Samba remote code execution exploit: what you need to know

This is going to hurt home users with Samba shares mounted on their SoHo routers or NAS, among other things. 

Samba is a file sharing service for Linux, similar to Windows SMB file shares (yes, the same SMB that was exploited in the recent WannaCry ransomware worm). A vulnerability in Samba could enable a similar attack on Linux systems. A malicious actor with access to upload files to a Samba share, can upload malicious code and then use this vulnerability to cause the server to execute it.

Unlike SMB, Samba exists on a wide variety of systems from different makers - servers, laptops, home routers, network storage systems, media servers, and many IoT devices. And unlike Windows, those devices may not automatically install an update - even if the manufacturer provides one. 

A quick query of Internet scanner Shodan shows that nearly a half million devices running Samba are publicly accessible on the Internet. Interestingly, the large majority of those appear to be in the United Arab Emirates, leading one to wonder if Emirates Telecommunications Corporation is equipping its customers with a gateway router that has Samba enabled by default.

What can you do?

Update Samba

The best course of action is to update Samba to a non-vulnerable version (4.6.4 or newer; 4.5.10 or newer; or 4.4.14 or newer, according to the Samba Project advisory).

For most IoT devices, you are likely dependent on the manufacturer to release a firmware update that includes this fix.

Disable writable shares

This vulnerability can only be exploited using shares that allow uploading or writing files; read-only shares cannot be exploited.

Disable "named pipe endpoints" in your Samba config file

Similar to the way port numbers allow multiple layer 4 connections to the same layer 3 network address, named pipes allow multiple layer 5 (SMB) connections to the same layer 4 port (TCP 445). This is also the feature that can be exploited due to this vulnerability. Disabling named pipes prevents exploitation, though it may also disable expected functionality in some cases.

To disable named pipes, add the parameter:

nt pipe support = no

to the [global] section of your smb.conf file and restart smbd. You can modify smb.conf on a couple of IoT devices as follows:

• ASUSWRT wireless routers
• Seagate GoFlex network storage devices

Double-check that Samba is not exposed to the Internet

• Browse to http://www.ipchicken.com/ to check your public Internet address
• Browse to https://shodan.io and search for your address. You do not want to see the following - if you do, you'll need to check your router or firewall and disable public (or WAN) access to port 445:

Hit by WannaCry? It may also be a HIPAA breach

Ransomware is a common form of malware, designed to encrypt personal and business data, making it unusable unless the victim pays a "ransom" fee to the attacker to purchase the recovery key. It most often affects one person at a time, delivered by email or a malicious web browser download. 

Beginning May 12 however, the "WannaCry" or "WannaCrypt" ransomware spread rapidly by exploiting a flaw in the Windows operating system -- a flaw patched by Microsoft in March, but that nonetheless remained exposed in many organizations that had not yet updated their systems.

Under guidance issued by the US Department of Health and Human Services (HHS) last summer:

The presence of ransomware (or any malware) on a covered entity’s or business associate’s computer systems is a security incident under the HIPAA Security Rule.

When electronic protected health information (ePHI) is encrypted as the result of a ransomware attack, a breach has occurred because the ePHI encrypted by the ransomware was acquired (i.e., unauthorized individuals have taken possession or control of the information), and thus is a “disclosure” not permitted under the HIPAA Privacy Rule.

Unless the covered entity or business associate can demonstrate that there is a “…low probability that the PHI has been compromised,” based on the factors set forth in the Breach Notification Rule, a breach of PHI is presumed to have occurred. The entity must then comply with the applicable breach notification provisions, including notification to affected individuals without unreasonable delay, to the Secretary of HHS, and to the media (for breaches affecting over 500 individuals) in accordance with HIPAA breach notification requirements. See 45 C.F.R. 164.400-414. 

The HHS ransomware fact sheet (PDF download) includes the following Q&A:

Is it a HIPAA breach if ransomware infects a covered entity’s or business associate’s computer system?

When electronic protected health information (ePHI) is encrypted as the result of a ransomware attack, a breach has occurred because the ePHI encrypted by the ransomware was acquired (i.e., unauthorized individuals have taken possession or control of the information), and thus is a “disclosure” not permitted under the HIPAA Privacy Rule. Unless the covered entity or business associate can demonstrate that there is a “…low probability that the PHI has been compromised,” based on the factors set forth in the Breach Notification Rule, a breach of PHI is presumed to have occurred. The entity must then comply with the applicable breach notification provisions, including notification...in accordance with HIPAA breach notification requirements.

How can covered entities or business associates demonstrate “…that there is a low probability that the PHI has been compromised” such that breach notification would not be required?

To demonstrate that there is a low probability that the protected health information (PHI) has been compromised because of a breach, a risk assessment considering at least the following four factors must be conducted: 

1. The nature and extent of the PHI involved, including the types of identifiers and the likelihood of re-identification
2. The unauthorized person who used the PHI or to whom the disclosure was made
3. Whether the PHI was actually acquired or viewed; and
4. The extent to which the risk to the PHI has been mitigated. 

A thorough and accurate evaluation of the evidence acquired and analyzed as a result of security incident response activities could help entities with the risk assessment process above by revealing, for example: the exact type and variant of malware discovered; the algorithmic steps undertaken by the malware; communications, including exfiltration attempts between the malware and attackers’ command and control servers; and whether or not the malware propagated to other systems, potentially affecting additional sources of electronic PHI (ePHI). Correctly identifying the malware involved can assist an entity to determine what algorithmic steps the malware is programmed to perform. Understanding what a particular strain of malware is programmed to do can help determine how or if a particular malware variant may laterally propagate throughout an entity’s enterprise, what types of data the malware is searching for, whether or not the malware may attempt to exfiltrate data, or whether or not the malware deposits hidden malicious software or exploits vulnerabilities to provide future unauthorized access, among other factors

Ransomware now comes in worm flavor

If you have SMBv1 in your enterprise, and haven't completed deploying MS17-010 (released in March), now would be a good time to expedite that. Multiple news outlets are reporting a widespread outbreak of the "WannaCry" ransomware. 

Ransomware is malware that encrypts all the data on a computer, holding it hostage until the victim pays a ransom fee. This particular attack is especially insidious because it acts as a "worm" - it spreads from computer to computer on its own, without any interaction from users.

The saving grace is that the vulnerability it exploits to spread, was fixed by Microsoft in March. Most home users are safe because Windows Updates apply automatically (yes, it's annoying to have a computer reboot when you do not want it to, but today you are thanking your lucky stars).

Some reports of note:

CCN-CERT, the computer emergency response team for Spain, first issued a warning (in Spanish) of this outbreak Friday morning.

Spanish telecommunications company Telefónica reported (in Spanish) that they too have been affected.

The British Broadcasting Company has a running commentary on effects in the UK, and specifically the effects on the National Heathcare Service of the UK.

The Register reports that UK hospitals have effectively shutdown, and are not accepting new patients.

Global delivery company FedEx reported that it has been affected, but has not specified what locations or if deliveries have been interrupted. At least one FedEx customer reported Customer Service being unable to provide support due to server outages.

What can you do:

Home users by and large are not affected by this. If you follow the basic steps I recommend in https://securityforrealpeople.com/cybertips (in particular, setting Windows to automatically install updates), Windows lomng ago installed the patch to protect you from this worm.

For corporate and small business readers:

• Block TCP 445 and 135 inbound from the Internet
   
• Install MS17-010 everywhere. Note that the April and May cumulative updates for Windows include this patch
   
• Kill off SMBv1. SMB version 1 is a 30-year-old protocol that has outlived its usefulness. Every modern operating system - including all supported Windows variants, MacOS and OS X, and the Samba product for Linux file sharing, supports the newer v2 and v3 versions.
  
  SMBv1 can be disabled by creating or editing the following value in the Windows Registry:
  
  
  
  HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\LanmanServer\Parameters
  Name: SMB1
  Type: DWORD
  Value: 0
  
  
  
  Then run the following command to disable SMBv1 on the client side:
  
  
  
  sc.exe config lanmanworkstation depend= bowser/mrxsmb20/nsi
  sc.exe config mrxsmb10 start= disabled
  
  
  
  
• Block client-to-client SMB (TCP 445) traffic. Generally speaking, laptops don't need to map file shares of other laptops. Blocing lateral SMB traffic prevents this malware from spreading laptop-to-laptop. Then focus on patching your domain controllers and enterprise file servers - which genuinely do need to share services on TCP 445.
   
• Run Windows Firewall and block inbound TCP 445 connections when on an untrusted network (public WiFi, for example).

Hacking the SIEM

Day 1 of Security B-Sides Austin is in the books. One talk in particular stuck with me: "Hack the SIEM" by John Griggs of Meta Studios, Inc.

Your SIEM is an aggregation of lots of data about your company - it contains information about endpoints, network controls, detective capabilities, and incidents. To an attacker, it is a gold mine of recon.

John brought up a different point, one I had not considered: your Security Information and Event Management system, or SIEM, may also be the single pane of glass that your SOC relies on. If an attacker doesn't show up in the SIEM, your SOC may not be aware of the incident - even if the originating network control is squawking at the top of its lungs.

Ergo, an attacker doesn't have to cover all of its tracks - they only need to stop their actions from showing up in the SIEM. Sure, original logs will show the attacker's trail in the post-mortem, but depending on their objectives, avoiding real-time detection may be all the attacker needs.

Is your SIEM locked down to prevent it from being used and abused by an attacker?