[openib-general] [PATCH] Initial checkin of userspace MAD access

[Hal Rosenstock]

On Wed, 2004-11-03 at 12:14, Roland Dreier wrote:
> Another option is to revise the kernel MAD code so that it does not
> need to register an agent for the SM classes (ie pass all MADs to
> low-level driver first).

That's what I'm working on now (eliminate snoop_mad and replace with
process_mad). I should have the first cut by COB today. After that, I
will work on SMI restructure that needs to be done so that outgoing SMI
updating becomes part of ib_post_send_mad rather than a precursor to it.

-- Hal