If register_candev fails, xcan_probe does not clean the napi
created by netif_napi_add.
Fix this by adding error handling code to clean napi when
register_candev fails.
Signed-off-by: Dongliang Mu <[email protected]>
---
drivers/net/can/xilinx_can.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/net/can/xilinx_can.c b/drivers/net/can/xilinx_can.c
index 3b883e607d8b..6ee0b5a8cdfc 100644
--- a/drivers/net/can/xilinx_can.c
+++ b/drivers/net/can/xilinx_can.c
@@ -1807,7 +1807,7 @@ static int xcan_probe(struct platform_device *pdev)
ret = register_candev(ndev);
if (ret) {
dev_err(&pdev->dev, "fail to register failed (err=%d)\n", ret);
- goto err_disableclks;
+ goto err_del_napi;
}
devm_can_led_init(ndev);
@@ -1825,6 +1825,8 @@ static int xcan_probe(struct platform_device *pdev)
return 0;
+err_del_napi:
+ netif_napi_del(&priv->napi);
err_disableclks:
pm_runtime_put(priv->dev);
pm_runtime_disable(&pdev->dev);
--
2.25.1
On Wed, Oct 13, 2021 at 12:04 PM Dongliang Mu <[email protected]> wrote:
>
> If register_candev fails, xcan_probe does not clean the napi
> created by netif_napi_add.
>
It seems the netif_napi_del operation is done in the free_candev
(free_netdev precisely).
list_for_each_entry_safe(p, n, &dev->napi_list, dev_list)
netif_napi_del(p);
And list_add_rcu(&napi->dev_list, &dev->napi_list) is done in the
netif_napi_add.
Therefore, I suggest removing "netif_napi_del" operation in the
xcan_remove to match probe and remove function.
> Fix this by adding error handling code to clean napi when
> register_candev fails.
>
> Signed-off-by: Dongliang Mu <[email protected]>
> ---
> drivers/net/can/xilinx_can.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/can/xilinx_can.c b/drivers/net/can/xilinx_can.c
> index 3b883e607d8b..6ee0b5a8cdfc 100644
> --- a/drivers/net/can/xilinx_can.c
> +++ b/drivers/net/can/xilinx_can.c
> @@ -1807,7 +1807,7 @@ static int xcan_probe(struct platform_device *pdev)
> ret = register_candev(ndev);
> if (ret) {
> dev_err(&pdev->dev, "fail to register failed (err=%d)\n", ret);
> - goto err_disableclks;
> + goto err_del_napi;
> }
>
> devm_can_led_init(ndev);
> @@ -1825,6 +1825,8 @@ static int xcan_probe(struct platform_device *pdev)
>
> return 0;
>
> +err_del_napi:
> + netif_napi_del(&priv->napi);
> err_disableclks:
> pm_runtime_put(priv->dev);
> pm_runtime_disable(&pdev->dev);
> --
> 2.25.1
>
On 13.10.2021 13:21:09, Dongliang Mu wrote:
> On Wed, Oct 13, 2021 at 12:04 PM Dongliang Mu <[email protected]> wrote:
> >
> > If register_candev fails, xcan_probe does not clean the napi
> > created by netif_napi_add.
> >
>
> It seems the netif_napi_del operation is done in the free_candev
> (free_netdev precisely).
>
> list_for_each_entry_safe(p, n, &dev->napi_list, dev_list)
> netif_napi_del(p);
>
> And list_add_rcu(&napi->dev_list, &dev->napi_list) is done in the
> netif_napi_add.
>
> Therefore, I suggest removing "netif_napi_del" operation in the
> xcan_remove to match probe and remove function.
Sounds reasonable, can you create a patch for this.
regards,
Marc
--
Pengutronix e.K. | Marc Kleine-Budde |
Embedded Linux | https://www.pengutronix.de |
Vertretung West/Dortmund | Phone: +49-231-2826-924 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
On Sun, Oct 17, 2021 at 8:36 PM Marc Kleine-Budde <[email protected]> wrote:
>
> On 13.10.2021 13:21:09, Dongliang Mu wrote:
> > On Wed, Oct 13, 2021 at 12:04 PM Dongliang Mu <[email protected]> wrote:
> > >
> > > If register_candev fails, xcan_probe does not clean the napi
> > > created by netif_napi_add.
> > >
> >
> > It seems the netif_napi_del operation is done in the free_candev
> > (free_netdev precisely).
> >
> > list_for_each_entry_safe(p, n, &dev->napi_list, dev_list)
> > netif_napi_del(p);
> >
> > And list_add_rcu(&napi->dev_list, &dev->napi_list) is done in the
> > netif_napi_add.
> >
> > Therefore, I suggest removing "netif_napi_del" operation in the
> > xcan_remove to match probe and remove function.
>
> Sounds reasonable, can you create a patch for this.
I have submitted one patch - https://lkml.org/lkml/2021/10/17/181
>
> regards,
> Marc
>
> --
> Pengutronix e.K. | Marc Kleine-Budde |
> Embedded Linux | https://www.pengutronix.de |
> Vertretung West/Dortmund | Phone: +49-231-2826-924 |
> Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
On 17.10.2021 20:52:14, Dongliang Mu wrote:
> On Sun, Oct 17, 2021 at 8:36 PM Marc Kleine-Budde <[email protected]> wrote:
> >
> > On 13.10.2021 13:21:09, Dongliang Mu wrote:
> > > On Wed, Oct 13, 2021 at 12:04 PM Dongliang Mu <[email protected]> wrote:
> > > >
> > > > If register_candev fails, xcan_probe does not clean the napi
> > > > created by netif_napi_add.
> > > >
> > >
> > > It seems the netif_napi_del operation is done in the free_candev
> > > (free_netdev precisely).
> > >
> > > list_for_each_entry_safe(p, n, &dev->napi_list, dev_list)
> > > netif_napi_del(p);
> > >
> > > And list_add_rcu(&napi->dev_list, &dev->napi_list) is done in the
> > > netif_napi_add.
> > >
> > > Therefore, I suggest removing "netif_napi_del" operation in the
> > > xcan_remove to match probe and remove function.
> >
> > Sounds reasonable, can you create a patch for this.
>
> I have submitted one patch - https://lkml.org/lkml/2021/10/17/181
Thanks for the patch.
Regards,
Marc
BTW: Do you know the new kernel.org mailing list archive available at
https://lore.kernel.org ?
You can reference a mail using its Message-ID, in you case it's:
https://lore.kernel.org/all/[email protected]
--
Pengutronix e.K. | Marc Kleine-Budde |
Embedded Linux | https://www.pengutronix.de |
Vertretung West/Dortmund | Phone: +49-231-2826-924 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |