Merge tag 'net-5.13-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Networking fixes, including fixes from bpf, wireless, netfilter and
  wireguard trees.

  The bpf vs lockdown+audit fix is the most notable.

  Things haven't slowed down just yet, both in terms of regressions in
  current release and largish fixes for older code, but we usually see a
  slowdown only after -rc5.

  Current release - regressions:

   - virtio-net: fix page faults and crashes when XDP is enabled

   - mlx5e: fix HW timestamping with CQE compression, and make sure they
     are only allowed to coexist with capable devices

   - stmmac:
      - fix kernel panic due to NULL pointer dereference of
        mdio_bus_data
      - fix double clk unprepare when no PHY device is connected

  Current release - new code bugs:

   - mt76: a few fixes for the recent MT7921 devices and runtime power
     management

  Previous releases - regressions:

   - ice:
      - track AF_XDP ZC enabled queues in bitmap to fix copy mode Tx
      - fix allowing VF to request more/less queues via virtchnl
      - correct supported and advertised autoneg by using PHY
        capabilities
      - allow all LLDP packets from PF to Tx

   - kbuild: quote OBJCOPY var to avoid a pahole call break the build

  Previous releases - always broken:

   - bpf, lockdown, audit: fix buggy SELinux lockdown permission checks

   - mt76: address the recent FragAttack vulnerabilities not covered by
     generic fixes

   - ipv6: fix KASAN: slab-out-of-bounds Read in
     fib6_nh_flush_exceptions

   - Bluetooth:
      - fix the erroneous flush_work() order, to avoid double free
      - use correct lock to prevent UAF of hdev object

   - nfc: fix NULL ptr dereference in llcp_sock_getname() after failed
     connect

   - ieee802154: multiple fixes to error checking and return values

   - igb: fix XDP with PTP enabled

   - intel: add correct exception tracing for XDP

   - tls: fix use-after-free when TLS offload device goes down and back
     up

   - ipvs: ignore IP_VS_SVC_F_HASHED flag when adding service

   - netfilter: nft_ct: skip expectations for confirmed conntrack

   - mptcp: fix falling back to TCP in presence of out of order packets
     early in connection lifetime

   - wireguard: switch from O(n) to a O(1) algorithm for maintaining
     peers, fixing stalls and a large memory leak in the process

  Misc:

   - devlink: correct VIRTUAL port to not have phys_port attributes

   - Bluetooth: fix VIRTIO_ID_BT assigned number

   - net: return the correct errno code ENOBUF -> ENOMEM

   - wireguard:
      - peer: allocate in kmem_cache saving 25% on peer memory
      - do not use -O3"

* tag 'net-5.13-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (91 commits)
  cxgb4: avoid link re-train during TC-MQPRIO configuration
  sch_htb: fix refcount leak in htb_parent_to_leaf_offload
  wireguard: allowedips: free empty intermediate nodes when removing single node
  wireguard: allowedips: allocate nodes in kmem_cache
  wireguard: allowedips: remove nodes in O(1)
  wireguard: allowedips: initialize list head in selftest
  wireguard: peer: allocate in kmem_cache
  wireguard: use synchronize_net rather than synchronize_rcu
  wireguard: do not use -O3
  wireguard: selftests: make sure rp_filter is disabled on vethc
  wireguard: selftests: remove old conntrack kconfig value
  virtchnl: Add missing padding to virtchnl_proto_hdrs
  ice: Allow all LLDP packets from PF to Tx
  ice: report supported and advertised autoneg using PHY capabilities
  ice: handle the VF VSI rebuild failure
  ice: Fix VFR issues for AVF drivers that expect ATQLEN cleared
  ice: Fix allowing VF to request more/less queues via virtchnl
  virtio-net: fix for skb_over_panic inside big mode
  ipv6: Fix KASAN: slab-out-of-bounds Read in fib6_nh_flush_exceptions
  fib: Return the correct errno code
  ...

1 files changed, 40 insertions, 39 deletions

diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
index bde6be77ea73..ef3d037f984a 100644
--- a/net/mptcp/subflow.c
+++ b/net/mptcp/subflow.c
@@ -630,21 +630,20 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 
 	/* if the sk is MP_CAPABLE, we try to fetch the client key */
 	if (subflow_req->mp_capable) {
-		if (TCP_SKB_CB(skb)->seq != subflow_req->ssn_offset + 1) {
-			/* here we can receive and accept an in-window,
-			 * out-of-order pkt, which will not carry the MP_CAPABLE
-			 * opt even on mptcp enabled paths
-			 */
-			goto create_msk;
-		}
-
+		/* we can receive and accept an in-window, out-of-order pkt,
+		 * which may not carry the MP_CAPABLE opt even on mptcp enabled
+		 * paths: always try to extract the peer key, and fallback
+		 * for packets missing it.
+		 * Even OoO DSS packets coming legitly after dropped or
+		 * reordered MPC will cause fallback, but we don't have other
+		 * options.
+		 */
 		mptcp_get_options(skb, &mp_opt);
 		if (!mp_opt.mp_capable) {
 			fallback = true;
 			goto create_child;
 		}
 
-create_msk:
 		new_msk = mptcp_sk_clone(listener->conn, &mp_opt, req);
 		if (!new_msk)
 			fallback = true;
@@ -1012,21 +1011,11 @@ static bool subflow_check_data_avail(struct sock *ssk)
 
 		status = get_mapping_status(ssk, msk);
 		trace_subflow_check_data_avail(status, skb_peek(&ssk->sk_receive_queue));
-		if (status == MAPPING_INVALID) {
-			ssk->sk_err = EBADMSG;
-			goto fatal;
-		}
-		if (status == MAPPING_DUMMY) {
-			__mptcp_do_fallback(msk);
-			skb = skb_peek(&ssk->sk_receive_queue);
-			subflow->map_valid = 1;
-			subflow->map_seq = READ_ONCE(msk->ack_seq);
-			subflow->map_data_len = skb->len;
-			subflow->map_subflow_seq = tcp_sk(ssk)->copied_seq -
-						   subflow->ssn_offset;
-			subflow->data_avail = MPTCP_SUBFLOW_DATA_AVAIL;
-			return true;
-		}
+		if (unlikely(status == MAPPING_INVALID))
+			goto fallback;
+
+		if (unlikely(status == MAPPING_DUMMY))
+			goto fallback;
 
 		if (status != MAPPING_OK)
 			goto no_data;
@@ -1039,10 +1028,8 @@ static bool subflow_check_data_avail(struct sock *ssk)
 		 * MP_CAPABLE-based mapping
 		 */
 		if (unlikely(!READ_ONCE(msk->can_ack))) {
-			if (!subflow->mpc_map) {
-				ssk->sk_err = EBADMSG;
-				goto fatal;
-			}
+			if (!subflow->mpc_map)
+				goto fallback;
 			WRITE_ONCE(msk->remote_key, subflow->remote_key);
 			WRITE_ONCE(msk->ack_seq, subflow->map_seq);
 			WRITE_ONCE(msk->can_ack, true);
@@ -1070,17 +1057,31 @@ static bool subflow_check_data_avail(struct sock *ssk)
 no_data:
 	subflow_sched_work_if_closed(msk, ssk);
 	return false;
-fatal:
-	/* fatal protocol error, close the socket */
-	/* This barrier is coupled with smp_rmb() in tcp_poll() */
-	smp_wmb();
-	ssk->sk_error_report(ssk);
-	tcp_set_state(ssk, TCP_CLOSE);
-	subflow->reset_transient = 0;
-	subflow->reset_reason = MPTCP_RST_EMPTCP;
-	tcp_send_active_reset(ssk, GFP_ATOMIC);
-	subflow->data_avail = 0;
-	return false;
+
+fallback:
+	/* RFC 8684 section 3.7. */
+	if (subflow->mp_join || subflow->fully_established) {
+		/* fatal protocol error, close the socket.
+		 * subflow_error_report() will introduce the appropriate barriers
+		 */
+		ssk->sk_err = EBADMSG;
+		ssk->sk_error_report(ssk);
+		tcp_set_state(ssk, TCP_CLOSE);
+		subflow->reset_transient = 0;
+		subflow->reset_reason = MPTCP_RST_EMPTCP;
+		tcp_send_active_reset(ssk, GFP_ATOMIC);
+		subflow->data_avail = 0;
+		return false;
+	}
+
+	__mptcp_do_fallback(msk);
+	skb = skb_peek(&ssk->sk_receive_queue);
+	subflow->map_valid = 1;
+	subflow->map_seq = READ_ONCE(msk->ack_seq);
+	subflow->map_data_len = skb->len;
+	subflow->map_subflow_seq = tcp_sk(ssk)->copied_seq - subflow->ssn_offset;
+	subflow->data_avail = MPTCP_SUBFLOW_DATA_AVAIL;
+	return true;
 }
 
 bool mptcp_subflow_data_available(struct sock *sk)