[PATCH] net: af_key: Add error check in set_sadb_address()

[PATCH] net: af_key: Add error check in set_sadb_address()

[Wentao Liang]

The function set_sadb_address() calls the function
pfkey_sockaddr_fill(), but does not check its return value.
A proper implementation can be found in set_sadb_kmaddress().

Add an error check for set_sadb_address(), return error code
if the function fails.

Fixes: e5b56652c11b ("key: Share common code path to fill sockaddr{}.")
Cc: stable@vger.kernel.org # v2.6
Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>
---
 net/key/af_key.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/net/key/af_key.c b/net/key/af_key.c
index c56bb4f451e6..537c9604e356 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -3474,15 +3474,17 @@ static int set_sadb_address(struct sk_buff *skb, int sasize, int type,
 	switch (type) {
 	case SADB_EXT_ADDRESS_SRC:
 		addr->sadb_address_prefixlen = sel->prefixlen_s;
-		pfkey_sockaddr_fill(&sel->saddr, 0,
-				    (struct sockaddr *)(addr + 1),
-				    sel->family);
+		if (!pfkey_sockaddr_fill(&sel->saddr, 0,
+					 (struct sockaddr *)(addr + 1),
+					 sel->family))
+			return -EINVAL;
 		break;
 	case SADB_EXT_ADDRESS_DST:
 		addr->sadb_address_prefixlen = sel->prefixlen_d;
-		pfkey_sockaddr_fill(&sel->daddr, 0,
-				    (struct sockaddr *)(addr + 1),
-				    sel->family);
+		if (!pfkey_sockaddr_fill(&sel->daddr, 0,
+					 (struct sockaddr *)(addr + 1),
+					 sel->family))
+			return -EINVAL;
 		break;
 	default:
 		return -EINVAL;
-- 
2.42.0.windows.2

Re: [PATCH] net: af_key: Add error check in set_sadb_address()

[Michal Swiatkowski]

On Sun, May 25, 2025 at 11:53:50PM +0800, Wentao Liang wrote:
> The function set_sadb_address() calls the function
> pfkey_sockaddr_fill(), but does not check its return value.
> A proper implementation can be found in set_sadb_kmaddress().
> 
> Add an error check for set_sadb_address(), return error code
> if the function fails.
> 
> Fixes: e5b56652c11b ("key: Share common code path to fill sockaddr{}.")
> Cc: stable@vger.kernel.org # v2.6
> Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>
> ---
>  net/key/af_key.c | 14 ++++++++------
>  1 file changed, 8 insertions(+), 6 deletions(-)
> 
> diff --git a/net/key/af_key.c b/net/key/af_key.c
> index c56bb4f451e6..537c9604e356 100644
> --- a/net/key/af_key.c
> +++ b/net/key/af_key.c
> @@ -3474,15 +3474,17 @@ static int set_sadb_address(struct sk_buff *skb, int sasize, int type,
>  	switch (type) {
>  	case SADB_EXT_ADDRESS_SRC:
>  		addr->sadb_address_prefixlen = sel->prefixlen_s;
> -		pfkey_sockaddr_fill(&sel->saddr, 0,
> -				    (struct sockaddr *)(addr + 1),
> -				    sel->family);
> +		if (!pfkey_sockaddr_fill(&sel->saddr, 0,
> +					 (struct sockaddr *)(addr + 1),
> +					 sel->family))
> +			return -EINVAL;
>  		break;
>  	case SADB_EXT_ADDRESS_DST:
>  		addr->sadb_address_prefixlen = sel->prefixlen_d;
> -		pfkey_sockaddr_fill(&sel->daddr, 0,
> -				    (struct sockaddr *)(addr + 1),
> -				    sel->family);
> +		if (!pfkey_sockaddr_fill(&sel->daddr, 0,
> +					 (struct sockaddr *)(addr + 1),
> +					 sel->family))
> +			return -EINVAL;
>  		break;
>  	default:
>  		return -EINVAL;

There are few other calls to pfkey_sockaddr_fill() without checking, but
family is already checked in such case, so it is fine.

Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>

I am not sure if it should be a fix. If family is set there is no
problem. Probably it is set in all cases. Maybe you should target it to
net-next, but as I said, I am not sure.

Thanks

> -- 
> 2.42.0.windows.2

Re: [PATCH] net: af_key: Add error check in set_sadb_address()

[Jakub Kicinski]

On Sun, 25 May 2025 23:53:50 +0800 Wentao Liang wrote:
> The function set_sadb_address() calls the function
> pfkey_sockaddr_fill(), but does not check its return value.
> A proper implementation can be found in set_sadb_kmaddress().
> 
> Add an error check for set_sadb_address(), return error code
> if the function fails.

Please look at the callers, and you'll find out that the family has
already been validated.
-- 
pw-bot: reject