Whamcloud - gitweb
fanyong [Fri, 17 Nov 2006 03:14:29 +0000 (03:14 +0000)]
For change file time operation (utime),
file owner and CAP_FOWNER privilege are preferential.
vitaly [Thu, 16 Nov 2006 10:54:39 +0000 (10:54 +0000)]
Branch b_new_cmd
b=11103
Do not let dentry to be unhashed after create,
some callers, e.g. do_coredump, expect it to be hashed.
fanyong [Thu, 16 Nov 2006 10:22:06 +0000 (10:22 +0000)]
Do UCRED_OLD authentication when gss disabled.
fanyong [Thu, 16 Nov 2006 07:41:42 +0000 (07:41 +0000)]
(1) Remove fs privilege (CAP_FS_MASK) for non-root user, it can result in
invalid authorization.
(2) Do UCRED_OLD authentication for inter-mdt ops and for the case that
client and mdt on the same node. (gss can not distinguish the two cases)
(3) Fix Bug of authority promote for inter-mdt ops.
ericm [Thu, 16 Nov 2006 07:34:58 +0000 (07:34 +0000)]
branch: b_new_cmd
fix gss daemons build, introduced by last commit.
fanyong [Thu, 16 Nov 2006 06:41:44 +0000 (06:41 +0000)]
Drop unnecessary warning message in the scripts, and make it more compatible.
ericm [Thu, 16 Nov 2006 04:33:12 +0000 (04:33 +0000)]
branch: b_new_cmd
port from b1_8_gss: gss daemons sync to nfs-utils-1.0.10-CITI-3
ericm [Thu, 16 Nov 2006 04:03:09 +0000 (04:03 +0000)]
file context_lucid.c was added on branch b_new_cmd on 2006-11-16 04:33:12 +0000
huanghua [Thu, 16 Nov 2006 03:28:25 +0000 (03:28 +0000)]
change debug subsystem according to nathan's cleanup
nikita [Wed, 15 Nov 2006 22:06:10 +0000 (22:06 +0000)]
iam: add more debugging based on dynlock_is_locked().
nikita [Wed, 15 Nov 2006 22:05:15 +0000 (22:05 +0000)]
dynlocks: add dynlock_is_locked() function.
nikita [Wed, 15 Nov 2006 22:04:38 +0000 (22:04 +0000)]
mdd: put md_ucred() back.
yury [Wed, 15 Nov 2006 18:54:50 +0000 (18:54 +0000)]
- small cleanups.
nikita [Wed, 15 Nov 2006 17:26:02 +0000 (17:26 +0000)]
iam: 0. fix assertion in split_index_node(). 1. remove ext3_journal_get_write_access() from ext3_append().
nikita [Wed, 15 Nov 2006 17:24:38 +0000 (17:24 +0000)]
mkfs_lustre: let us try smaller inodes.
yury [Wed, 15 Nov 2006 14:14:51 +0000 (14:14 +0000)]
- fixes in lov to make it a bit more parallel creates friendly.
nikita [Wed, 15 Nov 2006 13:27:28 +0000 (13:27 +0000)]
lu_object: return back old hash: hash_long() proved to be a disadvantage.
nikita [Wed, 15 Nov 2006 13:26:03 +0000 (13:26 +0000)]
iam: disable dx_bug11027_check().
yury [Wed, 15 Nov 2006 11:49:11 +0000 (11:49 +0000)]
- modules order fixed
wangdi [Wed, 15 Nov 2006 11:16:00 +0000 (11:16 +0000)]
Branch:b_new_cmd
obd_recoverying should be set to 0 before obd_notify in mdd_recovery_complete.
otherwise llog may not be initialized correctly in recovery.
b:11217
huanghua [Wed, 15 Nov 2006 10:28:34 +0000 (10:28 +0000)]
(1) mdd_name_insert(), the return value is lost;
(2) The cmr_create() should check the child existence at very first.
nikita [Wed, 15 Nov 2006 00:13:19 +0000 (00:13 +0000)]
iam: fix pdirops locking during leaf split.
yury [Tue, 14 Nov 2006 17:39:19 +0000 (17:39 +0000)]
- added more stats to mdt operations;
- fixes in modules load order.
brian [Tue, 14 Nov 2006 16:28:23 +0000 (16:28 +0000)]
Test the value of the failo{ver,ut} flags before deciding to add a format
param.
yury [Tue, 14 Nov 2006 16:21:20 +0000 (16:21 +0000)]
- fixed race in lmv about fid allocation. It allocated fid on one target and sent create operation to anorher mds if dir got split in the middle;
- added proc stats to mdt.
nikita [Tue, 14 Nov 2006 13:13:55 +0000 (13:13 +0000)]
add more debugging for unlink -ENOENT bug on buffalo.
ericm [Tue, 14 Nov 2006 04:13:08 +0000 (04:13 +0000)]
branch: b_new_cmd
b=11215
fix wrong assertion.
nikita [Mon, 13 Nov 2006 22:37:37 +0000 (22:37 +0000)]
jbd: add jbd-checkpoint-on-commit.patch.
nikita [Mon, 13 Nov 2006 17:34:14 +0000 (17:34 +0000)]
osd: debugging dump index leaf node when osd_inode_remove() fails to find a record in oi.
nikita [Mon, 13 Nov 2006 17:33:32 +0000 (17:33 +0000)]
iam: debugging code for oi -ENOENT error.
nikita [Mon, 13 Nov 2006 14:14:29 +0000 (14:14 +0000)]
lu_object: use hash_long() to hash fids.
nikita [Mon, 13 Nov 2006 12:17:39 +0000 (12:17 +0000)]
osd: put back fid_is_local() assertion.
nikita [Mon, 13 Nov 2006 12:17:17 +0000 (12:17 +0000)]
mdd_unlink(): fix format.
vitaly [Mon, 13 Nov 2006 10:48:26 +0000 (10:48 +0000)]
Branch b_new_cmd
b=22564
eliminate the access to freed by ll_done_writing @och in ll_close_inode_openhandle.
fanyong [Mon, 13 Nov 2006 06:15:54 +0000 (06:15 +0000)]
(1) add some condition check.
(2) simplify lprocfs_wr_identity_flush.
(3) add some comment.
fanyong [Mon, 13 Nov 2006 06:03:40 +0000 (06:03 +0000)]
Drop unnecessary check, update remote perm anyway.
huanghua [Mon, 13 Nov 2006 03:58:35 +0000 (03:58 +0000)]
fix memleak in error case.
tappro [Sun, 12 Nov 2006 10:17:00 +0000 (10:17 +0000)]
- fix small typo
tappro [Sun, 12 Nov 2006 10:15:29 +0000 (10:15 +0000)]
- return MIN_PRECREATE to the initial value - 32
- check the llog and echo groups and do real create always for them
tappro [Sat, 11 Nov 2006 15:08:33 +0000 (15:08 +0000)]
remove voluntary preempt from configs
tappro [Sat, 11 Nov 2006 08:32:58 +0000 (08:32 +0000)]
disable lookup() in sanity checks for create temporary
tappro [Sat, 11 Nov 2006 08:26:34 +0000 (08:26 +0000)]
add debug info
nikita [Sat, 11 Nov 2006 01:02:31 +0000 (01:02 +0000)]
iam: add debugging for bug 11027.
tappro [Fri, 10 Nov 2006 20:10:59 +0000 (20:10 +0000)]
- avoid assert case in mdd_unlink() while building the transaction parameters.
The issue is sanity check was done too late.
brian [Fri, 10 Nov 2006 18:11:44 +0000 (18:11 +0000)]
Support the lmc --failout argument and set the failover mode in the csv
file accordingly.
nikita [Fri, 10 Nov 2006 16:46:00 +0000 (16:46 +0000)]
iam: take dx lock when inserting new pointer into index node.
yury [Fri, 10 Nov 2006 16:17:17 +0000 (16:17 +0000)]
- detect cross-ref cases in MDT using special flag rather than by using if (name[0] == 0);
- fixes in check split optimization.
tappro [Fri, 10 Nov 2006 09:52:42 +0000 (09:52 +0000)]
- add more stats in mdd
ericm [Fri, 10 Nov 2006 00:31:44 +0000 (00:31 +0000)]
branch: b_new_cmd
b=11185
only register gss policy after it is ready.
nikita [Thu, 9 Nov 2006 10:35:57 +0000 (10:35 +0000)]
mkfs_lustre.c: cap journal size at 2GB to avoid overflows.
wangdi [Thu, 9 Nov 2006 08:19:14 +0000 (08:19 +0000)]
Branch: b_new_cmd
it should return -ENOENT instead of ESTALE, if it is dead obj in mdd_lookup.
found by nikita
lsy [Thu, 9 Nov 2006 07:51:57 +0000 (07:51 +0000)]
update comment according to code change.
ericm [Thu, 9 Nov 2006 03:22:54 +0000 (03:22 +0000)]
branch: b_new_cmd
debug on mountain.
ericm [Wed, 8 Nov 2006 20:28:16 +0000 (20:28 +0000)]
branch: b_new_cmd
fix null pointer reference in CERROR.
nikita [Wed, 8 Nov 2006 20:25:25 +0000 (20:25 +0000)]
kgdb: use correct macro version.
nikita [Wed, 8 Nov 2006 20:24:54 +0000 (20:24 +0000)]
jbd: add a patch to create shadow-copies for all transaction buffers.
lsy [Wed, 8 Nov 2006 16:04:33 +0000 (16:04 +0000)]
forgot pack capa for iocontrol EXT3_IOC_SETFLAGS.
in filter_auth_capa fid might be NULL.
yury [Wed, 8 Nov 2006 15:59:43 +0000 (15:59 +0000)]
- fixed possible race in ll_close_inode_openhandle()
yury [Wed, 8 Nov 2006 15:33:17 +0000 (15:33 +0000)]
- remove mkfs hjournal upper limit again after mk2fs is patched.
lsy [Wed, 8 Nov 2006 15:07:05 +0000 (15:07 +0000)]
more verbose message.
nikita [Wed, 8 Nov 2006 13:01:50 +0000 (13:01 +0000)]
mkfs_lustre: return journal size limit back: buffalo tests fail without this.
lsy [Wed, 8 Nov 2006 10:49:22 +0000 (10:49 +0000)]
add debug code for capa cleanup failure.
rename filter_verify_capa to filter_auth_capa.
yury [Wed, 8 Nov 2006 08:44:25 +0000 (08:44 +0000)]
- fixed pssible NULL pointer access in lprocfs_rd_conn_uuid();
yury [Wed, 8 Nov 2006 08:30:42 +0000 (08:30 +0000)]
- fixed pssibly using freed och in ll_close_inode_openhandle().
yury [Tue, 7 Nov 2006 20:07:27 +0000 (20:07 +0000)]
- do not check dir for split if cleint already knows that it is split.
tappro [Tue, 7 Nov 2006 18:25:41 +0000 (18:25 +0000)]
fix in patch
tappro [Tue, 7 Nov 2006 16:26:52 +0000 (16:26 +0000)]
return minor changes in config back
yury [Tue, 7 Nov 2006 16:18:44 +0000 (16:18 +0000)]
- added -ERESTART handling for link, unlink and rename on client. This is not case for tests, because tests do not contain concurrent create vs. rename, but this is correctly from overall point of view of FS functionaling. So it should be.
yury [Tue, 7 Nov 2006 14:21:44 +0000 (14:21 +0000)]
- added stats to cmm split;
- get rid of second attr_get() in cmm_split_check().
fanyong [Tue, 7 Nov 2006 14:07:09 +0000 (14:07 +0000)]
(1) provide locked and unlocked version for mdd_permission.
(2) fix some outdated comment.
(3) other small fix.
tappro [Tue, 7 Nov 2006 12:47:13 +0000 (12:47 +0000)]
fix broken compile
huanghua [Tue, 7 Nov 2006 10:34:00 +0000 (10:34 +0000)]
fix error handling; kill possible memleak.
tappro [Tue, 7 Nov 2006 07:32:19 +0000 (07:32 +0000)]
- update default configs
tappro [Tue, 7 Nov 2006 07:31:06 +0000 (07:31 +0000)]
add highmem-split patch from HEAD to series
fanyong [Tue, 7 Nov 2006 06:51:37 +0000 (06:51 +0000)]
(1) drop unnecessary mdd_read_lock.
(2) add some comment.
(3) cleanup the code.
fanyong [Tue, 7 Nov 2006 05:06:07 +0000 (05:06 +0000)]
Pass lu_attr as arg for __mdd_permission_internal explicitly.
fanyong [Tue, 7 Nov 2006 03:31:05 +0000 (03:31 +0000)]
Small fix, cleanup the code.
lsy [Tue, 7 Nov 2006 03:05:32 +0000 (03:05 +0000)]
rename op_mod_capa? to op_capa?. (original mod already means md_op_data)
nikita [Mon, 6 Nov 2006 22:33:43 +0000 (22:33 +0000)]
cmm,osd: add debugging to catch bug11150.
yury [Mon, 6 Nov 2006 18:40:11 +0000 (18:40 +0000)]
- use separate create lock for each group on OST, so that, creates from different MDS may be handled in parallel;
- use separate destroy_in_progress flag for each group on OST.
nikita [Mon, 6 Nov 2006 16:53:46 +0000 (16:53 +0000)]
mkfs.lustre: remove upper limit on journal size: we want larger journal.
yury [Mon, 6 Nov 2006 16:24:03 +0000 (16:24 +0000)]
- all md_op_data field have prefix op_;
- use name[0] == 0 to check if this is partial operation.
tappro [Mon, 6 Nov 2006 14:36:49 +0000 (14:36 +0000)]
return TEA hash back
tappro [Mon, 6 Nov 2006 14:34:34 +0000 (14:34 +0000)]
another part of -ERESTART handling
tappro [Mon, 6 Nov 2006 14:20:56 +0000 (14:20 +0000)]
- finish reqp if -ERESTART received, add debug
nikita [Mon, 6 Nov 2006 11:21:29 +0000 (11:21 +0000)]
osd: fix ipd handling for iterators.
wangdi [Mon, 6 Nov 2006 10:36:49 +0000 (10:36 +0000)]
Branch: b_new_cmd
add debug_info for 11123
fanyong [Mon, 6 Nov 2006 07:19:06 +0000 (07:19 +0000)]
Check WRITE permission of parent when create.
yury [Sun, 5 Nov 2006 14:12:23 +0000 (14:12 +0000)]
- optimization for open. In case mdt_reint_open() sees that child object does not exist and should be created, tell mdd do not lookup child object again in sanity check.
tappro [Sun, 5 Nov 2006 12:37:04 +0000 (12:37 +0000)]
correct block/inode ratio for OST, don't use defaults
tappro [Sun, 5 Nov 2006 12:31:13 +0000 (12:31 +0000)]
- enable 4G of memory usage, disable preemt
nikita [Sat, 4 Nov 2006 23:24:09 +0000 (23:24 +0000)]
osd/mdd: new (weaker) locking for ->do_attr_{g,e}set(), see message to colibri@ for details.
nikita [Sat, 4 Nov 2006 21:14:47 +0000 (21:14 +0000)]
mdd: little cleanup.
nikita [Sat, 4 Nov 2006 21:14:11 +0000 (21:14 +0000)]
0. iam: pub r5 hash back.
1. iam/osd: fix a race that caused 11150.
yury [Sat, 4 Nov 2006 15:33:38 +0000 (15:33 +0000)]
- remove debug stuff for b11150
tappro [Sat, 4 Nov 2006 15:30:35 +0000 (15:30 +0000)]
- make 16K bytes/inode ratio only for really big devices > 100Gb. This is
workaround against -ENOSPC on OST due to lack of inodes
yury [Sat, 4 Nov 2006 08:27:23 +0000 (08:27 +0000)]
- cleanups and fixes in fid/lproc* stuff;
- in mdd_dir_page_build() added check for wrong hash.
- in mdd_dir_page_build() fixed using not swabbed fid.
huanghua [Sat, 4 Nov 2006 07:31:39 +0000 (07:31 +0000)]
(1) check whether reply buffer has been packed;
(2) shrink reply buffer if needed;
(3) call ptlrpc_lprocfs_unregister_obd() while cleanup.
huanghua [Sat, 4 Nov 2006 06:08:00 +0000 (06:08 +0000)]
fix a typo
ericm [Fri, 3 Nov 2006 21:06:07 +0000 (21:06 +0000)]
branch: b_new_cmd
after negotiation handling might invalidate the whole negotiation, do
proper cleanup, and client don't repeatedly try (bug found by fanyong).