[openib-general][PATCH][RFC]: CMA IB implementation

Guy German guyg at voltaire.com
Tue Sep 20 03:18:47 PDT 2005


Roland Dreier wrote:
> the only use of it is bogus as well.  You never use timeout again,
> which is good because qp_attr->timeout is not in units of
> microseconds; it's already in the IB logarithmic scale.

You are right. It is a dead code - I will remove it.
The first implementation takes default values.

Guy




More information about the general mailing list