On Wed, Jul 16, 2025 at 11:47:09AM +0200, Florian Westphal wrote:
> Phil Sutter <phil@xxxxxx> wrote:
> > + len = strlen(dev_array[0].ifname) + 1;
> > + if (dev_array[0].ifname[len - 2] == '*')
> > + len -= 2;
>
> Not obvious to me, is there a guarantee that 'len' is 2?
> And, what if len yields 0 here?
Oh, right. Same goes for the introduced libnftnl helpers.
> > + if (dev_array[i].ifname[len - 2] == '*')
> > + len -= 2;
> > + mnl_attr_put(nlh, NFTA_DEVICE_NAME, len, dev_array[i].ifname);
> > mnl_attr_nest_end(nlh, nest_dev);
> > }
> > }
> > @@ -2084,14 +2090,17 @@ static void mnl_nft_ft_devs_build(struct nlmsghdr *nlh, struct cmd *cmd)
> > {
> > const struct expr *dev_expr = cmd->flowtable->dev_expr;
> > const struct nft_dev *dev_array;
> > + int i, len, num_devs = 0;
> > struct nlattr *nest_dev;
> > - int i, num_devs= 0;
> >
> > dev_array = nft_dev_array(dev_expr, &num_devs);
> > nest_dev = mnl_attr_nest_start(nlh, NFTA_FLOWTABLE_HOOK_DEVS);
> > for (i = 0; i < num_devs; i++) {
> > cmd_add_loc(cmd, nlh, dev_array[i].location);
> > - mnl_attr_put_strz(nlh, NFTA_DEVICE_NAME, dev_array[i].ifname);
> > + len = strlen(dev_array[i].ifname) + 1;
> > + if (dev_array[i].ifname[len - 2] == '*')
> > + len -= 2;
> > + mnl_attr_put(nlh, NFTA_DEVICE_NAME, len, dev_array[i].ifname);
>
> This (test, subtract, put) is a repeating pattern, perhaps this warrants a helper?
It's even there already, (still) named 'mnl_attr_put_ifname' in my
libnftnl patch. I'll export it (with correct prefix) for use in
nftables.
Thanks, Phil
