Patchwork [Lucid,Precise,CVE-2013-3222] atm: update msg_namelen in vcc_recvmsg()

login
register
mail settings
Submitter Luis Henriques
Date May 9, 2013, 10:01 a.m.
Message ID <1368093700-6678-1-git-send-email-luis.henriques@canonical.com>
Download mbox | patch
Permalink /patch/242741/
State New
Headers show

Comments

Luis Henriques - May 9, 2013, 10:01 a.m.
From: Mathias Krause <minipli@googlemail.com>

CVE-2013-3222

BugLink: https://bugs.launchpad.net/bugs/1172365

The current code does not fill the msg_name member in case it is set.
It also does not set the msg_namelen member to 0 and therefore makes
net/socket.c leak the local, uninitialized sockaddr_storage variable
to userland -- 128 bytes of kernel stack memory.

Fix that by simply setting msg_namelen to 0 as obviously nobody cared
about vcc_recvmsg() not filling the msg_name in case it was set.

Signed-off-by: Mathias Krause <minipli@googlemail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit 9b3e617f3df53822345a8573b6d358f6b9e5ed87)

Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
---
 net/atm/common.c | 2 ++
 1 file changed, 2 insertions(+)
Andy Whitcroft - May 9, 2013, 10:45 a.m.
On Thu, May 09, 2013 at 11:01:39AM +0100, Luis Henriques wrote:
> From: Mathias Krause <minipli@googlemail.com>
> 
> CVE-2013-3222
> 
> BugLink: https://bugs.launchpad.net/bugs/1172365
> 
> The current code does not fill the msg_name member in case it is set.
> It also does not set the msg_namelen member to 0 and therefore makes
> net/socket.c leak the local, uninitialized sockaddr_storage variable
> to userland -- 128 bytes of kernel stack memory.
> 
> Fix that by simply setting msg_namelen to 0 as obviously nobody cared
> about vcc_recvmsg() not filling the msg_name in case it was set.
> 
> Signed-off-by: Mathias Krause <minipli@googlemail.com>
> Signed-off-by: David S. Miller <davem@davemloft.net>
> (cherry picked from commit 9b3e617f3df53822345a8573b6d358f6b9e5ed87)
> 
> Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
> ---
>  net/atm/common.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/net/atm/common.c b/net/atm/common.c
> index 0baf05e..65737b8 100644
> --- a/net/atm/common.c
> +++ b/net/atm/common.c
> @@ -473,6 +473,8 @@ int vcc_recvmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg,
>  	struct sk_buff *skb;
>  	int copied, error = -EINVAL;
>  
> +	msg->msg_namelen = 0;
> +
>  	if (sock->state != SS_CONNECTED)
>  		return -ENOTCONN;
>  	if (flags & ~MSG_DONTWAIT)		/* only handle MSG_DONTWAIT */

Is an upstream cherry-pick as claimed, seems to do what is expected.

Acked-by: Andy Whitcroft <apw@canonical.com>

-apw
Andy Whitcroft - May 9, 2013, 10:49 a.m.
On Thu, May 09, 2013 at 11:45:16AM +0100, Andy Whitcroft wrote:
> On Thu, May 09, 2013 at 11:01:39AM +0100, Luis Henriques wrote:
> > From: Mathias Krause <minipli@googlemail.com>
> > 
> > CVE-2013-3222
> > 
> > BugLink: https://bugs.launchpad.net/bugs/1172365
> > 
> > The current code does not fill the msg_name member in case it is set.
> > It also does not set the msg_namelen member to 0 and therefore makes
> > net/socket.c leak the local, uninitialized sockaddr_storage variable
> > to userland -- 128 bytes of kernel stack memory.
> > 
> > Fix that by simply setting msg_namelen to 0 as obviously nobody cared
> > about vcc_recvmsg() not filling the msg_name in case it was set.
> > 
> > Signed-off-by: Mathias Krause <minipli@googlemail.com>
> > Signed-off-by: David S. Miller <davem@davemloft.net>
> > (cherry picked from commit 9b3e617f3df53822345a8573b6d358f6b9e5ed87)
> > 
> > Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
> > ---
> >  net/atm/common.c | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/net/atm/common.c b/net/atm/common.c
> > index 0baf05e..65737b8 100644
> > --- a/net/atm/common.c
> > +++ b/net/atm/common.c
> > @@ -473,6 +473,8 @@ int vcc_recvmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg,
> >  	struct sk_buff *skb;
> >  	int copied, error = -EINVAL;
> >  
> > +	msg->msg_namelen = 0;
> > +
> >  	if (sock->state != SS_CONNECTED)
> >  		return -ENOTCONN;
> >  	if (flags & ~MSG_DONTWAIT)		/* only handle MSG_DONTWAIT */
> 
> Is an upstream cherry-pick as claimed, seems to do what is expected.

I note there is a slight difference in context here, but I assume that
is cherry-pick being smart, rather than this being a backport.

> Acked-by: Andy Whitcroft <apw@canonical.com>

-apw
Stefan Bader - May 9, 2013, 1:05 p.m.
Guess un-fuzzed. Not sure whether this then is already a backport or still
(mostly) a cherry-pick.
Luis Henriques - May 9, 2013, 1:19 p.m.
On Thu, May 09, 2013 at 03:05:14PM +0200, Stefan Bader wrote:
> Guess un-fuzzed. Not sure whether this then is already a backport or still
> (mostly) a cherry-pick.
> 

This is a simple cherry-pick.  As Andy stated in his email, its just
git being smart and not a backport.

Cheers,
--
Luis
Tim Gardner - May 9, 2013, 3:24 p.m.

Patch

diff --git a/net/atm/common.c b/net/atm/common.c
index 0baf05e..65737b8 100644
--- a/net/atm/common.c
+++ b/net/atm/common.c
@@ -473,6 +473,8 @@  int vcc_recvmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg,
 	struct sk_buff *skb;
 	int copied, error = -EINVAL;
 
+	msg->msg_namelen = 0;
+
 	if (sock->state != SS_CONNECTED)
 		return -ENOTCONN;
 	if (flags & ~MSG_DONTWAIT)		/* only handle MSG_DONTWAIT */