[GIT PULL] Ceph updates for 4.18-rc1

From: Ilya Dryomov
Date: Thu Jun 14 2018 - 06:58:32 EST


Hi Linus,

The following changes since commit 29dcea88779c856c7dc92040a0c01233263101d4:

Linux 4.17 (2018-06-03 14:15:21 -0700)

are available in the git repository at:

https://github.com/ceph/ceph-client.git tags/ceph-for-4.18-rc1

for you to fetch changes up to 23edca864951250af845a11da86bb3ea63522ed2:

rbd: flush rbd_dev->watch_dwork after watch is unregistered (2018-06-04 20:46:02 +0200)

----------------------------------------------------------------
The main piece is a set of libceph changes that revamps how OSD
requests are aborted, improving CephFS ENOSPC handling and making
"umount -f" actually work (Zheng and myself). The rest is mostly
mount option handling cleanups from Chengguang and assorted fixes
from Zheng, Luis and Dongsheng.

----------------------------------------------------------------
Chengguang Xu (5):
libceph, rbd: add error handling for osd_req_op_cls_init()
ceph: fix alignment of rasize
ceph: strengthen rsize/wsize/readdir_max_bytes validation
ceph: show ino32 if the value is different with default
ceph: update description of some mount options

Dongsheng Yang (1):
rbd: flush rbd_dev->watch_dwork after watch is unregistered

Ilya Dryomov (13):
libceph: get rid of more_kvec in try_write()
libceph: use MSG_TRUNC for discarding received bytes
ceph: show wsize only if non-default
libceph: introduce ceph_osdc_abort_requests()
libceph: no need to call flush_workqueue() before destruction
libceph: move more code into __complete_request()
libceph: defer __complete_request() to a workqueue
libceph: use for_each_request() in ceph_osdc_abort_on_full()
libceph: don't warn if req->r_abort_on_full is set
libceph: avoid a use-after-free during map check
libceph: don't abort reads in ceph_osdc_abort_on_full()
libceph: make abort_on_full a per-osdc setting
libceph: allocate the locator string with GFP_NOFAIL

Luis Henriques (2):
ceph: fix st_nlink stat for directories
ceph: fix use-after-free in ceph_statfs()

Yan, Zheng (10):
ceph: use bit flags to define vxattr attributes
ceph: always get rstat from auth mds
ceph: update i_files/i_subdirs only when Fs cap is issued
ceph: define argument structure for handle_cap_grant
ceph: handle the new nfiles/nsubdirs fields in cap message
ceph: support file lock on directory
ceph: abort osd requests on force umount
ceph: flush pending works before shutdown super
ceph: fix wrong check for the case of updating link count
ceph: prevent i_version from going back

Documentation/filesystems/ceph.txt | 8 +-
drivers/block/rbd.c | 11 +-
fs/ceph/addr.c | 1 -
fs/ceph/caps.c | 160 ++++++++++++++++-----------
fs/ceph/dir.c | 2 +
fs/ceph/file.c | 1 -
fs/ceph/inode.c | 67 +++++++-----
fs/ceph/super.c | 35 ++++--
fs/ceph/xattr.c | 60 ++++++-----
include/linux/ceph/ceph_fs.h | 1 +
include/linux/ceph/osd_client.h | 8 +-
include/linux/ceph/osdmap.h | 8 +-
net/ceph/messenger.c | 31 ++----
net/ceph/osd_client.c | 216 ++++++++++++++++++++++---------------
net/ceph/osdmap.c | 19 ++--
15 files changed, 372 insertions(+), 256 deletions(-)