| 1 |
On Sat, 19 Nov 2005 11:41:54 +0000 Mike Auty <m.auty@××××××××.net> |
| 2 |
wrote: |
| 3 |
| So now the task is for an untrusted user to get a dev to sign |
| 4 |
| some long random value that hashes to one thing, and then replace it |
| 5 |
| with a different long random value when a user finally downloads it. |
| 6 |
| It must then be interpreted by portage and execute some malicious |
| 7 |
| code (and if I've read news about savior correctly, this will be |
| 8 |
| sandboxed until the merge stage, so most of the ebuild must be |
| 9 |
| valid). Does this not seem a little unlikely? |
| 10 |
|
| 11 |
Ah, see, you're missing the attack vector. You don't need to replace a |
| 12 |
whole segment of code. You just need to flip one bit which is used as a |
| 13 |
trigger. |
| 14 |
|
| 15 |
-- |
| 16 |
Ciaran McCreesh : Gentoo Developer (Look! Shiny things!) |
| 17 |
Mail : ciaranm at gentoo.org |
| 18 |
Web : http://dev.gentoo.org/~ciaranm |
Archives