[pull request][net 00/14] Mellanox, mlx5 fixes 2017-07-27

[pull request][net 00/14] Mellanox, mlx5 fixes 2017-07-27

[Saeed Mahameed]

Hi Dave,

This series contains some misc fixes to the mlx5 driver.

Please pull and let me know if there's any problem.

for -stable:
4.7
net/mlx5: Fix command bad flow on command entry allocation failure

4.9
net/mlx5: Consider tx_enabled in all modes on remap
net/mlx5e: Fix outer_header_zero() check size

4.10
net/mlx5: Fix mlx5_add_flow_rules call with correct num of dests

4.11
net/mlx5: Fix mlx5_ifc_mtpps_reg_bits structure size
net/mlx5e: Add field select to MTPPS register
net/mlx5e: Fix broken disable 1PPS flow
net/mlx5e: Change 1PPS out scheme
net/mlx5e: Add missing support for PTP_CLK_REQ_PPS request
net/mlx5e: Fix wrong delay calculation for overflow check scheduling
net/mlx5e: Schedule overflow check work to mlx5e workqueue

4.12
net/mlx5: Fix command completion after timeout access invalid structure
net/mlx5e: IPoIB, Modify add/remove underlay QPN flows

I hope this is not too much, but most of the patches do apply cleanly on -stable.

Thanks,
Saeed.

---

The following changes since commit 6b84202c946cd3da3a8daa92c682510e9ed80321:

  sctp: fix the check for _sctp_walk_params and _sctp_walk_errors (2017-07-27 00:03:12 -0700)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-fixes-2017-07-27

for you to fetch changes up to db1e7fa982d0b183501f95010118a1c6f796c838:

  net/mlx5: Fix mlx5_add_flow_rules call with correct num of dests (2017-07-27 11:11:37 +0300)

----------------------------------------------------------------
mlx5-fixes-2017-07-27

----------------------------------------------------------------
Alex Vesker (1):
      net/mlx5e: IPoIB, Modify add/remove underlay QPN flows

Aviv Heller (1):
      net/mlx5: Consider tx_enabled in all modes on remap

Eran Ben Elisha (1):
      net/mlx5: Clean SRIOV eswitch resources upon VF creation failure

Eugenia Emantayev (7):
      net/mlx5: Fix mlx5_ifc_mtpps_reg_bits structure size
      net/mlx5e: Add field select to MTPPS register
      net/mlx5e: Fix broken disable 1PPS flow
      net/mlx5e: Change 1PPS out scheme
      net/mlx5e: Add missing support for PTP_CLK_REQ_PPS request
      net/mlx5e: Fix wrong delay calculation for overflow check scheduling
      net/mlx5e: Schedule overflow check work to mlx5e workqueue

Ilan Tayari (1):
      net/mlx5e: Fix outer_header_zero() check size

Moshe Shemesh (2):
      net/mlx5: Fix command completion after timeout access invalid structure
      net/mlx5: Fix command bad flow on command entry allocation failure

Paul Blakey (1):
      net/mlx5: Fix mlx5_add_flow_rules call with correct num of dests

 drivers/net/ethernet/mellanox/mlx5/core/cmd.c      |  25 ++-
 drivers/net/ethernet/mellanox/mlx5/core/en.h       |  10 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_clock.c | 222 +++++++++++++++------
 .../ethernet/mellanox/mlx5/core/en_fs_ethtool.c    |   4 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c  |   1 -
 drivers/net/ethernet/mellanox/mlx5/core/eq.c       |   2 +-
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c  |   3 +-
 .../net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c  |  16 +-
 drivers/net/ethernet/mellanox/mlx5/core/lag.c      |  26 +--
 .../net/ethernet/mellanox/mlx5/core/mlx5_core.h    |   5 +
 drivers/net/ethernet/mellanox/mlx5/core/sriov.c    |   5 +
 include/linux/mlx5/mlx5_ifc.h                      |  10 +-
 12 files changed, 233 insertions(+), 96 deletions(-)

[net 01/14] net/mlx5: Clean SRIOV eswitch resources upon VF creation failure

[Saeed Mahameed]

From: Eran Ben Elisha <eranbe@mellanox.com>

Upon sriov enable, eswitch is always enabled.
Currently, if enable hca failed over all VFs, we would skip eswitch
disable as part of sriov disable, which will lead to resources leak.

Fix it by disabling eswitch if it was enabled (use indication from
eswitch mode).

Fixes: 6b6adee3dad2 ('net/mlx5: SRIOV core code refactoring')
Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Noa Osherovich <noaos@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 3 ++-
 drivers/net/ethernet/mellanox/mlx5/core/sriov.c   | 5 +++++
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 89bfda419efe..8b18cc9ec026 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1668,7 +1668,8 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw)
 	int i;
 
 	if (!esw || !MLX5_CAP_GEN(esw->dev, vport_group_manager) ||
-	    MLX5_CAP_GEN(esw->dev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+	    MLX5_CAP_GEN(esw->dev, port_type) != MLX5_CAP_PORT_TYPE_ETH ||
+	    esw->mode == SRIOV_NONE)
 		return;
 
 	esw_info(esw->dev, "disable SRIOV: active vports(%d) mode(%d)\n",
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c
index bcdf7779c48d..bf99d40e30b4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c
@@ -88,7 +88,11 @@ static void mlx5_device_disable_sriov(struct mlx5_core_dev *dev)
 	int vf;
 
 	if (!sriov->enabled_vfs)
+#ifdef CONFIG_MLX5_CORE_EN
+		goto disable_sriov_resources;
+#else
 		return;
+#endif
 
 	for (vf = 0; vf < sriov->num_vfs; vf++) {
 		if (!sriov->vfs_ctx[vf].enabled)
@@ -103,6 +107,7 @@ static void mlx5_device_disable_sriov(struct mlx5_core_dev *dev)
 	}
 
 #ifdef CONFIG_MLX5_CORE_EN
+disable_sriov_resources:
 	mlx5_eswitch_disable_sriov(dev->priv.eswitch);
 #endif
 
-- 
2.13.0

[net 02/14] net/mlx5: Consider tx_enabled in all modes on remap

[Saeed Mahameed]

From: Aviv Heller <avivh@mellanox.com>

The tx_enabled lag event field is used to determine whether a slave is
active.
Current logic uses this value only if the mode is active-backup.

However, LACP mode, although considered a load balancing mode, can mark
a slave as inactive in certain situations (e.g., LACP timeout).

This fix takes the tx_enabled value into account when remapping, with
no respect to the LAG mode (this should not affect the behavior in XOR
mode, since in this mode both slaves are marked as active).

Fixes: 7907f23adc18 (net/mlx5: Implement RoCE LAG feature)
Signed-off-by: Aviv Heller <avivh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/lag.c | 26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
index a3a836bdcfd2..654b02e7d1aa 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -162,21 +162,17 @@ static bool mlx5_lag_is_bonded(struct mlx5_lag *ldev)
 static void mlx5_infer_tx_affinity_mapping(struct lag_tracker *tracker,
 					   u8 *port1, u8 *port2)
 {
-	if (tracker->tx_type == NETDEV_LAG_TX_TYPE_ACTIVEBACKUP) {
-		if (tracker->netdev_state[0].tx_enabled) {
-			*port1 = 1;
-			*port2 = 1;
-		} else {
-			*port1 = 2;
-			*port2 = 2;
-		}
-	} else {
-		*port1 = 1;
-		*port2 = 2;
-		if (!tracker->netdev_state[0].link_up)
-			*port1 = 2;
-		else if (!tracker->netdev_state[1].link_up)
-			*port2 = 1;
+	*port1 = 1;
+	*port2 = 2;
+	if (!tracker->netdev_state[0].tx_enabled ||
+	    !tracker->netdev_state[0].link_up) {
+		*port1 = 2;
+		return;
+	}
+
+	if (!tracker->netdev_state[1].tx_enabled ||
+	    !tracker->netdev_state[1].link_up) {
+		*port2 = 1;
 	}
 }
 
-- 
2.13.0

Re: [net 02/14] net/mlx5: Consider tx_enabled in all modes on remap

[Sergei Shtylyov]

Hello!

On 07/27/2017 01:25 PM, Saeed Mahameed wrote:

> From: Aviv Heller <avivh@mellanox.com>
>
> The tx_enabled lag event field is used to determine whether a slave is
> active.
> Current logic uses this value only if the mode is active-backup.
>
> However, LACP mode, although considered a load balancing mode, can mark
> a slave as inactive in certain situations (e.g., LACP timeout).
>
> This fix takes the tx_enabled value into account when remapping, with
> no respect to the LAG mode (this should not affect the behavior in XOR
> mode, since in this mode both slaves are marked as active).
>
> Fixes: 7907f23adc18 (net/mlx5: Implement RoCE LAG feature)
> Signed-off-by: Aviv Heller <avivh@mellanox.com>
> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
> ---
>  drivers/net/ethernet/mellanox/mlx5/core/lag.c | 26 +++++++++++---------------
>  1 file changed, 11 insertions(+), 15 deletions(-)
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
> index a3a836bdcfd2..654b02e7d1aa 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
> @@ -162,21 +162,17 @@ static bool mlx5_lag_is_bonded(struct mlx5_lag *ldev)
>  static void mlx5_infer_tx_affinity_mapping(struct lag_tracker *tracker,
>  					   u8 *port1, u8 *port2)
>  {
> -	if (tracker->tx_type == NETDEV_LAG_TX_TYPE_ACTIVEBACKUP) {
> -		if (tracker->netdev_state[0].tx_enabled) {
> -			*port1 = 1;
> -			*port2 = 1;
> -		} else {
> -			*port1 = 2;
> -			*port2 = 2;
> -		}
> -	} else {
> -		*port1 = 1;
> -		*port2 = 2;
> -		if (!tracker->netdev_state[0].link_up)
> -			*port1 = 2;
> -		else if (!tracker->netdev_state[1].link_up)
> -			*port2 = 1;
> +	*port1 = 1;
> +	*port2 = 2;
> +	if (!tracker->netdev_state[0].tx_enabled ||
> +	    !tracker->netdev_state[0].link_up) {
> +		*port1 = 2;
> +		return;
> +	}
> +
> +	if (!tracker->netdev_state[1].tx_enabled ||
> +	    !tracker->netdev_state[1].link_up) {
> +		*port2 = 1;
>  	}

    {} not needed here.

[...]

MBR, Sergei

Re: [net 02/14] net/mlx5: Consider tx_enabled in all modes on remap

[Saeed Mahameed]

On Thu, Jul 27, 2017 at 2:55 PM, Sergei Shtylyov
<sergei.shtylyov@cogentembedded.com> wrote:
> Hello!
>
>
> On 07/27/2017 01:25 PM, Saeed Mahameed wrote:
>
>> From: Aviv Heller <avivh@mellanox.com>
>>
>> The tx_enabled lag event field is used to determine whether a slave is
>> active.
>> Current logic uses this value only if the mode is active-backup.
>>
>> However, LACP mode, although considered a load balancing mode, can mark
>> a slave as inactive in certain situations (e.g., LACP timeout).
>>
>> This fix takes the tx_enabled value into account when remapping, with
>> no respect to the LAG mode (this should not affect the behavior in XOR
>> mode, since in this mode both slaves are marked as active).
>>
>> Fixes: 7907f23adc18 (net/mlx5: Implement RoCE LAG feature)
>> Signed-off-by: Aviv Heller <avivh@mellanox.com>
>> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
>> ---
>>  drivers/net/ethernet/mellanox/mlx5/core/lag.c | 26
>> +++++++++++---------------
>>  1 file changed, 11 insertions(+), 15 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
>> b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
>> index a3a836bdcfd2..654b02e7d1aa 100644
>> --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
>> +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
>> @@ -162,21 +162,17 @@ static bool mlx5_lag_is_bonded(struct mlx5_lag
>> *ldev)
>>  static void mlx5_infer_tx_affinity_mapping(struct lag_tracker *tracker,
>>                                            u8 *port1, u8 *port2)
>>  {
>> -       if (tracker->tx_type == NETDEV_LAG_TX_TYPE_ACTIVEBACKUP) {
>> -               if (tracker->netdev_state[0].tx_enabled) {
>> -                       *port1 = 1;
>> -                       *port2 = 1;
>> -               } else {
>> -                       *port1 = 2;
>> -                       *port2 = 2;
>> -               }
>> -       } else {
>> -               *port1 = 1;
>> -               *port2 = 2;
>> -               if (!tracker->netdev_state[0].link_up)
>> -                       *port1 = 2;
>> -               else if (!tracker->netdev_state[1].link_up)
>> -                       *port2 = 1;
>> +       *port1 = 1;
>> +       *port2 = 2;
>> +       if (!tracker->netdev_state[0].tx_enabled ||
>> +           !tracker->netdev_state[0].link_up) {
>> +               *port1 = 2;
>> +               return;
>> +       }
>> +
>> +       if (!tracker->netdev_state[1].tx_enabled ||
>> +           !tracker->netdev_state[1].link_up) {
>> +               *port2 = 1;
>>         }
>
>
>    {} not needed here.

Right, will fix and resubmit, Thanks Sergei.

>
> [...]
>
> MBR, Sergei
>

[net 03/14] net/mlx5: Fix command completion after timeout access invalid structure

[Saeed Mahameed]

From: Moshe Shemesh <moshe@mellanox.com>

Completion on timeout should not free the driver command entry structure
as it will need to access it again once real completion event from FW
will occur.

Fixes: 73dd3a4839c1 ('net/mlx5: Avoid using pending command interface slots')
Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Cc: kernel-team@fb.com
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
index f5a2c605749f..25fd32fcdf79 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
@@ -967,7 +967,7 @@ static int mlx5_cmd_invoke(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *in,
 
 	err = wait_func(dev, ent);
 	if (err == -ETIMEDOUT)
-		goto out_free;
+		goto out;
 
 	ds = ent->ts2 - ent->ts1;
 	op = MLX5_GET(mbox_in, in->first.data, opcode);
@@ -1430,6 +1430,7 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced)
 					mlx5_core_err(dev, "Command completion arrived after timeout (entry idx = %d).\n",
 						      ent->idx);
 					free_ent(cmd, ent->idx);
+					free_cmd(ent);
 				}
 				continue;
 			}
@@ -1488,7 +1489,8 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced)
 				free_msg(dev, ent->in);
 
 				err = err ? err : ent->status;
-				free_cmd(ent);
+				if (!forced)
+					free_cmd(ent);
 				callback(err, context);
 			} else {
 				complete(&ent->done);
-- 
2.13.0

[net 04/14] net/mlx5: Fix command bad flow on command entry allocation failure

[Saeed Mahameed]

From: Moshe Shemesh <moshe@mellanox.com>

When driver fail to allocate an entry to send command to FW, it must
notify the calling function and release the memory allocated for
this command.

Fixes: e126ba97dba9e ('mlx5: Add driver for Mellanox Connect-IB adapters')
Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Cc: kernel-team@fb.com
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
index 25fd32fcdf79..31cbe5e86a01 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
@@ -786,6 +786,10 @@ static void cb_timeout_handler(struct work_struct *work)
 	mlx5_cmd_comp_handler(dev, 1UL << ent->idx, true);
 }
 
+static void free_msg(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *msg);
+static void mlx5_free_cmd_msg(struct mlx5_core_dev *dev,
+			      struct mlx5_cmd_msg *msg);
+
 static void cmd_work_handler(struct work_struct *work)
 {
 	struct mlx5_cmd_work_ent *ent = container_of(work, struct mlx5_cmd_work_ent, work);
@@ -796,17 +800,28 @@ static void cmd_work_handler(struct work_struct *work)
 	struct semaphore *sem;
 	unsigned long flags;
 	bool poll_cmd = ent->polling;
+	int alloc_ret;
 
 
 	sem = ent->page_queue ? &cmd->pages_sem : &cmd->sem;
 	down(sem);
 	if (!ent->page_queue) {
-		ent->idx = alloc_ent(cmd);
-		if (ent->idx < 0) {
+		alloc_ret = alloc_ent(cmd);
+		if (alloc_ret < 0) {
 			mlx5_core_err(dev, "failed to allocate command entry\n");
+			if (ent->callback) {
+				ent->callback(-EAGAIN, ent->context);
+				mlx5_free_cmd_msg(dev, ent->out);
+				free_msg(dev, ent->in);
+				free_cmd(ent);
+			} else {
+				ent->ret = -EAGAIN;
+				complete(&ent->done);
+			}
 			up(sem);
 			return;
 		}
+		ent->idx = alloc_ret;
 	} else {
 		ent->idx = cmd->max_reg_cmds;
 		spin_lock_irqsave(&cmd->alloc_lock, flags);
-- 
2.13.0

[net 05/14] net/mlx5e: IPoIB, Modify add/remove underlay QPN flows

[Saeed Mahameed]

From: Alex Vesker <valex@mellanox.com>

On interface remove, the clean-up was done incorrectly causing
an error in the log:
"SET_FLOW_TABLE_ROOT(0x92f) op_mod(0x0) failed...syndrome (0x7e9f14)"

This was caused by the following flow:
-ndo_uninit:
 Move QP state to RST (this disconnects the QP from FT),
 the QP cannot be attached to any FT unless it is in RTS.

-mlx5_rdma_netdev_free:
 cleanup_rx: Destroy FT
 cleanup_tx: Destroy QP and remove QPN from FT

This caused a problem when destroying current FT we tried to
re-attach the QP to the next FT which is not needed.

The correct flow is:
-mlx5_rdma_netdev_free:
	cleanup_rx: remove QPN from FT & Destroy FT
	cleanup_tx: Destroy QP

Fixes: 508541146af1 ("net/mlx5: Use underlay QPN from the root name space")
Signed-off-by: Alex Vesker <valex@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index 1ee5bce85901..85298051a3e4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -178,8 +178,6 @@ static int mlx5i_create_underlay_qp(struct mlx5_core_dev *mdev, struct mlx5_core
 
 static void mlx5i_destroy_underlay_qp(struct mlx5_core_dev *mdev, struct mlx5_core_qp *qp)
 {
-	mlx5_fs_remove_rx_underlay_qpn(mdev, qp->qpn);
-
 	mlx5_core_destroy_qp(mdev, qp);
 }
 
@@ -194,8 +192,6 @@ static int mlx5i_init_tx(struct mlx5e_priv *priv)
 		return err;
 	}
 
-	mlx5_fs_add_rx_underlay_qpn(priv->mdev, ipriv->qp.qpn);
-
 	err = mlx5e_create_tis(priv->mdev, 0 /* tc */, ipriv->qp.qpn, &priv->tisn[0]);
 	if (err) {
 		mlx5_core_warn(priv->mdev, "create tis failed, %d\n", err);
@@ -253,6 +249,7 @@ static void mlx5i_destroy_flow_steering(struct mlx5e_priv *priv)
 
 static int mlx5i_init_rx(struct mlx5e_priv *priv)
 {
+	struct mlx5i_priv *ipriv  = priv->ppriv;
 	int err;
 
 	err = mlx5e_create_indirect_rqt(priv);
@@ -271,12 +268,18 @@ static int mlx5i_init_rx(struct mlx5e_priv *priv)
 	if (err)
 		goto err_destroy_indirect_tirs;
 
-	err = mlx5i_create_flow_steering(priv);
+	err = mlx5_fs_add_rx_underlay_qpn(priv->mdev, ipriv->qp.qpn);
 	if (err)
 		goto err_destroy_direct_tirs;
 
+	err = mlx5i_create_flow_steering(priv);
+	if (err)
+		goto err_remove_rx_underlay_qpn;
+
 	return 0;
 
+err_remove_rx_underlay_qpn:
+	mlx5_fs_remove_rx_underlay_qpn(priv->mdev, ipriv->qp.qpn);
 err_destroy_direct_tirs:
 	mlx5e_destroy_direct_tirs(priv);
 err_destroy_indirect_tirs:
@@ -290,6 +293,9 @@ static int mlx5i_init_rx(struct mlx5e_priv *priv)
 
 static void mlx5i_cleanup_rx(struct mlx5e_priv *priv)
 {
+	struct mlx5i_priv *ipriv  = priv->ppriv;
+
+	mlx5_fs_remove_rx_underlay_qpn(priv->mdev, ipriv->qp.qpn);
 	mlx5i_destroy_flow_steering(priv);
 	mlx5e_destroy_direct_tirs(priv);
 	mlx5e_destroy_indirect_tirs(priv);
-- 
2.13.0

[net 06/14] net/mlx5e: Fix outer_header_zero() check size

[Saeed Mahameed]

From: Ilan Tayari <ilant@mellanox.com>

outer_header_zero() routine checks if the outer_headers match of a
flow-table entry are all zero.

This function uses the size of whole fte_match_param, instead of just
the outer_headers member, causing failure to detect all-zeros if
any other members of the fte_match_param are non-zero.

Use the correct size for zero check.

Fixes: 6dc6071cfcde ("net/mlx5e: Add ethtool flow steering support")
Signed-off-by: Ilan Tayari <ilant@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
index bdd82c9b3992..22fb993987b4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
@@ -276,7 +276,7 @@ static void add_rule_to_list(struct mlx5e_priv *priv,
 
 static bool outer_header_zero(u32 *match_criteria)
 {
-	int size = MLX5_ST_SZ_BYTES(fte_match_param);
+	int size = MLX5_FLD_SZ_BYTES(fte_match_param, outer_headers);
 	char *outer_headers_c = MLX5_ADDR_OF(fte_match_param, match_criteria,
 					     outer_headers);
 
-- 
2.13.0

[net 07/14] net/mlx5: Fix mlx5_ifc_mtpps_reg_bits structure size

[Saeed Mahameed]

From: Eugenia Emantayev <eugenia@mellanox.com>

Fix miscalculation in reserved_at_1a0 field.

Fixes: ee7f12205abc ('net/mlx5e: Implement 1PPS support')
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 include/linux/mlx5/mlx5_ifc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index 87869c04849a..fd98aef4545c 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -8175,7 +8175,7 @@ struct mlx5_ifc_mtpps_reg_bits {
 	u8         out_pulse_duration[0x10];
 	u8         out_periodic_adjustment[0x10];
 
-	u8         reserved_at_1a0[0x60];
+	u8         reserved_at_1a0[0x40];
 };
 
 struct mlx5_ifc_mtppse_reg_bits {
-- 
2.13.0

[net 08/14] net/mlx5e: Add field select to MTPPS register

[Saeed Mahameed]

From: Eugenia Emantayev <eugenia@mellanox.com>

In order to mark relevant fields while setting the MTPPS register
add field select. Otherwise it can cause a misconfiguration in
firmware.

Fixes: ee7f12205abc ('net/mlx5e: Implement 1PPS support')
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_clock.c | 29 +++++++++++++++++-----
 drivers/net/ethernet/mellanox/mlx5/core/eq.c       |  2 +-
 .../net/ethernet/mellanox/mlx5/core/mlx5_core.h    |  5 ++++
 include/linux/mlx5/mlx5_ifc.h                      | 10 +++++---
 4 files changed, 36 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
index 66f432385dbb..ab07233e2faa 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
@@ -53,6 +53,15 @@ enum {
 	MLX5E_EVENT_MODE_ONCE_TILL_ARM	= 0x2,
 };
 
+enum {
+	MLX5E_MTPPS_FS_ENABLE			= BIT(0x0),
+	MLX5E_MTPPS_FS_PATTERN			= BIT(0x2),
+	MLX5E_MTPPS_FS_PIN_MODE			= BIT(0x3),
+	MLX5E_MTPPS_FS_TIME_STAMP		= BIT(0x4),
+	MLX5E_MTPPS_FS_OUT_PULSE_DURATION	= BIT(0x5),
+	MLX5E_MTPPS_FS_ENH_OUT_PER_ADJ		= BIT(0x7),
+};
+
 void mlx5e_fill_hwstamp(struct mlx5e_tstamp *tstamp, u64 timestamp,
 			struct skb_shared_hwtstamps *hwts)
 {
@@ -221,7 +230,10 @@ static int mlx5e_ptp_adjfreq(struct ptp_clock_info *ptp, s32 delta)
 
 		/* For future use need to add a loop for finding all 1PPS out pins */
 		MLX5_SET(mtpps_reg, in, pin_mode, MLX5E_PIN_MODE_OUT);
-		MLX5_SET(mtpps_reg, in, out_periodic_adjustment, delta & 0xFFFF);
+		MLX5_SET(mtpps_reg, in, enhanced_out_periodic_adjustment, delta);
+		MLX5_SET(mtpps_reg, in, field_select,
+			 MLX5E_MTPPS_FS_PIN_MODE |
+			 MLX5E_MTPPS_FS_ENH_OUT_PER_ADJ);
 
 		mlx5_set_mtpps(priv->mdev, in, sizeof(in));
 	}
@@ -257,8 +269,7 @@ static int mlx5e_extts_configure(struct ptp_clock_info *ptp,
 	int pin = -1;
 	int err = 0;
 
-	if (!MLX5_CAP_GEN(priv->mdev, pps) ||
-	    !MLX5_CAP_GEN(priv->mdev, pps_modify))
+	if (!MLX5_PPS_CAP(priv->mdev))
 		return -EOPNOTSUPP;
 
 	if (rq->extts.index >= tstamp->ptp_info.n_pins)
@@ -277,6 +288,9 @@ static int mlx5e_extts_configure(struct ptp_clock_info *ptp,
 	MLX5_SET(mtpps_reg, in, pin_mode, MLX5E_PIN_MODE_IN);
 	MLX5_SET(mtpps_reg, in, pattern, pattern);
 	MLX5_SET(mtpps_reg, in, enable, on);
+	MLX5_SET(mtpps_reg, in, field_select, MLX5E_MTPPS_FS_PIN_MODE |
+					      MLX5E_MTPPS_FS_PATTERN |
+					      MLX5E_MTPPS_FS_ENABLE);
 
 	err = mlx5_set_mtpps(priv->mdev, in, sizeof(in));
 	if (err)
@@ -302,7 +316,7 @@ static int mlx5e_perout_configure(struct ptp_clock_info *ptp,
 	int pin = -1;
 	s64 ns;
 
-	if (!MLX5_CAP_GEN(priv->mdev, pps_modify))
+	if (!MLX5_PPS_CAP(priv->mdev))
 		return -EOPNOTSUPP;
 
 	if (rq->perout.index >= tstamp->ptp_info.n_pins)
@@ -337,7 +351,10 @@ static int mlx5e_perout_configure(struct ptp_clock_info *ptp,
 	MLX5_SET(mtpps_reg, in, pattern, MLX5E_OUT_PATTERN_PERIODIC);
 	MLX5_SET(mtpps_reg, in, enable, on);
 	MLX5_SET64(mtpps_reg, in, time_stamp, time_stamp);
-
+	MLX5_SET(mtpps_reg, in, field_select, MLX5E_MTPPS_FS_PIN_MODE |
+					      MLX5E_MTPPS_FS_PATTERN |
+					      MLX5E_MTPPS_FS_ENABLE |
+					      MLX5E_MTPPS_FS_TIME_STAMP);
 	return mlx5_set_mtpps(priv->mdev, in, sizeof(in));
 }
 
@@ -487,7 +504,7 @@ void mlx5e_timestamp_init(struct mlx5e_priv *priv)
 #define MAX_PIN_NUM	8
 	tstamp->pps_pin_caps = kzalloc(sizeof(u8) * MAX_PIN_NUM, GFP_KERNEL);
 	if (tstamp->pps_pin_caps) {
-		if (MLX5_CAP_GEN(priv->mdev, pps))
+		if (MLX5_PPS_CAP(priv->mdev))
 			mlx5e_get_pps_caps(priv, tstamp);
 		if (tstamp->ptp_info.n_pins)
 			mlx5e_init_pin_config(tstamp);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
index af51a5d2b912..52b9a64cd3a2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
@@ -698,7 +698,7 @@ int mlx5_start_eqs(struct mlx5_core_dev *dev)
 	else
 		mlx5_core_dbg(dev, "port_module_event is not set\n");
 
-	if (MLX5_CAP_GEN(dev, pps))
+	if (MLX5_PPS_CAP(dev))
 		async_event_mask |= (1ull << MLX5_EVENT_TYPE_PPS_EVENT);
 
 	if (MLX5_CAP_GEN(dev, fpga))
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index 6a3d6bef7dd4..6a263e8d883a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -154,6 +154,11 @@ int mlx5_set_mtpps(struct mlx5_core_dev *mdev, u32 *mtpps, u32 mtpps_size);
 int mlx5_query_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 *arm, u8 *mode);
 int mlx5_set_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 arm, u8 mode);
 
+#define MLX5_PPS_CAP(mdev) (MLX5_CAP_GEN((mdev), pps) &&		\
+			    MLX5_CAP_GEN((mdev), pps_modify) &&		\
+			    MLX5_CAP_MCAM_FEATURE((mdev), mtpps_fs) &&	\
+			    MLX5_CAP_MCAM_FEATURE((mdev), mtpps_enh_out_per_adj))
+
 int mlx5_firmware_flash(struct mlx5_core_dev *dev, const struct firmware *fw);
 
 void mlx5e_init(void);
diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index fd98aef4545c..3030121b4746 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -7749,8 +7749,10 @@ struct mlx5_ifc_pcam_reg_bits {
 };
 
 struct mlx5_ifc_mcam_enhanced_features_bits {
-	u8         reserved_at_0[0x7f];
+	u8         reserved_at_0[0x7d];
 
+	u8         mtpps_enh_out_per_adj[0x1];
+	u8         mtpps_fs[0x1];
 	u8         pcie_performance_group[0x1];
 };
 
@@ -8159,7 +8161,8 @@ struct mlx5_ifc_mtpps_reg_bits {
 	u8         reserved_at_78[0x4];
 	u8         cap_pin_4_mode[0x4];
 
-	u8         reserved_at_80[0x80];
+	u8         field_select[0x20];
+	u8         reserved_at_a0[0x60];
 
 	u8         enable[0x1];
 	u8         reserved_at_101[0xb];
@@ -8174,8 +8177,9 @@ struct mlx5_ifc_mtpps_reg_bits {
 
 	u8         out_pulse_duration[0x10];
 	u8         out_periodic_adjustment[0x10];
+	u8         enhanced_out_periodic_adjustment[0x20];
 
-	u8         reserved_at_1a0[0x40];
+	u8         reserved_at_1c0[0x20];
 };
 
 struct mlx5_ifc_mtppse_reg_bits {
-- 
2.13.0

[net 09/14] net/mlx5e: Fix broken disable 1PPS flow

[Saeed Mahameed]

From: Eugenia Emantayev <eugenia@mellanox.com>

Need to disable the MTPPS and unsubscribe from the pulse events
when user disables the 1PPS functionality.

Fixes: ee7f12205abc ('net/mlx5e: Implement 1PPS support')
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_clock.c | 75 +++++++++++++---------
 1 file changed, 46 insertions(+), 29 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
index ab07233e2faa..25d43767c888 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
@@ -265,6 +265,8 @@ static int mlx5e_extts_configure(struct ptp_clock_info *ptp,
 	struct mlx5e_priv *priv =
 		container_of(tstamp, struct mlx5e_priv, tstamp);
 	u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
+	u32 field_select = 0;
+	u8 pin_mode = 0;
 	u8 pattern = 0;
 	int pin = -1;
 	int err = 0;
@@ -279,18 +281,21 @@ static int mlx5e_extts_configure(struct ptp_clock_info *ptp,
 		pin = ptp_find_pin(tstamp->ptp, PTP_PF_EXTTS, rq->extts.index);
 		if (pin < 0)
 			return -EBUSY;
+		pin_mode = MLX5E_PIN_MODE_IN;
+		pattern = !!(rq->extts.flags & PTP_FALLING_EDGE);
+		field_select = MLX5E_MTPPS_FS_PIN_MODE |
+			       MLX5E_MTPPS_FS_PATTERN |
+			       MLX5E_MTPPS_FS_ENABLE;
+	} else {
+		pin = rq->extts.index;
+		field_select = MLX5E_MTPPS_FS_ENABLE;
 	}
 
-	if (rq->extts.flags & PTP_FALLING_EDGE)
-		pattern = 1;
-
 	MLX5_SET(mtpps_reg, in, pin, pin);
-	MLX5_SET(mtpps_reg, in, pin_mode, MLX5E_PIN_MODE_IN);
+	MLX5_SET(mtpps_reg, in, pin_mode, pin_mode);
 	MLX5_SET(mtpps_reg, in, pattern, pattern);
 	MLX5_SET(mtpps_reg, in, enable, on);
-	MLX5_SET(mtpps_reg, in, field_select, MLX5E_MTPPS_FS_PIN_MODE |
-					      MLX5E_MTPPS_FS_PATTERN |
-					      MLX5E_MTPPS_FS_ENABLE);
+	MLX5_SET(mtpps_reg, in, field_select, field_select);
 
 	err = mlx5_set_mtpps(priv->mdev, in, sizeof(in));
 	if (err)
@@ -313,6 +318,9 @@ static int mlx5e_perout_configure(struct ptp_clock_info *ptp,
 	u64 cycles_now, cycles_delta;
 	struct timespec64 ts;
 	unsigned long flags;
+	u32 field_select = 0;
+	u8 pin_mode = 0;
+	u8 pattern = 0;
 	int pin = -1;
 	s64 ns;
 
@@ -327,34 +335,43 @@ static int mlx5e_perout_configure(struct ptp_clock_info *ptp,
 				   rq->perout.index);
 		if (pin < 0)
 			return -EBUSY;
-	}
 
-	ts.tv_sec = rq->perout.period.sec;
-	ts.tv_nsec = rq->perout.period.nsec;
-	ns = timespec64_to_ns(&ts);
-	if (on)
+		pin_mode = MLX5E_PIN_MODE_OUT;
+		pattern = MLX5E_OUT_PATTERN_PERIODIC;
+		ts.tv_sec = rq->perout.period.sec;
+		ts.tv_nsec = rq->perout.period.nsec;
+		ns = timespec64_to_ns(&ts);
+
 		if ((ns >> 1) != 500000000LL)
 			return -EINVAL;
-	ts.tv_sec = rq->perout.start.sec;
-	ts.tv_nsec = rq->perout.start.nsec;
-	ns = timespec64_to_ns(&ts);
-	cycles_now = mlx5_read_internal_timer(tstamp->mdev);
-	write_lock_irqsave(&tstamp->lock, flags);
-	nsec_now = timecounter_cyc2time(&tstamp->clock, cycles_now);
-	nsec_delta = ns - nsec_now;
-	cycles_delta = div64_u64(nsec_delta << tstamp->cycles.shift,
-				 tstamp->cycles.mult);
-	write_unlock_irqrestore(&tstamp->lock, flags);
-	time_stamp = cycles_now + cycles_delta;
+
+		ts.tv_sec = rq->perout.start.sec;
+		ts.tv_nsec = rq->perout.start.nsec;
+		ns = timespec64_to_ns(&ts);
+		cycles_now = mlx5_read_internal_timer(tstamp->mdev);
+		write_lock_irqsave(&tstamp->lock, flags);
+		nsec_now = timecounter_cyc2time(&tstamp->clock, cycles_now);
+		nsec_delta = ns - nsec_now;
+		cycles_delta = div64_u64(nsec_delta << tstamp->cycles.shift,
+					 tstamp->cycles.mult);
+		write_unlock_irqrestore(&tstamp->lock, flags);
+		time_stamp = cycles_now + cycles_delta;
+		field_select = MLX5E_MTPPS_FS_PIN_MODE |
+			       MLX5E_MTPPS_FS_PATTERN |
+			       MLX5E_MTPPS_FS_ENABLE |
+			       MLX5E_MTPPS_FS_TIME_STAMP;
+	} else {
+		pin = rq->perout.index;
+		field_select = MLX5E_MTPPS_FS_ENABLE;
+	}
+
 	MLX5_SET(mtpps_reg, in, pin, pin);
-	MLX5_SET(mtpps_reg, in, pin_mode, MLX5E_PIN_MODE_OUT);
-	MLX5_SET(mtpps_reg, in, pattern, MLX5E_OUT_PATTERN_PERIODIC);
+	MLX5_SET(mtpps_reg, in, pin_mode, pin_mode);
+	MLX5_SET(mtpps_reg, in, pattern, pattern);
 	MLX5_SET(mtpps_reg, in, enable, on);
 	MLX5_SET64(mtpps_reg, in, time_stamp, time_stamp);
-	MLX5_SET(mtpps_reg, in, field_select, MLX5E_MTPPS_FS_PIN_MODE |
-					      MLX5E_MTPPS_FS_PATTERN |
-					      MLX5E_MTPPS_FS_ENABLE |
-					      MLX5E_MTPPS_FS_TIME_STAMP);
+	MLX5_SET(mtpps_reg, in, field_select, field_select);
+
 	return mlx5_set_mtpps(priv->mdev, in, sizeof(in));
 }
 
-- 
2.13.0

[net 10/14] net/mlx5e: Change 1PPS out scheme

[Saeed Mahameed]

From: Eugenia Emantayev <eugenia@mellanox.com>

In order to fix the drift in 1PPS out need to adjust the next pulse.
On each 1PPS out falling edge driver gets the event, then the event
handler adjusts the next pulse starting time.

Fixes: ee7f12205abc ('net/mlx5e: Implement 1PPS support')
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h       |   9 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_clock.c | 116 ++++++++++++++-------
 2 files changed, 87 insertions(+), 38 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index e1b7ddfecd01..d42826f0f0d6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -266,6 +266,13 @@ struct mlx5e_dcbx {
 };
 #endif
 
+#define MAX_PIN_NUM	8
+struct mlx5e_pps {
+	u8                         pin_caps[MAX_PIN_NUM];
+	struct work_struct         out_work;
+	u64                        start[MAX_PIN_NUM];
+};
+
 struct mlx5e_tstamp {
 	rwlock_t                   lock;
 	struct cyclecounter        cycles;
@@ -277,7 +284,7 @@ struct mlx5e_tstamp {
 	struct mlx5_core_dev      *mdev;
 	struct ptp_clock          *ptp;
 	struct ptp_clock_info      ptp_info;
-	u8                        *pps_pin_caps;
+	struct mlx5e_pps           pps_info;
 };
 
 enum {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
index 25d43767c888..464ddd10ebbc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
@@ -82,6 +82,33 @@ static u64 mlx5e_read_internal_timer(const struct cyclecounter *cc)
 	return mlx5_read_internal_timer(tstamp->mdev) & cc->mask;
 }
 
+static void mlx5e_pps_out(struct work_struct *work)
+{
+	struct mlx5e_pps *pps_info = container_of(work, struct mlx5e_pps,
+						  out_work);
+	struct mlx5e_tstamp *tstamp = container_of(pps_info, struct mlx5e_tstamp,
+						   pps_info);
+	u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
+	unsigned long flags;
+	int i;
+
+	for (i = 0; i < tstamp->ptp_info.n_pins; i++) {
+		u64 tstart;
+
+		write_lock_irqsave(&tstamp->lock, flags);
+		tstart = tstamp->pps_info.start[i];
+		tstamp->pps_info.start[i] = 0;
+		write_unlock_irqrestore(&tstamp->lock, flags);
+		if (!tstart)
+			continue;
+
+		MLX5_SET(mtpps_reg, in, pin, i);
+		MLX5_SET64(mtpps_reg, in, time_stamp, tstart);
+		MLX5_SET(mtpps_reg, in, field_select, MLX5E_MTPPS_FS_TIME_STAMP);
+		mlx5_set_mtpps(tstamp->mdev, in, sizeof(in));
+	}
+}
+
 static void mlx5e_timestamp_overflow(struct work_struct *work)
 {
 	struct delayed_work *dwork = to_delayed_work(work);
@@ -222,21 +249,6 @@ static int mlx5e_ptp_adjfreq(struct ptp_clock_info *ptp, s32 delta)
 	int neg_adj = 0;
 	struct mlx5e_tstamp *tstamp = container_of(ptp, struct mlx5e_tstamp,
 						  ptp_info);
-	struct mlx5e_priv *priv =
-		container_of(tstamp, struct mlx5e_priv, tstamp);
-
-	if (MLX5_CAP_GEN(priv->mdev, pps_modify)) {
-		u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
-
-		/* For future use need to add a loop for finding all 1PPS out pins */
-		MLX5_SET(mtpps_reg, in, pin_mode, MLX5E_PIN_MODE_OUT);
-		MLX5_SET(mtpps_reg, in, enhanced_out_periodic_adjustment, delta);
-		MLX5_SET(mtpps_reg, in, field_select,
-			 MLX5E_MTPPS_FS_PIN_MODE |
-			 MLX5E_MTPPS_FS_ENH_OUT_PER_ADJ);
-
-		mlx5_set_mtpps(priv->mdev, in, sizeof(in));
-	}
 
 	if (delta < 0) {
 		neg_adj = 1;
@@ -314,7 +326,7 @@ static int mlx5e_perout_configure(struct ptp_clock_info *ptp,
 	struct mlx5e_priv *priv =
 		container_of(tstamp, struct mlx5e_priv, tstamp);
 	u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
-	u64 nsec_now, nsec_delta, time_stamp;
+	u64 nsec_now, nsec_delta, time_stamp = 0;
 	u64 cycles_now, cycles_delta;
 	struct timespec64 ts;
 	unsigned long flags;
@@ -322,6 +334,7 @@ static int mlx5e_perout_configure(struct ptp_clock_info *ptp,
 	u8 pin_mode = 0;
 	u8 pattern = 0;
 	int pin = -1;
+	int err = 0;
 	s64 ns;
 
 	if (!MLX5_PPS_CAP(priv->mdev))
@@ -372,7 +385,12 @@ static int mlx5e_perout_configure(struct ptp_clock_info *ptp,
 	MLX5_SET64(mtpps_reg, in, time_stamp, time_stamp);
 	MLX5_SET(mtpps_reg, in, field_select, field_select);
 
-	return mlx5_set_mtpps(priv->mdev, in, sizeof(in));
+	err = mlx5_set_mtpps(priv->mdev, in, sizeof(in));
+	if (err)
+		return err;
+
+	return mlx5_set_mtppse(priv->mdev, pin, 0,
+			       MLX5E_EVENT_MODE_REPETETIVE & on);
 }
 
 static int mlx5e_ptp_enable(struct ptp_clock_info *ptp,
@@ -456,22 +474,50 @@ static void mlx5e_get_pps_caps(struct mlx5e_priv *priv,
 	tstamp->ptp_info.n_per_out = MLX5_GET(mtpps_reg, out,
 					      cap_max_num_of_pps_out_pins);
 
-	tstamp->pps_pin_caps[0] = MLX5_GET(mtpps_reg, out, cap_pin_0_mode);
-	tstamp->pps_pin_caps[1] = MLX5_GET(mtpps_reg, out, cap_pin_1_mode);
-	tstamp->pps_pin_caps[2] = MLX5_GET(mtpps_reg, out, cap_pin_2_mode);
-	tstamp->pps_pin_caps[3] = MLX5_GET(mtpps_reg, out, cap_pin_3_mode);
-	tstamp->pps_pin_caps[4] = MLX5_GET(mtpps_reg, out, cap_pin_4_mode);
-	tstamp->pps_pin_caps[5] = MLX5_GET(mtpps_reg, out, cap_pin_5_mode);
-	tstamp->pps_pin_caps[6] = MLX5_GET(mtpps_reg, out, cap_pin_6_mode);
-	tstamp->pps_pin_caps[7] = MLX5_GET(mtpps_reg, out, cap_pin_7_mode);
+	tstamp->pps_info.pin_caps[0] = MLX5_GET(mtpps_reg, out, cap_pin_0_mode);
+	tstamp->pps_info.pin_caps[1] = MLX5_GET(mtpps_reg, out, cap_pin_1_mode);
+	tstamp->pps_info.pin_caps[2] = MLX5_GET(mtpps_reg, out, cap_pin_2_mode);
+	tstamp->pps_info.pin_caps[3] = MLX5_GET(mtpps_reg, out, cap_pin_3_mode);
+	tstamp->pps_info.pin_caps[4] = MLX5_GET(mtpps_reg, out, cap_pin_4_mode);
+	tstamp->pps_info.pin_caps[5] = MLX5_GET(mtpps_reg, out, cap_pin_5_mode);
+	tstamp->pps_info.pin_caps[6] = MLX5_GET(mtpps_reg, out, cap_pin_6_mode);
+	tstamp->pps_info.pin_caps[7] = MLX5_GET(mtpps_reg, out, cap_pin_7_mode);
 }
 
 void mlx5e_pps_event_handler(struct mlx5e_priv *priv,
 			     struct ptp_clock_event *event)
 {
+	struct net_device *netdev = priv->netdev;
 	struct mlx5e_tstamp *tstamp = &priv->tstamp;
+	struct timespec64 ts;
+	u64 nsec_now, nsec_delta;
+	u64 cycles_now, cycles_delta;
+	int pin = event->index;
+	s64 ns;
+	unsigned long flags;
 
-	ptp_clock_event(tstamp->ptp, event);
+	switch (tstamp->ptp_info.pin_config[pin].func) {
+	case PTP_PF_EXTTS:
+		ptp_clock_event(tstamp->ptp, event);
+		break;
+	case PTP_PF_PEROUT:
+		mlx5e_ptp_gettime(&tstamp->ptp_info, &ts);
+		cycles_now = mlx5_read_internal_timer(tstamp->mdev);
+		ts.tv_sec += 1;
+		ts.tv_nsec = 0;
+		ns = timespec64_to_ns(&ts);
+		write_lock_irqsave(&tstamp->lock, flags);
+		nsec_now = timecounter_cyc2time(&tstamp->clock, cycles_now);
+		nsec_delta = ns - nsec_now;
+		cycles_delta = div64_u64(nsec_delta << tstamp->cycles.shift,
+					 tstamp->cycles.mult);
+		tstamp->pps_info.start[pin] = cycles_now + cycles_delta;
+		queue_work(priv->wq, &tstamp->pps_info.out_work);
+		write_unlock_irqrestore(&tstamp->lock, flags);
+		break;
+	default:
+		netdev_err(netdev, "%s: Unhandled event\n", __func__);
+	}
 }
 
 void mlx5e_timestamp_init(struct mlx5e_priv *priv)
@@ -507,6 +553,7 @@ void mlx5e_timestamp_init(struct mlx5e_priv *priv)
 	do_div(ns, NSEC_PER_SEC / 2 / HZ);
 	tstamp->overflow_period = ns;
 
+	INIT_WORK(&tstamp->pps_info.out_work, mlx5e_pps_out);
 	INIT_DELAYED_WORK(&tstamp->overflow_work, mlx5e_timestamp_overflow);
 	if (tstamp->overflow_period)
 		schedule_delayed_work(&tstamp->overflow_work, 0);
@@ -518,16 +565,10 @@ void mlx5e_timestamp_init(struct mlx5e_priv *priv)
 	snprintf(tstamp->ptp_info.name, 16, "mlx5 ptp");
 
 	/* Initialize 1PPS data structures */
-#define MAX_PIN_NUM	8
-	tstamp->pps_pin_caps = kzalloc(sizeof(u8) * MAX_PIN_NUM, GFP_KERNEL);
-	if (tstamp->pps_pin_caps) {
-		if (MLX5_PPS_CAP(priv->mdev))
-			mlx5e_get_pps_caps(priv, tstamp);
-		if (tstamp->ptp_info.n_pins)
-			mlx5e_init_pin_config(tstamp);
-	} else {
-		mlx5_core_warn(priv->mdev, "1PPS initialization failed\n");
-	}
+	if (MLX5_PPS_CAP(priv->mdev))
+		mlx5e_get_pps_caps(priv, tstamp);
+	if (tstamp->ptp_info.n_pins)
+		mlx5e_init_pin_config(tstamp);
 
 	tstamp->ptp = ptp_clock_register(&tstamp->ptp_info,
 					 &priv->mdev->pdev->dev);
@@ -550,7 +591,8 @@ void mlx5e_timestamp_cleanup(struct mlx5e_priv *priv)
 		priv->tstamp.ptp = NULL;
 	}
 
-	kfree(tstamp->pps_pin_caps);
+	cancel_work_sync(&tstamp->pps_info.out_work);
+
 	kfree(tstamp->ptp_info.pin_config);
 
 	cancel_delayed_work_sync(&tstamp->overflow_work);
-- 
2.13.0

[net 11/14] net/mlx5e: Add missing support for PTP_CLK_REQ_PPS request

[Saeed Mahameed]

From: Eugenia Emantayev <eugenia@mellanox.com>

Add the missing option to enable the PTP_CLK_PPS function.
In this case pin should be configured as 1PPS IN first and
then it will be connected to PPS mechanism.
Events will be reported as PTP_CLOCK_PPSUSR events to relevant sysfs.

Fixes: ee7f12205abc ('net/mlx5e: Implement 1PPS support')
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h       |  1 +
 drivers/net/ethernet/mellanox/mlx5/core/en_clock.c | 20 ++++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c  |  1 -
 3 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index d42826f0f0d6..0039b4725405 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -271,6 +271,7 @@ struct mlx5e_pps {
 	u8                         pin_caps[MAX_PIN_NUM];
 	struct work_struct         out_work;
 	u64                        start[MAX_PIN_NUM];
+	u8                         enabled;
 };
 
 struct mlx5e_tstamp {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
index 464ddd10ebbc..ba355c6b6e1f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
@@ -393,6 +393,17 @@ static int mlx5e_perout_configure(struct ptp_clock_info *ptp,
 			       MLX5E_EVENT_MODE_REPETETIVE & on);
 }
 
+static int mlx5e_pps_configure(struct ptp_clock_info *ptp,
+			       struct ptp_clock_request *rq,
+			       int on)
+{
+	struct mlx5e_tstamp *tstamp =
+		container_of(ptp, struct mlx5e_tstamp, ptp_info);
+
+	tstamp->pps_info.enabled = !!on;
+	return 0;
+}
+
 static int mlx5e_ptp_enable(struct ptp_clock_info *ptp,
 			    struct ptp_clock_request *rq,
 			    int on)
@@ -402,6 +413,8 @@ static int mlx5e_ptp_enable(struct ptp_clock_info *ptp,
 		return mlx5e_extts_configure(ptp, rq, on);
 	case PTP_CLK_REQ_PEROUT:
 		return mlx5e_perout_configure(ptp, rq, on);
+	case PTP_CLK_REQ_PPS:
+		return mlx5e_pps_configure(ptp, rq, on);
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -447,6 +460,7 @@ static int mlx5e_init_pin_config(struct mlx5e_tstamp *tstamp)
 		return -ENOMEM;
 	tstamp->ptp_info.enable = mlx5e_ptp_enable;
 	tstamp->ptp_info.verify = mlx5e_ptp_verify;
+	tstamp->ptp_info.pps = 1;
 
 	for (i = 0; i < tstamp->ptp_info.n_pins; i++) {
 		snprintf(tstamp->ptp_info.pin_config[i].name,
@@ -498,6 +512,12 @@ void mlx5e_pps_event_handler(struct mlx5e_priv *priv,
 
 	switch (tstamp->ptp_info.pin_config[pin].func) {
 	case PTP_PF_EXTTS:
+		if (tstamp->pps_info.enabled) {
+			event->type = PTP_CLOCK_PPSUSR;
+			event->pps_times.ts_real = ns_to_timespec64(event->timestamp);
+		} else {
+			event->type = PTP_CLOCK_EXTTS;
+		}
 		ptp_clock_event(tstamp->ptp, event);
 		break;
 	case PTP_PF_PEROUT:
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 1eac5003084f..57f31fa478ce 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -377,7 +377,6 @@ static void mlx5e_async_event(struct mlx5_core_dev *mdev, void *vpriv,
 		break;
 	case MLX5_DEV_EVENT_PPS:
 		eqe = (struct mlx5_eqe *)param;
-		ptp_event.type = PTP_CLOCK_EXTTS;
 		ptp_event.index = eqe->data.pps.pin;
 		ptp_event.timestamp =
 			timecounter_cyc2time(&priv->tstamp.clock,
-- 
2.13.0

[net 12/14] net/mlx5e: Fix wrong delay calculation for overflow check scheduling

[Saeed Mahameed]

From: Eugenia Emantayev <eugenia@mellanox.com>

The overflow_period is calculated in seconds. In order to use it
for delayed work scheduling translation to jiffies is needed.

Fixes: ef9814deafd0 ('net/mlx5e: Add HW timestamping (TS) support')
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_clock.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
index ba355c6b6e1f..cde8e8e772ec 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
@@ -119,7 +119,8 @@ static void mlx5e_timestamp_overflow(struct work_struct *work)
 	write_lock_irqsave(&tstamp->lock, flags);
 	timecounter_read(&tstamp->clock);
 	write_unlock_irqrestore(&tstamp->lock, flags);
-	schedule_delayed_work(&tstamp->overflow_work, tstamp->overflow_period);
+	schedule_delayed_work(&tstamp->overflow_work,
+			      msecs_to_jiffies(tstamp->overflow_period * 1000));
 }
 
 int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
-- 
2.13.0

[net 13/14] net/mlx5e: Schedule overflow check work to mlx5e workqueue

[Saeed Mahameed]

From: Eugenia Emantayev <eugenia@mellanox.com>

This is done in order to ensure that work will not run after the cleanup.

Fixes: ef9814deafd0 ('net/mlx5e: Add HW timestamping (TS) support')
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_clock.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
index cde8e8e772ec..84dd63e74041 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c
@@ -114,13 +114,14 @@ static void mlx5e_timestamp_overflow(struct work_struct *work)
 	struct delayed_work *dwork = to_delayed_work(work);
 	struct mlx5e_tstamp *tstamp = container_of(dwork, struct mlx5e_tstamp,
 						   overflow_work);
+	struct mlx5e_priv *priv = container_of(tstamp, struct mlx5e_priv, tstamp);
 	unsigned long flags;
 
 	write_lock_irqsave(&tstamp->lock, flags);
 	timecounter_read(&tstamp->clock);
 	write_unlock_irqrestore(&tstamp->lock, flags);
-	schedule_delayed_work(&tstamp->overflow_work,
-			      msecs_to_jiffies(tstamp->overflow_period * 1000));
+	queue_delayed_work(priv->wq, &tstamp->overflow_work,
+			   msecs_to_jiffies(tstamp->overflow_period * 1000));
 }
 
 int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
@@ -577,7 +578,7 @@ void mlx5e_timestamp_init(struct mlx5e_priv *priv)
 	INIT_WORK(&tstamp->pps_info.out_work, mlx5e_pps_out);
 	INIT_DELAYED_WORK(&tstamp->overflow_work, mlx5e_timestamp_overflow);
 	if (tstamp->overflow_period)
-		schedule_delayed_work(&tstamp->overflow_work, 0);
+		queue_delayed_work(priv->wq, &tstamp->overflow_work, 0);
 	else
 		mlx5_core_warn(priv->mdev, "invalid overflow period, overflow_work is not scheduled\n");
 
@@ -613,8 +614,6 @@ void mlx5e_timestamp_cleanup(struct mlx5e_priv *priv)
 	}
 
 	cancel_work_sync(&tstamp->pps_info.out_work);
-
-	kfree(tstamp->ptp_info.pin_config);
-
 	cancel_delayed_work_sync(&tstamp->overflow_work);
+	kfree(tstamp->ptp_info.pin_config);
 }
-- 
2.13.0

[net 14/14] net/mlx5: Fix mlx5_add_flow_rules call with correct num of dests

[Saeed Mahameed]

From: Paul Blakey <paulb@mellanox.com>

When adding ethtool steering rule with action DISCARD we wrongly
pass a NULL dest with dest_num 1 to mlx5_add_flow_rules().
What this error seems to have caused is sending VPORT 0
(MLX5_FLOW_DESTINATION_TYPE_VPORT) as the fte dest instead of no dests.
We have fte action correctly set to DROP so it might been ignored
anyways.

To reproduce use:
 # sudo ethtool --config-nfc <dev> flow-type ether \
   dst aa:bb:cc:dd:ee:ff action -1

Fixes: 74491de93712 ("net/mlx5: Add multi dest support")
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
index 22fb993987b4..eafc59280ada 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
@@ -320,7 +320,7 @@ add_ethtool_flow_rule(struct mlx5e_priv *priv,
 
 	spec->match_criteria_enable = (!outer_header_zero(spec->match_criteria));
 	flow_act.flow_tag = MLX5_FS_DEFAULT_FLOW_TAG;
-	rule = mlx5_add_flow_rules(ft, spec, &flow_act, dst, 1);
+	rule = mlx5_add_flow_rules(ft, spec, &flow_act, dst, dst ? 1 : 0);
 	if (IS_ERR(rule)) {
 		err = PTR_ERR(rule);
 		netdev_err(priv->netdev, "%s: failed to add ethtool steering rule: %d\n",
-- 
2.13.0