[Patch v2] net/mlx4: fix verbs fd leak in the secondary process

[Patch v2] net/mlx4: fix verbs fd leak in the secondary process

[longli]

From: Long Li <longli@microsoft.com>

FDs passed from rte_mp_msg are duplicated to the secondary process and
need to be closed.

Fixes: 0203d33a10 ("net/mlx4: support secondary process")
Signed-off-by: Long Li <longli@microsoft.com>
---
Change in v2: handle error case where mlx4_proc_priv_init() might fail 

 drivers/net/mlx4/mlx4.c    | 9 ++++++---
 drivers/net/mlx4/mlx4_mp.c | 7 ++++++-
 2 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index 3f3c4a7c72..2e0b4a17e2 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/mlx4.c
@@ -877,6 +877,8 @@ mlx4_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		snprintf(name, sizeof(name), "%s port %u",
 			 mlx4_glue->get_device_name(ibv_dev), port);
 		if (rte_eal_process_type() == RTE_PROC_SECONDARY) {
+			int fd;
+
 			eth_dev = rte_eth_dev_attach_secondary(name);
 			if (eth_dev == NULL) {
 				ERROR("can not attach rte ethdev");
@@ -899,13 +901,14 @@ mlx4_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 			if (err)
 				goto err_secondary;
 			/* Receive command fd from primary process. */
-			err = mlx4_mp_req_verbs_cmd_fd(eth_dev);
-			if (err < 0) {
+			fd = mlx4_mp_req_verbs_cmd_fd(eth_dev);
+			if (fd < 0) {
 				err = rte_errno;
 				goto err_secondary;
 			}
 			/* Remap UAR for Tx queues. */
-			err = mlx4_tx_uar_init_secondary(eth_dev, err);
+			err = mlx4_tx_uar_init_secondary(eth_dev, fd);
+			close(fd);
 			if (err) {
 				err = rte_errno;
 				goto err_secondary;
diff --git a/drivers/net/mlx4/mlx4_mp.c b/drivers/net/mlx4/mlx4_mp.c
index 8fcfb5490e..b0bb48c8f1 100644
--- a/drivers/net/mlx4/mlx4_mp.c
+++ b/drivers/net/mlx4/mlx4_mp.c
@@ -5,6 +5,7 @@
 
 #include <stdio.h>
 #include <time.h>
+#include <unistd.h>
 
 #include <rte_eal.h>
 #include <ethdev_driver.h>
@@ -134,15 +135,19 @@ mp_secondary_handle(const struct rte_mp_msg *mp_msg, const void *peer)
 			mlx4_tx_uar_uninit_secondary(dev);
 			mlx4_proc_priv_uninit(dev);
 			ret = mlx4_proc_priv_init(dev);
-			if (ret)
+			if (ret) {
+				close(mp_msg->fds[0]);
 				return -rte_errno;
+			}
 			ret = mlx4_tx_uar_init_secondary(dev, mp_msg->fds[0]);
 			if (ret) {
+				close(mp_msg->fds[0]);
 				mlx4_proc_priv_uninit(dev);
 				return -rte_errno;
 			}
 		}
 #endif
+		close(mp_msg->fds[0]);
 		rte_mb();
 		mp_init_msg(dev, &mp_res, param->type);
 		res->result = 0;
-- 
2.17.1

[Patch v2] net/mlx5: fix verbs fd leak in the secondary process

[longli]

From: Long Li <longli@microsoft.com>

FDs passed from rte_mp_msg are duplicated to the secondary process and
need to be closed.

Fixes: 9a8ab29b84 ("net/mlx5: replace IPC socket with EAL API")
Signed-off-by: Long Li <longli@microsoft.com>
---
Change in v2: handle error case where mlx5_proc_priv_init() might fail

 drivers/net/mlx5/linux/mlx5_mp_os.c | 6 +++++-
 drivers/net/mlx5/linux/mlx5_os.c    | 8 +++++---
 2 files changed, 10 insertions(+), 4 deletions(-)

diff --git a/drivers/net/mlx5/linux/mlx5_mp_os.c b/drivers/net/mlx5/linux/mlx5_mp_os.c
index c448a3e9eb..0ba2208fe0 100644
--- a/drivers/net/mlx5/linux/mlx5_mp_os.c
+++ b/drivers/net/mlx5/linux/mlx5_mp_os.c
@@ -177,14 +177,18 @@ struct rte_mp_msg mp_res;
 			mlx5_tx_uar_uninit_secondary(dev);
 			mlx5_proc_priv_uninit(dev);
 			ret = mlx5_proc_priv_init(dev);
-			if (ret)
+			if (ret) {
+				close(mp_msg->fds[0]);
 				return -rte_errno;
+			}
 			ret = mlx5_tx_uar_init_secondary(dev, mp_msg->fds[0]);
 			if (ret) {
+				close(mp_msg->fds[0]);
 				mlx5_proc_priv_uninit(dev);
 				return -rte_errno;
 			}
 		}
+		close(mp_msg->fds[0]);
 		rte_mb();
 		mp_init_msg(&priv->mp_id, &mp_res, param->type);
 		res->result = 0;
diff --git a/drivers/net/mlx5/linux/mlx5_os.c b/drivers/net/mlx5/linux/mlx5_os.c
index c29fe3d92b..69ae3c944b 100644
--- a/drivers/net/mlx5/linux/mlx5_os.c
+++ b/drivers/net/mlx5/linux/mlx5_os.c
@@ -933,6 +933,7 @@ mlx5_dev_spawn(struct rte_device *dpdk_dev,
 	DRV_LOG(DEBUG, "naming Ethernet device \"%s\"", name);
 	if (rte_eal_process_type() == RTE_PROC_SECONDARY) {
 		struct mlx5_mp_id mp_id;
+		int fd;
 
 		eth_dev = rte_eth_dev_attach_secondary(name);
 		if (eth_dev == NULL) {
@@ -949,11 +950,12 @@ mlx5_dev_spawn(struct rte_device *dpdk_dev,
 			return NULL;
 		mlx5_mp_id_init(&mp_id, eth_dev->data->port_id);
 		/* Receive command fd from primary process */
-		err = mlx5_mp_req_verbs_cmd_fd(&mp_id);
-		if (err < 0)
+		fd = mlx5_mp_req_verbs_cmd_fd(&mp_id);
+		if (fd < 0)
 			goto err_secondary;
 		/* Remap UAR for Tx queues. */
-		err = mlx5_tx_uar_init_secondary(eth_dev, err);
+		err = mlx5_tx_uar_init_secondary(eth_dev, fd);
+		close(fd);
 		if (err)
 			goto err_secondary;
 		/*
-- 
2.17.1

RE: [Patch v2] net/mlx5: fix verbs fd leak in the secondary process

[Slava Ovsiienko]

> -----Original Message-----
> From: longli@linuxonhyperv.com <longli@linuxonhyperv.com>
> Sent: Wednesday, July 6, 2022 20:49
> To: Matan Azrad <matan@nvidia.com>; Ferruh Yigit <ferruh.yigit@xilinx.com>
> Cc: dev@dpdk.org; Karanjot Singh <ksingh1@paloaltonetworks.com>; NBU-
> Contact-longli (EXTERNAL) <longli@microsoft.com>
> Subject: [Patch v2] net/mlx5: fix verbs fd leak in the secondary process
> 
> From: Long Li <longli@microsoft.com>
> 
> FDs passed from rte_mp_msg are duplicated to the secondary process and need
> to be closed.
> 
> Fixes: 9a8ab29b84 ("net/mlx5: replace IPC socket with EAL API")
> Signed-off-by: Long Li <longli@microsoft.com>
> ---
> Change in v2: handle error case where mlx5_proc_priv_init() might fail
> 
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>

PS. Nice catch, thank you. Nice catch, thank you, Long

RE: [Patch v2] net/mlx4: fix verbs fd leak in the secondary process

[Slava Ovsiienko]

> -----Original Message-----
> From: longli@linuxonhyperv.com <longli@linuxonhyperv.com>
> Sent: Wednesday, July 6, 2022 20:49
> To: Matan Azrad <matan@nvidia.com>; Ferruh Yigit <ferruh.yigit@xilinx.com>
> Cc: dev@dpdk.org; Karanjot Singh <ksingh1@paloaltonetworks.com>; NBU-
> Contact-longli (EXTERNAL) <longli@microsoft.com>
> Subject: [Patch v2] net/mlx4: fix verbs fd leak in the secondary process
> 
> From: Long Li <longli@microsoft.com>
> 
> FDs passed from rte_mp_msg are duplicated to the secondary process and need
> to be closed.
> 
> Fixes: 0203d33a10 ("net/mlx4: support secondary process")
> Signed-off-by: Long Li <longli@microsoft.com>
> ---
> Change in v2: handle error case where mlx4_proc_priv_init() might fail
>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>

RE: [Patch v2] net/mlx4: fix verbs fd leak in the secondary process

[Raslan Darawsheh]

Hi,

> -----Original Message-----
> From: longli@linuxonhyperv.com <longli@linuxonhyperv.com>
> Sent: Wednesday, July 6, 2022 8:49 PM
> To: Matan Azrad <matan@nvidia.com>; Ferruh Yigit
> <ferruh.yigit@xilinx.com>
> Cc: dev@dpdk.org; Karanjot Singh <ksingh1@paloaltonetworks.com>; NBU-
> Contact-longli (EXTERNAL) <longli@microsoft.com>
> Subject: [Patch v2] net/mlx4: fix verbs fd leak in the secondary process
> 
> From: Long Li <longli@microsoft.com>
> 
> FDs passed from rte_mp_msg are duplicated to the secondary process and
> need to be closed.
> 
> Fixes: 0203d33a10 ("net/mlx4: support secondary process")
> Signed-off-by: Long Li <longli@microsoft.com>

Patch applied to next-net-mlx,

Kindest regards,
Raslan Darawsheh

RE: [Patch v2] net/mlx5: fix verbs fd leak in the secondary process

[Raslan Darawsheh]

Hi,

> -----Original Message-----
> From: longli@linuxonhyperv.com <longli@linuxonhyperv.com>
> Sent: Wednesday, July 6, 2022 8:49 PM
> To: Matan Azrad <matan@nvidia.com>; Ferruh Yigit
> <ferruh.yigit@xilinx.com>
> Cc: dev@dpdk.org; Karanjot Singh <ksingh1@paloaltonetworks.com>; NBU-
> Contact-longli (EXTERNAL) <longli@microsoft.com>
> Subject: [Patch v2] net/mlx5: fix verbs fd leak in the secondary process
> 
> From: Long Li <longli@microsoft.com>
> 
> FDs passed from rte_mp_msg are duplicated to the secondary process and
> need to be closed.
> 
> Fixes: 9a8ab29b84 ("net/mlx5: replace IPC socket with EAL API")
> Signed-off-by: Long Li <longli@microsoft.com>

Patch applied to next-net-mlx,

Kindest regards,
Raslan Darawsheh