[openib-general] [PATCH 1 of 2] core extensions to fix race at module unloading

[Sean Hefty]

Michael S. Tsirkin wrote:
>  /**
> + * ib_flush_mad_agent - flush any callbacks in flight for this client.
> + * @mad_agent: Corresponding MAD registration request to flush.
> + */
> +int ib_flush_mad_agent(struct ib_mad_agent *mad_agent);

To restate what I've mentioned before, I don't see any issue with the ib_mad 
interface.  It doesn't make sense to me to change this API, when the problem is 
in the client.  But I'd like to hear what others think.

- Sean