Ohh, crap :) Ok, maybe Jan? @Jan: if this is wrong again, can you point me to a kernel developer that is working with the I/O scheduler?