Hi!
The following patches will make the cephfs kernel client behave the same
way as the fuse client when doing renames across different quota realms.
Changes since v1:
- Dropped 'old == new' check in ceph_quota_check_rename() and added back
optimization in ceph_rename(), to only check realms if old_dir and
new_dir are different.
Luis Henriques (2):
ceph: normalize 'delta' parameter usage in check_quota_exceeded
ceph: allow rename operation under different quota realms
fs/ceph/dir.c | 9 +++----
fs/ceph/quota.c | 62 +++++++++++++++++++++++++++++++++++++++++++++----
fs/ceph/super.h | 3 ++-
3 files changed, 65 insertions(+), 9 deletions(-)
On Tue, 2020-04-07 at 11:30 +0100, Luis Henriques wrote:
> Hi!
>
> The following patches will make the cephfs kernel client behave the same
> way as the fuse client when doing renames across different quota realms.
>
> Changes since v1:
>
> - Dropped 'old == new' check in ceph_quota_check_rename() and added back
> optimization in ceph_rename(), to only check realms if old_dir and
> new_dir are different.
>
> Luis Henriques (2):
> ceph: normalize 'delta' parameter usage in check_quota_exceeded
> ceph: allow rename operation under different quota realms
>
> fs/ceph/dir.c | 9 +++----
> fs/ceph/quota.c | 62 +++++++++++++++++++++++++++++++++++++++++++++----
> fs/ceph/super.h | 3 ++-
> 3 files changed, 65 insertions(+), 9 deletions(-)
>
Looks good. Merged into ceph-client/testing.
Thanks!
--
Jeff Layton <[email protected]>