Merge tag 'net-6.13-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

Pull networking fixes from Paolo Abeni:
 "Notably this includes fixes for a few regressions spotted very
  recently. No known outstanding ones.

  Current release - regressions:

   - core: avoid CFI problems with sock priv helpers

   - xsk: bring back busy polling support

   - netpoll: ensure skb_pool list is always initialized

  Current release - new code bugs:

   - core: make page_pool_ref_netmem work with net iovs

   - ipv4: route: fix drop reason being overridden in
     ip_route_input_slow

   - udp: make rehash4 independent in udp_lib_rehash()

  Previous releases - regressions:

   - bpf: fix bpf_sk_select_reuseport() memory leak

   - openvswitch: fix lockup on tx to unregistering netdev with carrier

   - mptcp: be sure to send ack when mptcp-level window re-opens

   - eth:
      - bnxt: always recalculate features after XDP clearing, fix
        null-deref
      - mlx5: fix sub-function add port error handling
      - fec: handle page_pool_dev_alloc_pages error

  Previous releases - always broken:

   - vsock: some fixes due to transport de-assignment

   - eth:
      - ice: fix E825 initialization
      - mlx5e: fix inversion dependency warning while enabling IPsec
        tunnel
      - gtp: destroy device along with udp socket's netns dismantle.
      - xilinx: axienet: Fix IRQ coalescing packet count overflow"

* tag 'net-6.13-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (44 commits)
  netdev: avoid CFI problems with sock priv helpers
  net/mlx5e: Always start IPsec sequence number from 1
  net/mlx5e: Rely on reqid in IPsec tunnel mode
  net/mlx5e: Fix inversion dependency warning while enabling IPsec tunnel
  net/mlx5: Clear port select structure when fail to create
  net/mlx5: SF, Fix add port error handling
  net/mlx5: Fix a lockdep warning as part of the write combining test
  net/mlx5: Fix RDMA TX steering prio
  net: make page_pool_ref_netmem work with net iovs
  net: ethernet: xgbe: re-add aneg to supported features in PHY quirks
  net: pcs: xpcs: actively unset DW_VR_MII_DIG_CTRL1_2G5_EN for 1G SGMII
  net: pcs: xpcs: fix DW_VR_MII_DIG_CTRL1_2G5_EN bit being set for 1G SGMII w/o inband
  selftests: net: Adapt ethtool mq tests to fix in qdisc graft
  net: fec: handle page_pool_dev_alloc_pages error
  net: netpoll: ensure skb_pool list is always initialized
  net: xilinx: axienet: Fix IRQ coalescing packet count overflow
  nfp: bpf: prevent integer overflow in nfp_bpf_event_output()
  selftests: mptcp: avoid spurious errors on disconnect
  mptcp: fix spurious wake-up on under memory pressure
  mptcp: be sure to send ack when mptcp-level window re-opens
  ...

1 files changed, 27 insertions, 19 deletions

diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index e8953e88efef..86d282618515 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -533,7 +533,7 @@ begin:
 	return NULL;
 }
 
-/* In hash4, rehash can happen in connect(), where hash4_cnt keeps unchanged. */
+/* udp_rehash4() only checks hslot4, and hash4_cnt is not processed. */
 static void udp_rehash4(struct udp_table *udptable, struct sock *sk,
 			u16 newhash4)
 {
@@ -582,15 +582,13 @@ void udp_lib_hash4(struct sock *sk, u16 hash)
 	struct net *net = sock_net(sk);
 	struct udp_table *udptable;
 
-	/* Connected udp socket can re-connect to another remote address,
-	 * so rehash4 is needed.
+	/* Connected udp socket can re-connect to another remote address, which
+	 * will be handled by rehash. Thus no need to redo hash4 here.
 	 */
-	udptable = net->ipv4.udp_table;
-	if (udp_hashed4(sk)) {
-		udp_rehash4(udptable, sk, hash);
+	if (udp_hashed4(sk))
 		return;
-	}
 
+	udptable = net->ipv4.udp_table;
 	hslot = udp_hashslot(udptable, net, udp_sk(sk)->udp_port_hash);
 	hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash);
 	hslot4 = udp_hashslot4(udptable, hash);
@@ -2173,14 +2171,14 @@ void udp_lib_rehash(struct sock *sk, u16 newhash, u16 newhash4)
 		struct udp_table *udptable = udp_get_table_prot(sk);
 		struct udp_hslot *hslot, *hslot2, *nhslot2;
 
+		hslot = udp_hashslot(udptable, sock_net(sk),
+				     udp_sk(sk)->udp_port_hash);
 		hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash);
 		nhslot2 = udp_hashslot2(udptable, newhash);
 		udp_sk(sk)->udp_portaddr_hash = newhash;
 
 		if (hslot2 != nhslot2 ||
 		    rcu_access_pointer(sk->sk_reuseport_cb)) {
-			hslot = udp_hashslot(udptable, sock_net(sk),
-					     udp_sk(sk)->udp_port_hash);
 			/* we must lock primary chain too */
 			spin_lock_bh(&hslot->lock);
 			if (rcu_access_pointer(sk->sk_reuseport_cb))
@@ -2199,19 +2197,29 @@ void udp_lib_rehash(struct sock *sk, u16 newhash, u16 newhash4)
 				spin_unlock(&nhslot2->lock);
 			}
 
-			if (udp_hashed4(sk)) {
-				udp_rehash4(udptable, sk, newhash4);
+			spin_unlock_bh(&hslot->lock);
+		}
+
+		/* Now process hash4 if necessary:
+		 * (1) update hslot4;
+		 * (2) update hslot2->hash4_cnt.
+		 * Note that hslot2/hslot4 should be checked separately, as
+		 * either of them may change with the other unchanged.
+		 */
+		if (udp_hashed4(sk)) {
+			spin_lock_bh(&hslot->lock);
 
-				if (hslot2 != nhslot2) {
-					spin_lock(&hslot2->lock);
-					udp_hash4_dec(hslot2);
-					spin_unlock(&hslot2->lock);
+			udp_rehash4(udptable, sk, newhash4);
+			if (hslot2 != nhslot2) {
+				spin_lock(&hslot2->lock);
+				udp_hash4_dec(hslot2);
+				spin_unlock(&hslot2->lock);
 
-					spin_lock(&nhslot2->lock);
-					udp_hash4_inc(nhslot2);
-					spin_unlock(&nhslot2->lock);
-				}
+				spin_lock(&nhslot2->lock);
+				udp_hash4_inc(nhslot2);
+				spin_unlock(&nhslot2->lock);
 			}
+
 			spin_unlock_bh(&hslot->lock);
 		}
 	}