[ovs-dev] [PATCH 2/5] dpif-netdev: Implement the API functions to allow multiple handler threads read upcall.
Alex Wang
alexw at nicira.com
Thu Mar 6 01:38:41 UTC 2014
Hey Ben,
> It would be nice to add a sentence or two to the commit message
> explaining why we think it's OK to drop the distinction between the
> two queues we had before (as already discussed).
>
Yes, I'll do that,
> The choice of hash function, in flow_hash_5tuple(), seems curious. I
> would have expected something like:
>
> hash = mhash_add(basis, flow->nw_src);
> hash = mhash_add(hash, flow->nw_dst);
> hash = mhash_add(hash, (flow->tp_src << 16) | flow->tp_dst);
> hash = mhash_add(hash, flow->nw_proto);
> return mhash_finish(hash, 13);
>
> with some OVS_FORCE casts thrown in to make sparse happy. Any
> particular reason to just xor everything together? (Do we need a
> symmetric hash?)
>
Yeah, I did xor when prototype datapath fairness project and kept using
them.
I'll use mhash.
> I didn't look over the details of the queuing changes yet.
>
I'll post a V2 once we settle down on the API patch.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openvswitch.org/pipermail/ovs-dev/attachments/20140305/afd4a666/attachment-0005.html>
More information about the dev
mailing list