February 29, 2012
This article was contributed by Nathan Willis
Fedora is exploring a new security feature in the upcoming Fedora 17 release that is intended to make firewall configuration simpler for the average user. Based on features found in the new, D-Bus-enabled firewall tool, the Network Zones feature will automatically activate different sets of firewall rules based on the network location, and provide sensible defaults for different classes of network — from trusted connections at home to potentially dangerous open WiFi networks in public.
Background
At the heart of Network Zones is the notion that each network has an associated trust level — fully trusted, mostly trusted, mostly untrusted, or fully untrusted. To serve the needs of users that regularly move between network locations, the firewall should ratchet up the netfilter rules to a stricter level when in an untrusted location such as a restaurant or public park, but ratchet them back down when reconnecting to the home or office network. Exactly which rules are applied at each trust level is up to the user or administrator.
Existing firewall tools, however, fall short in two ways. First, they know only about network interfaces, not about the networks themselves (trust information in particular). Second, they require stopping and restarting the firewall in order to change any settings. Red Hat's Thomas Woerner developed a solution tackling both issues based on GNOME's NetworkManager, and a new firewall application named Firewalld.
NetworkManager, which was designed to keep track of WiFi networks and opportunistically switch between them when roaming, already implemented some of the required functionality. It maintains a history of networks visited, and allows users to associate preferences with each saved network connection. Adding a trust level to each network connection would allow NetworkManager to instruct the firewall to change to a different rule set.
Firewalld solves the restarting problem by running as a D-Bus enabled daemon, listening for state change commands, and emitting information on the current status of the firewall. NetworkManager also uses D-Bus, so it can simply send a message to Firewalld telling it to raise or lower specific firewall rules as the situation dictates. To keep the firewall state consistent, though, Firewalld must be the only application modifying the settings (thus requiring that other firewall management packages be disabled), and for security reasons, all rule change requests accepted by Firewalld must be authenticated with PolicyKit.
In order for Firewalld to expose a straightforward set of configuration
options to users (and other applications) over D-Bus, the firewall rule set
had to be refactored into a clean set of distinct
chains that correspond to individual features and services. Thus,
adding or removing a rule in one chain will not interfere with the others.
Firewalld implements chains for feature sets like virtualization,
masquerading (NAT), port forwarding, and open ports, plus predefined chains for individual services like Samba or FTP.
The Zones zone
The official Network Zones proposal on the Fedora wiki defines nine initial network trust levels: trusted (meaning fully trusted, with all incoming traffic permitted), home, work, and internal (all three of which are for mostly trusted networks), dmz, public, and external (which are mostly untrusted), and block and drop (which are both fully untrusted).
Firewalld provides an initial configuration for each zone in an XML
file. By default, the mostly-trusted zones open only a select set of
ports (SSH, mDNS, Samba, and Internet Printing Protocol (IPP) for the
home and internal zones, and SSH and IPP for
work). The mostly-untrusted zones are more restricted, with SSH
being the only allowed protocol for public and dmz, but
SSH allowed and IP Masquerading enabled for external — though it's
not exactly clear what the benefit of the latter is. The block zone rejects all incoming connections, while the drop zone drops them silently.
Of course, the intent is for users or system administrators to customize each of the zones' configuration as desired, allowing some differentiation between the otherwise-identical offerings. There are GUI and command-line utilities (named firewall-config and firewall-cmd, respectively) for examining and altering the configuration options; however as of today not all of the Firewalld rules are supported in either tool.
There does not appear to be a DTD or XML Schema for the zone configuration or firewall rules yet, but the syntax is straightforward. Individual services are enabled with a <service name=servicename> element; the other available firewall options are enabled by adding an element specific to the feature, such as <masquerade enabled="True"/>.
Network Zone integration is available in the NetworkManager 0.9.4
release, which will be part of Fedora 17, allowing users to assign a trust
level to each of their saved networks, as well as a default zone to apply
to unknown network connections. A system tray applet will display the
current firewall state in GNOME Shell. The project has discussed
adding the same functionality to KDE's network manager as well.
Firewalld was first made available in Fedora 15, but with the completion of the Network Zones support, it is slated to become the default firewall configuration tool in Fedora 17 (scheduled for release in early May 2012). Network zone support is not the only benefit of the daemon-like firewall approach — D-Bus controls open the door for other dynamic features in the future, like triggering temporary firewall rules without manual intervention, and desktop notifications triggered by firewall events.
The Firewalld project is not resting on its laurels, however. The future plans include support for granting or limiting access to the configuration tools on a per-user basis, and more abstract firewall rules based on metadata — such as "allow external access to music sharing applications." Network Zones is of clear benefit to laptop users, who both expose their systems to the greatest risk while roaming, and have had the hardest time finding a balanced firewall policy. But the possibilities enabled by a dynamically controlled firewall extend further; only time will tell what roles it can fill that a static configuration hasn't.
Comments (12 posted)
Brief items
7. We are right now looking at you through your webcam. Do you always move your lips like that when you read? We also recorded what you were doing last week and are sending the video to (you know who). If the prior statements are not true, it's because in addition to everything else, we reserve the right to lie to you, and you agree to believe us and hold us harmless for any and all such lies. Furthermore, if we are not recording everything you're doing through your webcam, it's either because we haven't figured out how, you're just not that interesting, or both.
8. We are serious about all of the above. So don't go trying to sue us later with some nonsense like "I thought that was all satire." All your privacy are belong to us. We mean it.
-- Parts of the
Skipity privacy policy
The aim of our sponsorship is simple: we have a big learning opportunity when we receive full end-to-end exploits. Not only can we fix the bugs, but by studying the vulnerability and exploit techniques we can enhance our mitigations, automated testing, and sandboxing. This enables us to better protect our users.
While we’re proud of Chrome’s leading track record in past competitions,
the fact is that not receiving exploits means that it’s harder to learn and
improve. To maximize our chances of receiving exploits this year, we’ve
upped the ante. We will directly sponsor up to $1 million worth of rewards [...]
--
Google
ponies up for Chrome browser exploits
It doesn't take more than a few minutes of thought to see the utterly disastrous ramifications of the "right to be forgotten" approach, and the cascading damage to free speech that could easily spread malignantly across the global Internet as a result.
The crux of the matter is simple enough. Even if search engine results are selectively expunged on demand, the "upsetting" material in question will still likely exist on the Internet itself, still subject to being located by other means, including via sites that merely discuss related topics, situations, companies, or individuals.
--
Lauren Weinstein
Comments (5 posted)
A proposal to add a DRM layer for web audio and video has been rather controversial on the W3C HTML mailing list, as
reported by ars technica. The
Encrypted Media Extensions proposal authored by Google, Microsoft, and Netflix would add an optional layer for protected media content, but Mozilla and others, including Google's Ian Hickson who is the WHATWG HTML specification editor, have spoken up against the proposal. "
'I believe this proposal is unethical and that we should not pursue it,' he [Hickson] wrote in response to a message that Microsoft's Adrian Bateman posted on the mailing list about the draft. 'The proposal above does not provide robust content protection, so it would not address this use case even if it wasn't unethical.'"
Comments (90 posted)
The Mozilla Foundation has announced the availability of the
Collusion add-on for
Firefox. "
Collusion is an experimental add-on for Firefox and allows
you to see all the third parties that are tracking your movements across
the Web. It will show, in real time, how that data creates a spider-web of
interaction between companies and other trackers."
Comments (15 posted)
New vulnerabilities
asterisk: denial of service
| Package(s): | asterisk |
CVE #(s): | CVE-2012-0885
|
| Created: | February 23, 2012 |
Updated: | February 29, 2012 |
| Description: |
From the Gentoo advisory:
A vulnerability has been found in Asterisk's handling of certain
encrypted streams where the res_srtp module has been loaded but video
support has not been enabled.
A remote attacker could send a specially crafted SDP message to the
Asterisk daemon, possibly resulting in a Denial of Service condition. |
| Alerts: |
|
Comments (none posted)
csound: code execution
| Package(s): | csound |
CVE #(s): | CVE-2012-0270
|
| Created: | February 28, 2012 |
Updated: | March 14, 2012 |
| Description: |
From the Secunia advisory:
Secunia Research has discovered two vulnerabilities in Csound, which
can be exploited by malicious people to compromise a user's system.
1) A boundary error within the "getnum()" function (util/heti_main.c)
can be exploited to cause a stack-based buffer overflow via a
specially crafted hetro file.
2) A boundary error within the "getnum()" function (util/pv_import.c)
can be exploited to cause a stack-based buffer overflow via a
specially crafted PVOC file.
Successful exploitation allows execution of arbitrary code, but
requires tricking a user into converting a malicious file. |
| Alerts: |
|
Comments (none posted)
drupal6: multiple vulnerabilities
| Package(s): | drupal6 |
CVE #(s): | |
| Created: | February 27, 2012 |
Updated: | February 29, 2012 |
| Description: |
Multiple vulnerabilities were fixed in Drupal 6.23. Drupal 6.24 contains additional bug fixes. |
| Alerts: |
|
Comments (none posted)
drupal7: multiple vulnerabilities
| Package(s): | drupal7 |
CVE #(s): | |
| Created: | February 27, 2012 |
Updated: | February 29, 2012 |
| Description: |
Drupal 7.11 fixes multiple vulnerabilities. Drupal 7.12 contains additional bug fixes.
|
| Alerts: |
|
Comments (none posted)
fex: fixes a regression in a previous update
| Package(s): | fex |
CVE #(s): | CVE-2012-0869
|
| Created: | February 27, 2012 |
Updated: | February 29, 2012 |
| Description: |
From the Debian advisory:
It was discovered that the last security update for F*X, DSA-2414-1,
introduced a regression. Updated packages are now available to address
this problem. |
| Alerts: |
|
Comments (none posted)
glibc: format string protection mechanism bypass
| Package(s): | glibc |
CVE #(s): | CVE-2012-0864
|
| Created: | February 27, 2012 |
Updated: | March 22, 2012 |
| Description: |
From the Red Hat bugzilla:
In the Phrack article "A Eulogy for Format Strings", a researcher using
nickname "Captain Planet" reported an integer overflow flaw in the format
string protection mechanism offered by FORTIFY_SOURCE. A remote attacker could provide a specially crafted executable, leading to FORTIFY_SOURCE format string protection mechanism bypass, when executed. |
| Alerts: |
|
Comments (none posted)
kernel: denial of service
| Package(s): | kernel |
CVE #(s): | CVE-2011-2518
|
| Created: | February 29, 2012 |
Updated: | February 29, 2012 |
| Description: |
The TOMOYO Linux security module does not properly handle mount() calls, allowing an unprivileged local process to oops the kernel. |
| Alerts: |
|
Comments (none posted)
kernel: denial of service
| Package(s): | kernel |
CVE #(s): | CVE-2011-4097
|
| Created: | February 29, 2012 |
Updated: | March 6, 2012 |
| Description: |
The kernel's calculation of out-of-memory scores could result in the untimely demise of the wrong process. A local user could use this error to kill an unrelated process. |
| Alerts: |
|
Comments (none posted)
libvpx: denial of service
| Package(s): | libvpx |
CVE #(s): | CVE-2012-0823
|
| Created: | February 27, 2012 |
Updated: | February 29, 2012 |
| Description: |
From the CVE entry:
VP8 Codec SDK (libvpx) before 1.0.0 "Duclair" allows remote attackers to cause a denial of service (application crash) via (1) unspecified "corrupt input" or (2) by "starting decoding from a P-frame," which triggers an out-of-bounds read, related to "the clamping of motion vectors in SPLITMV blocks". |
| Alerts: |
|
Comments (none posted)
maradns: denial of service
| Package(s): | maradns |
CVE #(s): | CVE-2012-0024
|
| Created: | February 23, 2012 |
Updated: | February 29, 2012 |
| Description: |
From the Gentoo advisory:
MaraDNS does not properly randomize hash functions to protect against
hash collision attacks.
A remote attacker could send many specially crafted DNS recursive
queries, possibly resulting in a Denial of Service condition. |
| Alerts: |
|
Comments (none posted)
notmuch: information disclosure
| Package(s): | notmuch |
CVE #(s): | CVE-2011-1103
|
| Created: | February 23, 2012 |
Updated: | March 19, 2012 |
| Description: |
From the Debian advisory:
It was discovered that Notmuch, an email indexer, did not sufficiently
escape Emacs MML tags. When using the Emacs interface, a user could
be tricked into replying to a maliciously formatted message which could
lead to files from the local machine being attached to the outgoing
message. |
| Alerts: |
|
Comments (none posted)
openjdk: sandbox bypass
| Package(s): | openjdk |
CVE #(s): | CVE-2012-0507
|
| Created: | February 29, 2012 |
Updated: | May 10, 2012 |
| Description: |
The openjdk AtomicReferenceArray class does not check the type of the incoming array, leading to a JVM crash or sandbox bypass. |
| Alerts: |
|
Comments (none posted)
postgresql: multiple vulnerabilities
| Package(s): | postgresql |
CVE #(s): | CVE-2012-0866
CVE-2012-0867
CVE-2012-0868
|
| Created: | February 27, 2012 |
Updated: | September 28, 2012 |
| Description: |
From the Debian advisory:
CVE-2012-0866:
It was discovered that the permissions of a function called by a
trigger are not checked. This could result in privilege escalation.
CVE-2012-0867:
It was discovered that only the first 32 characters of a host name
are checked when validating host names through SSL certificates.
This could result in spoofing the connection in limited
circumstances.
CVE-2012-0868:
It was discovered that pg_dump did not sanitise object names.
This could result in arbitrary SQL command execution if a
malformed dump file is opened.
See the PostgreSQL 9.1.3, 9.0.7, 8.4.11 and 8.3.18 update announcement for more information. |
| Alerts: |
|
Comments (none posted)
puppet: two privilege escalations
| Package(s): | puppet |
CVE #(s): | CVE-2012-1053
CVE-2012-1054
|
| Created: | February 23, 2012 |
Updated: | July 4, 2012 |
| Description: |
From the Ubuntu advisory:
It was discovered that Puppet did not drop privileges when executing
commands as different users. If an attacker had control of the execution
manifests or the executed command, this could be used to execute code with
elevated group permissions (typically root). (CVE-2012-1053)
It was discovered that Puppet unsafely opened files when the k5login type
is used to manage files. A local attacker could exploit this to overwrite
arbitrary files and escalate privileges. (CVE-2012-1054) |
| Alerts: |
|
Comments (none posted)
python-httplib2: information disclosure
| Package(s): | python-httplib2 |
CVE #(s): | |
| Created: | February 27, 2012 |
Updated: | February 29, 2012 |
| Description: |
From the Ubuntu advisory:
The httplib2 Python library earlier than version 0.7.0 did not perform any
server certificate validation when using HTTPS connections. If a remote
attacker were able to perform a man-in-the-middle attack, this flaw could
be exploited to alter or compromise confidential information in
applications that used the httplib2 library. |
| Alerts: |
|
Comments (none posted)
samba: remote code execution
| Package(s): | samba |
CVE #(s): | CVE-2012-0870
|
| Created: | February 24, 2012 |
Updated: | March 12, 2012 |
| Description: |
From the Red Hat advisory:
An input validation flaw was found in the way Samba handled Any Batched
(AndX) requests. A remote, unauthenticated attacker could send a
specially-crafted SMB packet to the Samba server, possibly resulting in
arbitrary code execution with the privileges of the Samba server (root). |
| Alerts: |
|
Comments (none posted)
systemd: arbitrary file creation
| Package(s): | systemd |
CVE #(s): | CVE-2012-0871
|
| Created: | February 29, 2012 |
Updated: | March 12, 2012 |
| Description: |
The systemd-logind process creates files under /run/user in an insecure manner, allowing a local attacker to create symbolic links in arbitrary locations. |
| Alerts: |
|
Comments (none posted)
systemtap: denial of service
| Package(s): | systemtap |
CVE #(s): | CVE-2012-0875
|
| Created: | February 27, 2012 |
Updated: | March 18, 2013 |
| Description: |
From the Red Hat bugzilla:
A flaw was discovered in how systemtap handled DWARF expressions when
unwinding the stack. This could result in an invalid pointer read, leading to reading kernel memory, or a kernel panic (and if the kernel reboot on panic flag was set (panic_on_oops), it would cause the system to reboot).
In order to trigger this flaw, an admin would have to enable unprivileged mode (giving users membership in the 'stapusr' group and configuring the local machine with 'signer,all-users' stap-server trust). If an admin has enabled unprivileged mode, a user with such access could use this to crash the local machine. |
| Alerts: |
|
Comments (none posted)
webcalendar: cross-site scripting
| Package(s): | WebCalendar |
CVE #(s): | CVE-2012-0846
|
| Created: | February 28, 2012 |
Updated: | February 29, 2012 |
| Description: |
From the Red Hat bugzilla:
It was reported that WebCalendar suffers from a stored XSS flaw in the
location variable. |
| Alerts: |
|
Comments (none posted)
Page editor: Jake Edge
Next page: Kernel development>>