<rsalvaterra>
Realtek chips are usually terrible. I have a couple of RTL8821AE cards, they've only became usable in the last 2 years, or so (I have them since 2014).
<mangix>
no upstream driver. lovely.
<stintel>
realtek is shit in general yeah, but I do like the realtek target :)
<stintel>
I should receive 2 Edge-Core ECS4100-12PH this week
<rsalvaterra>
Sure, I was only speaking of Wi-Fi chips. :)
<mangix>
small google search shows ustream patch was denied
<stintel>
I wasn't ;)
<stintel>
RTL8139 -> trash
<stintel>
R8169 -> trash
<rsalvaterra>
Can't deny facts. They don't even support byte queue limits.
<rsalvaterra>
(The GbE ones, I don't know/care about RTL8139. :P)
<rsalvaterra>
Ok, off to sleep now. Over and out. :)
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
JuniorJPDJ has joined #openwrt-devel
t4h4[m] has joined #openwrt-devel
<mangix>
rsalvaterra: you need to use ccache
hexagonwin[m] has joined #openwrt-devel
fda- has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
aleksander has quit [Ping timeout: 480 seconds]
aleksander has joined #openwrt-devel
Tapper has joined #openwrt-devel
<Tusker>
anyone here have much experience with logic analyzers ? I tried the "buck50 - bluepill logic analyzer" and it showed the timing of the different pins, but it doesn't reveal any more information than I already knew (3.3v on one pin and ~2.5v on another pin). What should I be looking for ?
victhor has quit [Ping timeout: 480 seconds]
clayface has joined #openwrt-devel
clayface_ has quit [Ping timeout: 480 seconds]
pavlix has joined #openwrt-devel
<aparcar[m]>
mangix: ping
<aparcar[m]>
mangix: is it possible mac just needs forever to build the kernel? https://paste.debian.net/
<digitalcircuit>
rsalvaterra: Brief belated note - nice to see another Quassel user! If you're ever around Libera Chat/#quassel I'll have to say hi :) (I'm just a community contributor, not project lead or anything)
<Tusker>
i have some memory of having a tool that can detect uart ports, like jtagulator but running on the bluepill... does such a thing exist, or am I dreaming ?
<Tusker>
uartfuzz isn't what I remember, but it might be good enough...
agb[m] has joined #openwrt-devel
gnustomp[m] has joined #openwrt-devel
<mangix>
aparcar[m]: that looks like the build didn't even start
<aparcar[m]>
mangix: what do?
<mangix>
no idea
musashino has joined #openwrt-devel
<aparcar[m]>
☹️
Rikimp has joined #openwrt-devel
Rikimp has quit [Quit: Page closed]
olmari has joined #openwrt-devel
rmilecki has joined #openwrt-devel
valku has quit [Quit: valku]
goliath has joined #openwrt-devel
guidosarducci has quit [Remote host closed the connection]
guidosarducci has joined #openwrt-devel
slh has quit [Ping timeout: 480 seconds]
slh64 has quit [Read error: No route to host]
nitroshift has joined #openwrt-devel
Dgrey has quit [Remote host closed the connection]
mangix has quit [Remote host closed the connection]
mangix has joined #openwrt-devel
<Tusker>
Extreme Networks are finally committed to sending me a GPL Source Code bundle for the AP3935i :) Still fighting with WatchGuard, their support is saying that their legal team is still looking at it. I now have a case with the WatchGuard legal team open too... and have one of the legal team member "email read notification"
musashino has quit [Ping timeout: 480 seconds]
<aleksander>
jow, it really was just adding a script in /etc/hotplug.d/wwan :)
slh has joined #openwrt-devel
slh64 has joined #openwrt-devel
musashino has joined #openwrt-devel
pmelange has joined #openwrt-devel
danitool has joined #openwrt-devel
musashino has quit [Quit: Ping timeout (120 seconds)]
swiftgeek has quit [Ping timeout: 480 seconds]
mangix_ has joined #openwrt-devel
mangix has quit [Read error: Connection reset by peer]
guidosarducci has quit [Remote host closed the connection]
guidosarducci has joined #openwrt-devel
pmelange has joined #openwrt-devel
victhor has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
rsalvaterra has quit [Quit: rsalvaterra]
rsalvaterra has joined #openwrt-devel
rsalvaterra has quit []
rsalvaterra has joined #openwrt-devel
<stintel>
and the macbook lost its network again. what a piece of crap os
<rsalvaterra>
stintel: I have a macOS machine that from time to time, out of the blue, slows down to a crawl. I can barely move the mouse pointer. I have absolutely no idea what it's doing.
<rsalvaterra>
I honestly consider macOS to be much worse than Windows.
gladiac is now known as Guest1889
gladiac has joined #openwrt-devel
Guest1889 has quit [Ping timeout: 480 seconds]
<stintel>
rsalvaterra: 100%
Tusker has quit [Quit: Time wasted on IRC: 14 hours 37 minutes 21 seconds]
* ldir
loves his macos devices but realises everyone is entitled to an opinion
minimal has joined #openwrt-devel
* ldir
says 'systemd' to watch the linux group split into factions
nitroshift has quit [Quit: Gone that way --->]
jbowen has joined #openwrt-devel
<Slimey>
lol name a useless device that runs off of Poe
<karlp>
gives you controland power of lighting over one system, instead of running in dali and controllers and junk
<Slimey>
you know those usb dongles that are basicly led lights, promo items you get at shows and what now thats what im thinking of but poe powered :P
<PaulFertser>
I'd guess it's somewhat suboptimal, having relatively low voltage (so requiring thick cables) and humble per-port budget.
<PaulFertser>
karlp: I challenge you to provide an example of something more useless than that.
<PaulFertser>
A PoE ceiling fan?
aleksander has quit [Quit: Leaving]
<karlp>
paul, well, when a bunch of led lighting is just doing cheap and nasty ac dc supplies in each one, is it _really_ that bad? you don't need "thick cables" you're staying in the same power budget of PoE.... ?
<PaulFertser>
UTP is much thicker and costs more than you'd need for 230 AC line for the same power.
<karlp>
yes, but you get the control as well....
dwmw2_gone has quit [Remote host closed the connection]
pmelange has joined #openwrt-devel
dwmw2_gone has joined #openwrt-devel
<PaulFertser>
karlp: I do not argue the idea that PoE lighting might make sense in certain specific conditions. However I can't find anything that's more obviously useless.
<karlp>
I dunno, poe lighting sounds great
<karlp>
alternative is trusting them to get wireless mesh shit to work properly and provision them nicely?
<karlp>
and besides, how else could FB work in the dark to try and turn it back on again, if they were using plain old lights, they might stay running? :)
<nick[m]1234>
@dangowrt: if I try to setup ibss or 802.11s mesh on mt7622 it always says Tue Oct 5 14:10:27 2021 kern.warn kernel: [ 909.747490] netlink: 'iw': attribute type 302 has an invalid length.
swiftgeek has quit [Ping timeout: 480 seconds]
<minimal>
PaulFertser: more useless? PoE-powered door card readers where the PoE switch is inside the room the locks are protecting (assuming they are "fail locked")? lol
<minimal>
not that far off Facebook's problem yesterday - the apparently had to use an angle grinder to cut their way into their server security cage as the card readers weren't working for some strange reason lol
<PaulFertser>
minimal: :)
<rmilecki>
ynezz: i'm pushing your firmware-utils commit :)
dwmw2_gone has quit [Quit: Coyote finally caught me]
dwmw2_gone has joined #openwrt-devel
swalker has quit [Remote host closed the connection]
swalker has joined #openwrt-devel
Dgrey has joined #openwrt-devel
clayface has joined #openwrt-devel
clayface_ has quit [Ping timeout: 480 seconds]
mangix_ has quit []
mangix has joined #openwrt-devel
decke has quit [Quit: Leaving.]
<mangix>
ldir: nothing wrong with factions. As far as macOS is concerned, keypress speed is horrible. Deleting stuff is painful.
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<hauke>
mangix: are you sure it is nto related to your VM setup?
<hauke>
*not
<mangix>
I don't think so. It was just as slow with a bare metal install
<rsalvaterra>
stintel: Fetching the popcorn, heh? :P
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
<stintel>
nah
<stintel>
still at "work"
<stintel>
ordered pizza, I'll probably be going for another while
<rsalvaterra>
Heh… today is holiday here. :)
<stintel>
enjoy!
<rsalvaterra>
stintel: Partying hard with the git bisect. :P
<rsalvaterra>
I don't know if there are any users running OpenWrt on NVIDIA ION systems, but I'm sure they would like their AHCI controllers working when Linux is bumped to 5.15…
<stintel>
:P
rsalvaterra has quit [Quit: rsalvaterra]
rsalvaterra has joined #openwrt-devel
cp- has joined #openwrt-devel
<rsalvaterra>
About 3 steps to go and I'm deep into PCI commit territory. Hmm…
<stintel>
I've had a similar issue where iommu broke a SATA controller
<stintel>
but that was on an i7 3930k :P
<stintel>
not as painful as the ion
<stintel>
I sold my ion a loooooong time ago
<stintel>
these first gens of atom are trash
<rsalvaterra>
I have mine since around 2009… :)
<rsalvaterra>
The first gen atoms are immune to Spectre/Meltdown. ;)
<stintel>
pfft :)
<stintel>
yet without the mitigations they're still slower than anything from the same era :P
* stintel
must drink more water
<rsalvaterra>
Also more power-efficient. :)
<rsalvaterra>
If I could, I'd DCC you a bottle. And get a couple of Nobel prizes in physics, in the process. :P
<ldir>
I'd write you a letter but I couldn't spell prrrpprprpppprp and that's about all I got saaayy
<stintel>
does anyone know if I can configure an l2tp connection to reconnect after being disconnected ?
<rsalvaterra>
stintel: Never touched L2TP, I'm afraid… :/
<stintel>
it works fine but after some time it disconnects, I'd just like it to reconnect instead of leaving the connection down
<stintel>
oh well, I guess I can try to remember if "AP x does not show up in controller, check L2TP first"
<stintel>
unfortunately it only happens maybe once every 2 weeks or so and I keep forgetting that :P
<stintel>
pppoe seems to do that already
<stintel>
maxfail 0
<stintel>
ok so combination of maxfail 0 and persist
<aparcar[m]>
neoraider: ping
<stintel>
is there a way to tell netifd to reload protocol handler scripts without /etc/init.d/network restart ?
<rsalvaterra>
stintel: I was going to say I did it with PPtP, but it's irrelevant. :P
<stintel>
it takes the new script actions instantly, when doing ifup
<stintel>
but new protocol options are ignored until netifd reloads the protocol
<stintel>
I just stopped keepalived on the backup router to prevent havoc
<stintel>
now ... to debug the next problem
<stintel>
keepalived segfault during stop
<philipp64|work>
stintel: I’m trying to use an external USB drive for /var on my Archer A7 v5 but it hangs during boot for some reason…
<stintel>
no experience with that
<stintel>
hmmm, now how do I stop this keepalived thing when it's running via gdbserver :/
<philipp64|work>
oh… why did I think you did the persistent var changes…
<stintel>
philipp64|work: to have /var not on tmpfs
<philipp64|work>
Okay, so how do you manage it?
<philipp64|work>
or do you just have enough flash that you can have a large overlay FS?
<stintel>
yes
<stintel>
actually not even overlay, just ext4 image, remove the /var symlink so /var is just part of the root ext4 fs
<stintel>
I should probably do some follow up commits that move some stuff that expects /var to be a symlink to /tmp
<stintel>
like procd creates a bunch of /var/run/foo.lock
<philipp64|work>
Ah. I’m trying to mount the USB ext2 FS into /var but that’s not working and I can’t tell why because there’s no console on my Archer…
<stintel>
philipp64|work: try it in a VM first?
<philipp64|work>
Do I need an Arm Mac Book to do that?
<stintel>
nope
<philipp64|work>
all of my KVM servers are x86.
<stintel>
I doubt it's arch specific
<philipp64|work>
Hmm.
<stintel>
probably some critical service tries to write something in /var/foo and doesn't exist
<stintel>
make a lock dir on that USB drive you are mounting so you will have /var/lock root:root 755\
<stintel>
and also lib etc log
<philipp64|work>
Well, with your change /var is an empty directory in the ROM image (squashfs in my case) that would get mounted over when /etc/config/fstab gets processed…
<stintel>
ah and these dirs are mkdir'd in /etc/init.d/boot
<philipp64|work>
So I’m thinking that (a) the file would get created too early, live on the /overlay and then get hidden when /var gets mounted over, or (b) /var gets mounted relatively early on (like S11?) but then the subdirs don’t get created soon enough…
<philipp64|work>
* lives
<aparcar[m]>
mangix: could you make a meson backport PR and we see what people think?
<stintel>
but then the files that have already been created would be gone too
<stintel>
I'm going to conclude that you can't mount /var on a separate filesystem like that
<stintel>
this is probably why we have extroot ?
<philipp64|work>
Why are those /tmp and not /run with run being a symlink (for now) to /tmp?
rmilecki has quit [Ping timeout: 480 seconds]
<stintel>
-EPARSE
<philipp64|work>
it would be nice if /etc/config/fstab supported “bind” mounts…
<stintel>
it would be nice if /etc/config/fstab supported cifs mounts ;)
<philipp64|work>
81-83…. should be mkdir(“/run/lock”); and mkdir(“/run/state”); etc.
<stintel>
I'm actually considering adding /var/lock -> /tmp/lock and /var/state -> /tmp/state yes, but still won't solve your problem?
<stintel>
only if you make those yourself, in advance, on the USB drive
<mangix>
aparcar[m]: why?
<aparcar[m]>
backport apk long term maybe.
<mangix>
I don't recall SDK changes being backported.
<mangix>
why would you backport apk longterm?
<aparcar[m]>
21.02 lives easily for another 18 month. Would be nice to have a base for APK
<mangix>
I'm not the person to ask. that being said, major SDK changes like this I have not seen being some before.
<mangix>
/s/some/done
swalker has quit [Remote host closed the connection]
goliath has quit [Quit: SIGSEGV]
<aparcar[m]>
nevermind I'l lmaintain my own patches for now
<philipp64|work>
stintel: Yeah, we should probably have a short script that initializes a /var FS properly and then pivots it for testing before a reboot…
<stintel>
XDP just allows to run eBPF in hardware (NICs) that supports it afaik?
<stintel>
so XDP on its own is useless, we'd need eBPF support first
<rsalvaterra>
Not sure. From what I read/understood, there's nothing stopping XDP from running in software.
<stintel>
> The idea behind XDP is to add an early hook in the RX path of the kernel, and let a user supplied eBPF program decide the fate of the packet.
<stintel>
so yeah, how I understand it, we need eBPF support first
<stintel>
I doubt we'll have anything in the next release
<rsalvaterra>
That's not an issue, we're in it for the long run. :)
<stintel>
the issue is going to be manpower again :P
<stintel>
if only we could make a couple of clones of Felix, Jo, John, etc :)