git://git.whamcloud.com - fs/lustre-release.git/log

LU-2136 tests: build posix test suite on i686 client

This patch improves prep() in lustre/tests/posix/posix.cfg
to support building posix test suite on i686 client.

Test-Parameters: envdefinitions=SLOW=yes \
clientdistro=el6 serverdistro=el6 clientarch=i686 \
serverarch=x86_64 testlist=posix

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I7b2fdc5552f0d2eb5cc7ef38b1192ec5a4fdefcf
Reviewed-on: http://review.whamcloud.com/5831
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Wei Liu <wei3.liu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2598 tests: add version check code to sanity test 180c

This patch adds Lustre version check codes into sanity test
180c and obdfilter-survey to make the tests work with servers
which do not have the patch of commit 25170d2.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
clientdistro=el6 serverdistro=el6 clientarch=x86_64 \
serverarch=x86_64 serverjob=lustre-b2_1 serverbuildno=164 \
testlist=sanity,obdfilter-survey

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: Iad7dddf21b223d78efcc285b13b9381e5f0e0ccd
Reviewed-on: http://review.whamcloud.com/5902
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>

LU-3378 kernel: Kernel update [RHEL5.9 2.6.18-348.6.1.el5]

Update RHEL 5.9 patch to 2.6.18-348.6.1.el5.

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I5d75972265c22fae3410106ca443c1a1fb4d9466
Reviewed-on: http://review.whamcloud.com/6439
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3376 build: b2_1 copyrights need to be updated

This is the mass update from the updatacw.sh tool.

I added the previous copyright commit to the list
of commits to ignore by hand.

There were only a small number of changes so the
copyright updates are small as well.

Signed-off-by: Keith Mannthey <keith.mannthey@intel.com>
Change-Id: I62aad23e70ca1e8280f755539f02281ae5fd1da3
Reviewed-on: http://review.whamcloud.com/6410
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>

LU-3354 kernel: Kernel update [RHEL6.4 2.6.32-358.6.2.el6]

Update RHEL6.4 kernel to 2.6.32-358.6.2.el6.

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I120e766c5555472b5c856189dca9504d9a88e824
Reviewed-on: http://review.whamcloud.com/6381
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3187 obdecho: always set o_id non-zero

Always initialize the o_id field in obdecho requests so the server
doesn't get confused between o_id == 0 and f_seq == 0. This isn't
strictly necessary for protocol correctness (the server assigns
the actual OID in the end), but makes life a bit easier for cross
compatibility with FID-on-OST.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: Ifc318333df97326fb33366cd9f06d33e3f5139bb
Reviewed-on: http://review.whamcloud.com/6319
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2797 osc: GPF in osc_send_oap_rpc

Due to cl_loi_list_lock release in osc_send_oap_rpc(), this
leaves a window where lov_oinfo may be freed/poisoned between both
write/read steps/calls of osc_send_oap_rpc() in osc_check_rpcs(),
thus it needs to be refreshed.

Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Change-Id: Id3056c1fca91bade65b6d2b556b6b47a3d3d089e
Reviewed-on: http://review.whamcloud.com/6089
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>

LU-1468 o2iblnd: Support OFED-3.5 for o2ib

OFED is having new structure based on linux kernel code + backports
and packaging. Here is detailed information.
http://lists.openfabrics.org/pipermail/ewg/2011-December/017156.html

This patches are the lustre build improvements to support OFED 3.5,
3.x whatever future OFED release.

Porting from: Id4ffc39bc7fc24cc591bf6fb47e9b0e662993bda

Signed-off-by: Minh Diep <minh.diep@intel.com>
Change-Id: Icb64c3ab57be4bcbbdd8120aec6b58d02ff7253f
Reviewed-on: http://review.whamcloud.com/5688
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Brian J. Murrell <brian.murrell@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3219 ost: Ensure dirty flushed on fiemap ioctl

Data corruption is possible if cp(coreutils) uses FIEMAP to obtain
data holes in shared file, since there could be dirty cache on
other clients which hasn't been flushed back.

To ensure all the dirty on remote clients being flushed back on
fiemap ioctl, we'd acquire ldlm lock on server side for fiemap,
unless the local client (which invoke fiemap) has cached lock.

Xyratex-bug-id: MRP-1001
Signed-off-by: Artem Blagodarenko <artem_blagodarenko@xyratex.com>
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I6debe07acd015eb990d5b3748e50cc88d61065f8
Reviewed-on: http://review.whamcloud.com/6377
Tested-by: Hudson
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-2368 recovery: fix for obd_stale_clients counter

class_fail_export() occured during recovery process, when
the mds connect to ost from new IP(failover happened), and cause
perpetual recovery.
- in class_fail_export() update obd_stale_clients counter if
recovery is in progress.

Signed-off-by: Mikhail Pershin <tappro@whamcloud.com>
Signed-off-by: Alexander Boyko <alexander_boyko@xyratex.com>
Xyratex-bug-id: MRP-738
Change-Id: I2b2fe5853a9b0713a0a9357713f612b331505c6f
Reviewed-on: http://review.whamcloud.com/4641
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Keith Mannthey <keith.mannthey@intel.com>
Reviewed-by: Keith Mannthey <DELETED>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3279 changelog: fix CL_LAYOUT, accept all types

In order to avoid compatibility issues with older ChangeLog consumers,
change the new CL_LAYOUT record to use the value previously assigned
to CL_IOCTL.  The CL_IOCTL type was never used anywhere, and it didn't
really make any sense as a ChangeLog record, since it could really
mean anything at all.

The changelog_show_cb() function is renamed to changelog_kkuc_cb(),
since it is really about consuming the ChangeLog and passing it up to
the kernel-user-coms interface.  At some point we should consider to
implement a DBUS-based ChangeLog interface as well.

The changelog_kkuc_cb() "sanity check" on cr_type was removed, since
there is no reason the client kernel needs to know every record type
that is being passed to userspace.  It is up to the client tool to
determine what records that it can process.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: Id8e6471827a89d8c62808e6b3cfd9693d63ebbe5
Reviewed-on: http://review.whamcloud.com/6335
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: jacques-Charles Lafoucriere <jacques-charles.lafoucriere@cea.fr>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3218 kernel: Kernel update [RHEL5.9 2.6.18-348.4.1.el5]

Update RHEL5.9 patch to 2.6.18-348.4.1.el5.

Signed-off-by: yang sheng <yang.sheng@intel.com>
Change-Id: Ib9741a9323fff58926a12b8f46ddae43950ea964
Reviewed-on: http://review.whamcloud.com/6279
Tested-by: Hudson
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3029 osd-ldiskfs: clear old FMODE_32BITHASH for readdir

On MDS, the RPC service thread's "osd_thread_info" will be reused
without completely reset when switches from one RPC processing to
another RPC processing.

Some old client may not claim "OBD_CONNECT_64BITHASH" when connect
to the MDS, then it will be regarded as 32-bit client. For readdir
RPC from such old client, the MDS will use 32-bit dirhash, the RPC
service thread's "osd_thread_info::oti_it_ea::oie_file::f_mode" is
set as "FMODE_32BITHASH", which will not be dropped until restart.

If some RPC service threads (for readdir) are "FMODE_32BITHASH",
but some NOT, then for new client, which support 64-bit dirhash,
it may get trouble when traverses large directroy as following:

1) The first readdir RPC is served by the RPC service thread1,
   which is marked as "FMODE_32BITHASH" because it ever served
   readdir RPC from old client. So the thread1 still generates
   32-bit dirhash (major hash only) for the new client readdir.

2) The new client triggers another readdir RPC for the same dir
   with the 32-bit dirhash as cookie, which was returned by the
   thread1.

3) This time, the readdir RPC is served by another RPC service
   thread2, which is NOT marked as "FMODE_32BITHASH". Then the
   thread2 explains the readdir cookie as 64-bit dirhash, It's
   wrong. So the thread2 can NOT locate the position correctly.

So we need to clear some fields in "osd_thread_info" to avoid
to be reused when switch from one RPC processing to another.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I1b0468fe1d3bed5b3563b17c9722009514bbea2e
Reviewed-on: http://review.whamcloud.com/6176
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Ned Bass <bass6@llnl.gov>
Reviewed-by: Prakash Surya <surya1@llnl.gov>

LU-1291 mds: kernel crashed while umount mds

The obd_notify needs protect by mds_notify_lock. Else LOV
maybe gone while __mds_lov_synchronize be called.

Signed-off-by: yang sheng <yang.sheng@intel.com>
Change-Id: Ia96e0df2e7c8489741c8cbfe9f4fa35ecdb8f453
Reviewed-on: http://review.whamcloud.com/5789
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3216 kernel: kernel update [RHEL6.4 2.6.32-358.6.1.el6]

--update RHEL6.4 kernel to 2.6.32-358.6.1.el6
--resolved ldiskfs patches conflict
--update kernel config files
--include LU-2967 mentioned patch

Signed-off-by: yang sheng <yang.sheng@intel.com>
Change-Id: I462035db390c05712875c6c90d88de94153278ec
Reviewed-on: http://review.whamcloud.com/5504
Tested-by: Hudson
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2888 osd: OI implemented within OSD

This patch is mostly port of master commit f2f09b6e, which uses
ldiskfs directly to implement OI, so that be align with 2.4 lustre.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: If83cd61b6c90e195a7a364153819bbec8aa0f9c1
Reviewed-on: http://review.whamcloud.com/5731
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Hudson
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-506 tests: check lustre.conf for modprobe

The newer userspace modprobe requires that the /etc/modprobe.d/
files end in ".conf", so that there is no confusion between config
files that are supposed to be parsed, and files that are left
after editing (e.g. lustre.conf.orig or lustre.conf~).

Add a check for /etc/modprobe.d/lustre.conf to get lnet module
parameters during testing.

Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ie408f5334b2570f13e225038488df06ff7b524c4
Reviewed-on: http://review.whamcloud.com/1877
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

Lustre 2.1.5-RC1

Change-Id: Ibb844135405ae52ec9be02a55b878978007005dc
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2694 test: fix is_empty_fs in t-f

The original is_empty_fs is incorrect, which can cause unexpected
error when running lfsck.sh, because lfsck sets the max error
level based on whether filesystem is empty.

The test directory of lfsck.sh contains some files referencing
same object, which could cause error when removing the directory
on test cleanup.

This patch is backported from commit 48c78f6 and 2df0104.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
testlist=lfsck

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I869ce947f5dc3a5e1c8923017bcfcc6902015927
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/5762
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2857 tests: disable sanity test_76 for LOD/OSP

Client inode cache shrinking does not work for LOD/OSP anymore.
Disable test when running in interop mode with 2.4 servers.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I01286dcefea0df8dd06869e5dcabcf690e3ebbe5
Reviewed-on: http://review.whamcloud.com/5755
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2517 kernel: Kernel update [RHEL6.3 2.6.32-279.19.1.el6]

Update RHEL6.3 kernel to 2.6.32-279.19.1.el6.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
clientdistro=el6 serverdistro=el6 testgroup=review

Signed-off-by: yang sheng <yang.sheng@intel.com>
Change-Id: I683327c6dd0f8ba383c9bcbb215cbce8f4b88e22
Reviewed-on: http://review.whamcloud.com/5710
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2864 test: waiting import state in fail().

Anyway, There still has a rare chance that the request meet
a invalid import after fail() return. So we should waiting
import restore to a certain state and doing next operation.
Add wait_import_state_mount() to check import state while
client has a mount point.

Signed-off-by: yang sheng <yang.sheng@intel.com>
Change-Id: I0b4a79c877bd6f240020d629fd672ba87ee6c99b
Reviewed-on: http://review.whamcloud.com/5743
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-588 ldiskfs: Don't release super block buffer_head too early

If the super block buffer_head is released prior to MMP stopping
kmmpd can get a zeroed buffer_head and exit. The following code
causes kmmpd to exit when the s_feature_incompat is zero:

if (!(le32_to_cpu(es->s_feature_incompat) &
LDISKFS_FEATURE_INCOMPAT_MMP)) {
ldiskfs_warning(sb, "kmmpd being stopped since MMP feature"
" has been disabled.");
LDISKFS_SB(sb)->s_mmp_tsk = NULL;
goto failed;
}

A deadlock can occur with the kthread_stop_lock mutex because
ldiskfs_put_super calls kthread_stop on an already stopped thread
(kmmpd) so it waits for completion of the thread stopping before
releasing the kthread_stop_lock.

This is the result of a race with the kmmpd thread setting s_mmp_tsk
to NULL and ldiskfs_put_super in another thread to checks s_mmp_tsk
for NULL prior to calling kthread_stop.

Signed-off-by: Jeremy Filizetti <jeremy.filizetti@gmail.com>
Change-Id: Ia15d8ff829705a5d51dea4f86e40ba7c5745a9c5
Reviewed-on: http://review.whamcloud.com/3172
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Yu Jian <yujian@whamcloud.com>

LU-2598 tests: check npages in filter_preprw()

This patch checks the number of bulk I/O RPC pages with
PTLRPC_MAX_BRW_PAGES in filter_preprw() to avoid
LASSERT(iobuf->dr_npages < iobuf->dr_max_pages) occurring
while larger I/O size is specified.

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I1b85866045d753359c420ed9dfd8703eec14f18e
Reviewed-on: http://review.whamcloud.com/5715
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>

LU-2592 tests: do not remove $TMP/*active in rpc.sh

The $TMP/*active files record the current active facets info
under failover test configuration. They are removed from
init_test_env() initially. If failover tests use do_rpc_nodes()
which performs rpc.sh and in which init_test_env() is used,
then the active facets info will be lost during the testing.

This patch fixes the above issue by introducing an RPC_MODE
variable which controls that the $TMP/*active files will not
be removed from init_test_env() in rpc.sh.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
clientcount=4 osscount=2 mdscount=2 austeroptions=-R \
failover=true useiscsi=true testlist=replay-vbr

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I7f23c58e06e76d23ee6be8f779218e8d8ff32ae1
Reviewed-on: http://review.whamcloud.com/5126
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Wei Liu <wei3.liu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1189 tests: run save_lustre_params() on facets

This patch fixes save_lustre_params() and restore_lustre_params()
to get and set Lustre parameters through do_facet() instead of
do_node(). This will fix the issue that restore_lustre_params() still
tries to restore the saved params on the original node which will
likely become inactive under failover test environment.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
clientcount=4 osscount=2 mdscount=2 austeroptions=-R \
failover=true useiscsi=true \
testlist=replay-vbr,replay-dual

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: If73adb84c4f1b2f6d6667a3865f2cc0221f0e687
Reviewed-on: http://review.whamcloud.com/5627
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2928 build: update commit exclusion list for 2.1.5

The mass update should be excluded from future
looks at the copyrights. Copyright updates are
not a reason to update a copyright.

Signed-off-by: Keith Mannthey <keith.mannthey@intel.com>
Change-Id: Ibf15639882dd2f3f13593f4ad7c4cfd7171f6b7d
Reviewed-on: http://review.whamcloud.com/5651
Tested-by: Hudson
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2928 build: update the copyrights for b2_1

These are the changed produced by the updatecw.sh tool
and manual edits to lustre/liblustre/llite_cl.c and
lustre/obdclass/hash.c .

With this commit all files report they are upto date and
have proper Intel Copyrights where they are deserved.

Signed-off-by: Keith Mannthey <keith.mannthey@intel.com>
Change-Id: I2c01062c612f9f968cabc2a890a094a2b2ceb36c
Reviewed-on: http://review.whamcloud.com/5650
Tested-by: Hudson
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2928 build: update updatecw.sh script

Sync this tool with the curren master version.
There were some minor improvments since the last
2.1.X release.

Signed-off-by: Keith Mannthey <keith.mannthey@intel.com>
Change-Id: Iba98883d8b1f76712b349dd40f7974272445529c
Reviewed-on: http://review.whamcloud.com/5649
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2922 ldiskfs: depend on e2fsprogs-1.42.6.wc2

This patch updates lustre-ldiskfs.spec.in to depend on
a newer e2fsprogs-1.42.6.wc2 release instead of the
old 1.41.12.2.ora1 version.

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I58828fa5427e5f50cfb3b285f980e946e7eb62e7
Reviewed-on: http://review.whamcloud.com/5623
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Brian J. Murrell <brian.murrell@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-2620 tests: Flush changes from test 5 in replay-ost-single

Use wait_delete_completed_mds to flush out and wait for all the
effects on kbytesavail from test 5 to finish before test 6 starts.

Test-Parameters: testlist=replay-ost-single envdefinitions=SLOW=yes

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: Id37cf23402dd609e6743c36a1aff8e71dd246776
Reviewed-on: http://review.whamcloud.com/5042
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2012 tests: disable replay-dual test_14b

Disable replay-dual.sh test_14b until OST gap handling is fixed.

The test is modified to make the pass/fail result more clear for when
it is re-enabled, since any blocks that are allocated on the OST while
the test is running will cause it to fail. This could happen because
of updated config llog, OI updates, etc. Allow some small margin of
space to be allocated on the OST before declaring failure. To make
failed the orphan handling totally clear, ensure the orphan object is
much larger than the margin of error.

This patch is ported from commit b72c4e8.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
clientdistro=el6 serverdistro=el6 testlist=replay-dual

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: Ie48791c110c54fd6fae45ae7321370e35c16822e
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/5571
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>

LU-2861 tests: use mkfs_opts() instead of *_MKFS_OPTS

This patch fixes conf-sanity.sh to use mkfs_opts()
instead of *_MKFS_OPTS for adding new server targets.

This patch also ports the conf-sanity.sh changes from
commit 648b69c to support the interoperability with
2.4 server.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
clientdistro=el6 serverdistro=el6 \
serverjob=lustre-master serverbuildno=1297 \
testlist=conf-sanity

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I62d99ef70ada044eeda6f873fb835e603dc48df2
Reviewed-on: http://review.whamcloud.com/5587
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Li Wei <wei.g.li@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2769 kernel: Kernel update to RHEL6.3 2.6.32-279.22.1.el6

Update kernel for 2.6.32-279.22.1.el6.

Signed-off-by: yang sheng <yang.sheng@intel.com>
Change-Id: I05c5270230963d0dde4fb0add0237a1682ecbefd
Reviewed-on: http://review.whamcloud.com/5529
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1571 mdt: Do not update xid for open replay req

Do not update last_xid for open replay req,
otherwise the following resend(after replay)
can not be matched with correct xid.

Remove unnecessary mti_transo zero check in
mdt_empty_transno.

Original Change-Id: I2a05f3ac05b301ae31641a1dc51f8c4eed96427d
Original Reviewed-on: http://review.whamcloud.com/3195
cherry-picked from 3402e6c6517cc15f159b19bf4462279285cd83f2

Change form Original:
Fixed mutex change in mdt_empty_transno()

Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Signed-off-by: wang di <di.wang@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I9a26b25ae88e0ae2299975ff4abce9b47289fb17
Reviewed-on: http://review.whamcloud.com/5003
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-971 mdt: fix open resent issue with last_xid

req_xid_is_last() compares xid with one in client data,
but in master it is not updated in case of open without create.
Patch updates export lcd with last xid, transno and rc for open
requests.

Original Change-Id: I9e769fecc1e93c7a556d4b9eeadc80a70369e901
Original Reviewed-on: http://review.whamcloud.com/1930
Cherry-picked ed5e47ac21b02abb16b10446dac7dae722259f0a

Signed-off-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Li Wei <liwei@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Change-Id: I015aa109bef65f6207f462358e40cae6e6092cc7
Reviewed-on: http://review.whamcloud.com/5002
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>

LU-744 obdclass: revise cl_page refcount

By holding a refcount for radix tree, we can remove a couple of
spin lock and atomic operation for cl_page.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Gregoire Pichon <gregoire.pichon@bull.net>
Change-Id: I4d8dc04b9b4ef066b54a90be3cfbd0c09d96c320
Reviewed-on: http://review.whamcloud.com/5429
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2591 lov: race between ptlrpc_rcv and umount/umount thread

The race which is refered here happens in the following scenario

1) mount runs but fails to communicate with some OSTs. Then the
   import objects which represent the OSTs are registered to
   a pinger list.
2) pinger succeeds to communicate with, at least, ONE OST. Then
   ptlrpc_rcv calls lov_set_osc_active to activate the OST and
   holds lov_refcount.
3) For some reason ... possibly mount finally fails or umount runs,
   ll_put_super is called
4) ll_put_super tries to disconnect all the OSTs with
   lov_disconnect and this func calls lov_del_target to set all
   the OSC's target->ltd_reap flags in order for lov_putref to
   handle all of them.
5) ptlrpc_rcv thread puts lov_refcount and if lov_refcount becomes
   0 here, the thread has to disconnect all the OSCs whose ltd->reap
   has been set.
6) Some OSCs' imports have still been LUSTRE_IMP_CONNECTING state
   because of (2), so ptlrpc_rcv thread has to wait for these
   import states to be changed to non-recovery states, such as FULL,
   CLOSED or DISCON at ptlrpc_disconnect_import.

Now that ptlrpc_rcv thread is waiting for the import states to be
changed to non-recovery states but ptlrpc_rcv is the one who is
supposed to change a recovery state to a non-recovery state, So
ptlrpc_rcv must hung, And mount/umount thread which has called
ll_put_super has to also wait for changing state at
ptlrpc_disconnect_import, so that thread must hung too.

Signed-off-by: Hiroya Nozaki <nozaki.hiroya@jp.fujitsu.com>
Change-Id: I1d74967b883f079aafd11454dca44960a8f7a3f2
Reviewed-on: http://review.whamcloud.com/5527
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2183 quota: update tests for DNE

Update test_13 of sq since the osp is changed to lwp in DNE.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I4ce0522aa1de46a02d2084f329471551fce3ee90
Reviewed-on: http://review.whamcloud.com/5240
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1727 lprocfs: fix an old assertion

lprocfs_counter_add() can be called in interrupt context, as a memory
allocation can trigger memory shrinker call ldlm_pool_shrink(), which
calls lprocfs_counter_add().

Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: I0f5d3e07c234f7d615ca87761a7bd6367e72beac
Reviewed-on: http://review.whamcloud.com/3729
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2072 ldiskfs: Leak of s_mb_prealloc_table in ldiskfs

ext4-prealloc-rhel[5,6] patches adds kmalloc'ed
sbi->s_mb_prealloc_table field which is never freed
in [ldiskfs,ext4]_mb_release()/ext4-vmalloc-rhel[5,6].patch

port of master patch 536464e609c46a3950d7742bc3860b87d6bfc726

Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: I0acd9f62d6ffc516540d707830b40c0bd211a43b
Reviewed-on: http://review.whamcloud.com/5474
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1484 lprocfs: enable HAVE_PROCFS_USERS for rhel kernels

For rhel kernels, given only the kernel-devel rpm it is not possible
to tell if proc_dir_entry_aux is defined, so assume it is. It has
been included since late in the 5.x release cycle.

Originally Change-Id: If17ecb0902ec90a1af6228d2b9b1b72bc68a6672

Test-Parameters: envdefinitions=SLOW=yes clientdistro=el5 serverdistro=el5 clientarch=x86_64 serverarch=x86_64 testlist=recovery-small
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: I7cde85592759abe020ff31fbf4840717b3e95bef
Reviewed-on: http://review.whamcloud.com/5468
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2601 kernel: kernel update [RHEL5.9 2.6.18-348.1.1.el5]

Update RHEL5.9 kernel to 2.6.18-348.1.1.el5.
Adding scsi/fc_compat.h to lustre-lnet.m4 and o2iblnd.h to build
o2ib modules.

Signed-off-by: Shuichi Ihara <sihara@ddn.com>
Signed-off-by: yang sheng <ys@whamcloud.com>
Change-Id: I1d66d12f64238cb410288715ff389ab96b15f0f3
Reviewed-on: http://review.whamcloud.com/5451
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2824 mds: don't deactivate OSC if osc_create returns -EBUSY

During MDS<->OST orphan recovery, osc_create() will likely return
-EBUSY while OSCC_FLAG_SYNC_IN_PROGRESS flag is still set
due to slow ll_mdt service thread. The __mds_lov_synchronize() will
then deactivate the OSC, which will cause mds_create_objects() get
-EIO error.

This patch fixes the above issue by checking the return value of
mds_lov_clear_orphans(). If it's -EBUSY, then do not mark the
OSC as inactive.

Test-Parameters: envdefinitions=DURATION=14400 clientdistro=el5 serverdistro=el5 clientarch=x86_64 serverarch=x86_64 clientcount=4 osscount=2 mdscount=2 austeroptions=-R failover=true useiscsi=true testlist=recovery-mds-scale
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I7b3376e3b6087dd35c5f0e33e98b1e538ba711fc
Reviewed-on: http://review.whamcloud.com/5462
Tested-by: Hudson
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1666 obdclass: reduce lock contention on coh_page_guard

Define a per-page spinlock to get and put a cl_page instead of
grabbing per-object lock coh_page_guard.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Gregoire Pichon <gregoire.pichon@bull.net>
Change-Id: I3c63f0e33f3d4d44fbb5de06bcd0670e76b872bd
Reviewed-on: http://review.whamcloud.com/5428
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2775 fid: allow FID-on-OST in fid_seq_is_mdt()

The old 2.1 LASSERT_SEQ_IS_MDT() macro used fid_seq_is_mdt() in
several places to verify that a FID was "sane" for where it was
being used. However, fid_seq_is_mdt() only checked FID_SEQ_OST_MDT0
through FID_SEQ_OST_MDT9 values, which immediately broke 2.1 clients
when FID-on-OST was enabled on 2.4 MDTs.

The use of FID-on-OST will be limited to only DNE MDTs for 2.4, but
clients should not LASSERT() on a value that it got from the network.
The use of LASSERT_SEQ_IS_MDT() is removed from clients, since this
is "validating" data from the network. Even with this removal, the
fid_seq_is_mdt() macro is updated to allow FID_SEQ_NORMAL sequences,
as in 2.4, in case it is needed elsewhere.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I74bc9198799045b8bd91510cb45e8f8760300c1e
Reviewed-on: http://review.whamcloud.com/5304
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2555 fsfilt: fix race in ldiskfs_ext_new_extent_cb

In ldiskfs_ext_calc_credits_for_insert(), we should use the 'depth'
stored in the 'path' instead from inode, because the extent tree
could have been changed when ldiskfs_ext_calc_credits_for_insert()
is called (by ldiskfs_ext_new_extent_cb()).

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: Ie28f7dfc393c8feb8336f6b391cebb0adc814283
Reviewed-on: http://review.whamcloud.com/4977
Tested-by: Hudson
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2555 ldiskfs: fix ext4_mb_add_n_trim()

In ext4_mb_add_n_trim(), lg_prealloc_lock should be taken
when changing the lg_prealloc_list.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I4f3a4b8bc9697b7ccdeade43f5bfd10960533fca
Reviewed-on: http://review.whamcloud.com/4978
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-2634 ldiskfs: clear extent flag for fast symlink

Fast symlinks on MDT filesystems formatted with the "extents" feature
are created with EXT4_EXTENTS_FL. e2fsck considers this a corruption
and deletes the symlink.

The patch is to clear extent flag in osd_ldiskfs_writelink() just as
ext4_symlink() does.

Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: Ic0c711cc82de140f86baa71bacd1721fd4a19e07
Reviewed-on: http://review.whamcloud.com/5458
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-2657 recovery: don't delete objects

In mds_lov_update_objids(), when an data object id gap is
detected during recovery, it'll delete all the objects in
the gap, which isn't quite correct, because we can't guarantee
that the id is always increased by the transno order,
furthermore, that could also bring big trouble when the
lov_objid file was removed manually (to rebuild the corrupted
lov_objid file).

Fix type defect in filter_recov_log_unlink_cb(), where
oa->o_id should be increased by each loop cycle.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I7553614f30fe60675b8dab405d9100cb63069b23
Reviewed-on: http://review.whamcloud.com/5206
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1714 lnet: Properly initialize sg_magic value

When the CONFIG_DEBUG_SG flag is enabled in the kernel, we must ensure
the sg_magic field is properly initialized. Otherwise, internal kernel
assertions will fail when trying to verify this field. As a result,
certain calls to sg_* function had to be changed or inserted to ensure
the sg_init_table function would be called, initializing the magic
value. Also, we need to ensure this value isn't zeroed out in the
kiblnd_setup_rd_kiov function.

port of master patch d3972843ae3ba829f821bd135d5ba97d4b0356f3

Signed-off-by: Prakash Surya <surya1@llnl.gov>
Signed-off-by: Alexander Boyko <alexander_boyko@xyratex.com>
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: I82b19348bcf981b5d843ae4bf07cbd2a6d51b701
Reviewed-on: http://review.whamcloud.com/5452
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2638 ldiskfs: update dotdot entry for rename

On master, when rename a directory, its old dotdot entry will
be removed firstly, then insert the new dotdot entry, and try
to append FID-in-dirent. But the space for dotdot entry maybe
not enough to hold the new dotdot with FID-in-dirent, such as
the MDT device restored from file-level backup, or the device
is upgraded from 1.8. If we do not move data in the directory
block, the FID-in-dirent will overwrite the dx_root and cause
the directory crashed. Currently, we do not want to introduce
complex logic to handle directory data moving, instead, under
such case, ignore the FID-in-dirent for the new dotdot entry,
and just insert the new dotdot entry.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: Id38f5136fb2bc607344c37ca10a32e63ab427e3d
Reviewed-on: http://review.whamcloud.com/5179
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>

LU-1486 tests: fix sanityn test_45g() to add error check

This patch fixes sanityn test_45g() to add error check for
"stat must fail".

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Iff65f2684244ab58e0d0cab68db0795f5604187e
Reviewed-on: http://review.whamcloud.com/3826
Tested-by: Hudson
Reviewed-by: Wei Liu <wei3.liu@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-1448 llite: Prevent NULL pointer dereference on disabled OSC

When a file system is mounted with a disabled OSC reading the import
information from the proc file system can result in a NULL pointer
dereference. The Lustre import on a disabled OSC with remain
in the LUSTRE_IMP_NEW state and imp_connection will remain NULL.

Signed-off-by: Jeremy Filizetti <jeremy.filizetti@gmail.com>
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: Ia6df51a36efbcd5a7fc7668bb23455b253ae4855
Reviewed-on: http://review.whamcloud.com/2995
Reviewed-by: Liang Zhen <liang@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-582 test: fix bug in get_ost_param in sanity.sh

in get_ost_param(), the ouput should be gl_sum, which is
the sum of 'gl' in all OSS.

Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Change-Id: I51d3aa591c3ca875f8a559974cf05d250b0c032a
Reviewed-on: http://review.whamcloud.com/3693
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Jian Yu <jian.yu@intel.com>
Tested-by: Hudson
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2550 osc: set resend count properly

The resend count of new io request should be set properly
in osc_brw_redo_request().

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: If02f0f9abeb316f0718d34d0d1858985209acbb0
Reviewed-on: http://review.whamcloud.com/5184
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-2526 quota: writer proper size in test_9

Write proper size (4.5G) in test_9 of s-q.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I3dbb9f9b5e0b053529b97b90156dea1b75e366b1
Reviewed-on: http://review.whamcloud.com/5185
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-2309 tests: ignore unknown configuration param

This patch improves conf-sanity test 42 to verify that
invalid sys config param should not prevent client/server
from mounting.

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I083f0c143753ff2b990c91319db057ef8221992b
Reviewed-on: http://review.whamcloud.com/5031
Tested-by: Hudson
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1887 ptlrpc: grant shrink rpc format is special

Grant shrink rpc is packed in OST_SET_INFO request, but its format is
different from others.

Change-Id: I24fef30504847d26699df427f787d7c1d594c7c5
Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/4269
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1331 changelog: allow changelog to extend record

Previously rename is split into two records, this isn't natural and
makes replication hard because rename needs two records, but they
may not be consecutive.

To solve this, allow Changelog to extend record, add fields sfid/spfid
in changelog record to store source/parent fid, and pack both source
and target name into the record.

Added changelog flag CLF_RENAME_LAST: rename unlink of last hardlink
of the target.

Added macro HAVE_CHANGELOG_EXTEND_REC: liblustreapi receives all
changelog records in changelog_ext_rec format.

Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: I97b990687726e45e661cfb11609f80132c84825d
Reviewed-on: http://review.whamcloud.com/3835
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Thomas LEIBOVICI - CEA <thomas.leibovici@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-676 tests: machinefile option for mpirun via a variable

Not all MPI implementations pass the host file to mpirun via the same
option. Common options are -machinefile and -hostfile.

This problem can be resolved by using a variable MACHINEFILE_OPTION
instead. A default value is assigned if the variable not defined.

Signed-off-by: Jay J Lan <jay.j.lan@nasa.gov>
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: I3362a6e62a27318cff733aea2f99b1356b3ff02e
Reviewed-on: http://review.whamcloud.com/3874
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-1234 dcache: don't drop invalid dentry arbitrarily

This is a backport of part of LU-506 dcache scalability support:
* remove super hack d_rehash_cond(), and treats DCACHE_LUSTRE_INVALID
  similar to DCACHE_DISCONNECTED, therefore dentry doesn't need to
  be dropped and rehashed frequently.
* .lookup(LOOKUP_CREATE) calls d_add() dentry directly, and .create
  only needs to d_instantiate() this dentry.
* other cleanups.

Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: Ie169bc7e763e6891084999041aac9f62c8dee9f0
Reviewed-on: http://review.whamcloud.com/2400
Tested-by: Hudson
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Peng Tao <bergwolf@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2237 tests: rename conf-sanity::test_61 as test_65

To make it to be consistent with master branch.

Signed-off-by: Fan Yong <yong.fan@whamcloud.com>
Change-Id: I5cf281c74dbe66f742353843c813f6724340442a
Reviewed-on: http://review.whamcloud.com/4456
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>

LU-1018 tests: reduce runtime value for compilebench

We are reducing the cbench_IDIRS and cbench_RUNS to 2.
These variables can be adjusted to other runs beside
sanity check.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes testlist=parallel-scale
Signed-off-by: Minh Diep <minh.diep@intel.com>
Change-Id: I63694bd26cdf8cecdbabebf601c014718870af21
Reviewed-on: http://review.whamcloud.com/5053
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Keith Mannthey <keith.mannthey@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-2467 protocol: Add OBD_CONNECT_PINGLESS

Reserve a bit for OBD_CONNECT_PINGLESS, which indicates a client is
capable of suppressing keep-alive OBD_PINGs. If granted by a server,
it means the server does not require (but still allows) pings.

Change-Id: Ie10f505badd4250582656295eac2fd6844bb76b3
Signed-off-by: Li Wei <wei.g.li@intel.com>
Reviewed-on: http://review.whamcloud.com/5232
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-1770 idl: reserve new OBD_CONNECT feature flags

Reserve protocol feature flags on b2_1 to match the definitions
on the master branch:
- OBD_CONNECT_FLOCK_OWNER
- OBD_CONNECT_LVB_TYPE
- OBD_CONNECT_NANOSEC_TIME
- OBD_CONNECT_LIGHTWEIGHT
- OBD_CONNECT_SHORTIO

The OBD_CONNECT_NANOSEC_TIME was originally added in commit
596816513b0603cb159353e1444c97d8bc3bcacf (on master it was
528feb1793e744c63c2cd137d11418c513eb92d), but this was changed
in master commit f100d508311d16a09040a057438ee06fd9e6d7fe due
to a conflict with OBD_CONNECT_FLOCK_OWNER in a customer patch.

As of the time of this patch, the nanosecond timestamp flag
is not yet in use on any branch, so changing the flag value
does not cause any protocol incompatibility.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: Iae34aeec54cdfeb39fe932aa132a3cac982d1ee5
Reviewed-on: http://review.whamcloud.com/5162
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>

LU-20 ldiskfs: remove spurious warning message

Remove the spurious warning message that we added for the missing
extents option since the MDS never gets the extents option enabled,
nor do we want it to be enabled for the MDT filesystem either.

ldiskfs_fill_super: extents feature not enabled on this
filesystem, use tune2fs

Original-Commit: 6541ea42978e73e0262273f6bd5ce0b71689fdc2

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I9d896258b89ec3db528db59094572daac8dccab0
Reviewed-on: http://review.whamcloud.com/4028
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>

2.1.4rc2

Change-Id: Ie51d791502006d094219643e9b0286cba8ebe2dd

LU-1526 tests: Support for MDS-initiated OST_DESTROYs

This patch is backported from commit af5f388 of LU-1303 to
support interoperating with 2.4 server.

The patch makes sure the tests work with MDSs that destroy OST
objects asynchronously on behalf of clients.

Signed-off-by: Li Wei <liwei@whamcloud.com>
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I375d33b53e88a1649b68c1a3999d273d759d17b7
Reviewed-on: http://review.whamcloud.com/4832
Tested-by: Hudson
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-1526 tests: new sanity-quota tests

This patch is ported from commit d6f2a9f to add new sanity-quota.sh
for the interoprability testing with new quota architecture,
meanwhile, the old sanity-quota.sh is reserved for the interoprability
with old server.

The patch also contains the fixups for LU-2174, LU-2283, LU-2284
and LU-2329.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes testlist=conf-sanity,sanity-quota
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I334890a4c7a233311e8f4075a8bf8dd904563525
Reviewed-on: http://review.whamcloud.com/4767
Tested-by: Hudson
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-691 tests: Fix OST index errors in test suite

Several tests run do_facet ost which assumes ost is ost0 which does
not exist according to the way the test suite works. This patch
address several areas where the wrong ost index is used. This patch
also introduces some short hand functions to get OST properties
from the index. Those functions ensure ost1, ost10 and ost100 are
seen as different which is not always the case.

This patch is backported from commit 8afd35c of LU-691.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes testlist=lfsck,sanity,replay-single
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I97d9554a99be25db0e4f550bf8071ddc981dddc4
Reviewed-on: http://review.whamcloud.com/4821
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-1526 tests: Handle OFD procfs changes

In order to interop with 2.4 server, the following procfs entry
changes need to be handled:

- obdfilter.*.mntdev -> osd-*.*.mntdev
- obdfilter.*.<cache_related> -> osd-*.*.<cache_related>

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes testlist=lfsck,sanity
Signed-off-by: Li Wei <liwei@whamcloud.com>
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I0bc3fc5a2f6d3a3ff6653175a8b00cd51d48534c
Reviewed-on: http://review.whamcloud.com/4783
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Hudson
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2483 build: Update Copyrights for b2_1

In prepration of the release of 2.1.4 the Copyrights need to be
updated.

As I have bound the script to the changes so that it will not be lost
from the changes it helped make.

The accounting works for both Intel and Whamcloud email addresses.
See: lustre/ldlm/ldlm_request.c
It is able to add new Copyrights when aproprate
See: lustre/ptlrpc/layout.c
I didn't find any Intel, Inc signed files.
Two man pages were edited by hand and are included in this commit.

Signed-off-by: Keith Mannthey <keith.mannthey@intel.com>
Change-Id: I0c6daac2c6bfd2b6d5ed1f14bb5993519d099b91
Reviewed-on: http://review.whamcloud.com/4857
Tested-by: Hudson
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>

LU-1484 kernel: pass RHEL5 build for 2.6.18-308

For vanilla kernel, proc_dir_entry::deleted and ::pde_users co-exists
from 2.6.23 to 2.6.23.17.

For some RHEL5 kernels, it defines co-existings
proc_dir_entry::deleted and proc_dir_entry_aux::pde_users.

Signed-off-by: Bobi Jam <bobijam@gmail.com>
Change-Id: I107e15e6e7d41680c99b324c35ddac0a7c3716e6
Reviewed-on: http://review.whamcloud.com/4794
Tested-by: Hudson
Reviewed-by: Peng Tao <bergwolf@gmail.com>
Reviewed-by: Stephen Champion <schamp@sgi.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-797 tests: improve test_23b of ost-pools.sh

- fail the test once the written space is greater than the
available space.
- use it's own maxfree limit, and increase the limit a lot.
- terminate the test when the write speed is too slow.

Test-Parameters: envdefinitions=SLOW=yes testlist=ost-pools
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: Id9aef77ef07358a03133ed6d34c299609540ef42
Reviewed-on: http://review.whamcloud.com/4831
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2505 osc: use GFP_ATOMIC for ll_dirty_page_discard_warn()

Because it can be called inside client obd list lock as follows:

ll_dirty_page_discard_warn at ffffffffa0a3d252 [lustre]
vvp_page_completion_common at ffffffffa0a7adfc [lustre]
vvp_page_completion_write_common at ffffffffa0a7ae6b [lustre]
vvp_page_completion_write at ffffffffa0a7b83e [lustre]
cl_page_completion at ffffffffa05eed8f [obdclass]
osc_completion at ffffffffa0880812 [osc]
osc_ap_completion at ffffffffa086a544 [osc]
brw_interpret at ffffffffa0876d69 [osc]

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: Ia1ee6c9885941068a358c5a6aeaa9bd47a9f9b7b
Reviewed-on: http://review.whamcloud.com/4866
Tested-by: Hudson
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Tested-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

Lustre 2.1.4-rc1

Change-Id: I53a3d90a9118d083b52bc2f08f0b7b99c70d48f4
Signed-off-by: Oleg Drokin <green@whamcloud.com>

LU-797 tests: process lfs df output properly

There is a space in the "filesystem summary" of lfs df output,
which could cause trouble for the parsing script, test_23b of
ost-pool.sh is an example.

We introduced lfs_df() function to remove this space internally,
this will allow fixing the "lfs df" output line in the future.

port of master patch fb1a1ec6300a5ec3925b725d5d2b783314dff3f8

Signed-off-by: Niu Yawei <niu@whamcloud.com>
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: Iba57b4aac714e28ebdd985f15ff3e2444e64e007
Reviewed-on: http://review.whamcloud.com/4784
Tested-by: Hudson
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-2463 security: osd_mknod shouldn't alter mode after creation

osd_mknod() calls init_special_inode(), but it should use
inode->i_mode instead of the one from client.

Test-Parameters: envdefinitions=SLOW=yes testlist=posix
Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: Iee0f49dc4beef4aae6b5b6f57679de2643887e25
Reviewed-on: http://review.whamcloud.com/4820
Tested-by: Hudson
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Jian Yu <jian.yu@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2304 lov: check enq flags when matching top lock

cld_enq_flags must be checked when matching top lock, otherwise
if a sublock is missing then it will be reenqueued with wrong enq
flags.

For LU-2304, a previous truncate to zero lock(with CEF_DISCARD_DATA)
flags was matched by not-to-zero truncate and then wrongly reenqueued
with discard flag. This caused dirty pages to be discarded on the
other clients.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: I0575509b3ec30982fe7e5a7ff018ff4dd3dbe911
Reviewed-on: http://review.whamcloud.com/4818
Tested-by: Hudson
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2309 config: ignore unknown configuration param

Client or server should not fail to mount if it hits
a configuration parameter that it doesn't understand.
This patch fixes class_process_config() to meet
the above requirement.

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ida537193459f08d804fc4c0f1f72630b416440e3
Reviewed-on: http://review.whamcloud.com/4663
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-1075 tests: auster detect directory as script name

Auster should use -f to detect the existing of
script name rather use -e which will treat directory
same as script.

Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ic05aeeb6cdda183dddfdc1bec2646546898922dd
Reviewed-on: http://review.whamcloud.com/4795
Tested-by: Hudson
Reviewed-by: Wei Liu <wei3.liu@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2420 tests: add add_group() and add_user()

This patch adds add_group() and add_user() functions into the
test-framework.sh. They are used by setup_posix_users() in
posix.sh.

This patch is backported from commit 34a6092 for LU-1661.
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I33baec4abe60752ba236205e03d607a5da250216
Reviewed-on: http://review.whamcloud.com/4797
Tested-by: Hudson
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2448 build: add {prepare-,}commit-msg git hooks

Add commit message and patch format hooks for b2_1 to match other
development branches. This avoids problems when checking out b2_1
from master, where the .git/hooks/{prepare-,}commit-msg symlinks
are installed but not pointing at anything.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I9bea14402d1f5668855b55230aaac919e5d0d90e
Reviewed-on: http://review.whamcloud.com/4782
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-1299 clio: a combo patch to fix cl_lock

In this patch, the following issues are fixed:
* redefine cl_lock_error(). There are two kinds of error when
  enqueuing a cl_lock: internal errors return from server side and
  local error for example interrupted by signal when waiting for a
  lock. Obviously the latter one is not fatal so we shouldn't err the
  lock out;
* for fault handling, only signal SIGKILL and SIGTERM is not allowed;
* handle return code of ll_fault correctly;
* redefine ->unuse() to release resource it held if enqueuing lock
  failed;
* in cl_lock_hold_release(), only CLS_CACHED mode lock will be cached,
  for locks in other states, they will be cancelled voluntarily;
* fix the problem that cfs_block_sigs() is wrongly used as
  cfs_restore_sigs() and implementation of cfs_block_sigs() was wrong
* Unuse method releases the underneath resources even error occurs, so
  we should set the lock to be CLS_NEW anyway.
* other minor fixes.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: Iffce8be356723781b8f33ec9bdc2cf73e9e07138
Reviewed-on: http://review.whamcloud.com/2574
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-1833 util: don't update mtab if linked to /proc

Backport patch of LU-1833 for b2_1.
This is needed for the diskless environment since their
mtab is linking to /proc/mount.

Signed-off-by: Shuichi Ihara <sihara@ddn.com>
Signed-off-by: Peng Tao <tao.peng@emc.com>
Change-Id: Id18cd6a820dff905f0d61ecc1f71f4ce7b48cf77
Reviewed-on: http://review.whamcloud.com/4785
Tested-by: Hudson
Reviewed-by: Peng Tao <bergwolf@gmail.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-1526 tests: add version_code() and lustre_version_code()

This patch adds the version_code() and lustre_version_code()
functions into the test framework.

It also fixes the version number for the interoperability
of quota codes and reads the quota_type from MDS instead of
MGS.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes testlist=conf-sanity,ost-pools,sanity-quota
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I5a4d26fb71a7e56fa8497688ee2492ec90c446f8
Reviewed-on: http://review.whamcloud.com/4754
Tested-by: Hudson
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-847 quota: client retrieve quota usage directly

Current 'lfs quota' sends getquota RPC to MDS, and MDS is responsible
for retrieving disk usage from all targets, this scheme will be
changed to client retrieving disk usage from all targets directly.

This patch addresses the compatibility issue as well: If the getquota
returned by MDS has QIF_SPACE, client just trust the disk usage
returned by MDS, otherwise, client has to issue RPCs to all OSTs to
collect disk usage by itself.

Signed-off-by: Niu Yawei <niu@whamcloud.com>
Signed-off-by: Bob Glossman <bogl@whamcloud.com>
Change-Id: Ia05e588e1a28abd3057274b38684057ecc887f5b
Reviewed-on: http://review.whamcloud.com/2794
Tested-by: Hudson
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-1693 obdfilter: Set bi_rw before calling bio_add_page()

In filter_do_bio() set bi_rw before calling bio_add_page()
so that the correct bio direction is used in mgrge_bvec_fn().

port of master patch 79badd39c6584ae136b14d919eaf699ee812b666

Signed-off-by: John L. Hammond <jhammond@tacc.utexas.edu>
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: Ib43a5c198db691f8ed4001c41af7b984a264922b
Reviewed-on: http://review.whamcloud.com/4765
Tested-by: Hudson
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-2420 tests: Have POSIX testing on b2_1

Have POSIX testing automated on b2_1 as it is on master.
LU-2274 port is included to change the baseline filesystem to ext3
for POSIX testing on SLES distro.

Test-Parameters: envdefinitions=SLOW=yes clientdistro=sles11 testlist=posix
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: If20262a37d77053bda3a1e143d74c4ebd11b7ec5
Reviewed-on: http://review.whamcloud.com/4731
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2405 tests: verify off-by-one symlink length fix

Add regression test for off-by-one bug in symlink code. Ensures boundary
condition of 60 character symlinks are readable after creation.

original commit 6c3606b79a4c25946255a680f79ff2d58df66bbc

Whamcloud-bug-id: LU-2241
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: I4264558a5b638e73e7631b988b05b92405d13288
Reviewed-on: http://review.whamcloud.com/4711
Tested-by: Hudson
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-2371 ptlrpc: add support for -EINPROGRESS

Backport patches from LU-904, LU-1329 and LU-1788 to introduce
support for -EINPROGRESS in lustre 2.1. This is needed for
quota interoperability with 2.4 servers.

Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Change-Id: I17689380c8d42a8daef915342f63d96f777dc6fa
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/4645
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2170 osc: set osc_lock attribute only once

Set osc_lock's attribute by lock allocator, otherwise if this lock is
matched and enqueued by a glimpse thread, the osc_lock's ols_glimpse
will be set to true and the lock state will be messed in
osc_lock_upcall().

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: I2aff3d372d059474fed2c5d0983bdb116e932701
Reviewed-on: http://review.whamcloud.com/4482
Tested-by: Hudson
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-1648 llog: prevent journal to restart for llog cancel

Journal credit for llog cancel should include the blocks for
deleting the inode. Otherwise the journal handle may need to
be restarted during the transaction for more credit. Then it
may be blocked with holding llog_handle::lgh_lock, which may
cause deadlock with other llog operations those have started
journal handle but waiting for the llog_handle::lgh_lock.

port of master patch 8d334f67052ccf77713c780a8040593728977599

Signed-off-by: Fan Yong <yong.fan@whamcloud.com>
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: I4013eb764485c282ed2443d02556c3377e16c936
Reviewed-on: http://review.whamcloud.com/4743
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-1749 llog: MDT should work if cancel llog failed (-ESTALE)

Cancel catlog may be stale, in this case MDT should continue to work.

port of b2_3 patch e5d5cd20169b476cdb4fe4857d64aef4bf03d497

Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: I85d571a953b2a34d51d586595e271933a016b04f
Reviewed-on: http://review.whamcloud.com/4742
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-1650 obdclass: find the lock by index of subpage

In check_and_discard_cb(), subpage's index should be used to find if
there exists another lock to cover the same page.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: Id8e1da89ef81fad4c7917360d591fc7bd1309a00
Reviewed-on: http://review.whamcloud.com/4698
Tested-by: Hudson
Reviewed-by: Bobi Jam <bobijam@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>

LU-2373 kernel: Kernel update [RHEL 6.3 2.6.32-279.14.1.el6]

Update RHEL6.3 kernel to 2.6.32-279.14.1.el6.

Signed-off-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: yangsheng <ys@whamcloud.com>
Change-Id: I43a6721663cfab794882f220f7c64a705be20c0f
Reviewed-on: http://review.whamcloud.com/3811
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-921 llite: warning in case of discarding dirty pages

when a client is evicted, dirty pages may get silently discarded,
the caller of successful write(2) will not know that the data he
wrote have been discarded due to eviction before it can be flushed
to the OSS.
test 24 of recovery-small is now reenabled back.

port of master patch 624a3ac233938153e889ae81f639cf46c0a8312b
port of master patch 4cfcf91efa06decc155d8f12eafc8ede3e850fd5

Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Signed-off-by: Liu Ying <emoly.liu@intel.com>
Change-Id: If8aeed507890a656cbb8231879d0e31aa9f4ea02
Reviewed-on: http://review.whamcloud.com/4716
Tested-by: Hudson
Reviewed-by: Jinshan Xiong <jinshan.xiong@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-2372 kernel: kernel update [RHEL5.8 2.6.18-308.20.1.el5]

Update RHEL5.8 kernel to 2.6.18-308.20.1.el5.

Signed-off-by: yang sheng <ys@whamcloud.com>
Change-Id: Iafb4088227fea098840b70e574f2afe5a1f4067c
Reviewed-on: http://review.whamcloud.com/4658
Tested-by: Hudson
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>