[openib-general] Re: Some Initial RMPP Comments and Questions

[Sean Hefty]

Hal Rosenstock wrote:
>>>That would make the receive side of user space much easier. Shall I
>>>defer implementing this part until coalesce is implemented ?
>>
>>If you want to go ahead and do this part of the implementation, you can 
>>just call coalesce and assume that it'll be there shortly...
> 
> 
> Sounds good. Thanks.

I'm considering changing the API to this:

ib_coalesce_recv_mad(mad_agent, mad_recv_wc, user_buffer, buf)

The mad_agent is needed in order to know if there's an RMPP header.  If 
there is, then there may be padding at the end of the MAD which the 
user didn't allocate buffer space for.  The user_buffer parameter would 
be used to indicate which copy routine to user: memcpy or copy_to_user.

- Sean