bpf: Devmap adjust uapi for attach bpf program

V2:
- Defer changing BPF-syscall to start at file-descriptor 1
- Use {} to zero initialise struct.

The recent commit fbee97f ("bpf: Add support to attach bpf program to a
devmap entry"), introduced ability to attach (and run) a separate XDP
bpf_prog for each devmap entry. A bpf_prog is added via a file-descriptor.
As zero were a valid FD, not using the feature requires using value minus-1.
The UAPI is extended via tail-extending struct bpf_devmap_val and using
map->value_size to determine the feature set.

This will break older userspace applications not using the bpf_prog feature.
Consider an old userspace app that is compiled against newer kernel
uapi/bpf.h, it will not know that it need to initialise the member
bpf_prog.fd to minus-1. Thus, users will be forced to update source code to
get program running on newer kernels.

This patch remove the minus-1 checks, and have zero mean feature isn't used.

Followup patches either for kernel or libbpf should handle and avoid
returning file-descriptor zero in the first place.

Fixes: fbee97f ("bpf: Add support to attach bpf program to a devmap entry")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/159170950687.2102545.7235914718298050113.stgit@firesoul

include/uapi/linux/bpf.h

@@ -3761,6 +3761,19 @@ struct xdp_md {
 	__u32 egress_ifindex;  /* txq->dev->ifindex */
 };
 
+/* DEVMAP map-value layout
+ *
+ * The struct data-layout of map-value is a configuration interface.
+ * New members can only be added to the end of this structure.
+ */
+struct bpf_devmap_val {
+	__u32 ifindex;   /* device index */
+	union {
+		int   fd;  /* prog fd on map write */
+		__u32 id;  /* prog id on map read */
+	} bpf_prog;
+};
+
 enum sk_action {
 	SK_DROP = 0,
 	SK_PASS,

kernel/bpf/devmap.c

@@ -60,15 +60,6 @@ struct xdp_dev_bulk_queue {
 	unsigned int count;
 };
 
-/* DEVMAP values */
-struct bpf_devmap_val {
-	u32 ifindex;   /* device index */
-	union {
-		int fd;  /* prog fd on map write */
-		u32 id;  /* prog id on map read */
-	} bpf_prog;
-};
-
 struct bpf_dtab_netdev {
 	struct net_device *dev; /* must be first member, due to tracepoint */
 	struct hlist_node index_hlist;
@@ -619,7 +610,7 @@ static struct bpf_dtab_netdev *__dev_map_alloc_node(struct net *net,
 	if (!dev->dev)
 		goto err_out;
 
-	if (val->bpf_prog.fd >= 0) {
+	if (val->bpf_prog.fd > 0) {
 		prog = bpf_prog_get_type_dev(val->bpf_prog.fd,
 					     BPF_PROG_TYPE_XDP, false);
 		if (IS_ERR(prog))
@@ -653,8 +644,8 @@ static int __dev_map_update_elem(struct net *net, struct bpf_map *map,
 				 void *key, void *value, u64 map_flags)
 {
 	struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
-	struct bpf_devmap_val val = { .bpf_prog.fd = -1 };
 	struct bpf_dtab_netdev *dev, *old_dev;
+	struct bpf_devmap_val val = {};
 	u32 i = *(u32 *)key;
 
 	if (unlikely(map_flags > BPF_EXIST))
@@ -670,7 +661,7 @@ static int __dev_map_update_elem(struct net *net, struct bpf_map *map,
 	if (!val.ifindex) {
 		dev = NULL;
 		/* can not specify fd if ifindex is 0 */
-		if (val.bpf_prog.fd != -1)
+		if (val.bpf_prog.fd > 0)
 			return -EINVAL;
 	} else {
 		dev = __dev_map_alloc_node(net, dtab, &val, i);
@@ -700,8 +691,8 @@ static int __dev_map_hash_update_elem(struct net *net, struct bpf_map *map,
 				     void *key, void *value, u64 map_flags)
 {
 	struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
-	struct bpf_devmap_val val = { .bpf_prog.fd = -1 };
 	struct bpf_dtab_netdev *dev, *old_dev;
+	struct bpf_devmap_val val = {};
 	u32 idx = *(u32 *)key;
 	unsigned long flags;
 	int err = -EEXIST;