[dpdk-stable] [PATCH v2 00/67] net/mlx5: backport patches for v17.11.3 LTS

[dpdk-stable] [PATCH v2 00/67] net/mlx5: backport patches for v17.11.3 LTS

[Yongseok Koh]

This patchset includes patches for mlx4 and mlx5. All the patches are from
dpdk/master and properly backported based on stable/17.11.

v2:
* no change, company mail server failed to send emails, just resend.

Adrien Mazarguil (6):
  net/mlx: control netdevices through ioctl only
  net/mlx4: fix RSS resource leak in case of error
  net/mlx5: fix RSS flow action bounds check
  net/mlx5: fix flow director rule deletion crash
  net/mlx4: fix Rx resource leak in case of error
  net/mlx4: fix UDP flow rule limitation enforcement

Dahir Osman (1):
  net/mlx5: setup RSS regardless of queue count

David Marchand (1):
  net/mlx5: fix count in xstats

Gavin Hu (1):
  net/mlx5: fix build with clang on ARM

Moti Haimovsky (2):
  net/mlx4: fix broadcast Rx
  net/mlx4: fix removal detection of stopped port

Nélio Laranjeiro (26):
  net/mlx5: remove get priv internal function
  net/mlx5: fix flow director conversion
  net/mlx5: fix reception of multiple MAC addresses
  net/mlx5: fix sriov flag
  net/mlx5: name parameters in function prototypes
  net/mlx5: mark parameters with unused attribute
  net/mlx5: normalize function prototypes
  net/mlx5: add missing function documentation
  net/mlx5: remove useless empty lines
  net/mlx5: remove control path locks
  net/mlx5: prefix all functions with mlx5
  net/mlx5: change non failing function return values
  net/mlx5: standardize on negative errno values
  net/mlx5: use port id in PMD log
  net/mlx5: use dynamic logging
  net/mlx5: remove kernel version check
  net/mlx5: fix link status behavior
  net/mlx5: fix link status to use wait to complete
  net/mlx5: change tunnel flow priority
  net/mlx5: improve flow error explanation
  net/mlx5: refuse empty VLAN flow specification
  net/mlx5: fix icc build
  net/mlx5: remove 32-bit support
  net/mlx5: split L3/L4 in flow director
  net/mlx5: fix flow director mask
  net/mlx5: fix flow validation

Olivier Matz (1):
  net/mlx5: fix allocation when no memory on device NUMA node

Ophir Munk (2):
  net/mlx4: fix single port configuration
  net/mlx4: avoid constant recreations in function

Rami Rosen (1):
  net/mlx4: fix a typo in header file

Raslan Darawsheh (3):
  net/mlx4: store RSS hash result in mbufs
  net/mlx5: fix double free on error handling
  net/mlx5: fix resource leak in case of error

Shachar Beiser (2):
  net: add IPv6 header fields macros
  net/mlx5: fix IPv6 header fields

Shahaf Shuler (12):
  net/mlx5: fix secondary process mempool registration
  net/mlx5: remove assert un-accessible from secondary process
  net/mlx5: warn for unsuccessful memory registration
  net/mlx5: fix CRC strip capability query
  net/mlx5: change pkt burst select function prototype
  net/mlx5: enforce RSS key length limitation
  net/mlx5: fix RSS key length query
  net/mlx5: fix link status initialization
  net/mlx5: fix ethtool link setting call order
  net/mlx5: fix socket connection return value
  net/mlx5: fix probe return value polarity
  net/mlx5: fix flow director drop rule deletion crash

Xueming Li (4):
  net/mlx5: map UAR address around huge pages
  net/mlx5: fix close after start failure
  net/mlx5: fix invalid flow item check
  net/mlx5: add data-plane debug message macro

Yongseok Koh (5):
  net/mlx5: fix synchronization on polling Rx completions
  net/mlx5: fix disabling Tx packet inlining
  net/mlx5: remove excessive data prefetch
  net/mlx5: fix calculation of Tx TSO inline room size
  net/mlx5: change device reference for secondary process

 doc/guides/nics/mlx5.rst              |    6 +-
 drivers/net/mlx4/mlx4.c               |   18 +-
 drivers/net/mlx4/mlx4.h               |    3 +
 drivers/net/mlx4/mlx4_ethdev.c        |  192 +-----
 drivers/net/mlx4/mlx4_flow.c          |   34 +-
 drivers/net/mlx4/mlx4_flow.h          |    1 +
 drivers/net/mlx4/mlx4_intr.c          |   41 +-
 drivers/net/mlx4/mlx4_rxq.c           |    7 +
 drivers/net/mlx4/mlx4_rxtx.c          |    3 +-
 drivers/net/mlx4/mlx4_rxtx.h          |    2 +-
 drivers/net/mlx5/mlx5.c               |  525 +++++++++------
 drivers/net/mlx5/mlx5.h               |  267 ++++----
 drivers/net/mlx5/mlx5_defs.h          |   14 +-
 drivers/net/mlx5/mlx5_ethdev.c        | 1018 +++++++++--------------------
 drivers/net/mlx5/mlx5_flow.c          | 1132 ++++++++++++++++++---------------
 drivers/net/mlx5/mlx5_mac.c           |   49 +-
 drivers/net/mlx5/mlx5_mr.c            |  197 +++---
 drivers/net/mlx5/mlx5_rss.c           |  176 +++--
 drivers/net/mlx5/mlx5_rxmode.c        |   28 +-
 drivers/net/mlx5/mlx5_rxq.c           |  712 ++++++++++++---------
 drivers/net/mlx5/mlx5_rxtx.c          |   76 +--
 drivers/net/mlx5/mlx5_rxtx.h          |  179 +++---
 drivers/net/mlx5/mlx5_rxtx_vec.c      |   26 +-
 drivers/net/mlx5/mlx5_rxtx_vec_neon.h |   65 +-
 drivers/net/mlx5/mlx5_rxtx_vec_sse.h  |    2 +-
 drivers/net/mlx5/mlx5_socket.c        |  167 +++--
 drivers/net/mlx5/mlx5_stats.c         |  248 ++++----
 drivers/net/mlx5/mlx5_trigger.c       |  258 ++++----
 drivers/net/mlx5/mlx5_txq.c           |  379 ++++++-----
 drivers/net/mlx5/mlx5_utils.h         |   29 +-
 drivers/net/mlx5/mlx5_vlan.c          |  106 ++-
 lib/librte_net/rte_ip.h               |    6 +
 32 files changed, 2947 insertions(+), 3019 deletions(-)

-- 
2.11.0

[dpdk-stable] [PATCH v2 01/67] net/mlx5: remove get priv internal function

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 01d79216e68e5fa5f79ef09bfcb9344b9d56800e ]

mlx5_get_priv() is barely use across the driver.  To avoid mixing access,
this function is definitely removed.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5.c        |  2 +-
 drivers/net/mlx5/mlx5_ethdev.c | 21 +++------------------
 drivers/net/mlx5/mlx5_rxq.c    |  4 ++--
 drivers/net/mlx5/mlx5_stats.c  |  8 ++++----
 4 files changed, 10 insertions(+), 25 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 45e0e8dba..7f44bb169 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -191,7 +191,7 @@ mlx5_free_verbs_buf(void *ptr, void *data __rte_unused)
 static void
 mlx5_dev_close(struct rte_eth_dev *dev)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 	int ret;
 
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index ffe1cdd64..3e5a2922d 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -119,21 +119,6 @@ struct ethtool_link_settings {
 #endif
 
 /**
- * Return private structure associated with an Ethernet device.
- *
- * @param dev
- *   Pointer to Ethernet device structure.
- *
- * @return
- *   Pointer to private structure.
- */
-struct priv *
-mlx5_get_priv(struct rte_eth_dev *dev)
-{
-	return dev->data->dev_private;
-}
-
-/**
  * Get interface name from private structure.
  *
  * @param[in] priv
@@ -655,7 +640,7 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 void
 mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	unsigned int max;
 	char ifname[IF_NAMESIZE];
 
@@ -747,7 +732,7 @@ mlx5_dev_supported_ptypes_get(struct rte_eth_dev *dev)
 static int
 mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev, int wait_to_complete)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	struct ethtool_cmd edata = {
 		.cmd = ETHTOOL_GSET /* Deprecated since Linux v4.5. */
 	};
@@ -813,7 +798,7 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev, int wait_to_complete)
 static int
 mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev, int wait_to_complete)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	struct ethtool_link_settings gcmd = { .cmd = ETHTOOL_GLINKSETTINGS };
 	struct ifreq ifr;
 	struct rte_eth_link dev_link;
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 20f3ec6c9..a4cdd374a 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -454,7 +454,7 @@ mlx5_arm_cq(struct mlx5_rxq_data *rxq, int sq_n_rxq)
 int
 mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_data *rxq_data;
 	struct mlx5_rxq_ctrl *rxq_ctrl;
 	int ret = 0;
@@ -498,7 +498,7 @@ mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 int
 mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_data *rxq_data;
 	struct mlx5_rxq_ctrl *rxq_ctrl;
 	struct mlx5_rxq_ibv *rxq_ibv = NULL;
diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index 2427585f7..c11ee57f0 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -319,7 +319,7 @@ priv_xstats_reset(struct priv *priv)
 int
 mlx5_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	struct rte_eth_stats tmp = {0};
 	unsigned int i;
 	unsigned int idx;
@@ -425,7 +425,7 @@ int
 mlx5_xstats_get(struct rte_eth_dev *dev,
 		struct rte_eth_xstat *stats, unsigned int n)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	int ret = xstats_n;
 
 	if (n >= xstats_n && stats) {
@@ -455,7 +455,7 @@ mlx5_xstats_get(struct rte_eth_dev *dev,
 void
 mlx5_xstats_reset(struct rte_eth_dev *dev)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
 	int stats_n;
 
@@ -487,7 +487,7 @@ int
 mlx5_xstats_get_names(struct rte_eth_dev *dev,
 		struct rte_eth_xstat_name *xstats_names, unsigned int n)
 {
-	struct priv *priv = mlx5_get_priv(dev);
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
 	if (n >= xstats_n && xstats_names) {
-- 
2.11.0

[dpdk-stable] [PATCH v2 02/67] net/mlx4: store RSS hash result in mbufs

[Yongseok Koh]

From: Raslan Darawsheh <rasland@mellanox.com>

[ upstream commit 8937c9f1de1b85c5b21f701d5e87caa8f54496d0 ]

Add RSS hash result from CQE to mbuf,
Also, set PKT_RX_RSS_HASH in the ol_flags.

Signed-off-by: Raslan Darawsheh <rasland@mellanox.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx4/mlx4_rxtx.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mlx4/mlx4_rxtx.c b/drivers/net/mlx4/mlx4_rxtx.c
index 92b625775..05c489207 100644
--- a/drivers/net/mlx4/mlx4_rxtx.c
+++ b/drivers/net/mlx4/mlx4_rxtx.c
@@ -961,7 +961,8 @@ mlx4_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
 			/* Update packet information. */
 			pkt->packet_type =
 				rxq_cq_to_pkt_type(cqe, rxq->l2tun_offload);
-			pkt->ol_flags = 0;
+			pkt->ol_flags = PKT_RX_RSS_HASH;
+			pkt->hash.rss = cqe->immed_rss_invalid;
 			pkt->pkt_len = len;
 			if (rxq->csum | rxq->csum_l2tun) {
 				uint32_t flags =
-- 
2.11.0

[dpdk-stable] [PATCH v2 03/67] net/mlx5: fix synchronization on polling Rx completions

[Yongseok Koh]

[ backported from upstream commit 1742c2d9fab07e66209f2d14e7daa50829fc4423 ]

Polling a new packet is basically sensing the generation bit in a
completion entry. For some processors not having strongly-ordered memory
model, there has to be a memory barrier between reading the generation bit
and other fields of the entry in order to guarantee data is not stale.

Fixes: 570acdb1da8a ("net/mlx5: add vectorized Rx/Tx burst for ARM")
Cc: stable@dpdk.org

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_rxtx.c          |  1 +
 drivers/net/mlx5/mlx5_rxtx_vec_neon.h | 53 ++++++++++++++++++++---------------
 drivers/net/mlx5/mlx5_rxtx_vec_sse.h  |  2 +-
 3 files changed, 32 insertions(+), 24 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 9f1478cdb..c2dab1f4c 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -1717,6 +1717,7 @@ mlx5_rx_poll_len(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cqe,
 			return 0;
 		++rxq->cq_ci;
 		op_own = cqe->op_own;
+		rte_io_rmb();
 		if (MLX5_CQE_FORMAT(op_own) == MLX5_COMPRESSED) {
 			volatile struct mlx5_mini_cqe8 (*mc)[8] =
 				(volatile struct mlx5_mini_cqe8 (*)[8])
diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
index 06f83ef14..0b842f9a0 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
+++ b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
@@ -806,6 +806,7 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n,
 		uint16x4_t mask;
 		uint16x4_t byte_cnt;
 		uint32x4_t ptype_info, flow_tag;
+		register uint64x2_t c0, c1, c2, c3;
 		uint8_t *p0, *p1, *p2, *p3;
 		uint8_t *e0 = (void *)&elts[pos]->pkt_len;
 		uint8_t *e1 = (void *)&elts[pos + 1]->pkt_len;
@@ -822,6 +823,16 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n,
 		p1 = p0 + (pkts_n - pos > 1) * sizeof(struct mlx5_cqe);
 		p2 = p1 + (pkts_n - pos > 2) * sizeof(struct mlx5_cqe);
 		p3 = p2 + (pkts_n - pos > 3) * sizeof(struct mlx5_cqe);
+		/* B.0 (CQE 3) load a block having op_own. */
+		c3 = vld1q_u64((uint64_t *)(p3 + 48));
+		/* B.0 (CQE 2) load a block having op_own. */
+		c2 = vld1q_u64((uint64_t *)(p2 + 48));
+		/* B.0 (CQE 1) load a block having op_own. */
+		c1 = vld1q_u64((uint64_t *)(p1 + 48));
+		/* B.0 (CQE 0) load a block having op_own. */
+		c0 = vld1q_u64((uint64_t *)(p0 + 48));
+		/* Synchronize for loading the rest of blocks. */
+		rte_io_rmb();
 		/* Prefetch next 4 CQEs. */
 		if (pkts_n - pos >= 2 * MLX5_VPMD_DESCS_PER_LOOP) {
 			unsigned int next = pos + MLX5_VPMD_DESCS_PER_LOOP;
@@ -831,50 +842,46 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n,
 			rte_prefetch_non_temporal(&cq[next + 3]);
 		}
 		__asm__ volatile (
-		/* B.1 (CQE 3) load a block having op_own. */
-		"ld1 {v19.16b}, [%[p3]] \n\t"
-		"sub %[p3], %[p3], #48 \n\t"
-		/* B.2 (CQE 3) load the rest blocks. */
+		/* B.1 (CQE 3) load the rest of blocks. */
 		"ld1 {v16.16b - v18.16b}, [%[p3]] \n\t"
+		/* B.2 (CQE 3) move the block having op_own. */
+		"mov v19.16b, %[c3].16b \n\t"
 		/* B.3 (CQE 3) extract 16B fields. */
 		"tbl v23.16b, {v16.16b - v19.16b}, %[cqe_shuf_m].16b \n\t"
+		/* B.1 (CQE 2) load the rest of blocks. */
+		"ld1 {v16.16b - v18.16b}, [%[p2]] \n\t"
 		/* B.4 (CQE 3) adjust CRC length. */
 		"sub v23.8h, v23.8h, %[crc_adj].8h \n\t"
-		/* B.1 (CQE 2) load a block having op_own. */
-		"ld1 {v19.16b}, [%[p2]] \n\t"
-		"sub %[p2], %[p2], #48 \n\t"
 		/* C.1 (CQE 3) generate final structure for mbuf. */
 		"tbl v15.16b, {v23.16b}, %[mb_shuf_m].16b \n\t"
-		/* B.2 (CQE 2) load the rest blocks. */
-		"ld1 {v16.16b - v18.16b}, [%[p2]] \n\t"
+		/* B.2 (CQE 2) move the block having op_own. */
+		"mov v19.16b, %[c2].16b \n\t"
 		/* B.3 (CQE 2) extract 16B fields. */
 		"tbl v22.16b, {v16.16b - v19.16b}, %[cqe_shuf_m].16b \n\t"
+		/* B.1 (CQE 1) load the rest of blocks. */
+		"ld1 {v16.16b - v18.16b}, [%[p1]] \n\t"
 		/* B.4 (CQE 2) adjust CRC length. */
 		"sub v22.8h, v22.8h, %[crc_adj].8h \n\t"
-		/* B.1 (CQE 1) load a block having op_own. */
-		"ld1 {v19.16b}, [%[p1]] \n\t"
-		"sub %[p1], %[p1], #48 \n\t"
 		/* C.1 (CQE 2) generate final structure for mbuf. */
 		"tbl v14.16b, {v22.16b}, %[mb_shuf_m].16b \n\t"
-		/* B.2 (CQE 1) load the rest blocks. */
-		"ld1 {v16.16b - v18.16b}, [%[p1]] \n\t"
+		/* B.2 (CQE 1) move the block having op_own. */
+		"mov v19.16b, %[c1].16b \n\t"
 		/* B.3 (CQE 1) extract 16B fields. */
 		"tbl v21.16b, {v16.16b - v19.16b}, %[cqe_shuf_m].16b \n\t"
+		/* B.1 (CQE 0) load the rest of blocks. */
+		"ld1 {v16.16b - v18.16b}, [%[p0]] \n\t"
 		/* B.4 (CQE 1) adjust CRC length. */
 		"sub v21.8h, v21.8h, %[crc_adj].8h \n\t"
-		/* B.1 (CQE 0) load a block having op_own. */
-		"ld1 {v19.16b}, [%[p0]] \n\t"
-		"sub %[p0], %[p0], #48 \n\t"
 		/* C.1 (CQE 1) generate final structure for mbuf. */
 		"tbl v13.16b, {v21.16b}, %[mb_shuf_m].16b \n\t"
-		/* B.2 (CQE 0) load the rest blocks. */
-		"ld1 {v16.16b - v18.16b}, [%[p0]] \n\t"
+		/* B.2 (CQE 0) move the block having op_own. */
+		"mov v19.16b, %[c0].16b \n\t"
+		/* A.1 load mbuf pointers. */
+		"ld1 {v24.2d - v25.2d}, [%[elts_p]] \n\t"
 		/* B.3 (CQE 0) extract 16B fields. */
 		"tbl v20.16b, {v16.16b - v19.16b}, %[cqe_shuf_m].16b \n\t"
 		/* B.4 (CQE 0) adjust CRC length. */
 		"sub v20.8h, v20.8h, %[crc_adj].8h \n\t"
-		/* A.1 load mbuf pointers. */
-		"ld1 {v24.2d - v25.2d}, [%[elts_p]] \n\t"
 		/* D.1 extract op_own byte. */
 		"tbl %[op_own].8b, {v20.16b - v23.16b}, %[owner_shuf_m].8b \n\t"
 		/* C.2 (CQE 3) adjust flow mark. */
@@ -909,9 +916,9 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n,
 		 [byte_cnt]"=&w"(byte_cnt),
 		 [ptype_info]"=&w"(ptype_info),
 		 [flow_tag]"=&w"(flow_tag)
-		:[p3]"r"(p3 + 48), [p2]"r"(p2 + 48),
-		 [p1]"r"(p1 + 48), [p0]"r"(p0 + 48),
+		:[p3]"r"(p3), [p2]"r"(p2), [p1]"r"(p1), [p0]"r"(p0),
 		 [e3]"r"(e3), [e2]"r"(e2), [e1]"r"(e1), [e0]"r"(e0),
+		 [c3]"w"(c3), [c2]"w"(c2), [c1]"w"(c1), [c0]"w"(c0),
 		 [elts_p]"r"(elts_p),
 		 [pkts_p]"r"(pkts_p),
 		 [cqe_shuf_m]"w"(cqe_shuf_m),
diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
index 7ef2c59e1..793142922 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
+++ b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
@@ -825,7 +825,7 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n,
 		/* B.2 copy mbuf pointers. */
 		_mm_storeu_si128((__m128i *)&pkts[pos], mbp1);
 		_mm_storeu_si128((__m128i *)&pkts[pos + 2], mbp2);
-		rte_compiler_barrier();
+		rte_io_rmb();
 		/* C.1 load remained CQE data and extract necessary fields. */
 		cqe_tmp2 = _mm_load_si128((__m128i *)&cq[pos + p3]);
 		cqe_tmp1 = _mm_load_si128((__m128i *)&cq[pos + p2]);
-- 
2.11.0

[dpdk-stable] [PATCH v2 04/67] net/mlx5: fix allocation when no memory on device NUMA node

[Yongseok Koh]

From: Olivier Matz <olivier.matz@6wind.com>

[ backported from upstream commit d10b09db0a455a27f1f1fa5e4a0fa7b1350706a5 ]

When no memory is available on the same numa node than the device, the
initialization of the device fails. However, the use case where the
cores and memory are on a different socket than the device is valid,
even if not optimal.

To fix this issue, this commit introduces an infrastructure to select
the socket on which to allocate the verbs objects based on the ethdev
configuration and the object type, rather than the PCI numa node.

Fixes: 1e3a39f72d5d ("net/mlx5: allocate verbs object into shared memory")
Cc: stable@dpdk.org

Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5.c     | 14 ++++++++++++--
 drivers/net/mlx5/mlx5.h     | 20 ++++++++++++++++++++
 drivers/net/mlx5/mlx5_rxq.c |  4 ++++
 drivers/net/mlx5/mlx5_txq.c |  4 ++++
 4 files changed, 40 insertions(+), 2 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 7f44bb169..16ba5b724 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -156,10 +156,20 @@ mlx5_alloc_verbs_buf(size_t size, void *data)
 	struct priv *priv = data;
 	void *ret;
 	size_t alignment = sysconf(_SC_PAGESIZE);
+	unsigned int socket = SOCKET_ID_ANY;
 
+	if (priv->verbs_alloc_ctx.type == MLX5_VERBS_ALLOC_TYPE_TX_QUEUE) {
+		const struct mlx5_txq_ctrl *ctrl = priv->verbs_alloc_ctx.obj;
+
+		socket = ctrl->socket;
+	} else if (priv->verbs_alloc_ctx.type ==
+		   MLX5_VERBS_ALLOC_TYPE_RX_QUEUE) {
+		const struct mlx5_rxq_ctrl *ctrl = priv->verbs_alloc_ctx.obj;
+
+		socket = ctrl->socket;
+	}
 	assert(data != NULL);
-	ret = rte_malloc_socket(__func__, size, alignment,
-				priv->dev->device->numa_node);
+	ret = rte_malloc_socket(__func__, size, alignment, socket);
 	DEBUG("Extern alloc size: %lu, align: %lu: %p", size, alignment, ret);
 	return ret;
 }
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index d49595bc9..004e83670 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -90,6 +90,24 @@ struct mlx5_xstats_ctrl {
 /* Flow list . */
 TAILQ_HEAD(mlx5_flows, rte_flow);
 
+/**
+ * Type of objet being allocated.
+ */
+enum mlx5_verbs_alloc_type {
+	MLX5_VERBS_ALLOC_TYPE_NONE,
+	MLX5_VERBS_ALLOC_TYPE_TX_QUEUE,
+	MLX5_VERBS_ALLOC_TYPE_RX_QUEUE,
+};
+
+/**
+ * Verbs allocator needs a context to know in the callback which kind of
+ * resources it is allocating.
+ */
+struct mlx5_verbs_alloc_ctx {
+	enum mlx5_verbs_alloc_type type; /* Kind of object being allocated. */
+	const void *obj; /* Pointer to the DPDK object. */
+};
+
 struct priv {
 	struct rte_eth_dev *dev; /* Ethernet device of master process. */
 	struct ibv_context *ctx; /* Verbs context. */
@@ -149,6 +167,8 @@ struct priv {
 	rte_spinlock_t lock; /* Lock for control functions. */
 	int primary_socket; /* Unix socket for primary process. */
 	struct rte_intr_handle intr_handle_socket; /* Interrupt handler. */
+	struct mlx5_verbs_alloc_ctx verbs_alloc_ctx;
+	/* Context for Verbs allocator. */
 };
 
 /**
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index a4cdd374a..457f919b4 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -573,6 +573,8 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 
 	assert(rxq_data);
 	assert(!rxq_ctrl->ibv);
+	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_RX_QUEUE;
+	priv->verbs_alloc_ctx.obj = rxq_ctrl;
 	tmpl = rte_calloc_socket(__func__, 1, sizeof(*tmpl), 0,
 				 rxq_ctrl->socket);
 	if (!tmpl) {
@@ -736,6 +738,7 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 	DEBUG("%p: Verbs Rx queue %p: refcnt %d", (void *)priv,
 	      (void *)tmpl, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->rxqsibv, tmpl, next);
+	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return tmpl;
 error:
 	if (tmpl->wq)
@@ -746,6 +749,7 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 		claim_zero(ibv_destroy_comp_channel(tmpl->channel));
 	if (tmpl->mr)
 		priv_mr_release(priv, tmpl->mr);
+	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return NULL;
 }
 
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 7ca99f5aa..12f258bbb 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -307,6 +307,8 @@ mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
 	int ret = 0;
 
 	assert(txq_data);
+	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_TX_QUEUE;
+	priv->verbs_alloc_ctx.obj = txq_ctrl;
 	if (mlx5_getenv_int("MLX5_ENABLE_CQE_COMPRESSION")) {
 		ERROR("MLX5_ENABLE_CQE_COMPRESSION must never be set");
 		goto error;
@@ -435,12 +437,14 @@ mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
 	DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)priv,
 	      (void *)txq_ibv, rte_atomic32_read(&txq_ibv->refcnt));
 	LIST_INSERT_HEAD(&priv->txqsibv, txq_ibv, next);
+	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return txq_ibv;
 error:
 	if (tmpl.cq)
 		claim_zero(ibv_destroy_cq(tmpl.cq));
 	if (tmpl.qp)
 		claim_zero(ibv_destroy_qp(tmpl.qp));
+	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return NULL;
 }
 
-- 
2.11.0

[dpdk-stable] [PATCH v2 05/67] net/mlx5: fix flow director conversion

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit d10f422fd59a7a6d01da1b2a79975e9776fba840 ]

Flow director must provide the same spec and mask to be sure to be
validated.

Fixes: 4c3e9bcdd52e ("net/mlx5: support flow director")
Cc: stable@dpdk.org

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5_flow.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 3a9407b00..76c4354bb 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -2650,10 +2650,12 @@ priv_fdir_filter_convert(struct priv *priv,
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV4,
 			.spec = &attributes->l3,
+			.mask = &attributes->l3,
 		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_UDP,
 			.spec = &attributes->l4,
+			.mask = &attributes->l4,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV4_TCP:
@@ -2671,10 +2673,12 @@ priv_fdir_filter_convert(struct priv *priv,
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV4,
 			.spec = &attributes->l3,
+			.mask = &attributes->l3,
 		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_TCP,
 			.spec = &attributes->l4,
+			.mask = &attributes->l4,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV4_OTHER:
@@ -2688,6 +2692,7 @@ priv_fdir_filter_convert(struct priv *priv,
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV4,
 			.spec = &attributes->l3,
+			.mask = &attributes->l3,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV6_UDP:
@@ -2708,10 +2713,12 @@ priv_fdir_filter_convert(struct priv *priv,
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV6,
 			.spec = &attributes->l3,
+			.mask = &attributes->l3,
 		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_UDP,
 			.spec = &attributes->l4,
+			.mask = &attributes->l4,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
@@ -2732,10 +2739,12 @@ priv_fdir_filter_convert(struct priv *priv,
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV6,
 			.spec = &attributes->l3,
+			.mask = &attributes->l3,
 		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_TCP,
 			.spec = &attributes->l4,
+			.mask = &attributes->l4,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV6_OTHER:
@@ -2752,6 +2761,7 @@ priv_fdir_filter_convert(struct priv *priv,
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV6,
 			.spec = &attributes->l3,
+			.mask = &attributes->l3,
 		};
 		break;
 	default:
-- 
2.11.0

[dpdk-stable] [PATCH v2 06/67] net/mlx5: fix reception of multiple MAC addresses

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit 5a29b9a35a7214e2e36764ea67e13d7573eb925a ]

When promiscuous is disabled, adding/removing a mac address is ignored
causing the packet to not be received or still being received corresponding
to the add or remove request.

Fixes: 272733b5ebfd ("net/mlx5: use flow to enable unicast traffic")
Cc: stable@dpdk.org

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
---
 drivers/net/mlx5/mlx5_mac.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_mac.c b/drivers/net/mlx5/mlx5_mac.c
index 9fb5ba5e7..3ad21c000 100644
--- a/drivers/net/mlx5/mlx5_mac.c
+++ b/drivers/net/mlx5/mlx5_mac.c
@@ -95,7 +95,7 @@ mlx5_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index)
 {
 	assert(index < MLX5_MAX_MAC_ADDRESSES);
 	memset(&dev->data->mac_addrs[index], 0, sizeof(struct ether_addr));
-	if (!dev->data->promiscuous && !dev->data->all_multicast)
+	if (!dev->data->promiscuous)
 		mlx5_traffic_restart(dev);
 }
 
@@ -134,7 +134,7 @@ mlx5_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac,
 		return EADDRINUSE;
 	}
 	dev->data->mac_addrs[index] = *mac;
-	if (!dev->data->promiscuous && !dev->data->all_multicast)
+	if (!dev->data->promiscuous)
 		mlx5_traffic_restart(dev);
 	return ret;
 }
-- 
2.11.0

[dpdk-stable] [PATCH v2 07/67] net/mlx5: fix secondary process mempool registration

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ upstream commit a482a41a638749b2a5b9a1f37c8774453274a8f8 ]

Secondary process is not allowed to register mempools on the flight.

The code will return invalid memory key for such case.

Fixes: 87ec44ce1651 ("net/mlx5: add operations for secondary process")
Cc: stable@dpdk.org

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 doc/guides/nics/mlx5.rst   | 6 +++++-
 drivers/net/mlx5/mlx5_mr.c | 9 ++++++++-
 2 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/doc/guides/nics/mlx5.rst b/doc/guides/nics/mlx5.rst
index f9558da89..50fced3f4 100644
--- a/doc/guides/nics/mlx5.rst
+++ b/doc/guides/nics/mlx5.rst
@@ -108,7 +108,11 @@ Limitations
 - Port statistics through software counters only. Flow statistics are
   supported by hardware counters.
 - Hardware checksum RX offloads for VXLAN inner header are not supported yet.
-- Forked secondary process not supported.
+- For secondary process:
+
+  - Forked secondary process not supported.
+  - All mempools must be initialized before rte_eth_dev_start().
+
 - Flow pattern without any specific vlan will match for vlan packets as well:
 
   When VLAN spec is not specified in the pattern, the matching rule will be created with VLAN as a wild card.
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 2776dc700..33d037d28 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -141,8 +141,15 @@ priv_txq_mp2mr_reg(struct priv *priv, struct mlx5_txq_data *txq,
 	DEBUG("%p: discovered new memory pool \"%s\" (%p)",
 	      (void *)txq_ctrl, mp->name, (void *)mp);
 	mr = priv_mr_get(priv, mp);
-	if (mr == NULL)
+	if (mr == NULL) {
+		if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
+			DEBUG("Using unregistered mempool 0x%p(%s) in secondary process,"
+			     " please create mempool before rte_eth_dev_start()",
+			     (void *)mp, mp->name);
+			return NULL;
+		}
 		mr = priv_mr_new(priv, mp);
+	}
 	if (unlikely(mr == NULL)) {
 		DEBUG("%p: unable to configure MR, ibv_reg_mr() failed.",
 		      (void *)txq_ctrl);
-- 
2.11.0

[dpdk-stable] [PATCH v2 08/67] net/mlx5: remove assert un-accessible from secondary process

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ upstream commit 1b2a3baaed8a8c5165bc6d8ef9164a10183b2310 ]

Verbs structs such as ibv_mr are not accessible from the secondary
process.

Choose to remove the assert in favor of performing more checks on the
critical data path.

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_rxtx.h | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index de5b769ef..9132d49c2 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -558,8 +558,6 @@ mlx5_tx_mb2mr(struct mlx5_txq_data *txq, struct rte_mbuf *mb)
 		if (txq->mp2mr[i]->start <= addr &&
 		    txq->mp2mr[i]->end > addr) {
 			assert(txq->mp2mr[i]->lkey != (uint32_t)-1);
-			assert(rte_cpu_to_be_32(txq->mp2mr[i]->mr->lkey) ==
-			       txq->mp2mr[i]->lkey);
 			txq->mr_cache_idx = i;
 			return txq->mp2mr[i]->lkey;
 		}
-- 
2.11.0

[dpdk-stable] [PATCH v2 09/67] net/mlx5: warn for unsuccessful memory registration

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ upstream commit bb6883120a95861c1d4dae27839871ce5805e381 ]

Memory registration can fail, add the proper warning for such scenario
for it at least to be visible in debug mode.

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_rxtx.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index 9132d49c2..e35a6ab34 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -571,6 +571,11 @@ mlx5_tx_mb2mr(struct mlx5_txq_data *txq, struct rte_mbuf *mb)
 		rte_atomic32_inc(&mr->refcnt);
 		txq->mr_cache_idx = i >= RTE_DIM(txq->mp2mr) ? i - 1 : i;
 		return mr->lkey;
+	} else {
+		struct rte_mempool *mp = mlx5_tx_mb2mp(mb);
+
+		WARN("Failed to register mempool 0x%p(%s)",
+		      (void *)mp, mp->name);
 	}
 	return (uint32_t)-1;
 }
-- 
2.11.0

[dpdk-stable] [PATCH v2 10/67] net/mlx5: map UAR address around huge pages

[Yongseok Koh]

From: Xueming Li <xuemingl@mellanox.com>

[ backported from upstream commit 4a984153430cbdafe2b971e999bba8d96610b299 ]

Reserving the memory space for the UAR near huge pages helps to
**reduce** the cases where the secondary process cannot start. Those
pages being physical pages they must be mapped at the same virtual
address as in the primary process to have a
working secondary process.

As this remap is almost the latest being done by the processes
(libraries, heaps, stacks are already loaded), similar to huge pages,
there is **no guarantee** this mechanism will always work.

Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5.c         | 114 ++++++++++++++++++++++++++++++++++++++--
 drivers/net/mlx5/mlx5.h         |   1 +
 drivers/net/mlx5/mlx5_defs.h    |  10 ++++
 drivers/net/mlx5/mlx5_rxtx.h    |   3 +-
 drivers/net/mlx5/mlx5_trigger.c |   7 ++-
 drivers/net/mlx5/mlx5_txq.c     |  51 +++++++++++++-----
 6 files changed, 167 insertions(+), 19 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 16ba5b724..7dc3dc399 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -39,6 +39,7 @@
 #include <stdlib.h>
 #include <errno.h>
 #include <net/if.h>
+#include <sys/mman.h>
 
 /* Verbs header. */
 /* ISO C doesn't support unnamed structs/unions, disabling -pedantic. */
@@ -56,6 +57,7 @@
 #include <rte_pci.h>
 #include <rte_bus_pci.h>
 #include <rte_common.h>
+#include <rte_eal_memconfig.h>
 #include <rte_kvargs.h>
 
 #include "mlx5.h"
@@ -496,6 +498,106 @@ mlx5_args(struct mlx5_args *args, struct rte_devargs *devargs)
 
 static struct rte_pci_driver mlx5_driver;
 
+/*
+ * Reserved UAR address space for TXQ UAR(hw doorbell) mapping, process
+ * local resource used by both primary and secondary to avoid duplicate
+ * reservation.
+ * The space has to be available on both primary and secondary process,
+ * TXQ UAR maps to this area using fixed mmap w/o double check.
+ */
+static void *uar_base;
+
+/**
+ * Reserve UAR address space for primary process.
+ *
+ * @param[in] priv
+ *   Pointer to private structure.
+ *
+ * @return
+ *   0 on success, errno value on failure.
+ */
+static int
+priv_uar_init_primary(struct priv *priv)
+{
+	void *addr = (void *)0;
+	int i;
+	const struct rte_mem_config *mcfg;
+	int ret;
+
+	if (uar_base) { /* UAR address space mapped. */
+		priv->uar_base = uar_base;
+		return 0;
+	}
+	/* find out lower bound of hugepage segments */
+	mcfg = rte_eal_get_configuration()->mem_config;
+	for (i = 0; i < RTE_MAX_MEMSEG && mcfg->memseg[i].addr; i++) {
+		if (addr)
+			addr = RTE_MIN(addr, mcfg->memseg[i].addr);
+		else
+			addr = mcfg->memseg[i].addr;
+	}
+	/* keep distance to hugepages to minimize potential conflicts. */
+	addr = RTE_PTR_SUB(addr, MLX5_UAR_OFFSET + MLX5_UAR_SIZE);
+	/* anonymous mmap, no real memory consumption. */
+	addr = mmap(addr, MLX5_UAR_SIZE,
+		    PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+	if (addr == MAP_FAILED) {
+		ERROR("Failed to reserve UAR address space, please adjust "
+		      "MLX5_UAR_SIZE or try --base-virtaddr");
+		ret = ENOMEM;
+		return ret;
+	}
+	/* Accept either same addr or a new addr returned from mmap if target
+	 * range occupied.
+	 */
+	INFO("Reserved UAR address space: %p", addr);
+	priv->uar_base = addr; /* for primary and secondary UAR re-mmap. */
+	uar_base = addr; /* process local, don't reserve again. */
+	return 0;
+}
+
+/**
+ * Reserve UAR address space for secondary process, align with
+ * primary process.
+ *
+ * @param[in] priv
+ *   Pointer to private structure.
+ *
+ * @return
+ *   0 on success, errno value on failure.
+ */
+static int
+priv_uar_init_secondary(struct priv *priv)
+{
+	void *addr;
+	int ret;
+
+	assert(priv->uar_base);
+	if (uar_base) { /* already reserved. */
+		assert(uar_base == priv->uar_base);
+		return 0;
+	}
+	/* anonymous mmap, no real memory consumption. */
+	addr = mmap(priv->uar_base, MLX5_UAR_SIZE,
+		    PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+	if (addr == MAP_FAILED) {
+		ERROR("UAR mmap failed: %p size: %llu",
+		      priv->uar_base, MLX5_UAR_SIZE);
+		ret = ENXIO;
+		return ret;
+	}
+	if (priv->uar_base != addr) {
+		ERROR("UAR address %p size %llu occupied, please adjust "
+		      "MLX5_UAR_OFFSET or try EAL parameter --base-virtaddr",
+		      priv->uar_base, MLX5_UAR_SIZE);
+		ret = ENXIO;
+		return ret;
+	}
+	uar_base = addr; /* process local, don't reserve again */
+	INFO("Reserved UAR address space: %p", addr);
+	return 0;
+}
+
 /**
  * Assign parameters from args into priv, only non default
  * values are considered.
@@ -710,6 +812,11 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 			eth_dev->device = &pci_dev->device;
 			eth_dev->dev_ops = &mlx5_dev_sec_ops;
 			priv = eth_dev->data->dev_private;
+			err = priv_uar_init_secondary(priv);
+			if (err < 0) {
+				err = -err;
+				goto error;
+			}
 			/* Receive command fd from primary process */
 			err = priv_socket_connect(priv);
 			if (err < 0) {
@@ -718,10 +825,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 			}
 			/* Remap UAR for Tx queues. */
 			err = priv_tx_uar_remap(priv, err);
-			if (err < 0) {
-				err = -err;
+			if (err)
 				goto error;
-			}
 			priv_dev_select_rx_function(priv, eth_dev);
 			priv_dev_select_tx_function(priv, eth_dev);
 			continue;
@@ -883,6 +988,9 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 			WARN("Rx CQE compression isn't supported");
 			priv->cqe_comp = 0;
 		}
+		err = priv_uar_init_primary(priv);
+		if (err)
+			goto port_error;
 		/* Configure the first MAC address by default. */
 		if (priv_get_mac(priv, &mac.addr_bytes)) {
 			ERROR("cannot get MAC address, is mlx5_en loaded?"
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 004e83670..8f49cecc8 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -166,6 +166,7 @@ struct priv {
 	struct mlx5_xstats_ctrl xstats_ctrl; /* Extended stats control. */
 	rte_spinlock_t lock; /* Lock for control functions. */
 	int primary_socket; /* Unix socket for primary process. */
+	void *uar_base; /* Reserved address space for UAR mapping */
 	struct rte_intr_handle intr_handle_socket; /* Interrupt handler. */
 	struct mlx5_verbs_alloc_ctx verbs_alloc_ctx;
 	/* Context for Verbs allocator. */
diff --git a/drivers/net/mlx5/mlx5_defs.h b/drivers/net/mlx5/mlx5_defs.h
index 24caf7e79..9c740f5e1 100644
--- a/drivers/net/mlx5/mlx5_defs.h
+++ b/drivers/net/mlx5/mlx5_defs.h
@@ -113,4 +113,14 @@
 /* Maximum number of attempts to query link status before giving up. */
 #define MLX5_MAX_LINK_QUERY_ATTEMPTS 5
 
+/* Reserved address space for UAR mapping. */
+#define MLX5_UAR_SIZE (1ULL << 32)
+
+/* Offset of reserved UAR address space to hugepage memory. Offset is used here
+ * to minimize possibility of address next to hugepage being used by other code
+ * in either primary or secondary process, failing to map TX UAR would make TX
+ * packets invisible to HW.
+ */
+#define MLX5_UAR_OFFSET (1ULL << 32)
+
 #endif /* RTE_PMD_MLX5_DEFS_H_ */
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index e35a6ab34..28761c299 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -204,7 +204,7 @@ struct mlx5_txq_data {
 	volatile void *wqes; /* Work queue (use volatile to write into). */
 	volatile uint32_t *qp_db; /* Work queue doorbell. */
 	volatile uint32_t *cq_db; /* Completion queue doorbell. */
-	volatile void *bf_reg; /* Blueflame register. */
+	volatile void *bf_reg; /* Blueflame register remapped. */
 	struct mlx5_mr *mp2mr[MLX5_PMD_TX_MP_CACHE]; /* MR translation table. */
 	struct rte_mbuf *(*elts)[]; /* TX elements. */
 	struct mlx5_txq_stats stats; /* TX queue counters. */
@@ -229,6 +229,7 @@ struct mlx5_txq_ctrl {
 	struct mlx5_txq_ibv *ibv; /* Verbs queue object. */
 	struct mlx5_txq_data txq; /* Data path structure. */
 	off_t uar_mmap_offset; /* UAR mmap offset for non-primary process. */
+	volatile void *bf_reg_orig; /* Blueflame register from verbs. */
 };
 
 /* mlx5_rxq.c */
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index d682ea2c9..da9494eb8 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -76,10 +76,13 @@ priv_txq_start(struct priv *priv)
 			goto error;
 		}
 	}
-	return -ret;
+	ret = priv_tx_uar_remap(priv, priv->ctx->cmd_fd);
+	if (ret)
+		goto error;
+	return ret;
 error:
 	priv_txq_stop(priv);
-	return -ret;
+	return ret;
 }
 
 static void
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 12f258bbb..f39b98a9a 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -217,7 +217,9 @@ mlx5_tx_queue_release(void *dpdk_txq)
 
 
 /**
- * Map locally UAR used in Tx queues for BlueFlame doorbell.
+ * Mmap TX UAR(HW doorbell) pages into reserved UAR address space.
+ * Both primary and secondary process do mmap to make UAR address
+ * aligned.
  *
  * @param[in] priv
  *   Pointer to private structure.
@@ -234,11 +236,14 @@ priv_tx_uar_remap(struct priv *priv, int fd)
 	uintptr_t pages[priv->txqs_n];
 	unsigned int pages_n = 0;
 	uintptr_t uar_va;
+	uintptr_t off;
 	void *addr;
+	void *ret;
 	struct mlx5_txq_data *txq;
 	struct mlx5_txq_ctrl *txq_ctrl;
 	int already_mapped;
 	size_t page_size = sysconf(_SC_PAGESIZE);
+	int r;
 
 	memset(pages, 0, priv->txqs_n * sizeof(uintptr_t));
 	/*
@@ -251,8 +256,10 @@ priv_tx_uar_remap(struct priv *priv, int fd)
 			continue;
 		txq = (*priv->txqs)[i];
 		txq_ctrl = container_of(txq, struct mlx5_txq_ctrl, txq);
-		uar_va = (uintptr_t)txq_ctrl->txq.bf_reg;
-		uar_va = RTE_ALIGN_FLOOR(uar_va, page_size);
+		/* UAR addr form verbs used to find dup and offset in page. */
+		uar_va = (uintptr_t)txq_ctrl->bf_reg_orig;
+		off = uar_va & (page_size - 1); /* offset in page. */
+		uar_va = RTE_ALIGN_FLOOR(uar_va, page_size); /* page addr. */
 		already_mapped = 0;
 		for (j = 0; j != pages_n; ++j) {
 			if (pages[j] == uar_va) {
@@ -260,16 +267,30 @@ priv_tx_uar_remap(struct priv *priv, int fd)
 				break;
 			}
 		}
-		if (already_mapped)
-			continue;
-		pages[pages_n++] = uar_va;
-		addr = mmap((void *)uar_va, page_size,
-			    PROT_WRITE, MAP_FIXED | MAP_SHARED, fd,
-			    txq_ctrl->uar_mmap_offset);
-		if (addr != (void *)uar_va) {
-			ERROR("call to mmap failed on UAR for txq %d\n", i);
-			return -1;
+		/* new address in reserved UAR address space. */
+		addr = RTE_PTR_ADD(priv->uar_base,
+				   uar_va & (MLX5_UAR_SIZE - 1));
+		if (!already_mapped) {
+			pages[pages_n++] = uar_va;
+			/* fixed mmap to specified address in reserved
+			 * address space.
+			 */
+			ret = mmap(addr, page_size,
+				   PROT_WRITE, MAP_FIXED | MAP_SHARED, fd,
+				   txq_ctrl->uar_mmap_offset);
+			if (ret != addr) {
+				/* fixed mmap have to return same address */
+				ERROR("call to mmap failed on UAR for txq %d\n",
+				      i);
+				r = ENXIO;
+				return r;
+			}
 		}
+		if (rte_eal_process_type() == RTE_PROC_PRIMARY) /* save once */
+			txq_ctrl->txq.bf_reg = RTE_PTR_ADD((void *)addr, off);
+		else
+			assert(txq_ctrl->txq.bf_reg ==
+			       RTE_PTR_ADD((void *)addr, off));
 	}
 	return 0;
 }
@@ -416,7 +437,7 @@ mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
 	txq_data->wqes = qp.sq.buf;
 	txq_data->wqe_n = log2above(qp.sq.wqe_cnt);
 	txq_data->qp_db = &qp.dbrec[MLX5_SND_DBR];
-	txq_data->bf_reg = qp.bf.reg;
+	txq_ctrl->bf_reg_orig = qp.bf.reg;
 	txq_data->cq_db = cq_info.dbrec;
 	txq_data->cqes =
 		(volatile struct mlx5_cqe (*)[])
@@ -717,6 +738,7 @@ mlx5_priv_txq_release(struct priv *priv, uint16_t idx)
 {
 	unsigned int i;
 	struct mlx5_txq_ctrl *txq;
+	size_t page_size = sysconf(_SC_PAGESIZE);
 
 	if (!(*priv->txqs)[idx])
 		return 0;
@@ -736,6 +758,9 @@ mlx5_priv_txq_release(struct priv *priv, uint16_t idx)
 			txq->txq.mp2mr[i] = NULL;
 		}
 	}
+	if (priv->uar_base)
+		munmap((void *)RTE_ALIGN_FLOOR((uintptr_t)txq->txq.bf_reg,
+		       page_size), page_size);
 	if (rte_atomic32_dec_and_test(&txq->refcnt)) {
 		txq_free_elts(txq);
 		LIST_REMOVE(txq, next);
-- 
2.11.0

[dpdk-stable] [PATCH v2 11/67] net/mlx4: fix single port configuration

[Yongseok Koh]

From: Ophir Munk <ophirmu@mellanox.com>

[ upstream commit a43fba2c1ea2e2f602c603f8338542f165b1c06b ]

The number of mlx4 present ports is calculated as follows:
conf.ports.present |= (UINT64_C(1) << device_attr.phys_port_cnt) - 1;

That is - all ones sequence (due to -1 subtraction)
When retrieving the number of ports, 1 must be added in order to obtain
the correct number of ports to the power of 2, as follows:
uint32_t ports = rte_log2_u32(conf->ports.present + 1);
If 1 was not added, in the case of one port, the number of ports would
be falsely calculated as 0.

Fixes: 8264279967dc ("net/mlx4: check max number of ports dynamically")
Cc: stable@dpdk.org

Signed-off-by: Ophir Munk <ophirmu@mellanox.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx4/mlx4.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index 97dac6439..9a5773381 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/mlx4.c
@@ -336,7 +336,7 @@ mlx4_arg_parse(const char *key, const char *val, struct mlx4_conf *conf)
 		return -rte_errno;
 	}
 	if (strcmp(MLX4_PMD_PORT_KVARG, key) == 0) {
-		uint32_t ports = rte_log2_u32(conf->ports.present);
+		uint32_t ports = rte_log2_u32(conf->ports.present + 1);
 
 		if (tmp >= ports) {
 			ERROR("port index %lu outside range [0,%" PRIu32 ")",
-- 
2.11.0

[dpdk-stable] [PATCH v2 12/67] net/mlx4: fix broadcast Rx

[Yongseok Koh]

From: Moti Haimovsky <motih@mellanox.com>

[ upstream commit 643958cf910f9d2b6a22307a331834a93722884d ]

This patch fixes the issue of mlx4 not receiving broadcast packets
when configured to work promiscuous or allmulticast modes.

Fixes: eacaac7bae36 ("net/mlx4: restore promisc and allmulti support")
Cc: stable@dpdk.org

Signed-off-by: Moti Haimovsky <motih@mellanox.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx4/mlx4_flow.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/net/mlx4/mlx4_flow.c b/drivers/net/mlx4/mlx4_flow.c
index e81e24dbb..e8ebeebe3 100644
--- a/drivers/net/mlx4/mlx4_flow.c
+++ b/drivers/net/mlx4/mlx4_flow.c
@@ -1217,9 +1217,12 @@ mlx4_flow_internal_next_vlan(struct priv *priv, uint16_t vlan)
  *
  * Various flow rules are created depending on the mode the device is in:
  *
- * 1. Promiscuous: port MAC + catch-all (VLAN filtering is ignored).
- * 2. All multicast: port MAC/VLAN + catch-all multicast.
- * 3. Otherwise: port MAC/VLAN + broadcast MAC/VLAN.
+ * 1. Promiscuous:
+ *       port MAC + broadcast + catch-all (VLAN filtering is ignored).
+ * 2. All multicast:
+ *       port MAC/VLAN + broadcast + catch-all multicast.
+ * 3. Otherwise:
+ *       port MAC/VLAN + broadcast MAC/VLAN.
  *
  * About MAC flow rules:
  *
@@ -1298,9 +1301,6 @@ mlx4_flow_internal(struct priv *priv, struct rte_flow_error *error)
 		!priv->dev->data->promiscuous ?
 		&vlan_spec.tci :
 		NULL;
-	int broadcast =
-		!priv->dev->data->promiscuous &&
-		!priv->dev->data->all_multicast;
 	uint16_t vlan = 0;
 	struct rte_flow *flow;
 	unsigned int i;
@@ -1334,7 +1334,7 @@ mlx4_flow_internal(struct priv *priv, struct rte_flow_error *error)
 			rule_vlan = NULL;
 		}
 	}
-	for (i = 0; i != RTE_DIM(priv->mac) + broadcast; ++i) {
+	for (i = 0; i != RTE_DIM(priv->mac) + 1; ++i) {
 		const struct ether_addr *mac;
 
 		/* Broadcasts are handled by an extra iteration. */
@@ -1398,7 +1398,7 @@ mlx4_flow_internal(struct priv *priv, struct rte_flow_error *error)
 			goto next_vlan;
 	}
 	/* Take care of promiscuous and all multicast flow rules. */
-	if (!broadcast) {
+	if (priv->dev->data->promiscuous || priv->dev->data->all_multicast) {
 		for (flow = LIST_FIRST(&priv->flows);
 		     flow && flow->internal;
 		     flow = LIST_NEXT(flow, next)) {
-- 
2.11.0

[dpdk-stable] [PATCH v2 13/67] net/mlx4: fix removal detection of stopped port

[Yongseok Koh]

From: Moti Haimovsky <motih@mellanox.com>

[ upstream commit fc1b5ec522ee92f303cb74e683a5efe477cf9b56 ]

In failsafe device start can be called for ports/devices that
had been plugged out.
The mlx4 PMD detects device removal by listening to the device RMV
events, when the mlx4 port is being stopped, the PMD no longer
listens to these events causing the PMD to stop detecting device
removals.
This patch fixes this issue by moving installation of the interrupt
handler to device configuration, and toggle only the Rx-queue
interrupts on start/stop.

Fixes: a6e8b01c3c26 ("net/mlx4: compact interrupt functions")
Cc: stable@dpdk.org

Signed-off-by: Moti Haimovsky <motih@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
---
 drivers/net/mlx4/mlx4.c      | 10 ++++++++--
 drivers/net/mlx4/mlx4.h      |  2 ++
 drivers/net/mlx4/mlx4_intr.c | 41 ++++++++++++++++++++++++++++++++++++++---
 3 files changed, 48 insertions(+), 5 deletions(-)

diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index 9a5773381..26a0c85ad 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/mlx4.c
@@ -108,7 +108,13 @@ mlx4_dev_configure(struct rte_eth_dev *dev)
 		      " flow error type %d, cause %p, message: %s",
 		      -ret, strerror(-ret), error.type, error.cause,
 		      error.message ? error.message : "(unspecified)");
+		goto exit;
 	}
+	ret = mlx4_intr_install(priv);
+	if (ret)
+		ERROR("%p: interrupt handler installation failed",
+		      (void *)dev);
+exit:
 	return ret;
 }
 
@@ -141,7 +147,7 @@ mlx4_dev_start(struct rte_eth_dev *dev)
 		      (void *)dev, strerror(-ret));
 		goto err;
 	}
-	ret = mlx4_intr_install(priv);
+	ret = mlx4_rxq_intr_enable(priv);
 	if (ret) {
 		ERROR("%p: interrupt handler installation failed",
 		     (void *)dev);
@@ -187,7 +193,7 @@ mlx4_dev_stop(struct rte_eth_dev *dev)
 	dev->rx_pkt_burst = mlx4_rx_burst_removed;
 	rte_wmb();
 	mlx4_flow_sync(priv, NULL);
-	mlx4_intr_uninstall(priv);
+	mlx4_rxq_intr_disable(priv);
 	mlx4_rss_deinit(priv);
 }
 
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index 3aeef87e9..82573d4ab 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -170,6 +170,8 @@ const uint32_t *mlx4_dev_supported_ptypes_get(struct rte_eth_dev *dev);
 
 int mlx4_intr_uninstall(struct priv *priv);
 int mlx4_intr_install(struct priv *priv);
+int mlx4_rxq_intr_enable(struct priv *priv);
+void mlx4_rxq_intr_disable(struct priv *priv);
 int mlx4_rx_intr_disable(struct rte_eth_dev *dev, uint16_t idx);
 int mlx4_rx_intr_enable(struct rte_eth_dev *dev, uint16_t idx);
 
diff --git a/drivers/net/mlx4/mlx4_intr.c b/drivers/net/mlx4/mlx4_intr.c
index 50d197698..2364cb203 100644
--- a/drivers/net/mlx4/mlx4_intr.c
+++ b/drivers/net/mlx4/mlx4_intr.c
@@ -291,7 +291,7 @@ mlx4_intr_uninstall(struct priv *priv)
 	}
 	rte_eal_alarm_cancel((void (*)(void *))mlx4_link_status_alarm, priv);
 	priv->intr_alarm = 0;
-	mlx4_rx_intr_vec_disable(priv);
+	mlx4_rxq_intr_disable(priv);
 	rte_errno = err;
 	return 0;
 }
@@ -313,8 +313,6 @@ mlx4_intr_install(struct priv *priv)
 	int rc;
 
 	mlx4_intr_uninstall(priv);
-	if (intr_conf->rxq && mlx4_rx_intr_vec_enable(priv) < 0)
-		goto error;
 	if (intr_conf->lsc | intr_conf->rmv) {
 		priv->intr_handle.fd = priv->ctx->async_fd;
 		rc = rte_intr_callback_register(&priv->intr_handle,
@@ -395,3 +393,40 @@ mlx4_rx_intr_enable(struct rte_eth_dev *dev, uint16_t idx)
 	}
 	return -ret;
 }
+
+/**
+ * Enable datapath interrupts.
+ *
+ * @param priv
+ *   Pointer to private structure.
+ *
+ * @return
+ *   0 on success, negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx4_rxq_intr_enable(struct priv *priv)
+{
+	const struct rte_intr_conf *const intr_conf =
+		&priv->dev->data->dev_conf.intr_conf;
+
+	if (intr_conf->rxq && mlx4_rx_intr_vec_enable(priv) < 0)
+		goto error;
+	return 0;
+error:
+	return -rte_errno;
+}
+
+/**
+ * Disable datapath interrupts, keeping other interrupts intact.
+ *
+ * @param priv
+ *   Pointer to private structure.
+ */
+void
+mlx4_rxq_intr_disable(struct priv *priv)
+{
+	int err = rte_errno; /* Make sure rte_errno remains unchanged. */
+
+	mlx4_rx_intr_vec_disable(priv);
+	rte_errno = err;
+}
-- 
2.11.0

[dpdk-stable] [PATCH v2 14/67] net/mlx5: fix CRC strip capability query

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ backported from upstream commit cd230a3ed985894efafe326cbbf5db7f9e45b25c ]

IBV_WQ_FLAGS_SCATTER_FCS is a WQ flag to be used to configure the CRC
strip on a queue upon creation.

Using IBV_RAW_PACKET_CAP_SCATTER_FCS instead to query to capability.
Even though this is RAW_QP capability, it is being used by rdma-core to
indicate for both RAW_QP and WQ.

Fixes: 43e9d9794cde ("net/mlx5: support upstream rdma-core")
Cc: stable@dpdk.org

Reported-by: Alex Rosenbaum <alexr@mellanox.com>
Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 7dc3dc399..f83508cf9 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -939,9 +939,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		DEBUG("VLAN stripping is %ssupported",
 		      (priv->hw_vlan_strip ? "" : "not "));
 
-		priv->hw_fcs_strip =
-				!!(device_attr_ex.orig_attr.device_cap_flags &
-				IBV_WQ_FLAGS_SCATTER_FCS);
+		priv->hw_fcs_strip = !!(device_attr_ex.raw_packet_caps &
+					IBV_RAW_PACKET_CAP_SCATTER_FCS);
 		DEBUG("FCS stripping configuration is %ssupported",
 		      (priv->hw_fcs_strip ? "" : "not "));
 
-- 
2.11.0

[dpdk-stable] [PATCH v2 15/67] net/mlx5: fix close after start failure

[Yongseok Koh]

From: Xueming Li <xuemingl@mellanox.com>

[ upstream commit 8c5bca92c92cce90239ca813f293f3df53427775 ]

This patch fixed primary socket assertion error during close on a device
that failed to start.

Fixes: f8b9a3bad467 ("net/mlx5: install a socket to exchange a file descriptor")
Cc: stable@dpdk.org

Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index f83508cf9..6ad5aa7e3 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -243,7 +243,8 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 		rte_free(priv->rss_conf.rss_key);
 	if (priv->reta_idx != NULL)
 		rte_free(priv->reta_idx);
-	priv_socket_uninit(priv);
+	if (priv->primary_socket)
+		priv_socket_uninit(priv);
 	ret = mlx5_priv_hrxq_ibv_verify(priv);
 	if (ret)
 		WARN("%p: some Hash Rx queue still remain", (void *)priv);
-- 
2.11.0

[dpdk-stable] [PATCH v2 16/67] net/mlx: control netdevices through ioctl only

[Yongseok Koh]

From: Adrien Mazarguil <adrien.mazarguil@6wind.com>

[ upstream commit fc40db997323bb0e9b725a6e8d65eae95372446c ]

Several control operations implemented by these PMDs affect netdevices
through sysfs, itself subject to file system permission checks enforced by
the kernel, which limits their use for most purposes to applications
running with root privileges.

Since performing the same operations through ioctl() requires fewer
capabilities (only CAP_NET_ADMIN) and given the remaining operations are
already implemented this way, this patch standardizes on ioctl() and gets
rid of redundant code.

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
 drivers/net/mlx4/mlx4_ethdev.c | 192 ++-----------------------------
 drivers/net/mlx5/mlx5.h        |   2 -
 drivers/net/mlx5/mlx5_ethdev.c | 255 +++++------------------------------------
 drivers/net/mlx5/mlx5_stats.c  |  28 ++++-
 4 files changed, 63 insertions(+), 414 deletions(-)

diff --git a/drivers/net/mlx4/mlx4_ethdev.c b/drivers/net/mlx4/mlx4_ethdev.c
index 2f69e7d4f..89f552c87 100644
--- a/drivers/net/mlx4/mlx4_ethdev.c
+++ b/drivers/net/mlx4/mlx4_ethdev.c
@@ -159,167 +159,6 @@ mlx4_get_ifname(const struct priv *priv, char (*ifname)[IF_NAMESIZE])
 }
 
 /**
- * Read from sysfs entry.
- *
- * @param[in] priv
- *   Pointer to private structure.
- * @param[in] entry
- *   Entry name relative to sysfs path.
- * @param[out] buf
- *   Data output buffer.
- * @param size
- *   Buffer size.
- *
- * @return
- *   Number of bytes read on success, negative errno value otherwise and
- *   rte_errno is set.
- */
-static int
-mlx4_sysfs_read(const struct priv *priv, const char *entry,
-		char *buf, size_t size)
-{
-	char ifname[IF_NAMESIZE];
-	FILE *file;
-	int ret;
-
-	ret = mlx4_get_ifname(priv, &ifname);
-	if (ret)
-		return ret;
-
-	MKSTR(path, "%s/device/net/%s/%s", priv->ctx->device->ibdev_path,
-	      ifname, entry);
-
-	file = fopen(path, "rb");
-	if (file == NULL) {
-		rte_errno = errno;
-		return -rte_errno;
-	}
-	ret = fread(buf, 1, size, file);
-	if ((size_t)ret < size && ferror(file)) {
-		rte_errno = EIO;
-		ret = -rte_errno;
-	} else {
-		ret = size;
-	}
-	fclose(file);
-	return ret;
-}
-
-/**
- * Write to sysfs entry.
- *
- * @param[in] priv
- *   Pointer to private structure.
- * @param[in] entry
- *   Entry name relative to sysfs path.
- * @param[in] buf
- *   Data buffer.
- * @param size
- *   Buffer size.
- *
- * @return
- *   Number of bytes written on success, negative errno value otherwise and
- *   rte_errno is set.
- */
-static int
-mlx4_sysfs_write(const struct priv *priv, const char *entry,
-		 char *buf, size_t size)
-{
-	char ifname[IF_NAMESIZE];
-	FILE *file;
-	int ret;
-
-	ret = mlx4_get_ifname(priv, &ifname);
-	if (ret)
-		return ret;
-
-	MKSTR(path, "%s/device/net/%s/%s", priv->ctx->device->ibdev_path,
-	      ifname, entry);
-
-	file = fopen(path, "wb");
-	if (file == NULL) {
-		rte_errno = errno;
-		return -rte_errno;
-	}
-	ret = fwrite(buf, 1, size, file);
-	if ((size_t)ret < size || ferror(file)) {
-		rte_errno = EIO;
-		ret = -rte_errno;
-	} else {
-		ret = size;
-	}
-	fclose(file);
-	return ret;
-}
-
-/**
- * Get unsigned long sysfs property.
- *
- * @param priv
- *   Pointer to private structure.
- * @param[in] name
- *   Entry name relative to sysfs path.
- * @param[out] value
- *   Value output buffer.
- *
- * @return
- *   0 on success, negative errno value otherwise and rte_errno is set.
- */
-static int
-mlx4_get_sysfs_ulong(struct priv *priv, const char *name, unsigned long *value)
-{
-	int ret;
-	unsigned long value_ret;
-	char value_str[32];
-
-	ret = mlx4_sysfs_read(priv, name, value_str, (sizeof(value_str) - 1));
-	if (ret < 0) {
-		DEBUG("cannot read %s value from sysfs: %s",
-		      name, strerror(rte_errno));
-		return ret;
-	}
-	value_str[ret] = '\0';
-	errno = 0;
-	value_ret = strtoul(value_str, NULL, 0);
-	if (errno) {
-		rte_errno = errno;
-		DEBUG("invalid %s value `%s': %s", name, value_str,
-		      strerror(rte_errno));
-		return -rte_errno;
-	}
-	*value = value_ret;
-	return 0;
-}
-
-/**
- * Set unsigned long sysfs property.
- *
- * @param priv
- *   Pointer to private structure.
- * @param[in] name
- *   Entry name relative to sysfs path.
- * @param value
- *   Value to set.
- *
- * @return
- *   0 on success, negative errno value otherwise and rte_errno is set.
- */
-static int
-mlx4_set_sysfs_ulong(struct priv *priv, const char *name, unsigned long value)
-{
-	int ret;
-	MKSTR(value_str, "%lu", value);
-
-	ret = mlx4_sysfs_write(priv, name, value_str, (sizeof(value_str) - 1));
-	if (ret < 0) {
-		DEBUG("cannot write %s `%s' (%lu) to sysfs: %s",
-		      name, value_str, value, strerror(rte_errno));
-		return ret;
-	}
-	return 0;
-}
-
-/**
  * Perform ifreq ioctl() on associated Ethernet device.
  *
  * @param[in] priv
@@ -388,12 +227,12 @@ mlx4_get_mac(struct priv *priv, uint8_t (*mac)[ETHER_ADDR_LEN])
 int
 mlx4_mtu_get(struct priv *priv, uint16_t *mtu)
 {
-	unsigned long ulong_mtu = 0;
-	int ret = mlx4_get_sysfs_ulong(priv, "mtu", &ulong_mtu);
+	struct ifreq request;
+	int ret = mlx4_ifreq(priv, SIOCGIFMTU, &request);
 
 	if (ret)
 		return ret;
-	*mtu = ulong_mtu;
+	*mtu = request.ifr_mtu;
 	return 0;
 }
 
@@ -412,20 +251,13 @@ int
 mlx4_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
 {
 	struct priv *priv = dev->data->dev_private;
-	uint16_t new_mtu;
-	int ret = mlx4_set_sysfs_ulong(priv, "mtu", mtu);
+	struct ifreq request = { .ifr_mtu = mtu, };
+	int ret = mlx4_ifreq(priv, SIOCSIFMTU, &request);
 
 	if (ret)
 		return ret;
-	ret = mlx4_mtu_get(priv, &new_mtu);
-	if (ret)
-		return ret;
-	if (new_mtu == mtu) {
-		priv->mtu = mtu;
-		return 0;
-	}
-	rte_errno = EINVAL;
-	return -rte_errno;
+	priv->mtu = mtu;
+	return 0;
 }
 
 /**
@@ -444,14 +276,14 @@ mlx4_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
 static int
 mlx4_set_flags(struct priv *priv, unsigned int keep, unsigned int flags)
 {
-	unsigned long tmp = 0;
-	int ret = mlx4_get_sysfs_ulong(priv, "flags", &tmp);
+	struct ifreq request;
+	int ret = mlx4_ifreq(priv, SIOCGIFFLAGS, &request);
 
 	if (ret)
 		return ret;
-	tmp &= keep;
-	tmp |= (flags & (~keep));
-	return mlx4_set_sysfs_ulong(priv, "flags", tmp);
+	request.ifr_flags &= keep;
+	request.ifr_flags |= flags & ~keep;
+	return mlx4_ifreq(priv, SIOCSIFFLAGS, &request);
 }
 
 /**
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 8f49cecc8..1b01f4603 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -223,8 +223,6 @@ struct priv *mlx5_get_priv(struct rte_eth_dev *dev);
 int mlx5_is_secondary(void);
 int priv_get_ifname(const struct priv *, char (*)[IF_NAMESIZE]);
 int priv_ifreq(const struct priv *, int req, struct ifreq *);
-int priv_is_ib_cntr(const char *);
-int priv_get_cntr_sysfs(struct priv *, const char *, uint64_t *);
 int priv_get_num_vfs(struct priv *, uint16_t *);
 int priv_get_mtu(struct priv *, uint16_t *);
 int priv_set_flags(struct priv *, unsigned int, unsigned int);
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 3e5a2922d..0a36a839c 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -35,6 +35,7 @@
 
 #include <stddef.h>
 #include <assert.h>
+#include <inttypes.h>
 #include <unistd.h>
 #include <stdint.h>
 #include <stdio.h>
@@ -200,181 +201,6 @@ priv_get_ifname(const struct priv *priv, char (*ifname)[IF_NAMESIZE])
 }
 
 /**
- * Check if the counter is located on ib counters file.
- *
- * @param[in] cntr
- *   Counter name.
- *
- * @return
- *   1 if counter is located on ib counters file , 0 otherwise.
- */
-int
-priv_is_ib_cntr(const char *cntr)
-{
-	if (!strcmp(cntr, "out_of_buffer"))
-		return 1;
-	return 0;
-}
-
-/**
- * Read from sysfs entry.
- *
- * @param[in] priv
- *   Pointer to private structure.
- * @param[in] entry
- *   Entry name relative to sysfs path.
- * @param[out] buf
- *   Data output buffer.
- * @param size
- *   Buffer size.
- *
- * @return
- *   0 on success, -1 on failure and errno is set.
- */
-static int
-priv_sysfs_read(const struct priv *priv, const char *entry,
-		char *buf, size_t size)
-{
-	char ifname[IF_NAMESIZE];
-	FILE *file;
-	int ret;
-	int err;
-
-	if (priv_get_ifname(priv, &ifname))
-		return -1;
-
-	if (priv_is_ib_cntr(entry)) {
-		MKSTR(path, "%s/ports/1/hw_counters/%s",
-		      priv->ibdev_path, entry);
-		file = fopen(path, "rb");
-	} else {
-		MKSTR(path, "%s/device/net/%s/%s",
-		      priv->ibdev_path, ifname, entry);
-		file = fopen(path, "rb");
-	}
-	if (file == NULL)
-		return -1;
-	ret = fread(buf, 1, size, file);
-	err = errno;
-	if (((size_t)ret < size) && (ferror(file)))
-		ret = -1;
-	else
-		ret = size;
-	fclose(file);
-	errno = err;
-	return ret;
-}
-
-/**
- * Write to sysfs entry.
- *
- * @param[in] priv
- *   Pointer to private structure.
- * @param[in] entry
- *   Entry name relative to sysfs path.
- * @param[in] buf
- *   Data buffer.
- * @param size
- *   Buffer size.
- *
- * @return
- *   0 on success, -1 on failure and errno is set.
- */
-static int
-priv_sysfs_write(const struct priv *priv, const char *entry,
-		 char *buf, size_t size)
-{
-	char ifname[IF_NAMESIZE];
-	FILE *file;
-	int ret;
-	int err;
-
-	if (priv_get_ifname(priv, &ifname))
-		return -1;
-
-	MKSTR(path, "%s/device/net/%s/%s", priv->ibdev_path, ifname, entry);
-
-	file = fopen(path, "wb");
-	if (file == NULL)
-		return -1;
-	ret = fwrite(buf, 1, size, file);
-	err = errno;
-	if (((size_t)ret < size) || (ferror(file)))
-		ret = -1;
-	else
-		ret = size;
-	fclose(file);
-	errno = err;
-	return ret;
-}
-
-/**
- * Get unsigned long sysfs property.
- *
- * @param priv
- *   Pointer to private structure.
- * @param[in] name
- *   Entry name relative to sysfs path.
- * @param[out] value
- *   Value output buffer.
- *
- * @return
- *   0 on success, -1 on failure and errno is set.
- */
-static int
-priv_get_sysfs_ulong(struct priv *priv, const char *name, unsigned long *value)
-{
-	int ret;
-	unsigned long value_ret;
-	char value_str[32];
-
-	ret = priv_sysfs_read(priv, name, value_str, (sizeof(value_str) - 1));
-	if (ret == -1) {
-		DEBUG("cannot read %s value from sysfs: %s",
-		      name, strerror(errno));
-		return -1;
-	}
-	value_str[ret] = '\0';
-	errno = 0;
-	value_ret = strtoul(value_str, NULL, 0);
-	if (errno) {
-		DEBUG("invalid %s value `%s': %s", name, value_str,
-		      strerror(errno));
-		return -1;
-	}
-	*value = value_ret;
-	return 0;
-}
-
-/**
- * Set unsigned long sysfs property.
- *
- * @param priv
- *   Pointer to private structure.
- * @param[in] name
- *   Entry name relative to sysfs path.
- * @param value
- *   Value to set.
- *
- * @return
- *   0 on success, -1 on failure and errno is set.
- */
-static int
-priv_set_sysfs_ulong(struct priv *priv, const char *name, unsigned long value)
-{
-	int ret;
-	MKSTR(value_str, "%lu", value);
-
-	ret = priv_sysfs_write(priv, name, value_str, (sizeof(value_str) - 1));
-	if (ret == -1) {
-		DEBUG("cannot write %s `%s' (%lu) to sysfs: %s",
-		      name, value_str, value, strerror(errno));
-		return -1;
-	}
-	return 0;
-}
-
-/**
  * Perform ifreq ioctl() on associated Ethernet device.
  *
  * @param[in] priv
@@ -417,20 +243,25 @@ priv_get_num_vfs(struct priv *priv, uint16_t *num_vfs)
 {
 	/* The sysfs entry name depends on the operating system. */
 	const char **name = (const char *[]){
-		"device/sriov_numvfs",
-		"device/mlx5_num_vfs",
+		"sriov_numvfs",
+		"mlx5_num_vfs",
 		NULL,
 	};
-	int ret;
 
 	do {
-		unsigned long ulong_num_vfs;
+		int n;
+		FILE *file;
+		MKSTR(path, "%s/device/%s", priv->ibdev_path, *name);
 
-		ret = priv_get_sysfs_ulong(priv, *name, &ulong_num_vfs);
-		if (!ret)
-			*num_vfs = ulong_num_vfs;
-	} while (*(++name) && ret);
-	return ret;
+		file = fopen(path, "rb");
+		if (!file)
+			continue;
+		n = fscanf(file, "%" SCNu16, num_vfs);
+		fclose(file);
+		if (n == 1)
+			return 0;
+	} while (*(++name));
+	return -1;
 }
 
 /**
@@ -447,35 +278,12 @@ priv_get_num_vfs(struct priv *priv, uint16_t *num_vfs)
 int
 priv_get_mtu(struct priv *priv, uint16_t *mtu)
 {
-	unsigned long ulong_mtu;
+	struct ifreq request;
+	int ret = priv_ifreq(priv, SIOCGIFMTU, &request);
 
-	if (priv_get_sysfs_ulong(priv, "mtu", &ulong_mtu) == -1)
-		return -1;
-	*mtu = ulong_mtu;
-	return 0;
-}
-
-/**
- * Read device counter from sysfs.
- *
- * @param priv
- *   Pointer to private structure.
- * @param name
- *   Counter name.
- * @param[out] cntr
- *   Counter output buffer.
- *
- * @return
- *   0 on success, -1 on failure and errno is set.
- */
-int
-priv_get_cntr_sysfs(struct priv *priv, const char *name, uint64_t *cntr)
-{
-	unsigned long ulong_ctr;
-
-	if (priv_get_sysfs_ulong(priv, name, &ulong_ctr) == -1)
-		return -1;
-	*cntr = ulong_ctr;
+	if (ret)
+		return ret;
+	*mtu = request.ifr_mtu;
 	return 0;
 }
 
@@ -493,15 +301,9 @@ priv_get_cntr_sysfs(struct priv *priv, const char *name, uint64_t *cntr)
 static int
 priv_set_mtu(struct priv *priv, uint16_t mtu)
 {
-	uint16_t new_mtu;
+	struct ifreq request = { .ifr_mtu = mtu, };
 
-	if (priv_set_sysfs_ulong(priv, "mtu", mtu) ||
-	    priv_get_mtu(priv, &new_mtu))
-		return -1;
-	if (new_mtu == mtu)
-		return 0;
-	errno = EINVAL;
-	return -1;
+	return priv_ifreq(priv, SIOCSIFMTU, &request);
 }
 
 /**
@@ -520,13 +322,14 @@ priv_set_mtu(struct priv *priv, uint16_t mtu)
 int
 priv_set_flags(struct priv *priv, unsigned int keep, unsigned int flags)
 {
-	unsigned long tmp;
+	struct ifreq request;
+	int ret = priv_ifreq(priv, SIOCGIFFLAGS, &request);
 
-	if (priv_get_sysfs_ulong(priv, "flags", &tmp) == -1)
-		return -1;
-	tmp &= keep;
-	tmp |= (flags & (~keep));
-	return priv_set_sysfs_ulong(priv, "flags", tmp);
+	if (ret)
+		return ret;
+	request.ifr_flags &= keep;
+	request.ifr_flags |= flags & ~keep;
+	return priv_ifreq(priv, SIOCSIFFLAGS, &request);
 }
 
 /**
diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index c11ee57f0..7943c4c3f 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -31,8 +31,11 @@
  *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 
+#include <inttypes.h>
 #include <linux/sockios.h>
 #include <linux/ethtool.h>
+#include <stdint.h>
+#include <stdio.h>
 
 #include <rte_ethdev.h>
 #include <rte_common.h>
@@ -47,6 +50,7 @@ struct mlx5_counter_ctrl {
 	char dpdk_name[RTE_ETH_XSTATS_NAME_SIZE];
 	/* Name of the counter on the device table. */
 	char ctr_name[RTE_ETH_XSTATS_NAME_SIZE];
+	uint32_t ib:1; /**< Nonzero for IB counters. */
 };
 
 static const struct mlx5_counter_ctrl mlx5_counters_init[] = {
@@ -121,6 +125,7 @@ static const struct mlx5_counter_ctrl mlx5_counters_init[] = {
 	{
 		.dpdk_name = "rx_out_of_buffer",
 		.ctr_name = "out_of_buffer",
+		.ib = 1,
 	},
 };
 
@@ -155,13 +160,24 @@ priv_read_dev_counters(struct priv *priv, uint64_t *stats)
 		return -1;
 	}
 	for (i = 0; i != xstats_n; ++i) {
-		if (priv_is_ib_cntr(mlx5_counters_init[i].ctr_name))
-			priv_get_cntr_sysfs(priv,
-					    mlx5_counters_init[i].ctr_name,
-					    &stats[i]);
-		else
+		if (mlx5_counters_init[i].ib) {
+			FILE *file;
+			MKSTR(path, "%s/ports/1/hw_counters/%s",
+			      priv->ibdev_path,
+			      mlx5_counters_init[i].ctr_name);
+
+			file = fopen(path, "rb");
+			if (file) {
+				int n = fscanf(file, "%" SCNu64, &stats[i]);
+
+				fclose(file);
+				if (n != 1)
+					stats[i] = 0;
+			}
+		} else {
 			stats[i] = (uint64_t)
 				et_stats->data[xstats_ctrl->dev_table_idx[i]];
+		}
 	}
 	return 0;
 }
@@ -244,7 +260,7 @@ priv_xstats_init(struct priv *priv)
 		}
 	}
 	for (j = 0; j != xstats_n; ++j) {
-		if (priv_is_ib_cntr(mlx5_counters_init[j].ctr_name))
+		if (mlx5_counters_init[j].ib)
 			continue;
 		if (xstats_ctrl->dev_table_idx[j] >= dev_stats_n) {
 			WARN("counter \"%s\" is not recognized",
-- 
2.11.0

[dpdk-stable] [PATCH v2 17/67] net/mlx5: fix disabling Tx packet inlining

[Yongseok Koh]

[ backported from upstream commit 24a8f52455ff69826da460068ba1a5db17cc1085 ]

Adding 'txq_inline=0' to PMD parameter should disable Tx packet inlining
but it doesn't work properly for Enhanced Multi-Packet Send.

Fixes: 6ce84bd88919 ("net/mlx5: add enhanced multi-packet send for ConnectX-5")
Cc: stable@dpdk.org

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_rxtx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index c2dab1f4c..9955faedf 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -1499,7 +1499,7 @@ mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 				max_wqe--;
 			else
 				max_wqe -= 2;
-		} else if (do_inline) {
+		} else if (max_inline && do_inline) {
 			/* Inline packet into WQE. */
 			unsigned int max;
 
-- 
2.11.0

[dpdk-stable] [PATCH v2 18/67] net/mlx5: fix sriov flag

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit a61888c8f297eac7e917850e2a4f66601bb715c1 ]

priv_get_num_vfs() was used to help the PMD in prefetching the mbuf in
datapath when the PMD was behaving in VF mode.
This knowledge is no more used.

Fixes: 528a9fbec6de ("net/mlx5: support ConnectX-5 devices")
Cc: stable@dpdk.org

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.c        | 18 ++----------------
 drivers/net/mlx5/mlx5.h        |  2 --
 drivers/net/mlx5/mlx5_ethdev.c | 37 -------------------------------------
 3 files changed, 2 insertions(+), 55 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 6ad5aa7e3..c3d57c58c 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -653,7 +653,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 	int err = 0;
 	struct ibv_context *attr_ctx = NULL;
 	struct ibv_device_attr_ex device_attr;
-	unsigned int sriov;
 	unsigned int mps;
 	unsigned int cqe_comp;
 	unsigned int tunnel_en = 0;
@@ -700,14 +699,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		    (pci_dev->addr.devid != pci_addr.devid) ||
 		    (pci_dev->addr.function != pci_addr.function))
 			continue;
-		sriov = ((pci_dev->id.device_id ==
-		       PCI_DEVICE_ID_MELLANOX_CONNECTX4VF) ||
-		      (pci_dev->id.device_id ==
-		       PCI_DEVICE_ID_MELLANOX_CONNECTX4LXVF) ||
-		      (pci_dev->id.device_id ==
-		       PCI_DEVICE_ID_MELLANOX_CONNECTX5VF) ||
-		      (pci_dev->id.device_id ==
-		       PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF));
 		switch (pci_dev->id.device_id) {
 		case PCI_DEVICE_ID_MELLANOX_CONNECTX4:
 			tunnel_en = 1;
@@ -722,10 +713,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		default:
 			break;
 		}
-		INFO("PCI information matches, using device \"%s\""
-		     " (SR-IOV: %s)",
-		     list[i]->name,
-		     sriov ? "true" : "false");
+		INFO("PCI information matches, using device \"%s\"",
+		     list[i]->name);
 		attr_ctx = ibv_open_device(list[i]);
 		err = errno;
 		break;
@@ -783,7 +772,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		struct rte_eth_dev *eth_dev;
 		struct ibv_device_attr_ex device_attr_ex;
 		struct ether_addr mac;
-		uint16_t num_vfs = 0;
 		struct ibv_device_attr_ex device_attr;
 		struct mlx5_args args = {
 			.cqe_comp = MLX5_ARG_UNSET,
@@ -951,8 +939,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		DEBUG("hardware RX end alignment padding is %ssupported",
 		      (priv->hw_padding ? "" : "not "));
 
-		priv_get_num_vfs(priv, &num_vfs);
-		priv->sriov = (num_vfs || sriov);
 		priv->tso = ((priv->tso) &&
 			    (device_attr_ex.tso_caps.max_tso > 0) &&
 			    (device_attr_ex.tso_caps.supported_qpts &
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 1b01f4603..2fa0aa274 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -125,7 +125,6 @@ struct priv {
 	unsigned int hw_vlan_strip:1; /* VLAN stripping is supported. */
 	unsigned int hw_fcs_strip:1; /* FCS stripping is supported. */
 	unsigned int hw_padding:1; /* End alignment padding is supported. */
-	unsigned int sriov:1; /* This is a VF or PF with VF devices. */
 	unsigned int mps:2; /* Multi-packet send mode (0: disabled). */
 	unsigned int mpw_hdr_dseg:1; /* Enable DSEGs in the title WQEBB. */
 	unsigned int cqe_comp:1; /* Whether CQE compression is enabled. */
@@ -223,7 +222,6 @@ struct priv *mlx5_get_priv(struct rte_eth_dev *dev);
 int mlx5_is_secondary(void);
 int priv_get_ifname(const struct priv *, char (*)[IF_NAMESIZE]);
 int priv_ifreq(const struct priv *, int req, struct ifreq *);
-int priv_get_num_vfs(struct priv *, uint16_t *);
 int priv_get_mtu(struct priv *, uint16_t *);
 int priv_set_flags(struct priv *, unsigned int, unsigned int);
 int mlx5_dev_configure(struct rte_eth_dev *);
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 0a36a839c..f9f04ec6d 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -228,43 +228,6 @@ priv_ifreq(const struct priv *priv, int req, struct ifreq *ifr)
 }
 
 /**
- * Return the number of active VFs for the current device.
- *
- * @param[in] priv
- *   Pointer to private structure.
- * @param[out] num_vfs
- *   Number of active VFs.
- *
- * @return
- *   0 on success, -1 on failure and errno is set.
- */
-int
-priv_get_num_vfs(struct priv *priv, uint16_t *num_vfs)
-{
-	/* The sysfs entry name depends on the operating system. */
-	const char **name = (const char *[]){
-		"sriov_numvfs",
-		"mlx5_num_vfs",
-		NULL,
-	};
-
-	do {
-		int n;
-		FILE *file;
-		MKSTR(path, "%s/device/%s", priv->ibdev_path, *name);
-
-		file = fopen(path, "rb");
-		if (!file)
-			continue;
-		n = fscanf(file, "%" SCNu16, num_vfs);
-		fclose(file);
-		if (n == 1)
-			return 0;
-	} while (*(++name));
-	return -1;
-}
-
-/**
  * Get device MTU.
  *
  * @param priv
-- 
2.11.0

[dpdk-stable] [PATCH v2 19/67] net/mlx5: name parameters in function prototypes

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 3692c7ec9e97efeefb497cbbdb99aa766a163221 ]

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.h      | 185 +++++++++++++++++++++++--------------------
 drivers/net/mlx5/mlx5_rxtx.h | 154 +++++++++++++++++++----------------
 2 files changed, 187 insertions(+), 152 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 2fa0aa274..dc165a8a3 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -220,25 +220,29 @@ int mlx5_getenv_int(const char *);
 
 struct priv *mlx5_get_priv(struct rte_eth_dev *dev);
 int mlx5_is_secondary(void);
-int priv_get_ifname(const struct priv *, char (*)[IF_NAMESIZE]);
-int priv_ifreq(const struct priv *, int req, struct ifreq *);
-int priv_get_mtu(struct priv *, uint16_t *);
-int priv_set_flags(struct priv *, unsigned int, unsigned int);
-int mlx5_dev_configure(struct rte_eth_dev *);
-void mlx5_dev_infos_get(struct rte_eth_dev *, struct rte_eth_dev_info *);
+int priv_get_ifname(const struct priv *priv, char (*ifname)[IF_NAMESIZE]);
+int priv_ifreq(const struct priv *priv, int req, struct ifreq *ifr);
+int priv_get_mtu(struct priv *priv, uint16_t *mtu);
+int priv_set_flags(struct priv *priv, unsigned int keep, unsigned int flags);
+int mlx5_dev_configure(struct rte_eth_dev *dev);
+void mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info);
 const uint32_t *mlx5_dev_supported_ptypes_get(struct rte_eth_dev *dev);
-int priv_link_update(struct priv *, int);
-int priv_force_link_status_change(struct priv *, int);
-int mlx5_link_update(struct rte_eth_dev *, int);
-int mlx5_dev_set_mtu(struct rte_eth_dev *, uint16_t);
-int mlx5_dev_get_flow_ctrl(struct rte_eth_dev *, struct rte_eth_fc_conf *);
-int mlx5_dev_set_flow_ctrl(struct rte_eth_dev *, struct rte_eth_fc_conf *);
-int mlx5_ibv_device_to_pci_addr(const struct ibv_device *,
-				struct rte_pci_addr *);
-void mlx5_dev_link_status_handler(void *);
-void mlx5_dev_interrupt_handler(void *);
-void priv_dev_interrupt_handler_uninstall(struct priv *, struct rte_eth_dev *);
-void priv_dev_interrupt_handler_install(struct priv *, struct rte_eth_dev *);
+int priv_link_update(struct priv *priv, int wait_to_complete);
+int priv_force_link_status_change(struct priv *priv, int status);
+int mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete);
+int mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu);
+int mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev,
+			   struct rte_eth_fc_conf *fc_conf);
+int mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev,
+			   struct rte_eth_fc_conf *fc_conf);
+int mlx5_ibv_device_to_pci_addr(const struct ibv_device *device,
+				struct rte_pci_addr *pci_addr);
+void mlx5_dev_link_status_handler(void *arg);
+void mlx5_dev_interrupt_handler(void *cb_arg);
+void priv_dev_interrupt_handler_uninstall(struct priv *priv,
+					  struct rte_eth_dev *dev);
+void priv_dev_interrupt_handler_install(struct priv *priv,
+					struct rte_eth_dev *dev);
 int mlx5_set_link_down(struct rte_eth_dev *dev);
 int mlx5_set_link_up(struct rte_eth_dev *dev);
 void priv_dev_select_tx_function(struct priv *priv, struct rte_eth_dev *dev);
@@ -246,86 +250,99 @@ void priv_dev_select_rx_function(struct priv *priv, struct rte_eth_dev *dev);
 
 /* mlx5_mac.c */
 
-int priv_get_mac(struct priv *, uint8_t (*)[ETHER_ADDR_LEN]);
-void mlx5_mac_addr_remove(struct rte_eth_dev *, uint32_t);
-int mlx5_mac_addr_add(struct rte_eth_dev *, struct ether_addr *, uint32_t,
-		      uint32_t);
-void mlx5_mac_addr_set(struct rte_eth_dev *, struct ether_addr *);
+int priv_get_mac(struct priv *priv, uint8_t (*mac)[ETHER_ADDR_LEN]);
+void mlx5_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index);
+int mlx5_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac,
+		      uint32_t index, uint32_t vmdq);
+void mlx5_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr);
 
 /* mlx5_rss.c */
 
-int mlx5_rss_hash_update(struct rte_eth_dev *, struct rte_eth_rss_conf *);
-int mlx5_rss_hash_conf_get(struct rte_eth_dev *, struct rte_eth_rss_conf *);
-int priv_rss_reta_index_resize(struct priv *, unsigned int);
-int mlx5_dev_rss_reta_query(struct rte_eth_dev *,
-			    struct rte_eth_rss_reta_entry64 *, uint16_t);
-int mlx5_dev_rss_reta_update(struct rte_eth_dev *,
-			     struct rte_eth_rss_reta_entry64 *, uint16_t);
+int mlx5_rss_hash_update(struct rte_eth_dev *dev,
+			 struct rte_eth_rss_conf *rss_conf);
+int mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
+			   struct rte_eth_rss_conf *rss_conf);
+int priv_rss_reta_index_resize(struct priv *priv, unsigned int reta_size);
+int mlx5_dev_rss_reta_query(struct rte_eth_dev *dev,
+			    struct rte_eth_rss_reta_entry64 *reta_conf,
+			    uint16_t reta_size);
+int mlx5_dev_rss_reta_update(struct rte_eth_dev *dev,
+			     struct rte_eth_rss_reta_entry64 *reta_conf,
+			     uint16_t reta_size);
 
 /* mlx5_rxmode.c */
 
-void mlx5_promiscuous_enable(struct rte_eth_dev *);
-void mlx5_promiscuous_disable(struct rte_eth_dev *);
-void mlx5_allmulticast_enable(struct rte_eth_dev *);
-void mlx5_allmulticast_disable(struct rte_eth_dev *);
+void mlx5_promiscuous_enable(struct rte_eth_dev *dev);
+void mlx5_promiscuous_disable(struct rte_eth_dev *dev);
+void mlx5_allmulticast_enable(struct rte_eth_dev *dev);
+void mlx5_allmulticast_disable(struct rte_eth_dev *dev);
 
 /* mlx5_stats.c */
 
-void priv_xstats_init(struct priv *);
-int mlx5_stats_get(struct rte_eth_dev *, struct rte_eth_stats *);
-void mlx5_stats_reset(struct rte_eth_dev *);
-int mlx5_xstats_get(struct rte_eth_dev *,
-		    struct rte_eth_xstat *, unsigned int);
-void mlx5_xstats_reset(struct rte_eth_dev *);
-int mlx5_xstats_get_names(struct rte_eth_dev *,
-			  struct rte_eth_xstat_name *, unsigned int);
+void priv_xstats_init(struct priv *priv);
+int mlx5_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats);
+void mlx5_stats_reset(struct rte_eth_dev *dev);
+int mlx5_xstats_get(struct rte_eth_dev *dev,
+		    struct rte_eth_xstat *stats, unsigned int n);
+void mlx5_xstats_reset(struct rte_eth_dev *dev);
+int mlx5_xstats_get_names(struct rte_eth_dev *dev,
+			  struct rte_eth_xstat_name *xstats_names,
+			  unsigned int n);
 
 /* mlx5_vlan.c */
 
-int mlx5_vlan_filter_set(struct rte_eth_dev *, uint16_t, int);
-int mlx5_vlan_offload_set(struct rte_eth_dev *, int);
-void mlx5_vlan_strip_queue_set(struct rte_eth_dev *, uint16_t, int);
+int mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on);
+void mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on);
+int mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask);
 
 /* mlx5_trigger.c */
 
-int mlx5_dev_start(struct rte_eth_dev *);
-void mlx5_dev_stop(struct rte_eth_dev *);
-int priv_dev_traffic_enable(struct priv *, struct rte_eth_dev *);
-int priv_dev_traffic_disable(struct priv *, struct rte_eth_dev *);
-int priv_dev_traffic_restart(struct priv *, struct rte_eth_dev *);
-int mlx5_traffic_restart(struct rte_eth_dev *);
+int mlx5_dev_start(struct rte_eth_dev *dev);
+void mlx5_dev_stop(struct rte_eth_dev *dev);
+int priv_dev_traffic_enable(struct priv *priv, struct rte_eth_dev *dev);
+int priv_dev_traffic_disable(struct priv *priv, struct rte_eth_dev *dev);
+int priv_dev_traffic_restart(struct priv *priv, struct rte_eth_dev *dev);
+int mlx5_traffic_restart(struct rte_eth_dev *dev);
 
 /* mlx5_flow.c */
 
-int mlx5_dev_filter_ctrl(struct rte_eth_dev *, enum rte_filter_type,
-			 enum rte_filter_op, void *);
-int mlx5_flow_validate(struct rte_eth_dev *, const struct rte_flow_attr *,
-		       const struct rte_flow_item [],
-		       const struct rte_flow_action [],
-		       struct rte_flow_error *);
-struct rte_flow *mlx5_flow_create(struct rte_eth_dev *,
-				  const struct rte_flow_attr *,
-				  const struct rte_flow_item [],
-				  const struct rte_flow_action [],
-				  struct rte_flow_error *);
-int mlx5_flow_destroy(struct rte_eth_dev *, struct rte_flow *,
-		      struct rte_flow_error *);
-void priv_flow_flush(struct priv *, struct mlx5_flows *);
-int mlx5_flow_flush(struct rte_eth_dev *, struct rte_flow_error *);
-int mlx5_flow_query(struct rte_eth_dev *, struct rte_flow *,
-		    enum rte_flow_action_type, void *,
-		    struct rte_flow_error *);
-int mlx5_flow_isolate(struct rte_eth_dev *, int, struct rte_flow_error *);
-int priv_flow_start(struct priv *, struct mlx5_flows *);
-void priv_flow_stop(struct priv *, struct mlx5_flows *);
-int priv_flow_verify(struct priv *);
-int mlx5_ctrl_flow_vlan(struct rte_eth_dev *, struct rte_flow_item_eth *,
-			struct rte_flow_item_eth *, struct rte_flow_item_vlan *,
-			struct rte_flow_item_vlan *);
-int mlx5_ctrl_flow(struct rte_eth_dev *, struct rte_flow_item_eth *,
-		   struct rte_flow_item_eth *);
-int priv_flow_create_drop_queue(struct priv *);
-void priv_flow_delete_drop_queue(struct priv *);
+int mlx5_flow_validate(struct rte_eth_dev *dev,
+		       const struct rte_flow_attr *attr,
+		       const struct rte_flow_item items[],
+		       const struct rte_flow_action actions[],
+		       struct rte_flow_error *error);
+void priv_flow_flush(struct priv *priv, struct mlx5_flows *list);
+int priv_flow_create_drop_queue(struct priv *priv);
+void priv_flow_stop(struct priv *priv, struct mlx5_flows *list);
+int priv_flow_start(struct priv *priv, struct mlx5_flows *list);
+int priv_flow_verify(struct priv *priv);
+int priv_flow_create_drop_queue(struct priv *priv);
+void priv_flow_delete_drop_queue(struct priv *priv);
+int mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
+			struct rte_flow_item_eth *eth_spec,
+			struct rte_flow_item_eth *eth_mask,
+			struct rte_flow_item_vlan *vlan_spec,
+			struct rte_flow_item_vlan *vlan_mask);
+int mlx5_ctrl_flow(struct rte_eth_dev *dev,
+		   struct rte_flow_item_eth *eth_spec,
+		   struct rte_flow_item_eth *eth_mask);
+struct rte_flow *mlx5_flow_create(struct rte_eth_dev *dev,
+				  const struct rte_flow_attr *attr,
+				  const struct rte_flow_item items[],
+				  const struct rte_flow_action actions[],
+				  struct rte_flow_error *error);
+int mlx5_flow_destroy(struct rte_eth_dev *dev, struct rte_flow *flow,
+		      struct rte_flow_error *error);
+int mlx5_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error);
+int mlx5_flow_query(struct rte_eth_dev *dev, struct rte_flow *flow,
+		    enum rte_flow_action_type action, void *data,
+		    struct rte_flow_error *error);
+int mlx5_flow_isolate(struct rte_eth_dev *dev, int enable,
+		      struct rte_flow_error *error);
+int mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
+			 enum rte_filter_type filter_type,
+			 enum rte_filter_op filter_op,
+			 void *arg);
 
 /* mlx5_socket.c */
 
@@ -336,9 +353,9 @@ int priv_socket_connect(struct priv *priv);
 
 /* mlx5_mr.c */
 
-struct mlx5_mr *priv_mr_new(struct priv *, struct rte_mempool *);
-struct mlx5_mr *priv_mr_get(struct priv *, struct rte_mempool *);
-int priv_mr_release(struct priv *, struct mlx5_mr *);
-int priv_mr_verify(struct priv *);
+struct mlx5_mr *priv_mr_new(struct priv *priv, struct rte_mempool *mp);
+struct mlx5_mr *priv_mr_get(struct priv *priv, struct rte_mempool *mp);
+int priv_mr_release(struct priv *priv, struct mlx5_mr *mr);
+int priv_mr_verify(struct priv *priv);
 
 #endif /* RTE_PMD_MLX5_H_ */
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index 28761c299..9d5edeedd 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -237,93 +237,111 @@ struct mlx5_txq_ctrl {
 extern uint8_t rss_hash_default_key[];
 extern const size_t rss_hash_default_key_len;
 
-void mlx5_rxq_cleanup(struct mlx5_rxq_ctrl *);
-int mlx5_rx_queue_setup(struct rte_eth_dev *, uint16_t, uint16_t, unsigned int,
-			const struct rte_eth_rxconf *, struct rte_mempool *);
-void mlx5_rx_queue_release(void *);
+void mlx5_rxq_cleanup(struct mlx5_rxq_ctrl *rxq_ctrl);
+int mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
+			unsigned int socket, const struct rte_eth_rxconf *conf,
+			struct rte_mempool *mp);
+void mlx5_rx_queue_release(void *dpdk_rxq);
 int priv_rx_intr_vec_enable(struct priv *priv);
 void priv_rx_intr_vec_disable(struct priv *priv);
 int mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id);
 int mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id);
-struct mlx5_rxq_ibv *mlx5_priv_rxq_ibv_new(struct priv *, uint16_t);
-struct mlx5_rxq_ibv *mlx5_priv_rxq_ibv_get(struct priv *, uint16_t);
-int mlx5_priv_rxq_ibv_release(struct priv *, struct mlx5_rxq_ibv *);
-int mlx5_priv_rxq_ibv_releasable(struct priv *, struct mlx5_rxq_ibv *);
-int mlx5_priv_rxq_ibv_verify(struct priv *);
-struct mlx5_rxq_ctrl *mlx5_priv_rxq_new(struct priv *, uint16_t,
-					uint16_t, unsigned int,
-					struct rte_mempool *);
-struct mlx5_rxq_ctrl *mlx5_priv_rxq_get(struct priv *, uint16_t);
-int mlx5_priv_rxq_release(struct priv *, uint16_t);
-int mlx5_priv_rxq_releasable(struct priv *, uint16_t);
-int mlx5_priv_rxq_verify(struct priv *);
-int rxq_alloc_elts(struct mlx5_rxq_ctrl *);
-struct mlx5_ind_table_ibv *mlx5_priv_ind_table_ibv_new(struct priv *,
-						       uint16_t [],
-						       uint16_t);
-struct mlx5_ind_table_ibv *mlx5_priv_ind_table_ibv_get(struct priv *,
-						       uint16_t [],
-						       uint16_t);
-int mlx5_priv_ind_table_ibv_release(struct priv *, struct mlx5_ind_table_ibv *);
-int mlx5_priv_ind_table_ibv_verify(struct priv *);
-struct mlx5_hrxq *mlx5_priv_hrxq_new(struct priv *, uint8_t *, uint8_t,
-				     uint64_t, uint16_t [], uint16_t);
-struct mlx5_hrxq *mlx5_priv_hrxq_get(struct priv *, uint8_t *, uint8_t,
-				     uint64_t, uint16_t [], uint16_t);
-int mlx5_priv_hrxq_release(struct priv *, struct mlx5_hrxq *);
-int mlx5_priv_hrxq_ibv_verify(struct priv *);
+struct mlx5_rxq_ibv *mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx);
+struct mlx5_rxq_ibv *mlx5_priv_rxq_ibv_get(struct priv *priv, uint16_t idx);
+int mlx5_priv_rxq_ibv_release(struct priv *priv, struct mlx5_rxq_ibv *rxq_ibv);
+int mlx5_priv_rxq_ibv_releasable(struct priv *priv,
+				 struct mlx5_rxq_ibv *rxq_ibv);
+int mlx5_priv_rxq_ibv_verify(struct priv *priv);
+struct mlx5_rxq_ctrl *mlx5_priv_rxq_new(struct priv *priv, uint16_t idx,
+					uint16_t desc,
+					unsigned int socket,
+					struct rte_mempool *mp);
+struct mlx5_rxq_ctrl *mlx5_priv_rxq_get(struct priv *priv, uint16_t idx);
+int mlx5_priv_rxq_release(struct priv *priv, uint16_t idx);
+int mlx5_priv_rxq_releasable(struct priv *priv, uint16_t idx);
+int mlx5_priv_rxq_verify(struct priv *priv);
+int rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl);
+struct mlx5_ind_table_ibv *mlx5_priv_ind_table_ibv_new(struct priv *priv,
+						       uint16_t queues[],
+						       uint16_t queues_n);
+struct mlx5_ind_table_ibv *mlx5_priv_ind_table_ibv_get(struct priv *priv,
+						       uint16_t queues[],
+						       uint16_t queues_n);
+int mlx5_priv_ind_table_ibv_release(struct priv *priv,
+				    struct mlx5_ind_table_ibv *ind_tbl);
+int mlx5_priv_ind_table_ibv_verify(struct priv *priv);
+struct mlx5_hrxq *mlx5_priv_hrxq_new(struct priv *priv, uint8_t *rss_key,
+				     uint8_t rss_key_len, uint64_t hash_fields,
+				     uint16_t queues[], uint16_t queues_n);
+struct mlx5_hrxq *mlx5_priv_hrxq_get(struct priv *priv, uint8_t *rss_key,
+				     uint8_t rss_key_len, uint64_t hash_fields,
+				     uint16_t queues[], uint16_t queues_n);
+int mlx5_priv_hrxq_release(struct priv *priv, struct mlx5_hrxq *hrxq);
+int mlx5_priv_hrxq_ibv_verify(struct priv *priv);
 
 /* mlx5_txq.c */
 
-int mlx5_tx_queue_setup(struct rte_eth_dev *, uint16_t, uint16_t, unsigned int,
-			const struct rte_eth_txconf *);
-void mlx5_tx_queue_release(void *);
+int mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
+			unsigned int socket, const struct rte_eth_txconf *conf);
+void mlx5_tx_queue_release(void *dpdk_txq);
 int priv_tx_uar_remap(struct priv *priv, int fd);
-struct mlx5_txq_ibv *mlx5_priv_txq_ibv_new(struct priv *, uint16_t);
-struct mlx5_txq_ibv *mlx5_priv_txq_ibv_get(struct priv *, uint16_t);
-int mlx5_priv_txq_ibv_release(struct priv *, struct mlx5_txq_ibv *);
-int mlx5_priv_txq_ibv_releasable(struct priv *, struct mlx5_txq_ibv *);
-int mlx5_priv_txq_ibv_verify(struct priv *);
-struct mlx5_txq_ctrl *mlx5_priv_txq_new(struct priv *, uint16_t,
-					uint16_t, unsigned int,
-					const struct rte_eth_txconf *);
-struct mlx5_txq_ctrl *mlx5_priv_txq_get(struct priv *, uint16_t);
-int mlx5_priv_txq_release(struct priv *, uint16_t);
-int mlx5_priv_txq_releasable(struct priv *, uint16_t);
-int mlx5_priv_txq_verify(struct priv *);
-void txq_alloc_elts(struct mlx5_txq_ctrl *);
+struct mlx5_txq_ibv *mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx);
+struct mlx5_txq_ibv *mlx5_priv_txq_ibv_get(struct priv *priv, uint16_t idx);
+int mlx5_priv_txq_ibv_release(struct priv *priv, struct mlx5_txq_ibv *txq_ibv);
+int mlx5_priv_txq_ibv_releasable(struct priv *priv,
+				 struct mlx5_txq_ibv *txq_ibv);
+int mlx5_priv_txq_ibv_verify(struct priv *priv);
+struct mlx5_txq_ctrl *mlx5_priv_txq_new(struct priv *priv, uint16_t idx,
+					uint16_t desc, unsigned int socket,
+					const struct rte_eth_txconf *conf);
+struct mlx5_txq_ctrl *mlx5_priv_txq_get(struct priv *priv, uint16_t idx);
+int mlx5_priv_txq_release(struct priv *priv, uint16_t idx);
+int mlx5_priv_txq_releasable(struct priv *priv, uint16_t idx);
+int mlx5_priv_txq_verify(struct priv *priv);
+void txq_alloc_elts(struct mlx5_txq_ctrl *txq_ctrl);
 
 /* mlx5_rxtx.c */
 
 extern uint32_t mlx5_ptype_table[];
 
 void mlx5_set_ptype_table(void);
-uint16_t mlx5_tx_burst(void *, struct rte_mbuf **, uint16_t);
-uint16_t mlx5_tx_burst_mpw(void *, struct rte_mbuf **, uint16_t);
-uint16_t mlx5_tx_burst_mpw_inline(void *, struct rte_mbuf **, uint16_t);
-uint16_t mlx5_tx_burst_empw(void *, struct rte_mbuf **, uint16_t);
-uint16_t mlx5_rx_burst(void *, struct rte_mbuf **, uint16_t);
-uint16_t removed_tx_burst(void *, struct rte_mbuf **, uint16_t);
-uint16_t removed_rx_burst(void *, struct rte_mbuf **, uint16_t);
-int mlx5_rx_descriptor_status(void *, uint16_t);
-int mlx5_tx_descriptor_status(void *, uint16_t);
+uint16_t mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts,
+		       uint16_t pkts_n);
+uint16_t mlx5_tx_burst_mpw(void *dpdk_txq, struct rte_mbuf **pkts,
+			   uint16_t pkts_n);
+uint16_t mlx5_tx_burst_mpw_inline(void *dpdk_txq, struct rte_mbuf **pkts,
+				  uint16_t pkts_n);
+uint16_t mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts,
+			    uint16_t pkts_n);
+uint16_t mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n);
+uint16_t removed_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts,
+			  uint16_t pkts_n);
+uint16_t removed_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts,
+			  uint16_t pkts_n);
+int mlx5_rx_descriptor_status(void *rx_queue, uint16_t offset);
+int mlx5_tx_descriptor_status(void *tx_queue, uint16_t offset);
 
 /* Vectorized version of mlx5_rxtx.c */
-int priv_check_raw_vec_tx_support(struct priv *);
-int priv_check_vec_tx_support(struct priv *);
-int rxq_check_vec_support(struct mlx5_rxq_data *);
-int priv_check_vec_rx_support(struct priv *);
-uint16_t mlx5_tx_burst_raw_vec(void *, struct rte_mbuf **, uint16_t);
-uint16_t mlx5_tx_burst_vec(void *, struct rte_mbuf **, uint16_t);
-uint16_t mlx5_rx_burst_vec(void *, struct rte_mbuf **, uint16_t);
+
+int priv_check_raw_vec_tx_support(struct priv *priv);
+int priv_check_vec_tx_support(struct priv *priv);
+int rxq_check_vec_support(struct mlx5_rxq_data *rxq);
+int priv_check_vec_rx_support(struct priv *priv);
+uint16_t mlx5_tx_burst_raw_vec(void *dpdk_txq, struct rte_mbuf **pkts,
+			       uint16_t pkts_n);
+uint16_t mlx5_tx_burst_vec(void *dpdk_txq, struct rte_mbuf **pkts,
+			   uint16_t pkts_n);
+uint16_t mlx5_rx_burst_vec(void *dpdk_rxq, struct rte_mbuf **pkts,
+			   uint16_t pkts_n);
 
 /* mlx5_mr.c */
 
-void mlx5_mp2mr_iter(struct rte_mempool *, void *);
-struct mlx5_mr *priv_txq_mp2mr_reg(struct priv *priv, struct mlx5_txq_data *,
-				   struct rte_mempool *, unsigned int);
-struct mlx5_mr *mlx5_txq_mp2mr_reg(struct mlx5_txq_data *, struct rte_mempool *,
-				   unsigned int);
+void mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg);
+struct mlx5_mr *priv_txq_mp2mr_reg(struct priv *priv, struct mlx5_txq_data *txq,
+				   struct rte_mempool *mp, unsigned int idx);
+struct mlx5_mr *mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq,
+				   struct rte_mempool *mp,
+				   unsigned int idx);
 
 #ifndef NDEBUG
 /**
-- 
2.11.0

[dpdk-stable] [PATCH v2 20/67] net/mlx5: mark parameters with unused attribute

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 56f08e1671f99a15cf6b00027c2f7d81d69c4f5f ]

Replaces all (void)foo; by __rte_unused macro except when variables are
under #if statements.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.c         |  4 ++--
 drivers/net/mlx5/mlx5_ethdev.c  | 18 ++++++----------
 drivers/net/mlx5/mlx5_flow.c    | 25 ++++++++--------------
 drivers/net/mlx5/mlx5_mac.c     |  3 +--
 drivers/net/mlx5/mlx5_mr.c      | 10 +++------
 drivers/net/mlx5/mlx5_rxq.c     |  8 +++----
 drivers/net/mlx5/mlx5_rxtx.c    | 47 +++++++++++++++++------------------------
 drivers/net/mlx5/mlx5_stats.c   |  2 +-
 drivers/net/mlx5/mlx5_trigger.c |  4 ++--
 drivers/net/mlx5/mlx5_txq.c     | 19 +++++++----------
 10 files changed, 55 insertions(+), 85 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index c3d57c58c..efcf86ba4 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -646,7 +646,8 @@ mlx5_args_assign(struct priv *priv, struct mlx5_args *args)
  *   0 on success, negative errno value on failure.
  */
 static int
-mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
+mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
+	       struct rte_pci_device *pci_dev)
 {
 	struct ibv_device **list;
 	struct ibv_device *ibv_dev;
@@ -663,7 +664,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 	struct ibv_counter_set_description cs_desc;
 #endif
 
-	(void)pci_drv;
 	assert(pci_drv == &mlx5_driver);
 	/* Get mlx5_dev[] index. */
 	idx = mlx5_dev_idx(&pci_dev->addr);
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index f9f04ec6d..0266a4fe5 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -492,11 +492,9 @@ mlx5_dev_supported_ptypes_get(struct rte_eth_dev *dev)
  *
  * @param dev
  *   Pointer to Ethernet device structure.
- * @param wait_to_complete
- *   Wait for request completion (ignored).
  */
 static int
-mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev, int wait_to_complete)
+mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 	struct ethtool_cmd edata = {
@@ -508,7 +506,6 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev, int wait_to_complete)
 
 	/* priv_lock() is not taken to allow concurrent calls. */
 
-	(void)wait_to_complete;
 	if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) {
 		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
 		return -1;
@@ -558,11 +555,9 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev, int wait_to_complete)
  *
  * @param dev
  *   Pointer to Ethernet device structure.
- * @param wait_to_complete
- *   Wait for request completion (ignored).
  */
 static int
-mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev, int wait_to_complete)
+mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 	struct ethtool_link_settings gcmd = { .cmd = ETHTOOL_GLINKSETTINGS };
@@ -570,7 +565,6 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev, int wait_to_complete)
 	struct rte_eth_link dev_link;
 	uint64_t sc;
 
-	(void)wait_to_complete;
 	if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) {
 		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
 		return -1;
@@ -700,7 +694,7 @@ priv_link_stop(struct priv *priv)
  *   Wait for request completion (ignored).
  */
 int
-priv_link_update(struct priv *priv, int wait_to_complete)
+priv_link_update(struct priv *priv, int wait_to_complete __rte_unused)
 {
 	struct rte_eth_dev *dev = priv->dev;
 	struct utsname utsname;
@@ -712,9 +706,9 @@ priv_link_update(struct priv *priv, int wait_to_complete)
 	    sscanf(utsname.release, "%d.%d.%d",
 		   &ver[0], &ver[1], &ver[2]) != 3 ||
 	    KERNEL_VERSION(ver[0], ver[1], ver[2]) < KERNEL_VERSION(4, 9, 0))
-		ret = mlx5_link_update_unlocked_gset(dev, wait_to_complete);
+		ret = mlx5_link_update_unlocked_gset(dev);
 	else
-		ret = mlx5_link_update_unlocked_gs(dev, wait_to_complete);
+		ret = mlx5_link_update_unlocked_gs(dev);
 	/* If lsc interrupt is disabled, should always be ready for traffic. */
 	if (!dev->data->dev_conf.intr_conf.lsc) {
 		priv_link_start(priv);
@@ -766,7 +760,7 @@ priv_force_link_status_change(struct priv *priv, int status)
  *   Wait for request completion (ignored).
  */
 int
-mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete)
+mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
 {
 	struct priv *priv = dev->data->dev_private;
 	int ret;
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 76c4354bb..0103d9342 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -565,7 +565,7 @@ mlx5_flow_item_validate(const struct rte_flow_item *item,
  *   0 on success, errno value on failure.
  */
 static int
-priv_flow_convert_rss_conf(struct priv *priv,
+priv_flow_convert_rss_conf(struct priv *priv __rte_unused,
 			   struct mlx5_flow_parse *parser,
 			   const struct rte_eth_rss_conf *rss_conf)
 {
@@ -574,7 +574,6 @@ priv_flow_convert_rss_conf(struct priv *priv,
 	 * priv_flow_convert_actions() to initialize the parser with the
 	 * device default RSS configuration.
 	 */
-	(void)priv;
 	if (rss_conf) {
 		if (rss_conf->rss_hf & MLX5_RSS_HF_MASK)
 			return EINVAL;
@@ -607,13 +606,11 @@ priv_flow_convert_rss_conf(struct priv *priv,
  *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_convert_attributes(struct priv *priv,
+priv_flow_convert_attributes(struct priv *priv __rte_unused,
 			     const struct rte_flow_attr *attr,
 			     struct rte_flow_error *error,
-			     struct mlx5_flow_parse *parser)
+			     struct mlx5_flow_parse *parser __rte_unused)
 {
-	(void)priv;
-	(void)parser;
 	if (attr->group) {
 		rte_flow_error_set(error, ENOTSUP,
 				   RTE_FLOW_ERROR_TYPE_ATTR_GROUP,
@@ -818,7 +815,7 @@ priv_flow_convert_actions(struct priv *priv,
  *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_convert_items_validate(struct priv *priv,
+priv_flow_convert_items_validate(struct priv *priv __rte_unused,
 				 const struct rte_flow_item items[],
 				 struct rte_flow_error *error,
 				 struct mlx5_flow_parse *parser)
@@ -826,7 +823,6 @@ priv_flow_convert_items_validate(struct priv *priv,
 	const struct mlx5_flow_items *cur_item = mlx5_flow_items;
 	unsigned int i;
 
-	(void)priv;
 	/* Initialise the offsets to start after verbs attribute. */
 	for (i = 0; i != hash_rxq_init_n; ++i)
 		parser->queue[i].offset = sizeof(struct ibv_flow_attr);
@@ -910,14 +906,13 @@ priv_flow_convert_items_validate(struct priv *priv,
  *   A verbs flow attribute on success, NULL otherwise.
  */
 static struct ibv_flow_attr*
-priv_flow_convert_allocate(struct priv *priv,
+priv_flow_convert_allocate(struct priv *priv __rte_unused,
 			   unsigned int priority,
 			   unsigned int size,
 			   struct rte_flow_error *error)
 {
 	struct ibv_flow_attr *ibv_attr;
 
-	(void)priv;
 	ibv_attr = rte_calloc(__func__, 1, size, 0);
 	if (!ibv_attr) {
 		rte_flow_error_set(error, ENOMEM,
@@ -939,7 +934,8 @@ priv_flow_convert_allocate(struct priv *priv,
  *   Internal parser structure.
  */
 static void
-priv_flow_convert_finalise(struct priv *priv, struct mlx5_flow_parse *parser)
+priv_flow_convert_finalise(struct priv *priv __rte_unused,
+			   struct mlx5_flow_parse *parser)
 {
 	const unsigned int ipv4 =
 		hash_rxq_init[parser->layer].ip_version == MLX5_IPV4;
@@ -950,7 +946,6 @@ priv_flow_convert_finalise(struct priv *priv, struct mlx5_flow_parse *parser)
 	const enum hash_rxq_type ip = ipv4 ? HASH_RXQ_IPV4 : HASH_RXQ_IPV6;
 	unsigned int i;
 
-	(void)priv;
 	/* Remove any other flow not matching the pattern. */
 	if (parser->queues_n == 1) {
 		for (i = 0; i != hash_rxq_init_n; ++i) {
@@ -2445,11 +2440,10 @@ mlx5_ctrl_flow(struct rte_eth_dev *dev,
 int
 mlx5_flow_destroy(struct rte_eth_dev *dev,
 		  struct rte_flow *flow,
-		  struct rte_flow_error *error)
+		  struct rte_flow_error *error __rte_unused)
 {
 	struct priv *priv = dev->data->dev_private;
 
-	(void)error;
 	priv_lock(priv);
 	priv_flow_destroy(priv, &priv->flows, flow);
 	priv_unlock(priv);
@@ -2464,11 +2458,10 @@ mlx5_flow_destroy(struct rte_eth_dev *dev,
  */
 int
 mlx5_flow_flush(struct rte_eth_dev *dev,
-		struct rte_flow_error *error)
+		struct rte_flow_error *error __rte_unused)
 {
 	struct priv *priv = dev->data->dev_private;
 
-	(void)error;
 	priv_lock(priv);
 	priv_flow_flush(priv, &priv->flows);
 	priv_unlock(priv);
diff --git a/drivers/net/mlx5/mlx5_mac.c b/drivers/net/mlx5/mlx5_mac.c
index 3ad21c000..8e7b098ae 100644
--- a/drivers/net/mlx5/mlx5_mac.c
+++ b/drivers/net/mlx5/mlx5_mac.c
@@ -116,12 +116,11 @@ mlx5_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index)
  */
 int
 mlx5_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac,
-		  uint32_t index, uint32_t vmdq)
+		  uint32_t index, uint32_t vmdq __rte_unused)
 {
 	unsigned int i;
 	int ret = 0;
 
-	(void)vmdq;
 	assert(index < MLX5_MAX_MAC_ADDRESSES);
 	/* First, make sure this address isn't already configured. */
 	for (i = 0; (i != MLX5_MAX_MAC_ADDRESSES); ++i) {
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 33d037d28..0bf65ec0e 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -55,15 +55,12 @@ struct mlx5_check_mempool_data {
 
 /* Called by mlx5_check_mempool() when iterating the memory chunks. */
 static void
-mlx5_check_mempool_cb(struct rte_mempool *mp,
+mlx5_check_mempool_cb(struct rte_mempool *mp __rte_unused,
 		      void *opaque, struct rte_mempool_memhdr *memhdr,
-		      unsigned int mem_idx)
+		      unsigned int mem_idx __rte_unused)
 {
 	struct mlx5_check_mempool_data *data = opaque;
 
-	(void)mp;
-	(void)mem_idx;
-
 	/* It already failed, skip the next chunks. */
 	if (data->ret != 0)
 		return;
@@ -365,9 +362,8 @@ priv_mr_get(struct priv *priv, struct rte_mempool *mp)
  *   0 on success, errno on failure.
  */
 int
-priv_mr_release(struct priv *priv, struct mlx5_mr *mr)
+priv_mr_release(struct priv *priv __rte_unused, struct mlx5_mr *mr)
 {
-	(void)priv;
 	assert(mr);
 	DEBUG("Memory Region %p refcnt: %d",
 	      (void *)mr, rte_atomic32_read(&mr->refcnt));
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 457f919b4..342dad18b 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -232,7 +232,8 @@ mlx5_rxq_cleanup(struct mlx5_rxq_ctrl *rxq_ctrl)
  */
 int
 mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
-		    unsigned int socket, const struct rte_eth_rxconf *conf,
+		    unsigned int socket,
+		    const struct rte_eth_rxconf *conf __rte_unused,
 		    struct rte_mempool *mp)
 {
 	struct priv *priv = dev->data->dev_private;
@@ -241,7 +242,6 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		container_of(rxq, struct mlx5_rxq_ctrl, rxq);
 	int ret = 0;
 
-	(void)conf;
 	priv_lock(priv);
 	if (!rte_is_power_of_2(desc)) {
 		desc = 1 << log2above(desc);
@@ -854,9 +854,9 @@ mlx5_priv_rxq_ibv_verify(struct priv *priv)
  *   Verbs Rx queue object.
  */
 int
-mlx5_priv_rxq_ibv_releasable(struct priv *priv, struct mlx5_rxq_ibv *rxq_ibv)
+mlx5_priv_rxq_ibv_releasable(struct priv *priv __rte_unused,
+			     struct mlx5_rxq_ibv *rxq_ibv)
 {
-	(void)priv;
 	assert(rxq_ibv);
 	return (rte_atomic32_read(&rxq_ibv->refcnt) == 1);
 }
diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 9955faedf..40a271a68 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -1975,11 +1975,10 @@ mlx5_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
  *   Number of packets successfully transmitted (<= pkts_n).
  */
 uint16_t
-removed_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
+removed_tx_burst(void *dpdk_txq __rte_unused,
+		 struct rte_mbuf **pkts __rte_unused,
+		 uint16_t pkts_n __rte_unused)
 {
-	(void)dpdk_txq;
-	(void)pkts;
-	(void)pkts_n;
 	return 0;
 }
 
@@ -2000,11 +1999,10 @@ removed_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
  *   Number of packets successfully received (<= pkts_n).
  */
 uint16_t
-removed_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
+removed_rx_burst(void *dpdk_txq __rte_unused,
+		 struct rte_mbuf **pkts __rte_unused,
+		 uint16_t pkts_n __rte_unused)
 {
-	(void)dpdk_rxq;
-	(void)pkts;
-	(void)pkts_n;
 	return 0;
 }
 
@@ -2016,56 +2014,49 @@ removed_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
  */
 
 uint16_t __attribute__((weak))
-mlx5_tx_burst_raw_vec(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
+mlx5_tx_burst_raw_vec(void *dpdk_txq __rte_unused,
+		      struct rte_mbuf **pkts __rte_unused,
+		      uint16_t pkts_n __rte_unused)
 {
-	(void)dpdk_txq;
-	(void)pkts;
-	(void)pkts_n;
 	return 0;
 }
 
 uint16_t __attribute__((weak))
-mlx5_tx_burst_vec(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
+mlx5_tx_burst_vec(void *dpdk_txq __rte_unused,
+		  struct rte_mbuf **pkts __rte_unused,
+		  uint16_t pkts_n __rte_unused)
 {
-	(void)dpdk_txq;
-	(void)pkts;
-	(void)pkts_n;
 	return 0;
 }
 
 uint16_t __attribute__((weak))
-mlx5_rx_burst_vec(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
+mlx5_rx_burst_vec(void *dpdk_txq __rte_unused,
+		  struct rte_mbuf **pkts __rte_unused,
+		  uint16_t pkts_n __rte_unused)
 {
-	(void)dpdk_rxq;
-	(void)pkts;
-	(void)pkts_n;
 	return 0;
 }
 
 int __attribute__((weak))
-priv_check_raw_vec_tx_support(struct priv *priv)
+priv_check_raw_vec_tx_support(struct priv *priv __rte_unused)
 {
-	(void)priv;
 	return -ENOTSUP;
 }
 
 int __attribute__((weak))
-priv_check_vec_tx_support(struct priv *priv)
+priv_check_vec_tx_support(struct priv *priv __rte_unused)
 {
-	(void)priv;
 	return -ENOTSUP;
 }
 
 int __attribute__((weak))
-rxq_check_vec_support(struct mlx5_rxq_data *rxq)
+rxq_check_vec_support(struct mlx5_rxq_data *rxq __rte_unused)
 {
-	(void)rxq;
 	return -ENOTSUP;
 }
 
 int __attribute__((weak))
-priv_check_vec_rx_support(struct priv *priv)
+priv_check_vec_rx_support(struct priv *priv __rte_unused)
 {
-	(void)priv;
 	return -ENOTSUP;
 }
diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index 7943c4c3f..50e3ea5f2 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -500,7 +500,7 @@ mlx5_xstats_reset(struct rte_eth_dev *dev)
  *   Number of xstats names.
  */
 int
-mlx5_xstats_get_names(struct rte_eth_dev *dev,
+mlx5_xstats_get_names(struct rte_eth_dev *dev __rte_unused,
 		struct rte_eth_xstat_name *xstats_names, unsigned int n)
 {
 	struct priv *priv = dev->data->dev_private;
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index da9494eb8..2d9bb86ea 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -367,9 +367,9 @@ priv_dev_traffic_enable(struct priv *priv, struct rte_eth_dev *dev)
  *   0 on success.
  */
 int
-priv_dev_traffic_disable(struct priv *priv, struct rte_eth_dev *dev)
+priv_dev_traffic_disable(struct priv *priv,
+			 struct rte_eth_dev *dev __rte_unused)
 {
-	(void)dev;
 	priv_flow_flush(priv, &priv->ctrl_flows);
 	return 0;
 }
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index f39b98a9a..a1c3e9d83 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -511,9 +511,9 @@ mlx5_priv_txq_ibv_get(struct priv *priv, uint16_t idx)
  *   0 on success, errno on failure.
  */
 int
-mlx5_priv_txq_ibv_release(struct priv *priv, struct mlx5_txq_ibv *txq_ibv)
+mlx5_priv_txq_ibv_release(struct priv *priv __rte_unused,
+			  struct mlx5_txq_ibv *txq_ibv)
 {
-	(void)priv;
 	assert(txq_ibv);
 	DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)priv,
 	      (void *)txq_ibv, rte_atomic32_read(&txq_ibv->refcnt));
@@ -536,9 +536,9 @@ mlx5_priv_txq_ibv_release(struct priv *priv, struct mlx5_txq_ibv *txq_ibv)
  *   Verbs Tx queue object.
  */
 int
-mlx5_priv_txq_ibv_releasable(struct priv *priv, struct mlx5_txq_ibv *txq_ibv)
+mlx5_priv_txq_ibv_releasable(struct priv *priv __rte_unused,
+			     struct mlx5_txq_ibv *txq_ibv)
 {
-	(void)priv;
 	assert(txq_ibv);
 	return (rte_atomic32_read(&txq_ibv->refcnt) == 1);
 }
@@ -707,13 +707,10 @@ mlx5_priv_txq_get(struct priv *priv, uint16_t idx)
 
 		mlx5_priv_txq_ibv_get(priv, idx);
 		for (i = 0; i != MLX5_PMD_TX_MP_CACHE; ++i) {
-			struct mlx5_mr *mr = NULL;
-
-			(void)mr;
-			if (ctrl->txq.mp2mr[i]) {
-				mr = priv_mr_get(priv, ctrl->txq.mp2mr[i]->mp);
-				assert(mr);
-			}
+			if (ctrl->txq.mp2mr[i])
+				claim_nonzero
+					(priv_mr_get(priv,
+						     ctrl->txq.mp2mr[i]->mp));
 		}
 		rte_atomic32_inc(&ctrl->refcnt);
 		DEBUG("%p: Tx queue %p: refcnt %d", (void *)priv,
-- 
2.11.0

[dpdk-stable] [PATCH v2 21/67] net/mlx5: normalize function prototypes

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit c9e88d35daf9a19bf9b653dd96c216be4a214d32 ]

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c |  2 +-
 drivers/net/mlx5/mlx5_mr.c   | 11 ++++++-----
 drivers/net/mlx5/mlx5_rxq.c  | 16 ++++++++--------
 drivers/net/mlx5/mlx5_txq.c  |  8 ++++----
 4 files changed, 19 insertions(+), 18 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 0103d9342..9f48ccf3d 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -905,7 +905,7 @@ priv_flow_convert_items_validate(struct priv *priv __rte_unused,
  * @return
  *   A verbs flow attribute on success, NULL otherwise.
  */
-static struct ibv_flow_attr*
+static struct ibv_flow_attr *
 priv_flow_convert_allocate(struct priv *priv __rte_unused,
 			   unsigned int priority,
 			   unsigned int size,
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 0bf65ec0e..ee4848446 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -95,8 +95,9 @@ mlx5_check_mempool_cb(struct rte_mempool *mp __rte_unused,
  * @return
  *   0 on success (mempool is virtually contiguous), -1 on error.
  */
-static int mlx5_check_mempool(struct rte_mempool *mp, uintptr_t *start,
-	uintptr_t *end)
+static int
+mlx5_check_mempool(struct rte_mempool *mp, uintptr_t *start,
+		   uintptr_t *end)
 {
 	struct mlx5_check_mempool_data data;
 
@@ -126,7 +127,7 @@ static int mlx5_check_mempool(struct rte_mempool *mp, uintptr_t *start,
  * @return
  *   mr on success, NULL on failure.
  */
-struct mlx5_mr*
+struct mlx5_mr *
 priv_txq_mp2mr_reg(struct priv *priv, struct mlx5_txq_data *txq,
 		   struct rte_mempool *mp, unsigned int idx)
 {
@@ -273,7 +274,7 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
  * @return
  *   The memory region on success.
  */
-struct mlx5_mr*
+struct mlx5_mr *
 priv_mr_new(struct priv *priv, struct rte_mempool *mp)
 {
 	const struct rte_memseg *ms = rte_eal_get_physmem_layout();
@@ -333,7 +334,7 @@ priv_mr_new(struct priv *priv, struct rte_mempool *mp)
  * @return
  *   The memory region on success.
  */
-struct mlx5_mr*
+struct mlx5_mr *
 priv_mr_get(struct priv *priv, struct rte_mempool *mp)
 {
 	struct mlx5_mr *mr;
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 342dad18b..003959213 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -548,7 +548,7 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
  * @return
  *   The Verbs object initialised if it can be created.
  */
-struct mlx5_rxq_ibv*
+struct mlx5_rxq_ibv *
 mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 {
 	struct mlx5_rxq_data *rxq_data = (*priv->rxqs)[idx];
@@ -764,7 +764,7 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
  * @return
  *   The Verbs object if it exists.
  */
-struct mlx5_rxq_ibv*
+struct mlx5_rxq_ibv *
 mlx5_priv_rxq_ibv_get(struct priv *priv, uint16_t idx)
 {
 	struct mlx5_rxq_data *rxq_data = (*priv->rxqs)[idx];
@@ -876,7 +876,7 @@ mlx5_priv_rxq_ibv_releasable(struct priv *priv __rte_unused,
  * @return
  *   A DPDK queue object on success.
  */
-struct mlx5_rxq_ctrl*
+struct mlx5_rxq_ctrl *
 mlx5_priv_rxq_new(struct priv *priv, uint16_t idx, uint16_t desc,
 		  unsigned int socket, struct rte_mempool *mp)
 {
@@ -999,7 +999,7 @@ mlx5_priv_rxq_new(struct priv *priv, uint16_t idx, uint16_t desc,
  * @return
  *   A pointer to the queue if it exists.
  */
-struct mlx5_rxq_ctrl*
+struct mlx5_rxq_ctrl *
 mlx5_priv_rxq_get(struct priv *priv, uint16_t idx)
 {
 	struct mlx5_rxq_ctrl *rxq_ctrl = NULL;
@@ -1112,7 +1112,7 @@ mlx5_priv_rxq_verify(struct priv *priv)
  * @return
  *   A new indirection table.
  */
-struct mlx5_ind_table_ibv*
+struct mlx5_ind_table_ibv *
 mlx5_priv_ind_table_ibv_new(struct priv *priv, uint16_t queues[],
 			    uint16_t queues_n)
 {
@@ -1174,7 +1174,7 @@ mlx5_priv_ind_table_ibv_new(struct priv *priv, uint16_t queues[],
  * @return
  *   An indirection table if found.
  */
-struct mlx5_ind_table_ibv*
+struct mlx5_ind_table_ibv *
 mlx5_priv_ind_table_ibv_get(struct priv *priv, uint16_t queues[],
 			    uint16_t queues_n)
 {
@@ -1272,7 +1272,7 @@ mlx5_priv_ind_table_ibv_verify(struct priv *priv)
  * @return
  *   An hash Rx queue on success.
  */
-struct mlx5_hrxq*
+struct mlx5_hrxq *
 mlx5_priv_hrxq_new(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
 		   uint64_t hash_fields, uint16_t queues[], uint16_t queues_n)
 {
@@ -1341,7 +1341,7 @@ mlx5_priv_hrxq_new(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
  * @return
  *   An hash Rx queue on success.
  */
-struct mlx5_hrxq*
+struct mlx5_hrxq *
 mlx5_priv_hrxq_get(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
 		   uint64_t hash_fields, uint16_t queues[], uint16_t queues_n)
 {
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index a1c3e9d83..3741103d7 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -306,7 +306,7 @@ priv_tx_uar_remap(struct priv *priv, int fd)
  * @return
  *   The Verbs object initialised if it can be created.
  */
-struct mlx5_txq_ibv*
+struct mlx5_txq_ibv *
 mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
 {
 	struct mlx5_txq_data *txq_data = (*priv->txqs)[idx];
@@ -480,7 +480,7 @@ mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
  * @return
  *   The Verbs object if it exists.
  */
-struct mlx5_txq_ibv*
+struct mlx5_txq_ibv *
 mlx5_priv_txq_ibv_get(struct priv *priv, uint16_t idx)
 {
 	struct mlx5_txq_ctrl *txq_ctrl;
@@ -582,7 +582,7 @@ mlx5_priv_txq_ibv_verify(struct priv *priv)
  * @return
  *   A DPDK queue object on success.
  */
-struct mlx5_txq_ctrl*
+struct mlx5_txq_ctrl *
 mlx5_priv_txq_new(struct priv *priv, uint16_t idx, uint16_t desc,
 		  unsigned int socket,
 		  const struct rte_eth_txconf *conf)
@@ -695,7 +695,7 @@ mlx5_priv_txq_new(struct priv *priv, uint16_t idx, uint16_t desc,
  * @return
  *   A pointer to the queue if it exists.
  */
-struct mlx5_txq_ctrl*
+struct mlx5_txq_ctrl *
 mlx5_priv_txq_get(struct priv *priv, uint16_t idx)
 {
 	struct mlx5_txq_ctrl *ctrl = NULL;
-- 
2.11.0

[dpdk-stable] [PATCH v2 22/67] net/mlx5: add missing function documentation

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit fb732b0a49c22ee7aeae4a23aab834feb5dfd1d4 ]

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5_ethdev.c  | 18 ++++++++++++++++++
 drivers/net/mlx5/mlx5_mr.c      |  7 +++++--
 drivers/net/mlx5/mlx5_rxq.c     | 20 ++++++++++++--------
 drivers/net/mlx5/mlx5_trigger.c | 30 ++++++++++++++++++++++++++++++
 drivers/net/mlx5/mlx5_txq.c     | 10 ++++++----
 5 files changed, 71 insertions(+), 14 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 0266a4fe5..764549b63 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -460,6 +460,15 @@ mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info)
 	priv_unlock(priv);
 }
 
+/**
+ * Get supported packet types.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   A pointer to the supported Packet types array.
+ */
 const uint32_t *
 mlx5_dev_supported_ptypes_get(struct rte_eth_dev *dev)
 {
@@ -492,6 +501,9 @@ mlx5_dev_supported_ptypes_get(struct rte_eth_dev *dev)
  *
  * @param dev
  *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   0 on success, -1 on error.
  */
 static int
 mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
@@ -555,6 +567,9 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
  *
  * @param dev
  *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   0 on success, -1 on error.
  */
 static int
 mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
@@ -758,6 +773,9 @@ priv_force_link_status_change(struct priv *priv, int status)
  *   Pointer to Ethernet device structure.
  * @param wait_to_complete
  *   Wait for request completion (ignored).
+ *
+ * @return
+ *   0 on success, -1 on error.
  */
 int
 mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index ee4848446..75de5a44d 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -271,6 +271,7 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
  *   Pointer to private structure.
  * @param mp
  *   Pointer to the memory pool to register.
+ *
  * @return
  *   The memory region on success.
  */
@@ -331,6 +332,7 @@ priv_mr_new(struct priv *priv, struct rte_mempool *mp)
  *   Pointer to private structure.
  * @param mp
  *   Pointer to the memory pool to register.
+ *
  * @return
  *   The memory region on success.
  */
@@ -381,9 +383,10 @@ priv_mr_release(struct priv *priv __rte_unused, struct mlx5_mr *mr)
  * Verify the flow list is empty
  *
  * @param priv
- *  Pointer to private structure.
+ *   Pointer to private structure.
  *
- * @return the number of object not released.
+ * @return
+ *   The number of object not released.
  */
 int
 priv_mr_verify(struct priv *priv)
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 003959213..68b03d246 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -827,9 +827,10 @@ mlx5_priv_rxq_ibv_release(struct priv *priv, struct mlx5_rxq_ibv *rxq_ibv)
  * Verify the Verbs Rx queue list is empty
  *
  * @param priv
- *  Pointer to private structure.
+ *   Pointer to private structure.
  *
- * @return the number of object not released.
+ * @return
+ *   The number of object not released.
  */
 int
 mlx5_priv_rxq_ibv_verify(struct priv *priv)
@@ -1081,9 +1082,10 @@ mlx5_priv_rxq_releasable(struct priv *priv, uint16_t idx)
  * Verify the Rx Queue list is empty
  *
  * @param priv
- *  Pointer to private structure.
+ *   Pointer to private structure.
  *
- * @return the number of object not released.
+ * @return
+ *   The number of object not released.
  */
 int
 mlx5_priv_rxq_verify(struct priv *priv)
@@ -1234,9 +1236,10 @@ mlx5_priv_ind_table_ibv_release(struct priv *priv,
  * Verify the Rx Queue list is empty
  *
  * @param priv
- *  Pointer to private structure.
+ *   Pointer to private structure.
  *
- * @return the number of object not released.
+ * @return
+ *   The number of object not released.
  */
 int
 mlx5_priv_ind_table_ibv_verify(struct priv *priv)
@@ -1403,9 +1406,10 @@ mlx5_priv_hrxq_release(struct priv *priv, struct mlx5_hrxq *hrxq)
  * Verify the Rx Queue list is empty
  *
  * @param priv
- *  Pointer to private structure.
+ *   Pointer to private structure.
  *
- * @return the number of object not released.
+ * @return
+ *   The number of object not released.
  */
 int
 mlx5_priv_hrxq_ibv_verify(struct priv *priv)
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index 2d9bb86ea..ad15158fe 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -41,6 +41,12 @@
 #include "mlx5_rxtx.h"
 #include "mlx5_utils.h"
 
+/**
+ * Stop traffic on Tx queues.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ */
 static void
 priv_txq_stop(struct priv *priv)
 {
@@ -50,6 +56,15 @@ priv_txq_stop(struct priv *priv)
 		mlx5_priv_txq_release(priv, i);
 }
 
+/**
+ * Start traffic on Tx queues.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   0 on success, errno on error.
+ */
 static int
 priv_txq_start(struct priv *priv)
 {
@@ -85,6 +100,12 @@ priv_txq_start(struct priv *priv)
 	return ret;
 }
 
+/**
+ * Stop traffic on Rx queues.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ */
 static void
 priv_rxq_stop(struct priv *priv)
 {
@@ -94,6 +115,15 @@ priv_rxq_stop(struct priv *priv)
 		mlx5_priv_rxq_release(priv, i);
 }
 
+/**
+ * Start traffic on Rx queues.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   0 on success, errno on error.
+ */
 static int
 priv_rxq_start(struct priv *priv)
 {
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 3741103d7..09d62a95b 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -547,9 +547,10 @@ mlx5_priv_txq_ibv_releasable(struct priv *priv __rte_unused,
  * Verify the Verbs Tx queue list is empty
  *
  * @param priv
- *  Pointer to private structure.
+ *   Pointer to private structure.
  *
- * @return the number of object not released.
+ * @return
+ *   The number of object not released.
  */
 int
 mlx5_priv_txq_ibv_verify(struct priv *priv)
@@ -794,9 +795,10 @@ mlx5_priv_txq_releasable(struct priv *priv, uint16_t idx)
  * Verify the Tx Queue list is empty
  *
  * @param priv
- *  Pointer to private structure.
+ *   Pointer to private structure.
  *
- * @return the number of object not released.
+ * @return
+ *   The number of object not released.
  */
 int
 mlx5_priv_txq_verify(struct priv *priv)
-- 
2.11.0

[dpdk-stable] [PATCH v2 23/67] net/mlx5: remove useless empty lines

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 0b3456e391a6518e26c60ab2aa6546964d46a6f6 ]

Some empty lines have been added in the middle of the code without any
reason.  This commit removes them.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.c        | 22 ----------------------
 drivers/net/mlx5/mlx5_ethdev.c |  7 -------
 drivers/net/mlx5/mlx5_mr.c     |  1 -
 drivers/net/mlx5/mlx5_rss.c    |  2 --
 drivers/net/mlx5/mlx5_rxq.c    |  1 -
 drivers/net/mlx5/mlx5_vlan.c   |  6 ------
 6 files changed, 39 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index efcf86ba4..9ee9d2c36 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -672,7 +672,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		return -ENOMEM;
 	}
 	DEBUG("using driver device index %d", idx);
-
 	/* Save PCI address. */
 	mlx5_dev[idx].pci_addr = pci_dev->addr;
 	list = ibv_get_device_list(&i);
@@ -733,7 +732,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		return -err;
 	}
 	ibv_dev = list[i];
-
 	DEBUG("device opened");
 	/*
 	 * Multi-packet send is supported by ConnectX-4 Lx PF as well
@@ -760,7 +758,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	if (ibv_query_device_ex(attr_ctx, NULL, &device_attr))
 		goto error;
 	INFO("%u port(s) detected", device_attr.orig_attr.phys_port_cnt);
-
 	for (i = 0; i < device_attr.orig_attr.phys_port_cnt; i++) {
 		char name[RTE_ETH_NAME_MAX_LEN];
 		uint32_t port = i + 1; /* ports are indexed from one */
@@ -788,9 +785,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		snprintf(name, sizeof(name), PCI_PRI_FMT,
 			 pci_dev->addr.domain, pci_dev->addr.bus,
 			 pci_dev->addr.devid, pci_dev->addr.function);
-
 		mlx5_dev[idx].ports |= test;
-
 		if (rte_eal_process_type() == RTE_PROC_SECONDARY) {
 			eth_dev = rte_eth_dev_attach_secondary(name);
 			if (eth_dev == NULL) {
@@ -820,15 +815,12 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			priv_dev_select_tx_function(priv, eth_dev);
 			continue;
 		}
-
 		DEBUG("using port %u (%08" PRIx32 ")", port, test);
-
 		ctx = ibv_open_device(ibv_dev);
 		if (ctx == NULL) {
 			err = ENODEV;
 			goto port_error;
 		}
-
 		ibv_query_device_ex(ctx, NULL, &device_attr);
 		/* Check port status. */
 		err = ibv_query_port(ctx, port, &port_attr);
@@ -836,19 +828,16 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			ERROR("port query failed: %s", strerror(err));
 			goto port_error;
 		}
-
 		if (port_attr.link_layer != IBV_LINK_LAYER_ETHERNET) {
 			ERROR("port %d is not configured in Ethernet mode",
 			      port);
 			err = EINVAL;
 			goto port_error;
 		}
-
 		if (port_attr.state != IBV_PORT_ACTIVE)
 			DEBUG("port %d is not active: \"%s\" (%d)",
 			      port, ibv_port_state_str(port_attr.state),
 			      port_attr.state);
-
 		/* Allocate protection domain. */
 		pd = ibv_alloc_pd(ctx);
 		if (pd == NULL) {
@@ -856,9 +845,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			err = ENOMEM;
 			goto port_error;
 		}
-
 		mlx5_dev[idx].ports |= test;
-
 		/* from rte_ethdev.c */
 		priv = rte_zmalloc("ethdev private structure",
 				   sizeof(*priv),
@@ -868,7 +855,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			err = ENOMEM;
 			goto port_error;
 		}
-
 		priv->ctx = ctx;
 		strncpy(priv->ibdev_path, priv->ctx->device->ibdev_path,
 			sizeof(priv->ibdev_path));
@@ -893,7 +879,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			ERROR("ibv_query_device_ex() failed");
 			goto port_error;
 		}
-
 		priv->hw_csum =
 			!!(device_attr_ex.device_cap_flags_ex &
 			   IBV_DEVICE_RAW_IP_CSUM);
@@ -938,7 +923,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 #endif
 		DEBUG("hardware RX end alignment padding is %ssupported",
 		      (priv->hw_padding ? "" : "not "));
-
 		priv->tso = ((priv->tso) &&
 			    (device_attr_ex.tso_caps.max_tso > 0) &&
 			    (device_attr_ex.tso_caps.supported_qpts &
@@ -1003,7 +987,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		/* Get actual MTU if possible. */
 		priv_get_mtu(priv, &priv->mtu);
 		DEBUG("port %u MTU is %u", priv->port, priv->mtu);
-
 		eth_dev = rte_eth_dev_allocate(name);
 		if (eth_dev == NULL) {
 			ERROR("can not allocate rte ethdev");
@@ -1026,7 +1009,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		claim_zero(mlx5_mac_addr_add(eth_dev, &mac, 0, 0));
 		TAILQ_INIT(&priv->flows);
 		TAILQ_INIT(&priv->ctrl_flows);
-
 		/* Hint libmlx5 to use PMD allocator for data plane resources */
 		struct mlx5dv_ctx_allocators alctr = {
 			.alloc = &mlx5_alloc_verbs_buf,
@@ -1040,7 +1022,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		DEBUG("forcing Ethernet interface up");
 		priv_set_flags(priv, ~IFF_UP, IFF_UP);
 		continue;
-
 port_error:
 		if (priv)
 			rte_free(priv);
@@ -1050,20 +1031,17 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			claim_zero(ibv_close_device(ctx));
 		break;
 	}
-
 	/*
 	 * XXX if something went wrong in the loop above, there is a resource
 	 * leak (ctx, pd, priv, dpdk ethdev) but we can do nothing about it as
 	 * long as the dpdk does not provide a way to deallocate a ethdev and a
 	 * way to enumerate the registered ethdevs to free the previous ones.
 	 */
-
 	/* no port found, complain */
 	if (!mlx5_dev[idx].ports) {
 		err = ENODEV;
 		goto error;
 	}
-
 error:
 	if (attr_ctx)
 		claim_zero(ibv_close_device(attr_ctx));
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 764549b63..79d62bbc1 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -411,7 +411,6 @@ mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info)
 	char ifname[IF_NAMESIZE];
 
 	info->pci_dev = RTE_ETH_DEV_TO_PCI(dev);
-
 	priv_lock(priv);
 	/* FIXME: we should ask the device for these values. */
 	info->min_rx_bufsize = 32;
@@ -517,7 +516,6 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 	int link_speed = 0;
 
 	/* priv_lock() is not taken to allow concurrent calls. */
-
 	if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) {
 		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
 		return -1;
@@ -863,7 +861,6 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		     strerror(ret));
 		goto out;
 	}
-
 	fc_conf->autoneg = ethpause.autoneg;
 	if (ethpause.rx_pause && ethpause.tx_pause)
 		fc_conf->mode = RTE_FC_FULL;
@@ -874,7 +871,6 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 	else
 		fc_conf->mode = RTE_FC_NONE;
 	ret = 0;
-
 out:
 	priv_unlock(priv);
 	assert(ret >= 0);
@@ -915,7 +911,6 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		ethpause.tx_pause = 1;
 	else
 		ethpause.tx_pause = 0;
-
 	priv_lock(priv);
 	if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
 		ret = errno;
@@ -925,7 +920,6 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		goto out;
 	}
 	ret = 0;
-
 out:
 	priv_unlock(priv);
 	assert(ret >= 0);
@@ -1183,7 +1177,6 @@ priv_dev_interrupt_handler_install(struct priv *priv, struct rte_eth_dev *dev)
 		rte_intr_callback_register(&priv->intr_handle,
 					   mlx5_dev_interrupt_handler, dev);
 	}
-
 	rc = priv_socket_init(priv);
 	if (!rc && priv->primary_socket) {
 		priv->intr_handle_socket.fd = priv->primary_socket;
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 75de5a44d..7da54758c 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -105,7 +105,6 @@ mlx5_check_mempool(struct rte_mempool *mp, uintptr_t *start,
 	rte_mempool_mem_iter(mp, mlx5_check_mempool_cb, &data);
 	*start = (uintptr_t)data.start;
 	*end = (uintptr_t)data.end;
-
 	return data.ret;
 }
 
diff --git a/drivers/net/mlx5/mlx5_rss.c b/drivers/net/mlx5/mlx5_rss.c
index f47bda667..8f0882c89 100644
--- a/drivers/net/mlx5/mlx5_rss.c
+++ b/drivers/net/mlx5/mlx5_rss.c
@@ -151,7 +151,6 @@ priv_rss_reta_index_resize(struct priv *priv, unsigned int reta_size)
 		return ENOMEM;
 	priv->reta_idx = mem;
 	priv->reta_idx_n = reta_size;
-
 	if (old_size < reta_size)
 		memset(&(*priv->reta_idx)[old_size], 0,
 		       (reta_size - old_size) *
@@ -219,7 +218,6 @@ priv_dev_rss_reta_update(struct priv *priv,
 	ret = priv_rss_reta_index_resize(priv, reta_size);
 	if (ret)
 		return ret;
-
 	for (idx = 0, i = 0; (i != reta_size); ++i) {
 		idx = i / RTE_RETA_GROUP_SIZE;
 		pos = i % RTE_RETA_GROUP_SIZE;
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 68b03d246..26eb30a40 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -1009,7 +1009,6 @@ mlx5_priv_rxq_get(struct priv *priv, uint16_t idx)
 		rxq_ctrl = container_of((*priv->rxqs)[idx],
 					struct mlx5_rxq_ctrl,
 					rxq);
-
 		mlx5_priv_rxq_ibv_get(priv, idx);
 		rte_atomic32_inc(&rxq_ctrl->refcnt);
 		DEBUG("%p: Rx queue %p: refcnt %d", (void *)priv,
diff --git a/drivers/net/mlx5/mlx5_vlan.c b/drivers/net/mlx5/mlx5_vlan.c
index 198a69e3c..aa8b319d7 100644
--- a/drivers/net/mlx5/mlx5_vlan.c
+++ b/drivers/net/mlx5/mlx5_vlan.c
@@ -137,14 +137,12 @@ priv_vlan_strip_queue_set(struct priv *priv, uint16_t idx, int on)
 		.flags_mask = IBV_WQ_FLAGS_CVLAN_STRIPPING,
 		.flags = vlan_offloads,
 	};
-
 	err = ibv_modify_wq(rxq_ctrl->ibv->wq, &mod);
 	if (err) {
 		ERROR("%p: failed to modified stripping mode: %s",
 		      (void *)priv, strerror(err));
 		return;
 	}
-
 	/* Update related bits in RX queue. */
 	rxq->vlan_strip = !!on;
 }
@@ -169,13 +167,11 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 		ERROR("VLAN stripping is not supported");
 		return;
 	}
-
 	/* Validate queue number */
 	if (queue >= priv->rxqs_n) {
 		ERROR("VLAN stripping, invalid queue number %d", queue);
 		return;
 	}
-
 	priv_lock(priv);
 	priv_vlan_strip_queue_set(priv, queue, on);
 	priv_unlock(priv);
@@ -202,13 +198,11 @@ mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask)
 			ERROR("VLAN stripping is not supported");
 			return 0;
 		}
-
 		/* Run on every RX queue and set/reset VLAN stripping. */
 		priv_lock(priv);
 		for (i = 0; (i != priv->rxqs_n); i++)
 			priv_vlan_strip_queue_set(priv, i, hw_vlan_strip);
 		priv_unlock(priv);
 	}
-
 	return 0;
 }
-- 
2.11.0

[dpdk-stable] [PATCH v2 24/67] net/mlx5: remove control path locks

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 7b2423cd2e84c61d3052147cea7643d15d49d1e9 ]

In priv struct only the memory region needs to be protected against
concurrent access between the control plane and the data plane.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.c         |  2 --
 drivers/net/mlx5/mlx5.h         | 43 +-----------------------------
 drivers/net/mlx5/mlx5_ethdev.c  | 58 +++--------------------------------------
 drivers/net/mlx5/mlx5_flow.c    | 18 +------------
 drivers/net/mlx5/mlx5_mr.c      |  4 +--
 drivers/net/mlx5/mlx5_rss.c     |  8 ------
 drivers/net/mlx5/mlx5_rxq.c     |  9 -------
 drivers/net/mlx5/mlx5_stats.c   | 15 +----------
 drivers/net/mlx5/mlx5_trigger.c |  7 -----
 drivers/net/mlx5/mlx5_txq.c     |  5 ----
 drivers/net/mlx5/mlx5_vlan.c    |  6 -----
 11 files changed, 9 insertions(+), 166 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 9ee9d2c36..98e877040 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -207,7 +207,6 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 	unsigned int i;
 	int ret;
 
-	priv_lock(priv);
 	DEBUG("%p: closing device \"%s\"",
 	      (void *)dev,
 	      ((priv->ctx != NULL) ? priv->ctx->device->name : ""));
@@ -269,7 +268,6 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 	ret = priv_mr_verify(priv);
 	if (ret)
 		WARN("%p: some Memory Region still remain", (void *)priv);
-	priv_unlock(priv);
 	memset(priv, 0, sizeof(*priv));
 }
 
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index dc165a8a3..6eacc3738 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -163,7 +163,7 @@ struct priv {
 	LIST_HEAD(ind_tables, mlx5_ind_table_ibv) ind_tbls;
 	uint32_t link_speed_capa; /* Link speed capabilities. */
 	struct mlx5_xstats_ctrl xstats_ctrl; /* Extended stats control. */
-	rte_spinlock_t lock; /* Lock for control functions. */
+	rte_spinlock_t mr_lock; /* MR Lock. */
 	int primary_socket; /* Unix socket for primary process. */
 	void *uar_base; /* Reserved address space for UAR mapping */
 	struct rte_intr_handle intr_handle_socket; /* Interrupt handler. */
@@ -171,47 +171,6 @@ struct priv {
 	/* Context for Verbs allocator. */
 };
 
-/**
- * Lock private structure to protect it from concurrent access in the
- * control path.
- *
- * @param priv
- *   Pointer to private structure.
- */
-static inline void
-priv_lock(struct priv *priv)
-{
-	rte_spinlock_lock(&priv->lock);
-}
-
-/**
- * Try to lock private structure to protect it from concurrent access in the
- * control path.
- *
- * @param priv
- *   Pointer to private structure.
- *
- * @return
- *   1 if the lock is successfully taken; 0 otherwise.
- */
-static inline int
-priv_trylock(struct priv *priv)
-{
-	return rte_spinlock_trylock(&priv->lock);
-}
-
-/**
- * Unlock private structure.
- *
- * @param priv
- *   Pointer to private structure.
- */
-static inline void
-priv_unlock(struct priv *priv)
-{
-	rte_spinlock_unlock(&priv->lock);
-}
-
 /* mlx5.c */
 
 int mlx5_getenv_int(const char *);
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 79d62bbc1..650b6287c 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -296,18 +296,16 @@ priv_set_flags(struct priv *priv, unsigned int keep, unsigned int flags)
 }
 
 /**
- * Ethernet device configuration.
- *
- * Prepare the driver for a given number of TX and RX queues.
+ * DPDK callback for Ethernet device configuration.
  *
  * @param dev
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, negative errno value on failure.
  */
-static int
-dev_configure(struct rte_eth_dev *dev)
+int
+mlx5_dev_configure(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 	unsigned int rxqs_n = dev->data->nb_rx_queues;
@@ -371,28 +369,7 @@ dev_configure(struct rte_eth_dev *dev)
 			j = 0;
 	}
 	return 0;
-}
-
-/**
- * DPDK callback for Ethernet device configuration.
- *
- * @param dev
- *   Pointer to Ethernet device structure.
- *
- * @return
- *   0 on success, negative errno value on failure.
- */
-int
-mlx5_dev_configure(struct rte_eth_dev *dev)
-{
-	struct priv *priv = dev->data->dev_private;
-	int ret;
 
-	priv_lock(priv);
-	ret = dev_configure(dev);
-	assert(ret >= 0);
-	priv_unlock(priv);
-	return -ret;
 }
 
 /**
@@ -411,7 +388,6 @@ mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info)
 	char ifname[IF_NAMESIZE];
 
 	info->pci_dev = RTE_ETH_DEV_TO_PCI(dev);
-	priv_lock(priv);
 	/* FIXME: we should ask the device for these values. */
 	info->min_rx_bufsize = 32;
 	info->max_rx_pktlen = 65536;
@@ -456,7 +432,6 @@ mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info)
 	info->hash_key_size = priv->rss_conf.rss_key_len;
 	info->speed_capa = priv->link_speed_capa;
 	info->flow_type_rss_offloads = ~MLX5_RSS_HF_MASK;
-	priv_unlock(priv);
 }
 
 /**
@@ -515,7 +490,6 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 	struct rte_eth_link dev_link;
 	int link_speed = 0;
 
-	/* priv_lock() is not taken to allow concurrent calls. */
 	if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) {
 		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
 		return -1;
@@ -781,9 +755,7 @@ mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
 	struct priv *priv = dev->data->dev_private;
 	int ret;
 
-	priv_lock(priv);
 	ret = priv_link_update(priv, wait_to_complete);
-	priv_unlock(priv);
 	return ret;
 }
 
@@ -805,7 +777,6 @@ mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 	uint16_t kern_mtu;
 	int ret = 0;
 
-	priv_lock(priv);
 	ret = priv_get_mtu(priv, &kern_mtu);
 	if (ret)
 		goto out;
@@ -820,13 +791,11 @@ mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 		priv->mtu = mtu;
 		DEBUG("adapter port %u MTU set to %u", priv->port, mtu);
 	}
-	priv_unlock(priv);
 	return 0;
 out:
 	ret = errno;
 	WARN("cannot set port %u MTU to %u: %s", priv->port, mtu,
 	     strerror(ret));
-	priv_unlock(priv);
 	assert(ret >= 0);
 	return -ret;
 }
@@ -853,7 +822,6 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 	int ret;
 
 	ifr.ifr_data = (void *)&ethpause;
-	priv_lock(priv);
 	if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
 		ret = errno;
 		WARN("ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM)"
@@ -872,7 +840,6 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		fc_conf->mode = RTE_FC_NONE;
 	ret = 0;
 out:
-	priv_unlock(priv);
 	assert(ret >= 0);
 	return -ret;
 }
@@ -911,7 +878,6 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		ethpause.tx_pause = 1;
 	else
 		ethpause.tx_pause = 0;
-	priv_lock(priv);
 	if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
 		ret = errno;
 		WARN("ioctl(SIOCETHTOOL, ETHTOOL_SPAUSEPARAM)"
@@ -921,7 +887,6 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 	}
 	ret = 0;
 out:
-	priv_unlock(priv);
 	assert(ret >= 0);
 	return -ret;
 }
@@ -1064,15 +1029,8 @@ mlx5_dev_link_status_handler(void *arg)
 	struct priv *priv = dev->data->dev_private;
 	int ret;
 
-	while (!priv_trylock(priv)) {
-		/* Alarm is being canceled. */
-		if (priv->pending_alarm == 0)
-			return;
-		rte_pause();
-	}
 	priv->pending_alarm = 0;
 	ret = priv_link_status_update(priv);
-	priv_unlock(priv);
 	if (!ret)
 		_rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL,
 					      NULL);
@@ -1093,9 +1051,7 @@ mlx5_dev_interrupt_handler(void *cb_arg)
 	struct priv *priv = dev->data->dev_private;
 	uint32_t events;
 
-	priv_lock(priv);
 	events = priv_dev_status_handler(priv);
-	priv_unlock(priv);
 	if (events & (1 << RTE_ETH_EVENT_INTR_LSC))
 		_rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL,
 					      NULL);
@@ -1116,9 +1072,7 @@ mlx5_dev_handler_socket(void *cb_arg)
 	struct rte_eth_dev *dev = cb_arg;
 	struct priv *priv = dev->data->dev_private;
 
-	priv_lock(priv);
 	priv_socket_handle(priv);
-	priv_unlock(priv);
 }
 
 /**
@@ -1218,9 +1172,7 @@ mlx5_set_link_down(struct rte_eth_dev *dev)
 	struct priv *priv = dev->data->dev_private;
 	int err;
 
-	priv_lock(priv);
 	err = priv_dev_set_link(priv, 0);
-	priv_unlock(priv);
 	return err;
 }
 
@@ -1239,9 +1191,7 @@ mlx5_set_link_up(struct rte_eth_dev *dev)
 	struct priv *priv = dev->data->dev_private;
 	int err;
 
-	priv_lock(priv);
 	err = priv_dev_set_link(priv, 1);
-	priv_unlock(priv);
 	return err;
 }
 
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 9f48ccf3d..d51f0ece5 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -1937,9 +1937,7 @@ mlx5_flow_validate(struct rte_eth_dev *dev,
 	int ret;
 	struct mlx5_flow_parse parser = { .create = 0, };
 
-	priv_lock(priv);
 	ret = priv_flow_convert(priv, attr, items, actions, error, &parser);
-	priv_unlock(priv);
 	return ret;
 }
 
@@ -1959,10 +1957,8 @@ mlx5_flow_create(struct rte_eth_dev *dev,
 	struct priv *priv = dev->data->dev_private;
 	struct rte_flow *flow;
 
-	priv_lock(priv);
 	flow = priv_flow_create(priv, &priv->flows, attr, items, actions,
 				error);
-	priv_unlock(priv);
 	return flow;
 }
 
@@ -2444,9 +2440,7 @@ mlx5_flow_destroy(struct rte_eth_dev *dev,
 {
 	struct priv *priv = dev->data->dev_private;
 
-	priv_lock(priv);
 	priv_flow_destroy(priv, &priv->flows, flow);
-	priv_unlock(priv);
 	return 0;
 }
 
@@ -2462,9 +2456,7 @@ mlx5_flow_flush(struct rte_eth_dev *dev,
 {
 	struct priv *priv = dev->data->dev_private;
 
-	priv_lock(priv);
 	priv_flow_flush(priv, &priv->flows);
-	priv_unlock(priv);
 	return 0;
 }
 
@@ -2522,16 +2514,14 @@ priv_flow_query_count(struct ibv_counter_set *cs,
  * @see rte_flow_ops
  */
 int
-mlx5_flow_query(struct rte_eth_dev *dev,
+mlx5_flow_query(struct rte_eth_dev *dev __rte_unused,
 		struct rte_flow *flow,
 		enum rte_flow_action_type action __rte_unused,
 		void *data,
 		struct rte_flow_error *error)
 {
-	struct priv *priv = dev->data->dev_private;
 	int res = EINVAL;
 
-	priv_lock(priv);
 	if (flow->cs) {
 		res = priv_flow_query_count(flow->cs,
 					&flow->counter_stats,
@@ -2543,7 +2533,6 @@ mlx5_flow_query(struct rte_eth_dev *dev,
 				   NULL,
 				   "no counter found for flow");
 	}
-	priv_unlock(priv);
 	return -res;
 }
 #endif
@@ -2561,13 +2550,11 @@ mlx5_flow_isolate(struct rte_eth_dev *dev,
 {
 	struct priv *priv = dev->data->dev_private;
 
-	priv_lock(priv);
 	if (dev->data->dev_started) {
 		rte_flow_error_set(error, EBUSY,
 				   RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
 				   NULL,
 				   "port must be stopped first");
-		priv_unlock(priv);
 		return -rte_errno;
 	}
 	priv->isolated = !!enable;
@@ -2575,7 +2562,6 @@ mlx5_flow_isolate(struct rte_eth_dev *dev,
 		priv->dev->dev_ops = &mlx5_dev_ops_isolate;
 	else
 		priv->dev->dev_ops = &mlx5_dev_ops;
-	priv_unlock(priv);
 	return 0;
 }
 
@@ -3057,9 +3043,7 @@ mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
 		*(const void **)arg = &mlx5_flow_ops;
 		return 0;
 	case RTE_ETH_FILTER_FDIR:
-		priv_lock(priv);
 		ret = priv_fdir_ctrl_func(priv, filter_op, arg);
-		priv_unlock(priv);
 		break;
 	default:
 		ERROR("%p: filter type (%d) not supported",
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 7da54758c..12d9b2f8e 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -193,9 +193,9 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 		container_of(txq, struct mlx5_txq_ctrl, txq);
 	struct mlx5_mr *mr;
 
-	priv_lock(txq_ctrl->priv);
+	rte_spinlock_lock(&txq_ctrl->priv->mr_lock);
 	mr = priv_txq_mp2mr_reg(txq_ctrl->priv, txq, mp, idx);
-	priv_unlock(txq_ctrl->priv);
+	rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 	return mr;
 }
 
diff --git a/drivers/net/mlx5/mlx5_rss.c b/drivers/net/mlx5/mlx5_rss.c
index 8f0882c89..45838f0b3 100644
--- a/drivers/net/mlx5/mlx5_rss.c
+++ b/drivers/net/mlx5/mlx5_rss.c
@@ -72,7 +72,6 @@ mlx5_rss_hash_update(struct rte_eth_dev *dev,
 	struct priv *priv = dev->data->dev_private;
 	int ret = 0;
 
-	priv_lock(priv);
 	if (rss_conf->rss_hf & MLX5_RSS_HF_MASK) {
 		ret = -EINVAL;
 		goto out;
@@ -90,7 +89,6 @@ mlx5_rss_hash_update(struct rte_eth_dev *dev,
 	}
 	priv->rss_conf.rss_hf = rss_conf->rss_hf;
 out:
-	priv_unlock(priv);
 	return ret;
 }
 
@@ -113,7 +111,6 @@ mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
 
 	if (!rss_conf)
 		return -EINVAL;
-	priv_lock(priv);
 	if (rss_conf->rss_key &&
 	    (rss_conf->rss_key_len >= priv->rss_conf.rss_key_len)) {
 		memcpy(rss_conf->rss_key, priv->rss_conf.rss_key,
@@ -121,7 +118,6 @@ mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
 	}
 	rss_conf->rss_key_len = priv->rss_conf.rss_key_len;
 	rss_conf->rss_hf = priv->rss_conf.rss_hf;
-	priv_unlock(priv);
 	return 0;
 }
 
@@ -250,9 +246,7 @@ mlx5_dev_rss_reta_query(struct rte_eth_dev *dev,
 	int ret;
 	struct priv *priv = dev->data->dev_private;
 
-	priv_lock(priv);
 	ret = priv_dev_rss_reta_query(priv, reta_conf, reta_size);
-	priv_unlock(priv);
 	return -ret;
 }
 
@@ -277,9 +271,7 @@ mlx5_dev_rss_reta_update(struct rte_eth_dev *dev,
 	int ret;
 	struct priv *priv = dev->data->dev_private;
 
-	priv_lock(priv);
 	ret = priv_dev_rss_reta_update(priv, reta_conf, reta_size);
-	priv_unlock(priv);
 	if (dev->data->dev_started) {
 		mlx5_dev_stop(dev);
 		mlx5_dev_start(dev);
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 26eb30a40..91f598c18 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -242,7 +242,6 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		container_of(rxq, struct mlx5_rxq_ctrl, rxq);
 	int ret = 0;
 
-	priv_lock(priv);
 	if (!rte_is_power_of_2(desc)) {
 		desc = 1 << log2above(desc);
 		WARN("%p: increased number of descriptors in RX queue %u"
@@ -254,7 +253,6 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	if (idx >= priv->rxqs_n) {
 		ERROR("%p: queue index out of range (%u >= %u)",
 		      (void *)dev, idx, priv->rxqs_n);
-		priv_unlock(priv);
 		return -EOVERFLOW;
 	}
 	if (!mlx5_priv_rxq_releasable(priv, idx)) {
@@ -275,7 +273,6 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	      (void *)dev, (void *)rxq_ctrl);
 	(*priv->rxqs)[idx] = &rxq_ctrl->rxq;
 out:
-	priv_unlock(priv);
 	return -ret;
 }
 
@@ -296,12 +293,10 @@ mlx5_rx_queue_release(void *dpdk_rxq)
 		return;
 	rxq_ctrl = container_of(rxq, struct mlx5_rxq_ctrl, rxq);
 	priv = rxq_ctrl->priv;
-	priv_lock(priv);
 	if (!mlx5_priv_rxq_releasable(priv, rxq_ctrl->rxq.stats.idx))
 		rte_panic("Rx queue %p is still used by a flow and cannot be"
 			  " removed\n", (void *)rxq_ctrl);
 	mlx5_priv_rxq_release(priv, rxq_ctrl->rxq.stats.idx);
-	priv_unlock(priv);
 }
 
 /**
@@ -459,7 +454,6 @@ mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	struct mlx5_rxq_ctrl *rxq_ctrl;
 	int ret = 0;
 
-	priv_lock(priv);
 	rxq_data = (*priv->rxqs)[rx_queue_id];
 	if (!rxq_data) {
 		ret = EINVAL;
@@ -478,7 +472,6 @@ mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 		mlx5_priv_rxq_ibv_release(priv, rxq_ibv);
 	}
 exit:
-	priv_unlock(priv);
 	if (ret)
 		WARN("unable to arm interrupt on rx queue %d", rx_queue_id);
 	return -ret;
@@ -506,7 +499,6 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	void *ev_ctx;
 	int ret = 0;
 
-	priv_lock(priv);
 	rxq_data = (*priv->rxqs)[rx_queue_id];
 	if (!rxq_data) {
 		ret = EINVAL;
@@ -530,7 +522,6 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 exit:
 	if (rxq_ibv)
 		mlx5_priv_rxq_ibv_release(priv, rxq_ibv);
-	priv_unlock(priv);
 	if (ret)
 		WARN("unable to disable interrupt on rx queue %d",
 		     rx_queue_id);
diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index 50e3ea5f2..785563227 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -340,7 +340,6 @@ mlx5_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
 	unsigned int i;
 	unsigned int idx;
 
-	priv_lock(priv);
 	/* Add software counters. */
 	for (i = 0; (i != priv->rxqs_n); ++i) {
 		struct mlx5_rxq_data *rxq = (*priv->rxqs)[i];
@@ -386,7 +385,6 @@ mlx5_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
 	/* FIXME: retrieve and add hardware counters. */
 #endif
 	*stats = tmp;
-	priv_unlock(priv);
 	return 0;
 }
 
@@ -403,7 +401,6 @@ mlx5_stats_reset(struct rte_eth_dev *dev)
 	unsigned int i;
 	unsigned int idx;
 
-	priv_lock(priv);
 	for (i = 0; (i != priv->rxqs_n); ++i) {
 		if ((*priv->rxqs)[i] == NULL)
 			continue;
@@ -421,7 +418,6 @@ mlx5_stats_reset(struct rte_eth_dev *dev)
 #ifndef MLX5_PMD_SOFT_COUNTERS
 	/* FIXME: reset hardware counters. */
 #endif
-	priv_unlock(priv);
 }
 
 /**
@@ -448,16 +444,13 @@ mlx5_xstats_get(struct rte_eth_dev *dev,
 		struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
 		int stats_n;
 
-		priv_lock(priv);
 		stats_n = priv_ethtool_get_stats_n(priv);
 		if (stats_n < 0) {
-			priv_unlock(priv);
 			return -1;
 		}
 		if (xstats_ctrl->stats_n != stats_n)
 			priv_xstats_init(priv);
 		ret = priv_xstats_get(priv, stats);
-		priv_unlock(priv);
 	}
 	return ret;
 }
@@ -475,15 +468,12 @@ mlx5_xstats_reset(struct rte_eth_dev *dev)
 	struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
 	int stats_n;
 
-	priv_lock(priv);
 	stats_n = priv_ethtool_get_stats_n(priv);
 	if (stats_n < 0)
-		goto unlock;
+		return;
 	if (xstats_ctrl->stats_n != stats_n)
 		priv_xstats_init(priv);
 	priv_xstats_reset(priv);
-unlock:
-	priv_unlock(priv);
 }
 
 /**
@@ -503,18 +493,15 @@ int
 mlx5_xstats_get_names(struct rte_eth_dev *dev __rte_unused,
 		struct rte_eth_xstat_name *xstats_names, unsigned int n)
 {
-	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
 	if (n >= xstats_n && xstats_names) {
-		priv_lock(priv);
 		for (i = 0; i != xstats_n; ++i) {
 			strncpy(xstats_names[i].name,
 				mlx5_counters_init[i].dpdk_name,
 				RTE_ETH_XSTATS_NAME_SIZE);
 			xstats_names[i].name[RTE_ETH_XSTATS_NAME_SIZE - 1] = 0;
 		}
-		priv_unlock(priv);
 	}
 	return xstats_n;
 }
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index ad15158fe..2ae532d92 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -169,7 +169,6 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 	int err;
 
 	dev->data->dev_started = 1;
-	priv_lock(priv);
 	err = priv_flow_create_drop_queue(priv);
 	if (err) {
 		ERROR("%p: Drop queue allocation failed: %s",
@@ -207,7 +206,6 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 		goto error;
 	}
 	priv_dev_interrupt_handler_install(priv, dev);
-	priv_unlock(priv);
 	return 0;
 error:
 	/* Rollback. */
@@ -219,7 +217,6 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 	priv_txq_stop(priv);
 	priv_rxq_stop(priv);
 	priv_flow_delete_drop_queue(priv);
-	priv_unlock(priv);
 	return err;
 }
 
@@ -237,7 +234,6 @@ mlx5_dev_stop(struct rte_eth_dev *dev)
 	struct priv *priv = dev->data->dev_private;
 	struct mlx5_mr *mr;
 
-	priv_lock(priv);
 	dev->data->dev_started = 0;
 	/* Prevent crashes when queues are still in use. */
 	dev->rx_pkt_burst = removed_rx_burst;
@@ -254,7 +250,6 @@ mlx5_dev_stop(struct rte_eth_dev *dev)
 	for (mr = LIST_FIRST(&priv->mr); mr; mr = LIST_FIRST(&priv->mr))
 		priv_mr_release(priv, mr);
 	priv_flow_delete_drop_queue(priv);
-	priv_unlock(priv);
 }
 
 /**
@@ -439,8 +434,6 @@ mlx5_traffic_restart(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 
-	priv_lock(priv);
 	priv_dev_traffic_restart(priv, dev);
-	priv_unlock(priv);
 	return 0;
 }
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 09d62a95b..f3b3daecb 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -142,7 +142,6 @@ mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		container_of(txq, struct mlx5_txq_ctrl, txq);
 	int ret = 0;
 
-	priv_lock(priv);
 	if (desc <= MLX5_TX_COMP_THRESH) {
 		WARN("%p: number of descriptors requested for TX queue %u"
 		     " must be higher than MLX5_TX_COMP_THRESH, using"
@@ -161,7 +160,6 @@ mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	if (idx >= priv->txqs_n) {
 		ERROR("%p: queue index out of range (%u >= %u)",
 		      (void *)dev, idx, priv->txqs_n);
-		priv_unlock(priv);
 		return -EOVERFLOW;
 	}
 	if (!mlx5_priv_txq_releasable(priv, idx)) {
@@ -182,7 +180,6 @@ mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	      (void *)dev, (void *)txq_ctrl);
 	(*priv->txqs)[idx] = &txq_ctrl->txq;
 out:
-	priv_unlock(priv);
 	return -ret;
 }
 
@@ -204,7 +201,6 @@ mlx5_tx_queue_release(void *dpdk_txq)
 		return;
 	txq_ctrl = container_of(txq, struct mlx5_txq_ctrl, txq);
 	priv = txq_ctrl->priv;
-	priv_lock(priv);
 	for (i = 0; (i != priv->txqs_n); ++i)
 		if ((*priv->txqs)[i] == txq) {
 			DEBUG("%p: removing TX queue %p from list",
@@ -212,7 +208,6 @@ mlx5_tx_queue_release(void *dpdk_txq)
 			mlx5_priv_txq_release(priv, i);
 			break;
 		}
-	priv_unlock(priv);
 }
 
 
diff --git a/drivers/net/mlx5/mlx5_vlan.c b/drivers/net/mlx5/mlx5_vlan.c
index aa8b319d7..653c5fef4 100644
--- a/drivers/net/mlx5/mlx5_vlan.c
+++ b/drivers/net/mlx5/mlx5_vlan.c
@@ -63,7 +63,6 @@ mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
 	unsigned int i;
 	int ret = 0;
 
-	priv_lock(priv);
 	DEBUG("%p: %s VLAN filter ID %" PRIu16,
 	      (void *)dev, (on ? "enable" : "disable"), vlan_id);
 	assert(priv->vlan_filter_n <= RTE_DIM(priv->vlan_filter));
@@ -99,7 +98,6 @@ mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
 	if (dev->data->dev_started)
 		priv_dev_traffic_restart(priv, dev);
 out:
-	priv_unlock(priv);
 	return ret;
 }
 
@@ -172,9 +170,7 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 		ERROR("VLAN stripping, invalid queue number %d", queue);
 		return;
 	}
-	priv_lock(priv);
 	priv_vlan_strip_queue_set(priv, queue, on);
-	priv_unlock(priv);
 }
 
 /**
@@ -199,10 +195,8 @@ mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask)
 			return 0;
 		}
 		/* Run on every RX queue and set/reset VLAN stripping. */
-		priv_lock(priv);
 		for (i = 0; (i != priv->rxqs_n); i++)
 			priv_vlan_strip_queue_set(priv, i, hw_vlan_strip);
-		priv_unlock(priv);
 	}
 	return 0;
 }
-- 
2.11.0

[dpdk-stable] [PATCH v2 25/67] net/mlx5: prefix all functions with mlx5

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit af4f09f28294fac762ff413fbf14b48c42c128fd ]

This change removes the need to distinguish unlocked priv_*() functions
which are therefore renamed using a mlx5_*() prefix for consistency.

At the same time, all functions from mlx5 uses a pointer to the ETH device
instead of the one to the PMD private data.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.c          | 104 ++++++------
 drivers/net/mlx5/mlx5.h          |  88 +++++-----
 drivers/net/mlx5/mlx5_ethdev.c   | 273 +++++++++++++------------------
 drivers/net/mlx5/mlx5_flow.c     | 343 +++++++++++++++++++--------------------
 drivers/net/mlx5/mlx5_mac.c      |   8 +-
 drivers/net/mlx5/mlx5_mr.c       |  86 ++++------
 drivers/net/mlx5/mlx5_rss.c      | 107 ++++--------
 drivers/net/mlx5/mlx5_rxq.c      | 253 +++++++++++++++--------------
 drivers/net/mlx5/mlx5_rxtx.c     |   8 +-
 drivers/net/mlx5/mlx5_rxtx.h     | 107 ++++++------
 drivers/net/mlx5/mlx5_rxtx_vec.c |  26 +--
 drivers/net/mlx5/mlx5_socket.c   |  29 ++--
 drivers/net/mlx5/mlx5_stats.c    | 135 ++++++---------
 drivers/net/mlx5/mlx5_trigger.c  | 118 ++++++--------
 drivers/net/mlx5/mlx5_txq.c      | 108 ++++++------
 drivers/net/mlx5/mlx5_vlan.c     |  61 +++----
 16 files changed, 832 insertions(+), 1022 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 98e877040..ebb778826 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -211,8 +211,8 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 	      (void *)dev,
 	      ((priv->ctx != NULL) ? priv->ctx->device->name : ""));
 	/* In case mlx5_dev_stop() has not been called. */
-	priv_dev_interrupt_handler_uninstall(priv, dev);
-	priv_dev_traffic_disable(priv, dev);
+	mlx5_dev_interrupt_handler_uninstall(dev);
+	mlx5_traffic_disable(dev);
 	/* Prevent crashes when queues are still in use. */
 	dev->rx_pkt_burst = removed_rx_burst;
 	dev->tx_pkt_burst = removed_tx_burst;
@@ -220,7 +220,7 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 		/* XXX race condition if mlx5_rx_burst() is still running. */
 		usleep(1000);
 		for (i = 0; (i != priv->rxqs_n); ++i)
-			mlx5_priv_rxq_release(priv, i);
+			mlx5_rxq_release(dev, i);
 		priv->rxqs_n = 0;
 		priv->rxqs = NULL;
 	}
@@ -228,7 +228,7 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 		/* XXX race condition if mlx5_tx_burst() is still running. */
 		usleep(1000);
 		for (i = 0; (i != priv->txqs_n); ++i)
-			mlx5_priv_txq_release(priv, i);
+			mlx5_txq_release(dev, i);
 		priv->txqs_n = 0;
 		priv->txqs = NULL;
 	}
@@ -243,31 +243,31 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 	if (priv->reta_idx != NULL)
 		rte_free(priv->reta_idx);
 	if (priv->primary_socket)
-		priv_socket_uninit(priv);
-	ret = mlx5_priv_hrxq_ibv_verify(priv);
+		mlx5_socket_uninit(dev);
+	ret = mlx5_hrxq_ibv_verify(dev);
 	if (ret)
-		WARN("%p: some Hash Rx queue still remain", (void *)priv);
-	ret = mlx5_priv_ind_table_ibv_verify(priv);
+		WARN("%p: some Hash Rx queue still remain", (void *)dev);
+	ret = mlx5_ind_table_ibv_verify(dev);
 	if (ret)
-		WARN("%p: some Indirection table still remain", (void *)priv);
-	ret = mlx5_priv_rxq_ibv_verify(priv);
+		WARN("%p: some Indirection table still remain", (void *)dev);
+	ret = mlx5_rxq_ibv_verify(dev);
 	if (ret)
-		WARN("%p: some Verbs Rx queue still remain", (void *)priv);
-	ret = mlx5_priv_rxq_verify(priv);
+		WARN("%p: some Verbs Rx queue still remain", (void *)dev);
+	ret = mlx5_rxq_verify(dev);
 	if (ret)
-		WARN("%p: some Rx Queues still remain", (void *)priv);
-	ret = mlx5_priv_txq_ibv_verify(priv);
+		WARN("%p: some Rx Queues still remain", (void *)dev);
+	ret = mlx5_txq_ibv_verify(dev);
 	if (ret)
-		WARN("%p: some Verbs Tx queue still remain", (void *)priv);
-	ret = mlx5_priv_txq_verify(priv);
+		WARN("%p: some Verbs Tx queue still remain", (void *)dev);
+	ret = mlx5_txq_verify(dev);
 	if (ret)
-		WARN("%p: some Tx Queues still remain", (void *)priv);
-	ret = priv_flow_verify(priv);
+		WARN("%p: some Tx Queues still remain", (void *)dev);
+	ret = mlx5_flow_verify(dev);
 	if (ret)
-		WARN("%p: some flows still remain", (void *)priv);
-	ret = priv_mr_verify(priv);
+		WARN("%p: some flows still remain", (void *)dev);
+	ret = mlx5_mr_verify(dev);
 	if (ret)
-		WARN("%p: some Memory Region still remain", (void *)priv);
+		WARN("%p: some Memory Region still remain", (void *)dev);
 	memset(priv, 0, sizeof(*priv));
 }
 
@@ -509,15 +509,16 @@ static void *uar_base;
 /**
  * Reserve UAR address space for primary process.
  *
- * @param[in] priv
- *   Pointer to private structure.
+ * @param[in] dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   0 on success, errno value on failure.
  */
 static int
-priv_uar_init_primary(struct priv *priv)
+mlx5_uar_init_primary(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	void *addr = (void *)0;
 	int i;
 	const struct rte_mem_config *mcfg;
@@ -559,15 +560,16 @@ priv_uar_init_primary(struct priv *priv)
  * Reserve UAR address space for secondary process, align with
  * primary process.
  *
- * @param[in] priv
- *   Pointer to private structure.
+ * @param[in] dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   0 on success, errno value on failure.
  */
 static int
-priv_uar_init_secondary(struct priv *priv)
+mlx5_uar_init_secondary(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	void *addr;
 	int ret;
 
@@ -764,7 +766,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		struct ibv_port_attr port_attr;
 		struct ibv_pd *pd = NULL;
 		struct priv *priv = NULL;
-		struct rte_eth_dev *eth_dev;
+		struct rte_eth_dev *eth_dev = NULL;
 		struct ibv_device_attr_ex device_attr_ex;
 		struct ether_addr mac;
 		struct ibv_device_attr_ex device_attr;
@@ -793,24 +795,23 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			}
 			eth_dev->device = &pci_dev->device;
 			eth_dev->dev_ops = &mlx5_dev_sec_ops;
-			priv = eth_dev->data->dev_private;
-			err = priv_uar_init_secondary(priv);
+			err = mlx5_uar_init_secondary(eth_dev);
 			if (err < 0) {
 				err = -err;
 				goto error;
 			}
 			/* Receive command fd from primary process */
-			err = priv_socket_connect(priv);
+			err = mlx5_socket_connect(eth_dev);
 			if (err < 0) {
 				err = -err;
 				goto error;
 			}
 			/* Remap UAR for Tx queues. */
-			err = priv_tx_uar_remap(priv, err);
+			err = mlx5_tx_uar_remap(eth_dev, err);
 			if (err)
 				goto error;
-			priv_dev_select_rx_function(priv, eth_dev);
-			priv_dev_select_tx_function(priv, eth_dev);
+			mlx5_select_rx_function(eth_dev);
+			mlx5_select_tx_function(eth_dev);
 			continue;
 		}
 		DEBUG("using port %u (%08" PRIx32 ")", port, test);
@@ -956,11 +957,23 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			WARN("Rx CQE compression isn't supported");
 			priv->cqe_comp = 0;
 		}
-		err = priv_uar_init_primary(priv);
+		eth_dev = rte_eth_dev_allocate(name);
+		if (eth_dev == NULL) {
+			ERROR("can not allocate rte ethdev");
+			err = ENOMEM;
+			goto port_error;
+		}
+		eth_dev->data->dev_private = priv;
+		priv->dev = eth_dev;
+		eth_dev->data->mac_addrs = priv->mac;
+		eth_dev->device = &pci_dev->device;
+		rte_eth_copy_pci_info(eth_dev, pci_dev);
+		eth_dev->device->driver = &mlx5_driver.driver;
+		err = mlx5_uar_init_primary(eth_dev);
 		if (err)
 			goto port_error;
 		/* Configure the first MAC address by default. */
-		if (priv_get_mac(priv, &mac.addr_bytes)) {
+		if (mlx5_get_mac(eth_dev, &mac.addr_bytes)) {
 			ERROR("cannot get MAC address, is mlx5_en loaded?"
 			      " (errno: %s)", strerror(errno));
 			err = ENODEV;
@@ -975,7 +988,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		{
 			char ifname[IF_NAMESIZE];
 
-			if (priv_get_ifname(priv, &ifname) == 0)
+			if (mlx5_get_ifname(eth_dev, &ifname) == 0)
 				DEBUG("port %u ifname is \"%s\"",
 				      priv->port, ifname);
 			else
@@ -983,25 +996,13 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		}
 #endif
 		/* Get actual MTU if possible. */
-		priv_get_mtu(priv, &priv->mtu);
+		mlx5_get_mtu(eth_dev, &priv->mtu);
 		DEBUG("port %u MTU is %u", priv->port, priv->mtu);
-		eth_dev = rte_eth_dev_allocate(name);
-		if (eth_dev == NULL) {
-			ERROR("can not allocate rte ethdev");
-			err = ENOMEM;
-			goto port_error;
-		}
-		eth_dev->data->dev_private = priv;
-		eth_dev->data->mac_addrs = priv->mac;
-		eth_dev->device = &pci_dev->device;
-		rte_eth_copy_pci_info(eth_dev, pci_dev);
-		eth_dev->device->driver = &mlx5_driver.driver;
 		/*
 		 * Initialize burst functions to prevent crashes before link-up.
 		 */
 		eth_dev->rx_pkt_burst = removed_rx_burst;
 		eth_dev->tx_pkt_burst = removed_tx_burst;
-		priv->dev = eth_dev;
 		eth_dev->dev_ops = &mlx5_dev_ops;
 		/* Register MAC address. */
 		claim_zero(mlx5_mac_addr_add(eth_dev, &mac, 0, 0));
@@ -1015,10 +1016,9 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		};
 		mlx5dv_set_context_attr(ctx, MLX5DV_CTX_ATTR_BUF_ALLOCATORS,
 					(void *)((uintptr_t)&alctr));
-
 		/* Bring Ethernet device up. */
 		DEBUG("forcing Ethernet interface up");
-		priv_set_flags(priv, ~IFF_UP, IFF_UP);
+		mlx5_set_flags(eth_dev, ~IFF_UP, IFF_UP);
 		continue;
 port_error:
 		if (priv)
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 6eacc3738..5814ec051 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -177,18 +177,16 @@ int mlx5_getenv_int(const char *);
 
 /* mlx5_ethdev.c */
 
-struct priv *mlx5_get_priv(struct rte_eth_dev *dev);
-int mlx5_is_secondary(void);
-int priv_get_ifname(const struct priv *priv, char (*ifname)[IF_NAMESIZE]);
-int priv_ifreq(const struct priv *priv, int req, struct ifreq *ifr);
-int priv_get_mtu(struct priv *priv, uint16_t *mtu);
-int priv_set_flags(struct priv *priv, unsigned int keep, unsigned int flags);
+int mlx5_get_ifname(const struct rte_eth_dev *dev, char (*ifname)[IF_NAMESIZE]);
+int mlx5_ifreq(const struct rte_eth_dev *dev, int req, struct ifreq *ifr);
+int mlx5_get_mtu(struct rte_eth_dev *dev, uint16_t *mtu);
+int mlx5_set_flags(struct rte_eth_dev *dev, unsigned int keep,
+		   unsigned int flags);
 int mlx5_dev_configure(struct rte_eth_dev *dev);
 void mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info);
 const uint32_t *mlx5_dev_supported_ptypes_get(struct rte_eth_dev *dev);
-int priv_link_update(struct priv *priv, int wait_to_complete);
-int priv_force_link_status_change(struct priv *priv, int status);
 int mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete);
+int mlx5_force_link_status_change(struct rte_eth_dev *dev, int status);
 int mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu);
 int mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev,
 			   struct rte_eth_fc_conf *fc_conf);
@@ -197,19 +195,17 @@ int mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev,
 int mlx5_ibv_device_to_pci_addr(const struct ibv_device *device,
 				struct rte_pci_addr *pci_addr);
 void mlx5_dev_link_status_handler(void *arg);
-void mlx5_dev_interrupt_handler(void *cb_arg);
-void priv_dev_interrupt_handler_uninstall(struct priv *priv,
-					  struct rte_eth_dev *dev);
-void priv_dev_interrupt_handler_install(struct priv *priv,
-					struct rte_eth_dev *dev);
+void mlx5_dev_interrupt_handler(void *arg);
+void mlx5_dev_interrupt_handler_uninstall(struct rte_eth_dev *dev);
+void mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev);
 int mlx5_set_link_down(struct rte_eth_dev *dev);
 int mlx5_set_link_up(struct rte_eth_dev *dev);
-void priv_dev_select_tx_function(struct priv *priv, struct rte_eth_dev *dev);
-void priv_dev_select_rx_function(struct priv *priv, struct rte_eth_dev *dev);
+void mlx5_select_tx_function(struct rte_eth_dev *dev);
+void mlx5_select_rx_function(struct rte_eth_dev *dev);
 
 /* mlx5_mac.c */
 
-int priv_get_mac(struct priv *priv, uint8_t (*mac)[ETHER_ADDR_LEN]);
+int mlx5_get_mac(struct rte_eth_dev *dev, uint8_t (*mac)[ETHER_ADDR_LEN]);
 void mlx5_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index);
 int mlx5_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac,
 		      uint32_t index, uint32_t vmdq);
@@ -221,7 +217,7 @@ int mlx5_rss_hash_update(struct rte_eth_dev *dev,
 			 struct rte_eth_rss_conf *rss_conf);
 int mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
 			   struct rte_eth_rss_conf *rss_conf);
-int priv_rss_reta_index_resize(struct priv *priv, unsigned int reta_size);
+int mlx5_rss_reta_index_resize(struct rte_eth_dev *dev, unsigned int reta_size);
 int mlx5_dev_rss_reta_query(struct rte_eth_dev *dev,
 			    struct rte_eth_rss_reta_entry64 *reta_conf,
 			    uint16_t reta_size);
@@ -238,13 +234,13 @@ void mlx5_allmulticast_disable(struct rte_eth_dev *dev);
 
 /* mlx5_stats.c */
 
-void priv_xstats_init(struct priv *priv);
+void mlx5_xstats_init(struct rte_eth_dev *dev);
 int mlx5_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats);
 void mlx5_stats_reset(struct rte_eth_dev *dev);
-int mlx5_xstats_get(struct rte_eth_dev *dev,
-		    struct rte_eth_xstat *stats, unsigned int n);
+int mlx5_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *stats,
+		    unsigned int n);
 void mlx5_xstats_reset(struct rte_eth_dev *dev);
-int mlx5_xstats_get_names(struct rte_eth_dev *dev,
+int mlx5_xstats_get_names(struct rte_eth_dev *dev __rte_unused,
 			  struct rte_eth_xstat_name *xstats_names,
 			  unsigned int n);
 
@@ -258,9 +254,8 @@ int mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask);
 
 int mlx5_dev_start(struct rte_eth_dev *dev);
 void mlx5_dev_stop(struct rte_eth_dev *dev);
-int priv_dev_traffic_enable(struct priv *priv, struct rte_eth_dev *dev);
-int priv_dev_traffic_disable(struct priv *priv, struct rte_eth_dev *dev);
-int priv_dev_traffic_restart(struct priv *priv, struct rte_eth_dev *dev);
+int mlx5_traffic_enable(struct rte_eth_dev *dev);
+int mlx5_traffic_disable(struct rte_eth_dev *dev);
 int mlx5_traffic_restart(struct rte_eth_dev *dev);
 
 /* mlx5_flow.c */
@@ -270,21 +265,6 @@ int mlx5_flow_validate(struct rte_eth_dev *dev,
 		       const struct rte_flow_item items[],
 		       const struct rte_flow_action actions[],
 		       struct rte_flow_error *error);
-void priv_flow_flush(struct priv *priv, struct mlx5_flows *list);
-int priv_flow_create_drop_queue(struct priv *priv);
-void priv_flow_stop(struct priv *priv, struct mlx5_flows *list);
-int priv_flow_start(struct priv *priv, struct mlx5_flows *list);
-int priv_flow_verify(struct priv *priv);
-int priv_flow_create_drop_queue(struct priv *priv);
-void priv_flow_delete_drop_queue(struct priv *priv);
-int mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
-			struct rte_flow_item_eth *eth_spec,
-			struct rte_flow_item_eth *eth_mask,
-			struct rte_flow_item_vlan *vlan_spec,
-			struct rte_flow_item_vlan *vlan_mask);
-int mlx5_ctrl_flow(struct rte_eth_dev *dev,
-		   struct rte_flow_item_eth *eth_spec,
-		   struct rte_flow_item_eth *eth_mask);
 struct rte_flow *mlx5_flow_create(struct rte_eth_dev *dev,
 				  const struct rte_flow_attr *attr,
 				  const struct rte_flow_item items[],
@@ -292,6 +272,7 @@ struct rte_flow *mlx5_flow_create(struct rte_eth_dev *dev,
 				  struct rte_flow_error *error);
 int mlx5_flow_destroy(struct rte_eth_dev *dev, struct rte_flow *flow,
 		      struct rte_flow_error *error);
+void mlx5_flow_list_flush(struct rte_eth_dev *dev, struct mlx5_flows *list);
 int mlx5_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error);
 int mlx5_flow_query(struct rte_eth_dev *dev, struct rte_flow *flow,
 		    enum rte_flow_action_type action, void *data,
@@ -302,19 +283,32 @@ int mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
 			 enum rte_filter_type filter_type,
 			 enum rte_filter_op filter_op,
 			 void *arg);
+int mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list);
+void mlx5_flow_stop(struct rte_eth_dev *dev, struct mlx5_flows *list);
+int mlx5_flow_verify(struct rte_eth_dev *dev);
+int mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
+			struct rte_flow_item_eth *eth_spec,
+			struct rte_flow_item_eth *eth_mask,
+			struct rte_flow_item_vlan *vlan_spec,
+			struct rte_flow_item_vlan *vlan_mask);
+int mlx5_ctrl_flow(struct rte_eth_dev *dev,
+		   struct rte_flow_item_eth *eth_spec,
+		   struct rte_flow_item_eth *eth_mask);
+int mlx5_flow_create_drop_queue(struct rte_eth_dev *dev);
+void mlx5_flow_delete_drop_queue(struct rte_eth_dev *dev);
 
 /* mlx5_socket.c */
 
-int priv_socket_init(struct priv *priv);
-int priv_socket_uninit(struct priv *priv);
-void priv_socket_handle(struct priv *priv);
-int priv_socket_connect(struct priv *priv);
+int mlx5_socket_init(struct rte_eth_dev *priv);
+int mlx5_socket_uninit(struct rte_eth_dev *priv);
+void mlx5_socket_handle(struct rte_eth_dev *priv);
+int mlx5_socket_connect(struct rte_eth_dev *priv);
 
 /* mlx5_mr.c */
 
-struct mlx5_mr *priv_mr_new(struct priv *priv, struct rte_mempool *mp);
-struct mlx5_mr *priv_mr_get(struct priv *priv, struct rte_mempool *mp);
-int priv_mr_release(struct priv *priv, struct mlx5_mr *mr);
-int priv_mr_verify(struct priv *priv);
+struct mlx5_mr *mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp);
+struct mlx5_mr *mlx5_mr_get(struct rte_eth_dev *dev, struct rte_mempool *mp);
+int mlx5_mr_release(struct mlx5_mr *mr);
+int mlx5_mr_verify(struct rte_eth_dev *dev);
 
 #endif /* RTE_PMD_MLX5_H_ */
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 650b6287c..3435bf338 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -122,8 +122,8 @@ struct ethtool_link_settings {
 /**
  * Get interface name from private structure.
  *
- * @param[in] priv
- *   Pointer to private structure.
+ * @param[in] dev
+ *   Pointer to Ethernet device.
  * @param[out] ifname
  *   Interface name output buffer.
  *
@@ -131,8 +131,9 @@ struct ethtool_link_settings {
  *   0 on success, -1 on failure and errno is set.
  */
 int
-priv_get_ifname(const struct priv *priv, char (*ifname)[IF_NAMESIZE])
+mlx5_get_ifname(const struct rte_eth_dev *dev, char (*ifname)[IF_NAMESIZE])
 {
+	struct priv *priv = dev->data->dev_private;
 	DIR *dir;
 	struct dirent *dent;
 	unsigned int dev_type = 0;
@@ -203,8 +204,8 @@ priv_get_ifname(const struct priv *priv, char (*ifname)[IF_NAMESIZE])
 /**
  * Perform ifreq ioctl() on associated Ethernet device.
  *
- * @param[in] priv
- *   Pointer to private structure.
+ * @param[in] dev
+ *   Pointer to Ethernet device.
  * @param req
  *   Request number to pass to ioctl().
  * @param[out] ifr
@@ -214,14 +215,14 @@ priv_get_ifname(const struct priv *priv, char (*ifname)[IF_NAMESIZE])
  *   0 on success, -1 on failure and errno is set.
  */
 int
-priv_ifreq(const struct priv *priv, int req, struct ifreq *ifr)
+mlx5_ifreq(const struct rte_eth_dev *dev, int req, struct ifreq *ifr)
 {
 	int sock = socket(PF_INET, SOCK_DGRAM, IPPROTO_IP);
 	int ret = -1;
 
 	if (sock == -1)
 		return ret;
-	if (priv_get_ifname(priv, &ifr->ifr_name) == 0)
+	if (mlx5_get_ifname(dev, &ifr->ifr_name) == 0)
 		ret = ioctl(sock, req, ifr);
 	close(sock);
 	return ret;
@@ -230,8 +231,8 @@ priv_ifreq(const struct priv *priv, int req, struct ifreq *ifr)
 /**
  * Get device MTU.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param[out] mtu
  *   MTU value output buffer.
  *
@@ -239,10 +240,10 @@ priv_ifreq(const struct priv *priv, int req, struct ifreq *ifr)
  *   0 on success, -1 on failure and errno is set.
  */
 int
-priv_get_mtu(struct priv *priv, uint16_t *mtu)
+mlx5_get_mtu(struct rte_eth_dev *dev, uint16_t *mtu)
 {
 	struct ifreq request;
-	int ret = priv_ifreq(priv, SIOCGIFMTU, &request);
+	int ret = mlx5_ifreq(dev, SIOCGIFMTU, &request);
 
 	if (ret)
 		return ret;
@@ -253,8 +254,8 @@ priv_get_mtu(struct priv *priv, uint16_t *mtu)
 /**
  * Set device MTU.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param mtu
  *   MTU value to set.
  *
@@ -262,18 +263,18 @@ priv_get_mtu(struct priv *priv, uint16_t *mtu)
  *   0 on success, -1 on failure and errno is set.
  */
 static int
-priv_set_mtu(struct priv *priv, uint16_t mtu)
+mlx5_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 {
 	struct ifreq request = { .ifr_mtu = mtu, };
 
-	return priv_ifreq(priv, SIOCSIFMTU, &request);
+	return mlx5_ifreq(dev, SIOCSIFMTU, &request);
 }
 
 /**
  * Set device flags.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param keep
  *   Bitmask for flags that must remain untouched.
  * @param flags
@@ -283,16 +284,16 @@ priv_set_mtu(struct priv *priv, uint16_t mtu)
  *   0 on success, -1 on failure and errno is set.
  */
 int
-priv_set_flags(struct priv *priv, unsigned int keep, unsigned int flags)
+mlx5_set_flags(struct rte_eth_dev *dev, unsigned int keep, unsigned int flags)
 {
 	struct ifreq request;
-	int ret = priv_ifreq(priv, SIOCGIFFLAGS, &request);
+	int ret = mlx5_ifreq(dev, SIOCGIFFLAGS, &request);
 
 	if (ret)
 		return ret;
 	request.ifr_flags &= keep;
 	request.ifr_flags |= flags & ~keep;
-	return priv_ifreq(priv, SIOCSIFFLAGS, &request);
+	return mlx5_ifreq(dev, SIOCSIFFLAGS, &request);
 }
 
 /**
@@ -358,7 +359,7 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 	reta_idx_n = (1 << log2above((rxqs_n & (rxqs_n - 1)) ?
 				     priv->ind_table_max_size :
 				     rxqs_n));
-	if (priv_rss_reta_index_resize(priv, reta_idx_n))
+	if (mlx5_rss_reta_index_resize(dev, reta_idx_n))
 		return ENOMEM;
 	/* When the number of RX queues is not a power of two, the remaining
 	 * table entries are padded with reused WQs and hashes are not spread
@@ -425,7 +426,7 @@ mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info)
 		info->tx_offload_capa |= (DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM |
 					  DEV_TX_OFFLOAD_VXLAN_TNL_TSO |
 					  DEV_TX_OFFLOAD_GRE_TNL_TSO);
-	if (priv_get_ifname(priv, &ifname) == 0)
+	if (mlx5_get_ifname(dev, &ifname) == 0)
 		info->if_index = if_nametoindex(ifname);
 	info->reta_size = priv->reta_idx_n ?
 		priv->reta_idx_n : priv->ind_table_max_size;
@@ -490,7 +491,7 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 	struct rte_eth_link dev_link;
 	int link_speed = 0;
 
-	if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) {
+	if (mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr)) {
 		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
 		return -1;
 	}
@@ -498,7 +499,7 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 	dev_link.link_status = ((ifr.ifr_flags & IFF_UP) &&
 				(ifr.ifr_flags & IFF_RUNNING));
 	ifr.ifr_data = (void *)&edata;
-	if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
+	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
 		WARN("ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: %s",
 		     strerror(errno));
 		return -1;
@@ -552,7 +553,7 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 	struct rte_eth_link dev_link;
 	uint64_t sc;
 
-	if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) {
+	if (mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr)) {
 		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
 		return -1;
 	}
@@ -560,7 +561,7 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 	dev_link.link_status = ((ifr.ifr_flags & IFF_UP) &&
 				(ifr.ifr_flags & IFF_RUNNING));
 	ifr.ifr_data = (void *)&gcmd;
-	if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
+	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
 		DEBUG("ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS) failed: %s",
 		      strerror(errno));
 		return -1;
@@ -574,7 +575,7 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 
 	*ecmd = gcmd;
 	ifr.ifr_data = (void *)ecmd;
-	if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
+	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
 		DEBUG("ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS) failed: %s",
 		      strerror(errno));
 		return -1;
@@ -633,90 +634,50 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 /**
  * Enable receiving and transmitting traffic.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  */
 static void
-priv_link_start(struct priv *priv)
+mlx5_link_start(struct rte_eth_dev *dev)
 {
-	struct rte_eth_dev *dev = priv->dev;
+	struct priv *priv = dev->data->dev_private;
 	int err;
 
-	priv_dev_select_tx_function(priv, dev);
-	priv_dev_select_rx_function(priv, dev);
-	err = priv_dev_traffic_enable(priv, dev);
+	mlx5_select_tx_function(dev);
+	mlx5_select_rx_function(dev);
+	err = mlx5_traffic_enable(dev);
 	if (err)
 		ERROR("%p: error occurred while configuring control flows: %s",
-		      (void *)priv, strerror(err));
-	err = priv_flow_start(priv, &priv->flows);
+		      (void *)dev, strerror(err));
+	err = mlx5_flow_start(dev, &priv->flows);
 	if (err)
 		ERROR("%p: error occurred while configuring flows: %s",
-		      (void *)priv, strerror(err));
+		      (void *)dev, strerror(err));
 }
 
 /**
  * Disable receiving and transmitting traffic.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  */
 static void
-priv_link_stop(struct priv *priv)
+mlx5_link_stop(struct rte_eth_dev *dev)
 {
-	struct rte_eth_dev *dev = priv->dev;
+	struct priv *priv = dev->data->dev_private;
 
-	priv_flow_stop(priv, &priv->flows);
-	priv_dev_traffic_disable(priv, dev);
+	mlx5_flow_stop(dev, &priv->flows);
+	mlx5_traffic_disable(dev);
 	dev->rx_pkt_burst = removed_rx_burst;
 	dev->tx_pkt_burst = removed_tx_burst;
 }
 
 /**
- * Retrieve physical link information and update rx/tx_pkt_burst callbacks
- * accordingly.
- *
- * @param priv
- *   Pointer to private structure.
- * @param wait_to_complete
- *   Wait for request completion (ignored).
- */
-int
-priv_link_update(struct priv *priv, int wait_to_complete __rte_unused)
-{
-	struct rte_eth_dev *dev = priv->dev;
-	struct utsname utsname;
-	int ver[3];
-	int ret;
-	struct rte_eth_link dev_link = dev->data->dev_link;
-
-	if (uname(&utsname) == -1 ||
-	    sscanf(utsname.release, "%d.%d.%d",
-		   &ver[0], &ver[1], &ver[2]) != 3 ||
-	    KERNEL_VERSION(ver[0], ver[1], ver[2]) < KERNEL_VERSION(4, 9, 0))
-		ret = mlx5_link_update_unlocked_gset(dev);
-	else
-		ret = mlx5_link_update_unlocked_gs(dev);
-	/* If lsc interrupt is disabled, should always be ready for traffic. */
-	if (!dev->data->dev_conf.intr_conf.lsc) {
-		priv_link_start(priv);
-		return ret;
-	}
-	/* Re-select burst callbacks only if link status has been changed. */
-	if (!ret && dev_link.link_status != dev->data->dev_link.link_status) {
-		if (dev->data->dev_link.link_status == ETH_LINK_UP)
-			priv_link_start(priv);
-		else
-			priv_link_stop(priv);
-	}
-	return ret;
-}
-
-/**
  * Querying the link status till it changes to the desired state.
  * Number of query attempts is bounded by MLX5_MAX_LINK_QUERY_ATTEMPTS.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param status
  *   Link desired status.
  *
@@ -724,13 +685,13 @@ priv_link_update(struct priv *priv, int wait_to_complete __rte_unused)
  *   0 on success, negative errno value on failure.
  */
 int
-priv_force_link_status_change(struct priv *priv, int status)
+mlx5_force_link_status_change(struct rte_eth_dev *dev, int status)
 {
 	int try = 0;
 
 	while (try < MLX5_MAX_LINK_QUERY_ATTEMPTS) {
-		priv_link_update(priv, 0);
-		if (priv->dev->data->dev_link.link_status == status)
+		mlx5_link_update(dev, 0);
+		if (dev->data->dev_link.link_status == status)
 			return 0;
 		try++;
 		sleep(1);
@@ -752,10 +713,30 @@ priv_force_link_status_change(struct priv *priv, int status)
 int
 mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
 {
-	struct priv *priv = dev->data->dev_private;
+	struct utsname utsname;
+	int ver[3];
 	int ret;
+	struct rte_eth_link dev_link = dev->data->dev_link;
 
-	ret = priv_link_update(priv, wait_to_complete);
+	if (uname(&utsname) == -1 ||
+	    sscanf(utsname.release, "%d.%d.%d",
+		   &ver[0], &ver[1], &ver[2]) != 3 ||
+	    KERNEL_VERSION(ver[0], ver[1], ver[2]) < KERNEL_VERSION(4, 9, 0))
+		ret = mlx5_link_update_unlocked_gset(dev);
+	else
+		ret = mlx5_link_update_unlocked_gs(dev);
+	/* If lsc interrupt is disabled, should always be ready for traffic. */
+	if (!dev->data->dev_conf.intr_conf.lsc) {
+		mlx5_link_start(dev);
+		return ret;
+	}
+	/* Re-select burst callbacks only if link status has been changed. */
+	if (!ret && dev_link.link_status != dev->data->dev_link.link_status) {
+		if (dev->data->dev_link.link_status == ETH_LINK_UP)
+			mlx5_link_start(dev);
+		else
+			mlx5_link_stop(dev);
+	}
 	return ret;
 }
 
@@ -777,14 +758,14 @@ mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 	uint16_t kern_mtu;
 	int ret = 0;
 
-	ret = priv_get_mtu(priv, &kern_mtu);
+	ret = mlx5_get_mtu(dev, &kern_mtu);
 	if (ret)
 		goto out;
 	/* Set kernel interface MTU first. */
-	ret = priv_set_mtu(priv, mtu);
+	ret = mlx5_set_mtu(dev, mtu);
 	if (ret)
 		goto out;
-	ret = priv_get_mtu(priv, &kern_mtu);
+	ret = mlx5_get_mtu(dev, &kern_mtu);
 	if (ret)
 		goto out;
 	if (kern_mtu == mtu) {
@@ -814,7 +795,6 @@ mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 int
 mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 {
-	struct priv *priv = dev->data->dev_private;
 	struct ifreq ifr;
 	struct ethtool_pauseparam ethpause = {
 		.cmd = ETHTOOL_GPAUSEPARAM
@@ -822,10 +802,9 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 	int ret;
 
 	ifr.ifr_data = (void *)&ethpause;
-	if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
+	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
 		ret = errno;
-		WARN("ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM)"
-		     " failed: %s",
+		WARN("ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM) failed: %s",
 		     strerror(ret));
 		goto out;
 	}
@@ -858,7 +837,6 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 int
 mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 {
-	struct priv *priv = dev->data->dev_private;
 	struct ifreq ifr;
 	struct ethtool_pauseparam ethpause = {
 		.cmd = ETHTOOL_SPAUSEPARAM
@@ -878,7 +856,7 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		ethpause.tx_pause = 1;
 	else
 		ethpause.tx_pause = 0;
-	if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
+	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
 		ret = errno;
 		WARN("ioctl(SIOCETHTOOL, ETHTOOL_SPAUSEPARAM)"
 		     " failed: %s",
@@ -944,18 +922,19 @@ mlx5_ibv_device_to_pci_addr(const struct ibv_device *device,
 /**
  * Update the link status.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   Zero if the callback process can be called immediately.
  */
 static int
-priv_link_status_update(struct priv *priv)
+mlx5_link_status_update(struct rte_eth_dev *dev)
 {
-	struct rte_eth_link *link = &priv->dev->data->dev_link;
+	struct priv *priv = dev->data->dev_private;
+	struct rte_eth_link *link = &dev->data->dev_link;
 
-	priv_link_update(priv, 0);
+	mlx5_link_update(dev, 0);
 	if (((link->link_speed == 0) && link->link_status) ||
 		((link->link_speed != 0) && !link->link_status)) {
 		/*
@@ -980,8 +959,8 @@ priv_link_status_update(struct priv *priv)
 /**
  * Device status handler.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param events
  *   Pointer to event flags holder.
  *
@@ -989,8 +968,9 @@ priv_link_status_update(struct priv *priv)
  *   Events bitmap of callback process which can be called immediately.
  */
 static uint32_t
-priv_dev_status_handler(struct priv *priv)
+mlx5_dev_status_handler(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct ibv_async_event event;
 	uint32_t ret = 0;
 
@@ -1000,10 +980,10 @@ priv_dev_status_handler(struct priv *priv)
 			break;
 		if ((event.event_type == IBV_EVENT_PORT_ACTIVE ||
 			event.event_type == IBV_EVENT_PORT_ERR) &&
-			(priv->dev->data->dev_conf.intr_conf.lsc == 1))
+			(dev->data->dev_conf.intr_conf.lsc == 1))
 			ret |= (1 << RTE_ETH_EVENT_INTR_LSC);
 		else if (event.event_type == IBV_EVENT_DEVICE_FATAL &&
-			priv->dev->data->dev_conf.intr_conf.rmv == 1)
+			dev->data->dev_conf.intr_conf.rmv == 1)
 			ret |= (1 << RTE_ETH_EVENT_INTR_RMV);
 		else
 			DEBUG("event type %d on port %d not handled",
@@ -1011,7 +991,7 @@ priv_dev_status_handler(struct priv *priv)
 		ibv_ack_async_event(&event);
 	}
 	if (ret & (1 << RTE_ETH_EVENT_INTR_LSC))
-		if (priv_link_status_update(priv))
+		if (mlx5_link_status_update(dev))
 			ret &= ~(1 << RTE_ETH_EVENT_INTR_LSC);
 	return ret;
 }
@@ -1030,7 +1010,7 @@ mlx5_dev_link_status_handler(void *arg)
 	int ret;
 
 	priv->pending_alarm = 0;
-	ret = priv_link_status_update(priv);
+	ret = mlx5_link_status_update(dev);
 	if (!ret)
 		_rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL,
 					      NULL);
@@ -1048,10 +1028,9 @@ void
 mlx5_dev_interrupt_handler(void *cb_arg)
 {
 	struct rte_eth_dev *dev = cb_arg;
-	struct priv *priv = dev->data->dev_private;
 	uint32_t events;
 
-	events = priv_dev_status_handler(priv);
+	events = mlx5_dev_status_handler(dev);
 	if (events & (1 << RTE_ETH_EVENT_INTR_LSC))
 		_rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL,
 					      NULL);
@@ -1070,22 +1049,21 @@ static void
 mlx5_dev_handler_socket(void *cb_arg)
 {
 	struct rte_eth_dev *dev = cb_arg;
-	struct priv *priv = dev->data->dev_private;
 
-	priv_socket_handle(priv);
+	mlx5_socket_handle(dev);
 }
 
 /**
  * Uninstall interrupt handler.
  *
- * @param priv
- *   Pointer to private structure.
  * @param dev
- *   Pointer to the rte_eth_dev structure.
+ *   Pointer to Ethernet device.
  */
 void
-priv_dev_interrupt_handler_uninstall(struct priv *priv, struct rte_eth_dev *dev)
+mlx5_dev_interrupt_handler_uninstall(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
+
 	if (dev->data->dev_conf.intr_conf.lsc ||
 	    dev->data->dev_conf.intr_conf.rmv)
 		rte_intr_callback_unregister(&priv->intr_handle,
@@ -1106,14 +1084,13 @@ priv_dev_interrupt_handler_uninstall(struct priv *priv, struct rte_eth_dev *dev)
 /**
  * Install interrupt handler.
  *
- * @param priv
- *   Pointer to private structure.
  * @param dev
- *   Pointer to the rte_eth_dev structure.
+ *   Pointer to Ethernet device.
  */
 void
-priv_dev_interrupt_handler_install(struct priv *priv, struct rte_eth_dev *dev)
+mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	int rc, flags;
 
 	assert(priv->ctx->async_fd > 0);
@@ -1131,7 +1108,7 @@ priv_dev_interrupt_handler_install(struct priv *priv, struct rte_eth_dev *dev)
 		rte_intr_callback_register(&priv->intr_handle,
 					   mlx5_dev_interrupt_handler, dev);
 	}
-	rc = priv_socket_init(priv);
+	rc = mlx5_socket_init(dev);
 	if (!rc && priv->primary_socket) {
 		priv->intr_handle_socket.fd = priv->primary_socket;
 		priv->intr_handle_socket.type = RTE_INTR_HANDLE_EXT;
@@ -1141,23 +1118,6 @@ priv_dev_interrupt_handler_install(struct priv *priv, struct rte_eth_dev *dev)
 }
 
 /**
- * Change the link state (UP / DOWN).
- *
- * @param priv
- *   Pointer to private data structure.
- * @param up
- *   Nonzero for link up, otherwise link down.
- *
- * @return
- *   0 on success, errno value on failure.
- */
-static int
-priv_dev_set_link(struct priv *priv, int up)
-{
-	return priv_set_flags(priv, ~IFF_UP, up ? IFF_UP : ~IFF_UP);
-}
-
-/**
  * DPDK callback to bring the link DOWN.
  *
  * @param dev
@@ -1169,11 +1129,7 @@ priv_dev_set_link(struct priv *priv, int up)
 int
 mlx5_set_link_down(struct rte_eth_dev *dev)
 {
-	struct priv *priv = dev->data->dev_private;
-	int err;
-
-	err = priv_dev_set_link(priv, 0);
-	return err;
+	return mlx5_set_flags(dev, ~IFF_UP, ~IFF_UP);
 }
 
 /**
@@ -1188,31 +1144,25 @@ mlx5_set_link_down(struct rte_eth_dev *dev)
 int
 mlx5_set_link_up(struct rte_eth_dev *dev)
 {
-	struct priv *priv = dev->data->dev_private;
-	int err;
-
-	err = priv_dev_set_link(priv, 1);
-	return err;
+	return mlx5_set_flags(dev, ~IFF_UP, IFF_UP);
 }
 
 /**
  * Configure the TX function to use.
  *
- * @param priv
- *   Pointer to private data structure.
  * @param dev
  *   Pointer to rte_eth_dev structure.
  */
 void
-priv_dev_select_tx_function(struct priv *priv, struct rte_eth_dev *dev)
+mlx5_select_tx_function(struct rte_eth_dev *dev)
 {
-	assert(priv != NULL);
-	assert(dev != NULL);
+	struct priv *priv = dev->data->dev_private;
+
 	dev->tx_pkt_burst = mlx5_tx_burst;
 	/* Select appropriate TX function. */
 	if (priv->mps == MLX5_MPW_ENHANCED) {
-		if (priv_check_vec_tx_support(priv) > 0) {
-			if (priv_check_raw_vec_tx_support(priv) > 0)
+		if (mlx5_check_vec_tx_support(dev) > 0) {
+			if (mlx5_check_raw_vec_tx_support(dev) > 0)
 				dev->tx_pkt_burst = mlx5_tx_burst_raw_vec;
 			else
 				dev->tx_pkt_burst = mlx5_tx_burst_vec;
@@ -1233,17 +1183,14 @@ priv_dev_select_tx_function(struct priv *priv, struct rte_eth_dev *dev)
 /**
  * Configure the RX function to use.
  *
- * @param priv
- *   Pointer to private data structure.
  * @param dev
  *   Pointer to rte_eth_dev structure.
  */
 void
-priv_dev_select_rx_function(struct priv *priv, struct rte_eth_dev *dev)
+mlx5_select_rx_function(struct rte_eth_dev *dev)
 {
-	assert(priv != NULL);
 	assert(dev != NULL);
-	if (priv_check_vec_rx_support(priv) > 0) {
+	if (mlx5_check_vec_rx_support(dev) > 0) {
 		dev->rx_pkt_burst = mlx5_rx_burst_vec;
 		DEBUG("selected RX vectorized function");
 	} else {
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index d51f0ece5..aaa8727ee 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -128,7 +128,7 @@ static int
 mlx5_flow_create_flag_mark(struct mlx5_flow_parse *parser, uint32_t mark_id);
 
 static int
-mlx5_flow_create_count(struct priv *priv, struct mlx5_flow_parse *parser);
+mlx5_flow_create_count(struct rte_eth_dev *dev, struct mlx5_flow_parse *parser);
 
 /* Hash RX queue types. */
 enum hash_rxq_type {
@@ -554,8 +554,6 @@ mlx5_flow_item_validate(const struct rte_flow_item *item,
  * Copy the RSS configuration from the user ones, of the rss_conf is null,
  * uses the driver one.
  *
- * @param priv
- *   Pointer to private structure.
  * @param parser
  *   Internal parser structure.
  * @param rss_conf
@@ -565,13 +563,12 @@ mlx5_flow_item_validate(const struct rte_flow_item *item,
  *   0 on success, errno value on failure.
  */
 static int
-priv_flow_convert_rss_conf(struct priv *priv __rte_unused,
-			   struct mlx5_flow_parse *parser,
+mlx5_flow_convert_rss_conf(struct mlx5_flow_parse *parser,
 			   const struct rte_eth_rss_conf *rss_conf)
 {
 	/*
 	 * This function is also called at the beginning of
-	 * priv_flow_convert_actions() to initialize the parser with the
+	 * mlx5_flow_convert_actions() to initialize the parser with the
 	 * device default RSS configuration.
 	 */
 	if (rss_conf) {
@@ -593,23 +590,17 @@ priv_flow_convert_rss_conf(struct priv *priv __rte_unused,
 /**
  * Extract attribute to the parser.
  *
- * @param priv
- *   Pointer to private structure.
  * @param[in] attr
  *   Flow rule attributes.
  * @param[out] error
  *   Perform verbose error reporting if not NULL.
- * @param[in, out] parser
- *   Internal parser structure.
  *
  * @return
  *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_convert_attributes(struct priv *priv __rte_unused,
-			     const struct rte_flow_attr *attr,
-			     struct rte_flow_error *error,
-			     struct mlx5_flow_parse *parser __rte_unused)
+mlx5_flow_convert_attributes(const struct rte_flow_attr *attr,
+			     struct rte_flow_error *error)
 {
 	if (attr->group) {
 		rte_flow_error_set(error, ENOTSUP,
@@ -645,8 +636,8 @@ priv_flow_convert_attributes(struct priv *priv __rte_unused,
 /**
  * Extract actions request to the parser.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param[in] actions
  *   Associated actions (list terminated by the END action).
  * @param[out] error
@@ -658,16 +649,18 @@ priv_flow_convert_attributes(struct priv *priv __rte_unused,
  *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_convert_actions(struct priv *priv,
+mlx5_flow_convert_actions(struct rte_eth_dev *dev,
 			  const struct rte_flow_action actions[],
 			  struct rte_flow_error *error,
 			  struct mlx5_flow_parse *parser)
 {
+	struct priv *priv = dev->data->dev_private;
+
 	/*
 	 * Add default RSS configuration necessary for Verbs to create QP even
 	 * if no RSS is necessary.
 	 */
-	priv_flow_convert_rss_conf(priv, parser,
+	mlx5_flow_convert_rss_conf(parser,
 				   (const struct rte_eth_rss_conf *)
 				   &priv->rss_conf);
 	for (; actions->type != RTE_FLOW_ACTION_TYPE_END; ++actions) {
@@ -747,8 +740,7 @@ priv_flow_convert_actions(struct priv *priv,
 			for (n = 0; n < rss->num; ++n)
 				parser->queues[n] = rss->queue[n];
 			parser->queues_n = rss->num;
-			if (priv_flow_convert_rss_conf(priv, parser,
-						       rss->rss_conf)) {
+			if (mlx5_flow_convert_rss_conf(parser, rss->rss_conf)) {
 				rte_flow_error_set(error, EINVAL,
 						   RTE_FLOW_ERROR_TYPE_ACTION,
 						   actions,
@@ -802,8 +794,6 @@ priv_flow_convert_actions(struct priv *priv,
 /**
  * Validate items.
  *
- * @param priv
- *   Pointer to private structure.
  * @param[in] items
  *   Pattern specification (list terminated by the END pattern item).
  * @param[out] error
@@ -815,8 +805,7 @@ priv_flow_convert_actions(struct priv *priv,
  *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_convert_items_validate(struct priv *priv __rte_unused,
-				 const struct rte_flow_item items[],
+mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
 				 struct rte_flow_error *error,
 				 struct mlx5_flow_parse *parser)
 {
@@ -893,8 +882,6 @@ priv_flow_convert_items_validate(struct priv *priv __rte_unused,
 /**
  * Allocate memory space to store verbs flow attributes.
  *
- * @param priv
- *   Pointer to private structure.
  * @param[in] priority
  *   Flow priority.
  * @param[in] size
@@ -906,8 +893,7 @@ priv_flow_convert_items_validate(struct priv *priv __rte_unused,
  *   A verbs flow attribute on success, NULL otherwise.
  */
 static struct ibv_flow_attr *
-priv_flow_convert_allocate(struct priv *priv __rte_unused,
-			   unsigned int priority,
+mlx5_flow_convert_allocate(unsigned int priority,
 			   unsigned int size,
 			   struct rte_flow_error *error)
 {
@@ -928,14 +914,11 @@ priv_flow_convert_allocate(struct priv *priv __rte_unused,
 /**
  * Finalise verbs flow attributes.
  *
- * @param priv
- *   Pointer to private structure.
  * @param[in, out] parser
  *   Internal parser structure.
  */
 static void
-priv_flow_convert_finalise(struct priv *priv __rte_unused,
-			   struct mlx5_flow_parse *parser)
+mlx5_flow_convert_finalise(struct mlx5_flow_parse *parser)
 {
 	const unsigned int ipv4 =
 		hash_rxq_init[parser->layer].ip_version == MLX5_IPV4;
@@ -1054,8 +1037,8 @@ priv_flow_convert_finalise(struct priv *priv __rte_unused,
 /**
  * Validate and convert a flow supported by the NIC.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param[in] attr
  *   Flow rule attributes.
  * @param[in] pattern
@@ -1071,7 +1054,7 @@ priv_flow_convert_finalise(struct priv *priv __rte_unused,
  *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_convert(struct priv *priv,
+mlx5_flow_convert(struct rte_eth_dev *dev,
 		  const struct rte_flow_attr *attr,
 		  const struct rte_flow_item items[],
 		  const struct rte_flow_action actions[],
@@ -1088,24 +1071,24 @@ priv_flow_convert(struct priv *priv,
 		.layer = HASH_RXQ_ETH,
 		.mark_id = MLX5_FLOW_MARK_DEFAULT,
 	};
-	ret = priv_flow_convert_attributes(priv, attr, error, parser);
+	ret = mlx5_flow_convert_attributes(attr, error);
 	if (ret)
 		return ret;
-	ret = priv_flow_convert_actions(priv, actions, error, parser);
+	ret = mlx5_flow_convert_actions(dev, actions, error, parser);
 	if (ret)
 		return ret;
-	ret = priv_flow_convert_items_validate(priv, items, error, parser);
+	ret = mlx5_flow_convert_items_validate(items, error, parser);
 	if (ret)
 		return ret;
-	priv_flow_convert_finalise(priv, parser);
+	mlx5_flow_convert_finalise(parser);
 	/*
 	 * Second step.
 	 * Allocate the memory space to store verbs specifications.
 	 */
 	if (parser->drop) {
 		parser->queue[HASH_RXQ_ETH].ibv_attr =
-			priv_flow_convert_allocate
-			(priv, attr->priority,
+			mlx5_flow_convert_allocate
+			(attr->priority,
 			 parser->queue[HASH_RXQ_ETH].offset,
 			 error);
 		if (!parser->queue[HASH_RXQ_ETH].ibv_attr)
@@ -1125,7 +1108,7 @@ priv_flow_convert(struct priv *priv,
 				continue;
 			offset = parser->queue[i].offset;
 			parser->queue[i].ibv_attr =
-				priv_flow_convert_allocate(priv, priority,
+				mlx5_flow_convert_allocate(priority,
 							   offset, error);
 			if (!parser->queue[i].ibv_attr)
 				goto exit_enomem;
@@ -1153,7 +1136,7 @@ priv_flow_convert(struct priv *priv,
 	if (parser->mark)
 		mlx5_flow_create_flag_mark(parser, parser->mark_id);
 	if (parser->count && parser->create) {
-		mlx5_flow_create_count(priv, parser);
+		mlx5_flow_create_count(dev, parser);
 		if (!parser->cs)
 			goto exit_count_error;
 	}
@@ -1162,7 +1145,7 @@ priv_flow_convert(struct priv *priv,
 	 * configuration.
 	 */
 	if (!parser->drop) {
-		priv_flow_convert_finalise(priv, parser);
+		mlx5_flow_convert_finalise(parser);
 	} else {
 		parser->queue[HASH_RXQ_ETH].ibv_attr->priority =
 			attr->priority +
@@ -1598,8 +1581,8 @@ mlx5_flow_create_flag_mark(struct mlx5_flow_parse *parser, uint32_t mark_id)
 /**
  * Convert count action to Verbs specification.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param parser
  *   Pointer to MLX5 flow parser structure.
  *
@@ -1607,10 +1590,11 @@ mlx5_flow_create_flag_mark(struct mlx5_flow_parse *parser, uint32_t mark_id)
  *   0 on success, errno value on failure.
  */
 static int
-mlx5_flow_create_count(struct priv *priv __rte_unused,
+mlx5_flow_create_count(struct rte_eth_dev *dev __rte_unused,
 		       struct mlx5_flow_parse *parser __rte_unused)
 {
 #ifdef HAVE_IBV_DEVICE_COUNTERS_SET_SUPPORT
+	struct priv *priv = dev->data->dev_private;
 	unsigned int size = sizeof(struct ibv_flow_spec_counter_action);
 	struct ibv_counter_set_init_attr init_attr = {0};
 	struct ibv_flow_spec_counter_action counter = {
@@ -1632,8 +1616,8 @@ mlx5_flow_create_count(struct priv *priv __rte_unused,
 /**
  * Complete flow rule creation with a drop queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param parser
  *   Internal parser structure.
  * @param flow
@@ -1645,11 +1629,12 @@ mlx5_flow_create_count(struct priv *priv __rte_unused,
  *   0 on success, errno value on failure.
  */
 static int
-priv_flow_create_action_queue_drop(struct priv *priv,
+mlx5_flow_create_action_queue_drop(struct rte_eth_dev *dev,
 				   struct mlx5_flow_parse *parser,
 				   struct rte_flow *flow,
 				   struct rte_flow_error *error)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct ibv_flow_spec_action_drop *drop;
 	unsigned int size = sizeof(struct ibv_flow_spec_action_drop);
 	int err = 0;
@@ -1703,8 +1688,8 @@ priv_flow_create_action_queue_drop(struct priv *priv,
 /**
  * Create hash Rx queues when RSS is enabled.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param parser
  *   Internal parser structure.
  * @param flow
@@ -1716,11 +1701,12 @@ priv_flow_create_action_queue_drop(struct priv *priv,
  *   0 on success, a errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_create_action_queue_rss(struct priv *priv,
+mlx5_flow_create_action_queue_rss(struct rte_eth_dev *dev,
 				  struct mlx5_flow_parse *parser,
 				  struct rte_flow *flow,
 				  struct rte_flow_error *error)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
 	for (i = 0; i != hash_rxq_init_n; ++i) {
@@ -1734,21 +1720,21 @@ priv_flow_create_action_queue_rss(struct priv *priv,
 		if (!priv->dev->data->dev_started)
 			continue;
 		flow->frxq[i].hrxq =
-			mlx5_priv_hrxq_get(priv,
-					   parser->rss_conf.rss_key,
-					   parser->rss_conf.rss_key_len,
-					   hash_fields,
-					   parser->queues,
-					   parser->queues_n);
+			mlx5_hrxq_get(dev,
+				      parser->rss_conf.rss_key,
+				      parser->rss_conf.rss_key_len,
+				      hash_fields,
+				      parser->queues,
+				      parser->queues_n);
 		if (flow->frxq[i].hrxq)
 			continue;
 		flow->frxq[i].hrxq =
-			mlx5_priv_hrxq_new(priv,
-					   parser->rss_conf.rss_key,
-					   parser->rss_conf.rss_key_len,
-					   hash_fields,
-					   parser->queues,
-					   parser->queues_n);
+			mlx5_hrxq_new(dev,
+				      parser->rss_conf.rss_key,
+				      parser->rss_conf.rss_key_len,
+				      hash_fields,
+				      parser->queues,
+				      parser->queues_n);
 		if (!flow->frxq[i].hrxq) {
 			rte_flow_error_set(error, ENOMEM,
 					   RTE_FLOW_ERROR_TYPE_HANDLE,
@@ -1762,8 +1748,8 @@ priv_flow_create_action_queue_rss(struct priv *priv,
 /**
  * Complete flow rule creation.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param parser
  *   Internal parser structure.
  * @param flow
@@ -1775,11 +1761,12 @@ priv_flow_create_action_queue_rss(struct priv *priv,
  *   0 on success, a errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_create_action_queue(struct priv *priv,
+mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 			      struct mlx5_flow_parse *parser,
 			      struct rte_flow *flow,
 			      struct rte_flow_error *error)
 {
+	struct priv *priv = dev->data->dev_private;
 	int err = 0;
 	unsigned int i;
 	unsigned int flows_n = 0;
@@ -1787,7 +1774,7 @@ priv_flow_create_action_queue(struct priv *priv,
 	assert(priv->pd);
 	assert(priv->ctx);
 	assert(!parser->drop);
-	err = priv_flow_create_action_queue_rss(priv, parser, flow, error);
+	err = mlx5_flow_create_action_queue_rss(dev, parser, flow, error);
 	if (err)
 		goto error;
 	if (parser->count)
@@ -1834,7 +1821,7 @@ priv_flow_create_action_queue(struct priv *priv,
 			claim_zero(ibv_destroy_flow(ibv_flow));
 		}
 		if (flow->frxq[i].hrxq)
-			mlx5_priv_hrxq_release(priv, flow->frxq[i].hrxq);
+			mlx5_hrxq_release(dev, flow->frxq[i].hrxq);
 		if (flow->frxq[i].ibv_attr)
 			rte_free(flow->frxq[i].ibv_attr);
 	}
@@ -1849,8 +1836,8 @@ priv_flow_create_action_queue(struct priv *priv,
 /**
  * Convert a flow.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param list
  *   Pointer to a TAILQ flow list.
  * @param[in] attr
@@ -1866,19 +1853,19 @@ priv_flow_create_action_queue(struct priv *priv,
  *   A flow on success, NULL otherwise.
  */
 static struct rte_flow *
-priv_flow_create(struct priv *priv,
-		 struct mlx5_flows *list,
-		 const struct rte_flow_attr *attr,
-		 const struct rte_flow_item items[],
-		 const struct rte_flow_action actions[],
-		 struct rte_flow_error *error)
+mlx5_flow_list_create(struct rte_eth_dev *dev,
+		      struct mlx5_flows *list,
+		      const struct rte_flow_attr *attr,
+		      const struct rte_flow_item items[],
+		      const struct rte_flow_action actions[],
+		      struct rte_flow_error *error)
 {
 	struct mlx5_flow_parse parser = { .create = 1, };
 	struct rte_flow *flow = NULL;
 	unsigned int i;
 	int err;
 
-	err = priv_flow_convert(priv, attr, items, actions, error, &parser);
+	err = mlx5_flow_convert(dev, attr, items, actions, error, &parser);
 	if (err)
 		goto exit;
 	flow = rte_calloc(__func__, 1,
@@ -1902,10 +1889,10 @@ priv_flow_create(struct priv *priv,
 	memcpy(flow->rss_key, parser.rss_key, parser.rss_conf.rss_key_len);
 	/* finalise the flow. */
 	if (parser.drop)
-		err = priv_flow_create_action_queue_drop(priv, &parser, flow,
+		err = mlx5_flow_create_action_queue_drop(dev, &parser, flow,
 							 error);
 	else
-		err = priv_flow_create_action_queue(priv, &parser, flow, error);
+		err = mlx5_flow_create_action_queue(dev, &parser, flow, error);
 	if (err)
 		goto exit;
 	TAILQ_INSERT_TAIL(list, flow, next);
@@ -1933,11 +1920,10 @@ mlx5_flow_validate(struct rte_eth_dev *dev,
 		   const struct rte_flow_action actions[],
 		   struct rte_flow_error *error)
 {
-	struct priv *priv = dev->data->dev_private;
 	int ret;
 	struct mlx5_flow_parse parser = { .create = 0, };
 
-	ret = priv_flow_convert(priv, attr, items, actions, error, &parser);
+	ret = mlx5_flow_convert(dev, attr, items, actions, error, &parser);
 	return ret;
 }
 
@@ -1955,28 +1941,26 @@ mlx5_flow_create(struct rte_eth_dev *dev,
 		 struct rte_flow_error *error)
 {
 	struct priv *priv = dev->data->dev_private;
-	struct rte_flow *flow;
 
-	flow = priv_flow_create(priv, &priv->flows, attr, items, actions,
-				error);
-	return flow;
+	return mlx5_flow_list_create(dev, &priv->flows, attr, items, actions,
+				     error);
 }
 
 /**
- * Destroy a flow.
+ * Destroy a flow in a list.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param list
  *   Pointer to a TAILQ flow list.
  * @param[in] flow
  *   Flow to destroy.
  */
 static void
-priv_flow_destroy(struct priv *priv,
-		  struct mlx5_flows *list,
-		  struct rte_flow *flow)
+mlx5_flow_list_destroy(struct rte_eth_dev *dev, struct mlx5_flows *list,
+		       struct rte_flow *flow)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
 	if (flow->drop || !flow->mark)
@@ -2023,7 +2007,7 @@ priv_flow_destroy(struct priv *priv,
 			if (frxq->ibv_flow)
 				claim_zero(ibv_destroy_flow(frxq->ibv_flow));
 			if (frxq->hrxq)
-				mlx5_priv_hrxq_release(priv, frxq->hrxq);
+				mlx5_hrxq_release(dev, frxq->hrxq);
 			if (frxq->ibv_attr)
 				rte_free(frxq->ibv_attr);
 		}
@@ -2040,34 +2024,35 @@ priv_flow_destroy(struct priv *priv,
 /**
  * Destroy all flows.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param list
  *   Pointer to a TAILQ flow list.
  */
 void
-priv_flow_flush(struct priv *priv, struct mlx5_flows *list)
+mlx5_flow_list_flush(struct rte_eth_dev *dev, struct mlx5_flows *list)
 {
 	while (!TAILQ_EMPTY(list)) {
 		struct rte_flow *flow;
 
 		flow = TAILQ_FIRST(list);
-		priv_flow_destroy(priv, list, flow);
+		mlx5_flow_list_destroy(dev, list, flow);
 	}
 }
 
 /**
  * Create drop queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   0 on success.
  */
 int
-priv_flow_create_drop_queue(struct priv *priv)
+mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_hrxq_drop *fdq = NULL;
 
 	assert(priv->pd);
@@ -2145,12 +2130,13 @@ priv_flow_create_drop_queue(struct priv *priv)
 /**
  * Delete drop queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  */
 void
-priv_flow_delete_drop_queue(struct priv *priv)
+mlx5_flow_delete_drop_queue(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_hrxq_drop *fdq = priv->flow_drop_queue;
 
 	if (!fdq)
@@ -2170,14 +2156,15 @@ priv_flow_delete_drop_queue(struct priv *priv)
 /**
  * Remove all flows.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param list
  *   Pointer to a TAILQ flow list.
  */
 void
-priv_flow_stop(struct priv *priv, struct mlx5_flows *list)
+mlx5_flow_stop(struct rte_eth_dev *dev, struct mlx5_flows *list)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct rte_flow *flow;
 
 	TAILQ_FOREACH_REVERSE(flow, list, mlx5_flows, next) {
@@ -2219,7 +2206,7 @@ priv_flow_stop(struct priv *priv, struct mlx5_flows *list)
 				continue;
 			claim_zero(ibv_destroy_flow(flow->frxq[i].ibv_flow));
 			flow->frxq[i].ibv_flow = NULL;
-			mlx5_priv_hrxq_release(priv, flow->frxq[i].hrxq);
+			mlx5_hrxq_release(dev, flow->frxq[i].hrxq);
 			flow->frxq[i].hrxq = NULL;
 		}
 		DEBUG("Flow %p removed", (void *)flow);
@@ -2229,8 +2216,8 @@ priv_flow_stop(struct priv *priv, struct mlx5_flows *list)
 /**
  * Add all flows.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param list
  *   Pointer to a TAILQ flow list.
  *
@@ -2238,8 +2225,9 @@ priv_flow_stop(struct priv *priv, struct mlx5_flows *list)
  *   0 on success, a errno value otherwise and rte_errno is set.
  */
 int
-priv_flow_start(struct priv *priv, struct mlx5_flows *list)
+mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct rte_flow *flow;
 
 	TAILQ_FOREACH(flow, list, next) {
@@ -2264,19 +2252,19 @@ priv_flow_start(struct priv *priv, struct mlx5_flows *list)
 			if (!flow->frxq[i].ibv_attr)
 				continue;
 			flow->frxq[i].hrxq =
-				mlx5_priv_hrxq_get(priv, flow->rss_conf.rss_key,
-						   flow->rss_conf.rss_key_len,
-						   hash_rxq_init[i].hash_fields,
-						   (*flow->queues),
-						   flow->queues_n);
+				mlx5_hrxq_get(dev, flow->rss_conf.rss_key,
+					      flow->rss_conf.rss_key_len,
+					      hash_rxq_init[i].hash_fields,
+					      (*flow->queues),
+					      flow->queues_n);
 			if (flow->frxq[i].hrxq)
 				goto flow_create;
 			flow->frxq[i].hrxq =
-				mlx5_priv_hrxq_new(priv, flow->rss_conf.rss_key,
-						   flow->rss_conf.rss_key_len,
-						   hash_rxq_init[i].hash_fields,
-						   (*flow->queues),
-						   flow->queues_n);
+				mlx5_hrxq_new(dev, flow->rss_conf.rss_key,
+					      flow->rss_conf.rss_key_len,
+					      hash_rxq_init[i].hash_fields,
+					      (*flow->queues),
+					      flow->queues_n);
 			if (!flow->frxq[i].hrxq) {
 				DEBUG("Flow %p cannot be applied",
 				      (void *)flow);
@@ -2306,19 +2294,20 @@ priv_flow_start(struct priv *priv, struct mlx5_flows *list)
 /**
  * Verify the flow list is empty
  *
- * @param priv
- *  Pointer to private structure.
+ * @param dev
+ *  Pointer to Ethernet device.
  *
  * @return the number of flows not released.
  */
 int
-priv_flow_verify(struct priv *priv)
+mlx5_flow_verify(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct rte_flow *flow;
 	int ret = 0;
 
 	TAILQ_FOREACH(flow, &priv->flows, next) {
-		DEBUG("%p: flow %p still referenced", (void *)priv,
+		DEBUG("%p: flow %p still referenced", (void *)dev,
 		      (void *)flow);
 		++ret;
 	}
@@ -2399,8 +2388,8 @@ mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
 	action_rss.local.rss_conf = &priv->rss_conf;
 	action_rss.local.num = priv->reta_idx_n;
 	actions[0].conf = (const void *)&action_rss.rss;
-	flow = priv_flow_create(priv, &priv->ctrl_flows, &attr, items, actions,
-				&error);
+	flow = mlx5_flow_list_create(dev, &priv->ctrl_flows, &attr, items,
+				     actions, &error);
 	if (!flow)
 		return rte_errno;
 	return 0;
@@ -2440,7 +2429,7 @@ mlx5_flow_destroy(struct rte_eth_dev *dev,
 {
 	struct priv *priv = dev->data->dev_private;
 
-	priv_flow_destroy(priv, &priv->flows, flow);
+	mlx5_flow_list_destroy(dev, &priv->flows, flow);
 	return 0;
 }
 
@@ -2456,7 +2445,7 @@ mlx5_flow_flush(struct rte_eth_dev *dev,
 {
 	struct priv *priv = dev->data->dev_private;
 
-	priv_flow_flush(priv, &priv->flows);
+	mlx5_flow_list_flush(dev, &priv->flows);
 	return 0;
 }
 
@@ -2473,7 +2462,7 @@ mlx5_flow_flush(struct rte_eth_dev *dev,
  *   0 on success, a errno value otherwise and rte_errno is set.
  */
 static int
-priv_flow_query_count(struct ibv_counter_set *cs,
+mlx5_flow_query_count(struct ibv_counter_set *cs,
 		      struct mlx5_flow_counter_stats *counter_stats,
 		      struct rte_flow_query_count *query_count,
 		      struct rte_flow_error *error)
@@ -2523,7 +2512,7 @@ mlx5_flow_query(struct rte_eth_dev *dev __rte_unused,
 	int res = EINVAL;
 
 	if (flow->cs) {
-		res = priv_flow_query_count(flow->cs,
+		res = mlx5_flow_query_count(flow->cs,
 					&flow->counter_stats,
 					(struct rte_flow_query_count *)data,
 					error);
@@ -2568,8 +2557,8 @@ mlx5_flow_isolate(struct rte_eth_dev *dev,
 /**
  * Convert a flow director filter to a generic flow.
  *
- * @param priv
- *   Private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param fdir_filter
  *   Flow director filter to add.
  * @param attributes
@@ -2579,10 +2568,11 @@ mlx5_flow_isolate(struct rte_eth_dev *dev,
  *  0 on success, errno value on error.
  */
 static int
-priv_fdir_filter_convert(struct priv *priv,
+mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 			 const struct rte_eth_fdir_filter *fdir_filter,
 			 struct mlx5_fdir *attributes)
 {
+	struct priv *priv = dev->data->dev_private;
 	const struct rte_eth_fdir_input *input = &fdir_filter->input;
 
 	/* Validate queue number. */
@@ -2754,8 +2744,8 @@ priv_fdir_filter_convert(struct priv *priv,
 /**
  * Add new flow director filter and store it in list.
  *
- * @param priv
- *   Private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param fdir_filter
  *   Flow director filter to add.
  *
@@ -2763,9 +2753,10 @@ priv_fdir_filter_convert(struct priv *priv,
  *   0 on success, errno value on failure.
  */
 static int
-priv_fdir_filter_add(struct priv *priv,
+mlx5_fdir_filter_add(struct rte_eth_dev *dev,
 		     const struct rte_eth_fdir_filter *fdir_filter)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_fdir attributes = {
 		.attr.group = 0,
 		.l2_mask = {
@@ -2781,19 +2772,16 @@ priv_fdir_filter_add(struct priv *priv,
 	struct rte_flow *flow;
 	int ret;
 
-	ret = priv_fdir_filter_convert(priv, fdir_filter, &attributes);
+	ret = mlx5_fdir_filter_convert(dev, fdir_filter, &attributes);
 	if (ret)
 		return -ret;
-	ret = priv_flow_convert(priv, &attributes.attr, attributes.items,
+	ret = mlx5_flow_convert(dev, &attributes.attr, attributes.items,
 				attributes.actions, &error, &parser);
 	if (ret)
 		return -ret;
-	flow = priv_flow_create(priv,
-				&priv->flows,
-				&attributes.attr,
-				attributes.items,
-				attributes.actions,
-				&error);
+	flow = mlx5_flow_list_create(dev, &priv->flows, &attributes.attr,
+				     attributes.items, attributes.actions,
+				     &error);
 	if (flow) {
 		DEBUG("FDIR created %p", (void *)flow);
 		return 0;
@@ -2804,8 +2792,8 @@ priv_fdir_filter_add(struct priv *priv,
 /**
  * Delete specific filter.
  *
- * @param priv
- *   Private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param fdir_filter
  *   Filter to be deleted.
  *
@@ -2813,9 +2801,10 @@ priv_fdir_filter_add(struct priv *priv,
  *   0 on success, errno value on failure.
  */
 static int
-priv_fdir_filter_delete(struct priv *priv,
+mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
 			const struct rte_eth_fdir_filter *fdir_filter)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_fdir attributes = {
 		.attr.group = 0,
 	};
@@ -2828,10 +2817,10 @@ priv_fdir_filter_delete(struct priv *priv,
 	unsigned int i;
 	int ret;
 
-	ret = priv_fdir_filter_convert(priv, fdir_filter, &attributes);
+	ret = mlx5_fdir_filter_convert(dev, fdir_filter, &attributes);
 	if (ret)
 		return -ret;
-	ret = priv_flow_convert(priv, &attributes.attr, attributes.items,
+	ret = mlx5_flow_convert(dev, &attributes.attr, attributes.items,
 				attributes.actions, &error, &parser);
 	if (ret)
 		goto exit;
@@ -2889,7 +2878,7 @@ priv_fdir_filter_delete(struct priv *priv,
 		continue;
 	}
 	if (flow)
-		priv_flow_destroy(priv, &priv->flows, flow);
+		mlx5_flow_list_destroy(dev, &priv->flows, flow);
 exit:
 	for (i = 0; i != hash_rxq_init_n; ++i) {
 		if (parser.queue[i].ibv_attr)
@@ -2901,8 +2890,8 @@ priv_fdir_filter_delete(struct priv *priv,
 /**
  * Update queue for specific filter.
  *
- * @param priv
- *   Private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param fdir_filter
  *   Filter to be updated.
  *
@@ -2910,41 +2899,44 @@ priv_fdir_filter_delete(struct priv *priv,
  *   0 on success, errno value on failure.
  */
 static int
-priv_fdir_filter_update(struct priv *priv,
+mlx5_fdir_filter_update(struct rte_eth_dev *dev,
 			const struct rte_eth_fdir_filter *fdir_filter)
 {
 	int ret;
 
-	ret = priv_fdir_filter_delete(priv, fdir_filter);
+	ret = mlx5_fdir_filter_delete(dev, fdir_filter);
 	if (ret)
 		return ret;
-	ret = priv_fdir_filter_add(priv, fdir_filter);
+	ret = mlx5_fdir_filter_add(dev, fdir_filter);
 	return ret;
 }
 
 /**
  * Flush all filters.
  *
- * @param priv
- *   Private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  */
 static void
-priv_fdir_filter_flush(struct priv *priv)
+mlx5_fdir_filter_flush(struct rte_eth_dev *dev)
 {
-	priv_flow_flush(priv, &priv->flows);
+	struct priv *priv = dev->data->dev_private;
+
+	mlx5_flow_list_flush(dev, &priv->flows);
 }
 
 /**
  * Get flow director information.
  *
- * @param priv
- *   Private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param[out] fdir_info
  *   Resulting flow director information.
  */
 static void
-priv_fdir_info_get(struct priv *priv, struct rte_eth_fdir_info *fdir_info)
+mlx5_fdir_info_get(struct rte_eth_dev *dev, struct rte_eth_fdir_info *fdir_info)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct rte_eth_fdir_masks *mask =
 		&priv->dev->data->dev_conf.fdir_conf.mask;
 
@@ -2962,8 +2954,8 @@ priv_fdir_info_get(struct priv *priv, struct rte_eth_fdir_info *fdir_info)
 /**
  * Deal with flow director operations.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param filter_op
  *   Operation to perform.
  * @param arg
@@ -2973,8 +2965,10 @@ priv_fdir_info_get(struct priv *priv, struct rte_eth_fdir_info *fdir_info)
  *   0 on success, errno value on failure.
  */
 static int
-priv_fdir_ctrl_func(struct priv *priv, enum rte_filter_op filter_op, void *arg)
+mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
+		    void *arg)
 {
+	struct priv *priv = dev->data->dev_private;
 	enum rte_fdir_mode fdir_mode =
 		priv->dev->data->dev_conf.fdir_conf.mode;
 	int ret = 0;
@@ -2984,27 +2978,27 @@ priv_fdir_ctrl_func(struct priv *priv, enum rte_filter_op filter_op, void *arg)
 	if (fdir_mode != RTE_FDIR_MODE_PERFECT &&
 	    fdir_mode != RTE_FDIR_MODE_PERFECT_MAC_VLAN) {
 		ERROR("%p: flow director mode %d not supported",
-		      (void *)priv, fdir_mode);
+		      (void *)dev, fdir_mode);
 		return EINVAL;
 	}
 	switch (filter_op) {
 	case RTE_ETH_FILTER_ADD:
-		ret = priv_fdir_filter_add(priv, arg);
+		ret = mlx5_fdir_filter_add(dev, arg);
 		break;
 	case RTE_ETH_FILTER_UPDATE:
-		ret = priv_fdir_filter_update(priv, arg);
+		ret = mlx5_fdir_filter_update(dev, arg);
 		break;
 	case RTE_ETH_FILTER_DELETE:
-		ret = priv_fdir_filter_delete(priv, arg);
+		ret = mlx5_fdir_filter_delete(dev, arg);
 		break;
 	case RTE_ETH_FILTER_FLUSH:
-		priv_fdir_filter_flush(priv);
+		mlx5_fdir_filter_flush(dev);
 		break;
 	case RTE_ETH_FILTER_INFO:
-		priv_fdir_info_get(priv, arg);
+		mlx5_fdir_info_get(dev, arg);
 		break;
 	default:
-		DEBUG("%p: unknown operation %u", (void *)priv,
+		DEBUG("%p: unknown operation %u", (void *)dev,
 		      filter_op);
 		ret = EINVAL;
 		break;
@@ -3034,7 +3028,6 @@ mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
 		     void *arg)
 {
 	int ret = EINVAL;
-	struct priv *priv = dev->data->dev_private;
 
 	switch (filter_type) {
 	case RTE_ETH_FILTER_GENERIC:
@@ -3043,7 +3036,7 @@ mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
 		*(const void **)arg = &mlx5_flow_ops;
 		return 0;
 	case RTE_ETH_FILTER_FDIR:
-		ret = priv_fdir_ctrl_func(priv, filter_op, arg);
+		ret = mlx5_fdir_ctrl_func(dev, filter_op, arg);
 		break;
 	default:
 		ERROR("%p: filter type (%d) not supported",
diff --git a/drivers/net/mlx5/mlx5_mac.c b/drivers/net/mlx5/mlx5_mac.c
index 8e7b098ae..20fed527b 100644
--- a/drivers/net/mlx5/mlx5_mac.c
+++ b/drivers/net/mlx5/mlx5_mac.c
@@ -63,8 +63,8 @@
 /**
  * Get MAC address by querying netdevice.
  *
- * @param[in] priv
- *   struct priv for the requested device.
+ * @param[in] dev
+ *   Pointer to Ethernet device.
  * @param[out] mac
  *   MAC address output buffer.
  *
@@ -72,11 +72,11 @@
  *   0 on success, -1 on failure and errno is set.
  */
 int
-priv_get_mac(struct priv *priv, uint8_t (*mac)[ETHER_ADDR_LEN])
+mlx5_get_mac(struct rte_eth_dev *dev, uint8_t (*mac)[ETHER_ADDR_LEN])
 {
 	struct ifreq request;
 
-	if (priv_ifreq(priv, SIOCGIFHWADDR, &request))
+	if (mlx5_ifreq(dev, SIOCGIFHWADDR, &request))
 		return -1;
 	memcpy(mac, request.ifr_hwaddr.sa_data, ETHER_ADDR_LEN);
 	return 0;
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 12d9b2f8e..6f60aa1c5 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -112,10 +112,6 @@ mlx5_check_mempool(struct rte_mempool *mp, uintptr_t *start,
  * Register a Memory Region (MR) <-> Memory Pool (MP) association in
  * txq->mp2mr[]. If mp2mr[] is full, remove an entry first.
  *
- * This function should only be called by txq_mp2mr().
- *
- * @param priv
- *   Pointer to private structure.
  * @param txq
  *   Pointer to TX queue structure.
  * @param[in] mp
@@ -127,29 +123,35 @@ mlx5_check_mempool(struct rte_mempool *mp, uintptr_t *start,
  *   mr on success, NULL on failure.
  */
 struct mlx5_mr *
-priv_txq_mp2mr_reg(struct priv *priv, struct mlx5_txq_data *txq,
-		   struct rte_mempool *mp, unsigned int idx)
+mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
+		   unsigned int idx)
 {
 	struct mlx5_txq_ctrl *txq_ctrl =
 		container_of(txq, struct mlx5_txq_ctrl, txq);
+	struct rte_eth_dev *dev;
 	struct mlx5_mr *mr;
 
+	rte_spinlock_lock(&txq_ctrl->priv->mr_lock);
 	/* Add a new entry, register MR first. */
 	DEBUG("%p: discovered new memory pool \"%s\" (%p)",
 	      (void *)txq_ctrl, mp->name, (void *)mp);
-	mr = priv_mr_get(priv, mp);
+	dev = txq_ctrl->priv->dev;
+	mr = mlx5_mr_get(dev, mp);
 	if (mr == NULL) {
 		if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
-			DEBUG("Using unregistered mempool 0x%p(%s) in secondary process,"
-			     " please create mempool before rte_eth_dev_start()",
+			DEBUG("Using unregistered mempool 0x%p(%s) in "
+			      "secondary process, please create mempool before "
+			      " rte_eth_dev_start()",
 			     (void *)mp, mp->name);
+			rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 			return NULL;
 		}
-		mr = priv_mr_new(priv, mp);
+		mr = mlx5_mr_new(dev, mp);
 	}
 	if (unlikely(mr == NULL)) {
 		DEBUG("%p: unable to configure MR, ibv_reg_mr() failed.",
 		      (void *)txq_ctrl);
+		rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 		return NULL;
 	}
 	if (unlikely(idx == RTE_DIM(txq->mp2mr))) {
@@ -157,7 +159,7 @@ priv_txq_mp2mr_reg(struct priv *priv, struct mlx5_txq_data *txq,
 		DEBUG("%p: MR <-> MP table full, dropping oldest entry.",
 		      (void *)txq_ctrl);
 		--idx;
-		priv_mr_release(priv, txq->mp2mr[0]);
+		mlx5_mr_release(txq->mp2mr[0]);
 		memmove(&txq->mp2mr[0], &txq->mp2mr[1],
 			(sizeof(txq->mp2mr) - sizeof(txq->mp2mr[0])));
 	}
@@ -166,35 +168,6 @@ priv_txq_mp2mr_reg(struct priv *priv, struct mlx5_txq_data *txq,
 	DEBUG("%p: new MR lkey for MP \"%s\" (%p): 0x%08" PRIu32,
 	      (void *)txq_ctrl, mp->name, (void *)mp,
 	      txq_ctrl->txq.mp2mr[idx]->lkey);
-	return mr;
-}
-
-/**
- * Register a Memory Region (MR) <-> Memory Pool (MP) association in
- * txq->mp2mr[]. If mp2mr[] is full, remove an entry first.
- *
- * This function should only be called by txq_mp2mr().
- *
- * @param txq
- *   Pointer to TX queue structure.
- * @param[in] mp
- *   Memory Pool for which a Memory Region lkey must be returned.
- * @param idx
- *   Index of the next available entry.
- *
- * @return
- *   mr on success, NULL on failure.
- */
-struct mlx5_mr*
-mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
-		   unsigned int idx)
-{
-	struct mlx5_txq_ctrl *txq_ctrl =
-		container_of(txq, struct mlx5_txq_ctrl, txq);
-	struct mlx5_mr *mr;
-
-	rte_spinlock_lock(&txq_ctrl->priv->mr_lock);
-	mr = priv_txq_mp2mr_reg(txq_ctrl->priv, txq, mp, idx);
 	rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 	return mr;
 }
@@ -254,20 +227,20 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
 	if (rte_mempool_obj_iter(mp, txq_mp2mr_mbuf_check, &data) == 0 ||
 			data.ret == -1)
 		return;
-	mr = priv_mr_get(priv, mp);
+	mr = mlx5_mr_get(priv->dev, mp);
 	if (mr) {
-		priv_mr_release(priv, mr);
+		mlx5_mr_release(mr);
 		return;
 	}
-	priv_mr_new(priv, mp);
+	mlx5_mr_new(priv->dev, mp);
 }
 
 /**
  * Register a new memory region from the mempool and store it in the memory
  * region list.
  *
- * @param  priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param mp
  *   Pointer to the memory pool to register.
  *
@@ -275,8 +248,9 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
  *   The memory region on success.
  */
 struct mlx5_mr *
-priv_mr_new(struct priv *priv, struct rte_mempool *mp)
+mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 {
+	struct priv *priv = dev->data->dev_private;
 	const struct rte_memseg *ms = rte_eal_get_physmem_layout();
 	uintptr_t start;
 	uintptr_t end;
@@ -318,7 +292,7 @@ priv_mr_new(struct priv *priv, struct rte_mempool *mp)
 	mr->mp = mp;
 	mr->lkey = rte_cpu_to_be_32(mr->mr->lkey);
 	rte_atomic32_inc(&mr->refcnt);
-	DEBUG("%p: new Memory Region %p refcnt: %d", (void *)priv,
+	DEBUG("%p: new Memory Region %p refcnt: %d", (void *)dev,
 	      (void *)mr, rte_atomic32_read(&mr->refcnt));
 	LIST_INSERT_HEAD(&priv->mr, mr, next);
 	return mr;
@@ -327,8 +301,8 @@ priv_mr_new(struct priv *priv, struct rte_mempool *mp)
 /**
  * Search the memory region object in the memory region list.
  *
- * @param  priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param mp
  *   Pointer to the memory pool to register.
  *
@@ -336,8 +310,9 @@ priv_mr_new(struct priv *priv, struct rte_mempool *mp)
  *   The memory region on success.
  */
 struct mlx5_mr *
-priv_mr_get(struct priv *priv, struct rte_mempool *mp)
+mlx5_mr_get(struct rte_eth_dev *dev, struct rte_mempool *mp)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_mr *mr;
 
 	assert(mp);
@@ -364,7 +339,7 @@ priv_mr_get(struct priv *priv, struct rte_mempool *mp)
  *   0 on success, errno on failure.
  */
 int
-priv_mr_release(struct priv *priv __rte_unused, struct mlx5_mr *mr)
+mlx5_mr_release(struct mlx5_mr *mr)
 {
 	assert(mr);
 	DEBUG("Memory Region %p refcnt: %d",
@@ -381,20 +356,21 @@ priv_mr_release(struct priv *priv __rte_unused, struct mlx5_mr *mr)
 /**
  * Verify the flow list is empty
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   The number of object not released.
  */
 int
-priv_mr_verify(struct priv *priv)
+mlx5_mr_verify(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	int ret = 0;
 	struct mlx5_mr *mr;
 
 	LIST_FOREACH(mr, &priv->mr, next) {
-		DEBUG("%p: mr %p still referenced", (void *)priv,
+		DEBUG("%p: mr %p still referenced", (void *)dev,
 		      (void *)mr);
 		++ret;
 	}
diff --git a/drivers/net/mlx5/mlx5_rss.c b/drivers/net/mlx5/mlx5_rss.c
index 45838f0b3..badf0c0f9 100644
--- a/drivers/net/mlx5/mlx5_rss.c
+++ b/drivers/net/mlx5/mlx5_rss.c
@@ -124,8 +124,8 @@ mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
 /**
  * Allocate/reallocate RETA index table.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @praram reta_size
  *   The size of the array to allocate.
  *
@@ -133,8 +133,9 @@ mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
  *   0 on success, errno value on failure.
  */
 int
-priv_rss_reta_index_resize(struct priv *priv, unsigned int reta_size)
+mlx5_rss_reta_index_resize(struct rte_eth_dev *dev, unsigned int reta_size)
 {
+	struct priv *priv = dev->data->dev_private;
 	void *mem;
 	unsigned int old_size = priv->reta_idx_n;
 
@@ -155,28 +156,29 @@ priv_rss_reta_index_resize(struct priv *priv, unsigned int reta_size)
 }
 
 /**
- * Query RETA table.
+ * DPDK callback to get the RETA indirection table.
  *
- * @param priv
- *   Pointer to private structure.
- * @param[in, out] reta_conf
- *   Pointer to the first RETA configuration structure.
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ * @param reta_conf
+ *   Pointer to RETA configuration structure array.
  * @param reta_size
- *   Number of entries.
+ *   Size of the RETA table.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, negative errno value on failure.
  */
-static int
-priv_dev_rss_reta_query(struct priv *priv,
+int
+mlx5_dev_rss_reta_query(struct rte_eth_dev *dev,
 			struct rte_eth_rss_reta_entry64 *reta_conf,
-			unsigned int reta_size)
+			uint16_t reta_size)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int idx;
 	unsigned int i;
 
 	if (!reta_size || reta_size > priv->reta_idx_n)
-		return EINVAL;
+		return -EINVAL;
 	/* Fill each entry of the table even if its bit is not set. */
 	for (idx = 0, i = 0; (i != reta_size); ++i) {
 		idx = i / RTE_RETA_GROUP_SIZE;
@@ -187,31 +189,32 @@ priv_dev_rss_reta_query(struct priv *priv,
 }
 
 /**
- * Update RETA table.
+ * DPDK callback to update the RETA indirection table.
  *
- * @param priv
- *   Pointer to private structure.
- * @param[in] reta_conf
- *   Pointer to the first RETA configuration structure.
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ * @param reta_conf
+ *   Pointer to RETA configuration structure array.
  * @param reta_size
- *   Number of entries.
+ *   Size of the RETA table.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, negative errno value on failure.
  */
-static int
-priv_dev_rss_reta_update(struct priv *priv,
+int
+mlx5_dev_rss_reta_update(struct rte_eth_dev *dev,
 			 struct rte_eth_rss_reta_entry64 *reta_conf,
-			 unsigned int reta_size)
+			 uint16_t reta_size)
 {
+	int ret;
+	struct priv *priv = dev->data->dev_private;
 	unsigned int idx;
 	unsigned int i;
 	unsigned int pos;
-	int ret;
 
 	if (!reta_size)
-		return EINVAL;
-	ret = priv_rss_reta_index_resize(priv, reta_size);
+		return -EINVAL;
+	ret = mlx5_rss_reta_index_resize(dev, reta_size);
 	if (ret)
 		return ret;
 	for (idx = 0, i = 0; (i != reta_size); ++i) {
@@ -222,56 +225,6 @@ priv_dev_rss_reta_update(struct priv *priv,
 		assert(reta_conf[idx].reta[pos] < priv->rxqs_n);
 		(*priv->reta_idx)[i] = reta_conf[idx].reta[pos];
 	}
-	return 0;
-}
-
-/**
- * DPDK callback to get the RETA indirection table.
- *
- * @param dev
- *   Pointer to Ethernet device structure.
- * @param reta_conf
- *   Pointer to RETA configuration structure array.
- * @param reta_size
- *   Size of the RETA table.
- *
- * @return
- *   0 on success, negative errno value on failure.
- */
-int
-mlx5_dev_rss_reta_query(struct rte_eth_dev *dev,
-			struct rte_eth_rss_reta_entry64 *reta_conf,
-			uint16_t reta_size)
-{
-	int ret;
-	struct priv *priv = dev->data->dev_private;
-
-	ret = priv_dev_rss_reta_query(priv, reta_conf, reta_size);
-	return -ret;
-}
-
-/**
- * DPDK callback to update the RETA indirection table.
- *
- * @param dev
- *   Pointer to Ethernet device structure.
- * @param reta_conf
- *   Pointer to RETA configuration structure array.
- * @param reta_size
- *   Size of the RETA table.
- *
- * @return
- *   0 on success, negative errno value on failure.
- */
-int
-mlx5_dev_rss_reta_update(struct rte_eth_dev *dev,
-			 struct rte_eth_rss_reta_entry64 *reta_conf,
-			 uint16_t reta_size)
-{
-	int ret;
-	struct priv *priv = dev->data->dev_private;
-
-	ret = priv_dev_rss_reta_update(priv, reta_conf, reta_size);
 	if (dev->data->dev_started) {
 		mlx5_dev_stop(dev);
 		mlx5_dev_start(dev);
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 91f598c18..637092d44 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -124,7 +124,7 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 		(*rxq_ctrl->rxq.elts)[i] = buf;
 	}
 	/* If Rx vector is activated. */
-	if (rxq_check_vec_support(&rxq_ctrl->rxq) > 0) {
+	if (mlx5_rxq_check_vec_support(&rxq_ctrl->rxq) > 0) {
 		struct mlx5_rxq_data *rxq = &rxq_ctrl->rxq;
 		struct rte_mbuf *mbuf_init = &rxq->fake_mbuf;
 		int j;
@@ -183,7 +183,7 @@ rxq_free_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 	 * Some mbuf in the Ring belongs to the application.  They cannot be
 	 * freed.
 	 */
-	if (rxq_check_vec_support(rxq) > 0) {
+	if (mlx5_rxq_check_vec_support(rxq) > 0) {
 		for (i = 0; i < used; ++i)
 			(*rxq->elts)[(rxq->rq_ci + i) & q_mask] = NULL;
 		rxq->rq_pi = rxq->rq_ci;
@@ -208,7 +208,7 @@ mlx5_rxq_cleanup(struct mlx5_rxq_ctrl *rxq_ctrl)
 {
 	DEBUG("cleaning up %p", (void *)rxq_ctrl);
 	if (rxq_ctrl->ibv)
-		mlx5_priv_rxq_ibv_release(rxq_ctrl->priv, rxq_ctrl->ibv);
+		mlx5_rxq_ibv_release(rxq_ctrl->ibv);
 	memset(rxq_ctrl, 0, sizeof(*rxq_ctrl));
 }
 
@@ -255,14 +255,14 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		      (void *)dev, idx, priv->rxqs_n);
 		return -EOVERFLOW;
 	}
-	if (!mlx5_priv_rxq_releasable(priv, idx)) {
+	if (!mlx5_rxq_releasable(dev, idx)) {
 		ret = EBUSY;
 		ERROR("%p: unable to release queue index %u",
 		      (void *)dev, idx);
 		goto out;
 	}
-	mlx5_priv_rxq_release(priv, idx);
-	rxq_ctrl = mlx5_priv_rxq_new(priv, idx, desc, socket, mp);
+	mlx5_rxq_release(dev, idx);
+	rxq_ctrl = mlx5_rxq_new(dev, idx, desc, socket, mp);
 	if (!rxq_ctrl) {
 		ERROR("%p: unable to allocate queue index %u",
 		      (void *)dev, idx);
@@ -293,24 +293,25 @@ mlx5_rx_queue_release(void *dpdk_rxq)
 		return;
 	rxq_ctrl = container_of(rxq, struct mlx5_rxq_ctrl, rxq);
 	priv = rxq_ctrl->priv;
-	if (!mlx5_priv_rxq_releasable(priv, rxq_ctrl->rxq.stats.idx))
+	if (!mlx5_rxq_releasable(priv->dev, rxq_ctrl->rxq.stats.idx))
 		rte_panic("Rx queue %p is still used by a flow and cannot be"
 			  " removed\n", (void *)rxq_ctrl);
-	mlx5_priv_rxq_release(priv, rxq_ctrl->rxq.stats.idx);
+	mlx5_rxq_release(priv->dev, rxq_ctrl->rxq.stats.idx);
 }
 
 /**
  * Allocate queue vector and fill epoll fd list for Rx interrupts.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   0 on success, negative on failure.
  */
 int
-priv_rx_intr_vec_enable(struct priv *priv)
+mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 	unsigned int rxqs_n = priv->rxqs_n;
 	unsigned int n = RTE_MIN(rxqs_n, (uint32_t)RTE_MAX_RXTX_INTR_VEC_ID);
@@ -319,7 +320,7 @@ priv_rx_intr_vec_enable(struct priv *priv)
 
 	if (!priv->dev->data->dev_conf.intr_conf.rxq)
 		return 0;
-	priv_rx_intr_vec_disable(priv);
+	mlx5_rx_intr_vec_disable(dev);
 	intr_handle->intr_vec = malloc(n * sizeof(intr_handle->intr_vec[0]));
 	if (intr_handle->intr_vec == NULL) {
 		ERROR("failed to allocate memory for interrupt vector,"
@@ -329,7 +330,7 @@ priv_rx_intr_vec_enable(struct priv *priv)
 	intr_handle->type = RTE_INTR_HANDLE_EXT;
 	for (i = 0; i != n; ++i) {
 		/* This rxq ibv must not be released in this function. */
-		struct mlx5_rxq_ibv *rxq_ibv = mlx5_priv_rxq_ibv_get(priv, i);
+		struct mlx5_rxq_ibv *rxq_ibv = mlx5_rxq_ibv_get(dev, i);
 		int fd;
 		int flags;
 		int rc;
@@ -346,7 +347,7 @@ priv_rx_intr_vec_enable(struct priv *priv)
 			ERROR("too many Rx queues for interrupt vector size"
 			      " (%d), Rx interrupts cannot be enabled",
 			      RTE_MAX_RXTX_INTR_VEC_ID);
-			priv_rx_intr_vec_disable(priv);
+			mlx5_rx_intr_vec_disable(dev);
 			return -1;
 		}
 		fd = rxq_ibv->channel->fd;
@@ -355,7 +356,7 @@ priv_rx_intr_vec_enable(struct priv *priv)
 		if (rc < 0) {
 			ERROR("failed to make Rx interrupt file descriptor"
 			      " %d non-blocking for queue index %d", fd, i);
-			priv_rx_intr_vec_disable(priv);
+			mlx5_rx_intr_vec_disable(dev);
 			return -1;
 		}
 		intr_handle->intr_vec[i] = RTE_INTR_VEC_RXTX_OFFSET + count;
@@ -363,7 +364,7 @@ priv_rx_intr_vec_enable(struct priv *priv)
 		count++;
 	}
 	if (!count)
-		priv_rx_intr_vec_disable(priv);
+		mlx5_rx_intr_vec_disable(dev);
 	else
 		intr_handle->nb_efd = count;
 	return 0;
@@ -372,12 +373,13 @@ priv_rx_intr_vec_enable(struct priv *priv)
 /**
  * Clean up Rx interrupts handler.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  */
 void
-priv_rx_intr_vec_disable(struct priv *priv)
+mlx5_rx_intr_vec_disable(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct rte_intr_handle *intr_handle = priv->dev->intr_handle;
 	unsigned int i;
 	unsigned int rxqs_n = priv->rxqs_n;
@@ -400,7 +402,7 @@ priv_rx_intr_vec_disable(struct priv *priv)
 		 */
 		rxq_data = (*priv->rxqs)[i];
 		rxq_ctrl = container_of(rxq_data, struct mlx5_rxq_ctrl, rxq);
-		mlx5_priv_rxq_ibv_release(priv, rxq_ctrl->ibv);
+		mlx5_rxq_ibv_release(rxq_ctrl->ibv);
 	}
 free:
 	rte_intr_free_epoll_fd(intr_handle);
@@ -463,13 +465,13 @@ mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	if (rxq_ctrl->irq) {
 		struct mlx5_rxq_ibv *rxq_ibv;
 
-		rxq_ibv = mlx5_priv_rxq_ibv_get(priv, rx_queue_id);
+		rxq_ibv = mlx5_rxq_ibv_get(dev, rx_queue_id);
 		if (!rxq_ibv) {
 			ret = EINVAL;
 			goto exit;
 		}
 		mlx5_arm_cq(rxq_data, rxq_data->cq_arm_sn);
-		mlx5_priv_rxq_ibv_release(priv, rxq_ibv);
+		mlx5_rxq_ibv_release(rxq_ibv);
 	}
 exit:
 	if (ret)
@@ -507,7 +509,7 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	rxq_ctrl = container_of(rxq_data, struct mlx5_rxq_ctrl, rxq);
 	if (!rxq_ctrl->irq)
 		goto exit;
-	rxq_ibv = mlx5_priv_rxq_ibv_get(priv, rx_queue_id);
+	rxq_ibv = mlx5_rxq_ibv_get(dev, rx_queue_id);
 	if (!rxq_ibv) {
 		ret = EINVAL;
 		goto exit;
@@ -521,7 +523,7 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	ibv_ack_cq_events(rxq_ibv->cq, 1);
 exit:
 	if (rxq_ibv)
-		mlx5_priv_rxq_ibv_release(priv, rxq_ibv);
+		mlx5_rxq_ibv_release(rxq_ibv);
 	if (ret)
 		WARN("unable to disable interrupt on rx queue %d",
 		     rx_queue_id);
@@ -531,8 +533,8 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 /**
  * Create the Rx queue Verbs object.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   Queue index in DPDK Rx queue array
  *
@@ -540,8 +542,9 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
  *   The Verbs object initialised if it can be created.
  */
 struct mlx5_rxq_ibv *
-mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
+mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_data *rxq_data = (*priv->rxqs)[idx];
 	struct mlx5_rxq_ctrl *rxq_ctrl =
 		container_of(rxq_data, struct mlx5_rxq_ctrl, rxq);
@@ -575,9 +578,9 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 	}
 	tmpl->rxq_ctrl = rxq_ctrl;
 	/* Use the entire RX mempool as the memory region. */
-	tmpl->mr = priv_mr_get(priv, rxq_data->mp);
+	tmpl->mr = mlx5_mr_get(dev, rxq_data->mp);
 	if (!tmpl->mr) {
-		tmpl->mr = priv_mr_new(priv, rxq_data->mp);
+		tmpl->mr = mlx5_mr_new(dev, rxq_data->mp);
 		if (!tmpl->mr) {
 			ERROR("%p: MR creation failure", (void *)rxq_ctrl);
 			goto error;
@@ -607,7 +610,7 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 		 * For vectorized Rx, it must not be doubled in order to
 		 * make cq_ci and rq_ci aligned.
 		 */
-		if (rxq_check_vec_support(rxq_data) < 0)
+		if (mlx5_rxq_check_vec_support(rxq_data) < 0)
 			attr.cq.ibv.cqe *= 2;
 	} else if (priv->cqe_comp && rxq_data->hw_timestamp) {
 		DEBUG("Rx CQE compression is disabled for HW timestamp");
@@ -726,7 +729,7 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 	*rxq_data->rq_db = rte_cpu_to_be_32(rxq_data->rq_ci);
 	DEBUG("%p: rxq updated with %p", (void *)rxq_ctrl, (void *)&tmpl);
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("%p: Verbs Rx queue %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Verbs Rx queue %p: refcnt %d", (void *)dev,
 	      (void *)tmpl, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->rxqsibv, tmpl, next);
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
@@ -739,7 +742,7 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 	if (tmpl->channel)
 		claim_zero(ibv_destroy_comp_channel(tmpl->channel));
 	if (tmpl->mr)
-		priv_mr_release(priv, tmpl->mr);
+		mlx5_mr_release(tmpl->mr);
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return NULL;
 }
@@ -747,8 +750,8 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
 /**
  * Get an Rx queue Verbs object.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   Queue index in DPDK Rx queue array
  *
@@ -756,8 +759,9 @@ mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx)
  *   The Verbs object if it exists.
  */
 struct mlx5_rxq_ibv *
-mlx5_priv_rxq_ibv_get(struct priv *priv, uint16_t idx)
+mlx5_rxq_ibv_get(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_data *rxq_data = (*priv->rxqs)[idx];
 	struct mlx5_rxq_ctrl *rxq_ctrl;
 
@@ -767,9 +771,9 @@ mlx5_priv_rxq_ibv_get(struct priv *priv, uint16_t idx)
 		return NULL;
 	rxq_ctrl = container_of(rxq_data, struct mlx5_rxq_ctrl, rxq);
 	if (rxq_ctrl->ibv) {
-		priv_mr_get(priv, rxq_data->mp);
+		mlx5_mr_get(dev, rxq_data->mp);
 		rte_atomic32_inc(&rxq_ctrl->ibv->refcnt);
-		DEBUG("%p: Verbs Rx queue %p: refcnt %d", (void *)priv,
+		DEBUG("%p: Verbs Rx queue %p: refcnt %d", (void *)dev,
 		      (void *)rxq_ctrl->ibv,
 		      rte_atomic32_read(&rxq_ctrl->ibv->refcnt));
 	}
@@ -779,8 +783,6 @@ mlx5_priv_rxq_ibv_get(struct priv *priv, uint16_t idx)
 /**
  * Release an Rx verbs queue object.
  *
- * @param priv
- *   Pointer to private structure.
  * @param rxq_ibv
  *   Verbs Rx queue object.
  *
@@ -788,7 +790,7 @@ mlx5_priv_rxq_ibv_get(struct priv *priv, uint16_t idx)
  *   0 on success, errno value on failure.
  */
 int
-mlx5_priv_rxq_ibv_release(struct priv *priv, struct mlx5_rxq_ibv *rxq_ibv)
+mlx5_rxq_ibv_release(struct mlx5_rxq_ibv *rxq_ibv)
 {
 	int ret;
 
@@ -796,10 +798,10 @@ mlx5_priv_rxq_ibv_release(struct priv *priv, struct mlx5_rxq_ibv *rxq_ibv)
 	assert(rxq_ibv->wq);
 	assert(rxq_ibv->cq);
 	assert(rxq_ibv->mr);
-	ret = priv_mr_release(priv, rxq_ibv->mr);
+	ret = mlx5_mr_release(rxq_ibv->mr);
 	if (!ret)
 		rxq_ibv->mr = NULL;
-	DEBUG("%p: Verbs Rx queue %p: refcnt %d", (void *)priv,
+	DEBUG("Verbs Rx queue %p: refcnt %d",
 	      (void *)rxq_ibv, rte_atomic32_read(&rxq_ibv->refcnt));
 	if (rte_atomic32_dec_and_test(&rxq_ibv->refcnt)) {
 		rxq_free_elts(rxq_ibv->rxq_ctrl);
@@ -817,20 +819,21 @@ mlx5_priv_rxq_ibv_release(struct priv *priv, struct mlx5_rxq_ibv *rxq_ibv)
 /**
  * Verify the Verbs Rx queue list is empty
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   The number of object not released.
  */
 int
-mlx5_priv_rxq_ibv_verify(struct priv *priv)
+mlx5_rxq_ibv_verify(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	int ret = 0;
 	struct mlx5_rxq_ibv *rxq_ibv;
 
 	LIST_FOREACH(rxq_ibv, &priv->rxqsibv, next) {
-		DEBUG("%p: Verbs Rx queue %p still referenced", (void *)priv,
+		DEBUG("%p: Verbs Rx queue %p still referenced", (void *)dev,
 		      (void *)rxq_ibv);
 		++ret;
 	}
@@ -840,14 +843,11 @@ mlx5_priv_rxq_ibv_verify(struct priv *priv)
 /**
  * Return true if a single reference exists on the object.
  *
- * @param priv
- *   Pointer to private structure.
  * @param rxq_ibv
  *   Verbs Rx queue object.
  */
 int
-mlx5_priv_rxq_ibv_releasable(struct priv *priv __rte_unused,
-			     struct mlx5_rxq_ibv *rxq_ibv)
+mlx5_rxq_ibv_releasable(struct mlx5_rxq_ibv *rxq_ibv)
 {
 	assert(rxq_ibv);
 	return (rte_atomic32_read(&rxq_ibv->refcnt) == 1);
@@ -856,8 +856,8 @@ mlx5_priv_rxq_ibv_releasable(struct priv *priv __rte_unused,
 /**
  * Create a DPDK Rx queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   TX queue index.
  * @param desc
@@ -869,10 +869,10 @@ mlx5_priv_rxq_ibv_releasable(struct priv *priv __rte_unused,
  *   A DPDK queue object on success.
  */
 struct mlx5_rxq_ctrl *
-mlx5_priv_rxq_new(struct priv *priv, uint16_t idx, uint16_t desc,
-		  unsigned int socket, struct rte_mempool *mp)
+mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
+	     unsigned int socket, struct rte_mempool *mp)
 {
-	struct rte_eth_dev *dev = priv->dev;
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_ctrl *tmpl;
 	const uint16_t desc_n =
 		desc + priv->rx_vec_en * MLX5_VPMD_DESCS_PER_LOOP;
@@ -971,7 +971,7 @@ mlx5_priv_rxq_new(struct priv *priv, uint16_t idx, uint16_t desc,
 	tmpl->rxq.elts =
 		(struct rte_mbuf *(*)[1 << tmpl->rxq.elts_n])(tmpl + 1);
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("%p: Rx queue %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Rx queue %p: refcnt %d", (void *)dev,
 	      (void *)tmpl, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->rxqsctrl, tmpl, next);
 	return tmpl;
@@ -983,8 +983,8 @@ mlx5_priv_rxq_new(struct priv *priv, uint16_t idx, uint16_t desc,
 /**
  * Get a Rx queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   TX queue index.
  *
@@ -992,17 +992,18 @@ mlx5_priv_rxq_new(struct priv *priv, uint16_t idx, uint16_t desc,
  *   A pointer to the queue if it exists.
  */
 struct mlx5_rxq_ctrl *
-mlx5_priv_rxq_get(struct priv *priv, uint16_t idx)
+mlx5_rxq_get(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_ctrl *rxq_ctrl = NULL;
 
 	if ((*priv->rxqs)[idx]) {
 		rxq_ctrl = container_of((*priv->rxqs)[idx],
 					struct mlx5_rxq_ctrl,
 					rxq);
-		mlx5_priv_rxq_ibv_get(priv, idx);
+		mlx5_rxq_ibv_get(dev, idx);
 		rte_atomic32_inc(&rxq_ctrl->refcnt);
-		DEBUG("%p: Rx queue %p: refcnt %d", (void *)priv,
+		DEBUG("%p: Rx queue %p: refcnt %d", (void *)dev,
 		      (void *)rxq_ctrl, rte_atomic32_read(&rxq_ctrl->refcnt));
 	}
 	return rxq_ctrl;
@@ -1011,8 +1012,8 @@ mlx5_priv_rxq_get(struct priv *priv, uint16_t idx)
 /**
  * Release a Rx queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   TX queue index.
  *
@@ -1020,8 +1021,9 @@ mlx5_priv_rxq_get(struct priv *priv, uint16_t idx)
  *   0 on success, errno value on failure.
  */
 int
-mlx5_priv_rxq_release(struct priv *priv, uint16_t idx)
+mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_ctrl *rxq_ctrl;
 
 	if (!(*priv->rxqs)[idx])
@@ -1031,11 +1033,11 @@ mlx5_priv_rxq_release(struct priv *priv, uint16_t idx)
 	if (rxq_ctrl->ibv) {
 		int ret;
 
-		ret = mlx5_priv_rxq_ibv_release(rxq_ctrl->priv, rxq_ctrl->ibv);
+		ret = mlx5_rxq_ibv_release(rxq_ctrl->ibv);
 		if (!ret)
 			rxq_ctrl->ibv = NULL;
 	}
-	DEBUG("%p: Rx queue %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Rx queue %p: refcnt %d", (void *)dev,
 	      (void *)rxq_ctrl, rte_atomic32_read(&rxq_ctrl->refcnt));
 	if (rte_atomic32_dec_and_test(&rxq_ctrl->refcnt)) {
 		LIST_REMOVE(rxq_ctrl, next);
@@ -1049,8 +1051,8 @@ mlx5_priv_rxq_release(struct priv *priv, uint16_t idx)
 /**
  * Verify if the queue can be released.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   TX queue index.
  *
@@ -1058,8 +1060,9 @@ mlx5_priv_rxq_release(struct priv *priv, uint16_t idx)
  *   1 if the queue can be released.
  */
 int
-mlx5_priv_rxq_releasable(struct priv *priv, uint16_t idx)
+mlx5_rxq_releasable(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_ctrl *rxq_ctrl;
 
 	if (!(*priv->rxqs)[idx])
@@ -1071,20 +1074,21 @@ mlx5_priv_rxq_releasable(struct priv *priv, uint16_t idx)
 /**
  * Verify the Rx Queue list is empty
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   The number of object not released.
  */
 int
-mlx5_priv_rxq_verify(struct priv *priv)
+mlx5_rxq_verify(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_ctrl *rxq_ctrl;
 	int ret = 0;
 
 	LIST_FOREACH(rxq_ctrl, &priv->rxqsctrl, next) {
-		DEBUG("%p: Rx Queue %p still referenced", (void *)priv,
+		DEBUG("%p: Rx Queue %p still referenced", (void *)dev,
 		      (void *)rxq_ctrl);
 		++ret;
 	}
@@ -1094,8 +1098,8 @@ mlx5_priv_rxq_verify(struct priv *priv)
 /**
  * Create an indirection table.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param queues
  *   Queues entering in the indirection table.
  * @param queues_n
@@ -1105,9 +1109,10 @@ mlx5_priv_rxq_verify(struct priv *priv)
  *   A new indirection table.
  */
 struct mlx5_ind_table_ibv *
-mlx5_priv_ind_table_ibv_new(struct priv *priv, uint16_t queues[],
-			    uint16_t queues_n)
+mlx5_ind_table_ibv_new(struct rte_eth_dev *dev, uint16_t queues[],
+		       uint16_t queues_n)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_ind_table_ibv *ind_tbl;
 	const unsigned int wq_n = rte_is_power_of_2(queues_n) ?
 		log2above(queues_n) :
@@ -1121,8 +1126,7 @@ mlx5_priv_ind_table_ibv_new(struct priv *priv, uint16_t queues[],
 	if (!ind_tbl)
 		return NULL;
 	for (i = 0; i != queues_n; ++i) {
-		struct mlx5_rxq_ctrl *rxq =
-			mlx5_priv_rxq_get(priv, queues[i]);
+		struct mlx5_rxq_ctrl *rxq = mlx5_rxq_get(dev, queues[i]);
 
 		if (!rxq)
 			goto error;
@@ -1144,20 +1148,20 @@ mlx5_priv_ind_table_ibv_new(struct priv *priv, uint16_t queues[],
 		goto error;
 	rte_atomic32_inc(&ind_tbl->refcnt);
 	LIST_INSERT_HEAD(&priv->ind_tbls, ind_tbl, next);
-	DEBUG("%p: Indirection table %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Indirection table %p: refcnt %d", (void *)dev,
 	      (void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
 	return ind_tbl;
 error:
 	rte_free(ind_tbl);
-	DEBUG("%p cannot create indirection table", (void *)priv);
+	DEBUG("%p cannot create indirection table", (void *)dev);
 	return NULL;
 }
 
 /**
  * Get an indirection table.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param queues
  *   Queues entering in the indirection table.
  * @param queues_n
@@ -1167,9 +1171,10 @@ mlx5_priv_ind_table_ibv_new(struct priv *priv, uint16_t queues[],
  *   An indirection table if found.
  */
 struct mlx5_ind_table_ibv *
-mlx5_priv_ind_table_ibv_get(struct priv *priv, uint16_t queues[],
-			    uint16_t queues_n)
+mlx5_ind_table_ibv_get(struct rte_eth_dev *dev, uint16_t queues[],
+		       uint16_t queues_n)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_ind_table_ibv *ind_tbl;
 
 	LIST_FOREACH(ind_tbl, &priv->ind_tbls, next) {
@@ -1183,10 +1188,10 @@ mlx5_priv_ind_table_ibv_get(struct priv *priv, uint16_t queues[],
 		unsigned int i;
 
 		rte_atomic32_inc(&ind_tbl->refcnt);
-		DEBUG("%p: Indirection table %p: refcnt %d", (void *)priv,
+		DEBUG("%p: Indirection table %p: refcnt %d", (void *)dev,
 		      (void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
 		for (i = 0; i != ind_tbl->queues_n; ++i)
-			mlx5_priv_rxq_get(priv, ind_tbl->queues[i]);
+			mlx5_rxq_get(dev, ind_tbl->queues[i]);
 	}
 	return ind_tbl;
 }
@@ -1194,8 +1199,8 @@ mlx5_priv_ind_table_ibv_get(struct priv *priv, uint16_t queues[],
 /**
  * Release an indirection table.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param ind_table
  *   Indirection table to release.
  *
@@ -1203,17 +1208,17 @@ mlx5_priv_ind_table_ibv_get(struct priv *priv, uint16_t queues[],
  *   0 on success, errno value on failure.
  */
 int
-mlx5_priv_ind_table_ibv_release(struct priv *priv,
-				struct mlx5_ind_table_ibv *ind_tbl)
+mlx5_ind_table_ibv_release(struct rte_eth_dev *dev,
+			   struct mlx5_ind_table_ibv *ind_tbl)
 {
 	unsigned int i;
 
-	DEBUG("%p: Indirection table %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Indirection table %p: refcnt %d", (void *)dev,
 	      (void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
 	if (rte_atomic32_dec_and_test(&ind_tbl->refcnt))
 		claim_zero(ibv_destroy_rwq_ind_table(ind_tbl->ind_table));
 	for (i = 0; i != ind_tbl->queues_n; ++i)
-		claim_nonzero(mlx5_priv_rxq_release(priv, ind_tbl->queues[i]));
+		claim_nonzero(mlx5_rxq_release(dev, ind_tbl->queues[i]));
 	if (!rte_atomic32_read(&ind_tbl->refcnt)) {
 		LIST_REMOVE(ind_tbl, next);
 		rte_free(ind_tbl);
@@ -1225,21 +1230,22 @@ mlx5_priv_ind_table_ibv_release(struct priv *priv,
 /**
  * Verify the Rx Queue list is empty
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   The number of object not released.
  */
 int
-mlx5_priv_ind_table_ibv_verify(struct priv *priv)
+mlx5_ind_table_ibv_verify(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_ind_table_ibv *ind_tbl;
 	int ret = 0;
 
 	LIST_FOREACH(ind_tbl, &priv->ind_tbls, next) {
 		DEBUG("%p: Verbs indirection table %p still referenced",
-		      (void *)priv, (void *)ind_tbl);
+		      (void *)dev, (void *)ind_tbl);
 		++ret;
 	}
 	return ret;
@@ -1248,8 +1254,8 @@ mlx5_priv_ind_table_ibv_verify(struct priv *priv)
 /**
  * Create an Rx Hash queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param rss_key
  *   RSS key for the Rx hash queue.
  * @param rss_key_len
@@ -1266,17 +1272,18 @@ mlx5_priv_ind_table_ibv_verify(struct priv *priv)
  *   An hash Rx queue on success.
  */
 struct mlx5_hrxq *
-mlx5_priv_hrxq_new(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
-		   uint64_t hash_fields, uint16_t queues[], uint16_t queues_n)
+mlx5_hrxq_new(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
+	      uint64_t hash_fields, uint16_t queues[], uint16_t queues_n)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_hrxq *hrxq;
 	struct mlx5_ind_table_ibv *ind_tbl;
 	struct ibv_qp *qp;
 
 	queues_n = hash_fields ? queues_n : 1;
-	ind_tbl = mlx5_priv_ind_table_ibv_get(priv, queues, queues_n);
+	ind_tbl = mlx5_ind_table_ibv_get(dev, queues, queues_n);
 	if (!ind_tbl)
-		ind_tbl = mlx5_priv_ind_table_ibv_new(priv, queues, queues_n);
+		ind_tbl = mlx5_ind_table_ibv_new(dev, queues, queues_n);
 	if (!ind_tbl)
 		return NULL;
 	qp = ibv_create_qp_ex(
@@ -1308,11 +1315,11 @@ mlx5_priv_hrxq_new(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
 	memcpy(hrxq->rss_key, rss_key, rss_key_len);
 	rte_atomic32_inc(&hrxq->refcnt);
 	LIST_INSERT_HEAD(&priv->hrxqs, hrxq, next);
-	DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)dev,
 	      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
 	return hrxq;
 error:
-	mlx5_priv_ind_table_ibv_release(priv, ind_tbl);
+	mlx5_ind_table_ibv_release(dev, ind_tbl);
 	if (qp)
 		claim_zero(ibv_destroy_qp(qp));
 	return NULL;
@@ -1321,8 +1328,8 @@ mlx5_priv_hrxq_new(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
 /**
  * Get an Rx Hash queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param rss_conf
  *   RSS configuration for the Rx hash queue.
  * @param queues
@@ -1335,9 +1342,10 @@ mlx5_priv_hrxq_new(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
  *   An hash Rx queue on success.
  */
 struct mlx5_hrxq *
-mlx5_priv_hrxq_get(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
-		   uint64_t hash_fields, uint16_t queues[], uint16_t queues_n)
+mlx5_hrxq_get(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
+	      uint64_t hash_fields, uint16_t queues[], uint16_t queues_n)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_hrxq *hrxq;
 
 	queues_n = hash_fields ? queues_n : 1;
@@ -1350,15 +1358,15 @@ mlx5_priv_hrxq_get(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
 			continue;
 		if (hrxq->hash_fields != hash_fields)
 			continue;
-		ind_tbl = mlx5_priv_ind_table_ibv_get(priv, queues, queues_n);
+		ind_tbl = mlx5_ind_table_ibv_get(dev, queues, queues_n);
 		if (!ind_tbl)
 			continue;
 		if (ind_tbl != hrxq->ind_table) {
-			mlx5_priv_ind_table_ibv_release(priv, ind_tbl);
+			mlx5_ind_table_ibv_release(dev, ind_tbl);
 			continue;
 		}
 		rte_atomic32_inc(&hrxq->refcnt);
-		DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)priv,
+		DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)dev,
 		      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
 		return hrxq;
 	}
@@ -1368,8 +1376,8 @@ mlx5_priv_hrxq_get(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
 /**
  * Release the hash Rx queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param hrxq
  *   Pointer to Hash Rx queue to release.
  *
@@ -1377,39 +1385,40 @@ mlx5_priv_hrxq_get(struct priv *priv, uint8_t *rss_key, uint8_t rss_key_len,
  *   0 on success, errno value on failure.
  */
 int
-mlx5_priv_hrxq_release(struct priv *priv, struct mlx5_hrxq *hrxq)
+mlx5_hrxq_release(struct rte_eth_dev *dev, struct mlx5_hrxq *hrxq)
 {
-	DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)dev,
 	      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
 	if (rte_atomic32_dec_and_test(&hrxq->refcnt)) {
 		claim_zero(ibv_destroy_qp(hrxq->qp));
-		mlx5_priv_ind_table_ibv_release(priv, hrxq->ind_table);
+		mlx5_ind_table_ibv_release(dev, hrxq->ind_table);
 		LIST_REMOVE(hrxq, next);
 		rte_free(hrxq);
 		return 0;
 	}
-	claim_nonzero(mlx5_priv_ind_table_ibv_release(priv, hrxq->ind_table));
+	claim_nonzero(mlx5_ind_table_ibv_release(dev, hrxq->ind_table));
 	return EBUSY;
 }
 
 /**
  * Verify the Rx Queue list is empty
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   The number of object not released.
  */
 int
-mlx5_priv_hrxq_ibv_verify(struct priv *priv)
+mlx5_hrxq_ibv_verify(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_hrxq *hrxq;
 	int ret = 0;
 
 	LIST_FOREACH(hrxq, &priv->hrxqs, next) {
 		DEBUG("%p: Verbs Hash Rx queue %p still referenced",
-		      (void *)priv, (void *)hrxq);
+		      (void *)dev, (void *)hrxq);
 		++ret;
 	}
 	return ret;
diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 40a271a68..62ff1a553 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -2038,25 +2038,25 @@ mlx5_rx_burst_vec(void *dpdk_txq __rte_unused,
 }
 
 int __attribute__((weak))
-priv_check_raw_vec_tx_support(struct priv *priv __rte_unused)
+mlx5_check_raw_vec_tx_support(struct rte_eth_dev *dev __rte_unused)
 {
 	return -ENOTSUP;
 }
 
 int __attribute__((weak))
-priv_check_vec_tx_support(struct priv *priv __rte_unused)
+mlx5_check_vec_tx_support(struct rte_eth_dev *dev __rte_unused)
 {
 	return -ENOTSUP;
 }
 
 int __attribute__((weak))
-rxq_check_vec_support(struct mlx5_rxq_data *rxq __rte_unused)
+mlx5_rxq_check_vec_support(struct mlx5_rxq_data *rxq __rte_unused)
 {
 	return -ENOTSUP;
 }
 
 int __attribute__((weak))
-priv_check_vec_rx_support(struct priv *priv __rte_unused)
+mlx5_check_vec_rx_support(struct rte_eth_dev *dev __rte_unused)
 {
 	return -ENOTSUP;
 }
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index 9d5edeedd..d91828498 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -242,62 +242,59 @@ int mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 			unsigned int socket, const struct rte_eth_rxconf *conf,
 			struct rte_mempool *mp);
 void mlx5_rx_queue_release(void *dpdk_rxq);
-int priv_rx_intr_vec_enable(struct priv *priv);
-void priv_rx_intr_vec_disable(struct priv *priv);
+int mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev);
+void mlx5_rx_intr_vec_disable(struct rte_eth_dev *dev);
 int mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id);
 int mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id);
-struct mlx5_rxq_ibv *mlx5_priv_rxq_ibv_new(struct priv *priv, uint16_t idx);
-struct mlx5_rxq_ibv *mlx5_priv_rxq_ibv_get(struct priv *priv, uint16_t idx);
-int mlx5_priv_rxq_ibv_release(struct priv *priv, struct mlx5_rxq_ibv *rxq_ibv);
-int mlx5_priv_rxq_ibv_releasable(struct priv *priv,
-				 struct mlx5_rxq_ibv *rxq_ibv);
-int mlx5_priv_rxq_ibv_verify(struct priv *priv);
-struct mlx5_rxq_ctrl *mlx5_priv_rxq_new(struct priv *priv, uint16_t idx,
-					uint16_t desc,
-					unsigned int socket,
-					struct rte_mempool *mp);
-struct mlx5_rxq_ctrl *mlx5_priv_rxq_get(struct priv *priv, uint16_t idx);
-int mlx5_priv_rxq_release(struct priv *priv, uint16_t idx);
-int mlx5_priv_rxq_releasable(struct priv *priv, uint16_t idx);
-int mlx5_priv_rxq_verify(struct priv *priv);
+struct mlx5_rxq_ibv *mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx);
+struct mlx5_rxq_ibv *mlx5_rxq_ibv_get(struct rte_eth_dev *dev, uint16_t idx);
+int mlx5_rxq_ibv_release(struct mlx5_rxq_ibv *rxq_ibv);
+int mlx5_rxq_ibv_releasable(struct mlx5_rxq_ibv *rxq_ibv);
+int mlx5_rxq_ibv_verify(struct rte_eth_dev *dev);
+struct mlx5_rxq_ctrl *mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx,
+				   uint16_t desc, unsigned int socket,
+				   struct rte_mempool *mp);
+struct mlx5_rxq_ctrl *mlx5_rxq_get(struct rte_eth_dev *dev, uint16_t idx);
+int mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx);
+int mlx5_rxq_releasable(struct rte_eth_dev *dev, uint16_t idx);
+int mlx5_rxq_verify(struct rte_eth_dev *dev);
 int rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl);
-struct mlx5_ind_table_ibv *mlx5_priv_ind_table_ibv_new(struct priv *priv,
-						       uint16_t queues[],
-						       uint16_t queues_n);
-struct mlx5_ind_table_ibv *mlx5_priv_ind_table_ibv_get(struct priv *priv,
-						       uint16_t queues[],
-						       uint16_t queues_n);
-int mlx5_priv_ind_table_ibv_release(struct priv *priv,
-				    struct mlx5_ind_table_ibv *ind_tbl);
-int mlx5_priv_ind_table_ibv_verify(struct priv *priv);
-struct mlx5_hrxq *mlx5_priv_hrxq_new(struct priv *priv, uint8_t *rss_key,
-				     uint8_t rss_key_len, uint64_t hash_fields,
-				     uint16_t queues[], uint16_t queues_n);
-struct mlx5_hrxq *mlx5_priv_hrxq_get(struct priv *priv, uint8_t *rss_key,
-				     uint8_t rss_key_len, uint64_t hash_fields,
-				     uint16_t queues[], uint16_t queues_n);
-int mlx5_priv_hrxq_release(struct priv *priv, struct mlx5_hrxq *hrxq);
-int mlx5_priv_hrxq_ibv_verify(struct priv *priv);
+struct mlx5_ind_table_ibv *mlx5_ind_table_ibv_new(struct rte_eth_dev *dev,
+						  uint16_t queues[],
+						  uint16_t queues_n);
+struct mlx5_ind_table_ibv *mlx5_ind_table_ibv_get(struct rte_eth_dev *dev,
+						  uint16_t queues[],
+						  uint16_t queues_n);
+int mlx5_ind_table_ibv_release(struct rte_eth_dev *dev,
+			       struct mlx5_ind_table_ibv *ind_tbl);
+int mlx5_ind_table_ibv_verify(struct rte_eth_dev *dev);
+struct mlx5_hrxq *mlx5_hrxq_new(struct rte_eth_dev *dev, uint8_t *rss_key,
+				uint8_t rss_key_len, uint64_t hash_fields,
+				uint16_t queues[], uint16_t queues_n);
+struct mlx5_hrxq *mlx5_hrxq_get(struct rte_eth_dev *dev, uint8_t *rss_key,
+				uint8_t rss_key_len, uint64_t hash_fields,
+				uint16_t queues[], uint16_t queues_n);
+int mlx5_hrxq_release(struct rte_eth_dev *dev, struct mlx5_hrxq *hxrq);
+int mlx5_hrxq_ibv_verify(struct rte_eth_dev *dev);
 
 /* mlx5_txq.c */
 
 int mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 			unsigned int socket, const struct rte_eth_txconf *conf);
 void mlx5_tx_queue_release(void *dpdk_txq);
-int priv_tx_uar_remap(struct priv *priv, int fd);
-struct mlx5_txq_ibv *mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx);
-struct mlx5_txq_ibv *mlx5_priv_txq_ibv_get(struct priv *priv, uint16_t idx);
-int mlx5_priv_txq_ibv_release(struct priv *priv, struct mlx5_txq_ibv *txq_ibv);
-int mlx5_priv_txq_ibv_releasable(struct priv *priv,
-				 struct mlx5_txq_ibv *txq_ibv);
-int mlx5_priv_txq_ibv_verify(struct priv *priv);
-struct mlx5_txq_ctrl *mlx5_priv_txq_new(struct priv *priv, uint16_t idx,
-					uint16_t desc, unsigned int socket,
-					const struct rte_eth_txconf *conf);
-struct mlx5_txq_ctrl *mlx5_priv_txq_get(struct priv *priv, uint16_t idx);
-int mlx5_priv_txq_release(struct priv *priv, uint16_t idx);
-int mlx5_priv_txq_releasable(struct priv *priv, uint16_t idx);
-int mlx5_priv_txq_verify(struct priv *priv);
+int mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd);
+struct mlx5_txq_ibv *mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx);
+struct mlx5_txq_ibv *mlx5_txq_ibv_get(struct rte_eth_dev *dev, uint16_t idx);
+int mlx5_txq_ibv_release(struct mlx5_txq_ibv *txq_ibv);
+int mlx5_txq_ibv_releasable(struct mlx5_txq_ibv *txq_ibv);
+int mlx5_txq_ibv_verify(struct rte_eth_dev *dev);
+struct mlx5_txq_ctrl *mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx,
+				   uint16_t desc, unsigned int socket,
+				   const struct rte_eth_txconf *conf);
+struct mlx5_txq_ctrl *mlx5_txq_get(struct rte_eth_dev *dev, uint16_t idx);
+int mlx5_txq_release(struct rte_eth_dev *dev, uint16_t idx);
+int mlx5_txq_releasable(struct rte_eth_dev *dev, uint16_t idx);
+int mlx5_txq_verify(struct rte_eth_dev *dev);
 void txq_alloc_elts(struct mlx5_txq_ctrl *txq_ctrl);
 
 /* mlx5_rxtx.c */
@@ -322,26 +319,22 @@ int mlx5_rx_descriptor_status(void *rx_queue, uint16_t offset);
 int mlx5_tx_descriptor_status(void *tx_queue, uint16_t offset);
 
 /* Vectorized version of mlx5_rxtx.c */
-
-int priv_check_raw_vec_tx_support(struct priv *priv);
-int priv_check_vec_tx_support(struct priv *priv);
-int rxq_check_vec_support(struct mlx5_rxq_data *rxq);
-int priv_check_vec_rx_support(struct priv *priv);
+int mlx5_check_raw_vec_tx_support(struct rte_eth_dev *dev);
+int mlx5_check_vec_tx_support(struct rte_eth_dev *dev);
+int mlx5_rxq_check_vec_support(struct mlx5_rxq_data *rxq_data);
+int mlx5_check_vec_rx_support(struct rte_eth_dev *dev);
 uint16_t mlx5_tx_burst_raw_vec(void *dpdk_txq, struct rte_mbuf **pkts,
 			       uint16_t pkts_n);
 uint16_t mlx5_tx_burst_vec(void *dpdk_txq, struct rte_mbuf **pkts,
 			   uint16_t pkts_n);
-uint16_t mlx5_rx_burst_vec(void *dpdk_rxq, struct rte_mbuf **pkts,
+uint16_t mlx5_rx_burst_vec(void *dpdk_txq, struct rte_mbuf **pkts,
 			   uint16_t pkts_n);
 
 /* mlx5_mr.c */
 
 void mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg);
-struct mlx5_mr *priv_txq_mp2mr_reg(struct priv *priv, struct mlx5_txq_data *txq,
-				   struct rte_mempool *mp, unsigned int idx);
 struct mlx5_mr *mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq,
-				   struct rte_mempool *mp,
-				   unsigned int idx);
+				   struct rte_mempool *mp, unsigned int idx);
 
 #ifndef NDEBUG
 /**
diff --git a/drivers/net/mlx5/mlx5_rxtx_vec.c b/drivers/net/mlx5/mlx5_rxtx_vec.c
index 101aa1567..982b8f1fe 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec.c
+++ b/drivers/net/mlx5/mlx5_rxtx_vec.c
@@ -276,15 +276,16 @@ mlx5_rx_burst_vec(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
 /**
  * Check Tx queue flags are set for raw vectorized Tx.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   1 if supported, negative errno value if not.
  */
 int __attribute__((cold))
-priv_check_raw_vec_tx_support(struct priv *priv)
+mlx5_check_raw_vec_tx_support(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	uint16_t i;
 
 	/* All the configured queues should support. */
@@ -303,15 +304,17 @@ priv_check_raw_vec_tx_support(struct priv *priv)
 /**
  * Check a device can support vectorized TX.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   1 if supported, negative errno value if not.
  */
 int __attribute__((cold))
-priv_check_vec_tx_support(struct priv *priv)
+mlx5_check_vec_tx_support(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
+
 	if (!priv->tx_vec_en ||
 	    priv->txqs_n > MLX5_VPMD_MIN_TXQS ||
 	    priv->mps != MLX5_MPW_ENHANCED ||
@@ -330,7 +333,7 @@ priv_check_vec_tx_support(struct priv *priv)
  *   1 if supported, negative errno value if not.
  */
 int __attribute__((cold))
-rxq_check_vec_support(struct mlx5_rxq_data *rxq)
+mlx5_rxq_check_vec_support(struct mlx5_rxq_data *rxq)
 {
 	struct mlx5_rxq_ctrl *ctrl =
 		container_of(rxq, struct mlx5_rxq_ctrl, rxq);
@@ -343,15 +346,16 @@ rxq_check_vec_support(struct mlx5_rxq_data *rxq)
 /**
  * Check a device can support vectorized RX.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   1 if supported, negative errno value if not.
  */
 int __attribute__((cold))
-priv_check_vec_rx_support(struct priv *priv)
+mlx5_check_vec_rx_support(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	uint16_t i;
 
 	if (!priv->rx_vec_en)
@@ -362,7 +366,7 @@ priv_check_vec_rx_support(struct priv *priv)
 
 		if (!rxq)
 			continue;
-		if (rxq_check_vec_support(rxq) < 0)
+		if (mlx5_rxq_check_vec_support(rxq) < 0)
 			break;
 	}
 	if (i != priv->rxqs_n)
diff --git a/drivers/net/mlx5/mlx5_socket.c b/drivers/net/mlx5/mlx5_socket.c
index ddfdc9a8e..57089cb5e 100644
--- a/drivers/net/mlx5/mlx5_socket.c
+++ b/drivers/net/mlx5/mlx5_socket.c
@@ -45,15 +45,16 @@
 /**
  * Initialise the socket to communicate with the secondary process
  *
- * @param[in] priv
- *   Pointer to private structure.
+ * @param[in] dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   0 on success, errno value on failure.
  */
 int
-priv_socket_init(struct priv *priv)
+mlx5_socket_init(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct sockaddr_un sun = {
 		.sun_family = AF_UNIX,
 	};
@@ -103,15 +104,17 @@ priv_socket_init(struct priv *priv)
 /**
  * Un-Initialise the socket to communicate with the secondary process
  *
- * @param[in] priv
- *   Pointer to private structure.
+ * @param[in] dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   0 on success, errno value on failure.
  */
 int
-priv_socket_uninit(struct priv *priv)
+mlx5_socket_uninit(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
+
 	MKSTR(path, "/var/tmp/%s_%d", MLX5_DRIVER_NAME, priv->primary_socket);
 	claim_zero(close(priv->primary_socket));
 	priv->primary_socket = 0;
@@ -122,12 +125,13 @@ priv_socket_uninit(struct priv *priv)
 /**
  * Handle socket interrupts.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  */
 void
-priv_socket_handle(struct priv *priv)
+mlx5_socket_handle(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	int conn_sock;
 	int ret = 0;
 	struct cmsghdr *cmsg = NULL;
@@ -203,15 +207,16 @@ priv_socket_handle(struct priv *priv)
 /**
  * Connect to the primary process.
  *
- * @param[in] priv
- *   Pointer to private structure.
+ * @param[in] dev
+ *   Pointer to Ethernet structure.
  *
  * @return
  *   fd on success, negative errno value on failure.
  */
 int
-priv_socket_connect(struct priv *priv)
+mlx5_socket_connect(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct sockaddr_un sun = {
 		.sun_family = AF_UNIX,
 	};
diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index 785563227..6d454e5e8 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -134,8 +134,8 @@ static const unsigned int xstats_n = RTE_DIM(mlx5_counters_init);
 /**
  * Read device counters table.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param[out] stats
  *   Counters table output buffer.
  *
@@ -143,8 +143,9 @@ static const unsigned int xstats_n = RTE_DIM(mlx5_counters_init);
  *   0 on success and stats is filled, negative on error.
  */
 static int
-priv_read_dev_counters(struct priv *priv, uint64_t *stats)
+mlx5_read_dev_counters(struct rte_eth_dev *dev, uint64_t *stats)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
 	unsigned int i;
 	struct ifreq ifr;
@@ -155,7 +156,7 @@ priv_read_dev_counters(struct priv *priv, uint64_t *stats)
 	et_stats->cmd = ETHTOOL_GSTATS;
 	et_stats->n_stats = xstats_ctrl->stats_n;
 	ifr.ifr_data = (caddr_t)et_stats;
-	if (priv_ifreq(priv, SIOCETHTOOL, &ifr) != 0) {
+	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr) != 0) {
 		WARN("unable to read statistic values from device");
 		return -1;
 	}
@@ -185,20 +186,20 @@ priv_read_dev_counters(struct priv *priv, uint64_t *stats)
 /**
  * Query the number of statistics provided by ETHTOOL.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   Number of statistics on success, -1 on error.
  */
 static int
-priv_ethtool_get_stats_n(struct priv *priv) {
+mlx5_ethtool_get_stats_n(struct rte_eth_dev *dev) {
 	struct ethtool_drvinfo drvinfo;
 	struct ifreq ifr;
 
 	drvinfo.cmd = ETHTOOL_GDRVINFO;
 	ifr.ifr_data = (caddr_t)&drvinfo;
-	if (priv_ifreq(priv, SIOCETHTOOL, &ifr) != 0) {
+	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr) != 0) {
 		WARN("unable to query number of statistics");
 		return -1;
 	}
@@ -208,12 +209,13 @@ priv_ethtool_get_stats_n(struct priv *priv) {
 /**
  * Init the structures to read device counters.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  */
 void
-priv_xstats_init(struct priv *priv)
+mlx5_xstats_init(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
 	unsigned int i;
 	unsigned int j;
@@ -222,7 +224,7 @@ priv_xstats_init(struct priv *priv)
 	unsigned int dev_stats_n;
 	unsigned int str_sz;
 
-	dev_stats_n = priv_ethtool_get_stats_n(priv);
+	dev_stats_n = mlx5_ethtool_get_stats_n(dev);
 	if (dev_stats_n < 1) {
 		WARN("no extended statistics available");
 		return;
@@ -241,7 +243,7 @@ priv_xstats_init(struct priv *priv)
 	strings->string_set = ETH_SS_STATS;
 	strings->len = dev_stats_n;
 	ifr.ifr_data = (caddr_t)strings;
-	if (priv_ifreq(priv, SIOCETHTOOL, &ifr) != 0) {
+	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr) != 0) {
 		WARN("unable to get statistic names");
 		goto free;
 	}
@@ -270,61 +272,55 @@ priv_xstats_init(struct priv *priv)
 	}
 	/* Copy to base at first time. */
 	assert(xstats_n <= MLX5_MAX_XSTATS);
-	priv_read_dev_counters(priv, xstats_ctrl->base);
+	mlx5_read_dev_counters(dev, xstats_ctrl->base);
 free:
 	rte_free(strings);
 }
 
 /**
- * Get device extended statistics.
+ * DPDK callback to get extended device statistics.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param[out] stats
  *   Pointer to rte extended stats table.
+ * @param n
+ *   The size of the stats table.
  *
  * @return
  *   Number of extended stats on success and stats is filled,
  *   negative on error.
  */
-static int
-priv_xstats_get(struct priv *priv, struct rte_eth_xstat *stats)
+int
+mlx5_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *stats,
+		unsigned int n)
 {
-	struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
-	unsigned int n = xstats_n;
 	uint64_t counters[n];
+	int ret = 0;
 
-	if (priv_read_dev_counters(priv, counters) < 0)
-		return -1;
-	for (i = 0; i != xstats_n; ++i) {
-		stats[i].id = i;
-		stats[i].value = (counters[i] - xstats_ctrl->base[i]);
+	if (n >= xstats_n && stats) {
+		struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
+		int stats_n;
+
+		stats_n = mlx5_ethtool_get_stats_n(dev);
+		if (stats_n < 0)
+			return -1;
+		if (xstats_ctrl->stats_n != stats_n)
+			mlx5_xstats_init(dev);
+		ret = mlx5_read_dev_counters(dev, counters);
+		if (ret)
+			return ret;
+		for (i = 0; i != xstats_n; ++i) {
+			stats[i].id = i;
+			stats[i].value = (counters[i] - xstats_ctrl->base[i]);
+		}
 	}
 	return n;
 }
 
 /**
- * Reset device extended statistics.
- *
- * @param priv
- *   Pointer to private structure.
- */
-static void
-priv_xstats_reset(struct priv *priv)
-{
-	struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
-	unsigned int i;
-	unsigned int n = xstats_n;
-	uint64_t counters[n];
-
-	if (priv_read_dev_counters(priv, counters) < 0)
-		return;
-	for (i = 0; i != n; ++i)
-		xstats_ctrl->base[i] = counters[i];
-}
-
-/**
  * DPDK callback to get device statistics.
  *
  * @param dev
@@ -421,41 +417,6 @@ mlx5_stats_reset(struct rte_eth_dev *dev)
 }
 
 /**
- * DPDK callback to get extended device statistics.
- *
- * @param dev
- *   Pointer to Ethernet device structure.
- * @param[out] stats
- *   Stats table output buffer.
- * @param n
- *   The size of the stats table.
- *
- * @return
- *   Number of xstats on success, negative on failure.
- */
-int
-mlx5_xstats_get(struct rte_eth_dev *dev,
-		struct rte_eth_xstat *stats, unsigned int n)
-{
-	struct priv *priv = dev->data->dev_private;
-	int ret = xstats_n;
-
-	if (n >= xstats_n && stats) {
-		struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
-		int stats_n;
-
-		stats_n = priv_ethtool_get_stats_n(priv);
-		if (stats_n < 0) {
-			return -1;
-		}
-		if (xstats_ctrl->stats_n != stats_n)
-			priv_xstats_init(priv);
-		ret = priv_xstats_get(priv, stats);
-	}
-	return ret;
-}
-
-/**
  * DPDK callback to clear device extended statistics.
  *
  * @param dev
@@ -467,13 +428,19 @@ mlx5_xstats_reset(struct rte_eth_dev *dev)
 	struct priv *priv = dev->data->dev_private;
 	struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
 	int stats_n;
+	unsigned int i;
+	unsigned int n = xstats_n;
+	uint64_t counters[n];
 
-	stats_n = priv_ethtool_get_stats_n(priv);
+	stats_n = mlx5_ethtool_get_stats_n(dev);
 	if (stats_n < 0)
 		return;
 	if (xstats_ctrl->stats_n != stats_n)
-		priv_xstats_init(priv);
-	priv_xstats_reset(priv);
+		mlx5_xstats_init(dev);
+	if (mlx5_read_dev_counters(dev, counters) < 0)
+		return;
+	for (i = 0; i != n; ++i)
+		xstats_ctrl->base[i] = counters[i];
 }
 
 /**
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index 2ae532d92..1941586d3 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -48,12 +48,13 @@
  *   Pointer to Ethernet device structure.
  */
 static void
-priv_txq_stop(struct priv *priv)
+mlx5_txq_stop(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
 	for (i = 0; i != priv->txqs_n; ++i)
-		mlx5_priv_txq_release(priv, i);
+		mlx5_txq_release(dev, i);
 }
 
 /**
@@ -66,8 +67,9 @@ priv_txq_stop(struct priv *priv)
  *   0 on success, errno on error.
  */
 static int
-priv_txq_start(struct priv *priv)
+mlx5_txq_start(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 	int ret = 0;
 
@@ -75,28 +77,28 @@ priv_txq_start(struct priv *priv)
 	for (i = 0; i != priv->txqs_n; ++i) {
 		unsigned int idx = 0;
 		struct mlx5_mr *mr;
-		struct mlx5_txq_ctrl *txq_ctrl = mlx5_priv_txq_get(priv, i);
+		struct mlx5_txq_ctrl *txq_ctrl = mlx5_txq_get(dev, i);
 
 		if (!txq_ctrl)
 			continue;
 		LIST_FOREACH(mr, &priv->mr, next) {
-			priv_txq_mp2mr_reg(priv, &txq_ctrl->txq, mr->mp, idx++);
+			mlx5_txq_mp2mr_reg(&txq_ctrl->txq, mr->mp, idx++);
 			if (idx == MLX5_PMD_TX_MP_CACHE)
 				break;
 		}
 		txq_alloc_elts(txq_ctrl);
-		txq_ctrl->ibv = mlx5_priv_txq_ibv_new(priv, i);
+		txq_ctrl->ibv = mlx5_txq_ibv_new(dev, i);
 		if (!txq_ctrl->ibv) {
 			ret = ENOMEM;
 			goto error;
 		}
 	}
-	ret = priv_tx_uar_remap(priv, priv->ctx->cmd_fd);
+	ret = mlx5_tx_uar_remap(dev, priv->ctx->cmd_fd);
 	if (ret)
 		goto error;
 	return ret;
 error:
-	priv_txq_stop(priv);
+	mlx5_txq_stop(dev);
 	return ret;
 }
 
@@ -107,12 +109,13 @@ priv_txq_start(struct priv *priv)
  *   Pointer to Ethernet device structure.
  */
 static void
-priv_rxq_stop(struct priv *priv)
+mlx5_rxq_stop(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
 	for (i = 0; i != priv->rxqs_n; ++i)
-		mlx5_priv_rxq_release(priv, i);
+		mlx5_rxq_release(dev, i);
 }
 
 /**
@@ -125,20 +128,21 @@ priv_rxq_stop(struct priv *priv)
  *   0 on success, errno on error.
  */
 static int
-priv_rxq_start(struct priv *priv)
+mlx5_rxq_start(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 	int ret = 0;
 
 	for (i = 0; i != priv->rxqs_n; ++i) {
-		struct mlx5_rxq_ctrl *rxq_ctrl = mlx5_priv_rxq_get(priv, i);
+		struct mlx5_rxq_ctrl *rxq_ctrl = mlx5_rxq_get(dev, i);
 
 		if (!rxq_ctrl)
 			continue;
 		ret = rxq_alloc_elts(rxq_ctrl);
 		if (ret)
 			goto error;
-		rxq_ctrl->ibv = mlx5_priv_rxq_ibv_new(priv, i);
+		rxq_ctrl->ibv = mlx5_rxq_ibv_new(dev, i);
 		if (!rxq_ctrl->ibv) {
 			ret = ENOMEM;
 			goto error;
@@ -146,7 +150,7 @@ priv_rxq_start(struct priv *priv)
 	}
 	return -ret;
 error:
-	priv_rxq_stop(priv);
+	mlx5_rxq_stop(dev);
 	return -ret;
 }
 
@@ -169,7 +173,7 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 	int err;
 
 	dev->data->dev_started = 1;
-	err = priv_flow_create_drop_queue(priv);
+	err = mlx5_flow_create_drop_queue(dev);
 	if (err) {
 		ERROR("%p: Drop queue allocation failed: %s",
 		      (void *)dev, strerror(err));
@@ -177,46 +181,46 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 	}
 	DEBUG("%p: allocating and configuring hash RX queues", (void *)dev);
 	rte_mempool_walk(mlx5_mp2mr_iter, priv);
-	err = priv_txq_start(priv);
+	err = mlx5_txq_start(dev);
 	if (err) {
 		ERROR("%p: TXQ allocation failed: %s",
 		      (void *)dev, strerror(err));
 		goto error;
 	}
-	err = priv_rxq_start(priv);
+	err = mlx5_rxq_start(dev);
 	if (err) {
 		ERROR("%p: RXQ allocation failed: %s",
 		      (void *)dev, strerror(err));
 		goto error;
 	}
-	err = priv_rx_intr_vec_enable(priv);
+	err = mlx5_rx_intr_vec_enable(dev);
 	if (err) {
 		ERROR("%p: RX interrupt vector creation failed",
 		      (void *)priv);
 		goto error;
 	}
-	priv_xstats_init(priv);
+	mlx5_xstats_init(dev);
 	/* Update link status and Tx/Rx callbacks for the first time. */
 	memset(&dev->data->dev_link, 0, sizeof(struct rte_eth_link));
 	INFO("Forcing port %u link to be up", dev->data->port_id);
-	err = priv_force_link_status_change(priv, ETH_LINK_UP);
+	err = mlx5_force_link_status_change(dev, ETH_LINK_UP);
 	if (err) {
 		DEBUG("Failed to set port %u link to be up",
 		      dev->data->port_id);
 		goto error;
 	}
-	priv_dev_interrupt_handler_install(priv, dev);
+	mlx5_dev_interrupt_handler_install(dev);
 	return 0;
 error:
 	/* Rollback. */
 	dev->data->dev_started = 0;
 	for (mr = LIST_FIRST(&priv->mr); mr; mr = LIST_FIRST(&priv->mr))
-		priv_mr_release(priv, mr);
-	priv_flow_stop(priv, &priv->flows);
-	priv_dev_traffic_disable(priv, dev);
-	priv_txq_stop(priv);
-	priv_rxq_stop(priv);
-	priv_flow_delete_drop_queue(priv);
+		mlx5_mr_release(mr);
+	mlx5_flow_stop(dev, &priv->flows);
+	mlx5_traffic_disable(dev);
+	mlx5_txq_stop(dev);
+	mlx5_rxq_stop(dev);
+	mlx5_flow_delete_drop_queue(dev);
 	return err;
 }
 
@@ -241,21 +245,21 @@ mlx5_dev_stop(struct rte_eth_dev *dev)
 	rte_wmb();
 	usleep(1000 * priv->rxqs_n);
 	DEBUG("%p: cleaning up and destroying hash RX queues", (void *)dev);
-	priv_flow_stop(priv, &priv->flows);
-	priv_dev_traffic_disable(priv, dev);
-	priv_rx_intr_vec_disable(priv);
-	priv_dev_interrupt_handler_uninstall(priv, dev);
-	priv_txq_stop(priv);
-	priv_rxq_stop(priv);
+	mlx5_flow_stop(dev, &priv->flows);
+	mlx5_traffic_disable(dev);
+	mlx5_rx_intr_vec_disable(dev);
+	mlx5_dev_interrupt_handler_uninstall(dev);
+	mlx5_txq_stop(dev);
+	mlx5_rxq_stop(dev);
 	for (mr = LIST_FIRST(&priv->mr); mr; mr = LIST_FIRST(&priv->mr))
-		priv_mr_release(priv, mr);
-	priv_flow_delete_drop_queue(priv);
+		mlx5_mr_release(mr);
+	mlx5_flow_delete_drop_queue(dev);
 }
 
 /**
  * Enable traffic flows configured by control plane
  *
- * @param priv
+ * @param dev
  *   Pointer to Ethernet device private data.
  * @param dev
  *   Pointer to Ethernet device structure.
@@ -264,8 +268,9 @@ mlx5_dev_stop(struct rte_eth_dev *dev)
  *   0 on success.
  */
 int
-priv_dev_traffic_enable(struct priv *priv, struct rte_eth_dev *dev)
+mlx5_traffic_enable(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct rte_flow_item_eth bcast = {
 		.dst.addr_bytes = "\xff\xff\xff\xff\xff\xff",
 	};
@@ -383,40 +388,18 @@ priv_dev_traffic_enable(struct priv *priv, struct rte_eth_dev *dev)
 /**
  * Disable traffic flows configured by control plane
  *
- * @param priv
- *   Pointer to Ethernet device private data.
  * @param dev
- *   Pointer to Ethernet device structure.
- *
- * @return
- *   0 on success.
- */
-int
-priv_dev_traffic_disable(struct priv *priv,
-			 struct rte_eth_dev *dev __rte_unused)
-{
-	priv_flow_flush(priv, &priv->ctrl_flows);
-	return 0;
-}
-
-/**
- * Restart traffic flows configured by control plane
- *
- * @param priv
  *   Pointer to Ethernet device private data.
- * @param dev
- *   Pointer to Ethernet device structure.
  *
  * @return
  *   0 on success.
  */
 int
-priv_dev_traffic_restart(struct priv *priv, struct rte_eth_dev *dev)
+mlx5_traffic_disable(struct rte_eth_dev *dev)
 {
-	if (dev->data->dev_started) {
-		priv_dev_traffic_disable(priv, dev);
-		priv_dev_traffic_enable(priv, dev);
-	}
+	struct priv *priv = dev->data->dev_private;
+
+	mlx5_flow_list_flush(dev, &priv->ctrl_flows);
 	return 0;
 }
 
@@ -424,7 +407,7 @@ priv_dev_traffic_restart(struct priv *priv, struct rte_eth_dev *dev)
  * Restart traffic flows configured by control plane
  *
  * @param dev
- *   Pointer to Ethernet device structure.
+ *   Pointer to Ethernet device private data.
  *
  * @return
  *   0 on success.
@@ -432,8 +415,9 @@ priv_dev_traffic_restart(struct priv *priv, struct rte_eth_dev *dev)
 int
 mlx5_traffic_restart(struct rte_eth_dev *dev)
 {
-	struct priv *priv = dev->data->dev_private;
-
-	priv_dev_traffic_restart(priv, dev);
+	if (dev->data->dev_started) {
+		mlx5_traffic_disable(dev);
+		mlx5_traffic_enable(dev);
+	}
 	return 0;
 }
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index f3b3daecb..29bbc0278 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -162,14 +162,14 @@ mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		      (void *)dev, idx, priv->txqs_n);
 		return -EOVERFLOW;
 	}
-	if (!mlx5_priv_txq_releasable(priv, idx)) {
+	if (!mlx5_txq_releasable(dev, idx)) {
 		ret = EBUSY;
 		ERROR("%p: unable to release queue index %u",
 		      (void *)dev, idx);
 		goto out;
 	}
-	mlx5_priv_txq_release(priv, idx);
-	txq_ctrl = mlx5_priv_txq_new(priv, idx, desc, socket, conf);
+	mlx5_txq_release(dev, idx);
+	txq_ctrl = mlx5_txq_new(dev, idx, desc, socket, conf);
 	if (!txq_ctrl) {
 		ERROR("%p: unable to allocate queue index %u",
 		      (void *)dev, idx);
@@ -205,7 +205,7 @@ mlx5_tx_queue_release(void *dpdk_txq)
 		if ((*priv->txqs)[i] == txq) {
 			DEBUG("%p: removing TX queue %p from list",
 			      (void *)priv->dev, (void *)txq_ctrl);
-			mlx5_priv_txq_release(priv, i);
+			mlx5_txq_release(priv->dev, i);
 			break;
 		}
 }
@@ -216,8 +216,8 @@ mlx5_tx_queue_release(void *dpdk_txq)
  * Both primary and secondary process do mmap to make UAR address
  * aligned.
  *
- * @param[in] priv
- *   Pointer to private structure.
+ * @param[in] dev
+ *   Pointer to Ethernet device.
  * @param fd
  *   Verbs file descriptor to map UAR pages.
  *
@@ -225,8 +225,9 @@ mlx5_tx_queue_release(void *dpdk_txq)
  *   0 on success, errno value on failure.
  */
 int
-priv_tx_uar_remap(struct priv *priv, int fd)
+mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i, j;
 	uintptr_t pages[priv->txqs_n];
 	unsigned int pages_n = 0;
@@ -293,8 +294,8 @@ priv_tx_uar_remap(struct priv *priv, int fd)
 /**
  * Create the Tx queue Verbs object.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   Queue index in DPDK Rx queue array
  *
@@ -302,8 +303,9 @@ priv_tx_uar_remap(struct priv *priv, int fd)
  *   The Verbs object initialised if it can be created.
  */
 struct mlx5_txq_ibv *
-mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
+mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_txq_data *txq_data = (*priv->txqs)[idx];
 	struct mlx5_txq_ctrl *txq_ctrl =
 		container_of(txq_data, struct mlx5_txq_ctrl, txq);
@@ -450,7 +452,7 @@ mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
 		ERROR("Failed to retrieve UAR info, invalid libmlx5.so version");
 		goto error;
 	}
-	DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)dev,
 	      (void *)txq_ibv, rte_atomic32_read(&txq_ibv->refcnt));
 	LIST_INSERT_HEAD(&priv->txqsibv, txq_ibv, next);
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
@@ -467,8 +469,8 @@ mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
 /**
  * Get an Tx queue Verbs object.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   Queue index in DPDK Rx queue array
  *
@@ -476,8 +478,9 @@ mlx5_priv_txq_ibv_new(struct priv *priv, uint16_t idx)
  *   The Verbs object if it exists.
  */
 struct mlx5_txq_ibv *
-mlx5_priv_txq_ibv_get(struct priv *priv, uint16_t idx)
+mlx5_txq_ibv_get(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_txq_ctrl *txq_ctrl;
 
 	if (idx >= priv->txqs_n)
@@ -487,7 +490,7 @@ mlx5_priv_txq_ibv_get(struct priv *priv, uint16_t idx)
 	txq_ctrl = container_of((*priv->txqs)[idx], struct mlx5_txq_ctrl, txq);
 	if (txq_ctrl->ibv) {
 		rte_atomic32_inc(&txq_ctrl->ibv->refcnt);
-		DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)priv,
+		DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)dev,
 		      (void *)txq_ctrl->ibv,
 		      rte_atomic32_read(&txq_ctrl->ibv->refcnt));
 	}
@@ -497,8 +500,6 @@ mlx5_priv_txq_ibv_get(struct priv *priv, uint16_t idx)
 /**
  * Release an Tx verbs queue object.
  *
- * @param priv
- *   Pointer to private structure.
  * @param txq_ibv
  *   Verbs Tx queue object.
  *
@@ -506,11 +507,10 @@ mlx5_priv_txq_ibv_get(struct priv *priv, uint16_t idx)
  *   0 on success, errno on failure.
  */
 int
-mlx5_priv_txq_ibv_release(struct priv *priv __rte_unused,
-			  struct mlx5_txq_ibv *txq_ibv)
+mlx5_txq_ibv_release(struct mlx5_txq_ibv *txq_ibv)
 {
 	assert(txq_ibv);
-	DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)priv,
+	DEBUG("Verbs Tx queue %p: refcnt %d",
 	      (void *)txq_ibv, rte_atomic32_read(&txq_ibv->refcnt));
 	if (rte_atomic32_dec_and_test(&txq_ibv->refcnt)) {
 		claim_zero(ibv_destroy_qp(txq_ibv->qp));
@@ -525,14 +525,11 @@ mlx5_priv_txq_ibv_release(struct priv *priv __rte_unused,
 /**
  * Return true if a single reference exists on the object.
  *
- * @param priv
- *   Pointer to private structure.
  * @param txq_ibv
  *   Verbs Tx queue object.
  */
 int
-mlx5_priv_txq_ibv_releasable(struct priv *priv __rte_unused,
-			     struct mlx5_txq_ibv *txq_ibv)
+mlx5_txq_ibv_releasable(struct mlx5_txq_ibv *txq_ibv)
 {
 	assert(txq_ibv);
 	return (rte_atomic32_read(&txq_ibv->refcnt) == 1);
@@ -541,20 +538,21 @@ mlx5_priv_txq_ibv_releasable(struct priv *priv __rte_unused,
 /**
  * Verify the Verbs Tx queue list is empty
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   The number of object not released.
  */
 int
-mlx5_priv_txq_ibv_verify(struct priv *priv)
+mlx5_txq_ibv_verify(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	int ret = 0;
 	struct mlx5_txq_ibv *txq_ibv;
 
 	LIST_FOREACH(txq_ibv, &priv->txqsibv, next) {
-		DEBUG("%p: Verbs Tx queue %p still referenced", (void *)priv,
+		DEBUG("%p: Verbs Tx queue %p still referenced", (void *)dev,
 		      (void *)txq_ibv);
 		++ret;
 	}
@@ -564,8 +562,8 @@ mlx5_priv_txq_ibv_verify(struct priv *priv)
 /**
  * Create a DPDK Tx queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   TX queue index.
  * @param desc
@@ -579,10 +577,10 @@ mlx5_priv_txq_ibv_verify(struct priv *priv)
  *   A DPDK queue object on success.
  */
 struct mlx5_txq_ctrl *
-mlx5_priv_txq_new(struct priv *priv, uint16_t idx, uint16_t desc,
-		  unsigned int socket,
-		  const struct rte_eth_txconf *conf)
+mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
+	     unsigned int socket, const struct rte_eth_txconf *conf)
 {
+	struct priv *priv = dev->data->dev_private;
 	const unsigned int max_tso_inline =
 		((MLX5_MAX_TSO_HEADER + (RTE_CACHE_LINE_SIZE - 1)) /
 		 RTE_CACHE_LINE_SIZE);
@@ -674,7 +672,7 @@ mlx5_priv_txq_new(struct priv *priv, uint16_t idx, uint16_t desc,
 		(struct rte_mbuf *(*)[1 << tmpl->txq.elts_n])(tmpl + 1);
 	tmpl->txq.stats.idx = idx;
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("%p: Tx queue %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Tx queue %p: refcnt %d", (void *)dev,
 	      (void *)tmpl, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->txqsctrl, tmpl, next);
 	return tmpl;
@@ -683,8 +681,8 @@ mlx5_priv_txq_new(struct priv *priv, uint16_t idx, uint16_t desc,
 /**
  * Get a Tx queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   TX queue index.
  *
@@ -692,8 +690,9 @@ mlx5_priv_txq_new(struct priv *priv, uint16_t idx, uint16_t desc,
  *   A pointer to the queue if it exists.
  */
 struct mlx5_txq_ctrl *
-mlx5_priv_txq_get(struct priv *priv, uint16_t idx)
+mlx5_txq_get(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_txq_ctrl *ctrl = NULL;
 
 	if ((*priv->txqs)[idx]) {
@@ -701,15 +700,15 @@ mlx5_priv_txq_get(struct priv *priv, uint16_t idx)
 				    txq);
 		unsigned int i;
 
-		mlx5_priv_txq_ibv_get(priv, idx);
+		mlx5_txq_ibv_get(dev, idx);
 		for (i = 0; i != MLX5_PMD_TX_MP_CACHE; ++i) {
 			if (ctrl->txq.mp2mr[i])
 				claim_nonzero
-					(priv_mr_get(priv,
+					(mlx5_mr_get(dev,
 						     ctrl->txq.mp2mr[i]->mp));
 		}
 		rte_atomic32_inc(&ctrl->refcnt);
-		DEBUG("%p: Tx queue %p: refcnt %d", (void *)priv,
+		DEBUG("%p: Tx queue %p: refcnt %d", (void *)dev,
 		      (void *)ctrl, rte_atomic32_read(&ctrl->refcnt));
 	}
 	return ctrl;
@@ -718,8 +717,8 @@ mlx5_priv_txq_get(struct priv *priv, uint16_t idx)
 /**
  * Release a Tx queue.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   TX queue index.
  *
@@ -727,8 +726,9 @@ mlx5_priv_txq_get(struct priv *priv, uint16_t idx)
  *   0 on success, errno on failure.
  */
 int
-mlx5_priv_txq_release(struct priv *priv, uint16_t idx)
+mlx5_txq_release(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 	struct mlx5_txq_ctrl *txq;
 	size_t page_size = sysconf(_SC_PAGESIZE);
@@ -736,18 +736,18 @@ mlx5_priv_txq_release(struct priv *priv, uint16_t idx)
 	if (!(*priv->txqs)[idx])
 		return 0;
 	txq = container_of((*priv->txqs)[idx], struct mlx5_txq_ctrl, txq);
-	DEBUG("%p: Tx queue %p: refcnt %d", (void *)priv,
+	DEBUG("%p: Tx queue %p: refcnt %d", (void *)dev,
 	      (void *)txq, rte_atomic32_read(&txq->refcnt));
 	if (txq->ibv) {
 		int ret;
 
-		ret = mlx5_priv_txq_ibv_release(priv, txq->ibv);
+		ret = mlx5_txq_ibv_release(txq->ibv);
 		if (!ret)
 			txq->ibv = NULL;
 	}
 	for (i = 0; i != MLX5_PMD_TX_MP_CACHE; ++i) {
 		if (txq->txq.mp2mr[i]) {
-			priv_mr_release(priv, txq->txq.mp2mr[i]);
+			mlx5_mr_release(txq->txq.mp2mr[i]);
 			txq->txq.mp2mr[i] = NULL;
 		}
 	}
@@ -767,8 +767,8 @@ mlx5_priv_txq_release(struct priv *priv, uint16_t idx)
 /**
  * Verify if the queue can be released.
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  * @param idx
  *   TX queue index.
  *
@@ -776,8 +776,9 @@ mlx5_priv_txq_release(struct priv *priv, uint16_t idx)
  *   1 if the queue can be released.
  */
 int
-mlx5_priv_txq_releasable(struct priv *priv, uint16_t idx)
+mlx5_txq_releasable(struct rte_eth_dev *dev, uint16_t idx)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_txq_ctrl *txq;
 
 	if (!(*priv->txqs)[idx])
@@ -789,20 +790,21 @@ mlx5_priv_txq_releasable(struct priv *priv, uint16_t idx)
 /**
  * Verify the Tx Queue list is empty
  *
- * @param priv
- *   Pointer to private structure.
+ * @param dev
+ *   Pointer to Ethernet device.
  *
  * @return
  *   The number of object not released.
  */
 int
-mlx5_priv_txq_verify(struct priv *priv)
+mlx5_txq_verify(struct rte_eth_dev *dev)
 {
+	struct priv *priv = dev->data->dev_private;
 	struct mlx5_txq_ctrl *txq;
 	int ret = 0;
 
 	LIST_FOREACH(txq, &priv->txqsctrl, next) {
-		DEBUG("%p: Tx Queue %p still referenced", (void *)priv,
+		DEBUG("%p: Tx Queue %p still referenced", (void *)dev,
 		      (void *)txq);
 		++ret;
 	}
diff --git a/drivers/net/mlx5/mlx5_vlan.c b/drivers/net/mlx5/mlx5_vlan.c
index 653c5fef4..7f070058e 100644
--- a/drivers/net/mlx5/mlx5_vlan.c
+++ b/drivers/net/mlx5/mlx5_vlan.c
@@ -96,25 +96,26 @@ mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
 		++priv->vlan_filter_n;
 	}
 	if (dev->data->dev_started)
-		priv_dev_traffic_restart(priv, dev);
+		mlx5_traffic_restart(dev);
 out:
 	return ret;
 }
 
 /**
- * Set/reset VLAN stripping for a specific queue.
+ * Callback to set/reset VLAN stripping for a specific queue.
  *
- * @param priv
- *   Pointer to private structure.
- * @param idx
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ * @param queue
  *   RX queue index.
  * @param on
  *   Enable/disable VLAN stripping.
  */
-static void
-priv_vlan_strip_queue_set(struct priv *priv, uint16_t idx, int on)
+void
+mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 {
-	struct mlx5_rxq_data *rxq = (*priv->rxqs)[idx];
+	struct priv *priv = dev->data->dev_private;
+	struct mlx5_rxq_data *rxq = (*priv->rxqs)[queue];
 	struct mlx5_rxq_ctrl *rxq_ctrl =
 		container_of(rxq, struct mlx5_rxq_ctrl, rxq);
 	struct ibv_wq_attr mod;
@@ -123,8 +124,18 @@ priv_vlan_strip_queue_set(struct priv *priv, uint16_t idx, int on)
 		0;
 	int err;
 
+	/* Validate hw support */
+	if (!priv->hw_vlan_strip) {
+		ERROR("VLAN stripping is not supported");
+		return;
+	}
+	/* Validate queue number */
+	if (queue >= priv->rxqs_n) {
+		ERROR("VLAN stripping, invalid queue number %d", queue);
+		return;
+	}
 	DEBUG("set VLAN offloads 0x%x for port %d queue %d",
-	      vlan_offloads, rxq->port_id, idx);
+	      vlan_offloads, rxq->port_id, queue);
 	if (!rxq_ctrl->ibv) {
 		/* Update related bits in RX queue. */
 		rxq->vlan_strip = !!on;
@@ -138,7 +149,7 @@ priv_vlan_strip_queue_set(struct priv *priv, uint16_t idx, int on)
 	err = ibv_modify_wq(rxq_ctrl->ibv->wq, &mod);
 	if (err) {
 		ERROR("%p: failed to modified stripping mode: %s",
-		      (void *)priv, strerror(err));
+		      (void *)dev, strerror(err));
 		return;
 	}
 	/* Update related bits in RX queue. */
@@ -146,34 +157,6 @@ priv_vlan_strip_queue_set(struct priv *priv, uint16_t idx, int on)
 }
 
 /**
- * Callback to set/reset VLAN stripping for a specific queue.
- *
- * @param dev
- *   Pointer to Ethernet device structure.
- * @param queue
- *   RX queue index.
- * @param on
- *   Enable/disable VLAN stripping.
- */
-void
-mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
-{
-	struct priv *priv = dev->data->dev_private;
-
-	/* Validate hw support */
-	if (!priv->hw_vlan_strip) {
-		ERROR("VLAN stripping is not supported");
-		return;
-	}
-	/* Validate queue number */
-	if (queue >= priv->rxqs_n) {
-		ERROR("VLAN stripping, invalid queue number %d", queue);
-		return;
-	}
-	priv_vlan_strip_queue_set(priv, queue, on);
-}
-
-/**
  * Callback to set/reset VLAN offloads for a port.
  *
  * @param dev
@@ -196,7 +179,7 @@ mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask)
 		}
 		/* Run on every RX queue and set/reset VLAN stripping. */
 		for (i = 0; (i != priv->rxqs_n); i++)
-			priv_vlan_strip_queue_set(priv, i, hw_vlan_strip);
+			mlx5_vlan_strip_queue_set(dev, i, hw_vlan_strip);
 	}
 	return 0;
 }
-- 
2.11.0

[dpdk-stable] [PATCH v2 26/67] net/mlx5: change non failing function return values

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit 925061b58b487fba57f55847b1447417fed715fb ]

These functions return int although they are not supposed to fail,
resulting in unnecessary checks in their callers.
Some are returning error where is should be a boolean.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.h         |  4 ++--
 drivers/net/mlx5/mlx5_mr.c      |  4 ++--
 drivers/net/mlx5/mlx5_rxq.c     | 25 ++++++++++---------------
 drivers/net/mlx5/mlx5_socket.c  |  6 +-----
 drivers/net/mlx5/mlx5_trigger.c |  6 +-----
 drivers/net/mlx5/mlx5_txq.c     | 17 ++++++-----------
 6 files changed, 22 insertions(+), 40 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 5814ec051..8ecee0212 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -255,7 +255,7 @@ int mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask);
 int mlx5_dev_start(struct rte_eth_dev *dev);
 void mlx5_dev_stop(struct rte_eth_dev *dev);
 int mlx5_traffic_enable(struct rte_eth_dev *dev);
-int mlx5_traffic_disable(struct rte_eth_dev *dev);
+void mlx5_traffic_disable(struct rte_eth_dev *dev);
 int mlx5_traffic_restart(struct rte_eth_dev *dev);
 
 /* mlx5_flow.c */
@@ -300,7 +300,7 @@ void mlx5_flow_delete_drop_queue(struct rte_eth_dev *dev);
 /* mlx5_socket.c */
 
 int mlx5_socket_init(struct rte_eth_dev *priv);
-int mlx5_socket_uninit(struct rte_eth_dev *priv);
+void mlx5_socket_uninit(struct rte_eth_dev *priv);
 void mlx5_socket_handle(struct rte_eth_dev *priv);
 int mlx5_socket_connect(struct rte_eth_dev *priv);
 
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 6f60aa1c5..cfad4798b 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -336,7 +336,7 @@ mlx5_mr_get(struct rte_eth_dev *dev, struct rte_mempool *mp)
  *   Pointer to memory region to release.
  *
  * @return
- *   0 on success, errno on failure.
+ *   1 while a reference on it exists, 0 when freed.
  */
 int
 mlx5_mr_release(struct mlx5_mr *mr)
@@ -350,7 +350,7 @@ mlx5_mr_release(struct mlx5_mr *mr)
 		rte_free(mr);
 		return 0;
 	}
-	return EBUSY;
+	return 1;
 }
 
 /**
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 637092d44..cc1d7ba5d 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -787,7 +787,7 @@ mlx5_rxq_ibv_get(struct rte_eth_dev *dev, uint16_t idx)
  *   Verbs Rx queue object.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   1 while a reference on it exists, 0 when freed.
  */
 int
 mlx5_rxq_ibv_release(struct mlx5_rxq_ibv *rxq_ibv)
@@ -813,7 +813,7 @@ mlx5_rxq_ibv_release(struct mlx5_rxq_ibv *rxq_ibv)
 		rte_free(rxq_ibv);
 		return 0;
 	}
-	return EBUSY;
+	return 1;
 }
 
 /**
@@ -1018,7 +1018,7 @@ mlx5_rxq_get(struct rte_eth_dev *dev, uint16_t idx)
  *   TX queue index.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   1 while a reference on it exists, 0 when freed.
  */
 int
 mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx)
@@ -1030,13 +1030,8 @@ mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx)
 		return 0;
 	rxq_ctrl = container_of((*priv->rxqs)[idx], struct mlx5_rxq_ctrl, rxq);
 	assert(rxq_ctrl->priv);
-	if (rxq_ctrl->ibv) {
-		int ret;
-
-		ret = mlx5_rxq_ibv_release(rxq_ctrl->ibv);
-		if (!ret)
-			rxq_ctrl->ibv = NULL;
-	}
+	if (rxq_ctrl->ibv && !mlx5_rxq_ibv_release(rxq_ctrl->ibv))
+		rxq_ctrl->ibv = NULL;
 	DEBUG("%p: Rx queue %p: refcnt %d", (void *)dev,
 	      (void *)rxq_ctrl, rte_atomic32_read(&rxq_ctrl->refcnt));
 	if (rte_atomic32_dec_and_test(&rxq_ctrl->refcnt)) {
@@ -1045,7 +1040,7 @@ mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx)
 		(*priv->rxqs)[idx] = NULL;
 		return 0;
 	}
-	return EBUSY;
+	return 1;
 }
 
 /**
@@ -1205,7 +1200,7 @@ mlx5_ind_table_ibv_get(struct rte_eth_dev *dev, uint16_t queues[],
  *   Indirection table to release.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   1 while a reference on it exists, 0 when freed.
  */
 int
 mlx5_ind_table_ibv_release(struct rte_eth_dev *dev,
@@ -1224,7 +1219,7 @@ mlx5_ind_table_ibv_release(struct rte_eth_dev *dev,
 		rte_free(ind_tbl);
 		return 0;
 	}
-	return EBUSY;
+	return 1;
 }
 
 /**
@@ -1382,7 +1377,7 @@ mlx5_hrxq_get(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
  *   Pointer to Hash Rx queue to release.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   1 while a reference on it exists, 0 when freed.
  */
 int
 mlx5_hrxq_release(struct rte_eth_dev *dev, struct mlx5_hrxq *hrxq)
@@ -1397,7 +1392,7 @@ mlx5_hrxq_release(struct rte_eth_dev *dev, struct mlx5_hrxq *hrxq)
 		return 0;
 	}
 	claim_nonzero(mlx5_ind_table_ibv_release(dev, hrxq->ind_table));
-	return EBUSY;
+	return 1;
 }
 
 /**
diff --git a/drivers/net/mlx5/mlx5_socket.c b/drivers/net/mlx5/mlx5_socket.c
index 57089cb5e..8f400d06e 100644
--- a/drivers/net/mlx5/mlx5_socket.c
+++ b/drivers/net/mlx5/mlx5_socket.c
@@ -106,11 +106,8 @@ mlx5_socket_init(struct rte_eth_dev *dev)
  *
  * @param[in] dev
  *   Pointer to Ethernet device.
- *
- * @return
- *   0 on success, errno value on failure.
  */
-int
+void
 mlx5_socket_uninit(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
@@ -119,7 +116,6 @@ mlx5_socket_uninit(struct rte_eth_dev *dev)
 	claim_zero(close(priv->primary_socket));
 	priv->primary_socket = 0;
 	claim_zero(remove(path));
-	return 0;
 }
 
 /**
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index 1941586d3..19434b921 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -390,17 +390,13 @@ mlx5_traffic_enable(struct rte_eth_dev *dev)
  *
  * @param dev
  *   Pointer to Ethernet device private data.
- *
- * @return
- *   0 on success.
  */
-int
+void
 mlx5_traffic_disable(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 
 	mlx5_flow_list_flush(dev, &priv->ctrl_flows);
-	return 0;
 }
 
 /**
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 29bbc0278..53a21c259 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -504,7 +504,7 @@ mlx5_txq_ibv_get(struct rte_eth_dev *dev, uint16_t idx)
  *   Verbs Tx queue object.
  *
  * @return
- *   0 on success, errno on failure.
+ *   1 while a reference on it exists, 0 when freed.
  */
 int
 mlx5_txq_ibv_release(struct mlx5_txq_ibv *txq_ibv)
@@ -519,7 +519,7 @@ mlx5_txq_ibv_release(struct mlx5_txq_ibv *txq_ibv)
 		rte_free(txq_ibv);
 		return 0;
 	}
-	return EBUSY;
+	return 1;
 }
 
 /**
@@ -723,7 +723,7 @@ mlx5_txq_get(struct rte_eth_dev *dev, uint16_t idx)
  *   TX queue index.
  *
  * @return
- *   0 on success, errno on failure.
+ *   1 while a reference on it exists, 0 when freed.
  */
 int
 mlx5_txq_release(struct rte_eth_dev *dev, uint16_t idx)
@@ -738,13 +738,8 @@ mlx5_txq_release(struct rte_eth_dev *dev, uint16_t idx)
 	txq = container_of((*priv->txqs)[idx], struct mlx5_txq_ctrl, txq);
 	DEBUG("%p: Tx queue %p: refcnt %d", (void *)dev,
 	      (void *)txq, rte_atomic32_read(&txq->refcnt));
-	if (txq->ibv) {
-		int ret;
-
-		ret = mlx5_txq_ibv_release(txq->ibv);
-		if (!ret)
-			txq->ibv = NULL;
-	}
+	if (txq->ibv && !mlx5_txq_ibv_release(txq->ibv))
+		txq->ibv = NULL;
 	for (i = 0; i != MLX5_PMD_TX_MP_CACHE; ++i) {
 		if (txq->txq.mp2mr[i]) {
 			mlx5_mr_release(txq->txq.mp2mr[i]);
@@ -761,7 +756,7 @@ mlx5_txq_release(struct rte_eth_dev *dev, uint16_t idx)
 		(*priv->txqs)[idx] = NULL;
 		return 0;
 	}
-	return EBUSY;
+	return 1;
 }
 
 /**
-- 
2.11.0

[dpdk-stable] [PATCH v2 27/67] net/mlx5: standardize on negative errno values

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit a6d83b6a9209a198fa5a7d2f9cbb37190e256f9c ]

Set rte_errno systematically as well.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.c         |  88 ++++++-----
 drivers/net/mlx5/mlx5_ethdev.c  | 225 ++++++++++++++++------------
 drivers/net/mlx5/mlx5_flow.c    | 317 +++++++++++++++++++++++-----------------
 drivers/net/mlx5/mlx5_mac.c     |  33 +++--
 drivers/net/mlx5/mlx5_mr.c      |  15 +-
 drivers/net/mlx5/mlx5_rss.c     |  50 ++++---
 drivers/net/mlx5/mlx5_rxmode.c  |  28 +++-
 drivers/net/mlx5/mlx5_rxq.c     | 138 ++++++++++-------
 drivers/net/mlx5/mlx5_socket.c  |  82 +++++++----
 drivers/net/mlx5/mlx5_stats.c   |  53 +++++--
 drivers/net/mlx5/mlx5_trigger.c |  89 ++++++-----
 drivers/net/mlx5/mlx5_txq.c     |  50 ++++---
 drivers/net/mlx5/mlx5_vlan.c    |  24 +--
 13 files changed, 711 insertions(+), 481 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index ebb778826..9319effcb 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -150,7 +150,7 @@ mlx5_getenv_int(const char *name)
  *   A pointer to the callback data.
  *
  * @return
- *   a pointer to the allocate space.
+ *   Allocated buffer, NULL otherwise and rte_errno is set.
  */
 static void *
 mlx5_alloc_verbs_buf(size_t size, void *data)
@@ -172,6 +172,8 @@ mlx5_alloc_verbs_buf(size_t size, void *data)
 	}
 	assert(data != NULL);
 	ret = rte_malloc_socket(__func__, size, alignment, socket);
+	if (!ret && size)
+		rte_errno = ENOMEM;
 	DEBUG("Extern alloc size: %lu, align: %lu: %p", size, alignment, ret);
 	return ret;
 }
@@ -405,7 +407,7 @@ mlx5_dev_idx(struct rte_pci_addr *pci_addr)
  *   User data.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_args_check(const char *key, const char *val, void *opaque)
@@ -416,8 +418,9 @@ mlx5_args_check(const char *key, const char *val, void *opaque)
 	errno = 0;
 	tmp = strtoul(val, NULL, 0);
 	if (errno) {
+		rte_errno = errno;
 		WARN("%s: \"%s\" is not a valid integer", key, val);
-		return errno;
+		return -rte_errno;
 	}
 	if (strcmp(MLX5_RXQ_CQE_COMP_EN, key) == 0) {
 		args->cqe_comp = !!tmp;
@@ -439,7 +442,8 @@ mlx5_args_check(const char *key, const char *val, void *opaque)
 		args->rx_vec_en = !!tmp;
 	} else {
 		WARN("%s: unknown parameter", key);
-		return -EINVAL;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	return 0;
 }
@@ -453,7 +457,7 @@ mlx5_args_check(const char *key, const char *val, void *opaque)
  *   Device arguments structure.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_args(struct mlx5_args *args, struct rte_devargs *devargs)
@@ -485,9 +489,10 @@ mlx5_args(struct mlx5_args *args, struct rte_devargs *devargs)
 		if (rte_kvargs_count(kvlist, params[i])) {
 			ret = rte_kvargs_process(kvlist, params[i],
 						 mlx5_args_check, args);
-			if (ret != 0) {
+			if (ret) {
+				rte_errno = EINVAL;
 				rte_kvargs_free(kvlist);
-				return ret;
+				return -rte_errno;
 			}
 		}
 	}
@@ -513,7 +518,7 @@ static void *uar_base;
  *   Pointer to Ethernet device.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_uar_init_primary(struct rte_eth_dev *dev)
@@ -522,7 +527,6 @@ mlx5_uar_init_primary(struct rte_eth_dev *dev)
 	void *addr = (void *)0;
 	int i;
 	const struct rte_mem_config *mcfg;
-	int ret;
 
 	if (uar_base) { /* UAR address space mapped. */
 		priv->uar_base = uar_base;
@@ -544,8 +548,8 @@ mlx5_uar_init_primary(struct rte_eth_dev *dev)
 	if (addr == MAP_FAILED) {
 		ERROR("Failed to reserve UAR address space, please adjust "
 		      "MLX5_UAR_SIZE or try --base-virtaddr");
-		ret = ENOMEM;
-		return ret;
+		rte_errno = ENOMEM;
+		return -rte_errno;
 	}
 	/* Accept either same addr or a new addr returned from mmap if target
 	 * range occupied.
@@ -564,14 +568,13 @@ mlx5_uar_init_primary(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_uar_init_secondary(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 	void *addr;
-	int ret;
 
 	assert(priv->uar_base);
 	if (uar_base) { /* already reserved. */
@@ -584,15 +587,15 @@ mlx5_uar_init_secondary(struct rte_eth_dev *dev)
 	if (addr == MAP_FAILED) {
 		ERROR("UAR mmap failed: %p size: %llu",
 		      priv->uar_base, MLX5_UAR_SIZE);
-		ret = ENXIO;
-		return ret;
+		rte_errno = ENXIO;
+		return -rte_errno;
 	}
 	if (priv->uar_base != addr) {
 		ERROR("UAR address %p size %llu occupied, please adjust "
 		      "MLX5_UAR_OFFSET or try EAL parameter --base-virtaddr",
 		      priv->uar_base, MLX5_UAR_SIZE);
-		ret = ENXIO;
-		return ret;
+		rte_errno = ENXIO;
+		return -rte_errno;
 	}
 	uar_base = addr; /* process local, don't reserve again */
 	INFO("Reserved UAR address space: %p", addr);
@@ -643,13 +646,13 @@ mlx5_args_assign(struct priv *priv, struct mlx5_args *args)
  *   PCI device information.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	       struct rte_pci_device *pci_dev)
 {
-	struct ibv_device **list;
+	struct ibv_device **list = NULL;
 	struct ibv_device *ibv_dev;
 	int err = 0;
 	struct ibv_context *attr_ctx = NULL;
@@ -669,7 +672,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	idx = mlx5_dev_idx(&pci_dev->addr);
 	if (idx == -1) {
 		ERROR("this driver cannot support any more adapters");
-		return -ENOMEM;
+		err = ENOMEM;
+		goto error;
 	}
 	DEBUG("using driver device index %d", idx);
 	/* Save PCI address. */
@@ -677,9 +681,10 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	list = ibv_get_device_list(&i);
 	if (list == NULL) {
 		assert(errno);
+		err = errno;
 		if (errno == ENOSYS)
 			ERROR("cannot list devices, is ib_uverbs loaded?");
-		return -errno;
+		goto error;
 	}
 	assert(i >= 0);
 	/*
@@ -715,7 +720,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		INFO("PCI information matches, using device \"%s\"",
 		     list[i]->name);
 		attr_ctx = ibv_open_device(list[i]);
-		err = errno;
+		rte_errno = errno;
+		err = rte_errno;
 		break;
 	}
 	if (attr_ctx == NULL) {
@@ -723,13 +729,12 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		switch (err) {
 		case 0:
 			ERROR("cannot access device, is mlx5_ib loaded?");
-			return -ENODEV;
+			err = ENODEV;
+			goto error;
 		case EINVAL:
 			ERROR("cannot use device, are drivers up to date?");
-			return -EINVAL;
+			goto error;
 		}
-		assert(err > 0);
-		return -err;
 	}
 	ibv_dev = list[i];
 	DEBUG("device opened");
@@ -755,8 +760,10 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		cqe_comp = 0;
 	else
 		cqe_comp = 1;
-	if (ibv_query_device_ex(attr_ctx, NULL, &device_attr))
+	if (ibv_query_device_ex(attr_ctx, NULL, &device_attr)) {
+		err = errno;
 		goto error;
+	}
 	INFO("%u port(s) detected", device_attr.orig_attr.phys_port_cnt);
 	for (i = 0; i < device_attr.orig_attr.phys_port_cnt; i++) {
 		char name[RTE_ETH_NAME_MAX_LEN];
@@ -790,22 +797,19 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			eth_dev = rte_eth_dev_attach_secondary(name);
 			if (eth_dev == NULL) {
 				ERROR("can not attach rte ethdev");
-				err = ENOMEM;
+				rte_errno = ENOMEM;
+				err = rte_errno;
 				goto error;
 			}
 			eth_dev->device = &pci_dev->device;
 			eth_dev->dev_ops = &mlx5_dev_sec_ops;
 			err = mlx5_uar_init_secondary(eth_dev);
-			if (err < 0) {
-				err = -err;
+			if (err)
 				goto error;
-			}
 			/* Receive command fd from primary process */
 			err = mlx5_socket_connect(eth_dev);
-			if (err < 0) {
-				err = -err;
+			if (err)
 				goto error;
-			}
 			/* Remap UAR for Tx queues. */
 			err = mlx5_tx_uar_remap(eth_dev, err);
 			if (err)
@@ -876,6 +880,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		mlx5_args_assign(priv, &args);
 		if (ibv_query_device_ex(ctx, NULL, &device_attr_ex)) {
 			ERROR("ibv_query_device_ex() failed");
+			err = errno;
 			goto port_error;
 		}
 		priv->hw_csum =
@@ -996,7 +1001,9 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		}
 #endif
 		/* Get actual MTU if possible. */
-		mlx5_get_mtu(eth_dev, &priv->mtu);
+		err = mlx5_get_mtu(eth_dev, &priv->mtu);
+		if (err)
+			goto port_error;
 		DEBUG("port %u MTU is %u", priv->port, priv->mtu);
 		/*
 		 * Initialize burst functions to prevent crashes before link-up.
@@ -1037,16 +1044,19 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	 */
 	/* no port found, complain */
 	if (!mlx5_dev[idx].ports) {
-		err = ENODEV;
-		goto error;
+		rte_errno = ENODEV;
+		err = rte_errno;
 	}
 error:
 	if (attr_ctx)
 		claim_zero(ibv_close_device(attr_ctx));
 	if (list)
 		ibv_free_device_list(list);
-	assert(err >= 0);
-	return -err;
+	if (err) {
+		rte_errno = err;
+		return -rte_errno;
+	}
+	return 0;
 }
 
 static const struct rte_pci_id mlx5_pci_id_map[] = {
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 3435bf338..d0be35570 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -128,7 +128,7 @@ struct ethtool_link_settings {
  *   Interface name output buffer.
  *
  * @return
- *   0 on success, -1 on failure and errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_get_ifname(const struct rte_eth_dev *dev, char (*ifname)[IF_NAMESIZE])
@@ -144,8 +144,10 @@ mlx5_get_ifname(const struct rte_eth_dev *dev, char (*ifname)[IF_NAMESIZE])
 		MKSTR(path, "%s/device/net", priv->ibdev_path);
 
 		dir = opendir(path);
-		if (dir == NULL)
-			return -1;
+		if (dir == NULL) {
+			rte_errno = errno;
+			return -rte_errno;
+		}
 	}
 	while ((dent = readdir(dir)) != NULL) {
 		char *name = dent->d_name;
@@ -195,8 +197,10 @@ mlx5_get_ifname(const struct rte_eth_dev *dev, char (*ifname)[IF_NAMESIZE])
 			snprintf(match, sizeof(match), "%s", name);
 	}
 	closedir(dir);
-	if (match[0] == '\0')
-		return -1;
+	if (match[0] == '\0') {
+		rte_errno = ENOENT;
+		return -rte_errno;
+	}
 	strncpy(*ifname, match, sizeof(*ifname));
 	return 0;
 }
@@ -212,20 +216,31 @@ mlx5_get_ifname(const struct rte_eth_dev *dev, char (*ifname)[IF_NAMESIZE])
  *   Interface request structure output buffer.
  *
  * @return
- *   0 on success, -1 on failure and errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_ifreq(const struct rte_eth_dev *dev, int req, struct ifreq *ifr)
 {
 	int sock = socket(PF_INET, SOCK_DGRAM, IPPROTO_IP);
-	int ret = -1;
+	int ret = 0;
 
-	if (sock == -1)
-		return ret;
-	if (mlx5_get_ifname(dev, &ifr->ifr_name) == 0)
-		ret = ioctl(sock, req, ifr);
+	if (sock == -1) {
+		rte_errno = errno;
+		return -rte_errno;
+	}
+	ret = mlx5_get_ifname(dev, &ifr->ifr_name);
+	if (ret)
+		goto error;
+	ret = ioctl(sock, req, ifr);
+	if (ret == -1) {
+		rte_errno = errno;
+		goto error;
+	}
 	close(sock);
-	return ret;
+	return 0;
+error:
+	close(sock);
+	return -rte_errno;
 }
 
 /**
@@ -237,7 +252,7 @@ mlx5_ifreq(const struct rte_eth_dev *dev, int req, struct ifreq *ifr)
  *   MTU value output buffer.
  *
  * @return
- *   0 on success, -1 on failure and errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_get_mtu(struct rte_eth_dev *dev, uint16_t *mtu)
@@ -260,7 +275,7 @@ mlx5_get_mtu(struct rte_eth_dev *dev, uint16_t *mtu)
  *   MTU value to set.
  *
  * @return
- *   0 on success, -1 on failure and errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
@@ -281,7 +296,7 @@ mlx5_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
  *   Bitmask for flags to modify.
  *
  * @return
- *   0 on success, -1 on failure and errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_set_flags(struct rte_eth_dev *dev, unsigned int keep, unsigned int flags)
@@ -303,7 +318,7 @@ mlx5_set_flags(struct rte_eth_dev *dev, unsigned int keep, unsigned int flags)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_dev_configure(struct rte_eth_dev *dev)
@@ -316,19 +331,22 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 	unsigned int reta_idx_n;
 	const uint8_t use_app_rss_key =
 		!!dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key;
+	int ret = 0;
 
 	if (use_app_rss_key &&
 	    (dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key_len !=
 	     rss_hash_default_key_len)) {
 		/* MLX5 RSS only support 40bytes key. */
-		return EINVAL;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	priv->rss_conf.rss_key =
 		rte_realloc(priv->rss_conf.rss_key,
 			    rss_hash_default_key_len, 0);
 	if (!priv->rss_conf.rss_key) {
 		ERROR("cannot allocate RSS hash key memory (%u)", rxqs_n);
-		return ENOMEM;
+		rte_errno = ENOMEM;
+		return -rte_errno;
 	}
 	memcpy(priv->rss_conf.rss_key,
 	       use_app_rss_key ?
@@ -346,7 +364,8 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 	}
 	if (rxqs_n > priv->ind_table_max_size) {
 		ERROR("cannot handle this many RX queues (%u)", rxqs_n);
-		return EINVAL;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	if (rxqs_n == priv->rxqs_n)
 		return 0;
@@ -359,8 +378,9 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 	reta_idx_n = (1 << log2above((rxqs_n & (rxqs_n - 1)) ?
 				     priv->ind_table_max_size :
 				     rxqs_n));
-	if (mlx5_rss_reta_index_resize(dev, reta_idx_n))
-		return ENOMEM;
+	ret = mlx5_rss_reta_index_resize(dev, reta_idx_n);
+	if (ret)
+		return ret;
 	/* When the number of RX queues is not a power of two, the remaining
 	 * table entries are padded with reused WQs and hashes are not spread
 	 * uniformly. */
@@ -370,7 +390,6 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 			j = 0;
 	}
 	return 0;
-
 }
 
 /**
@@ -478,7 +497,7 @@ mlx5_dev_supported_ptypes_get(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, -1 on error.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
@@ -490,19 +509,22 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 	struct ifreq ifr;
 	struct rte_eth_link dev_link;
 	int link_speed = 0;
+	int ret;
 
-	if (mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr)) {
-		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
-		return -1;
+	ret = mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr);
+	if (ret) {
+		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(rte_errno));
+		return ret;
 	}
 	memset(&dev_link, 0, sizeof(dev_link));
 	dev_link.link_status = ((ifr.ifr_flags & IFF_UP) &&
 				(ifr.ifr_flags & IFF_RUNNING));
 	ifr.ifr_data = (void *)&edata;
-	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
+	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+	if (ret) {
 		WARN("ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: %s",
-		     strerror(errno));
-		return -1;
+		     strerror(rte_errno));
+		return ret;
 	}
 	link_speed = ethtool_cmd_speed(&edata);
 	if (link_speed == -1)
@@ -532,7 +554,8 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 		return 0;
 	}
 	/* Link status is still the same. */
-	return -1;
+	rte_errno = EAGAIN;
+	return -rte_errno;
 }
 
 /**
@@ -542,7 +565,7 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, -1 on error.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
@@ -552,19 +575,22 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 	struct ifreq ifr;
 	struct rte_eth_link dev_link;
 	uint64_t sc;
+	int ret;
 
-	if (mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr)) {
-		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
-		return -1;
+	ret = mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr);
+	if (ret) {
+		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(rte_errno));
+		return ret;
 	}
 	memset(&dev_link, 0, sizeof(dev_link));
 	dev_link.link_status = ((ifr.ifr_flags & IFF_UP) &&
 				(ifr.ifr_flags & IFF_RUNNING));
 	ifr.ifr_data = (void *)&gcmd;
-	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
+	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+	if (ret) {
 		DEBUG("ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS) failed: %s",
-		      strerror(errno));
-		return -1;
+		      strerror(rte_errno));
+		return ret;
 	}
 	gcmd.link_mode_masks_nwords = -gcmd.link_mode_masks_nwords;
 
@@ -575,10 +601,11 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 
 	*ecmd = gcmd;
 	ifr.ifr_data = (void *)ecmd;
-	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
+	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+	if (ret) {
 		DEBUG("ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS) failed: %s",
-		      strerror(errno));
-		return -1;
+		      strerror(rte_errno));
+		return ret;
 	}
 	dev_link.link_speed = ecmd->speed;
 	sc = ecmd->link_mode_masks[0] |
@@ -628,7 +655,8 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 		return 0;
 	}
 	/* Link status is still the same. */
-	return -1;
+	rte_errno = EAGAIN;
+	return -rte_errno;
 }
 
 /**
@@ -641,18 +669,21 @@ static void
 mlx5_link_start(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
-	int err;
+	int ret;
 
 	mlx5_select_tx_function(dev);
 	mlx5_select_rx_function(dev);
-	err = mlx5_traffic_enable(dev);
-	if (err)
+	ret = mlx5_traffic_enable(dev);
+	if (ret) {
 		ERROR("%p: error occurred while configuring control flows: %s",
-		      (void *)dev, strerror(err));
-	err = mlx5_flow_start(dev, &priv->flows);
-	if (err)
+		      (void *)dev, strerror(rte_errno));
+		return;
+	}
+	ret = mlx5_flow_start(dev, &priv->flows);
+	if (ret) {
 		ERROR("%p: error occurred while configuring flows: %s",
-		      (void *)dev, strerror(err));
+		      (void *)dev, strerror(rte_errno));
+	}
 }
 
 /**
@@ -682,7 +713,7 @@ mlx5_link_stop(struct rte_eth_dev *dev)
  *   Link desired status.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_force_link_status_change(struct rte_eth_dev *dev, int status)
@@ -696,7 +727,8 @@ mlx5_force_link_status_change(struct rte_eth_dev *dev, int status)
 		try++;
 		sleep(1);
 	}
-	return -EAGAIN;
+	rte_errno = EAGAIN;
+	return -rte_errno;
 }
 
 /**
@@ -708,7 +740,7 @@ mlx5_force_link_status_change(struct rte_eth_dev *dev, int status)
  *   Wait for request completion (ignored).
  *
  * @return
- *   0 on success, -1 on error.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
@@ -725,10 +757,12 @@ mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
 		ret = mlx5_link_update_unlocked_gset(dev);
 	else
 		ret = mlx5_link_update_unlocked_gs(dev);
+	if (ret)
+		return ret;
 	/* If lsc interrupt is disabled, should always be ready for traffic. */
 	if (!dev->data->dev_conf.intr_conf.lsc) {
 		mlx5_link_start(dev);
-		return ret;
+		return 0;
 	}
 	/* Re-select burst callbacks only if link status has been changed. */
 	if (!ret && dev_link.link_status != dev->data->dev_link.link_status) {
@@ -737,7 +771,7 @@ mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
 		else
 			mlx5_link_stop(dev);
 	}
-	return ret;
+	return 0;
 }
 
 /**
@@ -749,36 +783,32 @@ mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
  *   New MTU.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 {
 	struct priv *priv = dev->data->dev_private;
-	uint16_t kern_mtu;
-	int ret = 0;
+	uint16_t kern_mtu = 0;
+	int ret;
 
 	ret = mlx5_get_mtu(dev, &kern_mtu);
 	if (ret)
-		goto out;
+		return ret;
 	/* Set kernel interface MTU first. */
 	ret = mlx5_set_mtu(dev, mtu);
 	if (ret)
-		goto out;
+		return ret;
 	ret = mlx5_get_mtu(dev, &kern_mtu);
 	if (ret)
-		goto out;
+		return ret;
 	if (kern_mtu == mtu) {
 		priv->mtu = mtu;
 		DEBUG("adapter port %u MTU set to %u", priv->port, mtu);
+		return 0;
 	}
-	return 0;
-out:
-	ret = errno;
-	WARN("cannot set port %u MTU to %u: %s", priv->port, mtu,
-	     strerror(ret));
-	assert(ret >= 0);
-	return -ret;
+	rte_errno = EAGAIN;
+	return -rte_errno;
 }
 
 /**
@@ -790,7 +820,7 @@ mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
  *   Flow control output buffer.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
@@ -802,11 +832,11 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 	int ret;
 
 	ifr.ifr_data = (void *)&ethpause;
-	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
-		ret = errno;
+	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+	if (ret) {
 		WARN("ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM) failed: %s",
-		     strerror(ret));
-		goto out;
+		     strerror(rte_errno));
+		return ret;
 	}
 	fc_conf->autoneg = ethpause.autoneg;
 	if (ethpause.rx_pause && ethpause.tx_pause)
@@ -817,10 +847,7 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		fc_conf->mode = RTE_FC_TX_PAUSE;
 	else
 		fc_conf->mode = RTE_FC_NONE;
-	ret = 0;
-out:
-	assert(ret >= 0);
-	return -ret;
+	return 0;
 }
 
 /**
@@ -832,7 +859,7 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
  *   Flow control parameters.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
@@ -856,17 +883,14 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		ethpause.tx_pause = 1;
 	else
 		ethpause.tx_pause = 0;
-	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr)) {
-		ret = errno;
+	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+	if (ret) {
 		WARN("ioctl(SIOCETHTOOL, ETHTOOL_SPAUSEPARAM)"
 		     " failed: %s",
-		     strerror(ret));
-		goto out;
+		     strerror(rte_errno));
+		return ret;
 	}
-	ret = 0;
-out:
-	assert(ret >= 0);
-	return -ret;
+	return 0;
 }
 
 /**
@@ -878,7 +902,7 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
  *   PCI bus address output buffer.
  *
  * @return
- *   0 on success, -1 on failure and errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_ibv_device_to_pci_addr(const struct ibv_device *device,
@@ -889,8 +913,10 @@ mlx5_ibv_device_to_pci_addr(const struct ibv_device *device,
 	MKSTR(path, "%s/device/uevent", device->ibdev_path);
 
 	file = fopen(path, "rb");
-	if (file == NULL)
-		return -1;
+	if (file == NULL) {
+		rte_errno = errno;
+		return -rte_errno;
+	}
 	while (fgets(line, sizeof(line), file) == line) {
 		size_t len = strlen(line);
 		int ret;
@@ -926,15 +952,19 @@ mlx5_ibv_device_to_pci_addr(const struct ibv_device *device,
  *   Pointer to Ethernet device.
  *
  * @return
- *   Zero if the callback process can be called immediately.
+ *   Zero if the callback process can be called immediately, negative errno
+ *   value otherwise and rte_errno is set.
  */
 static int
 mlx5_link_status_update(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 	struct rte_eth_link *link = &dev->data->dev_link;
+	int ret;
 
-	mlx5_link_update(dev, 0);
+	ret = mlx5_link_update(dev, 0);
+	if (ret)
+		return ret;
 	if (((link->link_speed == 0) && link->link_status) ||
 		((link->link_speed != 0) && !link->link_status)) {
 		/*
@@ -1091,12 +1121,13 @@ void
 mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
-	int rc, flags;
+	int ret;
+	int flags;
 
 	assert(priv->ctx->async_fd > 0);
 	flags = fcntl(priv->ctx->async_fd, F_GETFL);
-	rc = fcntl(priv->ctx->async_fd, F_SETFL, flags | O_NONBLOCK);
-	if (rc < 0) {
+	ret = fcntl(priv->ctx->async_fd, F_SETFL, flags | O_NONBLOCK);
+	if (ret) {
 		INFO("failed to change file descriptor async event queue");
 		dev->data->dev_conf.intr_conf.lsc = 0;
 		dev->data->dev_conf.intr_conf.rmv = 0;
@@ -1108,8 +1139,10 @@ mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
 		rte_intr_callback_register(&priv->intr_handle,
 					   mlx5_dev_interrupt_handler, dev);
 	}
-	rc = mlx5_socket_init(dev);
-	if (!rc && priv->primary_socket) {
+	ret = mlx5_socket_init(dev);
+	if (ret)
+		ERROR("cannot initialise socket: %s", strerror(rte_errno));
+	else if (priv->primary_socket) {
 		priv->intr_handle_socket.fd = priv->primary_socket;
 		priv->intr_handle_socket.type = RTE_INTR_HANDLE_EXT;
 		rte_intr_callback_register(&priv->intr_handle_socket,
@@ -1124,7 +1157,7 @@ mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_set_link_down(struct rte_eth_dev *dev)
@@ -1139,7 +1172,7 @@ mlx5_set_link_down(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_set_link_up(struct rte_eth_dev *dev)
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index aaa8727ee..09a798924 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -286,7 +286,8 @@ struct mlx5_flow_items {
 	 *   Internal structure to store the conversion.
 	 *
 	 * @return
-	 *   0 on success, negative value otherwise.
+	 *   0 on success, a negative errno value otherwise and rte_errno is
+	 *   set.
 	 */
 	int (*convert)(const struct rte_flow_item *item,
 		       const void *default_mask,
@@ -499,45 +500,52 @@ struct ibv_spec_header {
  *   Bit-Mask size in bytes.
  *
  * @return
- *   0 on success.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_item_validate(const struct rte_flow_item *item,
 			const uint8_t *mask, unsigned int size)
 {
-	int ret = 0;
-
-	if (!item->spec && (item->mask || item->last))
-		return -1;
+	if (!item->spec && (item->mask || item->last)) {
+		rte_errno = EINVAL;
+		return -rte_errno;
+	}
 	if (item->spec && !item->mask) {
 		unsigned int i;
 		const uint8_t *spec = item->spec;
 
 		for (i = 0; i < size; ++i)
-			if ((spec[i] | mask[i]) != mask[i])
-				return -1;
+			if ((spec[i] | mask[i]) != mask[i]) {
+				rte_errno = EINVAL;
+				return -rte_errno;
+			}
 	}
 	if (item->last && !item->mask) {
 		unsigned int i;
 		const uint8_t *spec = item->last;
 
 		for (i = 0; i < size; ++i)
-			if ((spec[i] | mask[i]) != mask[i])
-				return -1;
+			if ((spec[i] | mask[i]) != mask[i]) {
+				rte_errno = EINVAL;
+				return -rte_errno;
+			}
 	}
 	if (item->mask) {
 		unsigned int i;
 		const uint8_t *spec = item->spec;
 
 		for (i = 0; i < size; ++i)
-			if ((spec[i] | mask[i]) != mask[i])
-				return -1;
+			if ((spec[i] | mask[i]) != mask[i]) {
+				rte_errno = EINVAL;
+				return -rte_errno;
+			}
 	}
 	if (item->spec && item->last) {
 		uint8_t spec[size];
 		uint8_t last[size];
 		const uint8_t *apply = mask;
 		unsigned int i;
+		int ret;
 
 		if (item->mask)
 			apply = item->mask;
@@ -546,8 +554,12 @@ mlx5_flow_item_validate(const struct rte_flow_item *item,
 			last[i] = ((const uint8_t *)item->last)[i] & apply[i];
 		}
 		ret = memcmp(spec, last, size);
+		if (ret != 0) {
+			rte_errno = EINVAL;
+			return -rte_errno;
+		}
 	}
-	return ret;
+	return 0;
 }
 
 /**
@@ -560,7 +572,7 @@ mlx5_flow_item_validate(const struct rte_flow_item *item,
  *   User RSS configuration to save.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_convert_rss_conf(struct mlx5_flow_parse *parser,
@@ -572,10 +584,14 @@ mlx5_flow_convert_rss_conf(struct mlx5_flow_parse *parser,
 	 * device default RSS configuration.
 	 */
 	if (rss_conf) {
-		if (rss_conf->rss_hf & MLX5_RSS_HF_MASK)
-			return EINVAL;
-		if (rss_conf->rss_key_len != 40)
-			return EINVAL;
+		if (rss_conf->rss_hf & MLX5_RSS_HF_MASK) {
+			rte_errno = EINVAL;
+			return -rte_errno;
+		}
+		if (rss_conf->rss_key_len != 40) {
+			rte_errno = EINVAL;
+			return -rte_errno;
+		}
 		if (rss_conf->rss_key_len && rss_conf->rss_key) {
 			parser->rss_conf.rss_key_len = rss_conf->rss_key_len;
 			memcpy(parser->rss_key, rss_conf->rss_key,
@@ -655,14 +671,17 @@ mlx5_flow_convert_actions(struct rte_eth_dev *dev,
 			  struct mlx5_flow_parse *parser)
 {
 	struct priv *priv = dev->data->dev_private;
+	int ret;
 
 	/*
 	 * Add default RSS configuration necessary for Verbs to create QP even
 	 * if no RSS is necessary.
 	 */
-	mlx5_flow_convert_rss_conf(parser,
-				   (const struct rte_eth_rss_conf *)
-				   &priv->rss_conf);
+	ret = mlx5_flow_convert_rss_conf(parser,
+					 (const struct rte_eth_rss_conf *)
+					 &priv->rss_conf);
+	if (ret)
+		return ret;
 	for (; actions->type != RTE_FLOW_ACTION_TYPE_END; ++actions) {
 		if (actions->type == RTE_FLOW_ACTION_TYPE_VOID) {
 			continue;
@@ -811,6 +830,7 @@ mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
 {
 	const struct mlx5_flow_items *cur_item = mlx5_flow_items;
 	unsigned int i;
+	int ret = 0;
 
 	/* Initialise the offsets to start after verbs attribute. */
 	for (i = 0; i != hash_rxq_init_n; ++i)
@@ -818,7 +838,6 @@ mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
 	for (; items->type != RTE_FLOW_ITEM_TYPE_END; ++items) {
 		const struct mlx5_flow_items *token = NULL;
 		unsigned int n;
-		int err;
 
 		if (items->type == RTE_FLOW_ITEM_TYPE_VOID)
 			continue;
@@ -834,10 +853,10 @@ mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
 		if (!token)
 			goto exit_item_not_supported;
 		cur_item = token;
-		err = mlx5_flow_item_validate(items,
+		ret = mlx5_flow_item_validate(items,
 					      (const uint8_t *)cur_item->mask,
 					      cur_item->mask_sz);
-		if (err)
+		if (ret)
 			goto exit_item_not_supported;
 		if (items->type == RTE_FLOW_ITEM_TYPE_VXLAN) {
 			if (parser->inner) {
@@ -874,9 +893,8 @@ mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
 	}
 	return 0;
 exit_item_not_supported:
-	rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM,
-			   items, "item not supported");
-	return -rte_errno;
+	return rte_flow_error_set(error, -ret, RTE_FLOW_ERROR_TYPE_ITEM,
+				  items, "item not supported");
 }
 
 /**
@@ -890,7 +908,7 @@ mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
  *   Perform verbose error reporting if not NULL.
  *
  * @return
- *   A verbs flow attribute on success, NULL otherwise.
+ *   A verbs flow attribute on success, NULL otherwise and rte_errno is set.
  */
 static struct ibv_flow_attr *
 mlx5_flow_convert_allocate(unsigned int priority,
@@ -1092,7 +1110,7 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 			 parser->queue[HASH_RXQ_ETH].offset,
 			 error);
 		if (!parser->queue[HASH_RXQ_ETH].ibv_attr)
-			return ENOMEM;
+			goto exit_enomem;
 		parser->queue[HASH_RXQ_ETH].offset =
 			sizeof(struct ibv_flow_attr);
 	} else {
@@ -1127,7 +1145,7 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 					 cur_item->mask),
 					parser);
 		if (ret) {
-			rte_flow_error_set(error, ret,
+			rte_flow_error_set(error, rte_errno,
 					   RTE_FLOW_ERROR_TYPE_ITEM,
 					   items, "item not supported");
 			goto exit_free;
@@ -1169,13 +1187,13 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 			parser->queue[i].ibv_attr = NULL;
 		}
 	}
-	rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+	rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
 			   NULL, "cannot allocate verbs spec attributes.");
-	return ret;
+	return -rte_errno;
 exit_count_error:
 	rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
 			   NULL, "cannot create counter.");
-	return rte_errno;
+	return -rte_errno;
 }
 
 /**
@@ -1221,6 +1239,9 @@ mlx5_flow_create_copy(struct mlx5_flow_parse *parser, void *src,
  *   Default bit-masks to use when item->mask is not provided.
  * @param data[in, out]
  *   User structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_eth(const struct rte_flow_item *item,
@@ -1270,6 +1291,9 @@ mlx5_flow_create_eth(const struct rte_flow_item *item,
  *   Default bit-masks to use when item->mask is not provided.
  * @param data[in, out]
  *   User structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_vlan(const struct rte_flow_item *item,
@@ -1310,6 +1334,9 @@ mlx5_flow_create_vlan(const struct rte_flow_item *item,
  *   Default bit-masks to use when item->mask is not provided.
  * @param data[in, out]
  *   User structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_ipv4(const struct rte_flow_item *item,
@@ -1362,6 +1389,9 @@ mlx5_flow_create_ipv4(const struct rte_flow_item *item,
  *   Default bit-masks to use when item->mask is not provided.
  * @param data[in, out]
  *   User structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_ipv6(const struct rte_flow_item *item,
@@ -1418,6 +1448,9 @@ mlx5_flow_create_ipv6(const struct rte_flow_item *item,
  *   Default bit-masks to use when item->mask is not provided.
  * @param data[in, out]
  *   User structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_udp(const struct rte_flow_item *item,
@@ -1464,6 +1497,9 @@ mlx5_flow_create_udp(const struct rte_flow_item *item,
  *   Default bit-masks to use when item->mask is not provided.
  * @param data[in, out]
  *   User structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_tcp(const struct rte_flow_item *item,
@@ -1510,6 +1546,9 @@ mlx5_flow_create_tcp(const struct rte_flow_item *item,
  *   Default bit-masks to use when item->mask is not provided.
  * @param data[in, out]
  *   User structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_vxlan(const struct rte_flow_item *item,
@@ -1549,8 +1588,10 @@ mlx5_flow_create_vxlan(const struct rte_flow_item *item,
 	 * before will also match this rule.
 	 * To avoid such situation, VNI 0 is currently refused.
 	 */
-	if (!vxlan.val.tunnel_id)
-		return EINVAL;
+	if (!vxlan.val.tunnel_id) {
+		rte_errno = EINVAL;
+		return -rte_errno;
+	}
 	mlx5_flow_create_copy(parser, &vxlan, size);
 	return 0;
 }
@@ -1562,6 +1603,9 @@ mlx5_flow_create_vxlan(const struct rte_flow_item *item,
  *   Internal parser structure.
  * @param mark_id
  *   Mark identifier.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_flag_mark(struct mlx5_flow_parse *parser, uint32_t mark_id)
@@ -1587,7 +1631,7 @@ mlx5_flow_create_flag_mark(struct mlx5_flow_parse *parser, uint32_t mark_id)
  *   Pointer to MLX5 flow parser structure.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_count(struct rte_eth_dev *dev __rte_unused,
@@ -1605,8 +1649,10 @@ mlx5_flow_create_count(struct rte_eth_dev *dev __rte_unused,
 
 	init_attr.counter_set_id = 0;
 	parser->cs = ibv_create_counter_set(priv->ctx, &init_attr);
-	if (!parser->cs)
-		return EINVAL;
+	if (!parser->cs) {
+		rte_errno = EINVAL;
+		return -rte_errno;
+	}
 	counter.counter_set_handle = parser->cs->handle;
 	mlx5_flow_create_copy(parser, &counter, size);
 #endif
@@ -1626,7 +1672,7 @@ mlx5_flow_create_count(struct rte_eth_dev *dev __rte_unused,
  *   Perform verbose error reporting if not NULL.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_action_queue_drop(struct rte_eth_dev *dev,
@@ -1637,7 +1683,6 @@ mlx5_flow_create_action_queue_drop(struct rte_eth_dev *dev,
 	struct priv *priv = dev->data->dev_private;
 	struct ibv_flow_spec_action_drop *drop;
 	unsigned int size = sizeof(struct ibv_flow_spec_action_drop);
-	int err = 0;
 
 	assert(priv->pd);
 	assert(priv->ctx);
@@ -1663,7 +1708,6 @@ mlx5_flow_create_action_queue_drop(struct rte_eth_dev *dev,
 	if (!flow->frxq[HASH_RXQ_ETH].ibv_flow) {
 		rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_HANDLE,
 				   NULL, "flow rule creation failure");
-		err = ENOMEM;
 		goto error;
 	}
 	return 0;
@@ -1682,7 +1726,7 @@ mlx5_flow_create_action_queue_drop(struct rte_eth_dev *dev,
 		flow->cs = NULL;
 		parser->cs = NULL;
 	}
-	return err;
+	return -rte_errno;
 }
 
 /**
@@ -1698,7 +1742,7 @@ mlx5_flow_create_action_queue_drop(struct rte_eth_dev *dev,
  *   Perform verbose error reporting if not NULL.
  *
  * @return
- *   0 on success, a errno value otherwise and rte_errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_action_queue_rss(struct rte_eth_dev *dev,
@@ -1736,10 +1780,10 @@ mlx5_flow_create_action_queue_rss(struct rte_eth_dev *dev,
 				      parser->queues,
 				      parser->queues_n);
 		if (!flow->frxq[i].hrxq) {
-			rte_flow_error_set(error, ENOMEM,
-					   RTE_FLOW_ERROR_TYPE_HANDLE,
-					   NULL, "cannot create hash rxq");
-			return ENOMEM;
+			return rte_flow_error_set(error, ENOMEM,
+						  RTE_FLOW_ERROR_TYPE_HANDLE,
+						  NULL,
+						  "cannot create hash rxq");
 		}
 	}
 	return 0;
@@ -1758,7 +1802,7 @@ mlx5_flow_create_action_queue_rss(struct rte_eth_dev *dev,
  *   Perform verbose error reporting if not NULL.
  *
  * @return
- *   0 on success, a errno value otherwise and rte_errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
@@ -1767,15 +1811,15 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 			      struct rte_flow_error *error)
 {
 	struct priv *priv = dev->data->dev_private;
-	int err = 0;
+	int ret;
 	unsigned int i;
 	unsigned int flows_n = 0;
 
 	assert(priv->pd);
 	assert(priv->ctx);
 	assert(!parser->drop);
-	err = mlx5_flow_create_action_queue_rss(dev, parser, flow, error);
-	if (err)
+	ret = mlx5_flow_create_action_queue_rss(dev, parser, flow, error);
+	if (ret)
 		goto error;
 	if (parser->count)
 		flow->cs = parser->cs;
@@ -1791,7 +1835,6 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 			rte_flow_error_set(error, ENOMEM,
 					   RTE_FLOW_ERROR_TYPE_HANDLE,
 					   NULL, "flow rule creation failure");
-			err = ENOMEM;
 			goto error;
 		}
 		++flows_n;
@@ -1813,6 +1856,7 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 	}
 	return 0;
 error:
+	ret = rte_errno; /* Save rte_errno before cleanup. */
 	assert(flow);
 	for (i = 0; i != hash_rxq_init_n; ++i) {
 		if (flow->frxq[i].ibv_flow) {
@@ -1830,7 +1874,8 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 		flow->cs = NULL;
 		parser->cs = NULL;
 	}
-	return err;
+	rte_errno = ret; /* Restore rte_errno. */
+	return -rte_errno;
 }
 
 /**
@@ -1850,7 +1895,7 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
  *   Perform verbose error reporting if not NULL.
  *
  * @return
- *   A flow on success, NULL otherwise.
+ *   A flow on success, NULL otherwise and rte_errno is set.
  */
 static struct rte_flow *
 mlx5_flow_list_create(struct rte_eth_dev *dev,
@@ -1863,10 +1908,10 @@ mlx5_flow_list_create(struct rte_eth_dev *dev,
 	struct mlx5_flow_parse parser = { .create = 1, };
 	struct rte_flow *flow = NULL;
 	unsigned int i;
-	int err;
+	int ret;
 
-	err = mlx5_flow_convert(dev, attr, items, actions, error, &parser);
-	if (err)
+	ret = mlx5_flow_convert(dev, attr, items, actions, error, &parser);
+	if (ret)
 		goto exit;
 	flow = rte_calloc(__func__, 1,
 			  sizeof(*flow) + parser.queues_n * sizeof(uint16_t),
@@ -1889,11 +1934,11 @@ mlx5_flow_list_create(struct rte_eth_dev *dev,
 	memcpy(flow->rss_key, parser.rss_key, parser.rss_conf.rss_key_len);
 	/* finalise the flow. */
 	if (parser.drop)
-		err = mlx5_flow_create_action_queue_drop(dev, &parser, flow,
+		ret = mlx5_flow_create_action_queue_drop(dev, &parser, flow,
 							 error);
 	else
-		err = mlx5_flow_create_action_queue(dev, &parser, flow, error);
-	if (err)
+		ret = mlx5_flow_create_action_queue(dev, &parser, flow, error);
+	if (ret)
 		goto exit;
 	TAILQ_INSERT_TAIL(list, flow, next);
 	DEBUG("Flow created %p", (void *)flow);
@@ -1920,11 +1965,9 @@ mlx5_flow_validate(struct rte_eth_dev *dev,
 		   const struct rte_flow_action actions[],
 		   struct rte_flow_error *error)
 {
-	int ret;
 	struct mlx5_flow_parse parser = { .create = 0, };
 
-	ret = mlx5_flow_convert(dev, attr, items, actions, error, &parser);
-	return ret;
+	return mlx5_flow_convert(dev, attr, items, actions, error, &parser);
 }
 
 /**
@@ -2047,7 +2090,7 @@ mlx5_flow_list_flush(struct rte_eth_dev *dev, struct mlx5_flows *list)
  *   Pointer to Ethernet device.
  *
  * @return
- *   0 on success.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
@@ -2060,11 +2103,13 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 	fdq = rte_calloc(__func__, 1, sizeof(*fdq), 0);
 	if (!fdq) {
 		WARN("cannot allocate memory for drop queue");
-		goto error;
+		rte_errno = ENOMEM;
+		return -rte_errno;
 	}
 	fdq->cq = ibv_create_cq(priv->ctx, 1, NULL, NULL, 0);
 	if (!fdq->cq) {
 		WARN("cannot allocate CQ for drop queue");
+		rte_errno = errno;
 		goto error;
 	}
 	fdq->wq = ibv_create_wq(priv->ctx,
@@ -2077,6 +2122,7 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 			});
 	if (!fdq->wq) {
 		WARN("cannot allocate WQ for drop queue");
+		rte_errno = errno;
 		goto error;
 	}
 	fdq->ind_table = ibv_create_rwq_ind_table(priv->ctx,
@@ -2087,6 +2133,7 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 			});
 	if (!fdq->ind_table) {
 		WARN("cannot allocate indirection table for drop queue");
+		rte_errno = errno;
 		goto error;
 	}
 	fdq->qp = ibv_create_qp_ex(priv->ctx,
@@ -2108,6 +2155,7 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 		});
 	if (!fdq->qp) {
 		WARN("cannot allocate QP for drop queue");
+		rte_errno = errno;
 		goto error;
 	}
 	priv->flow_drop_queue = fdq;
@@ -2124,7 +2172,7 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 	if (fdq)
 		rte_free(fdq);
 	priv->flow_drop_queue = NULL;
-	return -1;
+	return -rte_errno;
 }
 
 /**
@@ -2222,7 +2270,7 @@ mlx5_flow_stop(struct rte_eth_dev *dev, struct mlx5_flows *list)
  *   Pointer to a TAILQ flow list.
  *
  * @return
- *   0 on success, a errno value otherwise and rte_errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
@@ -2242,7 +2290,7 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 				DEBUG("Flow %p cannot be applied",
 				      (void *)flow);
 				rte_errno = EINVAL;
-				return rte_errno;
+				return -rte_errno;
 			}
 			DEBUG("Flow %p applied", (void *)flow);
 			/* Next flow. */
@@ -2269,7 +2317,7 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 				DEBUG("Flow %p cannot be applied",
 				      (void *)flow);
 				rte_errno = EINVAL;
-				return rte_errno;
+				return -rte_errno;
 			}
 flow_create:
 			flow->frxq[i].ibv_flow =
@@ -2279,7 +2327,7 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 				DEBUG("Flow %p cannot be applied",
 				      (void *)flow);
 				rte_errno = EINVAL;
-				return rte_errno;
+				return -rte_errno;
 			}
 			DEBUG("Flow %p applied", (void *)flow);
 		}
@@ -2329,7 +2377,7 @@ mlx5_flow_verify(struct rte_eth_dev *dev)
  *   A VLAN flow mask to apply.
  *
  * @return
- *   0 on success.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
@@ -2381,8 +2429,10 @@ mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
 		} local;
 	} action_rss;
 
-	if (!priv->reta_idx_n)
-		return EINVAL;
+	if (!priv->reta_idx_n) {
+		rte_errno = EINVAL;
+		return -rte_errno;
+	}
 	for (i = 0; i != priv->reta_idx_n; ++i)
 		action_rss.local.queue[i] = (*priv->reta_idx)[i];
 	action_rss.local.rss_conf = &priv->rss_conf;
@@ -2391,7 +2441,7 @@ mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
 	flow = mlx5_flow_list_create(dev, &priv->ctrl_flows, &attr, items,
 				     actions, &error);
 	if (!flow)
-		return rte_errno;
+		return -rte_errno;
 	return 0;
 }
 
@@ -2406,7 +2456,7 @@ mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
  *   An Ethernet flow mask to apply.
  *
  * @return
- *   0 on success.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_ctrl_flow(struct rte_eth_dev *dev,
@@ -2459,7 +2509,7 @@ mlx5_flow_flush(struct rte_eth_dev *dev,
  *   returned data from the counter.
  *
  * @return
- *   0 on success, a errno value otherwise and rte_errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_flow_query_count(struct ibv_counter_set *cs,
@@ -2476,15 +2526,13 @@ mlx5_flow_query_count(struct ibv_counter_set *cs,
 		.out = counters,
 		.outlen = 2 * sizeof(uint64_t),
 	};
-	int res = ibv_query_counter_set(&query_cs_attr, &query_out);
+	int err = ibv_query_counter_set(&query_cs_attr, &query_out);
 
-	if (res) {
-		rte_flow_error_set(error, -res,
-				   RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
-				   NULL,
-				   "cannot read counter");
-		return -res;
-	}
+	if (err)
+		return rte_flow_error_set(error, err,
+					  RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+					  NULL,
+					  "cannot read counter");
 	query_count->hits_set = 1;
 	query_count->bytes_set = 1;
 	query_count->hits = counters[0] - counter_stats->hits;
@@ -2509,20 +2557,22 @@ mlx5_flow_query(struct rte_eth_dev *dev __rte_unused,
 		void *data,
 		struct rte_flow_error *error)
 {
-	int res = EINVAL;
-
 	if (flow->cs) {
-		res = mlx5_flow_query_count(flow->cs,
-					&flow->counter_stats,
-					(struct rte_flow_query_count *)data,
-					error);
+		int ret;
+
+		ret = mlx5_flow_query_count(flow->cs,
+					    &flow->counter_stats,
+					    (struct rte_flow_query_count *)data,
+					    error);
+		if (ret)
+			return ret;
 	} else {
-		rte_flow_error_set(error, res,
-				   RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
-				   NULL,
-				   "no counter found for flow");
+		return rte_flow_error_set(error, EINVAL,
+					  RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+					  NULL,
+					  "no counter found for flow");
 	}
-	return -res;
+	return 0;
 }
 #endif
 
@@ -2565,7 +2615,7 @@ mlx5_flow_isolate(struct rte_eth_dev *dev,
  *   Generic flow parameters structure.
  *
  * @return
- *  0 on success, errno value on error.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
@@ -2578,7 +2628,8 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 	/* Validate queue number. */
 	if (fdir_filter->action.rx_queue >= priv->rxqs_n) {
 		ERROR("invalid queue number %d", fdir_filter->action.rx_queue);
-		return EINVAL;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	attributes->attr.ingress = 1;
 	attributes->items[0] = (struct rte_flow_item) {
@@ -2600,7 +2651,8 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 		break;
 	default:
 		ERROR("invalid behavior %d", fdir_filter->action.behavior);
-		return ENOTSUP;
+		rte_errno = ENOTSUP;
+		return -rte_errno;
 	}
 	attributes->queue.index = fdir_filter->action.rx_queue;
 	switch (fdir_filter->input.flow_type) {
@@ -2734,9 +2786,9 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 		};
 		break;
 	default:
-		ERROR("invalid flow type%d",
-		      fdir_filter->input.flow_type);
-		return ENOTSUP;
+		ERROR("invalid flow type%d", fdir_filter->input.flow_type);
+		rte_errno = ENOTSUP;
+		return -rte_errno;
 	}
 	return 0;
 }
@@ -2750,7 +2802,7 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
  *   Flow director filter to add.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_fdir_filter_add(struct rte_eth_dev *dev,
@@ -2774,11 +2826,11 @@ mlx5_fdir_filter_add(struct rte_eth_dev *dev,
 
 	ret = mlx5_fdir_filter_convert(dev, fdir_filter, &attributes);
 	if (ret)
-		return -ret;
+		return ret;
 	ret = mlx5_flow_convert(dev, &attributes.attr, attributes.items,
 				attributes.actions, &error, &parser);
 	if (ret)
-		return -ret;
+		return ret;
 	flow = mlx5_flow_list_create(dev, &priv->flows, &attributes.attr,
 				     attributes.items, attributes.actions,
 				     &error);
@@ -2786,7 +2838,7 @@ mlx5_fdir_filter_add(struct rte_eth_dev *dev,
 		DEBUG("FDIR created %p", (void *)flow);
 		return 0;
 	}
-	return ENOTSUP;
+	return -rte_errno;
 }
 
 /**
@@ -2798,7 +2850,7 @@ mlx5_fdir_filter_add(struct rte_eth_dev *dev,
  *   Filter to be deleted.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
@@ -2819,7 +2871,7 @@ mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
 
 	ret = mlx5_fdir_filter_convert(dev, fdir_filter, &attributes);
 	if (ret)
-		return -ret;
+		return ret;
 	ret = mlx5_flow_convert(dev, &attributes.attr, attributes.items,
 				attributes.actions, &error, &parser);
 	if (ret)
@@ -2877,6 +2929,7 @@ mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
 		/* The flow does not match. */
 		continue;
 	}
+	ret = rte_errno; /* Save rte_errno before cleanup. */
 	if (flow)
 		mlx5_flow_list_destroy(dev, &priv->flows, flow);
 exit:
@@ -2884,7 +2937,8 @@ mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
 		if (parser.queue[i].ibv_attr)
 			rte_free(parser.queue[i].ibv_attr);
 	}
-	return -ret;
+	rte_errno = ret; /* Restore rte_errno. */
+	return -rte_errno;
 }
 
 /**
@@ -2896,7 +2950,7 @@ mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
  *   Filter to be updated.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_fdir_filter_update(struct rte_eth_dev *dev,
@@ -2907,8 +2961,7 @@ mlx5_fdir_filter_update(struct rte_eth_dev *dev,
 	ret = mlx5_fdir_filter_delete(dev, fdir_filter);
 	if (ret)
 		return ret;
-	ret = mlx5_fdir_filter_add(dev, fdir_filter);
-	return ret;
+	return mlx5_fdir_filter_add(dev, fdir_filter);
 }
 
 /**
@@ -2962,7 +3015,7 @@ mlx5_fdir_info_get(struct rte_eth_dev *dev, struct rte_eth_fdir_info *fdir_info)
  *   Pointer to operation-specific structure.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
@@ -2971,7 +3024,6 @@ mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
 	struct priv *priv = dev->data->dev_private;
 	enum rte_fdir_mode fdir_mode =
 		priv->dev->data->dev_conf.fdir_conf.mode;
-	int ret = 0;
 
 	if (filter_op == RTE_ETH_FILTER_NOP)
 		return 0;
@@ -2979,18 +3031,16 @@ mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
 	    fdir_mode != RTE_FDIR_MODE_PERFECT_MAC_VLAN) {
 		ERROR("%p: flow director mode %d not supported",
 		      (void *)dev, fdir_mode);
-		return EINVAL;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	switch (filter_op) {
 	case RTE_ETH_FILTER_ADD:
-		ret = mlx5_fdir_filter_add(dev, arg);
-		break;
+		return mlx5_fdir_filter_add(dev, arg);
 	case RTE_ETH_FILTER_UPDATE:
-		ret = mlx5_fdir_filter_update(dev, arg);
-		break;
+		return mlx5_fdir_filter_update(dev, arg);
 	case RTE_ETH_FILTER_DELETE:
-		ret = mlx5_fdir_filter_delete(dev, arg);
-		break;
+		return mlx5_fdir_filter_delete(dev, arg);
 	case RTE_ETH_FILTER_FLUSH:
 		mlx5_fdir_filter_flush(dev);
 		break;
@@ -2998,12 +3048,11 @@ mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
 		mlx5_fdir_info_get(dev, arg);
 		break;
 	default:
-		DEBUG("%p: unknown operation %u", (void *)dev,
-		      filter_op);
-		ret = EINVAL;
-		break;
+		DEBUG("%p: unknown operation %u", (void *)dev, filter_op);
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
-	return ret;
+	return 0;
 }
 
 /**
@@ -3019,7 +3068,7 @@ mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
  *   Pointer to operation-specific structure.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
@@ -3027,21 +3076,21 @@ mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
 		     enum rte_filter_op filter_op,
 		     void *arg)
 {
-	int ret = EINVAL;
-
 	switch (filter_type) {
 	case RTE_ETH_FILTER_GENERIC:
-		if (filter_op != RTE_ETH_FILTER_GET)
-			return -EINVAL;
+		if (filter_op != RTE_ETH_FILTER_GET) {
+			rte_errno = EINVAL;
+			return -rte_errno;
+		}
 		*(const void **)arg = &mlx5_flow_ops;
 		return 0;
 	case RTE_ETH_FILTER_FDIR:
-		ret = mlx5_fdir_ctrl_func(dev, filter_op, arg);
-		break;
+		return mlx5_fdir_ctrl_func(dev, filter_op, arg);
 	default:
 		ERROR("%p: filter type (%d) not supported",
 		      (void *)dev, filter_type);
-		break;
+		rte_errno = ENOTSUP;
+		return -rte_errno;
 	}
-	return -ret;
+	return 0;
 }
diff --git a/drivers/net/mlx5/mlx5_mac.c b/drivers/net/mlx5/mlx5_mac.c
index 20fed527b..e9d9c67e9 100644
--- a/drivers/net/mlx5/mlx5_mac.c
+++ b/drivers/net/mlx5/mlx5_mac.c
@@ -69,15 +69,17 @@
  *   MAC address output buffer.
  *
  * @return
- *   0 on success, -1 on failure and errno is set.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_get_mac(struct rte_eth_dev *dev, uint8_t (*mac)[ETHER_ADDR_LEN])
 {
 	struct ifreq request;
+	int ret;
 
-	if (mlx5_ifreq(dev, SIOCGIFHWADDR, &request))
-		return -1;
+	ret = mlx5_ifreq(dev, SIOCGIFHWADDR, &request);
+	if (ret)
+		return ret;
 	memcpy(mac, request.ifr_hwaddr.sa_data, ETHER_ADDR_LEN);
 	return 0;
 }
@@ -95,8 +97,13 @@ mlx5_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index)
 {
 	assert(index < MLX5_MAX_MAC_ADDRESSES);
 	memset(&dev->data->mac_addrs[index], 0, sizeof(struct ether_addr));
-	if (!dev->data->promiscuous)
-		mlx5_traffic_restart(dev);
+	if (!dev->data->promiscuous) {
+		int ret = mlx5_traffic_restart(dev);
+
+		if (ret)
+			ERROR("%p cannot remove mac address: %s", (void *)dev,
+			      strerror(rte_errno));
+	}
 }
 
 /**
@@ -112,14 +119,13 @@ mlx5_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index)
  *   VMDq pool index to associate address with (ignored).
  *
  * @return
- *   0 on success.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac,
 		  uint32_t index, uint32_t vmdq __rte_unused)
 {
 	unsigned int i;
-	int ret = 0;
 
 	assert(index < MLX5_MAX_MAC_ADDRESSES);
 	/* First, make sure this address isn't already configured. */
@@ -130,12 +136,13 @@ mlx5_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac,
 		if (memcmp(&dev->data->mac_addrs[i], mac, sizeof(*mac)))
 			continue;
 		/* Address already configured elsewhere, return with error. */
-		return EADDRINUSE;
+		rte_errno = EADDRINUSE;
+		return -rte_errno;
 	}
 	dev->data->mac_addrs[index] = *mac;
 	if (!dev->data->promiscuous)
-		mlx5_traffic_restart(dev);
-	return ret;
+		return mlx5_traffic_restart(dev);
+	return 0;
 }
 
 /**
@@ -149,6 +156,10 @@ mlx5_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac,
 void
 mlx5_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr)
 {
+	int ret;
+
 	DEBUG("%p: setting primary MAC address", (void *)dev);
-	mlx5_mac_addr_add(dev, mac_addr, 0, 0);
+	ret = mlx5_mac_addr_add(dev, mac_addr, 0, 0);
+	if (ret)
+		ERROR("cannot set mac address: %s", strerror(rte_errno));
 }
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index cfad4798b..3a4e46f9a 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -120,7 +120,7 @@ mlx5_check_mempool(struct rte_mempool *mp, uintptr_t *start,
  *   Index of the next available entry.
  *
  * @return
- *   mr on success, NULL on failure.
+ *   mr on success, NULL on failure and rte_errno is set.
  */
 struct mlx5_mr *
 mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
@@ -144,6 +144,7 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 			      " rte_eth_dev_start()",
 			     (void *)mp, mp->name);
 			rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
+			rte_errno = ENOTSUP;
 			return NULL;
 		}
 		mr = mlx5_mr_new(dev, mp);
@@ -232,7 +233,9 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
 		mlx5_mr_release(mr);
 		return;
 	}
-	mlx5_mr_new(priv->dev, mp);
+	mr = mlx5_mr_new(priv->dev, mp);
+	if (!mr)
+		ERROR("cannot create memory region: %s", strerror(rte_errno));
 }
 
 /**
@@ -245,7 +248,7 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
  *   Pointer to the memory pool to register.
  *
  * @return
- *   The memory region on success.
+ *   The memory region on success, NULL on failure and rte_errno is set.
  */
 struct mlx5_mr *
 mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
@@ -260,11 +263,13 @@ mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 	mr = rte_zmalloc_socket(__func__, sizeof(*mr), 0, mp->socket_id);
 	if (!mr) {
 		DEBUG("unable to configure MR, ibv_reg_mr() failed.");
+		rte_errno = ENOMEM;
 		return NULL;
 	}
 	if (mlx5_check_mempool(mp, &start, &end) != 0) {
 		ERROR("mempool %p: not virtually contiguous",
 		      (void *)mp);
+		rte_errno = ENOMEM;
 		return NULL;
 	}
 	DEBUG("mempool %p area start=%p end=%p size=%zu",
@@ -289,6 +294,10 @@ mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 	      (size_t)(end - start));
 	mr->mr = ibv_reg_mr(priv->pd, (void *)start, end - start,
 			    IBV_ACCESS_LOCAL_WRITE);
+	if (!mr->mr) {
+		rte_errno = ENOMEM;
+		return NULL;
+	}
 	mr->mp = mp;
 	mr->lkey = rte_cpu_to_be_32(mr->mr->lkey);
 	rte_atomic32_inc(&mr->refcnt);
diff --git a/drivers/net/mlx5/mlx5_rss.c b/drivers/net/mlx5/mlx5_rss.c
index badf0c0f9..90682a308 100644
--- a/drivers/net/mlx5/mlx5_rss.c
+++ b/drivers/net/mlx5/mlx5_rss.c
@@ -63,33 +63,31 @@
  *   RSS configuration data.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_rss_hash_update(struct rte_eth_dev *dev,
 		     struct rte_eth_rss_conf *rss_conf)
 {
 	struct priv *priv = dev->data->dev_private;
-	int ret = 0;
 
 	if (rss_conf->rss_hf & MLX5_RSS_HF_MASK) {
-		ret = -EINVAL;
-		goto out;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	if (rss_conf->rss_key && rss_conf->rss_key_len) {
 		priv->rss_conf.rss_key = rte_realloc(priv->rss_conf.rss_key,
 						     rss_conf->rss_key_len, 0);
 		if (!priv->rss_conf.rss_key) {
-			ret = -ENOMEM;
-			goto out;
+			rte_errno = ENOMEM;
+			return -rte_errno;
 		}
 		memcpy(priv->rss_conf.rss_key, rss_conf->rss_key,
 		       rss_conf->rss_key_len);
 		priv->rss_conf.rss_key_len = rss_conf->rss_key_len;
 	}
 	priv->rss_conf.rss_hf = rss_conf->rss_hf;
-out:
-	return ret;
+	return 0;
 }
 
 /**
@@ -101,7 +99,7 @@ mlx5_rss_hash_update(struct rte_eth_dev *dev,
  *   RSS configuration data.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
@@ -109,8 +107,10 @@ mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
 {
 	struct priv *priv = dev->data->dev_private;
 
-	if (!rss_conf)
-		return -EINVAL;
+	if (!rss_conf) {
+		rte_errno = EINVAL;
+		return -rte_errno;
+	}
 	if (rss_conf->rss_key &&
 	    (rss_conf->rss_key_len >= priv->rss_conf.rss_key_len)) {
 		memcpy(rss_conf->rss_key, priv->rss_conf.rss_key,
@@ -130,7 +130,7 @@ mlx5_rss_hash_conf_get(struct rte_eth_dev *dev,
  *   The size of the array to allocate.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_rss_reta_index_resize(struct rte_eth_dev *dev, unsigned int reta_size)
@@ -144,8 +144,10 @@ mlx5_rss_reta_index_resize(struct rte_eth_dev *dev, unsigned int reta_size)
 
 	mem = rte_realloc(priv->reta_idx,
 			  reta_size * sizeof((*priv->reta_idx)[0]), 0);
-	if (!mem)
-		return ENOMEM;
+	if (!mem) {
+		rte_errno = ENOMEM;
+		return -rte_errno;
+	}
 	priv->reta_idx = mem;
 	priv->reta_idx_n = reta_size;
 	if (old_size < reta_size)
@@ -166,7 +168,7 @@ mlx5_rss_reta_index_resize(struct rte_eth_dev *dev, unsigned int reta_size)
  *   Size of the RETA table.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_dev_rss_reta_query(struct rte_eth_dev *dev,
@@ -177,8 +179,10 @@ mlx5_dev_rss_reta_query(struct rte_eth_dev *dev,
 	unsigned int idx;
 	unsigned int i;
 
-	if (!reta_size || reta_size > priv->reta_idx_n)
-		return -EINVAL;
+	if (!reta_size || reta_size > priv->reta_idx_n) {
+		rte_errno = EINVAL;
+		return -rte_errno;
+	}
 	/* Fill each entry of the table even if its bit is not set. */
 	for (idx = 0, i = 0; (i != reta_size); ++i) {
 		idx = i / RTE_RETA_GROUP_SIZE;
@@ -199,7 +203,7 @@ mlx5_dev_rss_reta_query(struct rte_eth_dev *dev,
  *   Size of the RETA table.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_dev_rss_reta_update(struct rte_eth_dev *dev,
@@ -212,8 +216,10 @@ mlx5_dev_rss_reta_update(struct rte_eth_dev *dev,
 	unsigned int i;
 	unsigned int pos;
 
-	if (!reta_size)
-		return -EINVAL;
+	if (!reta_size) {
+		rte_errno = EINVAL;
+		return -rte_errno;
+	}
 	ret = mlx5_rss_reta_index_resize(dev, reta_size);
 	if (ret)
 		return ret;
@@ -227,7 +233,7 @@ mlx5_dev_rss_reta_update(struct rte_eth_dev *dev,
 	}
 	if (dev->data->dev_started) {
 		mlx5_dev_stop(dev);
-		mlx5_dev_start(dev);
+		return mlx5_dev_start(dev);
 	}
-	return -ret;
+	return 0;
 }
diff --git a/drivers/net/mlx5/mlx5_rxmode.c b/drivers/net/mlx5/mlx5_rxmode.c
index 6fb245ba1..09808a39a 100644
--- a/drivers/net/mlx5/mlx5_rxmode.c
+++ b/drivers/net/mlx5/mlx5_rxmode.c
@@ -60,8 +60,13 @@
 void
 mlx5_promiscuous_enable(struct rte_eth_dev *dev)
 {
+	int ret;
+
 	dev->data->promiscuous = 1;
-	mlx5_traffic_restart(dev);
+	ret = mlx5_traffic_restart(dev);
+	if (ret)
+		ERROR("%p cannot enable promiscuous mode: %s", (void *)dev,
+		      strerror(rte_errno));
 }
 
 /**
@@ -73,8 +78,13 @@ mlx5_promiscuous_enable(struct rte_eth_dev *dev)
 void
 mlx5_promiscuous_disable(struct rte_eth_dev *dev)
 {
+	int ret;
+
 	dev->data->promiscuous = 0;
-	mlx5_traffic_restart(dev);
+	ret = mlx5_traffic_restart(dev);
+	if (ret)
+		ERROR("%p cannot disable promiscuous mode: %s", (void *)dev,
+		      strerror(rte_errno));
 }
 
 /**
@@ -86,8 +96,13 @@ mlx5_promiscuous_disable(struct rte_eth_dev *dev)
 void
 mlx5_allmulticast_enable(struct rte_eth_dev *dev)
 {
+	int ret;
+
 	dev->data->all_multicast = 1;
-	mlx5_traffic_restart(dev);
+	ret = mlx5_traffic_restart(dev);
+	if (ret)
+		ERROR("%p cannot enable allmulicast mode: %s", (void *)dev,
+		      strerror(rte_errno));
 }
 
 /**
@@ -99,6 +114,11 @@ mlx5_allmulticast_enable(struct rte_eth_dev *dev)
 void
 mlx5_allmulticast_disable(struct rte_eth_dev *dev)
 {
+	int ret;
+
 	dev->data->all_multicast = 0;
-	mlx5_traffic_restart(dev);
+	ret = mlx5_traffic_restart(dev);
+	if (ret)
+		ERROR("%p cannot disable allmulicast mode: %s", (void *)dev,
+		      strerror(rte_errno));
 }
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index cc1d7ba5d..40a8b72e7 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -88,7 +88,7 @@ const size_t rss_hash_default_key_len = sizeof(rss_hash_default_key);
  *   Pointer to RX queue structure.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
@@ -96,7 +96,7 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 	const unsigned int sges_n = 1 << rxq_ctrl->rxq.sges_n;
 	unsigned int elts_n = 1 << rxq_ctrl->rxq.elts_n;
 	unsigned int i;
-	int ret = 0;
+	int err;
 
 	/* Iterate on segments. */
 	for (i = 0; (i != elts_n); ++i) {
@@ -105,7 +105,7 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 		buf = rte_pktmbuf_alloc(rxq_ctrl->rxq.mp);
 		if (buf == NULL) {
 			ERROR("%p: empty mbuf pool", (void *)rxq_ctrl);
-			ret = ENOMEM;
+			rte_errno = ENOMEM;
 			goto error;
 		}
 		/* Headroom is reserved by rte_pktmbuf_alloc(). */
@@ -147,9 +147,9 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 	}
 	DEBUG("%p: allocated and configured %u segments (max %u packets)",
 	      (void *)rxq_ctrl, elts_n, elts_n / (1 << rxq_ctrl->rxq.sges_n));
-	assert(ret == 0);
 	return 0;
 error:
+	err = rte_errno; /* Save rte_errno before cleanup. */
 	elts_n = i;
 	for (i = 0; (i != elts_n); ++i) {
 		if ((*rxq_ctrl->rxq.elts)[i] != NULL)
@@ -157,8 +157,8 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 		(*rxq_ctrl->rxq.elts)[i] = NULL;
 	}
 	DEBUG("%p: failed, freed everything", (void *)rxq_ctrl);
-	assert(ret > 0);
-	return ret;
+	rte_errno = err; /* Restore rte_errno. */
+	return -rte_errno;
 }
 
 /**
@@ -228,7 +228,7 @@ mlx5_rxq_cleanup(struct mlx5_rxq_ctrl *rxq_ctrl)
  *   Memory pool for buffer allocations.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
@@ -240,7 +240,6 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	struct mlx5_rxq_data *rxq = (*priv->rxqs)[idx];
 	struct mlx5_rxq_ctrl *rxq_ctrl =
 		container_of(rxq, struct mlx5_rxq_ctrl, rxq);
-	int ret = 0;
 
 	if (!rte_is_power_of_2(desc)) {
 		desc = 1 << log2above(desc);
@@ -253,27 +252,27 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	if (idx >= priv->rxqs_n) {
 		ERROR("%p: queue index out of range (%u >= %u)",
 		      (void *)dev, idx, priv->rxqs_n);
-		return -EOVERFLOW;
+		rte_errno = EOVERFLOW;
+		return -rte_errno;
 	}
 	if (!mlx5_rxq_releasable(dev, idx)) {
-		ret = EBUSY;
 		ERROR("%p: unable to release queue index %u",
 		      (void *)dev, idx);
-		goto out;
+		rte_errno = EBUSY;
+		return -rte_errno;
 	}
 	mlx5_rxq_release(dev, idx);
 	rxq_ctrl = mlx5_rxq_new(dev, idx, desc, socket, mp);
 	if (!rxq_ctrl) {
 		ERROR("%p: unable to allocate queue index %u",
 		      (void *)dev, idx);
-		ret = ENOMEM;
-		goto out;
+		rte_errno = ENOMEM;
+		return -rte_errno;
 	}
 	DEBUG("%p: adding RX queue %p to list",
 	      (void *)dev, (void *)rxq_ctrl);
 	(*priv->rxqs)[idx] = &rxq_ctrl->rxq;
-out:
-	return -ret;
+	return 0;
 }
 
 /**
@@ -306,7 +305,7 @@ mlx5_rx_queue_release(void *dpdk_rxq)
  *   Pointer to Ethernet device.
  *
  * @return
- *   0 on success, negative on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
@@ -325,7 +324,8 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 	if (intr_handle->intr_vec == NULL) {
 		ERROR("failed to allocate memory for interrupt vector,"
 		      " Rx interrupts will not be supported");
-		return -ENOMEM;
+		rte_errno = ENOMEM;
+		return -rte_errno;
 	}
 	intr_handle->type = RTE_INTR_HANDLE_EXT;
 	for (i = 0; i != n; ++i) {
@@ -348,16 +348,18 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 			      " (%d), Rx interrupts cannot be enabled",
 			      RTE_MAX_RXTX_INTR_VEC_ID);
 			mlx5_rx_intr_vec_disable(dev);
-			return -1;
+			rte_errno = ENOMEM;
+			return -rte_errno;
 		}
 		fd = rxq_ibv->channel->fd;
 		flags = fcntl(fd, F_GETFL);
 		rc = fcntl(fd, F_SETFL, flags | O_NONBLOCK);
 		if (rc < 0) {
+			rte_errno = errno;
 			ERROR("failed to make Rx interrupt file descriptor"
 			      " %d non-blocking for queue index %d", fd, i);
 			mlx5_rx_intr_vec_disable(dev);
-			return -1;
+			return -rte_errno;
 		}
 		intr_handle->intr_vec[i] = RTE_INTR_VEC_RXTX_OFFSET + count;
 		intr_handle->efds[count] = fd;
@@ -446,7 +448,7 @@ mlx5_arm_cq(struct mlx5_rxq_data *rxq, int sq_n_rxq)
  *   Rx queue number.
  *
  * @return
- *   0 on success, negative on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
@@ -454,12 +456,11 @@ mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_data *rxq_data;
 	struct mlx5_rxq_ctrl *rxq_ctrl;
-	int ret = 0;
 
 	rxq_data = (*priv->rxqs)[rx_queue_id];
 	if (!rxq_data) {
-		ret = EINVAL;
-		goto exit;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	rxq_ctrl = container_of(rxq_data, struct mlx5_rxq_ctrl, rxq);
 	if (rxq_ctrl->irq) {
@@ -467,16 +468,13 @@ mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 
 		rxq_ibv = mlx5_rxq_ibv_get(dev, rx_queue_id);
 		if (!rxq_ibv) {
-			ret = EINVAL;
-			goto exit;
+			rte_errno = EINVAL;
+			return -rte_errno;
 		}
 		mlx5_arm_cq(rxq_data, rxq_data->cq_arm_sn);
 		mlx5_rxq_ibv_release(rxq_ibv);
 	}
-exit:
-	if (ret)
-		WARN("unable to arm interrupt on rx queue %d", rx_queue_id);
-	return -ret;
+	return 0;
 }
 
 /**
@@ -488,7 +486,7 @@ mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
  *   Rx queue number.
  *
  * @return
- *   0 on success, negative on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
@@ -499,35 +497,36 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	struct mlx5_rxq_ibv *rxq_ibv = NULL;
 	struct ibv_cq *ev_cq;
 	void *ev_ctx;
-	int ret = 0;
+	int ret;
 
 	rxq_data = (*priv->rxqs)[rx_queue_id];
 	if (!rxq_data) {
-		ret = EINVAL;
-		goto exit;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	rxq_ctrl = container_of(rxq_data, struct mlx5_rxq_ctrl, rxq);
 	if (!rxq_ctrl->irq)
-		goto exit;
+		return 0;
 	rxq_ibv = mlx5_rxq_ibv_get(dev, rx_queue_id);
 	if (!rxq_ibv) {
-		ret = EINVAL;
-		goto exit;
+		rte_errno = EINVAL;
+		return -rte_errno;
 	}
 	ret = ibv_get_cq_event(rxq_ibv->channel, &ev_cq, &ev_ctx);
 	if (ret || ev_cq != rxq_ibv->cq) {
-		ret = EINVAL;
+		rte_errno = EINVAL;
 		goto exit;
 	}
 	rxq_data->cq_arm_sn++;
 	ibv_ack_cq_events(rxq_ibv->cq, 1);
+	return 0;
 exit:
+	ret = rte_errno; /* Save rte_errno before cleanup. */
 	if (rxq_ibv)
 		mlx5_rxq_ibv_release(rxq_ibv);
-	if (ret)
-		WARN("unable to disable interrupt on rx queue %d",
-		     rx_queue_id);
-	return -ret;
+	WARN("unable to disable interrupt on rx queue %d", rx_queue_id);
+	rte_errno = ret; /* Restore rte_errno. */
+	return -rte_errno;
 }
 
 /**
@@ -539,7 +538,7 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
  *   Queue index in DPDK Rx queue array
  *
  * @return
- *   The Verbs object initialised if it can be created.
+ *   The Verbs object initialised, NULL otherwise and rte_errno is set.
  */
 struct mlx5_rxq_ibv *
 mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
@@ -574,6 +573,7 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (!tmpl) {
 		ERROR("%p: cannot allocate verbs resources",
 		       (void *)rxq_ctrl);
+		rte_errno = ENOMEM;
 		goto error;
 	}
 	tmpl->rxq_ctrl = rxq_ctrl;
@@ -591,6 +591,7 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		if (!tmpl->channel) {
 			ERROR("%p: Comp Channel creation failure",
 			      (void *)rxq_ctrl);
+			rte_errno = ENOMEM;
 			goto error;
 		}
 	}
@@ -619,6 +620,7 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 						    &attr.cq.mlx5));
 	if (tmpl->cq == NULL) {
 		ERROR("%p: CQ creation failure", (void *)rxq_ctrl);
+		rte_errno = ENOMEM;
 		goto error;
 	}
 	DEBUG("priv->device_attr.max_qp_wr is %d",
@@ -655,6 +657,7 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	tmpl->wq = ibv_create_wq(priv->ctx, &attr.wq);
 	if (tmpl->wq == NULL) {
 		ERROR("%p: WQ creation failure", (void *)rxq_ctrl);
+		rte_errno = ENOMEM;
 		goto error;
 	}
 	/*
@@ -669,6 +672,7 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		      ((1 << rxq_data->elts_n) >> rxq_data->sges_n),
 		      (1 << rxq_data->sges_n),
 		      attr.wq.max_wr, attr.wq.max_sge);
+		rte_errno = EINVAL;
 		goto error;
 	}
 	/* Change queue state to ready. */
@@ -680,6 +684,7 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (ret) {
 		ERROR("%p: WQ state to IBV_WQS_RDY failed",
 		      (void *)rxq_ctrl);
+		rte_errno = ret;
 		goto error;
 	}
 	obj.cq.in = tmpl->cq;
@@ -687,11 +692,14 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	obj.rwq.in = tmpl->wq;
 	obj.rwq.out = &rwq;
 	ret = mlx5dv_init_obj(&obj, MLX5DV_OBJ_CQ | MLX5DV_OBJ_RWQ);
-	if (ret != 0)
+	if (ret) {
+		rte_errno = ret;
 		goto error;
+	}
 	if (cq_info.cqe_size != RTE_CACHE_LINE_SIZE) {
 		ERROR("Wrong MLX5_CQE_SIZE environment variable value: "
 		      "it should be set to %u", RTE_CACHE_LINE_SIZE);
+		rte_errno = EINVAL;
 		goto error;
 	}
 	/* Fill the rings. */
@@ -735,6 +743,7 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return tmpl;
 error:
+	ret = rte_errno; /* Save rte_errno before cleanup. */
 	if (tmpl->wq)
 		claim_zero(ibv_destroy_wq(tmpl->wq));
 	if (tmpl->cq)
@@ -744,6 +753,7 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (tmpl->mr)
 		mlx5_mr_release(tmpl->mr);
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
+	rte_errno = ret; /* Restore rte_errno. */
 	return NULL;
 }
 
@@ -866,7 +876,7 @@ mlx5_rxq_ibv_releasable(struct mlx5_rxq_ibv *rxq_ibv)
  *   NUMA socket on which memory must be allocated.
  *
  * @return
- *   A DPDK queue object on success.
+ *   A DPDK queue object on success, NULL otherwise and rte_errno is set.
  */
 struct mlx5_rxq_ctrl *
 mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
@@ -882,8 +892,10 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 				 sizeof(*tmpl) +
 				 desc_n * sizeof(struct rte_mbuf *),
 				 0, socket);
-	if (!tmpl)
+	if (!tmpl) {
+		rte_errno = ENOMEM;
 		return NULL;
+	}
 	tmpl->socket = socket;
 	if (priv->dev->data->dev_conf.intr_conf.rxq)
 		tmpl->irq = 1;
@@ -913,6 +925,7 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 			      (void *)dev,
 			      1 << sges_n,
 			      dev->data->dev_conf.rxmode.max_rx_pkt_len);
+			rte_errno = EOVERFLOW;
 			goto error;
 		}
 	} else {
@@ -931,6 +944,7 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		      (void *)dev,
 		      desc,
 		      1 << tmpl->rxq.sges_n);
+		rte_errno = EINVAL;
 		goto error;
 	}
 	/* Toggle RX checksum offload if hardware supports it. */
@@ -989,7 +1003,7 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
  *   TX queue index.
  *
  * @return
- *   A pointer to the queue if it exists.
+ *   A pointer to the queue if it exists, NULL otherwise.
  */
 struct mlx5_rxq_ctrl *
 mlx5_rxq_get(struct rte_eth_dev *dev, uint16_t idx)
@@ -1052,7 +1066,8 @@ mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx)
  *   TX queue index.
  *
  * @return
- *   1 if the queue can be released.
+ *   1 if the queue can be released, negative errno otherwise and rte_errno is
+ *   set.
  */
 int
 mlx5_rxq_releasable(struct rte_eth_dev *dev, uint16_t idx)
@@ -1060,8 +1075,10 @@ mlx5_rxq_releasable(struct rte_eth_dev *dev, uint16_t idx)
 	struct priv *priv = dev->data->dev_private;
 	struct mlx5_rxq_ctrl *rxq_ctrl;
 
-	if (!(*priv->rxqs)[idx])
-		return -1;
+	if (!(*priv->rxqs)[idx]) {
+		rte_errno = EINVAL;
+		return -rte_errno;
+	}
 	rxq_ctrl = container_of((*priv->rxqs)[idx], struct mlx5_rxq_ctrl, rxq);
 	return (rte_atomic32_read(&rxq_ctrl->refcnt) == 1);
 }
@@ -1101,7 +1118,7 @@ mlx5_rxq_verify(struct rte_eth_dev *dev)
  *   Number of queues in the array.
  *
  * @return
- *   A new indirection table.
+ *   The Verbs object initialised, NULL otherwise and rte_errno is set.
  */
 struct mlx5_ind_table_ibv *
 mlx5_ind_table_ibv_new(struct rte_eth_dev *dev, uint16_t queues[],
@@ -1118,8 +1135,10 @@ mlx5_ind_table_ibv_new(struct rte_eth_dev *dev, uint16_t queues[],
 
 	ind_tbl = rte_calloc(__func__, 1, sizeof(*ind_tbl) +
 			     queues_n * sizeof(uint16_t), 0);
-	if (!ind_tbl)
+	if (!ind_tbl) {
+		rte_errno = ENOMEM;
 		return NULL;
+	}
 	for (i = 0; i != queues_n; ++i) {
 		struct mlx5_rxq_ctrl *rxq = mlx5_rxq_get(dev, queues[i]);
 
@@ -1139,8 +1158,10 @@ mlx5_ind_table_ibv_new(struct rte_eth_dev *dev, uint16_t queues[],
 			.ind_tbl = wq,
 			.comp_mask = 0,
 		});
-	if (!ind_tbl->ind_table)
+	if (!ind_tbl->ind_table) {
+		rte_errno = errno;
 		goto error;
+	}
 	rte_atomic32_inc(&ind_tbl->refcnt);
 	LIST_INSERT_HEAD(&priv->ind_tbls, ind_tbl, next);
 	DEBUG("%p: Indirection table %p: refcnt %d", (void *)dev,
@@ -1264,7 +1285,7 @@ mlx5_ind_table_ibv_verify(struct rte_eth_dev *dev)
  *   Number of queues.
  *
  * @return
- *   An hash Rx queue on success.
+ *   The Verbs object initialised, NULL otherwise and rte_errno is set.
  */
 struct mlx5_hrxq *
 mlx5_hrxq_new(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
@@ -1274,13 +1295,16 @@ mlx5_hrxq_new(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 	struct mlx5_hrxq *hrxq;
 	struct mlx5_ind_table_ibv *ind_tbl;
 	struct ibv_qp *qp;
+	int err;
 
 	queues_n = hash_fields ? queues_n : 1;
 	ind_tbl = mlx5_ind_table_ibv_get(dev, queues, queues_n);
 	if (!ind_tbl)
 		ind_tbl = mlx5_ind_table_ibv_new(dev, queues, queues_n);
-	if (!ind_tbl)
+	if (!ind_tbl) {
+		rte_errno = ENOMEM;
 		return NULL;
+	}
 	qp = ibv_create_qp_ex(
 		priv->ctx,
 		&(struct ibv_qp_init_attr_ex){
@@ -1298,8 +1322,10 @@ mlx5_hrxq_new(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 			.rwq_ind_tbl = ind_tbl->ind_table,
 			.pd = priv->pd,
 		});
-	if (!qp)
+	if (!qp) {
+		rte_errno = errno;
 		goto error;
+	}
 	hrxq = rte_calloc(__func__, 1, sizeof(*hrxq) + rss_key_len, 0);
 	if (!hrxq)
 		goto error;
@@ -1314,9 +1340,11 @@ mlx5_hrxq_new(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 	      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
 	return hrxq;
 error:
+	err = rte_errno; /* Save rte_errno before cleanup. */
 	mlx5_ind_table_ibv_release(dev, ind_tbl);
 	if (qp)
 		claim_zero(ibv_destroy_qp(qp));
+	rte_errno = err; /* Restore rte_errno. */
 	return NULL;
 }
 
diff --git a/drivers/net/mlx5/mlx5_socket.c b/drivers/net/mlx5/mlx5_socket.c
index 8f400d06e..5499a01b9 100644
--- a/drivers/net/mlx5/mlx5_socket.c
+++ b/drivers/net/mlx5/mlx5_socket.c
@@ -49,7 +49,7 @@
  *   Pointer to Ethernet device.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_socket_init(struct rte_eth_dev *dev)
@@ -67,45 +67,51 @@ mlx5_socket_init(struct rte_eth_dev *dev)
 	 */
 	ret = socket(AF_UNIX, SOCK_STREAM, 0);
 	if (ret < 0) {
+		rte_errno = errno;
 		WARN("secondary process not supported: %s", strerror(errno));
-		return ret;
+		goto error;
 	}
 	priv->primary_socket = ret;
 	flags = fcntl(priv->primary_socket, F_GETFL, 0);
-	if (flags == -1)
-		goto out;
+	if (flags == -1) {
+		rte_errno = errno;
+		goto error;
+	}
 	ret = fcntl(priv->primary_socket, F_SETFL, flags | O_NONBLOCK);
-	if (ret < 0)
-		goto out;
+	if (ret < 0) {
+		rte_errno = errno;
+		goto error;
+	}
 	snprintf(sun.sun_path, sizeof(sun.sun_path), "/var/tmp/%s_%d",
 		 MLX5_DRIVER_NAME, priv->primary_socket);
 	remove(sun.sun_path);
 	ret = bind(priv->primary_socket, (const struct sockaddr *)&sun,
 		   sizeof(sun));
 	if (ret < 0) {
+		rte_errno = errno;
 		WARN("cannot bind socket, secondary process not supported: %s",
 		     strerror(errno));
 		goto close;
 	}
 	ret = listen(priv->primary_socket, 0);
 	if (ret < 0) {
+		rte_errno = errno;
 		WARN("Secondary process not supported: %s", strerror(errno));
 		goto close;
 	}
-	return ret;
+	return 0;
 close:
 	remove(sun.sun_path);
-out:
+error:
 	claim_zero(close(priv->primary_socket));
 	priv->primary_socket = 0;
-	return -(ret);
+	return -rte_errno;
 }
 
 /**
  * Un-Initialise the socket to communicate with the secondary process
  *
  * @param[in] dev
- *   Pointer to Ethernet device.
  */
 void
 mlx5_socket_uninit(struct rte_eth_dev *dev)
@@ -155,19 +161,21 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	ret = setsockopt(conn_sock, SOL_SOCKET, SO_PASSCRED, &(int){1},
 					 sizeof(int));
 	if (ret < 0) {
-		WARN("cannot change socket options");
-		goto out;
+		ret = errno;
+		WARN("cannot change socket options: %s", strerror(rte_errno));
+		goto error;
 	}
 	ret = recvmsg(conn_sock, &msg, MSG_WAITALL);
 	if (ret < 0) {
-		WARN("received an empty message: %s", strerror(errno));
-		goto out;
+		ret = errno;
+		WARN("received an empty message: %s", strerror(rte_errno));
+		goto error;
 	}
 	/* Expect to receive credentials only. */
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
 		WARN("no message");
-		goto out;
+		goto error;
 	}
 	if ((cmsg->cmsg_type == SCM_CREDENTIALS) &&
 		(cmsg->cmsg_len >= sizeof(*cred))) {
@@ -177,13 +185,13 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	cmsg = CMSG_NXTHDR(&msg, cmsg);
 	if (cmsg != NULL) {
 		WARN("Message wrongly formatted");
-		goto out;
+		goto error;
 	}
 	/* Make sure all the ancillary data was received and valid. */
 	if ((cred == NULL) || (cred->uid != getuid()) ||
 	    (cred->gid != getgid())) {
 		WARN("wrong credentials");
-		goto out;
+		goto error;
 	}
 	/* Set-up the ancillary data. */
 	cmsg = CMSG_FIRSTHDR(&msg);
@@ -196,7 +204,7 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	ret = sendmsg(conn_sock, &msg, 0);
 	if (ret < 0)
 		WARN("cannot send response");
-out:
+error:
 	close(conn_sock);
 }
 
@@ -207,7 +215,7 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
  *   Pointer to Ethernet structure.
  *
  * @return
- *   fd on success, negative errno value on failure.
+ *   fd on success, negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_socket_connect(struct rte_eth_dev *dev)
@@ -216,7 +224,7 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	struct sockaddr_un sun = {
 		.sun_family = AF_UNIX,
 	};
-	int socket_fd;
+	int socket_fd = -1;
 	int *fd = NULL;
 	int ret;
 	struct ucred *cred;
@@ -236,57 +244,67 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 
 	ret = socket(AF_UNIX, SOCK_STREAM, 0);
 	if (ret < 0) {
+		rte_errno = errno;
 		WARN("cannot connect to primary");
-		return ret;
+		goto error;
 	}
 	socket_fd = ret;
 	snprintf(sun.sun_path, sizeof(sun.sun_path), "/var/tmp/%s_%d",
 		 MLX5_DRIVER_NAME, priv->primary_socket);
 	ret = connect(socket_fd, (const struct sockaddr *)&sun, sizeof(sun));
 	if (ret < 0) {
+		rte_errno = errno;
 		WARN("cannot connect to primary");
-		goto out;
+		goto error;
 	}
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
+		rte_errno = EINVAL;
 		DEBUG("cannot get first message");
-		goto out;
+		goto error;
 	}
 	cmsg->cmsg_level = SOL_SOCKET;
 	cmsg->cmsg_type = SCM_CREDENTIALS;
 	cmsg->cmsg_len = CMSG_LEN(sizeof(*cred));
 	cred = (struct ucred *)CMSG_DATA(cmsg);
 	if (cred == NULL) {
+		rte_errno = EINVAL;
 		DEBUG("no credentials received");
-		goto out;
+		goto error;
 	}
 	cred->pid = getpid();
 	cred->uid = getuid();
 	cred->gid = getgid();
 	ret = sendmsg(socket_fd, &msg, MSG_DONTWAIT);
 	if (ret < 0) {
+		rte_errno = errno;
 		WARN("cannot send credentials to primary: %s",
 		     strerror(errno));
-		goto out;
+		goto error;
 	}
 	ret = recvmsg(socket_fd, &msg, MSG_WAITALL);
 	if (ret <= 0) {
+		rte_errno = errno;
 		WARN("no message from primary: %s", strerror(errno));
-		goto out;
+		goto error;
 	}
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
+		rte_errno = EINVAL;
 		WARN("No file descriptor received");
-		goto out;
+		goto error;
 	}
 	fd = (int *)CMSG_DATA(cmsg);
-	if (*fd <= 0) {
+	if (*fd < 0) {
 		WARN("no file descriptor received: %s", strerror(errno));
-		ret = *fd;
-		goto out;
+		rte_errno = *fd;
+		goto error;
 	}
 	ret = *fd;
-out:
 	close(socket_fd);
-	return ret;
+	return 0;
+error:
+	if (socket_fd != -1)
+		close(socket_fd);
+	return -rte_errno;
 }
diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index 6d454e5e8..b3500df97 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -140,7 +140,8 @@ static const unsigned int xstats_n = RTE_DIM(mlx5_counters_init);
  *   Counters table output buffer.
  *
  * @return
- *   0 on success and stats is filled, negative on error.
+ *   0 on success and stats is filled, negative errno value otherwise and
+ *   rte_errno is set.
  */
 static int
 mlx5_read_dev_counters(struct rte_eth_dev *dev, uint64_t *stats)
@@ -152,13 +153,15 @@ mlx5_read_dev_counters(struct rte_eth_dev *dev, uint64_t *stats)
 	unsigned int stats_sz = xstats_ctrl->stats_n * sizeof(uint64_t);
 	unsigned char et_stat_buf[sizeof(struct ethtool_stats) + stats_sz];
 	struct ethtool_stats *et_stats = (struct ethtool_stats *)et_stat_buf;
+	int ret;
 
 	et_stats->cmd = ETHTOOL_GSTATS;
 	et_stats->n_stats = xstats_ctrl->stats_n;
 	ifr.ifr_data = (caddr_t)et_stats;
-	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr) != 0) {
+	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+	if (ret) {
 		WARN("unable to read statistic values from device");
-		return -1;
+		return ret;
 	}
 	for (i = 0; i != xstats_n; ++i) {
 		if (mlx5_counters_init[i].ib) {
@@ -190,18 +193,21 @@ mlx5_read_dev_counters(struct rte_eth_dev *dev, uint64_t *stats)
  *   Pointer to Ethernet device.
  *
  * @return
- *   Number of statistics on success, -1 on error.
+ *   Number of statistics on success, negative errno value otherwise and
+ *   rte_errno is set.
  */
 static int
 mlx5_ethtool_get_stats_n(struct rte_eth_dev *dev) {
 	struct ethtool_drvinfo drvinfo;
 	struct ifreq ifr;
+	int ret;
 
 	drvinfo.cmd = ETHTOOL_GDRVINFO;
 	ifr.ifr_data = (caddr_t)&drvinfo;
-	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr) != 0) {
+	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+	if (ret) {
 		WARN("unable to query number of statistics");
-		return -1;
+		return ret;
 	}
 	return drvinfo.n_stats;
 }
@@ -223,12 +229,14 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 	struct ethtool_gstrings *strings = NULL;
 	unsigned int dev_stats_n;
 	unsigned int str_sz;
+	int ret;
 
-	dev_stats_n = mlx5_ethtool_get_stats_n(dev);
-	if (dev_stats_n < 1) {
+	ret = mlx5_ethtool_get_stats_n(dev);
+	if (ret < 0) {
 		WARN("no extended statistics available");
 		return;
 	}
+	dev_stats_n = ret;
 	xstats_ctrl->stats_n = dev_stats_n;
 	/* Allocate memory to grab stat names and values. */
 	str_sz = dev_stats_n * ETH_GSTRING_LEN;
@@ -243,7 +251,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 	strings->string_set = ETH_SS_STATS;
 	strings->len = dev_stats_n;
 	ifr.ifr_data = (caddr_t)strings;
-	if (mlx5_ifreq(dev, SIOCETHTOOL, &ifr) != 0) {
+	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+	if (ret) {
 		WARN("unable to get statistic names");
 		goto free;
 	}
@@ -272,7 +281,9 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 	}
 	/* Copy to base at first time. */
 	assert(xstats_n <= MLX5_MAX_XSTATS);
-	mlx5_read_dev_counters(dev, xstats_ctrl->base);
+	ret = mlx5_read_dev_counters(dev, xstats_ctrl->base);
+	if (ret)
+		ERROR("cannot read device counters: %s", strerror(rte_errno));
 free:
 	rte_free(strings);
 }
@@ -289,7 +300,7 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
  *
  * @return
  *   Number of extended stats on success and stats is filled,
- *   negative on error.
+ *   negative on error and rte_errno is set.
  */
 int
 mlx5_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *stats,
@@ -298,15 +309,15 @@ mlx5_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *stats,
 	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 	uint64_t counters[n];
-	int ret = 0;
 
 	if (n >= xstats_n && stats) {
 		struct mlx5_xstats_ctrl *xstats_ctrl = &priv->xstats_ctrl;
 		int stats_n;
+		int ret;
 
 		stats_n = mlx5_ethtool_get_stats_n(dev);
 		if (stats_n < 0)
-			return -1;
+			return stats_n;
 		if (xstats_ctrl->stats_n != stats_n)
 			mlx5_xstats_init(dev);
 		ret = mlx5_read_dev_counters(dev, counters);
@@ -327,6 +338,10 @@ mlx5_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *stats,
  *   Pointer to Ethernet device structure.
  * @param[out] stats
  *   Stats structure output buffer.
+ *
+ * @return
+ *   0 on success and stats is filled, negative errno value otherwise and
+ *   rte_errno is set.
  */
 int
 mlx5_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
@@ -431,14 +446,22 @@ mlx5_xstats_reset(struct rte_eth_dev *dev)
 	unsigned int i;
 	unsigned int n = xstats_n;
 	uint64_t counters[n];
+	int ret;
 
 	stats_n = mlx5_ethtool_get_stats_n(dev);
-	if (stats_n < 0)
+	if (stats_n < 0) {
+		ERROR("%p cannot get stats: %s", (void *)dev,
+		      strerror(-stats_n));
 		return;
+	}
 	if (xstats_ctrl->stats_n != stats_n)
 		mlx5_xstats_init(dev);
-	if (mlx5_read_dev_counters(dev, counters) < 0)
+	ret = mlx5_read_dev_counters(dev, counters);
+	if (ret) {
+		ERROR("%p cannot read device counters: %s", (void *)dev,
+		      strerror(rte_errno));
 		return;
+	}
 	for (i = 0; i != n; ++i)
 		xstats_ctrl->base[i] = counters[i];
 }
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index 19434b921..5d2eff506 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -64,14 +64,14 @@ mlx5_txq_stop(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, errno on error.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_txq_start(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
-	int ret = 0;
+	int ret;
 
 	/* Add memory regions to Tx queues. */
 	for (i = 0; i != priv->txqs_n; ++i) {
@@ -89,17 +89,19 @@ mlx5_txq_start(struct rte_eth_dev *dev)
 		txq_alloc_elts(txq_ctrl);
 		txq_ctrl->ibv = mlx5_txq_ibv_new(dev, i);
 		if (!txq_ctrl->ibv) {
-			ret = ENOMEM;
+			rte_errno = ENOMEM;
 			goto error;
 		}
 	}
 	ret = mlx5_tx_uar_remap(dev, priv->ctx->cmd_fd);
 	if (ret)
 		goto error;
-	return ret;
+	return 0;
 error:
+	ret = rte_errno; /* Save rte_errno before cleanup. */
 	mlx5_txq_stop(dev);
-	return ret;
+	rte_errno = ret; /* Restore rte_errno. */
+	return -rte_errno;
 }
 
 /**
@@ -125,7 +127,7 @@ mlx5_rxq_stop(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, errno on error.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
 mlx5_rxq_start(struct rte_eth_dev *dev)
@@ -143,15 +145,15 @@ mlx5_rxq_start(struct rte_eth_dev *dev)
 		if (ret)
 			goto error;
 		rxq_ctrl->ibv = mlx5_rxq_ibv_new(dev, i);
-		if (!rxq_ctrl->ibv) {
-			ret = ENOMEM;
+		if (!rxq_ctrl->ibv)
 			goto error;
-		}
 	}
-	return -ret;
+	return 0;
 error:
+	ret = rte_errno; /* Save rte_errno before cleanup. */
 	mlx5_rxq_stop(dev);
-	return -ret;
+	rte_errno = ret; /* Restore rte_errno. */
+	return -rte_errno;
 }
 
 /**
@@ -163,48 +165,48 @@ mlx5_rxq_start(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_dev_start(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
 	struct mlx5_mr *mr = NULL;
-	int err;
+	int ret;
 
 	dev->data->dev_started = 1;
-	err = mlx5_flow_create_drop_queue(dev);
-	if (err) {
+	ret = mlx5_flow_create_drop_queue(dev);
+	if (ret) {
 		ERROR("%p: Drop queue allocation failed: %s",
-		      (void *)dev, strerror(err));
+		      (void *)dev, strerror(rte_errno));
 		goto error;
 	}
 	DEBUG("%p: allocating and configuring hash RX queues", (void *)dev);
 	rte_mempool_walk(mlx5_mp2mr_iter, priv);
-	err = mlx5_txq_start(dev);
-	if (err) {
-		ERROR("%p: TXQ allocation failed: %s",
-		      (void *)dev, strerror(err));
+	ret = mlx5_txq_start(dev);
+	if (ret) {
+		ERROR("%p: Tx Queue allocation failed: %s",
+		      (void *)dev, strerror(rte_errno));
 		goto error;
 	}
-	err = mlx5_rxq_start(dev);
-	if (err) {
-		ERROR("%p: RXQ allocation failed: %s",
-		      (void *)dev, strerror(err));
+	ret = mlx5_rxq_start(dev);
+	if (ret) {
+		ERROR("%p: Rx Queue allocation failed: %s",
+		      (void *)dev, strerror(rte_errno));
 		goto error;
 	}
-	err = mlx5_rx_intr_vec_enable(dev);
-	if (err) {
-		ERROR("%p: RX interrupt vector creation failed",
-		      (void *)priv);
+	ret = mlx5_rx_intr_vec_enable(dev);
+	if (ret) {
+		ERROR("%p: Rx interrupt vector creation failed",
+		      (void *)dev);
 		goto error;
 	}
 	mlx5_xstats_init(dev);
 	/* Update link status and Tx/Rx callbacks for the first time. */
 	memset(&dev->data->dev_link, 0, sizeof(struct rte_eth_link));
 	INFO("Forcing port %u link to be up", dev->data->port_id);
-	err = mlx5_force_link_status_change(dev, ETH_LINK_UP);
-	if (err) {
+	ret = mlx5_force_link_status_change(dev, ETH_LINK_UP);
+	if (ret) {
 		DEBUG("Failed to set port %u link to be up",
 		      dev->data->port_id);
 		goto error;
@@ -212,6 +214,7 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 	mlx5_dev_interrupt_handler_install(dev);
 	return 0;
 error:
+	ret = rte_errno; /* Save rte_errno before cleanup. */
 	/* Rollback. */
 	dev->data->dev_started = 0;
 	for (mr = LIST_FIRST(&priv->mr); mr; mr = LIST_FIRST(&priv->mr))
@@ -221,7 +224,8 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 	mlx5_txq_stop(dev);
 	mlx5_rxq_stop(dev);
 	mlx5_flow_delete_drop_queue(dev);
-	return err;
+	rte_errno = ret; /* Restore rte_errno. */
+	return -rte_errno;
 }
 
 /**
@@ -265,7 +269,7 @@ mlx5_dev_stop(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device structure.
  *
  * @return
- *   0 on success.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_traffic_enable(struct rte_eth_dev *dev)
@@ -303,8 +307,9 @@ mlx5_traffic_enable(struct rte_eth_dev *dev)
 			.type = 0,
 		};
 
-		claim_zero(mlx5_ctrl_flow(dev, &promisc, &promisc));
-		return 0;
+		ret = mlx5_ctrl_flow(dev, &promisc, &promisc);
+		if (ret)
+			goto error;
 	}
 	if (dev->data->all_multicast) {
 		struct rte_flow_item_eth multicast = {
@@ -313,7 +318,9 @@ mlx5_traffic_enable(struct rte_eth_dev *dev)
 			.type = 0,
 		};
 
-		claim_zero(mlx5_ctrl_flow(dev, &multicast, &multicast));
+		ret = mlx5_ctrl_flow(dev, &multicast, &multicast);
+		if (ret)
+			goto error;
 	} else {
 		/* Add broadcast/multicast flows. */
 		for (i = 0; i != vlan_filter_n; ++i) {
@@ -373,15 +380,17 @@ mlx5_traffic_enable(struct rte_eth_dev *dev)
 				goto error;
 		}
 		if (!vlan_filter_n) {
-			ret = mlx5_ctrl_flow(dev, &unicast,
-					     &unicast_mask);
+			ret = mlx5_ctrl_flow(dev, &unicast, &unicast_mask);
 			if (ret)
 				goto error;
 		}
 	}
 	return 0;
 error:
-	return rte_errno;
+	ret = rte_errno; /* Save rte_errno before cleanup. */
+	mlx5_flow_list_flush(dev, &priv->ctrl_flows);
+	rte_errno = ret; /* Restore rte_errno. */
+	return -rte_errno;
 }
 
 
@@ -406,14 +415,14 @@ mlx5_traffic_disable(struct rte_eth_dev *dev)
  *   Pointer to Ethernet device private data.
  *
  * @return
- *   0 on success.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_traffic_restart(struct rte_eth_dev *dev)
 {
 	if (dev->data->dev_started) {
 		mlx5_traffic_disable(dev);
-		mlx5_traffic_enable(dev);
+		return mlx5_traffic_enable(dev);
 	}
 	return 0;
 }
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 53a21c259..baf3fe984 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -130,7 +130,7 @@ txq_free_elts(struct mlx5_txq_ctrl *txq_ctrl)
  *   Thresholds parameters.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
@@ -140,7 +140,6 @@ mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	struct mlx5_txq_data *txq = (*priv->txqs)[idx];
 	struct mlx5_txq_ctrl *txq_ctrl =
 		container_of(txq, struct mlx5_txq_ctrl, txq);
-	int ret = 0;
 
 	if (desc <= MLX5_TX_COMP_THRESH) {
 		WARN("%p: number of descriptors requested for TX queue %u"
@@ -160,27 +159,26 @@ mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	if (idx >= priv->txqs_n) {
 		ERROR("%p: queue index out of range (%u >= %u)",
 		      (void *)dev, idx, priv->txqs_n);
-		return -EOVERFLOW;
+		rte_errno = EOVERFLOW;
+		return -rte_errno;
 	}
 	if (!mlx5_txq_releasable(dev, idx)) {
-		ret = EBUSY;
+		rte_errno = EBUSY;
 		ERROR("%p: unable to release queue index %u",
 		      (void *)dev, idx);
-		goto out;
+		return -rte_errno;
 	}
 	mlx5_txq_release(dev, idx);
 	txq_ctrl = mlx5_txq_new(dev, idx, desc, socket, conf);
 	if (!txq_ctrl) {
 		ERROR("%p: unable to allocate queue index %u",
 		      (void *)dev, idx);
-		ret = ENOMEM;
-		goto out;
+		return -rte_errno;
 	}
 	DEBUG("%p: adding TX queue %p to list",
 	      (void *)dev, (void *)txq_ctrl);
 	(*priv->txqs)[idx] = &txq_ctrl->txq;
-out:
-	return -ret;
+	return 0;
 }
 
 /**
@@ -203,9 +201,9 @@ mlx5_tx_queue_release(void *dpdk_txq)
 	priv = txq_ctrl->priv;
 	for (i = 0; (i != priv->txqs_n); ++i)
 		if ((*priv->txqs)[i] == txq) {
+			mlx5_txq_release(priv->dev, i);
 			DEBUG("%p: removing TX queue %p from list",
 			      (void *)priv->dev, (void *)txq_ctrl);
-			mlx5_txq_release(priv->dev, i);
 			break;
 		}
 }
@@ -222,7 +220,7 @@ mlx5_tx_queue_release(void *dpdk_txq)
  *   Verbs file descriptor to map UAR pages.
  *
  * @return
- *   0 on success, errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd)
@@ -239,7 +237,6 @@ mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd)
 	struct mlx5_txq_ctrl *txq_ctrl;
 	int already_mapped;
 	size_t page_size = sysconf(_SC_PAGESIZE);
-	int r;
 
 	memset(pages, 0, priv->txqs_n * sizeof(uintptr_t));
 	/*
@@ -278,8 +275,8 @@ mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd)
 				/* fixed mmap have to return same address */
 				ERROR("call to mmap failed on UAR for txq %d\n",
 				      i);
-				r = ENXIO;
-				return r;
+				rte_errno = ENXIO;
+				return -rte_errno;
 			}
 		}
 		if (rte_eal_process_type() == RTE_PROC_PRIMARY) /* save once */
@@ -300,7 +297,7 @@ mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd)
  *   Queue index in DPDK Rx queue array
  *
  * @return
- *   The Verbs object initialised if it can be created.
+ *   The Verbs object initialised, NULL otherwise and rte_errno is set.
  */
 struct mlx5_txq_ibv *
 mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
@@ -329,7 +326,8 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	priv->verbs_alloc_ctx.obj = txq_ctrl;
 	if (mlx5_getenv_int("MLX5_ENABLE_CQE_COMPRESSION")) {
 		ERROR("MLX5_ENABLE_CQE_COMPRESSION must never be set");
-		goto error;
+		rte_errno = EINVAL;
+		return NULL;
 	}
 	memset(&tmpl, 0, sizeof(struct mlx5_txq_ibv));
 	/* MRs will be registered in mp2mr[] later. */
@@ -343,6 +341,7 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	tmpl.cq = ibv_create_cq(priv->ctx, cqe_n, NULL, NULL, 0);
 	if (tmpl.cq == NULL) {
 		ERROR("%p: CQ creation failure", (void *)txq_ctrl);
+		rte_errno = errno;
 		goto error;
 	}
 	attr.init = (struct ibv_qp_init_attr_ex){
@@ -384,6 +383,7 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	tmpl.qp = ibv_create_qp_ex(priv->ctx, &attr.init);
 	if (tmpl.qp == NULL) {
 		ERROR("%p: QP creation failure", (void *)txq_ctrl);
+		rte_errno = errno;
 		goto error;
 	}
 	attr.mod = (struct ibv_qp_attr){
@@ -395,6 +395,7 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, (IBV_QP_STATE | IBV_QP_PORT));
 	if (ret) {
 		ERROR("%p: QP state to IBV_QPS_INIT failed", (void *)txq_ctrl);
+		rte_errno = errno;
 		goto error;
 	}
 	attr.mod = (struct ibv_qp_attr){
@@ -403,18 +404,21 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, IBV_QP_STATE);
 	if (ret) {
 		ERROR("%p: QP state to IBV_QPS_RTR failed", (void *)txq_ctrl);
+		rte_errno = errno;
 		goto error;
 	}
 	attr.mod.qp_state = IBV_QPS_RTS;
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, IBV_QP_STATE);
 	if (ret) {
 		ERROR("%p: QP state to IBV_QPS_RTS failed", (void *)txq_ctrl);
+		rte_errno = errno;
 		goto error;
 	}
 	txq_ibv = rte_calloc_socket(__func__, 1, sizeof(struct mlx5_txq_ibv), 0,
 				    txq_ctrl->socket);
 	if (!txq_ibv) {
 		ERROR("%p: cannot allocate memory", (void *)txq_ctrl);
+		rte_errno = ENOMEM;
 		goto error;
 	}
 	obj.cq.in = tmpl.cq;
@@ -422,11 +426,14 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	obj.qp.in = tmpl.qp;
 	obj.qp.out = &qp;
 	ret = mlx5dv_init_obj(&obj, MLX5DV_OBJ_CQ | MLX5DV_OBJ_QP);
-	if (ret != 0)
+	if (ret != 0) {
+		rte_errno = errno;
 		goto error;
+	}
 	if (cq_info.cqe_size != RTE_CACHE_LINE_SIZE) {
 		ERROR("Wrong MLX5_CQE_SIZE environment variable value: "
 		      "it should be set to %u", RTE_CACHE_LINE_SIZE);
+		rte_errno = EINVAL;
 		goto error;
 	}
 	txq_data->cqe_n = log2above(cq_info.cqe_cnt);
@@ -450,6 +457,7 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		txq_ctrl->uar_mmap_offset = qp.uar_mmap_offset;
 	} else {
 		ERROR("Failed to retrieve UAR info, invalid libmlx5.so version");
+		rte_errno = EINVAL;
 		goto error;
 	}
 	DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)dev,
@@ -458,11 +466,13 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return txq_ibv;
 error:
+	ret = rte_errno; /* Save rte_errno before cleanup. */
 	if (tmpl.cq)
 		claim_zero(ibv_destroy_cq(tmpl.cq));
 	if (tmpl.qp)
 		claim_zero(ibv_destroy_qp(tmpl.qp));
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
+	rte_errno = ret; /* Restore rte_errno. */
 	return NULL;
 }
 
@@ -574,7 +584,7 @@ mlx5_txq_ibv_verify(struct rte_eth_dev *dev)
  *  Thresholds parameters.
  *
  * @return
- *   A DPDK queue object on success.
+ *   A DPDK queue object on success, NULL otherwise and rte_errno is set.
  */
 struct mlx5_txq_ctrl *
 mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
@@ -590,8 +600,10 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 				 sizeof(*tmpl) +
 				 desc * sizeof(struct rte_mbuf *),
 				 0, socket);
-	if (!tmpl)
+	if (!tmpl) {
+		rte_errno = ENOMEM;
 		return NULL;
+	}
 	assert(desc > MLX5_TX_COMP_THRESH);
 	tmpl->txq.flags = conf->txq_flags;
 	tmpl->priv = priv;
diff --git a/drivers/net/mlx5/mlx5_vlan.c b/drivers/net/mlx5/mlx5_vlan.c
index 7f070058e..7e4830138 100644
--- a/drivers/net/mlx5/mlx5_vlan.c
+++ b/drivers/net/mlx5/mlx5_vlan.c
@@ -54,14 +54,13 @@
  *   Toggle filter.
  *
  * @return
- *   0 on success, negative errno value on failure.
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
 {
 	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
-	int ret = 0;
 
 	DEBUG("%p: %s VLAN filter ID %" PRIu16,
 	      (void *)dev, (on ? "enable" : "disable"), vlan_id);
@@ -71,8 +70,8 @@ mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
 			break;
 	/* Check if there's room for another VLAN filter. */
 	if (i == RTE_DIM(priv->vlan_filter)) {
-		ret = -ENOMEM;
-		goto out;
+		rte_errno = ENOMEM;
+		return -rte_errno;
 	}
 	if (i < priv->vlan_filter_n) {
 		assert(priv->vlan_filter_n != 0);
@@ -95,10 +94,10 @@ mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
 		priv->vlan_filter[priv->vlan_filter_n] = vlan_id;
 		++priv->vlan_filter_n;
 	}
-	if (dev->data->dev_started)
-		mlx5_traffic_restart(dev);
 out:
-	return ret;
+	if (dev->data->dev_started)
+		return mlx5_traffic_restart(dev);
+	return 0;
 }
 
 /**
@@ -122,7 +121,7 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 	uint16_t vlan_offloads =
 		(on ? IBV_WQ_FLAGS_CVLAN_STRIPPING : 0) |
 		0;
-	int err;
+	int ret;
 
 	/* Validate hw support */
 	if (!priv->hw_vlan_strip) {
@@ -146,10 +145,10 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 		.flags_mask = IBV_WQ_FLAGS_CVLAN_STRIPPING,
 		.flags = vlan_offloads,
 	};
-	err = ibv_modify_wq(rxq_ctrl->ibv->wq, &mod);
-	if (err) {
+	ret = ibv_modify_wq(rxq_ctrl->ibv->wq, &mod);
+	if (ret) {
 		ERROR("%p: failed to modified stripping mode: %s",
-		      (void *)dev, strerror(err));
+		      (void *)dev, strerror(rte_errno));
 		return;
 	}
 	/* Update related bits in RX queue. */
@@ -163,6 +162,9 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
  *   Pointer to Ethernet device structure.
  * @param mask
  *   VLAN offload bit mask.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 int
 mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask)
-- 
2.11.0

[dpdk-stable] [PATCH v2 28/67] net/mlx5: use port id in PMD log

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 0f99970b4adc943264df0487904d340124765e68 ]

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.c         |  75 +++++++++-------
 drivers/net/mlx5/mlx5_ethdev.c  |  78 ++++++++--------
 drivers/net/mlx5/mlx5_flow.c    |  79 +++++++++-------
 drivers/net/mlx5/mlx5_mac.c     |   9 +-
 drivers/net/mlx5/mlx5_mr.c      |  58 ++++++------
 drivers/net/mlx5/mlx5_rxmode.c  |  16 ++--
 drivers/net/mlx5/mlx5_rxq.c     | 194 ++++++++++++++++++++++------------------
 drivers/net/mlx5/mlx5_rxtx.h    |   7 +-
 drivers/net/mlx5/mlx5_socket.c  |  47 +++++-----
 drivers/net/mlx5/mlx5_stats.c   |  29 +++---
 drivers/net/mlx5/mlx5_trigger.c |  26 +++---
 drivers/net/mlx5/mlx5_txq.c     | 119 +++++++++++++-----------
 drivers/net/mlx5/mlx5_vlan.c    |  21 +++--
 13 files changed, 431 insertions(+), 327 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 9319effcb..8518fa588 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -174,7 +174,6 @@ mlx5_alloc_verbs_buf(size_t size, void *data)
 	ret = rte_malloc_socket(__func__, size, alignment, socket);
 	if (!ret && size)
 		rte_errno = ENOMEM;
-	DEBUG("Extern alloc size: %lu, align: %lu: %p", size, alignment, ret);
 	return ret;
 }
 
@@ -190,7 +189,6 @@ static void
 mlx5_free_verbs_buf(void *ptr, void *data __rte_unused)
 {
 	assert(data != NULL);
-	DEBUG("Extern free request: %p", ptr);
 	rte_free(ptr);
 }
 
@@ -209,8 +207,8 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 	unsigned int i;
 	int ret;
 
-	DEBUG("%p: closing device \"%s\"",
-	      (void *)dev,
+	DEBUG("port %u closing device \"%s\"",
+	      dev->data->port_id,
 	      ((priv->ctx != NULL) ? priv->ctx->device->name : ""));
 	/* In case mlx5_dev_stop() has not been called. */
 	mlx5_dev_interrupt_handler_uninstall(dev);
@@ -248,28 +246,35 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 		mlx5_socket_uninit(dev);
 	ret = mlx5_hrxq_ibv_verify(dev);
 	if (ret)
-		WARN("%p: some Hash Rx queue still remain", (void *)dev);
+		WARN("port %u some hash Rx queue still remain",
+		     dev->data->port_id);
 	ret = mlx5_ind_table_ibv_verify(dev);
 	if (ret)
-		WARN("%p: some Indirection table still remain", (void *)dev);
+		WARN("port %u some indirection table still remain",
+		     dev->data->port_id);
 	ret = mlx5_rxq_ibv_verify(dev);
 	if (ret)
-		WARN("%p: some Verbs Rx queue still remain", (void *)dev);
+		WARN("port %u some Verbs Rx queue still remain",
+		     dev->data->port_id);
 	ret = mlx5_rxq_verify(dev);
 	if (ret)
-		WARN("%p: some Rx Queues still remain", (void *)dev);
+		WARN("port %u some Rx queues still remain",
+		     dev->data->port_id);
 	ret = mlx5_txq_ibv_verify(dev);
 	if (ret)
-		WARN("%p: some Verbs Tx queue still remain", (void *)dev);
+		WARN("port %u some Verbs Tx queue still remain",
+		     dev->data->port_id);
 	ret = mlx5_txq_verify(dev);
 	if (ret)
-		WARN("%p: some Tx Queues still remain", (void *)dev);
+		WARN("port %u some Tx queues still remain",
+		     dev->data->port_id);
 	ret = mlx5_flow_verify(dev);
 	if (ret)
-		WARN("%p: some flows still remain", (void *)dev);
+		WARN("port %u some flows still remain", dev->data->port_id);
 	ret = mlx5_mr_verify(dev);
 	if (ret)
-		WARN("%p: some Memory Region still remain", (void *)dev);
+		WARN("port %u some memory region still remain",
+		     dev->data->port_id);
 	memset(priv, 0, sizeof(*priv));
 }
 
@@ -546,15 +551,17 @@ mlx5_uar_init_primary(struct rte_eth_dev *dev)
 	addr = mmap(addr, MLX5_UAR_SIZE,
 		    PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
 	if (addr == MAP_FAILED) {
-		ERROR("Failed to reserve UAR address space, please adjust "
-		      "MLX5_UAR_SIZE or try --base-virtaddr");
+		ERROR("port %u failed to reserve UAR address space, please"
+		      " adjust MLX5_UAR_SIZE or try --base-virtaddr",
+		      dev->data->port_id);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
 	/* Accept either same addr or a new addr returned from mmap if target
 	 * range occupied.
 	 */
-	INFO("Reserved UAR address space: %p", addr);
+	INFO("port %u reserved UAR address space: %p", dev->data->port_id,
+	     addr);
 	priv->uar_base = addr; /* for primary and secondary UAR re-mmap. */
 	uar_base = addr; /* process local, don't reserve again. */
 	return 0;
@@ -585,20 +592,21 @@ mlx5_uar_init_secondary(struct rte_eth_dev *dev)
 	addr = mmap(priv->uar_base, MLX5_UAR_SIZE,
 		    PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
 	if (addr == MAP_FAILED) {
-		ERROR("UAR mmap failed: %p size: %llu",
-		      priv->uar_base, MLX5_UAR_SIZE);
+		ERROR("port %u UAR mmap failed: %p size: %llu",
+		      dev->data->port_id, priv->uar_base, MLX5_UAR_SIZE);
 		rte_errno = ENXIO;
 		return -rte_errno;
 	}
 	if (priv->uar_base != addr) {
-		ERROR("UAR address %p size %llu occupied, please adjust "
+		ERROR("port %u UAR address %p size %llu occupied, please adjust "
 		      "MLX5_UAR_OFFSET or try EAL parameter --base-virtaddr",
-		      priv->uar_base, MLX5_UAR_SIZE);
+		      dev->data->port_id, priv->uar_base, MLX5_UAR_SIZE);
 		rte_errno = ENXIO;
 		return -rte_errno;
 	}
 	uar_base = addr; /* process local, don't reserve again */
-	INFO("Reserved UAR address space: %p", addr);
+	INFO("port %u reserved UAR address space: %p", dev->data->port_id,
+	     addr);
 	return 0;
 }
 
@@ -745,7 +753,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	mlx5dv_query_device(attr_ctx, &attrs_out);
 	if (attrs_out.flags & MLX5DV_CONTEXT_FLAGS_MPW_ALLOWED) {
 		if (attrs_out.flags & MLX5DV_CONTEXT_FLAGS_ENHANCED_MPW) {
-			DEBUG("Enhanced MPW is supported");
+			DEBUG("enhanced MPW is supported");
 			mps = MLX5_MPW_ENHANCED;
 		} else {
 			DEBUG("MPW is supported");
@@ -910,7 +918,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		if (priv->ind_table_max_size >
 				(unsigned int)ETH_RSS_RETA_SIZE_512)
 			priv->ind_table_max_size = ETH_RSS_RETA_SIZE_512;
-		DEBUG("maximum RX indirection table size is %u",
+		DEBUG("maximum Rx indirection table size is %u",
 		      priv->ind_table_max_size);
 		priv->hw_vlan_strip = !!(device_attr_ex.raw_packet_caps &
 					 IBV_RAW_PACKET_CAP_CVLAN_STRIPPING);
@@ -925,7 +933,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 #ifdef HAVE_IBV_WQ_FLAG_RX_END_PADDING
 		priv->hw_padding = !!device_attr_ex.rx_pad_end_addr_align;
 #endif
-		DEBUG("hardware RX end alignment padding is %ssupported",
+		DEBUG("hardware Rx end alignment padding is %ssupported",
 		      (priv->hw_padding ? "" : "not "));
 		priv->tso = ((priv->tso) &&
 			    (device_attr_ex.tso_caps.max_tso > 0) &&
@@ -944,8 +952,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			      "with TSO. MPS disabled");
 			priv->mps = 0;
 		}
-		INFO("%sMPS is %s",
-		     priv->mps == MLX5_MPW_ENHANCED ? "Enhanced " : "",
+		INFO("%s MPS is %s",
+		     priv->mps == MLX5_MPW_ENHANCED ? "enhanced " : "",
 		     priv->mps != MLX5_MPW_DISABLED ? "enabled" : "disabled");
 		/* Set default values for Enhanced MPW, a.k.a MPWv2. */
 		if (priv->mps == MLX5_MPW_ENHANCED) {
@@ -979,13 +987,14 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			goto port_error;
 		/* Configure the first MAC address by default. */
 		if (mlx5_get_mac(eth_dev, &mac.addr_bytes)) {
-			ERROR("cannot get MAC address, is mlx5_en loaded?"
-			      " (errno: %s)", strerror(errno));
+			ERROR("port %u cannot get MAC address, is mlx5_en"
+			      " loaded? (errno: %s)", eth_dev->data->port_id,
+			      strerror(errno));
 			err = ENODEV;
 			goto port_error;
 		}
 		INFO("port %u MAC address is %02x:%02x:%02x:%02x:%02x:%02x",
-		     priv->port,
+		     eth_dev->data->port_id,
 		     mac.addr_bytes[0], mac.addr_bytes[1],
 		     mac.addr_bytes[2], mac.addr_bytes[3],
 		     mac.addr_bytes[4], mac.addr_bytes[5]);
@@ -995,16 +1004,17 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 
 			if (mlx5_get_ifname(eth_dev, &ifname) == 0)
 				DEBUG("port %u ifname is \"%s\"",
-				      priv->port, ifname);
+				      eth_dev->data->port_id, ifname);
 			else
-				DEBUG("port %u ifname is unknown", priv->port);
+				DEBUG("port %u ifname is unknown",
+				      eth_dev->data->port_id);
 		}
 #endif
 		/* Get actual MTU if possible. */
 		err = mlx5_get_mtu(eth_dev, &priv->mtu);
 		if (err)
 			goto port_error;
-		DEBUG("port %u MTU is %u", priv->port, priv->mtu);
+		DEBUG("port %u MTU is %u", eth_dev->data->port_id, priv->mtu);
 		/*
 		 * Initialize burst functions to prevent crashes before link-up.
 		 */
@@ -1024,7 +1034,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		mlx5dv_set_context_attr(ctx, MLX5DV_CTX_ATTR_BUF_ALLOCATORS,
 					(void *)((uintptr_t)&alctr));
 		/* Bring Ethernet device up. */
-		DEBUG("forcing Ethernet interface up");
+		DEBUG("port %u forcing Ethernet interface up",
+		      eth_dev->data->port_id);
 		mlx5_set_flags(eth_dev, ~IFF_UP, IFF_UP);
 		continue;
 port_error:
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index d0be35570..8696c2d45 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -344,7 +344,8 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 		rte_realloc(priv->rss_conf.rss_key,
 			    rss_hash_default_key_len, 0);
 	if (!priv->rss_conf.rss_key) {
-		ERROR("cannot allocate RSS hash key memory (%u)", rxqs_n);
+		ERROR("port %u cannot allocate RSS hash key memory (%u)",
+		      dev->data->port_id, rxqs_n);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
@@ -358,19 +359,20 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 	priv->rxqs = (void *)dev->data->rx_queues;
 	priv->txqs = (void *)dev->data->tx_queues;
 	if (txqs_n != priv->txqs_n) {
-		INFO("%p: TX queues number update: %u -> %u",
-		     (void *)dev, priv->txqs_n, txqs_n);
+		INFO("port %u Tx queues number update: %u -> %u",
+		     dev->data->port_id, priv->txqs_n, txqs_n);
 		priv->txqs_n = txqs_n;
 	}
 	if (rxqs_n > priv->ind_table_max_size) {
-		ERROR("cannot handle this many RX queues (%u)", rxqs_n);
+		ERROR("port %u cannot handle this many Rx queues (%u)",
+		      dev->data->port_id, rxqs_n);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
 	if (rxqs_n == priv->rxqs_n)
 		return 0;
-	INFO("%p: RX queues number update: %u -> %u",
-	     (void *)dev, priv->rxqs_n, rxqs_n);
+	INFO("port %u Rx queues number update: %u -> %u",
+	     dev->data->port_id, priv->rxqs_n, rxqs_n);
 	priv->rxqs_n = rxqs_n;
 	/* If the requested number of RX queues is not a power of two, use the
 	 * maximum indirection table size for better balancing.
@@ -513,7 +515,8 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 
 	ret = mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr);
 	if (ret) {
-		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(rte_errno));
+		WARN("port %u ioctl(SIOCGIFFLAGS) failed: %s",
+		     dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	memset(&dev_link, 0, sizeof(dev_link));
@@ -522,8 +525,8 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 	ifr.ifr_data = (void *)&edata;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: %s",
-		     strerror(rte_errno));
+		WARN("port %u ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: %s",
+		     dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	link_speed = ethtool_cmd_speed(&edata);
@@ -579,7 +582,8 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 
 	ret = mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr);
 	if (ret) {
-		WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(rte_errno));
+		WARN("port %u ioctl(SIOCGIFFLAGS) failed: %s",
+		     dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	memset(&dev_link, 0, sizeof(dev_link));
@@ -588,8 +592,8 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 	ifr.ifr_data = (void *)&gcmd;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		DEBUG("ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS) failed: %s",
-		      strerror(rte_errno));
+		DEBUG("port %u ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS)"
+		      " failed: %s", dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	gcmd.link_mode_masks_nwords = -gcmd.link_mode_masks_nwords;
@@ -603,8 +607,8 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 	ifr.ifr_data = (void *)ecmd;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		DEBUG("ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS) failed: %s",
-		      strerror(rte_errno));
+		DEBUG("port %u ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS)"
+		      " failed: %s", dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	dev_link.link_speed = ecmd->speed;
@@ -675,15 +679,14 @@ mlx5_link_start(struct rte_eth_dev *dev)
 	mlx5_select_rx_function(dev);
 	ret = mlx5_traffic_enable(dev);
 	if (ret) {
-		ERROR("%p: error occurred while configuring control flows: %s",
-		      (void *)dev, strerror(rte_errno));
+		ERROR("port %u error occurred while configuring control flows:"
+		      " %s", dev->data->port_id, strerror(rte_errno));
 		return;
 	}
 	ret = mlx5_flow_start(dev, &priv->flows);
-	if (ret) {
-		ERROR("%p: error occurred while configuring flows: %s",
-		      (void *)dev, strerror(rte_errno));
-	}
+	if (ret)
+		ERROR("port %u error occurred while configuring flows: %s",
+		       dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -804,7 +807,7 @@ mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 		return ret;
 	if (kern_mtu == mtu) {
 		priv->mtu = mtu;
-		DEBUG("adapter port %u MTU set to %u", priv->port, mtu);
+		DEBUG("port %u adapter MTU set to %u", dev->data->port_id, mtu);
 		return 0;
 	}
 	rte_errno = EAGAIN;
@@ -834,8 +837,8 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 	ifr.ifr_data = (void *)&ethpause;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM) failed: %s",
-		     strerror(rte_errno));
+		WARN("port %u ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM) failed:"
+		     " %s", dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	fc_conf->autoneg = ethpause.autoneg;
@@ -885,9 +888,8 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		ethpause.tx_pause = 0;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("ioctl(SIOCETHTOOL, ETHTOOL_SPAUSEPARAM)"
-		     " failed: %s",
-		     strerror(rte_errno));
+		WARN("port %u ioctl(SIOCETHTOOL, ETHTOOL_SPAUSEPARAM)"
+		     " failed: %s", dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	return 0;
@@ -1016,8 +1018,8 @@ mlx5_dev_status_handler(struct rte_eth_dev *dev)
 			dev->data->dev_conf.intr_conf.rmv == 1)
 			ret |= (1 << RTE_ETH_EVENT_INTR_RMV);
 		else
-			DEBUG("event type %d on port %d not handled",
-			      event.event_type, event.element.port_num);
+			DEBUG("port %u event type %d on not handled",
+			      dev->data->port_id, event.event_type);
 		ibv_ack_async_event(&event);
 	}
 	if (ret & (1 << RTE_ETH_EVENT_INTR_LSC))
@@ -1128,7 +1130,8 @@ mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
 	flags = fcntl(priv->ctx->async_fd, F_GETFL);
 	ret = fcntl(priv->ctx->async_fd, F_SETFL, flags | O_NONBLOCK);
 	if (ret) {
-		INFO("failed to change file descriptor async event queue");
+		INFO("port %u failed to change file descriptor async event"
+		     " queue", dev->data->port_id);
 		dev->data->dev_conf.intr_conf.lsc = 0;
 		dev->data->dev_conf.intr_conf.rmv = 0;
 	}
@@ -1141,7 +1144,8 @@ mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
 	}
 	ret = mlx5_socket_init(dev);
 	if (ret)
-		ERROR("cannot initialise socket: %s", strerror(rte_errno));
+		ERROR("port %u cannot initialise socket: %s",
+		      dev->data->port_id, strerror(rte_errno));
 	else if (priv->primary_socket) {
 		priv->intr_handle_socket.fd = priv->primary_socket;
 		priv->intr_handle_socket.type = RTE_INTR_HANDLE_EXT;
@@ -1199,17 +1203,20 @@ mlx5_select_tx_function(struct rte_eth_dev *dev)
 				dev->tx_pkt_burst = mlx5_tx_burst_raw_vec;
 			else
 				dev->tx_pkt_burst = mlx5_tx_burst_vec;
-			DEBUG("selected Enhanced MPW TX vectorized function");
+			DEBUG("port %u selected enhanced MPW Tx vectorized"
+			      " function", dev->data->port_id);
 		} else {
 			dev->tx_pkt_burst = mlx5_tx_burst_empw;
-			DEBUG("selected Enhanced MPW TX function");
+			DEBUG("port %u selected enhanced MPW Tx function",
+			      dev->data->port_id);
 		}
 	} else if (priv->mps && priv->txq_inline) {
 		dev->tx_pkt_burst = mlx5_tx_burst_mpw_inline;
-		DEBUG("selected MPW inline TX function");
+		DEBUG("port %u selected MPW inline Tx function",
+		      dev->data->port_id);
 	} else if (priv->mps) {
 		dev->tx_pkt_burst = mlx5_tx_burst_mpw;
-		DEBUG("selected MPW TX function");
+		DEBUG("port %u selected MPW Tx function", dev->data->port_id);
 	}
 }
 
@@ -1225,7 +1232,8 @@ mlx5_select_rx_function(struct rte_eth_dev *dev)
 	assert(dev != NULL);
 	if (mlx5_check_vec_rx_support(dev) > 0) {
 		dev->rx_pkt_burst = mlx5_rx_burst_vec;
-		DEBUG("selected RX vectorized function");
+		DEBUG("port %u selected Rx vectorized function",
+		      dev->data->port_id);
 	} else {
 		dev->rx_pkt_burst = mlx5_rx_burst;
 	}
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 09a798924..326392798 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -922,7 +922,7 @@ mlx5_flow_convert_allocate(unsigned int priority,
 		rte_flow_error_set(error, ENOMEM,
 				   RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
 				   NULL,
-				   "cannot allocate verbs spec attributes.");
+				   "cannot allocate verbs spec attributes");
 		return NULL;
 	}
 	ibv_attr->priority = priority;
@@ -1188,11 +1188,11 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 		}
 	}
 	rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
-			   NULL, "cannot allocate verbs spec attributes.");
+			   NULL, "cannot allocate verbs spec attributes");
 	return -rte_errno;
 exit_count_error:
 	rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
-			   NULL, "cannot create counter.");
+			   NULL, "cannot create counter");
 	return -rte_errno;
 }
 
@@ -1838,7 +1838,8 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 			goto error;
 		}
 		++flows_n;
-		DEBUG("%p type %d QP %p ibv_flow %p",
+		DEBUG("port %u %p type %d QP %p ibv_flow %p",
+		      dev->data->port_id,
 		      (void *)flow, i,
 		      (void *)flow->frxq[i].hrxq,
 		      (void *)flow->frxq[i].ibv_flow);
@@ -1941,7 +1942,7 @@ mlx5_flow_list_create(struct rte_eth_dev *dev,
 	if (ret)
 		goto exit;
 	TAILQ_INSERT_TAIL(list, flow, next);
-	DEBUG("Flow created %p", (void *)flow);
+	DEBUG("port %u flow created %p", dev->data->port_id, (void *)flow);
 	return flow;
 exit:
 	for (i = 0; i != hash_rxq_init_n; ++i) {
@@ -2060,7 +2061,7 @@ mlx5_flow_list_destroy(struct rte_eth_dev *dev, struct mlx5_flows *list,
 		flow->cs = NULL;
 	}
 	TAILQ_REMOVE(list, flow, next);
-	DEBUG("Flow destroyed %p", (void *)flow);
+	DEBUG("port %u flow destroyed %p", dev->data->port_id, (void *)flow);
 	rte_free(flow);
 }
 
@@ -2102,13 +2103,15 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 	assert(priv->ctx);
 	fdq = rte_calloc(__func__, 1, sizeof(*fdq), 0);
 	if (!fdq) {
-		WARN("cannot allocate memory for drop queue");
+		WARN("port %u cannot allocate memory for drop queue",
+		     dev->data->port_id);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
 	fdq->cq = ibv_create_cq(priv->ctx, 1, NULL, NULL, 0);
 	if (!fdq->cq) {
-		WARN("cannot allocate CQ for drop queue");
+		WARN("port %u cannot allocate CQ for drop queue",
+		     dev->data->port_id);
 		rte_errno = errno;
 		goto error;
 	}
@@ -2121,7 +2124,8 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 			.cq = fdq->cq,
 			});
 	if (!fdq->wq) {
-		WARN("cannot allocate WQ for drop queue");
+		WARN("port %u cannot allocate WQ for drop queue",
+		     dev->data->port_id);
 		rte_errno = errno;
 		goto error;
 	}
@@ -2132,7 +2136,8 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 			.comp_mask = 0,
 			});
 	if (!fdq->ind_table) {
-		WARN("cannot allocate indirection table for drop queue");
+		WARN("port %u cannot allocate indirection table for drop"
+		     " queue", dev->data->port_id);
 		rte_errno = errno;
 		goto error;
 	}
@@ -2154,7 +2159,8 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 			.pd = priv->pd
 		});
 	if (!fdq->qp) {
-		WARN("cannot allocate QP for drop queue");
+		WARN("port %u cannot allocate QP for drop queue",
+		     dev->data->port_id);
 		rte_errno = errno;
 		goto error;
 	}
@@ -2225,7 +2231,8 @@ mlx5_flow_stop(struct rte_eth_dev *dev, struct mlx5_flows *list)
 			claim_zero(ibv_destroy_flow
 				   (flow->frxq[HASH_RXQ_ETH].ibv_flow));
 			flow->frxq[HASH_RXQ_ETH].ibv_flow = NULL;
-			DEBUG("Flow %p removed", (void *)flow);
+			DEBUG("port %u flow %p removed", dev->data->port_id,
+			      (void *)flow);
 			/* Next flow. */
 			continue;
 		}
@@ -2257,7 +2264,8 @@ mlx5_flow_stop(struct rte_eth_dev *dev, struct mlx5_flows *list)
 			mlx5_hrxq_release(dev, flow->frxq[i].hrxq);
 			flow->frxq[i].hrxq = NULL;
 		}
-		DEBUG("Flow %p removed", (void *)flow);
+		DEBUG("port %u flow %p removed", dev->data->port_id,
+		      (void *)flow);
 	}
 }
 
@@ -2287,12 +2295,14 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 				(priv->flow_drop_queue->qp,
 				 flow->frxq[HASH_RXQ_ETH].ibv_attr);
 			if (!flow->frxq[HASH_RXQ_ETH].ibv_flow) {
-				DEBUG("Flow %p cannot be applied",
+				DEBUG("port %u flow %p cannot be applied",
+				      dev->data->port_id,
 				      (void *)flow);
 				rte_errno = EINVAL;
 				return -rte_errno;
 			}
-			DEBUG("Flow %p applied", (void *)flow);
+			DEBUG("port %u flow %p applied", dev->data->port_id,
+			      (void *)flow);
 			/* Next flow. */
 			continue;
 		}
@@ -2314,8 +2324,8 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 					      (*flow->queues),
 					      flow->queues_n);
 			if (!flow->frxq[i].hrxq) {
-				DEBUG("Flow %p cannot be applied",
-				      (void *)flow);
+				DEBUG("port %u flow %p cannot be applied",
+				      dev->data->port_id, (void *)flow);
 				rte_errno = EINVAL;
 				return -rte_errno;
 			}
@@ -2324,12 +2334,13 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 				ibv_create_flow(flow->frxq[i].hrxq->qp,
 						flow->frxq[i].ibv_attr);
 			if (!flow->frxq[i].ibv_flow) {
-				DEBUG("Flow %p cannot be applied",
-				      (void *)flow);
+				DEBUG("port %u flow %p cannot be applied",
+				      dev->data->port_id, (void *)flow);
 				rte_errno = EINVAL;
 				return -rte_errno;
 			}
-			DEBUG("Flow %p applied", (void *)flow);
+			DEBUG("port %u flow %p applied",
+			      dev->data->port_id, (void *)flow);
 		}
 		if (!flow->mark)
 			continue;
@@ -2355,8 +2366,8 @@ mlx5_flow_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	TAILQ_FOREACH(flow, &priv->flows, next) {
-		DEBUG("%p: flow %p still referenced", (void *)dev,
-		      (void *)flow);
+		DEBUG("port %u flow %p still referenced",
+		      dev->data->port_id, (void *)flow);
 		++ret;
 	}
 	return ret;
@@ -2627,7 +2638,8 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 
 	/* Validate queue number. */
 	if (fdir_filter->action.rx_queue >= priv->rxqs_n) {
-		ERROR("invalid queue number %d", fdir_filter->action.rx_queue);
+		ERROR("port %u invalid queue number %d",
+		      dev->data->port_id, fdir_filter->action.rx_queue);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
@@ -2650,7 +2662,9 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 		};
 		break;
 	default:
-		ERROR("invalid behavior %d", fdir_filter->action.behavior);
+		ERROR("port %u invalid behavior %d",
+		      dev->data->port_id,
+		      fdir_filter->action.behavior);
 		rte_errno = ENOTSUP;
 		return -rte_errno;
 	}
@@ -2786,7 +2800,8 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 		};
 		break;
 	default:
-		ERROR("invalid flow type%d", fdir_filter->input.flow_type);
+		ERROR("port %u invalid flow type%d",
+		      dev->data->port_id, fdir_filter->input.flow_type);
 		rte_errno = ENOTSUP;
 		return -rte_errno;
 	}
@@ -2835,7 +2850,8 @@ mlx5_fdir_filter_add(struct rte_eth_dev *dev,
 				     attributes.items, attributes.actions,
 				     &error);
 	if (flow) {
-		DEBUG("FDIR created %p", (void *)flow);
+		DEBUG("port %u FDIR created %p", dev->data->port_id,
+		      (void *)flow);
 		return 0;
 	}
 	return -rte_errno;
@@ -3029,8 +3045,8 @@ mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
 		return 0;
 	if (fdir_mode != RTE_FDIR_MODE_PERFECT &&
 	    fdir_mode != RTE_FDIR_MODE_PERFECT_MAC_VLAN) {
-		ERROR("%p: flow director mode %d not supported",
-		      (void *)dev, fdir_mode);
+		ERROR("port %u flow director mode %d not supported",
+		      dev->data->port_id, fdir_mode);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
@@ -3048,7 +3064,8 @@ mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
 		mlx5_fdir_info_get(dev, arg);
 		break;
 	default:
-		DEBUG("%p: unknown operation %u", (void *)dev, filter_op);
+		DEBUG("port %u unknown operation %u", dev->data->port_id,
+		      filter_op);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
@@ -3087,8 +3104,8 @@ mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
 	case RTE_ETH_FILTER_FDIR:
 		return mlx5_fdir_ctrl_func(dev, filter_op, arg);
 	default:
-		ERROR("%p: filter type (%d) not supported",
-		      (void *)dev, filter_type);
+		ERROR("port %u filter type (%d) not supported",
+		      dev->data->port_id, filter_type);
 		rte_errno = ENOTSUP;
 		return -rte_errno;
 	}
diff --git a/drivers/net/mlx5/mlx5_mac.c b/drivers/net/mlx5/mlx5_mac.c
index e9d9c67e9..69fc06897 100644
--- a/drivers/net/mlx5/mlx5_mac.c
+++ b/drivers/net/mlx5/mlx5_mac.c
@@ -101,8 +101,8 @@ mlx5_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index)
 		int ret = mlx5_traffic_restart(dev);
 
 		if (ret)
-			ERROR("%p cannot remove mac address: %s", (void *)dev,
-			      strerror(rte_errno));
+			ERROR("port %u cannot remove mac address: %s",
+			      dev->data->port_id, strerror(rte_errno));
 	}
 }
 
@@ -158,8 +158,9 @@ mlx5_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr)
 {
 	int ret;
 
-	DEBUG("%p: setting primary MAC address", (void *)dev);
+	DEBUG("port %u setting primary MAC address", dev->data->port_id);
 	ret = mlx5_mac_addr_add(dev, mac_addr, 0, 0);
 	if (ret)
-		ERROR("cannot set mac address: %s", strerror(rte_errno));
+		ERROR("port %u cannot set mac address: %s",
+		      dev->data->port_id, strerror(rte_errno));
 }
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 3a4e46f9a..42109a6a4 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -133,15 +133,16 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 
 	rte_spinlock_lock(&txq_ctrl->priv->mr_lock);
 	/* Add a new entry, register MR first. */
-	DEBUG("%p: discovered new memory pool \"%s\" (%p)",
-	      (void *)txq_ctrl, mp->name, (void *)mp);
+	DEBUG("port %u discovered new memory pool \"%s\" (%p)",
+	      txq_ctrl->priv->dev->data->port_id, mp->name, (void *)mp);
 	dev = txq_ctrl->priv->dev;
 	mr = mlx5_mr_get(dev, mp);
 	if (mr == NULL) {
 		if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
-			DEBUG("Using unregistered mempool 0x%p(%s) in "
+			DEBUG("port %u using unregistered mempool 0x%p(%s) in "
 			      "secondary process, please create mempool before "
 			      " rte_eth_dev_start()",
+			      txq_ctrl->priv->dev->data->port_id,
 			     (void *)mp, mp->name);
 			rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 			rte_errno = ENOTSUP;
@@ -150,15 +151,17 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 		mr = mlx5_mr_new(dev, mp);
 	}
 	if (unlikely(mr == NULL)) {
-		DEBUG("%p: unable to configure MR, ibv_reg_mr() failed.",
-		      (void *)txq_ctrl);
 		rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
+		DEBUG("port %u unable to configure memory region, ibv_reg_mr()"
+		      " failed",
+		      txq_ctrl->priv->dev->data->port_id);
 		return NULL;
 	}
 	if (unlikely(idx == RTE_DIM(txq->mp2mr))) {
 		/* Table is full, remove oldest entry. */
-		DEBUG("%p: MR <-> MP table full, dropping oldest entry.",
-		      (void *)txq_ctrl);
+		DEBUG("port %u memroy region <-> memory pool table full, "
+		      " dropping oldest entry",
+		      txq_ctrl->priv->dev->data->port_id);
 		--idx;
 		mlx5_mr_release(txq->mp2mr[0]);
 		memmove(&txq->mp2mr[0], &txq->mp2mr[1],
@@ -166,8 +169,8 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 	}
 	/* Store the new entry. */
 	txq_ctrl->txq.mp2mr[idx] = mr;
-	DEBUG("%p: new MR lkey for MP \"%s\" (%p): 0x%08" PRIu32,
-	      (void *)txq_ctrl, mp->name, (void *)mp,
+	DEBUG("port %u new memory region lkey for MP \"%s\" (%p): 0x%08" PRIu32,
+	      txq_ctrl->priv->dev->data->port_id, mp->name, (void *)mp,
 	      txq_ctrl->txq.mp2mr[idx]->lkey);
 	rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 	return mr;
@@ -235,7 +238,8 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
 	}
 	mr = mlx5_mr_new(priv->dev, mp);
 	if (!mr)
-		ERROR("cannot create memory region: %s", strerror(rte_errno));
+		ERROR("port %u cannot create memory region: %s",
+		      priv->dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -262,18 +266,20 @@ mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 
 	mr = rte_zmalloc_socket(__func__, sizeof(*mr), 0, mp->socket_id);
 	if (!mr) {
-		DEBUG("unable to configure MR, ibv_reg_mr() failed.");
+		DEBUG("port %u unable to configure memory region, ibv_reg_mr()"
+		      " failed",
+		      dev->data->port_id);
 		rte_errno = ENOMEM;
 		return NULL;
 	}
 	if (mlx5_check_mempool(mp, &start, &end) != 0) {
-		ERROR("mempool %p: not virtually contiguous",
-		      (void *)mp);
+		ERROR("port %u mempool %p: not virtually contiguous",
+		      dev->data->port_id, (void *)mp);
 		rte_errno = ENOMEM;
 		return NULL;
 	}
-	DEBUG("mempool %p area start=%p end=%p size=%zu",
-	      (void *)mp, (void *)start, (void *)end,
+	DEBUG("port %u mempool %p area start=%p end=%p size=%zu",
+	      dev->data->port_id, (void *)mp, (void *)start, (void *)end,
 	      (size_t)(end - start));
 	/* Save original addresses for exact MR lookup. */
 	mr->start = start;
@@ -289,8 +295,9 @@ mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 		if ((end > addr) && (end < addr + len))
 			end = RTE_ALIGN_CEIL(end, align);
 	}
-	DEBUG("mempool %p using start=%p end=%p size=%zu for MR",
-	      (void *)mp, (void *)start, (void *)end,
+	DEBUG("port %u mempool %p using start=%p end=%p size=%zu for memory"
+	      " region",
+	      dev->data->port_id, (void *)mp, (void *)start, (void *)end,
 	      (size_t)(end - start));
 	mr->mr = ibv_reg_mr(priv->pd, (void *)start, end - start,
 			    IBV_ACCESS_LOCAL_WRITE);
@@ -301,8 +308,8 @@ mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 	mr->mp = mp;
 	mr->lkey = rte_cpu_to_be_32(mr->mr->lkey);
 	rte_atomic32_inc(&mr->refcnt);
-	DEBUG("%p: new Memory Region %p refcnt: %d", (void *)dev,
-	      (void *)mr, rte_atomic32_read(&mr->refcnt));
+	DEBUG("port %u new memory region %p refcnt: %d",
+	      dev->data->port_id, (void *)mr, rte_atomic32_read(&mr->refcnt));
 	LIST_INSERT_HEAD(&priv->mr, mr, next);
 	return mr;
 }
@@ -330,8 +337,9 @@ mlx5_mr_get(struct rte_eth_dev *dev, struct rte_mempool *mp)
 	LIST_FOREACH(mr, &priv->mr, next) {
 		if (mr->mp == mp) {
 			rte_atomic32_inc(&mr->refcnt);
-			DEBUG("Memory Region %p refcnt: %d",
-			      (void *)mr, rte_atomic32_read(&mr->refcnt));
+			DEBUG("port %u memory region %p refcnt: %d",
+			      dev->data->port_id, (void *)mr,
+			      rte_atomic32_read(&mr->refcnt));
 			return mr;
 		}
 	}
@@ -351,8 +359,8 @@ int
 mlx5_mr_release(struct mlx5_mr *mr)
 {
 	assert(mr);
-	DEBUG("Memory Region %p refcnt: %d",
-	      (void *)mr, rte_atomic32_read(&mr->refcnt));
+	DEBUG("memory region %p refcnt: %d", (void *)mr,
+	      rte_atomic32_read(&mr->refcnt));
 	if (rte_atomic32_dec_and_test(&mr->refcnt)) {
 		claim_zero(ibv_dereg_mr(mr->mr));
 		LIST_REMOVE(mr, next);
@@ -379,8 +387,8 @@ mlx5_mr_verify(struct rte_eth_dev *dev)
 	struct mlx5_mr *mr;
 
 	LIST_FOREACH(mr, &priv->mr, next) {
-		DEBUG("%p: mr %p still referenced", (void *)dev,
-		      (void *)mr);
+		DEBUG("port %u memory region %p still referenced",
+		      dev->data->port_id, (void *)mr);
 		++ret;
 	}
 	return ret;
diff --git a/drivers/net/mlx5/mlx5_rxmode.c b/drivers/net/mlx5/mlx5_rxmode.c
index 09808a39a..f92ce8ef8 100644
--- a/drivers/net/mlx5/mlx5_rxmode.c
+++ b/drivers/net/mlx5/mlx5_rxmode.c
@@ -65,8 +65,8 @@ mlx5_promiscuous_enable(struct rte_eth_dev *dev)
 	dev->data->promiscuous = 1;
 	ret = mlx5_traffic_restart(dev);
 	if (ret)
-		ERROR("%p cannot enable promiscuous mode: %s", (void *)dev,
-		      strerror(rte_errno));
+		ERROR("port %u cannot enable promiscuous mode: %s",
+		      dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -83,8 +83,8 @@ mlx5_promiscuous_disable(struct rte_eth_dev *dev)
 	dev->data->promiscuous = 0;
 	ret = mlx5_traffic_restart(dev);
 	if (ret)
-		ERROR("%p cannot disable promiscuous mode: %s", (void *)dev,
-		      strerror(rte_errno));
+		ERROR("port %u cannot disable promiscuous mode: %s",
+		      dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -101,8 +101,8 @@ mlx5_allmulticast_enable(struct rte_eth_dev *dev)
 	dev->data->all_multicast = 1;
 	ret = mlx5_traffic_restart(dev);
 	if (ret)
-		ERROR("%p cannot enable allmulicast mode: %s", (void *)dev,
-		      strerror(rte_errno));
+		ERROR("port %u cannot enable allmulicast mode: %s",
+		      dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -119,6 +119,6 @@ mlx5_allmulticast_disable(struct rte_eth_dev *dev)
 	dev->data->all_multicast = 0;
 	ret = mlx5_traffic_restart(dev);
 	if (ret)
-		ERROR("%p cannot disable allmulicast mode: %s", (void *)dev,
-		      strerror(rte_errno));
+		ERROR("port %u cannot disable allmulicast mode: %s",
+		      dev->data->port_id, strerror(rte_errno));
 }
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 40a8b72e7..c97844f63 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -104,7 +104,8 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 
 		buf = rte_pktmbuf_alloc(rxq_ctrl->rxq.mp);
 		if (buf == NULL) {
-			ERROR("%p: empty mbuf pool", (void *)rxq_ctrl);
+			ERROR("port %u empty mbuf pool",
+			      rxq_ctrl->priv->dev->data->port_id);
 			rte_errno = ENOMEM;
 			goto error;
 		}
@@ -145,8 +146,9 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 		for (j = 0; j < MLX5_VPMD_DESCS_PER_LOOP; ++j)
 			(*rxq->elts)[elts_n + j] = &rxq->fake_mbuf;
 	}
-	DEBUG("%p: allocated and configured %u segments (max %u packets)",
-	      (void *)rxq_ctrl, elts_n, elts_n / (1 << rxq_ctrl->rxq.sges_n));
+	DEBUG("port %u Rx queue %u allocated and configured %u segments"
+	      " (max %u packets)", rxq_ctrl->priv->dev->data->port_id,
+	      rxq_ctrl->idx, elts_n, elts_n / (1 << rxq_ctrl->rxq.sges_n));
 	return 0;
 error:
 	err = rte_errno; /* Save rte_errno before cleanup. */
@@ -156,7 +158,8 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 			rte_pktmbuf_free_seg((*rxq_ctrl->rxq.elts)[i]);
 		(*rxq_ctrl->rxq.elts)[i] = NULL;
 	}
-	DEBUG("%p: failed, freed everything", (void *)rxq_ctrl);
+	DEBUG("port %u Rx queue %u failed, freed everything",
+	      rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
 	rte_errno = err; /* Restore rte_errno. */
 	return -rte_errno;
 }
@@ -176,7 +179,8 @@ rxq_free_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 	uint16_t used = q_n - (rxq->rq_ci - rxq->rq_pi);
 	uint16_t i;
 
-	DEBUG("%p: freeing WRs", (void *)rxq_ctrl);
+	DEBUG("port %u Rx queue %u freeing WRs",
+	      rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
 	if (rxq->elts == NULL)
 		return;
 	/**
@@ -206,7 +210,8 @@ rxq_free_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 void
 mlx5_rxq_cleanup(struct mlx5_rxq_ctrl *rxq_ctrl)
 {
-	DEBUG("cleaning up %p", (void *)rxq_ctrl);
+	DEBUG("port %u cleaning up Rx queue %u",
+	      rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
 	if (rxq_ctrl->ibv)
 		mlx5_rxq_ibv_release(rxq_ctrl->ibv);
 	memset(rxq_ctrl, 0, sizeof(*rxq_ctrl));
@@ -243,34 +248,33 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 
 	if (!rte_is_power_of_2(desc)) {
 		desc = 1 << log2above(desc);
-		WARN("%p: increased number of descriptors in RX queue %u"
+		WARN("port %u increased number of descriptors in Rx queue %u"
 		     " to the next power of two (%d)",
-		     (void *)dev, idx, desc);
+		     dev->data->port_id, idx, desc);
 	}
-	DEBUG("%p: configuring queue %u for %u descriptors",
-	      (void *)dev, idx, desc);
+	DEBUG("port %u configuring Rx queue %u for %u descriptors",
+	      dev->data->port_id, idx, desc);
 	if (idx >= priv->rxqs_n) {
-		ERROR("%p: queue index out of range (%u >= %u)",
-		      (void *)dev, idx, priv->rxqs_n);
+		ERROR("port %u Rx queue index out of range (%u >= %u)",
+		      dev->data->port_id, idx, priv->rxqs_n);
 		rte_errno = EOVERFLOW;
 		return -rte_errno;
 	}
 	if (!mlx5_rxq_releasable(dev, idx)) {
-		ERROR("%p: unable to release queue index %u",
-		      (void *)dev, idx);
+		ERROR("port %u unable to release queue index %u",
+		      dev->data->port_id, idx);
 		rte_errno = EBUSY;
 		return -rte_errno;
 	}
 	mlx5_rxq_release(dev, idx);
 	rxq_ctrl = mlx5_rxq_new(dev, idx, desc, socket, mp);
 	if (!rxq_ctrl) {
-		ERROR("%p: unable to allocate queue index %u",
-		      (void *)dev, idx);
+		ERROR("port %u unable to allocate queue index %u",
+		      dev->data->port_id, idx);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
-	DEBUG("%p: adding RX queue %p to list",
-	      (void *)dev, (void *)rxq_ctrl);
+	DEBUG("port %u adding Rx queue %u to list", dev->data->port_id, idx);
 	(*priv->rxqs)[idx] = &rxq_ctrl->rxq;
 	return 0;
 }
@@ -293,8 +297,9 @@ mlx5_rx_queue_release(void *dpdk_rxq)
 	rxq_ctrl = container_of(rxq, struct mlx5_rxq_ctrl, rxq);
 	priv = rxq_ctrl->priv;
 	if (!mlx5_rxq_releasable(priv->dev, rxq_ctrl->rxq.stats.idx))
-		rte_panic("Rx queue %p is still used by a flow and cannot be"
-			  " removed\n", (void *)rxq_ctrl);
+		rte_panic("port %u Rx queue %u is still used by a flow and"
+			  " cannot be removed\n", priv->dev->data->port_id,
+			  rxq_ctrl->idx);
 	mlx5_rxq_release(priv->dev, rxq_ctrl->rxq.stats.idx);
 }
 
@@ -322,8 +327,9 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 	mlx5_rx_intr_vec_disable(dev);
 	intr_handle->intr_vec = malloc(n * sizeof(intr_handle->intr_vec[0]));
 	if (intr_handle->intr_vec == NULL) {
-		ERROR("failed to allocate memory for interrupt vector,"
-		      " Rx interrupts will not be supported");
+		ERROR("port %u failed to allocate memory for interrupt vector,"
+		      " Rx interrupts will not be supported",
+		      dev->data->port_id);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
@@ -344,9 +350,9 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 			continue;
 		}
 		if (count >= RTE_MAX_RXTX_INTR_VEC_ID) {
-			ERROR("too many Rx queues for interrupt vector size"
-			      " (%d), Rx interrupts cannot be enabled",
-			      RTE_MAX_RXTX_INTR_VEC_ID);
+			ERROR("port %u too many Rx queues for interrupt vector"
+			      " size (%d), Rx interrupts cannot be enabled",
+			      dev->data->port_id, RTE_MAX_RXTX_INTR_VEC_ID);
 			mlx5_rx_intr_vec_disable(dev);
 			rte_errno = ENOMEM;
 			return -rte_errno;
@@ -356,8 +362,9 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 		rc = fcntl(fd, F_SETFL, flags | O_NONBLOCK);
 		if (rc < 0) {
 			rte_errno = errno;
-			ERROR("failed to make Rx interrupt file descriptor"
-			      " %d non-blocking for queue index %d", fd, i);
+			ERROR("port %u failed to make Rx interrupt file"
+			      " descriptor %d non-blocking for queue index %d",
+			      dev->data->port_id, fd, i);
 			mlx5_rx_intr_vec_disable(dev);
 			return -rte_errno;
 		}
@@ -524,7 +531,8 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	ret = rte_errno; /* Save rte_errno before cleanup. */
 	if (rxq_ibv)
 		mlx5_rxq_ibv_release(rxq_ibv);
-	WARN("unable to disable interrupt on rx queue %d", rx_queue_id);
+	WARN("port %u unable to disable interrupt on Rx queue %d",
+	     dev->data->port_id, rx_queue_id);
 	rte_errno = ret; /* Restore rte_errno. */
 	return -rte_errno;
 }
@@ -571,8 +579,8 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	tmpl = rte_calloc_socket(__func__, 1, sizeof(*tmpl), 0,
 				 rxq_ctrl->socket);
 	if (!tmpl) {
-		ERROR("%p: cannot allocate verbs resources",
-		       (void *)rxq_ctrl);
+		ERROR("port %u Rx queue %u cannot allocate verbs resources",
+		      dev->data->port_id, rxq_ctrl->idx);
 		rte_errno = ENOMEM;
 		goto error;
 	}
@@ -582,15 +590,16 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (!tmpl->mr) {
 		tmpl->mr = mlx5_mr_new(dev, rxq_data->mp);
 		if (!tmpl->mr) {
-			ERROR("%p: MR creation failure", (void *)rxq_ctrl);
+			ERROR("port %u: memory region creation failure",
+			      dev->data->port_id);
 			goto error;
 		}
 	}
 	if (rxq_ctrl->irq) {
 		tmpl->channel = ibv_create_comp_channel(priv->ctx);
 		if (!tmpl->channel) {
-			ERROR("%p: Comp Channel creation failure",
-			      (void *)rxq_ctrl);
+			ERROR("port %u: comp channel creation failure",
+			      dev->data->port_id);
 			rte_errno = ENOMEM;
 			goto error;
 		}
@@ -614,19 +623,21 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		if (mlx5_rxq_check_vec_support(rxq_data) < 0)
 			attr.cq.ibv.cqe *= 2;
 	} else if (priv->cqe_comp && rxq_data->hw_timestamp) {
-		DEBUG("Rx CQE compression is disabled for HW timestamp");
+		DEBUG("port %u Rx CQE compression is disabled for HW timestamp",
+		      dev->data->port_id);
 	}
 	tmpl->cq = ibv_cq_ex_to_cq(mlx5dv_create_cq(priv->ctx, &attr.cq.ibv,
 						    &attr.cq.mlx5));
 	if (tmpl->cq == NULL) {
-		ERROR("%p: CQ creation failure", (void *)rxq_ctrl);
+		ERROR("port %u Rx queue %u CQ creation failure",
+		      dev->data->port_id, idx);
 		rte_errno = ENOMEM;
 		goto error;
 	}
-	DEBUG("priv->device_attr.max_qp_wr is %d",
-	      priv->device_attr.orig_attr.max_qp_wr);
-	DEBUG("priv->device_attr.max_sge is %d",
-	      priv->device_attr.orig_attr.max_sge);
+	DEBUG("port %u priv->device_attr.max_qp_wr is %d",
+	      dev->data->port_id, priv->device_attr.orig_attr.max_qp_wr);
+	DEBUG("port %u priv->device_attr.max_sge is %d",
+	      dev->data->port_id, priv->device_attr.orig_attr.max_sge);
 	attr.wq = (struct ibv_wq_init_attr){
 		.wq_context = NULL, /* Could be useful in the future. */
 		.wq_type = IBV_WQT_RQ,
@@ -656,7 +667,8 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 #endif
 	tmpl->wq = ibv_create_wq(priv->ctx, &attr.wq);
 	if (tmpl->wq == NULL) {
-		ERROR("%p: WQ creation failure", (void *)rxq_ctrl);
+		ERROR("port %u Rx queue %u WQ creation failure",
+		      dev->data->port_id, idx);
 		rte_errno = ENOMEM;
 		goto error;
 	}
@@ -667,8 +679,9 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (((int)attr.wq.max_wr !=
 	     ((1 << rxq_data->elts_n) >> rxq_data->sges_n)) ||
 	    ((int)attr.wq.max_sge != (1 << rxq_data->sges_n))) {
-		ERROR("%p: requested %u*%u but got %u*%u WRs*SGEs",
-		      (void *)rxq_ctrl,
+		ERROR("port %u Rx queue %u requested %u*%u but got %u*%u"
+		      " WRs*SGEs",
+		      dev->data->port_id, idx,
 		      ((1 << rxq_data->elts_n) >> rxq_data->sges_n),
 		      (1 << rxq_data->sges_n),
 		      attr.wq.max_wr, attr.wq.max_sge);
@@ -682,8 +695,8 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	};
 	ret = ibv_modify_wq(tmpl->wq, &mod);
 	if (ret) {
-		ERROR("%p: WQ state to IBV_WQS_RDY failed",
-		      (void *)rxq_ctrl);
+		ERROR("port %u Rx queue %u WQ state to IBV_WQS_RDY failed",
+		      dev->data->port_id, idx);
 		rte_errno = ret;
 		goto error;
 	}
@@ -697,8 +710,9 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		goto error;
 	}
 	if (cq_info.cqe_size != RTE_CACHE_LINE_SIZE) {
-		ERROR("Wrong MLX5_CQE_SIZE environment variable value: "
-		      "it should be set to %u", RTE_CACHE_LINE_SIZE);
+		ERROR("port %u wrong MLX5_CQE_SIZE environment variable value: "
+		      "it should be set to %u", dev->data->port_id,
+		      RTE_CACHE_LINE_SIZE);
 		rte_errno = EINVAL;
 		goto error;
 	}
@@ -735,10 +749,11 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	rxq_data->rq_ci = (1 << rxq_data->elts_n) >> rxq_data->sges_n;
 	rte_wmb();
 	*rxq_data->rq_db = rte_cpu_to_be_32(rxq_data->rq_ci);
-	DEBUG("%p: rxq updated with %p", (void *)rxq_ctrl, (void *)&tmpl);
+	DEBUG("port %u rxq %u updated with %p", dev->data->port_id, idx,
+	      (void *)&tmpl);
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("%p: Verbs Rx queue %p: refcnt %d", (void *)dev,
-	      (void *)tmpl, rte_atomic32_read(&tmpl->refcnt));
+	DEBUG("port %u Verbs Rx queue %u: refcnt %d", dev->data->port_id, idx,
+	      rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->rxqsibv, tmpl, next);
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return tmpl;
@@ -783,8 +798,8 @@ mlx5_rxq_ibv_get(struct rte_eth_dev *dev, uint16_t idx)
 	if (rxq_ctrl->ibv) {
 		mlx5_mr_get(dev, rxq_data->mp);
 		rte_atomic32_inc(&rxq_ctrl->ibv->refcnt);
-		DEBUG("%p: Verbs Rx queue %p: refcnt %d", (void *)dev,
-		      (void *)rxq_ctrl->ibv,
+		DEBUG("port %u Verbs Rx queue %u: refcnt %d",
+		      dev->data->port_id, rxq_ctrl->idx,
 		      rte_atomic32_read(&rxq_ctrl->ibv->refcnt));
 	}
 	return rxq_ctrl->ibv;
@@ -811,8 +826,9 @@ mlx5_rxq_ibv_release(struct mlx5_rxq_ibv *rxq_ibv)
 	ret = mlx5_mr_release(rxq_ibv->mr);
 	if (!ret)
 		rxq_ibv->mr = NULL;
-	DEBUG("Verbs Rx queue %p: refcnt %d",
-	      (void *)rxq_ibv, rte_atomic32_read(&rxq_ibv->refcnt));
+	DEBUG("port %u Verbs Rx queue %u: refcnt %d",
+	      rxq_ibv->rxq_ctrl->priv->dev->data->port_id,
+	      rxq_ibv->rxq_ctrl->idx, rte_atomic32_read(&rxq_ibv->refcnt));
 	if (rte_atomic32_dec_and_test(&rxq_ibv->refcnt)) {
 		rxq_free_elts(rxq_ibv->rxq_ctrl);
 		claim_zero(ibv_destroy_wq(rxq_ibv->wq));
@@ -843,8 +859,8 @@ mlx5_rxq_ibv_verify(struct rte_eth_dev *dev)
 	struct mlx5_rxq_ibv *rxq_ibv;
 
 	LIST_FOREACH(rxq_ibv, &priv->rxqsibv, next) {
-		DEBUG("%p: Verbs Rx queue %p still referenced", (void *)dev,
-		      (void *)rxq_ibv);
+		DEBUG("port %u Verbs Rx queue %u still referenced",
+		      dev->data->port_id, rxq_ibv->rxq_ctrl->idx);
 		++ret;
 	}
 	return ret;
@@ -920,28 +936,28 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		size = mb_len * (1 << tmpl->rxq.sges_n);
 		size -= RTE_PKTMBUF_HEADROOM;
 		if (size < dev->data->dev_conf.rxmode.max_rx_pkt_len) {
-			ERROR("%p: too many SGEs (%u) needed to handle"
+			ERROR("port %u too many SGEs (%u) needed to handle"
 			      " requested maximum packet size %u",
-			      (void *)dev,
+			      dev->data->port_id,
 			      1 << sges_n,
 			      dev->data->dev_conf.rxmode.max_rx_pkt_len);
 			rte_errno = EOVERFLOW;
 			goto error;
 		}
 	} else {
-		WARN("%p: the requested maximum Rx packet size (%u) is"
+		WARN("port %u the requested maximum Rx packet size (%u) is"
 		     " larger than a single mbuf (%u) and scattered"
 		     " mode has not been requested",
-		     (void *)dev,
+		     dev->data->port_id,
 		     dev->data->dev_conf.rxmode.max_rx_pkt_len,
 		     mb_len - RTE_PKTMBUF_HEADROOM);
 	}
-	DEBUG("%p: maximum number of segments per packet: %u",
-	      (void *)dev, 1 << tmpl->rxq.sges_n);
+	DEBUG("port %u maximum number of segments per packet: %u",
+	      dev->data->port_id, 1 << tmpl->rxq.sges_n);
 	if (desc % (1 << tmpl->rxq.sges_n)) {
-		ERROR("%p: number of RX queue descriptors (%u) is not a"
+		ERROR("port %u number of Rx queue descriptors (%u) is not a"
 		      " multiple of SGEs per packet (%u)",
-		      (void *)dev,
+		      dev->data->port_id,
 		      desc,
 		      1 << tmpl->rxq.sges_n);
 		rte_errno = EINVAL;
@@ -964,15 +980,15 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	} else if (priv->hw_fcs_strip) {
 		tmpl->rxq.crc_present = 1;
 	} else {
-		WARN("%p: CRC stripping has been disabled but will still"
+		WARN("port %u CRC stripping has been disabled but will still"
 		     " be performed by hardware, make sure MLNX_OFED and"
 		     " firmware are up to date",
-		     (void *)dev);
+		     dev->data->port_id);
 		tmpl->rxq.crc_present = 0;
 	}
-	DEBUG("%p: CRC stripping is %s, %u bytes will be subtracted from"
+	DEBUG("port %u CRC stripping is %s, %u bytes will be subtracted from"
 	      " incoming frames to hide it",
-	      (void *)dev,
+	      dev->data->port_id,
 	      tmpl->rxq.crc_present ? "disabled" : "enabled",
 	      tmpl->rxq.crc_present << 2);
 	/* Save port ID. */
@@ -984,9 +1000,10 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	tmpl->rxq.elts_n = log2above(desc);
 	tmpl->rxq.elts =
 		(struct rte_mbuf *(*)[1 << tmpl->rxq.elts_n])(tmpl + 1);
+	tmpl->idx = idx;
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("%p: Rx queue %p: refcnt %d", (void *)dev,
-	      (void *)tmpl, rte_atomic32_read(&tmpl->refcnt));
+	DEBUG("port %u Rx queue %u: refcnt %d", dev->data->port_id,
+	      idx, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->rxqsctrl, tmpl, next);
 	return tmpl;
 error:
@@ -1017,8 +1034,8 @@ mlx5_rxq_get(struct rte_eth_dev *dev, uint16_t idx)
 					rxq);
 		mlx5_rxq_ibv_get(dev, idx);
 		rte_atomic32_inc(&rxq_ctrl->refcnt);
-		DEBUG("%p: Rx queue %p: refcnt %d", (void *)dev,
-		      (void *)rxq_ctrl, rte_atomic32_read(&rxq_ctrl->refcnt));
+		DEBUG("port %u Rx queue %u: refcnt %d", dev->data->port_id,
+		      rxq_ctrl->idx, rte_atomic32_read(&rxq_ctrl->refcnt));
 	}
 	return rxq_ctrl;
 }
@@ -1046,8 +1063,8 @@ mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx)
 	assert(rxq_ctrl->priv);
 	if (rxq_ctrl->ibv && !mlx5_rxq_ibv_release(rxq_ctrl->ibv))
 		rxq_ctrl->ibv = NULL;
-	DEBUG("%p: Rx queue %p: refcnt %d", (void *)dev,
-	      (void *)rxq_ctrl, rte_atomic32_read(&rxq_ctrl->refcnt));
+	DEBUG("port %u Rx queue %u: refcnt %d", dev->data->port_id,
+	      rxq_ctrl->idx, rte_atomic32_read(&rxq_ctrl->refcnt));
 	if (rte_atomic32_dec_and_test(&rxq_ctrl->refcnt)) {
 		LIST_REMOVE(rxq_ctrl, next);
 		rte_free(rxq_ctrl);
@@ -1100,8 +1117,8 @@ mlx5_rxq_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	LIST_FOREACH(rxq_ctrl, &priv->rxqsctrl, next) {
-		DEBUG("%p: Rx Queue %p still referenced", (void *)dev,
-		      (void *)rxq_ctrl);
+		DEBUG("port %u Rx queue %u still referenced",
+		      dev->data->port_id, rxq_ctrl->idx);
 		++ret;
 	}
 	return ret;
@@ -1164,12 +1181,12 @@ mlx5_ind_table_ibv_new(struct rte_eth_dev *dev, uint16_t queues[],
 	}
 	rte_atomic32_inc(&ind_tbl->refcnt);
 	LIST_INSERT_HEAD(&priv->ind_tbls, ind_tbl, next);
-	DEBUG("%p: Indirection table %p: refcnt %d", (void *)dev,
+	DEBUG("port %u indirection table %p: refcnt %d", dev->data->port_id,
 	      (void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
 	return ind_tbl;
 error:
 	rte_free(ind_tbl);
-	DEBUG("%p cannot create indirection table", (void *)dev);
+	DEBUG("port %u cannot create indirection table", dev->data->port_id);
 	return NULL;
 }
 
@@ -1204,8 +1221,9 @@ mlx5_ind_table_ibv_get(struct rte_eth_dev *dev, uint16_t queues[],
 		unsigned int i;
 
 		rte_atomic32_inc(&ind_tbl->refcnt);
-		DEBUG("%p: Indirection table %p: refcnt %d", (void *)dev,
-		      (void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
+		DEBUG("port %u indirection table %p: refcnt %d",
+		      dev->data->port_id, (void *)ind_tbl,
+		      rte_atomic32_read(&ind_tbl->refcnt));
 		for (i = 0; i != ind_tbl->queues_n; ++i)
 			mlx5_rxq_get(dev, ind_tbl->queues[i]);
 	}
@@ -1229,7 +1247,8 @@ mlx5_ind_table_ibv_release(struct rte_eth_dev *dev,
 {
 	unsigned int i;
 
-	DEBUG("%p: Indirection table %p: refcnt %d", (void *)dev,
+	DEBUG("port %u indirection table %p: refcnt %d",
+	      ((struct priv *)dev->data->dev_private)->port,
 	      (void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
 	if (rte_atomic32_dec_and_test(&ind_tbl->refcnt))
 		claim_zero(ibv_destroy_rwq_ind_table(ind_tbl->ind_table));
@@ -1260,8 +1279,8 @@ mlx5_ind_table_ibv_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	LIST_FOREACH(ind_tbl, &priv->ind_tbls, next) {
-		DEBUG("%p: Verbs indirection table %p still referenced",
-		      (void *)dev, (void *)ind_tbl);
+		DEBUG("port %u Verbs indirection table %p still referenced",
+		      dev->data->port_id, (void *)ind_tbl);
 		++ret;
 	}
 	return ret;
@@ -1336,7 +1355,7 @@ mlx5_hrxq_new(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 	memcpy(hrxq->rss_key, rss_key, rss_key_len);
 	rte_atomic32_inc(&hrxq->refcnt);
 	LIST_INSERT_HEAD(&priv->hrxqs, hrxq, next);
-	DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)dev,
+	DEBUG("port %u hash Rx queue %p: refcnt %d", dev->data->port_id,
 	      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
 	return hrxq;
 error:
@@ -1389,7 +1408,7 @@ mlx5_hrxq_get(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 			continue;
 		}
 		rte_atomic32_inc(&hrxq->refcnt);
-		DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)dev,
+		DEBUG("port %u hash Rx queue %p: refcnt %d", dev->data->port_id,
 		      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
 		return hrxq;
 	}
@@ -1410,7 +1429,8 @@ mlx5_hrxq_get(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 int
 mlx5_hrxq_release(struct rte_eth_dev *dev, struct mlx5_hrxq *hrxq)
 {
-	DEBUG("%p: Hash Rx queue %p: refcnt %d", (void *)dev,
+	DEBUG("port %u hash Rx queue %p: refcnt %d",
+	      ((struct priv *)dev->data->dev_private)->port,
 	      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
 	if (rte_atomic32_dec_and_test(&hrxq->refcnt)) {
 		claim_zero(ibv_destroy_qp(hrxq->qp));
@@ -1440,8 +1460,8 @@ mlx5_hrxq_ibv_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	LIST_FOREACH(hrxq, &priv->hrxqs, next) {
-		DEBUG("%p: Verbs Hash Rx queue %p still referenced",
-		      (void *)dev, (void *)hrxq);
+		DEBUG("port %u Verbs hash Rx queue %p still referenced",
+		      dev->data->port_id, (void *)hrxq);
 		++ret;
 	}
 	return ret;
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index d91828498..47a8729a8 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -154,6 +154,7 @@ struct mlx5_rxq_ctrl {
 	struct mlx5_rxq_data rxq; /* Data path structure. */
 	unsigned int socket; /* CPU socket ID for allocations. */
 	unsigned int irq:1; /* Whether IRQ is enabled. */
+	uint16_t idx; /* Queue index. */
 };
 
 /* Indirection table. */
@@ -214,6 +215,7 @@ struct mlx5_txq_data {
 struct mlx5_txq_ibv {
 	LIST_ENTRY(mlx5_txq_ibv) next; /* Pointer to the next element. */
 	rte_atomic32_t refcnt; /* Reference counter. */
+	struct mlx5_txq_ctrl *txq_ctrl; /* Pointer to the control queue. */
 	struct ibv_cq *cq; /* Completion Queue. */
 	struct ibv_qp *qp; /* Queue Pair. */
 };
@@ -230,6 +232,7 @@ struct mlx5_txq_ctrl {
 	struct mlx5_txq_data txq; /* Data path structure. */
 	off_t uar_mmap_offset; /* UAR mmap offset for non-primary process. */
 	volatile void *bf_reg_orig; /* Blueflame register from verbs. */
+	uint16_t idx; /* Queue index. */
 };
 
 /* mlx5_rxq.c */
@@ -469,7 +472,7 @@ mlx5_tx_complete(struct mlx5_txq_data *txq)
 	if ((MLX5_CQE_OPCODE(cqe->op_own) == MLX5_CQE_RESP_ERR) ||
 	    (MLX5_CQE_OPCODE(cqe->op_own) == MLX5_CQE_REQ_ERR)) {
 		if (!check_cqe_seen(cqe)) {
-			ERROR("unexpected error CQE, TX stopped");
+			ERROR("unexpected error CQE, Tx stopped");
 			rte_hexdump(stderr, "MLX5 TXQ:",
 				    (const void *)((uintptr_t)txq->wqes),
 				    ((1 << txq->wqe_n) *
@@ -586,7 +589,7 @@ mlx5_tx_mb2mr(struct mlx5_txq_data *txq, struct rte_mbuf *mb)
 	} else {
 		struct rte_mempool *mp = mlx5_tx_mb2mp(mb);
 
-		WARN("Failed to register mempool 0x%p(%s)",
+		WARN("failed to register mempool 0x%p(%s)",
 		      (void *)mp, mp->name);
 	}
 	return (uint32_t)-1;
diff --git a/drivers/net/mlx5/mlx5_socket.c b/drivers/net/mlx5/mlx5_socket.c
index 5499a01b9..f4a5c835e 100644
--- a/drivers/net/mlx5/mlx5_socket.c
+++ b/drivers/net/mlx5/mlx5_socket.c
@@ -68,7 +68,8 @@ mlx5_socket_init(struct rte_eth_dev *dev)
 	ret = socket(AF_UNIX, SOCK_STREAM, 0);
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("secondary process not supported: %s", strerror(errno));
+		WARN("port %u secondary process not supported: %s",
+		     dev->data->port_id, strerror(errno));
 		goto error;
 	}
 	priv->primary_socket = ret;
@@ -89,14 +90,15 @@ mlx5_socket_init(struct rte_eth_dev *dev)
 		   sizeof(sun));
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("cannot bind socket, secondary process not supported: %s",
-		     strerror(errno));
+		WARN("port %u cannot bind socket, secondary process not"
+		     " supported: %s", dev->data->port_id, strerror(errno));
 		goto close;
 	}
 	ret = listen(priv->primary_socket, 0);
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("Secondary process not supported: %s", strerror(errno));
+		WARN("port %u secondary process not supported: %s",
+		     dev->data->port_id, strerror(errno));
 		goto close;
 	}
 	return 0;
@@ -155,26 +157,29 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	/* Accept the connection from the client. */
 	conn_sock = accept(priv->primary_socket, NULL, NULL);
 	if (conn_sock < 0) {
-		WARN("connection failed: %s", strerror(errno));
+		WARN("port %u connection failed: %s", dev->data->port_id,
+		     strerror(errno));
 		return;
 	}
 	ret = setsockopt(conn_sock, SOL_SOCKET, SO_PASSCRED, &(int){1},
 					 sizeof(int));
 	if (ret < 0) {
 		ret = errno;
-		WARN("cannot change socket options: %s", strerror(rte_errno));
+		WARN("port %u cannot change socket options: %s",
+		     dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
 	ret = recvmsg(conn_sock, &msg, MSG_WAITALL);
 	if (ret < 0) {
 		ret = errno;
-		WARN("received an empty message: %s", strerror(rte_errno));
+		WARN("port %u received an empty message: %s",
+		     dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
 	/* Expect to receive credentials only. */
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
-		WARN("no message");
+		WARN("port %u no message", dev->data->port_id);
 		goto error;
 	}
 	if ((cmsg->cmsg_type == SCM_CREDENTIALS) &&
@@ -184,13 +189,13 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	}
 	cmsg = CMSG_NXTHDR(&msg, cmsg);
 	if (cmsg != NULL) {
-		WARN("Message wrongly formatted");
+		WARN("port %u message wrongly formatted", dev->data->port_id);
 		goto error;
 	}
 	/* Make sure all the ancillary data was received and valid. */
 	if ((cred == NULL) || (cred->uid != getuid()) ||
 	    (cred->gid != getgid())) {
-		WARN("wrong credentials");
+		WARN("port %u wrong credentials", dev->data->port_id);
 		goto error;
 	}
 	/* Set-up the ancillary data. */
@@ -203,7 +208,7 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	*fd = priv->ctx->cmd_fd;
 	ret = sendmsg(conn_sock, &msg, 0);
 	if (ret < 0)
-		WARN("cannot send response");
+		WARN("port %u cannot send response", dev->data->port_id);
 error:
 	close(conn_sock);
 }
@@ -245,7 +250,7 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	ret = socket(AF_UNIX, SOCK_STREAM, 0);
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("cannot connect to primary");
+		WARN("port %u cannot connect to primary", dev->data->port_id);
 		goto error;
 	}
 	socket_fd = ret;
@@ -254,13 +259,13 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	ret = connect(socket_fd, (const struct sockaddr *)&sun, sizeof(sun));
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("cannot connect to primary");
+		WARN("port %u cannot connect to primary", dev->data->port_id);
 		goto error;
 	}
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
 		rte_errno = EINVAL;
-		DEBUG("cannot get first message");
+		DEBUG("port %u cannot get first message", dev->data->port_id);
 		goto error;
 	}
 	cmsg->cmsg_level = SOL_SOCKET;
@@ -269,7 +274,7 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	cred = (struct ucred *)CMSG_DATA(cmsg);
 	if (cred == NULL) {
 		rte_errno = EINVAL;
-		DEBUG("no credentials received");
+		DEBUG("port %u no credentials received", dev->data->port_id);
 		goto error;
 	}
 	cred->pid = getpid();
@@ -278,25 +283,27 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	ret = sendmsg(socket_fd, &msg, MSG_DONTWAIT);
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("cannot send credentials to primary: %s",
-		     strerror(errno));
+		WARN("port %u cannot send credentials to primary: %s",
+		     dev->data->port_id, strerror(errno));
 		goto error;
 	}
 	ret = recvmsg(socket_fd, &msg, MSG_WAITALL);
 	if (ret <= 0) {
 		rte_errno = errno;
-		WARN("no message from primary: %s", strerror(errno));
+		WARN("port %u no message from primary: %s",
+		     dev->data->port_id, strerror(errno));
 		goto error;
 	}
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
 		rte_errno = EINVAL;
-		WARN("No file descriptor received");
+		WARN("port %u no file descriptor received", dev->data->port_id);
 		goto error;
 	}
 	fd = (int *)CMSG_DATA(cmsg);
 	if (*fd < 0) {
-		WARN("no file descriptor received: %s", strerror(errno));
+		WARN("port %u no file descriptor received: %s",
+		     dev->data->port_id, strerror(errno));
 		rte_errno = *fd;
 		goto error;
 	}
diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index b3500df97..cd8a94a48 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -160,7 +160,8 @@ mlx5_read_dev_counters(struct rte_eth_dev *dev, uint64_t *stats)
 	ifr.ifr_data = (caddr_t)et_stats;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("unable to read statistic values from device");
+		WARN("port %u unable to read statistic values from device",
+		     dev->data->port_id);
 		return ret;
 	}
 	for (i = 0; i != xstats_n; ++i) {
@@ -206,7 +207,8 @@ mlx5_ethtool_get_stats_n(struct rte_eth_dev *dev) {
 	ifr.ifr_data = (caddr_t)&drvinfo;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("unable to query number of statistics");
+		WARN("port %u unable to query number of statistics",
+		     dev->data->port_id);
 		return ret;
 	}
 	return drvinfo.n_stats;
@@ -233,7 +235,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 
 	ret = mlx5_ethtool_get_stats_n(dev);
 	if (ret < 0) {
-		WARN("no extended statistics available");
+		WARN("port %u no extended statistics available",
+		     dev->data->port_id);
 		return;
 	}
 	dev_stats_n = ret;
@@ -244,7 +247,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 		  rte_malloc("xstats_strings",
 			     str_sz + sizeof(struct ethtool_gstrings), 0);
 	if (!strings) {
-		WARN("unable to allocate memory for xstats");
+		WARN("port %u unable to allocate memory for xstats",
+		     dev->data->port_id);
 		return;
 	}
 	strings->cmd = ETHTOOL_GSTRINGS;
@@ -253,7 +257,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 	ifr.ifr_data = (caddr_t)strings;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("unable to get statistic names");
+		WARN("port %u unable to get statistic names",
+		     dev->data->port_id);
 		goto free;
 	}
 	for (j = 0; j != xstats_n; ++j)
@@ -274,7 +279,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 		if (mlx5_counters_init[j].ib)
 			continue;
 		if (xstats_ctrl->dev_table_idx[j] >= dev_stats_n) {
-			WARN("counter \"%s\" is not recognized",
+			WARN("port %u counter \"%s\" is not recognized",
+			     dev->data->port_id,
 			     mlx5_counters_init[j].dpdk_name);
 			goto free;
 		}
@@ -283,7 +289,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 	assert(xstats_n <= MLX5_MAX_XSTATS);
 	ret = mlx5_read_dev_counters(dev, xstats_ctrl->base);
 	if (ret)
-		ERROR("cannot read device counters: %s", strerror(rte_errno));
+		ERROR("port %u cannot read device counters: %s",
+		      dev->data->port_id, strerror(rte_errno));
 free:
 	rte_free(strings);
 }
@@ -450,7 +457,7 @@ mlx5_xstats_reset(struct rte_eth_dev *dev)
 
 	stats_n = mlx5_ethtool_get_stats_n(dev);
 	if (stats_n < 0) {
-		ERROR("%p cannot get stats: %s", (void *)dev,
+		ERROR("port %u cannot get stats: %s", dev->data->port_id,
 		      strerror(-stats_n));
 		return;
 	}
@@ -458,8 +465,8 @@ mlx5_xstats_reset(struct rte_eth_dev *dev)
 		mlx5_xstats_init(dev);
 	ret = mlx5_read_dev_counters(dev, counters);
 	if (ret) {
-		ERROR("%p cannot read device counters: %s", (void *)dev,
-		      strerror(rte_errno));
+		ERROR("port %u cannot read device counters: %s",
+		      dev->data->port_id, strerror(rte_errno));
 		return;
 	}
 	for (i = 0; i != n; ++i)
@@ -481,7 +488,7 @@ mlx5_xstats_reset(struct rte_eth_dev *dev)
  */
 int
 mlx5_xstats_get_names(struct rte_eth_dev *dev __rte_unused,
-		struct rte_eth_xstat_name *xstats_names, unsigned int n)
+		      struct rte_eth_xstat_name *xstats_names, unsigned int n)
 {
 	unsigned int i;
 
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index 5d2eff506..fd9b62251 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -177,37 +177,38 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 	dev->data->dev_started = 1;
 	ret = mlx5_flow_create_drop_queue(dev);
 	if (ret) {
-		ERROR("%p: Drop queue allocation failed: %s",
-		      (void *)dev, strerror(rte_errno));
+		ERROR("port %u drop queue allocation failed: %s",
+		      dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
-	DEBUG("%p: allocating and configuring hash RX queues", (void *)dev);
+	DEBUG("port %u allocating and configuring hash Rx queues",
+	      dev->data->port_id);
 	rte_mempool_walk(mlx5_mp2mr_iter, priv);
 	ret = mlx5_txq_start(dev);
 	if (ret) {
-		ERROR("%p: Tx Queue allocation failed: %s",
-		      (void *)dev, strerror(rte_errno));
+		ERROR("port %u Tx queue allocation failed: %s",
+		      dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
 	ret = mlx5_rxq_start(dev);
 	if (ret) {
-		ERROR("%p: Rx Queue allocation failed: %s",
-		      (void *)dev, strerror(rte_errno));
+		ERROR("port %u Rx queue allocation failed: %s",
+		      dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
 	ret = mlx5_rx_intr_vec_enable(dev);
 	if (ret) {
-		ERROR("%p: Rx interrupt vector creation failed",
-		      (void *)dev);
+		ERROR("port %u Rx interrupt vector creation failed",
+		      dev->data->port_id);
 		goto error;
 	}
 	mlx5_xstats_init(dev);
 	/* Update link status and Tx/Rx callbacks for the first time. */
 	memset(&dev->data->dev_link, 0, sizeof(struct rte_eth_link));
-	INFO("Forcing port %u link to be up", dev->data->port_id);
+	INFO("port %u forcing link to be up", dev->data->port_id);
 	ret = mlx5_force_link_status_change(dev, ETH_LINK_UP);
 	if (ret) {
-		DEBUG("Failed to set port %u link to be up",
+		DEBUG("failed to set port %u link to be up",
 		      dev->data->port_id);
 		goto error;
 	}
@@ -248,7 +249,8 @@ mlx5_dev_stop(struct rte_eth_dev *dev)
 	dev->tx_pkt_burst = removed_tx_burst;
 	rte_wmb();
 	usleep(1000 * priv->rxqs_n);
-	DEBUG("%p: cleaning up and destroying hash RX queues", (void *)dev);
+	DEBUG("port %u cleaning up and destroying hash Rx queues",
+	      dev->data->port_id);
 	mlx5_flow_stop(dev, &priv->flows);
 	mlx5_traffic_disable(dev);
 	mlx5_rx_intr_vec_disable(dev);
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index baf3fe984..b6d0066fc 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -74,7 +74,8 @@ txq_alloc_elts(struct mlx5_txq_ctrl *txq_ctrl)
 
 	for (i = 0; (i != elts_n); ++i)
 		(*txq_ctrl->txq.elts)[i] = NULL;
-	DEBUG("%p: allocated and configured %u WRs", (void *)txq_ctrl, elts_n);
+	DEBUG("port %u Tx queue %u allocated and configured %u WRs",
+	      txq_ctrl->priv->dev->data->port_id, txq_ctrl->idx, elts_n);
 	txq_ctrl->txq.elts_head = 0;
 	txq_ctrl->txq.elts_tail = 0;
 	txq_ctrl->txq.elts_comp = 0;
@@ -95,7 +96,8 @@ txq_free_elts(struct mlx5_txq_ctrl *txq_ctrl)
 	uint16_t elts_tail = txq_ctrl->txq.elts_tail;
 	struct rte_mbuf *(*elts)[elts_n] = txq_ctrl->txq.elts;
 
-	DEBUG("%p: freeing WRs", (void *)txq_ctrl);
+	DEBUG("port %u Tx queue %u freeing WRs",
+	      txq_ctrl->priv->dev->data->port_id, txq_ctrl->idx);
 	txq_ctrl->txq.elts_head = 0;
 	txq_ctrl->txq.elts_tail = 0;
 	txq_ctrl->txq.elts_comp = 0;
@@ -142,41 +144,40 @@ mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		container_of(txq, struct mlx5_txq_ctrl, txq);
 
 	if (desc <= MLX5_TX_COMP_THRESH) {
-		WARN("%p: number of descriptors requested for TX queue %u"
+		WARN("port %u number of descriptors requested for Tx queue %u"
 		     " must be higher than MLX5_TX_COMP_THRESH, using"
 		     " %u instead of %u",
-		     (void *)dev, idx, MLX5_TX_COMP_THRESH + 1, desc);
+		     dev->data->port_id, idx, MLX5_TX_COMP_THRESH + 1, desc);
 		desc = MLX5_TX_COMP_THRESH + 1;
 	}
 	if (!rte_is_power_of_2(desc)) {
 		desc = 1 << log2above(desc);
-		WARN("%p: increased number of descriptors in TX queue %u"
+		WARN("port %u increased number of descriptors in Tx queue %u"
 		     " to the next power of two (%d)",
-		     (void *)dev, idx, desc);
+		     dev->data->port_id, idx, desc);
 	}
-	DEBUG("%p: configuring queue %u for %u descriptors",
-	      (void *)dev, idx, desc);
+	DEBUG("port %u configuring queue %u for %u descriptors",
+	      dev->data->port_id, idx, desc);
 	if (idx >= priv->txqs_n) {
-		ERROR("%p: queue index out of range (%u >= %u)",
-		      (void *)dev, idx, priv->txqs_n);
+		ERROR("port %u Tx queue index out of range (%u >= %u)",
+		      dev->data->port_id, idx, priv->txqs_n);
 		rte_errno = EOVERFLOW;
 		return -rte_errno;
 	}
 	if (!mlx5_txq_releasable(dev, idx)) {
 		rte_errno = EBUSY;
-		ERROR("%p: unable to release queue index %u",
-		      (void *)dev, idx);
+		ERROR("port %u unable to release queue index %u",
+		      dev->data->port_id, idx);
 		return -rte_errno;
 	}
 	mlx5_txq_release(dev, idx);
 	txq_ctrl = mlx5_txq_new(dev, idx, desc, socket, conf);
 	if (!txq_ctrl) {
-		ERROR("%p: unable to allocate queue index %u",
-		      (void *)dev, idx);
+		ERROR("port %u unable to allocate queue index %u",
+		      dev->data->port_id, idx);
 		return -rte_errno;
 	}
-	DEBUG("%p: adding TX queue %p to list",
-	      (void *)dev, (void *)txq_ctrl);
+	DEBUG("port %u adding Tx queue %u to list", dev->data->port_id, idx);
 	(*priv->txqs)[idx] = &txq_ctrl->txq;
 	return 0;
 }
@@ -202,8 +203,8 @@ mlx5_tx_queue_release(void *dpdk_txq)
 	for (i = 0; (i != priv->txqs_n); ++i)
 		if ((*priv->txqs)[i] == txq) {
 			mlx5_txq_release(priv->dev, i);
-			DEBUG("%p: removing TX queue %p from list",
-			      (void *)priv->dev, (void *)txq_ctrl);
+			DEBUG("port %u removing Tx queue %u from list",
+			      priv->dev->data->port_id, txq_ctrl->idx);
 			break;
 		}
 }
@@ -249,6 +250,7 @@ mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd)
 			continue;
 		txq = (*priv->txqs)[i];
 		txq_ctrl = container_of(txq, struct mlx5_txq_ctrl, txq);
+		assert(txq_ctrl->idx == (uint16_t)i);
 		/* UAR addr form verbs used to find dup and offset in page. */
 		uar_va = (uintptr_t)txq_ctrl->bf_reg_orig;
 		off = uar_va & (page_size - 1); /* offset in page. */
@@ -273,8 +275,9 @@ mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd)
 				   txq_ctrl->uar_mmap_offset);
 			if (ret != addr) {
 				/* fixed mmap have to return same address */
-				ERROR("call to mmap failed on UAR for txq %d\n",
-				      i);
+				ERROR("port %u call to mmap failed on UAR for"
+				      " txq %u", dev->data->port_id,
+				      txq_ctrl->idx);
 				rte_errno = ENXIO;
 				return -rte_errno;
 			}
@@ -325,7 +328,8 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_TX_QUEUE;
 	priv->verbs_alloc_ctx.obj = txq_ctrl;
 	if (mlx5_getenv_int("MLX5_ENABLE_CQE_COMPRESSION")) {
-		ERROR("MLX5_ENABLE_CQE_COMPRESSION must never be set");
+		ERROR("port %u MLX5_ENABLE_CQE_COMPRESSION must never be set",
+		      dev->data->port_id);
 		rte_errno = EINVAL;
 		return NULL;
 	}
@@ -340,7 +344,8 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		cqe_n += MLX5_TX_COMP_THRESH_INLINE_DIV;
 	tmpl.cq = ibv_create_cq(priv->ctx, cqe_n, NULL, NULL, 0);
 	if (tmpl.cq == NULL) {
-		ERROR("%p: CQ creation failure", (void *)txq_ctrl);
+		ERROR("port %u Tx queue %u CQ creation failure",
+		      dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
@@ -382,7 +387,8 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	}
 	tmpl.qp = ibv_create_qp_ex(priv->ctx, &attr.init);
 	if (tmpl.qp == NULL) {
-		ERROR("%p: QP creation failure", (void *)txq_ctrl);
+		ERROR("port %u Tx queue %u QP creation failure",
+		      dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
@@ -394,7 +400,8 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	};
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, (IBV_QP_STATE | IBV_QP_PORT));
 	if (ret) {
-		ERROR("%p: QP state to IBV_QPS_INIT failed", (void *)txq_ctrl);
+		ERROR("port %u Tx queue %u QP state to IBV_QPS_INIT failed",
+		      dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
@@ -403,21 +410,24 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	};
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, IBV_QP_STATE);
 	if (ret) {
-		ERROR("%p: QP state to IBV_QPS_RTR failed", (void *)txq_ctrl);
+		ERROR("port %u Tx queue %u QP state to IBV_QPS_RTR failed",
+		      dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
 	attr.mod.qp_state = IBV_QPS_RTS;
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, IBV_QP_STATE);
 	if (ret) {
-		ERROR("%p: QP state to IBV_QPS_RTS failed", (void *)txq_ctrl);
+		ERROR("port %u Tx queue %u QP state to IBV_QPS_RTS failed",
+		      dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
 	txq_ibv = rte_calloc_socket(__func__, 1, sizeof(struct mlx5_txq_ibv), 0,
 				    txq_ctrl->socket);
 	if (!txq_ibv) {
-		ERROR("%p: cannot allocate memory", (void *)txq_ctrl);
+		ERROR("port %u Tx queue %u cannot allocate memory",
+		      dev->data->port_id, idx);
 		rte_errno = ENOMEM;
 		goto error;
 	}
@@ -431,8 +441,9 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		goto error;
 	}
 	if (cq_info.cqe_size != RTE_CACHE_LINE_SIZE) {
-		ERROR("Wrong MLX5_CQE_SIZE environment variable value: "
-		      "it should be set to %u", RTE_CACHE_LINE_SIZE);
+		ERROR("port %u wrong MLX5_CQE_SIZE environment variable value: "
+		      "it should be set to %u", dev->data->port_id,
+		      RTE_CACHE_LINE_SIZE);
 		rte_errno = EINVAL;
 		goto error;
 	}
@@ -456,13 +467,15 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (qp.comp_mask & MLX5DV_QP_MASK_UAR_MMAP_OFFSET) {
 		txq_ctrl->uar_mmap_offset = qp.uar_mmap_offset;
 	} else {
-		ERROR("Failed to retrieve UAR info, invalid libmlx5.so version");
+		ERROR("port %u failed to retrieve UAR info, invalid libmlx5.so",
+		      dev->data->port_id);
 		rte_errno = EINVAL;
 		goto error;
 	}
-	DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)dev,
-	      (void *)txq_ibv, rte_atomic32_read(&txq_ibv->refcnt));
+	DEBUG("port %u Verbs Tx queue %u: refcnt %d", dev->data->port_id, idx,
+	      rte_atomic32_read(&txq_ibv->refcnt));
 	LIST_INSERT_HEAD(&priv->txqsibv, txq_ibv, next);
+	txq_ibv->txq_ctrl = txq_ctrl;
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return txq_ibv;
 error:
@@ -500,8 +513,8 @@ mlx5_txq_ibv_get(struct rte_eth_dev *dev, uint16_t idx)
 	txq_ctrl = container_of((*priv->txqs)[idx], struct mlx5_txq_ctrl, txq);
 	if (txq_ctrl->ibv) {
 		rte_atomic32_inc(&txq_ctrl->ibv->refcnt);
-		DEBUG("%p: Verbs Tx queue %p: refcnt %d", (void *)dev,
-		      (void *)txq_ctrl->ibv,
+		DEBUG("port %u Verbs Tx queue %u: refcnt %d",
+		      dev->data->port_id, txq_ctrl->idx,
 		      rte_atomic32_read(&txq_ctrl->ibv->refcnt));
 	}
 	return txq_ctrl->ibv;
@@ -520,8 +533,9 @@ int
 mlx5_txq_ibv_release(struct mlx5_txq_ibv *txq_ibv)
 {
 	assert(txq_ibv);
-	DEBUG("Verbs Tx queue %p: refcnt %d",
-	      (void *)txq_ibv, rte_atomic32_read(&txq_ibv->refcnt));
+	DEBUG("port %u Verbs Tx queue %u: refcnt %d",
+	      txq_ibv->txq_ctrl->priv->dev->data->port_id,
+	      txq_ibv->txq_ctrl->idx, rte_atomic32_read(&txq_ibv->refcnt));
 	if (rte_atomic32_dec_and_test(&txq_ibv->refcnt)) {
 		claim_zero(ibv_destroy_qp(txq_ibv->qp));
 		claim_zero(ibv_destroy_cq(txq_ibv->cq));
@@ -562,8 +576,9 @@ mlx5_txq_ibv_verify(struct rte_eth_dev *dev)
 	struct mlx5_txq_ibv *txq_ibv;
 
 	LIST_FOREACH(txq_ibv, &priv->txqsibv, next) {
-		DEBUG("%p: Verbs Tx queue %p still referenced", (void *)dev,
-		      (void *)txq_ibv);
+		DEBUG("port %u Verbs Tx queue %u still referenced",
+		      dev->data->port_id,
+		      txq_ibv->txq_ctrl->idx);
 		++ret;
 	}
 	return ret;
@@ -609,12 +624,13 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	tmpl->priv = priv;
 	tmpl->socket = socket;
 	tmpl->txq.elts_n = log2above(desc);
+	tmpl->idx = idx;
 	if (priv->mps == MLX5_MPW_ENHANCED)
 		tmpl->txq.mpw_hdr_dseg = priv->mpw_hdr_dseg;
 	/* MRs will be registered in mp2mr[] later. */
-	DEBUG("priv->device_attr.max_qp_wr is %d",
+	DEBUG("port %u priv->device_attr.max_qp_wr is %d", dev->data->port_id,
 	      priv->device_attr.orig_attr.max_qp_wr);
-	DEBUG("priv->device_attr.max_sge is %d",
+	DEBUG("port %u priv->device_attr.max_sge is %d", dev->data->port_id,
 	      priv->device_attr.orig_attr.max_sge);
 	if (priv->txq_inline && (priv->txqs_n >= priv->txqs_inline)) {
 		unsigned int ds_cnt;
@@ -666,9 +682,10 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 
 			max_inline = max_inline - (max_inline %
 						   RTE_CACHE_LINE_SIZE);
-			WARN("txq inline is too large (%d) setting it to "
-			     "the maximum possible: %d\n",
-			     priv->txq_inline, max_inline);
+			WARN("port %u txq inline is too large (%d) setting it"
+			     " to the maximum possible: %d\n",
+			     priv->dev->data->port_id, priv->txq_inline,
+			     max_inline);
 			tmpl->txq.max_inline = max_inline / RTE_CACHE_LINE_SIZE;
 		}
 	}
@@ -684,8 +701,8 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		(struct rte_mbuf *(*)[1 << tmpl->txq.elts_n])(tmpl + 1);
 	tmpl->txq.stats.idx = idx;
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("%p: Tx queue %p: refcnt %d", (void *)dev,
-	      (void *)tmpl, rte_atomic32_read(&tmpl->refcnt));
+	DEBUG("port %u Tx queue %u: refcnt %d", dev->data->port_id,
+	      idx, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->txqsctrl, tmpl, next);
 	return tmpl;
 }
@@ -720,8 +737,8 @@ mlx5_txq_get(struct rte_eth_dev *dev, uint16_t idx)
 						     ctrl->txq.mp2mr[i]->mp));
 		}
 		rte_atomic32_inc(&ctrl->refcnt);
-		DEBUG("%p: Tx queue %p: refcnt %d", (void *)dev,
-		      (void *)ctrl, rte_atomic32_read(&ctrl->refcnt));
+		DEBUG("port %u Tx queue %u refcnt %d", dev->data->port_id,
+		      ctrl->idx, rte_atomic32_read(&ctrl->refcnt));
 	}
 	return ctrl;
 }
@@ -748,8 +765,8 @@ mlx5_txq_release(struct rte_eth_dev *dev, uint16_t idx)
 	if (!(*priv->txqs)[idx])
 		return 0;
 	txq = container_of((*priv->txqs)[idx], struct mlx5_txq_ctrl, txq);
-	DEBUG("%p: Tx queue %p: refcnt %d", (void *)dev,
-	      (void *)txq, rte_atomic32_read(&txq->refcnt));
+	DEBUG("port %u Tx queue %u: refcnt %d", dev->data->port_id,
+	      txq->idx, rte_atomic32_read(&txq->refcnt));
 	if (txq->ibv && !mlx5_txq_ibv_release(txq->ibv))
 		txq->ibv = NULL;
 	for (i = 0; i != MLX5_PMD_TX_MP_CACHE; ++i) {
@@ -811,8 +828,8 @@ mlx5_txq_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	LIST_FOREACH(txq, &priv->txqsctrl, next) {
-		DEBUG("%p: Tx Queue %p still referenced", (void *)dev,
-		      (void *)txq);
+		DEBUG("port %u Tx queue %u still referenced",
+		      dev->data->port_id, txq->idx);
 		++ret;
 	}
 	return ret;
diff --git a/drivers/net/mlx5/mlx5_vlan.c b/drivers/net/mlx5/mlx5_vlan.c
index 7e4830138..377d06884 100644
--- a/drivers/net/mlx5/mlx5_vlan.c
+++ b/drivers/net/mlx5/mlx5_vlan.c
@@ -62,8 +62,8 @@ mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
 	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
-	DEBUG("%p: %s VLAN filter ID %" PRIu16,
-	      (void *)dev, (on ? "enable" : "disable"), vlan_id);
+	DEBUG("port %u %s VLAN filter ID %" PRIu16,
+	      dev->data->port_id, (on ? "enable" : "disable"), vlan_id);
 	assert(priv->vlan_filter_n <= RTE_DIM(priv->vlan_filter));
 	for (i = 0; (i != priv->vlan_filter_n); ++i)
 		if (priv->vlan_filter[i] == vlan_id)
@@ -125,16 +125,18 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 
 	/* Validate hw support */
 	if (!priv->hw_vlan_strip) {
-		ERROR("VLAN stripping is not supported");
+		ERROR("port %u VLAN stripping is not supported",
+		      dev->data->port_id);
 		return;
 	}
 	/* Validate queue number */
 	if (queue >= priv->rxqs_n) {
-		ERROR("VLAN stripping, invalid queue number %d", queue);
+		ERROR("port %u VLAN stripping, invalid queue number %d",
+		      dev->data->port_id, queue);
 		return;
 	}
-	DEBUG("set VLAN offloads 0x%x for port %d queue %d",
-	      vlan_offloads, rxq->port_id, queue);
+	DEBUG("port %u set VLAN offloads 0x%x for port %uqueue %d",
+	      dev->data->port_id, vlan_offloads, rxq->port_id, queue);
 	if (!rxq_ctrl->ibv) {
 		/* Update related bits in RX queue. */
 		rxq->vlan_strip = !!on;
@@ -147,8 +149,8 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 	};
 	ret = ibv_modify_wq(rxq_ctrl->ibv->wq, &mod);
 	if (ret) {
-		ERROR("%p: failed to modified stripping mode: %s",
-		      (void *)dev, strerror(rte_errno));
+		ERROR("port %u failed to modified stripping mode: %s",
+		      dev->data->port_id, strerror(rte_errno));
 		return;
 	}
 	/* Update related bits in RX queue. */
@@ -176,7 +178,8 @@ mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask)
 		int hw_vlan_strip = !!dev->data->dev_conf.rxmode.hw_vlan_strip;
 
 		if (!priv->hw_vlan_strip) {
-			ERROR("VLAN stripping is not supported");
+			ERROR("port %u VLAN stripping is not supported",
+			      dev->data->port_id);
 			return 0;
 		}
 		/* Run on every RX queue and set/reset VLAN stripping. */
-- 
2.11.0

[dpdk-stable] [PATCH v2 29/67] net/mlx5: use dynamic logging

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit a170a30d22a8c34c36541d0dd6bcc2fcc4c9ee2f ]

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.c         | 221 +++++++++++++++++--------------
 drivers/net/mlx5/mlx5_ethdev.c  | 104 +++++++++------
 drivers/net/mlx5/mlx5_flow.c    | 105 ++++++++-------
 drivers/net/mlx5/mlx5_mac.c     |  12 +-
 drivers/net/mlx5/mlx5_mr.c      |  85 ++++++------
 drivers/net/mlx5/mlx5_rxmode.c  |  16 +--
 drivers/net/mlx5/mlx5_rxq.c     | 283 ++++++++++++++++++++++------------------
 drivers/net/mlx5/mlx5_rxtx.h    |  17 +--
 drivers/net/mlx5/mlx5_socket.c  |  65 +++++----
 drivers/net/mlx5/mlx5_stats.c   |  38 +++---
 drivers/net/mlx5/mlx5_trigger.c |  30 ++---
 drivers/net/mlx5/mlx5_txq.c     | 152 +++++++++++----------
 drivers/net/mlx5/mlx5_utils.h   |  27 ++--
 drivers/net/mlx5/mlx5_vlan.c    |  24 ++--
 14 files changed, 646 insertions(+), 533 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 8518fa588..911d4cf65 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -119,6 +119,10 @@ struct mlx5_args {
 	int tx_vec_en;
 	int rx_vec_en;
 };
+
+/** Driver-specific log messages type. */
+int mlx5_logtype;
+
 /**
  * Retrieve integer value from environment variable.
  *
@@ -207,9 +211,9 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 	unsigned int i;
 	int ret;
 
-	DEBUG("port %u closing device \"%s\"",
-	      dev->data->port_id,
-	      ((priv->ctx != NULL) ? priv->ctx->device->name : ""));
+	DRV_LOG(DEBUG, "port %u closing device \"%s\"",
+		dev->data->port_id,
+		((priv->ctx != NULL) ? priv->ctx->device->name : ""));
 	/* In case mlx5_dev_stop() has not been called. */
 	mlx5_dev_interrupt_handler_uninstall(dev);
 	mlx5_traffic_disable(dev);
@@ -246,35 +250,36 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 		mlx5_socket_uninit(dev);
 	ret = mlx5_hrxq_ibv_verify(dev);
 	if (ret)
-		WARN("port %u some hash Rx queue still remain",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u some hash Rx queue still remain",
+			dev->data->port_id);
 	ret = mlx5_ind_table_ibv_verify(dev);
 	if (ret)
-		WARN("port %u some indirection table still remain",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u some indirection table still remain",
+			dev->data->port_id);
 	ret = mlx5_rxq_ibv_verify(dev);
 	if (ret)
-		WARN("port %u some Verbs Rx queue still remain",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u some Verbs Rx queue still remain",
+			dev->data->port_id);
 	ret = mlx5_rxq_verify(dev);
 	if (ret)
-		WARN("port %u some Rx queues still remain",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u some Rx queues still remain",
+			dev->data->port_id);
 	ret = mlx5_txq_ibv_verify(dev);
 	if (ret)
-		WARN("port %u some Verbs Tx queue still remain",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u some Verbs Tx queue still remain",
+			dev->data->port_id);
 	ret = mlx5_txq_verify(dev);
 	if (ret)
-		WARN("port %u some Tx queues still remain",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u some Tx queues still remain",
+			dev->data->port_id);
 	ret = mlx5_flow_verify(dev);
 	if (ret)
-		WARN("port %u some flows still remain", dev->data->port_id);
+		DRV_LOG(WARNING, "port %u some flows still remain",
+			dev->data->port_id);
 	ret = mlx5_mr_verify(dev);
 	if (ret)
-		WARN("port %u some memory region still remain",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u some memory region still remain",
+			dev->data->port_id);
 	memset(priv, 0, sizeof(*priv));
 }
 
@@ -424,7 +429,7 @@ mlx5_args_check(const char *key, const char *val, void *opaque)
 	tmp = strtoul(val, NULL, 0);
 	if (errno) {
 		rte_errno = errno;
-		WARN("%s: \"%s\" is not a valid integer", key, val);
+		DRV_LOG(WARNING, "%s: \"%s\" is not a valid integer", key, val);
 		return -rte_errno;
 	}
 	if (strcmp(MLX5_RXQ_CQE_COMP_EN, key) == 0) {
@@ -446,7 +451,7 @@ mlx5_args_check(const char *key, const char *val, void *opaque)
 	} else if (strcmp(MLX5_RX_VEC_EN, key) == 0) {
 		args->rx_vec_en = !!tmp;
 	} else {
-		WARN("%s: unknown parameter", key);
+		DRV_LOG(WARNING, "%s: unknown parameter", key);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
@@ -551,17 +556,18 @@ mlx5_uar_init_primary(struct rte_eth_dev *dev)
 	addr = mmap(addr, MLX5_UAR_SIZE,
 		    PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
 	if (addr == MAP_FAILED) {
-		ERROR("port %u failed to reserve UAR address space, please"
-		      " adjust MLX5_UAR_SIZE or try --base-virtaddr",
-		      dev->data->port_id);
+		DRV_LOG(ERR,
+			"port %u failed to reserve UAR address space, please"
+			" adjust MLX5_UAR_SIZE or try --base-virtaddr",
+			dev->data->port_id);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
 	/* Accept either same addr or a new addr returned from mmap if target
 	 * range occupied.
 	 */
-	INFO("port %u reserved UAR address space: %p", dev->data->port_id,
-	     addr);
+	DRV_LOG(INFO, "port %u reserved UAR address space: %p",
+		dev->data->port_id, addr);
 	priv->uar_base = addr; /* for primary and secondary UAR re-mmap. */
 	uar_base = addr; /* process local, don't reserve again. */
 	return 0;
@@ -592,21 +598,23 @@ mlx5_uar_init_secondary(struct rte_eth_dev *dev)
 	addr = mmap(priv->uar_base, MLX5_UAR_SIZE,
 		    PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
 	if (addr == MAP_FAILED) {
-		ERROR("port %u UAR mmap failed: %p size: %llu",
-		      dev->data->port_id, priv->uar_base, MLX5_UAR_SIZE);
+		DRV_LOG(ERR, "port %u UAR mmap failed: %p size: %llu",
+			dev->data->port_id, priv->uar_base, MLX5_UAR_SIZE);
 		rte_errno = ENXIO;
 		return -rte_errno;
 	}
 	if (priv->uar_base != addr) {
-		ERROR("port %u UAR address %p size %llu occupied, please adjust "
-		      "MLX5_UAR_OFFSET or try EAL parameter --base-virtaddr",
-		      dev->data->port_id, priv->uar_base, MLX5_UAR_SIZE);
+		DRV_LOG(ERR,
+			"port %u UAR address %p size %llu occupied, please"
+			" adjust MLX5_UAR_OFFSET or try EAL parameter"
+			" --base-virtaddr",
+			dev->data->port_id, priv->uar_base, MLX5_UAR_SIZE);
 		rte_errno = ENXIO;
 		return -rte_errno;
 	}
 	uar_base = addr; /* process local, don't reserve again */
-	INFO("port %u reserved UAR address space: %p", dev->data->port_id,
-	     addr);
+	DRV_LOG(INFO, "port %u reserved UAR address space: %p",
+		dev->data->port_id, addr);
 	return 0;
 }
 
@@ -679,11 +687,11 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	/* Get mlx5_dev[] index. */
 	idx = mlx5_dev_idx(&pci_dev->addr);
 	if (idx == -1) {
-		ERROR("this driver cannot support any more adapters");
+		DRV_LOG(ERR, "this driver cannot support any more adapters");
 		err = ENOMEM;
 		goto error;
 	}
-	DEBUG("using driver device index %d", idx);
+	DRV_LOG(DEBUG, "using driver device index %d", idx);
 	/* Save PCI address. */
 	mlx5_dev[idx].pci_addr = pci_dev->addr;
 	list = ibv_get_device_list(&i);
@@ -691,7 +699,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		assert(errno);
 		err = errno;
 		if (errno == ENOSYS)
-			ERROR("cannot list devices, is ib_uverbs loaded?");
+			DRV_LOG(ERR,
+				"cannot list devices, is ib_uverbs loaded?");
 		goto error;
 	}
 	assert(i >= 0);
@@ -703,7 +712,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		struct rte_pci_addr pci_addr;
 
 		--i;
-		DEBUG("checking device \"%s\"", list[i]->name);
+		DRV_LOG(DEBUG, "checking device \"%s\"", list[i]->name);
 		if (mlx5_ibv_device_to_pci_addr(list[i], &pci_addr))
 			continue;
 		if ((pci_dev->addr.domain != pci_addr.domain) ||
@@ -725,7 +734,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		default:
 			break;
 		}
-		INFO("PCI information matches, using device \"%s\"",
+		DRV_LOG(INFO, "PCI information matches, using device \"%s\"",
 		     list[i]->name);
 		attr_ctx = ibv_open_device(list[i]);
 		rte_errno = errno;
@@ -736,16 +745,18 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		ibv_free_device_list(list);
 		switch (err) {
 		case 0:
-			ERROR("cannot access device, is mlx5_ib loaded?");
+			DRV_LOG(ERR,
+				"cannot access device, is mlx5_ib loaded?");
 			err = ENODEV;
 			goto error;
 		case EINVAL:
-			ERROR("cannot use device, are drivers up to date?");
+			DRV_LOG(ERR,
+				"cannot use device, are drivers up to date?");
 			goto error;
 		}
 	}
 	ibv_dev = list[i];
-	DEBUG("device opened");
+	DRV_LOG(DEBUG, "device opened");
 	/*
 	 * Multi-packet send is supported by ConnectX-4 Lx PF as well
 	 * as all ConnectX-5 devices.
@@ -753,14 +764,14 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	mlx5dv_query_device(attr_ctx, &attrs_out);
 	if (attrs_out.flags & MLX5DV_CONTEXT_FLAGS_MPW_ALLOWED) {
 		if (attrs_out.flags & MLX5DV_CONTEXT_FLAGS_ENHANCED_MPW) {
-			DEBUG("enhanced MPW is supported");
+			DRV_LOG(DEBUG, "enhanced MPW is supported");
 			mps = MLX5_MPW_ENHANCED;
 		} else {
-			DEBUG("MPW is supported");
+			DRV_LOG(DEBUG, "MPW is supported");
 			mps = MLX5_MPW;
 		}
 	} else {
-		DEBUG("MPW isn't supported");
+		DRV_LOG(DEBUG, "MPW isn't supported");
 		mps = MLX5_MPW_DISABLED;
 	}
 	if (RTE_CACHE_LINE_SIZE == 128 &&
@@ -772,7 +783,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		err = errno;
 		goto error;
 	}
-	INFO("%u port(s) detected", device_attr.orig_attr.phys_port_cnt);
+	DRV_LOG(INFO, "%u port(s) detected",
+		device_attr.orig_attr.phys_port_cnt);
 	for (i = 0; i < device_attr.orig_attr.phys_port_cnt; i++) {
 		char name[RTE_ETH_NAME_MAX_LEN];
 		uint32_t port = i + 1; /* ports are indexed from one */
@@ -804,7 +816,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		if (rte_eal_process_type() == RTE_PROC_SECONDARY) {
 			eth_dev = rte_eth_dev_attach_secondary(name);
 			if (eth_dev == NULL) {
-				ERROR("can not attach rte ethdev");
+				DRV_LOG(ERR, "can not attach rte ethdev");
 				rte_errno = ENOMEM;
 				err = rte_errno;
 				goto error;
@@ -826,7 +838,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			mlx5_select_tx_function(eth_dev);
 			continue;
 		}
-		DEBUG("using port %u (%08" PRIx32 ")", port, test);
+		DRV_LOG(DEBUG, "using port %u (%08" PRIx32 ")", port, test);
 		ctx = ibv_open_device(ibv_dev);
 		if (ctx == NULL) {
 			err = ENODEV;
@@ -836,23 +848,24 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		/* Check port status. */
 		err = ibv_query_port(ctx, port, &port_attr);
 		if (err) {
-			ERROR("port query failed: %s", strerror(err));
+			DRV_LOG(ERR, "port query failed: %s", strerror(err));
 			goto port_error;
 		}
 		if (port_attr.link_layer != IBV_LINK_LAYER_ETHERNET) {
-			ERROR("port %d is not configured in Ethernet mode",
-			      port);
+			DRV_LOG(ERR,
+				"port %d is not configured in Ethernet mode",
+				port);
 			err = EINVAL;
 			goto port_error;
 		}
 		if (port_attr.state != IBV_PORT_ACTIVE)
-			DEBUG("port %d is not active: \"%s\" (%d)",
-			      port, ibv_port_state_str(port_attr.state),
-			      port_attr.state);
+			DRV_LOG(DEBUG, "port %d is not active: \"%s\" (%d)",
+				port, ibv_port_state_str(port_attr.state),
+				port_attr.state);
 		/* Allocate protection domain. */
 		pd = ibv_alloc_pd(ctx);
 		if (pd == NULL) {
-			ERROR("PD allocation failure");
+			DRV_LOG(ERR, "PD allocation failure");
 			err = ENOMEM;
 			goto port_error;
 		}
@@ -862,7 +875,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 				   sizeof(*priv),
 				   RTE_CACHE_LINE_SIZE);
 		if (priv == NULL) {
-			ERROR("priv allocation failure");
+			DRV_LOG(ERR, "priv allocation failure");
 			err = ENOMEM;
 			goto port_error;
 		}
@@ -881,35 +894,36 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		priv->rx_vec_en = 1;
 		err = mlx5_args(&args, pci_dev->device.devargs);
 		if (err) {
-			ERROR("failed to process device arguments: %s",
-			      strerror(err));
+			DRV_LOG(ERR, "failed to process device arguments: %s",
+				strerror(err));
 			goto port_error;
 		}
 		mlx5_args_assign(priv, &args);
 		if (ibv_query_device_ex(ctx, NULL, &device_attr_ex)) {
-			ERROR("ibv_query_device_ex() failed");
+			DRV_LOG(ERR, "ibv_query_device_ex() failed");
 			err = errno;
 			goto port_error;
 		}
 		priv->hw_csum =
 			!!(device_attr_ex.device_cap_flags_ex &
 			   IBV_DEVICE_RAW_IP_CSUM);
-		DEBUG("checksum offloading is %ssupported",
-		      (priv->hw_csum ? "" : "not "));
+		DRV_LOG(DEBUG, "checksum offloading is %ssupported",
+			(priv->hw_csum ? "" : "not "));
 
 #ifdef HAVE_IBV_DEVICE_VXLAN_SUPPORT
 		priv->hw_csum_l2tun = !!(exp_device_attr.exp_device_cap_flags &
 					 IBV_DEVICE_VXLAN_SUPPORT);
 #endif
-		DEBUG("Rx L2 tunnel checksum offloads are %ssupported",
-		      (priv->hw_csum_l2tun ? "" : "not "));
+		DRV_LOG(DEBUG, "Rx L2 tunnel checksum offloads are %ssupported",
+			(priv->hw_csum_l2tun ? "" : "not "));
 
 #ifdef HAVE_IBV_DEVICE_COUNTERS_SET_SUPPORT
 		priv->counter_set_supported = !!(device_attr.max_counter_sets);
 		ibv_describe_counter_set(ctx, 0, &cs_desc);
-		DEBUG("counter type = %d, num of cs = %ld, attributes = %d",
-		      cs_desc.counter_type, cs_desc.num_of_cs,
-		      cs_desc.attributes);
+		DRV_LOG(DEBUG,
+			"counter type = %d, num of cs = %ld, attributes = %d",
+			cs_desc.counter_type, cs_desc.num_of_cs,
+			cs_desc.attributes);
 #endif
 		priv->ind_table_max_size =
 			device_attr_ex.rss_caps.max_rwq_indirection_table_size;
@@ -918,23 +932,24 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		if (priv->ind_table_max_size >
 				(unsigned int)ETH_RSS_RETA_SIZE_512)
 			priv->ind_table_max_size = ETH_RSS_RETA_SIZE_512;
-		DEBUG("maximum Rx indirection table size is %u",
-		      priv->ind_table_max_size);
+		DRV_LOG(DEBUG, "maximum Rx indirection table size is %u",
+			priv->ind_table_max_size);
 		priv->hw_vlan_strip = !!(device_attr_ex.raw_packet_caps &
 					 IBV_RAW_PACKET_CAP_CVLAN_STRIPPING);
-		DEBUG("VLAN stripping is %ssupported",
-		      (priv->hw_vlan_strip ? "" : "not "));
+		DRV_LOG(DEBUG, "VLAN stripping is %ssupported",
+			(priv->hw_vlan_strip ? "" : "not "));
 
 		priv->hw_fcs_strip = !!(device_attr_ex.raw_packet_caps &
 					IBV_RAW_PACKET_CAP_SCATTER_FCS);
-		DEBUG("FCS stripping configuration is %ssupported",
-		      (priv->hw_fcs_strip ? "" : "not "));
+		DRV_LOG(DEBUG, "FCS stripping configuration is %ssupported",
+			(priv->hw_fcs_strip ? "" : "not "));
 
 #ifdef HAVE_IBV_WQ_FLAG_RX_END_PADDING
 		priv->hw_padding = !!device_attr_ex.rx_pad_end_addr_align;
 #endif
-		DEBUG("hardware Rx end alignment padding is %ssupported",
-		      (priv->hw_padding ? "" : "not "));
+		DRV_LOG(DEBUG,
+			"hardware Rx end alignment padding is %ssupported",
+			(priv->hw_padding ? "" : "not "));
 		priv->tso = ((priv->tso) &&
 			    (device_attr_ex.tso_caps.max_tso > 0) &&
 			    (device_attr_ex.tso_caps.supported_qpts &
@@ -943,18 +958,21 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			priv->max_tso_payload_sz =
 				device_attr_ex.tso_caps.max_tso;
 		if (priv->mps && !mps) {
-			ERROR("multi-packet send not supported on this device"
-			      " (" MLX5_TXQ_MPW_EN ")");
+			DRV_LOG(ERR,
+				"multi-packet send not supported on this device"
+				" (" MLX5_TXQ_MPW_EN ")");
 			err = ENOTSUP;
 			goto port_error;
 		} else if (priv->mps && priv->tso) {
-			WARN("multi-packet send not supported in conjunction "
-			      "with TSO. MPS disabled");
+			DRV_LOG(WARNING,
+				"multi-packet send not supported in conjunction"
+				" with TSO. MPS disabled");
 			priv->mps = 0;
 		}
-		INFO("%s MPS is %s",
-		     priv->mps == MLX5_MPW_ENHANCED ? "enhanced " : "",
-		     priv->mps != MLX5_MPW_DISABLED ? "enabled" : "disabled");
+		DRV_LOG(INFO, "%s MPS is %s",
+			priv->mps == MLX5_MPW_ENHANCED ? "enhanced " : "",
+			priv->mps != MLX5_MPW_DISABLED ? "enabled" :
+							 "disabled");
 		/* Set default values for Enhanced MPW, a.k.a MPWv2. */
 		if (priv->mps == MLX5_MPW_ENHANCED) {
 			if (args.txqs_inline == MLX5_ARG_UNSET)
@@ -967,12 +985,12 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 						   MLX5_WQE_SIZE;
 		}
 		if (priv->cqe_comp && !cqe_comp) {
-			WARN("Rx CQE compression isn't supported");
+			DRV_LOG(WARNING, "Rx CQE compression isn't supported");
 			priv->cqe_comp = 0;
 		}
 		eth_dev = rte_eth_dev_allocate(name);
 		if (eth_dev == NULL) {
-			ERROR("can not allocate rte ethdev");
+			DRV_LOG(ERR, "can not allocate rte ethdev");
 			err = ENOMEM;
 			goto port_error;
 		}
@@ -987,34 +1005,37 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			goto port_error;
 		/* Configure the first MAC address by default. */
 		if (mlx5_get_mac(eth_dev, &mac.addr_bytes)) {
-			ERROR("port %u cannot get MAC address, is mlx5_en"
-			      " loaded? (errno: %s)", eth_dev->data->port_id,
-			      strerror(errno));
+			DRV_LOG(ERR,
+				"port %u cannot get MAC address, is mlx5_en"
+				" loaded? (errno: %s)",
+				eth_dev->data->port_id, strerror(errno));
 			err = ENODEV;
 			goto port_error;
 		}
-		INFO("port %u MAC address is %02x:%02x:%02x:%02x:%02x:%02x",
-		     eth_dev->data->port_id,
-		     mac.addr_bytes[0], mac.addr_bytes[1],
-		     mac.addr_bytes[2], mac.addr_bytes[3],
-		     mac.addr_bytes[4], mac.addr_bytes[5]);
+		DRV_LOG(INFO,
+			"port %u MAC address is %02x:%02x:%02x:%02x:%02x:%02x",
+			eth_dev->data->port_id,
+			mac.addr_bytes[0], mac.addr_bytes[1],
+			mac.addr_bytes[2], mac.addr_bytes[3],
+			mac.addr_bytes[4], mac.addr_bytes[5]);
 #ifndef NDEBUG
 		{
 			char ifname[IF_NAMESIZE];
 
 			if (mlx5_get_ifname(eth_dev, &ifname) == 0)
-				DEBUG("port %u ifname is \"%s\"",
-				      eth_dev->data->port_id, ifname);
+				DRV_LOG(DEBUG, "port %u ifname is \"%s\"",
+					eth_dev->data->port_id, ifname);
 			else
-				DEBUG("port %u ifname is unknown",
-				      eth_dev->data->port_id);
+				DRV_LOG(DEBUG, "port %u ifname is unknown",
+					eth_dev->data->port_id);
 		}
 #endif
 		/* Get actual MTU if possible. */
 		err = mlx5_get_mtu(eth_dev, &priv->mtu);
 		if (err)
 			goto port_error;
-		DEBUG("port %u MTU is %u", eth_dev->data->port_id, priv->mtu);
+		DRV_LOG(DEBUG, "port %u MTU is %u", eth_dev->data->port_id,
+			priv->mtu);
 		/*
 		 * Initialize burst functions to prevent crashes before link-up.
 		 */
@@ -1034,8 +1055,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		mlx5dv_set_context_attr(ctx, MLX5DV_CTX_ATTR_BUF_ALLOCATORS,
 					(void *)((uintptr_t)&alctr));
 		/* Bring Ethernet device up. */
-		DEBUG("port %u forcing Ethernet interface up",
-		      eth_dev->data->port_id);
+		DRV_LOG(DEBUG, "port %u forcing Ethernet interface up",
+			eth_dev->data->port_id);
 		mlx5_set_flags(eth_dev, ~IFF_UP, IFF_UP);
 		continue;
 port_error:
@@ -1143,3 +1164,11 @@ rte_mlx5_pmd_init(void)
 RTE_PMD_EXPORT_NAME(net_mlx5, __COUNTER__);
 RTE_PMD_REGISTER_PCI_TABLE(net_mlx5, mlx5_pci_id_map);
 RTE_PMD_REGISTER_KMOD_DEP(net_mlx5, "* ib_uverbs & mlx5_core & mlx5_ib");
+
+/** Initialize driver log type. */
+RTE_INIT(vdev_netvsc_init_log)
+{
+	mlx5_logtype = rte_log_register("pmd.net.mlx5");
+	if (mlx5_logtype >= 0)
+		rte_log_set_level(mlx5_logtype, RTE_LOG_NOTICE);
+}
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 8696c2d45..b78756efc 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -344,8 +344,8 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 		rte_realloc(priv->rss_conf.rss_key,
 			    rss_hash_default_key_len, 0);
 	if (!priv->rss_conf.rss_key) {
-		ERROR("port %u cannot allocate RSS hash key memory (%u)",
-		      dev->data->port_id, rxqs_n);
+		DRV_LOG(ERR, "port %u cannot allocate RSS hash key memory (%u)",
+			dev->data->port_id, rxqs_n);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
@@ -359,20 +359,20 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 	priv->rxqs = (void *)dev->data->rx_queues;
 	priv->txqs = (void *)dev->data->tx_queues;
 	if (txqs_n != priv->txqs_n) {
-		INFO("port %u Tx queues number update: %u -> %u",
-		     dev->data->port_id, priv->txqs_n, txqs_n);
+		DRV_LOG(INFO, "port %u Tx queues number update: %u -> %u",
+			dev->data->port_id, priv->txqs_n, txqs_n);
 		priv->txqs_n = txqs_n;
 	}
 	if (rxqs_n > priv->ind_table_max_size) {
-		ERROR("port %u cannot handle this many Rx queues (%u)",
-		      dev->data->port_id, rxqs_n);
+		DRV_LOG(ERR, "port %u cannot handle this many Rx queues (%u)",
+			dev->data->port_id, rxqs_n);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
 	if (rxqs_n == priv->rxqs_n)
 		return 0;
-	INFO("port %u Rx queues number update: %u -> %u",
-	     dev->data->port_id, priv->rxqs_n, rxqs_n);
+	DRV_LOG(INFO, "port %u Rx queues number update: %u -> %u",
+		dev->data->port_id, priv->rxqs_n, rxqs_n);
 	priv->rxqs_n = rxqs_n;
 	/* If the requested number of RX queues is not a power of two, use the
 	 * maximum indirection table size for better balancing.
@@ -515,8 +515,8 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 
 	ret = mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr);
 	if (ret) {
-		WARN("port %u ioctl(SIOCGIFFLAGS) failed: %s",
-		     dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(WARNING, "port %u ioctl(SIOCGIFFLAGS) failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	memset(&dev_link, 0, sizeof(dev_link));
@@ -525,8 +525,9 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 	ifr.ifr_data = (void *)&edata;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("port %u ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: %s",
-		     dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(WARNING,
+			"port %u ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	link_speed = ethtool_cmd_speed(&edata);
@@ -582,8 +583,8 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 
 	ret = mlx5_ifreq(dev, SIOCGIFFLAGS, &ifr);
 	if (ret) {
-		WARN("port %u ioctl(SIOCGIFFLAGS) failed: %s",
-		     dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(WARNING, "port %u ioctl(SIOCGIFFLAGS) failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	memset(&dev_link, 0, sizeof(dev_link));
@@ -592,8 +593,10 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 	ifr.ifr_data = (void *)&gcmd;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		DEBUG("port %u ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS)"
-		      " failed: %s", dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(DEBUG,
+			"port %u ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS)"
+			" failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	gcmd.link_mode_masks_nwords = -gcmd.link_mode_masks_nwords;
@@ -607,8 +610,10 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 	ifr.ifr_data = (void *)ecmd;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		DEBUG("port %u ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS)"
-		      " failed: %s", dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(DEBUG,
+			"port %u ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS)"
+			" failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	dev_link.link_speed = ecmd->speed;
@@ -679,14 +684,17 @@ mlx5_link_start(struct rte_eth_dev *dev)
 	mlx5_select_rx_function(dev);
 	ret = mlx5_traffic_enable(dev);
 	if (ret) {
-		ERROR("port %u error occurred while configuring control flows:"
-		      " %s", dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR,
+			"port %u error occurred while configuring control"
+			" flows: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return;
 	}
 	ret = mlx5_flow_start(dev, &priv->flows);
 	if (ret)
-		ERROR("port %u error occurred while configuring flows: %s",
-		       dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR,
+			"port %u error occurred while configuring flows: %s",
+			dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -807,7 +815,8 @@ mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 		return ret;
 	if (kern_mtu == mtu) {
 		priv->mtu = mtu;
-		DEBUG("port %u adapter MTU set to %u", dev->data->port_id, mtu);
+		DRV_LOG(DEBUG, "port %u adapter MTU set to %u",
+			dev->data->port_id, mtu);
 		return 0;
 	}
 	rte_errno = EAGAIN;
@@ -837,8 +846,10 @@ mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 	ifr.ifr_data = (void *)&ethpause;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("port %u ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM) failed:"
-		     " %s", dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(WARNING,
+			"port %u ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM) failed:"
+			" %s",
+			dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	fc_conf->autoneg = ethpause.autoneg;
@@ -888,8 +899,10 @@ mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
 		ethpause.tx_pause = 0;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("port %u ioctl(SIOCETHTOOL, ETHTOOL_SPAUSEPARAM)"
-		     " failed: %s", dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(WARNING,
+			"port %u ioctl(SIOCETHTOOL, ETHTOOL_SPAUSEPARAM)"
+			" failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return ret;
 	}
 	return 0;
@@ -1018,8 +1031,9 @@ mlx5_dev_status_handler(struct rte_eth_dev *dev)
 			dev->data->dev_conf.intr_conf.rmv == 1)
 			ret |= (1 << RTE_ETH_EVENT_INTR_RMV);
 		else
-			DEBUG("port %u event type %d on not handled",
-			      dev->data->port_id, event.event_type);
+			DRV_LOG(DEBUG,
+				"port %u event type %d on not handled",
+				dev->data->port_id, event.event_type);
 		ibv_ack_async_event(&event);
 	}
 	if (ret & (1 << RTE_ETH_EVENT_INTR_LSC))
@@ -1130,8 +1144,10 @@ mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
 	flags = fcntl(priv->ctx->async_fd, F_GETFL);
 	ret = fcntl(priv->ctx->async_fd, F_SETFL, flags | O_NONBLOCK);
 	if (ret) {
-		INFO("port %u failed to change file descriptor async event"
-		     " queue", dev->data->port_id);
+		DRV_LOG(INFO,
+			"port %u failed to change file descriptor async event"
+			" queue",
+			dev->data->port_id);
 		dev->data->dev_conf.intr_conf.lsc = 0;
 		dev->data->dev_conf.intr_conf.rmv = 0;
 	}
@@ -1144,8 +1160,8 @@ mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
 	}
 	ret = mlx5_socket_init(dev);
 	if (ret)
-		ERROR("port %u cannot initialise socket: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot initialise socket: %s",
+			dev->data->port_id, strerror(rte_errno));
 	else if (priv->primary_socket) {
 		priv->intr_handle_socket.fd = priv->primary_socket;
 		priv->intr_handle_socket.type = RTE_INTR_HANDLE_EXT;
@@ -1203,20 +1219,24 @@ mlx5_select_tx_function(struct rte_eth_dev *dev)
 				dev->tx_pkt_burst = mlx5_tx_burst_raw_vec;
 			else
 				dev->tx_pkt_burst = mlx5_tx_burst_vec;
-			DEBUG("port %u selected enhanced MPW Tx vectorized"
-			      " function", dev->data->port_id);
+			DRV_LOG(DEBUG,
+				"port %u selected enhanced MPW Tx vectorized"
+				" function",
+				dev->data->port_id);
 		} else {
 			dev->tx_pkt_burst = mlx5_tx_burst_empw;
-			DEBUG("port %u selected enhanced MPW Tx function",
-			      dev->data->port_id);
+			DRV_LOG(DEBUG,
+				"port %u selected enhanced MPW Tx function",
+				dev->data->port_id);
 		}
 	} else if (priv->mps && priv->txq_inline) {
 		dev->tx_pkt_burst = mlx5_tx_burst_mpw_inline;
-		DEBUG("port %u selected MPW inline Tx function",
-		      dev->data->port_id);
+		DRV_LOG(DEBUG, "port %u selected MPW inline Tx function",
+			dev->data->port_id);
 	} else if (priv->mps) {
 		dev->tx_pkt_burst = mlx5_tx_burst_mpw;
-		DEBUG("port %u selected MPW Tx function", dev->data->port_id);
+		DRV_LOG(DEBUG, "port %u selected MPW Tx function",
+			dev->data->port_id);
 	}
 }
 
@@ -1232,8 +1252,8 @@ mlx5_select_rx_function(struct rte_eth_dev *dev)
 	assert(dev != NULL);
 	if (mlx5_check_vec_rx_support(dev) > 0) {
 		dev->rx_pkt_burst = mlx5_rx_burst_vec;
-		DEBUG("port %u selected Rx vectorized function",
-		      dev->data->port_id);
+		DRV_LOG(DEBUG, "port %u selected Rx vectorized function",
+			dev->data->port_id);
 	} else {
 		dev->rx_pkt_burst = mlx5_rx_burst;
 	}
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 326392798..9a3fcf43e 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -1838,11 +1838,11 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 			goto error;
 		}
 		++flows_n;
-		DEBUG("port %u %p type %d QP %p ibv_flow %p",
-		      dev->data->port_id,
-		      (void *)flow, i,
-		      (void *)flow->frxq[i].hrxq,
-		      (void *)flow->frxq[i].ibv_flow);
+		DRV_LOG(DEBUG, "port %u %p type %d QP %p ibv_flow %p",
+			dev->data->port_id,
+			(void *)flow, i,
+			(void *)flow->frxq[i].hrxq,
+			(void *)flow->frxq[i].ibv_flow);
 	}
 	if (!flows_n) {
 		rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_HANDLE,
@@ -1942,7 +1942,8 @@ mlx5_flow_list_create(struct rte_eth_dev *dev,
 	if (ret)
 		goto exit;
 	TAILQ_INSERT_TAIL(list, flow, next);
-	DEBUG("port %u flow created %p", dev->data->port_id, (void *)flow);
+	DRV_LOG(DEBUG, "port %u flow created %p", dev->data->port_id,
+		(void *)flow);
 	return flow;
 exit:
 	for (i = 0; i != hash_rxq_init_n; ++i) {
@@ -2061,7 +2062,8 @@ mlx5_flow_list_destroy(struct rte_eth_dev *dev, struct mlx5_flows *list,
 		flow->cs = NULL;
 	}
 	TAILQ_REMOVE(list, flow, next);
-	DEBUG("port %u flow destroyed %p", dev->data->port_id, (void *)flow);
+	DRV_LOG(DEBUG, "port %u flow destroyed %p", dev->data->port_id,
+		(void *)flow);
 	rte_free(flow);
 }
 
@@ -2103,15 +2105,16 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 	assert(priv->ctx);
 	fdq = rte_calloc(__func__, 1, sizeof(*fdq), 0);
 	if (!fdq) {
-		WARN("port %u cannot allocate memory for drop queue",
-		     dev->data->port_id);
+		DRV_LOG(WARNING,
+			"port %u cannot allocate memory for drop queue",
+			dev->data->port_id);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
 	fdq->cq = ibv_create_cq(priv->ctx, 1, NULL, NULL, 0);
 	if (!fdq->cq) {
-		WARN("port %u cannot allocate CQ for drop queue",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u cannot allocate CQ for drop queue",
+			dev->data->port_id);
 		rte_errno = errno;
 		goto error;
 	}
@@ -2124,8 +2127,8 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 			.cq = fdq->cq,
 			});
 	if (!fdq->wq) {
-		WARN("port %u cannot allocate WQ for drop queue",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u cannot allocate WQ for drop queue",
+			dev->data->port_id);
 		rte_errno = errno;
 		goto error;
 	}
@@ -2136,8 +2139,10 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 			.comp_mask = 0,
 			});
 	if (!fdq->ind_table) {
-		WARN("port %u cannot allocate indirection table for drop"
-		     " queue", dev->data->port_id);
+		DRV_LOG(WARNING,
+			"port %u cannot allocate indirection table for drop"
+			" queue",
+			dev->data->port_id);
 		rte_errno = errno;
 		goto error;
 	}
@@ -2159,8 +2164,8 @@ mlx5_flow_create_drop_queue(struct rte_eth_dev *dev)
 			.pd = priv->pd
 		});
 	if (!fdq->qp) {
-		WARN("port %u cannot allocate QP for drop queue",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u cannot allocate QP for drop queue",
+			dev->data->port_id);
 		rte_errno = errno;
 		goto error;
 	}
@@ -2231,8 +2236,8 @@ mlx5_flow_stop(struct rte_eth_dev *dev, struct mlx5_flows *list)
 			claim_zero(ibv_destroy_flow
 				   (flow->frxq[HASH_RXQ_ETH].ibv_flow));
 			flow->frxq[HASH_RXQ_ETH].ibv_flow = NULL;
-			DEBUG("port %u flow %p removed", dev->data->port_id,
-			      (void *)flow);
+			DRV_LOG(DEBUG, "port %u flow %p removed",
+				dev->data->port_id, (void *)flow);
 			/* Next flow. */
 			continue;
 		}
@@ -2264,8 +2269,8 @@ mlx5_flow_stop(struct rte_eth_dev *dev, struct mlx5_flows *list)
 			mlx5_hrxq_release(dev, flow->frxq[i].hrxq);
 			flow->frxq[i].hrxq = NULL;
 		}
-		DEBUG("port %u flow %p removed", dev->data->port_id,
-		      (void *)flow);
+		DRV_LOG(DEBUG, "port %u flow %p removed", dev->data->port_id,
+			(void *)flow);
 	}
 }
 
@@ -2295,14 +2300,14 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 				(priv->flow_drop_queue->qp,
 				 flow->frxq[HASH_RXQ_ETH].ibv_attr);
 			if (!flow->frxq[HASH_RXQ_ETH].ibv_flow) {
-				DEBUG("port %u flow %p cannot be applied",
-				      dev->data->port_id,
-				      (void *)flow);
+				DRV_LOG(DEBUG,
+					"port %u flow %p cannot be applied",
+					dev->data->port_id, (void *)flow);
 				rte_errno = EINVAL;
 				return -rte_errno;
 			}
-			DEBUG("port %u flow %p applied", dev->data->port_id,
-			      (void *)flow);
+			DRV_LOG(DEBUG, "port %u flow %p applied",
+				dev->data->port_id, (void *)flow);
 			/* Next flow. */
 			continue;
 		}
@@ -2324,8 +2329,9 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 					      (*flow->queues),
 					      flow->queues_n);
 			if (!flow->frxq[i].hrxq) {
-				DEBUG("port %u flow %p cannot be applied",
-				      dev->data->port_id, (void *)flow);
+				DRV_LOG(DEBUG,
+					"port %u flow %p cannot be applied",
+					dev->data->port_id, (void *)flow);
 				rte_errno = EINVAL;
 				return -rte_errno;
 			}
@@ -2334,13 +2340,14 @@ mlx5_flow_start(struct rte_eth_dev *dev, struct mlx5_flows *list)
 				ibv_create_flow(flow->frxq[i].hrxq->qp,
 						flow->frxq[i].ibv_attr);
 			if (!flow->frxq[i].ibv_flow) {
-				DEBUG("port %u flow %p cannot be applied",
-				      dev->data->port_id, (void *)flow);
+				DRV_LOG(DEBUG,
+					"port %u flow %p cannot be applied",
+					dev->data->port_id, (void *)flow);
 				rte_errno = EINVAL;
 				return -rte_errno;
 			}
-			DEBUG("port %u flow %p applied",
-			      dev->data->port_id, (void *)flow);
+			DRV_LOG(DEBUG, "port %u flow %p applied",
+				dev->data->port_id, (void *)flow);
 		}
 		if (!flow->mark)
 			continue;
@@ -2366,8 +2373,8 @@ mlx5_flow_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	TAILQ_FOREACH(flow, &priv->flows, next) {
-		DEBUG("port %u flow %p still referenced",
-		      dev->data->port_id, (void *)flow);
+		DRV_LOG(DEBUG, "port %u flow %p still referenced",
+			dev->data->port_id, (void *)flow);
 		++ret;
 	}
 	return ret;
@@ -2638,8 +2645,8 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 
 	/* Validate queue number. */
 	if (fdir_filter->action.rx_queue >= priv->rxqs_n) {
-		ERROR("port %u invalid queue number %d",
-		      dev->data->port_id, fdir_filter->action.rx_queue);
+		DRV_LOG(ERR, "port %u invalid queue number %d",
+			dev->data->port_id, fdir_filter->action.rx_queue);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
@@ -2662,9 +2669,9 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 		};
 		break;
 	default:
-		ERROR("port %u invalid behavior %d",
-		      dev->data->port_id,
-		      fdir_filter->action.behavior);
+		DRV_LOG(ERR, "port %u invalid behavior %d",
+			dev->data->port_id,
+			fdir_filter->action.behavior);
 		rte_errno = ENOTSUP;
 		return -rte_errno;
 	}
@@ -2800,8 +2807,8 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 		};
 		break;
 	default:
-		ERROR("port %u invalid flow type%d",
-		      dev->data->port_id, fdir_filter->input.flow_type);
+		DRV_LOG(ERR, "port %u invalid flow type%d",
+			dev->data->port_id, fdir_filter->input.flow_type);
 		rte_errno = ENOTSUP;
 		return -rte_errno;
 	}
@@ -2850,8 +2857,8 @@ mlx5_fdir_filter_add(struct rte_eth_dev *dev,
 				     attributes.items, attributes.actions,
 				     &error);
 	if (flow) {
-		DEBUG("port %u FDIR created %p", dev->data->port_id,
-		      (void *)flow);
+		DRV_LOG(DEBUG, "port %u FDIR created %p", dev->data->port_id,
+			(void *)flow);
 		return 0;
 	}
 	return -rte_errno;
@@ -3045,8 +3052,8 @@ mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
 		return 0;
 	if (fdir_mode != RTE_FDIR_MODE_PERFECT &&
 	    fdir_mode != RTE_FDIR_MODE_PERFECT_MAC_VLAN) {
-		ERROR("port %u flow director mode %d not supported",
-		      dev->data->port_id, fdir_mode);
+		DRV_LOG(ERR, "port %u flow director mode %d not supported",
+			dev->data->port_id, fdir_mode);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
@@ -3064,8 +3071,8 @@ mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
 		mlx5_fdir_info_get(dev, arg);
 		break;
 	default:
-		DEBUG("port %u unknown operation %u", dev->data->port_id,
-		      filter_op);
+		DRV_LOG(DEBUG, "port %u unknown operation %u",
+			dev->data->port_id, filter_op);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
@@ -3104,8 +3111,8 @@ mlx5_dev_filter_ctrl(struct rte_eth_dev *dev,
 	case RTE_ETH_FILTER_FDIR:
 		return mlx5_fdir_ctrl_func(dev, filter_op, arg);
 	default:
-		ERROR("port %u filter type (%d) not supported",
-		      dev->data->port_id, filter_type);
+		DRV_LOG(ERR, "port %u filter type (%d) not supported",
+			dev->data->port_id, filter_type);
 		rte_errno = ENOTSUP;
 		return -rte_errno;
 	}
diff --git a/drivers/net/mlx5/mlx5_mac.c b/drivers/net/mlx5/mlx5_mac.c
index 69fc06897..9de351426 100644
--- a/drivers/net/mlx5/mlx5_mac.c
+++ b/drivers/net/mlx5/mlx5_mac.c
@@ -101,8 +101,8 @@ mlx5_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index)
 		int ret = mlx5_traffic_restart(dev);
 
 		if (ret)
-			ERROR("port %u cannot remove mac address: %s",
-			      dev->data->port_id, strerror(rte_errno));
+			DRV_LOG(ERR, "port %u cannot remove mac address: %s",
+				dev->data->port_id, strerror(rte_errno));
 	}
 }
 
@@ -158,9 +158,11 @@ mlx5_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr)
 {
 	int ret;
 
-	DEBUG("port %u setting primary MAC address", dev->data->port_id);
+	DRV_LOG(DEBUG, "port %u setting primary MAC address",
+		dev->data->port_id);
+
 	ret = mlx5_mac_addr_add(dev, mac_addr, 0, 0);
 	if (ret)
-		ERROR("port %u cannot set mac address: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot set mac address: %s",
+			dev->data->port_id, strerror(rte_errno));
 }
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 42109a6a4..933bfe395 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -133,17 +133,18 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 
 	rte_spinlock_lock(&txq_ctrl->priv->mr_lock);
 	/* Add a new entry, register MR first. */
-	DEBUG("port %u discovered new memory pool \"%s\" (%p)",
-	      txq_ctrl->priv->dev->data->port_id, mp->name, (void *)mp);
+	DRV_LOG(DEBUG, "port %u discovered new memory pool \"%s\" (%p)",
+		txq_ctrl->priv->dev->data->port_id, mp->name, (void *)mp);
 	dev = txq_ctrl->priv->dev;
 	mr = mlx5_mr_get(dev, mp);
 	if (mr == NULL) {
 		if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
-			DEBUG("port %u using unregistered mempool 0x%p(%s) in "
-			      "secondary process, please create mempool before "
-			      " rte_eth_dev_start()",
-			      txq_ctrl->priv->dev->data->port_id,
-			     (void *)mp, mp->name);
+			DRV_LOG(DEBUG,
+				"port %u using unregistered mempool 0x%p(%s)"
+				" in secondary process, please create mempool"
+				" before rte_eth_dev_start()",
+				txq_ctrl->priv->dev->data->port_id,
+				(void *)mp, mp->name);
 			rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 			rte_errno = ENOTSUP;
 			return NULL;
@@ -151,17 +152,19 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 		mr = mlx5_mr_new(dev, mp);
 	}
 	if (unlikely(mr == NULL)) {
+		DRV_LOG(DEBUG,
+			"port %u unable to configure memory region,"
+			" ibv_reg_mr() failed.",
+			txq_ctrl->priv->dev->data->port_id);
 		rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
-		DEBUG("port %u unable to configure memory region, ibv_reg_mr()"
-		      " failed",
-		      txq_ctrl->priv->dev->data->port_id);
 		return NULL;
 	}
 	if (unlikely(idx == RTE_DIM(txq->mp2mr))) {
 		/* Table is full, remove oldest entry. */
-		DEBUG("port %u memroy region <-> memory pool table full, "
-		      " dropping oldest entry",
-		      txq_ctrl->priv->dev->data->port_id);
+		DRV_LOG(DEBUG,
+			"port %u memory region <-> memory pool table full, "
+			" dropping oldest entry",
+			txq_ctrl->priv->dev->data->port_id);
 		--idx;
 		mlx5_mr_release(txq->mp2mr[0]);
 		memmove(&txq->mp2mr[0], &txq->mp2mr[1],
@@ -169,9 +172,11 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 	}
 	/* Store the new entry. */
 	txq_ctrl->txq.mp2mr[idx] = mr;
-	DEBUG("port %u new memory region lkey for MP \"%s\" (%p): 0x%08" PRIu32,
-	      txq_ctrl->priv->dev->data->port_id, mp->name, (void *)mp,
-	      txq_ctrl->txq.mp2mr[idx]->lkey);
+	DRV_LOG(DEBUG,
+		"port %u new memory region lkey for MP \"%s\" (%p): 0x%08"
+		PRIu32,
+		txq_ctrl->priv->dev->data->port_id, mp->name, (void *)mp,
+		txq_ctrl->txq.mp2mr[idx]->lkey);
 	rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 	return mr;
 }
@@ -238,8 +243,8 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
 	}
 	mr = mlx5_mr_new(priv->dev, mp);
 	if (!mr)
-		ERROR("port %u cannot create memory region: %s",
-		      priv->dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot create memory region: %s",
+			priv->dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -266,21 +271,22 @@ mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 
 	mr = rte_zmalloc_socket(__func__, sizeof(*mr), 0, mp->socket_id);
 	if (!mr) {
-		DEBUG("port %u unable to configure memory region, ibv_reg_mr()"
-		      " failed",
-		      dev->data->port_id);
+		DRV_LOG(DEBUG,
+			"port %u unable to configure memory region,"
+			" ibv_reg_mr() failed.",
+			dev->data->port_id);
 		rte_errno = ENOMEM;
 		return NULL;
 	}
 	if (mlx5_check_mempool(mp, &start, &end) != 0) {
-		ERROR("port %u mempool %p: not virtually contiguous",
-		      dev->data->port_id, (void *)mp);
+		DRV_LOG(ERR, "port %u mempool %p: not virtually contiguous",
+			dev->data->port_id, (void *)mp);
 		rte_errno = ENOMEM;
 		return NULL;
 	}
-	DEBUG("port %u mempool %p area start=%p end=%p size=%zu",
-	      dev->data->port_id, (void *)mp, (void *)start, (void *)end,
-	      (size_t)(end - start));
+	DRV_LOG(DEBUG, "port %u mempool %p area start=%p end=%p size=%zu",
+		dev->data->port_id, (void *)mp, (void *)start, (void *)end,
+		(size_t)(end - start));
 	/* Save original addresses for exact MR lookup. */
 	mr->start = start;
 	mr->end = end;
@@ -295,10 +301,11 @@ mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 		if ((end > addr) && (end < addr + len))
 			end = RTE_ALIGN_CEIL(end, align);
 	}
-	DEBUG("port %u mempool %p using start=%p end=%p size=%zu for memory"
-	      " region",
-	      dev->data->port_id, (void *)mp, (void *)start, (void *)end,
-	      (size_t)(end - start));
+	DRV_LOG(DEBUG,
+		"port %u mempool %p using start=%p end=%p size=%zu for memory"
+		" region",
+		dev->data->port_id, (void *)mp, (void *)start, (void *)end,
+		(size_t)(end - start));
 	mr->mr = ibv_reg_mr(priv->pd, (void *)start, end - start,
 			    IBV_ACCESS_LOCAL_WRITE);
 	if (!mr->mr) {
@@ -308,8 +315,8 @@ mlx5_mr_new(struct rte_eth_dev *dev, struct rte_mempool *mp)
 	mr->mp = mp;
 	mr->lkey = rte_cpu_to_be_32(mr->mr->lkey);
 	rte_atomic32_inc(&mr->refcnt);
-	DEBUG("port %u new memory region %p refcnt: %d",
-	      dev->data->port_id, (void *)mr, rte_atomic32_read(&mr->refcnt));
+	DRV_LOG(DEBUG, "port %u new memory Region %p refcnt: %d",
+		dev->data->port_id, (void *)mr, rte_atomic32_read(&mr->refcnt));
 	LIST_INSERT_HEAD(&priv->mr, mr, next);
 	return mr;
 }
@@ -337,9 +344,9 @@ mlx5_mr_get(struct rte_eth_dev *dev, struct rte_mempool *mp)
 	LIST_FOREACH(mr, &priv->mr, next) {
 		if (mr->mp == mp) {
 			rte_atomic32_inc(&mr->refcnt);
-			DEBUG("port %u memory region %p refcnt: %d",
-			      dev->data->port_id, (void *)mr,
-			      rte_atomic32_read(&mr->refcnt));
+			DRV_LOG(DEBUG, "port %u memory region %p refcnt: %d",
+				dev->data->port_id, (void *)mr,
+				rte_atomic32_read(&mr->refcnt));
 			return mr;
 		}
 	}
@@ -359,8 +366,8 @@ int
 mlx5_mr_release(struct mlx5_mr *mr)
 {
 	assert(mr);
-	DEBUG("memory region %p refcnt: %d", (void *)mr,
-	      rte_atomic32_read(&mr->refcnt));
+	DRV_LOG(DEBUG, "memory region %p refcnt: %d", (void *)mr,
+		rte_atomic32_read(&mr->refcnt));
 	if (rte_atomic32_dec_and_test(&mr->refcnt)) {
 		claim_zero(ibv_dereg_mr(mr->mr));
 		LIST_REMOVE(mr, next);
@@ -387,8 +394,8 @@ mlx5_mr_verify(struct rte_eth_dev *dev)
 	struct mlx5_mr *mr;
 
 	LIST_FOREACH(mr, &priv->mr, next) {
-		DEBUG("port %u memory region %p still referenced",
-		      dev->data->port_id, (void *)mr);
+		DRV_LOG(DEBUG, "port %u memory region %p still referenced",
+			dev->data->port_id, (void *)mr);
 		++ret;
 	}
 	return ret;
diff --git a/drivers/net/mlx5/mlx5_rxmode.c b/drivers/net/mlx5/mlx5_rxmode.c
index f92ce8ef8..23eae7c12 100644
--- a/drivers/net/mlx5/mlx5_rxmode.c
+++ b/drivers/net/mlx5/mlx5_rxmode.c
@@ -65,8 +65,8 @@ mlx5_promiscuous_enable(struct rte_eth_dev *dev)
 	dev->data->promiscuous = 1;
 	ret = mlx5_traffic_restart(dev);
 	if (ret)
-		ERROR("port %u cannot enable promiscuous mode: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot enable promiscuous mode: %s",
+			dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -83,8 +83,8 @@ mlx5_promiscuous_disable(struct rte_eth_dev *dev)
 	dev->data->promiscuous = 0;
 	ret = mlx5_traffic_restart(dev);
 	if (ret)
-		ERROR("port %u cannot disable promiscuous mode: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot disable promiscuous mode: %s",
+			dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -101,8 +101,8 @@ mlx5_allmulticast_enable(struct rte_eth_dev *dev)
 	dev->data->all_multicast = 1;
 	ret = mlx5_traffic_restart(dev);
 	if (ret)
-		ERROR("port %u cannot enable allmulicast mode: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot enable allmulicast mode: %s",
+			dev->data->port_id, strerror(rte_errno));
 }
 
 /**
@@ -119,6 +119,6 @@ mlx5_allmulticast_disable(struct rte_eth_dev *dev)
 	dev->data->all_multicast = 0;
 	ret = mlx5_traffic_restart(dev);
 	if (ret)
-		ERROR("port %u cannot disable allmulicast mode: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot disable allmulicast mode: %s",
+			dev->data->port_id, strerror(rte_errno));
 }
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index c97844f63..1b0a95e0a 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -104,8 +104,8 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 
 		buf = rte_pktmbuf_alloc(rxq_ctrl->rxq.mp);
 		if (buf == NULL) {
-			ERROR("port %u empty mbuf pool",
-			      rxq_ctrl->priv->dev->data->port_id);
+			DRV_LOG(ERR, "port %u empty mbuf pool",
+				rxq_ctrl->priv->dev->data->port_id);
 			rte_errno = ENOMEM;
 			goto error;
 		}
@@ -146,9 +146,11 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 		for (j = 0; j < MLX5_VPMD_DESCS_PER_LOOP; ++j)
 			(*rxq->elts)[elts_n + j] = &rxq->fake_mbuf;
 	}
-	DEBUG("port %u Rx queue %u allocated and configured %u segments"
-	      " (max %u packets)", rxq_ctrl->priv->dev->data->port_id,
-	      rxq_ctrl->idx, elts_n, elts_n / (1 << rxq_ctrl->rxq.sges_n));
+	DRV_LOG(DEBUG,
+		"port %u Rx queue %u allocated and configured %u segments"
+		" (max %u packets)",
+		rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx, elts_n,
+		elts_n / (1 << rxq_ctrl->rxq.sges_n));
 	return 0;
 error:
 	err = rte_errno; /* Save rte_errno before cleanup. */
@@ -158,8 +160,8 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 			rte_pktmbuf_free_seg((*rxq_ctrl->rxq.elts)[i]);
 		(*rxq_ctrl->rxq.elts)[i] = NULL;
 	}
-	DEBUG("port %u Rx queue %u failed, freed everything",
-	      rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
+	DRV_LOG(DEBUG, "port %u Rx queue %u failed, freed everything",
+		rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
 	rte_errno = err; /* Restore rte_errno. */
 	return -rte_errno;
 }
@@ -179,8 +181,8 @@ rxq_free_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 	uint16_t used = q_n - (rxq->rq_ci - rxq->rq_pi);
 	uint16_t i;
 
-	DEBUG("port %u Rx queue %u freeing WRs",
-	      rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
+	DRV_LOG(DEBUG, "port %u Rx queue %u freeing WRs",
+		rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
 	if (rxq->elts == NULL)
 		return;
 	/**
@@ -210,8 +212,8 @@ rxq_free_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 void
 mlx5_rxq_cleanup(struct mlx5_rxq_ctrl *rxq_ctrl)
 {
-	DEBUG("port %u cleaning up Rx queue %u",
-	      rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
+	DRV_LOG(DEBUG, "port %u cleaning up Rx queue %u",
+		rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
 	if (rxq_ctrl->ibv)
 		mlx5_rxq_ibv_release(rxq_ctrl->ibv);
 	memset(rxq_ctrl, 0, sizeof(*rxq_ctrl));
@@ -248,33 +250,35 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 
 	if (!rte_is_power_of_2(desc)) {
 		desc = 1 << log2above(desc);
-		WARN("port %u increased number of descriptors in Rx queue %u"
-		     " to the next power of two (%d)",
-		     dev->data->port_id, idx, desc);
+		DRV_LOG(WARNING,
+			"port %u increased number of descriptors in Rx queue %u"
+			" to the next power of two (%d)",
+			dev->data->port_id, idx, desc);
 	}
-	DEBUG("port %u configuring Rx queue %u for %u descriptors",
-	      dev->data->port_id, idx, desc);
+	DRV_LOG(DEBUG, "port %u configuring Rx queue %u for %u descriptors",
+		dev->data->port_id, idx, desc);
 	if (idx >= priv->rxqs_n) {
-		ERROR("port %u Rx queue index out of range (%u >= %u)",
-		      dev->data->port_id, idx, priv->rxqs_n);
+		DRV_LOG(ERR, "port %u Rx queue index out of range (%u >= %u)",
+			dev->data->port_id, idx, priv->rxqs_n);
 		rte_errno = EOVERFLOW;
 		return -rte_errno;
 	}
 	if (!mlx5_rxq_releasable(dev, idx)) {
-		ERROR("port %u unable to release queue index %u",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u unable to release queue index %u",
+			dev->data->port_id, idx);
 		rte_errno = EBUSY;
 		return -rte_errno;
 	}
 	mlx5_rxq_release(dev, idx);
 	rxq_ctrl = mlx5_rxq_new(dev, idx, desc, socket, mp);
 	if (!rxq_ctrl) {
-		ERROR("port %u unable to allocate queue index %u",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u unable to allocate queue index %u",
+			dev->data->port_id, idx);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
-	DEBUG("port %u adding Rx queue %u to list", dev->data->port_id, idx);
+	DRV_LOG(DEBUG, "port %u adding Rx queue %u to list",
+		dev->data->port_id, idx);
 	(*priv->rxqs)[idx] = &rxq_ctrl->rxq;
 	return 0;
 }
@@ -327,9 +331,10 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 	mlx5_rx_intr_vec_disable(dev);
 	intr_handle->intr_vec = malloc(n * sizeof(intr_handle->intr_vec[0]));
 	if (intr_handle->intr_vec == NULL) {
-		ERROR("port %u failed to allocate memory for interrupt vector,"
-		      " Rx interrupts will not be supported",
-		      dev->data->port_id);
+		DRV_LOG(ERR,
+			"port %u failed to allocate memory for interrupt"
+			" vector, Rx interrupts will not be supported",
+			dev->data->port_id);
 		rte_errno = ENOMEM;
 		return -rte_errno;
 	}
@@ -350,9 +355,11 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 			continue;
 		}
 		if (count >= RTE_MAX_RXTX_INTR_VEC_ID) {
-			ERROR("port %u too many Rx queues for interrupt vector"
-			      " size (%d), Rx interrupts cannot be enabled",
-			      dev->data->port_id, RTE_MAX_RXTX_INTR_VEC_ID);
+			DRV_LOG(ERR,
+				"port %u too many Rx queues for interrupt"
+				" vector size (%d), Rx interrupts cannot be"
+				" enabled",
+				dev->data->port_id, RTE_MAX_RXTX_INTR_VEC_ID);
 			mlx5_rx_intr_vec_disable(dev);
 			rte_errno = ENOMEM;
 			return -rte_errno;
@@ -362,9 +369,11 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 		rc = fcntl(fd, F_SETFL, flags | O_NONBLOCK);
 		if (rc < 0) {
 			rte_errno = errno;
-			ERROR("port %u failed to make Rx interrupt file"
-			      " descriptor %d non-blocking for queue index %d",
-			      dev->data->port_id, fd, i);
+			DRV_LOG(ERR,
+				"port %u failed to make Rx interrupt file"
+				" descriptor %d non-blocking for queue index"
+				" %d",
+				dev->data->port_id, fd, i);
 			mlx5_rx_intr_vec_disable(dev);
 			return -rte_errno;
 		}
@@ -531,8 +540,8 @@ mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	ret = rte_errno; /* Save rte_errno before cleanup. */
 	if (rxq_ibv)
 		mlx5_rxq_ibv_release(rxq_ibv);
-	WARN("port %u unable to disable interrupt on Rx queue %d",
-	     dev->data->port_id, rx_queue_id);
+	DRV_LOG(WARNING, "port %u unable to disable interrupt on Rx queue %d",
+		dev->data->port_id, rx_queue_id);
 	rte_errno = ret; /* Restore rte_errno. */
 	return -rte_errno;
 }
@@ -579,8 +588,9 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	tmpl = rte_calloc_socket(__func__, 1, sizeof(*tmpl), 0,
 				 rxq_ctrl->socket);
 	if (!tmpl) {
-		ERROR("port %u Rx queue %u cannot allocate verbs resources",
-		      dev->data->port_id, rxq_ctrl->idx);
+		DRV_LOG(ERR,
+			"port %u Rx queue %u cannot allocate verbs resources",
+			dev->data->port_id, rxq_ctrl->idx);
 		rte_errno = ENOMEM;
 		goto error;
 	}
@@ -590,16 +600,16 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (!tmpl->mr) {
 		tmpl->mr = mlx5_mr_new(dev, rxq_data->mp);
 		if (!tmpl->mr) {
-			ERROR("port %u: memory region creation failure",
-			      dev->data->port_id);
+			DRV_LOG(ERR, "port %u: memeroy region creation failure",
+				dev->data->port_id);
 			goto error;
 		}
 	}
 	if (rxq_ctrl->irq) {
 		tmpl->channel = ibv_create_comp_channel(priv->ctx);
 		if (!tmpl->channel) {
-			ERROR("port %u: comp channel creation failure",
-			      dev->data->port_id);
+			DRV_LOG(ERR, "port %u: comp channel creation failure",
+				dev->data->port_id);
 			rte_errno = ENOMEM;
 			goto error;
 		}
@@ -623,21 +633,23 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		if (mlx5_rxq_check_vec_support(rxq_data) < 0)
 			attr.cq.ibv.cqe *= 2;
 	} else if (priv->cqe_comp && rxq_data->hw_timestamp) {
-		DEBUG("port %u Rx CQE compression is disabled for HW timestamp",
-		      dev->data->port_id);
+		DRV_LOG(DEBUG,
+			"port %u Rx CQE compression is disabled for HW"
+			" timestamp",
+			dev->data->port_id);
 	}
 	tmpl->cq = ibv_cq_ex_to_cq(mlx5dv_create_cq(priv->ctx, &attr.cq.ibv,
 						    &attr.cq.mlx5));
 	if (tmpl->cq == NULL) {
-		ERROR("port %u Rx queue %u CQ creation failure",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u Rx queue %u CQ creation failure",
+			dev->data->port_id, idx);
 		rte_errno = ENOMEM;
 		goto error;
 	}
-	DEBUG("port %u priv->device_attr.max_qp_wr is %d",
-	      dev->data->port_id, priv->device_attr.orig_attr.max_qp_wr);
-	DEBUG("port %u priv->device_attr.max_sge is %d",
-	      dev->data->port_id, priv->device_attr.orig_attr.max_sge);
+	DRV_LOG(DEBUG, "port %u priv->device_attr.max_qp_wr is %d",
+		dev->data->port_id, priv->device_attr.orig_attr.max_qp_wr);
+	DRV_LOG(DEBUG, "port %u priv->device_attr.max_sge is %d",
+		dev->data->port_id, priv->device_attr.orig_attr.max_sge);
 	attr.wq = (struct ibv_wq_init_attr){
 		.wq_context = NULL, /* Could be useful in the future. */
 		.wq_type = IBV_WQT_RQ,
@@ -667,8 +679,8 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 #endif
 	tmpl->wq = ibv_create_wq(priv->ctx, &attr.wq);
 	if (tmpl->wq == NULL) {
-		ERROR("port %u Rx queue %u WQ creation failure",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u Rx queue %u WQ creation failure",
+			dev->data->port_id, idx);
 		rte_errno = ENOMEM;
 		goto error;
 	}
@@ -679,12 +691,13 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (((int)attr.wq.max_wr !=
 	     ((1 << rxq_data->elts_n) >> rxq_data->sges_n)) ||
 	    ((int)attr.wq.max_sge != (1 << rxq_data->sges_n))) {
-		ERROR("port %u Rx queue %u requested %u*%u but got %u*%u"
-		      " WRs*SGEs",
-		      dev->data->port_id, idx,
-		      ((1 << rxq_data->elts_n) >> rxq_data->sges_n),
-		      (1 << rxq_data->sges_n),
-		      attr.wq.max_wr, attr.wq.max_sge);
+		DRV_LOG(ERR,
+			"port %u Rx queue %u requested %u*%u but got %u*%u"
+			" WRs*SGEs",
+			dev->data->port_id, idx,
+			((1 << rxq_data->elts_n) >> rxq_data->sges_n),
+			(1 << rxq_data->sges_n),
+			attr.wq.max_wr, attr.wq.max_sge);
 		rte_errno = EINVAL;
 		goto error;
 	}
@@ -695,8 +708,9 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	};
 	ret = ibv_modify_wq(tmpl->wq, &mod);
 	if (ret) {
-		ERROR("port %u Rx queue %u WQ state to IBV_WQS_RDY failed",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR,
+			"port %u Rx queue %u WQ state to IBV_WQS_RDY failed",
+			dev->data->port_id, idx);
 		rte_errno = ret;
 		goto error;
 	}
@@ -710,9 +724,10 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		goto error;
 	}
 	if (cq_info.cqe_size != RTE_CACHE_LINE_SIZE) {
-		ERROR("port %u wrong MLX5_CQE_SIZE environment variable value: "
-		      "it should be set to %u", dev->data->port_id,
-		      RTE_CACHE_LINE_SIZE);
+		DRV_LOG(ERR,
+			"port %u wrong MLX5_CQE_SIZE environment variable"
+			" value: it should be set to %u",
+			dev->data->port_id, RTE_CACHE_LINE_SIZE);
 		rte_errno = EINVAL;
 		goto error;
 	}
@@ -749,11 +764,11 @@ mlx5_rxq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	rxq_data->rq_ci = (1 << rxq_data->elts_n) >> rxq_data->sges_n;
 	rte_wmb();
 	*rxq_data->rq_db = rte_cpu_to_be_32(rxq_data->rq_ci);
-	DEBUG("port %u rxq %u updated with %p", dev->data->port_id, idx,
-	      (void *)&tmpl);
+	DRV_LOG(DEBUG, "port %u rxq %u updated with %p", dev->data->port_id,
+		idx, (void *)&tmpl);
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("port %u Verbs Rx queue %u: refcnt %d", dev->data->port_id, idx,
-	      rte_atomic32_read(&tmpl->refcnt));
+	DRV_LOG(DEBUG, "port %u Verbs Rx queue %u: refcnt %d",
+		dev->data->port_id, idx, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->rxqsibv, tmpl, next);
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
 	return tmpl;
@@ -798,9 +813,9 @@ mlx5_rxq_ibv_get(struct rte_eth_dev *dev, uint16_t idx)
 	if (rxq_ctrl->ibv) {
 		mlx5_mr_get(dev, rxq_data->mp);
 		rte_atomic32_inc(&rxq_ctrl->ibv->refcnt);
-		DEBUG("port %u Verbs Rx queue %u: refcnt %d",
-		      dev->data->port_id, rxq_ctrl->idx,
-		      rte_atomic32_read(&rxq_ctrl->ibv->refcnt));
+		DRV_LOG(DEBUG, "port %u Verbs Rx queue %u: refcnt %d",
+			dev->data->port_id, rxq_ctrl->idx,
+			rte_atomic32_read(&rxq_ctrl->ibv->refcnt));
 	}
 	return rxq_ctrl->ibv;
 }
@@ -826,9 +841,9 @@ mlx5_rxq_ibv_release(struct mlx5_rxq_ibv *rxq_ibv)
 	ret = mlx5_mr_release(rxq_ibv->mr);
 	if (!ret)
 		rxq_ibv->mr = NULL;
-	DEBUG("port %u Verbs Rx queue %u: refcnt %d",
-	      rxq_ibv->rxq_ctrl->priv->dev->data->port_id,
-	      rxq_ibv->rxq_ctrl->idx, rte_atomic32_read(&rxq_ibv->refcnt));
+	DRV_LOG(DEBUG, "port %u Verbs Rx queue %u: refcnt %d",
+		rxq_ibv->rxq_ctrl->priv->dev->data->port_id,
+		rxq_ibv->rxq_ctrl->idx, rte_atomic32_read(&rxq_ibv->refcnt));
 	if (rte_atomic32_dec_and_test(&rxq_ibv->refcnt)) {
 		rxq_free_elts(rxq_ibv->rxq_ctrl);
 		claim_zero(ibv_destroy_wq(rxq_ibv->wq));
@@ -859,8 +874,8 @@ mlx5_rxq_ibv_verify(struct rte_eth_dev *dev)
 	struct mlx5_rxq_ibv *rxq_ibv;
 
 	LIST_FOREACH(rxq_ibv, &priv->rxqsibv, next) {
-		DEBUG("port %u Verbs Rx queue %u still referenced",
-		      dev->data->port_id, rxq_ibv->rxq_ctrl->idx);
+		DRV_LOG(DEBUG, "port %u Verbs Rx queue %u still referenced",
+			dev->data->port_id, rxq_ibv->rxq_ctrl->idx);
 		++ret;
 	}
 	return ret;
@@ -936,30 +951,33 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		size = mb_len * (1 << tmpl->rxq.sges_n);
 		size -= RTE_PKTMBUF_HEADROOM;
 		if (size < dev->data->dev_conf.rxmode.max_rx_pkt_len) {
-			ERROR("port %u too many SGEs (%u) needed to handle"
-			      " requested maximum packet size %u",
-			      dev->data->port_id,
-			      1 << sges_n,
-			      dev->data->dev_conf.rxmode.max_rx_pkt_len);
+			DRV_LOG(ERR,
+				"port %u too many SGEs (%u) needed to handle"
+				" requested maximum packet size %u",
+				dev->data->port_id,
+				1 << sges_n,
+				dev->data->dev_conf.rxmode.max_rx_pkt_len);
 			rte_errno = EOVERFLOW;
 			goto error;
 		}
 	} else {
-		WARN("port %u the requested maximum Rx packet size (%u) is"
-		     " larger than a single mbuf (%u) and scattered"
-		     " mode has not been requested",
-		     dev->data->port_id,
-		     dev->data->dev_conf.rxmode.max_rx_pkt_len,
-		     mb_len - RTE_PKTMBUF_HEADROOM);
+		DRV_LOG(WARNING,
+			"port %u the requested maximum Rx packet size (%u) is"
+			" larger than a single mbuf (%u) and scattered mode has"
+			" not been requested",
+			dev->data->port_id,
+			dev->data->dev_conf.rxmode.max_rx_pkt_len,
+			mb_len - RTE_PKTMBUF_HEADROOM);
 	}
-	DEBUG("port %u maximum number of segments per packet: %u",
-	      dev->data->port_id, 1 << tmpl->rxq.sges_n);
+	DRV_LOG(DEBUG, "port %u maximum number of segments per packet: %u",
+		dev->data->port_id, 1 << tmpl->rxq.sges_n);
 	if (desc % (1 << tmpl->rxq.sges_n)) {
-		ERROR("port %u number of Rx queue descriptors (%u) is not a"
-		      " multiple of SGEs per packet (%u)",
-		      dev->data->port_id,
-		      desc,
-		      1 << tmpl->rxq.sges_n);
+		DRV_LOG(ERR,
+			"port %u number of Rx queue descriptors (%u) is not a"
+			" multiple of SGEs per packet (%u)",
+			dev->data->port_id,
+			desc,
+			1 << tmpl->rxq.sges_n);
 		rte_errno = EINVAL;
 		goto error;
 	}
@@ -980,17 +998,19 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	} else if (priv->hw_fcs_strip) {
 		tmpl->rxq.crc_present = 1;
 	} else {
-		WARN("port %u CRC stripping has been disabled but will still"
-		     " be performed by hardware, make sure MLNX_OFED and"
-		     " firmware are up to date",
-		     dev->data->port_id);
+		DRV_LOG(WARNING,
+			"port %u CRC stripping has been disabled but will"
+			" still be performed by hardware, make sure MLNX_OFED"
+			" and firmware are up to date",
+			dev->data->port_id);
 		tmpl->rxq.crc_present = 0;
 	}
-	DEBUG("port %u CRC stripping is %s, %u bytes will be subtracted from"
-	      " incoming frames to hide it",
-	      dev->data->port_id,
-	      tmpl->rxq.crc_present ? "disabled" : "enabled",
-	      tmpl->rxq.crc_present << 2);
+	DRV_LOG(DEBUG,
+		"port %u CRC stripping is %s, %u bytes will be subtracted from"
+		" incoming frames to hide it",
+		dev->data->port_id,
+		tmpl->rxq.crc_present ? "disabled" : "enabled",
+		tmpl->rxq.crc_present << 2);
 	/* Save port ID. */
 	tmpl->rxq.rss_hash = priv->rxqs_n > 1;
 	tmpl->rxq.port_id = dev->data->port_id;
@@ -1002,8 +1022,8 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		(struct rte_mbuf *(*)[1 << tmpl->rxq.elts_n])(tmpl + 1);
 	tmpl->idx = idx;
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("port %u Rx queue %u: refcnt %d", dev->data->port_id,
-	      idx, rte_atomic32_read(&tmpl->refcnt));
+	DRV_LOG(DEBUG, "port %u Rx queue %u: refcnt %d", dev->data->port_id,
+		idx, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->rxqsctrl, tmpl, next);
 	return tmpl;
 error:
@@ -1034,8 +1054,9 @@ mlx5_rxq_get(struct rte_eth_dev *dev, uint16_t idx)
 					rxq);
 		mlx5_rxq_ibv_get(dev, idx);
 		rte_atomic32_inc(&rxq_ctrl->refcnt);
-		DEBUG("port %u Rx queue %u: refcnt %d", dev->data->port_id,
-		      rxq_ctrl->idx, rte_atomic32_read(&rxq_ctrl->refcnt));
+		DRV_LOG(DEBUG, "port %u Rx queue %u: refcnt %d",
+			dev->data->port_id, rxq_ctrl->idx,
+			rte_atomic32_read(&rxq_ctrl->refcnt));
 	}
 	return rxq_ctrl;
 }
@@ -1063,8 +1084,8 @@ mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx)
 	assert(rxq_ctrl->priv);
 	if (rxq_ctrl->ibv && !mlx5_rxq_ibv_release(rxq_ctrl->ibv))
 		rxq_ctrl->ibv = NULL;
-	DEBUG("port %u Rx queue %u: refcnt %d", dev->data->port_id,
-	      rxq_ctrl->idx, rte_atomic32_read(&rxq_ctrl->refcnt));
+	DRV_LOG(DEBUG, "port %u Rx queue %u: refcnt %d", dev->data->port_id,
+		rxq_ctrl->idx, rte_atomic32_read(&rxq_ctrl->refcnt));
 	if (rte_atomic32_dec_and_test(&rxq_ctrl->refcnt)) {
 		LIST_REMOVE(rxq_ctrl, next);
 		rte_free(rxq_ctrl);
@@ -1117,8 +1138,8 @@ mlx5_rxq_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	LIST_FOREACH(rxq_ctrl, &priv->rxqsctrl, next) {
-		DEBUG("port %u Rx queue %u still referenced",
-		      dev->data->port_id, rxq_ctrl->idx);
+		DRV_LOG(DEBUG, "port %u Rx Queue %u still referenced",
+			dev->data->port_id, rxq_ctrl->idx);
 		++ret;
 	}
 	return ret;
@@ -1181,12 +1202,14 @@ mlx5_ind_table_ibv_new(struct rte_eth_dev *dev, uint16_t queues[],
 	}
 	rte_atomic32_inc(&ind_tbl->refcnt);
 	LIST_INSERT_HEAD(&priv->ind_tbls, ind_tbl, next);
-	DEBUG("port %u indirection table %p: refcnt %d", dev->data->port_id,
-	      (void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
+	DRV_LOG(DEBUG, "port %u indirection table %p: refcnt %d",
+		dev->data->port_id, (void *)ind_tbl,
+		rte_atomic32_read(&ind_tbl->refcnt));
 	return ind_tbl;
 error:
 	rte_free(ind_tbl);
-	DEBUG("port %u cannot create indirection table", dev->data->port_id);
+	DRV_LOG(DEBUG, "port %u cannot create indirection table",
+		dev->data->port_id);
 	return NULL;
 }
 
@@ -1221,9 +1244,9 @@ mlx5_ind_table_ibv_get(struct rte_eth_dev *dev, uint16_t queues[],
 		unsigned int i;
 
 		rte_atomic32_inc(&ind_tbl->refcnt);
-		DEBUG("port %u indirection table %p: refcnt %d",
-		      dev->data->port_id, (void *)ind_tbl,
-		      rte_atomic32_read(&ind_tbl->refcnt));
+		DRV_LOG(DEBUG, "port %u indirection table %p: refcnt %d",
+			dev->data->port_id, (void *)ind_tbl,
+			rte_atomic32_read(&ind_tbl->refcnt));
 		for (i = 0; i != ind_tbl->queues_n; ++i)
 			mlx5_rxq_get(dev, ind_tbl->queues[i]);
 	}
@@ -1247,9 +1270,9 @@ mlx5_ind_table_ibv_release(struct rte_eth_dev *dev,
 {
 	unsigned int i;
 
-	DEBUG("port %u indirection table %p: refcnt %d",
-	      ((struct priv *)dev->data->dev_private)->port,
-	      (void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
+	DRV_LOG(DEBUG, "port %u indirection table %p: refcnt %d",
+		((struct priv *)dev->data->dev_private)->port,
+		(void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
 	if (rte_atomic32_dec_and_test(&ind_tbl->refcnt))
 		claim_zero(ibv_destroy_rwq_ind_table(ind_tbl->ind_table));
 	for (i = 0; i != ind_tbl->queues_n; ++i)
@@ -1279,8 +1302,9 @@ mlx5_ind_table_ibv_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	LIST_FOREACH(ind_tbl, &priv->ind_tbls, next) {
-		DEBUG("port %u Verbs indirection table %p still referenced",
-		      dev->data->port_id, (void *)ind_tbl);
+		DRV_LOG(DEBUG,
+			"port %u Verbs indirection table %p still referenced",
+			dev->data->port_id, (void *)ind_tbl);
 		++ret;
 	}
 	return ret;
@@ -1355,8 +1379,9 @@ mlx5_hrxq_new(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 	memcpy(hrxq->rss_key, rss_key, rss_key_len);
 	rte_atomic32_inc(&hrxq->refcnt);
 	LIST_INSERT_HEAD(&priv->hrxqs, hrxq, next);
-	DEBUG("port %u hash Rx queue %p: refcnt %d", dev->data->port_id,
-	      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
+	DRV_LOG(DEBUG, "port %u hash Rx queue %p: refcnt %d",
+		dev->data->port_id, (void *)hrxq,
+		rte_atomic32_read(&hrxq->refcnt));
 	return hrxq;
 error:
 	err = rte_errno; /* Save rte_errno before cleanup. */
@@ -1408,8 +1433,9 @@ mlx5_hrxq_get(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 			continue;
 		}
 		rte_atomic32_inc(&hrxq->refcnt);
-		DEBUG("port %u hash Rx queue %p: refcnt %d", dev->data->port_id,
-		      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
+		DRV_LOG(DEBUG, "port %u hash Rx queue %p: refcnt %d",
+			dev->data->port_id, (void *)hrxq,
+			rte_atomic32_read(&hrxq->refcnt));
 		return hrxq;
 	}
 	return NULL;
@@ -1429,9 +1455,9 @@ mlx5_hrxq_get(struct rte_eth_dev *dev, uint8_t *rss_key, uint8_t rss_key_len,
 int
 mlx5_hrxq_release(struct rte_eth_dev *dev, struct mlx5_hrxq *hrxq)
 {
-	DEBUG("port %u hash Rx queue %p: refcnt %d",
-	      ((struct priv *)dev->data->dev_private)->port,
-	      (void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
+	DRV_LOG(DEBUG, "port %u hash Rx queue %p: refcnt %d",
+		((struct priv *)dev->data->dev_private)->port,
+		(void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
 	if (rte_atomic32_dec_and_test(&hrxq->refcnt)) {
 		claim_zero(ibv_destroy_qp(hrxq->qp));
 		mlx5_ind_table_ibv_release(dev, hrxq->ind_table);
@@ -1460,8 +1486,9 @@ mlx5_hrxq_ibv_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	LIST_FOREACH(hrxq, &priv->hrxqs, next) {
-		DEBUG("port %u Verbs hash Rx queue %p still referenced",
-		      dev->data->port_id, (void *)hrxq);
+		DRV_LOG(DEBUG,
+			"port %u Verbs hash Rx queue %p still referenced",
+			dev->data->port_id, (void *)hrxq);
 		++ret;
 	}
 	return ret;
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index 47a8729a8..29019f792 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -400,9 +400,10 @@ check_cqe(volatile struct mlx5_cqe *cqe,
 		    (syndrome == MLX5_CQE_SYNDROME_REMOTE_ABORTED_ERR))
 			return 0;
 		if (!check_cqe_seen(cqe)) {
-			ERROR("unexpected CQE error %u (0x%02x)"
-			      " syndrome 0x%02x",
-			      op_code, op_code, syndrome);
+			DRV_LOG(ERR,
+				"unexpected CQE error %u (0x%02x) syndrome"
+				" 0x%02x",
+				op_code, op_code, syndrome);
 			rte_hexdump(stderr, "MLX5 Error CQE:",
 				    (const void *)((uintptr_t)err_cqe),
 				    sizeof(*err_cqe));
@@ -411,8 +412,8 @@ check_cqe(volatile struct mlx5_cqe *cqe,
 	} else if ((op_code != MLX5_CQE_RESP_SEND) &&
 		   (op_code != MLX5_CQE_REQ)) {
 		if (!check_cqe_seen(cqe)) {
-			ERROR("unexpected CQE opcode %u (0x%02x)",
-			      op_code, op_code);
+			DRV_LOG(ERR, "unexpected CQE opcode %u (0x%02x)",
+				op_code, op_code);
 			rte_hexdump(stderr, "MLX5 CQE:",
 				    (const void *)((uintptr_t)cqe),
 				    sizeof(*cqe));
@@ -472,7 +473,7 @@ mlx5_tx_complete(struct mlx5_txq_data *txq)
 	if ((MLX5_CQE_OPCODE(cqe->op_own) == MLX5_CQE_RESP_ERR) ||
 	    (MLX5_CQE_OPCODE(cqe->op_own) == MLX5_CQE_REQ_ERR)) {
 		if (!check_cqe_seen(cqe)) {
-			ERROR("unexpected error CQE, Tx stopped");
+			DRV_LOG(ERR, "unexpected error CQE, Tx stopped");
 			rte_hexdump(stderr, "MLX5 TXQ:",
 				    (const void *)((uintptr_t)txq->wqes),
 				    ((1 << txq->wqe_n) *
@@ -589,8 +590,8 @@ mlx5_tx_mb2mr(struct mlx5_txq_data *txq, struct rte_mbuf *mb)
 	} else {
 		struct rte_mempool *mp = mlx5_tx_mb2mp(mb);
 
-		WARN("failed to register mempool 0x%p(%s)",
-		      (void *)mp, mp->name);
+		DRV_LOG(WARNING, "failed to register mempool 0x%p(%s)",
+			(void *)mp, mp->name);
 	}
 	return (uint32_t)-1;
 }
diff --git a/drivers/net/mlx5/mlx5_socket.c b/drivers/net/mlx5/mlx5_socket.c
index f4a5c835e..bdbd390d1 100644
--- a/drivers/net/mlx5/mlx5_socket.c
+++ b/drivers/net/mlx5/mlx5_socket.c
@@ -68,8 +68,8 @@ mlx5_socket_init(struct rte_eth_dev *dev)
 	ret = socket(AF_UNIX, SOCK_STREAM, 0);
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("port %u secondary process not supported: %s",
-		     dev->data->port_id, strerror(errno));
+		DRV_LOG(WARNING, "port %u secondary process not supported: %s",
+			dev->data->port_id, strerror(errno));
 		goto error;
 	}
 	priv->primary_socket = ret;
@@ -90,15 +90,17 @@ mlx5_socket_init(struct rte_eth_dev *dev)
 		   sizeof(sun));
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("port %u cannot bind socket, secondary process not"
-		     " supported: %s", dev->data->port_id, strerror(errno));
+		DRV_LOG(WARNING,
+			"port %u cannot bind socket, secondary process not"
+			" supported: %s",
+			dev->data->port_id, strerror(errno));
 		goto close;
 	}
 	ret = listen(priv->primary_socket, 0);
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("port %u secondary process not supported: %s",
-		     dev->data->port_id, strerror(errno));
+		DRV_LOG(WARNING, "port %u secondary process not supported: %s",
+			dev->data->port_id, strerror(errno));
 		goto close;
 	}
 	return 0;
@@ -157,29 +159,29 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	/* Accept the connection from the client. */
 	conn_sock = accept(priv->primary_socket, NULL, NULL);
 	if (conn_sock < 0) {
-		WARN("port %u connection failed: %s", dev->data->port_id,
-		     strerror(errno));
+		DRV_LOG(WARNING, "port %u connection failed: %s",
+			dev->data->port_id, strerror(errno));
 		return;
 	}
 	ret = setsockopt(conn_sock, SOL_SOCKET, SO_PASSCRED, &(int){1},
 					 sizeof(int));
 	if (ret < 0) {
 		ret = errno;
-		WARN("port %u cannot change socket options: %s",
-		     dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(WARNING, "port %u cannot change socket options: %s",
+			dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
 	ret = recvmsg(conn_sock, &msg, MSG_WAITALL);
 	if (ret < 0) {
 		ret = errno;
-		WARN("port %u received an empty message: %s",
-		     dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(WARNING, "port %u received an empty message: %s",
+			dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
 	/* Expect to receive credentials only. */
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
-		WARN("port %u no message", dev->data->port_id);
+		DRV_LOG(WARNING, "port %u no message", dev->data->port_id);
 		goto error;
 	}
 	if ((cmsg->cmsg_type == SCM_CREDENTIALS) &&
@@ -189,13 +191,15 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	}
 	cmsg = CMSG_NXTHDR(&msg, cmsg);
 	if (cmsg != NULL) {
-		WARN("port %u message wrongly formatted", dev->data->port_id);
+		DRV_LOG(WARNING, "port %u message wrongly formatted",
+			dev->data->port_id);
 		goto error;
 	}
 	/* Make sure all the ancillary data was received and valid. */
 	if ((cred == NULL) || (cred->uid != getuid()) ||
 	    (cred->gid != getgid())) {
-		WARN("port %u wrong credentials", dev->data->port_id);
+		DRV_LOG(WARNING, "port %u wrong credentials",
+			dev->data->port_id);
 		goto error;
 	}
 	/* Set-up the ancillary data. */
@@ -208,7 +212,8 @@ mlx5_socket_handle(struct rte_eth_dev *dev)
 	*fd = priv->ctx->cmd_fd;
 	ret = sendmsg(conn_sock, &msg, 0);
 	if (ret < 0)
-		WARN("port %u cannot send response", dev->data->port_id);
+		DRV_LOG(WARNING, "port %u cannot send response",
+			dev->data->port_id);
 error:
 	close(conn_sock);
 }
@@ -250,7 +255,8 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	ret = socket(AF_UNIX, SOCK_STREAM, 0);
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("port %u cannot connect to primary", dev->data->port_id);
+		DRV_LOG(WARNING, "port %u cannot connect to primary",
+			dev->data->port_id);
 		goto error;
 	}
 	socket_fd = ret;
@@ -259,13 +265,15 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	ret = connect(socket_fd, (const struct sockaddr *)&sun, sizeof(sun));
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("port %u cannot connect to primary", dev->data->port_id);
+		DRV_LOG(WARNING, "port %u cannot connect to primary",
+			dev->data->port_id);
 		goto error;
 	}
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
 		rte_errno = EINVAL;
-		DEBUG("port %u cannot get first message", dev->data->port_id);
+		DRV_LOG(DEBUG, "port %u cannot get first message",
+			dev->data->port_id);
 		goto error;
 	}
 	cmsg->cmsg_level = SOL_SOCKET;
@@ -274,7 +282,8 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	cred = (struct ucred *)CMSG_DATA(cmsg);
 	if (cred == NULL) {
 		rte_errno = EINVAL;
-		DEBUG("port %u no credentials received", dev->data->port_id);
+		DRV_LOG(DEBUG, "port %u no credentials received",
+			dev->data->port_id);
 		goto error;
 	}
 	cred->pid = getpid();
@@ -283,27 +292,29 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	ret = sendmsg(socket_fd, &msg, MSG_DONTWAIT);
 	if (ret < 0) {
 		rte_errno = errno;
-		WARN("port %u cannot send credentials to primary: %s",
-		     dev->data->port_id, strerror(errno));
+		DRV_LOG(WARNING,
+			"port %u cannot send credentials to primary: %s",
+			dev->data->port_id, strerror(errno));
 		goto error;
 	}
 	ret = recvmsg(socket_fd, &msg, MSG_WAITALL);
 	if (ret <= 0) {
 		rte_errno = errno;
-		WARN("port %u no message from primary: %s",
-		     dev->data->port_id, strerror(errno));
+		DRV_LOG(WARNING, "port %u no message from primary: %s",
+			dev->data->port_id, strerror(errno));
 		goto error;
 	}
 	cmsg = CMSG_FIRSTHDR(&msg);
 	if (cmsg == NULL) {
 		rte_errno = EINVAL;
-		WARN("port %u no file descriptor received", dev->data->port_id);
+		DRV_LOG(WARNING, "port %u no file descriptor received",
+			dev->data->port_id);
 		goto error;
 	}
 	fd = (int *)CMSG_DATA(cmsg);
 	if (*fd < 0) {
-		WARN("port %u no file descriptor received: %s",
-		     dev->data->port_id, strerror(errno));
+		DRV_LOG(WARNING, "port %u no file descriptor received: %s",
+			dev->data->port_id, strerror(errno));
 		rte_errno = *fd;
 		goto error;
 	}
diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index cd8a94a48..7dda2691d 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -160,8 +160,9 @@ mlx5_read_dev_counters(struct rte_eth_dev *dev, uint64_t *stats)
 	ifr.ifr_data = (caddr_t)et_stats;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("port %u unable to read statistic values from device",
-		     dev->data->port_id);
+		DRV_LOG(WARNING,
+			"port %u unable to read statistic values from device",
+			dev->data->port_id);
 		return ret;
 	}
 	for (i = 0; i != xstats_n; ++i) {
@@ -207,8 +208,8 @@ mlx5_ethtool_get_stats_n(struct rte_eth_dev *dev) {
 	ifr.ifr_data = (caddr_t)&drvinfo;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("port %u unable to query number of statistics",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u unable to query number of statistics",
+			dev->data->port_id);
 		return ret;
 	}
 	return drvinfo.n_stats;
@@ -235,8 +236,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 
 	ret = mlx5_ethtool_get_stats_n(dev);
 	if (ret < 0) {
-		WARN("port %u no extended statistics available",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u no extended statistics available",
+			dev->data->port_id);
 		return;
 	}
 	dev_stats_n = ret;
@@ -247,7 +248,7 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 		  rte_malloc("xstats_strings",
 			     str_sz + sizeof(struct ethtool_gstrings), 0);
 	if (!strings) {
-		WARN("port %u unable to allocate memory for xstats",
+		DRV_LOG(WARNING, "port %u unable to allocate memory for xstats",
 		     dev->data->port_id);
 		return;
 	}
@@ -257,8 +258,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 	ifr.ifr_data = (caddr_t)strings;
 	ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
 	if (ret) {
-		WARN("port %u unable to get statistic names",
-		     dev->data->port_id);
+		DRV_LOG(WARNING, "port %u unable to get statistic names",
+			dev->data->port_id);
 		goto free;
 	}
 	for (j = 0; j != xstats_n; ++j)
@@ -279,9 +280,10 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 		if (mlx5_counters_init[j].ib)
 			continue;
 		if (xstats_ctrl->dev_table_idx[j] >= dev_stats_n) {
-			WARN("port %u counter \"%s\" is not recognized",
-			     dev->data->port_id,
-			     mlx5_counters_init[j].dpdk_name);
+			DRV_LOG(WARNING,
+				"port %u counter \"%s\" is not recognized",
+				dev->data->port_id,
+				mlx5_counters_init[j].dpdk_name);
 			goto free;
 		}
 	}
@@ -289,8 +291,8 @@ mlx5_xstats_init(struct rte_eth_dev *dev)
 	assert(xstats_n <= MLX5_MAX_XSTATS);
 	ret = mlx5_read_dev_counters(dev, xstats_ctrl->base);
 	if (ret)
-		ERROR("port %u cannot read device counters: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot read device counters: %s",
+			dev->data->port_id, strerror(rte_errno));
 free:
 	rte_free(strings);
 }
@@ -457,16 +459,16 @@ mlx5_xstats_reset(struct rte_eth_dev *dev)
 
 	stats_n = mlx5_ethtool_get_stats_n(dev);
 	if (stats_n < 0) {
-		ERROR("port %u cannot get stats: %s", dev->data->port_id,
-		      strerror(-stats_n));
+		DRV_LOG(ERR, "port %u cannot get stats: %s", dev->data->port_id,
+			strerror(-stats_n));
 		return;
 	}
 	if (xstats_ctrl->stats_n != stats_n)
 		mlx5_xstats_init(dev);
 	ret = mlx5_read_dev_counters(dev, counters);
 	if (ret) {
-		ERROR("port %u cannot read device counters: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u cannot read device counters: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return;
 	}
 	for (i = 0; i != n; ++i)
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index fd9b62251..b83c2b900 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -177,39 +177,39 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 	dev->data->dev_started = 1;
 	ret = mlx5_flow_create_drop_queue(dev);
 	if (ret) {
-		ERROR("port %u drop queue allocation failed: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u drop queue allocation failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
-	DEBUG("port %u allocating and configuring hash Rx queues",
-	      dev->data->port_id);
+	DRV_LOG(DEBUG, "port %u allocating and configuring hash Rx queues",
+		dev->data->port_id);
 	rte_mempool_walk(mlx5_mp2mr_iter, priv);
 	ret = mlx5_txq_start(dev);
 	if (ret) {
-		ERROR("port %u Tx queue allocation failed: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u Tx queue allocation failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
 	ret = mlx5_rxq_start(dev);
 	if (ret) {
-		ERROR("port %u Rx queue allocation failed: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u Rx queue allocation failed: %s",
+			dev->data->port_id, strerror(rte_errno));
 		goto error;
 	}
 	ret = mlx5_rx_intr_vec_enable(dev);
 	if (ret) {
-		ERROR("port %u Rx interrupt vector creation failed",
-		      dev->data->port_id);
+		DRV_LOG(ERR, "port %u Rx interrupt vector creation failed",
+			dev->data->port_id);
 		goto error;
 	}
 	mlx5_xstats_init(dev);
 	/* Update link status and Tx/Rx callbacks for the first time. */
 	memset(&dev->data->dev_link, 0, sizeof(struct rte_eth_link));
-	INFO("port %u forcing link to be up", dev->data->port_id);
+	DRV_LOG(INFO, "forcing port %u link to be up", dev->data->port_id);
 	ret = mlx5_force_link_status_change(dev, ETH_LINK_UP);
 	if (ret) {
-		DEBUG("failed to set port %u link to be up",
-		      dev->data->port_id);
+		DRV_LOG(DEBUG, "failed to set port %u link to be up",
+			dev->data->port_id);
 		goto error;
 	}
 	mlx5_dev_interrupt_handler_install(dev);
@@ -249,8 +249,8 @@ mlx5_dev_stop(struct rte_eth_dev *dev)
 	dev->tx_pkt_burst = removed_tx_burst;
 	rte_wmb();
 	usleep(1000 * priv->rxqs_n);
-	DEBUG("port %u cleaning up and destroying hash Rx queues",
-	      dev->data->port_id);
+	DRV_LOG(DEBUG, "port %u cleaning up and destroying hash Rx queues",
+		dev->data->port_id);
 	mlx5_flow_stop(dev, &priv->flows);
 	mlx5_traffic_disable(dev);
 	mlx5_rx_intr_vec_disable(dev);
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index b6d0066fc..4e54ff33d 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -74,8 +74,8 @@ txq_alloc_elts(struct mlx5_txq_ctrl *txq_ctrl)
 
 	for (i = 0; (i != elts_n); ++i)
 		(*txq_ctrl->txq.elts)[i] = NULL;
-	DEBUG("port %u Tx queue %u allocated and configured %u WRs",
-	      txq_ctrl->priv->dev->data->port_id, txq_ctrl->idx, elts_n);
+	DRV_LOG(DEBUG, "port %u Tx queue %u allocated and configured %u WRs",
+		txq_ctrl->priv->dev->data->port_id, txq_ctrl->idx, elts_n);
 	txq_ctrl->txq.elts_head = 0;
 	txq_ctrl->txq.elts_tail = 0;
 	txq_ctrl->txq.elts_comp = 0;
@@ -96,8 +96,8 @@ txq_free_elts(struct mlx5_txq_ctrl *txq_ctrl)
 	uint16_t elts_tail = txq_ctrl->txq.elts_tail;
 	struct rte_mbuf *(*elts)[elts_n] = txq_ctrl->txq.elts;
 
-	DEBUG("port %u Tx queue %u freeing WRs",
-	      txq_ctrl->priv->dev->data->port_id, txq_ctrl->idx);
+	DRV_LOG(DEBUG, "port %u Tx queue %u freeing WRs",
+		txq_ctrl->priv->dev->data->port_id, txq_ctrl->idx);
 	txq_ctrl->txq.elts_head = 0;
 	txq_ctrl->txq.elts_tail = 0;
 	txq_ctrl->txq.elts_comp = 0;
@@ -144,40 +144,43 @@ mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		container_of(txq, struct mlx5_txq_ctrl, txq);
 
 	if (desc <= MLX5_TX_COMP_THRESH) {
-		WARN("port %u number of descriptors requested for Tx queue %u"
-		     " must be higher than MLX5_TX_COMP_THRESH, using"
-		     " %u instead of %u",
-		     dev->data->port_id, idx, MLX5_TX_COMP_THRESH + 1, desc);
+		DRV_LOG(WARNING,
+			"port %u number of descriptors requested for Tx queue"
+			" %u must be higher than MLX5_TX_COMP_THRESH, using %u"
+			" instead of %u",
+			dev->data->port_id, idx, MLX5_TX_COMP_THRESH + 1, desc);
 		desc = MLX5_TX_COMP_THRESH + 1;
 	}
 	if (!rte_is_power_of_2(desc)) {
 		desc = 1 << log2above(desc);
-		WARN("port %u increased number of descriptors in Tx queue %u"
-		     " to the next power of two (%d)",
-		     dev->data->port_id, idx, desc);
+		DRV_LOG(WARNING,
+			"port %u increased number of descriptors in Tx queue"
+			" %u to the next power of two (%d)",
+			dev->data->port_id, idx, desc);
 	}
-	DEBUG("port %u configuring queue %u for %u descriptors",
-	      dev->data->port_id, idx, desc);
+	DRV_LOG(DEBUG, "port %u configuring queue %u for %u descriptors",
+		dev->data->port_id, idx, desc);
 	if (idx >= priv->txqs_n) {
-		ERROR("port %u Tx queue index out of range (%u >= %u)",
-		      dev->data->port_id, idx, priv->txqs_n);
+		DRV_LOG(ERR, "port %u Tx queue index out of range (%u >= %u)",
+			dev->data->port_id, idx, priv->txqs_n);
 		rte_errno = EOVERFLOW;
 		return -rte_errno;
 	}
 	if (!mlx5_txq_releasable(dev, idx)) {
 		rte_errno = EBUSY;
-		ERROR("port %u unable to release queue index %u",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u unable to release queue index %u",
+			dev->data->port_id, idx);
 		return -rte_errno;
 	}
 	mlx5_txq_release(dev, idx);
 	txq_ctrl = mlx5_txq_new(dev, idx, desc, socket, conf);
 	if (!txq_ctrl) {
-		ERROR("port %u unable to allocate queue index %u",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u unable to allocate queue index %u",
+			dev->data->port_id, idx);
 		return -rte_errno;
 	}
-	DEBUG("port %u adding Tx queue %u to list", dev->data->port_id, idx);
+	DRV_LOG(DEBUG, "port %u adding Tx queue %u to list",
+		dev->data->port_id, idx);
 	(*priv->txqs)[idx] = &txq_ctrl->txq;
 	return 0;
 }
@@ -203,8 +206,8 @@ mlx5_tx_queue_release(void *dpdk_txq)
 	for (i = 0; (i != priv->txqs_n); ++i)
 		if ((*priv->txqs)[i] == txq) {
 			mlx5_txq_release(priv->dev, i);
-			DEBUG("port %u removing Tx queue %u from list",
-			      priv->dev->data->port_id, txq_ctrl->idx);
+			DRV_LOG(DEBUG, "port %u removing Tx queue %u from list",
+				priv->dev->data->port_id, txq_ctrl->idx);
 			break;
 		}
 }
@@ -275,9 +278,10 @@ mlx5_tx_uar_remap(struct rte_eth_dev *dev, int fd)
 				   txq_ctrl->uar_mmap_offset);
 			if (ret != addr) {
 				/* fixed mmap have to return same address */
-				ERROR("port %u call to mmap failed on UAR for"
-				      " txq %u", dev->data->port_id,
-				      txq_ctrl->idx);
+				DRV_LOG(ERR,
+					"port %u call to mmap failed on UAR"
+					" for txq %u",
+					dev->data->port_id, txq_ctrl->idx);
 				rte_errno = ENXIO;
 				return -rte_errno;
 			}
@@ -328,8 +332,9 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_TX_QUEUE;
 	priv->verbs_alloc_ctx.obj = txq_ctrl;
 	if (mlx5_getenv_int("MLX5_ENABLE_CQE_COMPRESSION")) {
-		ERROR("port %u MLX5_ENABLE_CQE_COMPRESSION must never be set",
-		      dev->data->port_id);
+		DRV_LOG(ERR,
+			"port %u MLX5_ENABLE_CQE_COMPRESSION must never be set",
+			dev->data->port_id);
 		rte_errno = EINVAL;
 		return NULL;
 	}
@@ -344,8 +349,8 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		cqe_n += MLX5_TX_COMP_THRESH_INLINE_DIV;
 	tmpl.cq = ibv_create_cq(priv->ctx, cqe_n, NULL, NULL, 0);
 	if (tmpl.cq == NULL) {
-		ERROR("port %u Tx queue %u CQ creation failure",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u Tx queue %u CQ creation failure",
+			dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
@@ -387,8 +392,8 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	}
 	tmpl.qp = ibv_create_qp_ex(priv->ctx, &attr.init);
 	if (tmpl.qp == NULL) {
-		ERROR("port %u Tx queue %u QP creation failure",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u Tx queue %u QP creation failure",
+			dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
@@ -400,8 +405,9 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	};
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, (IBV_QP_STATE | IBV_QP_PORT));
 	if (ret) {
-		ERROR("port %u Tx queue %u QP state to IBV_QPS_INIT failed",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR,
+			"port %u Tx queue %u QP state to IBV_QPS_INIT failed",
+			dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
@@ -410,24 +416,26 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	};
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, IBV_QP_STATE);
 	if (ret) {
-		ERROR("port %u Tx queue %u QP state to IBV_QPS_RTR failed",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR,
+			"port %u Tx queue %u QP state to IBV_QPS_RTR failed",
+			dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
 	attr.mod.qp_state = IBV_QPS_RTS;
 	ret = ibv_modify_qp(tmpl.qp, &attr.mod, IBV_QP_STATE);
 	if (ret) {
-		ERROR("port %u Tx queue %u QP state to IBV_QPS_RTS failed",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR,
+			"port %u Tx queue %u QP state to IBV_QPS_RTS failed",
+			dev->data->port_id, idx);
 		rte_errno = errno;
 		goto error;
 	}
 	txq_ibv = rte_calloc_socket(__func__, 1, sizeof(struct mlx5_txq_ibv), 0,
 				    txq_ctrl->socket);
 	if (!txq_ibv) {
-		ERROR("port %u Tx queue %u cannot allocate memory",
-		      dev->data->port_id, idx);
+		DRV_LOG(ERR, "port %u Tx queue %u cannot allocate memory",
+			dev->data->port_id, idx);
 		rte_errno = ENOMEM;
 		goto error;
 	}
@@ -441,9 +449,10 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		goto error;
 	}
 	if (cq_info.cqe_size != RTE_CACHE_LINE_SIZE) {
-		ERROR("port %u wrong MLX5_CQE_SIZE environment variable value: "
-		      "it should be set to %u", dev->data->port_id,
-		      RTE_CACHE_LINE_SIZE);
+		DRV_LOG(ERR,
+			"port %u wrong MLX5_CQE_SIZE environment variable"
+			" value: it should be set to %u",
+			dev->data->port_id, RTE_CACHE_LINE_SIZE);
 		rte_errno = EINVAL;
 		goto error;
 	}
@@ -467,13 +476,15 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 	if (qp.comp_mask & MLX5DV_QP_MASK_UAR_MMAP_OFFSET) {
 		txq_ctrl->uar_mmap_offset = qp.uar_mmap_offset;
 	} else {
-		ERROR("port %u failed to retrieve UAR info, invalid libmlx5.so",
-		      dev->data->port_id);
+		DRV_LOG(ERR,
+			"port %u failed to retrieve UAR info, invalid"
+			" libmlx5.so",
+			dev->data->port_id);
 		rte_errno = EINVAL;
 		goto error;
 	}
-	DEBUG("port %u Verbs Tx queue %u: refcnt %d", dev->data->port_id, idx,
-	      rte_atomic32_read(&txq_ibv->refcnt));
+	DRV_LOG(DEBUG, "port %u Verbs Tx queue %u: refcnt %d",
+		dev->data->port_id, idx, rte_atomic32_read(&txq_ibv->refcnt));
 	LIST_INSERT_HEAD(&priv->txqsibv, txq_ibv, next);
 	txq_ibv->txq_ctrl = txq_ctrl;
 	priv->verbs_alloc_ctx.type = MLX5_VERBS_ALLOC_TYPE_NONE;
@@ -513,8 +524,8 @@ mlx5_txq_ibv_get(struct rte_eth_dev *dev, uint16_t idx)
 	txq_ctrl = container_of((*priv->txqs)[idx], struct mlx5_txq_ctrl, txq);
 	if (txq_ctrl->ibv) {
 		rte_atomic32_inc(&txq_ctrl->ibv->refcnt);
-		DEBUG("port %u Verbs Tx queue %u: refcnt %d",
-		      dev->data->port_id, txq_ctrl->idx,
+		DRV_LOG(DEBUG, "port %u Verbs Tx queue %u: refcnt %d",
+			dev->data->port_id, txq_ctrl->idx,
 		      rte_atomic32_read(&txq_ctrl->ibv->refcnt));
 	}
 	return txq_ctrl->ibv;
@@ -533,9 +544,9 @@ int
 mlx5_txq_ibv_release(struct mlx5_txq_ibv *txq_ibv)
 {
 	assert(txq_ibv);
-	DEBUG("port %u Verbs Tx queue %u: refcnt %d",
-	      txq_ibv->txq_ctrl->priv->dev->data->port_id,
-	      txq_ibv->txq_ctrl->idx, rte_atomic32_read(&txq_ibv->refcnt));
+	DRV_LOG(DEBUG, "port %u Verbs Tx queue %u: refcnt %d",
+		txq_ibv->txq_ctrl->priv->dev->data->port_id,
+		txq_ibv->txq_ctrl->idx, rte_atomic32_read(&txq_ibv->refcnt));
 	if (rte_atomic32_dec_and_test(&txq_ibv->refcnt)) {
 		claim_zero(ibv_destroy_qp(txq_ibv->qp));
 		claim_zero(ibv_destroy_cq(txq_ibv->cq));
@@ -576,9 +587,8 @@ mlx5_txq_ibv_verify(struct rte_eth_dev *dev)
 	struct mlx5_txq_ibv *txq_ibv;
 
 	LIST_FOREACH(txq_ibv, &priv->txqsibv, next) {
-		DEBUG("port %u Verbs Tx queue %u still referenced",
-		      dev->data->port_id,
-		      txq_ibv->txq_ctrl->idx);
+		DRV_LOG(DEBUG, "port %u Verbs Tx queue %u still referenced",
+			dev->data->port_id, txq_ibv->txq_ctrl->idx);
 		++ret;
 	}
 	return ret;
@@ -628,10 +638,10 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 	if (priv->mps == MLX5_MPW_ENHANCED)
 		tmpl->txq.mpw_hdr_dseg = priv->mpw_hdr_dseg;
 	/* MRs will be registered in mp2mr[] later. */
-	DEBUG("port %u priv->device_attr.max_qp_wr is %d", dev->data->port_id,
-	      priv->device_attr.orig_attr.max_qp_wr);
-	DEBUG("port %u priv->device_attr.max_sge is %d", dev->data->port_id,
-	      priv->device_attr.orig_attr.max_sge);
+	DRV_LOG(DEBUG, "port %u priv->device_attr.max_qp_wr is %d",
+		dev->data->port_id, priv->device_attr.orig_attr.max_qp_wr);
+	DRV_LOG(DEBUG, "port %u priv->device_attr.max_sge is %d",
+		dev->data->port_id, priv->device_attr.orig_attr.max_sge);
 	if (priv->txq_inline && (priv->txqs_n >= priv->txqs_inline)) {
 		unsigned int ds_cnt;
 
@@ -682,10 +692,11 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 
 			max_inline = max_inline - (max_inline %
 						   RTE_CACHE_LINE_SIZE);
-			WARN("port %u txq inline is too large (%d) setting it"
-			     " to the maximum possible: %d\n",
-			     priv->dev->data->port_id, priv->txq_inline,
-			     max_inline);
+			DRV_LOG(WARNING,
+				"port %u txq inline is too large (%d) setting it"
+				" to the maximum possible: %d\n",
+				priv->dev->data->port_id, priv->txq_inline,
+				max_inline);
 			tmpl->txq.max_inline = max_inline / RTE_CACHE_LINE_SIZE;
 		}
 	}
@@ -701,8 +712,8 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		(struct rte_mbuf *(*)[1 << tmpl->txq.elts_n])(tmpl + 1);
 	tmpl->txq.stats.idx = idx;
 	rte_atomic32_inc(&tmpl->refcnt);
-	DEBUG("port %u Tx queue %u: refcnt %d", dev->data->port_id,
-	      idx, rte_atomic32_read(&tmpl->refcnt));
+	DRV_LOG(DEBUG, "port %u Tx queue %u: refcnt %d", dev->data->port_id,
+		idx, rte_atomic32_read(&tmpl->refcnt));
 	LIST_INSERT_HEAD(&priv->txqsctrl, tmpl, next);
 	return tmpl;
 }
@@ -737,8 +748,9 @@ mlx5_txq_get(struct rte_eth_dev *dev, uint16_t idx)
 						     ctrl->txq.mp2mr[i]->mp));
 		}
 		rte_atomic32_inc(&ctrl->refcnt);
-		DEBUG("port %u Tx queue %u refcnt %d", dev->data->port_id,
-		      ctrl->idx, rte_atomic32_read(&ctrl->refcnt));
+		DRV_LOG(DEBUG, "port %u Tx queue %u refcnt %d",
+			dev->data->port_id,
+			ctrl->idx, rte_atomic32_read(&ctrl->refcnt));
 	}
 	return ctrl;
 }
@@ -765,8 +777,8 @@ mlx5_txq_release(struct rte_eth_dev *dev, uint16_t idx)
 	if (!(*priv->txqs)[idx])
 		return 0;
 	txq = container_of((*priv->txqs)[idx], struct mlx5_txq_ctrl, txq);
-	DEBUG("port %u Tx queue %u: refcnt %d", dev->data->port_id,
-	      txq->idx, rte_atomic32_read(&txq->refcnt));
+	DRV_LOG(DEBUG, "port %u Tx queue %u: refcnt %d", dev->data->port_id,
+		txq->idx, rte_atomic32_read(&txq->refcnt));
 	if (txq->ibv && !mlx5_txq_ibv_release(txq->ibv))
 		txq->ibv = NULL;
 	for (i = 0; i != MLX5_PMD_TX_MP_CACHE; ++i) {
@@ -828,8 +840,8 @@ mlx5_txq_verify(struct rte_eth_dev *dev)
 	int ret = 0;
 
 	LIST_FOREACH(txq, &priv->txqsctrl, next) {
-		DEBUG("port %u Tx queue %u still referenced",
-		      dev->data->port_id, txq->idx);
+		DRV_LOG(DEBUG, "port %u Tx queue %u still referenced",
+			dev->data->port_id, txq->idx);
 		++ret;
 	}
 	return ret;
diff --git a/drivers/net/mlx5/mlx5_utils.h b/drivers/net/mlx5/mlx5_utils.h
index 2fbd10b18..6c85c0739 100644
--- a/drivers/net/mlx5/mlx5_utils.h
+++ b/drivers/net/mlx5/mlx5_utils.h
@@ -89,14 +89,21 @@ pmd_drv_log_basename(const char *s)
 	return s;
 }
 
+extern int mlx5_logtype;
+
+#define PMD_DRV_LOG___(level, ...) \
+	rte_log(RTE_LOG_ ## level, \
+		mlx5_logtype, \
+		RTE_FMT(MLX5_DRIVER_NAME ": " \
+			RTE_FMT_HEAD(__VA_ARGS__,), \
+		RTE_FMT_TAIL(__VA_ARGS__,)))
+
 /*
  * When debugging is enabled (NDEBUG not defined), file, line and function
  * information replace the driver name (MLX5_DRIVER_NAME) in log messages.
  */
 #ifndef NDEBUG
 
-#define PMD_DRV_LOG___(level, ...) \
-	ERRNO_SAFE(RTE_LOG(level, PMD, __VA_ARGS__))
 #define PMD_DRV_LOG__(level, ...) \
 	PMD_DRV_LOG___(level, "%s:%u: %s(): " __VA_ARGS__)
 #define PMD_DRV_LOG_(level, s, ...) \
@@ -108,9 +115,6 @@ pmd_drv_log_basename(const char *s)
 		__VA_ARGS__)
 
 #else /* NDEBUG */
-
-#define PMD_DRV_LOG___(level, ...) \
-	ERRNO_SAFE(RTE_LOG(level, PMD, MLX5_DRIVER_NAME ": " __VA_ARGS__))
 #define PMD_DRV_LOG__(level, ...) \
 	PMD_DRV_LOG___(level, __VA_ARGS__)
 #define PMD_DRV_LOG_(level, s, ...) \
@@ -119,33 +123,24 @@ pmd_drv_log_basename(const char *s)
 #endif /* NDEBUG */
 
 /* Generic printf()-like logging macro with automatic line feed. */
-#define PMD_DRV_LOG(level, ...) \
+#define DRV_LOG(level, ...) \
 	PMD_DRV_LOG_(level, \
 		__VA_ARGS__ PMD_DRV_LOG_STRIP PMD_DRV_LOG_OPAREN, \
 		PMD_DRV_LOG_CPAREN)
 
-/*
- * Like assert(), DEBUG() becomes a no-op and claim_zero() does not perform
- * any check when debugging is disabled.
- */
+/* claim_zero() does not perform any check when debugging is disabled. */
 #ifndef NDEBUG
 
-#define DEBUG(...) PMD_DRV_LOG(DEBUG, __VA_ARGS__)
 #define claim_zero(...) assert((__VA_ARGS__) == 0)
 #define claim_nonzero(...) assert((__VA_ARGS__) != 0)
 
 #else /* NDEBUG */
 
-#define DEBUG(...) (void)0
 #define claim_zero(...) (__VA_ARGS__)
 #define claim_nonzero(...) (__VA_ARGS__)
 
 #endif /* NDEBUG */
 
-#define INFO(...) PMD_DRV_LOG(INFO, __VA_ARGS__)
-#define WARN(...) PMD_DRV_LOG(WARNING, __VA_ARGS__)
-#define ERROR(...) PMD_DRV_LOG(ERR, __VA_ARGS__)
-
 /* Convenience macros for accessing mbuf fields. */
 #define NEXT(m) ((m)->next)
 #define DATA_LEN(m) ((m)->data_len)
diff --git a/drivers/net/mlx5/mlx5_vlan.c b/drivers/net/mlx5/mlx5_vlan.c
index 377d06884..dbfa8a0c9 100644
--- a/drivers/net/mlx5/mlx5_vlan.c
+++ b/drivers/net/mlx5/mlx5_vlan.c
@@ -62,8 +62,8 @@ mlx5_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
 	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
-	DEBUG("port %u %s VLAN filter ID %" PRIu16,
-	      dev->data->port_id, (on ? "enable" : "disable"), vlan_id);
+	DRV_LOG(DEBUG, "port %u %s VLAN filter ID %" PRIu16,
+		dev->data->port_id, (on ? "enable" : "disable"), vlan_id);
 	assert(priv->vlan_filter_n <= RTE_DIM(priv->vlan_filter));
 	for (i = 0; (i != priv->vlan_filter_n); ++i)
 		if (priv->vlan_filter[i] == vlan_id)
@@ -125,18 +125,18 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 
 	/* Validate hw support */
 	if (!priv->hw_vlan_strip) {
-		ERROR("port %u VLAN stripping is not supported",
-		      dev->data->port_id);
+		DRV_LOG(ERR, "port %u VLAN stripping is not supported",
+			dev->data->port_id);
 		return;
 	}
 	/* Validate queue number */
 	if (queue >= priv->rxqs_n) {
-		ERROR("port %u VLAN stripping, invalid queue number %d",
-		      dev->data->port_id, queue);
+		DRV_LOG(ERR, "port %u VLAN stripping, invalid queue number %d",
+			dev->data->port_id, queue);
 		return;
 	}
-	DEBUG("port %u set VLAN offloads 0x%x for port %uqueue %d",
-	      dev->data->port_id, vlan_offloads, rxq->port_id, queue);
+	DRV_LOG(DEBUG, "port %u set VLAN offloads 0x%x for port %uqueue %d",
+		dev->data->port_id, vlan_offloads, rxq->port_id, queue);
 	if (!rxq_ctrl->ibv) {
 		/* Update related bits in RX queue. */
 		rxq->vlan_strip = !!on;
@@ -149,8 +149,8 @@ mlx5_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on)
 	};
 	ret = ibv_modify_wq(rxq_ctrl->ibv->wq, &mod);
 	if (ret) {
-		ERROR("port %u failed to modified stripping mode: %s",
-		      dev->data->port_id, strerror(rte_errno));
+		DRV_LOG(ERR, "port %u failed to modified stripping mode: %s",
+			dev->data->port_id, strerror(rte_errno));
 		return;
 	}
 	/* Update related bits in RX queue. */
@@ -178,8 +178,8 @@ mlx5_vlan_offload_set(struct rte_eth_dev *dev, int mask)
 		int hw_vlan_strip = !!dev->data->dev_conf.rxmode.hw_vlan_strip;
 
 		if (!priv->hw_vlan_strip) {
-			ERROR("port %u VLAN stripping is not supported",
-			      dev->data->port_id);
+			DRV_LOG(ERR, "port %u VLAN stripping is not supported",
+				dev->data->port_id);
 			return 0;
 		}
 		/* Run on every RX queue and set/reset VLAN stripping. */
-- 
2.11.0

[dpdk-stable] [PATCH v2 30/67] net/mlx5: remove kernel version check

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit f47ba80080ab81ad513f843407bec7178fed8857 ]

Kernel version check was introduced in
commit 3a49ffe38a95 ("net/mlx5: fix link status query")
due to a bug fixed by
commit ef09a7fc7620 ("net/mlx5: fix inconsistent link status query")

This patch restore the previous behavior as described in Linux API.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5_ethdev.c | 17 +++++------------
 1 file changed, 5 insertions(+), 12 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index b78756efc..7ae4e8577 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -46,11 +46,9 @@
 #include <net/if.h>
 #include <sys/ioctl.h>
 #include <sys/socket.h>
-#include <sys/utsname.h>
 #include <netinet/in.h>
 #include <linux/ethtool.h>
 #include <linux/sockios.h>
-#include <linux/version.h>
 #include <fcntl.h>
 #include <stdalign.h>
 #include <sys/un.h>
@@ -756,20 +754,15 @@ mlx5_force_link_status_change(struct rte_eth_dev *dev, int status)
 int
 mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
 {
-	struct utsname utsname;
-	int ver[3];
 	int ret;
 	struct rte_eth_link dev_link = dev->data->dev_link;
 
-	if (uname(&utsname) == -1 ||
-	    sscanf(utsname.release, "%d.%d.%d",
-		   &ver[0], &ver[1], &ver[2]) != 3 ||
-	    KERNEL_VERSION(ver[0], ver[1], ver[2]) < KERNEL_VERSION(4, 9, 0))
-		ret = mlx5_link_update_unlocked_gset(dev);
-	else
+	ret = mlx5_link_update_unlocked_gset(dev);
+	if (ret) {
 		ret = mlx5_link_update_unlocked_gs(dev);
-	if (ret)
-		return ret;
+		if (ret)
+			return ret;
+	}
 	/* If lsc interrupt is disabled, should always be ready for traffic. */
 	if (!dev->data->dev_conf.intr_conf.lsc) {
 		mlx5_link_start(dev);
-- 
2.11.0

[dpdk-stable] [PATCH v2 31/67] net/mlx5: change pkt burst select function prototype

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ backported from upstream commit 1cfa649ba614693165fc4659ace525d8b45bf110 ]

Change the function prototype to return the function pointer of the
selected Tx/Rx burst function instead of assigning it directly to the
device context.

Such change will enable to use those select functions to query the burst
function that will be selected according to the device configuration.

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5.c        | 11 +++++++++--
 drivers/net/mlx5/mlx5.h        |  4 ++--
 drivers/net/mlx5/mlx5_ethdev.c | 34 +++++++++++++++++++++-------------
 3 files changed, 32 insertions(+), 17 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 911d4cf65..141d46dae 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -834,8 +834,15 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			err = mlx5_tx_uar_remap(eth_dev, err);
 			if (err)
 				goto error;
-			mlx5_select_rx_function(eth_dev);
-			mlx5_select_tx_function(eth_dev);
+			/*
+			 * Ethdev pointer is still required as input since
+			 * the primary device is not accessible from the
+			 * secondary process.
+			 */
+			eth_dev->rx_pkt_burst =
+				mlx5_select_rx_function(eth_dev);
+			eth_dev->tx_pkt_burst =
+				mlx5_select_tx_function(eth_dev);
 			continue;
 		}
 		DRV_LOG(DEBUG, "using port %u (%08" PRIx32 ")", port, test);
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 8ecee0212..58713a136 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -200,8 +200,8 @@ void mlx5_dev_interrupt_handler_uninstall(struct rte_eth_dev *dev);
 void mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev);
 int mlx5_set_link_down(struct rte_eth_dev *dev);
 int mlx5_set_link_up(struct rte_eth_dev *dev);
-void mlx5_select_tx_function(struct rte_eth_dev *dev);
-void mlx5_select_rx_function(struct rte_eth_dev *dev);
+eth_tx_burst_t mlx5_select_tx_function(struct rte_eth_dev *dev);
+eth_rx_burst_t mlx5_select_rx_function(struct rte_eth_dev *dev);
 
 /* mlx5_mac.c */
 
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 7ae4e8577..eacea0aff 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -678,8 +678,8 @@ mlx5_link_start(struct rte_eth_dev *dev)
 	struct priv *priv = dev->data->dev_private;
 	int ret;
 
-	mlx5_select_tx_function(dev);
-	mlx5_select_rx_function(dev);
+	dev->tx_pkt_burst = mlx5_select_tx_function(dev);
+	dev->rx_pkt_burst = mlx5_select_rx_function(dev);
 	ret = mlx5_traffic_enable(dev);
 	if (ret) {
 		DRV_LOG(ERR,
@@ -1198,39 +1198,43 @@ mlx5_set_link_up(struct rte_eth_dev *dev)
  *
  * @param dev
  *   Pointer to rte_eth_dev structure.
+ *
+ * @return
+ *   Pointer to selected Tx burst function.
  */
-void
+eth_tx_burst_t
 mlx5_select_tx_function(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
+	eth_tx_burst_t tx_pkt_burst = mlx5_tx_burst;
 
-	dev->tx_pkt_burst = mlx5_tx_burst;
 	/* Select appropriate TX function. */
 	if (priv->mps == MLX5_MPW_ENHANCED) {
 		if (mlx5_check_vec_tx_support(dev) > 0) {
 			if (mlx5_check_raw_vec_tx_support(dev) > 0)
-				dev->tx_pkt_burst = mlx5_tx_burst_raw_vec;
+				tx_pkt_burst = mlx5_tx_burst_raw_vec;
 			else
-				dev->tx_pkt_burst = mlx5_tx_burst_vec;
+				tx_pkt_burst = mlx5_tx_burst_vec;
 			DRV_LOG(DEBUG,
 				"port %u selected enhanced MPW Tx vectorized"
 				" function",
 				dev->data->port_id);
 		} else {
-			dev->tx_pkt_burst = mlx5_tx_burst_empw;
+			tx_pkt_burst = mlx5_tx_burst_empw;
 			DRV_LOG(DEBUG,
 				"port %u selected enhanced MPW Tx function",
 				dev->data->port_id);
 		}
 	} else if (priv->mps && priv->txq_inline) {
-		dev->tx_pkt_burst = mlx5_tx_burst_mpw_inline;
+		tx_pkt_burst = mlx5_tx_burst_mpw_inline;
 		DRV_LOG(DEBUG, "port %u selected MPW inline Tx function",
 			dev->data->port_id);
 	} else if (priv->mps) {
-		dev->tx_pkt_burst = mlx5_tx_burst_mpw;
+		tx_pkt_burst = mlx5_tx_burst_mpw;
 		DRV_LOG(DEBUG, "port %u selected MPW Tx function",
 			dev->data->port_id);
 	}
+	return tx_pkt_burst;
 }
 
 /**
@@ -1238,16 +1242,20 @@ mlx5_select_tx_function(struct rte_eth_dev *dev)
  *
  * @param dev
  *   Pointer to rte_eth_dev structure.
+ *
+ * @return
+ *   Pointer to selected Rx burst function.
  */
-void
+eth_rx_burst_t
 mlx5_select_rx_function(struct rte_eth_dev *dev)
 {
+	eth_rx_burst_t rx_pkt_burst = mlx5_rx_burst;
+
 	assert(dev != NULL);
 	if (mlx5_check_vec_rx_support(dev) > 0) {
-		dev->rx_pkt_burst = mlx5_rx_burst_vec;
+		rx_pkt_burst = mlx5_rx_burst_vec;
 		DRV_LOG(DEBUG, "port %u selected Rx vectorized function",
 			dev->data->port_id);
-	} else {
-		dev->rx_pkt_burst = mlx5_rx_burst;
 	}
+	return rx_pkt_burst;
 }
-- 
2.11.0

[dpdk-stable] [PATCH v2 32/67] net/mlx5: fix link status behavior

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 7ba5320baa3285bf8beca96cc025ff56b494060d ]

This behavior is mixed between what should be handled by the application
and what is under PMD responsibility.

According to DPDK API:
- link_update() should only query the link status [1]
- link_set_{up,down}() should only set the link to the according status [1]
- dev_{start,stop}() should enable/disable traffic reception/emission [2]

On this PMD, the link status is retrieved from the net device associated
owned by the Linux Kernel, it does not means that even when this interface
is down, the PMD cannot send/receive traffic from the NIC those two
information are unrelated, until the physical port is active and has a
link, the PMD can receive/send traffic on the wire.

According to DPDK API, calling the rte_eth_dev_start() even when the Linux
interface link is down is then possible and allowed, as the traffic will
flow between the DPDK application and the Physical port.

This also means that a synchronization between the Linux interface and the
DPDK application remains under the DPDK application responsibility.

To handle such synchronization the application should behave as the
following scheme, to start:

 rte_eth_get_link(port_id, &link);
 if (link.link_status == ETH_DOWN)
	rte_eth_dev_set_link_up(port_id);
 rte_eth_dev_start(port_id);

Taking in account the possible returned values for each function.

and to stop:

 rte_eth_dev_stop(port_id);
 rte_eth_dev_set_link_down(port_id);

The application should also set the LSC interrupt callbacks to catch and
behave accordingly when the administrator set the Linux device down/up.
The same callbacks are called when the link on the medium falls/raise.

[1] https://dpdk.org/browse/dpdk/tree/lib/librte_ether/rte_ethdev_core.h
[2] https://dpdk.org/browse/dpdk/tree/lib/librte_ether/rte_ethdev.h#n1677

Fixes: c7bf62255edf ("net/mlx5: fix handling link status event")
Fixes: e313ef4c2fe8 ("net/mlx5: fix link state on device start")
Cc: stable@dpdk.org

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5.c         |  2 +-
 drivers/net/mlx5/mlx5_ethdev.c  | 92 +----------------------------------------
 drivers/net/mlx5/mlx5_trigger.c | 15 ++++---
 3 files changed, 12 insertions(+), 97 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 141d46dae..853599c25 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -1064,7 +1064,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		/* Bring Ethernet device up. */
 		DRV_LOG(DEBUG, "port %u forcing Ethernet interface up",
 			eth_dev->data->port_id);
-		mlx5_set_flags(eth_dev, ~IFF_UP, IFF_UP);
+		mlx5_set_link_up(eth_dev);
 		continue;
 port_error:
 		if (priv)
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index eacea0aff..d0bf5a788 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -667,80 +667,6 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 }
 
 /**
- * Enable receiving and transmitting traffic.
- *
- * @param dev
- *   Pointer to Ethernet device.
- */
-static void
-mlx5_link_start(struct rte_eth_dev *dev)
-{
-	struct priv *priv = dev->data->dev_private;
-	int ret;
-
-	dev->tx_pkt_burst = mlx5_select_tx_function(dev);
-	dev->rx_pkt_burst = mlx5_select_rx_function(dev);
-	ret = mlx5_traffic_enable(dev);
-	if (ret) {
-		DRV_LOG(ERR,
-			"port %u error occurred while configuring control"
-			" flows: %s",
-			dev->data->port_id, strerror(rte_errno));
-		return;
-	}
-	ret = mlx5_flow_start(dev, &priv->flows);
-	if (ret)
-		DRV_LOG(ERR,
-			"port %u error occurred while configuring flows: %s",
-			dev->data->port_id, strerror(rte_errno));
-}
-
-/**
- * Disable receiving and transmitting traffic.
- *
- * @param dev
- *   Pointer to Ethernet device.
- */
-static void
-mlx5_link_stop(struct rte_eth_dev *dev)
-{
-	struct priv *priv = dev->data->dev_private;
-
-	mlx5_flow_stop(dev, &priv->flows);
-	mlx5_traffic_disable(dev);
-	dev->rx_pkt_burst = removed_rx_burst;
-	dev->tx_pkt_burst = removed_tx_burst;
-}
-
-/**
- * Querying the link status till it changes to the desired state.
- * Number of query attempts is bounded by MLX5_MAX_LINK_QUERY_ATTEMPTS.
- *
- * @param dev
- *   Pointer to Ethernet device.
- * @param status
- *   Link desired status.
- *
- * @return
- *   0 on success, a negative errno value otherwise and rte_errno is set.
- */
-int
-mlx5_force_link_status_change(struct rte_eth_dev *dev, int status)
-{
-	int try = 0;
-
-	while (try < MLX5_MAX_LINK_QUERY_ATTEMPTS) {
-		mlx5_link_update(dev, 0);
-		if (dev->data->dev_link.link_status == status)
-			return 0;
-		try++;
-		sleep(1);
-	}
-	rte_errno = EAGAIN;
-	return -rte_errno;
-}
-
-/**
  * DPDK callback to retrieve physical link information.
  *
  * @param dev
@@ -755,26 +681,10 @@ int
 mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
 {
 	int ret;
-	struct rte_eth_link dev_link = dev->data->dev_link;
 
 	ret = mlx5_link_update_unlocked_gset(dev);
-	if (ret) {
+	if (ret)
 		ret = mlx5_link_update_unlocked_gs(dev);
-		if (ret)
-			return ret;
-	}
-	/* If lsc interrupt is disabled, should always be ready for traffic. */
-	if (!dev->data->dev_conf.intr_conf.lsc) {
-		mlx5_link_start(dev);
-		return 0;
-	}
-	/* Re-select burst callbacks only if link status has been changed. */
-	if (!ret && dev_link.link_status != dev->data->dev_link.link_status) {
-		if (dev->data->dev_link.link_status == ETH_LINK_UP)
-			mlx5_link_start(dev);
-		else
-			mlx5_link_stop(dev);
-	}
 	return 0;
 }
 
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index b83c2b900..214543f84 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -203,15 +203,20 @@ mlx5_dev_start(struct rte_eth_dev *dev)
 		goto error;
 	}
 	mlx5_xstats_init(dev);
-	/* Update link status and Tx/Rx callbacks for the first time. */
-	memset(&dev->data->dev_link, 0, sizeof(struct rte_eth_link));
-	DRV_LOG(INFO, "forcing port %u link to be up", dev->data->port_id);
-	ret = mlx5_force_link_status_change(dev, ETH_LINK_UP);
+	ret = mlx5_traffic_enable(dev);
 	if (ret) {
-		DRV_LOG(DEBUG, "failed to set port %u link to be up",
+		DRV_LOG(DEBUG, "port %u failed to set defaults flows",
 			dev->data->port_id);
 		goto error;
 	}
+	ret = mlx5_flow_start(dev, &priv->flows);
+	if (ret) {
+		DRV_LOG(DEBUG, "port %u failed to set flows",
+			dev->data->port_id);
+		goto error;
+	}
+	dev->tx_pkt_burst = mlx5_select_tx_function(dev);
+	dev->rx_pkt_burst = mlx5_select_rx_function(dev);
 	mlx5_dev_interrupt_handler_install(dev);
 	return 0;
 error:
-- 
2.11.0

[dpdk-stable] [PATCH v2 33/67] net/mlx5: fix link status to use wait to complete

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit cfee94752b8f8f09b984335ae77b29af15b967fd ]

Wait to complete is present to let the application get a correct status
when it requires it, it should not be ignored.

Fixes: e313ef4c2fe8 ("net/mlx5: fix link state on device start")
Fixes: cb8faed7dde8 ("mlx5: support link status update")
Cc: stable@dpdk.org

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5.h        |   1 -
 drivers/net/mlx5/mlx5_defs.h   |   4 +-
 drivers/net/mlx5/mlx5_ethdev.c | 148 ++++++++++++++++-------------------------
 3 files changed, 58 insertions(+), 95 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 58713a136..75aa853f2 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -128,7 +128,6 @@ struct priv {
 	unsigned int mps:2; /* Multi-packet send mode (0: disabled). */
 	unsigned int mpw_hdr_dseg:1; /* Enable DSEGs in the title WQEBB. */
 	unsigned int cqe_comp:1; /* Whether CQE compression is enabled. */
-	unsigned int pending_alarm:1; /* An alarm is pending. */
 	unsigned int tso:1; /* Whether TSO is supported. */
 	unsigned int tunnel_en:1;
 	unsigned int isolated:1; /* Whether isolated mode is enabled. */
diff --git a/drivers/net/mlx5/mlx5_defs.h b/drivers/net/mlx5/mlx5_defs.h
index 9c740f5e1..d70635767 100644
--- a/drivers/net/mlx5/mlx5_defs.h
+++ b/drivers/net/mlx5/mlx5_defs.h
@@ -110,8 +110,8 @@
 /* Supported RSS */
 #define MLX5_RSS_HF_MASK (~(ETH_RSS_IP | ETH_RSS_UDP | ETH_RSS_TCP))
 
-/* Maximum number of attempts to query link status before giving up. */
-#define MLX5_MAX_LINK_QUERY_ATTEMPTS 5
+/* Timeout in seconds to get a valid link status. */
+#define MLX5_LINK_STATUS_TIMEOUT 10
 
 /* Reserved address space for UAR mapping. */
 #define MLX5_UAR_SIZE (1ULL << 32)
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index d0bf5a788..18e1e534c 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -52,6 +52,7 @@
 #include <fcntl.h>
 #include <stdalign.h>
 #include <sys/un.h>
+#include <time.h>
 
 #include <rte_atomic.h>
 #include <rte_ethdev.h>
@@ -59,7 +60,6 @@
 #include <rte_mbuf.h>
 #include <rte_common.h>
 #include <rte_interrupts.h>
-#include <rte_alarm.h>
 #include <rte_malloc.h>
 
 #include "mlx5.h"
@@ -495,12 +495,15 @@ mlx5_dev_supported_ptypes_get(struct rte_eth_dev *dev)
  *
  * @param dev
  *   Pointer to Ethernet device structure.
+ * @param[out] link
+ *   Storage for current link status.
  *
  * @return
  *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
-mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
+mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev,
+			       struct rte_eth_link *link)
 {
 	struct priv *priv = dev->data->dev_private;
 	struct ethtool_cmd edata = {
@@ -550,14 +553,13 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
 				ETH_LINK_HALF_DUPLEX : ETH_LINK_FULL_DUPLEX);
 	dev_link.link_autoneg = !(dev->data->dev_conf.link_speeds &
 			ETH_LINK_SPEED_FIXED);
-	if (memcmp(&dev_link, &dev->data->dev_link, sizeof(dev_link))) {
-		/* Link status changed. */
-		dev->data->dev_link = dev_link;
-		return 0;
+	if ((dev_link.link_speed && !dev_link.link_status) ||
+	    (!dev_link.link_speed && dev_link.link_status)) {
+		rte_errno = EAGAIN;
+		return -rte_errno;
 	}
-	/* Link status is still the same. */
-	rte_errno = EAGAIN;
-	return -rte_errno;
+	*link = dev_link;
+	return 0;
 }
 
 /**
@@ -565,12 +567,16 @@ mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev)
  *
  * @param dev
  *   Pointer to Ethernet device structure.
+ * @param[out] link
+ *   Storage for current link status.
  *
  * @return
  *   0 on success, a negative errno value otherwise and rte_errno is set.
  */
 static int
-mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
+mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev,
+			     struct rte_eth_link *link)
+
 {
 	struct priv *priv = dev->data->dev_private;
 	struct ethtool_link_settings gcmd = { .cmd = ETHTOOL_GLINKSETTINGS };
@@ -656,14 +662,13 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
 				ETH_LINK_HALF_DUPLEX : ETH_LINK_FULL_DUPLEX);
 	dev_link.link_autoneg = !(dev->data->dev_conf.link_speeds &
 				  ETH_LINK_SPEED_FIXED);
-	if (memcmp(&dev_link, &dev->data->dev_link, sizeof(dev_link))) {
-		/* Link status changed. */
-		dev->data->dev_link = dev_link;
-		return 0;
+	if ((dev_link.link_speed && !dev_link.link_status) ||
+	    (!dev_link.link_speed && dev_link.link_status)) {
+		rte_errno = EAGAIN;
+		return -rte_errno;
 	}
-	/* Link status is still the same. */
-	rte_errno = EAGAIN;
-	return -rte_errno;
+	*link = dev_link;
+	return 0;
 }
 
 /**
@@ -672,20 +677,43 @@ mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev)
  * @param dev
  *   Pointer to Ethernet device structure.
  * @param wait_to_complete
- *   Wait for request completion (ignored).
+ *   Wait for request completion.
  *
  * @return
- *   0 on success, a negative errno value otherwise and rte_errno is set.
+ *   0 if link status was not updated, positive if it was, a negative errno
+ *   value otherwise and rte_errno is set.
  */
 int
-mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused)
+mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete)
 {
 	int ret;
+	struct rte_eth_link dev_link;
+	time_t start_time = time(NULL);
 
-	ret = mlx5_link_update_unlocked_gset(dev);
-	if (ret)
-		ret = mlx5_link_update_unlocked_gs(dev);
-	return 0;
+	do {
+		ret = mlx5_link_update_unlocked_gset(dev, &dev_link);
+		if (ret)
+			ret = mlx5_link_update_unlocked_gs(dev, &dev_link);
+		if (ret == 0)
+			break;
+		/* Handle wait to complete situation. */
+		if (wait_to_complete && ret == -EAGAIN) {
+			if (abs((int)difftime(time(NULL), start_time)) <
+			    MLX5_LINK_STATUS_TIMEOUT) {
+				usleep(0);
+				continue;
+			} else {
+				rte_errno = EBUSY;
+				return -rte_errno;
+			}
+		} else if (ret < 0) {
+			return ret;
+		}
+	} while (wait_to_complete);
+	ret = !!memcmp(&dev->data->dev_link, &dev_link,
+		       sizeof(struct rte_eth_link));
+	dev->data->dev_link = dev_link;
+	return ret;
 }
 
 /**
@@ -864,47 +892,6 @@ mlx5_ibv_device_to_pci_addr(const struct ibv_device *device,
 }
 
 /**
- * Update the link status.
- *
- * @param dev
- *   Pointer to Ethernet device.
- *
- * @return
- *   Zero if the callback process can be called immediately, negative errno
- *   value otherwise and rte_errno is set.
- */
-static int
-mlx5_link_status_update(struct rte_eth_dev *dev)
-{
-	struct priv *priv = dev->data->dev_private;
-	struct rte_eth_link *link = &dev->data->dev_link;
-	int ret;
-
-	ret = mlx5_link_update(dev, 0);
-	if (ret)
-		return ret;
-	if (((link->link_speed == 0) && link->link_status) ||
-		((link->link_speed != 0) && !link->link_status)) {
-		/*
-		 * Inconsistent status. Event likely occurred before the
-		 * kernel netdevice exposes the new status.
-		 */
-		if (!priv->pending_alarm) {
-			priv->pending_alarm = 1;
-			rte_eal_alarm_set(MLX5_ALARM_TIMEOUT_US,
-					  mlx5_dev_link_status_handler,
-					  priv->dev);
-		}
-		return 1;
-	} else if (unlikely(priv->pending_alarm)) {
-		/* Link interrupt occurred while alarm is already scheduled. */
-		priv->pending_alarm = 0;
-		rte_eal_alarm_cancel(mlx5_dev_link_status_handler, priv->dev);
-	}
-	return 0;
-}
-
-/**
  * Device status handler.
  *
  * @param dev
@@ -922,6 +909,10 @@ mlx5_dev_status_handler(struct rte_eth_dev *dev)
 	struct ibv_async_event event;
 	uint32_t ret = 0;
 
+	if (mlx5_link_update(dev, 0) == -EAGAIN) {
+		usleep(0);
+		return 0;
+	}
 	/* Read all message and acknowledge them. */
 	for (;;) {
 		if (ibv_get_async_event(priv->ctx, &event))
@@ -939,33 +930,10 @@ mlx5_dev_status_handler(struct rte_eth_dev *dev)
 				dev->data->port_id, event.event_type);
 		ibv_ack_async_event(&event);
 	}
-	if (ret & (1 << RTE_ETH_EVENT_INTR_LSC))
-		if (mlx5_link_status_update(dev))
-			ret &= ~(1 << RTE_ETH_EVENT_INTR_LSC);
 	return ret;
 }
 
 /**
- * Handle delayed link status event.
- *
- * @param arg
- *   Registered argument.
- */
-void
-mlx5_dev_link_status_handler(void *arg)
-{
-	struct rte_eth_dev *dev = arg;
-	struct priv *priv = dev->data->dev_private;
-	int ret;
-
-	priv->pending_alarm = 0;
-	ret = mlx5_link_status_update(dev);
-	if (!ret)
-		_rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL,
-					      NULL);
-}
-
-/**
  * Handle interrupts from the NIC.
  *
  * @param[in] intr_handle
@@ -1020,10 +988,6 @@ mlx5_dev_interrupt_handler_uninstall(struct rte_eth_dev *dev)
 	if (priv->primary_socket)
 		rte_intr_callback_unregister(&priv->intr_handle_socket,
 					     mlx5_dev_handler_socket, dev);
-	if (priv->pending_alarm) {
-		priv->pending_alarm = 0;
-		rte_eal_alarm_cancel(mlx5_dev_link_status_handler, dev);
-	}
 	priv->intr_handle.fd = 0;
 	priv->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
 	priv->intr_handle_socket.fd = 0;
-- 
2.11.0

[dpdk-stable] [PATCH v2 34/67] net/mlx5: change tunnel flow priority

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 749365717f5c5ff850f8aa93f512458ab23c72ff ]

Packet matching inner and outer flow rules are caught by the first one
added in the device as both flows are configured with the same priority.
To avoid such situation, the inner flow can have an higher priority than
the outer ones as their pattern matching will otherwise collide.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 73 +++++++++++++++++++++++++++-----------------
 1 file changed, 45 insertions(+), 28 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 9a3fcf43e..c150ad52b 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -157,7 +157,7 @@ const struct hash_rxq_init hash_rxq_init[] = {
 				IBV_RX_HASH_SRC_PORT_TCP |
 				IBV_RX_HASH_DST_PORT_TCP),
 		.dpdk_rss_hf = ETH_RSS_NONFRAG_IPV4_TCP,
-		.flow_priority = 0,
+		.flow_priority = 1,
 		.ip_version = MLX5_IPV4,
 	},
 	[HASH_RXQ_UDPV4] = {
@@ -166,7 +166,7 @@ const struct hash_rxq_init hash_rxq_init[] = {
 				IBV_RX_HASH_SRC_PORT_UDP |
 				IBV_RX_HASH_DST_PORT_UDP),
 		.dpdk_rss_hf = ETH_RSS_NONFRAG_IPV4_UDP,
-		.flow_priority = 0,
+		.flow_priority = 1,
 		.ip_version = MLX5_IPV4,
 	},
 	[HASH_RXQ_IPV4] = {
@@ -174,7 +174,7 @@ const struct hash_rxq_init hash_rxq_init[] = {
 				IBV_RX_HASH_DST_IPV4),
 		.dpdk_rss_hf = (ETH_RSS_IPV4 |
 				ETH_RSS_FRAG_IPV4),
-		.flow_priority = 1,
+		.flow_priority = 2,
 		.ip_version = MLX5_IPV4,
 	},
 	[HASH_RXQ_TCPV6] = {
@@ -183,7 +183,7 @@ const struct hash_rxq_init hash_rxq_init[] = {
 				IBV_RX_HASH_SRC_PORT_TCP |
 				IBV_RX_HASH_DST_PORT_TCP),
 		.dpdk_rss_hf = ETH_RSS_NONFRAG_IPV6_TCP,
-		.flow_priority = 0,
+		.flow_priority = 1,
 		.ip_version = MLX5_IPV6,
 	},
 	[HASH_RXQ_UDPV6] = {
@@ -192,7 +192,7 @@ const struct hash_rxq_init hash_rxq_init[] = {
 				IBV_RX_HASH_SRC_PORT_UDP |
 				IBV_RX_HASH_DST_PORT_UDP),
 		.dpdk_rss_hf = ETH_RSS_NONFRAG_IPV6_UDP,
-		.flow_priority = 0,
+		.flow_priority = 1,
 		.ip_version = MLX5_IPV6,
 	},
 	[HASH_RXQ_IPV6] = {
@@ -200,13 +200,13 @@ const struct hash_rxq_init hash_rxq_init[] = {
 				IBV_RX_HASH_DST_IPV6),
 		.dpdk_rss_hf = (ETH_RSS_IPV6 |
 				ETH_RSS_FRAG_IPV6),
-		.flow_priority = 1,
+		.flow_priority = 2,
 		.ip_version = MLX5_IPV6,
 	},
 	[HASH_RXQ_ETH] = {
 		.hash_fields = 0,
 		.dpdk_rss_hf = 0,
-		.flow_priority = 2,
+		.flow_priority = 3,
 	},
 };
 
@@ -900,8 +900,6 @@ mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
 /**
  * Allocate memory space to store verbs flow attributes.
  *
- * @param[in] priority
- *   Flow priority.
  * @param[in] size
  *   Amount of byte to allocate.
  * @param[out] error
@@ -911,9 +909,7 @@ mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
  *   A verbs flow attribute on success, NULL otherwise and rte_errno is set.
  */
 static struct ibv_flow_attr *
-mlx5_flow_convert_allocate(unsigned int priority,
-			   unsigned int size,
-			   struct rte_flow_error *error)
+mlx5_flow_convert_allocate(unsigned int size, struct rte_flow_error *error)
 {
 	struct ibv_flow_attr *ibv_attr;
 
@@ -925,11 +921,41 @@ mlx5_flow_convert_allocate(unsigned int priority,
 				   "cannot allocate verbs spec attributes");
 		return NULL;
 	}
-	ibv_attr->priority = priority;
 	return ibv_attr;
 }
 
 /**
+ * Make inner packet matching with an higher priority from the non Inner
+ * matching.
+ *
+ * @param[in, out] parser
+ *   Internal parser structure.
+ * @param attr
+ *   User flow attribute.
+ */
+static void
+mlx5_flow_update_priority(struct mlx5_flow_parse *parser,
+			  const struct rte_flow_attr *attr)
+{
+	unsigned int i;
+
+	if (parser->drop) {
+		parser->queue[HASH_RXQ_ETH].ibv_attr->priority =
+			attr->priority +
+			hash_rxq_init[HASH_RXQ_ETH].flow_priority;
+		return;
+	}
+	for (i = 0; i != hash_rxq_init_n; ++i) {
+		if (parser->queue[i].ibv_attr) {
+			parser->queue[i].ibv_attr->priority =
+				attr->priority +
+				hash_rxq_init[i].flow_priority -
+				(parser->inner ? 1 : 0);
+		}
+	}
+}
+
+/**
  * Finalise verbs flow attributes.
  *
  * @param[in, out] parser
@@ -1104,20 +1130,16 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 	 * Allocate the memory space to store verbs specifications.
 	 */
 	if (parser->drop) {
+		unsigned int offset = parser->queue[HASH_RXQ_ETH].offset;
+
 		parser->queue[HASH_RXQ_ETH].ibv_attr =
-			mlx5_flow_convert_allocate
-			(attr->priority,
-			 parser->queue[HASH_RXQ_ETH].offset,
-			 error);
+			mlx5_flow_convert_allocate(offset, error);
 		if (!parser->queue[HASH_RXQ_ETH].ibv_attr)
 			goto exit_enomem;
 		parser->queue[HASH_RXQ_ETH].offset =
 			sizeof(struct ibv_flow_attr);
 	} else {
 		for (i = 0; i != hash_rxq_init_n; ++i) {
-			unsigned int priority =
-				attr->priority +
-				hash_rxq_init[i].flow_priority;
 			unsigned int offset;
 
 			if (!(parser->rss_conf.rss_hf &
@@ -1126,8 +1148,7 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 				continue;
 			offset = parser->queue[i].offset;
 			parser->queue[i].ibv_attr =
-				mlx5_flow_convert_allocate(priority,
-							   offset, error);
+				mlx5_flow_convert_allocate(offset, error);
 			if (!parser->queue[i].ibv_attr)
 				goto exit_enomem;
 			parser->queue[i].offset = sizeof(struct ibv_flow_attr);
@@ -1162,13 +1183,9 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 	 * Last step. Complete missing specification to reach the RSS
 	 * configuration.
 	 */
-	if (!parser->drop) {
+	if (!parser->drop)
 		mlx5_flow_convert_finalise(parser);
-	} else {
-		parser->queue[HASH_RXQ_ETH].ibv_attr->priority =
-			attr->priority +
-			hash_rxq_init[parser->layer].flow_priority;
-	}
+	mlx5_flow_update_priority(parser, attr);
 exit_free:
 	/* Only verification is expected, all resources should be released. */
 	if (!parser->create) {
-- 
2.11.0

[dpdk-stable] [PATCH v2 35/67] net/mlx5: improve flow error explanation

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit fca13017680e538c5fe01f44e69fe01ce1609181 ]

Fill the error context in conversion function to provide a better reason on
why it cannot be done to the user.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 72 ++++++++++++++++++++++++--------------------
 1 file changed, 40 insertions(+), 32 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index c150ad52b..c5111de7b 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -83,40 +83,46 @@ ibv_destroy_counter_set(struct ibv_counter_set *cs)
 extern const struct eth_dev_ops mlx5_dev_ops;
 extern const struct eth_dev_ops mlx5_dev_ops_isolate;
 
+/** Structure give to the conversion functions. */
+struct mlx5_flow_data {
+	struct mlx5_flow_parse *parser; /** Parser context. */
+	struct rte_flow_error *error; /** Error context. */
+};
+
 static int
 mlx5_flow_create_eth(const struct rte_flow_item *item,
 		     const void *default_mask,
-		     void *data);
+		     struct mlx5_flow_data *data);
 
 static int
 mlx5_flow_create_vlan(const struct rte_flow_item *item,
 		      const void *default_mask,
-		      void *data);
+		      struct mlx5_flow_data *data);
 
 static int
 mlx5_flow_create_ipv4(const struct rte_flow_item *item,
 		      const void *default_mask,
-		      void *data);
+		      struct mlx5_flow_data *data);
 
 static int
 mlx5_flow_create_ipv6(const struct rte_flow_item *item,
 		      const void *default_mask,
-		      void *data);
+		      struct mlx5_flow_data *data);
 
 static int
 mlx5_flow_create_udp(const struct rte_flow_item *item,
 		     const void *default_mask,
-		     void *data);
+		     struct mlx5_flow_data *data);
 
 static int
 mlx5_flow_create_tcp(const struct rte_flow_item *item,
 		     const void *default_mask,
-		     void *data);
+		     struct mlx5_flow_data *data);
 
 static int
 mlx5_flow_create_vxlan(const struct rte_flow_item *item,
 		       const void *default_mask,
-		       void *data);
+		       struct mlx5_flow_data *data);
 
 struct mlx5_flow_parse;
 
@@ -291,7 +297,7 @@ struct mlx5_flow_items {
 	 */
 	int (*convert)(const struct rte_flow_item *item,
 		       const void *default_mask,
-		       void *data);
+		       struct mlx5_flow_data *data);
 	/** Size in bytes of the destination structure. */
 	const unsigned int dst_sz;
 	/** List of possible following items.  */
@@ -1157,6 +1163,11 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 	/* Third step. Conversion parse, fill the specifications. */
 	parser->inner = 0;
 	for (; items->type != RTE_FLOW_ITEM_TYPE_END; ++items) {
+		struct mlx5_flow_data data = {
+			.parser = parser,
+			.error = error,
+		};
+
 		if (items->type == RTE_FLOW_ITEM_TYPE_VOID)
 			continue;
 		cur_item = &mlx5_flow_items[items->type];
@@ -1164,13 +1175,9 @@ mlx5_flow_convert(struct rte_eth_dev *dev,
 					(cur_item->default_mask ?
 					 cur_item->default_mask :
 					 cur_item->mask),
-					parser);
-		if (ret) {
-			rte_flow_error_set(error, rte_errno,
-					   RTE_FLOW_ERROR_TYPE_ITEM,
-					   items, "item not supported");
+					 &data);
+		if (ret)
 			goto exit_free;
-		}
 	}
 	if (parser->mark)
 		mlx5_flow_create_flag_mark(parser, parser->mark_id);
@@ -1263,11 +1270,11 @@ mlx5_flow_create_copy(struct mlx5_flow_parse *parser, void *src,
 static int
 mlx5_flow_create_eth(const struct rte_flow_item *item,
 		     const void *default_mask,
-		     void *data)
+		     struct mlx5_flow_data *data)
 {
 	const struct rte_flow_item_eth *spec = item->spec;
 	const struct rte_flow_item_eth *mask = item->mask;
-	struct mlx5_flow_parse *parser = (struct mlx5_flow_parse *)data;
+	struct mlx5_flow_parse *parser = data->parser;
 	const unsigned int eth_size = sizeof(struct ibv_flow_spec_eth);
 	struct ibv_flow_spec_eth eth = {
 		.type = parser->inner | IBV_FLOW_SPEC_ETH,
@@ -1315,11 +1322,11 @@ mlx5_flow_create_eth(const struct rte_flow_item *item,
 static int
 mlx5_flow_create_vlan(const struct rte_flow_item *item,
 		      const void *default_mask,
-		      void *data)
+		      struct mlx5_flow_data *data)
 {
 	const struct rte_flow_item_vlan *spec = item->spec;
 	const struct rte_flow_item_vlan *mask = item->mask;
-	struct mlx5_flow_parse *parser = (struct mlx5_flow_parse *)data;
+	struct mlx5_flow_parse *parser = data->parser;
 	struct ibv_flow_spec_eth *eth;
 	const unsigned int eth_size = sizeof(struct ibv_flow_spec_eth);
 
@@ -1358,11 +1365,11 @@ mlx5_flow_create_vlan(const struct rte_flow_item *item,
 static int
 mlx5_flow_create_ipv4(const struct rte_flow_item *item,
 		      const void *default_mask,
-		      void *data)
+		      struct mlx5_flow_data *data)
 {
 	const struct rte_flow_item_ipv4 *spec = item->spec;
 	const struct rte_flow_item_ipv4 *mask = item->mask;
-	struct mlx5_flow_parse *parser = (struct mlx5_flow_parse *)data;
+	struct mlx5_flow_parse *parser = data->parser;
 	unsigned int ipv4_size = sizeof(struct ibv_flow_spec_ipv4_ext);
 	struct ibv_flow_spec_ipv4_ext ipv4 = {
 		.type = parser->inner | IBV_FLOW_SPEC_IPV4_EXT,
@@ -1413,11 +1420,11 @@ mlx5_flow_create_ipv4(const struct rte_flow_item *item,
 static int
 mlx5_flow_create_ipv6(const struct rte_flow_item *item,
 		      const void *default_mask,
-		      void *data)
+		      struct mlx5_flow_data *data)
 {
 	const struct rte_flow_item_ipv6 *spec = item->spec;
 	const struct rte_flow_item_ipv6 *mask = item->mask;
-	struct mlx5_flow_parse *parser = (struct mlx5_flow_parse *)data;
+	struct mlx5_flow_parse *parser = data->parser;
 	unsigned int ipv6_size = sizeof(struct ibv_flow_spec_ipv6);
 	struct ibv_flow_spec_ipv6 ipv6 = {
 		.type = parser->inner | IBV_FLOW_SPEC_IPV6,
@@ -1472,11 +1479,11 @@ mlx5_flow_create_ipv6(const struct rte_flow_item *item,
 static int
 mlx5_flow_create_udp(const struct rte_flow_item *item,
 		     const void *default_mask,
-		     void *data)
+		     struct mlx5_flow_data *data)
 {
 	const struct rte_flow_item_udp *spec = item->spec;
 	const struct rte_flow_item_udp *mask = item->mask;
-	struct mlx5_flow_parse *parser = (struct mlx5_flow_parse *)data;
+	struct mlx5_flow_parse *parser = data->parser;
 	unsigned int udp_size = sizeof(struct ibv_flow_spec_tcp_udp);
 	struct ibv_flow_spec_tcp_udp udp = {
 		.type = parser->inner | IBV_FLOW_SPEC_UDP,
@@ -1521,11 +1528,11 @@ mlx5_flow_create_udp(const struct rte_flow_item *item,
 static int
 mlx5_flow_create_tcp(const struct rte_flow_item *item,
 		     const void *default_mask,
-		     void *data)
+		     struct mlx5_flow_data *data)
 {
 	const struct rte_flow_item_tcp *spec = item->spec;
 	const struct rte_flow_item_tcp *mask = item->mask;
-	struct mlx5_flow_parse *parser = (struct mlx5_flow_parse *)data;
+	struct mlx5_flow_parse *parser = data->parser;
 	unsigned int tcp_size = sizeof(struct ibv_flow_spec_tcp_udp);
 	struct ibv_flow_spec_tcp_udp tcp = {
 		.type = parser->inner | IBV_FLOW_SPEC_TCP,
@@ -1570,11 +1577,11 @@ mlx5_flow_create_tcp(const struct rte_flow_item *item,
 static int
 mlx5_flow_create_vxlan(const struct rte_flow_item *item,
 		       const void *default_mask,
-		       void *data)
+		       struct mlx5_flow_data *data)
 {
 	const struct rte_flow_item_vxlan *spec = item->spec;
 	const struct rte_flow_item_vxlan *mask = item->mask;
-	struct mlx5_flow_parse *parser = (struct mlx5_flow_parse *)data;
+	struct mlx5_flow_parse *parser = data->parser;
 	unsigned int size = sizeof(struct ibv_flow_spec_tunnel);
 	struct ibv_flow_spec_tunnel vxlan = {
 		.type = parser->inner | IBV_FLOW_SPEC_VXLAN_TUNNEL,
@@ -1605,10 +1612,11 @@ mlx5_flow_create_vxlan(const struct rte_flow_item *item,
 	 * before will also match this rule.
 	 * To avoid such situation, VNI 0 is currently refused.
 	 */
-	if (!vxlan.val.tunnel_id) {
-		rte_errno = EINVAL;
-		return -rte_errno;
-	}
+	if (!vxlan.val.tunnel_id)
+		return rte_flow_error_set(data->error, EINVAL,
+					  RTE_FLOW_ERROR_TYPE_ITEM,
+					  item,
+					  "VxLAN vni cannot be 0");
 	mlx5_flow_create_copy(parser, &vxlan, size);
 	return 0;
 }
-- 
2.11.0

[dpdk-stable] [PATCH v2 36/67] net/mlx5: refuse empty VLAN flow specification

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit 0b1edd21cd78c48e944489231707cbd9a59273ec ]

Verbs specification doesn't help to distinguish between packets having an
VLAN and those which do not have, this ends by having flow rule which does
not react as the user expects e.g.

 flow create 0 ingress pattern eth / vlan / end action queue index 0 / end
 flow create 0 ingress pattern eth / end action queue index 1 / end

are colliding in Verbs definition as in both rule are matching packets with
or without VLAN.
For this reason, the VLAN specification must not be empty, otherwise the
PMD has to refuse it.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index c5111de7b..444b92484 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -1344,9 +1344,18 @@ mlx5_flow_create_vlan(const struct rte_flow_item *item,
 			eth->val.vlan_tag = spec->tci;
 			eth->mask.vlan_tag = mask->tci;
 			eth->val.vlan_tag &= eth->mask.vlan_tag;
+			/*
+			 * From verbs perspective an empty VLAN is equivalent
+			 * to a packet without VLAN layer.
+			 */
+			if (!eth->mask.vlan_tag)
+				goto error;
 		}
+		return 0;
 	}
-	return 0;
+error:
+	return rte_flow_error_set(data->error, EINVAL, RTE_FLOW_ERROR_TYPE_ITEM,
+				  item, "VLAN cannot be empty");
 }
 
 /**
-- 
2.11.0

[dpdk-stable] [PATCH v2 37/67] net/mlx5: fix icc build

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ backported from upstream commit 7b2207afe86df58b571e722dcee14734e91cdb2e ]

Remove the second declaration of device_attr [1] inside the loop as well as
the query_device_ex() which has already been done outside of the loop.

[1] https://dpdk.org/ml/archives/dev/2018-March/091744.html

Fixes: 9a761de8ea14 ("net/mlx5: flow counter support")
Cc: stable@dpdk.org

Reported-by: Ferruh Yigit <ferruh.yigit@intel.com>
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
Tested-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
 drivers/net/mlx5/mlx5.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 853599c25..858a6c108 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -796,7 +796,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		struct rte_eth_dev *eth_dev = NULL;
 		struct ibv_device_attr_ex device_attr_ex;
 		struct ether_addr mac;
-		struct ibv_device_attr_ex device_attr;
 		struct mlx5_args args = {
 			.cqe_comp = MLX5_ARG_UNSET,
 			.txq_inline = MLX5_ARG_UNSET,
@@ -851,7 +850,6 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			err = ENODEV;
 			goto port_error;
 		}
-		ibv_query_device_ex(ctx, NULL, &device_attr);
 		/* Check port status. */
 		err = ibv_query_port(ctx, port, &port_attr);
 		if (err) {
-- 
2.11.0

[dpdk-stable] [PATCH v2 38/67] net/mlx5: setup RSS regardless of queue count

[Yongseok Koh]

From: Dahir Osman <dahir.osman@windriver.com>

[ upstream commit 66669155da323e6d8516c3b276888ee0f3cbd799 ]

In some environments it is desirable to have the NIC perform RSS
normally on the packet regardless of the number of queues configured.
The RSS hash result that is stored in the mbuf can then be used by
the application to make decisions about how to distribute workloads
to threads, secondary processes, or even virtual machines if the
application is a virtual switch.  This change to the mlx5 driver
aligns with how other drivers in the Intel family work.

Signed-off-by: Allain Legacy <allain.legacy@windriver.com>
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
Tested-by: Allain Legacy <allain.legacy@windriver.com>
---
 drivers/net/mlx5/mlx5_flow.c |  2 +-
 drivers/net/mlx5/mlx5_rss.c  | 10 ++++++++++
 drivers/net/mlx5/mlx5_rxq.c  |  3 ++-
 3 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 444b92484..9bf1c570c 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -980,7 +980,7 @@ mlx5_flow_convert_finalise(struct mlx5_flow_parse *parser)
 	unsigned int i;
 
 	/* Remove any other flow not matching the pattern. */
-	if (parser->queues_n == 1) {
+	if (parser->queues_n == 1 && !parser->rss_conf.rss_hf) {
 		for (i = 0; i != hash_rxq_init_n; ++i) {
 			if (i == HASH_RXQ_ETH)
 				continue;
diff --git a/drivers/net/mlx5/mlx5_rss.c b/drivers/net/mlx5/mlx5_rss.c
index 90682a308..daadfbe5d 100644
--- a/drivers/net/mlx5/mlx5_rss.c
+++ b/drivers/net/mlx5/mlx5_rss.c
@@ -70,6 +70,8 @@ mlx5_rss_hash_update(struct rte_eth_dev *dev,
 		     struct rte_eth_rss_conf *rss_conf)
 {
 	struct priv *priv = dev->data->dev_private;
+	unsigned int i;
+	unsigned int idx;
 
 	if (rss_conf->rss_hf & MLX5_RSS_HF_MASK) {
 		rte_errno = EINVAL;
@@ -87,6 +89,14 @@ mlx5_rss_hash_update(struct rte_eth_dev *dev,
 		priv->rss_conf.rss_key_len = rss_conf->rss_key_len;
 	}
 	priv->rss_conf.rss_hf = rss_conf->rss_hf;
+	/* Enable the RSS hash in all Rx queues. */
+	for (i = 0, idx = 0; idx != priv->rxqs_n; ++i) {
+		if (!(*priv->rxqs)[i])
+			continue;
+		(*priv->rxqs)[i]->rss_hash = !!rss_conf->rss_hf &&
+			!!(dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS);
+		++idx;
+	}
 	return 0;
 }
 
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 1b0a95e0a..17b70dc9a 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -1012,7 +1012,8 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		tmpl->rxq.crc_present ? "disabled" : "enabled",
 		tmpl->rxq.crc_present << 2);
 	/* Save port ID. */
-	tmpl->rxq.rss_hash = priv->rxqs_n > 1;
+	tmpl->rxq.rss_hash = !!priv->rss_conf.rss_hf &&
+		(!!(dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS));
 	tmpl->rxq.port_id = dev->data->port_id;
 	tmpl->priv = priv;
 	tmpl->rxq.mp = mp;
-- 
2.11.0

[dpdk-stable] [PATCH v2 39/67] net/mlx5: enforce RSS key length limitation

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ upstream commit a1572312f713f0b218063c6213499e36a8b47c86 ]

RSS hash key must be 40 Bytes long.

Cc: stable@dpdk.org

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_ethdev.c | 3 ++-
 drivers/net/mlx5/mlx5_rss.c    | 7 +++++++
 2 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 18e1e534c..1e869db52 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -334,7 +334,8 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
 	if (use_app_rss_key &&
 	    (dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key_len !=
 	     rss_hash_default_key_len)) {
-		/* MLX5 RSS only support 40bytes key. */
+		DRV_LOG(ERR, "port %u RSS key len must be %zu Bytes long",
+			dev->data->port_id, rss_hash_default_key_len);
 		rte_errno = EINVAL;
 		return -rte_errno;
 	}
diff --git a/drivers/net/mlx5/mlx5_rss.c b/drivers/net/mlx5/mlx5_rss.c
index daadfbe5d..029e0ec49 100644
--- a/drivers/net/mlx5/mlx5_rss.c
+++ b/drivers/net/mlx5/mlx5_rss.c
@@ -78,6 +78,13 @@ mlx5_rss_hash_update(struct rte_eth_dev *dev,
 		return -rte_errno;
 	}
 	if (rss_conf->rss_key && rss_conf->rss_key_len) {
+		if (rss_conf->rss_key_len != rss_hash_default_key_len) {
+			DRV_LOG(ERR,
+				"port %u RSS key len must be %zu Bytes long",
+				dev->data->port_id, rss_hash_default_key_len);
+			rte_errno = EINVAL;
+			return -rte_errno;
+		}
 		priv->rss_conf.rss_key = rte_realloc(priv->rss_conf.rss_key,
 						     rss_conf->rss_key_len, 0);
 		if (!priv->rss_conf.rss_key) {
-- 
2.11.0

[dpdk-stable] [PATCH v2 40/67] net/mlx5: fix RSS key length query

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ backported from upstream commit e7041f5529d129eef57d6ac9405a510b1df69ae4 ]

The RSS key length returned by rte_eth_dev_info_get command was taken
from the
PMD private structure. This structure initialization was done only after
the port configuration.

Considering Mellanox device supports only 40B long RSS key, reporting
the fixed number instead.

Fixes: 29c1d8bb3e79 ("net/mlx5: handle a single RSS hash key for all protocols")
Cc: stable@dpdk.org

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_ethdev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 1e869db52..e9e56a2e0 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -450,7 +450,7 @@ mlx5_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info)
 		info->if_index = if_nametoindex(ifname);
 	info->reta_size = priv->reta_idx_n ?
 		priv->reta_idx_n : priv->ind_table_max_size;
-	info->hash_key_size = priv->rss_conf.rss_key_len;
+	info->hash_key_size = rss_hash_default_key_len;
 	info->speed_capa = priv->link_speed_capa;
 	info->flow_type_rss_offloads = ~MLX5_RSS_HF_MASK;
 }
-- 
2.11.0

[dpdk-stable] [PATCH v2 41/67] net/mlx4: fix a typo in header file

[Yongseok Koh]

From: Rami Rosen <rami.rosen@intel.com>

[ upstream commit 4db261fc88b2d17848227dd06452967816a56db2 ]

This patch fixes a trivial typo in mlx4 header file.

Fixes: 3d555728c933 ("net/mlx4: separate Rx/Tx definitions")
Cc: stable@dpdk.org

Signed-off-by: Rami Rosen <rami.rosen@intel.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx4/mlx4_rxtx.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/mlx4/mlx4_rxtx.h b/drivers/net/mlx4/mlx4_rxtx.h
index 463df2b0b..5a390e856 100644
--- a/drivers/net/mlx4/mlx4_rxtx.h
+++ b/drivers/net/mlx4/mlx4_rxtx.h
@@ -107,7 +107,7 @@ struct txq_elt {
 	struct rte_mbuf *buf; /**< Buffer. */
 };
 
-/** Rx queue counters. */
+/** Tx queue counters. */
 struct mlx4_txq_stats {
 	unsigned int idx; /**< Mapping index. */
 	uint64_t opackets; /**< Total of successfully sent packets. */
-- 
2.11.0

[dpdk-stable] [PATCH v2 42/67] net/mlx5: remove 32-bit support

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit ebbb81eb27daca0a89ee8f228fcf141d9eb6ef1c ]

naddr variable was introduced in
commit 9a7fa9f76d9e ("net/mlx5: use vector types to speed up processing")
to avoid compilation errors on 32bits compilation, as x86_32 is no more
supported by rdma-core nor by MLNX_OFED, this variable becomes useless and
can be safely removed.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5_rxtx.c | 20 +++++++++-----------
 1 file changed, 9 insertions(+), 11 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 62ff1a553..9c571f160 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -410,7 +410,6 @@ mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 		unsigned int ds = 0;
 		unsigned int sg = 0; /* counter of additional segs attached. */
 		uintptr_t addr;
-		uint64_t naddr;
 		uint16_t pkt_inline_sz = MLX5_WQE_DWORD_SIZE + 2;
 		uint16_t tso_header_sz = 0;
 		uint16_t ehdr;
@@ -634,12 +633,12 @@ mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 			ds = 3;
 use_dseg:
 			/* Add the remaining packet as a simple ds. */
-			naddr = rte_cpu_to_be_64(addr);
+			addr = rte_cpu_to_be_64(addr);
 			*dseg = (rte_v128u32_t){
 				rte_cpu_to_be_32(length),
 				mlx5_tx_mb2mr(txq, buf),
-				naddr,
-				naddr >> 32,
+				addr,
+				addr >> 32,
 			};
 			++ds;
 			if (!segs_n)
@@ -673,12 +672,12 @@ mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 		total_length += length;
 #endif
 		/* Store segment information. */
-		naddr = rte_cpu_to_be_64(rte_pktmbuf_mtod(buf, uintptr_t));
+		addr = rte_cpu_to_be_64(rte_pktmbuf_mtod(buf, uintptr_t));
 		*dseg = (rte_v128u32_t){
 			rte_cpu_to_be_32(length),
 			mlx5_tx_mb2mr(txq, buf),
-			naddr,
-			naddr >> 32,
+			addr,
+			addr >> 32,
 		};
 		(*txq->elts)[++elts_head & elts_m] = buf;
 		++sg;
@@ -1379,7 +1378,6 @@ mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	do {
 		struct rte_mbuf *buf = *(pkts++);
 		uintptr_t addr;
-		uint64_t naddr;
 		unsigned int n;
 		unsigned int do_inline = 0; /* Whether inline is possible. */
 		uint32_t length;
@@ -1561,12 +1559,12 @@ mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 			for (n = 0; n * RTE_CACHE_LINE_SIZE < length; n++)
 				rte_prefetch2((void *)(addr +
 						n * RTE_CACHE_LINE_SIZE));
-			naddr = rte_cpu_to_be_64(addr);
+			addr = rte_cpu_to_be_64(addr);
 			*dseg = (rte_v128u32_t) {
 				rte_cpu_to_be_32(length),
 				mlx5_tx_mb2mr(txq, buf),
-				naddr,
-				naddr >> 32,
+				addr,
+				addr >> 32,
 			};
 			mpw.data.raw = (volatile void *)(dseg + 1);
 			mpw.total_len += (inl_pad + sizeof(*dseg));
-- 
2.11.0

[dpdk-stable] [PATCH v2 43/67] net/mlx5: remove excessive data prefetch

[Yongseok Koh]

[ upstream commit f84411be9e307d2723052a8ae8ee80a1fd9f559b ]

In Enhanced Multi-Packet Send (eMPW), entire packet data is prefetched to
LLC if it isn't inlined. Even though this helps reducing jitter when HW
fetches data by DMA, this can thresh the LLC with evicting precious data.
And if the size of queue is large and there are many queues, this might not
be effective. Also, if application runs on a remote node from the PCIe
link, it may not be helpful and can even cause bad results.

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5_rxtx.c | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 9c571f160..2e003aeac 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -1378,7 +1378,6 @@ mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	do {
 		struct rte_mbuf *buf = *(pkts++);
 		uintptr_t addr;
-		unsigned int n;
 		unsigned int do_inline = 0; /* Whether inline is possible. */
 		uint32_t length;
 		unsigned int segs_n = buf->nb_segs;
@@ -1555,11 +1554,8 @@ mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 					((uintptr_t)mpw.data.raw +
 					 inl_pad);
 			(*txq->elts)[elts_head++ & elts_m] = buf;
-			addr = rte_pktmbuf_mtod(buf, uintptr_t);
-			for (n = 0; n * RTE_CACHE_LINE_SIZE < length; n++)
-				rte_prefetch2((void *)(addr +
-						n * RTE_CACHE_LINE_SIZE));
-			addr = rte_cpu_to_be_64(addr);
+			addr = rte_cpu_to_be_64(rte_pktmbuf_mtod(buf,
+								 uintptr_t));
 			*dseg = (rte_v128u32_t) {
 				rte_cpu_to_be_32(length),
 				mlx5_tx_mb2mr(txq, buf),
-- 
2.11.0

[dpdk-stable] [PATCH v2 44/67] net/mlx5: fix link status initialization

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ backported from upstream commit a85a606ca501296c81c4b3a5e6b359c7ad7e0a77 ]

Following commit 7ba5320baa32 ("net/mlx5: fix link status behavior")
The initial link status is no longer set as part of the port start.

When LSC interrupts are enabled, ethdev layer reads the link status
directly from the device data instead of using the PMD callback.
This may cause application to query the link as down while in fact it was
already up before the DPDK application start (and no interrupt to fix
it).

Fixes: 7ba5320baa32 ("net/mlx5: fix link status behavior")
Cc: stable@dpdk.org

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 858a6c108..d7cf5a63f 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -1063,6 +1063,12 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		DRV_LOG(DEBUG, "port %u forcing Ethernet interface up",
 			eth_dev->data->port_id);
 		mlx5_set_link_up(eth_dev);
+		/*
+		 * Even though the interrupt handler is not installed yet,
+		 * interrupts will still trigger on the asyn_fd from
+		 * Verbs context returned by ibv_open_device().
+		 */
+		mlx5_link_update(eth_dev, 0);
 		continue;
 port_error:
 		if (priv)
-- 
2.11.0

[dpdk-stable] [PATCH v2 45/67] net/mlx4: fix RSS resource leak in case of error

[Yongseok Koh]

From: Adrien Mazarguil <adrien.mazarguil@6wind.com>

[ upstream commit cb43322fbd00c7f7c13ad2095fb0331bf8a5f262 ]

When memory cannot be allocated for a flow rule, its RSS context reference
is not dropped.

Fixes: 078b8b452e6b ("net/mlx4: add RSS flow rule action support")
Cc: stable@dpdk.org

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
 drivers/net/mlx4/mlx4_flow.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mlx4/mlx4_flow.c b/drivers/net/mlx4/mlx4_flow.c
index e8ebeebe3..daa44eba9 100644
--- a/drivers/net/mlx4/mlx4_flow.c
+++ b/drivers/net/mlx4/mlx4_flow.c
@@ -839,11 +839,14 @@ mlx4_flow_prepare(struct priv *priv,
 			},
 		};
 
-		if (!mlx4_zmallocv(__func__, vec, RTE_DIM(vec)))
+		if (!mlx4_zmallocv(__func__, vec, RTE_DIM(vec))) {
+			if (temp.rss)
+				mlx4_rss_put(temp.rss);
 			return rte_flow_error_set
 				(error, -rte_errno,
 				 RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
 				 "flow rule handle allocation failure");
+		}
 		/* Most fields will be updated by second pass. */
 		*flow = (struct rte_flow){
 			.ibv_attr = temp.ibv_attr,
-- 
2.11.0

[dpdk-stable] [PATCH v2 46/67] net/mlx5: fix RSS flow action bounds check

[Yongseok Koh]

From: Adrien Mazarguil <adrien.mazarguil@6wind.com>

[ upstream commit e68744e53ed209c0419925171c5aa510f6643adb ]

The number of queues provided by the application is not checked against
parser's supported maximum.

Fixes: 3d821d6fea40 ("net/mlx5: support RSS action flow rule")
Cc: stable@dpdk.org

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
 drivers/net/mlx5/mlx5_flow.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 9bf1c570c..26f61ebbe 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -44,6 +44,7 @@
 #pragma GCC diagnostic error "-Wpedantic"
 #endif
 
+#include <rte_common.h>
 #include <rte_ethdev.h>
 #include <rte_flow.h>
 #include <rte_flow_driver.h>
@@ -752,6 +753,14 @@ mlx5_flow_convert_actions(struct rte_eth_dev *dev,
 					return -rte_errno;
 				}
 			}
+			if (rss->num > RTE_DIM(parser->queues)) {
+				rte_flow_error_set(error, EINVAL,
+						   RTE_FLOW_ERROR_TYPE_ACTION,
+						   actions,
+						   "too many queues for RSS"
+						   " context");
+				return -rte_errno;
+			}
 			for (n = 0; n < rss->num; ++n) {
 				if (rss->queue[n] >= priv->rxqs_n) {
 					rte_flow_error_set(error, EINVAL,
-- 
2.11.0

[dpdk-stable] [PATCH v2 47/67] net/mlx5: fix invalid flow item check

[Yongseok Koh]

From: Xueming Li <xuemingl@mellanox.com>

[ upstream commit 2323cc3c2e68bad29cc25d3924dbe4c55d358499 ]

This patch fixed invalid flow item check.

Fixes: a6d83b6a9209 ("net/mlx5: standardize on negative errno values")

Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 26f61ebbe..a5e20dfe9 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -865,8 +865,10 @@ mlx5_flow_convert_items_validate(const struct rte_flow_item items[],
 				break;
 			}
 		}
-		if (!token)
+		if (!token) {
+			ret = -ENOTSUP;
 			goto exit_item_not_supported;
+		}
 		cur_item = token;
 		ret = mlx5_flow_item_validate(items,
 					      (const uint8_t *)cur_item->mask,
-- 
2.11.0

[dpdk-stable] [PATCH v2 48/67] net/mlx5: split L3/L4 in flow director

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit ca42b8a8b74956d179a25c94ccb7f5a516013da0 ]

This will help to bring back the mask handler which was removed when this
feature was rewritten on top of rte_flow.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 122 +++++++++++++++----------------------------
 1 file changed, 42 insertions(+), 80 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index a5e20dfe9..9591009b6 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -2721,53 +2721,10 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 		return -rte_errno;
 	}
 	attributes->queue.index = fdir_filter->action.rx_queue;
+	/* Handle L3. */
 	switch (fdir_filter->input.flow_type) {
 	case RTE_ETH_FLOW_NONFRAG_IPV4_UDP:
-		attributes->l3.ipv4.hdr = (struct ipv4_hdr){
-			.src_addr = input->flow.udp4_flow.ip.src_ip,
-			.dst_addr = input->flow.udp4_flow.ip.dst_ip,
-			.time_to_live = input->flow.udp4_flow.ip.ttl,
-			.type_of_service = input->flow.udp4_flow.ip.tos,
-			.next_proto_id = input->flow.udp4_flow.ip.proto,
-		};
-		attributes->l4.udp.hdr = (struct udp_hdr){
-			.src_port = input->flow.udp4_flow.src_port,
-			.dst_port = input->flow.udp4_flow.dst_port,
-		};
-		attributes->items[1] = (struct rte_flow_item){
-			.type = RTE_FLOW_ITEM_TYPE_IPV4,
-			.spec = &attributes->l3,
-			.mask = &attributes->l3,
-		};
-		attributes->items[2] = (struct rte_flow_item){
-			.type = RTE_FLOW_ITEM_TYPE_UDP,
-			.spec = &attributes->l4,
-			.mask = &attributes->l4,
-		};
-		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV4_TCP:
-		attributes->l3.ipv4.hdr = (struct ipv4_hdr){
-			.src_addr = input->flow.tcp4_flow.ip.src_ip,
-			.dst_addr = input->flow.tcp4_flow.ip.dst_ip,
-			.time_to_live = input->flow.tcp4_flow.ip.ttl,
-			.type_of_service = input->flow.tcp4_flow.ip.tos,
-			.next_proto_id = input->flow.tcp4_flow.ip.proto,
-		};
-		attributes->l4.tcp.hdr = (struct tcp_hdr){
-			.src_port = input->flow.tcp4_flow.src_port,
-			.dst_port = input->flow.tcp4_flow.dst_port,
-		};
-		attributes->items[1] = (struct rte_flow_item){
-			.type = RTE_FLOW_ITEM_TYPE_IPV4,
-			.spec = &attributes->l3,
-			.mask = &attributes->l3,
-		};
-		attributes->items[2] = (struct rte_flow_item){
-			.type = RTE_FLOW_ITEM_TYPE_TCP,
-			.spec = &attributes->l4,
-			.mask = &attributes->l4,
-		};
-		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV4_OTHER:
 		attributes->l3.ipv4.hdr = (struct ipv4_hdr){
 			.src_addr = input->flow.ip4_flow.src_ip,
@@ -2783,73 +2740,78 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV6_UDP:
+	case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
+	case RTE_ETH_FLOW_NONFRAG_IPV6_OTHER:
 		attributes->l3.ipv6.hdr = (struct ipv6_hdr){
 			.hop_limits = input->flow.udp6_flow.ip.hop_limits,
 			.proto = input->flow.udp6_flow.ip.proto,
 		};
 		memcpy(attributes->l3.ipv6.hdr.src_addr,
-		       input->flow.udp6_flow.ip.src_ip,
+		       input->flow.ipv6_flow.src_ip,
 		       RTE_DIM(attributes->l3.ipv6.hdr.src_addr));
 		memcpy(attributes->l3.ipv6.hdr.dst_addr,
-		       input->flow.udp6_flow.ip.dst_ip,
+		       input->flow.ipv6_flow.dst_ip,
 		       RTE_DIM(attributes->l3.ipv6.hdr.src_addr));
-		attributes->l4.udp.hdr = (struct udp_hdr){
-			.src_port = input->flow.udp6_flow.src_port,
-			.dst_port = input->flow.udp6_flow.dst_port,
-		};
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV6,
 			.spec = &attributes->l3,
 			.mask = &attributes->l3,
 		};
+		break;
+	default:
+		DRV_LOG(ERR, "port %u invalid flow type%d",
+			dev->data->port_id, fdir_filter->input.flow_type);
+		rte_errno = ENOTSUP;
+		return -rte_errno;
+	}
+	/* Handle L4. */
+	switch (fdir_filter->input.flow_type) {
+	case RTE_ETH_FLOW_NONFRAG_IPV4_UDP:
+		attributes->l4.udp.hdr = (struct udp_hdr){
+			.src_port = input->flow.udp4_flow.src_port,
+			.dst_port = input->flow.udp4_flow.dst_port,
+		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_UDP,
 			.spec = &attributes->l4,
 			.mask = &attributes->l4,
 		};
 		break;
-	case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
-		attributes->l3.ipv6.hdr = (struct ipv6_hdr){
-			.hop_limits = input->flow.tcp6_flow.ip.hop_limits,
-			.proto = input->flow.tcp6_flow.ip.proto,
+	case RTE_ETH_FLOW_NONFRAG_IPV4_TCP:
+		attributes->l4.tcp.hdr = (struct tcp_hdr){
+			.src_port = input->flow.tcp4_flow.src_port,
+			.dst_port = input->flow.tcp4_flow.dst_port,
 		};
-		memcpy(attributes->l3.ipv6.hdr.src_addr,
-		       input->flow.tcp6_flow.ip.src_ip,
-		       RTE_DIM(attributes->l3.ipv6.hdr.src_addr));
-		memcpy(attributes->l3.ipv6.hdr.dst_addr,
-		       input->flow.tcp6_flow.ip.dst_ip,
-		       RTE_DIM(attributes->l3.ipv6.hdr.src_addr));
+		attributes->items[2] = (struct rte_flow_item){
+			.type = RTE_FLOW_ITEM_TYPE_TCP,
+			.spec = &attributes->l4,
+			.mask = &attributes->l4,
+		};
+		break;
+	case RTE_ETH_FLOW_NONFRAG_IPV6_UDP:
+		attributes->l4.udp.hdr = (struct udp_hdr){
+			.src_port = input->flow.udp6_flow.src_port,
+			.dst_port = input->flow.udp6_flow.dst_port,
+		};
+		attributes->items[2] = (struct rte_flow_item){
+			.type = RTE_FLOW_ITEM_TYPE_UDP,
+			.spec = &attributes->l4,
+			.mask = &attributes->l4,
+		};
+		break;
+	case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
 		attributes->l4.tcp.hdr = (struct tcp_hdr){
 			.src_port = input->flow.tcp6_flow.src_port,
 			.dst_port = input->flow.tcp6_flow.dst_port,
 		};
-		attributes->items[1] = (struct rte_flow_item){
-			.type = RTE_FLOW_ITEM_TYPE_IPV6,
-			.spec = &attributes->l3,
-			.mask = &attributes->l3,
-		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_TCP,
 			.spec = &attributes->l4,
 			.mask = &attributes->l4,
 		};
 		break;
+	case RTE_ETH_FLOW_NONFRAG_IPV4_OTHER:
 	case RTE_ETH_FLOW_NONFRAG_IPV6_OTHER:
-		attributes->l3.ipv6.hdr = (struct ipv6_hdr){
-			.hop_limits = input->flow.ipv6_flow.hop_limits,
-			.proto = input->flow.ipv6_flow.proto,
-		};
-		memcpy(attributes->l3.ipv6.hdr.src_addr,
-		       input->flow.ipv6_flow.src_ip,
-		       RTE_DIM(attributes->l3.ipv6.hdr.src_addr));
-		memcpy(attributes->l3.ipv6.hdr.dst_addr,
-		       input->flow.ipv6_flow.dst_ip,
-		       RTE_DIM(attributes->l3.ipv6.hdr.src_addr));
-		attributes->items[1] = (struct rte_flow_item){
-			.type = RTE_FLOW_ITEM_TYPE_IPV6,
-			.spec = &attributes->l3,
-			.mask = &attributes->l3,
-		};
 		break;
 	default:
 		DRV_LOG(ERR, "port %u invalid flow type%d",
-- 
2.11.0

[dpdk-stable] [PATCH v2 49/67] net/mlx5: fix flow director mask

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit 3c0db1ab516498e96b42fa804a3cf14e0ccf7647 ]

During the transition to resurrect flow director on top of rte_flow, mask
handling was removed by mistake.

Fixes: 4c3e9bcdd52e ("net/mlx5: support flow director")
Cc: stable@dpdk.org

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 56 +++++++++++++++++++++++++++++++++++++-------
 1 file changed, 48 insertions(+), 8 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 9591009b6..46fc7c49e 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -483,9 +483,17 @@ struct mlx5_fdir {
 		struct rte_flow_item_ipv6 ipv6;
 	} l3;
 	union {
+		struct rte_flow_item_ipv4 ipv4;
+		struct rte_flow_item_ipv6 ipv6;
+	} l3_mask;
+	union {
 		struct rte_flow_item_udp udp;
 		struct rte_flow_item_tcp tcp;
 	} l4;
+	union {
+		struct rte_flow_item_udp udp;
+		struct rte_flow_item_tcp tcp;
+	} l4_mask;
 	struct rte_flow_action_queue queue;
 };
 
@@ -2687,6 +2695,8 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 {
 	struct priv *priv = dev->data->dev_private;
 	const struct rte_eth_fdir_input *input = &fdir_filter->input;
+	const struct rte_eth_fdir_masks *mask =
+		&dev->data->dev_conf.fdir_conf.mask;
 
 	/* Validate queue number. */
 	if (fdir_filter->action.rx_queue >= priv->rxqs_n) {
@@ -2733,29 +2743,43 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 			.type_of_service = input->flow.ip4_flow.tos,
 			.next_proto_id = input->flow.ip4_flow.proto,
 		};
+		attributes->l3_mask.ipv4.hdr = (struct ipv4_hdr){
+			.src_addr = mask->ipv4_mask.src_ip,
+			.dst_addr = mask->ipv4_mask.dst_ip,
+			.time_to_live = mask->ipv4_mask.ttl,
+			.type_of_service = mask->ipv4_mask.tos,
+			.next_proto_id = mask->ipv4_mask.proto,
+		};
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV4,
 			.spec = &attributes->l3,
-			.mask = &attributes->l3,
+			.mask = &attributes->l3_mask,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV6_UDP:
 	case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
 	case RTE_ETH_FLOW_NONFRAG_IPV6_OTHER:
 		attributes->l3.ipv6.hdr = (struct ipv6_hdr){
-			.hop_limits = input->flow.udp6_flow.ip.hop_limits,
-			.proto = input->flow.udp6_flow.ip.proto,
+			.hop_limits = input->flow.ipv6_flow.hop_limits,
+			.proto = input->flow.ipv6_flow.proto,
 		};
+
 		memcpy(attributes->l3.ipv6.hdr.src_addr,
 		       input->flow.ipv6_flow.src_ip,
 		       RTE_DIM(attributes->l3.ipv6.hdr.src_addr));
 		memcpy(attributes->l3.ipv6.hdr.dst_addr,
 		       input->flow.ipv6_flow.dst_ip,
 		       RTE_DIM(attributes->l3.ipv6.hdr.src_addr));
+		memcpy(attributes->l3_mask.ipv6.hdr.src_addr,
+		       mask->ipv6_mask.src_ip,
+		       RTE_DIM(attributes->l3_mask.ipv6.hdr.src_addr));
+		memcpy(attributes->l3_mask.ipv6.hdr.dst_addr,
+		       mask->ipv6_mask.dst_ip,
+		       RTE_DIM(attributes->l3_mask.ipv6.hdr.src_addr));
 		attributes->items[1] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_IPV6,
 			.spec = &attributes->l3,
-			.mask = &attributes->l3,
+			.mask = &attributes->l3_mask,
 		};
 		break;
 	default:
@@ -2771,10 +2795,14 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 			.src_port = input->flow.udp4_flow.src_port,
 			.dst_port = input->flow.udp4_flow.dst_port,
 		};
+		attributes->l4_mask.udp.hdr = (struct udp_hdr){
+			.src_port = mask->src_port_mask,
+			.dst_port = mask->dst_port_mask,
+		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_UDP,
 			.spec = &attributes->l4,
-			.mask = &attributes->l4,
+			.mask = &attributes->l4_mask,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV4_TCP:
@@ -2782,10 +2810,14 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 			.src_port = input->flow.tcp4_flow.src_port,
 			.dst_port = input->flow.tcp4_flow.dst_port,
 		};
+		attributes->l4_mask.tcp.hdr = (struct tcp_hdr){
+			.src_port = mask->src_port_mask,
+			.dst_port = mask->dst_port_mask,
+		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_TCP,
 			.spec = &attributes->l4,
-			.mask = &attributes->l4,
+			.mask = &attributes->l4_mask,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV6_UDP:
@@ -2793,10 +2825,14 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 			.src_port = input->flow.udp6_flow.src_port,
 			.dst_port = input->flow.udp6_flow.dst_port,
 		};
+		attributes->l4_mask.udp.hdr = (struct udp_hdr){
+			.src_port = mask->src_port_mask,
+			.dst_port = mask->dst_port_mask,
+		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_UDP,
 			.spec = &attributes->l4,
-			.mask = &attributes->l4,
+			.mask = &attributes->l4_mask,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
@@ -2804,10 +2840,14 @@ mlx5_fdir_filter_convert(struct rte_eth_dev *dev,
 			.src_port = input->flow.tcp6_flow.src_port,
 			.dst_port = input->flow.tcp6_flow.dst_port,
 		};
+		attributes->l4_mask.tcp.hdr = (struct tcp_hdr){
+			.src_port = mask->src_port_mask,
+			.dst_port = mask->dst_port_mask,
+		};
 		attributes->items[2] = (struct rte_flow_item){
 			.type = RTE_FLOW_ITEM_TYPE_TCP,
 			.spec = &attributes->l4,
-			.mask = &attributes->l4,
+			.mask = &attributes->l4_mask,
 		};
 		break;
 	case RTE_ETH_FLOW_NONFRAG_IPV4_OTHER:
-- 
2.11.0

[dpdk-stable] [PATCH v2 50/67] net/mlx5: fix flow director rule deletion crash

[Yongseok Koh]

From: Adrien Mazarguil <adrien.mazarguil@6wind.com>

[ upstream commit 6f2f4948b23643d279f38d38f988b122f393270f ]

Flow director rules matching traffic properties above layer 2 do not
target a fixed hash Rx queue (HASH_RXQ_ETH), it actually depends on the
highest protocol layer specified by each flow rule.

mlx5_fdir_filter_delete() makes this wrong assumption and causes a crash
when attempting to destroy flow rules with L3/L4 specifications.

Fixes: 4c3e9bcdd52e ("net/mlx5: support flow director")
Cc: stable@dpdk.org

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 46fc7c49e..425fbeb40 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -2954,13 +2954,13 @@ mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
 	if (parser.drop) {
 		struct ibv_flow_spec_action_drop *drop;
 
-		drop = (void *)((uintptr_t)parser.queue[HASH_RXQ_ETH].ibv_attr +
-				parser.queue[HASH_RXQ_ETH].offset);
+		drop = (void *)((uintptr_t)parser.queue[parser.layer].ibv_attr +
+				parser.queue[parser.layer].offset);
 		*drop = (struct ibv_flow_spec_action_drop){
 			.type = IBV_FLOW_SPEC_ACTION_DROP,
 			.size = sizeof(struct ibv_flow_spec_action_drop),
 		};
-		parser.queue[HASH_RXQ_ETH].ibv_attr->num_of_specs++;
+		parser.queue[parser.layer].ibv_attr->num_of_specs++;
 	}
 	TAILQ_FOREACH(flow, &priv->flows, next) {
 		struct ibv_flow_attr *attr;
@@ -2971,8 +2971,8 @@ mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
 		void *flow_spec;
 		unsigned int specs_n;
 
-		attr = parser.queue[HASH_RXQ_ETH].ibv_attr;
-		flow_attr = flow->frxq[HASH_RXQ_ETH].ibv_attr;
+		attr = parser.queue[parser.layer].ibv_attr;
+		flow_attr = flow->frxq[parser.layer].ibv_attr;
 		/* Compare first the attributes. */
 		if (memcmp(attr, flow_attr, sizeof(struct ibv_flow_attr)))
 			continue;
-- 
2.11.0

[dpdk-stable] [PATCH v2 51/67] net/mlx4: fix Rx resource leak in case of error

[Yongseok Koh]

From: Adrien Mazarguil <adrien.mazarguil@6wind.com>

[ backported from upstream commit 84a684862f502f2d885bc5fa112da26265d5a0f4 ]

When creation of a flow rule fails during dev_start(), the usage count of
the common RSS context is not decremented, which triggers an assertion
failure in debug mode during dev_close().

This is addressed by tracking the initialization status of the common RSS
context in order to add missing cleanup code.

A similar issue exists in mlx4_rxq_attach(), where usage count is
incremented on a Rx queue but not released in case of error. This may lead
to the above issue since RSS contexts created by flow rules attach
themselves to Rx queues, incrementing their usage count.

Fixes: 5697a4142107 ("net/mlx4: relax Rx queue configuration order")
Cc: stable@dpdk.org

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx4/mlx4.c     | 6 ++++--
 drivers/net/mlx4/mlx4.h     | 1 +
 drivers/net/mlx4/mlx4_rxq.c | 7 +++++++
 3 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index 26a0c85ad..4d7bd5f0f 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/mlx4.c
@@ -85,6 +85,8 @@ const char *pmd_mlx4_init_params[] = {
 	NULL,
 };
 
+static void mlx4_dev_stop(struct rte_eth_dev *dev);
+
 /**
  * DPDK callback for Ethernet device configuration.
  *
@@ -167,8 +169,7 @@ mlx4_dev_start(struct rte_eth_dev *dev)
 	dev->rx_pkt_burst = mlx4_rx_burst;
 	return 0;
 err:
-	/* Rollback. */
-	priv->started = 0;
+	mlx4_dev_stop(dev);
 	return ret;
 }
 
@@ -218,6 +219,7 @@ mlx4_dev_close(struct rte_eth_dev *dev)
 	dev->tx_pkt_burst = mlx4_tx_burst_removed;
 	rte_wmb();
 	mlx4_flow_clean(priv);
+	mlx4_rss_deinit(priv);
 	for (i = 0; i != dev->data->nb_rx_queues; ++i)
 		mlx4_rx_queue_release(dev->data->rx_queues[i]);
 	for (i = 0; i != dev->data->nb_tx_queues; ++i)
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index 82573d4ab..41d652ba8 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -126,6 +126,7 @@ struct priv {
 	uint32_t vf:1; /**< This is a VF device. */
 	uint32_t intr_alarm:1; /**< An interrupt alarm is scheduled. */
 	uint32_t isolated:1; /**< Toggle isolated mode. */
+	uint32_t rss_init:1; /**< Common RSS context is initialized. */
 	uint32_t hw_csum:1; /* Checksum offload is supported. */
 	uint32_t hw_csum_l2tun:1; /* Checksum support for L2 tunnels. */
 	struct rte_intr_handle intr_handle; /**< Port interrupt handle. */
diff --git a/drivers/net/mlx4/mlx4_rxq.c b/drivers/net/mlx4/mlx4_rxq.c
index 53313c56f..06030c2c5 100644
--- a/drivers/net/mlx4/mlx4_rxq.c
+++ b/drivers/net/mlx4/mlx4_rxq.c
@@ -363,6 +363,8 @@ mlx4_rss_init(struct priv *priv)
 	unsigned int i;
 	int ret;
 
+	if (priv->rss_init)
+		return 0;
 	/* Prepare range for RSS contexts before creating the first WQ. */
 	ret = mlx4dv_set_context_attr(priv->ctx,
 				      MLX4DV_SET_CTX_ATTR_LOG_WQS_RANGE_SZ,
@@ -444,6 +446,7 @@ mlx4_rss_init(struct priv *priv)
 		}
 		wq_num_prev = wq_num;
 	}
+	priv->rss_init = 1;
 	return 0;
 error:
 	ERROR("cannot initialize common RSS resources (queue %u): %s: %s",
@@ -472,6 +475,8 @@ mlx4_rss_deinit(struct priv *priv)
 {
 	unsigned int i;
 
+	if (!priv->rss_init)
+		return;
 	for (i = 0; i != priv->dev->data->nb_rx_queues; ++i) {
 		struct rxq *rxq = priv->dev->data->rx_queues[i];
 
@@ -480,6 +485,7 @@ mlx4_rss_deinit(struct priv *priv)
 			mlx4_rxq_detach(rxq);
 		}
 	}
+	priv->rss_init = 0;
 }
 
 /**
@@ -622,6 +628,7 @@ mlx4_rxq_attach(struct rxq *rxq)
 		claim_zero(ibv_destroy_wq(wq));
 	if (cq)
 		claim_zero(ibv_destroy_cq(cq));
+	--rxq->usecnt;
 	rte_errno = ret;
 	ERROR("error while attaching Rx queue %p: %s: %s",
 	      (void *)rxq, msg, strerror(ret));
-- 
2.11.0

[dpdk-stable] [PATCH v2 52/67] net/mlx5: fix ethtool link setting call order

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ upstream commit 7d2e32f76cfccdbc565738a9bfb9ca4512e7f238 ]

According to ethtool_link_setting API recommendation ETHTOOL_GLINKSETTINGS
should be called before ETHTOOL_GSET as the later one deprecated.

Fixes: f47ba80080ab ("net/mlx5: remove kernel version check")

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_ethdev.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index e9e56a2e0..5edc75114 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -692,9 +692,9 @@ mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete)
 	time_t start_time = time(NULL);
 
 	do {
-		ret = mlx5_link_update_unlocked_gset(dev, &dev_link);
+		ret = mlx5_link_update_unlocked_gs(dev, &dev_link);
 		if (ret)
-			ret = mlx5_link_update_unlocked_gs(dev, &dev_link);
+			ret = mlx5_link_update_unlocked_gset(dev, &dev_link);
 		if (ret == 0)
 			break;
 		/* Handle wait to complete situation. */
-- 
2.11.0

[dpdk-stable] [PATCH v2 53/67] net/mlx5: fix socket connection return value

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ upstream commit eac9cd58dec1b08d241657b1a81deebed9d4db9d ]

Upon success, mlx5_socket_connect should return the fd descriptor of the
primary process

Fixes: a6d83b6a9209 ("net/mlx5: standardize on negative errno values")

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5.c        | 2 +-
 drivers/net/mlx5/mlx5_socket.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index d7cf5a63f..ac43d80aa 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -827,7 +827,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 				goto error;
 			/* Receive command fd from primary process */
 			err = mlx5_socket_connect(eth_dev);
-			if (err)
+			if (err < 0)
 				goto error;
 			/* Remap UAR for Tx queues. */
 			err = mlx5_tx_uar_remap(eth_dev, err);
diff --git a/drivers/net/mlx5/mlx5_socket.c b/drivers/net/mlx5/mlx5_socket.c
index bdbd390d1..7ab310007 100644
--- a/drivers/net/mlx5/mlx5_socket.c
+++ b/drivers/net/mlx5/mlx5_socket.c
@@ -320,7 +320,7 @@ mlx5_socket_connect(struct rte_eth_dev *dev)
 	}
 	ret = *fd;
 	close(socket_fd);
-	return 0;
+	return ret;
 error:
 	if (socket_fd != -1)
 		close(socket_fd);
-- 
2.11.0

[dpdk-stable] [PATCH v2 54/67] net/mlx5: add data-plane debug message macro

[Yongseok Koh]

From: Xueming Li <xuemingl@mellanox.com>

[ backported from upstream commit 80f2d0ed7ff95fe6feedaa27740decd948d78235 ]

Reinstate DEBUG() macro to print out debug messages only if it is enabled
in compile time.

Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_utils.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/net/mlx5/mlx5_utils.h b/drivers/net/mlx5/mlx5_utils.h
index 6c85c0739..e48352f37 100644
--- a/drivers/net/mlx5/mlx5_utils.h
+++ b/drivers/net/mlx5/mlx5_utils.h
@@ -131,16 +131,22 @@ extern int mlx5_logtype;
 /* claim_zero() does not perform any check when debugging is disabled. */
 #ifndef NDEBUG
 
+#define DEBUG(...) DRV_LOG(DEBUG, __VA_ARGS__)
 #define claim_zero(...) assert((__VA_ARGS__) == 0)
 #define claim_nonzero(...) assert((__VA_ARGS__) != 0)
 
 #else /* NDEBUG */
 
+#define DEBUG(...) (void)0
 #define claim_zero(...) (__VA_ARGS__)
 #define claim_nonzero(...) (__VA_ARGS__)
 
 #endif /* NDEBUG */
 
+#define INFO(...) DRV_LOG(INFO, __VA_ARGS__)
+#define WARN(...) DRV_LOG(WARNING, __VA_ARGS__)
+#define ERROR(...) DRV_LOG(ERR, __VA_ARGS__)
+
 /* Convenience macros for accessing mbuf fields. */
 #define NEXT(m) ((m)->next)
 #define DATA_LEN(m) ((m)->data_len)
-- 
2.11.0

[dpdk-stable] [PATCH v2 55/67] net/mlx5: fix probe return value polarity

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ backported from upstream commit 012ad9944dfc470c17d8f555c89dadd5473e25d4 ]

mlx5 prefixed function returns a negative errno value.
the error handler on mlx5_pci_probe is doing the same.

Fixes: a6d83b6a9209 ("net/mlx5: standardize on negative errno values")

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5.c | 30 +++++++++++++++++++++---------
 1 file changed, 21 insertions(+), 9 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index ac43d80aa..bd21454cb 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -779,8 +779,9 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		cqe_comp = 0;
 	else
 		cqe_comp = 1;
-	if (ibv_query_device_ex(attr_ctx, NULL, &device_attr)) {
-		err = errno;
+	err = ibv_query_device_ex(attr_ctx, NULL, &device_attr);
+	if (err) {
+		DEBUG("ibv_query_device_ex() failed");
 		goto error;
 	}
 	DRV_LOG(INFO, "%u port(s) detected",
@@ -823,16 +824,22 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			eth_dev->device = &pci_dev->device;
 			eth_dev->dev_ops = &mlx5_dev_sec_ops;
 			err = mlx5_uar_init_secondary(eth_dev);
-			if (err)
+			if (err) {
+				err = rte_errno;
 				goto error;
+			}
 			/* Receive command fd from primary process */
 			err = mlx5_socket_connect(eth_dev);
-			if (err < 0)
+			if (err < 0) {
+				err = rte_errno;
 				goto error;
+			}
 			/* Remap UAR for Tx queues. */
 			err = mlx5_tx_uar_remap(eth_dev, err);
-			if (err)
+			if (err) {
+				err = rte_errno;
 				goto error;
+			}
 			/*
 			 * Ethdev pointer is still required as input since
 			 * the primary device is not accessible from the
@@ -901,12 +908,13 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		if (err) {
 			DRV_LOG(ERR, "failed to process device arguments: %s",
 				strerror(err));
+			err = rte_errno;
 			goto port_error;
 		}
 		mlx5_args_assign(priv, &args);
-		if (ibv_query_device_ex(ctx, NULL, &device_attr_ex)) {
+		err = ibv_query_device_ex(ctx, NULL, &device_attr_ex);
+		if (err) {
 			DRV_LOG(ERR, "ibv_query_device_ex() failed");
-			err = errno;
 			goto port_error;
 		}
 		priv->hw_csum =
@@ -1006,8 +1014,10 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		rte_eth_copy_pci_info(eth_dev, pci_dev);
 		eth_dev->device->driver = &mlx5_driver.driver;
 		err = mlx5_uar_init_primary(eth_dev);
-		if (err)
+		if (err) {
+			err = rte_errno;
 			goto port_error;
+		}
 		/* Configure the first MAC address by default. */
 		if (mlx5_get_mac(eth_dev, &mac.addr_bytes)) {
 			DRV_LOG(ERR,
@@ -1037,8 +1047,10 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 #endif
 		/* Get actual MTU if possible. */
 		err = mlx5_get_mtu(eth_dev, &priv->mtu);
-		if (err)
+		if (err) {
+			err = rte_errno;
 			goto port_error;
+		}
 		DRV_LOG(DEBUG, "port %u MTU is %u", eth_dev->data->port_id,
 			priv->mtu);
 		/*
-- 
2.11.0

[dpdk-stable] [PATCH v2 56/67] net/mlx5: fix flow validation

[Yongseok Koh]

From: Nélio Laranjeiro <nelio.laranjeiro@6wind.com>

[ upstream commit b7a7c97a40cf8c513ca4526c4734d7401907692c ]

Item spec and last are wrongly compared to the NIC capability causing a
validation failure when the mask is null.
This validation function should only verify the user is not configuring
unsupported matching fields.

Fixes: 2097d0d1e2cc ("net/mlx5: support basic flow items and actions")
Cc: stable@dpdk.org

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 77 ++++++++++++++------------------------------
 1 file changed, 25 insertions(+), 52 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 425fbeb40..f9a931438 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -504,7 +504,7 @@ struct ibv_spec_header {
 };
 
 /**
- * Check support for a given item.
+ * Check item is fully supported by the NIC matching capability.
  *
  * @param item[in]
  *   Item specification.
@@ -521,60 +521,33 @@ static int
 mlx5_flow_item_validate(const struct rte_flow_item *item,
 			const uint8_t *mask, unsigned int size)
 {
-	if (!item->spec && (item->mask || item->last)) {
-		rte_errno = EINVAL;
-		return -rte_errno;
-	}
-	if (item->spec && !item->mask) {
-		unsigned int i;
-		const uint8_t *spec = item->spec;
-
-		for (i = 0; i < size; ++i)
-			if ((spec[i] | mask[i]) != mask[i]) {
-				rte_errno = EINVAL;
-				return -rte_errno;
-			}
-	}
-	if (item->last && !item->mask) {
-		unsigned int i;
-		const uint8_t *spec = item->last;
-
-		for (i = 0; i < size; ++i)
-			if ((spec[i] | mask[i]) != mask[i]) {
-				rte_errno = EINVAL;
-				return -rte_errno;
-			}
-	}
-	if (item->mask) {
-		unsigned int i;
-		const uint8_t *spec = item->spec;
-
-		for (i = 0; i < size; ++i)
-			if ((spec[i] | mask[i]) != mask[i]) {
-				rte_errno = EINVAL;
-				return -rte_errno;
-			}
-	}
-	if (item->spec && item->last) {
-		uint8_t spec[size];
-		uint8_t last[size];
-		const uint8_t *apply = mask;
-		unsigned int i;
-		int ret;
+	unsigned int i;
+	const uint8_t *spec = item->spec;
+	const uint8_t *last = item->last;
+	const uint8_t *m = item->mask ? item->mask : mask;
 
-		if (item->mask)
-			apply = item->mask;
-		for (i = 0; i < size; ++i) {
-			spec[i] = ((const uint8_t *)item->spec)[i] & apply[i];
-			last[i] = ((const uint8_t *)item->last)[i] & apply[i];
-		}
-		ret = memcmp(spec, last, size);
-		if (ret != 0) {
-			rte_errno = EINVAL;
-			return -rte_errno;
-		}
+	if (!spec && (item->mask || last))
+		goto error;
+	if (!spec)
+		return 0;
+	/*
+	 * Single-pass check to make sure that:
+	 * - item->mask is supported, no bits are set outside mask.
+	 * - Both masked item->spec and item->last are equal (no range
+	 *   supported).
+	 */
+	for (i = 0; i < size; i++) {
+		if (!m[i])
+			continue;
+		if ((m[i] | mask[i]) != mask[i])
+			goto error;
+		if (last && ((spec[i] & m[i]) != (last[i] & m[i])))
+			goto error;
 	}
 	return 0;
+error:
+	rte_errno = ENOTSUP;
+	return -rte_errno;
 }
 
 /**
-- 
2.11.0

[dpdk-stable] [PATCH v2 57/67] net/mlx4: fix UDP flow rule limitation enforcement

[Yongseok Koh]

From: Adrien Mazarguil <adrien.mazarguil@6wind.com>

[ upstream commit bf959ec58b5d9127f187e98de6d2d0976c111360 ]

For some unknown reason, priorities do not have any effect on flow rules
that happen to match UDP destination ports. Those are always matched first
regardless.

This patch is a workaround that enforces this limitation at the PMD level;
such flow rules can only be created at the highest priority level for
correctness.

Fixes: a5171594fc3b ("net/mlx4: expose support for flow rule priorities")
Cc: stable@dpdk.org

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx4/mlx4_flow.c | 9 +++++++++
 drivers/net/mlx4/mlx4_flow.h | 1 +
 2 files changed, 10 insertions(+)

diff --git a/drivers/net/mlx4/mlx4_flow.c b/drivers/net/mlx4/mlx4_flow.c
index daa44eba9..472cadc92 100644
--- a/drivers/net/mlx4/mlx4_flow.c
+++ b/drivers/net/mlx4/mlx4_flow.c
@@ -379,6 +379,9 @@ mlx4_flow_merge_ipv4(struct rte_flow *flow,
  * Additional mlx4-specific constraints on supported fields:
  *
  * - No support for partial masks.
+ * - Due to HW/FW limitation, flow rule priority is not taken into account
+ *   when matching UDP destination ports, doing is therefore only supported
+ *   at the highest priority level (0).
  *
  * @param[in, out] flow
  *   Flow rule handle to update.
@@ -410,6 +413,11 @@ mlx4_flow_merge_udp(struct rte_flow *flow,
 		msg = "mlx4 does not support matching partial UDP fields";
 		goto error;
 	}
+	if (mask && mask->hdr.dst_port && flow->priority) {
+		msg = "combining UDP destination port matching with a nonzero"
+			" priority level is not supported";
+		goto error;
+	}
 	if (!flow->ibv_attr)
 		return 0;
 	++flow->ibv_attr->num_of_specs;
@@ -674,6 +682,7 @@ mlx4_flow_prepare(struct priv *priv,
 			 NULL, "only ingress is supported");
 fill:
 	proc = mlx4_flow_proc_item_list;
+	flow->priority = attr->priority;
 	/* Go over pattern. */
 	for (item = pattern; item->type; ++item) {
 		const struct mlx4_flow_proc_item *next = NULL;
diff --git a/drivers/net/mlx4/mlx4_flow.h b/drivers/net/mlx4/mlx4_flow.h
index 651fd37b6..5e1f9ea77 100644
--- a/drivers/net/mlx4/mlx4_flow.h
+++ b/drivers/net/mlx4/mlx4_flow.h
@@ -70,6 +70,7 @@ struct rte_flow {
 	uint32_t promisc:1; /**< This rule matches everything. */
 	uint32_t allmulti:1; /**< This rule matches all multicast traffic. */
 	uint32_t drop:1; /**< This rule drops packets. */
+	uint32_t priority; /**< Flow rule priority. */
 	struct mlx4_rss *rss; /**< Rx target. */
 };
 
-- 
2.11.0

[dpdk-stable] [PATCH v2 58/67] net/mlx5: fix double free on error handling

[Yongseok Koh]

From: Raslan Darawsheh <rasland@mellanox.com>

[ backported from upstream commit e9f4166014cb7abf8c74b913ff046487d0e7bea2 ]

When attr_ctx is NULL it will attempt to free the list of devices twice.
Avoid double freeing the list by directly going to error handling.

Fixes: 771fa900b73a ("mlx5: introduce new driver for Mellanox ConnectX-4 adapters")
Cc: stable@dpdk.org

Signed-off-by: Raslan Darawsheh <rasland@mellanox.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index bd21454cb..ef17ec9a1 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -742,18 +742,18 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 		break;
 	}
 	if (attr_ctx == NULL) {
-		ibv_free_device_list(list);
 		switch (err) {
 		case 0:
 			DRV_LOG(ERR,
 				"cannot access device, is mlx5_ib loaded?");
 			err = ENODEV;
-			goto error;
+			break;
 		case EINVAL:
 			DRV_LOG(ERR,
 				"cannot use device, are drivers up to date?");
-			goto error;
+			break;
 		}
+		goto error;
 	}
 	ibv_dev = list[i];
 	DRV_LOG(DEBUG, "device opened");
-- 
2.11.0

[dpdk-stable] [PATCH v2 59/67] net/mlx5: fix resource leak in case of error

[Yongseok Koh]

From: Raslan Darawsheh <rasland@mellanox.com>

[ backported from upstream commit 690de2850b26466ad5325fde3e1aff040e3868f8 ]

If something went wrong in mlx5_pci_prob the allocated eth dev
will cause a memory leak.

This commit release the eth dev that was previously allocated.

Fixes: 771fa900b73a ("mlx5: introduce new driver for Mellanox ConnectX-4 adapters")
Cc: stable@dpdk.org

Signed-off-by: Raslan Darawsheh <rasland@mellanox.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index ef17ec9a1..bb56bd22d 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -1089,6 +1089,8 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			claim_zero(ibv_dealloc_pd(pd));
 		if (ctx)
 			claim_zero(ibv_close_device(ctx));
+		if (eth_dev && rte_eal_process_type() == RTE_PROC_PRIMARY)
+			rte_eth_dev_release_port(eth_dev);
 		break;
 	}
 	/*
-- 
2.11.0

[dpdk-stable] [PATCH v2 60/67] net/mlx5: fix calculation of Tx TSO inline room size

[Yongseok Koh]

[ backported from upstream commit 95d7e115be89f071d672d9cfd5c0eb1b23fce5fc ]

rdma-core doesn't add up max_tso_header size to max_inline_data size. The
library takes bigger value between the two.

Fixes: 43e9d9794cde ("net/mlx5: support upstream rdma-core")
Cc: stable@dpdk.org

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
---
 drivers/net/mlx5/mlx5_txq.c | 11 -----------
 1 file changed, 11 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 4e54ff33d..7abfd565b 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -662,17 +662,6 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 					  priv->inline_max_packet_sz) +
 				  (RTE_CACHE_LINE_SIZE - 1)) /
 				 RTE_CACHE_LINE_SIZE) * RTE_CACHE_LINE_SIZE;
-		} else if (priv->tso) {
-			int inline_diff = tmpl->txq.max_inline - max_tso_inline;
-
-			/*
-			 * Adjust inline value as Verbs aggregates
-			 * tso_inline and txq_inline fields.
-			 */
-			tmpl->max_inline_data = inline_diff > 0 ?
-					       inline_diff *
-					       RTE_CACHE_LINE_SIZE :
-					       0;
 		} else {
 			tmpl->max_inline_data =
 				tmpl->txq.max_inline * RTE_CACHE_LINE_SIZE;
-- 
2.11.0

[dpdk-stable] [PATCH v2 61/67] net/mlx5: change device reference for secondary process

[Yongseok Koh]

[ backported from upstream commit df428ceef4fdbceacda8d50341c25ddd46a76a39 ]

rte_eth_devices[] is not shared between primary and secondary process, but
a static array to each process. The reverse pointer of device (priv->dev)
is invalid. Instead, priv has the pointer to shared data of the device,
  struct rte_eth_dev_data *dev_data;

Two macros are added,
  #define PORT_ID(priv) ((priv)->dev_data->port_id)
  #define ETH_DEV(priv) (&rte_eth_devices[PORT_ID(priv)])

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5.c      |  2 +-
 drivers/net/mlx5/mlx5.h      |  5 ++++-
 drivers/net/mlx5/mlx5_flow.c | 21 +++++++++------------
 drivers/net/mlx5/mlx5_mr.c   | 19 +++++++++----------
 drivers/net/mlx5/mlx5_rxq.c  | 38 +++++++++++++++++++-------------------
 drivers/net/mlx5/mlx5_txq.c  | 13 ++++++-------
 6 files changed, 48 insertions(+), 50 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index bb56bd22d..10ce33592 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -1008,7 +1008,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			goto port_error;
 		}
 		eth_dev->data->dev_private = priv;
-		priv->dev = eth_dev;
+		priv->dev_data = eth_dev->data;
 		eth_dev->data->mac_addrs = priv->mac;
 		eth_dev->device = &pci_dev->device;
 		rte_eth_copy_pci_info(eth_dev, pci_dev);
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 75aa853f2..5e6027b82 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -109,7 +109,7 @@ struct mlx5_verbs_alloc_ctx {
 };
 
 struct priv {
-	struct rte_eth_dev *dev; /* Ethernet device of master process. */
+	struct rte_eth_dev_data *dev_data;  /* Pointer to device data. */
 	struct ibv_context *ctx; /* Verbs context. */
 	struct ibv_device_attr_ex device_attr; /* Device properties. */
 	struct ibv_pd *pd; /* Protection Domain. */
@@ -170,6 +170,9 @@ struct priv {
 	/* Context for Verbs allocator. */
 };
 
+#define PORT_ID(priv) ((priv)->dev_data->port_id)
+#define ETH_DEV(priv) (&rte_eth_devices[PORT_ID(priv)])
+
 /* mlx5.c */
 
 int mlx5_getenv_int(const char *);
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index f9a931438..54e0dbc57 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -1725,7 +1725,7 @@ mlx5_flow_create_action_queue_drop(struct rte_eth_dev *dev,
 		parser->queue[HASH_RXQ_ETH].ibv_attr;
 	if (parser->count)
 		flow->cs = parser->cs;
-	if (!priv->dev->data->dev_started)
+	if (!dev->data->dev_started)
 		return 0;
 	parser->queue[HASH_RXQ_ETH].ibv_attr = NULL;
 	flow->frxq[HASH_RXQ_ETH].ibv_flow =
@@ -1776,7 +1776,6 @@ mlx5_flow_create_action_queue_rss(struct rte_eth_dev *dev,
 				  struct rte_flow *flow,
 				  struct rte_flow_error *error)
 {
-	struct priv *priv = dev->data->dev_private;
 	unsigned int i;
 
 	for (i = 0; i != hash_rxq_init_n; ++i) {
@@ -1787,7 +1786,7 @@ mlx5_flow_create_action_queue_rss(struct rte_eth_dev *dev,
 		flow->frxq[i].ibv_attr = parser->queue[i].ibv_attr;
 		parser->queue[i].ibv_attr = NULL;
 		hash_fields = hash_rxq_init[i].hash_fields;
-		if (!priv->dev->data->dev_started)
+		if (!dev->data->dev_started)
 			continue;
 		flow->frxq[i].hrxq =
 			mlx5_hrxq_get(dev,
@@ -1836,7 +1835,7 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 			      struct rte_flow *flow,
 			      struct rte_flow_error *error)
 {
-	struct priv *priv = dev->data->dev_private;
+	struct priv *priv __rte_unused = dev->data->dev_private;
 	int ret;
 	unsigned int i;
 	unsigned int flows_n = 0;
@@ -1849,7 +1848,7 @@ mlx5_flow_create_action_queue(struct rte_eth_dev *dev,
 		goto error;
 	if (parser->count)
 		flow->cs = parser->cs;
-	if (!priv->dev->data->dev_started)
+	if (!dev->data->dev_started)
 		return 0;
 	for (i = 0; i != hash_rxq_init_n; ++i) {
 		if (!flow->frxq[i].hrxq)
@@ -2642,9 +2641,9 @@ mlx5_flow_isolate(struct rte_eth_dev *dev,
 	}
 	priv->isolated = !!enable;
 	if (enable)
-		priv->dev->dev_ops = &mlx5_dev_ops_isolate;
+		dev->dev_ops = &mlx5_dev_ops_isolate;
 	else
-		priv->dev->dev_ops = &mlx5_dev_ops;
+		dev->dev_ops = &mlx5_dev_ops;
 	return 0;
 }
 
@@ -3032,11 +3031,10 @@ mlx5_fdir_filter_flush(struct rte_eth_dev *dev)
 static void
 mlx5_fdir_info_get(struct rte_eth_dev *dev, struct rte_eth_fdir_info *fdir_info)
 {
-	struct priv *priv = dev->data->dev_private;
 	struct rte_eth_fdir_masks *mask =
-		&priv->dev->data->dev_conf.fdir_conf.mask;
+		&dev->data->dev_conf.fdir_conf.mask;
 
-	fdir_info->mode = priv->dev->data->dev_conf.fdir_conf.mode;
+	fdir_info->mode = dev->data->dev_conf.fdir_conf.mode;
 	fdir_info->guarant_spc = 0;
 	rte_memcpy(&fdir_info->mask, mask, sizeof(fdir_info->mask));
 	fdir_info->max_flexpayload = 0;
@@ -3064,9 +3062,8 @@ static int
 mlx5_fdir_ctrl_func(struct rte_eth_dev *dev, enum rte_filter_op filter_op,
 		    void *arg)
 {
-	struct priv *priv = dev->data->dev_private;
 	enum rte_fdir_mode fdir_mode =
-		priv->dev->data->dev_conf.fdir_conf.mode;
+		dev->data->dev_conf.fdir_conf.mode;
 
 	if (filter_op == RTE_ETH_FILTER_NOP)
 		return 0;
diff --git a/drivers/net/mlx5/mlx5_mr.c b/drivers/net/mlx5/mlx5_mr.c
index 933bfe395..a50c52088 100644
--- a/drivers/net/mlx5/mlx5_mr.c
+++ b/drivers/net/mlx5/mlx5_mr.c
@@ -134,8 +134,8 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 	rte_spinlock_lock(&txq_ctrl->priv->mr_lock);
 	/* Add a new entry, register MR first. */
 	DRV_LOG(DEBUG, "port %u discovered new memory pool \"%s\" (%p)",
-		txq_ctrl->priv->dev->data->port_id, mp->name, (void *)mp);
-	dev = txq_ctrl->priv->dev;
+		PORT_ID(txq_ctrl->priv), mp->name, (void *)mp);
+	dev = ETH_DEV(txq_ctrl->priv);
 	mr = mlx5_mr_get(dev, mp);
 	if (mr == NULL) {
 		if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
@@ -143,8 +143,7 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 				"port %u using unregistered mempool 0x%p(%s)"
 				" in secondary process, please create mempool"
 				" before rte_eth_dev_start()",
-				txq_ctrl->priv->dev->data->port_id,
-				(void *)mp, mp->name);
+				PORT_ID(txq_ctrl->priv), (void *)mp, mp->name);
 			rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 			rte_errno = ENOTSUP;
 			return NULL;
@@ -155,7 +154,7 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 		DRV_LOG(DEBUG,
 			"port %u unable to configure memory region,"
 			" ibv_reg_mr() failed.",
-			txq_ctrl->priv->dev->data->port_id);
+			PORT_ID(txq_ctrl->priv));
 		rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 		return NULL;
 	}
@@ -164,7 +163,7 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 		DRV_LOG(DEBUG,
 			"port %u memory region <-> memory pool table full, "
 			" dropping oldest entry",
-			txq_ctrl->priv->dev->data->port_id);
+			PORT_ID(txq_ctrl->priv));
 		--idx;
 		mlx5_mr_release(txq->mp2mr[0]);
 		memmove(&txq->mp2mr[0], &txq->mp2mr[1],
@@ -175,7 +174,7 @@ mlx5_txq_mp2mr_reg(struct mlx5_txq_data *txq, struct rte_mempool *mp,
 	DRV_LOG(DEBUG,
 		"port %u new memory region lkey for MP \"%s\" (%p): 0x%08"
 		PRIu32,
-		txq_ctrl->priv->dev->data->port_id, mp->name, (void *)mp,
+		PORT_ID(txq_ctrl->priv), mp->name, (void *)mp,
 		txq_ctrl->txq.mp2mr[idx]->lkey);
 	rte_spinlock_unlock(&txq_ctrl->priv->mr_lock);
 	return mr;
@@ -236,15 +235,15 @@ mlx5_mp2mr_iter(struct rte_mempool *mp, void *arg)
 	if (rte_mempool_obj_iter(mp, txq_mp2mr_mbuf_check, &data) == 0 ||
 			data.ret == -1)
 		return;
-	mr = mlx5_mr_get(priv->dev, mp);
+	mr = mlx5_mr_get(ETH_DEV(priv), mp);
 	if (mr) {
 		mlx5_mr_release(mr);
 		return;
 	}
-	mr = mlx5_mr_new(priv->dev, mp);
+	mr = mlx5_mr_new(ETH_DEV(priv), mp);
 	if (!mr)
 		DRV_LOG(ERR, "port %u cannot create memory region: %s",
-			priv->dev->data->port_id, strerror(rte_errno));
+			PORT_ID(priv), strerror(rte_errno));
 }
 
 /**
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 17b70dc9a..dcc5a87b6 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -105,7 +105,7 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 		buf = rte_pktmbuf_alloc(rxq_ctrl->rxq.mp);
 		if (buf == NULL) {
 			DRV_LOG(ERR, "port %u empty mbuf pool",
-				rxq_ctrl->priv->dev->data->port_id);
+				PORT_ID(rxq_ctrl->priv));
 			rte_errno = ENOMEM;
 			goto error;
 		}
@@ -149,7 +149,7 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 	DRV_LOG(DEBUG,
 		"port %u Rx queue %u allocated and configured %u segments"
 		" (max %u packets)",
-		rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx, elts_n,
+		PORT_ID(rxq_ctrl->priv), rxq_ctrl->idx, elts_n,
 		elts_n / (1 << rxq_ctrl->rxq.sges_n));
 	return 0;
 error:
@@ -161,7 +161,7 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 		(*rxq_ctrl->rxq.elts)[i] = NULL;
 	}
 	DRV_LOG(DEBUG, "port %u Rx queue %u failed, freed everything",
-		rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
+		PORT_ID(rxq_ctrl->priv), rxq_ctrl->idx);
 	rte_errno = err; /* Restore rte_errno. */
 	return -rte_errno;
 }
@@ -182,7 +182,7 @@ rxq_free_elts(struct mlx5_rxq_ctrl *rxq_ctrl)
 	uint16_t i;
 
 	DRV_LOG(DEBUG, "port %u Rx queue %u freeing WRs",
-		rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
+		PORT_ID(rxq_ctrl->priv), rxq_ctrl->idx);
 	if (rxq->elts == NULL)
 		return;
 	/**
@@ -213,7 +213,7 @@ void
 mlx5_rxq_cleanup(struct mlx5_rxq_ctrl *rxq_ctrl)
 {
 	DRV_LOG(DEBUG, "port %u cleaning up Rx queue %u",
-		rxq_ctrl->priv->dev->data->port_id, rxq_ctrl->idx);
+		PORT_ID(rxq_ctrl->priv), rxq_ctrl->idx);
 	if (rxq_ctrl->ibv)
 		mlx5_rxq_ibv_release(rxq_ctrl->ibv);
 	memset(rxq_ctrl, 0, sizeof(*rxq_ctrl));
@@ -300,11 +300,11 @@ mlx5_rx_queue_release(void *dpdk_rxq)
 		return;
 	rxq_ctrl = container_of(rxq, struct mlx5_rxq_ctrl, rxq);
 	priv = rxq_ctrl->priv;
-	if (!mlx5_rxq_releasable(priv->dev, rxq_ctrl->rxq.stats.idx))
+	if (!mlx5_rxq_releasable(ETH_DEV(priv), rxq_ctrl->rxq.stats.idx))
 		rte_panic("port %u Rx queue %u is still used by a flow and"
-			  " cannot be removed\n", priv->dev->data->port_id,
-			  rxq_ctrl->idx);
-	mlx5_rxq_release(priv->dev, rxq_ctrl->rxq.stats.idx);
+			  " cannot be removed\n",
+			  PORT_ID(priv), rxq_ctrl->idx);
+	mlx5_rxq_release(ETH_DEV(priv), rxq_ctrl->rxq.stats.idx);
 }
 
 /**
@@ -324,9 +324,9 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev)
 	unsigned int rxqs_n = priv->rxqs_n;
 	unsigned int n = RTE_MIN(rxqs_n, (uint32_t)RTE_MAX_RXTX_INTR_VEC_ID);
 	unsigned int count = 0;
-	struct rte_intr_handle *intr_handle = priv->dev->intr_handle;
+	struct rte_intr_handle *intr_handle = dev->intr_handle;
 
-	if (!priv->dev->data->dev_conf.intr_conf.rxq)
+	if (!dev->data->dev_conf.intr_conf.rxq)
 		return 0;
 	mlx5_rx_intr_vec_disable(dev);
 	intr_handle->intr_vec = malloc(n * sizeof(intr_handle->intr_vec[0]));
@@ -398,12 +398,12 @@ void
 mlx5_rx_intr_vec_disable(struct rte_eth_dev *dev)
 {
 	struct priv *priv = dev->data->dev_private;
-	struct rte_intr_handle *intr_handle = priv->dev->intr_handle;
+	struct rte_intr_handle *intr_handle = dev->intr_handle;
 	unsigned int i;
 	unsigned int rxqs_n = priv->rxqs_n;
 	unsigned int n = RTE_MIN(rxqs_n, (uint32_t)RTE_MAX_RXTX_INTR_VEC_ID);
 
-	if (!priv->dev->data->dev_conf.intr_conf.rxq)
+	if (!dev->data->dev_conf.intr_conf.rxq)
 		return;
 	if (!intr_handle->intr_vec)
 		goto free;
@@ -842,7 +842,7 @@ mlx5_rxq_ibv_release(struct mlx5_rxq_ibv *rxq_ibv)
 	if (!ret)
 		rxq_ibv->mr = NULL;
 	DRV_LOG(DEBUG, "port %u Verbs Rx queue %u: refcnt %d",
-		rxq_ibv->rxq_ctrl->priv->dev->data->port_id,
+		PORT_ID(rxq_ibv->rxq_ctrl->priv),
 		rxq_ibv->rxq_ctrl->idx, rte_atomic32_read(&rxq_ibv->refcnt));
 	if (rte_atomic32_dec_and_test(&rxq_ibv->refcnt)) {
 		rxq_free_elts(rxq_ibv->rxq_ctrl);
@@ -928,7 +928,7 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 		return NULL;
 	}
 	tmpl->socket = socket;
-	if (priv->dev->data->dev_conf.intr_conf.rxq)
+	if (dev->data->dev_conf.intr_conf.rxq)
 		tmpl->irq = 1;
 	/* Enable scattered packets support for this queue if necessary. */
 	assert(mb_len >= RTE_PKTMBUF_HEADROOM);
@@ -1272,8 +1272,8 @@ mlx5_ind_table_ibv_release(struct rte_eth_dev *dev,
 	unsigned int i;
 
 	DRV_LOG(DEBUG, "port %u indirection table %p: refcnt %d",
-		((struct priv *)dev->data->dev_private)->port,
-		(void *)ind_tbl, rte_atomic32_read(&ind_tbl->refcnt));
+		dev->data->port_id, (void *)ind_tbl,
+		rte_atomic32_read(&ind_tbl->refcnt));
 	if (rte_atomic32_dec_and_test(&ind_tbl->refcnt))
 		claim_zero(ibv_destroy_rwq_ind_table(ind_tbl->ind_table));
 	for (i = 0; i != ind_tbl->queues_n; ++i)
@@ -1457,8 +1457,8 @@ int
 mlx5_hrxq_release(struct rte_eth_dev *dev, struct mlx5_hrxq *hrxq)
 {
 	DRV_LOG(DEBUG, "port %u hash Rx queue %p: refcnt %d",
-		((struct priv *)dev->data->dev_private)->port,
-		(void *)hrxq, rte_atomic32_read(&hrxq->refcnt));
+		dev->data->port_id, (void *)hrxq,
+		rte_atomic32_read(&hrxq->refcnt));
 	if (rte_atomic32_dec_and_test(&hrxq->refcnt)) {
 		claim_zero(ibv_destroy_qp(hrxq->qp));
 		mlx5_ind_table_ibv_release(dev, hrxq->ind_table);
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 7abfd565b..a5c6b5851 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -75,7 +75,7 @@ txq_alloc_elts(struct mlx5_txq_ctrl *txq_ctrl)
 	for (i = 0; (i != elts_n); ++i)
 		(*txq_ctrl->txq.elts)[i] = NULL;
 	DRV_LOG(DEBUG, "port %u Tx queue %u allocated and configured %u WRs",
-		txq_ctrl->priv->dev->data->port_id, txq_ctrl->idx, elts_n);
+		PORT_ID(txq_ctrl->priv), txq_ctrl->idx, elts_n);
 	txq_ctrl->txq.elts_head = 0;
 	txq_ctrl->txq.elts_tail = 0;
 	txq_ctrl->txq.elts_comp = 0;
@@ -97,7 +97,7 @@ txq_free_elts(struct mlx5_txq_ctrl *txq_ctrl)
 	struct rte_mbuf *(*elts)[elts_n] = txq_ctrl->txq.elts;
 
 	DRV_LOG(DEBUG, "port %u Tx queue %u freeing WRs",
-		txq_ctrl->priv->dev->data->port_id, txq_ctrl->idx);
+		PORT_ID(txq_ctrl->priv), txq_ctrl->idx);
 	txq_ctrl->txq.elts_head = 0;
 	txq_ctrl->txq.elts_tail = 0;
 	txq_ctrl->txq.elts_comp = 0;
@@ -205,9 +205,9 @@ mlx5_tx_queue_release(void *dpdk_txq)
 	priv = txq_ctrl->priv;
 	for (i = 0; (i != priv->txqs_n); ++i)
 		if ((*priv->txqs)[i] == txq) {
-			mlx5_txq_release(priv->dev, i);
+			mlx5_txq_release(ETH_DEV(priv), i);
 			DRV_LOG(DEBUG, "port %u removing Tx queue %u from list",
-				priv->dev->data->port_id, txq_ctrl->idx);
+				PORT_ID(priv), txq_ctrl->idx);
 			break;
 		}
 }
@@ -545,7 +545,7 @@ mlx5_txq_ibv_release(struct mlx5_txq_ibv *txq_ibv)
 {
 	assert(txq_ibv);
 	DRV_LOG(DEBUG, "port %u Verbs Tx queue %u: refcnt %d",
-		txq_ibv->txq_ctrl->priv->dev->data->port_id,
+		PORT_ID(txq_ibv->txq_ctrl->priv),
 		txq_ibv->txq_ctrl->idx, rte_atomic32_read(&txq_ibv->refcnt));
 	if (rte_atomic32_dec_and_test(&txq_ibv->refcnt)) {
 		claim_zero(ibv_destroy_qp(txq_ibv->qp));
@@ -684,8 +684,7 @@ mlx5_txq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
 			DRV_LOG(WARNING,
 				"port %u txq inline is too large (%d) setting it"
 				" to the maximum possible: %d\n",
-				priv->dev->data->port_id, priv->txq_inline,
-				max_inline);
+				PORT_ID(priv), priv->txq_inline, max_inline);
 			tmpl->txq.max_inline = max_inline / RTE_CACHE_LINE_SIZE;
 		}
 	}
-- 
2.11.0

[dpdk-stable] [PATCH v2 62/67] net/mlx4: avoid constant recreations in function

[Yongseok Koh]

From: Ophir Munk <ophirmu@mellanox.com>

[ upstream commit cbd737416c34c2f71227ac450db2c85a789dac30 ]

Function mlx4_conv_rss_types() contains constant arrays variables
which are recreated with every call to the function. By changing the
arrays definitions from "const" to "static const" these recreations
can be saved.

Signed-off-by: Ophir Munk <ophirmu@mellanox.com>
---
 drivers/net/mlx4/mlx4_flow.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/mlx4/mlx4_flow.c b/drivers/net/mlx4/mlx4_flow.c
index 472cadc92..fc0f06147 100644
--- a/drivers/net/mlx4/mlx4_flow.c
+++ b/drivers/net/mlx4/mlx4_flow.c
@@ -116,7 +116,7 @@ static uint64_t
 mlx4_conv_rss_hf(uint64_t rss_hf)
 {
 	enum { IPV4, IPV6, TCP, UDP, };
-	const uint64_t in[] = {
+	static const uint64_t in[] = {
 		[IPV4] = (ETH_RSS_IPV4 |
 			  ETH_RSS_FRAG_IPV4 |
 			  ETH_RSS_NONFRAG_IPV4_TCP |
@@ -139,7 +139,7 @@ mlx4_conv_rss_hf(uint64_t rss_hf)
 		 */
 		[UDP] = 0,
 	};
-	const uint64_t out[RTE_DIM(in)] = {
+	static const uint64_t out[RTE_DIM(in)] = {
 		[IPV4] = IBV_RX_HASH_SRC_IPV4 | IBV_RX_HASH_DST_IPV4,
 		[IPV6] = IBV_RX_HASH_SRC_IPV6 | IBV_RX_HASH_DST_IPV6,
 		[TCP] = IBV_RX_HASH_SRC_PORT_TCP | IBV_RX_HASH_DST_PORT_TCP,
-- 
2.11.0

[dpdk-stable] [PATCH v2 63/67] net/mlx5: fix flow director drop rule deletion crash

[Yongseok Koh]

From: Shahaf Shuler <shahafs@mellanox.com>

[ upstream commit 8fe576ad17a6f091c100dc1c61c11c5fb857f864 ]

Drop flow rules are created on the ETH queue even though the parser layer
matches the flow rule layer (L3/L4)

Fixes: 6f2f4948b236 ("net/mlx5: fix flow director rule deletion crash")
Cc: stable@dpdk.org

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 54e0dbc57..271c71e01 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -2926,13 +2926,13 @@ mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
 	if (parser.drop) {
 		struct ibv_flow_spec_action_drop *drop;
 
-		drop = (void *)((uintptr_t)parser.queue[parser.layer].ibv_attr +
-				parser.queue[parser.layer].offset);
+		drop = (void *)((uintptr_t)parser.queue[HASH_RXQ_ETH].ibv_attr +
+				parser.queue[HASH_RXQ_ETH].offset);
 		*drop = (struct ibv_flow_spec_action_drop){
 			.type = IBV_FLOW_SPEC_ACTION_DROP,
 			.size = sizeof(struct ibv_flow_spec_action_drop),
 		};
-		parser.queue[parser.layer].ibv_attr->num_of_specs++;
+		parser.queue[HASH_RXQ_ETH].ibv_attr->num_of_specs++;
 	}
 	TAILQ_FOREACH(flow, &priv->flows, next) {
 		struct ibv_flow_attr *attr;
@@ -2942,11 +2942,14 @@ mlx5_fdir_filter_delete(struct rte_eth_dev *dev,
 		struct ibv_spec_header *flow_h;
 		void *flow_spec;
 		unsigned int specs_n;
+		unsigned int queue_id = parser.drop ? HASH_RXQ_ETH :
+						      parser.layer;
 
-		attr = parser.queue[parser.layer].ibv_attr;
-		flow_attr = flow->frxq[parser.layer].ibv_attr;
+		attr = parser.queue[queue_id].ibv_attr;
+		flow_attr = flow->frxq[queue_id].ibv_attr;
 		/* Compare first the attributes. */
-		if (memcmp(attr, flow_attr, sizeof(struct ibv_flow_attr)))
+		if (!flow_attr ||
+		    memcmp(attr, flow_attr, sizeof(struct ibv_flow_attr)))
 			continue;
 		if (attr->num_of_specs == 0)
 			continue;
-- 
2.11.0

[dpdk-stable] [PATCH v2 64/67] net/mlx5: fix build with clang on ARM

[Yongseok Koh]

From: Gavin Hu <gavin.hu@arm.com>

[ upstream commit 74572f23cd6003d93007ebd3131d93833215463e ]

This patch adds a pair of "()" to embrace the argument
input to the function-like macro invocation.

drivers/net/mlx5/mlx5_rxtx_vec.c:37:
drivers/net/mlx5/mlx5_rxtx_vec_neon.h:170:24: error: too many arguments
provided to function-like macro invocation
	(uint16x8_t) { 0, 0, cs_flags, rte_cpu_to_be_16(len),

Fixes: 570acdb1da ("net/mlx5: add vectorized Rx/Tx burst for ARM")
Cc: stable@dpdk.org

Signed-off-by: Gavin Hu <gavin.hu@arm.com>
Signed-off-by: Sirshak Das <sirshak.das@arm.com>
Reviewed-by: Phil Yang <phil.yang@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5_rxtx_vec_neon.h | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
index 0b842f9a0..cf424778a 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
+++ b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
@@ -193,8 +193,8 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 		vst1q_u8((void *)t_wqe, ctrl);
 		/* Fill ESEG in the header. */
 		vst1q_u16((void *)(t_wqe + 1),
-			  (uint16x8_t) { 0, 0, cs_flags, rte_cpu_to_be_16(len),
-					 0, 0, 0, 0 });
+			  ((uint16x8_t) { 0, 0, cs_flags, rte_cpu_to_be_16(len),
+					  0, 0, 0, 0 }));
 		txq->wqe_ci = wqe_ci;
 	}
 	if (!n)
@@ -320,10 +320,10 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	vst1q_u8((void *)t_wqe, ctrl);
 	/* Fill ESEG in the header. */
 	vst1q_u8((void *)(t_wqe + 1),
-		 (uint8x16_t) { 0, 0, 0, 0,
-				cs_flags, 0, 0, 0,
-				0, 0, 0, 0,
-				0, 0, 0, 0 });
+		 ((uint8x16_t) { 0, 0, 0, 0,
+				 cs_flags, 0, 0, 0,
+				 0, 0, 0, 0,
+				 0, 0, 0, 0 }));
 #ifdef MLX5_PMD_SOFT_COUNTERS
 	txq->stats.opackets += pkts_n;
 #endif
-- 
2.11.0

[dpdk-stable] [PATCH v2 65/67] net/mlx5: fix count in xstats

[Yongseok Koh]

From: David Marchand <david.marchand@6wind.com>

[ upstream commit 5fd4d049692b2fde8bf49c7461b18180a8fd2545 ]

With the commit af4f09f28294 ("net/mlx5: prefix all functions with mlx5"),
mlx5_xstats_get() is not compliant any longer with the api.
It always returns the caller max entries count while it should return how
many entries it wrote/wanted to write.

Fixes: af4f09f28294 ("net/mlx5: prefix all functions with mlx5")

Signed-off-by: David Marchand <david.marchand@6wind.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_stats.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5_stats.c b/drivers/net/mlx5/mlx5_stats.c
index 7dda2691d..345ed7075 100644
--- a/drivers/net/mlx5/mlx5_stats.c
+++ b/drivers/net/mlx5/mlx5_stats.c
@@ -337,7 +337,7 @@ mlx5_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *stats,
 			stats[i].value = (counters[i] - xstats_ctrl->base[i]);
 		}
 	}
-	return n;
+	return xstats_n;
 }
 
 /**
-- 
2.11.0

[dpdk-stable] [PATCH v2 66/67] net: add IPv6 header fields macros

[Yongseok Koh]

From: Shachar Beiser <shacharbe@mellanox.com>

[ upstream commit 9b20c6e9028eb903ffc5e49cbf5c63e25a7c568f ]

Support IPv6 header vtc_flow fields : tc , flow_label

Signed-off-by: Shachar Beiser <shacharbe@mellanox.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
---
 lib/librte_net/rte_ip.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/lib/librte_net/rte_ip.h b/lib/librte_net/rte_ip.h
index 73ec398fc..23468cb99 100644
--- a/lib/librte_net/rte_ip.h
+++ b/lib/librte_net/rte_ip.h
@@ -406,6 +406,12 @@ struct ipv6_hdr {
 	uint8_t  dst_addr[16]; /**< IP address of destination host(s). */
 } __attribute__((__packed__));
 
+/* IPv6 vtc_flow: IPv / TC / flow_label */
+#define IPV6_HDR_FL_SHIFT 0
+#define IPV6_HDR_TC_SHIFT 20
+#define IPV6_HDR_FL_MASK ((1u << IPV6_HDR_TC_SHIFT) - 1)
+#define IPV6_HDR_TC_MASK (0xf << IPV6_HDR_TC_SHIFT)
+
 /**
  * Process the pseudo-header checksum of an IPv6 header.
  *
-- 
2.11.0

[dpdk-stable] [PATCH v2 67/67] net/mlx5: fix IPv6 header fields

[Yongseok Koh]

From: Shachar Beiser <shacharbe@mellanox.com>

[ upstream commit e8aafd9405b9f1da345abd2d5ff0d7aabc1ce6f3 ]

There are parameters that are not copy from
spec to verbs structure in the vtc_label

Fixes: 43e9d9794cde ("net/mlx5: support upstream rdma-core")
Cc: stable@dpdk.org

Signed-off-by: Shachar Beiser <shacharbe@mellanox.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5_flow.c | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 271c71e01..57b654c35 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -49,6 +49,7 @@
 #include <rte_flow.h>
 #include <rte_flow_driver.h>
 #include <rte_malloc.h>
+#include <rte_ip.h>
 
 #include "mlx5.h"
 #include "mlx5_defs.h"
@@ -1437,6 +1438,8 @@ mlx5_flow_create_ipv6(const struct rte_flow_item *item,
 		parser->layer = HASH_RXQ_IPV6;
 	if (spec) {
 		unsigned int i;
+		uint32_t vtc_flow_val;
+		uint32_t vtc_flow_mask;
 
 		if (!mask)
 			mask = default_mask;
@@ -1448,7 +1451,20 @@ mlx5_flow_create_ipv6(const struct rte_flow_item *item,
 		       RTE_DIM(ipv6.mask.src_ip));
 		memcpy(&ipv6.mask.dst_ip, mask->hdr.dst_addr,
 		       RTE_DIM(ipv6.mask.dst_ip));
-		ipv6.mask.flow_label = mask->hdr.vtc_flow;
+		vtc_flow_val = rte_be_to_cpu_32(spec->hdr.vtc_flow);
+		vtc_flow_mask = rte_be_to_cpu_32(mask->hdr.vtc_flow);
+		ipv6.val.flow_label =
+			rte_cpu_to_be_32((vtc_flow_val & IPV6_HDR_FL_MASK) >>
+					 IPV6_HDR_FL_SHIFT);
+		ipv6.val.traffic_class = (vtc_flow_val & IPV6_HDR_TC_MASK) >>
+					 IPV6_HDR_TC_SHIFT;
+		ipv6.val.next_hdr = spec->hdr.proto;
+		ipv6.val.hop_limit = spec->hdr.hop_limits;
+		ipv6.mask.flow_label =
+			rte_cpu_to_be_32((vtc_flow_mask & IPV6_HDR_FL_MASK) >>
+					 IPV6_HDR_FL_SHIFT);
+		ipv6.mask.traffic_class = (vtc_flow_mask & IPV6_HDR_TC_MASK) >>
+					  IPV6_HDR_TC_SHIFT;
 		ipv6.mask.next_hdr = mask->hdr.proto;
 		ipv6.mask.hop_limit = mask->hdr.hop_limits;
 		/* Remove unwanted bits from values. */
@@ -1457,6 +1473,7 @@ mlx5_flow_create_ipv6(const struct rte_flow_item *item,
 			ipv6.val.dst_ip[i] &= ipv6.mask.dst_ip[i];
 		}
 		ipv6.val.flow_label &= ipv6.mask.flow_label;
+		ipv6.val.traffic_class &= ipv6.mask.traffic_class;
 		ipv6.val.next_hdr &= ipv6.mask.next_hdr;
 		ipv6.val.hop_limit &= ipv6.mask.hop_limit;
 	}
-- 
2.11.0

Re: [dpdk-stable] [PATCH v2 00/67] net/mlx5: backport patches for v17.11.3 LTS

[Yuanhan Liu]

On Mon, Jun 04, 2018 at 05:27:23PM -0700, Yongseok Koh wrote:
> This patchset includes patches for mlx4 and mlx5. All the patches are from
> dpdk/master and properly backported based on stable/17.11.

Thank you, and series applied dpdk-stable/17.11.

	--yliu
> 
> v2:
> * no change, company mail server failed to send emails, just resend.
> 
> Adrien Mazarguil (6):
>   net/mlx: control netdevices through ioctl only
>   net/mlx4: fix RSS resource leak in case of error
>   net/mlx5: fix RSS flow action bounds check
>   net/mlx5: fix flow director rule deletion crash
>   net/mlx4: fix Rx resource leak in case of error
>   net/mlx4: fix UDP flow rule limitation enforcement
> 
> Dahir Osman (1):
>   net/mlx5: setup RSS regardless of queue count
> 
> David Marchand (1):
>   net/mlx5: fix count in xstats
> 
> Gavin Hu (1):
>   net/mlx5: fix build with clang on ARM
> 
> Moti Haimovsky (2):
>   net/mlx4: fix broadcast Rx
>   net/mlx4: fix removal detection of stopped port
> 
> Nélio Laranjeiro (26):
>   net/mlx5: remove get priv internal function
>   net/mlx5: fix flow director conversion
>   net/mlx5: fix reception of multiple MAC addresses
>   net/mlx5: fix sriov flag
>   net/mlx5: name parameters in function prototypes
>   net/mlx5: mark parameters with unused attribute
>   net/mlx5: normalize function prototypes
>   net/mlx5: add missing function documentation
>   net/mlx5: remove useless empty lines
>   net/mlx5: remove control path locks
>   net/mlx5: prefix all functions with mlx5
>   net/mlx5: change non failing function return values
>   net/mlx5: standardize on negative errno values
>   net/mlx5: use port id in PMD log
>   net/mlx5: use dynamic logging
>   net/mlx5: remove kernel version check
>   net/mlx5: fix link status behavior
>   net/mlx5: fix link status to use wait to complete
>   net/mlx5: change tunnel flow priority
>   net/mlx5: improve flow error explanation
>   net/mlx5: refuse empty VLAN flow specification
>   net/mlx5: fix icc build
>   net/mlx5: remove 32-bit support
>   net/mlx5: split L3/L4 in flow director
>   net/mlx5: fix flow director mask
>   net/mlx5: fix flow validation
> 
> Olivier Matz (1):
>   net/mlx5: fix allocation when no memory on device NUMA node
> 
> Ophir Munk (2):
>   net/mlx4: fix single port configuration
>   net/mlx4: avoid constant recreations in function
> 
> Rami Rosen (1):
>   net/mlx4: fix a typo in header file
> 
> Raslan Darawsheh (3):
>   net/mlx4: store RSS hash result in mbufs
>   net/mlx5: fix double free on error handling
>   net/mlx5: fix resource leak in case of error
> 
> Shachar Beiser (2):
>   net: add IPv6 header fields macros
>   net/mlx5: fix IPv6 header fields
> 
> Shahaf Shuler (12):
>   net/mlx5: fix secondary process mempool registration
>   net/mlx5: remove assert un-accessible from secondary process
>   net/mlx5: warn for unsuccessful memory registration
>   net/mlx5: fix CRC strip capability query
>   net/mlx5: change pkt burst select function prototype
>   net/mlx5: enforce RSS key length limitation
>   net/mlx5: fix RSS key length query
>   net/mlx5: fix link status initialization
>   net/mlx5: fix ethtool link setting call order
>   net/mlx5: fix socket connection return value
>   net/mlx5: fix probe return value polarity
>   net/mlx5: fix flow director drop rule deletion crash
> 
> Xueming Li (4):
>   net/mlx5: map UAR address around huge pages
>   net/mlx5: fix close after start failure
>   net/mlx5: fix invalid flow item check
>   net/mlx5: add data-plane debug message macro
> 
> Yongseok Koh (5):
>   net/mlx5: fix synchronization on polling Rx completions
>   net/mlx5: fix disabling Tx packet inlining
>   net/mlx5: remove excessive data prefetch
>   net/mlx5: fix calculation of Tx TSO inline room size
>   net/mlx5: change device reference for secondary process
> 
>  doc/guides/nics/mlx5.rst              |    6 +-
>  drivers/net/mlx4/mlx4.c               |   18 +-
>  drivers/net/mlx4/mlx4.h               |    3 +
>  drivers/net/mlx4/mlx4_ethdev.c        |  192 +-----
>  drivers/net/mlx4/mlx4_flow.c          |   34 +-
>  drivers/net/mlx4/mlx4_flow.h          |    1 +
>  drivers/net/mlx4/mlx4_intr.c          |   41 +-
>  drivers/net/mlx4/mlx4_rxq.c           |    7 +
>  drivers/net/mlx4/mlx4_rxtx.c          |    3 +-
>  drivers/net/mlx4/mlx4_rxtx.h          |    2 +-
>  drivers/net/mlx5/mlx5.c               |  525 +++++++++------
>  drivers/net/mlx5/mlx5.h               |  267 ++++----
>  drivers/net/mlx5/mlx5_defs.h          |   14 +-
>  drivers/net/mlx5/mlx5_ethdev.c        | 1018 +++++++++--------------------
>  drivers/net/mlx5/mlx5_flow.c          | 1132 ++++++++++++++++++---------------
>  drivers/net/mlx5/mlx5_mac.c           |   49 +-
>  drivers/net/mlx5/mlx5_mr.c            |  197 +++---
>  drivers/net/mlx5/mlx5_rss.c           |  176 +++--
>  drivers/net/mlx5/mlx5_rxmode.c        |   28 +-
>  drivers/net/mlx5/mlx5_rxq.c           |  712 ++++++++++++---------
>  drivers/net/mlx5/mlx5_rxtx.c          |   76 +--
>  drivers/net/mlx5/mlx5_rxtx.h          |  179 +++---
>  drivers/net/mlx5/mlx5_rxtx_vec.c      |   26 +-
>  drivers/net/mlx5/mlx5_rxtx_vec_neon.h |   65 +-
>  drivers/net/mlx5/mlx5_rxtx_vec_sse.h  |    2 +-
>  drivers/net/mlx5/mlx5_socket.c        |  167 +++--
>  drivers/net/mlx5/mlx5_stats.c         |  248 ++++----
>  drivers/net/mlx5/mlx5_trigger.c       |  258 ++++----
>  drivers/net/mlx5/mlx5_txq.c           |  379 ++++++-----
>  drivers/net/mlx5/mlx5_utils.h         |   29 +-
>  drivers/net/mlx5/mlx5_vlan.c          |  106 ++-
>  lib/librte_net/rte_ip.h               |    6 +
>  32 files changed, 2947 insertions(+), 3019 deletions(-)
> 
> -- 
> 2.11.0