[dpdk-dev] [PATCH v6 2/2] app/testpmd: fix port stop

[Wu, Jingjing]

> -----Original Message-----
> From: Iremonger, Bernard
> Sent: Friday, February 3, 2017 6:38 PM
> To: Wu, Jingjing <jingjing.wu at intel.com>; dev at dpdk.org
> Cc: stable at dpdk.org
> Subject: RE: [PATCH v6 2/2] app/testpmd: fix port stop
> 
> Hi Jingjing
> 
> > -----Original Message-----
> > From: Wu, Jingjing
> > Sent: Friday, February 3, 2017 8:22 AM
> > To: Iremonger, Bernard <bernard.iremonger at intel.com>; dev at dpdk.org
> > Cc: stable at dpdk.org
> > Subject: RE: [PATCH v6 2/2] app/testpmd: fix port stop
> >
> >
> >
> > > -----Original Message-----
> > > From: Iremonger, Bernard
> > > Sent: Friday, January 27, 2017 6:50 PM
> > > To: dev at dpdk.org; Wu, Jingjing <jingjing.wu at intel.com>
> > > Cc: Iremonger, Bernard <bernard.iremonger at intel.com>;
> > stable at dpdk.org
> > > Subject: [PATCH v6 2/2] app/testpmd: fix port stop
> > >
> > > The rte_eth_dev_stop function is not called if the port_status is
> > > not RTE_PORT_STARTED. This can happen if the rte_eth_dev_start
> > > function is called directly, ie not through the start_port function.
> > >
> > > Make sure rte_eth_dev_stop is always called in stop_port function.
> > >
> > > Fixes: ce8d561418d4 ("app/testpmd: add port configuration settings")
> > >
> > > CC: stable at dpdk.org
> > > Signed-off-by: Bernard Iremonger <bernard.iremonger at intel.com>
> > > ---
> > >  app/test-pmd/testpmd.c | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index
> > > 3d25436..0d7a4d4 100644
> > > --- a/app/test-pmd/testpmd.c
> > > +++ b/app/test-pmd/testpmd.c
> > > @@ -1490,13 +1490,13 @@ stop_port(portid_t pid)
> > >  			continue;
> > >  		}
> > >
> > > +		rte_eth_dev_stop(pi);
> > > +
> > >  		port = &ports[pi];
> > >  		if (rte_atomic16_cmpset(&(port->port_status),
> > > RTE_PORT_STARTED,
> > >  						RTE_PORT_HANDLING) == 0)
> > >  			continue;
> > >
> > > -		rte_eth_dev_stop(pi);
> > > -
> >
> > I don't think this fix is correct to move rte_eth_dev_stop above.
> >
> > We need to make sure rte_eth_dev_start is called in start_port. For
> > vmdq configuration, You just need to change the configuration when
> > port is stopped.
> 
> I think the stop_port() function should always stop the port even if the
> port_status is not correct for any reason.
> At present stop_port() returns without stopping the port if the port_status is not
> RTE_PORT_STARTED.
>
This is testpmd's design. If you think it is an issue, maybe you need to prepare patch to optimize it. But for VMDQ configuration, I'd like to make it independent but not mixed.
 
> The VMDq configuration is done whet the port is stopped, however to the
> complete the VMDq configuration the port must be started.
> 

To change minor, we can stop port, then configure VMDQ and then start port.

You make port started in VMDQ config, the Symmetry of stop/start command is broken and it is not easy to maintain.

> Regards,
> 
> Bernard.