bpf: btf: Add pretty print support to the basic arraymap

This patch adds pretty print support to the basic arraymap.
Support for other bpf maps can be added later.

This patch adds new attrs to the BPF_MAP_CREATE command to allow
specifying the btf_fd, btf_key_id and btf_value_id.  The
BPF_MAP_CREATE can then associate the btf to the map if
the creating map supports BTF.

A BTF supported map needs to implement two new map ops,
map_seq_show_elem() and map_check_btf().  This patch has
implemented these new map ops for the basic arraymap.

It also adds file_operations, bpffs_map_fops, to the pinned
map such that the pinned map can be opened and read.
After that, the user has an intuitive way to do
"cat bpffs/pathto/a-pinned-map" instead of getting
an error.

bpffs_map_fops should not be extended further to support
other operations.  Other operations (e.g. write/key-lookup...)
should be realized by the userspace tools (e.g. bpftool) through
the BPF_OBJ_GET_INFO_BY_FD, map's lookup/update interface...etc.
Follow up patches will allow the userspace to obtain
the BTF from a map-fd.

Here is a sample output when reading a pinned arraymap
with the following map's value:

struct map_value {
	int count_a;
	int count_b;
};

cat /sys/fs/bpf/pinned_array_map:

0: {1,2}
1: {3,4}
2: {5,6}
...

Signed-off-by: Martin KaFai Lau <[email protected]>
Acked-by: Alexei Starovoitov <[email protected]>
Signed-off-by: Daniel Borkmann <[email protected]>

Author: iamkafai

include/linux/bpf.h

@@ -22,6 +22,8 @@ struct perf_event;
 struct bpf_prog;
 struct bpf_map;
 struct sock;
+struct seq_file;
+struct btf;
 
 /* map is generic key/value storage optionally accesible by eBPF programs */
 struct bpf_map_ops {
@@ -43,10 +45,14 @@ struct bpf_map_ops {
 	void (*map_fd_put_ptr)(void *ptr);
 	u32 (*map_gen_lookup)(struct bpf_map *map, struct bpf_insn *insn_buf);
 	u32 (*map_fd_sys_lookup_elem)(void *ptr);
+	void (*map_seq_show_elem)(struct bpf_map *map, void *key,
+				  struct seq_file *m);
+	int (*map_check_btf)(const struct bpf_map *map, const struct btf *btf,
+			     u32 key_type_id, u32 value_type_id);
 };
 
 struct bpf_map {
-	/* 1st cacheline with read-mostly members of which some
+	/* The first two cachelines with read-mostly members of which some
 	 * are also accessed in fast-path (e.g. ops, max_entries).
 	 */
 	const struct bpf_map_ops *ops ____cacheline_aligned;
@@ -62,10 +68,13 @@ struct bpf_map {
 	u32 pages;
 	u32 id;
 	int numa_node;
+	u32 btf_key_id;
+	u32 btf_value_id;
+	struct btf *btf;
 	bool unpriv_array;
-	/* 7 bytes hole */
+	/* 55 bytes hole */
 
-	/* 2nd cacheline with misc members to avoid false sharing
+	/* The 3rd and 4th cacheline with misc members to avoid false sharing
 	 * particularly with refcounting.
 	 */
 	struct user_struct *user ____cacheline_aligned;
@@ -100,6 +109,11 @@ static inline struct bpf_offloaded_map *map_to_offmap(struct bpf_map *map)
 	return container_of(map, struct bpf_offloaded_map, map);
 }
 
+static inline bool bpf_map_support_seq_show(const struct bpf_map *map)
+{
+	return map->ops->map_seq_show_elem && map->ops->map_check_btf;
+}
+
 extern const struct bpf_map_ops bpf_map_offload_ops;
 
 /* function argument constraints */

include/uapi/linux/bpf.h

@@ -280,6 +280,9 @@ union bpf_attr {
 					 */
 		char	map_name[BPF_OBJ_NAME_LEN];
 		__u32	map_ifindex;	/* ifindex of netdev to create on */
+		__u32	btf_fd;		/* fd pointing to a BTF type data */
+		__u32	btf_key_id;	/* BTF type_id of the key */
+		__u32	btf_value_id;	/* BTF type_id of the value */
 	};
 
 	struct { /* anonymous struct used by BPF_MAP_*_ELEM commands */

kernel/bpf/arraymap.c

@@ -11,11 +11,13 @@
  * General Public License for more details.
  */
 #include <linux/bpf.h>
+#include <linux/btf.h>
 #include <linux/err.h>
 #include <linux/slab.h>
 #include <linux/mm.h>
 #include <linux/filter.h>
 #include <linux/perf_event.h>
+#include <uapi/linux/btf.h>
 
 #include "map_in_map.h"
 
@@ -336,6 +338,52 @@ static void array_map_free(struct bpf_map *map)
 	bpf_map_area_free(array);
 }
 
+static void array_map_seq_show_elem(struct bpf_map *map, void *key,
+				    struct seq_file *m)
+{
+	void *value;
+
+	rcu_read_lock();
+
+	value = array_map_lookup_elem(map, key);
+	if (!value) {
+		rcu_read_unlock();
+		return;
+	}
+
+	seq_printf(m, "%u: ", *(u32 *)key);
+	btf_type_seq_show(map->btf, map->btf_value_id, value, m);
+	seq_puts(m, "\n");
+
+	rcu_read_unlock();
+}
+
+static int array_map_check_btf(const struct bpf_map *map, const struct btf *btf,
+			       u32 btf_key_id, u32 btf_value_id)
+{
+	const struct btf_type *key_type, *value_type;
+	u32 key_size, value_size;
+	u32 int_data;
+
+	key_type = btf_type_id_size(btf, &btf_key_id, &key_size);
+	if (!key_type || BTF_INFO_KIND(key_type->info) != BTF_KIND_INT)
+		return -EINVAL;
+
+	int_data = *(u32 *)(key_type + 1);
+	/* bpf array can only take a u32 key.  This check makes
+	 * sure that the btf matches the attr used during map_create.
+	 */
+	if (BTF_INT_BITS(int_data) != 32 || key_size != 4 ||
+	    BTF_INT_OFFSET(int_data))
+		return -EINVAL;
+
+	value_type = btf_type_id_size(btf, &btf_value_id, &value_size);
+	if (!value_type || value_size > map->value_size)
+		return -EINVAL;
+
+	return 0;
+}
+
 const struct bpf_map_ops array_map_ops = {
 	.map_alloc_check = array_map_alloc_check,
 	.map_alloc = array_map_alloc,
@@ -345,6 +393,8 @@ const struct bpf_map_ops array_map_ops = {
 	.map_update_elem = array_map_update_elem,
 	.map_delete_elem = array_map_delete_elem,
 	.map_gen_lookup = array_map_gen_lookup,
+	.map_seq_show_elem = array_map_seq_show_elem,
+	.map_check_btf = array_map_check_btf,
 };
 
 const struct bpf_map_ops percpu_array_map_ops = {

kernel/bpf/inode.c

@@ -150,15 +150,162 @@ static int bpf_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
 	return 0;
 }
 
+struct map_iter {
+	void *key;
+	bool done;
+};
+
+static struct map_iter *map_iter(struct seq_file *m)
+{
+	return m->private;
+}
+
+static struct bpf_map *seq_file_to_map(struct seq_file *m)
+{
+	return file_inode(m->file)->i_private;
+}
+
+static void map_iter_free(struct map_iter *iter)
+{
+	if (iter) {
+		kfree(iter->key);
+		kfree(iter);
+	}
+}
+
+static struct map_iter *map_iter_alloc(struct bpf_map *map)
+{
+	struct map_iter *iter;
+
+	iter = kzalloc(sizeof(*iter), GFP_KERNEL | __GFP_NOWARN);
+	if (!iter)
+		goto error;
+
+	iter->key = kzalloc(map->key_size, GFP_KERNEL | __GFP_NOWARN);
+	if (!iter->key)
+		goto error;
+
+	return iter;
+
+error:
+	map_iter_free(iter);
+	return NULL;
+}
+
+static void *map_seq_next(struct seq_file *m, void *v, loff_t *pos)
+{
+	struct bpf_map *map = seq_file_to_map(m);
+	void *key = map_iter(m)->key;
+
+	if (map_iter(m)->done)
+		return NULL;
+
+	if (unlikely(v == SEQ_START_TOKEN))
+		goto done;
+
+	if (map->ops->map_get_next_key(map, key, key)) {
+		map_iter(m)->done = true;
+		return NULL;
+	}
+
+done:
+	++(*pos);
+	return key;
+}
+
+static void *map_seq_start(struct seq_file *m, loff_t *pos)
+{
+	if (map_iter(m)->done)
+		return NULL;
+
+	return *pos ? map_iter(m)->key : SEQ_START_TOKEN;
+}
+
+static void map_seq_stop(struct seq_file *m, void *v)
+{
+}
+
+static int map_seq_show(struct seq_file *m, void *v)
+{
+	struct bpf_map *map = seq_file_to_map(m);
+	void *key = map_iter(m)->key;
+
+	if (unlikely(v == SEQ_START_TOKEN)) {
+		seq_puts(m, "# WARNING!! The output is for debug purpose only\n");
+		seq_puts(m, "# WARNING!! The output format will change\n");
+	} else {
+		map->ops->map_seq_show_elem(map, key, m);
+	}
+
+	return 0;
+}
+
+static const struct seq_operations bpffs_map_seq_ops = {
+	.start	= map_seq_start,
+	.next	= map_seq_next,
+	.show	= map_seq_show,
+	.stop	= map_seq_stop,
+};
+
+static int bpffs_map_open(struct inode *inode, struct file *file)
+{
+	struct bpf_map *map = inode->i_private;
+	struct map_iter *iter;
+	struct seq_file *m;
+	int err;
+
+	iter = map_iter_alloc(map);
+	if (!iter)
+		return -ENOMEM;
+
+	err = seq_open(file, &bpffs_map_seq_ops);
+	if (err) {
+		map_iter_free(iter);
+		return err;
+	}
+
+	m = file->private_data;
+	m->private = iter;
+
+	return 0;
+}
+
+static int bpffs_map_release(struct inode *inode, struct file *file)
+{
+	struct seq_file *m = file->private_data;
+
+	map_iter_free(map_iter(m));
+
+	return seq_release(inode, file);
+}
+
+/* bpffs_map_fops should only implement the basic
+ * read operation for a BPF map.  The purpose is to
+ * provide a simple user intuitive way to do
+ * "cat bpffs/pathto/a-pinned-map".
+ *
+ * Other operations (e.g. write, lookup...) should be realized by
+ * the userspace tools (e.g. bpftool) through the
+ * BPF_OBJ_GET_INFO_BY_FD and the map's lookup/update
+ * interface.
+ */
+static const struct file_operations bpffs_map_fops = {
+	.open		= bpffs_map_open,
+	.read		= seq_read,
+	.release	= bpffs_map_release,
+};
+
 static int bpf_mkobj_ops(struct dentry *dentry, umode_t mode, void *raw,
-			 const struct inode_operations *iops)
+			 const struct inode_operations *iops,
+			 const struct file_operations *fops)
 {
 	struct inode *dir = dentry->d_parent->d_inode;
 	struct inode *inode = bpf_get_inode(dir->i_sb, dir, mode);
 	if (IS_ERR(inode))
 		return PTR_ERR(inode);
 
 	inode->i_op = iops;
+	inode->i_fop = fops;
 	inode->i_private = raw;
 
 	bpf_dentry_finalize(dentry, inode, dir);
@@ -167,12 +314,15 @@ static int bpf_mkobj_ops(struct dentry *dentry, umode_t mode, void *raw,
 
 static int bpf_mkprog(struct dentry *dentry, umode_t mode, void *arg)
 {
-	return bpf_mkobj_ops(dentry, mode, arg, &bpf_prog_iops);
+	return bpf_mkobj_ops(dentry, mode, arg, &bpf_prog_iops, NULL);
 }
 
 static int bpf_mkmap(struct dentry *dentry, umode_t mode, void *arg)
 {
-	return bpf_mkobj_ops(dentry, mode, arg, &bpf_map_iops);
+	struct bpf_map *map = arg;
+
+	return bpf_mkobj_ops(dentry, mode, arg, &bpf_map_iops,
+			     map->btf ? &bpffs_map_fops : NULL);
 }
 
 static struct dentry *

kernel/bpf/syscall.c

@@ -27,6 +27,7 @@
 #include <linux/cred.h>
 #include <linux/timekeeping.h>
 #include <linux/ctype.h>
+#include <linux/btf.h>
 
 #define IS_FD_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PROG_ARRAY || \
 			   (map)->map_type == BPF_MAP_TYPE_PERF_EVENT_ARRAY || \
@@ -251,6 +252,7 @@ static void bpf_map_free_deferred(struct work_struct *work)
 
 	bpf_map_uncharge_memlock(map);
 	security_bpf_map_free(map);
+	btf_put(map->btf);
 	/* implementation dependent freeing */
 	map->ops->map_free(map);
 }
@@ -416,7 +418,7 @@ static int bpf_obj_name_cpy(char *dst, const char *src)
 	return 0;
 }
 
-#define BPF_MAP_CREATE_LAST_FIELD map_ifindex
+#define BPF_MAP_CREATE_LAST_FIELD btf_value_id
 /* called via syscall */
 static int map_create(union bpf_attr *attr)
 {
@@ -450,6 +452,33 @@ static int map_create(union bpf_attr *attr)
 	atomic_set(&map->refcnt, 1);
 	atomic_set(&map->usercnt, 1);
 
+	if (bpf_map_support_seq_show(map) &&
+	    (attr->btf_key_id || attr->btf_value_id)) {
+		struct btf *btf;
+
+		if (!attr->btf_key_id || !attr->btf_value_id) {
+			err = -EINVAL;
+			goto free_map_nouncharge;
+		}
+
+		btf = btf_get_by_fd(attr->btf_fd);
+		if (IS_ERR(btf)) {
+			err = PTR_ERR(btf);
+			goto free_map_nouncharge;
+		}
+
+		err = map->ops->map_check_btf(map, btf, attr->btf_key_id,
+					      attr->btf_value_id);
+		if (err) {
+			btf_put(btf);
+			goto free_map_nouncharge;
+		}
+
+		map->btf = btf;
+		map->btf_key_id = attr->btf_key_id;
+		map->btf_value_id = attr->btf_value_id;
+	}
+
 	err = security_bpf_map_alloc(map);
 	if (err)
 		goto free_map_nouncharge;
@@ -482,6 +511,7 @@ static int map_create(union bpf_attr *attr)
 free_map_sec:
 	security_bpf_map_free(map);
 free_map_nouncharge:
+	btf_put(map->btf);
 	map->ops->map_free(map);
 	return err;
 }