On Mon, 2007-04-02 at 00:04 -0400, Dmitry Torokhov wrote: > Actually I want to keep input subystm out of the loop here, since LEDs > such as mail, charging, etc have nothing to do with user input but > rather reflect overall system/application state. >
What if I just added a HID_QUIRK_HIDDEV for the vendor/product id's in the blacklist table in hiddev.c. That would force the creation of a hiddev device, through which a user-space program could access the HID usages directly. It would be awkward, though, since there would also be an input event device, but one through which the full functionality is not accessible. Is there a HID_QUIRK that would force creation of a hiddev *instead of* (rather than in addition to) an event dev? Two things to keep in mind: 1) The device in question (it's a Belkin USB Flip KVM switch, just to restate from the OP) has nothing to do with LED's. It just overloads the LED HID usage codes to receive control from the PC. 2) No matter what we do in terms of presenting this device to user-land, under the hood it's a HID-class USB device, and it probably doesn't make sense to not make use of that. Thanks, -Dan
signature.asc
Description: This is a digitally signed message part