By Jake Edge
September 2, 2009
A recent report
about a Skype trojan that could extract voice calls as mp3 files and
ship them off to other locations led to
an interesting discussion on the Fedora users mailing list. The trojan
itself is somewhat unsurprising as there have been persistent rumors about
wiretapping back doors in Skype for some time. The trojan is
Windows-only, but it does come with most of the source code, which makes it
interesting to those who study malware. While not a direct threat to Linux
users, it does highlight a number of privacy and security issues to ponder.
Skype is a popular voice over IP (VoIP) application that runs on Linux, Mac
OS X, and Windows. Part of its appeal is that there are many users of the
free (as in beer) software, so folks can make free phone calls to many of
their friends and family. But it is a closed source tool that resists
attempts to reverse-engineer its protocol, so there are no interoperable
free (as in freedom) equivalents.
Daniel B. Thurman brought up the
trojan and wondered if it was an example of the back doors or interception
facilities that governments
have long been rumored to be pushing for Skype. That set off a thread
in which "black helicopters" made a tongue-in-cheek appearance, but there
were also more serious postings. Marko Vojinovic asks about whether there are ongoing attempts
to reverse-engineer the Skype protocol:
I have a feeling that the majority of Linux users would switch to Ekiga or
something else open source, if only it gave them support to communicate with
skype peers on the other end. Linux folks (myself included) use skype mainly
because all their friends and other contacts also use it, and it is completely
impossible to convince them all to go the open source way. But if Ekiga would
support the protocol, it would eliminate the need to install or use skype
binary itself, while functionality would be preserved. Not to mention better
support for sound and video hardware etc.
There are a number of problems with that, as was pointed out, including
the likelihood that Skype would change the protocol to cripple
interoperability, much as instant messaging companies have done along the
way. Roberto Ragusa noted that there have
been people who looked at Skype, but they "found that it contains
tons and tons of cryptography, obfuscation and countermeasures against
debugging or reverse engineering." That is of concern he said
because one cannot be sure of exactly what it's doing: "A closed
source code like that and with an explicit purpose to build a crypted P2P
network bypassing firewalls with every trick possible is something to be
nervous about."
Alan Cox had some additional thoughts
on reverse-engineering the code: "The person who completely reverse
engineers skype probably destroys it. If you can write a skype client [then]
the spammers can write skype spam tools as well." He also mentions
the "mostly circumstantial" evidence that law enforcement has
added intercept facilities to Skype itself. Furthermore, anyone who might
be working on the problem has good reason to do it quietly, he said:
I would imagine anyone doing so is keeping fairly quiet - there is big
money on offer from the bad guys for skype trojans, intercepts and
clients, while anyone on the good side fiddling with it faces jail and
[harassment] - a fine example of perverse economic incentives.
So, we have a closed source application, which uses malware-like techniques
to obfuscate its functioning, and folks willingly run it on their
computers. In some ways, that's no different than any other closed source
application, but there are a few differences. Skype, by its very nature,
must use the network to send encrypted data to multiple untrusted machines
elsewhere. While it may not be compromised by governmental authorities in
the standard binary, it is a known target of those entities, and this
trojan demonstrates a way that it might be compromised. Overall, it would
seem there are a few risks to both security and privacy from that kind of
application—more so than a closed source word processor or
non-networked game.
Free software solutions, like Ekiga, may be
able to overcome some of the shortcomings of Skype. But, if those
solutions become popular, they are likely to run afoul of the spammers and
scammers that Cox warns about. It's likely to be true of regular and
cellular phone service as well, but a warning from "Tim" in the thread is worth
repeating:
Moral of the story; don't conduct illegal business over it, don't
conduct legal but confidential conversations over it; and if you're in
one of those places where criticising the government has nasty
repercussions, I wouldn't do that over it, either.
While Skype provides a nice service—without charge in many
cases—it does present a bit of a privacy headache. If it can be
subverted for wiretapping purposes, it can undoubtedly be subverted for
other reasons. Some of those could present security headaches as well.
Since we don't really know what the Skype code does when it isn't infected,
it will be
difficult to determine if its behavior changes in a malicious way. That
should be a little worrisome.
Comments (21 posted)
Brief items
A new site at
whattheinternetknowsaboutyou.com
is an interesting demonstration of CSS-related browser history disclosure
vulnerabilities. This site is able to produce a surprisingly comprehensive
list of sites that one has visited, down to the level of specific pages on
social networking sites and such. No JavaScript required. There's also
information on just how the
site works and how the disclosure of information can be minimized.
"
It is a source of amazement to us that such an obvious and
well-documented history sniffing channel has been allowed to exist for so
many years. We cannot help but wonder why, despite all the malicious
potential, such a hole has not yet been closed."
Comments (25 posted)
Ramon de Carvalho Valle has released an exploit for the Linux
sock_sendpage null pointer dereference vulnerability. The exploit was originally written to determine whether it was exploitable on the Power/Cell architecture, but was later expanded for i386 and x86_64. Many distribution kernels were tested using the exploit, and the results are included in the report to the bugtraq mailing list. The code may be of general interest, but also could be used on other kernels to determine if the problem has been addressed. Click below for the full report along with a link to the code.
Full Story (comments: 3)
The Apache project has
suffered a server compromise which took the site off the net for some hours. "
To the best of our knowledge at this time, no end users were affected by this incident, and the attackers were not able to escalate their privileges on any machines.
While we have no evidence that downloads were affected, users are always advised to check digital signatures where provided."
Comments (17 posted)
New vulnerabilities
dnsmasq: heap overflow, NULL pointer dereference
| Package(s): | dnsmasq |
CVE #(s): | CVE-2009-2957
CVE-2009-2958
|
| Created: | September 1, 2009 |
Updated: | October 14, 2009 |
| Description: |
From the Red Hat advisory:
Core Security Technologies discovered a heap overflow flaw in dnsmasq when
the TFTP service is enabled (the "--enable-tftp" command line option, or by
enabling "enable-tftp" in "/etc/dnsmasq.conf"). If the configured tftp-root
is sufficiently long, and a remote user sends a request that sends a long
file name, dnsmasq could crash or, possibly, execute arbitrary code with
the privileges of the dnsmasq service (usually the unprivileged "nobody"
user). (CVE-2009-2957)
A NULL pointer dereference flaw was discovered in dnsmasq when the TFTP
service is enabled. This flaw could allow a malicious TFTP client to crash
the dnsmasq service. (CVE-2009-2958)
|
| Alerts: |
|
Comments (none posted)
gfs2-utils: temporary file vulnerabilities
| Package(s): | gfs2-utils |
CVE #(s): | CVE-2008-6552
|
| Created: | September 2, 2009 |
Updated: | February 16, 2011 |
| Description: |
The gfs2-utils package suffers from multiple temporary file vulnerabilities which could be exploited by a local hacker to overwrite arbitrary files. |
| Alerts: |
|
Comments (none posted)
htmldoc: stack-based buffer overflow
| Package(s): | htmldoc |
CVE #(s): | |
| Created: | September 1, 2009 |
Updated: | September 2, 2009 |
| Description: |
From the Red
Hat bugzilla:
A stack-based buffer overflow by processing user-supplied input was found
in HTMLDOC's routine, used to set the result page output size for custom
page sizes. A remote attacker could provide a specially-crafted HTML file,
which once opened by an unsuspecting user, would lead to denial of service
(htmldoc crash). |
| Alerts: |
|
Comments (none posted)
ikiwiki: information disclosure
| Package(s): | ikiwiki |
CVE #(s): | CVE-2009-2944
|
| Created: | September 1, 2009 |
Updated: | April 1, 2010 |
| Description: |
From the Debian advisory:
Josh Triplett discovered that the blacklist for potentially harmful TeX
code of the teximg module of the Ikiwiki wiki compiler was incomplete,
resulting in information disclosure.
|
| Alerts: |
|
Comments (none posted)
kernel: denial of service
| Package(s): | kernel |
CVE #(s): | CVE-2009-2691
|
| Created: | August 27, 2009 |
Updated: | March 23, 2010 |
| Description: |
From the National Vulnerability Database
entry:
"The mm_for_maps function in fs/proc/base.c in the Linux kernel 2.6.30.4 and earlier allows local users to read (1) maps and (2) smaps files under proc/ via vectors related to ELF loading, a setuid process, and a race condition." |
| Alerts: |
|
Comments (none posted)
kernel: denial of service
| Package(s): | kernel |
CVE #(s): | CVE-2009-2767
|
| Created: | August 27, 2009 |
Updated: | October 22, 2009 |
| Description: |
From the National Vulnerability Database
entry:
"The init_posix_timers function in kernel/posix-timers.c in the Linux kernel before 2.6.31-rc6 allows local users to cause a denial of service (OOPS) or possibly gain privileges via a CLOCK_MONOTONIC_RAW clock_nanosleep call that triggers a NULL pointer dereference." |
| Alerts: |
|
Comments (none posted)
libmikmod: two denial of service vulnerabilities
| Package(s): | libmikmod |
CVE #(s): | CVE-2007-6720
CVE-2009-0179
|
| Created: | August 31, 2009 |
Updated: | October 11, 2010 |
| Description: |
From the Red Hat bugzilla entries [1 and 2]:
CVE-2009-0179: A denial of service flaw was found in the MikMod player, used for playing
MOD files. If an attacker would trick the mikmod user to load an XM file,
this could lead to denial of service (application crash).
CVE-2007-6720: A denial of service flaw was found in the MikMod player, used for playing
MOD files. If an attacker would trick the mikmod user to play multiple
MOD using files with varying number of channels, this could lead to
denial of service (application crash or abort).
|
| Alerts: |
|
Comments (none posted)
mono: cross-site scripting vulnerabilities
| Package(s): | mono |
CVE #(s): | CVE-2008-3422
|
| Created: | August 27, 2009 |
Updated: | December 7, 2009 |
| Description: |
From the National Vulnerability Database
entry:
"Multiple cross-site scripting (XSS) vulnerabilities in the ASP.net class libraries in Mono 2.0 and earlier allow remote attackers to inject arbitrary web script or HTML via crafted attributes related to (1) HtmlControl.cs (PreProcessRelativeReference), (2) HtmlForm.cs (RenderAttributes), (3) HtmlInputButton (RenderAttributes), (4) HtmlInputRadioButton (RenderAttributes), and (5) HtmlSelect (RenderChildren). " |
| Alerts: |
|
Comments (none posted)
openssh: information disclosure
| Package(s): | openssh |
CVE #(s): | CVE-2008-5161
|
| Created: | September 2, 2009 |
Updated: | March 8, 2010 |
| Description: |
Openssh is vulnerable to a specific man-in-the-middle attack which could be able to obtain a piece of plain text when the CBC cipher mode is used. |
| Alerts: |
|
Comments (none posted)
squirrelmail: cross-site request forgery
| Package(s): | squirrelmail |
CVE #(s): | CVE-2009-2964
|
| Created: | August 31, 2009 |
Updated: | August 13, 2010 |
| Description: |
From the Mandriva advisory:
All form submissions (send message, change preferences, etc.) in
SquirrelMail were previously subject to cross-site request forgery
(CSRF), wherein data could be sent to them from an offsite location,
which could allow an attacker to inject malicious content into
user preferences or possibly send emails without user consent
(CVE-2009-2964).
|
| Alerts: |
|
Comments (none posted)
wordpress: open redirect vulnerability
| Package(s): | wordpress |
CVE #(s): | CVE-2008-6762
|
| Created: | August 27, 2009 |
Updated: | September 2, 2009 |
| Description: |
From the National Vulnerability Database
entry:
"Open redirect vulnerability in wp-admin/upgrade.php in WordPress, probably 2.6.x, allows remote attackers to redirect users to arbitrary web sites and conduct phishing attacks via a URL in the backto parameter." |
| Alerts: |
|
Comments (none posted)
wordpress: denial of service
| Package(s): | wordpress |
CVE #(s): | CVE-2008-6767
|
| Created: | August 27, 2009 |
Updated: | September 2, 2009 |
| Description: |
From the National Vulnerability Database
entry:
"wp-admin/upgrade.php in WordPress, probably 2.6.x, allows remote attackers to upgrade the application, and possibly cause a denial of service (application outage), via a direct request." |
| Alerts: |
|
Comments (none posted)
wordpress: password vulnerability
| Package(s): | wordpress |
CVE #(s): | CVE-2008-4106
|
| Created: | August 27, 2009 |
Updated: | September 2, 2009 |
| Description: |
From the National Vulnerability Database
entry:
"WordPress before 2.6.2 does not properly handle MySQL warnings about insertion of username strings that exceed the maximum column width of the user_login column, and does not properly handle space characters when comparing usernames, which allows remote attackers to change an arbitrary user's password to a random value by registering a similar username and then requesting a password reset, related to a "SQL column truncation vulnerability."" |
| Alerts: |
|
Comments (none posted)
wordpress: directory traversal vulnerability
| Package(s): | wordpress |
CVE #(s): | CVE-2008-4769
|
| Created: | August 27, 2009 |
Updated: | September 2, 2009 |
| Description: |
From the National Vulnerability Database
entry:
"Directory traversal vulnerability in the get_category_template function in wp-includes/theme.php in WordPress 2.3.3 and earlier, and 2.5, allows remote attackers to include and possibly execute arbitrary PHP files via the cat parameter in index.php. NOTE: some of these details are obtained from third party information." |
| Alerts: |
|
Comments (none posted)
wordpress: cross-site request forgery vulnerability
| Package(s): | wordpress |
CVE #(s): | CVE-2008-5113
|
| Created: | August 27, 2009 |
Updated: | September 2, 2009 |
| Description: |
From the National Vulnerability Database
entry:
"WordPress 2.6.3 relies on the REQUEST superglobal array in certain dangerous situations, which makes it easier for remote attackers to conduct delayed and persistent cross-site request forgery (CSRF) attacks via crafted cookies, as demonstrated by attacks that (1) delete user accounts or (2) cause a denial of service (loss of application access). NOTE: this issue relies on the presence of an independent vulnerability that allows cookie injection." |
| Alerts: |
|
Comments (none posted)
Page editor: Jake Edge
Next page: Kernel development>>