Ok, now I'm sending this here since linux-usb-devel and Pavel seem to be ignoring my emails. Note, this is against 2.4.21-rc1, but I am prepared to finish this off for 2.5 aswell. I just need someone who knows USB to look this over and give me some freaking feedback. The multi-input patch seems to be of significant importance. --------------------------------------------------------------------- I've been working on getting my recently purchased Happ Controls UGCI Fighting interface working. Aside from getting it working with the standard input layer (it implements a mouse, 2 joysticks and a keyboard), it also has some specialized inputs/outputs that I am programming from userspace using the event layer. The first problem I had with the initial HID was that the hid-input layer only has per-interface granularity with the input layer. This device has only two interfaces. The second is the mouse, the first is 7 other inputs including the 2 joysticks and the keyboard. The hid-input layer would lump all the fields in the first interface (7 axis and 13 buttons in all) to one input. So I got one joystick device and that was it. I did up a patch that created an input for each input in the interfaces listing. IMO, this is the correct way since each input descibes the fields available in it, and is a logical grouping. Now, the second problem is that my joysticks look like this: Field(0) Physical(GenericDesktop.Pointer) Usage(2) GenericDesktop.X GenericDesktop.Y Logical Minimum(-128) Logical Maximum(127) Physical Minimum(0) Physical Maximum(255) Report Size(8) Report Count(2) Report Offset(0) Flags( Variable Absolute ) Notice the logical/physical min/max are different, even though they are the same range. As far as I can tell this is within HID specs. Windows uses the joystick just fine. However Linux's HID and HID-input layer both use the logic min/max in this case, while the device is sending values in the physical min/max range. This is digital, so we get 0,128,255 from the device, but HID interprets that as 0,-128,-1. Obviously that doesn't work right. I'm not sure what the correct thing to do is, so my patch for this is only RFC status. With this patch my mouse and keyboard still work, and the joysticks work correctly. Maybe the correct thing is to map the physical values read from the device to the logical values. -- Debian - http://www.debian.org/ Linux 1394 - http://www.linux1394.org/ Subversion - http://subversion.tigris.org/ Deqo - http://www.deqo.com/