If CONFIG_EXT4_FS_POSIX_ACL=n then the fallback version of ext4_init_acl()
will mask off the umask bits from the new inode's i_mode. This should not
be done if the inode is a symlink. If CONFIG_EXT4_FS_POSIX_ACL=y, then we
go through posix_acl_create() instead which does the right thing with
symlinks.
However, this is actually unnecessary now as vfs_prepare_mode() has already
done this where appropriate, so fix this by making the fallback version of
ext4_init_acl() do nothing.
Fixes: 484fd6c1de13 ("ext4: apply umask if ACL support is disabled")
Suggested-by: Miklos Szeredi <miklos@xxxxxxxxxx>
Signed-off-by: David Howells <dhowells@xxxxxxxxxx>
cc: Max Kellermann <max.kellermann@xxxxxxxxx>
cc: Jan Kara <jack@xxxxxxxx>
cc: Christian Brauner <brauner@xxxxxxxxxx>
cc: linux-ext4@xxxxxxxxxxxxxxx
cc: linux-fsdevel@xxxxxxxxxxxxxxx
---
fs/ext4/acl.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/fs/ext4/acl.h b/fs/ext4/acl.h
index ef4c19e5f570..566625286442 100644
--- a/fs/ext4/acl.h
+++ b/fs/ext4/acl.h
@@ -71,7 +71,8 @@ ext4_init_acl(handle_t *handle, struct inode *inode, struct inode *dir)
/* usually, the umask is applied by posix_acl_create(), but if
ext4 ACL support is disabled at compile time, we need to do
it here, because posix_acl_create() will never be called */
- inode->i_mode &= ~current_umask();
+ if (!S_ISLNK(inode->i_mode))
+ inode->i_mode &= ~current_umask();
return 0;
}
