Skip to contentSkip to conversation

2.6.3-mm1

DISCUSSION Andi Kleen 3 participants 7 messages Feb 18, 2004 – Feb 19, 2004 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Andi Kleen

author
#1 lore raw
Quoted 12 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.3/2.6.3-mm1/

- Added the dm-crypt driver: a crypto layer for device-mapper.

People need to test and use this please. There is documentation at
http://www.saout.de/misc/dm-crypt/.

We should get this tested and merged up. We can then remove the nasty bio remapping code from the loop driver. This will remove the current ordering guarantees which the loop driver provides for journalled filesystems. ie: ext3 on cryptoloop will no longer be crash-proof.

Quoted 4 lines

After that we should remove cryptoloop altogether.

It's a bit late but cyptoloop hasn't been there for long anyway and it doesn't even work right with highmem systems (that part is fixed in -mm).

Is it guaranteed that this thing will be disk format compatible to cryptoloop? (mainly in IVs and crypto algorithms)

While 2.3 and 2.4 have broken the on disk format of crypto loop several times (each time to a new "improved and ultimately perfect format") I don't think that's acceptable for a mature OS anymore.

-Andi

#2 lore raw
Quoted 15 lines · Andi Kleen
Quoted 12 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.3/2.6.3-mm1/

- Added the dm-crypt driver: a crypto layer for device-mapper.

People need to test and use this please. There is documentation at
http://www.saout.de/misc/dm-crypt/.

We should get this tested and merged up. We can then remove the nasty bio remapping code from the loop driver. This will remove the current ordering guarantees which the loop driver provides for journalled filesystems. ie: ext3 on cryptoloop will no longer be crash-proof.

Quoted 7 lines
Quoted 4 lines

After that we should remove cryptoloop altogether.

It's a bit late but cyptoloop hasn't been there for long anyway and it doesn't even work right with highmem systems (that part is fixed in -mm).

Is it guaranteed that this thing will be disk format compatible to cryptoloop? (mainly in IVs and crypto algorithms)

Allegedly. Of course, doing this will simply retain crypto-loop's security weaknesses.

Quoted 3 lines

While 2.3 and 2.4 have broken the on disk format of crypto loop several times (each time to a new "improved and ultimately perfect format") I don't think that's acceptable for a mature OS anymore.

Well I guess people are free to do that sort of thing with out-of-kernel patches.

One question which needs to be adressed is whether dm-crypt adequately addresses crypto-loop's security weaknesses, and if so, how one should set it up to do so.

author
#3 lore raw
Quoted 31 lines · On Wed, 18 Feb 2004 02:55:49 -0800 Andrew Morton <[email protected]> wrote:
Quoted 15 lines · Andi Kleen
Quoted 12 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.3/2.6.3-mm1/

- Added the dm-crypt driver: a crypto layer for device-mapper.

People need to test and use this please. There is documentation at
http://www.saout.de/misc/dm-crypt/.

We should get this tested and merged up. We can then remove the nasty bio remapping code from the loop driver. This will remove the current ordering guarantees which the loop driver provides for journalled filesystems. ie: ext3 on cryptoloop will no longer be crash-proof.

Quoted 7 lines
Quoted 4 lines

After that we should remove cryptoloop altogether.

It's a bit late but cyptoloop hasn't been there for long anyway and it doesn't even work right with highmem systems (that part is fixed in -mm).

Is it guaranteed that this thing will be disk format compatible to cryptoloop? (mainly in IVs and crypto algorithms)

Allegedly. Of course, doing this will simply retain crypto-loop's security weaknesses.

Quoted 3 lines

While 2.3 and 2.4 have broken the on disk format of crypto loop several times (each time to a new "improved and ultimately perfect format") I don't think that's acceptable for a mature OS anymore.

Quoted 6 lines

Well I guess people are free to do that sort of thing with out-of-kernel patches.

One question which needs to be adressed is whether dm-crypt adequately addresses crypto-loop's security weaknesses, and if so, how one should set it up to do so.

AFAIK the two big security weaknesses in most version of cryptoloop are:
(note that some versions have it fixed with various hacks)

- Weak IV
- Extremly bad key management

The first can be addressed in an crypto API module e.g. with an hashed IV, but it needs a stable IV format from dm-crypto (that is one of the things I asked for)

The second one is more a user space problem. However to solve it you need metadata. It would be much nicer if it was possible to store it on the block device directly (with a special losetup flag for compatibility). Otherwise you get into nasty chicken and egg problems with fully encrypted systems.

Supporting metadata can be quite simple - e.g. a standard header on the first blocks that has a length and a number of records with unique IDs. And the kernel driver would need to skip over these headers.

-Andi

#4 lore raw
Quoted 3 lines · Andi Kleen

Supporting metadata can be quite simple - e.g. a standard header on the first blocks that has a length and a number of records with unique IDs. And the kernel driver would need to skip over these headers.

The target already takes an offset into the device, so you have what you want.

- Joe

author
#5 lore raw
Quoted 6 lines · On Wed, 18 Feb 2004 13:45:58 +0000 Joe Thornber <[email protected]> wrote:
Quoted 3 lines · Andi Kleen

Supporting metadata can be quite simple - e.g. a standard header on the first blocks that has a length and a number of records with unique IDs. And the kernel driver would need to skip over these headers.

The target already takes an offset into the device, so you have what you want.

Ok fine. The only requirement would be compatible IVs then.

-Andi

#6 lore raw
Quoted 4 lines · Andi Kleen

On Wed, 18 Feb 2004 13:45:58 +0000
Joe Thornber <[email protected]> wrote:

Quoted 9 lines
Quoted 6 lines · Andi Kleen
Quoted 3 lines

Supporting metadata can be quite simple - e.g. a standard header on the first blocks that has a length and a number of records with unique IDs. And the kernel driver would need to skip over these headers.

The target already takes an offset into the device, so you have what you want.

Ok fine. The only requirement would be compatible IVs then.

Would it be correct to say that until someone does this development, dm-crypt has the same vulnerabilities as cryptoloop? Or is there some different way of using dm-crypt which is incompatible with cryptoloop, but is more secure?

author
#7 lore raw
Quoted 10 lines · On Wed, 18 Feb 2004 15:27:45 -0800 Andrew Morton <[email protected]> wrote:
Quoted 9 lines · Andi Kleen
Quoted 5 lines · On Wed, 18 Feb 2004 13:45:58 +0000 Joe Thornber <[email protected]> wrote:
Quoted 3 lines · Andi Kleen

Supporting metadata can be quite simple - e.g. a standard header on the first blocks that has a length and a number of records with unique IDs. And the kernel driver would need to skip over these headers.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

The target already takes an offset into the device, so you have what you want.

Ok fine. The only requirement would be compatible IVs then.

Would it be correct to say that until someone does this development, dm-crypt has the same vulnerabilities as cryptoloop? Or is there some

It all depends on what the crypto API module and the user space utils implement. That is why calling it dm-crypt is misleading, dm-filter or somesuch would be better.

Quoted 2 lines

different way of using dm-crypt which is incompatible with cryptoloop, but is more secure?

More secure will be incompatible.

-Andi