Whamcloud - gitweb
grev [Sun, 23 Dec 2007 23:36:48 +0000 (23:36 +0000)]
b=12499
i=Adilger
i=Alexey.Lyashkov
CLIENTMODSONLY for patchless
shadow [Fri, 21 Dec 2007 09:32:48 +0000 (09:32 +0000)]
cleanup in userland bitops code.
b=14390
i=green
i=deen
tianzy [Fri, 21 Dec 2007 05:50:21 +0000 (05:50 +0000)]
it is an obvious typo
b=14210
walter [Fri, 21 Dec 2007 04:46:23 +0000 (04:46 +0000)]
revert bz-14390 commit which is causing build failures on buffalo
walter [Thu, 20 Dec 2007 23:13:30 +0000 (23:13 +0000)]
b=14390
i=adilger
i=green
fix problem where b1_6 does not build on xt3
due to missing uint32_t defn in lnet/libcfs/user-bitops.c
grev [Thu, 20 Dec 2007 12:38:44 +0000 (12:38 +0000)]
b=14512
i=Adilger
use new-style setstripe args
grev [Thu, 20 Dec 2007 11:48:28 +0000 (11:48 +0000)]
b=14348 (made by Adilger)
i=grev
remove obsolete test scripts
grev [Thu, 20 Dec 2007 11:41:14 +0000 (11:41 +0000)]
b=14348 (made by Adilger)
i=grev
remove obsolete test scripts
shadow [Thu, 20 Dec 2007 09:13:10 +0000 (09:13 +0000)]
Cleanup for lov objid code, remove scability problems and
wrong locking. Fix sending last_id into ost.
b=12702
i=johann
i=nathan
yangsheng [Thu, 20 Dec 2007 08:27:54 +0000 (08:27 +0000)]
Branch b1_6
b=14388
i=shadow
i=scjody
Update to SLES10 SP1 latest kernel-2.6.16.54-0.2.3.
bobijam [Thu, 20 Dec 2007 01:58:43 +0000 (01:58 +0000)]
Branch b1_6
b=13715
i=nathan
i=johann
Activate import on intial connection.
grev [Wed, 19 Dec 2007 19:40:31 +0000 (19:40 +0000)]
b=14462
i=Adilger
i=Alexey.Lyashkov
error if dd is not started; use tfile instead of tdir
ericm [Wed, 19 Dec 2007 19:15:35 +0000 (19:15 +0000)]
branch: b1_6
fix AT test: using O_SYNC to trigger bulk write.
b=13659
r=johann, nathan
ericm [Wed, 19 Dec 2007 18:59:48 +0000 (18:59 +0000)]
branch: b1_6
- fix AT timer might not be restarted properly.
- fix AT checksum be ruined by further change in reply.
b=14374
r=nathan, adilger
cfaber [Wed, 19 Dec 2007 07:14:19 +0000 (07:14 +0000)]
test
cfaber [Wed, 19 Dec 2007 07:10:16 +0000 (07:10 +0000)]
test
cfaber [Wed, 19 Dec 2007 07:08:26 +0000 (07:08 +0000)]
test
cfaber [Mon, 17 Dec 2007 19:10:24 +0000 (19:10 +0000)]
test
grev [Mon, 17 Dec 2007 10:21:01 +0000 (10:21 +0000)]
b=13774
i=Adilger
use file with major,minor from /dev/console
yangsheng [Mon, 17 Dec 2007 08:04:26 +0000 (08:04 +0000)]
Branch b1_6
b=14289
i=johann
i=scjody
Update to RHEL5 Update-1 kernel-2.6.18-53.el5.
green [Sun, 16 Dec 2007 04:18:25 +0000 (04:18 +0000)]
b=14379
r=adilger,vitaly
Properly find out duplicate locks when comparing newly receiwed server locks.
mds_open cannot use accmode because it is affected by owneroverride flag
cfaber [Fri, 14 Dec 2007 21:01:34 +0000 (21:01 +0000)]
testing
cfaber [Fri, 14 Dec 2007 20:59:05 +0000 (20:59 +0000)]
testing
cfaber [Fri, 14 Dec 2007 20:30:13 +0000 (20:30 +0000)]
testing
cfaber [Fri, 14 Dec 2007 19:52:07 +0000 (19:52 +0000)]
testing again
cfaber [Fri, 14 Dec 2007 17:25:50 +0000 (17:25 +0000)]
This is a test
This is the note.
scjody [Fri, 14 Dec 2007 16:50:42 +0000 (16:50 +0000)]
cvs test
scjody [Fri, 14 Dec 2007 16:26:23 +0000 (16:26 +0000)]
cvs test
girish [Fri, 14 Dec 2007 15:14:27 +0000 (15:14 +0000)]
Update recommended e2fsprogs version to 1.40.2-cfs5
grev [Fri, 14 Dec 2007 10:05:04 +0000 (10:05 +0000)]
b=13306
i=Adilger
wait delete completed before check kbytesavail
scjody [Thu, 13 Dec 2007 21:00:31 +0000 (21:00 +0000)]
cvs test
scjody [Thu, 13 Dec 2007 20:46:37 +0000 (20:46 +0000)]
cvs test
scjody [Thu, 13 Dec 2007 18:41:23 +0000 (18:41 +0000)]
cvs test rm
scjody [Thu, 13 Dec 2007 18:38:37 +0000 (18:38 +0000)]
cvs test 4
scjody [Thu, 13 Dec 2007 18:16:11 +0000 (18:16 +0000)]
cvs test
scjody [Thu, 13 Dec 2007 18:05:11 +0000 (18:05 +0000)]
cvs test
scjody [Thu, 13 Dec 2007 17:49:17 +0000 (17:49 +0000)]
cvs test
shadow [Thu, 13 Dec 2007 10:20:23 +0000 (10:20 +0000)]
if client_disconnect_export was called without force flag set,
and exist connect request in flight, this can produce access to
NULL pointer (or already free pointer) when connect_interpret
store ocd flags in obd_namespace.
b=14260
i=adilger
i=johann
shadow [Thu, 13 Dec 2007 08:59:01 +0000 (08:59 +0000)]
not alloc memory with spinlock held.
b=14418
i=green
i=johann
bobijam [Thu, 13 Dec 2007 05:51:37 +0000 (05:51 +0000)]
Branch b1_6
b=12211
origianl patch producer=adilger
i=green, bobijam
move obd_fail_check() from being a static inline to being a function.
We already protect this function from being called needlessly by use of
unlikely() and checking obd_fail_loc != 0 before calling it.
Having such a large function inline bloats the code and likely reduces
performance by putting extra (though unlikely) code in every function.
bobijam [Thu, 13 Dec 2007 03:44:04 +0000 (03:44 +0000)]
Branch b1_6
b=14215
i=johann, adilger
user pointer parameter (struct ldlm_res_id*) in ldlm_cli_enqueue_local()
to reduce the stack usage.
grev [Wed, 12 Dec 2007 21:08:16 +0000 (21:08 +0000)]
b=13932
i=Adilger
remove useless "echo"
grev [Wed, 12 Dec 2007 20:31:54 +0000 (20:31 +0000)]
b=13932
i=Adilger
report the slow and excluded tests separately
ericm [Wed, 12 Dec 2007 20:13:42 +0000 (20:13 +0000)]
branch: b1_6
fix a race condition of early reply release.
b=14274
r=nathan, andreas
jxiong [Wed, 12 Dec 2007 07:03:20 +0000 (07:03 +0000)]
fixed some minor issues.
r=alex
grev [Tue, 11 Dec 2007 15:33:00 +0000 (15:33 +0000)]
b=14385
i=Adilger
force test_5c to use mount_client
grev [Tue, 11 Dec 2007 09:27:41 +0000 (09:27 +0000)]
b=13932
i=Adilger
typo fixed (found by Huang Hua)
yury [Mon, 10 Dec 2007 19:21:55 +0000 (19:21 +0000)]
b=14433
r=adilger,shadow,deen
- enable OBD_CONNECT_LRU_RESIZE only when lru resize is enabled in configure time.
grev [Mon, 10 Dec 2007 17:25:05 +0000 (17:25 +0000)]
b=13932
i=Adilger
test_29 cleanup
vitaly [Mon, 10 Dec 2007 15:05:28 +0000 (15:05 +0000)]
Branch b1_6
b=13563
i=green
i=shadow
1) cancel lru locks in ldlm_cli_cancel() and send 1 batched cancel RPC;
2) pass the whole list of canceled locks to the async blocking thread, not
lock-by-lock, and send cancels in batched cancel RPCs.
yury [Sun, 9 Dec 2007 14:25:56 +0000 (14:25 +0000)]
- fixed a typo.
yury [Sun, 9 Dec 2007 13:16:01 +0000 (13:16 +0000)]
b=13766
r=adilger,nikita,shadow
- lots of fixes in ldlm pools code for smoother work and correct memory pressure handling.
scjody [Sun, 9 Dec 2007 00:20:21 +0000 (00:20 +0000)]
Update w/ actual release date
green [Fri, 7 Dec 2007 07:28:16 +0000 (07:28 +0000)]
b=14422
r=johann,adilger
Commit missing parts of bug 10718 to correctly clear lock cache on liblustre.
johann [Thu, 6 Dec 2007 11:32:42 +0000 (11:32 +0000)]
Branch b1_6
b=13843
i=adilger
i=nathan
Fix coding style.
(Dynamically create DLM blocking callback threads)
johann [Thu, 6 Dec 2007 11:27:49 +0000 (11:27 +0000)]
Branch b1_6
b=13843
i=adilger
i=nathan
Dynamically create DLM blocking callback threads
+ fix an issue with MacOS implementation of atomic_sub_and_test().
bobijam [Thu, 6 Dec 2007 04:40:03 +0000 (04:40 +0000)]
Branch b1_6
b=12646
i=johann, shadow
Add opc parameter to ost_checksum_bulk().
bobijam [Thu, 6 Dec 2007 03:49:23 +0000 (03:49 +0000)]
Branch b1_6
b=13715
i=johann, nathan
Don't let out-dated timedout request invalidate the import.
bobijam [Thu, 6 Dec 2007 02:49:02 +0000 (02:49 +0000)]
Branch b1_6
b=14208
i=zhenyu.xu, johann
Replace macro __LINUX__ with __linux__.
green [Thu, 6 Dec 2007 00:58:18 +0000 (00:58 +0000)]
b=10718
r=vitaly,shadow,adilger(some arly versions)
Attach cached pagers to locks that protect those pages. This helps to quickly
find those pages for flush and removal purposes on AST reception and
significantly speeds up lock revocation on big files with pages at
high offsets.
deen [Wed, 5 Dec 2007 20:37:01 +0000 (20:37 +0000)]
Implementation of NFS-like root squash capability.
b=12749
i=adilger
i=johann
johann [Wed, 5 Dec 2007 14:12:15 +0000 (14:12 +0000)]
Branch b1_6
b=12829
i=adilger
i=bobijam
Fix tiny nit in the changelog entry.
johann [Wed, 5 Dec 2007 13:58:17 +0000 (13:58 +0000)]
Branch b1_6
b=12829
i=adilger
i=bobijam
When CRAY_XT3 is defined, the fsgid supplied by the client is
overridden with the primary group provided by the group upcall,
whereas the supplied fsgid can be trust if it is in the list of
supplementary groups returned by the group upcall.
deen [Sat, 1 Dec 2007 15:05:16 +0000 (15:05 +0000)]
Little fixes proposed by Andreas after inspection.
b=11301
grev [Fri, 30 Nov 2007 10:45:27 +0000 (10:45 +0000)]
b=13798
i=nathan.rutman
cleanup: remove tdir creation
grev [Thu, 29 Nov 2007 12:06:09 +0000 (12:06 +0000)]
b=14351
i=Adilger
get rid of relative path in setup_test102
bwzhou [Thu, 29 Nov 2007 05:39:20 +0000 (05:39 +0000)]
Branch b1_6
b=14326
r=johann,bwzhou
Use old size assignment to avoid deadlock caused by ll_inode_size_lock calls down semaphore under spinlock held.
fanyong [Thu, 29 Nov 2007 04:06:24 +0000 (04:06 +0000)]
Branch b1_6
b=14135
i=nathan
i=huanghua
Original "mgc_logname2resid" process both "logname" (which consists of fsname-nodetype)
and "fsname", it can not distinguish "-" contained fsname and the logname.
Split it into two functions: "mgc_logname2resid" and "mgs_fsname2resid" for that.
bobijam [Thu, 29 Nov 2007 01:45:59 +0000 (01:45 +0000)]
Branch b1_6
b=14283
i=zhenyu.xu, oleg.drokin
show "used" as only the used space, to match "df".
grev [Wed, 28 Nov 2007 22:09:01 +0000 (22:09 +0000)]
b=13932
i=Adilger
acc-sm "SLOW" list (all scripts)
walter [Wed, 28 Nov 2007 18:31:29 +0000 (18:31 +0000)]
b=14286
i=grev
i=adilger
branch=b1_6
change rundbench script to use contents of DBENCH_LIB variable to
find the location of the dbench executable (like acceptance-small.sh does)
so that we use the same environment variables to specify the location
of the dbench executable in both acceptance-small.sh and rundbench
instead of using different environment variables to configure the
same thing in the two different scripts.
tianzy [Wed, 28 Nov 2007 12:34:46 +0000 (12:34 +0000)]
A minor fix for bug14036
bobijam [Wed, 28 Nov 2007 02:25:56 +0000 (02:25 +0000)]
Branch b1_6
b=14270
i=johann, kalpak.shah
Description: lfs find does not continue on file error
Details : Continue other files processing when a file/dir is absent.
wangdi [Tue, 27 Nov 2007 23:01:06 +0000 (23:01 +0000)]
Branch: b1_6
b=14016
Fix BUGON when do per-client proc cleanup. Patch from shadow
i=Andreas,Nathan
grev [Tue, 27 Nov 2007 20:20:21 +0000 (20:20 +0000)]
b=14148
i=Adilger
get rid of not default stripe size; use new-style setstripe args
grev [Tue, 27 Nov 2007 14:31:09 +0000 (14:31 +0000)]
b=10825
typo fixed
tianzy [Tue, 27 Nov 2007 11:36:19 +0000 (11:36 +0000)]
b=14036
r=andrew.perepechko
r=johann
Branch b1_6
With this patch, three improvements are included:
1. detete the softlimit in mds and osts when use "lfs quota".
2. display the inaccurate data in the output of "lfs quota".
3. try to get quota info when "lfs quota" is executed.
kalpak [Tue, 27 Nov 2007 11:03:20 +0000 (11:03 +0000)]
b=13919
i=adilger,johann
llapi_semantic_traverse may report wrong filetype for some older kernels and we assumed that suchc files were either directories or regular files. This patch ensures that if DT_UNKNOWN is returned as filetype then it is correctly handled by finding out the correct filetype.
adilger [Tue, 27 Nov 2007 00:30:46 +0000 (00:30 +0000)]
Branch b1_6
Skip swap-on-lustre test 68, it can deadlock on TCP due to inability to
allocate receive buffers.
b=14033
i=jay
grev [Mon, 26 Nov 2007 13:26:22 +0000 (13:26 +0000)]
b=12649
escape "|" on log message to protect interpretation by shell
kalpak [Mon, 26 Nov 2007 09:22:16 +0000 (09:22 +0000)]
b=6334
i=adilger, scjody
Enable MMP and uninit_groups features if the available e2fsprogs and ldiskfs support them
grev [Fri, 23 Nov 2007 13:17:10 +0000 (13:17 +0000)]
b=14148
i=Adilger
use st_blksize instead of 1k while write big file
johann [Fri, 23 Nov 2007 08:42:57 +0000 (08:42 +0000)]
Branch b1_6
b=14022
i=tianzy
i=andrew
By default, enable/disable quotas for both user & group.
anserper [Thu, 22 Nov 2007 22:59:39 +0000 (22:59 +0000)]
Branch b1_6
b=13915
i=johann
i=tianzy
patch 13584
implementation of "lfs quota -t" and "lfs setquota -t" time format string
grev [Thu, 22 Nov 2007 21:41:42 +0000 (21:41 +0000)]
b=14148
i=Adilger
match sanity-quota to t-f; check_runas_id() modification
grev [Thu, 22 Nov 2007 17:09:39 +0000 (17:09 +0000)]
b=10825
i=Scjody
improve the output a bit
komaln [Thu, 22 Nov 2007 11:17:49 +0000 (11:17 +0000)]
b=11230
a small change left undone.
komaln [Thu, 22 Nov 2007 10:53:23 +0000 (10:53 +0000)]
b=11230
r=Adilger
add a test in conf-sanity.sh to verify if set tunables works for symlink device
bobijam [Thu, 22 Nov 2007 03:02:04 +0000 (03:02 +0000)]
Branch b1_6
b=11791
i=adilger, nathan.rutman
Description: Inconsistent usage of lustre_pack_reply()
Details : Standardize the usage of lustre_pack_reply() such that it
always generate a CERROR on failure.
grev [Thu, 22 Nov 2007 01:32:29 +0000 (01:32 +0000)]
b=10825
i=Scjody
print summary of acc-sm tests run
grev [Wed, 21 Nov 2007 21:20:05 +0000 (21:20 +0000)]
b=14054
i=Nathan
force runtests to use 1 mount point
kalpak [Wed, 21 Nov 2007 20:22:12 +0000 (20:22 +0000)]
b=13128
i=adilger
regression test for lfs find -mtime
kalpak [Wed, 21 Nov 2007 20:06:55 +0000 (20:06 +0000)]
b=13128
i=adilger, rupesh
check MDS timestamp only after stat of file.
nathan [Wed, 21 Nov 2007 17:47:58 +0000 (17:47 +0000)]
b=14236
i=adilger
i=johann
don't increase connect time on MGC
grev [Wed, 21 Nov 2007 15:16:49 +0000 (15:16 +0000)]
b=13979
i=Adilger
force replay-ost-single to reformat lustre; remove obsolete fn
shadow [Wed, 21 Nov 2007 12:11:10 +0000 (12:11 +0000)]
fix building problem with 2.6.22 kernel.
b=14255
i=adilger
grev [Wed, 21 Nov 2007 11:04:42 +0000 (11:04 +0000)]
b=14065
i=Nathan
remove tdir creation from sanityN subtests
grev [Wed, 21 Nov 2007 10:24:03 +0000 (10:24 +0000)]
b=13974
i=Nathan
test_43: change exec to multiop
adilger [Tue, 20 Nov 2007 22:59:06 +0000 (22:59 +0000)]
Branch b1_6
Add module license to llite_loop.c so load_modules does not taint kernel.
b=5498
i=jinshan.xiong
i=nathan.rutman
grev [Tue, 20 Nov 2007 21:39:26 +0000 (21:39 +0000)]
b=14087
i=Adilger
set debug level on remote nodes only if modules already loaded