Re: [PATCH 1/7] device: introduce resource structure to simplify resource declaration

[Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>]

> > +#endif /* __ASSEMBLY__ */
> > +#endif	/* _LINUX_IOPORT_H */
> > diff --git a/lib/driver.c b/lib/driver.c
> > index 4c10a49..ee5e850 100644
> > --- a/lib/driver.c
> > +++ b/lib/driver.c
> > @@ -103,6 +103,15 @@ int register_device(struct device_d *new_device)
> >  {
> >  	struct driver_d *drv;
> >  
> > +	/* if no map_base available use the first resource if available
> > +	 * so we do not need to duplicate it
> > +	 * Temporary fixup until we get rid of map_base and size
> > +	 */
> > +	if (new_device->map_base == 0 && new_device->resource) {
> > +		new_device->map_base = new_device->resource[0].start;
> > +		new_device->size = new_device->resource[0].size;
> > +	}
> 
> I think we don't need this. What we need though is the other way round:
> 
> 	if (new_device->map_base) {
> 		dev_warn(new_device, "uses map_base. Please convert to use resources\n")
> 		new_device->resource = xzalloc(sizeof(struct resource));
> 		new_device->resource[0].start = new_device->map_base;
> 		new_device->resource[0].size = new_device->size;
> 	}
> 
> This way all devices have resources, we can convert the drivers and
> devices to use resources one by one.
ok for the warning

but if we do not keep the map_base uptodate the core device drivers will be
broken so if no map_base is specified we must update it with the first
resource

how about this

	if (new_device->map_base) {
		if (new_device->resource) {
			dev_err(new_device, "map_base and resource specifed\n");
			return -EIO;
		}
		dev_warn(new_device, "uses map_base. Please convert to use resources\n")
		new_device->resource = xzalloc(sizeof(struct resource));
		new_device->resource[0].start = new_device->map_base;
		new_device->resource[0].size = new_device->size;
		new_device->numresources = 1;
	} else if (new_device->resource) {
		new_device->map_base = new_device->resource[0].start;
		new_device->size = new_device->resource[0].size;
	}

Best Regards,
J.

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox