Re: [PATCH] netfilter: ipset: harden payload calculation in call_ad()

[Pablo Neira Ayuso <pablo@netfilter.org>]

On Fri, Apr 10, 2026 at 03:05:18AM +0200, Pablo Neira Ayuso wrote:
> On Fri, Mar 13, 2026 at 01:01:32PM -0500, David Baum wrote:
> > call_ad() computes the netlink error payload size with
> > min(SIZE_MAX, sizeof(*errmsg) + nlmsg_len(nlh)), but min(SIZE_MAX, x)
> > is always x, so the guard is a no-op.
> > 
> > Replace it with an explicit negative-length check and
> > check_add_overflow() so the addition is validated before being passed
> > to nlmsg_new().
> > 
> > Signed-off-by: David Baum <davidbaum461@gmail.com>
> > ---
> >  net/netfilter/ipset/ip_set_core.c | 10 ++++++++--
> >  1 file changed, 8 insertions(+), 2 deletions(-)
> > 
> > diff --git a/net/netfilter/ipset/ip_set_core.c b/net/netfilter/ipset/ip_set_core.c
> > index a2fe711cb5e3..11d3854d9b11 100644
> > --- a/net/netfilter/ipset/ip_set_core.c
> > +++ b/net/netfilter/ipset/ip_set_core.c
> > @@ -10,6 +10,7 @@
> >  #include <linux/module.h>
> >  #include <linux/moduleparam.h>
> >  #include <linux/ip.h>
> > +#include <linux/overflow.h>
> >  #include <linux/skbuff.h>
> >  #include <linux/spinlock.h>
> >  #include <linux/rculist.h>
> > @@ -1763,13 +1764,18 @@ call_ad(struct net *net, struct sock *ctnl, struct sk_buff *skb,
> >  		struct nlmsghdr *rep, *nlh = nlmsg_hdr(skb);
> >  		struct sk_buff *skb2;
> >  		struct nlmsgerr *errmsg;
> > -		size_t payload = min(SIZE_MAX,
> > -				     sizeof(*errmsg) + nlmsg_len(nlh));
> > +		int nlmsg_payload_len = nlmsg_len(nlh);
> > +		size_t payload;
> >  		int min_len = nlmsg_total_size(sizeof(struct nfgenmsg));
> >  		struct nlattr *cda[IPSET_ATTR_CMD_MAX + 1];
> >  		struct nlattr *cmdattr;
> >  		u32 *errline;
> >  
> > +		if (nlmsg_payload_len < 0 ||
> 
> Hm, nlh was already sanitized by nfnetlink?
> 
> > +		    check_add_overflow(sizeof(*errmsg),
> > +				       (size_t)nlmsg_payload_len, &payload))
> 
> Maybe cap this to int:
> 
>                 int payload = sizeof(struct nlmsgerr);
>                 ...
>         
>                 if (check_add_overflow(payload, nlmsg_payload_len, &payload))

Wait, then payload and nlmsg_payload_len should be __u32, not int.

> 
> > +			return -ENOMEM;
> > +
> >  		skb2 = nlmsg_new(payload, GFP_KERNEL);
> >  		if (!skb2)
> >  			return -ENOMEM;
>