![]() | |||||||||||
Messages in this thread |
On Tue, Oct 18, 2005 at 02:22:41AM -0500, Dmitry Torokhov wrote: > On Tuesday 18 October 2005 02:17, Greg KH wrote: > > > > Because before my patch, any class_device created for the input class, > > had the name, phys, and uniq attributes created for them, including the > > "simple" class device structures event0, event1, and so on. The kobject > > being passed back to those callback functions was not of the same type > > of object as input0, input1 and so on. So bad things happened. > > Oh, I see. That's what you get for mixing devices of different classes > into one class ;) I agree, it's tough, nasty, and full of horrible issues like this. I got away with much the same mess with the USB core code where we mix devices of different types on the same bus, but I wouldn't recommend ever doing that again. So, when we move everything over to being only devices, I'll try to make problems like this impossible to occur, as it's a pain... thanks, greg k-h - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ | ||||||||||
| Last update: 2005-10-18 09:43 [from the cache] ©2003-2008 | |||||||||||