If it is something serious, you can create a ticket on the svn timeline, but you will need to provide more information there, or brainslayer or Kong will just close the ticket and ignore the problem.
Start by posting in the build thread.
I am currently waiting for the duplicate MAC problem to be fixed on several of my routers. _________________ Before asking a question on the forums, update dd-wrt: Where do I download firmware? I suggest reading it all.
QCA Best WiFi Settings
Some dd-wrt wiki pages are up to date, others are not. PM me if you find an old one.
Atheros:
Netgear R7800 x3 - WDS AP / station, gateway, QoS
TP-Link Archer C7 v2 x2 - WDS Station
TP-Link TL-WDR3600 v1 - WDS Station
TP-Link 841nd v8 - NU
D-Link 615 C1/E3/I1 x 7 - 1 WDS station
D-Link 825 B1 - NU
D-Link 862L A1 x2 - WDS Station
Netgear WNDR3700v2 - NU
UBNT loco M2 x2 - airOS
Broadcom
Linksys EA6400 - Gateway, QoS
Asus N66U - AP
Netgear WNDR3700v3 - not used
MediaTek
UBNT EdgeRouter X - switch
- Upgrade to Factory-to-DD-wrt.bin FIRST
- Upgrade to tl-wr940ndv3-webflash.bin
Now, your router will work, but only in Wi-Fi.
To make Ethernet usable, you need to go (with Wifi) in the Web admin page, Basic Setup > WAN Connection Type > Connection Type > Disabled
Voilà ! Your ethernet ports are working
We need to figure if we can make WAN functions usable without bugging the ethernet switch though.
Posted: Fri Nov 04, 2016 19:47 Post subject: TL-WR940N v3: no functional ethernet
Hi,
I have a TL-WR940N v3, presumably international version (bought from Amazon in the EU), which is equivalent to the TL-WR941ND v6, save for the detachable antennas.
I have the same problems as everyone else and did a bit of digging. A quick look at the dmesg output seems to hint what's going on:
Code:
<6>[ 0.900000] Atheros AR71xx hardware watchdog driver version 0.1.0
<6>[ 0.900000] ar71xx-wdt: timeout=15 secs (max=17) ref freq=250000000
<7>[ 2.970000] ar71xx-wdt: enabling watchdog timer
<6>[ 3.070000] libphy: ag71xx_mdio: probed
<6>[ 3.670000] eth0: Atheros AG71xx at 0xba000000, irq 5, mode:GMII
<3>[ 3.670000] ag71xx ag71xx.0: no MII bus device specified
<4>[ 3.680000] ag71xx: probe of ag71xx.0 failed with error -22
<6>[ 3.690000] eth0: link up (1000Mbps/Full duplex)
<6>[ 3.910000] Loading modules backported from Linux version wt-2016-10-03-1-g6fcb1a6
<6>[ 3.920000] Backport generated by backports.git backports-20160324-9-g0e38f5c
<7>[ 4.600000] ath: EEPROM regdomain: 0x0
<7>[ 4.600000] ath: EEPROM indicates default country code should be used
<7>[ 4.600000] ath: doing EEPROM country->regdmn map search
<7>[ 4.600000] ath: country maps to regdmn code: 0x3a
<7>[ 4.600000] ath: Country alpha2 being used: US
<7>[ 4.600000] ath: Regpair used: 0x3a
<7>[ 4.610000] ieee80211 phy0: Selected rate control algorithm 'minstrel_ht'
<3>[ 4.620000] gpiochip_find_base: cannot find free range
<3>[ 4.620000] gpiochip_add: GPIOs -1..21 (ath9k-phy0) failed to register
<6>[ 4.630000] ieee80211 phy0: Atheros AR9561 Rev:0 mem=0xb8100000, irq=79
<6>[ 7.420000] device br0 entered promiscuous mode
<6>[ 7.440000] device br0 left promiscuous mode
<6>[ 7.450000] device br0 entered promiscuous mode
<6>[ 7.470000] device br0 left promiscuous mode
<6>[ 14.850000] device ath0 entered promiscuous mode
<6>[ 14.850000] br0: port 1(ath0) entered forwarding state
<6>[ 14.860000] br0: port 1(ath0) entered forwarding state
<6>[ 15.360000] device eth0 entered promiscuous mode
<6>[ 15.900000] eth0: link down
<6>[ 15.910000] eth0: link up (1000Mbps/Full duplex)
<6>[ 15.920000] device eth0 left promiscuous mode
<6>[ 16.860000] br0: port 1(ath0) entered forwarding state
The first noticeable problem is here:
Code:
<3>[ 3.670000] ag71xx ag71xx.0: no MII bus device specified
<4>[ 3.680000] ag71xx: probe of ag71xx.0 failed with error -22
By the look of things it's failing to correctly probe the eth chip: it doesn't detect the embedded switch, doesn't detect and bring up the second (WAN) interface and thus ethernet doesn't work. Interestingly, all detected interfaces end up sharing the same MAC address. Anyway, this is clearly a kernel driver problem (error -22 is EINVAL if I'm not mistaken).
The second issue is here:
Code:
<3>[ 4.620000] gpiochip_find_base: cannot find free range
<3>[ 4.620000] gpiochip_add: GPIOs -1..21 (ath9k-phy0) failed to register
I'm not entirely sure what the implications are, but I can report that the LEDs on my device don't work as they should: the Wireless LED works correctly, but none of the LAN LEDs work (they don't turn on when a device is connected), while the WAN indicator gets both the blue and orange LEDs turned on (despite no cable being connected), with the blue LED blinking as if there was activity.
Anyway, I'm interested in getting this device to work and I have some kernel debugging experience so I will look into this more in the coming days. (Meanwhile I don't think the device should be listed as "supported" on the wiki because a router without properly working ethernet ports isn't super useful ;P)
Posted: Sat Nov 05, 2016 14:40 Post subject: Re: TL-WR940N v3: no functional ethernet
If you were the man who got this router to work properly - you would be a very popular guy!
f00b4r wrote:
Hi,
Anyway, I'm interested in getting this device to work and I have some kernel debugging experience so I will look into this more in the coming days. (Meanwhile I don't think the device should be listed as "supported" on the wiki because a router without properly working ethernet ports isn't super useful ;P)
For information, I succeeded to flash OpenWRT on my WR940NDv3. Everything is working great (Wifi/WAN/LAN).
Yes I'm aware OpenWRT works on this device (I've installed it on mine and confirmed it does work). Which means getting current DD-WRT to work is mostly a matter of back-porting the necessary bits into its kernel (OpenWRT uses a 4.4 kernel, whereas the snapshot dd-wrt I tried is 3.x).
However I also noticed that dd-wrt's svn already has a 4.4 kernel (which is also the kernel version into which brainslayer is apparently working, judging by the changesets), so it's possible the work is already "done" and it's just a matter of building a 4.4-based dd-wrt image...
I'm not really familiar with dd-wrt's development process so I don't know which would be the favored way (backport vs moving forward), maybe a guru could chime in?
I'm a bit skeptical this will work though since none of these changes touch drivers/net/ethernet/ag71xx, which is where the problem lies: the integrated switch on that device is an AR9343.
For reference, here's the log of the successful switch probe on OpenWRT:
Code:
[ 0.827680] libphy: ag71xx_mdio: probed
[ 1.420977] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.1:04 [uid=004dd042, driver=Generic PHY]
[ 1.431079] eth0: Atheros AG71xx at 0xb9000000, irq 4, mode:MII
[ 2.020763] ag71xx-mdio.1: Found an AR934X built-in switch
[ 2.062988] eth1: Atheros AG71xx at 0xba000000, irq 5, mode:GMII
).
