diff mbox series

[v5,net-next,11/15] net: dsa: sync static FDB entries on foreign interfaces to hardware

Message ID 20210629140658.2510288-12-olteanv@gmail.com (mailing list archive)
State Accepted
Commit 3068d466a67ec96a6972f248f5c7a7b6763dbeb1
Delegated to: Netdev Maintainers
Headers show
Series RX filtering in DSA | expand

Checks

Context Check Description
netdev/cover_letter success Link
netdev/fixes_present success Link
netdev/patch_count success Link
netdev/tree_selection success Clearly marked for net-next
netdev/subject_prefix success Link
netdev/cc_maintainers success CCed 7 of 7 maintainers
netdev/source_inline success Was 0 now: 0
netdev/verify_signedoff success Link
netdev/module_param success Was 0 now: 0
netdev/build_32bit success Errors and warnings before: 0 this patch: 0
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/verify_fixes success Link
netdev/checkpatch success total: 0 errors, 0 warnings, 0 checks, 24 lines checked
netdev/build_allmodconfig_warn success Errors and warnings before: 0 this patch: 0
netdev/header_inline success Link

Commit Message

Vladimir Oltean June 29, 2021, 2:06 p.m. UTC 
From: Vladimir Oltean <vladimir.oltean@nxp.com>

DSA is able to install FDB entries towards the CPU port for addresses
which were dynamically learnt by the software bridge on foreign
interfaces that are in the same bridge with a DSA switch interface.
Since this behavior is opportunistic, it is guarded by the
"assisted_learning_on_cpu_port" property which can be enabled by drivers
and is not done automatically (since certain switches may support
address learning of packets coming from the CPU port).

But if those FDB entries added on the foreign interfaces are static
(added by the user) instead of dynamically learnt, currently DSA does
not do anything (and arguably it should).

Because static FDB entries are not supposed to move on their own, there
is no downside in reusing the "assisted_learning_on_cpu_port" logic to
sync static FDB entries to the DSA CPU port unconditionally, even if
assisted_learning_on_cpu_port is not requested by the driver.

For example, this situation:

   br0
   / \
swp0 dummy0

$ bridge fdb add 02:00:de:ad:00:01 dev dummy0 vlan 1 master static

Results in DSA adding an entry in the hardware FDB, pointing this
address towards the CPU port.

The same is true for entries added to the bridge itself, e.g:

$ bridge fdb add 02:00:de:ad:00:01 dev br0 vlan 1 self local

(except that right now, DSA still ignores 'local' FDB entries, this will
be changed in a later patch)

Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
v4->v5: small fixups to commit message

 net/dsa/slave.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)
diff mbox series

Patch

diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index ac7f4f200ab1..ea9a7c1ce83e 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -2403,9 +2403,12 @@  static int dsa_slave_switchdev_event(struct notifier_block *unused,
 
 			dp = dsa_slave_to_port(dev);
 		} else {
-			/* Snoop addresses learnt on foreign interfaces
-			 * bridged with us, for switches that don't
-			 * automatically learn SA from CPU-injected traffic
+			/* Snoop addresses added to foreign interfaces
+			 * bridged with us, or the bridge
+			 * itself. Dynamically learned addresses can
+			 * also be added for switches that don't
+			 * automatically learn SA from CPU-injected
+			 * traffic.
 			 */
 			struct net_device *br_dev;
 			struct dsa_slave_priv *p;
@@ -2424,7 +2427,8 @@  static int dsa_slave_switchdev_event(struct notifier_block *unused,
 			dp = p->dp;
 			host_addr = true;
 
-			if (!dp->ds->assisted_learning_on_cpu_port)
+			if (!fdb_info->added_by_user &&
+			    !dp->ds->assisted_learning_on_cpu_port)
 				return NOTIFY_DONE;
 
 			/* When the bridge learns an address on an offloaded