From: Kuniyuki Iwashima <kuniyu@google.com>
Date: Thu, 4 Jun 2026 22:46:32 +0000 (+0000)
Subject: ip6mr: Replace RTNL with a dedicated mutex for MFC.
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=a4c98421720a79587ad6c528f81fe7976a45976b;p=thirdparty%2Flinux.git

ip6mr: Replace RTNL with a dedicated mutex for MFC.

ip6mr does not have rtnetlink interface for MFC unlike ipmr,
which uses dev_get_by_index_rcu() to set struct mfcctl.mfcc_parent.

ip6mr_mfc_add() and ip6mr_mfc_delete() are called under RTNL
from ip6_mroute_setsockopt() only.

There are no RTNL dependant, but ip6_mroute_setsockopt() reuses
RTNL just for mrt->mfc_hash and mrt->mfc_cache_list.

Let's replace RTNL with a new per-netns mutex.

Later, ip6mr_notifier_ops and ipmr_seq will be moved under
CONFIG_IPV6_MROUTE.

Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20260604224712.3209821-15-kuniyu@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---

diff --git a/include/net/netns/ipv6.h b/include/net/netns/ipv6.h
index 875916d60bfe4..668f10498c5c7 100644
--- a/include/net/netns/ipv6.h
+++ b/include/net/netns/ipv6.h
@@ -112,6 +112,7 @@ struct netns_ipv6 {
 	struct list_head	mr6_tables;
 	struct fib_rules_ops	*mr6_rules_ops;
 #endif
+	struct mutex		mfc_mutex;
 #endif
 	atomic_t		dev_addr_genid;
 	atomic_t		fib6_sernum;
diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
index b13ce9c2c463b..5a4816225fb1b 100644
--- a/net/ipv6/ip6mr.c
+++ b/net/ipv6/ip6mr.c
@@ -1266,7 +1266,6 @@ static int ip6mr_mfc_delete(struct mr_table *mrt, struct mf6cctl *mfc,
 {
 	struct mfc6_cache *c;
 
-	/* The entries are added/deleted only under RTNL */
 	rcu_read_lock();
 	c = ip6mr_cache_find_parent(mrt, &mfc->mf6cc_origin.sin6_addr,
 				    &mfc->mf6cc_mcastgrp.sin6_addr, parent);
@@ -1358,6 +1357,8 @@ static int __net_init ip6mr_net_init(struct net *net)
 #endif
 	int err;
 
+	mutex_init(&net->ipv6.mfc_mutex);
+
 	err = ip6mr_notifier_init(net);
 	if (err)
 		return err;
@@ -1486,7 +1487,6 @@ static int ip6mr_mfc_add(struct net *net, struct mr_table *mrt,
 			ttls[i] = 1;
 	}
 
-	/* The entries are added/deleted only under RTNL */
 	rcu_read_lock();
 	c = ip6mr_cache_find_parent(mrt, &mfc->mf6cc_origin.sin6_addr,
 				    &mfc->mf6cc_mcastgrp.sin6_addr, parent);
@@ -1581,6 +1581,8 @@ static void mroute_clean_tables(struct mr_table *mrt, int flags,
 
 	/* Wipe the cache */
 	if (flags & (MRT6_FLUSH_MFC | MRT6_FLUSH_MFC_STATIC)) {
+		mutex_lock(&net->ipv6.mfc_mutex);
+
 		list_for_each_entry_safe(c, tmp, &mrt->mfc_cache_list, list) {
 			if (((c->mfc_flags & MFC_STATIC) && !(flags & MRT6_FLUSH_MFC_STATIC)) ||
 			    (!(c->mfc_flags & MFC_STATIC) && !(flags & MRT6_FLUSH_MFC)))
@@ -1592,6 +1594,8 @@ static void mroute_clean_tables(struct mr_table *mrt, int flags,
 			mr6_netlink_event(mrt, (struct mfc6_cache *)c, RTM_DELROUTE);
 			mr_cache_put(c);
 		}
+
+		mutex_unlock(&net->ipv6.mfc_mutex);
 	}
 
 	if (flags & MRT6_FLUSH_MFC) {
@@ -1775,15 +1779,18 @@ int ip6_mroute_setsockopt(struct sock *sk, int optname, sockptr_t optval,
 			return -EFAULT;
 		if (parent == 0)
 			parent = mfc.mf6cc_parent;
-		rtnl_lock();
+
+		mutex_lock(&net->ipv6.mfc_mutex);
+
 		if (optname == MRT6_DEL_MFC || optname == MRT6_DEL_MFC_PROXY)
 			ret = ip6mr_mfc_delete(mrt, &mfc, parent);
 		else
 			ret = ip6mr_mfc_add(net, mrt, &mfc,
 					    sk ==
-					    rtnl_dereference(mrt->mroute_sk),
+					    rcu_access_pointer(mrt->mroute_sk),
 					    parent);
-		rtnl_unlock();
+
+		mutex_unlock(&net->ipv6.mfc_mutex);
 		return ret;
 
 	case MRT6_FLUSH: