[Severity: Medium]
Can a concurrent transmit still put the old remote's route back into the
cache after this reset?
Yes, the window exists: a transmit that read the old remote_ip before
the reset and reaches dst_cache_get() in a later jiffy refills the
cache with the old route. vxlan_fdb_replace() and the other tunnel
users that reset the cache after changing parameters have the same
window. Closing it needs a change to how dst_cache tracks freshness,
which is out of scope for this fix. Without the patch every move uses
the stale route; with it only this race does.
[Severity: Medium]
Does this overwrite a user-configured egress interface with 0?
Yes, this is a regression, and v4 has the same code. v5 compares
remote_ifindex in the early return only for link-local sources, so
traffic from the remote of an entry with a "via" device takes the early
return and the device stays set. The selftest gets a case for it.
pw-bot: cr