FS#72789 - [linux-zen] system can't boot on kernel 5.15.3, linux-xanmod 5.14.20 as well

Attached to Project: Arch Linux
Opened by Maxim Sheviakov (mradermaxlol) - Friday, 19 November 2021, 20:22 GMT
Last edited by Jan Alexander Steffens (heftig) - Monday, 22 November 2021, 14:45 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To Jan Alexander Steffens (heftig)
Architecture All
Severity Critical
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 3
Private No

Details

Description: after upgrading linux-zen to 5.15.3 from testing my system (ASUS TUF FX505DY, Ryzen 5 3550H) doesn't boot at all, I can only see the EFI splash screen. It reacts to REISUB with sysrq, though - the screen is cleared and regular REISUB messages on the tty are shown.
linux-zen 5.15.2 works fine; strangely enough, linux-xanmod 5.14.20 (which is my main kernel right now) is also broken in the same way, although 5.14.18 with the very same config worked fine.


Additional info:
* package version(s): linux-zen 5.15.3, linux-xanmod 5.14.20
This task depends upon

Closed by  Jan Alexander Steffens (heftig)
Monday, 22 November 2021, 14:45 GMT
Reason for closing:  Fixed
Additional comments about closing:  linux 5.15.4
Comment by Mthw (jari_45) - Sunday, 21 November 2021, 10:24 GMT
Same problem on the same laptop, during boot systemd shows (not exactly): "Waiting for /dev/mapper/luks" and "Waiting for /dev/nvme0n1p5" (which are the same device). I am assuming opening encrypted drives is broken.
Comment by Mthw (jari_45) - Sunday, 21 November 2021, 10:30 GMT
More info: The partition contains btrfs sub volumes one of which is the root. I also use tpm to unlock it. Older kernels work correctly.
Comment by Carlos Santacruz (cfsa900) - Sunday, 21 November 2021, 12:44 GMT
Exact same symptoms on the exact laptop model. Needed to downgrade to 5.15.2 to get a bootable system.
No volumes, encrypting or anything, plain old ext4 partitions on my setup.
Comment by Mthw (jari_45) - Sunday, 21 November 2021, 13:10 GMT
It looks like either the SSD or NVMe in general is not working correctly. I wonder if the drive would be accessible if I booted from a USB and tried to mount it. Also linux-mainline 5.16rc1 is broken. There must've been a bad commit between 5.15.2 and 5.15.3 that causes this.
Comment by Carlos Santacruz (cfsa900) - Sunday, 21 November 2021, 13:25 GMT
>I wonder if the drive would be accessible if I booted from a USB and tried to mount it
I had to chroot in order to downgrade. The drives responded fine.
You mention an SSD, when I was checking around on the emergency shell, my secondary drive (regular HDD) showed up correctly on /dev/disk/ so it has to be something related to that.
Comment by Maxim Sheviakov (mradermaxlol) - Sunday, 21 November 2021, 19:38 GMT
Yeah, seems like the common denominator here is the usage of an encrypted SSD. It also seems like the bad commit has got backported into older stable trees as well (5.14.20, for example).
Comment by loqs (loqs) - Sunday, 21 November 2021, 21:06 GMT
Can you bisect between 5.15.2 and 5.15.3 and find the causal commit? Or try reverting 54718ee9b8eed43ab5e20df936a6e638416a4bcf

https://bugzilla.kernel.org/show_bug.cgi?id=215081
Comment by Laurențiu Nicola (lnicola) - Monday, 22 November 2021, 07:53 GMT
It might also be371c6dde-ee04-cf2e-1463-113ba8a2ac22@googlemail.com/T/"> https://lore.kernel.org/lkml/371c6dde-ee04-cf2e-1463-113ba8a2ac22@googlemail.com/T/, reverted in 5.15.4.
Comment by Mthw (jari_45) - Monday, 22 November 2021, 08:54 GMT
Kernel 5.15.4 fixes this issue, can anyone confirm?
Comment by Maxim Sheviakov (mradermaxlol) - Monday, 22 November 2021, 10:22 GMT
Indeed, linux-zen 5.15.4 from testing boots fine. No unusual journal entries, either.
linux-xanmod 5.14.21 works fine as well.
Comment by Carlos Santacruz (cfsa900) - Monday, 22 November 2021, 10:34 GMT
Working here as well.

Loading...