[openib-general] [RFC] CM reject handling and redirection

[Hal Rosenstock]

On Wed, 2005-02-16 at 22:50, Roland Dreier wrote:
> Code 25 (port redirection only) raises a separate issue that I think
> is orthogonal to REJ handling.  The CM spec always allows the CM to
> run on a different port from the destination of a connection so if we
> want to implement the whole spec we do need a way for the consumer to
> specify a primary and alternate path for the connection itself as well
> as a CM path.

How much complexity is added to the CM to support this ? Do we really need to
(and if so what is the application which requires this) ? 

> As you say, separating the CM path from the connection path does raise
> the question of what to do for failover and APM.  As far as I can tell
> this is an area that the IB spec is silent on and I'm not sure that
> anyone ever came up with a good answer.

I'm not sure I get the "connection" between the CM path and data path.
Doesn't failover and APM apply only to the data path ? Is it purely the
detected of a failed path between CMs (which seems orthogonal to this to
me) ? That's not to say that this is not an issue.

-- Hal