On Fri, 2006-07-07 at 14:30, Sean Hefty wrote:
> The following check in user_mad is done when sending a MAD.
>
> /*
> * If userspace is generating a request that will generate a
> * response, we need to make sure the high-order part of the
> * transaction ID matches the agent being used to send the
> * MAD.
> */
> method = ((struct ib_mad_hdr *) packet->msg->mad)->method;
>
> if (!(method & IB_MGMT_METHOD_RESP) &&
> method != IB_MGMT_METHOD_TRAP_REPRESS &&
> method != IB_MGMT_METHOD_SEND) {
> tid = &((struct ib_mad_hdr *) packet->msg->mad)->tid;
> *tid = cpu_to_be64(((u64) agent->hi_tid) << 32 |
> (be64_to_cpup(tid) & 0xffffffff));
> }
>
> Does anyone know the reason for the last check (method != SEND)?
Because there is no required response to a send method.
> Isn't it still a requirement that the TID be unique?
I'm not sure if duplicate detection is based on the TID or some upper
layer protocol but if send is an operation under C13-18.1.1, each
different send is a different TID. Some SENDs have "responses" which are
sent by SEND method and others don't. I'm not sure if the matching is
transaction ID based or not.
-- Hal
> I'm trying to add some validation checks against MADs sent from userspace to
> protect against a client sending non-data RMPP MADs, duplicate MADs, etc., and
> the above check looks incorrect to me.
>
> - Sean
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
