[RESOLVED] Latest liquorix amd64 kernels (5.10 & 5.11) won’t boot
DigitalG
Status: New User - Welcome
Joined: 01 Mar 2021
Posts: 3
Reply Quote
On my PC, the latest liquorix kernels will only boot once in a blue moon. There seem to be no difference between a reboot and a cold boot, usually it will be stuck in grub with the following message:

:: Quote ::
Loading 5.11.0-2.3-liquorix-amd64 …
Loading initial ramdisk ...


Occasionally it will boot, but this normally only happens after several previous unsuccessful attempts.

This affects all kernels from linux-image-5.10.0-13.2-liquorix-amd64 up to the latest linux-image-5.11.0-2.3-liquorix-amd64 .

The latest kernel that works 100% is linux-image-5.10.0-12.1-liquorix-amd64, so I guess the problem must be a change introduced with the 5.10.0-13.2-liquorix kernel…

My system:

:: Code ::
➜ inxi -bxx
System:    Host: debian-gnome Kernel: 5.10.0-12.1-liquorix-amd64 x86_64 bits: 64 compiler: gcc v: 10.2.1 Desktop: GNOME 3.38.3
           tk: GTK 3.24.24 wm: gnome-shell dm: GDM3 Distro: Debian GNU/Linux bullseye/sid
Machine:   Type: Desktop Mobo: ASUSTeK model: ROG STRIX B350-F GAMING v: Rev X.0x serial: <superuser required>
           UEFI: American Megatrends v: 5406 date: 11/13/2019
CPU:       Info: 8-Core AMD Ryzen 7 3700X [MT MCP] arch: Zen 2 speed: 3531 MHz min/max: 2200/3600 MHz
Graphics:  Device-1: Advanced Micro Devices [AMD/ATI] Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] vendor: Gigabyte
           driver: amdgpu v: kernel bus ID: 08:00.0 chip ID: 1002:67df
           Display: wayland server: X.Org 1.20.10 compositor: gnome-shell driver: loaded: amdgpu,ati
           unloaded: fbdev,modesetting,vesa resolution: 1: 2560x1440~144Hz 2: 1920x1080~60Hz s-dpi: 96
           OpenGL: renderer: Radeon RX 580 Series (POLARIS10 DRM 3.40.0 5.10.0-12.1-liquorix-amd64 LLVM 11.0.1)
           v: 4.6 Mesa 20.3.4 direct render: Yes
Network:   Device-1: Intel I211 Gigabit Network vendor: ASUSTeK driver: igb v: kernel port: e000 bus ID: 03:00.0
           chip ID: 8086:1539
Drives:    Local Storage: total: 2.47 TiB used: 640.94 GiB (25.3%)
Info:      Processes: 402 Uptime: 56m Memory: 15.57 GiB used: 5.32 GiB (34.2%) Init: systemd v: 247 runlevel: 5 Compilers:
           gcc: 10.2.1 alt: 10/8/9 Packages: 4009 apt: 3986 flatpak: 14 snap: 9 Shell: Zsh v: 5.8 running in: gnome-terminal
           inxi: 3.3.01


Thanks!
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 906
Reply Quote
Here's the changelog between those versions:

:: Code ::
linux-liquorix (5.10-20) unstable; urgency=medium

  * merge 5.10.14 rc patches
  * disable most/all trace config options
  * update version to 5.10.0-13.2

 -- Steven Barrett <steven@liquorix.net>  Fri, 05 Feb 2021 09:12:49 -0600

linux-liquorix (5.10-19) unstable; urgency=medium

  * merge 5.10.13
  * update version to 5.10.0-13.1

 -- Steven Barrett <steven@liquorix.net>  Wed, 03 Feb 2021 19:31:21 -0600

linux-liquorix (5.10-18) unstable; urgency=medium

  * merge 5.10.12
  * update version to 5.10.0-12.1

 -- Steven Barrett <steven@liquorix.net>  Sat, 30 Jan 2021 09:18:28 -0600


I think a stable patch that entered 5.10.14 (part of the 5.10.0-13.2 release), caused the issue. It definitely shouldn't be disabling trace options, unless that somehow triggered a new bug that only happens with tracing options off.
Back to top
DigitalG
Status: New User - Welcome
Joined: 01 Mar 2021
Posts: 3
Reply Quote
Many thanks for your reply!

So I guess the bug is upstream then? Not sure if the following is of any help, I do know that liquorix differs from normal kernels... But anways, I did grab the generic 5.11 kernel debs from Ubuntu’s mainline kernel build service, just to test if a standard kernel build would work: 5.11.0-051100-generic downloaded from https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.11/amd64/

And this build actually works without any issues at all. Cold boot + multiple reboots both work without any freeze.
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 906
Reply Quote
Yes, I believe this bug is upstream, but MuQSS triggers it more frequently due to differences in process scheduling that don't affect CFS in mainline. Here's a github issue from one report in Zen Kernel where the original author decided to drop Zen Kernel entirely before answering any questions: github.com/zen-kernel/zen-kernel/issues/201

However, I did just push out an update yesterday that enables tracing. From an anecdotal report, this does mitigate the number of boot hangs but they still occur on AMD systems. Can you verify it makes any difference on your system?
Back to top
DigitalG
Status: New User - Welcome
Joined: 01 Mar 2021
Posts: 3
Reply Quote
I can confirm that the issue actually seem to be resolved in today's update (5.11.0-2.4-liquorix-amd64)! Tried a cold boot + and I've also done several reboots, and it boots every time now... I haven't had a hang yet; fingers crossed! :) Guess it's kind of weird that re-enabling tracing would make such a difference. Though if I do experience hangs I'll make sure to report back.

Many thanks, your work is much appreciated!
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 906
Reply Quote
Thanks! Glad your system boots properly now. Marking thread as resolved.
Back to top
Display posts from previous:   

All times are GMT - 8 Hours