To add some items to that list:
- USB 2.0 support! I think we can probably have "basic" support
working as a clean layer over 2.4, but full support (including the
transaction translator support) will probably take a while to finish,
then it'll need a shake down cruise and performance tuning.
- Update usbdevfs. Files should be "per-interface", not "per-device",
to provide a model that's more consistent with PCI ("per-function").
Names should be "stable" and based on topology. (That's a much
less evolutionary approach than I think Greg was thinking of... the
new dev_t and hotplug work may provide a good context for this.)
- Some mechanism to arbitrate which driver gets to issue a control
transfer at any given moment. That issue comes up periodically
(it's not the fault of usbdevfs that "cat devices" can cause trouble).
- Clean up configuration changing support ... disconnect all current
drivers (their interfaces are going away!) then rebind, at least.
- Likely the power management work ("ACPI") will run into the
need for USB to do some power budgeting.
Sounds like some of that "common driver" framework work would
be a good thing for diagnostics and supportability too -- not just to get
better power management.
- Dave
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
