[PATCH] reset: fix potential null pointer dereference on pointer dev

[PATCH] reset: fix potential null pointer dereference on pointer dev

[Colin King]

From: Colin Ian King <colin.king@canonical.com>

Pointer dev is being dereferenced when passed to the inlined
functon dev_name, however, dev is later being null checked.
Thus there is a potential null pointer dereference on a null
dev. Fix this by performing the null check on dev before
dereferencing it.

Addresses-Coverity: ("Dereference before null check")
Fixes: 6691dffab0ab ("reset: add support for non-DT systems")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
---
 drivers/reset/core.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/reset/core.c b/drivers/reset/core.c
index 81ea77cba123..83f1a1d5ee67 100644
--- a/drivers/reset/core.c
+++ b/drivers/reset/core.c
@@ -691,12 +691,13 @@ __reset_control_get_from_lookup(struct device *dev, const char *con_id,
 {
 	const struct reset_control_lookup *lookup;
 	struct reset_controller_dev *rcdev;
-	const char *dev_id = dev_name(dev);
+	const char *dev_id;
 	struct reset_control *rstc = NULL;
 
 	if (!dev)
 		return ERR_PTR(-EINVAL);
 
+	dev_id = dev_name(dev);
 	mutex_lock(&reset_lookup_mutex);
 
 	list_for_each_entry(lookup, &reset_lookup_list, list) {
-- 
2.20.1

Re: [PATCH] reset: fix potential null pointer dereference on pointer dev

[Philipp Zabel]

Hi Colin,

[added Bartosz to Cc:]

On Thu, 2019-05-09 at 17:00 +0100, Colin King wrote:
> From: Colin Ian King <colin.king@canonical.com>
> 
> Pointer dev is being dereferenced when passed to the inlined
> functon dev_name, however, dev is later being null checked.
> Thus there is a potential null pointer dereference on a null
> dev. Fix this by performing the null check on dev before
> dereferencing it.
> 
> Addresses-Coverity: ("Dereference before null check")
> Fixes: 6691dffab0ab ("reset: add support for non-DT systems")
> Signed-off-by: Colin Ian King <colin.king@canonical.com>
> ---
>  drivers/reset/core.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/reset/core.c b/drivers/reset/core.c
> index 81ea77cba123..83f1a1d5ee67 100644
> --- a/drivers/reset/core.c
> +++ b/drivers/reset/core.c
> @@ -691,12 +691,13 @@ __reset_control_get_from_lookup(struct device *dev, const char *con_id,
>  {
>  	const struct reset_control_lookup *lookup;
>  	struct reset_controller_dev *rcdev;
> -	const char *dev_id = dev_name(dev);
> +	const char *dev_id;
>  	struct reset_control *rstc = NULL;
>  
>  	if (!dev)
>  		return ERR_PTR(-EINVAL);

Thank you for the patch. I think this check should be removed instead,
though, as __reset_control_get_from_lookup is only ever called from
__reset_control_get, right after checking dev->of_node. So dev can not
be NULL here.

regards
Philipp

Re: [PATCH] reset: fix potential null pointer dereference on pointer dev

[Colin Ian King]

On 10/05/2019 10:14, Philipp Zabel wrote:
> Hi Colin,
> 
> [added Bartosz to Cc:]
> 
> On Thu, 2019-05-09 at 17:00 +0100, Colin King wrote:
>> From: Colin Ian King <colin.king@canonical.com>
>>
>> Pointer dev is being dereferenced when passed to the inlined
>> functon dev_name, however, dev is later being null checked.
>> Thus there is a potential null pointer dereference on a null
>> dev. Fix this by performing the null check on dev before
>> dereferencing it.
>>
>> Addresses-Coverity: ("Dereference before null check")
>> Fixes: 6691dffab0ab ("reset: add support for non-DT systems")
>> Signed-off-by: Colin Ian King <colin.king@canonical.com>
>> ---
>>  drivers/reset/core.c | 3 ++-
>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/reset/core.c b/drivers/reset/core.c
>> index 81ea77cba123..83f1a1d5ee67 100644
>> --- a/drivers/reset/core.c
>> +++ b/drivers/reset/core.c
>> @@ -691,12 +691,13 @@ __reset_control_get_from_lookup(struct device *dev, const char *con_id,
>>  {
>>  	const struct reset_control_lookup *lookup;
>>  	struct reset_controller_dev *rcdev;
>> -	const char *dev_id = dev_name(dev);
>> +	const char *dev_id;
>>  	struct reset_control *rstc = NULL;
>>  
>>  	if (!dev)
>>  		return ERR_PTR(-EINVAL);
> 
> Thank you for the patch. I think this check should be removed instead,
> though, as __reset_control_get_from_lookup is only ever called from
> __reset_control_get, right after checking dev->of_node. So dev can not
> be NULL here.

Good point, I totally overlooked that. I'll send a V2 shortly.

> 
> regards
> Philipp
>