On 11/18/2015 4:10 PM, Christoph Hellwig wrote:
There was absolutely nothing relating to bidi in either the initial iSER checking
This is wrong assertion.
Look on the code throughout the iser path done from iser_send_command,
we allowed the command associated with the
iscsi task to be IN, OUT, both or none, when we do all the dma-mapping,
memory registrations and such for either of the
needed directions and same on the completion path.
Sagi, do we still do IN, OUT, both or none? if not, where this stopped
to be supported? how large would be the fix?
Or.
[1] maybe start with 2.6.20, I guess we improved later... till the
point where things were potentially started
to be caught down, as I realized now that Sagi wasn't fully aligned on
that aspect of the driver
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
