Merge branch 'mld-sleepable'

Taehee Yoo says:

====================
mld: change context from atomic to sleepable

This patchset changes the context of MLD module.
Before this patchset, MLD functions are atomic context so it couldn't use
sleepable functions and flags.

There are several reasons why MLD functions are under atomic context.
1. It uses timer API.
Timer expiration functions are executed in the atomic context.
2. atomic locks
MLD functions use rwlock and spinlock to protect their own resources.

So, in order to switch context, this patchset converts resources to use
RCU and removes atomic locks and timer API.

1. The first patch convert from the timer API to delayed work.
Timer API is used for delaying some works.
MLD protocol has a delay mechanism, which is used for replying to a query.
If a listener receives a query from a router, it should send a response
after some delay. But because of timer expire function is executed in
the atomic context, this patch convert from timer API to the delayed work.

2. The fourth patch deletes inet6_dev->mc_lock.
The mc_lock has protected inet6_dev->mc_tomb pointer.
But this pointer is already protected by RTNL and it isn't be used by
datapath. So, it isn't be needed and because of this, many atomic context
critical sections are deleted.

3. The fifth patch convert ip6_sf_socklist to RCU.
ip6_sf_socklist has been protected by ipv6_mc_socklist->sflock(rwlock).
But this is already protected by RTNL So if it is converted to use RCU
in order to be used in the datapath, the sflock is no more needed.
So, its control path context can be switched to sleepable.

4. The sixth patch convert ip6_sf_list to RCU.
The reason for this patch is the same as the previous patch.

5. The seventh patch convert ifmcaddr6 to RCU.
The reason for this patch is the same as the previous patch.

6. Add new workqueues for processing query/report event.
By this patch, query and report events are processed by workqueue
So context is sleepable, not atomic.
While this logic, it acquires RTNL.

7. Add new mc_lock.
The purpose of this lock is to protect per-interface mld data.
Per-interface mld data is usually used by query/report event handler.
So, query/report event workers need only this lock instead of RTNL.
Therefore, it could reduce bottleneck.

Changelog:
v2 -> v3:
1. Do not use msecs_to_jiffies().
(by Cong Wang)
2. Do not add unnecessary rtnl_lock() and rtnl_unlock().
(by Cong Wang)
3. Fix sparse warnings because of rcu annotation.
(by kernel test robot)
   - Remove some rcu_assign_pointer(), which was used for non-rcu pointer.
   - Add union for rcu pointer.
   - Use rcu API in mld_clear_zeros().
   - Remove remained rcu_read_unlock().
   - Use rcu API for tomb resources.
4. withdraw prevopus 2nd and 3rd patch.
   - "separate two flags from ifmcaddr6->mca_flags"
   - "add a new delayed_work, mc_delrec_work"
5. Add 6th and 7th patch.

v1 -> v2:
1. Withdraw unnecessary refactoring patches.
(by Cong Wang, Eric Dumazet, David Ahern)
    a) convert from array to list.
    b) function rename.
2. Separate big one patch into small several patches.
3. Do not rename 'ifmcaddr6->mca_lock'.
In the v1 patch, this variable was changed to 'ifmcaddr6->mca_work_lock'.
But this is actually not needed.
4. Do not use atomic_t for 'ifmcaddr6->mca_sfcount' and
'ipv6_mc_socklist'->sf_count'.
5. Do not add mld_check_leave_group() function.
6. Do not add ip6_mc_del_src_bulk() function.
7. Do not add ip6_mc_add_src_bulk() function.
8. Do not use rcu_read_lock() in the qeth_l3_add_mcast_rtnl().
(by Julian Wiedmann)
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

9 files changed, 678 insertions, 471 deletions

diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index 35b42275a06c..d308ff744a29 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -1098,8 +1098,9 @@ walk_ipv6:
 	tmp.disp_flag = QETH_DISP_ADDR_ADD;
 	tmp.is_multicast = 1;
 
-	read_lock_bh(&in6_dev->lock);
-	for (im6 = in6_dev->mc_list; im6 != NULL; im6 = im6->next) {
+	for (im6 = rtnl_dereference(in6_dev->mc_list);
+	     im6;
+	     im6 = rtnl_dereference(im6->next)) {
 		tmp.u.a6.addr = im6->mca_addr;
 
 		ipm = qeth_l3_find_addr_by_ip(card, &tmp);
@@ -1117,7 +1118,6 @@ walk_ipv6:
 			 qeth_l3_ipaddr_hash(ipm));
 
 	}
-	read_unlock_bh(&in6_dev->lock);
 
 out:
 	return 0;
diff --git a/include/net/if_inet6.h b/include/net/if_inet6.h
index 8bf5906073bc..71bb4cc4d05d 100644
--- a/include/net/if_inet6.h
+++ b/include/net/if_inet6.h
@@ -78,6 +78,7 @@ struct inet6_ifaddr {
 struct ip6_sf_socklist {
 	unsigned int		sl_max;
 	unsigned int		sl_count;
+	struct rcu_head		rcu;
 	struct in6_addr		sl_addr[];
 };
 
@@ -91,18 +92,18 @@ struct ipv6_mc_socklist {
 	int			ifindex;
 	unsigned int		sfmode;		/* MCAST_{INCLUDE,EXCLUDE} */
 	struct ipv6_mc_socklist __rcu *next;
-	rwlock_t		sflock;
-	struct ip6_sf_socklist	*sflist;
+	struct ip6_sf_socklist	__rcu *sflist;
 	struct rcu_head		rcu;
 };
 
 struct ip6_sf_list {
-	struct ip6_sf_list	*sf_next;
+	struct ip6_sf_list __rcu *sf_next;
 	struct in6_addr		sf_addr;
 	unsigned long		sf_count[2];	/* include/exclude counts */
 	unsigned char		sf_gsresp;	/* include in g & s response? */
 	unsigned char		sf_oldin;	/* change state */
 	unsigned char		sf_crcount;	/* retrans. left to send */
+	struct rcu_head		rcu;
 };
 
 #define MAF_TIMER_RUNNING	0x01
@@ -114,19 +115,19 @@ struct ip6_sf_list {
 struct ifmcaddr6 {
 	struct in6_addr		mca_addr;
 	struct inet6_dev	*idev;
-	struct ifmcaddr6	*next;
-	struct ip6_sf_list	*mca_sources;
-	struct ip6_sf_list	*mca_tomb;
+	struct ifmcaddr6	__rcu *next;
+	struct ip6_sf_list	__rcu *mca_sources;
+	struct ip6_sf_list	__rcu *mca_tomb;
 	unsigned int		mca_sfmode;
 	unsigned char		mca_crcount;
 	unsigned long		mca_sfcount[2];
-	struct timer_list	mca_timer;
+	struct delayed_work	mca_work;
 	unsigned int		mca_flags;
 	int			mca_users;
 	refcount_t		mca_refcnt;
-	spinlock_t		mca_lock;
 	unsigned long		mca_cstamp;
 	unsigned long		mca_tstamp;
+	struct rcu_head		rcu;
 };
 
 /* Anycast stuff */
@@ -165,9 +166,8 @@ struct inet6_dev {
 
 	struct list_head	addr_list;
 
-	struct ifmcaddr6	*mc_list;
-	struct ifmcaddr6	*mc_tomb;
-	spinlock_t		mc_lock;
+	struct ifmcaddr6	__rcu *mc_list;
+	struct ifmcaddr6	__rcu *mc_tomb;
 
 	unsigned char		mc_qrv;		/* Query Robustness Variable */
 	unsigned char		mc_gq_running;
@@ -179,9 +179,18 @@ struct inet6_dev {
 	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 */
+	struct delayed_work	mc_gq_work;	/* general query work */
+	struct delayed_work	mc_ifc_work;	/* interface change work */
+	struct delayed_work	mc_dad_work;	/* dad complete mc work */
+	struct delayed_work	mc_query_work;	/* mld query work */
+	struct delayed_work	mc_report_work;	/* mld report work */
+
+	struct sk_buff_head	mc_query_queue;		/* mld query queue */
+	struct sk_buff_head	mc_report_queue;	/* mld report queue */
+
+	spinlock_t		mc_query_lock;	/* mld query queue lock */
+	spinlock_t		mc_report_lock;	/* mld query report lock */
+	struct mutex		mc_lock;	/* mld global lock */
 
 	struct ifacaddr6	*ac_list;
 	rwlock_t		lock;
diff --git a/include/net/mld.h b/include/net/mld.h
index 496bddb59942..c07359808493 100644
--- a/include/net/mld.h
+++ b/include/net/mld.h
@@ -92,6 +92,9 @@ struct mld2_query {
 #define MLD_EXP_MIN_LIMIT	32768UL
 #define MLDV1_MRD_MAX_COMPAT	(MLD_EXP_MIN_LIMIT - 1)
 
+#define MLD_MAX_QUEUE		8
+#define MLD_MAX_SKBS		32
+
 static inline unsigned long mldv2_mrc(const struct mld2_query *mlh2)
 {
 	/* RFC3810, 5.1.3. Maximum Response Code */
diff --git a/net/batman-adv/multicast.c b/net/batman-adv/multicast.c
index 28166402d30c..1d63c8cbbfe7 100644
--- a/net/batman-adv/multicast.c
+++ b/net/batman-adv/multicast.c
@@ -454,8 +454,9 @@ batadv_mcast_mla_softif_get_ipv6(struct net_device *dev,
 		return 0;
 	}
 
-	read_lock_bh(&in6_dev->lock);
-	for (pmc6 = in6_dev->mc_list; pmc6; pmc6 = pmc6->next) {
+	for (pmc6 = rcu_dereference(in6_dev->mc_list);
+	     pmc6;
+	     pmc6 = rcu_dereference(pmc6->next)) {
 		if (IPV6_ADDR_MC_SCOPE(&pmc6->mca_addr) <
 		    IPV6_ADDR_SCOPE_LINKLOCAL)
 			continue;
@@ -484,7 +485,6 @@ batadv_mcast_mla_softif_get_ipv6(struct net_device *dev,
 		hlist_add_head(&new->list, mcast_list);
 		ret++;
 	}
-	read_unlock_bh(&in6_dev->lock);
 	rcu_read_unlock();
 
 	return ret;
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index f2337fb756ac..b502f78d5091 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -5107,17 +5107,20 @@ next:
 		break;
 	}
 	case MULTICAST_ADDR:
+		read_unlock_bh(&idev->lock);
 		fillargs->event = RTM_GETMULTICAST;
 
 		/* multicast address */
-		for (ifmca = idev->mc_list; ifmca;
-		     ifmca = ifmca->next, ip_idx++) {
+		for (ifmca = rcu_dereference(idev->mc_list);
+		     ifmca;
+		     ifmca = rcu_dereference(ifmca->next), ip_idx++) {
 			if (ip_idx < s_ip_idx)
 				continue;
 			err = inet6_fill_ifmcaddr(skb, ifmca, fillargs);
 			if (err < 0)
 				break;
 		}
+		read_lock_bh(&idev->lock);
 		break;
 	case ANYCAST_ADDR:
 		fillargs->event = RTM_GETANYCAST;
@@ -6093,10 +6096,8 @@ static void __ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
 
 static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
 {
-	rcu_read_lock_bh();
 	if (likely(ifp->idev->dead == 0))
 		__ipv6_ifa_notify(event, ifp);
-	rcu_read_unlock_bh();
 }
 
 #ifdef CONFIG_SYSCTL
diff --git a/net/ipv6/addrconf_core.c b/net/ipv6/addrconf_core.c
index c70c192bc91b..a36626afbc02 100644
--- a/net/ipv6/addrconf_core.c
+++ b/net/ipv6/addrconf_core.c
@@ -250,7 +250,7 @@ void in6_dev_finish_destroy(struct inet6_dev *idev)
 	struct net_device *dev = idev->dev;
 
 	WARN_ON(!list_empty(&idev->addr_list));
-	WARN_ON(idev->mc_list);
+	WARN_ON(rcu_access_pointer(idev->mc_list));
 	WARN_ON(timer_pending(&idev->rs_timer));
 
 #ifdef NET_REFCNT_DEBUG
diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
index 802f5111805a..3c9bacffc9c3 100644
--- a/net/ipv6/af_inet6.c
+++ b/net/ipv6/af_inet6.c
@@ -222,7 +222,7 @@ lookup_protocol:
 	inet->mc_loop	= 1;
 	inet->mc_ttl	= 1;
 	inet->mc_index	= 0;
-	inet->mc_list	= NULL;
+	RCU_INIT_POINTER(inet->mc_list, NULL);
 	inet->rcv_tos	= 0;
 
 	if (net->ipv4.sysctl_ip_no_pmtu_disc)
diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c
index fd1f896115c1..29d38d6b55fb 100644
--- a/net/ipv6/icmp.c
+++ b/net/ipv6/icmp.c
@@ -944,11 +944,11 @@ static int icmpv6_rcv(struct sk_buff *skb)
 
 	case ICMPV6_MGM_QUERY:
 		igmp6_event_query(skb);
-		break;
+		return 0;
 
 	case ICMPV6_MGM_REPORT:
 		igmp6_event_report(skb);
-		break;
+		return 0;
 
 	case ICMPV6_MGM_REDUCTION:
 	case ICMPV6_NI_QUERY:
diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c
index 6c8604390266..49b0cebfdcdc 100644
--- a/net/ipv6/mcast.c
+++ b/net/ipv6/mcast.c
@@ -29,7 +29,6 @@
 #include <linux/socket.h>
 #include <linux/sockios.h>
 #include <linux/jiffies.h>
-#include <linux/times.h>
 #include <linux/net.h>
 #include <linux/in.h>
 #include <linux/in6.h>
@@ -42,6 +41,7 @@
 #include <linux/slab.h>
 #include <linux/pkt_sched.h>
 #include <net/mld.h>
+#include <linux/workqueue.h>
 
 #include <linux/netfilter.h>
 #include <linux/netfilter_ipv6.h>
@@ -67,14 +67,13 @@ static int __mld2_query_bugs[] __attribute__((__unused__)) = {
 	BUILD_BUG_ON_ZERO(offsetof(struct mld2_grec, grec_mca) % 4)
 };
 
+static struct workqueue_struct *mld_wq;
 static struct in6_addr mld2_all_mcr = MLD2_ALL_MCR_INIT;
 
 static void igmp6_join_group(struct ifmcaddr6 *ma);
 static void igmp6_leave_group(struct ifmcaddr6 *ma);
-static void igmp6_timer_handler(struct timer_list *t);
+static void mld_mca_work(struct work_struct *work);
 
-static void mld_gq_timer_expire(struct timer_list *t);
-static void mld_ifc_timer_expire(struct timer_list *t);
 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, struct ifmcaddr6 *pmc);
@@ -112,12 +111,49 @@ int sysctl_mld_qrv __read_mostly = MLD_QRV_DEFAULT;
 /*
  *	socket join on multicast group
  */
+#define mc_dereference(e, idev) \
+	rcu_dereference_protected(e, lockdep_is_held(&(idev)->mc_lock))
+
+#define for_each_pmc_rtnl(np, pmc)				\
+	for (pmc = rtnl_dereference((np)->ipv6_mc_list);	\
+	     pmc;						\
+	     pmc = rtnl_dereference(pmc->next))
 
 #define for_each_pmc_rcu(np, pmc)				\
-	for (pmc = rcu_dereference(np->ipv6_mc_list);		\
-	     pmc != NULL;					\
+	for (pmc = rcu_dereference((np)->ipv6_mc_list);		\
+	     pmc;						\
 	     pmc = rcu_dereference(pmc->next))
 
+#define for_each_psf_mclock(mc, psf)				\
+	for (psf = mc_dereference((mc)->mca_sources, mc->idev);	\
+	     psf;						\
+	     psf = mc_dereference(psf->sf_next, mc->idev))
+
+#define for_each_psf_rcu(mc, psf)				\
+	for (psf = rcu_dereference((mc)->mca_sources);		\
+	     psf;						\
+	     psf = rcu_dereference(psf->sf_next))
+
+#define for_each_psf_tomb(mc, psf)				\
+	for (psf = mc_dereference((mc)->mca_tomb, mc->idev);	\
+	     psf;						\
+	     psf = mc_dereference(psf->sf_next, mc->idev))
+
+#define for_each_mc_mclock(idev, mc)				\
+	for (mc = mc_dereference((idev)->mc_list, idev);	\
+	     mc;						\
+	     mc = mc_dereference(mc->next, idev))
+
+#define for_each_mc_rcu(idev, mc)				\
+	for (mc = rcu_dereference((idev)->mc_list);             \
+	     mc;                                                \
+	     mc = rcu_dereference(mc->next))
+
+#define for_each_mc_tomb(idev, mc)				\
+	for (mc = mc_dereference((idev)->mc_tomb, idev);	\
+	     mc;						\
+	     mc = mc_dereference(mc->next, idev))
+
 static int unsolicited_report_interval(struct inet6_dev *idev)
 {
 	int iv;
@@ -144,15 +180,11 @@ static int __ipv6_sock_mc_join(struct sock *sk, int ifindex,
 	if (!ipv6_addr_is_multicast(addr))
 		return -EINVAL;
 
-	rcu_read_lock();
-	for_each_pmc_rcu(np, mc_lst) {
+	for_each_pmc_rtnl(np, mc_lst) {
 		if ((ifindex == 0 || mc_lst->ifindex == ifindex) &&
-		    ipv6_addr_equal(&mc_lst->addr, addr)) {
-			rcu_read_unlock();
+		    ipv6_addr_equal(&mc_lst->addr, addr))
 			return -EADDRINUSE;
-		}
 	}
-	rcu_read_unlock();
 
 	mc_lst = sock_kmalloc(sk, sizeof(struct ipv6_mc_socklist), GFP_KERNEL);
 
@@ -179,8 +211,7 @@ static int __ipv6_sock_mc_join(struct sock *sk, int ifindex,
 
 	mc_lst->ifindex = dev->ifindex;
 	mc_lst->sfmode = mode;
-	rwlock_init(&mc_lst->sflock);
-	mc_lst->sflist = NULL;
+	RCU_INIT_POINTER(mc_lst->sflist, NULL);
 
 	/*
 	 *	now add/increase the group membership on the device
@@ -239,11 +270,12 @@ int ipv6_sock_mc_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
 			if (dev) {
 				struct inet6_dev *idev = __in6_dev_get(dev);
 
-				(void) ip6_mc_leave_src(sk, mc_lst, idev);
+				ip6_mc_leave_src(sk, mc_lst, idev);
 				if (idev)
 					__ipv6_dev_mc_dec(idev, &mc_lst->addr);
-			} else
-				(void) ip6_mc_leave_src(sk, mc_lst, NULL);
+			} else {
+				ip6_mc_leave_src(sk, mc_lst, NULL);
+			}
 
 			atomic_sub(sizeof(*mc_lst), &sk->sk_omem_alloc);
 			kfree_rcu(mc_lst, rcu);
@@ -255,10 +287,9 @@ int ipv6_sock_mc_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
 }
 EXPORT_SYMBOL(ipv6_sock_mc_drop);
 
-/* called with rcu_read_lock() */
-static struct inet6_dev *ip6_mc_find_dev_rcu(struct net *net,
-					     const struct in6_addr *group,
-					     int ifindex)
+static struct inet6_dev *ip6_mc_find_dev_rtnl(struct net *net,
+					      const struct in6_addr *group,
+					      int ifindex)
 {
 	struct net_device *dev = NULL;
 	struct inet6_dev *idev = NULL;
@@ -270,19 +301,17 @@ static struct inet6_dev *ip6_mc_find_dev_rcu(struct net *net,
 			dev = rt->dst.dev;
 			ip6_rt_put(rt);
 		}
-	} else
-		dev = dev_get_by_index_rcu(net, ifindex);
+	} else {
+		dev = __dev_get_by_index(net, ifindex);
+	}
 
 	if (!dev)
 		return NULL;
 	idev = __in6_dev_get(dev);
 	if (!idev)
 		return NULL;
-	read_lock_bh(&idev->lock);
-	if (idev->dead) {
-		read_unlock_bh(&idev->lock);
+	if (idev->dead)
 		return NULL;
-	}
 	return idev;
 }
 
@@ -303,11 +332,12 @@ void __ipv6_sock_mc_close(struct sock *sk)
 		if (dev) {
 			struct inet6_dev *idev = __in6_dev_get(dev);
 
-			(void) ip6_mc_leave_src(sk, mc_lst, idev);
+			ip6_mc_leave_src(sk, mc_lst, idev);
 			if (idev)
 				__ipv6_dev_mc_dec(idev, &mc_lst->addr);
-		} else
-			(void) ip6_mc_leave_src(sk, mc_lst, NULL);
+		} else {
+			ip6_mc_leave_src(sk, mc_lst, NULL);
+		}
 
 		atomic_sub(sizeof(*mc_lst), &sk->sk_omem_alloc);
 		kfree_rcu(mc_lst, rcu);
@@ -336,7 +366,6 @@ int ip6_mc_source(int add, int omode, struct sock *sk,
 	struct net *net = sock_net(sk);
 	int i, j, rv;
 	int leavegroup = 0;
-	int pmclocked = 0;
 	int err;
 
 	source = &((struct sockaddr_in6 *)&pgsr->gsr_source)->sin6_addr;
@@ -345,16 +374,14 @@ int ip6_mc_source(int add, int omode, struct sock *sk,
 	if (!ipv6_addr_is_multicast(group))
 		return -EINVAL;
 
-	rcu_read_lock();
-	idev = ip6_mc_find_dev_rcu(net, group, pgsr->gsr_interface);
-	if (!idev) {
-		rcu_read_unlock();
+	idev = ip6_mc_find_dev_rtnl(net, group, pgsr->gsr_interface);
+	if (!idev)
 		return -ENODEV;
-	}
 
 	err = -EADDRNOTAVAIL;
 
-	for_each_pmc_rcu(inet6, pmc) {
+	mutex_lock(&idev->mc_lock);
+	for_each_pmc_rtnl(inet6, pmc) {
 		if (pgsr->gsr_interface && pmc->ifindex != pgsr->gsr_interface)
 			continue;
 		if (ipv6_addr_equal(&pmc->addr, group))
@@ -365,7 +392,7 @@ int ip6_mc_source(int add, int omode, struct sock *sk,
 		goto done;
 	}
 	/* if a source filter was set, must be the same mode as before */
-	if (pmc->sflist) {
+	if (rcu_access_pointer(pmc->sflist)) {
 		if (pmc->sfmode != omode) {
 			err = -EINVAL;
 			goto done;
@@ -377,10 +404,7 @@ int ip6_mc_source(int add, int omode, struct sock *sk,
 		pmc->sfmode = omode;
 	}
 
-	write_lock(&pmc->sflock);
-	pmclocked = 1;
-
-	psl = pmc->sflist;
+	psl = rtnl_dereference(pmc->sflist);
 	if (!add) {
 		if (!psl)
 			goto done;	/* err = -EADDRNOTAVAIL */
@@ -420,7 +444,7 @@ int ip6_mc_source(int add, int omode, struct sock *sk,
 
 		if (psl)
 			count += psl->sl_max;
-		newpsl = sock_kmalloc(sk, IP6_SFLSIZE(count), GFP_ATOMIC);
+		newpsl = sock_kmalloc(sk, IP6_SFLSIZE(count), GFP_KERNEL);
 		if (!newpsl) {
 			err = -ENOBUFS;
 			goto done;
@@ -430,9 +454,11 @@ int ip6_mc_source(int add, int omode, struct sock *sk,
 		if (psl) {
 			for (i = 0; i < psl->sl_count; i++)
 				newpsl->sl_addr[i] = psl->sl_addr[i];
-			sock_kfree_s(sk, psl, IP6_SFLSIZE(psl->sl_max));
+			atomic_sub(IP6_SFLSIZE(psl->sl_max), &sk->sk_omem_alloc);
+			kfree_rcu(psl, rcu);
 		}
-		pmc->sflist = psl = newpsl;
+		psl = newpsl;
+		rcu_assign_pointer(pmc->sflist, psl);
 	}
 	rv = 1;	/* > 0 for insert logic below if sl_count is 0 */
 	for (i = 0; i < psl->sl_count; i++) {
@@ -448,10 +474,7 @@ int ip6_mc_source(int add, int omode, struct sock *sk,
 	/* update the interface list */
 	ip6_mc_add_src(idev, group, omode, 1, source, 1);
 done:
-	if (pmclocked)
-		write_unlock(&pmc->sflock);
-	read_unlock_bh(&idev->lock);
-	rcu_read_unlock();
+	mutex_unlock(&idev->mc_lock);
 	if (leavegroup)
 		err = ipv6_sock_mc_drop(sk, pgsr->gsr_interface, group);
 	return err;
@@ -477,13 +500,9 @@ int ip6_mc_msfilter(struct sock *sk, struct group_filter *gsf,
 	    gsf->gf_fmode != MCAST_EXCLUDE)
 		return -EINVAL;
 
-	rcu_read_lock();
-	idev = ip6_mc_find_dev_rcu(net, group, gsf->gf_interface);
-
-	if (!idev) {
-		rcu_read_unlock();
+	idev = ip6_mc_find_dev_rtnl(net, group, gsf->gf_interface);
+	if (!idev)
 		return -ENODEV;
-	}
 
 	err = 0;
 
@@ -492,7 +511,7 @@ int ip6_mc_msfilter(struct sock *sk, struct group_filter *gsf,
 		goto done;
 	}
 
-	for_each_pmc_rcu(inet6, pmc) {
+	for_each_pmc_rtnl(inet6, pmc) {
 		if (pmc->ifindex != gsf->gf_interface)
 			continue;
 		if (ipv6_addr_equal(&pmc->addr, group))
@@ -504,7 +523,7 @@ int ip6_mc_msfilter(struct sock *sk, struct group_filter *gsf,
 	}
 	if (gsf->gf_numsrc) {
 		newpsl = sock_kmalloc(sk, IP6_SFLSIZE(gsf->gf_numsrc),
-							  GFP_ATOMIC);
+							  GFP_KERNEL);
 		if (!newpsl) {
 			err = -ENOBUFS;
 			goto done;
@@ -516,32 +535,37 @@ int ip6_mc_msfilter(struct sock *sk, struct group_filter *gsf,
 			psin6 = (struct sockaddr_in6 *)list;
 			newpsl->sl_addr[i] = psin6->sin6_addr;
 		}
+		mutex_lock(&idev->mc_lock);
 		err = ip6_mc_add_src(idev, group, gsf->gf_fmode,
-			newpsl->sl_count, newpsl->sl_addr, 0);
+				     newpsl->sl_count, newpsl->sl_addr, 0);
 		if (err) {
+			mutex_unlock(&idev->mc_lock);
 			sock_kfree_s(sk, newpsl, IP6_SFLSIZE(newpsl->sl_max));
 			goto done;
 		}
+		mutex_unlock(&idev->mc_lock);
 	} else {
 		newpsl = NULL;
-		(void) ip6_mc_add_src(idev, group, gsf->gf_fmode, 0, NULL, 0);
+		mutex_lock(&idev->mc_lock);
+		ip6_mc_add_src(idev, group, gsf->gf_fmode, 0, NULL, 0);
+		mutex_unlock(&idev->mc_lock);
 	}
 
-	write_lock(&pmc->sflock);
-	psl = pmc->sflist;
+	mutex_lock(&idev->mc_lock);
+	psl = rtnl_dereference(pmc->sflist);
 	if (psl) {
-		(void) ip6_mc_del_src(idev, group, pmc->sfmode,
-			psl->sl_count, psl->sl_addr, 0);
-		sock_kfree_s(sk, psl, IP6_SFLSIZE(psl->sl_max));
-	} else
-		(void) ip6_mc_del_src(idev, group, pmc->sfmode, 0, NULL, 0);
-	pmc->sflist = newpsl;
+		ip6_mc_del_src(idev, group, pmc->sfmode,
+			       psl->sl_count, psl->sl_addr, 0);
+		atomic_sub(IP6_SFLSIZE(psl->sl_max), &sk->sk_omem_alloc);
+		kfree_rcu(psl, rcu);
+	} else {
+		ip6_mc_del_src(idev, group, pmc->sfmode, 0, NULL, 0);
+	}
+	mutex_unlock(&idev->mc_lock);
+	rcu_assign_pointer(pmc->sflist, newpsl);
 	pmc->sfmode = gsf->gf_fmode;
-	write_unlock(&pmc->sflock);
 	err = 0;
 done:
-	read_unlock_bh(&idev->lock);
-	rcu_read_unlock();
 	if (leavegroup)
 		err = ipv6_sock_mc_drop(sk, gsf->gf_interface, group);
 	return err;
@@ -563,13 +587,9 @@ int ip6_mc_msfget(struct sock *sk, struct group_filter *gsf,
 	if (!ipv6_addr_is_multicast(group))
 		return -EINVAL;
 
-	rcu_read_lock();
-	idev = ip6_mc_find_dev_rcu(net, group, gsf->gf_interface);
-
-	if (!idev) {
-		rcu_read_unlock();
+	idev = ip6_mc_find_dev_rtnl(net, group, gsf->gf_interface);
+	if (!idev)
 		return -ENODEV;
-	}
 
 	err = -EADDRNOTAVAIL;
 	/* changes to the ipv6_mc_list require the socket lock and
@@ -577,25 +597,22 @@ int ip6_mc_msfget(struct sock *sk, struct group_filter *gsf,
 	 * so reading the list is safe.
 	 */
 
-	for_each_pmc_rcu(inet6, pmc) {
+	for_each_pmc_rtnl(inet6, pmc) {
 		if (pmc->ifindex != gsf->gf_interface)
 			continue;
 		if (ipv6_addr_equal(group, &pmc->addr))
 			break;
 	}
 	if (!pmc)		/* must have a prior join */
-		goto done;
+		return err;
+
 	gsf->gf_fmode = pmc->sfmode;
-	psl = pmc->sflist;
+	psl = rtnl_dereference(pmc->sflist);
 	count = psl ? psl->sl_count : 0;
-	read_unlock_bh(&idev->lock);
-	rcu_read_unlock();
 
 	copycount = count < gsf->gf_numsrc ? count : gsf->gf_numsrc;
 	gsf->gf_numsrc = count;
-	/* changes to psl require the socket lock, and a write lock
-	 * on pmc->sflock. We have the socket lock so reading here is safe.
-	 */
+
 	for (i = 0; i < copycount; i++, p++) {
 		struct sockaddr_in6 *psin6;
 		struct sockaddr_storage ss;
@@ -608,10 +625,6 @@ int ip6_mc_msfget(struct sock *sk, struct group_filter *gsf,
 			return -EFAULT;
 	}
 	return 0;
-done:
-	read_unlock_bh(&idev->lock);
-	rcu_read_unlock();
-	return err;
 }
 
 bool inet6_mc_check(struct sock *sk, const struct in6_addr *mc_addr,
@@ -631,8 +644,7 @@ bool inet6_mc_check(struct sock *sk, const struct in6_addr *mc_addr,
 		rcu_read_unlock();
 		return np->mc_all;
 	}
-	read_lock(&mc->sflock);
-	psl = mc->sflist;
+	psl = rcu_dereference(mc->sflist);
 	if (!psl) {
 		rv = mc->sfmode == MCAST_EXCLUDE;
 	} else {
@@ -647,12 +659,12 @@ bool inet6_mc_check(struct sock *sk, const struct in6_addr *mc_addr,
 		if (mc->sfmode == MCAST_EXCLUDE && i < psl->sl_count)
 			rv = false;
 	}
-	read_unlock(&mc->sflock);
 	rcu_read_unlock();
 
 	return rv;
 }
 
+/* called with mc_lock */
 static void igmp6_group_added(struct ifmcaddr6 *mc)
 {
 	struct net_device *dev = mc->idev->dev;
@@ -662,13 +674,11 @@ static void igmp6_group_added(struct ifmcaddr6 *mc)
 	    IPV6_ADDR_SCOPE_LINKLOCAL)
 		return;
 
-	spin_lock_bh(&mc->mca_lock);
 	if (!(mc->mca_flags&MAF_LOADED)) {
 		mc->mca_flags |= MAF_LOADED;
 		if (ndisc_mc_map(&mc->mca_addr, buf, dev, 0) == 0)
 			dev_mc_add(dev, buf);
 	}
-	spin_unlock_bh(&mc->mca_lock);
 
 	if (!(dev->flags & IFF_UP) || (mc->mca_flags & MAF_NOREPORT))
 		return;
@@ -689,6 +699,7 @@ static void igmp6_group_added(struct ifmcaddr6 *mc)
 	mld_ifc_event(mc->idev);
 }
 
+/* called with mc_lock */
 static void igmp6_group_dropped(struct ifmcaddr6 *mc)
 {
 	struct net_device *dev = mc->idev->dev;
@@ -698,28 +709,25 @@ static void igmp6_group_dropped(struct ifmcaddr6 *mc)
 	    IPV6_ADDR_SCOPE_LINKLOCAL)
 		return;
 
-	spin_lock_bh(&mc->mca_lock);
 	if (mc->mca_flags&MAF_LOADED) {
 		mc->mca_flags &= ~MAF_LOADED;
 		if (ndisc_mc_map(&mc->mca_addr, buf, dev, 0) == 0)
 			dev_mc_del(dev, buf);
 	}
 
-	spin_unlock_bh(&mc->mca_lock);
 	if (mc->mca_flags & MAF_NOREPORT)
 		return;
 
 	if (!mc->idev->dead)
 		igmp6_leave_group(mc);
 
-	spin_lock_bh(&mc->mca_lock);
-	if (del_timer(&mc->mca_timer))
+	if (cancel_delayed_work(&mc->mca_work))
 		refcount_dec(&mc->mca_refcnt);
-	spin_unlock_bh(&mc->mca_lock);
 }
 
 /*
  * deleted ifmcaddr6 manipulation
+ * called with mc_lock
  */
 static void mld_add_delrec(struct inet6_dev *idev, struct ifmcaddr6 *im)
 {
@@ -731,12 +739,10 @@ static void mld_add_delrec(struct inet6_dev *idev, struct ifmcaddr6 *im)
 	 * for deleted items allows change reports to use common code with
 	 * non-deleted or query-response MCA's.
 	 */
-	pmc = kzalloc(sizeof(*pmc), GFP_ATOMIC);
+	pmc = kzalloc(sizeof(*pmc), GFP_KERNEL);
 	if (!pmc)
 		return;
 
-	spin_lock_bh(&im->mca_lock);
-	spin_lock_init(&pmc->mca_lock);
 	pmc->idev = im->idev;
 	in6_dev_hold(idev);
 	pmc->mca_addr = im->mca_addr;
@@ -745,90 +751,110 @@ static void mld_add_delrec(struct inet6_dev *idev, struct ifmcaddr6 *im)
 	if (pmc->mca_sfmode == MCAST_INCLUDE) {
 		struct ip6_sf_list *psf;
 
-		pmc->mca_tomb = im->mca_tomb;
-		pmc->mca_sources = im->mca_sources;
-		im->mca_tomb = im->mca_sources = NULL;
-		for (psf = pmc->mca_sources; psf; psf = psf->sf_next)
+		rcu_assign_pointer(pmc->mca_tomb,
+				   mc_dereference(im->mca_tomb, idev));
+		rcu_assign_pointer(pmc->mca_sources,
+				   mc_dereference(im->mca_sources, idev));
+		RCU_INIT_POINTER(im->mca_tomb, NULL);
+		RCU_INIT_POINTER(im->mca_sources, NULL);
+
+		for_each_psf_mclock(pmc, psf)
 			psf->sf_crcount = pmc->mca_crcount;
 	}
-	spin_unlock_bh(&im->mca_lock);
 
-	spin_lock_bh(&idev->mc_lock);
-	pmc->next = idev->mc_tomb;
-	idev->mc_tomb = pmc;
-	spin_unlock_bh(&idev->mc_lock);
+	rcu_assign_pointer(pmc->next, idev->mc_tomb);
+	rcu_assign_pointer(idev->mc_tomb, pmc);
 }
 
+/* called with mc_lock */
 static void mld_del_delrec(struct inet6_dev *idev, struct ifmcaddr6 *im)
 {
-	struct ifmcaddr6 *pmc, *pmc_prev;
-	struct ip6_sf_list *psf;
+	struct ip6_sf_list *psf, *sources, *tomb;
 	struct in6_addr *pmca = &im->mca_addr;
+	struct ifmcaddr6 *pmc, *pmc_prev;
 
-	spin_lock_bh(&idev->mc_lock);
 	pmc_prev = NULL;
-	for (pmc = idev->mc_tomb; pmc; pmc = pmc->next) {
+	for_each_mc_tomb(idev, pmc) {
 		if (ipv6_addr_equal(&pmc->mca_addr, pmca))
 			break;
 		pmc_prev = pmc;
 	}
 	if (pmc) {
 		if (pmc_prev)
-			pmc_prev->next = pmc->next;
+			rcu_assign_pointer(pmc_prev->next, pmc->next);
 		else
-			idev->mc_tomb = pmc->next;
+			rcu_assign_pointer(idev->mc_tomb, pmc->next);
 	}
-	spin_unlock_bh(&idev->mc_lock);
 
-	spin_lock_bh(&im->mca_lock);
 	if (pmc) {
 		im->idev = pmc->idev;
 		if (im->mca_sfmode == MCAST_INCLUDE) {
-			swap(im->mca_tomb, pmc->mca_tomb);
-			swap(im->mca_sources, pmc->mca_sources);
-			for (psf = im->mca_sources; psf; psf = psf->sf_next)
+			tomb = rcu_replace_pointer(im->mca_tomb,
+						   mc_dereference(pmc->mca_tomb, pmc->idev),
+						   lockdep_is_held(&im->idev->mc_lock));
+			rcu_assign_pointer(pmc->mca_tomb, tomb);
+
+			sources = rcu_replace_pointer(im->mca_sources,
+						      mc_dereference(pmc->mca_sources, pmc->idev),
+						      lockdep_is_held(&im->idev->mc_lock));
+			rcu_assign_pointer(pmc->mca_sources, sources);
+			for_each_psf_mclock(im, psf)
 				psf->sf_crcount = idev->mc_qrv;
 		} else {
 			im->mca_crcount = idev->mc_qrv;
 		}
 		in6_dev_put(pmc->idev);
 		ip6_mc_clear_src(pmc);
-		kfree(pmc);
+		kfree_rcu(pmc, rcu);
 	}
-	spin_unlock_bh(&im->mca_lock);
 }
 
+/* called with mc_lock */
 static void mld_clear_delrec(struct inet6_dev *idev)
 {
 	struct ifmcaddr6 *pmc, *nextpmc;
 
-	spin_lock_bh(&idev->mc_lock);
-	pmc = idev->mc_tomb;
-	idev->mc_tomb = NULL;
-	spin_unlock_bh(&idev->mc_lock);
+	pmc = mc_dereference(idev->mc_tomb, idev);
+	RCU_INIT_POINTER(idev->mc_tomb, NULL);
 
 	for (; pmc; pmc = nextpmc) {
-		nextpmc = pmc->next;
+		nextpmc = mc_dereference(pmc->next, idev);
 		ip6_mc_clear_src(pmc);
 		in6_dev_put(pmc->idev);
-		kfree(pmc);
+		kfree_rcu(pmc, rcu);
 	}
 
 	/* clear dead sources, too */
-	read_lock_bh(&idev->lock);
-	for (pmc = idev->mc_list; pmc; pmc = pmc->next) {
+	for_each_mc_mclock(idev, pmc) {
 		struct ip6_sf_list *psf, *psf_next;
 
-		spin_lock_bh(&pmc->mca_lock);
-		psf = pmc->mca_tomb;
-		pmc->mca_tomb = NULL;
-		spin_unlock_bh(&pmc->mca_lock);
+		psf = mc_dereference(pmc->mca_tomb, idev);
+		RCU_INIT_POINTER(pmc->mca_tomb, NULL);
 		for (; psf; psf = psf_next) {
-			psf_next = psf->sf_next;
-			kfree(psf);
+			psf_next = mc_dereference(psf->sf_next, idev);
+			kfree_rcu(psf, rcu);
 		}
 	}
-	read_unlock_bh(&idev->lock);
+}
+
+static void mld_clear_query(struct inet6_dev *idev)
+{
+	struct sk_buff *skb;
+
+	spin_lock_bh(&idev->mc_query_lock);
+	while ((skb = __skb_dequeue(&idev->mc_query_queue)))
+		kfree_skb(skb);
+	spin_unlock_bh(&idev->mc_query_lock);
+}
+
+static void mld_clear_report(struct inet6_dev *idev)
+{
+	struct sk_buff *skb;
+
+	spin_lock_bh(&idev->mc_report_lock);
+	while ((skb = __skb_dequeue(&idev->mc_report_queue)))
+		kfree_skb(skb);
+	spin_unlock_bh(&idev->mc_report_lock);
 }
 
 static void mca_get(struct ifmcaddr6 *mc)
@@ -840,21 +866,22 @@ static void ma_put(struct ifmcaddr6 *mc)
 {
 	if (refcount_dec_and_test(&mc->mca_refcnt)) {
 		in6_dev_put(mc->idev);
-		kfree(mc);
+		kfree_rcu(mc, rcu);
 	}
 }
 
+/* called with mc_lock */
 static struct ifmcaddr6 *mca_alloc(struct inet6_dev *idev,
 				   const struct in6_addr *addr,
 				   unsigned int mode)
 {
 	struct ifmcaddr6 *mc;
 
-	mc = kzalloc(sizeof(*mc), GFP_ATOMIC);
+	mc = kzalloc(sizeof(*mc), GFP_KERNEL);
 	if (!mc)
 		return NULL;
 
-	timer_setup(&mc->mca_timer, igmp6_timer_handler, 0);
+	INIT_DELAYED_WORK(&mc->mca_work, mld_mca_work);
 
 	mc->mca_addr = *addr;
 	mc->idev = idev; /* reference taken by caller */
@@ -862,7 +889,6 @@ static struct ifmcaddr6 *mca_alloc(struct inet6_dev *idev,
 	/* mca_stamp should be updated upon changes */
 	mc->mca_cstamp = mc->mca_tstamp = jiffies;
 	refcount_set(&mc->mca_refcnt, 1);
-	spin_lock_init(&mc->mca_lock);
 
 	mc->mca_sfmode = mode;
 	mc->mca_sfcount[mode] = 1;
@@ -891,18 +917,17 @@ static int __ipv6_dev_mc_inc(struct net_device *dev,
 	if (!idev)
 		return -EINVAL;
 
-	write_lock_bh(&idev->lock);
 	if (idev->dead) {
-		write_unlock_bh(&idev->lock);
 		in6_dev_put(idev);
 		return -ENODEV;
 	}
 
-	for (mc = idev->mc_list; mc; mc = mc->next) {
+	mutex_lock(&idev->mc_lock);
+	for_each_mc_mclock(idev, mc) {
 		if (ipv6_addr_equal(&mc->mca_addr, addr)) {
 			mc->mca_users++;
-			write_unlock_bh(&idev->lock);
 			ip6_mc_add_src(idev, &mc->mca_addr, mode, 0, NULL, 0);
+			mutex_unlock(&idev->mc_lock);
 			in6_dev_put(idev);
 			return 0;
 		}
@@ -910,22 +935,19 @@ static int __ipv6_dev_mc_inc(struct net_device *dev,
 
 	mc = mca_alloc(idev, addr, mode);
 	if (!mc) {
-		write_unlock_bh(&idev->lock);
+		mutex_unlock(&idev->mc_lock);
 		in6_dev_put(idev);
 		return -ENOMEM;
 	}
 
-	mc->next = idev->mc_list;
-	idev->mc_list = mc;
+	rcu_assign_pointer(mc->next, idev->mc_list);
+	rcu_assign_pointer(idev->mc_list, mc);
 
-	/* Hold this for the code below before we unlock,
-	 * it is already exposed via idev->mc_list.
-	 */
 	mca_get(mc);
-	write_unlock_bh(&idev->lock);
 
 	mld_del_delrec(idev, mc);
 	igmp6_group_added(mc);
+	mutex_unlock(&idev->mc_lock);
 	ma_put(mc);
 	return 0;
 }
@@ -937,33 +959,35 @@ int ipv6_dev_mc_inc(struct net_device *dev, const struct in6_addr *addr)
 EXPORT_SYMBOL(ipv6_dev_mc_inc);
 
 /*
- *	device multicast group del
+ * device multicast group del
  */
 int __ipv6_dev_mc_dec(struct inet6_dev *idev, const struct in6_addr *addr)
 {
-	struct ifmcaddr6 *ma, **map;
+	struct ifmcaddr6 *ma, __rcu **map;
 
 	ASSERT_RTNL();
 
-	write_lock_bh(&idev->lock);
-	for (map = &idev->mc_list; (ma = *map) != NULL; map = &ma->next) {
+	mutex_lock(&idev->mc_lock);
+	for (map = &idev->mc_list;
+	     (ma = mc_dereference(*map, idev));
+	     map = &ma->next) {
 		if (ipv6_addr_equal(&ma->mca_addr, addr)) {
 			if (--ma->mca_users == 0) {
 				*map = ma->next;
-				write_unlock_bh(&idev->lock);
 
 				igmp6_group_dropped(ma);
 				ip6_mc_clear_src(ma);
+				mutex_unlock(&idev->mc_lock);
 
 				ma_put(ma);
 				return 0;
 			}
-			write_unlock_bh(&idev->lock);
+			mutex_unlock(&idev->mc_lock);
 			return 0;
 		}
 	}
-	write_unlock_bh(&idev->lock);
 
+	mutex_unlock(&idev->mc_lock);
 	return -ENOENT;
 }
 
@@ -997,8 +1021,7 @@ bool ipv6_chk_mcast_addr(struct net_device *dev, const struct in6_addr *group,
 	rcu_read_lock();
 	idev = __in6_dev_get(dev);
 	if (idev) {
-		read_lock_bh(&idev->lock);
-		for (mc = idev->mc_list; mc; mc = mc->next) {
+		for_each_mc_rcu(idev, mc) {
 			if (ipv6_addr_equal(&mc->mca_addr, group))
 				break;
 		}
@@ -1006,8 +1029,7 @@ bool ipv6_chk_mcast_addr(struct net_device *dev, const struct in6_addr *group,
 			if (src_addr && !ipv6_addr_any(src_addr)) {
 				struct ip6_sf_list *psf;
 
-				spin_lock_bh(&mc->mca_lock);
-				for (psf = mc->mca_sources; psf; psf = psf->sf_next) {
+				for_each_psf_rcu(mc, psf) {
 					if (ipv6_addr_equal(&psf->sf_addr, src_addr))
 						break;
 				}
@@ -1017,89 +1039,107 @@ bool ipv6_chk_mcast_addr(struct net_device *dev, const struct in6_addr *group,
 						mc->mca_sfcount[MCAST_EXCLUDE];
 				else
 					rv = mc->mca_sfcount[MCAST_EXCLUDE] != 0;
-				spin_unlock_bh(&mc->mca_lock);
 			} else
 				rv = true; /* don't filter unspecified source */
 		}
-		read_unlock_bh(&idev->lock);
 	}
 	rcu_read_unlock();
 	return rv;
 }
 
-static void mld_gq_start_timer(struct inet6_dev *idev)
+/* called with mc_lock */
+static void mld_gq_start_work(struct inet6_dev *idev)
 {
 	unsigned long tv = prandom_u32() % idev->mc_maxdelay;
 
 	idev->mc_gq_running = 1;
-	if (!mod_timer(&idev->mc_gq_timer, jiffies+tv+2))
+	if (!mod_delayed_work(mld_wq, &idev->mc_gq_work, tv + 2))
 		in6_dev_hold(idev);
 }
 
-static void mld_gq_stop_timer(struct inet6_dev *idev)
+/* called with mc_lock */
+static void mld_gq_stop_work(struct inet6_dev *idev)
 {
 	idev->mc_gq_running = 0;
-	if (del_timer(&idev->mc_gq_timer))
+	if (cancel_delayed_work(&idev->mc_gq_work))
 		__in6_dev_put(idev);
 }
 
-static void mld_ifc_start_timer(struct inet6_dev *idev, unsigned long delay)
+/* called with mc_lock */
+static void mld_ifc_start_work(struct inet6_dev *idev, unsigned long delay)
 {
 	unsigned long tv = prandom_u32() % delay;
 
-	if (!mod_timer(&idev->mc_ifc_timer, jiffies+tv+2))
+	if (!mod_delayed_work(mld_wq, &idev->mc_ifc_work, tv + 2))
 		in6_dev_hold(idev);
 }
 
-static void mld_ifc_stop_timer(struct inet6_dev *idev)
+/* called with mc_lock */
+static void mld_ifc_stop_work(struct inet6_dev *idev)
 {
 	idev->mc_ifc_count = 0;
-	if (del_timer(&idev->mc_ifc_timer))
+	if (cancel_delayed_work(&idev->mc_ifc_work))
 		__in6_dev_put(idev);
 }
 
-static void mld_dad_start_timer(struct inet6_dev *idev, unsigned long delay)
+/* called with mc_lock */
+static void mld_dad_start_work(struct inet6_dev *idev, unsigned long delay)
 {
 	unsigned long tv = prandom_u32() % delay;
 
-	if (!mod_timer(&idev->mc_dad_timer, jiffies+tv+2))
+	if (!mod_delayed_work(mld_wq, &idev->mc_dad_work, tv + 2))
 		in6_dev_hold(idev);
 }
 
-static void mld_dad_stop_timer(struct inet6_dev *idev)
+static void mld_dad_stop_work(struct inet6_dev *idev)
+{
+	if (cancel_delayed_work(&idev->mc_dad_work))
+		__in6_dev_put(idev);
+}
+
+static void mld_query_stop_work(struct inet6_dev *idev)
 {
-	if (del_timer(&idev->mc_dad_timer))
+	spin_lock_bh(&idev->mc_query_lock);
+	if (cancel_delayed_work(&idev->mc_query_work))
+		__in6_dev_put(idev);
+	spin_unlock_bh(&idev->mc_query_lock);
+}
+
+static void mld_report_stop_work(struct inet6_dev *idev)
+{
+	if (cancel_delayed_work_sync(&idev->mc_report_work))
 		__in6_dev_put(idev);
 }
 
 /*
- *	IGMP handling (alias multicast ICMPv6 messages)
+ * IGMP handling (alias multicast ICMPv6 messages)
+ * called with mc_lock
  */
-
 static void igmp6_group_queried(struct ifmcaddr6 *ma, unsigned long resptime)
 {
 	unsigned long delay = resptime;
 
-	/* Do not start timer for these addresses */
+	/* Do not start work for these addresses */
 	if (ipv6_addr_is_ll_all_nodes(&ma->mca_addr) ||
 	    IPV6_ADDR_MC_SCOPE(&ma->mca_addr) < IPV6_ADDR_SCOPE_LINKLOCAL)
 		return;
 
-	if (del_timer(&ma->mca_timer)) {
+	if (cancel_delayed_work(&ma->mca_work)) {
 		refcount_dec(&ma->mca_refcnt);
-		delay = ma->mca_timer.expires - jiffies;
+		delay = ma->mca_work.timer.expires - jiffies;
 	}
 
 	if (delay >= resptime)
 		delay = prandom_u32() % resptime;
 
-	ma->mca_timer.expires = jiffies + delay;
-	if (!mod_timer(&ma->mca_timer, jiffies + delay))
+	if (!mod_delayed_work(mld_wq, &ma->mca_work, delay))
 		refcount_inc(&ma->mca_refcnt);
 	ma->mca_flags |= MAF_TIMER_RUNNING;
 }
 
-/* mark EXCLUDE-mode sources */
+/* mark EXCLUDE-mode sources
+ * called with mc_lock
+ */
 static bool mld_xmarksources(struct ifmcaddr6 *pmc, int nsrcs,
 			     const struct in6_addr *srcs)
 {
@@ -1107,7 +1147,7 @@ static bool mld_xmarksources(struct ifmcaddr6 *pmc, int nsrcs,
 	int i, scount;
 
 	scount = 0;
-	for (psf = pmc->mca_sources; psf; psf = psf->sf_next) {
+	for_each_psf_mclock(pmc, psf) {
 		if (scount == nsrcs)
 			break;
 		for (i = 0; i < nsrcs; i++) {
@@ -1128,6 +1168,7 @@ static bool mld_xmarksources(struct ifmcaddr6 *pmc, int nsrcs,
 	return true;
 }
 
+/* called with mc_lock */
 static bool mld_marksources(struct ifmcaddr6 *pmc, int nsrcs,
 			    const struct in6_addr *srcs)
 {
@@ -1140,7 +1181,7 @@ static bool mld_marksources(struct ifmcaddr6 *pmc, int nsrcs,
 	/* mark INCLUDE-mode sources */
 
 	scount = 0;
-	for (psf = pmc->mca_sources; psf; psf = psf->sf_next) {
+	for_each_psf_mclock(pmc, psf) {
 		if (scount == nsrcs)
 			break;
 		for (i = 0; i < nsrcs; i++) {
@@ -1305,10 +1346,10 @@ static int mld_process_v1(struct inet6_dev *idev, struct mld_msg *mld,
 	if (v1_query)
 		mld_set_v1_mode(idev);
 
-	/* cancel MLDv2 report timer */
-	mld_gq_stop_timer(idev);
-	/* cancel the interface change timer */
-	mld_ifc_stop_timer(idev);
+	/* cancel MLDv2 report work */
+	mld_gq_stop_work(idev);
+	/* cancel the interface change work */
+	mld_ifc_stop_work(idev);
 	/* clear deleted report items */
 	mld_clear_delrec(idev);
 
@@ -1332,18 +1373,41 @@ static int mld_process_v2(struct inet6_dev *idev, struct mld2_query *mld,
 /* called with rcu_read_lock() */
 int igmp6_event_query(struct sk_buff *skb)
 {
+	struct inet6_dev *idev = __in6_dev_get(skb->dev);
+
+	if (!idev)
+		return -EINVAL;
+
+	if (idev->dead) {
+		kfree_skb(skb);
+		return -ENODEV;
+	}
+
+	spin_lock_bh(&idev->mc_query_lock);
+	if (skb_queue_len(&idev->mc_query_queue) < MLD_MAX_SKBS) {
+		__skb_queue_tail(&idev->mc_query_queue, skb);
+		if (!mod_delayed_work(mld_wq, &idev->mc_query_work, 0))
+			in6_dev_hold(idev);
+	}
+	spin_unlock_bh(&idev->mc_query_lock);
+
+	return 0;
+}
+
+static void __mld_query_work(struct sk_buff *skb)
+{
 	struct mld2_query *mlh2 = NULL;
-	struct ifmcaddr6 *ma;
 	const struct in6_addr *group;
 	unsigned long max_delay;
 	struct inet6_dev *idev;
+	struct ifmcaddr6 *ma;
 	struct mld_msg *mld;
 	int group_type;
 	int mark = 0;
 	int len, err;
 
 	if (!pskb_may_pull(skb, sizeof(struct in6_addr)))
-		return -EINVAL;
+		goto kfree_skb;
 
 	/* compute payload length excluding extension headers */
 	len = ntohs(ipv6_hdr(skb)->payload_len) + sizeof(struct ipv6hdr);
@@ -1360,11 +1424,11 @@ int igmp6_event_query(struct sk_buff *skb)
 	    ipv6_hdr(skb)->hop_limit != 1 ||
 	    !(IP6CB(skb)->flags & IP6SKB_ROUTERALERT) ||
 	    IP6CB(skb)->ra != htons(IPV6_OPT_ROUTERALERT_MLD))
-		return -EINVAL;
+		goto kfree_skb;
 
-	idev = __in6_dev_get(skb->dev);
+	idev = in6_dev_get(skb->dev);
 	if (!idev)
-		return 0;
+		goto kfree_skb;
 
 	mld = (struct mld_msg *)icmp6_hdr(skb);
 	group = &mld->mld_mca;
@@ -1372,60 +1436,56 @@ int igmp6_event_query(struct sk_buff *skb)
 
 	if (group_type != IPV6_ADDR_ANY &&
 	    !(group_type&IPV6_ADDR_MULTICAST))
-		return -EINVAL;
+		goto out;
 
 	if (len < MLD_V1_QUERY_LEN) {
-		return -EINVAL;
+		goto out;
 	} else if (len == MLD_V1_QUERY_LEN || mld_in_v1_mode(idev)) {
 		err = mld_process_v1(idev, mld, &max_delay,
 				     len == MLD_V1_QUERY_LEN);
 		if (err < 0)
-			return err;
+			goto out;
 	} 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;
+			goto out;
 
 		mlh2 = (struct mld2_query *)skb_transport_header(skb);
 
 		err = mld_process_v2(idev, mlh2, &max_delay);
 		if (err < 0)
-			return err;
+			goto out;
 
 		if (group_type == IPV6_ADDR_ANY) { /* general query */
 			if (mlh2->mld2q_nsrcs)
-				return -EINVAL; /* no sources allowed */
+				goto out; /* no sources allowed */
 
-			mld_gq_start_timer(idev);
-			return 0;
+			mld_gq_start_work(idev);
+			goto out;
 		}
 		/* mark sources to include, if group & source-specific */
 		if (mlh2->mld2q_nsrcs != 0) {
 			if (!pskb_may_pull(skb, srcs_offset +
 			    ntohs(mlh2->mld2q_nsrcs) * sizeof(struct in6_addr)))
-				return -EINVAL;
+				goto out;
 
 			mlh2 = (struct mld2_query *)skb_transport_header(skb);
 			mark = 1;
 		}
 	} else {
-		return -EINVAL;
+		goto out;
 	}
 
-	read_lock_bh(&idev->lock);
 	if (group_type == IPV6_ADDR_ANY) {
-		for (ma = idev->mc_list; ma; ma = ma->next) {
-			spin_lock_bh(&ma->mca_lock);
+		for_each_mc_mclock(idev, ma) {
 			igmp6_group_queried(ma, max_delay);
-			spin_unlock_bh(&ma->mca_lock);
 		}
 	} else {
-		for (ma = idev->mc_list; ma; ma = ma->next) {
+		for_each_mc_mclock(idev, ma) {
 			if (!ipv6_addr_equal(group, &ma->mca_addr))
 				continue;
-			spin_lock_bh(&ma->mca_lock);
 			if (ma->mca_flags & MAF_TIMER_RUNNING) {
 				/* gsquery <- gsquery && mark */
 				if (!mark)
@@ -1440,34 +1500,91 @@ int igmp6_event_query(struct sk_buff *skb)
 			if (!(ma->mca_flags & MAF_GSQUERY) ||
 			    mld_marksources(ma, ntohs(mlh2->mld2q_nsrcs), mlh2->mld2q_srcs))
 				igmp6_group_queried(ma, max_delay);
-			spin_unlock_bh(&ma->mca_lock);
 			break;
 		}
 	}
-	read_unlock_bh(&idev->lock);
 
-	return 0;
+out:
+	in6_dev_put(idev);
+kfree_skb:
+	consume_skb(skb);
+}
+
+static void mld_query_work(struct work_struct *work)
+{
+	struct inet6_dev *idev = container_of(to_delayed_work(work),
+					      struct inet6_dev,
+					      mc_query_work);
+	struct sk_buff_head q;
+	struct sk_buff *skb;
+	bool rework = false;
+	int cnt = 0;
+
+	skb_queue_head_init(&q);
+
+	spin_lock_bh(&idev->mc_query_lock);
+	while ((skb = __skb_dequeue(&idev->mc_query_queue))) {
+		__skb_queue_tail(&q, skb);
+
+		if (++cnt >= MLD_MAX_QUEUE) {
+			rework = true;
+			schedule_delayed_work(&idev->mc_query_work, 0);
+			break;
+		}
+	}
+	spin_unlock_bh(&idev->mc_query_lock);
+
+	mutex_lock(&idev->mc_lock);
+	while ((skb = __skb_dequeue(&q)))
+		__mld_query_work(skb);
+	mutex_unlock(&idev->mc_lock);
+
+	if (!rework)
+		in6_dev_put(idev);
 }
 
 /* called with rcu_read_lock() */
 int igmp6_event_report(struct sk_buff *skb)
 {
-	struct ifmcaddr6 *ma;
+	struct inet6_dev *idev = __in6_dev_get(skb->dev);
+
+	if (!idev)
+		return -EINVAL;
+
+	if (idev->dead) {
+		kfree_skb(skb);
+		return -ENODEV;
+	}
+
+	spin_lock_bh(&idev->mc_report_lock);
+	if (skb_queue_len(&idev->mc_report_queue) < MLD_MAX_SKBS) {
+		__skb_queue_tail(&idev->mc_report_queue, skb);
+		if (!mod_delayed_work(mld_wq, &idev->mc_report_work, 0))
+			in6_dev_hold(idev);
+	}
+	spin_unlock_bh(&idev->mc_report_lock);
+
+	return 0;
+}
+
+static void __mld_report_work(struct sk_buff *skb)
+{
 	struct inet6_dev *idev;
+	struct ifmcaddr6 *ma;
 	struct mld_msg *mld;
 	int addr_type;
 
 	/* Our own report looped back. Ignore it. */
 	if (skb->pkt_type == PACKET_LOOPBACK)
-		return 0;
+		goto kfree_skb;
 
 	/* send our report if the MC router may not have heard this report */
 	if (skb->pkt_type != PACKET_MULTICAST &&
 	    skb->pkt_type != PACKET_BROADCAST)
-		return 0;
+		goto kfree_skb;
 
 	if (!pskb_may_pull(skb, sizeof(*mld) - sizeof(struct icmp6hdr)))
-		return -EINVAL;
+		goto kfree_skb;
 
 	mld = (struct mld_msg *)icmp6_hdr(skb);
 
@@ -1475,29 +1592,61 @@ int igmp6_event_report(struct sk_buff *skb)
 	addr_type = ipv6_addr_type(&ipv6_hdr(skb)->saddr);
 	if (addr_type != IPV6_ADDR_ANY &&
 	    !(addr_type&IPV6_ADDR_LINKLOCAL))
-		return -EINVAL;
+		goto kfree_skb;
 
-	idev = __in6_dev_get(skb->dev);
+	idev = in6_dev_get(skb->dev);
 	if (!idev)
-		return -ENODEV;
+		goto kfree_skb;
 
 	/*
-	 *	Cancel the timer for this group
+	 *	Cancel the work for this group
 	 */
 
-	read_lock_bh(&idev->lock);
-	for (ma = idev->mc_list; ma; ma = ma->next) {
+	for_each_mc_mclock(idev, ma) {
 		if (ipv6_addr_equal(&ma->mca_addr, &mld->mld_mca)) {
-			spin_lock(&ma->mca_lock);
-			if (del_timer(&ma->mca_timer))
+			if (cancel_delayed_work(&ma->mca_work))
 				refcount_dec(&ma->mca_refcnt);
-			ma->mca_flags &= ~(MAF_LAST_REPORTER|MAF_TIMER_RUNNING);
-			spin_unlock(&ma->mca_lock);
+			ma->mca_flags &= ~(MAF_LAST_REPORTER |
+					   MAF_TIMER_RUNNING);
 			break;
 		}
 	}
-	read_unlock_bh(&idev->lock);
-	return 0;
+
+	in6_dev_put(idev);
+kfree_skb:
+	consume_skb(skb);
+}
+
+static void mld_report_work(struct work_struct *work)
+{
+	struct inet6_dev *idev = container_of(to_delayed_work(work),
+					      struct inet6_dev,
+					      mc_report_work);
+	struct sk_buff_head q;
+	struct sk_buff *skb;
+	bool rework = false;
+	int cnt = 0;
+
+	skb_queue_head_init(&q);
+	spin_lock_bh(&idev->mc_report_lock);
+	while ((skb = __skb_dequeue(&idev->mc_report_queue))) {
+		__skb_queue_tail(&q, skb);
+
+		if (++cnt >= MLD_MAX_QUEUE) {
+			rework = true;
+			schedule_delayed_work(&idev->mc_report_work, 0);
+			break;
+		}
+	}
+	spin_unlock_bh(&idev->mc_report_lock);
+
+	mutex_lock(&idev->mc_lock);
+	while ((skb = __skb_dequeue(&q)))
+		__mld_report_work(skb);
+	mutex_unlock(&idev->mc_lock);
+
+	if (!rework)
+		in6_dev_put(idev);
 }
 
 static bool is_in(struct ifmcaddr6 *pmc, struct ip6_sf_list *psf, int type,
@@ -1550,7 +1699,7 @@ mld_scount(struct ifmcaddr6 *pmc, int type, int gdeleted, int sdeleted)
 	struct ip6_sf_list *psf;
 	int scount = 0;
 
-	for (psf = pmc->mca_sources; psf; psf = psf->sf_next) {
+	for_each_psf_mclock(pmc, psf) {
 		if (!is_in(pmc, psf, type, gdeleted, sdeleted))
 			continue;
 		scount++;
@@ -1724,15 +1873,18 @@ static struct sk_buff *add_grhead(struct sk_buff *skb, struct ifmcaddr6 *pmc,
 
 #define AVAILABLE(skb)	((skb) ? skb_availroom(skb) : 0)
 
+/* called with mc_lock */
 static struct sk_buff *add_grec(struct sk_buff *skb, struct ifmcaddr6 *pmc,
-	int type, int gdeleted, int sdeleted, int crsend)
+				int type, int gdeleted, int sdeleted,
+				int crsend)
 {
+	struct ip6_sf_list *psf, *psf_prev, *psf_next;
+	int scount, stotal, first, isquery, truncate;
+	struct ip6_sf_list __rcu **psf_list;
 	struct inet6_dev *idev = pmc->idev;
 	struct net_device *dev = idev->dev;
-	struct mld2_report *pmr;
 	struct mld2_grec *pgr = NULL;
-	struct ip6_sf_list *psf, *psf_next, *psf_prev, **psf_list;
-	int scount, stotal, first, isquery, truncate;
+	struct mld2_report *pmr;
 	unsigned int mtu;
 
 	if (pmc->mca_flags & MAF_NOREPORT)
@@ -1751,7 +1903,7 @@ static struct sk_buff *add_grec(struct sk_buff *skb, struct ifmcaddr6 *pmc,
 
 	psf_list = sdeleted ? &pmc->mca_tomb : &pmc->mca_sources;
 
-	if (!*psf_list)
+	if (!rcu_access_pointer(*psf_list))
 		goto empty_source;
 
 	pmr = skb ? (struct mld2_report *)skb_transport_header(skb) : NULL;
@@ -1767,10 +1919,12 @@ static struct sk_buff *add_grec(struct sk_buff *skb, struct ifmcaddr6 *pmc,
 	}
 	first = 1;
 	psf_prev = NULL;
-	for (psf = *psf_list; psf; psf = psf_next) {
+	for (psf = mc_dereference(*psf_list, idev);
+	     psf;
+	     psf = psf_next) {
 		struct in6_addr *psrc;
 
-		psf_next = psf->sf_next;
+		psf_next = mc_dereference(psf->sf_next, idev);
 
 		if (!is_in(pmc, psf, type, gdeleted, sdeleted) && !crsend) {
 			psf_prev = psf;
@@ -1817,10 +1971,12 @@ decrease_sf_crcount:
 			psf->sf_crcount--;
 			if ((sdeleted || gdeleted) && psf->sf_crcount == 0) {
 				if (psf_prev)
-					psf_prev->sf_next = psf->sf_next;
+					rcu_assign_pointer(psf_prev->sf_next,
+							   mc_dereference(psf->sf_next, idev));
 				else
-					*psf_list = psf->sf_next;
-				kfree(psf);
+					rcu_assign_pointer(*psf_list,
+							   mc_dereference(psf->sf_next, idev));
+				kfree_rcu(psf, rcu);
 				continue;
 			}
 		}
@@ -1849,72 +2005,73 @@ empty_source:
 	return skb;
 }
 
+/* called with mc_lock */
 static void mld_send_report(struct inet6_dev *idev, struct ifmcaddr6 *pmc)
 {
 	struct sk_buff *skb = NULL;
 	int type;
 
-	read_lock_bh(&idev->lock);
 	if (!pmc) {
-		for (pmc = idev->mc_list; pmc; pmc = pmc->next) {
+		for_each_mc_mclock(idev, pmc) {
 			if (pmc->mca_flags & MAF_NOREPORT)
 				continue;
-			spin_lock_bh(&pmc->mca_lock);
 			if (pmc->mca_sfcount[MCAST_EXCLUDE])
 				type = MLD2_MODE_IS_EXCLUDE;
 			else
 				type = MLD2_MODE_IS_INCLUDE;
 			skb = add_grec(skb, pmc, type, 0, 0, 0);
-			spin_unlock_bh(&pmc->mca_lock);
 		}
 	} else {
-		spin_lock_bh(&pmc->mca_lock);
 		if (pmc->mca_sfcount[MCAST_EXCLUDE])
 			type = MLD2_MODE_IS_EXCLUDE;
 		else
 			type = MLD2_MODE_IS_INCLUDE;
 		skb = add_grec(skb, pmc, type, 0, 0, 0);
-		spin_unlock_bh(&pmc->mca_lock);
 	}
-	read_unlock_bh(&idev->lock);
 	if (skb)
 		mld_sendpack(skb);
 }
 
 /*
  * remove zero-count source records from a source filter list
+ * called with mc_lock
  */
-static void mld_clear_zeros(struct ip6_sf_list **ppsf)
+static void mld_clear_zeros(struct ip6_sf_list __rcu **ppsf, struct inet6_dev *idev)
 {
 	struct ip6_sf_list *psf_prev, *psf_next, *psf;
 
 	psf_prev = NULL;
-	for (psf = *ppsf; psf; psf = psf_next) {
-		psf_next = psf->sf_next;
+	for (psf = mc_dereference(*ppsf, idev);
+	     psf;
+	     psf = psf_next) {
+		psf_next = mc_dereference(psf->sf_next, idev);
 		if (psf->sf_crcount == 0) {
 			if (psf_prev)
-				psf_prev->sf_next = psf->sf_next;
+				rcu_assign_pointer(psf_prev->sf_next,
+						   mc_dereference(psf->sf_next, idev));
 			else
-				*ppsf = psf->sf_next;
-			kfree(psf);
-		} else
+				rcu_assign_pointer(*ppsf,
+						   mc_dereference(psf->sf_next, idev));
+			kfree_rcu(psf, rcu);
+		} else {
 			psf_prev = psf;
+		}
 	}
 }
 
+/* called with mc_lock */
 static void mld_send_cr(struct inet6_dev *idev)
 {
 	struct ifmcaddr6 *pmc, *pmc_prev, *pmc_next;
 	struct sk_buff *skb = NULL;
 	int type, dtype;
 
-	read_lock_bh(&idev->lock);
-	spin_lock(&idev->mc_lock);
-
 	/* deleted MCA's */
 	pmc_prev = NULL;
-	for (pmc = idev->mc_tomb; pmc; pmc = pmc_next) {
-		pmc_next = pmc->next;
+	for (pmc = mc_dereference(idev->mc_tomb, idev);
+	     pmc;
+	     pmc = pmc_next) {
+		pmc_next = mc_dereference(pmc->next, idev);
 		if (pmc->mca_sfmode == MCAST_INCLUDE) {
 			type = MLD2_BLOCK_OLD_SOURCES;
 			dtype = MLD2_BLOCK_OLD_SOURCES;
@@ -1928,26 +2085,25 @@ static void mld_send_cr(struct inet6_dev *idev)
 			}
 			pmc->mca_crcount--;
 			if (pmc->mca_crcount == 0) {
-				mld_clear_zeros(&pmc->mca_tomb);
-				mld_clear_zeros(&pmc->mca_sources);
+				mld_clear_zeros(&pmc->mca_tomb, idev);
+				mld_clear_zeros(&pmc->mca_sources, idev);
 			}
 		}
-		if (pmc->mca_crcount == 0 && !pmc->mca_tomb &&
-		    !pmc->mca_sources) {
+		if (pmc->mca_crcount == 0 &&
+		    !rcu_access_pointer(pmc->mca_tomb) &&
+		    !rcu_access_pointer(pmc->mca_sources)) {
 			if (pmc_prev)
-				pmc_prev->next = pmc_next;
+				rcu_assign_pointer(pmc_prev->next, pmc_next);
 			else
-				idev->mc_tomb = pmc_next;
+				rcu_assign_pointer(idev->mc_tomb, pmc_next);
 			in6_dev_put(pmc->idev);
-			kfree(pmc);
+			kfree_rcu(pmc, rcu);
 		} else
 			pmc_prev = pmc;
 	}
-	spin_unlock(&idev->mc_lock);
 
 	/* change recs */
-	for (pmc = idev->mc_list; pmc; pmc = pmc->next) {
-		spin_lock_bh(&pmc->mca_lock);
+	for_each_mc_mclock(idev, pmc) {
 		if (pmc->mca_sfcount[MCAST_EXCLUDE]) {
 			type = MLD2_BLOCK_OLD_SOURCES;
 			dtype = MLD2_ALLOW_NEW_SOURCES;
@@ -1967,9 +2123,7 @@ static void mld_send_cr(struct inet6_dev *idev)
 			skb = add_grec(skb, pmc, type, 0, 0, 0);
 			pmc->mca_crcount--;
 		}
-		spin_unlock_bh(&pmc->mca_lock);
 	}
-	read_unlock_bh(&idev->lock);
 	if (!skb)
 		return;
 	(void) mld_sendpack(skb);
@@ -2071,6 +2225,7 @@ err_out:
 	goto out;
 }
 
+/* called with mc_lock */
 static void mld_send_initial_cr(struct inet6_dev *idev)
 {
 	struct sk_buff *skb;
@@ -2081,47 +2236,49 @@ static void mld_send_initial_cr(struct inet6_dev *idev)
 		return;
 
 	skb = NULL;
-	read_lock_bh(&idev->lock);
-	for (pmc = idev->mc_list; pmc; pmc = pmc->next) {
-		spin_lock_bh(&pmc->mca_lock);
+	for_each_mc_mclock(idev, pmc) {
 		if (pmc->mca_sfcount[MCAST_EXCLUDE])
 			type = MLD2_CHANGE_TO_EXCLUDE;
 		else
 			type = MLD2_ALLOW_NEW_SOURCES;
 		skb = add_grec(skb, pmc, type, 0, 0, 1);
-		spin_unlock_bh(&pmc->mca_lock);
 	}
-	read_unlock_bh(&idev->lock);
 	if (skb)
 		mld_sendpack(skb);
 }
 
 void ipv6_mc_dad_complete(struct inet6_dev *idev)
 {
+	mutex_lock(&idev->mc_lock);
 	idev->mc_dad_count = idev->mc_qrv;
 	if (idev->mc_dad_count) {
 		mld_send_initial_cr(idev);
 		idev->mc_dad_count--;
 		if (idev->mc_dad_count)
-			mld_dad_start_timer(idev,
-					    unsolicited_report_interval(idev));
+			mld_dad_start_work(idev,
+					   unsolicited_report_interval(idev));
 	}
+	mutex_unlock(&idev->mc_lock);
 }
 
-static void mld_dad_timer_expire(struct timer_list *t)
+static void mld_dad_work(struct work_struct *work)
 {
-	struct inet6_dev *idev = from_timer(idev, t, mc_dad_timer);
-
+	struct inet6_dev *idev = container_of(to_delayed_work(work),
+					      struct inet6_dev,
+					      mc_dad_work);
+	mutex_lock(&idev->mc_lock);
 	mld_send_initial_cr(idev);
 	if (idev->mc_dad_count) {
 		idev->mc_dad_count--;
 		if (idev->mc_dad_count)
-			mld_dad_start_timer(idev,
-					    unsolicited_report_interval(idev));
+			mld_dad_start_work(idev,
+					   unsolicited_report_interval(idev));
 	}
+	mutex_unlock(&idev->mc_lock);
 	in6_dev_put(idev);
 }
 
+/* called with mc_lock */
 static int ip6_mc_del1_src(struct ifmcaddr6 *pmc, int sfmode,
 	const struct in6_addr *psfsrc)
 {
@@ -2129,7 +2286,7 @@ static int ip6_mc_del1_src(struct ifmcaddr6 *pmc, int sfmode,
 	int rv = 0;
 
 	psf_prev = NULL;
-	for (psf = pmc->mca_sources; psf; psf = psf->sf_next) {
+	for_each_psf_mclock(pmc, psf) {
 		if (ipv6_addr_equal(&psf->sf_addr, psfsrc))
 			break;
 		psf_prev = psf;
@@ -2144,21 +2301,27 @@ static int ip6_mc_del1_src(struct ifmcaddr6 *pmc, int sfmode,
 
 		/* no more filters for this source */
 		if (psf_prev)
-			psf_prev->sf_next = psf->sf_next;
+			rcu_assign_pointer(psf_prev->sf_next,
+					   mc_dereference(psf->sf_next, idev));
 		else
-			pmc->mca_sources = psf->sf_next;
+			rcu_assign_pointer(pmc->mca_sources,
+					   mc_dereference(psf->sf_next, idev));
+
 		if (psf->sf_oldin && !(pmc->mca_flags & MAF_NOREPORT) &&
 		    !mld_in_v1_mode(idev)) {
 			psf->sf_crcount = idev->mc_qrv;
-			psf->sf_next = pmc->mca_tomb;
-			pmc->mca_tomb = psf;
+			rcu_assign_pointer(psf->sf_next,
+					   mc_dereference(pmc->mca_tomb, idev));
+			rcu_assign_pointer(pmc->mca_tomb, psf);
 			rv = 1;
-		} else
-			kfree(psf);
+		} else {
+			kfree_rcu(psf, rcu);
+		}
 	}
 	return rv;
 }
 
+/* called with mc_lock */
 static int ip6_mc_del_src(struct inet6_dev *idev, const struct in6_addr *pmca,
 			  int sfmode, int sfcount, const struct in6_addr *psfsrc,
 			  int delta)
@@ -2169,24 +2332,19 @@ static int ip6_mc_del_src(struct inet6_dev *idev, const struct in6_addr *pmca,
 
 	if (!idev)
 		return -ENODEV;
-	read_lock_bh(&idev->lock);
-	for (pmc = idev->mc_list; pmc; pmc = pmc->next) {
+
+	for_each_mc_mclock(idev, pmc) {
 		if (ipv6_addr_equal(pmca, &pmc->mca_addr))
 			break;
 	}
-	if (!pmc) {
-		/* MCA not found?? bug */
-		read_unlock_bh(&idev->lock);
+	if (!pmc)
 		return -ESRCH;
-	}
-	spin_lock_bh(&pmc->mca_lock);
+
 	sf_markstate(pmc);
 	if (!delta) {
-		if (!pmc->mca_sfcount[sfmode]) {
-			spin_unlock_bh(&pmc->mca_lock);
-			read_unlock_bh(&idev->lock);
+		if (!pmc->mca_sfcount[sfmode])
 			return -EINVAL;
-		}
+
 		pmc->mca_sfcount[sfmode]--;
 	}
 	err = 0;
@@ -2206,18 +2364,19 @@ static int ip6_mc_del_src(struct inet6_dev *idev, const struct in6_addr *pmca,
 		pmc->mca_sfmode = MCAST_INCLUDE;
 		pmc->mca_crcount = idev->mc_qrv;
 		idev->mc_ifc_count = pmc->mca_crcount;
-		for (psf = pmc->mca_sources; psf; psf = psf->sf_next)
+		for_each_psf_mclock(pmc, psf)
 			psf->sf_crcount = 0;
 		mld_ifc_event(pmc->idev);
-	} else if (sf_setstate(pmc) || changerec)
+	} else if (sf_setstate(pmc) || changerec) {
 		mld_ifc_event(pmc->idev);
-	spin_unlock_bh(&pmc->mca_lock);
-	read_unlock_bh(&idev->lock);
+	}
+
 	return err;
 }
 
 /*
  * Add multicast single-source filter to the interface list
+ * called with mc_lock
  */
 static int ip6_mc_add1_src(struct ifmcaddr6 *pmc, int sfmode,
 	const struct in6_addr *psfsrc)
@@ -2225,40 +2384,45 @@ static int ip6_mc_add1_src(struct ifmcaddr6 *pmc, int sfmode,
 	struct ip6_sf_list *psf, *psf_prev;
 
 	psf_prev = NULL;
-	for (psf = pmc->mca_sources; psf; psf = psf->sf_next) {
+	for_each_psf_mclock(pmc, psf) {
 		if (ipv6_addr_equal(&psf->sf_addr, psfsrc))
 			break;
 		psf_prev = psf;
 	}
 	if (!psf) {
-		psf = kzalloc(sizeof(*psf), GFP_ATOMIC);
+		psf = kzalloc(sizeof(*psf), GFP_KERNEL);
 		if (!psf)
 			return -ENOBUFS;
 
 		psf->sf_addr = *psfsrc;
 		if (psf_prev) {
-			psf_prev->sf_next = psf;
-		} else
-			pmc->mca_sources = psf;
+			rcu_assign_pointer(psf_prev->sf_next, psf);
+		} else {
+			rcu_assign_pointer(pmc->mca_sources, psf);
+		}
 	}
 	psf->sf_count[sfmode]++;
 	return 0;
 }
 
+/* called with mc_lock */
 static void sf_markstate(struct ifmcaddr6 *pmc)
 {
 	struct ip6_sf_list *psf;
 	int mca_xcount = pmc->mca_sfcount[MCAST_EXCLUDE];
 
-	for (psf = pmc->mca_sources; psf; psf = psf->sf_next)
+	for_each_psf_mclock(pmc, psf) {
 		if (pmc->mca_sfcount[MCAST_EXCLUDE]) {
 			psf->sf_oldin = mca_xcount ==
 				psf->sf_count[MCAST_EXCLUDE] &&
 				!psf->sf_count[MCAST_INCLUDE];
-		} else
+		} else {
 			psf->sf_oldin = psf->sf_count[MCAST_INCLUDE] != 0;
+		}
+	}
 }
 
+/* called with mc_lock */
 static int sf_setstate(struct ifmcaddr6 *pmc)
 {
 	struct ip6_sf_list *psf, *dpsf;
@@ -2267,7 +2431,7 @@ static int sf_setstate(struct ifmcaddr6 *pmc)
 	int new_in, rv;
 
 	rv = 0;
-	for (psf = pmc->mca_sources; psf; psf = psf->sf_next) {
+	for_each_psf_mclock(pmc, psf) {
 		if (pmc->mca_sfcount[MCAST_EXCLUDE]) {
 			new_in = mca_xcount == psf->sf_count[MCAST_EXCLUDE] &&
 				!psf->sf_count[MCAST_INCLUDE];
@@ -2277,8 +2441,7 @@ static int sf_setstate(struct ifmcaddr6 *pmc)
 			if (!psf->sf_oldin) {
 				struct ip6_sf_list *prev = NULL;
 
-				for (dpsf = pmc->mca_tomb; dpsf;
-				     dpsf = dpsf->sf_next) {
+				for_each_psf_tomb(pmc, dpsf) {
 					if (ipv6_addr_equal(&dpsf->sf_addr,
 					    &psf->sf_addr))
 						break;
@@ -2286,10 +2449,14 @@ static int sf_setstate(struct ifmcaddr6 *pmc)
 				}
 				if (dpsf) {
 					if (prev)
-						prev->sf_next = dpsf->sf_next;
+						rcu_assign_pointer(prev->sf_next,
+								   mc_dereference(dpsf->sf_next,
+										  pmc->idev));
 					else
-						pmc->mca_tomb = dpsf->sf_next;
-					kfree(dpsf);
+						rcu_assign_pointer(pmc->mca_tomb,
+								   mc_dereference(dpsf->sf_next,
+										  pmc->idev));
+					kfree_rcu(dpsf, rcu);
 				}
 				psf->sf_crcount = qrv;
 				rv++;
@@ -2300,18 +2467,19 @@ static int sf_setstate(struct ifmcaddr6 *pmc)
 			 * add or update "delete" records if an active filter
 			 * is now inactive
 			 */
-			for (dpsf = pmc->mca_tomb; dpsf; dpsf = dpsf->sf_next)
+
+			for_each_psf_tomb(pmc, dpsf)
 				if (ipv6_addr_equal(&dpsf->sf_addr,
 				    &psf->sf_addr))
 					break;
 			if (!dpsf) {
-				dpsf = kmalloc(sizeof(*dpsf), GFP_ATOMIC);
+				dpsf = kmalloc(sizeof(*dpsf), GFP_KERNEL);
 				if (!dpsf)
 					continue;
 				*dpsf = *psf;
-				/* pmc->mca_lock held by callers */
-				dpsf->sf_next = pmc->mca_tomb;
-				pmc->mca_tomb = dpsf;
+				rcu_assign_pointer(dpsf->sf_next,
+						   mc_dereference(pmc->mca_tomb, pmc->idev));
+				rcu_assign_pointer(pmc->mca_tomb, dpsf);
 			}
 			dpsf->sf_crcount = qrv;
 			rv++;
@@ -2322,6 +2490,7 @@ static int sf_setstate(struct ifmcaddr6 *pmc)
 
 /*
  * Add multicast source filter list to the interface list
+ * called with mc_lock
  */
 static int ip6_mc_add_src(struct inet6_dev *idev, const struct in6_addr *pmca,
 			  int sfmode, int sfcount, const struct in6_addr *psfsrc,
@@ -2333,17 +2502,13 @@ static int ip6_mc_add_src(struct inet6_dev *idev, const struct in6_addr *pmca,
 
 	if (!idev)
 		return -ENODEV;
-	read_lock_bh(&idev->lock);
-	for (pmc = idev->mc_list; pmc; pmc = pmc->next) {
+
+	for_each_mc_mclock(idev, pmc) {
 		if (ipv6_addr_equal(pmca, &pmc->mca_addr))
 			break;
 	}
-	if (!pmc) {
-		/* MCA not found?? bug */
-		read_unlock_bh(&idev->lock);
+	if (!pmc)
 		return -ESRCH;
-	}
-	spin_lock_bh(&pmc->mca_lock);
 
 	sf_markstate(pmc);
 	isexclude = pmc->mca_sfmode == MCAST_EXCLUDE;
@@ -2374,36 +2539,40 @@ static int ip6_mc_add_src(struct inet6_dev *idev, const struct in6_addr *pmca,
 
 		pmc->mca_crcount = idev->mc_qrv;
 		idev->mc_ifc_count = pmc->mca_crcount;
-		for (psf = pmc->mca_sources; psf; psf = psf->sf_next)
+		for_each_psf_mclock(pmc, psf)
 			psf->sf_crcount = 0;
 		mld_ifc_event(idev);
-	} else if (sf_setstate(pmc))
+	} else if (sf_setstate(pmc)) {
 		mld_ifc_event(idev);
-	spin_unlock_bh(&pmc->mca_lock);
-	read_unlock_bh(&idev->lock);
+	}
 	return err;
 }
 
+/* called with mc_lock */
 static void ip6_mc_clear_src(struct ifmcaddr6 *pmc)
 {
 	struct ip6_sf_list *psf, *nextpsf;
 
-	for (psf = pmc->mca_tomb; psf; psf = nextpsf) {
-		nextpsf = psf->sf_next;
-		kfree(psf);
+	for (psf = mc_dereference(pmc->mca_tomb, pmc->idev);
+	     psf;
+	     psf = nextpsf) {
+		nextpsf = mc_dereference(psf->sf_next, pmc->idev);
+		kfree_rcu(psf, rcu);
 	}
-	pmc->mca_tomb = NULL;
-	for (psf = pmc->mca_sources; psf; psf = nextpsf) {
-		nextpsf = psf->sf_next;
-		kfree(psf);
+	RCU_INIT_POINTER(pmc->mca_tomb, NULL);
+	for (psf = mc_dereference(pmc->mca_sources, pmc->idev);
+	     psf;
+	     psf = nextpsf) {
+		nextpsf = mc_dereference(psf->sf_next, pmc->idev);
+		kfree_rcu(psf, rcu);
 	}
-	pmc->mca_sources = NULL;
+	RCU_INIT_POINTER(pmc->mca_sources, NULL);
 	pmc->mca_sfmode = MCAST_EXCLUDE;
 	pmc->mca_sfcount[MCAST_INCLUDE] = 0;
 	pmc->mca_sfcount[MCAST_EXCLUDE] = 1;
 }
 
-
+/* called with mc_lock */
 static void igmp6_join_group(struct ifmcaddr6 *ma)
 {
 	unsigned long delay;
@@ -2415,93 +2584,115 @@ static void igmp6_join_group(struct ifmcaddr6 *ma)
 
 	delay = prandom_u32() % unsolicited_report_interval(ma->idev);
 
-	spin_lock_bh(&ma->mca_lock);
-	if (del_timer(&ma->mca_timer)) {
+	if (cancel_delayed_work(&ma->mca_work)) {
 		refcount_dec(&ma->mca_refcnt);
-		delay = ma->mca_timer.expires - jiffies;
+		delay = ma->mca_work.timer.expires - jiffies;
 	}
 
-	if (!mod_timer(&ma->mca_timer, jiffies + delay))
+	if (!mod_delayed_work(mld_wq, &ma->mca_work, delay))
 		refcount_inc(&ma->mca_refcnt);
 	ma->mca_flags |= MAF_TIMER_RUNNING | MAF_LAST_REPORTER;
-	spin_unlock_bh(&ma->mca_lock);
 }
 
 static int ip6_mc_leave_src(struct sock *sk, struct ipv6_mc_socklist *iml,
 			    struct inet6_dev *idev)
 {
+	struct ip6_sf_socklist *psl;
 	int err;
 
-	write_lock_bh(&iml->sflock);
-	if (!iml->sflist) {
+	psl = rtnl_dereference(iml->sflist);
+
+	if (idev)
+		mutex_lock(&idev->mc_lock);
+
+	if (!psl) {
 		/* any-source empty exclude case */
 		err = ip6_mc_del_src(idev, &iml->addr, iml->sfmode, 0, NULL, 0);
 	} else {
 		err = ip6_mc_del_src(idev, &iml->addr, iml->sfmode,
-				iml->sflist->sl_count, iml->sflist->sl_addr, 0);
-		sock_kfree_s(sk, iml->sflist, IP6_SFLSIZE(iml->sflist->sl_max));
-		iml->sflist = NULL;
+				     psl->sl_count, psl->sl_addr, 0);
+		RCU_INIT_POINTER(iml->sflist, NULL);
+		atomic_sub(IP6_SFLSIZE(psl->sl_max), &sk->sk_omem_alloc);
+		kfree_rcu(psl, rcu);
 	}
-	write_unlock_bh(&iml->sflock);
+
+	if (idev)
+		mutex_unlock(&idev->mc_lock);
+
 	return err;
 }
 
+/* called with mc_lock */
 static void igmp6_leave_group(struct ifmcaddr6 *ma)
 {
 	if (mld_in_v1_mode(ma->idev)) {
-		if (ma->mca_flags & MAF_LAST_REPORTER)
+		if (ma->mca_flags & MAF_LAST_REPORTER) {
 			igmp6_send(&ma->mca_addr, ma->idev->dev,
 				ICMPV6_MGM_REDUCTION);
+		}
 	} else {
 		mld_add_delrec(ma->idev, ma);
 		mld_ifc_event(ma->idev);
 	}
 }
 
-static void mld_gq_timer_expire(struct timer_list *t)
+static void mld_gq_work(struct work_struct *work)
 {
-	struct inet6_dev *idev = from_timer(idev, t, mc_gq_timer);
+	struct inet6_dev *idev = container_of(to_delayed_work(work),
+					      struct inet6_dev,
+					      mc_gq_work);
 
-	idev->mc_gq_running = 0;
+	mutex_lock(&idev->mc_lock);
 	mld_send_report(idev, NULL);
+	idev->mc_gq_running = 0;
+	mutex_unlock(&idev->mc_lock);
+
 	in6_dev_put(idev);
 }
 
-static void mld_ifc_timer_expire(struct timer_list *t)
+static void mld_ifc_work(struct work_struct *work)
 {
-	struct inet6_dev *idev = from_timer(idev, t, mc_ifc_timer);
+	struct inet6_dev *idev = container_of(to_delayed_work(work),
+					      struct inet6_dev,
+					      mc_ifc_work);
 
+	mutex_lock(&idev->mc_lock);
 	mld_send_cr(idev);
+
 	if (idev->mc_ifc_count) {
 		idev->mc_ifc_count--;
 		if (idev->mc_ifc_count)
-			mld_ifc_start_timer(idev,
-					    unsolicited_report_interval(idev));
+			mld_ifc_start_work(idev,
+					   unsolicited_report_interval(idev));
 	}
+	mutex_unlock(&idev->mc_lock);
 	in6_dev_put(idev);
 }
 
+/* called with mc_lock */
 static void mld_ifc_event(struct inet6_dev *idev)
 {
 	if (mld_in_v1_mode(idev))
 		return;
+
 	idev->mc_ifc_count = idev->mc_qrv;
-	mld_ifc_start_timer(idev, 1);
+	mld_ifc_start_work(idev, 1);
 }
 
-static void igmp6_timer_handler(struct timer_list *t)
+static void mld_mca_work(struct work_struct *work)
 {
-	struct ifmcaddr6 *ma = from_timer(ma, t, mca_timer);
+	struct ifmcaddr6 *ma = container_of(to_delayed_work(work),
+					    struct ifmcaddr6, mca_work);
 
+	mutex_lock(&ma->idev->mc_lock);
 	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);
-
-	spin_lock(&ma->mca_lock);
 	ma->mca_flags |=  MAF_LAST_REPORTER;
 	ma->mca_flags &= ~MAF_TIMER_RUNNING;
-	spin_unlock(&ma->mca_lock);
+	mutex_unlock(&ma->idev->mc_lock);
+
 	ma_put(ma);
 }
 
@@ -2513,10 +2704,10 @@ void ipv6_mc_unmap(struct inet6_dev *idev)
 
 	/* Install multicast list, except for all-nodes (already installed) */
 
-	read_lock_bh(&idev->lock);
-	for (i = idev->mc_list; i; i = i->next)
+	mutex_lock(&idev->mc_lock);
+	for_each_mc_mclock(idev, i)
 		igmp6_group_dropped(i);
-	read_unlock_bh(&idev->lock);
+	mutex_unlock(&idev->mc_lock);
 }
 
 void ipv6_mc_remap(struct inet6_dev *idev)
@@ -2525,25 +2716,25 @@ void ipv6_mc_remap(struct inet6_dev *idev)
 }
 
 /* Device going down */
-
 void ipv6_mc_down(struct inet6_dev *idev)
 {
 	struct ifmcaddr6 *i;
 
+	mutex_lock(&idev->mc_lock);
 	/* Withdraw multicast list */
-
-	read_lock_bh(&idev->lock);
-
-	for (i = idev->mc_list; i; i = i->next)
+	for_each_mc_mclock(idev, i)
 		igmp6_group_dropped(i);
+	mutex_unlock(&idev->mc_lock);
 
-	/* Should stop timer after group drop. or we will
-	 * start timer again in mld_ifc_event()
+	/* Should stop work after group drop. or we will
+	 * start work again in mld_ifc_event()
 	 */
-	mld_ifc_stop_timer(idev);
-	mld_gq_stop_timer(idev);
-	mld_dad_stop_timer(idev);
-	read_unlock_bh(&idev->lock);
+	synchronize_net();
+	mld_query_stop_work(idev);
+	mld_report_stop_work(idev);
+	mld_ifc_stop_work(idev);
+	mld_gq_stop_work(idev);
+	mld_dad_stop_work(idev);
 }
 
 static void ipv6_mc_reset(struct inet6_dev *idev)
@@ -2563,29 +2754,33 @@ void ipv6_mc_up(struct inet6_dev *idev)
 
 	/* Install multicast list, except for all-nodes (already installed) */
 
-	read_lock_bh(&idev->lock);
 	ipv6_mc_reset(idev);
-	for (i = idev->mc_list; i; i = i->next) {
+	mutex_lock(&idev->mc_lock);
+	for_each_mc_mclock(idev, i) {
 		mld_del_delrec(idev, i);
 		igmp6_group_added(i);
 	}
-	read_unlock_bh(&idev->lock);
+	mutex_unlock(&idev->mc_lock);
 }
 
 /* IPv6 device initialization. */
 
 void ipv6_mc_init_dev(struct inet6_dev *idev)
 {
-	write_lock_bh(&idev->lock);
-	spin_lock_init(&idev->mc_lock);
 	idev->mc_gq_running = 0;
-	timer_setup(&idev->mc_gq_timer, mld_gq_timer_expire, 0);
-	idev->mc_tomb = NULL;
+	INIT_DELAYED_WORK(&idev->mc_gq_work, mld_gq_work);
+	RCU_INIT_POINTER(idev->mc_tomb, NULL);
 	idev->mc_ifc_count = 0;
-	timer_setup(&idev->mc_ifc_timer, mld_ifc_timer_expire, 0);
-	timer_setup(&idev->mc_dad_timer, mld_dad_timer_expire, 0);
+	INIT_DELAYED_WORK(&idev->mc_ifc_work, mld_ifc_work);
+	INIT_DELAYED_WORK(&idev->mc_dad_work, mld_dad_work);
+	INIT_DELAYED_WORK(&idev->mc_query_work, mld_query_work);
+	INIT_DELAYED_WORK(&idev->mc_report_work, mld_report_work);
+	skb_queue_head_init(&idev->mc_query_queue);
+	skb_queue_head_init(&idev->mc_report_queue);
+	spin_lock_init(&idev->mc_query_lock);
+	spin_lock_init(&idev->mc_report_lock);
+	mutex_init(&idev->mc_lock);
 	ipv6_mc_reset(idev);
-	write_unlock_bh(&idev->lock);
 }
 
 /*
@@ -2596,9 +2791,13 @@ void ipv6_mc_destroy_dev(struct inet6_dev *idev)
 {
 	struct ifmcaddr6 *i;
 
-	/* Deactivate timers */
+	/* Deactivate works */
 	ipv6_mc_down(idev);
+	mutex_lock(&idev->mc_lock);
 	mld_clear_delrec(idev);
+	mutex_unlock(&idev->mc_lock);
+	mld_clear_query(idev);
+	mld_clear_report(idev);
 
 	/* Delete all-nodes address. */
 	/* We cannot call ipv6_dev_mc_dec() directly, our caller in
@@ -2610,16 +2809,14 @@ void ipv6_mc_destroy_dev(struct inet6_dev *idev)
 	if (idev->cnf.forwarding)
 		__ipv6_dev_mc_dec(idev, &in6addr_linklocal_allrouters);
 
-	write_lock_bh(&idev->lock);
-	while ((i = idev->mc_list) != NULL) {
-		idev->mc_list = i->next;
+	mutex_lock(&idev->mc_lock);
+	while ((i = mc_dereference(idev->mc_list, idev))) {
+		rcu_assign_pointer(idev->mc_list, mc_dereference(i->next, idev));
 
-		write_unlock_bh(&idev->lock);
 		ip6_mc_clear_src(i);
 		ma_put(i);
-		write_lock_bh(&idev->lock);
 	}
-	write_unlock_bh(&idev->lock);
+	mutex_unlock(&idev->mc_lock);
 }
 
 static void ipv6_mc_rejoin_groups(struct inet6_dev *idev)
@@ -2628,13 +2825,14 @@ static void ipv6_mc_rejoin_groups(struct inet6_dev *idev)
 
 	ASSERT_RTNL();
 
+	mutex_lock(&idev->mc_lock);
 	if (mld_in_v1_mode(idev)) {
-		read_lock_bh(&idev->lock);
-		for (pmc = idev->mc_list; pmc; pmc = pmc->next)
+		for_each_mc_mclock(idev, pmc)
 			igmp6_join_group(pmc);
-		read_unlock_bh(&idev->lock);
-	} else
+	} else {
 		mld_send_report(idev, NULL);
+	}
+	mutex_unlock(&idev->mc_lock);
 }
 
 static int ipv6_mc_netdev_event(struct notifier_block *this,
@@ -2681,13 +2879,12 @@ static inline struct ifmcaddr6 *igmp6_mc_get_first(struct seq_file *seq)
 		idev = __in6_dev_get(state->dev);
 		if (!idev)
 			continue;
-		read_lock_bh(&idev->lock);
-		im = idev->mc_list;
+
+		im = rcu_dereference(idev->mc_list);
 		if (im) {
 			state->idev = idev;
 			break;
 		}
-		read_unlock_bh(&idev->lock);
 	}
 	return im;
 }
@@ -2696,11 +2893,8 @@ static struct ifmcaddr6 *igmp6_mc_get_next(struct seq_file *seq, struct ifmcaddr
 {
 	struct igmp6_mc_iter_state *state = igmp6_mc_seq_private(seq);
 
-	im = im->next;
+	im = rcu_dereference(im->next);
 	while (!im) {
-		if (likely(state->idev))
-			read_unlock_bh(&state->idev->lock);
-
 		state->dev = next_net_device_rcu(state->dev);
 		if (!state->dev) {
 			state->idev = NULL;
@@ -2709,8 +2903,7 @@ static struct ifmcaddr6 *igmp6_mc_get_next(struct seq_file *seq, struct ifmcaddr
 		state->idev = __in6_dev_get(state->dev);
 		if (!state->idev)
 			continue;
-		read_lock_bh(&state->idev->lock);
-		im = state->idev->mc_list;
+		im = rcu_dereference(state->idev->mc_list);
 	}
 	return im;
 }
@@ -2744,10 +2937,8 @@ static void igmp6_mc_seq_stop(struct seq_file *seq, void *v)
 {
 	struct igmp6_mc_iter_state *state = igmp6_mc_seq_private(seq);
 
-	if (likely(state->idev)) {
-		read_unlock_bh(&state->idev->lock);
+	if (likely(state->idev))
 		state->idev = NULL;
-	}
 	state->dev = NULL;
 	rcu_read_unlock();
 }
@@ -2762,8 +2953,8 @@ static int igmp6_mc_seq_show(struct seq_file *seq, void *v)
 		   state->dev->ifindex, state->dev->name,
 		   &im->mca_addr,
 		   im->mca_users, im->mca_flags,
-		   (im->mca_flags&MAF_TIMER_RUNNING) ?
-		   jiffies_to_clock_t(im->mca_timer.expires-jiffies) : 0);
+		   (im->mca_flags & MAF_TIMER_RUNNING) ?
+		   jiffies_to_clock_t(im->mca_work.timer.expires - jiffies) : 0);
 	return 0;
 }
 
@@ -2797,19 +2988,16 @@ static inline struct ip6_sf_list *igmp6_mcf_get_first(struct seq_file *seq)
 		idev = __in6_dev_get(state->dev);
 		if (unlikely(idev == NULL))
 			continue;
-		read_lock_bh(&idev->lock);
-		im = idev->mc_list;
+
+		im = rcu_dereference(idev->mc_list);
 		if (likely(im)) {
-			spin_lock_bh(&im->mca_lock);
-			psf = im->mca_sources;
+			psf = rcu_dereference(im->mca_sources);
 			if (likely(psf)) {
 				state->im = im;
 				state->idev = idev;
 				break;
 			}
-			spin_unlock_bh(&im->mca_lock);
 		}
-		read_unlock_bh(&idev->lock);
 	}
 	return psf;
 }
@@ -2818,14 +3006,10 @@ static struct ip6_sf_list *igmp6_mcf_get_next(struct seq_file *seq, struct ip6_s
 {
 	struct igmp6_mcf_iter_state *state = igmp6_mcf_seq_private(seq);
 
-	psf = psf->sf_next;
+	psf = rcu_dereference(psf->sf_next);
 	while (!psf) {
-		spin_unlock_bh(&state->im->mca_lock);
-		state->im = state->im->next;
+		state->im = rcu_dereference(state->im->next);
 		while (!state->im) {
-			if (likely(state->idev))
-				read_unlock_bh(&state->idev->lock);
-
 			state->dev = next_net_device_rcu(state->dev);
 			if (!state->dev) {
 				state->idev = NULL;
@@ -2834,13 +3018,11 @@ static struct ip6_sf_list *igmp6_mcf_get_next(struct seq_file *seq, struct ip6_s
 			state->idev = __in6_dev_get(state->dev);
 			if (!state->idev)
 				continue;
-			read_lock_bh(&state->idev->lock);
-			state->im = state->idev->mc_list;
+			state->im = rcu_dereference(state->idev->mc_list);
 		}
 		if (!state->im)
 			break;
-		spin_lock_bh(&state->im->mca_lock);
-		psf = state->im->mca_sources;
+		psf = rcu_dereference(state->im->mca_sources);
 	}
 out:
 	return psf;
@@ -2877,14 +3059,12 @@ static void igmp6_mcf_seq_stop(struct seq_file *seq, void *v)
 	__releases(RCU)
 {
 	struct igmp6_mcf_iter_state *state = igmp6_mcf_seq_private(seq);
-	if (likely(state->im)) {
-		spin_unlock_bh(&state->im->mca_lock);
+
+	if (likely(state->im))
 		state->im = NULL;
-	}
-	if (likely(state->idev)) {
-		read_unlock_bh(&state->idev->lock);
+	if (likely(state->idev))
 		state->idev = NULL;
-	}
+
 	state->dev = NULL;
 	rcu_read_unlock();
 }
@@ -2965,6 +3145,7 @@ static int __net_init igmp6_net_init(struct net *net)
 	}
 
 	inet6_sk(net->ipv6.igmp_sk)->hop_limit = 1;
+	net->ipv6.igmp_sk->sk_allocation = GFP_KERNEL;
 
 	err = inet_ctl_sock_create(&net->ipv6.mc_autojoin_sk, PF_INET6,
 				   SOCK_RAW, IPPROTO_ICMPV6, net);
@@ -3002,7 +3183,19 @@ static struct pernet_operations igmp6_net_ops = {
 
 int __init igmp6_init(void)
 {
-	return register_pernet_subsys(&igmp6_net_ops);
+	int err;
+
+	err = register_pernet_subsys(&igmp6_net_ops);
+	if (err)
+		return err;
+
+	mld_wq = create_workqueue("mld");
+	if (!mld_wq) {
+		unregister_pernet_subsys(&igmp6_net_ops);
+		return -ENOMEM;
+	}
+
+	return err;
 }
 
 int __init igmp6_late_init(void)
@@ -3013,6 +3206,7 @@ int __init igmp6_late_init(void)
 void igmp6_cleanup(void)
 {
 	unregister_pernet_subsys(&igmp6_net_ops);
+	destroy_workqueue(mld_wq);
 }
 
 void igmp6_late_cleanup(void)