Kevinjil has quit [Ping timeout: 480 seconds]
slh has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
<KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_lantiq.html has been updated. (96.8% images and 98.6% packages reproducible in our current test framework.)
Daanct12 has joined #openwrt-devel
dermoth has quit [Remote host closed the connection]
thejoker8814 is now known as Guest5132
thejoker8814 has joined #openwrt-devel
Guest5132 has quit [Ping timeout: 480 seconds]
tSYS has quit [Quit: *squeak*]
tSYS has joined #openwrt-devel
Borromini has joined #openwrt-devel
Daanct12 has quit [Quit: WeeChat 4.2.2]
Daanct12 has joined #openwrt-devel
rua has quit [Quit: Leaving.]
gch981213 has joined #openwrt-devel
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openwrt-devel
Borromini has quit [Ping timeout: 480 seconds]
robimarko has joined #openwrt-devel
noltari has quit [Quit: Bye ~ Happy Hacking!]
noltari has joined #openwrt-devel
Daanct12 has quit [Quit: WeeChat 4.2.2]
Borromini has joined #openwrt-devel
Borromini has quit [Ping timeout: 480 seconds]
Borromini has joined #openwrt-devel
Forst has quit [Quit: Quitting]
bluew has quit [Ping timeout: 480 seconds]
Borromini has quit [Ping timeout: 480 seconds]
KGB-2 has quit [Remote host closed the connection]
KGB-2 has joined #openwrt-devel
minimal has joined #openwrt-devel
goliath has joined #openwrt-devel
rmilecki has quit [Quit: Konversation terminated!]
slh64 has joined #openwrt-devel
rua has joined #openwrt-devel
Forst has joined #openwrt-devel
rua has quit [Remote host closed the connection]
rua has joined #openwrt-devel
<j4cbo> so i'm looking at getting a Zyxel XGS1210, which is not currently supported in main, but the XGS1250 is... it seems like it would mostly just involve bringing the dts files from https://github.com/openwrt/openwrt/pull/10453, right? (cc svanheule since you've merged some XGS1250 PRs recently)
<j4cbo> (probably going to be replacing my GS110TUP)
esters has joined #openwrt-devel
<esters> Hi, I recently bought another TP-Link Archer C2600 v1 (ipq806x target) for testing purposes. How can I contribute or test OpenWRT on the device ?
mentalow has quit [Quit: :]]
mentalow has joined #openwrt-devel
minimal has quit [Quit: Leaving]
kenny has quit [Quit: WeeChat 4.2.2]
<Fijxu> Hello, is there any way to create my own OpenWRT image for a GPON router that uses U-Boot, Quantenna and some sort of "topaz-linux.lzma.img" Linux kernel on it? There is any good resouces about router moding?
<jakllsch> depends, but you'd probably be better off with a dumb ONT and a seperate router
<jakllsch> if the ONT function is being done in the same system as the router and wifi, it's likely you'll find it almost impossible to get into
<Fijxu> jakllsch: dumb ONT?
<Fijxu> And yeah the ONT runs the whole system, offering wifi and ethernet
<Fijxu> I even have the manual soo
Mangix has quit [Read error: Connection reset by peer]
<stintel> the nicest thing would be an OpenWrt device with SFP slot and being able to plug in the fiber directly
* stintel continues dreaming
<f00b4r0> problem is most ISPs will not provide you with SFP ONT
<f00b4r0> would be too nice otherwise (the Mochabin would be perfect in that setup)
<stintel> I'm considering asking the local guy who offers fiber here if this is an option
<Ansuel> on italy we are almost lucky... they provide ONT that looks to be designed for advanced user to use SOIC8 clip, dump the nand and get all the needed data (ISP use password and mac...)
<Ansuel> with those info it's only a matter of buy compatible SFP and you are done
<Ansuel> https://hack-gpon.org/ for the task
<f00b4r0> yep
<f00b4r0> I tried that here (FR), bought a number of (Italian-sourced) SFP ONTs, but I could never get it to work. Seems French ISPs have more controls in place :P
<f00b4r0> the best I got was I could receive some data, but I was sending to a blackhole
<f00b4r0> I should ebay these modules to recoup the expenses ;P
<Ansuel> i guess you took the tecnicolor one
<f00b4r0> I took all of them
<Ansuel> LOL
<f00b4r0> I don't even remember the references
<f00b4r0> i have like 5 or 6 different types, at least 2 copies of each, just to rule out duds
<Ansuel> anyway the controls most of the time is some shit put in the version or vendor path...
<Ansuel> if they use certificate to auth the connection (done by comcast in us) you should not receive any packet in theory
<f00b4r0> I'll give it another try when I move out, I'll use a different ISP then, will be a good opportunity to try again
Forst has quit [Quit: Quitting]
Forst has joined #openwrt-devel
<schmars[m]> stintel i do thinkcentre m720q + mellanox pcie card
<schmars[m]> cheap, and huge refurb market
<KanjiMonster> Fijxu: Quantenna Topaz is a wifi chip running a full linux stack on it (which is where the u-boot etc is used for). There is likely a different main system where the wifi is attached to.
<Fijxu> yeah I know
<Fijxu> there is two UART ports where I can attach a serial terminal
<Fijxu> And when I power up the router it, one console outputs the U-boot thing and then later the other one starts
noltari has quit [Ping timeout: 480 seconds]
Mangix has joined #openwrt-devel
rz has quit [Ping timeout: 480 seconds]
<Mangix> Ansuel: ping
<Ansuel> boing ?
<Mangix> for qca8k, what's the address for the switch?
<Ansuel> address? it's all from mdio
<Mangix> In my tree, I seem to have it defined at 10, but some definitions are at 1f
<Mangix> I'm guessing 1f is for the builtin one
<Ansuel> mdio tools should help to check what is on the bus i guess
<Ansuel> what soc is that?
<Ansuel> ath79 ?
<Mangix> yeah
<Ansuel> anyway in theory dmesg should give you the address
<Mangix> I'm seeing these mt7530 address fixes. Wonder if qca8k is the same way
<Ansuel> btw with some collaboration we can bisect the problem with qca8k
<Mangix> the other thing is that in my tree, I have switch0@10 but reg = <0> ???
<Ansuel> well yes reg are a bit random as mdio is actually used and then compatible i guess
<Mangix> right. these mt7530 fixes match reg and address. Actually I'm pretty sure that's also a DTC warning when mismatched
<Mangix> Ansuel: which problem btw?
<Ansuel> endianess problem....
<Mangix> ah you mean all ports except 1 dying after a while
<Ansuel> yes
<Ansuel> the info needed are with nothing connected pcap to the cpu port (eth0) and bridge fdb show output (while pcap collecting)
<Mangix> hmm? tcpdump?
<Ansuel> yes
<Ansuel> nothing connected ideally the thing run from serial
<Mangix> hmm what would that track? I/O with lo?
<Ansuel> mgmt eth
<Ansuel> mdio command sent in eth packet
<Mangix> well, hopefully there's traffic when no ethernet is connected
<Ansuel> in theory there should be both packet sent and received
<Ansuel> kernel creating the mgmt packet sent to the switch and switch answering them
<Ansuel> (if they are correct)
<Mangix> is tcpdump to be pointed at eth0, eth1, br-lan, or lo?
<Ansuel> at eth0
<Ansuel> or whatever is the cpu port set in dts
<Mangix> well, this has two
<Ansuel> the first one
<Mangix> right
<Mangix> just looked at ipq806x. switch@10 with reg = <0x10>
<Mangix> I should probably match that here...
<Ansuel> i wonder if this is the problematic code memcpy(val + 1, mgmt_eth_data->data + 1, len - QCA_HDR_MGMT_DATA1_LEN);
<Ansuel> mhhh actually no... but maybe yes
<Ansuel> bridge fdb show show that... since the value returned from that will be inconsistent
<Mangix> alright. flashing new build
<Mangix> let's see if it'll work
<Mangix> kernel 6.1. great.
<Mangix> /bin/ash: bridge: not found
<Ansuel> you need to install
<Ansuel> ip-bridge
<Ansuel> it's a different package than ip-full and ip-tiny
<Mangix> how do I run tcpdump in the background? screen?
<Mangix> tcpdump -i eth0
<Mangix> tcpdump: eth0: That device is not up
<Mangix> uhhhh
<Mangix> ERROR: package/feeds/packages/screen failed to build. ......
<Ansuel> o.O Mangix dmesg pls ?
<Mangix> error: implicit declaration of function 'openpty'; did you mean 'OpenPTY'? [-Wimplicit-function-declaration]
<Mangix> missing header...
<Ansuel> main problem of this warning is cmake tests
<Ansuel> that will probably disable all kind of stuff and enable workaround and other stuff
<Ansuel> had to catch problems...
<Ansuel> hard to *
<Mangix> ERROR: package/network/utils/iproute2 failed to build (build variant: tctiny).
<Mangix> note: expected 'const struct timeval *' but argument is of type 'const struct timeval *'
<Mangix> go home GCC, you're drunk
rmilecki has joined #openwrt-devel
robimarko has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
<Mangix> Ansuel: k running tcpdump on eth1 with nothing connected. what am I looking for?
<Ansuel> can you give me dmesg just to make sure eth1 is the correct port?
<Mangix> last two entries are me disconnecting the cable
<Ansuel> yes eth1 should be the first cpu port
<Mangix> yeah, my reading of it is eth1 == port0 and eth0 == port6
<Mangix> based on dts
<Ansuel> ok by interrupting the tcpdump
<Ansuel> do you have packet?
<Ansuel> or 0 ?
<Mangix> yeah, file is 5587012 bytes right now
<Ansuel> that is a lot
<Ansuel> please do this... run tcpdump run the bridge command
<Ansuel> and terminate tcpdump
<Mangix> almost 6 megabytes now
<Ansuel> and please pass me the .pcap
<Mangix> both or the latter?
<Ansuel> latter
<Mangix> k
<Ansuel> well i have to manually parse it
<Mangix> doesn't sound fun
<Ansuel> do you also have ethtool ?
<Mangix> i can compile it in
<Ansuel> yes please
<Ansuel> anyway it seems the switch is answering to packet correctly
<Ansuel> but maybe there is a problem with extended packets
<Mangix> bigger than 1500?
<Ansuel> no
<Ansuel> the mgmt eth can write up to 16bytes of data
<Ansuel> we normally write only 4 byte but for some command like vlan or fdb
<Mangix> alright. what you need with ethtool?
<Ansuel> tcpdump tcpdump -S eth0 and then stop tcpdump
<Mangix> hmm?
<Mangix> tcpdump: can't parse filter expression: syntax error
<Slimey> just once
<Ansuel> sorry
<Ansuel> the secondo is ethtool not tcpdump
<Ansuel> and what was the output?
<Mangix> I lost it. I think I should have used eth1
<Ansuel> mhhh strange maybe we need more advanced hack to bisect this...
<Mangix> well
<Mangix> lan1-4 currently works
<Ansuel> can you by chance apply that patch the user provided with the #if 0 ?
<Ansuel> and compare output of bridge fdb show ?
<Mangix> OK
bluew has joined #openwrt-devel
<Ansuel> in the meantime i will do something i wanted to do from a long time
<Ansuel> wireshark custom dissector for ath packets
<Ansuel> EHEHEH
<Mangix> Ansuel: flashed. You want tcpdump?
<Ansuel> no should be empty
<Ansuel> just compare the output
<Ansuel> i assume there should be extra lines
<Mangix> identical output
<Ansuel> ugh... and the problem was that after some time ports just stops working?
<Mangix> yep. around ~30 minutes of idling
<Mangix> all ports except port 2
goliath has quit [Quit: SIGSEGV]
<Mangix> the other user is able to get it to fail faster, but with traffic
<Mangix> oh boy, 6.6 for ath79 got merged
<Ansuel> Mangix I have to prepare a patch to validate if data is getting applied correctly
<Ansuel> but it's a bit late here so i can do it tomorrow
<Ansuel> hope it's not a problem
<Mangix> OK
<Ansuel> idea is to make mgmt eth write the command
<Ansuel> and read the reg to check if it got correctly applied
<Ansuel> for sure it's really strange that disabling mgmt eth makes the thing work again... i bet it's something wrong with big endian...
<Ansuel> maybe the counter
<Ansuel> and after a while sequence gets messed up and switch stop ack packets
<Ansuel> i will prepare a dissector so you can inspect packet easily
<Mangix> OK
<Mangix> unfortunately I have no oscilloscope
<Mangix> and IIRC, kabel still had issues with it
<Mangix> in the meantime, I'll try to build 6.6 and test to see if I still actually get the ports being unusable behavior
gch981213 has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
minimal has joined #openwrt-devel
<Mangix> kernel 6.6, still goes down
<damo22> TP-Link Archer MR200 v6 is a piece of crap, no FCCID and its put together with hot melt glue, only one screw to hold it together, but the rest are plastic clips that are impossible to take apart without a special tool
<damo22> i almost broke the case trying to crack it open