1 |
Hi! |
2 |
|
3 |
On Wed, 22 Sep 2004, John Richard Moser wrote: |
4 |
> The most immediately apparent route to take would be to have ebuilds |
5 |
> such as openssh, apache, and su stack smash protected. This would |
6 |
> prevent common buffer overflow attacks from being used to compromise |
7 |
> security; such attacks would only cause the program attacked to abort, |
8 |
> which could still be used as a Denial of Service attack, but would not |
9 |
> allow successful intrusion. |
10 |
|
11 |
One might make a case for "this provides a (false) sense of |
12 |
security". But I wouldn't think of this kind of view as feasible. |
13 |
Using that argument, one could drop just about every security |
14 |
measure that isn't perfect (and which is?). |
15 |
|
16 |
> Gentoo ships gcc with stack smash protection built in. This is |
17 |
> activated by -fstack-protector or -fstack-protector-all. It would be |
18 |
> feasible to add one of these flags to an ebuild based on a FEATURES or |
19 |
> USE setting. |
20 |
|
21 |
Nice idea, I agree. |
22 |
|
23 |
> I believe it would be a good idea to have such a FEATURES or USE flag on |
24 |
> by default in all profiles where SSP is supported. In this manner, the |
25 |
> major targets of security attacks would automatically be protected; |
26 |
> while still allowing the user to disable the protection if the user |
27 |
> desires. Users wanting more protection can simply add -fstack-protector |
28 |
> to CFLAGS, or use Hardened Gentoo. |
29 |
> |
30 |
> Any comments? Would this be more suitable as a USE or a FEATURES setting? |
31 |
|
32 |
I *think* it'd be better used as a USE flag for the "daemon-only" |
33 |
approach. As for the "whole system" approach, I'd lean towards |
34 |
FEATURE. But then I'm not well-versed in the policies regarding |
35 |
those two settings/variables. |
36 |
|
37 |
One problem (albeit small) I see is upstream maintainers refusing |
38 |
to support users who use these settings with their software. A |
39 |
similar example is the Mplayer team's stance towards |
40 |
optimization. Usually, this could be solved the same way as in |
41 |
the mplayer case (if the afflicted^Waffected maintainers aren't |
42 |
too many). |
43 |
|
44 |
Just my two eurocents, |
45 |
Tobias |
46 |
-- |
47 |
export DISPLAY=vt100 |