Hans Dedecker <dedec...@gmail.com> wrote: > By the introduction of linksensing the user experience changed when > using the proto static; before the static IP address was present in > the kernel when the link was down after the introduction of > linksensing the static IP address is not present anymore. I'm not > aware of other bugs. > >
Given that this is the way regular desktop linux works too, are we digging a hole here where OpenWrt keeps providing the old behaviour? My desktop has a static ip on one interface, and when I unplug the cable, I get... NetworkManager[995]: <info> (enp0s20u1u1): link disconnected (deferring action for 4 seconds) NetworkManager[995]: <info> (enp0s20u1u1): link disconnected (calling deferred action) NetworkManager[995]: <info> (enp0s20u1u1): device state change: activated -> unavailable (reason 'carrier-changed') [100 20 40] NetworkManager[995]: <info> (enp0s20u1u1): deactivating device (reason 'carrier-changed') [40] And then the static IP is gone too. It might have been a "behaviour change" but lots of things change. Was this really something critical enough to try and keep? > > Further, if netifd is the one adding force_link, then surely it should > > be removing it when you change to dhcp, not just leaving there like a > > land mine > Tested this on the netifd trunk version; switching proto static to > dhcp and did not observe an issue regarding force_link as the dhcp > client was teared down when the link cable was unplugged. Do you use > the netifd trunk version ? no idea, what version would that be exactly? netifd packages just have a date and a git tag :) I'm going to run some more tests on this, thanks for the information and background so far. Cheers, Karl Palsson
_______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel