[dpdk-dev] [PATCH v2] lib/efd: fix to free tail queue entry after use

[Maxime Coquelin]

On 11/14/18 12:17 PM, Hari Kumar Vemula wrote:
> In rte_efd_create() allocated memory for tail queue entry but
> not freed.
> Added freeing the tail queue entry.
> 
> Fixes: 56b6ef874f80 ("efd: new Elastic Flow Distributor library")
> Cc: stable at dpdk.org
> 
> Signed-off-by: Hari Kumar Vemula <hari.kumarx.vemula at intel.com>
> Acked-by: Reshma Pattan <reshma.pattan at intel.com>
> 
> ---
> v2: Updated commit message.
> ---
> 
>   lib/librte_efd/rte_efd.c | 21 +++++++++++++++++++++
>   1 file changed, 21 insertions(+)
> 
> diff --git a/lib/librte_efd/rte_efd.c b/lib/librte_efd/rte_efd.c
> index a780e2fe8..f8c6c447f 100644
> --- a/lib/librte_efd/rte_efd.c
> +++ b/lib/librte_efd/rte_efd.c
> @@ -739,17 +739,38 @@ void
>   rte_efd_free(struct rte_efd_table *table)
>   {
>   	uint8_t socket_id;
> +	struct rte_efd_list *efd_list = NULL;

NULL init seems useless here.

> +	struct rte_tailq_entry *te;
>   
>   	if (table == NULL)
>   		return;
>   
> +	efd_list = RTE_TAILQ_CAST(rte_efd_tailq.head, rte_efd_list);
> +
>   	for (socket_id = 0; socket_id < RTE_MAX_NUMA_NODES; socket_id++)
>   		rte_free(table->chunks[socket_id]);
>   
> +	rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK);
> +
> +	/* find our tailq entry */
> +	TAILQ_FOREACH(te, efd_list, next) {
> +		if (te->data == (void *) table)
> +			break;
> +	}
> +
> +	if (te == NULL) {
> +		rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
> +		return;
> +	}
> +
> +	TAILQ_REMOVE(efd_list, te, next);
> +	rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);

Wouldn't this be simpler by using TAILQ_FOREACH_SAFE() instead?
The element could removed from the list and freed within the loop.

> +
>   	rte_ring_free(table->free_slots);
>   	rte_free(table->offline_chunks);
>   	rte_free(table->keys);
>   	rte_free(table);
> +	rte_free(te);
>   }
>   
>   /**
>