From: Tal Lossos <tallossos@xxxxxxxxx>
[ Upstream commit 769c18b254ca191b45047e1fcb3b2ce56fada0b6 ]
bpf_fd_inode_storage_lookup_elem() returned NULL when getting a bad FD,
which caused -ENOENT in bpf_map_copy_value. -EBADF error is better than
-ENOENT for a bad FD behaviour.
The patch was partially contributed by CyberArk Software, Inc.
Fixes: 8ea636848aca ("bpf: Implement bpf_local_storage for inodes")
Signed-off-by: Tal Lossos <tallossos@xxxxxxxxx>
Signed-off-by: Daniel Borkmann <daniel@xxxxxxxxxxxxx>
Acked-by: Yonghong Song <yhs@xxxxxx>
Acked-by: KP Singh <kpsingh@xxxxxxxxxx>
Link: https://lore.kernel.org/bpf/20210307120948.61414-1-tallossos@xxxxxxxxx
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
kernel/bpf/bpf_inode_storage.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/bpf/bpf_inode_storage.c b/kernel/bpf/bpf_inode_storage.c
index 6639640523c0..b58b2efb9b43 100644
--- a/kernel/bpf/bpf_inode_storage.c
+++ b/kernel/bpf/bpf_inode_storage.c
@@ -109,7 +109,7 @@ static void *bpf_fd_inode_storage_lookup_elem(struct bpf_map *map, void *key)
fd = *(int *)key;
f = fget_raw(fd);
if (!f)
- return NULL;
+ return ERR_PTR(-EBADF);
sdata = inode_storage_lookup(f->f_inode, map, true);
fput(f);
--
2.30.1
