Hi,
I went through all changes in batman-adv since v4.14 with a Fixes: line and checked whether they were backported to the LTS kernels. The ones which weren't ported and applied to this branch are now part of this patch series.
There are also following three patches included:
* batman-adv: Consider fragmentation for needed_headroom * batman-adv: Reserve needed_*room for fragments * batman-adv: Don't always reallocate the fragmentation skb head
which could in some circumstances cause packet loss but which were created to fix high CPU load/low throughput problems. But I've added them here anyway because the corresponding VXLAN patches were also added to stable. And some stable kernels also got these fixes a while back.
Kind regards, Sven
Linus Lüssing (2): batman-adv: mcast: fix duplicate mcast packets in BLA backbone from LAN batman-adv: mcast: fix duplicate mcast packets from BLA backbone to mesh
Sven Eckelmann (3): batman-adv: Consider fragmentation for needed_headroom batman-adv: Reserve needed_*room for fragments batman-adv: Don't always reallocate the fragmentation skb head
net/batman-adv/bridge_loop_avoidance.c | 103 +++++++++++++++++++++---- net/batman-adv/fragmentation.c | 26 ++++--- net/batman-adv/hard-interface.c | 3 + net/batman-adv/multicast.c | 31 ++++++++ net/batman-adv/multicast.h | 15 ++++ net/batman-adv/soft-interface.c | 5 +- 6 files changed, 154 insertions(+), 29 deletions(-)
From: Linus Lüssing linus.luessing@c0d3.blue
commit 3236d215ad38a3f5372e65cd1e0a52cf93d3c6a2 upstream.
Scenario: * Multicast frame send from a BLA backbone (multiple nodes with their bat0 bridged together, with BLA enabled)
Issue: * BLA backbone nodes receive the frame multiple times on bat0
For multicast frames received via batman-adv broadcast packets the originator of the broadcast packet is checked before decapsulating and forwarding the frame to bat0 (batadv_bla_is_backbone_gw()-> batadv_recv_bcast_packet()). If it came from a node which shares the same BLA backbone with us then it is not forwarded to bat0 to avoid a loop.
When sending a multicast frame in a non-4-address batman-adv unicast packet we are currently missing this check - and cannot do so because the batman-adv unicast packet has no originator address field.
However, we can simply fix this on the sender side by only sending the multicast frame via unicasts to interested nodes which do not share the same BLA backbone with us. This also nicely avoids some unnecessary transmissions on mesh side.
Note that no infinite loop was observed, probably because of dropping via batadv_interface_tx()->batadv_bla_tx(). However the duplicates still utterly confuse switches/bridges, ICMPv6 duplicate address detection and neighbor discovery and therefore leads to long delays before being able to establish TCP connections, for instance. And it also leads to the Linux bridge printing messages like: "br-lan: received packet on eth1 with own address as source address ..."
Fixes: 1d8ab8d3c176 ("batman-adv: Modified forwarding behaviour for multicast packets") Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue Signed-off-by: Simon Wunderlich sw@simonwunderlich.de [ bp: 4.14 backport: drop usage in non-existing batadv_mcast_forw_*, correct fixes line ] Signed-off-by: Sven Eckelmann sven@narfation.org --- net/batman-adv/multicast.c | 31 +++++++++++++++++++++++++++++++ net/batman-adv/multicast.h | 15 +++++++++++++++ net/batman-adv/soft-interface.c | 5 ++--- 3 files changed, 48 insertions(+), 3 deletions(-)
diff --git a/net/batman-adv/multicast.c b/net/batman-adv/multicast.c index d47865e0e697..fc62f325f9ec 100644 --- a/net/batman-adv/multicast.c +++ b/net/batman-adv/multicast.c @@ -55,10 +55,12 @@ #include <net/ip.h> #include <net/ipv6.h>
+#include "bridge_loop_avoidance.h" #include "hard-interface.h" #include "hash.h" #include "log.h" #include "packet.h" +#include "send.h" #include "translation-table.h" #include "tvlv.h"
@@ -1290,6 +1292,35 @@ void batadv_mcast_free(struct batadv_priv *bat_priv) batadv_mcast_mla_tt_retract(bat_priv, NULL); }
+/** + * batadv_mcast_forw_send_orig() - send a multicast packet to an originator + * @bat_priv: the bat priv with all the soft interface information + * @skb: the multicast packet to send + * @vid: the vlan identifier + * @orig_node: the originator to send the packet to + * + * Return: NET_XMIT_DROP in case of error or NET_XMIT_SUCCESS otherwise. + */ +int batadv_mcast_forw_send_orig(struct batadv_priv *bat_priv, + struct sk_buff *skb, + unsigned short vid, + struct batadv_orig_node *orig_node) +{ + /* Avoid sending multicast-in-unicast packets to other BLA + * gateways - they already got the frame from the LAN side + * we share with them. + * TODO: Refactor to take BLA into account earlier, to avoid + * reducing the mcast_fanout count. + */ + if (batadv_bla_is_backbone_gw_orig(bat_priv, orig_node->orig, vid)) { + dev_kfree_skb(skb); + return NET_XMIT_SUCCESS; + } + + return batadv_send_skb_unicast(bat_priv, skb, BATADV_UNICAST, 0, + orig_node, vid); +} + /** * batadv_mcast_purge_orig - reset originator global mcast state modifications * @orig: the originator which is going to get purged diff --git a/net/batman-adv/multicast.h b/net/batman-adv/multicast.h index 2a78cddab0e9..8faf4f4e0c75 100644 --- a/net/batman-adv/multicast.h +++ b/net/batman-adv/multicast.h @@ -43,6 +43,11 @@ enum batadv_forw_mode batadv_mcast_forw_mode(struct batadv_priv *bat_priv, struct sk_buff *skb, struct batadv_orig_node **mcast_single_orig);
+int batadv_mcast_forw_send_orig(struct batadv_priv *bat_priv, + struct sk_buff *skb, + unsigned short vid, + struct batadv_orig_node *orig_node); + void batadv_mcast_init(struct batadv_priv *bat_priv);
int batadv_mcast_flags_seq_print_text(struct seq_file *seq, void *offset); @@ -65,6 +70,16 @@ static inline int batadv_mcast_init(struct batadv_priv *bat_priv) return 0; }
+static inline int +batadv_mcast_forw_send_orig(struct batadv_priv *bat_priv, + struct sk_buff *skb, + unsigned short vid, + struct batadv_orig_node *orig_node) +{ + kfree_skb(skb); + return NET_XMIT_DROP; +} + static inline void batadv_mcast_free(struct batadv_priv *bat_priv) { } diff --git a/net/batman-adv/soft-interface.c b/net/batman-adv/soft-interface.c index ba9dce04343a..7a04ac96b121 100644 --- a/net/batman-adv/soft-interface.c +++ b/net/batman-adv/soft-interface.c @@ -359,9 +359,8 @@ static int batadv_interface_tx(struct sk_buff *skb, goto dropped; ret = batadv_send_skb_via_gw(bat_priv, skb, vid); } else if (mcast_single_orig) { - ret = batadv_send_skb_unicast(bat_priv, skb, - BATADV_UNICAST, 0, - mcast_single_orig, vid); + ret = batadv_mcast_forw_send_orig(bat_priv, skb, vid, + mcast_single_orig); } else { if (batadv_dat_snoop_outgoing_arp_request(bat_priv, skb))
From: Linus Lüssing linus.luessing@c0d3.blue
commit 2369e827046920ef0599e6a36b975ac5c0a359c2 upstream.
Scenario: * Multicast frame send from BLA backbone gateways (multiple nodes with their bat0 bridged together, with BLA enabled) sharing the same LAN to nodes in the mesh
Issue: * Nodes receive the frame multiple times on bat0 from the mesh, once from each foreign BLA backbone gateway which shares the same LAN with another
For multicast frames via batman-adv broadcast packets coming from the same BLA backbone but from different backbone gateways duplicates are currently detected via a CRC history of previously received packets.
However this CRC so far was not performed for multicast frames received via batman-adv unicast packets. Fixing this by appyling the same check for such packets, too.
Room for improvements in the future: Ideally we would introduce the possibility to not only claim a client, but a complete originator, too. This would allow us to only send a multicast-in-unicast packet from a BLA backbone gateway claiming the node and by that avoid potential redundant transmissions in the first place.
Fixes: fe2da6ff27c7 ("batman-adv: add broadcast duplicate check") Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue Signed-off-by: Simon Wunderlich sw@simonwunderlich.de [ bp: 4.14 backported: adjust context, correct fixes line ] Signed-off-by: Sven Eckelmann sven@narfation.org --- net/batman-adv/bridge_loop_avoidance.c | 103 +++++++++++++++++++++---- 1 file changed, 87 insertions(+), 16 deletions(-)
diff --git a/net/batman-adv/bridge_loop_avoidance.c b/net/batman-adv/bridge_loop_avoidance.c index 12ecde765003..89afb90458f4 100644 --- a/net/batman-adv/bridge_loop_avoidance.c +++ b/net/batman-adv/bridge_loop_avoidance.c @@ -1598,13 +1598,16 @@ int batadv_bla_init(struct batadv_priv *bat_priv) }
/** - * batadv_bla_check_bcast_duplist - Check if a frame is in the broadcast dup. + * batadv_bla_check_duplist() - Check if a frame is in the broadcast dup. * @bat_priv: the bat priv with all the soft interface information - * @skb: contains the bcast_packet to be checked + * @skb: contains the multicast packet to be checked + * @payload_ptr: pointer to position inside the head buffer of the skb + * marking the start of the data to be CRC'ed + * @orig: originator mac address, NULL if unknown * - * check if it is on our broadcast list. Another gateway might - * have sent the same packet because it is connected to the same backbone, - * so we have to remove this duplicate. + * Check if it is on our broadcast list. Another gateway might have sent the + * same packet because it is connected to the same backbone, so we have to + * remove this duplicate. * * This is performed by checking the CRC, which will tell us * with a good chance that it is the same packet. If it is furthermore @@ -1613,19 +1616,17 @@ int batadv_bla_init(struct batadv_priv *bat_priv) * * Return: true if a packet is in the duplicate list, false otherwise. */ -bool batadv_bla_check_bcast_duplist(struct batadv_priv *bat_priv, - struct sk_buff *skb) +static bool batadv_bla_check_duplist(struct batadv_priv *bat_priv, + struct sk_buff *skb, u8 *payload_ptr, + const u8 *orig) { - int i, curr; - __be32 crc; - struct batadv_bcast_packet *bcast_packet; struct batadv_bcast_duplist_entry *entry; bool ret = false; - - bcast_packet = (struct batadv_bcast_packet *)skb->data; + int i, curr; + __be32 crc;
/* calculate the crc ... */ - crc = batadv_skb_crc32(skb, (u8 *)(bcast_packet + 1)); + crc = batadv_skb_crc32(skb, payload_ptr);
spin_lock_bh(&bat_priv->bla.bcast_duplist_lock);
@@ -1644,8 +1645,21 @@ bool batadv_bla_check_bcast_duplist(struct batadv_priv *bat_priv, if (entry->crc != crc) continue;
- if (batadv_compare_eth(entry->orig, bcast_packet->orig)) - continue; + /* are the originators both known and not anonymous? */ + if (orig && !is_zero_ether_addr(orig) && + !is_zero_ether_addr(entry->orig)) { + /* If known, check if the new frame came from + * the same originator: + * We are safe to take identical frames from the + * same orig, if known, as multiplications in + * the mesh are detected via the (orig, seqno) pair. + * So we can be a bit more liberal here and allow + * identical frames from the same orig which the source + * host might have sent multiple times on purpose. + */ + if (batadv_compare_eth(entry->orig, orig)) + continue; + }
/* this entry seems to match: same crc, not too old, * and from another gw. therefore return true to forbid it. @@ -1661,7 +1675,14 @@ bool batadv_bla_check_bcast_duplist(struct batadv_priv *bat_priv, entry = &bat_priv->bla.bcast_duplist[curr]; entry->crc = crc; entry->entrytime = jiffies; - ether_addr_copy(entry->orig, bcast_packet->orig); + + /* known originator */ + if (orig) + ether_addr_copy(entry->orig, orig); + /* anonymous originator */ + else + eth_zero_addr(entry->orig); + bat_priv->bla.bcast_duplist_curr = curr;
out: @@ -1670,6 +1691,48 @@ bool batadv_bla_check_bcast_duplist(struct batadv_priv *bat_priv, return ret; }
+/** + * batadv_bla_check_ucast_duplist() - Check if a frame is in the broadcast dup. + * @bat_priv: the bat priv with all the soft interface information + * @skb: contains the multicast packet to be checked, decapsulated from a + * unicast_packet + * + * Check if it is on our broadcast list. Another gateway might have sent the + * same packet because it is connected to the same backbone, so we have to + * remove this duplicate. + * + * Return: true if a packet is in the duplicate list, false otherwise. + */ +static bool batadv_bla_check_ucast_duplist(struct batadv_priv *bat_priv, + struct sk_buff *skb) +{ + return batadv_bla_check_duplist(bat_priv, skb, (u8 *)skb->data, NULL); +} + +/** + * batadv_bla_check_bcast_duplist() - Check if a frame is in the broadcast dup. + * @bat_priv: the bat priv with all the soft interface information + * @skb: contains the bcast_packet to be checked + * + * Check if it is on our broadcast list. Another gateway might have sent the + * same packet because it is connected to the same backbone, so we have to + * remove this duplicate. + * + * Return: true if a packet is in the duplicate list, false otherwise. + */ +bool batadv_bla_check_bcast_duplist(struct batadv_priv *bat_priv, + struct sk_buff *skb) +{ + struct batadv_bcast_packet *bcast_packet; + u8 *payload_ptr; + + bcast_packet = (struct batadv_bcast_packet *)skb->data; + payload_ptr = (u8 *)(bcast_packet + 1); + + return batadv_bla_check_duplist(bat_priv, skb, payload_ptr, + bcast_packet->orig); +} + /** * batadv_bla_is_backbone_gw_orig - Check if the originator is a gateway for * the VLAN identified by vid. @@ -1884,6 +1947,14 @@ bool batadv_bla_rx(struct batadv_priv *bat_priv, struct sk_buff *skb, packet_type == BATADV_UNICAST) goto handled;
+ /* potential duplicates from foreign BLA backbone gateways via + * multicast-in-unicast packets + */ + if (is_multicast_ether_addr(ethhdr->h_dest) && + packet_type == BATADV_UNICAST && + batadv_bla_check_ucast_duplist(bat_priv, skb)) + goto handled; + ether_addr_copy(search_claim.addr, ethhdr->h_source); search_claim.vid = vid; claim = batadv_claim_hash_find(bat_priv, &search_claim);
commit 4ca23e2c2074465bff55ea14221175fecdf63c5f upstream.
If a batman-adv packets has to be fragmented, then the original batman-adv packet header is not stripped away. Instead, only a new header is added in front of the packet after it was split.
This size must be considered to avoid cost intensive reallocations during the transmission through the various device layers.
Fixes: 7bca68c7844b ("batman-adv: Add lower layer needed_(head|tail)room to own ones") Reported-by: Linus Lüssing linus.luessing@c0d3.blue Signed-off-by: Sven Eckelmann sven@narfation.org Signed-off-by: Simon Wunderlich sw@simonwunderlich.de Signed-off-by: Sven Eckelmann sven@narfation.org --- net/batman-adv/hard-interface.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c index 9fdfa9984f02..7bdc5f26442e 100644 --- a/net/batman-adv/hard-interface.c +++ b/net/batman-adv/hard-interface.c @@ -558,6 +558,9 @@ static void batadv_hardif_recalc_extra_skbroom(struct net_device *soft_iface) needed_headroom = lower_headroom + (lower_header_len - ETH_HLEN); needed_headroom += batadv_max_header_len();
+ /* fragmentation headers don't strip the unicast/... header */ + needed_headroom += sizeof(struct batadv_frag_packet); + soft_iface->needed_headroom = needed_headroom; soft_iface->needed_tailroom = lower_tailroom; }
commit c5cbfc87558168ef4c3c27ce36eba6b83391db19 upstream.
The batadv net_device is trying to propagate the needed_headroom and needed_tailroom from the lower devices. This is needed to avoid cost intensive reallocations using pskb_expand_head during the transmission.
But the fragmentation code split the skb's without adding extra room at the end/beginning of the various fragments. This reduced the performance of transmissions over complex scenarios (batadv on vxlan on wireguard) because the lower devices had to perform the reallocations at least once.
Fixes: ee75ed88879a ("batman-adv: Fragment and send skbs larger than mtu") Signed-off-by: Sven Eckelmann sven@narfation.org Signed-off-by: Simon Wunderlich sw@simonwunderlich.de [ bp: 4.14 backported: adjust context. ] Signed-off-by: Sven Eckelmann sven@narfation.org --- net/batman-adv/fragmentation.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/net/batman-adv/fragmentation.c b/net/batman-adv/fragmentation.c index 788d62073964..f1446a04f055 100644 --- a/net/batman-adv/fragmentation.c +++ b/net/batman-adv/fragmentation.c @@ -402,6 +402,7 @@ bool batadv_frag_skb_fwd(struct sk_buff *skb,
/** * batadv_frag_create - create a fragment from skb + * @net_dev: outgoing device for fragment * @skb: skb to create fragment from * @frag_head: header to use in new fragment * @fragment_size: size of new fragment @@ -412,22 +413,25 @@ bool batadv_frag_skb_fwd(struct sk_buff *skb, * * Return: the new fragment, NULL on error. */ -static struct sk_buff *batadv_frag_create(struct sk_buff *skb, +static struct sk_buff *batadv_frag_create(struct net_device *net_dev, + struct sk_buff *skb, struct batadv_frag_packet *frag_head, unsigned int fragment_size) { + unsigned int ll_reserved = LL_RESERVED_SPACE(net_dev); + unsigned int tailroom = net_dev->needed_tailroom; struct sk_buff *skb_fragment; unsigned int header_size = sizeof(*frag_head); unsigned int mtu = fragment_size + header_size;
- skb_fragment = netdev_alloc_skb(NULL, mtu + ETH_HLEN); + skb_fragment = dev_alloc_skb(ll_reserved + mtu + tailroom); if (!skb_fragment) goto err;
skb_fragment->priority = skb->priority;
/* Eat the last mtu-bytes of the skb */ - skb_reserve(skb_fragment, header_size + ETH_HLEN); + skb_reserve(skb_fragment, ll_reserved + header_size); skb_split(skb, skb_fragment, skb->len - fragment_size);
/* Add the header */ @@ -450,11 +454,12 @@ int batadv_frag_send_packet(struct sk_buff *skb, struct batadv_orig_node *orig_node, struct batadv_neigh_node *neigh_node) { + struct net_device *net_dev = neigh_node->if_incoming->net_dev; struct batadv_priv *bat_priv; struct batadv_hard_iface *primary_if = NULL; struct batadv_frag_packet frag_header; struct sk_buff *skb_fragment; - unsigned int mtu = neigh_node->if_incoming->net_dev->mtu; + unsigned int mtu = net_dev->mtu; unsigned int header_size = sizeof(frag_header); unsigned int max_fragment_size, num_fragments; int ret; @@ -514,7 +519,7 @@ int batadv_frag_send_packet(struct sk_buff *skb, goto put_primary_if; }
- skb_fragment = batadv_frag_create(skb, &frag_header, + skb_fragment = batadv_frag_create(net_dev, skb, &frag_header, max_fragment_size); if (!skb_fragment) { ret = -ENOMEM;
commit 992b03b88e36254e26e9a4977ab948683e21bd9f upstream.
When a packet is fragmented by batman-adv, the original batman-adv header is not modified. Only a new fragmentation is inserted between the original one and the ethernet header. The code must therefore make sure that it has a writable region of this size in the skbuff head.
But it is not useful to always reallocate the skbuff by this size even when there would be more than enough headroom still in the skb. The reallocation is just to costly during in this codepath.
Fixes: ee75ed88879a ("batman-adv: Fragment and send skbs larger than mtu") Signed-off-by: Sven Eckelmann sven@narfation.org Signed-off-by: Simon Wunderlich sw@simonwunderlich.de Signed-off-by: Sven Eckelmann sven@narfation.org --- net/batman-adv/fragmentation.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/net/batman-adv/fragmentation.c b/net/batman-adv/fragmentation.c index f1446a04f055..4842436c55f3 100644 --- a/net/batman-adv/fragmentation.c +++ b/net/batman-adv/fragmentation.c @@ -538,13 +538,14 @@ int batadv_frag_send_packet(struct sk_buff *skb, frag_header.no++; }
- /* Make room for the fragment header. */ - if (batadv_skb_head_push(skb, header_size) < 0 || - pskb_expand_head(skb, header_size + ETH_HLEN, 0, GFP_ATOMIC) < 0) { - ret = -ENOMEM; + /* make sure that there is at least enough head for the fragmentation + * and ethernet headers + */ + ret = skb_cow_head(skb, ETH_HLEN + header_size); + if (ret < 0) goto put_primary_if; - }
+ skb_push(skb, header_size); memcpy(skb->data, &frag_header, header_size);
/* Send the last fragment */
linux-stable-mirror@lists.linaro.org
-
Sven Eckelmann