[Pkg-libvirt-maintainers] Bug#1024753: libvirtd doesn't run: "SCHED_CORE not supported by kernel"

[Salvatore Bonaccorso]

Control: reassign -1 src:libvirt 8.9.0-1

On Thu, Nov 24, 2022 at 10:10:49AM +0100, Philipp Marek wrote:
> Package: src:linux
> Version: 6.0.8-1
> Severity: important
> X-Debbugs-Cc: philipp at marek.priv.at
> 
> 
> Sorry about the German messages.
> 
>     libvirt version: 8.9.0, package: 1 (Andrea Bolognani <eof at kiyuko.org>
> Sat, 19 Nov 2022 23:00:34 +0100)
> 
>     Nicht unterstützte Konfiguration: SCHED_CORE not supported by kernel
>     Initialisierung des QEMU Status-Treibers fehlgeschlagen: Nicht
> unterstützte Konfiguration: SCHED_CORE not supported by kernel
>     Treiber Status Initialisierung fehlgeschlagen
> 
> My libvirt-daemon is 8.9.0-1; downgrading to 8.5.0-1 makes it work
> again, should the error be reported against that one,
> or does it make sense to recompile the kernel with that config?

Doing for now the reassignment to src:libvirt. In fact SCHED_CORE has
been disabled by default with upstream commit, d2343cb8d154
("sched/core: Disable CONFIG_SCHED_CORE by default"), in v5.14-rc1:

| commit d2343cb8d154fe20c4499711bb3a9af2095b2b4b
| Author: Ingo Molnar <mingo at kernel.org>
| Date:   Mon Jun 28 21:55:16 2021 +0200
| 
|     sched/core: Disable CONFIG_SCHED_CORE by default
| 
|     This option at minimum adds extra code to the scheduler - even if
|     it's default unused - and most users wouldn't want it.
| 
|     Reported-by: Linus Torvalds <torvalds at linux-foundation.org>
|     Signed-off-by: Ingo Molnar <mingo at kernel.org>
| 
| diff --git a/kernel/Kconfig.preempt b/kernel/Kconfig.preempt
| index bd7c4147b9a8..5876e30c5740 100644
| --- a/kernel/Kconfig.preempt
| +++ b/kernel/Kconfig.preempt
| @@ -102,7 +102,6 @@ config PREEMPT_DYNAMIC
| 
|  config SCHED_CORE
| 	bool "Core Scheduling for SMT"
| -       default y
| 	depends on SCHED_SMT
| 	help
| 	  This option permits Core Scheduling, a means of coordinated task
| @@ -115,7 +114,8 @@ config SCHED_CORE
| 	   - mitigation of some (not all) SMT side channels;
| 	   - limiting SMT interference to improve determinism and/or performance.
| 
| -         SCHED_CORE is default enabled when SCHED_SMT is enabled -- when
| -         unused there should be no impact on performance.
| +         SCHED_CORE is default disabled. When it is enabled and unused,
| +         which is the likely usage by Linux distributions, there should
| +         be no measurable impact on performance.

I have not checked what other distributions do, but given the above it
looks to me that we are doing already the okayish thing for
distributions. Libvirt-maintainers, what is your take on it?

How is your /etc/libvirt/qemu.conf configured with respect to the
"sched_core" setting?

Regards,
Salvatore