| - mjw (QUIT: Ping timeout: 265 seconds) (~mjw@gnu.wildebeest.org) | 00:31 | |
| + siviq (~siviq@user/siviq) | 00:40 | |
| - siviq (QUIT: Client Quit) (~siviq@user/siviq) | 00:43 | |
| chartreuse | ovra: They're already all set to true. | 00:55 |
|---|---|---|
| chartreuse | Does seem like the webrenderer is possibly being blacklisted for the reason "OpenGL less than 3" | 00:55 |
| - pomel0 (QUIT: Ping timeout: 248 seconds) (~pomel0@user/pomel0) | 00:56 | |
| chartreuse | Seems like it's reverting to software gl which is why it's painfully slow? | 00:57 |
| minute | josch: 57 you say | 01:02 |
| minute | chartreuse: even if icky, i would suggest chromium-ish browser on imx8m, generally better perf | 01:03 |
| chartreuse | minute: Fair, though I had them basically comparable performance a while back before setting back up the Reform and upgrading. Just seems like something degraded | 01:04 |
| minute | chartreuse: maybe firefox got slower | 01:06 |
| minute | chartreuse: i wonder if maybe cpu rendering would be faster... what happens with LIBGL_ALWAYS_SOFTWARE? | 01:07 |
| chartreuse | Probably, I'm guessing some GL stuff got changed either requiring a newer version, or in some way the GPU got blacklisted | 01:07 |
| minute | chartreuse: webrender never worked because it always needed gles 3.x | 01:08 |
| minute | chartreuse: i think gpu compositing used to work | 01:08 |
| chartreuse | I think it's already in that since it shows under WebRenderer Blocklisted; failure code FEATURE_FAILURE_SOFTWARE_GL | 01:09 |
| minute | josch: btw what's the output of sudo gpioget 73? | 01:09 |
| chartreuse | Feels the same with LIBGL_ALWAYS_SOFTWARE set | 01:10 |
| minute | chartreuse: webrender sw doesn't mean that all gpu features are off. they separate gpu rendering of content vs gpu compositing of rendered content | 01:10 |
| minute | chartreuse: ok | 01:10 |
| minute | chartreuse: does about:support say anything interesting about compositing without libgl_always_software? | 01:11 |
| minute | josch: if a reset line is gone that was there in a previous kernel, than that is 100% a problem | 01:12 |
| minute | josch: stuff like that happens when a devicetree property is renamed or moved but the .dts is not changed, for example | 01:13 |
| chartreuse | HW_Compositing and OpenGL_Compositing are forced on and no errors there. WebRender_compositor has a failure on Blocklisted; failure code FEATURE_FAILURE_NO_DMABUF | 01:13 |
| minute | josch: were there any changes in our upstream .dts / banapi cm4 .dts(i) in linux? | 01:13 |
| minute | chartreuse: smooth scrolling off? | 01:14 |
| chartreuse | I don't remember it being this bad before even with software libgl, I do remember doing a bunch of tweaking and preferences to get it to force using HW compositing before | 01:14 |
| chartreuse | minute: It's on, but the lag is all in the UI, like moving the mouse over the tab bar slows down, or trying to type in the url bar. I definitly had managed to get smooth scrolling smooth before | 01:15 |
| chartreuse | I should have just ordered the 3588(?) cpu module and got the increased performance from that. Guess it's a challenge now to find out why the browser got so slow. No difference with a fresh firefox profile or upgrading to 148 instead of 140-esr | 01:17 |
| minute | chartreuse: wow ok, sounds bad... sounds almost like your whole desktop is sw rendered | 01:18 |
| minute | chartreuse: are other apps normal? | 01:18 |
| chartreuse | Everything else seems fine. I had kicad open up the reform pcb layout and while a little slow just seemed cpu limited and not the toolkit slowing down | 01:19 |
| minute | very strange | 01:19 |
| + siviq (~siviq@user/siviq) | 01:20 | |
| chartreuse | Might just end up doing a fresh install, or at least put the latest image on the sd card. It was a bit of a chore getting it to upgrade so it's possible some stuff is a bit borked | 01:21 |
| minute | chartreuse: have you tried overriding GLES version for fun? | 01:21 |
| chartreuse | Not yet, don't remember all the magic environment variables and didn't seem I kept a document of them | 01:22 |
| chartreuse | libEGL warning: failed to get driver name for fd -1 | 01:22 |
| chartreuse | libEGL warning: MESA-LOADER: failed to retrieve device information | 01:22 |
| minute | chartreuse: export MESA_GLES_VERSION_OVERRIDE=3.0; export MESA_GLSL_VERSION_OVERRIDE=300 | 01:22 |
| chartreuse | MESA: error: ZINK: vkCreateInstance failed (VK_ERROR_INCOMPATIBLE_DRIVER) | 01:22 |
| chartreuse | libEGL warning: egl: failed to create dri2 screen | 01:22 |
| chartreuse | GetShaderInfoLog() -> | 01:22 |
| chartreuse | 0:2(12): error: extension `GL_EXT_shader_texture_lod' unsupported in fragment shader | 01:22 |
| minute | lol @ zink... well, i don't think you have vulkan on etnaviv | 01:23 |
| chartreuse | Feels about the same with the override, no changes in what about:support reports | 01:24 |
| chartreuse | Those were just the few warnings that were in console starting firefox. I think those have always been there though | 01:24 |
| minute | ok. and export MESA_GL_VERSION_OVERRIDE=3.2; export MESA_GLSL_VERSION_OVERRIDE=320; ? | 01:25 |
| minute | anyway, probably can't squeeze much more from that gpu at the moment | 01:25 |
| minute | maybe you could find an older version of FF | 01:25 |
| chartreuse | About the same, no new errors with that | 01:26 |
| minute | ok... bummer | 01:26 |
| - lidstah (QUIT: Ping timeout: 258 seconds) (~lidstah@gateway/tor-sasl/lidstah) | 01:26 | |
| chartreuse | I did try the firefox 115 before I upgraded and even the MNT forums couldn't really be used as it was an "unsupported browser" | 01:26 |
| chartreuse | I'll try with a fresh image and see if maybe that helps | 01:27 |
| minute | uff, damn "modern" web software | 01:27 |
| minute | chartreuse: the fastest web engine for that kind of chip is probably WPE WebKit, debian even has the "cog" package, but afaik it's borked (or last time i checked) | 01:28 |
| chartreuse | Modern web browsers and sites requiring them are the only reason to need newer computers. Besides the internet using 20+ year old computers feels just fine | 01:29 |
| minute | chartreuse: i have to say i'm embarassed that the forum has a browser version requirement | 01:29 |
| minute | that's what i get for not NIHing | 01:29 |
| chartreuse | You can load it, but it's in a kind of almost basic HTML view and you can't log in or comment or such | 01:30 |
| minute | if i had unlimited time i would make sure that all the websites ran OK on an amiga 1200... | 01:30 |
| chartreuse | Clearly just needs to be phpBB :P | 01:30 |
| minute | lmao... unfortunately i have php allergy | 01:30 |
| chartreuse | Hmm, need some forum written in Perl/CGI | 01:30 |
| minute | yeah, that's how i started webshit development back in the day with! | 01:31 |
| kfx | this is a job for d-lang | 01:31 |
| minute | had an o'reilly book about that | 01:31 |
| minute | well, at least you can go to this IRC channel on an amiga 1200. so all is not lost | 01:31 |
| kfx | oh no, dfeed is slopware now :( | 01:32 |
| - vagrantc (QUIT: Ping timeout: 244 seconds) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 01:32 | |
| chartreuse | Yeah IRC is great like that | 01:32 |
| minute | kfx: ;/ | 01:32 |
| kfx | afaik it was the last webforum/mailinglist combo | 01:32 |
| minute | ircv3 is awesome too... i recently implemented a comment/discussion function in our internal knowledge base sw by deploying ergo ircv3 server and wrote a tiny bit of react/javascript with manual irc command parsing including the chat history feature | 01:33 |
| minute | ergo has builtin irc-over-websockets too... | 01:34 |
| minute | i wonder if the cog package in debian will ever be fixed... even on this 8550 mega-platform i get > Cog-Wayland:ERROR:../platform/wayland/cog-platform-wl.c:1727:on_export_wl_egl_image: assertion failed: (s_eglCreateWaylandBufferFromImageWL) | 01:35 |
| chartreuse | Did just see someone talking about ircv3 recently, sounded like it is designed to be backwards compatible in that it would support original irc clients just without the advanced features? | 01:36 |
| chartreuse | Need to look into it more | 01:36 |
| minute | oh /usr/lib/aarch64-linux-gnu/wpe-webkit-2.0/MiniBrowser works | 01:36 |
| minute | chartreuse: correct. the client can negotiate more features | 01:36 |
| minute | including tagged messages | 01:36 |
| minute | and history | 01:36 |
| minute | chartreuse: for fun, try installing libwpewebkit-2.0-1 | 01:37 |
| chartreuse | Yeah the history seems nice (though does then require the moderation features) instead of running bouncers, or leaving irssi running in screen on a server | 01:37 |
| minute | chartreuse: and launch /usr/lib/aarch64-linux-gnu/wpe-webkit-2.0/MiniBrowser | 01:37 |
| minute | it doesn't have an address bar but you can pass an url, like /usr/lib/aarch64-linux-gnu/wpe-webkit-2.0/MiniBrowser https://mntre.com | 01:38 |
| minute | someone should make a little chrome (no pun intended) for wpe | 01:39 |
| - siviq (QUIT: Quit: Client closed) (~siviq@user/siviq) | 01:39 | |
| minute | chartreuse: that thing renders community.mnt.re fine for me! :D | 01:39 |
| chartreuse | minute: It gives mostly the same warnings in console as firefox did, but never appears, last message shown is ** (MiniBrowser:20531): WARNING **: 18:41:24.108: WindowViewBackend: cannot yet handle wpe_fdo_shm_exported_buffer. | 01:42 |
| minute | chartreuse: i hope you did that in a new terminal without those MESA overrides etc | 01:42 |
| chartreuse | It's in a different terminal yeah | 01:42 |
| minute | chartreuse: is that on sway? | 01:42 |
| chartreuse | Yeah | 01:43 |
| minute | chartreuse: mind giving me a pastebin of glxinfo ? | 01:43 |
| - voltaire28_ (QUIT: Quit: WeeChat 4.8.1) (~jlafon@28.162.2.93.rev.sfr.net) | 01:43 | |
| chartreuse | http://pastie.org/p/0myMTRwmInxT8m2haIntdA | 01:46 |
| chartreuse | (Was trying to use pastebin but every time I hit create it would do a cloudflare check, and then after than bring me back to an empty paste >_>) | 01:47 |
| minute | thanks! | 01:48 |
| minute | chartreuse: you're on sw rendering | 01:48 |
| minute | chartreuse: GPU is not in use at all | 01:49 |
| chartreuse | Huh... Wonder how I got into that state | 01:49 |
| minute | > Device: llvmpipe (LLVM 21.1.8, 128 bits) (0xffffffff) | 01:49 |
| minute | so the cpu is emulating a gpu | 01:49 |
| minute | chartreuse: lsmod | grep etnaviv | 01:50 |
| chartreuse | I don't see anything in env. etnaviv 122880 0 | 01:50 |
| minute | ok, driver is loaded | 01:50 |
| minute | (kernel driver) | 01:50 |
| minute | chartreuse: apt install libgl1-mesa-dri libgles2 | 01:52 |
| minute | chartreuse: and make sure no weird stuff in /usr/local/lib | 01:52 |
| minute | chartreuse: and there's an off chance that sway itself is in software (pixman) mode | 01:52 |
| chartreuse | Both already installed. Only python3.13 and 3.14 in /usr/local/lib | 01:53 |
| minute | maybe sway can't access the gpu | 01:53 |
| minute | could be permission/group issue. maybe you're not in "render" group | 01:53 |
| minute | wasn't there a change in debian at some point about that josch ? | 01:54 |
| minute | chartreuse: check if you're in the render group, and if not, add your account to it | 01:54 |
| chartreuse | I'm not in the render group | 01:54 |
| minute | chartreuse: aha | 01:54 |
| minute | chartreuse: that's probably your issue then | 01:54 |
| chartreuse | Guess that changed at some point as that group wasn't a thing before | 01:55 |
| minute | yep | 01:55 |
| minute | chartreuse: if you do `ls -l /dev/dri` | 01:55 |
| minute | chartreuse: i think you'll see `crw-rw----+ 1 root render 226, 128 Mar 19 01:17 renderD128` | 01:55 |
| chartreuse | Yeah it's owned by root:render | 01:55 |
| minute | chartreuse: but card0 has group "video" which existed before | 01:55 |
| chartreuse | While card0/card1 are in video which is where I'm already in | 01:56 |
| minute | that's why you can open the display. but not render with the render node | 01:56 |
| minute | very tricky trap | 01:56 |
| minute | why do i have this useless stuff in my head... | 01:56 |
| chartreuse | I guess it's because we're on unstable, surprised it wasn't a note during the upgrade. Didn't see any of the notes | 01:56 |
| chartreuse | alright I'll be back in a second | 01:56 |
| - chartreuse (QUIT: Quit: leaving) (~chartreus@104.246.217.181) | 01:56 | |
| minute | chartreuse: no it's been like that for a longer while @ render group. but maybe last year | 01:57 |
| minute | lol linux from scratch just replaces the render group with video in udev https://www.linuxfromscratch.org/lfs/view/12.3-systemd/chapter08/systemd.html | 01:59 |
| + chartreuse (~chartreus@104.246.217.181) | 01:59 | |
| chartreuse | That did it, firefox actually is usable now | 01:59 |
| + pomel0 (~pomel0@user/pomel0) | 02:00 | |
| chartreuse | Firefox reports WebRender (Software) now and actually scrolls reasonably | 02:00 |
| minute | chartreuse: great, happy to help! | 02:04 |
| chartreuse | Guessing a fresh install would have worked then, but good to know what the change actually was | 02:06 |
| minute | yeah. also in case other people run into this | 02:07 |
| chartreuse | Next thing I need to do is find my needle file set and modify this steel port cover. Missed the note on the shop that it had changed for USB-C on the newer ones and to ask support when ordering. At least it shouldn't be too hard to enlarge | 02:09 |
| minute | oh right. | 02:09 |
| minute | another option might be mouser/crowd supply, i think they have old steel port cover in stock | 02:10 |
| minute | (maybe pricey though, don't know atm... filing is surely cheaper) | 02:10 |
| minute | i just installed epiphany-browser here to test (gnome web), seems snappy at least on this platform | 02:12 |
| chartreuse | I'll just file for now, and save the money to put towards buying the RK3588 and maybe upgrading other parts like the keyboard | 02:15 |
| minute | good plan :3 | 02:16 |
| minute | josch: i realized that we're not shipping our own .dts for meson-g12b-bananapi-cm4-mnt-reform2.dts, instead relying on an old upstream copy and patching that file a bit. i guess the problem is that noone is making sure that that old upstreamed dts still works. | 02:36 |
| + pengu1nx7 (~pengu1nx@user/pengu1nx) | 02:41 | |
| - kfx (QUIT: Remote host closed the connection) (~kfx@104.255.228.215) | 02:42 | |
| - pengu1nx (QUIT: Ping timeout: 268 seconds) (~pengu1nx@user/pengu1nx) | 02:43 | |
| * pengu1nx7 -> pengu1nx | 02:43 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 02:50 | |
| - paperManu_ (QUIT: Ping timeout: 245 seconds) (~paperManu@173.206.232.56) | 02:54 | |
| - paperManu (QUIT: Ping timeout: 248 seconds) (~paperManu@173.206.232.56) | 02:54 | |
| + kfx (~kfx@grendel.sciops.net) | 03:38 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 04:40 | |
| - vagrantc (QUIT: Ping timeout: 248 seconds) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 05:04 | |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.91.73) | 05:10 | |
| + Ar|stote|is (~linx@149.210.91.73) | 05:19 | |
| - potash (QUIT: Quit: The Lounge - https://thelounge.chat) (~potash@user/foghorn) | 05:22 | |
| - robin (QUIT: Ping timeout: 256 seconds) (~robin@user/terpri) | 05:42 | |
| + robin (~robin@user/terpri) | 05:43 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 07:01 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 07:26 | |
| josch | minute: if the upstream copy doesn't change then patching that has the same effect as shipping our own static copy, no? | 07:27 |
| josch | if your hunch is, that this is is a dts change, let me show you a diff later today | 07:28 |
| josch | chartreuse, minute: the /dev/dri situation can be worked around by changing groups but the intended way is to not go via unix permissions | 07:35 |
| josch | if i do this: getfacl /dev/dri/renderD128 | 07:35 |
| josch | then i get: user:josch:rw- | 07:36 |
| josch | you can see that facl rules apply from the ls -l output of /dev/dri/renderD128 which includes a "+" at the end of the permission line | 07:36 |
| josch | and this should be set up by /usr/lib/udev/rules.d/70-uaccess.rules which has: | 07:37 |
| josch | SUBSYSTEM=="drm", KERNEL=="renderD*", TAG+="uaccess" | 07:37 |
| josch | and at the bottom it says TAGS=:uaccess:seat: which is what gives it the right ACL for my user | 07:37 |
| josch | we want to the bottom of this with wickedshell but that was nearly exactly a year ago: https://mntre.com/reform-irc-logs/2025-03-21.log.html | 07:39 |
| - lidstah (QUIT: Ping timeout: 258 seconds) (~lidstah@gateway/tor-sasl/lidstah) | 08:14 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 08:15 | |
| + voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net) | 08:22 | |
| - voltaire28_ (QUIT: Quit: WeeChat 4.8.1) (~jlafon@28.162.2.93.rev.sfr.net) | 09:22 | |
| josch | *went | 09:33 |
| orva | TIL that hw acceleration is gated seperalaty from video access | 09:34 |
| orva | Like, it probably does make sense in hindsight in security sense but that was not obvious for me | 09:34 |
| orva | Ugh, today seems to be dyslexia on day | 09:35 |
| orva | Format previous messages in a way that they make sense | 09:35 |
| + mjw (~mjw@gnu.wildebeest.org) | 10:01 | |
| - chartreuse (QUIT: Ping timeout: 272 seconds) (~chartreus@104.246.217.181) | 10:29 | |
| + chartreuse (~chartreus@192.252.234.186) | 10:30 | |
| - chartreuse (QUIT: Ping timeout: 264 seconds) (~chartreus@192.252.234.186) | 10:38 | |
| + chartreuse (~chartreus@104.246.216.98) | 10:40 | |
| - wickedshell (QUIT: Ping timeout: 264 seconds) (~wickedshe@2601:8c0:c7c:3572:b584:d88e:dcfc:5963) | 10:44 | |
| josch | minute: please hold off for a bit with investigating a311d ethernet. I now know why I thought that this was fixed in 6.19 | 10:54 |
| josch | Linux m 6.19.0-dirty #4 SMP PREEMPT Thu Mar 19 07:49:51 CET 2026 aarch64 GNU/Linux | 10:54 |
| josch | I'm on 6.19 and both ethernet and usb work just fine | 10:54 |
| josch | So maybe this is a change between 6.19 and 6.19.8 or it is because i only applied the a311d patch stack and the problem is introduced because of one of the other patches or it is because i'm not building with the full Debian kernel config but with "make localmodconfig" | 10:55 |
| josch | minute: in any case, you see that there is still more that i can do and i'll update you once i find out more | 10:56 |
| + paperManu (~paperManu@173.206.232.56) | 11:40 | |
| - cow321 (QUIT: Ping timeout: 268 seconds) (~deflated8@user/meow/deflated8837) | 11:59 | |
| + cow321 (~deflated8@user/meow/deflated8837) | 12:06 | |
| - ex-parrot (QUIT: Ping timeout: 276 seconds) (~fincham@user/ex-parrot) | 12:26 | |
| + ex-parrot (~fincham@user/ex-parrot) | 12:26 | |
| - b0 (QUIT: Read error: Connection reset by peer) (~b0@user/b0) | 12:49 | |
| - johl (QUIT: Read error: Connection reset by peer) (~johl@wikidata/Jens-Ohlig) | 12:49 | |
| + johl (~johl@leo.uberspace.de) | 12:49 | |
| + wickedshell (~wickedshe@2601:8c0:c7c:3572:130:9754:8db6:ff97) | 12:50 | |
| + b0 (~b0@user/b0) | 12:51 | |
| - paperManu (QUIT: Ping timeout: 244 seconds) (~paperManu@173.206.232.56) | 13:09 | |
| - mjw (QUIT: Ping timeout: 255 seconds) (~mjw@gnu.wildebeest.org) | 13:18 | |
| * Guest6628 -> mjw | 13:28 | |
| + siviq (~siviq@user/siviq) | 13:34 | |
| - siviq (QUIT: Ping timeout: 240 seconds) (~siviq@user/siviq) | 13:38 | |
| + paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 13:42 | |
| + hlep (~hlep@ip-046-223-162-180.um13.pools.vodafone-ip.de) | 13:58 | |
| - hlep (QUIT: Changing host) (~hlep@ip-046-223-162-180.um13.pools.vodafone-ip.de) | 14:08 | |
| + hlep (~hlep@user/hlep) | 14:08 | |
| + eery (~eery@192.150.253.237) | 14:36 | |
| - hlep (QUIT: Quit: hlep) (~hlep@user/hlep) | 14:49 | |
| + hlep (~hlep@ip-046-223-162-180.um13.pools.vodafone-ip.de) | 14:50 | |
| - hlep (QUIT: Changing host) (~hlep@ip-046-223-162-180.um13.pools.vodafone-ip.de) | 14:50 | |
| + hlep (~hlep@user/hlep) | 14:50 | |
| - hlep (QUIT: Client Quit) (~hlep@user/hlep) | 14:51 | |
| + hlep (~hlep@user/hlep) | 14:51 | |
| + wielaard (~mjw@gnu.wildebeest.org) | 14:52 | |
| minute | josch: ok thanks! | 15:02 |
| minute | josch: wakest will now test 6.19 package on imx8mp pocket reform | 15:23 |
| josch | nice, thank you wakest__ ! | 15:23 |
| - marty (QUIT: Quit: WeeChat 4.6.3) (~marty@143.244.47.98) | 15:36 | |
| - andreas-e (QUIT: Quit: Leaving) (~Andreas@2a02-8434-b6a3-e901-facc-8e87-8e54-890d.rev.sfr.net) | 15:43 | |
| - hlep (QUIT: Quit: hlep) (~hlep@user/hlep) | 15:47 | |
| - paperManu (QUIT: Ping timeout: 264 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 15:54 | |
| + paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 15:56 | |
| + marty (~marty@146.70.168.163) | 15:59 | |
| + robin_ (~robin@user/terpri) | 16:01 | |
| - robin (QUIT: Ping timeout: 264 seconds) (~robin@user/terpri) | 16:04 | |
| - pomel0 (QUIT: Ping timeout: 276 seconds) (~pomel0@user/pomel0) | 16:24 | |
| + pomel0 (~pomel0@user/pomel0) | 16:24 | |
| minute | josch: as our debian expert, could you explain why this would happen? https://minute.social/secret-room/20260319-dkms-619-lpc-weirdness.txt | 16:29 |
| minute | josch: ah, when writing down what we did i think i can answer my own question, sorry | 16:30 |
| minute | josch: we didn't install kernel headers :3 | 16:30 |
| minute | josch: also, audio is broken with 6.19 on imx8mp pocket, because: https://minute.social/secret-room/20260319-imx8mp-sound-parse-error.txt | 16:31 |
| minute | josch: otherwise it's fine. gnome performance is pretty acceptable even (with heatsink on, not with heatsink off ^^) | 16:31 |
| - murphnj (QUIT: Ping timeout: 272 seconds) (~murphnj@user/murphnj) | 17:12 | |
| + murphnj (~murphnj@user/murphnj) | 17:12 | |
| - elb (QUIT: Remote host closed the connection) (~elb@2600:4041:6682:ea00:4d81:6f3b:64f2:a848) | 17:18 | |
| + b05 (~b0@user/b0) | 17:21 | |
| - b0 (QUIT: Ping timeout: 248 seconds) (~b0@user/b0) | 17:23 | |
| * b05 -> b0 | 17:23 | |
| minute | ah the "asoc-simple-card: parse error" is misleading, it probably means that a device dependency is not probing | 17:24 |
| minute | i.e. one of the codecs | 17:24 |
| minute | > i2c 2-0018: deferred probe pending: i2c: wait for supplier /soc@0/bus@30df0000/clock-controller@30e20000 | 17:29 |
| minute | > fsl,imx8mp-audio-blk-ctrl | 17:30 |
| minute | but that gets built as part of CONFIG_CLK_IMX8MP | 17:31 |
| minute | which is a =y | 17:32 |
| josch | minute: running "sudo reform-check" should've listed it if kernel headers were missing. If that wasn't part of reform-check output, then that would be a bug in reform-check. Generally you want to have the linux-headers-mnt-reform-arm64 meta-package installed which will pull in the latest version | 17:41 |
| minute | josch: yeah, sorry for the noise! | 17:42 |
| minute | i pushed a git rm on a imx8mp related audio patch, hope that'll just fix the audio issue | 17:43 |
| minute | otherwise there are no blockers re:6.19 | 17:43 |
| minute | btw we saw that imx8mp has issues with certain HDMI modes/freqs, but others works fine. it wasn't able to output correctly on 1080p @ 74.9Hz, but 60/50Hz worked fine | 17:43 |
| josch | minute: no worries, better ask early :) | 17:43 |
| minute | this problem was there before though. i just didn't attribute it to the mode itself | 17:44 |
| minute | most of the time you get just a single color on screen, or crash, or no signal | 17:44 |
| minute | (with that mode) | 17:44 |
| minute | so if someone runs into this issue, it's recommended to just change the mode to another Hz | 17:44 |
| + AndrewJ (~AndrewJ@user/AndrewJ) | 18:01 | |
| AndrewJ | Hello all. Is the reform next still on track to ship at the end of the month? | 18:05 |
| * mjw -> Guest7485 | 18:06 | |
| * wielaard -> mjw | 18:06 | |
| - AndrewJ (QUIT: Quit: Client closed) (~AndrewJ@user/AndrewJ) | 18:22 | |
| + robin (~robin@user/terpri) | 18:43 | |
| - robin_ (QUIT: Ping timeout: 255 seconds) (~robin@user/terpri) | 18:44 | |
| josch | inux m 6.19.8-dirty #5 SMP PREEMPT Thu Mar 19 17:55:12 CET 2026 aarch64 GNU/Linux | 18:59 |
| josch | meson8b-dwmac ff3f0000.ethernet end0: Link is Up - 1Gbps/Full - flow control rx/tx | 18:59 |
| josch | right, so *this* is why i previously observed that 6.19 fixed the ethernet issue -- i'm running 6.19.8 and ethernet works just fine | 18:59 |
| josch | but evidently it doesn't work with the MNT gitlab CI pipeline artifacts so next i'm trying to add the remaining patches and see what happens | 19:00 |
| + siviq (~siviq@user/siviq) | 19:14 | |
| + elb (~elb@2600:4041:6682:ea00:3d07:26f7:5eeb:a869) | 19:34 | |
| minute | josch: does it also work when you comment out that weird workaround in reform-hw-setup? | 19:48 |
| josch | i didn't try that yet. I'll try that next. I'm currently building 6.19.8 with all patches | 19:49 |
| josch | (but config is still a minimal one generated with "make localmodconfig") | 19:49 |
| + ephase1 (~ephase@2a01:e0a:2a:5300:3049:c4ff:fe72:a63c) | 19:57 | |
| - pomel0 (QUIT: Ping timeout: 245 seconds) (~pomel0@user/pomel0) | 20:11 | |
| + pomel0 (~pomel0@user/pomel0) | 20:16 | |
| + ephase (~ephase@82.66.198.11) | 20:38 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 20:39 | |
| minute | if we're lucky, the imx8mp audio problem with 6.19 is a missing new kernel driver/option > # CONFIG_VIDEO_DW100 is not set | 20:45 |
| - pomel0 (QUIT: Read error: Connection reset by peer) (~pomel0@user/pomel0) | 21:05 | |
| + pomel0 (~pomel0@user/pomel0) | 21:06 | |
| - ephase (QUIT: Remote host closed the connection) (~ephase@82.66.198.11) | 21:20 | |
| + ephase (~ephase@82.66.198.11) | 21:27 | |
| - ephase (QUIT: Remote host closed the connection) (~ephase@82.66.198.11) | 21:28 | |
| + ephase (~ephase@82.66.198.11) | 21:28 | |
| josch | i have 6.19.8 with all MNT patches and ethernet is still working | 21:32 |
| josch | minute: but it needs the workaround in reform-hw-setup to work. Without it, the interface is still there but i get no link up | 21:32 |
| josch | now the last difference between my working kernel and the non-working on *should* be the kernel config -- i build a kernel with the full config next overnight and see what happens then... | 21:33 |
| ephase | josch: with an A311D? | 21:33 |
| - ephase1 (QUIT: Quit: WeeChat 4.8.1) (~ephase@2a01:e0a:2a:5300:3049:c4ff:fe72:a63c) | 21:33 | |
| - paperManu (QUIT: Ping timeout: 244 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 22:05 | |
| josch | ephase: yes, a311d | 22:06 |
| josch | ephase: you are also on a311d and are facing the ethernet problem, right? | 22:07 |
| ephase | exact! | 22:07 |
| josch | i'm installing the MNT 6.19 kernel just to make sure i can still reproduce it with that one | 22:08 |
| ephase | great thanks! | 22:10 |
| - timide (QUIT: Read error: Connection reset by peer) (~timide@user/timide) | 22:13 | |
| + timide (~timide@user/timide) | 22:13 | |
| - ephase (QUIT: Read error: Connection reset by peer) (~ephase@82.66.198.11) | 22:17 | |
| josch | noooo ephase don't leave yet! XD | 22:18 |
| josch | somehow i am only downloading with 200 kb/s from mnt gitlab ci so it takes a while... | 22:18 |
| + paperManu (~paperManu@173.206.232.56) | 22:31 | |
| + paperManu_ (~paperManu@173.206.232.56) | 22:34 | |
| josch | meson8b-dwmac ff3f0000.ethernet end0: cannot attach to PHY (error: -ENODEV) | 22:39 |
| josch | Linux m 6.19.8-mnt-reform-arm64 #1 SMP PREEMPT Debian 6.19.8-1+reform20260319T195408Z (2026-03-1 aarch64 GNU/Linux | 22:39 |
| josch | so yes, the bug *does* exist on linux from mnt gitlab ci | 22:39 |
| josch | i forgot another possible source of errors which i'm not accounting for in my builds: i'm building without the Debian patch stack | 22:40 |
| josch | so lets see... | 22:40 |
| - pomel0 (QUIT: Ping timeout: 264 seconds) (~pomel0@user/pomel0) | 22:41 | |
| + pomel0 (~pomel0@user/pomel0) | 22:41 | |
| - pomel0 (QUIT: Ping timeout: 255 seconds) (~pomel0@user/pomel0) | 22:48 | |
| - Gooberpatrol66 (QUIT: Excess Flood) (~Gooberpat@user/gooberpatrol66) | 22:52 | |
| minute | josch: if gitlab ci download is slow, try turning off ipv6. it usually helps with that | 22:54 |
| josch | oh funny :) | 22:54 |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 22:55 | |
| + ephase (~ephase@82.66.198.11) | 23:54 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!