Also mind HS is 40MB/s which means that if you would like to process that with 120MHz uC, you have to do it in 3 clocks per byte, average!! Thus Cortex M3 is not capable of any reasonalble processing of that stream.
Sure, but there are plenty of "average data rates" in between ~1MB/s ("full speed") and the ~40MB of "HS" that are sensible on a typical CM3...
Exactly. The 480 mbit/s does not have to be constantly piped into your micro, it is only the peak speed for one USB controller (!= endpoint). If you want to send 16mbit/sec, you need high speed. For example a 16 bit sample, every 100 clock cycles.
But the way I see it, small micros dont have Ethernet Phy and USB Phy built it, because they use low power semiconductor nodes, and the frequencies are large.