diff options
| author |   Yunhong Jiang <yunhong.jiang@intel.com> | 2015-08-18 11:23:10 -0700 |
|---|---|---|
| committer |   Yunhong Jiang <yunhong.jiang@linux.intel.com> | 2016-07-21 17:49:30 -0700 |
| commit | fafab8722263f4ab08894e35f4385febbce09dd1 (patch) | |
| tree | 018aafc8ddc8b37a6b7fedbea1086f0531fba090 /kernel | |
| parent | ef4e798bc8761c401451649ed17a52e3e1c638e8 (diff) | |
Restore the non_hz_full
Please refer to http://comments.gmane.org/gmane.linux.kernel/1891417, the
last two mails on this discussion.
The non_hz_full is important for RT, so take it back temporarily.
Unluckily I didn't find the corresponding reverting patch on the RT tree.
Upstream status: pending
Change-Id: I2748a8f9b7a98ef2185a1da60089984432393eff
Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
Diffstat (limited to 'kernel')
| -rw-r--r-- | kernel/kernel/time/timer.c | 10 |
1 files changed, 8 insertions, 2 deletions
diff --git a/kernel/kernel/time/timer.c b/kernel/kernel/time/timer.c index 76a301b24..016a0bf52 100644 --- a/kernel/kernel/time/timer.c +++ b/kernel/kernel/time/timer.c @@ -1453,9 +1453,11 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem) * the base lock to check when the next timer is pending and so * we assume the next jiffy. */ - return basem + TICK_NSEC; -#endif + if (!spin_do_trylock(&base->lock)) + return basem + TICK_NSEC; +#else spin_lock(&base->lock); +#endif if (base->active_timers) { if (time_before_eq(base->next_timer, base->timer_jiffies)) base->next_timer = __next_timer_interrupt(base); @@ -1465,7 +1467,11 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem) else expires = basem + (nextevt - basej) * TICK_NSEC; } +#ifdef CONFIG_PREEMPT_RT_FULL + rt_spin_unlock(&base->lock); +#else spin_unlock(&base->lock); +#endif return cmp_next_hrtimer_event(basem, expires); } |