[PATCH net 0/3] mlx4 fixes

[PATCH net 0/3] mlx4 fixes

[Or Gerlitz]

Hi Dave,

Few mlx4 driver fixes for net.

thanks,

Or.


Dotan Barak (1):
  net/mlx4_core: Unmap UAR also in the case of error flow

Jack Morgenstein (1):
  net/mlx4_en: Fix double-release-range in tx-rings

Moni Shoua (1):
  net/mlx4_en: Don't use vlan tag value as an indication for vlan presence

 drivers/net/ethernet/mellanox/mlx4/en_tx.c |    3 +--
 drivers/net/ethernet/mellanox/mlx4/eq.c    |   18 ++++++++++++++----
 2 files changed, 15 insertions(+), 6 deletions(-)

[PATCH net 1/3] net/mlx4_en: Fix double-release-range in tx-rings

[Or Gerlitz]

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

The QP range is reserved as a single block. However, when freeing the
en resources, the tx-ring QPs are released both in mlx4_en_destroy_tx_ring
(one at a time) and in mlx4_en_free_resources (as a block release).

Fix by eliminating the one-at-a-time release in mlx4_en_destroy_tx_ring.

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

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index c10e3a6..0a51095 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -143,7 +143,6 @@ void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv,
 		mlx4_bf_free(mdev->dev, &ring->bf);
 	mlx4_qp_remove(mdev->dev, &ring->qp);
 	mlx4_qp_free(mdev->dev, &ring->qp);
-	mlx4_qp_release_range(mdev->dev, ring->qpn, 1);
 	mlx4_en_unmap_buffer(&ring->wqres.buf);
 	mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size);
 	kfree(ring->bounce_buf);
-- 
1.7.1

[PATCH net 2/3] net/mlx4_en: Don't use vlan tag value as an indication for vlan presence

[Or Gerlitz]

From: Moni Shoua <monis@mellanox.co.il>

The vlan tag can be zero. This is why it can't serve as an indication
that packet requires VLAN header in the TX flow.

Signed-off-by: Moni Shoua <monis@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_tx.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index 0a51095..b35094c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -711,7 +711,7 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 	if (bounce)
 		tx_desc = mlx4_en_bounce_to_desc(priv, ring, index, desc_size);
 
-	if (ring->bf_enabled && desc_size <= MAX_BF && !bounce && !vlan_tag) {
+	if (ring->bf_enabled && desc_size <= MAX_BF && !bounce && !vlan_tx_tag_present(skb)) {
 		*(__be32 *) (&tx_desc->ctrl.vlan_tag) |= cpu_to_be32(ring->doorbell_qpn);
 		op_own |= htonl((bf_index & 0xffff) << 8);
 		/* Ensure new descirptor hits memory
-- 
1.7.1

[PATCH net 3/3] net/mlx4_core: Unmap UAR also in the case of error flow

[Or Gerlitz]

From: Dotan Barak <dotanb@dev.mellanox.co.il>

If a failure takes place during the EQ creation, we need to unmap the
UAR memory block too.

Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Signed-off-by: Uri Habusha <urih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/eq.c |   18 ++++++++++++++----
 1 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/eq.c b/drivers/net/ethernet/mellanox/mlx4/eq.c
index 51c7649..083767b 100644
--- a/drivers/net/ethernet/mellanox/mlx4/eq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/eq.c
@@ -843,6 +843,18 @@ static void __iomem *mlx4_get_eq_uar(struct mlx4_dev *dev, struct mlx4_eq *eq)
 	return priv->eq_table.uar_map[index] + 0x800 + 8 * (eq->eqn % 4);
 }
 
+static void mlx4_unmap_uar(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+
+	for (i = 0; i < mlx4_num_eq_uar(dev); ++i)
+		if (priv->eq_table.uar_map[i]) {
+			iounmap(priv->eq_table.uar_map[i]);
+			priv->eq_table.uar_map[i] = NULL;
+		}
+}
+
 static int mlx4_create_eq(struct mlx4_dev *dev, int nent,
 			  u8 intr, struct mlx4_eq *eq)
 {
@@ -1207,6 +1219,7 @@ err_out_unmap:
 	mlx4_free_irqs(dev);
 
 err_out_bitmap:
+	mlx4_unmap_uar(dev);
 	mlx4_bitmap_cleanup(&priv->eq_table.bitmap);
 
 err_out_free:
@@ -1231,10 +1244,7 @@ void mlx4_cleanup_eq_table(struct mlx4_dev *dev)
 	if (!mlx4_is_slave(dev))
 		mlx4_unmap_clr_int(dev);
 
-	for (i = 0; i < mlx4_num_eq_uar(dev); ++i)
-		if (priv->eq_table.uar_map[i])
-			iounmap(priv->eq_table.uar_map[i]);
-
+	mlx4_unmap_uar(dev);
 	mlx4_bitmap_cleanup(&priv->eq_table.bitmap);
 
 	kfree(priv->eq_table.uar_map);
-- 
1.7.1

Re: [PATCH net 0/3] mlx4 fixes

[David Miller]

From: Or Gerlitz <ogerlitz@mellanox.com>
Date: Thu, 25 Oct 2012 13:12:46 +0200

> Dotan Barak (1):
>   net/mlx4_core: Unmap UAR also in the case of error flow
> 
> Jack Morgenstein (1):
>   net/mlx4_en: Fix double-release-range in tx-rings
> 
> Moni Shoua (1):
>   net/mlx4_en: Don't use vlan tag value as an indication for vlan presence

All applied, thanks.

[PATCH net 0/3] mlx4 fixes

[Or Gerlitz]

Hi Dave,

These short series fixes two bugs related to the vxlan support and 
a missing req module call for the IB driver which is needed to support 
IB/RDMA over Ethernet.

Pathes done over the net tree, commit dd38743 "vlan: Set correct source MAC address 
with TX VLAN offload enabled"

Or.

Or Gerlitz (3):
  net/mlx4_core: Fix wrong dump of the vxlan offloads device capability
  net/mlx4_en: Handle vxlan steering rules for mac address changes
  net/mlx4_core: Load the IB driver when the device supports IBoE

 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |    8 ++++++++
 drivers/net/ethernet/mellanox/mlx4/fw.c        |    5 +++--
 drivers/net/ethernet/mellanox/mlx4/main.c      |    2 +-
 3 files changed, 12 insertions(+), 3 deletions(-)

Re: [PATCH net 0/3] mlx4 fixes

[David Miller]

From: Or Gerlitz <ogerlitz@mellanox.com>
Date: Wed, 12 Mar 2014 17:16:29 +0200

> These short series fixes two bugs related to the vxlan support and a
> missing req module call for the IB driver which is needed to support
> IB/RDMA over Ethernet.
> 
> Pathes done over the net tree, commit dd38743 "vlan: Set correct
> source MAC address with TX VLAN offload enabled"

Series applied, thanks Or.

I wonder though, if a MAC address change results in a VLAN steering
rule insertion failure, won't we be left in an inconsistent state?

Re: [PATCH net 0/3] mlx4 fixes

[Or Gerlitz]

On Wed, Mar 12, 2014 at 10:14 PM, David Miller <davem@davemloft.net> wrote:
> From: Or Gerlitz <ogerlitz@mellanox.com>
> Date: Wed, 12 Mar 2014 17:16:29 +0200
>
>> These short series fixes two bugs related to the vxlan support and a
>> missing req module call for the IB driver which is needed to support
>> IB/RDMA over Ethernet.
>>
>> Pathes done over the net tree, commit dd38743 "vlan: Set correct
>> source MAC address with TX VLAN offload enabled"
>
> Series applied, thanks Or.
>
> I wonder though, if a MAC address change results in a VLAN steering
> rule insertion failure, won't we be left in an inconsistent state?

in that case non-VXLAN traffic will fly in OK, but not VXLAN traffic.
I don't think we want to close the shop when this happens but probably
print some warning or a like, will send a patch...

Re: [PATCH net 0/3] mlx4 fixes

[Or Gerlitz]

On 12/03/2014 22:14, David Miller wrote:
> I wonder though, if a MAC address change results in a VLAN steering
> rule insertion failure, won't we be left in an inconsistent state?

So eventually when mlx4_en_replace_mac()  returns with error, there is a 
printout, so users
will at least be notified that something went wrong.

Or.