[dpdk-dev] [PATCH v4] null: fake PMD capabilities

[Michał Mirosław]

On Mon, Jan 09, 2017 at 12:07:36PM +0000, Ferruh Yigit wrote:
> On 12/14/2016 7:16 PM, Michał Mirosław wrote:
> > From: Paweł Małachowski <pawel.malachowski at atendesoftware.pl>
> > 
> > This allows for testing code which needs offloads to be supported.
> > 
> > Signed-off-by: Michał Mirosław <michal.miroslaw at atendesoftware.pl>
> > ---
> >  drivers/net/null/rte_eth_null.c | 23 ++++++++++++++++++++++-
> >  1 file changed, 22 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/net/null/rte_eth_null.c b/drivers/net/null/rte_eth_null.c
> > index 836d982..e60516f 100644
> > --- a/drivers/net/null/rte_eth_null.c
> > +++ b/drivers/net/null/rte_eth_null.c
> > @@ -284,6 +284,9 @@ eth_tx_queue_setup(struct rte_eth_dev *dev, uint16_t tx_queue_id,
> >  	return 0;
> >  }
> >  
> > +static void
> > +eth_dev_void_ok(struct rte_eth_dev *dev __rte_unused) { return; }
> > +
> >  
> >  static void
> >  eth_dev_info(struct rte_eth_dev *dev,
> > @@ -304,6 +307,19 @@ eth_dev_info(struct rte_eth_dev *dev,
> >  	dev_info->pci_dev = NULL;
> >  	dev_info->reta_size = internals->reta_size;
> >  	dev_info->flow_type_rss_offloads = internals->flow_type_rss_offloads;
> > +	/* We hereby declare we can RX offload VLAN-s out of thin air (they are not there)
> > +	 */
> > +	dev_info->rx_offload_capa = DEV_RX_OFFLOAD_VLAN_STRIP |
> > +		DEV_RX_OFFLOAD_QINQ_STRIP;
> > +	/* We promise we will do any TX offloads before passing packets to /dev/null
> > +	 */
> > +	dev_info->tx_offload_capa = DEV_TX_OFFLOAD_VLAN_INSERT |
> > +		DEV_TX_OFFLOAD_IPV4_CKSUM | DEV_TX_OFFLOAD_UDP_CKSUM |
> > +		DEV_TX_OFFLOAD_TCP_CKSUM | DEV_TX_OFFLOAD_SCTP_CKSUM |
> > +		DEV_TX_OFFLOAD_TCP_TSO | DEV_TX_OFFLOAD_UDP_TSO |
> > +		DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM | DEV_TX_OFFLOAD_QINQ_INSERT |
> > +		DEV_TX_OFFLOAD_VXLAN_TNL_TSO | DEV_TX_OFFLOAD_GRE_TNL_TSO |
> > +		DEV_TX_OFFLOAD_IPIP_TNL_TSO | DEV_TX_OFFLOAD_GENEVE_TNL_TSO;
> 
> What do you think reporting offload capa based on user provided devargs,
> as suggested by Konstantin?And by default not report any.

I can see use of it for people that want to test software fallbacks for
their app. For those who need only a /dev/null port, it doesn't add any value.

Best Regards,
Michał Mirosław