[PATCH net] net/x25: prevent a couple of overflows

[PATCH net] net/x25: prevent a couple of overflows

[Dan Carpenter]

From: "kiyin(尹亮)" <kiyin@tencent.com>

The .x25_addr[] address comes from the user and is not necessarily
NUL terminated.  This leads to a couple problems.  The first problem is
that the strlen() in x25_bind() can read beyond the end of the buffer.

The second problem is more subtle and could result in memory corruption.
The call tree is:
  x25_connect()
  --> x25_write_internal()
      --> x25_addr_aton()

The .x25_addr[] buffers are copied to the "addresses" buffer from
x25_write_internal() so it will lead to stack corruption.

The x25 protocol only allows 15 character addresses so putting a NUL
terminator as the 16th character is safe and obviously preferable to
reading out of bounds.

Signed-off-by: "kiyin(尹亮)" <kiyin@tencent.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---

 net/x25/af_x25.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c
index 0bbb283f23c9..3180f15942fe 100644
--- a/net/x25/af_x25.c
+++ b/net/x25/af_x25.c
@@ -686,6 +686,8 @@ static int x25_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
 		goto out;
 	}
 
+	addr->sx25_addr.x25_addr[X25_ADDR_LEN - 1] = '\0';
+
 	/* check for the null_x25_address */
 	if (strcmp(addr->sx25_addr.x25_addr, null_x25_address.x25_addr)) {
 
@@ -779,6 +781,7 @@ static int x25_connect(struct socket *sock, struct sockaddr *uaddr,
 		goto out;
 
 	rc = -ENETUNREACH;
+	addr->sx25_addr.x25_addr[X25_ADDR_LEN - 1] = '\0';
 	rt = x25_get_route(&addr->sx25_addr);
 	if (!rt)
 		goto out;
-- 
2.28.0

Re: [PATCH net] net/x25: prevent a couple of overflows

[Martin Schiller]

On 2020-11-30 11:04, Dan Carpenter wrote:
> From: "kiyin(尹亮)" <kiyin@tencent.com>
> 
> The .x25_addr[] address comes from the user and is not necessarily
> NUL terminated.  This leads to a couple problems.  The first problem is
> that the strlen() in x25_bind() can read beyond the end of the buffer.
> 
> The second problem is more subtle and could result in memory 
> corruption.
> The call tree is:
>   x25_connect()
>   --> x25_write_internal()
>       --> x25_addr_aton()
> 
> The .x25_addr[] buffers are copied to the "addresses" buffer from
> x25_write_internal() so it will lead to stack corruption.
> 
> The x25 protocol only allows 15 character addresses so putting a NUL
> terminator as the 16th character is safe and obviously preferable to
> reading out of bounds.
> 

OK, I see the potential danger. I'm just wondering what is the better
approach here to counteract it:
1. check if the string is terminated or exceeds the maximum allowed
    length and report an error if necessary.
2. always terminate the string at byte 15 as you suggested.

> Signed-off-by: "kiyin(尹亮)" <kiyin@tencent.com>
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> ---
> 
>  net/x25/af_x25.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c
> index 0bbb283f23c9..3180f15942fe 100644
> --- a/net/x25/af_x25.c
> +++ b/net/x25/af_x25.c
> @@ -686,6 +686,8 @@ static int x25_bind(struct socket *sock, struct
> sockaddr *uaddr, int addr_len)
>  		goto out;
>  	}
> 
> +	addr->sx25_addr.x25_addr[X25_ADDR_LEN - 1] = '\0';
> +
>  	/* check for the null_x25_address */
>  	if (strcmp(addr->sx25_addr.x25_addr, null_x25_address.x25_addr)) {
> 
> @@ -779,6 +781,7 @@ static int x25_connect(struct socket *sock, struct
> sockaddr *uaddr,
>  		goto out;
> 
>  	rc = -ENETUNREACH;
> +	addr->sx25_addr.x25_addr[X25_ADDR_LEN - 1] = '\0';
>  	rt = x25_get_route(&addr->sx25_addr);
>  	if (!rt)
>  		goto out;

[PATCH net v2] net/x25: prevent a couple of overflows

[Dan Carpenter]

The .x25_addr[] address comes from the user and is not necessarily
NUL terminated.  This leads to a couple problems.  The first problem is
that the strlen() in x25_bind() can read beyond the end of the buffer.

The second problem is more subtle and could result in memory corruption.
The call tree is:
  x25_connect()
  --> x25_write_internal()
      --> x25_addr_aton()

The .x25_addr[] buffers are copied to the "addresses" buffer from
x25_write_internal() so it will lead to stack corruption.

Verify that the strings are NUL terminated and return -EINVAL if they
are not.

Reported-by: "kiyin(尹亮)" <kiyin@tencent.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
The first patch put a NUL terminator on the end of the string and this
patch returns an error instead.  I don't have a strong preference, which
patch to go with.

 net/x25/af_x25.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c
index 9232cdb42ad9..d41fffb2507b 100644
--- a/net/x25/af_x25.c
+++ b/net/x25/af_x25.c
@@ -675,7 +675,8 @@ static int x25_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
 	int len, i, rc = 0;
 
 	if (addr_len != sizeof(struct sockaddr_x25) ||
-	    addr->sx25_family != AF_X25) {
+	    addr->sx25_family != AF_X25 ||
+	    strnlen(addr->sx25_addr.x25_addr, X25_ADDR_LEN) == X25_ADDR_LEN) {
 		rc = -EINVAL;
 		goto out;
 	}
@@ -769,7 +770,8 @@ static int x25_connect(struct socket *sock, struct sockaddr *uaddr,
 
 	rc = -EINVAL;
 	if (addr_len != sizeof(struct sockaddr_x25) ||
-	    addr->sx25_family != AF_X25)
+	    addr->sx25_family != AF_X25 ||
+	    strnlen(addr->sx25_addr.x25_addr, X25_ADDR_LEN) == X25_ADDR_LEN)
 		goto out;
 
 	rc = -ENETUNREACH;
-- 
2.29.2

RE: [PATCH net v2] net/x25: prevent a couple of overflows(Internet mail)

[kiyin(尹亮)]

Hi Dan,
    I think the strnlen is better. the kernel doesn't need to adjust user land mistake by putting a NULL terminator. just return an error to let the user land program fix the wrong address.

Regards,
kiyin

> -----Original Message-----
> From: Dan Carpenter [mailto:dan.carpenter@oracle.com]
> Sent: Tuesday, December 1, 2020 11:15 PM
> To: Martin Schiller <ms@dev.tdt.de>
> Cc: David S. Miller <davem@davemloft.net>; Jakub Kicinski <kuba@kernel.org>;
> linux-x25@vger.kernel.org; netdev@vger.kernel.org; Andrew Hendry
> <andrew.hendry@gmail.com>; kiyin(尹亮) <kiyin@tencent.com>;
> security@kernel.org; linux-distros@vs.openwall.org; huntchen(陈阳)
> <huntchen@tencent.com>; dannywang(王宇) <dannywang@tencent.com>;
> kernel-janitors@vger.kernel.org
> Subject: [PATCH net v2] net/x25: prevent a couple of overflows(Internet mail)
> 
> The .x25_addr[] address comes from the user and is not necessarily NUL
> terminated.  This leads to a couple problems.  The first problem is that the
> strlen() in x25_bind() can read beyond the end of the buffer.
> 
> The second problem is more subtle and could result in memory corruption.
> The call tree is:
>   x25_connect()
>   --> x25_write_internal()
>       --> x25_addr_aton()
> 
> The .x25_addr[] buffers are copied to the "addresses" buffer from
> x25_write_internal() so it will lead to stack corruption.
> 
> Verify that the strings are NUL terminated and return -EINVAL if they are not.
> 
> Reported-by: "kiyin(尹亮)" <kiyin@tencent.com>
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> ---
> The first patch put a NUL terminator on the end of the string and this patch
> returns an error instead.  I don't have a strong preference, which patch to go
> with.
> 
>  net/x25/af_x25.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c index
> 9232cdb42ad9..d41fffb2507b 100644
> --- a/net/x25/af_x25.c
> +++ b/net/x25/af_x25.c
> @@ -675,7 +675,8 @@ static int x25_bind(struct socket *sock, struct
> sockaddr *uaddr, int addr_len)
>  	int len, i, rc = 0;
> 
>  	if (addr_len != sizeof(struct sockaddr_x25) ||
> -	    addr->sx25_family != AF_X25) {
> +	    addr->sx25_family != AF_X25 ||
> +	    strnlen(addr->sx25_addr.x25_addr, X25_ADDR_LEN) ==
> X25_ADDR_LEN) {
>  		rc = -EINVAL;
>  		goto out;
>  	}
> @@ -769,7 +770,8 @@ static int x25_connect(struct socket *sock, struct
> sockaddr *uaddr,
> 
>  	rc = -EINVAL;
>  	if (addr_len != sizeof(struct sockaddr_x25) ||
> -	    addr->sx25_family != AF_X25)
> +	    addr->sx25_family != AF_X25 ||
> +	    strnlen(addr->sx25_addr.x25_addr, X25_ADDR_LEN) ==
> X25_ADDR_LEN)
>  		goto out;
> 
>  	rc = -ENETUNREACH;
> --
> 2.29.2

Re: [PATCH net v2] net/x25: prevent a couple of overflows

[Martin Schiller]

On 2020-12-01 16:15, Dan Carpenter wrote:
> The .x25_addr[] address comes from the user and is not necessarily
> NUL terminated.  This leads to a couple problems.  The first problem is
> that the strlen() in x25_bind() can read beyond the end of the buffer.
> 
> The second problem is more subtle and could result in memory 
> corruption.
> The call tree is:
>   x25_connect()
>   --> x25_write_internal()
>       --> x25_addr_aton()
> 
> The .x25_addr[] buffers are copied to the "addresses" buffer from
> x25_write_internal() so it will lead to stack corruption.
> 
> Verify that the strings are NUL terminated and return -EINVAL if they
> are not.
> 
> Reported-by: "kiyin(尹亮)" <kiyin@tencent.com>
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> ---
> The first patch put a NUL terminator on the end of the string and this
> patch returns an error instead.  I don't have a strong preference, 
> which
> patch to go with.
> 
>  net/x25/af_x25.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c
> index 9232cdb42ad9..d41fffb2507b 100644
> --- a/net/x25/af_x25.c
> +++ b/net/x25/af_x25.c
> @@ -675,7 +675,8 @@ static int x25_bind(struct socket *sock, struct
> sockaddr *uaddr, int addr_len)
>  	int len, i, rc = 0;
> 
>  	if (addr_len != sizeof(struct sockaddr_x25) ||
> -	    addr->sx25_family != AF_X25) {
> +	    addr->sx25_family != AF_X25 ||
> +	    strnlen(addr->sx25_addr.x25_addr, X25_ADDR_LEN) == X25_ADDR_LEN) 
> {
>  		rc = -EINVAL;
>  		goto out;
>  	}
> @@ -769,7 +770,8 @@ static int x25_connect(struct socket *sock, struct
> sockaddr *uaddr,
> 
>  	rc = -EINVAL;
>  	if (addr_len != sizeof(struct sockaddr_x25) ||
> -	    addr->sx25_family != AF_X25)
> +	    addr->sx25_family != AF_X25 ||
> +	    strnlen(addr->sx25_addr.x25_addr, X25_ADDR_LEN) == X25_ADDR_LEN)
>  		goto out;
> 
>  	rc = -ENETUNREACH;

Acked-by: Martin Schiller <ms@dev.tdt.de>

Re: [PATCH net v2] net/x25: prevent a couple of overflows

[Jakub Kicinski]

On Wed, 02 Dec 2020 10:27:18 +0100 Martin Schiller wrote:
> On 2020-12-01 16:15, Dan Carpenter wrote:
> > The .x25_addr[] address comes from the user and is not necessarily
> > NUL terminated.  This leads to a couple problems.  The first problem is
> > that the strlen() in x25_bind() can read beyond the end of the buffer.
> > 
> > The second problem is more subtle and could result in memory 
> > corruption.
> > The call tree is:
> >   x25_connect()  
> >   --> x25_write_internal()
> >       --> x25_addr_aton()  
> > 
> > The .x25_addr[] buffers are copied to the "addresses" buffer from
> > x25_write_internal() so it will lead to stack corruption.
> > 
> > Verify that the strings are NUL terminated and return -EINVAL if they
> > are not.
> > 
> > Reported-by: "kiyin(尹亮)" <kiyin@tencent.com>
> > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> 
> Acked-by: Martin Schiller <ms@dev.tdt.de>

Applied, thanks!