[PATCH 5.15 05/77] netlink: Fix wraparounds of sk->sk_rmem_alloc.

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

5.15-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Kuniyuki Iwashima <kuniyu@google.com>

[ Upstream commit ae8f160e7eb24240a2a79fc4c815c6a0d4ee16cc ]

Netlink has this pattern in some places

  if (atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf)
  	atomic_add(skb->truesize, &sk->sk_rmem_alloc);

, which has the same problem fixed by commit 5a465a0da13e ("udp:
Fix multiple wraparounds of sk->sk_rmem_alloc.").

For example, if we set INT_MAX to SO_RCVBUFFORCE, the condition
is always false as the two operands are of int.

Then, a single socket can eat as many skb as possible until OOM
happens, and we can see multiple wraparounds of sk->sk_rmem_alloc.

Let's fix it by using atomic_add_return() and comparing the two
variables as unsigned int.

Before:
  [root@fedora ~]# ss -f netlink
  Recv-Q      Send-Q Local Address:Port                Peer Address:Port
  -1668710080 0               rtnl:nl_wraparound/293               *

After:
  [root@fedora ~]# ss -f netlink
  Recv-Q     Send-Q Local Address:Port                Peer Address:Port
  2147483072 0               rtnl:nl_wraparound/290               *
  ^
  `--- INT_MAX - 576

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-by: Jason Baron <jbaron@akamai.com>
Closes: https://lore.kernel.org/netdev/cover.1750285100.git.jbaron@akamai.com/
Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20250704054824.1580222-1-kuniyu@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 net/netlink/af_netlink.c | 81 ++++++++++++++++++++++++----------------
 1 file changed, 49 insertions(+), 32 deletions(-)

diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
index d3852526ef52e..ae83e56478998 100644
--- a/net/netlink/af_netlink.c
+++ b/net/netlink/af_netlink.c
@@ -387,7 +387,6 @@ static void netlink_skb_set_owner_r(struct sk_buff *skb, struct sock *sk)
 	WARN_ON(skb->sk != NULL);
 	skb->sk = sk;
 	skb->destructor = netlink_skb_destructor;
-	atomic_add(skb->truesize, &sk->sk_rmem_alloc);
 	sk_mem_charge(sk, skb->truesize);
 }
 
@@ -1211,41 +1210,48 @@ static struct sk_buff *netlink_alloc_large_skb(unsigned int size,
 int netlink_attachskb(struct sock *sk, struct sk_buff *skb,
 		      long *timeo, struct sock *ssk)
 {
+	DECLARE_WAITQUEUE(wait, current);
 	struct netlink_sock *nlk;
+	unsigned int rmem;
 
 	nlk = nlk_sk(sk);
+	rmem = atomic_add_return(skb->truesize, &sk->sk_rmem_alloc);
 
-	if ((atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf ||
-	     test_bit(NETLINK_S_CONGESTED, &nlk->state))) {
-		DECLARE_WAITQUEUE(wait, current);
-		if (!*timeo) {
-			if (!ssk || netlink_is_kernel(ssk))
-				netlink_overrun(sk);
-			sock_put(sk);
-			kfree_skb(skb);
-			return -EAGAIN;
-		}
-
-		__set_current_state(TASK_INTERRUPTIBLE);
-		add_wait_queue(&nlk->wait, &wait);
+	if ((rmem == skb->truesize || rmem < READ_ONCE(sk->sk_rcvbuf)) &&
+	    !test_bit(NETLINK_S_CONGESTED, &nlk->state)) {
+		netlink_skb_set_owner_r(skb, sk);
+		return 0;
+	}
 
-		if ((atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf ||
-		     test_bit(NETLINK_S_CONGESTED, &nlk->state)) &&
-		    !sock_flag(sk, SOCK_DEAD))
-			*timeo = schedule_timeout(*timeo);
+	atomic_sub(skb->truesize, &sk->sk_rmem_alloc);
 
-		__set_current_state(TASK_RUNNING);
-		remove_wait_queue(&nlk->wait, &wait);
+	if (!*timeo) {
+		if (!ssk || netlink_is_kernel(ssk))
+			netlink_overrun(sk);
 		sock_put(sk);
+		kfree_skb(skb);
+		return -EAGAIN;
+	}
 
-		if (signal_pending(current)) {
-			kfree_skb(skb);
-			return sock_intr_errno(*timeo);
-		}
-		return 1;
+	__set_current_state(TASK_INTERRUPTIBLE);
+	add_wait_queue(&nlk->wait, &wait);
+	rmem = atomic_read(&sk->sk_rmem_alloc);
+
+	if (((rmem && rmem + skb->truesize > READ_ONCE(sk->sk_rcvbuf)) ||
+	     test_bit(NETLINK_S_CONGESTED, &nlk->state)) &&
+	    !sock_flag(sk, SOCK_DEAD))
+		*timeo = schedule_timeout(*timeo);
+
+	__set_current_state(TASK_RUNNING);
+	remove_wait_queue(&nlk->wait, &wait);
+	sock_put(sk);
+
+	if (signal_pending(current)) {
+		kfree_skb(skb);
+		return sock_intr_errno(*timeo);
 	}
-	netlink_skb_set_owner_r(skb, sk);
-	return 0;
+
+	return 1;
 }
 
 static int __netlink_sendskb(struct sock *sk, struct sk_buff *skb)
@@ -1305,6 +1311,7 @@ static int netlink_unicast_kernel(struct sock *sk, struct sk_buff *skb,
 	ret = -ECONNREFUSED;
 	if (nlk->netlink_rcv != NULL) {
 		ret = skb->len;
+		atomic_add(skb->truesize, &sk->sk_rmem_alloc);
 		netlink_skb_set_owner_r(skb, sk);
 		NETLINK_CB(skb).sk = ssk;
 		netlink_deliver_tap_kernel(sk, ssk, skb);
@@ -1383,13 +1390,19 @@ EXPORT_SYMBOL_GPL(netlink_strict_get_check);
 static int netlink_broadcast_deliver(struct sock *sk, struct sk_buff *skb)
 {
 	struct netlink_sock *nlk = nlk_sk(sk);
+	unsigned int rmem, rcvbuf;
 
-	if (atomic_read(&sk->sk_rmem_alloc) <= sk->sk_rcvbuf &&
+	rmem = atomic_add_return(skb->truesize, &sk->sk_rmem_alloc);
+	rcvbuf = READ_ONCE(sk->sk_rcvbuf);
+
+	if ((rmem != skb->truesize || rmem <= rcvbuf) &&
 	    !test_bit(NETLINK_S_CONGESTED, &nlk->state)) {
 		netlink_skb_set_owner_r(skb, sk);
 		__netlink_sendskb(sk, skb);
-		return atomic_read(&sk->sk_rmem_alloc) > (sk->sk_rcvbuf >> 1);
+		return rmem > (rcvbuf >> 1);
 	}
+
+	atomic_sub(skb->truesize, &sk->sk_rmem_alloc);
 	return -1;
 }
 
@@ -2206,6 +2219,7 @@ static int netlink_dump(struct sock *sk, bool lock_taken)
 	struct module *module;
 	int err = -ENOBUFS;
 	int alloc_min_size;
+	unsigned int rmem;
 	int alloc_size;
 
 	if (!lock_taken)
@@ -2215,9 +2229,6 @@ static int netlink_dump(struct sock *sk, bool lock_taken)
 		goto errout_skb;
 	}
 
-	if (atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf)
-		goto errout_skb;
-
 	/* NLMSG_GOODSIZE is small to avoid high order allocations being
 	 * required, but it makes sense to _attempt_ a 16K bytes allocation
 	 * to reduce number of system calls on dump operations, if user
@@ -2240,6 +2251,12 @@ static int netlink_dump(struct sock *sk, bool lock_taken)
 	if (!skb)
 		goto errout_skb;
 
+	rmem = atomic_add_return(skb->truesize, &sk->sk_rmem_alloc);
+	if (rmem >= READ_ONCE(sk->sk_rcvbuf)) {
+		atomic_sub(skb->truesize, &sk->sk_rmem_alloc);
+		goto errout_skb;
+	}
+
 	/* Trim skb to allocated size. User is expected to provide buffer as
 	 * large as max(min_dump_alloc, 16KiB (mac_recvmsg_len capped at
 	 * netlink_recvmsg())). dump will pack as many smaller messages as
-- 
2.39.5