Archer D7 support

Hi,

if you want to use the serial connection to normal way the rx, TX and GND has to be connect. The logical level is 3.3 Volts.

Regards,
Wolfgang

Thank you very much I'll try to install it with Rpi and share the result here later.

I finally managed to get the serial connection. There is one problem the environment I get after serial connection seems to be cfe. Not u-boot. Any suggestions?

There are two complete CPU groups inside the D7, a Qualcomm controlling the main routing functions and a Broadcom as the DSL modem. CFE is generally associated with Broadcom. Find the other serial port that is connected to the Qualcomm CPU.

Thank you very much. Now I have another problem. I opened the case so looking at the board from left to right we have vcc, gnd, rx, tx. If I connect the wires using this layout it doesn't work. It work properly for Broadcom part though.

Update:

Finally I could connect to Qualcomm but there is strange problem. Unlike Broadcom which shows CFE bootloader and boots normally, the u-boot stops after two lines. This is what I get and after this it just stops working.

U-Boot 1.1.4 (Jan 21 2016 - 16:02:58)

ap135 - Scorpion 1.0
DRAM:  

I also checked both router mode and ADSL modem mode work perfectly under TP-link firmware. Any ideas how to solve this?

The Vcc and GND pins are usually easy to tell because they connect to larger areas of copper than the Tx and Rx. Positively identify GND and Vcc and double check with an ohm meter if necessary. The GND will be zero ohms to other grounded parts like the antenna cable shields. The Vcc will have very low resistance to the inductor in the 3.3 volt regulator. Ohmmeter tests are always done with the power to the unit disconnected.

Once you have those positively identified, put your serial converter ground on GND and leave it, and leave Vcc alone.

This leaves Tx and Rx which will have small copper traces to the CPU. It is OK to try connecting Tx and RX both ways to the serial converter. It will only work one way. TX from the router needs to cross over to Rx on the converter and vise versa. If you don't get any data swap the two wires.

2 Likes

Probably the problem is caused by the adapter. I tried many USB-SERIAL adapters and many of them shown similar behaviour. The adapters based on the CP2102 work better for me.
However if you connect the adapter after the start of the bootloader (1 or 2 seconds after you switch the router on) you should be able to bypass the problem :wink: Once the serial connection works properly you can soft reboot the router and you will be able to stop the boot

@suppenkasper0815 @powo01 or every other user with the "b" version, could you please help with this?
If you need further instructions, I can help you.

Thanks

Hello Davide ,

at my device the mac from the stick is show at

 /proc/device-tree/ahb/eth@1a000000/mac-address

The real position where the device is able to the mac grep it is inside the rom partition at the offset I have written before

There are the following mac files inside /proc/device-tree/ahb

/proc/device-tree/ahb/eth@19000000
/proc/device-tree/ahb/eth@1a000000
/proc/device-tree/ahb/wmac@18100000

My current build on device

OpenWrt SNAPSHOT, r10073+284-0293aa72d1

Regards,
Wolfgang

1 Like

Thank you for the quick response! :slight_smile:

Can you confirm that
/proc/device-tree/ahb/eth@1a000000
and
/proc/device-tree/ahb/wmac@18100000
shows the same (LAN) mac addres, while
/proc/device-tree/ahb/eth@19000000
shows the WAN mac address?

@numero53

as requested I confirm

.../eth@1a000000 => sticker mac address
../wmac@18100000 same like  eth@1a000000
../eth@19000000 =>  eth@1a000000 incremented by 1

Regeards,
Wolfgang

1 Like

Perfect. Thank you! :slight_smile:

@numero53 sorry for the late response bro, I wasn't able to get access to my D7 yet...

Here are the results of my D50 with snapshot image


BusyBox v1.31.0 () built-in shell (ash)

  _______                     ________        __
 |       |.-----.-----.-----.|  |  |  |.----.|  |_
 |   -   ||  _  |  -__|     ||  |  |  ||   _||   _|
 |_______||   __|_____|__|__||________||__|  |____|
          |__| W I R E L E S S   F R E E D O M
 -----------------------------------------------------
 OpenWrt SNAPSHOT, r10494-e05310b9b8
 -----------------------------------------------------
root@D50v1:~#  /proc/device-tree/ahb/eth@1a000000/mac-addre
s
-ash: /proc/device-tree/ahb/eth@1a000000/mac-address: Permi
sion denied
root@D50v1:~# /proc/device-tree/ahb/eth@19000000
-ash: /proc/device-tree/ahb/eth@19000000: Permission denied
root@D50v1:~# /proc/device-tree/ahb/eth@1a000000
-ash: /proc/device-tree/ahb/eth@1a000000: Permission denied
root@D50v1:~# /proc/device-tree/ahb/wmac@18100000
-ash: /proc/device-tree/ahb/wmac@18100000: not found
root@D50v1:~# su
-ash: su: not found
root@D50v1:~#

Don't worry, we already solved for the D7 :slight_smile:
However you can be very useful for the D50, but you typed the wrong commands.

I wrote clearer instructions in the D50 thread so we don't go off-topic here. :wink:

Thank you :slight_smile:

1 Like

Hi @numero53,
could you possibly explain how you got the 5GHz wifi working on your excellent precompiled ath79 image above? I've tried multiple merges of the commits from your fork on github, and even just building your repo directly, but in every case I can't get a 5GHz wifi device to even be recognised. On first boot the wireless config file only contains one device, the 2.4GHz one.

I noticed that on your prebuilt image the Wireless page in LuCI actually shows it as a
Qualcomm Atheros QCA9880 802.11nac
But I can't find anything in the source tree with that text!

Really stuck, but loving the prebuilt image!
Thanks!

Hi @jcd92!
Are you sure you are compiling from the right branch (this)?
For what concerns the "Qualcomm Atheros QCA9880 802.11nac" string, you won't find anything like this in my commit because it's automatically recognised by the driver.
However I hope that the patch will be accepted soon.

Hi @numero53,
thanks! I was either not using the correct branch, or more likely I was not doing a clean recompile.
Either way I now get the correct 5GHz network on my fresh build as I was seeing on your precompiled image. I also see there are some important changes to LuCI and also the wifi config file that I wasn't getting before, so it must have kept some old parts of the build.

Many thanks!

how to compile the firmware? after executing 'make menuconfig' do i need changing some settings or can i just select "exit". i am new to this. thanks

Using the command make will start it building however you can also use make -j4 which will build it quicker of you have a Quad-Core system.

Do I need to change some settings when I use " make menuconfig"?
Where is the new firmware stored? Thanks.