[FS#1431] hostapd: VLAN: br_delif: Failure determining interface index with dynamic vlans

[LEDE Bugs]

A new Flyspray task has been opened.  Details are below. 

User who did this - Ian MacDonald (imac) 

Attached to Project - OpenWrt/LEDE Project
Summary - hostapd: VLAN: br_delif: Failure determining interface index with dynamic vlans
Task Type - Bug Report
Category - Base system
Status - Unconfirmed
Assigned To - 
Operating System - All
Severity - Low
Priority - Very Low
Reported Version - lede-17.01
Due in Version - Undecided
Due Date - Undecided
Details - We see some error messages related to bridge port add/remove activities when using dynamic vlans, though the dynamic vlans themselves function correctly. 

The platforms include (Archer C7, Archer C2600, WD N600, BT HH5A) with standard 17.01.4 images, with the following package changes:

opkg remove wpad-mini hostapd
opkg install hostapd-common wpad

In our case, our captive portal system sends CoA requests to initiate VLAN changes, and passes the VLAN via radius-accept messages, which all works fine. Since radius and vlans are not normally used on open SSIDs, we have manually added the dae_ auth_ and acct_ variables to the encryption 'none' section of /lib/netifd/hostapd.sh 

Everything works wrt to the dynamic vlans and radius, but we see the following messages when the first client connects to a dynamic VLAN and when the last client disconnects from a vlan. Two examples are described below: 

1) When last client disconnects from dynamic vlan 71, connected via wlan0.71 on br-vlan71: 

   Wed Mar 14 15:31:54 2018 daemon.err hostapd: VLAN: br_delif: Failure determining interface index for 'wlan0.71'

2) When the first client connects to dynamic vlan 71, connected via wlan0.71 on br-vlan71:

   Wed Mar 14 15:33:19 2018 daemon.err hostapd: VLAN: vlan_add: ADD_VLAN_CMD failed for eth0: File exists

This only occurs when it is the first client on the vlan, or the last client to disconnect from the vlan, otherwise the dynamic vlan changes produce no errors related to the station-to-interface binding/unbinding. 

Despite the errors, the interfaces appear to be added and removed as expected.  i.e. after the first client connects we see 

root at AP:~# brctl show
bridge name	bridge id		STP enabled	interfaces
...
br-vlan71		7fff.c025e940533c	no		eth0.71
							wlan0.71

And after the last client disconnects we see

root at AP:~# brctl show
bridge name	bridge id		STP enabled	interfaces
...
br-vlan71		7fff.c025e940533c	no		eth0.71


In the case used for the example we have the following vlan related interface settings in /etc/config/wireless and /etc/config/network, noting that eth0 is the trunk over which the VLANs are connected to the OpenWRT/LEDE device in this instance:

config wifi-iface 'default_radio0'
	option device 'radio0'
	option mode 'ap'
	option encryption 'none'
	option vlan_file '/etc/config/hostapd.vlan'
	option vlan_tagged_interface 'eth0'
	option vlan_bridge 'br-vlan'
	option vlan_naming '0'
        ...

config interface 'vlan71'
	option type 'bridge'
	option ifname 'eth0.71'
        




More information can be found at the following URL:
https://bugs.openwrt.org/index.php?do=details&task_id=1431