minimal has quit [Quit: Leaving]
danitool has quit [Ping timeout: 480 seconds]
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
SlimeyX_ has joined #openwrt-devel
valku has quit [Remote host closed the connection]
valku has joined #openwrt-devel
SlimeyX has quit [Ping timeout: 480 seconds]
SlimeyX_ is now known as SlimeyX
valku has quit [Remote host closed the connection]
dangole has quit [Ping timeout: 480 seconds]
cbeznea has quit [Ping timeout: 480 seconds]
cbeznea has joined #openwrt-devel
srslypascal is now known as Guest1064
srslypascal has joined #openwrt-devel
Guest1064 has quit [Ping timeout: 480 seconds]
hanetzer has joined #openwrt-devel
Tapper has joined #openwrt-devel
cbeznea has quit [Ping timeout: 480 seconds]
cbeznea has joined #openwrt-devel
cbeznea has quit [Ping timeout: 480 seconds]
cbeznea has joined #openwrt-devel
danitool has joined #openwrt-devel
danitool has quit [Remote host closed the connection]
MaxSoniX has joined #openwrt-devel
robimarko has joined #openwrt-devel
<aparcar[m]> is there a way to compile specific packages statically?
<aparcar[m]> Mangix: -^
hanetzer1 has joined #openwrt-devel
hanetzer has quit [Ping timeout: 480 seconds]
Piraty_ has quit [Ping timeout: 480 seconds]
Piraty has joined #openwrt-devel
Piraty has quit []
cbeznea has quit [Ping timeout: 480 seconds]
cbeznea has joined #openwrt-devel
<f00b4r0> rmilecki: seems mtd/linux.git is back alive, do you still want to ping them or should I? :)
<robimarko> f00b4r0: Use the momentum
<f00b4r0> heh :)
<f00b4r0> let's see if I can dig up the patch in my email history ;P
<robimarko> Everybody is back from ELCE so its a short window of activity
<f00b4r0> hmm, patchwork says the patches are now delegated.
<f00b4r0> I suppose there's no point in pinging just now then
<f00b4r0> still state "NEW". Older patches in that state are from last May
<f00b4r0> oldest*
sorinello has quit [Quit: Leaving]
sorinello has joined #openwrt-devel
<robimarko> Patchwork is hit and mis when it comes to the status
<Habbie> alpine ports has 'abump', which bumps the version of a port and updates the hash after fetching the tarball - is there a similar tool hiding somewhere in the openwrt tree?
<robimarko> make package/path/check FIXUP=1 should update the hash
<stintel> no, bump manually, make package/foo/download PKG_HASH=skip; make package/foo/check FIXUP=1
<Habbie> ah neat
<stintel> anyone ran into situation where mikrotik routers ignore DHCP requests from OpenWrt devices ?
<f00b4r0> robimarko: frankly i find it a bit depressing that a relatively simple patch doesn't receive a single comment in ~2 months. I'd have expected the mtd subsystem to be a little more "lively" than that ;P
<robimarko> f00b4r0: well, like many subsystems there is a lack of manpower
<Habbie> stintel, that worked perfectly, thanks
<stintel> Habbie: welcome
<robimarko> Even those that are maintainers do it part-time only and not for the whole month
<stintel> pfff stupid hex S doesn't even negiotate PoE on my 10G switch
<stintel> fucking mikrotik crap
<robimarko> Maybe they got out of spec resistor
<robimarko> I personally am not fan of their 802.3at/af implementation
<stintel> got tempted to use 2 hex S to replace the APU2 at my parents' place and do a HA setup there as well
<stintel> maybe I'll just throw them in the trash instead
<f00b4r0> mail them to me :3
<robimarko> f00b4r0 has a love and hate relationship with mtik gear
<f00b4r0> haha
<stintel> reminds me of my 2 RB2011L-IN I gave those away too
<robimarko> I am slowly shifting to the hate side
<f00b4r0> same, truth told
<robimarko> With the sh*t they are pulling now
<stintel> ufff and that dreaded cli
<robimarko> Its one of the rare instances where using Winbox is faster than CLI
<robimarko> Though, I am hating them more and more due to pulling more and more stunts to prevent 3rd party SW
<robimarko> Like literaly disabling UART through RouterBoot and ignoring their usual NO_UART bit in hard_config
<rmilecki> f00b4r0: i'm chatting with mtd guys
<rmilecki> started looking at their progress this morning
<rmilecki> I didn't forget about you :)
<f00b4r0> rmilecki: cool :)
<f00b4r0> thx!
<stintel> looks like mikrotik routeros does not respond to DHCP requests from unifi AP 6 LR
<stintel> U6LR running OpenWrt
<robimarko> Thats weird, I am using ROS for DHCP
<robimarko> And its working fine with OpenWrt
<stintel> colleague had the same problem, we've been trying for a day to get his U6LR get an IP via DHCP
<stintel> he tried 2 mikrotik, same problem, then replace mikrotik with gl-inet and problem went away
<stintel> I am now trying the same at home, same problem apparently
<robimarko> Whats the ROS version?
<stintel> how do I check that in cli ?
<stintel> fuck how anti intuitive this thing
<robimarko> "/system/package/print"
<stintel> MikroTik 6.48.6 (long-term)
<stintel> just logged out and logged in again it showed in the banner
<robimarko> I am using 7.5 on RB4011 and it works fine with all of the OpenWrt devices I tried with
<robimarko> v6 worked fine before as well
<stintel> I'm seeing the usteer broadcast when sniffing on the mikrotik, not the DHCP request
<stintel> switching the U6LR to a vlan with OpenWrt DHCP it instantly gets an IP
<stintel> vague shizzle
dangole has joined #openwrt-devel
<rmilecki> f00b4r0: can you join #mtd ?
<f00b4r0> rmilecki: sure
hanetzer1 is now known as hanetzer
<stintel> 33 time=21.757 num=34 direction=rx src-mac=F4:92:BF:A9:F4:AB dst-mac=01:80:C2:00:00:00 interface=ether1 protocol=802.2 size=64 cpu=1 fp=no
<stintel> so I see traffic from the unifi on the mikrotik, I can ping the u6lr from the mikrotik, but no DHCP request
<stintel> wtf
<stintel> lan has default config, lan1 is alias on same physical interface as lan, with DHCP
Ansuel has joined #openwrt-devel
<Ansuel> erm openwrt.org is down ?
<\x> The page you are looking for is temporarily unavailable.
<\x> Please try again later.
<Ansuel> yep
<f00b4r0> rmilecki: I'm afraid this is going nowhere :(
<stintel> mariadb is stuck again
<stintel> stupid wiki
<stintel> I restarted mariadb and php-fpm
<stintel> we really need to find a better alternative for the TOH crap
<f00b4r0> run it on the client side?
* f00b4r0 hides :D
* stintel throws a server at f00b4r0
<f00b4r0> ^o^
* stintel disconnects the mikrotik and gets on with life
<f00b4r0> that's the spirit :)
<stintel> ain't nobody got time for debugging mikrotik shit :P
<stintel> could be interesting for target practice, see if i can hit from 20m with a 9mm :P
<f00b4r0> see if you can fit 9mm slugs in each ethernet port at 20m :D
<stintel> :D
<stintel> gonna take some practice :P
<f00b4r0> i'm sure you like a good challenge :)
<stintel> # iw phy0 interface add wlan0 type managed
<stintel> command failed: No such device (-19)
<stintel> welp, nothing works today
<stintel> iw phy lists phy0
* Ansuel is scared
<Ansuel> poor router
<\x> so Ansuel, about that 1k qam over VHT (NitroQAM), will that hack work on non-broadcom stuff?
<Ansuel> \x link ?
<\x> oh, i thought you have it or something lmao, i was only searching on chatlogs here
Strykar has quit [Quit: /quit]
Strykar has joined #openwrt-devel
<Mangix> aparcar[m]: depends
tlj_ has joined #openwrt-devel
tlj has quit [Ping timeout: 480 seconds]
lynxis has joined #openwrt-devel
<\x> ei robimarko, about that ipq40xx threaded napi thing, is there a recommended way of spreading the interrupts or its like whatever like this https://i.imgur.com/A6lKmHD.png
Ansuel has quit [Read error: Connection reset by peer]
minimal has joined #openwrt-devel
CrazyLemon has joined #openwrt-devel
CrazyLemon has quit []
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
rua has quit [Quit: Leaving.]
<robimarko> \x: As long as you spread them out evently it should be fine
tlj_ has quit [Remote host closed the connection]
tlj_ has joined #openwrt-devel
valku has joined #openwrt-devel
<\x> maaan, renewed MR7350 (ipq60xx) is only 40$ on amazon
<\x> kind of tempting now, hope the guy doing that 60xx support succeeds at it
<robimarko> \x: Thats a steal for that price
<mrnuke> \x: " the guy doing that 60xx support " -- I like that callname
<\x> currently it can be built with qsdk, I do wonder how usable is it. I need vlans, but if its not available i can use gre.
cbeznea1 has joined #openwrt-devel
<\x> i coped for like 6 months with gre with ipq40xx
<hurricos> jow: First seeds planted RE: making OpenWISP consume more upstream OpenWrt stuff: https://communitynetworks.group/t/openwisp-an-open-source-network-management-system/843/6
<hurricos> unrelated: anyone here familar with TI's davinci silicon? It's armv5l but with some v4l2- and media-controller-compatible silicon for h264
<hurricos> I saw https://github.com/CamWRT but nothing since then. Seems like an easily supported target with a lot of upstream Linux support.
<\x> now the question is, will 512MB be good enough for two ath11k radios mrnuke robimarko
<robimarko> \x: 2?
<hurricos> (and tons of ram on the platforms I've seen). The downside, of course, is cameras.
<robimarko> \x: Its effectively a single radio
<robimarko> As its handled by the same remoteproc
<\x> oh so its like dbdc?
<\x> nice
<\x> think thats enough?
<robimarko> 512MB is fine, AX6/3600 work fine with 512
<robimarko> We had to add a memory profile for it, but it will work fine
<\x> I want to run some services, nothing huge but unbound in recursive mode, so lets say 120MB
<hurricos> pepes: ram's OK, it's cpu that struggles pushing full bandwidth :D
<mrnuke> I my (uninformed) ipinion, 512 MB is a metric tonne of RAM for an AP
<robimarko> \x: Check the AX3600 thread for a lot of RAM graphs
<hurricos> p2020 might be fast enough, much higehr base clock I think?
<robimarko> mrnuke: Well, not really
<\x> thanks ill have a go robimarko
<robimarko> Q6 in 1G mode will eat 85MB only for itself
<hurricos> I know P1010 comes with single core, 1GHz, not sure what P2020 does
<hurricos> P2041 I saw @ 1.5GHz, twice as fast as Cavium's mips64 cores on octeon2 (which were higher freq anyways)
<hurricos> in terms of raw perf
<mrnuke> BTW, \x, I just placed an order for that linksys
<mrnuke> hurricos: hi
<hurricos> so likely just faster
<hurricos> mrnuke: Hi, I saw the PR. Talk to me about the "manage" node in ubus
<hurricos> or is that relevant as of that CI?
cbeznea has quit [Ping timeout: 480 seconds]
<\x> mrnuke I might join you in some time, its still has to travel halfway around the globe https://i.imgur.com/y6ezADr.png
<mrnuke> hurricos: which PR/CI? the manage is just the patch to shut up a port
<hurricos> got it
<hurricos> 67 or 68. I read these updates bleary while in the morning, then go to work and then come back and forget :^)
<mrnuke> robimarko: Someone mentioned that qca-nss-dp might not be necessarry with a realtek-phy. Is there an alternative driver that I completely missed?
<robimarko> mrnuke: obviously not
<robimarko> You must use it
<robimarko> As its the ethernet controller driver
<robimarko> \x: Found that Italian Amazon had that Linksys for 55EUR
<robimarko> Had to order it obviously
Piraty has joined #openwrt-devel
<hurricos> booting up switch now. Unfortunately due to dead serial I have to test the switch from the network, using an 11s bridge on an AP which is powered by the switch :^)
<mrnuke> hurricos: Oh, CI58 ? Story time> CI58 is the realtek-poe v1.0. The packages feed is not on v1.0. There is another PR for that. So many PR's, so little time
<hurricos> Yes, that's the PR I was CC'd on
<hurricos> sorry
<\x> oh yeah knowing linksys, will this have any issues with country code something?
<hurricos> that's what I'm trying to refer to and judge quickly before I have to go do more web hackery
<\x> like can you override the country embedded on ART
<\x> i heard thats an issue on linksys wrt-ac stuff
<robimarko> \x: Yeah, everybody ignores the EEPROM country
Piraty has quit []
Piraty has joined #openwrt-devel
<robimarko> As its set to WORLD anyway 99% of the time, even QCA does that
<mrnuke> hurricos: walterav also mentioned he's still getting the UFO on boot-up. That might be a mode of failure we haven't considered yet -- I have my suspicions, but it looks different than the ff-reply issue that we already fixed
<mrnuke> robimarko: Is it feasible to make the qca-ssdk use the linux PHY drivers instead of their own crapware?
<hurricos> Have a link? If it's the one I'm thinking of, then we're still just waiting on a dump
<mrnuke> hurricos: https://github.com/Hurricos/realtek-poe/issues/10 <- Is this the link you're looking for?
<robimarko> mrnuke: Yes, for Aquantia already the in-kernel driver is used
<mrnuke> robimarko: Is there a patch for that? How does that work?
<robimarko> mrnuke: No patch needed, kernel has the driver for AQR already
<robimarko> And it attaches that based on PHY ID
<hurricos> mrnuke: Yes, thanks for link.
<robimarko> SSDK has some basic support for AQR internal
<hurricos> Let's see if I can reproduce this on my h/w
<robimarko> So, it can still mess with the PHY directly
<robimarko> But untagling QCA8075 or QCA8081 would be a mess from SSDK
<robimarko> As they are poking directly
<mrnuke> robimarko: qca-ssdk/src/adpt/hppe/adpt_hppe_portctrl.c : qca_hppe_mac_sw_sync_task()
<mrnuke> robimarko: the call to adpt_hppe_port_phy_status_get() fails, an that causes the code to leave the MAC in the disabled state
<mrnuke> robimarko: that call fails because there is no SSDK driver for the phy
<robimarko> Ahh
<robimarko> In that case, it can be patched in
<robimarko> But honestly, I would rather get rid of that POS
<mrnuke> me too
<mrnuke> That would need a fresh upstream-style rewrite, right?
Piraty has quit [Quit: --]
robimarko has quit [Remote host closed the connection]
robimarko has joined #openwrt-devel
Piraty has joined #openwrt-devel
cbeznea1 has quit [Ping timeout: 480 seconds]
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
cbeznea has joined #openwrt-devel
<robimarko> mrnuke: unfortunately yes
<robimarko> Switch especially needs to be rewritten
<robimarko> But good luck without datasheets
<\x> robimarko: yeah, this is kinda nice, checked and trying to build lean for it, and it does have wifi support unlike the 128MB ipq60xx one (ZN-M2)
<\x> so for the meantime maybe qsdk it is
<robimarko> \x: What, is there seriously a 128MB model?
<robimarko> Not even QCA supports that
<mrnuke> robimarko: who needs datasheets when everything is nicely documented in the SSDK?
<mrnuke> ^sarcasm
<robimarko> almost past my sarcasm alarm
<\x> robimarko: 256 my bad
<robimarko> \x: 256 is barely doable
<\x> and no bootloader mods needed
<hurricos> mrnuke: install is clean :phew:
<hurricos> rebooting now.
<\x> as for recovery, maybe the linksys failsafe is enough, thats like 4 failed boots and you get into fiormware2
<robimarko> \x: Yeah, wouldnt have even ordered it if it wasnt 512MB
<robimarko> Though it looks like they went for the bottom tier IPQ6000
<robimarko> Do you maybe know its FCC code?
<\x> seems theres two variants or aiamadeus is wrong?
<\x> wikidevi says 6028, aiamadeus says 6000
<\x> but anyway, they put an OC on it anyway?
<robimarko> Its gonna be 6000 as official spec sheet says 1.2GHz
<\x> so hmmm
<robimarko> 1.6GHz is not supported
<robimarko> It may work on your specific example
<\x> so this is fake OC?
<robimarko> But they are binning them for a reason
<\x> keks
<\x> like new operating point but it wont hit it?
<\x> ehh its fine, ill just lower the clock if it wont do it
<robimarko> \x: Its probably gonna be unstable
<robimarko> But like I said, varies on your example and if you want to fiddle with it
<\x> ill just lower it then
<\x> yep dw, i actually do xOC stuff on desktop
<robimarko> Then feel free to try, there is nothing preventing you
<\x> ^^
<\x> how much of a hack is qsdk's kernel https://i.imgur.com/4QFsKWj.png
<\x> 4.4 is so jurassic
<robimarko> They are using 5.4 now
<\x> oooh that sounds good
<\x> ill join you guys once i have it ahaha, maybe i wont be deploying it for some time
<robimarko> Its much better than 4.4 but still your usually vendor hackshow
<mrnuke> robimarko: Should I just patch out the PHY checks from the ssdk instead of patching in linux-phy support? Or will noth be equally painful?
<robimarko> mrnuke: Whats the exact PHY?
<\x> it might take a week or so, halfway around the globe man, and it seems heavy? last time i ordered a cpu on amazon (5600G) it was only like 12$ shipping. this one is 20$, half the item price.
<robimarko> I assume it would be easier to add its ID and a basic "driver" that just fetches info from the kernel
<robimarko> \x: I am lucky then, its from EU, Amazon says its arriving next week
<\x> theres pads for spi
<robimarko> acwifi is like the best portal for scoping out HW in advance
<\x> oof, its hard to remove those heatsinks
<\x> seems some chonky ones though
<robimarko> Soldered in
<\x> i thought this is a wisoc, why are the radios external
<robimarko> I dont care, UART has a nice header
<robimarko> Its like a hybrid
<robimarko> Q6 remoteproc runs the FW and thats in the SoC
<robimarko> But the RF IC is external
<robimarko> This way you can have different specs with the same CPU
<robimarko> Or do 6E for example
<robimarko> IPQ5018 is a proper WiSoC as it has built-in single band AX radio
<\x> theyre doind proper shielding on usb3 now huh
<robimarko> Well, it started messing with the WLAN performance
<\x> >The measured standby power is 5.8W, which is not high
<\x> oof
<\x> my 40xx is like 4W maaan
<\x> but i guess its worth it
<robimarko> IPQ40xx is like a kids toy compared to IPQ60xx, let alone 807x
<\x> ill use this as my router, keep the other 40xx with nvme on it
<\x> maybe in the future i can get another ap with mpcie and then replace
<robimarko> I have 2 Edgecore EAP101-s
<robimarko> They are cool as both ports are 2.5G
<robimarko> And it uses IPQ6018 instead of the bottom tier
<\x> how much am I losing with 1.2 and 1.8GHz, it shouldnt be much aye? like not factoring overclocking yet and tbh i dont think itll matter the ethernet ports are the main bottleneck
<\x> 1.2GHz A53 is like rpi3
<\x> thats a lot for my 100Mbps internet lmao
<robimarko> For that for sure
<\x> but hey now i can ditch my old stuff and recoup costs for the upgrade
<mrnuke> robimarko: RTL8211F (according to linux)
<robimarko> mrnuke: Is there in-kernel driver?
<mrnuke> robimarko: of course there is :) !
<mrnuke> It's CONFIG_REALTEK_PHY=y
Tapper has quit [Ping timeout: 480 seconds]
Piraty has quit [Quit: No Ping reply in 180 seconds.]
Piraty has joined #openwrt-devel
<robimarko> mrnuke: Then it should be possible to just get the stuff they manually parse from "drivers" from the kernel directly
<hurricos> mrnuke: ACK
<hurricos> thanks for being patient
dangole has quit [Ping timeout: 480 seconds]
<mrnuke> robimarko: more ssdk hacking! How great a vendor quallcrom is!
<robimarko> mrnuke: To be fair, SSDK is the worst
<robimarko> Newer stuff is much better
<robimarko> Though still shit
Ansuel has joined #openwrt-devel
<mrnuke> robimarko: SSDK puts Italy's spaghetti to shame
ahf has joined #openwrt-devel
dangole has joined #openwrt-devel
gladiac has joined #openwrt-devel
<robimarko> mrnuke: Well pu
<robimarko> *put
Ansuel has quit [Ping timeout: 480 seconds]
Strykar has quit [Quit: /quit]
Strykar has joined #openwrt-devel
<svanheule> mrnuke: I bought a MP3924 PSE controller dev board; the one that's in the SG2210P v5
<svanheule> mrnuke: maybe, some day, I'll have time to write driver for that chip then. If someone else doesn't beat me to it
<mrnuke> svanheule: I2C, eh? Hopefully it's as simple as TPS23861
<svanheule> mrnuke: yup, and single chip too. The documentation seemed detailed enough to be able to write a driver without too much digging
<svanheule> mrnuke: although admittedly, maybe that dev board wasn't the best investment at this time; when I have plenty of other, more important, things on my list :^)
<svanheule> but I was too happy to see an affordable board, I suppose
Strykar has quit [Quit: /quit]
Strykar has joined #openwrt-devel
danitool has joined #openwrt-devel
Tapper has joined #openwrt-devel
MaxSoniX has quit [Read error: Connection reset by peer]
<mrnuke> svanheule: Was said board more affordable than a SG2210P-v5 ?
Ansuel has joined #openwrt-devel
dangole has quit [Remote host closed the connection]
cbeznea has quit [Read error: Connection reset by peer]
dangole has joined #openwrt-devel
MaxSoniX has joined #openwrt-devel
<svanheule> mrnuke: yes, although shippping reduced the ratio to about 2.5
<svanheule> mrnuke: board's model number is EV3924-U-00A
<mrnuke> svanheule: oh, then it was pretty cheap. Look at the bright side: You can now add PoE to _any_ switch :p
<svanheule> mrnuke: exactly! all I need is an I2C bus isolator, and a bunch of wires to hook everyting up :P
<mrnuke> svanheule: That never stopped you before :p
* svanheule glances at the modded GS1900-8 on his desk
<mrnuke> so that's where the I2C wires went
Ansuel has quit [Ping timeout: 480 seconds]
cbeznea has joined #openwrt-devel
dangole has quit [Ping timeout: 480 seconds]
robimarko has quit [Quit: Leaving]
Ansuel has joined #openwrt-devel
cbeznea has quit [Quit: Leaving.]
goliath has quit [Quit: SIGSEGV]
Piraty has quit [Remote host closed the connection]
Piraty has joined #openwrt-devel
Ansuel has quit [Ping timeout: 480 seconds]
philipp64 has quit [Read error: Connection reset by peer]
valku has quit [Quit: valku]
Ansuel has joined #openwrt-devel
Ansuel has quit [Quit: Probably my PC decided to sleep or I decided to sleep.]
MaxSoniX has quit [Quit: Konversation terminated!]
rua has joined #openwrt-devel
<KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_x86.html has been updated. (100.0% images and 99.6% packages reproducible in our current test framework.)
<pepes> hurricos: Hmm, with P2020 it was good, but you will see once you get Turris 1.1 router! :)
philipp64 has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
cbeznea has joined #openwrt-devel
cbeznea has quit []
<mrnuke> I made in the wiki a wrong techdata page (https://openwrt.org/toh/hwdata/tp-link/tp-link_sg2210p_3), It's wrong in that the version is "3" instead of "v3". I since created the correct page, but I don't know how to delete the wrong one
<dwfreed> PaulFertser is the resident wiki expert
minimal has quit [Quit: Leaving]