Hi, Norbert, On Nov 25, 2014, at 4:36 AM, Norbert Bonnici wrote:
> Initially I was getting the full boot sequence with the ROACH going to > the NFS boot up sequence, obtaining some IP address and loading a > weird image and failing at it. The message "DHCP client bound to address 192.168.55.101" means that a DHCP server responded to its DHCP request, but the next message of "no boot file name" means that the DHCP reply from the server did not include the name of a boot image file (i.e. the DHCP server is not properly configured to netboot the ROACH). Is this the address you expected the ROACH to get? The weird file name (6537A8C0.img) is base on the byte-swapped IP address in hex. Apparently u-boot defaults to this in the absence of a boot file name from the DHCP server. Then u-boot tries to tftp this weird file from IP address 192.168.55.1, which is probably the DHCP server (and/or the "next server" that the DHCP server provided, if any). The "T" in "Loading: T" means timeout. I don't know if this means the tftp connection timed out or if this is the response to a "file not found" condition, but in any case it's not good. > This morning I tried the same but it's failing to do the boot up > countdown and killing the serial port with weird data. It manages to > crash minicom. This is almost certainly a separate problem. I'm pretty sure a misconfigured DHCP server cannot break a ROACH. Have you tried power cycling the ROACH? Maybe the baud rates have become mismatched? What do you mean by "crash minicom"? HTH, Dave
