On Fri, 26 Oct 2001, Adam Warner wrote: > Most of the kernel patch problems may have been overcome. I don't think > Jari's approach has ever caused data corruption. If you read through the > archives you'll be up to speed with developments. I understand that > sometimes if you upgrade your kernel to 2.4 with the International patch > you can't access a volume generated with a 2.2 kernel! At the time I > read that a solution had not been found.
The old code used absolute disc sector numbers as an initialisation vector; the new code can be switched between using absolute end relative numbers. Using absolute numbers means that you cannot copy the encrypted filesystem to a different place. There seems to be another problem with the crypto patch and kernels >=2.4.10: It simply doesn't work! (Probably because of the block device in page cache changes) Walter
