Merge tag 'net-5.12-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Networking fixes for 5.12-rc7, including fixes from can, ipsec,
  mac80211, wireless, and bpf trees.

  No scary regressions here or in the works, but small fixes for 5.12
  changes keep coming.

  Current release - regressions:

   - virtio: do not pull payload in skb->head

   - virtio: ensure mac header is set in virtio_net_hdr_to_skb()

   - Revert "net: correct sk_acceptq_is_full()"

   - mptcp: revert "mptcp: provide subflow aware release function"

   - ethernet: lan743x: fix ethernet frame cutoff issue

   - dsa: fix type was not set for devlink port

   - ethtool: remove link_mode param and derive link params from driver

   - sched: htb: fix null pointer dereference on a null new_q

   - wireless: iwlwifi: Fix softirq/hardirq disabling in
     iwl_pcie_enqueue_hcmd()

   - wireless: iwlwifi: fw: fix notification wait locking

   - wireless: brcmfmac: p2p: Fix deadlock introduced by avoiding the
     rtnl dependency

  Current release - new code bugs:

   - napi: fix hangup on napi_disable for threaded napi

   - bpf: take module reference for trampoline in module

   - wireless: mt76: mt7921: fix airtime reporting and related tx hangs

   - wireless: iwlwifi: mvm: rfi: don't lock mvm->mutex when sending
     config command

  Previous releases - regressions:

   - rfkill: revert back to old userspace API by default

   - nfc: fix infinite loop, refcount & memory leaks in LLCP sockets

   - let skb_orphan_partial wake-up waiters

   - xfrm/compat: Cleanup WARN()s that can be user-triggered

   - vxlan, geneve: do not modify the shared tunnel info when PMTU
     triggers an ICMP reply

   - can: fix msg_namelen values depending on CAN_REQUIRED_SIZE

   - can: uapi: mark union inside struct can_frame packed

   - sched: cls: fix action overwrite reference counting

   - sched: cls: fix err handler in tcf_action_init()

   - ethernet: mlxsw: fix ECN marking in tunnel decapsulation

   - ethernet: nfp: Fix a use after free in nfp_bpf_ctrl_msg_rx

   - ethernet: i40e: fix receiving of single packets in xsk zero-copy
     mode

   - ethernet: cxgb4: avoid collecting SGE_QBASE regs during traffic

  Previous releases - always broken:

   - bpf: Refuse non-O_RDWR flags in BPF_OBJ_GET

   - bpf: Refcount task stack in bpf_get_task_stack

   - bpf, x86: Validate computation of branch displacements

   - ieee802154: fix many similar syzbot-found bugs
       - fix NULL dereferences in netlink attribute handling
       - reject unsupported operations on monitor interfaces
       - fix error handling in llsec_key_alloc()

   - xfrm: make ipv4 pmtu check honor ip header df

   - xfrm: make hash generation lock per network namespace

   - xfrm: esp: delete NETIF_F_SCTP_CRC bit from features for esp
     offload

   - ethtool: fix incorrect datatype in set_eee ops

   - xdp: fix xdp_return_frame() kernel BUG throw for page_pool memory
     model

   - openvswitch: fix send of uninitialized stack memory in ct limit
     reply

  Misc:

   - udp: add get handling for UDP_GRO sockopt"

* tag 'net-5.12-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (182 commits)
  net: fix hangup on napi_disable for threaded napi
  net: hns3: Trivial spell fix in hns3 driver
  lan743x: fix ethernet frame cutoff issue
  net: ipv6: check for validity before dereferencing cfg->fc_nlinfo.nlh
  net: dsa: lantiq_gswip: Configure all remaining GSWIP_MII_CFG bits
  net: dsa: lantiq_gswip: Don't use PHY auto polling
  net: sched: sch_teql: fix null-pointer dereference
  ipv6: report errors for iftoken via netlink extack
  net: sched: fix err handler in tcf_action_init()
  net: sched: fix action overwrite reference counting
  Revert "net: sched: bump refcount for new action in ACT replace mode"
  ice: fix memory leak of aRFS after resuming from suspend
  i40e: Fix sparse warning: missing error code 'err'
  i40e: Fix sparse error: 'vsi->netdev' could be null
  i40e: Fix sparse error: uninitialized symbol 'ring'
  i40e: Fix sparse errors in i40e_txrx.c
  i40e: Fix parameters in aq_get_phy_register()
  nl80211: fix beacon head validation
  bpf, x86: Validate computation of branch displacements for x86-32
  bpf, x86: Validate computation of branch displacements for x86-64
  ...

2 files changed, 16 insertions, 7 deletions

diff --git a/net/dsa/dsa2.c b/net/dsa/dsa2.c
index d142eb2b288b..3c3e56a1f34d 100644
--- a/net/dsa/dsa2.c
+++ b/net/dsa/dsa2.c
@@ -795,8 +795,14 @@ static int dsa_tree_setup_switches(struct dsa_switch_tree *dst)
 
 	list_for_each_entry(dp, &dst->ports, list) {
 		err = dsa_port_setup(dp);
-		if (err)
+		if (err) {
+			dsa_port_devlink_teardown(dp);
+			dp->type = DSA_PORT_TYPE_UNUSED;
+			err = dsa_port_devlink_setup(dp);
+			if (err)
+				goto teardown;
 			continue;
+		}
 	}
 
 	return 0;
diff --git a/net/dsa/switch.c b/net/dsa/switch.c
index 4b5da89dc27a..32963276452f 100644
--- a/net/dsa/switch.c
+++ b/net/dsa/switch.c
@@ -107,7 +107,7 @@ static int dsa_switch_bridge_leave(struct dsa_switch *ds,
 	bool unset_vlan_filtering = br_vlan_enabled(info->br);
 	struct dsa_switch_tree *dst = ds->dst;
 	struct netlink_ext_ack extack = {0};
-	int err, i;
+	int err, port;
 
 	if (dst->index == info->tree_index && ds->index == info->sw_index &&
 	    ds->ops->port_bridge_join)
@@ -124,13 +124,16 @@ static int dsa_switch_bridge_leave(struct dsa_switch *ds,
 	 * it. That is a good thing, because that lets us handle it and also
 	 * handle the case where the switch's vlan_filtering setting is global
 	 * (not per port). When that happens, the correct moment to trigger the
-	 * vlan_filtering callback is only when the last port left this bridge.
+	 * vlan_filtering callback is only when the last port leaves the last
+	 * VLAN-aware bridge.
 	 */
 	if (unset_vlan_filtering && ds->vlan_filtering_is_global) {
-		for (i = 0; i < ds->num_ports; i++) {
-			if (i == info->port)
-				continue;
-			if (dsa_to_port(ds, i)->bridge_dev == info->br) {
+		for (port = 0; port < ds->num_ports; port++) {
+			struct net_device *bridge_dev;
+
+			bridge_dev = dsa_to_port(ds, port)->bridge_dev;
+
+			if (bridge_dev && br_vlan_enabled(bridge_dev)) {
 				unset_vlan_filtering = false;
 				break;
 			}