[Solved]Blackscreen after upgrade via chroot. kernel 6.12.30-1

rebornKidz · 2025-10-30 06:48:13

Assalamualaikum ٱلسَّلَامُ عَلَيْكُمْ

I try to get into another/ secondary tty, but it seems that passing the nomodeset does nothing.

linux /boot/vmlinuz-linux root=UUID=0a3407de-014b-458b-b5c1-848e92a327a3 rw nomodeset
linux /boot/vmlinuz-linux root=UUID=0a3407de-014b-458b-b5c1-848e92a327a3 rw quite nomodeset

I just changed the order, hoping it works.

I am using pc not a laptop, Naively, it is possible that config keyboard might cause an issue? I try

Keyboard shortcuts: Try the correct key combinations for your hardware. On most laptops, you may need to use Fn + Alt + F2 instead of Ctrl + Alt + F2 to switch to a TTY.

But still it black screen and I am unable to spawn another tty.

My last arch setup was install the kde plasma, which already. Before this chaos. I stuck on enabling the network, and already asked on another site

uname -r
6.12.30-1-lts

I can't remember correctly which kernel which works fine last time i login.
Just yesterday i chroot using systemrescue from usb and do the "pacman -Syyu".
What i notice was some cached header file missing if i want to downgrade to that version.

here is the cached package

Last edited by rebornKidz (2026-01-16 23:16:57)

rebornKidz · 2025-11-15 13:13:12

Somehow i so desperate to solve this.
I just chroot using systemrescue, download the latest kernel and kernel header. but it does not fix anything. Should i trigger the grub config update manually?
Actualy my kde with old kernel still bootable, just no internet connection.

when i try to toubleshoot the network i got new error

lspci: Unable to load libkmod resources: error -2

seth · 2025-11-15 14:54:44

https://wiki.archlinux.org/title/Genera … l_messages
And can you boot rescue or multi-user.target, https://wiki.archlinux.org/title/System … _boot_into ?

But still it black screen and I am unable to spawn another tty.

So where did you run

uname -r
6.12.30-1-lts

?

rebornKidz · 2025-11-15 17:38:28

Thanks Seth, if i recall properly, it's after i do chrooted.

By using multi-user, it's a try to get into the tty?
Will try to get the printing mode enabled.

Already try add `debug` and `rescue` but still got blackscreen, downgrading the kernel also has no effect.
Should i remove the kernel param which set lvl to 3 ?

Last edited by rebornKidz (2025-11-16 12:45:34)

seth · 2025-11-15 22:01:15

uname from a chroot will just show the kernel of the system from where you chrooted - it's meaningless for these purposes.

"debug ignore_loglevel earlyprintk=vga,keep nomodeset systemd.unit=multi-user.target" should do.

rebornKidz · 2025-11-16 13:29:08

Sorry for the meaninglessness, i already did these param.

The main tty is printing arch 6.15.3-arch1-1

Main tty

try to spawn another tty, so can gather necessary information but

tty2 stuck

I am unable to spawn tty9

I forgot to mention that, perhaps after do upgrade via chrooted I can reset the boot with ctrl + alt delete. on advanced linux boot with old kernel, it is still bootable with KDE Plasma dekstop, but I have trouble with the networking.

Is using this old kernel have easier way to fix my arch?

Last edited by rebornKidz (2025-11-16 18:37:47)

seth · 2025-11-16 13:38:02

6.15.3 is from Jun 19, 2025 and was outdated a week later w/ 6.15.4
Try to install that kernel from the chroot, does the system boot afterwards?

pacman -U https://archive.archlinux.org/packages/l/linux/linux-6.15.3.arch1-1-x86_64.pkg.tar.zst

If so, please post the outputs of "lsblk -f" and "cat /proc/cmdline"
Also make sure to remove the quiet marameter next to adding the other ones.

rebornKidz · 2025-11-16 17:10:33

Thanks for the reminder about removing 'quiet', will do what you ask.
But I only need to reinstall the Linux package without the Linux header?

What I do is edit the grub directly with `e` mode then directy boot it use ctrl + x

lsblk -f

sda                                                                                        
|-sda1      vfat   FAT32               30C5-ACC5                                           
|-sda2                                                                                     
|-sda3      ntfs         System        4A60D16C60D15F6F                                    
`-sda4      ntfs                       DAA40FCEA40FAC57                                    
nvme0n1                                                                                    
|-nvme0n1p1 ext4   1.0   /             759f6fc2-0535-4342-9f93-5f6bb1486fb3   34.9G    38% /home
|                                                                                          /
|-nvme0n1p2 ntfs         /cross-system 68DFD0FE55198462                      100.7G    51% /run/media/plasma/_cross-system
|-nvme0n1p3 ext4   1.0   /arch-media   1256b93c-8d3d-44cd-b630-13dbf8a0bcbb                
`-nvme0n1p4 vfat   FAT32 ESP           0F5A-6664                             764.7M    25% /efi

cat proc/cmdline

BOOT_IMAGE=/vmlinuz-linux root=UUID=759f6fc2-0535-4342-9f93-5f6bb1486fb3 rw loglevel=3 quiet

Last edited by rebornKidz (2025-11-16 18:35:21)

seth · 2025-11-16 17:14:24

But I only need to reinstall the Linux package without the Linux header?

https://archlinux.org/packages/core/x86 … i-headers/ will only become relevant once you start to compile userspace stuff (and typically isn't that update-sensitive) and https://archlinux.org/packages/core/x86 … x-headers/ is generally only required for OOT packages (dkms or otherwise)
Just to get (probably) installed and booting header aligned doesn't require changes to either and if you get the system to boot this way, we'll fix it from there anyway.

rebornKidz · 2025-11-16 18:57:33

Ahhh, I guess it wrongly I thought I needed the header because plasma needs it. I think stupidly that it was what caused the problem.
Installing from the pacman cache linux-6.15.3 already fixes the boot problem.
Thanks for helping me and being kind Seth.

Before I mark the question as resolved, would you mind explaining what actually happened?
I mean my wrong move, things I should avoid in the future.

seth · 2025-11-16 19:04:02

Most common and likely scenario is a disparity between booting and installed kernel because you're not installing the kernel into the boot location.

lsblk -f
cat /proc/cmdline
cat /etc/fstab

rebornKidz · 2025-11-16 21:52:08

Yes, I misunderstood about the boot process, and addressed here.
I commented out the boot partition at that time.

Last edited by rebornKidz (2025-11-16 21:53:07)

seth · 2025-11-16 22:16:41

Can you please post the outputs of those three commands?

rebornKidz · 2025-11-16 22:55:56

here is the fstab

# /dev/nvme0n1p1 LABEL=/
UUID=759f6fc2-0535-4342-9f93-5f6bb1486fb3	/         	ext4      	rw,relatime	0 1
# /dev/sda1
# UUID=30C5-ACC5      	/win-boot 	vfat      	rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro	0 2

# /dev/nvme0n1p1 LABEL=/
UUID=759f6fc2-0535-4342-9f93-5f6bb1486fb3	/         	ext4      	rw,relatime	0 1

# /dev/nvme0n1p4 LABEL=ESP
# UUID=0F5A-6664      	/boot     	vfat      	rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro	0 2
# /dev/nvme0n1p1 LABEL=/
UUID=759f6fc2-0535-4342-9f93-5f6bb1486fb3	/home     	ext4      	rw,nodev,relatime	0 2

# /dev/sda1
# UUID=30C5-ACC5      	/win-boot 	vfat      	rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro	0 2

and i do `pacman -Syu` and it breaks again. Will change with boot enabled and try again.

the other 2 remain unchanged. But I will double check soon.
let me know if it wrong or stupid

seth · 2025-11-16 23:05:16

Ah, sorry - missed your edit.

You#re mounting 759f6fc2-0535-4342-9f93-5f6bb1486fb3 (a 60GB partition) twice as root (which is wrong) and then as /home (what's likely not what you want?) and you're booting from a boot partition which is either 30C5-ACC5 or 0F5A-6664

ls /efi

And what's the plan about the home partition?

rebornKidz · 2025-11-17 02:18:30

ls /efi via chrooted

no file

the /boot contain 3 files.

To be sure about either 30C5-ACC5 or 0F5A-6664 , can I verify it by the loaded current grub menu?

Sorry at this time not yet double check `/proc/cmdline` , not do any formating partition `lsblk -f` should be the same.

Ah this is my sin,

nvme0n1p3   /arch-media   1256b93c-8d3d-44cd-b630-13dbf8a0bcbb

It should be the home. Thanks for help me realize that. I confirm the partition is fully unused, and only contains lost+found

Back then I just backed up the old fstab to give context on what had happened previously
But now, plasma and some installed packages are already made. Can i just magically do `mv` and trigger mkinitcpio, grub?
I am expecting the newest kernel if it is possible and not too complicated. The saved boot option, which uses linux-6.15.3 cannot start the NetworkManager service.

Last edited by rebornKidz (2025-11-17 03:24:06)

seth · 2025-11-17 09:18:46

ls /efi [quote]via chrooted[/quote]
no file

the /boot contain 3 files.

The /efi situation is irrelevant because you likely just didn't mount anything there.
/boot probably contains the kernel you falsely installed there because your fstab doesn't mount anything into /boot

On the install iso or from the booted system or wherever:

mkdir /mnt/nvme0n1p4 /mnt/sda1
mount /dev/nvme0n1p4 /mnt/nvme0n1p4
mount /dev/sda1 /mnt/sda1
ls -lR /mnt > /tmp/mnt.ls
less /tmp/mnt.ls # check whether this looks ok
cat /tmp/mnt.ls | curl -F 'file=@-' 0x0.st

Ie. mount the two fat partitions, log their contents and upload that to 0x0.st (or wherever else you can)

rebornKidz · 2025-11-17 13:26:46

This is what I do before your last post, which I hope to correct the boot process.

chroot nvme0n1/
mount /nvme0n1p4 mnt/boot
mount /nvme0n1p1 mnt/arch
mount /nvme0n1p3 mnt/home
genfstab -U /mnt /mnt/
pacman -U kernel 6.17.8

which is still wrong, and yield `fstab`

# /dev/nvme0n1p1 LABEL=/
UUID=759f6fc2-0535-4342-9f93-5f6bb1486fb3	/         	ext4      	rw,relatime	0 1

# /dev/sda1
# UUID=30C5-ACC5      	/win-boot 	vfat      	rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro	0 2

# /dev/nvme0n1p1 LABEL=/
UUID=759f6fc2-0535-4342-9f93-5f6bb1486fb3	/         	ext4      	rw,relatime	0 1

# /dev/nvme0n1p4 LABEL=ESP
# UUID=0F5A-6664      	/boot     	vfat      	rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro	0 2

# /dev/nvme0n1p1 LABEL=/
UUID=759f6fc2-0535-4342-9f93-5f6bb1486fb3	/home     	ext4      	rw,nodev,relatime	0 2

# /dev/sda1
# UUID=30C5-ACC5      	/win-boot 	vfat      	rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro	0 2

will do what you ask shortly, but I don't think I need to mount the sda. But here is the output of mnt.ls

Last edited by rebornKidz (2025-11-17 13:50:46)

seth · 2025-11-17 14:57:49

Mounting anything into /mnt after chrooting is wrong for sure and we're not gonna rely on fstab to fix this.
You also mounted your root partiton int /mnt/p2 for some reason, blowing up some hundred lines to a 3.4MB text.

Please stop flailing around.

You're boot partition is nvme0n1p4

-rwxr-xr-x 1 root root    184320 Sep 17 18:51 amd-ucode.img
drwxr-xr-x 3 root root      4096 Jun 25 12:10 EFI
drwxr-xr-x 6 root root      4096 Jul  1 19:36 grub
-rwxr-xr-x 1 root root 193619211 Oct  5 06:50 initramfs-linux-fallback.img
-rwxr-xr-x 1 root root  44621207 Oct  5 06:50 initramfs-linux.img
drwxr-xr-x 2 root root      4096 Jun 22 10:58 System Volume Information
-rwxr-xr-x 1 root root  15942144 Oct  5 06:50 vmlinuz-linux

So your fstab should look somewhat like

# /dev/nvme0n1p1 LABEL=/
UUID=759f6fc2-0535-4342-9f93-5f6bb1486fb3	/         	ext4      	rw,relatime	0 1

# /dev/nvme0n1p4 LABEL=ESP
UUID=0F5A-6664      	/boot     	vfat      	rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro	0 2

# /dev/nvme0n1p3 - this was /arch-media
UUID=1256b93c-8d3d-44cd-b630-13dbf8a0bcbb	/home     	ext4      	rw,nodev,relatime	0 2

# /dev/sda1 - this is commented, optional and really not necessary
#UUID=30C5-ACC5      	/win-boot 	vfat      	rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro	0 2

nb. that you'll likely lose your previous $HOME which is not most likely (still) a directory right under / and you need to transfer that to /home which now mounts nvme0n1p3

rebornKidz · 2025-11-17 20:41:36

The reason I am mounting p2 is that I make this partition to work with both windows. When on liveUSB I output to this partition.
Will try that out with genfstab, thanks for helping.

I am not sure, which order is correct.
I think I use genfstab wrongly, after chroot the genfstab just like what you said. but cat etc/fstab show diffently. Like it not changed.
But now the kernel on rescue mode if i boot from it. with failure to mount boot, because it was vfat partition.

Sorry

Last edited by rebornKidz (2025-11-17 20:59:13)

seth · 2025-11-17 20:46:26

Do not run genfstab and in particular not when mounting stuff into /mnt after the chroot.
Restore

pacman -U https://archive.archlinux.org/packages/l/linux/linux-6.15.3.arch1-1-x86_64.pkg.tar.zst

fix the fstab, boot the system, check "lsblk -f" to see that everything's correctly mounted into place (especially nvme0n1p4 at /boot) then update the kernel (or entire system) again and reboot.

rebornKidz · 2025-11-17 21:19:56

cannot restore linux-6.15.3 , i try twice. one with only nvme0n1p1 mounted before chroot. sencod i try mount both nvme0n1p1 and p4 as boot.
both blackscreen and un-inspectable.

seth · 2025-11-17 21:30:07

It worked before?
https://bbs.archlinux.org/viewtopic.php … 3#p2273063

Boot the install iso

mount /dev/nvme0n1p1 /mnt
mount /dev/nvme0n1p4 /mnt/boot
arch-chroot /mnt
# make sure  /etc/fstab looks like in #19
pacman -Syu # pay attention to any kind of errors
exit # the chrrot
reboot # into the installed system

rebornKidz · 2025-11-18 00:32:33

Yes, the i do the `pacman -Syu`. It broke after it. I already tried triggering offline install via cache too, as I remember 6.17.8. Yes, a couple of warnings replacing some packages.

Can it pacman -U 6.17.8 instead of pacman -Syu in this state?

Attempt with pacman -U 6.17.8

still fail when mounting boot, yield error unknown filesystem type 'vfat'. I have /etc/modules-load.d/vfat.conf exactly same with preload vfat module on boot

Attempt to revert with pacman -U 6.15.3

With both attempts, I double-checked that my fstab should look like #19.
I am still unable to boot normally now.
I suspect my old fstab because I experienced it in Aug 2025.
I tried to remove the boot from fstab, which I thought was the solution back then.
Now I know it's not, it can be the cause of my black screen.
I tried a couple of times with it enabled and disabled.

I try to boot with logLevel 7 debug rescue, and I notice it looks trying to load it twice. Is it just a retry mechanism, or is it my fault?

I am reading the journal log with ` journalctl -xe` and `journalctl -xb`. With logLevel 7, i am kinda of overwhelmed by the output, but I do not want to miss any clue / important error log.
Please let me know if you have a better idea how to examine this because I lack knowledge and experience in the rescue mode.

I also do grub-mkconfig, which I assume is a missing step that causes the black screen. This one is wrong?

Last edited by rebornKidz (2025-11-19 02:59:49)

rebornKidz · 2025-11-27 22:04:00

kindly locate the missing kernel package, install it from the cache. try to look for both version. Try again and report to us

Arch Linux

#1 2025-10-30 06:48:13