Thomas suggested a few month ago to make the tick dependency check
more synchronous. Instead of checking asynchronously from each interrupt
if the tick can be stopped, every subsystem that may have a tick
dependency should set itself a flag specifying the state of that
dependency. This way we can verify if we can stop the tick with a single
lightweight mask check.
This patchset introduces that. There is still room for optimizations.
Also in a further series I would like to piggyback on the scheduler
interrupt to restart the tick in case of remote wakeups. The tick
restart from Irq exit prepares for that.
git://git.kernel.org/pub/scm/linux/kernel/git/frederic/linux-dynticks.git
timers/core-v2
HEAD: e5470c6cd483edb77d6c8398b3f1479aa0eba215
Thanks,
Frederic
---
Frederic Weisbecker (8):
jiffies: Remove HZ > USEC_PER_SEC special case
apm32: Fix cputime == jiffies assumption
alpha: Fix jiffies based cputime assumption
nohz: Remove idle task special case
nohz: Restart the tick from irq exit
nohz: Move tick_nohz_restart_sched_tick() above its users
nohz: Evaluate tick dependency once on context switch
nohz: Remove useless argument on tick_nohz_task_switch()
arch/alpha/kernel/osf_sys.c | 13 +++--
arch/x86/kernel/apm_32.c | 2 +-
include/linux/tick.h | 16 ++----
kernel/sched/core.c | 2 +-
kernel/time/tick-sched.c | 127 ++++++++++++++++++++++----------------------
kernel/time/tick-sched.h | 6 +++
kernel/time/time.c | 14 ++---
7 files changed, 93 insertions(+), 87 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
