By Jake Edge
November 3, 2010
The recent release of Firesheep—a Firefox
extension that captures others' cookies on open WiFi networks—has
set off something of a firestorm. The particular hole that Firesheep
exploits is not anything new, we looked at an EFF-sponsored workaround for the
problem back in July, but the particulars of the Firesheep implementation
are fairly eye-opening. It would seem that Firesheep developer Eric Butler
was wildly successful in doing what he set out to do: increase
the visibility of insecure session cookie handling by major web sites.
It is fairly standard for web sites to protect their login screens by using
HTTPS (i.e. SSL/TLS encrypted connections) so that usernames and passwords
cannot be intercepted. But once the login has been completed, a session is
created, and sites
typically hand out a cookie—a (hopefully) opaque value that the
server can use to associate a request with a particular session
(i.e. user). Each time the user's browser sends a request to the site, it
also sends any cookies that have been set by that site. Those cookies are
valid for a server-selectable period of
time, and while they are valid, they can be used by anyone to appear to the
server as the user who logged in. The problem is that the cookies are
often transmitted via unencrypted HTTP.
So Firesheep, which was released
at ToorcCon 12 on October 24, can intercept these cookie values for
various high-profile web sites (e.g. Facebook, Twitter, Amazon, Google,
Github, and so on). It does the cookie interception by sniffing the network
traffic on open WiFi networks, and once it has them, it offers the user the
ability to connect to those services using the captured cookies. So someone
sitting in a coffeeshop can run Firesheep and potentially access
Facebook as some other unsuspecting customer.
The ability to do a one-click takeover of someone's account is clearly
Firesheep's most controversial feature. But it certainly serves the
purpose of alerting the public to this particular problem. Packaging the
program as a Firefox extension is also a clever touch. There is no reason
that Firesheep couldn't be a standalone program, but making it available in
the browser eases the installation process so that it can get in the hands
of more (ab)users.
Butler's intent is to shame (or scare) web site operators into switching to
HTTPS. It is the same end goal that the EFF had with its HTTPS Everywhere Firefox
extension, but Firesheep definitely grabbed a lot more attention than the
EFF's tool did. HTTPS Everywhere uses rules to rewrite http://
URLs to https:// URLs, which is useful—but not
particularly striking, at least to casual users and the press.
People have expressed ethical concerns about the release of Firesheep, but
like many security-oriented tools, it can be used for good or ill. There
are also reports that Microsoft's anti-virus software is marking Firesheep
as a threat. This firestorm has caused Butler to strongly
defend Firesheep and its release:
In addition to questioning Firesheep's legality, some people have
questioned the ethics of its release. Similar tools have existed for years,
so big companies, especially Facebook and Twitter, cannot claim they are
unaware of these issues. They have knowingly placed user privacy on the
back burner, and I'd be interested to hear some discussion about the ethics
of these decisions, which have left users at risk since long before
Firesheep.
Web sites can fix the problem by converting over to HTTPS and marking their
session cookies as HTTPS-only, but it's not quite as simple as just
flipping a switch. HTTPS will definitely require more server resources to
encrypt and decrypt all of its traffic, but there are other potential
problem areas as well. Various internal links in existing content may need
to be converted or
handled by the web server rewrite engine, and there is a class of content
that web site operators may not have any control over: advertisements.
Ad networks run by Google and others often do
not offer HTTPS for serving ads. That results in a warning from many
web browsers because there is insecure (i.e. HTTP) content in an HTTPS
page. The last thing many web site operators want is for their new users
to be greeted with a scary warning about the site.
We have been running some experiments here at LWN and plan to have
HTTPS-only cookies soon, though we haven't quite figured out how to handle
the Google ad problem. It is really something we (and lots of other sites)
should have done a long time ago. Thanks to Firesheep, there are now even more
compelling reasons to make that switch happen.
Comments (72 posted)
New vulnerabilities
clamav: code execution
| Package(s): | clamav |
CVE #(s): | CVE-2010-3434
|
| Created: | October 29, 2010 |
Updated: | November 3, 2010 |
| Description: |
From the CVE entry:
Buffer overflow in the find_stream_bounds function in pdf.c in libclamav in ClamAV before 0.96.3 allows remote attackers to cause a denial of service (application crash) or possibly execute arbitrary code via a crafted PDF document. NOTE: some of these details are obtained from third party information. |
| Alerts: |
|
Comments (none posted)
cups: code execution
| Package(s): | cups |
CVE #(s): | CVE-2010-2941
|
| Created: | October 29, 2010 |
Updated: | March 2, 2011 |
| Description: |
From the Red Hat advisory:
A use-after-free flaw was found in the way the CUPS server parsed Internet
Printing Protocol (IPP) packets. A malicious user able to send IPP requests
to the CUPS server could use this flaw to crash the CUPS server or,
potentially, execute arbitrary code with the privileges of the CUPS server.
|
| Alerts: |
|
Comments (none posted)
cvs: code execution
| Package(s): | cvs |
CVE #(s): | CVE-2010-3846
|
| Created: | October 29, 2010 |
Updated: | November 30, 2010 |
| Description: |
From the Red Hat bugzilla:
An array index error, leading to heap-based buffer overflow was found
in the way CVS version control system applied certain delta fragments
changes from input file in the RCS (Revision Control System file)
format. A local attacker could store a specially-crafted RCS file into
the CVS repository and trick the remote victim to checkout (update their
CVS repository tree) with this file, which could lead to arbitrary code
execution with the privileges of the user running cvs client executable.
|
| Alerts: |
|
Comments (none posted)
dovecot: restriction bypass
| Package(s): | dovecot |
CVE #(s): | CVE-2010-3706
CVE-2010-3707
|
| Created: | October 29, 2010 |
Updated: | May 19, 2011 |
| Description: |
From the CVE entries:
plugins/acl/acl-backend-vfile.c in Dovecot 1.2.x before 1.2.15 and 2.0.x before 2.0.5 interprets an ACL entry as a directive to add to the permissions granted by another ACL entry, instead of a directive to replace the permissions granted by another ACL entry, in certain circumstances involving the private namespace of a user, which allows remote authenticated users to bypass intended access restrictions via a request to read or modify a mailbox. (CVE-2010-3706)
plugins/acl/acl-backend-vfile.c in Dovecot 1.2.x before 1.2.15 and 2.0.x before 2.0.5 interprets an ACL entry as a directive to add to the permissions granted by another ACL entry, instead of a directive to replace the permissions granted by another ACL entry, in certain circumstances involving more specific entries that occur after less specific entries, which allows remote authenticated users to bypass intended access restrictions via a request to read or modify a mailbox. (CVE-2010-3707) |
| Alerts: |
|
Comments (none posted)
dovecot: multiple vulnerabilities
| Package(s): | dovecot |
CVE #(s): | CVE-2010-3779
CVE-2010-3780
|
| Created: | November 1, 2010 |
Updated: | May 19, 2011 |
| Description: |
From the Mandriva advisory:
Dovecot 1.2.x before 1.2.15 and 2.0.x before 2.0.beta2 grants the admin
permission to the owner of each mailbox in a non-public namespace,
which might allow remote authenticated users to bypass intended access
restrictions by changing the ACL of a mailbox, as demonstrated by a
symlinked shared mailbox (CVE-2010-3779).
Dovecot 1.2.x before 1.2.15 allows remote authenticated users to
cause a denial of service (master process outage) by simultaneously
disconnecting many (1) IMAP or (2) POP3 sessions (CVE-2010-3780).
|
| Alerts: |
|
Comments (none posted)
gnucash: arbitrary code execution
| Package(s): | gnucash |
CVE #(s): | CVE-2010-3999
|
| Created: | November 1, 2010 |
Updated: | November 25, 2010 |
| Description: |
From the Red Hat bugzilla:
Ludwig Nussel discovered that gnucash contained a script that could be abused
by an attacker to execute arbitrary code.
The vulnerability is due to an insecure change to LD_LIBRARY_PATH, and
environment variable used by ld.so(8) to look for libraries in directories
other than the standard paths. When there is an empty item in the
colon-separated list of directories in LD_LIBRARY_PATH, ld.so(8) treats it as a
'.' (current working directory). If the given script is executed from a
directory where a local attacker could write files, there is a chance for
exploitation.
|
| Alerts: |
|
Comments (none posted)
libguestfs: possible host corruption
| Package(s): | libguestfs |
CVE #(s): | CVE-2010-3851
|
| Created: | November 3, 2010 |
Updated: | July 7, 2011 |
| Description: |
From the Red Hat bugzilla:
libguestfs doesn't currently allow the format of a disk to be specified
explicitly, and therefore always uses automatic format detection. It takes disk images as arguments, and can therefore only be run by the virtualisation administrator. However, if a malicious guest administrator knows that libguestfs will run against their image, they could still use this technique to corrupt the host.
|
| Alerts: |
|
Comments (1 posted)
luci: authentication bypass
| Package(s): | luci |
CVE #(s): | CVE-2010-3852
|
| Created: | November 3, 2010 |
Updated: | November 5, 2010 |
| Description: |
From the Red Hat bugzilla:
A security flaw was found in the way Luci administration application
processed ticket cookies. A remote attacker, with certain knowledge
of running Luci instance environment details could use this flaw to
bypass standard Luci authentication mechanism (access resources which
should be otherwise protected by authentication). |
| Alerts: |
|
Comments (none posted)
Mozilla products: remote code execution
| Package(s): | firefox seamonkey thunderbird xulrunner |
CVE #(s): | CVE-2010-3765
|
| Created: | October 28, 2010 |
Updated: | November 17, 2010 |
| Description: |
A race condition in Mozilla's document object model handling can be exploited (and is being exploited) to execute arbitrary code. The Firefox 3.6.12/3.5.15, Thunderbird 3.1.6/3.0.10, and Seamonkey 2.0.10 releases fix the problem. |
| Alerts: |
|
Comments (none posted)
pam: privilege escalation
| Package(s): | pam |
CVE #(s): | CVE-2010-3316
CVE-2010-3435
CVE-2010-3853
|
| Created: | November 2, 2010 |
Updated: | November 3, 2011 |
| Description: |
From the Red Hat advisory:
It was discovered that the pam_namespace module executed the external
script namespace.init with an unchanged environment inherited from an
application calling PAM. In cases where such an environment was untrusted
(for example, when pam_namespace was configured for setuid applications
such as su or sudo), a local, unprivileged user could possibly use this
flaw to escalate their privileges. (CVE-2010-3853)
It was discovered that the pam_mail module used root privileges while
accessing users' files. In certain configurations, a local, unprivileged
user could use this flaw to obtain limited information about files or
directories that they do not have access to. (CVE-2010-3435)
It was discovered that the pam_xauth module did not verify the return
values of the setuid() and setgid() system calls. A local, unprivileged
user could use this flaw to execute the xauth command with root privileges
and make it read an arbitrary input file. (CVE-2010-3316)
|
| Alerts: |
|
Comments (none posted)
php: multiple vulnerabilities
| Package(s): | php |
CVE #(s): | CVE-2010-3710
CVE-2010-3709
CVE-2010-3436
|
| Created: | November 1, 2010 |
Updated: | April 15, 2011 |
| Description: |
From the Mandriva advisory:
Stack consumption vulnerability in the filter_var function in PHP 5.2.x
through 5.2.14 and 5.3.x through 5.3.3, when FILTER_VALIDATE_EMAIL
mode is used, allows remote attackers to cause a denial of service
(memory consumption and application crash) via a long e-mail address
string (CVE-2010-3710).
A NULL pointer dereference was discovered in
ZipArchive::getArchiveComment (CVE-2010-3709).
A possible flaw was discovered in open_basedir (CVE-2010-3436).
|
| Alerts: |
|
Comments (none posted)
proftpd: arbitrary code execution
| Package(s): | proftpd |
CVE #(s): | CVE-2010-3867
|
| Created: | November 2, 2010 |
Updated: | March 15, 2011 |
| Description: |
From the Slackware advisory:
Fixed Telnet IAC stack overflow vulnerability (ZDI-CAN-925), which can
allow remote execution of arbitrary code as the user running the
ProFTPD daemon. Thanks to TippingPoint and the Zero Day Initiative (ZDI).
|
| Alerts: |
|
Comments (1 posted)
python: multiple vulnerabilities
| Package(s): | python |
CVE #(s): | CVE-2009-4134
CVE-2010-1449
CVE-2010-1450
CVE-2010-3492
CVE-2010-3493
|
| Created: | November 1, 2010 |
Updated: | October 18, 2012 |
| Description: |
From the Mandriva advisory:
Buffer underflow in the rgbimg module in Python 2.5 allows remote
attackers to cause a denial of service (application crash) via a large
ZSIZE value in a black-and-white (aka B/W) RGB image that triggers
an invalid pointer dereference (CVE-2009-4134).
Integer overflow in rgbimgmodule.c in the rgbimg module in Python
2.5 allows remote attackers to have an unspecified impact via a large
image that triggers a buffer overflow. NOTE: this vulnerability exists
because of an incomplete fix for CVE-2008-3143.12 (CVE-2010-1449).
Multiple buffer overflows in the RLE decoder in the rgbimg module in
Python 2.5 allow remote attackers to have an unspecified impact via an
image file containing crafted data that triggers improper processing
within the (1) longimagedata or (2) expandrow function (CVE-2010-1450).
The asyncore module in Python before 3.2 does not properly handle
unsuccessful calls to the accept function, and does not have
accompanying documentation describing how daemon applications should
handle unsuccessful calls to the accept function, which makes it
easier for remote attackers to conduct denial of service attacks that
terminate these applications via network connections (CVE-2010-3492).
Multiple race conditions in smtpd.py in the smtpd module in Python 2.6,
2.7, 3.1, and 3.2 alpha allow remote attackers to cause a denial of
service (daemon outage) by establishing and then immediately closing
a TCP connection, leading to the accept function having an unexpected
return value of None, an unexpected value of None for the address,
or an ECONNABORTED, EAGAIN, or EWOULDBLOCK error, or the getpeername
function having an ENOTCONN error, a related issue to CVE-2010-3492
(CVE-2010-3493).
|
| Alerts: |
|
Comments (none posted)
Page editor: Jake Edge
Next page: Kernel development>>