You are not logged in.
I'm really hoping someone can help me.
Yesterday I did an update with `pacman -Syu` and rebooted. When I reboot, it stop doing anything with the screen saying...
/dev/nvme0n1p3: clean, 175050/1638400 files, 2485941/6553600 blocks
And the usually blinking cursor, was not blinking.
How can I undo what has happened and get my computer to boot again?
Offline
Are you using an NVIDIA graphics card? Which packages did get upgraded during the transaction?
Offline
I have no ideas, I never actually watch what gets updated. They are always beyond the scope of my understanding.
I can't load the drive as is, but I am able to load the arch install disc.
Is it possible to undo a pacman update on the hard drive without the OS loading?
EDIT: No, the computer has Ryzen 5 7520U.
Last edited by NeedHelpFast (2024-11-21 20:27:56)
Offline
It is possible to undo the update but first you need to determine what has been updated. chroot in and check /var/log/pacman.log.
I never actually watch what gets updated.
Does that mean you never read the messages relating to manual intervention or failures during during pacman updates?
Last edited by loqs (2024-11-21 20:34:53)
Offline
Does that mean you never read the messages relating to manual intervention or failures during during pacman updates?
I've never seen an error when updating. But if it errored out, it wouldn't have rebooted with the command:
pacman -Syu && reboot
And manual interventions wait for a y/n, don't they?
Offline
The `/var/log/pacman.log` says it updated:
[2024-11-20T15:28:24-0800] [PACMAN] Running 'pacman -Syu'
[2024-11-20T15:28:24-0800] [PACMAN] synchronizing package lists
[2024-11-20T15:28:24-0800] [PACMAN] starting full system upgrade
[2024-11-20T15:28:46-0800] [ALPM] transaction started
[2024-11-20T15:28:47-0800] [ALPM] upgraded chromium (131.0.6778.69-1 -> 131.0.6778.85-1)
[2024-11-20T15:28:47-0800] [ALPM] upgraded exfatprogs (1.2.5-1 -> 1.2.6-1)
[2024-11-20T15:28:47-0800] [ALPM] upgraded iproute2 (6.11.0-1 -> 6.12.0-1)
[2024-11-20T15:28:47-0800] [ALPM] upgraded libinput (1.26.2-1 -> 1.27.0-1)
[2024-11-20T15:28:47-0800] [ALPM] upgraded wireshark-cli (4.4.1-1 -> 4.4.2-1)
[2024-11-20T15:28:47-0800] [ALPM] transaction completed
[2024-11-20T15:28:47-0800] [ALPM] running '20-systemd-sysusers.hook'...
[2024-11-20T15:28:47-0800] [ALPM] running '30-systemd-udev-reload.hook'...
[2024-11-20T15:28:48-0800] [ALPM] running '30-systemd-update.hook'...
[2024-11-20T15:28:48-0800] [ALPM] running 'gtk-update-icon-cache.hook'...
[2024-11-20T15:28:48-0800] [ALPM] running 'update-desktop-database.hook'...
i
Says transaction completed and there is no mention of an error.
EDIT: Add log ending.
Last edited by NeedHelpFast (2024-11-21 21:07:12)
Offline
After chrooting in can you post the journal for system's last boot?
# journalctl -b-1
Offline
After chrooting in can you post the journal for system's last boot?
# journalctl -b-1
I'm still figuring out how to get this text in a file I can move to a working computer, but in the display the last line reads:
systemname kernel: BUG: kernel NULL pointer dereference, address: 0000000000000010
Last edited by NeedHelpFast (2024-11-21 22:24:02)
Offline
You can use a pastebin:
# journalctl -b-1 | curl -F 'file=@-' 0x0.st
Offline
You can use a pastebin:
# journalctl -b-1 | curl -F 'file=@-' 0x0.st
Ty, wasn't easy without your help.
But this is from `cat text.file` as problem computer has no internet.
Offline
Can you as a test drop the ipv6.disable and nowatchdog parameters? I'd not put it past a kernel bug in that region, but if that's all there is it's not completely clear what causes the issue
Offline
Can you as a test drop the ipv6.disable and nowatchdog parameters? I'd not put it past a kernel bug in that region, but if that's all there is it's not completely clear what causes the issue
I can try. Do I just use grub-mkconfig like usual from arch-chroot after editing the '/etc/default/grub` file? Sorry, not too familiar with chroot or the damage it may cause.
Offline
you can just live edit the cmdline while in grub with e https://wiki.archlinux.org/title/Kernel_parameters#GRUB but technically that would work as well if you mounted your partitions correctly under the chroot
Last edited by V1del (2024-11-21 23:14:45)
Offline
No, removing those two parameters didn't change anything. Still same behaviour.
Offline
I assume, with all my data listed and the lack of responses, that I'm screwed with no way back?!?
Offline
no way back is always a bit extreme, if it just happened in the last kernel update you can downgrade the kernel to a known working version from the cache, or the ALA if need be: https://wiki.archlinux.org/title/Downgrading_packages
Offline
Downgrading did nothing to fix this problem. It still freezes in the same place.
No way back means I need to reinstall, losing all my configurations. And if I can't get this to go I probably need a different distro. But up until recently, I have enjoyed this one.
Offline
To which version did you downgrade? Did you also already try the linux-lts package?
Offline
1. Try to install the linux-lts kernel and boot that
2. remove the "quiet" parameter, add "nomodeset earlyprintk=vga,keep systemd.unit=multi-user.target initcall_blacklist=simpledrm_platform_driver_init", https://wiki.archlinux.org/title/Kernel_parameters
See whether either of that lets you boot and/or gets you a visual of the nullptr deref backtrace.
In case of the latter, make a photo, upload it to imgur and link it here (please don't embed it)
On a formal note, please stop bumping and edit your previous posts to mend them if nobody has yet replied.
Offline
To which version did you downgrade?
There wasn't a lot to go on, so I listed the directory based on date, and chose the linux and linux-headers package with the date `Nov 15`. So from 6.11.9 to 6.11.8. Did I not go far enough? I figured any source older than the problem should suffice.
Did you also already try the linux-lts package?
Not yet, first I've heard of this. Will give this a look.
EDIT:
1. Try to install the linux-lts kernel and boot that
2. remove the "quiet" parameter, add "nomodeset earlyprintk=vga,keep systemd.unit=multi-user.target initcall_blacklist=simpledrm_platform_driver_init", https://wiki.archlinux.org/title/Kernel_parameters
Can't install the "lts" because it doesn't seem to list safe step by step instructions on what to do.
Here is the screen shot of the changed parameters booting:
https://imgur.com/a/ToGE2nO
On a formal note, please stop bumping and edit your previous posts to mend them if nobody has yet replied.
I apologize, it's hard to lose two days of productivity because of a software update. Kind of maddening, really.
EDIT 2: img tags aren't working even though it says "(img) tag: on", switched to url tag.
Last edited by NeedHelpFast (2024-11-22 15:46:06)
Offline
Can't install the "lts" because it doesn't seem to list safe step by step instructions on what to do.
You do it the same way you installed and set up the other kernel
Offline
If you blacklist snd_hda_intel and amdgpu and set the target to boot into to multi-user.target does that work. If not does rescue.target work?
Offline
You do it the same way you installed and set up the other kernel
The other kernel was done by changing two packages. This tarball seems to be everything. What do I install?
Offline
If you blacklist snd_hda_intel and amdgpu and set the target to boot into to multi-user.target does that work. If not does rescue.target work?
This got me to a login screen! But what do those parameters actually do? Will I have reduced capabilities? Target: multi-user.target.
EDIT: Oops, sorry about the double post, Seth. Was just so happy.
Last edited by NeedHelpFast (2024-11-22 16:28:49)
Offline
loqs wrote:If you blacklist snd_hda_intel and amdgpu and set the target to boot into to multi-user.target does that work. If not does rescue.target work?
This got me to a login screen! But what do those parameters actually do? Will I have reduced capabilities? Target: multi-user.target.
Swapping to multi-user.target prevents any systemd based display manager from being started which would fail with the amdgpu module blacklisted. You can try only blacklisting snd_hda_intel and see if that is enough to avoid the issue.
Offline