[PATCH 07/10] locking/qspinlock: Use smp_cond_load_relaxed to wait for next node

From: Will Deacon
Date: Thu Apr 05 2018 - 13:00:20 EST


When a locker reaches the head of the queue and takes the lock, a
concurrent locker may enqueue and force the lock holder to spin
whilst its node->next field is initialised. Rather than open-code
a READ_ONCE/cpu_relax() loop, this can be implemented using
smp_cond_load_relaxed instead.

Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Signed-off-by: Will Deacon <will.deacon@xxxxxxx>
---
kernel/locking/qspinlock.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/kernel/locking/qspinlock.c b/kernel/locking/qspinlock.c
index 291e1526d27b..c8b57d375b49 100644
--- a/kernel/locking/qspinlock.c
+++ b/kernel/locking/qspinlock.c
@@ -475,10 +475,8 @@ void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val)
/*
* contended path; wait for next if not observed yet, release.
*/
- if (!next) {
- while (!(next = READ_ONCE(node->next)))
- cpu_relax();
- }
+ if (!next)
+ next = smp_cond_load_relaxed(&node->next, (VAL));

arch_mcs_spin_unlock_contended(&next->locked);
pv_kick_node(lock, next);
--
2.1.4