[ofa-general] rdma cm timeout option, was [iWARP issues]

[Caitlin Bestler]

On Nov 2, 2007 11:17 AM, Kanevsky, Arkady <Arkady.Kanevsky at netapp.com> wrote:
> Yes on first.
> No on second yet.
> Maybe for iWARP the only acceptable value will be "default"?
> But this still feels that ULPs need to do something transport specific.
> except for "default" case.
> But implementing TCP style timeout for IB looks like overkill.
>

Keep in mind that the transport layer timeout only helps iWARP for establishing
the TCP connection. It does not deal with stalled MPA Request exchanges.

So there is need for *some* mechanism to timeout a stalled iWARP connectionion
process even after a valid TCP connection is established. A vendor
specific, non-
configurable, method is probably more than adequate. But *something* has to be
there, you cannot just rely on the TCP mechanisms. Nor can you assume that the
TCP stack servicing RDMA connections has the same defaults as the host stack.