On 9/12/05, Greg KH <gregkh at suse.de> wrote:
> +static void hdaps_mousedev_enable(void)
> +{
> + down(&hdaps_sem);
> +
> + /* calibrate the device before enabling */
> + hdaps_calibrate();
> +
> + /* initialize the input class */
> + init_input_dev(&hdaps_idev);
> + hdaps_idev.dev = &pdev->dev;
> + hdaps_idev.evbit[0] = BIT(EV_KEY) | BIT(EV_REL);
> + hdaps_idev.relbit[0] = BIT(REL_X) | BIT(REL_Y);
> + hdaps_idev.keybit[LONG(BTN_LEFT)] = BIT(BTN_LEFT);
> + input_register_device(&hdaps_idev);
> +
I do not see the device ever reporting BTN_LEFT, why is its presence advertised?
> +
> +/*
> + * hdaps_mousedev_disable - disable the input class device. Caller must hold
> + * hdaps_sem.
> + */
> +static void hdaps_mousedev_disable(void)
> +{
> + down(&hdaps_sem);
Bad comment. The function itself holds the semaphore.
> +
> + if (hdaps_mousedev)
> + hdaps_mousedev_enable();
> +
Please consider converting the input device to an absolute device
(reporting ABS_X/ABS_Y). This way it will be bound to a joystick
interface (/dev/input/jsX) and won't mess your /dev/input/mice so you
won't have to fiddle with enabling/disabling and will be able to have
the device turned on unconditionally.
--
Dmitry
