Werner changed the topic of #armbian to: armbian - Linux for ARM development boards | www.armbian.com | Github: github.com/armbian | Commits: #armbian-commits | Help our crowdfunding: https://bit.ly/3kcuCvj | Type 'help' for help | Logs: -> irc.armbian.com
<lanefu> yeah ifb is just for old school stuff. need a qos tool oriented around nftables
<stipa> it doesn't seems old
<stipa> lanefu: it's used here https://github.com/tohojo/sqm-scripts
<stipa> it's constantly mentioned all over the place for todays qos open solution
<lanefu> for like openwrt?
<stipa> yes
<TRS-80> yes this is a CAKE implementation
<TRS-80> fq_codel, etc
<stipa> sch_cake is available as a module, i checked it
<stipa> it should be compiled in
<TRS-80> Perhaps that is an nfstables based solution, as lanefu mentioned if preferable?
<TRS-80> s/if/is/
<ArmbianHelper> TRS-80 meant to say: Perhaps that is an nfstables based solution, as lanefu mentioned is preferable?
<stipa> TRS-80: I have no idea
* archetech is doing an ipfire setup tonight 3 nics
<stipa> CAKE is like new hot thing
<stipa> lanefu: yeah, we had this conversation, ifb isn't available in the current build for my card
<stipa> but i did research after our chat and it led be to ifb again
<lanefu> yeah doesnt look like a lot of options its a wtf for sure
<lanefu> what SBC?
<stipa> orangepi lite 2
<lanefu> you're traffic shaping on the wifi board lol
<stipa> not yet
<stipa> i hooked usb3 hub, usb nic and usb wifi to it
<stipa> nic is gigabit, wifi is two antenna N ht40
<stipa> swifi should be AC soon
<stipa> nic connects to the 4G modem
<stipa> wifi inbard is not in use
<stipa> onboard*
<stipa> that works
<stipa> need a qos now
<stipa> but ifb module is fucking me
<stipa> had to go to 5.4.75, i hope it works.
<stipa> i hope it'll boot :D
<stipa> can't do it now, it's running the wifi network
<stipa> i'll leave it for saturday
<TRS-80> buy more devices, so you have spares (when you can)
<TRS-80> I mean in general, it's just a good idea
<stipa> yeah, plan to buy one more and make a wifi repeater, for now laptop is doing it
<TRS-80> I think I have 2 good,tested of these WNDR3800 that I settled on, and wouldn't mind 1-2 more
<TRS-80> but I plan to expand out network to other locations
<stipa> yeah, i think i got hooked up as well
<stipa> but i want to keep as much stuff on one sbc as possible
<stipa> don't know how long that idea is going to last...
<stipa> but i need one SBC more for the repeater, that's for sure...
<lanefu> stipa: OMG i found it in the kernel finaly
<stipa> lanefu: what?
* TRS-80 intensly awaits exciting conclusion
<stipa> ifb?
<stipa> :/
<lanefu> config IFB
<lanefu> tristate "Intermediate Functional Block support"
<lanefu> depends on NET_CLS_ACT
<lanefu> select NET_REDIRECT
<lanefu> help
<lanefu> This is an intermediate driver that allows sharing of
<lanefu> resources.
<lanefu> To compile this driver as a module, choose M here: the module
<lanefu> will be called ifb. If you want to use more than one ifb
<lanefu> device at a time, you need to compile this driver as a module.
<lanefu> Instead of 'ifb', the devices will then be called 'ifb0',
<lanefu> 'ifb1' etc.
<lanefu> Look at the iproute2 documentation directory for usage etc
<lanefu> stipa: using debian or ubuntu?
<stipa> i'm compiling on ubuntu server version
<stipa> 5.4.0-52-generic
<stipa> inside of the Virtual Box
DigitalMan1983_ has joined #armbian
<DigitalMan1983_> hey gang
<stipa> yo
<TRS-80> hey DigitalMan1983_
<DigitalMan1983_> how's everyone doing tonight?
<TRS-80> Hungry! But Mrs. should be home soon with pizza.
<DigitalMan1983_> nice
<TRS-80> and there is the message; time to go make drinks and eat pizza
<TRS-80> afk
* [TheBug] slaps DigitalMan1983_ around a bit with a large trout
<DigitalMan1983_> hahahahaha
<DigitalMan1983_> that will never cease to be funny
<lanefu> stipa: building with ifb
<lanefu> what a goosechase
<stipa> lanefu: i wouldn't know, i would be fine with legacy image that has ifb module running on my board but if it can be current it would be nice.
* DigitalMan1983_ chases gooses
<lanefu> stipa: kernel config used http://ix.io/2ETq
<stipa> lanefu: it's downloading
<lanefu> needless to say i was wrong about ifb being deprecated
<stipa> np lanefu
<lanefu> how fast is xfer?
<stipa> around 2 megabytes
<stipa> per second
<lanefu> not bad.. and you're in like the eu?
<stipa> yes
<stipa> in Croatia
<stipa> but in my hometown the speed is better
<stipa> more than 50Mbs download on 4G
<lanefu> yeah thats pretty good
<lanefu> but i'll aktake uploading 20mbit overseas on a friday night as a win
<stipa> it's max my temporary gsm cell can offer to me
<lanefu> stipa: if it works... run armbian-config and freeze kernel
<stipa> ok
<stipa> i'll try it tomorrow
<lanefu> cool good luck
<stipa> thanks
<lanefu> ya glad to help
ChriChri_ has joined #armbian
ChriChri has quit [Ping timeout: 264 seconds]
ChriChri_ is now known as ChriChri
archetech has quit [Quit: Konversation terminated!]
archetech has joined #armbian
<lanefu> who was the person thats into sdr?
DigitalMan1983_ has quit [Ping timeout: 260 seconds]
<TRS-80> I am, but maybe others too?
<lanefu> wondering if htere's any sdr data crunching you can do in a distrubted fashion
<lanefu> not liek multiple radios, but just distrubte whatever crunching from what you recorded
<TRS-80> Well, there are a number of distributed computing projects already like SETI, folding at home, etc. so I think maybe? Even probably?
<TRS-80> In fact that is exactly what SETI is (I think?)/
DaRock has joined #armbian
<lanefu> you know what woudl be cool a bash autocomplete script for armbian builder
<lanefu> TRS-80: yeah i know that's out there
<lanefu> but my buddy adn I have a reallly nerdy idea
<lanefu> he's been workng on a portable nomad cluster made from SBCs
<lanefu> and since covid we thought maybe we'd meet in the middle and go camping
<lanefu> but bring that cluster adn like sniff the air lol
<TRS-80> I don't know how much you know about radio stuff in general, but the antenna will be by far the most important part, espescially if you are out in the country.
<lanefu> yeah makes sense
<lanefu> radio is his job
<lanefu> fortunaetly
<TRS-80> Last field day we were at the local FD training center and ran couple huge antennar from the top of the fire tower (for example).
<TRS-80> you talking about some sort of seeking / auto tuning / multiple channel monitoring software?
DigitalMan1983_ has joined #armbian
<lanefu> i have no idea
<lanefu> just looking for ways to crunch data
<lanefu> in a distribued way
<TRS-80> There is a whole entire section of Debian repository dedicated to Ham Radio: https://packages.debian.org/stable/hamradio/
<TRS-80> Distributed? I dunno. But maybe walk before running? ;)
archetech has quit [Quit: Konversation terminated!]
<TRS-80> I have the cheap and famous RTL-SDR, but haven't played with it too much yet.
sunshavi has quit [Remote host closed the connection]
<TRS-80> and that's a first post!
<lanefu> yeah thats pretty dope
s_frit_ has joined #armbian
s_frit has quit [Ping timeout: 260 seconds]
TRS-80 has quit [Quit: WeeChat 2.9]
drobo_00 has quit [Ping timeout: 260 seconds]
<Tonymac32> wut
<Tonymac32> object detection without letter agency inspection, hmmmm
<Tonymac32> that's on an H3, so with a real SoC...
<Tonymac32> (Not a dig at the H3, that just isn't it's forte)
* Tonymac32 needs to get his license for radio
DigitalMan1983_ has quit [Ping timeout: 256 seconds]
DigitalMan1983_ has joined #armbian
drobo_00 has joined #armbian
drobo_01 has joined #armbian
drobo_00 has quit [Ping timeout: 256 seconds]
drobo_01 is now known as drobo_00
archetech has joined #armbian
Elpaulo has joined #armbian
drobo_00 has quit [Ping timeout: 260 seconds]
<archetech> moved router >router setup to ipfire behind isp router browser spped doubled
<archetech> speed*
<archetech> and blocking whole countries :)
Elpaulo has quit [Quit: Elpaulo]
DaRock has quit [Ping timeout: 240 seconds]
DaRock has joined #armbian
DaRock has quit [Ping timeout: 265 seconds]
DaRock has joined #armbian
archetech has quit [Read error: Connection reset by peer]
archetech has joined #armbian
manpaz has quit [Remote host closed the connection]
raver has joined #armbian
Net147 has quit [Read error: Connection reset by peer]
Net147 has joined #armbian
Naka has joined #armbian
Nakaori has quit [Ping timeout: 264 seconds]
stipa has quit [Remote host closed the connection]
archetech has quit [Ping timeout: 240 seconds]
<ArmbianTwitter> @jerry_hopper (jerry hopper): The latest the Altis Times! https://t.co/DE89LdVbzQ Thanks to @trico_arma @armbian @StratSentinel #armyfamily #dyk (9s ago)
archetech has joined #armbian
DaRock has quit [Ping timeout: 260 seconds]
DaRock has joined #armbian
Net147 has quit [Read error: Connection reset by peer]
Net147 has joined #armbian
Heisath has joined #armbian
stipa has joined #armbian
stipa has quit [Quit: WeeChat 2.8]
archetech has quit [Ping timeout: 256 seconds]
stipa has joined #armbian
stipa has quit [Client Quit]
RzR has joined #armbian
RzR has quit [Changing host]
stipa has joined #armbian
<stipa> lanefu: your IFB image is running ok. First time it booted ok and i transfered network config files to the new sdcard through serial console. Second time network booted up but serial console doesn't allow me to login, it hangs and spits out kernel messages if for example i plug something into USB it registers it. I can shell into it with putty but can't login into it via serial console. I suspect
<stipa> that's normal, here is the conosle screenshot https://ibb.co/dj0HVx4
<fromport> i am running one of my odroid hc2 with a more recent kernel. just checked it and found this is dmesg: https://termbin.com/jw6av
<fromport> is it worth reporting that somewhere ?
<stipa> lanefu:if i unplug usb hub where all my network interfaces are then it boots with option that allows me to login into machine through serial console again.
<stipa> lanefu: but then i don't have network :D
<stipa> fromport: i guess your problem activates at [56282.584430] swapper/0: page allocation failure: order:0, mode:0xa20(GFP_ATOMIC), nodemask=(null),cpuset=/,mems_allowed=0
<fromport> Mem: 1.9Gi 730Mi 836Mi total used free
<stipa> lanefu: the board booted up with all interfaces connected but when i brought them up or something serial console doesn't allow me to login into the machine, it just spits kernel messages.
<stipa> fromport:yeah, but something happened at [56282.584430] , that's seconds after you booted your board
<stipa> maybe something eated all your ram
<stipa> just a guess
<stipa> maybe a memory leak or something
<stipa> in some software or python script
<fromport> i am running lizardsfs distributed filesystem. rest of my 'chunckservers' are running 5.4.72-odroidxu4, this one is running 5.8.16-odroidxu4. the other units i checked don't have any messages in dmesg.
<stipa> fromport: some erors aren't logged, maybe a forum thread like this https://askubuntu.com/questions/152716/how-to-detect-a-memory-leak could help
<fromport> like i said: i have 10+ of these machines running the same code/task none of them except the one with the experimental kernel gave me this error
<fromport> i will downgrade the kernel to same one as the others. I thought it might be useful to whoever compiles the experimental kernels to get feedback of events
<stipa> it's great you didn't upgrade all boards
<stipa> fromport: also, if you didn't post it on Arbian forum
<stipa> and stick around, peeps here a re sleeping probably...
DaRock has quit [Ping timeout: 256 seconds]
Net147 has quit [Read error: Connection reset by peer]
Net147 has joined #armbian
<lanefu> fromport: worth sharing some where along with an armbianmonitor -u
redentor has joined #armbian
<lanefu> stipa: that's certainly odd. I can't spend anymore time on it tho. You could experiment with that kernel config I gave you with armbian builder
<fromport> lanefu: open a new topic on the forum?
<lanefu> Yeah. Make sure your clear about that file system you use
<lanefu> Probably best if you can do the armbianmonitor thing after it happens again
<stipa> lanefu:oh, i just replace your config with the original one linux-sunxi64-current.config ?
<fromport> i hadn't rebooted it yet: http://ix.io/2EXt
<fromport> am using xfs file system, will mention that as well
<lanefu> stipa: yes or lead it from the kernel menu config and make sure you resave as .config
<stipa> lanefu:ok, thanks
lanefu has quit [Ping timeout: 256 seconds]
stipa has quit [Quit: WeeChat 2.8]
stipa has joined #armbian
Net147 has quit [Read error: Connection reset by peer]
Net147 has joined #armbian
<fromport> just opened a topic on the forum with my kernel swap message: https://forum.armbian.com/topic/16064-kernel-swap-error-with-5816-odroidxu4-running-on-odroid-hc2/
lanefu has joined #armbian
<lanefu> pineH64 crashed.. lemme tell you how happy i am aotu that
<Werner> :/
<lanefu> its funny cuz nicod says his runns cooler than his opi3
<lanefu> and this thign runs hot
<Werner> same kernel?
<lanefu> probably not
<lanefu> stipa: 's problem with the image i gave him makes me slightly weary of 5.9 sunxi
lanefu has quit [Client Quit]
<fromport> werner what does csc mean ?
oida has joined #armbian
<Werner> no official support/community support only
raver has quit [Remote host closed the connection]
lanefu has joined #armbian
<fromport> would it have been better to post it in p2p?
<lanefu> ... and my kernel update bricked it
* lanefu might be possible it was a kernel i was messing with to fix opi3
<Werner> fromport, if you mean the hc2 topic: no. was just a mistake
Net147 has quit [Read error: Connection reset by peer]
<Werner> hc4 is csc, hc2 is fine
Net147 has joined #armbian
<stipa> lanefu: the board boots now into login prompt via serial console. I didnt send hostapd in to the background like '& > /dev/null 2>&1' and boot hanged at that place. It's all good now, thanks again.
drobo_00 has joined #armbian
<lanefu> stipa: glad its working... don't forget to freeze kernel
redentor has quit [Quit: Leaving]
<stipa> lanefu:i did
<lanefu> historically bulletproof sunxi64 is getting on my bad side
<lanefu> (not because if you stipa ) my opi3 and pine-h64b are being troublesome
sunshavi has joined #armbian
igorp_ has joined #armbian
drobo_00 has quit [Ping timeout: 264 seconds]
<stipa> lanefu: yeah, i'm aware taht H6 has some silicone bugs... It must be a mess for devs and i guess there's a lot of hacking to make it work.
<lanefu> Yeah my opi3 was bulletproof until kernel 5.9
<igorp_> so we go out with 5.8.y ?
<lanefu> That would be my preference
<igorp_> ok, this gives us some space
<igorp_> i guess this Orangepi R1 networking issues is also tied to 5.9?
<igorp_> and works well with 5.8.
<Tonymac32> Does anyone remember off the top of their head what the compression algo was that worked best on ARM? I know tkaiser had some stuff about it in the forums but I'm having a hard time filtering for it
<lanefu> Tonymac32: lzo i believe
<lanefu> hmm mauybe thats not true
<Tonymac32> I remember the results seems backward from what you'd expect
<lanefu> Tonymac32: look on his github he has a repo called "knowlledge"
<Tonymac32> somehow I remember LZ4 somehow sucked
<lanefu> lzo is what armbian zram uses
<Tonymac32> that would be a good bet then
<lanefu> Tonymac32: yeah i double checked on teh forum adn tried to filter through the scolding and I think lzo is in fact the winner
<lanefu> and if lzo-rle is an option you may want to try that
raver has joined #armbian
<lanefu> man he hated us for a long time. tearing the project a new asshole in 2018
<Tonymac32> O-RLY?
<Tonymac32> that is just personality type lol
<Tonymac32> I know a *lot* of people like that in autom,otive
<buZz> didnt they just remove lzo from zram?
<Tonymac32> *they* being?
<[TheBug]> if your t alking about tkaiser, yeah, thats just his personailty type
<lanefu> buZz: good question.. tell us what you know
<lanefu> spill the beans
<lanefu> someoen should mine all the tkaiser posts, and translated them into lessons learned and testing procedures
<[TheBug]> buZz: I don't know if they removed it but I though lz4 became the default
<lanefu> its all great shit
<buZz> lzo-rle is the current default
<buZz> hmmz
<buZz> cant find the commit msg
<buZz> maybe plain lzo got removed, lzo-rle is the only remaining
<lanefu> k the armbian-zram script makes allowances for lzo-rle
<[TheBug]> only say that cause I spent 2 hours the other day trying to make zram work for swap on a Centos 7 server and ened up throwing in the towel, but I recall it wanting to use lz4
<lanefu> per me looking at it 5 minutes ago
<Tonymac32> well remember on x86 LZ4 is faster than LZO
<Tonymac32> but my brain is remembering testing via Herr Kaiser saying nope on ARM
<[TheBug]> hehe
<[TheBug]> strange how his words penetrate your brain.. maybe we need a vaccine for that :D
<[TheBug]> lol
<Tonymac32> The Goron Ramsay of SBC's
<igorp_> you faggot
<igorp_> :)
<buZz> :D
<buZz> out of all the cooks-that-insult, he's one of the better
<buZz> but Uncle Roger beats him
<igorp_> not familiar
<igorp_> those kitchen shows are not my area
<igorp_> but its hard to not know who gordon ramsey is
<Tonymac32> haha yeah
<[TheBug]> nah, I imagine your area is out by the fire pit with a cold beer IgorPec ;)
<igorp_> speaking on beer. i only got one beer left at home :(
<buZz> just someone responding to cooking videos
<[TheBug]> IgorPec :(
<buZz> its pretty damn good :P
<stipa> gordon ramsey ?
<buZz> Uncle Roger
<[TheBug]> Martha Stewart
<buZz> [TheBug]: ooo the weed woman
<[TheBug]> LOL, nm, I was trying to be funny and make a pattern of random cook show hosts names, stipa had started... haha
<buZz> :)
<buZz> martha stewart made videos together with snoop dogg
<[TheBug]> indeed
<igorp_> tonymac32: did you find time for N2 u-boot? Otherwise I might try it
<[TheBug]> gotta make that money
<igorp_> it should not be much different than C4
<[TheBug]> if Snoop showed at your house and asked you to make him dinner, I am pretty sure you would too, ahahah
<Tonymac32> My job kept me 60 hours this week :'(
<igorp_> yeah, likewise
<Tonymac32> and no it shouldn't be, I think it should only need the boot script you put in C4
<buZz> [TheBug]: i'd make him better food that this martha stewart garbage :P
<igorp_> it has been insane week. only today i spent half of the day with kids
<buZz> than*
<Tonymac32> I'm not sure why the person in the PR is having trouble, but then again some other stuff has been moving
<igorp_> tonymac32: yeah, i think so
<stipa> Uncle Roger : what could be a BBC? LOOOOL
<Tonymac32> le sigh
<[TheBug]> Deep breathe
<[TheBug]> ;)
<stipa> gordon ramsy had to fire all his employes due to COVID, none likes him anymore...
<Tonymac32> hmmm, NicoD and another guy were showing NVMe speed numbers for the M4V2, and they sucked, like 700 MB/s I need to double check the paramters they used
<lp0-on-fire> ugh I gave an honest try to the xr819 just for the fun of having something to do
<lp0-on-fire> lol
<lp0-on-fire> I think I shouldn't have persisted ;)
<lp0-on-fire> [ 9.410063] Unable to handle kernel NULL pointer dereference at virtual address 00000000
<lp0-on-fire> [ 9.406983] 8<--- cut here ---
<lp0-on-fire> [ 9.424512] Internal error: Oops: 5 [#1] PREEMPT SMP ARM
<lp0-on-fire> that was fun
<Tonymac32> I got almost 1400 out of my T4, and I see tkaiser results agree...
<Tonymac32> and lol xradio
* lp0-on-fire had nothing to do while waiting for a wifi dongle to arrive
<lp0-on-fire> yeah lol
<lp0-on-fire> I think the driver is allergic to kernel 5.9
<[TheBug]> lp0-on-fire: planning to re-write the driver? otherwise... its like trying to fixa botched carving... ( https://www.smithsonianmag.com/smart-news/amateur-restorer-botches-carving-spain-180976271/ )
* Tonymac32 has a NanoPi DUO
<lp0-on-fire> some folks reported that wpa_supplicant wouldn't work so I was like… might as well toy with iwd
<lp0-on-fire> I added it, and it was nice enough to tell me that my kernel was missing some settings, and even gave their names
<lp0-on-fire> so I enabled them and I got a kernel oops for my trouble LOL
<[TheBug]> best quote from that article... "Now, the carving is essentially ruined, with some observers saying that it more closely resembles Donald Trump or Mr. Potato Head than a female figure."
<lp0-on-fire> hah, sadly nop. I'm not knowledgeable enough in code, let alone coding low level things like drivers
<Tonymac32> lanefu botfap came in with some notes saing lz4 is far better on intel than ARM. HOWEVER time has passed 2+ years
<lp0-on-fire> my coding is limited to some tiny amount of shell script, but oneliners with absolutely no check aside from set -e and && between commands
<Tonymac32> It is likely time for a re-run
<lanefu> yeah we probalby need to re-valute a lot of optimizations
<lp0-on-fire> better than nothing :D
<Tonymac32> that's the topic where it was discussed
<lp0-on-fire> from what I heard the hardware itself ain't very nice, and the driver just makes it worse
<[TheBug]> lp0-on-fire: yeah thats kinda what I was inferring there ;)
<Tonymac32> :D
<igorp_> we should really go for 1TB of RAM :)
<[TheBug]> wooo
<[TheBug]> big iron
<buZz> :) my mini-itx server takes up to 0.25TB ram, but ECC only
<lanefu> yeah like NVME pricing vs ram pricing
<buZz> nonECC just 64gb max iirc
<lanefu> i always wonder what teh sweet spot is
<igorp_> this one is also maxed out
<lp0-on-fire> driver that much I knew, but the hardware, all I have is fact from others, and no way to check them for curiosity's sake
<lp0-on-fire> I like trying to get hardware working or trying to find why it doesn't work, but sometimes its not a clear way to it
<[TheBug]> lp0-on-fire: for H3droid devs there spent some time trying to make that driver usable in even legacy kernel for Android, let alone Linux, an thats what driver was originally written for - another reason for it being so cuddy, it appears like it was "just get this out the door"
<[TheBug]> s/cuddy/cruddy
<ArmbianHelper> [TheBug] meant to say: lp0-on-fire: for H3droid devs there spent some time trying to make that driver usable in even legacy kernel for Android, let alone Linux, an thats what driver was originally written for - another reason for it being so cruddy, it appears like it was "just get this out the door"
<lp0-on-fire> heh, sad
<lp0-on-fire> but is the thing as bad as is just because of the driver, really ?
<lp0-on-fire> surely not
<[TheBug]> no specifically, you hit it on the nose, the hardware it's self is already 'flakey' and then they attempt to make it usable with a hacky driver set
<lp0-on-fire> well heh that's disapointing
<lp0-on-fire> any thoughts about the xr829 ? I reckon same chip with bluetooth lol
<[TheBug]> thats why everyone who starts down the trek you are taking eventually bails and that driver stays in same state of usefulnes ;Z
<Tonymac32> if it starts with xr I think it's doomed
<lp0-on-fire> yeah
<lp0-on-fire> I can understand them
* Tonymac32 has an Opi IoT
<lp0-on-fire> just like mwifiex
<[TheBug]> yeah Xunlong had a bad habit of just throwing those xr based wifi chips on everything they were producing
<lp0-on-fire> LOL
<Tonymac32> the 96-boards one is worse
<Tonymac32> it's an XR SoC too
<igorp_> [thebug] a new generation is here
<igorp_> of xr :)
<lp0-on-fire> heh… Apparently the opi 0-2 is different though and is using some kind of bradcom chip. I don't have a clue what its worth
<[TheBug]> but my assumption is they were "cheap as chips" and it allowed them to drop their BOM by 1-2$, so... why not use them on... everything..
<igorp_> opi zero2 has allwinner wifi chip
<igorp_> AW559A
<lp0-on-fire> yeah but the thing iside appears to be broadcom based
<igorp_> 859a
<lp0-on-fire> that or buildroot has gone crazy selecting the bcm43xxx firmwares form linux firmwares
<lp0-on-fire> or, who knows, allwinner stole something from bcm too :p
<[TheBug]> igorp_: well they had to do something, they produced so many board with that crap xr chipset maybe tyhey finally ran out of them -- or had a corp purchaser finally wear them out with tickets on it so they finally replaced ..
<lp0-on-fire> they stole from st, why not steal from bcm
<igorp_> [thebug] i am affraid this is xradio v2.0 :)
<[TheBug]> lol
<[TheBug]> well,at least you can say Xunlong is consistent ;D
<igorp_> also with confising naming :)
<[TheBug]> haha
<[TheBug]> bbiaf
<lp0-on-fire> I had bad experience with mwifiex too, and decided I would stick away from marvell. Might as well add xr to my list of bad things hehe
Heisath has quit [Quit: Leaving]
<lp0-on-fire> what's worse miwfiex is mainlined… the driver is just, hum
<lp0-on-fire> spaghetti code ?
<igorp_> xr mad thing? noo :)
<igorp_> s/mad/bad
<ArmbianHelper> igorp_ meant to say: xr bad thing? noo :)
<[TheBug]> lp0-on-fire: usually your better off just getting a decent dongle -- some of the ones I have no problem sugesting are the ones that HardKernel sells directly for their units, they have actually tested them in most case to be stable / have stable drivers and my success rate has been very high -- if you in USA you can check Ameridroid as they also sell them, but they have several modules for
<[TheBug]> somewhere around $20 or less
<igorp_> but i guess both is valid :)
<lp0-on-fire> yeah I saw this one
<igorp_> this one is well supported
<lp0-on-fire> hardkernel seems to be lacking, or I searched at the wrong place… they only seem to sell two wifi modules, 0 and 5a
<igorp_> but you can easily find 8811 on Ebay/Aliexpress
<lp0-on-fire> works well, but worth noting that you need some kind of out of tree driver
<igorp_> no
<lp0-on-fire> really ? Hmm
<igorp_> we pack driver to the kernel
<lp0-on-fire> yeah
<lp0-on-fire> but its not in the mainline kernel I mean
<igorp_> but its not a part of the kernel, true
<lp0-on-fire> if you use vanilla
<igorp_> yep, we add lots of stuff on top of vanilla
<igorp_> not sure if those drivers will get to the mainline. some perhaps, all certainly not
<lp0-on-fire> my project uses buildroot, so we try to stick as close to vanilla as possible and only grab things we absolutely need so its sometimes tryal and error process
<igorp_> for embedded dev, things are fundamentally different, yeah
<lp0-on-fire> but I like to check what armbian is doing so I can replicate the same kind of things
<lp0-on-fire> I'm payed for this :p that's why I really tried my best to get the wifi to behave on opi 0 but, at soime point you're forced to admit defeat
<lp0-on-fire> *some point
<igorp_> we already lost enormous of time for this shit wifi
<lp0-on-fire> either we use ethernet, or the thing will have a dongle sticking out at this point
<lp0-on-fire> yeah I bet
<lp0-on-fire> well I was probably a bit harsh with it myself when I tried to get it to run on 5.9 kernel :D
<lp0-on-fire> 5.8 might have worked
<lp0-on-fire> at least my boss had it working just fine on armbian buster
<igorp_> that's smallest problem
<lp0-on-fire> don't ask me how
<igorp_> it works up to some degree. for IOT data exchange is perhaps enough
<lp0-on-fire> yeah its probably fine to listen to some low bitrate internet radio as demo or something, when you can get it to work
<lp0-on-fire> but I've heard rumors it can even lockup the whole kernel
<igorp_> thats ofc possible
<lp0-on-fire> nothing can really surprise me where I'm at, that's how I see things
<igorp_> likewise :)
<lp0-on-fire> I mean, I once crashed 200 laptops in a single shot because of a bogus wifi packet
<lp0-on-fire> direct kernel panic LOL
<igorp_> that's a good one
<lp0-on-fire> a cascade of failures :D
<lp0-on-fire> just because of one bogus multicast packet
<lp0-on-fire> that was embarrassing actually, what with my laptop being the only one still running ;) there was no doubt it was me
<igorp_> they know who is the boss :)
igorp_ has quit [Quit: Leaving]
igorp_ has joined #armbian
igorp_ has quit [Client Quit]
<lp0-on-fire> lol
<lp0-on-fire> sure that was a big fial and totally not intentional on my end, but it could be used as a DOS technic :p
<lp0-on-fire> *fail
<[TheBug]> lp0-on-fire: though let me make an argument to you to make you possibly rethink your path - even if you got that wifi working 'ok' in one condition, because of how poor the device and drive is in general, how you can you rely on this in a production enviroment? What happens next kernel when now your 'fix' or magic patch, just no longer works? If this is really for work, my suggestion is
<[TheBug]> coming up with a good third party solution (dongle) or maybe consider a diferent similar board / SoC that uses a differnt wifi
<lp0-on-fire> indeed
<lp0-on-fire> that's what I told my boss
<lp0-on-fire> frankly I just tried to get it to behave since I was bored while waiting for my dongle to arrive
<[TheBug]> and thats logical -- though somtimes still business just comes to $, so I understand push back -- though my argument here is that your time is valuable as well and you have proven thats not the solution, so now is the time to invest in finding what your options are, don't just keep beating dead horse ;)
<lp0-on-fire> :) yeah
<[TheBug]> ahh well, see, if for fun thats another thing, hahaha
<[TheBug]> if its a hobby for your self, then enjoy, if it is work and you are seeing issues like that where its expected for production, I would take other path
<lp0-on-fire> well, its both for work and for keeping myself busy
<lp0-on-fire> still, since this is buggy as all hell, it sure won't be used in prod
<lp0-on-fire> might go as far as totally removing it from the device tree
<Werner> The moment you use your non-existent gfx skills to create a new animated emoji because you are bored...
<[TheBug]> lp0-on-fire: to be honest, if you are in a position where you are going to need a large amount of boards for a project that all do the same thing > 1000 then another path can be to ask Xulong (or board vendor) do they have other more reliable options for the wifi chipset -- sometimes there are some updrade options they just don't offer on dev board, but they will produce for testing if you
<[TheBug]> intend for commitment
<lp0-on-fire> hmm
<lp0-on-fire> they suggested the opi 0-2 to us, but it drives up the cost a bit. $12 per unit vs $16, sounds stupid like that but for mass prod
<[TheBug]> cause really a nicer chipset only effects BOM by few $ but when it comes to produccing a dev board and having a budget.. corners get cut
<[TheBug]> haha, see I was already on that same wavelength before you said it ;)
<[TheBug]> but yeah, cost is always one of the thoughts ;)
<lp0-on-fire> plus since the wifi is also allwinner chip, I'm not convinced
<lp0-on-fire> I'm not sure if it works well or not
<[TheBug]> but if you are serious and can show them so, they will provide some different options for testing to find what you need.
<lp0-on-fire> and it does a lot of things we will honestly never need like hdmi port, bluetooth
<lp0-on-fire> I see
<[TheBug]> have you discussed then your own board? maybe you could have trade off, removed undeeded items to make budget for better wifi?
<lp0-on-fire> its my first job so not sure, but it does sound like something my boss would have to do… though maybe not
<lp0-on-fire> yeah the guy has been trying to make his own board for 3 years before he gave it up as a bad job, his previous dev left the project while still having a huge chunk of the company in his name
<lp0-on-fire> probably a miracle he found me
<[TheBug]> Well you seem to have the right interest level, build up your own knowledge, it sounds like you may enjoy it
lanefu has quit [Quit: my empire is curmbling]
<lp0-on-fire> they had a board desgiend by some guy, based on a v3s, but the deal totally failed and the hardware was crap
<[TheBug]> interesting
<[TheBug]> v3s has some interesting use case
<[TheBug]> what failed in their attempt there?
<lp0-on-fire> so they tried the rpi 0, realize how annoying it is to get rpi in bulk which is to say, not possible at all, and stuck to the orange pi 0
<[TheBug]> lp0-on-fire: I ask because I spend some time on v3s occasionally when I have time -- https://licheepizero.us ;)
<lp0-on-fire> but the previous dev totally disagreed on sticking with opi 0 and left
<lp0-on-fire> I'm not sure, I think it was the hardware design itself that was crap
<[TheBug]> intersting
lanefu has joined #armbian
<[TheBug]> what is the actual end goal for your project, or is that something you can't share yet?
<lp0-on-fire> oh, it is to make a speedsaver device
<[TheBug]> can y ou provide more context?
<lp0-on-fire> you put it in your car, and it uses a map and gps to find the local speed limit and your own speed and warn you if you go over the limit
<[TheBug]> interesting
<lp0-on-fire> but the map isn't displayed, only data on a tiny oled display
<[TheBug]> what actual devices do you need to accomplish your goal (wifi, memory, etc
<lp0-on-fire> gps, wifi, storage medium obviously… Memory I'd say even 512 mb of ram is overkill
<lp0-on-fire> wifi would be for updates
<[TheBug]> whats funny / not funny is other than GPS that almost sounds like a use case for C.H.I.P. however, they don't even produce it anymore so..
<lp0-on-fire> don't they ? Well that's no fun
<[TheBug]> well it was a crouwd funded ordeal and then the company had problems
<[TheBug]> you can still find a gross of CHIP Pockets out there on ebay
<lp0-on-fire> the plan was originally to solder the gps via spi or such at first, that was done on the v3s
<lp0-on-fire> but on opi 0 he went with a ublox gps dongle
<[TheBug]> but its because they actually failed to deliver them, went bankrupt and someone bought the devices in auction and then resold them
<[TheBug]> but the company it's self is gone
<lp0-on-fire> damn
<[TheBug]> let me think a moment
<[TheBug]> you say now your using OPi 0?
<[TheBug]> and your cost is what, 12$
<lp0-on-fire> yeah
<[TheBug]> is that with the gps dongle?
<[TheBug]> or is that additional
<lp0-on-fire> nah just the board
<lp0-on-fire> xunlong used to sell the opi 0 for $10, but they raised to $12
<lp0-on-fire> so guessing that, plus gps dongle plus wifi dongle if we go that route, plus oled display… Yeah
<lp0-on-fire> my boss say he wants to sell the end product for $40 afair
<lanefu> [TheBug]: RIP CHIP. I use my pocket chip for telnet bbs
<Tonymac32> transmission was being an ass about being relocated to a different drive lol, I think I got it to behave
<Tonymac32> lp0-on-fire ESP32 + SD with maps + GPS chip + OLED
<Tonymac32> I would not put myself through Linux for that
lanefu has quit [Quit: brb]
<Tonymac32> I can point you in the direction of a good develpment platform though XD
<Tonymac32> j/k j/k
lanefu has joined #armbian
<Tonymac32> lanefu did you see my comment about the gps speedo thingy
<lanefu> I'm glad I keep my home directory on a usb drive
<lanefu> Tonymac32: nope sorry
* lanefu 's IRC _should_ be stable now
<Tonymac32> yeah an ESP32 might actually hit the cost targetat $3.50 a module, and you avoid dealing with Linux for something so simple
<Tonymac32> there's no magic to GPS or OLED or SD storage
<Tonymac32> that requires linux anyway
<Tonymac32> sheesh GPS chips are pricey
<Tonymac32> figure out what module that is
<lanefu> 10 bucks outrageous
<lanefu> but yeah def agree esp32 for simple gps doodad
<Tonymac32> you could probably do this whole thing in micropython TBH
<Tonymac32> :P
<Tonymac32> ffs how many images do we run now? I'm bringing mine back online lol
<Tonymac32> there you go
oida has quit [Remote host closed the connection]
bzyx has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
bzyx has joined #armbian
Elpaulo has joined #armbian
<[TheBug]> Tonymac32: interestingly your not the first to make the ESP32 suggestiojn to lp0-on-fire
<[TheBug]> but yeah I think it was more from area of comfort for the development platform from what I gathered
Tonymac32 has quit [Ping timeout: 256 seconds]
<lp0-on-fire> well I have no idea how esp32 even work :p
<lp0-on-fire> plus also we already have a linux software that does what we ask, and I don't know if there's even such a thing on esp32
<lp0-on-fire> I'm good at linux, but esp32 I don't know it
<lp0-on-fire> but I'm in no way refusing the suggestion ;)
<buZz> esp is fun
<buZz> i just made a 230v extension cord with built-in ntp clock based switching on hours :)
<buZz> esp8266 though, didnt need anything fancy
<lp0-on-fire> the way I see it, it looks like you have to code your own things if you use esp, and that's definitely way above my head
<buZz> yeah its just running a single program
<buZz> but that could be a python interpreter
<buZz> or lua
<buZz> or you could write in c++ and have alllll the power
<lp0-on-fire> and also it would make it a bit meh because we would then need to write our own updater, you can't go around without ever updating the device obviously if only for the maps
<buZz> hmhm
<buZz> thats all true
<lp0-on-fire> whereas in linux we already have a good software base… even an updater
<lp0-on-fire> rauc ftw ;)
drobo_00 has joined #armbian
<ArmbianTwitter> @edrex (Eric Drechsel): Testing host OSes and container runtimes for edge devices. Ruled out #k3s (👆idle CPU usage), #armbian (stale host pkgs). Docker-compose seems solid, and a host OS based on either @AlpineLinux or @yoctoproject. About to test @balena_io, wary of lockin. Any experiences? (28s ago)
drobo_01 has joined #armbian
drobo_00 has quit [Ping timeout: 246 seconds]
drobo_01 is now known as drobo_00
oida has joined #armbian
drobo_00 has quit [Ping timeout: 260 seconds]
drobo_00 has joined #armbian
Tonymac32 has joined #armbian
* stipa Tinker Board 2 SBC advances to RK3399 http://linuxgizmos.com/tinker-board-2-sbc-advances-to-rk3399/
<Tonymac32> yes it does :)
<Tonymac32> and it's a K so it can go 2.0 GHz reliably and shouldn't be as hot
<Tonymac32> also notive wifi via Mini-pcie, and a beefy power jack
<Tonymac32> as long as whatever heat sink they bring can move the heat in that SoC-on-top config it should be a nice board
<stipa> looks good
<stipa> it should crunch some big data
<Tonymac32> better power management than a rockpi most likely, but missing the possibility to use NVMe
<stipa> 20Gbps USB
<Tonymac32> I hope the power management on that board takes advantage of that huge power plug and can actually supply some power to those USB jacks
<stipa> not sure if it's shared by three USB 3.2 ports or per one
<Tonymac32> I don't have one (yet)
<stipa> it's refreshing
<stipa> it's Ausus, right
<Tonymac32> I wish there were newer Rockchip SoC's on the high end, but a process-improved RK3399 is alright for now
<Tonymac32> yes
<Tonymac32> ASUS
<stipa> i dont like that there are no holes for heatsink on the pcb
<Tonymac32> The RK3399 has a few USB 3 ports, so that type C is probably at least independant
<Tonymac32> I agree
<stipa> oh if it's 3x20Gbs thats sweet
<stipa> mcoie
<stipa> + mpcie *
archetech has joined #armbian
c0rnelius has quit [Ping timeout: 272 seconds]
c0rnelius has joined #armbian
<Tonymac32> the connectivity options are usually better on *not TV box* SoC's
<archetech> odroid n2 panfrost Linux groovy-server 5.10.0-odroid-panfrost-arm64
<archetech> grab their kernel and have some fun
<stipa> Tonymac32: I have absoulutely no experience with TV boxes, i'll trust your info.
<stipa> but i got your idea, chips that are used in TV boxes are also in majority of SBC's if that's what you have said.
<stipa> but, it's cool to have a linux on a TV box, i wouldn't mind it at all if i had one.
<stipa> i hear some SBC's like rpi 4 can spit out 4K video, that's not that bad.
<c0rnelius> most of those tv boxes are just garbage from understanding
<stipa> that's all that TV box needs
<stipa> c0rnelius: there must be some gems somewhere
<c0rnelius> stipa: i would imagine, yes.
<c0rnelius> somewhere :D
<stipa> yeah
<c0rnelius> I tend to just use Kodi or Roku's
<c0rnelius> even firesticks suck hard
<stipa> i have fisrst mode of the chromecast
<stipa> model*
<c0rnelius> I had one. I ended up giving to it to someone who needed it more than I.
<stipa> i use it to watch movies on a TV
<stipa> or YT
<stipa> it's integrated into Chrome
<c0rnelius> thats what I do with the whatever SBC I may have Kodi installed on.
<c0rnelius> So the chromecast just became whatever.
<stipa> i see
<c0rnelius> you can even push youtube vids to Kodi
<c0rnelius> Yatse is a great app
<c0rnelius> and there are a few for firefox and chrome you can use to do the same
<stipa> i would probably use those if i had SBC connected to the TV
<stipa> i have only one SBC and it's CLI only and headless ATM
<stipa> but i doubt it would do video reliably because it has only 1GB of ram
<c0rnelius> I tend to just go the headless route myself.
<c0rnelius> I have so many headless boards doing something around the house. Be it used as a cam, or embedded in a speaker box to be used as pandora player or bluetooth player.
<c0rnelius> or both
<stipa> IoT guy
<c0rnelius> I have broken rpi zero w I use as a clock on my desk
<stipa> so it still works?
<c0rnelius> yeah works fine
<c0rnelius> I even soldered a wifi dongle to it :D
<c0rnelius> for shits and giggles
<c0rnelius> i was playing around with putting a battery on it yesterday
<stipa> like UPS for SBC?
<c0rnelius> sorta yeah
lanefu has quit [Quit: WeeChat 2.8]
<stipa> that's cool
lanefu has joined #armbian
<c0rnelius> A lot of garbage tech can be reused. Say you have old bluetooth speaker ur no longer using "speaker died on it". The bits inside are still usable though.
<c0rnelius> including the rechargeable battery.
<stipa> they are, but if you don't know what to do with them they are just garbage
<stipa> it's fascniating how much you need to know and learn to see value in garbage
<c0rnelius> even board for them usually have holes for where headers wouold be "i guess for debugging?", they just need to be soldered in.
<stipa> serial port i guess
<c0rnelius> something like this
<stipa> designers leave it in
<stipa> they had used it while prototyping the board
<stipa> product*
<c0rnelius> yeap
<stipa> it's an easy way to access the processor
<c0rnelius> The hard part is finding how they interfaced with it and what software was used... and is it available online?
<stipa> programming/debugging port of the main processor
<c0rnelius> and if not... can I hack myself in?
<stipa> yeah, you'll have to reverse enginner something like that
<stipa> gather the information about the chips
<c0rnelius> yes
<stipa> as much info as you can
<stipa> research it
<stipa> and eventually you'll be able to read the code from the flash if it's not locked
<c0rnelius> so the lil shit board become harder to deal with... the reusable parts though, like the rechargeable battery. Not to much of a science behind this.
<stipa> but some lock can be broken
<stipa> yeah, chips are smaller and smaller
<stipa> and everything is going SoC
<stipa> into one chip
<stipa> but there will always be markets for electronics with "big" parts that can be managable by a human.
<stipa> you can't fix laptop or a phone without a microscope
<stipa> and how to get parts is a nightmare
<stipa> but people are doing it
<stipa> c0rnelius: if you haven't seen how they fix iphone X you haven't seen nothing :D
<stipa> it's fun