Hi,
On Sat, Oct 14, 2023 at 05:15:04PM -0400, Jeffrey Mark Siskind wrote:
> When first installed, it ran kernel 6.1.0.12. That kernel found
> all 384 "CPUs". All were reported in /proc/cpuinfo. I subsequently
> did an apt upgrade which upgraded to 6.1.0.13. Upon boot, dmesg
> -lerr reports:
>
>root@poto:~# dmesg -lerr
>[ 11.080833] smpboot: do_boot_cpu failed(-1) to wakeup CPU#192
>
> Further, /proc/cpuinfo is missing processor 192.
>
> I subsequently installed 6.4.0-0-deb12.2-amd64 from bookwork
> backports. The issue still occurs. If I boot 6.1.0.12 I do not get the issue.
I think you should use "reportbug" to report the bug to the kernel
team.
You may like to try building a Debian kernel package from upstream
source:
https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-kernel-org-package
If that results in a package that works for you, that's something to
report in the bug.
If it doesn't, you can go back through the 6.1 kernel releases until
you find one that does work. Then you can report that.
Once you find an upstream source version that doesn't work, you can
use "git bisect" to go back and find the version that DOES work, and
then further bisect things until you find the exact commit that
causes the issue for you.
The Debian kernel team may then be able to backport a fix to the
bookworm stable kernel, and submit it upstream if necessary.
Alternatively, the kernel team may have a way for you to step
through individual changes in between Debian packages 6.1.0.12 and
6.1.0.13.
Good hunting!
Thanks,
Andy
--
https://bitfolk.com/ -- No-nonsense VPS hosting
