| 1 |
On Tuesday 10 February 2004 16:36, Daniel Brandt wrote: |
| 2 |
> > Hmm, let's say the attacker gains access to the machine, the firewall |
| 3 |
> > blocks all binary transfer (I know uuencode/decode, but lets think the |
| 4 |
> > attacker is not in the position to transfer executables onto the |
| 5 |
> > compromised system, perhaps he can't transfer any files) and the |
| 6 |
> > attacker only needs 10 lines of c-code to exploit the kernel or |
| 7 |
> > whatever - don't worry about if he can compile the 10 lines or not? |
| 8 |
> > Perhaps also the system runs on alpha hardware but the attacker only |
| 9 |
> > has x86 binaries etc.. |
| 10 |
> |
| 11 |
> Too hypothetic, transfer of a binary file will be possible if there is |
| 12 |
> two-way communication. |
| 13 |
Not really, e.g. a lot of dedicated systems' serial consoles are connected |
| 14 |
to a central server to provide a login possibilty if login via network is |
| 15 |
not possible. Using such a console does not provide the possibility to |
| 16 |
transfer binary data, I think. |
| 17 |
|
| 18 |
> Non x86 hardware is not a problem because of cross compiling. |
| 19 |
ACK (for Linux - other Unixes may still run on hardware the attacker has no |
| 20 |
crosscompiler for or the ABI is different). |
| 21 |
|
| 22 |
Regards |
| 23 |
Daniel |
| 24 |
|
| 25 |
-- |
| 26 |
"Those who would give up essential liberty, to purchase a little temporary |
| 27 |
safety, deserve neither liberty nor safety." - Benjamin Franklin |
| 28 |
|
| 29 |
|
| 30 |
-- |
| 31 |
gentoo-security@g.o mailing list |
Archives