[PATCH net] eth: gve: add missing netdev locks on reset and shutdown paths

[PATCH net] eth: gve: add missing netdev locks on reset and shutdown paths

[Jakub Kicinski]

All the misc entry points end up calling into either gve_open()
or gve_close(), they take rtnl_lock today but since the recent
instance locking changes should also take the instance lock.

Found by code inspection and untested.

Fixes: cae03e5bdd9e ("net: hold netdev instance lock during queue operations")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
CC: jeroendb@google.com
CC: hramamurthy@google.com
CC: pkaligineedi@google.com
CC: willemb@google.com
CC: shailend@google.com
CC: joshwash@google.com
CC: sdf@fomichev.me
---
 drivers/net/ethernet/google/gve/gve_main.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c
index cb2f9978f45e..f9a73c956861 100644
--- a/drivers/net/ethernet/google/gve/gve_main.c
+++ b/drivers/net/ethernet/google/gve/gve_main.c
@@ -2077,7 +2077,9 @@ static void gve_handle_reset(struct gve_priv *priv)
 
 	if (gve_get_do_reset(priv)) {
 		rtnl_lock();
+		netdev_lock(priv->dev);
 		gve_reset(priv, false);
+		netdev_unlock(priv->dev);
 		rtnl_unlock();
 	}
 }
@@ -2714,6 +2716,7 @@ static void gve_shutdown(struct pci_dev *pdev)
 	bool was_up = netif_running(priv->dev);
 
 	rtnl_lock();
+	netdev_lock(netdev);
 	if (was_up && gve_close(priv->dev)) {
 		/* If the dev was up, attempt to close, if close fails, reset */
 		gve_reset_and_teardown(priv, was_up);
@@ -2721,6 +2724,7 @@ static void gve_shutdown(struct pci_dev *pdev)
 		/* If the dev wasn't up or close worked, finish tearing down */
 		gve_teardown_priv_resources(priv);
 	}
+	netdev_unlock(netdev);
 	rtnl_unlock();
 }
 
-- 
2.49.0

Re: [PATCH net] eth: gve: add missing netdev locks on reset and shutdown paths

[Stanislav Fomichev]

On 03/28, Jakub Kicinski wrote:
> All the misc entry points end up calling into either gve_open()
> or gve_close(), they take rtnl_lock today but since the recent
> instance locking changes should also take the instance lock.
> 
> Found by code inspection and untested.
> 
> Fixes: cae03e5bdd9e ("net: hold netdev instance lock during queue operations")
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Acked-by: Stanislav Fomichev <sdf@fomichev.me>

Re: [PATCH net] eth: gve: add missing netdev locks on reset and shutdown paths

[Harshitha Ramamurthy]

On Fri, Mar 28, 2025 at 9:47 AM Jakub Kicinski <kuba@kernel.org> wrote:
>
> All the misc entry points end up calling into either gve_open()
> or gve_close(), they take rtnl_lock today but since the recent
> instance locking changes should also take the instance lock.
>
> Found by code inspection and untested.
>
> Fixes: cae03e5bdd9e ("net: hold netdev instance lock during queue operations")
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Reviewed-by: Harshitha Ramamurthy <hramamurthy@google.com>
> ---
> CC: jeroendb@google.com
> CC: hramamurthy@google.com
> CC: pkaligineedi@google.com
> CC: willemb@google.com
> CC: shailend@google.com
> CC: joshwash@google.com
> CC: sdf@fomichev.me
> ---
>  drivers/net/ethernet/google/gve/gve_main.c | 4 ++++
>  1 file changed, 4 insertions(+)
>
> diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c
> index cb2f9978f45e..f9a73c956861 100644
> --- a/drivers/net/ethernet/google/gve/gve_main.c
> +++ b/drivers/net/ethernet/google/gve/gve_main.c
> @@ -2077,7 +2077,9 @@ static void gve_handle_reset(struct gve_priv *priv)
>
>         if (gve_get_do_reset(priv)) {
>                 rtnl_lock();
> +               netdev_lock(priv->dev);
>                 gve_reset(priv, false);
> +               netdev_unlock(priv->dev);
>                 rtnl_unlock();
>         }
>  }
> @@ -2714,6 +2716,7 @@ static void gve_shutdown(struct pci_dev *pdev)
>         bool was_up = netif_running(priv->dev);
>
>         rtnl_lock();
> +       netdev_lock(netdev);
>         if (was_up && gve_close(priv->dev)) {
>                 /* If the dev was up, attempt to close, if close fails, reset */
>                 gve_reset_and_teardown(priv, was_up);
> @@ -2721,6 +2724,7 @@ static void gve_shutdown(struct pci_dev *pdev)
>                 /* If the dev wasn't up or close worked, finish tearing down */
>                 gve_teardown_priv_resources(priv);
>         }
> +       netdev_unlock(netdev);
>         rtnl_unlock();
>  }
>
> --
> 2.49.0
>

Re: [PATCH net] eth: gve: add missing netdev locks on reset and shutdown paths

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Fri, 28 Mar 2025 09:47:42 -0700 you wrote:
> All the misc entry points end up calling into either gve_open()
> or gve_close(), they take rtnl_lock today but since the recent
> instance locking changes should also take the instance lock.
> 
> Found by code inspection and untested.
> 
> Fixes: cae03e5bdd9e ("net: hold netdev instance lock during queue operations")
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
> 
> [...]

Here is the summary with links:
  - [net] eth: gve: add missing netdev locks on reset and shutdown paths
    https://git.kernel.org/netdev/net/c/9e3267cf02c2

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html