| 1 |
On Thursday 14 October 2004 08:15 pm, Dan Margolis wrote: |
| 2 |
>Following on your previous e-mail to the documentation editors, I would |
| 3 |
>appreciate it if you'd file a bug at |
| 4 |
>http://bugs.gentoo.org/enter_bug.cgi?product=Docs-user&format=guided. |
| 5 |
|
| 6 |
Done: http://bugs.gentoo.org/show_bug.cgi?id=67701 |
| 7 |
|
| 8 |
>As for this being widely known, it is not something I ever considered, |
| 9 |
>but it also is not, in and of itself, an exploit. |
| 10 |
|
| 11 |
True - It's not an exploit - But it's definitely a well-known weakness. |
| 12 |
|
| 13 |
>Presumably the attacker first used some other vulnerability to gain access to |
| 14 |
>your server. |
| 15 |
|
| 16 |
A valid user was able to install/execute a root-kit from /dev/shm. |
| 17 |
|
| 18 |
>Further, if I were to set up a server where I did not want to allow |
| 19 |
>users to execute anything but my own chosen binaries, I would probably |
| 20 |
>use GRSecurity's Trusted Path Execution, which allows a whitelist |
| 21 |
>approach (i.e. allow only root owned binaries in /bin or /usr/local/bin) |
| 22 |
>rather than trying to make sure our partitions are all mounted correctly. |
| 23 |
|
| 24 |
Had I known how easily Linux's security could be circumvented, I surely would |
| 25 |
have invested more time hardening the system. :( |
| 26 |
|
| 27 |
>That said, this is certainly a lapse in the Security Guide, and if you |
| 28 |
>file a bug I am sure it will be corrected shortly (if by nobody else, |
| 29 |
>than, soon as I have the time, I will write an update). |
| 30 |
|
| 31 |
Thanks, Dan! |
| 32 |
|
| 33 |
-- |
| 34 |
Eric P. |
| 35 |
|
| 36 |
-- |
| 37 |
gentoo-hardened@g.o mailing list |
Archives