rmilecki has quit [Quit: Konversation terminated!]
rmilecki has joined #openwrt-devel
rmilecki has quit []
slh64 has joined #openwrt-devel
xes has quit [Ping timeout: 480 seconds]
rmilecki has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
<Mangix>
soxrok2212: nyet
<rmilecki>
nbd: with brcmfmac there are no beacons at all after my test
<rmilecki>
scanning doesn't provide OpenWrtB SSID anymore
<rmilecki>
(untill i call "wifi")
gladiac has joined #openwrt-devel
rz has joined #openwrt-devel
goliath has joined #openwrt-devel
<rmilecki>
nbd: i'm not sure if you can find anything interesting there, but still, got the log with debugging for you: https://pastebin.com/raw/UxhmHgia
xavifr has joined #openwrt-devel
<nbd>
rmilecki: okay, maybe i need to find a way to change the order in which interfaces are reconfigured
<rmilecki>
nbd: i'm afraid it may be some brcmfmac trickery
<rmilecki>
it seems reconfig went as expected
<rmilecki>
interface is setup
<rmilecki>
iw dev X info looks fine
<rmilecki>
maybe brcmfmac doesn't send something to the firmware
xes has joined #openwrt-devel
tom- has quit [Quit: WeeChat 4.0.0]
<nbd>
i guess i will have to change the code to force a restart if the first interface is affected
<nbd>
it seems that various drivers (especially brcmfmac) are rather picky about up/down order when it comes to the first AP interface
<nbd>
because they all rely on hostapd not supporting anything dynamic there by default
Misanthropos has quit [Read error: Connection reset by peer]
philipp64 is now known as Guest6925
philipp64 has joined #openwrt-devel
Guest6925 has quit [Ping timeout: 480 seconds]
drikus_ has quit [Ping timeout: 480 seconds]
Misanthropos has joined #openwrt-devel
rz has quit [Remote host closed the connection]
danitool has joined #openwrt-devel
robimarko has joined #openwrt-devel
Tapper has joined #openwrt-devel
dangole has joined #openwrt-devel
<nbd>
rmilecki: pushed a fix
<rmilecki>
nbd: for mt7603 or brcmfmac?
<nbd>
should fix both
<nbd>
though with mt7603 some corner cases remain
<nbd>
but i found the bug
<rmilecki>
nice
<nbd>
the bug on mt7603 is that if a vif gets disabled, beacons get disabled
<nbd>
very simple fix
<rmilecki>
nbd: fix confirmed for brcmfmac
bluew has quit [Ping timeout: 480 seconds]
<nbd>
can you please try adding 3 ap interfaces and disabling the middle one?
<nbd>
that should trigger the reload path and it should only cause the last one to be recreated
Ansuel has quit [Quit: ZNC 1.9.x-git-253-750eac08 - https://znc.in]
Ansuel has joined #openwrt-devel
Ansuel has quit []
Ansuel has joined #openwrt-devel
Ansuel has quit [Quit: ZNC 1.9.x-git-253-750eac08 - https://znc.in]
Ansuel has joined #openwrt-devel
<Ansuel>
hi from oracle arm istance.... they finally didn't reject my card -.-
<Ansuel>
was impressed by how ""easy"" was to setup compared to aws stuff... the only blocker was discovering that they don't love ufw and just ask you to open stuff with iptables
<hauke>
Ansuel: opening a port took me ~2 hours in the oracle clould ;-)
<shibboleth>
isn't ufw a wrapper on top of iptables/netfilter?
<colo>
hauke: lol o_o
<colo>
shibboleth: yes
<Ansuel>
well it's not immediate to understand that ufw is disabled and discouraged
<colo>
iptables and nowadays nft, I guess
<shibboleth>
true
<hauke>
they have a por filter in addition outside of the VM
<Ansuel>
and you need to manually edit v4.rules for iptables
<hauke>
*port
<Ansuel>
oh yes also that
<colo>
Ansuel: they ship/set up Oracle Unbreakable stuff aka RHEL, I guess?
<colo>
then I would assume they have firewalld?
<hauke>
I used ubuntu
<colo>
ah OK
<Ansuel>
colo you can chose multiple thing i chose ubuntu for the lulz
<colo>
I am actually surprised by that degree of freedom :D
gladiac has quit [Quit: k thx bye]
gladiac has joined #openwrt-devel
<Ansuel>
colo you pay taht freedom with having to contact support for some special feature
<colo>
they will probably monetize it all some day, by suitably updating the Oracle JRE/JDK license, and shoehorning that into each of the "free" instances on Oracle Cloud (:
gladiac has quit []
<Ansuel>
like rDNS and to support is not available with the always free tier and you need to switch to pay as you go (aka a pact with the devil as soon as you use anything that is not free)
<hauke>
I think they are about 10 times more expensive than hetzner if you have to pay
<hauke>
for the same CPU speed, traffic probably even more
<Ansuel>
yep
<Ansuel>
but mhe i'm happy... for znc is just ok
<Ansuel>
was having too much problem with my local istance at my parents house as with summer the blackouts are daily -.-
<Ansuel>
and stupid ups just turn off and doesn't swiotch on again if the battery go dead -.-
<robimarko>
I am interested as well, that sounds weird
<ynezz>
stintel: Re: #13156, are you as well using filogic AP/device?
<ynezz>
stintel: or is that some other chipset/driver then mt76?
<ynezz>
stintel: do you experience same SAE issues on iOS devices (client reports wrong key until you restart the VAP)
<ynezz>
hauke: ok, I'll do my best to test/commit it soon
<stintel>
ynezz: ramips/mt7621 EAP615-Wall
minimal has joined #openwrt-devel
<ynezz>
stintel: ok, so according to ToH its MT7905DAN/MT7975DN so mt76, thanks
danitool has quit [Ping timeout: 480 seconds]
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<stintel>
ynezz: I don't have iOS devices. I do have an iPad but I rarely use it so didn't notice issues
<stintel>
I could do some more debugging probably on 1/08
valku has joined #openwrt-devel
schwicht has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
xavifr has quit [Ping timeout: 480 seconds]
<Ansuel>
haule south Italy grid doesn't keep up with the spike AC load
<Ansuel>
In Naples cable are catching fire
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<hauke>
Ansuel: oh, I didn't know that the grid is so bad in south Italy
<robimarko>
Grid in Dalmatia is bad, but its somehow keeping up with AC-s
<robimarko>
Didnt know situation was that bad
<hauke>
I only visited Rome and there I didn't notice any problems with the grid
<Ansuel>
It's not that it's bad, it wasn't bad but we weren't used to use AC and everyone is switching to it... No idea how bad the situation is but for Naples is just bad maintenance for my parents house it might be something minor
<johnf>
is there a way to check on NOR flash health? eMMC has nice tools though it's also far more likely to have problems
<johnf>
that is, for types of flash far more likely to have problems
<hauke>
Ansuel: In Germany ACs are still uncommon in private houeses and also in many offices
<Ansuel>
johnf you should check bad blocks but it's not trivial as they can be handled by hw
<robimarko>
I am working on adding FW loading support to the Aquantia PHY driver to make deving easier and not have to depend on U-boot
<Ansuel>
hauke same here but the state gave some bonus so installing AC was free (free for real) but the electricity company didn't cared to update the grid and the situation is getting worse every year...
<Ansuel>
robimarko what API you would use?
<robimarko>
Ansuel: Just the regular kernel firmware APi
<robimarko>
I am just wondering if its ok to have it as downstream patch
<robimarko>
As it would get shot upstream since the FW is not public
<Ansuel>
Are you sure it won't be accepted?
<robimarko>
Yeah, cause they shot FW loading for Marvell10G PHY driver for that exact reason
<Ansuel>
Also the user would be Qualcomm ax?
<Ansuel>
Or also other?
<robimarko>
Anything that uses AQR PHY-s would be able to use it
<Ansuel>
If there isn't a chance for upstream I would provide as file
<Ansuel>
Or even better a kmod
<hauke>
I added FW loading for PHYs in lantiq GSWIP and it was no problem
<johnf>
Ansuel: using the badblocks program to test? this additional writing would eventually ensure I had bad blocks :)
<johnf>
or did you mean something else?
<robimarko>
Ansuel: Its not possible to detach it from the PHY driver nor it makes sense
<hauke>
I think you should try to get it upstream
<Ansuel>
Well at least free review for better code
<robimarko>
hauke: I can try, thats not the issue
<hauke>
maybe reference some obsurce page which provides a FW
<hauke>
I thiunk it is also ok for OpenWrt to carry a patch for this
<Ansuel>
robimarko idea was to make it external so we don't have to put stuff in the target directory
<hauke>
Can you replace the PHY FW on all the chips?
<hauke>
normally they even restrict some features using the FW
<hauke>
so OpenWrt could circumvent this
<robimarko>
Ansuel: But how, as its modifying the PHY driver itself, I plan to just include it as a generic patch
schwicht has joined #openwrt-devel
<robimarko>
hauke: Yeah, even if AQR is loading the FW from NOR if you want it can be updated
<robimarko>
Currently we depend on having U-boot load the FW before booting or you have to use the patched userspace tool which is not shipped in OpenWrt
<robimarko>
Ansuel: I wanted to use NVMEM to be able to load the FW from NOR for example so the exact version that vendor uses can be loaded
<robimarko>
Cause, for example Qnap 301W has different FW for both AQR-s as their LED config is reversed
<robimarko>
But currently we use U-boot to load just one FW to both of them before booting
<Ansuel>
That is doable
<Ansuel>
Actually the fact that the Fw might be present in nand
<Ansuel>
Is a chance for upstream
<Ansuel>
You can justify the patch that way
<robimarko>
Thats a great idea
<robimarko>
Cause QCA made a separate partition by default to ship the PHY FW in there
<Ansuel>
Just add support for nvmem and you get sysfs load support
<Ansuel>
Indirectly
<robimarko>
Most of the vendors just left it empty
<Ansuel>
Yes but if at least one does than we have an user eheheheh
<robimarko>
QCA reference designs use it
<robimarko>
That should be enough of a use case
<Ansuel>
Yep no excuse to reject it
<Ansuel>
At worst they will ask to support only mtd
<Ansuel>
And we will use a patch to support sysfs fallback downstream
<robimarko>
Regular fw_request will be a fallback for deving mostly
<robimarko>
As the FW is not freely redistributed anyway
<robimarko>
So we dont ship it
<robimarko>
I have the FW loading working, ported it from U-boot, now its just a matter of cleaning it up
<hauke>
nice
soxrok2212 has quit [Read error: Connection reset by peer]
<schmars[m]>
Tianling Shen: i took my nanopi r5s with me on holiday, with a large ssd and a complete downloads.openwrt.org mirror - works pretty nicely 🥰 one issue i'm seeing is the following: i power it from my thinkpad laptop on battery, and if any ethernet port is plugged in while booting, the network wont activate. Unfortunately i dont have any tools with me for debugging. It works if i plug in the cable after the wan leds has blinked very
<schmars[m]>
briefly, right after the failsafe led blinking during early boot. Maybe this info helps, i can debug more in a few days when i'm back home
<hauke>
does QCA store the AQR FW on the same flash as the Linux system or do they attach a SPI NOR to the PHY?
<Ansuel>
Feel free to send the patch maybe I can catch some nitpick so we speedup the review but it's net-next so should be fast
<robimarko>
hauke: They have an ETHPHYFW partition on NOR or NAND
<Ansuel>
hauke just additional partition in the boot nand
<robimarko>
Cause 99% of the vendors dont want to spend money on attaching a NOR to AQR PHY-s
<robimarko>
I have just one board that has NOR attached to AQR so it loads the FW itself
<robimarko>
Ansuel: Sure, I need to get the FW name per model done and I can send it to you
<hauke>
ok
<Ansuel>
Also remember DT support for Fw name should be better than hardcoding
<Mangix>
Ansuel: that grid discussion was confusing when talking about AC.
<Ansuel>
?
<Mangix>
Made it sound like Italy switched to using Alternating Current
<Mangix>
Anyway, RIP 64MB on mt7621. No idea if you will be missed.
<Ansuel>
Oh lol actually we should all switch to DC :)
<grid>
grid discussion is definitely confusion :o
<Mangix>
Conversion to DC is too expensive. And dangerous.
<Ansuel>
grid we are not talking about you
<grid>
heh ;)
<Ansuel>
For long distance it was discovered that very high DC is better than AC
<Ansuel>
Also solar system are DC already
<Ansuel>
So there is also that problem is independent house where there is a lot of loss in converting DC from battery and solar to AC and then to DC again
<Ansuel>
It's an interesting situation if you consider this scenario
<Mangix>
I asked ChatGPT. It has no such information.
<Ansuel>
Search for HVDC
<Mangix>
Ansuel: sure. DC only plugs have to be developed
<Mangix>
For solar it makes sense.
<Ansuel>
IMHO the best thing would be use AC for stove fridge oven and AC and DC for everthing else
<Ansuel>
I bet device would be much more efficient and simple with just transforming the voltage to the needed value and have directly DC supplied
<Ansuel>
Using a different plug is a way to skip mistake with using AC on DC devices
<Ansuel>
Also many device have already that (device with external power supply)
shibboleth has quit [Quit: shibboleth]
<Mangix>
Why AC for stove fridge oven?
<Ansuel>
Big W they are dangerous to use in DC
<Ansuel>
Also the components internally would use AC directly
<Ansuel>
Fridge compressor use AC, stove same, oven? Same. AC? They are fridge
<Mangix>
Hmm OK
<Ansuel>
America have a similar thing
<Ansuel>
They have 110v AC but fir big appliances they use 220v
<Ansuel>
Kitchen have 2 plug 110v and 220v
<Mangix>
120/240 actually
<Mangix>
It's that way because two phases are present.
<Ansuel>
Here it's mono 220 if I'm not wrong
<Mangix>
One room has phase 1 and the one next to it has phase 2. Makes power line networking extremely slow
<Mangix>
As for big appliances, it depends. I have no 240V plugs nor appliances in my house
<Mangix>
Fridge is connected to a 20A 120V socket
<Ansuel>
Oven?
<Mangix>
Mine is gas. Probably a regular 15A outlet
<Ansuel>
Also I think it's easier to do AC to DC than DC to AC
<Ansuel>
I also have methane but Russia made us discover that low price it's nothing permanent lol
<hauke>
changing voltages of AC is pretty easy and efficient using a transformer, charging voltages of DC is not so easy.
<f00b4r0>
Ansuel: not sure why you think DC is more dangerous than AC. It's quite the opposite in fact.
<Mangix>
f00b4r0: because the current is direct
<f00b4r0>
precisely. It's less dangerous.
<Mangix>
that's...what makes it more dangerous
<Mangix>
it doesn't go backwards at any point.
<robimarko>
?
<f00b4r0>
AC provokes muscular convulsion with each phase change
<f00b4r0>
DC shocks you out
<robimarko>
^Exactly
<Ansuel>
DC is not dangerous just because we use it at low W. Afaik for high W AC is better. And dangerous it's not only life risk but also fire hazard
<robimarko>
Well yeah, high enough of DC voltage and it can kill you
<Mangix>
Oh I wasn't thinking about danger to humans. I was thinking electronic failure
<f00b4r0>
Ansuel: you'll have to elaborate on that, because I can't quite correlate with my current knowledge :)
<Ansuel>
Run 3kw on DC and I expect more heat and resistance than AC
<f00b4r0>
it's the exact opposite.
<robimarko>
Yeah
<f00b4r0>
I'm afraid your expectations don't align with the underlying science ;D
<f00b4r0>
DC transmission is far more efficient than AC. That's why HVDC transmission lines are a thing.
<Mangix>
a bit unfair probably
<robimarko>
Unfair to who?
<Ansuel>
But AC main usage was low loss on long transmission o.O
<Mangix>
HVDC being more efficient than AC. is it more efficient than HVAC :)?
<f00b4r0>
;3
<Ansuel>
It is but HVDC is very high compared to AC from the few articles I have read at times
<f00b4r0>
Ansuel: no, not really. It's primarily a matter of convenience
<f00b4r0>
given the prevalent generation
<Ansuel>
So AC is used just because it's easier to transform to the required voltage in DC?
<robimarko>
Its a matter of history as well
<Ansuel>
History was DC is shit in long transport and AC solved the problem
<robimarko>
HVAC transformers were a thing way before HVDC ones
<robimarko>
Especially since HVDC uses higher voltage
<Ansuel>
I guess HVDC weren't a thing at all at times lol?
<Mangix>
well
<robimarko>
HVDC has a nice feature that it doesnt suffer from the skin effect
<Mangix>
here in the US, split-phase 240V is used. Probably a result of earlier cables being subpar
<Mangix>
every electrical decision is historical
<f00b4r0>
Ansuel: definitely not the case. HVAC comes with all sorts of annoyances such as skin effect, hysteresis losses, necessary ballast loads, synchronization, etc.
<Mangix>
when I was in Europe, I saw a lot of light coming from the wall outlets on connection. I guess because of higher voltage
<Mangix>
I don't see it here
<f00b4r0>
Mangix: that's more likely the sign of shitty equipment with high surge current being plugged in
<robimarko>
That is common for capacitive loads
<f00b4r0>
^
<Ansuel>
Ok from the map it's mostly underwater stuff so why not used for terrestrial? Probably the line are already there?
<f00b4r0>
they are
<f00b4r0>
the other primary issue is reliability
<f00b4r0>
can't really have your cake and eat it, you see :)
<robimarko>
Well, HVDC complicates things quite a bit
<Ansuel>
So AC is more reliable ?
<Mangix>
f00b4r0: laptop charger
<f00b4r0>
Mangix: typical.
<f00b4r0>
unlimited inrush current
<robimarko>
Yeah, that would explain the "spark"
<Mangix>
more light in europe than the US
<Ansuel>
My DELL charger love to make sparks
<robimarko>
It takes quite some juice to fill up those capacitors
<Mangix>
Ansuel: mine's DELL too
<robimarko>
Typical lets make it as small and cheap as possible SMPS
<Ansuel>
Same for gans usbc 130w charger
<f00b4r0>
Ansuel: AC is more reliable by virtue of not need active components to "operate"
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<f00b4r0>
it also is easier to build AC circuit breakers, these are more reliable
<Ansuel>
But using DC would make everything very efficient correct?
<f00b4r0>
quenching a HVDC arc isn't trivial, you see :)
<Mangix>
how many amps are mains breakers in europe? mine's 100A
<robimarko>
That varies widely
<hauke>
Ansuel: chageing voltage of DC is not efficent compared to AC
<Ansuel>
My parents house is single breaker and pray to God nothing bad happen lol
<hauke>
Mangix: the power gird is different per country in europe, but they are all interconnected
<Ansuel>
Actually 2 breaker one from the electrical company for 2.7kw max
<Ansuel>
haule my scenario was a solar panel system where you do double conversation
<f00b4r0>
Ansuel: sounds like a recipe for disaster ;P
<Ansuel>
haule*
<Ansuel>
Damn autocorrect
<Ansuel>
f00b4r0 also no earth
<f00b4r0>
to answer your previous question, as hauke pointed out, it's difficult to match the efficiency of AC transformers in DC. High power trafos can reach 99.5% efficiency. I don't think DC is quite there yet (but I haven't followed)
<f00b4r0>
Mangix: residential power distribution in France maxes out at 60A single phase, or 3x60A 3-phase
<f00b4r0>
it used to be possible to subscribe up to 90A single phase but that's no longer offered.
<Mangix>
it's 100A across both 120V phases here
<f00b4r0>
the US isn't the largest energy hog and polluter in the world for nothing ;3
<robimarko>
To me the weirdest thing is California with 208V
<Mangix>
f00b4r0: not sure what you mean
<Mangix>
robimarko: or you. I live in California. I've never heard of 208V
<f00b4r0>
Mangix: i mean "bigger is better" has a reverse side.
<robimarko>
Mangix: I follow some HVAC channels and they use 208V 3 phase all of the time for Air Conditioning
<Mangix>
oh 3 phase
<Mangix>
some industrial thing
<robimarko>
Seems to be spread out to all of the restaurants
<Mangix>
f00b4r0: still not sure what you mean
<f00b4r0>
¯\_(ツ)_/¯
<Ansuel>
No idea what I find so fascinating in HVAC repair channel
<f00b4r0>
bigger appliances means more power used means more amps means larger power plants means more pollution.
<f00b4r0>
clearer that way maybe? :)
<Ansuel>
There is one channel that I have like every damn video watched
<robimarko>
HVACR videos or?
<Ansuel>
But I would love to switch to UK channel... They have massive complex LG system
<f00b4r0>
just look at the size of a US fridge vs an European one to illustrate the above.
<Mangix>
hehe
<f00b4r0>
clearly targeting different species :)
<robimarko>
Well, that is one thing EU is good at
<Ansuel>
Yep hvacr
<robimarko>
Forcing energy efficiency
<robimarko>
Good luck selling your A class apliance if competition has A++
<Ansuel>
America is also pushing some of it
<Ansuel>
And hvac are hating them
<Ansuel>
For some reason VFD doesn't like sand and high temp
<Ansuel>
No idea why !?!!
<Ansuel>
Also the meme efficiency scale got updated... Meme A+++++ device got declassed to D ahahahah
<Mangix>
robimarko: to be fair, the greens are all over California, pushing electric cars and whatnot
<robimarko>
Mangix: That is not what I mean
<robimarko>
They have been imposing energy efficiency on home appliances for decades now
<robimarko>
And the effect is visible
<robimarko>
Ansuel: It was time to reset the scale, it makes sense
<Ansuel>
Yep was just curious how trash the scale was initially
<robimarko>
And shockingly FW loading works on mainline as well
<f00b4r0>
robimarko: I'm afraid they're pushing it a bit too far nowadays. Especially on lighting. They no longer consider the output/power ratio, but seem to primarily consider pure power usage only. Meaning if you actually need more light, you end up with silly ratings :P
<robimarko>
f00b4r0: I agree on that
danitool has joined #openwrt-devel
<f00b4r0>
they're also forcing LEDs down everyone's throats when there are better alternatives for specific use cases (such as HID or LPS)
<Ansuel>
Well at least that moves the LED tech even further
tchebb has joined #openwrt-devel
<Mangix>
to be fair, they have Energy Star ratings here as well
<robimarko>
Well, its hard to sell HID/Xenon when LED equivelant uses less power and that is all the end customer cares about
<f00b4r0>
Ansuel: to a degree. It displaces long proven tech, forces renewal on perfectly fine fixtures (think public lighting) and generates a lot of waste.
<f00b4r0>
I'm also dubious what kind of "recycling" used LED fixtures (which no longer use "bulbs") will face.
<f00b4r0>
of course Chinese LED manufacturers rejoice.
<Ansuel>
That is a problem on his own... Also company overvolting LED and reducing life of 80%...
<Ansuel>
All waste
<f00b4r0>
*nod*
<Ansuel>
Meanwhile in Arabia LED light that will probably last more than the building itself
gladiac has joined #openwrt-devel
<f00b4r0>
ah, you've seen that :)
<Ansuel>
Eheheh
<Ansuel>
I'm also very curious to modify some led lamp
<f00b4r0>
shows how corrupt mfgrs are and how screwed we are.
<Ansuel>
Cut the resistor
<Ansuel>
Bam +20 years
gladiac has quit []
<Mangix>
I have no LED in my house
* Mangix
hides
<Ansuel>
What do you have ?
<Mangix>
CFLs mainly
<f00b4r0>
i still use halogen in specific places where I want CRI=100
<f00b4r0>
and no flicker
<Ansuel>
CFL are ok
<Ansuel>
Just slow to emit the full light
<f00b4r0>
CFLs are awful.
<f00b4r0>
anyway, we've gone quite off topic and it's about dinner-o-clock here ;)