[PATCH] can: mpc5xxx_can: check of_iomap return before use

 of_iompa() can return NULL so that return needs to be checked and NULL
treated as failure. While at it also take care of the missing
of_node_put() in the error path.

Signed-off-by: Nicholas Mc Guire <redacted>
Fixes: commit afa17a500a36 ("net/can: add driver for mscan family & mpc52xx_mscan")
---

Problem was found by an experimental coccinelle script

Patch was compiletested with: mpc5200_defconfig + CONFIG_CAN=y,
CONFIG_CAN_MSCAN=y, CONFIG_CAN_MPC5XXX=y
(with a number of sparse warnings not related to the proposed change)

Patch is against 4.18-rc3 (localversion-next is next-20180706)

 drivers/net/can/mscan/mpc5xxx_can.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/can/mscan/mpc5xxx_can.c b/drivers/net/can/mscan/mpc5xxx_can.c
index c7427bd..2949a38 100644
--- a/drivers/net/can/mscan/mpc5xxx_can.c
+++ b/drivers/net/can/mscan/mpc5xxx_can.c

@@ -86,6 +86,11 @@ static u32 mpc52xx_can_get_clock(struct platform_device *ofdev,
 		return 0;
 	}
 	cdm = of_iomap(np_cdm, 0);
+	if (!cdm) {
+		of_node_put(np_cdm);
+		dev_err(&ofdev->dev, "can't map clock node!\n");
+		return 0;
+	}
 
 	if (in_8(&cdm->ipb_clk_sel) & 0x1)
 		freq *= 2;

-- 
2.1.4

On Mon, Jul 9, 2018 at 4:16 PM, Nicholas Mc Guire [off-list ref] wrote:

 of_iompa() can return NULL so that return needs to be checked and NULL

s/of_iompa/of_iomap/

I think you should return an error code here. -ENOMEM maybe?

On Mon, Jul 09, 2018 at 04:28:41PM -0300, Fabio Estevam wrote:

On Mon, Jul 9, 2018 at 4:16 PM, Nicholas Mc Guire [off-list ref] wrote:

quoted

 of_iompa() can return NULL so that return needs to be checked and NULL

s/of_iompa/of_iomap/

sorry -  thats a stupid one.

quoted

treated as failure. While at it also take care of the missing
of_node_put() in the error path.

Signed-off-by: Nicholas Mc Guire <redacted>
Fixes: commit afa17a500a36 ("net/can: add driver for mscan family & mpc52xx_mscan")
---

Problem was found by an experimental coccinelle script

Patch was compiletested with: mpc5200_defconfig + CONFIG_CAN=y,
CONFIG_CAN_MSCAN=y, CONFIG_CAN_MPC5XXX=y
(with a number of sparse warnings not related to the proposed change)

Patch is against 4.18-rc3 (localversion-next is next-20180706)

 drivers/net/can/mscan/mpc5xxx_can.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/can/mscan/mpc5xxx_can.c b/drivers/net/can/mscan/mpc5xxx_can.c
index c7427bd..2949a38 100644
--- a/drivers/net/can/mscan/mpc5xxx_can.c
+++ b/drivers/net/can/mscan/mpc5xxx_can.c

@@ -86,6 +86,11 @@ static u32 mpc52xx_can_get_clock(struct platform_device *ofdev,
                return 0;
        }
        cdm = of_iomap(np_cdm, 0);
+       if (!cdm) {
+               of_node_put(np_cdm);
+               dev_err(&ofdev->dev, "can't map clock node!\n");
+               return 0;

I think you should return an error code here. -ENOMEM maybe?

I don´t think so the code above this part return 0 on error and a
valid frequency on success so returning -ENOMEM would probably be
mistaken as a frequency !

thx!
hofrat

On Mon, Jul 9, 2018 at 4:31 PM, Nicholas Mc Guire [off-list ref] wrote:

I don´t think so the code above this part return 0 on error and a
valid frequency on success so returning -ENOMEM would probably be
mistaken as a frequency !

Now that I looked closely in the function context I see it makes sense
to return 0 :-)

Thanks

On 07/09/2018 09:28 PM, Fabio Estevam wrote:

On Mon, Jul 9, 2018 at 4:16 PM, Nicholas Mc Guire [off-list ref] wrote:

quoted

 of_iompa() can return NULL so that return needs to be checked and NULL

s/of_iompa/of_iomap/

quoted

treated as failure. While at it also take care of the missing
of_node_put() in the error path.

Signed-off-by: Nicholas Mc Guire <redacted>
Fixes: commit afa17a500a36 ("net/can: add driver for mscan family & mpc52xx_mscan")
---

Problem was found by an experimental coccinelle script

Patch was compiletested with: mpc5200_defconfig + CONFIG_CAN=y,
CONFIG_CAN_MSCAN=y, CONFIG_CAN_MPC5XXX=y
(with a number of sparse warnings not related to the proposed change)

Patch is against 4.18-rc3 (localversion-next is next-20180706)

 drivers/net/can/mscan/mpc5xxx_can.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/can/mscan/mpc5xxx_can.c b/drivers/net/can/mscan/mpc5xxx_can.c
index c7427bd..2949a38 100644
--- a/drivers/net/can/mscan/mpc5xxx_can.c
+++ b/drivers/net/can/mscan/mpc5xxx_can.c

@@ -86,6 +86,11 @@ static u32 mpc52xx_can_get_clock(struct platform_device *ofdev,
                return 0;
        }
        cdm = of_iomap(np_cdm, 0);
+       if (!cdm) {
+               of_node_put(np_cdm);
+               dev_err(&ofdev->dev, "can't map clock node!\n");
+               return 0;

I think you should return an error code here. -ENOMEM maybe?

Can you send an updated version?

Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |

On Mon, Jul 23, 2018 at 01:37:05PM +0200, Marc Kleine-Budde wrote:

On 07/09/2018 09:28 PM, Fabio Estevam wrote:

quoted

On Mon, Jul 9, 2018 at 4:16 PM, Nicholas Mc Guire [off-list ref] wrote:

quoted

 of_iompa() can return NULL so that return needs to be checked and NULL

s/of_iompa/of_iomap/

quoted

treated as failure. While at it also take care of the missing
of_node_put() in the error path.

Signed-off-by: Nicholas Mc Guire <redacted>
Fixes: commit afa17a500a36 ("net/can: add driver for mscan family & mpc52xx_mscan")
---

Problem was found by an experimental coccinelle script

Patch was compiletested with: mpc5200_defconfig + CONFIG_CAN=y,
CONFIG_CAN_MSCAN=y, CONFIG_CAN_MPC5XXX=y
(with a number of sparse warnings not related to the proposed change)

Patch is against 4.18-rc3 (localversion-next is next-20180706)

 drivers/net/can/mscan/mpc5xxx_can.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/can/mscan/mpc5xxx_can.c b/drivers/net/can/mscan/mpc5xxx_can.c
index c7427bd..2949a38 100644
--- a/drivers/net/can/mscan/mpc5xxx_can.c
+++ b/drivers/net/can/mscan/mpc5xxx_can.c

@@ -86,6 +86,11 @@ static u32 mpc52xx_can_get_clock(struct platform_device *ofdev,
                return 0;
        }
        cdm = of_iomap(np_cdm, 0);
+       if (!cdm) {
+               of_node_put(np_cdm);
+               dev_err(&ofdev->dev, "can't map clock node!\n");
+               return 0;

I think you should return an error code here. -ENOMEM maybe?

Can you send an updated version?

The way I understand the code the return value will be used
as a frequency - the only valid error valeu thus is 0 here,
so I think the return 0 is correct.

thx!
hofrat

On 07/23/2018 01:57 PM, Nicholas Mc Guire wrote:

On Mon, Jul 23, 2018 at 01:37:05PM +0200, Marc Kleine-Budde wrote:

quoted

On 07/09/2018 09:28 PM, Fabio Estevam wrote:

quoted

On Mon, Jul 9, 2018 at 4:16 PM, Nicholas Mc Guire [off-list ref] wrote:

quoted

 of_iompa() can return NULL so that return needs to be checked and NULL

s/of_iompa/of_iomap/

quoted

treated as failure. While at it also take care of the missing
of_node_put() in the error path.

Signed-off-by: Nicholas Mc Guire <redacted>
Fixes: commit afa17a500a36 ("net/can: add driver for mscan family & mpc52xx_mscan")
---

Problem was found by an experimental coccinelle script

Patch was compiletested with: mpc5200_defconfig + CONFIG_CAN=y,
CONFIG_CAN_MSCAN=y, CONFIG_CAN_MPC5XXX=y
(with a number of sparse warnings not related to the proposed change)

Patch is against 4.18-rc3 (localversion-next is next-20180706)

 drivers/net/can/mscan/mpc5xxx_can.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/can/mscan/mpc5xxx_can.c b/drivers/net/can/mscan/mpc5xxx_can.c
index c7427bd..2949a38 100644
--- a/drivers/net/can/mscan/mpc5xxx_can.c
+++ b/drivers/net/can/mscan/mpc5xxx_can.c

@@ -86,6 +86,11 @@ static u32 mpc52xx_can_get_clock(struct platform_device *ofdev,
                return 0;
        }
        cdm = of_iomap(np_cdm, 0);
+       if (!cdm) {
+               of_node_put(np_cdm);
+               dev_err(&ofdev->dev, "can't map clock node!\n");
+               return 0;

I think you should return an error code here. -ENOMEM maybe?

Can you send an updated version?

The way I understand the code the return value will be used
as a frequency - the only valid error valeu thus is 0 here,
so I think the return 0 is correct.

Yes. Looking at the code this makes sense.

Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |