Re: [PATCH v4 2/6] GPIO: pca953x.c: Set device platform_data pointerafter allocating it

[Grant Likely]

On Wed, Jun 08, 2011 at 02:48:30PM +0200, David Jander wrote:
> In the case that we obtain device-tree data to fill in platform_data,
> the dev.platform_data pointer needs to be set, since it is used by
> functions such as pca953x_irq_setup() later on.
> 
> Signed-off-by: David Jander <david@xxxxxxxxxxx>
> ---
>  drivers/gpio/pca953x.c |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/gpio/pca953x.c b/drivers/gpio/pca953x.c
> index b31f881..2dff562 100644
> --- a/drivers/gpio/pca953x.c
> +++ b/drivers/gpio/pca953x.c
> @@ -660,6 +660,7 @@ static int __devinit pca953x_probe(struct i2c_client *client,
>  		 * dynamically and must be freed in the driver
>  		 */
>  		chip->dyn_pdata = pdata;
> +		client->dev.platform_data = pdata;

No, this is illegal.  The dev->platform_data pointer must be
considered immutable by the device driver.  Otherwise you risk
breaking driver unbind/rebind use case.

Instead, the driver should be fixed so that it doesn't look at the
client->dev.platform_data pointer after .probe() has returned.  You
can do this by copying the relevant data into the driver's private
data structure.

g.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/