MaxSoniX has quit [Quit: Konversation terminated!]
torv has quit [Remote host closed the connection]
torv has joined #openwrt-devel
<karlp>
when / where is wireless disabled in 2203? I've had a uci-defaults/99_initial_install script that enables wireless and sets the AP name and things for a few years now, and I can see it running, and other changes it makes are all done, but I still /etc/config/wireless is still default and disabled?
<stintel>
I believe that happens in "wifi config", maybe that helps you dig further
<stintel>
e.g. rm /etc/config/wireless; wifi config # results in wifi disabled, and this happens on first boot as there is no /etc/config/wireless then
<stintel>
or at least that's my understanding - if I'm wrong or not helpful, just ignore me :P
schwicht has quit [Ping timeout: 480 seconds]
schwicht has joined #openwrt-devel
<karlp>
looks like it might be /rom/etc/hotplug.d/ieee80211/10-wifi-detect then...
<karlp>
which yes, is going to very likely be after the uci_defaults have finished running.
<karlp>
that's apparently been there since 2016 though, so maybe just timing changed?
rua has quit [Remote host closed the connection]
rua has joined #openwrt-devel
schwicht has quit [Ping timeout: 480 seconds]
__lore__ has joined #openwrt-devel
_lore_ has quit [Ping timeout: 480 seconds]
barhom has joined #openwrt-devel
bluse-blue[m] has joined #openwrt-devel
ctdvqgg445[m] has joined #openwrt-devel
decke[m] has joined #openwrt-devel
dfceaef[m] has joined #openwrt-devel
domon has joined #openwrt-devel
evils[m]1 has joined #openwrt-devel
fieryeagle954[m] has joined #openwrt-devel
fpsusername[m] has joined #openwrt-devel
gnustomp[m] has joined #openwrt-devel
hexagonwin[m] has joined #openwrt-devel
John[m]123456 has joined #openwrt-devel
Jonny[m] has joined #openwrt-devel
JuniorJPDJ has joined #openwrt-devel
Kiste has joined #openwrt-devel
Q__ has joined #openwrt-devel
lipnitsk has joined #openwrt-devel
MatMaul[m] has joined #openwrt-devel
Movedtomkg20001mkg20001io[m] has joined #openwrt-devel
<oliv3r[m]>
I'm hacking on the realtek stuff, and (at least for now) we see some duplication in headers under drivers/net/{ethernet,phy,sda}/*.h If defines are needed multiple locations, how is this usually done currently?
* f00b4r0
learns the hard way that sysupgrade -b doesn't work after mount_root
<yolo>
browsers are going to be default to DoH(dns over https), maybe dnsmasq will only act as a dhcp server one day.
<stintel>
yeah and goodbye ad blocking, dns-based dscp marking, etc
<jow>
stintel: it just means that *you* can't do it anymore :) The big corps controlling the Do-whatever infra can do
<stintel>
yep
<stintel>
advertisements become non-optional ;)
<stintel>
time to build that cabin in the woods
<f00b4r0>
:)
<yolo>
so /etc/hosts and dnsmasq soon to be meaningless on openwrt for adblocks, but, they can still provide an IP list to block, it's just IP is shared by multiple hosts so lots of overblocking expected
<jow>
how convenient for the big ones that everying is concentrated into a cloud hosting duopoly
<jow>
makes ip based blocking futile as well
* yolo
feels hopeless
<jow>
why? it's all for the privacy! it's for your own good!
guidosarducci has quit [Remote host closed the connection]
<yolo>
you mean my privacy is now concentrated by one or two big guns. with AI it will know me better than myself
<jow>
yes, exactly
<yolo>
next is vpn built into browsers, but very few players offer 'free' vpn servers, same thing
<jow>
commercial vpn is essentially paying someone for collecting all your data
<\x>
maybe you guys will be spammed a lot with adverts of like APs and uart usb converters
<jow>
in addition to also paying your ISP
guidosarducci has joined #openwrt-devel
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
<f00b4r0>
hmm I must be specially tired. I'm trying to reproduce a working network config on a different but very similar device (hap-acl to hap-ac2) and I'm thoroughly failing
<f00b4r0>
oh crap
guidosarducci has quit []
guidosarducci has joined #openwrt-devel
<f00b4r0>
robimarko: on hap-ac2 the wan port is on the switch, how is one supposed to use this thing with vlans? :) (21.02, pre-DSA)
schwicht has joined #openwrt-devel
<f00b4r0>
i don't even understand how the default configuration with eth0 + eth1 works
schwicht_ has quit [Read error: Connection reset by peer]
Tapper has quit [Read error: Connection reset by peer]
Tapper has joined #openwrt-devel
<robimarko>
f00b4r0: I did not really use VLAN-s pre DSA
<robimarko>
But AFAIK, you can change the ethernet defconfig to define a switch there and then its like on any other swconfig device
<f00b4r0>
robimarko: I see. I think there's simply no way to make this work. As soon as you setup vlans, the wan port no longer works
<f00b4r0>
no, because eth1 is apparently hardwired to the wan phy
<robimarko>
Not really
<robimarko>
All of the ports are on the switch
<f00b4r0>
I'm going to give a try to 22.03. I fear I'm going to regret that but heh :)
<robimarko>
22.03 will be the same
<f00b4r0>
robimarko: sure, but there's eth0 and eth1
<robimarko>
Well, thats just a EDMA driver fiction
<robimarko>
There is just one ethernet controller
<f00b4r0>
with eth1 being the "wan" device, and eth0 being the switch cpu port
<f00b4r0>
so how does this work
<f00b4r0>
oh, single controller
<robimarko>
Yeah
Gaspare has quit [Quit: Gaspare]
<f00b4r0>
well it's completely absurd to present two interface then, is it not?
<robimarko>
Its basically just so that "LAN" and "WAN" traffic dont mix
<robimarko>
As there was no proper switch driver
<f00b4r0>
because eth1 cannot do anything as soon as the switch is configured. And I'm going to bet you have leaks between lan and wan in the default setup
<robimarko>
So QCA did this crap with always tagging on both
<f00b4r0>
oh
<f00b4r0>
dear
<robimarko>
You shouldnt have any leaks as they use different VLAN-s
<f00b4r0>
well yeah but then you can't use vlans of your own :P
<robimarko>
Not properly as you will be double tagging
<robimarko>
Hence why DSA was a rather drastic improvement
<f00b4r0>
dsa should be better I suppose, at least the interfaces will "make sense"
<f00b4r0>
yeah
<f00b4r0>
i really didn't want to deal with nftables, but I guess so be it ;P
<robimarko>
With DSA even VLAN filtering works
<f00b4r0>
i'm kinda pressed to upgrade because since I flashed 21.02 on my hapacl, the wireguard cpu load has gone insane
<f00b4r0>
i couldn't find out what caused it, but it increased by x17
<robimarko>
That is rather weird
<f00b4r0>
hmm. the 21.02 luci happily accepted and flashed the 22.03 image with preserving settings. I suspect I'm going to regret that ;P
<f00b4r0>
robimarko: yeah it's crazy. I don't know what causes it
<f00b4r0>
oh
<f00b4r0>
22.03 doesn't have the DSA driver
<robimarko>
Yeah
<f00b4r0>
well fuck this then
<robimarko>
It only happened like a month ago
<robimarko>
It does work fine on 22.03 if you want to backport though
<f00b4r0>
but it needs 5.15 I guess?
<robimarko>
No, it works on 5.10 as well
<f00b4r0>
ah
<robimarko>
It was initally done on 5.10 and then just added compability for 5.15 as well
<f00b4r0>
i see. But even if I backport, such a patch will be (rightfully) NAKd: it would break service release upgrades
<robimarko>
Yeah
<f00b4r0>
well I guess I'm going to put that device back in the drawer and find something else
<robimarko>
That is not something for upstream backporting
<robimarko>
Why not use snapshots?
<f00b4r0>
still have that Mochabin to play with
<f00b4r0>
robimarko: because I don't run the routers I rely on for work on snapshots :)
<f00b4r0>
it's a policy.
<f00b4r0>
since I work from home.
<robimarko>
Ok, I wont argue that logic
<stintel>
have 2 in a HA setup ;)
* stintel
hides
<f00b4r0>
;)
<f00b4r0>
you're braver than I
<f00b4r0>
robimarko: we should probably document that VLAN setups on hap-ac2 are totally hosed though
<robimarko>
f00b4r0: Its a know issue
<f00b4r0>
ha
<f00b4r0>
I didn't RTFM then :)
<robimarko>
You can check the forums, there is dozen or so threads for VLAN-s on IPQ40xx
<f00b4r0>
i can see why, now
<robimarko>
Its been like that since the target was added
* stintel
gets the feeling he didn't miss out on ipq4 :P
<robimarko>
stintel: not like there was a lot of competition for ipq40xx
<yolo>
openwrt can still block DoH if adblock is a must: detect the traffic, if there are lots of 443/80 traffic without 53 traffic, block everything. meanwhile openwrt can provide https-dns-proxy if you _really_ need that, now you have both adblock and DoH
<stintel>
you could indeed do IP whitelisting based on the DNS requests you receive locally
<stintel>
good idea :)
<stintel>
afk, weekend, cocktail o clock
<robimarko>
Arent you a day early for weekend?
<f00b4r0>
hf!
<yolo>
IP whitelisting will overblock, one IP could be used by multiple sites esp via CDN etc
<yolo>
probably do not need detect traffic at all, just block anything that is not in dnsmasq cache, that way dnsmasq must be used for dns.
<yolo>
then add https-dns-proxy on the way out for "privacy", goal is to adblock+"privacy"
guidosarducci has quit [Remote host closed the connection]
guidosarducci has joined #openwrt-devel
<mrnuke>
svanheule: I was editing /etc/config/system to enable the LEDs with LED controller driver (rtl838x). I couldn't find a way to set 'rtl_hw_trigger'
MaxSoniX has joined #openwrt-devel
Gaspare has joined #openwrt-devel
Borromini has joined #openwrt-devel
Borromini has quit [Quit: Lost terminal]
srslypascal has quit [Remote host closed the connection]
srslypascal has joined #openwrt-devel
aiyion_ has quit [Remote host closed the connection]
<aiyion>
The first bad commit changes the patchversion from 5.4 to 5.10;
srslypascal is now known as Guest4235
srslypascal has joined #openwrt-devel
<aiyion>
I fail to see a relevant difference in the patch 100-sunxi-h3-add-support-for-nanopi-r1.patch
Guest4235 has quit [Ping timeout: 480 seconds]
Slimey has joined #openwrt-devel
srslypascal is now known as Guest4237
srslypascal has joined #openwrt-devel
Guest4237 has quit [Ping timeout: 480 seconds]
Tapper1 has joined #openwrt-devel
Tapper has quit [Read error: Connection reset by peer]
<Tapper1>
Hi I keep having reboots with master and my r7800
<Tapper1>
SNAPSHOT r21100
Borromini has quit [Quit: Lost terminal]
csrf has joined #openwrt-devel
<mrnuke>
svanheule: Need to update /etc/init.d/leds was my understanding
srslypascal has quit [Remote host closed the connection]
<f00b4r0>
zorun: so I just experienced the "ULA problem" on macos
<f00b4r0>
I don't even have a working ipv6 upstream.
<slh>
Tapper1: did locking the cpu frequency to max make any difference?
<Tapper1>
slh I did not make the change.
<Tapper1>
I never had a reboot befor the update to the ipq8x patches
<Tapper1>
ipq86x
srslypascal has joined #openwrt-devel
<slh>
Tapper1: please try it, the issue seems related to frequency transition, especially involving the lower frequencies (384 MHz) - might not fix it completely (might not even fix it at all, but it's still interesting to put to the test), but it should help a little. Ansuel is working hard on finding the real issue
<Tapper1>
can you tell me how to do it agin?
<slh>
I feared you'd ask that ;)
<Tapper1>
lol sorry
<slh>
(I'm currently physically apart from my ipq806x devices...)
<Tapper1>
slh OK mate I will flash latest master anyway and see how we go.
<slh>
I wouldn't expect a change there, but no harm trying
<Tapper1>
I do have a older build that does not reboot.
<Tapper1>
That I mite flash back to
<Tapper1>
Going to go to latest master first.
Tapper has joined #openwrt-devel
Tapper1 has quit [Read error: Connection reset by peer]
<slh>
Tapper1: cat /sys/devices/system/cpu/cpufreq/policy0/cpuinfo_max_freq should tell you the exact max value, in the past I had this in my /etc/rc.local (replace 600000 with the value found as max), http://paste.debian.net/1258503/ only lines 4 to 7
<slh>
you don't even need to change rc.local, just execute those 4 lines after booting (they're forgotten after the next reboot, unless you add them to rc.local)