Yossi Etigin wrote:
Sorry, I just saw a mistake in this patch (missing return value in
cma_set_qkey),
please use this code instead:
When doing rdma_resolve_addr() and relevant port is down, the
function fails
and rdma_cm id is not bound to the device. Therefore, application does
not have
device handle and cannot wait for the port to become active. The function
fails because ipoib is not joined to the multicast group and therefore
sa does not have a multicast record to take a qkey from.
The proposed patch is to make lazy qkey resolution - cma_set_qkey
will set id_priv->qkey if it was not set, and will be called just
before the qkey is
really required.
Signed-off-by: Yossi Etigin <[email protected]>
---
Applied,
Regards,
Vladimir
_______________________________________________
ewg mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
