* Re: [PATCH] ipv6: check return value of ipv6_skip_exthdr
2021-11-17 18:16 [PATCH] ipv6: check return value of ipv6_skip_exthdr Jordy Zomer
@ 2021-11-17 18:44 ` Kees Cook
2021-11-17 18:46 ` Kees Cook
2021-11-17 19:06 ` [PATCH v2] " Jordy Zomer
2 siblings, 0 replies; 5+ messages in thread
From: Kees Cook @ 2021-11-17 18:44 UTC (permalink / raw)
To: Jordy Zomer
Cc: linux-kernel, linux-hardening, Steffen Klassert, Herbert Xu,
David S. Miller, Hideaki YOSHIFUJI, David Ahern, Jakub Kicinski,
netdev
On Wed, Nov 17, 2021 at 07:16:10PM +0100, Jordy Zomer wrote:
> The offset value is used in pointer math on skb->data.
> Since ipv6_skip_exthdr may return -1 the pointer to uh and th
> may not point to the actual udp and tcp headers and potentially
> overwrite other stuff. This is why I think this should be checked.
>
> Signed-off-by: Jordy Zomer <jordy@pwning.systems>
> ---
> net/ipv6/esp6.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/net/ipv6/esp6.c b/net/ipv6/esp6.c
> index ed2f061b8768..dc4251655df9 100644
> --- a/net/ipv6/esp6.c
> +++ b/net/ipv6/esp6.c
> @@ -808,6 +808,11 @@ int esp6_input_done2(struct sk_buff *skb, int err)
> struct tcphdr *th;
>
> offset = ipv6_skip_exthdr(skb, offset, &nexthdr, &frag_off);
> +
> + if (offset < 0)
> + err = -EINVAL;
> + goto out;
> +
Ew. Yeah, it seems like ipv6_skip_exthdr() needs to be checked in a lot
of places. If this is part of protocol decoding, I'm surprised fuzzers
haven't found this? Is this state reachable?
I assume so, as there have been similar fixes in the past:
https://git.kernel.org/linus/92c6058024e87087cf1b99b0389d67c0a886360e
> uh = (void *)(skb->data + offset);
> th = (void *)(skb->data + offset);
> hdr_len += offset;
> --
> 2.27.0
>
--
Kees Cook
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] ipv6: check return value of ipv6_skip_exthdr
2021-11-17 18:16 [PATCH] ipv6: check return value of ipv6_skip_exthdr Jordy Zomer
2021-11-17 18:44 ` Kees Cook
@ 2021-11-17 18:46 ` Kees Cook
2021-11-17 19:06 ` [PATCH v2] " Jordy Zomer
2 siblings, 0 replies; 5+ messages in thread
From: Kees Cook @ 2021-11-17 18:46 UTC (permalink / raw)
To: Jordy Zomer
Cc: linux-kernel, linux-hardening, Steffen Klassert, Herbert Xu,
David S. Miller, Hideaki YOSHIFUJI, David Ahern, Jakub Kicinski,
netdev
On Wed, Nov 17, 2021 at 07:16:10PM +0100, Jordy Zomer wrote:
> The offset value is used in pointer math on skb->data.
> Since ipv6_skip_exthdr may return -1 the pointer to uh and th
> may not point to the actual udp and tcp headers and potentially
> overwrite other stuff. This is why I think this should be checked.
>
> Signed-off-by: Jordy Zomer <jordy@pwning.systems>
> ---
> net/ipv6/esp6.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/net/ipv6/esp6.c b/net/ipv6/esp6.c
> index ed2f061b8768..dc4251655df9 100644
> --- a/net/ipv6/esp6.c
> +++ b/net/ipv6/esp6.c
> @@ -808,6 +808,11 @@ int esp6_input_done2(struct sk_buff *skb, int err)
> struct tcphdr *th;
>
> offset = ipv6_skip_exthdr(skb, offset, &nexthdr, &frag_off);
> +
> + if (offset < 0)
> + err = -EINVAL;
> + goto out;
> +
Oh, and, this needs { }s. The CI noticed too:
https://patchwork.kernel.org/project/netdevbpf/patch/20211117181610.2731938-1-jordy@pwning.systems/
> uh = (void *)(skb->data + offset);
> th = (void *)(skb->data + offset);
> hdr_len += offset;
> --
> 2.27.0
>
--
Kees Cook
^ permalink raw reply [flat|nested] 5+ messages in thread* [PATCH v2] ipv6: check return value of ipv6_skip_exthdr
2021-11-17 18:16 [PATCH] ipv6: check return value of ipv6_skip_exthdr Jordy Zomer
2021-11-17 18:44 ` Kees Cook
2021-11-17 18:46 ` Kees Cook
@ 2021-11-17 19:06 ` Jordy Zomer
2021-11-18 11:50 ` patchwork-bot+netdevbpf
2 siblings, 1 reply; 5+ messages in thread
From: Jordy Zomer @ 2021-11-17 19:06 UTC (permalink / raw)
To: linux-kernel
Cc: Kees Cook, Jordy Zomer, Steffen Klassert, Herbert Xu,
David S. Miller, Hideaki YOSHIFUJI, David Ahern, Jakub Kicinski,
netdev
The offset value is used in pointer math on skb->data.
Since ipv6_skip_exthdr may return -1 the pointer to uh and th
may not point to the actual udp and tcp headers and potentially
overwrite other stuff. This is why I think this should be checked.
EDIT: added {}'s, thanks Kees
Signed-off-by: Jordy Zomer <jordy@pwning.systems>
---
net/ipv6/esp6.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/net/ipv6/esp6.c b/net/ipv6/esp6.c
index ed2f061b8768..f0bac6f7ab6b 100644
--- a/net/ipv6/esp6.c
+++ b/net/ipv6/esp6.c
@@ -808,6 +808,12 @@ int esp6_input_done2(struct sk_buff *skb, int err)
struct tcphdr *th;
offset = ipv6_skip_exthdr(skb, offset, &nexthdr, &frag_off);
+
+ if (offset < 0) {
+ err = -EINVAL;
+ goto out;
+ }
+
uh = (void *)(skb->data + offset);
th = (void *)(skb->data + offset);
hdr_len += offset;
--
2.27.0
^ permalink raw reply related [flat|nested] 5+ messages in thread* Re: [PATCH v2] ipv6: check return value of ipv6_skip_exthdr
2021-11-17 19:06 ` [PATCH v2] " Jordy Zomer
@ 2021-11-18 11:50 ` patchwork-bot+netdevbpf
0 siblings, 0 replies; 5+ messages in thread
From: patchwork-bot+netdevbpf @ 2021-11-18 11:50 UTC (permalink / raw)
To: Jordy Zomer
Cc: linux-kernel, keescook, steffen.klassert, herbert, davem,
yoshfuji, dsahern, kuba, netdev
Hello:
This patch was applied to netdev/net.git (master)
by David S. Miller <davem@davemloft.net>:
On Wed, 17 Nov 2021 20:06:48 +0100 you wrote:
> The offset value is used in pointer math on skb->data.
> Since ipv6_skip_exthdr may return -1 the pointer to uh and th
> may not point to the actual udp and tcp headers and potentially
> overwrite other stuff. This is why I think this should be checked.
>
> EDIT: added {}'s, thanks Kees
>
> [...]
Here is the summary with links:
- [v2] ipv6: check return value of ipv6_skip_exthdr
https://git.kernel.org/netdev/net/c/5f9c55c8066b
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 5+ messages in thread