>Some things unclear to me: >1. Is a user somehow *required* to call ib_cm_establish if he gets the event? > E.g. what about cma users?
A user is not required to call ib_cm_establish. The patch fixes the issue where CMA users receive a COMM_EST event, but do not have the ability to call ib_cm_establish directly. It does this by pushing COMM_EST event handling directly to the IB CM. >2. There are now two ways CM gets notified about Communication established > events - what happends if both the user calls ib_cm_establish, and CM > gets this directly? The second call will fail due to an invalid state. Note that this failure was already possible, since an RTU could have been processed before the user could have called ib_cm_establish. >3. If I understand it correctly, CM does not yet know how to switch to the > alternate path for CM messages. So, is the patch still useful by itself, > or is it more of a framework preparatory work? The patch is useful by itself for the COMM_EST event. It is my intent that the same infrastructure be used to notify the IB CM of path migration. PATH_MIG event handling is currently limited to mthca reporting path migration events to the CM event handler. The IB CM currently just ignores the event. If it makes it cleaner, I can remove the line from mthca that reports PATH_MIG events. - 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
