On Fri, Sep 15, 2017 at 06:30:06PM +0000, Darrell Ball wrote:
> That was my first reaction to remove such blind re-try. It could have
> been a good option if all the PMD driver report the error consistenly.
>
> And unfortunately, it's not true. For example, for MARK without QUEUE
> action, i40e reports RTE_FLOW_ERROR_TYPE_ACTION, which, IMO, is nothing
> wrong. While for mlx5, it reports RTE_FLOW_ERROR_TYPE_HANDLE. I don't
> know why it was set like this, and we may could fix this. But my point
> was, it's not that reliable to use rte_errno, at least it's true for now.
>
> [Darrell] I see what you mean looking at the drivers code and I imagine it
> makes you work difficult.
> It looks like only either RTE_FLOW_ERROR_TYPE_ACTION or
> RTE_FLOW_ERROR_TYPE_HANDLE
> is used though in these cases from what I saw; would it make
> sense to check for those specifically?
Yes, I think it's better than a blind retry (assuming it's failed due
to an unsupported protocol). Meanwhile, I will double check the DPDK
driver code.
--yliu
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
