3868,38 +3854,21 @@ lpfc_idiag_cqs_for_eq(struct lpfc_hba *phba, char *pbuffer,
 	struct lpfc_queue *qp;
 	int rc;
 
-	qp = phba->sli4_hba.hdwq[eqidx].fcp_cq;
+	qp = phba->sli4_hba.hdwq[eqidx].io_cq;
 
-	*len = __lpfc_idiag_print_cq(qp, "FCP", pbuffer, *len);
+	*len = __lpfc_idiag_print_cq(qp, "IO", pbuffer, *len);
 
 	/* Reset max counter */
 	qp->CQ_max_cqe = 0;
 
 	if (*len >= max_cnt)
 		return 1;
 
-	rc = lpfc_idiag_wqs_for_cq(phba, "FCP", pbuffer, len,
+	rc = lpfc_idiag_wqs_for_cq(phba, "IO", pbuffer, len,
 				   max_cnt, qp->queue_id);
 	if (rc)
 		return 1;
 
-	if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) {
-		qp = phba->sli4_hba.hdwq[eqidx].nvme_cq;
-
-		*len = __lpfc_idiag_print_cq(qp, "NVME", pbuffer, *len);
-
-		/* Reset max counter */
-		qp->CQ_max_cqe = 0;
-
-		if (*len >= max_cnt)
-			return 1;
-
-		rc = lpfc_idiag_wqs_for_cq(phba, "NVME", pbuffer, len,
-					   max_cnt, qp->queue_id);
-		if (rc)
-			return 1;
-	}
-
 	if ((eqidx < phba->cfg_nvmet_mrq) && phba->nvmet_support) {
 		/* NVMET CQset */
 		qp = phba->sli4_hba.nvmet_cqset[eqidx];
@@ -4348,7 +4317,7 @@ lpfc_idiag_queacc_write(struct file *file, const char __user *buf,
 		if (phba->sli4_hba.hdwq) {
 			for (qidx = 0; qidx < phba->cfg_hdw_queue;
 								qidx++) {
-				qp = phba->sli4_hba.hdwq[qidx].fcp_cq;
+				qp = phba->sli4_hba.hdwq[qidx].io_cq;
 				if (qp && qp->queue_id == queid) {
 					/* Sanity check */
 					rc = lpfc_idiag_que_param_check(
@@ -4360,22 +4329,6 @@ lpfc_idiag_queacc_write(struct file *file, const char __user *buf,
 				}
 			}
 		}
-		/* NVME complete queue */
-		if (phba->sli4_hba.hdwq) {
-			qidx = 0;
-			do {
-				qp = phba->sli4_hba.hdwq[qidx].nvme_cq;
-				if (qp && qp->queue_id == queid) {
-					/* Sanity check */
-					rc = lpfc_idiag_que_param_check(
-						qp, index, count);
-					if (rc)
-						goto error_out;
-					idiag.ptr_private = qp;
-					goto pass_check;
-				}
-			} while (++qidx < phba->cfg_hdw_queue);
-		}
 		goto error_out;
 		break;
 	case LPFC_IDIAG_MQ:
@@ -4419,20 +4372,7 @@ lpfc_idiag_queacc_write(struct file *file, const char __user *buf,
 		if (phba->sli4_hba.hdwq) {
 			/* FCP/SCSI work queue */
 			for (qidx = 0; qidx < phba->cfg_hdw_queue; qidx++) {
-				qp = phba->sli4_hba.hdwq[qidx].fcp_wq;
-				if (qp && qp->queue_id == queid) {
-					/* Sanity check */
-					rc = lpfc_idiag_que_param_check(
-						qp, index, count);
-					if (rc)
-						goto error_out;
-					idiag.ptr_private = qp;
-					goto pass_check;
-				}
-			}
-			/* NVME work queue */
-			for (qidx = 0; qidx < phba->cfg_hdw_queue; qidx++) {
-				qp = phba->sli4_hba.hdwq[qidx].nvme_wq;
+				qp = phba->sli4_hba.hdwq[qidx].io_wq;
 				if (qp && qp->queue_id == queid) {
 					/* Sanity check */
 					rc = lpfc_idiag_que_param_check(
@@ -6442,12 +6382,7 @@ lpfc_debug_dump_all_queues(struct lpfc_hba *phba)
 	lpfc_debug_dump_wq(phba, DUMP_NVMELS, 0);
 
 	for (idx = 0; idx < phba->cfg_hdw_queue; idx++)
-		lpfc_debug_dump_wq(phba, DUMP_FCP, idx);
-
-	if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) {
-		for (idx = 0; idx < phba->cfg_hdw_queue; idx++)
-			lpfc_debug_dump_wq(phba, DUMP_NVME, idx);
-	}
+		lpfc_debug_dump_wq(phba, DUMP_IO, idx);
 
 	lpfc_debug_dump_hdr_rq(phba);
 	lpfc_debug_dump_dat_rq(phba);
@@ -6459,12 +6394,7 @@ lpfc_debug_dump_all_queues(struct lpfc_hba *phba)
 	lpfc_debug_dump_cq(phba, DUMP_NVMELS, 0);
 
 	for (idx = 0; idx < phba->cfg_hdw_queue; idx++)
-		lpfc_debug_dump_cq(phba, DUMP_FCP, idx);
-
-	if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) {
-		for (idx = 0; idx < phba->cfg_hdw_queue; idx++)
-			lpfc_debug_dump_cq(phba, DUMP_NVME, idx);
-	}
+		lpfc_debug_dump_cq(phba, DUMP_IO, idx);
 
 	/*
 	 * Dump Event Queues (EQs)

drivers/scsi/lpfc/lpfc_debugfs.h

@@ -291,8 +291,7 @@ struct lpfc_idiag {
 #define LPFC_DUMP_MULTIXRIPOOL_SIZE 8192
 
 enum {
-	DUMP_FCP,
-	DUMP_NVME,
+	DUMP_IO,
 	DUMP_MBX,
 	DUMP_ELS,
 	DUMP_NVMELS,
@@ -415,12 +414,9 @@ lpfc_debug_dump_wq(struct lpfc_hba *phba, int qtype, int wqidx)
 	struct lpfc_queue *wq;
 	char *qtypestr;
 
-	if (qtype == DUMP_FCP) {
-		wq = phba->sli4_hba.hdwq[wqidx].fcp_wq;
-		qtypestr = "FCP";
-	} else if (qtype == DUMP_NVME) {
-		wq = phba->sli4_hba.hdwq[wqidx].nvme_wq;
-		qtypestr = "NVME";
+	if (qtype == DUMP_IO) {
+		wq = phba->sli4_hba.hdwq[wqidx].io_wq;
+		qtypestr = "IO";
 	} else if (qtype == DUMP_MBX) {
 		wq = phba->sli4_hba.mbx_wq;
 		qtypestr = "MBX";
@@ -433,7 +429,7 @@ lpfc_debug_dump_wq(struct lpfc_hba *phba, int qtype, int wqidx)
 	} else
 		return;
 
-	if (qtype == DUMP_FCP || qtype == DUMP_NVME)
+	if (qtype == DUMP_IO)
 		pr_err("%s WQ: WQ[Idx:%d|Qid:%d]\n",
 			qtypestr, wqidx, wq->queue_id);
 	else
@@ -459,17 +455,13 @@ lpfc_debug_dump_cq(struct lpfc_hba *phba, int qtype, int wqidx)
 	char *qtypestr;
 	int eqidx;
 
-	/* fcp/nvme wq and cq are 1:1, thus same indexes */
+	/* io wq and cq are 1:1, thus same indexes */
 	eq = NULL;
 
-	if (qtype == DUMP_FCP) {
-		wq = phba->sli4_hba.hdwq[wqidx].fcp_wq;
-		cq = phba->sli4_hba.hdwq[wqidx].fcp_cq;
-		qtypestr = "FCP";
-	} else if (qtype == DUMP_NVME) {
-		wq = phba->sli4_hba.hdwq[wqidx].nvme_wq;
-		cq = phba->sli4_hba.hdwq[wqidx].nvme_cq;
-		qtypestr = "NVME";
+	if (qtype == DUMP_IO) {
+		wq = phba->sli4_hba.hdwq[wqidx].io_wq;
+		cq = phba->sli4_hba.hdwq[wqidx].io_cq;
+		qtypestr = "IO";
 	} else if (qtype == DUMP_MBX) {
 		wq = phba->sli4_hba.mbx_wq;
 		cq = phba->sli4_hba.mbx_cq;
@@ -496,7 +488,7 @@ lpfc_debug_dump_cq(struct lpfc_hba *phba, int qtype, int wqidx)
 		eq = phba->sli4_hba.hdwq[0].hba_eq;
 	}
 
-	if (qtype == DUMP_FCP || qtype == DUMP_NVME)
+	if (qtype == DUMP_IO)
 		pr_err("%s CQ: WQ[Idx:%d|Qid%d]->CQ[Idx%d|Qid%d]"
 			"->EQ[Idx:%d|Qid:%d]:\n",
 			qtypestr, wqidx, wq->queue_id, wqidx, cq->queue_id,
@@ -572,20 +564,11 @@ lpfc_debug_dump_wq_by_id(struct lpfc_hba *phba, int qid)
 	int wq_idx;
 
 	for (wq_idx = 0; wq_idx < phba->cfg_hdw_queue; wq_idx++)
-		if (phba->sli4_hba.hdwq[wq_idx].fcp_wq->queue_id == qid)
+		if (phba->sli4_hba.hdwq[wq_idx].io_wq->queue_id == qid)
 			break;
 	if (wq_idx < phba->cfg_hdw_queue) {
-		pr_err("FCP WQ[Idx:%d|Qid:%d]\n", wq_idx, qid);
-		lpfc_debug_dump_q(phba->sli4_hba.hdwq[wq_idx].fcp_wq);
-		return;
-	}
-
-	for (wq_idx = 0; wq_idx < phba->cfg_hdw_queue; wq_idx++)
-		if (phba->sli4_hba.hdwq[wq_idx].nvme_wq->queue_id == qid)
-			break;
-	if (wq_idx < phba->cfg_hdw_queue) {
-		pr_err("NVME WQ[Idx:%d|Qid:%d]\n", wq_idx, qid);
-		lpfc_debug_dump_q(phba->sli4_hba.hdwq[wq_idx].nvme_wq);
+		pr_err("IO WQ[Idx:%d|Qid:%d]\n", wq_idx, qid);
+		lpfc_debug_dump_q(phba->sli4_hba.hdwq[wq_idx].io_wq);
 		return;
 	}
 
@@ -654,22 +637,12 @@ lpfc_debug_dump_cq_by_id(struct lpfc_hba *phba, int qid)
 	int cq_idx;
 
 	for (cq_idx = 0; cq_idx < phba->cfg_hdw_queue; cq_idx++)
-		if (phba->sli4_hba.hdwq[cq_idx].fcp_cq->queue_id == qid)
-			break;
-
-	if (cq_idx < phba->cfg_hdw_queue) {
-		pr_err("FCP CQ[Idx:%d|Qid:%d]\n", cq_idx, qid);
-		lpfc_debug_dump_q(phba->sli4_hba.hdwq[cq_idx].fcp_cq);
-		return;
-	}
-
-	for (cq_idx = 0; cq_idx < phba->cfg_hdw_queue; cq_idx++)
-		if (phba->sli4_hba.hdwq[cq_idx].nvme_cq->queue_id == qid)
+		if (phba->sli4_hba.hdwq[cq_idx].io_cq->queue_id == qid)
 			break;
 
 	if (cq_idx < phba->cfg_hdw_queue) {
-		pr_err("NVME CQ[Idx:%d|Qid:%d]\n", cq_idx, qid);
-		lpfc_debug_dump_q(phba->sli4_hba.hdwq[cq_idx].nvme_cq);
+		pr_err("IO CQ[Idx:%d|Qid:%d]\n", cq_idx, qid);
+		lpfc_debug_dump_q(phba->sli4_hba.hdwq[cq_idx].io_cq);
 		return;
 	}