guifipedro_ has quit [Ping timeout: 480 seconds]
shibboleth has joined #openwrt-devel
guifipedro has joined #openwrt-devel
<aparcar[m]> stintel: ping
<aparcar[m]> do you have a running ucentral setup?
bluew has quit [Quit: Leaving]
shibboleth has quit [Quit: shibboleth]
<Tusker> ok, got the DSA switch working on the T30-W - behaves differently to the XTM330 (annoyingly)
<aparcar[m]> Tusker: what target?
<Tusker> mpc85xx / p1020
<Tusker> just working on 02_network so it comes up sensibly
<Tusker> using the wrt32x wan/lanX naming convention
Tapper has joined #openwrt-devel
<Tusker> cool, that works well
<Tusker> now to get the flash working
Ansuel__ has quit [Quit: Probably my PC crashed or time to sleep.]
goliath has quit [Quit: SIGSEGV]
Tapper has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
victhor has quit [Ping timeout: 480 seconds]
Tapper has quit [Ping timeout: 480 seconds]
dangole has quit [Ping timeout: 480 seconds]
romany has quit [Quit: Ping timeout (120 seconds)]
romany has joined #openwrt-devel
romany2 has joined #openwrt-devel
romany has quit [Ping timeout: 480 seconds]
romany2 is now known as romany
Slimey has quit [Remote host closed the connection]
pmelange has joined #openwrt-devel
romany4 has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
romany has quit [Ping timeout: 480 seconds]
romany4 is now known as romany
rmilecki has joined #openwrt-devel
danitool has quit [Ping timeout: 480 seconds]
nitroshift has joined #openwrt-devel
<Tusker> stintel: I noticed that you had a "rename" issue with eth3 on your switch config, it can be "solved" by changing 02_network appropriately - https://github.com/dmascord/openwrt/commit/b4bd6e9bfe371065d74960248848a47fe41b895d
dedeckeh has joined #openwrt-devel
ecrn has joined #openwrt-devel
digitalcircuit has quit [Quit: Signing off from Quassel - see ya!]
digitalcircuit has joined #openwrt-devel
ecrn has quit [Remote host closed the connection]
Tapper has joined #openwrt-devel
ecrn has joined #openwrt-devel
<ecrn> hi, I have a linksys MR9600 router, which is based on bcm4908 - from what I found, there is no openwrt support for it as of now, but there is for some other routers based on the same soc
<ecrn> for example, netgear R8000P
<ecrn> the radio is based on bcm43684
<ecrn> is there any hope to run openwrt on it? I have UART connected and the modules loaded are https://nopaste.net/ANj915WdRY
<ecrn> the vendor firmware image which I downloaded from their website seems to have similar layout to the supported routers on bcm4908
<PaulFertser> ecrn: you can try running initramfs image and see what works and wht doesn't.
<ecrn> PaulFertser: with openwrt kernel and basing on some device tree from other bcm4908 routers?
<PaulFertser> ecrn: yes
rua has quit [Quit: Leaving.]
<ecrn> the CFE bootloader seems to be somewhat modified: https://nopaste.net/YmbDZapZ9X - I managed to load a file to RAM with tftp, so probably can try it
ecrn has quit [Remote host closed the connection]
ecrn has joined #openwrt-devel
<Tusker> hmm... looks like cfg80211 doesn't like -fcommon and refuses to load. Is there a way in an individual device to specify gcc level and needed kernel modules ?
<mangix> Nope. Just target
<slh64> ecrn: there is zero support for bcm43684 and I wouldn't expect that to change either
<ecrn> slh64: ok, thank you
<ecrn> so the router purchase was a really terrible decision to make: the vendor firmware not only expects one to use a smartphone app for setup (normal web interface is available only after clicking a hidden button on the served page - we had to call product support for that), but also you can't set even the tx power and channel, not to mention client modes for wifi, wds or more advanced bridging
<ecrn> and the USB is only for storage devices, no modem support, even with rndis/cdc_ecm modems
rua has joined #openwrt-devel
<russell--> ecrn: sounds like you should return it while there is still time
<mrkiko> ecrn: router name / model?
<mrkiko> ecrn: sorry, I did read.
<rsalvaterra> ecrn: Return that crap while there's still time.
<rsalvaterra> Oh, russell-- already said it. :)
<dhewg> rmilecki: do my patches fix the issue on your end?
rua has quit [Quit: Leaving.]
pmelange has joined #openwrt-devel
rua has joined #openwrt-devel
ecrn has quit [Quit: Page closed]
dangole has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
goliath has joined #openwrt-devel
rua has joined #openwrt-devel
<stintel> aparcar[m]: yes
<stintel> Tusker: I solved it with rearranging things in dts
Tapper has quit [Ping timeout: 480 seconds]
rmilecki has quit [Quit: Konversation terminated!]
Tapper has joined #openwrt-devel
Guest2049 is now known as ynezz
ynezz is now known as Guest2868
Guest2868 is now known as ynezz
<pmelange> Hi, is there any change of getting a 21.02.1 release in the near future?
<pmelange> *change -> chance
<jow> sure
<jow> usually 4-8 weeks after the first release
<jow> or 8-12 weeks rather
victhor has joined #openwrt-devel
indy has quit []
<embargo> Good morning, guys. I would like to ask, is there any chance to have the VRF implemented? So far I have found this FR https://bugs.openwrt.org/index.php?do=details&task_id=2344
rua has quit [Ping timeout: 480 seconds]
indy has joined #openwrt-devel
rua has joined #openwrt-devel
<pmelange> jow: we are already at 6 weeks, so I guess that means soon. We (freifunk-berlin) are looking forward to the kernel version and most of all the recent change to the olsrd init script in openwrt-routing. Thanks for the info.
hitech95 has joined #openwrt-devel
<hitech95> Hi, how can I specify a dependency of a kernel module? The module has been implemented on 5.3+ so I would like to show (in menuconfig) that only on 5.3+
Tapper has quit [Read error: No route to host]
Tapper has joined #openwrt-devel
<jow> hitech95: DEPENDS := @LINUX_X_Y or @!LINUX_X_Y or @(LINUX_X_Y||LINUX_X_Z)
<jow> to make it future proof it is usually best to use inverted depends, so depend on Linux not being one of the older supported versions
<jow> e.g. DEPENDS := @!LINUX_4_14
<hitech95> jow: so for 5.3+ i should do @(LINUX_5_4||LINUX_4_10), what is the oldest version of the kernel on openwrt? 4.14?
<hitech95> Typo @(LINUX_5_4||LINUX_5_10)
<jow> that will fail once linux 5.11 is added
<jow> better to rule out supported older kernels
<jow> you can look those up in include/kernel-version.mk
<jow> master currently defines 5.4 and 5.10
<jow> 21.02 defines 4.4, 4.9. 4.14 and 5.4
<jow> so you likely want @!(LINUX_4_4||LINUX_4_9||LINUX_4_14)
<hitech95> jow, yea, went with that. I'll see is this patch will work.
<hitech95> BTW I'm having an issue with developing on luci. (modal rendering) I'll have to ask you some infos about why :)
<jow> hitech95: sure, feel free to pm me
pmelange has left #openwrt-devel [#openwrt-devel]
<hitech95> jow, I'll provide the sources once I will cleanup them. But the issue I'm facing is with the network configuration modal. The body is always white whatever I try to do.
<hitech95> I've tried to add a new tab called "status", in this way a virtual protocols (yes I also have patched the Protocol class) can show some additional info.
<hitech95> (I'm on MAP-T so a lot of stuff is hidden in the UI, so i'm adding them: https://ibb.co/Jtgzwtp)
<hitech95> I've created a Status tab in the network configuration modal.
<hitech95> Then I've tried to write something in it, but the modal is always white even if a console.log() functions showed me that the "body render" was called. https://ibb.co/D4JTbvp
indy has quit []
minimal has joined #openwrt-devel
indy has joined #openwrt-devel
nitroshift has quit [Quit: Gone that way --->]
<jow> hitech95: yeah, best is if you can provide some code to look at
<jow> probably simple to fix
jbowen has joined #openwrt-devel
Ansuel has joined #openwrt-devel
Slimey has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
fda has joined #openwrt-devel
robimarko has joined #openwrt-devel
<robimarko> I saw that somebody was mentioning Unifi 6 LR here
<robimarko> What is the HW inside of it?
<robimarko> I am looking for something decent to run AX until the IPQ807x is stable enough
fda- has quit [Ping timeout: 480 seconds]
<Ansuel> robimarko mediatek probably
<stintel> robimarko: Ansuel: yes Unifi 6LR is mediatek
<stintel> few things to note: opening the case means breaking the case, afaik, and the 2.4GHz radio is wifi4
<Ansuel> wth only 5ghz is wifi ax o.O
<robimarko> stintel: Ok, so its not for me then as I always mess with stuff
<Ansuel> that doesn't make sense considering the improevemtn done by ax on 2.4... who design these device LOL
<robimarko> stintel: Is it MT7622 at least?
<Habbie> stintel, somebody recently suggested to me, 'for recent mobile phones', take some cloth and a 70C iron to weaken the glue
<stintel> Ansuel: yeah that's a bit of a bummer :(
<stintel> robimarko: it's the mediatek target. not ramips
<stintel> or did you mean wifi?
<Ansuel> Habbie: hot air gun and pray that it doesn't melt ?
<robimarko> Ansuel: I assume they reused existing SoC-s or a cheap radio fo 2G
<Habbie> Ansuel, well that's the beauty of using something with a thermostat, i guess
<robimarko> stintel: MT7622 is the ARM SoC, so I suppose it uses that then
<Habbie> Ansuel, i haven't tried any of these things myself
<stintel> mt7622-wmac and mt7915e is the wifi
<robimarko> stintel: Ok, so they just use the SoC built-in radio
<robimarko> Honestly, not that impresive for 180 EUR
<stintel> the 5GHz radio works well
<stintel> rx bitrate: 2161.3 MBit/s 160MHz HE-MCS 10 HE-NSS 2 HE-GI 0 HE-DCM 0
<stintel> tx bitrate: 1633.5 MBit/s 160MHz HE-MCS 8 HE-NSS 2 HE-GI 1 HE-DCM 0
<stintel> this is my intel ax100 client
<stintel> ax200*
<stintel> saturates the gigE uplink easily
<Ansuel> all offloaded ?
<stintel> offloaded?
<Ansuel> robimarko yes considering the ax9000 is 200€ now
<robimarko> mt7915e seems really good, but somehow they always crap out on the 2G radio or SoC
<Ansuel> stintel i mean about cpu load
<blocktrron> Ansuel: U6-LR does not require SW offload to saturate the ethernet uplink
<blocktrron> Even a U6-Lite can saturate the gigabit uplink using sw flow-offload
<stintel> hmmm
<stintel> I might have to redo this image ... at some point I threw iptables and everything related out of my dumbap builds
<blocktrron> unless someone figures out how to load firmware on the aquantia PHY to enable N-Base-T, i would opt for the U6-Lite (or Cudy X6 / Tenbay if 2.4GHz 802.11ax is required)
<stintel> blocktrron: it's the reason I bought the U6LR - at home I have NBaseT so I can test stuff
<robimarko> Which AQR is it using?
<robimarko> I assume one of the Gen4?
<blocktrron> No, older one
<blocktrron> AQR112C afair
<robimarko> Ok, that is fine for 2.5G
<robimarko> But the FW is a bit of an issue
<blocktrron> I'm not sure if they have even wired it up to the SoC. I was not able to find it on the MDIO bus and Ubiquiti does not interact with it neither in their bootloader, nor withing Linux
<Ansuel> o.O they have hardware that is actually not usable ?
<blocktrron> The Ubiquiti product selection is a mess
<Ansuel> prograde pricevy stuff so i assume they don't care
<Ansuel> having extra stuff not working
<blocktrron> Their AX gear was discontinued while they were selling off the discontinued products to their loyal fans thru early access
<Ansuel> robimarko: btw i think no luck about testing if switch interrupt is connected to ipq806x... don't know if i tested it right... but by enabling every interrupt trigger for the switch no interrupt was detected.... i added to qca8k every free interrupt for 32 to 400... 0 with everything.... (could be that i did something wrong)?
<Ansuel> it's sad considering how good it would work with interrupt...
<Ansuel> stupid move IMHO
<robimarko> Ansuel: I dont think that they connected the pin on most boards at all
<Ansuel> I will do it myself !??!?!... what if they connected the pin to the gmac interrupt ?
<Ansuel> still can't understand why that much interrupts are output by the gmac
pmelange has joined #openwrt-devel
<Ansuel> i mean i think nss have his separate interrupt... so whta it's used for ?
<Ansuel> dma ?
Tapper has quit [Quit: Tapper]
fda- has joined #openwrt-devel
cp-- has quit [Quit: Disappeared in a puff of smoke]
fda has quit [Ping timeout: 480 seconds]
cp- has joined #openwrt-devel
<stintel> jow: can we drop the luci-app-firewall dep from luci? I'd like to test firewall4 but keep luci
<mangix> robimarko: I bought the Linksys E7350 precisely because of 2.4ghz ax
<mangix> Fastest wifi I've had.
<robimarko> That one looks interesting because of dual MT7915-s
Dgrey has joined #openwrt-devel
<mangix> Yep. It's amazingly stable compared to the other mt76 wifi chips. Probably because mediatek employees work on the driver this time around.
<mangix> I'm still unable to get 160MHz working though...
<stintel> works for me on the U6LR
<stintel> jinxed it
<stintel> Thu Oct 14 17:30:08 2021 daemon.notice hostapd: dfs_downgrade_bandwidth: no DFS channels left, waiting for NOP to finish
<stintel> and bye bye 5GHz bssid
<stintel> sigh
<mangix> I get a different DFS error
<stintel> this was right after radar detection
<mangix> Speaking of DFS, 6ghz will not have it AFAIK
<Ansuel> can't wait for wifi 7 that will just use 2.4 5 and 6 at the same time...
<Ansuel> datarate will be a meme
<stintel> can't wait for 6E APs to arrive in EU
<stintel> because yes, 5GHz is basically a joke due to DFS
<Habbie> my ISP is currently widely advertising the next evolution in wifi
<Ansuel> not if you.... hack the reg LOL
<Habbie> they call it
<Habbie> 'wifi 2.0
<Habbie> '
<stintel> lolwut
<Habbie> it's so dumb
<mangix> It's funny. I'm the only one around my area that uses channel 100
<mangix> Everyone uses non DFS channels
<mangix> Which kinda kills the point.
<Habbie> the point is using that capacity as long as it's safe?
<stintel> does the mt7915 in the U6Lite support 2.4GHz?
<stintel> could do a U6LR with 5GHz 11ax and then a U6Lite with 2.4GHz 11ax if that's the case
<stintel> and keep the other radios for legacy non-11ax clients
<stintel> ok, somewhat expensive, but at least no QCA junk needed then
<mangix> stintel: wifi says yes. That's quite surprising though
<mangix> All other mt7622 devices do not.
<mangix> Habbie: the point is more bandwidth that 2.4ghz. but there are only 2 non DFS channels
<Habbie> oh only 2, wow
<mangix> 36 and 149
<lemmi> i rarely have dfs events here. and they should be more firmware issues than anything else
<lemmi> there is just no radar here operating at that frequency
<lemmi> (at least to my knowledge)
<PaulFertser> If mt7915 is dbdc why not all the vendors are using it for 2.4 GHz?
f00b4r0 has joined #openwrt-devel
<PaulFertser> Probably only MT7915D supports it.
<Slimey> a good SDR can tell you if you have things in the 5Ghz band
<PaulFertser> mangix: but is it probably 2x2 on both bands? Can you paste the iw list output please?
<f00b4r0> is 'chunkeey' usually around?
<Habbie> f00b4r0, don't think i've ever seen a chunkeey in here
<f00b4r0> k
<f00b4r0> thx
<PaulFertser> mangix: thank you! stintel see, you probably do not want 2x2.
<stintel> PaulFertser: why not ?
<stintel> for 2.4GHz I could care less
<PaulFertser> stintel: but it's for 5 GHz too.
<stintel> PaulFertser: I have a U6LR for 5GHz
<stintel> I am considering the U6LR for 2.4GHz 11ax
<PaulFertser> stintel: ah, I see now
<stintel> ehr. U6Lite*
<stintel> for 2.4
<stintel> actually, I've already ordered one :P
<mangix> Wonder when mt7921 units will arrive.
<stintel> so it looks like when I boot with HE160 it doesn't work, previously I was running with HE80 then changed to HE160 and then I got 160MHz working
<Slimey> heh 00000000: BADB100D
<stintel> ah, HE160 working again, I had channel 100, changed that to auto and now it's working
<robimarko> stintel: Isnt U6 Lite also AX only on 5G
<robimarko> And for 2.4 they are simply using the SoC which is N only
<stintel> I will know soon :P
<robimarko> stintel: Its for sure just old SoC WLAN, as its advertised as 300Mbps 2x2 so N norm
<stintel> either way, the U6 Lite can stay at my parents' place to replace the crashing ath10k junk, I'll take the U6LR home where I have NBaseT
<stintel> so fingers crossed someone can figure out how to make it work at 2.5GbE
<robimarko> stintel: Have you seen the U6 Pro?
<robimarko> They have only 2 terrible FCC internal pictures so I have no idea about the HW but it looks extremely similar to the LR
<stintel> have not
<robimarko> Its actually a bit cheaper then the LR
fda- has quit [Quit: ZNC - https://znc.in]
fda has joined #openwrt-devel
<Slimey> heh punks want to use CobraNet >:|
pmelange has left #openwrt-devel [#openwrt-devel]
jbowen has quit [Remote host closed the connection]
<Slimey> i should reply that im not sure if my ethernet cables are oxygen free for that
fda- has joined #openwrt-devel
<mangix> PaulFertser: that’s not the ubiquiti unit
jbowen has joined #openwrt-devel
<PaulFertser> mangix: but it's from mt7915d, right? So if the ubnt unit has the same chip it should be the same.
<mangix> No idea if it’s the same. Mine does 2.4ghz. I don’t think the u6lr one does.
<PaulFertser> mangix: interesting, I saw stintel asking "does the mt7915 in the U6Lite support 2.4GHz?" and you answered "stintel: wifi says yes." so I assumed it's an established fact.
fda has quit [Ping timeout: 480 seconds]
<mangix> i thought it was mt7622.
<PaulFertser> It also says MT7603E
<mangix> And it says wl2 has two chips???
<PaulFertser> https://wikidevi.wi-cat.ru/images/a/a0/U6LITE-reverse-below.jpg shows just two feeders going out from mt7915 chips
aiyion_ is now known as aiyion
<PaulFertser> stintel: so probably it's not the device you want to pay so much for, looks like mt7603 for 2.4 GHz :/
<robimarko> PaulFertser: That is what I was thinking, it has to be some old IC as its only advertised as 300Mbps
<PaulFertser> robimarko: right :/ somehow I assumed it got mt7915d, probably wasn't reading the discussion with enough attention, thank you.
<mangix> Quite strange
<robimarko> UBNT does have any Unifi dual band AX devices
<robimarko> Its all AX on 5GHz only
<PaulFertser> Better get rt3200 for half the price
<mangix> or rt1800
<mangix> the rt series are cheaper versions of the linksys versions AFAIK
<stintel> PaulFertser: too late :P
Ansuel has quit [Ping timeout: 480 seconds]
<stintel> well these alternatives can't be PoE powered so are not an option for me
minimal has quit []
mrkiko has quit [Ping timeout: 480 seconds]
mrkiko has joined #openwrt-devel
mrkiko has quit []
danitool has joined #openwrt-devel
pmelange has joined #openwrt-devel
eck has quit [Quit: PIRCH98:WIN 95/98/WIN NT:1.0 (build 1.0.1.1190)]
eck has joined #openwrt-devel
<fda-> has still nobody an idea what causes this message during every reboot? "daemon.notice procd: /etc/rc.d/S95done: Command failed: Not found"
<Habbie> fda-, ls -al /etc/rc.d/S95done
<Habbie> fda-, please :)
<fda-> Habbie: its a link /etc/rc.d/S95done -> ../init.d/done
<fda-> and the done does exists too
<fda-> (after finished reboot)
<stintel> the "Command failed: Not found" is not about the script itself, rather something called from within the script
<fda-> as i have this on multiple devices i think its not because of my configuratiion
<stintel> you get the same if you stop a service that isn't running
<stintel> root@ap0:~# /etc/init.d/snmpd stop
<stintel> root@ap0:~# /etc/init.d/snmpd stop
<stintel> Command failed: Not found
<stintel> e.g.
<Habbie> fda-, can you ls -al /etc/init.d/done for us?
<Habbie> stintel, ah
<fda-> its a 263 Bytes executable file
<stintel> it's a mildly annoying, useless message
<fda-> /etc/rc.d/S95done start ; echo $? -> 0
<fda-> can you do a logread | grep S95
<stintel> or `logread -e S95`
<stintel> it could be something in /etc/rc.local which is called from the done init script
<stintel> what I would probably do is change the shebang line to /bin/sh -x /etc/rc.common and reboot
<fda-> in the rc.local i have 4 lines starting wiht /etc/init.d/
<stintel> I think this should learn you where the message comes from
<stintel> fda-: ok, it's becoming very likely, see my above example with snmpd
<stintel> are you stopping / disabling services there perhaps ?
<fda-> wait. on another device with also this message there is only "wsdd2 restart ; exit 0" in rc.local
<fda-> (with full path)
<stintel> well, if you restart a service that is not running you would also trigger that message
<fda-> well ...
<fda-> wsdd2 script is not abel to start as i does not detect ksmbd/samba4 properly and fails. so i added a "restart" to rc.local...
<stintel> just change it into start
<fda-> i tested with stop+restart and i got this message ... thx
<fda-> maybe it should be change, maybe "not running" or so
<stintel> you should probably use /etc/init.d/wsdd2 status and do start or restart depending on the outcome
<fda-> i tested, "start" when its running does not cause an error and no 2 daemons
<fda-> but it needs real 0m 5.24s
<stintel> well, the best advice is probably to actually fix wsdd2 instead of using workarounds
<fda-> (5 seconds, regardingless of running or not
<fda-> that would be the best :)
<stintel> but you could do "/etc/init.d/wsdd2 status > /dev/null && /etc/init.d/wsdd2 restart || /etc/init.d/wsdd2 start"
<stintel> probably
<fda-> i know wsdd2 fails always: "daemon.err wsdd2: samba36/4 or ksmbd is not running, can't start wsdd2!"
<fda-> this is on devices with samba and also with ksmbd
<fda-> but anyway, wsdd2 segfaults currently and does not run at all ^^
<fda-> seems like much work
<fda-> so i'be running nmbd 2oo
<stintel> afaik strace is not useful for SIGSEGV
<stintel> try running it in gdb or gdbserver
<fda-> i never used it or know how..
<fda-> just selecte the package?
<stintel> it's documented in the wiki
aleksander has quit [Quit: Leaving]
<fda-> ill check
<fda-> you know the source of "Command failed: Not found"? i only found ./feeds/luci/applications/luci-app-commands/luasrc/view/commands.htm
<stintel> I don't
<fda-> or is it "msgid "Command failed"" in translations?
<stintel> I suspect it comes from procd and I doubt it procd reads luci translations
<stintel> as I said "mildly annoying" because no clue where the heck it actually comes from
<fda-> grep shows bin/ubus: Übereinstimmungen in Binärdatei
<fda-> https://git.openwrt.org/?p=project/ubus.git;a=blob;f=cli.c;h=81591ec44d11aa053d3461d774d58e59b0bfbbe0;hb=HEAD#l649
<fda-> funfact: you could STOP a server many time without any error. and START as much as you like. restart also, EXCEPT it is not running
<stintel> nah, if you stop when not running it will error
<fda-> not always ^^
<stintel> anyway, that error message should probably be Command %s failed: %s\n", cmd, ubus_strerror(ret)
<stintel> to make it at least useful
<fda-> wsdd2 dows segfauls as you know now. when i START and then STOP there is no error. additional STOP shows the error
<fda-> so the state seems not handled properly
<stintel> do know that you can also trigger the same message like this: root@ap0:~# ubus call foo bar
<stintel> Command failed: Not found
<stintel> so it's not only for services, it's more general
<stintel> keep that in mind in case you decide to send a patch :)
<fda-> "service wsdd2 status" shows "running"
<fda-> im writing not very good comments
<Habbie> re SIGSEGV and strace: -sometimes- helpful information comes out, but not often
<fda-> i cant read the stracktrace, so i just added it to issue...
<fda-> the ""Command failed" maybe could also be another error as "if (ret > 0)" is before
<fda-> so im not sure
<stintel> would be interesting to know if the problem was introduced by gcc 10/11 or musl 1.2
<fda-> i can say wsdd2 ran some weeks ago on all my devices
<stintel> that's not very useful
<fda-> somone here in thr channel mentioned
<stintel> so you should either go back to an image from some weeks ago, or do the effort to bisect what is actually causing it
jbowen has quit [Quit: leaving]
<fda-> stintel: i could checkout an old revision...
<fda-> as i wrote in the issue: some time ago it workes on all my 3 devices
<fda-> now it runs only on e8450
<fda-> e8450 has kernel 5.10, the other 5.4
<stintel> unfortunately that's not very helpful
<stintel> a lot changed between now and some time ago
<fda-> i have still some old .image files. how does it help?
<stintel> it probably won't help much
<stintel> but it could be a starting point for git bisect good
<fda-> i always build my images myself. with the opkg, could there be old versions be downloaded?
<fda-> i think i will try gdb first
robimarko has quit [Quit: Page closed]
<fda-> stintel: can you help me when gdb shell is running? compiling is still running
<fda-> should this be replaced by wsdd2? gdbserver :9000 /bin/ping example.org
<mangix> stintel: I believe the issue has to do with GCC11
<mangix> wssd2 crashes at what looks like a malloc call.
<mangix> I think because of some asprintf call
<mangix> I bet compiling with O2 is enough to fix.
<fda-> mangix: helps gdb to find it?
<mangix> What?
<fda-> if asprinf is the cause
<fda-> TARGET_CFLAGS += -O2 to disabel optimisation?
<mangix> No
<mangix> There's an example in the zstd Makefile
<mangix> Os has to be stripped out
<fda-> TARGET_CFLAGS:= $(filter-out -O%,$(TARGET_CFLAGS)) -O3
<mangix> Right.
<mangix> Replace O3 with O2.
<mangix> So in the other project that broke with Os, one of the solutions was to add a pointer. No idea if applicable here.
<fda-> im currently compiling with -O0
<mangix> In any case, that was a GCC bug. Might also be the case here. Who knows.
<fda-> i can tell it soon
<fda-> write the flash takes on my old test device 1 minute for the 8mb
<fda-> mangix: still do_page_fault(): sending SIGSEGV to wsdd2 for invalid read access from 00000001 with -O0
<fda-> but i have now gdb included
dedeckeh has quit [Remote host closed the connection]
<fda-> mangix / stintel gdb connections is working. and now?
<fda-> i see Reading symbols from ./build_dir/target-mips_24kc_musl/wsdd2-2021-08-09-bcc51c3e/wsdd2...
<fda-> helps this??
<fda-> 0x77fc176c in memchr (src=<optimized out>, c=0, n=2147483647) at src/string/memchr.c:16
<fda-> 16 for (; ((uintptr_t)s & ALIGN) && n && *s != c; s++, n--);
<fda-> this seems source of musl, in the #ifdef __GNUC__ block
<mangix> Can you run bt?
<fda-> what is bt?
<fda-> mangix: ah, its a command: https://pastebin.com/E8e7V5NQ
<fda-> its also after the UUID part like in strace
<fda-> :(
<mangix> stintel: so time_t bug
<mangix> lovely
<fda-> in a asprintf of wsdd2 as you told..
<stintel> I suspect my DAWN issue might be the same
<fda-> ah, it still runs on my E8540, which is the only aarch64
<mangix> goddamn apt is so slow
<fda-> switch to dnf ;-)
<dwfreed> dnf is short for "does not finish"
* dwfreed ducks
<mangix> dwfreed: false
<mangix> fda-: found the bug. probably...
<fda-> great! im also searching where to change the format-thing but i have no clue :)
<fda-> so its aa bug in wsdd2 - caused by a muslc bump?
<mangix> so strange that llu is used. time_t is signed.
<mangix> fda-: guess so
<stintel> s/caused/exposed/ ;)
<mangix> stintel: it's just more proof that format functions in C are trash
<fda-> mangix: i will test, compile+flash takes some minutes
<stintel> mangix: or that people don't know how to write code :)
<stintel> that includes myself lol
<mangix> stintel: no they're actually trash. C++ projects have mostly transitioned to fmtlib, which implements Python style format strings. Much easier to use.
Ansuel has joined #openwrt-devel
<Ansuel> robimarko: ping ?
<fda-> damn, openwrt does not automatically applies patches if there is a pathes/ directory was created :\
<fda-> .. even if correctly written
<mangix> uhhh
<mangix> sounds wrong
Ansuel_ has joined #openwrt-devel
<fda-> should it?
<fda-> then i did something wrong
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<karlp> depends what directory, its not magical.
Ansuel has quit [Ping timeout: 480 seconds]
<karlp> mangix: wht does "python style" format strings even mean? there's _many_ styles :)
<mangix> karlp: fmt::format("test: {}/{}/{}, a, b, c);
<mangix> the type is automatically deduced
<fda-> mangix: thx, i have the old patch now applying. there where tabs replaced by spaces
<fda-> mangix: wsdd2 works with the 1st patch version now
<fda-> (it still compiled with -O0)
<fda-> thx!
<fda-> and afaik openwrt uses musl 1.2.2
Acinonyx_ has joined #openwrt-devel
nindoja has quit [Remote host closed the connection]
Acinonyx has quit [Ping timeout: 480 seconds]
bluew has joined #openwrt-devel
<mangix> fda-: -O0 is not needed
<fda-> mangix: im now compiling without for another device (avm1200) to test
pmelange has left #openwrt-devel [#openwrt-devel]
<Tusker> hmm... are there any examples of a pinmux for gpio-led ? ie, two GPIO that control one LED, and it has various on/off states to control colour ?
<Tusker> at the moment, the off state of one colour is opposite of the other, so that it causes the LED to light in one colour
<fda-> mangix: wsdd2 works for avm1200/ipq40xx/generic too again with your patch!
<Ansuel_> any idea what would cause a missing watchdog ?
<Ansuel_> i just notice ipq8064 doesn't have a watchdog defined at all...
<fda-> me? i dont know about the watchdog
<Ansuel_> and the driver is used with the wrong reg
<Ansuel_> just asking to everyone
<mangix> Ansuel_: no idea
fda has joined #openwrt-devel
fda- has quit [Ping timeout: 480 seconds]
<fda> a new error message during boot: "E8450 : jail: can't read from parent" just after pppoe module load and before collectd network plugin loading
<fda> maybe interesting for dangole
<Ansuel_> wow we are starting to use jail '
<Ansuel_> ?
<mangix> it got enabled by default
<fda> i have no idea from which package it is :\
<fda> with dnsmasq i had also jail-proclems and i had to add a direcotry manually to init script
Ycarus_ has joined #openwrt-devel
<dangole> fda: in default installation it would be either dnsmasq, hostapd/wpa_supplicant or sysntpd which use any jailing by now. umdns and transmission are optional packages which make use of jail with seccomp.
<mangix> dangole: btw I tested transmission with that patch. works. I accidentally let the debiab mips ISO finish downloading
<fda> thx dangole, i will restart these. i have no umdns or transmission. dnsmasq uses unbound, sysntpd a local ntp
<dangole> mangix: beautiful. i didn't believe we would ever get it to work with wolfssl...
<mangix> the more confusing part is why options.h is an external header
<dangole> wolfssl is just weird
<fda> sysntpd does not write anything to syslog, dnsmasq no erors now on restart. how can i restart wifi?
<dangole> fda: /etc/init.d/wpad restart
<mangix> dangole: weird history too.
<dangole> mangix: i know it was called cyassl earlier and we had a long history of carrying it around before switching to use it by default. looking at the code I'd prefer mbedtls, by magnitudes.
Ycarus has quit [Ping timeout: 480 seconds]
<mangix> dangole: I do too. Looking at the wolfSSL code it looks like at some point it was forked from OpenSSL.
<fda> thx, also no jail errors, but multiple "hostapd: nl80211: Failed to remove interface wlanX-X from bridge br-lan: No such device" and "hostapd: rmdir[ctrl_interface=/var/run/hostapd]: Permission denied" im not sure if they are "normal"
<mangix> dangole: hah. even their docs suck. they say to include version.h first and in other places options.h should be first.
<dangole> anybody knows who is openwrt.club btw?
<Tusker> some guy called 萝卜maybe ?
<fda> like https://repo.xposed.info/module-overview , i which a asia filter for it as i cant read it
<Ansuel_> openwrt.club.......
<mangix> openwrt seems big in china
<fda> china has just 1400mio people
<slh> ...and lots of incentive to get 'real' internet, beyond the walled garden