ptr_ring: prevent queue load/store tearing
In theory compiler could tear queue loads or stores in two. It does not seem to be happening in practice but it seems easier to convert the cases where this would be a problem to READ/WRITE_ONCE than worry about it. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
parent
f417dc2818
commit
a07d29c672
1 changed files with 2 additions and 2 deletions
|
@ -114,7 +114,7 @@ static inline int __ptr_ring_produce(struct ptr_ring *r, void *ptr)
|
|||
/* Pairs with smp_read_barrier_depends in __ptr_ring_consume. */
|
||||
smp_wmb();
|
||||
|
||||
r->queue[r->producer++] = ptr;
|
||||
WRITE_ONCE(r->queue[r->producer++], ptr);
|
||||
if (unlikely(r->producer >= r->size))
|
||||
r->producer = 0;
|
||||
return 0;
|
||||
|
@ -173,7 +173,7 @@ static inline int ptr_ring_produce_bh(struct ptr_ring *r, void *ptr)
|
|||
static inline void *__ptr_ring_peek(struct ptr_ring *r)
|
||||
{
|
||||
if (likely(r->size))
|
||||
return r->queue[r->consumer_head];
|
||||
return READ_ONCE(r->queue[r->consumer_head]);
|
||||
return NULL;
|
||||
}
|
||||
|
||||
|
|
Loading…
Reference in a new issue