Great - so the uhci fault is being cause by that frame limit, that can be fixed (and even documented). Although that hangup needs more investigation...
The ehci fault is odd - if it's triggered by just 64 frames, it might not be the same thing that I've experienced. If you type 'continue' in ddb after the nmi is triggered (has to be typed twice I believe), does it post another nmi or continue like nothing happened? Also, can you upload the device descriptors? (I think usbutil can dump them). -- Thanks, Jeremy
