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

[Sagi Grimberg]

>> 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