Yeah, I removed the hack. As far as I understood, these two commits should properly fix the volume ID issue brought in by fuse wrapping:
The thing is though that CM has implemented support for ext4/exfat/ntfs/whatever sdcards (and some more changes) and I've a very strong hunch some of that is conflicting with these upstream changes. I planned to try to drop all that extra stuff and rebuild MediaProvider+vold from AOSP with only non-related commits from CM and see what happens. Evervolv also dropped the hack and haven't gone back to it, at least not yet, so I assume upstream has really fixed it.
P.S. It's so cool to see your browser again! :P hahahaha!