Am Donnerstag, 26. Januar 2012, 15:56:28 schrieb Alan Stern: > On Thu, 26 Jan 2012, Oliver Neukum wrote: > > > Second, it seems to me that in the long run, this is asking for trouble. > > IMHO the function should be represented in the device tree between > > the device and the interfaces. We have the same constraints as for > > interfaces and devices. A function's interfaces cannot do PM independently > > and suspending a device means suspending all its functions. > > This will require a significant amount of change to the entire USB > stack. There must be lots of places where we assume that the device is > the parent of the interface. (It wouldn't be surprising if userspace > makes that assumption as well.) Making it even more difficult is the > fact that USB-2 devices don't have functions whereas USB-3 devices do. > > > We'd get the handling of multifunction devices for free putting functions > > into the device tree. > > Yes, but adding functions to the driver model is far from free. I'm > not saying we shouldn't do it eventually, but it will be quite a big > job. True, but the alternative is not doing PM at the function level, isn't it? 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
