[ofa-general] Re: [PATCH] IB/core: fix null pointer dereference in local_completions()

[Ralph Campbell]

On Wed, 2009-02-25 at 16:53 -0800, Roland Dreier wrote:
> By the way, I didn't pay close attention to the previous discussion
> about this.  Did you and Hal reach agreement about the approach?
> 
>  - R.

The earlier patch I posted wasn't correct. I was looking for
comments about how kmem_cache_free() is called when
recv_mad_agent->agent.recv_handler() is called.
Hal didn't answer directly so I checked the code and I see that
the receive handler is responsible for calling ib_free_recv_mad()
which does the work.

Hal just wanted me to test it, which I did.