[PATCH v3 16/26] net/bnxt: reorg the query stats code

[Ajit Khaparde <ajit.khaparde@broadcom.com>]

1) Use hwrm_stat_ctx_query command to query statistics
Using hwrm_stat_ctx_query command will allow polling
the statistics from hardware instead of using the current push
model from the hardware which does a DMA of the stats to the host
at fixed intervals.
2) Use the rx_mbuf_alloc_fail to track mbuf alloc failures.
3) We were wrongly incrementing hwrm_cmd_seq in bnxt_hwrm_stat_clear
and bnxt_hwrm_stat_ctx_alloc functions.  This patch fixes that.

Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

--
v1->v2: regroup related patches and incorporate other review comments

v2->v3:
  - rebasing to next-net tree
  - Use net/bnxt instead of just bnxt in patch subject
---
 drivers/net/bnxt/bnxt.h        |  1 +
 drivers/net/bnxt/bnxt_ethdev.c |  2 ++
 drivers/net/bnxt/bnxt_hwrm.c   | 82 ++++++++++++++++++++++++++++++++++++++----
 drivers/net/bnxt/bnxt_hwrm.h   |  4 +++
 drivers/net/bnxt/bnxt_rxr.c    | 16 ++++++---
 drivers/net/bnxt/bnxt_stats.c  | 59 ++++--------------------------
 6 files changed, 102 insertions(+), 62 deletions(-)

diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
index adf0ab1..e411704 100644
--- a/drivers/net/bnxt/bnxt.h
+++ b/drivers/net/bnxt/bnxt.h
@@ -204,6 +204,7 @@ struct bnxt {
 	uint16_t		vxlan_fw_dst_port_id;
 	uint16_t		geneve_fw_dst_port_id;
 	uint32_t		fw_ver;
+	rte_atomic64_t		rx_mbuf_alloc_fail;
 };
 
 /*
diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c
index 178e581..4576156 100644
--- a/drivers/net/bnxt/bnxt_ethdev.c
+++ b/drivers/net/bnxt/bnxt_ethdev.c
@@ -1616,6 +1616,8 @@ bnxt_dev_init(struct rte_eth_dev *eth_dev)
 	eth_dev->data->dev_flags |= RTE_ETH_DEV_DETACHABLE;
 
 	bp = eth_dev->data->dev_private;
+
+	rte_atomic64_init(&bp->rx_mbuf_alloc_fail);
 	bp->dev_stopped = 1;
 
 	if (bnxt_vf_pciid(pci_dev->id.device_id))
diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c
index 3e595c3..a1aa80e 100644
--- a/drivers/net/bnxt/bnxt_hwrm.c
+++ b/drivers/net/bnxt/bnxt_hwrm.c
@@ -821,13 +821,12 @@ int bnxt_hwrm_stat_clear(struct bnxt *bp, struct bnxt_cp_ring_info *cpr)
 	struct hwrm_stat_ctx_clr_stats_input req = {.req_type = 0 };
 	struct hwrm_stat_ctx_clr_stats_output *resp = bp->hwrm_cmd_resp_addr;
 
-	HWRM_PREP(req, STAT_CTX_CLR_STATS, -1, resp);
-
 	if (cpr->hw_stats_ctx_id == (uint32_t)HWRM_NA_SIGNATURE)
 		return rc;
 
+	HWRM_PREP(req, STAT_CTX_CLR_STATS, -1, resp);
+
 	req.stat_ctx_id = rte_cpu_to_le_16(cpr->hw_stats_ctx_id);
-	req.seq_id = rte_cpu_to_le_16(bp->hwrm_cmd_seq++);
 
 	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
 
@@ -845,9 +844,8 @@ int bnxt_hwrm_stat_ctx_alloc(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 
 	HWRM_PREP(req, STAT_CTX_ALLOC, -1, resp);
 
-	req.update_period_ms = rte_cpu_to_le_32(1000);
+	req.update_period_ms = rte_cpu_to_le_32(0);
 
-	req.seq_id = rte_cpu_to_le_16(bp->hwrm_cmd_seq++);
 	req.stats_dma_addr =
 	    rte_cpu_to_le_64(cpr->hw_stats_map);
 
@@ -870,7 +868,6 @@ int bnxt_hwrm_stat_ctx_free(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 	HWRM_PREP(req, STAT_CTX_FREE, -1, resp);
 
 	req.stat_ctx_id = rte_cpu_to_le_16(cpr->hw_stats_ctx_id);
-	req.seq_id = rte_cpu_to_le_16(bp->hwrm_cmd_seq++);
 
 	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
 
@@ -1227,6 +1224,43 @@ int bnxt_hwrm_func_vf_mac(struct bnxt *bp, uint16_t vf, const uint8_t *mac_addr)
 	return rc;
 }
 
+int bnxt_hwrm_func_qstats(struct bnxt *bp, uint16_t fid,
+			  struct rte_eth_stats *stats)
+{
+	int rc = 0;
+	struct hwrm_func_qstats_input req = {.req_type = 0};
+	struct hwrm_func_qstats_output *resp = bp->hwrm_cmd_resp_addr;
+
+	HWRM_PREP(req, FUNC_QSTATS, -1, resp);
+
+	req.fid = rte_cpu_to_le_16(fid);
+
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+
+	HWRM_CHECK_RESULT;
+
+	stats->ipackets = rte_le_to_cpu_64(resp->rx_ucast_pkts);
+	stats->ipackets += rte_le_to_cpu_64(resp->rx_mcast_pkts);
+	stats->ipackets += rte_le_to_cpu_64(resp->rx_bcast_pkts);
+	stats->ibytes = rte_le_to_cpu_64(resp->rx_ucast_bytes);
+	stats->ibytes += rte_le_to_cpu_64(resp->rx_mcast_bytes);
+	stats->ibytes += rte_le_to_cpu_64(resp->rx_bcast_bytes);
+
+	stats->opackets = rte_le_to_cpu_64(resp->tx_ucast_pkts);
+	stats->opackets += rte_le_to_cpu_64(resp->tx_mcast_pkts);
+	stats->opackets += rte_le_to_cpu_64(resp->tx_bcast_pkts);
+	stats->obytes = rte_le_to_cpu_64(resp->tx_ucast_bytes);
+	stats->obytes += rte_le_to_cpu_64(resp->tx_mcast_bytes);
+	stats->obytes += rte_le_to_cpu_64(resp->tx_bcast_bytes);
+
+	stats->ierrors = rte_le_to_cpu_64(resp->rx_err_pkts);
+	stats->oerrors = rte_le_to_cpu_64(resp->tx_err_pkts);
+
+	stats->imissed = rte_le_to_cpu_64(resp->rx_drop_pkts);
+
+	return rc;
+}
+
 /*
  * HWRM utility functions
  */
@@ -2348,6 +2382,42 @@ int bnxt_hwrm_exec_fwd_resp(struct bnxt *bp, uint16_t target_id,
 	return rc;
 }
 
+int bnxt_hwrm_ctx_qstats(struct bnxt *bp, uint32_t cid, int idx,
+			 struct rte_eth_stats *stats)
+{
+	int rc = 0;
+	struct hwrm_stat_ctx_query_input req = {.req_type = 0};
+	struct hwrm_stat_ctx_query_output *resp = bp->hwrm_cmd_resp_addr;
+
+	HWRM_PREP(req, STAT_CTX_QUERY, -1, resp);
+
+	req.stat_ctx_id = rte_cpu_to_le_32(cid);
+
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+
+	HWRM_CHECK_RESULT;
+
+	stats->q_ipackets[idx] = rte_le_to_cpu_64(resp->rx_ucast_pkts);
+	stats->q_ipackets[idx] += rte_le_to_cpu_64(resp->rx_mcast_pkts);
+	stats->q_ipackets[idx] += rte_le_to_cpu_64(resp->rx_bcast_pkts);
+	stats->q_ibytes[idx] = rte_le_to_cpu_64(resp->rx_ucast_bytes);
+	stats->q_ibytes[idx] += rte_le_to_cpu_64(resp->rx_mcast_bytes);
+	stats->q_ibytes[idx] += rte_le_to_cpu_64(resp->rx_bcast_bytes);
+
+	stats->q_opackets[idx] = rte_le_to_cpu_64(resp->tx_ucast_pkts);
+	stats->q_opackets[idx] += rte_le_to_cpu_64(resp->tx_mcast_pkts);
+	stats->q_opackets[idx] += rte_le_to_cpu_64(resp->tx_bcast_pkts);
+	stats->q_obytes[idx] = rte_le_to_cpu_64(resp->tx_ucast_bytes);
+	stats->q_obytes[idx] += rte_le_to_cpu_64(resp->tx_mcast_bytes);
+	stats->q_obytes[idx] += rte_le_to_cpu_64(resp->tx_bcast_bytes);
+
+	stats->q_errors[idx] = rte_le_to_cpu_64(resp->rx_err_pkts);
+	stats->q_errors[idx] += rte_le_to_cpu_64(resp->tx_err_pkts);
+	stats->q_errors[idx] += rte_le_to_cpu_64(resp->rx_drop_pkts);
+
+	return rc;
+}
+
 int bnxt_hwrm_port_qstats(struct bnxt *bp)
 {
 	struct hwrm_port_qstats_input req = {0};
diff --git a/drivers/net/bnxt/bnxt_hwrm.h b/drivers/net/bnxt/bnxt_hwrm.h
index 1759187..ab42913 100644
--- a/drivers/net/bnxt/bnxt_hwrm.h
+++ b/drivers/net/bnxt/bnxt_hwrm.h
@@ -62,6 +62,8 @@ int bnxt_hwrm_func_driver_register(struct bnxt *bp);
 int bnxt_hwrm_func_qcaps(struct bnxt *bp);
 int bnxt_hwrm_func_reset(struct bnxt *bp);
 int bnxt_hwrm_func_driver_unregister(struct bnxt *bp, uint32_t flags);
+int bnxt_hwrm_func_qstats(struct bnxt *bp, uint16_t fid,
+			  struct rte_eth_stats *stats);
 int bnxt_hwrm_func_cfg_def_cp(struct bnxt *bp);
 int bnxt_hwrm_vf_func_cfg_def_cp(struct bnxt *bp);
 
@@ -81,6 +83,8 @@ int bnxt_hwrm_stat_ctx_alloc(struct bnxt *bp,
 			     struct bnxt_cp_ring_info *cpr, unsigned int idx);
 int bnxt_hwrm_stat_ctx_free(struct bnxt *bp,
 			    struct bnxt_cp_ring_info *cpr, unsigned int idx);
+int bnxt_hwrm_ctx_qstats(struct bnxt *bp, uint32_t cid, int idx,
+			 struct rte_eth_stats *stats);
 
 int bnxt_hwrm_ver_get(struct bnxt *bp);
 
diff --git a/drivers/net/bnxt/bnxt_rxr.c b/drivers/net/bnxt/bnxt_rxr.c
index 1f0afcb..3001d81 100644
--- a/drivers/net/bnxt/bnxt_rxr.c
+++ b/drivers/net/bnxt/bnxt_rxr.c
@@ -68,8 +68,10 @@ static inline int bnxt_alloc_rx_data(struct bnxt_rx_queue *rxq,
 	struct rte_mbuf *data;
 
 	data = __bnxt_alloc_rx_data(rxq->mb_pool);
-	if (!data)
+	if (!data) {
+		rte_atomic64_inc(&rxq->bp->rx_mbuf_alloc_fail);
 		return -ENOMEM;
+	}
 
 	rx_buf->mbuf = data;
 
@@ -87,8 +89,10 @@ static inline int bnxt_alloc_ag_data(struct bnxt_rx_queue *rxq,
 	struct rte_mbuf *data;
 
 	data = __bnxt_alloc_rx_data(rxq->mb_pool);
-	if (!data)
+	if (!data) {
+		rte_atomic64_inc(&rxq->bp->rx_mbuf_alloc_fail);
 		return -ENOMEM;
+	}
 
 	if (rxbd == NULL)
 		RTE_LOG(ERR, PMD, "Jumbo Frame. rxbd is NULL\n");
@@ -321,8 +325,10 @@ static inline struct rte_mbuf *bnxt_tpa_end(
 
 	struct rte_mbuf *new_data = __bnxt_alloc_rx_data(rxq->mb_pool);
 	RTE_ASSERT(new_data != NULL);
-	if (!new_data)
+	if (!new_data) {
+		rte_atomic64_inc(&rxq->bp->rx_mbuf_alloc_fail);
 		return NULL;
+	}
 	tpa_info->mbuf = new_data;
 
 	return mbuf;
@@ -679,8 +685,10 @@ int bnxt_init_one_rx_ring(struct bnxt_rx_queue *rxq)
 		for (i = 0; i < BNXT_TPA_MAX; i++) {
 			rxr->tpa_info[i].mbuf =
 				__bnxt_alloc_rx_data(rxq->mb_pool);
-			if (!rxr->tpa_info[i].mbuf)
+			if (!rxr->tpa_info[i].mbuf) {
+				rte_atomic64_inc(&rxq->bp->rx_mbuf_alloc_fail);
 				return -ENOMEM;
+			}
 		}
 	}
 	RTE_LOG(DEBUG, PMD, "%s TPA alloc Done!\n", __func__);
diff --git a/drivers/net/bnxt/bnxt_stats.c b/drivers/net/bnxt/bnxt_stats.c
index 154dc0c..ade8195 100644
--- a/drivers/net/bnxt/bnxt_stats.c
+++ b/drivers/net/bnxt/bnxt_stats.c
@@ -196,64 +196,18 @@ void bnxt_stats_get_op(struct rte_eth_dev *eth_dev,
 	for (i = 0; i < bp->rx_cp_nr_rings; i++) {
 		struct bnxt_rx_queue *rxq = bp->rx_queues[i];
 		struct bnxt_cp_ring_info *cpr = rxq->cp_ring;
-		struct ctx_hw_stats64 *hw_stats =
-		    (struct ctx_hw_stats64 *)cpr->hw_stats;
-
-		bnxt_stats->q_ipackets[i] +=
-		    rte_le_to_cpu_64(hw_stats->rx_ucast_pkts);
-		bnxt_stats->q_ipackets[i] +=
-		    rte_le_to_cpu_64(hw_stats->rx_mcast_pkts);
-		bnxt_stats->q_ipackets[i] +=
-		    rte_le_to_cpu_64(hw_stats->rx_bcast_pkts);
-
-		bnxt_stats->q_ibytes[i] +=
-		    rte_le_to_cpu_64(hw_stats->rx_ucast_bytes);
-		bnxt_stats->q_ibytes[i] +=
-		    rte_le_to_cpu_64(hw_stats->rx_mcast_bytes);
-		bnxt_stats->q_ibytes[i] +=
-		    rte_le_to_cpu_64(hw_stats->rx_bcast_bytes);
-
-		/*
-		 * TBD: No clear mapping to this... we don't seem
-		 * to have a stat specifically for dropped due to
-		 * insufficient mbufs.
-		 */
-		bnxt_stats->q_errors[i] = 0;
-
-		/* These get replaced once the *_QSTATS commands work */
-		bnxt_stats->ipackets += bnxt_stats->q_ipackets[i];
-		bnxt_stats->ibytes += bnxt_stats->q_ibytes[i];
-		bnxt_stats->imissed += bnxt_stats->q_errors[i];
-		bnxt_stats->ierrors +=
-				rte_le_to_cpu_64(hw_stats->rx_discard_pkts);
+
+		bnxt_hwrm_ctx_qstats(bp, cpr->hw_stats_ctx_id, i, bnxt_stats);
 	}
 
 	for (i = 0; i < bp->tx_cp_nr_rings; i++) {
 		struct bnxt_tx_queue *txq = bp->tx_queues[i];
 		struct bnxt_cp_ring_info *cpr = txq->cp_ring;
-		struct ctx_hw_stats64 *hw_stats =
-		    (struct ctx_hw_stats64 *)cpr->hw_stats;
-
-		bnxt_stats->q_opackets[i] +=
-		    rte_le_to_cpu_64(hw_stats->tx_ucast_pkts);
-		bnxt_stats->q_opackets[i] +=
-		    rte_le_to_cpu_64(hw_stats->tx_mcast_pkts);
-		bnxt_stats->q_opackets[i] +=
-		    rte_le_to_cpu_64(hw_stats->tx_bcast_pkts);
-
-		bnxt_stats->q_obytes[i] +=
-		    rte_le_to_cpu_64(hw_stats->tx_ucast_bytes);
-		bnxt_stats->q_obytes[i] +=
-		    rte_le_to_cpu_64(hw_stats->tx_mcast_bytes);
-		bnxt_stats->q_obytes[i] +=
-		    rte_le_to_cpu_64(hw_stats->tx_bcast_bytes);
-
-		/* These get replaced once the *_QSTATS commands work */
-		bnxt_stats->opackets += bnxt_stats->q_opackets[i];
-		bnxt_stats->obytes +=  bnxt_stats->q_obytes[i];
-		bnxt_stats->oerrors += rte_le_to_cpu_64(hw_stats->tx_drop_pkts);
-		bnxt_stats->oerrors += rte_le_to_cpu_64(hw_stats->tx_discard_pkts);
+
+		bnxt_hwrm_ctx_qstats(bp, cpr->hw_stats_ctx_id, i, bnxt_stats);
 	}
+	bnxt_hwrm_func_qstats(bp, 0xffff, bnxt_stats);
+	bnxt_stats->rx_nombuf = rte_atomic64_read(&bp->rx_mbuf_alloc_fail);
 }
 
 void bnxt_stats_reset_op(struct rte_eth_dev *eth_dev)
@@ -261,6 +215,7 @@ void bnxt_stats_reset_op(struct rte_eth_dev *eth_dev)
 	struct bnxt *bp = (struct bnxt *)eth_dev->data->dev_private;
 
 	bnxt_clear_all_hwrm_stat_ctxs(bp);
+	rte_atomic64_clear(&bp->rx_mbuf_alloc_fail);
 }
 
 int bnxt_dev_xstats_get_op(struct rte_eth_dev *eth_dev,
-- 
2.10.1 (Apple Git-78)