You are not logged in.
Hi,
My PC uses a Ryzen 3 1300X CPU running on a Gigabyte AB350M-D3V MB
The GPU is a Gigabyte NVIDIA GeForce GT 1030
I'm using Grub as the boot loader
I've had boot issues since upgrading from linux-6.2.2.arch2-1, with the system hanging after displaying:
Booting `Arch Linux'
Loading Linux linux ...
Loading initial ramdisk ...
I've tried upgrading to linux-6.2.6-arch1-1 and linux-6.2.8.arch1-1, both show the same error. Downgrading to linux-6.2.2.arch2-1 allows the boot-up to proceed as normal.
Other things I've tried:
Booting from the fallback image
Booting with and without "nvidia-drm.modeset=1"
It's not a new system and was built when /boot/efi was a suggested EFI system partition mount point and was wondering if this might be part of the problem, although Grub should be okay with it afaik.
Any advice will be gratefully received.
Thanks,
Rob
***
Solved - This is fixed in 6.2.11-arch1-1
***
Last edited by Painless (2023-04-15 13:29:30)
Offline
This very strongly looks like the "usual" issue of having a deviating /boot setup during installation and actual system runtime
After updating what is your output for
lsblk -f
cat /etc/fstab
cat /boot/grub/grub.cfg
file /boot/vmlinuz-linux
sudo lsinitcpio /boot/initramfs-linux.img | grep nvidia
Online
Thanks for getting back. I have posted the output after update/before reboot to https://pastebin.com/WEt29t7H
Offline
Looks fine, what are you expecting to boot into? If some graphical session/display manager, do you get to a terminal with the
systemd.unit=multi-user.target
parameter?
Online
I would expect it to boot into LightDM.
Unfortunately it just hangs with a flashing cursor after printing:
Loading initial ramdisk ...
It seems loading of initramfs does not complete properly.
Keyboard and mouse are both unlit/unresponsive so I can't check virtual consoles.
Offline
I think I may have a similar issue
Edit solved it reinstalling kernel and mkinitcpio
Last edited by afader (2023-03-25 03:48:03)
Offline
Can you boot "systemd.unit=multi-user.target nomodeset"?
Can you ping the non-booting system?
Offline
I think I may have a similar issue
Edit solved it reinstalling kernel and mkinitcpio
Unfortunately this didn't help. I also re-installed grub, re-ran grub-install and grub-mkconfig.
Can you boot "systemd.unit=multi-user.target nomodeset"?
Can you ping the non-booting system?
The PC hangs at the same point and cannot be pinged.
Offline
If you install linux-lts in addion to the linux package, can you boot into linux-lts? If you use the ALA to obtain the linux package releases between 6.2.2.arch2-1 and 6.2.8.arch1-1 do they all have the issue?
Offline
I just tried completely uninstalling nvidia (plus utils/settings/lib32), then installing linux-6.2.8.arch1 ... still hangs as before.
If you use the ALA to obtain the linux package releases between 6.2.2.arch2-1 and 6.2.8.arch1-1 do they all have the issue?
linux-6.2.2.arch2-1 - boots
linux-6.2.6.arch1-1 - hangs
linux-6.2.8.arch1-1 - hangs
I'll try linux-lts next.
Offline
If you install linux-lts in addion to the linux package, can you boot into linux-lts?
Unfortunately this hangs as before, same goes for linux-lts fallback.
I have also tried changing mkinitcpio compression to gzip, but no joy.
I'm very confused about this.
Offline
What about linux-6.2.1.arch1-1 or linux-6.2.arch1-1? If you change the systemd target to boot into to rescue.target or emergency.target does the system still hang?
Offline
linux-6.2.arch1-1 - boots
linux-6.2.1.arch1-1 - boots
linux-6.2.8.arch1-1
rescue.target - hangs
emergency.target - hangs
Offline
Something introduced between 6.2.2 and 6.2.6 appears to have introduced a regression. Apart from narrowing down that range I would suggest trying the kernel parameters iommu=off acpi=off nomodeset.
Offline
'acpi=off' allows the system to boot but with CPU tsc errors and without LightDM loading
'iommu=off' allows the system to boot, seems normal. I'll have to do some research on iommu.
Thanks
Offline
Do you get away w/ "iommu=soft"?
Offline
Do you get away w/ "iommu=soft"?
I'm afraid it hangs.
I've also tried 'amd_iommu=fullflush' which hangs and 'amd_iommu=off' which boots.
Offline
@seth any thoughts on what could have introduced in it linux-6.2.3.arch1-1 to avoid a bisection?
Offline
The same thing happened to me (AMD Ryzen 1500X, Gigabyte AB350-Gaming 3), only by disabling the IOMMU could it boot (or downgrading). But after enabling SVM in BIOS (M.I.T. --> Advanced Frequency Settings --> Advanced CPU Settings --> SVM Mode: Enabled), it has started to boot fine.
Offline
Given #15 & #17: https://github.com/gregkh/linux/commits … vers/iommu - the Mar 10, 2023 commits, particualirly the amd related ones. Then the global ones.
The SNP one kinda looks at me …
Edit: even more after the last post…
Last edited by seth (2023-03-29 21:45:03)
Offline
If you blacklist kvm and kvm-amd can you boot with SVM disabled?
module_blacklist=kvm,kvm-amd
Last edited by loqs (2023-03-29 22:02:43)
Offline
If you blacklist kvm and kvm-amd can you boot with SVM disabled?
module_blacklist=kvm,kvm-amd
It hangs like before.
For me it only works if I boot:
- With kernel "linux-6.2.2.arch2-1-x86_64" or "linux-lts-6.1.15-1-x86_64" (or earlier).
- IOMMU disabled, either from BIOS or with kernel parameter.
- SVM mode enabled in BIOS.
Last edited by ote, that in signing up t (2023-03-29 23:56:37)
Offline
$ git bisect start
status: waiting for both good and bad commits
$ git bisect bad v6.2.3
status: waiting for good commit(s), bad commit known
$ git bisect good v6.2.2
Bisecting: 498 revisions left to test after this (roughly 9 steps)
[065a789b413ab69b6a67c6af4970d997ac50c96c] coresight: etm4x: Fix accesses to TRCSEQRSTEVR and TRCSEQSTR
https://drive.google.com/file/d/1REWFPz … share_link linux-6.2.2.r499.g065a789b413a-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/18HzURF … share_link linux-headers-6.2.2.r499.g065a789b413a-1-x86_64.pkg.tar.zst
Edit:
Assuming the last kernel was good
$ git bisect good
Bisecting: 249 revisions left to test after this (roughly 8 steps)
[17e3c7f2b576bcca0a973a8f506f9e598d4fb704] ipmi_ssif: Rename idle state and check
https://drive.google.com/file/d/1BQRQuB … share_link linux-6.2.2.r748.g17e3c7f2b576-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1qFyHHK … share_link linux-headers-6.2.2.r748.g17e3c7f2b576-1-x86_64.pkg.tar.zst
Edit2:
Assuming the last kernel was good
$ git bisect good
Bisecting: 124 revisions left to test after this (roughly 7 steps)
[9930b568dccf197a2cd32183905ebe4df195dff9] selftests: vm: Fix incorrect kernel headers search path
https://drive.google.com/file/d/19xYmcc … share_link linux-6.2.2.r873.g9930b568dccf-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1IU3HjR … share_link linux-headers-6.2.2.r873.g9930b568dccf-1-x86_64.pkg.tar.zst
Last edited by loqs (2023-03-30 00:51:36)
Offline
If you blacklist kvm and kvm-amd can you boot with SVM disabled?
module_blacklist=kvm,kvm-amd
Likewise, hangs for me
Offline
Unfortunately the first kernel - linux-6.2.2.r499.g065a789b413a-1-x86_64.pkg.tar.zst - hangs
Offline