bugtraq - Re: Calife heap corrupt / potential local root exploit

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date: 27 Feb 2004 16:49:34 -0000
From: Ollivier Robert <roberto@...tia.freenix.fr>
To: bugtraq@...urityfocus.com
Subject: Re: Calife heap corrupt / potential local root exploit

In-Reply-To: <20040227091921.26210.qmail@....securityfocus.com>

>Calife heap corrupt / potential local root exploit
>--------------------------------------------------
>by Leon Juranic a.k.a DownBload <downbload@...mail.com> / II-Labs
>
>
>Version affected(tested): calife-2.8.4c and calife-2.8.5
>- calife can be found at packages.debian.org, FreeBSD 5.0 (security), ...

Thanks you for taking the time to contact me before sending such a mail to Bugtraq.  It is always nice to deal with such nice people [NOT!]

>[downbload@...alhost downbload]$ calife luser
>Password: "A" x 3000
>Password: real_user_password
>Segmentation fault
>[downbload@...alhost downbload]$

Interesting, on which plateform?  I just tried that on FreeBSD 4.9, 5.2 and could not reproduce.

On Linux/Debian, it does segfault. glibc problem?

>- "A" x 3000 will corrupt the heap.
>- If real_user_password isn't correct, calife will do exit()
>- If attacker wants to exploit calife, there must be at least one user "available" in /etc/calife.auth

Do you have such an exploit?  I'd like to see it.

>            pt_pass = (char *) getpass ("Password:");
>            memset (user_pass, '\0', l_size);
>            strcpy (user_pass, pt_pass); // <- BAD CODE

I could have used strlcpy but I assumed (and my reading of the FreeBSD source code confirm it) that getpass(3) was doing the size check.

In FreeBSD, it seems not possible to overflow that as the code verify the length.

I'll release 2.8.6 today.

Courtesy seems to go down the gutters these days apparently.

Ollivier, pissed off.