switchdev: Convert switchdev_port_obj_{add,del}() to notifiers
Subsystems (8)
cover switchdev: Convert switchdev_port_obj_{add,del}() to notifiers
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.
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.
The series starts with a clean-up patch #1, where SWITCHDEV_OBJ_PORT_{VLAN, MDB}() are fixed up to lift the constraint that the passed-in argument be a simple variable named "obj".
switchdev_port_obj_add and _del are invoked in a context that permits blocking. Not only that, at least for the VLAN notification, being able to signal failure is actually important. Therefore introduce a new blocking notifier chain that the new events will be sent on. That's done in patch #2. Retain the current (atomic) notifier chain for the preexisting notifications.
In patch #3, 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 correspond to the switchdev op arguments. 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.
In patches #4, #5, and #7, rocker, DSA resp. ethsw are updated to subscribe to the switchdev blocking notifier chain, and handle the new notifier types. #6 introduces a helper to determine whether a netdevice corresponds to a front panel port.
What these three drivers have in common is that their ports don't support any uppers besides bridge. That makes it possible to ignore any notifiers that don't reference a front-panel port device, because they are certainly out of scope.
Unlike the previous three, mlxsw and ocelot drivers admit stacked devices as uppers. While the current switchdev code recursively descends through layers of lower devices, eventually calling the op on a front-panel port device, the notifier would reference a stacking device that's one of front-panel ports uppers. The filtering is thus more complex.
For ocelot, such iteration is currently pretty much required, because there's no bookkeeping of LAG devices. mlxsw does keep the list of LAGs, however it iterates the lower devices anyway when deciding whether an event on a tunnel device pertains to the driver or not.
Therefore this patch set instead introduces, in patch #8, a helper to iterate through lowers, much like the current switchdev code does, looking for devices that match a given predicate.
Then in patches #9 and #10, first mlxsw and then ocelot are updated to dispatch the newly-added notifier types to the preexisting port_obj_add/_del handlers. The dispatch is done via the new helper, to recursively descend through lower devices.
Finally in patch #11, the actual switch is made, retiring the current SDO-based code in favor of a notifier.
Now that the event is distributed through a notifier, the explicit netdevice check in rocker, DSA and ethsw doesn't let through any events except those done on a front-panel port itself. It is therefore unnecessary to check in VLAN-handling code whether a VLAN was added to the bridge itself: such events will simply be ignored much sooner. Therefore remove it in patch #12.
Petr Machata (12):
switchdev: SWITCHDEV_OBJ_PORT_{VLAN, MDB}(): Sanitize
switchdev: Add a blocking notifier chain
switchdev: Add SWITCHDEV_PORT_OBJ_ADD, SWITCHDEV_PORT_OBJ_DEL
rocker: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
net: dsa: slave: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
staging: fsl-dpaa2: ethsw: Introduce ethsw_port_dev_check()
staging: fsl-dpaa2: ethsw: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
switchdev: Add helpers to aid traversal through lower devices
mlxsw: spectrum_switchdev: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
ocelot: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
switchdev: Replace port obj add/del SDO with a notification rocker, dsa, ethsw: Don't filter VLAN events on bridge itself
.../ethernet/mellanox/mlxsw/spectrum_switchdev.c | 47 ++++- drivers/net/ethernet/mscc/ocelot.c | 30 +++- drivers/net/ethernet/mscc/ocelot.h | 1 + drivers/net/ethernet/mscc/ocelot_board.c | 3 + drivers/net/ethernet/rocker/rocker_main.c | 60 ++++++- drivers/staging/fsl-dpaa2/ethsw/ethsw.c | 68 +++++++- include/net/switchdev.h | 87 ++++++++-- net/dsa/port.c | 3 - net/dsa/slave.c | 58 ++++++- net/switchdev/switchdev.c | 193 ++++++++++++++++----- 10 files changed, 474 insertions(+), 76 deletions(-)
Signature · 1 line
2.4.11
PATCH 1/12 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.
--- include/net/switchdev.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
include/net/switchdev.h +4 -4
@@ -95,8 +95,8 @@ struct switchdev_obj_port_vlan {u16 vid_end; }; -#define SWITCHDEV_OBJ_PORT_VLAN (obj) \ - container_of(obj, struct switchdev_obj_port_vlan, obj) +#define SWITCHDEV_OBJ_PORT_VLAN (OBJ ) \ + container_of((OBJ), struct switchdev_obj_port_vlan, obj)/* SWITCHDEV_OBJ_ID_PORT_MDB */ struct switchdev_obj_port_mdb {
@@ -105,8 +105,8 @@ struct switchdev_obj_port_mdb {u16 vid; }; -#define SWITCHDEV_OBJ_PORT_MDB (obj) \ - container_of(obj, struct switchdev_obj_port_mdb, obj) +#define SWITCHDEV_OBJ_PORT_MDB (OBJ ) \ + container_of((OBJ), struct switchdev_obj_port_mdb, obj) void switchdev_trans_item_enqueue(struct switchdev_trans *trans, void *data, void (*destructor)(void const *),
Signature · 1 line
2.4.11
PATCH 2/12 switchdev: Add a blocking notifier chain
In reply to Petr MachataIn general one can't assume that a switchdev notifier is called in a non-atomic context, and correspondingly, the switchdev notifier chain is an atomic one.
However, port object addition and deletion messages are delivered from a process context. Even the MDB addition messages, whose delivery is scheduled from atomic context, are queued and the delivery itself takes place in blocking context. For VLAN messages in particular, keeping the blocking nature is important for error reporting.
Therefore introduce a blocking notifier chain and related service functions to distribute the notifications for which a blocking context can be assumed.
--- include/net/switchdev.h | 27 +++++++++++++++++++++++++++ net/switchdev/switchdev.c | 26 ++++++++++++++++++++++++++ 2 files changed, 53 insertions(+)
include/net/switchdev.h +27 -0
@@ -182,10 +182,17 @@ int switchdev_port_obj_add(struct net_device *dev, const struct switchdev_obj *obj);int switchdev_port_obj_del (struct net_device *dev,const struct switchdev_obj *obj); +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,struct switchdev_notifier_info *info); + +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, +struct switchdev_notifier_info *info); +void switchdev_port_fwd_mark_set (struct net_device *dev,struct net_device *group_dev,bool joining);
@@ -241,6 +248,26 @@ static inline int call_switchdev_notifiers(unsigned long val,return NOTIFY_DONE ; } +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, +struct switchdev_notifier_info *info) +{ +return NOTIFY_DONE ; +} +static inline bool switchdev_port_same_parent_id (struct net_device *a,struct net_device *b) {
net/switchdev/switchdev.c +26 -0
@@ -535,6 +535,7 @@ int switchdev_port_obj_del(struct net_device *dev,EXPORT_SYMBOL_GPL (switchdev_port_obj_del);static ATOMIC_NOTIFIER_HEAD (switchdev_notif_chain); +static BLOCKING_NOTIFIER_HEAD (switchdev_blocking_notif_chain);/** * register_switchdev_notifier - Register notifier
@@ -576,6 +577,31 @@ int call_switchdev_notifiers(unsigned long val, struct net_device *dev, }EXPORT_SYMBOL_GPL (call_switchdev_notifiers); +int register_switchdev_blocking_notifier (struct notifier_block *nb) +{ +struct blocking_notifier_head *chain = &switchdev_blocking_notif_chain; + +return blocking_notifier_chain_register (chain, nb); +} +EXPORT_SYMBOL_GPL (register_switchdev_blocking_notifier); + +int unregister_switchdev_blocking_notifier (struct notifier_block *nb) +{ +struct blocking_notifier_head *chain = &switchdev_blocking_notif_chain; + +return blocking_notifier_chain_unregister (chain, nb); +} +EXPORT_SYMBOL_GPL (unregister_switchdev_blocking_notifier); + +int call_switchdev_blocking_notifiers (unsigned long val,struct net_device *dev, +struct switchdev_notifier_info *info) +{ + info->dev = dev; +return blocking_notifier_call_chain (&switchdev_blocking_notif_chain, + val, info); +} +EXPORT_SYMBOL_GPL (call_switchdev_blocking_notifiers); +bool switchdev_port_same_parent_id (struct net_device *a,struct net_device *b) {
Signature · 1 line
2.4.11
PATCH 3/12 switchdev: Add SWITCHDEV_PORT_OBJ_ADD, SWITCHDEV_PORT_OBJ_DEL
In reply to Petr MachataAn 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.
--- include/net/switchdev.h | 10 ++++++++++ 1 file changed, 10 insertions(+)
include/net/switchdev.h +10 -0
@@ -146,6 +146,9 @@ enum switchdev_notifier_type {SWITCHDEV_FDB_DEL_TO_DEVICE ,SWITCHDEV_FDB_OFFLOADED , +SWITCHDEV_PORT_OBJ_ADD ,/* Blocking. */ +SWITCHDEV_PORT_OBJ_DEL ,/* Blocking. */ +SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE ,SWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE , SWITCHDEV_VXLAN_FDB_ADD_TO_DEVICE,
@@ -165,6 +168,13 @@ struct switchdev_notifier_fdb_info { offloaded:1 ; }; +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; +}; +static inline struct net_device *switchdev_notifier_info_to_dev (const struct switchdev_notifier_info *info) {
Signature · 1 line
2.4.11
PATCH 4/12 rocker: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
In reply to Petr MachataFollowing patches will change the way of distributing port object changes from a switchdev operation to a switchdev notifier. The switchdev code currently recursively descends through layers of lower devices, eventually calling the op on a front-panel port device. The notifier will instead be sent referencing the bridge port device, which may be a stacking device that's one of front-panel ports uppers, or a completely unrelated device.
rocker currently doesn't support any uppers other than bridge. Thus the only case that a stacked device could be validly referenced by port object notifications are bridge notifications for VLAN objects added to the bridge itself. But the driver explicitly rejects such notifications in rocker_world_port_obj_vlan_add(). It is therefore safe to assume that the only interesting case is that the notification is on a front-panel port netdevice.
Subscribe to the blocking notifier chain. In the handler, filter out notifications on any foreign netdevices. Dispatch the new notifiers to rocker_port_obj_add() resp. _del() to maintain the behavior that the switchdev operation based code currently has.
--- drivers/net/ethernet/rocker/rocker_main.c | 55 +++++++++++++++++++++++++++++++ 1 file changed, 55 insertions(+)
drivers/net/ethernet/rocker/rocker_main.c +55 -0
@@ -2812,12 +2812,54 @@ static int rocker_switchdev_event(struct notifier_block *unused,return NOTIFY_DONE ; } +static int +rocker_switchdev_port_obj_event (unsigned long event,struct net_device *netdev, +struct switchdev_notifier_port_obj_info *port_obj_info) +{ +int err = -EOPNOTSUPP ; + +switch (event) { +case SWITCHDEV_PORT_OBJ_ADD : + err =rocker_port_obj_add (netdev, port_obj_info->obj, + port_obj_info->trans); +break ; +case SWITCHDEV_PORT_OBJ_DEL : + err =rocker_port_obj_del (netdev, port_obj_info->obj); +break ; + } + + port_obj_info->handled = true; +return notifier_from_errno (err); +} + +static int rocker_switchdev_blocking_event (struct notifier_block *unused, +unsigned long event,void *ptr) +{ +struct net_device *dev =switchdev_notifier_info_to_dev (ptr); + +if (!rocker_port_dev_check (dev)) +return NOTIFY_DONE ; + +switch (event) { +case SWITCHDEV_PORT_OBJ_ADD : +case SWITCHDEV_PORT_OBJ_DEL : +return rocker_switchdev_port_obj_event (event, dev, ptr); + } + +return NOTIFY_DONE ; +} +static struct notifier_block rocker_switchdev_notifier = { .notifier_call = rocker_switchdev_event, }; +static struct notifier_block rocker_switchdev_blocking_notifier = { + .notifier_call = rocker_switchdev_blocking_event, +}; +static int rocker_probe (struct pci_dev *pdev,const struct pci_device_id *id) { +struct notifier_block *nb;struct rocker *rocker;int err;
@@ -2933,6 +2975,13 @@ static int rocker_probe(struct pci_dev *pdev, const struct pci_device_id *id) goto err_register_switchdev_notifier; } + nb = &rocker_switchdev_blocking_notifier; + err =register_switchdev_blocking_notifier (nb); +if (err) { +dev_err (&pdev->dev,"Failed to register switchdev blocking notifier\n" ); + goto err_register_switchdev_blocking_notifier; + } + rocker->hw.id =rocker_read64 (rocker,SWITCH_ID ); dev_info(&pdev->dev, "Rocker switch with id %*phN\n",
@@ -2940,6 +2989,8 @@ static int rocker_probe(struct pci_dev *pdev, const struct pci_device_id *id)return 0 ; +err_register_switchdev_blocking_notifier: +unregister_switchdev_notifier (&rocker_switchdev_notifier); err_register_switchdev_notifier:unregister_fib_notifier (&rocker->fib_nb); err_register_fib_notifier:
@@ -2971,6 +3022,10 @@ static int rocker_probe(struct pci_dev *pdev, const struct pci_device_id *id)static void rocker_remove (struct pci_dev *pdev) {struct rocker *rocker =pci_get_drvdata (pdev); +struct notifier_block *nb; + + nb = &rocker_switchdev_blocking_notifier; +unregister_switchdev_blocking_notifier (nb);unregister_switchdev_notifier (&rocker_switchdev_notifier);unregister_fib_notifier (&rocker->fib_nb);
Signature · 1 line
2.4.11
PATCH 5/12 net: dsa: slave: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
In reply to Petr MachataFollowing patches will change the way of distributing port object changes from a switchdev operation to a switchdev notifier. The switchdev code currently recursively descends through layers of lower devices, eventually calling the op on a front-panel port device. The notifier will instead be sent referencing the bridge port device, which may be a stacking device that's one of front-panel ports uppers, or a completely unrelated device.
DSA currently doesn't support any other uppers than bridge. SWITCHDEV_OBJ_ID_HOST_MDB and _PORT_MDB objects are always notified on the bridge port device. Thus the only case that a stacked device could be validly referenced by port object notifications are bridge notifications for VLAN objects added to the bridge itself. But the driver explicitly rejects such notifications in dsa_port_vlan_add(). It is therefore safe to assume that the only interesting case is that the notification is on a front-panel port netdevice. Therefore keep the filtering by dsa_slave_dev_check() in place.
To handle SWITCHDEV_PORT_OBJ_ADD and _DEL, subscribe to the blocking notifier chain. Dispatch to rocker_port_obj_add() resp. _del() to maintain the behavior that the switchdev operation based code currently has.
--- net/dsa/slave.c | 56 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 56 insertions(+)
net/dsa/slave.c +56 -0
@@ -1557,6 +1557,44 @@ static int dsa_slave_switchdev_event(struct notifier_block *unused,return NOTIFY_BAD ; } +static int +dsa_slave_switchdev_port_obj_event (unsigned long event, +struct net_device *netdev, +struct switchdev_notifier_port_obj_info *port_obj_info) +{ +int err = -EOPNOTSUPP ; + +switch (event) { +case SWITCHDEV_PORT_OBJ_ADD : + err =dsa_slave_port_obj_add (netdev, port_obj_info->obj, + port_obj_info->trans); +break ; +case SWITCHDEV_PORT_OBJ_DEL : + err =dsa_slave_port_obj_del (netdev, port_obj_info->obj); +break ; + } + + port_obj_info->handled = true; +return notifier_from_errno (err); +} + +static int dsa_slave_switchdev_blocking_event (struct notifier_block *unused, +unsigned long event,void *ptr) +{ +struct net_device *dev =switchdev_notifier_info_to_dev (ptr); + +if (!dsa_slave_dev_check (dev)) +return NOTIFY_DONE ; + +switch (event) { +case SWITCHDEV_PORT_OBJ_ADD :/* fall through */ +case SWITCHDEV_PORT_OBJ_DEL : +return dsa_slave_switchdev_port_obj_event (event, dev, ptr); + } + +return NOTIFY_DONE ; +} +static struct notifier_block dsa_slave_nb__read_mostly = { .notifier_call = dsa_slave_netdevice_event, };
@@ -1565,8 +1603,13 @@ static struct notifier_block dsa_slave_switchdev_notifier = { .notifier_call = dsa_slave_switchdev_event, }; +static struct notifier_block dsa_slave_switchdev_blocking_notifier = { + .notifier_call = dsa_slave_switchdev_blocking_event, +}; +int dsa_slave_register_notifier (void ) { +struct notifier_block *nb;int err; err =register_netdevice_notifier (&dsa_slave_nb);
@@ -1577,8 +1620,15 @@ int dsa_slave_register_notifier(void)if (err) goto err_switchdev_nb; + nb = &dsa_slave_switchdev_blocking_notifier; + err =register_switchdev_blocking_notifier (nb); +if (err) + goto err_switchdev_blocking_nb; +return 0 ; +err_switchdev_blocking_nb: +unregister_switchdev_notifier (&dsa_slave_switchdev_notifier); err_switchdev_nb:unregister_netdevice_notifier (&dsa_slave_nb);return err;
@@ -1586,8 +1636,14 @@ int dsa_slave_register_notifier(void)void dsa_slave_unregister_notifier (void ) { +struct notifier_block *nb;int err; + nb = &dsa_slave_switchdev_blocking_notifier; + err =unregister_switchdev_blocking_notifier (nb); +if (err) +pr_err ("DSA: failed to unregister switchdev blocking notifier (%d)\n" , err); + err =unregister_switchdev_notifier (&dsa_slave_switchdev_notifier);if (err)pr_err ("DSA: failed to unregister switchdev notifier (%d)\n" , err);
Signature · 1 line
2.4.11
PATCH 6/12 staging: fsl-dpaa2: ethsw: Introduce ethsw_port_dev_check()
In reply to Petr Machataethsw currently uses an open-coded comparison of netdev_ops to determine whether whether a device represents a front panel port. Wrap this into a named function to simplify reuse.
--- drivers/staging/fsl-dpaa2/ethsw/ethsw.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/staging/fsl-dpaa2/ethsw/ethsw.c +6 -1
@@ -972,6 +972,11 @@ static int port_bridge_leave(struct net_device *netdev)return err; } +static bool ethsw_port_dev_check (const struct net_device *netdev) +{ +return netdev->netdev_ops == ðsw_port_ops; +} +static int port_netdevice_event (struct notifier_block *unused,unsigned long event,void *ptr) {
@@ -980,7 +985,7 @@ static int port_netdevice_event(struct notifier_block *unused,struct net_device *upper_dev;int err =0 ; -if (netdev->netdev_ops != ðsw_port_ops) +if (!ethsw_port_dev_check (netdev))return NOTIFY_DONE ;/* Handle just upper dev link/unlink for the moment */
Signature · 1 line
2.4.11
PATCH 7/12 staging: fsl-dpaa2: ethsw: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
In reply to Petr MachataFollowing patches will change the way of distributing port object changes from a switchdev operation to a switchdev notifier. The switchdev code currently recursively descends through layers of lower devices, eventually calling the op on a front-panel port device. The notifier will instead be sent referencing the bridge port device, which may be a stacking device that's one of front-panel ports uppers, or a completely unrelated device.
ethsw currently doesn't support any uppers other than bridge. SWITCHDEV_OBJ_ID_HOST_MDB and _PORT_MDB objects are always notified on the bridge port device. Thus the only case that a stacked device could be validly referenced by port object notifications are bridge notifications for VLAN objects added to the bridge itself. But the driver explicitly rejects such notifications in port_vlans_add(). It is therefore safe to assume that the only interesting case is that the notification is on a front-panel port netdevice.
To handle SWITCHDEV_PORT_OBJ_ADD and _DEL, subscribe to the blocking notifier chain. Dispatch to swdev_port_obj_add() resp. _del() to maintain the behavior that the switchdev operation based code currently has.
--- drivers/staging/fsl-dpaa2/ethsw/ethsw.c | 56 +++++++++++++++++++++++++++++++++ 1 file changed, 56 insertions(+)
drivers/staging/fsl-dpaa2/ethsw/ethsw.c +56 -0
@@ -1088,10 +1088,51 @@ static int port_switchdev_event(struct notifier_block *unused,return NOTIFY_BAD ; } +static int +ethsw_switchdev_port_obj_event (unsigned long event,struct net_device *netdev, +struct switchdev_notifier_port_obj_info *port_obj_info) +{ +int err = -EOPNOTSUPP ; + +switch (event) { +case SWITCHDEV_PORT_OBJ_ADD : + err =swdev_port_obj_add (netdev, port_obj_info->obj, + port_obj_info->trans); +break ; +case SWITCHDEV_PORT_OBJ_DEL : + err =swdev_port_obj_del (netdev, port_obj_info->obj); +break ; + } + + port_obj_info->handled = true; +return notifier_from_errno (err); +} + +static int port_switchdev_blocking_event (struct notifier_block *unused, +unsigned long event,void *ptr) +{ +struct net_device *dev =switchdev_notifier_info_to_dev (ptr); + +if (!ethsw_port_dev_check (dev)) +return NOTIFY_DONE ; + +switch (event) { +case SWITCHDEV_PORT_OBJ_ADD :/* fall through */ +case SWITCHDEV_PORT_OBJ_DEL : +return ethsw_switchdev_port_obj_event (event, dev, ptr); + } + +return NOTIFY_DONE ; +} +static struct notifier_block port_switchdev_nb = { .notifier_call = port_switchdev_event, }; +static struct notifier_block port_switchdev_blocking_nb = { + .notifier_call = port_switchdev_blocking_event, +}; +static int ethsw_register_notifier (struct device *dev) {int err;
@@ -1108,8 +1149,16 @@ static int ethsw_register_notifier(struct device *dev) goto err_switchdev_nb; } + err =register_switchdev_blocking_notifier (&port_switchdev_blocking_nb); +if (err) { +dev_err (dev,"Failed to register switchdev blocking notifier\n" ); + goto err_switchdev_blocking_nb; + } +return 0 ; +err_switchdev_blocking_nb: +unregister_switchdev_notifier (&port_switchdev_nb); err_switchdev_nb:unregister_netdevice_notifier (&port_nb);return err;
@@ -1296,8 +1345,15 @@ static int ethsw_port_init(struct ethsw_port_priv *port_priv, u16 port)static void ethsw_unregister_notifier (struct device *dev) { +struct notifier_block *nb;int err; + nb = &port_switchdev_blocking_nb; + err =unregister_switchdev_blocking_notifier (nb); +if (err) +dev_err (dev, +"Failed to unregister switchdev blocking notifier (%d)\n" , err); + err =unregister_switchdev_notifier (&port_switchdev_nb);if (err) dev_err(dev,
Signature · 1 line
2.4.11
PATCH 8/12 switchdev: Add helpers to aid traversal through lower devices
In reply to Petr MachataAfter the transition from switchdev operations to notifier chain (which will take place in following patches), the onus is on the driver to find its own devices below possible layer of LAG or other uppers.
The logic to do so is fairly repetitive: each driver is looking for its own devices among the lowers of the notified device. For those that it finds, it calls a handler. To indicate that the event was handled, struct switchdev_notifier_port_obj_info.handled is set. The differences lie only in what constitutes an "own" device and what handler to call.
Therefore abstract this logic into two helpers, switchdev_handle_port_obj_add() and switchdev_handle_port_obj_del(). If a driver only supports physical ports under a bridge device, it will simply avoid this layer of indirection.
One area where this helper diverges from the current switchdev behavior is the case of mixed lowers, some of which are switchdev ports and some of which are not. Previously, such scenario would fail with -EOPNOTSUPP. The helper could do that for lowers for which the passed-in predicate doesn't hold. That would however break the case that switchdev ports from several different drivers are stashed under one master, a scenario that switchdev currently happily supports. Therefore tolerate any and all unknown netdevices, whether they are backed by a switchdev driver or not.
--- include/net/switchdev.h | 33 +++++++++++++++ net/switchdev/switchdev.c | 100 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 133 insertions(+)
include/net/switchdev.h +33 -0
@@ -210,6 +210,18 @@ void switchdev_port_fwd_mark_set(struct net_device *dev,bool switchdev_port_same_parent_id (struct net_device *a,struct net_device *b); +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, +struct switchdev_trans *trans)); +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)); +#define SWITCHDEV_SET_OPS (netdev, ops) ((netdev)->switchdev_ops = (ops))#else
@@ -284,6 +296,27 @@ static inline bool switchdev_port_same_parent_id(struct net_device *a, return false; } +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, +struct switchdev_trans *trans)) +{ +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 ; +} +#define SWITCHDEV_SET_OPS (netdev, ops) do {} while (0)#endif
net/switchdev/switchdev.c +100 -0
103 lines not on this page · raw message
Signature · 1 line
2.4.11
PATCH 9/12 mlxsw: spectrum_switchdev: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
In reply to Petr MachataFollowing patches will change the way of distributing port object changes from a switchdev operation to a switchdev notifier. The switchdev code currently recursively descends through layers of lower devices, eventually calling the op on a front-panel port device. The notifier will instead be sent referencing the bridge port device, which may be a stacking device that's one of front-panel ports uppers, or a completely unrelated device.
To handle SWITCHDEV_PORT_OBJ_ADD and _DEL, subscribe to the blocking notifier chain. Dispatch to mlxsw_sp_port_obj_add() resp. _del() to maintain the behavior that the switchdev operation based code currently has. Defer to switchdev_handle_port_obj_add() / _del() to handle the recursive descend, because mlxsw supports a number of upper types.
--- .../ethernet/mellanox/mlxsw/spectrum_switchdev.c | 45 +++++++++++++++++++++- 1 file changed, 44 insertions(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c +44 -1
@@ -3118,6 +3118,32 @@ static struct notifier_block mlxsw_sp_switchdev_notifier = { .notifier_call = mlxsw_sp_switchdev_event, }; +static int mlxsw_sp_switchdev_blocking_event (struct notifier_block *unused, +unsigned long event,void *ptr) +{ +struct net_device *dev =switchdev_notifier_info_to_dev (ptr); +int err; + +switch (event) { +case SWITCHDEV_PORT_OBJ_ADD : + err =switchdev_handle_port_obj_add (dev, ptr, + mlxsw_sp_port_dev_check, + mlxsw_sp_port_obj_add); +return notifier_from_errno (err); +case SWITCHDEV_PORT_OBJ_DEL : + err =switchdev_handle_port_obj_del (dev, ptr, + mlxsw_sp_port_dev_check, + mlxsw_sp_port_obj_del); +return notifier_from_errno (err); + } + +return NOTIFY_DONE ; +} + +static struct notifier_block mlxsw_sp_switchdev_blocking_notifier = { + .notifier_call = mlxsw_sp_switchdev_blocking_event, +}; +u8 mlxsw_sp_bridge_port_stp_state (struct mlxsw_sp_bridge_port *bridge_port) {
@@ -3127,6 +3153,7 @@ mlxsw_sp_bridge_port_stp_state(struct mlxsw_sp_bridge_port *bridge_port)static int mlxsw_sp_fdb_init (struct mlxsw_sp *mlxsw_sp) {struct mlxsw_sp_bridge *bridge = mlxsw_sp->bridge; +struct notifier_block *nb;int err; err =mlxsw_sp_ageing_set (mlxsw_sp,MLXSW_SP_DEFAULT_AGEING_TIME );
@@ -3141,17 +3168,33 @@ static int mlxsw_sp_fdb_init(struct mlxsw_sp *mlxsw_sp)return err; } + nb = &mlxsw_sp_switchdev_blocking_notifier; + err =register_switchdev_blocking_notifier (nb); +if (err) { +dev_err (mlxsw_sp->bus_info->dev,"Failed to register switchdev blocking notifier\n" ); + goto err_register_switchdev_blocking_notifier; + } +INIT_DELAYED_WORK (&bridge->fdb_notify.dw, mlxsw_sp_fdb_notify_work); bridge->fdb_notify.interval =MLXSW_SP_DEFAULT_LEARNING_INTERVAL ;mlxsw_sp_fdb_notify_work_schedule (mlxsw_sp);return 0 ; + +err_register_switchdev_blocking_notifier: +unregister_switchdev_notifier (&mlxsw_sp_switchdev_notifier); +return err; }static void mlxsw_sp_fdb_fini (struct mlxsw_sp *mlxsw_sp) { +struct notifier_block *nb; +cancel_delayed_work_sync (&mlxsw_sp->bridge->fdb_notify.dw); -unregister_switchdev_notifier (&mlxsw_sp_switchdev_notifier); + nb = &mlxsw_sp_switchdev_blocking_notifier; +unregister_switchdev_blocking_notifier (nb); + +unregister_switchdev_notifier (&mlxsw_sp_switchdev_notifier); }int mlxsw_sp_switchdev_init (struct mlxsw_sp *mlxsw_sp)
Signature · 1 line
2.4.11
PATCH 10/12 ocelot: Handle SWITCHDEV_PORT_OBJ_ADD/_DEL
In reply to Petr MachataFollowing patches will change the way of distributing port object changes from a switchdev operation to a switchdev notifier. The switchdev code currently recursively descends through layers of lower devices, eventually calling the op on a front-panel port device. The notifier will instead be sent referencing the bridge port device, which may be a stacking device that's one of front-panel ports uppers, or a completely unrelated device.
Dispatch the new events to ocelot_port_obj_add() resp. _del() to maintain the same behavior that the switchdev operation based code currently has. Pass through switchdev_handle_port_obj_add() / _del() to handle the recursive descend, because Ocelot supports LAG uppers.
Register to the new switchdev blocking notifier chain to get the new events when they start getting distributed.
--- drivers/net/ethernet/mscc/ocelot.c | 28 ++++++++++++++++++++++++++++ drivers/net/ethernet/mscc/ocelot.h | 1 + drivers/net/ethernet/mscc/ocelot_board.c | 3 +++ 3 files changed, 32 insertions(+)
drivers/net/ethernet/mscc/ocelot.c +28 -0
@@ -1595,6 +1595,34 @@ struct notifier_block ocelot_netdevice_nb __read_mostly = { };EXPORT_SYMBOL (ocelot_netdevice_nb); +static int ocelot_switchdev_blocking_event (struct notifier_block *unused, +unsigned long event,void *ptr) +{ +struct net_device *dev =switchdev_notifier_info_to_dev (ptr); +int err; + +switch (event) { +/* Blocking events. */ +case SWITCHDEV_PORT_OBJ_ADD : + err =switchdev_handle_port_obj_add (dev, ptr, + ocelot_netdevice_dev_check, + ocelot_port_obj_add); +return notifier_from_errno (err); +case SWITCHDEV_PORT_OBJ_DEL : + err =switchdev_handle_port_obj_del (dev, ptr, + ocelot_netdevice_dev_check, + ocelot_port_obj_del); +return notifier_from_errno (err); + } + +return NOTIFY_DONE ; +} + +struct notifier_block ocelot_switchdev_blocking_nb__read_mostly = { + .notifier_call = ocelot_switchdev_blocking_event, +}; +EXPORT_SYMBOL (ocelot_switchdev_blocking_nb); +int ocelot_probe_port (struct ocelot *ocelot,u8 port,void __iomem *regs,struct phy_device *phy)
drivers/net/ethernet/mscc/ocelot.h +1 -0
@@ -499,5 +499,6 @@ int ocelot_probe_port(struct ocelot *ocelot, u8 port, struct phy_device *phy);extern struct notifier_block ocelot_netdevice_nb; +extern struct notifier_block ocelot_switchdev_blocking_nb;#endif
drivers/net/ethernet/mscc/ocelot_board.c +3 -0
@@ -12,6 +12,7 @@#include <linux/of_platform.h> #include <linux/mfd/syscon.h> #include <linux/skbuff.h> +#include <net/switchdev.h> #include "ocelot.h"
@@ -328,6 +329,7 @@ static int mscc_ocelot_probe(struct platform_device *pdev) }register_netdevice_notifier (&ocelot_netdevice_nb); +register_switchdev_blocking_notifier (&ocelot_switchdev_blocking_nb);dev_info (&pdev->dev,"Ocelot switch probed\n" );
@@ -342,6 +344,7 @@ static int mscc_ocelot_remove(struct platform_device *pdev)struct ocelot *ocelot =platform_get_drvdata (pdev);ocelot_deinit (ocelot); +unregister_switchdev_blocking_notifier (&ocelot_switchdev_blocking_nb);unregister_netdevice_notifier (&ocelot_netdevice_nb);return 0 ;
Signature · 1 line
2.4.11
PATCH 11/12 switchdev: Replace port obj add/del SDO with a notification
In reply to Petr MachataDrop switchdev_ops.switchdev_port_obj_add and _del. Drop the uses of this field from all clients, which were migrated to use switchdev notification in the previous patches.
Add a new function switchdev_port_obj_notify() that sends the switchdev notifications SWITCHDEV_PORT_OBJ_ADD and _DEL.
Update switchdev_port_obj_del_now() to dispatch to this new function. Drop __switchdev_port_obj_add() and update switchdev_port_obj_add() likewise.
--- .../ethernet/mellanox/mlxsw/spectrum_switchdev.c | 2 - drivers/net/ethernet/mscc/ocelot.c | 2 - drivers/net/ethernet/rocker/rocker_main.c | 2 - drivers/staging/fsl-dpaa2/ethsw/ethsw.c | 2 - include/net/switchdev.h | 9 --- net/dsa/slave.c | 2 - net/switchdev/switchdev.c | 67 ++++++++-------------- 7 files changed, 25 insertions(+), 61 deletions(-)
drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c +0 -2
@@ -1968,8 +1968,6 @@ static struct mlxsw_sp_port *mlxsw_sp_lag_rep_port(struct mlxsw_sp *mlxsw_sp,static const struct switchdev_ops mlxsw_sp_port_switchdev_ops = { .switchdev_port_attr_get = mlxsw_sp_port_attr_get, .switchdev_port_attr_set = mlxsw_sp_port_attr_set, - .switchdev_port_obj_add = mlxsw_sp_port_obj_add, - .switchdev_port_obj_del = mlxsw_sp_port_obj_del, }; static int
drivers/net/ethernet/mscc/ocelot.c +0 -2
@@ -1337,8 +1337,6 @@ static int ocelot_port_obj_del(struct net_device *dev,static const struct switchdev_ops ocelot_port_switchdev_ops = { .switchdev_port_attr_get = ocelot_port_attr_get, .switchdev_port_attr_set = ocelot_port_attr_set, - .switchdev_port_obj_add = ocelot_port_obj_add, - .switchdev_port_obj_del = ocelot_port_obj_del, }; static int ocelot_port_bridge_join(struct ocelot_port *ocelot_port,
drivers/net/ethernet/rocker/rocker_main.c +0 -2
@@ -2145,8 +2145,6 @@ static int rocker_port_obj_del(struct net_device *dev,static const struct switchdev_ops rocker_port_switchdev_ops = { .switchdev_port_attr_get = rocker_port_attr_get, .switchdev_port_attr_set = rocker_port_attr_set, - .switchdev_port_obj_add = rocker_port_obj_add, - .switchdev_port_obj_del = rocker_port_obj_del, };struct rocker_fib_event_work {
drivers/staging/fsl-dpaa2/ethsw/ethsw.c +0 -2
@@ -930,8 +930,6 @@ static int swdev_port_obj_del(struct net_device *netdev,static const struct switchdev_ops ethsw_port_switchdev_ops = { .switchdev_port_attr_get = swdev_port_attr_get, .switchdev_port_attr_set = swdev_port_attr_set, - .switchdev_port_obj_add = swdev_port_obj_add, - .switchdev_port_obj_del = swdev_port_obj_del, };/* For the moment, only flood setting needs to be updated */
include/net/switchdev.h +0 -9
@@ -121,10 +121,6 @@ typedef int switchdev_obj_dump_cb_t(struct switchdev_obj *obj);* @switchdev_port_attr_get: Get a port attribute (see switchdev_attr). * * @switchdev_port_attr_set: Set a port attribute (see switchdev_attr). -* -* @switchdev_port_obj_add: Add an object to port (see switchdev_obj_*). -* -* @switchdev_port_obj_del: Delete an object from port (see switchdev_obj_*). */ struct switchdev_ops { int (*switchdev_port_attr_get)(struct net_device *dev,
@@ -132,11 +128,6 @@ struct switchdev_ops {int (*switchdev_port_attr_set)(struct net_device *dev,const struct switchdev_attr *attr,struct switchdev_trans *trans); -int (*switchdev_port_obj_add)(struct net_device *dev, -const struct switchdev_obj *obj, -struct switchdev_trans *trans); -int (*switchdev_port_obj_del)(struct net_device *dev, -const struct switchdev_obj *obj); };enum switchdev_notifier_type {
net/dsa/slave.c +0 -2
@@ -1050,8 +1050,6 @@ static const struct net_device_ops dsa_slave_netdev_ops = {static const struct switchdev_ops dsa_slave_switchdev_ops = { .switchdev_port_attr_get = dsa_slave_port_attr_get, .switchdev_port_attr_set = dsa_slave_port_attr_set, - .switchdev_port_obj_add = dsa_slave_port_obj_add, - .switchdev_port_obj_del = dsa_slave_port_obj_del, };static struct device_type dsa_type = {
net/switchdev/switchdev.c +25 -42
95 lines not on this page · raw message
Signature · 1 line
2.4.11
PATCH 12/12 rocker, dsa, ethsw: Don't filter VLAN events on bridge itself
In reply to Petr MachataDue to an explicit check in rocker_world_port_obj_vlan_add(), dsa_slave_switchdev_event() resp. port_switchdev_event(), VLAN objects that are added to a device that is not a front-panel port device are ignored. Therefore this check is immaterial.
--- drivers/net/ethernet/rocker/rocker_main.c | 3 --- drivers/staging/fsl-dpaa2/ethsw/ethsw.c | 3 --- net/dsa/port.c | 3 --- 3 files changed, 9 deletions(-)
drivers/net/ethernet/rocker/rocker_main.c +0 -3
@@ -1632,9 +1632,6 @@ rocker_world_port_obj_vlan_add(struct rocker_port *rocker_port, {struct rocker_world_ops *wops = rocker_port->rocker->wops; -if (netif_is_bridge_master (vlan->obj.orig_dev)) -return -EOPNOTSUPP ; -if (!wops->port_obj_vlan_add)return -EOPNOTSUPP ;
drivers/staging/fsl-dpaa2/ethsw/ethsw.c +0 -3
@@ -719,9 +719,6 @@ static int port_vlans_add(struct net_device *netdev,struct ethsw_port_priv *port_priv =netdev_priv (netdev);int vid, err =0 ; -if (netif_is_bridge_master (vlan->obj.orig_dev)) -return -EOPNOTSUPP ; -if (switchdev_trans_ph_prepare (trans))return 0 ;
net/dsa/port.c +0 -3
@@ -252,9 +252,6 @@ int dsa_port_vlan_add(struct dsa_port *dp, .vlan = vlan, }; -if (netif_is_bridge_master (vlan->obj.orig_dev)) -return -EOPNOTSUPP ; -if (br_vlan_enabled (dp->bridge_dev))return dsa_port_notify (dp,DSA_NOTIFIER_VLAN_ADD , &info);
Signature · 1 line
2.4.11
Quoted 7 lines · Petr Machata
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.
mlxsw will use these notifications to offload VXLAN devices attached to a VLAN-aware bridge. The patches are available here should anyone wish to take a look:
https://github.com/idosch/linux/commits/vxlan
Thanks,
Petr
David Miller
maintainerFrom: Petr Machata <[email protected]>
Date: Thu, 22 Nov 2018 23:27:52 +0000
Quoted 8 lines
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.
Quoted 8 lines
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.
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.
...
Series applied, thank you.