On 27/10/2022 17:35, Luís Henriques wrote:
On Thu, Oct 27, 2022 at 11:20:23AM +0800, Zorro Lang wrote:
On Thu, Oct 27, 2022 at 11:00:21AM +0800, xiubli@xxxxxxxxxx wrote:
From: Xiubo Li <xiubli@xxxxxxxxxx>
Signed-off-by: Xiubo Li <xiubli@xxxxxxxxxx>
---
common/encrypt | 3 +++
1 file changed, 3 insertions(+)
diff --git a/common/encrypt b/common/encrypt
index 45ce0954..1a77e23b 100644
--- a/common/encrypt
+++ b/common/encrypt
@@ -153,6 +153,9 @@ _scratch_mkfs_encrypted()
# erase the UBI volume; reformated automatically on next mount
$UBIUPDATEVOL_PROG ${SCRATCH_DEV} -t
;;
+ ceph)
+ _scratch_cleanup_files
+ ;;
Any commits about that?
After I sent it out I found I forgot to add it.
Should I send V3 for it ?
Sorry I'm not familar with cephfs, is this patch enough to help ceph to test
encrypted ceph? Due to you tried to do some "checking" job last time.
Can "./check -g encrypt" work on ceph? May you paste this test result to help
to review? And welcome review points from ceph list.
I think Xiubo's patch is likely to be required for enabling encryption
testing in ceph. Simply doing a '_scratch_cleanup_files' is exactly what
network filesystems are doing on _scratch_mkfs(). Thus it makes sense for
ceph to do the same for testing fscrypt support, as we don't really have
an 'mkfs.ceph' tool.
Now, this patch alone is probably not enough to allow to do all the
validation we're looking for. (But note that I did *not* tried it myself,
so I may be wrong.) I think we'll need to go through each of the
'encrypt' tests, run it in ceph and see if they are really testing what
they are supposed to.
But that's just my two cents ;-)
This is my test output locally:
[root@lxbceph1 xfstests-dev]# ./check -g encrypt
FSTYP -- ceph
PLATFORM -- Linux/x86_64 lxbceph1 6.1.0-rc1+ #164 SMP
PREEMPT_DYNAMIC Mon Oct 24 10:18:33 CST 2022
MKFS_OPTIONS -- 10.72.47.117:40267:/testB
MOUNT_OPTIONS -- -o name=admin,nowsync,copyfrom,rasize=4096 -o
context=system_u:object_r:root_t:s0 10.72.47.117:40267:/testB /mnt/kcephfs.B
generic/395 187s ... 224s
generic/396 164s ... 219s
generic/397 177s ... 226s
generic/398 [not run] kernel doesn't support renameat2 syscall
generic/399 [not run] Filesystem ceph not supported in
_scratch_mkfs_sized_encrypted
generic/419 [not run] kernel doesn't support renameat2 syscall
generic/421 170s ... 219s
generic/429 179s ... 237s
generic/435
...
Not finished yet.
From my previous test only 395,396,397,421,429,440,593,595,598 test
case will pass and all the other will be skipped like the 398,399,419 above.
Before this patch or my V1 patches, all the test cases will be skipped.
Thanks!
- Xiubo
Cheers,
--
Luís
Thanks,
Zorro
[1]
$ grep -rsn _scratch_mkfs_encrypted tests/generic/
tests/generic/395:22:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/396:21:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/580:23:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/581:36:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/595:35:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/613:29:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/621:57:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/429:36:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/397:28:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/398:28:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/421:24:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/440:29:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/419:29:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/435:33:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/593:24:_scratch_mkfs_encrypted &>> $seqres.full
tests/generic/576:34:_scratch_mkfs_encrypted_verity &>> $seqres.full
*)
_notrun "No encryption support for $FSTYP"
;;
--
2.31.1
