[PATCH net 1/5] net, team, bonding: Add netdev_base_features helper

[PATCH net 1/5] net, team, bonding: Add netdev_base_features helper

[Daniel Borkmann]

Both bonding and team driver have logic to derive the base feature
flags before iterating over their slave devices to refine the set
via netdev_increment_features().

Add a small helper netdev_base_features() so this can be reused
instead of having it open-coded multiple times.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Nikolay Aleksandrov <razor@blackwall.org>
Cc: Ido Schimmel <idosch@idosch.org>
Cc: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/bonding/bond_main.c | 4 +---
 drivers/net/team/team_core.c    | 3 +--
 include/linux/netdev_features.h | 7 +++++++
 3 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 49dd4fe195e5..42c835c60cd8 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -1520,9 +1520,7 @@ static netdev_features_t bond_fix_features(struct net_device *dev,
 	struct slave *slave;
 
 	mask = features;
-
-	features &= ~NETIF_F_ONE_FOR_ALL;
-	features |= NETIF_F_ALL_FOR_ALL;
+	features = netdev_base_features(features);
 
 	bond_for_each_slave(bond, slave, iter) {
 		features = netdev_increment_features(features,
diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c
index a1b27b69f010..1df062c67640 100644
--- a/drivers/net/team/team_core.c
+++ b/drivers/net/team/team_core.c
@@ -2011,8 +2011,7 @@ static netdev_features_t team_fix_features(struct net_device *dev,
 	netdev_features_t mask;
 
 	mask = features;
-	features &= ~NETIF_F_ONE_FOR_ALL;
-	features |= NETIF_F_ALL_FOR_ALL;
+	features = netdev_base_features(features);
 
 	rcu_read_lock();
 	list_for_each_entry_rcu(port, &team->port_list, list) {
diff --git a/include/linux/netdev_features.h b/include/linux/netdev_features.h
index 66e7d26b70a4..11be70a7929f 100644
--- a/include/linux/netdev_features.h
+++ b/include/linux/netdev_features.h
@@ -253,4 +253,11 @@ static inline int find_next_netdev_feature(u64 feature, unsigned long start)
 				 NETIF_F_GSO_UDP_TUNNEL |		\
 				 NETIF_F_GSO_UDP_TUNNEL_CSUM)
 
+static inline netdev_features_t netdev_base_features(netdev_features_t features)
+{
+	features &= ~NETIF_F_ONE_FOR_ALL;
+	features |= NETIF_F_ALL_FOR_ALL;
+	return features;
+}
+
 #endif	/* _LINUX_NETDEV_FEATURES_H */
-- 
2.43.0

[PATCH net 2/5] bonding: Fix initial {vlan,mpls}_feature set in bond_compute_features

[Daniel Borkmann]

If a bonding device has slave devices, then the current logic to derive
the feature set for the master bond device is limited in that flags which
are fully supported by the underlying slave devices cannot be propagated
up to vlan devices which sit on top of bond devices. Instead, these get
blindly masked out via current NETIF_F_ALL_FOR_ALL logic.

vlan_features and mpls_features should reuse netdev_base_features() in
order derive the set in the same way as ndo_fix_features before iterating
through the slave devices to refine the feature set.

Fixes: a9b3ace44c7d ("bonding: fix vlan_features computing")
Fixes: 2e770b507ccd ("net: bonding: Inherit MPLS features from slave devices")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Nikolay Aleksandrov <razor@blackwall.org>
Cc: Ido Schimmel <idosch@idosch.org>
Cc: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/bonding/bond_main.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 42c835c60cd8..320dd71392ef 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -1563,8 +1563,9 @@ static void bond_compute_features(struct bonding *bond)
 
 	if (!bond_has_slaves(bond))
 		goto done;
-	vlan_features &= NETIF_F_ALL_FOR_ALL;
-	mpls_features &= NETIF_F_ALL_FOR_ALL;
+
+	vlan_features = netdev_base_features(vlan_features);
+	mpls_features = netdev_base_features(mpls_features);
 
 	bond_for_each_slave(bond, slave, iter) {
 		vlan_features = netdev_increment_features(vlan_features,
-- 
2.43.0

[PATCH net 3/5] bonding: Fix feature propagation of NETIF_F_GSO_ENCAP_ALL

[Daniel Borkmann]

Drivers like mlx5 expose NIC's vlan_features such as
NETIF_F_GSO_UDP_TUNNEL & NETIF_F_GSO_UDP_TUNNEL_CSUM which are
later not propagated when the underlying devices are bonded and
a vlan device created on top of the bond.

Right now, the more cumbersome workaround for this is to create
the vlan on top of the mlx5 and then enslave the vlan devices
to a bond.

To fix this, add NETIF_F_GSO_ENCAP_ALL to BOND_VLAN_FEATURES
such that bond_compute_features() can probe and propagate the
vlan_features from the slave devices up to the vlan device.

Given the following bond:

  # ethtool -i enp2s0f{0,1}np{0,1}
  driver: mlx5_core
  [...]

  # ethtool -k enp2s0f0np0 | grep udp
  tx-udp_tnl-segmentation: on
  tx-udp_tnl-csum-segmentation: on
  tx-udp-segmentation: on
  rx-udp_tunnel-port-offload: on
  rx-udp-gro-forwarding: off

  # ethtool -k enp2s0f1np1 | grep udp
  tx-udp_tnl-segmentation: on
  tx-udp_tnl-csum-segmentation: on
  tx-udp-segmentation: on
  rx-udp_tunnel-port-offload: on
  rx-udp-gro-forwarding: off

  # ethtool -k bond0 | grep udp
  tx-udp_tnl-segmentation: on
  tx-udp_tnl-csum-segmentation: on
  tx-udp-segmentation: on
  rx-udp_tunnel-port-offload: off [fixed]
  rx-udp-gro-forwarding: off

Before:

  # ethtool -k bond0.100 | grep udp
  tx-udp_tnl-segmentation: off [requested on]
  tx-udp_tnl-csum-segmentation: off [requested on]
  tx-udp-segmentation: on
  rx-udp_tunnel-port-offload: off [fixed]
  rx-udp-gro-forwarding: off

After:

  # ethtool -k bond0.100 | grep udp
  tx-udp_tnl-segmentation: on
  tx-udp_tnl-csum-segmentation: on
  tx-udp-segmentation: on
  rx-udp_tunnel-port-offload: off [fixed]
  rx-udp-gro-forwarding: off

Various users have run into this reporting performance issues when
configuring Cilium in vxlan tunneling mode and having the combination
of bond & vlan for the core devices connecting the Kubernetes cluster
to the outside world.

Fixes: a9b3ace44c7d ("bonding: fix vlan_features computing")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Nikolay Aleksandrov <razor@blackwall.org>
Cc: Ido Schimmel <idosch@idosch.org>
Cc: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/bonding/bond_main.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 320dd71392ef..7b78c2bada81 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -1534,6 +1534,7 @@ static netdev_features_t bond_fix_features(struct net_device *dev,
 
 #define BOND_VLAN_FEATURES	(NETIF_F_HW_CSUM | NETIF_F_SG | \
 				 NETIF_F_FRAGLIST | NETIF_F_GSO_SOFTWARE | \
+				 NETIF_F_GSO_ENCAP_ALL | \
 				 NETIF_F_HIGHDMA | NETIF_F_LRO)
 
 #define BOND_ENC_FEATURES	(NETIF_F_HW_CSUM | NETIF_F_SG | \
-- 
2.43.0

[PATCH net 4/5] team: Fix initial vlan_feature set in __team_compute_features

[Daniel Borkmann]

Similarly as with bonding, fix the calculation of vlan_features to reuse
netdev_base_features() in order derive the set in the same way as
ndo_fix_features before iterating through the slave devices to refine the
feature set.

Fixes: 3625920b62c3 ("teaming: fix vlan_features computing")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Nikolay Aleksandrov <razor@blackwall.org>
Cc: Ido Schimmel <idosch@idosch.org>
Cc: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/team/team_core.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c
index 1df062c67640..306416fc1db0 100644
--- a/drivers/net/team/team_core.c
+++ b/drivers/net/team/team_core.c
@@ -991,13 +991,14 @@ static void team_port_disable(struct team *team,
 static void __team_compute_features(struct team *team)
 {
 	struct team_port *port;
-	netdev_features_t vlan_features = TEAM_VLAN_FEATURES &
-					  NETIF_F_ALL_FOR_ALL;
+	netdev_features_t vlan_features = TEAM_VLAN_FEATURES;
 	netdev_features_t enc_features  = TEAM_ENC_FEATURES;
 	unsigned short max_hard_header_len = ETH_HLEN;
 	unsigned int dst_release_flag = IFF_XMIT_DST_RELEASE |
 					IFF_XMIT_DST_RELEASE_PERM;
 
+	vlan_features = netdev_base_features(vlan_features);
+
 	rcu_read_lock();
 	list_for_each_entry_rcu(port, &team->port_list, list) {
 		vlan_features = netdev_increment_features(vlan_features,
-- 
2.43.0

[PATCH net 5/5] team: Fix feature propagation of NETIF_F_GSO_ENCAP_ALL

[Daniel Borkmann]

Similar to bonding driver, add NETIF_F_GSO_ENCAP_ALL to TEAM_VLAN_FEATURES
in order to support slave devices which propagate NETIF_F_GSO_UDP_TUNNEL &
NETIF_F_GSO_UDP_TUNNEL_CSUM as vlan_features.

Fixes: 3625920b62c3 ("teaming: fix vlan_features computing")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Nikolay Aleksandrov <razor@blackwall.org>
Cc: Ido Schimmel <idosch@idosch.org>
Cc: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/team/team_core.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c
index 306416fc1db0..69ea2c3c76bf 100644
--- a/drivers/net/team/team_core.c
+++ b/drivers/net/team/team_core.c
@@ -983,7 +983,8 @@ static void team_port_disable(struct team *team,
 
 #define TEAM_VLAN_FEATURES (NETIF_F_HW_CSUM | NETIF_F_SG | \
 			    NETIF_F_FRAGLIST | NETIF_F_GSO_SOFTWARE | \
-			    NETIF_F_HIGHDMA | NETIF_F_LRO)
+			    NETIF_F_HIGHDMA | NETIF_F_LRO | \
+			    NETIF_F_GSO_ENCAP_ALL)
 
 #define TEAM_ENC_FEATURES	(NETIF_F_HW_CSUM | NETIF_F_SG | \
 				 NETIF_F_RXCSUM | NETIF_F_GSO_SOFTWARE)
-- 
2.43.0

Re: [PATCH net 1/5] net, team, bonding: Add netdev_base_features helper

[Hangbin Liu]

On Tue, Dec 10, 2024 at 03:12:41PM +0100, Daniel Borkmann wrote:
> Both bonding and team driver have logic to derive the base feature
> flags before iterating over their slave devices to refine the set
> via netdev_increment_features().
> 
> Add a small helper netdev_base_features() so this can be reused
> instead of having it open-coded multiple times.
> 
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/bonding/bond_main.c | 4 +---
>  drivers/net/team/team_core.c    | 3 +--
>  include/linux/netdev_features.h | 7 +++++++
>  3 files changed, 9 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 49dd4fe195e5..42c835c60cd8 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -1520,9 +1520,7 @@ static netdev_features_t bond_fix_features(struct net_device *dev,
>  	struct slave *slave;
>  
>  	mask = features;
> -
> -	features &= ~NETIF_F_ONE_FOR_ALL;
> -	features |= NETIF_F_ALL_FOR_ALL;
> +	features = netdev_base_features(features);
>  
>  	bond_for_each_slave(bond, slave, iter) {
>  		features = netdev_increment_features(features,
> diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c
> index a1b27b69f010..1df062c67640 100644
> --- a/drivers/net/team/team_core.c
> +++ b/drivers/net/team/team_core.c
> @@ -2011,8 +2011,7 @@ static netdev_features_t team_fix_features(struct net_device *dev,
>  	netdev_features_t mask;
>  
>  	mask = features;
> -	features &= ~NETIF_F_ONE_FOR_ALL;
> -	features |= NETIF_F_ALL_FOR_ALL;
> +	features = netdev_base_features(features);
>  
>  	rcu_read_lock();
>  	list_for_each_entry_rcu(port, &team->port_list, list) {
> diff --git a/include/linux/netdev_features.h b/include/linux/netdev_features.h
> index 66e7d26b70a4..11be70a7929f 100644
> --- a/include/linux/netdev_features.h
> +++ b/include/linux/netdev_features.h
> @@ -253,4 +253,11 @@ static inline int find_next_netdev_feature(u64 feature, unsigned long start)
>  				 NETIF_F_GSO_UDP_TUNNEL |		\
>  				 NETIF_F_GSO_UDP_TUNNEL_CSUM)
>  
> +static inline netdev_features_t netdev_base_features(netdev_features_t features)
> +{
> +	features &= ~NETIF_F_ONE_FOR_ALL;
> +	features |= NETIF_F_ALL_FOR_ALL;
> +	return features;
> +}
> +
>  #endif	/* _LINUX_NETDEV_FEATURES_H */
> -- 
> 2.43.0
> 

Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>

Re: [PATCH net 1/5] net, team, bonding: Add netdev_base_features helper

[Nikolay Aleksandrov]

On 12/10/24 16:12, Daniel Borkmann wrote:
> Both bonding and team driver have logic to derive the base feature
> flags before iterating over their slave devices to refine the set
> via netdev_increment_features().
> 
> Add a small helper netdev_base_features() so this can be reused
> instead of having it open-coded multiple times.
> 
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/bonding/bond_main.c | 4 +---
>  drivers/net/team/team_core.c    | 3 +--
>  include/linux/netdev_features.h | 7 +++++++
>  3 files changed, 9 insertions(+), 5 deletions(-)
> 

Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>

Re: [PATCH net 2/5] bonding: Fix initial {vlan,mpls}_feature set in bond_compute_features

[Nikolay Aleksandrov]

On 12/10/24 16:12, Daniel Borkmann wrote:
> If a bonding device has slave devices, then the current logic to derive
> the feature set for the master bond device is limited in that flags which
> are fully supported by the underlying slave devices cannot be propagated
> up to vlan devices which sit on top of bond devices. Instead, these get
> blindly masked out via current NETIF_F_ALL_FOR_ALL logic.
> 
> vlan_features and mpls_features should reuse netdev_base_features() in
> order derive the set in the same way as ndo_fix_features before iterating
> through the slave devices to refine the feature set.
> 
> Fixes: a9b3ace44c7d ("bonding: fix vlan_features computing")
> Fixes: 2e770b507ccd ("net: bonding: Inherit MPLS features from slave devices")
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/bonding/bond_main.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 42c835c60cd8..320dd71392ef 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -1563,8 +1563,9 @@ static void bond_compute_features(struct bonding *bond)
>  
>  	if (!bond_has_slaves(bond))
>  		goto done;
> -	vlan_features &= NETIF_F_ALL_FOR_ALL;
> -	mpls_features &= NETIF_F_ALL_FOR_ALL;
> +
> +	vlan_features = netdev_base_features(vlan_features);
> +	mpls_features = netdev_base_features(mpls_features);
>  
>  	bond_for_each_slave(bond, slave, iter) {
>  		vlan_features = netdev_increment_features(vlan_features,

Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>

Re: [PATCH net 3/5] bonding: Fix feature propagation of NETIF_F_GSO_ENCAP_ALL

[Nikolay Aleksandrov]

On 12/10/24 16:12, Daniel Borkmann wrote:
> Drivers like mlx5 expose NIC's vlan_features such as
> NETIF_F_GSO_UDP_TUNNEL & NETIF_F_GSO_UDP_TUNNEL_CSUM which are
> later not propagated when the underlying devices are bonded and
> a vlan device created on top of the bond.
> 
> Right now, the more cumbersome workaround for this is to create
> the vlan on top of the mlx5 and then enslave the vlan devices
> to a bond.
> 
> To fix this, add NETIF_F_GSO_ENCAP_ALL to BOND_VLAN_FEATURES
> such that bond_compute_features() can probe and propagate the
> vlan_features from the slave devices up to the vlan device.
> 
> Given the following bond:
> 
>   # ethtool -i enp2s0f{0,1}np{0,1}
>   driver: mlx5_core
>   [...]
> 
>   # ethtool -k enp2s0f0np0 | grep udp
>   tx-udp_tnl-segmentation: on
>   tx-udp_tnl-csum-segmentation: on
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: on
>   rx-udp-gro-forwarding: off
> 
>   # ethtool -k enp2s0f1np1 | grep udp
>   tx-udp_tnl-segmentation: on
>   tx-udp_tnl-csum-segmentation: on
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: on
>   rx-udp-gro-forwarding: off
> 
>   # ethtool -k bond0 | grep udp
>   tx-udp_tnl-segmentation: on
>   tx-udp_tnl-csum-segmentation: on
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: off [fixed]
>   rx-udp-gro-forwarding: off
> 
> Before:
> 
>   # ethtool -k bond0.100 | grep udp
>   tx-udp_tnl-segmentation: off [requested on]
>   tx-udp_tnl-csum-segmentation: off [requested on]
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: off [fixed]
>   rx-udp-gro-forwarding: off
> 
> After:
> 
>   # ethtool -k bond0.100 | grep udp
>   tx-udp_tnl-segmentation: on
>   tx-udp_tnl-csum-segmentation: on
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: off [fixed]
>   rx-udp-gro-forwarding: off
> 
> Various users have run into this reporting performance issues when
> configuring Cilium in vxlan tunneling mode and having the combination
> of bond & vlan for the core devices connecting the Kubernetes cluster
> to the outside world.
> 
> Fixes: a9b3ace44c7d ("bonding: fix vlan_features computing")
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/bonding/bond_main.c | 1 +
>  1 file changed, 1 insertion(+)
> 

Indeed, I've tested a similar change a year ago to get the expected performance.

Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>

Re: [PATCH net 4/5] team: Fix initial vlan_feature set in __team_compute_features

[Nikolay Aleksandrov]

On 12/10/24 16:12, Daniel Borkmann wrote:
> Similarly as with bonding, fix the calculation of vlan_features to reuse
> netdev_base_features() in order derive the set in the same way as
> ndo_fix_features before iterating through the slave devices to refine the
> feature set.
> 
> Fixes: 3625920b62c3 ("teaming: fix vlan_features computing")
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/team/team_core.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c
> index 1df062c67640..306416fc1db0 100644
> --- a/drivers/net/team/team_core.c
> +++ b/drivers/net/team/team_core.c
> @@ -991,13 +991,14 @@ static void team_port_disable(struct team *team,
>  static void __team_compute_features(struct team *team)
>  {
>  	struct team_port *port;
> -	netdev_features_t vlan_features = TEAM_VLAN_FEATURES &
> -					  NETIF_F_ALL_FOR_ALL;
> +	netdev_features_t vlan_features = TEAM_VLAN_FEATURES;
>  	netdev_features_t enc_features  = TEAM_ENC_FEATURES;
>  	unsigned short max_hard_header_len = ETH_HLEN;
>  	unsigned int dst_release_flag = IFF_XMIT_DST_RELEASE |
>  					IFF_XMIT_DST_RELEASE_PERM;
>  
> +	vlan_features = netdev_base_features(vlan_features);
> +
>  	rcu_read_lock();
>  	list_for_each_entry_rcu(port, &team->port_list, list) {
>  		vlan_features = netdev_increment_features(vlan_features,

Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>

Re: [PATCH net 5/5] team: Fix feature propagation of NETIF_F_GSO_ENCAP_ALL

[Nikolay Aleksandrov]

On 12/10/24 16:12, Daniel Borkmann wrote:
> Similar to bonding driver, add NETIF_F_GSO_ENCAP_ALL to TEAM_VLAN_FEATURES
> in order to support slave devices which propagate NETIF_F_GSO_UDP_TUNNEL &
> NETIF_F_GSO_UDP_TUNNEL_CSUM as vlan_features.
> 
> Fixes: 3625920b62c3 ("teaming: fix vlan_features computing")
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/team/team_core.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c
> index 306416fc1db0..69ea2c3c76bf 100644
> --- a/drivers/net/team/team_core.c
> +++ b/drivers/net/team/team_core.c
> @@ -983,7 +983,8 @@ static void team_port_disable(struct team *team,
>  
>  #define TEAM_VLAN_FEATURES (NETIF_F_HW_CSUM | NETIF_F_SG | \
>  			    NETIF_F_FRAGLIST | NETIF_F_GSO_SOFTWARE | \
> -			    NETIF_F_HIGHDMA | NETIF_F_LRO)
> +			    NETIF_F_HIGHDMA | NETIF_F_LRO | \
> +			    NETIF_F_GSO_ENCAP_ALL)
>  
>  #define TEAM_ENC_FEATURES	(NETIF_F_HW_CSUM | NETIF_F_SG | \
>  				 NETIF_F_RXCSUM | NETIF_F_GSO_SOFTWARE)

Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>

Re: [PATCH net 2/5] bonding: Fix initial {vlan,mpls}_feature set in bond_compute_features

[Hangbin Liu]

On Tue, Dec 10, 2024 at 03:12:42PM +0100, Daniel Borkmann wrote:
> If a bonding device has slave devices, then the current logic to derive
> the feature set for the master bond device is limited in that flags which
> are fully supported by the underlying slave devices cannot be propagated
> up to vlan devices which sit on top of bond devices. Instead, these get
> blindly masked out via current NETIF_F_ALL_FOR_ALL logic.
> 
> vlan_features and mpls_features should reuse netdev_base_features() in
> order derive the set in the same way as ndo_fix_features before iterating
> through the slave devices to refine the feature set.
> 
> Fixes: a9b3ace44c7d ("bonding: fix vlan_features computing")
> Fixes: 2e770b507ccd ("net: bonding: Inherit MPLS features from slave devices")
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/bonding/bond_main.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 42c835c60cd8..320dd71392ef 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -1563,8 +1563,9 @@ static void bond_compute_features(struct bonding *bond)
>  
>  	if (!bond_has_slaves(bond))
>  		goto done;
> -	vlan_features &= NETIF_F_ALL_FOR_ALL;
> -	mpls_features &= NETIF_F_ALL_FOR_ALL;
> +
> +	vlan_features = netdev_base_features(vlan_features);
> +	mpls_features = netdev_base_features(mpls_features);
>  
>  	bond_for_each_slave(bond, slave, iter) {
>  		vlan_features = netdev_increment_features(vlan_features,
> -- 
> 2.43.0
> 

Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>

Re: [PATCH net 4/5] team: Fix initial vlan_feature set in __team_compute_features

[Hangbin Liu]

On Tue, Dec 10, 2024 at 03:12:44PM +0100, Daniel Borkmann wrote:
> Similarly as with bonding, fix the calculation of vlan_features to reuse
> netdev_base_features() in order derive the set in the same way as
> ndo_fix_features before iterating through the slave devices to refine the
> feature set.
> 
> Fixes: 3625920b62c3 ("teaming: fix vlan_features computing")
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/team/team_core.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c
> index 1df062c67640..306416fc1db0 100644
> --- a/drivers/net/team/team_core.c
> +++ b/drivers/net/team/team_core.c
> @@ -991,13 +991,14 @@ static void team_port_disable(struct team *team,
>  static void __team_compute_features(struct team *team)
>  {
>  	struct team_port *port;
> -	netdev_features_t vlan_features = TEAM_VLAN_FEATURES &
> -					  NETIF_F_ALL_FOR_ALL;
> +	netdev_features_t vlan_features = TEAM_VLAN_FEATURES;
>  	netdev_features_t enc_features  = TEAM_ENC_FEATURES;
>  	unsigned short max_hard_header_len = ETH_HLEN;
>  	unsigned int dst_release_flag = IFF_XMIT_DST_RELEASE |
>  					IFF_XMIT_DST_RELEASE_PERM;
>  
> +	vlan_features = netdev_base_features(vlan_features);
> +
>  	rcu_read_lock();
>  	list_for_each_entry_rcu(port, &team->port_list, list) {
>  		vlan_features = netdev_increment_features(vlan_features,
> -- 
> 2.43.0
> 

Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>

Re: [PATCH net 3/5] bonding: Fix feature propagation of NETIF_F_GSO_ENCAP_ALL

[Hangbin Liu]

On Tue, Dec 10, 2024 at 03:12:43PM +0100, Daniel Borkmann wrote:
> Drivers like mlx5 expose NIC's vlan_features such as
> NETIF_F_GSO_UDP_TUNNEL & NETIF_F_GSO_UDP_TUNNEL_CSUM which are
> later not propagated when the underlying devices are bonded and
> a vlan device created on top of the bond.
> 
> Right now, the more cumbersome workaround for this is to create
> the vlan on top of the mlx5 and then enslave the vlan devices
> to a bond.
> 
> To fix this, add NETIF_F_GSO_ENCAP_ALL to BOND_VLAN_FEATURES
> such that bond_compute_features() can probe and propagate the
> vlan_features from the slave devices up to the vlan device.
> 
> Given the following bond:
> 
>   # ethtool -i enp2s0f{0,1}np{0,1}
>   driver: mlx5_core
>   [...]
> 
>   # ethtool -k enp2s0f0np0 | grep udp
>   tx-udp_tnl-segmentation: on
>   tx-udp_tnl-csum-segmentation: on
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: on
>   rx-udp-gro-forwarding: off
> 
>   # ethtool -k enp2s0f1np1 | grep udp
>   tx-udp_tnl-segmentation: on
>   tx-udp_tnl-csum-segmentation: on
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: on
>   rx-udp-gro-forwarding: off
> 
>   # ethtool -k bond0 | grep udp
>   tx-udp_tnl-segmentation: on
>   tx-udp_tnl-csum-segmentation: on
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: off [fixed]
>   rx-udp-gro-forwarding: off
> 
> Before:
> 
>   # ethtool -k bond0.100 | grep udp
>   tx-udp_tnl-segmentation: off [requested on]
>   tx-udp_tnl-csum-segmentation: off [requested on]
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: off [fixed]
>   rx-udp-gro-forwarding: off
> 
> After:
> 
>   # ethtool -k bond0.100 | grep udp
>   tx-udp_tnl-segmentation: on
>   tx-udp_tnl-csum-segmentation: on
>   tx-udp-segmentation: on
>   rx-udp_tunnel-port-offload: off [fixed]
>   rx-udp-gro-forwarding: off
> 
> Various users have run into this reporting performance issues when
> configuring Cilium in vxlan tunneling mode and having the combination
> of bond & vlan for the core devices connecting the Kubernetes cluster
> to the outside world.
> 
> Fixes: a9b3ace44c7d ("bonding: fix vlan_features computing")
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/bonding/bond_main.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 320dd71392ef..7b78c2bada81 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -1534,6 +1534,7 @@ static netdev_features_t bond_fix_features(struct net_device *dev,
>  
>  #define BOND_VLAN_FEATURES	(NETIF_F_HW_CSUM | NETIF_F_SG | \
>  				 NETIF_F_FRAGLIST | NETIF_F_GSO_SOFTWARE | \
> +				 NETIF_F_GSO_ENCAP_ALL | \
>  				 NETIF_F_HIGHDMA | NETIF_F_LRO)
>  
>  #define BOND_ENC_FEATURES	(NETIF_F_HW_CSUM | NETIF_F_SG | \
> -- 
> 2.43.0
> 
Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>

Re: [PATCH net 5/5] team: Fix feature propagation of NETIF_F_GSO_ENCAP_ALL

[Hangbin Liu]

On Tue, Dec 10, 2024 at 03:12:45PM +0100, Daniel Borkmann wrote:
> Similar to bonding driver, add NETIF_F_GSO_ENCAP_ALL to TEAM_VLAN_FEATURES
> in order to support slave devices which propagate NETIF_F_GSO_UDP_TUNNEL &
> NETIF_F_GSO_UDP_TUNNEL_CSUM as vlan_features.
> 
> Fixes: 3625920b62c3 ("teaming: fix vlan_features computing")
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>
> ---
>  drivers/net/team/team_core.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c
> index 306416fc1db0..69ea2c3c76bf 100644
> --- a/drivers/net/team/team_core.c
> +++ b/drivers/net/team/team_core.c
> @@ -983,7 +983,8 @@ static void team_port_disable(struct team *team,
>  
>  #define TEAM_VLAN_FEATURES (NETIF_F_HW_CSUM | NETIF_F_SG | \
>  			    NETIF_F_FRAGLIST | NETIF_F_GSO_SOFTWARE | \
> -			    NETIF_F_HIGHDMA | NETIF_F_LRO)
> +			    NETIF_F_HIGHDMA | NETIF_F_LRO | \
> +			    NETIF_F_GSO_ENCAP_ALL)
>  
>  #define TEAM_ENC_FEATURES	(NETIF_F_HW_CSUM | NETIF_F_SG | \
>  				 NETIF_F_RXCSUM | NETIF_F_GSO_SOFTWARE)
> -- 
> 2.43.0
> 

Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>

Re: [PATCH net 1/5] net, team, bonding: Add netdev_base_features helper

[Paolo Abeni]

On 12/10/24 15:12, Daniel Borkmann wrote:
> Both bonding and team driver have logic to derive the base feature
> flags before iterating over their slave devices to refine the set
> via netdev_increment_features().
> 
> Add a small helper netdev_base_features() so this can be reused
> instead of having it open-coded multiple times.
> 
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Nikolay Aleksandrov <razor@blackwall.org>
> Cc: Ido Schimmel <idosch@idosch.org>
> Cc: Jiri Pirko <jiri@nvidia.com>

The series looks good, I'm applying it right now, but please include a
(even small) cover letter in the next multi-patch submission, thanks!

Paolo

Re: [PATCH net 1/5] net, team, bonding: Add netdev_base_features helper

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net.git (main)
by Paolo Abeni <pabeni@redhat.com>:

On Tue, 10 Dec 2024 15:12:41 +0100 you wrote:
> Both bonding and team driver have logic to derive the base feature
> flags before iterating over their slave devices to refine the set
> via netdev_increment_features().
> 
> Add a small helper netdev_base_features() so this can be reused
> instead of having it open-coded multiple times.
> 
> [...]

Here is the summary with links:
  - [net,1/5] net, team, bonding: Add netdev_base_features helper
    https://git.kernel.org/netdev/net/c/d2516c3a5370
  - [net,2/5] bonding: Fix initial {vlan,mpls}_feature set in bond_compute_features
    https://git.kernel.org/netdev/net/c/d064ea7fe2a2
  - [net,3/5] bonding: Fix feature propagation of NETIF_F_GSO_ENCAP_ALL
    https://git.kernel.org/netdev/net/c/77b11c8bf3a2
  - [net,4/5] team: Fix initial vlan_feature set in __team_compute_features
    https://git.kernel.org/netdev/net/c/396699ac2cb1
  - [net,5/5] team: Fix feature propagation of NETIF_F_GSO_ENCAP_ALL
    https://git.kernel.org/netdev/net/c/98712844589e

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

Re: [PATCH net 1/5] net, team, bonding: Add netdev_base_features helper

[Daniel Borkmann]

On 12/12/24 11:58 AM, Paolo Abeni wrote:
> On 12/10/24 15:12, Daniel Borkmann wrote:
>> Both bonding and team driver have logic to derive the base feature
>> flags before iterating over their slave devices to refine the set
>> via netdev_increment_features().
>>
>> Add a small helper netdev_base_features() so this can be reused
>> instead of having it open-coded multiple times.
>>
>> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
>> Cc: Nikolay Aleksandrov <razor@blackwall.org>
>> Cc: Ido Schimmel <idosch@idosch.org>
>> Cc: Jiri Pirko <jiri@nvidia.com>
> 
> The series looks good, I'm applying it right now, but please include a
> (even small) cover letter in the next multi-patch submission, thanks!

Ok, thanks, will do next time!

Cheers,
Daniel