Add BPF_MAP_GET_FD_BY_ID command to allow user to get a fd
from a bpf_map's ID.
bpf_map_inc_not_zero() is added and is called with map_idr_lock
held.
__bpf_map_put() is also added which has the 'bool do_idr_lock'
param to decide if the map_idr_lock should be acquired when
freeing the map->id.
In the error path of bpf_map_inc_not_zero(), it may have to
call __bpf_map_put(map, false) which does not need
to take the map_idr_lock when freeing the map->id.
It is currently limited to CAP_SYS_ADMIN which we can
consider to lift it in followup patches.
Signed-off-by: Martin KaFai Lau <ka...@fb.com>
Acked-by: Alexei Starovoitov <a...@fb.com>
Acked-by: Daniel Borkmann <dan...@iogearbox.net>
---
include/uapi/linux/bpf.h | 2 +
kernel/bpf/syscall.c | 95 +++++++++++++++++++++++++++++++++++++++++++-----
2 files changed, 87 insertions(+), 10 deletions(-)
diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
index 6d4e1cc5bd18..cf704e8b6e65 100644
--- a/include/uapi/linux/bpf.h
+++ b/include/uapi/linux/bpf.h
@@ -85,6 +85,7 @@ enum bpf_cmd {
BPF_PROG_GET_NEXT_ID,
BPF_MAP_GET_NEXT_ID,
BPF_PROG_GET_FD_BY_ID,
+ BPF_MAP_GET_FD_BY_ID,
};
enum bpf_map_type {
@@ -217,6 +218,7 @@ union bpf_attr {
union {
__u32 start_id;
__u32 prog_id;
+ __u32 map_id;
};
__u32 next_id;
};
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 4f9ee57e7140..de8fe04a6539 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -135,11 +135,19 @@ static int bpf_map_alloc_id(struct bpf_map *map)
return id > 0 ? 0 : id;
}
-static void bpf_map_free_id(struct bpf_map *map)
+static void bpf_map_free_id(struct bpf_map *map, bool do_idr_lock)
{
- spin_lock_bh(&map_idr_lock);
+ if (do_idr_lock)
+ spin_lock_bh(&map_idr_lock);
+ else
+ __acquire(&map_idr_lock);
+
idr_remove(&map_idr, map->id);
- spin_unlock_bh(&map_idr_lock);
+
+ if (do_idr_lock)
+ spin_unlock_bh(&map_idr_lock);
+ else
+ __release(&map_idr_lock);
}
/* called from workqueue */
@@ -163,16 +171,21 @@ static void bpf_map_put_uref(struct bpf_map *map)
/* decrement map refcnt and schedule it for freeing via workqueue
* (unrelying map implementation ops->map_free() might sleep)
*/
-void bpf_map_put(struct bpf_map *map)
+static void __bpf_map_put(struct bpf_map *map, bool do_idr_lock)
{
if (atomic_dec_and_test(&map->refcnt)) {
/* bpf_map_free_id() must be called first */
- bpf_map_free_id(map);
+ bpf_map_free_id(map, do_idr_lock);
INIT_WORK(&map->work, bpf_map_free_deferred);
schedule_work(&map->work);
}
}
+void bpf_map_put(struct bpf_map *map)
+{
+ __bpf_map_put(map, true);
+}
+
void bpf_map_put_with_uref(struct bpf_map *map)
{
bpf_map_put_uref(map);
@@ -271,15 +284,20 @@ static int map_create(union bpf_attr *attr)
goto free_map;
err = bpf_map_new_fd(map);
- if (err < 0)
- /* failed to allocate fd */
- goto free_id;
+ if (err < 0) {
+ /* failed to allocate fd.
+ * bpf_map_put() is needed because the above
+ * bpf_map_alloc_id() has published the map
+ * to the userspace and the userspace may
+ * have refcnt-ed it through BPF_MAP_GET_FD_BY_ID.
+ */
+ bpf_map_put(map);
+ return err;
+ }
trace_bpf_map_create(map, err);
return err;
-free_id:
- bpf_map_free_id(map);
free_map:
bpf_map_uncharge_memlock(map);
free_map_nouncharge:
@@ -331,6 +349,28 @@ struct bpf_map *bpf_map_get_with_uref(u32 ufd)
return map;
}
+/* map_idr_lock should have been held */
+static struct bpf_map *bpf_map_inc_not_zero(struct bpf_map *map,
+ bool uref)
+{
+ int refold;
+
+ refold = __atomic_add_unless(&map->refcnt, 1, 0);
+
+ if (refold >= BPF_MAX_REFCNT) {
+ __bpf_map_put(map, false);
+ return ERR_PTR(-EBUSY);
+ }
+
+ if (!refold)
+ return ERR_PTR(-ENOENT);
+
+ if (uref)
+ atomic_inc(&map->usercnt);
+
+ return map;
+}
+
int __weak bpf_stackmap_copy(struct bpf_map *map, void *key, void *value)
{
return -ENOTSUPP;
@@ -1165,6 +1205,38 @@ static int bpf_prog_get_fd_by_id(const union bpf_attr
*attr)
return fd;
}
+#define BPF_MAP_GET_FD_BY_ID_LAST_FIELD map_id
+
+static int bpf_map_get_fd_by_id(const union bpf_attr *attr)
+{
+ struct bpf_map *map;
+ u32 id = attr->map_id;
+ int fd;
+
+ if (CHECK_ATTR(BPF_MAP_GET_FD_BY_ID))
+ return -EINVAL;
+
+ if (!capable(CAP_SYS_ADMIN))
+ return -EPERM;
+
+ spin_lock_bh(&map_idr_lock);
+ map = idr_find(&map_idr, id);
+ if (map)
+ map = bpf_map_inc_not_zero(map, true);
+ else
+ map = ERR_PTR(-ENOENT);
+ spin_unlock_bh(&map_idr_lock);
+
+ if (IS_ERR(map))
+ return PTR_ERR(map);
+
+ fd = bpf_map_new_fd(map);
+ if (fd < 0)
+ bpf_map_put(map);
+
+ return fd;
+}
+
SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int,
size)
{
union bpf_attr attr = {};
@@ -1253,6 +1325,9 @@ SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *,
uattr, unsigned int, siz
case BPF_PROG_GET_FD_BY_ID:
err = bpf_prog_get_fd_by_id(&attr);
break;
+ case BPF_MAP_GET_FD_BY_ID:
+ err = bpf_map_get_fd_by_id(&attr);
+ break;
default:
err = -EINVAL;
break;
--
2.9.3
