<mangix> wonder if it's because the DSA driver only does one interfaced
<mangix> *interface
<mangix> AFAIK stock interface assignment = eth0(WAN), eth1(LAN1-4)
<mangix> I think it's currently eth1(ALL)
<mangix> yep that's exactly it.
<mangix> I find it interesting that DSA eth interfaces have an MTU of 1502
<rsalvaterra> I have 1504 on my RM2100.
<mangix> "The DSA framework will automatically adjust the MTU of the master interface to accommodate for this extra size in order for DSA user ports to support the standard MTU"
<mangix> interesting...
<rsalvaterra> That reminds me…
<rsalvaterra> I have that MTU reservation patch in my tree.
<rsalvaterra> Yes, that one.
<mangix> yeah that one
<rsalvaterra> LOL
<mangix> ETH_SWITCH_HEADER_LEN becomes part of mtu with DSA
<mangix> guess it's 4 with mt7530
<rsalvaterra> Probably. Haven't dug in the code to confirm, though.
<mangix> different switches have different tagging protocols so it makes sense
<stintel> see "Switch tagging protocols" secrion
<stintel> unfortunately it's incomplete, but it does show that there are 2 and 4 byte DSA headers
<rsalvaterra> Yep, that explains it nicely.
<mangix> hmmm the upstream ag71xx driver is quite different from the one in openwrt. and it diverging even further.
<rsalvaterra> I don't know what's missing from the upstream driver for us to be able to use it… :/
<mangix> a bunch of stuff needed for stuff newer than ar934x
<mangix> including delays and whatnot
<rsalvaterra> Even if not upstreamable in the short term, wouldn't it be possible for us to carry the needed patches to the upstream ag71xx driver?
<mangix> IIRC this was attempted before and abandoned.
* rsalvaterra would rather deal with patches than wholesale replacement of files…
<rsalvaterra> Oh, well…
<mangix> yeah do a search for ath79: use downstream ag71xx for Kernel 5.4 on the mailing list
<mangix> hmmm rgmii delays are extremely important.
<stintel> finally set up netbox for my own infra
<blocktrron1> rsalvaterra mangix: upstream ag71xx needs massive reworks to support more than embedded-switch boards
<stintel> adding stuff to it proves my infra is getting out of hand :P
<blocktrron1> it has to be redesigned to maintain multiple shared states between the two actual GMAC blocks
<blocktrron1> given it works and upstream will most likely not progress by itself, i don't see the gain upsetting the cost
<blocktrron1> if something in ath79 should be upstreamed, it is the accident of usb-phy & hc driver.
<blocktrron1> stintel: as long as you dont maintain it like i do
<stintel> yeah once you set something like that up you should make it a habit of using it
<stintel> but I'm considering integrating my ansible plays with it
<stintel> so in this case I will have to add new stuff before I can properly use it
<blocktrron1> :)
Tapper has joined #openwrt-devel
<stintel> ooh tp-link claims they updated the GPL code for https://www.tp-link.com/en/support/download/eap615-wall/#GPL-Code
<rsalvaterra> blocktrron1: I see, not a priority, fair enough. Maybe one day our driver will replace the upstream one. :)
<blocktrron1> rsalvaterra: lol
<rsalvaterra> And now I'm off to bed. Will be back in about 7 ours, or so. :P
<rsalvaterra> *hours
<rsalvaterra> Damn phone…
<stintel> nn
danitool has quit [Read error: Connection reset by peer]
fda- has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
minimal has quit []
Tapper has quit [Ping timeout: 480 seconds]
mangix has quit [Read error: Connection reset by peer]
mangix has joined #openwrt-devel
valku has quit [Quit: valku]
<Pepes> mangix: I got very late here, but no. :-) Turris is not interested in Broadcom for now. I can confirm that we are still sticking to Marvell. I also did a small leak on OpenWrt forum, what Turris kernel developers are looking for. No planned change on this one, AFAIK. I don't want to make any doubts, etc. It's just me.
cooler has joined #openwrt-devel
cooler has quit []
nlowe has joined #openwrt-devel
nitroshift has joined #openwrt-devel
nitroshift has quit []
nitroshift has joined #openwrt-devel
Tapper has joined #openwrt-devel
nlowe has quit []
pmelange has joined #openwrt-devel
<rsalvaterra> Pepes: Thanks for the clarification! :)
rmilecki has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
srslypascal has quit [Remote host closed the connection]
srslypascal has joined #openwrt-devel
Borromini has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
cp-- has joined #openwrt-devel
cp- has quit [Ping timeout: 480 seconds]
pmelange has left #openwrt-devel [#openwrt-devel]
danitool has joined #openwrt-devel
snh has quit [Read error: Connection reset by peer]
snh has joined #openwrt-devel
victhor has joined #openwrt-devel
ecloud has quit [Ping timeout: 480 seconds]
ecloud has joined #openwrt-devel
<nitroshift> rsalvaterra, you around?
<rsalvaterra> nitroshift: o/
<nitroshift> yo! :)
<stintel> neat
<nitroshift> should i get it to replace the rt8822?
<stintel> but I avoid ordering stuff from outside EU
<nitroshift> stintel, if it's about taxes i'm not worried ;)
<stintel> it's about customs declaration being a pain in the ass
<rsalvaterra> Oooooh…!
<nitroshift> yeah, customs also are not a problem for me ;)
<stintel> another half-assed regulation from EU, just like the cookie law, gdpr, all it does is create shitty experience for 99% of europeans
<stintel> nitroshift: if customs are not a problem for you, order one extra for me
<rsalvaterra> stintel: I agree with the objective, but the implementation sucks, indeed. Data collection (for targeted advertising or any other non-functional purpose) should be strictly opt-in or outright banned.
<nitroshift> fsck! real price with shipping comes up at 30 euro :|
<stintel> rsalvaterra: I'm considering a dnsbl for sites that use annoying/slow/misleading gdpr cookie config crap
<stintel> but really EU should just make that illegal
<rsalvaterra> nitroshift: Can't fit it in an half-length mini-PCIe slot, though… :(
<nitroshift> rsalvaterra, i have a ngff to pci-x adapter
<rsalvaterra> Sure, but you're not sticking that inside a laptop. :)
<nitroshift> rsalvaterra, i'll use it in a mini pc ;)
<nitroshift> that's ngff to mini pci-x
<nitroshift> that's what i have right now
<rsalvaterra> nitroshift: You need to practice that link hygiene. ;)
<nitroshift> yeah, i hit enter right away, there's no "undo" button on irc :))
<rsalvaterra> Ironically, my Eee PC 901 has a full-length mini-PCI slot. My slowest laptop with the fastest Wi-Fi card… :P
<rsalvaterra> Especially considering it only does 100Base-T wired. :P
<stintel> geez, are you actually using that slowass thing?
<stintel> heck that's even a n270, I ditched my 330 system rather fast
<stintel> and later the d525 also
<rsalvaterra> stintel: It's a usable terminal… :)
<rsalvaterra> … and immune to Spectre/Meltdown. B)
SamantazFox has quit [Ping timeout: 480 seconds]
dangole has joined #openwrt-devel
<Borromini> anyone know of any affordable x86 mini PCs with 2x Gbe? all i'd need for a new router but hard to find it seems (or pricey)
<\x> rsalvaterra: just a heads up, best laptop motherboard for mini pcie is Thinkpad x200. #1. 3x mini pcie slots (2x full +1x half) #2. corebootable #3, soic16/soic8 (has pads for both) spi nor ;)
<\x> Borromini: maybe search for like "pfsense mini pc" on aliexpress or something, I think you can get one for as low as 70$ iirc
<nitroshift> Borromini, i'm using a converted fujitsu siemend esprimo e720
<nitroshift> it only has an intel gbe nic, but has 3 pci-x slots
<Borromini> thanks
<nitroshift> :)
<Borromini> yeah ideally i'd get something like an Optiplex USFF but no PCI-E slot etc
<\x> some small thinkstations have like a m.2 e-key (wifi mist of the times) or mini pcie (haswell and older) you can get an extra nic in on those
<\x> you can even get a realtek 8125 on mini pcie, but yeah you sometimes have to mod the case or remove that vga port cover
<\x> dont get the ones with like CNVio thing though as i dont think those accept generic pcie devices (not sure really)
<Borromini> :)
nitroshift has quit [Quit: Gone that way --->]
<\x> or you can always just use some laptop and use their mini pcie slots for that extra nic. it wont look nice but you get a machine with a UPS :D
<Borromini> :P
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<\x> oh and yeah, mt7921k has AP mode?
<rsalvaterra> \x: I don't think so… nbd?
<\x> yeah doesnt seem like it
<rsalvaterra> Right, managed and monitor only.
<rsalvaterra> Wait, it does 6 GHz too…?
<\x> rsalvaterra: mt7921 = 2.4/5, mt7921k = 2.4/5/6
<\x> "K"
<rsalvaterra> Oooh, didn't know about that, cool!
<rsalvaterra> 100 % useless to me, though, for the time being. :)
<\x> same
dangole has quit [Remote host closed the connection]
<Namidairo> it's absurdly cheaply priced, but it don't do 160
<rsalvaterra> Namidairo: Is that an issue? VHT160 is a big compromise. You pay the extra bandwidth with lower TX power.
<rsalvaterra> (At least on all devices I've seen out there.)
Tapper has quit [Ping timeout: 480 seconds]
srslypascal is now known as Guest9624
srslypascal has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
Guest9624 has quit [Ping timeout: 480 seconds]
<Namidairo> you have to remember that it's a 6e client
mattytap has quit [Quit: Leaving]
<Namidairo> in the above iw output it looks around the same tx power for whatever country they have it set to
rua has joined #openwrt-devel
<Namidairo> ignoring the fact that DE appears to allow the highest on a DFS range
Gaspare has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
hexagonwin has joined #openwrt-devel
rua has joined #openwrt-devel
valku has joined #openwrt-devel
hexagonwin has quit [Quit: [ Quitting client ]]
hexagonwin has joined #openwrt-devel
Gaspare has quit [Quit: Gaspare]
hexagonwin has quit []
rua has quit [Ping timeout: 480 seconds]
srslypascal has quit [Remote host closed the connection]
srslypascal has joined #openwrt-devel
minimal has joined #openwrt-devel
rua has joined #openwrt-devel
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
ecloud has quit [Ping timeout: 480 seconds]
ecloud has joined #openwrt-devel
goliath has joined #openwrt-devel
dedeckeh has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<Guest9490> mangix: ping
<Guest9490> mangix: I'm very lost with meson, it compiles fine on my local machine but using CI it recently started to act up. It wants some -fPIC stuff however that's already enabled for both Lua and APK. Any clues? https://paste.debian.net/1225274/
danitool has joined #openwrt-devel
onemarcfifty has quit [Ping timeout: 480 seconds]
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<nick[m]1234> hauke: you are also facing the binutils relink bug. Do you just go to a earlier binutils version? I guess every arch/manjaro user is currently facing the issue then compiling something to do with iproute2
Tapper has joined #openwrt-devel
srslypascal has quit [Remote host closed the connection]
srslypascal has joined #openwrt-devel
pmelange has joined #openwrt-devel
<mangix> Guest9490: the problem is liblua needs fPIC.
<mangix> As a workaround you could pass -Db_staticpie=false
<mangix> Sorry, b_staticpic
nlowe has joined #openwrt-devel
nlowe has quit []
nlowe has joined #openwrt-devel
<Guest9490> Thanks
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
nlowe has joined #openwrt-devel
Borromini has left #openwrt-devel [#openwrt-devel]
nlowe has quit []
Gaspare has joined #openwrt-devel
dedeckeh has quit [Remote host closed the connection]
<hauke> nick[m]1234: I was able to reproduce the problem on an Arch Linux test system, but I am using Debian on my main system
<Grommish> I need ncurses headers available to the build system.. I added ncurses/host and +libncurses-dev +libncurses to the Package Depends, but I still get errors pointed to missing ncurses headers (undefined reference to `setupterm'). What am I missing? I found https://forum.openwrt.org/t/solved-how-to-get-ncurses-host-and-pthread-host/17538 but the OP just changed the title to SOLVED without any info :/
<nick[m]1234> hauke: ah okay. However, I think there will be a lot of other people using arch and compiling openwrt. :S I have no idea why this only happens with iproute2. However, build process includes "-L/usr/lib". I guess it is accessing host libraries and that is why it breaks.
Gaspare has quit [Quit: Gaspare]
<nick[m]1234> I have no idea where this include happens :/
<hauke> nick[m]1234: I saw this problem when compiling binutils
<nick[m]1234> hauke: but only in combinations with iproute2
<nick[m]1234> ?
<hauke> iproute2 depends on binutils
<nick[m]1234> at least, I can compile everything if I don't include any additional package (fritzbox 7530)
<nick[m]1234> hauke: where is the binutils dependency?
<nick[m]1234> I can not find it in the makefile
<hauke> ip-full and tc-full are depending on libbpf, then bpftool-full depends on libbfd
<hauke> these are mostly build depdnecies
<nick[m]1234> is on debian the file /usr/lib/libctf.a also available?
<nick[m]1234> hauke: thanks
<slh> no, $ wget -qO- http://ftp.de.debian.org/debian/dists/unstable/main/Contents-all.gz | zgrep libctf.a #<-- not in the repos at all
<stintel> same on mips64
<nick[m]1234> hauke: yey, it compiles finally. now I have to figure out which change fixed it
<Habbie> stintel, i am
<Habbie> stintel, any idea what changed?
<nick[m]1234> hauke: fixed by https://github.com/openwrt/openwrt/pull/4905
<stintel> Habbie: boost was just bumped
<Habbie> ah, what old and new versions?
<stintel> or is this not a new issue? I just came across this faillog by accident
<Habbie> it is a recurring issue, often coinciding with boost version bumps :)
<stintel> 1.77 -> 1.78
<Habbie> thanks
<stintel> yw
<Habbie> is next week okay?
<stintel> sure
<Habbie> awesome
<stintel> next year also ;P
<Habbie> same thing!
<Habbie> i have a few days off and i suck at actually not doing work
<stintel> enjoy ;)
<Habbie> thanks :)
<stintel> sigh, opkg deps suck balls
Tapper has quit [Ping timeout: 480 seconds]
<mangix> Guest9490: looks like lua
<mangix> 's liblua.a generation does not pass fPIC
<mangix> hence the linking error
kenny has joined #openwrt-devel