WPA2 Security Flaw Puts Almost Every Wi-Fi Device at Risk of Hijack, Eavesdropping (zdnet.com)

A security protocol at the heart of most modern Wi-Fi devices, including computers, phones, and routers, has been broken, putting almost every wireless-enabled device at risk of attack. From a report: The bug, known as "KRACK" for Key Reinstallation Attack, exposes a fundamental flaw in WPA2, a common protocol used in securing most modern wireless networks. Mathy Vanhoef, a computer security academic, who found the flaw, said the weakness lies in the protocol's four-way handshake, which securely allows new devices with a pre-shared password to join the network. That weakness can, at its worst, allow an attacker to decrypt network traffic from a WPA2-enabled device, hijack connections, and inject content into the traffic stream. In other words: hackers can eavesdrop on your network traffic. The bug represents a complete breakdown of the WPA2 protocol, for both personal and enterprise devices -- putting every supported device at risk. "If your device supports Wi-Fi, it is most likely affected," said Vanhoef, on his website. News of the vulnerability was later confirmed on Monday by US Homeland Security's cyber-emergency unit US-CERT, which about two months ago had confidentially warned vendors and experts of the bug, ZDNet has learned.

Comments

Re:How serious is this? How exploitable is it?

[Racemaniac]

it's an attack on the client machine, not on the access point.
it means that any unpatched device connected to a wifi network with WPA2 encryption can expect the same security as being connected to an unsecured wifi (anyone can eavesdrop your communication, or could inject stuff, ...).
how easy it is to do, don't know yet, but it seems to be a feasible attack to carry out.

Re:What the fuck is Google going to do about Andro

[Anonymous Coward]

If this problem can be avoided with a software fix, I think that Google should do everything they possibly can to get this fix to as many Android devices as possible.

Google can't do anything about that.

It's the fucking telcos who are withholding updates from the end users. Even if you have the patched version on your hard drive, you can't install it, because your wireless provider won't let you. Verizon is the most egregious offender; as long as they continue to refuse to sell devices with unlocked bootloaders, the only way to install an update is when the telco feels like pushing it to the users.

Re:Finally!

[Solandri]

From the paper and blog:

In practice, some complications arise when executing the attack. First, not all Wi-Fi clients properly implement the state machine. In particular, Windows and iOS do not accept retransmissions of message 3 (see Table 1 column 2). This violates the 802.11 standard. As a result, these implementations are not vulnerable to our key reinstallation attack against the 4-way handshake. Unfortunately, from a defenders perspective, both iOS and Windows are still vulnerable to our attack against the group key handshake

So basically, Windows and iOS were protected for implementing 802.11 incorrectly.

Our attack is especially catastrophic against version 2.4 and above of wpa_supplicant, a Wi-Fi client commonly used on Linux. Here, the client will install an all-zero encryption key instead of reinstalling the real key. This vulnerability appears to be caused by a remark in the Wi-Fi standard that suggests to clear the encryption key from memory once it has been installed for the first time. When the client now receives a retransmitted message 3 of the 4-way handshake, it will reinstall the now-cleared encryption key, effectively installing an all-zero key. Because Android uses wpa_supplicant, Android 6.0 and above also contains this vulnerability. This makes it trivial to intercept and manipulate traffic sent by these Linux and Android devices.

While Android got screwed over by implementing it rigorously.

This should also become a programming example of the difference between setting something to NULL vs setting it to zero. Instead of implementing the encryption key as a string, it shouldn've been implemented as a pointer to the string. And when the standard called for the key to be cleared, the value shouldn've been zeroed out (to prevent it from being recovered in memory), memory released, and the pointer set to NULL (so the software would know the value didn't exist anymore and wouldn't try to use it).