Re: pam_unix.so and unix_chkpw setgid - does it work for regular users?
Wolfgang Draxinger <[email protected]>
| Newsgroups | gmane.linux.pam |
|---|---|
| Organization | LMU Fakultät für Physik |
| Message-ID | <[email protected]> |
On Thu, 2 Aug 2012 12:47:36 -0700 Steve Langasek <[email protected]> wrote: > It's so that a regular user can *self* authenticate. Allowing users > to call this setgid helper directly for other accounts would let them > use it for brute forcing of passwords. So no, what you're asking for > is disallowed by design. Well, how about making brute forcing passwords hard by using a hash function designed for the use with passwords like bcrypt. That way all the user did was wasting his CPU cycles without getting close to anything. I do understand the motivation for preventing a user to use unix_chkpw for brute forcing. But what does prevent said user from using `ssh localhost` for this? Well, the increasing retry delay, maybe a disallow for localhost (erm...). But one could also add a usleep(100000) after a negative result, and as such slowing down a brute force significantly. Wolfgang _______________________________________________ Pam-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/pam-list
signature.asc
(application/pgp-signature, 198 B)
-----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.11 (GNU/Linux) iEYEARECAAYFAlAbBQ4ACgkQBfWmRR/TvT6+8wCfad+azoo2nqveBPFM7lyQ4el7 DAYAniZrirPuwQWuDbMrImHehm8S1aUV =TCe9 -----END PGP SIGNATURE-----