linux.git/net/ceph, branch v3.18-rc2

Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client

2014-10-15T04:46:01+00:00

Pull Ceph updates from Sage Weil:
 "There is the long-awaited discard support for RBD (Guangliang Zhao,
  Josh Durgin), a pile of RBD bug fixes that didn't belong in late -rc's
  (Ilya Dryomov, Li RongQing), a pile of fs/ceph bug fixes and
  performance and debugging improvements (Yan, Zheng, John Spray), and a
  smattering of cleanups (Chao Yu, Fabian Frederick, Joe Perches)"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client: (40 commits)
  ceph: fix divide-by-zero in __validate_layout()
  rbd: rbd workqueues need a resque worker
  libceph: ceph-msgr workqueue needs a resque worker
  ceph: fix bool assignments
  libceph: separate multiple ops with commas in debugfs output
  libceph: sync osd op definitions in rados.h
  libceph: remove redundant declaration
  ceph: additional debugfs output
  ceph: export ceph_session_state_name function
  ceph: include the initial ACL in create/mkdir/mknod MDS requests
  ceph: use pagelist to present MDS request data
  libceph: reference counting pagelist
  ceph: fix llistxattr on symlink
  ceph: send client metadata to MDS
  ceph: remove redundant code for max file size verification
  ceph: remove redundant io_iter_advance()
  ceph: move ceph_find_inode() outside the s_mutex
  ceph: request xattrs if xattr_version is zero
  rbd: set the remaining discard properties to enable support
  rbd: use helpers to handle discard for layered images correctly
  ...

libceph: ceph-msgr workqueue needs a resque worker

2014-10-14T19:57:04+00:00

Commit f363e45fd118 ("net/ceph: make ceph_msgr_wq non-reentrant")
effectively removed WQ_MEM_RECLAIM flag from ceph_msgr_wq.  This is
wrong - libceph is very much a memory reclaim path, so restore it.

Cc: stable@vger.kernel.org # needs backporting for < 3.12
Signed-off-by: Ilya Dryomov 
Tested-by: Micha Krause 
Reviewed-by: Sage Weil

libceph: separate multiple ops with commas in debugfs output

2014-10-14T19:57:03+00:00

For requests with multiple ops, separate ops with commas instead of \t,
which is a field separator here.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Sage Weil

libceph: sync osd op definitions in rados.h

2014-10-14T19:57:02+00:00

Bring in missing osd ops and strings, use macros to eliminate multiple
points of maintenance.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Sage Weil

libceph: reference counting pagelist

2014-10-14T19:56:48+00:00

this allow pagelist to present data that may be sent multiple times.

Signed-off-by: Yan, Zheng 
Reviewed-by: Sage Weil

libceph: don't try checking queue_work() return value

2014-10-14T17:03:25+00:00

queue_work() doesn't "fail to queue", it returns false if work was
already on a queue, which can't happen here since we allocate
event_work right before we queue it.  So don't bother at all.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Alex Elder

libceph: Convert pr_warning to pr_warn

2014-10-14T17:03:23+00:00

Use the more common pr_warn.

Other miscellanea:

o Coalesce formats
o Realign arguments

Signed-off-by: Joe Perches 
Signed-off-by: Ilya Dryomov

libceph: fix a use after free issue in osdmap_set_max_osd

2014-10-14T17:03:21+00:00

If the state variable is krealloced successfully, map->osd_state will be
freed, once following two reallocation failed, and exit the function
without resetting map->osd_state, map->osd_state become a wild pointer.

fix it by resetting them after krealloc successfully.

Signed-off-by: Li RongQing 
Signed-off-by: Ilya Dryomov

libceph: select CRYPTO_CBC in addition to CRYPTO_AES

2014-10-14T17:03:20+00:00

We want "cbc(aes)" algorithm, so select CRYPTO_CBC too, not just
CRYPTO_AES.  Otherwise on !CRYPTO_CBC kernels we fail rbd map/mount
with

    libceph: error -2 building auth method x request

Signed-off-by: Ilya Dryomov

libceph: resend lingering requests with a new tid

2014-10-14T17:03:19+00:00

Both not yet registered (r_linger && list_empty(&r_linger_item)) and
registered linger requests should use the new tid on resend to avoid
the dup op detection logic on the OSDs, yet we were doing this only for
"registered" case.  Factor out and simplify the "registered" logic and
use the new helper for "not registered" case as well.

Fixes: http://tracker.ceph.com/issues/8806

Signed-off-by: Ilya Dryomov 
Reviewed-by: Alex Elder