[openib-general] [PATCH] rdma_cm oops: rdma_destroy_id

[Roland Dreier]

    Michael> Given that iwarp support on trunk is for now just a stub
    Michael> ...  I guess we can just fix the IB part, and worry about
    Michael> the iWARP part when its there.

My confusion comes from the fact that rdma_resolve_addr() just calls
rdma_resolve_ip() unconditionally (except for loopback addresses),
even on the iWARP branch.  So I'm confused about the design, since I
thought ib_addr was IB-specific.

 - R.