[PATCH] qib_qp: Replace rcu_assign_pointer() with RCU_INIT_POINTER()

[PATCH] qib_qp: Replace rcu_assign_pointer() with RCU_INIT_POINTER()

[Andreea-Cristina Bernat]

The uses of "rcu_assign_pointer()" are NULLing out the pointers.
According to RCU_INIT_POINTER()'s block comment:
"1.   This use of RCU_INIT_POINTER() is NULLing out the pointer"
it is better to use it instead of rcu_assign_pointer() because it has a
smaller overhead.

The following Coccinelle semantic patch was used:
@@
@@

- rcu_assign_pointer
+ RCU_INIT_POINTER
  (..., NULL)

Signed-off-by: Andreea-Cristina Bernat <bernat.ada@gmail.com>
---
 drivers/infiniband/hw/qib/qib_qp.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/qib/qib_qp.c b/drivers/infiniband/hw/qib/qib_qp.c
index 7fcc150..17daf42 100644
--- a/drivers/infiniband/hw/qib/qib_qp.c
+++ b/drivers/infiniband/hw/qib/qib_qp.c
@@ -255,10 +255,10 @@ static void remove_qp(struct qib_ibdev *dev, struct qib_qp *qp)
 
 	if (rcu_dereference_protected(ibp->qp0,
 			lockdep_is_held(&dev->qpt_lock)) == qp) {
-		rcu_assign_pointer(ibp->qp0, NULL);
+		RCU_INIT_POINTER(ibp->qp0, NULL);
 	} else if (rcu_dereference_protected(ibp->qp1,
 			lockdep_is_held(&dev->qpt_lock)) == qp) {
-		rcu_assign_pointer(ibp->qp1, NULL);
+		RCU_INIT_POINTER(ibp->qp1, NULL);
 	} else {
 		struct qib_qp *q;
 		struct qib_qp __rcu **qpp;
@@ -315,7 +315,7 @@ unsigned qib_free_all_qps(struct qib_devdata *dd)
 	for (n = 0; n < dev->qp_table_size; n++) {
 		qp = rcu_dereference_protected(dev->qp_table[n],
 			lockdep_is_held(&dev->qpt_lock));
-		rcu_assign_pointer(dev->qp_table[n], NULL);
+		RCU_INIT_POINTER(dev->qp_table[n], NULL);
 
 		for (; qp; qp = rcu_dereference_protected(qp->next,
 					lockdep_is_held(&dev->qpt_lock)))
-- 
1.9.1

RE: [PATCH] qib_qp: Replace rcu_assign_pointer() with RCU_INIT_POINTER()

[Marciniszyn, Mike]

> Subject: [PATCH] qib_qp: Replace rcu_assign_pointer() with
> RCU_INIT_POINTER()
> 

I would prefer the summary line be:
IB/qib: remove_qp() use RCU_INIT_POINTER() instead of rcu_assign_pointer()

Otherwise I agree with the patch and I have tested it with large numbers of QPs.

Mike