On Fri, Jan 28, 2011 at 02:02:15AM -0800, Yongqiang Liu wrote: > Hi Everyone: > > For running ovs on Android phone, so far i have installed openvswitch_mod.ko > on Android kernel and cross compiled ovs-dpctl. > It seems they work on Android: > 1. I can create a datapath dp0 > 2. i can add phone's wifi interface (tiwlan0) to the dp0 > > the following is the dp0 status: > # ./ovs-dpctl show > system@dp0: > flows: cur:0, soft-max:1024, hard-max:1048576 > ports: cur:2, max:1024 > groups: max:16 > lookups: frags:0, hit:0, missed:4, lost:0 > queues: max-miss:100, max-action:100 > port 0: dp0 (internal) > port 1: tiwlan0 > Jan 28 01:48:39|00001|dpif_linux|WARN|/dev/net/dp46388 is device 248:0 but > should be 248:1, fixing > Jan 28 01:48:39|00002|dpif_linux|WARN|/dev/net/dp46388 is device 248:1 but > should be 248:2, fixing > Jan 28 01:48:39|00003|dpif_linux|WARN|/dev/net/dp46388 is device 248:2 but > should be 248:3, fixing > Jan 28 01:48:39|00004|dpif_linux|WARN|/dev/net/dp46388 is device 248:3 but > should be 248:4, fixing > Jan 28 01:48:39|00005|dpif_linux|WARN|/dev/net/dp46388 is device 248:4 but > should be 248:5, fixing > > But i don't know reason of the dpif_linux WARN message, and is it a > potential problem, how to fix it?
That is a really strange error message. The openvswitch_mod kernel module creates a character device driver, and the userspace components create corresponding device nodes /dev/net/dp<N> as needed. But the device nodes are only numbered 0 through 255, so /dev/net/dp46388 doesn't make any sense and I don't understand why ovs-dpctl would be trying to create it. Could you run "ovs-dpctl -v show" and post the output? There will be a lot of output. The output might give some hints. _______________________________________________ discuss mailing list discuss@openvswitch.org http://openvswitch.org/mailman/listinfo/discuss_openvswitch.org
