On Mon, Jul 26, 2021 at 10:14 AM Klaus Kudielka <klaus.kudie...@gmail.com> wrote: > > > I tried to run OpenWrt with a bleeding-edge net-next kernel (commit > > 0e804326759d), then noticed that only the first port of a bridge is > > brought up automatically, the rest are down and not members of the > > bridge. > > > i.e. with the following settings > > network.@device > > <https://lists.openwrt.org/mailman/listinfo/openwrt-devel>[0]=device > > network.@device > <https://lists.openwrt.org/mailman/listinfo/openwrt-devel>[0].ports='lan1' > 'lan2' 'lan3' 'lan4' > > network.@device > <https://lists.openwrt.org/mailman/listinfo/openwrt-devel>[0].type='bridge' > > network.@device > <https://lists.openwrt.org/mailman/listinfo/openwrt-devel>[0].name='br-lan' > > > Only lan1 is in br-lan, and I have to manually configure the rest: > > Exactly the same happens to me on Turris Omnia with master, either 5.4 > OpenWrt kernel or 5.10 OpenWrt kernel. > > I bisected this to "2801fe6132c4e2e364e2d5a304594185351b501b netifd: > update to the latest version". > > And, I believe, Petr has noticed the same thing with 21.02. > > Regards, Klaus > > > _______________________________________________ > openwrt-devel mailing list > openwrt-devel@lists.openwrt.org > https://lists.openwrt.org/mailman/listinfo/openwrt-devel
I ran into this on our mt7622 (mediatek) target as well. It appears to be related to this netifd commit specifically: https://git.openwrt.org/?p=project/netifd.git;a=commit;h=85f01c44a950be8518ce5a7d251b5bba219348cf I’m not sure exactly where the issue is just yet but will take a look and see if it’s something I can identify. Thanks, -Chad _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel