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/Twitter feed: #armbian-rss | Type 'help' for help | Logs: -> irc.armbian.com
<Armbian-Discord> <T​onymac32> Overclock what?
psydroid has joined #armbian
solderfumes has joined #armbian
nekomancer[m] has joined #armbian
Michi[m] has joined #armbian
mrjpaxton[m] has joined #armbian
minnix[m] has joined #armbian
<Armbian-Discord> <T​onymac32> You aren't overclocking any AML device
<c0rnelius> Well that's not true. Sure amlogic lies but we can still pretend.
<c0rnelius> Plus that's mostly the s905* series and s912. You can see a diff in the s922x.
<c0rnelius> I don't own the s905y* so I have no clue what it can actually do but you can overclock it past what the dts suggests.
<c0rnelius> And please don't link me any cnx links. I've read them.
archetech has joined #armbian
<c0rnelius> Plus Armbian provides patches to push the S922X and S905X3 past what the DTS suggestion is. Are you saying thats pointless and the patches are BS?
TRS-80 has joined #armbian
<Armbian-Discord> <R​adxaYuntian> minnix[m]: I'm the maintainer from Radxa. Currently trying to figure out why armbian autotests won't find the build. Once that passes I think we can start producing nightlies
<Armbian-Discord> <R​adxaYuntian> If you just want to run Armbian now you can build off trunk. It was ported a few months ago.
<c0rnelius> Are you the working in the Radxa github on the 5.10.y branch for that board?
<c0rnelius> the one*
<c0rnelius> R​adxaYuntian: the above was for you.
<c0rnelius> Because I noticed the removal of this, which on its face wouldn't be a fix for much considering its still referenced else where in the dts: https://github.com/radxa/kernel/commit/fd159ba07d5cc71e6d9c11d9daf23a2a7d7543e9
<c0rnelius> And when checked by someone else `i don't have the board yet` using 5.15.y the error happens on dai-link-2
<Armbian-Discord> <R​adxaYuntian> I'm not working on kernel right now. lemme see if I can get stephen here.
<c0rnelius> chewitt, suggested it may be a simple scripting issue. `Force things to looks here and make adjustments upon boot`, but I did find that commit to be a bit odd.
elroo has joined #armbian
<Armbian-Discord> <R​adxaYuntian> c0rnelius: I have already forwarded all your concerns. However, he is currently working on another commit that prevents zero from rebooting.
<Armbian-Discord> <R​adxaYuntian> i'll keep a watch on this
<c0rnelius> Thats should allow the board to reboot
<c0rnelius> As for shutdown properly I don't know
<c0rnelius> steev, from Kali has already tested this for me.
<c0rnelius> But its been a problem for some time now.
<c0rnelius> tobetter and the hardkernel peps took it a whole step further and added an actual whole bit to the kernel for proper shutdown, but as I"m still waiting I haven't bothered trying to port it over and seeing if it works.
<c0rnelius> But regardless that revert is still apart of it.
TheCoffeMaker has quit [Ping timeout: 240 seconds]
<Armbian-Discord> <R​adxaYuntian> thanks. i have forwarded your message
<Armbian-Discord> <R​adxaYuntian> can you point me where they did the shutdown bit?
<c0rnelius> I believe I simple port won't do, but its a a good starting point.
archetech has quit [Quit: Konversation terminated!]
<c0rnelius> There doesn't seem to be a huge diff between G12A and B. Which is why I think the port should be doable.
<c0rnelius> For that matter it plays well on SM1... so?
<chewitt> It will stop the stuff spamming the log
<chewitt> *bu*
<chewitt> *but*
<chewitt> it will also prevent audio from workiinig
<chewitt> the solution is simply to set the mixer controls from userspace; this completes the routing and stops the spamming
<c0rnelius> chewitt: I know :) Its why its not apart of my patch set. It makes no sense.
<c0rnelius> chewitt: When i get a chance I was gonna try ur scripting out.
<chewitt> and C2/C4/HC4/N2/N2+ etc. reboot fine - when you dump all the hacks from HK repos and simply use upstream u-boot and current kernel
<c0rnelius> Plus on 5.10 vs 5.15 is errors diff.
<chewitt> or you have to infect your kernel with all kind of downstream crap
<c0rnelius> chewitt: What current kernel?
<chewitt> Amlogic stuff should track latest kernel .. it is the best option
<chewitt> 5.15 has more things present/fixed than 5.10
<c0rnelius> Doesn't Libre/Core use mostly vendor kernels?
<chewitt> Core = Only
<chewitt> Libre = None
TheCoffeMaker has joined #armbian
<chewitt> LE master is currently 5.15 across the board
<c0rnelius> In my own testing I've been using the tobetter patch set and most recent uboot and everything has been fine.
<c0rnelius> On odroids N2s/C4
<chewitt> It shouldn't be needed, and if there is an issue, it should be reported to maintainers and fixed
<chewitt> else you will be stuck with a crapload of downstream patches forever
<c0rnelius> chewitt: Yeah but isn't there huge diff between a base Debian install VS Libre?
<chewitt> Debian has wider use-cases to handle for sure, but the same statement applies
<chewitt> If there's a problem - report it - get it fixed
<c0rnelius> You guys throw in all kinds of tricks is what I'm getting at. Its very particular.
<chewitt> there's very little in my branch that's a true hack (and they are clearly labelled) .. it's mostly backports and stuff submitted upstream
<chewitt> plus anything I am currently working on
<chewitt> which is mostly device-tree tweaks since my ability to write actual code is rather low :)
<c0rnelius> Well I can only speak for what I've tested. For the n2 and c4 I ran both libre/core and then tried to work out why mainline kernels got fails and hit panics which turned out to be a uboot issue. Which turned out to be a nomap issue -> https://raw.githubusercontent.com/pyavitz/debian-image-builder/feature/patches/amlogic/009-prevent-random-reboots_allow-it-to-be-marked-nomap.patch
<c0rnelius> So I suspected you guys are running vendor uboot?
<chewitt> LE is 50/50 since we support TV boxes
<chewitt> Anything we produce a proper image for is 100% mainline u-boot
<chewitt> There is also a 'box' image which should work with vendor u-boot
<c0rnelius> How did you get around the nomap prob running most recent uboot?
<c0rnelius> Same junk I did :)
<chewitt> there was some mailing list convo on this, but it appears to have fizzled out without being fixed yet
<c0rnelius> Yeah I read a lot
<c0rnelius> Never was resolved
<c0rnelius> Seemed like peps just thru their hands up and said, fuck it. Subject was dropped and that was the fix.
<c0rnelius> Seems the common is to just do endless reverts.
TheCoffeMaker has quit [Ping timeout: 252 seconds]
<c0rnelius> For example the current stable uboot for the NanoPC-T4 doesn't work.
<c0rnelius> Need to use v2021.07
<c0rnelius> I guess they made changes for the Pinebook ( who cares? ) and decided to nothing else matters.
TheCoffeMaker has joined #armbian
<stipa> hookers
<Armbian-Discord> <T​onymac32> c0rnelius I wouldn't assume any such patches cause any real change. I don't know who added them either, so shrugs
<c0rnelius> T​onymac32: Ur entitled to ur opinion but I can say in my experience there is a diff on the S922X. As for the others I really don't know, but that could be said for a lot of SoCs? I can overclock the H5/A64, but that doesn't mean anything has much changed.
<c0rnelius> I used ur overlay to create a patch to overclock the RK3399. I can for sure say it runs hotter :)
kprasadvnsi has quit [Quit: Client limit exceeded: 20000]
<chewitt> G12A/B and SM1 are over-clockable
gordie[m] has quit [Quit: Client limit exceeded: 20000]
kprasadvnsi has joined #armbian
farooqkz[m] has quit [Quit: Client limit exceeded: 20000]
gordie[m] has joined #armbian
<chewitt> how much you can get away with depends on the chip revision and board/cooling .. of course
farooqkz[m] has joined #armbian
<c0rnelius> At least that way it defaults
<chewitt> that's a bad patch .. to explain
<c0rnelius> explain?
<chewitt> it should be moved to the zero board dts, not put in the SoC dtsi where it impacts all boards
<chewitt> never assume that all devices will overclock the same
<chewitt> because, err, that isn't the case
<c0rnelius> My builder is only applying to that board
<chewitt> it's still a good habit to put things in the right place
<c0rnelius> I would never willy just drop it on everything
<chewitt> I have learned that people raid my repo for patches, so I try to be careful with whats in public view
<c0rnelius> I'm just saying. The scripting only applies whats suggested
<chewitt> sure
<c0rnelius> plus even if the patch was applied across board it wouldn't matter. You would actually need to echo into the boost to make it do anything.
<c0rnelius> its null
<Armbian-Discord> <T​onymac32> unfortunately the only benchmarks here are between the xtra crusty Amlogic kernel and 5.14
<Armbian-Discord> <T​onymac32> would be nice to see this overclock vs no overclock on the same system
<Armbian-Discord> <T​onymac32> ah wait, that's n2 and N2+ 😛
<Armbian-Discord> <T​onymac32> so even worse
<Armbian-Discord> <T​onymac32> I'm not benchmarking mine because I'm mad at it's basic design and it gets to be my homeassistant until it quits working 😄
<c0rnelius> the old tkaiser standard.
<c0rnelius> well there is ur overlay applied in a actual patch in turbo-mode on a T4: Cores: 1512000 1512000 1512000 1512000 2016000 2016000
<c0rnelius> we can move forward and not use the same approach on a N2+: Cores: 1500000 1500000 2400000 2400000 2400000 2400000
<c0rnelius> This is not using vendor garbage
<c0rnelius> sprry that was on schedutils, here it performance: Cores: 2016000 2016000 2400000 2400000 2400000 2400000
<c0rnelius> We can also attempt to overclock some stupid a64/H5 boards :) doesn't seem to make a huge diff though.
<c0rnelius> I've always found it fascinating that within the Linux community any distortion created, creates a disagreement. But yet, most peoples work are founded on those who tried before us. A disagreement ensues but to only rap its self the original work of said author.
<c0rnelius> It is a silly world we live in.
<c0rnelius> I believe I once read `chewitt is the master of copy and paste` and I think that's pretty accurate amongst us all.
<chewitt> I'm getting quite good at it :)
<c0rnelius> and I am very grateful.
<c0rnelius> By the way, I liked ur old picture on ur github. Just saying.
<chewitt> I started to get customers and colleagues looking the profile so had to smarten it up a bit
TRS-80 has quit [Quit: WeeChat 3.0]
_whitelogger has joined #armbian
maknho has quit [Ping timeout: 240 seconds]
maknho has joined #armbian
stipa has quit [Quit: WeeChat 3.3]
BeachyHead has quit [Read error: Connection reset by peer]
BeachyHead has joined #armbian
stipa has joined #armbian
TRS-80 has joined #armbian
<Armbian-Discord> <r​pardini> please read through it carefully
<Armbian-Discord> <r​pardini> the background is at https://lore.kernel.org/linux-arm-kernel/7h4kfjpe07.fsf@baylibre.com/T/
<Armbian-Discord> <r​pardini> that said, that has only been fixed in current (5.10), edge (?) is still in a terrible state. I gotta find time
<c0rnelius> Thanx, r​pardini. I've read the Armbian commit before but never seen the discussion on lore. The commit is focused on the N2/+ correct? Which seems to not be working on the C4 -> https://forum.armbian.com/topic/19340-odroid-c4-will-not-reboot-after-any-sort-of-kernel-update-have-tried-running-nand-sata-install/
<c0rnelius> Not sure about the HC4 as I don't have it.
<c0rnelius> Interesting reading, thanx for the lore link.
<[TheBug]> If you have something that needs tested on HC4 let me know and I can pull mine our and test
<[TheBug]> out*
Guest5703 has left #armbian [#armbian]
<lanefu> If yes maybe toss em in the old armbian.lane-fu.com/linx/ pastbin
<[TheBug]> hmm
<[TheBug]> you know I haven't ever opened the case for it
<[TheBug]> but I am guessing TTL console is required to determine the issue?
<[TheBug]> maybe I can do that after work here
<[TheBug]> lanefu: ohh didn't even notice you were talking about Rock 3A
<c0rnelius> Its odd but with Bob Dole's passing and him being in the news now, all I can think about is those Family Guy episodes. Where he just repeats his name over and over again.
<[TheBug]> I thought you were referring to HC4
<[TheBug]> lanefu: is that the same one you linked me before? I think i may have already written a card just not tested it
<Armbian-Discord> <l​anefu> Freshly baked
<Armbian-Discord> <l​anefu> So probably the same but different
ziofork has quit [Quit: The Lounge - https://thelounge.chat]
ziofork has joined #armbian
sunshavi has quit [Remote host closed the connection]
<TRS-80> `/me gets his news from c0rnelius
<TRS-80> bah
* lanefu laughs at /me fail
sunshavi has joined #armbian
<TRS-80> `/me skulks away in shame
<c0rnelius> :)
<c0rnelius> Or when Peter meets Richard Keith Armey `Dick Armey` and he just starts laughing `like are you serious`? Whats your wife's name, Vagina Coast Guard? So many good Family Guy eps.
<TRS-80> I always was like wtf is this guy for real
<TRS-80> also John Boner, who they go out of the way to pronounce Baehner on the legacy corporate media
<Armbian-Discord> <r​pardini> c0rnelius: that reboot hang affected my HC4's as well. It'd suggest people try a build from trunk containing that commit. I don't get any more hangs, on the opposite, I can reboot freely, and I've an N2+ that has 30+ days uptime with this (never ever got beyond a few days before it). The HC4 too. The C4 I don't have to test, but I'd say the same...
<c0rnelius> TRS-80: Yeah that ones great too :)
<c0rnelius> I haven't tested ur commit yet so I can't say for sure. Other then that nomap patch which I've been using for a long time now, I just use the odroid patch set from tobetter. I could give a whirl on the C4 sometime to see if it works.
<c0rnelius> I also stopped using 5.10.y so that could be a problem.
<c0rnelius> r​pardini: above for you. oops
<c0rnelius> TRS-80: Pretty sure that was the dude who use to cry all the time? I think he came out with a book recently, both written and audio where he dropped the F-Bomb on Ted Cruz.
nexgen has joined #armbian
nexgen2 has joined #armbian
nexgen has quit [Client Quit]
nexgen2 has quit [Remote host closed the connection]
nexgen has joined #armbian
alekksander has joined #armbian
nexgen has quit [Quit: Leaving]
nexgen has joined #armbian
nexgen has quit [Read error: Connection reset by peer]
nexgen has joined #armbian
nexgen has quit [Read error: Connection reset by peer]
nexgen has joined #armbian
nexgen has quit [Remote host closed the connection]
nexgen has joined #armbian
TRS-80 has quit [Quit: WeeChat 3.0]
nexgen has quit [Quit: Leaving]
nexgen has joined #armbian
TRS-80 has joined #armbian
MoeIcenowy has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
MoeIcenowy has joined #armbian
loki_val has quit [Remote host closed the connection]
crabbedhaloablut has joined #armbian
MoeIcenowy has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
MoeIcenowy has joined #armbian
sunshavi has quit [Remote host closed the connection]
<Armbian-Discord> <r​pardini> c0rnelius: tobetter's tree? really? has he fixed it last 5 days?
<c0rnelius> Fixed what and which branch?
<Armbian-Discord> <r​pardini> That is his current 5.15.y branch
<Armbian-Discord> <r​pardini> he has been rebasing that shit for over a year now
<Armbian-Discord> <r​pardini> I dunno if the audio stuff is correct or not
<Armbian-Discord> <r​pardini> but why split the dtb?
<c0rnelius> He creates his own DTS files for legacy reasons I suppose? As far as I know it all works.
<Armbian-Discord> <r​pardini> yeah, mainline is not legacy
<Armbian-Discord> <r​pardini> his reboot driver + dts stuff is 🤮
<c0rnelius> I actually extract his patch set and pull out the XU4 stuff.
<c0rnelius> the edge one is for 5.16.y. The other 5.15 LTS.
<Armbian-Discord> <r​pardini> well. I'd love if he could show up and talk to us, maybe consider doing it proper, heading for mainline, I can't see why not.
<c0rnelius> Well I don't see why if you or I or someone wanted too they couldn't just add it all to the original mainline dts.
<c0rnelius> But I suppose he keeps it legacy as he works for them?
<c0rnelius> I've only interacted with him when I find bugs.
<c0rnelius> I don't get into the politics of it
<Armbian-Discord> <r​pardini> the SD card speed reset for high speed needs coordinates patches across u-boot and dts I think
<Armbian-Discord> <r​pardini> and they're the vendor, they should care. we're unpaid volunteers
<Armbian-Discord> <r​pardini> nice and what u-boot do you use?
<c0rnelius> I don't know for sure to what degree it actually works?
<c0rnelius> Mainline
<c0rnelius> Current
<c0rnelius> Current stable that is
<Armbian-Discord> <r​pardini> interesting. if you could get a log of sdr104 SD card actually rebooting from SD using mainline u-boot would be very useful
<c0rnelius> What u-boot are you using?
<Armbian-Discord> <r​pardini> gotta check, I think 2021.07
<c0rnelius> I'm on v2021.10
<c0rnelius> I only apply the the one patch for uboot and then add the nomap patch so the board doesn't lock up.
<Armbian-Discord> <r​pardini> yeah, Armbian meson64 default is 2021.07 plus a few patches from chewitt, including DT sync
<Armbian-Discord> <r​pardini> I mean, neil's patches in chewitt's tree
<c0rnelius> I gotcha
<Armbian-Discord> <r​pardini> Might be time for a resync. I personally don't care about SD cards, but I know users do.
f476_ has quit [Ping timeout: 256 seconds]
<Armbian-Discord> <r​pardini> (my N2+ boots from SPI to USB3 with no SD/eMMC)
<c0rnelius> My neither. I only using it for testing, after that flash to eMMC.
<c0rnelius> Thats nifty.
<Armbian-Discord> <r​pardini> I gotta find time, and find/write a tool for working these patches
<c0rnelius> I don't ever use the SPI as I mostly run headless. So for me it would be like adding grub... pointless.
<c0rnelius> My interest is more in stability and making sure everything is working on a basic level.
nexgen has quit [Quit: Leaving]
nexgen has joined #armbian
<c0rnelius> As far mainline is not Legacy most of the work done in his branch would probs not be accepted anyway. Its like the RPI foundations attempts of mainlining things. Its just not in the cards or up to snuff. Its a bunch hacks or re-working the problem in order to get said boards to work.
<c0rnelius> Its not that his DTS is the problem in my own opinion, its the fact everything else gets reverted or added that is the true problem.
<c0rnelius> So it all ends up being very specific.
<c0rnelius> From what I can tell this is a common theme across arm.
<c0rnelius> His patch set for the H/C4 N2/+ is 3226 lines. That's added and reverted code and thats not even including the DTS and Overlays. And so far thats just for 5.15.y. I think 5.16.y is larger? And that's me ripping out all the XU4 stuff :) If I even got it all?
<Armbian-Discord> <r​pardini> I dunno or care about the size, if it worked and made sense across meson64. otherwise, you can just build his tree, hack the dtb reference, and use that, if it works for you. but it's not proper.
<Armbian-Discord> <r​pardini> gotta go! cya
<c0rnelius> cya... note: show me one proper arm board? Like realistically.
sunshavi has joined #armbian
nexgen has quit [Read error: Connection reset by peer]
<TRS-80> c0rnelius: You ain't wrong! And this is sort of the problem here, hence me always telling (especially new) people "ARM world is not like x86 you are used to"
<TRS-80> luckily few people with some sense/taste are trying to remedy the situation as much as they can (Armbian), thank Goodness
<TRS-80> and few others who care about mainlining
nexgen has joined #armbian
alekksander has quit [Quit: Konversation terminated!]
<c0rnelius> TRS-80: Well that's an opinion.
<TRS-80> > That's just like, your opinon, maaaaaan :)
<c0rnelius> I didn't realize I was having one? I thought I was stating facts. As you said these aren't X86.
<c0rnelius> Even X86 sometimes need hacks.
<TRS-80> I was joking what you said to me back at you, the quote from The Dude IIRC
<TRS-80> relax mate
<c0rnelius> I'm feeling fine :) Our banter back and forth has always been good in my opinion.
* TRS-80 returns with 2 beers, cracks one open and hands it to c0rnelius then opens the other for himself
<TRS-80> happy Friday, cheers!
<c0rnelius> Thanx
* TRS-80 made some progress on tag gardening functions today
<c0rnelius> TRS-80: Have you ever seen the Movie?: Beyond the Black Rainbow
<TRS-80> I don't think so.
* TRS-80 goes to look it up
<c0rnelius> The ending of it is not `grand` in anyway shape or form. But I believe that was the point.
<c0rnelius> Worth a watch.
<TRS-80> sounds creepy, I would need to be in the right mood for that
<c0rnelius> It highlights the mechanisms of control in which we live in. Just like the need to make adjustments to code and revert this and that, it does so on a human perspective level. Its a very interesting film.
<c0rnelius> Right vs Wrong... Wrong vs Right? There is none... So like a SBC we do what we have to make it work.
<c0rnelius> In the end of course... mainline is the goal.