#1128632 linux-image-6.12.73+deb13-amd64: boot failures, ACPI errors, and sleep issues on ASUS Sabertooth Z77 (i7-3770K)

#1128632#5
Date:
2026-02-22 02:18:02 UTC
From:
To:

Dear Debian Kernel Team,


I would like to report several boot and power-management issues observed after installing kernel version 6.12.73+deb13-amd64 on older Ivy Bridge hardware.

System Information


Motherboard: ASUS Sabertooth Z77
CPU: Intel Core i7-3770K (Ivy Bridge)
Firmware: UEFI (legacy-compatible board)
Distribution: Debian (stable branch)
Architecture: amd64
GPU: NVIDIA (proprietary driver installed)


Problem Description

After upgrading to kernel 6.12.73+deb13-amd64, the system exhibits multiple boot and runtime problems:


Boot failures occur intermittently with the following messages:



EFI stub: ERROR: Failed to decompress kernel
EFI stub: ERROR: efi_stub_entry() failed



During successful boots, ACPI-related errors appear in dmesg involving unresolved symbols and SATA controller initialization (_GTF/DSSP related messages).


System suspend behavior is unreliable:
*When placing the system into sleep mode, the computer immediately wakes without user interaction.


Overall system startup appears noticeably slower compared to previous kernels.

Expected Behavior

The system boots reliably without EFI stub errors, ACPI warnings, or unexpected wake events during suspend.

Observed Behavior

Kernel 6.12.73+deb13-amd64 causes intermittent boot loader failures, ACPI-related warnings, and unreliable suspend functionality on this hardware.

Workaround

Booting an earlier kernel (e.g., 6.12.69+deb13-amd64) restores normal behavior and eliminates the EFI decompression failure and sleep issues.

Additional Notes

Reinstalling the kernel temporarily resolved the EFI decompression error, suggesting either a regression affecting older firmware platforms or sensitivity during installation on legacy UEFI implementations.


Please let me know if additional logs, dmesg output, or hardware information would be helpful for debugging.


Best regards,


Donovan

#1128632#10
Date:
2026-02-22 02:35:06 UTC
From:
To:
Hello,

As requested, attaching logs from the affected system. 

Best regards,
Donovan

#1128632#21
Date:
2026-02-27 21:56:01 UTC
From:
To:
Hello,

That is really surprising. This happens very early in the boot process
and if there was a problem in that code, this wouldn't be the only
report. The efi stub randomizes the memory location, that is the only
non-deterministic thing I'm aware of that could explain that (maybe in
combination with a BIOS issue where a memory area isn't marked as
reserved). Did that happen more than once and only with
6.12.73+deb13-amd64?

That is unfortunately normal and most of the time no problem. I would
expect you have these with earlier kernels, too.

https://www.kernel.org/doc/html/latest/power/basic-pm-debugging.html#a-test-modes-of-hibernation
contains some debug ideas. I'm not sure how up-to-date this document is,
but it is worth a try.

Can you please provide a kernel log for both 6.12.69+deb13-amd64 and
6.12.73+deb13-amd64?

Best regards
Uwe

#1128632#28
Date:
2026-03-04 17:28:35 UTC
From:
To:
Hello,

Forwarding to the bug address to have this info in the buglog. Please always keep the bug address on Cc: to increase the bus factor.

Best regards
Uwe

#1128632#33
Date:
2026-03-07 07:22:10 UTC
From:
To:
Hello Donovan,

I'm irritated about your wording. If 6.12.69 is "more reliable", that
means it crashes less often than 6.12.73 but not never, right?

Putting the Linux image and initramfs into RAM and decompressing them
are the first heavy RAM operations. So if you have overclocked your
system and RAM isn't 100% reliable any more it's expected that it fails
during these operations.

But you still saw failures?

It was already in your first mail, but there I missed it. The
proprietary nvidia driver is known to introduce problems that are not
debuggable due to the driver not being OSS. Can you please try to
reproduce the suspend issue without this driver? It wouldn't be the
first time the nvidia driver creates suspend issues. If not using it
fixes the suspend issue, there is unfortunately nothing we (Debian
kernel team) can do for you, and also upstream won't be able to help.

There are a few differences (e.g. irq assignment to the different SATA
devices, which might play a role, but my bet is on the nvidia driver
until proven otherwise for now.

Best regards
Uwe