Re: [PATCH rfc 6/6] nvme-rdma: use intelligent affinity based queue mappings

[Sagi Grimberg <sagi@xxxxxxxxxxx>]

> > Use the geneic block layer affinity mapping helper. Also,
>
>           generic
>
> >  	nr_io_queues = min(opts->nr_io_queues, num_online_cpus());
> > +	nr_io_queues = min_t(unsigned int, nr_io_queues,
> > +				ibdev->num_comp_vectors);
> > +
>
> Add a comment here?

Will do