Am Samstag, 27. Juni 2009 16:42:31 schrieb Alan Stern: > On Sat, 27 Jun 2009, Oliver Neukum wrote: > > Am Samstag, 27. Juni 2009 03:31:51 schrieb Alan Stern: > > > Oliver, maybe you should patch usbhid to make it avoid autosuspending > > > mice while the device file is open. That should prevent a lot of > > > complaints. Maybe keyboards too, I don't know -- people don't seem to > > > be having as much trouble with them. > > > > To have that you simply revert to the 2.6.29 version. > > People who have just installed 2.6.30 generally don't feel like > reverting to 2.6.29. :-) I meant reverting the autosuspend changes of 2.6.30 > Nevertheless, my point was that such a large percentage of mice do > poorly at remote wakeup, we should avoid autosuspending mice entirely > -- unless remote wakeup doesn't matter, i.e., when the device file is > closed. That means don't autosuspend regardless of the timeout value. As you noted, that case is very rare. So the only practical options are those 2.6.30 provides for mice. You either don't suspend at all, or you do it with remote wakeup like 2.6.30 does. Making such choices was exactly the reason we left the choice of enabling and disabling autosuspend to user space. > Also, usbhid can afford to be selective. Even though mice behave > badly, we can continue to autosuspend other sorts of devices. That means we effectively reduce autosuspend for mice to a nop which can be had by switching off autosuspend for a device and penalize those devices that work well. Regards Oliver -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
