I normally work on FreeBSD, and have had no problems with boards up to the Due - largely because those who've gone before have already figured out the tool chain.
I was wondering if anyone else is working on programming a Due from FreeBSD, and might have done the work I'm avoiding?
My normal route for hardware testing doesn't work in this case, either. That's to run the tool chain on Linux running on a VirtualBox VM. The programming phase keeps failing. I suspect this is related to the new USB behavior, as I see the same problems with the Leonardo (though that works fine with my FreeBSD tool chain).