diff options
author | Vladimir Oltean <vladimir.oltean@nxp.com> | 2022-03-22 02:37:01 +0200 |
---|---|---|
committer | Jakub Kicinski <kuba@kernel.org> | 2022-03-22 22:19:35 -0700 |
commit | 5077e2c8cf4d6b22a95a6c54a917f764e8887978 (patch) | |
tree | 6d9dd2960a983198a41c5c3aa3bab84f4a3d1ab0 /net/dsa/master.c | |
parent | 2af7e566a8616c278e1d7287ce86cd3900bed943 (diff) | |
download | linux-5077e2c8cf4d6b22a95a6c54a917f764e8887978.tar.bz2 |
net: dsa: fix missing host-filtered multicast addresses
DSA ports are stacked devices, so they use dev_mc_add() to sync their
address list to their lower interface (DSA master). But they are also
hardware devices, so they program those addresses to hardware using the
__dev_mc_add() sync and unsync callbacks.
Unfortunately both cannot work at the same time, and it seems that the
multicast addresses which are already present on the DSA master, like
33:33:00:00:00:01 (added by addrconf.c as in6addr_linklocal_allnodes)
are synced to the master via dev_mc_sync(), but not to hardware by
__dev_mc_sync().
This happens because both the dev_mc_sync() -> __hw_addr_sync_one()
code path, as well as __dev_mc_sync() -> __hw_addr_sync_dev(), operate
on the same variable: ha->sync_cnt, in a way that causes the "sync"
method (dsa_slave_sync_mc) to no longer be called.
To fix the issue we need to work with the API in the way in which it was
intended to be used, and therefore, call dev_uc_add() and friends for
each individual hardware address, from the sync and unsync callbacks.
Fixes: 5e8a1e03aa4d ("net: dsa: install secondary unicast and multicast addresses as host FDB/MDB")
Link: https://lore.kernel.org/netdev/20220321163213.lrn5sk7m6grighbl@skbuf/
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20220322003701.2056895-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Diffstat (limited to 'net/dsa/master.c')
0 files changed, 0 insertions, 0 deletions