| freaxeh | "ACPI Error: AE_NOT_FOUND, while resolving a named reference package element" | 11:00 |
|---|---|---|
| freaxeh | any hints? | 11:00 |
| djph | Usually either a bios in need of update; or ignorable (although "something" with ACPI won't work) | 11:04 |
| freaxeh | right | 11:04 |
| gnarface | freaxeh: yea, bios issue... i thought i remember there was some way to insert the missing data into acpi directly, but you would also have to locate the missing data somehow. in a fair world, it'd be fixable with a bios update, though usually it won't be ever fixed and often it's completely unimportant | 11:30 |
| buZz | often you can spoof 'the OS' to acpi too | 12:12 |
| buZz | manufacturers dont really test the 'linux' acpi often | 12:12 |
| freaxeh | thats good that its a non issue because the bios is 14 years old and came out in 2010 and its the latest bios for this motherboard :D | 13:52 |
| buZz | hehe, well , many ACPI tables in bios are jsut faulty ;) | 14:13 |
| buZz | its often not a big deal if you get such ACPI errors | 14:14 |
| buZz | or are you actually having issues, beside 'there is a message' ? | 14:14 |
| freaxeh | nope no issues just the message, I am the voice of the kernel! | 14:15 |
| freaxeh | not really... | 14:16 |
| buZz | its safe to ignore ;) | 14:16 |
| freaxeh | yep :) | 14:16 |
| drdozer | HI - I just updated my devuan and the nvidia driver failed to compile. | 16:53 |
| drdozer | It is apparently a now-fixed source incompattibility between the kernel version and driver source | 16:54 |
| drdozer | but I can't see how to encourage apt to use a newer version of the driver :( | 16:54 |
| drdozer | I am on daedolas and using the nvidia driver PPA | 16:55 |
| drdozer | deb https://ppa.launchpadcontent.net/graphics-drivers/ppa/ubuntu/ noble main | 16:55 |
| drdozer | Linux taiwaif 6.9.7+bpo-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.9.7-1~bpo12+1 (2024-07-03) x86_64 GNU/Linux | 16:56 |
| drdozer | I'm not sure what I should try -- I could perhaps stop the kernel updating to 6.10.*? | 16:59 |
| gnarface | drdozer: you should absolutely never have mixed ubuntu packages into your install | 17:01 |
| drdozer | gnarface, ok -- I can remove that ppa and update | 17:02 |
| gnarface | maybe | 17:02 |
| gnarface | you might have a mess on your hands | 17:02 |
| gnarface | the current stable devuan versions had that problem recently too, but i think it's been fixed | 17:03 |
| gnarface | which version of devuan are you using? daedalus? | 17:03 |
| drdozer | yes | 17:03 |
| drdozer | I think a fix for this has been pushed to debian, but not to devuan | 17:04 |
| gnarface | shouldn't be any difference | 17:04 |
| drdozer | It is this issue: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1077841 | 17:05 |
| gnarface | the problem is, ubuntu's versions are higher, so the package manager will think you're already up-to-date, so you might have to manually downgrade packages or remove them first | 17:05 |
| gnarface | yea, that's not the stable kernel either. looks like you're also using a backports kernel? did you get that kernel from the ubuntu backports or the debian backports? | 17:06 |
| gnarface | there are devuan backports repos you should use | 17:06 |
| gnarface | (only if you need them) | 17:06 |
| drdozer | My apt sources list is: | 17:07 |
| drdozer | deb http://mirrors.dotsrc.org/devuan/merged daedalus main non-free-firmware non-free contrib | 17:07 |
| drdozer | deb http://mirrors.dotsrc.org/devuan/merged daedalus-updates main non-free contrib | 17:07 |
| drdozer | deb http://mirrors.dotsrc.org/devuan/merged daedalus-security main non-free contrib | 17:07 |
| drdozer | deb http://mirrors.dotsrc.org/devuan/merged daedalus-backports main non-free contrib | 17:07 |
| gnarface | ah, that's also an issue | 17:07 |
| gnarface | i'm pretty sure you need "non-free-firmware" on all lines | 17:07 |
| drdozer | I have added non-free-firmware and tried again | 17:09 |
| gnarface | but for it not to ignore the daedalus-backports line, you might also have to add "-t daedalus-backports" to your apt-get command-line | 17:09 |
| gnarface | or apt-cache | 17:09 |
| gnarface | try: apt-cache -t daedalus-backports search ^nvidia | 17:09 |
| gnarface | or: apt-cache -t daedalus-backports search nvidia | 17:09 |
| gnarface | also: apt-cache -t daedalus-backports search linux-image | 17:10 |
| gnarface | (make sure you ran "apt-get update" once after changing your sources.list) | 17:10 |
| drdozer | should I see a difference with and without the -t daedalus-backports? | 17:11 |
| gnarface | uh... i think so? | 17:11 |
| gnarface | lemme make sure that's a official mirror... | 17:11 |
| gnarface | hmm, yes, appears to be | 17:12 |
| drdozer | diffing the nvidia searches says they are identical | 17:12 |
| gnarface | the difference would just be in version numbers | 17:12 |
| gnarface | what version to you see? | 17:12 |
| drdozer | the linux-image searches are different | 17:13 |
| gnarface | aah, hmm... maybe there are no backports nvidia drivers right at the moment then | 17:13 |
| gnarface | it doesn't always happen right away | 17:13 |
| gnarface | usually you're better off not having them though | 17:13 |
| drdozer | understood | 17:14 |
| gnarface | did you have a particular reason to go up beyond the daedalus version other than the build error? | 17:14 |
| gnarface | because it should be fixed by now but that unfortunately means downgrading a bunch of stuff, which can be a pain in the ass... | 17:14 |
| drdozer | just the build error -- I now have a half-built 6.10 kernel which worries me a bit | 17:15 |
| gnarface | if 6.10 has a similar issue that's probably easiest avoided by just going back to the current stable kernel, which should be 6.1.0 something | 17:15 |
| gnarface | you didn't have a reason to upgrade to 6.10 did you? | 17:16 |
| gnarface | if you did, at least use the devuan-backports one instead of the ubuntu one | 17:16 |
| gnarface | but it seems like it's not ready for nvidia yet | 17:16 |
| drdozer | 6.9 works fine -- it failed when trying to build the nvidia driver against 6.10 | 17:16 |
| gnarface | yea, that's bad news | 17:16 |
| gnarface | current stable should be 6.1.0-something | 17:17 |
| gnarface | i think | 17:17 |
| drdozer | yeah, the internet says 6.1.* | 17:17 |
| gnarface | but that one should work fine with the current stable nvidia driver versions, which should be 535.183.01-1~deb12u1 | 17:17 |
| drdozer | is there a magic incantation to tell apt to roll back the kernel version? | 17:18 |
| gnarface | it might not have actually even uninstalled the last one | 17:18 |
| gnarface | you might just be able to reboot and select it at the grub prompt | 17:19 |
| gnarface | dpkg -l |grep linux-image | 17:19 |
| gnarface | should show you which ones are installed | 17:19 |
| gnarface | assuming the right one is still there, just reboot to it, then uninstall the 6.10 one | 17:19 |
| gnarface | (then re-run "update-grub" to flush it out of your boot menu) | 17:20 |
| gnarface | don't forget your corresponding linux-headers-* packages | 17:21 |
| drdozer | it is indeed still there - so can I just do "apt remove linux-image-6.10.6-1~bpo12+1" and cross my fingers? | 17:22 |
| gnarface | uh, don't do it while you're running it. reboot to the old one first | 17:22 |
| gnarface | then no finger-crossing should be necessary | 17:22 |
| drdozer | I'm booted into 6.9 as the update never fully completed | 17:22 |
| gnarface | are you sure? that's what "uname -a" says? | 17:22 |
| drdozer | Linux taiwaif 6.9.7+bpo-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.9.7-1~bpo12+1 (2024-07-03) x86_64 GNU/Linux | 17:23 |
| gnarface | alright, if that one is working for you then i can't argue with it | 17:23 |
| drdozer | Ah, it looks like the package is calling itself linux-image-amd64, which is uselessly generic | 17:23 |
| gnarface | so that's not an actual kernel, that's just a meta-package that always refers to the current stable one | 17:25 |
| gnarface | if you have it installed, a regular "apt-get upgrade" will always pull in the newest kernel stable kernel, but note that it's still older than the 6.9 backports one you're using, so that will be irrelevant | 17:26 |
| drdozer | figured it out: sudo apt remove linux-image-6.10.6+bpo-amd64 | 17:26 |
| gnarface | oh, sorry, forgot to tell you the important part | 17:27 |
| drdozer | update-grub has done the right thing, and not added an entry for 6.10, so I think I'm good to go | 17:29 |
| drdozer | unless there's something else | 17:29 |
| gnarface | nothing important that i can think of. only other thing is, don't forget to remove linux-headers-6.10.6+bpo-amd64 and linux-headers-6.10.6+bpo-common too | 17:30 |
| gnarface | but they should be harmless, merely wasting some disk space | 17:30 |
| gnarface | and make sure "dpkg -l |grep nvidia" turns up a list of all the same versions for the ones that are formatted like 5**.***.**-... | 17:31 |
| drdozer | great, and an autoremove has got rid of some other junk | 17:31 |
| gnarface | (one of the big pitfalls of using the nvidia drivers is accidentally mixing adjacent versions, which sometimes even mostly works, but mostly working is the worst type of working, usually) | 17:32 |
| drdozer | yeah, apparently I'm 535 | 17:32 |
| drdozer | thanks - you're a life saver | 17:32 |
| gnarface | some of them have weird serial number or datestamp versions, you can't really do much but make sure they're the latest, but where it's bad is if for example you have a few 565.** or whatever ones mixed in with the 535.** ones | 17:33 |
| gnarface | no problem | 17:33 |
| plasma41 | freaxeh: For anything ACPI-related, I'd say your best bet is to ask in #osdev | 20:28 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!