| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@gnu.wildebeest.org) | 00:06 | |
| kfx | a few of the 6.15 patches don't apply cleanly to mainline linux :( | 00:07 |
|---|---|---|
| minute | mainline 6.15.4? | 00:20 |
| minute | btw there are at least 15 unnecessary patches | 00:21 |
| josch | minute: do we care? are they harmful? | 00:24 |
| kfx | minute: yeah, I get failures on the samsung hdptx stuff | 00:25 |
| kfx | 0009, 0011, 0012, 0013, 0014 at least | 00:26 |
| minute | josch: not really. just optics | 00:32 |
| minute | josch: there are other priorities at the moment indeed :D | 00:32 |
| minute | like, i need to organize releasing the latest version of pocket reform sysctl fw on lvfs, and release + make easily available the keyboard v4 fw update for classic reforms, and make sure the new lpc driver module is fine/tested and released | 00:34 |
| josch | minute: and there is also the random reboots of pocket reform with the latest sysctl, right? | 00:34 |
| minute | josch: afaik this is solved by the version with 10s watchdog timeout | 00:34 |
| josch | ah ince | 00:35 |
| minute | josch: at least esther gave positive feedback today and it isn't happening to me | 00:35 |
| josch | perfect | 00:35 |
| minute | but the version in lvfs is the worse one | 00:35 |
| minute | with 1sec timeout | 00:35 |
| minute | but i'm too tired/it's way too late to touch that tonight. hopefully tomorrow | 00:36 |
| josch | of course :) | 00:36 |
| josch | minute: i just had both rk3588 pocket reform and classic reform shut down successfully with reform2-lpc-dkms package with your patch \o/ | 00:59 |
| minute | josch: noice | 01:00 |
| josch | so cool -- thank you for your work! | 01:00 |
| josch | tomorrow i'll check whether it works well with 6.12 as well | 01:01 |
| josch | both rk3588 pocket reform and classic reform above were with 6.15 | 01:01 |
| josch | good night _o/ | 01:05 |
| kfx | the patches marked MERGED -- does anyone know what they've been merged into? | 01:08 |
| bremner | gordon1: I started some experiments with hibernate on the pocket reform (stock kernel). my notes are at https://www.cs.unb.ca/~bremner/blog/posts/hibernate-pocket/ . I have some side quests about noisy logging on the console, then I guess I'll search for more info about the wireless problem (or just use disable that and see if it gets further) | 01:17 |
| bremner | does anyone recognize the message "kernel: hdmi-audio-codec hdmi-audio-codec.4.auto: ASoC error (-19): at snd_soc_dai_prepare() on i2s-hifi" | 01:19 |
| bremner | my console is full of them, which is a bit inconvenient (when not running sway). | 01:20 |
| bremner | I'm running kernel 6.15.4-1~exp1+reform20250628T170930Z | 01:21 |
| minute | bremner: yes, harmless | 01:28 |
| minute | bremner: i'll see if i can patch them out | 01:28 |
| bremner | minute: great@ | 01:28 |
| bremner | s/@/!/ | 01:28 |
| minute | cc josch this is the stuff that motivated me earlier about loglevel | 01:29 |
| bremner | I guess switching to a different VT and logging there is a workaround | 01:37 |
| - chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:85f:9a3f:9300:1ac0:4dff:fedb:a3f1) | 01:40 | |
| bremner | I suppose the fact that I keep plugging into the wrong USB C port and running the batteries all the way down is a me problem | 01:41 |
| minute | bremner: aw, yeah. only the upper port does usb-c pd | 01:42 |
| bremner | so I discovered. again ;) | 01:43 |
| + pff (~pff@2600:4040:572f:5600:f1ff:dcc8:62cd:537) | 01:54 | |
| - shtrophic (QUIT: Remote host closed the connection) (~m-hrdsqi@user/shtrophic) | 03:05 | |
| - paperManu (QUIT: Ping timeout: 272 seconds) (~paperManu@72.10.128.164) | 03:18 | |
| - kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 03:56 | |
| bkeys | gnu131 | 03:57 |
| bkeys | minute: I will check with a311d tomorrow | 03:57 |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 03:58 | |
| - kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:03 | |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:04 | |
| - kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:05 | |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:05 | |
| - svp (QUIT: Ping timeout: 252 seconds) (~svp@host-79-7-240-189.business.telecomitalia.it) | 05:11 | |
| - chartreuse (QUIT: Read error: Connection reset by peer) (~chartreus@S0106908d78501d1d.cg.shawcable.net) | 05:18 | |
| + chartreuse (~chartreus@S0106908d78501d1d.cg.shawcable.net) | 05:22 | |
| + svp (~svp@host-79-7-240-189.business.telecomitalia.it) | 05:29 | |
| - switchy (QUIT: Ping timeout: 248 seconds) (~switchy@mechboards/switchy) | 06:03 | |
| + casparvitch (~casparvit@130.102.161.18) | 06:03 | |
| josch | kfx: torvalds linux git. They are part of v6.16-rc1 | 07:14 |
| josch | minute: i thought your motivation re loglevel was tuigreet? | 07:15 |
| josch | if printing of the hdmi message is wrong then the fix should be to not print the hdmi message -- the fix should not be to silence everything including legitimate messages, no? | 07:16 |
| + switchy (~switchy@mechboards/switchy) | 07:28 | |
| josch | or in other words: by hiding a bug today (noisy hdmi log messages) we make our life easy today (no need to fix this). But we do this at the expense of everybody in the future who actually wants log messages (for whatever thing they are working on) and then has to deal with these messages again. | 07:52 |
| - switchy (QUIT: Remote host closed the connection) (~switchy@mechboards/switchy) | 09:00 | |
| + switchy (~switchy@mechboards/switchy) | 09:01 | |
| - casparvitch (QUIT: Ping timeout: 244 seconds) (~casparvit@130.102.161.18) | 09:46 | |
| + mjw (~mjw@gnu.wildebeest.org) | 10:26 | |
| minute | josch: alright. | 10:28 |
| gordon1 | bremner: i'm pretty sure you can unload wifi module before hibernate and load it after and it should just workâ˘, are you testing your hibernate with nvme? | 10:34 |
| gordon1 | i also noticed usb problems but it probably also could be fixed by unloading the module | 10:36 |
| gordon1 | nvme cannot tho | 10:36 |
| josch | minute: i successfully tested your lpc changes on 6.12 and we can make a new reform-tools release whenever you are ready -- just press the merge button :) | 10:36 |
| gordon1 | it nice to see that graphics still works, that is something i couldn't test yet | 10:36 |
| minute | josch: awesome, many thanks for testing | 10:37 |
| josch | many thanks for fixing this :) | 10:37 |
| minute | josch: merged! | 10:38 |
| josch | sweet :) | 10:39 |
| grimmware | minute: I enjoyed your follow-along thread about the lpc fix last night :) | 10:47 |
| josch | minute: reform-tools 1.75 is tagged and uploaded to Debian experimental. In ~5 hours it will be on the mirrors and then we can re-trigger the reform-debian-packages pipeline. Should I include the tuigreet loglevel maintainer script snippet? | 11:13 |
| minute | grimmware: thx, maybe i should do more | 11:14 |
| minute | josch: cool, oh yeah, please put the script in | 11:14 |
| josch | The Liberux people chose RK3588 also because it will be used in the MNT Reform Next: https://linmob.net/liberux-nexx-an-interview-with-liberux/ | 11:19 |
| minute | that's interesting because i'm choosing qcs6490 for our mobile device because it's less hot :D | 11:36 |
| minute | (and pin compatible qcs8550, lets see how well that works) | 11:38 |
| minute | https://www.thundercomm.com/product/c8550-som/ | 11:39 |
| bremner | josch: minute: would it be hard to put tui-greeter on a different VT than where the log messages go? | 11:42 |
| minute | bremner: good question. somehow i thought they go to all vts, but probably they don't | 11:42 |
| bremner | minute: not afaict. I just switched VT to get away from them :) | 11:43 |
| bremner | let me test more... | 11:43 |
| minute | bremner: does tuigreet have a config option for vt? | 11:43 |
| minute | or the service file perhaps | 11:43 |
| josch | bremner: for me, the kernel log messages went to the tty which i currently was on -- if you find out how to change this, i'm all ears :) | 11:44 |
| bremner | oh, sure enough, I was optimistic. They follow me wherever I go | 11:45 |
| bremner | well, never mind then | 11:46 |
| minute | what i really want is sm8750, but not available just yet ^^ | 11:46 |
| grimmware | I was under the impression that you could send the kernel messages to a specific terminal but I could have made that up.. | 11:46 |
| grimmware | minute: I think there's a lot of value in those threads where you're actively debugging | 11:47 |
| grimmware | One of the things that I lament about the company I work for having gone fully remote is that nobody learns anything from watching over someone else's shoulder any more | 11:48 |
| josch | minute: i was really hoping for you to be able to catch a break one of these days... XD | 11:48 |
| grimmware | and honestly that's always the most practical knowledge | 11:49 |
| bremner | apparently you can give a kernel command console=tty3 | 11:49 |
| wose | josch: setting ForwardToConsole=yes and TTYPath=/dev/tty6 in /etc/systemd/journald.conf should do the trick | 11:49 |
| bremner | ah, instead of /dev/console. Let us see... | 11:50 |
| josch | bremner: with that setting, things like the luks passphrase prompt also would show up on tty3 | 11:50 |
| bremner | right | 11:50 |
| minute | grimmware: ok, will do more... i used to do more of this somehow before i got adhd meds, maybe as a form of public body doubling :D | 11:50 |
| + casparvitch (~casparvit@36.255.114.132) | 11:51 | |
| grimmware | a bit of public accountability to keep on task | 11:51 |
| grimmware | ? | 11:51 |
| minute | josch: one thing i'd also like to figure out one day is how to always have a non primary getty on uart as well, that doesn't claim the password prompt etc | 11:51 |
| josch | yes, same | 11:51 |
| minute | grimmware: also like, there were a lot of situation where i was seeking input from others who are more experienced | 11:52 |
| minute | grimmware: this got a bit more rare since the issues have become less exotic and i can read dry technical manuals better :D | 11:52 |
| minute | i.e. i was dealing with bizarro hardware bringup issues on ls1028a or the stdp2600 | 11:53 |
| minute | the rk3588 has been more forgiving so far (except for some pcie edge cases and the whole suspend thing) | 11:54 |
| bremner | hrm. there are different kinds of messages right? some of these things are not coming through journald, and those seem to be ending up on vt7 | 11:54 |
| minute | it feels like there was more attention/money/work poured into rk3588 mainlining than into some industrial socs that came before. i'm kind of still wondering how that happened | 11:54 |
| grimmware | I think the more that you talk about this debugging in the open the more you a) buy understanding from people as to why these things aren't "just fixed" but also b) give other people the opportunity to learn so they can become more proactive and helpful community members | 11:55 |
| + chomwitt (~chomwitt@2a02:85f:9a3f:9300:1ac0:4dff:fedb:a3f1) | 11:55 | |
| grimmware | I appear to have found a cafe in London that consistently stocks Club Mate <3 | 11:56 |
| minute | i.e. who decided to fund all this work @ collabora (i only know one person/company funding the gpu drivers), and/or maybe it's a side effect of ARM being suddenly interested in panthor... and now for the "tyr" rust driver suddenly google is involved | 11:56 |
| minute | grimmware: that makes a lot of sense | 11:56 |
| josch | grimmware: pfff Club Mate -- Mio Mio Mate is where it's at these days! ;) | 11:56 |
| minute | flora mate was the thing at gpn :D | 11:57 |
| josch | haha indeed :) | 11:57 |
| minute | i'm kind of decaffeinated nowadays due to above mentioned stuff though | 11:57 |
| grimmware | I will endeavor to develop opinions on some of these varieties | 11:57 |
| josch | i met the person behind the counter who was dissing club mate for every customer XD | 11:57 |
| minute | (drinking one decaf in the morning) | 11:57 |
| minute | josch: lol | 11:57 |
| josch | minute: i have no more input for why rk3588 became popular i just observed its popularity from the sidelines as the board has been popping up in virtually every channel i'm lurking in | 11:58 |
| josch | somehow it became a really interesting platform but i suspect that some company was interested in it but no clue who | 11:58 |
| grimmware | is availability a factor? | 11:58 |
| minute | i think performance of the chip itself X availability yeah | 11:59 |
| minute | and some skilled kernel devs / their company picking up the work to clean up the vendor code | 12:00 |
| grimmware | I'm currently looking closely at the state of linux phones again after a very disappointing time with the pinephone | 12:01 |
| minute | i wonder if rockchip can pull of another one like this with the rk3688... | 12:01 |
| bremner | dmesg -D is a big hammer, but maybe useful for someone (and resets on reboot) | 12:01 |
| grimmware | I really just want a handset with mainline kernel support so I can ride it til it dies | 12:01 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!