[openib-general] GSI compromise

[Sean Hefty]

On Wed, 04 Aug 2004 10:28:04 -0700
Roland Dreier <roland at topspin.com> wrote:

> How does the response bit get handled then?  (eg lots of consumers
> will want to see get responses but not gets).

Currently, there's not a way to hand a received MAD to more than one consumer, so responses go only to the requestor.  My preference would be to give ownership of a received MAD to one consumer in order to avoid data copies.

To allow other clients to see MADs, I was thinking of adding in some sort of snooping functionality (either by extending these APIs or adding a new one).  But I wanted to make it clear through the API which clients would be taking ownership of received MADs, versus those who were only allowed to view it.

> Also if it's just a bitmask of which method we want, we might as well
> make it a bitmap (ie use DECLARE_BITMAP(method_mask, 256) or something
> like that).

I've updated the file to use a bitmask.  I've also added the file to SVN under https://openib.org/svn/trunk/contrib/intel/ib_mad.h.
 
>     Roland> - should we use a "struct list_head" instead of our own type
>     Roland> for ib_mad_msg.next?
> 
>     Sean> I'm fine with this.  I followed what I had for the work
>     Sean> requests.  Maybe those should change as well?
> 
> That's a good idea.  It adds one more pointer to the work request
> (since struct list_head is doubly linked) but I think being able to
> use macros like list_for_each(), list_del(), etc. is worth it.

I will update the pointers.