You are right, i forgot about that 
If no clock is provided in pll node, the fallback is 40MHz...
@lucize
Do you have the time to port for example the Netgear WNR2000v3 device to the new platform? I have serial connection to the device and can test images and provice logfiles.
can you boot the initramfs image with tft, can you try (tft and bootm) ? and bear in mind that 4.14 kernel is larger so it may be possible that you won't be able to use LUCI on it
1 Like
I also posted an ar71xx dmesg in the first post, but maybe you missed it because is hidden 
Have you got a reference for gpio-keys-polled being deprecated in favour of gpio-keys?
https://www.kernel.org/doc/Documentation/devicetree/bindings/input/gpio-keys.txt has no mention of any properties for poll interval, and https://www.kernel.org/doc/Documentation/devicetree/bindings/input/gpio-keys-polled.txt has no mention of any deprecations.
I fixed a lot of the DTS warnings in my branch I pushed earlier today: https://github.com/openwrt/openwrt/pull/991
oh, I was just doing this, but yours is extended
did you tried the regs too <x 0 0>, are they working ?
should I give you the dts, or make a PR ?
can you manage to build yours ?
and also, this is only for the initramfs boot, don't write it to flash, you didn't provide a functional dmesg for me to get some flash values, and also I'm sure it will work
and a cat /proc/mtd
I just saw that you modified the defaults for mdio-bus so no more values
Ok, that's kinda different 
Totally not deprecated, and depends on a proposed upstream patch. So.... until that's merged/accepted and backported, trying to use gpio-keys on current ath79 targets would straight up not work
I'm glad to know someone's finally got around to fixing up gpio interrupts on ath79, but it has to land first!
for me the buttons work on 1043nd-v3 and tl-mr3020-v1, also on wr841n-v9 and I think it worked on 1043nd-v1 too
Hi all, I've been having a go at the archer c7 v2. At the moment, I got it to boot and the switch is working. The USB ports and both WIFI radios (and the corresponding LEDs) do not work yet though.
Could someone point me in the right direction?
I've got two dmesg:
- trunk (https://ghostbin.com/paste/s564k/raw)
- my build (https://ghostbin.com/paste/555nm/raw)
Here is my git repo: https://github.com/SibrenVasse/openwrt/tree/ath79-archer-c7
Any help is much appreciated.
On TP Link WDR4300(AR9344) it does also not work with interrupts... Irq request failed...
Did a very small review on your branch...
@lucize Did you maybe face a leds-gpio: probe of leds failed with error -16 error on QCA9533 or 9531?
I am porting CPE210 v2 and I get that weird error in the bootlog and off course LEDs dont work
And weirdly WLAN does not broadcast anything despite it looking properly configured and iw reading the power and noise floor.
no, they are ok, like gpoi-keys.
This is the actual status for incomplete type of devices that I own:
qca9533 is working fine on 841n-v9, except wan led that is always on because forced link, I couldn't make it work without that.
maybe they share the gpios like netgear devices, with jtag, for wifi, did you load the correct art data ?
the problems are on qca9588 wifi (crashing the kernel), booting qca9563 (maybe wifi also), for me 1043nd-v1 wired part is unstable, sometime it routes packets sometimes not, tried the new dsa driver for realtek switch but I face a problem with
[ 1.060958] realtek-smi switch: deasserted RESET
[ 1.065729] realtek-smi switch: found an RTL8366RB switch
[ 1.101071] realtek-smi switch: unable to register switch ret = -517
https://patchwork.ozlabs.org/cover/921642/
tried kernel 4.16 and 4.17-rc7 but of course I'm not so good at it, they all stop at
[ 0.000000] Movable zone start for each node
[ 0.000000] Early memory node ranges
[ 0.000000] node 0: [mem 0x0000000000000000-0x0000000003ffffff]
[ 0.000000] Initmem setup node 0 [mem 0x0000000000000000-0x0000000003ffffff]
[ 0.000000] random: get_random_bytes called from start_kernel+0x8c/0x46c with crng_init=0
I also tried a u4ae fusion 70 but didn't finished it because redboot would accept only gzip kernel and I didn't resized the kernel partition because I tried to make owl-loader work with the device but I couldn't make the lazma-loader find the magic signature of the kernel (like on 1043nd-v1)
You can try patches sent by Linus on the OpenWrt mailing list.
He even sent a example configuration for a Dlink using that switch,so you can reuse that.
Altough even he is having a lot of issues getting it to work.
You cant really test out 4.16 and anything other than 4.14 since there are a lot of OpenWrt custom patches running on top.
Without those you cant do anything
Yes, the patches I used are from mailinglist, but he wrote them for 4.17 that's why I tried to use it.
I modified most of the patches in openwrt to apply on 4.16/17 but I don't know what I missed, I ignored the patches for flowoffload (or netfilter related), mtd/nand and other arch so I not waste time adjusting them without having the kernel boot