2019-05-27 14:55:01 +08:00
|
|
|
/* SPDX-License-Identifier: GPL-2.0-or-later */
|
2014-11-28 21:34:17 +08:00
|
|
|
/*
|
|
|
|
|
* include/net/switchdev.h - Switch device API
|
2015-09-24 16:02:41 +08:00
|
|
|
* Copyright (c) 2014-2015 Jiri Pirko <jiri@resnulli.us>
|
2015-03-10 04:59:09 +08:00
|
|
|
* Copyright (c) 2014-2015 Scott Feldman <sfeldma@gmail.com>
|
2014-11-28 21:34:17 +08:00
|
|
|
*/
|
|
|
|
|
#ifndef _LINUX_SWITCHDEV_H_
|
|
|
|
|
#define _LINUX_SWITCHDEV_H_
|
|
|
|
|
|
|
|
|
|
#include <linux/netdevice.h>
|
2015-01-16 06:49:36 +08:00
|
|
|
#include <linux/notifier.h>
|
2015-09-24 16:02:41 +08:00
|
|
|
#include <linux/list.h>
|
2015-10-15 01:40:51 +08:00
|
|
|
#include <net/ip_fib.h>
|
2015-01-16 06:49:36 +08:00
|
|
|
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
#define SWITCHDEV_F_NO_RECURSE BIT(0)
|
2015-10-09 10:23:18 +08:00
|
|
|
#define SWITCHDEV_F_SKIP_EOPNOTSUPP BIT(1)
|
2015-10-15 01:40:50 +08:00
|
|
|
#define SWITCHDEV_F_DEFER BIT(2)
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
|
2015-09-24 16:02:41 +08:00
|
|
|
struct switchdev_trans {
|
2015-09-24 16:02:49 +08:00
|
|
|
bool ph_prepare;
|
2015-09-24 16:02:41 +08:00
|
|
|
};
|
|
|
|
|
|
2015-09-24 16:02:43 +08:00
|
|
|
static inline bool switchdev_trans_ph_prepare(struct switchdev_trans *trans)
|
|
|
|
|
{
|
2015-09-24 16:02:49 +08:00
|
|
|
return trans && trans->ph_prepare;
|
2015-09-24 16:02:43 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline bool switchdev_trans_ph_commit(struct switchdev_trans *trans)
|
|
|
|
|
{
|
2015-09-24 16:02:49 +08:00
|
|
|
return trans && !trans->ph_prepare;
|
2015-09-24 16:02:43 +08:00
|
|
|
}
|
|
|
|
|
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
enum switchdev_attr_id {
|
2015-10-01 17:03:42 +08:00
|
|
|
SWITCHDEV_ATTR_ID_UNDEFINED,
|
|
|
|
|
SWITCHDEV_ATTR_ID_PORT_STP_STATE,
|
|
|
|
|
SWITCHDEV_ATTR_ID_PORT_BRIDGE_FLAGS,
|
2019-02-21 08:58:19 +08:00
|
|
|
SWITCHDEV_ATTR_ID_PORT_PRE_BRIDGE_FLAGS,
|
2017-02-09 21:54:42 +08:00
|
|
|
SWITCHDEV_ATTR_ID_PORT_MROUTER,
|
2015-10-09 10:23:17 +08:00
|
|
|
SWITCHDEV_ATTR_ID_BRIDGE_AGEING_TIME,
|
2016-01-06 20:01:05 +08:00
|
|
|
SWITCHDEV_ATTR_ID_BRIDGE_VLAN_FILTERING,
|
2017-02-09 21:54:40 +08:00
|
|
|
SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED,
|
2017-10-09 17:15:31 +08:00
|
|
|
SWITCHDEV_ATTR_ID_BRIDGE_MROUTER,
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
struct switchdev_attr {
|
2015-12-15 23:03:35 +08:00
|
|
|
struct net_device *orig_dev;
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
enum switchdev_attr_id id;
|
|
|
|
|
u32 flags;
|
2016-04-21 18:52:43 +08:00
|
|
|
void *complete_priv;
|
|
|
|
|
void (*complete)(struct net_device *dev, int err, void *priv);
|
2015-05-11 00:47:49 +08:00
|
|
|
union {
|
2015-05-11 00:47:51 +08:00
|
|
|
u8 stp_state; /* PORT_STP_STATE */
|
2019-02-21 08:58:19 +08:00
|
|
|
unsigned long brport_flags; /* PORT_{PRE}_BRIDGE_FLAGS */
|
2017-02-09 21:54:42 +08:00
|
|
|
bool mrouter; /* PORT_MROUTER */
|
2016-07-19 03:02:06 +08:00
|
|
|
clock_t ageing_time; /* BRIDGE_AGEING_TIME */
|
2016-01-06 20:01:05 +08:00
|
|
|
bool vlan_filtering; /* BRIDGE_VLAN_FILTERING */
|
2017-02-09 21:54:40 +08:00
|
|
|
bool mc_disabled; /* MC_DISABLED */
|
2015-05-14 02:16:50 +08:00
|
|
|
} u;
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
};
|
|
|
|
|
|
2015-05-11 00:47:52 +08:00
|
|
|
enum switchdev_obj_id {
|
2015-10-01 17:03:41 +08:00
|
|
|
SWITCHDEV_OBJ_ID_UNDEFINED,
|
|
|
|
|
SWITCHDEV_OBJ_ID_PORT_VLAN,
|
2016-01-11 04:06:22 +08:00
|
|
|
SWITCHDEV_OBJ_ID_PORT_MDB,
|
2017-11-10 06:10:59 +08:00
|
|
|
SWITCHDEV_OBJ_ID_HOST_MDB,
|
2015-05-11 00:47:52 +08:00
|
|
|
};
|
|
|
|
|
|
2015-10-01 17:03:45 +08:00
|
|
|
struct switchdev_obj {
|
2015-12-15 23:03:35 +08:00
|
|
|
struct net_device *orig_dev;
|
2015-10-01 17:03:46 +08:00
|
|
|
enum switchdev_obj_id id;
|
2015-10-15 01:40:52 +08:00
|
|
|
u32 flags;
|
2016-04-21 18:52:43 +08:00
|
|
|
void *complete_priv;
|
|
|
|
|
void (*complete)(struct net_device *dev, int err, void *priv);
|
2015-10-01 17:03:45 +08:00
|
|
|
};
|
|
|
|
|
|
2015-10-01 17:03:41 +08:00
|
|
|
/* SWITCHDEV_OBJ_ID_PORT_VLAN */
|
2015-10-01 17:03:43 +08:00
|
|
|
struct switchdev_obj_port_vlan {
|
2015-10-01 17:03:45 +08:00
|
|
|
struct switchdev_obj obj;
|
2015-09-30 00:07:18 +08:00
|
|
|
u16 flags;
|
|
|
|
|
u16 vid_begin;
|
|
|
|
|
u16 vid_end;
|
|
|
|
|
};
|
|
|
|
|
|
switchdev: SWITCHDEV_OBJ_PORT_{VLAN, MDB}(): Sanitize
The two macros SWITCHDEV_OBJ_PORT_VLAN() and SWITCHDEV_OBJ_PORT_MDB()
expand to a container_of() call, yielding an appropriate container of
their sole argument. However, due to a name collision, the first
argument, i.e. the contained object pointer, is not the only one to get
expanded. The third argument, which is a structure member name, and
should be kept literal, gets expanded as well. The only safe way to use
these two macros is therefore to name the local variable passed to them
"obj".
To fix this, rename the sole argument of the two macros from
"obj" (which collides with the member name) to "OBJ". Additionally,
instead of passing "OBJ" to container_of() verbatim, parenthesize it, so
that a comma in the passed-in expression doesn't pollute the
container_of() invocation.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-11-23 07:28:07 +08:00
|
|
|
#define SWITCHDEV_OBJ_PORT_VLAN(OBJ) \
|
|
|
|
|
container_of((OBJ), struct switchdev_obj_port_vlan, obj)
|
2015-10-01 17:03:45 +08:00
|
|
|
|
2016-01-11 04:06:22 +08:00
|
|
|
/* SWITCHDEV_OBJ_ID_PORT_MDB */
|
|
|
|
|
struct switchdev_obj_port_mdb {
|
|
|
|
|
struct switchdev_obj obj;
|
|
|
|
|
unsigned char addr[ETH_ALEN];
|
|
|
|
|
u16 vid;
|
|
|
|
|
};
|
|
|
|
|
|
switchdev: SWITCHDEV_OBJ_PORT_{VLAN, MDB}(): Sanitize
The two macros SWITCHDEV_OBJ_PORT_VLAN() and SWITCHDEV_OBJ_PORT_MDB()
expand to a container_of() call, yielding an appropriate container of
their sole argument. However, due to a name collision, the first
argument, i.e. the contained object pointer, is not the only one to get
expanded. The third argument, which is a structure member name, and
should be kept literal, gets expanded as well. The only safe way to use
these two macros is therefore to name the local variable passed to them
"obj".
To fix this, rename the sole argument of the two macros from
"obj" (which collides with the member name) to "OBJ". Additionally,
instead of passing "OBJ" to container_of() verbatim, parenthesize it, so
that a comma in the passed-in expression doesn't pollute the
container_of() invocation.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-11-23 07:28:07 +08:00
|
|
|
#define SWITCHDEV_OBJ_PORT_MDB(OBJ) \
|
|
|
|
|
container_of((OBJ), struct switchdev_obj_port_mdb, obj)
|
2016-01-11 04:06:22 +08:00
|
|
|
|
2015-10-01 17:03:45 +08:00
|
|
|
typedef int switchdev_obj_dump_cb_t(struct switchdev_obj *obj);
|
|
|
|
|
|
2015-05-11 00:47:46 +08:00
|
|
|
enum switchdev_notifier_type {
|
2017-06-08 14:44:14 +08:00
|
|
|
SWITCHDEV_FDB_ADD_TO_BRIDGE = 1,
|
|
|
|
|
SWITCHDEV_FDB_DEL_TO_BRIDGE,
|
|
|
|
|
SWITCHDEV_FDB_ADD_TO_DEVICE,
|
|
|
|
|
SWITCHDEV_FDB_DEL_TO_DEVICE,
|
2017-06-08 14:44:15 +08:00
|
|
|
SWITCHDEV_FDB_OFFLOADED,
|
2018-10-17 16:53:22 +08:00
|
|
|
|
switchdev: Add SWITCHDEV_PORT_OBJ_ADD, SWITCHDEV_PORT_OBJ_DEL
An offloading driver may need to have access to switchdev events on
ports that aren't directly under its control. An example is a VXLAN port
attached to a bridge offloaded by a driver. The driver needs to know
about VLANs configured on the VXLAN device. However the VXLAN device
isn't stashed between the bridge and a front-panel-port device (such as
is the case e.g. for LAG devices), so the usual switchdev ops don't
reach the driver.
VXLAN is likely not the only device type like this: in theory any L2
tunnel device that needs offloading will prompt requirement of this
sort. This falsifies the assumption that only the lower devices of a
front panel port need to be notified to achieve flawless offloading.
A way to fix this is to give up the notion of port object addition /
deletion as a switchdev operation, which assumes somewhat tight coupling
between the message producer and consumer. And instead send the message
over a notifier chain.
To that end, introduce two new switchdev notifier types,
SWITCHDEV_PORT_OBJ_ADD and SWITCHDEV_PORT_OBJ_DEL. These notifier types
communicate the same event as the corresponding switchdev op, except in
a form of a notification. struct switchdev_notifier_port_obj_info was
added to carry the fields that the switchdev op carries. An additional
field, handled, will be used to communicate back to switchdev that the
event has reached an interested party, which will be important for the
two-phase commit.
The two switchdev operations themselves are kept in place. Following
patches first convert individual clients to the notifier protocol, and
only then are the operations removed.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-11-23 07:28:38 +08:00
|
|
|
SWITCHDEV_PORT_OBJ_ADD, /* Blocking. */
|
|
|
|
|
SWITCHDEV_PORT_OBJ_DEL, /* Blocking. */
|
2019-02-28 03:44:25 +08:00
|
|
|
SWITCHDEV_PORT_ATTR_SET, /* May be blocking . */
|
switchdev: Add SWITCHDEV_PORT_OBJ_ADD, SWITCHDEV_PORT_OBJ_DEL
An offloading driver may need to have access to switchdev events on
ports that aren't directly under its control. An example is a VXLAN port
attached to a bridge offloaded by a driver. The driver needs to know
about VLANs configured on the VXLAN device. However the VXLAN device
isn't stashed between the bridge and a front-panel-port device (such as
is the case e.g. for LAG devices), so the usual switchdev ops don't
reach the driver.
VXLAN is likely not the only device type like this: in theory any L2
tunnel device that needs offloading will prompt requirement of this
sort. This falsifies the assumption that only the lower devices of a
front panel port need to be notified to achieve flawless offloading.
A way to fix this is to give up the notion of port object addition /
deletion as a switchdev operation, which assumes somewhat tight coupling
between the message producer and consumer. And instead send the message
over a notifier chain.
To that end, introduce two new switchdev notifier types,
SWITCHDEV_PORT_OBJ_ADD and SWITCHDEV_PORT_OBJ_DEL. These notifier types
communicate the same event as the corresponding switchdev op, except in
a form of a notification. struct switchdev_notifier_port_obj_info was
added to carry the fields that the switchdev op carries. An additional
field, handled, will be used to communicate back to switchdev that the
event has reached an interested party, which will be important for the
two-phase commit.
The two switchdev operations themselves are kept in place. Following
patches first convert individual clients to the notifier protocol, and
only then are the operations removed.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-11-23 07:28:38 +08:00
|
|
|
|
2018-11-21 16:02:39 +08:00
|
|
|
SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE,
|
|
|
|
|
SWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE,
|
2018-10-17 16:53:22 +08:00
|
|
|
SWITCHDEV_VXLAN_FDB_ADD_TO_DEVICE,
|
|
|
|
|
SWITCHDEV_VXLAN_FDB_DEL_TO_DEVICE,
|
2018-10-17 16:53:26 +08:00
|
|
|
SWITCHDEV_VXLAN_FDB_OFFLOADED,
|
2015-01-16 06:49:37 +08:00
|
|
|
};
|
|
|
|
|
|
2015-05-11 00:47:46 +08:00
|
|
|
struct switchdev_notifier_info {
|
2015-01-16 06:49:36 +08:00
|
|
|
struct net_device *dev;
|
2018-12-13 01:02:54 +08:00
|
|
|
struct netlink_ext_ack *extack;
|
2015-01-16 06:49:36 +08:00
|
|
|
};
|
|
|
|
|
|
2015-05-11 00:47:46 +08:00
|
|
|
struct switchdev_notifier_fdb_info {
|
|
|
|
|
struct switchdev_notifier_info info; /* must be first */
|
2015-01-16 06:49:37 +08:00
|
|
|
const unsigned char *addr;
|
|
|
|
|
u16 vid;
|
2018-10-17 16:53:29 +08:00
|
|
|
u8 added_by_user:1,
|
|
|
|
|
offloaded:1;
|
2015-01-16 06:49:37 +08:00
|
|
|
};
|
|
|
|
|
|
switchdev: Add SWITCHDEV_PORT_OBJ_ADD, SWITCHDEV_PORT_OBJ_DEL
An offloading driver may need to have access to switchdev events on
ports that aren't directly under its control. An example is a VXLAN port
attached to a bridge offloaded by a driver. The driver needs to know
about VLANs configured on the VXLAN device. However the VXLAN device
isn't stashed between the bridge and a front-panel-port device (such as
is the case e.g. for LAG devices), so the usual switchdev ops don't
reach the driver.
VXLAN is likely not the only device type like this: in theory any L2
tunnel device that needs offloading will prompt requirement of this
sort. This falsifies the assumption that only the lower devices of a
front panel port need to be notified to achieve flawless offloading.
A way to fix this is to give up the notion of port object addition /
deletion as a switchdev operation, which assumes somewhat tight coupling
between the message producer and consumer. And instead send the message
over a notifier chain.
To that end, introduce two new switchdev notifier types,
SWITCHDEV_PORT_OBJ_ADD and SWITCHDEV_PORT_OBJ_DEL. These notifier types
communicate the same event as the corresponding switchdev op, except in
a form of a notification. struct switchdev_notifier_port_obj_info was
added to carry the fields that the switchdev op carries. An additional
field, handled, will be used to communicate back to switchdev that the
event has reached an interested party, which will be important for the
two-phase commit.
The two switchdev operations themselves are kept in place. Following
patches first convert individual clients to the notifier protocol, and
only then are the operations removed.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-11-23 07:28:38 +08:00
|
|
|
struct switchdev_notifier_port_obj_info {
|
|
|
|
|
struct switchdev_notifier_info info; /* must be first */
|
|
|
|
|
const struct switchdev_obj *obj;
|
|
|
|
|
struct switchdev_trans *trans;
|
|
|
|
|
bool handled;
|
|
|
|
|
};
|
|
|
|
|
|
2019-02-28 03:44:25 +08:00
|
|
|
struct switchdev_notifier_port_attr_info {
|
|
|
|
|
struct switchdev_notifier_info info; /* must be first */
|
|
|
|
|
const struct switchdev_attr *attr;
|
|
|
|
|
struct switchdev_trans *trans;
|
|
|
|
|
bool handled;
|
|
|
|
|
};
|
|
|
|
|
|
2015-01-16 06:49:36 +08:00
|
|
|
static inline struct net_device *
|
2015-05-11 00:47:46 +08:00
|
|
|
switchdev_notifier_info_to_dev(const struct switchdev_notifier_info *info)
|
2015-01-16 06:49:36 +08:00
|
|
|
{
|
|
|
|
|
return info->dev;
|
|
|
|
|
}
|
2014-11-28 21:34:17 +08:00
|
|
|
|
2018-12-13 01:02:54 +08:00
|
|
|
static inline struct netlink_ext_ack *
|
|
|
|
|
switchdev_notifier_info_to_extack(const struct switchdev_notifier_info *info)
|
|
|
|
|
{
|
|
|
|
|
return info->extack;
|
|
|
|
|
}
|
|
|
|
|
|
2014-11-28 21:34:17 +08:00
|
|
|
#ifdef CONFIG_NET_SWITCHDEV
|
|
|
|
|
|
2015-10-15 01:40:48 +08:00
|
|
|
void switchdev_deferred_process(void);
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
int switchdev_port_attr_set(struct net_device *dev,
|
2015-10-15 01:40:49 +08:00
|
|
|
const struct switchdev_attr *attr);
|
2015-10-01 17:03:46 +08:00
|
|
|
int switchdev_port_obj_add(struct net_device *dev,
|
2018-12-13 01:02:52 +08:00
|
|
|
const struct switchdev_obj *obj,
|
|
|
|
|
struct netlink_ext_ack *extack);
|
2015-10-01 17:03:46 +08:00
|
|
|
int switchdev_port_obj_del(struct net_device *dev,
|
2015-10-01 17:03:45 +08:00
|
|
|
const struct switchdev_obj *obj);
|
2018-11-23 07:28:25 +08:00
|
|
|
|
2015-05-11 00:47:46 +08:00
|
|
|
int register_switchdev_notifier(struct notifier_block *nb);
|
|
|
|
|
int unregister_switchdev_notifier(struct notifier_block *nb);
|
|
|
|
|
int call_switchdev_notifiers(unsigned long val, struct net_device *dev,
|
2019-01-17 07:06:56 +08:00
|
|
|
struct switchdev_notifier_info *info,
|
|
|
|
|
struct netlink_ext_ack *extack);
|
2018-11-23 07:28:25 +08:00
|
|
|
|
|
|
|
|
int register_switchdev_blocking_notifier(struct notifier_block *nb);
|
|
|
|
|
int unregister_switchdev_blocking_notifier(struct notifier_block *nb);
|
|
|
|
|
int call_switchdev_blocking_notifiers(unsigned long val, struct net_device *dev,
|
2018-12-13 01:02:54 +08:00
|
|
|
struct switchdev_notifier_info *info,
|
|
|
|
|
struct netlink_ext_ack *extack);
|
2018-11-23 07:28:25 +08:00
|
|
|
|
2015-07-19 09:24:50 +08:00
|
|
|
void switchdev_port_fwd_mark_set(struct net_device *dev,
|
|
|
|
|
struct net_device *group_dev,
|
|
|
|
|
bool joining);
|
2015-03-06 13:21:15 +08:00
|
|
|
|
2018-11-23 07:29:44 +08:00
|
|
|
int switchdev_handle_port_obj_add(struct net_device *dev,
|
|
|
|
|
struct switchdev_notifier_port_obj_info *port_obj_info,
|
|
|
|
|
bool (*check_cb)(const struct net_device *dev),
|
|
|
|
|
int (*add_cb)(struct net_device *dev,
|
|
|
|
|
const struct switchdev_obj *obj,
|
2018-12-13 01:02:56 +08:00
|
|
|
struct switchdev_trans *trans,
|
|
|
|
|
struct netlink_ext_ack *extack));
|
2018-11-23 07:29:44 +08:00
|
|
|
int switchdev_handle_port_obj_del(struct net_device *dev,
|
|
|
|
|
struct switchdev_notifier_port_obj_info *port_obj_info,
|
|
|
|
|
bool (*check_cb)(const struct net_device *dev),
|
|
|
|
|
int (*del_cb)(struct net_device *dev,
|
|
|
|
|
const struct switchdev_obj *obj));
|
|
|
|
|
|
2019-02-28 03:44:25 +08:00
|
|
|
int switchdev_handle_port_attr_set(struct net_device *dev,
|
|
|
|
|
struct switchdev_notifier_port_attr_info *port_attr_info,
|
|
|
|
|
bool (*check_cb)(const struct net_device *dev),
|
|
|
|
|
int (*set_cb)(struct net_device *dev,
|
|
|
|
|
const struct switchdev_attr *attr,
|
|
|
|
|
struct switchdev_trans *trans));
|
2014-11-28 21:34:17 +08:00
|
|
|
#else
|
|
|
|
|
|
2015-10-15 01:40:48 +08:00
|
|
|
static inline void switchdev_deferred_process(void)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
static inline int switchdev_port_attr_set(struct net_device *dev,
|
2015-10-15 01:40:49 +08:00
|
|
|
const struct switchdev_attr *attr)
|
switchdev: introduce get/set attrs ops
Add two new swdev ops for get/set switch port attributes. Most swdev
interactions on a port are gets or sets on port attributes, so rather than
adding ops for each attribute, let's define clean get/set ops for all
attributes, and then we can have clear, consistent rules on how attributes
propagate on stacked devs.
Add the basic algorithms for get/set attr ops. Use the same recusive algo
to walk lower devs we've used for STP updates, for example. For get,
compare attr value for each lower dev and only return success if attr
values match across all lower devs. For sets, set the same attr value for
all lower devs. We'll use a two-phase prepare-commit transaction model for
sets. In the first phase, the driver(s) are asked if attr set is OK. If
all OK, the commit attr set in second phase. A driver would NACK the
prepare phase if it can't set the attr due to lack of resources or support,
within it's control. RTNL lock must be held across both phases because
we'll recurse all lower devs first in prepare phase, and then recurse all
lower devs again in commit phase. If any lower dev fails the prepare
phase, we need to abort the transaction for all lower devs.
If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to
indicate get/set only work on port (lowest) device.
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-11 00:47:48 +08:00
|
|
|
{
|
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
|
}
|
|
|
|
|
|
2015-05-11 00:47:52 +08:00
|
|
|
static inline int switchdev_port_obj_add(struct net_device *dev,
|
2018-12-13 01:02:52 +08:00
|
|
|
const struct switchdev_obj *obj,
|
|
|
|
|
struct netlink_ext_ack *extack)
|
2015-05-11 00:47:52 +08:00
|
|
|
{
|
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline int switchdev_port_obj_del(struct net_device *dev,
|
2015-10-01 17:03:45 +08:00
|
|
|
const struct switchdev_obj *obj)
|
2015-05-11 00:47:52 +08:00
|
|
|
{
|
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
|
}
|
|
|
|
|
|
2015-05-11 00:47:46 +08:00
|
|
|
static inline int register_switchdev_notifier(struct notifier_block *nb)
|
2015-01-16 06:49:36 +08:00
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-05-11 00:47:46 +08:00
|
|
|
static inline int unregister_switchdev_notifier(struct notifier_block *nb)
|
2015-01-16 06:49:36 +08:00
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-05-11 00:47:46 +08:00
|
|
|
static inline int call_switchdev_notifiers(unsigned long val,
|
|
|
|
|
struct net_device *dev,
|
2019-01-17 07:06:56 +08:00
|
|
|
struct switchdev_notifier_info *info,
|
|
|
|
|
struct netlink_ext_ack *extack)
|
2015-01-16 06:49:36 +08:00
|
|
|
{
|
|
|
|
|
return NOTIFY_DONE;
|
|
|
|
|
}
|
|
|
|
|
|
2018-11-23 07:28:25 +08:00
|
|
|
static inline int
|
|
|
|
|
register_switchdev_blocking_notifier(struct notifier_block *nb)
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline int
|
|
|
|
|
unregister_switchdev_blocking_notifier(struct notifier_block *nb)
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline int
|
|
|
|
|
call_switchdev_blocking_notifiers(unsigned long val,
|
|
|
|
|
struct net_device *dev,
|
2018-12-13 01:02:54 +08:00
|
|
|
struct switchdev_notifier_info *info,
|
|
|
|
|
struct netlink_ext_ack *extack)
|
2018-11-23 07:28:25 +08:00
|
|
|
{
|
|
|
|
|
return NOTIFY_DONE;
|
|
|
|
|
}
|
|
|
|
|
|
2018-11-23 07:29:44 +08:00
|
|
|
static inline int
|
|
|
|
|
switchdev_handle_port_obj_add(struct net_device *dev,
|
|
|
|
|
struct switchdev_notifier_port_obj_info *port_obj_info,
|
|
|
|
|
bool (*check_cb)(const struct net_device *dev),
|
|
|
|
|
int (*add_cb)(struct net_device *dev,
|
|
|
|
|
const struct switchdev_obj *obj,
|
2018-12-13 01:02:56 +08:00
|
|
|
struct switchdev_trans *trans,
|
|
|
|
|
struct netlink_ext_ack *extack))
|
2018-11-23 07:29:44 +08:00
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline int
|
|
|
|
|
switchdev_handle_port_obj_del(struct net_device *dev,
|
|
|
|
|
struct switchdev_notifier_port_obj_info *port_obj_info,
|
|
|
|
|
bool (*check_cb)(const struct net_device *dev),
|
|
|
|
|
int (*del_cb)(struct net_device *dev,
|
|
|
|
|
const struct switchdev_obj *obj))
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2019-02-28 03:44:25 +08:00
|
|
|
static inline int
|
|
|
|
|
switchdev_handle_port_attr_set(struct net_device *dev,
|
|
|
|
|
struct switchdev_notifier_port_attr_info *port_attr_info,
|
|
|
|
|
bool (*check_cb)(const struct net_device *dev),
|
|
|
|
|
int (*set_cb)(struct net_device *dev,
|
|
|
|
|
const struct switchdev_attr *attr,
|
|
|
|
|
struct switchdev_trans *trans))
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2014-11-28 21:34:17 +08:00
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
#endif /* _LINUX_SWITCHDEV_H_ */
|
