maybe eEye will publish more details because as usual NGSS withholds the details of the hole :p, anyway might be good to start looking at this one nolimit ;) I will try under free time.



Mark Litchfield of NGSSoftware has discovered a high risk vulnerability in the
Veritas Backup Exec Server for Windows which can allow remote code execution.

Affected Products include -
Backup Exec 10.0 for Windows Servers rev. 5484
Backup Exec 9.1 for Windows Servers rev. 4691
Backup Exec 9.0 for Windows Servers rev. 4454
Backup Exec 9.0 for Windows Servers rev. 4367

A Patch can be obtained from - http://seer.support.veritas.com/docs/276607.htm
Please note that the documentation provided by this link is partially incorrect
and I believe it will be updated very shortly.

** Email Extract

In 9.x, the Admin Plus Pack has to be installed in order for the
vulnerability to exist. You are correct in 10.x it is vulnerable whether
the Admin Pack is installed or not.

So the workaround to disable the Admin Pack Option is only valid for a 9.x
installation. On 10.0, applying the patch is required

** End Email Extract

NGSSoftware are going to withhold details of this flaw for three months.
Full details will be published on the 23rd of September 2005. This three month
window will allow users of Veritas Backup Exec Server the time needed to apply the patch
before the details are released to the general public. This reflects
NGSSoftware's approach to responsible disclosure.

NGSSoftware Insight Security Research
http://www.ngssoftware.com
http://www.databasesecurity.com/
+44(0)208 401 0070

Indeed, so is v9.x and some v8. even , enterprise + ... seems port 6101 is the port attacked, so for a quickarse workaround (Im looking at a scanner-expoloit here..so its in use scanning), the fastest way to keep safe is block the port i guess.. then i dont really use that product to see why it needs so much remote, but it is definately being rapz0red ;o so yes, update.

ha no this is iDEFENSE this time rlsing much infos, hot :>



Veritas Backup Exec Agent CONNECT_CLIENT_AUTH Buffer Overflow
Vulnerability

iDEFENSE Security Advisory 06.23.05
www.idefense.com/application/poi/displa ... rabilities (http://www.idefense.com/application/poi/display?id=272&type=vulnerabilities)
June 23, 2005

I. BACKGROUND

VERITAS Backup Exec is an advanced backup and restore solution for
Microsoft Windows server environments. More information is available
from the vendor's website:

http://veritas.com/Products/www?c=product&refId=57

II. DESCRIPTION

Exploitation of a buffer overflow vulnerability in Veritas Software
Corp.'s Backup Exec allows remote attackers to execute arbitrary code.

The Veritas Backup Exec Agent listens on TCP port 10000 and is
responsible for accepting connections from the backup server when a
backup is to occur. Typically the agent would be installed on most
servers and important workstations in an enterprise environment.

Veritas Backup Exec uses the standard NMDP protocol to communicate with
the listening agents. The NMDP protocol allows multiple authentication
types, including support for Windows user credentials.

The vulnerability specifically exists because of insufficient input
validation on CONNECT_CLIENT_AUTH requests. CONNECT_CLIENT_AUTH requests
sent with an authentication method type "3," indicating Windows user
credentials, and an overly long password argument can overflow the
buffer and lead to arbitrary code execution. The overflow will copy the
user-supplied password to the stack until it attempts to run off the
page of memory. At this point, the SEH frame will be overwritten and the
program will jump to a user-defined location when the page fault occurs.
Debugger output showing control of execution from the SEH frame is
detailed as follows:

(1d8.b1c): Access violation - code c0000005 (first chance)
First chance exceptions are reported before any exception handling.
This exception may be expected and handled.

eax=00002001 ebx=0032ad78 ecx=0000037e
edx=00fbedf8 esi=01045928 edi=00fc0000
eip=0141b77f esp=00fbedd8 ebp=0032c040 iopl=0
nv up ei pl nz na po nc
cs=001b ss=0023 ds=0023 es=0023 fs=003b gs=0000 efl=00010206

*** ERROR: Symbol file could not be found. Defaulted to export symbols
for C:\Program Files\VERITAS\Backup Exec\NT\beclass.dll -
beclass!DeasciifyAndDecrypt+0xaf:

0141b77f f3a5 rep movsd ds:01045928=39784638 es:00fc0000=????????
0:005> g

(1d8.b1c): Access violation - code c0000005 (first chance)
First chance exceptions are reported before any exception handling.
This exception may be expected and handled.

eax=00000000 ebx=00000000 ecx=41424344
edx=7c9037d8 esi=00000000 edi=00000000
eip=41424344 esp=00fbea08 ebp=00fbea28 iopl=0
nv up ei pl zr na po nc
cs=001b ss=0023 ds=0023 es=0023 fs=003b gs=0000 efl=00010246
41424344 ?? ???
0:005> kp
ChildEBP RetAddr
WARNING: Frame IP not in any known module. Following frames may be
wrong.
00fbea04 7c9037bf 0x41424344
00fbea28 7c90378b ntdll!RtlConvertUlongToLargeInteger+0x7a
00fbead8 7c90eafa ntdll!RtlConvertUlongToLargeInteger+0x46
0032c040 0032b470 ntdll!KiUserExceptionDispatcher+0xe
00000003 00000000 0x32b470

III. ANALYSIS

Exploitation does not require authentication, thereby allowing any
remote attacker to execute arbitrary code under the privileges of the
Backup Exec Agent Browser (benetns.exe) process, which is usually a
domain administrative account. Exploitation can occur fairly reliably
since the overflow is able to control code execution via the structured
exception handler.

IV. DETECTION

iDEFENSE has confirmed the existence of this vulnerability in Backup
Exec for Windows Servers versions 9.1 and 10.0. It is suspected
that other versions are also vulnerable.

Several vendors ship or use Veritas software including Hitachi and NEC
Corporation.

V. WORKAROUND

Use a firewall to restrict incoming connections to trusted workstations
running the Backup Exec client software which uses port 10000 TCP.

VI. VENDOR RESPONSE

The vendor has addressed this vulnerability in security advisory
VX05-002:

http://seer.support.veritas.com/docs/276604.htm

VII. CVE INFORMATION

The Common Vulnerabilities and Exposures (CVE) project has assigned the
name CAN-2005-0773 to this issue. This is a candidate for inclusion
in the CVE list (http://cve.mitre.org), which standardizes names for
security problems.

VIII. DISCLOSURE TIMELINE

03/16/2005 Initial vendor notification
03/30/2005 Initial vendor response
06/22/2005 Coordinated public disclosure

IX. CREDIT

The discoverer of this vulnerability wishes to remain anonymous.

Get paid for vulnerability research
http://www.idefense.com/poi/teams/vcp.jsp

Free tools, research and upcoming events
http://labs.idefense.com

X. LEGAL NOTICES

Copyright (c) 2005 iDEFENSE, Inc.

Permission is granted for the redistribution of this alert
electronically. It may not be edited in any way without the express
written consent of iDEFENSE. If you wish to reprint the whole or any
part of this alert in any other medium other than electronically, please
email customerservice@idefense.com for permission.

Disclaimer: The information in the advisory is believed to be accurate
at the time of publishing based on currently available information. Use
of the information constitutes acceptance for use in an AS IS condition.
There are no warranties with regard to this information. Neither the
author nor the publisher accepts any liability for any direct, indirect,
or consequential loss or damage arising from use of, or reliance on,
this information.

hopefully this isn't the actual hole that I have in a veritas product, but quite similiar ;>

haha yeah I saw that. I love how idefense is so descriptive in their advisories, makes research that much easier. When I get a chance I might make a POC for it, but I'm kinda busy this weekand and I'd have to look over the whole NDMP protocol to form the proper request. That is unless I just knab it with a packetsniffer. Anywho, I'll update later when I do more research on it :)

to note that NDMP isnt a unique agent of BackupExec, there is prolly much more 'tas software vuln to this or similiar :>

btw kiddies are using http://www.metasploit.com/projects/Fram ... c_agent.pm (http://www.metasploit.com/projects/Framework/modules/exploits/backupexec_agent.pm) to exploit it :> will see if I can code a scanner of it under free time.

this is possible to retrieve many infos via the NDMP protocol, you can have datas such:

NDMP CONFIG_GET_HOST_INFO

00000048 00 00 00 0a 47 56 2d 58 4f 45 52 44 45 58 00 00 ....GV-X OERDEX..
00000058 00 00 00 0a 57 69 6e 64 6f 77 73 20 4e 54 00 00 ....Wind ows NT..
00000068 00 00 00 8d 4d 61 6a 6f 72 20 56 65 72 73 69 6f ....Majo r Versio
00000078 6e 3d 35 20 4d 69 6e 6f 72 20 56 65 72 73 69 6f n=5 Mino r Versio
00000088 6e 3d 30 20 42 75 69 6c 64 20 4e 75 6d 62 65 72 n=0 Buil d Number
00000098 3d 32 31 39 35 20 53 65 72 76 69 63 65 50 61 63 =2195 Se rvicePac
000000A8 6b 20 4d 61 6a 6f 72 3d 34 20 53 65 72 76 69 63 k Major= 4 Servic
000000B8 65 50 61 63 6b 20 4d 69 6e 6f 72 3d 30 20 53 75 ePack Mi nor=0 Su
000000C8 69 74 65 4d 61 73 6b 3d 32 37 34 20 50 72 6f 64 iteMask= 274 Prod
000000D8 75 63 74 54 79 70 65 3d 32 20 50 72 6f 63 65 73 uctType= 2 Proces
000000E8 73 6f 72 54 79 70 65 3d 49 6e 74 65 6c 20 78 38 sorType= Intel x8
000000F8 36 00 00 00 00 00 00 0d 58 58 58 2e 58 58 2e 58 6....... XXX.XX.X
00000108 58 58 2e 58 58 00 00 00 XX.XX...


look at NDMP Specs. (http://www.ndmp.org) for much informations, this is prolly possible to guess via this way if the server is vulnerable and why not to find much vulnerabilities linked to some other requests :)

this source took from ethereal0.10.11 helps so to understand how is working NDMP protocol (all various packets request/reply over TCP/IP)and so on this vulnerability exploitation.

have compared tons of requests with LOG, CONNECT, CONFIG, between bexec sp0-sp1 vuln/patched, havent caught something helping to determine the vulnerability, if anyone see let me know.

nothing valuable found with other interfaces (even if I didnt try to reset scsi bus, lol). Anyway on a xp sp2 en not patched the reply from backup exec agent port is ...$....B..~............................

lol , would be better in hexadecimal , and anyway in this first reply there is nothing interesting to grab, you can make some particular request to guss some os infos , etc

class have you tried looking for non RFC standard requests in veritas?

lo nolimit, I have tested nothing yet in veritas else I had to find the mark's hole (lol the big flamer :o)

btw kiddies are using http://www.metasploit.com/projects/Fram ... c_agent.pm (http://www.metasploit.com/projects/Framework/modules/exploits/backupexec_agent.pm) to exploit it :> will see if I can code a scanner of it under free time.



eheh class, it will be also for kiddies but it works very well :)
I tested it on some vuln ips (with banner $...B..) ... i don't say one shell for every ip but nearly.....Only one thing I don't understand, it works only on .us, .or and .au ranges....in eu it seems to have difficulty....what could it be?

btw you can get [url=ftp://ftp.support.veritas.com/pub/support/products/Backup_Exec_for_WindowsNT/Q175168.BEWS10.0.5520.ESD_277181.zip:deca6]here[/url:deca6] the new veritas 10.0 build 5520 for win servers patched (the #5520 for netware still need hotfix) because if you look in the download section , you can get at this time the highly critical build (5484) and you'll have to install several hotfix then. Better is to grab the build #5520 here: http://seer.support.veritas.com/docs/277181.htm

eheh class, it will be also for kiddies but it works very well :)
I tested it on some vuln ips (with banner $...B..) ... i don't say one shell for every ip but nearly.....Only one thing I don't understand, it works only on .us, .or and .au ranges....in eu it seems to have difficulty....what could it be?


Shell hacking is normally done to contribute to the warez community. As most of the real scene apparently is in europe, it's only logical that european servers would be the ones to be targetted first.
Abstract: .eu servers get hit first, and get patched first.