drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c | 3 +++ 1 file changed, 3 insertions(+)
of_device_get_match_data() can return NULL if of_match_device failed, and
the pointer 'data' was dereferenced without checking against NULL. Add
checking of pointer 'data' in qcom_ethqos_probe().
Cc: stable@vger.kernel.org
Fixes: a7c30e62d4b8 ("net: stmmac: Add driver for Qualcomm ethqos")
Signed-off-by: Ma Ke <make24@iscas.ac.cn>
---
drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
index 901a3c1959fa..f18393fe58a4 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
@@ -838,6 +838,9 @@ static int qcom_ethqos_probe(struct platform_device *pdev)
ethqos->mac_base = stmmac_res.addr;
data = of_device_get_match_data(dev);
+ if (!data)
+ return -ENODEV;
+
ethqos->por = data->por;
ethqos->num_por = data->num_por;
ethqos->rgmii_config_loopback_en = data->rgmii_config_loopback_en;
--
2.25.1
On Wed, Aug 21, 2024 at 09:19:49PM +0800, Ma Ke wrote:
> of_device_get_match_data() can return NULL if of_match_device failed, and
> the pointer 'data' was dereferenced without checking against NULL. Add
> checking of pointer 'data' in qcom_ethqos_probe().
>
> Cc: stable@vger.kernel.org
> Fixes: a7c30e62d4b8 ("net: stmmac: Add driver for Qualcomm ethqos")
> Signed-off-by: Ma Ke <make24@iscas.ac.cn>
Hi Ma Ke,
There is probably no need to repost just because of this.
But as a fix for Networking code it should be targeted at the net tree.
Subject: [PATCH net] ...
> ---
> drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
> index 901a3c1959fa..f18393fe58a4 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
> @@ -838,6 +838,9 @@ static int qcom_ethqos_probe(struct platform_device *pdev)
> ethqos->mac_base = stmmac_res.addr;
>
> data = of_device_get_match_data(dev);
> + if (!data)
> + return -ENODEV;
> +
In this function dev_err_probe() is used, I assume in cases
where a function that returns an error does not emit any logs.
For consistency, perhaps that is appropriate here too?
> ethqos->por = data->por;
> ethqos->num_por = data->num_por;
> ethqos->rgmii_config_loopback_en = data->rgmii_config_loopback_en;
> --
> 2.25.1
On Wed, Aug 21, 2024 at 06:37:30PM +0100, Simon Horman wrote:
> On Wed, Aug 21, 2024 at 09:19:49PM +0800, Ma Ke wrote:
> > of_device_get_match_data() can return NULL if of_match_device failed, and
> > the pointer 'data' was dereferenced without checking against NULL. Add
> > checking of pointer 'data' in qcom_ethqos_probe().
> >
> > Cc: stable@vger.kernel.org
> > Fixes: a7c30e62d4b8 ("net: stmmac: Add driver for Qualcomm ethqos")
> > Signed-off-by: Ma Ke <make24@iscas.ac.cn>
>
> Hi Ma Ke,
>
> There is probably no need to repost just because of this.
> But as a fix for Networking code it should be targeted at the net tree.
>
> Subject: [PATCH net] ...
>
> > ---
> > drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c | 3 +++
> > 1 file changed, 3 insertions(+)
> >
> > diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
> > index 901a3c1959fa..f18393fe58a4 100644
> > --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
> > +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
> > @@ -838,6 +838,9 @@ static int qcom_ethqos_probe(struct platform_device *pdev)
> > ethqos->mac_base = stmmac_res.addr;
> >
> > data = of_device_get_match_data(dev);
> > + if (!data)
> > + return -ENODEV;
> > +
>
> In this function dev_err_probe() is used, I assume in cases
> where a function that returns an error does not emit any logs.
>
> For consistency, perhaps that is appropriate here too?
>
Unless I'm missing something here this function can only ever be invoked
by a match against one of the entries in qcom_ethqos_match[], which all
of them have a non-NULL data pointer.
As such, if we somehow arrive here with data of NULL, the NULL pointer
dereference on the next line will provide a welcome large splat and a
callstack indicating that we have a problem.
If there's some use case I'm missing, I would prefer if this was
documented in the commit message.
Regards,
Bjorn
> > ethqos->por = data->por;
> > ethqos->num_por = data->num_por;
> > ethqos->rgmii_config_loopback_en = data->rgmii_config_loopback_en;
> > --
> > 2.25.1
© 2016 - 2026 Red Hat, Inc.