Re: [RFC PATCH 1/4] PLATFORM: introduce structure to bind async platform data to a dev path name

["Rafael J. Wysocki" <rjw@sisk.pl>]

Hi,

On Saturday, March 12, 2011, Andy Green wrote:
> This structure allows tagging arbitrary platform_data that can't be attached
> to a device until after it is probed, with the device path name that it is
> to be attached to.
> 
> Signed-off-by: Andy Green <andy.green@linaro.org>
> ---
> 
>  include/linux/platform_device.h |   17 +++++++++++++++++
>  1 files changed, 17 insertions(+), 0 deletions(-)
> 
> diff --git a/include/linux/platform_device.h b/include/linux/platform_device.h
> index 2e700ec..d8c0ba9 100644
> --- a/include/linux/platform_device.h
> +++ b/include/linux/platform_device.h
> @@ -193,4 +193,21 @@ static inline char *early_platform_driver_setup_func(void)		\
>  }
>  #endif /* MODULE */
>  
> +/**
> + * platform_async_platform_data - maps a known bus + device name on to
> + *				  platform_data to be attached to that device
> + *				  when it is eventually instantiated.  For use
> + *				  with onboard devices on buses that probe
> + *				  asynchronously.  Device path fields must
> + *				  be separated with '/'.
> + * @device_path:	bus / device path, eg, "usb1/1-1/1-1.1"
> + * @platform_data:	platform_data to attach to device matching the
> + *			device_path
> + */
> +
> +struct platform_async_platform_data {
> +	const char *device_path;
> +	void *platform_data;
> +};
> +
>  #endif /* _PLATFORM_DEVICE_H_ */

Using device paths for this purpose seems to be very fragile to me.  Isn't
there any better solution?

Rafael