[dpdk-dev] [PATCH v2] net/ixgbe: Restore vlan filter/extend for ixgbevf

[Zhao1, Wei]

> -----Original Message-----
> From: Ananyev, Konstantin
> Sent: Friday, March 15, 2019 9:00 PM
> To: Zhao1, Wei <wei.zhao1 at intel.com>; Zhang, Qi Z <qi.z.zhang at intel.com>;
> David Harton (dharton) <dharton at cisco.com>
> Cc: Lu, Wenzhuo <wenzhuo.lu at intel.com>; dev at dpdk.org
> Subject: RE: [PATCH v2] net/ixgbe: Restore vlan filter/extend for ixgbevf
> 
> Hi Wei,
> 
> > > > > > > >
> > > > > > > > Hi Qi,
> > > > > > > >
> > > > > > > > This patch was rejected with no feedback.
> > > > > > > >
> > > > > > > > Can you share why?
> > > > > > >
> > > > > > > Because as Zhao Wei explained in v1 thread, in
> > > > > > > ixgbe_vlan_offload_config , ETH_VLAN_FILTER_MASK and
> > > > > > > DEV_RX_OFFLOAD_VLAN_EXTEND are ignored This looks like
> > > > > VLAN_FILTER
> > > > > > AND
> > > > > > > VLAN_EXTEND offload is not supported and I didn't see you
> > > > > > > reply for that, so I
> > > > > > assume you agree with that.
> > > > > > >
> > > > > > >
> > > > > > > >
> > > > > > > > Without this patch users are not able to call the
> > > > > > > > rte_eth_dev_vlan_filter() API for ixgbevf based ports
> > > > > > > > because VLAN_FILTER cannot be enabled when calling
> > > rte_eth_dev_configure().
> > > > > > >
> > > > > > > OK, seems vlan filter is able to be configured for ixgbevf,
> > > > > > > but I don't know how this will work because vlan filter
> > > > > > > table needs to be enabled first
> > > > > > > (ixgbe_vlan_hw_filter_enable) So at least when it work with
> > > > > > > a DPDK PF, vlan filter will not work, maybe kernel PF driver
> > > > > > > will enable it by default? Could you share more
> > > insight?
> > > > > >
> > > > > > I can't answer that but ixgbevf_vlan_filter_set() has been
> > > > > > implemented for quite some time and it can no longer be called
> > > > > > since the new offload flags are being enforced now.
> > > > > >
> > > > > > I can share that
> > > > > > rte_eth_dev_vlan_filter()/ixgbevf_vlan_filter_set()
> > > > > > works properly with the propose change (VLAN traffic flows)
> > > > > > where it doesn't without it.
> > > > >
> > > > > Ok, if it is proved to work well with kernel PF, I think we
> > > > > should have this offload
> > > > > >
> > > > > > >
> > > > > > > And same question to VLAN_EXTEND
> > > > > >
> > > > > > I actually don't care about extend but simply changed it
> > > > > > because it seems to be in the same situation.  I'm happy to
> > > > > > leave it  out of the patch if that grants acceptance.
> > > > >
> > > > > Seems VLAN Extend is only supported by DPDK PF driver, I didn't
> > > > > see any related configuration in kernel driver.
> > > > > Properly we need to remove it.
> > > > >
> > > > > @Wei, would you help to confirm on that?
> > > >
> > > >
> > > > Yes, you are right, kernel pf do not has any code to set bit of
> > > IXGBE_DMATXCTL_GDV which enable double valn.
> > > > So, that means (kernel pf host + dpdk vf) can not enable double
> > > > valn
> > > feature. It seems only dpdk pf host + vf can do that.
> > >
> > > Wonder can vf request pf what is supported?
> > > Konstantin
> >
> > No such type of message defined in mailbox by now.
> 
> My thought was try to execute command (set vlan) with some dummy value,
> and analyze response to decide do we provide such capability or not.
> Or is too much hassle?

Maybe hassle to certain extent for implement, I think.


> 
> >
> > >
> > >
> > > >
> > > > Maybe we can change code to
> > > > "
> > > > 	offloads |= DEV_RX_OFFLOAD_VLAN_FILTER;
> > > >
> > > > 	if (ixgbe_is_vf(dev) == 0)
> > > > 		offloads |= DEV_RX_OFFLOAD_VLAN_EXTEND; "
> > > > But that means when in ENV (kernel pf host + dpdk vf),  this is
> > > > much
> > > depended on kernel pf default configuration.
> > > >
> > > >
> > > >
> > > > >
> > > > > Thanks
> > > > > Qi
> > > > >
> > > > > >
> > > > > > Thanks,
> > > > > > Dave
> > > > > >
> > > > > > >
> > > > > > >
> > > > > > > >
> > > > > > > > Or, am I missing something?
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > > Dave
> > > > > > > >
> > > > > > > > > -----Original Message-----
> > > > > > > > > From: David Harton <dharton at cisco.com>
> > > > > > > > > Sent: Friday, March 08, 2019 3:35 PM
> > > > > > > > > To: dev at dpdk.org
> > > > > > > > > Cc: wenzhuo.lu at intel.com; konstantin.ananyev at intel.com;
> > > > > > > > > qi.z.zhang at intel.com; David Harton (dharton)
> > > > > > > > > <dharton at cisco.com>
> > > > > > > > > Subject: [PATCH v2] net/ixgbe: Restore vlan
> > > > > > > > > filter/extend for ixgbevf
> > > > > > > > >
> > > > > > > > > ixgbevf vlan strip and extend capabilities were removed
> > > > > > > > > when migrating to the bit flags implementation.
> > > > > > > > >
> > > > > > > > > Restoring the capbility to enable these offloads at
> > > > > > > > > configuration
> > > > > > > time.
> > > > > > > > >
> > > > > > > > > Fixes: ec3b1124d14d (\"net/ixgbe: convert to new Rx
> > > > > > > > > offloads
> > > > > > > > > API\")
> > > > > > > > > Signed-off-by: David Harton <dharton at cisco.com>
> > > > > > > > > ---
> > > > > > > > >
> > > > > > > > > v2: removed unused function ixgbe_is_vf()
> > > > > > > > >
> > > > > > > > >  drivers/net/ixgbe/ixgbe_rxtx.c | 23
> > > > > > > > > ++---------------------
> > > > > > > > >  1 file changed, 2 insertions(+), 21 deletions(-)
> > > > > > > > >
> > > > > > > > > diff --git a/drivers/net/ixgbe/ixgbe_rxtx.c
> > > > > > > > > b/drivers/net/ixgbe/ixgbe_rxtx.c index
> > > > > > > > > e92a70fb3..b1b83613e
> > > > > > > > > 100644
> > > > > > > > > --- a/drivers/net/ixgbe/ixgbe_rxtx.c
> > > > > > > > > +++ b/drivers/net/ixgbe/ixgbe_rxtx.c
> > > > > > > > > @@ -2813,23 +2813,6 @@ ixgbe_reset_rx_queue(struct
> > > > > ixgbe_adapter
> > > > > > > > > *adapter, struct ixgbe_rx_queue *rxq)  #endif  }
> > > > > > > > >
> > > > > > > > > -static int
> > > > > > > > > -ixgbe_is_vf(struct rte_eth_dev *dev) -{
> > > > > > > > > -	struct ixgbe_hw *hw = IXGBE_DEV_PRIVATE_TO_HW(dev-
> > > > > >data-
> > > > > > > > > >dev_private);
> > > > > > > > > -
> > > > > > > > > -	switch (hw->mac.type) {
> > > > > > > > > -	case ixgbe_mac_82599_vf:
> > > > > > > > > -	case ixgbe_mac_X540_vf:
> > > > > > > > > -	case ixgbe_mac_X550_vf:
> > > > > > > > > -	case ixgbe_mac_X550EM_x_vf:
> > > > > > > > > -	case ixgbe_mac_X550EM_a_vf:
> > > > > > > > > -		return 1;
> > > > > > > > > -	default:
> > > > > > > > > -		return 0;
> > > > > > > > > -	}
> > > > > > > > > -}
> > > > > > > > > -
> > > > > > > > >  uint64_t
> > > > > > > > >  ixgbe_get_rx_queue_offloads(struct rte_eth_dev *dev)  {
> > > > > > > > > @@
> > > > > > > > > -2853,15
> > > > > > > > > +2836,13 @@ ixgbe_get_rx_port_offloads(struct
> > > > > > > > > +rte_eth_dev
> > > > > > > > > +*dev)
> > > > > > > > >  		   DEV_RX_OFFLOAD_TCP_CKSUM   |
> > > > > > > > >  		   DEV_RX_OFFLOAD_KEEP_CRC    |
> > > > > > > > >  		   DEV_RX_OFFLOAD_JUMBO_FRAME |
> > > > > > > > > +		   DEV_RX_OFFLOAD_VLAN_FILTER |
> > > > > > > > > +		   DEV_RX_OFFLOAD_VLAN_EXTEND |
> > > > > > > > >  		   DEV_RX_OFFLOAD_SCATTER;
> > > > > > > > >
> > > > > > > > >  	if (hw->mac.type == ixgbe_mac_82598EB)
> > > > > > > > >  		offloads |= DEV_RX_OFFLOAD_VLAN_STRIP;
> > > > > > > > >
> > > > > > > > > -	if (ixgbe_is_vf(dev) == 0)
> > > > > > > > > -		offloads |= (DEV_RX_OFFLOAD_VLAN_FILTER |
> > > > > > > > > -			     DEV_RX_OFFLOAD_VLAN_EXTEND);
> > > > > > > > > -
> > > > > > > > >  	/*
> > > > > > > > >  	 * RSC is only supported by 82599 and x540 PF devices
> > > > > > > > > in a
> > > > > > > > > non-SR- IOV
> > > > > > > > >  	 * mode.
> > > > > > > > > --
> > > > > > > > > 2.19.1