[PATCH net 0/6] Mellanox Core and Ethernet driver fixes 2013-03-07

[PATCH net 0/6] Mellanox Core and Ethernet driver fixes 2013-03-07

[Or Gerlitz]

Hi Dave,

Here's a batch of fixes to the mlx4 core and ethernet drivers for 3.9

The commit that disabled RFS when running in SRIOV mode fixes a regression which was 
introduced in 3.9-rc1 but actually present also in the 3.8 -stable series. It turns out 
that a slightly different fix is needed there and we will generate and submit it there.

Patches done against net commit 66d29cbc59433 "benet: Wait f/w POST until timeout"

Or.

Amir Vadai (1):
  net/mlx4_en: Disable RFS when running in SRIOV mode

Jack Morgenstein (3):
  net/mlx4_core: Disable mlx4_QP_ATTACH calls from guests if the host uses flow steering
  net/mlx4_core: Turn off device-managed FS bit in dev-cap wrapper if DMFS is not enabled
  net/mlx4_core: Fix endianness bug in set_param_l

Yan Burman (2):
  net/mlx4_en: Fix race when setting the device MAC address
  net/mlx4_en: Cleanup MAC resources on module unload or port stop

 drivers/net/ethernet/mellanox/mlx4/cq.c            |    2 +-
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c     |   86 +++++++++++---------
 drivers/net/ethernet/mellanox/mlx4/fw.c            |    8 ++
 drivers/net/ethernet/mellanox/mlx4/main.c          |    2 +-
 drivers/net/ethernet/mellanox/mlx4/mlx4.h          |    2 +-
 drivers/net/ethernet/mellanox/mlx4/mlx4_en.h       |    1 -
 drivers/net/ethernet/mellanox/mlx4/mr.c            |   10 +-
 drivers/net/ethernet/mellanox/mlx4/pd.c            |    2 +-
 drivers/net/ethernet/mellanox/mlx4/port.c          |    8 +-
 drivers/net/ethernet/mellanox/mlx4/qp.c            |    8 +-
 .../net/ethernet/mellanox/mlx4/resource_tracker.c  |    3 +
 drivers/net/ethernet/mellanox/mlx4/srq.c           |    2 +-
 12 files changed, 75 insertions(+), 59 deletions(-)

[PATCH net 1/6] net/mlx4_core: Disable mlx4_QP_ATTACH calls from guests if the host uses flow steering

[Or Gerlitz]

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

Guests kernels may not correctly detect if DMFS (device-enabled flow steering) is
activated by the host. If DMFS is activated, the master should return error to guests
which try to use the B0-steering flow calls (mlx4_QP_ATTACH).

Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 .../net/ethernet/mellanox/mlx4/resource_tracker.c  |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
index 083fb48..2995687 100644
--- a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
+++ b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
@@ -2990,6 +2990,9 @@ int mlx4_QP_ATTACH_wrapper(struct mlx4_dev *dev, int slave,
 	u8 steer_type_mask = 2;
 	enum mlx4_steer_type type = (gid[7] & steer_type_mask) >> 1;
 
+	if (dev->caps.steering_mode != MLX4_STEERING_MODE_B0)
+		return -EINVAL;
+
 	qpn = vhcr->in_modifier & 0xffffff;
 	err = get_res(dev, slave, qpn, RES_QP, &rqp);
 	if (err)
-- 
1.7.1

[PATCH net 2/6] net/mlx4_core: Turn off device-managed FS bit in dev-cap wrapper if DMFS is not enabled

[Or Gerlitz]

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

Older kernels detect DMFS (device-managed flow steering) from the HCA
device capability directly, regardless of whether the capability was
enabled in INIT_HCA, this is fixed by commit 7b8157bed "mlx4_core: Adjustments
to Flow Steering activation logic for SR-IOV"

To protect against guests running kernels without this fix, the host driver
should turn off the DMFS capability bit in mlx4_QUERY_DEV_CAP_wrapper.

Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/fw.c |    8 ++++++++
 1 files changed, 8 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 50917eb..f624557 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -787,6 +787,14 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
 	bmme_flags &= ~MLX4_BMME_FLAG_TYPE_2_WIN;
 	MLX4_PUT(outbox->buf, bmme_flags, QUERY_DEV_CAP_BMME_FLAGS_OFFSET);
 
+	/* turn off device-managed steering capability if not enabled */
+	if (dev->caps.steering_mode != MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		MLX4_GET(field, outbox->buf,
+			 QUERY_DEV_CAP_FLOW_STEERING_RANGE_EN_OFFSET);
+		field &= 0x7f;
+		MLX4_PUT(outbox->buf, field,
+			 QUERY_DEV_CAP_FLOW_STEERING_RANGE_EN_OFFSET);
+	}
 	return 0;
 }
 
-- 
1.7.1

[PATCH net 3/6] net/mlx4_core: Fix endianness bug in set_param_l

[Or Gerlitz]

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

The set_param_l function assumes casting a u64 pointer to a u32 pointer
allows to access the lower 32bits, but it results in writing the upper
32 bits on big endian systems.

The fixed function reads the upper 32 bits of the 64 argument, and or's
them with the 32 bits of the 32-bit value passed to the function.

Since this is now a "read-modify-write" operation, we got many
"unintialized variable" warnings which needed to be fixed as well.

Reported-by: Alexander Schmidt <alexschm@de.ibm.com>.
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/cq.c   |    2 +-
 drivers/net/ethernet/mellanox/mlx4/main.c |    2 +-
 drivers/net/ethernet/mellanox/mlx4/mlx4.h |    2 +-
 drivers/net/ethernet/mellanox/mlx4/mr.c   |   10 +++++-----
 drivers/net/ethernet/mellanox/mlx4/pd.c   |    2 +-
 drivers/net/ethernet/mellanox/mlx4/port.c |    8 ++++----
 drivers/net/ethernet/mellanox/mlx4/qp.c   |    8 ++++----
 drivers/net/ethernet/mellanox/mlx4/srq.c  |    2 +-
 8 files changed, 18 insertions(+), 18 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/cq.c b/drivers/net/ethernet/mellanox/mlx4/cq.c
index 7e64033..0706623 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cq.c
@@ -226,7 +226,7 @@ void __mlx4_cq_free_icm(struct mlx4_dev *dev, int cqn)
 
 static void mlx4_cq_free_icm(struct mlx4_dev *dev, int cqn)
 {
-	u64 in_param;
+	u64 in_param = 0;
 	int err;
 
 	if (mlx4_is_mfunc(dev)) {
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index d180bc4..16abde2 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -1555,7 +1555,7 @@ void __mlx4_counter_free(struct mlx4_dev *dev, u32 idx)
 
 void mlx4_counter_free(struct mlx4_dev *dev, u32 idx)
 {
-	u64 in_param;
+	u64 in_param = 0;
 
 	if (mlx4_is_mfunc(dev)) {
 		set_param_l(&in_param, idx);
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index cf88334..d738454 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -1235,7 +1235,7 @@ int mlx4_get_qp_per_mgm(struct mlx4_dev *dev);
 
 static inline void set_param_l(u64 *arg, u32 val)
 {
-	*((u32 *)arg) = val;
+	*arg = (*arg & 0xffffffff00000000ULL) | (u64) val;
 }
 
 static inline void set_param_h(u64 *arg, u32 val)
diff --git a/drivers/net/ethernet/mellanox/mlx4/mr.c b/drivers/net/ethernet/mellanox/mlx4/mr.c
index 602ca9b..f91719a 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mr.c
+++ b/drivers/net/ethernet/mellanox/mlx4/mr.c
@@ -183,7 +183,7 @@ u32 __mlx4_alloc_mtt_range(struct mlx4_dev *dev, int order)
 
 static u32 mlx4_alloc_mtt_range(struct mlx4_dev *dev, int order)
 {
-	u64 in_param;
+	u64 in_param = 0;
 	u64 out_param;
 	int err;
 
@@ -240,7 +240,7 @@ void __mlx4_free_mtt_range(struct mlx4_dev *dev, u32 offset, int order)
 
 static void mlx4_free_mtt_range(struct mlx4_dev *dev, u32 offset, int order)
 {
-	u64 in_param;
+	u64 in_param = 0;
 	int err;
 
 	if (mlx4_is_mfunc(dev)) {
@@ -351,7 +351,7 @@ void __mlx4_mpt_release(struct mlx4_dev *dev, u32 index)
 
 static void mlx4_mpt_release(struct mlx4_dev *dev, u32 index)
 {
-	u64 in_param;
+	u64 in_param = 0;
 
 	if (mlx4_is_mfunc(dev)) {
 		set_param_l(&in_param, index);
@@ -374,7 +374,7 @@ int __mlx4_mpt_alloc_icm(struct mlx4_dev *dev, u32 index)
 
 static int mlx4_mpt_alloc_icm(struct mlx4_dev *dev, u32 index)
 {
-	u64 param;
+	u64 param = 0;
 
 	if (mlx4_is_mfunc(dev)) {
 		set_param_l(&param, index);
@@ -395,7 +395,7 @@ void __mlx4_mpt_free_icm(struct mlx4_dev *dev, u32 index)
 
 static void mlx4_mpt_free_icm(struct mlx4_dev *dev, u32 index)
 {
-	u64 in_param;
+	u64 in_param = 0;
 
 	if (mlx4_is_mfunc(dev)) {
 		set_param_l(&in_param, index);
diff --git a/drivers/net/ethernet/mellanox/mlx4/pd.c b/drivers/net/ethernet/mellanox/mlx4/pd.c
index 1ac8863..00f223a 100644
--- a/drivers/net/ethernet/mellanox/mlx4/pd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/pd.c
@@ -101,7 +101,7 @@ void __mlx4_xrcd_free(struct mlx4_dev *dev, u32 xrcdn)
 
 void mlx4_xrcd_free(struct mlx4_dev *dev, u32 xrcdn)
 {
-	u64 in_param;
+	u64 in_param = 0;
 	int err;
 
 	if (mlx4_is_mfunc(dev)) {
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index 719ead1..10c57c8 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -175,7 +175,7 @@ EXPORT_SYMBOL_GPL(__mlx4_register_mac);
 
 int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac)
 {
-	u64 out_param;
+	u64 out_param = 0;
 	int err;
 
 	if (mlx4_is_mfunc(dev)) {
@@ -222,7 +222,7 @@ EXPORT_SYMBOL_GPL(__mlx4_unregister_mac);
 
 void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, u64 mac)
 {
-	u64 out_param;
+	u64 out_param = 0;
 
 	if (mlx4_is_mfunc(dev)) {
 		set_param_l(&out_param, port);
@@ -361,7 +361,7 @@ out:
 
 int mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index)
 {
-	u64 out_param;
+	u64 out_param = 0;
 	int err;
 
 	if (mlx4_is_mfunc(dev)) {
@@ -406,7 +406,7 @@ out:
 
 void mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, int index)
 {
-	u64 in_param;
+	u64 in_param = 0;
 	int err;
 
 	if (mlx4_is_mfunc(dev)) {
diff --git a/drivers/net/ethernet/mellanox/mlx4/qp.c b/drivers/net/ethernet/mellanox/mlx4/qp.c
index 81e2abe..e891b05 100644
--- a/drivers/net/ethernet/mellanox/mlx4/qp.c
+++ b/drivers/net/ethernet/mellanox/mlx4/qp.c
@@ -222,7 +222,7 @@ int __mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align,
 
 int mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align, int *base)
 {
-	u64 in_param;
+	u64 in_param = 0;
 	u64 out_param;
 	int err;
 
@@ -255,7 +255,7 @@ void __mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt)
 
 void mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt)
 {
-	u64 in_param;
+	u64 in_param = 0;
 	int err;
 
 	if (mlx4_is_mfunc(dev)) {
@@ -319,7 +319,7 @@ err_out:
 
 static int mlx4_qp_alloc_icm(struct mlx4_dev *dev, int qpn)
 {
-	u64 param;
+	u64 param = 0;
 
 	if (mlx4_is_mfunc(dev)) {
 		set_param_l(&param, qpn);
@@ -344,7 +344,7 @@ void __mlx4_qp_free_icm(struct mlx4_dev *dev, int qpn)
 
 static void mlx4_qp_free_icm(struct mlx4_dev *dev, int qpn)
 {
-	u64 in_param;
+	u64 in_param = 0;
 
 	if (mlx4_is_mfunc(dev)) {
 		set_param_l(&in_param, qpn);
diff --git a/drivers/net/ethernet/mellanox/mlx4/srq.c b/drivers/net/ethernet/mellanox/mlx4/srq.c
index feda6c0..e329fe1 100644
--- a/drivers/net/ethernet/mellanox/mlx4/srq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/srq.c
@@ -149,7 +149,7 @@ void __mlx4_srq_free_icm(struct mlx4_dev *dev, int srqn)
 
 static void mlx4_srq_free_icm(struct mlx4_dev *dev, int srqn)
 {
-	u64 in_param;
+	u64 in_param = 0;
 
 	if (mlx4_is_mfunc(dev)) {
 		set_param_l(&in_param, srqn);
-- 
1.7.1

[PATCH net 4/6] net/mlx4_en: Fix race when setting the device MAC address

[Or Gerlitz]

From: Yan Burman <yanb@mellanox.com>

Remove unnecessary use of workqueue for the device MAC address setting
flow, and fix a race when setting MAC address which was introduced by
commit c07cb4b0a "net/mlx4_en: Manage hash of MAC addresses per port"

The race happened when mlx4_en_replace_mac was being executed in parallel
with a successive call to ndo_set_mac_address, e.g witn an A/B/A MAC
setting configuration test, the third set fails.

With this change we also properly report an error if set MAC fails.

Signed-off-by: Yan Burman <yanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |   42 ++++++++++++------------
 drivers/net/ethernet/mellanox/mlx4/mlx4_en.h   |    1 -
 2 files changed, 21 insertions(+), 22 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index bb4d8d9..217e618 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -650,28 +650,10 @@ u64 mlx4_en_mac_to_u64(u8 *addr)
 	return mac;
 }
 
-static int mlx4_en_set_mac(struct net_device *dev, void *addr)
+static int mlx4_en_do_set_mac(struct mlx4_en_priv *priv)
 {
-	struct mlx4_en_priv *priv = netdev_priv(dev);
-	struct mlx4_en_dev *mdev = priv->mdev;
-	struct sockaddr *saddr = addr;
-
-	if (!is_valid_ether_addr(saddr->sa_data))
-		return -EADDRNOTAVAIL;
-
-	memcpy(dev->dev_addr, saddr->sa_data, ETH_ALEN);
-	queue_work(mdev->workqueue, &priv->mac_task);
-	return 0;
-}
-
-static void mlx4_en_do_set_mac(struct work_struct *work)
-{
-	struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv,
-						 mac_task);
-	struct mlx4_en_dev *mdev = priv->mdev;
 	int err = 0;
 
-	mutex_lock(&mdev->state_lock);
 	if (priv->port_up) {
 		/* Remove old MAC and insert the new one */
 		err = mlx4_en_replace_mac(priv, priv->base_qpn,
@@ -683,7 +665,26 @@ static void mlx4_en_do_set_mac(struct work_struct *work)
 	} else
 		en_dbg(HW, priv, "Port is down while registering mac, exiting...\n");
 
+	return err;
+}
+
+static int mlx4_en_set_mac(struct net_device *dev, void *addr)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct sockaddr *saddr = addr;
+	int err;
+
+	if (!is_valid_ether_addr(saddr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	memcpy(dev->dev_addr, saddr->sa_data, ETH_ALEN);
+
+	mutex_lock(&mdev->state_lock);
+	err = mlx4_en_do_set_mac(priv);
 	mutex_unlock(&mdev->state_lock);
+
+	return err;
 }
 
 static void mlx4_en_clear_list(struct net_device *dev)
@@ -1348,7 +1349,7 @@ static void mlx4_en_do_get_stats(struct work_struct *work)
 		queue_delayed_work(mdev->workqueue, &priv->stats_task, STATS_DELAY);
 	}
 	if (mdev->mac_removed[MLX4_MAX_PORTS + 1 - priv->port]) {
-		queue_work(mdev->workqueue, &priv->mac_task);
+		mlx4_en_do_set_mac(priv);
 		mdev->mac_removed[MLX4_MAX_PORTS + 1 - priv->port] = 0;
 	}
 	mutex_unlock(&mdev->state_lock);
@@ -2078,7 +2079,6 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	priv->msg_enable = MLX4_EN_MSG_LEVEL;
 	spin_lock_init(&priv->stats_lock);
 	INIT_WORK(&priv->rx_mode_task, mlx4_en_do_set_rx_mode);
-	INIT_WORK(&priv->mac_task, mlx4_en_do_set_mac);
 	INIT_WORK(&priv->watchdog_task, mlx4_en_restart);
 	INIT_WORK(&priv->linkstate_task, mlx4_en_linkstate);
 	INIT_DELAYED_WORK(&priv->stats_task, mlx4_en_do_get_stats);
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index c313d7e..f710b7c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -509,7 +509,6 @@ struct mlx4_en_priv {
 	struct mlx4_en_cq rx_cq[MAX_RX_RINGS];
 	struct mlx4_qp drop_qp;
 	struct work_struct rx_mode_task;
-	struct work_struct mac_task;
 	struct work_struct watchdog_task;
 	struct work_struct linkstate_task;
 	struct delayed_work stats_task;
-- 
1.7.1

[PATCH net 5/6] net/mlx4_en: Cleanup MAC resources on module unload or port stop

[Or Gerlitz]

From: Yan Burman <yanb@mellanox.com>

Make sure we cleanup all MAC related resources (entries in the port MAC
table and steering rules) when stopping a port or when the driver is unloaded.

The leak was introduced by commit 07cb4b0a "net/mlx4_en: Manage hash of MAC
addresses per port".

Signed-off-by: Yan Burman <yanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |   38 +++++++++++++----------
 1 files changed, 21 insertions(+), 17 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 217e618..7fd0936 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -565,34 +565,38 @@ static void mlx4_en_put_qp(struct mlx4_en_priv *priv)
 	struct mlx4_en_dev *mdev = priv->mdev;
 	struct mlx4_dev *dev = mdev->dev;
 	int qpn = priv->base_qpn;
-	u64 mac = mlx4_en_mac_to_u64(priv->dev->dev_addr);
-
-	en_dbg(DRV, priv, "Registering MAC: %pM for deleting\n",
-	       priv->dev->dev_addr);
-	mlx4_unregister_mac(dev, priv->port, mac);
+	u64 mac;
 
-	if (dev->caps.steering_mode != MLX4_STEERING_MODE_A0) {
+	if (dev->caps.steering_mode == MLX4_STEERING_MODE_A0) {
+		mac = mlx4_en_mac_to_u64(priv->dev->dev_addr);
+		en_dbg(DRV, priv, "Registering MAC: %pM for deleting\n",
+		       priv->dev->dev_addr);
+		mlx4_unregister_mac(dev, priv->port, mac);
+	} else {
 		struct mlx4_mac_entry *entry;
 		struct hlist_node *tmp;
 		struct hlist_head *bucket;
-		unsigned int mac_hash;
+		unsigned int i;
 
-		mac_hash = priv->dev->dev_addr[MLX4_EN_MAC_HASH_IDX];
-		bucket = &priv->mac_hash[mac_hash];
-		hlist_for_each_entry_safe(entry, tmp, bucket, hlist) {
-			if (ether_addr_equal_64bits(entry->mac,
-						    priv->dev->dev_addr)) {
-				en_dbg(DRV, priv, "Releasing qp: port %d, MAC %pM, qpn %d\n",
-				       priv->port, priv->dev->dev_addr, qpn);
+		for (i = 0; i < MLX4_EN_MAC_HASH_SIZE; ++i) {
+			bucket = &priv->mac_hash[i];
+			hlist_for_each_entry_safe(entry, tmp, bucket, hlist) {
+				mac = mlx4_en_mac_to_u64(entry->mac);
+				en_dbg(DRV, priv, "Registering MAC: %pM for deleting\n",
+				       entry->mac);
 				mlx4_en_uc_steer_release(priv, entry->mac,
 							 qpn, entry->reg_id);
-				mlx4_qp_release_range(dev, qpn, 1);
 
+				mlx4_unregister_mac(dev, priv->port, mac);
 				hlist_del_rcu(&entry->hlist);
 				kfree_rcu(entry, rcu);
-				break;
 			}
 		}
+
+		en_dbg(DRV, priv, "Releasing qp: port %d, qpn %d\n",
+		       priv->port, qpn);
+		mlx4_qp_release_range(dev, qpn, 1);
+		priv->flags &= ~MLX4_EN_FLAG_FORCE_PROMISC;
 	}
 }
 
-- 
1.7.1

[PATCH net 6/6] net/mlx4_en: Disable RFS when running in SRIOV mode

[Or Gerlitz]

From: Amir Vadai <amirv@mellanox.com>

Commit 37706996 "mlx4_en: fix allocation of CPU affinity reverse-map" fixed
a bug when mlx4_dev->caps.comp_pool is larger from the device rx rings, but
introduced a regression.

When the mlx4_core is activating its "legacy mode" (e.g when running in SRIOV
mode) w.r.t to EQs/IRQs usage, comp_pool becomes zero and we're crashing on
divide by zero alloc_cpu_rmap.

Fix that by enabling RFS only when running in non-legacy mode.

Reported-by: Yan Burman <yanb@mellanox.com>
Cc: Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---

This regression was introduced in 3.9-rc1 but actually present also in the 
3.8 -stable series. It turns out that a slightly different fix is needed there and 
we will generate and submit it there.

 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |    8 +++++---
 1 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 7fd0936..995d4b6 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -1833,9 +1833,11 @@ int mlx4_en_alloc_resources(struct mlx4_en_priv *priv)
 	}
 
 #ifdef CONFIG_RFS_ACCEL
-	priv->dev->rx_cpu_rmap = alloc_irq_cpu_rmap(priv->mdev->dev->caps.comp_pool);
-	if (!priv->dev->rx_cpu_rmap)
-		goto err;
+	if (priv->mdev->dev->caps.comp_pool) {
+		priv->dev->rx_cpu_rmap = alloc_irq_cpu_rmap(priv->mdev->dev->caps.comp_pool);
+		if (!priv->dev->rx_cpu_rmap)
+			goto err;
+	}
 #endif
 
 	return 0;
-- 
1.7.1

Re: [PATCH net 0/6] Mellanox Core and Ethernet driver fixes 2013-03-07

[David Miller]

From: Or Gerlitz <ogerlitz@mellanox.com>
Date: Thu,  7 Mar 2013 15:46:51 +0200

> Here's a batch of fixes to the mlx4 core and ethernet drivers for 3.9
> 
> The commit that disabled RFS when running in SRIOV mode fixes a regression which was 
> introduced in 3.9-rc1 but actually present also in the 3.8 -stable series. It turns out 
> that a slightly different fix is needed there and we will generate and submit it there.
> 
> Patches done against net commit 66d29cbc59433 "benet: Wait f/w POST until timeout"

Series applied, thanks.