Re: [PATCH] selftests/net/ipsec: Fix Null pointer dereference in rtattr_pack()

[Simon Horman]

On Tue, Jan 14, 2025 at 03:43:29PM +0800, liuye wrote:
>     Fix the following warning.

I think it is a bit more than a warning, I'd phrase this more like,
even as it repeats the subject. Also, it would be nice to cite
the tool that generates the warning.

Address Null pointer dereference in rtattr_pack.

Flagged by ??? as:

> 
>     tools/testing/selftests/net/ipsec.c:230:25: warning: Possible null pointer
>     dereference: payload [nullPointer]
>     memcpy(RTA_DATA(attr), payload, size);
>                            ^
>     tools/testing/selftests/net/ipsec.c:1618:54: note: Calling function 'rtattr_pack',
>     4th argument 'NULL' value is 0
>     if (rtattr_pack(&req.nh, sizeof(req), XFRMA_IF_ID, NULL, 0)) {
>                                                        ^
>     tools/testing/selftests/net/ipsec.c:230:25: note: Null pointer dereference
>     memcpy(RTA_DATA(attr), payload, size);
>                            ^
> 

And I wonder if a fixes tag is appropriate, and if so this one:

70bfdf62e93a ("selftests/net/ipsec: Add test for xfrm_spdattr_type_t")

And, accordingly if this patch should be targeted at net:

	[PATCH net] ...

> Signed-off-by: liuye <liuye@xxxxxxxxxx>

Please consider separating out your family and given name in hte
signed-off-by line. Perhaps Lin Ye (apologies if that is incorrect).

The above not withstanding, the code change looks good to me.
So feel free to include the following in a v2 with an updated patch
description.

> ---
>  tools/testing/selftests/net/ipsec.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/tools/testing/selftests/net/ipsec.c b/tools/testing/selftests/net/ipsec.c
> index be4a30a0d02a..725310ac26a9 100644
> --- a/tools/testing/selftests/net/ipsec.c
> +++ b/tools/testing/selftests/net/ipsec.c
> @@ -227,7 +227,8 @@ static int rtattr_pack(struct nlmsghdr *nh, size_t req_sz,
>  
>  	attr->rta_len = RTA_LENGTH(size);
>  	attr->rta_type = rta_type;
> -	memcpy(RTA_DATA(attr), payload, size);
> +	if (payload != NULL)

I think it would be more idiomatic to express this as:

	if (payload)

> +		memcpy(RTA_DATA(attr), payload, size);
>  
>  	return 0;
>  }
> -- 
> 2.25.1
>