[PATCH][next] reset: npcm: check for NULL return from syscon_regmap_lookup_by_compat

[PATCH][next] reset: npcm: check for NULL return from syscon_regmap_lookup_by_compat

[Colin King]

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

Function syscon_regmap_lookup_by_compat can return a NULL pointer, so
the IS_ERR check on the return is incorrect. Fix this by checking for
IS_ERR_OR_NULL and return -ENODEV if true.  This avoids a null pointer
dereference on gcr_regmap later on.

Addresses-Coverity: ("Dereference null return (stat)")
Fixes: b3f1d036f26d ("reset: npcm: add NPCM reset controller driver")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
---
 drivers/reset/reset-npcm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/reset/reset-npcm.c b/drivers/reset/reset-npcm.c
index 2ea4d3136e15..9febf8bed2f6 100644
--- a/drivers/reset/reset-npcm.c
+++ b/drivers/reset/reset-npcm.c
@@ -161,9 +161,9 @@ static int npcm_usb_reset(struct platform_device *pdev, struct npcm_rc_data *rc)
 	of_match_device(dev->driver->of_match_table, dev)->data;
 
 	gcr_regmap = syscon_regmap_lookup_by_compatible(gcr_dt);
-	if (IS_ERR(gcr_regmap)) {
+	if (IS_ERR_OR_NULL(gcr_regmap)) {
 		dev_err(&pdev->dev, "Failed to find %s\n", gcr_dt);
-		return PTR_ERR(gcr_regmap);
+		return -ENODEV;
 	}
 
 	/* checking which USB device is enabled */
-- 
2.20.1

Re: [PATCH][next] reset: npcm: check for NULL return from syscon_regmap_lookup_by_compat

[Philipp Zabel]

Hi Colin,

On Fri, 2019-11-08 at 15:55 +0000, Colin King wrote:
> From: Colin Ian King <colin.king@canonical.com>
> 
> Function syscon_regmap_lookup_by_compat can return a NULL pointer, so

Could you point out where that NULL pointer would come from? As far as I
understand, syscon_regmap_lookup_by_compatible() should either return a
negative error code, or syscon->regmap, which should never be NULL.

> the IS_ERR check on the return is incorrect. Fix this by checking for
> IS_ERR_OR_NULL and return -ENODEV if true.  This avoids a null pointer
> dereference on gcr_regmap later on.
> 
> Addresses-Coverity: ("Dereference null return (stat)")
> Fixes: b3f1d036f26d ("reset: npcm: add NPCM reset controller driver")
> Signed-off-by: Colin Ian King <colin.king@canonical.com>
> ---
>  drivers/reset/reset-npcm.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/reset/reset-npcm.c b/drivers/reset/reset-npcm.c
> index 2ea4d3136e15..9febf8bed2f6 100644
> --- a/drivers/reset/reset-npcm.c
> +++ b/drivers/reset/reset-npcm.c
> @@ -161,9 +161,9 @@ static int npcm_usb_reset(struct platform_device *pdev, struct npcm_rc_data *rc)
>  	of_match_device(dev->driver->of_match_table, dev)->data;
>  
>  	gcr_regmap = syscon_regmap_lookup_by_compatible(gcr_dt);
> -	if (IS_ERR(gcr_regmap)) {
> +	if (IS_ERR_OR_NULL(gcr_regmap)) {
>  		dev_err(&pdev->dev, "Failed to find %s\n", gcr_dt);
> -		return PTR_ERR(gcr_regmap);
> +		return -ENODEV;
>  	}
>  
>  	/* checking which USB device is enabled */

regards
Philipp

Re: [PATCH][next] reset: npcm: check for NULL return from syscon_regmap_lookup_by_compat

[Colin Ian King]

On 14/11/2019 13:22, Philipp Zabel wrote:
> Hi Colin,
> 
> On Fri, 2019-11-08 at 15:55 +0000, Colin King wrote:
>> From: Colin Ian King <colin.king@canonical.com>
>>
>> Function syscon_regmap_lookup_by_compat can return a NULL pointer, so
> 
> Could you point out where that NULL pointer would come from? As far as I
> understand, syscon_regmap_lookup_by_compatible() should either return a
> negative error code, or syscon->regmap, which should never be NULL.

Maybe I was over-zealous here with the results from the static analyzer.
Re-reading the call chain it does seem like I was mistaken. So NAK this
patch.

Apologies for the noise.

> 
>> the IS_ERR check on the return is incorrect. Fix this by checking for
>> IS_ERR_OR_NULL and return -ENODEV if true.  This avoids a null pointer
>> dereference on gcr_regmap later on.
>>
>> Addresses-Coverity: ("Dereference null return (stat)")
>> Fixes: b3f1d036f26d ("reset: npcm: add NPCM reset controller driver")
>> Signed-off-by: Colin Ian King <colin.king@canonical.com>
>> ---
>>  drivers/reset/reset-npcm.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/reset/reset-npcm.c b/drivers/reset/reset-npcm.c
>> index 2ea4d3136e15..9febf8bed2f6 100644
>> --- a/drivers/reset/reset-npcm.c
>> +++ b/drivers/reset/reset-npcm.c
>> @@ -161,9 +161,9 @@ static int npcm_usb_reset(struct platform_device *pdev, struct npcm_rc_data *rc)
>>  	of_match_device(dev->driver->of_match_table, dev)->data;
>>  
>>  	gcr_regmap = syscon_regmap_lookup_by_compatible(gcr_dt);
>> -	if (IS_ERR(gcr_regmap)) {
>> +	if (IS_ERR_OR_NULL(gcr_regmap)) {
>>  		dev_err(&pdev->dev, "Failed to find %s\n", gcr_dt);
>> -		return PTR_ERR(gcr_regmap);
>> +		return -ENODEV;
>>  	}
>>  
>>  	/* checking which USB device is enabled */
> 
> regards
> Philipp
>