[RFC PATCH net-next 0/7] IPv6 MLD updates

[RFC PATCH net-next 0/7] IPv6 MLD updates

[Daniel Borkmann]

If there are no further objections, I'll send out the non-RFC version on
Monday after some more testing. That's just a heads up. This set includes
my improved fix from yesterday as well besides other things.

Thanks!

Daniel Borkmann (7):
  net: ipv6: mld: fix v1/v2 switchback timeout to rfc3810, 9.12.
  net: ipv6: mld: clean up MLD_V1_SEEN macro
  net: ipv6: mld: get rid of MLDV2_MRC and simplify calculation
  net: ipv6: mld: implement RFC3810 MLDv2 mode only
  net: ipv6: mld: restrict min/max of sysctl force_mld_version
  net: ipv6: mld: similarly to MLDv2 have min max_delay of 1
  net: ipv6: mld: document force_mld_version in ip-sysctl.txt

 Documentation/networking/ip-sysctl.txt |   5 ++
 include/net/if_inet6.h                 |   9 +-
 include/net/mld.h                      |  51 +++++++++--
 net/bridge/br_multicast.c              |   3 +-
 net/ipv6/addrconf.c                    |   7 +-
 net/ipv6/mcast.c                       | 159 +++++++++++++++++++++++++++------
 6 files changed, 195 insertions(+), 39 deletions(-)

-- 
1.7.11.7

[RFC PATCH net-next 1/7] net: ipv6: mld: fix v1/v2 switchback timeout to rfc3810, 9.12.

[Daniel Borkmann]

i) RFC3810, 9.2. Query Interval [QI] says:

   The Query Interval variable denotes the interval between General
   Queries sent by the Querier. Default value: 125 seconds. [...]

ii) RFC3810, 9.3. Query Response Interval [QRI] says:

  The Maximum Response Delay used to calculate the Maximum Response
  Code inserted into the periodic General Queries. Default value:
  10000 (10 seconds) [...] The number of seconds represented by the
  [Query Response Interval] must be less than the [Query Interval].

iii) RFC3810, 9.12. Older Version Querier Present Timeout [OVQPT] says:

  The Older Version Querier Present Timeout is the time-out for
  transitioning a host back to MLDv2 Host Compatibility Mode. When an
  MLDv1 query is received, MLDv2 hosts set their Older Version Querier
  Present Timer to [Older Version Querier Present Timeout].

  This value MUST be ([Robustness Variable] times (the [Query Interval]
  in the last Query received)) plus ([Query Response Interval]).

Hence, on *default* the timeout results in:

  [RV] = 2, [QI] = 125sec, [QRI] = 10sec
  [OVQPT] = [RV] * [QI] + [QRI] = 260sec

Having that said, we currently calculate [OVQPT] (here given as 'switchback'
variable) as ...

  switchback = (idev->mc_qrv + 1) * max_delay

On the other hand, RFC3810, 9.12. says "the [Query Interval] in the last Query
received". In section "9.14. Configuring timers", it is said:

  This section is meant to provide advice to network administrators on
  how to tune these settings to their network. Ambitious router
  implementations might tune these settings dynamically based upon
  changing characteristics of the network. [...]

iv) RFC38010, 9.14.2. Query Interval:

  The overall level of periodic MLD traffic is inversely proportional
  to the Query Interval. A longer Query Interval results in a lower
  overall level of MLD traffic. The value of the Query Interval MUST
  be equal to or greater than the Maximum Response Delay used to
  calculate the Maximum Response Code inserted in General Query
  messages.

I assume that was why switchback is calculated as is (3 * max_delay), although
this setting seems to be meant for routers only to configure their [QI]
interval for non-default intervals. So usage here like this is wrong.

Concluding, the current behaviour in IPv6's multicast code is not conform
to the RFC as switch back is calculated wrongly. That is, it has a too small
value, so MLDv2 hosts switch back again to MLDv2 way too early.

Hence, introduce helper functions and fix this up properly as it should be.

Introduced in 06da92283 ("[IPV6]: Add MLDv2 support."). Credits to Hannes
Frederic Sowa who also had a hand in this as well.

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: David Stevens <dlstevens@us.ibm.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
---
 include/net/if_inet6.h |   9 ++++-
 include/net/mld.h      |  27 ++++++++++++-
 net/ipv6/mcast.c       | 108 ++++++++++++++++++++++++++++++++++++++++++++++---
 3 files changed, 135 insertions(+), 9 deletions(-)

diff --git a/include/net/if_inet6.h b/include/net/if_inet6.h
index 736b5fb..02ef772 100644
--- a/include/net/if_inet6.h
+++ b/include/net/if_inet6.h
@@ -171,12 +171,17 @@ struct inet6_dev {
 	struct ifmcaddr6	*mc_list;
 	struct ifmcaddr6	*mc_tomb;
 	spinlock_t		mc_lock;
-	unsigned char		mc_qrv;
+
+	unsigned char		mc_qrv;		/* Query Robustness Variable */
 	unsigned char		mc_gq_running;
 	unsigned char		mc_ifc_count;
 	unsigned char		mc_dad_count;
-	unsigned long		mc_v1_seen;
+
+	unsigned long		mc_v1_seen;	/* Max time we stay in MLDv1 mode */
+	unsigned long		mc_qi;		/* Query Interval */
+	unsigned long		mc_qri;		/* Query Response Interval */
 	unsigned long		mc_maxdelay;
+
 	struct timer_list	mc_gq_timer;	/* general query timer */
 	struct timer_list	mc_ifc_timer;	/* interface change timer */
 	struct timer_list	mc_dad_timer;	/* dad complete mc timer */
diff --git a/include/net/mld.h b/include/net/mld.h
index 467143c..2b5421f 100644
--- a/include/net/mld.h
+++ b/include/net/mld.h
@@ -63,7 +63,7 @@ struct mld2_query {
 #define mld2q_mrc		mld2q_hdr.icmp6_maxdelay
 #define mld2q_resv1		mld2q_hdr.icmp6_dataun.un_data16[1]
 
-/* Max Response Code */
+/* Max Response Code, TODO: transform this to use the below */
 #define MLDV2_MASK(value, nb) ((nb)>=32 ? (value) : ((1<<(nb))-1) & (value))
 #define MLDV2_EXP(thresh, nbmant, nbexp, value) \
 	((value) < (thresh) ? (value) : \
@@ -72,4 +72,29 @@ struct mld2_query {
 
 #define MLDV2_MRC(value) MLDV2_EXP(0x8000, 12, 3, value)
 
+/* RFC3810, 5.1.3. Maximum Response Code:
+ *
+ * If Maximum Response Code >= 32768, Maximum Response Code represents a
+ * floating-point value as follows:
+ *
+ *  0 1 2 3 4 5 6 7 8 9 A B C D E F
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * |1| exp |          mant         |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ */
+#define MLDV2_MRC_EXP(value)	(((value) >> 12) & 0x0007)
+#define MLDV2_MRC_MAN(value)	((value) & 0x0fff)
+
+/* RFC3810, 5.1.9. QQIC (Querier's Query Interval Code):
+ *
+ * If QQIC >= 128, QQIC represents a floating-point value as follows:
+ *
+ *  0 1 2 3 4 5 6 7
+ * +-+-+-+-+-+-+-+-+
+ * |1| exp | mant  |
+ * +-+-+-+-+-+-+-+-+
+ */
+#define MLDV2_QQIC_EXP(value)	(((value) >> 4) & 0x07)
+#define MLDV2_QQIC_MAN(value)	((value) & 0x0f)
+
 #endif
diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c
index 98ead2b..8d232af 100644
--- a/net/ipv6/mcast.c
+++ b/net/ipv6/mcast.c
@@ -108,6 +108,10 @@ static int ip6_mc_leave_src(struct sock *sk, struct ipv6_mc_socklist *iml,
 			    struct inet6_dev *idev);
 
 #define MLD_QRV_DEFAULT		2
+/* RFC3810, 9.2. Query Interval */
+#define MLD_QI_DEFAULT		(125 * HZ)
+/* RFC3810, 9.3. Query Response Interval */
+#define MLD_QRI_DEFAULT		(10 * HZ)
 
 /* RFC3810, 8.1 Query Version Distinctions */
 #define MLD_V1_QUERY_LEN	24
@@ -1112,6 +1116,93 @@ static bool mld_marksources(struct ifmcaddr6 *pmc, int nsrcs,
 	return true;
 }
 
+static void mld_set_v1_mode(struct inet6_dev *idev)
+{
+	/* RFC3810, relevant sections:
+	 *  - 9.1. Robustness Variable
+	 *  - 9.2. Query Interval
+	 *  - 9.3. Query Response Interval
+	 *  - 9.12. Older Version Querier Present Timeout
+	 */
+	unsigned long switchback;
+
+	switchback = (idev->mc_qrv * idev->mc_qi) + idev->mc_qri;
+
+	idev->mc_v1_seen = jiffies + switchback;
+}
+
+static void mld_update_qrv(struct inet6_dev *idev,
+			   const struct mld2_query *mlh2)
+{
+	/* RFC3810, relevant sections:
+	 *  - 5.1.8. QRV (Querier's Robustness Variable)
+	 *  - 9.1. Robustness Variable
+	 */
+
+	/* The value of the Robustness Variable MUST NOT be zero,
+	 * and SHOULD NOT be one. Catch this here if we ever run
+	 * into such a case in future.
+	 */
+	BUG_ON(idev->mc_qrv == 0);
+
+	if (mlh2->mld2q_qrv > 0)
+		idev->mc_qrv = mlh2->mld2q_qrv;
+
+	if (unlikely(idev->mc_qrv < 2)) {
+		net_warn_ratelimited("IPv6: MLD: clamping QRV from %u to %u!\n",
+				     idev->mc_qrv, MLD_QRV_DEFAULT);
+		idev->mc_qrv = MLD_QRV_DEFAULT;
+	}
+}
+
+static void mld_update_qi(struct inet6_dev *idev,
+			  const struct mld2_query *mlh2)
+{
+	/* RFC3810, relevant sections:
+	 *  - 5.1.9. QQIC (Querier's Query Interval Code)
+	 *  - 9.2. Query Interval
+	 *  - 9.12. Older Version Querier Present Timeout
+	 *    (the [Query Interval] in the last Query received)
+	 */
+	unsigned long mc_qqi;
+
+	if (mlh2->mld2q_qqic < 128) {
+		mc_qqi = mlh2->mld2q_qqic;
+	} else {
+		unsigned long mc_man, mc_exp;
+
+		mc_exp = MLDV2_QQIC_EXP(mlh2->mld2q_qqic);
+		mc_man = MLDV2_QQIC_MAN(mlh2->mld2q_qqic);
+
+		mc_qqi = (mc_man | 0x10) << (mc_exp + 3);
+	}
+
+	idev->mc_qi = mc_qqi * HZ;
+}
+
+static void mld_update_qri(struct inet6_dev *idev,
+			   const struct mld2_query *mlh2)
+{
+	/* RFC3810, relevant sections:
+	 *  - 5.1.3. Maximum Response Code
+	 *  - 9.3. Query Response Interval
+	 */
+	unsigned long mc_qri, mc_mrc = ntohs(mlh2->mld2q_mrc);
+
+	if (mc_mrc < 32768) {
+		mc_qri = mc_mrc;
+	} else {
+		unsigned long mc_man, mc_exp;
+
+		mc_exp = MLDV2_MRC_EXP(mc_mrc);
+		mc_man = MLDV2_MRC_MAN(mc_mrc);
+
+		mc_qri = (mc_man | 0x1000) << (mc_exp + 3);
+	}
+
+	idev->mc_qri = msecs_to_jiffies(mc_qri);
+}
+
 /* called with rcu_read_lock() */
 int igmp6_event_query(struct sk_buff *skb)
 {
@@ -1150,12 +1241,10 @@ int igmp6_event_query(struct sk_buff *skb)
 		return -EINVAL;
 
 	if (len == MLD_V1_QUERY_LEN) {
-		int switchback;
 		/* MLDv1 router present */
-
 		max_delay = msecs_to_jiffies(ntohs(mld->mld_maxdelay));
-		switchback = (idev->mc_qrv + 1) * max_delay;
-		idev->mc_v1_seen = jiffies + switchback;
+
+		mld_set_v1_mode(idev);
 
 		/* cancel the interface change timer */
 		idev->mc_ifc_count = 0;
@@ -1166,6 +1255,7 @@ int igmp6_event_query(struct sk_buff *skb)
 	} else if (len >= MLD_V2_QUERY_LEN_MIN) {
 		int srcs_offset = sizeof(struct mld2_query) -
 				  sizeof(struct icmp6hdr);
+
 		if (!pskb_may_pull(skb, srcs_offset))
 			return -EINVAL;
 
@@ -1175,8 +1265,10 @@ int igmp6_event_query(struct sk_buff *skb)
 
 		idev->mc_maxdelay = max_delay;
 
-		if (mlh2->mld2q_qrv)
-			idev->mc_qrv = mlh2->mld2q_qrv;
+		mld_update_qrv(idev, mlh2);
+		mld_update_qi(idev, mlh2);
+		mld_update_qri(idev, mlh2);
+
 		if (group_type == IPV6_ADDR_ANY) { /* general query */
 			if (mlh2->mld2q_nsrcs)
 				return -EINVAL; /* no sources allowed */
@@ -2337,7 +2429,11 @@ void ipv6_mc_init_dev(struct inet6_dev *idev)
 			(unsigned long)idev);
 	setup_timer(&idev->mc_dad_timer, mld_dad_timer_expire,
 		    (unsigned long)idev);
+
 	idev->mc_qrv = MLD_QRV_DEFAULT;
+	idev->mc_qi = MLD_QI_DEFAULT;
+	idev->mc_qri = MLD_QRI_DEFAULT;
+
 	idev->mc_maxdelay = unsolicited_report_interval(idev);
 	idev->mc_v1_seen = 0;
 	write_unlock_bh(&idev->lock);
-- 
1.7.11.7

[RFC PATCH net-next 2/7] net: ipv6: mld: clean up MLD_V1_SEEN macro

[Daniel Borkmann]

Replace the macro with a function to make it more readable. GCC will
eventually decide whether to inline this or not (also, that's not
fast-path anyway).

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
---
 net/ipv6/mcast.c | 32 ++++++++++++++++++++------------
 1 file changed, 20 insertions(+), 12 deletions(-)

diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c
index 8d232af..6e290f2 100644
--- a/net/ipv6/mcast.c
+++ b/net/ipv6/mcast.c
@@ -95,6 +95,7 @@ static void mld_ifc_event(struct inet6_dev *idev);
 static void mld_add_delrec(struct inet6_dev *idev, struct ifmcaddr6 *pmc);
 static void mld_del_delrec(struct inet6_dev *idev, const struct in6_addr *addr);
 static void mld_clear_delrec(struct inet6_dev *idev);
+static bool mld_in_v1_mode(const struct inet6_dev *idev);
 static int sf_setstate(struct ifmcaddr6 *pmc);
 static void sf_markstate(struct ifmcaddr6 *pmc);
 static void ip6_mc_clear_src(struct ifmcaddr6 *pmc);
@@ -117,11 +118,6 @@ static int ip6_mc_leave_src(struct sock *sk, struct ipv6_mc_socklist *iml,
 #define MLD_V1_QUERY_LEN	24
 #define MLD_V2_QUERY_LEN_MIN	28
 
-#define MLD_V1_SEEN(idev) (dev_net((idev)->dev)->ipv6.devconf_all->force_mld_version == 1 || \
-		(idev)->cnf.force_mld_version == 1 || \
-		((idev)->mc_v1_seen && \
-		time_before(jiffies, (idev)->mc_v1_seen)))
-
 #define IPV6_MLD_MAX_MSF	64
 
 int sysctl_mld_max_msf __read_mostly = IPV6_MLD_MAX_MSF;
@@ -139,7 +135,7 @@ static int unsolicited_report_interval(struct inet6_dev *idev)
 {
 	int iv;
 
-	if (MLD_V1_SEEN(idev))
+	if (mld_in_v1_mode(idev))
 		iv = idev->cnf.mldv1_unsolicited_report_interval;
 	else
 		iv = idev->cnf.mldv2_unsolicited_report_interval;
@@ -695,7 +691,7 @@ static void igmp6_group_added(struct ifmcaddr6 *mc)
 	if (!(dev->flags & IFF_UP) || (mc->mca_flags & MAF_NOREPORT))
 		return;
 
-	if (MLD_V1_SEEN(mc->idev)) {
+	if (mld_in_v1_mode(mc->idev)) {
 		igmp6_join_group(mc);
 		return;
 	}
@@ -1116,6 +1112,18 @@ static bool mld_marksources(struct ifmcaddr6 *pmc, int nsrcs,
 	return true;
 }
 
+static bool mld_in_v1_mode(const struct inet6_dev *idev)
+{
+	if (dev_net(idev->dev)->ipv6.devconf_all->force_mld_version == 1)
+		return true;
+	if (idev->cnf.force_mld_version == 1)
+		return true;
+	if (idev->mc_v1_seen && time_before(jiffies, idev->mc_v1_seen))
+		return true;
+
+	return false;
+}
+
 static void mld_set_v1_mode(struct inet6_dev *idev)
 {
 	/* RFC3810, relevant sections:
@@ -1934,7 +1942,7 @@ err_out:
 
 static void mld_resend_report(struct inet6_dev *idev)
 {
-	if (MLD_V1_SEEN(idev)) {
+	if (mld_in_v1_mode(idev)) {
 		struct ifmcaddr6 *mcaddr;
 		read_lock_bh(&idev->lock);
 		for (mcaddr = idev->mc_list; mcaddr; mcaddr = mcaddr->next) {
@@ -1998,7 +2006,7 @@ static int ip6_mc_del1_src(struct ifmcaddr6 *pmc, int sfmode,
 		else
 			pmc->mca_sources = psf->sf_next;
 		if (psf->sf_oldin && !(pmc->mca_flags & MAF_NOREPORT) &&
-		    !MLD_V1_SEEN(idev)) {
+		    !mld_in_v1_mode(idev)) {
 			psf->sf_crcount = idev->mc_qrv;
 			psf->sf_next = pmc->mca_tomb;
 			pmc->mca_tomb = psf;
@@ -2298,7 +2306,7 @@ static int ip6_mc_leave_src(struct sock *sk, struct ipv6_mc_socklist *iml,
 
 static void igmp6_leave_group(struct ifmcaddr6 *ma)
 {
-	if (MLD_V1_SEEN(ma->idev)) {
+	if (mld_in_v1_mode(ma->idev)) {
 		if (ma->mca_flags & MAF_LAST_REPORTER)
 			igmp6_send(&ma->mca_addr, ma->idev->dev,
 				ICMPV6_MGM_REDUCTION);
@@ -2332,7 +2340,7 @@ static void mld_ifc_timer_expire(unsigned long data)
 
 static void mld_ifc_event(struct inet6_dev *idev)
 {
-	if (MLD_V1_SEEN(idev))
+	if (mld_in_v1_mode(idev))
 		return;
 	idev->mc_ifc_count = idev->mc_qrv;
 	mld_ifc_start_timer(idev, 1);
@@ -2343,7 +2351,7 @@ static void igmp6_timer_handler(unsigned long data)
 {
 	struct ifmcaddr6 *ma = (struct ifmcaddr6 *) data;
 
-	if (MLD_V1_SEEN(ma->idev))
+	if (mld_in_v1_mode(ma->idev))
 		igmp6_send(&ma->mca_addr, ma->idev->dev, ICMPV6_MGM_REPORT);
 	else
 		mld_send_report(ma->idev, ma);
-- 
1.7.11.7

[RFC PATCH net-next 3/7] net: ipv6: mld: get rid of MLDV2_MRC and simplify calculation

[Daniel Borkmann]

Get rid of MLDV2_MRC and use our new macros for mantisse and
exponent to calculate Maximum Response Delay out of the Maximum
Response Code.

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
---
 include/net/mld.h         | 28 +++++++++++++++++++---------
 net/bridge/br_multicast.c |  3 ++-
 net/ipv6/mcast.c          | 18 ++----------------
 3 files changed, 23 insertions(+), 26 deletions(-)

diff --git a/include/net/mld.h b/include/net/mld.h
index 2b5421f..faa1d16 100644
--- a/include/net/mld.h
+++ b/include/net/mld.h
@@ -63,15 +63,6 @@ struct mld2_query {
 #define mld2q_mrc		mld2q_hdr.icmp6_maxdelay
 #define mld2q_resv1		mld2q_hdr.icmp6_dataun.un_data16[1]
 
-/* Max Response Code, TODO: transform this to use the below */
-#define MLDV2_MASK(value, nb) ((nb)>=32 ? (value) : ((1<<(nb))-1) & (value))
-#define MLDV2_EXP(thresh, nbmant, nbexp, value) \
-	((value) < (thresh) ? (value) : \
-	((MLDV2_MASK(value, nbmant) | (1<<(nbmant))) << \
-	(MLDV2_MASK((value) >> (nbmant), nbexp) + (nbexp))))
-
-#define MLDV2_MRC(value) MLDV2_EXP(0x8000, 12, 3, value)
-
 /* RFC3810, 5.1.3. Maximum Response Code:
  *
  * If Maximum Response Code >= 32768, Maximum Response Code represents a
@@ -97,4 +88,23 @@ struct mld2_query {
 #define MLDV2_QQIC_EXP(value)	(((value) >> 4) & 0x07)
 #define MLDV2_QQIC_MAN(value)	((value) & 0x0f)
 
+static inline unsigned long mldv2_mrc(const struct mld2_query *mlh2)
+{
+	/* RFC3810, 5.1.3. Maximum Response Code */
+	unsigned long ret, mc_mrc = ntohs(mlh2->mld2q_mrc);
+
+	if (mc_mrc < 32768) {
+		ret = mc_mrc;
+	} else {
+		unsigned long mc_man, mc_exp;
+
+		mc_exp = MLDV2_MRC_EXP(mc_mrc);
+		mc_man = MLDV2_MRC_MAN(mc_mrc);
+
+		ret = (mc_man | 0x1000) << (mc_exp + 3);
+	}
+
+	return ret;
+}
+
 #endif
diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 08e576a..4accd0d 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -1203,7 +1203,8 @@ static int br_ip6_multicast_query(struct net_bridge *br,
 		mld2q = (struct mld2_query *)icmp6_hdr(skb);
 		if (!mld2q->mld2q_nsrcs)
 			group = &mld2q->mld2q_mca;
-		max_delay = mld2q->mld2q_mrc ? MLDV2_MRC(ntohs(mld2q->mld2q_mrc)) : 1;
+
+		max_delay = max(msecs_to_jiffies(mldv2_mrc(mld2q)), 1UL);
 	}
 
 	br_multicast_query_received(br, port, !ipv6_addr_any(&ip6h->saddr),
diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c
index 6e290f2..b6d3652 100644
--- a/net/ipv6/mcast.c
+++ b/net/ipv6/mcast.c
@@ -1195,20 +1195,7 @@ static void mld_update_qri(struct inet6_dev *idev,
 	 *  - 5.1.3. Maximum Response Code
 	 *  - 9.3. Query Response Interval
 	 */
-	unsigned long mc_qri, mc_mrc = ntohs(mlh2->mld2q_mrc);
-
-	if (mc_mrc < 32768) {
-		mc_qri = mc_mrc;
-	} else {
-		unsigned long mc_man, mc_exp;
-
-		mc_exp = MLDV2_MRC_EXP(mc_mrc);
-		mc_man = MLDV2_MRC_MAN(mc_mrc);
-
-		mc_qri = (mc_man | 0x1000) << (mc_exp + 3);
-	}
-
-	idev->mc_qri = msecs_to_jiffies(mc_qri);
+	idev->mc_qri = msecs_to_jiffies(mldv2_mrc(mlh2));
 }
 
 /* called with rcu_read_lock() */
@@ -1269,8 +1256,7 @@ int igmp6_event_query(struct sk_buff *skb)
 
 		mlh2 = (struct mld2_query *)skb_transport_header(skb);
 
-		max_delay = max(msecs_to_jiffies(MLDV2_MRC(ntohs(mlh2->mld2q_mrc))), 1UL);
-
+		max_delay = max(msecs_to_jiffies(mldv2_mrc(mlh2)), 1UL);
 		idev->mc_maxdelay = max_delay;
 
 		mld_update_qrv(idev, mlh2);
-- 
1.7.11.7

[RFC PATCH net-next 4/7] net: ipv6: mld: implement RFC3810 MLDv2 mode only

[Daniel Borkmann]

RFC3810, 10. Security Considerations says under subsection 10.1.
Query Message:

  A forged Version 1 Query message will put MLDv2 listeners on that
  link in MLDv1 Host Compatibility Mode. This scenario can be avoided
  by providing MLDv2 hosts with a configuration option to ignore
  Version 1 messages completely.

Hence, implement a MLDv2-only mode that will ignore MLDv1 traffic:

  echo 2 > /proc/sys/net/ipv6/conf/ethX/force_mld_version

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
---
 net/ipv6/mcast.c | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c
index b6d3652..d6f2809 100644
--- a/net/ipv6/mcast.c
+++ b/net/ipv6/mcast.c
@@ -1112,9 +1112,21 @@ static bool mld_marksources(struct ifmcaddr6 *pmc, int nsrcs,
 	return true;
 }
 
+static bool mld_in_v2_mode_only(const struct inet6_dev *idev)
+{
+	return dev_net(idev->dev)->ipv6.devconf_all->force_mld_version == 2;
+}
+
+static bool mld_in_v1_mode_only(const struct inet6_dev *idev)
+{
+	return dev_net(idev->dev)->ipv6.devconf_all->force_mld_version == 1;
+}
+
 static bool mld_in_v1_mode(const struct inet6_dev *idev)
 {
-	if (dev_net(idev->dev)->ipv6.devconf_all->force_mld_version == 1)
+	if (mld_in_v2_mode_only(idev))
+		return false;
+	if (mld_in_v1_mode_only(idev))
 		return true;
 	if (idev->cnf.force_mld_version == 1)
 		return true;
@@ -1223,7 +1235,6 @@ int igmp6_event_query(struct sk_buff *skb)
 		return -EINVAL;
 
 	idev = __in6_dev_get(skb->dev);
-
 	if (idev == NULL)
 		return 0;
 
@@ -1236,6 +1247,10 @@ int igmp6_event_query(struct sk_buff *skb)
 		return -EINVAL;
 
 	if (len == MLD_V1_QUERY_LEN) {
+		/* Ignore v1 queries */
+		if (mld_in_v2_mode_only(idev))
+			return 0;
+
 		/* MLDv1 router present */
 		max_delay = msecs_to_jiffies(ntohs(mld->mld_maxdelay));
 
-- 
1.7.11.7

[RFC PATCH net-next 5/7] net: ipv6: mld: restrict min/max of sysctl force_mld_version

[Daniel Borkmann]

Restrict minimum and maximum values that are allowed to be set by the
administrator for force_mld_version into sysctl:

  0 - No enforcement of any MLD version
  1 - Enforce to use MLD version 1
  2 - Enforce to use MLD version 2

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
---
 net/ipv6/addrconf.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index a7183fc..dde76d4 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -106,6 +106,9 @@
 
 #define	INFINITY_LIFE_TIME	0xFFFFFFFF
 
+static int zero = 0;
+static int two = 2;
+
 static inline u32 cstamp_delta(unsigned long cstamp)
 {
 	return (cstamp - INITIAL_JIFFIES) * 100UL / HZ;
@@ -4842,7 +4845,9 @@ static struct addrconf_sysctl_table
 			.data		= &ipv6_devconf.force_mld_version,
 			.maxlen		= sizeof(int),
 			.mode		= 0644,
-			.proc_handler	= proc_dointvec,
+			.proc_handler	= proc_dointvec_minmax,
+			.extra1		= &zero,
+			.extra2		= &two,
 		},
 		{
 			.procname	= "mldv1_unsolicited_report_interval",
-- 
1.7.11.7

[RFC PATCH net-next 6/7] net: ipv6: mld: similarly to MLDv2 have min max_delay of 1

[Daniel Borkmann]

Similarly as we do in MLDv2 queries, set a forged MLDv1 query with
0 ms mld_maxdelay to minimum timer shot time of 1 jiffies. This is
eventually done in igmp6_group_queried() anyway, so we can simplify
a check there.

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
---
 net/ipv6/mcast.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c
index d6f2809..be22317 100644
--- a/net/ipv6/mcast.c
+++ b/net/ipv6/mcast.c
@@ -1040,12 +1040,9 @@ static void igmp6_group_queried(struct ifmcaddr6 *ma, unsigned long resptime)
 		delay = ma->mca_timer.expires - jiffies;
 	}
 
-	if (delay >= resptime) {
-		if (resptime)
-			delay = net_random() % resptime;
-		else
-			delay = 1;
-	}
+	if (delay >= resptime)
+		delay = net_random() % resptime;
+
 	ma->mca_timer.expires = jiffies + delay;
 	if (!mod_timer(&ma->mca_timer, jiffies + delay))
 		atomic_inc(&ma->mca_refcnt);
@@ -1247,12 +1244,15 @@ int igmp6_event_query(struct sk_buff *skb)
 		return -EINVAL;
 
 	if (len == MLD_V1_QUERY_LEN) {
+		unsigned long mldv1_md;
+
 		/* Ignore v1 queries */
 		if (mld_in_v2_mode_only(idev))
 			return 0;
 
 		/* MLDv1 router present */
-		max_delay = msecs_to_jiffies(ntohs(mld->mld_maxdelay));
+		mldv1_md = ntohs(mld->mld_maxdelay);
+		max_delay = max(msecs_to_jiffies(mldv1_md), 1UL);
 
 		mld_set_v1_mode(idev);
 
-- 
1.7.11.7

[RFC PATCH net-next 7/7] net: ipv6: mld: document force_mld_version in ip-sysctl.txt

[Daniel Borkmann]

Document force_mld_version parameter in ip-sysctl.txt.

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
---
 Documentation/networking/ip-sysctl.txt | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt
index 1cb3aeb..a46d785 100644
--- a/Documentation/networking/ip-sysctl.txt
+++ b/Documentation/networking/ip-sysctl.txt
@@ -1358,6 +1358,11 @@ mldv2_unsolicited_report_interval - INTEGER
 	MLDv2 report retransmit will take place.
 	Default: 1000 (1 second)
 
+force_mld_version - INTEGER
+	0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
+	1 - Enforce to use MLD version 1
+	2 - Enforce to use MLD version 2
+
 suppress_frag_ndisc - INTEGER
 	Control RFC 6980 (Security Implications of IPv6 Fragmentation
 	with IPv6 Neighbor Discovery) behavior:
-- 
1.7.11.7

Re: [RFC PATCH net-next 0/7] IPv6 MLD updates

[David Miller]

From: Daniel Borkmann <dborkman@redhat.com>
Date: Fri, 30 Aug 2013 18:58:21 +0200

> If there are no further objections, I'll send out the non-RFC version on
> Monday after some more testing. That's just a heads up. This set includes
> my improved fix from yesterday as well besides other things.
> 
> Thanks!

>From a coding style and general correctness perspective these changes
look fine to me.

I'll let IPV6 MLD experts comment on the semantic part.