| 1 |
>>> Goals: |
| 2 |
>>> ------ |
| 3 |
>>> - protect against compromised developer box / rogue developer |
| 4 |
>>> - protect against compromised rsync server |
| 5 |
>> |
| 6 |
>> Exactly how secure are we aiming at? The schemes presented here does |
| 7 |
>> nothing to secure gentoo boxes from malicious source code. |
| 8 |
> |
| 9 |
> That is completely not the intent of this system. This system is |
| 10 |
> intended to stop tampering with the portage tree. The portage tree is |
| 11 |
> dangerous in it's nature of being executable shell. |
| 12 |
|
| 13 |
Before discussing implementation, you should decide on the goals. |
| 14 |
|
| 15 |
I agree with Paul and Robin in that the goals should stay in reality. We |
| 16 |
have a big problem today : Gentoo global security depends on factors |
| 17 |
outside Gentoo's control : the security of each server in the rsync |
| 18 |
mirror servers network. It's a flaw that needs to be acted on very quickly. |
| 19 |
|
| 20 |
We are not trying to discuss ways of having the most secure Linux |
| 21 |
distribution out there, review all sources, treat the rogue developer |
| 22 |
problem ; we are trying to have a secure distribution mecanism which |
| 23 |
does not depend on outside factors. For the moment we don't have that. |
| 24 |
Once this is done, it will always be time to discuss better mecanisms to |
| 25 |
ensure better security. You will probably find in the end that the |
| 26 |
increased-security/work-overhead tradeoff to solve the rogue dev case is |
| 27 |
not acceptable. |
| 28 |
|
| 29 |
So for for me the only objective is : |
| 30 |
|
| 31 |
* protect against compromised rsync server |
| 32 |
|
| 33 |
|
| 34 |
Just my 2c |
| 35 |
|
| 36 |
-K |
| 37 |
|
| 38 |
-- |
| 39 |
gentoo-dev@g.o mailing list |
Archives