Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1323804 > unrolled thread

[PATCH V5] netfilter: h323: avoid potential attack

Started byZhouyi Zhou <zhouzhouyi@gmail.com>
First post2016-02-02 06:50 +0100
Last post2016-02-02 12:00 +0100
Articles 2 — 2 participants

Back to article view | Back to linux.kernel


Contents

  [PATCH V5] netfilter: h323: avoid potential attack Zhouyi Zhou <zhouzhouyi@gmail.com> - 2016-02-02 06:50 +0100
    Re: [PATCH V5] netfilter: h323: avoid potential attack Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> - 2016-02-02 12:00 +0100

#1323804 — [PATCH V5] netfilter: h323: avoid potential attack

FromZhouyi Zhou <zhouzhouyi@gmail.com>
Date2016-02-02 06:50 +0100
Subject[PATCH V5] netfilter: h323: avoid potential attack
Message-ID<qXC0G-6aS-7@gated-at.bofh.it>
I think hackers chould build a malicious h323 packet to overflow
the pointer p which will panic during the memcpy(addr, p, len)
For example, he may fabricate a very large taddr->ipAddress.ip.

In order to avoid this, I add a valid memory reference check in
 get_h2x5_addr functions.

As suggested by Eric, this module is protected by a lock (nf_h323_lock)
so adding a variable h323_buffer_valid_bytes that would contain
the number of valid bytes would not require to change prototypes of
get_h2x5_addr.

Signed-off-by: Zhouyi Zhou <yizhouzhou@ict.ac.cn>

---
 net/netfilter/nf_conntrack_h323_main.c | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

diff --git a/net/netfilter/nf_conntrack_h323_main.c b/net/netfilter/nf_conntrack_h323_main.c
index 9511af0..21665ec 100644
--- a/net/netfilter/nf_conntrack_h323_main.c
+++ b/net/netfilter/nf_conntrack_h323_main.c
@@ -110,6 +110,25 @@ int (*nat_q931_hook) (struct sk_buff *skb,
 
 static DEFINE_SPINLOCK(nf_h323_lock);
 static char *h323_buffer;
+static int h323_buffer_valid_bytes;
+
+static bool h323_buffer_ref_valid(void *p, int len)
+{
+
+	if ((unsigned long)len > h323_buffer_valid_bytes) {
+		return false; 
+	}
+
+	if (p + len > (void *)h323_buffer + h323_buffer_valid_bytes) {
+		return false;
+	}
+
+	if (p < (void *)h323_buffer) {
+		return false;
+	}
+
+	return true;
+}
 
 static struct nf_conntrack_helper nf_conntrack_helper_h245;
 static struct nf_conntrack_helper nf_conntrack_helper_q931[];
@@ -145,6 +164,7 @@ static int get_tpkt_data(struct sk_buff *skb, unsigned int protoff,
 
 	if (*data == NULL) {	/* first TPKT */
 		/* Get first TPKT pointer */
+		h323_buffer_valid_bytes = tcpdatalen;
 		tpkt = skb_header_pointer(skb, tcpdataoff, tcpdatalen,
 					  h323_buffer);
 		BUG_ON(tpkt == NULL);
@@ -247,6 +267,9 @@ static int get_h245_addr(struct nf_conn *ct, const unsigned char *data,
 		return 0;
 	}
 
+	if (!h323_buffer_ref_valid((void *)p, len + sizeof(__be16)))
+		return 0;
+
 	memcpy(addr, p, len);
 	memset((void *)addr + len, 0, sizeof(*addr) - len);
 	memcpy(port, p + len, sizeof(__be16));
@@ -669,6 +692,9 @@ int get_h225_addr(struct nf_conn *ct, unsigned char *data,
 		return 0;
 	}
 
+	if (!h323_buffer_ref_valid((void *)p, len + sizeof(__be16)))
+		return 0;
+
 	memcpy(addr, p, len);
 	memset((void *)addr + len, 0, sizeof(*addr) - len);
 	memcpy(port, p + len, sizeof(__be16));
@@ -1248,6 +1274,7 @@ static unsigned char *get_udp_data(struct sk_buff *skb, unsigned int protoff,
 	if (dataoff >= skb->len)
 		return NULL;
 	*datalen = skb->len - dataoff;
+	h323_buffer_valid_bytes = *datalen;
 	return skb_header_pointer(skb, dataoff, *datalen, h323_buffer);
 }
 
-- 
1.9.1

[toc] | [next] | [standalone]


#1323968

FromSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Date2016-02-02 12:00 +0100
Message-ID<qXGQH-1uV-19@gated-at.bofh.it>
In reply to#1323804
Hello.

On 2/2/2016 8:37 AM, Zhouyi Zhou wrote:

> I think hackers chould build a malicious h323 packet to overflow
> the pointer p which will panic during the memcpy(addr, p, len)
> For example, he may fabricate a very large taddr->ipAddress.ip.
>
> In order to avoid this, I add a valid memory reference check in
>   get_h2x5_addr functions.
>
> As suggested by Eric, this module is protected by a lock (nf_h323_lock)
> so adding a variable h323_buffer_valid_bytes that would contain
> the number of valid bytes would not require to change prototypes of
> get_h2x5_addr.
>
> Signed-off-by: Zhouyi Zhou <yizhouzhou@ict.ac.cn>
>
> ---
>   net/netfilter/nf_conntrack_h323_main.c | 27 +++++++++++++++++++++++++++
>   1 file changed, 27 insertions(+)
>
> diff --git a/net/netfilter/nf_conntrack_h323_main.c b/net/netfilter/nf_conntrack_h323_main.c
> index 9511af0..21665ec 100644
> --- a/net/netfilter/nf_conntrack_h323_main.c
> +++ b/net/netfilter/nf_conntrack_h323_main.c
> @@ -110,6 +110,25 @@ int (*nat_q931_hook) (struct sk_buff *skb,
>
>   static DEFINE_SPINLOCK(nf_h323_lock);
>   static char *h323_buffer;
> +static int h323_buffer_valid_bytes;
> +
> +static bool h323_buffer_ref_valid(void *p, int len)
> +{
> +
> +	if ((unsigned long)len > h323_buffer_valid_bytes) {
> +		return false;
> +	}

    {} not needed.

> +
> +	if (p + len > (void *)h323_buffer + h323_buffer_valid_bytes) {
> +		return false;
> +	}

    Likewise.

> +
> +	if (p < (void *)h323_buffer) {
> +		return false;
> +	}

    Likewise.

[...]

MBR, Sergei

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web