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

[Or Gerlitz]

On Wed, Feb 4, 2009 at 6:41 PM, Sean Hefty <sean.hefty at intel.com> wrote:
> There may be some way to defer setting the qkey if it's not available when binding, but how
> does allowing the bind to proceed help?  Without the qkey, the QP is basically unusable.

We have two usage cases:

- an rdma-cm based app wants to determine if the route for a multicast
group leads to IPoIB interface/device based on the outcome of
rdma_bind_addr etc

- for HA scheme, an app want to resolve the device/port and then use
IB events as a trigger to actually start doing things such as QP
creation, Joining multicast groups, etc

Or