This may be related to an earlier issue that I posted. Same device; I'm now on 25-12.0 stable. The original issue still exists.
I have a firewall zone called lan6 which includes the interfaces eth0 and wglan6.
eth0 has the IP address 2XXX:XXXX:XXXX:XXXf::1/64. wglan6 has the IP address 2XXX:XXXX:XXXX:YYY0::1/64.
When connected to wglan6 (over Wireguard), I can ping both IP addresses.
When connected to eth0 (over Ethernet), I cannot ping 2XXX:XXXX:XXXX:XXXf::1/64 (the address for eth0). I can ping wglan6's address and Internet destinations.
I created a firewall rule and put it at the top which accepts and logs all ICMP packets sent from the device to the router. The firewall rule was first set to capture packets with destination any zone (forward), and I could ping Internet destinations and see corresponding logs. It was then changed to capture packets with destination this device (input), and when I tried pinging the router using eth0's address, it didn't work and no logs were generated.
netstat -rn on the connected device includes the following:
where ZZZZ represents another device connected to the same port which I can ping.
I suspect the differing gateways are the culprit here, but I don't know anything beyond that.
With my extremely limited knowledge on how networking works below the IP level, could it perhaps be the case that unless a packet is addressed specifically to the device's link-local or MAC address (which maybe it isn't if netstat's gateway is the link itself) it just never gets accepted for processing? To test this I tried enabling promiscuous mode on the device, but after doing that I wasn't able to get an IP address assigned in the first place.