patches for DPDK stable branches
 help / color / mirror / Atom feed
* [dpdk-stable] [PATCH v2] hash: fix lock release on add
       [not found] <1495801830-36849-1-git-send-email-mike.stolarchuk@bigswitch.com>
@ 2017-07-07  5:54 ` Pablo de Lara
  2017-07-08 17:00   ` [dpdk-stable] [dpdk-dev] " Thomas Monjalon
  0 siblings, 1 reply; 2+ messages in thread
From: Pablo de Lara @ 2017-07-07  5:54 UTC (permalink / raw)
  To: dev; +Cc: mstolarchuk, stable

From: mstolarchuk <mike.stolarchuk@bigswitch.com>

When adding items to a hash table with multiple threads,
there is an spinlock used to prevent data corruption
(unless Transactional Memory is supported).

If there is a failure, the spinlock should be released,
but there were cases where that was not happening.

Fixes: be856325cba3 ("hash: add scalable multi-writer insertion with Intel TSX")
CC: stable@dpdk.org

Signed-off-by: mstolarchuk <mike.stolarchuk@bigswitch.com>
Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---

Changes in v2:

- Fixed commit message

 lib/librte_hash/rte_cuckoo_hash.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/lib/librte_hash/rte_cuckoo_hash.c b/lib/librte_hash/rte_cuckoo_hash.c
index 80391cf..5b2b8dd 100644
--- a/lib/librte_hash/rte_cuckoo_hash.c
+++ b/lib/librte_hash/rte_cuckoo_hash.c
@@ -539,8 +539,10 @@ __rte_hash_add_key_with_hash(const struct rte_hash *h, const void *key,
 			n_slots = rte_ring_mc_dequeue_burst(h->free_slots,
 					cached_free_slots->objs,
 					LCORE_CACHE_SIZE, NULL);
-			if (n_slots == 0)
-				return -ENOSPC;
+			if (n_slots == 0) {
+				ret = -ENOSPC;
+				goto failure;
+			}
 
 			cached_free_slots->len += n_slots;
 		}
@@ -549,8 +551,10 @@ __rte_hash_add_key_with_hash(const struct rte_hash *h, const void *key,
 		cached_free_slots->len--;
 		slot_id = cached_free_slots->objs[cached_free_slots->len];
 	} else {
-		if (rte_ring_sc_dequeue(h->free_slots, &slot_id) != 0)
-			return -ENOSPC;
+		if (rte_ring_sc_dequeue(h->free_slots, &slot_id) != 0) {
+			ret = -ENOSPC;
+			goto failure;
+		}
 	}
 
 	new_k = RTE_PTR_ADD(keys, (uintptr_t)slot_id * h->key_entry_size);
@@ -660,6 +664,7 @@ __rte_hash_add_key_with_hash(const struct rte_hash *h, const void *key,
 	/* Error in addition, store new slot back in the ring and return error */
 	enqueue_slot_back(h, cached_free_slots, (void *)((uintptr_t) new_idx));
 
+failure:
 	if (h->add_key == ADD_KEY_MULTIWRITER)
 		rte_spinlock_unlock(h->multiwriter_lock);
 	return ret;
-- 
2.9.4

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [dpdk-stable] [dpdk-dev] [PATCH v2] hash: fix lock release on add
  2017-07-07  5:54 ` [dpdk-stable] [PATCH v2] hash: fix lock release on add Pablo de Lara
@ 2017-07-08 17:00   ` Thomas Monjalon
  0 siblings, 0 replies; 2+ messages in thread
From: Thomas Monjalon @ 2017-07-08 17:00 UTC (permalink / raw)
  To: Pablo de Lara, mstolarchuk; +Cc: dev, stable

07/07/2017 07:54, Pablo de Lara:
> From: mstolarchuk <mike.stolarchuk@bigswitch.com>
> 
> When adding items to a hash table with multiple threads,
> there is an spinlock used to prevent data corruption
> (unless Transactional Memory is supported).
> 
> If there is a failure, the spinlock should be released,
> but there were cases where that was not happening.
> 
> Fixes: be856325cba3 ("hash: add scalable multi-writer insertion with Intel TSX")
> CC: stable@dpdk.org
> 
> Signed-off-by: mstolarchuk <mike.stolarchuk@bigswitch.com>
> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>

Applied, thanks

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2017-07-08 17:00 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <1495801830-36849-1-git-send-email-mike.stolarchuk@bigswitch.com>
2017-07-07  5:54 ` [dpdk-stable] [PATCH v2] hash: fix lock release on add Pablo de Lara
2017-07-08 17:00   ` [dpdk-stable] [dpdk-dev] " Thomas Monjalon

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).