Re: [PATCH v2 7/8] net/ring: add promisc and all-MC stubs

[Bruce Richardson <bruce.richardson@intel.com>]

On Fri, Feb 04, 2022 at 02:36:47PM +0000, Ferruh Yigit wrote:
> On 12/21/2021 7:57 PM, Robert Sanford wrote:
> > Add promiscuous_enable, promiscuous_disable, allmulticast_enable,
> > and allmulticast_disable API stubs.
> > This helps clean up errors in dpdk-test link_bonding_mode4_autotest.
> > 
> > Signed-off-by: Robert Sanford <rsanford@akamai.com>
> > ---
> >   drivers/net/ring/rte_eth_ring.c | 28 ++++++++++++++++++++++++++++
> >   1 file changed, 28 insertions(+)
> > 
> > diff --git a/drivers/net/ring/rte_eth_ring.c b/drivers/net/ring/rte_eth_ring.c
> > index db10f03..cfb81da 100644
> > --- a/drivers/net/ring/rte_eth_ring.c
> > +++ b/drivers/net/ring/rte_eth_ring.c
> > @@ -226,6 +226,30 @@ eth_mac_addr_add(struct rte_eth_dev *dev __rte_unused,
> >   }
> >   static int
> > +eth_promiscuous_enable(struct rte_eth_dev *dev __rte_unused)
> > +{
> > +	return 0;
> > +}
> > +
> > +static int
> > +eth_promiscuous_disable(struct rte_eth_dev *dev __rte_unused)
> > +{
> > +	return 0;
> > +}
> > +
> > +static int
> > +eth_allmulticast_enable(struct rte_eth_dev *dev __rte_unused)
> > +{
> > +	return 0;
> > +}
> > +
> > +static int
> > +eth_allmulticast_disable(struct rte_eth_dev *dev __rte_unused)
> > +{
> > +	return 0;
> > +}
> > +
> > +static int
> >   eth_link_update(struct rte_eth_dev *dev __rte_unused,
> >   		int wait_to_complete __rte_unused) { return 0; }
> > @@ -275,6 +299,10 @@ static const struct eth_dev_ops ops = {
> >   	.stats_reset = eth_stats_reset,
> >   	.mac_addr_remove = eth_mac_addr_remove,
> >   	.mac_addr_add = eth_mac_addr_add,
> > +	.promiscuous_enable = eth_promiscuous_enable,
> > +	.promiscuous_disable = eth_promiscuous_disable,
> > +	.allmulticast_enable = eth_allmulticast_enable,
> > +	.allmulticast_disable = eth_allmulticast_disable,
> 
> not sure about adding dummy dev_ops to the driver for the unit test,
> what about updating 'link_bonding_mode4_autotest' accordingly?
> 
> Bruce (net/ring maintainer), what do you think about dummy dev_ops?

For something like ring PMD, I don't see an issue with it, since they don't
really have any meaning for a ring PMD, they might as well just return
success rather than having application code have to handle errors from
them.


Acked-by: Bruce Richardson <bruce.richardson@intel.com>