Messages in this thread | | | Date | 26 Nov 1998 09:05:03 -0000 | From | HyperReal-Anon <> | Subject | Migration from Mileski to kerneli |
| |
On Mon, 23 Nov 1998 kernel@mallory.draper.net wrote:
> On Mon, Nov 23, 1998 at 12:13:59PM +0100, Alexander Kjeldaas wrote: > <snip> > > I don't think you should send your patches to me - I won't accept > > them. However, what you can do (if you bother to do so) is to go > > through your patch and describe the changes you've done. Send the > > description to me. I can take your description and reimplement your > > patch. This should probably work for your patch since it seems that > > the changes are pretty small. > > Ah, I see. Well, I don't think that would be very productive. I am > curious, are you refusing patches from just me, the USA, or the entire > world?
Tell me about it. I'd love to work on this, but I'm in the USA so I won't tempt fate.
This does bring me to a question. How would a person migrate gracefully from Mileski's loop patches for 2.1.98 to the 2.1.129 international patches? The 2.1.129 international patches don't seem to support IDEA or CAST-128, though they do support CAST-256, and a quick look at the included patches for mount-2.5l don't show support for some of the newer algorithms the international patch covers, like MARS.
I have a lot of data encrypted with Mileski's IDEA patch over here. I tried to hack IDEA support into 2.1.126 and while I learned a lot in the process, my implementation of the algorithm itself (which was ripped directly from Mileski's patch) didn't work. I'm assuming it's because I effectively used some other block chaining scheme and didn't realize it. I was trying to use the code in the newer loop driver for 2.1.126. I'm not a cryptographer.
For personal use only I could hack in whatever algorithms I wanted, but I hate to have to move several gigabytes of encrypted data spread about several partitions to another set of partitions because the methods are incompatable.
I suppose that ideally I could figure out the differences in implementation, then write a program which reads the cyphertext block file and spits out another which is 2.1.129-compliant. Of course I have to know what those differences are (if any).
Oh, and I might make a very minor and easy to implement suggestion for losetup, swapon, and mount: Add a flag which uses a totally random passphrase, then generates the appropriate filesystem. For instance, if I want 64MB of encrypted swap, I specify my algorithm, the system gets a random passphrase, and handles the setup (mkswap, swapon) for me. I did this with Mileski's patches and it actually works well.
- 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/
| |