[PATCH mlx5-next v1 00/11] Convert mlx5 to use auxiliary bus

[PATCH mlx5-next v1 00/11] Convert mlx5 to use auxiliary bus

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Changelog:
v1:
 * Renamed _mlx5_rescan_driver to be mlx5_rescan_driver_locked like in
   other parts of the mlx5 driver.
 * Renamed MLX5_INTERFACE_PROTOCOL_VDPA to tbe MLX5_INTERFACE_PROTOCOL_VNET as
   a preparation to coming series from Eli C.
 * Some small naming renames in mlx5_vdpa.
 * Refactored adev index code to make Parav's SF series to apply more easily.
 * Fixed devlink reload bug that caused to lost TCP connection.
v0:
https://lore.kernel.org/lkml/20201026111849.1035786-1-leon@kernel.org/

--------------------------------------------------------------

Hi,

This patch set converts mlx5 driver to use auxiliary bus [1].

In this series, we are connecting three subsystems (VDPA, netdev and
RDMA) through mlx5_core PCI driver. That driver is responsible to create
proper devices based on supported firmware.

First four patches are preparitions and fixes that were spotted during
code development, rest is the conversion itself.

Thanks

[1]
https://lore.kernel.org/lkml/20201023003338.1285642-1-david.m.ertman@intel.com

Leon Romanovsky (11):
  net/mlx5: Don't skip vport check
  net/mlx5: Properly convey driver version to firmware
  net/mlx5_core: Clean driver version and name
  vdpa/mlx5: Make hardware definitions visible to all mlx5 devices
  net/mlx5: Register mlx5 devices to auxiliary virtual bus
  vdpa/mlx5: Connect mlx5_vdpa to auxiliary bus
  net/mlx5e: Connect ethernet part to auxiliary bus
  RDMA/mlx5: Convert mlx5_ib to use auxiliary bus
  net/mlx5: Delete custom device management logic
  net/mlx5: Simplify eswitch mode check
  RDMA/mlx5: Remove IB representors dead code

 drivers/infiniband/hw/mlx5/counters.c         |   7 -
 drivers/infiniband/hw/mlx5/ib_rep.c           | 113 ++--
 drivers/infiniband/hw/mlx5/ib_rep.h           |  45 +-
 drivers/infiniband/hw/mlx5/main.c             | 148 +++--
 drivers/infiniband/hw/mlx5/mlx5_ib.h          |   4 +-
 .../net/ethernet/mellanox/mlx5/core/Kconfig   |   1 +
 drivers/net/ethernet/mellanox/mlx5/core/dev.c | 567 ++++++++++++------
 .../net/ethernet/mellanox/mlx5/core/devlink.c |   4 +-
 .../ethernet/mellanox/mlx5/core/en_ethtool.c  |   4 +-
 .../net/ethernet/mellanox/mlx5/core/en_main.c | 135 ++---
 .../net/ethernet/mellanox/mlx5/core/en_rep.c  |  42 +-
 .../net/ethernet/mellanox/mlx5/core/en_rep.h  |   6 +-
 .../net/ethernet/mellanox/mlx5/core/en_tc.c   |   8 +-
 .../mellanox/mlx5/core/esw/devlink_port.c     |   2 +-
 .../net/ethernet/mellanox/mlx5/core/eswitch.c |  28 +-
 .../mellanox/mlx5/core/eswitch_offloads.c     |   6 +
 .../mellanox/mlx5/core/ipoib/ethtool.c        |   2 +-
 drivers/net/ethernet/mellanox/mlx5/core/lag.c |  58 +-
 .../net/ethernet/mellanox/mlx5/core/main.c    |  50 +-
 .../ethernet/mellanox/mlx5/core/mlx5_core.h   |  33 +-
 drivers/vdpa/mlx5/Makefile                    |   2 +-
 drivers/vdpa/mlx5/net/main.c                  |  76 ---
 drivers/vdpa/mlx5/net/mlx5_vnet.c             |  53 +-
 drivers/vdpa/mlx5/net/mlx5_vnet.h             |  24 -
 include/linux/mlx5/driver.h                   |  34 +-
 include/linux/mlx5/eswitch.h                  |   8 +-
 .../linux/mlx5/mlx5_ifc_vdpa.h                |   6 +-
 27 files changed, 818 insertions(+), 648 deletions(-)
 delete mode 100644 drivers/vdpa/mlx5/net/main.c
 delete mode 100644 drivers/vdpa/mlx5/net/mlx5_vnet.h
 rename drivers/vdpa/mlx5/core/mlx5_vdpa_ifc.h => include/linux/mlx5/mlx5_ifc_vdpa.h (97%)

--
2.28.0

[PATCH mlx5-next v1 01/11] net/mlx5: Don't skip vport check

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Users of mlx5_eswitch_get_vport() are required to check return value
prior to passing mlx5_vport further. Fix all the places to do not skip
that check.

Fixes: c7eddc6092b4 ("net/mlx5: E-switch, Move devlink eswitch ports closer to eswitch")
Fixes: 5d9986a3947a ("net/mlx5: E-Switch, Fix the check of legal vport")
Reviewed-by: Eli Cohen <elic@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/esw/devlink_port.c | 2 +-
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c          | 7 +++++++
 drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 6 ++++++
 3 files changed, 14 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/devlink_port.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/devlink_port.c
index ffff11baa3d0..88688b84513b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/esw/devlink_port.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/devlink_port.c
@@ -120,5 +120,5 @@ struct devlink_port *mlx5_esw_offloads_devlink_port(struct mlx5_eswitch *esw, u1
 	struct mlx5_vport *vport;

 	vport = mlx5_eswitch_get_vport(esw, vport_num);
-	return vport->dl_port;
+	return IS_ERR(vport) ? ERR_CAST(vport) : vport->dl_port;
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 6e6a9a563992..2e14bf238588 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1142,6 +1142,9 @@ int mlx5_esw_modify_vport_rate(struct mlx5_eswitch *esw, u16 vport_num,
 	struct mlx5_vport *vport;

 	vport = mlx5_eswitch_get_vport(esw, vport_num);
+	if (IS_ERR(vport))
+		return PTR_ERR(vport);
+
 	MLX5_SET(scheduling_context, ctx, max_average_bw, rate_mbps);

 	return mlx5_modify_scheduling_element_cmd(esw->dev,
@@ -1276,6 +1279,8 @@ static int esw_enable_vport(struct mlx5_eswitch *esw, u16 vport_num,
 	int ret;

 	vport = mlx5_eswitch_get_vport(esw, vport_num);
+	if (IS_ERR(vport))
+		return PTR_ERR(vport);

 	mutex_lock(&esw->state_lock);
 	WARN_ON(vport->enabled);
@@ -1311,6 +1316,8 @@ static void esw_disable_vport(struct mlx5_eswitch *esw, u16 vport_num)
 	struct mlx5_vport *vport;

 	vport = mlx5_eswitch_get_vport(esw, vport_num);
+	if (IS_ERR(vport))
+		return;

 	mutex_lock(&esw->state_lock);
 	if (!vport->enabled)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index c9c2962ad49f..429dc613530b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -2159,6 +2159,9 @@ static int esw_create_uplink_offloads_acl_tables(struct mlx5_eswitch *esw)
 	struct mlx5_vport *vport;

 	vport = mlx5_eswitch_get_vport(esw, MLX5_VPORT_UPLINK);
+	if (IS_ERR(vport))
+		return PTR_ERR(vport);
+
 	return esw_vport_create_offloads_acl_tables(esw, vport);
 }

@@ -2167,6 +2170,9 @@ static void esw_destroy_uplink_offloads_acl_tables(struct mlx5_eswitch *esw)
 	struct mlx5_vport *vport;

 	vport = mlx5_eswitch_get_vport(esw, MLX5_VPORT_UPLINK);
+	if (IS_ERR(vport))
+		return;
+
 	esw_vport_destroy_offloads_acl_tables(esw, vport);
 }

--
2.28.0

[PATCH mlx5-next v1 02/11] net/mlx5: Properly convey driver version to firmware

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

mlx5 firmware expects driver version in specific format X.X.X, so
make it always correct and based on real kernel version aligned with
the driver.

Fixes: 012e50e109fd ("net/mlx5: Set driver version into firmware")
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/main.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 8ff207aa1479..71e210f22f69 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -50,6 +50,7 @@
 #ifdef CONFIG_RFS_ACCEL
 #include <linux/cpu_rmap.h>
 #endif
+#include <linux/version.h>
 #include <net/devlink.h>
 #include "mlx5_core.h"
 #include "lib/eq.h"
@@ -233,7 +234,10 @@ static void mlx5_set_driver_version(struct mlx5_core_dev *dev)
 	strncat(string, ",", remaining_size);

 	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
-	strncat(string, DRIVER_VERSION, remaining_size);
+
+	snprintf(string + strlen(string), remaining_size, "%u.%u.%u",
+		 (u8)(LINUX_VERSION_CODE >> 16), (u8)(LINUX_VERSION_CODE >> 8),
+		 (u16)(LINUX_VERSION_CODE & 0xff));

 	/*Send the command*/
 	MLX5_SET(set_driver_version_in, in, opcode,
--
2.28.0

Re: [PATCH mlx5-next v1 02/11] net/mlx5: Properly convey driver version to firmware

[Roi Dayan]

On 2020-11-01 10:15 PM, Leon Romanovsky wrote:
> From: Leon Romanovsky <leonro@nvidia.com>
> 
> mlx5 firmware expects driver version in specific format X.X.X, so
> make it always correct and based on real kernel version aligned with
> the driver.
> 
> Fixes: 012e50e109fd ("net/mlx5: Set driver version into firmware")
> Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
> ---
>   drivers/net/ethernet/mellanox/mlx5/core/main.c | 6 +++++-
>   1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
> index 8ff207aa1479..71e210f22f69 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
> @@ -50,6 +50,7 @@
>   #ifdef CONFIG_RFS_ACCEL
>   #include <linux/cpu_rmap.h>
>   #endif
> +#include <linux/version.h>
>   #include <net/devlink.h>
>   #include "mlx5_core.h"
>   #include "lib/eq.h"
> @@ -233,7 +234,10 @@ static void mlx5_set_driver_version(struct mlx5_core_dev *dev)
>   	strncat(string, ",", remaining_size);
> 
>   	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
> -	strncat(string, DRIVER_VERSION, remaining_size);
> +
> +	snprintf(string + strlen(string), remaining_size, "%u.%u.%u",
> +		 (u8)(LINUX_VERSION_CODE >> 16), (u8)(LINUX_VERSION_CODE >> 8),
> +		 (u16)(LINUX_VERSION_CODE & 0xff));
> 
>   	/*Send the command*/
>   	MLX5_SET(set_driver_version_in, in, opcode,
> --
> 2.28.0
> 

Reviewed-by: Roi Dayan <roid@nvidia.com>

[PATCH mlx5-next v1 03/11] net/mlx5_core: Clean driver version and name

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Remove exposed driver version as it was done in other drivers,
so module version will work correctly by displaying the kernel
version for which it is compiled.

And move mlx5_core module name to general include, so auxiliary drivers
will be able to use it as a basis for a name in their device ID tables.

Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/devlink.c     |  2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c  |  4 +---
 drivers/net/ethernet/mellanox/mlx5/core/en_rep.c      |  1 -
 .../net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c   |  2 +-
 drivers/net/ethernet/mellanox/mlx5/core/main.c        | 11 +++++++----
 drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h   |  3 ---
 include/linux/mlx5/driver.h                           |  2 ++
 7 files changed, 12 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
index a28f95df2901..1a351e2f6ace 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
@@ -52,7 +52,7 @@ mlx5_devlink_info_get(struct devlink *devlink, struct devlink_info_req *req,
 	u32 running_fw, stored_fw;
 	int err;

-	err = devlink_info_driver_name_put(req, DRIVER_NAME);
+	err = devlink_info_driver_name_put(req, KBUILD_MODNAME);
 	if (err)
 		return err;

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index d25a56ec6876..bcff18a87bcd 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -40,9 +40,7 @@ void mlx5e_ethtool_get_drvinfo(struct mlx5e_priv *priv,
 {
 	struct mlx5_core_dev *mdev = priv->mdev;

-	strlcpy(drvinfo->driver, DRIVER_NAME, sizeof(drvinfo->driver));
-	strlcpy(drvinfo->version, DRIVER_VERSION,
-		sizeof(drvinfo->version));
+	strlcpy(drvinfo->driver, KBUILD_MODNAME, sizeof(drvinfo->driver));
 	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
 		 "%d.%d.%04d (%.16s)",
 		 fw_rev_maj(mdev), fw_rev_min(mdev), fw_rev_sub(mdev),
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 67247c33b9fd..ef2f8889ba0f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -64,7 +64,6 @@ static void mlx5e_rep_get_drvinfo(struct net_device *dev,

 	strlcpy(drvinfo->driver, mlx5e_rep_driver_name,
 		sizeof(drvinfo->driver));
-	strlcpy(drvinfo->version, UTS_RELEASE, sizeof(drvinfo->version));
 	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
 		 "%d.%d.%04d (%.16s)",
 		 fw_rev_maj(mdev), fw_rev_min(mdev),
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
index cac8f085b16d..97d96fc38a65 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
@@ -39,7 +39,7 @@ static void mlx5i_get_drvinfo(struct net_device *dev,
 	struct mlx5e_priv *priv = mlx5i_epriv(dev);

 	mlx5e_ethtool_get_drvinfo(priv, drvinfo);
-	strlcpy(drvinfo->driver, DRIVER_NAME "[ib_ipoib]",
+	strlcpy(drvinfo->driver, KBUILD_MODNAME "[ib_ipoib]",
 		sizeof(drvinfo->driver));
 }

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 71e210f22f69..9827127cb674 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -77,7 +77,6 @@
 MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
 MODULE_DESCRIPTION("Mellanox 5th generation network adapters (ConnectX series) core driver");
 MODULE_LICENSE("Dual BSD/GPL");
-MODULE_VERSION(DRIVER_VERSION);

 unsigned int mlx5_core_debug_mask;
 module_param_named(debug_mask, mlx5_core_debug_mask, uint, 0644);
@@ -228,7 +227,7 @@ static void mlx5_set_driver_version(struct mlx5_core_dev *dev)
 	strncat(string, ",", remaining_size);

 	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
-	strncat(string, DRIVER_NAME, remaining_size);
+	strncat(string, KBUILD_MODNAME, remaining_size);

 	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
 	strncat(string, ",", remaining_size);
@@ -313,7 +312,7 @@ static int request_bar(struct pci_dev *pdev)
 		return -ENODEV;
 	}

-	err = pci_request_regions(pdev, DRIVER_NAME);
+	err = pci_request_regions(pdev, KBUILD_MODNAME);
 	if (err)
 		dev_err(&pdev->dev, "Couldn't get PCI resources, aborting\n");

@@ -1617,7 +1616,7 @@ void mlx5_recover_device(struct mlx5_core_dev *dev)
 }

 static struct pci_driver mlx5_core_driver = {
-	.name           = DRIVER_NAME,
+	.name           = KBUILD_MODNAME,
 	.id_table       = mlx5_core_pci_table,
 	.probe          = init_one,
 	.remove         = remove_one,
@@ -1643,6 +1642,10 @@ static int __init init(void)
 {
 	int err;

+	WARN_ONCE(strcmp(MLX5_ADEV_NAME, KBUILD_MODNAME) ||
+		  strlen(MLX5_ADEV_NAME) != strlen(KBUILD_MODNAME),
+		  "mlx5_core name not in sync with kernel module name");
+
 	get_random_bytes(&sw_owner_id, sizeof(sw_owner_id));

 	mlx5_core_verify_params();
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index 8cec85ab419d..b285f1515e4e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -42,9 +42,6 @@
 #include <linux/mlx5/fs.h>
 #include <linux/mlx5/driver.h>

-#define DRIVER_NAME "mlx5_core"
-#define DRIVER_VERSION "5.0-0"
-
 extern uint mlx5_core_debug_mask;

 #define mlx5_core_dbg(__dev, format, ...)				\
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 317257f8e0ad..ed1d030658d2 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -56,6 +56,8 @@
 #include <linux/ptp_clock_kernel.h>
 #include <net/devlink.h>

+#define MLX5_ADEV_NAME "mlx5_core"
+
 enum {
 	MLX5_BOARD_ID_LEN = 64,
 };
--
2.28.0

RE: [PATCH mlx5-next v1 03/11] net/mlx5_core: Clean driver version and name

[Parav Pandit]

> From: Leon Romanovsky <leon@kernel.org>
> Sent: Monday, November 2, 2020 1:46 AM
> 
> From: Leon Romanovsky <leonro@nvidia.com>
> 
> Remove exposed driver version as it was done in other drivers, so module
> version will work correctly by displaying the kernel version for which it is
> compiled.
> 
> And move mlx5_core module name to general include, so auxiliary drivers
> will be able to use it as a basis for a name in their device ID tables.
> 
> Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
> ---
>  drivers/net/ethernet/mellanox/mlx5/core/devlink.c     |  2 +-
>  drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c  |  4 +---
>  drivers/net/ethernet/mellanox/mlx5/core/en_rep.c      |  1 -
>  .../net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c   |  2 +-
>  drivers/net/ethernet/mellanox/mlx5/core/main.c        | 11 +++++++----
>  drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h   |  3 ---
>  include/linux/mlx5/driver.h                           |  2 ++
>  7 files changed, 12 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> index a28f95df2901..1a351e2f6ace 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> @@ -52,7 +52,7 @@ mlx5_devlink_info_get(struct devlink *devlink, struct
> devlink_info_req *req,
>  	u32 running_fw, stored_fw;
>  	int err;
> 
> -	err = devlink_info_driver_name_put(req, DRIVER_NAME);
> +	err = devlink_info_driver_name_put(req, KBUILD_MODNAME);
>  	if (err)
>  		return err;
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
> b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
> index d25a56ec6876..bcff18a87bcd 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
> @@ -40,9 +40,7 @@ void mlx5e_ethtool_get_drvinfo(struct mlx5e_priv
> *priv,  {
>  	struct mlx5_core_dev *mdev = priv->mdev;
> 
> -	strlcpy(drvinfo->driver, DRIVER_NAME, sizeof(drvinfo->driver));
> -	strlcpy(drvinfo->version, DRIVER_VERSION,
> -		sizeof(drvinfo->version));
> +	strlcpy(drvinfo->driver, KBUILD_MODNAME, sizeof(drvinfo-
> >driver));
>  	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
>  		 "%d.%d.%04d (%.16s)",
>  		 fw_rev_maj(mdev), fw_rev_min(mdev),
> fw_rev_sub(mdev), diff --git
> a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> index 67247c33b9fd..ef2f8889ba0f 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> @@ -64,7 +64,6 @@ static void mlx5e_rep_get_drvinfo(struct net_device
> *dev,
> 
>  	strlcpy(drvinfo->driver, mlx5e_rep_driver_name,
>  		sizeof(drvinfo->driver));
> -	strlcpy(drvinfo->version, UTS_RELEASE, sizeof(drvinfo->version));
>  	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
>  		 "%d.%d.%04d (%.16s)",
>  		 fw_rev_maj(mdev), fw_rev_min(mdev),
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
> b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
> index cac8f085b16d..97d96fc38a65 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
> @@ -39,7 +39,7 @@ static void mlx5i_get_drvinfo(struct net_device *dev,
>  	struct mlx5e_priv *priv = mlx5i_epriv(dev);
> 
>  	mlx5e_ethtool_get_drvinfo(priv, drvinfo);
> -	strlcpy(drvinfo->driver, DRIVER_NAME "[ib_ipoib]",
> +	strlcpy(drvinfo->driver, KBUILD_MODNAME "[ib_ipoib]",
>  		sizeof(drvinfo->driver));
>  }
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c
> b/drivers/net/ethernet/mellanox/mlx5/core/main.c
> index 71e210f22f69..9827127cb674 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
> @@ -77,7 +77,6 @@
>  MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
> MODULE_DESCRIPTION("Mellanox 5th generation network adapters
> (ConnectX series) core driver");  MODULE_LICENSE("Dual BSD/GPL"); -
> MODULE_VERSION(DRIVER_VERSION);
> 
>  unsigned int mlx5_core_debug_mask;
>  module_param_named(debug_mask, mlx5_core_debug_mask, uint, 0644);
> @@ -228,7 +227,7 @@ static void mlx5_set_driver_version(struct
> mlx5_core_dev *dev)
>  	strncat(string, ",", remaining_size);
> 
>  	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
> -	strncat(string, DRIVER_NAME, remaining_size);
> +	strncat(string, KBUILD_MODNAME, remaining_size);
> 
>  	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
>  	strncat(string, ",", remaining_size);
> @@ -313,7 +312,7 @@ static int request_bar(struct pci_dev *pdev)
>  		return -ENODEV;
>  	}
> 
> -	err = pci_request_regions(pdev, DRIVER_NAME);
> +	err = pci_request_regions(pdev, KBUILD_MODNAME);
>  	if (err)
>  		dev_err(&pdev->dev, "Couldn't get PCI resources,
> aborting\n");
> 
> @@ -1617,7 +1616,7 @@ void mlx5_recover_device(struct mlx5_core_dev
> *dev)  }
> 
>  static struct pci_driver mlx5_core_driver = {
> -	.name           = DRIVER_NAME,
> +	.name           = KBUILD_MODNAME,
>  	.id_table       = mlx5_core_pci_table,
>  	.probe          = init_one,
>  	.remove         = remove_one,
> @@ -1643,6 +1642,10 @@ static int __init init(void)  {
>  	int err;
> 
> +	WARN_ONCE(strcmp(MLX5_ADEV_NAME, KBUILD_MODNAME) ||
> +		  strlen(MLX5_ADEV_NAME) != strlen(KBUILD_MODNAME),
> +		  "mlx5_core name not in sync with kernel module name");
> +
In which case, both the strings are same but their length not?
You likely don't need the string length check.

>  	get_random_bytes(&sw_owner_id, sizeof(sw_owner_id));
> 
>  	mlx5_core_verify_params();
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
> b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
> index 8cec85ab419d..b285f1515e4e 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
> @@ -42,9 +42,6 @@
>  #include <linux/mlx5/fs.h>
>  #include <linux/mlx5/driver.h>
> 
> -#define DRIVER_NAME "mlx5_core"
> -#define DRIVER_VERSION "5.0-0"
> -
>  extern uint mlx5_core_debug_mask;
> 
>  #define mlx5_core_dbg(__dev, format, ...)				\
> diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index
> 317257f8e0ad..ed1d030658d2 100644
> --- a/include/linux/mlx5/driver.h
> +++ b/include/linux/mlx5/driver.h
> @@ -56,6 +56,8 @@
>  #include <linux/ptp_clock_kernel.h>
>  #include <net/devlink.h>
> 
> +#define MLX5_ADEV_NAME "mlx5_core"
> +
>  enum {
>  	MLX5_BOARD_ID_LEN = 64,
>  };
> --
> 2.28.0


Other than strlen removal check,
Reviewed-by: Parav Pandit <parav@nvidia.com>

Re: [PATCH mlx5-next v1 03/11] net/mlx5_core: Clean driver version and name

[Leon Romanovsky]

On Mon, Nov 02, 2020 at 05:07:59AM +0000, Parav Pandit wrote:
>
>
> > From: Leon Romanovsky <leon@kernel.org>
> > Sent: Monday, November 2, 2020 1:46 AM
> >
> > From: Leon Romanovsky <leonro@nvidia.com>
> >
> > Remove exposed driver version as it was done in other drivers, so module
> > version will work correctly by displaying the kernel version for which it is
> > compiled.
> >
> > And move mlx5_core module name to general include, so auxiliary drivers
> > will be able to use it as a basis for a name in their device ID tables.
> >
> > Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
> > ---
> >  drivers/net/ethernet/mellanox/mlx5/core/devlink.c     |  2 +-
> >  drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c  |  4 +---
> >  drivers/net/ethernet/mellanox/mlx5/core/en_rep.c      |  1 -
> >  .../net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c   |  2 +-
> >  drivers/net/ethernet/mellanox/mlx5/core/main.c        | 11 +++++++----
> >  drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h   |  3 ---
> >  include/linux/mlx5/driver.h                           |  2 ++
> >  7 files changed, 12 insertions(+), 13 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> > b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> > index a28f95df2901..1a351e2f6ace 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> > @@ -52,7 +52,7 @@ mlx5_devlink_info_get(struct devlink *devlink, struct
> > devlink_info_req *req,
> >  	u32 running_fw, stored_fw;
> >  	int err;
> >
> > -	err = devlink_info_driver_name_put(req, DRIVER_NAME);
> > +	err = devlink_info_driver_name_put(req, KBUILD_MODNAME);
> >  	if (err)
> >  		return err;
> >
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
> > b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
> > index d25a56ec6876..bcff18a87bcd 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
> > @@ -40,9 +40,7 @@ void mlx5e_ethtool_get_drvinfo(struct mlx5e_priv
> > *priv,  {
> >  	struct mlx5_core_dev *mdev = priv->mdev;
> >
> > -	strlcpy(drvinfo->driver, DRIVER_NAME, sizeof(drvinfo->driver));
> > -	strlcpy(drvinfo->version, DRIVER_VERSION,
> > -		sizeof(drvinfo->version));
> > +	strlcpy(drvinfo->driver, KBUILD_MODNAME, sizeof(drvinfo-
> > >driver));
> >  	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
> >  		 "%d.%d.%04d (%.16s)",
> >  		 fw_rev_maj(mdev), fw_rev_min(mdev),
> > fw_rev_sub(mdev), diff --git
> > a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> > b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> > index 67247c33b9fd..ef2f8889ba0f 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> > @@ -64,7 +64,6 @@ static void mlx5e_rep_get_drvinfo(struct net_device
> > *dev,
> >
> >  	strlcpy(drvinfo->driver, mlx5e_rep_driver_name,
> >  		sizeof(drvinfo->driver));
> > -	strlcpy(drvinfo->version, UTS_RELEASE, sizeof(drvinfo->version));
> >  	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
> >  		 "%d.%d.%04d (%.16s)",
> >  		 fw_rev_maj(mdev), fw_rev_min(mdev),
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
> > b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
> > index cac8f085b16d..97d96fc38a65 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
> > @@ -39,7 +39,7 @@ static void mlx5i_get_drvinfo(struct net_device *dev,
> >  	struct mlx5e_priv *priv = mlx5i_epriv(dev);
> >
> >  	mlx5e_ethtool_get_drvinfo(priv, drvinfo);
> > -	strlcpy(drvinfo->driver, DRIVER_NAME "[ib_ipoib]",
> > +	strlcpy(drvinfo->driver, KBUILD_MODNAME "[ib_ipoib]",
> >  		sizeof(drvinfo->driver));
> >  }
> >
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c
> > b/drivers/net/ethernet/mellanox/mlx5/core/main.c
> > index 71e210f22f69..9827127cb674 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
> > @@ -77,7 +77,6 @@
> >  MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
> > MODULE_DESCRIPTION("Mellanox 5th generation network adapters
> > (ConnectX series) core driver");  MODULE_LICENSE("Dual BSD/GPL"); -
> > MODULE_VERSION(DRIVER_VERSION);
> >
> >  unsigned int mlx5_core_debug_mask;
> >  module_param_named(debug_mask, mlx5_core_debug_mask, uint, 0644);
> > @@ -228,7 +227,7 @@ static void mlx5_set_driver_version(struct
> > mlx5_core_dev *dev)
> >  	strncat(string, ",", remaining_size);
> >
> >  	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
> > -	strncat(string, DRIVER_NAME, remaining_size);
> > +	strncat(string, KBUILD_MODNAME, remaining_size);
> >
> >  	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
> >  	strncat(string, ",", remaining_size);
> > @@ -313,7 +312,7 @@ static int request_bar(struct pci_dev *pdev)
> >  		return -ENODEV;
> >  	}
> >
> > -	err = pci_request_regions(pdev, DRIVER_NAME);
> > +	err = pci_request_regions(pdev, KBUILD_MODNAME);
> >  	if (err)
> >  		dev_err(&pdev->dev, "Couldn't get PCI resources,
> > aborting\n");
> >
> > @@ -1617,7 +1616,7 @@ void mlx5_recover_device(struct mlx5_core_dev
> > *dev)  }
> >
> >  static struct pci_driver mlx5_core_driver = {
> > -	.name           = DRIVER_NAME,
> > +	.name           = KBUILD_MODNAME,
> >  	.id_table       = mlx5_core_pci_table,
> >  	.probe          = init_one,
> >  	.remove         = remove_one,
> > @@ -1643,6 +1642,10 @@ static int __init init(void)  {
> >  	int err;
> >
> > +	WARN_ONCE(strcmp(MLX5_ADEV_NAME, KBUILD_MODNAME) ||
> > +		  strlen(MLX5_ADEV_NAME) != strlen(KBUILD_MODNAME),
> > +		  "mlx5_core name not in sync with kernel module name");
> > +
> In which case, both the strings are same but their length not?
> You likely don't need the string length check.

Yes, I was overzealous, I'll remove when will apply the series.

>
> >  	get_random_bytes(&sw_owner_id, sizeof(sw_owner_id));
> >
> >  	mlx5_core_verify_params();
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
> > b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
> > index 8cec85ab419d..b285f1515e4e 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
> > @@ -42,9 +42,6 @@
> >  #include <linux/mlx5/fs.h>
> >  #include <linux/mlx5/driver.h>
> >
> > -#define DRIVER_NAME "mlx5_core"
> > -#define DRIVER_VERSION "5.0-0"
> > -
> >  extern uint mlx5_core_debug_mask;
> >
> >  #define mlx5_core_dbg(__dev, format, ...)				\
> > diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index
> > 317257f8e0ad..ed1d030658d2 100644
> > --- a/include/linux/mlx5/driver.h
> > +++ b/include/linux/mlx5/driver.h
> > @@ -56,6 +56,8 @@
> >  #include <linux/ptp_clock_kernel.h>
> >  #include <net/devlink.h>
> >
> > +#define MLX5_ADEV_NAME "mlx5_core"
> > +
> >  enum {
> >  	MLX5_BOARD_ID_LEN = 64,
> >  };
> > --
> > 2.28.0
>
>
> Other than strlen removal check,
> Reviewed-by: Parav Pandit <parav@nvidia.com>
>

Thanks

[PATCH mlx5-next v1 04/11] vdpa/mlx5: Make hardware definitions visible to all mlx5 devices

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Move mlx5_vdpa IFC header file to the general include folder, so
mlx5_core will be able to reuse it to check if VDPA is supported
prior to creating an auxiliary device.

As part of this move, update the header file name to mlx5 general
naming scheme.

Reviewed-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/vdpa/mlx5/net/main.c                                | 2 +-
 drivers/vdpa/mlx5/net/mlx5_vnet.c                           | 2 +-
 .../mlx5_vdpa_ifc.h => include/linux/mlx5/mlx5_ifc_vdpa.h   | 6 +++---
 3 files changed, 5 insertions(+), 5 deletions(-)
 rename drivers/vdpa/mlx5/core/mlx5_vdpa_ifc.h => include/linux/mlx5/mlx5_ifc_vdpa.h (97%)

diff --git a/drivers/vdpa/mlx5/net/main.c b/drivers/vdpa/mlx5/net/main.c
index 838cd98386ff..4dd3f00f2306 100644
--- a/drivers/vdpa/mlx5/net/main.c
+++ b/drivers/vdpa/mlx5/net/main.c
@@ -4,7 +4,7 @@
 #include <linux/module.h>
 #include <linux/mlx5/driver.h>
 #include <linux/mlx5/device.h>
-#include "mlx5_vdpa_ifc.h"
+#include <linux/mlx5/mlx5_ifc_vdpa.h>
 #include "mlx5_vnet.h"

 MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
index 1fa6fcac8299..6c218b47b9f1 100644
--- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
+++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
@@ -9,8 +9,8 @@
 #include <linux/mlx5/vport.h>
 #include <linux/mlx5/fs.h>
 #include <linux/mlx5/device.h>
+#include <linux/mlx5/mlx5_ifc_vdpa.h>
 #include "mlx5_vnet.h"
-#include "mlx5_vdpa_ifc.h"
 #include "mlx5_vdpa.h"

 #define to_mvdev(__vdev) container_of((__vdev), struct mlx5_vdpa_dev, vdev)
diff --git a/drivers/vdpa/mlx5/core/mlx5_vdpa_ifc.h b/include/linux/mlx5/mlx5_ifc_vdpa.h
similarity index 97%
rename from drivers/vdpa/mlx5/core/mlx5_vdpa_ifc.h
rename to include/linux/mlx5/mlx5_ifc_vdpa.h
index f6f57a29b38e..97784098a992 100644
--- a/drivers/vdpa/mlx5/core/mlx5_vdpa_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc_vdpa.h
@@ -1,8 +1,8 @@
 /* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
 /* Copyright (c) 2020 Mellanox Technologies Ltd. */

-#ifndef __MLX5_VDPA_IFC_H_
-#define __MLX5_VDPA_IFC_H_
+#ifndef __MLX5_IFC_VDPA_H_
+#define __MLX5_IFC_VDPA_H_

 #include <linux/mlx5/mlx5_ifc.h>

@@ -165,4 +165,4 @@ struct mlx5_ifc_modify_virtio_net_q_out_bits {
 	struct mlx5_ifc_general_obj_out_cmd_hdr_bits general_obj_out_cmd_hdr;
 };

-#endif /* __MLX5_VDPA_IFC_H_ */
+#endif /* __MLX5_IFC_VDPA_H_ */
--
2.28.0

Re: [PATCH mlx5-next v1 04/11] vdpa/mlx5: Make hardware definitions visible to all mlx5 devices

[Saeed Mahameed]

On Sun, 2020-11-01 at 22:15 +0200, Leon Romanovsky wrote:
> From: Leon Romanovsky <leonro@nvidia.com>
> 
> Move mlx5_vdpa IFC header file to the general include folder, so
> mlx5_core will be able to reuse it to check if VDPA is supported
> prior to creating an auxiliary device.
> 

I don't really like this, the whole idea of aux devices is that they
get to do own logic and hide details, now we are exposing aux specific
stuff to the bus .. 
let's figure a way to avoid such exposure as we discussed yesterday.

is_supported check shouldn't belong to mlx5_core and each aux device
(en/ib/vdpa) should implement own is_supported op and keep the details
hidden in the aux driver like it was before this patch.

Re: [PATCH mlx5-next v1 04/11] vdpa/mlx5: Make hardware definitions visible to all mlx5 devices

[Jason Gunthorpe]

On Thu, Nov 05, 2020 at 12:31:52PM -0800, Saeed Mahameed wrote:
> On Sun, 2020-11-01 at 22:15 +0200, Leon Romanovsky wrote:
> > From: Leon Romanovsky <leonro@nvidia.com>
> > 
> > Move mlx5_vdpa IFC header file to the general include folder, so
> > mlx5_core will be able to reuse it to check if VDPA is supported
> > prior to creating an auxiliary device.
> > 
> 
> I don't really like this, the whole idea of aux devices is that they
> get to do own logic and hide details, now we are exposing aux
> specific stuff to the bus ..  let's figure a way to avoid such
> exposure as we discussed yesterday.

Not quite, the idea is we get to have a cleaner split between the two
sides.

The device side is responsible for things centric to the device, like
"does this device actually exists" which is what is_supported is
doing.

The driver side holds the driver specific logic.

> is_supported check shouldn't belong to mlx5_core and each aux device
> (en/ib/vdpa) should implement own is_supported op and keep the details
> hidden in the aux driver like it was before this patch.

No, it really should be in the device side.

Part of the point here is to properly fix module loading. That means
the core driver must only create devices that can actually have a
driver bound to them because creating a device triggers module
loading.

For instance we do not want to auto load vdpa modules on every mlx5
system for no reason, that is not clean at all.

Jason

Re: [PATCH mlx5-next v1 04/11] vdpa/mlx5: Make hardware definitions visible to all mlx5 devices

[Leon Romanovsky]

On Thu, Nov 05, 2020 at 04:36:57PM -0400, Jason Gunthorpe wrote:
> On Thu, Nov 05, 2020 at 12:31:52PM -0800, Saeed Mahameed wrote:
> > On Sun, 2020-11-01 at 22:15 +0200, Leon Romanovsky wrote:
> > > From: Leon Romanovsky <leonro@nvidia.com>
> > >
> > > Move mlx5_vdpa IFC header file to the general include folder, so
> > > mlx5_core will be able to reuse it to check if VDPA is supported
> > > prior to creating an auxiliary device.
> > >
> >
> > I don't really like this, the whole idea of aux devices is that they
> > get to do own logic and hide details, now we are exposing aux
> > specific stuff to the bus ..  let's figure a way to avoid such
> > exposure as we discussed yesterday.
>
> Not quite, the idea is we get to have a cleaner split between the two
> sides.
>
> The device side is responsible for things centric to the device, like
> "does this device actually exists" which is what is_supported is
> doing.
>
> The driver side holds the driver specific logic.
>
> > is_supported check shouldn't belong to mlx5_core and each aux device
> > (en/ib/vdpa) should implement own is_supported op and keep the details
> > hidden in the aux driver like it was before this patch.
>
> No, it really should be in the device side.
>
> Part of the point here is to properly fix module loading. That means
> the core driver must only create devices that can actually have a
> driver bound to them because creating a device triggers module
> loading.
>
> For instance we do not want to auto load vdpa modules on every mlx5
> system for no reason, that is not clean at all.

Saeed,

Jason gave very good example and it is not far from the real life requirement.
We have an internal task to make sure that mlx5_vdpa is loaded without any
other mlx5_* modules (ib and eth). This series solves it naturally.

Thanks

>
> Jason

[PATCH mlx5-next v1 05/11] net/mlx5: Register mlx5 devices to auxiliary virtual bus

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Create auxiliary devices under new virtual bus. This will replace
the custom-made mlx5 ->add()/->remove() interfaces and next patches
will fill the missing callback and remove the old interface logic.

The attachment of auxiliary drivers to the devices is possible in
1-to-1 manner only and it requires us to create device for every protocol,
so that device (module) will be able to connect to it.

System with 2 IB and 1 RoCE cards:
[leonro@vm ~]$ lspci |grep nox
00:09.0 Ethernet controller: Mellanox Technologies MT27800 Family [ConnectX-5]
00:0a.0 Ethernet controller: Mellanox Technologies MT28908 Family [ConnectX-6]
00:0b.0 Ethernet controller: Mellanox Technologies MT2910 Family [ConnectX-7]
[leonro@vm ~]$ ls -l /sys/bus/auxiliary/devices/
 mlx5_core.eth.2 -> ../../../devices/pci0000:00/0000:00:0b.0/mlx5_core.eth.2
 mlx5_core.rdma.0 -> ../../../devices/pci0000:00/0000:00:09.0/mlx5_core.rdma.0
 mlx5_core.rdma.1 -> ../../../devices/pci0000:00/0000:00:0a.0/mlx5_core.rdma.1
 mlx5_core.rdma.2 -> ../../../devices/pci0000:00/0000:00:0b.0/mlx5_core.rdma.2
 mlx5_core.vdpa.1 -> ../../../devices/pci0000:00/0000:00:0a.0/mlx5_core.vdpa.1
 mlx5_core.vdpa.2 -> ../../../devices/pci0000:00/0000:00:0b.0/mlx5_core.vdpa.2
[leonro@vm ~]$ rdma dev
0: ibp0s9: node_type ca fw 4.6.9999 node_guid 5254:00c0:fe12:3455 sys_image_guid 5254:00c0:fe12:3455
1: ibp0s10: node_type ca fw 4.6.9999 node_guid 5254:00c0:fe12:3456 sys_image_guid 5254:00c0:fe12:3456
2: rdmap0s11: node_type ca fw 4.6.9999 node_guid 5254:00c0:fe12:3457 sys_image_guid 5254:00c0:fe12:3457

System with RoCE SR-IOV card with 4 VFs:
[leonro@vm ~]$ lspci |grep nox
01:00.0 Ethernet controller: Mellanox Technologies MT28908 Family [ConnectX-6]
01:00.1 Ethernet controller: Mellanox Technologies MT28908 Family [ConnectX-6 Virtual Function]
01:00.2 Ethernet controller: Mellanox Technologies MT28908 Family [ConnectX-6 Virtual Function]
01:00.3 Ethernet controller: Mellanox Technologies MT28908 Family [ConnectX-6 Virtual Function]
01:00.4 Ethernet controller: Mellanox Technologies MT28908 Family [ConnectX-6 Virtual Function]
[leonro@vm ~]$ ls -l /sys/bus/auxiliary/devices/
 mlx5_core.eth.0 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.0/mlx5_core.eth.0
 mlx5_core.eth.1 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.1/mlx5_core.eth.1
 mlx5_core.eth.2 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.2/mlx5_core.eth.2
 mlx5_core.eth.3 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.3/mlx5_core.eth.3
 mlx5_core.eth.4 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.4/mlx5_core.eth.4
 mlx5_core.rdma.0 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.0/mlx5_core.rdma.0
 mlx5_core.rdma.1 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.1/mlx5_core.rdma.1
 mlx5_core.rdma.2 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.2/mlx5_core.rdma.2
 mlx5_core.rdma.3 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.3/mlx5_core.rdma.3
 mlx5_core.rdma.4 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.4/mlx5_core.rdma.4
 mlx5_core.vdpa.1 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.1/mlx5_core.vdpa.1
 mlx5_core.vdpa.2 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.2/mlx5_core.vdpa.2
 mlx5_core.vdpa.3 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.3/mlx5_core.vdpa.3
 mlx5_core.vdpa.4 -> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.4/mlx5_core.vdpa.4
[leonro@vm ~]$ rdma dev
0: rocep1s0f0: node_type ca fw 4.6.9999 node_guid 5254:00c0:fe12:3455 sys_image_guid 5254:00c0:fe12:3455
1: rocep1s0f0v0: node_type ca fw 4.6.9999 node_guid 0000:0000:0000:0000 sys_image_guid 5254:00c0:fe12:3456
2: rocep1s0f0v1: node_type ca fw 4.6.9999 node_guid 0000:0000:0000:0000 sys_image_guid 5254:00c0:fe12:3457
3: rocep1s0f0v2: node_type ca fw 4.6.9999 node_guid 0000:0000:0000:0000 sys_image_guid 5254:00c0:fe12:3458
4: rocep1s0f0v3: node_type ca fw 4.6.9999 node_guid 0000:0000:0000:0000 sys_image_guid 5254:00c0:fe12:3459

Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 .../net/ethernet/mellanox/mlx5/core/Kconfig   |   1 +
 drivers/net/ethernet/mellanox/mlx5/core/dev.c | 265 +++++++++++++++++-
 .../net/ethernet/mellanox/mlx5/core/main.c    |  24 +-
 .../ethernet/mellanox/mlx5/core/mlx5_core.h   |  20 +-
 include/linux/mlx5/driver.h                   |  26 +-
 5 files changed, 325 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
index 99f1ec3b2575..485478979b1a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
@@ -6,6 +6,7 @@
 config MLX5_CORE
 	tristate "Mellanox 5th generation network adapters (ConnectX series) core driver"
 	depends on PCI
+	select AUXILIARY_BUS
 	select NET_DEVLINK
 	depends on VXLAN || !VXLAN
 	depends on MLXFW || !MLXFW
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
index 1972ddd12704..8ddf469b2d05 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
@@ -37,6 +37,7 @@ static LIST_HEAD(intf_list);
 static LIST_HEAD(mlx5_dev_list);
 /* intf dev list mutex */
 static DEFINE_MUTEX(mlx5_intf_mutex);
+static DEFINE_IDA(mlx5_adev_ida);

 struct mlx5_device_context {
 	struct list_head	list;
@@ -50,6 +51,39 @@ enum {
 	MLX5_INTERFACE_ATTACHED,
 };

+static const struct mlx5_adev_device {
+	const char *suffix;
+	bool (*is_supported)(struct mlx5_core_dev *dev);
+} mlx5_adev_devices[1] = {};
+
+int mlx5_adev_idx_alloc(void)
+{
+	return ida_alloc(&mlx5_adev_ida, GFP_KERNEL);
+}
+
+void mlx5_adev_idx_free(int idx)
+{
+	ida_free(&mlx5_adev_ida, idx);
+}
+
+int mlx5_adev_init(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+
+	priv->adev = kcalloc(ARRAY_SIZE(mlx5_adev_devices),
+			     sizeof(struct mlx5_adev *), GFP_KERNEL);
+	if (!priv->adev)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx5_adev_cleanup(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+
+	kfree(priv->adev);
+}

 void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
 {
@@ -135,15 +169,99 @@ static void mlx5_attach_interface(struct mlx5_interface *intf, struct mlx5_priv
 	}
 }

-void mlx5_attach_device(struct mlx5_core_dev *dev)
+static void adev_release(struct device *dev)
+{
+	struct mlx5_adev *mlx5_adev =
+		container_of(dev, struct mlx5_adev, adev.dev);
+	struct mlx5_priv *priv = &mlx5_adev->mdev->priv;
+	int idx = mlx5_adev->idx;
+
+	kfree(mlx5_adev);
+	priv->adev[idx] = NULL;
+}
+
+static struct mlx5_adev *add_adev(struct mlx5_core_dev *dev, int idx)
+{
+	const char *suffix = mlx5_adev_devices[idx].suffix;
+	struct auxiliary_device *adev;
+	struct mlx5_adev *madev;
+	int ret;
+
+	madev = kzalloc(sizeof(*madev), GFP_KERNEL);
+	if (!madev)
+		return ERR_PTR(-ENOMEM);
+
+	adev = &madev->adev;
+	adev->id = dev->priv.adev_idx;
+	adev->name = suffix;
+	adev->dev.parent = dev->device;
+	adev->dev.release = adev_release;
+	madev->mdev = dev;
+	madev->idx = idx;
+
+	ret = auxiliary_device_init(adev);
+	if (ret) {
+		kfree(madev);
+		return ERR_PTR(ret);
+	}
+
+	ret = auxiliary_device_add(adev);
+	if (ret) {
+		auxiliary_device_uninit(adev);
+		return ERR_PTR(ret);
+	}
+	return madev;
+}
+
+static void del_adev(struct auxiliary_device *adev)
+{
+	auxiliary_device_delete(adev);
+	auxiliary_device_uninit(adev);
+}
+
+int mlx5_attach_device(struct mlx5_core_dev *dev)
 {
 	struct mlx5_priv *priv = &dev->priv;
+	struct auxiliary_device *adev;
+	struct auxiliary_driver *adrv;
 	struct mlx5_interface *intf;
+	int ret = 0, i;

 	mutex_lock(&mlx5_intf_mutex);
+	for (i = 0; i < ARRAY_SIZE(mlx5_adev_devices); i++) {
+		if (!priv->adev[i]) {
+			bool is_supported = false;
+
+			if (mlx5_adev_devices[i].is_supported)
+				is_supported = mlx5_adev_devices[i].is_supported(dev);
+
+			if (!is_supported)
+				continue;
+
+			priv->adev[i] = add_adev(dev, i);
+			if (IS_ERR(priv->adev[i])) {
+				ret = PTR_ERR(priv->adev[i]);
+				priv->adev[i] = NULL;
+			}
+		} else {
+			adev = &priv->adev[i]->adev;
+			adrv = to_auxiliary_drv(adev->dev.driver);
+
+			if (adrv->resume)
+				ret = adrv->resume(adev);
+		}
+		if (ret) {
+			mlx5_core_warn(dev, "Device[%d] (%s) failed to load\n",
+				       i, mlx5_adev_devices[i].suffix);
+
+			break;
+		}
+	}
+
 	list_for_each_entry(intf, &intf_list, list)
 		mlx5_attach_interface(intf, priv);
 	mutex_unlock(&mlx5_intf_mutex);
+	return ret;
 }

 static void mlx5_detach_interface(struct mlx5_interface *intf, struct mlx5_priv *priv)
@@ -171,9 +289,29 @@ static void mlx5_detach_interface(struct mlx5_interface *intf, struct mlx5_priv
 void mlx5_detach_device(struct mlx5_core_dev *dev)
 {
 	struct mlx5_priv *priv = &dev->priv;
+	struct auxiliary_device *adev;
+	struct auxiliary_driver *adrv;
 	struct mlx5_interface *intf;
+	pm_message_t pm = {};
+	int i;

 	mutex_lock(&mlx5_intf_mutex);
+	for (i = ARRAY_SIZE(mlx5_adev_devices) - 1; i >= 0; i--) {
+		if (!priv->adev[i])
+			continue;
+
+		adev = &priv->adev[i]->adev;
+		adrv = to_auxiliary_drv(adev->dev.driver);
+
+		if (adrv->suspend) {
+			adrv->suspend(adev, pm);
+			continue;
+		}
+
+		del_adev(&priv->adev[i]->adev);
+		priv->adev[i] = NULL;
+	}
+
 	list_for_each_entry(intf, &intf_list, list)
 		mlx5_detach_interface(intf, priv);
 	mutex_unlock(&mlx5_intf_mutex);
@@ -193,16 +331,30 @@ bool mlx5_device_registered(struct mlx5_core_dev *dev)
 	return found;
 }

-void mlx5_register_device(struct mlx5_core_dev *dev)
+int mlx5_register_device(struct mlx5_core_dev *dev)
 {
 	struct mlx5_priv *priv = &dev->priv;
 	struct mlx5_interface *intf;
+	int ret;
+
+	mutex_lock(&mlx5_intf_mutex);
+	dev->priv.flags &= ~MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV;
+	ret = mlx5_rescan_drivers_locked(dev);
+	mutex_unlock(&mlx5_intf_mutex);
+	if (ret)
+		goto add_err;

 	mutex_lock(&mlx5_intf_mutex);
 	list_add_tail(&priv->dev_list, &mlx5_dev_list);
 	list_for_each_entry(intf, &intf_list, list)
 		mlx5_add_device(intf, priv);
 	mutex_unlock(&mlx5_intf_mutex);
+
+	return 0;
+
+add_err:
+	mlx5_unregister_device(dev);
+	return ret;
 }

 void mlx5_unregister_device(struct mlx5_core_dev *dev)
@@ -214,6 +366,9 @@ void mlx5_unregister_device(struct mlx5_core_dev *dev)
 	list_for_each_entry_reverse(intf, &intf_list, list)
 		mlx5_remove_device(intf, priv);
 	list_del(&priv->dev_list);
+
+	dev->priv.flags |= MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV;
+	mlx5_rescan_drivers_locked(dev);
 	mutex_unlock(&mlx5_intf_mutex);
 }

@@ -246,6 +401,77 @@ void mlx5_unregister_interface(struct mlx5_interface *intf)
 }
 EXPORT_SYMBOL(mlx5_unregister_interface);

+static int add_drivers(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	int i, ret = 0;
+
+	for (i = 0; i < ARRAY_SIZE(mlx5_adev_devices); i++) {
+		bool is_supported = false;
+
+		if (priv->adev[i])
+			continue;
+
+		if (mlx5_adev_devices[i].is_supported)
+			is_supported = mlx5_adev_devices[i].is_supported(dev);
+
+		if (!is_supported)
+			continue;
+
+		priv->adev[i] = add_adev(dev, i);
+		if (IS_ERR(priv->adev[i])) {
+			mlx5_core_warn(dev, "Device[%d] (%s) failed to load\n",
+				       i, mlx5_adev_devices[i].suffix);
+			/* We continue to rescan drivers and leave to the caller
+			 * to make decision if to release everything or continue.
+			 */
+			ret = PTR_ERR(priv->adev[i]);
+			priv->adev[i] = NULL;
+		}
+	}
+	return ret;
+}
+
+static void delete_drivers(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	bool delete_all;
+	int i;
+
+	delete_all = priv->flags & MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV;
+
+	for (i = ARRAY_SIZE(mlx5_adev_devices) - 1; i >= 0; i--) {
+		bool is_supported = false;
+
+		if (!priv->adev[i])
+			continue;
+
+		if (mlx5_adev_devices[i].is_supported && !delete_all)
+			is_supported = mlx5_adev_devices[i].is_supported(dev);
+
+		if (is_supported)
+			continue;
+
+		del_adev(&priv->adev[i]->adev);
+		priv->adev[i] = NULL;
+	}
+}
+
+/* This function is used after mlx5_core_dev is reconfigured.
+ */
+int mlx5_rescan_drivers_locked(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+
+	lockdep_assert_held(&mlx5_intf_mutex);
+
+	delete_drivers(dev);
+	if (priv->flags & MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV)
+		return 0;
+
+	return add_drivers(dev);
+}
+
 /* Must be called with intf_mutex held */
 static bool mlx5_has_added_dev_by_protocol(struct mlx5_core_dev *mdev, int protocol)
 {
@@ -299,24 +525,55 @@ void mlx5_remove_dev_by_protocol(struct mlx5_core_dev *dev, int protocol)
 		}
 }

-static u32 mlx5_gen_pci_id(struct mlx5_core_dev *dev)
+static u32 mlx5_gen_pci_id(const struct mlx5_core_dev *dev)
 {
 	return (u32)((pci_domain_nr(dev->pdev->bus) << 16) |
 		     (dev->pdev->bus->number << 8) |
 		     PCI_SLOT(dev->pdev->devfn));
 }

-/* Must be called with intf_mutex held */
+static int next_phys_dev(struct device *dev, const void *data)
+{
+	struct mlx5_adev *madev = container_of(dev, struct mlx5_adev, adev.dev);
+	struct mlx5_core_dev *mdev = madev->mdev;
+	const struct mlx5_core_dev *curr = data;
+
+	if (!mlx5_core_is_pf(mdev))
+		return 0;
+
+	if (mdev == curr)
+		return 0;
+
+	if (mlx5_gen_pci_id(mdev) != mlx5_gen_pci_id(curr))
+		return 0;
+
+	return 1;
+}
+
+/* This function is called with two flows:
+ * 1. During initialization of mlx5_core_dev and we don't need to lock it.
+ * 2. During LAG configure stage and caller holds &mlx5_intf_mutex.
+ */
 struct mlx5_core_dev *mlx5_get_next_phys_dev(struct mlx5_core_dev *dev)
 {
 	struct mlx5_core_dev *res = NULL;
 	struct mlx5_core_dev *tmp_dev;
+	struct auxiliary_device *adev;
+	struct mlx5_adev *madev;
 	struct mlx5_priv *priv;
 	u32 pci_id;

 	if (!mlx5_core_is_pf(dev))
 		return NULL;

+	adev = auxiliary_find_device(NULL, dev, &next_phys_dev);
+	if (adev) {
+		madev = container_of(adev, struct mlx5_adev, adev);
+
+		put_device(&adev->dev);
+		return madev->mdev;
+	}
+
 	pci_id = mlx5_gen_pci_id(dev);
 	list_for_each_entry(priv, &mlx5_dev_list, dev_list) {
 		tmp_dev = container_of(priv, struct mlx5_core_dev, priv);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 9827127cb674..9397762c3a69 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -1222,14 +1222,21 @@ int mlx5_load_one(struct mlx5_core_dev *dev, bool boot)
 		err = mlx5_devlink_register(priv_to_devlink(dev), dev->device);
 		if (err)
 			goto err_devlink_reg;
-		mlx5_register_device(dev);
+
+		err = mlx5_register_device(dev);
 	} else {
-		mlx5_attach_device(dev);
+		err = mlx5_attach_device(dev);
 	}

+	if (err)
+		goto err_register;
+
 	mutex_unlock(&dev->intf_state_mutex);
 	return 0;

+err_register:
+	if (boot)
+		mlx5_devlink_unregister(priv_to_devlink(dev));
 err_devlink_reg:
 	clear_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state);
 	mlx5_unload(dev);
@@ -1306,8 +1313,14 @@ static int mlx5_mdev_init(struct mlx5_core_dev *dev, int profile_idx)
 	if (err)
 		goto err_pagealloc_init;

+	err = mlx5_adev_init(dev);
+	if (err)
+		goto err_adev_init;
+
 	return 0;

+err_adev_init:
+	mlx5_pagealloc_cleanup(dev);
 err_pagealloc_init:
 	mlx5_health_cleanup(dev);
 err_health_init:
@@ -1324,6 +1337,7 @@ static void mlx5_mdev_uninit(struct mlx5_core_dev *dev)
 {
 	struct mlx5_priv *priv = &dev->priv;

+	mlx5_adev_cleanup(dev);
 	mlx5_pagealloc_cleanup(dev);
 	mlx5_health_cleanup(dev);
 	debugfs_remove_recursive(dev->priv.dbg_root);
@@ -1354,6 +1368,10 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 	dev->coredev_type = id->driver_data & MLX5_PCI_DEV_IS_VF ?
 			 MLX5_COREDEV_VF : MLX5_COREDEV_PF;

+	dev->priv.adev_idx = mlx5_adev_idx_alloc();
+	if (dev->priv.adev_idx < 0)
+		return dev->priv.adev_idx;
+
 	err = mlx5_mdev_init(dev, prof_sel);
 	if (err)
 		goto mdev_init_err;
@@ -1387,6 +1405,7 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 pci_init_err:
 	mlx5_mdev_uninit(dev);
 mdev_init_err:
+	mlx5_adev_idx_free(dev->priv.adev_idx);
 	mlx5_devlink_free(devlink);

 	return err;
@@ -1403,6 +1422,7 @@ static void remove_one(struct pci_dev *pdev)
 	mlx5_unload_one(dev, true);
 	mlx5_pci_close(dev);
 	mlx5_mdev_uninit(dev);
+	mlx5_adev_idx_free(dev->priv.adev_idx);
 	mlx5_devlink_free(devlink);
 }

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index b285f1515e4e..11195381a870 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -178,12 +178,17 @@ void mlx5_events_cleanup(struct mlx5_core_dev *dev);
 void mlx5_events_start(struct mlx5_core_dev *dev);
 void mlx5_events_stop(struct mlx5_core_dev *dev);

+int mlx5_adev_idx_alloc(void);
+void mlx5_adev_idx_free(int idx);
+void mlx5_adev_cleanup(struct mlx5_core_dev *dev);
+int mlx5_adev_init(struct mlx5_core_dev *dev);
+
 void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv);
 void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv);
-void mlx5_attach_device(struct mlx5_core_dev *dev);
+int mlx5_attach_device(struct mlx5_core_dev *dev);
 void mlx5_detach_device(struct mlx5_core_dev *dev);
 bool mlx5_device_registered(struct mlx5_core_dev *dev);
-void mlx5_register_device(struct mlx5_core_dev *dev);
+int mlx5_register_device(struct mlx5_core_dev *dev);
 void mlx5_unregister_device(struct mlx5_core_dev *dev);
 void mlx5_add_dev_by_protocol(struct mlx5_core_dev *dev, int protocol);
 void mlx5_remove_dev_by_protocol(struct mlx5_core_dev *dev, int protocol);
@@ -232,6 +237,17 @@ static inline int mlx5_lag_is_lacp_owner(struct mlx5_core_dev *dev)
 		    MLX5_CAP_GEN(dev, lag_master);
 }

+int mlx5_rescan_drivers_locked(struct mlx5_core_dev *dev);
+static inline int mlx5_rescan_drivers(struct mlx5_core_dev *dev)
+{
+	int ret;
+
+	mlx5_dev_list_lock();
+	ret = mlx5_rescan_drivers_locked(dev);
+	mlx5_dev_list_unlock();
+	return ret;
+}
+
 void mlx5_reload_interface(struct mlx5_core_dev *mdev, int protocol);
 void mlx5_lag_update(struct mlx5_core_dev *dev);

diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index ed1d030658d2..f35ee4855777 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -48,6 +48,7 @@
 #include <linux/idr.h>
 #include <linux/notifier.h>
 #include <linux/refcount.h>
+#include <linux/auxiliary_bus.h>

 #include <linux/mlx5/device.h>
 #include <linux/mlx5/doorbell.h>
@@ -536,6 +537,17 @@ struct mlx5_core_roce {
 	struct mlx5_flow_handle *allow_rule;
 };

+enum {
+	MLX5_PRIV_FLAGS_DISABLE_IB_ADEV = 1 << 0,
+	MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV = 1 << 1,
+};
+
+struct mlx5_adev {
+	struct auxiliary_device adev;
+	struct mlx5_core_dev *mdev;
+	int idx;
+};
+
 struct mlx5_priv {
 	/* IRQ table valid only for real pci devices PF or VF */
 	struct mlx5_irq_table   *irq_table;
@@ -573,6 +585,8 @@ struct mlx5_priv {
 	struct list_head        dev_list;
 	struct list_head        ctx_list;
 	spinlock_t              ctx_lock;
+	struct mlx5_adev       **adev;
+	int			adev_idx;
 	struct mlx5_events      *events;

 	struct mlx5_flow_steering *steering;
@@ -580,6 +594,7 @@ struct mlx5_priv {
 	struct mlx5_eswitch     *eswitch;
 	struct mlx5_core_sriov	sriov;
 	struct mlx5_lag		*lag;
+	u32			flags;
 	struct mlx5_devcom	*devcom;
 	struct mlx5_fw_reset	*fw_reset;
 	struct mlx5_core_roce	roce;
@@ -1063,9 +1078,14 @@ enum {
 };

 enum {
-	MLX5_INTERFACE_PROTOCOL_IB  = 0,
-	MLX5_INTERFACE_PROTOCOL_ETH = 1,
-	MLX5_INTERFACE_PROTOCOL_VDPA = 2,
+	MLX5_INTERFACE_PROTOCOL_ETH_REP,
+	MLX5_INTERFACE_PROTOCOL_ETH,
+
+	MLX5_INTERFACE_PROTOCOL_IB_REP,
+	MLX5_INTERFACE_PROTOCOL_MPIB,
+	MLX5_INTERFACE_PROTOCOL_IB,
+
+	MLX5_INTERFACE_PROTOCOL_VDPA,
 };

 struct mlx5_interface {
--
2.28.0

Re: [PATCH mlx5-next v1 05/11] net/mlx5: Register mlx5 devices to auxiliary virtual bus

[Saeed Mahameed]

On Sun, 2020-11-01 at 22:15 +0200, Leon Romanovsky wrote:
> From: Leon Romanovsky <leonro@nvidia.com>
> 
> Create auxiliary devices under new virtual bus. This will replace
> the custom-made mlx5 ->add()/->remove() interfaces and next patches
> will fill the missing callback and remove the old interface logic.
> 
> The attachment of auxiliary drivers to the devices is possible in
> 1-to-1 manner only and it requires us to create device for every
> protocol,
> so that device (module) will be able to connect to it.
> 
> System with 2 IB and 1 RoCE cards:
> [leonro@vm ~]$ lspci |grep nox
> 00:09.0 Ethernet controller: Mellanox Technologies MT27800 Family
> [ConnectX-5]
> 00:0a.0 Ethernet controller: Mellanox Technologies MT28908 Family
> [ConnectX-6]
> 00:0b.0 Ethernet controller: Mellanox Technologies MT2910 Family
> [ConnectX-7]
> [leonro@vm ~]$ ls -l /sys/bus/auxiliary/devices/
>  mlx5_core.eth.2 ->
> ../../../devices/pci0000:00/0000:00:0b.0/mlx5_core.eth.2
>  mlx5_core.rdma.0 ->
> ../../../devices/pci0000:00/0000:00:09.0/mlx5_core.rdma.0
>  mlx5_core.rdma.1 ->
> ../../../devices/pci0000:00/0000:00:0a.0/mlx5_core.rdma.1
>  mlx5_core.rdma.2 ->
> ../../../devices/pci0000:00/0000:00:0b.0/mlx5_core.rdma.2
>  mlx5_core.vdpa.1 ->
> ../../../devices/pci0000:00/0000:00:0a.0/mlx5_core.vdpa.1
>  mlx5_core.vdpa.2 ->
> ../../../devices/pci0000:00/0000:00:0b.0/mlx5_core.vdpa.2
> [leonro@vm ~]$ rdma dev
> 0: ibp0s9: node_type ca fw 4.6.9999 node_guid 5254:00c0:fe12:3455
> sys_image_guid 5254:00c0:fe12:3455
> 1: ibp0s10: node_type ca fw 4.6.9999 node_guid 5254:00c0:fe12:3456
> sys_image_guid 5254:00c0:fe12:3456
> 2: rdmap0s11: node_type ca fw 4.6.9999 node_guid 5254:00c0:fe12:3457
> sys_image_guid 5254:00c0:fe12:3457
> 
> System with RoCE SR-IOV card with 4 VFs:
> [leonro@vm ~]$ lspci |grep nox
> 01:00.0 Ethernet controller: Mellanox Technologies MT28908 Family
> [ConnectX-6]
> 01:00.1 Ethernet controller: Mellanox Technologies MT28908 Family
> [ConnectX-6 Virtual Function]
> 01:00.2 Ethernet controller: Mellanox Technologies MT28908 Family
> [ConnectX-6 Virtual Function]
> 01:00.3 Ethernet controller: Mellanox Technologies MT28908 Family
> [ConnectX-6 Virtual Function]
> 01:00.4 Ethernet controller: Mellanox Technologies MT28908 Family
> [ConnectX-6 Virtual Function]
> [leonro@vm ~]$ ls -l /sys/bus/auxiliary/devices/
>  mlx5_core.eth.0 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.0/mlx5_core.eth.0
>  mlx5_core.eth.1 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.1/mlx5_core.eth.1
>  mlx5_core.eth.2 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.2/mlx5_core.eth.2
>  mlx5_core.eth.3 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.3/mlx5_core.eth.3
>  mlx5_core.eth.4 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.4/mlx5_core.eth.4
>  mlx5_core.rdma.0 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.0/mlx5_core.rdma.
> 0
>  mlx5_core.rdma.1 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.1/mlx5_core.rdma.
> 1
>  mlx5_core.rdma.2 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.2/mlx5_core.rdma.
> 2
>  mlx5_core.rdma.3 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.3/mlx5_core.rdma.
> 3
>  mlx5_core.rdma.4 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.4/mlx5_core.rdma.
> 4
>  mlx5_core.vdpa.1 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.1/mlx5_core.vdpa.
> 1
>  mlx5_core.vdpa.2 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.2/mlx5_core.vdpa.
> 2
>  mlx5_core.vdpa.3 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.3/mlx5_core.vdpa.
> 3
>  mlx5_core.vdpa.4 ->
> ../../../devices/pci0000:00/0000:00:09.0/0000:01:00.4/mlx5_core.vdpa.
> 4
> [leonro@vm ~]$ rdma dev
> 0: rocep1s0f0: node_type ca fw 4.6.9999 node_guid 5254:00c0:fe12:3455
> sys_image_guid 5254:00c0:fe12:3455
> 1: rocep1s0f0v0: node_type ca fw 4.6.9999 node_guid
> 0000:0000:0000:0000 sys_image_guid 5254:00c0:fe12:3456
> 2: rocep1s0f0v1: node_type ca fw 4.6.9999 node_guid
> 0000:0000:0000:0000 sys_image_guid 5254:00c0:fe12:3457
> 3: rocep1s0f0v2: node_type ca fw 4.6.9999 node_guid
> 0000:0000:0000:0000 sys_image_guid 5254:00c0:fe12:3458
> 4: rocep1s0f0v3: node_type ca fw 4.6.9999 node_guid
> 0000:0000:0000:0000 sys_image_guid 5254:00c0:fe12:3459
> 
> Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
> ---
>  .../net/ethernet/mellanox/mlx5/core/Kconfig   |   1 +
>  drivers/net/ethernet/mellanox/mlx5/core/dev.c | 265
> +++++++++++++++++-
>  .../net/ethernet/mellanox/mlx5/core/main.c    |  24 +-
>  .../ethernet/mellanox/mlx5/core/mlx5_core.h   |  20 +-
>  include/linux/mlx5/driver.h                   |  26 +-
>  5 files changed, 325 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
> b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
> index 99f1ec3b2575..485478979b1a 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
> @@ -6,6 +6,7 @@
>  config MLX5_CORE
>  	tristate "Mellanox 5th generation network adapters (ConnectX
> series) core driver"
>  	depends on PCI
> +	select AUXILIARY_BUS
>  	select NET_DEVLINK
>  	depends on VXLAN || !VXLAN
>  	depends on MLXFW || !MLXFW
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
> b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
> index 1972ddd12704..8ddf469b2d05 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
> @@ -37,6 +37,7 @@ static LIST_HEAD(intf_list);
>  static LIST_HEAD(mlx5_dev_list);
>  /* intf dev list mutex */
>  static DEFINE_MUTEX(mlx5_intf_mutex);
> +static DEFINE_IDA(mlx5_adev_ida);
> 
>  struct mlx5_device_context {
>  	struct list_head	list;
> @@ -50,6 +51,39 @@ enum {
>  	MLX5_INTERFACE_ATTACHED,
>  };
> 
> +static const struct mlx5_adev_device {
> +	const char *suffix;
> +	bool (*is_supported)(struct mlx5_core_dev *dev);
> +} mlx5_adev_devices[1] = {};
> +
> +int mlx5_adev_idx_alloc(void)
> +{
> +	return ida_alloc(&mlx5_adev_ida, GFP_KERNEL);
> +}
> +
> +void mlx5_adev_idx_free(int idx)
> +{
> +	ida_free(&mlx5_adev_ida, idx);
> +}
> +
> +int mlx5_adev_init(struct mlx5_core_dev *dev)
> +{
> +	struct mlx5_priv *priv = &dev->priv;
> +
> +	priv->adev = kcalloc(ARRAY_SIZE(mlx5_adev_devices),
> +			     sizeof(struct mlx5_adev *), GFP_KERNEL);
> +	if (!priv->adev)
> +		return -ENOMEM;
> +
> +	return 0;
> +}
> +
> +void mlx5_adev_cleanup(struct mlx5_core_dev *dev)
> +{
> +	struct mlx5_priv *priv = &dev->priv;
> +
> +	kfree(priv->adev);
> +}
> 
>  void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv
> *priv)
>  {
> @@ -135,15 +169,99 @@ static void mlx5_attach_interface(struct
> mlx5_interface *intf, struct mlx5_priv
>  	}
>  }
> 
> -void mlx5_attach_device(struct mlx5_core_dev *dev)
> +static void adev_release(struct device *dev)
> +{
> +	struct mlx5_adev *mlx5_adev =
> +		container_of(dev, struct mlx5_adev, adev.dev);
> +	struct mlx5_priv *priv = &mlx5_adev->mdev->priv;
> +	int idx = mlx5_adev->idx;
> +
> +	kfree(mlx5_adev);
> +	priv->adev[idx] = NULL;
> +}
> +
> +static struct mlx5_adev *add_adev(struct mlx5_core_dev *dev, int
> idx)
> +{
> +	const char *suffix = mlx5_adev_devices[idx].suffix;
> +	struct auxiliary_device *adev;
> +	struct mlx5_adev *madev;
> +	int ret;
> +
> +	madev = kzalloc(sizeof(*madev), GFP_KERNEL);
> +	if (!madev)
> +		return ERR_PTR(-ENOMEM);
> +
> +	adev = &madev->adev;
> +	adev->id = dev->priv.adev_idx;
> +	adev->name = suffix;
> +	adev->dev.parent = dev->device;
> +	adev->dev.release = adev_release;
> +	madev->mdev = dev;
> +	madev->idx = idx;
> +
> +	ret = auxiliary_device_init(adev);
> +	if (ret) {
> +		kfree(madev);
> +		return ERR_PTR(ret);
> +	}
> +
> +	ret = auxiliary_device_add(adev);
> +	if (ret) {
> +		auxiliary_device_uninit(adev);
> +		return ERR_PTR(ret);
> +	}
> +	return madev;
> +}
> +
> +static void del_adev(struct auxiliary_device *adev)
> +{
> +	auxiliary_device_delete(adev);
> +	auxiliary_device_uninit(adev);
> +}
> +
> +int mlx5_attach_device(struct mlx5_core_dev *dev)
>  {
>  	struct mlx5_priv *priv = &dev->priv;
> +	struct auxiliary_device *adev;
> +	struct auxiliary_driver *adrv;
>  	struct mlx5_interface *intf;
> +	int ret = 0, i;
> 
>  	mutex_lock(&mlx5_intf_mutex);
> +	for (i = 0; i < ARRAY_SIZE(mlx5_adev_devices); i++) {
> +		if (!priv->adev[i]) {
> +			bool is_supported = false;
> +
> +			if (mlx5_adev_devices[i].is_supported)
> +				is_supported =
> mlx5_adev_devices[i].is_supported(dev);
> +
> +			if (!is_supported)
> +				continue;
> +
> +			priv->adev[i] = add_adev(dev, i);
> +			if (IS_ERR(priv->adev[i])) {
> +				ret = PTR_ERR(priv->adev[i]);
> +				priv->adev[i] = NULL;
> +			}
> +		} else {
> +			adev = &priv->adev[i]->adev;
> +			adrv = to_auxiliary_drv(adev->dev.driver);
> +
> +			if (adrv->resume)
> +				ret = adrv->resume(adev);
> +		}
> +		if (ret) {
> +			mlx5_core_warn(dev, "Device[%d] (%s) failed to
> load\n",
> +				       i, mlx5_adev_devices[i].suffix);
> +
> +			break;
> +		}
> +	}
> +
>  	list_for_each_entry(intf, &intf_list, list)
>  		mlx5_attach_interface(intf, priv);
>  	mutex_unlock(&mlx5_intf_mutex);
> +	return ret;
>  }
> 
>  static void mlx5_detach_interface(struct mlx5_interface *intf,
> struct mlx5_priv *priv)
> @@ -171,9 +289,29 @@ static void mlx5_detach_interface(struct
> mlx5_interface *intf, struct mlx5_priv
>  void mlx5_detach_device(struct mlx5_core_dev *dev)
>  {
>  	struct mlx5_priv *priv = &dev->priv;
> +	struct auxiliary_device *adev;
> +	struct auxiliary_driver *adrv;
>  	struct mlx5_interface *intf;
> +	pm_message_t pm = {};
> +	int i;
> 
>  	mutex_lock(&mlx5_intf_mutex);
> +	for (i = ARRAY_SIZE(mlx5_adev_devices) - 1; i >= 0; i--) {
> +		if (!priv->adev[i])
> +			continue;
> +
> +		adev = &priv->adev[i]->adev;
> +		adrv = to_auxiliary_drv(adev->dev.driver);
> +
> +		if (adrv->suspend) {
> +			adrv->suspend(adev, pm);
> +			continue;
> +		}
> +
> +		del_adev(&priv->adev[i]->adev);
> +		priv->adev[i] = NULL;
> +	}
> +
>  	list_for_each_entry(intf, &intf_list, list)
>  		mlx5_detach_interface(intf, priv);
>  	mutex_unlock(&mlx5_intf_mutex);
> @@ -193,16 +331,30 @@ bool mlx5_device_registered(struct
> mlx5_core_dev *dev)
>  	return found;
>  }
> 
> -void mlx5_register_device(struct mlx5_core_dev *dev)
> +int mlx5_register_device(struct mlx5_core_dev *dev)
>  {
>  	struct mlx5_priv *priv = &dev->priv;
>  	struct mlx5_interface *intf;
> +	int ret;
> +
> +	mutex_lock(&mlx5_intf_mutex);
> +	dev->priv.flags &= ~MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV;
> +	ret = mlx5_rescan_drivers_locked(dev);
> +	mutex_unlock(&mlx5_intf_mutex);
> +	if (ret)
> +		goto add_err;
> 
>  	mutex_lock(&mlx5_intf_mutex);
>  	list_add_tail(&priv->dev_list, &mlx5_dev_list);
>  	list_for_each_entry(intf, &intf_list, list)
>  		mlx5_add_device(intf, priv);
>  	mutex_unlock(&mlx5_intf_mutex);
> +
> +	return 0;
> +
> +add_err:
> +	mlx5_unregister_device(dev);
> +	return ret;
>  }
> 
>  void mlx5_unregister_device(struct mlx5_core_dev *dev)
> @@ -214,6 +366,9 @@ void mlx5_unregister_device(struct mlx5_core_dev
> *dev)
>  	list_for_each_entry_reverse(intf, &intf_list, list)
>  		mlx5_remove_device(intf, priv);
>  	list_del(&priv->dev_list);
> +
> +	dev->priv.flags |= MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV;
> +	mlx5_rescan_drivers_locked(dev);
>  	mutex_unlock(&mlx5_intf_mutex);
>  }
> 
> @@ -246,6 +401,77 @@ void mlx5_unregister_interface(struct
> mlx5_interface *intf)
>  }
>  EXPORT_SYMBOL(mlx5_unregister_interface);
> 
> +static int add_drivers(struct mlx5_core_dev *dev)
> +{
> +	struct mlx5_priv *priv = &dev->priv;
> +	int i, ret = 0;
> +
> +	for (i = 0; i < ARRAY_SIZE(mlx5_adev_devices); i++) {
> +		bool is_supported = false;
> +
> +		if (priv->adev[i])
> +			continue;
> +
> +		if (mlx5_adev_devices[i].is_supported)
> +			is_supported =
> mlx5_adev_devices[i].is_supported(dev);
> +
> +		if (!is_supported)
> +			continue;
> +

I think this is wrong for two reasons.

1. is_supported should belong to the ulp aux device itself, and must be
performed before probe. drivers should be added unconditionally and
is_supproted should be checked only prior to probe.

2. you can always load a driver without its underlying device existed.
for example, you can load a pci device driver/module and it will load
and wait for pci devices to pop up, the subsysetem infrastructure will
match between drivers and devices and probe them.

Aux should be the same with the small change that all ulp aux devices
should implement is_supported if they need, since they are virtual
devices they might have some other constrains other than just matching
device ids.



I would suggest the following infra/API semantics changes:

Aux bus parent device:
mlx5_core pci device load/probe(pci_dev) 
{
  struct aux_device *mlx5_aux_dev = alloc_aux_device()

  mlx5_aux_dev->priv = pci_dev;
  register_aux_device("mlx5_core", mlx5_aux_dev);
}



Aux ULP driver:

struct aux_driver mlx5_vpda_aux_driver {

      .name = "vdpa",
       /* match this driver with mlx5_core devices */
      .id_table = {"mlx5_core"}, 
      .ops {
            /* called before probe on actual aux mlx5_core device */
           .is_supported(struct aux_device); 

           .probe = mlx5v_probe,
           .remove = mlx5v_remove,
        }
}

mlx5_vdpa_module_init():
    register_aux_driver(mlx5_vpda_aux_driver);



Aux infrastructure semantics:

  a) on  register_aux_device("mlx5_core", mlx5_aux_dev); 
     it will match all drivers and probe them if is_supported is true,
     reveres flow on unregister_aux_device()

  b) on register_aux_driver(); probe driver on all current devices with
matching ids if is_supported() returned ture


So you don't really need to re-implement mlx5_rescan_drivers_locked and
mlx5_add_drivers in each and every aux device/driver providers, this
should be a aux bus infra' logic.

Re: [PATCH mlx5-next v1 05/11] net/mlx5: Register mlx5 devices to auxiliary virtual bus

[Jason Gunthorpe]

On Thu, Nov 05, 2020 at 12:59:20PM -0800, Saeed Mahameed wrote:

> 2. you can always load a driver without its underlying device existed.
> for example, you can load a pci device driver/module and it will load
> and wait for pci devices to pop up, the subsysetem infrastructure will
> match between drivers and devices and probe them.

Yes, this works fine with this design

> struct aux_driver mlx5_vpda_aux_driver {
> 
>       .name = "vdpa",
>        /* match this driver with mlx5_core devices */
>       .id_table = {"mlx5_core"}, 
>       .ops {
>             /* called before probe on actual aux mlx5_core device */
>            .is_supported(struct aux_device); 

This means module auto loading is impossible, we can't tell to load
the module until we load the module to call the is_supported code ..

Jason

Re: [PATCH mlx5-next v1 05/11] net/mlx5: Register mlx5 devices to auxiliary virtual bus

[Leon Romanovsky]

On Thu, Nov 05, 2020 at 05:09:48PM -0400, Jason Gunthorpe wrote:
> On Thu, Nov 05, 2020 at 12:59:20PM -0800, Saeed Mahameed wrote:
>
> > 2. you can always load a driver without its underlying device existed.
> > for example, you can load a pci device driver/module and it will load
> > and wait for pci devices to pop up, the subsysetem infrastructure will
> > match between drivers and devices and probe them.
>
> Yes, this works fine with this design
>
> > struct aux_driver mlx5_vpda_aux_driver {
> >
> >       .name = "vdpa",
> >        /* match this driver with mlx5_core devices */
> >       .id_table = {"mlx5_core"},
> >       .ops {
> >             /* called before probe on actual aux mlx5_core device */
> >            .is_supported(struct aux_device);
>
> This means module auto loading is impossible, we can't tell to load
> the module until we load the module to call the is_supported code ..

Right, and if we can, it will be violation of everything we know in
driver model, because the call "is_supported" will need to be called
for every registered driver without any relation to existed devices.

And mlx5_rescan_drivers() came as a need to overcome LAG and eswitch
craziness in everything related to their reprobe flows. Once they will
be changed to work like normal drivers, we will be able to simplify it.

So let's talk offline to see how can we improve mlx5_core even more
after this series is merged.

Thanks

>
> Jason

[PATCH mlx5-next v1 06/11] vdpa/mlx5: Connect mlx5_vdpa to auxiliary bus

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Change module registration logic to use auxiliary bus instead of custom
made mlx5 register interface.

Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/dev.c | 28 ++++++-
 drivers/vdpa/mlx5/Makefile                    |  2 +-
 drivers/vdpa/mlx5/net/main.c                  | 76 -------------------
 drivers/vdpa/mlx5/net/mlx5_vnet.c             | 51 +++++++++++--
 drivers/vdpa/mlx5/net/mlx5_vnet.h             | 24 ------
 5 files changed, 72 insertions(+), 109 deletions(-)
 delete mode 100644 drivers/vdpa/mlx5/net/main.c
 delete mode 100644 drivers/vdpa/mlx5/net/mlx5_vnet.h

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
index 8ddf469b2d05..76eb09f4adba 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
@@ -31,6 +31,7 @@
  */

 #include <linux/mlx5/driver.h>
+#include <linux/mlx5/mlx5_ifc_vdpa.h>
 #include "mlx5_core.h"

 static LIST_HEAD(intf_list);
@@ -51,10 +52,35 @@ enum {
 	MLX5_INTERFACE_ATTACHED,
 };

+static bool is_vnet_supported(struct mlx5_core_dev *dev)
+{
+	if (!IS_ENABLED(CONFIG_MLX5_VDPA_NET))
+		return false;
+
+	if (mlx5_core_is_pf(dev))
+		return false;
+
+	if (!(MLX5_CAP_GEN_64(dev, general_obj_types) &
+	      MLX5_GENERAL_OBJ_TYPES_CAP_VIRTIO_NET_Q))
+		return false;
+
+	if (!(MLX5_CAP_DEV_VDPA_EMULATION(dev, event_mode) &
+	      MLX5_VIRTIO_Q_EVENT_MODE_QP_MODE))
+		return false;
+
+	if (!MLX5_CAP_DEV_VDPA_EMULATION(dev, eth_frame_offload_type))
+		return false;
+
+	return true;
+}
+
 static const struct mlx5_adev_device {
 	const char *suffix;
 	bool (*is_supported)(struct mlx5_core_dev *dev);
-} mlx5_adev_devices[1] = {};
+} mlx5_adev_devices[] = {
+	[MLX5_INTERFACE_PROTOCOL_VDPA] = { .suffix = "vnet",
+					   .is_supported = &is_vnet_supported },
+};

 int mlx5_adev_idx_alloc(void)
 {
diff --git a/drivers/vdpa/mlx5/Makefile b/drivers/vdpa/mlx5/Makefile
index 89a5bededc9f..f717978c83bf 100644
--- a/drivers/vdpa/mlx5/Makefile
+++ b/drivers/vdpa/mlx5/Makefile
@@ -1,4 +1,4 @@
 subdir-ccflags-y += -I$(srctree)/drivers/vdpa/mlx5/core

 obj-$(CONFIG_MLX5_VDPA_NET) += mlx5_vdpa.o
-mlx5_vdpa-$(CONFIG_MLX5_VDPA_NET) += net/main.o net/mlx5_vnet.o core/resources.o core/mr.o
+mlx5_vdpa-$(CONFIG_MLX5_VDPA_NET) += net/mlx5_vnet.o core/resources.o core/mr.o
diff --git a/drivers/vdpa/mlx5/net/main.c b/drivers/vdpa/mlx5/net/main.c
deleted file mode 100644
index 4dd3f00f2306..000000000000
--- a/drivers/vdpa/mlx5/net/main.c
+++ /dev/null
@@ -1,76 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
-/* Copyright (c) 2020 Mellanox Technologies Ltd. */
-
-#include <linux/module.h>
-#include <linux/mlx5/driver.h>
-#include <linux/mlx5/device.h>
-#include <linux/mlx5/mlx5_ifc_vdpa.h>
-#include "mlx5_vnet.h"
-
-MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
-MODULE_DESCRIPTION("Mellanox VDPA driver");
-MODULE_LICENSE("Dual BSD/GPL");
-
-static bool required_caps_supported(struct mlx5_core_dev *mdev)
-{
-	u8 event_mode;
-	u64 got;
-
-	got = MLX5_CAP_GEN_64(mdev, general_obj_types);
-
-	if (!(got & MLX5_GENERAL_OBJ_TYPES_CAP_VIRTIO_NET_Q))
-		return false;
-
-	event_mode = MLX5_CAP_DEV_VDPA_EMULATION(mdev, event_mode);
-	if (!(event_mode & MLX5_VIRTIO_Q_EVENT_MODE_QP_MODE))
-		return false;
-
-	if (!MLX5_CAP_DEV_VDPA_EMULATION(mdev, eth_frame_offload_type))
-		return false;
-
-	return true;
-}
-
-static void *mlx5_vdpa_add(struct mlx5_core_dev *mdev)
-{
-	struct mlx5_vdpa_dev *vdev;
-
-	if (mlx5_core_is_pf(mdev))
-		return NULL;
-
-	if (!required_caps_supported(mdev)) {
-		dev_info(mdev->device, "virtio net emulation not supported\n");
-		return NULL;
-	}
-	vdev = mlx5_vdpa_add_dev(mdev);
-	if (IS_ERR(vdev))
-		return NULL;
-
-	return vdev;
-}
-
-static void mlx5_vdpa_remove(struct mlx5_core_dev *mdev, void *context)
-{
-	struct mlx5_vdpa_dev *vdev = context;
-
-	mlx5_vdpa_remove_dev(vdev);
-}
-
-static struct mlx5_interface mlx5_vdpa_interface = {
-	.add = mlx5_vdpa_add,
-	.remove = mlx5_vdpa_remove,
-	.protocol = MLX5_INTERFACE_PROTOCOL_VDPA,
-};
-
-static int __init mlx5_vdpa_init(void)
-{
-	return mlx5_register_interface(&mlx5_vdpa_interface);
-}
-
-static void __exit mlx5_vdpa_exit(void)
-{
-	mlx5_unregister_interface(&mlx5_vdpa_interface);
-}
-
-module_init(mlx5_vdpa_init);
-module_exit(mlx5_vdpa_exit);
diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
index 6c218b47b9f1..5316e51e72d4 100644
--- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
+++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
@@ -1,18 +1,27 @@
 // SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
 /* Copyright (c) 2020 Mellanox Technologies Ltd. */

+#include <linux/module.h>
 #include <linux/vdpa.h>
+#include <linux/vringh.h>
+#include <uapi/linux/virtio_net.h>
 #include <uapi/linux/virtio_ids.h>
 #include <linux/virtio_config.h>
+#include <linux/auxiliary_bus.h>
+#include <linux/mlx5/cq.h>
 #include <linux/mlx5/qp.h>
 #include <linux/mlx5/device.h>
+#include <linux/mlx5/driver.h>
 #include <linux/mlx5/vport.h>
 #include <linux/mlx5/fs.h>
-#include <linux/mlx5/device.h>
 #include <linux/mlx5/mlx5_ifc_vdpa.h>
-#include "mlx5_vnet.h"
 #include "mlx5_vdpa.h"

+MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
+MODULE_DESCRIPTION("Mellanox VDPA driver");
+MODULE_LICENSE("Dual BSD/GPL");
+
+#define to_mlx5_vdpa_ndev(__mvdev) container_of(__mvdev, struct mlx5_vdpa_net, mvdev)
 #define to_mvdev(__vdev) container_of((__vdev), struct mlx5_vdpa_dev, vdev)

 #define VALID_FEATURES_MASK                                                                        \
@@ -159,6 +168,11 @@ static bool mlx5_vdpa_debug;
 			mlx5_vdpa_info(mvdev, "%s\n", #_status);                                   \
 	} while (0)

+static inline u32 mlx5_vdpa_max_qps(int max_vqs)
+{
+	return max_vqs / 2;
+}
+
 static void print_status(struct mlx5_vdpa_dev *mvdev, u8 status, bool set)
 {
 	if (status & ~VALID_STATUS_MASK)
@@ -1928,8 +1942,11 @@ static void init_mvqs(struct mlx5_vdpa_net *ndev)
 	}
 }

-void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev)
+static int mlx5v_probe(struct auxiliary_device *adev,
+		       const struct auxiliary_device_id *id)
 {
+	struct mlx5_adev *madev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5_core_dev *mdev = madev->mdev;
 	struct virtio_net_config *config;
 	struct mlx5_vdpa_dev *mvdev;
 	struct mlx5_vdpa_net *ndev;
@@ -1943,7 +1960,7 @@ void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev)
 	ndev = vdpa_alloc_device(struct mlx5_vdpa_net, mvdev.vdev, mdev->device, &mlx5_vdpa_ops,
 				 2 * mlx5_vdpa_max_qps(max_vqs));
 	if (IS_ERR(ndev))
-		return ndev;
+		return PTR_ERR(ndev);

 	ndev->mvdev.max_vqs = max_vqs;
 	mvdev = &ndev->mvdev;
@@ -1972,7 +1989,8 @@ void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev)
 	if (err)
 		goto err_reg;

-	return ndev;
+	dev_set_drvdata(&adev->dev, ndev);
+	return 0;

 err_reg:
 	free_resources(ndev);
@@ -1981,10 +1999,29 @@ void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev)
 err_mtu:
 	mutex_destroy(&ndev->reslock);
 	put_device(&mvdev->vdev.dev);
-	return ERR_PTR(err);
+	return err;
 }

-void mlx5_vdpa_remove_dev(struct mlx5_vdpa_dev *mvdev)
+static int mlx5v_remove(struct auxiliary_device *adev)
 {
+	struct mlx5_vdpa_dev *mvdev = dev_get_drvdata(&adev->dev);
+
 	vdpa_unregister_device(&mvdev->vdev);
+	return 0;
 }
+
+static const struct auxiliary_device_id mlx5v_id_table[] = {
+	{ .name = MLX5_ADEV_NAME ".vnet", },
+	{},
+};
+
+MODULE_DEVICE_TABLE(auxiliary, mlx5v_id_table);
+
+static struct auxiliary_driver mlx5v_driver = {
+	.name = "vnet",
+	.probe = mlx5v_probe,
+	.remove = mlx5v_remove,
+	.id_table = mlx5v_id_table,
+};
+
+module_auxiliary_driver(mlx5v_driver);
diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.h b/drivers/vdpa/mlx5/net/mlx5_vnet.h
deleted file mode 100644
index f2d6d68b020e..000000000000
--- a/drivers/vdpa/mlx5/net/mlx5_vnet.h
+++ /dev/null
@@ -1,24 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
-/* Copyright (c) 2020 Mellanox Technologies Ltd. */
-
-#ifndef __MLX5_VNET_H_
-#define __MLX5_VNET_H_
-
-#include <linux/vdpa.h>
-#include <linux/virtio_net.h>
-#include <linux/vringh.h>
-#include <linux/mlx5/driver.h>
-#include <linux/mlx5/cq.h>
-#include <linux/mlx5/qp.h>
-#include "mlx5_vdpa.h"
-
-static inline u32 mlx5_vdpa_max_qps(int max_vqs)
-{
-	return max_vqs / 2;
-}
-
-#define to_mlx5_vdpa_ndev(__mvdev) container_of(__mvdev, struct mlx5_vdpa_net, mvdev)
-void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev);
-void mlx5_vdpa_remove_dev(struct mlx5_vdpa_dev *mvdev);
-
-#endif /* __MLX5_VNET_H_ */
--
2.28.0

Re: [PATCH mlx5-next v1 06/11] vdpa/mlx5: Connect mlx5_vdpa to auxiliary bus

[Jason Gunthorpe]

On Sun, Nov 01, 2020 at 10:15:37PM +0200, Leon Romanovsky wrote:
> diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> index 6c218b47b9f1..5316e51e72d4 100644
> +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> @@ -1,18 +1,27 @@
>  // SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
>  /* Copyright (c) 2020 Mellanox Technologies Ltd. */
> 
> +#include <linux/module.h>
>  #include <linux/vdpa.h>
> +#include <linux/vringh.h>
> +#include <uapi/linux/virtio_net.h>
>  #include <uapi/linux/virtio_ids.h>
>  #include <linux/virtio_config.h>
> +#include <linux/auxiliary_bus.h>
> +#include <linux/mlx5/cq.h>
>  #include <linux/mlx5/qp.h>
>  #include <linux/mlx5/device.h>
> +#include <linux/mlx5/driver.h>
>  #include <linux/mlx5/vport.h>
>  #include <linux/mlx5/fs.h>
> -#include <linux/mlx5/device.h>
>  #include <linux/mlx5/mlx5_ifc_vdpa.h>
> -#include "mlx5_vnet.h"
>  #include "mlx5_vdpa.h"
> 
> +MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
> +MODULE_DESCRIPTION("Mellanox VDPA driver");
> +MODULE_LICENSE("Dual BSD/GPL");
> +
> +#define to_mlx5_vdpa_ndev(__mvdev) container_of(__mvdev, struct mlx5_vdpa_net, mvdev)
>  #define to_mvdev(__vdev) container_of((__vdev), struct mlx5_vdpa_dev, vdev)
> 
>  #define VALID_FEATURES_MASK                                                                        \
> @@ -159,6 +168,11 @@ static bool mlx5_vdpa_debug;
>  			mlx5_vdpa_info(mvdev, "%s\n", #_status);                                   \
>  	} while (0)
> 
> +static inline u32 mlx5_vdpa_max_qps(int max_vqs)
> +{
> +	return max_vqs / 2;
> +}
> +
>  static void print_status(struct mlx5_vdpa_dev *mvdev, u8 status, bool set)
>  {
>  	if (status & ~VALID_STATUS_MASK)
> @@ -1928,8 +1942,11 @@ static void init_mvqs(struct mlx5_vdpa_net *ndev)
>  	}
>  }
> 
> -void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev)
> +static int mlx5v_probe(struct auxiliary_device *adev,
> +		       const struct auxiliary_device_id *id)
>  {
> +	struct mlx5_adev *madev = container_of(adev, struct mlx5_adev, adev);
> +	struct mlx5_core_dev *mdev = madev->mdev;
>  	struct virtio_net_config *config;
>  	struct mlx5_vdpa_dev *mvdev;
>  	struct mlx5_vdpa_net *ndev;
> @@ -1943,7 +1960,7 @@ void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev)
>  	ndev = vdpa_alloc_device(struct mlx5_vdpa_net, mvdev.vdev, mdev->device, &mlx5_vdpa_ops,
>  				 2 * mlx5_vdpa_max_qps(max_vqs));
>  	if (IS_ERR(ndev))
> -		return ndev;
> +		return PTR_ERR(ndev);
> 
>  	ndev->mvdev.max_vqs = max_vqs;
>  	mvdev = &ndev->mvdev;
> @@ -1972,7 +1989,8 @@ void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev)
>  	if (err)
>  		goto err_reg;
> 
> -	return ndev;
> +	dev_set_drvdata(&adev->dev, ndev);
> +	return 0;
> 
>  err_reg:
>  	free_resources(ndev);
> @@ -1981,10 +1999,29 @@ void *mlx5_vdpa_add_dev(struct mlx5_core_dev *mdev)
>  err_mtu:
>  	mutex_destroy(&ndev->reslock);
>  	put_device(&mvdev->vdev.dev);
> -	return ERR_PTR(err);
> +	return err;
>  }
> 
> -void mlx5_vdpa_remove_dev(struct mlx5_vdpa_dev *mvdev)
> +static int mlx5v_remove(struct auxiliary_device *adev)
>  {
> +	struct mlx5_vdpa_dev *mvdev = dev_get_drvdata(&adev->dev);
> +
>  	vdpa_unregister_device(&mvdev->vdev);
> +	return 0;
>  }
> +
> +static const struct auxiliary_device_id mlx5v_id_table[] = {
> +	{ .name = MLX5_ADEV_NAME ".vnet", },
> +	{},
> +};
> +
> +MODULE_DEVICE_TABLE(auxiliary, mlx5v_id_table);
> +
> +static struct auxiliary_driver mlx5v_driver = {
> +	.name = "vnet",
> +	.probe = mlx5v_probe,
> +	.remove = mlx5v_remove,
> +	.id_table = mlx5v_id_table,
> +};

It is hard to see from the diff, but when this patch is applied the
vdpa module looks like I imagined things would look with the auxiliary
bus. It is very similar in structure to a PCI driver with the probe()
function cleanly registering with its subsystem. This is what I'd like
to see from the new Intel RDMA driver.

Greg, I think this patch is the best clean usage example.

I've looked over this series and it has the right idea and
parts. There is definitely more that can be done to improve mlx5 in
this area, but this series is well scoped and cleans a good part of
it.

Jason

Re: [PATCH mlx5-next v1 06/11] vdpa/mlx5: Connect mlx5_vdpa to auxiliary bus

[Dan Williams]

On Tue, Nov 3, 2020 at 7:45 AM Jason Gunthorpe <jgg@ziepe.ca> wrote:
[..]
> > +MODULE_DEVICE_TABLE(auxiliary, mlx5v_id_table);
> > +
> > +static struct auxiliary_driver mlx5v_driver = {
> > +     .name = "vnet",
> > +     .probe = mlx5v_probe,
> > +     .remove = mlx5v_remove,
> > +     .id_table = mlx5v_id_table,
> > +};
>
> It is hard to see from the diff, but when this patch is applied the
> vdpa module looks like I imagined things would look with the auxiliary
> bus. It is very similar in structure to a PCI driver with the probe()
> function cleanly registering with its subsystem. This is what I'd like
> to see from the new Intel RDMA driver.
>
> Greg, I think this patch is the best clean usage example.
>
> I've looked over this series and it has the right idea and
> parts. There is definitely more that can be done to improve mlx5 in
> this area, but this series is well scoped and cleans a good part of
> it.

Greg?

I know you alluded to going your own way if the auxiliary bus patches
did not shape up soon, but it seems they have and the stakeholders
have reached this consensus point.

Were there any additional changes you wanted to see happen? I'll go
give the final set another once over, but David has been diligently
fixing up all the declared major issues so I expect to find at most
minor incremental fixups.

Re: [PATCH mlx5-next v1 06/11] vdpa/mlx5: Connect mlx5_vdpa to auxiliary bus

[gregkh]

On Wed, Nov 04, 2020 at 03:21:23PM -0800, Dan Williams wrote:
> On Tue, Nov 3, 2020 at 7:45 AM Jason Gunthorpe <jgg@ziepe.ca> wrote:
> [..]
> > > +MODULE_DEVICE_TABLE(auxiliary, mlx5v_id_table);
> > > +
> > > +static struct auxiliary_driver mlx5v_driver = {
> > > +     .name = "vnet",
> > > +     .probe = mlx5v_probe,
> > > +     .remove = mlx5v_remove,
> > > +     .id_table = mlx5v_id_table,
> > > +};
> >
> > It is hard to see from the diff, but when this patch is applied the
> > vdpa module looks like I imagined things would look with the auxiliary
> > bus. It is very similar in structure to a PCI driver with the probe()
> > function cleanly registering with its subsystem. This is what I'd like
> > to see from the new Intel RDMA driver.
> >
> > Greg, I think this patch is the best clean usage example.
> >
> > I've looked over this series and it has the right idea and
> > parts. There is definitely more that can be done to improve mlx5 in
> > this area, but this series is well scoped and cleans a good part of
> > it.
> 
> Greg?
> 
> I know you alluded to going your own way if the auxiliary bus patches
> did not shape up soon, but it seems they have and the stakeholders
> have reached this consensus point.
> 
> Were there any additional changes you wanted to see happen? I'll go
> give the final set another once over, but David has been diligently
> fixing up all the declared major issues so I expect to find at most
> minor incremental fixups.

This is in my to-review pile, along with a load of other stuff at the
moment:
	$ ~/bin/mdfrm -c ~/mail/todo/
	1709 messages in /home/gregkh/mail/todo/

So give me a chance.  There is no rush on my side for this given the
huge delays that have happened here on the authorship side many times in
the past :)

If you can review it, or anyone else, that is always most appreciated.

thanks,

greg k-h

Re: [PATCH mlx5-next v1 06/11] vdpa/mlx5: Connect mlx5_vdpa to auxiliary bus

[Dan Williams]

On Wed, Nov 4, 2020 at 11:32 PM gregkh <gregkh@linuxfoundation.org> wrote:
>
> On Wed, Nov 04, 2020 at 03:21:23PM -0800, Dan Williams wrote:
> > On Tue, Nov 3, 2020 at 7:45 AM Jason Gunthorpe <jgg@ziepe.ca> wrote:
> > [..]
> > > > +MODULE_DEVICE_TABLE(auxiliary, mlx5v_id_table);
> > > > +
> > > > +static struct auxiliary_driver mlx5v_driver = {
> > > > +     .name = "vnet",
> > > > +     .probe = mlx5v_probe,
> > > > +     .remove = mlx5v_remove,
> > > > +     .id_table = mlx5v_id_table,
> > > > +};
> > >
> > > It is hard to see from the diff, but when this patch is applied the
> > > vdpa module looks like I imagined things would look with the auxiliary
> > > bus. It is very similar in structure to a PCI driver with the probe()
> > > function cleanly registering with its subsystem. This is what I'd like
> > > to see from the new Intel RDMA driver.
> > >
> > > Greg, I think this patch is the best clean usage example.
> > >
> > > I've looked over this series and it has the right idea and
> > > parts. There is definitely more that can be done to improve mlx5 in
> > > this area, but this series is well scoped and cleans a good part of
> > > it.
> >
> > Greg?
> >
> > I know you alluded to going your own way if the auxiliary bus patches
> > did not shape up soon, but it seems they have and the stakeholders
> > have reached this consensus point.
> >
> > Were there any additional changes you wanted to see happen? I'll go
> > give the final set another once over, but David has been diligently
> > fixing up all the declared major issues so I expect to find at most
> > minor incremental fixups.
>
> This is in my to-review pile, along with a load of other stuff at the
> moment:
>         $ ~/bin/mdfrm -c ~/mail/todo/
>         1709 messages in /home/gregkh/mail/todo/
>
> So give me a chance.  There is no rush on my side for this given the
> huge delays that have happened here on the authorship side many times in
> the past :)

Sure, I was more looking to confirm that it's worth continuing to
polish this set given your mention of possibly going a different
direction.

> If you can review it, or anyone else, that is always most appreciated.

Thanks, will do.

Re: [PATCH mlx5-next v1 06/11] vdpa/mlx5: Connect mlx5_vdpa to auxiliary bus

[Jason Gunthorpe]

On Thu, Nov 05, 2020 at 08:33:02AM +0100, gregkh wrote:
> > Were there any additional changes you wanted to see happen? I'll go
> > give the final set another once over, but David has been diligently
> > fixing up all the declared major issues so I expect to find at most
> > minor incremental fixups.
> 
> This is in my to-review pile, along with a load of other stuff at the
> moment:
> 	$ ~/bin/mdfrm -c ~/mail/todo/
> 	1709 messages in /home/gregkh/mail/todo/
> 
> So give me a chance.  There is no rush on my side for this given the
> huge delays that have happened here on the authorship side many times in
> the past :)

On the other hand Leon and his team did invest alot of time and
effort, very quickly, to build and QA this large mlx5 series here to
give a better/second example as you requested only a few weeks ago.

> If you can review it, or anyone else, that is always most appreciated.

Dan, Leon, Myself and others have looked at the auxiliary bus patch a
more than a few times now. Leon in particular went over it very
carefully and a number of bugs were fixed while he developed this
series.

There seems to be nothing fundamentally wrong with it, so long as
people are fine with the colour of the shed...

Jason

Re: [PATCH mlx5-next v1 06/11] vdpa/mlx5: Connect mlx5_vdpa to auxiliary bus

[gregkh]

On Thu, Nov 05, 2020 at 12:47:38PM -0400, Jason Gunthorpe wrote:
> On Thu, Nov 05, 2020 at 08:33:02AM +0100, gregkh wrote:
> > > Were there any additional changes you wanted to see happen? I'll go
> > > give the final set another once over, but David has been diligently
> > > fixing up all the declared major issues so I expect to find at most
> > > minor incremental fixups.
> > 
> > This is in my to-review pile, along with a load of other stuff at the
> > moment:
> > 	$ ~/bin/mdfrm -c ~/mail/todo/
> > 	1709 messages in /home/gregkh/mail/todo/
> > 
> > So give me a chance.  There is no rush on my side for this given the
> > huge delays that have happened here on the authorship side many times in
> > the past :)
> 
> On the other hand Leon and his team did invest alot of time and
> effort, very quickly, to build and QA this large mlx5 series here to
> give a better/second example as you requested only a few weeks ago.

Leon and his team have done a great job, and I never said otherwise.

greg k-h

[PATCH mlx5-next v1 07/11] net/mlx5e: Connect ethernet part to auxiliary bus

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Reuse auxiliary bus to perform device management of the
ethernet part of the mlx5 driver.

Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/dev.c |  74 ++++++++++
 .../net/ethernet/mellanox/mlx5/core/en_main.c | 133 ++++++++----------
 .../net/ethernet/mellanox/mlx5/core/en_rep.c  |  41 +++++-
 .../net/ethernet/mellanox/mlx5/core/en_rep.h  |   6 +-
 .../net/ethernet/mellanox/mlx5/core/eswitch.c |   6 +-
 drivers/net/ethernet/mellanox/mlx5/core/lag.c |   2 +
 .../net/ethernet/mellanox/mlx5/core/main.c    |   6 +-
 .../ethernet/mellanox/mlx5/core/mlx5_core.h   |   2 +-
 8 files changed, 181 insertions(+), 89 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
index 76eb09f4adba..6d413c622785 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
@@ -31,6 +31,7 @@
  */

 #include <linux/mlx5/driver.h>
+#include <linux/mlx5/eswitch.h>
 #include <linux/mlx5/mlx5_ifc_vdpa.h>
 #include "mlx5_core.h"

@@ -52,6 +53,75 @@ enum {
 	MLX5_INTERFACE_ATTACHED,
 };

+static bool is_eth_rep_supported(struct mlx5_core_dev *dev)
+{
+	if (!IS_ENABLED(CONFIG_MLX5_ESWITCH))
+		return false;
+
+	if (!MLX5_ESWITCH_MANAGER(dev))
+		return false;
+
+	if (mlx5_eswitch_mode(dev->priv.eswitch) != MLX5_ESWITCH_OFFLOADS)
+		return false;
+
+	return true;
+}
+
+static bool is_eth_supported(struct mlx5_core_dev *dev)
+{
+	if (!IS_ENABLED(CONFIG_MLX5_CORE_EN))
+		return false;
+
+	if (is_eth_rep_supported(dev))
+		return false;
+
+	if (MLX5_CAP_GEN(dev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+		return false;
+
+	if (!MLX5_CAP_GEN(dev, eth_net_offloads)) {
+		mlx5_core_warn(dev, "Missing eth_net_offloads capability\n");
+		return false;
+	}
+
+	if (!MLX5_CAP_GEN(dev, nic_flow_table)) {
+		mlx5_core_warn(dev, "Missing nic_flow_table capability\n");
+		return false;
+	}
+
+	if (!MLX5_CAP_ETH(dev, csum_cap)) {
+		mlx5_core_warn(dev, "Missing csum_cap capability\n");
+		return false;
+	}
+
+	if (!MLX5_CAP_ETH(dev, max_lso_cap)) {
+		mlx5_core_warn(dev, "Missing max_lso_cap capability\n");
+		return false;
+	}
+
+	if (!MLX5_CAP_ETH(dev, vlan_cap)) {
+		mlx5_core_warn(dev, "Missing vlan_cap capability\n");
+		return false;
+	}
+
+	if (!MLX5_CAP_ETH(dev, rss_ind_tbl_cap)) {
+		mlx5_core_warn(dev, "Missing rss_ind_tbl_cap capability\n");
+		return false;
+	}
+
+	if (MLX5_CAP_FLOWTABLE(dev,
+			       flow_table_properties_nic_receive.max_ft_level) < 3) {
+		mlx5_core_warn(dev, "max_ft_level < 3\n");
+		return false;
+	}
+
+	if (!MLX5_CAP_ETH(dev, self_lb_en_modifiable))
+		mlx5_core_warn(dev, "Self loop back prevention is not supported\n");
+	if (!MLX5_CAP_GEN(dev, cq_moderation))
+		mlx5_core_warn(dev, "CQ moderation is not supported\n");
+
+	return true;
+}
+
 static bool is_vnet_supported(struct mlx5_core_dev *dev)
 {
 	if (!IS_ENABLED(CONFIG_MLX5_VDPA_NET))
@@ -80,6 +150,10 @@ static const struct mlx5_adev_device {
 } mlx5_adev_devices[] = {
 	[MLX5_INTERFACE_PROTOCOL_VDPA] = { .suffix = "vnet",
 					   .is_supported = &is_vnet_supported },
+	[MLX5_INTERFACE_PROTOCOL_ETH] = { .suffix = "eth",
+					  .is_supported = &is_eth_supported },
+	[MLX5_INTERFACE_PROTOCOL_ETH_REP] = { .suffix = "eth-rep",
+					   .is_supported = &is_eth_rep_supported },
 };

 int mlx5_adev_idx_alloc(void)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index b3f02aac7f26..3c4f880c6329 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -4597,31 +4597,6 @@ const struct net_device_ops mlx5e_netdev_ops = {
 	.ndo_get_devlink_port    = mlx5e_get_devlink_port,
 };

-static int mlx5e_check_required_hca_cap(struct mlx5_core_dev *mdev)
-{
-	if (MLX5_CAP_GEN(mdev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
-		return -EOPNOTSUPP;
-	if (!MLX5_CAP_GEN(mdev, eth_net_offloads) ||
-	    !MLX5_CAP_GEN(mdev, nic_flow_table) ||
-	    !MLX5_CAP_ETH(mdev, csum_cap) ||
-	    !MLX5_CAP_ETH(mdev, max_lso_cap) ||
-	    !MLX5_CAP_ETH(mdev, vlan_cap) ||
-	    !MLX5_CAP_ETH(mdev, rss_ind_tbl_cap) ||
-	    MLX5_CAP_FLOWTABLE(mdev,
-			       flow_table_properties_nic_receive.max_ft_level)
-			       < 3) {
-		mlx5_core_warn(mdev,
-			       "Not creating net device, some required device capabilities are missing\n");
-		return -EOPNOTSUPP;
-	}
-	if (!MLX5_CAP_ETH(mdev, self_lb_en_modifiable))
-		mlx5_core_warn(mdev, "Self loop back prevention is not supported\n");
-	if (!MLX5_CAP_GEN(mdev, cq_moderation))
-		mlx5_core_warn(mdev, "CQ moderation is not supported\n");
-
-	return 0;
-}
-
 void mlx5e_build_default_indir_rqt(u32 *indirection_rqt, int len,
 				   int num_channels)
 {
@@ -5440,13 +5415,12 @@ void mlx5e_destroy_netdev(struct mlx5e_priv *priv)
 	free_netdev(netdev);
 }

-/* mlx5e_attach and mlx5e_detach scope should be only creating/destroying
- * hardware contexts and to connect it to the current netdev.
- */
-static int mlx5e_attach(struct mlx5_core_dev *mdev, void *vpriv)
+static int mlx5e_resume(struct auxiliary_device *adev)
 {
-	struct mlx5e_priv *priv = vpriv;
+	struct mlx5_adev *edev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5e_priv *priv = dev_get_drvdata(&adev->dev);
 	struct net_device *netdev = priv->netdev;
+	struct mlx5_core_dev *mdev = edev->mdev;
 	int err;

 	if (netif_device_present(netdev))
@@ -5465,109 +5439,112 @@ static int mlx5e_attach(struct mlx5_core_dev *mdev, void *vpriv)
 	return 0;
 }

-static void mlx5e_detach(struct mlx5_core_dev *mdev, void *vpriv)
+static int mlx5e_suspend(struct auxiliary_device *adev, pm_message_t state)
 {
-	struct mlx5e_priv *priv = vpriv;
+	struct mlx5e_priv *priv = dev_get_drvdata(&adev->dev);
 	struct net_device *netdev = priv->netdev;
-
-#ifdef CONFIG_MLX5_ESWITCH
-	if (MLX5_ESWITCH_MANAGER(mdev) && vpriv == mdev)
-		return;
-#endif
+	struct mlx5_core_dev *mdev = priv->mdev;

 	if (!netif_device_present(netdev))
-		return;
+		return -ENODEV;

 	mlx5e_detach_netdev(priv);
 	mlx5e_destroy_mdev_resources(mdev);
+	return 0;
 }

-static void *mlx5e_add(struct mlx5_core_dev *mdev)
+static int mlx5e_probe(struct auxiliary_device *adev,
+		       const struct auxiliary_device_id *id)
 {
+	struct mlx5_adev *edev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5_core_dev *mdev = edev->mdev;
 	struct net_device *netdev;
+	pm_message_t state = {};
 	void *priv;
 	int err;
 	int nch;

-	err = mlx5e_check_required_hca_cap(mdev);
-	if (err)
-		return NULL;
-
-#ifdef CONFIG_MLX5_ESWITCH
-	if (MLX5_ESWITCH_MANAGER(mdev) &&
-	    mlx5_eswitch_mode(mdev->priv.eswitch) == MLX5_ESWITCH_OFFLOADS) {
-		mlx5e_rep_register_vport_reps(mdev);
-		return mdev;
-	}
-#endif
-
 	nch = mlx5e_get_max_num_channels(mdev);
 	netdev = mlx5e_create_netdev(mdev, &mlx5e_nic_profile, nch, NULL);
 	if (!netdev) {
 		mlx5_core_err(mdev, "mlx5e_create_netdev failed\n");
-		return NULL;
+		return -ENOMEM;
 	}

 	dev_net_set(netdev, mlx5_core_net(mdev));
 	priv = netdev_priv(netdev);
+	dev_set_drvdata(&adev->dev, priv);

-	err = mlx5e_attach(mdev, priv);
+	err = mlx5e_resume(adev);
 	if (err) {
-		mlx5_core_err(mdev, "mlx5e_attach failed, %d\n", err);
+		mlx5_core_err(mdev, "mlx5e_resume failed, %d\n", err);
 		goto err_destroy_netdev;
 	}

 	err = register_netdev(netdev);
 	if (err) {
 		mlx5_core_err(mdev, "register_netdev failed, %d\n", err);
-		goto err_detach;
+		goto err_resume;
 	}

 	mlx5e_devlink_port_type_eth_set(priv);

 	mlx5e_dcbnl_init_app(priv);
-	return priv;
+	return 0;

-err_detach:
-	mlx5e_detach(mdev, priv);
+err_resume:
+	mlx5e_suspend(adev, state);
 err_destroy_netdev:
 	mlx5e_destroy_netdev(priv);
-	return NULL;
+	return err;
 }

-static void mlx5e_remove(struct mlx5_core_dev *mdev, void *vpriv)
+static int mlx5e_remove(struct auxiliary_device *adev)
 {
-	struct mlx5e_priv *priv;
+	struct mlx5e_priv *priv = dev_get_drvdata(&adev->dev);
+	pm_message_t state = {};

-#ifdef CONFIG_MLX5_ESWITCH
-	if (MLX5_ESWITCH_MANAGER(mdev) && vpriv == mdev) {
-		mlx5e_rep_unregister_vport_reps(mdev);
-		return;
-	}
-#endif
-	priv = vpriv;
 	mlx5e_dcbnl_delete_app(priv);
 	unregister_netdev(priv->netdev);
-	mlx5e_detach(mdev, vpriv);
+	mlx5e_suspend(adev, state);
 	mlx5e_destroy_netdev(priv);
+	return 0;
 }

-static struct mlx5_interface mlx5e_interface = {
-	.add       = mlx5e_add,
-	.remove    = mlx5e_remove,
-	.attach    = mlx5e_attach,
-	.detach    = mlx5e_detach,
-	.protocol  = MLX5_INTERFACE_PROTOCOL_ETH,
+static const struct auxiliary_device_id mlx5e_id_table[] = {
+	{ .name = MLX5_ADEV_NAME ".eth", },
+	{},
 };

-void mlx5e_init(void)
+MODULE_DEVICE_TABLE(auxiliary, mlx5e_id_table);
+
+static struct auxiliary_driver mlx5e_driver = {
+	.name = "eth",
+	.probe = mlx5e_probe,
+	.remove = mlx5e_remove,
+	.suspend = mlx5e_suspend,
+	.resume = mlx5e_resume,
+	.id_table = mlx5e_id_table,
+};
+
+int mlx5e_init(void)
 {
+	int ret;
+
 	mlx5e_ipsec_build_inverse_table();
 	mlx5e_build_ptys2ethtool_map();
-	mlx5_register_interface(&mlx5e_interface);
+	ret = mlx5e_rep_init();
+	if (ret)
+		return ret;
+
+	ret = auxiliary_driver_register(&mlx5e_driver);
+	if (ret)
+		mlx5e_rep_cleanup();
+	return ret;
 }

 void mlx5e_cleanup(void)
 {
-	mlx5_unregister_interface(&mlx5e_interface);
+	auxiliary_driver_unregister(&mlx5e_driver);
+	mlx5e_rep_cleanup();
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index ef2f8889ba0f..6a65a732de56 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -1315,16 +1315,49 @@ static const struct mlx5_eswitch_rep_ops rep_ops = {
 	.get_proto_dev = mlx5e_vport_rep_get_proto_dev
 };

-void mlx5e_rep_register_vport_reps(struct mlx5_core_dev *mdev)
+static int mlx5e_rep_probe(struct auxiliary_device *adev,
+			   const struct auxiliary_device_id *id)
 {
-	struct mlx5_eswitch *esw = mdev->priv.eswitch;
+	struct mlx5_adev *edev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5_core_dev *mdev = edev->mdev;
+	struct mlx5_eswitch *esw;

+	esw = mdev->priv.eswitch;
 	mlx5_eswitch_register_vport_reps(esw, &rep_ops, REP_ETH);
+	return 0;
 }

-void mlx5e_rep_unregister_vport_reps(struct mlx5_core_dev *mdev)
+static int mlx5e_rep_remove(struct auxiliary_device *adev)
 {
-	struct mlx5_eswitch *esw = mdev->priv.eswitch;
+	struct mlx5_adev *vdev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5_core_dev *mdev = vdev->mdev;
+	struct mlx5_eswitch *esw;

+	esw = mdev->priv.eswitch;
 	mlx5_eswitch_unregister_vport_reps(esw, REP_ETH);
+	return 0;
+}
+
+static const struct auxiliary_device_id mlx5e_rep_id_table[] = {
+	{ .name = MLX5_ADEV_NAME ".eth-rep", },
+	{},
+};
+
+MODULE_DEVICE_TABLE(auxiliary, mlx5e_rep_id_table);
+
+static struct auxiliary_driver mlx5e_rep_driver = {
+	.name = "eth-rep",
+	.probe = mlx5e_rep_probe,
+	.remove = mlx5e_rep_remove,
+	.id_table = mlx5e_rep_id_table,
+};
+
+int mlx5e_rep_init(void)
+{
+	return auxiliary_driver_register(&mlx5e_rep_driver);
+}
+
+void mlx5e_rep_cleanup(void)
+{
+	auxiliary_driver_unregister(&mlx5e_rep_driver);
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
index 9020d1419bcf..5709b30bf4e1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
@@ -203,8 +203,8 @@ struct mlx5e_rep_sq {
 	struct list_head	 list;
 };

-void mlx5e_rep_register_vport_reps(struct mlx5_core_dev *mdev);
-void mlx5e_rep_unregister_vport_reps(struct mlx5_core_dev *mdev);
+int mlx5e_rep_init(void);
+void mlx5e_rep_cleanup(void);
 int mlx5e_rep_bond_init(struct mlx5e_rep_priv *rpriv);
 void mlx5e_rep_bond_cleanup(struct mlx5e_rep_priv *rpriv);
 int mlx5e_rep_bond_enslave(struct mlx5_eswitch *esw, struct net_device *netdev,
@@ -232,6 +232,8 @@ static inline bool mlx5e_eswitch_rep(struct net_device *netdev)
 static inline bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv) { return false; }
 static inline int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv) { return 0; }
 static inline void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv) {}
+static inline int mlx5e_rep_init(void) { return 0; };
+static inline void mlx5e_rep_cleanup(void) {};
 #endif

 static inline bool mlx5e_is_vport_rep(struct mlx5e_priv *priv)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 2e14bf238588..78a854926b00 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1616,7 +1616,7 @@ int mlx5_eswitch_enable_locked(struct mlx5_eswitch *esw, int mode, int num_vfs)
 	if (mode == MLX5_ESWITCH_LEGACY) {
 		err = esw_legacy_enable(esw);
 	} else {
-		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH);
+		mlx5_rescan_drivers(esw->dev);
 		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
 		err = esw_offloads_enable(esw);
 	}
@@ -1637,7 +1637,7 @@ int mlx5_eswitch_enable_locked(struct mlx5_eswitch *esw, int mode, int num_vfs)

 	if (mode == MLX5_ESWITCH_OFFLOADS) {
 		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
-		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH);
+		mlx5_rescan_drivers(esw->dev);
 	}
 	esw_destroy_tsar(esw);
 	return err;
@@ -1701,7 +1701,7 @@ void mlx5_eswitch_disable_locked(struct mlx5_eswitch *esw, bool clear_vf)

 	if (old_mode == MLX5_ESWITCH_OFFLOADS) {
 		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
-		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH);
+		mlx5_rescan_drivers(esw->dev);
 	}
 	esw_destroy_tsar(esw);

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
index 33081b24f10a..e4d4de1719bd 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -596,6 +596,8 @@ void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev)
 	if (err)
 		mlx5_core_err(dev, "Failed to init multipath lag err=%d\n",
 			      err);
+
+	return;
 }

 /* Must be called with intf_mutex held */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 9397762c3a69..0b4793aff274 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -1677,7 +1677,11 @@ static int __init init(void)
 		goto err_debug;

 #ifdef CONFIG_MLX5_CORE_EN
-	mlx5e_init();
+	err = mlx5e_init();
+	if (err) {
+		pci_unregister_driver(&mlx5_core_driver);
+		goto err_debug;
+	}
 #endif

 	return 0;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index 11195381a870..f7e44e04d465 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -217,7 +217,7 @@ int mlx5_firmware_flash(struct mlx5_core_dev *dev, const struct firmware *fw,
 int mlx5_fw_version_query(struct mlx5_core_dev *dev,
 			  u32 *running_ver, u32 *stored_ver);

-void mlx5e_init(void);
+int mlx5e_init(void);
 void mlx5e_cleanup(void);

 static inline bool mlx5_sriov_is_enabled(struct mlx5_core_dev *dev)
--
2.28.0

[PATCH mlx5-next v1 08/11] RDMA/mlx5: Convert mlx5_ib to use auxiliary bus

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

The conversion to auxiliary bus solves long standing issue with
existing mlx5_ib<->mlx5_core coupling. It required to have both
modules in initramfs if one of them needed for the boot.

Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/mlx5/ib_rep.c           |  77 ++++++---
 drivers/infiniband/hw/mlx5/ib_rep.h           |   8 +-
 drivers/infiniband/hw/mlx5/main.c             | 148 +++++++++++-------
 drivers/infiniband/hw/mlx5/mlx5_ib.h          |   4 +-
 drivers/net/ethernet/mellanox/mlx5/core/dev.c |  66 ++++++++
 .../net/ethernet/mellanox/mlx5/core/eswitch.c |  11 +-
 drivers/net/ethernet/mellanox/mlx5/core/lag.c |  38 +++--
 .../net/ethernet/mellanox/mlx5/core/main.c    |   3 -
 8 files changed, 251 insertions(+), 104 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/ib_rep.c b/drivers/infiniband/hw/mlx5/ib_rep.c
index 5c3d052ac30b..0dc15757cc66 100644
--- a/drivers/infiniband/hw/mlx5/ib_rep.c
+++ b/drivers/infiniband/hw/mlx5/ib_rep.c
@@ -33,6 +33,7 @@ mlx5_ib_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 	const struct mlx5_ib_profile *profile;
 	struct mlx5_ib_dev *ibdev;
 	int vport_index;
+	int ret;

 	if (rep->vport == MLX5_VPORT_UPLINK)
 		profile = &raw_eth_profile;
@@ -46,8 +47,8 @@ mlx5_ib_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 	ibdev->port = kcalloc(num_ports, sizeof(*ibdev->port),
 			      GFP_KERNEL);
 	if (!ibdev->port) {
-		ib_dealloc_device(&ibdev->ib_dev);
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto fail_port;
 	}

 	ibdev->is_rep = true;
@@ -58,12 +59,19 @@ mlx5_ib_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 	ibdev->mdev = dev;
 	ibdev->num_ports = num_ports;

-	if (!__mlx5_ib_add(ibdev, profile))
-		return -EINVAL;
+	ret = __mlx5_ib_add(ibdev, profile);
+	if (ret)
+		goto fail_add;

 	rep->rep_data[REP_IB].priv = ibdev;

 	return 0;
+
+fail_add:
+	kfree(ibdev->port);
+fail_port:
+	ib_dealloc_device(&ibdev->ib_dev);
+	return ret;
 }

 static void
@@ -94,20 +102,6 @@ static const struct mlx5_eswitch_rep_ops rep_ops = {
 	.get_proto_dev = mlx5_ib_vport_get_proto_dev,
 };

-void mlx5_ib_register_vport_reps(struct mlx5_core_dev *mdev)
-{
-	struct mlx5_eswitch *esw = mdev->priv.eswitch;
-
-	mlx5_eswitch_register_vport_reps(esw, &rep_ops, REP_IB);
-}
-
-void mlx5_ib_unregister_vport_reps(struct mlx5_core_dev *mdev)
-{
-	struct mlx5_eswitch *esw = mdev->priv.eswitch;
-
-	mlx5_eswitch_unregister_vport_reps(esw, REP_IB);
-}
-
 u8 mlx5_ib_eswitch_mode(struct mlx5_eswitch *esw)
 {
 	return mlx5_eswitch_mode(esw);
@@ -154,3 +148,50 @@ struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
 	return mlx5_eswitch_add_send_to_vport_rule(esw, rep->vport,
 						   sq->base.mqp.qpn);
 }
+
+static int mlx5r_rep_probe(struct auxiliary_device *adev,
+			   const struct auxiliary_device_id *id)
+{
+	struct mlx5_adev *idev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5_core_dev *mdev = idev->mdev;
+	struct mlx5_eswitch *esw;
+
+	esw = mdev->priv.eswitch;
+	mlx5_eswitch_register_vport_reps(esw, &rep_ops, REP_IB);
+	return 0;
+}
+
+static int mlx5r_rep_remove(struct auxiliary_device *adev)
+{
+	struct mlx5_adev *idev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5_core_dev *mdev = idev->mdev;
+	struct mlx5_eswitch *esw;
+
+	esw = mdev->priv.eswitch;
+	mlx5_eswitch_unregister_vport_reps(esw, REP_IB);
+	return 0;
+}
+
+static const struct auxiliary_device_id mlx5r_rep_id_table[] = {
+	{ .name = MLX5_ADEV_NAME ".rdma-rep", },
+	{},
+};
+
+MODULE_DEVICE_TABLE(auxiliary, mlx5r_rep_id_table);
+
+static struct auxiliary_driver mlx5r_rep_driver = {
+	.name = "rep",
+	.probe = mlx5r_rep_probe,
+	.remove = mlx5r_rep_remove,
+	.id_table = mlx5r_rep_id_table,
+};
+
+int mlx5r_rep_init(void)
+{
+	return auxiliary_driver_register(&mlx5r_rep_driver);
+}
+
+void mlx5r_rep_cleanup(void)
+{
+	auxiliary_driver_unregister(&mlx5r_rep_driver);
+}
diff --git a/drivers/infiniband/hw/mlx5/ib_rep.h b/drivers/infiniband/hw/mlx5/ib_rep.h
index 5b30d3fa8f8d..94bf51ddd422 100644
--- a/drivers/infiniband/hw/mlx5/ib_rep.h
+++ b/drivers/infiniband/hw/mlx5/ib_rep.h
@@ -18,8 +18,8 @@ struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
 struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw);
 struct mlx5_eswitch_rep *mlx5_ib_vport_rep(struct mlx5_eswitch *esw,
 					   u16 vport_num);
-void mlx5_ib_register_vport_reps(struct mlx5_core_dev *mdev);
-void mlx5_ib_unregister_vport_reps(struct mlx5_core_dev *mdev);
+int mlx5r_rep_init(void);
+void mlx5r_rep_cleanup(void);
 struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
 						   struct mlx5_ib_sq *sq,
 						   u16 port);
@@ -51,8 +51,8 @@ struct mlx5_eswitch_rep *mlx5_ib_vport_rep(struct mlx5_eswitch *esw,
 	return NULL;
 }

-static inline void mlx5_ib_register_vport_reps(struct mlx5_core_dev *mdev) {}
-static inline void mlx5_ib_unregister_vport_reps(struct mlx5_core_dev *mdev) {}
+static inline int mlx5r_rep_init(void) { return 0; }
+static inline void mlx5r_rep_cleanup(void) {}
 static inline
 struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
 						   struct mlx5_ib_sq *sq,
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index a6d4692cf352..7e259823c714 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -4562,8 +4562,8 @@ void __mlx5_ib_remove(struct mlx5_ib_dev *dev,
 	ib_dealloc_device(&dev->ib_dev);
 }

-void *__mlx5_ib_add(struct mlx5_ib_dev *dev,
-		    const struct mlx5_ib_profile *profile)
+int __mlx5_ib_add(struct mlx5_ib_dev *dev,
+		  const struct mlx5_ib_profile *profile)
 {
 	int err;
 	int i;
@@ -4579,13 +4579,11 @@ void *__mlx5_ib_add(struct mlx5_ib_dev *dev,
 	}

 	dev->ib_active = true;
-
-	return dev;
+	return 0;

 err_out:
 	__mlx5_ib_remove(dev, profile, i);
-
-	return NULL;
+	return -ENOMEM;
 }

 static const struct mlx5_ib_profile pf_profile = {
@@ -4708,8 +4706,11 @@ const struct mlx5_ib_profile raw_eth_profile = {
 		     NULL),
 };

-static void *mlx5_ib_add_slave_port(struct mlx5_core_dev *mdev)
+static int mlx5r_mp_probe(struct auxiliary_device *adev,
+			  const struct auxiliary_device_id *id)
 {
+	struct mlx5_adev *idev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5_core_dev *mdev = idev->mdev;
 	struct mlx5_ib_multiport_info *mpi;
 	struct mlx5_ib_dev *dev;
 	bool bound = false;
@@ -4717,15 +4718,14 @@ static void *mlx5_ib_add_slave_port(struct mlx5_core_dev *mdev)

 	mpi = kzalloc(sizeof(*mpi), GFP_KERNEL);
 	if (!mpi)
-		return NULL;
+		return -ENOMEM;

 	mpi->mdev = mdev;
-
 	err = mlx5_query_nic_vport_system_image_guid(mdev,
 						     &mpi->sys_image_guid);
 	if (err) {
 		kfree(mpi);
-		return NULL;
+		return err;
 	}

 	mutex_lock(&mlx5_ib_multiport_mutex);
@@ -4746,40 +4746,47 @@ static void *mlx5_ib_add_slave_port(struct mlx5_core_dev *mdev)
 	}
 	mutex_unlock(&mlx5_ib_multiport_mutex);

-	return mpi;
+	dev_set_drvdata(&adev->dev, mpi);
+	return 0;
 }

-static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
+static int mlx5r_mp_remove(struct auxiliary_device *adev)
 {
+	struct mlx5_ib_multiport_info *mpi;
+
+	mpi = dev_get_drvdata(&adev->dev);
+	mutex_lock(&mlx5_ib_multiport_mutex);
+	if (mpi->ibdev)
+		mlx5_ib_unbind_slave_port(mpi->ibdev, mpi);
+	list_del(&mpi->list);
+	mutex_unlock(&mlx5_ib_multiport_mutex);
+	kfree(mpi);
+	return 0;
+}
+
+static int mlx5r_probe(struct auxiliary_device *adev,
+		       const struct auxiliary_device_id *id)
+{
+	struct mlx5_adev *idev = container_of(adev, struct mlx5_adev, adev);
+	struct mlx5_core_dev *mdev = idev->mdev;
 	const struct mlx5_ib_profile *profile;
+	int port_type_cap, num_ports, ret;
 	enum rdma_link_layer ll;
 	struct mlx5_ib_dev *dev;
-	int port_type_cap;
-	int num_ports;
-
-	if (MLX5_ESWITCH_MANAGER(mdev) &&
-	    mlx5_ib_eswitch_mode(mdev->priv.eswitch) == MLX5_ESWITCH_OFFLOADS) {
-		if (!mlx5_core_mp_enabled(mdev))
-			mlx5_ib_register_vport_reps(mdev);
-		return mdev;
-	}

 	port_type_cap = MLX5_CAP_GEN(mdev, port_type);
 	ll = mlx5_port_type_cap_to_rdma_ll(port_type_cap);

-	if (mlx5_core_is_mp_slave(mdev) && ll == IB_LINK_LAYER_ETHERNET)
-		return mlx5_ib_add_slave_port(mdev);
-
 	num_ports = max(MLX5_CAP_GEN(mdev, num_ports),
 			MLX5_CAP_GEN(mdev, num_vhca_ports));
 	dev = ib_alloc_device(mlx5_ib_dev, ib_dev);
 	if (!dev)
-		return NULL;
+		return -ENOMEM;
 	dev->port = kcalloc(num_ports, sizeof(*dev->port),
 			     GFP_KERNEL);
 	if (!dev->port) {
 		ib_dealloc_device(&dev->ib_dev);
-		return NULL;
+		return -ENOMEM;
 	}

 	dev->mdev = mdev;
@@ -4790,43 +4797,56 @@ static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 	else
 		profile = &pf_profile;

-	return __mlx5_ib_add(dev, profile);
+	ret = __mlx5_ib_add(dev, profile);
+	if (ret) {
+		kfree(dev->port);
+		ib_dealloc_device(&dev->ib_dev);
+		return ret;
+	}
+
+	dev_set_drvdata(&adev->dev, dev);
+	return 0;
 }

-static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)
+static int mlx5r_remove(struct auxiliary_device *adev)
 {
-	struct mlx5_ib_multiport_info *mpi;
 	struct mlx5_ib_dev *dev;

-	if (MLX5_ESWITCH_MANAGER(mdev) && context == mdev) {
-		mlx5_ib_unregister_vport_reps(mdev);
-		return;
-	}
-
-	if (mlx5_core_is_mp_slave(mdev)) {
-		mpi = context;
-		mutex_lock(&mlx5_ib_multiport_mutex);
-		if (mpi->ibdev)
-			mlx5_ib_unbind_slave_port(mpi->ibdev, mpi);
-		list_del(&mpi->list);
-		mutex_unlock(&mlx5_ib_multiport_mutex);
-		kfree(mpi);
-		return;
-	}
-
-	dev = context;
+	dev = dev_get_drvdata(&adev->dev);
 	__mlx5_ib_remove(dev, dev->profile, MLX5_IB_STAGE_MAX);
+	return 0;
 }

-static struct mlx5_interface mlx5_ib_interface = {
-	.add            = mlx5_ib_add,
-	.remove         = mlx5_ib_remove,
-	.protocol	= MLX5_INTERFACE_PROTOCOL_IB,
+static const struct auxiliary_device_id mlx5r_mp_id_table[] = {
+	{ .name = MLX5_ADEV_NAME ".multiport", },
+	{},
+};
+
+static const struct auxiliary_device_id mlx5r_id_table[] = {
+	{ .name = MLX5_ADEV_NAME ".rdma", },
+	{},
+};
+
+MODULE_DEVICE_TABLE(auxiliary, mlx5r_mp_id_table);
+MODULE_DEVICE_TABLE(auxiliary, mlx5r_id_table);
+
+static struct auxiliary_driver mlx5r_mp_driver = {
+	.name = "multiport",
+	.probe = mlx5r_mp_probe,
+	.remove = mlx5r_mp_remove,
+	.id_table = mlx5r_mp_id_table,
+};
+
+static struct auxiliary_driver mlx5r_driver = {
+	.name = "rdma",
+	.probe = mlx5r_probe,
+	.remove = mlx5r_remove,
+	.id_table = mlx5r_id_table,
 };

 static int __init mlx5_ib_init(void)
 {
-	int err;
+	int ret;

 	xlt_emergency_page = (void *)__get_free_page(GFP_KERNEL);
 	if (!xlt_emergency_page)
@@ -4839,15 +4859,33 @@ static int __init mlx5_ib_init(void)
 	}

 	mlx5_ib_odp_init();
+	ret = mlx5r_rep_init();
+	if (ret)
+		goto rep_err;
+	ret = auxiliary_driver_register(&mlx5r_mp_driver);
+	if (ret)
+		goto mp_err;
+	ret = auxiliary_driver_register(&mlx5r_driver);
+	if (ret)
+		goto drv_err;
+	return 0;

-	err = mlx5_register_interface(&mlx5_ib_interface);
-
-	return err;
+drv_err:
+	auxiliary_driver_unregister(&mlx5r_mp_driver);
+mp_err:
+	mlx5r_rep_cleanup();
+rep_err:
+	destroy_workqueue(mlx5_ib_event_wq);
+	free_page((unsigned long)xlt_emergency_page);
+	return ret;
 }

 static void __exit mlx5_ib_cleanup(void)
 {
-	mlx5_unregister_interface(&mlx5_ib_interface);
+	auxiliary_driver_unregister(&mlx5r_driver);
+	auxiliary_driver_unregister(&mlx5r_mp_driver);
+	mlx5r_rep_cleanup();
+
 	destroy_workqueue(mlx5_ib_event_wq);
 	free_page((unsigned long)xlt_emergency_page);
 }
diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h
index 254668d31988..200df80393de 100644
--- a/drivers/infiniband/hw/mlx5/mlx5_ib.h
+++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h
@@ -1383,8 +1383,8 @@ extern const struct mmu_interval_notifier_ops mlx5_mn_ops;
 void __mlx5_ib_remove(struct mlx5_ib_dev *dev,
 		      const struct mlx5_ib_profile *profile,
 		      int stage);
-void *__mlx5_ib_add(struct mlx5_ib_dev *dev,
-		    const struct mlx5_ib_profile *profile);
+int __mlx5_ib_add(struct mlx5_ib_dev *dev,
+		  const struct mlx5_ib_profile *profile);

 int mlx5_ib_get_vf_config(struct ib_device *device, int vf,
 			  u8 port, struct ifla_vf_info *info);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
index 6d413c622785..843a8579d8c8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
@@ -144,16 +144,82 @@ static bool is_vnet_supported(struct mlx5_core_dev *dev)
 	return true;
 }

+static bool is_ib_rep_supported(struct mlx5_core_dev *dev)
+{
+	if (!IS_ENABLED(CONFIG_MLX5_INFINIBAND))
+		return false;
+
+	if (dev->priv.flags & MLX5_PRIV_FLAGS_DISABLE_IB_ADEV)
+		return false;
+
+	if (!is_eth_rep_supported(dev))
+		return false;
+
+	if (!MLX5_ESWITCH_MANAGER(dev))
+		return false;
+
+	if (mlx5_eswitch_mode(dev->priv.eswitch) != MLX5_ESWITCH_OFFLOADS)
+		return false;
+
+	if (mlx5_core_mp_enabled(dev))
+		return false;
+
+	return true;
+}
+
+static bool is_mp_supported(struct mlx5_core_dev *dev)
+{
+	if (!IS_ENABLED(CONFIG_MLX5_INFINIBAND))
+		return false;
+
+	if (dev->priv.flags & MLX5_PRIV_FLAGS_DISABLE_IB_ADEV)
+		return false;
+
+	if (is_ib_rep_supported(dev))
+		return false;
+
+	if (MLX5_CAP_GEN(dev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+		return false;
+
+	if (!mlx5_core_is_mp_slave(dev))
+		return false;
+
+	return true;
+}
+
+static bool is_ib_supported(struct mlx5_core_dev *dev)
+{
+	if (!IS_ENABLED(CONFIG_MLX5_INFINIBAND))
+		return false;
+
+	if (dev->priv.flags & MLX5_PRIV_FLAGS_DISABLE_IB_ADEV)
+		return false;
+
+	if (is_ib_rep_supported(dev))
+		return false;
+
+	if (is_mp_supported(dev))
+		return false;
+
+	return true;
+}
+
 static const struct mlx5_adev_device {
 	const char *suffix;
 	bool (*is_supported)(struct mlx5_core_dev *dev);
 } mlx5_adev_devices[] = {
 	[MLX5_INTERFACE_PROTOCOL_VDPA] = { .suffix = "vnet",
 					   .is_supported = &is_vnet_supported },
+	[MLX5_INTERFACE_PROTOCOL_IB] = { .suffix = "rdma",
+					 .is_supported = &is_ib_supported },
 	[MLX5_INTERFACE_PROTOCOL_ETH] = { .suffix = "eth",
 					  .is_supported = &is_eth_supported },
 	[MLX5_INTERFACE_PROTOCOL_ETH_REP] = { .suffix = "eth-rep",
 					   .is_supported = &is_eth_rep_supported },
+	[MLX5_INTERFACE_PROTOCOL_IB_REP] = { .suffix = "rdma-rep",
+					   .is_supported = &is_ib_rep_supported },
+	[MLX5_INTERFACE_PROTOCOL_MPIB] = { .suffix = "multiport",
+					   .is_supported = &is_mp_supported },
 };

 int mlx5_adev_idx_alloc(void)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 78a854926b00..b652b4bde733 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1617,7 +1617,6 @@ int mlx5_eswitch_enable_locked(struct mlx5_eswitch *esw, int mode, int num_vfs)
 		err = esw_legacy_enable(esw);
 	} else {
 		mlx5_rescan_drivers(esw->dev);
-		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
 		err = esw_offloads_enable(esw);
 	}

@@ -1635,10 +1634,9 @@ int mlx5_eswitch_enable_locked(struct mlx5_eswitch *esw, int mode, int num_vfs)
 abort:
 	esw->mode = MLX5_ESWITCH_NONE;

-	if (mode == MLX5_ESWITCH_OFFLOADS) {
-		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
+	if (mode == MLX5_ESWITCH_OFFLOADS)
 		mlx5_rescan_drivers(esw->dev);
-	}
+
 	esw_destroy_tsar(esw);
 	return err;
 }
@@ -1699,10 +1697,9 @@ void mlx5_eswitch_disable_locked(struct mlx5_eswitch *esw, bool clear_vf)

 	mlx5_lag_update(esw->dev);

-	if (old_mode == MLX5_ESWITCH_OFFLOADS) {
-		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
+	if (old_mode == MLX5_ESWITCH_OFFLOADS)
 		mlx5_rescan_drivers(esw->dev);
-	}
+
 	esw_destroy_tsar(esw);

 	if (clear_vf)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
index e4d4de1719bd..325f32b9525c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -243,24 +243,30 @@ static bool mlx5_lag_check_prereq(struct mlx5_lag *ldev)
 #endif
 }

-static void mlx5_lag_add_ib_devices(struct mlx5_lag *ldev)
+static void mlx5_lag_add_devices(struct mlx5_lag *ldev)
 {
 	int i;

-	for (i = 0; i < MLX5_MAX_PORTS; i++)
-		if (ldev->pf[i].dev)
-			mlx5_add_dev_by_protocol(ldev->pf[i].dev,
-						 MLX5_INTERFACE_PROTOCOL_IB);
+	for (i = 0; i < MLX5_MAX_PORTS; i++) {
+		if (!ldev->pf[i].dev)
+			continue;
+
+		ldev->pf[i].dev->priv.flags &= ~MLX5_PRIV_FLAGS_DISABLE_IB_ADEV;
+		mlx5_rescan_drivers_locked(ldev->pf[i].dev);
+	}
 }

-static void mlx5_lag_remove_ib_devices(struct mlx5_lag *ldev)
+static void mlx5_lag_remove_devices(struct mlx5_lag *ldev)
 {
 	int i;

-	for (i = 0; i < MLX5_MAX_PORTS; i++)
-		if (ldev->pf[i].dev)
-			mlx5_remove_dev_by_protocol(ldev->pf[i].dev,
-						    MLX5_INTERFACE_PROTOCOL_IB);
+	for (i = 0; i < MLX5_MAX_PORTS; i++) {
+		if (!ldev->pf[i].dev)
+			continue;
+
+		ldev->pf[i].dev->priv.flags |= MLX5_PRIV_FLAGS_DISABLE_IB_ADEV;
+		mlx5_rescan_drivers_locked(ldev->pf[i].dev);
+	}
 }

 static void mlx5_do_bond(struct mlx5_lag *ldev)
@@ -290,20 +296,21 @@ static void mlx5_do_bond(struct mlx5_lag *ldev)
 #endif

 		if (roce_lag)
-			mlx5_lag_remove_ib_devices(ldev);
+			mlx5_lag_remove_devices(ldev);

 		err = mlx5_activate_lag(ldev, &tracker,
 					roce_lag ? MLX5_LAG_FLAG_ROCE :
 					MLX5_LAG_FLAG_SRIOV);
 		if (err) {
 			if (roce_lag)
-				mlx5_lag_add_ib_devices(ldev);
+				mlx5_lag_add_devices(ldev);

 			return;
 		}

 		if (roce_lag) {
-			mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB);
+			dev0->priv.flags &= ~MLX5_PRIV_FLAGS_DISABLE_IB_ADEV;
+			mlx5_rescan_drivers_locked(dev0);
 			mlx5_nic_vport_enable_roce(dev1);
 		}
 	} else if (do_bond && __mlx5_lag_is_active(ldev)) {
@@ -312,7 +319,8 @@ static void mlx5_do_bond(struct mlx5_lag *ldev)
 		roce_lag = __mlx5_lag_is_roce(ldev);

 		if (roce_lag) {
-			mlx5_remove_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB);
+			dev0->priv.flags |= MLX5_PRIV_FLAGS_DISABLE_IB_ADEV;
+			mlx5_rescan_drivers_locked(dev0);
 			mlx5_nic_vport_disable_roce(dev1);
 		}

@@ -321,7 +329,7 @@ static void mlx5_do_bond(struct mlx5_lag *ldev)
 			return;

 		if (roce_lag)
-			mlx5_lag_add_ib_devices(ldev);
+			mlx5_lag_add_devices(ldev);
 	}
 }

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 0b4793aff274..8f5556e49b00 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -1348,7 +1348,6 @@ static void mlx5_mdev_uninit(struct mlx5_core_dev *dev)
 	mutex_destroy(&dev->intf_state_mutex);
 }

-#define MLX5_IB_MOD "mlx5_ib"
 static int init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 {
 	struct mlx5_core_dev *dev;
@@ -1390,8 +1389,6 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 		goto err_load_one;
 	}

-	request_module_nowait(MLX5_IB_MOD);
-
 	err = mlx5_crdump_enable(dev);
 	if (err)
 		dev_err(&pdev->dev, "mlx5_crdump_enable failed with error code %d\n", err);
--
2.28.0

[PATCH mlx5-next v1 09/11] net/mlx5: Delete custom device management logic

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

After conversion to use auxiliary bus, all custom device management is
not needed anymore, delete it.

Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/dev.c | 288 ++----------------
 drivers/net/ethernet/mellanox/mlx5/core/lag.c |  18 --
 .../ethernet/mellanox/mlx5/core/mlx5_core.h   |   8 -
 include/linux/mlx5/driver.h                   |  22 --
 4 files changed, 18 insertions(+), 318 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
index 843a8579d8c8..3a81c2f1971b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
@@ -35,24 +35,10 @@
 #include <linux/mlx5/mlx5_ifc_vdpa.h>
 #include "mlx5_core.h"

-static LIST_HEAD(intf_list);
-static LIST_HEAD(mlx5_dev_list);
 /* intf dev list mutex */
 static DEFINE_MUTEX(mlx5_intf_mutex);
 static DEFINE_IDA(mlx5_adev_ida);

-struct mlx5_device_context {
-	struct list_head	list;
-	struct mlx5_interface  *intf;
-	void		       *context;
-	unsigned long		state;
-};
-
-enum {
-	MLX5_INTERFACE_ADDED,
-	MLX5_INTERFACE_ATTACHED,
-};
-
 static bool is_eth_rep_supported(struct mlx5_core_dev *dev)
 {
 	if (!IS_ENABLED(CONFIG_MLX5_ESWITCH))
@@ -204,11 +190,22 @@ static bool is_ib_supported(struct mlx5_core_dev *dev)
 	return true;
 }

+enum {
+	MLX5_INTERFACE_PROTOCOL_ETH_REP,
+	MLX5_INTERFACE_PROTOCOL_ETH,
+
+	MLX5_INTERFACE_PROTOCOL_IB_REP,
+	MLX5_INTERFACE_PROTOCOL_MPIB,
+	MLX5_INTERFACE_PROTOCOL_IB,
+
+	MLX5_INTERFACE_PROTOCOL_VNET,
+};
+
 static const struct mlx5_adev_device {
 	const char *suffix;
 	bool (*is_supported)(struct mlx5_core_dev *dev);
 } mlx5_adev_devices[] = {
-	[MLX5_INTERFACE_PROTOCOL_VDPA] = { .suffix = "vnet",
+	[MLX5_INTERFACE_PROTOCOL_VNET] = { .suffix = "vnet",
 					   .is_supported = &is_vnet_supported },
 	[MLX5_INTERFACE_PROTOCOL_IB] = { .suffix = "rdma",
 					 .is_supported = &is_ib_supported },
@@ -251,90 +248,6 @@ void mlx5_adev_cleanup(struct mlx5_core_dev *dev)
 	kfree(priv->adev);
 }

-void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
-{
-	struct mlx5_device_context *dev_ctx;
-	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
-
-	if (!mlx5_lag_intf_add(intf, priv))
-		return;
-
-	dev_ctx = kzalloc(sizeof(*dev_ctx), GFP_KERNEL);
-	if (!dev_ctx)
-		return;
-
-	dev_ctx->intf = intf;
-
-	dev_ctx->context = intf->add(dev);
-	if (dev_ctx->context) {
-		set_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state);
-		if (intf->attach)
-			set_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state);
-
-		spin_lock_irq(&priv->ctx_lock);
-		list_add_tail(&dev_ctx->list, &priv->ctx_list);
-		spin_unlock_irq(&priv->ctx_lock);
-	}
-
-	if (!dev_ctx->context)
-		kfree(dev_ctx);
-}
-
-static struct mlx5_device_context *mlx5_get_device(struct mlx5_interface *intf,
-						   struct mlx5_priv *priv)
-{
-	struct mlx5_device_context *dev_ctx;
-
-	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
-		if (dev_ctx->intf == intf)
-			return dev_ctx;
-	return NULL;
-}
-
-void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
-{
-	struct mlx5_device_context *dev_ctx;
-	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
-
-	dev_ctx = mlx5_get_device(intf, priv);
-	if (!dev_ctx)
-		return;
-
-	spin_lock_irq(&priv->ctx_lock);
-	list_del(&dev_ctx->list);
-	spin_unlock_irq(&priv->ctx_lock);
-
-	if (test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state))
-		intf->remove(dev, dev_ctx->context);
-
-	kfree(dev_ctx);
-}
-
-static void mlx5_attach_interface(struct mlx5_interface *intf, struct mlx5_priv *priv)
-{
-	struct mlx5_device_context *dev_ctx;
-	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
-
-	dev_ctx = mlx5_get_device(intf, priv);
-	if (!dev_ctx)
-		return;
-
-	if (intf->attach) {
-		if (test_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state))
-			return;
-		if (intf->attach(dev, dev_ctx->context))
-			return;
-		set_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state);
-	} else {
-		if (test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state))
-			return;
-		dev_ctx->context = intf->add(dev);
-		if (!dev_ctx->context)
-			return;
-		set_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state);
-	}
-}
-
 static void adev_release(struct device *dev)
 {
 	struct mlx5_adev *mlx5_adev =
@@ -390,7 +303,6 @@ int mlx5_attach_device(struct mlx5_core_dev *dev)
 	struct mlx5_priv *priv = &dev->priv;
 	struct auxiliary_device *adev;
 	struct auxiliary_driver *adrv;
-	struct mlx5_interface *intf;
 	int ret = 0, i;

 	mutex_lock(&mlx5_intf_mutex);
@@ -423,41 +335,15 @@ int mlx5_attach_device(struct mlx5_core_dev *dev)
 			break;
 		}
 	}
-
-	list_for_each_entry(intf, &intf_list, list)
-		mlx5_attach_interface(intf, priv);
 	mutex_unlock(&mlx5_intf_mutex);
 	return ret;
 }

-static void mlx5_detach_interface(struct mlx5_interface *intf, struct mlx5_priv *priv)
-{
-	struct mlx5_device_context *dev_ctx;
-	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
-
-	dev_ctx = mlx5_get_device(intf, priv);
-	if (!dev_ctx)
-		return;
-
-	if (intf->detach) {
-		if (!test_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state))
-			return;
-		intf->detach(dev, dev_ctx->context);
-		clear_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state);
-	} else {
-		if (!test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state))
-			return;
-		intf->remove(dev, dev_ctx->context);
-		clear_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state);
-	}
-}
-
 void mlx5_detach_device(struct mlx5_core_dev *dev)
 {
 	struct mlx5_priv *priv = &dev->priv;
 	struct auxiliary_device *adev;
 	struct auxiliary_driver *adrv;
-	struct mlx5_interface *intf;
 	pm_message_t pm = {};
 	int i;

@@ -477,30 +363,11 @@ void mlx5_detach_device(struct mlx5_core_dev *dev)
 		del_adev(&priv->adev[i]->adev);
 		priv->adev[i] = NULL;
 	}
-
-	list_for_each_entry(intf, &intf_list, list)
-		mlx5_detach_interface(intf, priv);
 	mutex_unlock(&mlx5_intf_mutex);
 }

-bool mlx5_device_registered(struct mlx5_core_dev *dev)
-{
-	struct mlx5_priv *priv;
-	bool found = false;
-
-	mutex_lock(&mlx5_intf_mutex);
-	list_for_each_entry(priv, &mlx5_dev_list, dev_list)
-		if (priv == &dev->priv)
-			found = true;
-	mutex_unlock(&mlx5_intf_mutex);
-
-	return found;
-}
-
 int mlx5_register_device(struct mlx5_core_dev *dev)
 {
-	struct mlx5_priv *priv = &dev->priv;
-	struct mlx5_interface *intf;
 	int ret;

 	mutex_lock(&mlx5_intf_mutex);
@@ -508,65 +375,19 @@ int mlx5_register_device(struct mlx5_core_dev *dev)
 	ret = mlx5_rescan_drivers_locked(dev);
 	mutex_unlock(&mlx5_intf_mutex);
 	if (ret)
-		goto add_err;
-
-	mutex_lock(&mlx5_intf_mutex);
-	list_add_tail(&priv->dev_list, &mlx5_dev_list);
-	list_for_each_entry(intf, &intf_list, list)
-		mlx5_add_device(intf, priv);
-	mutex_unlock(&mlx5_intf_mutex);
-
-	return 0;
+		mlx5_unregister_device(dev);

-add_err:
-	mlx5_unregister_device(dev);
 	return ret;
 }

 void mlx5_unregister_device(struct mlx5_core_dev *dev)
 {
-	struct mlx5_priv *priv = &dev->priv;
-	struct mlx5_interface *intf;
-
 	mutex_lock(&mlx5_intf_mutex);
-	list_for_each_entry_reverse(intf, &intf_list, list)
-		mlx5_remove_device(intf, priv);
-	list_del(&priv->dev_list);
-
 	dev->priv.flags |= MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV;
 	mlx5_rescan_drivers_locked(dev);
 	mutex_unlock(&mlx5_intf_mutex);
 }

-int mlx5_register_interface(struct mlx5_interface *intf)
-{
-	struct mlx5_priv *priv;
-
-	if (!intf->add || !intf->remove)
-		return -EINVAL;
-
-	mutex_lock(&mlx5_intf_mutex);
-	list_add_tail(&intf->list, &intf_list);
-	list_for_each_entry(priv, &mlx5_dev_list, dev_list)
-		mlx5_add_device(intf, priv);
-	mutex_unlock(&mlx5_intf_mutex);
-
-	return 0;
-}
-EXPORT_SYMBOL(mlx5_register_interface);
-
-void mlx5_unregister_interface(struct mlx5_interface *intf)
-{
-	struct mlx5_priv *priv;
-
-	mutex_lock(&mlx5_intf_mutex);
-	list_for_each_entry(priv, &mlx5_dev_list, dev_list)
-		mlx5_remove_device(intf, priv);
-	list_del(&intf->list);
-	mutex_unlock(&mlx5_intf_mutex);
-}
-EXPORT_SYMBOL(mlx5_unregister_interface);
-
 static int add_drivers(struct mlx5_core_dev *dev)
 {
 	struct mlx5_priv *priv = &dev->priv;
@@ -638,59 +459,6 @@ int mlx5_rescan_drivers_locked(struct mlx5_core_dev *dev)
 	return add_drivers(dev);
 }

-/* Must be called with intf_mutex held */
-static bool mlx5_has_added_dev_by_protocol(struct mlx5_core_dev *mdev, int protocol)
-{
-	struct mlx5_device_context *dev_ctx;
-	struct mlx5_interface *intf;
-	bool found = false;
-
-	list_for_each_entry(intf, &intf_list, list) {
-		if (intf->protocol == protocol) {
-			dev_ctx = mlx5_get_device(intf, &mdev->priv);
-			if (dev_ctx && test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state))
-				found = true;
-			break;
-		}
-	}
-
-	return found;
-}
-
-void mlx5_reload_interface(struct mlx5_core_dev *mdev, int protocol)
-{
-	mutex_lock(&mlx5_intf_mutex);
-	if (mlx5_has_added_dev_by_protocol(mdev, protocol)) {
-		mlx5_remove_dev_by_protocol(mdev, protocol);
-		mlx5_add_dev_by_protocol(mdev, protocol);
-	}
-	mutex_unlock(&mlx5_intf_mutex);
-}
-
-/* Must be called with intf_mutex held */
-void mlx5_add_dev_by_protocol(struct mlx5_core_dev *dev, int protocol)
-{
-	struct mlx5_interface *intf;
-
-	list_for_each_entry(intf, &intf_list, list)
-		if (intf->protocol == protocol) {
-			mlx5_add_device(intf, &dev->priv);
-			break;
-		}
-}
-
-/* Must be called with intf_mutex held */
-void mlx5_remove_dev_by_protocol(struct mlx5_core_dev *dev, int protocol)
-{
-	struct mlx5_interface *intf;
-
-	list_for_each_entry(intf, &intf_list, list)
-		if (intf->protocol == protocol) {
-			mlx5_remove_device(intf, &dev->priv);
-			break;
-		}
-}
-
 static u32 mlx5_gen_pci_id(const struct mlx5_core_dev *dev)
 {
 	return (u32)((pci_domain_nr(dev->pdev->bus) << 16) |
@@ -722,45 +490,25 @@ static int next_phys_dev(struct device *dev, const void *data)
  */
 struct mlx5_core_dev *mlx5_get_next_phys_dev(struct mlx5_core_dev *dev)
 {
-	struct mlx5_core_dev *res = NULL;
-	struct mlx5_core_dev *tmp_dev;
 	struct auxiliary_device *adev;
 	struct mlx5_adev *madev;
-	struct mlx5_priv *priv;
-	u32 pci_id;

 	if (!mlx5_core_is_pf(dev))
 		return NULL;

 	adev = auxiliary_find_device(NULL, dev, &next_phys_dev);
-	if (adev) {
-		madev = container_of(adev, struct mlx5_adev, adev);
-
-		put_device(&adev->dev);
-		return madev->mdev;
-	}
-
-	pci_id = mlx5_gen_pci_id(dev);
-	list_for_each_entry(priv, &mlx5_dev_list, dev_list) {
-		tmp_dev = container_of(priv, struct mlx5_core_dev, priv);
-		if (!mlx5_core_is_pf(tmp_dev))
-			continue;
-
-		if ((dev != tmp_dev) && (mlx5_gen_pci_id(tmp_dev) == pci_id)) {
-			res = tmp_dev;
-			break;
-		}
-	}
+	if (!adev)
+		return NULL;

-	return res;
+	madev = container_of(adev, struct mlx5_adev, adev);
+	put_device(&adev->dev);
+	return madev->mdev;
 }

-
 void mlx5_dev_list_lock(void)
 {
 	mutex_lock(&mlx5_intf_mutex);
 }
-
 void mlx5_dev_list_unlock(void)
 {
 	mutex_unlock(&mlx5_intf_mutex);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
index 325f32b9525c..f3d45ef082cd 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -749,24 +749,6 @@ u8 mlx5_lag_get_slave_port(struct mlx5_core_dev *dev,
 }
 EXPORT_SYMBOL(mlx5_lag_get_slave_port);

-bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv)
-{
-	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev,
-						 priv);
-	struct mlx5_lag *ldev;
-
-	if (intf->protocol != MLX5_INTERFACE_PROTOCOL_IB)
-		return true;
-
-	ldev = mlx5_lag_dev_get(dev);
-	if (!ldev || !__mlx5_lag_is_roce(ldev) ||
-	    ldev->pf[MLX5_LAG_P1].dev == dev)
-		return true;
-
-	/* If bonded, we do not add an IB device for PF1. */
-	return false;
-}
-
 int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev,
 				 u64 *values,
 				 int num_counters,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index f7e44e04d465..dd7312621d0d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -183,22 +183,15 @@ void mlx5_adev_idx_free(int idx);
 void mlx5_adev_cleanup(struct mlx5_core_dev *dev);
 int mlx5_adev_init(struct mlx5_core_dev *dev);

-void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv);
-void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv);
 int mlx5_attach_device(struct mlx5_core_dev *dev);
 void mlx5_detach_device(struct mlx5_core_dev *dev);
-bool mlx5_device_registered(struct mlx5_core_dev *dev);
 int mlx5_register_device(struct mlx5_core_dev *dev);
 void mlx5_unregister_device(struct mlx5_core_dev *dev);
-void mlx5_add_dev_by_protocol(struct mlx5_core_dev *dev, int protocol);
-void mlx5_remove_dev_by_protocol(struct mlx5_core_dev *dev, int protocol);
 struct mlx5_core_dev *mlx5_get_next_phys_dev(struct mlx5_core_dev *dev);
 void mlx5_dev_list_lock(void);
 void mlx5_dev_list_unlock(void);
 int mlx5_dev_list_trylock(void);

-bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv);
-
 int mlx5_query_mtpps(struct mlx5_core_dev *dev, u32 *mtpps, u32 mtpps_size);
 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);
@@ -248,7 +241,6 @@ static inline int mlx5_rescan_drivers(struct mlx5_core_dev *dev)
 	return ret;
 }

-void mlx5_reload_interface(struct mlx5_core_dev *mdev, int protocol);
 void mlx5_lag_update(struct mlx5_core_dev *dev);

 enum {
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index f35ee4855777..e616e3a656ca 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -1077,28 +1077,6 @@ enum {
 	MAX_MR_CACHE_ENTRIES
 };

-enum {
-	MLX5_INTERFACE_PROTOCOL_ETH_REP,
-	MLX5_INTERFACE_PROTOCOL_ETH,
-
-	MLX5_INTERFACE_PROTOCOL_IB_REP,
-	MLX5_INTERFACE_PROTOCOL_MPIB,
-	MLX5_INTERFACE_PROTOCOL_IB,
-
-	MLX5_INTERFACE_PROTOCOL_VDPA,
-};
-
-struct mlx5_interface {
-	void *			(*add)(struct mlx5_core_dev *dev);
-	void			(*remove)(struct mlx5_core_dev *dev, void *context);
-	int			(*attach)(struct mlx5_core_dev *dev, void *context);
-	void			(*detach)(struct mlx5_core_dev *dev, void *context);
-	int			protocol;
-	struct list_head	list;
-};
-
-int mlx5_register_interface(struct mlx5_interface *intf);
-void mlx5_unregister_interface(struct mlx5_interface *intf);
 int mlx5_notifier_register(struct mlx5_core_dev *dev, struct notifier_block *nb);
 int mlx5_notifier_unregister(struct mlx5_core_dev *dev, struct notifier_block *nb);
 int mlx5_eq_notifier_register(struct mlx5_core_dev *dev, struct mlx5_nb *nb);
--
2.28.0

[PATCH mlx5-next v1 10/11] net/mlx5: Simplify eswitch mode check

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Provide mlx5_core device instead of "priv" pointer while checking
eswith mode.

Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/mlx5/counters.c             | 7 -------
 drivers/infiniband/hw/mlx5/ib_rep.c               | 5 -----
 drivers/infiniband/hw/mlx5/ib_rep.h               | 6 ------
 drivers/net/ethernet/mellanox/mlx5/core/dev.c     | 4 ++--
 drivers/net/ethernet/mellanox/mlx5/core/devlink.c | 2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c   | 8 +++-----
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 4 +++-
 include/linux/mlx5/eswitch.h                      | 8 ++++++--
 9 files changed, 16 insertions(+), 30 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/counters.c b/drivers/infiniband/hw/mlx5/counters.c
index 70c8fd67ee2f..084652e2b15a 100644
--- a/drivers/infiniband/hw/mlx5/counters.c
+++ b/drivers/infiniband/hw/mlx5/counters.c
@@ -138,13 +138,6 @@ static int mlx5_ib_create_counters(struct ib_counters *counters,
 }


-static bool is_mdev_switchdev_mode(const struct mlx5_core_dev *mdev)
-{
-	return MLX5_ESWITCH_MANAGER(mdev) &&
-	       mlx5_ib_eswitch_mode(mdev->priv.eswitch) ==
-		       MLX5_ESWITCH_OFFLOADS;
-}
-
 static const struct mlx5_ib_counters *get_counters(struct mlx5_ib_dev *dev,
 						   u8 port_num)
 {
diff --git a/drivers/infiniband/hw/mlx5/ib_rep.c b/drivers/infiniband/hw/mlx5/ib_rep.c
index 0dc15757cc66..9810bdd7f3bc 100644
--- a/drivers/infiniband/hw/mlx5/ib_rep.c
+++ b/drivers/infiniband/hw/mlx5/ib_rep.c
@@ -102,11 +102,6 @@ static const struct mlx5_eswitch_rep_ops rep_ops = {
 	.get_proto_dev = mlx5_ib_vport_get_proto_dev,
 };

-u8 mlx5_ib_eswitch_mode(struct mlx5_eswitch *esw)
-{
-	return mlx5_eswitch_mode(esw);
-}
-
 struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
 					  u16 vport_num)
 {
diff --git a/drivers/infiniband/hw/mlx5/ib_rep.h b/drivers/infiniband/hw/mlx5/ib_rep.h
index 94bf51ddd422..93f562735e89 100644
--- a/drivers/infiniband/hw/mlx5/ib_rep.h
+++ b/drivers/infiniband/hw/mlx5/ib_rep.h
@@ -12,7 +12,6 @@
 extern const struct mlx5_ib_profile raw_eth_profile;

 #ifdef CONFIG_MLX5_ESWITCH
-u8 mlx5_ib_eswitch_mode(struct mlx5_eswitch *esw);
 struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
 					  u16 vport_num);
 struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw);
@@ -26,11 +25,6 @@ struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
 struct net_device *mlx5_ib_get_rep_netdev(struct mlx5_eswitch *esw,
 					  u16 vport_num);
 #else /* CONFIG_MLX5_ESWITCH */
-static inline u8 mlx5_ib_eswitch_mode(struct mlx5_eswitch *esw)
-{
-	return MLX5_ESWITCH_NONE;
-}
-
 static inline
 struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
 					  u16 vport_num)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
index 3a81c2f1971b..b051417ede67 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
@@ -47,7 +47,7 @@ static bool is_eth_rep_supported(struct mlx5_core_dev *dev)
 	if (!MLX5_ESWITCH_MANAGER(dev))
 		return false;

-	if (mlx5_eswitch_mode(dev->priv.eswitch) != MLX5_ESWITCH_OFFLOADS)
+	if (!is_mdev_switchdev_mode(dev))
 		return false;

 	return true;
@@ -144,7 +144,7 @@ static bool is_ib_rep_supported(struct mlx5_core_dev *dev)
 	if (!MLX5_ESWITCH_MANAGER(dev))
 		return false;

-	if (mlx5_eswitch_mode(dev->priv.eswitch) != MLX5_ESWITCH_OFFLOADS)
+	if (!is_mdev_switchdev_mode(dev))
 		return false;

 	if (mlx5_core_mp_enabled(dev))
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
index 1a351e2f6ace..aeffb6b135ee 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
@@ -221,7 +221,7 @@ static int mlx5_devlink_fs_mode_validate(struct devlink *devlink, u32 id,
 		u8 eswitch_mode;
 		bool smfs_cap;

-		eswitch_mode = mlx5_eswitch_mode(dev->priv.eswitch);
+		eswitch_mode = mlx5_eswitch_mode(dev);
 		smfs_cap = mlx5_fs_dr_is_supported(dev);

 		if (!smfs_cap) {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 3c4f880c6329..8d65ac888a28 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -3135,7 +3135,7 @@ static void mlx5e_modify_admin_state(struct mlx5_core_dev *mdev,

 	mlx5_set_port_admin_status(mdev, state);

-	if (!MLX5_ESWITCH_MANAGER(mdev) ||  mlx5_eswitch_mode(esw) == MLX5_ESWITCH_OFFLOADS)
+	if (mlx5_eswitch_mode(mdev) != MLX5_ESWITCH_LEGACY)
 		return;

 	if (state == MLX5_PORT_UP)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index e3a968e9e2a0..7548bab78654 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -271,8 +271,6 @@ mlx5e_tc_match_to_reg_set(struct mlx5_core_dev *mdev,
 	return 0;
 }

-#define esw_offloads_mode(esw) (mlx5_eswitch_mode(esw) == MLX5_ESWITCH_OFFLOADS)
-
 static struct mlx5_tc_ct_priv *
 get_ct_priv(struct mlx5e_priv *priv)
 {
@@ -280,7 +278,7 @@ get_ct_priv(struct mlx5e_priv *priv)
 	struct mlx5_rep_uplink_priv *uplink_priv;
 	struct mlx5e_rep_priv *uplink_rpriv;

-	if (esw_offloads_mode(esw)) {
+	if (is_mdev_switchdev_mode(priv->mdev)) {
 		uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
 		uplink_priv = &uplink_rpriv->uplink_priv;

@@ -297,7 +295,7 @@ mlx5_tc_rule_insert(struct mlx5e_priv *priv,
 {
 	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;

-	if (esw_offloads_mode(esw))
+	if (is_mdev_switchdev_mode(priv->mdev))
 		return mlx5_eswitch_add_offloaded_rule(esw, spec, attr);

 	return	mlx5e_add_offloaded_nic_rule(priv, spec, attr);
@@ -310,7 +308,7 @@ mlx5_tc_rule_delete(struct mlx5e_priv *priv,
 {
 	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;

-	if (esw_offloads_mode(esw)) {
+	if (is_mdev_switchdev_mode(priv->mdev)) {
 		mlx5_eswitch_del_offloaded_rule(esw, rule, attr);

 		return;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index b652b4bde733..b44f28fb5518 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -2439,8 +2439,10 @@ int mlx5_eswitch_get_vport_stats(struct mlx5_eswitch *esw,
 	return err;
 }

-u8 mlx5_eswitch_mode(struct mlx5_eswitch *esw)
+u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev)
 {
+	struct mlx5_eswitch *esw = dev->priv.eswitch;
+
 	return ESW_ALLOWED(esw) ? esw->mode : MLX5_ESWITCH_NONE;
 }
 EXPORT_SYMBOL_GPL(mlx5_eswitch_mode);
diff --git a/include/linux/mlx5/eswitch.h b/include/linux/mlx5/eswitch.h
index b0ae8020f13e..29fd832950e0 100644
--- a/include/linux/mlx5/eswitch.h
+++ b/include/linux/mlx5/eswitch.h
@@ -96,10 +96,10 @@ static inline u32 mlx5_eswitch_get_vport_metadata_mask(void)

 u32 mlx5_eswitch_get_vport_metadata_for_match(struct mlx5_eswitch *esw,
 					      u16 vport_num);
-u8 mlx5_eswitch_mode(struct mlx5_eswitch *esw);
+u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev);
 #else  /* CONFIG_MLX5_ESWITCH */

-static inline u8 mlx5_eswitch_mode(struct mlx5_eswitch *esw)
+static inline u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev)
 {
 	return MLX5_ESWITCH_NONE;
 }
@@ -136,4 +136,8 @@ mlx5_eswitch_get_vport_metadata_mask(void)
 }
 #endif /* CONFIG_MLX5_ESWITCH */

+static inline bool is_mdev_switchdev_mode(struct mlx5_core_dev *dev)
+{
+	return mlx5_eswitch_mode(dev) == MLX5_ESWITCH_OFFLOADS;
+}
 #endif
--
2.28.0

Re: [PATCH mlx5-next v1 10/11] net/mlx5: Simplify eswitch mode check

[Roi Dayan]

On 2020-11-01 10:15 PM, Leon Romanovsky wrote:
> From: Leon Romanovsky <leonro@nvidia.com>
> 
> Provide mlx5_core device instead of "priv" pointer while checking
> eswith mode.
> 
> Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
> ---
>   drivers/infiniband/hw/mlx5/counters.c             | 7 -------
>   drivers/infiniband/hw/mlx5/ib_rep.c               | 5 -----
>   drivers/infiniband/hw/mlx5/ib_rep.h               | 6 ------
>   drivers/net/ethernet/mellanox/mlx5/core/dev.c     | 4 ++--
>   drivers/net/ethernet/mellanox/mlx5/core/devlink.c | 2 +-
>   drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 +-
>   drivers/net/ethernet/mellanox/mlx5/core/en_tc.c   | 8 +++-----
>   drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 4 +++-
>   include/linux/mlx5/eswitch.h                      | 8 ++++++--
>   9 files changed, 16 insertions(+), 30 deletions(-)
> 
> diff --git a/drivers/infiniband/hw/mlx5/counters.c b/drivers/infiniband/hw/mlx5/counters.c
> index 70c8fd67ee2f..084652e2b15a 100644
> --- a/drivers/infiniband/hw/mlx5/counters.c
> +++ b/drivers/infiniband/hw/mlx5/counters.c
> @@ -138,13 +138,6 @@ static int mlx5_ib_create_counters(struct ib_counters *counters,
>   }
> 
> 
> -static bool is_mdev_switchdev_mode(const struct mlx5_core_dev *mdev)
> -{
> -	return MLX5_ESWITCH_MANAGER(mdev) &&
> -	       mlx5_ib_eswitch_mode(mdev->priv.eswitch) ==
> -		       MLX5_ESWITCH_OFFLOADS;
> -}
> -
>   static const struct mlx5_ib_counters *get_counters(struct mlx5_ib_dev *dev,
>   						   u8 port_num)
>   {
> diff --git a/drivers/infiniband/hw/mlx5/ib_rep.c b/drivers/infiniband/hw/mlx5/ib_rep.c
> index 0dc15757cc66..9810bdd7f3bc 100644
> --- a/drivers/infiniband/hw/mlx5/ib_rep.c
> +++ b/drivers/infiniband/hw/mlx5/ib_rep.c
> @@ -102,11 +102,6 @@ static const struct mlx5_eswitch_rep_ops rep_ops = {
>   	.get_proto_dev = mlx5_ib_vport_get_proto_dev,
>   };
> 
> -u8 mlx5_ib_eswitch_mode(struct mlx5_eswitch *esw)
> -{
> -	return mlx5_eswitch_mode(esw);
> -}
> -
>   struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
>   					  u16 vport_num)
>   {
> diff --git a/drivers/infiniband/hw/mlx5/ib_rep.h b/drivers/infiniband/hw/mlx5/ib_rep.h
> index 94bf51ddd422..93f562735e89 100644
> --- a/drivers/infiniband/hw/mlx5/ib_rep.h
> +++ b/drivers/infiniband/hw/mlx5/ib_rep.h
> @@ -12,7 +12,6 @@
>   extern const struct mlx5_ib_profile raw_eth_profile;
> 
>   #ifdef CONFIG_MLX5_ESWITCH
> -u8 mlx5_ib_eswitch_mode(struct mlx5_eswitch *esw);
>   struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
>   					  u16 vport_num);
>   struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw);
> @@ -26,11 +25,6 @@ struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
>   struct net_device *mlx5_ib_get_rep_netdev(struct mlx5_eswitch *esw,
>   					  u16 vport_num);
>   #else /* CONFIG_MLX5_ESWITCH */
> -static inline u8 mlx5_ib_eswitch_mode(struct mlx5_eswitch *esw)
> -{
> -	return MLX5_ESWITCH_NONE;
> -}
> -
>   static inline
>   struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
>   					  u16 vport_num)
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
> index 3a81c2f1971b..b051417ede67 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
> @@ -47,7 +47,7 @@ static bool is_eth_rep_supported(struct mlx5_core_dev *dev)
>   	if (!MLX5_ESWITCH_MANAGER(dev))
>   		return false;
> 
> -	if (mlx5_eswitch_mode(dev->priv.eswitch) != MLX5_ESWITCH_OFFLOADS)
> +	if (!is_mdev_switchdev_mode(dev))
>   		return false;
> 
>   	return true;
> @@ -144,7 +144,7 @@ static bool is_ib_rep_supported(struct mlx5_core_dev *dev)
>   	if (!MLX5_ESWITCH_MANAGER(dev))
>   		return false;
> 
> -	if (mlx5_eswitch_mode(dev->priv.eswitch) != MLX5_ESWITCH_OFFLOADS)
> +	if (!is_mdev_switchdev_mode(dev))
>   		return false;
> 
>   	if (mlx5_core_mp_enabled(dev))
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> index 1a351e2f6ace..aeffb6b135ee 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
> @@ -221,7 +221,7 @@ static int mlx5_devlink_fs_mode_validate(struct devlink *devlink, u32 id,
>   		u8 eswitch_mode;
>   		bool smfs_cap;
> 
> -		eswitch_mode = mlx5_eswitch_mode(dev->priv.eswitch);
> +		eswitch_mode = mlx5_eswitch_mode(dev);
>   		smfs_cap = mlx5_fs_dr_is_supported(dev);
> 
>   		if (!smfs_cap) {
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
> index 3c4f880c6329..8d65ac888a28 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
> @@ -3135,7 +3135,7 @@ static void mlx5e_modify_admin_state(struct mlx5_core_dev *mdev,
> 
>   	mlx5_set_port_admin_status(mdev, state);
> 
> -	if (!MLX5_ESWITCH_MANAGER(mdev) ||  mlx5_eswitch_mode(esw) == MLX5_ESWITCH_OFFLOADS)
> +	if (mlx5_eswitch_mode(mdev) != MLX5_ESWITCH_LEGACY)
>   		return;
> 
>   	if (state == MLX5_PORT_UP)
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> index e3a968e9e2a0..7548bab78654 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> @@ -271,8 +271,6 @@ mlx5e_tc_match_to_reg_set(struct mlx5_core_dev *mdev,
>   	return 0;
>   }
> 
> -#define esw_offloads_mode(esw) (mlx5_eswitch_mode(esw) == MLX5_ESWITCH_OFFLOADS)
> -
>   static struct mlx5_tc_ct_priv *
>   get_ct_priv(struct mlx5e_priv *priv)
>   {
> @@ -280,7 +278,7 @@ get_ct_priv(struct mlx5e_priv *priv)
>   	struct mlx5_rep_uplink_priv *uplink_priv;
>   	struct mlx5e_rep_priv *uplink_rpriv;
> 
> -	if (esw_offloads_mode(esw)) {
> +	if (is_mdev_switchdev_mode(priv->mdev)) {
>   		uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
>   		uplink_priv = &uplink_rpriv->uplink_priv;
> 
> @@ -297,7 +295,7 @@ mlx5_tc_rule_insert(struct mlx5e_priv *priv,
>   {
>   	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
> 
> -	if (esw_offloads_mode(esw))
> +	if (is_mdev_switchdev_mode(priv->mdev))
>   		return mlx5_eswitch_add_offloaded_rule(esw, spec, attr);
> 
>   	return	mlx5e_add_offloaded_nic_rule(priv, spec, attr);
> @@ -310,7 +308,7 @@ mlx5_tc_rule_delete(struct mlx5e_priv *priv,
>   {
>   	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
> 
> -	if (esw_offloads_mode(esw)) {
> +	if (is_mdev_switchdev_mode(priv->mdev)) {
>   		mlx5_eswitch_del_offloaded_rule(esw, rule, attr);
> 
>   		return;
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
> index b652b4bde733..b44f28fb5518 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
> @@ -2439,8 +2439,10 @@ int mlx5_eswitch_get_vport_stats(struct mlx5_eswitch *esw,
>   	return err;
>   }
> 
> -u8 mlx5_eswitch_mode(struct mlx5_eswitch *esw)
> +u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev)
>   {
> +	struct mlx5_eswitch *esw = dev->priv.eswitch;
> +
>   	return ESW_ALLOWED(esw) ? esw->mode : MLX5_ESWITCH_NONE;
>   }
>   EXPORT_SYMBOL_GPL(mlx5_eswitch_mode);
> diff --git a/include/linux/mlx5/eswitch.h b/include/linux/mlx5/eswitch.h
> index b0ae8020f13e..29fd832950e0 100644
> --- a/include/linux/mlx5/eswitch.h
> +++ b/include/linux/mlx5/eswitch.h
> @@ -96,10 +96,10 @@ static inline u32 mlx5_eswitch_get_vport_metadata_mask(void)
> 
>   u32 mlx5_eswitch_get_vport_metadata_for_match(struct mlx5_eswitch *esw,
>   					      u16 vport_num);
> -u8 mlx5_eswitch_mode(struct mlx5_eswitch *esw);
> +u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev);
>   #else  /* CONFIG_MLX5_ESWITCH */
> 
> -static inline u8 mlx5_eswitch_mode(struct mlx5_eswitch *esw)
> +static inline u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev)
>   {
>   	return MLX5_ESWITCH_NONE;
>   }
> @@ -136,4 +136,8 @@ mlx5_eswitch_get_vport_metadata_mask(void)
>   }
>   #endif /* CONFIG_MLX5_ESWITCH */
> 
> +static inline bool is_mdev_switchdev_mode(struct mlx5_core_dev *dev)
> +{
> +	return mlx5_eswitch_mode(dev) == MLX5_ESWITCH_OFFLOADS;
> +}
>   #endif
> --
> 2.28.0
> 

Reviewed-by: Roi Dayan <roid@nvidia.com>

[PATCH mlx5-next v1 11/11] RDMA/mlx5: Remove IB representors dead code

[Leon Romanovsky]

From: Leon Romanovsky <leonro@nvidia.com>

Delete dead code.

Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/mlx5/ib_rep.c | 31 +++++++----------------------
 drivers/infiniband/hw/mlx5/ib_rep.h | 31 -----------------------------
 2 files changed, 7 insertions(+), 55 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/ib_rep.c b/drivers/infiniband/hw/mlx5/ib_rep.c
index 9810bdd7f3bc..a1a9450ed92c 100644
--- a/drivers/infiniband/hw/mlx5/ib_rep.c
+++ b/drivers/infiniband/hw/mlx5/ib_rep.c
@@ -13,7 +13,7 @@ mlx5_ib_set_vport_rep(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 	struct mlx5_ib_dev *ibdev;
 	int vport_index;

-	ibdev = mlx5_ib_get_uplink_ibdev(dev->priv.eswitch);
+	ibdev = mlx5_eswitch_uplink_get_proto_dev(dev->priv.eswitch, REP_IB);
 	vport_index = rep->vport_index;

 	ibdev->port[vport_index].rep = rep;
@@ -74,6 +74,11 @@ mlx5_ib_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 	return ret;
 }

+static void *mlx5_ib_rep_to_dev(struct mlx5_eswitch_rep *rep)
+{
+	return rep->rep_data[REP_IB].priv;
+}
+
 static void
 mlx5_ib_vport_rep_unload(struct mlx5_eswitch_rep *rep)
 {
@@ -91,40 +96,18 @@ mlx5_ib_vport_rep_unload(struct mlx5_eswitch_rep *rep)
 		__mlx5_ib_remove(dev, dev->profile, MLX5_IB_STAGE_MAX);
 }

-static void *mlx5_ib_vport_get_proto_dev(struct mlx5_eswitch_rep *rep)
-{
-	return mlx5_ib_rep_to_dev(rep);
-}
-
 static const struct mlx5_eswitch_rep_ops rep_ops = {
 	.load = mlx5_ib_vport_rep_load,
 	.unload = mlx5_ib_vport_rep_unload,
-	.get_proto_dev = mlx5_ib_vport_get_proto_dev,
+	.get_proto_dev = mlx5_ib_rep_to_dev,
 };

-struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
-					  u16 vport_num)
-{
-	return mlx5_eswitch_get_proto_dev(esw, vport_num, REP_IB);
-}
-
 struct net_device *mlx5_ib_get_rep_netdev(struct mlx5_eswitch *esw,
 					  u16 vport_num)
 {
 	return mlx5_eswitch_get_proto_dev(esw, vport_num, REP_ETH);
 }

-struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw)
-{
-	return mlx5_eswitch_uplink_get_proto_dev(esw, REP_IB);
-}
-
-struct mlx5_eswitch_rep *mlx5_ib_vport_rep(struct mlx5_eswitch *esw,
-					   u16 vport_num)
-{
-	return mlx5_eswitch_vport_rep(esw, vport_num);
-}
-
 struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
 						   struct mlx5_ib_sq *sq,
 						   u16 port)
diff --git a/drivers/infiniband/hw/mlx5/ib_rep.h b/drivers/infiniband/hw/mlx5/ib_rep.h
index 93f562735e89..ce1dcb105dbd 100644
--- a/drivers/infiniband/hw/mlx5/ib_rep.h
+++ b/drivers/infiniband/hw/mlx5/ib_rep.h
@@ -12,11 +12,6 @@
 extern const struct mlx5_ib_profile raw_eth_profile;

 #ifdef CONFIG_MLX5_ESWITCH
-struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
-					  u16 vport_num);
-struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw);
-struct mlx5_eswitch_rep *mlx5_ib_vport_rep(struct mlx5_eswitch *esw,
-					   u16 vport_num);
 int mlx5r_rep_init(void);
 void mlx5r_rep_cleanup(void);
 struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
@@ -25,26 +20,6 @@ struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
 struct net_device *mlx5_ib_get_rep_netdev(struct mlx5_eswitch *esw,
 					  u16 vport_num);
 #else /* CONFIG_MLX5_ESWITCH */
-static inline
-struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
-					  u16 vport_num)
-{
-	return NULL;
-}
-
-static inline
-struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw)
-{
-	return NULL;
-}
-
-static inline
-struct mlx5_eswitch_rep *mlx5_ib_vport_rep(struct mlx5_eswitch *esw,
-					   u16 vport_num)
-{
-	return NULL;
-}
-
 static inline int mlx5r_rep_init(void) { return 0; }
 static inline void mlx5r_rep_cleanup(void) {}
 static inline
@@ -62,10 +37,4 @@ struct net_device *mlx5_ib_get_rep_netdev(struct mlx5_eswitch *esw,
 	return NULL;
 }
 #endif
-
-static inline
-struct mlx5_ib_dev *mlx5_ib_rep_to_dev(struct mlx5_eswitch_rep *rep)
-{
-	return rep->rep_data[REP_IB].priv;
-}
 #endif /* __MLX5_IB_REP_H__ */
--
2.28.0

Re: [PATCH mlx5-next v1 11/11] RDMA/mlx5: Remove IB representors dead code

[Roi Dayan]

On 2020-11-01 10:15 PM, Leon Romanovsky wrote:
> From: Leon Romanovsky <leonro@nvidia.com>
> 
> Delete dead code.
> 
> Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
> ---
>   drivers/infiniband/hw/mlx5/ib_rep.c | 31 +++++++----------------------
>   drivers/infiniband/hw/mlx5/ib_rep.h | 31 -----------------------------
>   2 files changed, 7 insertions(+), 55 deletions(-)
> 
> diff --git a/drivers/infiniband/hw/mlx5/ib_rep.c b/drivers/infiniband/hw/mlx5/ib_rep.c
> index 9810bdd7f3bc..a1a9450ed92c 100644
> --- a/drivers/infiniband/hw/mlx5/ib_rep.c
> +++ b/drivers/infiniband/hw/mlx5/ib_rep.c
> @@ -13,7 +13,7 @@ mlx5_ib_set_vport_rep(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
>   	struct mlx5_ib_dev *ibdev;
>   	int vport_index;
> 
> -	ibdev = mlx5_ib_get_uplink_ibdev(dev->priv.eswitch);
> +	ibdev = mlx5_eswitch_uplink_get_proto_dev(dev->priv.eswitch, REP_IB);
>   	vport_index = rep->vport_index;
> 
>   	ibdev->port[vport_index].rep = rep;
> @@ -74,6 +74,11 @@ mlx5_ib_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
>   	return ret;
>   }
> 
> +static void *mlx5_ib_rep_to_dev(struct mlx5_eswitch_rep *rep)
> +{
> +	return rep->rep_data[REP_IB].priv;
> +}
> +
>   static void
>   mlx5_ib_vport_rep_unload(struct mlx5_eswitch_rep *rep)
>   {
> @@ -91,40 +96,18 @@ mlx5_ib_vport_rep_unload(struct mlx5_eswitch_rep *rep)
>   		__mlx5_ib_remove(dev, dev->profile, MLX5_IB_STAGE_MAX);
>   }
> 
> -static void *mlx5_ib_vport_get_proto_dev(struct mlx5_eswitch_rep *rep)
> -{
> -	return mlx5_ib_rep_to_dev(rep);
> -}
> -
>   static const struct mlx5_eswitch_rep_ops rep_ops = {
>   	.load = mlx5_ib_vport_rep_load,
>   	.unload = mlx5_ib_vport_rep_unload,
> -	.get_proto_dev = mlx5_ib_vport_get_proto_dev,
> +	.get_proto_dev = mlx5_ib_rep_to_dev,
>   };
> 
> -struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
> -					  u16 vport_num)
> -{
> -	return mlx5_eswitch_get_proto_dev(esw, vport_num, REP_IB);
> -}
> -
>   struct net_device *mlx5_ib_get_rep_netdev(struct mlx5_eswitch *esw,
>   					  u16 vport_num)
>   {
>   	return mlx5_eswitch_get_proto_dev(esw, vport_num, REP_ETH);
>   }
> 
> -struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw)
> -{
> -	return mlx5_eswitch_uplink_get_proto_dev(esw, REP_IB);
> -}
> -
> -struct mlx5_eswitch_rep *mlx5_ib_vport_rep(struct mlx5_eswitch *esw,
> -					   u16 vport_num)
> -{
> -	return mlx5_eswitch_vport_rep(esw, vport_num);
> -}
> -
>   struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
>   						   struct mlx5_ib_sq *sq,
>   						   u16 port)
> diff --git a/drivers/infiniband/hw/mlx5/ib_rep.h b/drivers/infiniband/hw/mlx5/ib_rep.h
> index 93f562735e89..ce1dcb105dbd 100644
> --- a/drivers/infiniband/hw/mlx5/ib_rep.h
> +++ b/drivers/infiniband/hw/mlx5/ib_rep.h
> @@ -12,11 +12,6 @@
>   extern const struct mlx5_ib_profile raw_eth_profile;
> 
>   #ifdef CONFIG_MLX5_ESWITCH
> -struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
> -					  u16 vport_num);
> -struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw);
> -struct mlx5_eswitch_rep *mlx5_ib_vport_rep(struct mlx5_eswitch *esw,
> -					   u16 vport_num);
>   int mlx5r_rep_init(void);
>   void mlx5r_rep_cleanup(void);
>   struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
> @@ -25,26 +20,6 @@ struct mlx5_flow_handle *create_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
>   struct net_device *mlx5_ib_get_rep_netdev(struct mlx5_eswitch *esw,
>   					  u16 vport_num);
>   #else /* CONFIG_MLX5_ESWITCH */
> -static inline
> -struct mlx5_ib_dev *mlx5_ib_get_rep_ibdev(struct mlx5_eswitch *esw,
> -					  u16 vport_num)
> -{
> -	return NULL;
> -}
> -
> -static inline
> -struct mlx5_ib_dev *mlx5_ib_get_uplink_ibdev(struct mlx5_eswitch *esw)
> -{
> -	return NULL;
> -}
> -
> -static inline
> -struct mlx5_eswitch_rep *mlx5_ib_vport_rep(struct mlx5_eswitch *esw,
> -					   u16 vport_num)
> -{
> -	return NULL;
> -}
> -
>   static inline int mlx5r_rep_init(void) { return 0; }
>   static inline void mlx5r_rep_cleanup(void) {}
>   static inline
> @@ -62,10 +37,4 @@ struct net_device *mlx5_ib_get_rep_netdev(struct mlx5_eswitch *esw,
>   	return NULL;
>   }
>   #endif
> -
> -static inline
> -struct mlx5_ib_dev *mlx5_ib_rep_to_dev(struct mlx5_eswitch_rep *rep)
> -{
> -	return rep->rep_data[REP_IB].priv;
> -}
>   #endif /* __MLX5_IB_REP_H__ */
> --
> 2.28.0
> 

Reviewed-by: Roi Dayan <roid@nvidia.com>