Another point that supports Apple's assertion is that according to the "Sex, Drugs & Unix" blog entry that Glenn linked to a while back, an unassociated station could only perpetrate this attack if the victim has drivers that were specially crafted. If that's the case, then this really isn't a significant flaw at all beyond a traditional peer-to-peer attack over a Wi-Fi network.

I don't see how Apple can say this isn't their problem. Unless they think that every Apple user uses all Apple equipment.

At Defcon they showed this as well; and they made it clear that they were using 3rd party equipment; however, they did hint that they could have done this with Apple's stuff.

[Editor's note: Tricky point here. I want to agree with you that Apple should have responsibility for an operating system not being able to be compromised by a faulty driver. But the Wi-Fi stuff is a very particular subcase. Apple ships Wi-Fi as a built-in feature with most of its computers now, and has done so for its laptops for over two years. It's unlikely that someone would use a third-party card or option with a newer Mac for Wi-Fi.

Further, Apple pursues a kind of openness that Microsoft does not in terms of drivers. Anyone can write and distribute hardware drivers for Mac OS X. Microsoft puts up big warnings for drivers that haven't gone through their testing and certification process, although there are apparently holes discovered in wireless drivers that are certified.

So you're in effect asking Apple to be responsible for any behavior on the part of third-party developers or to restrict drivers insertion through some paid labeling program.

Further, the researchers are no longer claiming that they have any Apple exploit. They said at one point that they'd been in touch with Apple; Apple denies that they received anything useful and the researchers haven't contradicted this.--gf]