: # cp /net/bogus/.../quake2-3.20-glibc-i386-unknown-linux2.0.tar.gz .
: # md5sum quake2-3.20-glibc-i386-unknown-linux2.0.tar.gz
: c545441d63e1fff78835ee7e871033c0
This is probably the 'more correct' value, because you read from
the cache.
: # sync
: # shutdown -r now
: [after reboot, login]
: # md5sum quake2-3.20-glibc-i386-unknown-linux2.0.tar.gz
: 7b5528c06d86ebf5f45016620c9014bd
This is not from the cache, but what you wrote to the disk.
: This is 100% reproducible.
By that you probably mean, more md5summing gives the same wrong value.
I had a similar problem, which took some time to chase down.
The ide cable for the second controller had no devices attached
and was working like an antenna, interfering with the first channel.
During hard disk writes sometimes 2 bytes got duplicated (a spike on a
control line probably). Later with a new sector the shift
was lost. So the file length were correct.
I got rid of the second cable, and voila, no more problems.
I would check the ide cable setup for your box.
Heiko
PS: a happy new year to all
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/