On Wed, 2012-08-15 at 20:08 +0700, Mulyadi Santosa wrote: > Hi Ankur :) Hi Mulyadi, Thanks for your reply! <snip> > > Just a quick test, what happen if you do: > insmod -f /path/to/the/module.ko > ? I didn't try the -f switch. (The man page I have in Fedora doesn't even mention it). Unfortunately, the robot is maintained by another laboratory: we need to take permission and make an appointment to use it. I'll try this the next time I have an appointment. > > IMHO, quite likely the message you saw happened since you compile the > kernel module with different version of gcc and/or different gcc > parameter, that somehow yield binary incompatibility. Probably :/ . The manufacturers haven't mentioned what version of gcc etc. they used. I used the one I had available: > gcc-4.7.0-5.fc17.x86_64 > > Another thought, what arch do you use when you compile that kernel > module? AMD Geode too? > No. I compiled it on my system, which is an x86_64 using the ARCH parameter to the make command: make ARCH=i386 ... (I got this from here[1]) The system they provide (the Geode), along with their file system etc. has no development tools at all, so I can't build the module there (unless I first compile gcc etc from source, like LFS?). It seems to use a live persistent USB (which has a data and system partition), although none of the partitions are marked bootable. I'm guessing it has a bootloader on-board, but again, no info on that either. [1] http://www.linuxtopia.org/online_books/linux_kernel/kernel_configuration/ch05s04.html
Attachment:
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Kernelnewbies mailing list Kernelnewbies@xxxxxxxxxxxxxxxxx http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
