Boot hangs - sort of [SOLVED]

[bluenuht]

The boot process stops after showing a green star next to Logical Volume Manager

But if I blindly ALT+F4 enter my user credentials and run

Code: Select all

startx

I get my normal GUI

I do not use LVM so I deleted /etc/runlevels/boot/lvm and all that changes is it stops on the completion of checking local filesystems (green star)

qlop hints towards either nvidia drivers or GRUB
https://bpa.st/OFCA

If I run the kernel with

Code: Select all

pci=noacpi

then I get the complete boot process and the tty login, but then startx fails

Currently I am running the kernel with

Code: Select all

pcie_aspm=off video=efifb:reprobes vga=current console=tty1

but still have to do my blind login

dmesg doesn't have anything I can latch on to
https://bpa.st/JZMA

About three weeks before this I had updated the kernel to 6.12.58 and had had multiple boots without issue.

Since exploring the issue I have also updated my motherboard firmware to the latest. In theory the issue was rebar and memory allocation, but I have toggled the hell out of those options and kernel parameters.

Any ideas where I should look next?

[NeddySeagoon]

bluenuht,

Please put your kernel .config file onto a pastebin, or tell us which binary kernel you use.
It sounds like you are missing a framebuffer console driver.
I thinkeit goes wrong here

Code: Select all

[Sat Feb 28 09:38:08 2026] [drm] Initialized nvidia-drm 0.0.0 for 0000:01:00.0 on minor 0
[Sat Feb 28 09:38:08 2026] Console: switching to colour dummy device 80x25
[Sat Feb 28 09:38:08 2026] nvidia 0000:01:00.0: vgaarb: deactivate vga console

[bluenuht]

sorry for the delay, I got stuck with net-libs/webkit-gtk slot conflicts that took 8-10 hours per build.

.config
https://pastebin.com/bSgu2tZh

The only thing I added since the issue started was

Code: Select all

CONFIG_FB_SIMPLE=y

and that made no difference

my usual grub boot line is

Code: Select all

BOOT_IMAGE=/vmlinuz-6.12.58-gentoo root=PARTUUID=21d71659-6646-467b-826d-27450fca2289 ro pcie_aspm=off

[pietinger]

bluenuht,

You have configured your graphics driver as built-in:

Code: Select all

CONFIG_DRM_I915=y

... but you have an i5-9400F ... F means no integrated GPU

At the moment 6.18 is the current Gentoo stable kernel ...

Maybe have a look into:
https://wiki.gentoo.org/wiki/User:Pieti ... figuration

and here a an old link to 6.12:
https://wiki.gentoo.org/wiki/User:Pieti ... rsion_6.12

but I suggest to install and use 6.18 ->
https://wiki.gentoo.org/wiki/User:Pieti ... rsion_6.18

What happens if you remove all the kernel command line parameters (except of course: root=PARTUUID=21d71659-6646-467b-826d-27450fca2289 ro)?


Internal notes:

Code: Select all

[Sat Feb 28 09:38:04 2026] smpboot: CPU0: Intel(R) Core(TM) i5-9400F CPU @ 2.90GHz (family: 0x6, model: 0x9e, stepping: 0xa)
# CONFIG_X86_INTEL_LPSS is not set
# CONFIG_MFD_INTEL_LPSS_ACPI is not set
# CONFIG_MFD_INTEL_LPSS_PCI is not set
# CONFIG_INTEL_IOMMU_DEFAULT_ON is not set
# CONFIG_IRQ_REMAP is not set
CONFIG_USB_PCI_AMD=y

[NeddySeagoon]

bluenuht,

Code: Select all

CONFIG_FB_SIMPLE=y

is so simple that the kernel cannot configure it. The kernel needs to be passed a ready to use framebuffer by the boot loader. I'm fairly sure that it won't help though.
You have a working console until the nvidia driver loads. Then kernel that switches away from the working console driver to a broken one, so your console stops updating.

pietinger hinted at some LPSS knobs you probably need in your kernel.
Can you share

Code: Select all

lspci -nnk

output please?

-nn provides the Vendor and Device IDs, which should be an unambiguous pointers to your devices.
The -k tells about the kernel modules loaded for the devices, if any.

I suspect you will need something in one or both of

Code: Select all

# CONFIG_PINCTRL is not set
# CONFIG_GPIOLIB is not set

of these menus.

[pietinger]

[...]
I suspect you will need something in one or both of

Code: Select all

# CONFIG_PINCTRL is not set
# CONFIG_GPIOLIB is not set

of these menus.

Neddy,

no ... these can not cause such a problem ... my guess is a missing CONFIG_X86_X2APIC=y

(we had already here: https://forums.gentoo.org/viewtopic-t-1174380.html )

... and he does not have it (and does not see this option) because there is this dependency:

Code: Select all

Depends on: X86_LOCAL_APIC [=y] && X86_64 [=y] && (IRQ_REMAP [=y] || HYPERVISOR_GUEST [=n])

... so it is necessary to enable IRQ_REMAP to get it ... or ... changeover to 6.18

(X86_X2APIC is NOT enabled by default in 6.12; it is enabled by default in 6.18; this was the reason I gave him the links)

[bluenuht]

Bottom line is that building 6.18.12 has fixed this issue. Thanks !

But 6.18.12 wasn't available at the end of January, only 6.12.58

So I would be really interested if anyone has any ideas how I could have
1. built the kernel (6.12.58) and had multiple weeks of "good" boots
2. an emerge -avDNut world would then cause this issue