Hello Yongqin, > is there any message that I could check to make sure if the > initialization is finished? > or like with adding some printk lines for some kernel functions to hack > >> Anyway, I will try to reproduce here and analyze it. > > Thanks very much! And please feel free to let me know if there is > anything I could help with on the Android build. I have finally managed to reproduce an error similar to the one mentioned during the boot stage. I created a systemd service with a similar configuration script to do the same and it works (I can reconfigure the mac address at boot time, the ip address is configured and the interface works), because the driver is completely initialized. In order to reproduce, I have introduced a big delay in the probe operation to get closer in time to the configuration script and the problem is there. Maybe, the script set_ethaddr.sh could be synchronized with the driver, but I think, if possible, it is better to check in a better way in the driver; I will try it. When I have something I can comment you, if you can test it, to be sure about the solution. By the way, I have tried with my other fix to avoid the spurious link messages, but it didn't help. Best regards José Ignacio
