(Resent as plain text, the first copy was HTML and did not reach the list.)
Hi Pablo, On Thu, Oct 15, 2026 at 12:34:06PM +0200, Pablo Neira Ayuso wrote: > > Assisted-by: Claude:claude-opus-5-5 > > I did not use this LLM assistance for my patch and I show as > co-developer here ;-) The tag covers my changes on top of yours, v8 says so below ---. > This comment is not needed? LLM likes to bloat code with unnecessary comments. Removed. > Why LLM decided to rename this function? :-) No reason, back to fdb_valid_dst(). > I think bridge_ifidx is not needed here and I am not sure we can cache > this value? Dropped. The gc now walks from out.ifidx to the bridge port through the vlan devices in the encap. A 2-bit port_vlans says how many there are. > Why do we need bridge_vid? See below to explain an alternative. [...] > A new flag might be need, but there is room for new flags to annotate > this VLAN header came from bridge filter filtering, I understand > info->encap is not enough. Dropped too, there is a bridge_vlan flag now. For an untagged port on a vlan_filtering bridge the vid is not in the encap, so the helper uses the port pvid. > Why LLM moved this away from nf_flow_dst_check()? Moved back, nf_flow_dst_valid() calls both. v8: https://lore.kernel.org/netfilter-devel/[email protected]/ Thanks, Julius
