13/05/2021 11:59, Bruce Richardson: > On Wed, May 12, 2021 at 09:43:57PM +0300, Medvedkin, Vladimir wrote: > > Hi Bruce, > > > > On 12/05/2021 19:32, Bruce Richardson wrote: > > > On Mon, May 10, 2021 at 06:53:19PM +0200, Heinrich Kuhn wrote: > > > > From: "Chaoyong.He" <chaoyong...@corigine.com> > > > > > > > > Set the Rx multi-queue mode to NONE when configuring a port that is > > > > associated with hardware that only supports a single Rx queue. > > > > > > > > Signed-off-by: Chaoyong He <chaoyong...@corigine.com> > > > > Signed-off-by: Heinrich Kuhn <heinrich.k...@netronome.com> > > > > Signed-off-by: Simon Horman <simon.hor...@netronome.com> > > > > --- > > > > + if (dev_info.max_rx_queues == 1) > > > > + local_port_conf.rxmode.mq_mode = ETH_MQ_RX_NONE; > > > > + > > > > > > While it makes sense to do this when the port only supports a single > > > queue, > > > would it not also make sense to do this when the requested queues are 1 > > > too? > > > > > > Adding some lookup library maintainers on CC - I assume that the RSS value > > > is not actually used for lookup anywhere in l3fwd. > > > > > > > As far as I can see the rss hash value is not used anywhere in l3fwd. In > > LPM/FIB this is not required at all, in EM CRC or Jenkins hash is used. > > > That's what I thought from looking at the code too. Since this is not > really a bug fix, I think it can be pushed till 21.08. > > With or without the change I suggest above: > Acked-by: Bruce Richardson <bruce.richard...@intel.com>
Applied, sorry it has been waiting so long.
