[PATCH] drivers: of: check input parameter name for __of_find_property

[PATCH] drivers: of: check input parameter name for __of_find_property

[Peng Fan]

Check input parameter 'name' for __of_find_property. If name is NULL,
of_prop_cmp->strcasecmp may trigger panic.

Signed-off-by: Peng Fan <van.freenix@gmail.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Frank Rowand <frowand.list@gmail.com>
Cc: Grant Likely <grant.likely@linaro.org>
---
 drivers/of/base.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/of/base.c b/drivers/of/base.c
index 8b5a187..e41436d 100644
--- a/drivers/of/base.c
+++ b/drivers/of/base.c
@@ -215,7 +215,7 @@ static struct property *__of_find_property(const struct device_node *np,
 {
 	struct property *pp;
 
-	if (!np)
+	if (!np || !name)
 		return NULL;
 
 	for (pp = np->properties; pp; pp = pp->next) {
-- 
1.8.4

[PATCH] drivers: of: check input parameter name for __of_find_property

[Peng Fan]

Check input parameter 'name' for __of_find_property. If name is NULL,
of_prop_cmp->strcasecmp may trigger panic.

Signed-off-by: Peng Fan <van.freenix-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: Rob Herring <robh+dt-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
Cc: Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: Grant Likely <grant.likely-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
---
 drivers/of/base.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/of/base.c b/drivers/of/base.c
index 8b5a187..e41436d 100644
--- a/drivers/of/base.c
+++ b/drivers/of/base.c
@@ -215,7 +215,7 @@ static struct property *__of_find_property(const struct device_node *np,
 {
 	struct property *pp;
 
-	if (!np)
+	if (!np || !name)
 		return NULL;
 
 	for (pp = np->properties; pp; pp = pp->next) {
-- 
1.8.4

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] drivers: of: check input parameter name for __of_find_property

[Rob Herring]

On 09/11/2015 08:44 AM, Peng Fan wrote:
> Check input parameter 'name' for __of_find_property. If name is NULL,
> of_prop_cmp->strcasecmp may trigger panic.

Arguably that could be a feature. Do you have a usecase where name being
NULL is valid and panicking is a problem?

Rob

> 
> Signed-off-by: Peng Fan <van.freenix@gmail.com>
> Cc: Rob Herring <robh+dt@kernel.org>
> Cc: Frank Rowand <frowand.list@gmail.com>
> Cc: Grant Likely <grant.likely@linaro.org>
> ---
>  drivers/of/base.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/of/base.c b/drivers/of/base.c
> index 8b5a187..e41436d 100644
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -215,7 +215,7 @@ static struct property *__of_find_property(const struct device_node *np,
>  {
>  	struct property *pp;
>  
> -	if (!np)
> +	if (!np || !name)
>  		return NULL;
>  
>  	for (pp = np->properties; pp; pp = pp->next) {
> 

Re: [PATCH] drivers: of: check input parameter name for __of_find_property

[Peng Fan]

Hi Rob,

On Tue, Sep 15, 2015 at 10:56:28AM -0500, Rob Herring wrote:
> On 09/11/2015 08:44 AM, Peng Fan wrote:
> > Check input parameter 'name' for __of_find_property. If name is NULL,
> > of_prop_cmp->strcasecmp may trigger panic.
> 
> Arguably that could be a feature. Do you have a usecase where name being
> NULL is valid and panicking is a problem?

In drivers/pinctrl/devicetree.c

195                 propname = kasprintf(GFP_KERNEL, "pinctrl-%d", state);
196                 prop = of_find_property(np, propname, &size);
197                 kfree(propname);
198                 if (!prop)
199                         break;

If propname is NULL, of_find_property may trigger panic. Anyway propname should be checked
before passing to of_find_property.

I did not met panic message. I wrote this patch when I was reading the piece code.
I think the name parameter should be checked before doing string compare.

Regards,
Peng.

> 
> Rob
> 
> > 
> > Signed-off-by: Peng Fan <van.freenix@gmail.com>
> > Cc: Rob Herring <robh+dt@kernel.org>
> > Cc: Frank Rowand <frowand.list@gmail.com>
> > Cc: Grant Likely <grant.likely@linaro.org>
> > ---
> >  drivers/of/base.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/of/base.c b/drivers/of/base.c
> > index 8b5a187..e41436d 100644
> > --- a/drivers/of/base.c
> > +++ b/drivers/of/base.c
> > @@ -215,7 +215,7 @@ static struct property *__of_find_property(const struct device_node *np,
> >  {
> >  	struct property *pp;
> >  
> > -	if (!np)
> > +	if (!np || !name)
> >  		return NULL;
> >  
> >  	for (pp = np->properties; pp; pp = pp->next) {
> > 
> 

Re: [PATCH] drivers: of: check input parameter name for __of_find_property

[Rob Herring]

On 09/15/2015 07:16 PM, Peng Fan wrote:
> Hi Rob,
> 
> On Tue, Sep 15, 2015 at 10:56:28AM -0500, Rob Herring wrote:
>> On 09/11/2015 08:44 AM, Peng Fan wrote:
>>> Check input parameter 'name' for __of_find_property. If name is NULL,
>>> of_prop_cmp->strcasecmp may trigger panic.
>>
>> Arguably that could be a feature. Do you have a usecase where name being
>> NULL is valid and panicking is a problem?
> 
> In drivers/pinctrl/devicetree.c
> 
> 195                 propname = kasprintf(GFP_KERNEL, "pinctrl-%d", state);
> 196                 prop = of_find_property(np, propname, &size);
> 197                 kfree(propname);
> 198                 if (!prop)
> 199                         break;
> 
> If propname is NULL, of_find_property may trigger panic. Anyway propname should be checked
> before passing to of_find_property.

propname will only be NULL if the memory allocation failed which gives
you a big warning message. I don't think you would want to continue on
in this case.

So I'm inclined to leave this as is with passing NULL to
of_find_property to always be an error and fatal.

Rob

> I did not met panic message. I wrote this patch when I was reading the piece code.
> I think the name parameter should be checked before doing string compare.
> 
> Regards,
> Peng.
> 
>>
>> Rob
>>
>>>
>>> Signed-off-by: Peng Fan <van.freenix@gmail.com>
>>> Cc: Rob Herring <robh+dt@kernel.org>
>>> Cc: Frank Rowand <frowand.list@gmail.com>
>>> Cc: Grant Likely <grant.likely@linaro.org>
>>> ---
>>>  drivers/of/base.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/of/base.c b/drivers/of/base.c
>>> index 8b5a187..e41436d 100644
>>> --- a/drivers/of/base.c
>>> +++ b/drivers/of/base.c
>>> @@ -215,7 +215,7 @@ static struct property *__of_find_property(const struct device_node *np,
>>>  {
>>>  	struct property *pp;
>>>  
>>> -	if (!np)
>>> +	if (!np || !name)
>>>  		return NULL;
>>>  
>>>  	for (pp = np->properties; pp; pp = pp->next) {
>>>
>>

Re: [PATCH] drivers: of: check input parameter name for __of_find_property

[Rob Herring]

On 09/15/2015 07:16 PM, Peng Fan wrote:
> Hi Rob,
> 
> On Tue, Sep 15, 2015 at 10:56:28AM -0500, Rob Herring wrote:
>> On 09/11/2015 08:44 AM, Peng Fan wrote:
>>> Check input parameter 'name' for __of_find_property. If name is NULL,
>>> of_prop_cmp->strcasecmp may trigger panic.
>>
>> Arguably that could be a feature. Do you have a usecase where name being
>> NULL is valid and panicking is a problem?
> 
> In drivers/pinctrl/devicetree.c
> 
> 195                 propname = kasprintf(GFP_KERNEL, "pinctrl-%d", state);
> 196                 prop = of_find_property(np, propname, &size);
> 197                 kfree(propname);
> 198                 if (!prop)
> 199                         break;
> 
> If propname is NULL, of_find_property may trigger panic. Anyway propname should be checked
> before passing to of_find_property.

propname will only be NULL if the memory allocation failed which gives
you a big warning message. I don't think you would want to continue on
in this case.

So I'm inclined to leave this as is with passing NULL to
of_find_property to always be an error and fatal.

Rob

> I did not met panic message. I wrote this patch when I was reading the piece code.
> I think the name parameter should be checked before doing string compare.
> 
> Regards,
> Peng.
> 
>>
>> Rob
>>
>>>
>>> Signed-off-by: Peng Fan <van.freenix-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>>> Cc: Rob Herring <robh+dt-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
>>> Cc: Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>>> Cc: Grant Likely <grant.likely-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
>>> ---
>>>  drivers/of/base.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/of/base.c b/drivers/of/base.c
>>> index 8b5a187..e41436d 100644
>>> --- a/drivers/of/base.c
>>> +++ b/drivers/of/base.c
>>> @@ -215,7 +215,7 @@ static struct property *__of_find_property(const struct device_node *np,
>>>  {
>>>  	struct property *pp;
>>>  
>>> -	if (!np)
>>> +	if (!np || !name)
>>>  		return NULL;
>>>  
>>>  	for (pp = np->properties; pp; pp = pp->next) {
>>>
>>

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] drivers: of: check input parameter name for __of_find_property

[Peng Fan]

On Tue, Sep 15, 2015 at 09:27:59PM -0500, Rob Herring wrote:
> On 09/15/2015 07:16 PM, Peng Fan wrote:
> > Hi Rob,
> > 
> > On Tue, Sep 15, 2015 at 10:56:28AM -0500, Rob Herring wrote:
> >> On 09/11/2015 08:44 AM, Peng Fan wrote:
> >>> Check input parameter 'name' for __of_find_property. If name is NULL,
> >>> of_prop_cmp->strcasecmp may trigger panic.
> >>
> >> Arguably that could be a feature. Do you have a usecase where name being
> >> NULL is valid and panicking is a problem?
> > 
> > In drivers/pinctrl/devicetree.c
> > 
> > 195                 propname = kasprintf(GFP_KERNEL, "pinctrl-%d", state);
> > 196                 prop = of_find_property(np, propname, &size);
> > 197                 kfree(propname);
> > 198                 if (!prop)
> > 199                         break;
> > 
> > If propname is NULL, of_find_property may trigger panic. Anyway propname should be checked
> > before passing to of_find_property.
> 
> propname will only be NULL if the memory allocation failed which gives
> you a big warning message. I don't think you would want to continue on
> in this case.
> 
> So I'm inclined to leave this as is with passing NULL to
> of_find_property to always be an error and fatal.

of_find_property has EXPORT_SYMBOL, it maybe called in different places. And it should handle the case that name is NULL, right?
If passing NULL to of_find_property is an error, then why check parameter 'np' like "if (!np)" in of_find_property?

Regards,
Peng.

> 
> Rob
> 
> > I did not met panic message. I wrote this patch when I was reading the piece code.
> > I think the name parameter should be checked before doing string compare.
> > 
> > Regards,
> > Peng.
> > 
> >>
> >> Rob
> >>
> >>>
> >>> Signed-off-by: Peng Fan <van.freenix@gmail.com>
> >>> Cc: Rob Herring <robh+dt@kernel.org>
> >>> Cc: Frank Rowand <frowand.list@gmail.com>
> >>> Cc: Grant Likely <grant.likely@linaro.org>
> >>> ---
> >>>  drivers/of/base.c | 2 +-
> >>>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>>
> >>> diff --git a/drivers/of/base.c b/drivers/of/base.c
> >>> index 8b5a187..e41436d 100644
> >>> --- a/drivers/of/base.c
> >>> +++ b/drivers/of/base.c
> >>> @@ -215,7 +215,7 @@ static struct property *__of_find_property(const struct device_node *np,
> >>>  {
> >>>  	struct property *pp;
> >>>  
> >>> -	if (!np)
> >>> +	if (!np || !name)
> >>>  		return NULL;
> >>>  
> >>>  	for (pp = np->properties; pp; pp = pp->next) {
> >>>
> >>
>