Bug #186

wndr3700 eth0 cannot be configured as a distinct ethernet port

Added by Dave Täht almost 2 years ago. Updated over 1 year ago.

Status:Closed Start date:05/29/2011
Priority:High Due date:
Assignee:Jo-Philipp Wich % Done:

100%
Category:- Spent time: 14.50 hours
Target version:1st Public Cerowrt release

Description

By default, openwrt configures the first radio + ethernet port as a bridge across the local switch, and wireless. It appears to be picking up the mac address of the wireless interface when doing so, and works pretty good, except that thus far I've been unable to coax greater than 60Mbit performance out of it.

It would be useful to be able to NOT configure the ethernet interface as a bridge, but as a real, separate wired interface for my ongoing investigation to QoS and multicast issues, in particular.

Attempting to do so, results in the following - even after disabling the wireless interfaces entirely.

root@aitne:~# ifconfig eth0
eth0 Link encap:Ethernet HWaddr C4:3D:C7:B0:AE:36
inet addr:192.168.130.1 Bcast:192.168.130.255 Mask:255.255.255.0
inet6 addr: fe80::c63d:c7ff:feb0:ae36/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1219 errors:0 dropped:1218 overruns:0 frame:0
TX packets:1296 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:16
RX bytes:130214 (127.1 KiB) TX bytes:338924 (330.9 KiB)
Interrupt:4

100% drops.

Even though, the other ethernet interface is fine

eth1 Link encap:Ethernet HWaddr C4:3D:C7:B0:AE:37
inet addr:192.168.115.171 Bcast:192.168.115.255 Mask:255.255.255.0
inet6 addr: fe80::c63d:c7ff:feb0:ae37/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:5081 errors:0 dropped:0 overruns:0 frame:0
TX packets:3399 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:5
RX bytes:669121 (653.4 KiB) TX bytes:312308 (304.9 KiB)
Interrupt:5

openwrt bleeding edge, r26878

There are no revealing errors in the log, this is all that is said about eth0 and eth1.

root@aitne:~# dmesg | grep eth0
eth0: Atheros AG71xx at 0xb9000000, irq 4
eth0: using fixed link parameters
eth0: link up (1000Mbps/Full duplex)
eth0: link down
eth0: link up (1000Mbps/Full duplex)

ar71xx-wndr3700-locally-administered-eth0mac.patch (1.1 kB) Jo-Philipp Wich, 06/16/2011 05:44 am

History

#1
Updated by Dave Täht almost 2 years ago

config switch
option name rtl8366s
option reset 1
option enable_vlan 1 # Blinkrate: 0=43ms; 1=84ms; 2=120ms; 3=170ms; 4=340ms; 5=670ms
option blinkrate 2

config switch_vlan
option device rtl8366s
option vlan 1
option ports "0 1 2 3 5t"

I'm about to turn off vlan and see what happens.

#2
Updated by Dave Täht almost 2 years ago

Turned off the vlan on the switch, I see packets YES! Bwhahahahaa

#3
Updated by Dave Täht almost 2 years ago

the wireless interfaces insist on coming up on the same mac as eth0

perhaps there is a mac wired into the ethernet interface that is not being seen?

#4
Updated by Dave Täht almost 2 years ago

The only cure for this, assuming there is no mac addr assigned, is to either
invent one, or buy a OUI.

I REALLY want to route, not bridge.

#5
Updated by Dave Täht almost 2 years ago

  • Assignee set to Dave Täht
  • % Done changed from 0 to 100

I am frustrated enough by bridging to want to go pure routing. I can easily, with various forms of traffic on the switch, totally saturate the wireless bridge. Regrettably there is no extra mac on the wndrs.

Buying a full set of 'bufferbloat' mac ids will cost 1750.00 dollars.

#6
Updated by Jo-Philipp Wich almost 2 years ago

This is a possible fix for the duplicate MAC address issue, the same trick is already applied to the Routerstation boards.

#7
Updated by Dave Täht almost 2 years ago

Oh, thank you. I did NOT see this fix when it went by.

#8
Updated by Dave Täht almost 2 years ago

  • Status changed from New to Resolved
  • Assignee changed from Dave Täht to Jo-Philipp Wich

automagic mac generation fix appears to work well. Would suggest pushing into openwrt head, if you haven't already.

#9
Updated by Dave Täht almost 2 years ago

  • Project changed from Project BISMark to Cerowrt

#10
Updated by Dave Täht almost 2 years ago

  • Status changed from Resolved to Closed

Well, I'll carry this patch forward if I have to.

#11
Updated by Dave Täht over 1 year ago

  • Target version changed from 2nd generation (worldwide bismark measurements) to 1st Public Cerowrt release

Also available in: Atom PDF