Re: cleanup of hiddev

[Jiri Slaby <jirislaby@gmail.com>]

On 11/04/2008 11:51 PM, Oliver Neukum wrote:
> If you call usb_register_dev() before you set
> hiddev_table[usbhid->intf->minor - HIDDEV_MINOR_BASE]
> this check in hiddev_open() may fail:
> 	if (i >= HIDDEV_MINORS || i < 0 || !hiddev_table[i])
> 		return -ENODEV;

I think, that's the point. You should also get ENODEV if you don't call
usb_register_dev(), create a node for non-existent minor manually and try to
open it. It behaves the same as if there was a lock you suggested...

Ex.:
# grep 137 /proc/devices
# mknod node c 137 0
# cat node
cat: node: No such device or address