[PATCH net 0/9] mlx4 fixes

[PATCH net 0/9] mlx4 fixes

[Tariq Toukan]

Hi Dave,

This patchset contains several bug fixes from the team to the
mlx4 Eth and core drivers.

Series generated against net commit:
9dd4aaef194e "MAINTAINERS: Update CPMAC email address"

Please push the following patch to -stable  >= 4.6 as well:
"net/mlx4_core: Fix to clean devlink resources"

Thanks,
Tariq.

Jack Morgenstein (1):
  net/mlx4_core: Fix deadlock when switching between polling and event
    fw commands

Kamal Heib (4):
  net/mlx4_en: Fix wrong indentation
  net/mlx4_en: Fix the return value of mlx4_en_dcbnl_set_all()
  net/mlx4_en: Fix the return value of mlx4_en_dcbnl_set_state()
  net/mlx4_core: Fix to clean devlink resources

Leon Romanovsky (1):
  net/mlx4_core: Use RCU to perform radix tree lookup for SRQ

Moshe Shemesh (1):
  net/mlx4_en: Fix panic on xmit while port is down

Tariq Toukan (2):
  net/mlx4_en: Add branch prediction hints in RX data-path
  net/mlx4_en: Fixes for DCBX

 drivers/net/ethernet/mellanox/mlx4/cmd.c       | 23 ++++++++---
 drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c | 57 ++++++++++++++------------
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 21 ++++------
 drivers/net/ethernet/mellanox/mlx4/en_rx.c     | 28 +++++++------
 drivers/net/ethernet/mellanox/mlx4/en_tx.c     | 12 +++---
 drivers/net/ethernet/mellanox/mlx4/main.c      |  3 ++
 drivers/net/ethernet/mellanox/mlx4/mlx4.h      |  2 +
 drivers/net/ethernet/mellanox/mlx4/mlx4_en.h   | 15 ++-----
 drivers/net/ethernet/mellanox/mlx4/port.c      |  4 +-
 drivers/net/ethernet/mellanox/mlx4/srq.c       | 14 +++----
 10 files changed, 92 insertions(+), 87 deletions(-)

-- 
1.8.3.1

[PATCH net 1/9] net/mlx4_en: Add branch prediction hints in RX data-path

[Tariq Toukan]

Add likely/unlikely hints to improve branch predictions
in the RX data-path.

Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_rx.c | 26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index 2040dad8611d..f9d14ef6adc1 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -72,7 +72,7 @@ static int mlx4_alloc_pages(struct mlx4_en_priv *priv,
 	}
 	dma = dma_map_page(priv->ddev, page, 0, PAGE_SIZE << order,
 			   frag_info->dma_dir);
-	if (dma_mapping_error(priv->ddev, dma)) {
+	if (unlikely(dma_mapping_error(priv->ddev, dma))) {
 		put_page(page);
 		return -ENOMEM;
 	}
@@ -108,7 +108,8 @@ static int mlx4_en_alloc_frags(struct mlx4_en_priv *priv,
 		    ring_alloc[i].page_size)
 			continue;
 
-		if (mlx4_alloc_pages(priv, &page_alloc[i], frag_info, gfp))
+		if (unlikely(mlx4_alloc_pages(priv, &page_alloc[i],
+					      frag_info, gfp)))
 			goto out;
 	}
 
@@ -583,7 +584,7 @@ static int mlx4_en_complete_rx_desc(struct mlx4_en_priv *priv,
 		frag_info = &priv->frag_info[nr];
 		if (length <= frag_info->frag_prefix_size)
 			break;
-		if (!frags[nr].page)
+		if (unlikely(!frags[nr].page))
 			goto fail;
 
 		dma = be64_to_cpu(rx_desc->data[nr].addr);
@@ -623,7 +624,7 @@ static struct sk_buff *mlx4_en_rx_skb(struct mlx4_en_priv *priv,
 	dma_addr_t dma;
 
 	skb = netdev_alloc_skb(priv->dev, SMALL_PACKET_SIZE + NET_IP_ALIGN);
-	if (!skb) {
+	if (unlikely(!skb)) {
 		en_dbg(RX_ERR, priv, "Failed allocating skb\n");
 		return NULL;
 	}
@@ -734,7 +735,8 @@ static int get_fixed_ipv6_csum(__wsum hw_checksum, struct sk_buff *skb,
 {
 	__wsum csum_pseudo_hdr = 0;
 
-	if (ipv6h->nexthdr == IPPROTO_FRAGMENT || ipv6h->nexthdr == IPPROTO_HOPOPTS)
+	if (unlikely(ipv6h->nexthdr == IPPROTO_FRAGMENT ||
+		     ipv6h->nexthdr == IPPROTO_HOPOPTS))
 		return -1;
 	hw_checksum = csum_add(hw_checksum, (__force __wsum)htons(ipv6h->nexthdr));
 
@@ -767,7 +769,7 @@ static int check_csum(struct mlx4_cqe *cqe, struct sk_buff *skb, void *va,
 		get_fixed_ipv4_csum(hw_checksum, skb, hdr);
 #if IS_ENABLED(CONFIG_IPV6)
 	else if (cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPV6))
-		if (get_fixed_ipv6_csum(hw_checksum, skb, hdr))
+		if (unlikely(get_fixed_ipv6_csum(hw_checksum, skb, hdr)))
 			return -1;
 #endif
 	return 0;
@@ -794,10 +796,10 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 	u64 timestamp;
 	bool l2_tunnel;
 
-	if (!priv->port_up)
+	if (unlikely(!priv->port_up))
 		return 0;
 
-	if (budget <= 0)
+	if (unlikely(budget <= 0))
 		return polled;
 
 	xdp_prog = READ_ONCE(ring->xdp_prog);
@@ -902,16 +904,16 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 			case XDP_PASS:
 				break;
 			case XDP_TX:
-				if (!mlx4_en_xmit_frame(frags, dev,
+				if (likely(!mlx4_en_xmit_frame(frags, dev,
 							length, tx_index,
-							&doorbell_pending))
+							&doorbell_pending)))
 					goto consumed;
 				break;
 			default:
 				bpf_warn_invalid_xdp_action(act);
 			case XDP_ABORTED:
 			case XDP_DROP:
-				if (mlx4_en_rx_recycle(ring, frags))
+				if (likely(mlx4_en_rx_recycle(ring, frags)))
 					goto consumed;
 				goto next;
 			}
@@ -1015,7 +1017,7 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 
 		/* GRO not possible, complete processing here */
 		skb = mlx4_en_rx_skb(priv, rx_desc, frags, length);
-		if (!skb) {
+		if (unlikely(!skb)) {
 			ring->dropped++;
 			goto next;
 		}
-- 
1.8.3.1

[PATCH net 2/9] net/mlx4_en: Fix wrong indentation

[Tariq Toukan]

From: Kamal Heib <kamalh@mellanox.com>

Use tabs instead of spaces before if statement, no functional change.

Fixes: e7c1c2c46201 ("mlx4_en: Added self diagnostics test implementation")
Signed-off-by: Kamal Heib <kamalh@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_rx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index f9d14ef6adc1..1eae305866d7 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -1022,7 +1022,7 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 			goto next;
 		}
 
-                if (unlikely(priv->validate_loopback)) {
+		if (unlikely(priv->validate_loopback)) {
 			validate_loopback(priv, skb);
 			goto next;
 		}
-- 
1.8.3.1

[PATCH net 3/9] net/mlx4_en: Fix the return value of mlx4_en_dcbnl_set_all()

[Tariq Toukan]

From: Kamal Heib <kamalh@mellanox.com>

mlx4_en_dcbnl_set_all() returns u8, so return value can't be negative in
case of failure.

Fixes: af7d51852631 ("net/mlx4_en: Add DCB PFC support through CEE netlink commands")
Signed-off-by: Kamal Heib <kamalh@mellanox.com>
Signed-off-by: Rana Shahout <ranas@mellanox.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
index 99c6bbdff501..97081e5bafd1 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
@@ -158,10 +158,9 @@ static u8 mlx4_en_dcbnl_set_all(struct net_device *netdev)
 	struct mlx4_en_priv *priv = netdev_priv(netdev);
 	struct mlx4_en_dev *mdev = priv->mdev;
 	struct mlx4_en_cee_config *dcb_cfg = &priv->cee_params.dcb_cfg;
-	int err = 0;
 
 	if (!(priv->cee_params.dcbx_cap & DCB_CAP_DCBX_VER_CEE))
-		return -EINVAL;
+		return 1;
 
 	if (dcb_cfg->pfc_state) {
 		int tc;
@@ -199,15 +198,17 @@ static u8 mlx4_en_dcbnl_set_all(struct net_device *netdev)
 		en_dbg(DRV, priv, "Set pfc off\n");
 	}
 
-	err = mlx4_SET_PORT_general(mdev->dev, priv->port,
-				    priv->rx_skb_size + ETH_FCS_LEN,
-				    priv->prof->tx_pause,
-				    priv->prof->tx_ppp,
-				    priv->prof->rx_pause,
-				    priv->prof->rx_ppp);
-	if (err)
+	if (mlx4_SET_PORT_general(mdev->dev, priv->port,
+				  priv->rx_skb_size + ETH_FCS_LEN,
+				  priv->prof->tx_pause,
+				  priv->prof->tx_ppp,
+				  priv->prof->rx_pause,
+				  priv->prof->rx_ppp)) {
 		en_err(priv, "Failed setting pause params\n");
-	return err;
+		return 1;
+	}
+
+	return 0;
 }
 
 static u8 mlx4_en_dcbnl_get_state(struct net_device *dev)
-- 
1.8.3.1

[PATCH net 4/9] net/mlx4_en: Fix the return value of mlx4_en_dcbnl_set_state()

[Tariq Toukan]

From: Kamal Heib <kamalh@mellanox.com>

mlx4_en_dcbnl_set_state() returns u8, the return value from
mlx4_en_setup_tc() could be negative in case of failure, so fix that.

Fixes: af7d51852631 ("net/mlx4_en: Add DCB PFC support through CEE netlink commands")
Signed-off-by: Kamal Heib <kamalh@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
index 97081e5bafd1..316a70714434 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
@@ -239,7 +239,10 @@ static u8 mlx4_en_dcbnl_set_state(struct net_device *dev, u8 state)
 		priv->flags &= ~MLX4_EN_FLAG_DCB_ENABLED;
 	}
 
-	return mlx4_en_setup_tc(dev, num_tcs);
+	if (mlx4_en_setup_tc(dev, num_tcs))
+		return 1;
+
+	return 0;
 }
 
 /* On success returns a non-zero 802.1p user priority bitmap
-- 
1.8.3.1

[PATCH net 5/9] net/mlx4_en: Fixes for DCBX

[Tariq Toukan]

This patch adds a capability check before enabling DCBX.
In addition, it re-organizes the relevant data structures,
and fixes a typo in a define.

Fixes: af7d51852631 ("net/mlx4_en: Add DCB PFC support through CEE netlink commands")
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c | 31 +++++++++++++-------------
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 21 +++++++----------
 drivers/net/ethernet/mellanox/mlx4/mlx4_en.h   | 15 +++----------
 drivers/net/ethernet/mellanox/mlx4/port.c      |  4 ++--
 4 files changed, 28 insertions(+), 43 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
index 316a70714434..b04760a5034b 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
@@ -94,7 +94,7 @@ static u8 mlx4_en_dcbnl_getcap(struct net_device *dev, int capid, u8 *cap)
 		*cap = true;
 		break;
 	case DCB_CAP_ATTR_DCBX:
-		*cap = priv->cee_params.dcbx_cap;
+		*cap = priv->dcbx_cap;
 		break;
 	case DCB_CAP_ATTR_PFC_TCS:
 		*cap = 1 <<  mlx4_max_tc(priv->mdev->dev);
@@ -111,14 +111,14 @@ static u8 mlx4_en_dcbnl_getpfcstate(struct net_device *netdev)
 {
 	struct mlx4_en_priv *priv = netdev_priv(netdev);
 
-	return priv->cee_params.dcb_cfg.pfc_state;
+	return priv->cee_config.pfc_state;
 }
 
 static void mlx4_en_dcbnl_setpfcstate(struct net_device *netdev, u8 state)
 {
 	struct mlx4_en_priv *priv = netdev_priv(netdev);
 
-	priv->cee_params.dcb_cfg.pfc_state = state;
+	priv->cee_config.pfc_state = state;
 }
 
 static void mlx4_en_dcbnl_get_pfc_cfg(struct net_device *netdev, int priority,
@@ -126,7 +126,7 @@ static void mlx4_en_dcbnl_get_pfc_cfg(struct net_device *netdev, int priority,
 {
 	struct mlx4_en_priv *priv = netdev_priv(netdev);
 
-	*setting = priv->cee_params.dcb_cfg.tc_config[priority].dcb_pfc;
+	*setting = priv->cee_config.dcb_pfc[priority];
 }
 
 static void mlx4_en_dcbnl_set_pfc_cfg(struct net_device *netdev, int priority,
@@ -134,8 +134,8 @@ static void mlx4_en_dcbnl_set_pfc_cfg(struct net_device *netdev, int priority,
 {
 	struct mlx4_en_priv *priv = netdev_priv(netdev);
 
-	priv->cee_params.dcb_cfg.tc_config[priority].dcb_pfc = setting;
-	priv->cee_params.dcb_cfg.pfc_state = true;
+	priv->cee_config.dcb_pfc[priority] = setting;
+	priv->cee_config.pfc_state = true;
 }
 
 static int mlx4_en_dcbnl_getnumtcs(struct net_device *netdev, int tcid, u8 *num)
@@ -157,12 +157,11 @@ static u8 mlx4_en_dcbnl_set_all(struct net_device *netdev)
 {
 	struct mlx4_en_priv *priv = netdev_priv(netdev);
 	struct mlx4_en_dev *mdev = priv->mdev;
-	struct mlx4_en_cee_config *dcb_cfg = &priv->cee_params.dcb_cfg;
 
-	if (!(priv->cee_params.dcbx_cap & DCB_CAP_DCBX_VER_CEE))
+	if (!(priv->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
 		return 1;
 
-	if (dcb_cfg->pfc_state) {
+	if (priv->cee_config.pfc_state) {
 		int tc;
 
 		priv->prof->rx_pause = 0;
@@ -170,7 +169,7 @@ static u8 mlx4_en_dcbnl_set_all(struct net_device *netdev)
 		for (tc = 0; tc < CEE_DCBX_MAX_PRIO; tc++) {
 			u8 tc_mask = 1 << tc;
 
-			switch (dcb_cfg->tc_config[tc].dcb_pfc) {
+			switch (priv->cee_config.dcb_pfc[tc]) {
 			case pfc_disabled:
 				priv->prof->tx_ppp &= ~tc_mask;
 				priv->prof->rx_ppp &= ~tc_mask;
@@ -226,7 +225,7 @@ static u8 mlx4_en_dcbnl_set_state(struct net_device *dev, u8 state)
 	struct mlx4_en_priv *priv = netdev_priv(dev);
 	int num_tcs = 0;
 
-	if (!(priv->cee_params.dcbx_cap & DCB_CAP_DCBX_VER_CEE))
+	if (!(priv->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
 		return 1;
 
 	if (!!(state) == !!(priv->flags & MLX4_EN_FLAG_DCB_ENABLED))
@@ -256,7 +255,7 @@ static int mlx4_en_dcbnl_getapp(struct net_device *netdev, u8 idtype, u16 id)
 				.selector = idtype,
 				.protocol = id,
 			     };
-	if (!(priv->cee_params.dcbx_cap & DCB_CAP_DCBX_VER_CEE))
+	if (!(priv->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
 		return 0;
 
 	return dcb_getapp(netdev, &app);
@@ -268,7 +267,7 @@ static int mlx4_en_dcbnl_setapp(struct net_device *netdev, u8 idtype,
 	struct mlx4_en_priv *priv = netdev_priv(netdev);
 	struct dcb_app app;
 
-	if (!(priv->cee_params.dcbx_cap & DCB_CAP_DCBX_VER_CEE))
+	if (!(priv->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
 		return -EINVAL;
 
 	memset(&app, 0, sizeof(struct dcb_app));
@@ -437,7 +436,7 @@ static u8 mlx4_en_dcbnl_getdcbx(struct net_device *dev)
 {
 	struct mlx4_en_priv *priv = netdev_priv(dev);
 
-	return priv->cee_params.dcbx_cap;
+	return priv->dcbx_cap;
 }
 
 static u8 mlx4_en_dcbnl_setdcbx(struct net_device *dev, u8 mode)
@@ -446,7 +445,7 @@ static u8 mlx4_en_dcbnl_setdcbx(struct net_device *dev, u8 mode)
 	struct ieee_ets ets = {0};
 	struct ieee_pfc pfc = {0};
 
-	if (mode == priv->cee_params.dcbx_cap)
+	if (mode == priv->dcbx_cap)
 		return 0;
 
 	if ((mode & DCB_CAP_DCBX_LLD_MANAGED) ||
@@ -455,7 +454,7 @@ static u8 mlx4_en_dcbnl_setdcbx(struct net_device *dev, u8 mode)
 	    !(mode & DCB_CAP_DCBX_HOST))
 		goto err;
 
-	priv->cee_params.dcbx_cap = mode;
+	priv->dcbx_cap = mode;
 
 	ets.ets_cap = IEEE_8021QAZ_MAX_TCS;
 	pfc.pfc_cap = IEEE_8021QAZ_MAX_TCS;
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 4198e9bf89d0..fedb829276f4 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -71,10 +71,11 @@ int mlx4_en_setup_tc(struct net_device *dev, u8 up)
 #ifdef CONFIG_MLX4_EN_DCB
 	if (!mlx4_is_slave(priv->mdev->dev)) {
 		if (up) {
-			priv->flags |= MLX4_EN_FLAG_DCB_ENABLED;
+			if (priv->dcbx_cap)
+				priv->flags |= MLX4_EN_FLAG_DCB_ENABLED;
 		} else {
 			priv->flags &= ~MLX4_EN_FLAG_DCB_ENABLED;
-			priv->cee_params.dcb_cfg.pfc_state = false;
+			priv->cee_config.pfc_state = false;
 		}
 	}
 #endif /* CONFIG_MLX4_EN_DCB */
@@ -3048,9 +3049,6 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	struct mlx4_en_priv *priv;
 	int i;
 	int err;
-#ifdef CONFIG_MLX4_EN_DCB
-	struct tc_configuration *tc;
-#endif
 
 	dev = alloc_etherdev_mqs(sizeof(struct mlx4_en_priv),
 				 MAX_TX_RINGS, MAX_RX_RINGS);
@@ -3117,16 +3115,13 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	priv->msg_enable = MLX4_EN_MSG_LEVEL;
 #ifdef CONFIG_MLX4_EN_DCB
 	if (!mlx4_is_slave(priv->mdev->dev)) {
-		priv->cee_params.dcbx_cap = DCB_CAP_DCBX_VER_CEE |
-					    DCB_CAP_DCBX_HOST |
-					    DCB_CAP_DCBX_VER_IEEE;
+		priv->dcbx_cap = DCB_CAP_DCBX_VER_CEE | DCB_CAP_DCBX_HOST |
+			DCB_CAP_DCBX_VER_IEEE;
 		priv->flags |= MLX4_EN_DCB_ENABLED;
-		priv->cee_params.dcb_cfg.pfc_state = false;
+		priv->cee_config.pfc_state = false;
 
-		for (i = 0; i < MLX4_EN_NUM_UP; i++) {
-			tc = &priv->cee_params.dcb_cfg.tc_config[i];
-			tc->dcb_pfc = pfc_disabled;
-		}
+		for (i = 0; i < MLX4_EN_NUM_UP; i++)
+			priv->cee_config.dcb_pfc[i] = pfc_disabled;
 
 		if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETS_CFG) {
 			dev->dcbnl_ops = &mlx4_en_dcbnl_ops;
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index 2c2913dcae98..9099dbd04951 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -482,20 +482,10 @@ enum dcb_pfc_type {
 	pfc_enabled_rx
 };
 
-struct tc_configuration {
-	enum dcb_pfc_type  dcb_pfc;
-};
-
 struct mlx4_en_cee_config {
 	bool	pfc_state;
-	struct	tc_configuration tc_config[MLX4_EN_NUM_UP];
+	enum	dcb_pfc_type dcb_pfc[MLX4_EN_NUM_UP];
 };
-
-struct mlx4_en_cee_params {
-	u8 dcbx_cap;
-	struct mlx4_en_cee_config dcb_cfg;
-};
-
 #endif
 
 struct ethtool_flow_id {
@@ -624,7 +614,8 @@ struct mlx4_en_priv {
 	struct ieee_ets ets;
 	u16 maxrate[IEEE_8021QAZ_MAX_TCS];
 	enum dcbnl_cndd_states cndd_state[IEEE_8021QAZ_MAX_TCS];
-	struct mlx4_en_cee_params cee_params;
+	struct mlx4_en_cee_config cee_config;
+	u8 dcbx_cap;
 #endif
 #ifdef CONFIG_RFS_ACCEL
 	spinlock_t filters_lock;
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index 3d2095e5c61c..c5b2064297a1 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -52,7 +52,7 @@
 
 #define MLX4_FLAG_V_IGNORE_FCS_MASK		0x2
 #define MLX4_IGNORE_FCS_MASK			0x1
-#define MLNX4_TX_MAX_NUMBER			8
+#define MLX4_TC_MAX_NUMBER			8
 
 void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table)
 {
@@ -2022,7 +2022,7 @@ int mlx4_max_tc(struct mlx4_dev *dev)
 	u8 num_tc = dev->caps.max_tc_eth;
 
 	if (!num_tc)
-		num_tc = MLNX4_TX_MAX_NUMBER;
+		num_tc = MLX4_TC_MAX_NUMBER;
 
 	return num_tc;
 }
-- 
1.8.3.1

[PATCH net 6/9] net/mlx4_core: Use RCU to perform radix tree lookup for SRQ

[Tariq Toukan]

From: Leon Romanovsky <leonro@mellanox.com>

Radix tree lookup can be performed without locking.

Fixes: 225c7b1feef1 ("IB/mlx4: Add a driver Mellanox ConnectX InfiniBand adapters")
Suggested-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/srq.c | 14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/srq.c b/drivers/net/ethernet/mellanox/mlx4/srq.c
index 67146624eb58..f44d089e2ca6 100644
--- a/drivers/net/ethernet/mellanox/mlx4/srq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/srq.c
@@ -45,15 +45,12 @@ void mlx4_srq_event(struct mlx4_dev *dev, u32 srqn, int event_type)
 	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
 	struct mlx4_srq *srq;
 
-	spin_lock(&srq_table->lock);
-
+	rcu_read_lock();
 	srq = radix_tree_lookup(&srq_table->tree, srqn & (dev->caps.num_srqs - 1));
+	rcu_read_unlock();
 	if (srq)
 		atomic_inc(&srq->refcount);
-
-	spin_unlock(&srq_table->lock);
-
-	if (!srq) {
+	else {
 		mlx4_warn(dev, "Async event for bogus SRQ %08x\n", srqn);
 		return;
 	}
@@ -301,12 +298,11 @@ struct mlx4_srq *mlx4_srq_lookup(struct mlx4_dev *dev, u32 srqn)
 {
 	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
 	struct mlx4_srq *srq;
-	unsigned long flags;
 
-	spin_lock_irqsave(&srq_table->lock, flags);
+	rcu_read_lock();
 	srq = radix_tree_lookup(&srq_table->tree,
 				srqn & (dev->caps.num_srqs - 1));
-	spin_unlock_irqrestore(&srq_table->lock, flags);
+	rcu_read_unlock();
 
 	return srq;
 }
-- 
1.8.3.1

[PATCH net 7/9] net/mlx4_core: Fix deadlock when switching between polling and event fw commands

[Tariq Toukan]

From: Jack Morgenstein <jackm@dev.mellanox.co.il>

When switching from polling-based fw commands to event-based fw
commands, there is a race condition which could cause a fw command
in another task to hang: that task will keep waiting for the polling
sempahore, but may never be able to acquire it. This is due to
mlx4_cmd_use_events, which "down"s the sempahore back to 0.

During driver initialization, this is not a problem, since no other
tasks which invoke FW commands are active.

However, there is a problem if the driver switches to polling mode
and then back to event mode during normal operation.

The "test_interrupts" feature does exactly that.
Running "ethtool -t <eth device> offline" causes the PF driver to
temporarily switch to polling mode, and then back to event mode.
(Note that for VF drivers, such switching is not performed).

Fix this by adding a read-write semaphore for protection when
switching between modes.

Fixes: 225c7b1feef1 ("IB/mlx4: Add a driver Mellanox ConnectX InfiniBand adapters")
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/cmd.c  | 23 +++++++++++++++++------
 drivers/net/ethernet/mellanox/mlx4/mlx4.h |  2 ++
 2 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index f04a423ff79d..a58d96cf1ed1 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -785,17 +785,23 @@ int __mlx4_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
 		return mlx4_cmd_reset_flow(dev, op, op_modifier, -EIO);
 
 	if (!mlx4_is_mfunc(dev) || (native && mlx4_is_master(dev))) {
+		int ret;
+
 		if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
 			return mlx4_internal_err_ret_value(dev, op,
 							  op_modifier);
+		down_read(&mlx4_priv(dev)->cmd.switch_sem);
 		if (mlx4_priv(dev)->cmd.use_events)
-			return mlx4_cmd_wait(dev, in_param, out_param,
-					     out_is_imm, in_modifier,
-					     op_modifier, op, timeout);
+			ret = mlx4_cmd_wait(dev, in_param, out_param,
+					    out_is_imm, in_modifier,
+					    op_modifier, op, timeout);
 		else
-			return mlx4_cmd_poll(dev, in_param, out_param,
-					     out_is_imm, in_modifier,
-					     op_modifier, op, timeout);
+			ret = mlx4_cmd_poll(dev, in_param, out_param,
+					    out_is_imm, in_modifier,
+					    op_modifier, op, timeout);
+
+		up_read(&mlx4_priv(dev)->cmd.switch_sem);
+		return ret;
 	}
 	return mlx4_slave_cmd(dev, in_param, out_param, out_is_imm,
 			      in_modifier, op_modifier, op, timeout);
@@ -2454,6 +2460,7 @@ int mlx4_cmd_init(struct mlx4_dev *dev)
 	int flags = 0;
 
 	if (!priv->cmd.initialized) {
+		init_rwsem(&priv->cmd.switch_sem);
 		mutex_init(&priv->cmd.slave_cmd_mutex);
 		sema_init(&priv->cmd.poll_sem, 1);
 		priv->cmd.use_events = 0;
@@ -2583,6 +2590,7 @@ int mlx4_cmd_use_events(struct mlx4_dev *dev)
 	if (!priv->cmd.context)
 		return -ENOMEM;
 
+	down_write(&priv->cmd.switch_sem);
 	for (i = 0; i < priv->cmd.max_cmds; ++i) {
 		priv->cmd.context[i].token = i;
 		priv->cmd.context[i].next  = i + 1;
@@ -2606,6 +2614,7 @@ int mlx4_cmd_use_events(struct mlx4_dev *dev)
 
 	down(&priv->cmd.poll_sem);
 	priv->cmd.use_events = 1;
+	up_write(&priv->cmd.switch_sem);
 
 	return err;
 }
@@ -2618,6 +2627,7 @@ void mlx4_cmd_use_polling(struct mlx4_dev *dev)
 	struct mlx4_priv *priv = mlx4_priv(dev);
 	int i;
 
+	down_write(&priv->cmd.switch_sem);
 	priv->cmd.use_events = 0;
 
 	for (i = 0; i < priv->cmd.max_cmds; ++i)
@@ -2626,6 +2636,7 @@ void mlx4_cmd_use_polling(struct mlx4_dev *dev)
 	kfree(priv->cmd.context);
 
 	up(&priv->cmd.poll_sem);
+	up_write(&priv->cmd.switch_sem);
 }
 
 struct mlx4_cmd_mailbox *mlx4_alloc_cmd_mailbox(struct mlx4_dev *dev)
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index c9d7fc5159f2..c128ba3ef014 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -46,6 +46,7 @@
 #include <linux/interrupt.h>
 #include <linux/spinlock.h>
 #include <net/devlink.h>
+#include <linux/rwsem.h>
 
 #include <linux/mlx4/device.h>
 #include <linux/mlx4/driver.h>
@@ -627,6 +628,7 @@ struct mlx4_cmd {
 	struct mutex		slave_cmd_mutex;
 	struct semaphore	poll_sem;
 	struct semaphore	event_sem;
+	struct rw_semaphore	switch_sem;
 	int			max_cmds;
 	spinlock_t		context_lock;
 	int			free_head;
-- 
1.8.3.1

[PATCH net 8/9] net/mlx4_core: Fix to clean devlink resources

[Tariq Toukan]

From: Kamal Heib <kamalh@mellanox.com>

This patch cleans devlink resources by calling devlink_port_unregister()
to avoid the following issues:

- Kernel panic when triggering reset flow.
- Memory leak due to unfreed resources in mlx4_init_port_info().

Fixes: 09d4d087cd48 ("mlx4: Implement devlink interface")
Signed-off-by: Kamal Heib <kamalh@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/main.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 75dd2e3d3059..7183ac4135d2 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -2970,6 +2970,7 @@ static int mlx4_init_port_info(struct mlx4_dev *dev, int port)
 		mlx4_err(dev, "Failed to create mtu file for port %d\n", port);
 		device_remove_file(&info->dev->persist->pdev->dev,
 				   &info->port_attr);
+		devlink_port_unregister(&info->devlink_port);
 		info->port = -1;
 	}
 
@@ -2984,6 +2985,8 @@ static void mlx4_cleanup_port_info(struct mlx4_port_info *info)
 	device_remove_file(&info->dev->persist->pdev->dev, &info->port_attr);
 	device_remove_file(&info->dev->persist->pdev->dev,
 			   &info->port_mtu_attr);
+	devlink_port_unregister(&info->devlink_port);
+
 #ifdef CONFIG_RFS_ACCEL
 	free_irq_cpu_rmap(info->rmap);
 	info->rmap = NULL;
-- 
1.8.3.1

[PATCH net 9/9] net/mlx4_en: Fix panic on xmit while port is down

[Tariq Toukan]

From: Moshe Shemesh <moshe@mellanox.com>

When port is down, tx drop counter update is not needed.
Updating the counter in this case can cause a kernel
panic as when the port is down, ring can be NULL.

Fixes: 63a664b7e92b ("net/mlx4_en: fix tx_dropped bug")
Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_tx.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index 9df87ca0515a..e2509bba3e7c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -818,7 +818,7 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 	real_size = get_real_size(skb, shinfo, dev, &lso_header_size,
 				  &inline_ok, &fragptr);
 	if (unlikely(!real_size))
-		goto tx_drop;
+		goto tx_drop_count;
 
 	/* Align descriptor to TXBB size */
 	desc_size = ALIGN(real_size, TXBB_SIZE);
@@ -826,7 +826,7 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 	if (unlikely(nr_txbb > MAX_DESC_TXBBS)) {
 		if (netif_msg_tx_err(priv))
 			en_warn(priv, "Oversized header or SG list\n");
-		goto tx_drop;
+		goto tx_drop_count;
 	}
 
 	bf_ok = ring->bf_enabled;
@@ -1071,9 +1071,10 @@ tx_drop_unmap:
 			       PCI_DMA_TODEVICE);
 	}
 
+tx_drop_count:
+	ring->tx_dropped++;
 tx_drop:
 	dev_kfree_skb_any(skb);
-	ring->tx_dropped++;
 	return NETDEV_TX_OK;
 }
 
@@ -1106,7 +1107,7 @@ netdev_tx_t mlx4_en_xmit_frame(struct mlx4_en_rx_alloc *frame,
 		goto tx_drop;
 
 	if (mlx4_en_is_tx_ring_full(ring))
-		goto tx_drop;
+		goto tx_drop_count;
 
 	/* fetch ring->cons far ahead before needing it to avoid stall */
 	ring_cons = READ_ONCE(ring->cons);
@@ -1176,7 +1177,8 @@ netdev_tx_t mlx4_en_xmit_frame(struct mlx4_en_rx_alloc *frame,
 
 	return NETDEV_TX_OK;
 
-tx_drop:
+tx_drop_count:
 	ring->tx_dropped++;
+tx_drop:
 	return NETDEV_TX_BUSY;
 }
-- 
1.8.3.1

Re: [PATCH net 1/9] net/mlx4_en: Add branch prediction hints in RX data-path

[David Miller]

From: Tariq Toukan <tariqt@mellanox.com>
Date: Thu,  8 Sep 2016 11:51:53 +0300

> Add likely/unlikely hints to improve branch predictions
> in the RX data-path.
> 
> Signed-off-by: Tariq Toukan <tariqt@mellanox.com>

This is absolutely not appropriate for 'net'.  This is not a bug
fix, it is a feature/enhancement.

Re: [PATCH net 2/9] net/mlx4_en: Fix wrong indentation

[David Miller]

From: Tariq Toukan <tariqt@mellanox.com>
Date: Thu,  8 Sep 2016 11:51:54 +0300

> From: Kamal Heib <kamalh@mellanox.com>
> 
> Use tabs instead of spaces before if statement, no functional change.
> 
> Fixes: e7c1c2c46201 ("mlx4_en: Added self diagnostics test implementation")
> Signed-off-by: Kamal Heib <kamalh@mellanox.com>
> Signed-off-by: Tariq Toukan <tariqt@mellanox.com>

Cleanups like this are not bug fixes, please don't put things like this
into changes targetting 'net'.

Re: [PATCH net 6/9] net/mlx4_core: Use RCU to perform radix tree lookup for SRQ

[David Miller]

From: Tariq Toukan <tariqt@mellanox.com>
Date: Thu,  8 Sep 2016 11:51:58 +0300

> From: Leon Romanovsky <leonro@mellanox.com>
> 
> Radix tree lookup can be performed without locking.
> 
> Fixes: 225c7b1feef1 ("IB/mlx4: Add a driver Mellanox ConnectX InfiniBand adapters")
> Suggested-by: Sagi Grimberg <sagi@grimberg.me>
> Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
> Signed-off-by: Tariq Toukan <tariqt@mellanox.com>

Unless this fixes a bug, it isn't appropriate for 'net'.

If it does fix a bug, you have to explain what that bug is and
how this fixes it.

Re: [PATCH net 6/9] net/mlx4_core: Use RCU to perform radix tree lookup for SRQ

[Tariq Toukan]

Hi Dave,

On 08/09/2016 11:36 PM, David Miller wrote:
> From: Tariq Toukan <tariqt@mellanox.com>
> Date: Thu,  8 Sep 2016 11:51:58 +0300
>
>> From: Leon Romanovsky <leonro@mellanox.com>
>>
>> Radix tree lookup can be performed without locking.
>>
>> Fixes: 225c7b1feef1 ("IB/mlx4: Add a driver Mellanox ConnectX InfiniBand adapters")
>> Suggested-by: Sagi Grimberg <sagi@grimberg.me>
>> Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
>> Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
> Unless this fixes a bug, it isn't appropriate for 'net'.
I see, I will exclude these patches and re-submit the others.
>
> If it does fix a bug, you have to explain what that bug is and
> how this fixes it.
Thanks,
Tariq