[dpdk-dev] [PATCH] net/mlx4: fix crash when configure is not called

[dpdk-dev] [PATCH] net/mlx4: fix crash when configure is not called

[Yongseok Koh]

Although uncommon, applications may destroy a device immediately after
probing it without going through dev_configure() first.

This patch addresses a crash which occurs when mlx4_dev_close() calls
mlx4_mr_release() due to an uninitialized entry in the private structure.

And the device's B-tree should be allocated once.

Fixes: 9797bfcce1c9 ("net/mlx4: add new memory region support")

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx4/mlx4.c | 31 +++++++++++++++++--------------
 1 file changed, 17 insertions(+), 14 deletions(-)

diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index 9f8ecd072..a29814b3b 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/mlx4.c
@@ -98,20 +98,6 @@ mlx4_dev_configure(struct rte_eth_dev *dev)
 	if (ret)
 		ERROR("%p: interrupt handler installation failed",
 		      (void *)dev);
-	/*
-	 * Once the device is added to the list of memory event callback, its
-	 * global MR cache table cannot be expanded on the fly because of
-	 * deadlock. If it overflows, lookup should be done by searching MR list
-	 * linearly, which is slow.
-	 */
-	if (mlx4_mr_btree_init(&priv->mr.cache, MLX4_MR_BTREE_CACHE_N * 2,
-			       dev->device->numa_node)) {
-		/* rte_errno is already set. */
-		return -rte_errno;
-	}
-	rte_rwlock_write_lock(&mlx4_mem_event_rwlock);
-	LIST_INSERT_HEAD(&mlx4_mem_event_cb_list, priv, mem_event_cb);
-	rte_rwlock_write_unlock(&mlx4_mem_event_rwlock);
 exit:
 	return ret;
 }
@@ -761,6 +747,23 @@ mlx4_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		/* Update link status once if waiting for LSC. */
 		if (eth_dev->data->dev_flags & RTE_ETH_DEV_INTR_LSC)
 			mlx4_link_update(eth_dev, 0);
+		/*
+		 * Once the device is added to the list of memory event
+		 * callback, its global MR cache table cannot be expanded
+		 * on the fly because of deadlock. If it overflows, lookup
+		 * should be done by searching MR list linearly, which is slow.
+		 */
+		err = mlx4_mr_btree_init(&priv->mr.cache,
+					 MLX4_MR_BTREE_CACHE_N * 2,
+					 eth_dev->device->numa_node);
+		if (err) {
+			/* rte_errno is already set. */
+			goto port_error;
+		}
+		/* Add device to memory callback list. */
+		rte_rwlock_write_lock(&mlx4_mem_event_rwlock);
+		LIST_INSERT_HEAD(&mlx4_mem_event_cb_list, priv, mem_event_cb);
+		rte_rwlock_write_unlock(&mlx4_mem_event_rwlock);
 		rte_eth_dev_probing_finish(eth_dev);
 		continue;
 port_error:
-- 
2.11.0

Re: [dpdk-dev] [PATCH] net/mlx4: fix crash when configure is not called

[Shahaf Shuler]

Sunday, May 27, 2018 10:05 AM, Yongseok Koh:
> Subject: [PATCH] net/mlx4: fix crash when configure is not called
> 
> Although uncommon, applications may destroy a device immediately after
> probing it without going through dev_configure() first.
> 
> This patch addresses a crash which occurs when mlx4_dev_close() calls
> mlx4_mr_release() due to an uninitialized entry in the private structure.
> 
> And the device's B-tree should be allocated once.
> 
> Fixes: 9797bfcce1c9 ("net/mlx4: add new memory region support")
> 
> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
> Signed-off-by: Xueming Li <xuemingl@mellanox.com>
> Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
> ---

Applied to next-net-mlx, thanks.