[PATCH AUTOSEL 4.14 02/27] netfilter: ipset: list:set: Decrease refcount synchronously on deletion and replace

[PATCH AUTOSEL 4.14 02/27] netfilter: ipset: list:set: Decrease refcount synchronously on deletion and replace

[Sasha Levin]

From: Stefano Brivio <sbrivio@redhat.com>

[ Upstream commit 439cd39ea136d2c026805264d58a91f36b6b64ca ]

Commit 45040978c899 ("netfilter: ipset: Fix set:list type crash
when flush/dump set in parallel") postponed decreasing set
reference counters to the RCU callback.

An 'ipset del' command can terminate before the RCU grace period
is elapsed, and if sets are listed before then, the reference
counter shown in userspace will be wrong:

 # ipset create h hash:ip; ipset create l list:set; ipset add l
 # ipset del l h; ipset list h
 Name: h
 Type: hash:ip
 Revision: 4
 Header: family inet hashsize 1024 maxelem 65536
 Size in memory: 88
 References: 1
 Number of entries: 0
 Members:
 # sleep 1; ipset list h
 Name: h
 Type: hash:ip
 Revision: 4
 Header: family inet hashsize 1024 maxelem 65536
 Size in memory: 88
 References: 0
 Number of entries: 0
 Members:

Fix this by making the reference count update synchronous again.

As a result, when sets are listed, ip_set_name_byindex() might
now fetch a set whose reference count is already zero. Instead
of relying on the reference count to protect against concurrent
set renaming, grab ip_set_ref_lock as reader and copy the name,
while holding the same lock in ip_set_rename() as writer
instead.

Reported-by: Li Shuang <shuali@redhat.com>
Fixes: 45040978c899 ("netfilter: ipset: Fix set:list type crash when flush/dump set in parallel")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 include/linux/netfilter/ipset/ip_set.h |  2 +-
 net/netfilter/ipset/ip_set_core.c      | 23 +++++++++++------------
 net/netfilter/ipset/ip_set_list_set.c  | 17 +++++++++++------
 3 files changed, 23 insertions(+), 19 deletions(-)

diff --git a/include/linux/netfilter/ipset/ip_set.h b/include/linux/netfilter/ipset/ip_set.h
index 8e42253e5d4d..91a533bd3eb1 100644
--- a/include/linux/netfilter/ipset/ip_set.h
+++ b/include/linux/netfilter/ipset/ip_set.h
@@ -312,7 +312,7 @@ enum {
 extern ip_set_id_t ip_set_get_byname(struct net *net,
 				     const char *name, struct ip_set **set);
 extern void ip_set_put_byindex(struct net *net, ip_set_id_t index);
-extern const char *ip_set_name_byindex(struct net *net, ip_set_id_t index);
+extern void ip_set_name_byindex(struct net *net, ip_set_id_t index, char *name);
 extern ip_set_id_t ip_set_nfnl_get_byindex(struct net *net, ip_set_id_t index);
 extern void ip_set_nfnl_put(struct net *net, ip_set_id_t index);
 
diff --git a/net/netfilter/ipset/ip_set_core.c b/net/netfilter/ipset/ip_set_core.c
index 9d2ce1459cec..a3f1dc7cf538 100644
--- a/net/netfilter/ipset/ip_set_core.c
+++ b/net/netfilter/ipset/ip_set_core.c
@@ -668,21 +668,20 @@ ip_set_put_byindex(struct net *net, ip_set_id_t index)
 EXPORT_SYMBOL_GPL(ip_set_put_byindex);
 
 /* Get the name of a set behind a set index.
- * We assume the set is referenced, so it does exist and
- * can't be destroyed. The set cannot be renamed due to
- * the referencing either.
- *
+ * Set itself is protected by RCU, but its name isn't: to protect against
+ * renaming, grab ip_set_ref_lock as reader (see ip_set_rename()) and copy the
+ * name.
  */
-const char *
-ip_set_name_byindex(struct net *net, ip_set_id_t index)
+void
+ip_set_name_byindex(struct net *net, ip_set_id_t index, char *name)
 {
-	const struct ip_set *set = ip_set_rcu_get(net, index);
+	struct ip_set *set = ip_set_rcu_get(net, index);
 
 	BUG_ON(!set);
-	BUG_ON(set->ref == 0);
 
-	/* Referenced, so it's safe */
-	return set->name;
+	read_lock_bh(&ip_set_ref_lock);
+	strncpy(name, set->name, IPSET_MAXNAMELEN);
+	read_unlock_bh(&ip_set_ref_lock);
 }
 EXPORT_SYMBOL_GPL(ip_set_name_byindex);
 
@@ -1128,7 +1127,7 @@ static int ip_set_rename(struct net *net, struct sock *ctnl,
 	if (!set)
 		return -ENOENT;
 
-	read_lock_bh(&ip_set_ref_lock);
+	write_lock_bh(&ip_set_ref_lock);
 	if (set->ref != 0) {
 		ret = -IPSET_ERR_REFERENCED;
 		goto out;
@@ -1145,7 +1144,7 @@ static int ip_set_rename(struct net *net, struct sock *ctnl,
 	strncpy(set->name, name2, IPSET_MAXNAMELEN);
 
 out:
-	read_unlock_bh(&ip_set_ref_lock);
+	write_unlock_bh(&ip_set_ref_lock);
 	return ret;
 }
 
diff --git a/net/netfilter/ipset/ip_set_list_set.c b/net/netfilter/ipset/ip_set_list_set.c
index 178d4eba013b..75d52aed6fdb 100644
--- a/net/netfilter/ipset/ip_set_list_set.c
+++ b/net/netfilter/ipset/ip_set_list_set.c
@@ -156,9 +156,7 @@ __list_set_del_rcu(struct rcu_head * rcu)
 {
 	struct set_elem *e = container_of(rcu, struct set_elem, rcu);
 	struct ip_set *set = e->set;
-	struct list_set *map = set->data;
 
-	ip_set_put_byindex(map->net, e->id);
 	ip_set_ext_destroy(set, e);
 	kfree(e);
 }
@@ -166,15 +164,21 @@ __list_set_del_rcu(struct rcu_head * rcu)
 static inline void
 list_set_del(struct ip_set *set, struct set_elem *e)
 {
+	struct list_set *map = set->data;
+
 	set->elements--;
 	list_del_rcu(&e->list);
+	ip_set_put_byindex(map->net, e->id);
 	call_rcu(&e->rcu, __list_set_del_rcu);
 }
 
 static inline void
-list_set_replace(struct set_elem *e, struct set_elem *old)
+list_set_replace(struct ip_set *set, struct set_elem *e, struct set_elem *old)
 {
+	struct list_set *map = set->data;
+
 	list_replace_rcu(&old->list, &e->list);
+	ip_set_put_byindex(map->net, old->id);
 	call_rcu(&old->rcu, __list_set_del_rcu);
 }
 
@@ -306,7 +310,7 @@ list_set_uadd(struct ip_set *set, void *value, const struct ip_set_ext *ext,
 	INIT_LIST_HEAD(&e->list);
 	list_set_init_extensions(set, ext, e);
 	if (n)
-		list_set_replace(e, n);
+		list_set_replace(set, e, n);
 	else if (next)
 		list_add_tail_rcu(&e->list, &next->list);
 	else if (prev)
@@ -497,6 +501,7 @@ list_set_list(const struct ip_set *set,
 	const struct list_set *map = set->data;
 	struct nlattr *atd, *nested;
 	u32 i = 0, first = cb->args[IPSET_CB_ARG0];
+	char name[IPSET_MAXNAMELEN];
 	struct set_elem *e;
 	int ret = 0;
 
@@ -515,8 +520,8 @@ list_set_list(const struct ip_set *set,
 		nested = ipset_nest_start(skb, IPSET_ATTR_DATA);
 		if (!nested)
 			goto nla_put_failure;
-		if (nla_put_string(skb, IPSET_ATTR_NAME,
-				   ip_set_name_byindex(map->net, e->id)))
+		ip_set_name_byindex(map->net, e->id, name);
+		if (nla_put_string(skb, IPSET_ATTR_NAME, name))
 			goto nla_put_failure;
 		if (ip_set_put_extensions(skb, set, e, true))
 			goto nla_put_failure;
-- 
2.17.1

[PATCH AUTOSEL 4.14 03/27] netfilter: ipset: actually allow allowable CIDR 0 in hash:net,port,net

[Sasha Levin]

From: Eric Westbrook <eric@westbrook.io>

[ Upstream commit 886503f34d63e681662057448819edb5b1057a97 ]

Allow /0 as advertised for hash:net,port,net sets.

For "hash:net,port,net", ipset(8) says that "either subnet
is permitted to be a /0 should you wish to match port
between all destinations."

Make that statement true.

Before:

    # ipset create cidrzero hash:net,port,net
    # ipset add cidrzero 0.0.0.0/0,12345,0.0.0.0/0
    ipset v6.34: The value of the CIDR parameter of the IP address is invalid

    # ipset create cidrzero6 hash:net,port,net family inet6
    # ipset add cidrzero6 ::/0,12345,::/0
    ipset v6.34: The value of the CIDR parameter of the IP address is invalid

After:

    # ipset create cidrzero hash:net,port,net
    # ipset add cidrzero 0.0.0.0/0,12345,0.0.0.0/0
    # ipset test cidrzero 192.168.205.129,12345,172.16.205.129
    192.168.205.129,tcp:12345,172.16.205.129 is in set cidrzero.

    # ipset create cidrzero6 hash:net,port,net family inet6
    # ipset add cidrzero6 ::/0,12345,::/0
    # ipset test cidrzero6 fe80::1,12345,ff00::1
    fe80::1,tcp:12345,ff00::1 is in set cidrzero6.

See also:

  https://bugzilla.kernel.org/show_bug.cgi?id=200897
  https://github.com/ewestbrook/linux/commit/df7ff6efb0934ab6acc11f003ff1a7580d6c1d9c

Signed-off-by: Eric Westbrook <linux@westbrook.io>
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 net/netfilter/ipset/ip_set_hash_netportnet.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/net/netfilter/ipset/ip_set_hash_netportnet.c b/net/netfilter/ipset/ip_set_hash_netportnet.c
index 8602f2595a1a..0e6e40c6f652 100644
--- a/net/netfilter/ipset/ip_set_hash_netportnet.c
+++ b/net/netfilter/ipset/ip_set_hash_netportnet.c
@@ -213,13 +213,13 @@ hash_netportnet4_uadt(struct ip_set *set, struct nlattr *tb[],
 
 	if (tb[IPSET_ATTR_CIDR]) {
 		e.cidr[0] = nla_get_u8(tb[IPSET_ATTR_CIDR]);
-		if (!e.cidr[0] || e.cidr[0] > HOST_MASK)
+		if (e.cidr[0] > HOST_MASK)
 			return -IPSET_ERR_INVALID_CIDR;
 	}
 
 	if (tb[IPSET_ATTR_CIDR2]) {
 		e.cidr[1] = nla_get_u8(tb[IPSET_ATTR_CIDR2]);
-		if (!e.cidr[1] || e.cidr[1] > HOST_MASK)
+		if (e.cidr[1] > HOST_MASK)
 			return -IPSET_ERR_INVALID_CIDR;
 	}
 
@@ -492,13 +492,13 @@ hash_netportnet6_uadt(struct ip_set *set, struct nlattr *tb[],
 
 	if (tb[IPSET_ATTR_CIDR]) {
 		e.cidr[0] = nla_get_u8(tb[IPSET_ATTR_CIDR]);
-		if (!e.cidr[0] || e.cidr[0] > HOST_MASK)
+		if (e.cidr[0] > HOST_MASK)
 			return -IPSET_ERR_INVALID_CIDR;
 	}
 
 	if (tb[IPSET_ATTR_CIDR2]) {
 		e.cidr[1] = nla_get_u8(tb[IPSET_ATTR_CIDR2]);
-		if (!e.cidr[1] || e.cidr[1] > HOST_MASK)
+		if (e.cidr[1] > HOST_MASK)
 			return -IPSET_ERR_INVALID_CIDR;
 	}
 
-- 
2.17.1

[PATCH AUTOSEL 4.14 08/27] netfilter: ipset: Correct rcu_dereference() call in ip_set_put_comment()

[Sasha Levin]

From: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>

[ Upstream commit 17b8b74c0f8dbf9b9e3301f9ca5b65dd1c079951 ]

The function is called when rcu_read_lock() is held and not
when rcu_read_lock_bh() is held.

Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 include/linux/netfilter/ipset/ip_set_comment.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/linux/netfilter/ipset/ip_set_comment.h b/include/linux/netfilter/ipset/ip_set_comment.h
index 8e2bab1e8e90..70877f8de7e9 100644
--- a/include/linux/netfilter/ipset/ip_set_comment.h
+++ b/include/linux/netfilter/ipset/ip_set_comment.h
@@ -43,11 +43,11 @@ ip_set_init_comment(struct ip_set *set, struct ip_set_comment *comment,
 	rcu_assign_pointer(comment->c, c);
 }
 
-/* Used only when dumping a set, protected by rcu_read_lock_bh() */
+/* Used only when dumping a set, protected by rcu_read_lock() */
 static inline int
 ip_set_put_comment(struct sk_buff *skb, const struct ip_set_comment *comment)
 {
-	struct ip_set_comment_rcu *c = rcu_dereference_bh(comment->c);
+	struct ip_set_comment_rcu *c = rcu_dereference(comment->c);
 
 	if (!c)
 		return 0;
-- 
2.17.1

[PATCH AUTOSEL 4.14 09/27] netfilter: xt_IDLETIMER: add sysfs filename checking routine

[Sasha Levin]

From: Taehee Yoo <ap420073@gmail.com>

[ Upstream commit 54451f60c8fa061af9051a53be9786393947367c ]

When IDLETIMER rule is added, sysfs file is created under
/sys/class/xt_idletimer/timers/
But some label name shouldn't be used.
".", "..", "power", "uevent", "subsystem", etc...
So that sysfs filename checking routine is needed.

test commands:
   %iptables -I INPUT -j IDLETIMER --timeout 1 --label "power"

splat looks like:
[95765.423132] sysfs: cannot create duplicate filename '/devices/virtual/xt_idletimer/timers/power'
[95765.433418] CPU: 0 PID: 8446 Comm: iptables Not tainted 4.19.0-rc6+ #20
[95765.449755] Call Trace:
[95765.449755]  dump_stack+0xc9/0x16b
[95765.449755]  ? show_regs_print_info+0x5/0x5
[95765.449755]  sysfs_warn_dup+0x74/0x90
[95765.449755]  sysfs_add_file_mode_ns+0x352/0x500
[95765.449755]  sysfs_create_file_ns+0x179/0x270
[95765.449755]  ? sysfs_add_file_mode_ns+0x500/0x500
[95765.449755]  ? idletimer_tg_checkentry+0x3e5/0xb1b [xt_IDLETIMER]
[95765.449755]  ? rcu_read_lock_sched_held+0x114/0x130
[95765.449755]  ? __kmalloc_track_caller+0x211/0x2b0
[95765.449755]  ? memcpy+0x34/0x50
[95765.449755]  idletimer_tg_checkentry+0x4e2/0xb1b [xt_IDLETIMER]
[ ... ]

Fixes: 0902b469bd25 ("netfilter: xtables: idletimer target implementation")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 net/netfilter/xt_IDLETIMER.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/net/netfilter/xt_IDLETIMER.c b/net/netfilter/xt_IDLETIMER.c
index 1141f08810b6..3fef8c2e545d 100644
--- a/net/netfilter/xt_IDLETIMER.c
+++ b/net/netfilter/xt_IDLETIMER.c
@@ -116,6 +116,22 @@ static void idletimer_tg_expired(unsigned long data)
 	schedule_work(&timer->work);
 }
 
+static int idletimer_check_sysfs_name(const char *name, unsigned int size)
+{
+	int ret;
+
+	ret = xt_check_proc_name(name, size);
+	if (ret < 0)
+		return ret;
+
+	if (!strcmp(name, "power") ||
+	    !strcmp(name, "subsystem") ||
+	    !strcmp(name, "uevent"))
+		return -EINVAL;
+
+	return 0;
+}
+
 static int idletimer_tg_create(struct idletimer_tg_info *info)
 {
 	int ret;
@@ -126,6 +142,10 @@ static int idletimer_tg_create(struct idletimer_tg_info *info)
 		goto out;
 	}
 
+	ret = idletimer_check_sysfs_name(info->label, sizeof(info->label));
+	if (ret < 0)
+		goto out_free_timer;
+
 	sysfs_attr_init(&info->timer->attr.attr);
 	info->timer->attr.attr.name = kstrdup(info->label, GFP_KERNEL);
 	if (!info->timer->attr.attr.name) {
-- 
2.17.1