LigoWave LigoDLB Propeller 5

Hi, everybody.
My boss bought a couple of LigoWawe LigoDLB Propeller 5 thinking that they were already supported by LEDE.
Unfortunately, he was wrong.
So I was given the task of trying to get LEDE running on it.
Stock firmware had a great troubleshooting feature that pretty much exports configuration of everything on the device including GPIO pins and rest.
So LEDs and reset button were easy, got MAC offsets and configured the single port that is connected via MII to AR9342
But I am now stuck, stock firmware does some kind of firmware verification before flashing.
TFTP flashes stock firmware without issues, I also tried to flash sysupgrade image over TFTP but after flashing all I got was LEDs flashing in a row.

It would really ease development if I got access to serial, but the case looks like I would need to destroy it in order to open it and access UART header.

Device specs are following:
16 MB of flash (Winbond W25Q128FVFG)
64 MB of RAM (Winbond W9751G6KB-25)
1 POE IN 100 Mbit port

So it should run LEDE without a problem, but LigoWawe uses dual firmware.
So when first firmware fails to boot 3 times the second one gets loaded.
So effective space is 8MB.

My current code can be found here:
Current firmware build, flash backup and dump of pretty much everything about device are here:

I was able to get online UI to accept the image.
Unfortunately, after some time it looks like bootloader just switches to firmware2.

Finally got the device open.
They use glue and plastic clips to hold the lid.
Then styrofoam fills most of the device space.

Antenna is glued on back side of PCB

Also,does anybody have AR9342 datasheet?

Here is the bootlog:
It looks like partition layout issue

your driver for the MTD is missing.

Use a mach-* file which is close enough to your board.
They should be plenty board with SPI Flash and 5GHz radio available.

Thanks, I was looking at mach files but I can't find where MTD driver is defined

I'm lazy to search now ...
Which board is close enough ??

CPE510 should be close enough.
Minus one RJ45 port.

looks good.
Also look at Open Mesh 5

because the startup script ask for

And this line is missing in ligdolb_5_setup(void)


@ElektromAn Thanks,but it is actually not missing.
Also, I did not understand your statement that startup script is asking for something.
Can you explain that?

Adding the mach* file is one part of LEDE support

again I'm too lazy to explain ...

git grep OM5P target/linux/ar71xx/base-files/
git grep CPE510 target/linux/ar71xx/base-files/

your machine name is parsed via startup

Well, I dont understand the issue.
Everything regarding machine name looks to be correctly setup.
In bootlog you can see it: MIPS: machine is LigoWawe DLB-5

Your Board name "DLB-5" is correct for boot the kernel.
After this some other things must happen.
i.e. for OM5P
look in
git grep -l om5p target/linux/ar71xx/base-files/
but all this happens in your rootfs.
But you need to fix the VFS issue to see this.

Your bootlog is missing i.e. these lines

[    2.994600] m25p80 spi0.0: found mx25l12805d, expected m25p80
[    3.000420] m25p80 spi0.0: mx25l12805d (16384 Kbytes)

Those lines are not missing.
Check line 70 and 71.
Only thing I can think of is to not patch cmdline or to include mtd in mach file

hack target/linux/ar71xx/config-4.4 CONFIG_CMDLINE="....." remove from it noinitrd. This allow boot to console.

Thanks,will try it in morning

Please try this

MTDPARTS := spi0.0:192k(u-boot)ro,64k(u-boot-env),7680k(firmware),7680k(firmware-backup),576k(data)ro,128k(cfg)ro,64k(art)ro

VFS can't be found because the startup scripts are searching for
kernel, rootfs and rootfs-data (for config)


not for kernel1, rootfs1 etc.

Will do, compiling it now.

@ElektromAn That helped a bit
But now there is fatal exception and reboot again.

It looks like the same issue was present here:

@robimarko show dts file which you use?

Since it is ar71xx dts is not used,but mach.
I have a feeling that that could be issue with AR8032 configuration.
In stock bootlog it loks like MII is used instead of RGMII