Whamcloud - gitweb
nikita [Thu, 12 Oct 2006 23:29:32 +0000 (23:29 +0000)]
lu_object: add lu_object cache shrinking.
nikita [Thu, 12 Oct 2006 20:18:02 +0000 (20:18 +0000)]
ptlrpc: mark branches in hot path as unlikely()
nikita [Thu, 12 Oct 2006 20:16:03 +0000 (20:16 +0000)]
iam: further cpu consumtion fixes
nikita [Thu, 12 Oct 2006 20:05:47 +0000 (20:05 +0000)]
lu_object: cpu consumption reduction: increase hash table size. (Should be automatically tuned, in fact.)
nikita [Thu, 12 Oct 2006 20:05:02 +0000 (20:05 +0000)]
ptlrpc: mark branches in hot path as unlikely()
yury [Thu, 12 Oct 2006 18:37:46 +0000 (18:37 +0000)]
- do not allocate super-seq in controller assign time. Do that later when that is really needed.
- save defult value on controller in strt time.
yury [Thu, 12 Oct 2006 18:23:05 +0000 (18:23 +0000)]
- added ldiskfs debug configure keys.
ericm [Thu, 12 Oct 2006 17:59:27 +0000 (17:59 +0000)]
branch: b_new_cmd
- first part of b11012: mdc_realloc_openmsg() should not manipulate
request buffer directly, should be done in sptlrpc layer. implement
enlarge_msg() in null and plain policy.
- restore default to plain for rpcs to MDTs
fanyong [Thu, 12 Oct 2006 16:08:22 +0000 (16:08 +0000)]
Roll back for the change of server_put_mount.
Use server_get_mount_2 instead of server_get_mount.
lsy [Thu, 12 Oct 2006 15:57:27 +0000 (15:57 +0000)]
give friendly error message if crypto failed.
lsy [Thu, 12 Oct 2006 15:55:05 +0000 (15:55 +0000)]
ll_revalid_it will pack fid2, while its capability might expire yet,
but actually mds will use its name to locate object, so this message
might not indicate an error.
lsy [Thu, 12 Oct 2006 15:35:35 +0000 (15:35 +0000)]
* all capabilities of opened inode should always renew because many file
operations do from file->f_dentry->d_inode directly.
* all dir capabilities should always renew because ll_revalid_it needs
parent dir capability.
* filter update capability key more intelligently.
* rename ll_i2mdscapa -> ll_mdscapa_get, ll_lookup_oss_capa -> ll_osscapa_get.
* other code cleanup.
nikita [Thu, 12 Oct 2006 14:15:23 +0000 (14:15 +0000)]
iam,cmm,llite: cleanup handling of corner cases in readdir, especially for split directories. Get rid of -ERANGE and -E2BIG errors.
yury [Thu, 12 Oct 2006 14:11:48 +0000 (14:11 +0000)]
- zero out used locks in MDT before finishing req handling;
- fixes in seq-mgr recovery (meta and super);
- cleanups, more debug and comments, fixes in comments.
fanyong [Thu, 12 Oct 2006 13:32:17 +0000 (13:32 +0000)]
(1) Do not call server_deregister_mount unless ref conut is zero.
(2) Enable acl under default mode.
tappro [Thu, 12 Oct 2006 11:58:34 +0000 (11:58 +0000)]
- serialize done_writting with close requests
- use 'close' fields in last_rcvd also for done_writting
yury [Thu, 12 Oct 2006 10:31:49 +0000 (10:31 +0000)]
- using find in sanity-lmv again.
wangdi [Thu, 12 Oct 2006 08:47:07 +0000 (08:47 +0000)]
Branch: b_new_cmd
ma_lmm_size is maxisam of lov md size initially, which should be update
in req2attr update
tappro [Thu, 12 Oct 2006 08:03:58 +0000 (08:03 +0000)]
- put some asserts for debug
- remove unneeded debug message
wangdi [Thu, 12 Oct 2006 07:49:54 +0000 (07:49 +0000)]
Branch: b_new_cmd
Log:
oops, typo again, sorry :(
wangdi [Thu, 12 Oct 2006 07:47:06 +0000 (07:47 +0000)]
Branch: b_new_cmd
Log:
fix typo in the last commit
wangdi [Thu, 12 Oct 2006 07:40:01 +0000 (07:40 +0000)]
correct bit size for computing hash in client
huanghua [Thu, 12 Oct 2006 07:30:29 +0000 (07:30 +0000)]
fix return value for mdt_is_subdir().
huanghua [Thu, 12 Oct 2006 02:28:57 +0000 (02:28 +0000)]
passing valid argument
ericm [Wed, 11 Oct 2006 18:07:30 +0000 (18:07 +0000)]
branch: b_new_cmd
hack version of b11058.
wangdi [Wed, 11 Oct 2006 16:08:05 +0000 (16:08 +0000)]
Branch: b_new_cmd
use ma in cmm_thread_info, not allocate MA
tappro [Wed, 11 Oct 2006 13:22:37 +0000 (13:22 +0000)]
- fix the checking stale objects in mdt_getattr_name_lock
yury [Wed, 11 Oct 2006 11:22:10 +0000 (11:22 +0000)]
- remove check for obd_configured from target_handle_connect(), this should be fixed some another way.
yury [Wed, 11 Oct 2006 10:24:51 +0000 (10:24 +0000)]
- stop recovery thare in obdfilter before freeing namespace;
- cleanups in error messages and comments;
- more verbosity to class_config_notify_end();
- set obd->obd_configured = 0 in class_setup() before doing anything;
- do not pre-allocate seq-ranges in mdc. It usually will not do that asap
as remote target is not yet configured. It will be allocated later in very
first opeartion time, when you fid is needed;
- zero out lock handle in mdt_intent_lock_replace() to not confuse mdt_info_thread_fini();
- do not set obd->obd_configured if log processing finishes with error;
- do not allow gss connection if obd is not configured;
- disable FLD relay test 0c for multi MDS config.
fanyong [Wed, 11 Oct 2006 08:48:12 +0000 (08:48 +0000)]
Drop security ctxt when umount, for avoiding ldlm_revoke_export_locks
accessing freed resource.
wangdi [Wed, 11 Oct 2006 08:08:51 +0000 (08:08 +0000)]
Branch: b_new_cmd
fix the problem in reset hash segment
ericm [Wed, 11 Oct 2006 04:14:22 +0000 (04:14 +0000)]
branch: b_new_cmd
mdc_realloc_openmsg() is violating rules of sptlrpc, LBUG it for this
moment. real fixes shold come asap.
tappro [Wed, 11 Oct 2006 00:26:32 +0000 (00:26 +0000)]
remove '\n' from the end of the debug message
tappro [Wed, 11 Oct 2006 00:25:24 +0000 (00:25 +0000)]
add debug
tappro [Wed, 11 Oct 2006 00:24:47 +0000 (00:24 +0000)]
- fix test 17
ericm [Tue, 10 Oct 2006 17:23:59 +0000 (17:23 +0000)]
branch: b_new_cmd
"body might be used uninitialized"
ericm [Tue, 10 Oct 2006 17:22:50 +0000 (17:22 +0000)]
branch: b_new_cmd
port from b1_8_gss:
- wake waiting threads upon return enc pages to pools.
wangdi [Tue, 10 Oct 2006 15:26:47 +0000 (15:26 +0000)]
Branch: b_new_cmd
remove useless debug info
yury [Tue, 10 Oct 2006 15:17:20 +0000 (15:17 +0000)]
- reverse wrong commit about open_req.
wangdi [Tue, 10 Oct 2006 15:11:52 +0000 (15:11 +0000)]
Branch: b_new_cmd
2 fixes of split
1)reset hash_segment until we do not get -E2BIG
2)do not create . and .. for slave obj
yury [Tue, 10 Oct 2006 14:19:32 +0000 (14:19 +0000)]
- fixes from 1_5 (from green and bzzz).
yury [Tue, 10 Oct 2006 14:07:14 +0000 (14:07 +0000)]
- compile fixes in fld_rrb_scan()
yury [Tue, 10 Oct 2006 14:02:10 +0000 (14:02 +0000)]
- fixed missed braces in mdc_clear_open_replay_data()
yury [Tue, 10 Oct 2006 13:58:49 +0000 (13:58 +0000)]
- fixes and cleanup in seq-mgr about setting controller client. Return -ENODEV if controller client is not set but we want allocate new super. Caller should repeat req later;
- in fld_rrb_scan() added detailed dump of targets for case no target found by seq-built hash;
- in fld_client_add_target() detailed info about added target;
- in target_handle_connect() do not allow to connect if obd configuratoin is not yet finished. MGS obd is always counted as finished configuration and may accept connections from MGC;
- fixed leak of och in ll_close_inode_openhandle(). och should be marked DEAD also for cases !(och->och_flags & FMODE_WRITE);
- in lmv_clear_open_replay_data() handle possible error if no target found by fid;
- detailed debug in mdc_realloc_openmsg() about req realloc activities and attributes (sizes);
- in mdc_enqueue() do not realloc failed open reqs;
- in mdc_enqueue() set open replay data only in cases open req is not errorneous and will be replayed (req->rq_replay == 1);
- in mdc_enqueue() try to realloc open req for EA not only for reg files. Directories also may have stripe info (default stripe info for children files). Cleanups;
- in mdc_set_open_replay_data() add ref to open request saved to open replay data to prevent its freeing underneath what may confuse caller (mdc_enqueue() or llite);. Balance refc in mdc_clear_open_replay_data() which is called when open req is not longer needed to be preserved for replay (close is running or open req is errorneous);
- fixed typo in message in mdc_fid_init();
- in mdt fixes in comments, small cleanups;
- in test-framework.sh, find "evict_client" in mdd lprocfs (not imlemented yet).
huanghua [Tue, 10 Oct 2006 07:48:11 +0000 (07:48 +0000)]
to handle stripecount == -1 correctly
ericm [Tue, 10 Oct 2006 07:21:26 +0000 (07:21 +0000)]
branch: b_new_cmd
temporarily use null policy by default to MDTs, for debugging.
huanghua [Tue, 10 Oct 2006 06:44:17 +0000 (06:44 +0000)]
directory maybe have been deleted, and need to check this.
That is: check the existance of parent directory.
yujian [Tue, 10 Oct 2006 04:19:20 +0000 (04:19 +0000)]
Add stripe parameters to MDTs.
tappro [Mon, 9 Oct 2006 20:45:34 +0000 (20:45 +0000)]
- set rq_reqmsg to NULL in null_free_req with rq_reqbuf
- add debug info
tappro [Mon, 9 Oct 2006 20:42:21 +0000 (20:42 +0000)]
- workaround for cross-ref replay. The requested operation can be done during
original request, just do nothing in that case.
tappro [Mon, 9 Oct 2006 20:39:23 +0000 (20:39 +0000)]
- correct error handling
nikita [Mon, 9 Oct 2006 18:32:57 +0000 (18:32 +0000)]
- iam: assure that hash value is always even, even (heh) for dot and dotdot.
- lvar_root(), lvar_leaf(): don't assume that new node is zeroed.
nikita [Mon, 9 Oct 2006 18:31:52 +0000 (18:31 +0000)]
osd: remove outdated comments
nikita [Mon, 9 Oct 2006 16:04:50 +0000 (16:04 +0000)]
iam: lvar_root(): do not assume newly allocated block is zeroed out.
lsy [Mon, 9 Oct 2006 14:30:39 +0000 (14:30 +0000)]
remove redundant code.
yury [Mon, 9 Oct 2006 14:00:48 +0000 (14:00 +0000)]
- fixes about using lmv_get_export(), fixes possible memory leaks;
- fixed bug with split. If after split repeat intent_lock() return error we have to free req and zero it out to not confuse client.
nikita [Mon, 9 Oct 2006 13:28:02 +0000 (13:28 +0000)]
small cleanups
yury [Mon, 9 Oct 2006 12:21:43 +0000 (12:21 +0000)]
- fixed to sanity-lmv.sh, we can't use find for that due to "Value is too big"
huanghua [Mon, 9 Oct 2006 09:21:49 +0000 (09:21 +0000)]
more comments on why we need to set nlink to zero for a dir.
lsy [Mon, 9 Oct 2006 08:31:48 +0000 (08:31 +0000)]
oss capability fully works now.
add setattr capability check for oss (including req from mds).
code cleanup.
fanyong [Mon, 9 Oct 2006 06:30:40 +0000 (06:30 +0000)]
Small fix and make the code more clean.
huanghua [Mon, 9 Oct 2006 05:39:45 +0000 (05:39 +0000)]
add stripecount parameter for mdtlov to stripe all data across all OSTs
huanghua [Mon, 9 Oct 2006 03:58:35 +0000 (03:58 +0000)]
(1) correctly unlink a dir: we have to set i_nlink to zero manually.
(2) some debug code;
wangdi [Mon, 9 Oct 2006 03:56:34 +0000 (03:56 +0000)]
Branch: b_new_cmd
update lmv slave object size.
fanyong [Mon, 9 Oct 2006 02:40:16 +0000 (02:40 +0000)]
Small bug fix and make the code more conciser.
yury [Sun, 8 Oct 2006 17:23:09 +0000 (17:23 +0000)]
- Mike suggested to rely that all remote objects in split are dirs.
yury [Sun, 8 Oct 2006 17:10:28 +0000 (17:10 +0000)]
- fixes and cleanups. Tom notice my comment in cmm_remove_dir_ent().
yury [Sun, 8 Oct 2006 16:07:32 +0000 (16:07 +0000)]
- move GSS CONDITIONAL to correct m4 file.
tappro [Sun, 8 Oct 2006 14:41:08 +0000 (14:41 +0000)]
update time only if there was no error
tappro [Sun, 8 Oct 2006 14:21:14 +0000 (14:21 +0000)]
- set placement policy to balanced algorithm from stressed
fanyong [Sun, 8 Oct 2006 10:21:06 +0000 (10:21 +0000)]
The rootsquash info will be changed without notify client.
So the client will update perm actively. (sanity-sec test_3)
fanyong [Sun, 8 Oct 2006 07:36:00 +0000 (07:36 +0000)]
(1) Adding squash flag in md_ucred will increase performance of idmap.
(2) Fix bug about mdt_body_reverse_idmap for remote user with rootsquash.
yury [Sun, 8 Oct 2006 07:19:27 +0000 (07:19 +0000)]
- revised seq-mgr, a lot of simplifications.
wangdi [Sun, 8 Oct 2006 06:18:42 +0000 (06:18 +0000)]
Branch: b_new_cmd
rec_len is 0 for the last rec of each page, so it can not be used to
check rec valid
huanghua [Sun, 8 Oct 2006 03:56:41 +0000 (03:56 +0000)]
some code cleanup
wangdi [Sun, 8 Oct 2006 02:31:07 +0000 (02:31 +0000)]
Branch: b_new_cmd
enable split, in case someone did not consider split when commit some stuff.
Mountain's corruption is not related with split, since there is no split at all
currently.
huanghua [Sun, 8 Oct 2006 01:59:02 +0000 (01:59 +0000)]
remove compiler warning.
tappro [Sat, 7 Oct 2006 22:13:22 +0000 (22:13 +0000)]
- remove unused code
tappro [Sat, 7 Oct 2006 22:08:21 +0000 (22:08 +0000)]
- fast fix of mountconf bug when mds is set up before osc are added.
This issues the race between recovery can start but not all osc are added yet.
Fix adds obd_configured flag to track when config llog parsing is finished and MDS is really set up and ready for recovery.
yury [Sat, 7 Oct 2006 17:42:24 +0000 (17:42 +0000)]
- changes in comments.
ericm [Sat, 7 Oct 2006 17:27:42 +0000 (17:27 +0000)]
branch: b_new_cmd
more fixes of configuration of capability and gss.
yury [Sat, 7 Oct 2006 15:44:57 +0000 (15:44 +0000)]
- disable split by default for a now to make sure that corruptions on mountain are not related to it.
fanyong [Sat, 7 Oct 2006 15:42:10 +0000 (15:42 +0000)]
Interface name fix.
lsy [Sat, 7 Oct 2006 15:26:59 +0000 (15:26 +0000)]
remove \n from parameter strings, otherwise it will confuse later pattern match.
also rename rootsquash_skips to nosquash_nids.
fanyong [Sat, 7 Oct 2006 15:10:07 +0000 (15:10 +0000)]
Change with the changing of rootsquash interface.
tappro [Sat, 7 Oct 2006 15:05:31 +0000 (15:05 +0000)]
use mdc RPC lock to queue the seq/fld requests, except FLD_LOOKUP
lsy [Sat, 7 Oct 2006 14:05:09 +0000 (14:05 +0000)]
fix capa renewal reply offset.
class_process_proc_param() compare lvar more accuratly.
code cleanup.
lsy [Sat, 7 Oct 2006 10:23:51 +0000 (10:23 +0000)]
add capability config interface. also modified the interface for rootsquash.
fanyong [Sat, 7 Oct 2006 09:14:15 +0000 (09:14 +0000)]
(1) drop unnecessary ops.
(2) small fix for remote_perm.
yury [Sat, 7 Oct 2006 08:59:06 +0000 (08:59 +0000)]
- fixes in GSS configuration
fanyong [Sat, 7 Oct 2006 08:54:53 +0000 (08:54 +0000)]
Invalid assert when cal is enabled.
tappro [Sat, 7 Oct 2006 08:50:58 +0000 (08:50 +0000)]
- temporary change to see request debug instead of assert. This is to 11012
debug
fanyong [Sat, 7 Oct 2006 07:27:37 +0000 (07:27 +0000)]
Small fix for remote_perm.
fanyong [Sat, 7 Oct 2006 07:24:36 +0000 (07:24 +0000)]
Small fix for remote client.
ericm [Fri, 6 Oct 2006 22:41:10 +0000 (22:41 +0000)]
branch: b_new_cmd (lustre), HEAD (build)
hmac checking in lustre-core.m4 break kerneless building. move to
lustre-build-linux.m4 and use existing LB_LINUX_CONFIG.
ericm [Fri, 6 Oct 2006 21:49:27 +0000 (21:49 +0000)]
branch: b_new_cmd
port gss fixes from b1_8_gss:
- pre-alloc page pools for bulk write in encrypting mode.
- improve procfs for sptlrpc/gss.
- various other sptlrpc fixes.
ericm [Fri, 6 Oct 2006 20:40:43 +0000 (20:40 +0000)]
branch: b_new_cmd
minor fix for gss configure
ericm [Fri, 6 Oct 2006 19:40:52 +0000 (19:40 +0000)]
file sec_lproc.c was added on branch b_new_cmd on 2006-10-06 21:49:25 +0000
ericm [Fri, 6 Oct 2006 19:40:50 +0000 (19:40 +0000)]
file sec_bulk.c was added on branch b_new_cmd on 2006-10-06 21:49:25 +0000
brian [Fri, 6 Oct 2006 18:57:01 +0000 (18:57 +0000)]
Pointing to $LINUX/.config is bad. $LINUX_CONFIG is where the .config file
can be found it's not always == $LINUX/.config.