I have a SFP+ card from delock. The device driver ixgbe is used.
The problem is that, if I set force_link to 0 and use the option proto dhcp for this interface. I get into a infinite loop, so that I see the following lines in the logs again and again.
Thu Mar 9 15:21:48 2023 kern.info kernel: [ 4409.718308] ixgbe 0000:01:00.0: registered PHC device on eth4
Thu Mar 9 15:21:48 2023 kern.info kernel: [ 4409.719227] 8021q: adding VLAN 0 to HW filter on device eth4
Thu Mar 9 15:21:48 2023 kern.info kernel: [ 4409.784399] ixgbe 0000:01:00.0 eth4: detected SFP+: 3
Thu Mar 9 15:21:48 2023 daemon.notice netifd: Network device 'eth4' link is up
Thu Mar 9 15:21:48 2023 daemon.notice netifd: Interface 'wan' has link connectivity
Thu Mar 9 15:21:48 2023 kern.info kernel: [ 4410.496141] ixgbe 0000:01:00.0 eth4: NIC Link is Up 10 Gbps, Flow Control: RX/TX
Thu Mar 9 15:21:48 2023 kern.info kernel: [ 4410.496310] IPv6: ADDRCONF(NETDEV_CHANGE): eth4: link becomes ready
Thu Mar 9 15:21:49 2023 daemon.notice netifd: Network device 'eth4' link is down
Thu Mar 9 15:21:49 2023 daemon.notice netifd: Interface 'wan' has link connectivity loss
Thu Mar 9 15:21:49 2023 kern.info kernel: [ 4411.513577] ixgbe 0000:01:00.0 eth4: NIC Link is Down
Thu Mar 9 15:21:51 2023 daemon.notice netifd: Network device 'eth4' link is up
Thu Mar 9 15:21:51 2023 daemon.notice netifd: Interface 'wan' has link connectivity
Thu Mar 9 15:21:51 2023 kern.info kernel: [ 4412.726137] ixgbe 0000:01:00.0 eth4: NIC Link is Up 10 Gbps, Flow Control: RX/TX
If I set the force_link option to 1, I get an IP from the DHCP server and the interface is working as expected.
My guess is that, if the force_link is set to 0, the netifd gets 2 up events from the Kernel. And if the netifd has already received an up event from this interface and then receives another one, then it first takes the interface down.
Does anyone else see this behaviour for the ixgbe device. I have tried both the owrt-21.02 with kernel 5.4 and the master with kernel 5.15. Same behavior.