Bug 1194455

Summary:	kernel-default-5.15.12-1.3 doesnt boot on VMware-Workstation15.5.1-player (running on windows10)
Product:	[openSUSE] openSUSE Tumbleweed	Reporter:	Detlef PÜltz <detlef.pueltz>
Component:	Kernel	Assignee:	openSUSE Kernel Bugs <kernel-bugs>
Status:	RESOLVED DUPLICATE	QA Contact:	E-mail List <qa-bugs>
Severity:	Normal
Priority:	P5 - None	CC:	ingo.goeppert+suse, krakopo, n.moendjen, nick.battle, shung-hsi.yu, skylos, tiwai, uemit.arslan
Version:	Current
Target Milestone:	---
Hardware:	x86-64
OS:	openSUSE Tumbleweed
Whiteboard:
Found By:	---	Services Priority:
Business Priority:		Blocker:	---
Marketing QA Status:	---	IT Deployment:	---

Description Detlef PÜltz 2022-01-10 11:34:28 UTC

User-Agent:       Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko
Build Identifier: 

after "zypper dup" on morining of 10.01.2022, the kernel stops booting with lot of Consolemessages like
BPF:[103117] Invalid Name_offset:1822002
BPF:[103160] FUNC
BPF:type_id=103163
BPF:
BPF: Invalid Name
BPF:

Reproducible: Always

Steps to Reproduce:
Every time I reboot the machine, if not choosing the last running kernel at GRUB

Comment 1 Kai Arif 2022-01-10 13:34:38 UTC

It's the same for my VM running on Proxmox. The hard drives go missing so the system can't start. Works perfectly with kernel-default-5.15.8-1.1.x86_64.

Comment 2 Detlef PÜltz 2022-01-10 14:25:30 UTC

Hi Kai,
what Kind of Diskcontroller provides proxmox to the guest?
my VMware-worksation provides "SCSI storage controller: Broadcom / LSI 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI", and I find no possibility to choose another, like "vmw_pvscsi" (which VMware ESX can do)

At home with qemu/KVM (choosed controllertype: virtio) I have no problems with the new kernel)

Comment 3 Kai Arif 2022-01-10 14:28:54 UTC

I get this:

 SCSI storage controller: Red Hat, Inc. Virtio SCSI

I can't change that either because is a VM running on a providers host.

Comment 4 Ingo Göppert 2022-01-10 15:16:34 UTC

I have the same problem on a Lenovo T480. But here Tumbleweed is installed on the internal NVMEe device without virtualisation. Kernel 5.15.8 boots without problems.

Comment 5 Nick Battle 2022-01-10 18:25:52 UTC

(In reply to Ingo Göppert from comment #4)
> I have the same problem on a Lenovo T480. But here Tumbleweed is installed
> on the internal NVMEe device without virtualisation. Kernel 5.15.8 boots
> without problems.

Confirmed. I have the same problem with a direct installation on a Lenovo T420. Tumbleweed is installed and updated this afternoon to 5.15.12. The previous 5.15.8 was fine (though somehow, I don't have the previous kernel to go back to, so now rather stuck).

BPF invalid name errors during boot, plus (in journalctl) many occurrences of "failed to validate module [<module name>] BTF: -22"

Comment 6 Norbert Möndjen 2022-01-10 23:03:50 UTC

Here also the same problem on bare hardware. Booting the last kernel makes the system work. As I always install the kernel Sources to do a compile of the kernel was doing the trick. Boot with old kernel than go to the new kernel tree and do:
make oldconfig
make
make modules
make install
make modules_install
And everything was working for me again.
It is a pity that QA has missed this problem.

Comment 7 Ingo Göppert 2022-01-11 07:09:59 UTC

Looks like https://bugzilla.opensuse.org/show_bug.cgi?id=1194501 describes the same bug.

Comment 8 Takashi Iwai 2022-01-11 07:20:22 UTC

Yes, very likely the same issue.  Let's close as dup.

*** This bug has been marked as a duplicate of bug 1194501 ***

Comment 9 Shung-Hsi Yu 2022-01-21 14:57:06 UTC

Hi, I'm working on it but couldn't reproduce the issue, could someone who can reproduce this send me (shung-hsi.yu@suse.com) whats mention in bug 1194501 comment 34? (and join the discussion in 1194501 if possible)

Thanks!