hrtimer: Don't take cpu_base::lock in hrtimer_get_next_event() when hres_active

hrtimer_get_next_event() runs on every tick-stop decision via get_next_timer_interrupt() -> cmp_next_hrtimer_event(). When high resolution timers are active it must return KTIME_MAX -- the caller documents and depends on this. The function takes cpu_base->lock, checks hres_active, and returns KTIME_MAX. Taking the lock is not required because cpu_base->hres_active is only written by the local CPU in hrtimer_switch_to_hres() from hard interrupt context and in hrtimers_cpu_starting() during bring-up. All callers of hrtimer_get_next_event() reach it from the tick-stop / cpuidle paths with interrupts disabled on that CPU. No writer can therefore race with the read, so an unlocked hres_active check is stable and the lock can be skipped in this case. On a 176-thread AMD EPYC 9D64 running a production workload, bucketing callers of native_queued_spin_lock_slowpath(), the slowpath had 199 samples in total, of which 54 are attributed to hrtimer_get_next_event(), i.e. this accounts for ~27% of slowpath hits on this specific workload. Move the hres_active check before the lock guard region to address this. [ tglx: Massaged change log and comments ]

Usama Arif committed 4 months ago

Commit c4415c91 parent 1d28a671 file · +13 −5Browse files

Showing the changes to kernel (1 of 1 file). Show all

Unified · Split

kernel/time/hrtimer.c+13−5

@@ -1786,13 +1786,21 @@ EXPORT_SYMBOL_GPL(__hrtimer_get_remaining); · 1785 unmodified lines
1786	1786	ktime_t hrtimer_get_next_event(void)
1787	1787	{
1788	1788	struct hrtimer_cpu_base *cpu_base = this_cpu_ptr(&hrtimer_bases);
1789		− ktime_t expires = KTIME_MAX;
1790	1789
1791		− guard(raw_spinlock_irqsave)(&cpu_base->lock);
1792		− if (!hrtimer_hres_active(cpu_base))
1793		− expires = __hrtimer_get_next_event(cpu_base, HRTIMER_ACTIVE_ALL);
	1790	+ /*
	1791	+ * When HRES is active cmp_next_hrtimer_event() expects KTIME_MAX.
	1792	+ *
	1793	+ * cpu_base->hres_active is written only by the local CPU in
	1794	+ * hrtimer_switch_to_hres() from hard interrupt context and in
	1795	+ * hrtimers_cpu_starting() during CPU bring-up, and all callers reach
	1796	+ * this with interrupts disabled on the same CPU, so an unlocked read is
	1797	+ * stable without holding the lock.
	1798	+ */
	1799	+ if (hrtimer_hres_active(cpu_base))
	1800	+ return KTIME_MAX;
1794	1801
1795		− return expires;
	1802	+ guard(raw_spinlock_irqsave)(&cpu_base->lock);
	1803	+ return __hrtimer_get_next_event(cpu_base, HRTIMER_ACTIVE_ALL);
1796	1804	}
1797	1805
1798	1806	/**