Bug 212545 - Toshiba Satellite S55 won't boot with Kernel >=5.11 unless acpi is set to off
Summary: Toshiba Satellite S55 won't boot with Kernel >=5.11 unless acpi is set to off
Status: RESOLVED PATCH_ALREADY_AVAILABLE
Alias: None
Product: Platform Specific/Hardware
Classification: Unclassified
Component: x86-64 (show other bugs)
Hardware: x86-64 Linux
: P1 normal
Assignee: acpi_other
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2021-04-04 23:24 UTC by aeder.redea
Modified: 2021-09-24 11:31 UTC (History)
0 users

See Also:
Kernel Version: 5.12
Subsystem:
Regression: No
Bisected commit-id:


Attachments
journalctl boot log with quiet and rhgb off (365.29 KB, text/plain)
2021-04-04 23:24 UTC, aeder.redea
Details
dmesg log (76.07 KB, text/plain)
2021-04-04 23:24 UTC, aeder.redea
Details
bisect log (3.23 KB, text/plain)
2021-04-23 06:01 UTC, aeder.redea
Details
liveusb journalctl boot log kernel 5.11.12 (1.16 MB, text/plain)
2021-04-30 19:31 UTC, aeder.redea
Details

Description aeder.redea 2021-04-04 23:24:03 UTC
Created attachment 296225 [details]
journalctl boot log with quiet and rhgb off

When booting with Kernel 5.11.10 and later this laptop will freeze before reaching SDDM. From that point onwards it won't respond to keyboard or mouse input until it is reset.

The laptop works fine up to Kernel 5.10.20 (no other kernels past this one in the 5.10.x series were tried).

The kernel will boot with the acpi=off option. Using  pci=acpioff has no effect.

Attached is systemd boot log with quiet and rhgb off of the crash

Distro: Fedora 33
CPU: Intel i5 4210U
GPU(s): Intel HD4000 (integrated), AMD Radeon R7 M260
Comment 1 aeder.redea 2021-04-04 23:24:56 UTC
Created attachment 296227 [details]
dmesg log
Comment 2 aeder.redea 2021-04-23 05:49:20 UTC
Ok, so I decided to run a git bisect on the upstream stable kernels seeing as it was very likely this is a hardware specific bug and the result was the following:

4932d37055e7a459a92424c4d21ebf8c2c0fb19a is the first bad commit
commit 4932d37055e7a459a92424c4d21ebf8c2c0fb19a
Author: Christian König <christian.koenig@amd.com>
Date:   Tue Dec 8 16:19:42 2020 +0100

    drm/amdgpu: limit the amdgpu_vm_update_ptes trace point
    
    The text output should not be more than a page, so only print the first
    32 page table entries.
    
    If we need all of them we can still look into the binary trace.
    
    Signed-off-by: Christian König <christian.koenig@amd.com>
    Reviewed-by: Shashank Sharma <shashank.sharma@amd.com>
    Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

 drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

I'm not entirely sure how to proceed from this point onwards though.
Comment 3 aeder.redea 2021-04-23 06:01:54 UTC
Created attachment 296467 [details]
bisect log
Comment 4 aeder.redea 2021-04-30 19:23:56 UTC
> When booting with Kernel 5.11.10 and later this laptop will freeze before
> reaching SDDM. From that point onwards it won't respond to keyboard or mouse
> input until it is reset.

It should be noted that even after updating everything else (and switching to an SSD), this error still happens. The only difference is that I make it to the login screen and then after 5 seconds it freezes.  


> The laptop works fine up to Kernel 5.10.20 (no other kernels past this one
> in the 5.10.x series were tried).

An update on this: I tried up to kernel 5.10.28 and it works, I also tried kernel 5.11.3 and it doesn't work. So those are the points I used for the bisection.

This problem is still present in kernel 5.12
Comment 5 aeder.redea 2021-04-30 19:30:57 UTC
Additionally using something like the Fedora LiveUSB (on kernel 5.11.12) gets past the login screen, but not without some errors ocurring during boot. I'm not sure if the boot log from the liveusb captured the exact same issue that I see flash past the screen, but im attaching the file in case it's useful.
Comment 6 aeder.redea 2021-04-30 19:31:50 UTC
Created attachment 296559 [details]
liveusb journalctl boot log kernel 5.11.12
Comment 7 aeder.redea 2021-05-28 15:40:51 UTC
Another update: Adding radeon.runpm=0 also allows it to boot with kernel 5.12, without losing access to battery information.
Comment 8 aeder.redea 2021-09-24 11:31:41 UTC
Kernel 5.13 no longer has this issue.

Note You need to log in before you can comment on or make changes to this bug.