rtla/osnoise: prevent NULL dereference in error handling

[PATCH] rtla/osnoise: prevent NULL dereference in error handling

Posted by Dan Carpenter 1 year, 6 months ago

If the "tool->data" allocation fails then there is no need to call
osnoise_free_top() and, in fact, doing so will lead to a NULL dereference.

Fixes: 1eceb2fc2ca5 ("rtla/osnoise: Add osnoise top mode")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
---
 tools/tracing/rtla/src/osnoise_top.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

diff --git a/tools/tracing/rtla/src/osnoise_top.c b/tools/tracing/rtla/src/osnoise_top.c
index f594a44df840..2f756628613d 100644
--- a/tools/tracing/rtla/src/osnoise_top.c
+++ b/tools/tracing/rtla/src/osnoise_top.c
@@ -651,8 +651,10 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
 		return NULL;
 
 	tool->data = osnoise_alloc_top(nr_cpus);
-	if (!tool->data)
-		goto out_err;
+	if (!tool->data) {
+		osnoise_destroy_tool(tool);
+		return NULL;
+	}
 
 	tool->params = params;
 
@@ -660,11 +662,6 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
 				   osnoise_top_handler, NULL);
 
 	return tool;
-
-out_err:
-	osnoise_free_top(tool->data);
-	osnoise_destroy_tool(tool);
-	return NULL;
 }
 
 static int stop_tracing;
-- 
2.43.0

Re: [PATCH] rtla/osnoise: prevent NULL dereference in error handling

Posted by Luis Claudio R. Goncalves 1 year, 5 months ago

On Fri, Aug 09, 2024 at 03:34:30PM +0300, Dan Carpenter wrote:
> If the "tool->data" allocation fails then there is no need to call
> osnoise_free_top() and, in fact, doing so will lead to a NULL dereference.
> 
> Fixes: 1eceb2fc2ca5 ("rtla/osnoise: Add osnoise top mode")
> Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
> ---
>  tools/tracing/rtla/src/osnoise_top.c | 11 ++++-------
>  1 file changed, 4 insertions(+), 7 deletions(-)

Looks good to me.

Reviewed-by: Luis Claudio R. Goncalves <lgoncalv@redhat.com>


> 
> diff --git a/tools/tracing/rtla/src/osnoise_top.c b/tools/tracing/rtla/src/osnoise_top.c
> index f594a44df840..2f756628613d 100644
> --- a/tools/tracing/rtla/src/osnoise_top.c
> +++ b/tools/tracing/rtla/src/osnoise_top.c
> @@ -651,8 +651,10 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
>  		return NULL;
>  
>  	tool->data = osnoise_alloc_top(nr_cpus);
> -	if (!tool->data)
> -		goto out_err;
> +	if (!tool->data) {
> +		osnoise_destroy_tool(tool);
> +		return NULL;
> +	}
>  
>  	tool->params = params;
>  
> @@ -660,11 +662,6 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
>  				   osnoise_top_handler, NULL);
>  
>  	return tool;
> -
> -out_err:
> -	osnoise_free_top(tool->data);
> -	osnoise_destroy_tool(tool);
> -	return NULL;
>  }
>  
>  static int stop_tracing;
> -- 
> 2.43.0
> 
> 
---end quoted text---

Re: [PATCH] rtla/osnoise: prevent NULL dereference in error handling

Posted by John Kacur 1 year, 6 months ago


On Fri, 9 Aug 2024, Dan Carpenter wrote:

> If the "tool->data" allocation fails then there is no need to call
> osnoise_free_top() and, in fact, doing so will lead to a NULL dereference.
> 
> Fixes: 1eceb2fc2ca5 ("rtla/osnoise: Add osnoise top mode")
> Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
> ---
>  tools/tracing/rtla/src/osnoise_top.c | 11 ++++-------
>  1 file changed, 4 insertions(+), 7 deletions(-)
> 
> diff --git a/tools/tracing/rtla/src/osnoise_top.c b/tools/tracing/rtla/src/osnoise_top.c
> index f594a44df840..2f756628613d 100644
> --- a/tools/tracing/rtla/src/osnoise_top.c
> +++ b/tools/tracing/rtla/src/osnoise_top.c
> @@ -651,8 +651,10 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
>  		return NULL;
>  
>  	tool->data = osnoise_alloc_top(nr_cpus);
> -	if (!tool->data)
> -		goto out_err;
> +	if (!tool->data) {
> +		osnoise_destroy_tool(tool);
> +		return NULL;
> +	}
>  
>  	tool->params = params;
>  
> @@ -660,11 +662,6 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
>  				   osnoise_top_handler, NULL);
>  
>  	return tool;
> -
> -out_err:
> -	osnoise_free_top(tool->data);
> -	osnoise_destroy_tool(tool);
> -	return NULL;
>  }
>  
>  static int stop_tracing;
> -- 

Although your fix appears to be correct, I wonder if it would be better to 
create a second error label, such as out_destroy_tool: as described in 
section 7 of the coding-style.rst

Thanks

John Kacur

Re: [PATCH] rtla/osnoise: prevent NULL dereference in error handling

Posted by Steven Rostedt 1 year, 6 months ago

On Fri, 9 Aug 2024 13:34:28 -0400 (EDT)
John Kacur <jkacur@redhat.com> wrote:

> On Fri, 9 Aug 2024, Dan Carpenter wrote:
> 
> > If the "tool->data" allocation fails then there is no need to call
> > osnoise_free_top() and, in fact, doing so will lead to a NULL dereference.
> > 
> > Fixes: 1eceb2fc2ca5 ("rtla/osnoise: Add osnoise top mode")
> > Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
> > ---
> >  tools/tracing/rtla/src/osnoise_top.c | 11 ++++-------
> >  1 file changed, 4 insertions(+), 7 deletions(-)
> > 
> > diff --git a/tools/tracing/rtla/src/osnoise_top.c b/tools/tracing/rtla/src/osnoise_top.c
> > index f594a44df840..2f756628613d 100644
> > --- a/tools/tracing/rtla/src/osnoise_top.c
> > +++ b/tools/tracing/rtla/src/osnoise_top.c
> > @@ -651,8 +651,10 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
> >  		return NULL;
> >  
> >  	tool->data = osnoise_alloc_top(nr_cpus);
> > -	if (!tool->data)
> > -		goto out_err;
> > +	if (!tool->data) {
> > +		osnoise_destroy_tool(tool);
> > +		return NULL;
> > +	}
> >  
> >  	tool->params = params;
> >  
> > @@ -660,11 +662,6 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
> >  				   osnoise_top_handler, NULL);
> >  
> >  	return tool;
> > -
> > -out_err:
> > -	osnoise_free_top(tool->data);
> > -	osnoise_destroy_tool(tool);
> > -	return NULL;
> >  }
> >  
> >  static int stop_tracing;
> > --   
> 
> Although your fix appears to be correct, I wonder if it would be better to 
> create a second error label, such as out_destroy_tool: as described in 
> section 7 of the coding-style.rst
> 

There's no reason for that. It's the only error path. That is, nothing
would jump to the original out_err:

And for a single error, an if statement is good enough.

-- Steve

Re: [PATCH] rtla/osnoise: prevent NULL dereference in error handling

Posted by John Kacur 1 year, 6 months ago


On Fri, 9 Aug 2024, Steven Rostedt wrote:

> On Fri, 9 Aug 2024 13:34:28 -0400 (EDT)
> John Kacur <jkacur@redhat.com> wrote:
> 
> > On Fri, 9 Aug 2024, Dan Carpenter wrote:
> > 
> > > If the "tool->data" allocation fails then there is no need to call
> > > osnoise_free_top() and, in fact, doing so will lead to a NULL dereference.
> > > 
> > > Fixes: 1eceb2fc2ca5 ("rtla/osnoise: Add osnoise top mode")
> > > Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
> > > ---
> > >  tools/tracing/rtla/src/osnoise_top.c | 11 ++++-------
> > >  1 file changed, 4 insertions(+), 7 deletions(-)
> > > 
> > > diff --git a/tools/tracing/rtla/src/osnoise_top.c b/tools/tracing/rtla/src/osnoise_top.c
> > > index f594a44df840..2f756628613d 100644
> > > --- a/tools/tracing/rtla/src/osnoise_top.c
> > > +++ b/tools/tracing/rtla/src/osnoise_top.c
> > > @@ -651,8 +651,10 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
> > >  		return NULL;
> > >  
> > >  	tool->data = osnoise_alloc_top(nr_cpus);
> > > -	if (!tool->data)
> > > -		goto out_err;
> > > +	if (!tool->data) {
> > > +		osnoise_destroy_tool(tool);
> > > +		return NULL;
> > > +	}
> > >  
> > >  	tool->params = params;
> > >  
> > > @@ -660,11 +662,6 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
> > >  				   osnoise_top_handler, NULL);
> > >  
> > >  	return tool;
> > > -
> > > -out_err:
> > > -	osnoise_free_top(tool->data);
> > > -	osnoise_destroy_tool(tool);
> > > -	return NULL;
> > >  }
> > >  
> > >  static int stop_tracing;
> > > --   
> > 
> > Although your fix appears to be correct, I wonder if it would be better to 
> > create a second error label, such as out_destroy_tool: as described in 
> > section 7 of the coding-style.rst
> > 
> 
> There's no reason for that. It's the only error path. That is, nothing
> would jump to the original out_err:
> 
> And for a single error, an if statement is good enough.
> 
> -- Steve
> 
> 

Ah, right of course.
Okay in that case, Signed-off-by: John Kacur <jkacur@redhat.com>
(applied the patch, built and ran)

Re: [PATCH] rtla/osnoise: prevent NULL dereference in error handling

Posted by Steven Rostedt 1 year, 6 months ago

On Fri, 9 Aug 2024 13:53:33 -0400 (EDT)
John Kacur <jkacur@redhat.com> wrote:


> > > Although your fix appears to be correct, I wonder if it would be better to 
> > > create a second error label, such as out_destroy_tool: as described in 
> > > section 7 of the coding-style.rst
> > >   
> > 
> > There's no reason for that. It's the only error path. That is, nothing
> > would jump to the original out_err:
> > 
> > And for a single error, an if statement is good enough.
> > 
> > -- Steve
> > 
> >   
> 
> Ah, right of course.
> Okay in that case, Signed-off-by: John Kacur <jkacur@redhat.com>
> (applied the patch, built and ran)

Note, "Signed-off-by" is for the author of a patch or someone pushing it
through their tree. I believe you want either "Acked-by" or "Reviewed-by",
and since you ran it you could also add "Tested-by".

-- Steve

Re: [PATCH] rtla/osnoise: prevent NULL dereference in error handling

Posted by John Kacur 1 year, 6 months ago


On Fri, 9 Aug 2024, Steven Rostedt wrote:

> On Fri, 9 Aug 2024 13:53:33 -0400 (EDT)
> John Kacur <jkacur@redhat.com> wrote:
> 
> 
> > > > Although your fix appears to be correct, I wonder if it would be better to 
> > > > create a second error label, such as out_destroy_tool: as described in 
> > > > section 7 of the coding-style.rst
> > > >   
> > > 
> > > There's no reason for that. It's the only error path. That is, nothing
> > > would jump to the original out_err:
> > > 
> > > And for a single error, an if statement is good enough.
> > > 
> > > -- Steve
> > > 
> > >   
> > 
> > Ah, right of course.
> > Okay in that case, Signed-off-by: John Kacur <jkacur@redhat.com>
> > (applied the patch, built and ran)
> 
> Note, "Signed-off-by" is for the author of a patch or someone pushing it
> through their tree. I believe you want either "Acked-by" or "Reviewed-by",
> and since you ran it you could also add "Tested-by".
> 
> -- Steve

Thanks Steve,

Reviewed-by: John Kacur <jkacur@redhat.com>
Tested-by: John Kacur <jkacur@redhat.com>