cpufreq: Avoid a bad reference count on CPU node

[PATCH] cpufreq: Avoid a bad reference count on CPU node

Posted by Miquel Sabaté Solà 2 months, 1 week ago

In the parse_perf_domain function, if the call to
of_parse_phandle_with_args returns an error, then the reference to the
CPU device node that was acquired at the start of the function would not
be properly decremented.

Address this by declaring the variable with the __free(device_node)
cleanup attribute.

Signed-off-by: Miquel Sabaté Solà <mikisabate@gmail.com>
---
 include/linux/cpufreq.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h
index d4d2f4d1d7cb..aabec598f79a 100644
--- a/include/linux/cpufreq.h
+++ b/include/linux/cpufreq.h
@@ -1113,10 +1113,9 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
 				    const char *cell_name,
 				    struct of_phandle_args *args)
 {
-	struct device_node *cpu_np;
 	int ret;
 
-	cpu_np = of_cpu_device_node_get(cpu);
+	struct device_node *cpu_np __free(device_node) = of_cpu_device_node_get(cpu);
 	if (!cpu_np)
 		return -ENODEV;
 
@@ -1124,9 +1123,6 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
 					 args);
 	if (ret < 0)
 		return ret;
-
-	of_node_put(cpu_np);
-
 	return 0;
 }
 
-- 
2.46.0

Re: [PATCH] cpufreq: Avoid a bad reference count on CPU node

Posted by Viresh Kumar 1 month, 4 weeks ago

On 17-09-24, 15:42, Miquel Sabaté Solà wrote:
> In the parse_perf_domain function, if the call to
> of_parse_phandle_with_args returns an error, then the reference to the
> CPU device node that was acquired at the start of the function would not
> be properly decremented.
> 
> Address this by declaring the variable with the __free(device_node)
> cleanup attribute.
> 
> Signed-off-by: Miquel Sabaté Solà <mikisabate@gmail.com>
> ---
>  include/linux/cpufreq.h | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
> 
> diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h
> index d4d2f4d1d7cb..aabec598f79a 100644
> --- a/include/linux/cpufreq.h
> +++ b/include/linux/cpufreq.h
> @@ -1113,10 +1113,9 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
>  				    const char *cell_name,
>  				    struct of_phandle_args *args)
>  {
> -	struct device_node *cpu_np;
>  	int ret;
>  
> -	cpu_np = of_cpu_device_node_get(cpu);
> +	struct device_node *cpu_np __free(device_node) = of_cpu_device_node_get(cpu);
>  	if (!cpu_np)
>  		return -ENODEV;
>  
> @@ -1124,9 +1123,6 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
>  					 args);
>  	if (ret < 0)
>  		return ret;
> -
> -	of_node_put(cpu_np);
> -
>  	return 0;
>  }

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>

-- 
viresh

Re: [PATCH] cpufreq: Avoid a bad reference count on CPU node

Posted by Rafael J. Wysocki 1 month, 4 weeks ago

On Tue, Oct 1, 2024 at 8:32 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
>
> On 17-09-24, 15:42, Miquel Sabaté Solà wrote:
> > In the parse_perf_domain function, if the call to
> > of_parse_phandle_with_args returns an error, then the reference to the
> > CPU device node that was acquired at the start of the function would not
> > be properly decremented.
> >
> > Address this by declaring the variable with the __free(device_node)
> > cleanup attribute.
> >
> > Signed-off-by: Miquel Sabaté Solà <mikisabate@gmail.com>
> > ---
> >  include/linux/cpufreq.h | 6 +-----
> >  1 file changed, 1 insertion(+), 5 deletions(-)
> >
> > diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h
> > index d4d2f4d1d7cb..aabec598f79a 100644
> > --- a/include/linux/cpufreq.h
> > +++ b/include/linux/cpufreq.h
> > @@ -1113,10 +1113,9 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
> >                                   const char *cell_name,
> >                                   struct of_phandle_args *args)
> >  {
> > -     struct device_node *cpu_np;
> >       int ret;
> >
> > -     cpu_np = of_cpu_device_node_get(cpu);
> > +     struct device_node *cpu_np __free(device_node) = of_cpu_device_node_get(cpu);
> >       if (!cpu_np)
> >               return -ENODEV;
> >
> > @@ -1124,9 +1123,6 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
> >                                        args);
> >       if (ret < 0)
> >               return ret;
> > -
> > -     of_node_put(cpu_np);
> > -
> >       return 0;
> >  }
>
> Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
>
> --

Applied as 6.12-rc material, thanks!

Re: [PATCH] cpufreq: Avoid a bad reference count on CPU node

Posted by Miquel Sabaté Solà 1 month, 4 weeks ago

On dt., d’oct. 01 2024, Rafael J. Wysocki wrote:

> On Tue, Oct 1, 2024 at 8:32 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
>>
>> On 17-09-24, 15:42, Miquel Sabaté Solà wrote:
>> > In the parse_perf_domain function, if the call to
>> > of_parse_phandle_with_args returns an error, then the reference to the
>> > CPU device node that was acquired at the start of the function would not
>> > be properly decremented.
>> >
>> > Address this by declaring the variable with the __free(device_node)
>> > cleanup attribute.
>> >
>> > Signed-off-by: Miquel Sabaté Solà <mikisabate@gmail.com>
>> > ---
>> >  include/linux/cpufreq.h | 6 +-----
>> >  1 file changed, 1 insertion(+), 5 deletions(-)
>> >
>> > diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h
>> > index d4d2f4d1d7cb..aabec598f79a 100644
>> > --- a/include/linux/cpufreq.h
>> > +++ b/include/linux/cpufreq.h
>> > @@ -1113,10 +1113,9 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
>> >                                   const char *cell_name,
>> >                                   struct of_phandle_args *args)
>> >  {
>> > -     struct device_node *cpu_np;
>> >       int ret;
>> >
>> > -     cpu_np = of_cpu_device_node_get(cpu);
>> > +     struct device_node *cpu_np __free(device_node) = of_cpu_device_node_get(cpu);
>> >       if (!cpu_np)
>> >               return -ENODEV;
>> >
>> > @@ -1124,9 +1123,6 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
>> >                                        args);
>> >       if (ret < 0)
>> >               return ret;
>> > -
>> > -     of_node_put(cpu_np);
>> > -
>> >       return 0;
>> >  }
>>
>> Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
>>
>> --
>
> Applied as 6.12-rc material, thanks!

Great, thanks!

Re: [PATCH] cpufreq: Avoid a bad reference count on CPU node

Posted by Miquel Sabaté Solà 1 month, 4 weeks ago

On dt., de set. 17 2024, Miquel Sabaté Solà wrote:

> In the parse_perf_domain function, if the call to
> of_parse_phandle_with_args returns an error, then the reference to the
> CPU device node that was acquired at the start of the function would not
> be properly decremented.
>
> Address this by declaring the variable with the __free(device_node)
> cleanup attribute.
>
> Signed-off-by: Miquel Sabaté Solà <mikisabate@gmail.com>
> ---
>  include/linux/cpufreq.h | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
>
> diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h
> index d4d2f4d1d7cb..aabec598f79a 100644
> --- a/include/linux/cpufreq.h
> +++ b/include/linux/cpufreq.h
> @@ -1113,10 +1113,9 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
>  				    const char *cell_name,
>  				    struct of_phandle_args *args)
>  {
> -	struct device_node *cpu_np;
>  	int ret;
>  
> -	cpu_np = of_cpu_device_node_get(cpu);
> +	struct device_node *cpu_np __free(device_node) = of_cpu_device_node_get(cpu);
>  	if (!cpu_np)
>  		return -ENODEV;
>  
> @@ -1124,9 +1123,6 @@ static inline int parse_perf_domain(int cpu, const char *list_name,
>  					 args);
>  	if (ret < 0)
>  		return ret;
> -
> -	of_node_put(cpu_np);
> -
>  	return 0;
>  }

Gently ping :)

Do you have some time to take a look at this fix?

Thanks,
Miquel