On 4/17/23 05:10, Jeff Layton wrote:
On Sun, 2023-04-16 at 22:01 +0200, Ilya Dryomov wrote:
On Wed, Apr 12, 2023 at 1:13 PM <xiubli@xxxxxxxxxx> wrote:
[...]
#endif
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index b9dd2fa36d8b..4b0a070d5c6d 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -591,6 +591,12 @@ static int ceph_parse_mount_param(struct fs_context *fc,
break;
case Opt_test_dummy_encryption:
#ifdef CONFIG_FS_ENCRYPTION
+ /* HACK: allow for cleartext "encryption" in files for testing */
+ if (param->string && !strcmp(param->string, "clear")) {
+ fsopt->flags |= CEPH_MOUNT_OPT_DUMMY_ENC_CLEAR;
I really wonder whether this is still needed? Having a mount option
that causes everything to be automatically encrypted with a dummy key
for testing purposes makes total sense. Making it possible to disable
encryption through the same -- not so much.
Does any other fscrypt-enabled filesystem in mainline do this?
I doubt it. It was totally a hack that I had in place to help debugging
when I was developing this. My intention was always to remove this
before merging it. I think doing that now would be a good idea.
Yeah, no this for other FSs and I will remove this option.
Thanks
- Xiubo
