diff mbox series

[net,01/12] net/mlx5e: Hold reference on mirred devices while accessing them

Message ID 20200728091035.112067-2-saeedm@mellanox.com
State Changes Requested
Delegated to: David Miller
Headers show
Series [net,01/12] net/mlx5e: Hold reference on mirred devices while accessing them | expand

Commit Message

Saeed Mahameed July 28, 2020, 9:10 a.m. UTC 
From: Eli Cohen <eli@mellanox.com>

Net devices might be removed. For example, a vxlan device could be
deleted and its ifnidex would become invalid. Use dev_get_by_index()
instead of __dev_get_by_index() to hold reference on the device while
accessing it and release after done.

Fixes: 3c37745ec614 ("net/mlx5e: Properly deal with encap flows add/del under neigh update")
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Vlad Buslov <vladbu@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

Comments

Or Gerlitz July 28, 2020, 10:27 a.m. UTC | #1 
On Tue, Jul 28, 2020 at 12:13 PM Saeed Mahameed <saeedm@mellanox.com> wrote:
> From: Eli Cohen <eli@mellanox.com>
>
> Net devices might be removed. For example, a vxlan device could be
> deleted and its ifnidex would become invalid. Use dev_get_by_index()
> instead of __dev_get_by_index() to hold reference on the device while
> accessing it and release after done.

haven't this patch sent in the past in the same form and we were in the middle
of discussing how to properly address this? if something changed, what?
Saeed Mahameed July 28, 2020, 7:32 p.m. UTC | #2 
On Tue, 2020-07-28 at 13:27 +0300, Or Gerlitz wrote:
> On Tue, Jul 28, 2020 at 12:13 PM Saeed Mahameed <saeedm@mellanox.com>
> wrote:
> > From: Eli Cohen <eli@mellanox.com>
> > 
> > Net devices might be removed. For example, a vxlan device could be
> > deleted and its ifnidex would become invalid. Use
> > dev_get_by_index()
> > instead of __dev_get_by_index() to hold reference on the device
> > while
> > accessing it and release after done.
> 
> haven't this patch sent in the past in the same form and we were in
> the middle
> of discussing how to properly address this? if something changed,
> what?

I thought the discussion was concluded ? 

anyway i will remove this patch from my trees, Eli please discuss with
Or, i am fine with this patch as is since it is a minimal fix to net.
diff mbox series

Patch

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index cc8412151ca09..518957d82b1ca 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1327,10 +1327,14 @@  mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
 			continue;
 
 		mirred_ifindex = parse_attr->mirred_ifindex[out_index];
-		out_dev = __dev_get_by_index(dev_net(priv->netdev),
-					     mirred_ifindex);
+		out_dev = dev_get_by_index(dev_net(priv->netdev),
+					   mirred_ifindex);
+		if (!out_dev)
+			return -ENODEV;
+
 		err = mlx5e_attach_encap(priv, flow, out_dev, out_index,
 					 extack, &encap_dev, &encap_valid);
+		dev_put(out_dev);
 		if (err)
 			return err;