[ofa-general] Re: [PATCH] ib/cm: fix stale connection detection

Michael S. Tsirkin mst at dev.mellanox.co.il
Wed May 23 12:57:00 PDT 2007


> Quoting Sean Hefty <sean.hefty at intel.com>:
> Subject: RE: [ofa-general] Re: [PATCH] ib/cm: fix stale connection detection
> 
> >> If two REQs are received with matching local IDs, but the REQs
> >> themselves differ on one or more fields, such as the QPN, the second REQ
> >> is dropped as a duplicate.
> >
> >Why do you speak about dropping duplicates as a valid response?
> 
> I was only mentioning the current behavior.
> 
> >As far as I can tell, the 2 legal responses to a duplicate REQ
> >are resending a REP and rejecting with code 30.
> 
> It's possible to receive a duplicate REQ before processing has completed and a
> REP generated to the first REQ.  In this case, it makes sense simply to discard
> the duplicate REQ.
> 
> When processing completes on the first REQ, the CM will generate either a REP or
> a REJ, so I believe that the behavior is compliant when handling an actual
> duplicate REQs.

Well, in case the second REQ differs from the first one, discarding
it might not be the best option: I think we might want to queue it
and process when you exit the ephemeural state.

-- 
MST



More information about the general mailing list