| 1 |
Ciaran McCreesh wrote: |
| 2 |
> On Thu, 03 Apr 2008 14:29:10 +0200 |
| 3 |
> Patrick Lauer <bugs@××××××××××××××××××××××.org> wrote: |
| 4 |
> |
| 5 |
>>> Nope. In fact, using such a system, there are ways of getting in |
| 6 |
>>> code that doesn't get triggered until someone's key gets |
| 7 |
>>> invalidated. |
| 8 |
>>> |
| 9 |
>> By this reasoning you shouldn't use passwords ... |
| 10 |
>> |
| 11 |
>> The idea is to limit the attack vectors and make simple attacks much |
| 12 |
>> harder. A sophisticated "hacker" could just rent a busload of angry |
| 13 |
>> serbians, kidnap 12 developers and force them to do some subtle |
| 14 |
>> changes in many places. But is that likely to happen? |
| 15 |
>> |
| 16 |
> |
| 17 |
> No no. The point is, there's no effective technological way of |
| 18 |
> preventing malicious developers from using the tree to screw over end |
| 19 |
> users. Signing isn't designed to and can't prevent that class of |
| 20 |
> attack (and nor can it protect against compromised end user systems). |
| 21 |
> What it *can* do is reduce the amount of damage done by a compromised |
| 22 |
> rsync server. |
| 23 |
> |
| 24 |
So then we should at first focus the discussion on a few things: |
| 25 |
- what classes of attackers are there |
| 26 |
- what defense mechanisms we can use |
| 27 |
- what the costs (complexity, time, extra code) of each defense is |
| 28 |
|
| 29 |
and then, from that design space, select the option(s) that have the |
| 30 |
best behaviour. If you get bored you can read the not-yet-GLEPs robbat2 |
| 31 |
has written with the help of a few others, which would cut out a large |
| 32 |
part of the discussion: |
| 33 |
http://viewcvs.gentoo.org/viewcvs.py/gentoo/users/robbat2/tree-signing-gleps/ |
| 34 |
|
| 35 |
That's exactly the thing under discussion -- the design of the system |
| 36 |
> necessitates trust in both the main repository and the end user system, |
| 37 |
> and signing does absolutely nothing to help there. No-one is suggesting |
| 38 |
> that anyone from infra is going to do anything to utterly screw over |
| 39 |
> Gentoo for petty personal reasons. |
| 40 |
> |
| 41 |
But if you don't trust anyone there is no reason why you would even try |
| 42 |
to interact with Gentoo. So at some point you will have to decide to |
| 43 |
arbitrarily trust a few entities, be it devs or servers or cryptographic |
| 44 |
keys ... |
| 45 |
|
| 46 |
|
| 47 |
|
| 48 |
-- |
| 49 |
gentoo-dev@l.g.o mailing list |
Archives