bpf: Replace bpf_lpm_trie_key 0-length array with flexible array

Replace deprecated 0-length array in struct bpf_lpm_trie_key with
flexible array. Found with GCC 13:

../kernel/bpf/lpm_trie.c:207:51: warning: array subscript i is outside array bounds of 'const __u8[0]' {aka 'const unsigned char[]'} [-Warray-bounds=]
207 | *(__be16 *)&key->data[i]);
| ^~~~~~~~~~~~~
../include/uapi/linux/swab.h:102:54: note: in definition of macro '__swab16'
102 | #define __swab16(x) (__u16)__builtin_bswap16((__u16)(x))
| ^
../include/linux/byteorder/generic.h:97:21: note: in expansion of macro '__be16_to_cpu'
97 | #define be16_to_cpu __be16_to_cpu
| ^~~~~~~~~~~~~
../kernel/bpf/lpm_trie.c:206:28: note: in expansion of macro 'be16_to_cpu'
206 | u16 diff = be16_to_cpu(*(__be16 *)&node->data[i]
^
| ^~~~~~~~~~~
In file included from ../include/linux/bpf.h:7:
../include/uapi/linux/bpf.h:82:17: note: while referencing 'data'
82 | __u8 data[0]; /* Arbitrary size */
| ^~~~

And found at run-time under CONFIG_FORTIFY_SOURCE:

UBSAN: array-index-out-of-bounds in kernel/bpf/lpm_trie.c:218:49
index 0 is out of range for type '__u8 [*]'

Changing struct bpf_lpm_trie_key is difficult since has been used by
userspace. For example, in Cilium:

struct egress_gw_policy_key {
struct bpf_lpm_trie_key lpm_key;
__u32 saddr;
__u32 daddr;
};

While direct references to the "data" member haven't been found, there
are static initializers what include the final member. For example,
the "{}" here:

struct egress_gw_policy_key in_key = {
.lpm_key = { 32 + 24, {} },
.saddr = CLIENT_IP,
.daddr = EXTERNAL_SVC_IP & 0Xffffff,
};

To avoid the build time and run time warnings seen with a 0-sized
trailing array for struct bpf_lpm_trie_key, introduce a new struct
that correctly uses a flexible array for the trailing bytes,
struct bpf_lpm_trie_key_u8. As part of this, include the "header"
portion (which is just the "prefixlen" member), so it can be used
by anything building a bpf_lpr_trie_key that has trailing members that
aren't a u8 flexible array (like the self-test[1]), which is named
struct bpf_lpm_trie_key_hdr.

Unfortunately, C++ refuses to parse the __struct_group() helper, so
it is not possible to define struct bpf_lpm_trie_key_hdr directly in
struct bpf_lpm_trie_key_u8, so we must open-code the union directly.

Adjust the kernel code to use struct bpf_lpm_trie_key_u8 through-out,
and for the selftest to use struct bpf_lpm_trie_key_hdr. Add a comment
to the UAPI header directing folks to the two new options.

Reported-by: Mark Rutland <[e-mail]>
Signed-off-by: Kees Cook <[e-mail]>
Signed-off-by: Daniel Borkmann <[e-mail]>
Acked-by: Gustavo A. R. Silva <[e-mail]>
Closes: https://paste.debian.net/hidden/ca500597/
Link: https://lore.kernel.org/all/202206281009.4332AA33@keescook/ [1]
Link: https://lore.kernel.org/bpf/[e-mail]

Documentation/bpf/map_lpm_trie.rst

···
 
 LPM tries may be created with a maximum prefix length that is a multiple
 of 8, in the range from 8 to 2048. The key used for lookup and update
-operations is a ``struct bpf_lpm_trie_key``, extended by
+operations is a ``struct bpf_lpm_trie_key_u8``, extended by
 ``max_prefixlen/8`` bytes.
 
 - For IPv4 addresses the data length is 4 bytes

include/uapi/linux/bpf.h

···
 	__s32	imm;		/* signed immediate constant */
 };
 
-/* Key of an a BPF_MAP_TYPE_LPM_TRIE entry */
+/* Deprecated: use struct bpf_lpm_trie_key_u8 (when the "data" member is needed for
+ * byte access) or struct bpf_lpm_trie_key_hdr (when using an alternative type for
+ * the trailing flexible array member) instead.
+ */
 struct bpf_lpm_trie_key {
 	__u32	prefixlen;	/* up to 32 for AF_INET, 128 for AF_INET6 */
 	__u8	data[0];	/* Arbitrary size */
+};
+
+/* Header for bpf_lpm_trie_key structs */
+struct bpf_lpm_trie_key_hdr {
+	__u32	prefixlen;
+};
+
+/* Key of an a BPF_MAP_TYPE_LPM_TRIE entry, with trailing byte array. */
+struct bpf_lpm_trie_key_u8 {
+	union {
+		struct bpf_lpm_trie_key_hdr	hdr;
+		__u32				prefixlen;
+	};
+	__u8	data[];		/* Arbitrary size */
 };
 
 struct bpf_cgroup_storage_key {

kernel/bpf/lpm_trie.c

···
  */
 static size_t longest_prefix_match(const struct lpm_trie *trie,
 				   const struct lpm_trie_node *node,
-				   const struct bpf_lpm_trie_key *key)
+				   const struct bpf_lpm_trie_key_u8 *key)
 {
 	u32 limit = min(node->prefixlen, key->prefixlen);
 	u32 prefixlen = 0, i = 0;
 
 	BUILD_BUG_ON(offsetof(struct lpm_trie_node, data) % sizeof(u32));
-	BUILD_BUG_ON(offsetof(struct bpf_lpm_trie_key, data) % sizeof(u32));
+	BUILD_BUG_ON(offsetof(struct bpf_lpm_trie_key_u8, data) % sizeof(u32));
 
 #if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) && defined(CONFIG_64BIT)
 
···
 {
 	struct lpm_trie *trie = container_of(map, struct lpm_trie, map);
 	struct lpm_trie_node *node, *found = NULL;
-	struct bpf_lpm_trie_key *key = _key;
+	struct bpf_lpm_trie_key_u8 *key = _key;
 
 	if (key->prefixlen > trie->max_prefixlen)
 		return NULL;
···
 	struct lpm_trie *trie = container_of(map, struct lpm_trie, map);
 	struct lpm_trie_node *node, *im_node = NULL, *new_node = NULL;
 	struct lpm_trie_node __rcu **slot;
-	struct bpf_lpm_trie_key *key = _key;
+	struct bpf_lpm_trie_key_u8 *key = _key;
 	unsigned long irq_flags;
 	unsigned int next_bit;
 	size_t matchlen = 0;
···
 static long trie_delete_elem(struct bpf_map *map, void *_key)
 {
 	struct lpm_trie *trie = container_of(map, struct lpm_trie, map);
-	struct bpf_lpm_trie_key *key = _key;
+	struct bpf_lpm_trie_key_u8 *key = _key;
 	struct lpm_trie_node __rcu **trim, **trim2;
 	struct lpm_trie_node *node, *parent;
 	unsigned long irq_flags;
···
 				 sizeof(struct lpm_trie_node))
 #define LPM_VAL_SIZE_MIN	1
 
-#define LPM_KEY_SIZE(X)		(sizeof(struct bpf_lpm_trie_key) + (X))
+#define LPM_KEY_SIZE(X)		(sizeof(struct bpf_lpm_trie_key_u8) + (X))
 #define LPM_KEY_SIZE_MAX	LPM_KEY_SIZE(LPM_DATA_SIZE_MAX)
 #define LPM_KEY_SIZE_MIN	LPM_KEY_SIZE(LPM_DATA_SIZE_MIN)
 
···
 	/* copy mandatory map attributes */
 	bpf_map_init_from_attr(&trie->map, attr);
 	trie->data_size = attr->key_size -
-			  offsetof(struct bpf_lpm_trie_key, data);
+			  offsetof(struct bpf_lpm_trie_key_u8, data);
 	trie->max_prefixlen = trie->data_size * 8;
 
 	spin_lock_init(&trie->lock);
···
 {
 	struct lpm_trie_node *node, *next_node = NULL, *parent, *search_root;
 	struct lpm_trie *trie = container_of(map, struct lpm_trie, map);
-	struct bpf_lpm_trie_key *key = _key, *next_key = _next_key;
+	struct bpf_lpm_trie_key_u8 *key = _key, *next_key = _next_key;
 	struct lpm_trie_node **node_stack = NULL;
 	int err = 0, stack_ptr = -1;
 	unsigned int next_bit;
···
 	}
 do_copy:
 	next_key->prefixlen = next_node->prefixlen;
-	memcpy((void *)next_key + offsetof(struct bpf_lpm_trie_key, data),
+	memcpy((void *)next_key + offsetof(struct bpf_lpm_trie_key_u8, data),
 	       next_node->data, trie->data_size);
 free_stack:
 	kfree(node_stack);
···
 			  const struct btf_type *key_type,
 			  const struct btf_type *value_type)
 {
-	/* Keys must have struct bpf_lpm_trie_key embedded. */
+	/* Keys must have struct bpf_lpm_trie_key_u8 embedded. */
 	return BTF_INFO_KIND(key_type->info) != BTF_KIND_STRUCT ?
 	       -EINVAL : 0;
 }

samples/bpf/map_perf_test_user.c

···
 
 static void fill_lpm_trie(void)
 {
-	struct bpf_lpm_trie_key *key;
+	struct bpf_lpm_trie_key_u8 *key;
 	unsigned long value = 0;
 	unsigned int i;
 	int r;

samples/bpf/xdp_router_ipv4_user.c

···
 static void read_route(struct nlmsghdr *nh, int nll)
 {
 	char dsts[24], gws[24], ifs[16], dsts_len[24], metrics[24];
-	struct bpf_lpm_trie_key *prefix_key;
+	struct bpf_lpm_trie_key_u8 *prefix_key;
 	struct rtattr *rt_attr;
 	struct rtmsg *rt_msg;
 	int rtm_family;

tools/include/uapi/linux/bpf.h

···
 	__s32	imm;		/* signed immediate constant */
 };
 
-/* Key of an a BPF_MAP_TYPE_LPM_TRIE entry */
+/* Deprecated: use struct bpf_lpm_trie_key_u8 (when the "data" member is needed for
+ * byte access) or struct bpf_lpm_trie_key_hdr (when using an alternative type for
+ * the trailing flexible array member) instead.
+ */
 struct bpf_lpm_trie_key {
 	__u32	prefixlen;	/* up to 32 for AF_INET, 128 for AF_INET6 */
 	__u8	data[0];	/* Arbitrary size */
+};
+
+/* Header for bpf_lpm_trie_key structs */
+struct bpf_lpm_trie_key_hdr {
+	__u32	prefixlen;
+};
+
+/* Key of an a BPF_MAP_TYPE_LPM_TRIE entry, with trailing byte array. */
+struct bpf_lpm_trie_key_u8 {
+	union {
+		struct bpf_lpm_trie_key_hdr	hdr;
+		__u32				prefixlen;
+	};
+	__u8	data[];		/* Arbitrary size */
 };
 
 struct bpf_cgroup_storage_key {

tools/testing/selftests/bpf/progs/map_ptr_kern.c

···
 } __attribute__((preserve_access_index));
 
 struct lpm_key {
-	struct bpf_lpm_trie_key trie_key;
+	struct bpf_lpm_trie_key_hdr trie_key;
 	__u32 data;
 };
 

tools/testing/selftests/bpf/test_lpm_map.c

···
 	volatile size_t n_matches, n_matches_after_delete;
 	size_t i, j, n_nodes, n_lookups;
 	struct tlpm_node *t, *list = NULL;
-	struct bpf_lpm_trie_key *key;
+	struct bpf_lpm_trie_key_u8 *key;
 	uint8_t *data, *value;
 	int r, map;
 
···
 static void test_lpm_ipaddr(void)
 {
 	LIBBPF_OPTS(bpf_map_create_opts, opts, .map_flags = BPF_F_NO_PREALLOC);
-	struct bpf_lpm_trie_key *key_ipv4;
-	struct bpf_lpm_trie_key *key_ipv6;
+	struct bpf_lpm_trie_key_u8 *key_ipv4;
+	struct bpf_lpm_trie_key_u8 *key_ipv6;
 	size_t key_size_ipv4;
 	size_t key_size_ipv6;
 	int map_fd_ipv4;
···
 static void test_lpm_delete(void)
 {
 	LIBBPF_OPTS(bpf_map_create_opts, opts, .map_flags = BPF_F_NO_PREALLOC);
-	struct bpf_lpm_trie_key *key;
+	struct bpf_lpm_trie_key_u8 *key;
 	size_t key_size;
 	int map_fd;
 	__u64 value;
···
 static void test_lpm_get_next_key(void)
 {
 	LIBBPF_OPTS(bpf_map_create_opts, opts, .map_flags = BPF_F_NO_PREALLOC);
-	struct bpf_lpm_trie_key *key_p, *next_key_p;
+	struct bpf_lpm_trie_key_u8 *key_p, *next_key_p;
 	size_t key_size;
 	__u32 value = 0;
 	int map_fd;
···
 {
 	int i, j, ret, iter, key_size;
 	struct lpm_mt_test_info *info = arg;
-	struct bpf_lpm_trie_key *key_p;
+	struct bpf_lpm_trie_key_u8 *key_p;
 
-	key_size = sizeof(struct bpf_lpm_trie_key) + sizeof(__u32);
+	key_size = sizeof(*key_p) + sizeof(__u32);
 	key_p = alloca(key_size);
 	for (iter = 0; iter < info->iter; iter++)
 		for (i = 0; i < MAX_TEST_KEYS; i++) {
···
 				ret = bpf_map_lookup_elem(info->map_fd, key_p, &value);
 				assert(ret == 0 || errno == ENOENT);
 			} else {
-				struct bpf_lpm_trie_key *next_key_p = alloca(key_size);
+				struct bpf_lpm_trie_key_u8 *next_key_p = alloca(key_size);
 				ret = bpf_map_get_next_key(info->map_fd, key_p, next_key_p);
 				assert(ret == 0 || errno == ENOENT || errno == ENOMEM);
 			}
···
 
 	/* create a trie */
 	value_size = sizeof(__u32);
-	key_size = sizeof(struct bpf_lpm_trie_key) + value_size;
+	key_size = sizeof(struct bpf_lpm_trie_key_hdr) + value_size;
 	map_fd = bpf_map_create(BPF_MAP_TYPE_LPM_TRIE, NULL, key_size, value_size, 100, &opts);
 
 	/* create 4 threads to test update, delete, lookup and get_next_key */