Whamcloud - gitweb
fs/lustre-release.git
19 years agofile 2.6-vanilla.target was initially added on branch b1_2_smallfix.
jacob [Wed, 23 Jun 2004 04:54:09 +0000 (04:54 +0000)]
file 2.6-vanilla.target was initially added on branch b1_2_smallfix.

19 years agofile ldiskfs-2.6-vanilla.series was initially added on branch b1_2_smallfix.
jacob [Wed, 23 Jun 2004 04:54:08 +0000 (04:54 +0000)]
file ldiskfs-2.6-vanilla.series was initially added on branch b1_2_smallfix.

19 years agofile 2.6-vanilla.series was initially added on branch b1_2_smallfix.
jacob [Wed, 23 Jun 2004 04:54:07 +0000 (04:54 +0000)]
file 2.6-vanilla.series was initially added on branch b1_2_smallfix.

19 years agofile vfs_nointent-2.6-vanilla.patch was initially added on branch b1_2_smallfix.
jacob [Wed, 23 Jun 2004 04:54:06 +0000 (04:54 +0000)]
file vfs_nointent-2.6-vanilla.patch was initially added on branch b1_2_smallfix.

19 years agofile vfs_intent-2.6-vanilla.patch was initially added on branch b1_2_smallfix.
jacob [Wed, 23 Jun 2004 04:54:05 +0000 (04:54 +0000)]
file vfs_intent-2.6-vanilla.patch was initially added on branch b1_2_smallfix.

19 years agofile removepage-2.6-vanilla.patch was initially added on branch b1_2_smallfix.
jacob [Wed, 23 Jun 2004 04:54:03 +0000 (04:54 +0000)]
file removepage-2.6-vanilla.patch was initially added on branch b1_2_smallfix.

19 years ago- mds_preprw() handles failed mds_fid2dentry() properly
alex [Wed, 23 Jun 2004 04:21:46 +0000 (04:21 +0000)]
- mds_preprw() handles failed mds_fid2dentry() properly

19 years ago- adding CONFIG_KALLSYMS to kksymoops patch
ccooper [Wed, 23 Jun 2004 02:33:14 +0000 (02:33 +0000)]
- adding CONFIG_KALLSYMS to kksymoops patch

19 years ago- three lmv obj leaks fixed
alex [Tue, 22 Jun 2004 15:50:42 +0000 (15:50 +0000)]
- three lmv obj leaks fixed
- cross-node mkdir() splits newly-created directory if requested
- remove_entries_from_orig_dir() accepts cross-node dir entries
- sanity check 3a tests splitting a directory with cross-node entries

19 years agoSmall fix about using LTIME_S macro
yury [Tue, 22 Jun 2004 10:47:54 +0000 (10:47 +0000)]
Small fix about using LTIME_S macro

19 years agofile ext3-mballoc-2.4.21-chaos.patch was initially added on branch b1_4_mballoc.
zab [Mon, 21 Jun 2004 21:59:24 +0000 (21:59 +0000)]
file ext3-mballoc-2.4.21-chaos.patch was initially added on branch b1_4_mballoc.

19 years agofile ext3-extents-asyncdel-2.4.21-chaos.patch was initially added on branch b1_4_mballoc.
zab [Mon, 21 Jun 2004 21:59:23 +0000 (21:59 +0000)]
file ext3-extents-asyncdel-2.4.21-chaos.patch was initially added on branch b1_4_mballoc.

19 years agofile ext3-extents-2.4.21-chaos.patch was initially added on branch b1_4_mballoc.
zab [Mon, 21 Jun 2004 21:59:22 +0000 (21:59 +0000)]
file ext3-extents-2.4.21-chaos.patch was initially added on branch b1_4_mballoc.

19 years agofile exports-2.4.21-chaos.patch was initially added on branch b1_4_mballoc.
zab [Mon, 21 Jun 2004 21:59:21 +0000 (21:59 +0000)]
file exports-2.4.21-chaos.patch was initially added on branch b1_4_mballoc.

19 years ago- minor fix: lmv object leak in lmv_get_mea_and_update_object()
alex [Mon, 21 Jun 2004 16:42:05 +0000 (16:42 +0000)]
- minor fix: lmv object leak in lmv_get_mea_and_update_object()
- CERROR() against leaked lmv objects

19 years agoFixes with refcounting.
yury [Mon, 21 Jun 2004 15:28:37 +0000 (15:28 +0000)]
Fixes with refcounting.

19 years ago- this test shows lmv obj refcounting problem
alex [Mon, 21 Jun 2004 14:54:16 +0000 (14:54 +0000)]
- this test shows lmv obj refcounting problem

19 years agoAdded splitted object handling in lmv_getattr().
yury [Mon, 21 Jun 2004 14:25:00 +0000 (14:25 +0000)]
Added splitted object handling in lmv_getattr().

19 years agoobk->freeing changed to obj->state and O_FREEING flag is added instead.
yury [Mon, 21 Jun 2004 12:16:29 +0000 (12:16 +0000)]
obk->freeing changed to obj->state and O_FREEING flag is added instead.

19 years agoFixed possible sleep in invalid context in LMV object manager.
yury [Mon, 21 Jun 2004 11:38:25 +0000 (11:38 +0000)]
Fixed possible sleep in invalid context in LMV object manager.

19 years agoCWARN() chnaged to CDEBUG in lmv_delete_object().
yury [Mon, 21 Jun 2004 11:07:25 +0000 (11:07 +0000)]
CWARN() chnaged to CDEBUG in lmv_delete_object().

19 years ago- added md_delete_object() and lmv_delete_object(). This is needed to remove LMV...
yury [Mon, 21 Jun 2004 11:03:06 +0000 (11:03 +0000)]
- added md_delete_object() and lmv_delete_object(). This is needed to remove LMV object on delete_inode() call.
- small fixes in md_*() methods (added missed ENTRY).

19 years agoLocking fix in lmv_destroy_obj().
yury [Sat, 19 Jun 2004 17:23:52 +0000 (17:23 +0000)]
Locking fix in lmv_destroy_obj().

19 years ago- added mutex to lmv_obj to protect slaves attrs.
yury [Sat, 19 Jun 2004 17:15:26 +0000 (17:15 +0000)]
- added mutex to lmv_obj to protect slaves attrs.

- changed semantic of lmv_create_obj(). By now it
returns created object, as mostly this is what all expect from it.

19 years agoFixes with managing objects in LMV object manager.
yury [Sat, 19 Jun 2004 13:29:21 +0000 (13:29 +0000)]
Fixes with managing objects in LMV object manager.

19 years ago- added lmv_destroy_obj() which will be used for destroying splitted object
yury [Fri, 18 Jun 2004 16:04:20 +0000 (16:04 +0000)]
- added lmv_destroy_obj() which will be used for destroying splitted object
on inode deleting.

19 years ago- added proper ref counting in lmv object manager.
yury [Fri, 18 Jun 2004 15:48:31 +0000 (15:48 +0000)]
- added proper ref counting in lmv object manager.
- fixed possible race in lmv_create_obj().
- cleanups in API of lmv object manager.
- removed some redundant checks for return code.
- added error handling for lmv_create_obj() in various places.
- coding style and indentation fixes in lmv code.

19 years agorelease the parent locks in error handler in mds_get_parents_children_locked
wangdi [Fri, 18 Jun 2004 10:22:26 +0000 (10:22 +0000)]
release the parent locks in error handler in mds_get_parents_children_locked

19 years agob=3643
green [Thu, 17 Jun 2004 16:40:56 +0000 (16:40 +0000)]
b=3643
Only try to look for GNS mount objects inside of directories.

19 years agoFixed incorrect attempt to read clients configuration when mds is mounted for writing...
yury [Thu, 17 Jun 2004 16:01:39 +0000 (16:01 +0000)]
Fixed incorrect attempt to read clients configuration when mds is mounted for writing logs.

19 years ago- added error handling in various places of lmv code.
yury [Thu, 17 Jun 2004 15:23:22 +0000 (15:23 +0000)]
- added error handling in various places of lmv code.

- fixed error code leakage in filter_precreate() what led to endless loop
in the case of -ENOSPC due to real lack of space or inability to allocate new inode.

- added various error messages in error code pathes in lmv code.
- small changes in existent error messages.

19 years agoadd jounal blocks in ext3_init_extents_ea
wangdi [Thu, 17 Jun 2004 07:49:14 +0000 (07:49 +0000)]
add jounal blocks in ext3_init_extents_ea

19 years agofile lov_request.c was initially added on branch b_lovcleanup.
niu [Thu, 17 Jun 2004 06:21:00 +0000 (06:21 +0000)]
file lov_request.c was initially added on branch b_lovcleanup.

19 years agofile lov_qos.c was initially added on branch b_lovcleanup.
niu [Thu, 17 Jun 2004 06:20:59 +0000 (06:20 +0000)]
file lov_qos.c was initially added on branch b_lovcleanup.

19 years agofile lov_offset.c was initially added on branch b_lovcleanup.
niu [Thu, 17 Jun 2004 06:20:58 +0000 (06:20 +0000)]
file lov_offset.c was initially added on branch b_lovcleanup.

19 years agofile lov_merge.c was initially added on branch b_lovcleanup.
niu [Thu, 17 Jun 2004 06:20:57 +0000 (06:20 +0000)]
file lov_merge.c was initially added on branch b_lovcleanup.

19 years ago- lmv_set_mdc_active() should skip NULL exports
alex [Wed, 16 Jun 2004 14:16:30 +0000 (14:16 +0000)]
- lmv_set_mdc_active() should skip NULL exports
  actually, such an export could be connection back to owner mds

19 years ago1) add cmobd support in lconf and lmc
wangdi [Wed, 16 Jun 2004 13:48:59 +0000 (13:48 +0000)]
1) add cmobd support in lconf and lmc
2) remove and modify some test scripts according to the new lconf and lmc

19 years ago* pass in the 3rd parameter for obdgens while getting lov stripe config.
niu [Wed, 16 Jun 2004 07:55:13 +0000 (07:55 +0000)]
* pass in the 3rd parameter for obdgens while getting lov stripe config.

19 years agofile configurable-x86-stack-2.4.21-chaos.patch was initially added on branch b1_2_sma...
adilger [Wed, 16 Jun 2004 07:46:05 +0000 (07:46 +0000)]
file configurable-x86-stack-2.4.21-chaos.patch was initially added on branch b1_2_smallfix.

19 years agob=2837
alex [Wed, 16 Jun 2004 07:40:59 +0000 (07:40 +0000)]
b=2837

- minor optimization: MDS doesn't take UPDATE lock to protect
  from splitting if splitting isn't possible (slave objects)

19 years agob=1742
uid723 [Wed, 16 Jun 2004 07:33:21 +0000 (07:33 +0000)]
b=1742

- landed on HEAD

19 years agob=1742
uid723 [Wed, 16 Jun 2004 07:24:02 +0000 (07:24 +0000)]
b=1742

- landed on HEAD

19 years agob=3063
alex [Wed, 16 Jun 2004 06:38:18 +0000 (06:38 +0000)]
b=3063

- 2nd patch of the bug landed

19 years agob=2837
alex [Tue, 15 Jun 2004 20:15:19 +0000 (20:15 +0000)]
b=2837

- revert back old change to IT_OPEN handling. it shouldn't take UPDATE
  lock on the inode

19 years ago- adding kksymoops patch to vanilla-2.4.24 series
ccooper [Tue, 15 Jun 2004 12:55:25 +0000 (12:55 +0000)]
- adding kksymoops patch to vanilla-2.4.24 series

19 years agoFixes in mdsadd.sh
yury [Tue, 15 Jun 2004 10:35:26 +0000 (10:35 +0000)]
Fixes in mdsadd.sh

19 years agoget rid of some unused structure for smfs
wangdi [Tue, 15 Jun 2004 10:34:09 +0000 (10:34 +0000)]
get rid of some unused structure for smfs

19 years agob=2837
alex [Tue, 15 Jun 2004 08:47:51 +0000 (08:47 +0000)]
b=2837

- initialize all fields in op_data

19 years agofailover.sh renamed to mds.sh. This is a script for testing mds addiotion. Currently...
yury [Tue, 15 Jun 2004 07:52:34 +0000 (07:52 +0000)]
failover.sh renamed to mds.sh. This is a script for testing mds addiotion. Currently it does not use failover at all.

19 years agob=2837
alex [Mon, 14 Jun 2004 15:27:21 +0000 (15:27 +0000)]
b=2837
b=3582

  inproper kunmap() usage fixed

19 years agoget rid of setting stripe info of parent inode in mds_create_objects. which will...
wangdi [Mon, 14 Jun 2004 14:37:45 +0000 (14:37 +0000)]
get rid of setting stripe info of parent inode in mds_create_objects. which will conflict with mea in dir

19 years agoSmall fix in liblustre.
yury [Mon, 14 Jun 2004 06:47:32 +0000 (06:47 +0000)]
Small fix in liblustre.

19 years agoland b_ost_amd onto HEAD.
phil [Mon, 14 Jun 2004 05:10:57 +0000 (05:10 +0000)]
land b_ost_amd onto HEAD.

I haven't re-tested the AMD functionality since landing, but I did
test iozone and dbench to ensure sanity.

19 years agoin create obj, the inode should get stripe info from parent for sanity 65b. get fix...
wangdi [Sun, 13 Jun 2004 17:58:06 +0000 (17:58 +0000)]
in create obj, the inode should get stripe info from parent for sanity 65b. get fix from b1_2smallfix

19 years agoAdded client disconnect error message.
yury [Sun, 13 Jun 2004 10:51:16 +0000 (10:51 +0000)]
Added client disconnect error message.

19 years agoAdded the following:
yury [Sun, 13 Jun 2004 10:24:38 +0000 (10:24 +0000)]
Added the following:
- lmv_iocontrol() which currently passes all commands to underlying mds's.
- some stuff in lproc_lmv.c, which shows now number of obds, number of active obds, etc.
- some cleanups, checks and error messages in lmv_obd.c.

19 years agomerge tiny patch from HP for lfsck
phil [Sat, 12 Jun 2004 14:33:47 +0000 (14:33 +0000)]
merge tiny patch from HP for lfsck

19 years ago- moved dir entries are deleted from the original dir (master object)
alex [Fri, 11 Jun 2004 15:08:41 +0000 (15:08 +0000)]
- moved dir entries are deleted from the original dir (master object)
- unlink for splitted dir has been implemented. it uses IT_UNLINK to check
  dir's emptiness and works as following: unlink requests comes to mds holding
  dir; mds recognizes splitted dir and issues LCK_EX with IT_UNLINK intent.
  each slave object is checked and locked. if all the slaves and master object
  are empty, then mds unlinks them and unlocks slave objects
- lmv_enqueue() handles splitted dir properly: issues given lock for each object
- lmv_unlink() handles splitted dir properly: unlinks each slave object
- mds_lock_slave_objs(), mds_unlock_slave_objs() and mds_unlink_slave_objs()
  have been introduced to implement splitted dir unlink
- mds_lock_and_check_slave() is IT_UNLINK handler on mds side
- mds_reint_unlink() sets MDS_MODE_REPLAY on "drop nlink" request in replay case
- mds_reint_unlink() recognizes MDS_MODE_DONT_LOCK and doesn't try to lock
  slave object being removed (it gets locked during earlier)
- minor cleanups in lmv to avoid needless debug messages
- sanity-lmv.sh has been added

19 years agoFixes and cleanups in lmv.
yury [Fri, 11 Jun 2004 11:47:19 +0000 (11:47 +0000)]
Fixes and cleanups in lmv.

Fixed:
* failure path in lmv_connect().
* added checks if late connect is success.
* added accounting of active mds.

19 years agominor fix in duplicate_file
wangdi [Fri, 11 Jun 2004 06:20:07 +0000 (06:20 +0000)]
minor fix in duplicate_file

19 years agoin smfs_fsync, file maybe NULL sometimes, so get cache_inode for dentry
wangdi [Fri, 11 Jun 2004 06:13:53 +0000 (06:13 +0000)]
in smfs_fsync, file maybe NULL sometimes, so get cache_inode for dentry

19 years agofile ext3-pdirops-2.4.20-rh.patch was initially added on branch b1_2_smallfix.
jacob [Thu, 10 Jun 2004 23:15:46 +0000 (23:15 +0000)]
file ext3-pdirops-2.4.20-rh.patch was initially added on branch b1_2_smallfix.

19 years ago1)add netconsole-2.4.24
wangdi [Thu, 10 Jun 2004 15:45:05 +0000 (15:45 +0000)]
1)add netconsole-2.4.24
2)add journal blocks in ext3_ext_in_ea_new_extent, for its reservation is not enough. (tmp fix here)

19 years agoMostly fixes related to using failover MDSs along lmv. This is also related to active...
yury [Thu, 10 Jun 2004 15:10:57 +0000 (15:10 +0000)]
Mostly fixes related to using failover MDSs along lmv. This is also related to active_ref handling, etc.

Also fixed the following:
* name of lproc entry where lustre upcall can be seen in error message if upcall is failed.
* removed --master option which was supposed to work along with lmv and now it is obsoleted.
* clanups and fixes both in lconf and lmc

19 years agominor fix in mds_finish_open
wangdi [Thu, 10 Jun 2004 09:07:17 +0000 (09:07 +0000)]
minor fix in mds_finish_open

19 years agob=3560
zab [Tue, 8 Jun 2004 16:22:56 +0000 (16:22 +0000)]
b=3560
filter_sync_llog() was double locking a spin lock

19 years ago- lmv should accept IT_CHDIR intent
alex [Tue, 8 Jun 2004 10:22:49 +0000 (10:22 +0000)]
- lmv should accept IT_CHDIR intent

19 years ago1)do precreate record in obdfilter
wangdi [Tue, 8 Jun 2004 09:17:39 +0000 (09:17 +0000)]
1)do precreate record in obdfilter
2)fix bugs in fsfilt_commit in smfs, sometime inode will be null sometime

19 years agodrivers/hotplug/pci_hotplug_core.c apparently calls lookup_hash, but
phil [Tue, 8 Jun 2004 06:00:19 +0000 (06:00 +0000)]
drivers/hotplug/pci_hotplug_core.c apparently calls lookup_hash, but
wasn't updated for new nameidata arg

19 years agoland b_gns onto HEAD. If you are working on CMD, you MUST UPDATE YOUR
phil [Mon, 7 Jun 2004 22:41:07 +0000 (22:41 +0000)]
land b_gns onto HEAD.  If you are working on CMD, you MUST UPDATE YOUR
KERNEL, or LUSTRE WILL NOT BUILD.

- full GNS kernel updates added to vanilla-2.4.24 series
- Oleg patched the API updates into the other kernel series, but none
  of the GNS functionality is present.  I don't know how well those
  other series have been tested.
- I tested normal functionality, but have not tested GNS yet.

19 years agoadd some cvsignore bits
phil [Mon, 7 Jun 2004 20:19:10 +0000 (20:19 +0000)]
add some cvsignore bits

19 years ago- import tracks when connection procedure started. if connection error happens
alex [Sun, 6 Jun 2004 09:06:52 +0000 (09:06 +0000)]
- import tracks when connection procedure started. if connection error happens
  after HZ then ptlrpc_connect_interpret() starts reconnection immediately.
  this speeds recovery up a bit
- pinger pings inter-mds connection more ofter then usual ones
- pinger finds the most sooner event and waits for it, not for obd_timeout * 2
- mds_getattr_name() needs not to check whether mds is choosen right if
  attrs are requested by fid

19 years ago- merge 2 weeks of b1_4 fixes onto HEAD
phil [Fri, 4 Jun 2004 15:14:58 +0000 (15:14 +0000)]
- merge 2 weeks of b1_4 fixes onto HEAD
- b1_4 is basically the parent of HEAD, because that's the direction
  that changes flow, as strange as that sounds.  so there's a
  HEAD_BASE tag which sits on b1_4

19 years ago- typo fixed
alex [Fri, 4 Jun 2004 13:00:09 +0000 (13:00 +0000)]
- typo fixed

19 years agomove mds_check_mds_num right after fill_inode to avoid race.
wangdi [Fri, 4 Jun 2004 11:47:07 +0000 (11:47 +0000)]
move mds_check_mds_num right after fill_inode to avoid race.

19 years agoadd -ERESTART handler for mds_getattr_name
wangdi [Fri, 4 Jun 2004 11:13:41 +0000 (11:13 +0000)]
add -ERESTART handler for mds_getattr_name

19 years agoIt should only be in b_apple_develop now, not in HEAD
liangzhen [Thu, 3 Jun 2004 17:39:52 +0000 (17:39 +0000)]
It should only be in b_apple_develop now, not in HEAD

19 years agoIt should only be in b_apple_develop
liangzhen [Thu, 3 Jun 2004 17:37:10 +0000 (17:37 +0000)]
It should only be in b_apple_develop

19 years agoProject file for xcodebuild
liangzhen [Thu, 3 Jun 2004 17:11:20 +0000 (17:11 +0000)]
Project file for xcodebuild

19 years agostatmany error if one of "-els" isn't given
phil [Thu, 3 Jun 2004 15:12:00 +0000 (15:12 +0000)]
statmany error if one of "-els" isn't given

19 years agochanges to protect directory splitting from concurrent modifies/lookups:
alex [Thu, 3 Jun 2004 11:23:51 +0000 (11:23 +0000)]
changes to protect directory splitting from concurrent modifies/lookups:

 - mds_fid2locked_dentry() takes UPDATE lock on the inode. depending on the
   given mode for LOOKUP lock new routine mds_lock_mode_for_dir() learns
   what mode to use for UPDATE lock. it could be:
   - LCK_CR - lookup case. it protects directory from concurrent splitting
     and don't invalidate client cache readdir()
   - LCK_CW - modify case. it protects directory from concurrent splitting
     and invalidate client cache for readdir()
   - LCK_EX - modify with possible splitting. protects from any parallel access

 - mds_getattr_name(), mds_open(), mds_reint_setattr(), mds_reint_create(),
   mds_get_parent_child_locked(), mds_reint_unlink(), mds_reint_rename() and
   mds_get_parents_children_lock() have been modified to play new game with
   directory locking

 - mds_splitting_expected() predicts splitting possibility

19 years agoget rid of --smfs-enable options after merge in HEAD
wangdi [Thu, 3 Jun 2004 09:05:49 +0000 (09:05 +0000)]
get rid of --smfs-enable options after merge in HEAD

19 years agoupdate vanilla-2.4.24 series
wangchao [Thu, 3 Jun 2004 04:38:30 +0000 (04:38 +0000)]
update vanilla-2.4.24 series

19 years agofile Makefile.mk was initially added on branch b1_4_openib.
eeb [Wed, 2 Jun 2004 19:15:43 +0000 (19:15 +0000)]
file Makefile.mk was initially added on branch b1_4_openib.

19 years agofile replay-sanity.sh was initially added on branch b1_4_smallfix.
rread [Wed, 2 Jun 2004 19:10:46 +0000 (19:10 +0000)]
file replay-sanity.sh was initially added on branch b1_4_smallfix.

19 years agolanding b_cmobd_merge on HEAD
phil [Wed, 2 Jun 2004 15:06:23 +0000 (15:06 +0000)]
landing b_cmobd_merge on HEAD

b_cmobd_merge should not be used for any further development or testing!

19 years ago- mds->lmv->mdc propagate lower timeout down to import
alex [Tue, 1 Jun 2004 15:19:34 +0000 (15:19 +0000)]
- mds->lmv->mdc propagate lower timeout down to import
  we need this becase client's timeout should be longer then inter-mds recovery may take
- few ugly hacks that allow client to reconnect with old (before failure took place) UUID
- mds_preprw() and mds_preprw() should use l_dput() instead of f_dput()
- SOCKNAL_IO_TIMEOUT has been set to 20 seconds to make discovery of
  stale connection faster
- lconf generates persisten UUIDs for LMV clients
  it each new LMV uses fresh UUID, then each recovered MDS looks like
  new client (target MDS doesn't recognize it's old because of new UUID).
  if target MDS gets restarted, then it'll find more clients in LAST_RCVD
  then it actually had

19 years agofile export-zap-page-range.patch was initially added on branch b_mmap.
zab [Sat, 29 May 2004 00:17:09 +0000 (00:17 +0000)]
file export-zap-page-range.patch was initially added on branch b_mmap.

19 years agofile mmap_sanity.c was initially added on branch b_mmap.
zab [Sat, 29 May 2004 00:13:20 +0000 (00:13 +0000)]
file mmap_sanity.c was initially added on branch b_mmap.

19 years agofile ext3-htree-rename_fix.patch was initially added on branch b1_2.
adilger [Wed, 26 May 2004 15:58:24 +0000 (15:58 +0000)]
file ext3-htree-rename_fix.patch was initially added on branch b1_2.

19 years agob=2950
alex [Tue, 25 May 2004 10:22:10 +0000 (10:22 +0000)]
b=2950

landed on HEAD

19 years agob=3063
alex [Tue, 25 May 2004 09:10:38 +0000 (09:10 +0000)]
b=3063

19 years ago- lmv_getattr_name() recognizes cross-node name and repeats getattr to proper
alex [Mon, 24 May 2004 18:39:04 +0000 (18:39 +0000)]
- lmv_getattr_name() recognizes cross-node name and repeats getattr to proper
  MDS node with gotten ino/generation couple
- warning for cross-node rename in lmv_rename(). it isn't implemented yet.
  hence, we'd like to know when we hit this path
- mdc_getattr_name() accepts empty name in order to support getattr by fid request
- mds_get_parent_child_locked() forces MDS_INODELOCK_LOOKUP to be used
  if found dentry points to another MDS node
- mea_name2idx() and raw_name2idx() checks for last not to be zero and warns if it's

NOTE: this should fix the cause of the test #5 failure

19 years ago- bugs in fsfilt_ext3_add_dir_entry() and fsfilt_ext3_del_dir_entryI() fixed:
alex [Mon, 24 May 2004 14:42:18 +0000 (14:42 +0000)]
- bugs in fsfilt_ext3_add_dir_entry() and fsfilt_ext3_del_dir_entryI() fixed:
  JBD routines must be called under BKL on 2.4

19 years ago- ll_update_inode() puts number of MDS holding inode to inode->i_dev. this
alex [Mon, 24 May 2004 08:57:47 +0000 (08:57 +0000)]
- ll_update_inode() puts number of MDS holding inode to inode->i_dev. this
  fix the problem some apps confuses on: they see the same inode number for
  different objects. thanks to Andreas for the idea.

19 years ago- mdt_obj_create() understands a request which is part of cross-node mkdir()
alex [Sun, 23 May 2004 16:22:08 +0000 (16:22 +0000)]
- mdt_obj_create() understands a request which is part of cross-node mkdir()
  recovery and looks for given ino/generation first. if no ino/generation are
  given then it simple recreates new object and returns ino/generation back
  to MDS holding a name. leaked inode is subject to rollback mechanism that
  is to be implemented yet
- when mds_reint_create() is asked to replay cross-node mkdir() it signals
  this to MDS holding dir inode and sends ino/generation couple if possible

NOTE: now cross-node mkdir() survives failure on any MDS node: holding the
      name and holding the inode

19 years ago- extN-wantedi accepts generation as well as ino to create an inode with
alex [Fri, 21 May 2004 17:16:44 +0000 (17:16 +0000)]
- extN-wantedi accepts generation as well as ino to create an inode with
  given ino/generation. we need this for cross-node mkdir() which creates
  inode on remote MDS and adds dir entry then. after that remote MDS may
  fail and primary MDS will have to recreate inode with already gotten
  ino and generation numbers
- export refcounting bug in lmv fixed
- mdc_obj_create() stores just received ino/generation back to request.
  if request replay will happen mdt_obj_create() will use stored numbers
  to recreate lost inode
- mds_filter_recovery_request() accepts OST_CREATE request. need to be
  discussed though
- mdt_obj_create() rewritten to support recovery phase
- needless mds_lmv_disconnect() in mds_fs_cleanup() removed
- possible bug in retrieve_generation_numbers() fixed: it should understand
  cross-node dir entries and put proper mdsnum's

NOTE: please, re-apply extN-wantedi patch

19 years ago- workaround for name collision in mdt_obj_create() + debug CERROR there
alex [Fri, 21 May 2004 08:43:35 +0000 (08:43 +0000)]
- workaround for name collision in mdt_obj_create() + debug CERROR there
- mds_commitrw() should use l_dput() instead of f_dput()
- removed needless CERROR() from mds_create_local_dentry()

19 years ago- pdirops patches against vanilla-2.4.20 series
alex [Thu, 20 May 2004 19:48:24 +0000 (19:48 +0000)]
- pdirops patches against vanilla-2.4.20 series

NOTE: vanilla-2.4.20 runs CMD now