Comment 70 for bug 870326

I was finally able to track down the lockdep issue I mentioned in comment #66. It was a bug with how the VFS was using lockdep. Since this was a lockdep false positive, it allows for the fix to be very clean and simple. The patch I'm attaching is against 3.2-rc1.

I've built an Oneiric test kernel, based upon 3.0.0-12.20, which contains this fix and a fix for bug #813146. Here are the links:
http://people.canonical.com/~tyhicks/ecryptfs/linux-headers-3.0.0-12-generic_3.0.0-12.20+ecryptfs1_amd64.deb
http://people.canonical.com/~tyhicks/ecryptfs/linux-image-3.0.0-12-generic_3.0.0-12.20+ecryptfs1_amd64.deb