libera/#devuan-arm/ Sunday, 2021-10-24

* Vall tips his hat to c0rnelius16:40
VallHello everyone16:40
c0rneliusHey16:41
VallDo we have an image for official Chimaera for RPi4 available somewhere?16:42
c0rneliusI have non-official ones -> https://github.com/pyavitz/binary/releases/tag/images16:43
c0rneliusAfter flashing and going through the setup you should upgrade the kernel `fetch -u`16:44
c0rneliusI found a bug and plus I think there was an audio issue that got fixed since I made those.16:44
Vallc0rnelius: thanks again. Your efforts on creating/maintaining these images are really appreciated.16:46
c0rneliusThere is also a noticeable /dev/log error, but that appears to be in Bullseye too.16:46
c0rneliusYou are welcome.16:46
Vallc0rnelius: I wonder why you aren't the Devuan official RPi image provider already16:50
VallAs you seem to be the "de facto" one16:51
c0rneliusI think adam takes care of that?16:51
VallNever heard about/from adam. Is he "adam_free2air"?16:52
c0rneliusBut I honestly don't know :) about the inner workings of the Devuan hierarchy and who does what exactly?16:52
c0rneliusyeah thats him16:52
Vallc0rnelius: :-)16:53
VallAnyway, I'm glad we Devuan RPi4 users have you around ;-)16:54
c0rneliusWell I'm happy you appreciate at it.16:54
VallMy super-dooper new SD card should arrive sometime this week and then I plan on resuming my RPi4 KVM host project16:57
c0rneliusNice.17:03
c0rneliusbrb. unless I broke something.17:13
Vallc0rnelius: looks like nothing broke ;-)17:33
c0rneliusVall: Also didn't fix the bluetooth problem.17:36
c0rneliusStupid NanoPi NEO Plus217:36
c0rneliusTrying to port this over to 5.14.y -> https://github.com/friendlyarm/linux/commit/de8acffc1a023f92f12fa9c9a4ff9656e9be14a2#diff-f03340c270d98b07cd53c0768aa766c817:38
c0rneliusI didn't the add the rtc bit though as I'm not sure how it relates.17:38
Vallc0rnelius: having a look at the commit17:38
Vallholy kernel patches, Batman!17:39
VallI wonder why these aren't submitted to upstream (Torvald/GregKH?)17:42
c0rneliusVery good question. Probs because it don't work :)17:46
VallLOL17:46
c0rneliusAll the allwinner wifi/bluetooth modules are kind of sketchy.17:46
VallExplains all observable facts ;-)17:46
c0rneliusLike I got the PineA64+ to work, which is just really a glorified H5, but even with that one the bluetooth only works if you power down the board. If you reboot it doesn't come back.17:47
VallMy impression is that most (all?) ARM boards aren't as well supported as the RPi series17:48
Vallbrb17:48
c0rneliusIt depends. Most of them have a group of people `The Vendor` that does work on them. But most of the work done doesn't make it upstream. Just like the Foundation kernels.17:49
c0rneliusThey make to many edits, that only make sense for said board.17:50
Vallback18:23
Vallc0rnelius: makes sense.  What doesn make sense is the fragmentation that results, IMHO18:24
hyrcanusit makes sense when you consider all the constraints involved18:27
hyrcanusit's just not the world as we'd wish it to be18:27
hyrcanusa simple example is the decision to put /tmp on a ramdrive18:28
hyrcanusthat was done by people who want to imagine that ram is no longer a constraint18:28
hyrcanusmuch like the wishful thinking behind Modern Monetary Theory :)18:29
Vallhyrcanus: hummmrmrmr... "a simple example is the decision to put /tmp on a ramdrive" -> I don't think this is an example as it's really not something done in the kernel18:39
hyrcanusby distro maintainers18:39
hyrcanussorry i switched context18:39
Vallno prob18:39
VallI'm not as worried about distro fragmentation (actually, I'm not worried at all)18:40
Vallbut kernel fragmentation really bothers me18:40
hyrcanusi wouldn't quite call it fragmentation even18:40
hyrcanusi mean, people do some work, for drivers18:40
hyrcanusoften that work isn't good enough18:40
hyrcanusthose are then available to build-in, if you want.  but they aren't up to quality for mainline18:41
hyrcanusand if you want to build them into a newer kernel, do the work, right18:41
hyrcanusand you can then publish the 'new gpu drivers for 5.15'18:41
hyrcanusthat work doesn't automatically get done if it's "in mainline"18:41
hyrcanusthe stuff that's maintained reliably in mainline is stuff that's needed, like filesystems18:43
Vallhyrcanus: agreed re:drivers. But from c0rnelius comments, I was under the impression he was referring also to edits *outside* pure drivers18:44
hyrcanusok sorry, my brain isn't maintaining the state of the conversation properly - I am multitasking here18:44
VallFor example, in the commit he referred to (https://github.com/friendlyarm/linux/commit/de8acffc1a023f92f12fa9c9a4ff9656e9be14a2#diff-f03340c270d98b07cd53c0768aa766c8), only one of the altered files is a driver18:45
Vallhyrcanus: no prob ;-)18:45
hyrcanusah, i am in agreement now18:45
VallWhat I think happens is that it's so difficult to get something in mainline that most developers don't bother18:46
Valls/most/many/18:46
hyrcanusi can't even get a patch into xscreensaver so... :P18:47
hyrcanusalso fuck you jamie, my configurable colors for glmatrix was a good patch18:48
Vallheh believe me, I know the feeling ;-)18:48
VallBack in the day, I developed some kernel patches for my own use. After lurking in (at the time) the kernel developers mailing list to see how I should proceed in submitting them, I just gave up18:50
VallA lot of aggravation and abuse was going on, and I would not sign up for any of it18:50
VallI've been a long time away from the kernel developing scene, but from what I see re: other people experiences, seems it just got worse18:51
Vallc0rnelius: BTW, if you ever need any beta-testers for RPi4 stuff, just drop me a /msg18:52
c0rneliusVall: Will do18:58

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!