[openib-general] Re: [PATCH] mad: add GID/class checking for matching received to sent MADs

[Michael S. Tsirkin]

Quoting r. Sean Hefty <mshefty at ichips.intel.com>:
> Subject: Re: [PATCH] mad: add GID/class checking for matching received to sent MADs
> 
> Michael S. Tsirkin wrote:
> >I think we return 1 here since the TID check performed earlier is 
> >sufficient
> >in this case since its locally unique.
> >As I said, same_gid might not be the best name for this function, but the
> >logic overall looks right to me.
> 
> Isn't a user permitted to re-use the same TID when sending out different 
> GIDs? I think that the function name is fine, I'm just not sure about the 
> check being completely correct.  (It's more correct than what was there 
> before, so the changes have been committed.)

Right. We assumed the user does not do this, managing TIDs per port.
If you want to enable this, you'll need to extend the check for this case.

-- 
Michael S. Tsirkin
Staff Engineer, Mellanox Technologies