> Hmm, I would like to quote the spec *literally*. Maybe > - and then invoke a Destroy QP or Reset QP.
That's fine ... in fact the spec has a bullet for that line too (I just looked). The way you have it formatted now is visually kind of strange (the last line looks odd): > + * - Put the QP in the Error State > + * - Wait for the Affiliated Asynchronous Last WQE Reached Event; > + * - either: > + * drain the CQ by invoking the Poll CQ verb and either wait for CQ > + * to be empty or the number of Poll CQ operations has exceeded > + * CQ capacity size; > + * - or > + * post another WR that completes on the same CQ and wait for this > + * WR to return as a WC; (NB: this is the option that we use) > + * and then invoke a Destroy QP or Reset QP. which doesn't really match the spec's formatting... I guess this is pretty minor, but I would write the comment as: * - Put the QP in the Error State; * - wait for the Affiliated Asynchronous Last WQE Reached Event; * - either: * - drain the CQ by invoking the Poll CQ verb and either wait for CQ * to be empty or the number of Poll CQ operations has exceeded * CQ capacity size; or * - post another WR that completes on the same CQ and wait for this * WR to return as a WC; * - and then invoke a Destroy QP or Reset QP. * * For IPoIB we choose the second option of posting another WR, and we * keep a dedicated QP in the error state for doing this. > > This actually seems like a good motivation for the mthca RESET -> > > ERROR fix. We could avoid the transition to INIT if we fixed mthca > > and mlx4, right? > > Yes. That was the motivation. OK, I just queued the mthca and mlx4 versions of the RESET->ERROR fix. So I guess you can drop the dummy transition to INIT within IPoIB for the final version of the WQE leakage patch. - R. _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
