Re: [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