ArmbianHelper changed the topic of #armbian to: armbian - Linux for ARM development boards | www.armbian.com | Github: github.com/armbian | Commits: #armbian-commits | Developer talk: #armbian-devel | Forum feed: #armbian-rss | Off-Topic: #armbian-offtopic | Logs: -> irc.armbian.com
aasami_ has quit [Quit: leaving]
MrFixIt has quit [Read error: Connection reset by peer]
archetech has quit [Quit: Leaving]
MrFixIt has joined #armbian
MrFixIt has quit [Max SendQ exceeded]
archetech has joined #armbian
MrFixIt has joined #armbian
norwich has quit [Ping timeout: 255 seconds]
norwich has joined #armbian
Suspect has quit [Ping timeout: 252 seconds]
<nekomancer[m]>
any pastebin with ANSI colors support?
oida has joined #armbian
p0g0 has joined #armbian
grriz has joined #armbian
jclsn has quit [Ping timeout: 248 seconds]
jclsn has joined #armbian
<Armbian-Discord>
<rpardini> lol
<Armbian-Discord>
<rpardini> I wrote one
<Armbian-Discord>
<rpardini> or, cobbled together
<Armbian-Discord>
<rpardini> was gonna call it armbin then armbiansi then gave up
stipa_ has joined #armbian
stipa has quit [Read error: Connection reset by peer]
stipa_ is now known as stipa
grriz has quit [Quit: Leaving]
archetech has quit [Quit: Leaving]
Dagger has quit [Ping timeout: 252 seconds]
Dagger has joined #armbian
<Kreyren>
ANSBINAN
* Kreyren
grins
<Kreyren>
hmm i figured out what the issue is on the encryption setup, but i dunno how to fix it
<Kreyren>
basically the framework is doing `# losetup /dev/loop0 path/to/built.img` like it mounts it, but it doesn't show the partitions like /dev/loop0p2, but it shows them in the host
ikmaak has quit [Read error: Connection reset by peer]
arch3r has quit [Ping timeout: 246 seconds]
arch3r has joined #armbian
ikmaak has joined #armbian
<CosmicDJ>
kreyren: try losetup -P
<Kreyren>
CosmicDJ, did, but it didn't help
<CosmicDJ>
kreyren: "Note that the partition table parsing depends on sector sizes. The default is sector size is 512 bytes" different sector size?
<Kreyren>
like losetup from docker mounts the /dev/loop0 including the partitions /dev/loop0pX on host, but only the /dev/loop0 is available on docker
<CosmicDJ>
ah sounds like a Docker problem then
<Kreyren>
yep
<Kreyren>
it has /dev set a device though
<Kreyren>
so like it should work
archetech has joined #armbian
ikmaak has quit [Ping timeout: 248 seconds]
ikmaak has joined #armbian
arch3r has quit [Quit: bye]
arch3r has joined #armbian
danilogondolfo has joined #armbian
flyback has quit [Ping timeout: 252 seconds]
flyback has joined #armbian
Suspect has joined #armbian
Suspect has quit [Quit: Leaving]
kolla has quit [Quit: %fog relay%]
kolla has joined #armbian
qqqhhh8619935 has joined #armbian
qqqhhh861993 has quit [Ping timeout: 252 seconds]
qqqhhh8619935 is now known as qqqhhh861993
arch3r has quit [Quit: bye]
arch3r has joined #armbian
arch3r has quit [Ping timeout: 256 seconds]
arch3r has joined #armbian
\` has quit [Ping timeout: 265 seconds]
LanDi has joined #armbian
LanDi has quit [Remote host closed the connection]
\` has joined #armbian
zeemate_ has joined #armbian
<Armbian-Discord>
<Tenkawa> You will want to use these in losetup btw too: -o, --offset <num> start at offset <num> into file
<Armbian-Discord>
<Tenkawa> -b, --sector-size <num> set the logical sector size to <num>
<Armbian-Discord>
<Tenkawa> If you need to get into offsets and the sector size
<Armbian-Discord>
<rpardini> Maybe just call partprobe inside Docker after losetup. Or reuse some of the helpers
Mangix has quit [Read error: Connection reset by peer]
Mangix has joined #armbian
lyri has joined #armbian
<Armbian-Discord>
<Tenkawa> @rpardini if they are having bind problems or offset problems with /dev a partprobe is just going to call a reload of that
<Armbian-Discord>
<Tenkawa> partprobe isn't that smart of a tool
<Armbian-Discord>
<Tenkawa> it just reloads the table
<Armbian-Discord>
<Tenkawa> docker stopping and starting the container itself should be doing that
alekksander has joined #armbian
<Armbian-Discord>
<rpardini> Nah, usually when in Docker and "device shows but partitions don't" it's because of a disconnect between udev (or non-udev'ness) inside Docker vs the host kernel, some event is missed. So the p1's etc show up on the host, but not inside Docker.
<Armbian-Discord>
<rpardini> when you partprobe it manually it "syncs" and the partitions show up.
<Armbian-Discord>
<Tenkawa> that's still not going to be inheritantly "fixed" by partprobe... thats a bandaid
<Armbian-Discord>
<rpardini> yep, /dev/loopX inside Docker is a huge patchwork of bandaids.
<Armbian-Discord>
<Tenkawa> works fine for me
<Armbian-Discord>
<rpardini> you're lucky. I've hit snags every other way. maybe you've some magical Docker option that makes it work?
<Armbian-Discord>
<Tenkawa> No.. you have to keep things "simple" in docker... don't get so overcomplicated
<Armbian-Discord>
<rpardini> Hmmm please instruct me
<Armbian-Discord>
<Tenkawa> Part of the armbian builds for example is they try to do too much at once
<Armbian-Discord>
<rpardini> hmmm but doing just that one /dev/loop stuff fails by itself
<Armbian-Discord>
<rpardini> so focus on that one issue, how do I fix it?
<Armbian-Discord>
<Tenkawa> If it was committing changes then syncing more it would be easier to manipulating disk changes
<Armbian-Discord>
<Tenkawa> do you chroot in or work from inside it?
<Armbian-Discord>
<Tenkawa> (I can't remember)
<Armbian-Discord>
<rpardini> specifically, for the case where docker (client) and dockerd (host) are not in the same host, eg, when using Docker Desktop on Mac
<Armbian-Discord>
<rpardini> chroot is irrelevant, /dev/loop with Docker in a VM happens with or without chroot
<Armbian-Discord>
<Tenkawa> That is odd...
<Armbian-Discord>
<rpardini> guess you've no idea, then.
<Armbian-Discord>
<Tenkawa> I'm on an M1 as well
<Armbian-Discord>
<Tenkawa> I'll have to see what our build is doing differently
<Armbian-Discord>
<rpardini> how do I test your build on a Mac M1?
<Armbian-Discord>
<rpardini> cos it's maybe easier for me to peek at yours (since it's better/simpler) than you to look at mine
<Armbian-Discord>
<Tenkawa> I've got to write up Docker instructs (mine has been setup for a while and I still prefer to use my full VM because its faster)
<Armbian-Discord>
<Tenkawa> Will give me a good chance to (Just realized I had to remove my last one because I was running out of disk space for a RISC-V test)
<Armbian-Discord>
<Tenkawa> That risc-v build is nuts
<Armbian-Discord>
<rpardini> yeah well. Armbian's decade-old secrets for remote-Docker lie here, if you wanna peek:
<Armbian-Discord>
<Tenkawa> Sifive internal build wants insane amounts of space
<Armbian-Discord>
<rpardini> any insight on how to make it "just work" with any "oddness" is welcome -- even if it's a full rewrite, I don't mind, if we find the fixing principle
<Armbian-Discord>
<Tenkawa> Ahhh
<Armbian-Discord>
<Tenkawa> I see one big diff
<Armbian-Discord>
<Tenkawa> You are using Ubuntu core
<Armbian-Discord>
<Tenkawa> I'm using Debian core
<Armbian-Discord>
<Tenkawa> makes a huge difference
<Armbian-Discord>
<rpardini> nah, that doesn't solve it, I can switch with 1 parameter, all fail the same way. (some very old Debian fail more even)
<Armbian-Discord>
<Tenkawa> no
<Armbian-Discord>
<Tenkawa> the tweaks you add for apparmor/etc are part of what I take out
<Armbian-Discord>
<Tenkawa> and cgroup
<Armbian-Discord>
<Tenkawa> I don't use any of that'
<Armbian-Discord>
<rpardini> that doesn't fix or break the loop issue
<Armbian-Discord>
<Tenkawa> It does for me
<Armbian-Discord>
<Tenkawa> And I also remove a ton of debug mess from the kernel... there's no point in having it
<Armbian-Discord>
<Tenkawa> (not on the build side)
<Armbian-Discord>
<rpardini> well if you have something concrete, any evidence of a any kind of build running on M1 with Docker sans bandaids, I'd love to know.
\` has quit [Ping timeout: 268 seconds]
<Armbian-Discord>
<Tenkawa> I will rebuild it and let you know.. is armbian-next good to pull to test with?
<Armbian-Discord>
<Tenkawa> or do you want me to use regular armbian?
<Armbian-Discord>
<rpardini> -next, for sure
<Armbian-Discord>
<rpardini> regular/master Armbian's funeral is.... soon, hopefully -- few weeks at max
<Armbian-Discord>
<Tenkawa> ok.. I "should" have some time either this weekend or Monday
<Armbian-Discord>
<Tenkawa> Just got back to normal today...
<Armbian-Discord>
<Tenkawa> (whatever normal is)
<Armbian-Discord>
<rpardini> no problem, any help is welcome. I've struggled with this for months and would love to get rid of those hacks
<Armbian-Discord>
<rpardini> maybe/probably there's something very simple I've overlooked
<Armbian-Discord>
<Tenkawa> In other news I'm disappointed but not suprised by what chewitt/narmstrong mentioned about the vim4... think I might hold off on getting one now
<Armbian-Discord>
<rpardini> they said that from the start when it was released. that's why I never got one...
<Armbian-Discord>
<Tenkawa> I do like the Edge2 though
<Armbian-Discord>
<Tenkawa> It has been actualy a decent surprise
<Armbian-Discord>
<Tenkawa> Either way other things are broke...
<Armbian-Discord>
<c0rnelius> i can only assume this got introduced into the 5.10 as well, would seem pretty random that you get 256.
<Armbian-Discord>
<Tenkawa> And Khadas doesn't support it
<Armbian-Discord>
<Tenkawa> Every SoC I have gets 256 except this one
<Armbian-Discord>
<Tenkawa> And my x86 has a non x86 number
<Armbian-Discord>
<Tenkawa> hmm interesting...
<Armbian-Discord>
<c0rnelius> thats what I mean. you said its getting stuck at 256. thats not a bug but a feature. now of course there could be something a miss on the source they use, but the overall implementation is intentional.
<Armbian-Discord>
<Tenkawa> I do agree its a good feature... I just would like to see more consistency (but we know thats not going to happen)
<Armbian-Discord>
<Tenkawa> If I had known that tree worked I could've upgraded a long time ago... I ran that test successfully a long time ago
<Armbian-Discord>
<Tenkawa> err... that behaviour
<Armbian-Discord>
<Tenkawa> problem is 110 is still ancient even on 5.10
<Armbian-Discord>
<c0rnelius> yeah its not really the height of amazing achievement is it? 😉
<Armbian-Discord>
<Tenkawa> unfortunately no
<Armbian-Discord>
<Tenkawa> But 6 is a PITA
<Armbian-Discord>
<Tenkawa> lol
<Armbian-Discord>
<rpardini> at least -- amazingfate is gonna unify all the things (66, 68, 88, all vendors) under a single rk 5.10 kernel for Armbian
<Armbian-Discord>
<rpardini> fact is, many ppl have tried to bring back boards from media, but they all disappear over time, while oleg is always there, fixing, and supporting users. and it works -- so I'm all for it.
<Armbian-Discord>
<Tenkawa> Problem is... the "support" word is relative... I've had/seen him asked to document his changes and procedures for people and not seen it done once... and I have no tolerance for that.
<Armbian-Discord>
<Tenkawa> If "insert person here" is going to take it upon themselves to be the center of contact/support for something... be ready to take responsibility and contact when the everyday person has questions if you don't do that documentation.
<Armbian-Discord>
<IgorPec> i had a long discussion with Oleg last week and I hope he will join at least with rk3588. His compromise goes into direction of maintaining our own fork of Rockchip kernel with integrating patches for radxa, opi, khadas ... is this doable and acceptable, I don't know?
<Armbian-Discord>
<IgorPec> his argument against was - he doesn't trust to base on radxa kernel source
arch3r has quit [Ping timeout: 256 seconds]
<Armbian-Discord>
<rpardini> it's doable -- @EfeCTN (and me, clapping) has proven it's possible
<Armbian-Discord>
<rpardini> but I think rk has to be the base, with Radxa / Xunlong / Khadas patches on top
<Armbian-Discord>
<IgorPec> yes, that way
arch3r has joined #armbian
<Armbian-Discord>
<rpardini> Amazingfate has many more, for the 66/68's too on same kernel
<Armbian-Discord>
<rpardini> so has to be concerted effort I think
<Armbian-Discord>
<rpardini> also: having those squashed/single commit trees Oleg normally does -- a no no
<Armbian-Discord>
<rpardini> we need clean patchset in files against a public well known base
<Armbian-Discord>
<IgorPec> yes.
<Armbian-Discord>
<IgorPec> his argument was that what we do now is not good enough 😉
<Armbian-Discord>
<IgorPec> could be lost in translation, so don't catch on this
<Armbian-Discord>
<rpardini> well it... works. welcome to do better, of course, but squashing in a hidden tree somewhere and hijacking any external contributions... is not "better"
<Armbian-Discord>
<IgorPec> i know
<Armbian-Discord>
<rpardini> so all for doing it better, as long as it is actually better
<Armbian-Discord>
<IgorPec> lets see if @amazingfate and @EfeCTN are also ok with this approach, then we can start thinking about
<Armbian-Discord>
<rpardini> I'd say amazingfate knows more and better, also introduced the 88 legacy originally, so I'd propose he's the lead on this
<Armbian-Discord>
<rpardini> (not to discount @EfeCTN heroic efforts, of course, and my clapping)
<Armbian-Discord>
<IgorPec> agree. Oleg is only adding dt for station m3 and some minor fixes
<Armbian-Discord>
<rpardini> yep that should be very ok. I think they all will coexist no problem... kedge2 might be more effort due to MCU
<Armbian-Discord>
<rpardini> gotta go, and will miss the Lounge tonite, have actual social event 🫡
<Armbian-Discord>
<EfeCTN> I think this is just what we're doing by patches. Seems good idea but not big improvement
<Armbian-Discord>
<IgorPec> great! enjoy it. i am still in the wilderness
<Armbian-Discord>
<IgorPec> improvement is that we somehow convince Oleg its better to work together 😉
<Armbian-Discord>
<EfeCTN> Nice deal
<Armbian-Discord>
<rpardini> Single legacy Rockchip 5.10 kernel for all rk3566/68/88(s) boards
<Armbian-Discord>
<rpardini> oh damn meant that for armbian-rockchip... sorry
<Kreyren>
Btw. I looked through the whole hackerspace and didn't find RK3588S so i will buy it @_@
<Armbian-Discord>
<rpardini> recommend a Mekotronics -- @monkaBlyat has DTs for those
<Armbian-Discord>
<amazingfate> I think tag linux-5.10-gen-rkr3.6 in my PR should be a good code base because it is an official release from the sdk of rockchip, and clean. We can update by tags when rockchip releases newer
<Armbian-Discord>
<EfeCTN> What differences 3.6 has over 3.4. I wonder why radxa doesn't rebase
<Armbian-Discord>
<amazingfate> I think radxa would update code to their rkr3.4 branch in the future. rkr3.6 fixes some bugs, for example rk3568 with pcie can boot with it.
<Armbian-Discord>
<amazingfate> This kernel is upoaded by radxa. It has been kept up to date in the past several mouths. And I think we can remind radxa when the repo need sync from rockchip.
arch3r has quit [Ping timeout: 246 seconds]
arch3r has joined #armbian
arch3r has quit [Excess Flood]
arch3r has joined #armbian
arch3r has quit [Ping timeout: 248 seconds]
arch3r- has joined #armbian
arch3r- is now known as arch3r
<Armbian-Discord>
<rpardini> another possibility, which I think @balbes150 has been exploring, is using edk2 for these boards.
<Armbian-Discord>
<rpardini> I've had big success with the quartz64a with it -- it's full UEFI/ACPI/DSDT though, no DeviceTree at all, and some drivers (eg GMAC) needs glue code, since they were not designed for ACPI, instead DT.
<Armbian-Discord>
<rpardini> Alternatively edk2 -> grub can boot linux passing the DT and acpi=off -- that allows ed2k-in-SPI -> NVMe with grub -> Linux with DT, all working 100%
<Armbian-Discord>
<rpardini> but yeah, building edk2 is... a real pain. all that C++ 😄
<Armbian-Discord>
<amazingfate> Does the edk2 have video output?
<Armbian-Discord>
<rpardini> Yep -- at least using jarredmcneil's edk2 on the quartz64a it does. What it doesn't have is any nvram, so any settings changed in the "BIOS" are lost with reboot.
<Armbian-Discord>
<Tenkawa> This repo is behind "another" active one Radxa one is already actively using though.. so I'm confused though how does that help?
LanDi has joined #armbian
<Armbian-Discord>
<rpardini> It also doesn't see the USB3, but USB2 works (for booting off of)
<Armbian-Discord>
<Tenkawa> This repo has got active rk3588 changes as of yesterday:
<Armbian-Discord>
<rpardini> Tenkawa: that is our current base. We want off out of it.
<Armbian-Discord>
<rpardini> we wanna be on rk base.
<Armbian-Discord>
<amazingfate> You can treat it as the upstream of legacy kernel from rockchip. All vendors are developing based on this.
<Armbian-Discord>
<Tenkawa> Ok.. your pai
<Armbian-Discord>
<amazingfate> And the radxa one is using a tag name as their developing branch's name
<Armbian-Discord>
<rpardini> the rebase is not too painful -- some hundred patches
<Armbian-Discord>
<rpardini> opi5 overlays very nicely on top of it
<Armbian-Discord>
<Tenkawa> Once its done I agree.. in its current state its horrible
<Armbian-Discord>
<rpardini> kedge2 is significantly more pain, since its base is that old .66
<Armbian-Discord>
<amazingfate> Most boards just need a dts patch
<Armbian-Discord>
<rpardini> yep.... Opi5 is mostly DTS stuff... some motorcomm for phy, etc
<Armbian-Discord>
<rpardini> main "work" is moving hacks the vendors do to .dtsi (which would impact the other boards) out of there and into the board's DT or some intermediaty dtsi
LanDi has quit [Quit: LanDi]
junaid_ has joined #armbian
Suspect has joined #armbian
danilogondolfo has quit [Remote host closed the connection]
<lanefu>
Linux used to have the alan cox kernel so why not have the Oleg kernel 😃
<lanefu>
Are the ar.bia
<lanefu>
Are the patches Armbian is providing on top of the vendor legacy kernel for rk3588 adding real value over the kernel trees maintained by Radxa and xunlong?
<lanefu>
Is it motivated by managing device tree overlays?
<archetech>
theres no rock5 maintainer yet who knows
norwich has quit [Ping timeout: 264 seconds]
alekksander has quit [Ping timeout: 248 seconds]
<yang2_>
I have the rock5
alekksander has joined #armbian
junaid_ has quit [Ping timeout: 256 seconds]
alekksander has quit [Quit: Konversation terminated!]
hectroid has quit [Read error: Connection reset by peer]
hectroid has joined #armbian
archetech has quit [Quit: Konversation terminated!]
archetech has joined #armbian
lyri has quit [Ping timeout: 268 seconds]
lyri has joined #armbian
junaid_ has joined #armbian
junaid_ has quit [Ping timeout: 255 seconds]
Suspect has quit [Remote host closed the connection]