[ofa-general] librdmacm 1.0.4 release

[Or Gerlitz]

Sean Hefty wrote:
> I've pushed out a release 1.0.4 of librdmacm that addresses some of the feedback
> from Doug.  Patches were posted previously to the list, with a small update
> based on that feedback.

> Changes from 1.0.3:
> librdmacm/cma: provide wrapper functions to extract src/dst addresses
> librdmacm/cma: provide sanity checks for max outstanding rdma ops
> librdmacm/man: update man pages to clarify connection request params

Hi Sean,

I think you have mentioned that some documentation update is planned?

Anyway, I know that both rdma_connect and rdma_accept get struct 
rdma_conn_param and following a question from a user, I wondered which 
of the fields are actually relevant in the passive side.

Doing a quick look at the kernel core code, I saw that:

- param.retry_count is ignored in the passive side rdma-cm code and the 
IB cm uses the one present in the req message.

- param.rnr_retry_count is not ignored in the passive side, but from 
looking in the code, I was not sure if the value used is the one present 
in the req or the one supplied by the passive consumer.

- param.flow_control is a pure SW field which does not get into the QP 
attr. My understanding is that IB RC flow-control means non zero rnr 
counter, is this all? if yes, maybe we need to expose only 
rnr_retry_count field

Or.