[openib-general] a question about the local/remote CM.REQ attruibutes

[Dotan Barak]

kalash nainwal wrote:

> Sorry for jumping in, but in "InfiniBand Architecture Specification
> Volume 1" page 594, they have explained -
>
> RESPONDER RESOURCES: ...The recipient of the REQ message shall choose
> a local Initiator Depth that does not exceed the Responder Resources
> offered in the REQ...
>
> INITIATOR DEPTH: ...The recipient of the REQ message should try to
> choose a number of local Responder Resources that is greater than or
> equal to the Initiator Depth
> in the REQ message...
>
> So if you can have a look at the code for cm_format_req_event() (which
> is called from cm_req_handler):
>
> .
> struct ib_cm_req_event_param *param;
> .
> param->responder_resources = cm_req_get_init_depth(req_msg);<<<<
> param->initiator_depth = cm_req_get_resp_res(req_msg);<<<<<
> param->local_cm_response_timeout 
> =cm_req_get_remote_resp_timeout(req_msg);<<<<
> param->remote_cm_response_timeout 
> =cm_req_get_local_resp_timeout(req_msg);<<<<
> .
> .
> So init_depth in the req message is simply chosen as the value of
> param->responder_resources, and vice versa.
>
> -Kalash

thanks, i saw this code, i just wondered if this is the right behavior ...

Dotan