Re: 2.4.19, USB_HID only works compiled in, not as module

[Greg KH <greg@kroah.com>]

On Tue, Aug 06, 2002 at 12:00:55AM +1000, Brad Hards wrote:
> 
> Greg: I think this was one of your patches, associated with the 
> HIDINPUT patch. It looks like the return value is wrong. See 
> below.
> 
> --- include/linux/hiddev.h.orig Mon Aug  5 23:19:54 2002
> +++ include/linux/hiddev.h      Mon Aug  5 23:56:34 2002
> @@ -183,7 +183,7 @@
>  int __init hiddev_init(void);
>  void __exit hiddev_exit(void);
>  #else
> -static inline void *hiddev_connect(struct hid_device *hid) { return NULL; }
> +static inline void *hiddev_connect(struct hid_device *hid) { return -1; }
>  static inline void hiddev_disconnect(struct hid_device *hid) { }
>  static inline void hiddev_hid_event(struct hid_device *hid, unsigned int usage, int value) { }
>  static inline int hiddev_init(void) { return 0; }

??? Why return -1 as a void *?

The only caller of hiddev_connect is:
	if (!hiddev_connect(hid))
		hid->claimed |= HID_CLAIMED_HIDDEV;

Hm, seems like you don't want a void * there at all, but a int, right?
And as the "original" hiddev_connect returns an int, this does look like
a bug on my part, sorry.  I'll go fix it.

But that doesn't explain the error people are having with the code
compiled in.

thanks,

greg k-h