The Linux Kernel Mailing List
 help / color / mirror / Atom feed
* [PATCH] ipv6: ndisc: fix potential out-of-bounds read in ndisc_router_discovery()
@ 2025-06-02  6:58 Rafal Bilkowski
  2025-06-02 21:32 ` Kuniyuki Iwashima
  0 siblings, 1 reply; 2+ messages in thread
From: Rafal Bilkowski @ 2025-06-02  6:58 UTC (permalink / raw)
  To: netdev
  Cc: linux-kernel, davem, dsahern, edumazet, kuba, pabeni, alex.aring,
	rafalbilkowski

This patch adds a length check at the start of ndisc_router_discovery()
to prevent a potential out-of-bounds read if a short packet is received.
Without this check, the function may dereference memory outside the
buffer.

Fixes: 8610c7c6e3bd ("net: ipv6: add support for rpl sr exthdr")
Signed-off-by: Rafal Bilkowski <rafalbilkowski@gmail.com>
---
 net/ipv6/ndisc.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
index 8d853971f2f6..bdaac5a195d6 100644
--- a/net/ipv6/ndisc.c
+++ b/net/ipv6/ndisc.c
@@ -1235,6 +1235,10 @@ static void ndisc_ra_useropt(struct sk_buff *ra, struct nd_opt_hdr *opt)
 
 static enum skb_drop_reason ndisc_router_discovery(struct sk_buff *skb)
 {
+	// Check if the buffer contains enough data for the struct ra_msg
+	if (!pskb_may_pull(skb, skb_transport_offset(skb) + sizeof(struct ra_msg)))
+		return SKB_DROP_REASON_PKT_TOO_SMALL;
+
 	struct ra_msg *ra_msg = (struct ra_msg *)skb_transport_header(skb);
 	bool send_ifinfo_notify = false;
 	struct neighbour *neigh = NULL;
-- 
2.43.0


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH] ipv6: ndisc: fix potential out-of-bounds read in ndisc_router_discovery()
  2025-06-02  6:58 [PATCH] ipv6: ndisc: fix potential out-of-bounds read in ndisc_router_discovery() Rafal Bilkowski
@ 2025-06-02 21:32 ` Kuniyuki Iwashima
  0 siblings, 0 replies; 2+ messages in thread
From: Kuniyuki Iwashima @ 2025-06-02 21:32 UTC (permalink / raw)
  To: rafalbilkowski
  Cc: alex.aring, davem, dsahern, edumazet, kuba, linux-kernel, netdev,
	pabeni

From: Rafal Bilkowski <rafalbilkowski@gmail.com>
Date: Mon,  2 Jun 2025 08:58:26 +0200
> This patch adds a length check at the start of ndisc_router_discovery()

I think it already has the check.

---8<---
        optlen = (skb_tail_pointer(skb) - skb_transport_header(skb)) -
                sizeof(struct ra_msg);
...
        if (optlen < 0)
                return SKB_DROP_REASON_PKT_TOO_SMALL;
---8<---


> to prevent a potential out-of-bounds read if a short packet is received.
> Without this check, the function may dereference memory outside the
> buffer.

Do you have a KMSAM splat ?


> 
> Fixes: 8610c7c6e3bd ("net: ipv6: add support for rpl sr exthdr")
> Signed-off-by: Rafal Bilkowski <rafalbilkowski@gmail.com>
> ---
>  net/ipv6/ndisc.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
> index 8d853971f2f6..bdaac5a195d6 100644
> --- a/net/ipv6/ndisc.c
> +++ b/net/ipv6/ndisc.c
> @@ -1235,6 +1235,10 @@ static void ndisc_ra_useropt(struct sk_buff *ra, struct nd_opt_hdr *opt)
>  
>  static enum skb_drop_reason ndisc_router_discovery(struct sk_buff *skb)
>  {
> +	// Check if the buffer contains enough data for the struct ra_msg
> +	if (!pskb_may_pull(skb, skb_transport_offset(skb) + sizeof(struct ra_msg)))
> +		return SKB_DROP_REASON_PKT_TOO_SMALL;

Please run checkpatch.pl.  We don't add code before variable declarations.

Also, the skb is linearized in ndisc_rcv().

> +
>  	struct ra_msg *ra_msg = (struct ra_msg *)skb_transport_header(skb);
>  	bool send_ifinfo_notify = false;
>  	struct neighbour *neigh = NULL;
> -- 
> 2.43.0

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2025-06-02 21:33 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-06-02  6:58 [PATCH] ipv6: ndisc: fix potential out-of-bounds read in ndisc_router_discovery() Rafal Bilkowski
2025-06-02 21:32 ` Kuniyuki Iwashima

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox