On Thu, May 29, 2014 at 02:23:59PM -0700, Sudeep Dutt wrote:
> +/**
> + * mbus_device - representation of a device using mbus
> + * @priv: private pointer for the driver's use.
> + * @mmio_va: virtual address of mmio space
> + * @hw_ops: the hardware ops supported by this device.
> + * @id: the device type identification (used to match it with a driver).
> + * @dev: underlying device.
> + * be used to communicate with.
> + * @index: unique position on the mbus bus
> + */
> +struct mbus_device {
> + void *priv;
What's wrong with the built-in "private" pointer that struct device has?
That is why it is there...
> + void __iomem *mmio_va;
> + struct mbus_hw_ops *hw_ops;
> + struct mbus_device_id id;
> + struct device dev;
> + int index;
> +};
On Wed, 2014-07-09 at 17:42 -0700, Greg Kroah-Hartman wrote:
> On Thu, May 29, 2014 at 02:23:59PM -0700, Sudeep Dutt wrote:
> > +/**
> > + * mbus_device - representation of a device using mbus
> > + * @priv: private pointer for the driver's use.
> > + * @mmio_va: virtual address of mmio space
> > + * @hw_ops: the hardware ops supported by this device.
> > + * @id: the device type identification (used to match it with a driver).
> > + * @dev: underlying device.
> > + * be used to communicate with.
> > + * @index: unique position on the mbus bus
> > + */
> > +struct mbus_device {
> > + void *priv;
>
> What's wrong with the built-in "private" pointer that struct device has?
> That is why it is there...
>
We were using both mbus_device priv and the device driver_data fields
unnecessarily. I have a patch which gets rid of the mbus_device priv
field and uses only the device driver_data field via
dev_set/get_drvdata(..). The next revision of the patch series will have
this cleaned up.
Thanks for the review!
Sudeep Dutt
> > + void __iomem *mmio_va;
> > + struct mbus_hw_ops *hw_ops;
> > + struct mbus_device_id id;
> > + struct device dev;
> > + int index;
> > +};
>