On 11/28/2018 4:25 PM, Jason Gunthorpe wrote: > On Wed, Nov 28, 2018 at 04:21:48PM -0600, Steve Wise wrote: > >>>> It does make sense to not require type. The name must be unique so that >>>> should be enough. I'll have to respin the kernel side though... >>> The delete_link really should be an operation on the ib_device, not >>> the link_ops thing. >>> >>> That directly prevents mis-matching function callbacks.. >>> >>> Jason >> Looking at the rtnetlink newlink/dellink, I see they cache the link_ops >> ptr in the net_device struct. So when the link is deleted, then >> appropriate driver-specific dellink function can be called after finding >> the device to be deleted. Should I do something along these lines? IE >> add a struct rdma_link_ops pointer to struct ib_device. > I don't see a problem with that either.. > > Jason Ok, I'll respin the kernel and user patches tomorrow. Thanks!
