I'll try to catch up, but since it is a testing device it doesn't follow the stable channel -- and hence cannot update regularly, although you can force the installation work opkg.
I will build a new version with an updated kernel and latest changes from the stable branch.
So far I am aware of some potential limitations:
saving on whw03v1 devices does not work, a change was introduced some commits ago and need to be reviewed
there is an updated whw03v2 with an unsupported nand chip
whw01 model is not supported
Other than that, devices were working fine and supported latest features (K,V)
v0.0.6 from May 2021 is working quite fine on whw03v1 with some tweaks to the board files.
Not sure why this is the only version that does not remove changes during reboot.
Let me know if I can test anything for you on my devices.
Glad to see you back here
For the whw03v1 the saving went wrong between the "Fix boot counter" release and the "move to 21.02 stable branch".
As there aren't any code commits I can't help searching what would cause that. I would guess something with the overlayfs or a drive mount during startup.
If working with the beta branch is easier for you I don't mind. I run a lot of beta software at home.
I do have another question: what's the difference between the assets? Which one to use in what situation?
I see initramfs-fit-zImage.itb, squashfs-factory.bin and squashfs-sysupgrade.bin
Would you mind sharing the board changes you've made?
Or, even better, do a PR to the repository (for now branch whw03-21.02) with your changes and I'll merge them.
Maybe I also found something interesting...
For some reason I have OpenWRT on both boort partitions.
I think I screwed the settings on partition 1. Can't access it any more. After several restarts OpenWRT started again on partition 2. But, then the settings weren't saved anymore.
I'll dig around, v1 are different as it has an eMMC (4Gb) vs the NAND on the v2 (both with plenty of space).
The assets generated:
initramfs-fit-zImage.itb: image file to load from bootloader, helpful to test changes without persisting them to disk
squashfs-factory.bin: firmware file to install from factory firmware by using the CA method
squashfs-sysupgrade: firmware file to update an OpenWRT installation
Hope it helps!
As a reference, this is the topic on which I log my steps to gain access to the devices and start figuring out how the devices worked (and I actually bricked one of my v1 devices).
And this is the topic explaining why some v2 versions do not work due to not identifying the NAND chip.
New release for whw03v2, merged with latest stable branch (21.02).
New release for whw03v1, based on tag 0.0.6 and merged with stable branch (21.02).
Haven't been able to test any of those, so if you can load the ITB file from the bootloader, share bootlog and results.
If you do not have serial access to the device, flash from the stock firmware to not overwrite.
I will post a procedure to manually flash v1 to stock from OpenWRT in a few days, v2 procedure is explained earlier in this post.
I've found out, that the sysupdate bin is written to the non active partition. That was the reason, why I did owerwrite my stock firmware.
So if you start sysupdate on part2 it is written to pat1 and vise versa.
The performance of the board files in your version 0.0.6 was not sufficient for radio 2.
Not sure if it is relevant, but for radion 0 the borad.bin file was a link to a non-existing file. I have removed the link and copied the new borad.bin file instead. With the new files, the information of the signal strength is not correct, but this is just a cosmetic issue.
Thank you for making OpenWrt available for newer "mesh" devices. I mainly bought Linksys Velops for their looks (bad reason?), since nowdays almost every (higher end) Wi-Fi router looks more of a "spaceship" than access point..
So, just bought pair of these Velop nodes (whw03v2) and of course wanted to replace official (bad, limited) f/w with OpenWrt (this device development to get supported seemed to be at the farthest from newer mesh devices), but unfortunately I'm experiencing exact same outcome as @jmceleney, @angdraug etc. discussed earlier. Guess my Velops too are using newer hardware with different NAND which is not yet supported, since:
Flashing any of whw03v2 OpenWrt releases on flipy's GitHub lead to solid purple or red led on nodes after "installing firmware" from stock gui (either CA-method or fwupdate.html, parent/child - doesn't matter - it does indicate that firmware was flashed successfully on both ways, however)
When waiting enough and switching device of and back on after flashing, it led stays solid blue and after boot count cycles, revert to stock firmware is performed
hw-version is indeed whw03v2, both in original package and stock f/w gui, with latest stock firmware inside when trying to flash
I really appriciate your quick work you are putting in this and just wanted to ask if making this "newer hardware" model of device with currently unsupported NAND available is how much work behind it? Is it how tough process and if there is something what can be tested/sorted out from end user perspective to speed up the progress?