Comment 24 for bug 904412

I haven't had a chance to test the upstream kernel yet, but I upgraded my workstation to Raring, and with 3.8.0-5-generic, I can now run my benchmark without the card-reader being powered off. However, it's slower than when I use the same card and card-reader plugged into a USB2 port.

With 3.8.0-4-generic, I was still experiencing the same problem with the card-reading powering off. So between the -4 and -5 revisions, something change that seemingly fixed the problem, aside from the performance issue.