[openib-general] Re: [PATCH] ib_mad: prevent duplicateoutstanding MADtransactions with same TID

[Michael S. Tsirkin]

Quoting r. Sean Hefty <mshefty at ichips.intel.com>:
> >I think we can use the method field for this.
> >C15-0.1.18:
> >
> >...
> >. The method used for all packets sent from the Receiver to the Sender
> >shall be SubnAdmGetTable() or SubnAdmGetTraceTable(), depending
> >on which initiated the transfer.
> >
> >. The method used for all packets sent from the Sender to the Receiver
> >shall be SubnAdmGetTableResp().
> 
> I'd like to resist putting protocol specific knowledge into the MAD layer 
> as much as possible.  Such a fix is unlikely to apply in more generic 
> cases.  I believe that this is an architectural issue that would be best 
> addressed at that level.

I dont see a way around this: the dirt is at the spec level.

You just said:
	I don't believe that there's any way to know if an
	abort is for an RMPP message that is being sent, versus received.

Did you change your mind then?

-- 
Michael S. Tsirkin
Staff Engineer, Mellanox Technologies