On Wed, 18 Jul 2012, Bjørn Mork wrote: > Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> writes: > > > Any attempt to use a suspended device will return an error, whether the > > port is powered off or not. The driver has to resume the device before > > using it, and Sarah is proposing that this should cause the port to > > power back on. > > Assuming that "powering off" a port means removing power from any bus > powered device, won't this be something completely different from doing > a USB resume or reset_resume? Complex devices may choose to keep parts > of the system running while "suspended" in the USB bus context. Probably not anything significant, if they are bus powered. If they are self powered then they continue to keep various parts running even when bus power is turned off. > Removing power will prevent this if the device is bus powered (which I > believe you never know as there are too many devices lying about this > for various reasons). > > Even if a driver support reset_resume, that only means that the driver > knows how to recreate the device state after a USB bus reset. It does > *not* mean that the driver knows how to recreate the device state after a > cold boot of the device. That is in general not possible. The only way > is usually to reinitialize from scratch. I.e. disconnect. Oliver pointed this out a few emails back. It's worth mentioning, however, that drivers _do_ have to face this possibility when resuming from hibernation. And as Oliver said, sometimes this does indeed result in a disconnect. > Somehow I am pretty sure that nothing of this is news to any of you, so > I am now sitting here wondering what part of this discussion I've > completely misunderstood... Alan Stern -- 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
