Re: [RFC PATCH 02/10] vfio: Update vfio_add_group_dev() API

[Jason Gunthorpe]

On Mon, Feb 22, 2021 at 09:50:47AM -0700, Alex Williamson wrote:
> diff --git a/drivers/vfio/vfio.c b/drivers/vfio/vfio.c
> index 464caef97aff..067cd843961c 100644
> +++ b/drivers/vfio/vfio.c
> @@ -848,8 +848,9 @@ static int vfio_iommu_group_notifier(struct notifier_block *nb,
>  /**
>   * VFIO driver API
>   */
> -int vfio_add_group_dev(struct device *dev,
> -		       const struct vfio_device_ops *ops, void *device_data)
> +struct vfio_device *vfio_add_group_dev(struct device *dev,
> +				       const struct vfio_device_ops *ops,
> +				       void *device_data)
>  {
>  	struct iommu_group *iommu_group;
>  	struct vfio_group *group;
> @@ -857,14 +858,14 @@ int vfio_add_group_dev(struct device *dev,
>  
>  	iommu_group = iommu_group_get(dev);
>  	if (!iommu_group)
> -		return -EINVAL;
> +		return ERR_PTR(-EINVAL);
>  
>  	group = vfio_group_get_from_iommu(iommu_group);
>  	if (!group) {
>  		group = vfio_create_group(iommu_group);
>  		if (IS_ERR(group)) {
>  			iommu_group_put(iommu_group);
> -			return PTR_ERR(group);
> +			return (struct vfio_device *)group;

Use ERR_CAST() here

Also, I've wrote a small series last week that goes further than this,
I made 'struct vfio_device *' the universal handle to refer to the
device, instead of using 'void *' or 'struct device *' as a surrogate.

It is interesting that you hit on the same issue as a blocker to this
series. So for I've found quite a few other things that are out of
sorts because of this.

Cheers,
Jason