From mboxrd@z Thu Jan 1 00:00:00 1970 From: Joe Stringer Subject: Re: [PATCH net 3/5] fm10k: Implement ndo_gso_check() Date: Thu, 6 Nov 2014 13:15:22 -0800 Message-ID: <20141106211522.GA55313@gmail.com> References: <1415138202-1197-1-git-send-email-joestringer@nicira.com> <1415138202-1197-4-git-send-email-joestringer@nicira.com> <545AE2C8.3070705@gmail.com> <20141106184115.GB18339@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Alexander Duyck , sathya.perla@emulex.com, jeffrey.t.kirsher@intel.com, linux.nics@intel.com, amirv@mellanox.com, shahed.shaikh@qlogic.com, therbert@google.com To: alexander.duyck@gmail.com, netdev@vger.kernel.org, Dept-GELinuxNICDev@qlogic.com, linux-kernel@vger.kernel.org Return-path: Received: from na3sys009aog128.obsmtp.com ([74.125.149.141]:58253 "HELO na3sys009aog128.obsmtp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1750828AbaKFVPj (ORCPT ); Thu, 6 Nov 2014 16:15:39 -0500 Received: by mail-pd0-f173.google.com with SMTP id v10so1880590pde.18 for ; Thu, 06 Nov 2014 13:15:38 -0800 (PST) Content-Disposition: inline In-Reply-To: <20141106184115.GB18339@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: On Thu, Nov 06, 2014 at 10:41:15AM -0800, Joe Stringer wrote: > On Wed, Nov 05, 2014 at 06:54:00PM -0800, Alexander Duyck wrote: > > On 11/04/2014 01:56 PM, Joe Stringer wrote: > > > ndo_gso_check() was recently introduced to allow NICs to report t= he > > > offloading support that they have on a per-skb basis. Add an > > > implementation for this driver which checks for something that lo= oks > > > like VXLAN. > > > > > > Implementation shamelessly stolen from Tom Herbert: > > > http://thread.gmane.org/gmane.linux.network/332428/focus=3D333111 > > > > > > Signed-off-by: Joe Stringer > > > --- > > > Should this driver report support for GSO on packets with tunnel = headers > > > up to 64B like the i40e driver does? > > > --- > > > drivers/net/ethernet/intel/fm10k/fm10k_netdev.c | 12 +++++++++= +++ > > > 1 file changed, 12 insertions(+) > > > > > > diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c b/dr= ivers/net/ethernet/intel/fm10k/fm10k_netdev.c > > > index 8811364..b9ef622 100644 > > > --- a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c > > > +++ b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c > > > @@ -1350,6 +1350,17 @@ static void fm10k_dfwd_del_station(struct = net_device *dev, void *priv) > > > } > > > } > > > =20 > > > +static bool fm10k_gso_check(struct sk_buff *skb, struct net_devi= ce *dev) > > > +{ > > > + if ((skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL) && > > > + (skb->inner_protocol_type !=3D ENCAP_TYPE_ETHER || > > > + skb->inner_protocol !=3D htons(ETH_P_TEB) || > > > + skb_inner_mac_header(skb) - skb_transport_header(skb) !=3D= 16)) > > > + return false; > > > + > > > + return true; > > > +} > > > + > > > static const struct net_device_ops fm10k_netdev_ops =3D { > > > .ndo_open =3D fm10k_open, > > > .ndo_stop =3D fm10k_close, > > > @@ -1372,6 +1383,7 @@ static const struct net_device_ops fm10k_ne= tdev_ops =3D { > > > .ndo_do_ioctl =3D fm10k_ioctl, > > > .ndo_dfwd_add_station =3D fm10k_dfwd_add_station, > > > .ndo_dfwd_del_station =3D fm10k_dfwd_del_station, > > > + .ndo_gso_check =3D fm10k_gso_check, > > > }; > > > =20 > > > #define DEFAULT_DEBUG_LEVEL_SHIFT 3 > >=20 > > I'm thinking this check is far too simplistic. If you look the fm1= 0k > > driver already has fm10k_tx_encap_offload() in the TSO function for > > verifying if it can support offloading tunnels or not. I would > > recommend starting there or possibly even just adapting that functi= on to > > suit your purpose. > >=20 > > Thanks, > >=20 > > Alex >=20 > Would it be enough to just call fm10k_tx_encap_offload() in a way tha= t echoes fm10k_tso()? >=20 > +static bool fm10k_gso_check(struct sk_buff *skb, struct net_device *= dev) > +{ > + =A0 =A0 =A0 if (skb->encapsulation && !fm10k_tx_encap_offload(skb)) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return false; > + > + =A0 =A0 =A0 return true; > +} Oh, I suppose we need to check the gso_type too. More like this? +static bool fm10k_gso_check(struct sk_buff *skb, struct net_device *de= v) +{ + if ((skb_shinfo(skb)->gso_type & (SKB_GSO_UDP_TUNNEL | SKB_GSO_= GRE)) && + !fm10k_tx_encap_offload(skb)) + return false; + + return true; +}