[PATCH] usb: phy: msm: fix a possible NULL dereference

[PATCH] usb: phy: msm: fix a possible NULL dereference

[LABBE Corentin]

of_match_device could return NULL, and so cause a NULL pointer
dereference later. Renaming id to of_id (like all others do) in the
process.

Reported-by: coverity (CID 1324133)
Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com>
---
 drivers/usb/phy/phy-msm-usb.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/usb/phy/phy-msm-usb.c b/drivers/usb/phy/phy-msm-usb.c
index 80eb991..c4a66cf 100644
--- a/drivers/usb/phy/phy-msm-usb.c
+++ b/drivers/usb/phy/phy-msm-usb.c
@@ -1506,20 +1506,23 @@ static int msm_otg_read_dt(struct platform_device *pdev, struct msm_otg *motg)
 {
 	struct msm_otg_platform_data *pdata;
 	struct extcon_dev *ext_id, *ext_vbus;
-	const struct of_device_id *id;
+	const struct of_device_id *of_id;
 	struct device_node *node = pdev->dev.of_node;
 	struct property *prop;
 	int len, ret, words;
 	u32 val, tmp[3];
 
+	of_id = of_match_device(msm_otg_dt_match, &pdev->dev);
+	if (!of_id)
+		return -ENODEV;
+
 	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
 	if (!pdata)
 		return -ENOMEM;
 
 	motg->pdata = pdata;
 
-	id = of_match_device(msm_otg_dt_match, &pdev->dev);
-	pdata->phy_type = (enum msm_usb_phy_type) id->data;
+	pdata->phy_type = (enum msm_usb_phy_type)of_id->data;
 
 	motg->link_rst = devm_reset_control_get(&pdev->dev, "link");
 	if (IS_ERR(motg->link_rst))
-- 
2.4.10

Re: [PATCH] usb: phy: msm: fix a possible NULL dereference

[Felipe Balbi]

[-- Attachment #1: Type: text/plain, Size: 992 bytes --]


Hi,

LABBE Corentin <clabbe.montjoie@gmail.com> writes:
> of_match_device could return NULL, and so cause a NULL pointer
> dereference later. Renaming id to of_id (like all others do) in the
> process.
>
> Reported-by: coverity (CID 1324133)
> Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com>
> ---
>  drivers/usb/phy/phy-msm-usb.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/usb/phy/phy-msm-usb.c b/drivers/usb/phy/phy-msm-usb.c
> index 80eb991..c4a66cf 100644
> --- a/drivers/usb/phy/phy-msm-usb.c
> +++ b/drivers/usb/phy/phy-msm-usb.c
> @@ -1506,20 +1506,23 @@ static int msm_otg_read_dt(struct platform_device *pdev, struct msm_otg *motg)
>  {
>  	struct msm_otg_platform_data *pdata;
>  	struct extcon_dev *ext_id, *ext_vbus;
> -	const struct of_device_id *id;
> +	const struct of_device_id *of_id;

and why did you need to rename this variable ? Please resend without the
variable rename.

-- 
balbi

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]

Re: [PATCH] usb: phy: msm: fix a possible NULL dereference

[Corentin LABBE]

Le 16/11/2015 18:07, Felipe Balbi a écrit :
> 
> Hi,
> 
> LABBE Corentin <clabbe.montjoie@gmail.com> writes:
>> of_match_device could return NULL, and so cause a NULL pointer
>> dereference later. Renaming id to of_id (like all others do) in the
>> process.
>>
>> Reported-by: coverity (CID 1324133)
>> Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com>
>> ---
>>  drivers/usb/phy/phy-msm-usb.c | 9 ++++++---
>>  1 file changed, 6 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/usb/phy/phy-msm-usb.c b/drivers/usb/phy/phy-msm-usb.c
>> index 80eb991..c4a66cf 100644
>> --- a/drivers/usb/phy/phy-msm-usb.c
>> +++ b/drivers/usb/phy/phy-msm-usb.c
>> @@ -1506,20 +1506,23 @@ static int msm_otg_read_dt(struct platform_device *pdev, struct msm_otg *motg)
>>  {
>>  	struct msm_otg_platform_data *pdata;
>>  	struct extcon_dev *ext_id, *ext_vbus;
>> -	const struct of_device_id *id;
>> +	const struct of_device_id *of_id;
> 
> and why did you need to rename this variable ? Please resend without the
> variable rename.
> 

I was renaming because all other drivers call it of_id, but I got hint from another thread to use of_device_get_match_data() which will permit to remove that variable, and still solve the issue.

Regards

Re: [PATCH] usb: phy: msm: fix a possible NULL dereference

[Felipe Balbi]

[-- Attachment #1: Type: text/plain, Size: 1756 bytes --]


Hi,

(you need to break your lines at 80 characters)

Corentin LABBE <clabbe.montjoie@gmail.com> writes:
> Le 16/11/2015 18:07, Felipe Balbi a écrit :
>> 
>> Hi,
>> 
>> LABBE Corentin <clabbe.montjoie@gmail.com> writes:
>>> of_match_device could return NULL, and so cause a NULL pointer
>>> dereference later. Renaming id to of_id (like all others do) in the
>>> process.
>>>
>>> Reported-by: coverity (CID 1324133)
>>> Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com>
>>> ---
>>>  drivers/usb/phy/phy-msm-usb.c | 9 ++++++---
>>>  1 file changed, 6 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/usb/phy/phy-msm-usb.c b/drivers/usb/phy/phy-msm-usb.c
>>> index 80eb991..c4a66cf 100644
>>> --- a/drivers/usb/phy/phy-msm-usb.c
>>> +++ b/drivers/usb/phy/phy-msm-usb.c
>>> @@ -1506,20 +1506,23 @@ static int msm_otg_read_dt(struct platform_device *pdev, struct msm_otg *motg)
>>>  {
>>>  	struct msm_otg_platform_data *pdata;
>>>  	struct extcon_dev *ext_id, *ext_vbus;
>>> -	const struct of_device_id *id;
>>> +	const struct of_device_id *of_id;
>> 
>> and why did you need to rename this variable ? Please resend without the
>> variable rename.
>> 
>
> I was renaming because all other drivers call it of_id, but I got hint
> from another thread to use of_device_get_match_data() which will
> permit to remove that variable, and still solve the issue.

for -rc we usually want the smallest patch needed to fix the issue. That
would be the two line if() condition. If converting to
of_device_get_match_data() is about the same (2 or 3 lines), then it
should be fine; but as a general rule of thumb: we don't fix and
cleanup/refactor on the same patch unless strictly necessary.

-- 
balbi

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]