[dpdk-dev] [PATCH] app/testpmd: fix memory leak

[dpdk-dev] [PATCH] app/testpmd: fix memory leak

[Pablo de Lara]

Free memory when port flow entry creation fails.

Coverity issue: 139600
Fixes: 938a184a1870 ("app/testpmd: implement basic support for flow API")

Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---
 app/test-pmd/config.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 5834498..5823100 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -954,6 +954,8 @@ port_flow_new(const struct rte_flow_attr *attr,
 		goto store;
 	}
 notsup:
+	if (pf)
+		free(pf);
 	rte_errno = err;
 	return NULL;
 }
-- 
2.7.4

[dpdk-dev] [PATCH v2] app/testpmd: fix memory leak

[Pablo de Lara]

Free memory when port flow entry creation fails.

Coverity issue: 139600
Fixes: 938a184a1870 ("app/testpmd: implement basic support for flow API")

Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---
Changes in v2:

- Removed unnecessary conditional

 app/test-pmd/config.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 5834498..467932f 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -954,6 +954,7 @@ port_flow_new(const struct rte_flow_attr *attr,
 		goto store;
 	}
 notsup:
+	free(pf);
 	rte_errno = err;
 	return NULL;
 }
-- 
2.7.4

Re: [dpdk-dev] [PATCH v2] app/testpmd: fix memory leak

[Adrien Mazarguil]

Hi Pablo,

On Fri, Jan 27, 2017 at 02:54:58PM +0000, Pablo de Lara wrote:
> Free memory when port flow entry creation fails.
> 
> Coverity issue: 139600
> Fixes: 938a184a1870 ("app/testpmd: implement basic support for flow API")
> 
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> ---
> Changes in v2:
> 
> - Removed unnecessary conditional
> 
>  app/test-pmd/config.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
> index 5834498..467932f 100644
> --- a/app/test-pmd/config.c
> +++ b/app/test-pmd/config.c
> @@ -954,6 +954,7 @@ port_flow_new(const struct rte_flow_attr *attr,
>  		goto store;
>  	}
>  notsup:
> +	free(pf);
>  	rte_errno = err;
>  	return NULL;
>  }
> -- 
> 2.7.4
> 

I think this is a false positive, which is why I did not address it during
the last round of Coverity issues.

As a two-pass function, errors are checked during the first pass, when pf is
not allocated yet. During the second pass, intermediate functions are not
supposed to return a different result and "notsup" cannot occur.

I think assert()/rte_assert() would make more sense and should fool Coverity
into understanding the expected behavior.

The commit log should reflect that we are addressing a false positive since
there is no problem with the code logic (of course unless I missed anything
obvious).

Thanks.

-- 
Adrien Mazarguil
6WIND

Re: [dpdk-dev] [PATCH v2] app/testpmd: fix memory leak

[De Lara Guarch, Pablo]

Hi Adrien,

> -----Original Message-----
> From: Adrien Mazarguil [mailto:adrien.mazarguil@6wind.com]
> Sent: Monday, January 30, 2017 9:33 AM
> To: De Lara Guarch, Pablo
> Cc: Wu, Jingjing; dev@dpdk.org
> Subject: Re: [PATCH v2] app/testpmd: fix memory leak
> 
> Hi Pablo,
> 
> On Fri, Jan 27, 2017 at 02:54:58PM +0000, Pablo de Lara wrote:
> > Free memory when port flow entry creation fails.
> >
> > Coverity issue: 139600
> > Fixes: 938a184a1870 ("app/testpmd: implement basic support for flow
> API")
> >
> > Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> > ---
> > Changes in v2:
> >
> > - Removed unnecessary conditional
> >
> >  app/test-pmd/config.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
> > index 5834498..467932f 100644
> > --- a/app/test-pmd/config.c
> > +++ b/app/test-pmd/config.c
> > @@ -954,6 +954,7 @@ port_flow_new(const struct rte_flow_attr *attr,
> >  		goto store;
> >  	}
> >  notsup:
> > +	free(pf);
> >  	rte_errno = err;
> >  	return NULL;
> >  }
> > --
> > 2.7.4
> >
> 
> I think this is a false positive, which is why I did not address it during
> the last round of Coverity issues.
> 
> As a two-pass function, errors are checked during the first pass, when pf is
> not allocated yet. During the second pass, intermediate functions are not
> supposed to return a different result and "notsup" cannot occur.
> 
> I think assert()/rte_assert() would make more sense and should fool
> Coverity
> into understanding the expected behavior.
> 
> The commit log should reflect that we are addressing a false positive since
> there is no problem with the code logic (of course unless I missed anything
> obvious).

Then, I would just ignore the coverity issue, changing the status to intentional.

Thanks,
Pablo

> 
> Thanks.
> 
> --
> Adrien Mazarguil
> 6WIND