LU-6142 lustre: change all 'struct seq_operations' to const

All 'struct seq_operations' are now 'static const'.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ia0f6a2863835ec02fb7d283a099eea489f65ab3a
Reviewed-on: https://review.whamcloud.com/39396
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-6142 lustre: change super/file/inode operations to const

All 'struct file_operations', 'struct inode_operations', 'struct
export_operations' and 'struct super_operations' are changed to
'const'.  This potenetially allows them to be placed in read-only
memory, and ensure they are never changed.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I8b236f0248eca11f91f11da02fe18be3f6d2e17c
Reviewed-on: https://review.whamcloud.com/39394
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-6142 obdclass: use cl_object_for_each more broadly.

There are several loops which follow the pattern of the macros
cl_object_for_each() or cl_object_for_each_reverse(), but
don't use those macros.

Change to use the macros.  In some case we change the name of an 'obj'
arg to a function to 'top', to avoid overloading the name 'obj'.

Also include some minor formatting improvements.

Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I544d4148719a9284f33fd61730bd733c0d40e97c
Reviewed-on: https://review.whamcloud.com/39391
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-6142 lustre: remove module_vars arg to class_register_type()

The module_vars arg to class_register_type() is always NULL.  So it
can be removed.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ie8d7e79075ba068dec606cc9dfcc38a90e371e5b
Reviewed-on: https://review.whamcloud.com/39383
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-6142 lov: style cleanups in lov_set_osc_active()

1/ don't pre-declare the function as it is not used before it
   is defined.
2/ " TEST ? 1 : 0" is identical to "TEST", so remove the "? 1 : 0".
3/ When the 'then' part of an 'if' ends with a GOTO, there is
   not point having an 'else', particularly if there is also
   code in the block following the "if".  The 'else' code and
   the following code should be together.

4/ other minor changes, and conversion of spaces to tabs.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I00cb8b1efe25c0ad9adc76ba21cca003aed203e2
Reviewed-on: https://review.whamcloud.com/39382
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-6142 lustre: change obd_ioctl_getdata() args

Instead of passing a pointer to 'char *' to obd_ioctl_getdata(),
pass a pointer to 'struct obd_ioctl_data *'.
This seems more natural, and avoid needing extra 'char *data'
variables wherever the function is called.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I0287cd2fafd071a2580d48a52498469ceded3e79
Reviewed-on: https://review.whamcloud.com/39381
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>

LU-6142 lov: chnage lsm_op_find() to a non-inline function.

Making lsm_op_find() non-inline improves encapsulation and means that
various _ops structures can now be static.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ifee2ab4f023e56bfa1e663a0106590cf7858856b
Reviewed-on: https://review.whamcloud.com/39377
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-6142 lustre: make various 'struct file_operations' static

These 'struct file_operations' are only used locally, so make them
static.
Except lprocfs_evict_client_fops() which isn't used at all and doesn't
exist, so discard the declaration.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ib6c51683c1e765db202b3f72d2accebe17191303
Reviewed-on: https://review.whamcloud.com/39741
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-9859 libcfs: discard TCD_MAX_TYPES

As well as CFS_TCD_TYPE_CNT we have TCD_MAX_TYPES which has a larger
value but a similar meaning.  Discard it and just use
CFS_TCD_TYPE_CNT.

Two places relied on the fact that TCD_MAX_TYPES was larger and so
there would be NULLs at the end of the array.  Change
them to check the array size properly.

Linux-commit: 22627ffa5e7fae0b0b62f90a8f30fc6dae0fff3c

Change-Id: I73a54a409674d3b569bcbaf72e2296209542b74c
Signed-off-by: Mr. NeilBrown <neilb@suse.de>
Reviewed-on: https://review.whamcloud.com/39668
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10391 lnet: allow lnet_connect() to use IPv6 addresses.

Now that we can created and bind IPv6 sockets, connecting to them is
trivial.

Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I6032ea8a5937711922d9c318f72bc64c7020cc6d
Reviewed-on: https://review.whamcloud.com/37706
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10391 lnet: allow creation of IPv6 socket.

With this patch, lnet_sock_create() can make IPv6 sockets.  If an
interface and destination is given, completely different code is
needed to bind the local address.

When no interface or destination is given, an IPv6 socket is always
created, so lnet_acceptor() needs to request that IPv4 connections are
accepted as well, and lnet_sock_getaddr() needs to detect mapped v4
addresses and present them as true v4 addresses.

Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Iaef1c08a9afd578b60f54cc83fd2d6882a021e6b
Reviewed-on: https://review.whamcloud.com/37705
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-9679 ptlrpc: list_for_each improvements.

1/ use list_for_each_entry_safe() instead of
   list_for_each_safe() and similar.

2/ use list_first_entry() and list_last_entry() where appropriate.

3/ When removing everything from a list, use
   while ((x = list_first_entry_or_null()) {
   as it makes the intent clear

Linux-commit: ef8e5dbbb09035a0c41aa47a328e6248702d4d2b

Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I490afcb70d18170e4ecc5d062d327d141668d3db
Reviewed-on: https://review.whamcloud.com/39133
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14275 tests: add ior_CLEANUP

Sometimes we need to run read IOR test separately from
write test. IOR -r test requires file exists before test starts.
If ior_CLEANUP set to false $testdir is not removed and
$testdir/iorData file can be used for IOR -r further test.

Test-Parameters: trivial testlist=parallel-scale
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-9560
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@hpe.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Change-Id: Iac790976a46db2bf5be374a42e3facdce074228b
Reviewed-on: https://review.whamcloud.com/41079
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-14388 utils: always enable ldiskfs project quota

Always enable project quota for newly-formatted ldiskfs filesystems.

Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I1b0f745bc04b5c42592bcc4fd9823d068fef2a79
Reviewed-on: https://review.whamcloud.com/41370
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@whamcloud.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14353 obd: move debug.c to obdecho

The code in obdclass/debug.c is only used from obdecho/, so move the
file there and clean it up a little on the way (e.g. spaces to TABs).

Similarly lustre_debug.h contains very little that is used outside of
obdecho.  Move that little (LL_CDEBUG_PAGE()) into lustre_lib.h, and
add the declarations needed for debug.c into echo_internal.h.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ia1bcf3f3e1453130ed1adc75918bdad23eb5b5c7
Reviewed-on: https://review.whamcloud.com/41290
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14305 ldiskfs: add parameters for mb_c123_threshold

Add mount options for /sys/fs/ldiskfs/*/mb_c[123]_threshold values
so that they can be set persistently via mount options.

The /sys/fs/ldiskfs/*/mb_c[123]_threshold values are always shown
rounded down to the next lower percentage value due to integer
division, since internal values are stored as blocks for efficiency.

Round up the values shown to the next percent to match what was
used to originally set these parameters.

Fixes: 95f8ae567749 ("LU-12103 ldiskfs: don't search large block range if disk full")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ie36a6667f8bca7481aa8179ab5b97c85d449d619
Reviewed-on: https://review.whamcloud.com/41193
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14289 ptlrpc: move heap.c from libcfs to ptlrpc

cfs_binheap is only used by the nrs code in ptlrpc, so move it out of
libcfs to the one place it is used.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I4687d4de0a83e0ce106d4f4cfd5ea95faef27dfc
Reviewed-on: https://review.whamcloud.com/41129
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14291 lustre: only include nrs headers when needed

Various lustre_nrs_*.h header are only needed on the server.  Make
this explicit, so it is easier to see which files are needed on the
client.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I9bc30acc4f62ea9af516055edbb7833919811089
Reviewed-on: https://review.whamcloud.com/41128
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14291 ptlrpc: support nrs_delay for client-only builds

nrs_delay is intended for use on the client, but is currently disabled
for client-only builds.  So move the ptlrpc_nrs_policy_register() call
outside of the #ifdef.

Also change the nrs modules that *are* server-only to not have #ifdef
in the C file, but that rather to be identified in the Makefile as
only being built when server code is wanted.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ie25ed5ac837ff5f891824fad31ab258923840ec0
Reviewed-on: https://review.whamcloud.com/41127
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14285 utils: Add error message when osd_init fails

It was observed that mkfs.lustre fails silently under
function osd_init(). One such case where it silently
fails is when back-end is 'zfs' and libz.so library
is not loaded or under search path.

This patch adds error message in case osd_init() call
fails. It makes end user observe the failure more easily.

Before Patch:
~~~~~~~~~~~~~
$ echo $LD_LIBRARY_PATH
/root/zfs_spl_install/0840/zfs/lib
$ unset LD_LIBRARY_PATH
$ mkfs.lustre --reformat --index 0 --backfstype=zfs --fsname=lustre --mgsnode=192.168.50.140@tcp --mgs --mdt gpool/metadata /dev/sdb1
$ echo $?
22

After Patch:
~~~~~~~~~~~~
$ echo $LD_LIBRARY_PATH
/root/zfs_spl_install/0840/zfs/lib
$ unset LD_LIBRARY_PATH
$ mkfs.lustre --reformat --index 0 --backfstype=zfs --fsname=lustre --mgsnode=192.168.50.140@tcp --mgs --mdt gpool/metadata /dev/sdb1
mkfs.lustre: osd_init() failed : 22 (Invalid argument)
$

Test-Parameters: trivial
Signed-off-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Change-Id: Ie75b00974852d7a4b95b52d26501e55b66db5891
Reviewed-on: https://review.whamcloud.com/41118
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-12477 llite: remove unused ll_teardown_mmaps()

The ll_teardown_mmaps() function is no longer used and can be
removed.

Fixes: f114d4dac6f8 ("LU-12477 llite: use generic_error_remove_page()")
Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I42506210e028da3bf8036a28d4d61b1e7d3ebbe5
Reviewed-on: https://review.whamcloud.com/41086
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-930 doc: fix format man page sections for lctl

For lctl man pages, add proper formatting for the NAME
and TITLE sections so that 'man -k lctl*' reports the
real explanation for the command/utility and not
"(unknown subject)".

Modifications in this patch include:
- add a dash between multi-word name commands in NAME
- format the date to YYYY-MM-DD in TITLE
- capitalize command in TITLE
- change hyphen (-) to underscore in command where necessary

Test-Parameters: trivial
Signed-off-by: James Nunez <jnunez@whamcloud.com>
Change-Id: I946a3683b0a48b662c4e3b7b5783d389fa710e08
Reviewed-on: https://review.whamcloud.com/41075
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>

LU-14272 tests: different mpirun options for different users

Patch adds the ability to use different mpirun options for
different mpi users, for example we use:
  --wdir=/home/quota15_1 for user quota15_1
and
  --wdir=/home/quota15_2 for user quota15_2.

Per mpi user option can be specified after ':' in ha_mpi_users:
Example:
  ha_mpi_users="quota15_1:--wdir=/home/quota15_1
                quota15_2:--wdir=/home/quota15_2
                quota15_3
                quota15_4:--wdir=/home/quota15_4"
Several options for user  can be set as:
  ha_mpi_users="user1:<option11>
                user1:<option12>
                user2:<option21>"
MPIRUN_OPTIONS will be applied for all users.

Patch adds the ability to execute the command before loads start.
Example: ha_precmd="for i in 1 2 3 4; do
                lfs quota -v -u quota15_$i $ha_test_dir; done"

Test-Parameters: trivial
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-9069
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@hpe.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Change-Id: I35feadcc856dbc3d6d6d5f2720ab49c6c111161b
Reviewed-on: https://review.whamcloud.com/41072
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14271 tests: add new node crash method

Patch adds a new node crash method: node crash
is simulated by command executed on the victim nodes.
Example:
  POWER_DOWN=sysrqcrash
triggers a crash  on all victim nodes.

Test-Parameters: trivial
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-8162
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@hpe.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Tested-by: Vitaly Fertman <vitaly.fertman@hpe.com>
Change-Id: Id39b0f58fbb59236b3af51516fe968c888b04a17
Reviewed-on: https://review.whamcloud.com/41071
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14270 tests: delay node's power up

In CS a node in UNCLEAN state can be STONITHed after
it has passed power down and power up already.

The patch adds ha_power_up_delay() in order to delay
node's power up until CRM state became OFFLINE when
failover pair is set and for $NODE_UP_DELAY seconds otherwise.
LOAD_TIMEOUT is added in order to ha_load_timeout be tunable.
The failover pair list for all victims are to be set via new
-f option.

ha_powermanage() is added to control node state by pm, (designed
by Maxim Storchak <maxim.storchak@hpe.com>).
PM_HOST is added in order to specify powermanage host, default
is $(hostname).

Test-Parameters: trivial
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-8610
Reviewed-by: Maxim Storchak <maxim.storchak@hpe.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Change-Id: Ia192200b9ff61c17a67fc16858938f24ade46e23
Reviewed-on: https://review.whamcloud.com/41070
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Wei Liu <sarah@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14262 utils: lfs to set component flags by pool name

so it'd be easy to set flags (like prefer) on the components
residing on specific OST identified by pool.

Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I733f92fe186682dc8d34512edf75b49e565c457f
Reviewed-on: https://review.whamcloud.com/41024
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14195 lustre: remove 'fs' from 'struct lvfs_run_ctxt'

The code protected by push_ctxt() and pop_ctx() never tries to access
any user-space data, so call set_fs() to KERNEL_DS is not needed.

So remove the 'fs' field and related code.

In linux-5.10 this code fails to compile as set_fs() is deprecated.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Idb2744d656dc4228375b6da54673e38cc1c112f5
Reviewed-on: https://review.whamcloud.com/40910
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14195 osd: don't use set_fs() for ->fiemap() calls.

->fiemap() only accesses kernel-space data, so does not need, and
never has needed, set_fs() calls.
In Linux 5.10, these calls are deprecated.
So remove the unnecessary code.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Id336855b4787ddbf656dfa3b8d0b12f663564795
Reviewed-on: https://review.whamcloud.com/40909
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14195 build: Adjust Makefile for Linux build changes.

Since v5.10-rc1~51^2~19, "KBUILD_BUILTIN" has been unset
for module builds.  This means that "targets-for-builtin"
isn't built, and that is how "extra-y" is built.

So we need another way to force LUSTRE_KERNEL_TEST to be built.

Since v5.6-rc1~1^2~5 any target listed in "always-y" will always get
built.  So we can assign LUSTRE_KERNEL_TEST to this macro.

Assigning both macros is safe, even for those kernels which include
both in the list of targets.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I508b3710579c068dec93baf81ee383f3f03bd370
Reviewed-on: https://review.whamcloud.com/40907
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14178 ldlm: return error from ldlm_namespace_new()

Return the underlying error in ldlm_namespace_new() from
ldlm_namespace_sysfs_register() to the caller instead of NULL.
Otherwise, the callers convert the NULL to -ENOMEM and this
is incorrectly reported as an allocation error to the user.

  sysfs: cannot create duplicate filename
     '/fs/lustre/ldlm/namespaces/lustre-OST0002-osc-ffff89f33be70000'
  mount.lustre: mount mgs:/lfs at /lfs failed: Cannot allocate memory

Change ldlm_namespace_new() to return errors via PTR_ERR() and
change the callers to use IS_ERR().

Fix associated CERROR() messages to follow proper code style.

Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I7d590c7242399549b32b1c4189e46ff8748c8096
Reviewed-on: https://review.whamcloud.com/40851
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14073 ofd: remove use of smp_read_barrier_depends()

Linux 5.9 removes smp_read_barrier_depends(), so lustre must stop
using it.

There is only one use: in ofd_access_log.c.
This use is unnecessary and can simply be removed.

The code is based on "Documentation/core-api/circular-buffers.rst"
which gives no indication that this barrier is needed.

The comment say its purpose is to ensure the index is read before the
data is read. This is unnecessary.
The data is written in osl_write_entry(), then a barrier is issued
(smp_store_release) before the ->head is written.
oal_read_entry() issues a barrier (smp_load_acquire()) before reading
that head.
'tail' is read without a barrer, but it then compared against ->head
in CIRC_CNT().  Even if reading ->tail was racey, the fact that
comparing it wilth ->head succeeded means that the data written at
->tail must have been safely written, and we can now read it without
any further barrier.

Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I9d0f0aeb67e1188d2012f4ae2e14b3656211c3e2
Reviewed-on: https://review.whamcloud.com/40394
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14047 lustre: change EWOULDBLOCK to EAGAIN

On linux, EWOULDBLOCK has always been defined as an alias for
EAGAIN. In the interest of readability we should not use two names for
the same thing. So change the remaining uses of EWOULDBLOCK to EAGAIN
and add EWOULDBLOCK||EAGAIN to spelling.txt.

Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: Ib48b8a1e58bfa961d2a4ba411c038c476bfc300d
Reviewed-on: https://review.whamcloud.com/40307
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Yingjin Qian <qian@ddn.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13783 libcfs: switch from ->mmap_sem to mmap_lock()

In Linux 5.8, ->mmap_sem is gone and the preferred interface
for locking the mmap is to suite of mmap*lock() functions.

So provide those functions when not available, and use them
as needed in Lustre.

Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I4ce3959f9e93eae10a7b7db03e2b0a1525723138
Reviewed-on: https://review.whamcloud.com/40288
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-9325 osd-ldisk: replace simple_strto* with kstr* functions

The parsing of mount parameters from the config llog is done in
some cases with simple_strto* which is considered obsolete.
Replace simple_strto* with the kstrto* equivalent functions.

Change-Id: I7c26d14d02828c9f9a96f31a086a65bb39f3ea87
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/40119
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-9859 lod: use linux kernel bitmap API

Now that modern Linux kernels support a bitmap API we can move
away from the libcfs specific bitmap API. This patch changes
the last bitmap in the lod module to use the Linux kernel
bitmap API.

Change-Id: I92c494bf2af62e31d7b9527b3f44580322e48fd3
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/39876
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-9859 libcfs: replace all CFS_CAP_* macros with CAP_*

Lustre defines a few CFS_CAP_* macros which are exactly the
same as the corresponding CAP_* macro, with one exception.

CFS_CAP_SYS_BOOT is 23
CAP_SYS_BOOT is 22.

CFS_CAP_SYS_BOOT is only used through CFS_CAP_FS_MASK and
causes capability 23 (CAP_SYS_NICE) to be dropped in certain
circumstances.
It is probable that the intention was to drop CAP_SYS_BOOT,
and this is what is now done.

CFS_CAP_CHOWN_MASK and CFS_CAP_SYS_RESOURCE_MASK are never
used, so they have been removed.

Linux-commit: 5ebaa2d14850205e44757c4d5fdd4097712d01ef

Change-Id: Ifb90c0a36e204c76b90ff23ac609345d11b878da
Signed-off-by: Mr. NeilBrown <neilb@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-on: https://review.whamcloud.com/39875
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Oleg Drokin <green@whamcloud.com>

LU-13100 lov: grant deadlock if same OSC in two components

The same osc can be involved in several components but osc layer
leaves active last used extent, so an RPC can't be sent if grants
are required from the same OST for another component.

Add cl_io_extent_release() to release active extent before
switching to the next component.

Change-Id: Idadda6eaecd1d47b78880c81e1fb7513d5be2419
Cray-bug-id: LUS-8038
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-by: Andrew Perepechko <c17827@cray.com>
Reviewed-on: https://review.whamcloud.com/37095
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-7853 lod: fixes bitfield in lod qos code

Updating bitfields in struct lod_qos struct is protected
by lq_rw_sem in most places but an update can be lost
due unprotected bitfield access from
lod_qos_thresholdrr_seq_write() and qos_prio_free_store().
This patch fixes it by replacing bitfields with named bits
and atomic bitops.

Cray-bug-id: LUS-4651
Signed-off-by: Rahul Deshmukh <rahul.deshmukh@seagate.com>
Signed-off-by: Alexander Zarochentsev <c17826@cray.com>
Change-Id: I28299ce4960e91be551d7f6e43a3b598daf4d7a2
Reviewed-on: https://review.whamcloud.com/18812
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14180 utils: verify setstripe comp_end is valid

Verify that the "lfs setstripe -E <component_end>" value is valid.
Otherwise, if "-S" is not specified at the same time, then an
invalid file layout can be created and the file cannot be deleted
normally, only via "lfs rmdif <FID>".

Allow values < 4096 (e.g. '64' or '128' which would all be invalid
anyway) to be interpreted as KiB units.

Update usage messages and man pages to match.

Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Change-Id: I47fe7729ffd447c1c1cc098e5117e456263ebbe5
Reviewed-on: https://review.whamcloud.com/41239
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14207 mgs: delete "add failnid" sections on replace_nids

Replace_nids left old nids in add_conn field of failnid
section of client llog. This leads to connecton errors.

Let's delete such sections. New failnids, if any, are
added by replace_nids.

Change-Id: I7fab00827035bd864aeb95fb4852a59c458bb2ba
HPE-bug-id: LUS-9440
Signed-off-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Reviewed-by: Vladimir Saveliev <c17830@cray.com>
Reviewed-by: Alexander Boyko <c17825@cray.com>
Tested-by: Jenkins Build User <nssreleng@cray.com>
Reviewed-on: https://review.whamcloud.com/40930
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13974 llog: check stale osp object

The logic of osp_attr_get has 2 path,
1) return attributes from a cache for health osp object
2) make an out update request and return attributes for stale
osp object, object lose stale state.

When some out update request with llog writes failed, osp object
become stale. But llog handle stay inconsistent (bitmap,count,
last_index), and a next llog_add->llog_osd_write_rec do dt_attr_get,
gets attributes and makes osp object valid, and uses wrong llog
handle data. The result is index jump at llog file - recX, recX+2.
And it makes an error during update log processing if failover take
a place.
The fix adds dt_object_stale function to check osp_object.
llog_osd_write_rec check it and return ESTALE. llog_add would fail
with ESTALE error and doesn't corrupt update log.

HPE-bug-id: LUS-9030
Signed-off-by: Alexander Boyko <alexander.boyko@hpe.com>
Change-Id: Iadf53fd816e1c5bde0a19d4c537f0408796c864a
Reviewed-on: https://review.whamcloud.com/40742
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-12961 mdd: avoid double call to mdd_changelog_fini()

the first call is done from mdd_prepare() as part of error
handling, another call is done from mdd_device_shutdown().

in the similar cases fini routines checks whether the state
is initialized. e.g. mdd_orphan_index_fini() releases the
object and sets mdd_orphans to NULL, then all subsequent
calls to mdd_orphan_index_fini() return immediately.
mdd_changelog_fini() can't do this way, so the excplicit
state has been introduced.

Change-Id: Ifd21569e68c836f44bb59adea4e8fed6ccef1c7b
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/36737
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>

LU-14439 build: require a newer version of e2fsprogs

Require a build version of e2fsprogs-1.44.3 for osd-ldiskfs builds to
get EXT2_FLAG_IGNORE_SB_ERRORS and ext2fs_has_feature_{mmp,quota}()
functions.

Require a minimum install version of ldiskfsprogs-1.45.6.wc1.

Test-Parameters: trivial
Fixes: 7dc8aa7e7848 ("LU-13241 utils: use libext2fs for ldiskfs operations")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I240823266c7342338c72126011485120543ebbe5
Reviewed-on: https://review.whamcloud.com/41682
Reviewed-by: Olaf Faaland-LLNL <faaland1@llnl.gov>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13609 mgs: fix config_log buffer handling

Fix buffer handling in mgs_list_logs() to list all MGS config_logs
using multiple ioctl calls when we have a large number of targets.

Fixes: 1d97a8b4cd3d ("LU-13609 llog: list all the log files correctly on MGS/MDT")
Signed-off-by: Stephane Thiell <sthiell@stanford.edu>
Change-Id: I1bf32e918e242f4da83c3d1624b7285a18a88d01
Reviewed-on: https://review.whamcloud.com/41478
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Emoly Liu <emoly@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-10391 socklnd: use sockaddr instead of __u32 addresses.

LNet/socklnd often uses __u32 to hold an ipv4 address.
As we want to extend socklnd to work with IPv6 addresses too,
this needs to change.

This patch changes many __u32s to variants of 'struct sockaddr'.

Library code from sunrpc is used for copying and comparing addresses
   rpc_copy_addr() and rpc_cmp_addr()
and for extracting or setting the port:
   rpc_get_port() and rpc_set_port().

The "%pIS" printf format is used for printing a sockaddr (works for
both IPv4 and IPv6), and "%pISp" for printing the address with the
port.

The __u32 is in host-byte-order, while addresses in sockaddr are
always network-order, so htonl and ntohl are used as needed.

When storing an address (e.g. in a structure), 'struct
sockaddr_storage' is used.  When passing an address to a function,
'struct sockaddr' is used.  When an address is known to be IPv4 (i.e.
when converting to or from __u32), 'struct sockaddr_in' is used.

The following functions are changed to take a 'struct sockaddr*'
argument:

 lnet_connect()
 lnet_connect_console_error()
 lnet_sock_getaddr()
 ksocknal_ip2iface()
 ksocknal_ip2index()
 ksocknal_create_route()
 ksocknal_connecting()
 ksocknal_close_peer_conns_locked()
 ksocknal_peer_del_interface_locked()

The following structures have had fields changed to 'struct
sockaddr_storage'

 struct ksock_interface:
      ksni_ipaddr -> ksni_addr
 struct ksock_conn
      ksnc_myipaddr -> ksnc_myaddr
      ksnc_ipaddr and ksnc_port -> ksnc_peeraddr
 struct ksock_route
      ksnr_myipaddr -> ksnr_myaddr
      ksnr_ipaddr and ksnr_port -> ksnr_addr

Assorted strings have been joined onto a single line, and various
indented have been changed from space to tabs.

There should be no behaviour changes, though the structures mentioned
above will now be a little larger.

Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I42d12260185638407b5b611391fc69bfd9f91754
Reviewed-on: https://review.whamcloud.com/37704
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13929 lnet: modify assertion in lnet_post_send_locked

Check that the pointer to the local interface is not NULL
before asserting. While checking if local ni is the destination,
the assertion may attempt to dereference pointer to local
interface after it has already been cleaned up on shutdown.

Test-Parameters: trivial
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: I0f4be04a728a7243823bec70f9efbe52bcb104b3
Reviewed-on: https://review.whamcloud.com/40749
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Cyril Bordage <cbordage@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14362 tests: sanity-flr to prepare stuff before checks

otherwise /mnt/lustre may be missing and sanity-flr fails

Test-Parameters: trivial testlist=sanity-flr
Change-Id: Ic5502ee07bab557162fbf16718ca5fb5beed45e9
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/41307
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-14423 osd: recognize holes in osd_is_mapped()

ldiskfs_fiemap() can return {0,0,0} for last non-allocated
region.  osd_is_mapped() should be able to recognize and
cache this state.

Fixes: 144b5a65c1 ("LU-7132 osd-ldiskfs: speedup rewrites")
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I03883038c2c0ec84754377a442c4947c7e3021a9
Reviewed-on: https://review.whamcloud.com/41481
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14398 llapi: add llapi_fid2path_at()

Add llapi_fid2path_at() which works like llapi_fid2path() takes an
open FD on the moint point instead of a 'fsname or dirirectory path'
and a const struct lu_fid * instead of a const char *.

Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: I76234bc28de231587b65c5d866954441e0893aac
Reviewed-on: https://review.whamcloud.com/41406
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14398 llapi: simplify llapi_fid2path()

Simplify llapi_fid2path(). Remove the fid_is_sane() check. Remove the
call to root_ioctl() and use get_root_path() directly.

Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: Ib70b8c9e239c77da8b46408de8341fc8aaf4d1c3
Reviewed-on: https://review.whamcloud.com/41405
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Emoly Liu <emoly@whamcloud.com>

LU-14390 gnilnd: Use DIV_ROUND_UP to calculate niov

Use DIV_ROUND_UP to calculate niov, also remove 'unlikely'
as this is the common case.

Fixes: 7a74d382 ("LU-13004 modules: replace lnet_kiov_t with struct bio_vec")
Test-Parameters: trivial
Signed-off-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Change-Id: Iaf62326743ebe5e92bb3aa4b5780b47a5cfdfb18
Reviewed-on: https://review.whamcloud.com/41371
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14301 lustre: add ENOTSUPP to spelling.txt

Add a spelling check for ENOTSUPP to suggest use of EOPNOTSUPP
instead. Note:

ENOTSUPP (524) and defined only in the kernel errno.h and is a NFSv3
specific errno. If ENOTSUPP is returned to userspace then strerror()
will print "Unknown error 524".

EOPNOTSUPP (95) is defined in kernel and userspace errno.h.

ENOTSUP is defined in userspace errno.h as an alias for EOPNOTSUPP.

Test-Parameters: trivial
Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: I13b0389c9ec0853f43d8ab4a8f6538eb24c8a2ad
Reviewed-on: https://review.whamcloud.com/41280
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>

LU-12766 test: convert time to seconds properly

According to test logs, grace time could be 3m56s,
and it will be converted to 21305s by wrongly which
caused timeout of test sytem.

Actually, grace time could be something like 1w2d3h4m5s
Fix calculations for this.

Test-Parameters: trivial testlist=sanity-quota
Signed-off-by: Wang Shilong <wshilong@ddn.com>
Change-Id: I1c02e83f5ec15de2a4a6b312dd6e36b55dd4a7bc
Reviewed-on: https://review.whamcloud.com/41259
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13903 build: make lustre-devel buildable for Linux client

Recently a new lustre-devel rpm was created which always has a
dependency on the kernel detected by the kmod macros. This doesn't
work in the case of running a custom Linux kernel. Addreess this
by not making use of the kmod macros when only building the Lustre
utilies i.e configure --disable-modules.

Test-Parameters: trivial
Fixes: 16af4e5ed634 ("LU-9215 build: Re-add the lustre-devel package")
Signed-off-by: James Simmons <jsimmons@infradead.org>
Change-Id: I324fbaa0f5b03e2095b493f3d8e00b74ca64298a
Reviewed-on: https://review.whamcloud.com/41188
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14313 utils: mount error when no server support

When the mount utility was built without server support,
and this causes the mount to fail, an error message should
say so.

The error values returned by main() should be positive as
they are exposed to the user.

Upon mount() failure, errno must be checked to determine
the reason for the failure.  The return value is -1.

Test-Parameters: trivial
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Change-Id: I49906340f6ddbe16b7503663cedabe0085daf113
Reviewed-on: https://review.whamcloud.com/41182
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14268 lod: fix layout generation inc for mirror split

Mirror split does not increase the layout generation properly.

Mirror split does not change FLR state of the file, even when it
contains 1 mirror afterwards, and FLR state should be LCM_FL_NONE
instead.

Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: I9c9621d67d901f2e9ca6ed3e0684cd308c396076
Reviewed-on: https://review.whamcloud.com/41068
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14098 obdclass: try to skip corrupted llog records

if llog's header or record is found corrupted, then
ignore the remaining records and try with the next one.

Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: If47ec1fc1e2eaf64be7ba08d3aa9c2b93903c0cf
Reviewed-on: https://review.whamcloud.com/40754
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-9820 osd-ldiskfs: OI scrub speed limit fix

The OI scrub speed limit is set at start time, and shouldn't be
changed in the middle, otherwise lfsck is run with speed control,
while OI scrub may finish before it, thus sanity-scrub 9 failed.

Allow wider margin in speed control since the speed control is
heuristic.

Test-Parameters: testlist=sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub,sanity-scrub
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: I995679db085a193c71fc7914f961b538b5930c69
Reviewed-on: https://review.whamcloud.com/40591
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14099 build: Fix for unconfigured arch_stackwalk

On aarch64 CONFIG_ARCH_STACKWALK is not defined and
print_stack_trace is not available.

Replace print_stack_trace with an open-coded variant
using %pB introduced in Linux v2.6.38-6557-g0f77a8d37825

This also fixes the symbols lookup of stack_trace_save_tsk
using kallsyms at module init time over the use of
symbol_get.

HPE-bug-id: LUS-9518
Signed-off-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Change-Id: I04c3a0a84bb1a05d813a90502d1ed0f5bb2e33ab
Reviewed-on: https://review.whamcloud.com/40503
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Tested-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14044 llog: check fid after convert

We should convert from llog_id and then check fid. Also
change fid-lookup to error check instead LASSERT.

Signed-off-by: Yang Sheng <ys@whamcloud.com>
Change-Id: I673d8f16ff9e57a0482d6a3ec3ee3db33699f57f
Reviewed-on: https://review.whamcloud.com/40294
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Emoly Liu <emoly@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13620 tests: pool_add_targets() fix

Fix pool_add_targets to don't fail if number of
OSTs is >= 10 - lctl expects them in a hex view.
Check result of "lctl pool_add". Pass only if
the result is either 0 or 17(EEXIST).
Finally, make pool_add_targets to check that were
added only requested OSTs, i.e. don't fail to add
OST0 if OST1 is already in a pool.

Test-Parameters: trivial testlist=ost-pools
HPE-bug-id: LUS-8723
HPE-bug-id: LUS-8941
Change-Id: I841b3db3a89dbc86075cd23b7d71764ffb849181
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/38803
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13584 tests: gather_logs() fix

Fix gather_logs() to work on real HW where server nodes
do not have the access to clients.

Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-8888
Reviewed-by: Alexander Lezhoev <alexander.lezhoev@hpe.com>
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@hpe.com>
Change-Id: Ifeea54e20d3123ee64582e32b92a4573e60ff33e
Reviewed-on: https://review.whamcloud.com/38660
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Vladimir Saveliev <c17830@cray.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13513 osp: make neterr not fatal for precreate_reserve

When OST_CREATE (not resendable rpc) sent by precreate thread fails
with network error, osp_pre_update_status() sets d->opd_pre_status to
EIO. osp_precreate_reserve() considers EIO as fatal and does not wait
for another attempt from precreate thread. That may make
mdt_intent_open() to return ENOSPC confusing a caller.  ENOSPC comes
from lod_alloc_rr().

osp_precreate_send(): in case of network error switch EIO to ENOTCONN.

Test to illustrate the issue is added.

Cray-bug-id: LUS-8811
Signed-off-by: Vladimir Saveliev <c17830@cray.com>
Change-Id: Iffaad9bd16f216f758c784b708e21b525c999b14
Reviewed-on: https://review.whamcloud.com/38472
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-13453 osd-ldiskfs: do not leak inode if OI insertion fails

osd_create() should destroy just created inode if OI insertion
fails.

also fixes lustre_index_restore() to drop nlink for object to
be removed.

the patch adds two tests:
 - ENOSPC on OI insertion
 - ENOSPC on .. insertion, i.e. directory block allocation

Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I2a5db657c7dab54b8dc2c50bc29365d5ee754a2e
Reviewed-on: https://review.whamcloud.com/38235
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>

Merge "LU-9121 lnet: User Defined Selection Policy (UDSP)"

LU-12682 llite: fake symlink type of foreign file/dir

This patch implements a "fake symlink" specific usage of
"foreign" LOV/LMV format. It basically allows these
particular type of foreign files/dirs to behave as a
symlink from VFS point of view, by allowing to construct
a relative path from the LOV/LMV foreign content, to
complement it with a prefix, and then to expose it to
the VFS as a symlink destination. The default/internal
mechanism simply takes the full foreign free string as
the relative path, and for more complex internal formats
an upcall has been implemented to provide format's
details (presently just in terms of constant strings
and substrings positions in EA, but this can be enhanced)
to llite layer.
Using this feature, instead of real symlinks or user EA,
will permit to benefit from the special features (lock,
prefetch, caches) already implemented to handle both
LOV/LMV EAs.

Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Change-Id: Id3c262e3b042447aa09aad25f682ff02787b350d
Reviewed-on: https://review.whamcloud.com/35856
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Ben Evans <beevans@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14444 gss: handle empty reqmsg in sptlrpc_req_ctx_switch

In sptlrpc_req_ctx_switch(), everything is already there to handle
the case of a ptlrpc_request that has an empty rq_reqmsg.
But assertions were left over at the beginning of the function, so
just remove them from here.

Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I6ae1f8b9da9600d3b57b9efc9018c2461114f2fe
Reviewed-on: https://review.whamcloud.com/41685
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Oleg Drokin <green@whamcloud.com>

LU-14455 mdt: fix DoM lock prolong logic

- don't stop at the first found lock if it is not PW or EX lock
- add LCK_GROUP lock as valid mode of lock to check

Signed-off-by: Mikhail Pershin <mpershin@whamcloud.com>
Change-Id: If947f8565008953cc34146b6f0ac1e0f0a038bb5
Reviewed-on: https://review.whamcloud.com/41701
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14436 tgt: only use T10PI guard when doing full sector read

The T10PI guard was generated on full sectors, if we
do we partial read and still use the guard, the rpc
checksum won't match.

Test-Parameters: trivial
Change-Id: I40d481d703a46b9711021a162208b86a956bd8d1
Signed-off-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-on: https://review.whamcloud.com/41677
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14435 doc: include lfs-flushctx manpage inside packages

lfs manpage redirects to lfs-flushctx(1), so it has to be
included in the Lustre packages.

Test-Parameters: trivial
Fixes: c246a9ba04 ("LU-14263 gss: unlink revoked key")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I5c55f8a74eb6dac20fa85b6ea0663ad701341006
Reviewed-on: https://review.whamcloud.com/41676
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>

LU-14430 mdd: fix inheritance of big default ACLs

If the number of default ACLs in directory is more than 31, then
mdd_acl_init() fails to inherit them for a newly created file.
This limitation is caused by using a fixed-size def_acl_buf buffer
in the mdd_create()->mdd_acl_init() call chain. Instead, the
default ACL buffer should be increased when it is needed.

Patch adds check for -ERANGE after mdd_acl_init(), reallocates
default ACL buffer with required size and calls mdd_acl_init()
again. Thus big default ACL are processed as expected.

Fixes: 6350af100c20 ("LU-3437 mdd: Fix ACL/def_ACL during object creation")
Signed-off-by: Mikhail Pershin <mpershin@whamcloud.com>
Change-Id: I700da90c09f824955fcb8dc7ca0bc2f581f916a0
Reviewed-on: https://review.whamcloud.com/41494
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

New tag 2.14.50

Change-Id: I760c8197bafa0f7eb8d1f5d78863900d647ccddc
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-9121 lnet: User Defined Selection Policy (UDSP)

User Defined Selection Policies (UDSP) are introduced to add
ability of fine traffic control. The policies are instantiated
on LNet constructs and allow preference of some constructs
over others as an extension of the selection algorithm.
The order of operation is defined by the selection
algorithm logical flow:

   1. Iterate over all the networks that a peer can be reached on
      and select the best local network
      - The remote network with the highest priority is examined
        (Network Rule)
      - The local network with the highest priority is selected
        (Network Rule)
      - The local NI with the highest priority is selected
        (NID Rule)
   2. If the peer is a remote peer and has no local networks,
      - then select the remote peer network with the highest
        priority (Network Rule)
      - Select the highest priority remote peer_ni on the network
        selected (NID Rule)
      - Now that the peer's network and NI are decided, select
        the router in round robin from the peer NI's preferred
        router list. (Router Rule)
      - Select the highest priority local NI on the local net of the
        selected route.
        (NID Rule)
   3. Otherwise for local peers, select the peer_ni from the peer.
      - highest priority peer NI is selected
        (NID Rule)
      - Select the peer NI which has the local NI selected on its
        preferred list.
        (NID Pair Rule)

   Accordingly, the User Interface allows for the following:
   - Adding a local network udsp: if multiple local networks are
     available, each one can have a priority.
   - Adding a local NID udsp: after a local network is chosen,
     if there are multiple NIs, each one can have a priority.
   - Adding a remote NID udsp: assign priority to a peer NID.
   - Adding a NID pair udsp: allows to specify local NIDs
     to be added on the list on the specified peer NIs
     When selecting a peer NI, the one with the
     local NID being used on its list is preferred.
   - Adding a Router udsp: similar to the NID pair udsp.
     Specified router NIDs are added on the list on the specified
     peer NIs. When sending to a remote peer, remote net is selected
     and the peer NID is selected. The router which has its nid on
     the peer NI list is preferred.

Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: I0d94d160bdb8c09c1cf467f1c155e14ab761a47e

LU-9121 lnet: Add info on udsp to lnetctl man page

Adding description of UDSP commands to lnetctl man page.
Listing each UDSP rule type with specific parameters.
Adding some examples of UDSP commands.

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: I1bde8a4da217c9ba89f31b5d9a1e7d26658bfb40
Reviewed-on: https://review.whamcloud.com/38698
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>

LU-9121 lnet: add show udsp command

Add the show udsp command in liblnetconfig and lnetctl

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: Ibddfa60a7b257b136a6be403e94d6f73fb444222
Reviewed-on: https://review.whamcloud.com/34580
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-9121 lnet: Delete a selection policy

This patch adds the function lustre_lnet_del_udsp()
which takes index of the udsp as input and calls
the IOC_LIBCFS_DEL_UDSP ioctl handler to
delete the udsp with that index value.

Change-Id: Ib81c79fd4e76f4e8fbd84709e08cb5d29a059e63
Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34553
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-9121 lnet: Add a selection policy

This patch adds the function lustre_lnet_add_udsp()
which marshals the input rules and makes an ioctl
call to add those rules in the lnet structures.

Change-Id: I5fa1e34d401b9eca58381e23cac2eb49c4dd3575
Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34529
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-9121 lnet: Add the userspace Marshalling API

Given a UDSP and a memory block of size
enough to hold a marshalled udsp, marshal
the UDSP pointed to by udsp into the memory block.

Change-Id: I67e9c2cc0d7f3dab1e968019d5ee546e6fefeba3
Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34513
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>

LU-9121 lnet: Add the userspace De-Marshalling API

Given a bulk containing a single UDSP, De-Marshalling
API demarshals it and populates a udsp structure.

Change-Id: Id2502dffc6fd259ad552c018b0cab415a12a8cfe
Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34501
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-9121 lnet: ioctl handler for get policy info

Add ioctl handler for GET_UDSP_SIZE and GET_UDSP

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: I4e188cac92ca98c26b26e00b8917ad3632d99d06
Reviewed-on: https://review.whamcloud.com/34579
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>

LU-9121 lnet: ioctl handler for "delete policy"

The ioctl handler for "delete policy" deletes
a policy with the given index value. It
returns 0 if a policy with that index is found
else it returns -EINVAL.

Change-Id: I037c6808e71e5c13ccf46c243145b0ce6f1229cb
Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34552
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>

LU-9121 lnet: Add the ioctl handler for "add policy"

The ioctl handler for "add policy" de-marshals the
udsp rules passed from userspace and then add the
rules if there is no copy of the same rules already
added. Apply the rules to the existing LNet
constructs.

Change-Id: Ia76ecdea6de94e1dbcfe71bfcb5a1753fb3c874d
Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34514
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>

LU-9121 lnet: Add the kernel level De-Marshalling API

Given a bulk allocated from userspace containing a
single UDSP, De-Marshalling API demarshals it
and populate the provided udsp structure.

Change-Id: I0a5a9148f8b2abc64284cb7780c37cbc8063b828
Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34488
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>

LU-9121 lnet: Add the kernel level Marshalling API

Given a UDSP, Marshal the UDSP pointed to by udsp
into the memory block that is allocated from userspace.

Change-Id: I325c977fd9c902f7ee31fceaaf07abf27ef7391e
Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Sonia Sharma <sharmaso@whamcloud.com>
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/34403
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>

LU-9121 lnet: Apply UDSP on local and remote NIs

When a peer net, peer ni, local net or local ni are created
apply the UDSPs in the system on these constructs.

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: I0ee427f1bc0d1c8d1ddc17072dcbd9442403fa0f
Reviewed-on: https://review.whamcloud.com/34355
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>

LU-9121 lnet: UDSP handling

This patch adds the following functionality:
1. Add UDSPs
2. Delete UDSPs
3. Apply UDSPs

- Adding a local network udsp: if multiple local networks are
available, each one can have a priority.
- Adding a local NID udsp: after a local network is chosen,
if there are multiple NIs, each one can have a priority.
- Adding a remote NID udsp: assign priority to peer NIDs.
- Adding a NID pair udsp: allows to specify local] NIDs
to be added to the list on the specified peer NIs. When
selecting a peer NI, the one with the local NID being used
on its list is preferred.
- Adding a Router udsp: similar to the NID pair udsp.
Specified router NIDs are added on the list on the specified
peer NIs. When sending to the remote peer, remote net is
selected and the peer NID is selected. The router which has
its nid on the peer NI list is preferred.
- Deleting a udsp: use the specified policy index to remove it
from the policy list.

Generally, the syntax is as follows
 lnetctl policy <add | del | show>
  --src: ip2nets syntax specifying the local NID to match
  --dst: ip2nets syntax specifying the remote NID to match
  --rte: ip2nets syntax specifying the router NID to match
  --priority: Priority to apply to rule matches
  --idx: Index of where to insert the rule. By default it appends
 to the end of the rule list

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: If588b78f3d5fce286270afb02e0e7683185d6898
Reviewed-on: https://review.whamcloud.com/34354
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>

LU-9121 lnet: select best peer and local net

Select the healthiest and highest priority peer and local net when
sending a message.

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: I42717e7fdc3226c6faa7c59c713f18422e27f2e5
Reviewed-on: https://review.whamcloud.com/34352
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>

LU-9121 lnet: Select NI/peer NI with highest prio

Modify the selection algorithm to select the highest priority
local and peer NI. Health always trumps all other selection
criteria

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: I487a706f4da30311d0bd59fe03f72dbe68a52425
Reviewed-on: https://review.whamcloud.com/34351
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>

LU-9121 lnet: Preferred gateway selection

Add mechanism for managing preferred gateway lists.
When selecting a route through a gateway, if there exists
a preferred gateway list for the destination peer, then choose
the preferred gateway. If there are multiple preferred
gateways, to make the selection, use in order of decreasing
importance: route priority, number of hops, number of available
tx credits on the associated lpni and route sequence counters.
If there are no preferred routes, select the best route
available using the same criteria.

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: If46920cf7b79aa8b211d6c0a35995edce9b1699a
Reviewed-on: https://review.whamcloud.com/34353
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>

LU-9121 lnet: foundation patch for selection mod

Add the priority and preferred NIDs fields in the lnet_ni,
lnet_net, lnet_peer_net and lnet_peer_ni. Switched
the implementation of the preferred NIDs list to list_head
instead of array, because the code is more straight forward.
There is more memory overhead due to list_head, but these lists
are expected to be small, so I chose code simplicity over memory.

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: I0c75855b736345c25e1604083eee2b65d38ef28d
Reviewed-on: https://review.whamcloud.com/34350
Reviewed-by: Chris Horn <chris.horn@hpe.com>

LU-9121 lnet: UDSP liblnetconfig structure def

Definition of the UDSP structures used to store parsed UDSPs

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: Ifc61a191de85d7904c70917d0c45648aecdd0310
Reviewed-on: https://review.whamcloud.com/34254
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>

LU-9121 lnet: UDSP storage and marshalled structs

Commit the structures which will be used by kernel space
to store the UDSPs. This commit also adds the IOCTL structures
which are used for marshalling the UDSPs between user and
kernel space.

Test-Parameters: trivial testlist=lnet-selftest,sanity-lnet
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: I490fa9a050cb1f8debc381be773cda4ce8abe29b
Reviewed-on: https://review.whamcloud.com/34253
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>

New release 2.14.0

Change-Id: I2eb99af8fbeaab80b6614e427b77949b1225b406
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-14345 misc: update e2fsprogs to 1.45.6.wc5

Update Changelog to reference new e2fsprogs release.

4aea203f LU-5949 e2fsck: call delete_inode() properly
8725134d LU-5949 e2fsck: simplify inode badness handling
71b74579 LU-14345 e2fsck: fix check of directories over 4GB

Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ie0845eeed410f6f9f8ef985342fc19d160aa8cb0
Reviewed-on: https://review.whamcloud.com/41433
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Reviewed-by: Peter Jones <pjones@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

New RC 2.14.0-RC3

Change-Id: I594b5c6d0da7f067bef69fa7a7027374d4434dd8
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-14424 Revert "LU-9679 osc: simplify osc_extent_find()"

It looks like there are performance regressions atttributed to this patch.

This reverts commit 80e21cce3dd6748fd760786cafe9c26d502fd74f.

Change-Id: I55e0abd50573dd82a9d216f9c3b01483f99c3223
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/41498
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>

New release candidate 2.14.0 RC2

Change-Id: Iad3d71e7dcf96173d192717ef4fef3f0dc12b051
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-13751 tests: remove read of changelog sanity 160j

sanity test 160j tries to read the changelog after one of two
client mounts is unmounted.  In this case, we can fail to read
the changelog and get a "Cannot send after transport endpoint
shutdown" error.

The intention of sanity test 160j is to check that
there is no LBUG due to missed obd device.  So, do not try to
read from the changelog after file system unmount.

Test-Parameters: trivial testlist=sanity env=ONLY=160j,ONLY_REPEAT=200
Signed-off-by: James Nunez <jnunez@whamcloud.com>
Change-Id: I1746a422b25d546b9aae38ae8438d9c08bce8827
Reviewed-on: https://review.whamcloud.com/41317
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>

LU-14355 ptlrpc: do not output error when imp_sec is freed

There is a race condition on client reconnect when the import is being
destroyed.  Some outstanding client bound requests are being processed
when the imp_sec has already been freed.
Ensure to output the error message in import_sec_validate_get() only
if import is not already in the zombie work queue.

Fixes: 135fea8fa9 ("LU-4423 obdclass: use workqueue for zombie management")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I4b431128e04f11b1e3ee7de47090af87538c3558
Reviewed-on: https://review.whamcloud.com/41310
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-14299 test: sleep to enable quota acquire again

sanity-quota test_61 fails with incorrect quota exceeded
errors because quota acquire will be disabled for 5 seconds
after edquot flag is set.  The test should introduce some
delay between the test of over quota and normal one.

Test-Parameters: trivial fstype=zfs testlist=sanity-quota env=ONLY=61,ONLY_REPEAT=20
Fixes: 530881fe4ee20 ("LU-7816 quota: add default quota setting support")
Change-Id: I8040ba960f32cf01cb7cee3a77c06ad4bd732f0e
Signed-off-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/41389
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>