net: bridge: add flush filtering support
Subsystems (4)
Nikolay Aleksandrov
maintainerv2 cover net: bridge: add flush filtering support
Hi,
This patch-set adds support to specify filtering conditions for a flush operation. This version has entirely different entry point (v1 had bridge-specific IFLA attribute, here I add new RTM_FLUSHNEIGH msg and netdev ndo_fdb_flush op) so I'll give a new overview altogether. After me and Ido discussed the feature offlist, we agreed that it would be best to add a new generic RTM_FLUSHNEIGH with a new ndo_fdb_flush callback which can be re-used for other drivers (e.g. vxlan). Patch 01 adds the new RTM_FLUSHNEIGH type, patch 02 then adds the new ndo_fdb_flush call. With this structure we need to add a generic rtnl_fdb_flush which will be used to do basic attribute validation and dispatch the call to the appropriate device based on the NTF_USE/MASTER flags (patch 03). Patch 04 then adds some common flush attributes which are used by the bridge and vxlan drivers (target ifindex, vlan id, ndm flags/state masks) with basic attribute validation, further validation can be done by the implementers of the ndo callback. Patch 05 adds a minimal ndo_fdb_flush to the bridge driver, it uses the current br_fdb_flush implementation to flush all entries similar to existing calls. Patch 06 adds filtering support to the new bridge flush op which supports target ifindex (port or bridge), vlan id and flags/state mask. Patch 07 converts ndm state/flags and their masks to bridge-private flags and fills them in the filter descriptor for matching. Finally patch 08 fills in the target ifindex (after validating it) and vlan id (already validated by rtnl_fdb_flush) for matching. Flush filtering is needed because user-space applications need a quick way to delete only a specific set of entries, e.g. mlag implementations need a way to flush only dynamic entries excluding externally learned ones or only externally learned ones without static entries etc. Also apps usually want to target only a specific vlan or port/vlan combination. The current 2 flush operations (per port and bridge-wide) are not extensible and cannot provide such filtering.
I decided against embedding new attrs into the old flush attributes for multiple reasons - proper error handling on unsupported attributes, older kernels silently flushing all, need for a second mechanism to signal that the attribute should be parsed (e.g. using boolopts), special treatment for permanent entries.
Examples:
$ bridge fdb flush dev bridge vlan 100 static < flush all static entries on vlan 100 > $ bridge fdb flush dev bridge vlan 1 dynamic < flush all dynamic entries on vlan 1 > $ bridge fdb flush dev bridge port ens16 vlan 1 dynamic < flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev ens16 vlan 1 dynamic master < as above: flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev bridge nooffloaded nopermanent self < flush all non-offloaded and non-permanent entries > $ bridge fdb flush dev bridge static noextern_learn < flush all static entries which are not externally learned > $ bridge fdb flush dev bridge permanent < flush all permanent entries > $ bridge fdb flush dev bridge port bridge permanent < flush all permanent entries pointing to the bridge itself >
Note that all flags have their negated version (static vs nostatic etc) and there are some tricky cases to handle like "static" which in flag terms means fdbs that have NUD_NOARP but *not* NUD_PERMANENT, so the mask matches on both but we need only NUD_NOARP to be set. That's because permanent entries have both set so we can't just match on NUD_NOARP. Also note that this flush operation doesn't treat permanent entries in a special way (fdb_delete vs fdb_delete_local), it will delete them regardless if any port is using them. We can extend the api with a flag to do that if needed in the future.
Patch-sets (in order):
- Initial flush infra and fdb flush filtering (this set)
- iproute2 support
- selftests
Future work:
- mdb flush support (RTM_FLUSHMDB type)
Thanks to Ido for the great discussion and feedback while working on this.
Thanks,
Nik
Nikolay Aleksandrov (8):
net: rtnetlink: add RTM_FLUSHNEIGH
net: add ndo_fdb_flush op
net: bridge: fdb: add ndo_fdb_flush op
net: rtnetlink: register a generic rtnl_fdb_flush call
net: rtnetlink: add common flush attributes
net: bridge: fdb: add support for fine-grained flushing
net: bridge: fdb: add support for flush filtering based on ndm flags
and state
net: bridge: fdb: add support for flush filtering based on ifindex and
vlan
include/linux/netdevice.h | 11 +++ include/uapi/linux/neighbour.h | 10 +++ include/uapi/linux/rtnetlink.h | 3 + net/bridge/br_device.c | 1 + net/bridge/br_fdb.c | 154 +++++++++++++++++++++++++++++++-- net/bridge/br_netlink.c | 9 +- net/bridge/br_private.h | 19 +++- net/bridge/br_sysfs_br.c | 6 +- net/core/rtnetlink.c | 62 +++++++++++++ security/selinux/nlmsgtab.c | 3 +- 10 files changed, 266 insertions(+), 12 deletions(-)
Signature · 1 line
2.35.1
Nikolay Aleksandrov
maintainerPATCH v2 1/8 net: rtnetlink: add RTM_FLUSHNEIGH
Add a new rtnetlink type used to flush neigh objects. It will be initially used to add flush with filtering support for bridge fdbs, but it also opens the door to add similar support to others (e.g. vxlan).
--- include/uapi/linux/rtnetlink.h | 3 +++ security/selinux/nlmsgtab.c | 3 ++- 2 files changed, 5 insertions(+), 1 deletion(-)
include/uapi/linux/rtnetlink.h +3 -0
@@ -194,6 +194,9 @@ enum { RTM_GETTUNNEL,#define RTM_GETTUNNEL RTM_GETTUNNEL + RTM_FLUSHNEIGH = 124, +#define RTM_FLUSHNEIGH RTM_FLUSHNEIGH + __RTM_MAX,#define RTM_MAX (((__RTM_MAX + 3) & ~3) - 1) };
security/selinux/nlmsgtab.c +2 -1
@@ -95,6 +95,7 @@ static const struct nlmsg_perm nlmsg_route_perms[] = { RTM_NEWTUNNEL, NETLINK_ROUTE_SOCKET__NLMSG_WRITE }, { RTM_DELTUNNEL, NETLINK_ROUTE_SOCKET__NLMSG_WRITE }, { RTM_GETTUNNEL, NETLINK_ROUTE_SOCKET__NLMSG_READ }, + { RTM_FLUSHNEIGH, NETLINK_ROUTE_SOCKET__NLMSG_WRITE }, }; static const struct nlmsg_perm nlmsg_tcpdiag_perms[] =
@@ -180,7 +181,7 @@ int selinux_nlmsg_lookup(u16 sclass, u16 nlmsg_type, u32 *perm)* structures at the top of this file with the new mappings * before updating the BUILD_BUG_ON() macro! */ -BUILD_BUG_ON (RTM_MAX != (RTM_NEWTUNNEL +3 )); +BUILD_BUG_ON (RTM_MAX != (RTM_FLUSHNEIGH +3 )); err =nlmsg_perm (nlmsg_type, perm, nlmsg_route_perms,sizeof (nlmsg_route_perms));break ;
Signature · 1 line
2.35.1
Re: [PATCH net-next v2 1/8] net: rtnetlink: add RTM_FLUSHNEIGH
Quoted 15 lines · Nikolay Aleksandrov
Add a new rtnetlink type used to flush neigh objects. It will be initially used to add flush with filtering support for bridge fdbs, but it also opens the door to add similar support to others (e.g. vxlan).
Signed-off-by Nikolay Aleksandrov <[email protected]>--- include/uapi/linux/rtnetlink.h | 3 +++ security/selinux/nlmsgtab.c | 3 ++- 2 files changed, 5 insertions(+), 1 deletion(-)include/uapi/linux/rtnetlink.h +0
-0
include/uapi/linux/rtnetlink.h enum { · line 199
Quoted 6 lines
include/uapi/linux/rtnetlink.h +3
-0@@ -194,6 +194,9 @@ enum { RTM_GETTUNNEL,#define RTM_GETTUNNEL RTM_GETTUNNEL + RTM_FLUSHNEIGH = 124, +#define RTM_FLUSHNEIGH RTM_FLUSHNEIGH +
rtm message types are "new, del, get, set" quadruplets; making this a flush breaks the current consistent style. Can this be done by adding a FLUSH flag to the RTM_DELNEIGH message?
Quoted 3 lines
include/uapi/linux/rtnetlink.h +0
-0__RTM_MAX,#define RTM_MAX (((__RTM_MAX + 3) & ~3) - 1) };
Nikolay Aleksandrov
maintainerPATCH v2 2/8 net: add ndo_fdb_flush op
In reply to Nikolay AleksandrovAdd a new netdev op called ndo_fdb_flush, it will be later used for driver-specific flush implementation dispatched from rtnetlink. The first user will be the bridge.
--- include/linux/netdevice.h | 11 +++++++++++ 1 file changed, 11 insertions(+)
include/linux/netdevice.h +11 -0
@@ -1265,6 +1265,12 @@ struct netdev_net_notifier {* int *idx) * Used to add FDB entries to dump requests. Implementers should add * entries to skb and update idx with the number of entries. +* int (*ndo_fdb_flush)(struct ndmsg *ndm, struct nlattr *tb[], +* struct net_device *dev, +* u16 vid, +* struct netlink_ext_ack *extack); +* Used to flush FDB entries. Filter attributes can be specified to delete +* only matching FDB entries if implementers support it. * * int (*ndo_bridge_setlink)(struct net_device *dev, struct nlmsghdr *nlh, * u16 flags, struct netlink_ext_ack *extack)
@@ -1515,6 +1521,11 @@ struct net_device_ops {struct net_device *dev, struct net_device *filter_dev, int *idx); +int (*ndo_fdb_flush)(struct ndmsg *ndm, +struct nlattr *tb[], +struct net_device *dev, +u16 vid, +struct netlink_ext_ack *extack); int (*ndo_fdb_get)(struct sk_buff *skb, struct nlattr *tb[], struct net_device *dev,
Signature · 1 line
2.35.1
Nikolay Aleksandrov
maintainerPATCH v2 3/8 net: bridge: fdb: add ndo_fdb_flush op
In reply to Nikolay AleksandrovAdd a minimal ndo_fdb_flush implementation which flushes all entries. Support for more fine-grained filtering will be added in the following patches.
--- net/bridge/br_device.c | 1 + net/bridge/br_fdb.c | 25 ++++++++++++++++++++++++- net/bridge/br_netlink.c | 2 +- net/bridge/br_private.h | 6 +++++- net/bridge/br_sysfs_br.c | 2 +- 5 files changed, 32 insertions(+), 4 deletions(-)
net/bridge/br_device.c +1 -0
@@ -466,6 +466,7 @@ static const struct net_device_ops br_netdev_ops = { .ndo_fdb_add = br_fdb_add, .ndo_fdb_del = br_fdb_delete, .ndo_fdb_dump = br_fdb_dump, + .ndo_fdb_flush = br_fdb_flush, .ndo_fdb_get = br_fdb_get, .ndo_bridge_getlink = br_getlink, .ndo_bridge_setlink = br_setlink,
net/bridge/br_fdb.c +24 -1
@@ -559,7 +559,7 @@ void br_fdb_cleanup(struct work_struct *work) }/* Completely flush all dynamic entries in forwarding database.*/ -void br_fdb_flush (struct net_bridge *br) +void __br_fdb_flush (struct net_bridge *br) {struct net_bridge_fdb_entry *f;struct hlist_node *tmp;
@@ -572,6 +572,29 @@ void br_fdb_flush(struct net_bridge *br)spin_unlock_bh (&br->hash_lock); } +int br_fdb_flush (struct ndmsg *ndm,struct nlattr *tb[], +struct net_device *dev,u16 vid, +struct netlink_ext_ack *extack) +{ +struct net_bridge *br; + +if (netif_is_bridge_master (dev)) { + br =netdev_priv (dev); + }else { +struct net_bridge_port *p =br_port_get_rtnl (dev); + +if (!p) { +NL_SET_ERR_MSG_MOD (extack,"Device is not a bridge port" ); +return -EINVAL ; + } + br = p->br; + } + +__br_fdb_flush (br); + +return 0 ; +} +/* Flush all entries referring to a specific port. * if do_all is set also flush static entries * if vid is set delete all entries that match the vlan_id
net/bridge/br_netlink.c +1 -1
@@ -1327,7 +1327,7 @@ static int br_changelink(struct net_device *brdev, struct nlattr *tb[], }if (data[IFLA_BR_FDB_FLUSH ]) -br_fdb_flush (br); +__br_fdb_flush (br);#ifdef CONFIG_BRIDGE_IGMP_SNOOPING if (data[IFLA_BR_MCAST_ROUTER ]) {
net/bridge/br_private.h +5 -1
@@ -759,7 +759,11 @@ int br_fdb_init(void);void br_fdb_fini (void );int br_fdb_hash_init (struct net_bridge *br);void br_fdb_hash_fini (struct net_bridge *br); -void br_fdb_flush (struct net_bridge *br); +void __br_fdb_flush (struct net_bridge *br); +int br_fdb_flush (struct ndmsg *ndm,struct nlattr *tb[], +struct net_device *dev,u16 vid, +struct netlink_ext_ack *extack); +void br_fdb_find_delete_local (struct net_bridge *br,const struct net_bridge_port *p,const unsigned char *addr,u16 vid);
net/bridge/br_sysfs_br.c +1 -1
@@ -344,7 +344,7 @@ static DEVICE_ATTR_RW(group_addr);static int set_flush (struct net_bridge *br,unsigned long val,struct netlink_ext_ack *extack) { -br_fdb_flush (br); +__br_fdb_flush (br);return 0 ; }
Signature · 1 line
2.35.1
Nikolay Aleksandrov
maintainerPATCH v2 4/8 net: rtnetlink: register a generic rtnl_fdb_flush call
In reply to Nikolay AleksandrovRegister a generic PF_BRIDGE rtnl_fdb_flush call which does basic validation and dispatches the call to the appropriate device based on ndm flags (NTF_MASTER and NTF_SELF). The flags are interepreted in a similar way to the already existing fdb add and del.
--- include/uapi/linux/neighbour.h | 6 ++++ net/core/rtnetlink.c | 52 ++++++++++++++++++++++++++++++++++ 2 files changed, 58 insertions(+)
include/uapi/linux/neighbour.h +6 -0
@@ -212,4 +212,10 @@ enum { };#define NFEA_MAX (__NFEA_MAX - 1) +enum { +NDFA_UNSPEC , + __NDFA_MAX +}; +#define NDFA_MAX (__NDFA_MAX - 1) +#endif
net/core/rtnetlink.c +52 -0
64 lines not on this page · raw message
Signature · 1 line
2.35.1
Nikolay Aleksandrov
maintainerPATCH v2 5/8 net: rtnetlink: add common flush attributes
In reply to Nikolay AleksandrovAdd common fdb flush attributes - ifindex, vlan id, ndm flags/state masks. All of these are used by the bridge and vxlan drivers. Also minimal attr policy validation is added, it is up to ndo_fdb_flush implementers to further validate them.
--- include/uapi/linux/neighbour.h | 4 ++++ net/core/rtnetlink.c | 16 +++++++++++++--- 2 files changed, 17 insertions(+), 3 deletions(-)
include/uapi/linux/neighbour.h +4 -0
@@ -214,6 +214,10 @@ enum {enum {NDFA_UNSPEC , +NDFA_IFINDEX , +NDFA_VLAN , +NDFA_NDM_STATE_MASK , +NDFA_NDM_FLAGS_MASK , __NDFA_MAX };#define NDFA_MAX (__NDFA_MAX - 1)
net/core/rtnetlink.c +13 -3
@@ -4660,7 +4660,11 @@ static int rtnl_fdb_get(struct sk_buff *in_skb, struct nlmsghdr *nlh, }static const struct nla_policy fdb_flush_policy[NDFA_MAX +1 ] = { - [NDFA_UNSPEC ] = { .type =NLA_REJECT }, + [NDFA_UNSPEC ] = { .type =NLA_REJECT }, + [NDFA_IFINDEX ] =NLA_POLICY_MIN (NLA_S32 ,1 ), + [NDFA_VLAN ] = { .type =NLA_U16 }, + [NDFA_NDM_STATE_MASK ] = { .type =NLA_U16 }, + [NDFA_NDM_FLAGS_MASK ] = { .type =NLA_U8 }, }; static int rtnl_fdb_flush(struct sk_buff *skb, struct nlmsghdr *nlh,
@@ -4670,6 +4674,7 @@ static int rtnl_fdb_flush(struct sk_buff *skb, struct nlmsghdr *nlh,struct nlattr *tb[NDFA_MAX +1 ];struct net_device *dev;struct ndmsg *ndm; +u16 vid;int err; err = nlmsg_parse(nlh, sizeof(*ndm), tb, NDFA_MAX, fdb_flush_policy,
@@ -4689,19 +4694,24 @@ static int rtnl_fdb_flush(struct sk_buff *skb, struct nlmsghdr *nlh,return -ENODEV ; } + err =fdb_vid_parse (tb[NDFA_VLAN ], &vid, extack); +if (err) +return err; + err = -EOPNOTSUPP ;if ((!ndm->ndm_flags || ndm->ndm_flags &NTF_MASTER ) &&netif_is_bridge_port (dev)) {struct net_device *br_dev =netdev_master_upper_dev_get (dev); - err = br_dev->netdev_ops->ndo_fdb_flush (ndm, tb, dev,0 , extack); + err = br_dev->netdev_ops->ndo_fdb_flush (ndm, tb, dev, vid, + extack);if (err) goto out;else ndm->ndm_flags &= ~NTF_MASTER ; }if ((ndm->ndm_flags &NTF_SELF ) && dev->netdev_ops->ndo_fdb_flush) { - err = dev->netdev_ops->ndo_fdb_flush (ndm, tb, dev,0 , extack); + err = dev->netdev_ops->ndo_fdb_flush (ndm, tb, dev, vid, extack);if (!err) ndm->ndm_flags &= ~NTF_SELF ; }
Signature · 1 line
2.35.1
Nikolay Aleksandrov
maintainerPATCH v2 6/8 net: bridge: fdb: add support for fine-grained flushing
In reply to Nikolay AleksandrovAdd the ability to specify exactly which fdbs to be flushed. They are described by a new structure - net_bridge_fdb_flush_desc. Currently it can match on port/bridge ifindex, vlan id and fdb flags. It is used to describe the existing dynamic fdb flush operation. Note that this flush operation doesn't treat permanent entries in a special way (fdb_delete vs fdb_delete_local), it will delete them regardless if any port is using them, so currently it can't directly replace deletes which need to handle that case, although we can extend it later for that too.
---
v2: changed the flush matches func for better readability (Ido)
net/bridge/br_fdb.c | 41 ++++++++++++++++++++++++++++++++-------- net/bridge/br_netlink.c | 9 +++++++-- net/bridge/br_private.h | 10 +++++++++- net/bridge/br_sysfs_br.c | 6 +++++- 4 files changed, 54 insertions(+), 12 deletions(-)
net/bridge/br_fdb.c +33 -8
@@ -558,24 +558,49 @@ void br_fdb_cleanup(struct work_struct *work)mod_delayed_work (system_long_wq, &br->gc_work, work_delay); } -/* Completely flush all dynamic entries in forwarding database.*/ -void __br_fdb_flush (struct net_bridge *br) +static bool __fdb_flush_matches (const struct net_bridge *br, +const struct net_bridge_fdb_entry *f, +const struct net_bridge_fdb_flush_desc *desc) +{ +const struct net_bridge_port *dst =READ_ONCE (f->dst); +int port_ifidx = dst ? dst->dev->ifindex : br->dev->ifindex; + +if (desc->vlan_id && desc->vlan_id != f->key.vlan_id) +return false; +if (desc->port_ifindex && desc->port_ifindex != port_ifidx) +return false; +if (desc->flags_mask && (f->flags & desc->flags_mask) != desc->flags) +return false; + +return true; +} + +/* Flush forwarding database entries matching the description */ +void __br_fdb_flush (struct net_bridge *br, +const struct net_bridge_fdb_flush_desc *desc) {struct net_bridge_fdb_entry *f; -struct hlist_node *tmp; -spin_lock_bh (&br->hash_lock); -hlist_for_each_entry_safe (f, tmp, &br->fdb_list, fdb_node) { -if (!test_bit (BR_FDB_STATIC , &f->flags)) +rcu_read_lock (); +hlist_for_each_entry_rcu (f, &br->fdb_list, fdb_node) { +if (!__fdb_flush_matches (br, f, desc)) +continue ; + +spin_lock_bh (&br->hash_lock); +if (!hlist_unhashed (&f->fdb_node))fdb_delete (br, f, true); +spin_unlock_bh (&br->hash_lock); } -spin_unlock_bh (&br->hash_lock); +rcu_read_unlock (); }int br_fdb_flush (struct ndmsg *ndm,struct nlattr *tb[],struct net_device *dev,u16 vid,struct netlink_ext_ack *extack) { +struct net_bridge_fdb_flush_desc desc = { + .flags_mask =BR_FDB_STATIC + };struct net_bridge *br;if (netif_is_bridge_master (dev)) {
@@ -590,7 +615,7 @@ int br_fdb_flush(struct ndmsg *ndm, struct nlattr *tb[], br = p->br; } -__br_fdb_flush (br); +__br_fdb_flush (br, &desc);return 0 ; }
net/bridge/br_netlink.c +7 -2
@@ -1326,8 +1326,13 @@ static int br_changelink(struct net_device *brdev, struct nlattr *tb[],br_recalculate_fwd_mask (br); } -if (data[IFLA_BR_FDB_FLUSH ]) -__br_fdb_flush (br); +if (data[IFLA_BR_FDB_FLUSH ]) { +struct net_bridge_fdb_flush_desc desc = { + .flags_mask =BR_FDB_STATIC + }; + +__br_fdb_flush (br, &desc); + }#ifdef CONFIG_BRIDGE_IGMP_SNOOPING if (data[IFLA_BR_MCAST_ROUTER ]) {
net/bridge/br_private.h +9 -1
@@ -274,6 +274,13 @@ struct net_bridge_fdb_entry {struct rcu_head rcu; }; +struct net_bridge_fdb_flush_desc { +unsigned long flags; +unsigned long flags_mask; +int port_ifindex; +u16 vlan_id; +}; +#define MDB_PG_FLAGS_PERMANENT BIT(0)#define MDB_PG_FLAGS_OFFLOAD BIT(1)#define MDB_PG_FLAGS_FAST_LEAVE BIT(2)
@@ -759,7 +766,8 @@ int br_fdb_init(void);void br_fdb_fini (void );int br_fdb_hash_init (struct net_bridge *br);void br_fdb_hash_fini (struct net_bridge *br); -void __br_fdb_flush (struct net_bridge *br); +void __br_fdb_flush (struct net_bridge *br, +const struct net_bridge_fdb_flush_desc *desc);int br_fdb_flush (struct ndmsg *ndm,struct nlattr *tb[],struct net_device *dev,u16 vid,struct netlink_ext_ack *extack);
net/bridge/br_sysfs_br.c +5 -1
@@ -344,7 +344,11 @@ static DEVICE_ATTR_RW(group_addr);static int set_flush (struct net_bridge *br,unsigned long val,struct netlink_ext_ack *extack) { -__br_fdb_flush (br); +struct net_bridge_fdb_flush_desc desc = { + .flags_mask =BR_FDB_STATIC + }; + +__br_fdb_flush (br, &desc);return 0 ; }
Signature · 1 line
2.35.1
Nikolay Aleksandrov
maintainerPATCH v2 7/8 net: bridge: fdb: add support for flush filtering based on ndm flags and state
In reply to Nikolay AleksandrovAdd support for fdb flush filtering based on ndm flags and state. NDM state and flags are mapped to bridge-specific flags and matched according to the specified masks. NTF_USE is used to represent added_by_user flag since it sets it on fdb add and we don't have a 1:1 mapping for it. Only allowed bits can be set, NTF_USE and NTF_MASTER are ignored.
---
v2: ignore NTF_USE/NTF_MASTER and reject unknown flags
net/bridge/br_fdb.c | 58 ++++++++++++++++++++++++++++++++++++++--- net/bridge/br_private.h | 5 ++++ 2 files changed, 60 insertions(+), 3 deletions(-)
net/bridge/br_fdb.c +55 -3
@@ -594,13 +594,40 @@ void __br_fdb_flush(struct net_bridge *br,rcu_read_unlock (); } +static unsigned long __ndm_state_to_fdb_flags (u16 ndm_state) +{ +unsigned long flags =0 ; + +if (ndm_state &NUD_PERMANENT ) +__set_bit (BR_FDB_LOCAL , &flags); +if (ndm_state &NUD_NOARP ) +__set_bit (BR_FDB_STATIC , &flags); + +return flags; +} + +static unsigned long __ndm_flags_to_fdb_flags (u8 ndm_flags) +{ +unsigned long flags =0 ; + +if (ndm_flags &NTF_USE ) +__set_bit (BR_FDB_ADDED_BY_USER , &flags); +if (ndm_flags &NTF_EXT_LEARNED ) +__set_bit (BR_FDB_ADDED_BY_EXT_LEARN , &flags); +if (ndm_flags &NTF_OFFLOADED ) +__set_bit (BR_FDB_OFFLOADED , &flags); +if (ndm_flags &NTF_STICKY ) +__set_bit (BR_FDB_STICKY , &flags); + +return flags; +} +int br_fdb_flush (struct ndmsg *ndm,struct nlattr *tb[],struct net_device *dev,u16 vid,struct netlink_ext_ack *extack) { -struct net_bridge_fdb_flush_desc desc = { - .flags_mask =BR_FDB_STATIC - }; +u8 ndm_flags = ndm->ndm_flags & ~FDB_FLUSH_IGNORED_NDM_FLAGS ; +struct net_bridge_fdb_flush_desc desc = {};struct net_bridge *br;if (netif_is_bridge_master (dev)) {
@@ -615,6 +642,31 @@ int br_fdb_flush(struct ndmsg *ndm, struct nlattr *tb[], br = p->br; } +if (ndm_flags & ~FDB_FLUSH_ALLOWED_NDM_FLAGS ) { +NL_SET_ERR_MSG (extack,"Unsupported fdb flush ndm flag bits set" ); +return -EINVAL ; + } +if (ndm->ndm_state & ~FDB_FLUSH_ALLOWED_NDM_STATES ) { +NL_SET_ERR_MSG (extack,"Unsupported fdb flush ndm state bits set" ); +return -EINVAL ; + } + + desc.flags |=__ndm_state_to_fdb_flags (ndm->ndm_state); + desc.flags |=__ndm_flags_to_fdb_flags (ndm_flags); +if (tb[NDFA_NDM_STATE_MASK ]) { +u16 ndm_state_mask =nla_get_u16 (tb[NDFA_NDM_STATE_MASK ]); + + desc.flags_mask |=__ndm_state_to_fdb_flags (ndm_state_mask); + } +if (tb[NDFA_NDM_FLAGS_MASK ]) { +u8 ndm_flags_mask =nla_get_u8 (tb[NDFA_NDM_FLAGS_MASK ]); + + desc.flags_mask |=__ndm_flags_to_fdb_flags (ndm_flags_mask); + } + +br_debug (br,"flushing port ifindex: %d vlan id: %u flags: 0x%lx flags mask: 0x%lx\n" , + desc.port_ifindex, desc.vlan_id, desc.flags, desc.flags_mask); +__br_fdb_flush (br, &desc);return 0 ;
net/bridge/br_private.h +5 -0
@@ -762,6 +762,11 @@ static inline void br_netpoll_disable(struct net_bridge_port *p)#endif /* br_fdb.c */ +#define FDB_FLUSH_IGNORED_NDM_FLAGS (NTF_MASTER | NTF_SELF) +#define FDB_FLUSH_ALLOWED_NDM_STATES (NUD_PERMANENT | NUD_NOARP) +#define FDB_FLUSH_ALLOWED_NDM_FLAGS (NTF_USE | NTF_EXT_LEARNED | \ + NTF_STICKY | NTF_OFFLOADED) +int br_fdb_init (void );void br_fdb_fini (void );int br_fdb_hash_init (struct net_bridge *br);
Signature · 1 line
2.35.1
Nikolay Aleksandrov
maintainerPATCH v2 8/8 net: bridge: fdb: add support for flush filtering based on ifindex and vlan
In reply to Nikolay AleksandrovAdd support for fdb flush filtering based on destination ifindex and vlan id. The ifindex must either match a port's device ifindex or the bridge's. The vlan support is trivial since it's already validated by rtnl_fdb_flush, we just need to fill it in.
---
v2: validate ifindex and fill in vlan id
net/bridge/br_fdb.c | 42 +++++++++++++++++++++++++++++++++++++++++- 1 file changed, 41 insertions(+), 1 deletion(-)
net/bridge/br_fdb.c +41 -1
@@ -622,12 +622,44 @@ static unsigned long __ndm_flags_to_fdb_flags(u8 ndm_flags)return flags; } +static int __fdb_flush_validate_ifindex (const struct net_bridge *br, +int ifindex, +struct netlink_ext_ack *extack) +{ +const struct net_device *dev; + + dev =__dev_get_by_index (dev_net (br->dev), ifindex); +if (!dev) { +NL_SET_ERR_MSG_MOD (extack,"Unknown flush device ifindex" ); +return -ENODEV ; + } +if (!netif_is_bridge_master (dev) && !netif_is_bridge_port (dev)) { +NL_SET_ERR_MSG_MOD (extack,"Flush device is not a bridge or bridge port" ); +return -EINVAL ; + } +if (netif_is_bridge_master (dev) && dev != br->dev) { +NL_SET_ERR_MSG_MOD (extack, +"Flush bridge device does not match target bridge device" ); +return -EINVAL ; + } +if (netif_is_bridge_port (dev)) { +struct net_bridge_port *p =br_port_get_rtnl (dev); + +if (p->br != br) { +NL_SET_ERR_MSG_MOD (extack,"Port belongs to a different bridge device" ); +return -EINVAL ; + } + } + +return 0 ; +} +int br_fdb_flush (struct ndmsg *ndm,struct nlattr *tb[],struct net_device *dev,u16 vid,struct netlink_ext_ack *extack) {u8 ndm_flags = ndm->ndm_flags & ~FDB_FLUSH_IGNORED_NDM_FLAGS ; -struct net_bridge_fdb_flush_desc desc = {}; +struct net_bridge_fdb_flush_desc desc = { .vlan_id = vid };struct net_bridge *br;if (netif_is_bridge_master (dev)) {
@@ -663,6 +695,14 @@ int br_fdb_flush(struct ndmsg *ndm, struct nlattr *tb[], desc.flags_mask |=__ndm_flags_to_fdb_flags (ndm_flags_mask); } +if (tb[NDFA_IFINDEX ]) { +int err, ifidx =nla_get_s32 (tb[NDFA_IFINDEX ]); + + err =__fdb_flush_validate_ifindex (br, ifidx, extack); +if (err) +return err; + desc.port_ifindex = ifidx; + }br_debug (br,"flushing port ifindex: %d vlan id: %u flags: 0x%lx flags mask: 0x%lx\n" , desc.port_ifindex, desc.vlan_id, desc.flags, desc.flags_mask);
Signature · 1 line
2.35.1
Nikolay Aleksandrov
maintainerQuoted 14 lines · Nikolay Aleksandrov
Hi,
This patch-set adds support to specify filtering conditions for a flush operation. This version has entirely different entry point (v1 had bridge-specific IFLA attribute, here I add newRTM_FLUSHNEIGHmsg and netdevndo_fdb_flushop) so I'll give a new overview altogether. After me and Ido discussed the feature offlist, we agreed that it would be best to add a new genericRTM_FLUSHNEIGHwith a newndo_fdb_flushcallback which can be re-used for other drivers (e.g. vxlan). Patch 01 adds the newRTM_FLUSHNEIGHtype, patch 02 then adds the newndo_fdb_flushcall. With this structure we need to add a genericrtnl_fdb_flushwhich will be used to do basic attribute validation and dispatch the call to the appropriate device based on the NTF_USE/MASTER flags (patch 03). Patch 04 then adds some common flush attributes which are used by the bridge and vxlan drivers (target ifindex, vlan id, ndm
Quoted 8 lines
flags/state masks) with basic attribute validation, further validation can be done by the implementers of the ndo callback. Patch 05 adds a minimal
ndo_fdb_flushto the bridge driver, it uses the currentbr_fdb_flushimplementation to flush all entries similar to existing calls. Patch 06 adds filtering support to the new bridge flush op which supports target ifindex (port or bridge), vlan id and flags/state mask. Patch 07 converts ndm state/flags and their masks to bridge-private flags and fills them in the filter descriptor for matching. Finally patch 08
Aargh.. I mixed up the patch numbers above. Patch 03 adds the minimal ndo_fdb_flush to the bridge driver (not patch 05), patch 04 adds the generic rtnl_fdb_flush (not patch 03) and patch 05 adds the common attributes (not patch 04).
Let me know if you'd like me to repost it with fixed numbers. I'll wait for feedback anyway.
Quoted 8 lines
fills in the target ifindex (after validating it) and vlan id (already
validated byrtnl_fdb_flush) for matching. Flush filtering is needed> because user-space applications need a quick way to delete only a specific set of entries, e.g. mlag implementations need a way to flush only
dynamic entries excluding externally learned ones or only externally
learned ones without static entries etc. Also apps usually want to target
only a specific vlan or port/vlan combination. The current 2 flush
operations (per port and bridge-wide) are not extensible and cannot
provide such filtering.
Quoted 71 lines · Nikolay Aleksandrov
Hi,
This patch-set adds support to specify filtering conditions for a flush operation. This version has entirely different entry point (v1 had bridge-specific IFLA attribute, here I add newRTM_FLUSHNEIGHmsg and netdevndo_fdb_flushop) so I'll give a new overview altogether. After me and Ido discussed the feature offlist, we agreed that it would be best to add a new genericRTM_FLUSHNEIGHwith a newndo_fdb_flushcallback which can be re-used for other drivers (e.g. vxlan). Patch 01 adds the newRTM_FLUSHNEIGHtype, patch 02 then adds the newndo_fdb_flushcall. With this structure we need to add a genericrtnl_fdb_flushwhich will be used to do basic attribute validation and dispatch the call to the appropriate device based on the NTF_USE/MASTER flags (patch 03). Patch 04 then adds some common flush attributes which are used by the bridge and vxlan drivers (target ifindex, vlan id, ndm flags/state masks) with basic attribute validation, further validation can be done by the implementers of the ndo callback. Patch 05 adds a minimalndo_fdb_flushto the bridge driver, it uses the currentbr_fdb_flushimplementation to flush all entries similar to existing calls. Patch 06 adds filtering support to the new bridge flush op which supports target ifindex (port or bridge), vlan id and flags/state mask. Patch 07 converts ndm state/flags and their masks to bridge-private flags and fills them in the filter descriptor for matching. Finally patch 08 fills in the target ifindex (after validating it) and vlan id (already validated byrtnl_fdb_flush) for matching. Flush filtering is needed because user-space applications need a quick way to delete only a specific set of entries, e.g. mlag implementations need a way to flush only dynamic entries excluding externally learned ones or only externally learned ones without static entries etc. Also apps usually want to target only a specific vlan or port/vlan combination. The current 2 flush operations (per port and bridge-wide) are not extensible and cannot provide such filtering.I decided against embedding new attrs into the old flush attributes for multiple reasons - proper error handling on unsupported attributes, older kernels silently flushing all, need for a second mechanism to signal that the attribute should be parsed (e.g. using boolopts), special treatment for permanent entries.
Examples:
$ bridge fdb flush dev bridge vlan 100 static < flush all static entries on vlan 100 > $ bridge fdb flush dev bridge vlan 1 dynamic < flush all dynamic entries on vlan 1 > $ bridge fdb flush dev bridge port ens16 vlan 1 dynamic < flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev ens16 vlan 1 dynamic master < as above: flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev bridge nooffloaded nopermanent self < flush all non-offloaded and non-permanent entries > $ bridge fdb flush dev bridge static noextern_learn < flush all static entries which are not externally learned > $ bridge fdb flush dev bridge permanent < flush all permanent entries > $ bridge fdb flush dev bridge port bridge permanent < flush all permanent entries pointing to the bridge itself >Note that all flags have their negated version (static vs nostatic etc) and there are some tricky cases to handle like "static" which in flag terms means fdbs that have
NUD_NOARPbut *not*NUD_PERMANENT, so the mask matches on both but we need onlyNUD_NOARPto be set. That's because permanent entries have both set so we can't just match onNUD_NOARP. Also note that this flush operation doesn't treat permanent entries in a special way (fdb_deletevsfdb_delete_local), it will delete them regardless if any port is using them. We can extend the api with a flag to do that if needed in the future.Patch-sets (in order):
- Initial flush infra and fdb flush filtering (this set)
- iproute2 support
- selftests
Quoted 5 lines
Future work:
- mdb flush support (RTM_FLUSHMDBtype)Thanks to Ido for the great discussion and feedback while working on this.
Cant we pile this on to RTM_DELNEIGH with a flush flag ?.
It is a bulk del, and sounds seems similar to the bulk dev del discussion on netdev a few months ago (i dont remember how that api ended up to be. unless i am misremembering).
neigh subsystem also needs this, curious how this api will work there.
(apologies if you guys already discussed this, did not have time to look through all the comments)
Nikolay Aleksandrov
maintainerQuoted 78 lines · Roopa Prabhu
Quoted 71 lines · Nikolay Aleksandrov
Hi,
This patch-set adds support to specify filtering conditions for a flush operation. This version has entirely different entry point (v1 had bridge-specific IFLA attribute, here I add newRTM_FLUSHNEIGHmsg and netdevndo_fdb_flushop) so I'll give a new overview altogether. After me and Ido discussed the feature offlist, we agreed that it would be best to add a new genericRTM_FLUSHNEIGHwith a newndo_fdb_flushcallback which can be re-used for other drivers (e.g. vxlan). Patch 01 adds the newRTM_FLUSHNEIGHtype, patch 02 then adds the newndo_fdb_flushcall. With this structure we need to add a genericrtnl_fdb_flushwhich will be used to do basic attribute validation and dispatch the call to the appropriate device based on the NTF_USE/MASTER flags (patch 03). Patch 04 then adds some common flush attributes which are used by the bridge and vxlan drivers (target ifindex, vlan id, ndm flags/state masks) with basic attribute validation, further validation can be done by the implementers of the ndo callback. Patch 05 adds a minimalndo_fdb_flushto the bridge driver, it uses the currentbr_fdb_flushimplementation to flush all entries similar to existing calls. Patch 06 adds filtering support to the new bridge flush op which supports target ifindex (port or bridge), vlan id and flags/state mask. Patch 07 converts ndm state/flags and their masks to bridge-private flags and fills them in the filter descriptor for matching. Finally patch 08 fills in the target ifindex (after validating it) and vlan id (already validated byrtnl_fdb_flush) for matching. Flush filtering is needed because user-space applications need a quick way to delete only a specific set of entries, e.g. mlag implementations need a way to flush only dynamic entries excluding externally learned ones or only externally learned ones without static entries etc. Also apps usually want to target only a specific vlan or port/vlan combination. The current 2 flush operations (per port and bridge-wide) are not extensible and cannot provide such filtering.I decided against embedding new attrs into the old flush attributes for multiple reasons - proper error handling on unsupported attributes, older kernels silently flushing all, need for a second mechanism to signal that the attribute should be parsed (e.g. using boolopts), special treatment for permanent entries.
Examples:
$ bridge fdb flush dev bridge vlan 100 static < flush all static entries on vlan 100 > $ bridge fdb flush dev bridge vlan 1 dynamic < flush all dynamic entries on vlan 1 > $ bridge fdb flush dev bridge port ens16 vlan 1 dynamic < flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev ens16 vlan 1 dynamic master < as above: flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev bridge nooffloaded nopermanent self < flush all non-offloaded and non-permanent entries > $ bridge fdb flush dev bridge static noextern_learn < flush all static entries which are not externally learned > $ bridge fdb flush dev bridge permanent < flush all permanent entries > $ bridge fdb flush dev bridge port bridge permanent < flush all permanent entries pointing to the bridge itself >Note that all flags have their negated version (static vs nostatic etc) and there are some tricky cases to handle like "static" which in flag terms means fdbs that have
NUD_NOARPbut *not*NUD_PERMANENT, so the mask matches on both but we need onlyNUD_NOARPto be set. That's because permanent entries have both set so we can't just match onNUD_NOARP. Also note that this flush operation doesn't treat permanent entries in a special way (fdb_deletevsfdb_delete_local), it will delete them regardless if any port is using them. We can extend the api with a flag to do that if needed in the future.Patch-sets (in order):
- Initial flush infra and fdb flush filtering (this set)
- iproute2 support
- selftestsQuoted 5 lines
Future work:
- mdb flush support (RTM_FLUSHMDBtype)Thanks to Ido for the great discussion and feedback while working on this.
Quoted 10 lines
Cant we pile this on to
RTM_DELNEIGHwith a flush flag ?.It is a bulk del, and sounds seems similar to the bulk dev del discussion on netdev a few months ago (i dont remember how that api ended up to be. unless i am misremembering).
neigh subsystem also needs this, curious how this api will work there.
(apologies if you guys already discussed this, did not have time to look through all the comments)
I thought about that option, but I didn't like overloading delneigh like that. del currently requires a mac address and we need to either signal the device supports a null mac, or we should push that verification to ndo_fdb_del users. Also we'll have attributes which are flush-specific and will work only when flushing as opposed to when deleting a specific mac, so handling them in the different cases can become a pain. MDBs will need DELMDB to be modified in a similar way.
IMO a separate flush op is cleaner, but I don't have a strong preference. This can very easily be adapted to delneigh with just a bit more mechanical changes if the mac check is pushed to the ndo implementers.
FLUSHNEIGH can easily work for neighs, just need another address family rtnl_register that implements it, the new ndo is just for PF_BRIDGE. :)
Cheers,
Nik
Nikolay Aleksandrov
maintainerQuoted 85 lines · Nikolay Aleksandrov
Quoted 84 lines · Roopa Prabhu
Quoted 71 lines · Nikolay Aleksandrov
Hi,
This patch-set adds support to specify filtering conditions for a flush operation. This version has entirely different entry point (v1 had bridge-specific IFLA attribute, here I add newRTM_FLUSHNEIGHmsg and netdevndo_fdb_flushop) so I'll give a new overview altogether. After me and Ido discussed the feature offlist, we agreed that it would be best to add a new genericRTM_FLUSHNEIGHwith a newndo_fdb_flushcallback which can be re-used for other drivers (e.g. vxlan). Patch 01 adds the newRTM_FLUSHNEIGHtype, patch 02 then adds the newndo_fdb_flushcall. With this structure we need to add a genericrtnl_fdb_flushwhich will be used to do basic attribute validation and dispatch the call to the appropriate device based on the NTF_USE/MASTER flags (patch 03). Patch 04 then adds some common flush attributes which are used by the bridge and vxlan drivers (target ifindex, vlan id, ndm flags/state masks) with basic attribute validation, further validation can be done by the implementers of the ndo callback. Patch 05 adds a minimalndo_fdb_flushto the bridge driver, it uses the currentbr_fdb_flushimplementation to flush all entries similar to existing calls. Patch 06 adds filtering support to the new bridge flush op which supports target ifindex (port or bridge), vlan id and flags/state mask. Patch 07 converts ndm state/flags and their masks to bridge-private flags and fills them in the filter descriptor for matching. Finally patch 08 fills in the target ifindex (after validating it) and vlan id (already validated byrtnl_fdb_flush) for matching. Flush filtering is needed because user-space applications need a quick way to delete only a specific set of entries, e.g. mlag implementations need a way to flush only dynamic entries excluding externally learned ones or only externally learned ones without static entries etc. Also apps usually want to target only a specific vlan or port/vlan combination. The current 2 flush operations (per port and bridge-wide) are not extensible and cannot provide such filtering.I decided against embedding new attrs into the old flush attributes for multiple reasons - proper error handling on unsupported attributes, older kernels silently flushing all, need for a second mechanism to signal that the attribute should be parsed (e.g. using boolopts), special treatment for permanent entries.
Examples:
$ bridge fdb flush dev bridge vlan 100 static < flush all static entries on vlan 100 > $ bridge fdb flush dev bridge vlan 1 dynamic < flush all dynamic entries on vlan 1 > $ bridge fdb flush dev bridge port ens16 vlan 1 dynamic < flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev ens16 vlan 1 dynamic master < as above: flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev bridge nooffloaded nopermanent self < flush all non-offloaded and non-permanent entries > $ bridge fdb flush dev bridge static noextern_learn < flush all static entries which are not externally learned > $ bridge fdb flush dev bridge permanent < flush all permanent entries > $ bridge fdb flush dev bridge port bridge permanent < flush all permanent entries pointing to the bridge itself >Note that all flags have their negated version (static vs nostatic etc) and there are some tricky cases to handle like "static" which in flag terms means fdbs that have
NUD_NOARPbut *not*NUD_PERMANENT, so the mask matches on both but we need onlyNUD_NOARPto be set. That's because permanent entries have both set so we can't just match onNUD_NOARP. Also note that this flush operation doesn't treat permanent entries in a special way (fdb_deletevsfdb_delete_local), it will delete them regardless if any port is using them. We can extend the api with a flag to do that if needed in the future.Patch-sets (in order):
- Initial flush infra and fdb flush filtering (this set)
- iproute2 support
- selftestsQuoted 5 lines
Future work:
- mdb flush support (RTM_FLUSHMDBtype)Thanks to Ido for the great discussion and feedback while working on this.
Cant we pile this on to
RTM_DELNEIGHwith a flush flag ?.It is a bulk del, and sounds seems similar to the bulk dev del discussion on netdev a few months ago (i dont remember how that api ended up to be. unless i am misremembering).
neigh subsystem also needs this, curious how this api will work there.
Quoted 7 lines
Quoted 4 lines
(apologies if you guys already discussed this, did not have time to look through all the comments)
I thought about that option, but I didn't like overloading delneigh like that.
del currently requires a mac address and we need to either signal the device supports> a null mac, or we should push that verification tondo_fdb_delusers. Also we'll have
that's the only thing, overloading delneigh with a flush-behaviour (multi-del or whatever) would require to push the mac check to ndo_fdb_del implementers
I don't mind going that road if others agree that we should do it through delneigh
+ a bit/option to signal flush, instead of a new rtm type.
Quoted 2 lines
attributes which are flush-specific and will work only when flushing as opposed to when deleting a specific mac, so handling them in the different cases can become a pain.
scratch the specific attributes, those can be adapted for both cases
Quoted 13 lines
MDBs will need DELMDB to be modified in a similar way.
IMO a separate flush op is cleaner, but I don't have a strong preference. This can very easily be adapted to delneigh with just a bit more mechanical changes if the mac check is pushed to the ndo implementers.
FLUSHNEIGH can easily work for neighs, just need another address family
rtnl_registerthat implements it, the new ndo is just forPF_BRIDGE. :)Cheers,
Nik
Quoted 94 lines · Nikolay Aleksandrov
Quoted 82 lines · Nikolay Aleksandrov
Quoted 81 lines · Roopa Prabhu
Quoted 71 lines · Nikolay Aleksandrov
Hi,
This patch-set adds support to specify filtering conditions for a flush operation. This version has entirely different entry point (v1 had bridge-specific IFLA attribute, here I add newRTM_FLUSHNEIGHmsg and netdevndo_fdb_flushop) so I'll give a new overview altogether. After me and Ido discussed the feature offlist, we agreed that it would be best to add a new genericRTM_FLUSHNEIGHwith a newndo_fdb_flushcallback which can be re-used for other drivers (e.g. vxlan). Patch 01 adds the newRTM_FLUSHNEIGHtype, patch 02 then adds the newndo_fdb_flushcall. With this structure we need to add a genericrtnl_fdb_flushwhich will be used to do basic attribute validation and dispatch the call to the appropriate device based on the NTF_USE/MASTER flags (patch 03). Patch 04 then adds some common flush attributes which are used by the bridge and vxlan drivers (target ifindex, vlan id, ndm flags/state masks) with basic attribute validation, further validation can be done by the implementers of the ndo callback. Patch 05 adds a minimalndo_fdb_flushto the bridge driver, it uses the currentbr_fdb_flushimplementation to flush all entries similar to existing calls. Patch 06 adds filtering support to the new bridge flush op which supports target ifindex (port or bridge), vlan id and flags/state mask. Patch 07 converts ndm state/flags and their masks to bridge-private flags and fills them in the filter descriptor for matching. Finally patch 08 fills in the target ifindex (after validating it) and vlan id (already validated byrtnl_fdb_flush) for matching. Flush filtering is needed because user-space applications need a quick way to delete only a specific set of entries, e.g. mlag implementations need a way to flush only dynamic entries excluding externally learned ones or only externally learned ones without static entries etc. Also apps usually want to target only a specific vlan or port/vlan combination. The current 2 flush operations (per port and bridge-wide) are not extensible and cannot provide such filtering.I decided against embedding new attrs into the old flush attributes for multiple reasons - proper error handling on unsupported attributes, older kernels silently flushing all, need for a second mechanism to signal that the attribute should be parsed (e.g. using boolopts), special treatment for permanent entries.
Examples:
$ bridge fdb flush dev bridge vlan 100 static < flush all static entries on vlan 100 > $ bridge fdb flush dev bridge vlan 1 dynamic < flush all dynamic entries on vlan 1 > $ bridge fdb flush dev bridge port ens16 vlan 1 dynamic < flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev ens16 vlan 1 dynamic master < as above: flush all dynamic entries on port ens16 and vlan 1 > $ bridge fdb flush dev bridge nooffloaded nopermanent self < flush all non-offloaded and non-permanent entries > $ bridge fdb flush dev bridge static noextern_learn < flush all static entries which are not externally learned > $ bridge fdb flush dev bridge permanent < flush all permanent entries > $ bridge fdb flush dev bridge port bridge permanent < flush all permanent entries pointing to the bridge itself >Note that all flags have their negated version (static vs nostatic etc) and there are some tricky cases to handle like "static" which in flag terms means fdbs that have
NUD_NOARPbut *not*NUD_PERMANENT, so the mask matches on both but we need onlyNUD_NOARPto be set. That's because permanent entries have both set so we can't just match onNUD_NOARP. Also note that this flush operation doesn't treat permanent entries in a special way (fdb_deletevsfdb_delete_local), it will delete them regardless if any port is using them. We can extend the api with a flag to do that if needed in the future.Patch-sets (in order):
- Initial flush infra and fdb flush filtering (this set)
- iproute2 support
- selftestsQuoted 5 lines
Future work:
- mdb flush support (RTM_FLUSHMDBtype)Thanks to Ido for the great discussion and feedback while working on this.
Cant we pile this on to
RTM_DELNEIGHwith a flush flag ?.It is a bulk del, and sounds seems similar to the bulk dev del discussion on netdev a few months ago (i dont remember how that api ended up to be. unless i am misremembering).
Quoted 8 lines
Quoted 6 lines
neigh subsystem also needs this, curious how this api will work there.
(apologies if you guys already discussed this, did not have time to look through all the comments)
I thought about that option, but I didn't like overloading delneigh like that.
del currently requires a mac address and we need to either signal the device supports> a null mac, or we should push that verification tondo_fdb_delusers. Also we'll havethat's the only thing, overloading delneigh with a flush-behaviour (multi-del or whatever) would require to push the mac check to
ndo_fdb_delimplementers
Quoted 15 lines
I don't mind going that road if others agree that we should do it through delneigh
+ a bit/option to signal flush, instead of a new rtm type.Quoted 2 lines
attributes which are flush-specific and will work only when flushing as opposed to when deleting a specific mac, so handling them in the different cases can become a pain.
scratch the specific attributes, those can be adapted for both cases
Quoted 8 lines
MDBs will need DELMDB to be modified in a similar way.
IMO a separate flush op is cleaner, but I don't have a strong preference. This can very easily be adapted to delneigh with just a bit more mechanical changes if the mac check is pushed to the ndo implementers.
FLUSHNEIGH can easily work for neighs, just need another address family
rtnl_registerthat implements it, the new ndo is just forPF_BRIDGE. :)
all great points. My only reason to explore RTM_DELNEIGH is to see if we can find a recipe to support similar bulk deletes of other objects handled via rtm msgs in the future. Plus, it allows you to maintain symmetry between flush requests and object delete notification msg types.
Lets see if there are other opinions.
Thanks Nikolay
Jakub Kicinski
maintainerQuoted 21 lines · On Mon, 11 Apr 2022 12:22:24 -0700 Roopa Prabhu wrote:
Quoted 5 lines
Quoted 2 lines
I thought about that option, but I didn't like overloading delneigh like that.
del currently requires a mac address and we need to either signal the device supports> a null mac, or we should push that verification tondo_fdb_delusers. Also we'll havethat's the only thing, overloading delneigh with a flush-behaviour (multi-del or whatever) would require to push the mac check to
ndo_fdb_delimplementersQuoted 15 lines
I don't mind going that road if others agree that we should do it through delneigh
+ a bit/option to signal flush, instead of a new rtm type.Quoted 2 lines
attributes which are flush-specific and will work only when flushing as opposed to when deleting a specific mac, so handling them in the different cases can become a pain.
scratch the specific attributes, those can be adapted for both cases
Quoted 8 lines
MDBs will need DELMDB to be modified in a similar way.
IMO a separate flush op is cleaner, but I don't have a strong preference. This can very easily be adapted to delneigh with just a bit more mechanical changes if the mac check is pushed to the ndo implementers.
FLUSHNEIGH can easily work for neighs, just need another address family
rtnl_registerthat implements it, the new ndo is just forPF_BRIDGE. :)
Quoted 6 lines
all great points. My only reason to explore
RTM_DELNEIGHis to see if we can find a recipe to support similar bulk deletes of other objects handled via rtm msgs in the future. Plus, it allows you to maintain symmetry between flush requests and object delete notification msg types.Lets see if there are other opinions.
I'd vote for reusing RTM_DELNEIGH, but that's purely based on intuition, I don't know this code. I'd also lean towards core creating struct net_bridge_fdb_flush_desc rather than piping raw netlink attrs thru. Lastly feels like fdb ops should find a new home rather than ndos, but that's largely unrelated..
Nikolay Aleksandrov
maintainerQuoted 22 lines · Jakub Kicinski
Quoted 21 lines · On Mon, 11 Apr 2022 12:22:24 -0700 Roopa Prabhu wrote:
Quoted 20 lines
Quoted 2 lines
I thought about that option, but I didn't like overloading delneigh like that.
del currently requires a mac address and we need to either signal the device supports> a null mac, or we should push that verification tondo_fdb_delusers. Also we'll havethat's the only thing, overloading delneigh with a flush-behaviour (multi-del or whatever) would require to push the mac check to
ndo_fdb_delimplementersI don't mind going that road if others agree that we should do it through delneigh
+ a bit/option to signal flush, instead of a new rtm type.Quoted 2 lines
attributes which are flush-specific and will work only when flushing as opposed to when deleting a specific mac, so handling them in the different cases can become a pain.
scratch the specific attributes, those can be adapted for both cases
Quoted 8 lines
MDBs will need DELMDB to be modified in a similar way.
IMO a separate flush op is cleaner, but I don't have a strong preference. This can very easily be adapted to delneigh with just a bit more mechanical changes if the mac check is pushed to the ndo implementers.
FLUSHNEIGH can easily work for neighs, just need another address family
rtnl_registerthat implements it, the new ndo is just forPF_BRIDGE. :)
Quoted 8 lines
Quoted 6 lines
all great points. My only reason to explore
RTM_DELNEIGHis to see if we can find a recipe to support similar bulk deletes of other objects handled via rtm msgs in the future. Plus, it allows you to maintain symmetry between flush requests and object delete notification msg types.Lets see if there are other opinions.
I'd vote for reusing
RTM_DELNEIGH, but that's purely based on
OK, I'll look into the delneigh solution. Note that for backwards compatibility we won't be able to return proper error because rtnl_fdb_del will be called without a mac address, so for old kernels with new iproute2 fdb flush will return "invalid address" as an error.
Quoted 3 lines
intuition, I don't know this code. I'd also lean towards core creating struct
net_bridge_fdb_flush_descrather than piping raw netlink attrs thru. Lastly feels like fdb ops should find
I don't think the struct can really be centralized, at least for the bridge case it contains private fields which parsed attributes get mapped to, specifically the ndm flags and state, and their maps are all mapped into bridge-private flags. Or did you mean pass the raw attribute vals through a struct instead of a nlattr table?
Quoted 1 line
a new home rather than ndos, but that's largely unrelated..
I like separating the ops idea. I'll add that to my bridge todo list. :)
Thanks,
Nik
Jakub Kicinski
maintainerQuoted 6 lines · On Mon, 11 Apr 2022 23:34:23 +0300 Nikolay Aleksandrov wrote:
Quoted 5 lines · Jakub Kicinski
Quoted 5 lines
all great points. My only reason to explore
RTM_DELNEIGHis to see if we can find a recipe to support similar bulk deletes of other objects handled via rtm msgs in the future. Plus, it allows you to maintain symmetry between flush requests and object delete notification msg types.
Quoted 8 lines
Quoted 3 lines
Quoted 1 line
Lets see if there are other opinions.
I'd vote for reusing
RTM_DELNEIGH, but that's purely based onOK, I'll look into the delneigh solution. Note that for backwards compatibility we won't be able to return proper error because
rtnl_fdb_delwill be called without a mac address, so for old kernels with new iproute2 fdb flush will return "invalid address" as an error.
If only we had policy dump for rtnl :) Another todo item, I guess.
Quoted 9 lines
Quoted 3 lines
intuition, I don't know this code. I'd also lean towards core creating struct
net_bridge_fdb_flush_descrather than piping raw netlink attrs thru. Lastly feels like fdb ops should findI don't think the struct can really be centralized, at least for the bridge case it contains private fields which parsed attributes get mapped to, specifically the ndm flags and state, and their maps are all mapped into bridge-private flags. Or did you mean pass the raw attribute vals through a struct instead of a nlattr table?
Yup, basically the policy is defined in the core, so the types are known. We can extract the fields from the message there, even if the exact meaning of the fields gets established in the callback.
BTW setting NLA_REJECT policy is not required, NLA_REJECT is 0 so it will be set automatically per C standard.
Quoted 7 lines
Quoted 1 line
a new home rather than ndos, but that's largely unrelated..
I like separating the ops idea. I'll add that to my bridge todo list. :)
Thanks,
Nik
Nikolay Aleksandrov
maintainerQuoted 9 lines · Jakub Kicinski
Quoted 8 lines · On Mon, 11 Apr 2022 23:34:23 +0300 Nikolay Aleksandrov wrote:
Quoted 7 lines · Jakub Kicinski
Quoted 6 lines
all great points. My only reason to explore
RTM_DELNEIGHis to see if we can find a recipe to support similar bulk deletes of other objects handled via rtm msgs in the future. Plus, it allows you to maintain symmetry between flush requests and object delete notification msg types.Lets see if there are other opinions.
Quoted 9 lines
Quoted 6 lines
Quoted 1 line
I'd vote for reusing
RTM_DELNEIGH, but that's purely based onOK, I'll look into the delneigh solution. Note that for backwards compatibility we won't be able to return proper error because
rtnl_fdb_delwill be called without a mac address, so for old kernels with new iproute2 fdb flush will return "invalid address" as an error.If only we had policy dump for rtnl :) Another todo item, I guess.
:)
Quoted 10 lines
Quoted 9 lines
Quoted 3 lines
intuition, I don't know this code. I'd also lean towards core creating struct
net_bridge_fdb_flush_descrather than piping raw netlink attrs thru. Lastly feels like fdb ops should findI don't think the struct can really be centralized, at least for the bridge case it contains private fields which parsed attributes get mapped to, specifically the ndm flags and state, and their maps are all mapped into bridge-private flags. Or did you mean pass the raw attribute vals through a struct instead of a nlattr table?
Quoted 4 lines
Yup, basically the policy is defined in the core, so the types are known. We can extract the fields from the message there, even if the exact meaning of the fields gets established in the callback.
That sounds nice, but there are a few catches, f.e. some ndo_fdb implementations check if attributes were set, i.e. they can also interpret 0, so it will require additional state (either special value, bitfield or some other way of telling them it was actually present but 0).
Anyway I think that is orthogonal to adding the flush support, it's a nice cleanup but can be done separately because it will have to be done for all ndo_fdb callbacks and I suspect the change will grow considerably.
OTOH the flush implementation via delneigh doesn't require a new ndo_fdb call way, would you mind if I finish that up without the struct conversion?
Quoted 3 lines
BTW setting
NLA_REJECTpolicy is not required,NLA_REJECTis 0 so it will be set automatically per C standard.
Indeed - habits, I'll drop it. :)
Quoted 8 lines
Quoted 7 lines
Quoted 1 line
a new home rather than ndos, but that's largely unrelated..
I like separating the ops idea. I'll add that to my bridge todo list. :)
Thanks,
Nik
Jakub Kicinski
maintainerQuoted 4 lines · On Tue, 12 Apr 2022 00:17:14 +0300 Nikolay Aleksandrov wrote:
Quoted 3 lines
Yup, basically the policy is defined in the core, so the types are known. We can extract the fields from the message there, even if the exact meaning of the fields gets established in the callback.
Quoted 9 lines
That sounds nice, but there are a few catches, f.e. some
ndo_fdbimplementations check if attributes were set, i.e. they can also interpret 0, so it will require additional state (either special value, bitfield or some other way of telling them it was actually present but 0).
Anyway I think that is orthogonal to adding the flush support, it's a nice cleanup but can be done separately because it will have to be done for allndo_fdbcallbacks and I suspect the change will grow considerably.
OTOH the flush implementation via delneigh doesn't require a newndo_fdbcall way, would you mind if I finish that up without the struct conversion?
Not terribly, go ahead.
Quoted 6 lines · Roopa Prabhu
all great points. My only reason to explore
RTM_DELNEIGHis to see if we can find a recipe to support similar bulk deletes of other objects handled via rtm msgs in the future. Plus, it allows you to maintain symmetry between flush requests and object delete notification msg types.Lets see if there are other opinions.
I guess I should have read the entire thread. :-) (still getting used to the new lei + mutt workflow). This was my thought - bulk delete is going to be a common need, and it is really just a mass delete. The GET message types are used for dumps and some allow attributes on the request as a means of coarse grain filtering. I think we should try something similar here for the flush case.