public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: Kuniyuki Iwashima <kuniyu@google.com>
To: "David S . Miller" <davem@davemloft.net>,
	David Ahern <dsahern@kernel.org>,
	 Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Cc: Simon Horman <horms@kernel.org>,
	Kuniyuki Iwashima <kuniyu@google.com>,
	 Kuniyuki Iwashima <kuni1840@gmail.com>,
	netdev@vger.kernel.org
Subject: [PATCH v3 net-next 15/15] ip6mr: Replace RTNL with a dedicated mutex for MFC.
Date: Tue,  7 Apr 2026 21:19:48 +0000	[thread overview]
Message-ID: <20260407212001.2368593-16-kuniyu@google.com> (raw)
In-Reply-To: <20260407212001.2368593-1-kuniyu@google.com>

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.

Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
---
 include/net/netns/ipv6.h |  1 +
 net/ipv6/ip6mr.c         | 21 ++++++++++++++-------
 2 files changed, 15 insertions(+), 7 deletions(-)

diff --git a/include/net/netns/ipv6.h b/include/net/netns/ipv6.h
index df00567374f4..6453d70d5946 100644
--- a/include/net/netns/ipv6.h
+++ b/include/net/netns/ipv6.h
@@ -114,6 +114,7 @@ struct netns_ipv6 {
 #endif
 	struct fib_notifier_ops	*ip6mr_notifier_ops;
 	atomic_t		ipmr_seq;
+	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 b1443fb65b40..e4c31d05744b 100644
--- a/net/ipv6/ip6mr.c
+++ b/net/ipv6/ip6mr.c
@@ -1256,7 +1256,6 @@ 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);
@@ -1346,6 +1345,8 @@ static int __net_init ip6mr_net_init(struct net *net)
 	LIST_HEAD(dev_kill_list);
 	int err;
 
+	mutex_init(&net->ipv6.mfc_mutex);
+
 	err = ip6mr_notifier_init(net);
 	if (err)
 		return err;
@@ -1474,7 +1475,6 @@ 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);
@@ -1553,6 +1553,7 @@ static int ip6mr_mfc_add(struct net *net, struct mr_table *mrt,
 static void mroute_clean_tables(struct mr_table *mrt, int flags,
 				struct list_head *dev_kill_list)
 {
+	struct net *net = read_pnet(&mrt->net);
 	struct mr_mfc *c, *tmp;
 	int i;
 
@@ -1569,18 +1570,21 @@ 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)))
 				continue;
 			rhltable_remove(&mrt->mfc_hash, &c->mnode, ip6mr_rht_params);
 			list_del_rcu(&c->list);
-			call_ip6mr_mfc_entry_notifiers(read_pnet(&mrt->net),
-						       FIB_EVENT_ENTRY_DEL,
+			call_ip6mr_mfc_entry_notifiers(net, FIB_EVENT_ENTRY_DEL,
 						       (struct mfc6_cache *)c, mrt->id);
 			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) {
@@ -1763,15 +1767,18 @@ 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:
-- 
2.53.0.1213.gd9a14994de-goog


  parent reply	other threads:[~2026-04-07 21:20 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-07 21:19 [PATCH v3 net-next 00/15] ip6mr: No RTNL for RTNL_FAMILY_IP6MR rtnetlink Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 01/15] selftest: net: Extend ipmr.c for IP6MR Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 02/15] ipmr: Convert mr_table.cache_resolve_queue_len to u32 Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 03/15] ip6mr: Annotate access to mrt->mroute_do_{pim,assert,wrvifwhole} Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 04/15] ip6mr: Use MAXMIFS in mr6_msgsize() Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 05/15] ip6mr: Allocate skb earlier in ip6mr_rtm_getroute() Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 06/15] ip6mr: Convert ip6mr_rtm_getroute() to RCU Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 07/15] ip6mr: Convert ip6mr_rtm_dumproute() " Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 08/15] net: Remove rtnl_held of struct fib_dump_filter Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 09/15] ip6mr: Move unregister_netdevice_many() out of mroute_clean_tables() Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 10/15] ip6mr: Move unregister_netdevice_many() out of ip6mr_free_table() Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 11/15] ip6mr: Convert ip6mr_net_exit_batch() to ->exit_rtnl() Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 12/15] ip6mr: Remove RTNL in ip6mr_rules_init() and ip6mr_net_init() Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 13/15] ip6mr: Call fib_rules_unregister() without RTNL Kuniyuki Iwashima
2026-04-07 21:19 ` [PATCH v3 net-next 14/15] ip6mr: Define net->ipv6.{ip6mr_notifier_ops,ipmr_seq} under CONFIG_IP_MROUTE Kuniyuki Iwashima
2026-04-07 21:19 ` Kuniyuki Iwashima [this message]
2026-04-08  1:42 ` [PATCH v3 net-next 00/15] ip6mr: No RTNL for RTNL_FAMILY_IP6MR rtnetlink Jakub Kicinski
2026-04-08  3:48   ` Kuniyuki Iwashima
2026-04-08  3:50     ` Kuniyuki Iwashima

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260407212001.2368593-16-kuniyu@google.com \
    --to=kuniyu@google.com \
    --cc=davem@davemloft.net \
    --cc=dsahern@kernel.org \
    --cc=edumazet@google.com \
    --cc=horms@kernel.org \
    --cc=kuba@kernel.org \
    --cc=kuni1840@gmail.com \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox