kernel/net/batman-adv/soft-interface.c, branch linux-rolling-stable

batman-adv: Use consistent name for mesh interface

2025-02-22T10:36:22Z

The way how the virtual interface is called inside the batman-adv source code is not consistent. The genl headers call it meshif and the rest of the code calls is (mostly) softif. The genl definitions cannot be touched because they are part of the UAPI. But the rest of the batman-adv code can be touched to have a consistent name again. The bulk of the renaming was done using sed -i -e 's/soft\(-\|\_\| \|\)i\([nf]\)/mesh\1i\2/g' \ -e 's/SOFT\(-\|\_\| \|\)I\([NF]\)/MESH\1I\2/g' and then it was adjusted slightly when proofreading the changes. Signed-off-by: Sven Eckelmann Signed-off-by: Simon Wunderlich

batman-adv: Add support for jumbo frames

2025-02-22T10:34:52Z

Since batman-adv is not actually depending on hardware capabilities, it has no limit on the MTU. Only the lower hard interfaces can limit it. In case these have an high enough MTU or fragmentation is enabled, a higher MTU than 1500 can be enabled. Signed-off-by: Sven Eckelmann Signed-off-by: Simon Wunderlich

rtnetlink: Pack newlink() params into struct

2025-02-21T23:28:02Z

There are 4 net namespaces involved when creating links: - source netns - where the netlink socket resides, - target netns - where to put the device being created, - link netns - netns associated with the device (backend), - peer netns - netns of peer device. Currently, two nets are passed to newlink() callback - "src_net" parameter and "dev_net" (implicitly in net_device). They are set as follows, depending on netlink attributes in the request. +------------+-------------------+---------+---------+ | peer netns | IFLA_LINK_NETNSID | src_net | dev_net | +------------+-------------------+---------+---------+ | | absent | source | target | | absent +-------------------+---------+---------+ | | present | link | link | +------------+-------------------+---------+---------+ | | absent | peer | target | | present +-------------------+---------+---------+ | | present | peer | link | +------------+-------------------+---------+---------+ When IFLA_LINK_NETNSID is present, the device is created in link netns first and then moved to target netns. This has some side effects, including extra ifindex allocation, ifname validation and link events. These could be avoided if we create it in target netns from the beginning. On the other hand, the meaning of src_net parameter is ambiguous. It varies depending on how parameters are passed. It is the effective link (or peer netns) by design, but some drivers ignore it and use dev_net instead. To provide more netns context for drivers, this patch packs existing newlink() parameters, along with the source netns, link netns and peer netns, into a struct. The old "src_net" is renamed to "net" to avoid confusion with real source netns, and will be deprecated later. The use of src_net are converted to params->net trivially. Signed-off-by: Xiao Liang Reviewed-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20250219125039.18024-3-shaw.leon@gmail.com Signed-off-by: Jakub Kicinski

batman-adv: Map VID 0 to untagged TT VLAN

2024-12-17T06:35:05Z

VID 0 is not a valid VLAN according to "802.1Q-2011" "Table 9-2—Reserved VID values". It is only used to indicate "priority tag" frames which only contain priority information and no VID. The 8021q is also redirecting the priority tagged frames to the underlying interface since commit ad1afb003939 ("vlan_dev: VLAN 0 should be treated as "no vlan tag" (802.1p packet)"). But at the same time, it automatically adds the VID 0 to all devices to ensure that VID 0 is in the allowed list of the HW filter. This resulted in a VLAN 0 which was always announced in OGM messages. batman-adv should therefore not create a new batadv_softif_vlan for VID 0 and handle all VID 0 related frames using the "untagged" global/local translation tables. Signed-off-by: Sven Eckelmann Acked-by: Antonio Quartulli Signed-off-by: Simon Wunderlich

batman-adv: Remove atomic usage for tt.local_changes

2024-12-05T21:38:43Z

The tt.local_changes atomic is either written with tt.changes_list_lock or close to it (see batadv_tt_local_event()). Thus the performance gain using an atomic was limited (or because of atomic_read() impact even negative). Using atomic also comes with the need to be wary of potential negative tt.local_changes value. Simplify the tt.local_changes usage by removing the atomic property and modifying it only with tt.changes_list_lock held. Signed-off-by: Remi Pommarel Signed-off-by: Sven Eckelmann Signed-off-by: Simon Wunderlich

netdev_features: convert NETIF_F_NETNS_LOCAL to dev->netns_local

2024-09-03T09:36:43Z

"Interface can't change network namespaces" is rather an attribute, not a feature, and it can't be changed via Ethtool. Make it a "cold" private flag instead of a netdev_feature and free one more bit. Signed-off-by: Alexander Lobakin Signed-off-by: Paolo Abeni

netdev_features: convert NETIF_F_LLTX to dev->lltx

2024-09-03T09:36:43Z

NETIF_F_LLTX can't be changed via Ethtool and is not a feature, rather an attribute, very similar to IFF_NO_QUEUE (and hot). Free one netdev_features_t bit and make it a "hot" private flag. Signed-off-by: Alexander Lobakin Signed-off-by: Paolo Abeni

net: annotate writes on dev->mtu from ndo_change_mtu()

2024-05-07T23:19:14Z

Simon reported that ndo_change_mtu() methods were never updated to use WRITE_ONCE(dev->mtu, new_mtu) as hinted in commit 501a90c94510 ("inet: protect against too small mtu values.") We read dev->mtu without holding RTNL in many places, with READ_ONCE() annotations. It is time to take care of ndo_change_mtu() methods to use corresponding WRITE_ONCE() Signed-off-by: Eric Dumazet Reported-by: Simon Horman Closes: https://lore.kernel.org/netdev/20240505144608.GB67882@kernel.org/ Reviewed-by: Jacob Keller Reviewed-by: Sabrina Dubroca Reviewed-by: Simon Horman Acked-by: Shannon Nelson Link: https://lore.kernel.org/r/20240506102812.3025432-1-edumazet@google.com Signed-off-by: Jakub Kicinski

batman-adv: mcast: implement multicast packet generation

2023-11-14T07:16:34Z

Implement the preparation of a batman-adv multicast packet and use this under certain conditions. For one thing this implements the capability to push a complete batman-adv multicast packet header, including a tracker TVLV with all originator destinations that have signaled interest in it, onto a given ethernet frame with an IP multicast packet inside. For another checks are implemented to determine if encapsulating a multicast packet in this new batman-adv multicast packet type and using it is feasible. Those checks are: 1) Have all nodes signaled that they are capable of handling the new batman-adv multicast packet type? 2) Do all active hard interfaces of all nodes, including us, have an MTU of at least 1280 bytes? 3) Does a complete multicast packet header with all its destination addresses fit onto the given multicast packet / ethernet frame and does not exceed 1280 bytes? If all checks passed then the new batman-adv multicast packet type will be used for transmission and distribution. Otherwise we fall back to one or more batman-adv unicast packet transmissions, if possible. Or if not possible we will fall back to classic flooding through a batman-adv broadcast packet. Signed-off-by: Linus Lüssing Signed-off-by: Simon Wunderlich

batman-adv: mcast: implement multicast packet reception and forwarding

2023-11-14T07:16:34Z

Implement functionality to receive and forward a new TVLV capable multicast packet type. The new batman-adv multicast packet type allows to contain several originator destination addresses within a TVLV. Routers on the way will potentially split the batman-adv multicast packet and adjust its tracker TVLV contents. Routing decisions are still based on the selected BATMAN IV or BATMAN V routing algorithm. So this new batman-adv multicast packet type retains the same loop-free properties. Also a new OGM multicast TVLV flag is introduced to signal to other nodes that we are capable of handling a batman-adv multicast packet and multicast tracker TVLV. And that all of our hard interfaces have an MTU of at least 1280 bytes (IPv6 minimum MTU), as a simple solution for now to avoid MTU issues while forwarding. Signed-off-by: Linus Lüssing Signed-off-by: Simon Wunderlich