[MPlayer-dev-eng] Fwd: [MPlayer-users] Mplayer exploit?!
Arpi
arpi at thot.banki.hu
Tue Jan 14 15:05:30 CET 2003
Hi,
While i (and others) think the bellow quoted bugtraq message is fake/joke,
i have to note that it could be true too, at least it's doable technically.
It's quite easy to find media files causing crash in any of the players,
just look at mphq incoming :)
Even if mplayer's mpeg demuxer & decoder is very error-resistant, the other
parsers (like the real or avi demuxers) aren't. Easy to get buffer overflow...
The same stays for subtitle or configfile parsers too.
Esp. that many users run mplayer (and not only that, xine too) as root, to
get DGA or Vidix or SVGAlib or VESA working.
So we should take it as a warning, that I'm not the only with this idea...
Btw, few words about why is it fake:
- riaa (imho) has no the right to develop virii/worms
- most of these players are distributed in source form, compiled with
various compilers and options, hard to make exploit which work everywhere.
- also (at least for mplayer) many versions are in use, imho impossible to
make exploit to work with all (most) of them
- it's (imho) impossible to make perfect exploit, and even if it fails only
1% of cases, it's easy to notice and we'd already heard about it then
- infecting all p2p clients? how?
ps: riaa.org was hacked few days ago, their homepage was replaced by fake
news page... so i doubt they have so good security experts there :)
A'rpi
--------- Forwarded message ---------
Introduction:
Several months ago, GOBBLES Security was recruited by the RIAA (riaa.org)
to invent, create, and finally deploy the future of antipiracy tools. We
focused on creating virii/worm hybrids to infect and spread over p2p nets.
Until we became RIAA contracters, the best they could do was to passively
monitor traffic. Our contributions to the RIAA have given them the power
to actively control the majority of hosts using these networks.
We focused our research on vulnerabilities in audio and video players.
The idea was to come up with holes in various programs, so that we could
spread malicious media through the p2p networks, and gain access to the
host when the media was viewed.
During our research, we auditted and developed our hydra for the following
media tools:
mplayer (www.mplayerhq.org)
WinAMP (www.winamp.com)
Windows Media Player (www.microsoft.com)
xine (xine.sourceforge.net)
mpg123 (www.mpg123.de)
xmms (www.xmms.org)
After developing robust exploits for each, we presented this first part of
our research to the RIAA. They were pleased, and approved us to continue
to phase two of the project -- development of the mechanism by which the
infection will spread.
It took us about a month to develop the complex hydra, and another month to
bring it up to the standards of excellence that the RIAA demanded of us. In
the end, we submitted them what is perhaps the most sophisticated tool for
compromising millions of computers in moments.
Our system works by first infecting a single host. It then fingerprints a
connecting host on the p2p network via passive traffic analysis, and
determines what the best possible method of infection for that host would
be. Then, the proper search results are sent back to the "victim" (not the
hard-working artists who p2p technology rapes, and the RIAA protects). The
user will then (hopefully) download the infected media file off the RIAA
server, and later play it on their own machine.
When the player is exploited, a few things happen. First, all p2p-serving
software on the machine is infected, which will allow it to infect other
hosts on the p2p network. Next, all media on the machine is cataloged, and
the full list is sent back to the RIAA headquarters (through specially
crafted requests over the p2p networks), where it is added to their records
and stored until a later time, when it can be used as evidence in criminal
proceedings against those criminals who think it's OK to break the law.
Our software worked better than even we hoped, and current reports indicate
that nearly 95% of all p2p-participating hosts are now infected with the
software that we developed for the RIAA.
Things to keep in mind:
1) If you participate in illegal file-sharing networks, your
computer now belongs to the RIAA.
2) Your BlackIce Defender(tm) firewall will not help you.
3) Snort, RealSecure, Dragon, NFR, and all that other crap
cannot detect this attack, or this type of attack.
4) Don't fuck with the RIAA again, scriptkids.
5) We have our own private version of this hydra actively
infecting p2p users, and building one giant ddosnet.
Due to our NDA with the RIAA, we are unable to give out any other details
concerning the technology that we developed for them, or the details on any
of the bugs that are exploited in our hydra.
However, as a demonstration of how this system works, we're providing the
academic security community with a single example exploit, for a mpg123 bug
that was found independantly of our work for the RIAA, and is not covered
under our agreement with the establishment.
Affected Software:
mpg123 (pre0.59s)
http://www.mpg123.de
Problem Type:
Local && Remote
Vendor Notification Status:
The professional staff of GOBBLES Security believe that by releasing our
advisories without vendor notification of any sort is cute and humorous, so
this is also the first time the vendor has been made aware of this problem.
We hope that you're as amused with our maturity as we are. ;PpPppPpPpPPPpP
Exploit Available:
Yes, attached below.
Technical Description of Problem:
Read the source.
Credits:
Special thanks to stran9er at openwall.com for the ethnic-cleansing shellcode.
More information about the MPlayer-dev-eng
mailing list