Alright everything's up. @ghoffman replace /lib/upgrade/platform.sh with this one to fix sysupgrade then upgrade to the new image.
@lmore377- thank you - all synced up and sysupgrade from your recent build does work.
i got a 5th device to play with - a SAC2V1K, also 256mb flash, machid 177e.
i think it's pretty consistent that 177d has 512mb (rev6) and 177e has 256mb (rev10)
these machid values are set in uboot environment at factory config.
i'm wondering if your modified bootcmd could check for the machid and issue the correct command based on the returned value?
thanks again!
I have a RAC2V1K REV 1 on firmware 1.1.32.
I used your 1.1.28 config file to get root, but am hesitant to continue. Do you think your build on github will work on my model? I won't be too disappointed if I brick it; I don't use it for anything important.
The revision on the web interface isn't accurate. Do fw_printenv to get the machine ID and reference the table here to see the actual revision. If it's 6 or 10 there's a pretty much 100% chance it work fine.
1 Like
I couldn't figure out how to get fw_printenv to say anything about the machine ID, but I found that mine is a REV 6 by looking at /proc/device-tree/model.
Thanks!
When you run fw_printenv one of the variables should me machid and it should be set to 177x where x is A thru E
@ghoffman I couldn't keep the whole command in one variable (I'm not too familiar with the way uboot does scripting) but these should work:
fw_setenv mtdcmd 'if test "$machid" = 177d; then setenv mtdparts mtdparts=nand0:0x1A000000@0x2400000(firmware) && echo 177d; elif test "$machid" = 177e; then setenv mtdparts mtdparts=nand0:0xDC00000@0x2400000(firmware) && echo 177e; else echo not compatible; fi'
fw_setenv bootcmd 'run mtdcmd && setenv mtdids nand0=nand0 && ubi part firmware && ubi read 0x44000000 kernel 0x6e0000 && bootm'
Run those commands from within openwrt. If you want to do it from uboot, just delete fw_ from fw_setenv
machid was definitely not one of the variables. Maybe that's new with fw 1.1.32? I wish I had saved the output, but now I have openwrt installed so I can't run it again.
Also, thank you for putting this build together and helping us out.
I know uboot sets it from smem on every boot and I've had this router for a while now so maybe I did saveenv while it was set. Can you do fw_printenv from openwrt so I can have a clean copy?
Here:
root@OpenWrt:~# fw_printenv
baudrate=115200
ipaddr=192.168.1.1
serverip=192.168.1.212
bootcmd=setenv mtdids nand0=nand0 && set mtdparts mtdparts=nand0:0x1A000000@0x2400000(firmware) && ubi part firmware && ubi read 0x44000000 kernel 0x6e0000 && bootm
bootdelay=2
root@OpenWrt:~#
Wow mine is definitely full of random stuff. Thank you!
@lmore377 Thank you very much for your work ...and for hosting the upgrade.
I had an image version of r13577 from you, as you made it for my request 5 months ago
https://forum.openwrt.org/t/askey-rac2v1k-rt4230w-rev6-support/15830/204?
I have used this firmware flawlessly with no hiccup at all and used wireguard and lots of VLAN and policies based routing
I upgraded the firmware today per your instructions by replacing platform.sh first
and upgraded via Luci as there is popup question to retain the config.
I upgraded to r15338 successfully. Thank you very much indeed.
The only thing, I have to add manually, I need to install the wireguard protocol which is not installed by default on r15338.
At first I change distfeeds.conf in /etc/opkg,
and install the wireguard package via luci and reboot, boom everything working perfectly. I plan to use these routers for dual wan and mesh wifi and I have been collecting 7 units of these routers.
also thank you to @ghoffman ,@efsg and everybody involve in this thread.
The amazing thing I experienced; the config was retained flawlessly including all my VPN setup, and ready to install any additional packages available without. warning incompatible kernel.
I have the same uboot from some units that came without QR code
I can't see machid in Uboot through fw_printenv
The router model is RAC2VIK FactoryID 3 with RAC2VIK firmware (256MiB) without QRcode the firmware is 1.1.23 , injecting with the config 1.1.28 without any problem.
I also have some of units RAC2VIK with QRcode
but the firmware is SAC2V1K without any option to inject the config through webgui.
My only option is by opening the router to get the serial console.
My experience to share related to FactoryID
FactoryID 2 most likely we can get 512MiB and the Machid 177d
FactoryID 3 most likely we get 256MiB and Machid 177e (can't confirm as @ghoffman has id3 =512MiB)
FactoryID 3 w/o QR code we can inject the config (depend of the spectrum hasn't upgraded the firmware through an automatic upgrade)
FactoryID 3 with QR code We need to open the router to get the serial console
i have reviewed my bootlogs from devices as far back as i have them.
it seems that one of my devices (rev 6, factory id 2, 512mb) did NOT have machid set in original environment as far as I can tell. i mistakenly thought I had cleaned the environment and then manually replaced it.
this would confirm what @eganov has found about one of his devices, and seems to make consistent device identification more problematic.
as @lmore377 has suggested, there is no consistent relationship between factory id and model
also - the results from cat /proc/device-tree/model would depend on the contents of the device tree file i think. my 256mb/machid 177e device was originally identified as rev10, but now since loading the recent firmware from @lmore377, the device-tree/model returns rev6.
It just says rev 6 because that's what I had put in the device tree and that name is static. The original firmware has the device tree files for all the revisions (that's how I found out about rev 1,3, and 5) and selects the correct one at boot based off of the machine ID.
makes sense now. thanks.
Did you get a chance to work on this? It would be exciting if the qca nss drivers could be made to work on this router!
I haven't really worked on it. I need to figure out which repo in that thread is the current/most updated so I can get started.
It is indeed a long thread. If it is of any help, my understanding is that the Netgear R7800 has had builds where these drivers have been incorporated and which have been working well. Kong - one of the well known developers from DD-WRT who recently moved over to OpenWRT development - also has his own builds using qca-nss cores, which have generally been well received. His qca-nss binaries are here - http://www.desipro.de/openwrt/trunk/ipq806x-nss/ - and his tweaks for his build are here - http://www.desipro.de/openwrt/sources/. The big thread on qca-nss above also has some discussions about it.
Let us know if any of us can be of any help to you. Thanks for all your work!
In your situation, I'd worry a lot more about rt4230w support getting merged into OpenWrt/ master (something lmore377 has been working on tirelessly for many months now), than the optional NSS acceleration (which is not part of OpenWrt yet, and will have a hard time to eventually -or not- getting merged). NSS dependencies would be more of a distraction of this first goal - and would actually make it harder to see official support for this device in OpenWrt.