2014-07-02 09:26:15

by Emil Goode

Subject: [PATCH] Staging: rtl8192e: Fix potential NULL pointer dereference

We need to make sure the struct rtllib_device pointer ieee is not NULL
after the goto rx_dropped label since it is dereferenced there.

Signed-off-by: Emil Goode <[email protected]>
---
drivers/staging/rtl8192e/rtllib_rx.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/staging/rtl8192e/rtllib_rx.c b/drivers/staging/rtl8192e/rtllib_rx.c
index 60de54c..7db3e74 100644
--- a/drivers/staging/rtl8192e/rtllib_rx.c
+++ b/drivers/staging/rtl8192e/rtllib_rx.c
@@ -1496,7 +1496,8 @@ int rtllib_rx(struct rtllib_device *ieee, struct sk_buff *skb,
return ret;

rx_dropped:
- ieee->stats.rx_dropped++;
+ if (ieee)
+ ieee->stats.rx_dropped++;
return 0;
}
EXPORT_SYMBOL(rtllib_rx);
--
1.7.10.4

2014-07-02 16:33:50

by Greg Kroah-Hartman

Subject: Re: [PATCH] Staging: rtl8192e: Fix potential NULL pointer dereference

On Wed, Jul 02, 2014 at 11:25:51AM +0200, Emil Goode wrote:
> We need to make sure the struct rtllib_device pointer ieee is not NULL
> after the goto rx_dropped label since it is dereferenced there.
>
> Signed-off-by: Emil Goode <[email protected]>
> ---
> drivers/staging/rtl8192e/rtllib_rx.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/staging/rtl8192e/rtllib_rx.c b/drivers/staging/rtl8192e/rtllib_rx.c
> index 60de54c..7db3e74 100644
> --- a/drivers/staging/rtl8192e/rtllib_rx.c
> +++ b/drivers/staging/rtl8192e/rtllib_rx.c
> @@ -1496,7 +1496,8 @@ int rtllib_rx(struct rtllib_device *ieee, struct sk_buff *skb,
> return ret;
>
> rx_dropped:
> - ieee->stats.rx_dropped++;
> + if (ieee)
> + ieee->stats.rx_dropped++;
> return 0;
> }
> EXPORT_SYMBOL(rtllib_rx);

Is this something that is hitting users today in the tree, or is this
just a bug you found looking at the code?

thanks,

greg k-h

2014-07-02 17:26:38

by Emil Goode

Subject: Re: [PATCH] Staging: rtl8192e: Fix potential NULL pointer dereference

Hello Greg,

On Wed, Jul 02, 2014 at 09:33:34AM -0700, Greg Kroah-Hartman wrote:
> On Wed, Jul 02, 2014 at 11:25:51AM +0200, Emil Goode wrote:
> > We need to make sure the struct rtllib_device pointer ieee is not NULL
> > after the goto rx_dropped label since it is dereferenced there.
> >
> > Signed-off-by: Emil Goode <[email protected]>
> > ---
> > drivers/staging/rtl8192e/rtllib_rx.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/staging/rtl8192e/rtllib_rx.c b/drivers/staging/rtl8192e/rtllib_rx.c
> > index 60de54c..7db3e74 100644
> > --- a/drivers/staging/rtl8192e/rtllib_rx.c
> > +++ b/drivers/staging/rtl8192e/rtllib_rx.c
> > @@ -1496,7 +1496,8 @@ int rtllib_rx(struct rtllib_device *ieee, struct sk_buff *skb,
> > return ret;
> >
> > rx_dropped:
> > - ieee->stats.rx_dropped++;
> > + if (ieee)
> > + ieee->stats.rx_dropped++;
> > return 0;
> > }
> > EXPORT_SYMBOL(rtllib_rx);
>
> Is this something that is hitting users today in the tree, or is this
> just a bug you found looking at the code?

It's a static checker fix and I'm not aware of any impact on users.
If you want I will resend with that information added?

Best regards,

Emil Goode

2014-07-02 17:33:12

by Greg Kroah-Hartman

Subject: Re: [PATCH] Staging: rtl8192e: Fix potential NULL pointer dereference

On Wed, Jul 02, 2014 at 07:26:20PM +0200, Emil Goode wrote:
> Hello Greg,
>
> On Wed, Jul 02, 2014 at 09:33:34AM -0700, Greg Kroah-Hartman wrote:
> > On Wed, Jul 02, 2014 at 11:25:51AM +0200, Emil Goode wrote:
> > > We need to make sure the struct rtllib_device pointer ieee is not NULL
> > > after the goto rx_dropped label since it is dereferenced there.
> > >
> > > Signed-off-by: Emil Goode <[email protected]>
> > > ---
> > > drivers/staging/rtl8192e/rtllib_rx.c | 3 ++-
> > > 1 file changed, 2 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/staging/rtl8192e/rtllib_rx.c b/drivers/staging/rtl8192e/rtllib_rx.c
> > > index 60de54c..7db3e74 100644
> > > --- a/drivers/staging/rtl8192e/rtllib_rx.c
> > > +++ b/drivers/staging/rtl8192e/rtllib_rx.c
> > > @@ -1496,7 +1496,8 @@ int rtllib_rx(struct rtllib_device *ieee, struct sk_buff *skb,
> > > return ret;
> > >
> > > rx_dropped:
> > > - ieee->stats.rx_dropped++;
> > > + if (ieee)
> > > + ieee->stats.rx_dropped++;
> > > return 0;
> > > }
> > > EXPORT_SYMBOL(rtllib_rx);
> >
> > Is this something that is hitting users today in the tree, or is this
> > just a bug you found looking at the code?
>
> It's a static checker fix and I'm not aware of any impact on users.
> If you want I will resend with that information added?

No, that's fine, just trying to see which kernel to apply this to...

thanks,

greg k-h