Open klemensn opened 1 year ago
Aborting with ^C
and rerunning above nixos-install
with --no-bootloader
this time seemingly finishes the installation:
[...]
setting up /etc...
setting up /etc...
setting root password...
New password:
Retype new password:
passwd: password updated successfully
installation finished!
(Yes, /etc is set up twice this time.)
Rebooting into the new installation does not work.
The boot loader complains about missing loader.conf
and fails to boot.
Reinstalling with the latest 23.05 unstable KDE GUI installer as of today looks like it worked, i.e. no error, ticking 'Restart Now' and clicking 'Done'...
but it is stuck in a boot loop:
EFI Default Loader
[long line of garbled, repeating glyphs]
Boot in 10 s.
is all I see. It counts down, briefly blanks (might be a boot attempt?) and reenters the same screen. Forever.
Same here with 23.05 Gnome installer. I am testing a manual install using Grub instead to see if it's bootloader related.
but it is stuck in a boot loop:
EFI Default Loader [long line of garbled, repeating glyphs] Boot in 10 s.
My PBP is stuck in the same spot. Anyone found a solution?
Edit: Got a little bit further by adding pine64/pinebook-pro
from the nixos-hardware
channel. Now it apparently loads the kernel but then crashes when starting it.
Some screenshots:
Booting from internal NVMe.
Even more success using Grub:
It cannot find the NVMe, I guess. I added noquiet
and debug
to the kernel cmdline via Grub but that didn’t yield any more details.
I read there might be issues with s2ram when booting from NVMe. Is that a thing? Might it be better to load the kernel from eMMC?
Edit: Should be initrd.kernelOptions
, but no difference.
I gave up on this hardware and used a real computer again.
Here's my [truncated] Pinebook Pro config which has worked fine for months, booting from NVMe:
{ pkgs, ... }:
{
boot.kernelPackages = pkgs.linuxPackages_latest;
boot.loader.generic-extlinux-compatible.enable = true;
boot.loader.grub.enable = false;
boot.kernelParams = [
"console=tty0"
"cma=32M"
"console=ttyS0,115200n8"
"console=ttyAMA0,115200n8"
];
boot.initrd.kernelModules = [
"nvme"
"pcie_rockchip_host"
"phy_rockchip_pcie"
# Rockchip modules
"rockchip_rga"
"rockchip_saradc"
"rockchip_thermal"
"rockchipdrm"
# GPU/Display modules
"analogix_dp"
"cec"
"drm"
"drm_kms_helper"
"dw_hdmi"
"dw_mipi_dsi"
"gpu_sched"
"panel_edp"
"panel_simple"
"panfrost"
"pwm_bl"
# USB / Type-C related modules
"fusb302"
"tcpm"
"typec"
# Misc. modules
"cw2015_battery"
"gpio_charger"
"rtc_rk808"
];
}
I recall someone else having difficult with booting NixOS on Pinebook Pro, and that the boot.kernelParams
console directives were surprisingly required. YMMV.
Describe the bug
First installation attempt failed at the end due to
efivarfs
being mounted read-only, which I only briefly noted in the GUI installer's text log.I pasted the logs but they are useless/lost due to https://github.com/NixOS/nixpkgs/issues/219326.
Steps To Reproduce
Steps to reproduce the behavior:
Expected behavior
Installation finished without errors.
Additional context
I retried the installation and successfully ran
mount -t efivarfs -o remount,rw /sys/firmware/efi/efivars
prior to starting the GUI installation.Installation failed again, this time with
bootctl
receivingSIGSEGV
.coredumpctl list
is empty.Inspecting the live system shows that the new installation is still mounted, so I manually ran
nixos-install --verbose --root /tmp/calamares-root-...
which now hangs atNothing has been happening for 30 minutes,
top
showsbootctl --esp-path=/boot/efi install
idling inS
state.