LU-11736 utils: don't set max_sectors_kb on MDT/MGT

The max_sectors_kb tunable should not be applied to MDT and MGT
devices. This tuning is needed for efficiency of large IOs for
spinning disks, but is not needed for SSDs or regular IO. It can
cause problems with DM Multipath configurations for minimal
benefits, so should be limited to OST devices.

This only applies to ldiskfs backend filesystems, no such tuning
is currently done for any ZFS devices.

Lustre-change: https://review.whamcloud.com/33796
Lustre-commit: 2f8d7b4679de3fa467040aa61733f262714e39c9

Test-Parameters: trivial testlist=conf-sanity
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I496603da13aae042f63cc37c0dea221a393ebbe5
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: Gu Zheng <gzheng@ddn.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34311
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-7631 tests: add debug info to conf-sanity 82a

In the routine check_stripe_count, the different error
messages need to be modified so when an error occurs,
a user can tell what error was hit. Also, print precreated
object information at the beginning of the test and on
error.

Test-Parameters: trivial mdscount=2 mdtcount=4 testlist=conf-sanity envdefinitions=ONLY=82a
Test-Parameters: mdscount=1 mdtcount=1 testlist=conf-sanity envdefinitions=ONLY=82a

Lustre-change: https://review.whamcloud.com/33689
Lustre-commit: e76683a5bd540cacd2271a969aa9acd9bf790ccf

Signed-off-by: James Nunez <jnunez@whamcloud.com>
Change-Id: Ifc75d52d38d9cb401118ef7baa4014bddf6298f2
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Emoly Liu <emoly@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34294
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-11757 lod: use calculated stripe count

lod_prep_md_striped_create() tries to allocat big
chunk of memory because
lum->lum_stripe_count == -1 and is converted to __u32.

ldo_dir_stripe_count was calculated already in lod_ah_init()

Lustre-change: https://review.whamcloud.com/33829
Lustre-commit: 622a94d5e27ed3e596918863c08b304a6be9a646

Change-Id: Id99d9e024638dfb1b34262840d2e543c808a9cdc
Cray-bug-id: LUS-6694
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Alexander Boyko <c17825@cray.com>
Tested-by: Elena Gryaznova <c17455@cray.com>
Reviewed-by: Vladimir Saveliev <c17830@cray.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34313
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-11737 lfsck: do not ignore dryrun

lfsck_layout_recreate_lovea() shouldn't ignore dryrun.

Lustre-change: https://review.whamcloud.com/33826
Lustre-commit: 875f3fc03aa15049892fe19d6a4fc1132848fced

Change-Id: Ia8bafc13f148b03573dee5db26b6aff9386b5b5f
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34312
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11625 ofd: handle upgraded filter_fid properly

Since there have been several iterations of struct filter_fid stored
on disk, the current code wasn't checking for all of the possible
cases when trying to decide what action to take when accessing and
upgrading the xattr for new capabilities.

Properly check for the various different struct filter_fid sizes and
handle them appropriately.  Add a more verbose description of the
various cases so that this is more clear to others in the future.

Add decoding of filter_fid fields added for FLR in 2.11.

We should already be testing for upgrading the filter_fid xattr
from different OST versions in conf-sanity test_32d.

Lustre-change: https://review.whamcloud.com/33627
Lustre-commit: 381a8cdd527ce4deccfc3f7eb461892f6f2f3fff

Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ifef2292296236cb06ff7e8cd50caff4b133ebbe5
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34304
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11620 lfsck: change llsd_rb_lock to rwsemaphore

llsd_rb_lock is taken in ->init, and released in ->fini, but during
this period it may getxattr which can sleep. Change it to rwsemaphore.

Lustre-change: https://review.whamcloud.com/33603
Lustre-commit: 925ce153979d6ac793a65e193181ec14a8281640

Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: Idc68eb886e60dc45ccfb7ac9bf5bf06db42d690d
Reviewed-by: Alexandr Boyko <c17825@cray.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34303
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-11712 osd-ldiskfs: Wrap blk integrity with config check

Build is currently broken for kernels without
CONFIG_BLK_DEV_INTEGRITY. Build failure introduced by LU-11096
commit c8505c2e70d03ba20edf9fcbf431888e87a21147
https://review.whamcloud.com/#/c/32725/
Use of blk integrity should be wrapped in the config check for
HAVE_BLK_INTEGRITY_ENABLED

Lustre-change: https://review.whamcloud.com/33745
Lustre-commit: fd193758bb95e3fbb4cd04e88f0d964f9cb510cf

Signed-off-by: Chris Horn <hornc@cray.com>
Change-Id: Iac9e4a2572024c026132c87c11042cf353b14d48
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34309
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11878 tests: don't fork-bomb sanity test_103b

Running sanity test_103b may start up to 512 parallel threads for
running the test, each of which starts two bash processes and lfs
or rm processes.

For the VMs running in our testbed (esp. ARM with 64KB PAGE_SIZE)
this can trigger the OOM killer and cause the test to fail if bash
is killed.  Limit the number of started bash processes to avoid this.

Lustre-commit: 42c5c9c2ca3e44cb1c3e8ecb144bdd20fb35cddb
Lustre-change: https://review.whamcloud.com/34082

Fixes: 543f1fbe260 ("LU-10830 utils: fix create mode for lfs")
Test-Parameters: trivial clientarch=aarch64
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I82c322013da91d717924e2c664fa57ad4e3ebbe5
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Elena Gryaznova <c17455@cray.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34202
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>

LU-11834 llite: fix temporary instance buffer size

The formatting of the cfg_instance variable was changed in LU-11809
to always use a fixed "%016llu" format, but the temporary buffer
allocations for the instance string were not changed to match the
printed value.

This results in string truncation in some situations.  Change the temp
buffer size to always have 16 bytes for the instance instead of using
sizeof(cfg_instance).

Fixes: cd294a1255 ("LU-11809 llite: don't use %p for cfg_instance")

Lustre-change: https://review.whamcloud.com/33951
Lustre-commit: 1db90b29ad676c2cf1888ef5a7c623161ff23bf9

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I5eca99afa2787cc57e739489b252b12af68cab07
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34315
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-10143 osd-zfs: allocate sequence in advance

on the controller, so that we have it ready before any potential
read-only makeup. this is what osd-ldiskfs is doing already.

Lustre-change: https://review.whamcloud.com/34069
Lustre-commit: 51c449b73994f2bba98ee27ac77f90c9aa846e88

Change-Id: I3d27f112b0d013ac923c5d250b296b5528b8112d
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34295
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-11721 tests: wait for statfs to update on DNE

Wait for the statfs to update properly when there are multiple
MDTs so that the test doesn't gratuitously fail.

Fixes: 757403191c3 ("LU-11721 utils: print used inodes ratio ...")
LU-11721 tests: wait for statfs to update on DNE

Lustre-commit: 263e80f4572b49044407b09f8a3e393677eafb5d
Lustre-change: https://review.whamcloud.com/34164

Test-Parameters: trivial testlist=sanity mdscount=2 mdtcount=4 ostcount=7
Test-Parameters: testlist=sanity fstype=zfs mdscount=2 mdtcount=4
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ia75f7bd4d3027c91f10ce990730b2bd7123ebbe5
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34321
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11721 utils: print used inodes ratio when using "lfs df -i"

"lfs df -i" prints the used blocks percentage, instead of the used
inodes percentage. Fix this by allowing obd_statfs_ratio() to
distinguish when "-i" is used.

Round up the ratio returned from obd_statfs_ratio() in a ceiling
manner, to match the output of df(1). Add a sanity test to check
that the outputs from df(1) and lfs df match.

Lustre-commit: 757403191c37db75ed35b02c971846dced5d5119
Lustre-change: https://review.whamcloud.com/33758

Signed-off-by: Nikitas Angelinas <nangelinas@cray.com>
Cray-bug-id: LUS-6748
Test-Parameters: trivial
Change-Id: I0b31ecb7371875c93bc07dda1f1c89e04d5b4576
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34136
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-11696 utils: "lfs getsom" returns "24" to userspace

The "lfs getsom" command always returns "24" to userspace because
"rc = 24" (sizeof(struct lustre_som_attrs)) after fetching the
xattr from the kernel.
In this patch, rc is set to 0 if the lfs_getsom()->lgetxattr()
call returns a positive value.

Lustre-change: https://review.whamcloud.com/33714
Lustre-commit: 9209e0e9428b6671c6bab9f901e04fdf5b29abc5

Signed-off-by: Qian Yingjin <qian@ddn.com>
Change-Id: Ie3151f67b5ce2b5b2bc35a4b6528ba9a20a5db9f
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34308
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-10384 mgs: replace_nids large string and failover support

Replace_nids uses nids list as new UUID. UUID string
length is limited by 38 symbols. So new nids list need
to be less then 38 symbols.

With this patch first nid in list string representation
is used for UUID as this done for failover nids.

Replace nids finds records for given device and regenerates
lines that contain old nids. add_uuid and add_conn lines for
failover used to be deleted during replace_nids which breaks
failover configuration.

This patch adds failover support to replace_nids command.
For example:

lctl replace_nids lustre-MDT0000 nid1,nid2:nid3,nid4:nid5,nid6

nid3,nid4 - nids from first failover node
nid5,nid6 - nids from second failover node

Lustre-change: https://review.whamcloud.com/30624
Lustre-commit: 09da1564d3794ca7b82e1c1791da253bee6178d4

Signed-off-by: Artem Blagodarenko <c17828@cray.com>
Cray-bug-id: MRP-4505
Change-Id: I4e9a35e8fa8781909ecbaa74785700f4ca04cf92
Reviewed-by: Vladimir Saveliev <c17830@cray.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34296
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-11739 lod: don't inherit default layout from root directory

There is no need to inherit the default directory layout from
the root directory when subdirectories are created therein.
This consumes xattr space on the subdirectories, and makes it
more complex to change the filesystem default layout in the future.

This patch fixes the above issue in lod_ah_init() to check if
the parent directory is the root directory and not copy
the default layout xattr to the new subdirectory.

Lustre-change: https://review.whamcloud.com/33956
Lustre-commit: 0a988cae95f99fee1a9c0d489ce00d0954d2a68e

Lustre-change: https://review.whamcloud.com/34175
Lustre-commit: ad1a74527f0ec59510bfa124b8280617a2b93840

Change-Id: Ie0d286785bdbcd73e2ae60b429e66d5d54b44eef
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34135
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-11161 tests: start running sanity 160g again

sanity test 160g was failing when run in a DNE configuration
and we stopped running this test meaning added to the
ALWAYS_EXCEPT list. The problem is that the test did not
write enough files to exceed changelog idle index threshold
for deregistering users.

Start running sanity test 160g with DNE testing again.

Lustre-change: https://review.whamcloud.com/33994
Lustre-commit: 22676740969314b1b08a31c24e5ebc4c403e08f2

Test-Parameters: trivial
Test-Parameters: ostfilesystemtype=ldiskfs mdtfilesystemtype=ldiskfs clientcount=2 mdscount=2 mdtcount=4 osscount=1 ostcount=8 testlist=sanity
Signed-off-by: James Nunez <jnunez@whamcloud.com>
Change-Id: I286ef8eb7c4638ff8f357db54c4926d5a2f20ac4
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Emoly Liu <emoly@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34297
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>

LU-8346 tests: remove spaces around fail_val

conf-sanity test 93 tries to set fail_loc and fail_val
with the command 'lctl set_param fail_val = 10 fail_loc...'.
fail_val should have no spaces before and after the
equals sign.

This patch is backport from master branch:
Lustre-commit: Iaa2bff1750a2afa96a73a452a0c098ae92f7616c
Lustre-change: https://review.whamcloud.com/34155

Test-Parameters: trivial mdscount=2 mdtcount=4 envdefinitions=ONLY=93 testlist=conf-sanity
Signed-off-by: James Nunez <jnunez@whamcloud.com>
Change-Id: Iaa2bff1750a2afa96a73a452a0c098ae92f7616c
Reviewed-on: https://review.whamcloud.com/34226
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-11605 osp: max_create_count and create_count changes

Setting max_create_count to 0 causes setting create_count
to 0. Set create_count to OST_MIN_PRECREATE when setting
back max_create_count.
Without the patch create_count remains equal to 0 despite
on changing max_create_count to something != 0.
This causes create to stuck in osp_precreate_reserve
because osp_precreate_send doesn't send new request to OST.
To understand the number of objects to precreate(grow) it
uses opd_pre_create_count that is equal to 0.

Lustre-commit: a531ab5f38a6da1de7948df979ae839aa847a370
Lustre-change: https://review.whamcloud.com/33559

Cray-bug-id: LUS-6435
Change-Id: I940c48f91e9c7d49b766bd85ea271ce229424c7f
Signed-off-by: Sergey Cheremencev <c17829@cray.com>
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-by: Artem Blagodarenko <c17828@cray.com>
Tested-by: Elena Gryaznova <c17455@cray.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34162
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11750 krb5: krb5int_derive_key has 'hash' extra parameter

From Kerberos 5 release 1.15, and introduction of support for
aes-sha2, krb5int_derive_key() groks an additional 'hash' parameter.

Lustre-change: https://review.whamcloud.com/33817
Lustre-commit: 4d1d6ed7849b0532e44f2fd742d4e07b649d6f66

Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I7c6ea5ac2d6844371b254b7361d28c462afe5afa
Reviewed-on: https://review.whamcloud.com/33961
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-11568 ldlm: Remove use of SLAB_DESTROY_BY_RCU for ldlm lock slab

Whatever it was doing does not appear to be necessary anymore
as evidenced with newer kernels where the define was removed,
but disabled in Lustre instead.
Another important reason to remove it is because rhel7.3+ seems
to have broken this RCU functionality and leads to frequent use
after frees.

This patch is back-port from:
Lustre-commit: 82d014e71e14671e876055851a0d37e98b4cc079
Lustre-change: https://review.whamcloud.com/34147

Change-Id: I50991b9daf4ef06b24cb65d7a04a5e9b86706d36
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/34434
Tested-by: Jenkins

New release 2.12.0

Change-Id: Icc5da4f5d1d032982a144ee5d13f214b04389d76
Signed-off-by: Oleg Drokin <green@whamcloud.com>

New release candidate 2.12.0-RC4

Change-Id: I5aa92f8a4232f5293e3253223d9447d23b3b0337
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-11809 llite: don't use %p to generate cfg_instance

In kernel 4.15 and later, using "%p" in a string format to
print a kernel pointer will result in the pointer being
hashed with a random value, and the high bytes will be masked
on 64-bit CPUs to prevent leaking kernel address-space info
to userspace to defeat ASLR.  In early boot, the "%p" pointer
may resolve to "        (ptrval)", if there is not enough
entropy in the system to generate a random hash value.

The superblock pointer is used on the client to uniquely
identify all of the OBD devices connected to it, and to
find the configuration llog that was used to mount the
filesystem, so that it can also be used at unmount time.
The sb pointer is also used in the OBD device names, and
the "        (ptrval)" expansion breaks /sys filenames,
and also breaks the uniqueness of the config instance.

On the server, there is also a pointer value used for the
FLDB SEQ servers of the OSTs.

For the short term, bypass the "%p" hashing, so that mount
continues to work properly, and this can be resolved in a
later patch to change ll_get_cfg_instance() to provide a
unique value that is not directly a kernel pointer.

In llapi_getname() don't depend on the cfg_instance being
exactly 16 characters long, if this changes in the future.

Test-Parameters: clientdistro=ubuntu1804 testlist=sanity
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I166de0248af8fe57535628a64bb770a4e03ebbe5
Reviewed-on: https://review.whamcloud.com/33900
Tested-by: Jenkins
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11783 build: remove lustre_user.h deprecation warning

The "lustre/lustre_user.h" header has been in use for many years.
The patch https://review.whamcloud.com/25246 "LU-6401 uapi: migrate
remaining uapi headers to uapi directory" moved the header to
"linux/lustre/lustre_user.h" and left a stub "lustre/lustre_user.h"
behind that generates a compiler warning that this header is
deprecated.

However, no window was given between the introduction of the new
header and the deprecation of the old header, which makes it harder
for applications to smoothly transition to the new header location.
Also, installing Lustre headers into the "linux/" directory before
Lustre is actually part of the kernel may potentially cause problems.

Disable the deprecation warning in the old header for several
releases, until the new header location has been available for a good
time and it is safe for applications to assume that it is available.

Test-Parameters: trivial
Fixes: 6712478e79588e73e28c7ccac3afc7ac2368a4f3
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: If5a62587e2d3627178a0f7a09c3a4c10801cab07
Reviewed-on: https://review.whamcloud.com/33872
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Ben Evans <bevans@cray.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

New release candidate 2.12.0-RC3

Change-Id: I841a274fd9ab3b35c8c3a5b93e33469a9e230af5
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-11753 obdclass: lu_dirent record length missing '0'

In lu_dirent packing, a '0' is appended after name, but it's not
counted in size calcuation, which may cause crash.

Add sanity test_230l.

Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: Iab4947dea8e26ea798d5f64e218268200a5fabe8
Reviewed-on: https://review.whamcloud.com/33865
Reviewed-by: Stephan Thiell <sthiell@stanford.edu>
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11753 obdclass: index_page support variable length rec

mdd_dir_is_empty() may readdir from other MDT if directory
is striped or remote, in this case, it will issue OBD_IDX_READ
RPC to fetch dir page, and on remote MDT dt_index_page_build()
is called to build page, but this function doesn't support
variable length record, so it may miscalculate offset in
reading, which may cause crash.

Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: Ia25a0aca52fb1323ea64a7ff72bf6022754af32c
Reviewed-on: https://review.whamcloud.com/33837
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Stephan Thiell <sthiell@stanford.edu>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11729 tests: skip sanity test 810 for ARM

Skip sanity.sh test_810 for ARM clients as it always failing.

Test-Parameters: trivial clientarch=aarch64 testlist=sanity
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I84117aebb277d4ddcb7787b715587e330f3ebbe5
Reviewed-on: https://review.whamcloud.com/33864
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11770 osd-ldiskfs: preserve bio_integrity API

Preserve the existing kernel API for bio_integrity when the T10-PI
patches are applied, so that any other code that may be using this
interface do not break.

In particular, keep the EXPORT_SYMBOL(bio_integrity_alloc) and
EXPORT_SYMBOL(bio_integrity_prep) in place to avoid module breakage.

In struct bio_integrity_payload put the *bip_generate_fn and
*bip_verify_fn pointers after *bip_vec, since bip_vec is the last
field directly accessed by callers.

In struct blk_integrity_exchg the bi_idx field only needs to be an
unsigned short since the bio->bi_idx and bio->bi_vcnt values used
with it are also unsigned short.  This saves 8 bytes of padding in
the struct and puts the added bi_bio field at the end to preserve
the structure field alignment for external callers.

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I9e29723d5d581a65b1c2ca2611d012c05b953514
Reviewed-on: https://review.whamcloud.com/33840
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11753 utils: print out DNE2 directory hash flags

There may be flags stored in the lmv_hash_type field, such as
"LMV_HASH_FLAG_MIGRATION" that is set while the directory is
being migrated.  Print out the flag from "lfs getdirstripe".

This is still missing support for "lfs find" to find directories
that have incomplete migration.

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ib6b362f9eb993b5fa0562b3a51b54eaee1ccab07
Reviewed-on: https://review.whamcloud.com/33843
Reviewed-by: Stephan Thiell <sthiell@stanford.edu>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11783 utils: fix warnings when lustre_user.h included

Checking for lustre/lustre_user.h in a configure script
generates a warning because of the included <sys/quota.h>

  checking lustre/lustre_user.h usability... no
  checking lustre/lustre_user.h presence... yes
  WARNING: present but cannot be compiled
  WARNING: check for missing prerequisite headers?
  WARNING: see the Autoconf documentation
  WARNING: section "Present But Cannot Be Compiled"
  WARNING: proceeding with the preprocessor's result
  WARNING: in the future, the compiler will take precedence

Looking into config.log it shows:

  In file included from /usr/include/lustre/lustre_user.h:59,
                   from conftest.c:91:
  /usr/include/sys/quota.h:221: error: expected declaration
    specifiers or '...' before 'caddr_t'

Since we don't really need much from the <sys/quota.h> header,
add conditional #defines for the few needed fields.

The FASYNC constant is not declared everywhere in userspace,
provide a compat declaration if unavailable.

Fix an unused variable warning in ll_dir_ioctl().

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I9cd2b0fcbaf16fe8a5a4a7a0309aada3a72cab07
Reviewed-on: https://review.whamcloud.com/33876
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

New release candidate 2.12.0-RC2

Change-Id: I84a1bcb460331928bd4987f33232c22a40b3d58c
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-11740 kernel: new kernel [RHEL7.6 4.14.0-115.2.2.el7a]

This patch support new RHEL 7.6 release on ARM

Test-Parameters: trivial

Change-Id: I79424f0759b79e0a2f45ea5337c3577f832dccb1
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33804
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11684 config: fix conf-sanity test_123

conf_param parameters go into FSNAME-MDT/OST/client files
set_parm -P parameters go into "params"

Change the test to set the conf_param parameter files first
followed by the set_param -P parameters, since there may be
overlap.  In this case, the test infrastructure is using
conf_param to set jobid_var, and test_123 is using set_param.
This causes collision and occasional failure.

Test-Parameters: trivial testlist=conf-sanity

Signed-off-by: Ben Evans <bevans@cray.com>
Change-Id: I5cbdbaf6cc0c1c55a870bd587e89b2cbdaf77c29
Reviewed-on: https://review.whamcloud.com/33720
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

Revert "LU-11152 lnd: test fpo_fmr_poool pointer instead of special bool"

This reverts commit 9b790ba0f5606c0a91563828fa43f5e4ae210425.

Change-Id: Ibca8e813ec7372510709578e33309140e8fc7b5f
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33802
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11734 lnet: handle multi-md usage

The MD can be used multiple times. The response tracker needs to have
the same lifespan as the MD. If we re-use the MD and a response
tracker has already been attached to it, then we'll update the
deadline for the response tracker. This means the deadline on the MD
is for its last user.

Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: I681630c3d599f66c007926525708e3004b343455
Reviewed-on: https://review.whamcloud.com/33794
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Olaf Weber <olaf.weber@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11440 misc: require ldiskfsprogs-1.44.3.wc1 or later

Require a current version of ldiskfsprogs to include support for
project quotas and large_dir.  Upstream now also includes ea_inode
support and many bug fixes in xattr verification and repair.

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I24eeb1c30d5c7b1daa1ad7d5d2f603d273054035
Reviewed-on: https://review.whamcloud.com/33766
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

New Release candidate 2.12.0-RC1

Change-Id: I5464c8f19a06e3dd35b7704bf4359726514e62ad
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-11389 lnet: increase lnet transaction timeout

Increase the new LNet Health transaction timeout to the original
50s value, to avoid spurious lnet-selftest failures and expected
false timeouts under load.

Fix the lnet_transaction_timeout module parameter description.

Test-Parameters: trivial clientarch=aarch64 testlist=lnet-selftest,lnet-selftest,lnet-selftest,lnet-selftest
Test-Parameters: clientarch=aarch64 testlist=lnet-selftest,lnet-selftest,lnet-selftest,lnet-selftest

Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Change-Id: Ic9df69ef8c7a4085815b54dc6741f37a73d36a75
Reviewed-on: https://review.whamcloud.com/33231
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>

LU-10576 tests: allow log files to be created/removed

Allow an llog file to be created or removed during the course of
the test, as this can happen due to internal housekeeping activity.

Also ensure that background cleanup has finished with ZFS before
fetching the number of objects from the MDT.

Test-Parameters: trivial
Test-Parameters: testlist=sanity mdscount=2 mdtcount=4 mdtfilesystemtype=zfs ostfilesystemtype=zfs
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I0a0968cfcd90c7493c67b54ba8a7f326163ebbe5
Reviewed-on: https://review.whamcloud.com/33677
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11697 ost: do not reuse T10PI guards of unaligned page write

If the write is partial page, the guards of RPC checksum should not
be reused for bio submission since the data might not be full-sector.
The bio guards will be generated later based on the full sectors. If
the sector size is 512B rather than 4 KB, or the page size on OST is
larger than 4KB, this might drop some useful guards for partial page
write, but it will only add minimal extra time of checksum calculation.

Change-Id: I868342df87c28ea91f5f8364fe377277595ecf6d
Signed-off-by: Li Xi <lixi@ddn.com>
Reviewed-on: https://review.whamcloud.com/33752
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11697 osc: wrong page offset for T10PI checksum

The page offset might could be non-zero value. Thus, when
calculating T10PI checksum, the offset should be correct value.

Change-Id: Ib32584eb47ea55ec3804e531ac02ffd252411886
Signed-off-by: Li Xi <lixi@ddn.com>
Reviewed-on: https://review.whamcloud.com/33727
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11663 osd-zfs: write partial pages with correct offset

otherwise non-aligned writes send wrong data to ZFS.

Change-Id: I1ae1f361981d548307d74344a5694f3ef39c0609
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33726
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11652 kernel: kernel update [SLES12 SP3 4.4.162-94.69]

Update SLES12 SP3 kernel to 4.4.162-94.69.

Test-Parameters: clientdistro=sles12sp3 ossdistro=sles12sp3 mdsdistro=sles12sp3

Change-Id: Iea1ec6def609059d67053d25360e8e986f2adbd9
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33637
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11668 mdt: check parent type in rename/migrate

Check parent existence and type in rename and migrate to avoid
potential race.

Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: I583e2f5a6f47073601e36c06890a6b22dfc734ad
Reviewed-on: https://review.whamcloud.com/33709
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-11668 debug: print object type in mdd_parent_fid

mdd_parent_fid() get parent fid for directory, but racer shows
the passed in object is not directory, print its type to help
debug.

Test-Parameters: trivial
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Signed-off-by: Peter Jones <pjones@whamcloud.com>
Change-Id: I4b3eedb159fc0efccc15e35cb59010fe02fa9e01
Reviewed-on: https://review.whamcloud.com/33700
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>

LU-10114 hsm: increase upper limit of maximum HSM backends registered with MDT

Lustre only supports at most 32 HSM backends, which limits HSM to be applied
to other features, such as LPCC. This patch breaks the limitation by allowing
the system take any interger number as a valid archive-id.

Test-Parameters: clientjob=lustre-b2_10 clientbuildno=136 testlist=sanity-hsm
Test-Parameters: mdsjob=lustre-b2_10 ossjob=lustre-b2_10 serverbuildno=136 testlist=sanity-hsm

Change-Id: I9523b92500b962db3e45a2bd6a67dba54eef5335
Signed-off-by: Teddy Zheng <teddy@ddn.com>
Signed-off-by: Li Xi <lixi@ddn.com>
Reviewed-on: https://review.whamcloud.com/32197
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-5152 quota: disable sync chgrp to OSTs

The syschronous chgrp to OSTs introduced by the previous patch
(the commit is 8a71fd5061bd073e055e6cbba1d238305e6827bb) causes
deadlock between MDT and OST, this patch disable it for now and
leave the updated patch to fix it.

Signed-off-by: Hongchao Zhang <hongchao@whamcloud.com>
Change-Id: I5ce48424f4d2011ce62e69047ace7f0b7c3ebbe5
Reviewed-on: https://review.whamcloud.com/33705
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11653 hsm: copytool registration wakes the coordinator

When a copytool registers to the MDS, it is possible there are
pending requests in the coordinator's llog that previously could not
be sent (either because there were not any copytools, or not any
compatible copytools).

With this patch, the coordinator will process those requests on its
next wake up (which happens every second).

Test-Parameters: trivial
Test-Parameters: mdscount=2 mdtcount=4 mdtfilesystemtype=zfs testlist=sanity-hsm
Test-Parameters: mdscount=2 mdtcount=4 mdtfilesystemtype=ldiskfs testlist=sanity-hsm
Test-Parameters: mdscount=2 mdtcount=4 mdtfilesystemtype=zfs testlist=sanity-hsm
Test-Parameters: mdscount=2 mdtcount=4 mdtfilesystemtype=ldiskfs testlist=sanity-hsm

Signed-off-by: Quentin Bouget <quentin.bouget@cea.fr>
Change-Id: Ie49b40d312f2f3e0d9c85dee27bb8813dc4dde40
Reviewed-on: https://review.whamcloud.com/33649
Tested-by: Jenkins
Reviewed-by: Ben Evans <bevans@cray.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11541 build: Use correct kernel version for DKMS MLNX OFED.

Check for $O2IBPATHS/${LINUXRELEASE} before $O2IBPATHS/default.
The "default" link is only created on the first dkms build of OFED.
So if doing dkms build for another kernel it may not be pointing
to the correct kernel.

Test-Parameters: trivial

Signed-off-by: Ake Sandgren <ake.sandgren@hpc2n.umu.se>
Change-Id: If8054ee64cf8795ed0e3ee50a8ef9ced067059d7
Reviewed-on: https://review.whamcloud.com/33702
Tested-by: Jenkins
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11567 utils: llog_reader print changelog index

When processing changelog type llogs, print the changelog index number
with each changelog record.  This allows one to compare the records on
disk with the output of lfs changelog or relatives.

Test-Parameters: trivial
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Change-Id: I0059cc34b39161462b3eadbb2512dc811c38705a
Reviewed-on: https://review.whamcloud.com/33473
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11329 utils: Add maintainer entries

Add clio reviewer entry.

Add Grant subsystem section.

Test-Parameters: trivial
Signed-off-by: Patrick Farrell <paf@cray.com>
Change-Id: I461b39fe0b65f8d48b3912086927c945b2da3db7
Reviewed-on: https://review.whamcloud.com/33684
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>

LU-11645 tests: fix sanity-sec test 31

In sanity-sec test 31, command to add new LNet network ${NETTYPE}999
may fail if servers have interface names different from the one used
on the client.
So fix the command so that it is run directly on each node.

Test-Parameters: trivial envdefinitions=ONLY=31 testlist=sanity-sec
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: Ibf9101524d94188b3beae3debe45e2ba151999ca
Reviewed-on: https://review.whamcloud.com/33622
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Sonia Sharma <sharmaso@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11572 tests: make sanity-hsm test_260c reliable

When the coordinator restarts, its first run does housekeeping.
For the following `loop_period' seconds, it will not run in
housekeeping mode.

This patch uses this to improve the reliability of sanity-hsm
test_260c.

Test-Parameters: trivial testlist=sanity-hsm,sanity-hsm,sanity-hsm
Signed-off-by: Quentin Bouget <quentin.bouget@cea.fr>
Change-Id: I8eeb8b6856d91b69495d592cdd1cb5f091b1cc2b
Reviewed-on: https://review.whamcloud.com/33478
Reviewed-by: Ben Evans <bevans@cray.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11519 hsm: improve the testing of hsm.max_requests

Modify sanity-hsm::test_250() to send more diverse HSM request.

This brings better code coverage (it would have caught the bug
reported in LU-11519).

Test-Parameters: trivial testlist=sanity-hsm,sanity-hsm,sanity-hsm
Signed-off-by: Quentin Bouget <quentin.bouget@cea.fr>
Change-Id: I0142b46dc804f649c33deb81efea7f68d1e29afa
Reviewed-on: https://review.whamcloud.com/33590
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Ben Evans <bevans@cray.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11662 llite: handle -ENODATA in ll_layout_fetch()

In ll_layout_fetch() handle -ENODATA returns from mdc_getxattr(). This
is needed for interop and restores the behavior from before commit
0f42b388432c4b898857660197ef13a40a82cd9d (LU-11380 mdc: move empty
xattr to mdc layer) landed.

Test-Parameters: clientjob=lustre-b2_10 clientbuildno=136 testlist=sanity-hsm
Test-Parameters: mdsjob=lustre-b2_10 ossjob=lustre-b2_10 serverbuildno=136 testlist=sanity-hsm
Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: I1fb85faf35c7d2303a1f61a5a9c5922988739817
Reviewed-on: https://review.whamcloud.com/33665
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11595 mdt: fix read-on-open for big PAGE_SIZE

Client PAGE_SIZE can be larger than server one so data returned
from server along with OPEN can be misaligned on client.

Patch replaces assertion on client with check and graceful exit,
changes MDC_DOM_DEF_INLINE_REPSIZE to be PAGE_SIZE at least and
updates mdt_dom_read_on_open() to return file tail for maximum
possible page size that can fit into reply.

Signed-off-by: Mikhail Pershin <mpershin@whamcloud.com>
Change-Id: Ic2c54b95c814d3b6df3b527527cac08488060651
Reviewed-on: https://review.whamcloud.com/33606
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11582 llite: protect reading inode->i_data.nrpages

truncate_inode_pages() looks up pages in the radix tree without
lock, and could miss finding pages removed from the radix tree
by __remove_mapping(), so that after calling truncate_inode_pages()
we need to read the nrpages of the inode->i_data with the protection
of tree_lock.

Since it could still be in the race window of __remove_mapping()->
__delete_from_page_cache()->page_cache_tree_delte(), before the
nrpages being decreased.

Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: I44ba6bea3dec4f0a110d1ae2a749514ec7dd0d12
Reviewed-on: https://review.whamcloud.com/33639
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-11466 mdt: Skip SOM xattr update for DoM-only files

When scan the MDT image, DoM-only file can be specialized handled,
the size and blocks can be got directly from the inode on MDT, no
need SOM xattr anymore.
Thus, there is no need to store the SOM xattr for DoM-only files.

Signed-off-by: Qian Yingjin <qian@ddn.com>
Change-Id: I0f871cde38fc846460dd1b6f92509dee9ea90bfc
Reviewed-on: https://review.whamcloud.com/33331
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-9793 ptlrpc: Do not map unrecognized ELDLM errnos to EIO

The lustre_errno_hton and lustre_errno_ntoh functions map between
host and network error numbers before they are sent over the network.
If an errno is unrecognized then it is mapped to EIO.

However an optimization for x86 and i386 architectures replaced the
functions with macros that simply return the original errno. The
result is that x86 and i386 return the original values for ELDLM
errnos and all other architectures return EIO. This difference is
known to break glimpse lock callback handling which depends on clients
responding with ELDLM_NO_LOCK_DATA. The difference in errnos may
result in other as yet unidentified bugs.

The fix defines mappings for the ELDLM errors that leaves the values
unchanged. Error numbers not found in the mapping tables are still
mapped to EIO.

Cray-bug-id: LUS-6057
Signed-off-by: Ann Koehler <amk@cray.com>
Change-Id: I0b4e1e0dc6de065729e18f2381ec9cfc58fe31db
Reviewed-on: https://review.whamcloud.com/33471
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11519 hsm: handle hsd_request_count == 0 properly

In mdt_cdt_waiting_cb() it may be that the coordinator has already
reached the limit of active requests and hsd contains no requests to
be started. Handle this properly when trying to prioritize a restore.

Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: Ic843b7672ae6a4509ac127c2d2f90bf3681f84fc
Reviewed-on: https://review.whamcloud.com/33580
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Quentin Bouget <quentin.bouget@cea.fr>
Reviewed-by: Ben Evans <bevans@cray.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11071 build: use --with-linux-obj for ubuntu

We can use --with-linux-obj instead of --with-linux-config
the config will be found just under LINUX_OBJ.
Otherwise the configure could fail if LINUX and LINUX_OBJ
are different paths. e.g. ubuntu18

Test-Parameters: trivial
Signed-off-by: Li Dongyang <dongyangli@ddn.com>
Change-Id: Ibaa159e5611c4054b3987da0ce3a8bca2992e057
Reviewed-on: https://review.whamcloud.com/33145
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11642 lmv: allocate fid on parent MDT in migrate

During directory migration, if the migrated file is not directory,
the target should be allocated on its parent MDT, not user specified
MDT. Because if it's parent is striped, this file should be migrated
to the MDT by its name hash, not the starting MDT of its parent.

Add sanity 230k to check file data not changed after migration.

Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: Ic7d3de8ea982b7cf4da758e4d3ab8d8ee15ecfdb
Reviewed-on: https://review.whamcloud.com/33641
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11642 mdt: revoke remote LOOKUP lock in dir layout shrink

mdt_dir_layout_shrink() should revoke remote LOOKUP lock if parent
is remote, because it will alter dir layout, which is refreshed
upon lookup.

Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: I26ae1af5da6142b44005e5d9ea11293af65ed7b5
Reviewed-on: https://review.whamcloud.com/33640
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11497 tests: improve ha.sh to set striped dirs

For DNE II testing we need the possibility to create
the striped directories, like it is done in t-f:test_mkdir().
This patch covers the following settings:
-c stripe_count -i mdt_index
-c stripe_count -i <random mdt_index>,
which allows to have the clients work directories with
the same stripe_count but different mdt indexes.

Test-Parameters:trivial
Signed-off-by: Elena Gryaznova <c17455@cray.com>
Cray-bug-id: LUS-5974
Reviewed-by: Vladimir Saveliev <c17830@cray.com>
Reviewed-by: Alexander Boyko <c17825@cray.com>
Change-Id: I6117601e741a95059750149d1c38b402fccb29b7
Reviewed-on: https://review.whamcloud.com/33340
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alexandr Boyko <c17825@cray.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11494 tests: sanity-quota/22 syntax error fix

Patch fixes test_22() trivial syntax error.

Test-Parameters:trivial testlist=sanity-quota envdefinitions=ONLY=22
Signed-off-by: Elena Gryaznova <c17455@cray.com>
Cray-bug-id: LUS-6011
Reviewed-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Vladimir Saveliev <c17830@cray.com>
Change-Id: Icb3474c7166d6b74b3092617bb314a561687df06
Reviewed-on: https://review.whamcloud.com/33337
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>

LU-11492 tests: fix thread_sanity() defect

tmin, tmax, tstarted are not set if do_facet fails by some reason.
This leads to the following failure:
Assertion 28 failed: (($tstarted >= $tmin && $tstarted <= $tmax ))
   (expanded: ((16 >=  && 16 <= 16 )))
Patch sets variables equal to 0 for cases of failed do_facet, like
it is done for other similar cases.

Test-Parameters:trivial testlist=sanity envdefinitions=ONLY="53a 53b"
Signed-off-by: Elena Gryaznova <c17455@cray.com>
Cray-bug-id: LUS-5638
Reviewed-by: Alexander Boyko <c17825@cray.com>
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Change-Id: If713ba20a0f71cb17208f776a9a4edd359c07c43
Reviewed-on: https://review.whamcloud.com/33335
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alexandr Boyko <c17825@cray.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11541 build: Adjust OFED check for DKMS enabled MLNX OFED.

The DKMS packaging of MLNX OFED, i.e. mlnx-ofed-kernel-dkms, (at least
v4.4 and newer), doesn't have /usr/src/ofa_kernel/default as part of
the package, it is created as a link during package installation.

Restore LB_USES_DPKG autoconf test that was removed since it is needed
in lustre-lnet.m4 so Debian / Ubuntu can build properly with external
infiniband stacks.

Test-Parameters: trivial
Signed-off-by: Ake Sandgren <ake.sandgren@hpc2n.umu.se>
Change-Id: I1739a67e16bc0dab8cf551b460b424441d81fc91
Reviewed-on: https://review.whamcloud.com/33396
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-9906 clio: use pagevec_release for many pages

When Lustre releases cached pages, it always uses
page_release, even when releasing many pages.

When clearing OST ldlm lock lrus in parallel with lots of
cached data, the ldlm_bl threads spend most of their time
contending for the zone lock taken by page_release.
Also, when osc_lru_reclaim kicks in when there's not enough
LRU slots during I/O, the contention on zone lock kills
I/O performance.

Switching to pagevec when we expect to actually release the
pages (discard_pages, truncate, lru reclaim) brings
significant performance benefits as shown below.

This patch introduces cl_pagevec_put() to release the pages
in batches using pagevec, which is essentially calling
release_pages().

  mpirun -np 48 ior -w -r -t 16m -b 16g -F -e -vv -o ... -i 1 [-B]

                mode         write (GB/s)    read (GB/s)
  master        O_DIRECT     20.8            21.8
  master+patch  O_DIRECT     20.7            22.2
  master        Buffered     11.6            12.3
  master+patch  Buffered     15.3            19.6

Also clean up the dead lovsub_page related code.

Signed-off-by: Patrick Farrell <paf@cray.com>
Signed-off-by: Li Dongyang <dongyangli@ddn.com>
Change-Id: I71447528db12858defb627c9c03b7193d116c935
Reviewed-on: https://review.whamcloud.com/28667
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alexey Lyashkov <c17817@cray.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-8777 mdt: add parameter to disable remote/striped dir

Restore mdt_enable_remote_dir to enable/disable remote directory
feature, and also add mdt_enable_striped_dir and
mdt_enable_dir_migration for striped directory and directory
migration.

Enable remote directory, striped directory and directory migration
by default.

Add sanity.sh 417.

Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: I583dbc5a5eb54de02c8a4459092d4a3f3b24b262
Reviewed-on: https://review.whamcloud.com/24498
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11596 tests: disable several sanity sub-tests for ARM

The following sanity sub-tests are consistently failing
on ARM clients:
- 42d, 42e, 63a, 63b, 64a, 64b and 64c (LU-11596)
- 45 (LU-11671)
- 101c (LU-11665)
- 103a (LU-11594)
- 317 (LU-11667)

This patch adds them to ALWAYS_EXCEPT list to exclude
the failures, so as to make ARM client test session pass.

They will be removed from the list and re-enabled in the
patches that fix the failures.

Test-Parameters: trivial

Change-Id: I1728dc665c8a52f0fd88de8173f88ccee0c8772d
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33652
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>

LU-11635 gss: proper gss build for client-only

Fix gss build system for both rpms and debs in case of client-only
build.
Fix consists in passing along '--{enable,disable}-gss' option, and
including lgssc.conf file in case of client-only build.

Test-Parameters: trivial
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: Id884d282e312177c262b1f4725b8f19e2c14fd37
Reviewed-on: https://review.whamcloud.com/33608
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11102 ldlm: don't skip bl_ast for local lock

The previous commit 954cc675 skips bl_ast for local lock but
there are cases on MDT when local lock can become a client lock,
see mdt_intent_lock_replace(). In that case the client should be
notified if lock is a blocking lock.

Patch reverts commit 954cc675 and provides alternative solution.
During downgrade to COS the lock renews own blocking AST states
and start reprocessing. Any new lock conflict will cause new
blocking AST and related async commit as needed.

Test-Parameters: mdssizegb=20 testlist=racer,racer,racer
Signed-off-by: Mikhail Pershin <mpershin@whamcloud.com>
Change-Id: I41adab5c805a59fdbeade8ae3556556b779dc3c0
Reviewed-on: https://review.whamcloud.com/33458
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10626 test: create custom udev rule

Some lustre developer do all their testing from their lustre
source tree instead of installing it into an image. Since lustre
is not installed into the image the udev rule for lctl set_param
doesn't work since its not in /etc/udev/rules.d. Additonally udev
rules require using the absolute paths for executables. To handle
this setup create a special udev rule on the fly using $LCTL to
be placed into /etc/udev/rules.d.

Change-Id: I84c55450c866977650b3c6d4190e233a24eb03d8
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/33143
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10335 test: enable sanity 130 tests for Ubuntu

The default e2fsprogs for Ubuntu 16 lacked the changes for fiemap.
This caused sanity test 130 to fail. Now with lustre using newer
e2fsprogs with the new Ubuntu server support this is no longer the
case. Lets re-enable sanity 130 test again.

Test-Parameters: clientdistro=ubuntu1604 trivial testlist=sanity envdefinitions=ONLY=130

Change-Id: Iad42e1f55855da904515b3fe00c8d047fa13b45c
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/33459
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2836 tests: Re-enable sanity-quota/3,6 for ZFS

Given the amount of fixes in Quota for ZFS 0.7.x and since Lustre 2.4.
Also ZFS backend is now not nearly as slow so sleep timeouts that work
for ldiskfs should also work for ZFS.
Re-enable these tests as they should consistantly pass now.

Test-Parameters: trivial
Test-Parameters: ostfilesystemtype=zfs mdtfilesystemtype=zfs ostcount=2 testlist=sanity-quota,sanity-quota,sanity-quota,sanity-quota,sanity-quota
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: I23714181a1aada7a32a540853cb2e0bd9d7211a2
Reviewed-on: https://review.whamcloud.com/32956
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11020 osp: fix race during lov_objids update

First thread can be delayed due to reading from disk, so it
will completed after second thread and overwrite the on-disk
lov_objids data with an older OID for that OST.

If the transaction commits during this window and then the
MDS crashes, it is possible that the stale lov_objids results
in an OST object being deleted during MDS->OSS recovery that
should have been kept.

Use a single buffer shared between threads to store lov_objids
so that even if multiple threads are updating the lov_objids
file at once, the latest OID will be written to disk even if
the threads commit their transactions out of order.

Cray-bug-id: LUS-5841
Change-Id: I0984e5f55d569260c1219bf87c82423cc5b8589b
Signed-off-by: Alexey Lyashkov <c17817@cray.com>
Reviewed-on: https://review.whamcloud.com/32867
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11654 mdd: check ucred before using it

On the code path of client eviction, lu_ucred in mdd_close() is not
initialized so we need to check it before trying to dereference it.

Signed-off-by: Jinshan Xiong <jinshan.xiong@uber.com>
Change-Id: I51b2a30e651761d66470fedf39da8e99b47ea6b4
Reviewed-on: https://review.whamcloud.com/33638
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11071 build: add files to .gitignore

For the debian / Ubuntu build lustre-resource-agent is not ignored
with git status. Also llsom_sync should be ignored.

Change-Id: Ic4ce9b2e097f7ede60cc8dc4d4ed63cb5a9ad692
Test-Parameters: trivial
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/33626
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>

LU-11071 osd-ldiskfs: support bio integrity with Ubuntu 18

The recent landed for bio integrity broke building osd-ldisk for
both Ubuntu 18 and RHEL alt kernels. The reason is struct bio no
longer contains a struct block_devices. Instead we can get the
block device from the osd_device instead.

Change-Id: Ic72db4b29f7d0d1921fd97cb5df5c6c77908c92d
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/33624
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-8602 gss: support OpenSSL 1.1

Add support for OpenSSL 1.1, used by default on Ubuntu 18.04.

It mainly consists in changing the way variables of type DH
and HMAC_CTX are used.

Test-Parameters: testlist=sanity-sec envdefinitions=SHARED_KEY=true
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I1fc9b77fc44976f4b1fb3a58ae7db9c20a96b3e9
Reviewed-on: https://review.whamcloud.com/33592
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-9795 gss: properly handle mgssec

Secured connection to MGS must have all 3 root, mds and oss flags set
by convention.

Also add regression tests for mgssec in sanity-sec.

Test-Parameters: testlist=sanity-sec envdefinitions=ONLY="32 33",SHARED_KEY=true
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I655913b15a551c205cdc3e16fffd48193c8da1fe
Reviewed-on: https://review.whamcloud.com/33357
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Jeremy Filizetti <jeremy.filizetti@gmail.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11071 build: kernel update Ubuntu 18.04

This enables ldiskfs support for kernels newer than
4.15.0-24.26

Signed-off-by: Li Dongyang <dongyangli@ddn.com>
Change-Id: I1d61534d7f5288495b0e26c3362211c10b9826dd
Reviewed-on: https://review.whamcloud.com/33322
Reviewed-by: Gu Zheng <gzheng@ddn.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-8066 lod: replace class_process_proc_param()

With the move to sysfs for lod every lctl conf_param option will
send out a udev event. Replace class_process_proc_param() with
class_modify_config() so sysfs attributes can be set directly.

This patch exposed an error in stripecount_store() which doesn't
allow setting stripe count to -1 which is wrong.

Change-Id: I4eee8b2b0f3661ac69242ff7a1da6346153a56e8
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/33180
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10965 doc: add mirror options to "lfs find" in lfs.1

This patch adds the following mirror related search
options to "lfs find" command in lfs man page:

[[!] --mirror-count|-N [+-]n]
[[!] --mirror-state <[^]state>]

The usage of the options are explained in separate
"lfs find" man page.

Change-Id: I13c19be01cb050cd5035d9058bbd6fc7d2f50564
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: https://review.whamcloud.com/32201
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>

LU-11483 ldlm ofd_lvbo_init() and mdt_lvbo_fill() create env

on demand as in very few cases it's called with env undefined.
it would be a very big patch to pass env through all the path
chains..

Change-Id: Ic6819215877891ad95f2937729941453abf981e7
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33321
Tested-by: Jenkins
Reviewed-by: Patrick Farrell <paf@cray.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11527 tests: fix sanity 270a for ARM

Fix sanity test_270a to work with ARM so that it does O_DIRECT
writes with a multiple of the local PAGE_SIZE (65536 bytes)
instead of 102400 bytes (which works fine on x64, but not ARM).

Generalize the test to use variables instead of precomputed values
so that it is clear why the particular values are used.

Test-Parameters: trivial envdefinitions=ONLY=270a
Test-Parameters: mdtfilesystemtype=zfs ostfilesystemtype=zfs testlist=sanity envdefinitions=ONLY=270a
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ic2a32215dc1423b0bbf1e5188a5e85420a3ebbe5
Reviewed-on: https://review.whamcloud.com/33633
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11597 tests: fix O_DIRECT test usage for ARM

Fix cases where dd [io]flag=direct is used but the blocksize is
not a multiple of the client PAGE_SIZE, which is required for
Lustre.

Add a global PAGE_SIZE variable in init_test_env() and use that
in the tests, rather than having multiple local functions/calls
to re-determine the PAGE_SIZE value, which will not change for
the duration of the test.

Keep the get_page_size() function to initialize the global
PAGE_SIZE value and for the few cases where tests use the server
PAGE_SIZE (e.g. obdecho testing).

Test-Parameters: trivial
Test-Parameters: mdtfilesystemtype=zfs ostfilesystemtype=zfs testlist=sanity
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I0bb9efbcc743f6508180c8496fc50a67363ebbe5
Reviewed-on: https://review.whamcloud.com/33636
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11418 llog: refresh remote llog upon -ESTALE

If a distributed transaction is aborted, it will invalidate all
objects involved, which include remote catalog and update logs.

So llog_cat_declare_add_rec() will refresh remote llog upon -ESTALE,
but it only does this for chd_current_log, not chd_next_log.

If an aborted transaction happens to invalidate catalog only, and
leave chd_current_log valid, which will cause subsequent operations
fail to create chd_next_log.

This patch prepares both current and next log of catalog before
declare_write in llog_cat_declare_add_rec().

Add sanity.sh 60g.

Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: Ie2d37686e910676587baefa1687ebb607be3c3a1
Reviewed-on: https://review.whamcloud.com/33401
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11522 tests: sanity-sec/27 unable to remount client

test_27() defect: test expects that lustre root is mounted on
$MOUNT which is not true if t-f runs witn FILESET set.

Patch fixes the test to have lustre root mounted before test
starts and restores the FILESET mount in the test cleanup.

Test-Parameters: trivial testlist=sanity-sec
Signed-off-by: Elena Gryaznova <c17455@cray.com>
Cray-bug-id: LUS-6560
Change-Id: I00cf50e00ed96d55552f010b5766a8bdab75a200
Reviewed-on: https://review.whamcloud.com/33389
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alexandr Boyko <c17825@cray.com>
Reviewed-by: Sergey Cheremencev <c17829@cray.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11489 tests: retry power off/on several times

Sometimes we have a BMC hangup for ~1 minute.
Patch improves ha.sh to try power off/on several
times with delay $ha_power_delay default 60 sec.

Test-Parameters:trivial
Signed-off-by: Elena Gryaznova <c17455@cray.com>
Cray-bug-id: LU-6486
Reviewed-by: Andrew Perepechko <c17827@cray.com>
Reviewed-by: Alexander Boyko <c17825@cray.com>
Change-Id: Id192e2a39a229979484b0d209fa8fdc5fafbbfc6
Reviewed-on: https://review.whamcloud.com/33332
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alexandr Boyko <c17825@cray.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11406 mdt: mdt_reint_setxattr() to check rr_eadata

when LMV EA is requested for removal mdt_reint_setxattr() shouldn't
try to access rr_eadata as it's NULL.

Change-Id: Iae2b6a96ba72ae54a3151f86c0f53af9517e3aad
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33273
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11425 quota: support quota for DoM

Support block quota enforcement on MDT to accommodate the data
writing introduced by DoM.

This patch adds a new qsd_instance on the OSD device at MDT to
implement the quota enforcement on block data, for there is only
one type of quota (meta data or block data) to be handled in one
qsd_instance in the current quota design of Lustre, it's much more
changes to adapt the qsd_instance to support more quota type than
to use a separated qsd_instance to enfore meta data on MDT.

Change-Id: I7c57f383d338cf0070e05ecc318d42586c39371e
Signed-off-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33257
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-11390 quota: mark over-quota flag correctly

Currently, if the current granted quota space is less than or
equal to the usage + pending_write + waiting_write, the over-quota
flag will be marked and sent back to OSC, it will damage the write
performance drastically. this patch will compare the current pending
and the waiting space to the threshold of over-quota check, and the
threshold will be taken into account only if the remain margin quota
space is larger than the threshold.

Test-Parameters: mdtfilesystemtype=zfs ostfilesystemtype=zfs \
testlist=sanity-quota,sanity-quota,sanity-quota,sanity-quota,sanity-quota

Change-Id: I0ecd134c3119ec1d86dbaa6e668091a68b7f5f54
Signed-off-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/33238
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Wang Shilong <wshilong@ddn.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-8602 gss: get rid of cfs_crypto_hash_desc

Get rid of cfs_crypto_hash_desc, and use ahash_request instead.

Test-Parameters: testlist=sanity-sec envdefinitions=SHARED_KEY=true
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: If47f5e66846459e8951779f4c6db22a9e4cf23a8
Reviewed-on: https://review.whamcloud.com/33493
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10030 hsm: make changelog flag argument an enum

Since the changelog record flag is being stored on disk, pass it
around as an enum instead of a signed int.  Also make it clear at
the caller that only the low 12 bits of the flag are normally
being stored in the changelog records, since this isn't obvious
to the reader.  For open and close records, the bottom 32 bits
of open flags are recorded.

Test-Parameters: trivial testlist=sanity-hsm
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ie676fb6bfd3661645dde845fd9ca8bad283ebbe5
Reviewed-on: https://review.whamcloud.com/32112
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10876 lnet: peer deletion code may hide error

lnet_peer_ni_del_locked might return -EBUSY if the
NID to be deleted is a gateway.

Check for the return value of lnet_peer_ni_del_locked
in lnet_peer_del_nid.

Change-Id: I72f0d9355284bcc14b2827df82765b95a6d589f0
Test-Parameters: trivial
Signed-off-by: Sonia Sharma <sonia.sharma@intel.com>
Reviewed-on: https://review.whamcloud.com/31861
Tested-by: Jenkins
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-8602 gss: Properly port gss to newer crypto api.

In newer kernels the old crypto_hash_* was removed which
requires the gss layer to now user crypto_ahash_*. Pieces
of the port were done incorrectly which this patch addresses.
In the process cleanup some cases where data was leaked
in the case of the sg_tables.

Address the limitation of the gss keys in that they are hard
coded to u16 variables that are not guaranteed to be constant
and it will prove to be challenge in the future to add new crypto
algorithms.
Keep u16 representation of hash and crypto algorithms in Shared-Keys
as they are stored on disk, to not break compatibility with already
existing keys. But pass in to kernel space the name string
of the crypto algorithm we want to use.

Also deal with timer_setup and key.usage new format introduced in
newer kernels.

Test-Parameters: testlist=sanity-sec envdefinitions=SHARED_KEY=true
Change-Id: I90ef58104a1955ce12603173964a6878f60b601e
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-on: https://review.whamcloud.com/28309
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10937 mgc: restore mgc binding for sptlrpc

The work for LU-9034 mapped config logs to separate mgc devices.
This change prevented the ability to configure sptlrpc. A later
work around was introduced in LU-9567. Recently it was reported
that the work around introduced can now cause a MGC failover
panic. This patch is the proper fix in that the sptlrpc is
properly bound to an mgc device.

The sptlrpc config record expects 2 pieces of data:

  *  [0]: fs_name/target_name,
  *  [1]: rule string

What was happening is that when you set cfg_instance it was used
to create a new instance name of the form fsname-%p. For sptlrpc
it expects it to only be fsname. The solution is to test if the
config record is for sptlrpc and in that can keep the first
record field as is. With this change we can drop cfg_obdname
which only sptlrpc used.

Test-Parameters: testlist=sanity-gss envdefinitions=ONLY=1,SHARED_KEY=true
Test-Parameters: testlist=sanity-sec envdefinitions=SHARED_KEY=true

Change-Id: I785f98264c6269f95c0d9a564b731d1b6ff0bcee
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/33311
Tested-by: Jenkins
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-9795 gss: fix gss-based integrity check for multi-rail

In case of multi-rail, rq_peer is not the NID from which request is
received, but primary NID of peer.
So gss_svc_upcall_handle_init() needs to pass rq_source, which
contains the NID actually in use, to lsvcgssd user-space daemon
in order to compute HMAC with correct key in handle_sk().

Test-Parameters: testlist=sanity-sec envdefinitions=ONLY=31,SHARED_KEY=true
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I70ee05355c3463a267a6faef724a5b7159ada014
Reviewed-on: https://review.whamcloud.com/33415
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>