gnarface | got everything working right now rhys? | 00:02 |
---|---|---|
gnarface | er, sorry rhy^ | 00:02 |
gnarface | i'm curious about your accounts of performance, i've heard some previous accounts that the nvidia cards in these devices are often not faster than the integrated intel gpu for many tasks since intel fixed their opengl support | 00:03 |
gnarface | when the hardware was originally released the intel drivers were seriously sandbagged crap on linux | 00:03 |
gnarface | but later that changed significantly when intel dumped a bunch of money into mesa support, trying to keep up with AMD | 00:04 |
gnarface | (nvidia's drivers don't use mesa, they have their own proprietary thing, so they can't benefit from those mesa improvements) | 00:04 |
hiddener | what is thinkpad console audio control in alsamixer? it seems it causes the warning at boot: failed to import hw:29 use case configuration | 07:19 |
hiddener | there are 2 cards: 0 and 29, 29 is thinkpad console audio control and can't control it at all, 0 works fine | 07:24 |
gnarface | hiddener: do you have all these packages? alsa-tools alsa-topology-conf alsa-ucm-conf alsa-utils | 07:28 |
hiddener | gnarface: i've just installed alsa-tools, the missing one | 07:32 |
darwin | when I try boot install using a radeon rx 6900 xt and select 'install', I always get a screen just full of disorganized lines like it wasn't initialized right | 07:34 |
darwin | installer | 07:34 |
darwin | regardless of monitor (4K, 1600x1200, 1920x1080) | 07:35 |
DashiePie | does Devuan have support for a 6900 XT? | 07:36 |
darwin | Linux kernels have had that for years | 07:37 |
darwin | it works fine with three *BSD UNIXes, Slackware, several other GNU/Linux, Windows, etc. | 07:37 |
DashiePie | what is Chimaera's kernel version, and what kernel version is required for a 6900 XT? | 07:37 |
brocashelm | if debian has support for it, devuan also | 07:37 |
hiddener | gnarface: i think alsactl --no-ucm restore would fix it, but how do I make the thinkpad console audio control work? i've selected it in alsamixer and I see 00 | 07:38 |
brocashelm | 5.10 for main chimaera kernel | 07:38 |
brocashelm | backports is 6.1 | 07:38 |
DashiePie | ah, okay, 6900 XT needs 5.9 | 07:39 |
hiddener | gnarface: doesn't go up or down with arrows or audio buttons | 07:39 |
DashiePie | I just remembered when I ordered a 5900 XT, Beowulf didn't have the required kernel version | 07:39 |
DashiePie | I still have not gotten my parts | 07:39 |
darwin | you mean 5700? | 07:39 |
DashiePie | yeah | 07:39 |
DashiePie | I forgot, since it's been over 2 1/2 years | 07:40 |
brocashelm | it is possible to run a newer kernel on beowulf if you have the right deb files to install with dpkg -i, but this is not supported and you can break your system if not careful | 07:40 |
gnarface | hiddener: try alsamixer -c 1, alsamixer -c 2, etc | 07:40 |
gnarface | hiddener: or try hitting F6 | 07:40 |
gnarface | hiddener: you may need to stop pulseaudio first | 07:40 |
hiddener | gnarface: i have no pulseaudio | 07:41 |
darwin | is there any other option I should try in installer to find out what the heck is going on? | 07:42 |
gnarface | darwin: daedalus test installers are available | 07:43 |
gnarface | there's also live isos | 07:43 |
gnarface | chimaera has a 5.10 kernel in the repo, but perhaps the installer itself is something older | 07:44 |
gnarface | just to be sure, which installer did you actually try? | 07:44 |
darwin | chimaera | 07:45 |
gnarface | no i mean show me the exact download link | 07:45 |
gnarface | so i can sanity check it | 07:45 |
darwin | i got it from devuan.org | 07:45 |
gnarface | how long ago? | 07:46 |
darwin | a few days ago | 07:46 |
gnarface | got time to try the daedalus one? | 07:46 |
gnarface | or one of the live isos either release? | 07:46 |
darwin | i'll get a live ISO | 07:51 |
gnarface | hiddener: the output of "aplay -l" will show you the current default, alsa defaults to whichever is #0 | 07:57 |
gnarface | you may need to set a different default in your alsa config, sometimes #0 is something useless | 07:58 |
gnarface | (like the HDMI jack) | 07:58 |
hiddener | gnarface: yes, i can default it with /etc/asound.conf like i do with the usb speakers | 08:07 |
hiddener | but what exactly does it do, the thinkpad console audio control? do I need to connect a device in order to control it? | 08:09 |
darwin | seems to be no link to the live ISO on devuan.org frontpage... where is it? | 11:17 |
darwin | i found it but didn't want to use it. I wanted the full 4GB+ desktop installer | 11:20 |
gnarface | darwin: you're looking for this: https://files.devuan.org/ | 12:03 |
gnarface | specifically, probably this: https://files.devuan.org/devuan_chimaera/desktop-live/ | 12:03 |
gnarface | or maybe this: https://files.devuan.org/devuan_daedalus/desktop-live/ | 12:04 |
gnarface | this is just for testing the drivers, i'm not telling you to also use the live disk as the installer | 12:04 |
gnarface | actual installer images are up on files.devuan.org too though | 12:04 |
gnarface | if hiddener comes back someone tell them i thought i knew what they were talking about but now realize i did not | 12:05 |
darwin | thanks | 12:47 |
darwin | i will try it tomorrow hopefully | 12:48 |
temp64 | is there a way to change init on a running system? at the moment i'm using this openrc-on-sysvinit thing, i'd like to change it to runit without nuking the whole installation | 17:21 |
rwp | temp64, You should be able to install runit and then reboot. | 18:03 |
rwp | I don't which of the three runit packages is needed to be installed though. | 18:03 |
fsmithred | temp64, install runit-init and optionally install runit-services | 18:06 |
fsmithred | the latter has more run scripts. | 18:07 |
fsmithred | runit will use sysvinit scripts if there are no run scripts for the service. | 18:07 |
rwp | I am looking at openrc and runit-init and can see where runit-init has a conflicts with systemd-sysv, sysvinit-core but not with openrc. | 18:09 |
fsmithred | so one could have both installed and boot one or the other? | 18:10 |
rwp | Hmm... I might guess that migrating that direction might need a manual remove of openrc on the way. "apt-get install runit-init openrc-" or something? I don't know. | 18:10 |
rwp | Or another alternative route from openrc to runit might be openrc to sysvinit and then sysvinit to runit. Since those paths should be well tested. | 18:11 |
fsmithred | oh, when installing openrc, do you still get told to make links yourself? | 18:12 |
fsmithred | and it provided the command to run. | 18:13 |
rwp | Good question. I don't know. I haven't played with openrc. | 18:13 |
fsmithred | afk | 18:14 |
rwp | temp64, You should be able to migrate but if there is a difficulty remember that the netinst iso is a good rescue boot tool which can be used to boot and make a repair if needed. Always good to make sure you have one of those available to boot first, just in case it is needed. | 18:16 |
rwp | Also I get a lot of use out of the refracta-nox live-boot system. | 18:17 |
temp64 | fsmithred, rwp: thanks, suddenly reinstalling with runit doesn't seem like that bad of an idea :P | 18:19 |
temp64 | i have a script that mounts my $HOME as tmpfs, reading from a disk backup on boot and saving it back on shutdown. i'd have to figure out how to make it work after switching | 18:21 |
fsmithred | why would it be different? | 18:24 |
fsmithred | hey, rwp - last week we talked about rsync -X for extended attributes. I use rsync -a and the man page say that -a includes -X. | 18:26 |
fsmithred | gotta go again. bbl. | 18:30 |
temp64 | fsmithred, i thought i had to make sure my $HOME is set up before the display manager gets to it but it probably shouldn't touch my user account unless i actually try to log in | 18:38 |
temp64 | btw, i use rsync -aO <src> <dest> --delete to back up and rsync -aO <src> <dest> to reload tmpfs | 18:39 |
rwp | fsmithred, My man page for rsync 3.2.7-1 says "--archive, -a archive mode is -rlptgoD (no -A,-X,-U,-N,-H)" with a *no -X* there. | 18:50 |
rwp | (It's strange they list what isn't there. Since they list what is there. As they add new options they will need to keep adding to the "no" list. Every versions gets longer there.) | 18:52 |
n4dir | speaking of rsync: i took the example from arch wiki, and it has an option: --exclude{"/dev/*", ...} | 19:18 |
n4dir | what is the purpose of the quotes around the exludes? | 19:18 |
n4dir | in other words: i did echo that command into an executable file, it somehow removed the quotes | 19:19 |
n4dir | now it seems rsync would copy all data again, instead only what has changed | 19:19 |
n4dir | i wonder if that relates to the quotes, or if i screwed it by other means | 19:20 |
rwp | The "*" character is a shell "meta-character" which means the shell (bash, zsh, other) will read it and try to expand it as a file glob. | 19:30 |
rwp | The * character is called a file glob character because it matches a glob of file characters from the directory. | 19:30 |
rwp | Putting quotes around shell meta-characters keeps the shell from interpreting those characters. | 19:30 |
rwp | If one uses echo /dev/* without quotes it will expand the * to match all of the files in /dev but echo "/dev/*" will echo out the literal /dev/* characters without file glob expansion. | 19:32 |
n4dir | interpreting the * as a literal * ? | 19:32 |
rwp | Right. Interpreting * as a literal * and not expanding it. | 19:32 |
n4dir | thanks. For sure not the source of the problem re-rsync'ing the whole shebang. God only knows why that happens | 19:32 |
rwp | Note that --exclude{"foo"} is not a syntax I recognize for rsync. But --exclude="/dev/*" would be. | 19:33 |
n4dir | i copied it from the arch wiki though. Let me check again, perhaps i made a typo when posting it here | 19:34 |
rwp | However I would not suggest --exclude="/dev/*" but perhaps instead --exclude=/dev or better use -x to avoid crossing mount points. | 19:34 |
rwp | URL? So I can look at that page too? | 19:34 |
n4dir | rsync -aAXHv --exclude={"/dev/*","/proc/*","/sys/*","/tmp/*","/run/*","/mnt/*","/media/*","/lost+found"} / /media/disk/ | 19:34 |
n4dir | that is the command, let me quickly look for the URL | 19:34 |
rwp | Okay. That's a better syntax that way. That is using bash curly brace expansion. | 19:35 |
n4dir | https://wiki.archlinux.org/title/Rsync#Full_system_backup | 19:35 |
rwp | The bash shell (and perhaps some others) will see the {a,b,c} part and make that into brace expansion. | 19:35 |
n4dir | not sure if i remember correct, i think i usually just did "rsync -auv ..." and that worked too. No clue what -AXH do | 19:35 |
rwp | Try this: echo /dev/{one,two,three} | 19:35 |
n4dir | yeah, that trick i do remember | 19:36 |
rwp | And see that it will basically be the same as: echo $(for f in one two three; do echo /dev/$f; done) | 19:36 |
rwp | So that --exclude={a,b,c} part is a bash expansion to avoid saying --exclude=/dev/* --exclude=/proc/* --exclude=/sys/* with the --exclude= part repeated for each argument. Instead the bash curly brace expansion does it. | 19:37 |
n4dir | ah, right. thanks | 19:38 |
rwp | Also if the file glob characters like * and ? do not match a file then the characters are passed through verbatim. | 19:38 |
n4dir | so nothing was excluded, but, say, /dev/* literally | 19:38 |
rwp | So... echo ./doesnotexist* will echo out the '*' character because there is no match. | 19:39 |
n4dir | still wouldn't explain why all data gets rsynced again | 19:39 |
rwp | But if I touch ./doesnotexist then it will match and won't. And if I touch ./doesnotexistfoofoofoo then it will echo out both of those. | 19:39 |
rwp | Creating what I would call a data dependent failure. So ALWAYS quote shell meta-characters if they are used in a way that needs to be passed through verbatim. | 19:40 |
n4dir | i usually do, at least i think so, here it got screwed by runing echo " full-rsync-command " >> backup_os | 19:40 |
rwp | Shell under-quoting is one of the most common bugs seen. It's a persistent problem. So everyone is overly sensitive to it. | 19:40 |
n4dir | when i saw all data gets rsynced again, i checked what i did echo into said file, backup_os, then saw the missing quotes | 19:41 |
rwp | So why would rsync copy things twice? Run it with the -n option to see what it would do without actually doing anything. I do that a very lot when testing. | 19:41 |
n4dir | fun fact: i ran the very same command for the other external hard-disk just before that, and there it worked as expected | 19:42 |
n4dir | once this rsync command finished, i will run it again, at least as a dry run | 19:42 |
rwp | Meanwhile... I don't like that suggested rsync command with all of those excludes. I would suggest instead using -x. | 19:42 |
* rwp has phone call... | 19:42 | |
n4dir | bye-bye and thanks | 19:42 |
n4dir | rwp: for later, running it again it works. I screwed it somehow, it seems, but am a step forward in what is not the problem. thanks again | 19:44 |
hacksenwerk | Hi. How can I disable the suspend mode on lid close? I want only the screen to turn off when I close the lid. I tried installing mate power manager to check if it changes some neccessary system files, that stay changed after removing mate power manager again, but tha is not the case... I also treid laptop-mode-tools but they work neither. When I search the web I get flooded with systemdeath tutorials | 19:58 |
hacksenwerk | for that workcase... -.- | 19:59 |
hacksenwerk | Oh and I tried xset -dpms, that works neither | 19:59 |
hacksenwerk | But I woluld like something that works under x _and_ on tty. But if not possible, at least it should work when running the xserver. | 20:00 |
fluffywolf | probably elogind being worse-than-worthless, as usual | 20:00 |
hacksenwerk | fluffywolf: was that for me? is elogind trigger that behaviour? | 20:02 |
fluffywolf | yep. | 20:02 |
hacksenwerk | I can remove it anyway as it seems, cause I donÄt need anything that it offers as far as I understand. | 20:02 |
fluffywolf | stop elogind and see if it stops suspending. if so, either change elogind's config, or, preferably, kill it with fire. | 20:03 |
rwp | Suspend on lid is most likely handled by your Desktop Environment. Check those settings first. Then elogind (logind replacment) does it. Then lastly folks like me with ACPI. | 20:03 |
hacksenwerk | rwp: I have no desktop enviroment. :) | 20:03 |
hacksenwerk | I use a bare xserver with evilwm as window manager and keylaunch as program starter and that's it. | 20:03 |
hacksenwerk | And xloadimager to set a background (fancy stuff though xD) | 20:04 |
hacksenwerk | *xloadimage | 20:04 |
rwp | In that case look for the ACPI configuration in /etc/acpi/lid.sh to see if you are getting suspended that way. | 20:05 |
hacksenwerk | I try removing elogind, it takes 25 other packages with it and it seems that this is stuff I do not need any longer (gvfs, polkit etc.) | 20:05 |
hacksenwerk | rwp: I don't have that file. | 20:05 |
fluffywolf | /etc/init.d/elogind stop, see if it still suspends | 20:05 |
hacksenwerk | I don't even have /etc/acpi | 20:06 |
rwp | DEs are heavy and almost all of them will use (e)logind for control-group functionality. | 20:06 |
hacksenwerk | fluffywolf: rigth good idea before removing :) | 20:06 |
rwp | I would also look in /var/log/syslog very closely to see if it is logging what is triggering the suspend action. | 20:06 |
hacksenwerk | rwp: yeah I used xfce before, so there's is still stuff there after the removal... | 20:06 |
rwp | Looking here I have /etc/elogind/logind.conf and the default listed is HandleLidSwitch=suspend here. So override that action. | 20:07 |
fluffywolf | imho, devuan's default .conf should disable it doing anything except the bare minimum required to make things that depend on it happy, and certainly not having it doing power management unexpectedly. | 20:08 |
hacksenwerk | Yuppie /etc/init.d/elogind stop does what I want! :D | 20:08 |
fluffywolf | figured. now either edit the config to disable suspending, or, as I said, kill it with extreme prejudice. I prefer the latter. | 20:09 |
hacksenwerk | fluffywolf: turning off the screen would be ok for default imho | 20:09 |
fluffywolf | doing nothing should be the default for a program that claims to manage logins. power management is not managing logins. | 20:10 |
hacksenwerk | I will change that config rwp posted and start the service again and then I will remove that packages | 20:10 |
hacksenwerk | fluffywolf: Oh! yeah right, I totally agreee! | 20:10 |
fluffywolf | it's breaking the "do one thing and do it right" rule. | 20:10 |
rwp | The problem is that people switching from windoze will have been trained already that closing the lid allows them to put it in their heat insulated bag and so if it does not suspend by default they will have an overheated machine accidentally. | 20:10 |
hacksenwerk | One programm one task | 20:10 |
hacksenwerk | rwp: lol | 20:10 |
fluffywolf | if there's a need for a program to be installed by default that does power management, then split it into a separate program that only does power management. | 20:11 |
hacksenwerk | I can already hear the windows fanboys screaming: "Linux burned my laptop down!11!" | 20:11 |
scorpion2185[m] | good rule | 20:11 |
rwp | Agreed. But elogind is a compromise replacement for systemd-logind which as we know violates that whole keep it simple philosophy. | 20:11 |
hacksenwerk | rwp: :( bad that we have to do such compromise... | 20:12 |
fluffywolf | I'm not against having a power management daemon installed by default nearly as much as I am against it being hidden in what seems like an entirely unrelated program that causes people to be unable to find how to configure it. this is like the tenth time I've helped someone kill elogind to fix a power management issue. | 20:13 |
rwp | I mostly run a simple window manager and so I don't need (e)logind on my main systems. I have yet to truly understand the DE stand that they *require* the functionality that systemd-logind and D-Bus provides. | 20:14 |
scorpion2185[m] | there is that that other powe manag. daemon acpi(?) | 20:14 |
rwp | Perhaps you are thinking of upowerd? | 20:14 |
fluffywolf | I run a simple window manager too, but if you try using any program starting with a "g" or a "k", or I think even an "x" these days, it becomes a dependency, even if you use no other part of those environments. | 20:15 |
scorpion2185[m] | acpid | 20:15 |
hacksenwerk | Ok I enabled HandleLidSwitchExternalPower=ignore in /etc/elogind/logind.conf and started the service again. Now it only power off the screen. | 20:15 |
fluffywolf | I ended up using equivs to make a fake package to keep apt happy. | 20:15 |
hacksenwerk | But I remove that elogind thing now | 20:15 |
rwp | acpid is a framework daemon for handling acpi events. It's not evil by itself. | 20:15 |
hacksenwerk | fluffywolf: rwp: Thank you very much! :) | 20:16 |
fluffywolf | so far not a single thing has broken, which means it's excessive dependencies. | 20:16 |
rwp | Lot's of hardware things generate acpi events. All of the hardware buttons and switches on the physical system. | 20:16 |
fluffywolf | yes, and some program with "power" or "acpi" in the name should handle them, not some login manager that got dragged in as a dependency. | 20:17 |
rwp | Also virtual machines guests are signaled by the host using acpi events as if it were an actual power button pressed to shutdown, suspend, reboot. VMs use ACPI events too. | 20:17 |
hacksenwerk | Ouch! After removing elogind and the 25 packges it took with it, I ran deborphan to check and I know that deborphan does stupid things sometimes, but it also show that package libelogind-compat and when I want to remove that one it takes a lot of pakacges with it that I want to keep, like cmus, ffmpeg, mpv, the whole xserver even and much more | 20:19 |
fluffywolf | again... I think we should handle them. I don't think elogind should be handling them. | 20:19 |
rwp | If you have a VM that does not respond to host controls to reboot and needs to be forced off then most likely acpid and acpi-support-base is not installed there. For example. It's a good thing to have installed. | 20:19 |
hacksenwerk | That seems like some weird dependency issue... | 20:20 |
rwp | I don't know about deborphan but for apt if those were installed as a dependency and that upstream package is gone then they will still be marked as automatically installed and will be candidates for autoremove. | 20:21 |
rwp | Meaning that you will need to mark them as manually installed for them to remain behind. | 20:21 |
rwp | Either "apt-get install foo" or "apt-mark manual foo". See the apt-mark man page. | 20:21 |
hacksenwerk | rwp: Ok thank you. I will check that tomorrow, I'm very tired now. ^^ | 20:22 |
rwp | I will usually "apt-get autoremove --purge" (--purge because etckeeper rocks!) and then REVIEW THE LIST, SAY NO, and then mark the things I want to keep and repeat until it does not want to remove something I want to keep. | 20:22 |
rwp | bbiab myself | 20:23 |
hacksenwerk | Yeah I use always apt-get autoremove --purge to remove packages | 20:23 |
hacksenwerk | Can I give apt-mark a list? | 20:25 |
hacksenwerk | Yes it does. ^^ | 20:27 |
hacksenwerk | Hmm.. but it still want to remove some packages I want to keep... Anyway I'll come back tomorrow. Good night! ^^ | 20:28 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!