Bug 216797 - Asus laptop won't boot after kernel upgrade above 6.0.8, acpi=off solves it...kind of!
Summary: Asus laptop won't boot after kernel upgrade above 6.0.8, acpi=off solves it.....
Status: NEW
Alias: None
Product: ACPI
Classification: Unclassified
Component: Config-Other (show other bugs)
Hardware: Intel Linux
: P1 blocking
Assignee: acpi_config-other
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2022-12-11 10:05 UTC by shayan.javani
Modified: 2023-01-13 05:14 UTC (History)
2 users (show)

See Also:
Kernel Version: after 6.0.8
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description shayan.javani 2022-12-11 10:05:22 UTC
At first I should confess that I'm not sure if the issue is in acpi, but I didn't know where else to put it, and also acpi seems related.

I have an Asus laptop(neofetch says the model is GL553VD 1.0). The cpu is intel i7-7700HQ with 16GB of RAM and a NVIDIA GeForce GTX 1050 Mobile GPU but I don't have a nvidia driver.
I have a 256GB SSD and a 1TB HDD. I have a windows 10 on my HDD that I use for gaming and a fedora 37 on my SSD as my main OS.

Now to the problem. I first started with fedora 36, and when fedora 37 came out, I directly upgraded it, no problem. But then I upgraded my kernel from 6.0.8 to 6.0.9 but then the system would not boot. After the grub menu, the fedora loading screen would appear but then it would just go to a black screen and nothing else. So I had to revert back to 6.0.8. I thought it would be fixed in the next upgrade, but no, after each upgrade the problem persisted and I had to revert back to 6.0.8 again.

Until today when I upgraded to 6.0.12 and the problem is still there. So I tried to find a way to solve it until I remembered my computer at work. I recently started working at a new company and they game me a desktop(don't know the hardware details) to work with with only windows 10 on it. So I installed fedora 36 but when you logged into it, it would just reset the system after a few seconds. After poking around a little, it turned out that as soon as I logged in, journald would start logging an error related to acpi(don't remember what it was) constantly and until the system would crash. So we set acpi=off in the grub config and the issue was solved. There was also no problem with direct upgrade to fedora 37 and with upgrading to new kernels. So I thought maybe somehow that would also solve the issue with my laptop.

So I set acpi=off as a grub arg for the kernel 6.0.12 entry, and it worked. The system booted and I got to the login screen. But now my keyboard and touchpad are not working and I can't login. So again, I had to revert back to 6.0.8.

I wish there is enough information here to solve the issue and feel free the move it to the appropriate place if this is not it.

Thanks for the great work
Comment 1 The Linux kernel's regression tracker (Thorsten Leemhuis) 2022-12-12 06:17:15 UTC
Just to ensure I understand this properly: So the system actual boots and you can even log in, but shortly after that it crashes? In that case please try to grab the output with the other computer. 

Side note: Would be good to know if 6.1 is affected as well. Also, 6.0 will likely be EOL in two or three weeks anyway. You hence might want to consider giving 6.1 a try (https://fedoraproject.org/wiki/Kernel_Vanilla_Repositories should have it within 24 hours, but you can try a pre-release that'S available there already now, too).
Comment 2 shayan.javani 2022-12-12 10:55:57 UTC
No, that was the problem I had with the computer at work that was solved with acpi=off. I just mentioned that because that was the reason I tried acpi=off in my laptop.

The problem with my laptop, which is the main reason for this bug report, is that it actually does not boot after 6.0.8. But then I tried acpi=off on 6.0.12 and I could get to the login screen, but my touchpad and keyboard were disabled.

I'll try 6.1 and report back.
Comment 3 shayan.javani 2022-12-16 09:26:38 UTC
I just downloaded kernel 6.1.0 from kernel.org and built and installed it on my laptop. But, again, after choosing it in the grub menu, it won't boot and just remains on a black screen with the cursor blinking. So the problem is still there even it 6.1.0!
It seems like the problem is with the kernel itself, but I may try to use another distribution to see what happens.
Comment 4 The Linux kernel's regression tracker (Thorsten Leemhuis) 2022-12-16 09:51:28 UTC
thx for your help, but I fear more work is needed from your side to find where the problem originates. Have you tried removing the parameters "rhgb quiet" to see if the kernel then shows some helpful error messages? And have you tried additionally using "nomodeset"?
Comment 5 shayan.javani 2022-12-16 12:47:21 UTC
I removed "rhgb quiet", but didn't get any error messages. And before getting the black screen and the blinking cursor, all the things that are listed have "[OK]" in the beginning.

I also added tried "nomodeset", but this time I only got the black screen, without the blinking cursor!
Comment 6 The Linux kernel's regression tracker (Thorsten Leemhuis) 2022-12-16 12:58:40 UTC
(In reply to shayan.javani from comment #5)
> I removed "rhgb quiet", but didn't get any error messages. And before
> getting the black screen and the blinking cursor, all the things that are
> listed have "[OK]" in the beginning.

So the system actually starts and hangs before or after the login manager? Try booting with kernel parameter "3"; if you can login there check "dmesg -w" for error messages. If that works, try to start the graphical session.

Sorry, I can't walk you through the details, but I hope you get the idea.
Comment 7 shayan.javani 2022-12-19 15:10:59 UTC
I did as you said and I could access a command line on 6.0.12. Then I ran "dmesg -w" and the only thing I can see resembling an error message is: "thermal thermal_zone3: failed to read out thermal zone (-61)" in bold!

Then I ran "sudo systemctl start gdm.service" which led me to the black screen with the blinking cursor!
Comment 8 shayan.javani 2022-12-19 15:11:31 UTC
I did as you said and I could access a command line on 6.0.12. Then I ran "dmesg -w" and the only thing I can see resembling an error message is: "thermal thermal_zone3: failed to read out thermal zone (-61)" in bold!

Then I ran "sudo systemctl start gdm.service" which led me to the black screen with the blinking cursor!
Comment 9 The Linux kernel's regression tracker (Thorsten Leemhuis) 2022-12-19 15:31:43 UTC
try "3" again and this time run "dmesg -w > foo" before starting gdm; then try to reboot with Ctrl+Alt+Del (or use Sysrq [google for it] with a sync if that doesn't work); after reboot check if any helpful messages can be found in that file. Alternatively you can try to grab the log from another machine as well.
Comment 10 shayan.javani 2023-01-13 05:14:44 UTC
Sorry, I've been really busy. Anyway, I changed my OS from fedora to opensuse tumbleweed and now I have no problem with kernel 6.1. It seems the problem is with the combination fedora+my laptop+kernel v > 6.0.8.

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