On 22-07-21, 16:52, Arnd Bergmann wrote:
> On Thu, Jul 22, 2021 at 11:56 AM Viresh Kumar <viresh.kumar@xxxxxxxxxx> wrote:
> >
> > +/* Virtio device compatibles and IDs */
> > +static const struct of_device_id of_virtio_devices[] = {
> > + { .compatible = "virtio,22", .data = (void *)VIRTIO_ID_I2C_ADAPTER },
> > + { .compatible = "virtio,29", .data = (void *)VIRTIO_ID_GPIO },
> > + { }
> > +};
> > +
> > +static int virtio_device_of_init(struct virtio_device *dev)
> > +{
> > + struct device_node *np, *pnode = dev->dev.parent->of_node;
> > + const struct of_device_id *match;
> > + int ret, count;
> > +
> > + if (!pnode)
> > + return 0;
> > +
> > + count = of_get_available_child_count(pnode);
> > + if (!count)
> > + return 0;
> > +
> > + /* There can be only 1 child node */
> > + if (WARN_ON(count > 1))
> > + return -EINVAL;
> > +
> > + np = of_get_next_available_child(pnode, NULL);
> > + if (WARN_ON(!np))
> > + return -ENODEV;
> > +
> > + match = of_match_node(of_virtio_devices, np);
> > + if (!match) {
> > + ret = -ENODEV;
> > + goto out;
> > + }
>
> I think it would be better not to have to enumerate the of_virtio_devices[]
> strings, but instead use of_device_is_compatible() to match against
> "virtio,%d". Otherwise we end up modifying this function for every
> virtio driver that needs a binding.
Yeah, will do that.
--
viresh
