LU-8056 llite: use inode_lock to access i_mutex

Linux kernel 4.5 introduced wrappers for i_mutex as the type
of i_mutex will be changed in the future.
This patch prepares us for the upcoming type change. It also
includes a patch from linux kernel removing unused declaration
to make the room for the wrappers mentioned above.

Note that this patch only touches the llite part. osd-ldiskfs still
calls mutex_{lock,unlock} on i_mutex.

Linux-commit:57b8f112cfe6622ddddb8c2641206bb5fa8a112d
Linux-commit:5955102c9984fa081b2d570cfac75c97eecf8f3b

Signed-off-by: Li Dongyang <dongyang.li@anu.edu.au>
Change-Id: Ia4c30443cddfaa5232e1dc1519c33c97dba2a271
Reviewed-on: http://review.whamcloud.com/20223
Tested-by: Jenkins
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8056 socklnd: NETIF_F_ALL_CSUM renamed to NETIF_F_CSUM_MASK

In kernel 4.5 NETIF_F_CSUM_MASK got renamed to NETIF_F_CSUM_MASK.
This patch handles the name change.

Linux-commit:a188222b6ed29404ac2d4232d35d1fe0e77af370

Signed-off-by: Li Dongyang <dongyang.li@anu.edu.au>
Change-Id: Id57505eeca613303c584d3cf74284920a837bb43
Reviewed-on: http://review.whamcloud.com/20222
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6401 headers: Move object_update functions to new header

The header lustre_idl.h is meant to be a uapi header but it
currently contains code that is only used by the lustre
kernel internals. This patch moves the code related to
object update handling out of lustre_idl.h into a separate
header that is only needed for the kernel server side code.

Signed-off-by: Ben Evans <bevans@cray.com>
Change-Id: Ie025f8424dad8cfc625258ad189c7b7576b3a9b2
Reviewed-on: http://review.whamcloud.com/20120
Reviewed-by: Frank Zago <fzago@cray.com>
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8113 build: Remove obsolete, unused meta packages

The meta packages, deps-rhel and deps-sles, were introduced
nine years ago and never used.  This removes those dead lines
in the configuration files.

We would not want to use this method in the future any way.
Better approaches would either be to use multiple spec files
or to use distro-specific macros and conditionals for expressing
dependencies.

Test-Parameters: trivial
Change-Id: I9ad7195e0504f638bca84984af499d959c37adb3
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/20045
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6696 llog: improve error handling

Handle error return value of llog_cat_process() in
osp_sync_thread().

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: I9b8e00dcc90e8d7d55648ce1b3a18af0a74c542b
Reviewed-on: http://review.whamcloud.com/19856
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>

LU-8021 tests: In interop, ensure to save/restore correct debug flags

debugsave() and debugrestore() in test-framework.sh save the
same debug flags of the client on servers as well.

In case of interop, esp between lustre 2.1.x server and
2.5.x clients, with PTLDEBUG=-1 the debugrestore() would fail as
there are some flags which are newly added to 2.5.x,
    e.g. "lfsck, hsm" are not defined on 2.1.x servers.

- The fix, makes sure to save the debug flags for servers and clients
separately and thus restore correctly.

- It restores the debug flags irrespective of, if the server node
is active or passive, failover node.

- It takes care,
if any of the node(server OR client) is down or unreachable,
if there are separate MDS MGS servers,
if any of the node is rebooted (FAILURE_MODE=hard)
but lustre modules are not loaded.

Also, there is defect introduced by LU-4480 in reset_fail_loc,
this patch fixes that as well.

Seagate-bug-id: MRP-1959
Change-Id: I6c6142cc113f8131bb89e6dc313215440f2242b9
Signed-off-by: Parinay Kondekar <parinay.kondekar@seagate.com>
Reviewed-by: Elena Gryaznova <elena.gryaznova@seagate.com>
Reviewed-by: Ujjwal Lanjewar <ujjwal.lanjewar@seagate.com>
Reviewed-on: http://review.whamcloud.com/19604
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-7961 build: Fix ldiskfs source autodetect for CentOS 6

CentOS 6 installed the kernel source in to the following path
which differs slightly from the RHEL version.  Expand the search
pattern to encompass both the CentOS and RHEL paths.

/usr/src/debug/*/linux-2.6.32-x.y.z.el6.centos.plus.x86_64

Change-Id: I0a102148f34d57b94dddfac7819303314c61a376
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-on: http://review.whamcloud.com/19248
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7899 osd: batch EA updates

during file creation we set number of EAs: LMA, VBR, LinkEA, LOVEA, ACLs.
calling into SA to refill spill again and again is expensive. thus it
makes sense to postpone this to osd_trans_stop() where all changed EAs
has been already collected in a temporary buffer.

Change-Id: I8f02a287b96615c3aa550d63ffd9dd3da51b39ee
Signed-off-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-on: http://review.whamcloud.com/19143
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7900 obdclass: lu_object_assign_fid() do not lookup

as the object is supposed to be new and unique by definition.
this method is used only with OST objects being created on MDT.

Change-Id: I9e641ab208654ea64eb42394a3618c9f30e7d751
Signed-off-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-on: http://review.whamcloud.com/19081
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7660 dne: support fs default stripe

On DNE system slave MDTs should honor fs default stripe, to achieve
this, slave MDT should lock fs root, and cache XATTR_NAME_LOV, below
functionalities are added for this:
* take and cache remote root XATTR lock on slave MDTs.
* add dt_invalidate operation to invalidate OSP attributes cache,
  normally it's called in mdt_object_unlock(), but if we cache this
  lock, it's called upon cross-MDT lock cancel.
* enable OSP attribute cache by default.
* migrate should invalidate remote target explictely if it's newly
  created, because it doesn't take its lock beforehand.
* OSP should cleanup ldlm namespace in device cleanup because there
  may be cached locks.
* add sanity.sh 405 test for this.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: I90e5f35f3b6a294a2a559c28df267f92ce188b9c
Reviewed-on: http://review.whamcloud.com/19041
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7782 scrub: handle slave obj of striped directory

When lookup item under striped directory, we need to locate the
master MDT-object of the striped directory firstly, then the
client will send lookup (getattr_by_name) RPC to the MDT with
some slave MDT-object's FID and the item's name. If the system
is restored from MDT file level backup, then before the OI scrub
completely built the OI files, the OI mappings of the master
MDT-object and slave MDT-object may be invalid. Usually, it is
not a problem for the master MDT-object. Because when locate the
master MDT-object, we will do name based lookup (for the striped
directory itself) firstly, during such process we can setup the
correct OI mapping for the master MDT-object. But it will be
trouble for the slave MDT-object. Because the client will not
trigger name based lookup on the MDT to locate the slave
MDT-object before locating item under the striped directory,
then when osd_fid_lookup(), it will find that the OI mapping
for the slave MDT-object is invalid and does not know what the
right OI mapping is, then the MDT has to return -EINPROGRESS to
the client to notify that the OI scrub is rebuiding the OI file,
related OI mapping is unknown yet, please try again later. And
then client will re-try the RPC again and again until related
OI mapping has been updated. That is quite inefficient.

To resolve above trouble, we will handle it as the following
two cases:

1) The slave MDT-object and the MDT-object are on different
   MDTs. It is relative easy. Be as one of remote MDT-objects,
   the slave MDT-object is linked under /REMOTE_PARENT_DIR
   with the name of its FID string. We can locate the slave
   MDT-object via lookup the /REMOTE_PARENT_DIR directly.

2) The slave MDT-object and the MDT-object reside on the same
   MDT. Under such case, during lookup the master MDT-object,
   we will lookup the slave MDT-object via readdir the master
   MDT-object, because the slave MDT-objects information are
   stored as sub-directories with the name "${FID}:${index}".
   Then when find the local slave MDT-object, its OI mapping
   will be recorded. Then subsequent osd_fid_lookup() will
   know the correct OI mapping for the slave MDT-object.

The patch also enhance sanity-scrub to avoid DNE in sanity-scrub
on one MDT.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I0bf12ac981017245e4c2da08176422a993086c18
Reviewed-on: http://review.whamcloud.com/18801
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7845 lnet: check if address is visible

In LNetDist(), check if ni's address is currently visible.
If not visible, assign order above 0xffff0000, to make this
ni not a priority.

Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I56735aa25ba632020858507939e12c4565c33d49
Reviewed-on: http://review.whamcloud.com/18782
Tested-by: Jenkins
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7674 lmv: remove unused placement parameter

Remove the unused lmv.*.placement parameter along with supporting
functions and struct members.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: I20274002ad429550d15c4c204899f897f1c69d0f
Reviewed-on: http://review.whamcloud.com/18019
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Ben Evans <bevans@cray.com>
Reviewed-by: Frank Zago <fzago@cray.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7619 obdclass: protect REFASSERT() with lu_ref::lf_guard

Some REFASSERT()s have been added by previous changes, but not
under required lu_ref::lf_guard spin-lock protection.
This can cause unexpected "self" dead-lock situations (vs
LBUG/ASSERT!!) upon lu_ref failure detection.

Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Change-Id: If23d43f6fca196a33602b6bb8f051c0a4e26a300
Reviewed-on: http://review.whamcloud.com/17756
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>

LU-1095 lod: quiet "wanted N, found M" console message

Quiet the relatively useless "wanted 1, found 0" console message from
lod_alloc_qos(), since this is not a fatal error - the MDS will fall
back to lod_alloc_rr() in this case.

Clean up lod_alloc_qos() to use a sensible variable name "lod" for the
LOD OBD device instead of "m", which doesn't really mean anything.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I882636a6b89fa2bcf76383ede8486be9c4300c1e
Reviewed-on: http://review.whamcloud.com/17131
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4739 obd: Fixing code inconsistencies and coding style

config_llog_instance is referenced differently in these functions
than anywhere else in the code.  This patch corrects that as well
as fixing some coding style errors.

Signed-off-by: Bobbie Lind <bobbie.j.lind@intel.com>
Change-Id: I9de1a37e397ffcf660422d3adfeb2f59881ea210
Reviewed-on: http://review.whamcloud.com/10420
Tested-by: Jenkins
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-6429 tests: fix typo in previous patch

Fix typo reformat_mdt to format_mdt.

Signed-off-by: Yang Sheng <yang.sheng@intel.com>
Change-Id: I0a78c44e16da496b071b649c10d072364f73d90d
Reviewed-on: http://review.whamcloud.com/21026
Tested-by: Jenkins
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8284 osd-ldiskfs: i_size updates from BRW should be atomic

In osd_write_commit(), it is possible that somewhere between
i_size_read() check and i_size_write() another write updates
i_size to a larger value. In the end, we will have a lower
i_size value than we should.

Change-Id: I720a6cc1214e3e6b32df4f9e2c1017a47e7962f0
Signed-off-by: Andrew Perepechko <andrew.perepechko@seagate.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@seagate.com>
Reviewed-by: Alexander Nikolaevich Boyko <alexander.boyko@seagate.com>
Seagate-bug-id: MRP-3528
Reviewed-on: http://review.whamcloud.com/20816
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alexander Boyko <alexander.boyko@seagate.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8199 ptlrpc: better lut_reply_bitmap handling

Use OBD_ALLOC_LARGE instead of OBD_ALLOC to allocate chunks because of
size.  Also ensure they are free'd correctly.
Better return checking of tgt_bitmap_chunk_alloc by callers.
Check lut_reply_bitmap[chunk] is non-null in tgt_clear_reply_slot (in
case tgt_bitmap_chunk_alloc did fail).
Don't leak lut_replay_bitmap chunks if tgt_init fails.

Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: Ib222300a8c6d0032834d774cde9a9cb3830fb289
Reviewed-on: http://review.whamcloud.com/20450
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Grégoire Pichon <gregoire.pichon@bull.net>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7234 mgs: For handling dash in FSNAME

For FSNAME 'lustre-1, function mgs_llog_open() in mgs_handler.c has
been modified to call strrchr instead of strchr so that 'logname'
variable is updated with entire FSNAME including '-'.

Seagate-bug-id: MRP-2790
Signed-off-by: Sundarraj Illindra <sundarraj.illindra@seagate.com>
Change-Id: I07b2b874f4cf881d0aa2c07586a113fcce8bebfe
Reviewed-on: http://review.whamcloud.com/16901
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8353 mdt: mdt unlink should lock before lookup

MDT unlink should lock parent before lookup child, otherwise the
child may be stale in use later.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: I70560beba7f689bb183ba83c698babc4ce024ece
Reviewed-on: http://review.whamcloud.com/21088
Tested-by: Jenkins
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8343 build: fix kmod_version on SLES 12

Starting SLES 12, the kernel-module-subpackage script added a 'k' in
Provides string, here is the sample

%package -n %{-n*}-kmp-%1
%define _this_kmp_version %{-v*}_k%(echo %2 | sed -r 'y/-/_/; \
%s/^(2\.6\.[0-9]+)_/\\1.0_/; # use 2.6.x.0 for mainline kernels')
...
Provides: %{-n*}-kmp = %_this_kmp_version

Change-Id: I5c4a8af03d70862aaca46e41b737a62491cc3cf7
Signed-off-by: Minh Diep <minh.diep@intel.com>
Reviewed-on: http://review.whamcloud.com/21021
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-by: Christopher J. Morrone <morrone2@llnl.gov>

LU-8319 fid: fix race in fid allocation

There is race condition when allocating fid/seq in parallel
as following:

The thread1 will release the lcs_mutex via seq_fid_alloc_prep(),
then another fid allocation thread2 can obtain the lcs_mutex and
allocate FID in the new sequence that has just been allocated by
the thread1 via seq_client_alloc_seq(); and then after thread2
released the lcs_mutex, the thread1 will re-allocate the current
FID in the new sequence without checking whether some others have
already taken such FID in the new sequence during it re-obtaining
the lcs_mutex.

Such race will cause two objects to use the same FID, and trigger
OI conflict and LMA verification failures.

This patch makes the fid allocation and lu_client_seq modification
to be protected by the lcs_mutex.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I79c250f57b73d0abd7d039eea02424ae438f1b56
Reviewed-on: http://review.whamcloud.com/20939
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-8377 build: clean up .rpmmacros

Landing LU-5614 causing ~/.rpmmacros to fill up on every build

Test-Parameters: trivial

Change-Id: Iecf6cc01e90ae951b878d9f0b99b0a3f4ae9ad31
Signed-off-by: Minh Diep <minh.diep@intel.com>
Reviewed-on: http://review.whamcloud.com/21199
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4039 tests: EXCEPT replay-single test 90

replay-single test 90 is failing multiple times a day on
the master branch. Add this test to the ALWAYS_EXCEPT list.

Test-Parameters: trivial testlist=replay-single

Signed-off-by: James Nunez <james.a.nunez@intel.com>
Change-Id: I8803b545268fe717d9f5d285bf59bfa360564d57
Reviewed-on: http://review.whamcloud.com/21224
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

Revert "LU-7334 lov: Cleanup lov_stripe proc files"

Causes problems with default striping described in LU-8338

This reverts commit d34b9dfbebc8d313be9c12a96b660a34ad45edcb.

Change-Id: Ib09cc449f9d826f31799ea0a2211820d3c443456
Reviewed-on: http://review.whamcloud.com/21129
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8055 ldev: Add %H and %b to command sub

Introduced %H and %b substitutions for hostname
and backend file system, respectively.

Test-Parameters: trivial testlist=conf-sanity
Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Change-Id: I9b50defee43307f614106de4ad75e1fa9267e23b
Reviewed-on: http://review.whamcloud.com/20040
Reviewed-by: Olaf Faaland-LLNL <faaland1@llnl.gov>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8055 ldev: File system label filtering

Introduced -F option to ldev. The -F requires
a file system name to be provided. All labels
associated with that file system will be
printed to stdout. The new option can also be
used with command substitution.

Test-Parameters: trivial testlist=conf-sanity

Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Change-Id: I50726371dce809e4d1b4f3de6f530d6f637856a1
Reviewed-on: http://review.whamcloud.com/19738
Tested-by: Jenkins
Reviewed-by: Olaf Faaland-LLNL <faaland1@llnl.gov>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7317 tests: Called the error function where it was missing.

There were bunch of files where error function was missing
Found this by executing grep '|| "' ./*.sh -r
and grep '&& "' ./*sh -r in lustre/tests directory.
Called the error function for those messages.

Seagate-bug-id: MRP-3028
Change-Id: I204008234e36dcb49e28e5256b93bbb0bbaeaea5
Signed-off-by: Aditya Pandit <aditya.pandit@seagate.com>
Signed-off-by: Ashish Purkar <ashish.purkar@seagate.com>
Reviewed-on: http://review.whamcloud.com/16888
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Jenkins
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8331 kernel: kernel update [SLES11 SP4 3.0.101-77]

Update SLES11 SP4 kernel to 3.0.101-77

Test-Parameters: mdsdistro=sles11sp4 ossdistro=sles11sp4 \
  clientdistro=sles11sp4 mdsfilesystemtype=ldiskfs \
  mdtfilesystemtype=ldiskfs ostfilesystemtype=ldiskfs \
  testgroup=review-ldiskfs

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: Iae1a185cfd054a43e4d4251e0e4f3bb4831065fd
Reviewed-on: http://review.whamcloud.com/21002
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8316 tgt: return -ENOMEM while kmalloc failed

Return valid value in tgt_init().

Signed-off-by: Yang Sheng <yang.sheng@intel.com>
Change-Id: Iaaa3697aca42d84011886a360fb31576ed4b53df
Reviewed-on: http://review.whamcloud.com/20918
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8270 ptlrpc: correct use of list_add_tail()

In sptlrpc_gc_add_sec() swap the arguments to list_add_tail() so that
it does what we meant it to do.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: I98f4cc0cdf3e271f003c1242b36a317025974c02
Reviewed-on: http://review.whamcloud.com/20784
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8056 lloop: fix bio_for_each_segment_all for newer kernels

Lustre patch http://review.whamcloud.com/20478 back ported
bio_for_each_segment_all from newer kernels but support
for newer kernels was done incorrectly. Update the code to
work with newer kernels.

Change-Id: I6a926320f80113169a13d2319190721c83d58b1d
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: http://review.whamcloud.com/20619
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: Li Dongyang <dongyang.li@anu.edu.au>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7334 lprocfs: Allow default multiplier of 1

A new default unit '1' is accepted by lprocfs_str_with_units_to_s64
to indicate a default multiplier of 1.

Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Change-Id: Iebf29f5d3c36e583729ea717288850729b99eeed
Reviewed-on: http://review.whamcloud.com/20455
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8174 llite: restore fd_och when putting lease

fd_och was not restored when putting back a file lease, preventing
from getting a lease, putting it back and taking it again on a FD.

Expand regression tests accordingly.

Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Change-Id: I1ad63c5332ee14bdac780f1d74794ac8827186cf
Reviewed-on: http://review.whamcloud.com/20331
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jean-Baptiste Riaux <riaux.jb@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8056 o2iblnd: ib_query_device removed in 4.5

We should use the cached attributes in ib_device instead of
calling ib_query_device since kernel 4.5

Linux-commit:182a2da0c768a9ec64abb0d6009667057f1c06af
Linux-commit:cebfe5ca038e2a0f677b41e9682501708ffa2ff3

Signed-off-by: Li Dongyang <dongyang.li@anu.edu.au>
Change-Id: Ie2ab681a3b508f0a1f89d3ff86bf2713b7257e62
Reviewed-on: http://review.whamcloud.com/20221
Tested-by: Jenkins
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8003 llite: remove obsolete code and comments

Remove obsolete comments about the behavior of ll_ddelete() and the
removed function ll_d_iput(). Remove '#if 0'-ed out code from
ll_ddelete().

Test-Parameters: trivial

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: Id2f2eed8e96ee07fdcd4ec14c4e82ed463752562
Reviewed-on: http://review.whamcloud.com/19881
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@seagate.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6401 obdclass: make get/set_wire_obdo non-inline

In http://review.whamcloud.com/#/c/16917/ it was requested that
lustre_get_wire_obdo and lustre_set_wire_obdo to not be
inlined functions.

Signed-off-by: Ben Evans <bevans@cray.com>
Change-Id: Ia0537dfbf2fdbd5e83fe49cda4ab6213d999e5e2
Reviewed-on: http://review.whamcloud.com/19266
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3105 llite: Remove unused lli_open_count from ll_inode_info

It was only used by now removed capabilities code.

Change-Id: Ica5617b0408cacbfbc7b08000e5d98a655f08e31
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/18544
Tested-by: Jenkins
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-5282 mdc: fix panic at mdc_free_open()

Assertion was happened for open request when rq_replay is set
to 1.
    ASSERTION(mod->mod_open_req->rq_replay == 0)
But this situation is not fatal for client, and could happened
when mdc_close() failed.
The fix allow to free such requests. If mdc_close fail, MDS doesn`t
receive close request from client. And in a worst case client would
be evicted.

The test recreates issue when mdc_close failed and
client asserts:
   ASSERTION( mod->mod_open_req->rq_replay == 0 ) failed

Signed-off-by: Alexander Boyko <alexander.boyko@seagate.com>
Seagate-bug-id: MRP-3156
Change-Id: I5f98901f633355849fc107149eadc9cf171819af
Reviewed-on: http://review.whamcloud.com/17495
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7931 tests: setup/cleanup after every test script

Setup and cleanup the filesystem after every acceptance-small.sh
test to try and catch problems with module references when they
happen instead of causing later tests to fail in mysterious ways.

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I0136ef81c601e7b9ee968ac62d674a8b16500c1e
Reviewed-on: http://review.whamcloud.com/6208
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8303 lnet: make connection more stable with packet loss

IB network may lose last connection handshake packet.
This problem isn't Lustre specific and described at
https://oss.oracle.com/pipermail/rds-devel/2007-December/000271.html
as example.
Solution is to make conection established if any packet is received
for it.

Signed-off-by: Alexander Boyko <alexander.boyko@seagate.com>
Signed-off-by: Alexey Lyashkov <alexey.lyashkov@seagate.com>
Seagate-bug-id: MRP-2883
Change-Id: I9bf69fdca24e51a5de06c92e9ad76b4e040c5d65
Reviewed-on: http://review.whamcloud.com/20874
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@seagate.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6215 build: announce linux kernel 4.4.6 support

Bump kernel version in ChangeLog to latest supported
kernel which is 4.4.6

Change-Id: I920e1139cc7abef82677e9ba79abfdc137dd205b
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: http://review.whamcloud.com/20543
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8125 nrs: pol_arg should be copied after the policy starts

NRS policy->pol_arg should be copied to a given policy argument
after the policy starts, otherwise a wrong argument will be saved
and it will cause LBUG when we run "lctl get_param" to get NRS
policy.
sanityn.sh test_77h is added to verify this patch.

Signed-off-by: Emoly Liu <emoly.liu@intel.com>
Change-Id: I5a015bb44fa88f69a4d5347cc2721eb72a57c8ad
Reviewed-on: http://review.whamcloud.com/20164
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Jean-Baptiste Riaux <riaux.jb@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7623 lnet: remove useless cast

In the cleanup of the user land / kernel pointers a
small cleanup was missed which removed a cast no
longer needed for a call of lnet_ping(). This brings
us into sync with the upstream client.

Change-Id: Ib281aaac7f2a1516d8586fbf3d10dee18cd77873
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: http://review.whamcloud.com/18950
Tested-by: Jenkins
Reviewed-by: Frank Zago <fzago@cray.com>
Reviewed-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8022 lnet: Correct position of lnet_ni_decref()

In fix http://review.whamcloud.com/#/c/19614/, the call
to lnet_ni_decref() should have followed the routines
which are using the NI.  This patch correct that.

Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Change-Id: I55fba4904cc6442f63cf03bf2e3614a7946200f7
Reviewed-on: http://review.whamcloud.com/21001
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4039 tests: ensure osc import in FULL state

The osp import may not in FULL state when
replay-single:test_90 start running. It
was caused by test_89, we should waiting
import valid then start next step.

Test-Parameters: trivial testlist=replay-single,replay-single,replay-single,replay-single
Signed-off-by: Yang Sheng <yang.sheng@intel.com>
Change-Id: I93746137c038183bf114619f53654af1bcaa68eb
Reviewed-on: http://review.whamcloud.com/20931
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-7759 llite: handle inactive OSTs better in statfs

Change the order of checks for inactive OSCs in lov_prep_statfs_set()
so that administratively disabled OSTs do not generate any output in
"lfs df" at all, to avoid needlessly cluttering the output.

Enable the lazystatfs mount option by default, so that "df" does not
hang when an OST is temporarily offline.

Fix test-framework.sh to use "lfs df $MOUNT" instead of "df $MOUNT" to
determine when recovery is complete, now that "df" does not block.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I993761a7eb120a36a1b80c2822cb6d8011ccab07
Reviewed-on: http://review.whamcloud.com/19195
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8259 osd-ldiskfs update all *_seq_write() to use "const char __user *buffer"

All *_seq_write functions in lustre/osd-ldiskfs/osd_lproc.c declare
buffer as "const char *buffer" and should use "const char __user *buffer"
instead.

This patch takes care of this change.

Change-Id: Ida44679f721a37cf71f327a6f00f3f8857ba0376
Signed-off-by: Parinay Kondekar <parinay.kondekar@seagate.com>
Reviewed-on: http://review.whamcloud.com/20887
Tested-by: Jenkins
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8176 libcfs: Add file_dentry() & file_inode() wrapper

Use file_dentry() & file_inode() wrappers to sync
with upstream.

Signed-off-by: Yang Sheng <yang.sheng@intel.com>
Change-Id: I7e62871789617a07dfe083a60374885099d25d51
Reviewed-on: http://review.whamcloud.com/20576
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8197 ptlrpc: do not reduce replay request deadline

New timeout for a replay request sent to a client with early reply is
calculated without taking into account that new deadline for the
replay request on server side is calculated based upon
obd_recovery_timeout. That results in reducing replay deadline which
forces to early disconnect of the client.

Calculate the timeout sent in early reply to a replay request in
according to new deadline of the replay request on server side.

Seagate-bug-id: MRP-3249
Signed-off-by: Vladimir Saveliev <vladimir.saveliev@seagate.com>
Change-Id: I478ca84a9627816be886a077170285f8af4bfc29
Reviewed-on: http://review.whamcloud.com/20399
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Hongchao Zhang <hongchao.zhang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8159 osd-ldiskfs: cache xattr in osd-ldiskfs

In LU-7660 LOD doesn't cache default striping, so each create will
get it from bottom. LU-7660 has enabled xattr cache in OSP, and
osd-zfs caches xattr already, this patch introduces a similar
mechanism in osd-ldiskfs.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: I3382119acaf9a819ee2be9737d1eae833e9ea2a4
Reviewed-on: http://review.whamcloud.com/20496
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7702 ldlm: skip lock if export failed

This patch resolves IO vs eviction race.
After eviction failed export stayed at stale list,
a client had IO processing and reconnected during it.
A client sent brw rpc with last lock cookie and new connection.
The lock with failed export was found and assert was happened.
(ost_handler.c:1812:ost_prolong_lock_one())
ASSERTION( lock->l_export == opd->opd_exp ) failed:

1. Skip the lock at ldlm_handle2lock if lock export failed.
2. Validation of lock for IO was added at hpreq_check(). The lock
   searching is based on granted interval tree. If server doesn`t
   have a valid lock, it reply to client with ESTALE.
3. A lock prolong was moved at ldlm layer.

Signed-off-by: Alexander Boyko <alexander.boyko@seagate.com>
Seagate-bug-id: MRP-2787
Change-Id: I2177736564c3b8164f1ad3e4cc02dca1704a0e6e
Reviewed-on: http://review.whamcloud.com/18120
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-5614 build: use %kernel_module_package in rpm spec

Replace module packages definitions with calls to the vendor defined
%kernel_module_package macro.  This results in many of the rpm packages
changing names from what we are used to in the past.  Each Linux
distribution may name the modules packages slightly differently, most
notably RHEL and SLES choose different naming schemes.  The
%kernel_module_package macro handles these details for us.

For instance, on RHEL7 the packages that contain lustre's kernel modules
become:

  kmod-lustre             (previously in "lustre-modules" package)
  kmod-lustre-osd-ldiskfs (previously in "lustre-osd-ldiskfs" package)
  kmod-lustre-osd-zfs     (previously in "lustre-osd-zfs" package)
  kmod-lustre-tests       (previously in "lustre-tests" package)

On SLES12, the same packages are instead named:

  lustre-kmp-default
  lustre-osd-ldiskfs-kmp-default
  lustre-osd-zfs-kmp-default
  lustre-tests-kmp-default

These kernel module packages give us what is known as "weak-updates".
Instead of compiling Lustre for every minor kernel that is installed,
the weak-updates will automatically symlink the Lustre modules into
and kernel's modules subdirectories if that kernel offers the required
kernel symbols with exactly the same symbol versions.

In addition to changing the package names for all of the Lustre packages
containing kernel modules, the location of the modules in /lib/modules
changes as well.  Each package will put its modules in a subdirectory
matching part of the package name itself (required by the macros on RHEL
systems).  The path name is also different on RHEL and SLES systems
(e.g. "extra" vs. "updates").

Some further path examples are:

  RHEL7, kmod-lustre package:
    /lib/modules/3.10.0-327.18.2.el7_lustre.x86_64/extra/lustre/fs/ost.ko

  SLES12, lustre-kmp-default package:
    /lib/modules/3.12.57-60.35_lustre-default/updates/lustre/fs/ost.ko

  RHEL7, kmod-lustre-osd-ldiskfs package:
    /lib/modules/3.10.0-327.18.2.el7_lustre.x86_64/extra/lustre-osd-ldiskfs/fs/osd_ldiskfs.ko

  SLES12, lustre-osd-ldiskfs-kmp-default package:
    /lib/modules/3.12.57-60.35_lustre-default/updates/lustre-osd-ldiskfs/fs/osd_ldiskfs.ko

Test-Parameters: trivial clientdistro=el6.7 mdsdistro=el6.7 ossdistro=el6.7
Signed-off-by: Stephen Champion <schamp@sgi.com>
Signed-off-by: Alexander Boyko <alexander.boyko@seagate.com>
Change-Id: Ic5351f4bda94eae80eaf349c8490a2c758b15c80
Signed-off-by: Minh Diep <minh.diep@intel.com>
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/12063
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7418 tests: avoid background writeout in sanity 29

Intermittent failures of sanity.sh test_29 are caused by background
writeout starting (or continuing) after the MDC DLM locks were
cancelled, requiring the layout lock to be re-fetched from the MDS.
Cancelling the DLM locks via "lru_size=clear" is not enough, since
this will not cancel any locks that are currently in use.

Instead, ensure that all the dirty pages from previous tests are
written out so that no new writes are started during the test.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: Ie45965a974e0cd170f969fa122e1762c2361c278
Reviewed-on: http://review.whamcloud.com/20929
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Revert "LU-5810 tests: add client hostname to lctl mark"

Adding hostname to individual lines in test logs broke autotest's
parsing of logs and resulted in no console logs being uploaded to
Maloo. In addition, the way that we added hostname to the log
output was nosier than expect and needs to be reworked.

This reverts commit 9c4156e6fc146a198bb342e28eb246f1076889bd.

Change-Id: I6d7c38fd37edc975bc4f5e1ff5c825d4899af474
Signed-off-by: James Nunez <james.a.nunez@intel.com>
Reviewed-on: http://review.whamcloud.com/20894
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Tested-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7542 obdclass: quiet lockdep recursive lock warning

Lockdep complains about potential recursive locking during mount
because the client configuration log is holding a lock on the MGC
obd_device to prevent it from being torn down, while also getting
locks on the MDC and OSC devices as they are instantiated:

 Lustre: Mounted myth-client
 =============================================
 [ INFO: possible recursive locking detected ]
 ---------------------------------------------

 2 locks held by ll_cfg_requeue/1460:
 #0:  (&cli->cl_sem){.+.+..}, at mgc_requeue_thread+0x27d/0x5d8 [mgc]
 #1:  (&cld->cld_lock){+.+...}, at mgc_process_log+0x88/0xc01 [mgc]

 CPU: 3 PID: 1460 Comm: ll_cfg_requeue
 Call Trace:
 [<ffffffff814cb360>] dump_stack+0x4f/0x6f
 [<ffffffff81081d73>] __lock_acquire+0xc06/0xc74
 [<ffffffff81081f4f>] lock_acquire+0x16e/0x194
 [<ffffffff814ce5d7>] down_read+0x4c/0x61
 [<ffffffffa029f370>] sptlrpc_conf_client_adapt+0x95/0x142 [ptlrpc]
 [<ffffffffa03a7a69>] mdc_set_info_async+0xeb/0x420 [mdc]
 [<ffffffffa01ad69e>] class_notify_sptlrpc_conf+0x2f4/0x347 [obdclass]
 [<ffffffffa04a805c>] mgc_process_log+0xa31/0xc01 [mgc]
 [<ffffffffa04a94ef>] mgc_requeue_thread+0x31f/0x5d8 [mgc]
 [<ffffffff81061f77>] kthread+0xfa/0x102

Add a separate lock class for the MGC callpath, since it will always
be held first, and none of the other obd_device locks should ever
be held concurrently.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: Ic36e64608d12dcd4f5b12056e3e2072ca6500c1e
Reviewed-on: http://review.whamcloud.com/17582
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7334 lov: Cleanup lov_stripe proc files

Make lov_stripe* proc files read only. Writing to
these proc files no longer has any impact.

Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Change-Id: I5adfbbe529208efef760df22c849e3684a34e187
Reviewed-on: http://review.whamcloud.com/20465
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-8203 tests: Wait for archive operation to be completed.

As bandwith control is set to 1048576B/s in the test, so
the archive operation takes <= 40 secs to copy 39000000
bytes.

If we dont wait for the on going archive action to finish
then status will remain as STARTED and never becomes
SUCCEED as we call copytool_cleanup soon after archiving
5MB of data and the request will be stuck there for ever.

Sometimes because of previous unclean shutdown of copytool,
test_60 results in failure because of stuck request.

Test-Parameters: trivial testlist=sanity-hsm,sanity-hsm

Change-Id: I2413c32a617e43fb24e7ca6696026a3c9e1cad00
Seagate-bug-id: MRP-2586
Signed-off-by: Kirtankumar Krishna Shetty <kirtan.shetty@seagate.com>
Reviewed-on: http://review.whamcloud.com/20442
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8058 lnet: Remove old commented out code

These #if 0 blocks have been in place for years. Assume
they are not used and remove them

Signed-off-by: Ben Evans <bevans@cray.com>
Change-Id: I3840513cc68786ded878ba8be6b92f0a180507a2
Reviewed-on: http://review.whamcloud.com/20414
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8150 mdt: Track open+create as mknod

Regular files created with mknod are tracked, but most files are
created as an open with create intent. Track this as a mknod also.

Signed-off-by: Matt Ezell <ezellma@ornl.gov>
Change-Id: I8f6ab9d91d0b31062c7f73f8394a1b12f7963f1c
Reviewed-on: http://review.whamcloud.com/20246
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8116 build: Fix libmount configure message

The configure messages for libmount checking were incorrectly nested resulting
in messages like this:

checking whether build with libmount... checking libmount/libmount.h usability... no
checking libmount/libmount.h presence... no
checking for libmount/libmount.h... no
no

This patch unwinds that mistake.

Change-Id: Iec716d8152880902d3e6961cd1d3b821e37c349a
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/20055
Tested-by: Jenkins
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-7896: do not call lu_site_purge() for single object exceed

first of all, this is expensive procedure including a global
mutex and per-bucket spinlocks. also, all the threads observed
exceed will be calling lu_site_purge() and essentially serialized
on that. instead we can let other threads to skip the whole
procedure.

Change-Id: Ib5d213c70b5a2f259153ab24d7c3f5a3e97239df
Signed-off-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-on: http://review.whamcloud.com/19082
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7664 test: set_conf_param_and_check() defect

set_conf_param_and_check() should wait update on facet_active_host
instead of facet_host; othewise tests failed on configuration
facet_HOST != facetfailover_HOST

Seagate-bug-id: MRP-2846
Signed-off-by: Elena Gryaznova <elena.gryaznova@seagate.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@seagate.com>
Reviewed-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Change-Id: I974f145b3783fec1adb2fc6219ef4bd025a31811
Reviewed-on: http://review.whamcloud.com/17995
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-5092 nodemap: remove nodemap_idx_action, only act on MGS

Only the MGS config needs to be modified on config changes.
The target configs will be loaded from the MGS and written
whole to disk.

Signed-off-by: Kit Westneat <kit.westneat@gmail.com>
Change-Id: I3114258d6e18d293cdf29ae47a28465f245387de
Reviewed-on: http://review.whamcloud.com/18554
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7403 obd: const correct md_set_lock_data()

Change the __u64 *cookie parameter of md_set_lock_data() to const
struct lustre_handle *lockh.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: Ie6de252a10c532ee395e407ad2aff258773478a8
Reviewed-on: http://review.whamcloud.com/17072
Reviewed-by: Frank Zago <fzago@cray.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8216 ldiskfs: fix journal quota files for RHEL6.8

When we hit Lustre crash, e2fsck always complains of Quota
accounting mismatch, and sometimes differences are huge.

The problem is that we don't journal quota updates properly, and
with current codes quota files will be only journaled when syncing
the filesystem, which only happens by sync() call or unmount.

RHEL6.8 version of the patch http://review.whamcloud.com/20503

Signed-off-by: Wang Shilong <wshilong@ddn.com>
Change-Id: Id5b580083197555029442338f7feaf59b1ad154c
Reviewed-on: http://review.whamcloud.com/20791
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-8220 tests: stripe file over all active OSTs

In test_iozone/parallel-scale-cifs, the iozone command creates the
temporary files under test with default stripe count=1, that is,
both the 256mb files reside on the same OST, over-utilizing the OST,
resulting in Use%=100%(No space left on device).

The fix is to create a sub-directory in the samba server mount
point with a stripe count of -1 (stripe over all available OSTs)
and then run iozone with this sub-directory path, so that all the
temporary files created within this sub-directory, during iozone,
reside on multiple OSTs and OST free space is utilized uniformly.

Since the test suite mounts lustre client on smb server, the "lfs
setstripe" command is executed on the smb server.

Test-Parameters: trivial

Seagate-bug-id: MRP-2179
Signed-off-by: Noopur Maheshwari <noopur.maheshwari@seagate.com>
Change-Id: If225b62848b5266dc51ef791ef0d9c8002f908a5
Reviewed-on: http://review.whamcloud.com/20515
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8106 lnet: Do not drop message when shutting down LNet

There is a case in lnet_parse() where we discover that LNet
is shutting down but we continue to use the NI when we drop the
message and end up calling ko2iblnd_check_send_locked() which tries to
allocate from the Tx pool which has been cleaned up already.
This triggers a NULL pointer dereference.

This fix just returns from lnet_parse() when we disover LNet is
shutting down.

Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Change-Id: Ia8dba848e88b02ed23e2e77adb702c7ad0e60628
Reviewed-on: http://review.whamcloud.com/19993
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7861 osp: replace the hot spinlock with atomic trackers

so that background scanning doesn't race with MDS_REINT_UNLINK

Change-Id: I4006be4298ffa841ca4d131ac37c201ff8ea729f
Signed-off-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-on: http://review.whamcloud.com/19211
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7204 test: handle FSNAME with dashes in replay-single 85b

If FSNAME contains a dash, test_85b incorrectly parses lctl dl
output. Fix that.
Also changed the checking of before recovery count value
by using numeric comparison instead of string comparison.
Used tabs in place of spaces.

Test-Parameters: trivial testlist=replay-single envdefinitions=ONLY=85b
Seagate-bug-id: MRP-2792
Signed-off-by: HemaHarish <hema.yarramilli@seagate.com>
Change-Id: I392921b837dee573a704a3fec4b35c7bb79136bc
Reviewed-on: http://review.whamcloud.com/16902
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7301 test: do cleanup only if required

change check_and_cleanup_lustre() to do cleanup only if required

Seagate-bug-id: MRP-1257
Signed-off-by: Chennaiah Palla <chennaiah.palla@seagate.com>
Change-Id: I22f3b2a448c78b0468c5d802a370a51aab3c9f8d
Reviewed-on: http://review.whamcloud.com/16887
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-8040 mgc: Apply changelog defaults values in  mdd_init0()

lctl set_param -P command stores options to params file. This options
are applied in server_start_targets() during target mount. Changelog
apply default options during initialization. So it reset options
applied by "lctl set_param -P".

This patch moves changelogs default value appying before params
processing. So new parameters aplied after default.

Signed-off-by: Artem Blagodarenko <artem.blagodarenko@seagate.com>
Change-Id: Ibb1a8d512844d4fcb66c051441c1864f994d4ded
Reviewed-on: http://review.whamcloud.com/19854
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

New tag 2.8.55

Change-Id: Ie0b88ee537b2f76d59a49e3512c834f5314b0eae
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7860 ldlm: revert part of commit 657bbc49

During ldlm_lock_cancel()->ldlm_cancel_callback(), the lock could
possibly be added back to waiting list with no resource lock holding.

Put back the second checking and removing from the waiting list.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: I8841422d6e2f6bf92efa24d542583936123ead66
Reviewed-on: http://review.whamcloud.com/20509
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-5564 mdt: skip permission check for close

When closing a file normally, skip the permission check
for open handle gives us permission.

Change-Id: I74271eb83cdd807940cb6ba37ede05fa9f9751ce
Signed-off-by: Hongchao Zhang <hongchao.zhang@intel.com>
Reviewed-on: http://review.whamcloud.com/15534
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-5969 procfs: restore missing newline for version param

In obd_proc_version_seq_show() restore the missing newline at the end
of the version param value.

Test-Parameters: trivial

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: I7f8eecc418ceb3191650d7dfba8d1e2545af631b
Reviewed-on: http://review.whamcloud.com/20809
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Frank Zago <fzago@cray.com>
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-5953 build: use installed OFED by default with dpkg

Add support for dpkg based systems (Ubuntu) when detecting installed
OFED drivers.

Lustre-change: http://review.whamcloud.com/16455

Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Minh Diep <minh.diep@intel.com>
Change-Id: Ifb0b1c41e2148a59e894fbf075ae9db5b8fe5297
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-on: http://review.whamcloud.com/20523
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>

LU-7004 utils: remove "lctl conf_param" deprecation

The "lctl set_param -P" functionality was intended to replace
"lctl conf_param" with a more uniform interface that matches
the existing "lctl set_param" and "lctl get_param" usage.
However, there are still unresolved issues that need to be
addressed and a longer transition period needed before the
old functionality can be removed.

Remove the deprecation warning for now, rather than bumping it
to some later release.  It can be re-added when "set_param -P"
has been fixed.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: If9a316bda4bf8d889c4072740f0879d9d43ebbe5
Reviewed-on: http://review.whamcloud.com/20573
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6971 cleanup: not support remote client anymore

There are several obsolete sub commands for lfs to work with
remote client. We do not support that anymore, and should be
deleted along with any kernel code related to remote client.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I06f4348b2b3def5ac4aa75676b1b5b7ccc4ab10e
Reviewed-on: http://review.whamcloud.com/19789
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8277 scripts: add missing commas in checkpatch.pl

In checkpatch.pl add some missing commas to the table of deprecated
symbols.

Test-Parameters: trivial
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: I409d930db5863b983370c3286d75b357bf4090a1
Reviewed-on: http://review.whamcloud.com/20786
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2766 llite: don't ignore layout for group lock request

ignore_layout can be set for operations that layout won't be changed,
typically page operations. Ignoring layout change in group lock
request will confuse layout change code at LOV layer and hit
assertion.

Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: I0f628e81d7f61a9b00ab2941ad3bcde4b32ecfe2
Reviewed-on: http://review.whamcloud.com/6828
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7893 osd-zfs: fix error handling in osd_mount()

Do not try to release objectset if we failed to open it.

Change-Id: I85926d0b0a7fb5faf832a3907eac2324e7e6c07f
Signed-off-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-on: http://review.whamcloud.com/15671
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>

LU-3289 gss: Cleanup gss code

Address some lingering style and code issues that were commented on
but not fixed.

Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: I2396bf3a6a6e3621aa37fd5592eb9eeaeedb06e7
Reviewed-on: http://review.whamcloud.com/20727
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-7434 ptlrpc: lost bulk leads to a hang

This is a combination of two commits. The original commit from:
http://review.whamcloud.com/17221
and a fix to that commit from:
http://review.whamcloud.com/19758

Description of the original patch:
The reverse order of request_out_callback() and reply_in_callback()
puts the RPC into UNREGISTERING state, which is waiting for RPC &
bulk md unlink, whereas only RPC md unlink has been called so far.
If bulk is lost, even expired_set does not check for UNREGISTERING
state.

The same for write if server returns an error.

This phase is ambiguous, split to UNREG_RPC and UNREG_BULK.

The fix to the original commit was originally pushed against LU-8062.
That fix is described thusly:
LU-8062 test: fix fail_val in recovery-small/115b

The fail_loc OBD_FAIL_OST_ENOSPC is caught in 2 places:
1. ofd_statfs() - where it is checked against the fail_val
and if fail_val matches the OST ID it is caught here
itself and hence test fails.
2. tgt_brw_write() - the actual place where the fail_loc should
be caught.

The patch makes the fail_loc to be caught at the appropriate place
by setting the fail_val to $OSTCOUNT. So even if fail_loc is set
in the statfs part it is not caught there and therefore caught in
tgt_brw_write().

Please note the author of the original patch was Vitaly Fertman
<vitaly.fertman@seagate.com>, and the author of the test fix was
Bhagyesh Dudhediya <bhagyesh.dudhediya@seagate.com>

Test-Parameters: testlist=recovery-small,recovery-small,recovery-small,recovery-small,recovery-small,recovery-small
Signed-off-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Signed-off-by: Bhagyesh Dudhediya <bhagyesh.dudhediya@seagate.com>
Signed-off-by: Chris Horn <hornc@cray.com>
Seagate-bug-id:  MRP-2953, MRP-3206, MRP-3150
Reviewed-by: Andriy Skulysh <andriy.skulysh@seagate.com>
Reviewed-by: Alexey Leonidovich Lyashkov <alexey.lyashkov@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Change-Id: I17319d40881c41f247c102aafc3a1b0db82d0b4a
Reviewed-on: http://review.whamcloud.com/19953
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Ann Koehler <amk@cray.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7846 tests: test for nodemap fileset in sanity-sec

Add new tests in sanity-sec.sh to test for fileset in nodemap.

Test-Parameters: testlist=sanity-sec
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I7903d0282d15df84e1f48028d3f44094cac20e98
Reviewed-on: http://review.whamcloud.com/19895
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Kit Westneat <kit.westneat@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7813 utils: improve lfs setstripe pool error message

When a specified OST is not in the requested pool, improve the error
message printed to the user.  Currently it only prints:

  $ lfs setstripe -p audio -i 1 /myth/foo
  error: setstripe: create file '/myth/foo' failed

but this doesn't indicate why it failed.  Add in more information:

  error: setstripe: create file '/myth/foo' failed: OST not in pool?
  error: setstripe: create file '/myth/foo' failed: permission denied

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I409739b1e3dad633f0b0a4c4f94a746c0b6485bd
Reviewed-on: http://review.whamcloud.com/18965
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Frank Zago <fzago@cray.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7445 test: combine dependent sanity 51b and 51ba tests

test_51ba() depends on test_51b() and skipped if run separately

Patch combines test_51b() and test_51ba() into test_51b().
$tdir is used for test dir name.

Test-Parameters: envdefinitions=ONLY=51b testlist=sanity
Seagate-bug-id: MRP-2290
Signed-off-by: Elena Gryaznova <elena.gryaznova@seagate.com>
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@seagate.com>
Reviewed-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Change-Id: I6ee019a03d398e36c7007973614375c9ca22e349
Reviewed-on: http://review.whamcloud.com/17239
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4048 build: fix 'control flow' errors

Fix "control flow" issues found by Coverity version 6.6.1:
Logically dead code (DEADCODE)
Execution cannot reach this statement.

Signed-off-by: Sebastien Buisson <sebastien.buisson@bull.net>
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I0be1b0f69b16a988e361185f6c4ef2d9dcee8d31
Reviewed-on: http://review.whamcloud.com/7824
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7934 osp: fix tr->otr_next_id overflow

The tracker next_id and current_id u32 type was based on
max llog records. But llog use cyclic store for records, so
llog could store infinite number of records and limited by
max number at moument of time. The u32 type could be
overflowed easy if a server isn`t rebooted.
osp_sync_id_get()) snx11126-OST0045-osc-MDT0000: next 0,
last synced 4294967205
This fix changes type u32 to u64 for *id. Now, we store only
low part current_id to llog record header id. And restore the
full u64 from record header later. It is possible because
llog catalog can store 64768^2 and it is less than u32 max value.

The patch adds test to check u32 overflow for otr_next_id field
of osp_id_tracker.

Recreate the next assertion
LustreError: 185667:0:(osp_sync.c:1544:osp_sync_id_get())
snx11126-OST0045-osc-MDT0000: next 0, last synced 4294967205
LustreError: 231396:0:(osp_sync.c:1545:osp_sync_id_get()) LBUG

Seagate-bug-id: MRP-3367
Signed-off-by: Alexander Boyko <alexander.boyko@seagate.com>
Change-Id: I89d70ecb068f8d0b5a1e1ac35b85a4b6e53211e5
Reviewed-on: http://review.whamcloud.com/19190
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@seagate.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7536 mdt: handling NULL dereference in mdt_reconstruct

mdt_reconstruct function corresponding to op-code
REINT_MIGRATE is not present, so as per request format
used in mdt_reint adding mdt reconstruct functions.
Also added assertion to check max opcodes and undefined
reconstruct function. Also updated opcode REINT_RMENTRY
and REINT_MIGRADE at missing places.

Seagate-bug-id: MRP-3205
Signed-off-by: Jadhav Vikram <jadhav.vikram@seagate.com>
Change-Id: I489ee3dfa5972c14f37aed5307103cf329d802f7
Reviewed-on: http://review.whamcloud.com/20395
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Gu Zheng <gzheng@ddn.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8279 scrub: fix inode reference leak

There was logic issue in osd_fid_lookup() to handle remote directory
under the case of MDT file-level backup and restore: the logic wanted
to trigger OI scrub when found the OI mapping for the remote directory
object invalid, but forgot to assign the valid inode to the object as
to such inode's reference cannot for released properly.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: Ib2c6dd7a3b43bd75b5e5aaadad91a1a72ef3d54e
Reviewed-on: http://review.whamcloud.com/20792
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4781 utils: handle EEXIST in lustre_rsync

In lustre_rsync if the file to be created in the target already exist
then handle this rather than returning an error.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: I8c04c80cf8618b3681be12d098be0f07a305df30
Reviewed-on: http://review.whamcloud.com/20649
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-8232 obdclass: remove dead code

No need to check for rc since it is never assigned after declaration
and sub-calls do not return error codes.

Test-Parameters: trivial

Change-Id: I040a45e8a779b11840c8c883fd4d5d87ee7f0fed
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-on: http://review.whamcloud.com/20587
Reviewed-by: Ben Evans <bevans@cray.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8227 kernel: kernel update RHEL6.8 [2.6.32-642.1.1.el6]

Update RHEL6.8 kernel to 2.6.32-642.1.1.el6

Test-Parameters: clientdistro=el6.8 mdsdistro=el6.8 ossdistro=el6.8 \
  mdsfilesystemtype=ldiskfs mdtfilesystemtype=ldiskfs \
  ostfilesystemtype=ldiskfs testgroup=review-ldiskfs

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I33789db4827b36a25332e5d7e85729ac036d0a4b
Reviewed-on: http://review.whamcloud.com/20542
Tested-by: Jenkins
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7747 lfs: fix help message for migrate

Combine two different help messages for migrate in one.

Test-Parameters: trivial

Change-Id: I2b44006dbd9114155cf7a1f65e68d3796f857ee3
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/20541
Tested-by: Jenkins
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-8196 obd: force very large allocations to use vmalloc

In cases where the requsted allocation size is very large
use OBD_VMALLOC().  This avoids warnings from el7 and newer
kernels due to very large kmalloc()s.

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I87e95b1f8e5e50bc35d713a1775313c10f0368a9
Reviewed-on: http://review.whamcloud.com/20420
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8058 utils: Remove old commented out code

These #if 0 blocks have been in place for years. Assume
they are not used and remove them

remove empty function handle_spkm3_upcall

Signed-off-by: Ben Evans <bevans@cray.com>
Change-Id: Ia660048adae000a506836fe226c475affd0149b8
Reviewed-on: http://review.whamcloud.com/20415
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8182 mdt: check connection data properly

For mdt_obd_connect() case, the parameter @data should never be
NULL, so drop the redundant check of "data != NULL", replace it
with "LASSERT(data != NULL);"

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: Ia6754752527e69c89e7a40d3d012232eaba94d3e
Reviewed-on: http://review.whamcloud.com/20391
Tested-by: Jenkins
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-8178 lproc: fix negative recovery_duration

There is a window in target_recovery_thread where
obd_recovery_end is not set despite of
obd_recovering becomes 0. In such case
recovery_duration shows difference between 0 and
obd_recovery_start like below:
recovery_duration: 18446744072323902150.

Change-Id: Id385156585c99d12726e89980b55c588e0bf3ceb
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@seagate.com>
Seagate-bug-id: MRP-1519
Reviewed-by: Alexander Nikolaevich Boyko <alexander.boyko@seagate.com>
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@seagate.com>
Tested-by: Parinay Vijayprakash Kondekar <parinay.kondekar@seagate.com>
Reviewed-on: http://review.whamcloud.com/20381
Tested-by: Jenkins
Reviewed-by: Alexander Boyko <alexander.boyko@seagate.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>