Security
Default "secrets"
One of the simplest principles of cryptography is that the secret keys which are used for encryption must be kept, well, secret. Exposing the key to anyone other than the intended recipient of the message can pretty obviously lead to a compromise of the encrypted data. So, for example, hardcoding a secret key into a firmware image is unlikely to lead to secure communications using that key. Unfortunately, networking device makers—and the creators of free software firmware replacements for those devices—seem to have missed, or ignored, this basic principle.
The problem stems from the SSL keys that are installed into the firmware images for the devices. In many cases, those keys—including the supposedly private key—are generated when the image is built and then flashed into hundreds or thousands of different devices. If one can get access to the private SSL key, traffic encrypted with it (which might include HTTPS or VPN traffic) can be trivially decrypted. As the announcement of the LittleBlackBox project describes, it is, unfortunately, rather easy to obtain said keys; in fact the project provides a database of thousands of private keys indexed by their public key.
In practical terms, that means an attacker can access a vulnerable SSL-protected web site, retrieve the public key certificate, look up the corresponding private key, and decrypt any traffic that is sent or received by the web site. An attacker could also do a man-in-the-middle attack by pretending to be the site in question, as there would be no way to determine that the spoofer wasn't the real site. In order to do either of those things, though, the attacker must get access to the encrypted data stream.
Open, or weakly secured, wireless networks are the easiest way for an attacker to get that access—or to become a man in the middle. As the concerns over Firesheep have shown, there is still a lot of traffic that travels unencrypted over wireless networks. Ironically, HTTPS is touted as a solution to that problem, but that only works if the private keys are kept secret. For the web applications targeted by Firesheep, that is not likely to be a problem, as their private keys were presumably generated individually and kept safe. But for others who might be using wireless networks to configure their routers—or connect to another network via VPN—it could be a much bigger problem.
While reconfiguring your router from the local coffee shop may be a pretty rare event, even having the HTTPS-enabled web server available over the internet gives an attacker the ability to retrieve the public key, which can then be looked up in the LittleBlackBox database. If that SSL key is used for other things like VPN—something that might well be used at an open WiFi hotspot—that traffic is at risk as well. The right solution seems clear: don't supply default "secrets". In some ways, this problem parallels the longstanding, but hopefully improving, situation with default administrative passwords.
Device manufacturers and firmware projects should not be shipping SSL keys and either generate them at "first boot" or provide a way for users to generate and upload their own keys. There are a few different reasons that it isn't always done that way today, from concerns over devices having enough entropy to generate a random key to the amount of time it can take to generate a key on a slow CPU, but those reasons aren't really offset by the damage that could be done. Users who enable HTTPS access to their devices do so with the idea that it will be more secure, and can be used in places where unencrypted communication doesn't make sense.
There are also hurdles to overcome in either creating a key for each device (and/or firmware image) or providing instructions for users but, once again, that really doesn't help users that are relying on the device for secure communications. While some in the DD-WRT community don't see it as a big problem it is likely more serious than they are crediting. It would make far more sense to disable HTTPS access entirely—perhaps requiring a manual process to generate keys and enable that access—than it does to provide well-known keys.
While the problem highlighted by LittleBlackBox isn't earth-shattering, it does show the sometimes cavalier attitude towards security that is shown by some in the embedded device arena. When you are selling (or providing) a device or firmware that is meant to secure someone's network, it makes sense to proceed carefully. And to keep secrets, secret.
Brief items
Security quotes of the week
Breaking GSM With a $15 Phone Plus Smarts (Wired)
Wired's Threat Level blog looks at a GSM cellphone eavesdropping demonstration made at the recent Chaos Computer Club Congress in Berlin. "As part of this background communication, GSM networks send out strings of identifying information, as well as essentially empty "Are you there?" messages. Empty space in these messages is filled with buffer bytes. Although a new GSM standard was put in place several years ago to turn these buffers into random bytes, they in fact remain largely identical today, under a much older standard. [...] This allows the researchers to predict with a high degree of probability the plain-text content of these encrypted system messages. This, combined with a two-terabyte table of precomputed encryption keys (a so-called rainbow table), allows a cracking program to discover the secret key to the sessions encryption in about 20 seconds."
Bareil: Linux Security, one year later...
Nicolas Bareil looks at GNU/Linux security in 2010. "Bug #1: Disabling frontier: The kernel has to validate each user-provided pointer to check if it is coming from user or kernel space. This is done by access_ok() with a simple comparison of the address against a limit (XXX). Sometimes, the kernel needs to use function which are normally designed to be called by userspace, and as such, theses functions checks the provenance of the pointer... which is embarrassing because the kernel only provides kernel pointers. So the kernel goes evil and cheats by manipulating the boundary via set_fs() in order to make access_ok() always successful. At this moment and until the kernel undoes its boundary manipulation, there is no protection against NULL pointer dereference attack." (Thanks to Patrick Guignot)
Spengler: False Boundaries and Arbitrary Code Execution
Brad Spengler has posted a review of Linux capabilities and how they can be leveraged for full root privileges on the grsecurity blog. In short, 20 of the 35 capabilities bits allow actions that can result in root privileges from an exploitable program. "As mentioned earlier, there are 35 capabilities currently implemented. I'll now discuss each capability that is effectively equal to root and a rough description of how each transition is made. I will try to make a distinction between cases that are generally applicable and those that are situational. Since we've already established that real uid 0 is equivalent to having full capabilities on any normal system, I'll assume we're a non-root user with only the mentioned capability raised." (Thanks to Dan Carpenter.)
New vulnerabilities
dbus: denial of service
| Package(s): | dbus | CVE #(s): | CVE-2010-4352 | ||||||||||||||||||||||||||||||||||||||||||||||||
| Created: | December 27, 2010 | Updated: | May 3, 2011 | ||||||||||||||||||||||||||||||||||||||||||||||||
| Description: | From the Red Hat bugzilla:
A stack overflow flaw was found in the way the D-BUS message bus service / messaging facility validated messages with excessive number of nested variants. A local, authenticated user could use this flaw to cause dbus daemon to crash (denial of service) via a specially-crafted message sent to the system bus. | ||||||||||||||||||||||||||||||||||||||||||||||||||
| Alerts: |
| ||||||||||||||||||||||||||||||||||||||||||||||||||
drupal-views: cross-site scripting
| Package(s): | drupal-views | CVE #(s): | |||||||||
| Created: | January 4, 2011 | Updated: | January 5, 2011 | ||||||||
| Description: | From the Drupal advisory:
The Views module provides a flexible method for Drupal site designers to control how lists and tables of content are presented. Under certain circumstances, Views could display parts of the page path without escaping, resulting in a relected Cross Site Scripting (XSS) vulnerability. An attacker could exploit this to gain full administrative access. | ||||||||||
| Alerts: |
| ||||||||||
eclipse: cross-site scripting
| Package(s): | eclipse | CVE #(s): | CVE-2010-4647 | ||||||||||||||||
| Created: | December 27, 2010 | Updated: | May 19, 2011 | ||||||||||||||||
| Description: | From the Red Hat bugzilla:
It was reported that the Eclipse Help Contents were vulnerable to Cross Site Scripting vulnerabilities in the /help/index.jsp and /help/advanced/content.jsp URLs that are served by the built-in Jetty Web Server plugin. | ||||||||||||||||||
| Alerts: |
| ||||||||||||||||||
evince: arbitrary code execution
| Package(s): | evince | CVE #(s): | CVE-2010-2640 CVE-2010-2641 CVE-2010-2642 CVE-2010-2643 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Created: | January 5, 2011 | Updated: | January 30, 2012 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Description: | From the Ubuntu advisory:
Jon Larimer discovered that Evince's font parsers incorrectly handled certain buffer lengths when rendering a DVI file. By tricking a user into opening or previewing a DVI file that uses a specially crafted font file, an attacker could crash evince or execute arbitrary code with the user's privileges. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Alerts: |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
gif2png: arbitrary code execution
| Package(s): | gif2png | CVE #(s): | CVE-2009-5018 | ||||||||
| Created: | January 5, 2011 | Updated: | January 17, 2011 | ||||||||
| Description: | From the Gentoo advisory:
gif2png contains a command line parsing vulnerability that may result in a stack overflow due to an unexpectedly long input filename. A remote attacker could entice a user to open a specially crafted image, possibly resulting in the execution of arbitrary code with the privileges of the user running the application, or a Denial of Service. Note that applications relying on gif2png to process images can also trigger the vulnerability. | ||||||||||
| Alerts: |
| ||||||||||
kernel: multiple vulnerabilities
| Package(s): | kernel | CVE #(s): | CVE-2010-4347 CVE-2010-4258 CVE-2010-4165 CVE-2010-4175 CVE-2010-4163 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Created: | January 3, 2011 | Updated: | August 9, 2011 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Description: | From the openSUSE advisory:
CVE-2010-4347: A local user could inject ACPI code into the kernel via the world-writable "custom_debug" file, allowing local privilege escalation. CVE-2010-4258: A local attacker could use a Oops (kernel crash) caused by other flaws to write a 0 byte to a attacker controlled address in the kernel. This could lead to privilege escalation together with other issues. CVE-2010-4165: The do_tcp_setsockopt function in net/ipv4/tcp.c in the Linux kernel did not properly restrict TCP_MAXSEG (aka MSS) values, which allows local users to cause a denial of service (OOPS) via a setsockopt call that specifies a small value, leading to a divide-by-zero error or incorrect use of a signed integer. CVE-2010-4175: A local attacker could cause memory overruns in the RDS protocol stack, potentially crashing the kernel. So far it is considered not to be exploitable. CVE-2010-4163: By submitting certain I/O requests with 0 length, a local user could have caused a kernel panic. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Alerts: |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
kernel: multiple vulnerabilities
| Package(s): | kernel | CVE #(s): | CVE-2010-3699 CVE-2010-4161 CVE-2010-4242 CVE-2010-4247 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Created: | January 4, 2011 | Updated: | July 15, 2011 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Description: | From the Red Hat advisory:
* A flaw was found in the Xenbus code for the unified block-device I/O interface back end. A privileged guest user could use this flaw to cause a denial of service on the host system running the Xen hypervisor. (CVE-2010-3699) * The fix for Red Hat Bugzilla bug 484590 as provided in RHSA-2009:1243 introduced a regression. A local, unprivileged user could use this flaw to cause a denial of service. (CVE-2010-4161) * A NULL pointer dereference flaw was found in the Bluetooth HCI UART driver in the Linux kernel. A local, unprivileged user could use this flaw to cause a denial of service. (CVE-2010-4242) * It was found that a malicious guest running on the Xen hypervisor could place invalid data in the memory that the guest shared with the blkback and blktap back-end drivers, resulting in a denial of service on the host system. (CVE-2010-4247) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Alerts: |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
libwmf: multiple vulnerabilities
| Package(s): | libwmf | CVE #(s): | CVE-2007-0455 CVE-2007-3472 CVE-2007-3473 CVE-2007-3474 CVE-2007-3475 CVE-2007-3476 CVE-2007-3477 CVE-2007-3478 | ||||||||||||
| Created: | January 5, 2011 | Updated: | January 5, 2011 | ||||||||||||
| Description: | From the Red Hat advisory:
libwmf embeds an old version of gd (2.0.1beta) which has a number of vulnerabilities associated with it. CVE-2007-0455 CVE-2007-3472 CVE-2007-3473 CVE-2007-3474 CVE-2007-3475 CVE-2007-3476 CVE-2007-3477 CVE-2007-3478 Cursory inspection of one of the patch diffs shows that no patches have been applied to libwmf. | ||||||||||||||
| Alerts: |
| ||||||||||||||
libxml2: arbitrary code execution
| Package(s): | libxml2 | CVE #(s): | CVE-2010-4494 | ||||||||||||||||||||||||||||||||||||||||||||||||||||
| Created: | December 27, 2010 | Updated: | April 15, 2011 | ||||||||||||||||||||||||||||||||||||||||||||||||||||
| Description: | From the Debian advisory:
Yang Dingning discovered a double free in libxml's Xpath processing, which might allow the execution of arbitrary code. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Alerts: |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
mantis: multiple vulnerabilities
| Package(s): | mantis | CVE #(s): | CVE-2010-3763 CVE-2010-4348 CVE-2010-4349 CVE-2010-4350 | ||||||||||||
| Created: | January 3, 2011 | Updated: | November 9, 2012 | ||||||||||||
| Description: | From the Red Hat bugzilla entry:
Cross-site scripting (XSS) vulnerability in core/summary_api.php in MantisBT before 1.2.3 allows remote attackers to inject arbitrary web script or HTML via the Summary field, a different vector than CVE-2010-3303. (CVE-2010-3763) From the Red Hat bugzilla entry: The MantisBT project was notified by Gjoko Krstic of Zero Science Lab (gjoko@zeroscience.mk) of multiple vulnerabilities affecting MantisBT <1.2.4. The two following advisories have been released explaining the vulnerabilities in greater detail: http://www.zeroscience.mk/en/vulnerabilities/ZSL-2010-4983.php http://www.zeroscience.mk/en/vulnerabilities/ZSL-2010-4984.php (CVE-2010-4348, CVE-2010-4349, CVE-2010-4350) | ||||||||||||||
| Alerts: |
| ||||||||||||||
opensc: arbitrary code execution
| Package(s): | opensc | CVE #(s): | CVE-2010-4523 | ||||||||||||||||||||||||
| Created: | January 4, 2011 | Updated: | January 22, 2014 | ||||||||||||||||||||||||
| Description: | From the Red Hat bugzilla:
Three stack-based buffer overflow flaws were found in the way OpenSC device drivers for A-Trust ACOS, ACS ACOS5 and STARCOS SPK 2.3 based smart cards processed certain values of card serial number. A local attacker could use this flaw to execute arbitrary code, with the privileges of the user running the opesc-tool or opensc-explorer binaries via a malicious smart card, with specially-crafted value of its serial number, inserted to the system. | ||||||||||||||||||||||||||
| Alerts: |
| ||||||||||||||||||||||||||
perl-IO-Socket-SSL: denial of service
| Package(s): | perl-IO-Socket-SSL | CVE #(s): | CVE-2010-4334 | ||||||||||||
| Created: | December 27, 2010 | Updated: | May 18, 2011 | ||||||||||||
| Description: | From the Red Hat bugzilla:
A Debian bug report indicated that using the IO::Socket::SSL perl module, if the verify_mode were set to 0x03 (verify peer, fail verification if no peer certificate exists), that the requests were removed unless either the ca_file or ca_path were supplied. This means that IO::Socket::SSL "fails open" if the user forgets to supply information about an acceptable set of trusted CAs, rather than "failing closed" (denying access by default, rather than allowing it). | ||||||||||||||
| Alerts: |
| ||||||||||||||
phpmyadmin: multiple vulnerabilities
| Package(s): | phpmyadmin | CVE #(s): | CVE-2010-4480 CVE-2010-4481 | ||||||||||||||||||||
| Created: | December 31, 2010 | Updated: | March 30, 2011 | ||||||||||||||||||||
| Description: | From the Debian advisory:
Cross site scripting was possible in errors, that allowed a remote attacker to inject arbitrary web script or HTML. (CVE-2010-4480) Display of PHP's phpinfo() function was available to world, but only if this functionality had been enabled (defaults to off). This may leak some information about the host system. (CVE-2010-4481) | ||||||||||||||||||||||
| Alerts: |
| ||||||||||||||||||||||
wordpress: SQL injection
| Package(s): | wordpress | CVE #(s): | CVE-2010-4257 | ||||||||||||||||||||
| Created: | December 29, 2010 | Updated: | January 10, 2011 | ||||||||||||||||||||
| Description: | From the Debian advisory: Vladimir Kolesnikov discovered a SQL injection vulnerability in wordpress, a weblog manager. An authenticated users could execute arbitrary SQL commands via the Send Trackbacks field. | ||||||||||||||||||||||
| Alerts: |
| ||||||||||||||||||||||
Page editor: Jake Edge
Next page:
Kernel development>>