> > and a series we would like to get into 2.8
> > netdev-dpdk: Use intermediate queue during packet transmission. Bhanu Jun
> > 29/V3
> > netdev: Add netdev_txq_flush function.
> > netdev-dpdk: Add netdev_dpdk_txq_flush function.
> > netdev-dpdk: Add netdev_dpdk_vhost_txq_flush function.
> > netdev-dpdk: Add intermediate queue support.
> > netdev-dpdk: Enable intermediate queue for vHost User port.
> > dpif-netdev: Flush the packets in intermediate queue.
> I think that we still not reached agreement about the level of implementation
> (netdev-dpdk or dpif-netdev). Just few people participate in discussion which
> is not very productive. I suggest not to target output batching for 2.8
> because of this and also lack of testing and review.
> As I understand, we have only 3 days merge window for the new features
> and I expect that we can't finish discussion, review and testing in time.
Ericsson is very much interested in time-based Tx batching, especially for
vhostuser ports as a means to reduce the virtio interrupts for non-PMD guests.
We do not see significant value of Tx batching within an Rx batch in realistic
use cases as Rx batches are fairly small until we reach PMD saturation.
We would like to take the time to evaluate and benchmark the two approaches
(Bhanu's above patches plus his RFC patch for time-based batching in
dpif-netdev and netdev-dpdk and Ilya's simpler patch set limited to
dpif-netdev). We will provide feedback on the ML next week.
In this situation we do not recommend merging the above patches in a rush for
> In addition I have a few general thoughts about merging via pull requests:
> 2. I'm a fan of plain git history. Could we use 'Rebase and merge' policy
> without merge commits ?
+1 for maintaining a plain git history w/o merge commits
dev mailing list