With this patch IGMP or MLD reports are always flooded. This is necessary for the upcoming bridge integration to function without multicast packet loss.

With the report handling so far bridges might miss interested multicast listeners, leading to wrongly excluding ports from multicast packet forwarding.

Currently we are treating IGMP/MLD reports, the messages bridges use to learn about interested multicast listeners, just as any other multicast packet: We try to send them to nodes matching its multicast destination.

Unfortunately, the destination address of reports of the older IGMPv2/MLDv1 protocol families do not strictly adhere to their own protocol: More precisely, the interested receiver, an IGMPv2 or MLDv1 querier, itself usually does not listen to the multicast destination address of any reports.

Therefore with this patch we are simply excluding IGMP/MLD reports from the multicast forwarding code path and keep flooding them. By that any bridge receives them and can properly learn about listeners.

To avoid compatibility issues with older nodes not yet implementing this report handling, we need to force them to flood reports: We do this by bumping the multicast TVLV version to 2, effectively disabling their multicast optimization.

Tested-by: Simon Wunderlich sw@simonwunderlich.de Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue --- Makefile | 4 + compat-include/linux/igmp.h | 13 ++ compat-include/linux/skbuff.h | 14 ++ compat-include/net/addrconf.h | 13 ++ compat-include/net/ip6_checksum.h | 18 +++ compat-include/net/ipv6.h | 17 +++ compat-sources/Makefile | 3 + compat-sources/net/core/skbuff.c | 212 +++++++++++++++++++++++++++++ compat-sources/net/ipv4/igmp.c | 241 +++++++++++++++++++++++++++++++++ compat-sources/net/ipv6/mcast_snoop.c | 216 +++++++++++++++++++++++++++++ net/batman-adv/Kconfig | 2 +- net/batman-adv/multicast.c | 85 ++++++++++-- 12 files changed, 825 insertions(+), 13 deletions(-) create mode 100644 compat-include/linux/igmp.h create mode 100644 compat-include/net/addrconf.h create mode 100644 compat-include/net/ip6_checksum.h create mode 100644 compat-include/net/ipv6.h create mode 100644 compat-sources/Makefile create mode 100644 compat-sources/net/core/skbuff.c create mode 100644 compat-sources/net/ipv4/igmp.c create mode 100644 compat-sources/net/ipv6/mcast_snoop.c

diff --git a/Makefile b/Makefile index 5d2c058..fd6237c 100644 --- a/Makefile +++ b/Makefile @@ -52,6 +52,10 @@ ifneq ($(REVISION),) NOSTDINC_FLAGS += -DBATADV_SOURCE_VERSION="$(REVISION)" endif

+include $(PWD)/compat-sources/Makefile + +export batman-adv-y + BUILD_FLAGS := \ M=$(PWD)/net/batman-adv \ CONFIG_BATMAN_ADV=m \ diff --git a/compat-include/linux/igmp.h b/compat-include/linux/igmp.h new file mode 100644 index 0000000..f61ab79 --- /dev/null +++ b/compat-include/linux/igmp.h @@ -0,0 +1,13 @@ +#ifndef _NET_BATMAN_ADV_COMPAT_LINUX_IGMP_H_ +#define _NET_BATMAN_ADV_COMPAT_LINUX_IGMP_H_ + +#include <linux/version.h> +#include_next <linux/igmp.h> + +#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 2, 0) + +int ip_mc_check_igmp(struct sk_buff *skb, struct sk_buff **skb_trimmed); + +#endif /* < KERNEL_VERSION(4, 2, 0) */ + +#endif /* _NET_BATMAN_ADV_COMPAT_LINUX_IGMP_H_ */ diff --git a/compat-include/linux/skbuff.h b/compat-include/linux/skbuff.h index b2cb161..9dd92a5 100644 --- a/compat-include/linux/skbuff.h +++ b/compat-include/linux/skbuff.h @@ -60,6 +60,20 @@

#define pskb_copy_for_clone pskb_copy

+__sum16 skb_checksum_simple_validate(struct sk_buff *skb); + +__sum16 +skb_checksum_validate(struct sk_buff *skb, int proto, + __wsum (*compute_pseudo)(struct sk_buff *skb, int proto)); + #endif /* < KERNEL_VERSION(3, 16, 0) */

+#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 2, 0) + +struct sk_buff *skb_checksum_trimmed(struct sk_buff *skb, + unsigned int transport_len, + __sum16(*skb_chkf)(struct sk_buff *skb)); + +#endif /* < KERNEL_VERSION(4, 2, 0) */ + #endif /* _NET_BATMAN_ADV_COMPAT_LINUX_SKBUFF_H_ */ diff --git a/compat-include/net/addrconf.h b/compat-include/net/addrconf.h new file mode 100644 index 0000000..69c45d0 --- /dev/null +++ b/compat-include/net/addrconf.h @@ -0,0 +1,13 @@ +#ifndef _NET_BATMAN_ADV_COMPAT_NET_ADDRCONF_H_ +#define _NET_BATMAN_ADV_COMPAT_NET_ADDRCONF_H_ + +#include <linux/version.h> +#include_next <net/addrconf.h> + +#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 2, 0) + +int ipv6_mc_check_mld(struct sk_buff *skb, struct sk_buff **skb_trimmed); + +#endif /* < KERNEL_VERSION(4, 2, 0) */ + +#endif /* _NET_BATMAN_ADV_COMPAT_NET_ADDRCONF_H_ */ diff --git a/compat-include/net/ip6_checksum.h b/compat-include/net/ip6_checksum.h new file mode 100644 index 0000000..fda0c07 --- /dev/null +++ b/compat-include/net/ip6_checksum.h @@ -0,0 +1,18 @@ +#ifndef _NET_BATMAN_ADV_COMPAT_NET_IP6_CHECKSUM_H_ +#define _NET_BATMAN_ADV_COMPAT_NET_IP6_CHECKSUM_H_ + +#include <linux/version.h> +#include_next <net/ip6_checksum.h> + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 16, 0) + +static inline __wsum ip6_compute_pseudo(struct sk_buff *skb, int proto) +{ + return ~csum_unfold(csum_ipv6_magic(&ipv6_hdr(skb)->saddr, + &ipv6_hdr(skb)->daddr, + skb->len, proto, 0)); +} + +#endif /* < KERNEL_VERSION(3, 16, 0) */ + +#endif /* _NET_BATMAN_ADV_COMPAT_NET_IP6_CHECKSUM_H_ */ diff --git a/compat-include/net/ipv6.h b/compat-include/net/ipv6.h new file mode 100644 index 0000000..1e190d8 --- /dev/null +++ b/compat-include/net/ipv6.h @@ -0,0 +1,17 @@ +#ifndef _NET_BATMAN_ADV_COMPAT_NET_IPV6_H_ +#define _NET_BATMAN_ADV_COMPAT_NET_IPV6_H_ + +#include <linux/version.h> +#include_next <net/ipv6.h> + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0) + +#define ipv6_skip_exthdr(skb, start, nexthdrp, frag_offp) \ + ({ \ + (void)frag_offp; \ + ipv6_skip_exthdr(skb, start, nexthdrp); \ + }) + +#endif /* < KERNEL_VERSION(3, 3, 0) */ + +#endif /* _NET_BATMAN_ADV_COMPAT_NET_IPV6_H_ */ diff --git a/compat-sources/Makefile b/compat-sources/Makefile new file mode 100644 index 0000000..a45c202 --- /dev/null +++ b/compat-sources/Makefile @@ -0,0 +1,3 @@ +batman-adv-$(CONFIG_BATMAN_ADV_MCAST) += ../../compat-sources/net/core/skbuff.o +batman-adv-$(CONFIG_BATMAN_ADV_MCAST) += ../../compat-sources/net/ipv4/igmp.o +batman-adv-$(CONFIG_BATMAN_ADV_MCAST) += ../../compat-sources/net/ipv6/mcast_snoop.o diff --git a/compat-sources/net/core/skbuff.c b/compat-sources/net/core/skbuff.c new file mode 100644 index 0000000..44d59dc --- /dev/null +++ b/compat-sources/net/core/skbuff.c @@ -0,0 +1,212 @@ +/* + * Routines having to do with the 'struct sk_buff' memory handlers. + * + * Authors: Alan Cox alan@lxorguk.ukuu.org.uk + * Florian La Roche rzsfl@rz.uni-sb.de + * + * Fixes: + * Alan Cox : Fixed the worst of the load + * balancer bugs. + * Dave Platt : Interrupt stacking fix. + * Richard Kooijman : Timestamp fixes. + * Alan Cox : Changed buffer format. + * Alan Cox : destructor hook for AF_UNIX etc. + * Linus Torvalds : Better skb_clone. + * Alan Cox : Added skb_copy. + * Alan Cox : Added all the changed routines Linus + * only put in the headers + * Ray VanTassle : Fixed --skb->lock in free + * Alan Cox : skb_copy copy arp field + * Andi Kleen : slabified it. + * Robert Olsson : Removed skb_head_pool + * + * NOTE: + * The __skb_ routines should be called with interrupts + * disabled, or you better be *real* sure that the operation is atomic + * with respect to whatever list is being frobbed (e.g. via lock_sock() + * or via disabling bottom half handlers, etc). + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#include <linux/ipv6.h> +#include <linux/skbuff.h> + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 16, 0) + +/* Compare with: + * "bridge: multicast: call skb_checksum_{simple_, }validate" + */ +__sum16 skb_checksum_simple_validate(struct sk_buff *skb) +{ + switch (skb->ip_summed) { + case CHECKSUM_COMPLETE: + if (!csum_fold(skb->csum)) + break; + /* fall through */ + case CHECKSUM_NONE: + skb->csum = 0; + return skb_checksum_complete(skb); + } + + return 0; +} + +/* Watch out: Not as generic as upstream + * - redefines this method to only fit with ICMPV6 + * + * Compare with: + * "bridge: multicast: call skb_checksum_{simple_, }validate" + */ +__sum16 +skb_checksum_validate(struct sk_buff *skb, int proto, + __wsum (*compute_pseudo)(struct sk_buff *skb, int proto)) +{ + const struct ipv6hdr *ip6h = ipv6_hdr(skb); + + switch (skb->ip_summed) { + case CHECKSUM_COMPLETE: + if (!csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr, skb->len, + IPPROTO_ICMPV6, skb->csum)) + break; + /*FALLTHROUGH*/ + case CHECKSUM_NONE: + skb->csum = ~csum_unfold(csum_ipv6_magic(&ip6h->saddr, + &ip6h->daddr, + skb->len, + IPPROTO_ICMPV6, 0)); + return __skb_checksum_complete(skb); + } + + return 0; +} + +#endif /* < KERNEL_VERSION(3, 16, 0) */ + +#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 2, 0) + +static inline void skb_postpush_rcsum(struct sk_buff *skb, + const void *start, unsigned int len) +{ + /* For performing the reverse operation to skb_postpull_rcsum(), + * we can instead of ... + * + * skb->csum = csum_add(skb->csum, csum_partial(start, len, 0)); + * + * ... just use this equivalent version here to save a few + * instructions. Feeding csum of 0 in csum_partial() and later + * on adding skb->csum is equivalent to feed skb->csum in the + * first place. + */ + if (skb->ip_summed == CHECKSUM_COMPLETE) + skb->csum = csum_partial(start, len, skb->csum); +} + +/** + * skb_push_rcsum - push skb and update receive checksum + * @skb: buffer to update + * @len: length of data pulled + * + * This function performs an skb_push on the packet and updates + * the CHECKSUM_COMPLETE checksum. It should be used on + * receive path processing instead of skb_push unless you know + * that the checksum difference is zero (e.g., a valid IP header) + * or you are setting ip_summed to CHECKSUM_NONE. + */ +static unsigned char *skb_push_rcsum(struct sk_buff *skb, unsigned len) +{ + skb_push(skb, len); + skb_postpush_rcsum(skb, skb->data, len); + return skb->data; +} + +/** + * skb_checksum_maybe_trim - maybe trims the given skb + * @skb: the skb to check + * @transport_len: the data length beyond the network header + * + * Checks whether the given skb has data beyond the given transport length. + * If so, returns a cloned skb trimmed to this transport length. + * Otherwise returns the provided skb. Returns NULL in error cases + * (e.g. transport_len exceeds skb length or out-of-memory). + * + * Caller needs to set the skb transport header and release the returned skb. + * Provided skb is consumed. + */ +static struct sk_buff *skb_checksum_maybe_trim(struct sk_buff *skb, + unsigned int transport_len) +{ + struct sk_buff *skb_chk; + unsigned int len = skb_transport_offset(skb) + transport_len; + int ret; + + if (skb->len < len) { + kfree_skb(skb); + return NULL; + } else if (skb->len == len) { + return skb; + } + + skb_chk = skb_clone(skb, GFP_ATOMIC); + kfree_skb(skb); + + if (!skb_chk) + return NULL; + + ret = pskb_trim_rcsum(skb_chk, len); + if (ret) { + kfree_skb(skb_chk); + return NULL; + } + + return skb_chk; +} + +/** + * skb_checksum_trimmed - validate checksum of an skb + * @skb: the skb to check + * @transport_len: the data length beyond the network header + * @skb_chkf: checksum function to use + * + * Applies the given checksum function skb_chkf to the provided skb. + * Returns a checked and maybe trimmed skb. Returns NULL on error. + * + * If the skb has data beyond the given transport length, then a + * trimmed & cloned skb is checked and returned. + * + * Caller needs to set the skb transport header and release the returned skb. + * Provided skb is consumed. + */ +struct sk_buff *skb_checksum_trimmed(struct sk_buff *skb, + unsigned int transport_len, + __sum16(*skb_chkf)(struct sk_buff *skb)) +{ + struct sk_buff *skb_chk; + unsigned int offset = skb_transport_offset(skb); + __sum16 ret; + + skb_chk = skb_checksum_maybe_trim(skb, transport_len); + if (!skb_chk) + return NULL; + + if (!pskb_may_pull(skb_chk, offset)) { + kfree_skb(skb_chk); + return NULL; + } + + skb_pull_rcsum(skb_chk, offset); + ret = skb_chkf(skb_chk); + skb_push_rcsum(skb_chk, offset); + + if (ret) { + kfree_skb(skb_chk); + return NULL; + } + + return skb_chk; +} + +#endif /* < KERNEL_VERSION(4, 2, 0) */ diff --git a/compat-sources/net/ipv4/igmp.c b/compat-sources/net/ipv4/igmp.c new file mode 100644 index 0000000..457a05e --- /dev/null +++ b/compat-sources/net/ipv4/igmp.c @@ -0,0 +1,241 @@ +/* + * Linux NET3: Internet Group Management Protocol [IGMP] + * + * This code implements the IGMP protocol as defined in RFC1112. There has + * been a further revision of this protocol since which is now supported. + * + * If you have trouble with this module be careful what gcc you have used, + * the older version didn't come out right using gcc 2.5.8, the newer one + * seems to fall out with gcc 2.6.2. + * + * Authors: + * Alan Cox alan@lxorguk.ukuu.org.uk + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + * + * Fixes: + * + * Alan Cox : Added lots of __inline__ to optimise + * the memory usage of all the tiny little + * functions. + * Alan Cox : Dumped the header building experiment. + * Alan Cox : Minor tweaks ready for multicast routing + * and extended IGMP protocol. + * Alan Cox : Removed a load of inline directives. Gcc 2.5.8 + * writes utterly bogus code otherwise (sigh) + * fixed IGMP loopback to behave in the manner + * desired by mrouted, fixed the fact it has been + * broken since 1.3.6 and cleaned up a few minor + * points. + * + * Chih-Jen Chang : Tried to revise IGMP to Version 2 + * Tsu-Sheng Tsao E-mail: chihjenc@scf.usc.edu and tsusheng@scf.usc.edu + * The enhancements are mainly based on Steve Deering's + * ipmulti-3.5 source code. + * Chih-Jen Chang : Added the igmp_get_mrouter_info and + * Tsu-Sheng Tsao igmp_set_mrouter_info to keep track of + * the mrouted version on that device. + * Chih-Jen Chang : Added the max_resp_time parameter to + * Tsu-Sheng Tsao igmp_heard_query(). Using this parameter + * to identify the multicast router version + * and do what the IGMP version 2 specified. + * Chih-Jen Chang : Added a timer to revert to IGMP V2 router + * Tsu-Sheng Tsao if the specified time expired. + * Alan Cox : Stop IGMP from 0.0.0.0 being accepted. + * Alan Cox : Use GFP_ATOMIC in the right places. + * Christian Daudt : igmp timer wasn't set for local group + * memberships but was being deleted, + * which caused a "del_timer() called + * from %p with timer not initialized\n" + * message (960131). + * Christian Daudt : removed del_timer from + * igmp_timer_expire function (960205). + * Christian Daudt : igmp_heard_report now only calls + * igmp_timer_expire if tm->running is + * true (960216). + * Malcolm Beattie : ttl comparison wrong in igmp_rcv made + * igmp_heard_query never trigger. Expiry + * miscalculation fixed in igmp_heard_query + * and random() made to return unsigned to + * prevent negative expiry times. + * Alexey Kuznetsov: Wrong group leaving behaviour, backport + * fix from pending 2.1.x patches. + * Alan Cox: Forget to enable FDDI support earlier. + * Alexey Kuznetsov: Fixed leaving groups on device down. + * Alexey Kuznetsov: Accordance to igmp-v2-06 draft. + * David L Stevens: IGMPv3 support, with help from + * Vinay Kulkarni + */ + +#include <linux/igmp.h> +#include <linux/ip.h> +#include <linux/skbuff.h> +#include <net/ip.h> + +#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 2, 0) + +static int ip_mc_check_iphdr(struct sk_buff *skb) +{ + const struct iphdr *iph; + unsigned int len; + unsigned int offset = skb_network_offset(skb) + sizeof(*iph); + + if (!pskb_may_pull(skb, offset)) + return -EINVAL; + + iph = ip_hdr(skb); + + if (iph->version != 4 || ip_hdrlen(skb) < sizeof(*iph)) + return -EINVAL; + + offset += ip_hdrlen(skb) - sizeof(*iph); + + if (!pskb_may_pull(skb, offset)) + return -EINVAL; + + iph = ip_hdr(skb); + + if (unlikely(ip_fast_csum((u8 *)iph, iph->ihl))) + return -EINVAL; + + len = skb_network_offset(skb) + ntohs(iph->tot_len); + if (skb->len < len || len < offset) + return -EINVAL; + + skb_set_transport_header(skb, offset); + + return 0; +} + +static int ip_mc_check_igmp_reportv3(struct sk_buff *skb) +{ + unsigned int len = skb_transport_offset(skb); + + len += sizeof(struct igmpv3_report); + + return pskb_may_pull(skb, len) ? 0 : -EINVAL; +} + +static int ip_mc_check_igmp_query(struct sk_buff *skb) +{ + unsigned int len = skb_transport_offset(skb); + + len += sizeof(struct igmphdr); + if (skb->len < len) + return -EINVAL; + + /* IGMPv{1,2}? */ + if (skb->len != len) { + /* or IGMPv3? */ + len += sizeof(struct igmpv3_query) - sizeof(struct igmphdr); + if (skb->len < len || !pskb_may_pull(skb, len)) + return -EINVAL; + } + + /* RFC2236+RFC3376 (IGMPv2+IGMPv3) require the multicast link layer + * all-systems destination addresses (224.0.0.1) for general queries + */ + if (!igmp_hdr(skb)->group && + ip_hdr(skb)->daddr != htonl(INADDR_ALLHOSTS_GROUP)) + return -EINVAL; + + return 0; +} + +static int ip_mc_check_igmp_msg(struct sk_buff *skb) +{ + switch (igmp_hdr(skb)->type) { + case IGMP_HOST_LEAVE_MESSAGE: + case IGMP_HOST_MEMBERSHIP_REPORT: + case IGMPV2_HOST_MEMBERSHIP_REPORT: + /* fall through */ + return 0; + case IGMPV3_HOST_MEMBERSHIP_REPORT: + return ip_mc_check_igmp_reportv3(skb); + case IGMP_HOST_MEMBERSHIP_QUERY: + return ip_mc_check_igmp_query(skb); + default: + return -ENOMSG; + } +} + +static inline __sum16 ip_mc_validate_checksum(struct sk_buff *skb) +{ + return skb_checksum_simple_validate(skb); +} + +static int __ip_mc_check_igmp(struct sk_buff *skb, struct sk_buff **skb_trimmed) + +{ + struct sk_buff *skb_chk; + unsigned int transport_len; + unsigned int len = skb_transport_offset(skb) + sizeof(struct igmphdr); + int ret; + + transport_len = ntohs(ip_hdr(skb)->tot_len) - ip_hdrlen(skb); + + skb_get(skb); + skb_chk = skb_checksum_trimmed(skb, transport_len, + ip_mc_validate_checksum); + if (!skb_chk) + return -EINVAL; + + if (!pskb_may_pull(skb_chk, len)) { + kfree_skb(skb_chk); + return -EINVAL; + } + + ret = ip_mc_check_igmp_msg(skb_chk); + if (ret) { + kfree_skb(skb_chk); + return ret; + } + + if (skb_trimmed) + *skb_trimmed = skb_chk; + else + kfree_skb(skb_chk); + + return 0; +} + +/** + * ip_mc_check_igmp - checks whether this is a sane IGMP packet + * @skb: the skb to validate + * @skb_trimmed: to store an skb pointer trimmed to IPv4 packet tail (optional) + * + * Checks whether an IPv4 packet is a valid IGMP packet. If so sets + * skb network and transport headers accordingly and returns zero. + * + * -EINVAL: A broken packet was detected, i.e. it violates some internet + * standard + * -ENOMSG: IP header validation succeeded but it is not an IGMP packet. + * -ENOMEM: A memory allocation failure happened. + * + * Optionally, an skb pointer might be provided via skb_trimmed (or set it + * to NULL): After parsing an IGMP packet successfully it will point to + * an skb which has its tail aligned to the IP packet end. This might + * either be the originally provided skb or a trimmed, cloned version if + * the skb frame had data beyond the IP packet. A cloned skb allows us + * to leave the original skb and its full frame unchanged (which might be + * desirable for layer 2 frame jugglers). + * + * The caller needs to release a reference count from any returned skb_trimmed. + */ +int ip_mc_check_igmp(struct sk_buff *skb, struct sk_buff **skb_trimmed) +{ + int ret = ip_mc_check_iphdr(skb); + + if (ret < 0) + return ret; + + if (ip_hdr(skb)->protocol != IPPROTO_IGMP) + return -ENOMSG; + + return __ip_mc_check_igmp(skb, skb_trimmed); +} + +#endif /* < KERNEL_VERSION(4, 2, 0) */ diff --git a/compat-sources/net/ipv6/mcast_snoop.c b/compat-sources/net/ipv6/mcast_snoop.c new file mode 100644 index 0000000..3f46ed3 --- /dev/null +++ b/compat-sources/net/ipv6/mcast_snoop.c @@ -0,0 +1,216 @@ +/* Copyright (C) 2010: YOSHIFUJI Hideaki yoshfuji@linux-ipv6.org + * Copyright (C) 2015: Linus Lüssing linus.luessing@c0d3.blue + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of version 2 of the GNU General Public + * License as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, see http://www.gnu.org/licenses/. + * + * + * Based on the MLD support added to br_multicast.c by YOSHIFUJI Hideaki. + */ + +#include <linux/skbuff.h> +#include <net/ipv6.h> +#include <net/mld.h> +#include <net/addrconf.h> +#include <net/ip6_checksum.h> + +#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 2, 0) + +static int ipv6_mc_check_ip6hdr(struct sk_buff *skb) +{ + const struct ipv6hdr *ip6h; + unsigned int len; + unsigned int offset = skb_network_offset(skb) + sizeof(*ip6h); + + if (!pskb_may_pull(skb, offset)) + return -EINVAL; + + ip6h = ipv6_hdr(skb); + + if (ip6h->version != 6) + return -EINVAL; + + len = offset + ntohs(ip6h->payload_len); + if (skb->len < len || len <= offset) + return -EINVAL; + + return 0; +} + +static int ipv6_mc_check_exthdrs(struct sk_buff *skb) +{ + const struct ipv6hdr *ip6h; + int offset; + u8 nexthdr; + __be16 frag_off; + + ip6h = ipv6_hdr(skb); + + if (ip6h->nexthdr != IPPROTO_HOPOPTS) + return -ENOMSG; + + nexthdr = ip6h->nexthdr; + offset = skb_network_offset(skb) + sizeof(*ip6h); + offset = ipv6_skip_exthdr(skb, offset, &nexthdr, &frag_off); + + if (offset < 0) + return -EINVAL; + + if (nexthdr != IPPROTO_ICMPV6) + return -ENOMSG; + + skb_set_transport_header(skb, offset); + + return 0; +} + +static int ipv6_mc_check_mld_reportv2(struct sk_buff *skb) +{ + unsigned int len = skb_transport_offset(skb); + + len += sizeof(struct mld2_report); + + return pskb_may_pull(skb, len) ? 0 : -EINVAL; +} + +static int ipv6_mc_check_mld_query(struct sk_buff *skb) +{ + struct mld_msg *mld; + unsigned int len = skb_transport_offset(skb); + + /* RFC2710+RFC3810 (MLDv1+MLDv2) require link-local source addresses */ + if (!(ipv6_addr_type(&ipv6_hdr(skb)->saddr) & IPV6_ADDR_LINKLOCAL)) + return -EINVAL; + + len += sizeof(struct mld_msg); + if (skb->len < len) + return -EINVAL; + + /* MLDv1? */ + if (skb->len != len) { + /* or MLDv2? */ + len += sizeof(struct mld2_query) - sizeof(struct mld_msg); + if (skb->len < len || !pskb_may_pull(skb, len)) + return -EINVAL; + } + + mld = (struct mld_msg *)skb_transport_header(skb); + + /* RFC2710+RFC3810 (MLDv1+MLDv2) require the multicast link layer + * all-nodes destination address (ff02::1) for general queries + */ + if (ipv6_addr_any(&mld->mld_mca) && + !ipv6_addr_is_ll_all_nodes(&ipv6_hdr(skb)->daddr)) + return -EINVAL; + + return 0; +} + +static int ipv6_mc_check_mld_msg(struct sk_buff *skb) +{ + struct mld_msg *mld = (struct mld_msg *)skb_transport_header(skb); + + switch (mld->mld_type) { + case ICMPV6_MGM_REDUCTION: + case ICMPV6_MGM_REPORT: + /* fall through */ + return 0; + case ICMPV6_MLD2_REPORT: + return ipv6_mc_check_mld_reportv2(skb); + case ICMPV6_MGM_QUERY: + return ipv6_mc_check_mld_query(skb); + default: + return -ENOMSG; + } +} + +static inline __sum16 ipv6_mc_validate_checksum(struct sk_buff *skb) +{ + return skb_checksum_validate(skb, IPPROTO_ICMPV6, ip6_compute_pseudo); +} + +static int __ipv6_mc_check_mld(struct sk_buff *skb, + struct sk_buff **skb_trimmed) + +{ + struct sk_buff *skb_chk = NULL; + unsigned int transport_len; + unsigned int len = skb_transport_offset(skb) + sizeof(struct mld_msg); + int ret; + + transport_len = ntohs(ipv6_hdr(skb)->payload_len); + transport_len -= skb_transport_offset(skb) - sizeof(struct ipv6hdr); + + skb_get(skb); + skb_chk = skb_checksum_trimmed(skb, transport_len, + ipv6_mc_validate_checksum); + if (!skb_chk) + return -EINVAL; + + if (!pskb_may_pull(skb_chk, len)) { + kfree_skb(skb_chk); + return -EINVAL; + } + + ret = ipv6_mc_check_mld_msg(skb_chk); + if (ret) { + kfree_skb(skb_chk); + return ret; + } + + if (skb_trimmed) + *skb_trimmed = skb_chk; + else + kfree_skb(skb_chk); + + return 0; +} + +/** + * ipv6_mc_check_mld - checks whether this is a sane MLD packet + * @skb: the skb to validate + * @skb_trimmed: to store an skb pointer trimmed to IPv6 packet tail (optional) + * + * Checks whether an IPv6 packet is a valid MLD packet. If so sets + * skb network and transport headers accordingly and returns zero. + * + * -EINVAL: A broken packet was detected, i.e. it violates some internet + * standard + * -ENOMSG: IP header validation succeeded but it is not an MLD packet. + * -ENOMEM: A memory allocation failure happened. + * + * Optionally, an skb pointer might be provided via skb_trimmed (or set it + * to NULL): After parsing an MLD packet successfully it will point to + * an skb which has its tail aligned to the IP packet end. This might + * either be the originally provided skb or a trimmed, cloned version if + * the skb frame had data beyond the IP packet. A cloned skb allows us + * to leave the original skb and its full frame unchanged (which might be + * desirable for layer 2 frame jugglers). + * + * The caller needs to release a reference count from any returned skb_trimmed. + */ +int ipv6_mc_check_mld(struct sk_buff *skb, struct sk_buff **skb_trimmed) +{ + int ret; + + ret = ipv6_mc_check_ip6hdr(skb); + if (ret < 0) + return ret; + + ret = ipv6_mc_check_exthdrs(skb); + if (ret < 0) + return ret; + + return __ipv6_mc_check_mld(skb, skb_trimmed); +} + +#endif /* < KERNEL_VERSION(4, 2, 0) */ diff --git a/net/batman-adv/Kconfig b/net/batman-adv/Kconfig index f66930e..b7ba97d 100644 --- a/net/batman-adv/Kconfig +++ b/net/batman-adv/Kconfig @@ -66,7 +66,7 @@ config BATMAN_ADV_NC

config BATMAN_ADV_MCAST bool "Multicast optimisation" - depends on BATMAN_ADV + depends on BATMAN_ADV && INET default n help This option enables the multicast optimisation which aims to diff --git a/net/batman-adv/multicast.c b/net/batman-adv/multicast.c index c32f24f..b8662f1 100644 --- a/net/batman-adv/multicast.c +++ b/net/batman-adv/multicast.c @@ -25,7 +25,9 @@ #include <linux/errno.h> #include <linux/etherdevice.h> #include <linux/fs.h> +#include <linux/icmpv6.h> #include <linux/if_ether.h> +#include <linux/igmp.h> #include <linux/in6.h> #include <linux/in.h> #include <linux/ip.h> @@ -236,7 +238,7 @@ static bool batadv_mcast_mla_tvlv_update(struct batadv_priv *bat_priv) if (batadv_mcast_has_bridge(bat_priv)) { if (bat_priv->mcast.enabled) { batadv_tvlv_container_unregister(bat_priv, - BATADV_TVLV_MCAST, 1); + BATADV_TVLV_MCAST, 2); bat_priv->mcast.enabled = false; }

@@ -245,7 +247,7 @@ static bool batadv_mcast_mla_tvlv_update(struct batadv_priv *bat_priv)

if (!bat_priv->mcast.enabled || mcast_data.flags != bat_priv->mcast.flags) { - batadv_tvlv_container_register(bat_priv, BATADV_TVLV_MCAST, 1, + batadv_tvlv_container_register(bat_priv, BATADV_TVLV_MCAST, 2, &mcast_data, sizeof(mcast_data)); bat_priv->mcast.flags = mcast_data.flags; bat_priv->mcast.enabled = true; @@ -283,6 +285,31 @@ out: }

/** + * batadv_mcast_is_report_ipv4 - check for IGMP reports + * @skb: the ethernet frame destined for the mesh + * + * This call might reallocate skb data. + * + * Checks whether the given frame is a valid IGMP report. + * + * Return: If so then true, otherwise false. + */ +static bool batadv_mcast_is_report_ipv4(struct sk_buff *skb) +{ + if (ip_mc_check_igmp(skb, NULL) < 0) + return false; + + switch (igmp_hdr(skb)->type) { + case IGMP_HOST_MEMBERSHIP_REPORT: + case IGMPV2_HOST_MEMBERSHIP_REPORT: + case IGMPV3_HOST_MEMBERSHIP_REPORT: + return true; + } + + return false; +} + +/** * batadv_mcast_forw_mode_check_ipv4 - check for optimized forwarding potential * @bat_priv: the bat priv with all the soft interface information * @skb: the IPv4 packet to check @@ -304,6 +331,9 @@ static int batadv_mcast_forw_mode_check_ipv4(struct batadv_priv *bat_priv, if (!pskb_may_pull(skb, sizeof(struct ethhdr) + sizeof(*iphdr))) return -ENOMEM;

+ if (batadv_mcast_is_report_ipv4(skb)) + return -EINVAL; + iphdr = ip_hdr(skb);

/* TODO: Implement Multicast Router Discovery (RFC4286), @@ -320,6 +350,31 @@ static int batadv_mcast_forw_mode_check_ipv4(struct batadv_priv *bat_priv, return 0; }

+#if IS_ENABLED(CONFIG_IPV6) +/** + * batadv_mcast_is_report_ipv6 - check for MLD reports + * @skb: the ethernet frame destined for the mesh + * + * This call might reallocate skb data. + * + * Checks whether the given frame is a valid MLD report. + * + * Return: If so then true, otherwise false. + */ +static bool batadv_mcast_is_report_ipv6(struct sk_buff *skb) +{ + if (ipv6_mc_check_mld(skb, NULL) < 0) + return false; + + switch (icmp6_hdr(skb)->icmp6_type) { + case ICMPV6_MGM_REPORT: + case ICMPV6_MLD2_REPORT: + return true; + } + + return false; +} + /** * batadv_mcast_forw_mode_check_ipv6 - check for optimized forwarding potential * @bat_priv: the bat priv with all the soft interface information @@ -341,6 +396,9 @@ static int batadv_mcast_forw_mode_check_ipv6(struct batadv_priv *bat_priv, if (!pskb_may_pull(skb, sizeof(struct ethhdr) + sizeof(*ip6hdr))) return -ENOMEM;

+ if (batadv_mcast_is_report_ipv6(skb)) + return -EINVAL; + ip6hdr = ipv6_hdr(skb);

/* TODO: Implement Multicast Router Discovery (RFC4286), @@ -357,6 +415,7 @@ static int batadv_mcast_forw_mode_check_ipv6(struct batadv_priv *bat_priv,

return 0; } +#endif

/** * batadv_mcast_forw_mode_check - check for optimized forwarding potential @@ -385,9 +444,11 @@ static int batadv_mcast_forw_mode_check(struct batadv_priv *bat_priv, case ETH_P_IP: return batadv_mcast_forw_mode_check_ipv4(bat_priv, skb, is_unsnoopable); +#if IS_ENABLED(CONFIG_IPV6) case ETH_P_IPV6: return batadv_mcast_forw_mode_check_ipv6(bat_priv, skb, is_unsnoopable); +#endif default: return -EINVAL; } @@ -728,18 +789,18 @@ static void batadv_mcast_want_ipv6_update(struct batadv_priv *bat_priv, }

/** - * batadv_mcast_tvlv_ogm_handler_v1 - process incoming multicast tvlv container + * batadv_mcast_tvlv_ogm_handler - process incoming multicast tvlv container * @bat_priv: the bat priv with all the soft interface information * @orig: the orig_node of the ogm * @flags: flags indicating the tvlv state (see batadv_tvlv_handler_flags) * @tvlv_value: tvlv buffer containing the multicast data * @tvlv_value_len: tvlv buffer length */ -static void batadv_mcast_tvlv_ogm_handler_v1(struct batadv_priv *bat_priv, - struct batadv_orig_node *orig, - u8 flags, - void *tvlv_value, - u16 tvlv_value_len) +static void batadv_mcast_tvlv_ogm_handler(struct batadv_priv *bat_priv, + struct batadv_orig_node *orig, + u8 flags, + void *tvlv_value, + u16 tvlv_value_len) { bool orig_mcast_enabled = !(flags & BATADV_TVLV_HANDLER_OGM_CIFNOTFND); u8 mcast_flags = BATADV_NO_FLAGS; @@ -789,8 +850,8 @@ static void batadv_mcast_tvlv_ogm_handler_v1(struct batadv_priv *bat_priv, */ void batadv_mcast_init(struct batadv_priv *bat_priv) { - batadv_tvlv_handler_register(bat_priv, batadv_mcast_tvlv_ogm_handler_v1, - NULL, BATADV_TVLV_MCAST, 1, + batadv_tvlv_handler_register(bat_priv, batadv_mcast_tvlv_ogm_handler, + NULL, BATADV_TVLV_MCAST, 2, BATADV_TVLV_HANDLER_OGM_CIFNOTFND); }

@@ -800,8 +861,8 @@ void batadv_mcast_init(struct batadv_priv *bat_priv) */ void batadv_mcast_free(struct batadv_priv *bat_priv) { - batadv_tvlv_container_unregister(bat_priv, BATADV_TVLV_MCAST, 1); - batadv_tvlv_handler_unregister(bat_priv, BATADV_TVLV_MCAST, 1); + batadv_tvlv_container_unregister(bat_priv, BATADV_TVLV_MCAST, 2); + batadv_tvlv_handler_unregister(bat_priv, BATADV_TVLV_MCAST, 2);

spin_lock_bh(&bat_priv->tt.commit_lock); batadv_mcast_mla_tt_retract(bat_priv, NULL);

With this patch we are finally able to support multicast optimizations in bridged setups, too. So far, if a bridge was added on top of a soft-interface (e.g. bat0) the batman-adv multicast optimizations needed to be disabled to avoid packetloss.

Current Linux bridge implementations and API can now provide us with the so far missing information about interested but "remote" multicast receivers behind bridge ports.

The Linux bridge performs the detection of remote participants interested in multicast packets with its own and mature so called IGMP and MLD snooping code and stores that in its database. With the new API provided by the bridge batman-adv can now simply hook into this database.

We then reliably announce the gathered multicast listeners to other nodes through the batman-adv translation table.

Additionally, the Linux bridge provides us with the information about whether an IGMP/MLD querier exists. If there is none then we need to disable multicast optimizations as we cannot learn about multicast listeners on external, bridged-in host then.

Tested-by: Simon Wunderlich sw@simonwunderlich.de Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue --- compat-include/linux/etherdevice.h | 9 ++ compat-include/linux/if_bridge.h | 68 ++++++++++++++ net/batman-adv/multicast.c | 181 ++++++++++++++++++++++++++++++++---- net/batman-adv/types.h | 13 +++ 4 files changed, 253 insertions(+), 18 deletions(-) create mode 100644 compat-include/linux/if_bridge.h

diff --git a/compat-include/linux/etherdevice.h b/compat-include/linux/etherdevice.h index f120202..721945c 100644 --- a/compat-include/linux/etherdevice.h +++ b/compat-include/linux/etherdevice.h @@ -35,6 +35,15 @@ static inline void batadv_eth_hw_addr_random(struct net_device *dev)

#endif /* < KERNEL_VERSION(3, 4, 0) */

+#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 7, 0) + +static inline void eth_zero_addr(u8 *addr) +{ + memset(addr, 0x00, ETH_ALEN); +} + +#endif /* < KERNEL_VERSION(3, 7, 0) */ + #if LINUX_VERSION_CODE < KERNEL_VERSION(3, 14, 0)

#define ether_addr_equal_unaligned(_a, _b) (memcmp(_a, _b, ETH_ALEN) == 0) diff --git a/compat-include/linux/if_bridge.h b/compat-include/linux/if_bridge.h new file mode 100644 index 0000000..9aae018 --- /dev/null +++ b/compat-include/linux/if_bridge.h @@ -0,0 +1,68 @@ +#ifndef _NET_BATMAN_ADV_COMPAT_LINUX_IF_BRIDGE_H_ +#define _NET_BATMAN_ADV_COMPAT_LINUX_IF_BRIDGE_H_ + +#include <linux/version.h> +#include_next <linux/if_bridge.h> + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 16, 0) + +struct br_ip { + union { + __be32 ip4; +#if IS_ENABLED(CONFIG_IPV6) + struct in6_addr ip6; +#endif + } u; + __be16 proto; + __u16 vid; +}; + +struct br_ip_list { + struct list_head list; + struct br_ip addr; +}; + +#endif /* < KERNEL_VERSION(3, 16, 0) */ + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 16, 0) || \ + LINUX_VERSION_CODE == KERNEL_VERSION(3, 16, 0) && \ + (!IS_ENABLED(CONFIG_BRIDGE) || \ + !IS_ENABLED(CONFIG_BRIDGE_IGMP_SNOOPING)) + +#define br_multicast_list_adjacent(dev, br_ip_list) \ + batadv_br_multicast_list_adjacent(dev, br_ip_list) + +#define br_multicast_has_querier_adjacent(dev, proto) \ + batadv_br_multicast_has_querier_adjacent(dev, proto) + +static inline int +batadv_br_multicast_list_adjacent(struct net_device *dev, + struct list_head *br_ip_list) +{ + return 0; +} + +static inline bool +batadv_br_multicast_has_querier_adjacent(struct net_device *dev, int proto) +{ + return false; +} + +#endif /* < KERNEL_VERSION(3, 16, 0) || + * == KERNEL_VERSION(3, 16, 0) && + * (!IS_ENABLED(CONFIG_BRIDGE) || + * !IS_ENABLED(CONFIG_BRIDGE_IGMP_SNOOPING)) */ + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 17, 0) + +static inline bool br_multicast_has_querier_anywhere(struct net_device *dev, + int proto) +{ + pr_warn_once("Old kernel detected (< 3.17) - multicast optimizations disabled\n"); + + return false; +} + +#endif /* < KERNEL_VERSION(3, 17, 0) */ + +#endif /* _NET_BATMAN_ADV_COMPAT_LINUX_IF_BRIDGE_H_ */ diff --git a/net/batman-adv/multicast.c b/net/batman-adv/multicast.c index b8662f1..a4e73f1 100644 --- a/net/batman-adv/multicast.c +++ b/net/batman-adv/multicast.c @@ -26,6 +26,7 @@ #include <linux/etherdevice.h> #include <linux/fs.h> #include <linux/icmpv6.h> +#include <linux/if_bridge.h> #include <linux/if_ether.h> #include <linux/igmp.h> #include <linux/in6.h> @@ -36,6 +37,7 @@ #include <linux/list.h> #include <linux/lockdep.h> #include <linux/netdevice.h> +#include <linux/printk.h> #include <linux/rculist.h> #include <linux/rcupdate.h> #include <linux/skbuff.h> @@ -45,18 +47,53 @@ #include <linux/string.h> #include <linux/types.h> #include <net/addrconf.h> +#include <net/if_inet6.h> +#include <net/ip.h> #include <net/ipv6.h>

#include "packet.h" #include "translation-table.h"

/** + * batadv_mcast_get_bridge - get the bridge on top of the softif if it exists + * @soft_iface: netdev struct of the mesh interface + * + * If the given soft interface has a bridge on top then the refcount + * of the according net device is increased. + * + * Return: NULL if no such bridge exists. Otherwise the net device of the + * bridge. + */ +static struct net_device *batadv_mcast_get_bridge(struct net_device *soft_iface) +{ + struct net_device *upper = soft_iface; + + rcu_read_lock(); + do { + upper = netdev_master_upper_dev_get_rcu(upper); + } while (upper && !(upper->priv_flags & IFF_EBRIDGE)); + + if (upper) + dev_hold(upper); + rcu_read_unlock(); + + return upper; +} + +/** * batadv_mcast_mla_softif_get - get softif multicast listeners * @dev: the device to collect multicast addresses from * @mcast_list: a list to put found addresses into * - * Collect multicast addresses of the local multicast listeners - * on the given soft interface, dev, in the given mcast_list. + * Collects multicast addresses of multicast listeners residing + * on this kernel on the given soft interface, dev, in + * the given mcast_list. In general, multicast listeners provided by + * your multicast receiving applications run directly on this node. + * + * If there is a bridge interface on top of dev, collects from that one + * instead. Just like with IP addresses and routes, multicast listeners + * will(/should) register to the bridge interface instead of an + * enslaved bat0. * * Return: -ENOMEM on memory allocation error or the number of * items added to the mcast_list otherwise. @@ -64,12 +101,13 @@ static int batadv_mcast_mla_softif_get(struct net_device *dev, struct hlist_head *mcast_list) { + struct net_device *bridge = batadv_mcast_get_bridge(dev); struct netdev_hw_addr *mc_list_entry; struct batadv_hw_addr *new; int ret = 0;

- netif_addr_lock_bh(dev); - netdev_for_each_mc_addr(mc_list_entry, dev) { + netif_addr_lock_bh(bridge ? bridge : dev); + netdev_for_each_mc_addr(mc_list_entry, bridge ? bridge : dev) { new = kmalloc(sizeof(*new), GFP_ATOMIC); if (!new) { ret = -ENOMEM; @@ -80,7 +118,10 @@ static int batadv_mcast_mla_softif_get(struct net_device *dev, hlist_add_head(&new->list, mcast_list); ret++; } - netif_addr_unlock_bh(dev); + netif_addr_unlock_bh(bridge ? bridge : dev); + + if (bridge) + dev_put(bridge);

return ret; } @@ -106,6 +147,83 @@ static bool batadv_mcast_mla_is_duplicate(u8 *mcast_addr, }

/** + * batadv_mcast_mla_br_addr_cpy - copy a bridge multicast address + * @dst: destination to write to - a multicast MAC address + * @src: source to read from - a multicast IP address + * + * Converts a given multicast IPv4/IPv6 address from a bridge + * to its matching multicast MAC address and copies it into the given + * destination buffer. + * + * Caller needs to make sure the destination buffer can hold + * at least ETH_ALEN bytes. + */ +static void batadv_mcast_mla_br_addr_cpy(char *dst, const struct br_ip *src) +{ + if (src->proto == htons(ETH_P_IP)) + ip_eth_mc_map(src->u.ip4, dst); +#if IS_ENABLED(CONFIG_IPV6) + else if (src->proto == htons(ETH_P_IPV6)) + ipv6_eth_mc_map(&src->u.ip6, dst); +#endif + else + eth_zero_addr(dst); +} + +/** + * batadv_mcast_mla_bridge_get - get bridged-in multicast listeners + * @dev: a bridge slave whose bridge to collect multicast addresses from + * @mcast_list: a list to put found addresses into + * + * Collects multicast addresses of multicast listeners residing + * on foreign, non-mesh devices which we gave access to our mesh via + * a bridge on top of the given soft interface, dev, in the given + * mcast_list. + * + * Return: -ENOMEM on memory allocation error or the number of + * items added to the mcast_list otherwise. + */ +static int batadv_mcast_mla_bridge_get(struct net_device *dev, + struct hlist_head *mcast_list) +{ + struct list_head bridge_mcast_list = LIST_HEAD_INIT(bridge_mcast_list); + struct br_ip_list *br_ip_entry, *tmp; + struct batadv_hw_addr *new; + u8 mcast_addr[ETH_ALEN]; + int ret; + + /* we don't need to detect these devices/listeners, the IGMP/MLD + * snooping code of the Linux bridge already does that for us + */ + ret = br_multicast_list_adjacent(dev, &bridge_mcast_list); + if (ret < 0) + goto out; + + list_for_each_entry(br_ip_entry, &bridge_mcast_list, list) { + batadv_mcast_mla_br_addr_cpy(mcast_addr, &br_ip_entry->addr); + if (batadv_mcast_mla_is_duplicate(mcast_addr, mcast_list)) + continue; + + new = kmalloc(sizeof(*new), GFP_ATOMIC); + if (!new) { + ret = -ENOMEM; + break; + } + + ether_addr_copy(new->addr, mcast_addr); + hlist_add_head(&new->list, mcast_list); + } + +out: + list_for_each_entry_safe(br_ip_entry, tmp, &bridge_mcast_list, list) { + list_del(&br_ip_entry->list); + kfree(br_ip_entry); + } + + return ret; +} + +/** * batadv_mcast_mla_list_free - free a list of multicast addresses * @bat_priv: the bat priv with all the soft interface information * @mcast_list: the list to free @@ -222,29 +340,51 @@ static bool batadv_mcast_has_bridge(struct batadv_priv *bat_priv) * Updates the own multicast tvlv with our current multicast related settings, * capabilities and inabilities. * - * Return: true if the tvlv container is registered afterwards. Otherwise - * returns false. + * Return: false if we want all IPv4 && IPv6 multicast traffic and true + * otherwise. */ static bool batadv_mcast_mla_tvlv_update(struct batadv_priv *bat_priv) { struct batadv_tvlv_mcast_data mcast_data; + struct batadv_mcast_querier_state querier4 = {false, false}; + struct batadv_mcast_querier_state querier6 = {false, false}; + struct net_device *dev = bat_priv->soft_iface;

mcast_data.flags = BATADV_NO_FLAGS; memset(mcast_data.reserved, 0, sizeof(mcast_data.reserved));

- /* Avoid attaching MLAs, if there is a bridge on top of our soft - * interface, we don't support that yet (TODO) + bat_priv->mcast.bridged = batadv_mcast_has_bridge(bat_priv); + if (!bat_priv->mcast.bridged) + goto update; + +#if !IS_ENABLED(CONFIG_BRIDGE_IGMP_SNOOPING) + pr_warn_once("No bridge IGMP snooping compiled - multicast optimizations disabled\n"); +#endif + + querier4.exists = br_multicast_has_querier_anywhere(dev, ETH_P_IP); + querier4.shadowing = br_multicast_has_querier_adjacent(dev, ETH_P_IP); + + querier6.exists = br_multicast_has_querier_anywhere(dev, ETH_P_IPV6); + querier6.shadowing = br_multicast_has_querier_adjacent(dev, ETH_P_IPV6); + + mcast_data.flags |= BATADV_MCAST_WANT_ALL_UNSNOOPABLES; + + /* 1) If no querier exists at all, then multicast listeners on + * our local TT clients behind the bridge will keep silent. + * 2) If the selected querier is on one of our local TT clients, + * behind the bridge, then this querier might shadow multicast + * listeners on our local TT clients, behind this bridge. + * + * In both cases, we will signalize other batman nodes that + * we need all multicast traffic of the according protocol. */ - if (batadv_mcast_has_bridge(bat_priv)) { - if (bat_priv->mcast.enabled) { - batadv_tvlv_container_unregister(bat_priv, - BATADV_TVLV_MCAST, 2); - bat_priv->mcast.enabled = false; - } + if (!querier4.exists || querier4.shadowing) + mcast_data.flags |= BATADV_MCAST_WANT_ALL_IPV4;

- return false; - } + if (!querier6.exists || querier6.shadowing) + mcast_data.flags |= BATADV_MCAST_WANT_ALL_IPV6;

+update: if (!bat_priv->mcast.enabled || mcast_data.flags != bat_priv->mcast.flags) { batadv_tvlv_container_register(bat_priv, BATADV_TVLV_MCAST, 2, @@ -253,7 +393,8 @@ static bool batadv_mcast_mla_tvlv_update(struct batadv_priv *bat_priv) bat_priv->mcast.enabled = true; }

- return true; + return !(mcast_data.flags & + (BATADV_MCAST_WANT_ALL_IPV4 + BATADV_MCAST_WANT_ALL_IPV6)); }

/** @@ -276,6 +417,10 @@ void batadv_mcast_mla_update(struct batadv_priv *bat_priv) if (ret < 0) goto out;

+ ret = batadv_mcast_mla_bridge_get(soft_iface, &mcast_list); + if (ret < 0) + goto out; + update: batadv_mcast_mla_tt_retract(bat_priv, &mcast_list); batadv_mcast_mla_tt_add(bat_priv, &mcast_list); diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h index 32c6d0e..83303c2 100644 --- a/net/batman-adv/types.h +++ b/net/batman-adv/types.h @@ -755,6 +755,17 @@ struct batadv_priv_dat {

#ifdef CONFIG_BATMAN_ADV_MCAST /** + * struct batadv_mcast_querier_state - IGMP/MLD querier state when bridged + * @exists: whether a querier exists in the mesh + * @shadowing: if a querier exists, whether it is potentially shadowing + * multicast listeners (i.e. querier is behind our own bridge segment) + */ +struct batadv_mcast_querier_state { + bool exists; + bool shadowing; +}; + +/** * struct batadv_priv_mcast - per mesh interface mcast data * @mla_list: list of multicast addresses we are currently announcing via TT * @want_all_unsnoopables_list: a list of orig_nodes wanting all unsnoopable @@ -763,6 +774,7 @@ struct batadv_priv_dat { * @want_all_ipv6_list: a list of orig_nodes wanting all IPv6 multicast traffic * @flags: the flags we have last sent in our mcast tvlv * @enabled: whether the multicast tvlv is currently enabled + * @bridged: whether the soft interface has a bridge on top * @num_disabled: number of nodes that have no mcast tvlv * @num_want_all_unsnoopables: number of nodes wanting unsnoopable IP traffic * @num_want_all_ipv4: counter for items in want_all_ipv4_list @@ -777,6 +789,7 @@ struct batadv_priv_mcast { struct hlist_head want_all_ipv6_list; u8 flags; bool enabled; + bool bridged; atomic_t num_disabled; atomic_t num_want_all_unsnoopables; atomic_t num_want_all_ipv4;

With this patch changes relevant to a node's own multicast flags are printed to the 'mcast' log level.

Tested-by: Simon Wunderlich sw@simonwunderlich.de Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue --- net/batman-adv/main.h | 4 +- net/batman-adv/multicast.c | 133 ++++++++++++++++++++++++++++++++++++++- net/batman-adv/soft-interface.c | 4 ++ net/batman-adv/types.h | 4 ++ 4 files changed, 142 insertions(+), 3 deletions(-)

diff --git a/net/batman-adv/main.h b/net/batman-adv/main.h index 7692526..832df0f 100644 --- a/net/batman-adv/main.h +++ b/net/batman-adv/main.h @@ -231,6 +231,7 @@ __be32 batadv_skb_crc32(struct sk_buff *skb, u8 *payload_ptr); * @BATADV_DBG_BLA: bridge loop avoidance messages * @BATADV_DBG_DAT: ARP snooping and DAT related messages * @BATADV_DBG_NC: network coding related messages + * @BATADV_DBG_MCAST: multicast related messages * @BATADV_DBG_ALL: the union of all the above log levels */ enum batadv_dbg_level { @@ -240,7 +241,8 @@ enum batadv_dbg_level { BATADV_DBG_BLA = BIT(3), BATADV_DBG_DAT = BIT(4), BATADV_DBG_NC = BIT(5), - BATADV_DBG_ALL = 63, + BATADV_DBG_MCAST = BIT(6), + BATADV_DBG_ALL = 127, };

#ifdef CONFIG_BATMAN_ADV_DEBUG diff --git a/net/batman-adv/multicast.c b/net/batman-adv/multicast.c index a4e73f1..d80178c 100644 --- a/net/batman-adv/multicast.c +++ b/net/batman-adv/multicast.c @@ -33,6 +33,7 @@ #include <linux/in.h> #include <linux/ip.h> #include <linux/ipv6.h> +#include <linux/kernel.h> #include <linux/kref.h> #include <linux/list.h> #include <linux/lockdep.h> @@ -334,6 +335,122 @@ static bool batadv_mcast_has_bridge(struct batadv_priv *bat_priv) }

/** + * batadv_mcast_querier_log - debug output regarding the querier status on link + * @bat_priv: the bat priv with all the soft interface information + * @str_proto: a string for the querier protocol (e.g. "IGMP" or "MLD") + * @old_state: the previous querier state on our link + * @new_state: the new querier state on our link + * + * Outputs debug messages to the logging facility with log level 'mcast' + * regarding changes to the querier status on the link which are relevant + * to our multicast optimizations. + * + * Usually this is about whether a querier appeared or vanished in + * our mesh or whether the querier is in the suboptimal position of being + * behind our local bridge segment: Snooping switches will directly + * forward listener reports to the querier, therefore batman-adv and + * the bridge will potentially not see these listeners - the querier is + * potentially shadowing listeners from us then. + * + * This is only interesting for nodes with a bridge on top of their + * soft interface. + */ +static void +batadv_mcast_querier_log(struct batadv_priv *bat_priv, char *str_proto, + struct batadv_mcast_querier_state *old_state, + struct batadv_mcast_querier_state *new_state) +{ + if (!old_state->exists && new_state->exists) + batadv_info(bat_priv->soft_iface, "%s Querier appeared\n", + str_proto); + else if (old_state->exists && !new_state->exists) + batadv_info(bat_priv->soft_iface, + "%s Querier disappeared - multicast optimizations disabled\n", + str_proto); + else if (!bat_priv->mcast.bridged && !new_state->exists) + batadv_info(bat_priv->soft_iface, + "No %s Querier present - multicast optimizations disabled\n", + str_proto); + + if (new_state->exists) { + if ((!old_state->shadowing && new_state->shadowing) || + (!old_state->exists && new_state->shadowing)) + batadv_dbg(BATADV_DBG_MCAST, bat_priv, + "%s Querier is behind our bridged segment: Might shadow listeners\n", + str_proto); + else if (old_state->shadowing && !new_state->shadowing) + batadv_dbg(BATADV_DBG_MCAST, bat_priv, + "%s Querier is not behind our bridged segment\n", + str_proto); + } +} + +/** + * batadv_mcast_bridge_log - debug output for topology changes in bridged setups + * @bat_priv: the bat priv with all the soft interface information + * @bridged: a flag about whether the soft interface is currently bridged or not + * @querier_ipv4: (maybe) new status of a potential, selected IGMP querier + * @querier_ipv6: (maybe) new status of a potential, selected MLD querier + * + * If no bridges are ever used on this node, then this function does nothing. + * + * Otherwise this function outputs debug information to the 'mcast' log level + * which might be relevant to our multicast optimizations. + * + * More precisely, it outputs information when a bridge interface is added or + * removed from a soft interface. And when a bridge is present, it further + * outputs information about the querier state which is relevant for the + * multicast flags this node is going to set. + */ +static void +batadv_mcast_bridge_log(struct batadv_priv *bat_priv, bool bridged, + struct batadv_mcast_querier_state *querier_ipv4, + struct batadv_mcast_querier_state *querier_ipv6) +{ + if (!bat_priv->mcast.bridged && bridged) + batadv_dbg(BATADV_DBG_MCAST, bat_priv, + "Bridge added: Setting Unsnoopables(U)-flag\n"); + else if (bat_priv->mcast.bridged && !bridged) + batadv_dbg(BATADV_DBG_MCAST, bat_priv, + "Bridge removed: Unsetting Unsnoopables(U)-flag\n"); + + if (bridged) { + batadv_mcast_querier_log(bat_priv, "IGMP", + &bat_priv->mcast.querier_ipv4, + querier_ipv4); + batadv_mcast_querier_log(bat_priv, "MLD", + &bat_priv->mcast.querier_ipv6, + querier_ipv6); + } +} + +/** + * batadv_mcast_flags_logs - output debug information about mcast flag changes + * @bat_priv: the bat priv with all the soft interface information + * @flags: flags indicating the new multicast state + * + * Whenever the multicast flags this nodes announces changes (@mcast_flags vs. + * bat_priv->mcast.flags), this notifies userspace via the 'mcast' log level. + */ +static void batadv_mcast_flags_log(struct batadv_priv *bat_priv, u8 flags) +{ + u8 old_flags = bat_priv->mcast.flags; + char str_old_flags[] = "[...]"; + + sprintf(str_old_flags, "[%c%c%c]", + (old_flags & BATADV_MCAST_WANT_ALL_UNSNOOPABLES) ? 'U' : '.', + (old_flags & BATADV_MCAST_WANT_ALL_IPV4) ? '4' : '.', + (old_flags & BATADV_MCAST_WANT_ALL_IPV6) ? '6' : '.'); + + batadv_dbg(BATADV_DBG_MCAST, bat_priv, + "Changing multicast flags from '%s' to '[%c%c%c]'\n", + bat_priv->mcast.enabled ? str_old_flags : "<undefined>", + (flags & BATADV_MCAST_WANT_ALL_UNSNOOPABLES) ? 'U' : '.', + (flags & BATADV_MCAST_WANT_ALL_IPV4) ? '4' : '.', + (flags & BATADV_MCAST_WANT_ALL_IPV6) ? '6' : '.'); +} + +/** * batadv_mcast_mla_tvlv_update - update multicast tvlv * @bat_priv: the bat priv with all the soft interface information * @@ -349,12 +466,13 @@ static bool batadv_mcast_mla_tvlv_update(struct batadv_priv *bat_priv) struct batadv_mcast_querier_state querier4 = {false, false}; struct batadv_mcast_querier_state querier6 = {false, false}; struct net_device *dev = bat_priv->soft_iface; + bool bridged;

mcast_data.flags = BATADV_NO_FLAGS; memset(mcast_data.reserved, 0, sizeof(mcast_data.reserved));

- bat_priv->mcast.bridged = batadv_mcast_has_bridge(bat_priv); - if (!bat_priv->mcast.bridged) + bridged = batadv_mcast_has_bridge(bat_priv); + if (!bridged) goto update;

#if !IS_ENABLED(CONFIG_BRIDGE_IGMP_SNOOPING) @@ -385,8 +503,19 @@ static bool batadv_mcast_mla_tvlv_update(struct batadv_priv *bat_priv) mcast_data.flags |= BATADV_MCAST_WANT_ALL_IPV6;

update: + batadv_mcast_bridge_log(bat_priv, bridged, &querier4, &querier6); + + bat_priv->mcast.querier_ipv4.exists = querier4.exists; + bat_priv->mcast.querier_ipv4.shadowing = querier4.shadowing; + + bat_priv->mcast.querier_ipv6.exists = querier6.exists; + bat_priv->mcast.querier_ipv6.shadowing = querier6.shadowing; + + bat_priv->mcast.bridged = bridged; + if (!bat_priv->mcast.enabled || mcast_data.flags != bat_priv->mcast.flags) { + batadv_mcast_flags_log(bat_priv, mcast_data.flags); batadv_tvlv_container_register(bat_priv, BATADV_TVLV_MCAST, 2, &mcast_data, sizeof(mcast_data)); bat_priv->mcast.flags = mcast_data.flags; diff --git a/net/batman-adv/soft-interface.c b/net/batman-adv/soft-interface.c index 81665b1..b60999d 100644 --- a/net/batman-adv/soft-interface.c +++ b/net/batman-adv/soft-interface.c @@ -808,6 +808,10 @@ static int batadv_softif_init_late(struct net_device *dev) atomic_set(&bat_priv->distributed_arp_table, 1); #endif #ifdef CONFIG_BATMAN_ADV_MCAST + bat_priv->mcast.querier_ipv4.exists = false; + bat_priv->mcast.querier_ipv4.shadowing = false; + bat_priv->mcast.querier_ipv6.exists = false; + bat_priv->mcast.querier_ipv6.shadowing = false; bat_priv->mcast.flags = BATADV_NO_FLAGS; atomic_set(&bat_priv->multicast_mode, 1); atomic_set(&bat_priv->mcast.num_disabled, 0); diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h index 83303c2..ab863a5 100644 --- a/net/batman-adv/types.h +++ b/net/batman-adv/types.h @@ -772,6 +772,8 @@ struct batadv_mcast_querier_state { * multicast traffic * @want_all_ipv4_list: a list of orig_nodes wanting all IPv4 multicast traffic * @want_all_ipv6_list: a list of orig_nodes wanting all IPv6 multicast traffic + * @querier_ipv4: the current state of an IGMP querier in the mesh + * @querier_ipv6: the current state of an MLD querier in the mesh * @flags: the flags we have last sent in our mcast tvlv * @enabled: whether the multicast tvlv is currently enabled * @bridged: whether the soft interface has a bridge on top @@ -787,6 +789,8 @@ struct batadv_priv_mcast { struct hlist_head want_all_unsnoopables_list; struct hlist_head want_all_ipv4_list; struct hlist_head want_all_ipv6_list; + struct batadv_mcast_querier_state querier_ipv4; + struct batadv_mcast_querier_state querier_ipv6; u8 flags; bool enabled; bool bridged;

This patch adds a debugfs table with originators and their according multicast flags to help users figure out why multicast optimizations might be enabled or disabled for them.

Tested-by: Simon Wunderlich sw@simonwunderlich.de Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue --- net/batman-adv/debugfs.c | 23 ++++++++++ net/batman-adv/multicast.c | 104 ++++++++++++++++++++++++++++++++++++++++++++ net/batman-adv/multicast.h | 3 ++ 3 files changed, 130 insertions(+)

diff --git a/net/batman-adv/debugfs.c b/net/batman-adv/debugfs.c index 9529004..f187a8f 100644 --- a/net/batman-adv/debugfs.c +++ b/net/batman-adv/debugfs.c @@ -48,6 +48,7 @@ #include "distributed-arp-table.h" #include "gateway_client.h" #include "icmp_socket.h" +#include "multicast.h" #include "network-coding.h" #include "originator.h" #include "translation-table.h" @@ -363,6 +364,22 @@ static int batadv_nc_nodes_open(struct inode *inode, struct file *file) } #endif

+#ifdef CONFIG_BATMAN_ADV_MCAST +/** + * batadv_mcast_flags_open - prepare file handler for reads from mcast_flags + * @inode: inode which was opened + * @file: file handle to be initialized + * + * Return: 0 on success or negative error number in case of failure + */ +static int batadv_mcast_flags_open(struct inode *inode, struct file *file) +{ + struct net_device *net_dev = (struct net_device *)inode->i_private; + + return single_open(file, batadv_mcast_flags_seq_print_text, net_dev); +} +#endif + #define BATADV_DEBUGINFO(_name, _mode, _open) \ struct batadv_debuginfo batadv_debuginfo_##_name = { \ .attr = { \ @@ -407,6 +424,9 @@ static BATADV_DEBUGINFO(transtable_local, S_IRUGO, #ifdef CONFIG_BATMAN_ADV_NC static BATADV_DEBUGINFO(nc_nodes, S_IRUGO, batadv_nc_nodes_open); #endif +#ifdef CONFIG_BATMAN_ADV_MCAST +static BATADV_DEBUGINFO(mcast_flags, S_IRUGO, batadv_mcast_flags_open); +#endif

static struct batadv_debuginfo *batadv_mesh_debuginfos[] = { &batadv_debuginfo_neighbors, @@ -424,6 +444,9 @@ static struct batadv_debuginfo *batadv_mesh_debuginfos[] = { #ifdef CONFIG_BATMAN_ADV_NC &batadv_debuginfo_nc_nodes, #endif +#ifdef CONFIG_BATMAN_ADV_MCAST + &batadv_debuginfo_mcast_flags, +#endif NULL, };

diff --git a/net/batman-adv/multicast.c b/net/batman-adv/multicast.c index d80178c..abb12f1 100644 --- a/net/batman-adv/multicast.c +++ b/net/batman-adv/multicast.c @@ -41,6 +41,7 @@ #include <linux/printk.h> #include <linux/rculist.h> #include <linux/rcupdate.h> +#include <linux/seq_file.h> #include <linux/skbuff.h> #include <linux/slab.h> #include <linux/spinlock.h> @@ -52,6 +53,8 @@ #include <net/ip.h> #include <net/ipv6.h>

+#include "hard-interface.h" +#include "hash.h" #include "packet.h" #include "translation-table.h"

@@ -1130,6 +1133,107 @@ void batadv_mcast_init(struct batadv_priv *bat_priv) }

/** + * batadv_mcast_flags_print_header - print own mcast flags to debugfs table + * @bat_priv: the bat priv with all the soft interface information + * @seq: debugfs table seq_file struct + * + * Prints our own multicast flags including a more specific reason why + * they are set, that is prints the bridge and querier state too, to + * the debugfs table specified via @seq. + */ +static void batadv_mcast_flags_print_header(struct batadv_priv *bat_priv, + struct seq_file *seq) +{ + u8 flags = bat_priv->mcast.flags; + char querier4, querier6, shadowing4, shadowing6; + bool bridged = bat_priv->mcast.bridged; + + if (bridged) { + querier4 = bat_priv->mcast.querier_ipv4.exists ? '.' : '4'; + querier6 = bat_priv->mcast.querier_ipv6.exists ? '.' : '6'; + shadowing4 = bat_priv->mcast.querier_ipv4.shadowing ? '4' : '.'; + shadowing6 = bat_priv->mcast.querier_ipv6.shadowing ? '6' : '.'; + } else { + querier4 = '?'; + querier6 = '?'; + shadowing4 = '?'; + shadowing6 = '?'; + } + + seq_printf(seq, "Multicast flags (own flags: [%c%c%c])\n", + (flags & BATADV_MCAST_WANT_ALL_UNSNOOPABLES) ? 'U' : '.', + (flags & BATADV_MCAST_WANT_ALL_IPV4) ? '4' : '.', + (flags & BATADV_MCAST_WANT_ALL_IPV6) ? '6' : '.'); + seq_printf(seq, "* Bridged [U]\t\t\t\t%c\n", bridged ? 'U' : '.'); + seq_printf(seq, "* No IGMP/MLD Querier [4/6]:\t\t%c/%c\n", + querier4, querier6); + seq_printf(seq, "* Shadowing IGMP/MLD Querier [4/6]:\t%c/%c\n", + shadowing4, shadowing6); + seq_puts(seq, "-------------------------------------------\n"); + seq_printf(seq, " %-10s %s\n", "Originator", "Flags"); +} + +/** + * batadv_mcast_flags_seq_print_text - print the mcast flags of other nodes + * @seq: seq file to print on + * @offset: not used + * + * This prints a table of (primary) originators and their according + * multicast flags, including (in the header) our own. + * + * Return: always 0 + */ +int batadv_mcast_flags_seq_print_text(struct seq_file *seq, void *offset) +{ + struct net_device *net_dev = (struct net_device *)seq->private; + struct batadv_priv *bat_priv = netdev_priv(net_dev); + struct batadv_hard_iface *primary_if; + struct batadv_hashtable *hash = bat_priv->orig_hash; + struct batadv_orig_node *orig_node; + struct hlist_head *head; + u8 flags; + u32 i; + + primary_if = batadv_seq_print_text_primary_if_get(seq); + if (!primary_if) + return 0; + + batadv_mcast_flags_print_header(bat_priv, seq); + + for (i = 0; i < hash->size; i++) { + head = &hash->table[i]; + + rcu_read_lock(); + hlist_for_each_entry_rcu(orig_node, head, hash_entry) { + if (!(orig_node->capa_initialized & + BATADV_ORIG_CAPA_HAS_MCAST)) + continue; + + if (!(orig_node->capabilities & + BATADV_ORIG_CAPA_HAS_MCAST)) { + seq_printf(seq, "%pM -\n", orig_node->orig); + continue; + } + + flags = orig_node->mcast_flags; + + seq_printf(seq, "%pM [%c%c%c]\n", orig_node->orig, + (flags & BATADV_MCAST_WANT_ALL_UNSNOOPABLES) + ? 'U' : '.', + (flags & BATADV_MCAST_WANT_ALL_IPV4) + ? '4' : '.', + (flags & BATADV_MCAST_WANT_ALL_IPV6) + ? '6' : '.'); + } + rcu_read_unlock(); + } + + batadv_hardif_put(primary_if); + + return 0; +} + +/** * batadv_mcast_free - free the multicast optimizations structures * @bat_priv: the bat priv with all the soft interface information */ diff --git a/net/batman-adv/multicast.h b/net/batman-adv/multicast.h index 80bceec..1fb00ba 100644 --- a/net/batman-adv/multicast.h +++ b/net/batman-adv/multicast.h @@ -20,6 +20,7 @@

#include "main.h"

+struct seq_file; struct sk_buff;

/** @@ -46,6 +47,8 @@ batadv_mcast_forw_mode(struct batadv_priv *bat_priv, struct sk_buff *skb,

void batadv_mcast_init(struct batadv_priv *bat_priv);

+int batadv_mcast_flags_seq_print_text(struct seq_file *seq, void *offset); + void batadv_mcast_free(struct batadv_priv *bat_priv);

void batadv_mcast_purge_orig(struct batadv_orig_node *orig_node);