[ofa-general] RE: impossibility to bind a device/port with the rdma-cm when the port is down

[Sean Hefty]

>cma_acquire_dev --> cma_set_qkey/ps=IPOIB --> ib_sa_get_mcmember_rec where the
>latter returns EADDRNOTAVAIL since when the port went down the core multicast
>code

Why is ib_sa_get_mcmember_rec being called?  Or is this issue separate from what
udaddy is showing?

>I  assume there must be a way to defer this resolving to a later stage such
>that binding would be possible when the port is down, thoughts?

Can you determine what call in the kernel is actually failing during the bind?
(I can try testing later, but I'm not near any systems currently.)  I'm
wondering if the failure is coming from rdma_translate_ip()->ip_dev_find().

- Sean