Hi: I'm sorry if this goes in Hardware. I got: The physical sector has, among other data, the user data (part of the kernel here) and a 32-bit[?] cyclic redundancy check number computed from the user data. So this is a hardware error. Either there was an I/O error or some bits spontaneously changed in the CRC or in the user data. Whatever, I did a 'e2fsck -c -k -C 0 /dev/hda1' and it output no errors. So I created /boot2, copied the three files (one of them previously I uncompressed it) in /kernels from the instalation CD (slackware 12.0), in an attempt to free myself from the CRC error by moving the kernel image to another area in the disk and reran lilo, pointing to him what the new /boot directory was. Now I got a clean boot. Latter on, I'll run 'e2fsck -cc -k -C 0 /dev/hda1', which does a non-destructive r/w test on the disk.

I hope this can serve somebody with a simmilar problem and wait for comments.

Are you using loadlin there?

No, I'm not. The hard disk unit was manufactured in 2006, a Western Digital Caviar family WD1200BB-55GU0. S.M.A.R.T. data:Then the unit has been in use for 4076 hours. Don't know what's the probability of getting one error in the first 4000 hours for a disk from 2006.