1 |
On Tue, 10 Feb 2004 15:24:34 +0100 |
2 |
Daniel Heemann <daniel.heemann@×××.de> wrote: |
3 |
|
4 |
> On Tuesday 10 February 2004 14:15, Daniel Brandt wrote: |
5 |
> > If you know you won't find an attacker _before_ he's playing with your |
6 |
> > compiler you should be more worried about your perimeter. |
7 |
> I do not know that - but do you know that you'll find him before? |
8 |
> |
9 |
|
10 |
I'm certain of it. As certain as I can be with honeytokens, host-based IDS, homebrew scripts and years of experience. |
11 |
|
12 |
> Hmm, let's say the attacker gains access to the machine, the firewall blocks |
13 |
> all binary transfer (I know uuencode/decode, but lets think the attacker is |
14 |
> not in the position to transfer executables onto the compromised system, |
15 |
> perhaps he can't transfer any files) and the attacker only needs 10 lines |
16 |
> of c-code to exploit the kernel or whatever - don't worry about if he can |
17 |
> compile the 10 lines or not? |
18 |
> Perhaps also the system runs on alpha hardware but the attacker only has x86 |
19 |
> binaries etc.. |
20 |
|
21 |
Too hypothetic, transfer of a binary file will be possible if there is two-way communication. Non x86 hardware is not a problem because of cross compiling. |
22 |
|
23 |
> never said that bsd is better/inferior btw... |
24 |
|
25 |
Understood. |
26 |
|
27 |
> I don't want to discuss this further on this list as it's not a gentoo |
28 |
> specific problem and I still think that for production systems there are |
29 |
> some alternatives available... |
30 |
> |
31 |
|
32 |
This is a gentoo issue since gentoo ships with a live compiler in it's default setting, making it a must. |
33 |
|
34 |
But whatever the case, alternatives exist and individual security measures are always of some use. |
35 |
|
36 |
Cheers, |
37 |
Daniel |
38 |
|
39 |
-- |
40 |
gentoo-security@g.o mailing list |