Hi Linus,
Please pull the following Ceph fixes from
git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client.git for-linus
The main item here is a workaround for a regression caused by the dcache
RCU stuff. d_kill now clears d_parent prior to the release dentry op,
which causes a NULL pointer dereference in the ceph code. Nick said this
should be easy enough to fix[1], but he's offline and it seems a bit late
to futz with the core dcache code. This patch basically just makes a copy
of d_parent in the private d_fsdata struct. We can revert it once the
d_kill/d_parent thing gets fixed up, presumably during the next merge
window.
There are also some error handling fixes and a fix for an possible
infinite loop in the snapshot notification handler.
Thanks-
sage
[1] http://marc.info/?l=linux-fsdevel&m=129610552517424&w=2
Sage Weil (3):
libceph: fix socket read error handling
libceph: fix socket write error handling
ceph: queue cap_snaps once per realm
Yehuda Sadeh (1):
ceph: keep reference to parent inode on ceph_dentry
fs/ceph/dir.c | 5 +++-
fs/ceph/snap.c | 14 ++++++++---
fs/ceph/super.h | 1 +
net/ceph/messenger.c | 62 +++++++++++++++++++++++++-------------------------
4 files changed, 46 insertions(+), 36 deletions(-)