LWN.net Logo

Security

A high-level view of the MeeGo security landscape

By Jake Edge
November 17, 2010

Several members of the MeeGo security team were on hand at the 2010 MeeGo conference to talk about what kinds of threats they will be trying to address—and why—as well as a security framework to enable MeeGo integrators and application developers to handle security tasks. MeeGo security architect Ryan Ware of Intel looked at the what and the why, while Elena Reshetova and Casey Schaufler of Nokia presented on the Mobile Simplified Security Framework (MSSF). As might be guessed from the presence of Schaufler, the Smack kernel security module plays a prominent role in the access control portion of MSSF. This week, we'll cover Ware's presentation and look at Reshetova and Schaufler's next week.

Ware started with a look back at 1990 by way of a justification of the need for MeeGo security solutions. In 1990, Intel had 25MHz 386 processors, the Simpsons were on TV, and there were all of 12 CERT security alerts for the year. All of those alerts "fit on one slide easily" and contain some amusing entries like "rumor of alleged attack" and "security probes from Italy". He listed, again on one slide, the conferences and other notable computer security news for the year. Things have changed just a little bit since then.

Fast-forwarding to the present, there have been 4221 CVEs so far this year, Intel has 3+GHz chips, and the Simpsons are still on TV. When looking at the growth of malware, there is an inflection point in 1996, which is probably associated with wider usage of the internet. "The internet is a petri dish" where all kinds of malware can grow and change. If you put a stock Windows XP system on the internet today without a firewall, it will be infected before you can get the updates installed; it only takes an average of four minutes before that happens, he said.

There is a huge financial incentive these days for those who write malware, which has changed the landscape significantly. You can now get "malware as a service" or rent botnets ($8-90/1000 bots "depending on quantity", he said). In the pwn2own contest at CanSecWest, someone with a working iPhone exploit was unwilling to release it for the $15,000 prize as they believed they could get more elsewhere—and did, with rumors of a six-figure sum.

There are also "spearphishing" efforts like Aurora that targeted Google and 30 other companies, including Intel, last year. It targeted specific individual employees, sending them an email that looked it came from someone they knew. When the PDF or JPG inside was opened, it appeared to be an innocuous file of that type, but actually infected their machine with a worm that looked for source code repositories. Once found, the contents of those repositories were slowly—so that intrusion detection systems weren't alerted—sent elsewhere. The Stuxnet worm/virus is another example of this new kind of "persistent" threat.

With MeeGo, there are new usage models where desktop data is migrating to mobile phones, which are much more easily lost, for example. People are doing banking from their phones as well. When Ware asked how many in the audience had used their phone for banking, he got quite a few hands; "you're all screwed", he said. Those credentials are stored somewhere in the phone for an attacker (or thief) to find. There are also various efforts to publish your location or turn your phone into a credit card, all of which have various dangers.

Because the number of Linux devices is growing quickly, it is becoming more of a target. For reference, he said there are more than a billion Windows-installed systems—some botnets have more than a million bots—but the smartphone market is growing at a rate (35.5%/year) that will go beyond that soon. At that rate, the expected sales of smartphones in 2014 is 506 million. In addition, the smartphone market is getting less fragmented and he sees iOS and Linux as likely to be the only players before too long.

The focus on mobile Linux security is growing, he said. He noted the recent Coverity study of the Android kernel that found 88 high-risk defects and there were "some interesting things in there". The report will not be available for a bit as Coverity gave Google 60 days to fix the problems before the report will be released. Ware noted that the study found that the defect rate for the code written for Android was "significantly higher than for the rest of the kernel".

MSSF was originally developed for smartphones, but has been broadened to support all of the MeeGo vertical markets (netbook, connected TV, in-vehicle-infotainment (IVI), ...). At a high level, the goals for MSSF are to provide protections for users of devices, the device itself, and for new services that are envisioned for MeeGo devices.

For users, that includes protecting things like login credentials and cookies, but also to try to prevent malicious software from being able to do things like making expensive phone calls without the knowledge or consent of the device owner. Protecting the device entails protecting the SIM lock and ensuring that regulatory requirements (for things like radio frequency emissions) are strictly adhered to. New services like mobile payment also need protection, he said.

The MeeGo security team is doing things beyond just MSSF. It ensures that the external facing MeeGo infrastructure is kept secure. That includes things like source code repositories and open build service packages. The team also ensures that MeeGo images are secure by not having insecure defaults on network services, patching packages for security vulnerabilities, and issuing MeeGo advisories.

MeeGo "can't be secure without you guys", he said. The team could do static analysis and code reviews for 80 hours a week and still not find everything. He asked that folks keep an eye out and point out any flaws they find to security@meego.com. There is also a new MeeGo-security-discussion mailing list and weekly IRC meetings of the security team are planned in the near future.

In answer to some audience questions, Ware said he was concerned about security issues surrounding "cloud" applications, but hadn't looked at it specifically yet. It is "something to look at in the future". He also was not interested in talking about DRM solutions, though some in the audience clearly were. He worked on DRM five years ago and was glad to not be working on it any more. "I don't want to fix someone's broken business model", he said. Others who need those kinds of "solutions" will undoubtedly come up with them.

Comments (10 posted)

Brief items

Security quote of the week

GSM equipment manufacturers and mobile operators have shown no interest in fixing gaping holes in their security system.
-- Harald Welte

Comments (none posted)

An OpenSSL race condition

The OpenSSL project has issued an advisory of a race condition which exists in versions prior to 0.9.8p or 1.0.0b. Successfully exploiting this race can enable a remote attacker to inject code into a server using OpenSSL. It's worth noting, though, that only servers which are (1) multi-threaded, and (2) using OpenSSL's internal caching are vulnerable. So, in particular, Apache servers are not at risk.

Full Story (comments: 1)

New vulnerabilities

banshee: privilege escalation

Package(s):banshee CVE #(s):CVE-2010-3998
Created:November 12, 2010 Updated:February 22, 2011
Description: From the CVE entry:

The (1) banshee-1 and (2) muinshee scripts in Banshee 1.8.0 and earlier place a zero-length directory name in the LD_LIBRARY_PATH, which allows local users to gain privileges via a Trojan horse shared library in the current working directory.

Alerts:
Mandriva MDVSA-2011:034 2011-02-21
Fedora FEDORA-2010-16907 2010-10-28
Fedora FEDORA-2010-16916 2010-10-28
Fedora FEDORA-2010-17021 2010-10-31

Comments (none posted)

bristol: privilege escalation

Package(s):bristol CVE #(s):CVE-2010-3351
Created:November 15, 2010 Updated:November 17, 2010
Description: From the CVE entry:

startBristol in Bristol 0.60.5 places a zero-length directory name in the LD_LIBRARY_PATH, which allows local users to gain privileges via a Trojan horse shared library in the current working directory.

Alerts:
Fedora FEDORA-2010-16676 2010-10-27
Fedora FEDORA-2010-16687 2010-10-27
Fedora FEDORA-2010-16714 2010-10-28

Comments (none posted)

bugzilla: multiple vulnerabilities

Package(s):bugzilla CVE #(s):CVE-2010-3764 CVE-2010-3172
Created:November 15, 2010 Updated:January 20, 2011
Description: From the CVE entries:

The Old Charts implementation in Bugzilla 2.12 through 3.2.8, 3.4.8, 3.6.2, 3.7.3, and 4.1 creates graph files with predictable names in graphs/, which allows remote attackers to obtain sensitive information via a modified URL. (CVE-2010-3764)

CRLF injection vulnerability in Bugzilla before 3.2.9, 3.4.x before 3.4.9, 3.6.x before 3.6.3, and 4.0.x before 4.0rc1, when Server Push is enabled in a web browser, allows remote attackers to inject arbitrary HTTP headers and content, and conduct HTTP response splitting attacks, via a crafted URL. (CVE-2010-3172)

Alerts:
Gentoo 201110-03 2011-10-10
openSUSE openSUSE-SU-2011:0020-1 2011-01-10
openSUSE openSUSE-SU-2011:0064-1 2011-01-20
Mandriva MDVSA-2010:252 2010-12-14
Fedora FEDORA-2010-17235 2010-11-04
Fedora FEDORA-2010-17280 2010-11-04
Fedora FEDORA-2010-17274 2010-11-04

Comments (none posted)

gromacs: code execution

Package(s):gromacs CVE #(s):CVE-2010-4001
Created:November 15, 2010 Updated:November 17, 2010
Description: From the Red Hat bugzilla:

Ludwig Nussel discovered that gromacs 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:
Fedora FEDORA-2010-17256 2010-11-04
Fedora FEDORA-2010-17248 2010-11-04

Comments (none posted)

kernel: privilege escalation

Package(s):kernel CVE #(s):CVE-2010-3865
Created:November 11, 2010 Updated:August 9, 2011
Description:

From the openSUSE advisory:

CVE-2010-3865: A iovec integer overflow in RDS sockets was fixed which could lead to local attackers gaining kernel privileges.

Alerts:
Ubuntu USN-1187-1 2011-08-09
Ubuntu USN-1164-1 2011-07-06
Ubuntu USN-1093-1 2011-03-25
Ubuntu USN-1119-1 2011-04-20
Ubuntu USN-1080-2 2011-03-02
Ubuntu USN-1081-1 2011-03-02
Ubuntu USN-1080-1 2011-03-01
Ubuntu USN-1073-1 2011-02-25
SUSE SUSE-SA:2011:007 2011-02-07
Red Hat RHSA-2011:0007-01 2011-01-11
CentOS CESA-2011:0004 2011-01-06
Red Hat RHSA-2011:0004-01 2011-01-04
openSUSE openSUSE-SU-2011:0003-1 2011-01-03
openSUSE openSUSE-SU-2011:0004-1 2011-01-03
SUSE SUSE-SA:2010:057 2010-11-11
openSUSE openSUSE-SU-2010:0933-1 2010-11-11

Comments (none posted)

kernel: denial of service

Package(s):kernel CVE #(s):CVE-2010-3698
Created:November 11, 2010 Updated:August 9, 2011
Description:

From the Red Hat advisory:

A flaw was found in the way KVM (Kernel-based Virtual Machine) handled the reloading of fs and gs segment registers when they had invalid selectors. A privileged host user with access to "/dev/kvm" could use this flaw to crash the host. (CVE-2010-3698, Moderate)

Alerts:
Ubuntu USN-1187-1 2011-08-09
Ubuntu USN-1081-1 2011-03-02
Ubuntu USN-1074-2 2011-02-28
Ubuntu USN-1074-1 2011-02-25
Ubuntu USN-1073-1 2011-02-25
Ubuntu USN-1072-1 2011-02-25
Mandriva MDVSA-2011:029 2011-02-17
Fedora FEDORA-2010-18983 2010-12-17
CentOS CESA-2010:0898 2010-12-14
Red Hat RHSA-2010:0898-01 2010-12-06
Red Hat RHSA-2010:0842-01 2010-11-10

Comments (none posted)

libxml2: code execution

Package(s):libxml2 CVE #(s):CVE-2010-4008
Created:November 11, 2010 Updated:December 8, 2010
Description:

From the Ubuntu advisory:

Bui Quang Minh discovered that libxml2 did not properly process XPath namespaces and attributes. If an application using libxml2 opened a specially crafted XML file, an attacker could cause a denial of service or possibly execute code as the user invoking the program.

Alerts:
Scientific Linux SL-libx-20111206 2011-12-06
Red Hat RHSA-2011:1749-03 2011-12-06
Gentoo 201110-26 2011-10-26
SUSE SUSE-SR:2010:023 2010-12-08
openSUSE openSUSE-SU-2010:1004-1 2010-12-02
Debian DSA-2128-1 2010-12-01
Mandriva MDVSA-2010:243 2010-11-29
Ubuntu USN-1016-1 2010-11-10
Red Hat RHSA-2012:0017-01 2012-01-11
CentOS CESA-2012:0017 2012-01-11
Scientific Linux SL-libx-20120112 2012-01-12
Oracle ELSA-2012-0017 2012-01-12
Oracle ELSA-2012-0324 2012-03-09
Red Hat RHSA-2013:0217-01 2013-01-31
CentOS CESA-2013:0217 2013-02-01
Oracle ELSA-2013-0217 2013-02-01
Scientific Linux SL-ming-20130201 2013-02-01

Comments (none posted)

mod_fcgid: buffer overflow

Package(s):mod_fcgid CVE #(s):CVE-2010-3872
Created:November 17, 2010 Updated:August 10, 2011
Description: The mod_fcgid Apache module is subject to a stack buffer overflow with uncertain effects (but code execution seems plausible).
Alerts:
SUSE SUSE-SU-2011:0885-1 2011-08-10
openSUSE openSUSE-SU-2011:0884-1 2011-08-10
Debian DSA-2140-1 2011-01-05
Fedora FEDORA-2010-17472 2010-11-08
Fedora FEDORA-2010-17434 2010-11-08
Fedora FEDORA-2010-17474 2010-11-08
Gentoo 201207-09 2012-07-09

Comments (none posted)

moodle: cross-site scripting

Package(s):moodle CVE #(s):CVE-2010-4207 CVE-2010-4208 CVE-2010-4209
Created:November 12, 2010 Updated:November 17, 2010
Description: From the openSUSE advisory:

CVE-2010-4207: Cross-site scripting vulnerability in the Flash component infrastructure in YUI allows remote attackers to inject arbitrary web script or HTML via charts/assets/charts.swf.

CVE-2010-4208: Cross-site scripting vulnerability in the Flash component infrastructure in YUI allows remote attackers to inject arbitrary web script or HTML via uploader/assets/uploader.swf.

CVE-2010-4209: Cross-site scripting vulnerability in the Flash component infrastructure in YUI allows remote attackers to inject arbitrary web script or HTML via swfstore/swfstore.swf.

Alerts:
SUSE SUSE-SR:2010:021 2010-11-16
Fedora FEDORA-2010-16845 2010-10-28
Fedora FEDORA-2010-16782 2010-10-28
Fedora FEDORA-2010-16764 2010-10-28
openSUSE openSUSE-SU-2010:0937-1 2010-11-12
Mageia MGASA-2013-0117 2013-04-18

Comments (none posted)

mysql: denial of service

Package(s):mysql-5.1, mysql-dfsg-5.0, mysql-dfsg-5.1 CVE #(s):CVE-2010-3834
Created:November 11, 2010 Updated:July 19, 2011
Description:

From the Ubuntu advisory:

It was discovered that MySQL incorrectly handled materializing a derived table that required a temporary table for grouping. An authenticated user could exploit this to make MySQL crash, causing a denial of service. (CVE-2010-3834)

Alerts:
openSUSE openSUSE-SU-2011:1250-1 2011-11-16
openSUSE openSUSE-SU-2011:0799-1 2011-07-19
openSUSE openSUSE-SU-2011:0774-1 2011-07-19
openSUSE openSUSE-SU-2011:0743-1 2011-07-06
Debian DSA-2143-1 2011-01-14
Ubuntu USN-1017-1 2010-11-11
Gentoo 201201-02 2012-01-05
Ubuntu USN-1397-1 2012-03-12

Comments (none posted)

openssl: remote code execution

Package(s):openssl CVE #(s):CVE-2010-3864
Created:November 17, 2010 Updated:November 30, 2010
Description: The OpenSSL project has issued an advisory of a race condition which exists in versions prior to 0.9.8p or 1.0.0b. Successfully exploiting this race can enable a remote attacker to inject code into a server using OpenSSL. It's worth noting, though, that only servers which are (1) multi-threaded, and (2) using OpenSSL's internal caching are vulnerable. So, in particular, Apache servers are not at risk. See this advisory for more information.
Alerts:
Gentoo 201110-01 2011-10-09
SUSE SUSE-SR:2010:022 2010-11-30
Ubuntu USN-1018-1 2010-11-18
Debian DSA-2125-1 2010-11-22
Slackware SSA:2010-326-01 2010-11-22
openSUSE openSUSE-SU-2010:0965-2 2010-11-22
Fedora FEDORA-2010-17847 2010-11-17
Fedora FEDORA-2010-17827 2010-11-17
openSUSE openSUSE-SU-2010:0965-1 2010-11-19
Mandriva MDVSA-2010:238 2010-11-17
Red Hat RHSA-2010:0888-01 2010-11-16

Comments (none posted)

openswan: code execution

Package(s):openswan CVE #(s):CVE-2010-3752 CVE-2010-3753
Created:November 17, 2010 Updated:November 17, 2010
Description: From the Red Hat advisory: two input sanitization flaws were found in the Openswan client-side handling of Cisco gateway banners. A malicious or compromised VPN gateway could use these flaws to execute arbitrary code on the connecting Openswan client.
Alerts:
Red Hat RHSA-2010:0892-01 2010-11-16
Mageia MGASA-2012-0300 2012-10-20

Comments (none posted)

perl-CGI: multiple vulnerabilities

Package(s):perl-CGI CVE #(s):
Created:November 16, 2010 Updated:November 17, 2010
Description: From the Mandriva advisory:

A new version of the CGI Perl module has been released to CPAN, which fixes several security bugs which directly affect Bugzilla (these two security bugs where first discovered as affecting Bugzilla, then identified as being bugs in CGI.pm itself).

Alerts:
Mandriva MDVSA-2010:237 2010-11-16

Comments (none posted)

proftpd: code execution

Package(s):proftpd CVE #(s):CVE-2010-4221
Created:November 11, 2010 Updated:December 24, 2010
Description:

From the proftpd bugzilla entry:

The flaw exists within the proftpd server component which listens by default on TCP port 21. When reading user input if a TELNET_IAC escape sequence is encountered the process miscalculates a buffer length counter value allowing a user controlled copy of data to a stack buffer. A remote attacker can exploit this vulnerability to execute arbitrary code under the context of the proftpd process.

Alerts:
Slackware SSA:2010-357-02 2010-12-24
Fedora FEDORA-2010-17220 2010-11-03
Mandriva MDVSA-2010:227 2010-11-11
Fedora FEDORA-2010-17091 2010-11-02
Fedora FEDORA-2010-17098 2010-11-02

Comments (none posted)

systemtap: privilege execution

Package(s):systemtap CVE #(s):CVE-2010-4170
Created:November 17, 2010 Updated:November 23, 2010
Description: The staprun utility contains two vulnerabilities which can be exploited for privilege escalation by local users; see this advisory for (a little) more information.
Alerts:
Debian DSA-2348-1 2011-11-17
CentOS CESA-2010:0895 2010-11-17
Fedora FEDORA-2010-17868 2010-11-18
Fedora FEDORA-2010-17873 2010-11-18
Fedora FEDORA-2010-17865 2010-11-18
CentOS CESA-2010:0894 2010-11-17
Red Hat RHSA-2010:0894-01 2010-11-17
Red Hat RHSA-2010:0895-01 2010-11-17

Comments (none posted)

Page editor: Jake Edge
Next page: Kernel development>>

Copyright © 2010, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds