Werner changed the topic of #armbian to: armbian - Linux for ARM development boards | Armbian 20.11 Tamandua released | www.armbian.com | Github: github.com/armbian | Commits: #armbian-commits | Developer talk: #armbian-devel | Forum feed: #armbian-rss | Type 'help' for help | Logs: -> irc.armbian.com
<stipa>
usb power?
<stipa>
switchabel USB power?
<stipa>
not yet...
<stipa>
you'll switch bpard on and off via rs232?
<stipa>
boards*
sunshavi has quit [Ping timeout: 260 seconds]
sunshavi has joined #armbian
<IgorPec>
lanefu: you use this for powering boards?
redentor has joined #armbian
<stipa>
i guess he's going to do some rs232 network betweene the boards and switch somehing via USB...
kanito has joined #armbian
<lanefu>
Yeah for SBC power
<IgorPec>
i can only power on/off all
Mony has quit [Ping timeout: 256 seconds]
<IgorPec>
together with switch
<stipa>
lanefu: has something fancy
<lanefu>
Yup 8 individually controlled.. and then serial console for some as well
<stipa>
how are you going to connect to the SBC that controls everything?
<lanefu>
I2c gpio expander. The 5 pin JST connector on that hat is for the ioexpander. 3 pins are uart
<stipa>
oh, you are going to send signals I2c signals to that board without shelling into it?
<lanefu>
No not quite. That SBC is going to control the relays and collect the serial consoles. That 8ch relay board had a gpio pin for each relay. So im just using a 8pin gpio expander to manage it via i2c
<lanefu>
It will be on the network
<stipa>
why is this for "The 5 pin JST connector on that hat is for the ioexpander." ?
<stipa>
additional GPIO?
<stipa>
yes
<stipa>
i got it
<lanefu>
Yup
<stipa>
so, the boards that will be connected to the master one will not be connected to the network?
<stipa>
that will be connected via uart*
<lanefu>
Everything will be on network. Power control and serial consoles is to just help with testing and debugging.... i might finally add another vlan for test devices like igor does. Still undecided
Mony has joined #armbian
<stipa>
why not
<IgorPec>
i got two vlans
<stipa>
fancy
<IgorPec>
to separate boards with more nic
sunshavi has quit [Ping timeout: 264 seconds]
<stipa>
i have no idea what's going on with vlan
<stipa>
is it if things crash and so that your main netwoek don't crash with boards under test?
<IgorPec>
no, its for security conerns
<stipa>
oh
<stipa>
it doesn't see the internet?
<IgorPec>
those networks are outside primary network
<stipa>
that vlan
<IgorPec>
it does, but it doesn't see my office network
<IgorPec>
but i see them
<stipa>
ph
<stipa>
oh*
<IgorPec>
we also have a plan to connect those test farms together
<stipa>
over internet?
<IgorPec>
yes
<IgorPec>
all those devices will be on the same network
<stipa>
that's an awesome idea
<IgorPec>
well, our internal network
<stipa>
i didn't experiment with vlan yet
<stipa>
sounds cool
<IgorPec>
vlan is just having more networks inside one physical
<stipa>
i can maybe understand what an virtual nic is, but not vlan, that's above my skills
<IgorPec>
virtual lan
<IgorPec>
physical is cable
<stipa>
i got that yes
<IgorPec>
and inside that cable you can have more virtual "cables"
<IgorPec>
you infrastructure must support that ofc
<stipa>
more than one vlan?
<stipa>
oh man, that gets complex
<IgorPec>
yes, virtually unlimited
<IgorPec>
no, that's not complicated :)
<stipa>
sounds interesting
redentor has quit [Remote host closed the connection]
<stipa>
maybe i should find an use for one
redentor has joined #armbian
<IgorPec>
you have some router or switch to experiment om?
<stipa>
so if one vlan crash others stay alive?
<stipa>
i have some "routers"
<IgorPec>
no, this is not failsafe
<IgorPec>
failsafe is something else
<IgorPec>
best example would be ..
<IgorPec>
that you have two rooms connected with one cable
<IgorPec>
but you want to drive several networks over it
redentor has quit [Client Quit]
<IgorPec>
lets say financial departnement and sales
<IgorPec>
they have their own IP space and they should be isolated
<[TheBug]>
stipa: thnk of vlan as a form of packet tagging
<IgorPec>
connection between rooms goes via one "trunk" connetion, which is a mixture of more VLANs
<IgorPec>
[thebug] hey.
<IgorPec>
trying to find simplest explanation :)
<stipa>
sounds lake many NAT over one cable
<IgorPec>
nat is again something else :)
<stipa>
that's the closest to vlan i can understand
<[TheBug]>
vlan is really more so at the router level, its adding specialized tags to the packet to indicate where it should be allowed to route, if the device doesn't allow or isn't setup to pass the vlan your on then you won't have access, if it allows your vlan, then is does. Usually vlans are a number, such as vlan 1, 2 , 15 etc.
<[TheBug]>
you're*
<[TheBug]>
so yes, you can do similar with iptables if you want, kinda similar concept to connmark
<stipa>
what does the tagging, some hardware?
<[TheBug]>
Router / switches
<[TheBug]>
managed switches can be programmed only to pass allowed vlans
<stipa>
that have "vlan" tech?
<stipa>
hmmm, i guess there is some linux hack for that?
<[TheBug]>
the vlan is specificed at the routing level, then you can have smart networking devices also manage packets using the same vlans -- this allows more of a phsyical seperation of networks even if they share the same actual switch hardware
<[TheBug]>
"In essence, a VLAN is a collection of devices or network nodes that communicate with one another as if they made up a single LAN, when in reality they exist in one or several LAN segments. In a technical sense, a segment is separated from the rest of the LAN by a bridge, router, or switch, and is typically used for a particular department. This means that when a workstation broadcasts packets,
<[TheBug]>
they reach all other workstations on the VLAN but none outside it. "
<[TheBug]>
I mean granted the page is full of sales garbage also
<[TheBug]>
but they kinda lay out what a vlan is in semi-laymens terms
<[TheBug]>
Usually you start to talk more about vlans as you start to build a large or cororate network where you need to have isolation for security
<[TheBug]>
corporate
<stipa>
got it
<stipa>
it's a necessity then i guess
kanito has quit [Ping timeout: 240 seconds]
<[TheBug]>
at your house... probably not.. in a small business, probably not a bad idea idea to investigate as you grow but probably not on your radar, a growing company with several departments or a hosting company with multiple large customers on the same equipment for example, then you start to discuss vlans
<stipa>
got it
<[TheBug]>
A lot of consumer routers 'fake' a vlan for you for using with 'guest' wifi network for example so that they can still access internet but not for example other machines on the network. This is kinda a dumbed down version of vlan as the person connecting to the limited guest network is tagged and the routing table knows that items tagged as such are not to have access to local devices and
<[TheBug]>
only over WAN.
<stipa>
neat
<[TheBug]>
likely though using connmark and not actual vlan, but similar concept
<stipa>
i would like to think of many vlans on a network as layers
<stipa>
but i think that's not so
<stipa>
like. layers of networks
oida has quit [Ping timeout: 240 seconds]
<[TheBug]>
well thats not the right term just because the work layer in networking has it's own meaning
<[TheBug]>
um how would I define it.. thats a good question
<[TheBug]>
s/work/word
<ArmbianHelper>
[TheBug] meant to say: well thats not the right term just because the word layer in networking has it's own meaning
<stipa>
it's weird
<[TheBug]>
all the networks are operating on the same 'layer' is the thing
<stipa>
to imagine
<stipa>
hehe
<stipa>
so, even if they're seperated they exist in the same bubble
<stipa>
which is correct
<stipa>
but since they're "Virtual" things get nasty of what it actually is
<stipa>
in virtual world
<stipa>
they're seperated
<stipa>
in physical not
<stipa>
better not to go there...
<[TheBug]>
stipa: see PM
<[TheBug]>
instead of spamming here
<stipa>
sorry
<[TheBug]>
no I just ddin't want to keep scrolling the chan here is all
<[TheBug]>
just provided some more explanation there instead
oida has joined #armbian
pikudoz has joined #armbian
pablocastellanos has quit [Quit: Leaving.]
pablocastellanos has joined #armbian
pablocastellanos has quit [Ping timeout: 240 seconds]
<lanefu>
now time to figure out i2c
sunshavi has joined #armbian
<stipa>
in a nutshell it's a bus that can communicate with more devices at once, not at the same time of course but
<stipa>
that's it's selling point
<stipa>
devices have adresses
<stipa>
and i2c uses that to find them on the bus
pikudoz has quit [Ping timeout: 256 seconds]
archetech has quit [Quit: Leaving]
<lanefu>
yah got my crappy expander working
<lanefu>
relays be clickin
<stipa>
there's nothing more soothing than a click of a relay
<[TheBug]>
unless its coming from a UPS
<[TheBug]>
that can sure tighten up your sphincter on short notice
<stipa>
click click
<stipa>
[TheBug]: i had tightened sphincter a wgile ago
<lanefu>
So scope is you looking at audio output or you're using scope data as some sort of input
<stipa>
that is output on the scope
<stipa>
of the signal generated in the software
<stipa>
right
<lanefu>
10 years ago i gave up on linux audio and just bought a mac and used garage band lol
<stipa>
yea
<stipa>
it sux
<stipa>
but there are some tricks to make it useful
<stipa>
yesterday i wasn't so optimistic
<stipa>
:D
<stipa>
the problem with linux audio is that delay compensation is not there yet
<stipa>
because some processors of the sound need a lot of time to process something
<stipa>
and it creates delay
<stipa>
so signals have to compensate
<stipa>
to be in time
<stipa>
and that sux on linux
<lanefu>
Are you running real time kernel?
<stipa>
i guess not
<stipa>
but that wouldn't help
<stipa>
because i use FIR filters
<stipa>
they by nature need to buffer signal
<stipa>
before they process it
<stipa>
so delay is there you want it or not
<stipa>
it's the nature of that filter, i guess realtime kernel wouldn't help much
<stipa>
but i don't know
<stipa>
maybe realtime kernel is needed to make it work, but for now noone is mentioning that real time kernel could fix that
<stipa>
it's like hard to do
<stipa>
hard to code
<stipa>
but it's working with commercial software on win and osx
<stipa>
payware*
<stipa>
the story is that noone has that skillset and that opensource guys have to pay someone to do it...
<stipa>
on windows for example newcomers don't even know what latency is
<stipa>
it's not a problem
<stipa>
if you stay virtual all the way of course
<stipa>
but, i'm making it to do what i want
<stipa>
i could work with how things are now, it could be done
<[TheBug]>
vwwwwww23333333e
<[TheBug]>
er doh :Z
<[TheBug]>
cleaning my keyboard haha, sorry
<stipa>
cleaning a mice while os is running can be nasty as well...
<stipa>
mouse*
<stipa>
but, i must admit, i have never celaned mouse while pc was off
<ArmbianTwitter>
@DieZuckerbude (Ben Zucker π°): @armbian Nice. It didn't when I tried yesterday against draft PR. Well then I'll give it another go. https://tinyurl.com/y8kkn7uo (24s ago)
oida has quit [Ping timeout: 240 seconds]
oida has joined #armbian
<Werner>
Good morning
<kprasadvnsi[m]>
Good morning
_whitelogger has joined #armbian
<ArmbianTwitter>
@DieZuckerbude (Ben Zucker π°): π Oddly enough I did a quick resource and found something a few years old by somebody that namespaces were missing. However this kernel has full namespaces support... I hope Armbian will not have this issue. https://t.co/hHpaXlefU9https://tinyurl.com/ybzrpd8u (21s ago)
_whitelogger has joined #armbian
_whitelogger has joined #armbian
<IgorPec>
good morning
<Werner>
trying to get 2nd nic of R4S up. r8168 module wont load for me though :(
<Werner>
r8168: module PLT section(s) missing
<IgorPec>
i got it working
<Werner>
how?
<IgorPec>
at least on 5.9.y
<Werner>
hm
<IgorPec>
perhaps not enabled in config
<Werner>
dkms 8168 or native 8169?
<IgorPec>
native
<Werner>
can you check lsmod for 8169?
<Werner>
bc this one is not set in dev, I am building atm
<IgorPec>
# CONFIG_R8169 is not set
<IgorPec>
CONFIG_R8169=m
<Werner>
ok so 8169 driver is good enough to catch 8111 too in the meantime. awesome
<IgorPec>
yes. i have 5.10.y on and second nic is not working
<IgorPec>
so this is it
<Werner>
confirmed. installed new kernel package and did not even need to reboot. modprobe and the interface poped right up
<IgorPec>
ok, make a change. i will make images. should we move it to .conf as well?
<Werner>
Not sure. Id consider mark it wip and move to conf once we get feedback from piter who also ordered one but1G model
<IgorPec>
ok
<Werner>
commited
<IgorPec>
building
<Werner>
Will go building too. My 3rd cup of coffee :)
<IgorPec>
i am on 1st :)
<IgorPec>
i am working too much with american shift :)
<Werner>
haha
<Werner>
Can you push new version of vscode to repo by chance?
<lanefu>
Freedom will make you tired
<IgorPec>
vscode update is scripted , so no big problem
<IgorPec>
i could hook it to auto
<IgorPec>
lanefu: good morning? :)
<Werner>
Hm there is a new version out for quite a while. How often is the script executed?
<lanefu>
Ha yeah im on werner time its 0500
<Werner>
11am here :)
<Werner>
kind a switched positions :P
<Werner>
any recommendation for rk3399 heat sink? my copper overkill thingies which perfectly fit H3 and H6 are too small
<IgorPec>
script is still manually executed. i'll do it asap
<Werner>
awesome
<IgorPec>
water cooling ?:)ΕΎ
<Werner>
will do with 2,2ghz oc overlay enabled :D
<Werner>
or just throw it into a bucket with distilled water
<lanefu>
Yeah id go bucket approach
<lanefu>
You guys see my video i posted last night? Got my relay control working with my janky io expander
<Xogium>
why would you put computers in water ?
<Werner>
water cooling :)
<IgorPec>
to heat water?:)
<Werner>
rofl
<Xogium>
I mean, even if it doesn't conduct electricity because it's perfectly distilled, wouldn't it heat up the water anyway ?
<Werner>
the beast could do that
<Werner>
sure it does but would take a while
<Xogium>
it's like the mineral oil I guess, except that for some reason it likes to climb up cables and get sucked dry
<Xogium>
hah speaking of that, I gave up putting the heatsink I bought with the stm board on it⦠Placing it is fine, but good god. it's super hard to line up correctly with the board, check that I don't touch any metal, and put it down in the exact same place that I did earlier with the heatsink still protected by its plastic film
<Xogium>
except if you remove that, and place it wrong, well you,re kinda toasted. It likes to stick
<Werner>
happened to me often enough that I shorted a running board by misplacing copper heat sink on it ^^ thankfully none died from it yet
<Xogium>
heh, in this case I have to mind the rows of gpio pins right next to the chip
<Xogium>
I'd say I have about the thickness of a popcycle stick of margin
<Xogium>
I think it's a bit easier if you at least what you're doing though hehe
<Xogium>
*at least see
<Werner>
just ordered the 20x20 copper heat sink. my other heat sinks seem to fit perfectly on the memory
<Xogium>
I guess I need to find someone who knows how to put up heatsink like these to do it for me
<Werner>
there is something wrong with cryptroot. fs does not expand if it is enabled
<Werner>
I also played with the build tools from fe yesterday a bit. their OS works with overlay. They have rootfs on a ro partition and all userfiles and changes are applied via overlay partition
<Werner>
quite interesting concept. though I do not like that they store kernel and uboot on non-readable separate partitions
<Werner>
so quite difficult to put usbstoragequirks for example
<IgorPec>
not sure if autoexpanding ever worked on cryptoroot
<Werner>
Probably not
<IgorPec>
if it works well, is already nice
<Werner>
It basically works (however --type= option seems broken) but expanding could become tricky. Did not further investigate
<ArmbianTwitter>
@DieZuckerbude (Ben Zucker π°): @FriendlyARM_ #nanopi R4S now running latest mainline kernel on #armbian. Also 2nd NIC can now be used. Very nice. https://tinyurl.com/y7r65b93 (8s ago)
kanito has joined #armbian
kanito has quit [Read error: Connection reset by peer]
hackersCardgame_ has joined #armbian
<ArmbianTwitter>
@armbian (armbian): RT @DieZuckerbude: @FriendlyARM_ #nanopi R4S now running latest mainline kernel on #armbian. Also 2nd NIC can now be used. Very nice. https://tinyurl.com/ydeue5m2https://tinyurl.com/y7hed7aw (4s ago)
kanito has joined #armbian
_whitelogger has joined #armbian
kanito has quit [Ping timeout: 272 seconds]
horiatm has joined #armbian
<horiatm>
hy
<lanefu>
IgorPec: do you think 5 sec power off is okay for SBC reboot, or should i let it "rest" longer
<IgorPec>
ok
horiatm has quit [Remote host closed the connection]
<Werner>
depends
archetech has joined #armbian
<ArmbianTwitter>
@lanefu (Lane Jennison): I made a hat for my @librecomputer TritiumH5 to connect to TTL console and control power to SBCs for @armbian testing. Since using i2c IO expander, made some quick docs and examples https://t.co/jlOPTpqcDZhttps://tinyurl.com/yay8sln9 (2s ago)
marcodiego has joined #armbian
<Tonymac32>
I wonder if this "Lane Jennison" person would want a PCB designed with the option of global shipping? :P
<Tonymac32>
I also have to wonder if that is a TI part, an NXP part, or an HGSEMI part
<Tonymac32>
I'm guessing NXP given video
archetech has quit [Quit: Konversation terminated!]
archetech has joined #armbian
stipa_ has joined #armbian
drobo_00 has joined #armbian
drobo_00 has quit [Client Quit]
drobo_00 has joined #armbian
stipa has quit [Ping timeout: 240 seconds]
<lanefu>
Tonymac32: TI Chip
<Tonymac32>
It's one of the fun IC's that is somehow made by 3 companies :P
<lanefu>
haha yeah i had a feeling
<Tonymac32>
probably doesn't matter which
<lanefu>
i wonder if lighter relays could be used
drobo_00 has quit [Quit: drobo_00]
<Tonymac32>
maybe, but given the price I don't see a point
<lanefu>
was mostly thinking of footprint
<marcodiego>
is there any sbc supported by armbian that can run entirely deblobed?
<Tonymac32>
I don't think there are any blobs in the Allwinner boards at this point, and the Asus Tinker should be as well (minus the Wifi)
<Tonymac32>
I guess it depends on how much function you want from completely deblobbed
<Tonymac32>
WiFi is always a problem
<Tonymac32>
and Mali FOSS drivers aren't super awesome yet
<marcodiego>
I follow the linux-sunxi effort. But A64 development seems to have slowed down recently
<Tonymac32>
The RK3399 in theory can have a blob-free bootloader, but that varies board/board since the support isn't so great yet, especially if it's DDR4
<Tonymac32>
Amlogic is a hopeless case
<marcodiego>
I have a rockpi4, the model without wifi. I think this is the closest to a fully deblobbed experience right now