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
jl4_ has quit [Ping timeout: 245 seconds]
mrueg has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
<nekomancer[m]> `sd 1:0:0:0: [sdb] tag#0 device offline or changed` trying to write to usb flash on odroid-n2. 2 different flash. no errors with usb-sata HDD.
<nekomancer[m]> wft?
<nekomancer[m]> same usb flash drives reads and writes without errors on RockPi-4.
Ntemis has joined #armbian
rneese has left #armbian ["Leaving"]
<[TheBug]> nekomancer[m]: what that makes me think is the USB port not putting our reliable power
<[TheBug]> nekomancer[m]: try powered USB hub and see same issue
Ntemis has quit [Read error: Connection reset by peer]
* nekomancer[m] not have powered usb hub around :(
<nekomancer[m]> but eeror with flashes — and no error with hdd??!
<nekomancer[m]> hdd usb powered too. spinning hdd
mrueg has joined #armbian
<[TheBug]> then other thought would be heat
<[TheBug]> those usb sticks can overheat
<[TheBug]> could be that because of how fast it writes the memory chip is getting really hot or the controller
<[TheBug]> could try taking it out of housing and see if same issue for example
<nekomancer[m]> yes, it hot. but it works with PC and rockpi4 SBC. being hot.
<[TheBug]> Yes, but could actually write to it slower? is it close to any other device when in n2?
<[TheBug]> just playing devils advocate here cause I can't think of another eason for that
<[TheBug]> reason*
<nekomancer[m]> I try to use only 1 and 2 flashes, occupies 1 and 2 usb ports on N2. one by one and both. Try different ports.
<nekomancer[m]> all thesame/
<nekomancer[m]> I have not so much usb flashes I can use to this test... generaly only 3
<nekomancer[m]> seems same for all
<nekomancer[m]> all usb fkashes I try generates io errors trying to read files on my N2
<nekomancer[m]> o! I found usb-microsd adapter. will try to do the same with sdcard in that usb adapter
Luc52 has joined #armbian
<Luc52> Hello guys have sombody an old debian 9 based image for bananapi m1?
<Luc52> the current one is slow as hell
<lanefu> nekomancer[m]: definitely try powered hub. N2 usb power is a huge failure
<lanefu> My only guess
<Luc52> thank u so much
<nekomancer[m]> =8-o
R0b0t1 has quit [Ping timeout: 260 seconds]
Elpaulo has quit [Read error: Connection reset by peer]
Elpaulo has joined #armbian
<Tonymac32> agree with Lanefu
<Tonymac32> I've experienced power issues on USB w/N2 as well
<lanefu> hell yeah have geoip redirects on 50% traffic
<Tonymac32> it's always awkward when lanefu talks dirty
<lanefu> imma freaky dude
<Tonymac32> An observation I made on RK3399 earlier and looked at more closely today, there is no clock-latency property on all but 1 opp
<Tonymac32> I'm going to add the property as it appears in the vender kernel since it isn't deprecated
<Tonymac32> and rk3328 has it
<Tonymac32> so why it's not in rk3399? no idea
<Tonymac32> "stuff happening before clock is stable" can cause ded lockups
<Tonymac32> D.E.D DED
<Tonymac32> :D
<lanefu> how on earth did you figure that out
<lanefu> but awesome!
<Tonymac32> I thought the device tree looked a little sparse when I was making the overlay XD
<Tonymac32> I honestly don't know if it will do a damned thing, but it looks like it should be there
<lanefu> works for me lol
<Tonymac32> A metric might be to test if you get performance differences before and after. This will slow down operations that cause DVFS to hop opps
<Tonymac32> as tkaiser has pointed out when showing performance vs other governors, you lose clock cycles during a freq change
<lanefu> oh interesting
<lanefu> makes sense tho
<Tonymac32> I'm also moving to just working on 5.10 since that is the LTS
<lanefu> so with performance it will still throttle down if temp is too hi
<lanefu> 5.10 flies
<lanefu> it turbo boosted PBP
<Tonymac32> excellent
<Tonymac32> I have been told the pcie took a hit though, do you have any hardware to test that on? I'm doing other stuff as described at the moment
<lanefu> afraid not.. Bug is your IO guy
<Tonymac32> haha he brought it up
<Tonymac32> ok
<lanefu> oh yes he kept me very informed
<Tonymac32> well I have potentially the most RK3399 devices ever
<Tonymac32> so I'll see if I can get some answers after opp season
<Tonymac32> XD
<lanefu> ha
<Tonymac32> I have rockpi 4b, NanopI M4V2, and PC T4, PBP, and a Tinker Edge R that I can't make boot anything but the asus image so far. :'(
<Tonymac32> oh, and Rockpro64's
<Tonymac32> but, well...
archetech has quit [Quit: Leaving]
Unit193 is now known as JackFrost
<lanefu> man you are rich in rk
<Tonymac32> I also have a pile of Tinker Boards. Rumor has it a new one will be joining them XD
<lanefu> tinkerpile
<Tonymac32> Makes a lot of heat
<lanefu> alright geoip is fully deployed
<Tonymac32> and can do a stupid amount of math
<Tonymac32> the Tinker is still a better desktop board than most of what we support
<Tonymac32> exceptions being N2 and RK3399 stuffs
<Tonymac32> Anything Mali450 is just doomed
<lanefu> i wonder how much more performan rk3399k is
<Tonymac32> I wouldn't worry about it unless it's the same thing as the OP1
<Tonymac32> it's 2.0/1,5 GHz
<Tonymac32> My understanding is it shares a lot with the Edge R
ChriChri_ has joined #armbian
ChriChri has quit [Ping timeout: 246 seconds]
ChriChri_ is now known as ChriChri
<lanefu> so is rk3399 op1 different than whats in my PBP?
<Tonymac32> yeah, it seems to be a minor revision fixing some random stuff
<Tonymac32> it uses lower voltages per frequency
djacer2k has joined #armbian
djacer2k has quit [Remote host closed the connection]
archetech has joined #armbian
<Tonymac32> we'll see if the wider audience does alright with it
_whitelogger has joined #armbian
_whitelogger has joined #armbian
<Sebastian[m]1> Tonymac32: Hey cool, I will test that patch for my media tree (https://git.linuxtv.org/media_tree.git/) build, on the NanoPC-T4. As you said above, this can be tested by checking the performance before and after. Do you have a favorite method of checking performance on ARM boards? Ah and additionally I maintain a set of patches in order to use the latest kernel version, while reading the chat messages I above I noticed
<Sebastian[m]1> that you also run the latest kernel for your rk3399 machines. Do you have a repository where you keep your up-to-date patches by any chance? Thanks, Sebastian.
<Sebastian[m]1> * Tonymac32: Hey cool, I will test that patch for my media tree (https://git.linuxtv.org/media_tree.git/) build, on the NanoPC-T4. As you said above, this can be tested by checking the performance before and after. Do you have a favorite method of checking performance on ARM boards? Ah and additionally I maintain a set of patches in order to use the latest kernel version, while reading the chat messages above I
<Sebastian[m]1> noticed that you also run the latest kernel for your rk3399 machines. Do you have a repository where you keep your up-to-date patches by any chance? Thanks, Sebastian.
archetech has quit [Quit: Leaving]
archetech has joined #armbian
a__pi has joined #armbian
<a__pi> i'm building armbian, and it's failing at E: Failed to fetch https://armbian.hosthatch.com/apt/dists/focal/main/binary-armhf/Packages.gz File has unexpected size (646150 != 645411). Mirror sync in progress? [IP: 443]
<a__pi> there are some of the packages from 5 days ago
<a__pi> is this correct?
<Werner> Building should work anyways. The image just does not come with up to date repository information
<a__pi> yes yes
<a__pi> image build correctly, but i think there is some problem in the sync script :)
<a__pi> not sure if it's know issue or no
<Werner> Yeah, we are aware of this but could not figure out the root cause yet
_whitelogger has joined #armbian
archetech_ has joined #armbian
archetech_ has quit [Client Quit]
archetech has joined #armbian
archetech has quit [Quit: Konversation terminated!]
archetech has joined #armbian
<ArmbianTwitter> @DieZuckerbude (Ben Zucker 🍰): @armbian @StationGeeker Station P1 review part 2: https://t.co/nxG8FTicag (20s ago)
<ArmbianTwitter> @motakasoft (motakasoft): GitHub Trending Archive, 18 Dec 2020, Shell. barrykn/big-sur-micropatcher, keycloak/keycloak-containers, confluentinc/examples, 21y4d/nmapAutomator, FydeOS/chromium_os-raspberry_pi, markshust/docker-magento, supercrabtree/k, armbian/build https://t.co/SqZ8LbOXXB (16s ago)
<ArmbianTwitter> @armbian (armbian): RT @DieZuckerbude: @armbian @StationGeeker Station P1 review part 2: https://t.co/nxG8FTicag (19s ago)
alexanderl has quit [Ping timeout: 260 seconds]
<Werner> Is it possible to get an 2.0/1.5GHz overlay to work on legacy too?
Luc52 has quit [Quit: Connection closed]
<Sebastian[m]1> Do these patches work for 5.11.1 ?
alexanderl has joined #armbian
<Sebastian[m]1> I thought the dev tree is always a few version after the most recent kernel version.
<Werner> In the future maybe
<Sebastian[m]1> Ok, yes that's what I meant in my question. I currently work on the most recent version of the liinux kernel in order to write & test patches on the correct base. And I wondered if some of you might maintain their own repos with the latest changes to make the build work.
<Sebastian[m]1> I currently maintain this for example: https://github.com/initBasti/NanoPC-T4_armbian_configuration (which only contains a subset of the patches)
alexanderl has quit [Ping timeout: 240 seconds]
alexanderl has joined #armbian
<Werner> You may want to talk to Tonymac32. He recently did a major cleanup of the rockchip-current patches and somewhen this will be necessary for dev as well
<Sebastian[m]1> Alright, cool thank you, will do :)
torv has quit [Remote host closed the connection]
torv has joined #armbian
margoaron has joined #armbian
_margoaron has joined #armbian
margoaron has quit [Ping timeout: 264 seconds]
Xogium has left #armbian ["leaving channel"]
Xogium has joined #armbian
<ArmbianTwitter> @imprashere (Prasanna Venkadesh): For more than a year had issues with my #OrangePiPc board. I thought it's literally dead since HDMI port was damaged and images of @armbian got struck on boot and I had no clue of why without a dispaly. SSH in armbian images are enabled but for some reason, it didn't work for me. (9s ago)
<ArmbianTwitter> @DieZuckerbude (Ben Zucker 🍰): @imprashere @armbian USB serial adapter is the way to go for debugging such issues. (13s ago)
<ArmbianTwitter> @imprashere (Prasanna Venkadesh): Only option was to debug via UART USB to TTL. But I did not have an UART convertor and so did not attempt. Armbian and OrangePi forums are full of such boot related issues. Yesterday dusted it off again and tried OrangePiPc+ image from @yunohost (16s ago)
emOne has quit [Ping timeout: 256 seconds]
<ArmbianTwitter> @imprashere (Prasanna Venkadesh): @DieZuckerbude @armbian True. I learnt about it much later but didn't try since I did not have one. I am planning to buy one and have it so it becomes handy for debugging next time. (5s ago)
emOne has joined #armbian
rneese has joined #armbian
<ArmbianTwitter> @imprashere (Prasanna Venkadesh): What made these possible? 1. Self hosting is impossible without FOSS (@yunohost, @armbian + all FOSS web services) 2. Open Source Hardware (Orange Pi) 3. Human labor for 1 and 2. Planning to donate to them and contribute code as well (22s ago)
R0b0t1 has joined #armbian
stipa has joined #armbian
rneese has left #armbian ["Leaving"]
<ArmbianTwitter> @armbian (armbian): We need help on testing new feature. Implement Device Tree Editor https://t.co/yxGMLeRDtV #armbian (0s ago)
<Tonymac32> good morning
<Werner> Morning Tony
<Tonymac32> This rockpi 4B power LED is blinding the $hit out of me
<archetech> were some shades
<archetech> wear
<archetech> or a welding mask lol
<Xogium> sometimes I'm glad to be blind
* Xogium hides
<Tonymac32> lol
<Tonymac32> you have to go lower power on green due to the eye being most sensitive to it.
<archetech> wife comes in sees mask on tony these led's are getting too much juice!"
<Tonymac32> they did not, and probably used too small of a resistor at the same time
<Tonymac32> XD
<Tonymac32> the old MiQi RK3288 board has a similar broblem with the Blue LED's
<Tonymac32> you can use it to light a small room
<Xogium> I heard the led on espressobin has the same thing
<Tonymac32> lanefu have you run any default desktop images on your RK3399 stuff? K5.10?
<Xogium> extremely bright, and you can't control it
<lanefu> Tonymac32: pbp
<lanefu> You want me to spin one?
<Tonymac32> the interface performance on trunk is just trash
<Tonymac32> I was curious if you'd seen that
<Tonymac32> it's like the RK3328 before I "noGPU'd it :'(
<lanefu> Hmm lemme flip kernels on my.opi4 and ill do some tests
<Tonymac32> Focal
<Tonymac32> as soon as you start doing anything with a browser X_X
archetech has quit [Quit: Konversation terminated!]
archetech has joined #armbian
<lanefu> Tonymac32: when you say interface you mean ethernets?
<Tonymac32> no, the UI
<Tonymac32> sorry
<Tonymac32> latency++
<lanefu> ane@pinebook-pro:~$ uname -a
<lanefu> Linux pinebook-pro 5.10.1-rockchip64 #trunk.21 SMP PREEMPT Wed Dec 16 01:19:52 CET 2020 aarch64 aarch64 aarch64 GNU/Linux
<lanefu> lane@pinebook-pro:~$
<lanefu> i'm humming along
<lanefu> are you using hte accelerated xorg with the modesetting driver? or fb?
<lanefu> Tonymac32: which board you using
<Tonymac32> Rockpi 4b
<Tonymac32> building an M4V2 image now
<lanefu> Tonymac32: from desktop branch?
<Tonymac32> No, from trunk.
<Tonymac32> the graphics were just an observation, I'm trying to murder the boards and am doing some benchmarking to test out my adjustments
archetech has quit [Quit: Konversation terminated!]
raver has quit [Remote host closed the connection]
<Tonymac32> from my build today: http://ix.io/2ITl look at dmesg output
<lanefu> Tonymac32: ahh panfrost noise
<Tonymac32> of lovely
<Tonymac32> lol
<lanefu> i mean mine does that
<lanefu> on pbp
<lanefu> but i'm humming along having a godo time
<lanefu> playing youtube music over bluetooth via firefox
<Tonymac32> so the image I ust DL'd what is it?
archetech has joined #armbian
<lanefu> that's desktop focal mate/ w nightly RK3399 dev kernel from repo
<Tonymac32> I'll try yours, my m4v2 dev kernel is bootlooping (no changes my side to dts)
<Tonymac32> it's the control image XD
<lanefu> yeah sometimes you need to use someobdy else's image then get made that theirs works
<Tonymac32> well the M4V2 is not a very stable piece of hardware in my experience
<ArmbianTwitter> @guidol70 (Guido Lehwalder): @armbian Also with the URL I couldnt see/find out how to use the Device Tree Editor and with which version of armbian it will be displayed in armbian-config :( (21s ago)
<Tonymac32> fark, your image is bootlooping too
<lanefu> hmm
<lanefu> i'm build a few other rk3399 desktops
<lanefu> will link yu to those
<Tonymac32> rp4b, roc-rk3399-pc, NanoPC T4, rockpro64 (maybe, usually won't boot)
<lanefu> Tonymac32: crap that's from an older checkout.... but kernel should stillbe nightly
<lanefu> anyway working on a rockpi 4b now
<Tonymac32> my RPi4B also booted funny just now, I think we need to look at 5.10 closely
<ArmbianTwitter> @useidel (Udo Seidel): #ubiquiti controller running on old #bananapi M1 using #armbian (#buster) with some help from #raspberrypios (11s ago)
<lanefu> Tonymac32: yeah sounds like some tuning is needed
<ArmbianTwitter> @meuon (Mike Harrison): @CommieGIR @hexadecim8 Armbian or straight up Debian on better hardware, especially with eMMC, rocks. Start here: https://t.co/8HdZrARhMn (1s ago)
<ArmbianTwitter> @armbian (armbian): @hexadecim8 me on a bit better hardware? (10s ago)
<ArmbianTwitter> @armbian (armbian): @guidol70 Changes were not merged anywhere - they only exists at a merge request atm. If this 1st testing goes well, we can push them up. (11s ago)
<lanefu> Tonymac32: roc-rk3399-pc renegade elite?
<Tonymac32> yes
rneese has joined #armbian
<rneese> ok are most of you enduseers or any of you desktop user/devs
<rneese> we need morre desktop devs
<lanefu> rneese: preachin to the choir homeboy
<rneese> rockpi
<rneese> wwe dont need ro rockpi
<Tonymac32> well my M4V2 won't boot, so
<Tonymac32> XD
<lanefu> Tonymac32: i'm just building rk3399 images for Tony to debug / compare to his
<lanefu> rneese: ^^
<rneese> what did you do to it tony
<Tonymac32> nothing, it's K5.10
<IgorPec> probably we need to put this - we need desktop devs - word outside our community
<Tonymac32> it booted 5.9 or 5.8, can't remember which
<Tonymac32> hi Igor
<rneese> yes IgorPec
<rneese> agree
<lanefu> Tonymac32: so weird thing is my PBP is rocking K5.10.1 no problema
bzyx has joined #armbian
<Tonymac32> I'm uncomfortably wondering if this is rams and routings related, building a NAnoPC-T4 image now
<rneese> we need a t4 with the 5.10 kernel
<rneese> post when its done
<ArmbianTwitter> @racerrupnik (Chris Rupnik): @armbian @hexadecim8 Armbian, on my multiple orangepi devices, has been wonderfully stable for me. Thank you for all the hard work! (4s ago)
<lanefu> Tonymac32: okay i'm using trunk.21 from dec 16th https://beta.armbian.com/pool/main/l/linux-5.10.1-rockchip64/
<Tonymac32> lanefu my build with my device tree adjustment didn't have the purge messages during benchmark. Probably coincidence but
<lanefu> lanefu: yeah.. those messages happen randomly.. it's like a buffer that flushes
<IgorPec> i think we have everything on 5.10, at least in dev
<Tonymac32> yeah I'm working out of dev since 5.10 is LTS
<rneese> doo we have a dl for dev iimgs or is it build only
<Tonymac32> but I'm having strange things happen
<Tonymac32> M4V2 bootloops
<Tonymac32> Rock Pi B 1.3 isn't booting every time just right either
<lanefu> rneese: build only... although i'm doing this right now https://armbian.lane-fu.com/desktop-preview/
<IgorPec> m4v2: i boot if succesfully few days ago.
<Tonymac32> I have an NVMe plugged in, so I have to wonder if it's tied to [TheBug]'s comment on speed issues there
<lanefu> IgorPec: I'm building with beta repo images.. is it safe to say that this image will have a broken kernel install because of the error? https://armbian.lane-fu.com/linx/gf3pdowo.txt
<lanefu> A: yes lol
bzyx has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
<Tonymac32> lanefu how random is the purge message?
<lanefu> Tonymac32: fairly frequent, but varies a lot.. more load, more rfrequent http://ix.io/2IUi
<IgorPec> lanefu: surely that's seom problem
<IgorPec> because we don't build roc-pc dev
<lanefu> archetech said its a panfrost-ism... it's been happening for a longn time
<lanefu> I think that's a red herring
<Tonymac32> I'm trying to make it happen and cant is all
<Tonymac32> XD
<lanefu> Tonymac32: is panfrost loaded?
<lanefu> are you desktoping?
<Tonymac32> it's identical to the image I got the errors on, just with the modified device tree
<lanefu> hmm
<Tonymac32> so watching some youtubes while mashing menus
<Tonymac32> nanoPC T4 flawless victory
<IgorPec> waht was it?
<Tonymac32> I don't know yet, but the thing booted perfectly
<Tonymac32> the DDR4 boards did not, could be coincidense though
<Tonymac32> The T4 is also recognizing it's NVMe
<Tonymac32> I2S is toast on RK3399-dev though it looks like
<Tonymac32> at least however it's set up on the Rock Pi and T4
raver has joined #armbian
<rneese> which omg
<rneese> tony
<rneese> bbiab
<Tonymac32> panfrost_gem_shrinker_scan is what generates the Purging message. I don't know it's purpose or what causes a purge, especially one that is rate limited on reporting
<Tonymac32> it's memory reclamation, yes, but I don't know why it's so damned noisy
<lanefu> probalby because htey left a debug message on
_margoaron has quit [Ping timeout: 265 seconds]
<Tonymac32> That's not my point
<Tonymac32> why so many consecutive calls to the same function so fast?
<Tonymac32> anyhow we can patch the message out no problem
<Tonymac32> Tinker has a Midgard, so does the H6, correct?
<Tonymac32> (waiting on sbc-becnh for the T4)
<Tonymac32> oh, and XU4
<a__pi> mali-t760 <- this is a a tinkerboard gpu
<a__pi> first tinkerboard
<Tonymac32> correct
<Tonymac32> XU4 is a T628
<Tonymac32> H6 is a T720
margoaron has joined #armbian
bzyx has joined #armbian
bzyx has quit [Client Quit]
<Tonymac32> I'm curious if anyone has memory reclamation seizures on the others
bzyx has joined #armbian
<lanefu> my opi4 did it last time i lokked
<Tonymac32> I expect all RK3399's do
<lanefu> ya
* Tonymac32 is chasing an invisible fault tree analysis in his head because waiting
* lanefu foolishly tried accelerated desktop on odroidn2 with kernel 5.9
* Tonymac32 just found opi3 under desk on floor oops
<lanefu> haha
<lanefu> odroidn2 makes me so mad lol
PipeItToDevNull has quit [Remote host closed the connection]
PipeItToDevNull has joined #armbian
<Tonymac32> Rock Pi 4B membench
<Tonymac32> standard memcpy : 1755.8 MB/s
<Tonymac32> standard memset : 8229.2 MB/s (0.1%)
<Tonymac32> That is DDR4
<Tonymac32> So NanoPC T4
<Tonymac32> standard memcpy : 1801.7 MB/s
<Tonymac32> standard memset : 7987.4 MB/s (0.2%)
<Tonymac32> DDR3
<lanefu> bency
<lanefu> benchy
<Tonymac32> bouncy?
<Tonymac32> :D
<Tonymac32> http://ix.io/2IU9
<Tonymac32> http://ix.io/2IUK
<Tonymac32> [TheBug]
<Tonymac32> My NVMe on the T4:
<Tonymac32> ok, got a few purge notifications on the T4. Still none on the RockPi 4
emOne has quit [Ping timeout: 240 seconds]
<lanefu> Wow. So there really is zero X11 acceleration with N2. Even wjth legacy
fuho has joined #armbian