LU-6210 utils: C99 initializer in llog_cancel_parse_optional()

This patch makes no functional changes.  The long_options struct
initializer in llog_cancel_parse_optional() is updated to C99 syntax.
The option struct and option string are renamed to long_opts and
short_opts for consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Ie6e1c823ebc59c3791c7c6b0360f33d1d0184c1a
Reviewed-on: https://review.whamcloud.com/27880
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>

LU-6210 utils: Use C99 initializer in jt_obd_md_common()

This patch makes no functional changes.  The long_opts struct
initializer in jt_obd_md_common() is updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Ifa7140e9980570940f8dfd1200307d02242a8c45
Reviewed-on: https://review.whamcloud.com/27879
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>

LU-9712 kernel: kernel update [SLES11 SP4 3.0.101-107]

Update SLES11 SP4 kernel to 3.0.101-107

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

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: Ic62b288becc936d54ee88539ab1635ca3a78cf7f
Reviewed-on: https://review.whamcloud.com/27841
Tested-by: Jenkins
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6051 utils: allow lfs_migrate to migrate links

When running on a server that supports "lfs migrate", it is possible
to migrate hard links transparently.  This is still not possible for
servers that do not support "lfs migrate" (layout swap) so we still
need to keep the hard-link check if "lfs migrate" fails.

Add test for migrating many hardlinks on supported kernels.

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I7fa1c70611aa7dd61f822844d5fe339f730640e5
Reviewed-on: https://review.whamcloud.com/27553
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Steve Guminski <stephenx.guminski@intel.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6210 utils: Use C99 struct initializers in lfs_hsm_request()

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.  Variables of
type struct option are renamed to long_opts for consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lfs_hsm_request() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I62b07e59a3f81cf8ee936ae46b1cd8f99913e53c
Reviewed-on: https://review.whamcloud.com/27526
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>

LU-6210 utils: Use C99 struct initializers in lfs_setquota()

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lfs_setquota() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I4bae3ccf40380d662b151b8ac53976f38c63daf6
Reviewed-on: https://review.whamcloud.com/27521
Tested-by: Jenkins
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: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6210 utils: Use C99 initializers in lfs_getstripe_internal

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lfs_getstripe_internal() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I865ee0200b58f649810452000d91fa455da4d7ba
Reviewed-on: https://review.whamcloud.com/27514
Tested-by: Jenkins
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: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9019 obdclass: migrate jobstats to 64 bit time

Change all time_t fields for jobstats to time64_t.
Replace cfs_time_current_sec() to avoid the overflow
issues in 2038 with ktime_get_real_seconds().

Change-Id: Icf3af60b9b373acbf5d921add9b4c4d33c77aa61
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/27911
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9728 osd: use GFP_HIGHUSER for non-local IO

When the obdfilter code was split into separate OFD and OSD modules,
the bulk IO page allocation was implemented to use GFP_NOFS to avoid
allocations recursing into the filesystem and causing deadlocks.

However, this is only possible if the RPC is coming from a local
client, as we might end up waiting on a page sent in the request we're
serving. Local RPCs use __GFP_HIGHMEM so that the pages can use all of
the available memory on the OSS on 32-bit machines.

It is possible to use more aggressive GFP_HIGHUSER flags for non-local
clients to be able to generate more memory pressure on the OSS and
allow inactive pages to be reclaimed, since the OSS doesn't have any
other processes or allocations that generate memory reclaim pressure.

See also b=17576 (bdf50dc9) and b=19529 (3dcf18d3) for details.

The patch also implements an LNet function to determine if a client NID
is local or not.  This becomes more complex in the LNet Multi-Rail world
and it is really LNet's job to handle NIDs, not that of Lustre.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I2806c9c5c2fe269669eafdafaf2310924c3ebbe5
Reviewed-on: https://review.whamcloud.com/27908
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8935 ptlrpc: missing barrier before wake_up

ptlrpc_client_wake_req() misses a memory barrier, which may cause
strange errors.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: Ic8e9cbaf8c07f503798b95c608477508204d9614
Reviewed-on: https://review.whamcloud.com/26583
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9019 lfsck: migrate to 64 bit time

Replace cfs_time_current_sec() to avoid the overflow
issues in 2038 with ktime_get_real_seconds(). The
reset of the lfsck code can move to time64_t as
well. Handle older lfsck on disk data formats that
contain 32 bit time formats. Newly generated on disk
data will always be 64 bit timestamps.

Change-Id: I1db47af3b9f462c31375f53a6216a095ab4767ad
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/26466
Reviewed-by: Fan Yong <fan.yong@intel.com>
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-7372 mgs: reprocess all locks at device fini

This is to avoid a case that IR lock revocation is going on when
the obd is being stopped, an extra ldlm_reprocess_recovery_done() is
required to make revocation process move forward.

Turn off 'set -e' in rundbench. Otherwise killing dbench process will
return an error to wait(1) in rundbench. Since test-framework has
turned on error on exit, it will set test result as failure, which
is actually a false alarm.

Test-Parameters: envdefinitions=SLOW=yes,ONLY=26 testlist=replay-dual,replay-dual,replay-dual,replay-dual
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: I43ab1df9c8fe5aea15da6c90175fd08a0b099ea2
Reviewed-on: https://review.whamcloud.com/17853
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8760 lib: avoid unexpected out of order execution

There is race condtion in __l_wait_event() because of the
out-of-order execution between changing thread state and
checking condition. It may block the thread (to be waken)
for ever. Consider the following real execution order:

1. Thread1 checks condition on CPU1, gets false.
2. Thread2 sets condition on CPU2.
3. Thread2 calls wake_up() on CPU2 to wake the threads with
   state TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE. But the
   Thread1'sstate is TASK_RUNNING at that time.
4. Thread1 sets its state as TASK_INTERRUPTIBLE on CPU1,
   then schedule.

If the '__timeout' variable is zero, the Thread1 will have
no chance to check the condition again.

Generally, the interval between out-of-ordered step1 and step4
is very tiny, as to above step2 and step3 cannot happen. On some
degree, it can explain why we seldom hit related trouble. But
such race really exists, especially consider that the step1 and
step4 can be interruptible.

The patch adds barrier between changing thread's state and
checking condition to avoid out-of-order execution.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I32caee6b332f037d864419ea8728112da563cce0
Reviewed-on: https://review.whamcloud.com/23564
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9776 lustre-client cannot be installed if both client and server repos are configured

Due to current Obsoletes: tags on lustre-client, the client cannot be
installed on nodes where the client and server repos are configured.

Update the Obsoletes: to only obsolete previous lustre-client versions,
not the current one.

Signed-off-by: Brian J. Murrell <brian.murrell@intel.com>
Change-Id: Ie2f022967fd6f65030feeb23ea9637dce505054a
Reviewed-on: https://review.whamcloud.com/28065
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-9731 kmods need to be limited to EL minor release kernel

Due to upstream RHBZ#1467319 kmods are not being populated with the full
kabi information needed to find a matching kernel for the kmod.

Until this is fixed, we need to apply a workaround to achieve the same
result.

Test-Parameters: trivial

Signed-off-by: Brian J. Murrell <brian.murrell@intel.com>
Change-Id: Ib2eab09719c75be8928eaf607efaa2d814baf5f2
Reviewed-on: https://review.whamcloud.com/28066
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9775 Look for kernel-devel in /usr/src/kernels

If one is building in a build [ch]root such as mock provides, one may
not have the kernel installed which corresponds to $(uname -r).  In such
a case, also try to look for the kernel-devel in /usr/src/kernels/ and
just build for the latest one.  Ideally there is only one installed in any case.

Test-Parameters: trivial

Signed-off-by: Brian J. Murrell <brian.murrell@intel.com>
Change-Id: I7704c6ce7078a507fd6b5f9178b07f750dc03789
Reviewed-on: https://review.whamcloud.com/28064
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9500 lnd: Don't Page Align remote_addr with FastReg

Trying to page align the remote_addr for IB_RDMA_WRITE work
requests is triggering "dump_cqe" errors from MOFED 4.x + mlx5.

This patch removes the address masking we were doing with FastReg
which was trying to page align remote_addr values. I am also
removing the setting of "mr->iova" with FastReg as this is being
done in the call to ib_map_mr_sg() and could cause problems.

Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Change-Id: If35baa467d8d60866f709b5feea7f619063c6da4
Reviewed-on: https://review.whamcloud.com/27149
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Gu Zheng <gzheng@ddn.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Amir Shehata <amir.shehata@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9749 llite: Reduce overhead for ll_do_fast_read

In ll_do_fast_read, looking up a cl_env adds some overhead,
and can also cause spinlock contention on older kernels.

Fast read can safely use the preallocated percpu cl_env, so
do that to reduce overhead.

SLES numbers on recent Xeon, CentOS numbers on VMs on
older hardware.  SLES has queued spinlocks and scales
perfectly with multiple threads, with or without this
patch.  CentOS scales poorly at small I/O sizes without
this patch.

SLES is SLES12SP2, CentOS is CentOS 7.3.

SLES:
1 thread
         8b   1K   1M
Without: 23   2200 6800
With:    27.5 2500 7200

4 threads
         8b   1K    1M
Without: 90   8700  27000
With:    108  10000 28000

Earlier kernel (CentOS 7.3):
1 thread
         8b  1K   1M
Without: 9   1000 5100
with:    12  1300 5800

4 threads
         8b  1K   1M
Without: 22  2400 17000
With:    48  4900 20000

Signed-off-by: Patrick Farrell <paf@cray.com>
Change-Id: Ice5d653ace5ce76bc8911501a9b15c11b7a3234a
Reviewed-on: https://review.whamcloud.com/27970
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>

LU-7129 tests: fsx with directio

This change adds O_DIRECT support to the existing fsx in lustre/tests.
It also add sanityn test_16 to excercise this option to fsx.

Upstream Xfstest-commit cad234da750aceb563992d95a0b14379740aedc6

Test-Parameters: trivial testlist=sanityn
Signed-off-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Signed-off-by: Parinay Kondekar <parinay.kondekar@seagate.com>
Change-Id: Ibd6a8dfeff8afaf08eeb3f2f8bddd7cdb688e8cf
Seagate-bug-id: MRP-852
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@seagate.com>
Reviewed-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Reviewed-on: https://review.whamcloud.com/16357
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-9772 utils: Enable new ZFS MMP on mkfs

ZFS 0.7.0 come with new multi-modifier protection, this patch
enables it by default, on mkfs.

This also ensures canmount is off for pools that were not just
created.

Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: If9b87e9786e0eaefe5ac9a536edcdca3d1012585
Reviewed-on: https://review.whamcloud.com/28051
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-9769 lnet: Fix lost lock

Unlock lnet_net_lock in case of error in function lnet_dyn_del_net().

Change-Id: Ie3fc8e27abc72077b9c0c330b5d325a47e5236e1
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: https://review.whamcloud.com/28039
Reviewed-by: Amir Shehata <amir.shehata@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-9019 target: migrate to 64 bit time

Replace cfs_time_current_sec() to avoid the overflow
issues in 2038 with ktime_get_real_seconds(). Change
time_t bi_deadline to time64_t. Do a direction compare
instead of cfs_time_beforeeq() wrapper.

Change-Id: I93d1cab9ebdcc1f910096e8a229d6c37b244f41d
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/27909
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-8849 ofd: Client hanges on ladvise with large start values

when the ladvise end is larger than file size, we don't judge
the ladvise start should be smaller than the ladvise end after
capped the end vaule to file size. This patch fixes the problem.

Change-Id: I30346c7ce69ad0d7ceca5ffc0d21d900afd3d2f1
Signed-off-by: Qian Yingjin <qian@ddn.com>
Reviewed-on: https://review.whamcloud.com/27209
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-6210 utils: Use C99 struct initializer for long_opt_start

This patch makes no functional changes.  The long_opt_start struct
initializer in lustre_lfsck.c is updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Id14efb47d29a5f16382422df6a3ad19cdf3156bf
Reviewed-on: https://review.whamcloud.com/27789
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: Fan Yong <fan.yong@intel.com>

LU-7760 ptlrpc: remove incorrect pid printing

Remove incorrect pid printing in ptlrpc_send_new_req().
It was always printing 0, and current pid is part of the message
header anyway.

Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: Ia786bc108c4fecb87d5cb16dab88253dc16ce28c
Reviewed-on: https://review.whamcloud.com/27336
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>

LU-9663 tests: replace “ls -s” with “stat -c '%s'” to get file size

This patch improves acl/2561{,_zfs}.test to use “stat -c '%s'” to
get file size instead of using “ls -s”, which prints the allocated
size of the file in blocks.

Test-Parameters: trivial \
envdefinitions=ONLY=103 \
mdtfilesystemtype=zfs ostfilesystemtype=zfs \
testlist=sanity

Change-Id: I8bfea5197102baeece244e20731c90172c5cc2f1
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: https://review.whamcloud.com/27961
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>

LU-7653 tests: replay-single/110f fails for mdts on same MDS

replay-single/110f fails when multiple mdts are present on same MDS
node. Changing/reversing the failover order of mdts in the script
replay-singl/110f helps fix this failure in such setup.

Signed-off-by: Yang Sheng <yang.sheng@intel.com>
Signed-off-by: Parinay Kondekar <parinay.kondekar@seagate.com>
Change-Id: I8b528c422f0a90dd3174a5ff6c7f8d938ed68a05
Reviewed-on: https://review.whamcloud.com/27940
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9738 kernel: kernel update RHEL7.3 [3.10.0-514.26.2.el7]

update RHEL 7.3 kernel to 3.10.0-514.26.2.el7

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I8eff2a9d0c732d97dac1df0c9233d88dbe564a4f
Reviewed-on: https://review.whamcloud.com/27934
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Tested-by: Jenkins
Reviewed-by: Minh Diep <minh.diep@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9729 lnet: correct locking in legacy add net

Make sure to unlock the api mutex properly
in lnet_dyn_add_net()

Test-Parameters: trivial
Signed-off-by: Amir Shehata <amir.shehata@intel.com>
Change-Id: I786545de690ea5966771be3e84d3561b794d55ec
Reviewed-on: https://review.whamcloud.com/27907
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Olaf Weber <olaf.weber@hpe.com>
Reviewed-by: Sonia Sharma <sonia.sharma@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9716 osc: osc_extent_tree_dump0() implementation is suboptimal

Avoid looping in osc_extent_tree_dump() if debugging is disabled.
This helps us save some cpu ticks.

Change-Id: I492429d8a6de79f67b5923895ffa58b7fe3a100d
Seagate-bug-id: MRP-4469
Signed-off-by: Andrew Perepechko <andrew.perepechko@seagate.com>
Reviewed-on: https://review.whamcloud.com/27866
Reviewed-by: Alexander Boyko <alexander.boyko@seagate.com>
Tested-by: Jenkins
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@seagate.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-6210 utils: Use C99 struct initializer in parse_opts()

This patch makes no functional changes.  The long_opt initializer in
parse_opts() is updated to C99 syntax, and renamed to long_opts for
consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Id3f1540c33cbd5db33a4a0eb69a89672eac7f713
Reviewed-on: https://review.whamcloud.com/27817
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>

LU-6210 utils: Use C99 struct initializer for opt_map

This patch makes no functional changes.  The opt_map struct
initializer in mount_lustre.c is updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I42532a6f6b3506e1893e137c054139f971241f61
Reviewed-on: https://review.whamcloud.com/27816
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>

LU-6210 utils: Use C99 struct initializer in mkfs_lustre.c

This patch makes no functional changes.  The long_opt_start
struct initializer in mkfs_lustre.c is updated to C99 syntax.
The option struct and option string have been renamed to long_opts
and short_opts for consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Icbc6be0eeaf6bf369a1b593917115378b2e3e159
Reviewed-on: https://review.whamcloud.com/27815
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>

LU-6210 utils: Use C99 struct initializer in lustre_c

This patch makes no functional changes.  The long_opts struct
initializer in lustre_rsync.c is updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Ic89ee776d5c2578759554446cdc33f5861316130
Reviewed-on: https://review.whamcloud.com/27814
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>

LU-6210 utils: Use consistent naming for option strings

This patch makes no functional changes.  The option strings in
jt_lfsck_start() and jt_lfsck_stop() are renamed to short_opts
for consistency.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Id29ef63aa2ef62223d72f31374eb0a93badc280e
Reviewed-on: https://review.whamcloud.com/27792
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Reviewed-by: Fan Yong <fan.yong@intel.com>

LU-6210 utils: Use C99 initializer for lfsck_types_names

This patch makes no functional changes.  The lfsck_types_names
struct initializer in lustre_lfsck.c is updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I8a3b44b56263b00a926550e4122193300c0c399f
Reviewed-on: https://review.whamcloud.com/27791
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: Fan Yong <fan.yong@intel.com>

LU-6210 utils: Use C99 struct initializers in lustre_fsck.c

This patch makes no functional changes.  The long_opt_stop and
long_opt_query struct initializers in lustre_lfsck.c are updated
to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Ie713904f31208447e4d5741023356aa1a37c0ea3
Reviewed-on: https://review.whamcloud.com/27790
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: Fan Yong <fan.yong@intel.com>

LU-6436 llite: NULL pointer dereference in cl_object_top()

During race condition iget5_locked() returns inode without
I_NEW bit for the second thread. If cl_file_inode_init()
failed for the first thread lli_clob isn't valid for
corresponding fid.

But failed inode is marked bad, check it.

Change-Id: I02e477486e51d2653c8b047a18c54a5528b440c5
Seagate-bug-id: MRP-4035
Signed-off-by: Andriy Skulysh <andriy.skulysh@seagate.com>
Reviewed-on: https://review.whamcloud.com/27777
Tested-by: Jenkins
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9558 lnet: kernel socket accept takes new bool agrument

During the development of the linux 4.11 kernel it was discovered
that the kernel socket layer could get into lockdep situation. To
handle this a new bool argument was added to the accept member
of struct socket. For LNet we can always pass false.

Change-Id: I420cda95b70cf927b1a6e3493b631bc5a3585d74
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/27642
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Doug Oucharek <doug@cadentcomputing.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-6210 utils: Use C99 struct initializers in lshowmount.c

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.  Variables of
type struct option are renamed to long_opts for consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lshowmount.c to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: If30db4ee6a825c326eaf246201f9173fc291a436
Reviewed-on: https://review.whamcloud.com/27532
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>

LU-6210 utils: Use C99 struct initializers in lr_reader.c

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.  Variables of
type struct option are renamed to long_opts for consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lr_reader.c to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Ie073206d6049fb63b23a30092b11672f8e7e9fd4
Reviewed-on: https://review.whamcloud.com/27531
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>

LU-6210 utils: Use C99 struct initializers in llverfs.c

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.  Variables of
type struct option are renamed to long_opts for consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates llverfs.c to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Ifa3882f435f546715c7a22c31934a9f17b1bfc01
Reviewed-on: https://review.whamcloud.com/27530
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>

LU-6210 utils: Use C99 initializers in lfs_get_ladvice()

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.  Variables of
type struct option are renamed to long_opts for consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lfs_get_ladvice() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Ida06e903a53aae37c878eebb8707b8062bef59dc
Reviewed-on: https://review.whamcloud.com/27527
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>

LU-6210 utils: Use C99 struct initializers in lfs_hsm_state()

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lfs_hsm_state() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I32f427396e99f1b86d329eb8457883a9aaaed181
Reviewed-on: https://review.whamcloud.com/27525
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>

LU-6210 utils: Use C99 struct initializers in lfs_fid2path()

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.  Variables of
type struct option are renamed to long_opts for consistency.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lfs_fid2path() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I65fb184a54aad142a87dd8f244e99a09a6482b71
Reviewed-on: https://review.whamcloud.com/27524
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>

LU-6210 utils: Use C99 struct initializers in mntdf()

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates mntdf() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: Ic4e446ecb03695b2526b0890f7bfeb76b1bfee88
Reviewed-on: https://review.whamcloud.com/27517
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>

LU-6210 utils: Use C99 struct initializers in lfs_mv()

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lfs_mv() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I81b9ace65b7f26766173199f6a14d947ccc86158
Reviewed-on: https://review.whamcloud.com/27516
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.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-6210 utils: Use C99 struct initializers in lfs_find()

This patch makes no functional changes.  Struct initializers that
use C89 or GCC-only syntax are updated to C99 syntax.

C89 positional initializers require values to be placed in the
correct order. This will cause errors if the fields of the struct
definition are reordered or fields are added or removed. C99 named
initializers avoid this problem, and also automatically clear any
values that are not explicitly set.

This patch updates lfs_find() to use the C99 syntax.

Test-Parameters: trivial
Signed-off-by: Steve Guminski <stephenx.guminski@intel.com>
Change-Id: I646e42ac5173f927c670a41cc56c13a7caa19a1e
Reviewed-on: https://review.whamcloud.com/27513
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>

LU-9620 tests: test race between pool_list and pool_remove

Test ost-pools/test_27 added to test race between
pool_list and pool_remove.

Change-Id: I8497836dd6e8b2c20ade3626b7a1f84b98f3a94f
Seagate-bug-id: MRP-3629
Signed-off-by: Jadhav Vikram <jadhav.vikram@seagate.com>
Reviewed-by: Ashish Purkar <ashish.purkar@seagate.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Reviewed-on: https://review.whamcloud.com/27505
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9620 lod: protected ost pool count updation

ASSERTION(iter->lpi_idx <= ((iter->lpi_pool)->pool_obds.op_count)
caused due to reading of ost pool count is not protected in
pool_proc_next and pool_proc_show, pool_proc_show get called when
op_count was zero.

Fix to protect ost pool count by taking lock at start sequence
function pool_proc_start and released lock in pool_proc_stop.
Rather than using down_read / up_read pairs around pool_proc_next
and pool_proc_show, this changes make sure ost pool data gets
protected throughout sequence operation.

Change-Id: I2954496d52dd703b600e69a494898c6266f026bd
Seagate-bug-id: MRP-3629
Signed-off-by: Jadhav Vikram <jadhav.vikram@seagate.com>
Reviewed-by: Ashish Purkar <ashish.purkar@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@seagate.com>
Reviewed-on: https://review.whamcloud.com/27506
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9564 build: Add server-build for Ubuntu with Kernel 4.4.0

This enables compatibility with the current LTS flavours of Ubuntu.
Do note that you need the Xenial HWE Kernel for Ubuntu 14.04.5, as
that distribution originally used a 3.x series Kernel.

The patches have been developed to apply cleanly to the kernel versions
4.4.0-45.66 to 4.4.0-85.108 from the Ubuntu Xenial (and its Trusty backports).

This change also adjusts the Debian scripting to produce the
ldiskfs modules and the server utilities.

To create the server modules run "./configure" with "--enable-server"
and specify "--enable-ldiskfs" and "--with-zfs/-spl" as appropriate.

The call to "make debs" will then produce the server modules and
utils instead of their client versions.

Test-Parameters: trivial
Signed-off-by: Martin Schroeder <martin.h.schroeder@intel.com>
Change-Id: Ia78702da304f735bb932738784f1346be0f39026
Reviewed-on: https://review.whamcloud.com/27323
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Gu Zheng <gzheng@ddn.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Thomas Stibor <t.stibor@gsi.de>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-282 tests: remove extra logging from sanity 27

Remove extra logging from sanity.sh test_27b to avoid confusing
autotest log parsing.

Replace some uses of $SETSTRIPE and $GETSTRIPE with $LFS in these
functions since this was only needed during the ancient transition
from the standalone "lstripe" binary.

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I270d2f75cd803bede5776117c9d5aaaa5b3ebbe5
Reviewed-on: https://review.whamcloud.com/27322
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Steve Guminski <stephenx.guminski@intel.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9509 tests: fix compilebench timeout

With ext4 and lustre found compilebench always takes more time in
create dir operation whenever previous same test aborted without
deleting directory. As per test board result, found compilebench
test ran from suite parallel-scale-nfs.sh timedout before running
from suite parallel-scale.sh, so 1st timedout compilebench not
clear data, rerunning compilebench again causes timeout.

Fix is creating different directory on each run of compilebench.

Seagate-bug-id: MRP-3753
Signed-off-by: Jadhav Vikram <jadhav.vikram@seagate.com>
Reviewed-by: Alexander Lezhoev <alexander.lezhoev@seagate.com>
Tested-by: Alexander Lezhoev <alexander.lezhoev@seagate.com>
Reviewed-by: Ashish Purkar <ashish.purkar@seagate.com>
Change-Id: I9b47e28fe5c638f4167c05fb68e46ec5e90de09e
Reviewed-on: https://review.whamcloud.com/27131
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>

LU-9511 utils: fix parallel_grouplock test timeout

In parallel_grouplock/subtest12, when process 2 acquire
grouplock before writing by process 0 there is deadlock
situation.
- Process 0 write will wait till process 2 release GR lock
  (gid=1).
- Process 0 will wait for process 2 to get its lock after write
  completed by process 0 using MPI_RECV and here write already
  waiting for process 2 to release grouplock.
- Process 2 release lock when process 0 notify process 2.
  This issue occured once on local setup, analysed issue through
  coredump and concluded following solution which helps to
  resolve this timeout.

Increased sleep time from microseconds to seconds for process 1
and 2 to make sure process 0 starts writing and get PW lock.

Seagate-bug-id: MRP-3751
Signed-off-by: Jadhav Vikram <jadhav.vikram@seagate.com>
Reviewed-by: Alexander Lezhoev <alexander.lezhoev@seagate.com>
Tested-by: Alexander Lezhoev <alexander.lezhoev@seagate.com>
Reviewed-by: Ashish Purkar <ashish.purkar@seagate.com>
Change-Id: I1100b86dba55b2eb01c9ae61ba8f1c8e376090c8
Reviewed-on: https://review.whamcloud.com/27127
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9345 tests: use hsm_remove with --mntpath for deleted files

In test_29d of sanity-hsm, to run "lfs hsm_remove" on a file deleted
from Lustre, one has to use the --mntpath option.

Test-Parameters: trivial testlist=sanity-hsm clientcount=3 envdefinitions="ONLY=29d"
Signed-off-by: Quentin Bouget <quentin.bouget@cea.fr>
Change-Id: I35865c059e498e1a0ced0cebeac22a8491231e00
Reviewed-on: https://review.whamcloud.com/27006
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9545 lfsck: report "inconsistent" under dryrun mode

It is confused to report the item as "fixed" under dryrun
mode LFSCK. Instead, report them as "inconsistent".

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I22e056d7143a55e0dc06d9a891f4126522b466c9
Reviewed-on: https://review.whamcloud.com/27606
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>

LU-9153 llog: consolidate common error checking

Consolidate error checking for corrupted and empty llog files
detected when the llog file is being opened into a single
routine llog_cat_process_common().

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I575343f486e7420c51466f9923b841c37f3ebbe5
Reviewed-on: https://review.whamcloud.com/26211
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Faccini Bruno <bruno.faccini@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9220 gss: support Kerberos auth from unprivileged container

When a container runs unprivileged, it cannot get access to /proc.
So, to be able to do the required ioctl to
/proc/fs/lustre/sptlrpc/gss/init_channel in order to negotiate
credentials, delegate this ioctl to a parent thread that does not run
in the container's namespace.
The rest of the authentication process is still carried out in the
container's namespace if relevant.

Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I445ccc82d9b853775bc0de73323699638f688dab
Reviewed-on: https://review.whamcloud.com/26035
Tested-by: Jenkins
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Kit Westneat <kit.westneat@gmail.com>
Reviewed-by: Jeremy Filizetti <jeremy.filizetti@gmail.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9219 tests: add missing mgs reformat to conf-sanity/56

conf-sanity/test_56 timedout while mounting client
the reason of timeout is mounting mds failed with -EADDRINUSE.
Registering mdt to MGS while mounting mds index of server was
already present in MGS configuration database so mds mount
failed with -EADDRINUSE and further test stuck while client
mount so test timedout.

This change adds missing mgs reformat when mgs and mds
separate before starting mdt and osts, This will make sure
index of mdt will not be present in MGS config database.

Seagate-bug-id: MRP-3806
Signed-off-by: Jadhav Vikram <jadhav.vikram@seagate.com>
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@seagate.com>
Reviewed-by: Alexandr Boyko <alexander.boyko@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Change-Id: Ie7d897534197af7e01d92d29613123a0290ffc4c
Reviewed-on: https://review.whamcloud.com/26029
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Alexander Boyko <alexander.boyko@seagate.com>
Tested-by: Jenkins
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9183 llite: add support set_acl method in inode operations

Linux kernel v3.14 adds set_acl method to inode operations.
See kernel commit 893d46e443346370cd4ea81d9d35f72952c62a37

Change-Id: Ia40d55364016fafa8633fdaecd317910505f8ad4
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: https://review.whamcloud.com/25965
Reviewed-by: Bob Glossman <bob.glossman@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-9183 llite: handle xattr with the xattr_handler infrastructure

In commit fd50ecaddf8372a1d96e0daeaac0f93cf04e4d42 for the linux
4.9 kernel the {get,set,remove}xattr inode operations were removed
and all xattr operations are now handled by xattr_handlers. For
the upsteam lustre client a port was already done with:

Linux-commit: 1e1f9ff406fd5f6003a5dab2ab5a26c4c5bb8cbd
Linux-commit: 2c563880ea8fdc900693ae372fa07b3894f8ff63

This patch brings this work to the OpenSFS/Intel branch. The
difference is that we also have to support older kernels which
means we need to handle the following changes to the
struct xattr_handler:

Linux-commit: e409de992e3ea3674393465f07cc71c948edd87a
Linux-commit: b296821a7c42fa58baa17513b2b7b30ae66f3336

Lastly the xattr_handler api for RHEL6 is too old for proper
support, lacks the flags in struct xattr_handler. Since this
is the case we have to carry around the pre xattr handler
code. Once RHEL6 support is dropped we can remove that code.

Test-Parameters: testgroup=review-ldiskfs
Test-Parameters: testgroup=review-zfs-part-1
Test-Parameters: testgroup=review-zfs-part-2
Test-Parameters: testgroup=review-dne-part-1
Test-Parameters: testgroup=review-dne-part-2
Test-Parameters: trivial clientselinux testlist=sanity-selinux

Change-Id: I7bdeb57c09f4a252f61737dbdbfa76939df7b5eb
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/27240
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9104 obd: Ignore unknown config param while mounting

class_process_proc_param() returns positive value when it encounters
unknown parameters in order to have the below levels process them.
At the very bottom layer the positive value returned by
class_process_proc_param() needs to be dropped.osd_process_config()
missed that, which resulted in target mount failure.

Make sure that osd_process_config() does not return positive value
returned by class_process_proc_param().

Test case has been added to check processing of unknown config
parameters.

Test-Parameters: testlist=conf-sanity
Signed-off-by: Parinay Kondekar <parinay.kondekar@seagate.com>
Signed-off-by: Rahul Deshmukh <rahul.deshmukh@seagate.com>
Seagate-bug-id: MRP-4162
Change-Id: I068b8f2aee4cee69629efc83745d7cb88aea268c
Reviewed-on: https://review.whamcloud.com/25368
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>

LU-9750 nrs: some code cleanup in NRS policies

This patch does some code cleanup in NRS CRR and ORR polices,
including:
- remove the useless NULL checks
- handle errors properly by multiple labels instead of a single one

Signed-off-by: Emoly Liu <emoly.liu@intel.com>
Change-Id: Iafe86ac94042547e83c69e4b46ff7bf1ca31f073
Reviewed-on: https://review.whamcloud.com/25319
Tested-by: Jenkins
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9184 ldlm: selective IBITS lock trying

There is code to get lock without blocking, e.g. 'try'
to get it. It enqueues lock with all wanted bits and
get it only when all bits are not conflicting.
The new code allows to get as much 'tried' bits as
possible. It also doesn't require extra locking attempt
for that.
Patch intoduces new inodebits field try_bits which contains
optinal bits to get if possible without blocking. Those
bits are checked along with the required bits and only
non-blocked are added to the new IBITS lock.

Signed-off-by: Mikhal Pershin <mike.pershin@intel.com>
Change-Id: I78d181243556e65d3bd702d332307d0723d5b665
Reviewed-on: https://review.whamcloud.com/25262
Tested-by: Jenkins
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8066 ptlrpc: migrate ptlrpc proc files to sysfs

move procfs high_priority_ratio file to sysfs

Move ptlrpc service high_priority_ratio entry from procfs to sysfs.
Currently in use only by ldlm callback service only in
/sys/fs/lustre/ldlm/services/ldlm_cbd

Linux-commit: b40881e579588360242bd3f8f79f978d2a871e0e

move procfs threads* files to sysfs

Move ptlrpc service threads_min, threads_max and threads_running
entries from procfs to sysfs. Currently in use only by ldlm
callback service only in /sys/fs/lustre/ldlm/services/ldlm_cbd

Linux-commit: 673a6796f290fe8079af6a688f20c87e7416bba5

Add infrastructure for sysfs migration

Added necessary plumbing for ptlrpc sysfs integration for registered
services, sysfs directory registration.

Linux-commit: 328676f823b49741675f8ae47643669a31c79558

Add support for server side for this change.

Change-Id: Ia82a11dabd672a0c36083d06303aff2a81f3c48f
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-on: https://review.whamcloud.com/25073
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>

LU-8066 obdclass : Add infrastructure for procfs to sysfs migration

Prepare for procfs to sysfs migration

Add necessary plumbing to register obd types and instances
under /sys/fs/lustre

Linux-commit : 9b8013023cb62360b56c04313687e93a1c2bf3d6

Move common obd proc files to sysfs

This moves uuid display and also underlying fs statistics.

Linux-commit : 0bc36cb06e4bc5e936f2f442535daf557c7a4200

Change-Id: Ieb4e2f60c7003d658126beac071c05f94fa8eab8
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-on: https://review.whamcloud.com/26020
Tested-by: Jenkins
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>

LU-8703 libcfs: change CPT estimate algorithm

The main idea to have more CPU partitions is based on KNL experience.
When a thread submit IO for network communication one of threads from
current CPT is used for network stack. Whith high parallelization many
threads become involved in network submission but having less CPU
partitions they will wait until single thread process them from network
queue. So, the bottleneck just moves into network layer in case of
small amount of CPU partitions. My experiments showed that the best
performance was when for each IO thread we have one network thread.
This condition can be provided having 2 real HW cores (without hyper
threads) per CPT. This is exactly what implemented in this patch.

Change CPT estimate algorithm from 2 * (N - 1)^2 < NCPUS <= 2 * N^2
to 2 HW cores per CPT. This is critical for machines with number of
cores different from 2^N.

Current algorithm splits CPTs in KNL:
LNet: HW CPU cores: 272, npartitions: 16
cpu_partition_table=
0       : 0-4,68-71,136-139,204-207
1       : 5-9,73-76,141-144,209-212
2       : 10-14,78-81,146-149,214-217
3       : 15-17,72,77,83-85,140,145,151-153,208,219-221
4       : 18-21,82,86-88,150,154-156,213,218,222-224
5       : 22-26,90-93,158-161,226-229
6       : 27-31,95-98,163-166,231-234
7       : 32-35,89,100-103,168-171,236-239
8       : 36-38,94,99,104-105,157,162,167,172-173,225,230,235,240-241
9       : 39-43,107-110,175-178,243-246
10      : 44-48,112-115,180-183,248-251
11      : 49-51,106,111,117-119,174,179,185-187,242,253-255
12      : 52-55,116,120-122,184,188-190,247,252,256-258
13      : 56-60,124-127,192-195,260-263
14      : 61-65,129-132,197-200,265-268
15      : 66-67,123,128,133-135,191,196,201-203,259,264,269-271

New algorithm will split CPTs in KNL:
LNet: HW CPU cores: 272, npartitions: 34
cpu_partition_table=
0       : 0-1,68-69,136-137,204-205
1       : 2-3,70-71,138-139,206-207
2       : 4-5,72-73,140-141,208-209
3       : 6-7,74-75,142-143,210-211
4       : 8-9,76-77,144-145,212-213
5       : 10-11,78-79,146-147,214-215
6       : 12-13,80-81,148-149,216-217
7       : 14-15,82-83,150-151,218-219
8       : 16-17,84-85,152-153,220-221
9       : 18-19,86-87,154-155,222-223
10      : 20-21,88-89,156-157,224-225
11      : 22-23,90-91,158-159,226-227
12      : 24-25,92-93,160-161,228-229
13      : 26-27,94-95,162-163,230-231
14      : 28-29,96-97,164-165,232-233
15      : 30-31,98-99,166-167,234-235
16      : 32-33,100-101,168-169,236-237
17      : 34-35,102-103,170-171,238-239
18      : 36-37,104-105,172-173,240-241
19      : 38-39,106-107,174-175,242-243
20      : 40-41,108-109,176-177,244-245
21      : 42-43,110-111,178-179,246-247
22      : 44-45,112-113,180-181,248-249
23      : 46-47,114-115,182-183,250-251
24      : 48-49,116-117,184-185,252-253
25      : 50-51,118-119,186-187,254-255
26      : 52-53,120-121,188-189,256-257
27      : 54-55,122-123,190-191,258-259
28      : 56-57,124-125,192-193,260-261
29      : 58-59,126-127,194-195,262-263
30      : 60-61,128-129,196-197,264-265
31      : 62-63,130-131,198-199,266-267
32      : 64-65,132-133,200-201,268-269
33      : 66-67,134-135,202-203,270-271

'N' pattern in KNL works is not always good.
in flat mode it will be one CPT with all CPUs inside.

in SNC-4 mode:
cpu_partition_table=
0       : 0-17,68-85,136-153,204-221
1       : 18-35,86-103,154-171,222-239
2       : 36-51,104-119,172-187,240-255
3       : 52-67,120-135,188-203,256-271

Change-Id: I07d1fa2e490dd5720497d438027f128df5f01773
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: https://review.whamcloud.com/24304
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>

LU-6671 wireshark: Process mds_md entries

In LDLM_ENQUEUE and MDS_GETATTR responses, process the v1 and v3
versions of the lov_mds_md structure.

Test-Parameters: trivial
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: I7bc75d6ae0eb7fe14c805bd0488369dd605521eb
Reviewed-on: https://review.whamcloud.com/18061
Reviewed-by: Doug Oucharek <doug@cadentcomputing.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9725 lwp: wait on deregister

When lustre_deregister_lwp_item() is being called, it should wait
for the inflight callback done before moving on to free the data
used by the callback.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I9c27a0ae4c765147fd183b78bf3693a66e7511dc
Reviewed-on: https://review.whamcloud.com/27987
Tested-by: Jenkins
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9753 ofd: 64-bits diff variable to avoid overflow

In ofd_create_hdl(), the logic will compare the OST stroed LAST_ID
with the MDT given one: if the difference exceeds some threshold,
then it will trust the OST LAST_ID directly and reset the MDT side
value with the OST one. Otherwise, the orphan OST-objects will be
destroyed.

Unfortunately, both the OST stored LAST_ID and MDT given one are
64 bits, but the @diff variable is only 32 bits, and if the OST
side value is too larger than the MDT side, then the @diff will
overflow. That will misguide the OST to destroy useful OST-objects
by wrong. This patch change the @diff as 64 bits variable.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: If75899cbab5754be4ede226e0463ba5f69d70e3d
Reviewed-on: https://review.whamcloud.com/27975
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9740 ldiskfs: more credits for non-append write

As code comment explained: for not append write, the split
may need to modify existing blocks moving entries into the
new ones. That needs more journal credits. The old logic
handled it as append case by typo.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I277f144ec056bb2f07ffd5e5ce19d9a6eee8e0ef
Reviewed-on: https://review.whamcloud.com/27947
Tested-by: Jenkins
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-9657 llapi: check if the file layout is composite

Add a new API llapi_layout_is_composite() to check if the file
layout is composite.

Signed-off-by: Emoly Liu <emoly.liu@intel.com>
Change-Id: Ie7d60f95feb9dfde5fbf52ea65ebd26b925bd33d
Reviewed-on: https://review.whamcloud.com/27752
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-3719 ldiskfs: adjust s_mb_prealloc_table_size correctly

When mb prealloc table item value is not valid, the
s_mb_prealloc_table_size should be adjust accordingly.
Port patch to rhel7 & sles12.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Alexander Boyko <alexander_boyko@xyratex.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Change-Id: I5aa3a32a6a3aedc70160409c6443746fd2ccbbc9
Reviewed-on: https://review.whamcloud.com/27748
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Ben Evans <bevans@cray.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-5106 readdir: improve striped readdir

Striped directory needs to build its directory page from sub-stripe
directory pages, current code iterate sub stripe directory pages
to search dirent for each hash, this is inefficient, as may cause
statahead fail because statahead thread is slow in readdir, while
'ls' is faster and can't find cached statahead entries, and finally
cause statahead fail.

This patch introduces a struct lmv_dir_ctxt which saves dir page and
current dirent for all stripes, to find the dirent which has the
closest hash value it only needs to compare the dirent of all stripes,
and then pop this dirent from its stripe, until all stripes reache
the end.

This patch contains another fix: previously LDP_COLLIDE is set by
default, change to set dir page end hash 'ldp_hash_end' to hash
of next dirent, and only set LDP_COLLIDE when 'ldp_hash_end' equals
last dirent hash 'lde_hash'. We should avoid dir hash collision.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: I99c33ea3c55772d3bb77571dcdc67312d386fa27
Reviewed-on: https://review.whamcloud.com/27663
Tested-by: Jenkins
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-5106 statahead: support striped directory

Current statahead will getattr by fid, but for striped directory,
it needs to fix its parent fid to stripe fid, which needs file
name, so pack file name in md_op_data.

Remove check for remote dir check in lmv_intent_getattr_async(),
which is not necessary, and not checking it won't cause anything
wrong, while the code is simpler.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: I275d866cb1ea157ef2b594ded0cc342264f0382c
Reviewed-on: https://review.whamcloud.com/27683
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: wangdi <di.wang@intel.com>

LU-6142 selftest: remove typedefs from lnet selftest

Change the typedefs in lnet selftest to proper structures.
Several other style changes to fix checkpatch issues with
code impacted by tyepdef change.

Test-Parameters: testlist=lnet-selftest

Change-Id: I941d460566673ccb5bccad9f3506f90ede47ed67
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/27298
Reviewed-by: Doug Oucharek <doug@cadentcomputing.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Olaf Weber <olaf.weber@hpe.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9442 osp: can't create IDIF fid number > 0xFFFFFFFF

fid_is_last_id didn't recognize IDIFes. This caused OST
to allocate new sequence despite of MDT still used
initial FID_SEQ_IDIF. Finally allocation failed with -115
returned from osd_check_lma:
osd_check_lma()) lustre-OST0000-osd: FID [0x100000001:0x0:0x0] != self_fid [0x100000000:0x0:0x0]

Patch has several typical "IDIF" fixes. Also it has lov_objid
fix to store all 48 IDIF bits instead of 32.
Finally it changes union fields order in ost_id.
Before the fix oi_fid.f_seq addressed oi_id instead of oi_seq.

Change-Id: Ifbda97a5b228254aedcb050c3d94d2ecb3a9590c
Seagate-bug-id: MRP-4392
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@seagate.com>
Reviewed-on: https://review.whamcloud.com/27225
Tested-by: Jenkins
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9505 llapi: treat MDT index as a hex number

Since MDT index is a hex number, "base" in strtol() should be 16.

Signed-off-by: Emoly Liu <emoly.liu@intel.com>
Change-Id: I50922f9eb5d1095f06a493628ef521d34969a59f
Reviewed-on: https://review.whamcloud.com/27156
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9417 mdc: excessive memory consumption by the xattr cache

The refill operation of the xattr cache does not know the
reply size in advance, so it makes a guess based on
the maxeasize value returned by the MDS.

In practice, it allocates 16 KiB for the common case and
4 MiB for the large xattr case. However, a typical reply
is just a few hundred bytes. Even worse, RHEL6 has a very
bad vmalloc()/vfree() design with global locks.

If we follow the conservative approach, we can prepare a
single memory page for the reply. It is large enough for
any reasonable xattr set and, at the same time, it does
not require multiple page memory reclaim, which can be
costly.

If, for a specific file, the reply is larger than a single
page, the client is prepared to handle that and will fall back
to non-cached xattr code. Indeed, if this happens often and
xattrs are often used to store large values, it makes sense to
disable the xattr cache at all since it wasn't designed for
such [mis]use.

Change-Id: I98d97ffea5c83bccbc9f254717af8d2c0ac4d77f
Signed-off-by: Andrew Perepechko <andrew.perepechko@seagate.com>
Reviewed-on: https://review.whamcloud.com/26887
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Ben Evans <bevans@cray.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8062 ptlrpc: increase sleep time in ptlrpc_request_bufs_pack()

schedule_timeout() does not necessarily expire. Increase the sleeping
time in ptlrpc_request_bufs_pack() as 2 seconds is too short, given
the 1 second sleep used for recovery-small test_115_write().

Test-Parameters: envdefinitions=ONLY=115 testlist=recovery-small,recovery-small,recovery-small,recovery-small
Test-Parameters: envdefinitions=ONLY=115 testlist=recovery-small,recovery-small,recovery-small,recovery-small
Test-Parameters: envdefinitions=ONLY=115 testlist=recovery-small,recovery-small,recovery-small,recovery-small
Test-Parameters: envdefinitions=ONLY=115 testlist=recovery-small,recovery-small,recovery-small,recovery-small
Signed-off-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: If1b6b147095e01663c0fa9a5b8f93c445d2061ee
Reviewed-on: https://review.whamcloud.com/26815
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>

LU-8885 tests: zconf_mount_clients() defect

Presently zconf_mount_clients() returns success(0)
in cases where the 'NFS' mount is already mounted
on a mount point which it is trying to mount again.
In this case, it silently ignores the mount, leading
to testcase failure. This patch addresses this defect
by allowing zconf_mount_clients() to compare that
nothing unexpected is mounted by comparing the result
of mount count with mount count of "type lustre". If
they are unequal the function exists with an error.

Test-Parameters: envdefinitions=PARALLEL_SCALE_EXCEPT=parallel_grouplock \
testlist=sanity,parallel-scale,parallel-scale-nfsv3
Signed-off-by: Arshad Hussain <arshad.hussain@seagate.com>
Change-Id: I55e4b2ef2a18985be4833fca017cc6c6b0c5410f
Seagate-bug-id: MRP-3773
Reviewed-by: Ashish Purkar <ashish.purkar@seagate.com>
Reviewed-by: Ujjwal Lanjewar <ujjwal.lanjewar@seagate.com>
Reviewed-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Reviewed-on: https://review.whamcloud.com/24054
Reviewed-by: Elena Gryaznova <elena.gryaznova@seagate.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-3285 osc: prepare OSC code to be used from MDC

Move osc/osc_cl_include.h to include/lustre_osc.h
to be included from both OSC and MDC.

Change-Id: Ic04995d8793cd64b998463aee59c6bf30445b64b
Signed-off-by: Mikhal Pershin <mike.pershin@intel.com>
Reviewed-on: https://review.whamcloud.com/22996
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-9671 nodemap: restore client's IDs for OST_WRITE

Client sets overquota flags for certain UID/GID based on the
IDs & flags in OST_WRITE reply, so we need to reply client IDs
instead of mapped IDs.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I375847fa734237f9bcea10fa676e09c471a0fcfb
Reviewed-on: https://review.whamcloud.com/27680
Tested-by: Jenkins
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Stephan Thiell <sthiell@stanford.edu>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9654 mdt: fix problem of RAoLU HSM policy

mdt_attr_get_complex() clears all known attributes even they are
already valid. So in mdt_handle_last_unlink(), the valid attributes
of HSM or nlink should be checked before calling that function.

Signed-off-by: Li Xi <lixi@ddn.com>
Change-Id: I9ba561cadcc40baf5e28172cfda699cdecce7ea8
Reviewed-on: https://review.whamcloud.com/27564
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Faccini Bruno <bruno.faccini@intel.com>
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7631 tests: wait_osts_up waits for MDS precreates

Fix wait_osts_up() to wait for the MDS to finish orphan cleanup and
precreate some OST objects so that there isn't a race to get all of
the OSTs available for conf-sanity test_82a.

Test-Parameters: trivial testlist=replay-single,replay-single
Test-Parameters: testlist=conf-sanity,conf-sanity,conf-sanity
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I534f0a1f36c3d00f702684041bfa991a4a3ebbe5
Reviewed-on: https://review.whamcloud.com/27441
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9514 ptlrpc: free reply buffer earlier for open RPC

It is unnecessary to keep the reply buffer for open RPC. Replay
related data has already been saved in the request buffer when
the RPC replied. If the open replay really happen, the replay
logic will alloc the reply buffer when needed.

On the other hand, the client always tries to alloc big enough
space to hold the open RPC reply since the client does not exactly
know how much data the server will reply to the client. So the reply
buffer may be quite larger than the real needed. Under such case,
keeping the large reply buffer for the open RPC will occupy a lot
of RAM as to OOM if the are too many open RPCs to be replayed.

This patch frees the reply buffer for the open RPC when only
the replay logic holds the last reference of the RPC.

Test-Parameters: envdefinitions=ONLY=51f testlist=sanity ostfilesystemtype=ldiskfs mdtfilesystemtype=ldiskfs
Test-Parameters: envdefinitions=ONLY=51f testlist=sanity ostfilesystemtype=zfs mdtfilesystemtype=zfs
Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I1bea2456b8aa4e53a0b65143a48e617f836181a0
Reviewed-on: https://review.whamcloud.com/27208
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-7988 hsm: mark the cdt as stopped when its thread exits

Use kthread_stop() to stop and join the coordinator thread. Only after
that step can the coordinator state be set to CDT_STOPPED. As a
coordinator doesn't stop instantly, this closes a race if the
coordinator is being restarted at the same time, leaving one thread
shutting down and a new one starting up.

Test-Parameters: trivial testlist=sanity-hsm
Signed-off-by: frank zago <fzago@cray.com>
Change-Id: I0a21d0d22403a56a8965441e1b57118073b6f210
Signed-off-by: Ben Evans <bevans@cray.com>
Reviewed-on: https://review.whamcloud.com/22762
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Quentin Bouget <quentin.bouget@cea.fr>
Reviewed-by: Faccini Bruno <bruno.faccini@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9274 ptlrpc: add replay request into unreplied list

ptlrpc_prepare_replay() may fail to add replay request into unreplied
list if the request hasn't been on replay list yet, so in
ptlrpc_replay_next() before sending replay, we'd always make sure the
replay request is on unreplied list.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I91757cd4fde1d85d146475e078db125acc2c821f
Reviewed-on: https://review.whamcloud.com/27920
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9305 osd: do not release pages twice

in case of blocksize mismatch dmu_assign_arcbuf() releases passed
abuf internally, including the pages. osd_bufs_put() can't detect
this and may call __free_page() on inappropriate pages (which can
be allocated to someone else already).

Change-Id: I454e56ee3de3d201a14e6ba7b4beabaad42d82ae
Signed-off-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-on: https://review.whamcloud.com/27950
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9693 tests: Adding sanity test_42a & 42c to always_except

Sanity test_42a and test_42c were removed from Always_Except
list earlier. But it seems the issues were not completely
resolved.

Hence, adding them back to Always_Except list.

Test-Parameters: trivial
Signed-off-by: Saurabh Tandan <saurabh.tandan@intel.com>
Change-Id: Ib82e0d788054f3ede58c2dbdf5af21227fb4e7f3
Reviewed-on: https://review.whamcloud.com/27740
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Wei Liu <wei3.liu@intel.com>

LU-9715 libcfs: crash in cpu_pattern parsing code

The for loop in cfs_cpt_table_create_pattern() that scans
for brackets to count the number of cpts is broken. It will
increment bracket beyond NULL and it will increment ncpt
beyond the number of available cpts. This has been fixed.

Test-Parameters: trivial
Signed-off-by: Amir Shehata <amir.shehata@intel.com>
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I87bc9de4c531c42c421e8e62edd881417dbcab07
Reviewed-on: https://review.whamcloud.com/27872
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Olaf Weber <olaf.weber@hpe.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>

Branch of 2.11 development branch new tag 2.10.50

Change-Id: Iea32c681312885f577421e1bf194c569a6b0da65
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9073 gss: remove newer kernel support

Revert the work to support newer kernels for GSS. For now
disable GSS support for kernels newer than 4.6 so this
doesn't block people on newer distros. Even is very basic
support for GSS is restored I wouldn't recommend this for
production systems at this time.

Change-Id: I7e1636bf695e1686bbdf968d088fcfc5a8f8f062
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/27823
Tested-by: Jenkins
Reviewed-by: Chris Hanna <hannac@iu.edu>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9678 osd-zfs: register arc_prune_func() after site init

Register arc_prune_func() after site init, otherwise it may be
called while object cache is not initialized yet.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: I822252da906f03899386fb0941cc11c1c3366fbf
Reviewed-on: https://review.whamcloud.com/27708
Tested-by: Jenkins
Reviewed-by: Fan Yong <fan.yong@intel.com>
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-9049 obdclass: change object lookup to no wait mode

Currently we set LU_OBJECT_HEARD_BANSHEE on object when we want
to remove object from cache, but this may lead to deadlock, because
when other process lookup such object, it needs to wait for this
object until release (done at last refcount put), while that process
maybe already hold an LDLM lock.

Now that current code can handle dying object correctly, we can just
return such object in lookup, thus the above deadlock can be avoided.

There is another case we need to make some changes:
objects created in OUT doesn't set dt_body_ops for LOD layer because
originally it's set by lod_create(), change to set dt_body_ops in
lod_object_init() so objects created in OUT is no different from
those created in MDT. To achieve this, functions in lod_body_ops
should check file type inside to avoid misuse.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: Ia31ab5f09f9bf80a9aa8fd7e7b60348b02400b25
Reviewed-on: https://review.whamcloud.com/26965
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Tested-by: Cliff White <cliff.white@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9485 llite: check the return value of cl_file_inode_init()

ll_update_inode() does not check the return value of
cl_file_inode_init(), and it should check.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: I4174e4f8166d7834a1d619aa8d0191d1f428c62c
Reviewed-on: https://review.whamcloud.com/27658
Tested-by: Jenkins
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-8703 libcfs: rework CPU pattern parsing code

Rewrite CPU pattern parsing code to avoid passed buffer change
add real errors propogation to caller function.

Change-Id: I8dfee2c0013fcfccd3d99c361d3ec626594689bd
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: https://review.whamcloud.com/23306
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: Patrick Farrell <paf@cray.com>
Reviewed-by: Olaf Weber <olaf.weber@hpe.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9597 test: wait setattr finished before checking accounting

Need wait setattr finished for ost objects, otherwise, accounting
for projects won't be accurate, also add failure check for setattr
ioctl.

Test-Parameters: testlist=sanity-quota,sanity-quota,sanity-quota,sanity-quota clientdistro=el7 serverdistro=el7 \
        ostfilesystemtype=ldiskfs mdtfilesystemtype=ldiskfs
Change-Id: I106689c224997f79eb779fdc6843704ae7e9ffe6
Signed-off-by: Wang Shilong <wshilong@ddn.com>
Reviewed-on: https://review.whamcloud.com/27425
Tested-by: Jenkins
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Yingjin Qian <qian@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9558 llite: use struct vma_area_struct address field

The field virtual_address of struct vma_area_struct was
removed since it provided no better benefit than just
using the address field directly.

Linux-commit: 1a29d85eb0f19b7d8271923d8917d7b4f5540b3e

Change-Id: I05068cdf27c93c5b3201c76ec043bc6c0e66df1f
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/27281
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-9210 statahead: missing barrier before wake_up

A barrier is missing before wake_up() in ll_statahead_interpret(),
which may cause 'ls' hang.

Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I391d6222e353fb27761ffd5412b52ce08f7a0592
Reviewed-on: https://review.whamcloud.com/27330
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>