[PATCH bpf 04/11] bpf: Add need_defer parameter to .map_fd_put_ptr()

[Hou Tao <houtao@xxxxxxxxxxxxxxx>]

From: Hou Tao <houtao1@xxxxxxxxxx>

need_defer tells the implementation to defer the reference release of
the passed element and ensure that the element is still alive before
the bpf program, which may manipulate it, exits.

The following three cases will invoke map_fd_put_ptr() and different
need_defer values will be passed to these callers:

1) release the reference of the old element in the map during map update
   or map deletion. The release must be deferred, otherwise the bpf
   program may incur use-after-free problem, so need_defer needs to be
   true.
2) release the reference of the to-be-added element in the error path of
   map update. The to-be-added element is not visible to any bpf
   program, so it is OK to pass false for need_defer parameter.
3) release the references of all elements in the map during map release.
   Any bpf program which has access to the map must have been exited and
   released, so need_defer=false will be OK.

The need_defer parameter will be used by the following patches to fix the
potential use-after-free problem for map-in-map.

Signed-off-by: Hou Tao <houtao1@xxxxxxxxxx>
---
 include/linux/bpf.h     |  6 +++++-
 kernel/bpf/arraymap.c   | 12 +++++++-----
 kernel/bpf/hashtab.c    |  6 +++---
 kernel/bpf/map_in_map.c |  2 +-
 kernel/bpf/map_in_map.h |  2 +-
 5 files changed, 17 insertions(+), 11 deletions(-)

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index b4825d3cdb292..88be6a4b0a315 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -106,7 +106,11 @@ struct bpf_map_ops {
 	/* funcs called by prog_array and perf_event_array map */
 	void *(*map_fd_get_ptr)(struct bpf_map *map, struct file *map_file,
 				int fd);
-	void (*map_fd_put_ptr)(void *ptr);
+	/* If need_defer is true, the implementation should guarantee that
+	 * the to-be-put element is still alive before the bpf program, which
+	 * may manipulate it, exists.
+	 */
+	void (*map_fd_put_ptr)(void *ptr, bool need_defer);
 	int (*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,
diff --git a/kernel/bpf/arraymap.c b/kernel/bpf/arraymap.c
index 2fcb7031fca87..c1124b71da158 100644
--- a/kernel/bpf/arraymap.c
+++ b/kernel/bpf/arraymap.c
@@ -867,7 +867,7 @@ int bpf_fd_array_map_update_elem(struct bpf_map *map, struct file *map_file,
 	}
 
 	if (old_ptr)
-		map->ops->map_fd_put_ptr(old_ptr);
+		map->ops->map_fd_put_ptr(old_ptr, true);
 	return 0;
 }
 
@@ -890,7 +890,7 @@ static long fd_array_map_delete_elem(struct bpf_map *map, void *key)
 	}
 
 	if (old_ptr) {
-		map->ops->map_fd_put_ptr(old_ptr);
+		map->ops->map_fd_put_ptr(old_ptr, true);
 		return 0;
 	} else {
 		return -ENOENT;
@@ -913,8 +913,9 @@ static void *prog_fd_array_get_ptr(struct bpf_map *map,
 	return prog;
 }
 
-static void prog_fd_array_put_ptr(void *ptr)
+static void prog_fd_array_put_ptr(void *ptr, bool deferred)
 {
+	/* bpf_prog is freed after one RCU or tasks trace grace period */
 	bpf_prog_put(ptr);
 }
 
@@ -1239,8 +1240,9 @@ static void *perf_event_fd_array_get_ptr(struct bpf_map *map,
 	return ee;
 }
 
-static void perf_event_fd_array_put_ptr(void *ptr)
+static void perf_event_fd_array_put_ptr(void *ptr, bool deferred)
 {
+	/* bpf_perf_event is freed after one RCU grace period */
 	bpf_event_entry_free_rcu(ptr);
 }
 
@@ -1294,7 +1296,7 @@ static void *cgroup_fd_array_get_ptr(struct bpf_map *map,
 	return cgroup_get_from_fd(fd);
 }
 
-static void cgroup_fd_array_put_ptr(void *ptr)
+static void cgroup_fd_array_put_ptr(void *ptr, bool deferred)
 {
 	/* cgroup_put free cgrp after a rcu grace period */
 	cgroup_put(ptr);
diff --git a/kernel/bpf/hashtab.c b/kernel/bpf/hashtab.c
index 7d1457360c99b..81b9f237c942b 100644
--- a/kernel/bpf/hashtab.c
+++ b/kernel/bpf/hashtab.c
@@ -897,7 +897,7 @@ static void htab_put_fd_value(struct bpf_htab *htab, struct htab_elem *l)
 
 	if (map->ops->map_fd_put_ptr) {
 		ptr = fd_htab_map_get_ptr(map, l);
-		map->ops->map_fd_put_ptr(ptr);
+		map->ops->map_fd_put_ptr(ptr, true);
 	}
 }
 
@@ -2484,7 +2484,7 @@ static void fd_htab_map_free(struct bpf_map *map)
 		hlist_nulls_for_each_entry_safe(l, n, head, hash_node) {
 			void *ptr = fd_htab_map_get_ptr(map, l);
 
-			map->ops->map_fd_put_ptr(ptr);
+			map->ops->map_fd_put_ptr(ptr, false);
 		}
 	}
 
@@ -2527,7 +2527,7 @@ int bpf_fd_htab_map_update_elem(struct bpf_map *map, struct file *map_file,
 	ret = htab_map_update_elem(map, key, &ptr, map_flags);
 	rcu_read_unlock();
 	if (ret)
-		map->ops->map_fd_put_ptr(ptr);
+		map->ops->map_fd_put_ptr(ptr, false);
 
 	return ret;
 }
diff --git a/kernel/bpf/map_in_map.c b/kernel/bpf/map_in_map.c
index cd5eafaba97e2..8323ce201159d 100644
--- a/kernel/bpf/map_in_map.c
+++ b/kernel/bpf/map_in_map.c
@@ -127,7 +127,7 @@ void *bpf_map_fd_get_ptr(struct bpf_map *map,
 	return inner_map;
 }
 
-void bpf_map_fd_put_ptr(void *ptr)
+void bpf_map_fd_put_ptr(void *ptr, bool deferred)
 {
 	/* ptr->ops->map_free() has to go through one
 	 * rcu grace period by itself.
diff --git a/kernel/bpf/map_in_map.h b/kernel/bpf/map_in_map.h
index bcb7534afb3c0..63872bffd9b3c 100644
--- a/kernel/bpf/map_in_map.h
+++ b/kernel/bpf/map_in_map.h
@@ -13,7 +13,7 @@ struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd);
 void bpf_map_meta_free(struct bpf_map *map_meta);
 void *bpf_map_fd_get_ptr(struct bpf_map *map, struct file *map_file,
 			 int ufd);
-void bpf_map_fd_put_ptr(void *ptr);
+void bpf_map_fd_put_ptr(void *ptr, bool need_defer);
 u32 bpf_map_fd_sys_lookup_elem(void *ptr);
 
 #endif
-- 
2.29.2