xisop | i3 v4.21.1 seems to not have the defect i reported earlier. | 00:07 |
---|---|---|
xisop | i'm going to re-enable hw accel in firefox | 00:07 |
ravehaver9000 | thanks for the advice with dpms! its worked pretty well, actually | 03:50 |
ravehaver9000 | also, i know this is probably more apt for the dev channel but since i dont have a bouncer rn i wanna ask here: will you add support for s6 soon? i want to try it out here in devuan. currently using sysvinit and i wan tto know if s6 is going to be supported soon alongside sysv, openrc and runit | 03:51 |
gnarface | ravehaver9000: you finally heard? really sorry about that typo, it was late | 04:20 |
ravehaver9000 | nah, i knew from days ago, only issue is that i weas eiother too busy or lazy to do it | 04:21 |
ravehaver9000 | i finally did it earlier today | 04:21 |
gnarface | oh my bad i'm really slipping, this is important, it's dkms not dpms | 04:22 |
gnarface | that was the typo | 04:22 |
ravehaver9000 | dkms?? | 04:22 |
ravehaver9000 | damn | 04:22 |
gnarface | sigh yes | 04:22 |
gnarface | dpms won't do shit | 04:22 |
ravehaver9000 | rthat explains why i still got modproble | 04:22 |
ravehaver9000 | i can just restart dkms with sysvinit right | 04:22 |
ravehaver9000 | or do i have to do anything extra | 04:22 |
gnarface | so here's what you do, you boot back into the new kernel, make sure you have the updated nvidia binary drivers that go with it (we were going with the backports stuff right?) then run dkms then reboot again | 04:23 |
gnarface | i think i have brain damage, i read it as dkms when you typed dpms again just now | 04:24 |
gnarface | it really should have been clear by context because i told you to also get the linux-headers package but i don't blame you | 04:24 |
gnarface | the point of the linux-headers package is you need it for dkms to rebuild the wrapper nvidia puts around their binary blob | 04:24 |
gnarface | i forget for sure if you were also updating to a backports kernel at the same time but assuming you were just make sure the nvidia packages and the linux-headers package all also come from backports | 04:25 |
ravehaver9000 | i have linux-headers, actually. ive always had it | 04:26 |
gnarface | dkms isn't a service, it's just a tool that recompiles the binary blob wrapper | 04:26 |
gnarface | you might also need to install build-essential | 04:26 |
gnarface | if it doesn't work it logs an error... somewhere, i forget where | 04:26 |
gnarface | but you'll see the same error at the terminal if you just run dkms | 04:27 |
ravehaver9000 | i see... build-essential is installed too | 04:27 |
gnarface | were you updating to backports too? | 04:27 |
ravehaver9000 | nah | 04:27 |
gnarface | or was that someone else? | 04:27 |
ravehaver9000 | 6.0.0-5 to 6.0.0-4 | 04:27 |
gnarface | alright, just run dkms right now and see what it says | 04:27 |
onefang | dkms gets run automatically by apt when needed, and complains during that process if it doesn't work. Been there, seen that. Think it was also for nVidia blob, but that was years ago. I changed to a pure AMD system over two years ago. | 04:28 |
gnarface | yea it should but if you're installing a bunch of stuff it might scroll by unnoticed | 04:28 |
onefang | That'll by why I scroll back and read it all. | 04:29 |
gnarface | and if you're missing linux-headers it'll definitely fail | 04:29 |
gnarface | or if the linux-headers package you have doesn't match | 04:29 |
ravehaver9000 | yeah i can very much see why. honestly if i had anticipated id go from full windows to full linux within 6 months of getting my current pc id have gone fuill amd instead | 04:29 |
ravehaver9000 | welp, at least my cpu is ryzen | 04:29 |
gnarface | or if the nvidia drivers packages you have are too old or new for the kernel it might also fail | 04:29 |
gnarface | or if you're missing some other build deps | 04:29 |
gnarface | yea, you don't have to mess with any of this stuff for AMD cards, it's much better integrated | 04:30 |
ravehaver9000 | yeah, and i could even try out wayland too | 04:30 |
ravehaver9000 | at most id need nvidia for stuff like gpt ai writing and i havent done that in a while | 04:30 |
gnarface | anyway, i'm stepping away again but let me know if that works, i'll read the scrollback | 04:30 |
ravehaver9000 | most of my gaming is emulation, thjis is why i see myself having no problem running daedalus even when it becomes oldoldstable as my daily driver instead of fucking off back to arch or some other gaming distro | 04:31 |
gnarface | incidentally, you don't need dpms for anything but screen blanking and you shouldn't have had to do anything special to get it, but make sure you have acpid, it's important for nvidia power management in linux and there's some weird cases you might be missing it | 04:31 |
ravehaver9000 | yeah just checked for acpid | 04:33 |
ravehaver9000 | i use sysvinit and acoid tries to install runit dpeendencies (runit-helper, namely) | 04:33 |
ravehaver9000 | do i just go with it or | 04:35 |
ravehaver9000 | in order to install acpid, which i dont have | 04:35 |
gnarface | ravehaver9000: how about dkms, what happened? acpid shouldn't need runit deps unless you already have part of runit i think... make sure you have these packages: sysv-rc, sysvinit, sysvinit-core, sysvinit-utils | 05:05 |
gnarface | i'm not sure what runit packages you might have but there must be something in as an alternate | 05:06 |
ravehaver9000 | just cheked for hte sysvinit packes | 05:08 |
ravehaver9000 | i have all four of them | 05:08 |
gnarface | hmm, anything with runit in the name? | 05:08 |
ravehaver9000 | none | 05:08 |
gnarface | hmm, it's gotta be something though | 05:09 |
gnarface | oh, and about the nvidia packges, make sure they're all the same version, sometimes previous versions can get left in there and have to be manually updated separately, they're not good at assigning dependencies like they're supposed to | 05:09 |
ravehaver9000 | all nvidia prackages i have are 510.85.02-6, except for nvidia0-legacy-check, which is at 515.48.07-1 instead | 05:11 |
gnarface | hmm | 05:11 |
ravehaver9000 | those with dates are all 20220217+1 | 05:12 |
gnarface | the ones with odd formats, just make sure they're the latest | 05:12 |
gnarface | but sometimes you have to prune the others, especially nvidia-persistenced | 05:12 |
gnarface | is this a testing/daedalus system? | 05:14 |
gnarface | is that how you got kernel 6.x? | 05:14 |
ravehaver9000 | eyah | 05:15 |
ravehaver9000 | and speaking of daedalus | 05:15 |
gnarface | there should be no 515.48.* anywhere | 05:15 |
ravehaver9000 | i wonder whats the naming theme for devuan versions | 05:15 |
gnarface | they're planets | 05:15 |
ravehaver9000 | damn really | 05:15 |
ravehaver9000 | i thought it was like, literature | 05:15 |
gnarface | well they're not planets picked at random | 05:16 |
gnarface | and they're being chosen alphabetically | 05:16 |
ravehaver9000 | yep i def noticed the alphabetical order too | 05:16 |
gnarface | there might be some other whimsical criteria but that might just be from the people who named the planets | 05:16 |
ravehaver9000 | yeah | 05:16 |
gnarface | https://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=%5Envidia-legacy-check&x=submit | 05:17 |
ravehaver9000 | cant believe theres a planet named ascii | 05:17 |
gnarface | https://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=%5Envidia-driver.*&x=submit | 05:17 |
ravehaver9000 | will devuan 6 be caleld earth | 05:17 |
gnarface | heh, no idea | 05:17 |
ravehaver9000 | yeah same | 05:17 |
ravehaver9000 | itd be funny if it was called earth tho | 05:18 |
gnarface | anyway, according to my initial searches it looks like these should all be 510.85.02-6 | 05:18 |
gnarface | maybe part of your issue is they put the wrong versions in then pulled them | 05:18 |
ravehaver9000 | yeah | 05:18 |
gnarface | it can happen in testing and unstable | 05:18 |
ravehaver9000 | that seems legit, itd have to then remove/purge the current version i have, then reinstall the corect one right | 05:19 |
gnarface | yea, and make sure dkms succeeds this time | 05:19 |
gnarface | then reboot and it should be good | 05:19 |
ravehaver9000 | ty a lot | 05:19 |
gnarface | good luck, i'll be back again later | 05:19 |
ravehaver9000 | ty | 05:19 |
ravehaver9000 | im honestly unironically planning to keep using daedalus even once its no longer even the curernt stable version | 05:20 |
ravehaver9000 | prob because i have strong opinions on the "rust in kernel" issue but i also feel like anything more thsat will be added in future versions of linux iwll be stuff i might not need anymore | 05:20 |
ravehaver9000 | except for maybe better nvidia support on wayland, but xorg does everything i want already | 05:21 |
ravehaver9000 | just restarted after changing the older nvidia-legacy-modules and rebuilding the module with dkms | 05:36 |
ravehaver9000 | i still get the "error running install command 'modproble -i nvidia-current' for module nvidia: retcode 1" error, but it also seems to have gotten rid of many other erros based on just how long it took to go from there to the login screen | 05:38 |
ravehaver9000 | of lightdm | 05:38 |
FatPhil | what's the best way of finding memory hogs if you've got so little ram you can't even run top ("-bash: fork: Cannot allocate memory | 10:41 |
Deknos | there are enough parameters for ps to look into it.. getting them right is the challenge :D | 10:43 |
FatPhil | I can't ssh in to that machine, and my g/f "Can't kill anything, can't su" | 10:45 |
FatPhil | so probably a ps will fail too | 10:45 |
FatPhil | wondering why OOMkiller hasn't kicked in, maybe I need to look into some better settings for that. | 10:47 |
FatPhil | can I read/display a file only using shell builtins? | 11:01 |
FatPhil | Oooh, this might be the start of a builtins-only ps: for i in /proc/[1-9]*; do read x < $i/comm; echo "$i: $x"; done | 11:09 |
FatPhil | however, even once I found a browser, I'm not sure I'd know how to kill it without forking a kill-alike. | 11:09 |
onefang | Reboot? | 11:10 |
FatPhil | bash at least has kill as a builtin | 11:28 |
FatPhil | but yes, the big red switch was the magical process-clearing technique we used. | 11:28 |
unclouded | /var/lib/urandom/random-seed did get generated automatically within this debootrap system, probably on boot | 20:27 |
unclouded | /etc/kernel-img.conf and /etc/initramfs-tools/conf.d/resume are still missing, with no ill effects | 20:28 |
gnarface | cool | 20:30 |
gnarface | good to know | 20:30 |
bb|hcb | unclouded: See /etc/init.d/urandom - the file is generated on reboot/halt and used to re-seed the urandom at boot time; from what I see it is safe to assume that file to be missing after install | 20:56 |
unclouded | thanks for the pointer to the script, I see it now | 21:47 |
gnarface | ah i should have guessed that, forgot about it | 21:49 |
gnarface | most the stuff you see done mysteriously on boot is gonna be from something in /etc/init.d/ | 21:49 |
gnarface | either directly or indirectly | 21:49 |
unclouded | I feel like a dolt for not thinking to grep -r | 21:50 |
unclouded | do_bootloader and do_initrd don't seem to be documented in kernel-img.conf(5) | 21:59 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!