<\x>
as for me, im fine with the small latency, so i just leave it as is
isak has quit [Quit: leaving]
isak has joined #openwrt-devel
Borromini has joined #openwrt-devel
goliath has joined #openwrt-devel
<Pokey>
Just picked up a ZBT-WG1602 V04 which runs OpenWRT out the box .No apparent documentation besides FCC stuff. As \x helpfully pointed out it could be ported to mainline, but although I am technically inclined I am not so in this field and I am not sure where to begin
zarzarzar_ has joined #openwrt-devel
zarzarzar has quit [Read error: Connection reset by peer]
g___ has joined #openwrt-devel
hgl has quit [Quit: Bye]
hgl has joined #openwrt-devel
diederik has joined #openwrt-devel
<Slimey>
Pokey do they supply sources
<Pokey>
Slimey: None that I can see no. Though I do have complete root access to the box as its running LEDE 17
<\x>
Pokey: please check first if it has that bootloader with a webui
<\x>
then you can test that already ported target board which is "ZBT-WG1602"
<\x>
normally you press the reset button at boot
<\x>
bootloader/recovery*
<Pokey>
\x Rgr. Its a single button device so there aren't too many possible scenarios. The board is a ZBT-WG1602 V04 so?
<\x>
normally the iterations are just ram and flash size
<Pokey>
I opted for the upgraded RAM and flash which is 512MB RAM 32MB flash
<\x>
if "ZBT-WG1602" boots and sime stuff are missing then we can fix things up
<\x>
yes as ive said the differences are just ram and flash
<Pokey>
Wonderbubble. Stand by, checking UI
<\x>
ram is autodected on on MT7621A, then we can just check if all the flash is accessed
<Pokey>
I also attached a serial console. Got a boot log whilst holding reset. Bootlog reports PandoraBox, says to access the WebUI for firmware upgrade, which I am just doing now...
<\x>
yeah it should be available on 192.168.1.1 most of the times
<Pokey>
Yep theres a web UI
<\x>
then you can try to flash ZBT-WG1602
<Pokey>
Its in chinese but its one big button in the middle which opens file selection so yea, I will do that
<robimarko>
And I can already see MAC setting regressions
<\x>
use this, might also be wise to include some stuff like pciutils and usbutils
<\x>
and have some pcie and usb devices at hand that way you can check if pcie and usb works
<Pokey>
rgr, added, build requested
<Pokey>
Got some WLAN cards and USB devices
<Pokey>
I discovered this is a 3 slot device. slot 1 is full PCIe and slots 2 and 3 are USB only, to which I have installed an adapter to turn them into USB ports already
<\x>
yeah whichever works, I use some old broadcom crystal hd to test these things
<Pokey>
Custom build succeded but only spat out a SYSUPGRADE. I suppose I flash the standard KERNEL bin via the WebUI and then later patch it with the custom SYSUPGRADE?
<\x>
sysupgrade will likely work
<Pokey>
rgr, will use that directly then
<\x>
pandorabox's bootloader works like that
<\x>
feel free to replace it with breed or something if you want overclocking but you need to know first which gpio is tied to that button, though I dont really recommend bootloader replacement for starters
<Pokey>
No exposed JTAG header on this unit and I don't feel like trying to find and solder one if it all goes badly, so I'll stick with Pandora
<\x>
its 32MB spi right?
<Pokey>
I am not certain
<\x>
then having no jtag is not that big of an issue
<Pokey>
Firmware flash done. Looks like its booting
<\x>
as spi-nor is easy to work with. either ch341a, XTW1000 or whatever
<\x>
Pokey: once youve booted, then check if all flash is accessed by going to the software menu, then proceed to test the eth ports on dmesg, usb ports on dmesg/lsusb
<\x>
if everything works then youre all good
<Pokey>
Right now I am not getting a link on LAN
<Pokey>
Well, link is up, but no IP
<\x>
weird. dhcp is enabled on openwrt
<\x>
try other ports
<\x>
maybe they changed something
<\x>
theres 5 ports on these so most of times either the 1st of the 5th one is WAN
<Pokey>
nothing. Link but no IP, all ports are a no including WAN
<\x>
try 21.xx
<\x>
maybe something is fucked up with the 2Gbps patch just like on newifi 3 d2
<\x>
if thats the issue then we can restore back the dts to pre-patch era
<Pokey>
Okay. I can on the serial console see that its connected to br-lan so I guess its not completely confused
<\x>
so you can replace that dtsi with the one from its genesis
<ynezz>
blocktrron: other then that, it looks pretty solid, so I would merge it sooner then later :)
<\x>
if it works then you should make a v04 variant and submit the changes
<Pokey>
Phew that's on the edge of my knowledge here. I'll have to do this later as I'm going shopping in 20 minutes. I hope you will not mind that I ask for guidance doing this. I will of course read the manual beforehand for the compilation of OpenWRT before bugging you/others in here
<blocktrron>
robimarko: with mac regressions you are referring incorectly configured mac-addresses or is there something wrong with the soc mac?
<Pokey>
No go on static IP set, by the way
<blocktrron>
I'm on ynezz side regarding a possible merge, in worst case we disable boards which are not yet done in case the migration is flimsy.
<blocktrron>
But i think it would be good if we have it in time for the next release. If we let it sit for the next 6 months, it probably won't reach a possible 22.x
<blocktrron>
*23.x
<\x>
Pokey: yeaah
<\x>
well you have to live with it, you have an unported board.
<\x>
what you can also try for now is to just install newifi 3 d2's firmware on it
<\x>
if that works then all thats left is enabling pcie and the usb ports
<\x>
newifi 3 d2 didnt receive that 2gbps patch on it as it bugs out on that
Borromini has quit [Quit: Lost terminal]
<Pokey>
I'm aware I'm in unchartered territory of my own doing. I'll get through it eventually and end up with a working device and a task to write a wiki page
<Pokey>
I'll try that firmware now then
<\x>
wifi will also not likely work but ethernet will likely work
<\x>
since it uses different chips for 5ghz or it is mish mashed on the eeprom
<Pokey>
Yea, it produced no WiFi network on the last image
<\x>
different variant yep since the one on wg1602's dts is a different 5ghz chip
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openwrt-devel
<\x>
>compatible = "pci14c3,7662";
<\x>
yours has 7612 I think
<Pokey>
MT7621A
<\x>
thats not the wifi chip
<Pokey>
Not immediately certain then
<\x>
we can check anyway on lspci
<Pokey>
No worries
<\x>
newifi d2's firmware booted?
<Pokey>
Booting
<Pokey>
I spotted 7612 in the log yes
<Pokey>
huh. Looks like it might have kernel paniced?
<\x>
no go then
<\x>
you have to compile wg1602 with the older "wg1602.dtsi"
<\x>
though for now since you have a shell anyway start profiling it, see what chips are installed for wifi
<hurricos>
blocktrron / robimarko: Is this a DSA migration for qca8k boards, or all boards? If all boards, which boards remain?
<hurricos>
I think apm821xx and mpc85xx are still out there, I can pick them up.
<blocktrron>
ipq40xx esw, Not qca8k
<hurricos>
OK, thanks.
<hurricos>
Sorry for the mistake, so it's specific to tohse
floof58 is now known as Guest1272
floof58 has joined #openwrt-devel
Guest1272 has quit [Ping timeout: 480 seconds]
<robimarko>
blocktrron: MAC adresses, I tried to ask everybody that converted to make sure they are the same
<robimarko>
But you know how it goes
<blocktrron>
Yes, I would see that as noise
<blocktrron>
And it can be fixed up later
<robimarko>
Yeah, for sure
<robimarko>
There were asks to just disable the boards that nobody converted until they are converted
<tmn505>
robimarko: disabling unconverted boards imho is prefereble, since You would lost ability interacting with them through ethernet and if wifi is disabled, one needs to do non-trivial recovery. While with converted ones, even with wrong configuration failsafe will be still available.
minimal has joined #openwrt-devel
<tmn505>
\x: setting to performance governor made huge difference, ping is average 0.3 ms. Thanks
<Pokey>
Hey \x, if the KERNEL has LuCI and starts up, whats the SYSUPGRADE add on top of that? Does it just install a bunch more packages?
<Pokey>
Oh, answered my own question. Initframfs mode. Changes are lost after rebooting
philipp64 has quit [Read error: No route to host]
* mrnuke
is having an annoying mind-bender with preinit
<mrnuke>
I do "boot_hook_add preinit_main my_hook" in /lib/preinit/81_..., but the hook does not run on boot
<mrnuke>
But if I run '/etc/preinit' after boot, then the hook runs just fine
<enyc>
Hrrm, OpenWRT 22.03.0 seems to (now) work fine on Plusnet Hub One, no "bootloop" warned about on wiki/ToH ...... is it like ly the issue is *actually* fixed, or just not happening for me, in this config, now ...?
<Pokey>
Erg, noob question, wht does a.dtsi which is present in the correct targets folder not appear in menuconfig, even though another in the same folder does?
<mrnuke>
Pokey: menuconfig is a big place. Where exactly does this dtsi appear?
<Pokey>
So, I don't know how much you're able to scroll up, but I have been discussing this with \x and I need to choose that specific dtsi to see if it resolved an issue for me. RIght now there is a variant of that dtsi with -16M on the end of it, which does indeed appear in menuconfig (by it's display name, of course) which is *not* the one I want
<mrnuke>
You don't typically "select" a dtsi in menuconfig. The typical fashion is for a devicetree to "#include "mt7621_zbtlink_zbt-wg1602.dtsi", which is exactly what mt7621_zbtlink_zbt-wg1602-16m.dts does.
<Pokey>
I see. Right. So I suppose I just need to find the .dts and point it at my preferred .dtsi
<mrnuke>
Normally, you would choose or add a target -- see target/linux/ramips/image/mt7621.mk. Each target has its own devicetree. That devicetree can then #Include dtsi s
<Pokey>
Or no, because ones a dts and ones a dtsi. Right. I think I know what I need to be doing now then. Just tweaking the dts a little and picking an older version of the dtsi
<Pokey>
hm. Can't even access the device over the network with the kernel image after rolling back the dts changes from GIT anymore... Thats problematic.
<mrnuke>
Serial cable?
ekathva has joined #openwrt-devel
ekathva has quit []
<Pokey>
mrnuke: Serial working fine
<mrnuke>
Pokey: so, can you recover over the serial port with TFTP?
<Pokey>
mrnuke: This system has a web based recovery, that is no isssue
<Pokey>
Right now I am trying to get a working firmware image
<Pokey>
the official openwrt-22.03.0-ramips-mt7621-zbtlink_zbt-wg1602-16m-initramfs-kernel works fine, but the sysupgrade image it comes with does not
<mrnuke>
Ooh! Those are fun to debug! Is it related to not being able to mount rootfs?
<Pokey>
\x pointed out to me that a DTSI change may have borked it. Nah, networking not working. Link is made but no IP is served and static IP is not working
<Pokey>
I locally built with the DTSI change reverted and the flash size increased (I have a version of the 1602 with double the flash), kernel doesn't even work. Boots, I get full busybox but no networking
<Pokey>
Put the new DTSI back, built, same result
<Pokey>
the official kernel one does work, gets me a WebUI
<Pokey>
applying the sysupgrade breaks it again
minimal has quit [Quit: Leaving]
gladiac has quit [Quit: k thx bye]
Gaspare has quit [Quit: Gaspare]
ekathva has joined #openwrt-devel
ekathva has quit [Remote host closed the connection]
MAbeeTT2 has joined #openwrt-devel
MAbeeTT has quit [Ping timeout: 480 seconds]
srslypascal is now known as Guest1282
srslypascal has joined #openwrt-devel
srslypascal has quit []
robimarko has quit [Quit: Leaving]
Guest1282 has quit [Ping timeout: 480 seconds]
<russell-->
Pokey: does dmesg tell you anything about your network device?
srslypascal has joined #openwrt-devel
<Pokey>
russell--: Everything on the surface appears to "work". It makes a link on the port, but provides no DHCP and static addresses are no use either. I have just built the bleeding edge latest commit to no success too.
<Pokey>
the SYSUPGRADE doesn't work and nothing I have built so far works. Though everything leads to a working BusyBox serial console, just networking issues
<mrkiko>
Pokey: did you try to revert the whole mt7621 refactoring patch, or did you just drop in the old dtsi + dts file?
<Pokey>
mrkiko: Dropin the DTSI file. Did not drop in the DTS file, but thats a tiny file anyway so not much point. did modify that DTS file though to increase the flash size, though I have since reverted all my changed, switched to the latest commit and compiled to no success still
srslypascal has quit [Quit: Leaving]
srslypascal has joined #openwrt-devel
<Pokey>
Manually configured an IP on a specific port after taking it out of the br-lan via serial console. It has an IP, it says UP under `ip a` but nada
<aiyion_>
Does someone own a cudy lt400 and can provide me with a partitiondump?
aiyion_ is now known as aiyion
Tapper has quit [Ping timeout: 480 seconds]
MaxSoniX has quit [Quit: Konversation terminated!]
cmonroe_ has joined #openwrt-devel
cmonroe has quit [Ping timeout: 480 seconds]
<mrkiko>
Any chance you guys know a software that detect duplicate files like fdupes does, but gives you the opportunity of creating symlinks or hardlinks of the copies?