Frederic, Can you take these patches. The first patch will need to get an ack by one of the x86 maintainers. That one adds a new schedule_preempt_user() that is called before going back to user space in order to safely switch from user_context before calling schedule.
The second patch prevents LOCKUP_DETECTION from being selected if NO_HZ_FULL is. That's because if LOCKUP_DETECTION is selected, it will prevent NO_HZ_FULL from functioning. I first thought about having NO_HZ_FULL depend on LOCKUP_DETECTION, but it would hide the feature. As NO_HZ_FULL is a feature and LOCKUP_DETECTION is debugging, I have precedence to the feature. The last patch produces a warning if nohz_full is used or NO_HZ_FULL_ALL is defined, and the system fails to go into nohz_full mode because the machine's clock is unstable. We don't want users thinking they have nohz_full but their machine can't handle it. -- Steve --- arch/x86/include/asm/context_tracking.h | 10 --------- b/arch/x86/kernel/entry_64.S | 26 +++++-------------------- b/kernel/sched/core.c | 33 +++++++++++++++++--------------- b/kernel/time/tick-sched.c | 5 ++++ b/lib/Kconfig.debug | 2 + 5 files changed, 31 insertions(+), 45 deletions(-) -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
