The `IFNAMSIZ` macro defines the required buffer size to hold a Linux
interface name including the terminating zero byte while netifd currently
uses an `IFNAMSIZ + 1` limit for interface name buffers.
This causes netifd to use overlong names (16 instead of 15 bytes) in
netlink communication with the kernel, leading to netlink failure replies
due to policy violations.
Fix this issue by applying the correct length, that is `IFNAMSIZ` directly,
to the corresponding buffers.
Ref: https://github.com/openwrt/openwrt/issues/11259
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
struct vlist_tree vlans;
struct kvlist vlan_aliases;
- char ifname[IFNAMSIZ + 1];
+ char ifname[IFNAMSIZ];
int ifindex;
struct blob_attr *config;
struct interface_user out_iface_user;
/* device name */
- char in_dev[IFNAMSIZ + 1];
- char out_dev[IFNAMSIZ + 1];
+ char in_dev[IFNAMSIZ];
+ char out_dev[IFNAMSIZ];
/* everything below is used as avl tree key */
/* don't change the order */
{
const char *port_name = port->ifname;
const char op_ch = add ? '+' : '-';
- char buf[IFNAMSIZ + 2];
+ char buf[IFNAMSIZ + 1];
snprintf(buf, sizeof(buf), "%c%s", op_ch, port_name);
system_if_down(port);
static void vlan_dev_cb(struct device_user *dep, enum device_event ev)
{
- char name[IFNAMSIZ + 1];
+ char name[IFNAMSIZ];
struct vlan_device *vldev;
vldev = container_of(dep, struct vlan_device, dep);
};
struct vlan_device *vldev;
struct device_user *dep;
- char name[IFNAMSIZ + 1];
+ char name[IFNAMSIZ];
char *err = NULL;
int id, *alias_id;