Jari Ruusu wrote: > > Peter Niemayer wrote: > > Yes, works great! Thanks a lot for fixing that bug! > > I have to ask one more question: Can you (or anyone else on the list) test > same changes for 2.2.20 kernel with 2048 hard sector size? I will post new > diff on the list if someone promises to test it. Alas, I don't have a system with tools old enough to allow for 2.2 boots - I'm just using too many fancy new features... I know... :-) I'll ask if some of the other people I know who own MO drives still uses 2.2 kernels. > 2.5 still has that > bug, and there is no easy way to fix that currently. In 2.5 all loop devices > _share_ same hard sector size... OUCH. Uh.. that reminds me of yet another 512-byte-nightmare not that long ago... with 2.2 kernels, FAT filesystems on 2048 byte/sector media worked just fine. Then came 2.4.0, and with it the inability of Linux to use FAT on such devices. It took loads of mails and 7 minor version numbers until this ugly bug was fixed, later... And then, there's yet another such bug in the kernel: The ufs filesystem makes the false assumption that some vendors variants of ufs are supporting 512 byte sectors, only. I had to do a quick dirty hack (hardwiring 2048 byte/sector) to read some old ufs media... It seems some automatic tool should reject any patch send to Linus that contains the constant "512" but not "1024", "2048" and "4096" as well... :-) Regards, Peter Niemayer - Linux-crypto: cryptography in and on the Linux system Archive: http://mail.nl.linux.org/linux-crypto/
