On Sun, May 11, 2003 at 06:35:02PM +0300, Jari Ruusu wrote:
> > I also brought it up on this list in 2001-September.
> >
> > Both Jari and HVR recommend against using cryptoapi (kerneli) for swap.
>
> I don't remember saying that, at least not for any recent version of
It wasn't recent. :) It was in 2001-September, here in this list. But
it's good to hear things have improved.
> Kerneli.org's loop-jari is a copy of old and buggy version of loop-AES' loop
> code. Last time I checked, loop-jari had nasty 'process stuck in D state'
> bug still present. That bug is fixed in loop-AES-v1.7b and later versions.
> v1.7b was released in January 2003.
-rw-rw-r-- 1 korg korg 33342 Dec 2 23:21 loop-jari-2.4.20.0.patch
Ah, so it's really "loop-jari-old". :) I assumed that HVR was taking the
loop-jari patches directly from your loop-AES releases, but I suppose he
hasn't done it very recently.
Again in the "works for me" category, I'm running the swap encryption on
2.4.17, and the loop-jari patch I used was a year prior to that. I've
probably just been lucky. I'll have to build a new kernel there soon ...
but not before it hits 200 days' uptime next week. :)
BTW I used to have problems with processes stuck in D state; I didn't
think to ask about that here. I never really figured out the problem,
but IIRC it went away with 2.4.17. Before that I used 2.4.10. I think I
might have used swap encryption then. <digging through /boot> Yes, it
seems I was. I built the 2.4.10 on 2001/09/29, and the present 2.4.17 on
2002/03/28.
Thanks for the reply.
Rob - /dev/rob0
-
Linux-crypto: cryptography in and on the Linux system
Archive: http://mail.nl.linux.org/linux-crypto/
