Linux – How to secure Linux systems against the BlueBorne remote attack

A bluetooth connection has significant latency compared to simple wired headphones or speakers. What's more, the connection latency can vary, depending on the properties of the bluetooth receiver and maybe even radio signal strength as the user moves around.

The interface between an application and PulseAudio can be as simple as "here's some PCM audio data; play this." But it can also be more complicated; something like "Here's some PCM audio data; play this and tell me every 50 ms how far you've got, so that I can tell you to skip ahead if it looks like you're falling out of lip-synch with the video stream I'm playing. Oh, and you'll need to resample it too, since the data has a sample rate your hardware won't directly support." In the latter case, PulseAudio needs to be able to give the application some feedback from the audio device to correctly determine how far the audio data is actually been played at any given time.

As a result, it makes sense for PulseAudio to be fairly deeply involved in Bluetooth audio processing: the more intervening layers there are, the more possibilities for data to be buffered without maintaining accurate feedback, causing lip-synch to be lost.

In fact, before PulseAudio existed, there used to be an ALSA backend for Bluetooth audio, but it was deprecated. I think the problem was that ALSA's interfaces at that time were designed mainly for traditional sound cards, and dealing with a potentially variable audio latency of Bluetooth was difficult.

PulseAudio's interfaces were designed from the ground up to handle various sound devices and even switching audio streams between them while the stream is playing, so it seems to me it has a pretty advanced concept of audio latency built-in too.

Yes, it could have been implemented in BlueZ rather than as a PulseAudio module; but then, BlueZ would have had to present an audio interface for the applications. And since PulseAudio wants to handle "all" the audio on a system (in order to be able to transfer the audio you've currently playing from speakers to Bluetooth or vice versa on-the-fly), it would have to interface with PulseAudio somehow anyway.

This Samba new vulnerability is already being called "Sambacry", while the exploit itself mentions "Eternal Red Samba", announced in twitter (sensationally) as:

Samba bug, the metasploit one-liner to trigger is just: simple.create_pipe("/path/to/target.so")

Potentially affected Samba versions are from Samba 3.5.0 to 4.5.4/4.5.10/4.4.14.

If your Samba installation meets the configurations described bellow, the fix/upgrade should be done ASAP as there are already exploits, other exploit in python and metasploit modules out there.

More interestingly enough, there are already add-ons to a know honeypot from the honeynet project, dionaea both to WannaCry and SambaCry plug-ins.

Samba cry seems to be already being (ab)used to install more crypto-miners "EternalMiner" or double down as a malware dropper in the future.

honeypots set up by the team of researchers from Kaspersky Lab have captured a malware campaign that is exploiting SambaCry vulnerability to infect Linux computers with cryptocurrency mining software. Another security researcher, Omri Ben Bassat‏, independently discovered the same campaign and named it "EternalMiner."

The advised workaround for systems with Samba installed (which also is present in the CVE notice) before updating it, is adding to smb.conf:

nt pipe support = no

(and restarting the Samba service)

This is supposed to disable a setting that turns on/off the ability to make anonymous connections to the windows IPC named pipes service. From man samba:

This global option is used by developers to allow or disallow Windows NT/2000/XP clients the ability to make connections to NT-specific SMB IPC$ pipes. As a user, you should never need to override the default.

However from our internal experience, it seems the fix is not compatible with older? Windows versions ( at least some? Windows 7 clients seem to not work with the nt pipe support = no), and as such the remediation route can go in extreme cases into installing or even compiling Samba.

More specifically, this fix disable shares listing from Windows clients, and if applied they have to manually specify the full path of the share to be able to use it.

Other known workaround is to make sure Samba shares are mounted with the noexec option. This will prevent the execution of binaries residing on the mounted filesystem.

The official security source code patch is here from the samba.org security page.

Debian already pushed yesterday (24/5) an update out the door, and the corresponding security notice DSA-3860-1 samba

To verify in if the vulnerability is corrected in Centos/RHEL/Fedora and derivates, do:

#rpm -q –changelog samba | grep -i CVE
– resolves: #1450782 – Fix CVE-2017-7494
– resolves: #1405356 – CVE-2016-2125 CVE-2016-2126
– related: #1322687 – Update CVE patchset

There is now an nmap detection script :samba-vuln-cve-2017-7494.nse for detecting Samba versions, or a much better nmap script that checks if the service is vulnerable at http://seclists.org/nmap-dev/2017/q2/att-110/samba-vuln-cve-2017-7494.nse , copy it to /usr/share/nmap/scripts and then update the nmap database , or run it as follows:

nmap --script /path/to/samba-vuln-cve-2017-7494.nse -p 445 <target>

About long term measures to protect the SAMBA service: The SMB protocol should never be offered directly to the Internet at large.

It goes also without saying that SMB has always been a convoluted protocol, and that these kind of services ought to be firewalled and restricted to the internal networks [to which they are being served].

When remote access is needed, either to home or specially to corporate networks, those accesses should be better done using VPN technology.

As usual, on this situations the Unix principle of only installing and activating the minimum services required does pay off.

Taken from the exploit itself:

Eternal Red Samba Exploit -- CVE-2017-7494.
Causes vulnerable Samba server to load a shared library in root context.
Credentials are not required if the server has a guest account.
For remote exploit you must have write permissions to at least one share.
Eternal Red will scan the Samba server for shares it can write to. It will also determine the fullpath of the remote share.

    For local exploit provide the full path to your shared library to load.  

    Your shared library should look something like this

    extern bool change_to_root_user(void);
    int samba_init_module(void)
    {
        change_to_root_user();
        /* Do what thou wilt */
    }

It is also known systems with SELinux enabled are not vulnerable to the exploit.

See 7-Year-Old Samba Flaw Lets Hackers Access Thousands of Linux PCs Remotely

According to the Shodan computer search engine, more than 485,000 Samba-enabled computers exposed port 445 on the Internet, and according to researchers at Rapid7, more than 104,000 internet-exposed endpoints appeared to be running vulnerable versions of Samba, out of which 92,000 are running unsupported versions of Samba.

Since Samba is the SMB protocol implemented on Linux and UNIX systems, so some experts are saying it is "Linux version of EternalBlue," used by the WannaCry ransomware.

...or should I say SambaCry?

Keeping in mind the number of vulnerable systems and ease of exploiting this vulnerability, the Samba flaw could be exploited at large scale with wormable capabilities.

Home networks with network-attached storage (NAS) devices [that also run Linux] could also be vulnerable to this flaw.

See also A wormable code-execution bug has lurked in Samba for 7 years. Patch now!

The seven-year-old flaw, indexed as CVE-2017-7494, can be reliably exploited with just one line of code to execute malicious code, as long as a few conditions are met. Those requirements include vulnerable computers that:

(a) make file- and printer-sharing port 445 reachable on the Internet,
(b) configure shared files to have write privileges, and
(c) use known or guessable server paths for those files.

When those conditions are satisfied, remote attackers can upload any code of their choosing and cause the server to execute it, possibly with unfettered root privileges, depending on the vulnerable platform.

Given the ease and reliability of exploits, this hole is worth plugging as soon as possible. It's likely only a matter of time until attackers begin actively targeting it.

Also Rapid 7 - Patching CVE-2017-7494 in Samba: It’s the Circle of Life

And more SambaCry: The Linux Sequel to WannaCry.

Need-to-Know Facts

CVE-2017-7494 has a CVSS Score of 7.5 (CVSS:3.0/AV:N/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H)3.

Threat Scope

A shodan.io query of "port:445 !os:windows" shows approximately one million non-Windows hosts that have tcp/445 open to the Internet, more than half of which exist in the United Arab Emirates (36%) and the U.S. (16%). While many of these may be running patched versions, have SELinux protections, or otherwise don't match the necessary criteria for running the exploit, the possible attack surface for this vulnerability is large.

P.S. The commit fix in the SAMBA github project appear to be commit 02a76d86db0cbe79fcaf1a500630e24d961fa149