Whamcloud - gitweb
fs/lustre-release.git
9 years agob=21846 add a test for testing rehash in sanity.sh
Fan Yong [Fri, 6 Aug 2010 02:22:50 +0000 (06:22 +0400)]
b=21846 add a test for testing rehash in sanity.sh

1. enable rehash on lqs
2. if long is 64bit, __fls() is wrong. Fix it.

i=landen
i=tappro

9 years agob=22842 do not use bash variable to pass test setup state between sub-test
Vladimir Saveliev [Fri, 6 Aug 2010 23:06:38 +0000 (03:06 +0400)]
b=22842 do not use bash variable to pass test setup state between sub-test

tests 42, 101, 102 are fixed

i=robert.read

9 years agob=18751 Move prng.c to libcfs
Maxim Patlasov [Fri, 6 Aug 2010 13:21:46 +0000 (17:21 +0400)]
b=18751 Move prng.c to libcfs

i=adilger
i=maxim

Note that some function names were changed:
  cfs_get_random_bytes --> cfs_get_random_bytes_prim
  ll_get_random_bytes --> cfs_get_random_bytes
  ll_rand --> cfs_rand
  ll_srand --> cfs_srand

9 years agob=19529 Avoid deadlock for local client writes.
Dmitry Zogin [Thu, 5 Aug 2010 19:32:40 +0000 (23:32 +0400)]
b=19529 Avoid deadlock for local client writes.

  Check the OBD_BRW_MEMALLOC flag correctly in the remote buffer.

 i=andreas.dilger
 i=johann.lombardi

9 years agob=21760 Applications hung in direct I/O
Dmitry Zogin [Thu, 5 Aug 2010 18:48:56 +0000 (22:48 +0400)]
b=21760 Applications hung in direct I/O

 Make sure that the bulk is aborted, if a request has been aborted in flight.
 Call ptlrpc_abort_bulk() out of ptlrpc_check_set()

 i=oleg.drokin
 i=andrew.perepechko

9 years agob=22612 dig out Module.symvers for sles10
Brian J. Murrell [Fri, 6 Aug 2010 17:59:04 +0000 (21:59 +0400)]
b=22612 dig out Module.symvers for sles10

When we build modules in lbuild for sles10 kernels, we point to the
contents of the kernel-source RPM.  That does not contain the
Module.symvers file though and thus we get a nice big warning that
the modules built (i.e. ldiskfs) will not contain versions.
This is because we need to unpack the /usr/src/linux-*obj/Module.symvers
file from the kernel-$flavour or kernel-syms RPM into the location pointed to
by --with-linux-obj.

Fix a bug in unpack_linux_devel_rpm-$DISTRO where we were not checking
for the success of the function.

i=mjmac
i=wangyb

9 years agob=23139 workaround to avoid assertion in osc_init_grant
hongchao.zhang [Wed, 28 Jul 2010 22:54:29 +0000 (02:54 +0400)]
b=23139 workaround to avoid assertion in osc_init_grant

workaround for 1.6 servers which don't have the patch
from bug20278 applied

i=oleg.drokin
i=eric.mei

9 years agob=22978 fix compiler warning error
Brian J. Murrell [Fri, 6 Aug 2010 14:08:59 +0000 (18:08 +0400)]
b=22978 fix compiler warning error

Fix the warning/error:
sanity.c: In function ‘t21’:
sanity.c:708: error: suggest parentheses around operand of ‘!’ or change ‘&’ to ‘&&’ or ‘!’ to ‘~’

i=panda
i=whitebear

9 years agob=23352 add debugging information to late lock cancels
hongchao.zhang [Wed, 28 Jul 2010 23:02:09 +0000 (03:02 +0400)]
b=23352 add debugging information to late lock cancels

print out the arrival time of the delayed cancel RPC, so we can
determine whether it was waiting in the queue a long time, or
whether it arrived late on the network

i=nathan.rutman
i=hongchao.zhang

9 years agob=20877 disable test_12 of sanity-quota.sh
Landen [Fri, 6 Aug 2010 02:33:07 +0000 (06:33 +0400)]
b=20877 disable test_12 of sanity-quota.sh

i=johann

9 years agob=23296 Fix the initialization of oti
Manoj Joseph [Fri, 6 Aug 2010 01:22:07 +0000 (05:22 +0400)]
b=23296 Fix the initialization of oti

Fix the initialization of obd_trans_info

i=qiang.wu
i=andreas.dilger

9 years agob=22818 exp_queued_rpc under a separate lock
Vitaly Fertman [Wed, 4 Aug 2010 22:17:50 +0000 (02:17 +0400)]
b=22818 exp_queued_rpc under a separate lock

i=green
i=liang

another bh lock is added just to protect exp_queued_rpc in the export

9 years agob=16890 a build fix for --disable-quota
Andrew Perepechko [Thu, 15 Jul 2010 14:49:08 +0000 (18:49 +0400)]
b=16890 a build fix for --disable-quota

i=ZhiYong Tian
i=Yong Fan

attachments 30599 and 30600

9 years agob=22481 man page for lfs_migrate
Andreas Dilger [Wed, 4 Aug 2010 20:10:16 +0000 (00:10 +0400)]
b=22481 man page for lfs_migrate

Add a manual page for lfs_migrate.
Minor formatting changes of the lfs_migrate::usage() message.

9 years agob=11739 eliminate lbug
Vitaly Fertman [Tue, 3 Aug 2010 09:49:27 +0000 (13:49 +0400)]
b=11739 eliminate lbug

i=green
i=adilger

don't LBUG on network errors in ldlm_handle_cancel().

9 years agob=21528 lower the message severety
Vitaly Fertman [Mon, 2 Aug 2010 10:45:23 +0000 (14:45 +0400)]
b=21528 lower the message severety

i=johann

9 years agob=22348 improve get_mdtosc_proc_path() to handle 2.0 server on 1.8 disk
Jian Yu [Mon, 19 Jul 2010 07:35:11 +0000 (11:35 +0400)]
b=22348 improve get_mdtosc_proc_path() to handle 2.0 server on 1.8 disk

If the 2.0 MDS was mounted on 1.8 device, then the OSC and LOV names
used by MDT would not be changed:
mdt lov: fsname-mdtlov
mdt osc: fsname-OSTXXXX-osc
The patch is to improve get_mdtosc_proc_path() to get correct OSC name
while 2.0 server being mounted on 1.8 disk.

i=robert.read
i=grev

9 years agob=19069 cleanup on nid_stats entry refcount
hongchao.zhang [Sat, 10 Jul 2010 08:39:37 +0000 (12:39 +0400)]
b=19069 cleanup on nid_stats entry refcount

introduce reconnect flag in "lprocfs_exp_setup" to cleanup
the refcount of nid_stats

i=andreas.dilger
i=yong.fan

9 years agob=16842 Libcfs common code changes
Maxim Patlasov [Thu, 22 Jul 2010 11:29:38 +0000 (15:29 +0400)]
b=16842 Libcfs common code changes

i=rread
i=alex
libcfs: get rid of cfs_alloc/free_pages and cfs_get_blockedsigs, fix __l_wait_event

9 years agob=17167 cfs_rcu_head cleanup in lustre source
Matt Wu [Thu, 29 Jul 2010 06:55:31 +0000 (10:55 +0400)]
b=17167 cfs_rcu_head cleanup in lustre source

i=andreas.dilger
i=maxim.patlasov

9 years agob=23382 t-f: do_nodes(): wrong sed RE
Elena Gryaznova [Fri, 30 Jul 2010 14:31:49 +0000 (18:31 +0400)]
b=23382 t-f: do_nodes(): wrong sed RE

i=Andrew.Perepechko

9 years agoMerge branch 'master' of git.lustre.org:prime/lustre
Mikhail Pershin [Fri, 30 Jul 2010 18:55:46 +0000 (22:55 +0400)]
Merge branch 'master' of git.lustre.org:prime/lustre

9 years agob=22977 re-struct obdfilter-survey.sh
Elena Gryaznova [Fri, 30 Jul 2010 13:36:36 +0000 (17:36 +0400)]
b=22977 re-struct obdfilter-survey.sh

i=Andrew.Perepechko

9 years agob=22976 fix bash syntax error
Brian J. Murrell [Thu, 29 Jul 2010 20:24:43 +0000 (00:24 +0400)]
b=22976 fix bash syntax error

When using "test" to compare strings, both sides of the comparison
should either be quoted or not quoted, not one of each.
= should be used for equality in portable tests.

i=mjmac
i=whitebear

9 years agoCorrected the field that represents pre release version numbers. 2.0.0.51 v2_0_0_51
Terry Rutledge [Fri, 30 Jul 2010 17:46:01 +0000 (10:46 -0700)]
Corrected the field that represents pre release version numbers.
This is for 2.1.0 build 01.

9 years agoUpdated version strings for 2.1.0 build 01. Updated ChangeLogs
Terry Rutledge [Fri, 30 Jul 2010 17:41:02 +0000 (10:41 -0700)]
Updated version strings for 2.1.0 build 01.  Updated ChangeLogs
with TBD, Oracle, Inc. and new version info.

9 years agoRevert "b=22464 add writeconf as mount option (includes bz 23228)"
Mikhail Pershin [Wed, 28 Jul 2010 17:09:48 +0000 (21:09 +0400)]
Revert "b=22464 add writeconf as mount option (includes bz 23228)"

This reverts commit adcf9aa295e5c0bf51bc8dd45dde1e900b0e1165.

9 years agob=20433 we should recycle dentries and inodes if only cancelling locks existing
Landen [Tue, 27 Jul 2010 11:11:04 +0000 (15:11 +0400)]
b=20433 we should recycle dentries and inodes if only cancelling locks existing

i=green
i=adilger

9 years agob=16919 Async journal commit support
Oleg Drokin [Fri, 23 Jul 2010 10:03:02 +0000 (14:03 +0400)]
b=16919 Async journal commit support

Async journal commit for 2.0 branch. Also reverts patch from bug 22190 as no
longer needed.
Also included changes from b1_8 that were introduced in bug 22241.
Async commit is ENABLED by default.
i=adilger
i=tappro

9 years agob=22252 allow multiple instances of the same nid in NID hash.
Oleg Drokin [Fri, 23 Jul 2010 09:54:01 +0000 (13:54 +0400)]
b=22252 allow multiple instances of the same nid in NID hash.

Case of multiple separate clients from the same NID (as with liblustre) is legitimate and so we
should allow multiple instances of the same NID in nid hash

i=rread
i=johann

9 years agob=22888 fix memory leak in handle_recovery_req() when export disconnected already
Fan Yong [Thu, 22 Jul 2010 01:22:12 +0000 (05:22 +0400)]
b=22888 fix memory leak in handle_recovery_req() when export disconnected already

fix memory leak in handle_recovery_req() when export disconnected already

i=Mikhail.Pershin
i=landen

9 years agob=22464 add writeconf as mount option (includes bz 23228)
Nathan Rutman [Wed, 21 Jul 2010 22:46:37 +0000 (02:46 +0400)]
b=22464 add writeconf as mount option (includes bz 23228)

i=alexg
i=adilger

9 years agob=22244
pravin [Mon, 19 Jul 2010 19:05:47 +0000 (23:05 +0400)]
b=22244

a=vitaly
i=oleg.drokin
i=andreas

fix stack overflow. (patch from attachment 29939 and 30120)

9 years agob=21128
pravin [Mon, 19 Jul 2010 19:05:20 +0000 (23:05 +0400)]
b=21128

i=oleg
i=rahul
avoid stack overflow by delegating work to separate thread

9 years agob=23335 validate echo_client ostid values
Mikhail Pershin [Wed, 21 Jul 2010 19:20:17 +0000 (23:20 +0400)]
b=23335 validate echo_client ostid values

i=tappro,alex

9 years agob=15685 fix obdfilter-survey script to work properly with remote oss-s
Elena Gryaznova [Wed, 21 Jul 2010 16:51:08 +0000 (20:51 +0400)]
b=15685 fix obdfilter-survey script to work properly with remote oss-s

a=David.Dillow dillowda@ornl.gov
i=grev

Fix obdfilter-survey to work on multiple OSSes at once

9 years agob=22402 fix whitespaces at the end of lines
Mikhail Pershin [Wed, 21 Jul 2010 14:42:34 +0000 (18:42 +0400)]
b=22402 fix whitespaces at the end of lines

9 years agob=19917 repeated atomic allocation failures in kptllnd_rx_alloc() causing OST to...
Dmitry Zogin [Wed, 21 Jul 2010 13:27:23 +0000 (17:27 +0400)]
b=19917 repeated atomic allocation failures in kptllnd_rx_alloc() causing OST to be unreachable

 Comment change.

9 years agob=20247 Dropping PUT error on first mount
Dmitry Zogin [Wed, 21 Jul 2010 13:19:38 +0000 (17:19 +0400)]
b=20247 Dropping PUT error on first mount

 Remove D_NETERROR from the default value of libcfs_printk

9 years agob=23182 lustre-iokit sgpdd-survey defect
Elena Gryaznova [Wed, 21 Jul 2010 12:21:30 +0000 (16:21 +0400)]
b=23182 lustre-iokit sgpdd-survey defect

i=Vladimir.Saveliev

9 years agob=22402 new OBDFILTER_SURVEY test suite
Elena Gryaznova [Tue, 20 Jul 2010 13:54:43 +0000 (17:54 +0400)]
b=22402 new OBDFILTER_SURVEY test suite

i=Andrew.Perepechko

9 years agob=13698 LL_IOC_RECREATE_FID
Vladimir Saveliev [Sat, 17 Jul 2010 00:14:12 +0000 (04:14 +0400)]
b=13698 LL_IOC_RECREATE_FID

define new ioctl for object replicate
it uses IDIF FID instead of truct ll_recreate_obj
old LL_IOC_RECREATE is kept for compatibility

i=andreas.dilger
i=di.wang

9 years agob=23335 Allocate echo objects that can be mapped to a valid FID
Andreas Dilger [Fri, 16 Jul 2010 22:17:05 +0000 (02:17 +0400)]
b=23335 Allocate echo objects that can be mapped to a valid FID

With the change to using valid FIDs for all OST objects in bug 19427,
the echo objid needs to be below 2^32, because regular FID numbers
are limited to 2^32 objects in a single sequence number.

i=aleksandr.guzovskiy
i=mikhail.pershin

9 years agob=22680 remove dead code in sanity.sh
Mikhail Pershin [Tue, 20 Jul 2010 19:05:18 +0000 (23:05 +0400)]
b=22680 remove dead code in sanity.sh

i=grev

9 years agob=23278 replay-single test 86 does not remount client
Elena Gryaznova [Mon, 19 Jul 2010 19:13:43 +0000 (23:13 +0400)]
b=23278 replay-single test 86 does not remount client

o=Oleg.Drokin
i=Elena.Gryaznova

9 years agob=21720 fix test 18 to interleave tests increase pass margin
Nathan Rutman [Mon, 19 Jul 2010 18:39:53 +0000 (22:39 +0400)]
b=21720 fix test 18 to interleave tests increase pass margin

i=rread

9 years agob=22944 remove conf-sanity test 16.
Eric Mei [Mon, 19 Jul 2010 16:16:11 +0000 (20:16 +0400)]
b=22944 remove conf-sanity test 16.

i=andreas.dilger
i=jian.yu

9 years agob=23076 fix for o2iblnd reconnect
Maxim Patlasov [Mon, 19 Jul 2010 09:37:57 +0000 (13:37 +0400)]
b=23076 fix for o2iblnd reconnect

i=isaac
Reconnect is still needed even when ibp_tx_queue is empty
if ibp_version != version because reconnect may be
initiated by kiblnd_query().

9 years agob=23293 Fix memory leak in lprocfs_fops_read().
Ricardo M. Correia [Fri, 16 Jul 2010 16:04:06 +0000 (20:04 +0400)]
b=23293 Fix memory leak in lprocfs_fops_read().

i=andreas.dilger
i=emoly.liu

It was possible for lprocfs_fops_read() to leak a page in some circumstances.
This small patch should fix it.

9 years agob=11742 FSX checksum false positves due to mmap IO
Dmitry Zogin [Fri, 16 Jul 2010 13:31:52 +0000 (17:31 +0400)]
b=11742 FSX checksum false positves due to mmap IO

 Use OBD_FL_MMAP flag for IOs on a memory mapped file. Do not print
 checksum errors, if the flag is set on a request.

 i=adilger
 i=alexey.lyashkov
 i=johann

9 years agoRevert "b=20355 formalize support for pthreads"
Mikhail Pershin [Mon, 19 Jul 2010 17:40:29 +0000 (21:40 +0400)]
Revert "b=20355 formalize support for pthreads"

This reverts commit 004cc2b4c6eb32672985e2f4b4bbf9413e314547.

9 years agob=23310 Partly matching the out message against the expected one to resolve the diffe...
Fan Yong [Fri, 16 Jul 2010 14:31:18 +0000 (18:31 +0400)]
b=23310 Partly matching the out message against the expected one to resolve the different output message format for getfacl/setfal on different Linux distributions

Partly matching the out message against the expected one to resolve the different output message format for getfacl/setfal on different Linux distributions.

i=tappro
i=wangyb

9 years agob=23244 unhash the new added export when such obd device is stopping
Fan Yong [Fri, 16 Jul 2010 14:48:11 +0000 (18:48 +0400)]
b=23244 unhash the new added export when such obd device is stopping

unhash the new added export when such obd device is stopping.

i=andreas.dilger
i=vitaly.fertman

9 years agob=13698 llapi_get_version
Vladimir Saveliev [Fri, 16 Jul 2010 13:51:18 +0000 (17:51 +0400)]
b=13698 llapi_get_version

this uses OBD_GET_VERSION ioctl to obtain lustre version

i=andreas.dilger

9 years agob=13698 rename osd to osd-ldiskfs
Vladimir Saveliev [Thu, 1 Jul 2010 21:39:55 +0000 (01:39 +0400)]
b=13698 rename osd to osd-ldiskfs

according to 20226

i=rahul

9 years agoImprove nn-check.py output to be less verbose.
Andreas Dilger [Wed, 14 Jul 2010 16:27:48 +0000 (20:27 +0400)]
Improve nn-check.py output to be less verbose.

Don't complain about common english words that may be in comments.  Require
that "up", "down", and "completion" be used as functions with opening "("
before they are flagged.

I thought about, and rejected, discarding all comments, but it still makes
sense to clean up non-portable function names in comments.

9 years agoRemove obsolete configuration tools.
Andreas Dilger [Wed, 14 Jul 2010 16:26:19 +0000 (20:26 +0400)]
Remove obsolete configuration tools.

Remove old 1.4 configuration tools from master.  They were removed from b1_8
a long time ago.

9 years agoUpdated version string for the start of development for 2.1.0.
Terry Rutledge [Wed, 14 Jul 2010 23:08:32 +0000 (16:08 -0700)]
Updated version string for the start of development for 2.1.0.

9 years agob=21054 code cleanup in lov_lock_delete()
Eric Mei [Thu, 8 Jul 2010 16:17:53 +0000 (10:17 -0600)]
b=21054 code cleanup in lov_lock_delete()

o=eeb
r=jay
r=wangdi

9 years agob=22934 fix writeconf in conf-sanity
Nathan Rutman [Wed, 7 Jul 2010 18:23:29 +0000 (11:23 -0700)]
b=22934 fix writeconf in conf-sanity

fix writeconf in conf-sanity
fix conf-sanity slow list
i=johann

9 years agob=14512 Fix mistake made in original patch
Robert Read [Tue, 6 Jul 2010 22:15:56 +0000 (15:15 -0700)]
b=14512 Fix mistake made in original patch

Change 39eabbad40 introduced an error in sanity.sh quite a while ago,
this cleans it up.

9 years agob=22745 Removed gmlnd from lnet.
Isaac Huang [Tue, 6 Jul 2010 17:24:13 +0000 (11:24 -0600)]
b=22745 Removed gmlnd from lnet.

Removed obsolete gmlnd from lnet.

i=maxim

9 years agob=21816 allow querying quota stats even when some OSTs are down
Andrew Perepechko [Tue, 6 Jul 2010 13:59:43 +0000 (17:59 +0400)]
b=21816 allow querying quota stats even when some OSTs are down

i=Yong Fan
i=Landen

9 years agob=19390 remove unneeded spinlock
Landen [Mon, 5 Jul 2010 15:34:26 +0000 (23:34 +0800)]
b=19390 remove unneeded spinlock

i=landen
i=johann

9 years agob=21696 create a new performance test in sanity-quota.sh
Landen [Mon, 5 Jul 2010 15:34:25 +0000 (23:34 +0800)]
b=21696 create a new performance test in sanity-quota.sh

i=yong.fan
i=robert.read

9 years agob=23009 Fix access past end of buffer in fsoptions_to_mdt_flags()
Ricardo M. Correia [Mon, 5 Jul 2010 14:39:53 +0000 (14:39 +0000)]
b=23009 Fix access past end of buffer in fsoptions_to_mdt_flags()

It seems fsoptions_to_mdt_flags() was accessing the "options" buffer past the
end of it due to a missing check for NULL. This should fix it.

i=alex.zhuravlev
i=tappro

9 years agob=19298 debug messages
Vitaly Fertman [Mon, 5 Jul 2010 10:56:38 +0000 (14:56 +0400)]
b=19298 debug messages

i=rread

some extra debug messages about i_size change

9 years agob=20908 wait on sendpage
Vitaly Fertman [Mon, 5 Jul 2010 10:56:47 +0000 (14:56 +0400)]
b=20908 wait on sendpage

i=rread
i=shadow

make the while loop for mdt_sendpage to wail until rpc deadline if it has been
prolonged meanwhile but the bulk timeout occurs.

9 years agob=23237 mount.lustre dies with SIGSEGV: Unable to read 1.8 config /tmp/lustre_tmp...
Dmitry Zogin [Sat, 3 Jul 2010 00:21:15 +0000 (20:21 -0400)]
b=23237 mount.lustre dies with SIGSEGV: Unable to read 1.8 config /tmp/lustre_tmp.IfgmBK/mountdata

 Do not try to close non-open file and return the error in get_mountdata()

 i=andreas.dilger
 i=johann

9 years agob=21563 Metadata performance has degraded for some operations between 1.6.5 and 1.8.1
Dmitry Zogin [Fri, 2 Jul 2010 23:53:49 +0000 (19:53 -0400)]
b=21563 Metadata performance has degraded for some operations between 1.6.5 and 1.8.1

 Change mdsrate tool to handle time with double resolution and use MPI_Wtime().
 Also stat operations measurements do not require --iter argument anymore, and do
 not use readdir() internally.

 i=johann
 i=iandrew.perepechko
 i=robert.read

9 years agob=21776 Use atomic memory allocation when VM is flushing to free memory.
Isaac Huang [Fri, 2 Jul 2010 21:15:52 +0000 (15:15 -0600)]
b=21776 Use atomic memory allocation when VM is flushing to free memory.

- Added rq_memalloc flag to RPC and set memory pressure when sending such RPCs
- Set memory pressure flag when osc creates RPC for flushing dirty pages.
- Added msg_vmflush to lnet_msg_t so that LNDs can set memory pressure
  flag when lnd_send called asynchronously. The o2iblnd probably doesn't
  need it as its TX are preallocated and TX pools grown asynchronously.

i=johann
i=dmitry.zoguine
i=maxim.patlasov

9 years agob=20978 Fix llog_process_lock scalability issues.
pravin [Fri, 2 Jul 2010 13:31:53 +0000 (19:01 +0530)]
b=20978 Fix llog_process_lock scalability issues.

i=andrew.perepechko
i=rahul.deshmukh

fix llog_process_lock scalability issues during normal start up due to
slow stale export processing.

9 years agob=15936 minor fixes noted in inspection
Mikhail Pershin [Fri, 2 Jul 2010 13:14:30 +0000 (17:14 +0400)]
b=15936 minor fixes noted in inspection

i=adilger,zam

Use single place for mount count, unify mntdev getting, introduce
obt_magic to make sure obt is correct.

9 years agob=20065 Add 1S statfs caching and cleanup CFS_HZ usage.
yangsheng [Fri, 2 Jul 2010 11:30:37 +0000 (19:30 +0800)]
b=20065 Add 1S statfs caching and cleanup CFS_HZ usage.

i=adilger
i=bobijam

9 years agob=23175 Use MPI_Bcast to send random value to all ranks.
minhdiep [Fri, 2 Jul 2010 00:27:22 +0000 (18:27 -0600)]
b=23175 Use MPI_Bcast to send random value to all ranks.

The file size will be different if each rank generate its own
chunksize. We need rank 0 to generate the random and pass it
to all ranks.

i=Johann

9 years agob=12197 The mount point passed into mtab should be the absolute path
minhdiep [Fri, 2 Jul 2010 00:02:46 +0000 (18:02 -0600)]
b=12197 The mount point passed into mtab should be the absolute path

Convert the mount point to real path.

i=Johann

9 years agob=21244 fix compilability on ia64
Vladimir Saveliev [Sat, 26 Jun 2010 13:54:28 +0000 (17:54 +0400)]
b=21244 fix compilability on ia64

use LPU64 macro to print __u64

i=zam

9 years agoRevert "b=20581 Handle directory entry hash collisions" v2_0_0-rc1a v2_0_0_RC1
Robert Read [Thu, 1 Jul 2010 06:46:55 +0000 (23:46 -0700)]
Revert "b=20581 Handle directory entry hash collisions"

This reverts commit 04e67d056c015dbe3f8ce69092489fe9a10d771d.

This patch is no longer required (as we think we understand the problem now),
and it introduced a rare NULL pointer exception, so best to remove it.

9 years agoUpdate ldiskfs version number. v2_0_0-rc1
Robert Read [Thu, 1 Jul 2010 00:42:20 +0000 (17:42 -0700)]
Update ldiskfs version number.

9 years agoPrepare for 2.0.0-rc1.
Robert Read [Wed, 30 Jun 2010 19:31:25 +0000 (12:31 -0700)]
Prepare for 2.0.0-rc1.

9 years agoRemove ia64 from oel5 target file.
Robert Read [Wed, 30 Jun 2010 19:29:26 +0000 (12:29 -0700)]
Remove ia64 from oel5 target file.

OEL5 doesn't support ia64.

9 years agob=20997 stop ost first in conf-sanity
Mikhail Pershin [Mon, 28 Jun 2010 18:24:18 +0000 (22:24 +0400)]
b=20997 stop ost first in conf-sanity

i=robert.read

9 years agob=20997 move obd_no_recov check back
Mikhail Pershin [Mon, 28 Jun 2010 18:19:17 +0000 (22:19 +0400)]
b=20997 move obd_no_recov check back

i=robert.read
i=oleg.drokin

9 years agoRevert unintended commit to build/updatecw.sh
Robert Read [Wed, 30 Jun 2010 19:00:00 +0000 (12:00 -0700)]
Revert unintended commit to build/updatecw.sh

(These hacks were needed to update every file.)

9 years agoMass conversion of all copyright messages to Oracle.
Robert Read [Wed, 30 Jun 2010 18:55:24 +0000 (11:55 -0700)]
Mass conversion of all copyright messages to Oracle.

9 years agoAdd missing .gitgnore files from new pthread directories.
Robert Read [Mon, 28 Jun 2010 23:12:35 +0000 (16:12 -0700)]
Add missing .gitgnore files from new pthread directories.

9 years agoRevert "b=21128 run sync ldlm_bl_to_thread_list() in separate thread to save stack...
Robert Read [Mon, 28 Jun 2010 18:04:28 +0000 (11:04 -0700)]
Revert "b=21128 run sync ldlm_bl_to_thread_list() in separate thread to save stack space."

This reverts commit 2cb1701012c4f1f0f1abdc99eb8ce6d4bb9c7846.

Still causing corruption, see comments on bug 21128.

9 years agoRevert "b=22244 delegate lock cancel to blocking thread"
Robert Read [Mon, 28 Jun 2010 18:04:00 +0000 (11:04 -0700)]
Revert "b=22244 delegate lock cancel to blocking thread"

This reverts commit f96aff96fae1fcb3d06fd40ae4837ba975df12e3.

Reverting because of dependency on bug 21128.

9 years agoRevert "b=22244 handle list of locks in ldlm_bl_thread_main properly"
Robert Read [Mon, 28 Jun 2010 18:00:18 +0000 (11:00 -0700)]
Revert "b=22244 handle list of locks in ldlm_bl_thread_main properly"

This reverts commit f2360643cbb8136f97ec750547d12c30f88e0d4d.

Reverting because of dependency on bug 21128.

9 years agob=20581 Handle directory entry hash collisions
Fan Yong [Fri, 25 Jun 2010 22:14:22 +0000 (15:14 -0700)]
b=20581 Handle directory entry hash collisions

i=rread
i=pravin

9 years agob=19151 use single-level test directory for large-scale test 3a
Jian Yu [Fri, 25 Jun 2010 09:00:28 +0000 (17:00 +0800)]
b=19151 use single-level test directory for large-scale test 3a

$tdir is a two-level directory which causes the mode setting issue
while running MPI programs. Let's use a single-level test directory
for large-scale test 3a.

i=grev
i=robert.read

9 years agob=22244 handle list of locks in ldlm_bl_thread_main properly
Vitaly Fertman [Thu, 24 Jun 2010 20:31:36 +0000 (00:31 +0400)]
b=22244 handle list of locks in ldlm_bl_thread_main properly

the fix to the original patch includes:
1. ldlm_bl_thread_main() cancels only locks left in the list after
ldlm_cli_cancel_list_local().
2. ldlm_bl_thread_main() and ldlm_cli_cancel_unused_resource() cancel
locks marked as LDLM_FL_BL_AST in the same RPC as others in the list.

i=green
i=johann

9 years agob=22244 delegate lock cancel to blocking thread
Vitaly Fertman [Thu, 24 Jun 2010 20:31:35 +0000 (00:31 +0400)]
b=22244 delegate lock cancel to blocking thread

i=adilger
i=green

re-landing: instead of cancelling locks locally in the shrinking thread,
deletate it to a separate blocking thread.

9 years agob=23004 bad cfs_user_write call causes oops
Nathan Rutman [Thu, 24 Jun 2010 17:49:30 +0000 (10:49 -0700)]
b=23004 bad cfs_user_write call causes oops

i=rread
i=breitz

9 years agob=22911 Fix incompatibilies between ext4 and ext3 based ldiskfs
Robert Read [Wed, 23 Jun 2010 16:30:24 +0000 (09:30 -0700)]
b=22911 Fix incompatibilies between ext4 and ext3 based ldiskfs

i=girish
i=rread
a=adilger
a=yangsheng

9 years agob=21128 run sync ldlm_bl_to_thread_list() in separate thread to save stack space.
pravin [Thu, 24 Jun 2010 16:14:45 +0000 (21:44 +0530)]
b=21128 run sync ldlm_bl_to_thread_list() in separate thread to save stack space.

 i=oleg.drokin
 i=rahul
 i=vitaly

9 years agoUpdate copyright messages.
Robert Read [Thu, 24 Jun 2010 17:02:12 +0000 (10:02 -0700)]
Update copyright messages.

9 years agob=21636 diagnostic patch to detect request delay send/reply
Fan Yong [Thu, 24 Jun 2010 06:02:33 +0000 (14:02 +0800)]
b=21636 diagnostic patch to detect request delay send/reply

diagnostic patch to detect request delay send/reply

i=robert.read
i=pravin.shelar

9 years agob=19427 add the missing swab field for "rr_flags" in "mdt_reint_rec"
Fan Yong [Thu, 24 Jun 2010 06:01:59 +0000 (14:01 +0800)]
b=19427 add the missing swab field for "rr_flags" in "mdt_reint_rec"

add the missing swab field for "rr_flags" in "mdt_reint_rec".

i=andreas.dilger
i=robert.read

9 years agob=20563
pravin [Wed, 23 Jun 2010 08:45:01 +0000 (14:15 +0530)]
b=20563

i=zam
i=andreas
i=rahul

fix values of LUSTRE_SEQ_MAX_WIDTH and LUSTRE_SEQ_META_WIDTH.
this patch fixes fid to inode mapping so that it will be independent of
LUSTRE_SEQ_MAX_WIDTH.
there is fix for seq flush also.