[openib-general] Re: [PATCH] request/response matching in MAD code

Sat Oct 2 04:28:04 PDT 2004

On Fri, 2004-10-01 at 12:41, Sean Hefty wrote:
> On Fri, 01 Oct 2004 12:33:44 -0400
> Hal Rosenstock <halr at voltaire.com> wrote:        
> > 2. Added the following to reassemble_recv (it was eliminated from
> > ib_mad_recv_done_handler):
> > 
> > INIT_LIST_HEAD(&recv->header.recv_buf.list);
> 
> I was going to get back to RMPP handling.  

I think we can wait on RMPP.

> 2. Added the following to reassemble_recv (it was eliminated from
> > ib_mad_recv_done_handler):
> > 
> > INIT_LIST_HEAD(&recv->header.recv_buf.list);
> I'm wasn't sure if we wanted to use a doubly linked list or singly linked one for this.

This isn't just for RMPP; it's needed to free the receive buffers
correctly.

I'm not sure either and have used doubly linked lists everywhere
although I too think that singly linked ones would suffice.

> > Also, should the TID be overwritten in the high 32 bits or do we trust
> > the client to set this properly ?
> 
> Based on our previous discussions on this, clients are responsible for setting the upper 32-bits of the TID correctly.  This should be more efficient, and only requires to use the TID provided through registration.
>  
> > Note that I only validated an incoming request and have not tested the
> > request/response matching as yet.
> 
> We may want to delay testing of this until I can get the timeout code added.

I will test this without timeouts first and then retest when timeouts
are implemented.

-- Hal