-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 10 Oct 2002 07:26, Douglas Roberts wrote:
> Brad Hards wrote:
> >>Loading these modules caused the directory hid to be created in
> >>/dev/usb, and the device file /dev/usb/hid/hiddev0 to be created (with
> >>the correct major & minor)
> >>
> >>lilly:[/dev/usb/hid]# ls -l
> >>total 0
> >>crw-rw-rw-    1 root     root     180,  96 Dec 31  1969 hiddev0
> >
> >So this device was not taken by the HID input code (either because you
> > didn't enable it in the kernel, or because the radio has the wrong
> > Application).
>
> Do you mean the input core support?  If so, that's turned on.
In later kernels, there is a config option for HID -> input layer. It is 
CONFIG_USB_HIDINPUT.

> The radio identifies itself as an hid device, so I don't understand what
> you meant by "wrong application", above.
You're going to have to read the HID spec some more times, before you'll be 
able to write a userspace app :-)
In the HID spec, Application (I'm using caps to denote a defined term) is a 
specific value returned by the HID interface. It is used to determine things 
like the usage pages. Look in the hid code (hid.h) and find the test for 
IS_INPUT_APPLICATION. If the Application(s) reported by the HID device 
matches, then the device is considered an input device, and gets associated 
with the hid-input code. Otherwise it goes to hiddev.

The dump-events code I gave a link to in my previous email can tell you what 
the Applications are for your device.

> >>3.  I ran the evtest code
> >
> >That won't help. You need something that understands hiddev events, not
> > input events.
>
> I thought that having evdev already loaded when the hid autoloaded would
> cause  input registration to occur via a call to event_register_device.
The HID code has two interfaces. I tried to draw it in 
Documentation/usb/hiddev.txt (in your kernel tree). One interface is to the 
input layer (where hid-input.c acts as the driver to the input core) and the 
other interface is to a char device, which has fairly raw HID.

You can use either interface, as long as you're willing to patch the kernel to 
override the IS_INPUT_APPLICATION test. That's not hard.

Brad
- -- 
http://linux.conf.au. 22-25Jan2003. Perth, Aust. I'm registered. Are you?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE9pKXNW6pHgIdAuOMRArglAJ4l6Uyei5RkuXxjsAlLE2GS5hjwOwCfQocT
3TtBaNLa1ViCnyvUmXu0BsM=
=P52x
-----END PGP SIGNATURE-----



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to