> My last version of this patch posted DEVICE_FATAL and you said it > wasn't right.
DEVICE_FATAL is right but it should be an unaffiliated event... ie rather than calling the QP event handler for every QP, just call ib_dispatch_event() with one DEVICE_FATAL event. See eg drivers/infiniband/hw/mlx4/main.c in mlx4_ib_event(). What I said last time was: | This doesn't match the IB driver behavior (or the IB spec) -- the | DEVICE_FATAL event is unaffiliated and delivered for the adapter as a | whole. QP events are supposed to be for events connected to a single | QP, not the whole adapter failing. and I stand by that... QP events are for things connected to one QP, but a PCI EEH event is not affiliated with any individual object but rather the adapter as a whole. I think you misunderstood my point -- I wasn't saying to make it a QP event, but rather to make it unaffiliated. - 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
