Why can a ppoe connection regularly fall off?

Question

linux

On the tplink TL-WR741ND (ver. 4.2) router, with the OpenWrt 12.09-rc1 firmware, the ppoe connection regularly falls off.
It started after the transition of my provider "VBT" (Kaliningrad) to the network "Golden Telecom" (Beeline).
If you connect directly from a PC (linux ubuntu) - there is Internet.
Here is the network config from the router /etc/config/network

config interface 'loopback'
  option ifname 'lo'
  option proto 'static'
  option ipaddr '127.0.0.1'
  option netmask '255.0.0.0'

config interface 'lan'
  option ifname 'eth0'
  option type 'bridge'
  option proto 'static'
  option ipaddr '192.168.1.1'
  option netmask '255.255.255.0'

config interface 'wan3'
  option proto 'pppoe'
  option password 'kel****'
  option ifname 'eth1'
  option username 'alexstep'
  option demand '0'

config switch
  option name 'eth0'
  option reset '1'
  option enable_vlan '1'

config switch_vlan
  option device 'eth0'
  option vlan '1'
  option ports '0 1 2 3 4'

The wan connection on the router regularly falls off and does not want to connect, rebooting the router does not help.
And it helps that it is interesting to rename the wan interface, for example, to wan2. After that, the same connection with a new name works.
Wan2 freezes again after a while, I rename it back to wan or wan3 or to anything else and there is a connection again.
This is some weird behavior.
Any idea what it is and why?
This is what dmesg outputs right after it "falls".

[   20.370000] br-lan: port 1(eth0) entered forwarding state
[   20.970000] eth1: link up (100Mbps/Half duplex)
[   21.730000] device wlan0 entered promiscuous mode
[   22.090000] br-lan: port 2(wlan0) entered forwarding state
[   22.100000] br-lan: port 2(wlan0) entered forwarding state
[   24.100000] br-lan: port 2(wlan0) entered forwarding state
[   77.070000] eth1: link down
[   78.030000] eth1: link up (100Mbps/Half duplex)
[ 2128.010000] ------------[ cut here ]------------
[ 2128.010000] WARNING: at net/sched/sch_generic.c:256 0x802011b0()
[ 2128.010000] NETDEV WATCHDOG: eth1 (ag71xx): transmit queue 0 timed out
[ 2128.020000] Modules linked in: ath79_wdt ohci_hcd ledtrig_usbdev ledtrig_netdev nf_nat_irc nf_conntrack_irc nf_nat_ftp nf_conntrack_ftp ipt_MASQUERADE iptable_nat nf_nat pppoe xt_conntrack xt_CT xt_NOTRACK iptable_raw xt_state nf_conntrack_ipv4 nf_defrag_ipv4 nf_conntrack ehci_hcd pppox ipt_REJECT xt_TCPMSS ipt_LOG xt_comment xt_multiport xt_mac xt_limit iptable_mangle iptable_filter ip_tables xt_tcpudp x_tables ppp_async ppp_generic slhc ath9k(O) ath9k_common(O) ath9k_hw(O) ath(O) mac80211(O) usbcore usb_common nls_base crc_ccitt cfg80211(O) compat(O) arc4 aes_generic crypto_algapi ledtrig_timer ledtrig_default_on leds_gpio gpio_button_hotplug(O)
[ 2128.080000] Call Trace:[<80283918>] 0x80283918
[ 2128.080000] [<80283918>] 0x80283918
[ 2128.090000] [<8007198c>] 0x8007198c
[ 2128.090000] [<802011b0>] 0x802011b0
[ 2128.090000] [<80071a40>] 0x80071a40
[ 2128.100000] [<802011b0>] 0x802011b0
[ 2128.100000] [<80201040>] 0x80201040
[ 2128.100000] [<8007c4ac>] 0x8007c4ac
[ 2128.110000] [<80076fb8>] 0x80076fb8
[ 2128.110000] [<800771c8>] 0x800771c8
[ 2128.110000] [<800e6188>] 0x800e6188
[ 2128.120000] [<800773fc>] 0x800773fc
[ 2128.120000] [<8006306c>] 0x8006306c
[ 2128.120000] 
[ 2128.130000] ---[ end trace 512fd51e622e6e05 ]---
[ 2128.130000] eth1: tx timeout
[ 2128.130000] eth1: link down
[ 2128.140000] eth1: link up (100Mbps/Half duplex)
[ 2138.010000] eth1: tx timeout
[ 2138.010000] eth1: link down

Here's an archive with all the configs just in case.