| Message ID | 20200210191850.240247-1-fbl@sysclose.org |
|---|---|
| State | Changes Requested |
| Delegated to: | Ilya Maximets |
| Headers | show |
| Series | [ovs-dev] netdev-dpdk: vhost: disable unsupported offload features. | expand |
On 2/10/20 8:18 PM, Flavio Leitner wrote: > Disable ECN and UFO since this is not supported yet. Also, disable > all other features when userspace_tso is not enabled. Disabling UFO saves us from receiving of a big unfragmented UDP packets, but UDP packets will still have bad checksums and will be dropped by the kernel on receive side. So, this is only part of a fix. > > Fixes: 29cf9c1b3b9c ("userspace: Add TCP Segmentation Offload support") > Signed-off-by: Flavio Leitner <fbl@sysclose.org> > --- > lib/netdev-dpdk.c | 25 ++++++++++++++++--------- > 1 file changed, 16 insertions(+), 9 deletions(-) > > > This should go to branch-2.13 if passes the review. > > > diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c > index 6187129c0..b3983b312 100644 > --- a/lib/netdev-dpdk.c > +++ b/lib/netdev-dpdk.c > @@ -5186,6 +5186,7 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) > struct netdev_dpdk *dev = netdev_dpdk_cast(netdev); > int err; > uint64_t vhost_flags = 0; > + uint64_t vhost_unsup_flags; > bool zc_enabled; > > ovs_mutex_lock(&dev->mutex); > @@ -5252,16 +5253,22 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) > netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_TSO; > netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_CKSUM; > netdev->ol_flags |= NETDEV_TX_OFFLOAD_IPV4_CKSUM; > + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_HOST_ECN > + | 1ULL << VIRTIO_NET_F_HOST_UFO; > } else { > - err = rte_vhost_driver_disable_features(dev->vhost_id, > - 1ULL << VIRTIO_NET_F_HOST_TSO4 > - | 1ULL << VIRTIO_NET_F_HOST_TSO6 > - | 1ULL << VIRTIO_NET_F_CSUM); > - if (err) { > - VLOG_ERR("rte_vhost_driver_disable_features failed for " > - "vhost user client port: %s\n", dev->up.name); > - goto unlock; > - } > + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_CSUM > + | 1ULL << VIRTIO_NET_F_HOST_UFO > + | 1ULL << VIRTIO_NET_F_HOST_TSO4 > + | 1ULL << VIRTIO_NET_F_HOST_TSO6 > + | 1ULL << VIRTIO_NET_F_HOST_ECN; All above features depends on VIRTIO_NET_F_CSUM, so, I think, we only need to disable this one. > + } > + > + err = rte_vhost_driver_disable_features(dev->vhost_id, > + vhost_unsup_flags); > + if (err) { > + VLOG_ERR("rte_vhost_driver_disable_features failed for " > + "vhost user client port: %s\n", dev->up.name); > + goto unlock; > } > > err = rte_vhost_driver_start(dev->vhost_id); >
On Wed, Feb 12, 2020 at 04:26:46PM +0100, Ilya Maximets wrote: > On 2/10/20 8:18 PM, Flavio Leitner wrote: > > Disable ECN and UFO since this is not supported yet. Also, disable > > all other features when userspace_tso is not enabled. > > Disabling UFO saves us from receiving of a big unfragmented > UDP packets, but UDP packets will still have bad checksums > and will be dropped by the kernel on receive side. > So, this is only part of a fix. Yes, this patch is not intended to fix the UDP checksum problem. Unless you see a way to do that with changing the flags below. > > Fixes: 29cf9c1b3b9c ("userspace: Add TCP Segmentation Offload support") > > Signed-off-by: Flavio Leitner <fbl@sysclose.org> > > --- > > lib/netdev-dpdk.c | 25 ++++++++++++++++--------- > > 1 file changed, 16 insertions(+), 9 deletions(-) > > > > > > This should go to branch-2.13 if passes the review. > > > > > > diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c > > index 6187129c0..b3983b312 100644 > > --- a/lib/netdev-dpdk.c > > +++ b/lib/netdev-dpdk.c > > @@ -5186,6 +5186,7 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) > > struct netdev_dpdk *dev = netdev_dpdk_cast(netdev); > > int err; > > uint64_t vhost_flags = 0; > > + uint64_t vhost_unsup_flags; > > bool zc_enabled; > > > > ovs_mutex_lock(&dev->mutex); > > @@ -5252,16 +5253,22 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) > > netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_TSO; > > netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_CKSUM; > > netdev->ol_flags |= NETDEV_TX_OFFLOAD_IPV4_CKSUM; > > + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_HOST_ECN > > + | 1ULL << VIRTIO_NET_F_HOST_UFO; > > } else { > > - err = rte_vhost_driver_disable_features(dev->vhost_id, > > - 1ULL << VIRTIO_NET_F_HOST_TSO4 > > - | 1ULL << VIRTIO_NET_F_HOST_TSO6 > > - | 1ULL << VIRTIO_NET_F_CSUM); > > - if (err) { > > - VLOG_ERR("rte_vhost_driver_disable_features failed for " > > - "vhost user client port: %s\n", dev->up.name); > > - goto unlock; > > - } > > + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_CSUM > > + | 1ULL << VIRTIO_NET_F_HOST_UFO > > + | 1ULL << VIRTIO_NET_F_HOST_TSO4 > > + | 1ULL << VIRTIO_NET_F_HOST_TSO6 > > + | 1ULL << VIRTIO_NET_F_HOST_ECN; > > All above features depends on VIRTIO_NET_F_CSUM, so, I think, we only > need to disable this one. You right. I added the others flags anyways to make it explicit. I thought either a comment or code would be needed. Since I see people grep'ing for the flags, there you go. fbl > > > + } > > + > > + err = rte_vhost_driver_disable_features(dev->vhost_id, > > + vhost_unsup_flags); > > + if (err) { > > + VLOG_ERR("rte_vhost_driver_disable_features failed for " > > + "vhost user client port: %s\n", dev->up.name); > > + goto unlock; > > } > > > > err = rte_vhost_driver_start(dev->vhost_id); > > > > _______________________________________________ > dev mailing list > dev@openvswitch.org > https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Wed, Feb 12, 2020 at 04:31:09PM -0300, Flavio Leitner wrote: > On Wed, Feb 12, 2020 at 04:26:46PM +0100, Ilya Maximets wrote: > > On 2/10/20 8:18 PM, Flavio Leitner wrote: > > > Disable ECN and UFO since this is not supported yet. Also, disable > > > all other features when userspace_tso is not enabled. > > > > Disabling UFO saves us from receiving of a big unfragmented > > UDP packets, but UDP packets will still have bad checksums > > and will be dropped by the kernel on receive side. > > So, this is only part of a fix. > > Yes, this patch is not intended to fix the UDP checksum problem. > Unless you see a way to do that with changing the flags below. BTW, why do you think there is a problem with UDP checksum? We set VIRTIO_NET_HDR_F_NEEDS_CSUM which should be all it is needed.
On 2/13/20 12:44 AM, Flavio Leitner wrote: > On Wed, Feb 12, 2020 at 04:31:09PM -0300, Flavio Leitner wrote: >> On Wed, Feb 12, 2020 at 04:26:46PM +0100, Ilya Maximets wrote: >>> On 2/10/20 8:18 PM, Flavio Leitner wrote: >>>> Disable ECN and UFO since this is not supported yet. Also, disable >>>> all other features when userspace_tso is not enabled. >>> >>> Disabling UFO saves us from receiving of a big unfragmented >>> UDP packets, but UDP packets will still have bad checksums >>> and will be dropped by the kernel on receive side. >>> So, this is only part of a fix. >> >> Yes, this patch is not intended to fix the UDP checksum problem. >> Unless you see a way to do that with changing the flags below. > > BTW, why do you think there is a problem with UDP checksum? > We set VIRTIO_NET_HDR_F_NEEDS_CSUM which should be all it is needed. At least we're sending UDP packets on a wire without requesting Tx UDP checksum offloading from HW NICs.
On Thu, Feb 13, 2020 at 11:10:15AM +0100, Ilya Maximets wrote: > On 2/13/20 12:44 AM, Flavio Leitner wrote: > > On Wed, Feb 12, 2020 at 04:31:09PM -0300, Flavio Leitner wrote: > >> On Wed, Feb 12, 2020 at 04:26:46PM +0100, Ilya Maximets wrote: > >>> On 2/10/20 8:18 PM, Flavio Leitner wrote: > >>>> Disable ECN and UFO since this is not supported yet. Also, disable > >>>> all other features when userspace_tso is not enabled. > >>> > >>> Disabling UFO saves us from receiving of a big unfragmented > >>> UDP packets, but UDP packets will still have bad checksums > >>> and will be dropped by the kernel on receive side. > >>> So, this is only part of a fix. > >> > >> Yes, this patch is not intended to fix the UDP checksum problem. > >> Unless you see a way to do that with changing the flags below. > > > > BTW, why do you think there is a problem with UDP checksum? > > We set VIRTIO_NET_HDR_F_NEEDS_CSUM which should be all it is needed. > > At least we're sending UDP packets on a wire without > requesting Tx UDP checksum offloading from HW NICs. Packets from virtual machines relies on DPDK vhost library, and as far as I can see, it does set PKT_TX_UDP_CKSUM properly. Now for packets coming from the kernel, then netdev_linux_parse_vnet_hdr() will set that as well. Packets from userspace to kernel are properly translated by netdev_linux_prepend_vnet_hdr().
On 2/13/20 12:30 PM, Flavio Leitner wrote: > On Thu, Feb 13, 2020 at 11:10:15AM +0100, Ilya Maximets wrote: >> On 2/13/20 12:44 AM, Flavio Leitner wrote: >>> On Wed, Feb 12, 2020 at 04:31:09PM -0300, Flavio Leitner wrote: >>>> On Wed, Feb 12, 2020 at 04:26:46PM +0100, Ilya Maximets wrote: >>>>> On 2/10/20 8:18 PM, Flavio Leitner wrote: >>>>>> Disable ECN and UFO since this is not supported yet. Also, disable >>>>>> all other features when userspace_tso is not enabled. >>>>> >>>>> Disabling UFO saves us from receiving of a big unfragmented >>>>> UDP packets, but UDP packets will still have bad checksums >>>>> and will be dropped by the kernel on receive side. >>>>> So, this is only part of a fix. >>>> >>>> Yes, this patch is not intended to fix the UDP checksum problem. >>>> Unless you see a way to do that with changing the flags below. >>> >>> BTW, why do you think there is a problem with UDP checksum? >>> We set VIRTIO_NET_HDR_F_NEEDS_CSUM which should be all it is needed. >> >> At least we're sending UDP packets on a wire without >> requesting Tx UDP checksum offloading from HW NICs. > > Packets from virtual machines relies on DPDK vhost library, and as > far as I can see, it does set PKT_TX_UDP_CKSUM properly. vhost library will properly set flags, however, OVS doesn't care about this case, e.g. we're not enabling DEV_TX_OFFLOAD_UDP_CKSUM while configuring physical devices. Even though most of the DPDK drivers will enable UDP checksum offload support regardless of this flag, this is still a logical error and we can't rely on this behavior. > > Now for packets coming from the kernel, then netdev_linux_parse_vnet_hdr() > will set that as well. > > Packets from userspace to kernel are properly translated by > netdev_linux_prepend_vnet_hdr(). >
On Thu, Feb 13, 2020 at 01:42:32PM +0100, Ilya Maximets wrote: > On 2/13/20 12:30 PM, Flavio Leitner wrote: > > On Thu, Feb 13, 2020 at 11:10:15AM +0100, Ilya Maximets wrote: > >> On 2/13/20 12:44 AM, Flavio Leitner wrote: > >>> On Wed, Feb 12, 2020 at 04:31:09PM -0300, Flavio Leitner wrote: > >>>> On Wed, Feb 12, 2020 at 04:26:46PM +0100, Ilya Maximets wrote: > >>>>> On 2/10/20 8:18 PM, Flavio Leitner wrote: > >>>>>> Disable ECN and UFO since this is not supported yet. Also, disable > >>>>>> all other features when userspace_tso is not enabled. > >>>>> > >>>>> Disabling UFO saves us from receiving of a big unfragmented > >>>>> UDP packets, but UDP packets will still have bad checksums > >>>>> and will be dropped by the kernel on receive side. > >>>>> So, this is only part of a fix. > >>>> > >>>> Yes, this patch is not intended to fix the UDP checksum problem. > >>>> Unless you see a way to do that with changing the flags below. > >>> > >>> BTW, why do you think there is a problem with UDP checksum? > >>> We set VIRTIO_NET_HDR_F_NEEDS_CSUM which should be all it is needed. > >> > >> At least we're sending UDP packets on a wire without > >> requesting Tx UDP checksum offloading from HW NICs. > > > > Packets from virtual machines relies on DPDK vhost library, and as > > far as I can see, it does set PKT_TX_UDP_CKSUM properly. > > vhost library will properly set flags, however, OVS doesn't care > about this case, e.g. we're not enabling DEV_TX_OFFLOAD_UDP_CKSUM > while configuring physical devices. Even though most of the DPDK > drivers will enable UDP checksum offload support regardless of this > flag, this is still a logical error and we can't rely on this > behavior. Ah, now I see what you're saying. I think we can support UDP csum, so will post another separate patch fixing it. Thanks,
On 2/12/20 8:31 PM, Flavio Leitner wrote: > On Wed, Feb 12, 2020 at 04:26:46PM +0100, Ilya Maximets wrote: >> On 2/10/20 8:18 PM, Flavio Leitner wrote: >>> Disable ECN and UFO since this is not supported yet. Also, disable >>> all other features when userspace_tso is not enabled. >> >> Disabling UFO saves us from receiving of a big unfragmented >> UDP packets, but UDP packets will still have bad checksums >> and will be dropped by the kernel on receive side. >> So, this is only part of a fix. > > Yes, this patch is not intended to fix the UDP checksum problem. > Unless you see a way to do that with changing the flags below. > >>> Fixes: 29cf9c1b3b9c ("userspace: Add TCP Segmentation Offload support") >>> Signed-off-by: Flavio Leitner <fbl@sysclose.org> >>> --- >>> lib/netdev-dpdk.c | 25 ++++++++++++++++--------- >>> 1 file changed, 16 insertions(+), 9 deletions(-) >>> >>> >>> This should go to branch-2.13 if passes the review. >>> >>> >>> diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c >>> index 6187129c0..b3983b312 100644 >>> --- a/lib/netdev-dpdk.c >>> +++ b/lib/netdev-dpdk.c >>> @@ -5186,6 +5186,7 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) >>> struct netdev_dpdk *dev = netdev_dpdk_cast(netdev); >>> int err; >>> uint64_t vhost_flags = 0; >>> + uint64_t vhost_unsup_flags; >>> bool zc_enabled; >>> >>> ovs_mutex_lock(&dev->mutex); >>> @@ -5252,16 +5253,22 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) >>> netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_TSO; >>> netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_CKSUM; >>> netdev->ol_flags |= NETDEV_TX_OFFLOAD_IPV4_CKSUM; >>> + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_HOST_ECN >>> + | 1ULL << VIRTIO_NET_F_HOST_UFO; >>> } else { >>> - err = rte_vhost_driver_disable_features(dev->vhost_id, >>> - 1ULL << VIRTIO_NET_F_HOST_TSO4 >>> - | 1ULL << VIRTIO_NET_F_HOST_TSO6 >>> - | 1ULL << VIRTIO_NET_F_CSUM); >>> - if (err) { >>> - VLOG_ERR("rte_vhost_driver_disable_features failed for " >>> - "vhost user client port: %s\n", dev->up.name); >>> - goto unlock; >>> - } >>> + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_CSUM >>> + | 1ULL << VIRTIO_NET_F_HOST_UFO >>> + | 1ULL << VIRTIO_NET_F_HOST_TSO4 >>> + | 1ULL << VIRTIO_NET_F_HOST_TSO6 >>> + | 1ULL << VIRTIO_NET_F_HOST_ECN; >> >> All above features depends on VIRTIO_NET_F_CSUM, so, I think, we only >> need to disable this one. > > You right. I added the others flags anyways to make it explicit. I thought > either a comment or code would be needed. Since I see people grep'ing > for the flags, there you go. I think, it's better to add a comment like this: /* This disables checksum offloading and all the features * that depends on it (TSO, UFO, ECN) according to virtio * specification. */ vhost_unsup_flags = 1ULL << VIRTIO_NET_F_CSUM; > > fbl > >> >>> + } >>> + >>> + err = rte_vhost_driver_disable_features(dev->vhost_id, >>> + vhost_unsup_flags); >>> + if (err) { >>> + VLOG_ERR("rte_vhost_driver_disable_features failed for " >>> + "vhost user client port: %s\n", dev->up.name); >>> + goto unlock; >>> } >>> >>> err = rte_vhost_driver_start(dev->vhost_id); >>> >> >> _______________________________________________ >> dev mailing list >> dev@openvswitch.org >> https://mail.openvswitch.org/mailman/listinfo/ovs-dev >
On Thu, Feb 13, 2020 at 04:18:35PM +0100, Ilya Maximets wrote: > On 2/12/20 8:31 PM, Flavio Leitner wrote: > > On Wed, Feb 12, 2020 at 04:26:46PM +0100, Ilya Maximets wrote: > >> On 2/10/20 8:18 PM, Flavio Leitner wrote: > >>> Disable ECN and UFO since this is not supported yet. Also, disable > >>> all other features when userspace_tso is not enabled. > >> > >> Disabling UFO saves us from receiving of a big unfragmented > >> UDP packets, but UDP packets will still have bad checksums > >> and will be dropped by the kernel on receive side. > >> So, this is only part of a fix. > > > > Yes, this patch is not intended to fix the UDP checksum problem. > > Unless you see a way to do that with changing the flags below. > > > >>> Fixes: 29cf9c1b3b9c ("userspace: Add TCP Segmentation Offload support") > >>> Signed-off-by: Flavio Leitner <fbl@sysclose.org> > >>> --- > >>> lib/netdev-dpdk.c | 25 ++++++++++++++++--------- > >>> 1 file changed, 16 insertions(+), 9 deletions(-) > >>> > >>> > >>> This should go to branch-2.13 if passes the review. > >>> > >>> > >>> diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c > >>> index 6187129c0..b3983b312 100644 > >>> --- a/lib/netdev-dpdk.c > >>> +++ b/lib/netdev-dpdk.c > >>> @@ -5186,6 +5186,7 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) > >>> struct netdev_dpdk *dev = netdev_dpdk_cast(netdev); > >>> int err; > >>> uint64_t vhost_flags = 0; > >>> + uint64_t vhost_unsup_flags; > >>> bool zc_enabled; > >>> > >>> ovs_mutex_lock(&dev->mutex); > >>> @@ -5252,16 +5253,22 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) > >>> netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_TSO; > >>> netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_CKSUM; > >>> netdev->ol_flags |= NETDEV_TX_OFFLOAD_IPV4_CKSUM; > >>> + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_HOST_ECN > >>> + | 1ULL << VIRTIO_NET_F_HOST_UFO; > >>> } else { > >>> - err = rte_vhost_driver_disable_features(dev->vhost_id, > >>> - 1ULL << VIRTIO_NET_F_HOST_TSO4 > >>> - | 1ULL << VIRTIO_NET_F_HOST_TSO6 > >>> - | 1ULL << VIRTIO_NET_F_CSUM); > >>> - if (err) { > >>> - VLOG_ERR("rte_vhost_driver_disable_features failed for " > >>> - "vhost user client port: %s\n", dev->up.name); > >>> - goto unlock; > >>> - } > >>> + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_CSUM > >>> + | 1ULL << VIRTIO_NET_F_HOST_UFO > >>> + | 1ULL << VIRTIO_NET_F_HOST_TSO4 > >>> + | 1ULL << VIRTIO_NET_F_HOST_TSO6 > >>> + | 1ULL << VIRTIO_NET_F_HOST_ECN; > >> > >> All above features depends on VIRTIO_NET_F_CSUM, so, I think, we only > >> need to disable this one. > > > > You right. I added the others flags anyways to make it explicit. I thought > > either a comment or code would be needed. Since I see people grep'ing > > for the flags, there you go. > > I think, it's better to add a comment like this: > > /* This disables checksum offloading and all the features > * that depends on it (TSO, UFO, ECN) according to virtio > * specification. */ > vhost_unsup_flags = 1ULL << VIRTIO_NET_F_CSUM; I will respin the patch as part of series fixing csum as well. Thanks,
diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c index 6187129c0..b3983b312 100644 --- a/lib/netdev-dpdk.c +++ b/lib/netdev-dpdk.c @@ -5186,6 +5186,7 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) struct netdev_dpdk *dev = netdev_dpdk_cast(netdev); int err; uint64_t vhost_flags = 0; + uint64_t vhost_unsup_flags; bool zc_enabled; ovs_mutex_lock(&dev->mutex); @@ -5252,16 +5253,22 @@ netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev) netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_TSO; netdev->ol_flags |= NETDEV_TX_OFFLOAD_TCP_CKSUM; netdev->ol_flags |= NETDEV_TX_OFFLOAD_IPV4_CKSUM; + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_HOST_ECN + | 1ULL << VIRTIO_NET_F_HOST_UFO; } else { - err = rte_vhost_driver_disable_features(dev->vhost_id, - 1ULL << VIRTIO_NET_F_HOST_TSO4 - | 1ULL << VIRTIO_NET_F_HOST_TSO6 - | 1ULL << VIRTIO_NET_F_CSUM); - if (err) { - VLOG_ERR("rte_vhost_driver_disable_features failed for " - "vhost user client port: %s\n", dev->up.name); - goto unlock; - } + vhost_unsup_flags = 1ULL << VIRTIO_NET_F_CSUM + | 1ULL << VIRTIO_NET_F_HOST_UFO + | 1ULL << VIRTIO_NET_F_HOST_TSO4 + | 1ULL << VIRTIO_NET_F_HOST_TSO6 + | 1ULL << VIRTIO_NET_F_HOST_ECN; + } + + err = rte_vhost_driver_disable_features(dev->vhost_id, + vhost_unsup_flags); + if (err) { + VLOG_ERR("rte_vhost_driver_disable_features failed for " + "vhost user client port: %s\n", dev->up.name); + goto unlock; } err = rte_vhost_driver_start(dev->vhost_id);
Disable ECN and UFO since this is not supported yet. Also, disable all other features when userspace_tso is not enabled. Fixes: 29cf9c1b3b9c ("userspace: Add TCP Segmentation Offload support") Signed-off-by: Flavio Leitner <fbl@sysclose.org> --- lib/netdev-dpdk.c | 25 ++++++++++++++++--------- 1 file changed, 16 insertions(+), 9 deletions(-) This should go to branch-2.13 if passes the review.