Werner changed the topic of #armbian to: armbian - Linux for ARM development boards | www.armbian.com | Github: github.com/armbian | Commits: #armbian-commits | Forums Feed: #armbian-rss | This channel is logged -> irc.armbian.com
jakogut has quit [Remote host closed the connection]
ChriChri_ has joined #armbian
ChriChri has quit [Ping timeout: 272 seconds]
ChriChri_ is now known as ChriChri
sunshavi has quit [Ping timeout: 264 seconds]
dddddd has quit [Remote host closed the connection]
xecuter has quit [Remote host closed the connection]
xecuter has joined #armbian
Tenkawa has quit [Quit: Leaving.]
sunshavi has joined #armbian
lanefu has joined #armbian
lanefu has quit [Changing host]
lanefu has joined #armbian
xec has joined #armbian
<R0b0t1> anyone have leads on how I can increase i2c bus speed?
xecuter has quit [Ping timeout: 244 seconds]
<R0b0t1> ok, revised question: if I compile a kernel how do I modify the dts?
<plntyk> add a patch file in patch/kernel/<architecture> of armbian build system that changes the dts
<plntyk> or use the userpatches infrastructure i think - https://docs.armbian.com/Developer-Guide_User-Configurations/
<R0b0t1> plntyk: I'm more confused about what I even patch
<R0b0t1> hold on it might be in arch/arm/mach-sunxi
<plntyk> you could search the git log of kernel or your board if it mentions "sunxi: add i2c" an look through that patch
<R0b0t1> ok, I am almost there
<R0b0t1> I just can't figure out which dtb gets distributed for my board
<R0b0t1> `find arch/arm/boot/ -name '*sun8i-h*dtb'` produces a lot of candidates
<R0b0t1> ah no I missed something about the build process, there is an exact dtb in my /boot
<R0b0t1> hm
<plntyk> ./patch/kernel/sunxi-current/board-h2plus-nanopi-duo-add-device.patch in armbian buildsystem adds that functionality because there is no mainline support yet it seem
<plntyk> or not for armbian-current variant
<R0b0t1> ah thanks appreciate it
<R0b0t1> ok this is probably the time I need to actually build armbian for real
<R0b0t1> alright I made a lot of progress
<R0b0t1> but I can't actually figure out where the clock prescaler or w/e is set
<R0b0t1> hmm
<Werner> Good morning
<IgorPec> good morning
valoga has quit [Ping timeout: 256 seconds]
<plntyk> R0b0t1: according to https://linux-sunxi.org/Linux_mainlining_effort I2C for H3 (which should be compatible to h2plus in nanopi du) was done in 4.9 so look for patches/changelog in that kernel to find where in the source this was done or sth should be in documentation of dtb / i2c subsystem
redentor has quit [Remote host closed the connection]
fizikz has quit [Quit: Leaving.]
DaRock has quit [Ping timeout: 272 seconds]
sbbg has quit [Quit: Konversation terminated!]
<ArmbianTwitter> @VeryMetal_dev (Henri ⌨️🖥️👨‍🔧🖋️✏️😷): @chriseppstein Just fixed the sound of my Orange Pi 3 with a generic external USB soundcard and listening radio via Bluetooth because of USB Bluetooth adapter. Gotta avoid getting caught up in config files any way possible. So yeah, the zone! Orange Pi 3 / Armbian configuration. (16s ago)
drobo_01 has joined #armbian
drobo_00 has quit [Ping timeout: 246 seconds]
drobo_01 is now known as drobo_00
macc24 has joined #armbian
fizikz has joined #armbian
JuniorJPDJ has quit [*.net *.split]
JuniorJPDJ has joined #armbian
JuniorJPDJ has quit [Changing host]
JuniorJPDJ has joined #armbian
gnom has quit [Ping timeout: 246 seconds]
mirage335 has quit [Ping timeout: 244 seconds]
<ArmbianTwitter> @orangepixunlong (OrangePi): RT @f5nlg: #spotnikV4Rc2 Orangepi0H2+ - H3 availlable to download : https://t.co/sUhdG0loLH #Analogique #RRF #FON #Echolink #C4FM #DMR #P… (18s ago)
andrevallestero[ has quit [Quit: killed]
psydruid has quit [Quit: killed]
t4h4[m] has quit [Quit: killed]
xlogik has quit [Quit: killed]
hyperreal has quit [Quit: killed]
solderfumes[m]1 has quit [Quit: killed]
phantomse[m] has quit [Quit: killed]
nekomancer[m] has quit [Quit: killed]
JuniorJPDJ has quit [Quit: killed]
t4h4[m] has joined #armbian
torv has quit [Ping timeout: 240 seconds]
torv has joined #armbian
JessiWilde has joined #armbian
JessiWilde has quit [Client Quit]
mirage335 has joined #armbian
phantomse[m] has joined #armbian
xlogik has joined #armbian
hyperreal has joined #armbian
solderfumes[m] has joined #armbian
psydruid has joined #armbian
andrevallestero[ has joined #armbian
JuniorJPDJ has joined #armbian
nekomancer[m] has joined #armbian
chewitt has quit [Read error: Connection reset by peer]
chewitt has joined #armbian
macc24 has quit [Ping timeout: 258 seconds]
macc24 has joined #armbian
macc24 has quit [Ping timeout: 256 seconds]
silven_ has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
silven has joined #armbian
macc24 has joined #armbian
Werner has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
Werner has joined #armbian
Werner has joined #armbian
Werner has quit [Changing host]
valoga has joined #armbian
Werner has quit [Read error: Connection reset by peer]
Werner has joined #armbian
Werner has quit [Changing host]
Werner has joined #armbian
Tenkawa has joined #armbian
macc24 has quit [Quit: WeeChat 2.8]
<R0b0t1> plntyk: I definitely found the relevant files there's just no mention of a clock speed, just a clock source, and no way to set a prescaler
<R0b0t1> will try a sunxi-specific channel if there is one
macc24 has joined #armbian
<ArmbianTwitter> @_arifinsantoso (YC2SAT): I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3bvg @Mubino1 and @amsat_id @lapansat https://t.co/OhPUb0VGXI (27s ago)
<ArmbianTwitter> @pujianto (pujianto): RT @_arifinsantoso: I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3b… (7s ago)
dddddd has joined #armbian
<ArmbianTwitter> @yc8nyj (JoeSant - YC8NYJ): RT @_arifinsantoso: I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3b… (12s ago)
<ArmbianTwitter> @armbian (armbian): RT @_arifinsantoso: I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3b… (21s ago)
macc24 has quit [Ping timeout: 246 seconds]
macc24 has joined #armbian
<ArmbianTwitter> @AdrianSuperNet (Adrián Ruiz Figueroa): RT @_arifinsantoso: I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3b… (28s ago)
<ArmbianTwitter> @_arifinsantoso (YC2SAT): Thankyu @armbian i use armbian 5.91 debian buster (27s ago)
<ArmbianTwitter> @Yomue_ (yomu): RT @_arifinsantoso: I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3b… (11s ago)
xec has quit [Remote host closed the connection]
xecuter has joined #armbian
<ArmbianTwitter> @DieZuckerbude (Ben Zucker 🍰): @_arifinsantoso @armbian A bit outdated though 😉 Release model has changed last year. You can find more info here: https://t.co/qNZiOwur15 You will find a changelog at the docs as well. (14s ago)
xec has joined #armbian
xecuter has quit [Ping timeout: 244 seconds]
xec has quit [Ping timeout: 260 seconds]
xecuter has joined #armbian
xecuter has quit [Ping timeout: 260 seconds]
<R0b0t1> where is the check for bionic?
<R0b0t1> I'd like to comment it out, I keep the pieces etc
<Werner> What do you mean?
<plntyk> created a pull request to fix some issue with odroidc1 images created by buildbot see https://github.com/armbian/build/pull/1949
<R0b0t1> ah NO_HOST_RELEASE_CHECK=yes ./compile.sh
<plntyk> hopefully the images on dl.armbian.com will work after that
<R0b0t1> ok... that's not working
<Werner> ./compile.sh NO_HOST_RELEASE_CHECK=yes should work
<R0b0t1> Werner: I'm on ubuntu cosmic (18.10) and the build aborts with invalid distro
<R0b0t1> ah
<R0b0t1> I mean, 18.10 vs 18.04 figuring it should work, if not I'll try docker
<R0b0t1> can't upgrade atm
<Werner> Just try but do not report any issues that may occur using this parameter ;)
<R0b0t1> yeah, wasn't gonna waste anyone's time with that
<R0b0t1> granted, I do use Gentoo most of the time, and find keeping everything super recent tends to reduce issues
<R0b0t1> it's just the cross toolchains are kinda broken (I think the project does arm builds on some servers for sbc stuff)
<R0b0t1> I don't have the time to fix the crosschain or wait for the builds, so...
<IgorPec> plntyk: i will remade images for c1 along with c4 which i just fixed some critical issues. tnx
<R0b0t1> ok build system is kinda nice
phantomse[m] has quit [Quit: Idle for 30+ days]
t4h4[m] has left #armbian ["Kicked by @appservice-irc:matrix.org : Idle for 30+ days"]
andrevallestero[ has quit [Quit: Idle for 30+ days]
<ArmbianTwitter> @ptwindu (Putu Windu Tenaya): RT @_arifinsantoso: I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3b… (5s ago)
<R0b0t1> if there were a way to avoid packaging the compilers and relying on upstream there would be no reason to tie to a distro release
<R0b0t1> am working through some errors, seems only real one is a libc++ issue, aside from some packages not being installed (but being otherwise available)
<R0b0t1> think I'll move to docker
Tenkawa has left #armbian [#armbian]
<IgorPec> i am compiling on linux mint and debian buster, natively
<IgorPec> but some proper inspection should be done to declare all as official.
<IgorPec> other distros are mainly problematic for their default packaging manager - more (un-needed) changes
<R0b0t1> yeah I think sticking to a single pm is fine
<R0b0t1> I like apt the best fwiw
sunshavi has quit [Remote host closed the connection]
sunshavi has joined #armbian
<R0b0t1> what's used to dl the torrents?
<Werner> aria2c AFAIK
<R0b0t1> nice
<R0b0t1> nice.jpg
<Werner> .great
<ArmbianHelper> great, adjective, synonym for Armbian
<R0b0t1> >download from two locations at once
<R0b0t1> lol
cardboard64 has left #armbian [#armbian]
<ArmbianTwitter> @Mubino1 (MubinYB3MBN): RT @_arifinsantoso: I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3b… (19s ago)
IgorPec has quit [Remote host closed the connection]
IgorPec has joined #armbian
IgorPec has joined #armbian
<R0b0t1> so I've created a patch
<R0b0t1> I drop it into patch/kernel/sunxi-current/board-nanopi-duo-i2c-speed.patch
<R0b0t1> but it fails to apply
<R0b0t1> any way to get better output? I'm using docker, so everything is in the docker image
<R0b0t1> I got it using git diff from a vanilla tree of a similar branch
<R0b0t1> I find a similar file in ./cache/sources/linux-mainline/orange-pi-5.4/arch/arm/boot/dts/sunxi-h3-h5.dtsi
<lanefu> R0b0t1: output/debug/*.log
<R0b0t1> lanefu: all I get is 1 out of 1 hunk FAILED -- saving rejects to file arch/arm/boot/dts/sunxi-h3-h5.dtsi.rej
drobo_00 has quit [Ping timeout: 260 seconds]
drobo_00 has joined #armbian
oida has quit [Quit: byez]
<R0b0t1> I tried removing the git hash
<R0b0t1> that is probably different
<R0b0t1> still fails
<ArmbianTwitter> @sam210723 (Sam): RT @_arifinsantoso: I was very surprised, my first picture from GK-2A was using Android TV Box using Armbian @sam210723 big thanks to @yc3b… (19s ago)
s_frit has quit []
torv has quit [Ping timeout: 240 seconds]
torv has joined #armbian
macc24 has quit [Ping timeout: 260 seconds]
oida has joined #armbian
<R0b0t1> yeah still at a loss
macc24 has joined #armbian
<R0b0t1> trying to bust into the docker container, not seeing it
<R0b0t1> ok I inspected the docker volume but am not having any better luck generating a patch that applies
<R0b0t1> the rej is just the contents of the failed patch file
<plntyk> just check out the kernel that armbian on that board uses and try to patch it
<plntyk> and then copy that patch file since it will work
<R0b0t1> the patch isn't against any source code just a dts, but yeah I'll try to grab the orangepi kernel armbian is using
<lanefu> where'd the patch come from
<lanefu> and what kernel is it made for
<lanefu> oh if i scroll up i can see yuo already said that
<R0b0t1> lanefu: I made the patch myself
<R0b0t1> it's made for any kernel, but specifically the one for my nanopi duo
<R0b0t1> lanefu: http://dpaste.com/0RZFW89
<R0b0t1> oooh actually this probably fails because there is more than i2c block, but shouldn't the line numbers fix that?
<lanefu> dunno about line number
<lanefu> was wondering if patch was applying too late and being impacted by others
<lanefu> but just sounds like you need to tweak teh patch
<plntyk> the patch might fail because the 3 lines before and after the "+" that add lines are not identical to those in the source probably
<plntyk> line numbers might not matter
<R0b0t1> no adding more context didn't help it apply
<R0b0t1> so yeah maybe there is another patch
<R0b0t1> I changed the name so it started with 0000, patch STILL fails
<R0b0t1> I'm kind of out of ideas here, I read a thing about failing to apply patches before but mfw
<R0b0t1> IgorPec: think you could help? I can't set clock-frequency for i2c0 in arch/arm/boot/dts/sunxi-h3-h5.dtsi
<R0b0t1> is there a way to maybe beat build process over head and skip patching, just insert by hand?
<nekomancer[m]> is it possible to stop script just before it should apply user patches — to make that user patches?
<plntyk> R0b0t1 which kernel version does your board/target use ?
<lanefu> given that teh changes are small, why dont you make a new patch by hand?
<plntyk> the patch wont work on vanilla 5.4.36 for example because the file is different there
<R0b0t1> lanefu: I am not sure how it would be different from output of `git diff` is my problem
<R0b0t1> ah ok
<R0b0t1> almost checkout orangepi
<plntyk> http://paste.debian.net/1145620/ for example in 5.4.36 should work
<plntyk> in mainline
<R0b0t1> that also does not apply, hmm
<plntyk> ah
<R0b0t1> I'm not sure I am picking out which version is selected in build, but I see linux-mainline/orange-pi-5.4
<plntyk> some other patch probably changes the file
<R0b0t1> it is allwinner H2+ based board
<R0b0t1> well, I put it right at start, name of file is now 0000-*
<plntyk> kernel/sunxi-current/patch-5.4.34-35.patch
<plntyk> in build dir already touches the dts that you change so better integrate the patch into that one / apply that patch first
<plntyk> dunno about the patch order in that directory
<R0b0t1> mine should go first, but I guess that one may be above the misc. patches
<R0b0t1> I add to the one you listed, seems to apply
<R0b0t1> nope
<R0b0t1> super confused the chunks are unrelated to each other, so it should apply even if the file is increased in size
<R0b0t1> and it's not
<R0b0t1> so, the patch that modifies that dts file goes after mine, and unrelated portion
<R0b0t1> hmm
torv has quit [Remote host closed the connection]
torv has joined #armbian
gnom has joined #armbian
<plntyk> i am now confused anyway why there are patch for linux sublevel version when there are the tags available in linux stable git tree
<lanefu> is this for allwinner?
<lanefu> if so i may have a half answer
<lanefu> :re sublevel patches
<R0b0t1> yeah I never figured this out either
<lanefu> yeah so with allwinner,and a few others..... armbian still isn't actaully using mainline.. it's using someone's fork..... so it's applying patches to megous's fork https://github.com/armbian/build/blob/master/config/sources/families/include/sunxi64_common.inc#L16
<lanefu> surprise!
<lanefu> :P
<plntyk> ah yes
<lanefu> megous's branch usually has a lot of the sunxi stuff that isnt accepted in mainline yet, so we start from there
<plntyk> having clones of mainline, stable, next, vendor branches and hacking/dev forks
<lanefu> sunxi is kind of an exception because hte project is so much.... like with amlogic i think we just patch dirtectly onto mainline
<lanefu> it's always a bit of a rats nest, but armbian gets more normalized with every release... but reality is a lot of these SBCs need extra care and feeding... which is why armbian exists
<plntyk> yeah with openwrt you have basically a stable checkout and then a huge number of numbered patches for generic and specific subarches/targets
<lanefu> yeah glancing around. it looks like sunxi is the main exception to us not using vanilla mainline
<plntyk> shitty github ... clicking 2 times and "You have triggered an abuse detection mechanism"
<lanefu> wtf really?
<nekomancer[m]> <plntyk "shitty github ... clicking 2 tim"> damn abuser!
<lanefu> nekomancer[m]: what IRC client do you use?
<lanefu> your text always has like 21st century enhancements
<plntyk> browsing linux clone on github is impossible anyway because the dts folder has too many files iirc
<R0b0t1> lanefu: matrix
<R0b0t1> he's a normie, can't see the desert of the real
<R0b0t1> lanefu: yeah I'm messing with armbian because it seems to have the best and most comprehensive support
<R0b0t1> it's just this ONE i2c issue
<R0b0t1> saldkjfalskjf
<plntyk> "Sorry we had to truncate this directory to 1,000 files"
<plntyk> almost 2000 dts files
<nekomancer[m]> <lanefu "nekomancer: what IRC client do y"> I use matrix client and matrix bridge. Why you ask?
<lanefu> cuz your stuff looks like this
<lanefu> <lanefu "nekomancer: what IRC client do y"> I use matrix client and matrix bridge. Why you ask?
<nekomancer[m]> Sorry :(((
<lanefu> lol
<lanefu> its alright
<lanefu> i mean i use slack all day long for other stuff
<lanefu> just have CLI purity with screen and weechat adn armbian irc
<lanefu> R0b0t1: so i thikn there's a seperate command line tool thats part of the patch suite, to fix line numbers..... i used it a few months ago, but i'd have to google all over again
<R0b0t1> can you think of maybe one or two crucial words
<lanefu> apparently not cuz my google is failing lol
<R0b0t1> my shitty keyboard depends on you, lanefu
<R0b0t1> Noooooooooooooo~
<plntyk> lanefu, dunno maybe you mean "quilt" ?
<lanefu> rediff
<lanefu> bash_history for the win
<R0b0t1> wow even if I editdiff rediff just fails
<R0b0t1> no error
<R0b0t1> ahh ok
<R0b0t1> sigh
<R0b0t1> editdiff works, but then it fails
<IgorPec> R0b0t1: did you try ./compile.sh CREATE_PATCHES="yes"
<R0b0t1> no, what are the preconditions for that
<R0b0t1> my linux source is somewhere else entirely
<R0b0t1> ahh ok
<R0b0t1> trying to find location in docker volume, that is a good feature
<IgorPec> docker is not most appropriate tool for development
<IgorPec> its more for building an image with small changes
<R0b0t1> unfortunately is only way I can run it atm
<IgorPec> this is worth studing if you plan to dig deeper
macc24 has quit [Ping timeout: 256 seconds]
<R0b0t1> finally
<R0b0t1> thanks IgorPec
drobo_00 has quit [Remote host closed the connection]
<ArmbianTwitter> @murdocklawless (Murdock Lawless): @nl_observant @The_Pi_Hole @armbian @lemakerbananapi I'm glad this helped. I still have banana pi (not pro) working. (3s ago)
<R0b0t1> niceee
<R0b0t1> yeah i2c stuff I'm doing is now 10x as fast
<R0b0t1> great
<IgorPec> now you can contribute :)
<IgorPec> somebody will be happy
<R0b0t1> well, 1mhz base speed is probably not a sane default, and considering it is in device tree...
<R0b0t1> the gadgetfs config maybe
<IgorPec> i don't know how those defaults are set, huh. but you can prepare a MR and lets see what people will say about
<nekomancer[m]> R0b0t1: how do you find final files to change?
<R0b0t1> I found a dts for my board in the armbian repos
<R0b0t1> and tracked through the includes until I found the i2c device definition
<R0b0t1> someone (maybe not here?) found a patch that indicated I could add clock-frequency, it was missing, default is 100kHz
<R0b0t1> I wasn't sure if I could just add it b/c sometimes you have to directly write a register or something
<R0b0t1> no no that was definitely here
<nekomancer[m]> but how about final version of kernel souces used by general build script?
<nekomancer[m]> how do you see what of many possible kernel sources it uses in your case?
R0b0t1` has joined #armbian
ed_peguillan has quit [Ping timeout: 256 seconds]