Dear Anthony Garland,
In message CACYg8QPf-d=_GrBpqRm9gC1GsANZ7A2D2w=redCSbOv5n_KMWQ@mail.gmail.com you wrote:
I'm working on a recovery tool for an overo COM where I can boot all the way into into Linux over USB. This is similar to the beagle-board recovery tools. I have a host computer send the overo a modified version of x-loader which will ask the host computer for more files to store at an address. I have the host computer send it u-boot and
Your version of U-Boot is pretty old, and you are still using X-Loader instead of SPL - these are too many old components to spend efforts in analyzing potential problems. Can you please update instead?
I thought that u-boot did a check sum on uMulti. If uMulti is being corrupted during the USB transfer by noise on the power supply, why isn't u-boot check sum catching it??
Maybe you switched it off?
I've attached the script I use to make uMulti, my Boot_usb.sh script, and a few logs of it randomly failing and succeeding.
You have other problems, like instable / incorrectly intialized memory.
# For whatever reason, the gzip files seem more error prone. With no compression I seem to get better results.
Wrong. Without compression errors don't get caought so easily. I bet they are still there.
...
Contents: Image 0: 2585928 Bytes = 2.5 MiB Image 1: 13631488 Bytes = 13 MiB Verifying Checksum ... OK
As you can see, it _does_ checksum verification.
[ 0.000000] Linux version 3.2.0 (anthony@linux-server) (gcc version 4.3.3 (GCC) ) #1 Thu Jun 28 16:13:37 CDT 2012
Uh, oh. GCC 4.3 - can you _please_ try a more recent and less buggy tool chain?
Recent kernel versions and ancient tools usually don't mix without problems.
[ 8.032104] alignment: ignoring faults is unsafe on this CPU. Defaulting to fixup mode. Remounting root file system... WARNING: -e needs -E or -F Segmentation fault
Alignments issues ?
Best regards,
Wolfgang Denk