Skip to contentSkip to conversation

RDMA_CM_EVENT_REJECTED

DISCUSSION Steve Wise 2 participants 2 messages Oct 25, 2016 InfiniBand

lore
Share & tools
lore

Steve Wise

author
#0 lore raw

While working on the reject helper functions, I notice that some of the rdma_cm kernel users have handlers for RDMA_CM_EVENT_REJECTED even for cm_ids that are spawned from listening cm_ids as the result of a RDMA_CM_EVENT_CONNECT_REQUEST, and that issue an rdma_accept() or rdma_reject() in response.

Q: is it possible for a REJECTED event to get posted to these cm_ids? It is not for iWARP, but I'm not sure about IB. I'm guessing this is just due to cut/paste process of adding event handlers, but I wanted to make sure.

Steve.

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#1 lore raw
Quoted 8 lines

While working on the reject helper functions, I notice that some of the rdma_cm
kernel users have handlers for RDMA_CM_EVENT_REJECTED even for cm_ids that are
spawned from listening cm_ids as the result of a RDMA_CM_EVENT_CONNECT_REQUEST,
and that issue an rdma_accept() or rdma_reject() in response.

Quoted 4 lines

Q: is it possible for a REJECTED event to get posted to these cm_ids? It is not
for iWARP, but I'm not sure about IB. I'm guessing this is just due to cut/paste process of adding event handlers, but I wanted to make sure.

If I'm following this correctly, then, yes, I believe those cm_id's can be rejected. The active side of a connection can give up waiting for the passive side to act. As a final gesture, the active side will send a reject message to notify the passive side that it's aborting the connection request.

- Sean
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html