Hi folks,

Thanks to suggestions from Mike, Frederic and Marco I ended up with
something that looks somewhat sane and with a minimal amount of crud.

Patches
=======

o Patch 1 is the meat of the topic and could be picked on its own if the
  rest is too icky.
o Patch 2 introduces helpers for the dynamic preempt state
o Patches 3-5 make use of said accessors where relevant.

Testing
=======

Briefly tested the dynamic part on an x86 kernel + QEMU. x86_64_defconfig
gets me:

  Dynamic Preempt: voluntary

and appending preempt=full gets me:

  Dynamic Preempt: full

Revisions
=========

v1: http://lore.kernel.org/r/20211105104035.3112162-1-valentin.schnei...@arm.com
v1.5: 
http://lore.kernel.org/r/20211109151057.3489223-1-valentin.schnei...@arm.com

This v2 is completely different from v1, so I felt like I could get away
without writing a version changelog...

Cheers,
Valentin

Valentin Schneider (5):
  preempt: Restore preemption model selection configs
  preempt/dynamic: Introduce preempt mode accessors
  powerpc: Use preemption model accessors
  kscan: Use preemption model accessors
  ftrace: Use preemption model accessors for trace header printout

 arch/powerpc/kernel/interrupt.c |  2 +-
 arch/powerpc/kernel/traps.c     |  2 +-
 include/linux/kernel.h          |  2 +-
 include/linux/sched.h           | 16 +++++++++++++
 include/linux/vermagic.h        |  2 +-
 init/Makefile                   |  2 +-
 kernel/Kconfig.preempt          | 42 ++++++++++++++++-----------------
 kernel/kcsan/kcsan_test.c       |  4 ++--
 kernel/sched/core.c             | 17 ++++++++++---
 kernel/trace/trace.c            | 14 ++++-------
 10 files changed, 62 insertions(+), 41 deletions(-)

--
2.25.1

Reply via email to