Merge branch 'bpf-type-format'

Martin KaFai Lau says:

====================
This patch introduces BPF Type Format (BTF).

BTF (BPF Type Format) is the meta data format which describes
the data types of BPF program/map.  Hence, it basically focus
on the C programming language which the modern BPF is primary
using.  The first use case is to provide a generic pretty print
capability for a BPF map.

A modified pahole that can convert dwarf to BTF is here:

  https://github.com/iamkafai/pahole/tree/btf

Please see individual patch for details.

v5:
- Remove BTF_KIND_FLOAT and BTF_KIND_FUNC which are not
  currently used.  They can be added in the future.
  Some bpf_df_xxx() are removed together.
- Add comment in patch 7 to clarify that the new bpffs_map_fops
  should not be extended further.

v4:
- Fix warning (remove unneeded semicolon)
- Remove a redundant variable (nr_bytes) from btf_int_check_meta() in
  patch 1.  Caught by W=1.

v3:
- Rebase to bpf-next
- Fix sparse warning (by adding static)
- Add BTF header logging: btf_verifier_log_hdr()
- Fix the alignment test on btf->type_off
- Add tests for the BTF header
- Lower the max BTF size to 16MB.  It should be enough
  for some time.  We could raise it later if it would
  be needed.

v2:
- Use kvfree where needed in patch 1 and 2
- Also consider BTF_INT_OFFSET() in the btf_int_check_meta()
  in patch 1
- Fix an incorrect goto target in map_create() during
  the btf-error-path in patch 7
- re-org some local vars to keep the rev xmas tree in btf.c
====================

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

1 files changed, 135 insertions, 13 deletions

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 3d35bacf656f..6513e0b08795 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -45,6 +45,7 @@
 
 #include "libbpf.h"
 #include "bpf.h"
+#include "btf.h"
 
 #ifndef EM_BPF
 #define EM_BPF 247
@@ -212,6 +213,8 @@ struct bpf_map {
 	char *name;
 	size_t offset;
 	struct bpf_map_def def;
+	uint32_t btf_key_id;
+	uint32_t btf_value_id;
 	void *priv;
 	bpf_map_clear_priv_t clear_priv;
 };
@@ -256,6 +259,8 @@ struct bpf_object {
 	 */
 	struct list_head list;
 
+	struct btf *btf;
+
 	void *priv;
 	bpf_object_clear_priv_t clear_priv;
 
@@ -819,7 +824,15 @@ static int bpf_object__elf_collect(struct bpf_object *obj)
 							data->d_size);
 		else if (strcmp(name, "maps") == 0)
 			obj->efile.maps_shndx = idx;
-		else if (sh.sh_type == SHT_SYMTAB) {
+		else if (strcmp(name, BTF_ELF_SEC) == 0) {
+			obj->btf = btf__new(data->d_buf, data->d_size,
+					    __pr_debug);
+			if (IS_ERR(obj->btf)) {
+				pr_warning("Error loading ELF section %s: %ld. Ignored and continue.\n",
+					   BTF_ELF_SEC, PTR_ERR(obj->btf));
+				obj->btf = NULL;
+			}
+		} else if (sh.sh_type == SHT_SYMTAB) {
 			if (obj->efile.symbols) {
 				pr_warning("bpf: multiple SYMTAB in %s\n",
 					   obj->path);
@@ -996,33 +1009,126 @@ bpf_program__collect_reloc(struct bpf_program *prog, GElf_Shdr *shdr,
 	return 0;
 }
 
+static int bpf_map_find_btf_info(struct bpf_map *map, const struct btf *btf)
+{
+	struct bpf_map_def *def = &map->def;
+	const size_t max_name = 256;
+	int64_t key_size, value_size;
+	int32_t key_id, value_id;
+	char name[max_name];
+
+	/* Find key type by name from BTF */
+	if (snprintf(name, max_name, "%s_key", map->name) == max_name) {
+		pr_warning("map:%s length of BTF key_type:%s_key is too long\n",
+			   map->name, map->name);
+		return -EINVAL;
+	}
+
+	key_id = btf__find_by_name(btf, name);
+	if (key_id < 0) {
+		pr_debug("map:%s key_type:%s cannot be found in BTF\n",
+			 map->name, name);
+		return key_id;
+	}
+
+	key_size = btf__resolve_size(btf, key_id);
+	if (key_size < 0) {
+		pr_warning("map:%s key_type:%s cannot get the BTF type_size\n",
+			   map->name, name);
+		return key_size;
+	}
+
+	if (def->key_size != key_size) {
+		pr_warning("map:%s key_type:%s has BTF type_size:%ld != key_size:%u\n",
+			   map->name, name, key_size, def->key_size);
+		return -EINVAL;
+	}
+
+	/* Find value type from BTF */
+	if (snprintf(name, max_name, "%s_value", map->name) == max_name) {
+		pr_warning("map:%s length of BTF value_type:%s_value is too long\n",
+			  map->name, map->name);
+		return -EINVAL;
+	}
+
+	value_id = btf__find_by_name(btf, name);
+	if (value_id < 0) {
+		pr_debug("map:%s value_type:%s cannot be found in BTF\n",
+			 map->name, name);
+		return value_id;
+	}
+
+	value_size = btf__resolve_size(btf, value_id);
+	if (value_size < 0) {
+		pr_warning("map:%s value_type:%s cannot get the BTF type_size\n",
+			   map->name, name);
+		return value_size;
+	}
+
+	if (def->value_size != value_size) {
+		pr_warning("map:%s value_type:%s has BTF type_size:%ld != value_size:%u\n",
+			   map->name, name, value_size, def->value_size);
+		return -EINVAL;
+	}
+
+	map->btf_key_id = key_id;
+	map->btf_value_id = value_id;
+
+	return 0;
+}
+
 static int
 bpf_object__create_maps(struct bpf_object *obj)
 {
+	struct bpf_create_map_attr create_attr = {};
 	unsigned int i;
+	int err;
 
 	for (i = 0; i < obj->nr_maps; i++) {
-		struct bpf_map_def *def = &obj->maps[i].def;
-		int *pfd = &obj->maps[i].fd;
-
-		*pfd = bpf_create_map_name(def->type,
-					   obj->maps[i].name,
-					   def->key_size,
-					   def->value_size,
-					   def->max_entries,
-					   def->map_flags);
+		struct bpf_map *map = &obj->maps[i];
+		struct bpf_map_def *def = &map->def;
+		int *pfd = &map->fd;
+
+		create_attr.name = map->name;
+		create_attr.map_type = def->type;
+		create_attr.map_flags = def->map_flags;
+		create_attr.key_size = def->key_size;
+		create_attr.value_size = def->value_size;
+		create_attr.max_entries = def->max_entries;
+		create_attr.btf_fd = 0;
+		create_attr.btf_key_id = 0;
+		create_attr.btf_value_id = 0;
+
+		if (obj->btf && !bpf_map_find_btf_info(map, obj->btf)) {
+			create_attr.btf_fd = btf__fd(obj->btf);
+			create_attr.btf_key_id = map->btf_key_id;
+			create_attr.btf_value_id = map->btf_value_id;
+		}
+
+		*pfd = bpf_create_map_xattr(&create_attr);
+		if (*pfd < 0 && create_attr.btf_key_id) {
+			pr_warning("Error in bpf_create_map_xattr(%s):%s(%d). Retrying without BTF.\n",
+				   map->name, strerror(errno), errno);
+			create_attr.btf_fd = 0;
+			create_attr.btf_key_id = 0;
+			create_attr.btf_value_id = 0;
+			map->btf_key_id = 0;
+			map->btf_value_id = 0;
+			*pfd = bpf_create_map_xattr(&create_attr);
+		}
+
 		if (*pfd < 0) {
 			size_t j;
-			int err = *pfd;
 
+			err = *pfd;
 			pr_warning("failed to create map (name: '%s'): %s\n",
-				   obj->maps[i].name,
+				   map->name,
 				   strerror(errno));
 			for (j = 0; j < i; j++)
 				zclose(obj->maps[j].fd);
 			return err;
 		}
-		pr_debug("create map %s: fd=%d\n", obj->maps[i].name, *pfd);
+		pr_debug("create map %s: fd=%d\n", map->name, *pfd);
 	}
 
 	return 0;
@@ -1641,6 +1747,7 @@ void bpf_object__close(struct bpf_object *obj)
 
 	bpf_object__elf_finish(obj);
 	bpf_object__unload(obj);
+	btf__free(obj->btf);
 
 	for (i = 0; i < obj->nr_maps; i++) {
 		zfree(&obj->maps[i].name);
@@ -1692,6 +1799,11 @@ unsigned int bpf_object__kversion(struct bpf_object *obj)
 	return obj ? obj->kern_version : 0;
 }
 
+int bpf_object__btf_fd(const struct bpf_object *obj)
+{
+	return obj->btf ? btf__fd(obj->btf) : -1;
+}
+
 int bpf_object__set_priv(struct bpf_object *obj, void *priv,
 			 bpf_object_clear_priv_t clear_priv)
 {
@@ -1937,6 +2049,16 @@ const char *bpf_map__name(struct bpf_map *map)
 	return map ? map->name : NULL;
 }
 
+uint32_t bpf_map__btf_key_id(const struct bpf_map *map)
+{
+	return map ? map->btf_key_id : 0;
+}
+
+uint32_t bpf_map__btf_value_id(const struct bpf_map *map)
+{
+	return map ? map->btf_value_id : 0;
+}
+
 int bpf_map__set_priv(struct bpf_map *map, void *priv,
 		     bpf_map_clear_priv_t clear_priv)
 {