[ofa-general] librdmacm and libmthca question

Sean Hefty sean.hefty at intel.com
Thu Oct 11 00:00:31 PDT 2007


>might be of value to document somewhere that the initiator depth and
>responder resources are not directly related to the actual work queue
>depth, and that without some sort of intervention, are not that high.

FYI - I am currently updating the librdmacm man pages based on your other
e-mail.  I will make sure to document these values.

>This leads one to believe that the values are OK, even though they fail
>when you use the same values in rdma_accept.  A note to this effect in
>the man pages would help.

I will at least note how to obtain the correct maximum for local
responder_resources by querying the local HCA.  The correct setting for
initiator_depth needs to come from the remote endpoint.


Please continue to let me know what other problems you run into.  I'd like to
address as many issues as possible for the next release.  (New features will be
pushed out some.)

- Sean



More information about the general mailing list