[PATCH V3 net 0/3] Fix checksum issues with Q-in-Q vlans

[PATCH V3 net 0/3] Fix checksum issues with Q-in-Q vlans

[Vladislav Yasevich]

TCP checksum appear broken on a lot of devices that
advertise NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM.  This problem
becomes very visible/reproducable since the series
commit afb0bc972b526 ("Merge branch 'stacked_vlan_tso'").

In particular, the issue appeared consistently on bnx2 and be2net
drivers (not all drivers were tested).

This short series corrects this by disabling checksum offload
support on packets sent through Q-in-Q vlans if the underlying HW only
enables IP specific checksum features.  We currently 'assume' that
any drivers setting NETIF_F_HW_CSUM can correclty pass checksum offsets
to HW.  It is up to individual drivers to enable it properly through
ndo_features_check if they have some support for Q-in-Q vlans.

Additionally, be2net driver was fixed to make the proper call.

While looking at the drivers, it was also found that virtio-net ended
up disabling accelerations, which is unnecessary.  

V3: Fixed checkpatch errors.

V2: Instead of disabling checksuming for all devices, only devices using
    NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM are now affected by this change.
    For drivers using NETIF_F_HW_CSUM, we will continue to use checksum
    offloading.  If any drivers are found to be broken, they would need
    be fixed individually.

Vladislav Yasevich (3):
  vlan: Fix tcp checksum offloads in Q-in-Q vlans
  be2net: Fix offload features for Q-in-Q packets
  virtio-net: enable TSO/checksum offloads for Q-in-Q vlans

 drivers/net/ethernet/emulex/benet/be_main.c |  4 +++-
 drivers/net/virtio_net.c                    |  1 +
 include/linux/if_vlan.h                     | 10 ++++++++--
 3 files changed, 12 insertions(+), 3 deletions(-)

-- 
2.7.4

[PATCH V3 net 1/3] vlan: Fix tcp checksum offloads in Q-in-Q vlans

[Vladislav Yasevich]

It appears that TCP checksum offloading has been broken for
Q-in-Q vlans.  The behavior was execerbated by the
series
    commit afb0bc972b52 ("Merge branch 'stacked_vlan_tso'")
that that enabled accleleration features on stacked vlans.

However, event without that series, it is possible to trigger
this issue.  It just requires a lot more specialized configuration.

The root cause is the interaction between how
netdev_intersect_features() works, the features actually set on
the vlan devices and HW having the ability to run checksum with
longer headers.

The issue starts when netdev_interesect_features() replaces
NETIF_F_HW_CSUM with a combination of NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM,
if the HW advertises IP|IPV6 specific checksums.  This happens
for tagged and multi-tagged packets.   However, HW that enables
IP|IPV6 checksum offloading doesn't gurantee that packets with
arbitrarily long headers can be checksummed.

This patch disables IP|IPV6 checksums on the packet for multi-tagged
packets.

CC: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
CC: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
 include/linux/if_vlan.h | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)

diff --git a/include/linux/if_vlan.h b/include/linux/if_vlan.h
index 8d5fcd6..6686d0f 100644
--- a/include/linux/if_vlan.h
+++ b/include/linux/if_vlan.h
@@ -614,14 +614,16 @@ static inline bool skb_vlan_tagged_multi(const struct sk_buff *skb)
 static inline netdev_features_t vlan_features_check(const struct sk_buff *skb,
 						    netdev_features_t features)
 {
-	if (skb_vlan_tagged_multi(skb))
-		features = netdev_intersect_features(features,
-						     NETIF_F_SG |
-						     NETIF_F_HIGHDMA |
-						     NETIF_F_FRAGLIST |
-						     NETIF_F_HW_CSUM |
-						     NETIF_F_HW_VLAN_CTAG_TX |
-						     NETIF_F_HW_VLAN_STAG_TX);
+	if (skb_vlan_tagged_multi(skb)) {
+		/* In the case of multi-tagged packets, use a direct mask
+		 * instead of using netdev_interesect_features(), to make
+		 * sure that only devices supporting NETIF_F_HW_CSUM will
+		 * have checksum offloading support.
+		 */
+		features &= NETIF_F_SG | NETIF_F_HIGHDMA | NETIF_F_HW_CSUM |
+			    NETIF_F_FRAGLIST | NETIF_F_HW_VLAN_CTAG_TX |
+			    NETIF_F_HW_VLAN_STAG_TX;
+	}
 
 	return features;
 }
-- 
2.7.4

[PATCH V3 net 2/3] be2net: Fix offload features for Q-in-Q packets

[Vladislav Yasevich]

At least some of the be2net cards do not seem to be capabled
of performing checksum offload computions on Q-in-Q packets.
In these case, the recevied checksum on the remote is invalid
and TCP syn packets are dropped.

This patch adds a call to check disbled acceleration features
on Q-in-Q tagged traffic.

CC: Sathya Perla <sathya.perla@broadcom.com>
CC: Ajit Khaparde <ajit.khaparde@broadcom.com>
CC: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com>
CC: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
 drivers/net/ethernet/emulex/benet/be_main.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index f3a09ab..4eee18c 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -5078,9 +5078,11 @@ static netdev_features_t be_features_check(struct sk_buff *skb,
 	struct be_adapter *adapter = netdev_priv(dev);
 	u8 l4_hdr = 0;
 
-	/* The code below restricts offload features for some tunneled packets.
+	/* The code below restricts offload features for some tunneled and
+	 * Q-in-Q packets.
 	 * Offload features for normal (non tunnel) packets are unchanged.
 	 */
+	features = vlan_features_check(skb, features);
 	if (!skb->encapsulation ||
 	    !(adapter->flags & BE_FLAGS_VXLAN_OFFLOADS))
 		return features;
-- 
2.7.4

[PATCH V3 net 3/3] virtio-net: enable TSO/checksum offloads for Q-in-Q vlans

[Vladislav Yasevich]

Since virtio does not provide it's own ndo_features_check handler,
TSO, and now checksum offload, are disabled for stacked vlans.
Re-enable the support and let the host take care of it.  This
restores/improves Guest-to-Guest performance over Q-in-Q vlans.

Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
 drivers/net/virtio_net.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 665627c..ead7a58 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -2020,6 +2020,7 @@ static const struct net_device_ops virtnet_netdev = {
 	.ndo_poll_controller = virtnet_netpoll,
 #endif
 	.ndo_xdp		= virtnet_xdp,
+	.ndo_features_check	= passthru_features_check,
 };
 
 static void virtnet_config_changed_work(struct work_struct *work)
-- 
2.7.4

Re: [PATCH V3 net 1/3] vlan: Fix tcp checksum offloads in Q-in-Q vlans

[Toshiaki Makita]

On 2017/05/24 2:38, Vladislav Yasevich wrote:
> It appears that TCP checksum offloading has been broken for
> Q-in-Q vlans.  The behavior was execerbated by the
> series
>     commit afb0bc972b52 ("Merge branch 'stacked_vlan_tso'")
> that that enabled accleleration features on stacked vlans.
> 
> However, event without that series, it is possible to trigger
> this issue.  It just requires a lot more specialized configuration.
> 
> The root cause is the interaction between how
> netdev_intersect_features() works, the features actually set on
> the vlan devices and HW having the ability to run checksum with
> longer headers.
> 
> The issue starts when netdev_interesect_features() replaces
> NETIF_F_HW_CSUM with a combination of NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM,
> if the HW advertises IP|IPV6 specific checksums.  This happens
> for tagged and multi-tagged packets.   However, HW that enables
> IP|IPV6 checksum offloading doesn't gurantee that packets with
> arbitrarily long headers can be checksummed.
> 
> This patch disables IP|IPV6 checksums on the packet for multi-tagged
> packets.
> 
> CC: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
> CC: Michal Kubecek <mkubecek@suse.cz>
> Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
> ---

Thank you for fixing it.
Acked-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>

Toshiaki Makita

Re: [PATCH V3 net 0/3] Fix checksum issues with Q-in-Q vlans

[David Miller]

From: Vladislav Yasevich <vyasevich@gmail.com>
Date: Tue, 23 May 2017 13:38:40 -0400

> TCP checksum appear broken on a lot of devices that
> advertise NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM.  This problem
> becomes very visible/reproducable since the series
> commit afb0bc972b526 ("Merge branch 'stacked_vlan_tso'").
> 
> In particular, the issue appeared consistently on bnx2 and be2net
> drivers (not all drivers were tested).
> 
> This short series corrects this by disabling checksum offload
> support on packets sent through Q-in-Q vlans if the underlying HW only
> enables IP specific checksum features.  We currently 'assume' that
> any drivers setting NETIF_F_HW_CSUM can correclty pass checksum offsets
> to HW.  It is up to individual drivers to enable it properly through
> ndo_features_check if they have some support for Q-in-Q vlans.
> 
> Additionally, be2net driver was fixed to make the proper call.
> 
> While looking at the drivers, it was also found that virtio-net ended
> up disabling accelerations, which is unnecessary.  
> 
> V3: Fixed checkpatch errors.
> 
> V2: Instead of disabling checksuming for all devices, only devices using
>     NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM are now affected by this change.
>     For drivers using NETIF_F_HW_CSUM, we will continue to use checksum
>     offloading.  If any drivers are found to be broken, they would need
>     be fixed individually.

Series applied and queued up for -stable, thanks.