On Thu, Mar 03, 2016 at 10:53:01AM -0700, Jason Gunthorpe wrote: > > I looked at the naming and I think a rdma_device_* or rdma_dev_* is > > simplify wrong - per the official decree we can support different > > protocols on different ports, and that's why we also pass the port_num > > to this function. How about rdma_cap_read_inv? > > We can't support different protocols on different ports. The ports > hung off a struct device must be similar. The rules might be bent a > bit for rocee/ib sharing (not a great idea, IMHO) but certainly iwap > and !iwarp cannot share a struct device. That's how I would like to see it. But the way how all the cap checks are done on a per port basis implies something different. And for this series I'm not going to changes this. In the long run I'd be much happier if this nonsense of passing a port number to the capability checks would go away. -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
